target-arm queue: this clears out a bunch of patches I'd sent over the last coupled of weeks that have now got reviewed. Mostly this is MPS2 device support improvements, put there is also more of the incremental work towards supporting AArch32 Hyp mode, a floating point bugfix, and the raspi framebuffer viewport support. thanks -- PMM The following changes since commit 5ccac548faf041ff5229a8e8342e3be14a34c8af: Merge remote-tracking branch 'remotes/cody/tags/block-pull-request' into staging (2018-08-23 17:35:48 +0100) are available in the Git repository at: https://git.linaro.org/people/pmaydell/qemu-arm.git tags/pull-target-arm-20180824 for you to fetch changes up to 30a719e3cb5c5367f3651eba8fa935634bfee286: hw/arm/mps2: Fix ID register errors on AN511 and AN385 (2018-08-24 10:22:44 +0100) ---------------------------------------------------------------- target-arm queue: * Fix rounding errors in scaling float-to-int and int-to-float operations * Connect virtualization-related IRQs and memory regions of GICv2 in boards that use Cortex-A7 or Cortex-A15 * Support taking exceptions to AArch32 Hyp mode * Clear CPSR.IL and CPSR.J on 32-bit exception entry (a minor bug fix that won't affect non-buggy guest code) * mps2-an505: Implement various missing devices: dual timer, watchdogs, counters in the FPGAIO registers, some missing ID/control registers, TrustZone Master Security Controllers, PL081 DMA controllers, PL022 SPI controllers * correct ID register values for mps2-an385, -an511, -an505 * fix some hardcoded tabs in untouched backwaters of the target/arm codebase * raspi: Refactor framebuffer property handling code and implement support for the virtual framebuffer/viewport ---------------------------------------------------------------- Peter Maydell (48): hw/intc/arm_gic: Make per-cpu GICH memory regions 0x200 bytes large hw/arm/vexpress: Connect VIRQ and VFIQ hw/arm/highbank: Connect VIRQ and VFIQ hw/arm/fsl-imx6ul: Connect VIRQ and VFIQ hw/arm/fsl-imx6ul: Connect VIRQ and VFIQ hw/cpu/a15mpcore: If CPU has EL2, enable it on the GIC and wire it up hw/arm/vexpress: Don't set info->secure_boot if CPU doesn't have EL3 hw/arm/vexpress: Add "virtualization" property controlling presence of EL2 target/arm: Implement RAZ/WI HACTLR2 target/arm: Implement AArch32 HCR and HCR2 target/arm: Factor out code for taking an AArch32 exception target/arm: Implement support for taking exceptions to Hyp mode target/arm: Clear CPSR.IL and CPSR.J on 32-bit exception entry hw/arm/boot: AArch32 kernels should be started in Hyp mode if available hw/misc/mps2-fpgaio: Implement 1Hz and 100Hz counters hw/misc/mps2-fpgaio: Implement PSCNTR and COUNTER hw/timer/cmsdk-apb-dualtimer: Implement CMSDK dual timer module hw/arm/iotkit: Wire up the dualtimer hw/arm/mps2: Wire up dual-timer in mps2-an385 and mps2-an511 hw/arm/iotkit: Wire up the watchdogs hw/arm/iotkit: Wire up the S32KTIMER hw/misc/iotkit-sysctl: Implement IoTKit system control element hw/misc/iotkit-sysinfo: Implement IoTKit system information block hw/misc/iotkit: Wire up the sysctl and sysinfo register blocks hw/misc/tz-msc: Model TrustZone Master Security Controller hw/misc/iotkit-secctl: Wire up registers for controlling MSCs hw/arm/iotkit: Wire up the lines for MSCs hw/arm/mps2-tz: Create PL081s and MSCs hw/ssi/pl022: Allow use as embedded-struct device hw/ssi/pl022: Set up reset function in class init hw/ssi/pl022: Don't directly call vmstate_register() hw/ssi/pl022: Use DeviceState::realize rather than SysBusDevice::init hw/ssi/pl022: Correct wrong value for PL022_INT_RT hw/ssi/pl022: Correct wrong DMACR and ICR handling hw/arm/mps2-tz: Instantiate SPI controllers hw/arm/mps2-tz: Fix MPS2 SCC config register values target/arm: Untabify translate.c target/arm: Untabify iwmmxt_helper.c target/arm: Remove a handful of stray tabs hw/misc/bcm2835_fb: Move config fields to their own struct hw/misc/bcm2835_property: Track fb settings using BCM2835FBConfig hw/display/bcm2835_fb: Drop unused size and pitch fields hw/display/bcm2835_fb: Reset resolution, etc correctly hw/display/bcm2835_fb: Abstract out calculation of pitch, size hw/display/bcm2835_fb: Fix handling of virtual framebuffer hw/display/bcm2835_fb: Validate config settings hw/display/bcm2835_fb: Validate bcm2835_fb_mbox_push() config hw/arm/mps2: Fix ID register errors on AN511 and AN385 Richard Henderson (4): softfloat: Add scaling int-to-float routines softfloat: Add scaling float-to-int routines target/arm: Use the int-to-float-scale softfloat routines target/arm: Use the float-to-int-scale softfloat routines hw/misc/Makefile.objs | 3 + hw/timer/Makefile.objs | 1 + include/fpu/softfloat.h | 169 +++++++--- include/hw/arm/iotkit.h | 25 +- include/hw/display/bcm2835_fb.h | 59 +++- include/hw/misc/iotkit-secctl.h | 14 + include/hw/misc/iotkit-sysctl.h | 49 +++ include/hw/misc/iotkit-sysinfo.h | 37 +++ include/hw/misc/mps2-fpgaio.h | 10 + include/hw/misc/tz-msc.h | 79 +++++ include/hw/ssi/pl022.h | 51 +++ include/hw/timer/cmsdk-apb-dualtimer.h | 72 ++++ target/arm/cpu.h | 16 +- fpu/softfloat.c | 579 ++++++++++++++++++++++++++------- hw/arm/boot.c | 11 + hw/arm/fsl-imx6ul.c | 4 + hw/arm/fsl-imx7.c | 4 + hw/arm/highbank.c | 6 + hw/arm/iotkit.c | 114 ++++++- hw/arm/mps2-tz.c | 142 +++++++- hw/arm/mps2.c | 17 +- hw/arm/vexpress.c | 64 +++- hw/cpu/a15mpcore.c | 31 +- hw/display/bcm2835_fb.c | 218 ++++++++----- hw/intc/arm_gic.c | 2 +- hw/misc/bcm2835_property.c | 123 ++++--- hw/misc/iotkit-secctl.c | 73 ++++- hw/misc/iotkit-sysctl.c | 261 +++++++++++++++ hw/misc/iotkit-sysinfo.c | 128 ++++++++ hw/misc/mps2-fpgaio.c | 146 ++++++++- hw/misc/tz-msc.c | 308 ++++++++++++++++++ hw/ssi/pl022.c | 57 ++-- hw/timer/cmsdk-apb-dualtimer.c | 515 +++++++++++++++++++++++++++++ target/arm/arm-semi.c | 2 +- target/arm/helper.c | 342 +++++++++++++------ target/arm/iwmmxt_helper.c | 234 ++++++------- target/arm/translate.c | 122 +++---- MAINTAINERS | 10 + default-configs/arm-softmmu.mak | 4 + hw/misc/trace-events | 16 + hw/timer/trace-events | 5 + 41 files changed, 3405 insertions(+), 718 deletions(-) create mode 100644 include/hw/misc/iotkit-sysctl.h create mode 100644 include/hw/misc/iotkit-sysinfo.h create mode 100644 include/hw/misc/tz-msc.h create mode 100644 include/hw/ssi/pl022.h create mode 100644 include/hw/timer/cmsdk-apb-dualtimer.h create mode 100644 hw/misc/iotkit-sysctl.c create mode 100644 hw/misc/iotkit-sysinfo.c create mode 100644 hw/misc/tz-msc.c create mode 100644 hw/timer/cmsdk-apb-dualtimer.c From: Richard Henderson <richard.henderson@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20180814002653.12828-2-richard.henderson@linaro.org Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- include/fpu/softfloat.h | 56 ++++++++---- fpu/softfloat.c | 188 +++++++++++++++++++++++++++++----------- 2 files changed, 179 insertions(+), 65 deletions(-) diff --git a/include/fpu/softfloat.h b/include/fpu/softfloat.h index XXXXXXX..XXXXXXX 100644 --- a/include/fpu/softfloat.h +++ b/include/fpu/softfloat.h @@ -XXX,XX +XXX,XX @@ enum { /*---------------------------------------------------------------------------- | Software IEC/IEEE integer-to-floating-point conversion routines. *----------------------------------------------------------------------------*/ + +float16 int16_to_float16_scalbn(int16_t a, int, float_status *status); +float16 int32_to_float16_scalbn(int32_t a, int, float_status *status); +float16 int64_to_float16_scalbn(int64_t a, int, float_status *status); +float16 uint16_to_float16_scalbn(uint16_t a, int, float_status *status); +float16 uint32_to_float16_scalbn(uint32_t a, int, float_status *status); +float16 uint64_to_float16_scalbn(uint64_t a, int, float_status *status); + +float16 int16_to_float16(int16_t a, float_status *status); +float16 int32_to_float16(int32_t a, float_status *status); +float16 int64_to_float16(int64_t a, float_status *status); +float16 uint16_to_float16(uint16_t a, float_status *status); +float16 uint32_to_float16(uint32_t a, float_status *status); +float16 uint64_to_float16(uint64_t a, float_status *status); + +float32 int16_to_float32_scalbn(int16_t, int, float_status *status); +float32 int32_to_float32_scalbn(int32_t, int, float_status *status); +float32 int64_to_float32_scalbn(int64_t, int, float_status *status); +float32 uint16_to_float32_scalbn(uint16_t, int, float_status *status); +float32 uint32_to_float32_scalbn(uint32_t, int, float_status *status); +float32 uint64_to_float32_scalbn(uint64_t, int, float_status *status); + float32 int16_to_float32(int16_t, float_status *status); float32 int32_to_float32(int32_t, float_status *status); -float64 int16_to_float64(int16_t, float_status *status); -float64 int32_to_float64(int32_t, float_status *status); +float32 int64_to_float32(int64_t, float_status *status); float32 uint16_to_float32(uint16_t, float_status *status); float32 uint32_to_float32(uint32_t, float_status *status); +float32 uint64_to_float32(uint64_t, float_status *status); + +float64 int16_to_float64_scalbn(int16_t, int, float_status *status); +float64 int32_to_float64_scalbn(int32_t, int, float_status *status); +float64 int64_to_float64_scalbn(int64_t, int, float_status *status); +float64 uint16_to_float64_scalbn(uint16_t, int, float_status *status); +float64 uint32_to_float64_scalbn(uint32_t, int, float_status *status); +float64 uint64_to_float64_scalbn(uint64_t, int, float_status *status); + +float64 int16_to_float64(int16_t, float_status *status); +float64 int32_to_float64(int32_t, float_status *status); +float64 int64_to_float64(int64_t, float_status *status); float64 uint16_to_float64(uint16_t, float_status *status); float64 uint32_to_float64(uint32_t, float_status *status); -floatx80 int32_to_floatx80(int32_t, float_status *status); -float128 int32_to_float128(int32_t, float_status *status); -float32 int64_to_float32(int64_t, float_status *status); -float64 int64_to_float64(int64_t, float_status *status); -floatx80 int64_to_floatx80(int64_t, float_status *status); -float128 int64_to_float128(int64_t, float_status *status); -float32 uint64_to_float32(uint64_t, float_status *status); float64 uint64_to_float64(uint64_t, float_status *status); + +floatx80 int32_to_floatx80(int32_t, float_status *status); +floatx80 int64_to_floatx80(int64_t, float_status *status); + +float128 int32_to_float128(int32_t, float_status *status); +float128 int64_to_float128(int64_t, float_status *status); float128 uint64_to_float128(uint64_t, float_status *status); /*---------------------------------------------------------------------------- @@ -XXX,XX +XXX,XX @@ int64_t float16_to_int64(float16, float_status *status); uint64_t float16_to_uint64(float16 a, float_status *status); int64_t float16_to_int64_round_to_zero(float16, float_status *status); uint64_t float16_to_uint64_round_to_zero(float16 a, float_status *status); -float16 int16_to_float16(int16_t a, float_status *status); -float16 int32_to_float16(int32_t a, float_status *status); -float16 int64_to_float16(int64_t a, float_status *status); -float16 uint16_to_float16(uint16_t a, float_status *status); -float16 uint32_to_float16(uint32_t a, float_status *status); -float16 uint64_to_float16(uint64_t a, float_status *status); /*---------------------------------------------------------------------------- | Software half-precision operations. diff --git a/fpu/softfloat.c b/fpu/softfloat.c index XXXXXXX..XXXXXXX 100644 --- a/fpu/softfloat.c +++ b/fpu/softfloat.c @@ -XXX,XX +XXX,XX @@ FLOAT_TO_UINT(64, 64) * to the IEC/IEEE Standard for Binary Floating-Point Arithmetic. */ -static FloatParts int_to_float(int64_t a, float_status *status) +static FloatParts int_to_float(int64_t a, int scale, float_status *status) { - FloatParts r = {}; + FloatParts r = { .sign = false }; + if (a == 0) { r.cls = float_class_zero; - r.sign = false; - } else if (a == (1ULL << 63)) { - r.cls = float_class_normal; - r.sign = true; - r.frac = DECOMPOSED_IMPLICIT_BIT; - r.exp = 63; } else { - uint64_t f; - if (a < 0) { - f = -a; - r.sign = true; - } else { - f = a; - r.sign = false; - } - int shift = clz64(f) - 1; + uint64_t f = a; + int shift; + r.cls = float_class_normal; - r.exp = (DECOMPOSED_BINARY_POINT - shift); - r.frac = f << shift; + if (a < 0) { + f = -f; + r.sign = true; + } + shift = clz64(f) - 1; + scale = MIN(MAX(scale, -0x10000), 0x10000); + + r.exp = DECOMPOSED_BINARY_POINT - shift + scale; + r.frac = (shift < 0 ? DECOMPOSED_IMPLICIT_BIT : f << shift); } return r; } +float16 int64_to_float16_scalbn(int64_t a, int scale, float_status *status) +{ + FloatParts pa = int_to_float(a, scale, status); + return float16_round_pack_canonical(pa, status); +} + +float16 int32_to_float16_scalbn(int32_t a, int scale, float_status *status) +{ + return int64_to_float16_scalbn(a, scale, status); +} + +float16 int16_to_float16_scalbn(int16_t a, int scale, float_status *status) +{ + return int64_to_float16_scalbn(a, scale, status); +} + float16 int64_to_float16(int64_t a, float_status *status) { - FloatParts pa = int_to_float(a, status); - return float16_round_pack_canonical(pa, status); + return int64_to_float16_scalbn(a, 0, status); } float16 int32_to_float16(int32_t a, float_status *status) { - return int64_to_float16(a, status); + return int64_to_float16_scalbn(a, 0, status); } float16 int16_to_float16(int16_t a, float_status *status) { - return int64_to_float16(a, status); + return int64_to_float16_scalbn(a, 0, status); +} + +float32 int64_to_float32_scalbn(int64_t a, int scale, float_status *status) +{ + FloatParts pa = int_to_float(a, scale, status); + return float32_round_pack_canonical(pa, status); +} + +float32 int32_to_float32_scalbn(int32_t a, int scale, float_status *status) +{ + return int64_to_float32_scalbn(a, scale, status); +} + +float32 int16_to_float32_scalbn(int16_t a, int scale, float_status *status) +{ + return int64_to_float32_scalbn(a, scale, status); } float32 int64_to_float32(int64_t a, float_status *status) { - FloatParts pa = int_to_float(a, status); - return float32_round_pack_canonical(pa, status); + return int64_to_float32_scalbn(a, 0, status); } float32 int32_to_float32(int32_t a, float_status *status) { - return int64_to_float32(a, status); + return int64_to_float32_scalbn(a, 0, status); } float32 int16_to_float32(int16_t a, float_status *status) { - return int64_to_float32(a, status); + return int64_to_float32_scalbn(a, 0, status); +} + +float64 int64_to_float64_scalbn(int64_t a, int scale, float_status *status) +{ + FloatParts pa = int_to_float(a, scale, status); + return float64_round_pack_canonical(pa, status); +} + +float64 int32_to_float64_scalbn(int32_t a, int scale, float_status *status) +{ + return int64_to_float64_scalbn(a, scale, status); +} + +float64 int16_to_float64_scalbn(int16_t a, int scale, float_status *status) +{ + return int64_to_float64_scalbn(a, scale, status); } float64 int64_to_float64(int64_t a, float_status *status) { - FloatParts pa = int_to_float(a, status); - return float64_round_pack_canonical(pa, status); + return int64_to_float64_scalbn(a, 0, status); } float64 int32_to_float64(int32_t a, float_status *status) { - return int64_to_float64(a, status); + return int64_to_float64_scalbn(a, 0, status); } float64 int16_to_float64(int16_t a, float_status *status) { - return int64_to_float64(a, status); + return int64_to_float64_scalbn(a, 0, status); } @@ -XXX,XX +XXX,XX @@ float64 int16_to_float64(int16_t a, float_status *status) * IEC/IEEE Standard for Binary Floating-Point Arithmetic. */ -static FloatParts uint_to_float(uint64_t a, float_status *status) +static FloatParts uint_to_float(uint64_t a, int scale, float_status *status) { - FloatParts r = { .sign = false}; + FloatParts r = { .sign = false }; if (a == 0) { r.cls = float_class_zero; } else { - int spare_bits = clz64(a) - 1; + scale = MIN(MAX(scale, -0x10000), 0x10000); r.cls = float_class_normal; - r.exp = DECOMPOSED_BINARY_POINT - spare_bits; - if (spare_bits < 0) { - shift64RightJamming(a, -spare_bits, &a); + if ((int64_t)a < 0) { + r.exp = DECOMPOSED_BINARY_POINT + 1 + scale; + shift64RightJamming(a, 1, &a); r.frac = a; } else { - r.frac = a << spare_bits; + int shift = clz64(a) - 1; + r.exp = DECOMPOSED_BINARY_POINT - shift + scale; + r.frac = a << shift; } } return r; } +float16 uint64_to_float16_scalbn(uint64_t a, int scale, float_status *status) +{ + FloatParts pa = uint_to_float(a, scale, status); + return float16_round_pack_canonical(pa, status); +} + +float16 uint32_to_float16_scalbn(uint32_t a, int scale, float_status *status) +{ + return uint64_to_float16_scalbn(a, scale, status); +} + +float16 uint16_to_float16_scalbn(uint16_t a, int scale, float_status *status) +{ + return uint64_to_float16_scalbn(a, scale, status); +} + float16 uint64_to_float16(uint64_t a, float_status *status) { - FloatParts pa = uint_to_float(a, status); - return float16_round_pack_canonical(pa, status); + return uint64_to_float16_scalbn(a, 0, status); } float16 uint32_to_float16(uint32_t a, float_status *status) { - return uint64_to_float16(a, status); + return uint64_to_float16_scalbn(a, 0, status); } float16 uint16_to_float16(uint16_t a, float_status *status) { - return uint64_to_float16(a, status); + return uint64_to_float16_scalbn(a, 0, status); +} + +float32 uint64_to_float32_scalbn(uint64_t a, int scale, float_status *status) +{ + FloatParts pa = uint_to_float(a, scale, status); + return float32_round_pack_canonical(pa, status); +} + +float32 uint32_to_float32_scalbn(uint32_t a, int scale, float_status *status) +{ + return uint64_to_float32_scalbn(a, scale, status); +} + +float32 uint16_to_float32_scalbn(uint16_t a, int scale, float_status *status) +{ + return uint64_to_float32_scalbn(a, scale, status); } float32 uint64_to_float32(uint64_t a, float_status *status) { - FloatParts pa = uint_to_float(a, status); - return float32_round_pack_canonical(pa, status); + return uint64_to_float32_scalbn(a, 0, status); } float32 uint32_to_float32(uint32_t a, float_status *status) { - return uint64_to_float32(a, status); + return uint64_to_float32_scalbn(a, 0, status); } float32 uint16_to_float32(uint16_t a, float_status *status) { - return uint64_to_float32(a, status); + return uint64_to_float32_scalbn(a, 0, status); +} + +float64 uint64_to_float64_scalbn(uint64_t a, int scale, float_status *status) +{ + FloatParts pa = uint_to_float(a, scale, status); + return float64_round_pack_canonical(pa, status); +} + +float64 uint32_to_float64_scalbn(uint32_t a, int scale, float_status *status) +{ + return uint64_to_float64_scalbn(a, scale, status); +} + +float64 uint16_to_float64_scalbn(uint16_t a, int scale, float_status *status) +{ + return uint64_to_float64_scalbn(a, scale, status); } float64 uint64_to_float64(uint64_t a, float_status *status) { - FloatParts pa = uint_to_float(a, status); - return float64_round_pack_canonical(pa, status); + return uint64_to_float64_scalbn(a, 0, status); } float64 uint32_to_float64(uint32_t a, float_status *status) { - return uint64_to_float64(a, status); + return uint64_to_float64_scalbn(a, 0, status); } float64 uint16_to_float64(uint16_t a, float_status *status) { - return uint64_to_float64(a, status); + return uint64_to_float64_scalbn(a, 0, status); } /* Float Min/Max */ -- 2.18.0 From: Richard Henderson <richard.henderson@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20180814002653.12828-3-richard.henderson@linaro.org Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- include/fpu/softfloat.h | 85 ++++++--- fpu/softfloat.c | 391 ++++++++++++++++++++++++++++++++-------- 2 files changed, 379 insertions(+), 97 deletions(-) diff --git a/include/fpu/softfloat.h b/include/fpu/softfloat.h index XXXXXXX..XXXXXXX 100644 --- a/include/fpu/softfloat.h +++ b/include/fpu/softfloat.h @@ -XXX,XX +XXX,XX @@ float128 uint64_to_float128(uint64_t, float_status *status); /*---------------------------------------------------------------------------- | Software half-precision conversion routines. *----------------------------------------------------------------------------*/ + float16 float32_to_float16(float32, bool ieee, float_status *status); float32 float16_to_float32(float16, bool ieee, float_status *status); float16 float64_to_float16(float64 a, bool ieee, float_status *status); float64 float16_to_float64(float16 a, bool ieee, float_status *status); + +int16_t float16_to_int16_scalbn(float16, int, int, float_status *status); +int32_t float16_to_int32_scalbn(float16, int, int, float_status *status); +int64_t float16_to_int64_scalbn(float16, int, int, float_status *status); + int16_t float16_to_int16(float16, float_status *status); -uint16_t float16_to_uint16(float16 a, float_status *status); -int16_t float16_to_int16_round_to_zero(float16, float_status *status); -uint16_t float16_to_uint16_round_to_zero(float16 a, float_status *status); int32_t float16_to_int32(float16, float_status *status); -uint32_t float16_to_uint32(float16 a, float_status *status); -int32_t float16_to_int32_round_to_zero(float16, float_status *status); -uint32_t float16_to_uint32_round_to_zero(float16 a, float_status *status); int64_t float16_to_int64(float16, float_status *status); -uint64_t float16_to_uint64(float16 a, float_status *status); + +int16_t float16_to_int16_round_to_zero(float16, float_status *status); +int32_t float16_to_int32_round_to_zero(float16, float_status *status); int64_t float16_to_int64_round_to_zero(float16, float_status *status); + +uint16_t float16_to_uint16_scalbn(float16 a, int, int, float_status *status); +uint32_t float16_to_uint32_scalbn(float16 a, int, int, float_status *status); +uint64_t float16_to_uint64_scalbn(float16 a, int, int, float_status *status); + +uint16_t float16_to_uint16(float16 a, float_status *status); +uint32_t float16_to_uint32(float16 a, float_status *status); +uint64_t float16_to_uint64(float16 a, float_status *status); + +uint16_t float16_to_uint16_round_to_zero(float16 a, float_status *status); +uint32_t float16_to_uint32_round_to_zero(float16 a, float_status *status); uint64_t float16_to_uint64_round_to_zero(float16 a, float_status *status); /*---------------------------------------------------------------------------- @@ -XXX,XX +XXX,XX @@ float16 float16_default_nan(float_status *status); /*---------------------------------------------------------------------------- | Software IEC/IEEE single-precision conversion routines. *----------------------------------------------------------------------------*/ + +int16_t float32_to_int16_scalbn(float32, int, int, float_status *status); +int32_t float32_to_int32_scalbn(float32, int, int, float_status *status); +int64_t float32_to_int64_scalbn(float32, int, int, float_status *status); + int16_t float32_to_int16(float32, float_status *status); -uint16_t float32_to_uint16(float32, float_status *status); -int16_t float32_to_int16_round_to_zero(float32, float_status *status); -uint16_t float32_to_uint16_round_to_zero(float32, float_status *status); int32_t float32_to_int32(float32, float_status *status); -int32_t float32_to_int32_round_to_zero(float32, float_status *status); -uint32_t float32_to_uint32(float32, float_status *status); -uint32_t float32_to_uint32_round_to_zero(float32, float_status *status); int64_t float32_to_int64(float32, float_status *status); -uint64_t float32_to_uint64(float32, float_status *status); -uint64_t float32_to_uint64_round_to_zero(float32, float_status *status); + +int16_t float32_to_int16_round_to_zero(float32, float_status *status); +int32_t float32_to_int32_round_to_zero(float32, float_status *status); int64_t float32_to_int64_round_to_zero(float32, float_status *status); + +uint16_t float32_to_uint16_scalbn(float32, int, int, float_status *status); +uint32_t float32_to_uint32_scalbn(float32, int, int, float_status *status); +uint64_t float32_to_uint64_scalbn(float32, int, int, float_status *status); + +uint16_t float32_to_uint16(float32, float_status *status); +uint32_t float32_to_uint32(float32, float_status *status); +uint64_t float32_to_uint64(float32, float_status *status); + +uint16_t float32_to_uint16_round_to_zero(float32, float_status *status); +uint32_t float32_to_uint32_round_to_zero(float32, float_status *status); +uint64_t float32_to_uint64_round_to_zero(float32, float_status *status); + float64 float32_to_float64(float32, float_status *status); floatx80 float32_to_floatx80(float32, float_status *status); float128 float32_to_float128(float32, float_status *status); @@ -XXX,XX +XXX,XX @@ float32 float32_default_nan(float_status *status); /*---------------------------------------------------------------------------- | Software IEC/IEEE double-precision conversion routines. *----------------------------------------------------------------------------*/ + +int16_t float64_to_int16_scalbn(float64, int, int, float_status *status); +int32_t float64_to_int32_scalbn(float64, int, int, float_status *status); +int64_t float64_to_int64_scalbn(float64, int, int, float_status *status); + int16_t float64_to_int16(float64, float_status *status); -uint16_t float64_to_uint16(float64, float_status *status); -int16_t float64_to_int16_round_to_zero(float64, float_status *status); -uint16_t float64_to_uint16_round_to_zero(float64, float_status *status); int32_t float64_to_int32(float64, float_status *status); -int32_t float64_to_int32_round_to_zero(float64, float_status *status); -uint32_t float64_to_uint32(float64, float_status *status); -uint32_t float64_to_uint32_round_to_zero(float64, float_status *status); int64_t float64_to_int64(float64, float_status *status); + +int16_t float64_to_int16_round_to_zero(float64, float_status *status); +int32_t float64_to_int32_round_to_zero(float64, float_status *status); int64_t float64_to_int64_round_to_zero(float64, float_status *status); -uint64_t float64_to_uint64(float64 a, float_status *status); -uint64_t float64_to_uint64_round_to_zero(float64 a, float_status *status); + +uint16_t float64_to_uint16_scalbn(float64, int, int, float_status *status); +uint32_t float64_to_uint32_scalbn(float64, int, int, float_status *status); +uint64_t float64_to_uint64_scalbn(float64, int, int, float_status *status); + +uint16_t float64_to_uint16(float64, float_status *status); +uint32_t float64_to_uint32(float64, float_status *status); +uint64_t float64_to_uint64(float64, float_status *status); + +uint16_t float64_to_uint16_round_to_zero(float64, float_status *status); +uint32_t float64_to_uint32_round_to_zero(float64, float_status *status); +uint64_t float64_to_uint64_round_to_zero(float64, float_status *status); + float32 float64_to_float32(float64, float_status *status); floatx80 float64_to_floatx80(float64, float_status *status); float128 float64_to_float128(float64, float_status *status); diff --git a/fpu/softfloat.c b/fpu/softfloat.c index XXXXXXX..XXXXXXX 100644 --- a/fpu/softfloat.c +++ b/fpu/softfloat.c @@ -XXX,XX +XXX,XX @@ float32 float64_to_float32(float64 a, float_status *s) * Arithmetic. */ -static FloatParts round_to_int(FloatParts a, int rounding_mode, float_status *s) +static FloatParts round_to_int(FloatParts a, int rmode, + int scale, float_status *s) { - if (is_nan(a.cls)) { - return return_nan(a, s); - } - switch (a.cls) { + case float_class_qnan: + case float_class_snan: + return return_nan(a, s); + case float_class_zero: case float_class_inf: - case float_class_qnan: /* already "integral" */ break; + case float_class_normal: + scale = MIN(MAX(scale, -0x10000), 0x10000); + a.exp += scale; + if (a.exp >= DECOMPOSED_BINARY_POINT) { /* already integral */ break; @@ -XXX,XX +XXX,XX @@ static FloatParts round_to_int(FloatParts a, int rounding_mode, float_status *s) bool one; /* all fractional */ s->float_exception_flags |= float_flag_inexact; - switch (rounding_mode) { + switch (rmode) { case float_round_nearest_even: one = a.exp == -1 && a.frac > DECOMPOSED_IMPLICIT_BIT; break; @@ -XXX,XX +XXX,XX @@ static FloatParts round_to_int(FloatParts a, int rounding_mode, float_status *s) uint64_t rnd_mask = rnd_even_mask >> 1; uint64_t inc; - switch (rounding_mode) { + switch (rmode) { case float_round_nearest_even: inc = ((a.frac & rnd_even_mask) != frac_lsbm1 ? frac_lsbm1 : 0); break; @@ -XXX,XX +XXX,XX @@ static FloatParts round_to_int(FloatParts a, int rounding_mode, float_status *s) float16 float16_round_to_int(float16 a, float_status *s) { FloatParts pa = float16_unpack_canonical(a, s); - FloatParts pr = round_to_int(pa, s->float_rounding_mode, s); + FloatParts pr = round_to_int(pa, s->float_rounding_mode, 0, s); return float16_round_pack_canonical(pr, s); } float32 float32_round_to_int(float32 a, float_status *s) { FloatParts pa = float32_unpack_canonical(a, s); - FloatParts pr = round_to_int(pa, s->float_rounding_mode, s); + FloatParts pr = round_to_int(pa, s->float_rounding_mode, 0, s); return float32_round_pack_canonical(pr, s); } float64 float64_round_to_int(float64 a, float_status *s) { FloatParts pa = float64_unpack_canonical(a, s); - FloatParts pr = round_to_int(pa, s->float_rounding_mode, s); + FloatParts pr = round_to_int(pa, s->float_rounding_mode, 0, s); return float64_round_pack_canonical(pr, s); } float64 float64_trunc_to_int(float64 a, float_status *s) { FloatParts pa = float64_unpack_canonical(a, s); - FloatParts pr = round_to_int(pa, float_round_to_zero, s); + FloatParts pr = round_to_int(pa, float_round_to_zero, 0, s); return float64_round_pack_canonical(pr, s); } @@ -XXX,XX +XXX,XX @@ float64 float64_trunc_to_int(float64 a, float_status *s) * is returned. */ -static int64_t round_to_int_and_pack(FloatParts in, int rmode, +static int64_t round_to_int_and_pack(FloatParts in, int rmode, int scale, int64_t min, int64_t max, float_status *s) { uint64_t r; int orig_flags = get_float_exception_flags(s); - FloatParts p = round_to_int(in, rmode, s); + FloatParts p = round_to_int(in, rmode, scale, s); switch (p.cls) { case float_class_snan: @@ -XXX,XX +XXX,XX @@ static int64_t round_to_int_and_pack(FloatParts in, int rmode, } } -#define FLOAT_TO_INT(fsz, isz) \ -int ## isz ## _t float ## fsz ## _to_int ## isz(float ## fsz a, \ - float_status *s) \ -{ \ - FloatParts p = float ## fsz ## _unpack_canonical(a, s); \ - return round_to_int_and_pack(p, s->float_rounding_mode, \ - INT ## isz ## _MIN, INT ## isz ## _MAX,\ - s); \ -} \ - \ -int ## isz ## _t float ## fsz ## _to_int ## isz ## _round_to_zero \ - (float ## fsz a, float_status *s) \ -{ \ - FloatParts p = float ## fsz ## _unpack_canonical(a, s); \ - return round_to_int_and_pack(p, float_round_to_zero, \ - INT ## isz ## _MIN, INT ## isz ## _MAX,\ - s); \ +int16_t float16_to_int16_scalbn(float16 a, int rmode, int scale, + float_status *s) +{ + return round_to_int_and_pack(float16_unpack_canonical(a, s), + rmode, scale, INT16_MIN, INT16_MAX, s); } -FLOAT_TO_INT(16, 16) -FLOAT_TO_INT(16, 32) -FLOAT_TO_INT(16, 64) +int32_t float16_to_int32_scalbn(float16 a, int rmode, int scale, + float_status *s) +{ + return round_to_int_and_pack(float16_unpack_canonical(a, s), + rmode, scale, INT32_MIN, INT32_MAX, s); +} -FLOAT_TO_INT(32, 16) -FLOAT_TO_INT(32, 32) -FLOAT_TO_INT(32, 64) +int64_t float16_to_int64_scalbn(float16 a, int rmode, int scale, + float_status *s) +{ + return round_to_int_and_pack(float16_unpack_canonical(a, s), + rmode, scale, INT64_MIN, INT64_MAX, s); +} -FLOAT_TO_INT(64, 16) -FLOAT_TO_INT(64, 32) -FLOAT_TO_INT(64, 64) +int16_t float32_to_int16_scalbn(float32 a, int rmode, int scale, + float_status *s) +{ + return round_to_int_and_pack(float32_unpack_canonical(a, s), + rmode, scale, INT16_MIN, INT16_MAX, s); +} -#undef FLOAT_TO_INT +int32_t float32_to_int32_scalbn(float32 a, int rmode, int scale, + float_status *s) +{ + return round_to_int_and_pack(float32_unpack_canonical(a, s), + rmode, scale, INT32_MIN, INT32_MAX, s); +} + +int64_t float32_to_int64_scalbn(float32 a, int rmode, int scale, + float_status *s) +{ + return round_to_int_and_pack(float32_unpack_canonical(a, s), + rmode, scale, INT64_MIN, INT64_MAX, s); +} + +int16_t float64_to_int16_scalbn(float64 a, int rmode, int scale, + float_status *s) +{ + return round_to_int_and_pack(float64_unpack_canonical(a, s), + rmode, scale, INT16_MIN, INT16_MAX, s); +} + +int32_t float64_to_int32_scalbn(float64 a, int rmode, int scale, + float_status *s) +{ + return round_to_int_and_pack(float64_unpack_canonical(a, s), + rmode, scale, INT32_MIN, INT32_MAX, s); +} + +int64_t float64_to_int64_scalbn(float64 a, int rmode, int scale, + float_status *s) +{ + return round_to_int_and_pack(float64_unpack_canonical(a, s), + rmode, scale, INT64_MIN, INT64_MAX, s); +} + +int16_t float16_to_int16(float16 a, float_status *s) +{ + return float16_to_int16_scalbn(a, s->float_rounding_mode, 0, s); +} + +int32_t float16_to_int32(float16 a, float_status *s) +{ + return float16_to_int32_scalbn(a, s->float_rounding_mode, 0, s); +} + +int64_t float16_to_int64(float16 a, float_status *s) +{ + return float16_to_int64_scalbn(a, s->float_rounding_mode, 0, s); +} + +int16_t float32_to_int16(float32 a, float_status *s) +{ + return float32_to_int16_scalbn(a, s->float_rounding_mode, 0, s); +} + +int32_t float32_to_int32(float32 a, float_status *s) +{ + return float32_to_int32_scalbn(a, s->float_rounding_mode, 0, s); +} + +int64_t float32_to_int64(float32 a, float_status *s) +{ + return float32_to_int64_scalbn(a, s->float_rounding_mode, 0, s); +} + +int16_t float64_to_int16(float64 a, float_status *s) +{ + return float64_to_int16_scalbn(a, s->float_rounding_mode, 0, s); +} + +int32_t float64_to_int32(float64 a, float_status *s) +{ + return float64_to_int32_scalbn(a, s->float_rounding_mode, 0, s); +} + +int64_t float64_to_int64(float64 a, float_status *s) +{ + return float64_to_int64_scalbn(a, s->float_rounding_mode, 0, s); +} + +int16_t float16_to_int16_round_to_zero(float16 a, float_status *s) +{ + return float16_to_int16_scalbn(a, float_round_to_zero, 0, s); +} + +int32_t float16_to_int32_round_to_zero(float16 a, float_status *s) +{ + return float16_to_int32_scalbn(a, float_round_to_zero, 0, s); +} + +int64_t float16_to_int64_round_to_zero(float16 a, float_status *s) +{ + return float16_to_int64_scalbn(a, float_round_to_zero, 0, s); +} + +int16_t float32_to_int16_round_to_zero(float32 a, float_status *s) +{ + return float32_to_int16_scalbn(a, float_round_to_zero, 0, s); +} + +int32_t float32_to_int32_round_to_zero(float32 a, float_status *s) +{ + return float32_to_int32_scalbn(a, float_round_to_zero, 0, s); +} + +int64_t float32_to_int64_round_to_zero(float32 a, float_status *s) +{ + return float32_to_int64_scalbn(a, float_round_to_zero, 0, s); +} + +int16_t float64_to_int16_round_to_zero(float64 a, float_status *s) +{ + return float64_to_int16_scalbn(a, float_round_to_zero, 0, s); +} + +int32_t float64_to_int32_round_to_zero(float64 a, float_status *s) +{ + return float64_to_int32_scalbn(a, float_round_to_zero, 0, s); +} + +int64_t float64_to_int64_round_to_zero(float64 a, float_status *s) +{ + return float64_to_int64_scalbn(a, float_round_to_zero, 0, s); +} /* * Returns the result of converting the floating-point value `a' to @@ -XXX,XX +XXX,XX @@ FLOAT_TO_INT(64, 64) * flag. */ -static uint64_t round_to_uint_and_pack(FloatParts in, int rmode, uint64_t max, - float_status *s) +static uint64_t round_to_uint_and_pack(FloatParts in, int rmode, int scale, + uint64_t max, float_status *s) { int orig_flags = get_float_exception_flags(s); - FloatParts p = round_to_int(in, rmode, s); + FloatParts p = round_to_int(in, rmode, scale, s); + uint64_t r; switch (p.cls) { case float_class_snan: @@ -XXX,XX +XXX,XX @@ static uint64_t round_to_uint_and_pack(FloatParts in, int rmode, uint64_t max, case float_class_zero: return 0; case float_class_normal: - { - uint64_t r; if (p.sign) { s->float_exception_flags = orig_flags | float_flag_invalid; return 0; @@ -XXX,XX +XXX,XX @@ static uint64_t round_to_uint_and_pack(FloatParts in, int rmode, uint64_t max, if (r > max) { s->float_exception_flags = orig_flags | float_flag_invalid; return max; - } else { - return r; } - } + return r; default: g_assert_not_reached(); } } -#define FLOAT_TO_UINT(fsz, isz) \ -uint ## isz ## _t float ## fsz ## _to_uint ## isz(float ## fsz a, \ - float_status *s) \ -{ \ - FloatParts p = float ## fsz ## _unpack_canonical(a, s); \ - return round_to_uint_and_pack(p, s->float_rounding_mode, \ - UINT ## isz ## _MAX, s); \ -} \ - \ -uint ## isz ## _t float ## fsz ## _to_uint ## isz ## _round_to_zero \ - (float ## fsz a, float_status *s) \ -{ \ - FloatParts p = float ## fsz ## _unpack_canonical(a, s); \ - return round_to_uint_and_pack(p, float_round_to_zero, \ - UINT ## isz ## _MAX, s); \ +uint16_t float16_to_uint16_scalbn(float16 a, int rmode, int scale, + float_status *s) +{ + return round_to_uint_and_pack(float16_unpack_canonical(a, s), + rmode, scale, UINT16_MAX, s); } -FLOAT_TO_UINT(16, 16) -FLOAT_TO_UINT(16, 32) -FLOAT_TO_UINT(16, 64) +uint32_t float16_to_uint32_scalbn(float16 a, int rmode, int scale, + float_status *s) +{ + return round_to_uint_and_pack(float16_unpack_canonical(a, s), + rmode, scale, UINT32_MAX, s); +} -FLOAT_TO_UINT(32, 16) -FLOAT_TO_UINT(32, 32) -FLOAT_TO_UINT(32, 64) +uint64_t float16_to_uint64_scalbn(float16 a, int rmode, int scale, + float_status *s) +{ + return round_to_uint_and_pack(float16_unpack_canonical(a, s), + rmode, scale, UINT64_MAX, s); +} -FLOAT_TO_UINT(64, 16) -FLOAT_TO_UINT(64, 32) -FLOAT_TO_UINT(64, 64) +uint16_t float32_to_uint16_scalbn(float32 a, int rmode, int scale, + float_status *s) +{ + return round_to_uint_and_pack(float32_unpack_canonical(a, s), + rmode, scale, UINT16_MAX, s); +} -#undef FLOAT_TO_UINT +uint32_t float32_to_uint32_scalbn(float32 a, int rmode, int scale, + float_status *s) +{ + return round_to_uint_and_pack(float32_unpack_canonical(a, s), + rmode, scale, UINT32_MAX, s); +} + +uint64_t float32_to_uint64_scalbn(float32 a, int rmode, int scale, + float_status *s) +{ + return round_to_uint_and_pack(float32_unpack_canonical(a, s), + rmode, scale, UINT64_MAX, s); +} + +uint16_t float64_to_uint16_scalbn(float64 a, int rmode, int scale, + float_status *s) +{ + return round_to_uint_and_pack(float64_unpack_canonical(a, s), + rmode, scale, UINT16_MAX, s); +} + +uint32_t float64_to_uint32_scalbn(float64 a, int rmode, int scale, + float_status *s) +{ + return round_to_uint_and_pack(float64_unpack_canonical(a, s), + rmode, scale, UINT32_MAX, s); +} + +uint64_t float64_to_uint64_scalbn(float64 a, int rmode, int scale, + float_status *s) +{ + return round_to_uint_and_pack(float64_unpack_canonical(a, s), + rmode, scale, UINT64_MAX, s); +} + +uint16_t float16_to_uint16(float16 a, float_status *s) +{ + return float16_to_uint16_scalbn(a, s->float_rounding_mode, 0, s); +} + +uint32_t float16_to_uint32(float16 a, float_status *s) +{ + return float16_to_uint32_scalbn(a, s->float_rounding_mode, 0, s); +} + +uint64_t float16_to_uint64(float16 a, float_status *s) +{ + return float16_to_uint64_scalbn(a, s->float_rounding_mode, 0, s); +} + +uint16_t float32_to_uint16(float32 a, float_status *s) +{ + return float32_to_uint16_scalbn(a, s->float_rounding_mode, 0, s); +} + +uint32_t float32_to_uint32(float32 a, float_status *s) +{ + return float32_to_uint32_scalbn(a, s->float_rounding_mode, 0, s); +} + +uint64_t float32_to_uint64(float32 a, float_status *s) +{ + return float32_to_uint64_scalbn(a, s->float_rounding_mode, 0, s); +} + +uint16_t float64_to_uint16(float64 a, float_status *s) +{ + return float64_to_uint16_scalbn(a, s->float_rounding_mode, 0, s); +} + +uint32_t float64_to_uint32(float64 a, float_status *s) +{ + return float64_to_uint32_scalbn(a, s->float_rounding_mode, 0, s); +} + +uint64_t float64_to_uint64(float64 a, float_status *s) +{ + return float64_to_uint64_scalbn(a, s->float_rounding_mode, 0, s); +} + +uint16_t float16_to_uint16_round_to_zero(float16 a, float_status *s) +{ + return float16_to_uint16_scalbn(a, float_round_to_zero, 0, s); +} + +uint32_t float16_to_uint32_round_to_zero(float16 a, float_status *s) +{ + return float16_to_uint32_scalbn(a, float_round_to_zero, 0, s); +} + +uint64_t float16_to_uint64_round_to_zero(float16 a, float_status *s) +{ + return float16_to_uint64_scalbn(a, float_round_to_zero, 0, s); +} + +uint16_t float32_to_uint16_round_to_zero(float32 a, float_status *s) +{ + return float32_to_uint16_scalbn(a, float_round_to_zero, 0, s); +} + +uint32_t float32_to_uint32_round_to_zero(float32 a, float_status *s) +{ + return float32_to_uint32_scalbn(a, float_round_to_zero, 0, s); +} + +uint64_t float32_to_uint64_round_to_zero(float32 a, float_status *s) +{ + return float32_to_uint64_scalbn(a, float_round_to_zero, 0, s); +} + +uint16_t float64_to_uint16_round_to_zero(float64 a, float_status *s) +{ + return float64_to_uint16_scalbn(a, float_round_to_zero, 0, s); +} + +uint32_t float64_to_uint32_round_to_zero(float64 a, float_status *s) +{ + return float64_to_uint32_scalbn(a, float_round_to_zero, 0, s); +} + +uint64_t float64_to_uint64_round_to_zero(float64 a, float_status *s) +{ + return float64_to_uint64_scalbn(a, float_round_to_zero, 0, s); +} /* * Integer to float conversions -- 2.18.0 From: Richard Henderson <richard.henderson@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20180814002653.12828-4-richard.henderson@linaro.org Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/helper.c | 29 +++++------------------------ 1 file changed, 5 insertions(+), 24 deletions(-) diff --git a/target/arm/helper.c b/target/arm/helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper.c +++ b/target/arm/helper.c @@ -XXX,XX +XXX,XX @@ float32 VFP_HELPER(fcvts, d)(float64 x, CPUARMState *env) #define VFP_CONV_FIX_FLOAT(name, p, fsz, isz, itype) \ float##fsz HELPER(vfp_##name##to##p)(uint##isz##_t x, uint32_t shift, \ void *fpstp) \ -{ \ - float_status *fpst = fpstp; \ - float##fsz tmp; \ - tmp = itype##_to_##float##fsz(x, fpst); \ - return float##fsz##_scalbn(tmp, -(int)shift, fpst); \ -} +{ return itype##_to_##float##fsz##_scalbn(x, -shift, fpstp); } /* Notice that we want only input-denormal exception flags from the * scalbn operation: the other possible flags (overflow+inexact if @@ -XXX,XX +XXX,XX @@ VFP_CONV_FIX_A64(uq, s, 32, 64, uint64) #undef VFP_CONV_FLOAT_FIX_ROUND #undef VFP_CONV_FIX_A64 -/* Conversion to/from f16 can overflow to infinity before/after scaling. - * Therefore we convert to f64, scale, and then convert f64 to f16; or - * vice versa for conversion to integer. - * - * For 16- and 32-bit integers, the conversion to f64 never rounds. - * For 64-bit integers, any integer that would cause rounding will also - * overflow to f16 infinity, so there is no double rounding problem. - */ - -static float16 do_postscale_fp16(float64 f, int shift, float_status *fpst) -{ - return float64_to_float16(float64_scalbn(f, -shift, fpst), true, fpst); -} - uint32_t HELPER(vfp_sltoh)(uint32_t x, uint32_t shift, void *fpst) { - return do_postscale_fp16(int32_to_float64(x, fpst), shift, fpst); + return int32_to_float16_scalbn(x, -shift, fpst); } uint32_t HELPER(vfp_ultoh)(uint32_t x, uint32_t shift, void *fpst) { - return do_postscale_fp16(uint32_to_float64(x, fpst), shift, fpst); + return uint32_to_float16_scalbn(x, -shift, fpst); } uint32_t HELPER(vfp_sqtoh)(uint64_t x, uint32_t shift, void *fpst) { - return do_postscale_fp16(int64_to_float64(x, fpst), shift, fpst); + return int64_to_float16_scalbn(x, -shift, fpst); } uint32_t HELPER(vfp_uqtoh)(uint64_t x, uint32_t shift, void *fpst) { - return do_postscale_fp16(uint64_to_float64(x, fpst), shift, fpst); + return uint64_to_float16_scalbn(x, -shift, fpst); } static float64 do_prescale_fp16(float16 f, int shift, float_status *fpst) -- 2.18.0 From: Richard Henderson <richard.henderson@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20180814002653.12828-5-richard.henderson@linaro.org Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/helper.c | 101 ++++++++++++++++++++++---------------------- 1 file changed, 51 insertions(+), 50 deletions(-) diff --git a/target/arm/helper.c b/target/arm/helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper.c +++ b/target/arm/helper.c @@ -XXX,XX +XXX,XX @@ float##fsz HELPER(vfp_##name##to##p)(uint##isz##_t x, uint32_t shift, \ void *fpstp) \ { return itype##_to_##float##fsz##_scalbn(x, -shift, fpstp); } -/* Notice that we want only input-denormal exception flags from the - * scalbn operation: the other possible flags (overflow+inexact if - * we overflow to infinity, output-denormal) aren't correct for the - * complete scale-and-convert operation. - */ -#define VFP_CONV_FLOAT_FIX_ROUND(name, p, fsz, isz, itype, round) \ -uint##isz##_t HELPER(vfp_to##name##p##round)(float##fsz x, \ - uint32_t shift, \ - void *fpstp) \ -{ \ - float_status *fpst = fpstp; \ - int old_exc_flags = get_float_exception_flags(fpst); \ - float##fsz tmp; \ - if (float##fsz##_is_any_nan(x)) { \ - float_raise(float_flag_invalid, fpst); \ - return 0; \ - } \ - tmp = float##fsz##_scalbn(x, shift, fpst); \ - old_exc_flags |= get_float_exception_flags(fpst) \ - & float_flag_input_denormal; \ - set_float_exception_flags(old_exc_flags, fpst); \ - return float##fsz##_to_##itype##round(tmp, fpst); \ +#define VFP_CONV_FLOAT_FIX_ROUND(name, p, fsz, isz, itype, ROUND, suff) \ +uint##isz##_t HELPER(vfp_to##name##p##suff)(float##fsz x, uint32_t shift, \ + void *fpst) \ +{ \ + if (unlikely(float##fsz##_is_any_nan(x))) { \ + float_raise(float_flag_invalid, fpst); \ + return 0; \ + } \ + return float##fsz##_to_##itype##_scalbn(x, ROUND, shift, fpst); \ } #define VFP_CONV_FIX(name, p, fsz, isz, itype) \ VFP_CONV_FIX_FLOAT(name, p, fsz, isz, itype) \ -VFP_CONV_FLOAT_FIX_ROUND(name, p, fsz, isz, itype, _round_to_zero) \ -VFP_CONV_FLOAT_FIX_ROUND(name, p, fsz, isz, itype, ) +VFP_CONV_FLOAT_FIX_ROUND(name, p, fsz, isz, itype, \ + float_round_to_zero, _round_to_zero) \ +VFP_CONV_FLOAT_FIX_ROUND(name, p, fsz, isz, itype, \ + get_float_rounding_mode(fpst), ) #define VFP_CONV_FIX_A64(name, p, fsz, isz, itype) \ VFP_CONV_FIX_FLOAT(name, p, fsz, isz, itype) \ -VFP_CONV_FLOAT_FIX_ROUND(name, p, fsz, isz, itype, ) +VFP_CONV_FLOAT_FIX_ROUND(name, p, fsz, isz, itype, \ + get_float_rounding_mode(fpst), ) VFP_CONV_FIX(sh, d, 64, 64, int16) VFP_CONV_FIX(sl, d, 64, 64, int32) @@ -XXX,XX +XXX,XX @@ uint32_t HELPER(vfp_uqtoh)(uint64_t x, uint32_t shift, void *fpst) return uint64_to_float16_scalbn(x, -shift, fpst); } -static float64 do_prescale_fp16(float16 f, int shift, float_status *fpst) -{ - if (unlikely(float16_is_any_nan(f))) { - float_raise(float_flag_invalid, fpst); - return 0; - } else { - int old_exc_flags = get_float_exception_flags(fpst); - float64 ret; - - ret = float16_to_float64(f, true, fpst); - ret = float64_scalbn(ret, shift, fpst); - old_exc_flags |= get_float_exception_flags(fpst) - & float_flag_input_denormal; - set_float_exception_flags(old_exc_flags, fpst); - - return ret; - } -} - uint32_t HELPER(vfp_toshh)(uint32_t x, uint32_t shift, void *fpst) { - return float64_to_int16(do_prescale_fp16(x, shift, fpst), fpst); + if (unlikely(float16_is_any_nan(x))) { + float_raise(float_flag_invalid, fpst); + return 0; + } + return float16_to_int16_scalbn(x, get_float_rounding_mode(fpst), + shift, fpst); } uint32_t HELPER(vfp_touhh)(uint32_t x, uint32_t shift, void *fpst) { - return float64_to_uint16(do_prescale_fp16(x, shift, fpst), fpst); + if (unlikely(float16_is_any_nan(x))) { + float_raise(float_flag_invalid, fpst); + return 0; + } + return float16_to_uint16_scalbn(x, get_float_rounding_mode(fpst), + shift, fpst); } uint32_t HELPER(vfp_toslh)(uint32_t x, uint32_t shift, void *fpst) { - return float64_to_int32(do_prescale_fp16(x, shift, fpst), fpst); + if (unlikely(float16_is_any_nan(x))) { + float_raise(float_flag_invalid, fpst); + return 0; + } + return float16_to_int32_scalbn(x, get_float_rounding_mode(fpst), + shift, fpst); } uint32_t HELPER(vfp_toulh)(uint32_t x, uint32_t shift, void *fpst) { - return float64_to_uint32(do_prescale_fp16(x, shift, fpst), fpst); + if (unlikely(float16_is_any_nan(x))) { + float_raise(float_flag_invalid, fpst); + return 0; + } + return float16_to_uint32_scalbn(x, get_float_rounding_mode(fpst), + shift, fpst); } uint64_t HELPER(vfp_tosqh)(uint32_t x, uint32_t shift, void *fpst) { - return float64_to_int64(do_prescale_fp16(x, shift, fpst), fpst); + if (unlikely(float16_is_any_nan(x))) { + float_raise(float_flag_invalid, fpst); + return 0; + } + return float16_to_int64_scalbn(x, get_float_rounding_mode(fpst), + shift, fpst); } uint64_t HELPER(vfp_touqh)(uint32_t x, uint32_t shift, void *fpst) { - return float64_to_uint64(do_prescale_fp16(x, shift, fpst), fpst); + if (unlikely(float16_is_any_nan(x))) { + float_raise(float_flag_invalid, fpst); + return 0; + } + return float16_to_uint64_scalbn(x, get_float_rounding_mode(fpst), + shift, fpst); } /* Set the current fp rounding mode and return the old one. -- 2.18.0 Reduce the size of the per-cpu GICH memory regions from 0x1000 to 0x200. The registers only cover 0x200 bytes, and the Cortex-A15 wants to map them at a spacing of 0x200 bytes apart. Having the region be too large interferes with mapping them like that, so reduce it. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Luc Michel <luc.michel@greensocs.com> Message-id: 20180821132811.17675-3-peter.maydell@linaro.org --- hw/intc/arm_gic.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/hw/intc/arm_gic.c b/hw/intc/arm_gic.c index XXXXXXX..XXXXXXX 100644 --- a/hw/intc/arm_gic.c +++ b/hw/intc/arm_gic.c @@ -XXX,XX +XXX,XX @@ static void arm_gic_realize(DeviceState *dev, Error **errp) for (i = 0; i < s->num_cpu; i++) { memory_region_init_io(&s->vifaceiomem[i + 1], OBJECT(s), &gic_viface_ops, &s->backref[i], - "gic_viface", 0x1000); + "gic_viface", 0x200); sysbus_init_mmio(sbd, &s->vifaceiomem[i + 1]); } } -- 2.18.0 Connect the VIRQ and VFIQ lines from the GIC to the CPU; these exist always for both CPU and GIC whether the virtualization extensions are enabled or not, so we can just unconditionally connect them. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Luc Michel <luc.michel@greensocs.com> Message-id: 20180821132811.17675-4-peter.maydell@linaro.org --- hw/arm/vexpress.c | 4 ++++ 1 file changed, 4 insertions(+) diff --git a/hw/arm/vexpress.c b/hw/arm/vexpress.c index XXXXXXX..XXXXXXX 100644 --- a/hw/arm/vexpress.c +++ b/hw/arm/vexpress.c @@ -XXX,XX +XXX,XX @@ static void init_cpus(const char *cpu_type, const char *privdev, sysbus_connect_irq(busdev, n, qdev_get_gpio_in(cpudev, ARM_CPU_IRQ)); sysbus_connect_irq(busdev, n + smp_cpus, qdev_get_gpio_in(cpudev, ARM_CPU_FIQ)); + sysbus_connect_irq(busdev, n + 2 * smp_cpus, + qdev_get_gpio_in(cpudev, ARM_CPU_VIRQ)); + sysbus_connect_irq(busdev, n + 3 * smp_cpus, + qdev_get_gpio_in(cpudev, ARM_CPU_VFIQ)); } } -- 2.18.0 Connect the VIRQ and VFIQ lines from the GIC to the CPU; these exist always for both CPU and GIC whether the virtualization extensions are enabled or not, so we can just unconditionally connect them. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Luc Michel <luc.michel@greensocs.com> Message-id: 20180821132811.17675-5-peter.maydell@linaro.org --- hw/arm/highbank.c | 6 ++++++ 1 file changed, 6 insertions(+) diff --git a/hw/arm/highbank.c b/hw/arm/highbank.c index XXXXXXX..XXXXXXX 100644 --- a/hw/arm/highbank.c +++ b/hw/arm/highbank.c @@ -XXX,XX +XXX,XX @@ static void calxeda_init(MachineState *machine, enum cxmachines machine_id) int n; qemu_irq cpu_irq[4]; qemu_irq cpu_fiq[4]; + qemu_irq cpu_virq[4]; + qemu_irq cpu_vfiq[4]; MemoryRegion *sysram; MemoryRegion *dram; MemoryRegion *sysmem; @@ -XXX,XX +XXX,XX @@ static void calxeda_init(MachineState *machine, enum cxmachines machine_id) object_property_set_bool(cpuobj, true, "realized", &error_fatal); cpu_irq[n] = qdev_get_gpio_in(DEVICE(cpu), ARM_CPU_IRQ); cpu_fiq[n] = qdev_get_gpio_in(DEVICE(cpu), ARM_CPU_FIQ); + cpu_virq[n] = qdev_get_gpio_in(DEVICE(cpu), ARM_CPU_VIRQ); + cpu_vfiq[n] = qdev_get_gpio_in(DEVICE(cpu), ARM_CPU_VFIQ); } sysmem = get_system_memory(); @@ -XXX,XX +XXX,XX @@ static void calxeda_init(MachineState *machine, enum cxmachines machine_id) for (n = 0; n < smp_cpus; n++) { sysbus_connect_irq(busdev, n, cpu_irq[n]); sysbus_connect_irq(busdev, n + smp_cpus, cpu_fiq[n]); + sysbus_connect_irq(busdev, n + 2 * smp_cpus, cpu_virq[n]); + sysbus_connect_irq(busdev, n + 3 * smp_cpus, cpu_vfiq[n]); } for (n = 0; n < 128; n++) { -- 2.18.0 Connect the VIRQ and VFIQ lines from the GIC to the CPU; these exist always for both CPU and GIC whether the virtualization extensions are enabled or not, so we can just unconditionally connect them. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Luc Michel <luc.michel@greensocs.com> Message-id: 20180821132811.17675-6-peter.maydell@linaro.org --- hw/arm/fsl-imx6ul.c | 4 ++++ 1 file changed, 4 insertions(+) diff --git a/hw/arm/fsl-imx6ul.c b/hw/arm/fsl-imx6ul.c index XXXXXXX..XXXXXXX 100644 --- a/hw/arm/fsl-imx6ul.c +++ b/hw/arm/fsl-imx6ul.c @@ -XXX,XX +XXX,XX @@ static void fsl_imx6ul_realize(DeviceState *dev, Error **errp) irq = qdev_get_gpio_in(d, ARM_CPU_IRQ); sysbus_connect_irq(sbd, i, irq); sysbus_connect_irq(sbd, i + smp_cpus, qdev_get_gpio_in(d, ARM_CPU_FIQ)); + sysbus_connect_irq(sbd, i + 2 * smp_cpus, + qdev_get_gpio_in(d, ARM_CPU_VIRQ)); + sysbus_connect_irq(sbd, i + 3 * smp_cpus, + qdev_get_gpio_in(d, ARM_CPU_VFIQ)); } /* -- 2.18.0 Connect the VIRQ and VFIQ lines from the GIC to the CPU; these exist always for both CPU and GIC whether the virtualization extensions are enabled or not, so we can just unconditionally connect them. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Luc Michel <luc.michel@greensocs.com> Message-id: 20180821132811.17675-7-peter.maydell@linaro.org --- hw/arm/fsl-imx7.c | 4 ++++ 1 file changed, 4 insertions(+) diff --git a/hw/arm/fsl-imx7.c b/hw/arm/fsl-imx7.c index XXXXXXX..XXXXXXX 100644 --- a/hw/arm/fsl-imx7.c +++ b/hw/arm/fsl-imx7.c @@ -XXX,XX +XXX,XX @@ static void fsl_imx7_realize(DeviceState *dev, Error **errp) sysbus_connect_irq(sbd, i, irq); irq = qdev_get_gpio_in(d, ARM_CPU_FIQ); sysbus_connect_irq(sbd, i + smp_cpus, irq); + irq = qdev_get_gpio_in(d, ARM_CPU_VIRQ); + sysbus_connect_irq(sbd, i + 2 * smp_cpus, irq); + irq = qdev_get_gpio_in(d, ARM_CPU_VFIQ); + sysbus_connect_irq(sbd, i + 3 * smp_cpus, irq); } /* -- 2.18.0 For the A15MPCore internal peripheral object, we handle GIC security extensions support by checking whether the CPUs have EL3 enabled; if so then we enable it also on the GIC. Handle the virtualization extensions in the same way: if the CPU has EL2 then enable it on the GIC and wire up the virtualization-specific memory regions and the maintenance interrupt. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Luc Michel <luc.michel@greensocs.com> Message-id: 20180821132811.17675-8-peter.maydell@linaro.org --- hw/cpu/a15mpcore.c | 31 ++++++++++++++++++++++++++++--- 1 file changed, 28 insertions(+), 3 deletions(-) diff --git a/hw/cpu/a15mpcore.c b/hw/cpu/a15mpcore.c index XXXXXXX..XXXXXXX 100644 --- a/hw/cpu/a15mpcore.c +++ b/hw/cpu/a15mpcore.c @@ -XXX,XX +XXX,XX @@ static void a15mp_priv_realize(DeviceState *dev, Error **errp) int i; Error *err = NULL; bool has_el3; + bool has_el2; Object *cpuobj; gicdev = DEVICE(&s->gic); @@ -XXX,XX +XXX,XX @@ static void a15mp_priv_realize(DeviceState *dev, Error **errp) has_el3 = object_property_find(cpuobj, "has_el3", NULL) && object_property_get_bool(cpuobj, "has_el3", &error_abort); qdev_prop_set_bit(gicdev, "has-security-extensions", has_el3); + /* Similarly for virtualization support */ + has_el2 = object_property_find(cpuobj, "has_el2", NULL) && + object_property_get_bool(cpuobj, "has_el2", &error_abort); + qdev_prop_set_bit(gicdev, "has-virtualization-extensions", has_el2); } object_property_set_bool(OBJECT(&s->gic), true, "realized", &err); @@ -XXX,XX +XXX,XX @@ static void a15mp_priv_realize(DeviceState *dev, Error **errp) qdev_get_gpio_in(gicdev, ppibase + timer_irq[irq])); } + if (has_el2) { + /* Connect the GIC maintenance interrupt to PPI ID 25 */ + sysbus_connect_irq(SYS_BUS_DEVICE(gicdev), i + 4 * s->num_cpu, + qdev_get_gpio_in(gicdev, ppibase + 25)); + } } /* Memory map (addresses are offsets from PERIPHBASE): * 0x0000-0x0fff -- reserved * 0x1000-0x1fff -- GIC Distributor * 0x2000-0x3fff -- GIC CPU interface - * 0x4000-0x4fff -- GIC virtual interface control (not modelled) - * 0x5000-0x5fff -- GIC virtual interface control (not modelled) - * 0x6000-0x7fff -- GIC virtual CPU interface (not modelled) + * 0x4000-0x4fff -- GIC virtual interface control for this CPU + * 0x5000-0x51ff -- GIC virtual interface control for CPU 0 + * 0x5200-0x53ff -- GIC virtual interface control for CPU 1 + * 0x5400-0x55ff -- GIC virtual interface control for CPU 2 + * 0x5600-0x57ff -- GIC virtual interface control for CPU 3 + * 0x6000-0x7fff -- GIC virtual CPU interface */ memory_region_add_subregion(&s->container, 0x1000, sysbus_mmio_get_region(busdev, 0)); memory_region_add_subregion(&s->container, 0x2000, sysbus_mmio_get_region(busdev, 1)); + if (has_el2) { + memory_region_add_subregion(&s->container, 0x4000, + sysbus_mmio_get_region(busdev, 2)); + memory_region_add_subregion(&s->container, 0x6000, + sysbus_mmio_get_region(busdev, 3)); + for (i = 0; i < s->num_cpu; i++) { + hwaddr base = 0x5000 + i * 0x200; + MemoryRegion *mr = sysbus_mmio_get_region(busdev, + 4 + s->num_cpu + i); + memory_region_add_subregion(&s->container, base, mr); + } + } } static Property a15mp_priv_properties[] = { -- 2.18.0 Don't request that the arm_load_kernel() code should boot in secure state if the CPU doesn't have a secure state. Currently this doesn't make a difference because the boot.c code only examines the secure_boot flag in code guarded by an ARM_FEATURE_EL3 check, but upcoming changes for supporting booting into Hyp mode will change that. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Luc Michel <luc.michel@greensocs.com> Message-id: 20180821132811.17675-9-peter.maydell@linaro.org --- hw/arm/vexpress.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/hw/arm/vexpress.c b/hw/arm/vexpress.c index XXXXXXX..XXXXXXX 100644 --- a/hw/arm/vexpress.c +++ b/hw/arm/vexpress.c @@ -XXX,XX +XXX,XX @@ static void vexpress_common_init(MachineState *machine) daughterboard->bootinfo.smp_bootreg_addr = map[VE_SYSREGS] + 0x30; daughterboard->bootinfo.gic_cpu_if_addr = daughterboard->gic_cpu_if_addr; daughterboard->bootinfo.modify_dtb = vexpress_modify_dtb; - /* Indicate that when booting Linux we should be in secure state */ - daughterboard->bootinfo.secure_boot = true; + /* When booting Linux we should be in secure state if the CPU has one. */ + daughterboard->bootinfo.secure_boot = vms->secure; arm_load_kernel(ARM_CPU(first_cpu), &daughterboard->bootinfo); } -- 2.18.0 Add a "virtualization" property to the vexpress-a15 board, controlling presence of EL2. As with EL3, we default to enabling it, but the user can disable it if they have an older guest which can't cope with it being present. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Luc Michel <luc.michel@greensocs.com> Message-id: 20180821132811.17675-10-peter.maydell@linaro.org --- hw/arm/vexpress.c | 56 ++++++++++++++++++++++++++++++++++++++++++++--- 1 file changed, 53 insertions(+), 3 deletions(-) diff --git a/hw/arm/vexpress.c b/hw/arm/vexpress.c index XXXXXXX..XXXXXXX 100644 --- a/hw/arm/vexpress.c +++ b/hw/arm/vexpress.c @@ -XXX,XX +XXX,XX @@ typedef struct { typedef struct { MachineState parent; bool secure; + bool virt; } VexpressMachineState; #define TYPE_VEXPRESS_MACHINE "vexpress" @@ -XXX,XX +XXX,XX @@ struct VEDBoardInfo { }; static void init_cpus(const char *cpu_type, const char *privdev, - hwaddr periphbase, qemu_irq *pic, bool secure) + hwaddr periphbase, qemu_irq *pic, bool secure, bool virt) { DeviceState *dev; SysBusDevice *busdev; @@ -XXX,XX +XXX,XX @@ static void init_cpus(const char *cpu_type, const char *privdev, if (!secure) { object_property_set_bool(cpuobj, false, "has_el3", NULL); } + if (!virt) { + if (object_property_find(cpuobj, "has_el2", NULL)) { + object_property_set_bool(cpuobj, false, "has_el2", NULL); + } + } if (object_property_find(cpuobj, "reset-cbar", NULL)) { object_property_set_int(cpuobj, periphbase, @@ -XXX,XX +XXX,XX @@ static void a9_daughterboard_init(const VexpressMachineState *vms, memory_region_add_subregion(sysmem, 0x60000000, ram); /* 0x1e000000 A9MPCore (SCU) private memory region */ - init_cpus(cpu_type, TYPE_A9MPCORE_PRIV, 0x1e000000, pic, vms->secure); + init_cpus(cpu_type, TYPE_A9MPCORE_PRIV, 0x1e000000, pic, + vms->secure, vms->virt); /* Daughterboard peripherals : 0x10020000 .. 0x20000000 */ @@ -XXX,XX +XXX,XX @@ static void a15_daughterboard_init(const VexpressMachineState *vms, memory_region_add_subregion(sysmem, 0x80000000, ram); /* 0x2c000000 A15MPCore private memory region (GIC) */ - init_cpus(cpu_type, TYPE_A15MPCORE_PRIV, 0x2c000000, pic, vms->secure); + init_cpus(cpu_type, TYPE_A15MPCORE_PRIV, 0x2c000000, pic, vms->secure, + vms->virt); /* A15 daughterboard peripherals: */ @@ -XXX,XX +XXX,XX @@ static void vexpress_set_secure(Object *obj, bool value, Error **errp) vms->secure = value; } +static bool vexpress_get_virt(Object *obj, Error **errp) +{ + VexpressMachineState *vms = VEXPRESS_MACHINE(obj); + + return vms->virt; +} + +static void vexpress_set_virt(Object *obj, bool value, Error **errp) +{ + VexpressMachineState *vms = VEXPRESS_MACHINE(obj); + + vms->virt = value; +} + static void vexpress_instance_init(Object *obj) { VexpressMachineState *vms = VEXPRESS_MACHINE(obj); @@ -XXX,XX +XXX,XX @@ static void vexpress_instance_init(Object *obj) NULL); } +static void vexpress_a15_instance_init(Object *obj) +{ + VexpressMachineState *vms = VEXPRESS_MACHINE(obj); + + /* + * For the vexpress-a15, EL2 is by default enabled if EL3 is, + * but can also be specifically set to on or off. + */ + vms->virt = true; + object_property_add_bool(obj, "virtualization", vexpress_get_virt, + vexpress_set_virt, NULL); + object_property_set_description(obj, "virtualization", + "Set on/off to enable/disable the ARM " + "Virtualization Extensions " + "(defaults to same as 'secure')", + NULL); +} + +static void vexpress_a9_instance_init(Object *obj) +{ + VexpressMachineState *vms = VEXPRESS_MACHINE(obj); + + /* The A9 doesn't have the virt extensions */ + vms->virt = false; +} + static void vexpress_class_init(ObjectClass *oc, void *data) { MachineClass *mc = MACHINE_CLASS(oc); @@ -XXX,XX +XXX,XX @@ static const TypeInfo vexpress_a9_info = { .name = TYPE_VEXPRESS_A9_MACHINE, .parent = TYPE_VEXPRESS_MACHINE, .class_init = vexpress_a9_class_init, + .instance_init = vexpress_a9_instance_init, }; static const TypeInfo vexpress_a15_info = { .name = TYPE_VEXPRESS_A15_MACHINE, .parent = TYPE_VEXPRESS_MACHINE, .class_init = vexpress_a15_class_init, + .instance_init = vexpress_a15_instance_init, }; static void vexpress_machine_init(void) -- 2.18.0 The v8 AArch32 HACTLR2 register maps to bits [63:32] of ACTLR_EL2. We implement ACTLR_EL2 as RAZ/WI, so make HACTLR2 also RAZ/WI. (We put the regdef next to ACTLR_EL2 as a reminder in case we ever make ACTLR_EL2 something other than RAZ/WI). Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Reviewed-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com> Reviewed-by: Luc Michel <luc.michel@greensocs.com> Message-id: 20180820153020.21478-2-peter.maydell@linaro.org --- target/arm/helper.c | 10 ++++++++++ 1 file changed, 10 insertions(+) diff --git a/target/arm/helper.c b/target/arm/helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper.c +++ b/target/arm/helper.c @@ -XXX,XX +XXX,XX @@ void register_cp_regs_for_features(ARMCPU *cpu) REGINFO_SENTINEL }; define_arm_cp_regs(cpu, auxcr_reginfo); + if (arm_feature(env, ARM_FEATURE_V8)) { + /* HACTLR2 maps to ACTLR_EL2[63:32] and is not in ARMv7 */ + ARMCPRegInfo hactlr2_reginfo = { + .name = "HACTLR2", .state = ARM_CP_STATE_AA32, + .cp = 15, .opc1 = 4, .crn = 1, .crm = 0, .opc2 = 3, + .access = PL2_RW, .type = ARM_CP_CONST, + .resetvalue = 0 + }; + define_one_arm_cp_reg(cpu, &hactlr2_reginfo); + } } if (arm_feature(env, ARM_FEATURE_CBAR)) { -- 2.18.0 The AArch32 HCR and HCR2 registers alias HCR_EL2 bits [31:0] and [63:32]; implement them. Since HCR2 exists in ARMv8 but not ARMv7, we need new regdef arrays for "we have EL3, not EL2, we're ARMv8" and "we have EL2, we're ARMv8" to hold the definitions. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Reviewed-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com> Reviewed-by: Luc Michel <luc.michel@greensocs.com> Message-id: 20180820153020.21478-3-peter.maydell@linaro.org --- target/arm/helper.c | 54 +++++++++++++++++++++++++++++++++++++++++---- 1 file changed, 50 insertions(+), 4 deletions(-) diff --git a/target/arm/helper.c b/target/arm/helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper.c +++ b/target/arm/helper.c @@ -XXX,XX +XXX,XX @@ static const ARMCPRegInfo el3_no_el2_cp_reginfo[] = { .opc0 = 3, .opc1 = 4, .crn = 12, .crm = 0, .opc2 = 0, .access = PL2_RW, .readfn = arm_cp_read_zero, .writefn = arm_cp_write_ignore }, - { .name = "HCR_EL2", .state = ARM_CP_STATE_AA64, + { .name = "HCR_EL2", .state = ARM_CP_STATE_BOTH, .type = ARM_CP_NO_RAW, .opc0 = 3, .opc1 = 4, .crn = 1, .crm = 1, .opc2 = 0, .access = PL2_RW, - .readfn = arm_cp_read_zero, .writefn = arm_cp_write_ignore }, + .type = ARM_CP_CONST, .resetvalue = 0 }, { .name = "ESR_EL2", .state = ARM_CP_STATE_BOTH, .opc0 = 3, .opc1 = 4, .crn = 5, .crm = 2, .opc2 = 0, .access = PL2_RW, @@ -XXX,XX +XXX,XX @@ static const ARMCPRegInfo el3_no_el2_cp_reginfo[] = { REGINFO_SENTINEL }; +/* Ditto, but for registers which exist in ARMv8 but not v7 */ +static const ARMCPRegInfo el3_no_el2_v8_cp_reginfo[] = { + { .name = "HCR2", .state = ARM_CP_STATE_AA32, + .cp = 15, .opc1 = 4, .crn = 1, .crm = 1, .opc2 = 4, + .access = PL2_RW, + .type = ARM_CP_CONST, .resetvalue = 0 }, + REGINFO_SENTINEL +}; + static void hcr_write(CPUARMState *env, const ARMCPRegInfo *ri, uint64_t value) { ARMCPU *cpu = arm_env_get_cpu(env); @@ -XXX,XX +XXX,XX @@ static void hcr_write(CPUARMState *env, const ARMCPRegInfo *ri, uint64_t value) * HCR_PTW forbids certain page-table setups * HCR_DC Disables stage1 and enables stage2 translation */ - if ((raw_read(env, ri) ^ value) & (HCR_VM | HCR_PTW | HCR_DC)) { + if ((env->cp15.hcr_el2 ^ value) & (HCR_VM | HCR_PTW | HCR_DC)) { tlb_flush(CPU(cpu)); } - raw_write(env, ri, value); + env->cp15.hcr_el2 = value; +} + +static void hcr_writehigh(CPUARMState *env, const ARMCPRegInfo *ri, + uint64_t value) +{ + /* Handle HCR2 write, i.e. write to high half of HCR_EL2 */ + value = deposit64(env->cp15.hcr_el2, 32, 32, value); + hcr_write(env, NULL, value); +} + +static void hcr_writelow(CPUARMState *env, const ARMCPRegInfo *ri, + uint64_t value) +{ + /* Handle HCR write, i.e. write to low half of HCR_EL2 */ + value = deposit64(env->cp15.hcr_el2, 0, 32, value); + hcr_write(env, NULL, value); } static const ARMCPRegInfo el2_cp_reginfo[] = { @@ -XXX,XX +XXX,XX @@ static const ARMCPRegInfo el2_cp_reginfo[] = { .opc0 = 3, .opc1 = 4, .crn = 1, .crm = 1, .opc2 = 0, .access = PL2_RW, .fieldoffset = offsetof(CPUARMState, cp15.hcr_el2), .writefn = hcr_write }, + { .name = "HCR", .state = ARM_CP_STATE_AA32, + .type = ARM_CP_ALIAS, + .cp = 15, .opc1 = 4, .crn = 1, .crm = 1, .opc2 = 0, + .access = PL2_RW, .fieldoffset = offsetof(CPUARMState, cp15.hcr_el2), + .writefn = hcr_writelow }, { .name = "ELR_EL2", .state = ARM_CP_STATE_AA64, .type = ARM_CP_ALIAS, .opc0 = 3, .opc1 = 4, .crn = 4, .crm = 0, .opc2 = 1, @@ -XXX,XX +XXX,XX @@ static const ARMCPRegInfo el2_cp_reginfo[] = { REGINFO_SENTINEL }; +static const ARMCPRegInfo el2_v8_cp_reginfo[] = { + { .name = "HCR2", .state = ARM_CP_STATE_AA32, + .type = ARM_CP_ALIAS, + .cp = 15, .opc1 = 4, .crn = 1, .crm = 1, .opc2 = 4, + .access = PL2_RW, + .fieldoffset = offsetofhigh32(CPUARMState, cp15.hcr_el2), + .writefn = hcr_writehigh }, + REGINFO_SENTINEL +}; + static CPAccessResult nsacr_access(CPUARMState *env, const ARMCPRegInfo *ri, bool isread) { @@ -XXX,XX +XXX,XX @@ void register_cp_regs_for_features(ARMCPU *cpu) }; define_arm_cp_regs(cpu, vpidr_regs); define_arm_cp_regs(cpu, el2_cp_reginfo); + if (arm_feature(env, ARM_FEATURE_V8)) { + define_arm_cp_regs(cpu, el2_v8_cp_reginfo); + } /* RVBAR_EL2 is only implemented if EL2 is the highest EL */ if (!arm_feature(env, ARM_FEATURE_EL3)) { ARMCPRegInfo rvbar = { @@ -XXX,XX +XXX,XX @@ void register_cp_regs_for_features(ARMCPU *cpu) }; define_arm_cp_regs(cpu, vpidr_regs); define_arm_cp_regs(cpu, el3_no_el2_cp_reginfo); + if (arm_feature(env, ARM_FEATURE_V8)) { + define_arm_cp_regs(cpu, el3_no_el2_v8_cp_reginfo); + } } } if (arm_feature(env, ARM_FEATURE_EL3)) { -- 2.18.0 Factor out the code which changes the CPU state so as to actually take an exception to AArch32. We're going to want to use this for handling exception entry to Hyp mode. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Reviewed-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com> Reviewed-by: Luc Michel <luc.michel@greensocs.com> Message-id: 20180820153020.21478-4-peter.maydell@linaro.org --- target/arm/helper.c | 64 +++++++++++++++++++++++++++++---------------- 1 file changed, 41 insertions(+), 23 deletions(-) diff --git a/target/arm/helper.c b/target/arm/helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper.c +++ b/target/arm/helper.c @@ -XXX,XX +XXX,XX @@ void aarch64_sync_64_to_32(CPUARMState *env) env->regs[15] = env->pc; } +static void take_aarch32_exception(CPUARMState *env, int new_mode, + uint32_t mask, uint32_t offset, + uint32_t newpc) +{ + /* Change the CPU state so as to actually take the exception. */ + switch_mode(env, new_mode); + /* + * For exceptions taken to AArch32 we must clear the SS bit in both + * PSTATE and in the old-state value we save to SPSR_<mode>, so zero it now. + */ + env->uncached_cpsr &= ~PSTATE_SS; + env->spsr = cpsr_read(env); + /* Clear IT bits. */ + env->condexec_bits = 0; + /* Switch to the new mode, and to the correct instruction set. */ + env->uncached_cpsr = (env->uncached_cpsr & ~CPSR_M) | new_mode; + /* Set new mode endianness */ + env->uncached_cpsr &= ~CPSR_E; + if (env->cp15.sctlr_el[arm_current_el(env)] & SCTLR_EE) { + env->uncached_cpsr |= CPSR_E; + } + env->daif |= mask; + + if (new_mode == ARM_CPU_MODE_HYP) { + env->thumb = (env->cp15.sctlr_el[2] & SCTLR_TE) != 0; + env->elr_el[2] = env->regs[15]; + } else { + /* + * this is a lie, as there was no c1_sys on V4T/V5, but who cares + * and we should just guard the thumb mode on V4 + */ + if (arm_feature(env, ARM_FEATURE_V4T)) { + env->thumb = + (A32_BANKED_CURRENT_REG_GET(env, sctlr) & SCTLR_TE) != 0; + } + env->regs[14] = env->regs[15] + offset; + } + env->regs[15] = newpc; +} + static void arm_cpu_do_interrupt_aarch32(CPUState *cs) { ARMCPU *cpu = ARM_CPU(cs); @@ -XXX,XX +XXX,XX @@ static void arm_cpu_do_interrupt_aarch32(CPUState *cs) env->cp15.scr_el3 &= ~SCR_NS; } - switch_mode (env, new_mode); - /* For exceptions taken to AArch32 we must clear the SS bit in both - * PSTATE and in the old-state value we save to SPSR_<mode>, so zero it now. - */ - env->uncached_cpsr &= ~PSTATE_SS; - env->spsr = cpsr_read(env); - /* Clear IT bits. */ - env->condexec_bits = 0; - /* Switch to the new mode, and to the correct instruction set. */ - env->uncached_cpsr = (env->uncached_cpsr & ~CPSR_M) | new_mode; - /* Set new mode endianness */ - env->uncached_cpsr &= ~CPSR_E; - if (env->cp15.sctlr_el[arm_current_el(env)] & SCTLR_EE) { - env->uncached_cpsr |= CPSR_E; - } - env->daif |= mask; - /* this is a lie, as the was no c1_sys on V4T/V5, but who cares - * and we should just guard the thumb mode on V4 */ - if (arm_feature(env, ARM_FEATURE_V4T)) { - env->thumb = (A32_BANKED_CURRENT_REG_GET(env, sctlr) & SCTLR_TE) != 0; - } - env->regs[14] = env->regs[15] + offset; - env->regs[15] = addr; + take_aarch32_exception(env, new_mode, mask, offset, addr); } /* Handle exception entry to a target EL which is using AArch64 */ -- 2.18.0 Implement the necessary support code for taking exceptions to Hyp mode in AArch32. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Reviewed-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com> Reviewed-by: Luc Michel <luc.michel@greensocs.com> Message-id: 20180820153020.21478-5-peter.maydell@linaro.org --- target/arm/helper.c | 82 +++++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 82 insertions(+) diff --git a/target/arm/helper.c b/target/arm/helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper.c +++ b/target/arm/helper.c @@ -XXX,XX +XXX,XX @@ static void take_aarch32_exception(CPUARMState *env, int new_mode, env->regs[15] = newpc; } +static void arm_cpu_do_interrupt_aarch32_hyp(CPUState *cs) +{ + /* + * Handle exception entry to Hyp mode; this is sufficiently + * different to entry to other AArch32 modes that we handle it + * separately here. + * + * The vector table entry used is always the 0x14 Hyp mode entry point, + * unless this is an UNDEF/HVC/abort taken from Hyp to Hyp. + * The offset applied to the preferred return address is always zero + * (see DDI0487C.a section G1.12.3). + * PSTATE A/I/F masks are set based only on the SCR.EA/IRQ/FIQ values. + */ + uint32_t addr, mask; + ARMCPU *cpu = ARM_CPU(cs); + CPUARMState *env = &cpu->env; + + switch (cs->exception_index) { + case EXCP_UDEF: + addr = 0x04; + break; + case EXCP_SWI: + addr = 0x14; + break; + case EXCP_BKPT: + /* Fall through to prefetch abort. */ + case EXCP_PREFETCH_ABORT: + env->cp15.ifar_s = env->exception.vaddress; + qemu_log_mask(CPU_LOG_INT, "...with HIFAR 0x%x\n", + (uint32_t)env->exception.vaddress); + addr = 0x0c; + break; + case EXCP_DATA_ABORT: + env->cp15.dfar_s = env->exception.vaddress; + qemu_log_mask(CPU_LOG_INT, "...with HDFAR 0x%x\n", + (uint32_t)env->exception.vaddress); + addr = 0x10; + break; + case EXCP_IRQ: + addr = 0x18; + break; + case EXCP_FIQ: + addr = 0x1c; + break; + case EXCP_HVC: + addr = 0x08; + break; + case EXCP_HYP_TRAP: + addr = 0x14; + default: + cpu_abort(cs, "Unhandled exception 0x%x\n", cs->exception_index); + } + + if (cs->exception_index != EXCP_IRQ && cs->exception_index != EXCP_FIQ) { + env->cp15.esr_el[2] = env->exception.syndrome; + } + + if (arm_current_el(env) != 2 && addr < 0x14) { + addr = 0x14; + } + + mask = 0; + if (!(env->cp15.scr_el3 & SCR_EA)) { + mask |= CPSR_A; + } + if (!(env->cp15.scr_el3 & SCR_IRQ)) { + mask |= CPSR_I; + } + if (!(env->cp15.scr_el3 & SCR_FIQ)) { + mask |= CPSR_F; + } + + addr += env->cp15.hvbar; + + take_aarch32_exception(env, ARM_CPU_MODE_HYP, mask, 0, addr); +} + static void arm_cpu_do_interrupt_aarch32(CPUState *cs) { ARMCPU *cpu = ARM_CPU(cs); @@ -XXX,XX +XXX,XX @@ static void arm_cpu_do_interrupt_aarch32(CPUState *cs) env->cp15.mdscr_el1 = deposit64(env->cp15.mdscr_el1, 2, 4, moe); } + if (env->exception.target_el == 2) { + arm_cpu_do_interrupt_aarch32_hyp(cs); + return; + } + /* TODO: Vectored interrupt controller. */ switch (cs->exception_index) { case EXCP_UDEF: -- 2.18.0 On 32-bit exception entry, CPSR.J must always be set to 0 (see v7A Arm ARM DDI0406C.c B1.8.5). CPSR.IL must also be cleared on 32-bit exception entry (see v8A Arm ARM DDI0487C.a G1.10). Clear these bits. (This fixes a bug which will never be noticed by non-buggy guests.) Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Reviewed-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com> Reviewed-by: Luc Michel <luc.michel@greensocs.com> Message-id: 20180820153020.21478-6-peter.maydell@linaro.org --- target/arm/helper.c | 2 ++ 1 file changed, 2 insertions(+) diff --git a/target/arm/helper.c b/target/arm/helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper.c +++ b/target/arm/helper.c @@ -XXX,XX +XXX,XX @@ static void take_aarch32_exception(CPUARMState *env, int new_mode, if (env->cp15.sctlr_el[arm_current_el(env)] & SCTLR_EE) { env->uncached_cpsr |= CPSR_E; } + /* J and IL must always be cleared for exception entry */ + env->uncached_cpsr &= ~(CPSR_IL | CPSR_J); env->daif |= mask; if (new_mode == ARM_CPU_MODE_HYP) { -- 2.18.0 The kernel booting specification for an AArch32 kernel requires that it is booted in Hyp mode if available; otherwise the kernel can't enable KVM. We were incorrectly leaving the kernel in SVC mode. If we're booting an AArch32 kernel in the Nonsecure state and Hyp mode is available, start in it. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Reviewed-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com> Reviewed-by: Luc Michel <luc.michel@greensocs.com> Message-id: 20180820153020.21478-7-peter.maydell@linaro.org --- hw/arm/boot.c | 11 +++++++++++ 1 file changed, 11 insertions(+) diff --git a/hw/arm/boot.c b/hw/arm/boot.c index XXXXXXX..XXXXXXX 100644 --- a/hw/arm/boot.c +++ b/hw/arm/boot.c @@ -XXX,XX +XXX,XX @@ static void do_cpu_reset(void *opaque) } } + if (!env->aarch64 && !info->secure_boot && + arm_feature(env, ARM_FEATURE_EL2)) { + /* + * This is an AArch32 boot not to Secure state, and + * we have Hyp mode available, so boot the kernel into + * Hyp mode. This is not how the CPU comes out of reset, + * so we need to manually put it there. + */ + cpsr_write(env, ARM_CPU_MODE_HYP, CPSR_M, CPSRWriteRaw); + } + if (cs == first_cpu) { AddressSpace *as = arm_boot_address_space(cpu, info); -- 2.18.0 The MPS2 FPGAIO block includes some simple free-running counters. Implement these. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20180820141116.9118-2-peter.maydell@linaro.org --- include/hw/misc/mps2-fpgaio.h | 4 +++ hw/misc/mps2-fpgaio.c | 53 ++++++++++++++++++++++++++++++++++- 2 files changed, 56 insertions(+), 1 deletion(-) diff --git a/include/hw/misc/mps2-fpgaio.h b/include/hw/misc/mps2-fpgaio.h index XXXXXXX..XXXXXXX 100644 --- a/include/hw/misc/mps2-fpgaio.h +++ b/include/hw/misc/mps2-fpgaio.h @@ -XXX,XX +XXX,XX @@ typedef struct { uint32_t misc; uint32_t prescale_clk; + + /* These hold the CLOCK_VIRTUAL ns tick when the CLK1HZ/CLK100HZ was zero */ + int64_t clk1hz_tick_offset; + int64_t clk100hz_tick_offset; } MPS2FPGAIO; #endif diff --git a/hw/misc/mps2-fpgaio.c b/hw/misc/mps2-fpgaio.c index XXXXXXX..XXXXXXX 100644 --- a/hw/misc/mps2-fpgaio.c +++ b/hw/misc/mps2-fpgaio.c @@ -XXX,XX +XXX,XX @@ #include "hw/sysbus.h" #include "hw/registerfields.h" #include "hw/misc/mps2-fpgaio.h" +#include "qemu/timer.h" REG32(LED0, 0) REG32(BUTTON, 8) @@ -XXX,XX +XXX,XX @@ REG32(PRESCALE, 0x1c) REG32(PSCNTR, 0x20) REG32(MISC, 0x4c) +static uint32_t counter_from_tickoff(int64_t now, int64_t tick_offset, int frq) +{ + return muldiv64(now - tick_offset, frq, NANOSECONDS_PER_SECOND); +} + +static int64_t tickoff_from_counter(int64_t now, uint32_t count, int frq) +{ + return now - muldiv64(count, NANOSECONDS_PER_SECOND, frq); +} + static uint64_t mps2_fpgaio_read(void *opaque, hwaddr offset, unsigned size) { MPS2FPGAIO *s = MPS2_FPGAIO(opaque); uint64_t r; + int64_t now; switch (offset) { case A_LED0: @@ -XXX,XX +XXX,XX @@ static uint64_t mps2_fpgaio_read(void *opaque, hwaddr offset, unsigned size) r = s->misc; break; case A_CLK1HZ: + now = qemu_clock_get_ns(QEMU_CLOCK_VIRTUAL); + r = counter_from_tickoff(now, s->clk1hz_tick_offset, 1); + break; case A_CLK100HZ: + now = qemu_clock_get_ns(QEMU_CLOCK_VIRTUAL); + r = counter_from_tickoff(now, s->clk100hz_tick_offset, 100); + break; case A_COUNTER: case A_PSCNTR: - /* These are all upcounters of various frequencies. */ qemu_log_mask(LOG_UNIMP, "MPS2 FPGAIO: counters unimplemented\n"); r = 0; break; @@ -XXX,XX +XXX,XX @@ static void mps2_fpgaio_write(void *opaque, hwaddr offset, uint64_t value, unsigned size) { MPS2FPGAIO *s = MPS2_FPGAIO(opaque); + int64_t now; trace_mps2_fpgaio_write(offset, value, size); @@ -XXX,XX +XXX,XX @@ static void mps2_fpgaio_write(void *opaque, hwaddr offset, uint64_t value, "MPS2 FPGAIO: MISC control bits unimplemented\n"); s->misc = value; break; + case A_CLK1HZ: + now = qemu_clock_get_ns(QEMU_CLOCK_VIRTUAL); + s->clk1hz_tick_offset = tickoff_from_counter(now, value, 1); + break; + case A_CLK100HZ: + now = qemu_clock_get_ns(QEMU_CLOCK_VIRTUAL); + s->clk100hz_tick_offset = tickoff_from_counter(now, value, 100); + break; default: qemu_log_mask(LOG_GUEST_ERROR, "MPS2 FPGAIO write: bad offset 0x%x\n", (int) offset); @@ -XXX,XX +XXX,XX @@ static const MemoryRegionOps mps2_fpgaio_ops = { static void mps2_fpgaio_reset(DeviceState *dev) { MPS2FPGAIO *s = MPS2_FPGAIO(dev); + int64_t now = qemu_clock_get_ns(QEMU_CLOCK_VIRTUAL); trace_mps2_fpgaio_reset(); s->led0 = 0; s->prescale = 0; s->misc = 0; + s->clk1hz_tick_offset = tickoff_from_counter(now, 0, 1); + s->clk100hz_tick_offset = tickoff_from_counter(now, 0, 100); } static void mps2_fpgaio_init(Object *obj) @@ -XXX,XX +XXX,XX @@ static void mps2_fpgaio_init(Object *obj) sysbus_init_mmio(sbd, &s->iomem); } +static bool mps2_fpgaio_counters_needed(void *opaque) +{ + /* Currently vmstate.c insists all subsections have a 'needed' function */ + return true; +} + +static const VMStateDescription mps2_fpgaio_counters_vmstate = { + .name = "mps2-fpgaio/counters", + .version_id = 1, + .minimum_version_id = 1, + .needed = mps2_fpgaio_counters_needed, + .fields = (VMStateField[]) { + VMSTATE_INT64(clk1hz_tick_offset, MPS2FPGAIO), + VMSTATE_INT64(clk100hz_tick_offset, MPS2FPGAIO), + VMSTATE_END_OF_LIST() + } +}; + static const VMStateDescription mps2_fpgaio_vmstate = { .name = "mps2-fpgaio", .version_id = 1, @@ -XXX,XX +XXX,XX @@ static const VMStateDescription mps2_fpgaio_vmstate = { VMSTATE_UINT32(prescale, MPS2FPGAIO), VMSTATE_UINT32(misc, MPS2FPGAIO), VMSTATE_END_OF_LIST() + }, + .subsections = (const VMStateDescription*[]) { + &mps2_fpgaio_counters_vmstate, + NULL } }; -- 2.18.0 In the MPS2 FPGAIO, PSCNTR is a free-running downcounter with a reload value configured via the PRESCALE register, and COUNTER counts up by 1 every time PSCNTR reaches zero. Implement these counters. We can just increment the counters migration subsection's version ID because we only added it in the previous commit, so no released QEMU versions will be using it. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Alistair Francis <alistair.francis@wdc.com> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20180820141116.9118-3-peter.maydell@linaro.org --- include/hw/misc/mps2-fpgaio.h | 6 +++ hw/misc/mps2-fpgaio.c | 97 +++++++++++++++++++++++++++++++++-- 2 files changed, 99 insertions(+), 4 deletions(-) diff --git a/include/hw/misc/mps2-fpgaio.h b/include/hw/misc/mps2-fpgaio.h index XXXXXXX..XXXXXXX 100644 --- a/include/hw/misc/mps2-fpgaio.h +++ b/include/hw/misc/mps2-fpgaio.h @@ -XXX,XX +XXX,XX @@ typedef struct { uint32_t prescale; uint32_t misc; + /* QEMU_CLOCK_VIRTUAL time at which counter and pscntr were last synced */ + int64_t pscntr_sync_ticks; + /* Values of COUNTER and PSCNTR at time pscntr_sync_ticks */ + uint32_t counter; + uint32_t pscntr; + uint32_t prescale_clk; /* These hold the CLOCK_VIRTUAL ns tick when the CLK1HZ/CLK100HZ was zero */ diff --git a/hw/misc/mps2-fpgaio.c b/hw/misc/mps2-fpgaio.c index XXXXXXX..XXXXXXX 100644 --- a/hw/misc/mps2-fpgaio.c +++ b/hw/misc/mps2-fpgaio.c @@ -XXX,XX +XXX,XX @@ static int64_t tickoff_from_counter(int64_t now, uint32_t count, int frq) return now - muldiv64(count, NANOSECONDS_PER_SECOND, frq); } +static void resync_counter(MPS2FPGAIO *s) +{ + /* + * Update s->counter and s->pscntr to their true current values + * by calculating how many times PSCNTR has ticked since the + * last time we did a resync. + */ + int64_t now = qemu_clock_get_ns(QEMU_CLOCK_VIRTUAL); + int64_t elapsed = now - s->pscntr_sync_ticks; + + /* + * Round elapsed down to a whole number of PSCNTR ticks, so we don't + * lose time if we do multiple resyncs in a single tick. + */ + uint64_t ticks = muldiv64(elapsed, s->prescale_clk, NANOSECONDS_PER_SECOND); + + /* + * Work out what PSCNTR and COUNTER have moved to. We assume that + * PSCNTR reloads from PRESCALE one tick-period after it hits zero, + * and that COUNTER increments at the same moment. + */ + if (ticks == 0) { + /* We haven't ticked since the last time we were asked */ + return; + } else if (ticks < s->pscntr) { + /* We haven't yet reached zero, just reduce the PSCNTR */ + s->pscntr -= ticks; + } else { + if (s->prescale == 0) { + /* + * If the reload value is zero then the PSCNTR will stick + * at zero once it reaches it, and so we will increment + * COUNTER every tick after that. + */ + s->counter += ticks - s->pscntr; + s->pscntr = 0; + } else { + /* + * This is the complicated bit. This ASCII art diagram gives an + * example with PRESCALE==5 PSCNTR==7: + * + * ticks 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 + * PSCNTR 7 6 5 4 3 2 1 0 5 4 3 2 1 0 5 + * cinc 1 2 + * y 0 1 2 3 4 5 6 7 8 9 10 11 12 + * x 0 1 2 3 4 5 0 1 2 3 4 5 0 + * + * where x = y % (s->prescale + 1) + * and so PSCNTR = s->prescale - x + * and COUNTER is incremented by y / (s->prescale + 1) + * + * The case where PSCNTR < PRESCALE works out the same, + * though we must be careful to calculate y as 64-bit unsigned + * for all parts of the expression. + * y < 0 is not possible because that implies ticks < s->pscntr. + */ + uint64_t y = ticks - s->pscntr + s->prescale; + s->pscntr = s->prescale - (y % (s->prescale + 1)); + s->counter += y / (s->prescale + 1); + } + } + + /* + * Only advance the sync time to the timestamp of the last PSCNTR tick, + * not all the way to 'now', so we don't lose time if we do multiple + * resyncs in a single tick. + */ + s->pscntr_sync_ticks += muldiv64(ticks, NANOSECONDS_PER_SECOND, + s->prescale_clk); +} + static uint64_t mps2_fpgaio_read(void *opaque, hwaddr offset, unsigned size) { MPS2FPGAIO *s = MPS2_FPGAIO(opaque); @@ -XXX,XX +XXX,XX @@ static uint64_t mps2_fpgaio_read(void *opaque, hwaddr offset, unsigned size) r = counter_from_tickoff(now, s->clk100hz_tick_offset, 100); break; case A_COUNTER: + resync_counter(s); + r = s->counter; + break; case A_PSCNTR: - qemu_log_mask(LOG_UNIMP, "MPS2 FPGAIO: counters unimplemented\n"); - r = 0; + resync_counter(s); + r = s->pscntr; break; default: qemu_log_mask(LOG_GUEST_ERROR, @@ -XXX,XX +XXX,XX @@ static void mps2_fpgaio_write(void *opaque, hwaddr offset, uint64_t value, s->led0 = value & 0x3; break; case A_PRESCALE: + resync_counter(s); s->prescale = value; break; case A_MISC: @@ -XXX,XX +XXX,XX @@ static void mps2_fpgaio_write(void *opaque, hwaddr offset, uint64_t value, now = qemu_clock_get_ns(QEMU_CLOCK_VIRTUAL); s->clk100hz_tick_offset = tickoff_from_counter(now, value, 100); break; + case A_COUNTER: + resync_counter(s); + s->counter = value; + break; + case A_PSCNTR: + resync_counter(s); + s->pscntr = value; + break; default: qemu_log_mask(LOG_GUEST_ERROR, "MPS2 FPGAIO write: bad offset 0x%x\n", (int) offset); @@ -XXX,XX +XXX,XX @@ static void mps2_fpgaio_reset(DeviceState *dev) s->misc = 0; s->clk1hz_tick_offset = tickoff_from_counter(now, 0, 1); s->clk100hz_tick_offset = tickoff_from_counter(now, 0, 100); + s->counter = 0; + s->pscntr = 0; + s->pscntr_sync_ticks = now; } static void mps2_fpgaio_init(Object *obj) @@ -XXX,XX +XXX,XX @@ static bool mps2_fpgaio_counters_needed(void *opaque) static const VMStateDescription mps2_fpgaio_counters_vmstate = { .name = "mps2-fpgaio/counters", - .version_id = 1, - .minimum_version_id = 1, + .version_id = 2, + .minimum_version_id = 2, .needed = mps2_fpgaio_counters_needed, .fields = (VMStateField[]) { VMSTATE_INT64(clk1hz_tick_offset, MPS2FPGAIO), VMSTATE_INT64(clk100hz_tick_offset, MPS2FPGAIO), + VMSTATE_UINT32(counter, MPS2FPGAIO), + VMSTATE_UINT32(pscntr, MPS2FPGAIO), + VMSTATE_INT64(pscntr_sync_ticks, MPS2FPGAIO), VMSTATE_END_OF_LIST() } }; -- 2.18.0 The Arm Cortex-M System Design Kit includes a "dual-input timer module" which combines two programmable down-counters. Implement a model of this device. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20180820141116.9118-4-peter.maydell@linaro.org --- hw/timer/Makefile.objs | 1 + include/hw/timer/cmsdk-apb-dualtimer.h | 72 ++++ hw/timer/cmsdk-apb-dualtimer.c | 515 +++++++++++++++++++++++++ MAINTAINERS | 2 + default-configs/arm-softmmu.mak | 1 + hw/timer/trace-events | 5 + 6 files changed, 596 insertions(+) create mode 100644 include/hw/timer/cmsdk-apb-dualtimer.h create mode 100644 hw/timer/cmsdk-apb-dualtimer.c diff --git a/hw/timer/Makefile.objs b/hw/timer/Makefile.objs index XXXXXXX..XXXXXXX 100644 --- a/hw/timer/Makefile.objs +++ b/hw/timer/Makefile.objs @@ -XXX,XX +XXX,XX @@ common-obj-$(CONFIG_ASPEED_SOC) += aspeed_timer.o common-obj-$(CONFIG_SUN4V_RTC) += sun4v-rtc.o common-obj-$(CONFIG_CMSDK_APB_TIMER) += cmsdk-apb-timer.o +common-obj-$(CONFIG_CMSDK_APB_DUALTIMER) += cmsdk-apb-dualtimer.o common-obj-$(CONFIG_MSF2) += mss-timer.o diff --git a/include/hw/timer/cmsdk-apb-dualtimer.h b/include/hw/timer/cmsdk-apb-dualtimer.h new file mode 100644 index XXXXXXX..XXXXXXX --- /dev/null +++ b/include/hw/timer/cmsdk-apb-dualtimer.h @@ -XXX,XX +XXX,XX @@ +/* + * ARM CMSDK APB dual-timer emulation + * + * Copyright (c) 2018 Linaro Limited + * Written by Peter Maydell + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License version 2 or + * (at your option) any later version. + */ + +/* + * This is a model of the "APB dual-input timer" which is part of the Cortex-M + * System Design Kit (CMSDK) and documented in the Cortex-M System + * Design Kit Technical Reference Manual (ARM DDI0479C): + * https://developer.arm.com/products/system-design/system-design-kits/cortex-m-system-design-kit + * + * QEMU interface: + * + QOM property "pclk-frq": frequency at which the timer is clocked + * + sysbus MMIO region 0: the register bank + * + sysbus IRQ 0: combined timer interrupt TIMINTC + * + sysbus IRO 1: timer block 1 interrupt TIMINT1 + * + sysbus IRQ 2: timer block 2 interrupt TIMINT2 + */ + +#ifndef CMSDK_APB_DUALTIMER_H +#define CMSDK_APB_DUALTIMER_H + +#include "hw/sysbus.h" +#include "hw/ptimer.h" + +#define TYPE_CMSDK_APB_DUALTIMER "cmsdk-apb-dualtimer" +#define CMSDK_APB_DUALTIMER(obj) OBJECT_CHECK(CMSDKAPBDualTimer, (obj), \ + TYPE_CMSDK_APB_DUALTIMER) + +typedef struct CMSDKAPBDualTimer CMSDKAPBDualTimer; + +/* One of the two identical timer modules in the dual-timer module */ +typedef struct CMSDKAPBDualTimerModule { + CMSDKAPBDualTimer *parent; + struct ptimer_state *timer; + qemu_irq timerint; + /* + * We must track the guest LOAD and VALUE register state by hand + * rather than leaving this state only in the ptimer limit/count, + * because if CONTROL.SIZE is 0 then only the low 16 bits of the + * counter actually counts, but the high half is still guest + * accessible. + */ + uint32_t load; + uint32_t value; + uint32_t control; + uint32_t intstatus; +} CMSDKAPBDualTimerModule; + +#define CMSDK_APB_DUALTIMER_NUM_MODULES 2 + +struct CMSDKAPBDualTimer { + /*< private >*/ + SysBusDevice parent_obj; + + /*< public >*/ + MemoryRegion iomem; + qemu_irq timerintc; + uint32_t pclk_frq; + + CMSDKAPBDualTimerModule timermod[CMSDK_APB_DUALTIMER_NUM_MODULES]; + uint32_t timeritcr; + uint32_t timeritop; +}; + +#endif diff --git a/hw/timer/cmsdk-apb-dualtimer.c b/hw/timer/cmsdk-apb-dualtimer.c new file mode 100644 index XXXXXXX..XXXXXXX --- /dev/null +++ b/hw/timer/cmsdk-apb-dualtimer.c @@ -XXX,XX +XXX,XX @@ +/* + * ARM CMSDK APB dual-timer emulation + * + * Copyright (c) 2018 Linaro Limited + * Written by Peter Maydell + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License version 2 or + * (at your option) any later version. + */ + +/* + * This is a model of the "APB dual-input timer" which is part of the Cortex-M + * System Design Kit (CMSDK) and documented in the Cortex-M System + * Design Kit Technical Reference Manual (ARM DDI0479C): + * https://developer.arm.com/products/system-design/system-design-kits/cortex-m-system-design-kit + */ + +#include "qemu/osdep.h" +#include "qemu/log.h" +#include "trace.h" +#include "qapi/error.h" +#include "qemu/main-loop.h" +#include "hw/sysbus.h" +#include "hw/registerfields.h" +#include "hw/timer/cmsdk-apb-dualtimer.h" + +REG32(TIMER1LOAD, 0x0) +REG32(TIMER1VALUE, 0x4) +REG32(TIMER1CONTROL, 0x8) + FIELD(CONTROL, ONESHOT, 0, 1) + FIELD(CONTROL, SIZE, 1, 1) + FIELD(CONTROL, PRESCALE, 2, 2) + FIELD(CONTROL, INTEN, 5, 1) + FIELD(CONTROL, MODE, 6, 1) + FIELD(CONTROL, ENABLE, 7, 1) +#define R_CONTROL_VALID_MASK (R_CONTROL_ONESHOT_MASK | R_CONTROL_SIZE_MASK | \ + R_CONTROL_PRESCALE_MASK | R_CONTROL_INTEN_MASK | \ + R_CONTROL_MODE_MASK | R_CONTROL_ENABLE_MASK) +REG32(TIMER1INTCLR, 0xc) +REG32(TIMER1RIS, 0x10) +REG32(TIMER1MIS, 0x14) +REG32(TIMER1BGLOAD, 0x18) +REG32(TIMER2LOAD, 0x20) +REG32(TIMER2VALUE, 0x24) +REG32(TIMER2CONTROL, 0x28) +REG32(TIMER2INTCLR, 0x2c) +REG32(TIMER2RIS, 0x30) +REG32(TIMER2MIS, 0x34) +REG32(TIMER2BGLOAD, 0x38) +REG32(TIMERITCR, 0xf00) + FIELD(TIMERITCR, ENABLE, 0, 1) +#define R_TIMERITCR_VALID_MASK R_TIMERITCR_ENABLE_MASK +REG32(TIMERITOP, 0xf04) + FIELD(TIMERITOP, TIMINT1, 0, 1) + FIELD(TIMERITOP, TIMINT2, 1, 1) +#define R_TIMERITOP_VALID_MASK (R_TIMERITOP_TIMINT1_MASK | \ + R_TIMERITOP_TIMINT2_MASK) +REG32(PID4, 0xfd0) +REG32(PID5, 0xfd4) +REG32(PID6, 0xfd8) +REG32(PID7, 0xfdc) +REG32(PID0, 0xfe0) +REG32(PID1, 0xfe4) +REG32(PID2, 0xfe8) +REG32(PID3, 0xfec) +REG32(CID0, 0xff0) +REG32(CID1, 0xff4) +REG32(CID2, 0xff8) +REG32(CID3, 0xffc) + +/* PID/CID values */ +static const int timer_id[] = { + 0x04, 0x00, 0x00, 0x00, /* PID4..PID7 */ + 0x23, 0xb8, 0x1b, 0x00, /* PID0..PID3 */ + 0x0d, 0xf0, 0x05, 0xb1, /* CID0..CID3 */ +}; + +static bool cmsdk_dualtimermod_intstatus(CMSDKAPBDualTimerModule *m) +{ + /* Return masked interrupt status for the timer module */ + return m->intstatus && (m->control & R_CONTROL_INTEN_MASK); +} + +static void cmsdk_apb_dualtimer_update(CMSDKAPBDualTimer *s) +{ + bool timint1, timint2, timintc; + + if (s->timeritcr) { + /* Integration test mode: outputs driven directly from TIMERITOP bits */ + timint1 = s->timeritop & R_TIMERITOP_TIMINT1_MASK; + timint2 = s->timeritop & R_TIMERITOP_TIMINT2_MASK; + } else { + timint1 = cmsdk_dualtimermod_intstatus(&s->timermod[0]); + timint2 = cmsdk_dualtimermod_intstatus(&s->timermod[1]); + } + + timintc = timint1 || timint2; + + qemu_set_irq(s->timermod[0].timerint, timint1); + qemu_set_irq(s->timermod[1].timerint, timint2); + qemu_set_irq(s->timerintc, timintc); +} + +static void cmsdk_dualtimermod_write_control(CMSDKAPBDualTimerModule *m, + uint32_t newctrl) +{ + /* Handle a write to the CONTROL register */ + uint32_t changed; + + newctrl &= R_CONTROL_VALID_MASK; + + changed = m->control ^ newctrl; + + if (changed & ~newctrl & R_CONTROL_ENABLE_MASK) { + /* ENABLE cleared, stop timer before any further changes */ + ptimer_stop(m->timer); + } + + if (changed & R_CONTROL_PRESCALE_MASK) { + int divisor; + + switch (FIELD_EX32(newctrl, CONTROL, PRESCALE)) { + case 0: + divisor = 1; + break; + case 1: + divisor = 16; + break; + case 2: + divisor = 256; + break; + case 3: + /* UNDEFINED; complain, and arbitrarily treat like 2 */ + qemu_log_mask(LOG_GUEST_ERROR, + "CMSDK APB dual-timer: CONTROL.PRESCALE==0b11" + " is undefined behaviour\n"); + divisor = 256; + break; + default: + g_assert_not_reached(); + } + ptimer_set_freq(m->timer, m->parent->pclk_frq / divisor); + } + + if (changed & R_CONTROL_MODE_MASK) { + uint32_t load; + if (newctrl & R_CONTROL_MODE_MASK) { + /* Periodic: the limit is the LOAD register value */ + load = m->load; + } else { + /* Free-running: counter wraps around */ + load = ptimer_get_limit(m->timer); + if (!(m->control & R_CONTROL_SIZE_MASK)) { + load = deposit32(m->load, 0, 16, load); + } + m->load = load; + load = 0xffffffff; + } + if (!(m->control & R_CONTROL_SIZE_MASK)) { + load &= 0xffff; + } + ptimer_set_limit(m->timer, load, 0); + } + + if (changed & R_CONTROL_SIZE_MASK) { + /* Timer switched between 16 and 32 bit count */ + uint32_t value, load; + + value = ptimer_get_count(m->timer); + load = ptimer_get_limit(m->timer); + if (newctrl & R_CONTROL_SIZE_MASK) { + /* 16 -> 32, top half of VALUE is in struct field */ + value = deposit32(m->value, 0, 16, value); + } else { + /* 32 -> 16: save top half to struct field and truncate */ + m->value = value; + value &= 0xffff; + } + + if (newctrl & R_CONTROL_MODE_MASK) { + /* Periodic, timer limit has LOAD value */ + if (newctrl & R_CONTROL_SIZE_MASK) { + load = deposit32(m->load, 0, 16, load); + } else { + m->load = load; + load &= 0xffff; + } + } else { + /* Free-running, timer limit is set to give wraparound */ + if (newctrl & R_CONTROL_SIZE_MASK) { + load = 0xffffffff; + } else { + load = 0xffff; + } + } + ptimer_set_count(m->timer, value); + ptimer_set_limit(m->timer, load, 0); + } + + if (newctrl & R_CONTROL_ENABLE_MASK) { + /* + * ENABLE is set; start the timer after all other changes. + * We start it even if the ENABLE bit didn't actually change, + * in case the timer was an expired one-shot timer that has + * now been changed into a free-running or periodic timer. + */ + ptimer_run(m->timer, !!(newctrl & R_CONTROL_ONESHOT_MASK)); + } + + m->control = newctrl; +} + +static uint64_t cmsdk_apb_dualtimer_read(void *opaque, hwaddr offset, + unsigned size) +{ + CMSDKAPBDualTimer *s = CMSDK_APB_DUALTIMER(opaque); + uint64_t r; + + if (offset >= A_TIMERITCR) { + switch (offset) { + case A_TIMERITCR: + r = s->timeritcr; + break; + case A_PID4 ... A_CID3: + r = timer_id[(offset - A_PID4) / 4]; + break; + default: + bad_offset: + qemu_log_mask(LOG_GUEST_ERROR, + "CMSDK APB dual-timer read: bad offset %x\n", + (int) offset); + r = 0; + break; + } + } else { + int timer = offset >> 5; + CMSDKAPBDualTimerModule *m; + + if (timer >= ARRAY_SIZE(s->timermod)) { + goto bad_offset; + } + + m = &s->timermod[timer]; + + switch (offset & 0x1F) { + case A_TIMER1LOAD: + case A_TIMER1BGLOAD: + if (m->control & R_CONTROL_MODE_MASK) { + /* + * Periodic: the ptimer limit is the LOAD register value, (or + * just the low 16 bits of it if the timer is in 16-bit mode) + */ + r = ptimer_get_limit(m->timer); + if (!(m->control & R_CONTROL_SIZE_MASK)) { + r = deposit32(m->load, 0, 16, r); + } + } else { + /* Free-running: LOAD register value is just in m->load */ + r = m->load; + } + break; + case A_TIMER1VALUE: + r = ptimer_get_count(m->timer); + if (!(m->control & R_CONTROL_SIZE_MASK)) { + r = deposit32(m->value, 0, 16, r); + } + break; + case A_TIMER1CONTROL: + r = m->control; + break; + case A_TIMER1RIS: + r = m->intstatus; + break; + case A_TIMER1MIS: + r = cmsdk_dualtimermod_intstatus(m); + break; + default: + goto bad_offset; + } + } + + trace_cmsdk_apb_dualtimer_read(offset, r, size); + return r; +} + +static void cmsdk_apb_dualtimer_write(void *opaque, hwaddr offset, + uint64_t value, unsigned size) +{ + CMSDKAPBDualTimer *s = CMSDK_APB_DUALTIMER(opaque); + + trace_cmsdk_apb_dualtimer_write(offset, value, size); + + if (offset >= A_TIMERITCR) { + switch (offset) { + case A_TIMERITCR: + s->timeritcr = value & R_TIMERITCR_VALID_MASK; + cmsdk_apb_dualtimer_update(s); + case A_TIMERITOP: + s->timeritop = value & R_TIMERITOP_VALID_MASK; + cmsdk_apb_dualtimer_update(s); + default: + bad_offset: + qemu_log_mask(LOG_GUEST_ERROR, + "CMSDK APB dual-timer write: bad offset %x\n", + (int) offset); + break; + } + } else { + int timer = offset >> 5; + CMSDKAPBDualTimerModule *m; + + if (timer >= ARRAY_SIZE(s->timermod)) { + goto bad_offset; + } + + m = &s->timermod[timer]; + + switch (offset & 0x1F) { + case A_TIMER1LOAD: + /* Set the limit, and immediately reload the count from it */ + m->load = value; + m->value = value; + if (!(m->control & R_CONTROL_SIZE_MASK)) { + value &= 0xffff; + } + if (!(m->control & R_CONTROL_MODE_MASK)) { + /* + * In free-running mode this won't set the limit but will + * still change the current count value. + */ + ptimer_set_count(m->timer, value); + } else { + if (!value) { + ptimer_stop(m->timer); + } + ptimer_set_limit(m->timer, value, 1); + if (value && (m->control & R_CONTROL_ENABLE_MASK)) { + /* Force possibly-expired oneshot timer to restart */ + ptimer_run(m->timer, 1); + } + } + break; + case A_TIMER1BGLOAD: + /* Set the limit, but not the current count */ + m->load = value; + if (!(m->control & R_CONTROL_MODE_MASK)) { + /* In free-running mode there is no limit */ + break; + } + if (!(m->control & R_CONTROL_SIZE_MASK)) { + value &= 0xffff; + } + ptimer_set_limit(m->timer, value, 0); + break; + case A_TIMER1CONTROL: + cmsdk_dualtimermod_write_control(m, value); + cmsdk_apb_dualtimer_update(s); + break; + case A_TIMER1INTCLR: + m->intstatus = 0; + cmsdk_apb_dualtimer_update(s); + break; + default: + goto bad_offset; + } + } +} + +static const MemoryRegionOps cmsdk_apb_dualtimer_ops = { + .read = cmsdk_apb_dualtimer_read, + .write = cmsdk_apb_dualtimer_write, + .endianness = DEVICE_LITTLE_ENDIAN, + /* byte/halfword accesses are just zero-padded on reads and writes */ + .impl.min_access_size = 4, + .impl.max_access_size = 4, + .valid.min_access_size = 1, + .valid.max_access_size = 4, +}; + +static void cmsdk_dualtimermod_tick(void *opaque) +{ + CMSDKAPBDualTimerModule *m = opaque; + + m->intstatus = 1; + cmsdk_apb_dualtimer_update(m->parent); +} + +static void cmsdk_dualtimermod_reset(CMSDKAPBDualTimerModule *m) +{ + m->control = R_CONTROL_INTEN_MASK; + m->intstatus = 0; + m->load = 0; + m->value = 0xffffffff; + ptimer_stop(m->timer); + /* + * We start in free-running mode, with VALUE at 0xffffffff, and + * in 16-bit counter mode. This means that the ptimer count and + * limit must both be set to 0xffff, so we wrap at 16 bits. + */ + ptimer_set_limit(m->timer, 0xffff, 1); + ptimer_set_freq(m->timer, m->parent->pclk_frq); +} + +static void cmsdk_apb_dualtimer_reset(DeviceState *dev) +{ + CMSDKAPBDualTimer *s = CMSDK_APB_DUALTIMER(dev); + int i; + + trace_cmsdk_apb_dualtimer_reset(); + + for (i = 0; i < ARRAY_SIZE(s->timermod); i++) { + cmsdk_dualtimermod_reset(&s->timermod[i]); + } + s->timeritcr = 0; + s->timeritop = 0; +} + +static void cmsdk_apb_dualtimer_init(Object *obj) +{ + SysBusDevice *sbd = SYS_BUS_DEVICE(obj); + CMSDKAPBDualTimer *s = CMSDK_APB_DUALTIMER(obj); + int i; + + memory_region_init_io(&s->iomem, obj, &cmsdk_apb_dualtimer_ops, + s, "cmsdk-apb-dualtimer", 0x1000); + sysbus_init_mmio(sbd, &s->iomem); + sysbus_init_irq(sbd, &s->timerintc); + + for (i = 0; i < ARRAY_SIZE(s->timermod); i++) { + sysbus_init_irq(sbd, &s->timermod[i].timerint); + } +} + +static void cmsdk_apb_dualtimer_realize(DeviceState *dev, Error **errp) +{ + CMSDKAPBDualTimer *s = CMSDK_APB_DUALTIMER(dev); + int i; + + if (s->pclk_frq == 0) { + error_setg(errp, "CMSDK APB timer: pclk-frq property must be set"); + return; + } + + for (i = 0; i < ARRAY_SIZE(s->timermod); i++) { + CMSDKAPBDualTimerModule *m = &s->timermod[i]; + QEMUBH *bh = qemu_bh_new(cmsdk_dualtimermod_tick, m); + + m->parent = s; + m->timer = ptimer_init(bh, + PTIMER_POLICY_WRAP_AFTER_ONE_PERIOD | + PTIMER_POLICY_TRIGGER_ONLY_ON_DECREMENT | + PTIMER_POLICY_NO_IMMEDIATE_RELOAD | + PTIMER_POLICY_NO_COUNTER_ROUND_DOWN); + } +} + +static const VMStateDescription cmsdk_dualtimermod_vmstate = { + .name = "cmsdk-apb-dualtimer-module", + .version_id = 1, + .minimum_version_id = 1, + .fields = (VMStateField[]) { + VMSTATE_PTIMER(timer, CMSDKAPBDualTimerModule), + VMSTATE_UINT32(load, CMSDKAPBDualTimerModule), + VMSTATE_UINT32(value, CMSDKAPBDualTimerModule), + VMSTATE_UINT32(control, CMSDKAPBDualTimerModule), + VMSTATE_UINT32(intstatus, CMSDKAPBDualTimerModule), + VMSTATE_END_OF_LIST() + } +}; + +static const VMStateDescription cmsdk_apb_dualtimer_vmstate = { + .name = "cmsdk-apb-dualtimer", + .version_id = 1, + .minimum_version_id = 1, + .fields = (VMStateField[]) { + VMSTATE_STRUCT_ARRAY(timermod, CMSDKAPBDualTimer, + CMSDK_APB_DUALTIMER_NUM_MODULES, + 1, cmsdk_dualtimermod_vmstate, + CMSDKAPBDualTimerModule), + VMSTATE_UINT32(timeritcr, CMSDKAPBDualTimer), + VMSTATE_UINT32(timeritop, CMSDKAPBDualTimer), + VMSTATE_END_OF_LIST() + } +}; + +static Property cmsdk_apb_dualtimer_properties[] = { + DEFINE_PROP_UINT32("pclk-frq", CMSDKAPBDualTimer, pclk_frq, 0), + DEFINE_PROP_END_OF_LIST(), +}; + +static void cmsdk_apb_dualtimer_class_init(ObjectClass *klass, void *data) +{ + DeviceClass *dc = DEVICE_CLASS(klass); + + dc->realize = cmsdk_apb_dualtimer_realize; + dc->vmsd = &cmsdk_apb_dualtimer_vmstate; + dc->reset = cmsdk_apb_dualtimer_reset; + dc->props = cmsdk_apb_dualtimer_properties; +} + +static const TypeInfo cmsdk_apb_dualtimer_info = { + .name = TYPE_CMSDK_APB_DUALTIMER, + .parent = TYPE_SYS_BUS_DEVICE, + .instance_size = sizeof(CMSDKAPBDualTimer), + .instance_init = cmsdk_apb_dualtimer_init, + .class_init = cmsdk_apb_dualtimer_class_init, +}; + +static void cmsdk_apb_dualtimer_register_types(void) +{ + type_register_static(&cmsdk_apb_dualtimer_info); +} + +type_init(cmsdk_apb_dualtimer_register_types); diff --git a/MAINTAINERS b/MAINTAINERS index XXXXXXX..XXXXXXX 100644 --- a/MAINTAINERS +++ b/MAINTAINERS @@ -XXX,XX +XXX,XX @@ F: hw/timer/pl031.c F: include/hw/arm/primecell.h F: hw/timer/cmsdk-apb-timer.c F: include/hw/timer/cmsdk-apb-timer.h +F: hw/timer/cmsdk-apb-dualtimer.c +F: include/hw/timer/cmsdk-apb-dualtimer.h F: hw/char/cmsdk-apb-uart.c F: include/hw/char/cmsdk-apb-uart.h F: hw/watchdog/cmsdk-apb-watchdog.c diff --git a/default-configs/arm-softmmu.mak b/default-configs/arm-softmmu.mak index XXXXXXX..XXXXXXX 100644 --- a/default-configs/arm-softmmu.mak +++ b/default-configs/arm-softmmu.mak @@ -XXX,XX +XXX,XX @@ CONFIG_STM32F2XX_SPI=y CONFIG_STM32F205_SOC=y CONFIG_CMSDK_APB_TIMER=y +CONFIG_CMSDK_APB_DUALTIMER=y CONFIG_CMSDK_APB_UART=y CONFIG_CMSDK_APB_WATCHDOG=y diff --git a/hw/timer/trace-events b/hw/timer/trace-events index XXXXXXX..XXXXXXX 100644 --- a/hw/timer/trace-events +++ b/hw/timer/trace-events @@ -XXX,XX +XXX,XX @@ cmsdk_apb_timer_read(uint64_t offset, uint64_t data, unsigned size) "CMSDK APB t cmsdk_apb_timer_write(uint64_t offset, uint64_t data, unsigned size) "CMSDK APB timer write: offset 0x%" PRIx64 " data 0x%" PRIx64 " size %u" cmsdk_apb_timer_reset(void) "CMSDK APB timer: reset" +# hw/timer/cmsdk_apb_dualtimer.c +cmsdk_apb_dualtimer_read(uint64_t offset, uint64_t data, unsigned size) "CMSDK APB dualtimer read: offset 0x%" PRIx64 " data 0x%" PRIx64 " size %u" +cmsdk_apb_dualtimer_write(uint64_t offset, uint64_t data, unsigned size) "CMSDK APB dualtimer write: offset 0x%" PRIx64 " data 0x%" PRIx64 " size %u" +cmsdk_apb_dualtimer_reset(void) "CMSDK APB dualtimer: reset" + # hw/timer/xlnx-zynqmp-rtc.c xlnx_zynqmp_rtc_gettime(int year, int month, int day, int hour, int min, int sec) "Get time from host: %d-%d-%d %2d:%02d:%02d" -- 2.18.0 Now we have a model of the CMSDK dual timer, we can wire it up in the IoTKit. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20180820141116.9118-5-peter.maydell@linaro.org --- include/hw/arm/iotkit.h | 3 ++- hw/arm/iotkit.c | 8 +++++--- 2 files changed, 7 insertions(+), 4 deletions(-) diff --git a/include/hw/arm/iotkit.h b/include/hw/arm/iotkit.h index XXXXXXX..XXXXXXX 100644 --- a/include/hw/arm/iotkit.h +++ b/include/hw/arm/iotkit.h @@ -XXX,XX +XXX,XX @@ #include "hw/misc/tz-ppc.h" #include "hw/misc/tz-mpc.h" #include "hw/timer/cmsdk-apb-timer.h" +#include "hw/timer/cmsdk-apb-dualtimer.h" #include "hw/misc/unimp.h" #include "hw/or-irq.h" #include "hw/core/split-irq.h" @@ -XXX,XX +XXX,XX @@ typedef struct IoTKit { SplitIRQ mpc_irq_splitter[IOTS_NUM_EXP_MPC + IOTS_NUM_MPC]; qemu_or_irq mpc_irq_orgate; - UnimplementedDeviceState dualtimer; + CMSDKAPBDualTimer dualtimer; UnimplementedDeviceState s32ktimer; MemoryRegion container; diff --git a/hw/arm/iotkit.c b/hw/arm/iotkit.c index XXXXXXX..XXXXXXX 100644 --- a/hw/arm/iotkit.c +++ b/hw/arm/iotkit.c @@ -XXX,XX +XXX,XX @@ static void iotkit_init(Object *obj) sysbus_init_child_obj(obj, "timer1", &s->timer1, sizeof(s->timer1), TYPE_CMSDK_APB_TIMER); sysbus_init_child_obj(obj, "dualtimer", &s->dualtimer, sizeof(s->dualtimer), - TYPE_UNIMPLEMENTED_DEVICE); + TYPE_CMSDK_APB_DUALTIMER); object_initialize_child(obj, "ppc-irq-orgate", &s->ppc_irq_orgate, sizeof(s->ppc_irq_orgate), TYPE_OR_IRQ, &error_abort, NULL); @@ -XXX,XX +XXX,XX @@ static void iotkit_realize(DeviceState *dev, Error **errp) return; } - qdev_prop_set_string(DEVICE(&s->dualtimer), "name", "Dual timer"); - qdev_prop_set_uint64(DEVICE(&s->dualtimer), "size", 0x1000); + + qdev_prop_set_uint32(DEVICE(&s->dualtimer), "pclk-frq", s->mainclk_frq); object_property_set_bool(OBJECT(&s->dualtimer), true, "realized", &err); if (err) { error_propagate(errp, err); return; } + sysbus_connect_irq(SYS_BUS_DEVICE(&s->dualtimer), 0, + qdev_get_gpio_in(DEVICE(&s->armv7m), 5)); mr = sysbus_mmio_get_region(SYS_BUS_DEVICE(&s->dualtimer), 0); object_property_set_link(OBJECT(&s->apb_ppc0), OBJECT(mr), "port[2]", &err); if (err) { -- 2.18.0 The MPS2 FPGA images for the Cortex-M3 (mps2-an385 and mps2-511) both include a CMSDK dual-timer module. Wire this up. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20180820141116.9118-6-peter.maydell@linaro.org --- hw/arm/mps2.c | 11 +++++++++++ 1 file changed, 11 insertions(+) diff --git a/hw/arm/mps2.c b/hw/arm/mps2.c index XXXXXXX..XXXXXXX 100644 --- a/hw/arm/mps2.c +++ b/hw/arm/mps2.c @@ -XXX,XX +XXX,XX @@ #include "hw/misc/unimp.h" #include "hw/char/cmsdk-apb-uart.h" #include "hw/timer/cmsdk-apb-timer.h" +#include "hw/timer/cmsdk-apb-dualtimer.h" #include "hw/misc/mps2-scc.h" #include "hw/devices.h" #include "net/net.h" @@ -XXX,XX +XXX,XX @@ typedef struct { MemoryRegion blockram_m3; MemoryRegion sram; MPS2SCC scc; + CMSDKAPBDualTimer dualtimer; } MPS2MachineState; #define TYPE_MPS2_MACHINE "mps2" @@ -XXX,XX +XXX,XX @@ static void mps2_common_init(MachineState *machine) cmsdk_apb_timer_create(0x40000000, qdev_get_gpio_in(armv7m, 8), SYSCLK_FRQ); cmsdk_apb_timer_create(0x40001000, qdev_get_gpio_in(armv7m, 9), SYSCLK_FRQ); + sysbus_init_child_obj(OBJECT(mms), "dualtimer", &mms->dualtimer, + sizeof(mms->dualtimer), TYPE_CMSDK_APB_DUALTIMER); + qdev_prop_set_uint32(DEVICE(&mms->dualtimer), "pclk-frq", SYSCLK_FRQ); + object_property_set_bool(OBJECT(&mms->dualtimer), true, "realized", + &error_fatal); + sysbus_connect_irq(SYS_BUS_DEVICE(&mms->dualtimer), 0, + qdev_get_gpio_in(armv7m, 10)); + sysbus_mmio_map(SYS_BUS_DEVICE(&mms->dualtimer), 0, 0x40002000); + object_initialize(&mms->scc, sizeof(mms->scc), TYPE_MPS2_SCC); sccdev = DEVICE(&mms->scc); qdev_set_parent_bus(sccdev, sysbus_get_default()); -- 2.18.0 The IoTKit includes three different instances of the CMSDK APB watchdog; create and wire them up. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20180820141116.9118-7-peter.maydell@linaro.org --- include/hw/arm/iotkit.h | 6 +++++ hw/arm/iotkit.c | 58 ++++++++++++++++++++++++++++++++++++++--- 2 files changed, 61 insertions(+), 3 deletions(-) diff --git a/include/hw/arm/iotkit.h b/include/hw/arm/iotkit.h index XXXXXXX..XXXXXXX 100644 --- a/include/hw/arm/iotkit.h +++ b/include/hw/arm/iotkit.h @@ -XXX,XX +XXX,XX @@ #include "hw/misc/tz-mpc.h" #include "hw/timer/cmsdk-apb-timer.h" #include "hw/timer/cmsdk-apb-dualtimer.h" +#include "hw/watchdog/cmsdk-apb-watchdog.h" #include "hw/misc/unimp.h" #include "hw/or-irq.h" #include "hw/core/split-irq.h" @@ -XXX,XX +XXX,XX @@ typedef struct IoTKit { SplitIRQ ppc_irq_splitter[NUM_PPCS]; SplitIRQ mpc_irq_splitter[IOTS_NUM_EXP_MPC + IOTS_NUM_MPC]; qemu_or_irq mpc_irq_orgate; + qemu_or_irq nmi_orgate; CMSDKAPBDualTimer dualtimer; UnimplementedDeviceState s32ktimer; + CMSDKAPBWatchdog s32kwatchdog; + CMSDKAPBWatchdog nswatchdog; + CMSDKAPBWatchdog swatchdog; + MemoryRegion container; MemoryRegion alias1; MemoryRegion alias2; diff --git a/hw/arm/iotkit.c b/hw/arm/iotkit.c index XXXXXXX..XXXXXXX 100644 --- a/hw/arm/iotkit.c +++ b/hw/arm/iotkit.c @@ -XXX,XX +XXX,XX @@ #include "hw/misc/unimp.h" #include "hw/arm/arm.h" +/* Clock frequency in HZ of the 32KHz "slow clock" */ +#define S32KCLK (32 * 1000) + /* Create an alias region of @size bytes starting at @base * which mirrors the memory starting at @orig. */ @@ -XXX,XX +XXX,XX @@ static void iotkit_init(Object *obj) TYPE_CMSDK_APB_TIMER); sysbus_init_child_obj(obj, "dualtimer", &s->dualtimer, sizeof(s->dualtimer), TYPE_CMSDK_APB_DUALTIMER); + sysbus_init_child_obj(obj, "s32kwatchdog", &s->s32kwatchdog, + sizeof(s->s32kwatchdog), TYPE_CMSDK_APB_WATCHDOG); + sysbus_init_child_obj(obj, "nswatchdog", &s->nswatchdog, + sizeof(s->nswatchdog), TYPE_CMSDK_APB_WATCHDOG); + sysbus_init_child_obj(obj, "swatchdog", &s->swatchdog, + sizeof(s->swatchdog), TYPE_CMSDK_APB_WATCHDOG); + object_initialize_child(obj, "nmi-orgate", &s->nmi_orgate, + sizeof(s->nmi_orgate), TYPE_OR_IRQ, + &error_abort, NULL); object_initialize_child(obj, "ppc-irq-orgate", &s->ppc_irq_orgate, sizeof(s->ppc_irq_orgate), TYPE_OR_IRQ, &error_abort, NULL); @@ -XXX,XX +XXX,XX @@ static void iotkit_realize(DeviceState *dev, Error **errp) create_unimplemented_device("SYSINFO", 0x40020000, 0x1000); create_unimplemented_device("SYSCONTROL", 0x50021000, 0x1000); - create_unimplemented_device("S32KWATCHDOG", 0x5002e000, 0x1000); + + /* This OR gate wires together outputs from the secure watchdogs to NMI */ + object_property_set_int(OBJECT(&s->nmi_orgate), 2, "num-lines", &err); + if (err) { + error_propagate(errp, err); + return; + } + object_property_set_bool(OBJECT(&s->nmi_orgate), true, "realized", &err); + if (err) { + error_propagate(errp, err); + return; + } + qdev_connect_gpio_out(DEVICE(&s->nmi_orgate), 0, + qdev_get_gpio_in_named(DEVICE(&s->armv7m), "NMI", 0)); + + qdev_prop_set_uint32(DEVICE(&s->s32kwatchdog), "wdogclk-frq", S32KCLK); + object_property_set_bool(OBJECT(&s->s32kwatchdog), true, "realized", &err); + if (err) { + error_propagate(errp, err); + return; + } + sysbus_connect_irq(SYS_BUS_DEVICE(&s->s32kwatchdog), 0, + qdev_get_gpio_in(DEVICE(&s->nmi_orgate), 0)); + sysbus_mmio_map(SYS_BUS_DEVICE(&s->s32kwatchdog), 0, 0x5002e000); /* 0x40080000 .. 0x4008ffff : IoTKit second Base peripheral region */ - create_unimplemented_device("NS watchdog", 0x40081000, 0x1000); - create_unimplemented_device("S watchdog", 0x50081000, 0x1000); + qdev_prop_set_uint32(DEVICE(&s->nswatchdog), "wdogclk-frq", s->mainclk_frq); + object_property_set_bool(OBJECT(&s->nswatchdog), true, "realized", &err); + if (err) { + error_propagate(errp, err); + return; + } + sysbus_connect_irq(SYS_BUS_DEVICE(&s->nswatchdog), 0, + qdev_get_gpio_in(DEVICE(&s->armv7m), 1)); + sysbus_mmio_map(SYS_BUS_DEVICE(&s->nswatchdog), 0, 0x40081000); + + qdev_prop_set_uint32(DEVICE(&s->swatchdog), "wdogclk-frq", s->mainclk_frq); + object_property_set_bool(OBJECT(&s->swatchdog), true, "realized", &err); + if (err) { + error_propagate(errp, err); + return; + } + sysbus_connect_irq(SYS_BUS_DEVICE(&s->swatchdog), 0, + qdev_get_gpio_in(DEVICE(&s->nmi_orgate), 1)); + sysbus_mmio_map(SYS_BUS_DEVICE(&s->swatchdog), 0, 0x50081000); for (i = 0; i < ARRAY_SIZE(s->ppc_irq_splitter); i++) { Object *splitter = OBJECT(&s->ppc_irq_splitter[i]); -- 2.18.0 The IoTKit has a CMSDK timer device that runs on the S32KCLK. Create this and wire it up. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20180820141116.9118-8-peter.maydell@linaro.org --- include/hw/arm/iotkit.h | 2 +- hw/arm/iotkit.c | 9 +++++---- 2 files changed, 6 insertions(+), 5 deletions(-) diff --git a/include/hw/arm/iotkit.h b/include/hw/arm/iotkit.h index XXXXXXX..XXXXXXX 100644 --- a/include/hw/arm/iotkit.h +++ b/include/hw/arm/iotkit.h @@ -XXX,XX +XXX,XX @@ typedef struct IoTKit { TZMPC mpc; CMSDKAPBTIMER timer0; CMSDKAPBTIMER timer1; + CMSDKAPBTIMER s32ktimer; qemu_or_irq ppc_irq_orgate; SplitIRQ sec_resp_splitter; SplitIRQ ppc_irq_splitter[NUM_PPCS]; @@ -XXX,XX +XXX,XX @@ typedef struct IoTKit { qemu_or_irq nmi_orgate; CMSDKAPBDualTimer dualtimer; - UnimplementedDeviceState s32ktimer; CMSDKAPBWatchdog s32kwatchdog; CMSDKAPBWatchdog nswatchdog; diff --git a/hw/arm/iotkit.c b/hw/arm/iotkit.c index XXXXXXX..XXXXXXX 100644 --- a/hw/arm/iotkit.c +++ b/hw/arm/iotkit.c @@ -XXX,XX +XXX,XX @@ static void iotkit_init(Object *obj) TYPE_CMSDK_APB_TIMER); sysbus_init_child_obj(obj, "timer1", &s->timer1, sizeof(s->timer1), TYPE_CMSDK_APB_TIMER); + sysbus_init_child_obj(obj, "s32ktimer", &s->s32ktimer, sizeof(s->s32ktimer), + TYPE_CMSDK_APB_TIMER); sysbus_init_child_obj(obj, "dualtimer", &s->dualtimer, sizeof(s->dualtimer), TYPE_CMSDK_APB_DUALTIMER); sysbus_init_child_obj(obj, "s32kwatchdog", &s->s32kwatchdog, @@ -XXX,XX +XXX,XX @@ static void iotkit_init(Object *obj) TYPE_SPLIT_IRQ, &error_abort, NULL); g_free(name); } - sysbus_init_child_obj(obj, "s32ktimer", &s->s32ktimer, sizeof(s->s32ktimer), - TYPE_UNIMPLEMENTED_DEVICE); } static void iotkit_exp_irq(void *opaque, int n, int level) @@ -XXX,XX +XXX,XX @@ static void iotkit_realize(DeviceState *dev, Error **errp) /* Devices behind APB PPC1: * 0x4002f000: S32K timer */ - qdev_prop_set_string(DEVICE(&s->s32ktimer), "name", "S32KTIMER"); - qdev_prop_set_uint64(DEVICE(&s->s32ktimer), "size", 0x1000); + qdev_prop_set_uint32(DEVICE(&s->s32ktimer), "pclk-frq", S32KCLK); object_property_set_bool(OBJECT(&s->s32ktimer), true, "realized", &err); if (err) { error_propagate(errp, err); return; } + sysbus_connect_irq(SYS_BUS_DEVICE(&s->s32ktimer), 0, + qdev_get_gpio_in(DEVICE(&s->armv7m), 2)); mr = sysbus_mmio_get_region(SYS_BUS_DEVICE(&s->s32ktimer), 0); object_property_set_link(OBJECT(&s->apb_ppc1), OBJECT(mr), "port[0]", &err); if (err) { -- 2.18.0 The Arm IoTKit includes a system control element which provides a block of read-only ID registers and a block of read-write control registers. Implement a minimal version of this. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20180820141116.9118-9-peter.maydell@linaro.org --- hw/misc/Makefile.objs | 1 + include/hw/misc/iotkit-sysctl.h | 49 ++++++ hw/misc/iotkit-sysctl.c | 261 ++++++++++++++++++++++++++++++++ MAINTAINERS | 2 + default-configs/arm-softmmu.mak | 1 + hw/misc/trace-events | 7 + 6 files changed, 321 insertions(+) create mode 100644 include/hw/misc/iotkit-sysctl.h create mode 100644 hw/misc/iotkit-sysctl.c diff --git a/hw/misc/Makefile.objs b/hw/misc/Makefile.objs index XXXXXXX..XXXXXXX 100644 --- a/hw/misc/Makefile.objs +++ b/hw/misc/Makefile.objs @@ -XXX,XX +XXX,XX @@ obj-$(CONFIG_MPS2_SCC) += mps2-scc.o obj-$(CONFIG_TZ_MPC) += tz-mpc.o obj-$(CONFIG_TZ_PPC) += tz-ppc.o obj-$(CONFIG_IOTKIT_SECCTL) += iotkit-secctl.o +obj-$(CONFIG_IOTKIT_SYSCTL) += iotkit-sysctl.o obj-$(CONFIG_PVPANIC) += pvpanic.o obj-$(CONFIG_HYPERV_TESTDEV) += hyperv_testdev.o diff --git a/include/hw/misc/iotkit-sysctl.h b/include/hw/misc/iotkit-sysctl.h new file mode 100644 index XXXXXXX..XXXXXXX --- /dev/null +++ b/include/hw/misc/iotkit-sysctl.h @@ -XXX,XX +XXX,XX @@ +/* + * ARM IoTKit system control element + * + * Copyright (c) 2018 Linaro Limited + * Written by Peter Maydell + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License version 2 or + * (at your option) any later version. + */ + +/* + * This is a model of the "system control element" which is part of the + * Arm IoTKit and documented in + * http://infocenter.arm.com/help/index.jsp?topic=/com.arm.doc.ecm0601256/index.html + * Specifically, it implements the "system information block" and + * "system control register" blocks. + * + * QEMU interface: + * + sysbus MMIO region 0: the system information register bank + * + sysbus MMIO region 1: the system control register bank + */ + +#ifndef HW_MISC_IOTKIT_SYSCTL_H +#define HW_MISC_IOTKIT_SYSCTL_H + +#include "hw/sysbus.h" + +#define TYPE_IOTKIT_SYSCTL "iotkit-sysctl" +#define IOTKIT_SYSCTL(obj) OBJECT_CHECK(IoTKitSysCtl, (obj), \ + TYPE_IOTKIT_SYSCTL) + +typedef struct IoTKitSysCtl { + /*< private >*/ + SysBusDevice parent_obj; + + /*< public >*/ + MemoryRegion iomem; + + uint32_t secure_debug; + uint32_t reset_syndrome; + uint32_t reset_mask; + uint32_t gretreg; + uint32_t initsvrtor0; + uint32_t cpuwait; + uint32_t wicctrl; +} IoTKitSysCtl; + +#endif diff --git a/hw/misc/iotkit-sysctl.c b/hw/misc/iotkit-sysctl.c new file mode 100644 index XXXXXXX..XXXXXXX --- /dev/null +++ b/hw/misc/iotkit-sysctl.c @@ -XXX,XX +XXX,XX @@ +/* + * ARM IoTKit system control element + * + * Copyright (c) 2018 Linaro Limited + * Written by Peter Maydell + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License version 2 or + * (at your option) any later version. + */ + +/* + * This is a model of the "system control element" which is part of the + * Arm IoTKit and documented in + * http://infocenter.arm.com/help/index.jsp?topic=/com.arm.doc.ecm0601256/index.html + * Specifically, it implements the "system control register" blocks. + */ + +#include "qemu/osdep.h" +#include "qemu/log.h" +#include "trace.h" +#include "qapi/error.h" +#include "sysemu/sysemu.h" +#include "hw/sysbus.h" +#include "hw/registerfields.h" +#include "hw/misc/iotkit-sysctl.h" + +REG32(SECDBGSTAT, 0x0) +REG32(SECDBGSET, 0x4) +REG32(SECDBGCLR, 0x8) +REG32(RESET_SYNDROME, 0x100) +REG32(RESET_MASK, 0x104) +REG32(SWRESET, 0x108) + FIELD(SWRESET, SWRESETREQ, 9, 1) +REG32(GRETREG, 0x10c) +REG32(INITSVRTOR0, 0x110) +REG32(CPUWAIT, 0x118) +REG32(BUSWAIT, 0x11c) +REG32(WICCTRL, 0x120) +REG32(PID4, 0xfd0) +REG32(PID5, 0xfd4) +REG32(PID6, 0xfd8) +REG32(PID7, 0xfdc) +REG32(PID0, 0xfe0) +REG32(PID1, 0xfe4) +REG32(PID2, 0xfe8) +REG32(PID3, 0xfec) +REG32(CID0, 0xff0) +REG32(CID1, 0xff4) +REG32(CID2, 0xff8) +REG32(CID3, 0xffc) + +/* PID/CID values */ +static const int sysctl_id[] = { + 0x04, 0x00, 0x00, 0x00, /* PID4..PID7 */ + 0x54, 0xb8, 0x0b, 0x00, /* PID0..PID3 */ + 0x0d, 0xf0, 0x05, 0xb1, /* CID0..CID3 */ +}; + +static uint64_t iotkit_sysctl_read(void *opaque, hwaddr offset, + unsigned size) +{ + IoTKitSysCtl *s = IOTKIT_SYSCTL(opaque); + uint64_t r; + + switch (offset) { + case A_SECDBGSTAT: + r = s->secure_debug; + break; + case A_RESET_SYNDROME: + r = s->reset_syndrome; + break; + case A_RESET_MASK: + r = s->reset_mask; + break; + case A_GRETREG: + r = s->gretreg; + break; + case A_INITSVRTOR0: + r = s->initsvrtor0; + break; + case A_CPUWAIT: + r = s->cpuwait; + break; + case A_BUSWAIT: + /* In IoTKit BUSWAIT is reserved, R/O, zero */ + r = 0; + break; + case A_WICCTRL: + r = s->wicctrl; + break; + case A_PID4 ... A_CID3: + r = sysctl_id[(offset - A_PID4) / 4]; + break; + case A_SECDBGSET: + case A_SECDBGCLR: + case A_SWRESET: + qemu_log_mask(LOG_GUEST_ERROR, + "IoTKit SysCtl read: read of WO offset %x\n", + (int)offset); + r = 0; + break; + default: + qemu_log_mask(LOG_GUEST_ERROR, + "IoTKit SysCtl read: bad offset %x\n", (int)offset); + r = 0; + break; + } + trace_iotkit_sysctl_read(offset, r, size); + return r; +} + +static void iotkit_sysctl_write(void *opaque, hwaddr offset, + uint64_t value, unsigned size) +{ + IoTKitSysCtl *s = IOTKIT_SYSCTL(opaque); + + trace_iotkit_sysctl_write(offset, value, size); + + /* + * Most of the state here has to do with control of reset and + * similar kinds of power up -- for instance the guest can ask + * what the reason for the last reset was, or forbid reset for + * some causes (like the non-secure watchdog). Most of this is + * not relevant to QEMU, which doesn't really model anything other + * than a full power-on reset. + * We just model the registers as reads-as-written. + */ + + switch (offset) { + case A_RESET_SYNDROME: + qemu_log_mask(LOG_UNIMP, + "IoTKit SysCtl RESET_SYNDROME unimplemented\n"); + s->reset_syndrome = value; + break; + case A_RESET_MASK: + qemu_log_mask(LOG_UNIMP, "IoTKit SysCtl RESET_MASK unimplemented\n"); + s->reset_mask = value; + break; + case A_GRETREG: + /* + * General retention register, which is only reset by a power-on + * reset. Technically this implementation is complete, since + * QEMU only supports power-on resets... + */ + s->gretreg = value; + break; + case A_INITSVRTOR0: + qemu_log_mask(LOG_UNIMP, "IoTKit SysCtl INITSVRTOR0 unimplemented\n"); + s->initsvrtor0 = value; + break; + case A_CPUWAIT: + qemu_log_mask(LOG_UNIMP, "IoTKit SysCtl CPUWAIT unimplemented\n"); + s->cpuwait = value; + break; + case A_WICCTRL: + qemu_log_mask(LOG_UNIMP, "IoTKit SysCtl WICCTRL unimplemented\n"); + s->wicctrl = value; + break; + case A_SECDBGSET: + /* write-1-to-set */ + qemu_log_mask(LOG_UNIMP, "IoTKit SysCtl SECDBGSET unimplemented\n"); + s->secure_debug |= value; + break; + case A_SECDBGCLR: + /* write-1-to-clear */ + s->secure_debug &= ~value; + break; + case A_SWRESET: + /* One w/o bit to request a reset; all other bits reserved */ + if (value & R_SWRESET_SWRESETREQ_MASK) { + qemu_system_reset_request(SHUTDOWN_CAUSE_GUEST_RESET); + } + break; + case A_BUSWAIT: /* In IoTKit BUSWAIT is reserved, R/O, zero */ + case A_SECDBGSTAT: + case A_PID4 ... A_CID3: + qemu_log_mask(LOG_GUEST_ERROR, + "IoTKit SysCtl write: write of RO offset %x\n", + (int)offset); + break; + default: + qemu_log_mask(LOG_GUEST_ERROR, + "IoTKit SysCtl write: bad offset %x\n", (int)offset); + break; + } +} + +static const MemoryRegionOps iotkit_sysctl_ops = { + .read = iotkit_sysctl_read, + .write = iotkit_sysctl_write, + .endianness = DEVICE_LITTLE_ENDIAN, + /* byte/halfword accesses are just zero-padded on reads and writes */ + .impl.min_access_size = 4, + .impl.max_access_size = 4, + .valid.min_access_size = 1, + .valid.max_access_size = 4, +}; + +static void iotkit_sysctl_reset(DeviceState *dev) +{ + IoTKitSysCtl *s = IOTKIT_SYSCTL(dev); + + trace_iotkit_sysctl_reset(); + s->secure_debug = 0; + s->reset_syndrome = 1; + s->reset_mask = 0; + s->gretreg = 0; + s->initsvrtor0 = 0x10000000; + s->cpuwait = 0; + s->wicctrl = 0; +} + +static void iotkit_sysctl_init(Object *obj) +{ + SysBusDevice *sbd = SYS_BUS_DEVICE(obj); + IoTKitSysCtl *s = IOTKIT_SYSCTL(obj); + + memory_region_init_io(&s->iomem, obj, &iotkit_sysctl_ops, + s, "iotkit-sysctl", 0x1000); + sysbus_init_mmio(sbd, &s->iomem); +} + +static const VMStateDescription iotkit_sysctl_vmstate = { + .name = "iotkit-sysctl", + .version_id = 1, + .minimum_version_id = 1, + .fields = (VMStateField[]) { + VMSTATE_UINT32(secure_debug, IoTKitSysCtl), + VMSTATE_UINT32(reset_syndrome, IoTKitSysCtl), + VMSTATE_UINT32(reset_mask, IoTKitSysCtl), + VMSTATE_UINT32(gretreg, IoTKitSysCtl), + VMSTATE_UINT32(initsvrtor0, IoTKitSysCtl), + VMSTATE_UINT32(cpuwait, IoTKitSysCtl), + VMSTATE_UINT32(wicctrl, IoTKitSysCtl), + VMSTATE_END_OF_LIST() + } +}; + +static void iotkit_sysctl_class_init(ObjectClass *klass, void *data) +{ + DeviceClass *dc = DEVICE_CLASS(klass); + + dc->vmsd = &iotkit_sysctl_vmstate; + dc->reset = iotkit_sysctl_reset; +} + +static const TypeInfo iotkit_sysctl_info = { + .name = TYPE_IOTKIT_SYSCTL, + .parent = TYPE_SYS_BUS_DEVICE, + .instance_size = sizeof(IoTKitSysCtl), + .instance_init = iotkit_sysctl_init, + .class_init = iotkit_sysctl_class_init, +}; + +static void iotkit_sysctl_register_types(void) +{ + type_register_static(&iotkit_sysctl_info); +} + +type_init(iotkit_sysctl_register_types); diff --git a/MAINTAINERS b/MAINTAINERS index XXXXXXX..XXXXXXX 100644 --- a/MAINTAINERS +++ b/MAINTAINERS @@ -XXX,XX +XXX,XX @@ F: hw/misc/mps2-*.c F: include/hw/misc/mps2-*.h F: hw/arm/iotkit.c F: include/hw/arm/iotkit.h +F: hw/misc/iotkit-sysctl.c +F: include/hw/misc/iotkit-sysctl.h Musicpal M: Jan Kiszka <jan.kiszka@web.de> diff --git a/default-configs/arm-softmmu.mak b/default-configs/arm-softmmu.mak index XXXXXXX..XXXXXXX 100644 --- a/default-configs/arm-softmmu.mak +++ b/default-configs/arm-softmmu.mak @@ -XXX,XX +XXX,XX @@ CONFIG_TZ_MPC=y CONFIG_TZ_PPC=y CONFIG_IOTKIT=y CONFIG_IOTKIT_SECCTL=y +CONFIG_IOTKIT_SYSCTL=y CONFIG_VERSATILE=y CONFIG_VERSATILE_PCI=y diff --git a/hw/misc/trace-events b/hw/misc/trace-events index XXXXXXX..XXXXXXX 100644 --- a/hw/misc/trace-events +++ b/hw/misc/trace-events @@ -XXX,XX +XXX,XX @@ ccm_freq(uint32_t freq) "freq = %d\n" ccm_clock_freq(uint32_t clock, uint32_t freq) "(Clock = %d) = %d\n" ccm_read_reg(const char *reg_name, uint32_t value) "reg[%s] <= 0x%" PRIx32 "\n" ccm_write_reg(const char *reg_name, uint32_t value) "reg[%s] => 0x%" PRIx32 "\n" + +# hw/misc/iotkit-sysctl.c +iotkit_sysinfo_read(uint64_t offset, uint64_t data, unsigned size) "IoTKit SysInfo read: offset 0x%" PRIx64 " data 0x%" PRIx64 " size %u" +iotkit_sysinfo_write(uint64_t offset, uint64_t data, unsigned size) "IoTKit SysInfo write: offset 0x%" PRIx64 " data 0x%" PRIx64 " size %u" +iotkit_sysctl_read(uint64_t offset, uint64_t data, unsigned size) "IoTKit SysCtl read: offset 0x%" PRIx64 " data 0x%" PRIx64 " size %u" +iotkit_sysctl_write(uint64_t offset, uint64_t data, unsigned size) "IoTKit SysCtl write: offset 0x%" PRIx64 " data 0x%" PRIx64 " size %u" +iotkit_sysctl_reset(void) "IoTKit SysCtl: reset" -- 2.18.0 Implement the IoTKit system control element's system information block; this is just a pair of read-only version/config registers, plus the usual PID/CID ID registers. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20180820141116.9118-10-peter.maydell@linaro.org --- hw/misc/Makefile.objs | 1 + include/hw/misc/iotkit-sysinfo.h | 37 +++++++++ hw/misc/iotkit-sysinfo.c | 128 +++++++++++++++++++++++++++++++ MAINTAINERS | 2 + default-configs/arm-softmmu.mak | 1 + 5 files changed, 169 insertions(+) create mode 100644 include/hw/misc/iotkit-sysinfo.h create mode 100644 hw/misc/iotkit-sysinfo.c diff --git a/hw/misc/Makefile.objs b/hw/misc/Makefile.objs index XXXXXXX..XXXXXXX 100644 --- a/hw/misc/Makefile.objs +++ b/hw/misc/Makefile.objs @@ -XXX,XX +XXX,XX @@ obj-$(CONFIG_TZ_MPC) += tz-mpc.o obj-$(CONFIG_TZ_PPC) += tz-ppc.o obj-$(CONFIG_IOTKIT_SECCTL) += iotkit-secctl.o obj-$(CONFIG_IOTKIT_SYSCTL) += iotkit-sysctl.o +obj-$(CONFIG_IOTKIT_SYSINFO) += iotkit-sysinfo.o obj-$(CONFIG_PVPANIC) += pvpanic.o obj-$(CONFIG_HYPERV_TESTDEV) += hyperv_testdev.o diff --git a/include/hw/misc/iotkit-sysinfo.h b/include/hw/misc/iotkit-sysinfo.h new file mode 100644 index XXXXXXX..XXXXXXX --- /dev/null +++ b/include/hw/misc/iotkit-sysinfo.h @@ -XXX,XX +XXX,XX @@ +/* + * ARM IoTKit system information block + * + * Copyright (c) 2018 Linaro Limited + * Written by Peter Maydell + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License version 2 or + * (at your option) any later version. + */ + +/* + * This is a model of the "system information block" which is part of the + * Arm IoTKit and documented in + * http://infocenter.arm.com/help/index.jsp?topic=/com.arm.doc.ecm0601256/index.html + * QEMU interface: + * + sysbus MMIO region 0: the system information register bank + */ + +#ifndef HW_MISC_IOTKIT_SYSINFO_H +#define HW_MISC_IOTKIT_SYSINFO_H + +#include "hw/sysbus.h" + +#define TYPE_IOTKIT_SYSINFO "iotkit-sysinfo" +#define IOTKIT_SYSINFO(obj) OBJECT_CHECK(IoTKitSysInfo, (obj), \ + TYPE_IOTKIT_SYSINFO) + +typedef struct IoTKitSysInfo { + /*< private >*/ + SysBusDevice parent_obj; + + /*< public >*/ + MemoryRegion iomem; +} IoTKitSysInfo; + +#endif diff --git a/hw/misc/iotkit-sysinfo.c b/hw/misc/iotkit-sysinfo.c new file mode 100644 index XXXXXXX..XXXXXXX --- /dev/null +++ b/hw/misc/iotkit-sysinfo.c @@ -XXX,XX +XXX,XX @@ +/* + * ARM IoTKit system information block + * + * Copyright (c) 2018 Linaro Limited + * Written by Peter Maydell + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License version 2 or + * (at your option) any later version. + */ + +/* + * This is a model of the "system information block" which is part of the + * Arm IoTKit and documented in + * http://infocenter.arm.com/help/index.jsp?topic=/com.arm.doc.ecm0601256/index.html + * It consists of 2 read-only version/config registers, plus the + * usual ID registers. + */ + +#include "qemu/osdep.h" +#include "qemu/log.h" +#include "trace.h" +#include "qapi/error.h" +#include "sysemu/sysemu.h" +#include "hw/sysbus.h" +#include "hw/registerfields.h" +#include "hw/misc/iotkit-sysinfo.h" + +REG32(SYS_VERSION, 0x0) +REG32(SYS_CONFIG, 0x4) +REG32(PID4, 0xfd0) +REG32(PID5, 0xfd4) +REG32(PID6, 0xfd8) +REG32(PID7, 0xfdc) +REG32(PID0, 0xfe0) +REG32(PID1, 0xfe4) +REG32(PID2, 0xfe8) +REG32(PID3, 0xfec) +REG32(CID0, 0xff0) +REG32(CID1, 0xff4) +REG32(CID2, 0xff8) +REG32(CID3, 0xffc) + +/* PID/CID values */ +static const int sysinfo_id[] = { + 0x04, 0x00, 0x00, 0x00, /* PID4..PID7 */ + 0x58, 0xb8, 0x0b, 0x00, /* PID0..PID3 */ + 0x0d, 0xf0, 0x05, 0xb1, /* CID0..CID3 */ +}; + +static uint64_t iotkit_sysinfo_read(void *opaque, hwaddr offset, + unsigned size) +{ + uint64_t r; + + switch (offset) { + case A_SYS_VERSION: + r = 0x41743; + break; + + case A_SYS_CONFIG: + r = 0x31; + break; + case A_PID4 ... A_CID3: + r = sysinfo_id[(offset - A_PID4) / 4]; + break; + default: + qemu_log_mask(LOG_GUEST_ERROR, + "IoTKit SysInfo read: bad offset %x\n", (int)offset); + r = 0; + break; + } + trace_iotkit_sysinfo_read(offset, r, size); + return r; +} + +static void iotkit_sysinfo_write(void *opaque, hwaddr offset, + uint64_t value, unsigned size) +{ + trace_iotkit_sysinfo_write(offset, value, size); + + qemu_log_mask(LOG_GUEST_ERROR, + "IoTKit SysInfo: write to RO offset 0x%x\n", (int)offset); +} + +static const MemoryRegionOps iotkit_sysinfo_ops = { + .read = iotkit_sysinfo_read, + .write = iotkit_sysinfo_write, + .endianness = DEVICE_LITTLE_ENDIAN, + /* byte/halfword accesses are just zero-padded on reads and writes */ + .impl.min_access_size = 4, + .impl.max_access_size = 4, + .valid.min_access_size = 1, + .valid.max_access_size = 4, +}; + +static void iotkit_sysinfo_init(Object *obj) +{ + SysBusDevice *sbd = SYS_BUS_DEVICE(obj); + IoTKitSysInfo *s = IOTKIT_SYSINFO(obj); + + memory_region_init_io(&s->iomem, obj, &iotkit_sysinfo_ops, + s, "iotkit-sysinfo", 0x1000); + sysbus_init_mmio(sbd, &s->iomem); +} + +static void iotkit_sysinfo_class_init(ObjectClass *klass, void *data) +{ + /* + * This device has no guest-modifiable state and so it + * does not need a reset function or VMState. + */ +} + +static const TypeInfo iotkit_sysinfo_info = { + .name = TYPE_IOTKIT_SYSINFO, + .parent = TYPE_SYS_BUS_DEVICE, + .instance_size = sizeof(IoTKitSysInfo), + .instance_init = iotkit_sysinfo_init, + .class_init = iotkit_sysinfo_class_init, +}; + +static void iotkit_sysinfo_register_types(void) +{ + type_register_static(&iotkit_sysinfo_info); +} + +type_init(iotkit_sysinfo_register_types); diff --git a/MAINTAINERS b/MAINTAINERS index XXXXXXX..XXXXXXX 100644 --- a/MAINTAINERS +++ b/MAINTAINERS @@ -XXX,XX +XXX,XX @@ F: hw/arm/iotkit.c F: include/hw/arm/iotkit.h F: hw/misc/iotkit-sysctl.c F: include/hw/misc/iotkit-sysctl.h +F: hw/misc/iotkit-sysinfo.c +F: include/hw/misc/iotkit-sysinfo.h Musicpal M: Jan Kiszka <jan.kiszka@web.de> diff --git a/default-configs/arm-softmmu.mak b/default-configs/arm-softmmu.mak index XXXXXXX..XXXXXXX 100644 --- a/default-configs/arm-softmmu.mak +++ b/default-configs/arm-softmmu.mak @@ -XXX,XX +XXX,XX @@ CONFIG_TZ_PPC=y CONFIG_IOTKIT=y CONFIG_IOTKIT_SECCTL=y CONFIG_IOTKIT_SYSCTL=y +CONFIG_IOTKIT_SYSINFO=y CONFIG_VERSATILE=y CONFIG_VERSATILE_PCI=y -- 2.18.0 Wire up the system control element's register banks (sysctl and sysinfo). This is the last of the previously completely unimplemented components in the IoTKit. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20180820141116.9118-11-peter.maydell@linaro.org --- include/hw/arm/iotkit.h | 6 +++++- hw/arm/iotkit.c | 26 ++++++++++++++++++-------- 2 files changed, 23 insertions(+), 9 deletions(-) diff --git a/include/hw/arm/iotkit.h b/include/hw/arm/iotkit.h index XXXXXXX..XXXXXXX 100644 --- a/include/hw/arm/iotkit.h +++ b/include/hw/arm/iotkit.h @@ -XXX,XX +XXX,XX @@ #include "hw/timer/cmsdk-apb-timer.h" #include "hw/timer/cmsdk-apb-dualtimer.h" #include "hw/watchdog/cmsdk-apb-watchdog.h" -#include "hw/misc/unimp.h" +#include "hw/misc/iotkit-sysctl.h" +#include "hw/misc/iotkit-sysinfo.h" #include "hw/or-irq.h" #include "hw/core/split-irq.h" @@ -XXX,XX +XXX,XX @@ typedef struct IoTKit { CMSDKAPBWatchdog nswatchdog; CMSDKAPBWatchdog swatchdog; + IoTKitSysCtl sysctl; + IoTKitSysCtl sysinfo; + MemoryRegion container; MemoryRegion alias1; MemoryRegion alias2; diff --git a/hw/arm/iotkit.c b/hw/arm/iotkit.c index XXXXXXX..XXXXXXX 100644 --- a/hw/arm/iotkit.c +++ b/hw/arm/iotkit.c @@ -XXX,XX +XXX,XX @@ #include "hw/sysbus.h" #include "hw/registerfields.h" #include "hw/arm/iotkit.h" -#include "hw/misc/unimp.h" #include "hw/arm/arm.h" /* Clock frequency in HZ of the 32KHz "slow clock" */ @@ -XXX,XX +XXX,XX @@ static void iotkit_init(Object *obj) sizeof(s->nswatchdog), TYPE_CMSDK_APB_WATCHDOG); sysbus_init_child_obj(obj, "swatchdog", &s->swatchdog, sizeof(s->swatchdog), TYPE_CMSDK_APB_WATCHDOG); + sysbus_init_child_obj(obj, "iotkit-sysctl", &s->sysctl, + sizeof(s->sysctl), TYPE_IOTKIT_SYSCTL); + sysbus_init_child_obj(obj, "iotkit-sysinfo", &s->sysinfo, + sizeof(s->sysinfo), TYPE_IOTKIT_SYSINFO); object_initialize_child(obj, "nmi-orgate", &s->nmi_orgate, sizeof(s->nmi_orgate), TYPE_OR_IRQ, &error_abort, NULL); @@ -XXX,XX +XXX,XX @@ static void iotkit_realize(DeviceState *dev, Error **errp) qdev_get_gpio_in_named(dev_apb_ppc1, "cfg_sec_resp", 0)); - /* Using create_unimplemented_device() maps the stub into the - * system address space rather than into our container, but the - * overall effect to the guest is the same. - */ - create_unimplemented_device("SYSINFO", 0x40020000, 0x1000); - - create_unimplemented_device("SYSCONTROL", 0x50021000, 0x1000); + object_property_set_bool(OBJECT(&s->sysinfo), true, "realized", &err); + if (err) { + error_propagate(errp, err); + return; + } + /* System information registers */ + sysbus_mmio_map(SYS_BUS_DEVICE(&s->sysinfo), 0, 0x40020000); + /* System control registers */ + object_property_set_bool(OBJECT(&s->sysctl), true, "realized", &err); + if (err) { + error_propagate(errp, err); + return; + } + sysbus_mmio_map(SYS_BUS_DEVICE(&s->sysctl), 0, 0x50021000); /* This OR gate wires together outputs from the secure watchdogs to NMI */ object_property_set_int(OBJECT(&s->nmi_orgate), 2, "num-lines", &err); -- 2.18.0 Implement a model of the TrustZone Master Securtiy Controller, as documented in the Arm CoreLink SIE-200 System IP for Embedded TRM (DDI0571G): https://developer.arm.com/products/architecture/m-profile/docs/ddi0571/g The MSC is intended to sit in front of a device which can be a bus master (eg a DMA controller) and programmably gate its transactions. This allows a bus-mastering device to be controlled by non-secure code but still restricted from making accesses to addresses which are secure-only. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Message-id: 20180820141116.9118-12-peter.maydell@linaro.org Reviewed-by: Richard Henderson <richard.henderson@linaro.org> --- hw/misc/Makefile.objs | 1 + include/hw/misc/tz-msc.h | 79 ++++++++ hw/misc/tz-msc.c | 308 ++++++++++++++++++++++++++++++++ MAINTAINERS | 2 + default-configs/arm-softmmu.mak | 1 + hw/misc/trace-events | 9 + 6 files changed, 400 insertions(+) create mode 100644 include/hw/misc/tz-msc.h create mode 100644 hw/misc/tz-msc.c diff --git a/hw/misc/Makefile.objs b/hw/misc/Makefile.objs index XXXXXXX..XXXXXXX 100644 --- a/hw/misc/Makefile.objs +++ b/hw/misc/Makefile.objs @@ -XXX,XX +XXX,XX @@ obj-$(CONFIG_MPS2_FPGAIO) += mps2-fpgaio.o obj-$(CONFIG_MPS2_SCC) += mps2-scc.o obj-$(CONFIG_TZ_MPC) += tz-mpc.o +obj-$(CONFIG_TZ_MSC) += tz-msc.o obj-$(CONFIG_TZ_PPC) += tz-ppc.o obj-$(CONFIG_IOTKIT_SECCTL) += iotkit-secctl.o obj-$(CONFIG_IOTKIT_SYSCTL) += iotkit-sysctl.o diff --git a/include/hw/misc/tz-msc.h b/include/hw/misc/tz-msc.h new file mode 100644 index XXXXXXX..XXXXXXX --- /dev/null +++ b/include/hw/misc/tz-msc.h @@ -XXX,XX +XXX,XX @@ +/* + * ARM TrustZone master security controller emulation + * + * Copyright (c) 2018 Linaro Limited + * Written by Peter Maydell + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License version 2 or + * (at your option) any later version. + */ + +/* + * This is a model of the TrustZone master security controller (MSC). + * It is documented in the ARM CoreLink SIE-200 System IP for Embedded TRM + * (DDI 0571G): + * https://developer.arm.com/products/architecture/m-profile/docs/ddi0571/g + * + * The MSC sits in front of a device which can be a bus master (such as + * a DMA controller) and allows secure software to configure it to either + * pass through or reject transactions made by that bus master. + * Rejected transactions may be configured to either be aborted, or to + * behave as RAZ/WI. An interrupt can be signalled for a rejected transaction. + * + * The MSC has no register interface -- it is configured purely by a + * collection of input signals from other hardware in the system. Typically + * they are either hardwired or exposed in an ad-hoc register interface by + * the SoC that uses the MSC. + * + * We don't currently implement the irq_enable GPIO input, because on + * the MPS2 FPGA images it is always tied high, which is awkward to + * implement in QEMU. + * + * QEMU interface: + * + Named GPIO input "cfg_nonsec": set to 1 if the bus master should be + * treated as nonsecure, or 0 for secure + * + Named GPIO input "cfg_sec_resp": set to 1 if a rejected transaction should + * result in a transaction error, or 0 for the transaction to RAZ/WI + * + Named GPIO input "irq_clear": set to 1 to clear a pending interrupt + * + Named GPIO output "irq": set for a transaction-failed interrupt + * + Property "downstream": MemoryRegion defining where bus master transactions + * are made if they are not blocked + * + Property "idau": an object implementing IDAUInterface, which defines which + * addresses should be treated as secure and which as non-secure. + * This need not be the same IDAU as the one used by the CPU. + * + sysbus MMIO region 0: MemoryRegion defining the upstream end of the MSC; + * this should be passed to the bus master device as the region it should + * make memory transactions to + */ + +#ifndef TZ_MSC_H +#define TZ_MSC_H + +#include "hw/sysbus.h" +#include "target/arm/idau.h" + +#define TYPE_TZ_MSC "tz-msc" +#define TZ_MSC(obj) OBJECT_CHECK(TZMSC, (obj), TYPE_TZ_MSC) + +typedef struct TZMSC { + /*< private >*/ + SysBusDevice parent_obj; + + /*< public >*/ + + /* State: these just track the values of our input signals */ + bool cfg_nonsec; + bool cfg_sec_resp; + bool irq_clear; + /* State: are we asserting irq ? */ + bool irq_status; + + qemu_irq irq; + MemoryRegion *downstream; + AddressSpace downstream_as; + MemoryRegion upstream; + IDAUInterface *idau; +} TZMSC; + +#endif diff --git a/hw/misc/tz-msc.c b/hw/misc/tz-msc.c new file mode 100644 index XXXXXXX..XXXXXXX --- /dev/null +++ b/hw/misc/tz-msc.c @@ -XXX,XX +XXX,XX @@ +/* + * ARM TrustZone master security controller emulation + * + * Copyright (c) 2018 Linaro Limited + * Written by Peter Maydell + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License version 2 or + * (at your option) any later version. + */ + +#include "qemu/osdep.h" +#include "qemu/log.h" +#include "qapi/error.h" +#include "trace.h" +#include "hw/sysbus.h" +#include "hw/registerfields.h" +#include "hw/misc/tz-msc.h" + +static void tz_msc_update_irq(TZMSC *s) +{ + bool level = s->irq_status; + + trace_tz_msc_update_irq(level); + qemu_set_irq(s->irq, level); +} + +static void tz_msc_cfg_nonsec(void *opaque, int n, int level) +{ + TZMSC *s = TZ_MSC(opaque); + + trace_tz_msc_cfg_nonsec(level); + s->cfg_nonsec = level; +} + +static void tz_msc_cfg_sec_resp(void *opaque, int n, int level) +{ + TZMSC *s = TZ_MSC(opaque); + + trace_tz_msc_cfg_sec_resp(level); + s->cfg_sec_resp = level; +} + +static void tz_msc_irq_clear(void *opaque, int n, int level) +{ + TZMSC *s = TZ_MSC(opaque); + + trace_tz_msc_irq_clear(level); + + s->irq_clear = level; + if (level) { + s->irq_status = false; + tz_msc_update_irq(s); + } +} + +/* The MSC may either block a transaction by aborting it, block a + * transaction by making it RAZ/WI, allow it through with + * MemTxAttrs indicating a secure transaction, or allow it with + * MemTxAttrs indicating a non-secure transaction. + */ +typedef enum MSCAction { + MSCBlockAbort, + MSCBlockRAZWI, + MSCAllowSecure, + MSCAllowNonSecure, +} MSCAction; + +static MSCAction tz_msc_check(TZMSC *s, hwaddr addr) +{ + /* + * Check whether to allow an access from the bus master, returning + * an MSCAction indicating the required behaviour. If the transaction + * is blocked, the caller must check cfg_sec_resp to determine + * whether to abort or RAZ/WI the transaction. + */ + IDAUInterfaceClass *iic = IDAU_INTERFACE_GET_CLASS(s->idau); + IDAUInterface *ii = IDAU_INTERFACE(s->idau); + bool idau_exempt = false, idau_ns = true, idau_nsc = true; + int idau_region = IREGION_NOTVALID; + + iic->check(ii, addr, &idau_region, &idau_exempt, &idau_ns, &idau_nsc); + + if (idau_exempt) { + /* + * Uncheck region -- OK, transaction type depends on + * whether bus master is configured as Secure or NonSecure + */ + return s->cfg_nonsec ? MSCAllowNonSecure : MSCAllowSecure; + } + + if (idau_ns) { + /* NonSecure region -- always forward as NS transaction */ + return MSCAllowNonSecure; + } + + if (!s->cfg_nonsec) { + /* Access to Secure region by Secure bus master: OK */ + return MSCAllowSecure; + } + + /* Attempted access to Secure region by NS bus master: block */ + trace_tz_msc_access_blocked(addr); + if (!s->cfg_sec_resp) { + return MSCBlockRAZWI; + } + + /* + * The TRM isn't clear on behaviour if irq_clear is high when a + * transaction is blocked. We assume that the MSC behaves like the + * PPC, where holding irq_clear high suppresses the interrupt. + */ + if (!s->irq_clear) { + s->irq_status = true; + tz_msc_update_irq(s); + } + return MSCBlockAbort; +} + +static MemTxResult tz_msc_read(void *opaque, hwaddr addr, uint64_t *pdata, + unsigned size, MemTxAttrs attrs) +{ + TZMSC *s = opaque; + AddressSpace *as = &s->downstream_as; + uint64_t data; + MemTxResult res; + + switch (tz_msc_check(s, addr)) { + case MSCBlockAbort: + return MEMTX_ERROR; + case MSCBlockRAZWI: + *pdata = 0; + return MEMTX_OK; + case MSCAllowSecure: + attrs.secure = 1; + attrs.unspecified = 0; + break; + case MSCAllowNonSecure: + attrs.secure = 0; + attrs.unspecified = 0; + break; + } + + switch (size) { + case 1: + data = address_space_ldub(as, addr, attrs, &res); + break; + case 2: + data = address_space_lduw_le(as, addr, attrs, &res); + break; + case 4: + data = address_space_ldl_le(as, addr, attrs, &res); + break; + case 8: + data = address_space_ldq_le(as, addr, attrs, &res); + break; + default: + g_assert_not_reached(); + } + *pdata = data; + return res; +} + +static MemTxResult tz_msc_write(void *opaque, hwaddr addr, uint64_t val, + unsigned size, MemTxAttrs attrs) +{ + TZMSC *s = opaque; + AddressSpace *as = &s->downstream_as; + MemTxResult res; + + switch (tz_msc_check(s, addr)) { + case MSCBlockAbort: + return MEMTX_ERROR; + case MSCBlockRAZWI: + return MEMTX_OK; + case MSCAllowSecure: + attrs.secure = 1; + attrs.unspecified = 0; + break; + case MSCAllowNonSecure: + attrs.secure = 0; + attrs.unspecified = 0; + break; + } + + switch (size) { + case 1: + address_space_stb(as, addr, val, attrs, &res); + break; + case 2: + address_space_stw_le(as, addr, val, attrs, &res); + break; + case 4: + address_space_stl_le(as, addr, val, attrs, &res); + break; + case 8: + address_space_stq_le(as, addr, val, attrs, &res); + break; + default: + g_assert_not_reached(); + } + return res; +} + +static const MemoryRegionOps tz_msc_ops = { + .read_with_attrs = tz_msc_read, + .write_with_attrs = tz_msc_write, + .endianness = DEVICE_LITTLE_ENDIAN, +}; + +static void tz_msc_reset(DeviceState *dev) +{ + TZMSC *s = TZ_MSC(dev); + + trace_tz_msc_reset(); + s->cfg_sec_resp = false; + s->cfg_nonsec = false; + s->irq_clear = 0; + s->irq_status = 0; +} + +static void tz_msc_init(Object *obj) +{ + DeviceState *dev = DEVICE(obj); + TZMSC *s = TZ_MSC(obj); + + qdev_init_gpio_in_named(dev, tz_msc_cfg_nonsec, "cfg_nonsec", 1); + qdev_init_gpio_in_named(dev, tz_msc_cfg_sec_resp, "cfg_sec_resp", 1); + qdev_init_gpio_in_named(dev, tz_msc_irq_clear, "irq_clear", 1); + qdev_init_gpio_out_named(dev, &s->irq, "irq", 1); +} + +static void tz_msc_realize(DeviceState *dev, Error **errp) +{ + Object *obj = OBJECT(dev); + SysBusDevice *sbd = SYS_BUS_DEVICE(dev); + TZMSC *s = TZ_MSC(dev); + const char *name = "tz-msc-downstream"; + uint64_t size; + + /* + * We can't create the upstream end of the port until realize, + * as we don't know the size of the MR used as the downstream until then. + * We insist on having a downstream, to avoid complicating the + * code with handling the "don't know how big this is" case. It's easy + * enough for the user to create an unimplemented_device as downstream + * if they have nothing else to plug into this. + */ + if (!s->downstream) { + error_setg(errp, "MSC 'downstream' link not set"); + return; + } + if (!s->idau) { + error_setg(errp, "MSC 'idau' link not set"); + return; + } + + size = memory_region_size(s->downstream); + address_space_init(&s->downstream_as, s->downstream, name); + memory_region_init_io(&s->upstream, obj, &tz_msc_ops, s, name, size); + sysbus_init_mmio(sbd, &s->upstream); +} + +static const VMStateDescription tz_msc_vmstate = { + .name = "tz-msc", + .version_id = 1, + .minimum_version_id = 1, + .fields = (VMStateField[]) { + VMSTATE_BOOL(cfg_nonsec, TZMSC), + VMSTATE_BOOL(cfg_sec_resp, TZMSC), + VMSTATE_BOOL(irq_clear, TZMSC), + VMSTATE_BOOL(irq_status, TZMSC), + VMSTATE_END_OF_LIST() + } +}; + +static Property tz_msc_properties[] = { + DEFINE_PROP_LINK("downstream", TZMSC, downstream, + TYPE_MEMORY_REGION, MemoryRegion *), + DEFINE_PROP_LINK("idau", TZMSC, idau, + TYPE_IDAU_INTERFACE, IDAUInterface *), + DEFINE_PROP_END_OF_LIST(), +}; + +static void tz_msc_class_init(ObjectClass *klass, void *data) +{ + DeviceClass *dc = DEVICE_CLASS(klass); + + dc->realize = tz_msc_realize; + dc->vmsd = &tz_msc_vmstate; + dc->reset = tz_msc_reset; + dc->props = tz_msc_properties; +} + +static const TypeInfo tz_msc_info = { + .name = TYPE_TZ_MSC, + .parent = TYPE_SYS_BUS_DEVICE, + .instance_size = sizeof(TZMSC), + .instance_init = tz_msc_init, + .class_init = tz_msc_class_init, +}; + +static void tz_msc_register_types(void) +{ + type_register_static(&tz_msc_info); +} + +type_init(tz_msc_register_types); diff --git a/MAINTAINERS b/MAINTAINERS index XXXXXXX..XXXXXXX 100644 --- a/MAINTAINERS +++ b/MAINTAINERS @@ -XXX,XX +XXX,XX @@ F: hw/misc/tz-ppc.c F: include/hw/misc/tz-ppc.h F: hw/misc/tz-mpc.c F: include/hw/misc/tz-mpc.h +F: hw/misc/tz-msc.c +F: include/hw/misc/tz-msc.h ARM cores M: Peter Maydell <peter.maydell@linaro.org> diff --git a/default-configs/arm-softmmu.mak b/default-configs/arm-softmmu.mak index XXXXXXX..XXXXXXX 100644 --- a/default-configs/arm-softmmu.mak +++ b/default-configs/arm-softmmu.mak @@ -XXX,XX +XXX,XX @@ CONFIG_MPS2_FPGAIO=y CONFIG_MPS2_SCC=y CONFIG_TZ_MPC=y +CONFIG_TZ_MSC=y CONFIG_TZ_PPC=y CONFIG_IOTKIT=y CONFIG_IOTKIT_SECCTL=y diff --git a/hw/misc/trace-events b/hw/misc/trace-events index XXXXXXX..XXXXXXX 100644 --- a/hw/misc/trace-events +++ b/hw/misc/trace-events @@ -XXX,XX +XXX,XX @@ tz_mpc_mem_blocked_write(uint64_t addr, uint64_t data, unsigned size, bool secur tz_mpc_translate(uint64_t addr, int flags, const char *idx, const char *res) "TZ MPC translate: addr 0x%" PRIx64 " flags 0x%x iommu_idx %s: %s" tz_mpc_iommu_notify(uint64_t addr) "TZ MPC iommu: notifying UNMAP/MAP for 0x%" PRIx64 +# hw/misc/tz-msc.c +tz_msc_reset(void) "TZ MSC: reset" +tz_msc_cfg_nonsec(int level) "TZ MSC: cfg_nonsec = %d" +tz_msc_cfg_sec_resp(int level) "TZ MSC: cfg_sec_resp = %d" +tz_msc_irq_enable(int level) "TZ MSC: int_enable = %d" +tz_msc_irq_clear(int level) "TZ MSC: int_clear = %d" +tz_msc_update_irq(int level) "TZ MSC: setting irq line to %d" +tz_msc_access_blocked(uint64_t offset) "TZ MSC: offset 0x%" PRIx64 " access blocked" + # hw/misc/tz-ppc.c tz_ppc_reset(void) "TZ PPC: reset" tz_ppc_cfg_nonsec(int n, int level) "TZ PPC: cfg_nonsec[%d] = %d" -- 2.18.0 The IoTKit does not have any Master Security Contollers itself, but it does provide registers in the secure privilege control block which allow control of MSCs in the external system. Add support for these registers. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Message-id: 20180820141116.9118-13-peter.maydell@linaro.org Reviewed-by: Richard Henderson <richard.henderson@linaro.org> --- include/hw/misc/iotkit-secctl.h | 14 +++++++ hw/misc/iotkit-secctl.c | 73 +++++++++++++++++++++++++++++---- 2 files changed, 79 insertions(+), 8 deletions(-) diff --git a/include/hw/misc/iotkit-secctl.h b/include/hw/misc/iotkit-secctl.h index XXXXXXX..XXXXXXX 100644 --- a/include/hw/misc/iotkit-secctl.h +++ b/include/hw/misc/iotkit-secctl.h @@ -XXX,XX +XXX,XX @@ * + named GPIO output "sec_resp_cfg" indicating whether blocked accesses * should RAZ/WI or bus error * + named GPIO output "nsc_cfg" whose value tracks the NSCCFG register value + * + named GPIO output "msc_irq" for the combined IRQ line from the MSCs * Controlling the 2 APB PPCs in the IoTKit: * + named GPIO outputs apb_ppc0_nonsec[0..2] and apb_ppc1_nonsec * + named GPIO outputs apb_ppc0_ap[0..2] and apb_ppc1_ap @@ -XXX,XX +XXX,XX @@ * Controlling each of the 16 expansion MPCs which a system using the IoTKit * might provide: * + named GPIO inputs mpcexp_status[0..15] + * Controlling each of the 16 expansion MSCs which a system using the IoTKit + * might provide: + * + named GPIO inputs mscexp_status[0..15] + * + named GPIO outputs mscexp_clear[0..15] + * + named GPIO outputs mscexp_ns[0..15] */ #ifndef IOTKIT_SECCTL_H @@ -XXX,XX +XXX,XX @@ #define IOTS_NUM_AHB_EXP_PPC 4 #define IOTS_NUM_EXP_MPC 16 #define IOTS_NUM_MPC 1 +#define IOTS_NUM_EXP_MSC 16 typedef struct IoTKitSecCtl IoTKitSecCtl; @@ -XXX,XX +XXX,XX @@ struct IoTKitSecCtl { uint32_t brginten; uint32_t mpcintstatus; + uint32_t secmscintstat; + uint32_t secmscinten; + uint32_t nsmscexp; + qemu_irq mscexp_clear[IOTS_NUM_EXP_MSC]; + qemu_irq mscexp_ns[IOTS_NUM_EXP_MSC]; + qemu_irq msc_irq; + IoTKitSecCtlPPC apb[IOTS_NUM_APB_PPC]; IoTKitSecCtlPPC apbexp[IOTS_NUM_APB_EXP_PPC]; IoTKitSecCtlPPC ahbexp[IOTS_NUM_APB_EXP_PPC]; diff --git a/hw/misc/iotkit-secctl.c b/hw/misc/iotkit-secctl.c index XXXXXXX..XXXXXXX 100644 --- a/hw/misc/iotkit-secctl.c +++ b/hw/misc/iotkit-secctl.c @@ -XXX,XX +XXX,XX @@ static MemTxResult iotkit_secctl_s_read(void *opaque, hwaddr addr, r = s->apbexp[offset_to_ppc_idx(offset)].sp; break; case A_SECMSCINTSTAT: + r = s->secmscintstat; + break; case A_SECMSCINTEN: + r = s->secmscinten; + break; case A_NSMSCEXP: - qemu_log_mask(LOG_UNIMP, - "IoTKit SecCtl S block read: " - "unimplemented offset 0x%x\n", offset); - r = 0; + r = s->nsmscexp; break; case A_PID4: case A_PID5: @@ -XXX,XX +XXX,XX @@ static void iotkit_secctl_ppc_update_irq_enable(IoTKitSecCtlPPC *ppc) qemu_set_irq(ppc->irq_enable, extract32(value, ppc->irq_bit_offset, 1)); } +static void iotkit_secctl_update_mscexp_irqs(qemu_irq *msc_irqs, uint32_t value) +{ + int i; + + for (i = 0; i < IOTS_NUM_EXP_MSC; i++) { + qemu_set_irq(msc_irqs[i], extract32(value, i + 16, 1)); + } +} + +static void iotkit_secctl_update_msc_irq(IoTKitSecCtl *s) +{ + /* Update the combined MSC IRQ, based on S_MSCEXP_STATUS and S_MSCEXP_EN */ + bool level = s->secmscintstat & s->secmscinten; + + qemu_set_irq(s->msc_irq, level); +} + static MemTxResult iotkit_secctl_s_write(void *opaque, hwaddr addr, uint64_t value, unsigned size, MemTxAttrs attrs) @@ -XXX,XX +XXX,XX @@ static MemTxResult iotkit_secctl_s_write(void *opaque, hwaddr addr, iotkit_secctl_ppc_sp_write(ppc, value); break; case A_SECMSCINTCLR: + iotkit_secctl_update_mscexp_irqs(s->mscexp_clear, value); + break; case A_SECMSCINTEN: - qemu_log_mask(LOG_UNIMP, - "IoTKit SecCtl S block write: " - "unimplemented offset 0x%x\n", offset); + s->secmscinten = value; + iotkit_secctl_update_msc_irq(s); + break; + case A_NSMSCEXP: + s->nsmscexp = value; + iotkit_secctl_update_mscexp_irqs(s->mscexp_ns, value); break; case A_SECMPCINTSTATUS: case A_SECPPCINTSTAT: @@ -XXX,XX +XXX,XX @@ static MemTxResult iotkit_secctl_s_write(void *opaque, hwaddr addr, case A_BRGINTSTAT: case A_AHBNSPPC0: case A_AHBSPPPC0: - case A_NSMSCEXP: case A_PID4: case A_PID5: case A_PID6: @@ -XXX,XX +XXX,XX @@ static void iotkit_secctl_mpcexp_status(void *opaque, int n, int level) s->mpcintstatus = deposit32(s->mpcintstatus, n + 16, 1, !!level); } +static void iotkit_secctl_mscexp_status(void *opaque, int n, int level) +{ + IoTKitSecCtl *s = IOTKIT_SECCTL(opaque); + + s->secmscintstat = deposit32(s->secmscintstat, n + 16, 1, !!level); + iotkit_secctl_update_msc_irq(s); +} + static void iotkit_secctl_ppc_irqstatus(void *opaque, int n, int level) { IoTKitSecCtlPPC *ppc = opaque; @@ -XXX,XX +XXX,XX @@ static void iotkit_secctl_init(Object *obj) qdev_init_gpio_in_named(dev, iotkit_secctl_mpcexp_status, "mpcexp_status", IOTS_NUM_EXP_MPC); + qdev_init_gpio_in_named(dev, iotkit_secctl_mscexp_status, + "mscexp_status", IOTS_NUM_EXP_MSC); + qdev_init_gpio_out_named(dev, s->mscexp_clear, "mscexp_clear", + IOTS_NUM_EXP_MSC); + qdev_init_gpio_out_named(dev, s->mscexp_ns, "mscexp_ns", + IOTS_NUM_EXP_MSC); + qdev_init_gpio_out_named(dev, &s->msc_irq, "msc_irq", 1); + memory_region_init_io(&s->s_regs, obj, &iotkit_secctl_s_ops, s, "iotkit-secctl-s-regs", 0x1000); memory_region_init_io(&s->ns_regs, obj, &iotkit_secctl_ns_ops, @@ -XXX,XX +XXX,XX @@ static const VMStateDescription iotkit_secctl_mpcintstatus_vmstate = { } }; +static bool needed_always(void *opaque) +{ + return true; +} + +static const VMStateDescription iotkit_secctl_msc_vmstate = { + .name = "iotkit-secctl/msc", + .version_id = 1, + .minimum_version_id = 1, + .needed = needed_always, + .fields = (VMStateField[]) { + VMSTATE_UINT32(secmscintstat, IoTKitSecCtl), + VMSTATE_UINT32(secmscinten, IoTKitSecCtl), + VMSTATE_UINT32(nsmscexp, IoTKitSecCtl), + VMSTATE_END_OF_LIST() + } +}; + static const VMStateDescription iotkit_secctl_vmstate = { .name = "iotkit-secctl", .version_id = 1, @@ -XXX,XX +XXX,XX @@ static const VMStateDescription iotkit_secctl_vmstate = { }, .subsections = (const VMStateDescription*[]) { &iotkit_secctl_mpcintstatus_vmstate, + &iotkit_secctl_msc_vmstate, NULL }, }; -- 2.18.0 The IoTKit doesn't have any MSCs itself but it does need some wiring to connect the external signals from MSCs in the outer board model up to the registers and the NVIC IRQ line. We also need to expose a MemoryRegion corresponding to the AHB bus, so that MSCs in the outer board model can use that as their downstream port. (In the FPGA this is the "AHB Slave Expansion" ports shown in the block diagram in the AN505 documentation.) Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Message-id: 20180820141116.9118-14-peter.maydell@linaro.org Reviewed-by: Richard Henderson <richard.henderson@linaro.org> --- include/hw/arm/iotkit.h | 8 ++++++++ hw/arm/iotkit.c | 15 +++++++++++++++ 2 files changed, 23 insertions(+) diff --git a/include/hw/arm/iotkit.h b/include/hw/arm/iotkit.h index XXXXXXX..XXXXXXX 100644 --- a/include/hw/arm/iotkit.h +++ b/include/hw/arm/iotkit.h @@ -XXX,XX +XXX,XX @@ * + QOM property "EXP_NUMIRQ" sets the number of expansion interrupts * + Named GPIO inputs "EXP_IRQ" 0..n are the expansion interrupts, which * are wired to the NVIC lines 32 .. n+32 + * + sysbus MMIO region 0 is the "AHB Slave Expansion" which allows + * bus master devices in the board model to make transactions into + * all the devices and memory areas in the IoTKit * Controlling up to 4 AHB expansion PPBs which a system using the IoTKit * might provide: * + named GPIO outputs apb_ppcexp{0,1,2,3}_nonsec[0..15] @@ -XXX,XX +XXX,XX @@ * Controlling each of the 16 expansion MPCs which a system using the IoTKit * might provide: * + named GPIO inputs mpcexp_status[0..15] + * Controlling each of the 16 expansion MSCs which a system using the IoTKit + * might provide: + * + named GPIO inputs mscexp_status[0..15] + * + named GPIO outputs mscexp_clear[0..15] + * + named GPIO outputs mscexp_ns[0..15] */ #ifndef IOTKIT_H diff --git a/hw/arm/iotkit.c b/hw/arm/iotkit.c index XXXXXXX..XXXXXXX 100644 --- a/hw/arm/iotkit.c +++ b/hw/arm/iotkit.c @@ -XXX,XX +XXX,XX @@ static void iotkit_realize(DeviceState *dev, Error **errp) iotkit_forward_sec_resp_cfg(s); + /* Forward the MSC related signals */ + qdev_pass_gpios(dev_secctl, dev, "mscexp_status"); + qdev_pass_gpios(dev_secctl, dev, "mscexp_clear"); + qdev_pass_gpios(dev_secctl, dev, "mscexp_ns"); + qdev_connect_gpio_out_named(dev_secctl, "msc_irq", 0, + qdev_get_gpio_in(DEVICE(&s->armv7m), 11)); + + /* + * Expose our container region to the board model; this corresponds + * to the AHB Slave Expansion ports which allow bus master devices + * (eg DMA controllers) in the board model to make transactions into + * devices in the IoTKit. + */ + sysbus_init_mmio(SYS_BUS_DEVICE(s), &s->container); + system_clock_scale = NANOSECONDS_PER_SECOND / s->mainclk_frq; } -- 2.18.0 The AN505 FPGA image includes four PL081 DMA controllers, each of which is gated by a Master Security Controller that allows the guest to prevent a non-secure DMA controller from accessing memory that is used by secure guest code. Create and wire up these devices. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Message-id: 20180820141116.9118-15-peter.maydell@linaro.org Reviewed-by: Richard Henderson <richard.henderson@linaro.org> --- hw/arm/mps2-tz.c | 100 +++++++++++++++++++++++++++++++++++++++++++---- 1 file changed, 93 insertions(+), 7 deletions(-) diff --git a/hw/arm/mps2-tz.c b/hw/arm/mps2-tz.c index XXXXXXX..XXXXXXX 100644 --- a/hw/arm/mps2-tz.c +++ b/hw/arm/mps2-tz.c @@ -XXX,XX +XXX,XX @@ #include "hw/misc/mps2-scc.h" #include "hw/misc/mps2-fpgaio.h" #include "hw/misc/tz-mpc.h" +#include "hw/misc/tz-msc.h" #include "hw/arm/iotkit.h" +#include "hw/dma/pl080.h" #include "hw/devices.h" #include "net/net.h" #include "hw/core/split-irq.h" @@ -XXX,XX +XXX,XX @@ typedef struct { UnimplementedDeviceState i2c[4]; UnimplementedDeviceState i2s_audio; UnimplementedDeviceState gpio[4]; - UnimplementedDeviceState dma[4]; UnimplementedDeviceState gfx; + PL080State dma[4]; + TZMSC msc[4]; CMSDKAPBUART uart[5]; SplitIRQ sec_resp_splitter; qemu_or_irq uart_irq_orgate; @@ -XXX,XX +XXX,XX @@ static MemoryRegion *make_mpc(MPS2TZMachineState *mms, void *opaque, return sysbus_mmio_get_region(SYS_BUS_DEVICE(mpc), 0); } +static MemoryRegion *make_dma(MPS2TZMachineState *mms, void *opaque, + const char *name, hwaddr size) +{ + PL080State *dma = opaque; + int i = dma - &mms->dma[0]; + SysBusDevice *s; + char *mscname = g_strdup_printf("%s-msc", name); + TZMSC *msc = &mms->msc[i]; + DeviceState *iotkitdev = DEVICE(&mms->iotkit); + MemoryRegion *msc_upstream; + MemoryRegion *msc_downstream; + + /* + * Each DMA device is a PL081 whose transaction master interface + * is guarded by a Master Security Controller. The downstream end of + * the MSC connects to the IoTKit AHB Slave Expansion port, so the + * DMA devices can see all devices and memory that the CPU does. + */ + sysbus_init_child_obj(OBJECT(mms), mscname, msc, sizeof(*msc), TYPE_TZ_MSC); + msc_downstream = sysbus_mmio_get_region(SYS_BUS_DEVICE(&mms->iotkit), 0); + object_property_set_link(OBJECT(msc), OBJECT(msc_downstream), + "downstream", &error_fatal); + object_property_set_link(OBJECT(msc), OBJECT(mms), + "idau", &error_fatal); + object_property_set_bool(OBJECT(msc), true, "realized", &error_fatal); + + qdev_connect_gpio_out_named(DEVICE(msc), "irq", 0, + qdev_get_gpio_in_named(iotkitdev, + "mscexp_status", i)); + qdev_connect_gpio_out_named(iotkitdev, "mscexp_clear", i, + qdev_get_gpio_in_named(DEVICE(msc), + "irq_clear", 0)); + qdev_connect_gpio_out_named(iotkitdev, "mscexp_ns", i, + qdev_get_gpio_in_named(DEVICE(msc), + "cfg_nonsec", 0)); + qdev_connect_gpio_out(DEVICE(&mms->sec_resp_splitter), + ARRAY_SIZE(mms->ppc) + i, + qdev_get_gpio_in_named(DEVICE(msc), + "cfg_sec_resp", 0)); + msc_upstream = sysbus_mmio_get_region(SYS_BUS_DEVICE(msc), 0); + + sysbus_init_child_obj(OBJECT(mms), name, dma, sizeof(*dma), TYPE_PL081); + object_property_set_link(OBJECT(dma), OBJECT(msc_upstream), + "downstream", &error_fatal); + object_property_set_bool(OBJECT(dma), true, "realized", &error_fatal); + + s = SYS_BUS_DEVICE(dma); + /* Wire up DMACINTR, DMACINTERR, DMACINTTC */ + sysbus_connect_irq(s, 0, qdev_get_gpio_in_named(iotkitdev, + "EXP_IRQ", 58 + i * 3)); + sysbus_connect_irq(s, 1, qdev_get_gpio_in_named(iotkitdev, + "EXP_IRQ", 56 + i * 3)); + sysbus_connect_irq(s, 2, qdev_get_gpio_in_named(iotkitdev, + "EXP_IRQ", 57 + i * 3)); + + return sysbus_mmio_get_region(s, 0); +} + static void mps2tz_common_init(MachineState *machine) { MPS2TZMachineState *mms = MPS2TZ_MACHINE(machine); @@ -XXX,XX +XXX,XX @@ static void mps2tz_common_init(MachineState *machine) &error_fatal); /* The sec_resp_cfg output from the IoTKit must be split into multiple - * lines, one for each of the PPCs we create here. + * lines, one for each of the PPCs we create here, plus one per MSC. */ object_initialize(&mms->sec_resp_splitter, sizeof(mms->sec_resp_splitter), TYPE_SPLIT_IRQ); object_property_add_child(OBJECT(machine), "sec-resp-splitter", OBJECT(&mms->sec_resp_splitter), &error_abort); - object_property_set_int(OBJECT(&mms->sec_resp_splitter), 5, + object_property_set_int(OBJECT(&mms->sec_resp_splitter), + ARRAY_SIZE(mms->ppc) + ARRAY_SIZE(mms->msc), "num-lines", &error_fatal); object_property_set_bool(OBJECT(&mms->sec_resp_splitter), true, "realized", &error_fatal); @@ -XXX,XX +XXX,XX @@ static void mps2tz_common_init(MachineState *machine) }, { .name = "ahb_ppcexp1", .ports = { - { "dma0", make_unimp_dev, &mms->dma[0], 0x40110000, 0x1000 }, - { "dma1", make_unimp_dev, &mms->dma[1], 0x40111000, 0x1000 }, - { "dma2", make_unimp_dev, &mms->dma[2], 0x40112000, 0x1000 }, - { "dma3", make_unimp_dev, &mms->dma[3], 0x40113000, 0x1000 }, + { "dma0", make_dma, &mms->dma[0], 0x40110000, 0x1000 }, + { "dma1", make_dma, &mms->dma[1], 0x40111000, 0x1000 }, + { "dma2", make_dma, &mms->dma[2], 0x40112000, 0x1000 }, + { "dma3", make_dma, &mms->dma[3], 0x40113000, 0x1000 }, }, }, }; @@ -XXX,XX +XXX,XX @@ static void mps2tz_common_init(MachineState *machine) armv7m_load_kernel(ARM_CPU(first_cpu), machine->kernel_filename, 0x400000); } +static void mps2_tz_idau_check(IDAUInterface *ii, uint32_t address, + int *iregion, bool *exempt, bool *ns, bool *nsc) +{ + /* + * The MPS2 TZ FPGA images have IDAUs in them which are connected to + * the Master Security Controllers. Thes have the same logic as + * is used by the IoTKit for the IDAU connected to the CPU, except + * that MSCs don't care about the NSC attribute. + */ + int region = extract32(address, 28, 4); + + *ns = !(region & 1); + *nsc = false; + /* 0xe0000000..0xe00fffff and 0xf0000000..0xf00fffff are exempt */ + *exempt = (address & 0xeff00000) == 0xe0000000; + *iregion = region; +} + static void mps2tz_class_init(ObjectClass *oc, void *data) { MachineClass *mc = MACHINE_CLASS(oc); + IDAUInterfaceClass *iic = IDAU_INTERFACE_CLASS(oc); mc->init = mps2tz_common_init; mc->max_cpus = 1; + iic->check = mps2_tz_idau_check; } static void mps2tz_an505_class_init(ObjectClass *oc, void *data) @@ -XXX,XX +XXX,XX @@ static const TypeInfo mps2tz_info = { .instance_size = sizeof(MPS2TZMachineState), .class_size = sizeof(MPS2TZMachineClass), .class_init = mps2tz_class_init, + .interfaces = (InterfaceInfo[]) { + { TYPE_IDAU_INTERFACE }, + { } + }, }; static const TypeInfo mps2tz_an505_info = { -- 2.18.0 Create a new include file for the pl022's device struct, type macros, etc, so that it can be instantiated using the "embedded struct" coding style. While we're adding the new file to MAINTAINERS, add also the .c file, which was missing an entry. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Message-id: 20180820141116.9118-16-peter.maydell@linaro.org Reviewed-by: Richard Henderson <richard.henderson@linaro.org> --- include/hw/ssi/pl022.h | 51 ++++++++++++++++++++++++++++++++++++++++++ hw/ssi/pl022.c | 26 +-------------------- MAINTAINERS | 2 ++ 3 files changed, 54 insertions(+), 25 deletions(-) create mode 100644 include/hw/ssi/pl022.h diff --git a/include/hw/ssi/pl022.h b/include/hw/ssi/pl022.h new file mode 100644 index XXXXXXX..XXXXXXX --- /dev/null +++ b/include/hw/ssi/pl022.h @@ -XXX,XX +XXX,XX @@ +/* + * ARM PrimeCell PL022 Synchronous Serial Port + * + * Copyright (c) 2007 CodeSourcery. + * Written by Paul Brook + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License version 2 or + * (at your option) any later version. + */ + +/* This is a model of the Arm PrimeCell PL022 synchronous serial port. + * The PL022 TRM is: + * http://infocenter.arm.com/help/topic/com.arm.doc.ddi0194h/DDI0194H_ssp_pl022_trm.pdf + * + * QEMU interface: + * + sysbus IRQ: SSPINTR combined interrupt line + * + sysbus MMIO region 0: MemoryRegion for the device's registers + */ + +#ifndef HW_SSI_PL022_H +#define HW_SSI_PL022_H + +#include "hw/sysbus.h" + +#define TYPE_PL022 "pl022" +#define PL022(obj) OBJECT_CHECK(PL022State, (obj), TYPE_PL022) + +typedef struct PL022State { + SysBusDevice parent_obj; + + MemoryRegion iomem; + uint32_t cr0; + uint32_t cr1; + uint32_t bitmask; + uint32_t sr; + uint32_t cpsr; + uint32_t is; + uint32_t im; + /* The FIFO head points to the next empty entry. */ + int tx_fifo_head; + int rx_fifo_head; + int tx_fifo_len; + int rx_fifo_len; + uint16_t tx_fifo[8]; + uint16_t rx_fifo[8]; + qemu_irq irq; + SSIBus *ssi; +} PL022State; + +#endif diff --git a/hw/ssi/pl022.c b/hw/ssi/pl022.c index XXXXXXX..XXXXXXX 100644 --- a/hw/ssi/pl022.c +++ b/hw/ssi/pl022.c @@ -XXX,XX +XXX,XX @@ #include "qemu/osdep.h" #include "hw/sysbus.h" +#include "hw/ssi/pl022.h" #include "hw/ssi/ssi.h" #include "qemu/log.h" @@ -XXX,XX +XXX,XX @@ do { fprintf(stderr, "pl022: error: " fmt , ## __VA_ARGS__);} while (0) #define PL022_INT_RX 0x04 #define PL022_INT_TX 0x08 -#define TYPE_PL022 "pl022" -#define PL022(obj) OBJECT_CHECK(PL022State, (obj), TYPE_PL022) - -typedef struct PL022State { - SysBusDevice parent_obj; - - MemoryRegion iomem; - uint32_t cr0; - uint32_t cr1; - uint32_t bitmask; - uint32_t sr; - uint32_t cpsr; - uint32_t is; - uint32_t im; - /* The FIFO head points to the next empty entry. */ - int tx_fifo_head; - int rx_fifo_head; - int tx_fifo_len; - int rx_fifo_len; - uint16_t tx_fifo[8]; - uint16_t rx_fifo[8]; - qemu_irq irq; - SSIBus *ssi; -} PL022State; - static const unsigned char pl022_id[8] = { 0x22, 0x10, 0x04, 0x00, 0x0d, 0xf0, 0x05, 0xb1 }; diff --git a/MAINTAINERS b/MAINTAINERS index XXXXXXX..XXXXXXX 100644 --- a/MAINTAINERS +++ b/MAINTAINERS @@ -XXX,XX +XXX,XX @@ F: hw/gpio/pl061.c F: hw/input/pl050.c F: hw/intc/pl190.c F: hw/sd/pl181.c +F: hw/ssi/pl022.c +F: include/hw/ssi/pl022.h F: hw/timer/pl031.c F: include/hw/arm/primecell.h F: hw/timer/cmsdk-apb-timer.c -- 2.18.0 Currently the PL022 calls pl022_reset() from its class init function. Make it register a DeviceState reset method instead, so that we reset the device on system reset. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Message-id: 20180820141116.9118-17-peter.maydell@linaro.org Reviewed-by: Richard Henderson <richard.henderson@linaro.org> --- hw/ssi/pl022.c | 7 +++++-- 1 file changed, 5 insertions(+), 2 deletions(-) diff --git a/hw/ssi/pl022.c b/hw/ssi/pl022.c index XXXXXXX..XXXXXXX 100644 --- a/hw/ssi/pl022.c +++ b/hw/ssi/pl022.c @@ -XXX,XX +XXX,XX @@ static void pl022_write(void *opaque, hwaddr offset, } } -static void pl022_reset(PL022State *s) +static void pl022_reset(DeviceState *dev) { + PL022State *s = PL022(dev); + s->rx_fifo_len = 0; s->tx_fifo_len = 0; s->im = 0; @@ -XXX,XX +XXX,XX @@ static int pl022_init(SysBusDevice *sbd) sysbus_init_mmio(sbd, &s->iomem); sysbus_init_irq(sbd, &s->irq); s->ssi = ssi_create_bus(dev, "ssi"); - pl022_reset(s); vmstate_register(dev, -1, &vmstate_pl022, s); return 0; } @@ -XXX,XX +XXX,XX @@ static int pl022_init(SysBusDevice *sbd) static void pl022_class_init(ObjectClass *klass, void *data) { SysBusDeviceClass *sdc = SYS_BUS_DEVICE_CLASS(klass); + DeviceClass *dc = DEVICE_CLASS(klass); sdc->init = pl022_init; + dc->reset = pl022_reset; } static const TypeInfo pl022_info = { -- 2.18.0 Use the DeviceState vmsd pointer rather than calling vmstate_register() directly. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Message-id: 20180820141116.9118-18-peter.maydell@linaro.org Reviewed-by: Richard Henderson <richard.henderson@linaro.org> --- hw/ssi/pl022.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/hw/ssi/pl022.c b/hw/ssi/pl022.c index XXXXXXX..XXXXXXX 100644 --- a/hw/ssi/pl022.c +++ b/hw/ssi/pl022.c @@ -XXX,XX +XXX,XX @@ static int pl022_init(SysBusDevice *sbd) sysbus_init_mmio(sbd, &s->iomem); sysbus_init_irq(sbd, &s->irq); s->ssi = ssi_create_bus(dev, "ssi"); - vmstate_register(dev, -1, &vmstate_pl022, s); return 0; } @@ -XXX,XX +XXX,XX @@ static void pl022_class_init(ObjectClass *klass, void *data) sdc->init = pl022_init; dc->reset = pl022_reset; + dc->vmsd = &vmstate_pl022; } static const TypeInfo pl022_info = { -- 2.18.0 Move from the legacy SysBusDevice::init method to using DeviceState::realize. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Message-id: 20180820141116.9118-19-peter.maydell@linaro.org Reviewed-by: Richard Henderson <richard.henderson@linaro.org> --- hw/ssi/pl022.c | 8 +++----- 1 file changed, 3 insertions(+), 5 deletions(-) diff --git a/hw/ssi/pl022.c b/hw/ssi/pl022.c index XXXXXXX..XXXXXXX 100644 --- a/hw/ssi/pl022.c +++ b/hw/ssi/pl022.c @@ -XXX,XX +XXX,XX @@ static const VMStateDescription vmstate_pl022 = { } }; -static int pl022_init(SysBusDevice *sbd) +static void pl022_realize(DeviceState *dev, Error **errp) { - DeviceState *dev = DEVICE(sbd); + SysBusDevice *sbd = SYS_BUS_DEVICE(dev); PL022State *s = PL022(dev); memory_region_init_io(&s->iomem, OBJECT(s), &pl022_ops, s, "pl022", 0x1000); sysbus_init_mmio(sbd, &s->iomem); sysbus_init_irq(sbd, &s->irq); s->ssi = ssi_create_bus(dev, "ssi"); - return 0; } static void pl022_class_init(ObjectClass *klass, void *data) { - SysBusDeviceClass *sdc = SYS_BUS_DEVICE_CLASS(klass); DeviceClass *dc = DEVICE_CLASS(klass); - sdc->init = pl022_init; dc->reset = pl022_reset; dc->vmsd = &vmstate_pl022; + dc->realize = pl022_realize; } static const TypeInfo pl022_info = { -- 2.18.0 The PL022 interrupt registers have bits allocated as: 0: ROR (receive overrun) 1: RT (receive timeout) 2: RX (receive FIFO half full or less) 3: TX (transmit FIFO half full or less) A cut and paste error meant we had the wrong value for the PL022_INT_RT constant. This bug doesn't affect device behaviour, because we don't implement the receive timeout feature and so never set that interrupt bit. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Message-id: 20180820141116.9118-20-peter.maydell@linaro.org Reviewed-by: Richard Henderson <richard.henderson@linaro.org> --- hw/ssi/pl022.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/hw/ssi/pl022.c b/hw/ssi/pl022.c index XXXXXXX..XXXXXXX 100644 --- a/hw/ssi/pl022.c +++ b/hw/ssi/pl022.c @@ -XXX,XX +XXX,XX @@ do { fprintf(stderr, "pl022: error: " fmt , ## __VA_ARGS__);} while (0) #define PL022_SR_BSY 0x10 #define PL022_INT_ROR 0x01 -#define PL022_INT_RT 0x04 +#define PL022_INT_RT 0x02 #define PL022_INT_RX 0x04 #define PL022_INT_TX 0x08 -- 2.18.0 In the PL022, register offset 0x20 is the ICR, a write-only interrupt-clear register. Register offset 0x24 is DMACR, the DMA control register. We were incorrectly implementing (a stub version of) DMACR at 0x20, and not implementing anything at 0x24. Fix this bug. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Message-id: 20180820141116.9118-21-peter.maydell@linaro.org Reviewed-by: Richard Henderson <richard.henderson@linaro.org> --- hw/ssi/pl022.c | 12 ++++++++++-- 1 file changed, 10 insertions(+), 2 deletions(-) diff --git a/hw/ssi/pl022.c b/hw/ssi/pl022.c index XXXXXXX..XXXXXXX 100644 --- a/hw/ssi/pl022.c +++ b/hw/ssi/pl022.c @@ -XXX,XX +XXX,XX @@ static uint64_t pl022_read(void *opaque, hwaddr offset, return s->is; case 0x1c: /* MIS */ return s->im & s->is; - case 0x20: /* DMACR */ + case 0x24: /* DMACR */ /* Not implemented. */ return 0; default: @@ -XXX,XX +XXX,XX @@ static void pl022_write(void *opaque, hwaddr offset, s->im = value; pl022_update(s); break; - case 0x20: /* DMACR */ + case 0x20: /* ICR */ + /* + * write-1-to-clear: bit 0 clears ROR, bit 1 clears RT; + * RX and TX interrupts cannot be cleared this way. + */ + value &= PL022_INT_ROR | PL022_INT_RT; + s->is &= ~value; + break; + case 0x24: /* DMACR */ if (value) { qemu_log_mask(LOG_UNIMP, "pl022: DMA not implemented\n"); } -- 2.18.0 The SPI controllers in the MPS2 AN505 board are PL022s. We have a model of the PL022, so create these devices. We don't currently model the LCD controller that sits behind one of the PL022s; the others are intended to control devices that sit on the FPGA's general purpose SPI connector or "shield" expansion connectors. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Message-id: 20180820141116.9118-22-peter.maydell@linaro.org Reviewed-by: Richard Henderson <richard.henderson@linaro.org> --- hw/arm/mps2-tz.c | 38 ++++++++++++++++++++++++++++++++------ 1 file changed, 32 insertions(+), 6 deletions(-) diff --git a/hw/arm/mps2-tz.c b/hw/arm/mps2-tz.c index XXXXXXX..XXXXXXX 100644 --- a/hw/arm/mps2-tz.c +++ b/hw/arm/mps2-tz.c @@ -XXX,XX +XXX,XX @@ #include "hw/misc/tz-msc.h" #include "hw/arm/iotkit.h" #include "hw/dma/pl080.h" +#include "hw/ssi/pl022.h" #include "hw/devices.h" #include "net/net.h" #include "hw/core/split-irq.h" @@ -XXX,XX +XXX,XX @@ typedef struct { MPS2FPGAIO fpgaio; TZPPC ppc[5]; TZMPC ssram_mpc[3]; - UnimplementedDeviceState spi[5]; + PL022State spi[5]; UnimplementedDeviceState i2c[4]; UnimplementedDeviceState i2s_audio; UnimplementedDeviceState gpio[4]; @@ -XXX,XX +XXX,XX @@ static MemoryRegion *make_dma(MPS2TZMachineState *mms, void *opaque, return sysbus_mmio_get_region(s, 0); } +static MemoryRegion *make_spi(MPS2TZMachineState *mms, void *opaque, + const char *name, hwaddr size) +{ + /* + * The AN505 has five PL022 SPI controllers. + * One of these should have the LCD controller behind it; the others + * are connected only to the FPGA's "general purpose SPI connector" + * or "shield" expansion connectors. + * Note that if we do implement devices behind SPI, the chip select + * lines are set via the "MISC" register in the MPS2 FPGAIO device. + */ + PL022State *spi = opaque; + int i = spi - &mms->spi[0]; + DeviceState *iotkitdev = DEVICE(&mms->iotkit); + SysBusDevice *s; + + sysbus_init_child_obj(OBJECT(mms), name, spi, sizeof(mms->spi[0]), + TYPE_PL022); + object_property_set_bool(OBJECT(spi), true, "realized", &error_fatal); + s = SYS_BUS_DEVICE(spi); + sysbus_connect_irq(s, 0, + qdev_get_gpio_in_named(iotkitdev, "EXP_IRQ", 51 + i)); + return sysbus_mmio_get_region(s, 0); +} + static void mps2tz_common_init(MachineState *machine) { MPS2TZMachineState *mms = MPS2TZ_MACHINE(machine); @@ -XXX,XX +XXX,XX @@ static void mps2tz_common_init(MachineState *machine) }, { .name = "apb_ppcexp1", .ports = { - { "spi0", make_unimp_dev, &mms->spi[0], 0x40205000, 0x1000 }, - { "spi1", make_unimp_dev, &mms->spi[1], 0x40206000, 0x1000 }, - { "spi2", make_unimp_dev, &mms->spi[2], 0x40209000, 0x1000 }, - { "spi3", make_unimp_dev, &mms->spi[3], 0x4020a000, 0x1000 }, - { "spi4", make_unimp_dev, &mms->spi[4], 0x4020b000, 0x1000 }, + { "spi0", make_spi, &mms->spi[0], 0x40205000, 0x1000 }, + { "spi1", make_spi, &mms->spi[1], 0x40206000, 0x1000 }, + { "spi2", make_spi, &mms->spi[2], 0x40209000, 0x1000 }, + { "spi3", make_spi, &mms->spi[3], 0x4020a000, 0x1000 }, + { "spi4", make_spi, &mms->spi[4], 0x4020b000, 0x1000 }, { "uart0", make_uart, &mms->uart[0], 0x40200000, 0x1000 }, { "uart1", make_uart, &mms->uart[1], 0x40201000, 0x1000 }, { "uart2", make_uart, &mms->uart[2], 0x40202000, 0x1000 }, -- 2.18.0 Some of the config register values we were setting for the MPS2 SCC weren't correct: * the SCC_AID bits [23:20] specify the FPGA build target board revision, and the SCC_CFG4 register specifies the actual board revision, so these should have matching values. Claim to be board revision C, consistently -- we had the revision in the wrong part of SCC_AID. * SCC_ID bits [15:4] should be 0x505, not decimal 505 Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Message-id: 20180820141116.9118-23-peter.maydell@linaro.org Reviewed-by: Richard Henderson <richard.henderson@linaro.org> --- hw/arm/mps2-tz.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/hw/arm/mps2-tz.c b/hw/arm/mps2-tz.c index XXXXXXX..XXXXXXX 100644 --- a/hw/arm/mps2-tz.c +++ b/hw/arm/mps2-tz.c @@ -XXX,XX +XXX,XX @@ static MemoryRegion *make_scc(MPS2TZMachineState *mms, void *opaque, sccdev = DEVICE(scc); qdev_set_parent_bus(sccdev, sysbus_get_default()); qdev_prop_set_uint32(sccdev, "scc-cfg4", 0x2); - qdev_prop_set_uint32(sccdev, "scc-aid", 0x02000008); + qdev_prop_set_uint32(sccdev, "scc-aid", 0x00200008); qdev_prop_set_uint32(sccdev, "scc-id", mmc->scc_id); object_property_set_bool(OBJECT(scc), true, "realized", &error_fatal); return sysbus_mmio_get_region(SYS_BUS_DEVICE(sccdev), 0); @@ -XXX,XX +XXX,XX @@ static void mps2tz_an505_class_init(ObjectClass *oc, void *data) mc->desc = "ARM MPS2 with AN505 FPGA image for Cortex-M33"; mmc->fpga_type = FPGA_AN505; mc->default_cpu_type = ARM_CPU_TYPE_NAME("cortex-m33"); - mmc->scc_id = 0x41040000 | (505 << 4); + mmc->scc_id = 0x41045050; } static const TypeInfo mps2tz_info = { -- 2.18.0 Untabify the arm translate.c. This affects only some lines, mostly comments, in the iwMMXt code. We've never touched that code in years, so it's not going to get fixed up by our "change when touched" process, and a bulk change is not going to be too disruptive. This commit was produced using Emacs "untabify"; it is a whitespace-only change. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Message-id: 20180821165215.29069-2-peter.maydell@linaro.org --- target/arm/translate.c | 122 ++++++++++++++++++++--------------------- 1 file changed, 61 insertions(+), 61 deletions(-) diff --git a/target/arm/translate.c b/target/arm/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate.c +++ b/target/arm/translate.c @@ -XXX,XX +XXX,XX @@ static inline void gen_mov_vreg_F0(int dp, int reg) tcg_gen_st_f32(cpu_F0s, cpu_env, vfp_reg_offset(dp, reg)); } -#define ARM_CP_RW_BIT (1 << 20) +#define ARM_CP_RW_BIT (1 << 20) static inline void iwmmxt_load_reg(TCGv_i64 var, int reg) { @@ -XXX,XX +XXX,XX @@ static int disas_iwmmxt_insn(DisasContext *s, uint32_t insn) wrd = insn & 0xf; rdlo = (insn >> 12) & 0xf; rdhi = (insn >> 16) & 0xf; - if (insn & ARM_CP_RW_BIT) { /* TMRRC */ + if (insn & ARM_CP_RW_BIT) { /* TMRRC */ iwmmxt_load_reg(cpu_V0, wrd); tcg_gen_extrl_i64_i32(cpu_R[rdlo], cpu_V0); tcg_gen_shri_i64(cpu_V0, cpu_V0, 32); tcg_gen_extrl_i64_i32(cpu_R[rdhi], cpu_V0); - } else { /* TMCRR */ + } else { /* TMCRR */ tcg_gen_concat_i32_i64(cpu_V0, cpu_R[rdlo], cpu_R[rdhi]); iwmmxt_store_reg(cpu_V0, wrd); gen_op_iwmmxt_set_mup(); @@ -XXX,XX +XXX,XX @@ static int disas_iwmmxt_insn(DisasContext *s, uint32_t insn) return 1; } if (insn & ARM_CP_RW_BIT) { - if ((insn >> 28) == 0xf) { /* WLDRW wCx */ + if ((insn >> 28) == 0xf) { /* WLDRW wCx */ tmp = tcg_temp_new_i32(); gen_aa32_ld32u(s, tmp, addr, get_mem_index(s)); iwmmxt_store_creg(wrd, tmp); } else { i = 1; if (insn & (1 << 8)) { - if (insn & (1 << 22)) { /* WLDRD */ + if (insn & (1 << 22)) { /* WLDRD */ gen_aa32_ld64(s, cpu_M0, addr, get_mem_index(s)); i = 0; - } else { /* WLDRW wRd */ + } else { /* WLDRW wRd */ tmp = tcg_temp_new_i32(); gen_aa32_ld32u(s, tmp, addr, get_mem_index(s)); } } else { tmp = tcg_temp_new_i32(); - if (insn & (1 << 22)) { /* WLDRH */ + if (insn & (1 << 22)) { /* WLDRH */ gen_aa32_ld16u(s, tmp, addr, get_mem_index(s)); - } else { /* WLDRB */ + } else { /* WLDRB */ gen_aa32_ld8u(s, tmp, addr, get_mem_index(s)); } } @@ -XXX,XX +XXX,XX @@ static int disas_iwmmxt_insn(DisasContext *s, uint32_t insn) gen_op_iwmmxt_movq_wRn_M0(wrd); } } else { - if ((insn >> 28) == 0xf) { /* WSTRW wCx */ + if ((insn >> 28) == 0xf) { /* WSTRW wCx */ tmp = iwmmxt_load_creg(wrd); gen_aa32_st32(s, tmp, addr, get_mem_index(s)); } else { gen_op_iwmmxt_movq_M0_wRn(wrd); tmp = tcg_temp_new_i32(); if (insn & (1 << 8)) { - if (insn & (1 << 22)) { /* WSTRD */ + if (insn & (1 << 22)) { /* WSTRD */ gen_aa32_st64(s, cpu_M0, addr, get_mem_index(s)); - } else { /* WSTRW wRd */ + } else { /* WSTRW wRd */ tcg_gen_extrl_i64_i32(tmp, cpu_M0); gen_aa32_st32(s, tmp, addr, get_mem_index(s)); } } else { - if (insn & (1 << 22)) { /* WSTRH */ + if (insn & (1 << 22)) { /* WSTRH */ tcg_gen_extrl_i64_i32(tmp, cpu_M0); gen_aa32_st16(s, tmp, addr, get_mem_index(s)); - } else { /* WSTRB */ + } else { /* WSTRB */ tcg_gen_extrl_i64_i32(tmp, cpu_M0); gen_aa32_st8(s, tmp, addr, get_mem_index(s)); } @@ -XXX,XX +XXX,XX @@ static int disas_iwmmxt_insn(DisasContext *s, uint32_t insn) return 1; switch (((insn >> 12) & 0xf00) | ((insn >> 4) & 0xff)) { - case 0x000: /* WOR */ + case 0x000: /* WOR */ wrd = (insn >> 12) & 0xf; rd0 = (insn >> 0) & 0xf; rd1 = (insn >> 16) & 0xf; @@ -XXX,XX +XXX,XX @@ static int disas_iwmmxt_insn(DisasContext *s, uint32_t insn) gen_op_iwmmxt_set_mup(); gen_op_iwmmxt_set_cup(); break; - case 0x011: /* TMCR */ + case 0x011: /* TMCR */ if (insn & 0xf) return 1; rd = (insn >> 12) & 0xf; @@ -XXX,XX +XXX,XX @@ static int disas_iwmmxt_insn(DisasContext *s, uint32_t insn) return 1; } break; - case 0x100: /* WXOR */ + case 0x100: /* WXOR */ wrd = (insn >> 12) & 0xf; rd0 = (insn >> 0) & 0xf; rd1 = (insn >> 16) & 0xf; @@ -XXX,XX +XXX,XX @@ static int disas_iwmmxt_insn(DisasContext *s, uint32_t insn) gen_op_iwmmxt_set_mup(); gen_op_iwmmxt_set_cup(); break; - case 0x111: /* TMRC */ + case 0x111: /* TMRC */ if (insn & 0xf) return 1; rd = (insn >> 12) & 0xf; @@ -XXX,XX +XXX,XX @@ static int disas_iwmmxt_insn(DisasContext *s, uint32_t insn) tmp = iwmmxt_load_creg(wrd); store_reg(s, rd, tmp); break; - case 0x300: /* WANDN */ + case 0x300: /* WANDN */ wrd = (insn >> 12) & 0xf; rd0 = (insn >> 0) & 0xf; rd1 = (insn >> 16) & 0xf; @@ -XXX,XX +XXX,XX @@ static int disas_iwmmxt_insn(DisasContext *s, uint32_t insn) gen_op_iwmmxt_set_mup(); gen_op_iwmmxt_set_cup(); break; - case 0x200: /* WAND */ + case 0x200: /* WAND */ wrd = (insn >> 12) & 0xf; rd0 = (insn >> 0) & 0xf; rd1 = (insn >> 16) & 0xf; @@ -XXX,XX +XXX,XX @@ static int disas_iwmmxt_insn(DisasContext *s, uint32_t insn) gen_op_iwmmxt_set_mup(); gen_op_iwmmxt_set_cup(); break; - case 0x810: case 0xa10: /* WMADD */ + case 0x810: case 0xa10: /* WMADD */ wrd = (insn >> 12) & 0xf; rd0 = (insn >> 0) & 0xf; rd1 = (insn >> 16) & 0xf; @@ -XXX,XX +XXX,XX @@ static int disas_iwmmxt_insn(DisasContext *s, uint32_t insn) gen_op_iwmmxt_movq_wRn_M0(wrd); gen_op_iwmmxt_set_mup(); break; - case 0x10e: case 0x50e: case 0x90e: case 0xd0e: /* WUNPCKIL */ + case 0x10e: case 0x50e: case 0x90e: case 0xd0e: /* WUNPCKIL */ wrd = (insn >> 12) & 0xf; rd0 = (insn >> 16) & 0xf; rd1 = (insn >> 0) & 0xf; @@ -XXX,XX +XXX,XX @@ static int disas_iwmmxt_insn(DisasContext *s, uint32_t insn) gen_op_iwmmxt_set_mup(); gen_op_iwmmxt_set_cup(); break; - case 0x10c: case 0x50c: case 0x90c: case 0xd0c: /* WUNPCKIH */ + case 0x10c: case 0x50c: case 0x90c: case 0xd0c: /* WUNPCKIH */ wrd = (insn >> 12) & 0xf; rd0 = (insn >> 16) & 0xf; rd1 = (insn >> 0) & 0xf; @@ -XXX,XX +XXX,XX @@ static int disas_iwmmxt_insn(DisasContext *s, uint32_t insn) gen_op_iwmmxt_set_mup(); gen_op_iwmmxt_set_cup(); break; - case 0x012: case 0x112: case 0x412: case 0x512: /* WSAD */ + case 0x012: case 0x112: case 0x412: case 0x512: /* WSAD */ wrd = (insn >> 12) & 0xf; rd0 = (insn >> 16) & 0xf; rd1 = (insn >> 0) & 0xf; @@ -XXX,XX +XXX,XX @@ static int disas_iwmmxt_insn(DisasContext *s, uint32_t insn) gen_op_iwmmxt_movq_wRn_M0(wrd); gen_op_iwmmxt_set_mup(); break; - case 0x010: case 0x110: case 0x210: case 0x310: /* WMUL */ + case 0x010: case 0x110: case 0x210: case 0x310: /* WMUL */ wrd = (insn >> 12) & 0xf; rd0 = (insn >> 16) & 0xf; rd1 = (insn >> 0) & 0xf; @@ -XXX,XX +XXX,XX @@ static int disas_iwmmxt_insn(DisasContext *s, uint32_t insn) gen_op_iwmmxt_movq_wRn_M0(wrd); gen_op_iwmmxt_set_mup(); break; - case 0x410: case 0x510: case 0x610: case 0x710: /* WMAC */ + case 0x410: case 0x510: case 0x610: case 0x710: /* WMAC */ wrd = (insn >> 12) & 0xf; rd0 = (insn >> 16) & 0xf; rd1 = (insn >> 0) & 0xf; @@ -XXX,XX +XXX,XX @@ static int disas_iwmmxt_insn(DisasContext *s, uint32_t insn) gen_op_iwmmxt_movq_wRn_M0(wrd); gen_op_iwmmxt_set_mup(); break; - case 0x006: case 0x406: case 0x806: case 0xc06: /* WCMPEQ */ + case 0x006: case 0x406: case 0x806: case 0xc06: /* WCMPEQ */ wrd = (insn >> 12) & 0xf; rd0 = (insn >> 16) & 0xf; rd1 = (insn >> 0) & 0xf; @@ -XXX,XX +XXX,XX @@ static int disas_iwmmxt_insn(DisasContext *s, uint32_t insn) gen_op_iwmmxt_set_mup(); gen_op_iwmmxt_set_cup(); break; - case 0x800: case 0x900: case 0xc00: case 0xd00: /* WAVG2 */ + case 0x800: case 0x900: case 0xc00: case 0xd00: /* WAVG2 */ wrd = (insn >> 12) & 0xf; rd0 = (insn >> 16) & 0xf; rd1 = (insn >> 0) & 0xf; @@ -XXX,XX +XXX,XX @@ static int disas_iwmmxt_insn(DisasContext *s, uint32_t insn) gen_op_iwmmxt_set_mup(); gen_op_iwmmxt_set_cup(); break; - case 0x802: case 0x902: case 0xa02: case 0xb02: /* WALIGNR */ + case 0x802: case 0x902: case 0xa02: case 0xb02: /* WALIGNR */ wrd = (insn >> 12) & 0xf; rd0 = (insn >> 16) & 0xf; rd1 = (insn >> 0) & 0xf; @@ -XXX,XX +XXX,XX @@ static int disas_iwmmxt_insn(DisasContext *s, uint32_t insn) gen_op_iwmmxt_movq_wRn_M0(wrd); gen_op_iwmmxt_set_mup(); break; - case 0x601: case 0x605: case 0x609: case 0x60d: /* TINSR */ + case 0x601: case 0x605: case 0x609: case 0x60d: /* TINSR */ if (((insn >> 6) & 3) == 3) return 1; rd = (insn >> 12) & 0xf; @@ -XXX,XX +XXX,XX @@ static int disas_iwmmxt_insn(DisasContext *s, uint32_t insn) gen_op_iwmmxt_movq_wRn_M0(wrd); gen_op_iwmmxt_set_mup(); break; - case 0x107: case 0x507: case 0x907: case 0xd07: /* TEXTRM */ + case 0x107: case 0x507: case 0x907: case 0xd07: /* TEXTRM */ rd = (insn >> 12) & 0xf; wrd = (insn >> 16) & 0xf; if (rd == 15 || ((insn >> 22) & 3) == 3) @@ -XXX,XX +XXX,XX @@ static int disas_iwmmxt_insn(DisasContext *s, uint32_t insn) } store_reg(s, rd, tmp); break; - case 0x117: case 0x517: case 0x917: case 0xd17: /* TEXTRC */ + case 0x117: case 0x517: case 0x917: case 0xd17: /* TEXTRC */ if ((insn & 0x000ff008) != 0x0003f000 || ((insn >> 22) & 3) == 3) return 1; tmp = iwmmxt_load_creg(ARM_IWMMXT_wCASF); @@ -XXX,XX +XXX,XX @@ static int disas_iwmmxt_insn(DisasContext *s, uint32_t insn) gen_set_nzcv(tmp); tcg_temp_free_i32(tmp); break; - case 0x401: case 0x405: case 0x409: case 0x40d: /* TBCST */ + case 0x401: case 0x405: case 0x409: case 0x40d: /* TBCST */ if (((insn >> 6) & 3) == 3) return 1; rd = (insn >> 12) & 0xf; @@ -XXX,XX +XXX,XX @@ static int disas_iwmmxt_insn(DisasContext *s, uint32_t insn) gen_op_iwmmxt_movq_wRn_M0(wrd); gen_op_iwmmxt_set_mup(); break; - case 0x113: case 0x513: case 0x913: case 0xd13: /* TANDC */ + case 0x113: case 0x513: case 0x913: case 0xd13: /* TANDC */ if ((insn & 0x000ff00f) != 0x0003f000 || ((insn >> 22) & 3) == 3) return 1; tmp = iwmmxt_load_creg(ARM_IWMMXT_wCASF); @@ -XXX,XX +XXX,XX @@ static int disas_iwmmxt_insn(DisasContext *s, uint32_t insn) tcg_temp_free_i32(tmp2); tcg_temp_free_i32(tmp); break; - case 0x01c: case 0x41c: case 0x81c: case 0xc1c: /* WACC */ + case 0x01c: case 0x41c: case 0x81c: case 0xc1c: /* WACC */ wrd = (insn >> 12) & 0xf; rd0 = (insn >> 16) & 0xf; gen_op_iwmmxt_movq_M0_wRn(rd0); @@ -XXX,XX +XXX,XX @@ static int disas_iwmmxt_insn(DisasContext *s, uint32_t insn) gen_op_iwmmxt_movq_wRn_M0(wrd); gen_op_iwmmxt_set_mup(); break; - case 0x115: case 0x515: case 0x915: case 0xd15: /* TORC */ + case 0x115: case 0x515: case 0x915: case 0xd15: /* TORC */ if ((insn & 0x000ff00f) != 0x0003f000 || ((insn >> 22) & 3) == 3) return 1; tmp = iwmmxt_load_creg(ARM_IWMMXT_wCASF); @@ -XXX,XX +XXX,XX @@ static int disas_iwmmxt_insn(DisasContext *s, uint32_t insn) tcg_temp_free_i32(tmp2); tcg_temp_free_i32(tmp); break; - case 0x103: case 0x503: case 0x903: case 0xd03: /* TMOVMSK */ + case 0x103: case 0x503: case 0x903: case 0xd03: /* TMOVMSK */ rd = (insn >> 12) & 0xf; rd0 = (insn >> 16) & 0xf; if ((insn & 0xf) != 0 || ((insn >> 22) & 3) == 3) @@ -XXX,XX +XXX,XX @@ static int disas_iwmmxt_insn(DisasContext *s, uint32_t insn) } store_reg(s, rd, tmp); break; - case 0x106: case 0x306: case 0x506: case 0x706: /* WCMPGT */ + case 0x106: case 0x306: case 0x506: case 0x706: /* WCMPGT */ case 0x906: case 0xb06: case 0xd06: case 0xf06: wrd = (insn >> 12) & 0xf; rd0 = (insn >> 16) & 0xf; @@ -XXX,XX +XXX,XX @@ static int disas_iwmmxt_insn(DisasContext *s, uint32_t insn) gen_op_iwmmxt_set_mup(); gen_op_iwmmxt_set_cup(); break; - case 0x00e: case 0x20e: case 0x40e: case 0x60e: /* WUNPCKEL */ + case 0x00e: case 0x20e: case 0x40e: case 0x60e: /* WUNPCKEL */ case 0x80e: case 0xa0e: case 0xc0e: case 0xe0e: wrd = (insn >> 12) & 0xf; rd0 = (insn >> 16) & 0xf; @@ -XXX,XX +XXX,XX @@ static int disas_iwmmxt_insn(DisasContext *s, uint32_t insn) gen_op_iwmmxt_set_mup(); gen_op_iwmmxt_set_cup(); break; - case 0x00c: case 0x20c: case 0x40c: case 0x60c: /* WUNPCKEH */ + case 0x00c: case 0x20c: case 0x40c: case 0x60c: /* WUNPCKEH */ case 0x80c: case 0xa0c: case 0xc0c: case 0xe0c: wrd = (insn >> 12) & 0xf; rd0 = (insn >> 16) & 0xf; @@ -XXX,XX +XXX,XX @@ static int disas_iwmmxt_insn(DisasContext *s, uint32_t insn) gen_op_iwmmxt_set_mup(); gen_op_iwmmxt_set_cup(); break; - case 0x204: case 0x604: case 0xa04: case 0xe04: /* WSRL */ + case 0x204: case 0x604: case 0xa04: case 0xe04: /* WSRL */ case 0x214: case 0x614: case 0xa14: case 0xe14: if (((insn >> 22) & 3) == 0) return 1; @@ -XXX,XX +XXX,XX @@ static int disas_iwmmxt_insn(DisasContext *s, uint32_t insn) gen_op_iwmmxt_set_mup(); gen_op_iwmmxt_set_cup(); break; - case 0x004: case 0x404: case 0x804: case 0xc04: /* WSRA */ + case 0x004: case 0x404: case 0x804: case 0xc04: /* WSRA */ case 0x014: case 0x414: case 0x814: case 0xc14: if (((insn >> 22) & 3) == 0) return 1; @@ -XXX,XX +XXX,XX @@ static int disas_iwmmxt_insn(DisasContext *s, uint32_t insn) gen_op_iwmmxt_set_mup(); gen_op_iwmmxt_set_cup(); break; - case 0x104: case 0x504: case 0x904: case 0xd04: /* WSLL */ + case 0x104: case 0x504: case 0x904: case 0xd04: /* WSLL */ case 0x114: case 0x514: case 0x914: case 0xd14: if (((insn >> 22) & 3) == 0) return 1; @@ -XXX,XX +XXX,XX @@ static int disas_iwmmxt_insn(DisasContext *s, uint32_t insn) gen_op_iwmmxt_set_mup(); gen_op_iwmmxt_set_cup(); break; - case 0x304: case 0x704: case 0xb04: case 0xf04: /* WROR */ + case 0x304: case 0x704: case 0xb04: case 0xf04: /* WROR */ case 0x314: case 0x714: case 0xb14: case 0xf14: if (((insn >> 22) & 3) == 0) return 1; @@ -XXX,XX +XXX,XX @@ static int disas_iwmmxt_insn(DisasContext *s, uint32_t insn) gen_op_iwmmxt_set_mup(); gen_op_iwmmxt_set_cup(); break; - case 0x116: case 0x316: case 0x516: case 0x716: /* WMIN */ + case 0x116: case 0x316: case 0x516: case 0x716: /* WMIN */ case 0x916: case 0xb16: case 0xd16: case 0xf16: wrd = (insn >> 12) & 0xf; rd0 = (insn >> 16) & 0xf; @@ -XXX,XX +XXX,XX @@ static int disas_iwmmxt_insn(DisasContext *s, uint32_t insn) gen_op_iwmmxt_movq_wRn_M0(wrd); gen_op_iwmmxt_set_mup(); break; - case 0x016: case 0x216: case 0x416: case 0x616: /* WMAX */ + case 0x016: case 0x216: case 0x416: case 0x616: /* WMAX */ case 0x816: case 0xa16: case 0xc16: case 0xe16: wrd = (insn >> 12) & 0xf; rd0 = (insn >> 16) & 0xf; @@ -XXX,XX +XXX,XX @@ static int disas_iwmmxt_insn(DisasContext *s, uint32_t insn) gen_op_iwmmxt_movq_wRn_M0(wrd); gen_op_iwmmxt_set_mup(); break; - case 0x002: case 0x102: case 0x202: case 0x302: /* WALIGNI */ + case 0x002: case 0x102: case 0x202: case 0x302: /* WALIGNI */ case 0x402: case 0x502: case 0x602: case 0x702: wrd = (insn >> 12) & 0xf; rd0 = (insn >> 16) & 0xf; @@ -XXX,XX +XXX,XX @@ static int disas_iwmmxt_insn(DisasContext *s, uint32_t insn) gen_op_iwmmxt_movq_wRn_M0(wrd); gen_op_iwmmxt_set_mup(); break; - case 0x01a: case 0x11a: case 0x21a: case 0x31a: /* WSUB */ + case 0x01a: case 0x11a: case 0x21a: case 0x31a: /* WSUB */ case 0x41a: case 0x51a: case 0x61a: case 0x71a: case 0x81a: case 0x91a: case 0xa1a: case 0xb1a: case 0xc1a: case 0xd1a: case 0xe1a: case 0xf1a: @@ -XXX,XX +XXX,XX @@ static int disas_iwmmxt_insn(DisasContext *s, uint32_t insn) gen_op_iwmmxt_set_mup(); gen_op_iwmmxt_set_cup(); break; - case 0x01e: case 0x11e: case 0x21e: case 0x31e: /* WSHUFH */ + case 0x01e: case 0x11e: case 0x21e: case 0x31e: /* WSHUFH */ case 0x41e: case 0x51e: case 0x61e: case 0x71e: case 0x81e: case 0x91e: case 0xa1e: case 0xb1e: case 0xc1e: case 0xd1e: case 0xe1e: case 0xf1e: @@ -XXX,XX +XXX,XX @@ static int disas_iwmmxt_insn(DisasContext *s, uint32_t insn) gen_op_iwmmxt_set_mup(); gen_op_iwmmxt_set_cup(); break; - case 0x018: case 0x118: case 0x218: case 0x318: /* WADD */ + case 0x018: case 0x118: case 0x218: case 0x318: /* WADD */ case 0x418: case 0x518: case 0x618: case 0x718: case 0x818: case 0x918: case 0xa18: case 0xb18: case 0xc18: case 0xd18: case 0xe18: case 0xf18: @@ -XXX,XX +XXX,XX @@ static int disas_iwmmxt_insn(DisasContext *s, uint32_t insn) gen_op_iwmmxt_set_mup(); gen_op_iwmmxt_set_cup(); break; - case 0x008: case 0x108: case 0x208: case 0x308: /* WPACK */ + case 0x008: case 0x108: case 0x208: case 0x308: /* WPACK */ case 0x408: case 0x508: case 0x608: case 0x708: case 0x808: case 0x908: case 0xa08: case 0xb08: case 0xc08: case 0xd08: case 0xe08: case 0xf08: @@ -XXX,XX +XXX,XX @@ static int disas_iwmmxt_insn(DisasContext *s, uint32_t insn) tmp = load_reg(s, rd0); tmp2 = load_reg(s, rd1); switch ((insn >> 16) & 0xf) { - case 0x0: /* TMIA */ + case 0x0: /* TMIA */ gen_helper_iwmmxt_muladdsl(cpu_M0, cpu_M0, tmp, tmp2); break; - case 0x8: /* TMIAPH */ + case 0x8: /* TMIAPH */ gen_helper_iwmmxt_muladdsw(cpu_M0, cpu_M0, tmp, tmp2); break; - case 0xc: case 0xd: case 0xe: case 0xf: /* TMIAxy */ + case 0xc: case 0xd: case 0xe: case 0xf: /* TMIAxy */ if (insn & (1 << 16)) tcg_gen_shri_i32(tmp, tmp, 16); if (insn & (1 << 17)) @@ -XXX,XX +XXX,XX @@ static int disas_dsp_insn(DisasContext *s, uint32_t insn) tmp = load_reg(s, rd0); tmp2 = load_reg(s, rd1); switch ((insn >> 16) & 0xf) { - case 0x0: /* MIA */ + case 0x0: /* MIA */ gen_helper_iwmmxt_muladdsl(cpu_M0, cpu_M0, tmp, tmp2); break; - case 0x8: /* MIAPH */ + case 0x8: /* MIAPH */ gen_helper_iwmmxt_muladdsw(cpu_M0, cpu_M0, tmp, tmp2); break; - case 0xc: /* MIABB */ - case 0xd: /* MIABT */ - case 0xe: /* MIATB */ - case 0xf: /* MIATT */ + case 0xc: /* MIABB */ + case 0xd: /* MIABT */ + case 0xe: /* MIATB */ + case 0xf: /* MIATT */ if (insn & (1 << 16)) tcg_gen_shri_i32(tmp, tmp, 16); if (insn & (1 << 17)) @@ -XXX,XX +XXX,XX @@ static int disas_dsp_insn(DisasContext *s, uint32_t insn) if (acc != 0) return 1; - if (insn & ARM_CP_RW_BIT) { /* MRA */ + if (insn & ARM_CP_RW_BIT) { /* MRA */ iwmmxt_load_reg(cpu_V0, acc); tcg_gen_extrl_i64_i32(cpu_R[rdlo], cpu_V0); tcg_gen_shri_i64(cpu_V0, cpu_V0, 32); tcg_gen_extrl_i64_i32(cpu_R[rdhi], cpu_V0); tcg_gen_andi_i32(cpu_R[rdhi], cpu_R[rdhi], (1 << (40 - 32)) - 1); - } else { /* MAR */ + } else { /* MAR */ tcg_gen_concat_i32_i64(cpu_V0, cpu_R[rdlo], cpu_R[rdhi]); iwmmxt_store_reg(cpu_V0, acc); } -- 2.18.0 Untabify the arm iwmmxt_helper.c. This affects only the iwMMXt code. We've never touched that code in years, so it's not going to get fixed up by our "change when touched" process, and a bulk change is not going to be too disruptive. This commit was produced using Emacs "untabify" (plus one by-hand removal of a space to fix a checkpatch nit); it is a whitespace-only change. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Message-id: 20180821165215.29069-3-peter.maydell@linaro.org --- target/arm/iwmmxt_helper.c | 234 ++++++++++++++++++------------------- 1 file changed, 117 insertions(+), 117 deletions(-) diff --git a/target/arm/iwmmxt_helper.c b/target/arm/iwmmxt_helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/iwmmxt_helper.c +++ b/target/arm/iwmmxt_helper.c @@ -XXX,XX +XXX,XX @@ /* iwMMXt macros extracted from GNU gdb. */ /* Set the SIMD wCASF flags for 8, 16, 32 or 64-bit operations. */ -#define SIMD8_SET( v, n, b) ((v != 0) << ((((b) + 1) * 4) + (n))) -#define SIMD16_SET(v, n, h) ((v != 0) << ((((h) + 1) * 8) + (n))) -#define SIMD32_SET(v, n, w) ((v != 0) << ((((w) + 1) * 16) + (n))) -#define SIMD64_SET(v, n) ((v != 0) << (32 + (n))) +#define SIMD8_SET(v, n, b) ((v != 0) << ((((b) + 1) * 4) + (n))) +#define SIMD16_SET(v, n, h) ((v != 0) << ((((h) + 1) * 8) + (n))) +#define SIMD32_SET(v, n, w) ((v != 0) << ((((w) + 1) * 16) + (n))) +#define SIMD64_SET(v, n) ((v != 0) << (32 + (n))) /* Flags to pass as "n" above. */ -#define SIMD_NBIT -1 -#define SIMD_ZBIT -2 -#define SIMD_CBIT -3 -#define SIMD_VBIT -4 +#define SIMD_NBIT -1 +#define SIMD_ZBIT -2 +#define SIMD_CBIT -3 +#define SIMD_VBIT -4 /* Various status bit macros. */ -#define NBIT8(x) ((x) & 0x80) -#define NBIT16(x) ((x) & 0x8000) -#define NBIT32(x) ((x) & 0x80000000) -#define NBIT64(x) ((x) & 0x8000000000000000ULL) -#define ZBIT8(x) (((x) & 0xff) == 0) -#define ZBIT16(x) (((x) & 0xffff) == 0) -#define ZBIT32(x) (((x) & 0xffffffff) == 0) -#define ZBIT64(x) (x == 0) +#define NBIT8(x) ((x) & 0x80) +#define NBIT16(x) ((x) & 0x8000) +#define NBIT32(x) ((x) & 0x80000000) +#define NBIT64(x) ((x) & 0x8000000000000000ULL) +#define ZBIT8(x) (((x) & 0xff) == 0) +#define ZBIT16(x) (((x) & 0xffff) == 0) +#define ZBIT32(x) (((x) & 0xffffffff) == 0) +#define ZBIT64(x) (x == 0) /* Sign extension macros. */ -#define EXTEND8H(a) ((uint16_t) (int8_t) (a)) -#define EXTEND8(a) ((uint32_t) (int8_t) (a)) -#define EXTEND16(a) ((uint32_t) (int16_t) (a)) -#define EXTEND16S(a) ((int32_t) (int16_t) (a)) -#define EXTEND32(a) ((uint64_t) (int32_t) (a)) +#define EXTEND8H(a) ((uint16_t) (int8_t) (a)) +#define EXTEND8(a) ((uint32_t) (int8_t) (a)) +#define EXTEND16(a) ((uint32_t) (int16_t) (a)) +#define EXTEND16S(a) ((int32_t) (int16_t) (a)) +#define EXTEND32(a) ((uint64_t) (int32_t) (a)) uint64_t HELPER(iwmmxt_maddsq)(uint64_t a, uint64_t b) { @@ -XXX,XX +XXX,XX @@ uint64_t HELPER(iwmmxt_macuw)(uint64_t a, uint64_t b) #define NZBIT64(x) \ SIMD64_SET(NBIT64(x), SIMD_NBIT) | \ SIMD64_SET(ZBIT64(x), SIMD_ZBIT) -#define IWMMXT_OP_UNPACK(S, SH0, SH1, SH2, SH3) \ +#define IWMMXT_OP_UNPACK(S, SH0, SH1, SH2, SH3) \ uint64_t HELPER(glue(iwmmxt_unpack, glue(S, b)))(CPUARMState *env, \ uint64_t a, uint64_t b) \ -{ \ - a = \ - (((a >> SH0) & 0xff) << 0) | (((b >> SH0) & 0xff) << 8) | \ - (((a >> SH1) & 0xff) << 16) | (((b >> SH1) & 0xff) << 24) | \ - (((a >> SH2) & 0xff) << 32) | (((b >> SH2) & 0xff) << 40) | \ - (((a >> SH3) & 0xff) << 48) | (((b >> SH3) & 0xff) << 56); \ - env->iwmmxt.cregs[ARM_IWMMXT_wCASF] = \ - NZBIT8(a >> 0, 0) | NZBIT8(a >> 8, 1) | \ - NZBIT8(a >> 16, 2) | NZBIT8(a >> 24, 3) | \ - NZBIT8(a >> 32, 4) | NZBIT8(a >> 40, 5) | \ - NZBIT8(a >> 48, 6) | NZBIT8(a >> 56, 7); \ +{ \ + a = \ + (((a >> SH0) & 0xff) << 0) | (((b >> SH0) & 0xff) << 8) | \ + (((a >> SH1) & 0xff) << 16) | (((b >> SH1) & 0xff) << 24) | \ + (((a >> SH2) & 0xff) << 32) | (((b >> SH2) & 0xff) << 40) | \ + (((a >> SH3) & 0xff) << 48) | (((b >> SH3) & 0xff) << 56); \ + env->iwmmxt.cregs[ARM_IWMMXT_wCASF] = \ + NZBIT8(a >> 0, 0) | NZBIT8(a >> 8, 1) | \ + NZBIT8(a >> 16, 2) | NZBIT8(a >> 24, 3) | \ + NZBIT8(a >> 32, 4) | NZBIT8(a >> 40, 5) | \ + NZBIT8(a >> 48, 6) | NZBIT8(a >> 56, 7); \ return a; \ -} \ +} \ uint64_t HELPER(glue(iwmmxt_unpack, glue(S, w)))(CPUARMState *env, \ uint64_t a, uint64_t b) \ -{ \ - a = \ - (((a >> SH0) & 0xffff) << 0) | \ - (((b >> SH0) & 0xffff) << 16) | \ - (((a >> SH2) & 0xffff) << 32) | \ - (((b >> SH2) & 0xffff) << 48); \ - env->iwmmxt.cregs[ARM_IWMMXT_wCASF] = \ - NZBIT8(a >> 0, 0) | NZBIT8(a >> 16, 1) | \ - NZBIT8(a >> 32, 2) | NZBIT8(a >> 48, 3); \ +{ \ + a = \ + (((a >> SH0) & 0xffff) << 0) | \ + (((b >> SH0) & 0xffff) << 16) | \ + (((a >> SH2) & 0xffff) << 32) | \ + (((b >> SH2) & 0xffff) << 48); \ + env->iwmmxt.cregs[ARM_IWMMXT_wCASF] = \ + NZBIT8(a >> 0, 0) | NZBIT8(a >> 16, 1) | \ + NZBIT8(a >> 32, 2) | NZBIT8(a >> 48, 3); \ return a; \ -} \ +} \ uint64_t HELPER(glue(iwmmxt_unpack, glue(S, l)))(CPUARMState *env, \ uint64_t a, uint64_t b) \ -{ \ - a = \ - (((a >> SH0) & 0xffffffff) << 0) | \ - (((b >> SH0) & 0xffffffff) << 32); \ - env->iwmmxt.cregs[ARM_IWMMXT_wCASF] = \ - NZBIT32(a >> 0, 0) | NZBIT32(a >> 32, 1); \ +{ \ + a = \ + (((a >> SH0) & 0xffffffff) << 0) | \ + (((b >> SH0) & 0xffffffff) << 32); \ + env->iwmmxt.cregs[ARM_IWMMXT_wCASF] = \ + NZBIT32(a >> 0, 0) | NZBIT32(a >> 32, 1); \ return a; \ -} \ +} \ uint64_t HELPER(glue(iwmmxt_unpack, glue(S, ub)))(CPUARMState *env, \ uint64_t x) \ -{ \ - x = \ - (((x >> SH0) & 0xff) << 0) | \ - (((x >> SH1) & 0xff) << 16) | \ - (((x >> SH2) & 0xff) << 32) | \ - (((x >> SH3) & 0xff) << 48); \ - env->iwmmxt.cregs[ARM_IWMMXT_wCASF] = \ - NZBIT16(x >> 0, 0) | NZBIT16(x >> 16, 1) | \ - NZBIT16(x >> 32, 2) | NZBIT16(x >> 48, 3); \ +{ \ + x = \ + (((x >> SH0) & 0xff) << 0) | \ + (((x >> SH1) & 0xff) << 16) | \ + (((x >> SH2) & 0xff) << 32) | \ + (((x >> SH3) & 0xff) << 48); \ + env->iwmmxt.cregs[ARM_IWMMXT_wCASF] = \ + NZBIT16(x >> 0, 0) | NZBIT16(x >> 16, 1) | \ + NZBIT16(x >> 32, 2) | NZBIT16(x >> 48, 3); \ return x; \ -} \ +} \ uint64_t HELPER(glue(iwmmxt_unpack, glue(S, uw)))(CPUARMState *env, \ uint64_t x) \ -{ \ - x = \ - (((x >> SH0) & 0xffff) << 0) | \ - (((x >> SH2) & 0xffff) << 32); \ - env->iwmmxt.cregs[ARM_IWMMXT_wCASF] = \ - NZBIT32(x >> 0, 0) | NZBIT32(x >> 32, 1); \ +{ \ + x = \ + (((x >> SH0) & 0xffff) << 0) | \ + (((x >> SH2) & 0xffff) << 32); \ + env->iwmmxt.cregs[ARM_IWMMXT_wCASF] = \ + NZBIT32(x >> 0, 0) | NZBIT32(x >> 32, 1); \ return x; \ -} \ +} \ uint64_t HELPER(glue(iwmmxt_unpack, glue(S, ul)))(CPUARMState *env, \ uint64_t x) \ -{ \ - x = (((x >> SH0) & 0xffffffff) << 0); \ - env->iwmmxt.cregs[ARM_IWMMXT_wCASF] = NZBIT64(x >> 0); \ +{ \ + x = (((x >> SH0) & 0xffffffff) << 0); \ + env->iwmmxt.cregs[ARM_IWMMXT_wCASF] = NZBIT64(x >> 0); \ return x; \ -} \ +} \ uint64_t HELPER(glue(iwmmxt_unpack, glue(S, sb)))(CPUARMState *env, \ uint64_t x) \ -{ \ - x = \ - ((uint64_t) EXTEND8H((x >> SH0) & 0xff) << 0) | \ - ((uint64_t) EXTEND8H((x >> SH1) & 0xff) << 16) | \ - ((uint64_t) EXTEND8H((x >> SH2) & 0xff) << 32) | \ - ((uint64_t) EXTEND8H((x >> SH3) & 0xff) << 48); \ - env->iwmmxt.cregs[ARM_IWMMXT_wCASF] = \ - NZBIT16(x >> 0, 0) | NZBIT16(x >> 16, 1) | \ - NZBIT16(x >> 32, 2) | NZBIT16(x >> 48, 3); \ +{ \ + x = \ + ((uint64_t) EXTEND8H((x >> SH0) & 0xff) << 0) | \ + ((uint64_t) EXTEND8H((x >> SH1) & 0xff) << 16) | \ + ((uint64_t) EXTEND8H((x >> SH2) & 0xff) << 32) | \ + ((uint64_t) EXTEND8H((x >> SH3) & 0xff) << 48); \ + env->iwmmxt.cregs[ARM_IWMMXT_wCASF] = \ + NZBIT16(x >> 0, 0) | NZBIT16(x >> 16, 1) | \ + NZBIT16(x >> 32, 2) | NZBIT16(x >> 48, 3); \ return x; \ -} \ +} \ uint64_t HELPER(glue(iwmmxt_unpack, glue(S, sw)))(CPUARMState *env, \ uint64_t x) \ -{ \ - x = \ - ((uint64_t) EXTEND16((x >> SH0) & 0xffff) << 0) | \ - ((uint64_t) EXTEND16((x >> SH2) & 0xffff) << 32); \ - env->iwmmxt.cregs[ARM_IWMMXT_wCASF] = \ - NZBIT32(x >> 0, 0) | NZBIT32(x >> 32, 1); \ +{ \ + x = \ + ((uint64_t) EXTEND16((x >> SH0) & 0xffff) << 0) | \ + ((uint64_t) EXTEND16((x >> SH2) & 0xffff) << 32); \ + env->iwmmxt.cregs[ARM_IWMMXT_wCASF] = \ + NZBIT32(x >> 0, 0) | NZBIT32(x >> 32, 1); \ return x; \ -} \ +} \ uint64_t HELPER(glue(iwmmxt_unpack, glue(S, sl)))(CPUARMState *env, \ uint64_t x) \ -{ \ - x = EXTEND32((x >> SH0) & 0xffffffff); \ - env->iwmmxt.cregs[ARM_IWMMXT_wCASF] = NZBIT64(x >> 0); \ +{ \ + x = EXTEND32((x >> SH0) & 0xffffffff); \ + env->iwmmxt.cregs[ARM_IWMMXT_wCASF] = NZBIT64(x >> 0); \ return x; \ } IWMMXT_OP_UNPACK(l, 0, 8, 16, 24) IWMMXT_OP_UNPACK(h, 32, 40, 48, 56) -#define IWMMXT_OP_CMP(SUFF, Tb, Tw, Tl, O) \ +#define IWMMXT_OP_CMP(SUFF, Tb, Tw, Tl, O) \ uint64_t HELPER(glue(iwmmxt_, glue(SUFF, b)))(CPUARMState *env, \ uint64_t a, uint64_t b) \ -{ \ - a = \ - CMP(0, Tb, O, 0xff) | CMP(8, Tb, O, 0xff) | \ - CMP(16, Tb, O, 0xff) | CMP(24, Tb, O, 0xff) | \ - CMP(32, Tb, O, 0xff) | CMP(40, Tb, O, 0xff) | \ - CMP(48, Tb, O, 0xff) | CMP(56, Tb, O, 0xff); \ - env->iwmmxt.cregs[ARM_IWMMXT_wCASF] = \ - NZBIT8(a >> 0, 0) | NZBIT8(a >> 8, 1) | \ - NZBIT8(a >> 16, 2) | NZBIT8(a >> 24, 3) | \ - NZBIT8(a >> 32, 4) | NZBIT8(a >> 40, 5) | \ - NZBIT8(a >> 48, 6) | NZBIT8(a >> 56, 7); \ +{ \ + a = \ + CMP(0, Tb, O, 0xff) | CMP(8, Tb, O, 0xff) | \ + CMP(16, Tb, O, 0xff) | CMP(24, Tb, O, 0xff) | \ + CMP(32, Tb, O, 0xff) | CMP(40, Tb, O, 0xff) | \ + CMP(48, Tb, O, 0xff) | CMP(56, Tb, O, 0xff); \ + env->iwmmxt.cregs[ARM_IWMMXT_wCASF] = \ + NZBIT8(a >> 0, 0) | NZBIT8(a >> 8, 1) | \ + NZBIT8(a >> 16, 2) | NZBIT8(a >> 24, 3) | \ + NZBIT8(a >> 32, 4) | NZBIT8(a >> 40, 5) | \ + NZBIT8(a >> 48, 6) | NZBIT8(a >> 56, 7); \ return a; \ -} \ +} \ uint64_t HELPER(glue(iwmmxt_, glue(SUFF, w)))(CPUARMState *env, \ uint64_t a, uint64_t b) \ -{ \ - a = CMP(0, Tw, O, 0xffff) | CMP(16, Tw, O, 0xffff) | \ - CMP(32, Tw, O, 0xffff) | CMP(48, Tw, O, 0xffff); \ - env->iwmmxt.cregs[ARM_IWMMXT_wCASF] = \ - NZBIT16(a >> 0, 0) | NZBIT16(a >> 16, 1) | \ - NZBIT16(a >> 32, 2) | NZBIT16(a >> 48, 3); \ +{ \ + a = CMP(0, Tw, O, 0xffff) | CMP(16, Tw, O, 0xffff) | \ + CMP(32, Tw, O, 0xffff) | CMP(48, Tw, O, 0xffff); \ + env->iwmmxt.cregs[ARM_IWMMXT_wCASF] = \ + NZBIT16(a >> 0, 0) | NZBIT16(a >> 16, 1) | \ + NZBIT16(a >> 32, 2) | NZBIT16(a >> 48, 3); \ return a; \ -} \ +} \ uint64_t HELPER(glue(iwmmxt_, glue(SUFF, l)))(CPUARMState *env, \ uint64_t a, uint64_t b) \ -{ \ - a = CMP(0, Tl, O, 0xffffffff) | \ - CMP(32, Tl, O, 0xffffffff); \ - env->iwmmxt.cregs[ARM_IWMMXT_wCASF] = \ - NZBIT32(a >> 0, 0) | NZBIT32(a >> 32, 1); \ +{ \ + a = CMP(0, Tl, O, 0xffffffff) | \ + CMP(32, Tl, O, 0xffffffff); \ + env->iwmmxt.cregs[ARM_IWMMXT_wCASF] = \ + NZBIT32(a >> 0, 0) | NZBIT32(a >> 32, 1); \ return a; \ } #define CMP(SHR, TYPE, OPER, MASK) ((((TYPE) ((a >> SHR) & MASK) OPER \ -- 2.18.0 Following the bulk conversion of the iwMMXt code, there are just a handful of hard coded tabs in target/arm; fix them. This is a whitespace-only patch. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Message-id: 20180821165215.29069-4-peter.maydell@linaro.org --- target/arm/cpu.h | 16 ++++++++-------- target/arm/arm-semi.c | 2 +- 2 files changed, 9 insertions(+), 9 deletions(-) diff --git a/target/arm/cpu.h b/target/arm/cpu.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/cpu.h +++ b/target/arm/cpu.h @@ -XXX,XX +XXX,XX @@ enum arm_cpu_mode { #define ARM_VFP_FPINST2 10 /* iwMMXt coprocessor control registers. */ -#define ARM_IWMMXT_wCID 0 -#define ARM_IWMMXT_wCon 1 -#define ARM_IWMMXT_wCSSF 2 -#define ARM_IWMMXT_wCASF 3 -#define ARM_IWMMXT_wCGR0 8 -#define ARM_IWMMXT_wCGR1 9 -#define ARM_IWMMXT_wCGR2 10 -#define ARM_IWMMXT_wCGR3 11 +#define ARM_IWMMXT_wCID 0 +#define ARM_IWMMXT_wCon 1 +#define ARM_IWMMXT_wCSSF 2 +#define ARM_IWMMXT_wCASF 3 +#define ARM_IWMMXT_wCGR0 8 +#define ARM_IWMMXT_wCGR1 9 +#define ARM_IWMMXT_wCGR2 10 +#define ARM_IWMMXT_wCGR3 11 /* V7M CCR bits */ FIELD(V7M_CCR, NONBASETHRDENA, 0, 1) diff --git a/target/arm/arm-semi.c b/target/arm/arm-semi.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/arm-semi.c +++ b/target/arm/arm-semi.c @@ -XXX,XX +XXX,XX @@ static void arm_semi_cb(CPUState *cs, target_ulong ret, target_ulong err) #ifdef CONFIG_USER_ONLY ts->swi_errno = err; #else - syscall_err = err; + syscall_err = err; #endif reg0 = ret; } else { -- 2.18.0 The handling of framebuffer properties in the bcm2835_property code is a bit clumsy, because for each of the many fb related properties we try to track the value we're about to set and whether we're going to be setting a value, and then we hand all the new values off to the framebuffer via a function which takes them all as separate arguments. It would be simpler if the property code could easily copy all the framebuffer's current settings, update them with the new specified values and then ask the framebuffer to switch to the new set. As the first part of this refactoring, pull all the fb config settings fields in BCM2835FBState out into their own struct. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20180814144436.679-2-peter.maydell@linaro.org --- include/hw/display/bcm2835_fb.h | 26 ++++++-- hw/display/bcm2835_fb.c | 114 +++++++++++++++++--------------- hw/misc/bcm2835_property.c | 28 ++++---- 3 files changed, 94 insertions(+), 74 deletions(-) diff --git a/include/hw/display/bcm2835_fb.h b/include/hw/display/bcm2835_fb.h index XXXXXXX..XXXXXXX 100644 --- a/include/hw/display/bcm2835_fb.h +++ b/include/hw/display/bcm2835_fb.h @@ -XXX,XX +XXX,XX @@ #define TYPE_BCM2835_FB "bcm2835-fb" #define BCM2835_FB(obj) OBJECT_CHECK(BCM2835FBState, (obj), TYPE_BCM2835_FB) +/* + * Configuration information about the fb which the guest can program + * via the mailbox property interface. + */ +typedef struct { + uint32_t xres, yres; + uint32_t xres_virtual, yres_virtual; + uint32_t xoffset, yoffset; + uint32_t bpp; + uint32_t base; + uint32_t pixo; + uint32_t alpha; +} BCM2835FBConfig; + typedef struct { /*< private >*/ SysBusDevice busdev; @@ -XXX,XX +XXX,XX @@ typedef struct { qemu_irq mbox_irq; bool lock, invalidate, pending; - uint32_t xres, yres; - uint32_t xres_virtual, yres_virtual; - uint32_t xoffset, yoffset; - uint32_t bpp; - uint32_t base, pitch, size; - uint32_t pixo, alpha; + + BCM2835FBConfig config; + + /* These are just cached values calculated from the config settings */ + uint32_t size; + uint32_t pitch; } BCM2835FBState; void bcm2835_fb_reconfigure(BCM2835FBState *s, uint32_t *xres, uint32_t *yres, diff --git a/hw/display/bcm2835_fb.c b/hw/display/bcm2835_fb.c index XXXXXXX..XXXXXXX 100644 --- a/hw/display/bcm2835_fb.c +++ b/hw/display/bcm2835_fb.c @@ -XXX,XX +XXX,XX @@ static void draw_line_src16(void *opaque, uint8_t *dst, const uint8_t *src, int bpp = surface_bits_per_pixel(surface); while (width--) { - switch (s->bpp) { + switch (s->config.bpp) { case 8: /* lookup palette starting at video ram base * TODO: cache translation, rather than doing this each time! @@ -XXX,XX +XXX,XX @@ static void draw_line_src16(void *opaque, uint8_t *dst, const uint8_t *src, break; } - if (s->pixo == 0) { + if (s->config.pixo == 0) { /* swap to BGR pixel format */ uint8_t tmp = r; r = b; @@ -XXX,XX +XXX,XX @@ static void fb_update_display(void *opaque) int src_width = 0; int dest_width = 0; - if (s->lock || !s->xres) { + if (s->lock || !s->config.xres) { return; } - src_width = s->xres * (s->bpp >> 3); - dest_width = s->xres; + src_width = s->config.xres * (s->config.bpp >> 3); + dest_width = s->config.xres; switch (surface_bits_per_pixel(surface)) { case 0: @@ -XXX,XX +XXX,XX @@ static void fb_update_display(void *opaque) } if (s->invalidate) { - framebuffer_update_memory_section(&s->fbsection, s->dma_mr, s->base, - s->yres, src_width); + framebuffer_update_memory_section(&s->fbsection, s->dma_mr, + s->config.base, + s->config.yres, src_width); } - framebuffer_update_display(surface, &s->fbsection, s->xres, s->yres, + framebuffer_update_display(surface, &s->fbsection, + s->config.xres, s->config.yres, src_width, dest_width, 0, s->invalidate, draw_line_src16, s, &first, &last); if (first >= 0) { - dpy_gfx_update(s->con, 0, first, s->xres, last - first + 1); + dpy_gfx_update(s->con, 0, first, s->config.xres, last - first + 1); } s->invalidate = false; @@ -XXX,XX +XXX,XX @@ static void bcm2835_fb_mbox_push(BCM2835FBState *s, uint32_t value) s->lock = true; - s->xres = ldl_le_phys(&s->dma_as, value); - s->yres = ldl_le_phys(&s->dma_as, value + 4); - s->xres_virtual = ldl_le_phys(&s->dma_as, value + 8); - s->yres_virtual = ldl_le_phys(&s->dma_as, value + 12); - s->bpp = ldl_le_phys(&s->dma_as, value + 20); - s->xoffset = ldl_le_phys(&s->dma_as, value + 24); - s->yoffset = ldl_le_phys(&s->dma_as, value + 28); + s->config.xres = ldl_le_phys(&s->dma_as, value); + s->config.yres = ldl_le_phys(&s->dma_as, value + 4); + s->config.xres_virtual = ldl_le_phys(&s->dma_as, value + 8); + s->config.yres_virtual = ldl_le_phys(&s->dma_as, value + 12); + s->config.bpp = ldl_le_phys(&s->dma_as, value + 20); + s->config.xoffset = ldl_le_phys(&s->dma_as, value + 24); + s->config.yoffset = ldl_le_phys(&s->dma_as, value + 28); - s->base = s->vcram_base | (value & 0xc0000000); - s->base += BCM2835_FB_OFFSET; + s->config.base = s->vcram_base | (value & 0xc0000000); + s->config.base += BCM2835_FB_OFFSET; /* TODO - Manage properly virtual resolution */ - s->pitch = s->xres * (s->bpp >> 3); - s->size = s->yres * s->pitch; + s->pitch = s->config.xres * (s->config.bpp >> 3); + s->size = s->config.yres * s->pitch; stl_le_phys(&s->dma_as, value + 16, s->pitch); - stl_le_phys(&s->dma_as, value + 32, s->base); + stl_le_phys(&s->dma_as, value + 32, s->config.base); stl_le_phys(&s->dma_as, value + 36, s->size); s->invalidate = true; - qemu_console_resize(s->con, s->xres, s->yres); + qemu_console_resize(s->con, s->config.xres, s->config.yres); s->lock = false; } @@ -XXX,XX +XXX,XX @@ void bcm2835_fb_reconfigure(BCM2835FBState *s, uint32_t *xres, uint32_t *yres, /* TODO: input validation! */ if (xres) { - s->xres = *xres; + s->config.xres = *xres; } if (yres) { - s->yres = *yres; + s->config.yres = *yres; } if (xoffset) { - s->xoffset = *xoffset; + s->config.xoffset = *xoffset; } if (yoffset) { - s->yoffset = *yoffset; + s->config.yoffset = *yoffset; } if (bpp) { - s->bpp = *bpp; + s->config.bpp = *bpp; } if (pixo) { - s->pixo = *pixo; + s->config.pixo = *pixo; } if (alpha) { - s->alpha = *alpha; + s->config.alpha = *alpha; } /* TODO - Manage properly virtual resolution */ - s->pitch = s->xres * (s->bpp >> 3); - s->size = s->yres * s->pitch; + s->pitch = s->config.xres * (s->config.bpp >> 3); + s->size = s->config.yres * s->pitch; s->invalidate = true; - qemu_console_resize(s->con, s->xres, s->yres); + qemu_console_resize(s->con, s->config.xres, s->config.yres); s->lock = false; } @@ -XXX,XX +XXX,XX @@ static const VMStateDescription vmstate_bcm2835_fb = { VMSTATE_BOOL(lock, BCM2835FBState), VMSTATE_BOOL(invalidate, BCM2835FBState), VMSTATE_BOOL(pending, BCM2835FBState), - VMSTATE_UINT32(xres, BCM2835FBState), - VMSTATE_UINT32(yres, BCM2835FBState), - VMSTATE_UINT32(xres_virtual, BCM2835FBState), - VMSTATE_UINT32(yres_virtual, BCM2835FBState), - VMSTATE_UINT32(xoffset, BCM2835FBState), - VMSTATE_UINT32(yoffset, BCM2835FBState), - VMSTATE_UINT32(bpp, BCM2835FBState), - VMSTATE_UINT32(base, BCM2835FBState), + VMSTATE_UINT32(config.xres, BCM2835FBState), + VMSTATE_UINT32(config.yres, BCM2835FBState), + VMSTATE_UINT32(config.xres_virtual, BCM2835FBState), + VMSTATE_UINT32(config.yres_virtual, BCM2835FBState), + VMSTATE_UINT32(config.xoffset, BCM2835FBState), + VMSTATE_UINT32(config.yoffset, BCM2835FBState), + VMSTATE_UINT32(config.bpp, BCM2835FBState), + VMSTATE_UINT32(config.base, BCM2835FBState), VMSTATE_UINT32(pitch, BCM2835FBState), VMSTATE_UINT32(size, BCM2835FBState), - VMSTATE_UINT32(pixo, BCM2835FBState), - VMSTATE_UINT32(alpha, BCM2835FBState), + VMSTATE_UINT32(config.pixo, BCM2835FBState), + VMSTATE_UINT32(config.alpha, BCM2835FBState), VMSTATE_END_OF_LIST() } }; @@ -XXX,XX +XXX,XX @@ static void bcm2835_fb_reset(DeviceState *dev) s->pending = false; - s->xres_virtual = s->xres; - s->yres_virtual = s->yres; - s->xoffset = 0; - s->yoffset = 0; - s->base = s->vcram_base + BCM2835_FB_OFFSET; - s->pitch = s->xres * (s->bpp >> 3); - s->size = s->yres * s->pitch; + s->config.xres_virtual = s->config.xres; + s->config.yres_virtual = s->config.yres; + s->config.xoffset = 0; + s->config.yoffset = 0; + s->config.base = s->vcram_base + BCM2835_FB_OFFSET; + s->pitch = s->config.xres * (s->config.bpp >> 3); + s->size = s->config.yres * s->pitch; s->invalidate = true; s->lock = false; @@ -XXX,XX +XXX,XX @@ static void bcm2835_fb_realize(DeviceState *dev, Error **errp) bcm2835_fb_reset(dev); s->con = graphic_console_init(dev, 0, &vgafb_ops, s); - qemu_console_resize(s->con, s->xres, s->yres); + qemu_console_resize(s->con, s->config.xres, s->config.yres); } static Property bcm2835_fb_props[] = { DEFINE_PROP_UINT32("vcram-base", BCM2835FBState, vcram_base, 0),/*required*/ DEFINE_PROP_UINT32("vcram-size", BCM2835FBState, vcram_size, DEFAULT_VCRAM_SIZE), - DEFINE_PROP_UINT32("xres", BCM2835FBState, xres, 640), - DEFINE_PROP_UINT32("yres", BCM2835FBState, yres, 480), - DEFINE_PROP_UINT32("bpp", BCM2835FBState, bpp, 16), - DEFINE_PROP_UINT32("pixo", BCM2835FBState, pixo, 1), /* 1=RGB, 0=BGR */ - DEFINE_PROP_UINT32("alpha", BCM2835FBState, alpha, 2), /* alpha ignored */ + DEFINE_PROP_UINT32("xres", BCM2835FBState, config.xres, 640), + DEFINE_PROP_UINT32("yres", BCM2835FBState, config.yres, 480), + DEFINE_PROP_UINT32("bpp", BCM2835FBState, config.bpp, 16), + DEFINE_PROP_UINT32("pixo", + BCM2835FBState, config.pixo, 1), /* 1=RGB, 0=BGR */ + DEFINE_PROP_UINT32("alpha", + BCM2835FBState, config.alpha, 2), /* alpha ignored */ DEFINE_PROP_END_OF_LIST() }; diff --git a/hw/misc/bcm2835_property.c b/hw/misc/bcm2835_property.c index XXXXXXX..XXXXXXX 100644 --- a/hw/misc/bcm2835_property.c +++ b/hw/misc/bcm2835_property.c @@ -XXX,XX +XXX,XX @@ static void bcm2835_property_mbox_push(BCM2835PropertyState *s, uint32_t value) /* Frame buffer */ case 0x00040001: /* Allocate buffer */ - stl_le_phys(&s->dma_as, value + 12, s->fbdev->base); - tmp_xres = newxres != NULL ? *newxres : s->fbdev->xres; - tmp_yres = newyres != NULL ? *newyres : s->fbdev->yres; - tmp_bpp = newbpp != NULL ? *newbpp : s->fbdev->bpp; + stl_le_phys(&s->dma_as, value + 12, s->fbdev->config.base); + tmp_xres = newxres != NULL ? *newxres : s->fbdev->config.xres; + tmp_yres = newyres != NULL ? *newyres : s->fbdev->config.yres; + tmp_bpp = newbpp != NULL ? *newbpp : s->fbdev->config.bpp; stl_le_phys(&s->dma_as, value + 16, tmp_xres * tmp_yres * tmp_bpp / 8); resplen = 8; @@ -XXX,XX +XXX,XX @@ static void bcm2835_property_mbox_push(BCM2835PropertyState *s, uint32_t value) break; case 0x00040003: /* Get display width/height */ case 0x00040004: - tmp_xres = newxres != NULL ? *newxres : s->fbdev->xres; - tmp_yres = newyres != NULL ? *newyres : s->fbdev->yres; + tmp_xres = newxres != NULL ? *newxres : s->fbdev->config.xres; + tmp_yres = newyres != NULL ? *newyres : s->fbdev->config.yres; stl_le_phys(&s->dma_as, value + 12, tmp_xres); stl_le_phys(&s->dma_as, value + 16, tmp_yres); resplen = 8; @@ -XXX,XX +XXX,XX @@ static void bcm2835_property_mbox_push(BCM2835PropertyState *s, uint32_t value) resplen = 8; break; case 0x00040005: /* Get depth */ - tmp_bpp = newbpp != NULL ? *newbpp : s->fbdev->bpp; + tmp_bpp = newbpp != NULL ? *newbpp : s->fbdev->config.bpp; stl_le_phys(&s->dma_as, value + 12, tmp_bpp); resplen = 4; break; @@ -XXX,XX +XXX,XX @@ static void bcm2835_property_mbox_push(BCM2835PropertyState *s, uint32_t value) resplen = 4; break; case 0x00040006: /* Get pixel order */ - tmp_pixo = newpixo != NULL ? *newpixo : s->fbdev->pixo; + tmp_pixo = newpixo != NULL ? *newpixo : s->fbdev->config.pixo; stl_le_phys(&s->dma_as, value + 12, tmp_pixo); resplen = 4; break; @@ -XXX,XX +XXX,XX @@ static void bcm2835_property_mbox_push(BCM2835PropertyState *s, uint32_t value) resplen = 4; break; case 0x00040007: /* Get alpha */ - tmp_alpha = newalpha != NULL ? *newalpha : s->fbdev->alpha; + tmp_alpha = newalpha != NULL ? *newalpha : s->fbdev->config.alpha; stl_le_phys(&s->dma_as, value + 12, tmp_alpha); resplen = 4; break; @@ -XXX,XX +XXX,XX @@ static void bcm2835_property_mbox_push(BCM2835PropertyState *s, uint32_t value) resplen = 4; break; case 0x00040008: /* Get pitch */ - tmp_xres = newxres != NULL ? *newxres : s->fbdev->xres; - tmp_bpp = newbpp != NULL ? *newbpp : s->fbdev->bpp; + tmp_xres = newxres != NULL ? *newxres : s->fbdev->config.xres; + tmp_bpp = newbpp != NULL ? *newbpp : s->fbdev->config.bpp; stl_le_phys(&s->dma_as, value + 12, tmp_xres * tmp_bpp / 8); resplen = 4; break; case 0x00040009: /* Get virtual offset */ - tmp_xoffset = newxoffset != NULL ? *newxoffset : s->fbdev->xoffset; - tmp_yoffset = newyoffset != NULL ? *newyoffset : s->fbdev->yoffset; + tmp_xoffset = newxoffset != NULL ? + *newxoffset : s->fbdev->config.xoffset; + tmp_yoffset = newyoffset != NULL ? + *newyoffset : s->fbdev->config.yoffset; stl_le_phys(&s->dma_as, value + 12, tmp_xoffset); stl_le_phys(&s->dma_as, value + 16, tmp_yoffset); resplen = 8; -- 2.18.0 Refactor the fb property setting code so that rather than using a set of pointers to local variables to track whether a config value has been updated in the current mbox and if so what its new value is, we just copy all the current settings of the fb at the start, and then update that copy as we go along, before asking the fb to switch to it at the end. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20180814144436.679-3-peter.maydell@linaro.org --- include/hw/display/bcm2835_fb.h | 4 +- hw/display/bcm2835_fb.c | 27 ++--------- hw/misc/bcm2835_property.c | 80 ++++++++++++++------------------- 3 files changed, 37 insertions(+), 74 deletions(-) diff --git a/include/hw/display/bcm2835_fb.h b/include/hw/display/bcm2835_fb.h index XXXXXXX..XXXXXXX 100644 --- a/include/hw/display/bcm2835_fb.h +++ b/include/hw/display/bcm2835_fb.h @@ -XXX,XX +XXX,XX @@ typedef struct { uint32_t pitch; } BCM2835FBState; -void bcm2835_fb_reconfigure(BCM2835FBState *s, uint32_t *xres, uint32_t *yres, - uint32_t *xoffset, uint32_t *yoffset, uint32_t *bpp, - uint32_t *pixo, uint32_t *alpha); +void bcm2835_fb_reconfigure(BCM2835FBState *s, BCM2835FBConfig *newconfig); #endif diff --git a/hw/display/bcm2835_fb.c b/hw/display/bcm2835_fb.c index XXXXXXX..XXXXXXX 100644 --- a/hw/display/bcm2835_fb.c +++ b/hw/display/bcm2835_fb.c @@ -XXX,XX +XXX,XX @@ static void bcm2835_fb_mbox_push(BCM2835FBState *s, uint32_t value) s->lock = false; } -void bcm2835_fb_reconfigure(BCM2835FBState *s, uint32_t *xres, uint32_t *yres, - uint32_t *xoffset, uint32_t *yoffset, uint32_t *bpp, - uint32_t *pixo, uint32_t *alpha) +void bcm2835_fb_reconfigure(BCM2835FBState *s, BCM2835FBConfig *newconfig) { s->lock = true; /* TODO: input validation! */ - if (xres) { - s->config.xres = *xres; - } - if (yres) { - s->config.yres = *yres; - } - if (xoffset) { - s->config.xoffset = *xoffset; - } - if (yoffset) { - s->config.yoffset = *yoffset; - } - if (bpp) { - s->config.bpp = *bpp; - } - if (pixo) { - s->config.pixo = *pixo; - } - if (alpha) { - s->config.alpha = *alpha; - } + + s->config = *newconfig; /* TODO - Manage properly virtual resolution */ diff --git a/hw/misc/bcm2835_property.c b/hw/misc/bcm2835_property.c index XXXXXXX..XXXXXXX 100644 --- a/hw/misc/bcm2835_property.c +++ b/hw/misc/bcm2835_property.c @@ -XXX,XX +XXX,XX @@ static void bcm2835_property_mbox_push(BCM2835PropertyState *s, uint32_t value) uint32_t tmp; int n; uint32_t offset, length, color; - uint32_t xres, yres, xoffset, yoffset, bpp, pixo, alpha; - uint32_t tmp_xres, tmp_yres, tmp_xoffset, tmp_yoffset; - uint32_t tmp_bpp, tmp_pixo, tmp_alpha; - uint32_t *newxres = NULL, *newyres = NULL, *newxoffset = NULL, - *newyoffset = NULL, *newbpp = NULL, *newpixo = NULL, *newalpha = NULL; + + /* + * Copy the current state of the framebuffer config; we will update + * this copy as we process tags and then ask the framebuffer to use + * it at the end. + */ + BCM2835FBConfig fbconfig = s->fbdev->config; + bool fbconfig_updated = false; value &= ~0xf; @@ -XXX,XX +XXX,XX @@ static void bcm2835_property_mbox_push(BCM2835PropertyState *s, uint32_t value) /* Frame buffer */ case 0x00040001: /* Allocate buffer */ - stl_le_phys(&s->dma_as, value + 12, s->fbdev->config.base); - tmp_xres = newxres != NULL ? *newxres : s->fbdev->config.xres; - tmp_yres = newyres != NULL ? *newyres : s->fbdev->config.yres; - tmp_bpp = newbpp != NULL ? *newbpp : s->fbdev->config.bpp; + stl_le_phys(&s->dma_as, value + 12, fbconfig.base); stl_le_phys(&s->dma_as, value + 16, - tmp_xres * tmp_yres * tmp_bpp / 8); + fbconfig.xres * fbconfig.yres * fbconfig.bpp / 8); resplen = 8; break; case 0x00048001: /* Release buffer */ @@ -XXX,XX +XXX,XX @@ static void bcm2835_property_mbox_push(BCM2835PropertyState *s, uint32_t value) break; case 0x00040003: /* Get display width/height */ case 0x00040004: - tmp_xres = newxres != NULL ? *newxres : s->fbdev->config.xres; - tmp_yres = newyres != NULL ? *newyres : s->fbdev->config.yres; - stl_le_phys(&s->dma_as, value + 12, tmp_xres); - stl_le_phys(&s->dma_as, value + 16, tmp_yres); + stl_le_phys(&s->dma_as, value + 12, fbconfig.xres); + stl_le_phys(&s->dma_as, value + 16, fbconfig.yres); resplen = 8; break; case 0x00044003: /* Test display width/height */ @@ -XXX,XX +XXX,XX @@ static void bcm2835_property_mbox_push(BCM2835PropertyState *s, uint32_t value) break; case 0x00048003: /* Set display width/height */ case 0x00048004: - xres = ldl_le_phys(&s->dma_as, value + 12); - newxres = &xres; - yres = ldl_le_phys(&s->dma_as, value + 16); - newyres = &yres; + fbconfig.xres = ldl_le_phys(&s->dma_as, value + 12); + fbconfig.yres = ldl_le_phys(&s->dma_as, value + 16); + fbconfig_updated = true; resplen = 8; break; case 0x00040005: /* Get depth */ - tmp_bpp = newbpp != NULL ? *newbpp : s->fbdev->config.bpp; - stl_le_phys(&s->dma_as, value + 12, tmp_bpp); + stl_le_phys(&s->dma_as, value + 12, fbconfig.bpp); resplen = 4; break; case 0x00044005: /* Test depth */ resplen = 4; break; case 0x00048005: /* Set depth */ - bpp = ldl_le_phys(&s->dma_as, value + 12); - newbpp = &bpp; + fbconfig.bpp = ldl_le_phys(&s->dma_as, value + 12); + fbconfig_updated = true; resplen = 4; break; case 0x00040006: /* Get pixel order */ - tmp_pixo = newpixo != NULL ? *newpixo : s->fbdev->config.pixo; - stl_le_phys(&s->dma_as, value + 12, tmp_pixo); + stl_le_phys(&s->dma_as, value + 12, fbconfig.pixo); resplen = 4; break; case 0x00044006: /* Test pixel order */ resplen = 4; break; case 0x00048006: /* Set pixel order */ - pixo = ldl_le_phys(&s->dma_as, value + 12); - newpixo = &pixo; + fbconfig.pixo = ldl_le_phys(&s->dma_as, value + 12); + fbconfig_updated = true; resplen = 4; break; case 0x00040007: /* Get alpha */ - tmp_alpha = newalpha != NULL ? *newalpha : s->fbdev->config.alpha; - stl_le_phys(&s->dma_as, value + 12, tmp_alpha); + stl_le_phys(&s->dma_as, value + 12, fbconfig.alpha); resplen = 4; break; case 0x00044007: /* Test pixel alpha */ resplen = 4; break; case 0x00048007: /* Set alpha */ - alpha = ldl_le_phys(&s->dma_as, value + 12); - newalpha = &alpha; + fbconfig.alpha = ldl_le_phys(&s->dma_as, value + 12); + fbconfig_updated = true; resplen = 4; break; case 0x00040008: /* Get pitch */ - tmp_xres = newxres != NULL ? *newxres : s->fbdev->config.xres; - tmp_bpp = newbpp != NULL ? *newbpp : s->fbdev->config.bpp; - stl_le_phys(&s->dma_as, value + 12, tmp_xres * tmp_bpp / 8); + stl_le_phys(&s->dma_as, value + 12, + fbconfig.xres * fbconfig.bpp / 8); resplen = 4; break; case 0x00040009: /* Get virtual offset */ - tmp_xoffset = newxoffset != NULL ? - *newxoffset : s->fbdev->config.xoffset; - tmp_yoffset = newyoffset != NULL ? - *newyoffset : s->fbdev->config.yoffset; - stl_le_phys(&s->dma_as, value + 12, tmp_xoffset); - stl_le_phys(&s->dma_as, value + 16, tmp_yoffset); + stl_le_phys(&s->dma_as, value + 12, fbconfig.xoffset); + stl_le_phys(&s->dma_as, value + 16, fbconfig.yoffset); resplen = 8; break; case 0x00044009: /* Test virtual offset */ resplen = 8; break; case 0x00048009: /* Set virtual offset */ - xoffset = ldl_le_phys(&s->dma_as, value + 12); - newxoffset = &xoffset; - yoffset = ldl_le_phys(&s->dma_as, value + 16); - newyoffset = &yoffset; + fbconfig.xoffset = ldl_le_phys(&s->dma_as, value + 12); + fbconfig.yoffset = ldl_le_phys(&s->dma_as, value + 16); + fbconfig_updated = true; resplen = 8; break; case 0x0004000a: /* Get/Test/Set overscan */ @@ -XXX,XX +XXX,XX @@ static void bcm2835_property_mbox_push(BCM2835PropertyState *s, uint32_t value) } /* Reconfigure framebuffer if required */ - if (newxres || newyres || newxoffset || newyoffset || newbpp || newpixo - || newalpha) { - bcm2835_fb_reconfigure(s->fbdev, newxres, newyres, newxoffset, - newyoffset, newbpp, newpixo, newalpha); + if (fbconfig_updated) { + bcm2835_fb_reconfigure(s->fbdev, &fbconfig); } /* Buffer response code */ -- 2.18.0 The BCM2835FBState struct has a 'pitch' field which is a cached copy of xres * (bpp >> 3), and a 'size' field which is a cached copy of pitch * yres. However we don't actually do anything with these fields; delete them. We retain the now-unused slots in the VMState struct for migration compatibility. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20180814144436.679-4-peter.maydell@linaro.org --- include/hw/display/bcm2835_fb.h | 4 ---- hw/display/bcm2835_fb.c | 19 ++++++++----------- 2 files changed, 8 insertions(+), 15 deletions(-) diff --git a/include/hw/display/bcm2835_fb.h b/include/hw/display/bcm2835_fb.h index XXXXXXX..XXXXXXX 100644 --- a/include/hw/display/bcm2835_fb.h +++ b/include/hw/display/bcm2835_fb.h @@ -XXX,XX +XXX,XX @@ typedef struct { bool lock, invalidate, pending; BCM2835FBConfig config; - - /* These are just cached values calculated from the config settings */ - uint32_t size; - uint32_t pitch; } BCM2835FBState; void bcm2835_fb_reconfigure(BCM2835FBState *s, BCM2835FBConfig *newconfig); diff --git a/hw/display/bcm2835_fb.c b/hw/display/bcm2835_fb.c index XXXXXXX..XXXXXXX 100644 --- a/hw/display/bcm2835_fb.c +++ b/hw/display/bcm2835_fb.c @@ -XXX,XX +XXX,XX @@ static void fb_update_display(void *opaque) static void bcm2835_fb_mbox_push(BCM2835FBState *s, uint32_t value) { + uint32_t pitch; + uint32_t size; + value &= ~0xf; s->lock = true; @@ -XXX,XX +XXX,XX @@ static void bcm2835_fb_mbox_push(BCM2835FBState *s, uint32_t value) /* TODO - Manage properly virtual resolution */ - s->pitch = s->config.xres * (s->config.bpp >> 3); - s->size = s->config.yres * s->pitch; + pitch = s->config.xres * (s->config.bpp >> 3); + size = s->config.yres * pitch; - stl_le_phys(&s->dma_as, value + 16, s->pitch); + stl_le_phys(&s->dma_as, value + 16, pitch); stl_le_phys(&s->dma_as, value + 32, s->config.base); - stl_le_phys(&s->dma_as, value + 36, s->size); + stl_le_phys(&s->dma_as, value + 36, size); s->invalidate = true; qemu_console_resize(s->con, s->config.xres, s->config.yres); @@ -XXX,XX +XXX,XX @@ void bcm2835_fb_reconfigure(BCM2835FBState *s, BCM2835FBConfig *newconfig) /* TODO - Manage properly virtual resolution */ - s->pitch = s->config.xres * (s->config.bpp >> 3); - s->size = s->config.yres * s->pitch; - s->invalidate = true; qemu_console_resize(s->con, s->config.xres, s->config.yres); s->lock = false; @@ -XXX,XX +XXX,XX @@ static const VMStateDescription vmstate_bcm2835_fb = { VMSTATE_UINT32(config.yoffset, BCM2835FBState), VMSTATE_UINT32(config.bpp, BCM2835FBState), VMSTATE_UINT32(config.base, BCM2835FBState), - VMSTATE_UINT32(pitch, BCM2835FBState), - VMSTATE_UINT32(size, BCM2835FBState), + VMSTATE_UNUSED(8), /* Was pitch and size */ VMSTATE_UINT32(config.pixo, BCM2835FBState), VMSTATE_UINT32(config.alpha, BCM2835FBState), VMSTATE_END_OF_LIST() @@ -XXX,XX +XXX,XX @@ static void bcm2835_fb_reset(DeviceState *dev) s->config.xoffset = 0; s->config.yoffset = 0; s->config.base = s->vcram_base + BCM2835_FB_OFFSET; - s->pitch = s->config.xres * (s->config.bpp >> 3); - s->size = s->config.yres * s->pitch; s->invalidate = true; s->lock = false; -- 2.18.0 The bcm2835_fb's initial resolution and other parameters are set via QOM properties. We should reset to those initial values on device reset, which means we need to save the QOM property values somewhere that they are not overwritten by guest changes to the framebuffer configuration. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20180814144436.679-5-peter.maydell@linaro.org --- include/hw/display/bcm2835_fb.h | 1 + hw/display/bcm2835_fb.c | 27 +++++++++++++++------------ 2 files changed, 16 insertions(+), 12 deletions(-) diff --git a/include/hw/display/bcm2835_fb.h b/include/hw/display/bcm2835_fb.h index XXXXXXX..XXXXXXX 100644 --- a/include/hw/display/bcm2835_fb.h +++ b/include/hw/display/bcm2835_fb.h @@ -XXX,XX +XXX,XX @@ typedef struct { bool lock, invalidate, pending; BCM2835FBConfig config; + BCM2835FBConfig initial_config; } BCM2835FBState; void bcm2835_fb_reconfigure(BCM2835FBState *s, BCM2835FBConfig *newconfig); diff --git a/hw/display/bcm2835_fb.c b/hw/display/bcm2835_fb.c index XXXXXXX..XXXXXXX 100644 --- a/hw/display/bcm2835_fb.c +++ b/hw/display/bcm2835_fb.c @@ -XXX,XX +XXX,XX @@ static void bcm2835_fb_reset(DeviceState *dev) s->pending = false; - s->config.xres_virtual = s->config.xres; - s->config.yres_virtual = s->config.yres; - s->config.xoffset = 0; - s->config.yoffset = 0; - s->config.base = s->vcram_base + BCM2835_FB_OFFSET; + s->config = s->initial_config; s->invalidate = true; s->lock = false; @@ -XXX,XX +XXX,XX @@ static void bcm2835_fb_realize(DeviceState *dev, Error **errp) return; } + /* Fill in the parts of initial_config that are not set by QOM properties */ + s->initial_config.xres_virtual = s->initial_config.xres; + s->initial_config.yres_virtual = s->initial_config.yres; + s->initial_config.xoffset = 0; + s->initial_config.yoffset = 0; + s->initial_config.base = s->vcram_base + BCM2835_FB_OFFSET; + s->dma_mr = MEMORY_REGION(obj); address_space_init(&s->dma_as, s->dma_mr, NULL); @@ -XXX,XX +XXX,XX @@ static Property bcm2835_fb_props[] = { DEFINE_PROP_UINT32("vcram-base", BCM2835FBState, vcram_base, 0),/*required*/ DEFINE_PROP_UINT32("vcram-size", BCM2835FBState, vcram_size, DEFAULT_VCRAM_SIZE), - DEFINE_PROP_UINT32("xres", BCM2835FBState, config.xres, 640), - DEFINE_PROP_UINT32("yres", BCM2835FBState, config.yres, 480), - DEFINE_PROP_UINT32("bpp", BCM2835FBState, config.bpp, 16), - DEFINE_PROP_UINT32("pixo", - BCM2835FBState, config.pixo, 1), /* 1=RGB, 0=BGR */ - DEFINE_PROP_UINT32("alpha", - BCM2835FBState, config.alpha, 2), /* alpha ignored */ + DEFINE_PROP_UINT32("xres", BCM2835FBState, initial_config.xres, 640), + DEFINE_PROP_UINT32("yres", BCM2835FBState, initial_config.yres, 480), + DEFINE_PROP_UINT32("bpp", BCM2835FBState, initial_config.bpp, 16), + DEFINE_PROP_UINT32("pixo", BCM2835FBState, + initial_config.pixo, 1), /* 1=RGB, 0=BGR */ + DEFINE_PROP_UINT32("alpha", BCM2835FBState, + initial_config.alpha, 2), /* alpha ignored */ DEFINE_PROP_END_OF_LIST() }; -- 2.18.0 Abstract out the calculation of the pitch and size of the framebuffer into functions that operate on the BCM2835FBConfig struct -- these are about to get a little more complicated when we add support for virtual and physical sizes differing. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20180814144436.679-6-peter.maydell@linaro.org --- include/hw/display/bcm2835_fb.h | 22 ++++++++++++++++++++++ hw/display/bcm2835_fb.c | 6 +++--- hw/misc/bcm2835_property.c | 4 ++-- 3 files changed, 27 insertions(+), 5 deletions(-) diff --git a/include/hw/display/bcm2835_fb.h b/include/hw/display/bcm2835_fb.h index XXXXXXX..XXXXXXX 100644 --- a/include/hw/display/bcm2835_fb.h +++ b/include/hw/display/bcm2835_fb.h @@ -XXX,XX +XXX,XX @@ typedef struct { void bcm2835_fb_reconfigure(BCM2835FBState *s, BCM2835FBConfig *newconfig); +/** + * bcm2835_fb_get_pitch: return number of bytes per line of the framebuffer + * @config: configuration info for the framebuffer + * + * Return the number of bytes per line of the framebuffer, ie the number + * that must be added to a pixel address to get the address of the pixel + * directly below it on screen. + */ +static inline uint32_t bcm2835_fb_get_pitch(BCM2835FBConfig *config) +{ + return config->xres * (config->bpp >> 3); +} + +/** + * bcm2835_fb_get_size: return total size of framebuffer in bytes + * @config: configuration info for the framebuffer + */ +static inline uint32_t bcm2835_fb_get_size(BCM2835FBConfig *config) +{ + return config->yres * bcm2835_fb_get_pitch(config); +} + #endif diff --git a/hw/display/bcm2835_fb.c b/hw/display/bcm2835_fb.c index XXXXXXX..XXXXXXX 100644 --- a/hw/display/bcm2835_fb.c +++ b/hw/display/bcm2835_fb.c @@ -XXX,XX +XXX,XX @@ static void fb_update_display(void *opaque) return; } - src_width = s->config.xres * (s->config.bpp >> 3); + src_width = bcm2835_fb_get_pitch(&s->config); dest_width = s->config.xres; switch (surface_bits_per_pixel(surface)) { @@ -XXX,XX +XXX,XX @@ static void bcm2835_fb_mbox_push(BCM2835FBState *s, uint32_t value) /* TODO - Manage properly virtual resolution */ - pitch = s->config.xres * (s->config.bpp >> 3); - size = s->config.yres * pitch; + pitch = bcm2835_fb_get_pitch(&s->config); + size = bcm2835_fb_get_size(&s->config); stl_le_phys(&s->dma_as, value + 16, pitch); stl_le_phys(&s->dma_as, value + 32, s->config.base); diff --git a/hw/misc/bcm2835_property.c b/hw/misc/bcm2835_property.c index XXXXXXX..XXXXXXX 100644 --- a/hw/misc/bcm2835_property.c +++ b/hw/misc/bcm2835_property.c @@ -XXX,XX +XXX,XX @@ static void bcm2835_property_mbox_push(BCM2835PropertyState *s, uint32_t value) case 0x00040001: /* Allocate buffer */ stl_le_phys(&s->dma_as, value + 12, fbconfig.base); stl_le_phys(&s->dma_as, value + 16, - fbconfig.xres * fbconfig.yres * fbconfig.bpp / 8); + bcm2835_fb_get_size(&fbconfig)); resplen = 8; break; case 0x00048001: /* Release buffer */ @@ -XXX,XX +XXX,XX @@ static void bcm2835_property_mbox_push(BCM2835PropertyState *s, uint32_t value) break; case 0x00040008: /* Get pitch */ stl_le_phys(&s->dma_as, value + 12, - fbconfig.xres * fbconfig.bpp / 8); + bcm2835_fb_get_pitch(&fbconfig)); resplen = 4; break; case 0x00040009: /* Get virtual offset */ -- 2.18.0 The raspi framebuffir in bcm2835_fb supports the definition of a virtual "viewport", which is smaller than the full physical framebuffer size and at an adjustable offset within it. Only the viewport area is sent to the screen. This allows the guest to do things like double buffering, or scrolling by adjusting the viewport origin. Currently QEMU doesn't implement this at all. Add support for this feature: * the property mailbox code needs to distinguish the virtual width/height from the physical width/height * the framebuffer code needs to do something with the virtual width/height/origin information Note that the wiki documentation on the semantics of the virtual and physical height and width has it the wrong way around -- the virtual size is the size of the allocated buffer, and the physical size is the size of the display, so the virtual size is always the same as or larger than the physical. If the viewport size is set smaller than the physical screen size, we ignore the viewport settings completely and just display the physical screen area. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20180814144436.679-7-peter.maydell@linaro.org --- include/hw/display/bcm2835_fb.h | 6 ++++-- hw/display/bcm2835_fb.c | 28 ++++++++++++++++++++++------ hw/misc/bcm2835_property.c | 21 +++++++++++++++------ 3 files changed, 41 insertions(+), 14 deletions(-) diff --git a/include/hw/display/bcm2835_fb.h b/include/hw/display/bcm2835_fb.h index XXXXXXX..XXXXXXX 100644 --- a/include/hw/display/bcm2835_fb.h +++ b/include/hw/display/bcm2835_fb.h @@ -XXX,XX +XXX,XX @@ void bcm2835_fb_reconfigure(BCM2835FBState *s, BCM2835FBConfig *newconfig); */ static inline uint32_t bcm2835_fb_get_pitch(BCM2835FBConfig *config) { - return config->xres * (config->bpp >> 3); + uint32_t xres = MAX(config->xres, config->xres_virtual); + return xres * (config->bpp >> 3); } /** @@ -XXX,XX +XXX,XX @@ static inline uint32_t bcm2835_fb_get_pitch(BCM2835FBConfig *config) */ static inline uint32_t bcm2835_fb_get_size(BCM2835FBConfig *config) { - return config->yres * bcm2835_fb_get_pitch(config); + uint32_t yres = MAX(config->yres, config->yres_virtual); + return yres * bcm2835_fb_get_pitch(config); } #endif diff --git a/hw/display/bcm2835_fb.c b/hw/display/bcm2835_fb.c index XXXXXXX..XXXXXXX 100644 --- a/hw/display/bcm2835_fb.c +++ b/hw/display/bcm2835_fb.c @@ -XXX,XX +XXX,XX @@ static void draw_line_src16(void *opaque, uint8_t *dst, const uint8_t *src, } } +static bool fb_use_offsets(BCM2835FBConfig *config) +{ + /* + * Return true if we should use the viewport offsets. + * Experimentally, the hardware seems to do this only if the + * viewport size is larger than the physical screen. (It doesn't + * prevent the guest setting this silly viewport setting, though...) + */ + return config->xres_virtual > config->xres && + config->yres_virtual > config->yres; +} + static void fb_update_display(void *opaque) { BCM2835FBState *s = opaque; @@ -XXX,XX +XXX,XX @@ static void fb_update_display(void *opaque) int last = 0; int src_width = 0; int dest_width = 0; + uint32_t xoff = 0, yoff = 0; if (s->lock || !s->config.xres) { return; } src_width = bcm2835_fb_get_pitch(&s->config); + if (fb_use_offsets(&s->config)) { + xoff = s->config.xoffset; + yoff = s->config.yoffset; + } + dest_width = s->config.xres; switch (surface_bits_per_pixel(surface)) { @@ -XXX,XX +XXX,XX @@ static void fb_update_display(void *opaque) } if (s->invalidate) { + hwaddr base = s->config.base + xoff + yoff * src_width; framebuffer_update_memory_section(&s->fbsection, s->dma_mr, - s->config.base, + base, s->config.yres, src_width); } @@ -XXX,XX +XXX,XX @@ static void fb_update_display(void *opaque) draw_line_src16, s, &first, &last); if (first >= 0) { - dpy_gfx_update(s->con, 0, first, s->config.xres, last - first + 1); + dpy_gfx_update(s->con, 0, first, s->config.xres, + last - first + 1); } s->invalidate = false; @@ -XXX,XX +XXX,XX @@ static void bcm2835_fb_mbox_push(BCM2835FBState *s, uint32_t value) s->config.base = s->vcram_base | (value & 0xc0000000); s->config.base += BCM2835_FB_OFFSET; - /* TODO - Manage properly virtual resolution */ - pitch = bcm2835_fb_get_pitch(&s->config); size = bcm2835_fb_get_size(&s->config); @@ -XXX,XX +XXX,XX @@ void bcm2835_fb_reconfigure(BCM2835FBState *s, BCM2835FBConfig *newconfig) s->config = *newconfig; - /* TODO - Manage properly virtual resolution */ - s->invalidate = true; qemu_console_resize(s->con, s->config.xres, s->config.yres); s->lock = false; diff --git a/hw/misc/bcm2835_property.c b/hw/misc/bcm2835_property.c index XXXXXXX..XXXXXXX 100644 --- a/hw/misc/bcm2835_property.c +++ b/hw/misc/bcm2835_property.c @@ -XXX,XX +XXX,XX @@ static void bcm2835_property_mbox_push(BCM2835PropertyState *s, uint32_t value) case 0x00040002: /* Blank screen */ resplen = 4; break; - case 0x00040003: /* Get display width/height */ - case 0x00040004: + case 0x00040003: /* Get physical display width/height */ stl_le_phys(&s->dma_as, value + 12, fbconfig.xres); stl_le_phys(&s->dma_as, value + 16, fbconfig.yres); resplen = 8; break; - case 0x00044003: /* Test display width/height */ - case 0x00044004: + case 0x00040004: /* Get virtual display width/height */ + stl_le_phys(&s->dma_as, value + 12, fbconfig.xres_virtual); + stl_le_phys(&s->dma_as, value + 16, fbconfig.yres_virtual); resplen = 8; break; - case 0x00048003: /* Set display width/height */ - case 0x00048004: + case 0x00044003: /* Test physical display width/height */ + case 0x00044004: /* Test virtual display width/height */ + resplen = 8; + break; + case 0x00048003: /* Set physical display width/height */ fbconfig.xres = ldl_le_phys(&s->dma_as, value + 12); fbconfig.yres = ldl_le_phys(&s->dma_as, value + 16); fbconfig_updated = true; resplen = 8; break; + case 0x00048004: /* Set virtual display width/height */ + fbconfig.xres_virtual = ldl_le_phys(&s->dma_as, value + 12); + fbconfig.yres_virtual = ldl_le_phys(&s->dma_as, value + 16); + fbconfig_updated = true; + resplen = 8; + break; case 0x00040005: /* Get depth */ stl_le_phys(&s->dma_as, value + 12, fbconfig.bpp); resplen = 4; -- 2.18.0 Validate the config settings that the guest tries to set. The wiki page documentation is not really accurate here: generally rather than failing requests to set bad parameters, the hardware will just clip them to something sensible. Validate the most important parameters: sizes and the viewport offsets. This prevents the framebuffer code from trying to read out-of-range memory. In the property handling code, we validate the new parameters every time we encounter a tag that sets them. This means we validate the config multiple times if the request includes multiple config-setting tags, but the code would require significant restructuring to do a validation only once but still return the clipped settings for get-parameter tags and the buffer allocation tag. Validation of settings made via the older bcm2835_fb_mbox_push() function will be done in the next commit. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20180814144436.679-8-peter.maydell@linaro.org --- include/hw/display/bcm2835_fb.h | 8 +++++ hw/display/bcm2835_fb.c | 48 +++++++++++++++++++++++++++-- hw/misc/bcm2835_property.c | 54 ++++++++++++++++----------------- 3 files changed, 81 insertions(+), 29 deletions(-) diff --git a/include/hw/display/bcm2835_fb.h b/include/hw/display/bcm2835_fb.h index XXXXXXX..XXXXXXX 100644 --- a/include/hw/display/bcm2835_fb.h +++ b/include/hw/display/bcm2835_fb.h @@ -XXX,XX +XXX,XX @@ static inline uint32_t bcm2835_fb_get_size(BCM2835FBConfig *config) return yres * bcm2835_fb_get_pitch(config); } +/** + * bcm2835_fb_validate_config: check provided config + * + * Validates the configuration information provided by the guest and + * adjusts it if necessary. + */ +void bcm2835_fb_validate_config(BCM2835FBConfig *config); + #endif diff --git a/hw/display/bcm2835_fb.c b/hw/display/bcm2835_fb.c index XXXXXXX..XXXXXXX 100644 --- a/hw/display/bcm2835_fb.c +++ b/hw/display/bcm2835_fb.c @@ -XXX,XX +XXX,XX @@ #define DEFAULT_VCRAM_SIZE 0x4000000 #define BCM2835_FB_OFFSET 0x00100000 +/* Maximum permitted framebuffer size; experimentally determined on an rpi2 */ +#define XRES_MAX 3840 +#define YRES_MAX 2560 +/* Framebuffer size used if guest requests zero size */ +#define XRES_SMALL 592 +#define YRES_SMALL 488 + static void fb_invalidate_display(void *opaque) { BCM2835FBState *s = BCM2835_FB(opaque); @@ -XXX,XX +XXX,XX @@ static void fb_update_display(void *opaque) s->invalidate = false; } +void bcm2835_fb_validate_config(BCM2835FBConfig *config) +{ + /* + * Validate the config, and clip any bogus values into range, + * as the hardware does. Note that fb_update_display() relies on + * this happening to prevent it from performing out-of-range + * accesses on redraw. + */ + config->xres = MIN(config->xres, XRES_MAX); + config->xres_virtual = MIN(config->xres_virtual, XRES_MAX); + config->yres = MIN(config->yres, YRES_MAX); + config->yres_virtual = MIN(config->yres_virtual, YRES_MAX); + + /* + * These are not minima: a 40x40 framebuffer will be accepted. + * They're only used as defaults if the guest asks for zero size. + */ + if (config->xres == 0) { + config->xres = XRES_SMALL; + } + if (config->yres == 0) { + config->yres = YRES_SMALL; + } + if (config->xres_virtual == 0) { + config->xres_virtual = config->xres; + } + if (config->yres_virtual == 0) { + config->yres_virtual = config->yres; + } + + if (fb_use_offsets(config)) { + /* Clip the offsets so the viewport is within the physical screen */ + config->xoffset = MIN(config->xoffset, + config->xres_virtual - config->xres); + config->yoffset = MIN(config->yoffset, + config->yres_virtual - config->yres); + } +} + static void bcm2835_fb_mbox_push(BCM2835FBState *s, uint32_t value) { uint32_t pitch; @@ -XXX,XX +XXX,XX @@ void bcm2835_fb_reconfigure(BCM2835FBState *s, BCM2835FBConfig *newconfig) { s->lock = true; - /* TODO: input validation! */ - s->config = *newconfig; s->invalidate = true; diff --git a/hw/misc/bcm2835_property.c b/hw/misc/bcm2835_property.c index XXXXXXX..XXXXXXX 100644 --- a/hw/misc/bcm2835_property.c +++ b/hw/misc/bcm2835_property.c @@ -XXX,XX +XXX,XX @@ static void bcm2835_property_mbox_push(BCM2835PropertyState *s, uint32_t value) case 0x00040002: /* Blank screen */ resplen = 4; break; - case 0x00040003: /* Get physical display width/height */ - stl_le_phys(&s->dma_as, value + 12, fbconfig.xres); - stl_le_phys(&s->dma_as, value + 16, fbconfig.yres); - resplen = 8; - break; - case 0x00040004: /* Get virtual display width/height */ - stl_le_phys(&s->dma_as, value + 12, fbconfig.xres_virtual); - stl_le_phys(&s->dma_as, value + 16, fbconfig.yres_virtual); - resplen = 8; - break; case 0x00044003: /* Test physical display width/height */ case 0x00044004: /* Test virtual display width/height */ resplen = 8; @@ -XXX,XX +XXX,XX @@ static void bcm2835_property_mbox_push(BCM2835PropertyState *s, uint32_t value) case 0x00048003: /* Set physical display width/height */ fbconfig.xres = ldl_le_phys(&s->dma_as, value + 12); fbconfig.yres = ldl_le_phys(&s->dma_as, value + 16); + bcm2835_fb_validate_config(&fbconfig); fbconfig_updated = true; + /* fall through */ + case 0x00040003: /* Get physical display width/height */ + stl_le_phys(&s->dma_as, value + 12, fbconfig.xres); + stl_le_phys(&s->dma_as, value + 16, fbconfig.yres); resplen = 8; break; case 0x00048004: /* Set virtual display width/height */ fbconfig.xres_virtual = ldl_le_phys(&s->dma_as, value + 12); fbconfig.yres_virtual = ldl_le_phys(&s->dma_as, value + 16); + bcm2835_fb_validate_config(&fbconfig); fbconfig_updated = true; + /* fall through */ + case 0x00040004: /* Get virtual display width/height */ + stl_le_phys(&s->dma_as, value + 12, fbconfig.xres_virtual); + stl_le_phys(&s->dma_as, value + 16, fbconfig.yres_virtual); resplen = 8; break; - case 0x00040005: /* Get depth */ - stl_le_phys(&s->dma_as, value + 12, fbconfig.bpp); - resplen = 4; - break; case 0x00044005: /* Test depth */ resplen = 4; break; case 0x00048005: /* Set depth */ fbconfig.bpp = ldl_le_phys(&s->dma_as, value + 12); + bcm2835_fb_validate_config(&fbconfig); fbconfig_updated = true; - resplen = 4; - break; - case 0x00040006: /* Get pixel order */ - stl_le_phys(&s->dma_as, value + 12, fbconfig.pixo); + /* fall through */ + case 0x00040005: /* Get depth */ + stl_le_phys(&s->dma_as, value + 12, fbconfig.bpp); resplen = 4; break; case 0x00044006: /* Test pixel order */ @@ -XXX,XX +XXX,XX @@ static void bcm2835_property_mbox_push(BCM2835PropertyState *s, uint32_t value) break; case 0x00048006: /* Set pixel order */ fbconfig.pixo = ldl_le_phys(&s->dma_as, value + 12); + bcm2835_fb_validate_config(&fbconfig); fbconfig_updated = true; - resplen = 4; - break; - case 0x00040007: /* Get alpha */ - stl_le_phys(&s->dma_as, value + 12, fbconfig.alpha); + /* fall through */ + case 0x00040006: /* Get pixel order */ + stl_le_phys(&s->dma_as, value + 12, fbconfig.pixo); resplen = 4; break; case 0x00044007: /* Test pixel alpha */ @@ -XXX,XX +XXX,XX @@ static void bcm2835_property_mbox_push(BCM2835PropertyState *s, uint32_t value) break; case 0x00048007: /* Set alpha */ fbconfig.alpha = ldl_le_phys(&s->dma_as, value + 12); + bcm2835_fb_validate_config(&fbconfig); fbconfig_updated = true; + /* fall through */ + case 0x00040007: /* Get alpha */ + stl_le_phys(&s->dma_as, value + 12, fbconfig.alpha); resplen = 4; break; case 0x00040008: /* Get pitch */ @@ -XXX,XX +XXX,XX @@ static void bcm2835_property_mbox_push(BCM2835PropertyState *s, uint32_t value) bcm2835_fb_get_pitch(&fbconfig)); resplen = 4; break; - case 0x00040009: /* Get virtual offset */ - stl_le_phys(&s->dma_as, value + 12, fbconfig.xoffset); - stl_le_phys(&s->dma_as, value + 16, fbconfig.yoffset); - resplen = 8; - break; case 0x00044009: /* Test virtual offset */ resplen = 8; break; case 0x00048009: /* Set virtual offset */ fbconfig.xoffset = ldl_le_phys(&s->dma_as, value + 12); fbconfig.yoffset = ldl_le_phys(&s->dma_as, value + 16); + bcm2835_fb_validate_config(&fbconfig); fbconfig_updated = true; + /* fall through */ + case 0x00040009: /* Get virtual offset */ + stl_le_phys(&s->dma_as, value + 12, fbconfig.xoffset); + stl_le_phys(&s->dma_as, value + 16, fbconfig.yoffset); resplen = 8; break; case 0x0004000a: /* Get/Test/Set overscan */ -- 2.18.0 Refactor bcm2835_fb_mbox_push() to work by calling bcm2835_fb_validate_config() and bcm2835_fb_reconfigure(), so that config set this way is also validated. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20180814144436.679-9-peter.maydell@linaro.org --- hw/display/bcm2835_fb.c | 63 ++++++++++++++++++++--------------------- 1 file changed, 31 insertions(+), 32 deletions(-) diff --git a/hw/display/bcm2835_fb.c b/hw/display/bcm2835_fb.c index XXXXXXX..XXXXXXX 100644 --- a/hw/display/bcm2835_fb.c +++ b/hw/display/bcm2835_fb.c @@ -XXX,XX +XXX,XX @@ void bcm2835_fb_validate_config(BCM2835FBConfig *config) } } -static void bcm2835_fb_mbox_push(BCM2835FBState *s, uint32_t value) -{ - uint32_t pitch; - uint32_t size; - - value &= ~0xf; - - s->lock = true; - - s->config.xres = ldl_le_phys(&s->dma_as, value); - s->config.yres = ldl_le_phys(&s->dma_as, value + 4); - s->config.xres_virtual = ldl_le_phys(&s->dma_as, value + 8); - s->config.yres_virtual = ldl_le_phys(&s->dma_as, value + 12); - s->config.bpp = ldl_le_phys(&s->dma_as, value + 20); - s->config.xoffset = ldl_le_phys(&s->dma_as, value + 24); - s->config.yoffset = ldl_le_phys(&s->dma_as, value + 28); - - s->config.base = s->vcram_base | (value & 0xc0000000); - s->config.base += BCM2835_FB_OFFSET; - - pitch = bcm2835_fb_get_pitch(&s->config); - size = bcm2835_fb_get_size(&s->config); - - stl_le_phys(&s->dma_as, value + 16, pitch); - stl_le_phys(&s->dma_as, value + 32, s->config.base); - stl_le_phys(&s->dma_as, value + 36, size); - - s->invalidate = true; - qemu_console_resize(s->con, s->config.xres, s->config.yres); - s->lock = false; -} - void bcm2835_fb_reconfigure(BCM2835FBState *s, BCM2835FBConfig *newconfig) { s->lock = true; @@ -XXX,XX +XXX,XX @@ void bcm2835_fb_reconfigure(BCM2835FBState *s, BCM2835FBConfig *newconfig) s->lock = false; } +static void bcm2835_fb_mbox_push(BCM2835FBState *s, uint32_t value) +{ + uint32_t pitch; + uint32_t size; + BCM2835FBConfig newconf; + + value &= ~0xf; + + newconf.xres = ldl_le_phys(&s->dma_as, value); + newconf.yres = ldl_le_phys(&s->dma_as, value + 4); + newconf.xres_virtual = ldl_le_phys(&s->dma_as, value + 8); + newconf.yres_virtual = ldl_le_phys(&s->dma_as, value + 12); + newconf.bpp = ldl_le_phys(&s->dma_as, value + 20); + newconf.xoffset = ldl_le_phys(&s->dma_as, value + 24); + newconf.yoffset = ldl_le_phys(&s->dma_as, value + 28); + + newconf.base = s->vcram_base | (value & 0xc0000000); + newconf.base += BCM2835_FB_OFFSET; + + bcm2835_fb_validate_config(&newconf); + + pitch = bcm2835_fb_get_pitch(&newconf); + size = bcm2835_fb_get_size(&newconf); + + stl_le_phys(&s->dma_as, value + 16, pitch); + stl_le_phys(&s->dma_as, value + 32, newconf.base); + stl_le_phys(&s->dma_as, value + 36, size); + + bcm2835_fb_reconfigure(s, &newconf); +} + static uint64_t bcm2835_fb_read(void *opaque, hwaddr offset, unsigned size) { BCM2835FBState *s = opaque; -- 2.18.0 Fix MPS2 SCC config register values for the mps2-an511 and mps2-an385 boards: * the SCC_AID bits [23:20] specify the FPGA build target board revision, and the SCC_CFG4 register specifies the actual board revision, so these should have matching values. Claim to be board revision C, consistently -- we had the revision in the wrong part of SCC_AID. * SCC_ID bits [15:4] should be the board number in hex, not decimal Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20180823175225.22612-1-peter.maydell@linaro.org --- hw/arm/mps2.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) diff --git a/hw/arm/mps2.c b/hw/arm/mps2.c index XXXXXXX..XXXXXXX 100644 --- a/hw/arm/mps2.c +++ b/hw/arm/mps2.c @@ -XXX,XX +XXX,XX @@ static void mps2_common_init(MachineState *machine) sccdev = DEVICE(&mms->scc); qdev_set_parent_bus(sccdev, sysbus_get_default()); qdev_prop_set_uint32(sccdev, "scc-cfg4", 0x2); - qdev_prop_set_uint32(sccdev, "scc-aid", 0x02000008); + qdev_prop_set_uint32(sccdev, "scc-aid", 0x00200008); qdev_prop_set_uint32(sccdev, "scc-id", mmc->scc_id); object_property_set_bool(OBJECT(&mms->scc), true, "realized", &error_fatal); @@ -XXX,XX +XXX,XX @@ static void mps2_an385_class_init(ObjectClass *oc, void *data) mc->desc = "ARM MPS2 with AN385 FPGA image for Cortex-M3"; mmc->fpga_type = FPGA_AN385; mc->default_cpu_type = ARM_CPU_TYPE_NAME("cortex-m3"); - mmc->scc_id = 0x41040000 | (385 << 4); + mmc->scc_id = 0x41043850; } static void mps2_an511_class_init(ObjectClass *oc, void *data) @@ -XXX,XX +XXX,XX @@ static void mps2_an511_class_init(ObjectClass *oc, void *data) mc->desc = "ARM MPS2 with AN511 DesignStart FPGA image for Cortex-M3"; mmc->fpga_type = FPGA_AN511; mc->default_cpu_type = ARM_CPU_TYPE_NAME("cortex-m3"); - mmc->scc_id = 0x4104000 | (511 << 4); + mmc->scc_id = 0x41045110; } static const TypeInfo mps2_info = { -- 2.18.0 Mostly this is RTH's memtag series, but there are also some cleanups from Philippe. thanks -- PMM The following changes since commit 10f7ffabf9c507fc02382b89912003b1c43c3231: Merge remote-tracking branch 'remotes/mcayland/tags/qemu-macppc-20200626' into staging (2020-06-26 12:14:18 +0100) are available in the Git repository at: https://git.linaro.org/people/pmaydell/qemu-arm.git tags/pull-target-arm-20200626 for you to fetch changes up to c7459633baa71d1781fde4a245d6ec9ce2f008cf: target/arm: Enable MTE (2020-06-26 14:32:24 +0100) ---------------------------------------------------------------- target-arm queue: * hw/arm/aspeed: improve QOM usage * hw/misc/pca9552: trace GPIO change events * target/arm: Implement ARMv8.5-MemTag for system emulation ---------------------------------------------------------------- Philippe Mathieu-Daudé (12): hw/arm/aspeed: Remove extraneous MemoryRegion object owner hw/arm/aspeed: Rename AspeedBoardState as AspeedMachineState hw/arm/aspeed: QOM'ify AspeedMachineState hw/i2c/core: Add i2c_try_create_slave() and i2c_realize_and_unref() hw/misc/pca9552: Rename 'nr_leds' as 'pin_count' hw/misc/pca9552: Rename generic code as pca955x hw/misc/pca9552: Add generic PCA955xClass, parent of TYPE_PCA9552 hw/misc/pca9552: Add a 'description' property for debugging purpose hw/misc/pca9552: Trace GPIO High/Low events hw/arm/aspeed: Describe each PCA9552 device hw/misc/pca9552: Trace GPIO change events hw/misc/pca9552: Model qdev output GPIOs Richard Henderson (45): target/arm: Add isar tests for mte target/arm: Improve masking of SCR RES0 bits target/arm: Add support for MTE to SCTLR_ELx target/arm: Add support for MTE to HCR_EL2 and SCR_EL3 target/arm: Rename DISAS_UPDATE to DISAS_UPDATE_EXIT target/arm: Add DISAS_UPDATE_NOCHAIN target/arm: Add MTE system registers target/arm: Add MTE bits to tb_flags target/arm: Implement the IRG instruction target/arm: Revise decoding for disas_add_sub_imm target/arm: Implement the ADDG, SUBG instructions target/arm: Implement the GMI instruction target/arm: Implement the SUBP instruction target/arm: Define arm_cpu_do_unaligned_access for user-only target/arm: Implement LDG, STG, ST2G instructions target/arm: Implement the STGP instruction target/arm: Restrict the values of DCZID.BS under TCG target/arm: Simplify DC_ZVA target/arm: Implement the LDGM, STGM, STZGM instructions target/arm: Implement the access tag cache flushes target/arm: Move regime_el to internals.h target/arm: Move regime_tcr to internals.h target/arm: Add gen_mte_check1 target/arm: Add gen_mte_checkN target/arm: Implement helper_mte_check1 target/arm: Implement helper_mte_checkN target/arm: Add helper_mte_check_zva target/arm: Use mte_checkN for sve unpredicated loads target/arm: Use mte_checkN for sve unpredicated stores target/arm: Use mte_check1 for sve LD1R target/arm: Tidy trans_LD1R_zpri target/arm: Add arm_tlb_bti_gp target/arm: Add mte helpers for sve scalar + int loads target/arm: Add mte helpers for sve scalar + int stores target/arm: Add mte helpers for sve scalar + int ff/nf loads target/arm: Handle TBI for sve scalar + int memory ops target/arm: Add mte helpers for sve scatter/gather memory ops target/arm: Complete TBI clearing for user-only for SVE target/arm: Implement data cache set allocation tags target/arm: Set PSTATE.TCO on exception entry target/arm: Always pass cacheattr to get_phys_addr target/arm: Cache the Tagged bit for a page in MemTxAttrs target/arm: Create tagged ram when MTE is enabled target/arm: Add allocation tag storage for system mode target/arm: Enable MTE include/hw/arm/aspeed.h | 12 +- include/hw/i2c/i2c.h | 2 + include/hw/misc/pca9552.h | 16 +- target/arm/cpu.h | 50 +- target/arm/helper-a64.h | 16 + target/arm/helper-sve.h | 488 ++++++++++++++ target/arm/helper.h | 2 + target/arm/internals.h | 153 ++++- target/arm/translate-a64.h | 5 + target/arm/translate.h | 23 +- hw/arm/aspeed.c | 46 +- hw/arm/virt.c | 55 +- hw/i2c/core.c | 18 +- hw/misc/pca9552.c | 216 +++++-- target/arm/cpu.c | 81 ++- target/arm/cpu64.c | 5 + target/arm/helper-a64.c | 94 +-- target/arm/helper.c | 423 ++++++++++--- target/arm/m_helper.c | 11 +- target/arm/mte_helper.c | 906 ++++++++++++++++++++++++++ target/arm/op_helper.c | 16 + target/arm/sve_helper.c | 616 ++++++++++++++---- target/arm/tlb_helper.c | 13 +- target/arm/translate-a64.c | 657 ++++++++++++++++--- target/arm/translate-sve.c | 1366 ++++++++++++++++++++++++++-------------- target/arm/translate-vfp.inc.c | 4 +- target/arm/translate.c | 16 +- hw/misc/trace-events | 4 + target/arm/Makefile.objs | 1 + 29 files changed, 4391 insertions(+), 924 deletions(-) create mode 100644 target/arm/mte_helper.c From: Philippe Mathieu-Daudé <f4bug@amsat.org> I'm confused by this code, 'bmc' is created as: bmc = g_new0(AspeedBoardState, 1); Then we use it as QOM owner for different MemoryRegion objects. But looking at memory_region_init_ram (similarly for ROM): void memory_region_init_ram(MemoryRegion *mr, struct Object *owner, const char *name, uint64_t size, Error **errp) { DeviceState *owner_dev; Error *err = NULL; memory_region_init_ram_nomigrate(mr, owner, name, size, &err); if (err) { error_propagate(errp, err); return; } /* This will assert if owner is neither NULL nor a DeviceState. * We only want the owner here for the purposes of defining a * unique name for migration. TODO: Ideally we should implement * a naming scheme for Objects which are not DeviceStates, in * which case we can relax this restriction. */ owner_dev = DEVICE(owner); vmstate_register_ram(mr, owner_dev); } The expected assertion is not triggered ('bmc' is not NULL neither a DeviceState). 'bmc' structure is defined as: struct AspeedBoardState { AspeedSoCState soc; MemoryRegion ram_container; MemoryRegion max_ram; }; What happens is when using 'OBJECT(bmc)', the QOM macros cast the memory pointed by bmc, which first member is 'soc', which is initialized ...: object_initialize_child(OBJECT(machine), "soc", &bmc->soc, amc->soc_name); The 'soc' object is indeed a DeviceState, so the assertion passes. Since this is fragile and only happens to work by luck, remove the dangerous OBJECT(bmc) owner argument. Note, this probably breaks migration for this machine. Reviewed-by: Cédric Le Goater <clg@kaod.org> Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Message-id: 20200623072132.2868-2-f4bug@amsat.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- hw/arm/aspeed.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) diff --git a/hw/arm/aspeed.c b/hw/arm/aspeed.c index XXXXXXX..XXXXXXX 100644 --- a/hw/arm/aspeed.c +++ b/hw/arm/aspeed.c @@ -XXX,XX +XXX,XX @@ static void aspeed_machine_init(MachineState *machine) * needed by the flash modules of the Aspeed machines. */ if (ASPEED_MACHINE(machine)->mmio_exec) { - memory_region_init_alias(boot_rom, OBJECT(bmc), "aspeed.boot_rom", + memory_region_init_alias(boot_rom, NULL, "aspeed.boot_rom", &fl->mmio, 0, fl->size); memory_region_add_subregion(get_system_memory(), FIRMWARE_ADDR, boot_rom); } else { - memory_region_init_rom(boot_rom, OBJECT(bmc), "aspeed.boot_rom", + memory_region_init_rom(boot_rom, NULL, "aspeed.boot_rom", fl->size, &error_abort); memory_region_add_subregion(get_system_memory(), FIRMWARE_ADDR, boot_rom); @@ -XXX,XX +XXX,XX @@ static void aspeed_machine_init(MachineState *machine) if (machine->kernel_filename && sc->num_cpus > 1) { /* With no u-boot we must set up a boot stub for the secondary CPU */ MemoryRegion *smpboot = g_new(MemoryRegion, 1); - memory_region_init_ram(smpboot, OBJECT(bmc), "aspeed.smpboot", + memory_region_init_ram(smpboot, NULL, "aspeed.smpboot", 0x80, &error_abort); memory_region_add_subregion(get_system_memory(), AST_SMP_MAILBOX_BASE, smpboot); -- 2.20.1 From: Philippe Mathieu-Daudé <f4bug@amsat.org> To have a more consistent naming, rename AspeedBoardState as AspeedMachineState. Suggested-by: Cédric Le Goater <clg@kaod.org> Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Reviewed-by: Cédric Le Goater <clg@kaod.org> Message-id: 20200623072132.2868-3-f4bug@amsat.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- include/hw/arm/aspeed.h | 4 ++-- hw/arm/aspeed.c | 20 ++++++++++---------- 2 files changed, 12 insertions(+), 12 deletions(-) diff --git a/include/hw/arm/aspeed.h b/include/hw/arm/aspeed.h index XXXXXXX..XXXXXXX 100644 --- a/include/hw/arm/aspeed.h +++ b/include/hw/arm/aspeed.h @@ -XXX,XX +XXX,XX @@ #include "hw/boards.h" -typedef struct AspeedBoardState AspeedBoardState; +typedef struct AspeedMachineState AspeedMachineState; #define TYPE_ASPEED_MACHINE MACHINE_TYPE_NAME("aspeed") #define ASPEED_MACHINE(obj) \ @@ -XXX,XX +XXX,XX @@ typedef struct AspeedMachineClass { const char *spi_model; uint32_t num_cs; uint32_t macs_mask; - void (*i2c_init)(AspeedBoardState *bmc); + void (*i2c_init)(AspeedMachineState *bmc); } AspeedMachineClass; diff --git a/hw/arm/aspeed.c b/hw/arm/aspeed.c index XXXXXXX..XXXXXXX 100644 --- a/hw/arm/aspeed.c +++ b/hw/arm/aspeed.c @@ -XXX,XX +XXX,XX @@ static struct arm_boot_info aspeed_board_binfo = { .board_id = -1, /* device-tree-only board */ }; -struct AspeedBoardState { +struct AspeedMachineState { AspeedSoCState soc; MemoryRegion ram_container; MemoryRegion max_ram; @@ -XXX,XX +XXX,XX @@ static void sdhci_attach_drive(SDHCIState *sdhci, DriveInfo *dinfo) static void aspeed_machine_init(MachineState *machine) { - AspeedBoardState *bmc; + AspeedMachineState *bmc; AspeedMachineClass *amc = ASPEED_MACHINE_GET_CLASS(machine); AspeedSoCClass *sc; DriveInfo *drive0 = drive_get(IF_MTD, 0, 0); @@ -XXX,XX +XXX,XX @@ static void aspeed_machine_init(MachineState *machine) int i; NICInfo *nd = &nd_table[0]; - bmc = g_new0(AspeedBoardState, 1); + bmc = g_new0(AspeedMachineState, 1); memory_region_init(&bmc->ram_container, NULL, "aspeed-ram-container", 4 * GiB); @@ -XXX,XX +XXX,XX @@ static void aspeed_machine_init(MachineState *machine) arm_load_kernel(ARM_CPU(first_cpu), machine, &aspeed_board_binfo); } -static void palmetto_bmc_i2c_init(AspeedBoardState *bmc) +static void palmetto_bmc_i2c_init(AspeedMachineState *bmc) { AspeedSoCState *soc = &bmc->soc; DeviceState *dev; @@ -XXX,XX +XXX,XX @@ static void palmetto_bmc_i2c_init(AspeedBoardState *bmc) object_property_set_int(OBJECT(dev), 110000, "temperature3", &error_abort); } -static void ast2500_evb_i2c_init(AspeedBoardState *bmc) +static void ast2500_evb_i2c_init(AspeedMachineState *bmc) { AspeedSoCState *soc = &bmc->soc; uint8_t *eeprom_buf = g_malloc0(8 * 1024); @@ -XXX,XX +XXX,XX @@ static void ast2500_evb_i2c_init(AspeedBoardState *bmc) i2c_create_slave(aspeed_i2c_get_bus(DEVICE(&soc->i2c), 11), "ds1338", 0x32); } -static void ast2600_evb_i2c_init(AspeedBoardState *bmc) +static void ast2600_evb_i2c_init(AspeedMachineState *bmc) { /* Start with some devices on our I2C busses */ ast2500_evb_i2c_init(bmc); } -static void romulus_bmc_i2c_init(AspeedBoardState *bmc) +static void romulus_bmc_i2c_init(AspeedMachineState *bmc) { AspeedSoCState *soc = &bmc->soc; @@ -XXX,XX +XXX,XX @@ static void romulus_bmc_i2c_init(AspeedBoardState *bmc) i2c_create_slave(aspeed_i2c_get_bus(DEVICE(&soc->i2c), 11), "ds1338", 0x32); } -static void swift_bmc_i2c_init(AspeedBoardState *bmc) +static void swift_bmc_i2c_init(AspeedMachineState *bmc) { AspeedSoCState *soc = &bmc->soc; @@ -XXX,XX +XXX,XX @@ static void swift_bmc_i2c_init(AspeedBoardState *bmc) i2c_create_slave(aspeed_i2c_get_bus(DEVICE(&soc->i2c), 12), "tmp105", 0x4a); } -static void sonorapass_bmc_i2c_init(AspeedBoardState *bmc) +static void sonorapass_bmc_i2c_init(AspeedMachineState *bmc) { AspeedSoCState *soc = &bmc->soc; @@ -XXX,XX +XXX,XX @@ static void sonorapass_bmc_i2c_init(AspeedBoardState *bmc) } -static void witherspoon_bmc_i2c_init(AspeedBoardState *bmc) +static void witherspoon_bmc_i2c_init(AspeedMachineState *bmc) { AspeedSoCState *soc = &bmc->soc; uint8_t *eeprom_buf = g_malloc0(8 * 1024); -- 2.20.1 From: Philippe Mathieu-Daudé <f4bug@amsat.org> AspeedMachineState seems crippled. We use incorrectly 2 different structures to do the same thing. Merge them altogether: - Move AspeedMachine fields to AspeedMachineState - AspeedMachineState is now QOM - Remove unused AspeedMachine structure Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Reviewed-by: Cédric Le Goater <clg@kaod.org> Message-id: 20200623072132.2868-4-f4bug@amsat.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- include/hw/arm/aspeed.h | 8 +------- hw/arm/aspeed.c | 11 +++++++---- 2 files changed, 8 insertions(+), 11 deletions(-) diff --git a/include/hw/arm/aspeed.h b/include/hw/arm/aspeed.h index XXXXXXX..XXXXXXX 100644 --- a/include/hw/arm/aspeed.h +++ b/include/hw/arm/aspeed.h @@ -XXX,XX +XXX,XX @@ typedef struct AspeedMachineState AspeedMachineState; #define TYPE_ASPEED_MACHINE MACHINE_TYPE_NAME("aspeed") #define ASPEED_MACHINE(obj) \ - OBJECT_CHECK(AspeedMachine, (obj), TYPE_ASPEED_MACHINE) - -typedef struct AspeedMachine { - MachineState parent_obj; - - bool mmio_exec; -} AspeedMachine; + OBJECT_CHECK(AspeedMachineState, (obj), TYPE_ASPEED_MACHINE) #define ASPEED_MAC0_ON (1 << 0) #define ASPEED_MAC1_ON (1 << 1) diff --git a/hw/arm/aspeed.c b/hw/arm/aspeed.c index XXXXXXX..XXXXXXX 100644 --- a/hw/arm/aspeed.c +++ b/hw/arm/aspeed.c @@ -XXX,XX +XXX,XX @@ static struct arm_boot_info aspeed_board_binfo = { }; struct AspeedMachineState { + /* Private */ + MachineState parent_obj; + /* Public */ + AspeedSoCState soc; MemoryRegion ram_container; MemoryRegion max_ram; + bool mmio_exec; }; /* Palmetto hardware value: 0x120CE416 */ @@ -XXX,XX +XXX,XX @@ static void sdhci_attach_drive(SDHCIState *sdhci, DriveInfo *dinfo) static void aspeed_machine_init(MachineState *machine) { - AspeedMachineState *bmc; + AspeedMachineState *bmc = ASPEED_MACHINE(machine); AspeedMachineClass *amc = ASPEED_MACHINE_GET_CLASS(machine); AspeedSoCClass *sc; DriveInfo *drive0 = drive_get(IF_MTD, 0, 0); @@ -XXX,XX +XXX,XX @@ static void aspeed_machine_init(MachineState *machine) int i; NICInfo *nd = &nd_table[0]; - bmc = g_new0(AspeedMachineState, 1); - memory_region_init(&bmc->ram_container, NULL, "aspeed-ram-container", 4 * GiB); memory_region_add_subregion(&bmc->ram_container, 0, machine->ram); @@ -XXX,XX +XXX,XX @@ static const TypeInfo aspeed_machine_types[] = { }, { .name = TYPE_ASPEED_MACHINE, .parent = TYPE_MACHINE, - .instance_size = sizeof(AspeedMachine), + .instance_size = sizeof(AspeedMachineState), .instance_init = aspeed_machine_instance_init, .class_size = sizeof(AspeedMachineClass), .class_init = aspeed_machine_class_init, -- 2.20.1 From: Philippe Mathieu-Daudé <f4bug@amsat.org> Extract i2c_try_create_slave() and i2c_realize_and_unref() from i2c_create_slave(). We can now set properties on a I2CSlave before it is realized. This is in line with the recent qdev/QOM changes merged in commit 6675a653d2e. Reviewed-by: Corey Minyard <cminyard@mvista.com> Reviewed-by: Cédric Le Goater <clg@kaod.org> Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Reviewed-by: Markus Armbruster <armbru@redhat.com> Tested-by: Cédric Le Goater <clg@kaod.org> Message-id: 20200623072723.6324-2-f4bug@amsat.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- include/hw/i2c/i2c.h | 2 ++ hw/i2c/core.c | 18 ++++++++++++++++-- 2 files changed, 18 insertions(+), 2 deletions(-) diff --git a/include/hw/i2c/i2c.h b/include/hw/i2c/i2c.h index XXXXXXX..XXXXXXX 100644 --- a/include/hw/i2c/i2c.h +++ b/include/hw/i2c/i2c.h @@ -XXX,XX +XXX,XX @@ int i2c_send(I2CBus *bus, uint8_t data); uint8_t i2c_recv(I2CBus *bus); DeviceState *i2c_create_slave(I2CBus *bus, const char *name, uint8_t addr); +DeviceState *i2c_try_create_slave(const char *name, uint8_t addr); +bool i2c_realize_and_unref(DeviceState *dev, I2CBus *bus, Error **errp); /* lm832x.c */ void lm832x_key_event(DeviceState *dev, int key, int state); diff --git a/hw/i2c/core.c b/hw/i2c/core.c index XXXXXXX..XXXXXXX 100644 --- a/hw/i2c/core.c +++ b/hw/i2c/core.c @@ -XXX,XX +XXX,XX @@ const VMStateDescription vmstate_i2c_slave = { } }; -DeviceState *i2c_create_slave(I2CBus *bus, const char *name, uint8_t addr) +DeviceState *i2c_try_create_slave(const char *name, uint8_t addr) { DeviceState *dev; dev = qdev_new(name); qdev_prop_set_uint8(dev, "address", addr); - qdev_realize_and_unref(dev, &bus->qbus, &error_fatal); + return dev; +} + +bool i2c_realize_and_unref(DeviceState *dev, I2CBus *bus, Error **errp) +{ + return qdev_realize_and_unref(dev, &bus->qbus, errp); +} + +DeviceState *i2c_create_slave(I2CBus *bus, const char *name, uint8_t addr) +{ + DeviceState *dev; + + dev = i2c_try_create_slave(name, addr); + i2c_realize_and_unref(dev, bus, &error_fatal); + return dev; } -- 2.20.1 From: Philippe Mathieu-Daudé <f4bug@amsat.org> The PCA9552 device does not expose LEDs, but simple pins to connnect LEDs to. To be clearer with the device model, rename 'nr_leds' as 'pin_count'. Reviewed-by: Cédric Le Goater <clg@kaod.org> Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Tested-by: Cédric Le Goater <clg@kaod.org> Message-id: 20200623072723.6324-3-f4bug@amsat.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- include/hw/misc/pca9552.h | 2 +- hw/misc/pca9552.c | 10 +++++----- 2 files changed, 6 insertions(+), 6 deletions(-) diff --git a/include/hw/misc/pca9552.h b/include/hw/misc/pca9552.h index XXXXXXX..XXXXXXX 100644 --- a/include/hw/misc/pca9552.h +++ b/include/hw/misc/pca9552.h @@ -XXX,XX +XXX,XX @@ typedef struct PCA9552State { uint8_t regs[PCA9552_NR_REGS]; uint8_t max_reg; - uint8_t nr_leds; + uint8_t pin_count; } PCA9552State; #endif diff --git a/hw/misc/pca9552.c b/hw/misc/pca9552.c index XXXXXXX..XXXXXXX 100644 --- a/hw/misc/pca9552.c +++ b/hw/misc/pca9552.c @@ -XXX,XX +XXX,XX @@ static void pca9552_update_pin_input(PCA9552State *s) { int i; - for (i = 0; i < s->nr_leds; i++) { + for (i = 0; i < s->pin_count; i++) { uint8_t input_reg = PCA9552_INPUT0 + (i / 8); uint8_t input_shift = (i % 8); uint8_t config = pca9552_pin_get_config(s, i); @@ -XXX,XX +XXX,XX @@ static void pca9552_get_led(Object *obj, Visitor *v, const char *name, error_setg(errp, "%s: error reading %s", __func__, name); return; } - if (led < 0 || led > s->nr_leds) { + if (led < 0 || led > s->pin_count) { error_setg(errp, "%s invalid led %s", __func__, name); return; } @@ -XXX,XX +XXX,XX @@ static void pca9552_set_led(Object *obj, Visitor *v, const char *name, error_setg(errp, "%s: error reading %s", __func__, name); return; } - if (led < 0 || led > s->nr_leds) { + if (led < 0 || led > s->pin_count) { error_setg(errp, "%s invalid led %s", __func__, name); return; } @@ -XXX,XX +XXX,XX @@ static void pca9552_initfn(Object *obj) * PCA955X device */ s->max_reg = PCA9552_LS3; - s->nr_leds = 16; + s->pin_count = 16; - for (led = 0; led < s->nr_leds; led++) { + for (led = 0; led < s->pin_count; led++) { char *name; name = g_strdup_printf("led%d", led); -- 2.20.1 From: Philippe Mathieu-Daudé <f4bug@amsat.org> Various code from the PCA9552 device model is generic to the PCA955X family. We'll split the generic code in a base class in the next commit. To ease review, first do a dumb renaming. Reviewed-by: Cédric Le Goater <clg@kaod.org> Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Tested-by: Cédric Le Goater <clg@kaod.org> Message-id: 20200623072723.6324-4-f4bug@amsat.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- include/hw/misc/pca9552.h | 10 ++--- hw/misc/pca9552.c | 80 +++++++++++++++++++-------------------- 2 files changed, 45 insertions(+), 45 deletions(-) diff --git a/include/hw/misc/pca9552.h b/include/hw/misc/pca9552.h index XXXXXXX..XXXXXXX 100644 --- a/include/hw/misc/pca9552.h +++ b/include/hw/misc/pca9552.h @@ -XXX,XX +XXX,XX @@ #include "hw/i2c/i2c.h" #define TYPE_PCA9552 "pca9552" -#define PCA9552(obj) OBJECT_CHECK(PCA9552State, (obj), TYPE_PCA9552) +#define PCA955X(obj) OBJECT_CHECK(PCA955xState, (obj), TYPE_PCA9552) -#define PCA9552_NR_REGS 10 +#define PCA955X_NR_REGS 10 -typedef struct PCA9552State { +typedef struct PCA955xState { /*< private >*/ I2CSlave i2c; /*< public >*/ @@ -XXX,XX +XXX,XX @@ typedef struct PCA9552State { uint8_t len; uint8_t pointer; - uint8_t regs[PCA9552_NR_REGS]; + uint8_t regs[PCA955X_NR_REGS]; uint8_t max_reg; uint8_t pin_count; -} PCA9552State; +} PCA955xState; #endif diff --git a/hw/misc/pca9552.c b/hw/misc/pca9552.c index XXXXXXX..XXXXXXX 100644 --- a/hw/misc/pca9552.c +++ b/hw/misc/pca9552.c @@ -XXX,XX +XXX,XX @@ static const char *led_state[] = {"on", "off", "pwm0", "pwm1"}; -static uint8_t pca9552_pin_get_config(PCA9552State *s, int pin) +static uint8_t pca955x_pin_get_config(PCA955xState *s, int pin) { uint8_t reg = PCA9552_LS0 + (pin / 4); uint8_t shift = (pin % 4) << 1; @@ -XXX,XX +XXX,XX @@ static uint8_t pca9552_pin_get_config(PCA9552State *s, int pin) return extract32(s->regs[reg], shift, 2); } -static void pca9552_update_pin_input(PCA9552State *s) +static void pca955x_update_pin_input(PCA955xState *s) { int i; for (i = 0; i < s->pin_count; i++) { uint8_t input_reg = PCA9552_INPUT0 + (i / 8); uint8_t input_shift = (i % 8); - uint8_t config = pca9552_pin_get_config(s, i); + uint8_t config = pca955x_pin_get_config(s, i); switch (config) { case PCA9552_LED_ON: @@ -XXX,XX +XXX,XX @@ static void pca9552_update_pin_input(PCA9552State *s) } } -static uint8_t pca9552_read(PCA9552State *s, uint8_t reg) +static uint8_t pca955x_read(PCA955xState *s, uint8_t reg) { switch (reg) { case PCA9552_INPUT0: @@ -XXX,XX +XXX,XX @@ static uint8_t pca9552_read(PCA9552State *s, uint8_t reg) } } -static void pca9552_write(PCA9552State *s, uint8_t reg, uint8_t data) +static void pca955x_write(PCA955xState *s, uint8_t reg, uint8_t data) { switch (reg) { case PCA9552_PSC0: @@ -XXX,XX +XXX,XX @@ static void pca9552_write(PCA9552State *s, uint8_t reg, uint8_t data) case PCA9552_LS2: case PCA9552_LS3: s->regs[reg] = data; - pca9552_update_pin_input(s); + pca955x_update_pin_input(s); break; case PCA9552_INPUT0: @@ -XXX,XX +XXX,XX @@ static void pca9552_write(PCA9552State *s, uint8_t reg, uint8_t data) * after each byte is sent to or received by the device. The index * rollovers to 0 when the maximum register address is reached. */ -static void pca9552_autoinc(PCA9552State *s) +static void pca955x_autoinc(PCA955xState *s) { if (s->pointer != 0xFF && s->pointer & PCA9552_AUTOINC) { uint8_t reg = s->pointer & 0xf; @@ -XXX,XX +XXX,XX @@ static void pca9552_autoinc(PCA9552State *s) } } -static uint8_t pca9552_recv(I2CSlave *i2c) +static uint8_t pca955x_recv(I2CSlave *i2c) { - PCA9552State *s = PCA9552(i2c); + PCA955xState *s = PCA955X(i2c); uint8_t ret; - ret = pca9552_read(s, s->pointer & 0xf); + ret = pca955x_read(s, s->pointer & 0xf); /* * From the Specs: @@ -XXX,XX +XXX,XX @@ static uint8_t pca9552_recv(I2CSlave *i2c) __func__); } - pca9552_autoinc(s); + pca955x_autoinc(s); return ret; } -static int pca9552_send(I2CSlave *i2c, uint8_t data) +static int pca955x_send(I2CSlave *i2c, uint8_t data) { - PCA9552State *s = PCA9552(i2c); + PCA955xState *s = PCA955X(i2c); /* First byte sent by is the register address */ if (s->len == 0) { s->pointer = data; s->len++; } else { - pca9552_write(s, s->pointer & 0xf, data); + pca955x_write(s, s->pointer & 0xf, data); - pca9552_autoinc(s); + pca955x_autoinc(s); } return 0; } -static int pca9552_event(I2CSlave *i2c, enum i2c_event event) +static int pca955x_event(I2CSlave *i2c, enum i2c_event event) { - PCA9552State *s = PCA9552(i2c); + PCA955xState *s = PCA955X(i2c); s->len = 0; return 0; } -static void pca9552_get_led(Object *obj, Visitor *v, const char *name, +static void pca955x_get_led(Object *obj, Visitor *v, const char *name, void *opaque, Error **errp) { - PCA9552State *s = PCA9552(obj); + PCA955xState *s = PCA955X(obj); int led, rc, reg; uint8_t state; @@ -XXX,XX +XXX,XX @@ static void pca9552_get_led(Object *obj, Visitor *v, const char *name, * reading the INPUTx reg */ reg = PCA9552_LS0 + led / 4; - state = (pca9552_read(s, reg) >> (led % 8)) & 0x3; + state = (pca955x_read(s, reg) >> (led % 8)) & 0x3; visit_type_str(v, name, (char **)&led_state[state], errp); } @@ -XXX,XX +XXX,XX @@ static inline uint8_t pca955x_ledsel(uint8_t oldval, int led_num, int state) ((state & 0x3) << (led_num << 1)); } -static void pca9552_set_led(Object *obj, Visitor *v, const char *name, +static void pca955x_set_led(Object *obj, Visitor *v, const char *name, void *opaque, Error **errp) { - PCA9552State *s = PCA9552(obj); + PCA955xState *s = PCA955X(obj); Error *local_err = NULL; int led, rc, reg, val; uint8_t state; @@ -XXX,XX +XXX,XX @@ static void pca9552_set_led(Object *obj, Visitor *v, const char *name, } reg = PCA9552_LS0 + led / 4; - val = pca9552_read(s, reg); + val = pca955x_read(s, reg); val = pca955x_ledsel(val, led % 4, state); - pca9552_write(s, reg, val); + pca955x_write(s, reg, val); } static const VMStateDescription pca9552_vmstate = { @@ -XXX,XX +XXX,XX @@ static const VMStateDescription pca9552_vmstate = { .version_id = 0, .minimum_version_id = 0, .fields = (VMStateField[]) { - VMSTATE_UINT8(len, PCA9552State), - VMSTATE_UINT8(pointer, PCA9552State), - VMSTATE_UINT8_ARRAY(regs, PCA9552State, PCA9552_NR_REGS), - VMSTATE_I2C_SLAVE(i2c, PCA9552State), + VMSTATE_UINT8(len, PCA955xState), + VMSTATE_UINT8(pointer, PCA955xState), + VMSTATE_UINT8_ARRAY(regs, PCA955xState, PCA955X_NR_REGS), + VMSTATE_I2C_SLAVE(i2c, PCA955xState), VMSTATE_END_OF_LIST() } }; static void pca9552_reset(DeviceState *dev) { - PCA9552State *s = PCA9552(dev); + PCA955xState *s = PCA955X(dev); s->regs[PCA9552_PSC0] = 0xFF; s->regs[PCA9552_PWM0] = 0x80; @@ -XXX,XX +XXX,XX @@ static void pca9552_reset(DeviceState *dev) s->regs[PCA9552_LS2] = 0x55; s->regs[PCA9552_LS3] = 0x55; - pca9552_update_pin_input(s); + pca955x_update_pin_input(s); s->pointer = 0xFF; s->len = 0; } -static void pca9552_initfn(Object *obj) +static void pca955x_initfn(Object *obj) { - PCA9552State *s = PCA9552(obj); + PCA955xState *s = PCA955X(obj); int led; /* If support for the other PCA955X devices are implemented, these @@ -XXX,XX +XXX,XX @@ static void pca9552_initfn(Object *obj) char *name; name = g_strdup_printf("led%d", led); - object_property_add(obj, name, "bool", pca9552_get_led, pca9552_set_led, + object_property_add(obj, name, "bool", pca955x_get_led, pca955x_set_led, NULL, NULL); g_free(name); } @@ -XXX,XX +XXX,XX @@ static void pca9552_class_init(ObjectClass *klass, void *data) DeviceClass *dc = DEVICE_CLASS(klass); I2CSlaveClass *k = I2C_SLAVE_CLASS(klass); - k->event = pca9552_event; - k->recv = pca9552_recv; - k->send = pca9552_send; + k->event = pca955x_event; + k->recv = pca955x_recv; + k->send = pca955x_send; dc->reset = pca9552_reset; dc->vmsd = &pca9552_vmstate; } @@ -XXX,XX +XXX,XX @@ static void pca9552_class_init(ObjectClass *klass, void *data) static const TypeInfo pca9552_info = { .name = TYPE_PCA9552, .parent = TYPE_I2C_SLAVE, - .instance_init = pca9552_initfn, - .instance_size = sizeof(PCA9552State), + .instance_init = pca955x_initfn, + .instance_size = sizeof(PCA955xState), .class_init = pca9552_class_init, }; -static void pca9552_register_types(void) +static void pca955x_register_types(void) { type_register_static(&pca9552_info); } -type_init(pca9552_register_types) +type_init(pca955x_register_types) -- 2.20.1 From: Philippe Mathieu-Daudé <f4bug@amsat.org> Extract the code common to the PCA955x family in PCA955xClass, keeping the PCA9552 specific parts into pca9552_class_init(). Remove the 'TODO' comment added in commit 5141d4158cf. Suggested-by: Cédric Le Goater <clg@kaod.org> Reviewed-by: Cédric Le Goater <clg@kaod.org> Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Tested-by: Cédric Le Goater <clg@kaod.org> Message-id: 20200623072723.6324-5-f4bug@amsat.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- include/hw/misc/pca9552.h | 6 ++-- hw/misc/pca9552.c | 66 ++++++++++++++++++++++++++++----------- 2 files changed, 51 insertions(+), 21 deletions(-) diff --git a/include/hw/misc/pca9552.h b/include/hw/misc/pca9552.h index XXXXXXX..XXXXXXX 100644 --- a/include/hw/misc/pca9552.h +++ b/include/hw/misc/pca9552.h @@ -XXX,XX +XXX,XX @@ #include "hw/i2c/i2c.h" #define TYPE_PCA9552 "pca9552" -#define PCA955X(obj) OBJECT_CHECK(PCA955xState, (obj), TYPE_PCA9552) +#define TYPE_PCA955X "pca955x" +#define PCA955X(obj) OBJECT_CHECK(PCA955xState, (obj), TYPE_PCA955X) #define PCA955X_NR_REGS 10 +#define PCA955X_PIN_COUNT_MAX 16 typedef struct PCA955xState { /*< private >*/ @@ -XXX,XX +XXX,XX @@ typedef struct PCA955xState { uint8_t pointer; uint8_t regs[PCA955X_NR_REGS]; - uint8_t max_reg; - uint8_t pin_count; } PCA955xState; #endif diff --git a/hw/misc/pca9552.c b/hw/misc/pca9552.c index XXXXXXX..XXXXXXX 100644 --- a/hw/misc/pca9552.c +++ b/hw/misc/pca9552.c @@ -XXX,XX +XXX,XX @@ * https://www.nxp.com/docs/en/application-note/AN264.pdf * * Copyright (c) 2017-2018, IBM Corporation. + * Copyright (c) 2020 Philippe Mathieu-Daudé * * This work is licensed under the terms of the GNU GPL, version 2 or * later. See the COPYING file in the top-level directory. @@ -XXX,XX +XXX,XX @@ #include "qapi/error.h" #include "qapi/visitor.h" +typedef struct PCA955xClass { + /*< private >*/ + I2CSlaveClass parent_class; + /*< public >*/ + + uint8_t pin_count; + uint8_t max_reg; +} PCA955xClass; + +#define PCA955X_CLASS(klass) \ + OBJECT_CLASS_CHECK(PCA955xClass, (klass), TYPE_PCA955X) +#define PCA955X_GET_CLASS(obj) \ + OBJECT_GET_CLASS(PCA955xClass, (obj), TYPE_PCA955X) + #define PCA9552_LED_ON 0x0 #define PCA9552_LED_OFF 0x1 #define PCA9552_LED_PWM0 0x2 @@ -XXX,XX +XXX,XX @@ static uint8_t pca955x_pin_get_config(PCA955xState *s, int pin) static void pca955x_update_pin_input(PCA955xState *s) { + PCA955xClass *k = PCA955X_GET_CLASS(s); int i; - for (i = 0; i < s->pin_count; i++) { + for (i = 0; i < k->pin_count; i++) { uint8_t input_reg = PCA9552_INPUT0 + (i / 8); uint8_t input_shift = (i % 8); uint8_t config = pca955x_pin_get_config(s, i); @@ -XXX,XX +XXX,XX @@ static void pca955x_write(PCA955xState *s, uint8_t reg, uint8_t data) */ static void pca955x_autoinc(PCA955xState *s) { + PCA955xClass *k = PCA955X_GET_CLASS(s); + if (s->pointer != 0xFF && s->pointer & PCA9552_AUTOINC) { uint8_t reg = s->pointer & 0xf; - reg = (reg + 1) % (s->max_reg + 1); + reg = (reg + 1) % (k->max_reg + 1); s->pointer = reg | PCA9552_AUTOINC; } } @@ -XXX,XX +XXX,XX @@ static int pca955x_event(I2CSlave *i2c, enum i2c_event event) static void pca955x_get_led(Object *obj, Visitor *v, const char *name, void *opaque, Error **errp) { + PCA955xClass *k = PCA955X_GET_CLASS(obj); PCA955xState *s = PCA955X(obj); int led, rc, reg; uint8_t state; @@ -XXX,XX +XXX,XX @@ static void pca955x_get_led(Object *obj, Visitor *v, const char *name, error_setg(errp, "%s: error reading %s", __func__, name); return; } - if (led < 0 || led > s->pin_count) { + if (led < 0 || led > k->pin_count) { error_setg(errp, "%s invalid led %s", __func__, name); return; } @@ -XXX,XX +XXX,XX @@ static inline uint8_t pca955x_ledsel(uint8_t oldval, int led_num, int state) static void pca955x_set_led(Object *obj, Visitor *v, const char *name, void *opaque, Error **errp) { + PCA955xClass *k = PCA955X_GET_CLASS(obj); PCA955xState *s = PCA955X(obj); Error *local_err = NULL; int led, rc, reg, val; @@ -XXX,XX +XXX,XX @@ static void pca955x_set_led(Object *obj, Visitor *v, const char *name, error_setg(errp, "%s: error reading %s", __func__, name); return; } - if (led < 0 || led > s->pin_count) { + if (led < 0 || led > k->pin_count) { error_setg(errp, "%s invalid led %s", __func__, name); return; } @@ -XXX,XX +XXX,XX @@ static void pca9552_reset(DeviceState *dev) static void pca955x_initfn(Object *obj) { - PCA955xState *s = PCA955X(obj); + PCA955xClass *k = PCA955X_GET_CLASS(obj); int led; - /* If support for the other PCA955X devices are implemented, these - * constant values might be part of class structure describing the - * PCA955X device - */ - s->max_reg = PCA9552_LS3; - s->pin_count = 16; - - for (led = 0; led < s->pin_count; led++) { + assert(k->pin_count <= PCA955X_PIN_COUNT_MAX); + for (led = 0; led < k->pin_count; led++) { char *name; name = g_strdup_printf("led%d", led); @@ -XXX,XX +XXX,XX @@ static void pca955x_initfn(Object *obj) } } -static void pca9552_class_init(ObjectClass *klass, void *data) +static void pca955x_class_init(ObjectClass *klass, void *data) { - DeviceClass *dc = DEVICE_CLASS(klass); I2CSlaveClass *k = I2C_SLAVE_CLASS(klass); k->event = pca955x_event; k->recv = pca955x_recv; k->send = pca955x_send; +} + +static const TypeInfo pca955x_info = { + .name = TYPE_PCA955X, + .parent = TYPE_I2C_SLAVE, + .instance_init = pca955x_initfn, + .instance_size = sizeof(PCA955xState), + .class_init = pca955x_class_init, + .abstract = true, +}; + +static void pca9552_class_init(ObjectClass *oc, void *data) +{ + DeviceClass *dc = DEVICE_CLASS(oc); + PCA955xClass *pc = PCA955X_CLASS(oc); + dc->reset = pca9552_reset; dc->vmsd = &pca9552_vmstate; + pc->max_reg = PCA9552_LS3; + pc->pin_count = 16; } static const TypeInfo pca9552_info = { .name = TYPE_PCA9552, - .parent = TYPE_I2C_SLAVE, - .instance_init = pca955x_initfn, - .instance_size = sizeof(PCA955xState), + .parent = TYPE_PCA955X, .class_init = pca9552_class_init, }; static void pca955x_register_types(void) { + type_register_static(&pca955x_info); type_register_static(&pca9552_info); } -- 2.20.1 From: Philippe Mathieu-Daudé <f4bug@amsat.org> Add a description field to distinguish between multiple devices. Reviewed-by: Cédric Le Goater <clg@kaod.org> Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Tested-by: Cédric Le Goater <clg@kaod.org> Message-id: 20200623072723.6324-6-f4bug@amsat.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- include/hw/misc/pca9552.h | 1 + hw/misc/pca9552.c | 18 ++++++++++++++++++ 2 files changed, 19 insertions(+) diff --git a/include/hw/misc/pca9552.h b/include/hw/misc/pca9552.h index XXXXXXX..XXXXXXX 100644 --- a/include/hw/misc/pca9552.h +++ b/include/hw/misc/pca9552.h @@ -XXX,XX +XXX,XX @@ typedef struct PCA955xState { uint8_t pointer; uint8_t regs[PCA955X_NR_REGS]; + char *description; /* For debugging purpose only */ } PCA955xState; #endif diff --git a/hw/misc/pca9552.c b/hw/misc/pca9552.c index XXXXXXX..XXXXXXX 100644 --- a/hw/misc/pca9552.c +++ b/hw/misc/pca9552.c @@ -XXX,XX +XXX,XX @@ #include "qemu/osdep.h" #include "qemu/log.h" #include "qemu/module.h" +#include "hw/qdev-properties.h" #include "hw/misc/pca9552.h" #include "hw/misc/pca9552_regs.h" #include "migration/vmstate.h" @@ -XXX,XX +XXX,XX @@ static void pca955x_initfn(Object *obj) } } +static void pca955x_realize(DeviceState *dev, Error **errp) +{ + PCA955xState *s = PCA955X(dev); + + if (!s->description) { + s->description = g_strdup("pca-unspecified"); + } +} + +static Property pca955x_properties[] = { + DEFINE_PROP_STRING("description", PCA955xState, description), + DEFINE_PROP_END_OF_LIST(), +}; + static void pca955x_class_init(ObjectClass *klass, void *data) { + DeviceClass *dc = DEVICE_CLASS(klass); I2CSlaveClass *k = I2C_SLAVE_CLASS(klass); k->event = pca955x_event; k->recv = pca955x_recv; k->send = pca955x_send; + dc->realize = pca955x_realize; + device_class_set_props(dc, pca955x_properties); } static const TypeInfo pca955x_info = { -- 2.20.1 From: Philippe Mathieu-Daudé <f4bug@amsat.org> Add a trivial representation of the PCA9552 GPIOs. Example booting obmc-phosphor-image: $ qemu-system-arm -M witherspoon-bmc -trace pca955x_gpio_status 1592689902.327837:pca955x_gpio_status pca-unspecified GPIOs 0-15 [*...............] 1592689902.329934:pca955x_gpio_status pca-unspecified GPIOs 0-15 [**..............] 1592689902.330717:pca955x_gpio_status pca-unspecified GPIOs 0-15 [***.............] 1592689902.331431:pca955x_gpio_status pca-unspecified GPIOs 0-15 [****............] 1592689902.332163:pca955x_gpio_status pca-unspecified GPIOs 0-15 [****.........*..] 1592689902.332888:pca955x_gpio_status pca-unspecified GPIOs 0-15 [****.........**.] 1592689902.333629:pca955x_gpio_status pca-unspecified GPIOs 0-15 [****.........***] 1592690032.793289:pca955x_gpio_status pca-unspecified GPIOs 0-15 [****.........*.*] 1592690033.303163:pca955x_gpio_status pca-unspecified GPIOs 0-15 [****.........***] 1592690033.812962:pca955x_gpio_status pca-unspecified GPIOs 0-15 [****.........*.*] 1592690034.323234:pca955x_gpio_status pca-unspecified GPIOs 0-15 [****.........***] 1592690034.832922:pca955x_gpio_status pca-unspecified GPIOs 0-15 [****.........*.*] We notice the GPIO #14 (front-power LED) starts to blink. This LED is described in the witherspoon device-tree [*]: front-power { retain-state-shutdown; default-state = "keep"; gpios = <&pca0 14 GPIO_ACTIVE_LOW>; }; [*] https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/arch/arm/boot/dts/aspeed-bmc-opp-witherspoon.dts?id=b1f9be9392f0#n140 Suggested-by: Cédric Le Goater <clg@kaod.org> Reviewed-by: Cédric Le Goater <clg@kaod.org> Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Tested-by: Cédric Le Goater <clg@kaod.org> Message-id: 20200623072723.6324-7-f4bug@amsat.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- hw/misc/pca9552.c | 39 +++++++++++++++++++++++++++++++++++++++ hw/misc/trace-events | 3 +++ 2 files changed, 42 insertions(+) diff --git a/hw/misc/pca9552.c b/hw/misc/pca9552.c index XXXXXXX..XXXXXXX 100644 --- a/hw/misc/pca9552.c +++ b/hw/misc/pca9552.c @@ -XXX,XX +XXX,XX @@ #include "qemu/osdep.h" #include "qemu/log.h" #include "qemu/module.h" +#include "qemu/bitops.h" #include "hw/qdev-properties.h" #include "hw/misc/pca9552.h" #include "hw/misc/pca9552_regs.h" #include "migration/vmstate.h" #include "qapi/error.h" #include "qapi/visitor.h" +#include "trace.h" typedef struct PCA955xClass { /*< private >*/ @@ -XXX,XX +XXX,XX @@ static uint8_t pca955x_pin_get_config(PCA955xState *s, int pin) return extract32(s->regs[reg], shift, 2); } +/* Return INPUT status (bit #N belongs to GPIO #N) */ +static uint16_t pca955x_pins_get_status(PCA955xState *s) +{ + return (s->regs[PCA9552_INPUT1] << 8) | s->regs[PCA9552_INPUT0]; +} + +static void pca955x_display_pins_status(PCA955xState *s, + uint16_t previous_pins_status) +{ + PCA955xClass *k = PCA955X_GET_CLASS(s); + uint16_t pins_status, pins_changed; + int i; + + pins_status = pca955x_pins_get_status(s); + pins_changed = previous_pins_status ^ pins_status; + if (!pins_changed) { + return; + } + if (trace_event_get_state_backends(TRACE_PCA955X_GPIO_STATUS)) { + char *buf = g_newa(char, k->pin_count + 1); + + for (i = 0; i < k->pin_count; i++) { + if (extract32(pins_status, i, 1)) { + buf[i] = '*'; + } else { + buf[i] = '.'; + } + } + buf[i] = '\0'; + trace_pca955x_gpio_status(s->description, buf); + } +} + static void pca955x_update_pin_input(PCA955xState *s) { PCA955xClass *k = PCA955X_GET_CLASS(s); @@ -XXX,XX +XXX,XX @@ static uint8_t pca955x_read(PCA955xState *s, uint8_t reg) static void pca955x_write(PCA955xState *s, uint8_t reg, uint8_t data) { + uint16_t pins_status; + switch (reg) { case PCA9552_PSC0: case PCA9552_PWM0: @@ -XXX,XX +XXX,XX @@ static void pca955x_write(PCA955xState *s, uint8_t reg, uint8_t data) case PCA9552_LS1: case PCA9552_LS2: case PCA9552_LS3: + pins_status = pca955x_pins_get_status(s); s->regs[reg] = data; pca955x_update_pin_input(s); + pca955x_display_pins_status(s, pins_status); break; case PCA9552_INPUT0: diff --git a/hw/misc/trace-events b/hw/misc/trace-events index XXXXXXX..XXXXXXX 100644 --- a/hw/misc/trace-events +++ b/hw/misc/trace-events @@ -XXX,XX +XXX,XX @@ via1_adb_poll(uint8_t data, const char *vadbint, int status, int index, int size # grlib_ahb_apb_pnp.c grlib_ahb_pnp_read(uint64_t addr, uint32_t value) "AHB PnP read addr:0x%03"PRIx64" data:0x%08x" grlib_apb_pnp_read(uint64_t addr, uint32_t value) "APB PnP read addr:0x%03"PRIx64" data:0x%08x" + +# pca9552.c +pca955x_gpio_status(const char *description, const char *buf) "%s GPIOs 0-15 [%s]" -- 2.20.1 From: Philippe Mathieu-Daudé <f4bug@amsat.org> We have 2 distinct PCA9552 devices. Set their description to distinguish them when looking at the trace events. Description name taken from: https://github.com/open-power/witherspoon-xml/blob/master/witherspoon.xml Reviewed-by: Cédric Le Goater <clg@kaod.org> Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Reviewed-by: Corey Minyard <cminyard@mvista.com> Reviewed-by: Markus Armbruster <armbru@redhat.com> Tested-by: Cédric Le Goater <clg@kaod.org> Message-id: 20200623072723.6324-8-f4bug@amsat.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- hw/arm/aspeed.c | 13 +++++++++---- 1 file changed, 9 insertions(+), 4 deletions(-) diff --git a/hw/arm/aspeed.c b/hw/arm/aspeed.c index XXXXXXX..XXXXXXX 100644 --- a/hw/arm/aspeed.c +++ b/hw/arm/aspeed.c @@ -XXX,XX +XXX,XX @@ static void witherspoon_bmc_i2c_init(AspeedMachineState *bmc) { AspeedSoCState *soc = &bmc->soc; uint8_t *eeprom_buf = g_malloc0(8 * 1024); + DeviceState *dev; /* Bus 3: TODO bmp280@77 */ /* Bus 3: TODO max31785@52 */ /* Bus 3: TODO dps310@76 */ - i2c_create_slave(aspeed_i2c_get_bus(DEVICE(&soc->i2c), 3), TYPE_PCA9552, - 0x60); + dev = i2c_try_create_slave(TYPE_PCA9552, 0x60); + qdev_prop_set_string(dev, "description", "pca1"); + i2c_realize_and_unref(dev, aspeed_i2c_get_bus(DEVICE(&soc->i2c), 3), + &error_fatal); i2c_create_slave(aspeed_i2c_get_bus(DEVICE(&soc->i2c), 4), "tmp423", 0x4c); i2c_create_slave(aspeed_i2c_get_bus(DEVICE(&soc->i2c), 5), "tmp423", 0x4c); @@ -XXX,XX +XXX,XX @@ static void witherspoon_bmc_i2c_init(AspeedMachineState *bmc) smbus_eeprom_init_one(aspeed_i2c_get_bus(DEVICE(&soc->i2c), 11), 0x51, eeprom_buf); - i2c_create_slave(aspeed_i2c_get_bus(DEVICE(&soc->i2c), 11), TYPE_PCA9552, - 0x60); + dev = i2c_try_create_slave(TYPE_PCA9552, 0x60); + qdev_prop_set_string(dev, "description", "pca0"); + i2c_realize_and_unref(dev, aspeed_i2c_get_bus(DEVICE(&soc->i2c), 11), + &error_fatal); /* Bus 11: TODO ucd90160@64 */ } -- 2.20.1 From: Philippe Mathieu-Daudé <f4bug@amsat.org> Emit a trace event when a GPIO change its state. Example booting obmc-phosphor-image: $ qemu-system-arm -M witherspoon-bmc -trace pca955x_gpio_change 1592690552.687372:pca955x_gpio_change pca1 GPIO id:0 status: 0 -> 1 1592690552.690169:pca955x_gpio_change pca1 GPIO id:1 status: 0 -> 1 1592690552.691673:pca955x_gpio_change pca1 GPIO id:2 status: 0 -> 1 1592690552.696886:pca955x_gpio_change pca1 GPIO id:3 status: 0 -> 1 1592690552.698614:pca955x_gpio_change pca1 GPIO id:13 status: 0 -> 1 1592690552.699833:pca955x_gpio_change pca1 GPIO id:14 status: 0 -> 1 1592690552.700842:pca955x_gpio_change pca1 GPIO id:15 status: 0 -> 1 1592690683.841921:pca955x_gpio_change pca1 GPIO id:14 status: 1 -> 0 1592690683.861660:pca955x_gpio_change pca1 GPIO id:14 status: 0 -> 1 1592690684.371460:pca955x_gpio_change pca1 GPIO id:14 status: 1 -> 0 1592690684.882115:pca955x_gpio_change pca1 GPIO id:14 status: 0 -> 1 1592690685.391411:pca955x_gpio_change pca1 GPIO id:14 status: 1 -> 0 1592690685.901391:pca955x_gpio_change pca1 GPIO id:14 status: 0 -> 1 1592690686.411678:pca955x_gpio_change pca1 GPIO id:14 status: 1 -> 0 1592690686.921279:pca955x_gpio_change pca1 GPIO id:14 status: 0 -> 1 We notice the GPIO #14 (front-power LED) starts to blink. This LED is described in the witherspoon device-tree [*]: front-power { retain-state-shutdown; default-state = "keep"; gpios = <&pca0 14 GPIO_ACTIVE_LOW>; }; [*] https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/arch/arm/boot/dts/aspeed-bmc-opp-witherspoon.dts?id=b1f9be9392f0#n140 Reviewed-by: Cédric Le Goater <clg@kaod.org> Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Tested-by: Cédric Le Goater <clg@kaod.org> Message-id: 20200623072723.6324-9-f4bug@amsat.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- hw/misc/pca9552.c | 15 +++++++++++++++ hw/misc/trace-events | 1 + 2 files changed, 16 insertions(+) diff --git a/hw/misc/pca9552.c b/hw/misc/pca9552.c index XXXXXXX..XXXXXXX 100644 --- a/hw/misc/pca9552.c +++ b/hw/misc/pca9552.c @@ -XXX,XX +XXX,XX @@ static void pca955x_display_pins_status(PCA955xState *s, buf[i] = '\0'; trace_pca955x_gpio_status(s->description, buf); } + if (trace_event_get_state_backends(TRACE_PCA955X_GPIO_CHANGE)) { + for (i = 0; i < k->pin_count; i++) { + if (extract32(pins_changed, i, 1)) { + unsigned new_state = extract32(pins_status, i, 1); + + /* + * We display the state using the PCA logic ("active-high"). + * This is not the state of the LED, which signal might be + * wired "active-low" on the board. + */ + trace_pca955x_gpio_change(s->description, i, + !new_state, new_state); + } + } + } } static void pca955x_update_pin_input(PCA955xState *s) diff --git a/hw/misc/trace-events b/hw/misc/trace-events index XXXXXXX..XXXXXXX 100644 --- a/hw/misc/trace-events +++ b/hw/misc/trace-events @@ -XXX,XX +XXX,XX @@ grlib_apb_pnp_read(uint64_t addr, uint32_t value) "APB PnP read addr:0x%03"PRIx6 # pca9552.c pca955x_gpio_status(const char *description, const char *buf) "%s GPIOs 0-15 [%s]" +pca955x_gpio_change(const char *description, unsigned id, unsigned prev_state, unsigned current_state) "%s GPIO id:%u status: %u -> %u" -- 2.20.1 From: Philippe Mathieu-Daudé <f4bug@amsat.org> The PCA9552 has 16 GPIOs which can be used as input, output or PWM mode. QEMU models the output GPIO with the qemu_irq type. Let the device expose the 16 GPIOs to allow us to later connect LEDs to these outputs. Reviewed-by: Cédric Le Goater <clg@kaod.org> Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Tested-by: Cédric Le Goater <clg@kaod.org> Message-id: 20200623072723.6324-10-f4bug@amsat.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- include/hw/misc/pca9552.h | 1 + hw/misc/pca9552.c | 6 ++++++ 2 files changed, 7 insertions(+) diff --git a/include/hw/misc/pca9552.h b/include/hw/misc/pca9552.h index XXXXXXX..XXXXXXX 100644 --- a/include/hw/misc/pca9552.h +++ b/include/hw/misc/pca9552.h @@ -XXX,XX +XXX,XX @@ typedef struct PCA955xState { uint8_t pointer; uint8_t regs[PCA955X_NR_REGS]; + qemu_irq gpio[PCA955X_PIN_COUNT_MAX]; char *description; /* For debugging purpose only */ } PCA955xState; diff --git a/hw/misc/pca9552.c b/hw/misc/pca9552.c index XXXXXXX..XXXXXXX 100644 --- a/hw/misc/pca9552.c +++ b/hw/misc/pca9552.c @@ -XXX,XX +XXX,XX @@ #include "hw/qdev-properties.h" #include "hw/misc/pca9552.h" #include "hw/misc/pca9552_regs.h" +#include "hw/irq.h" #include "migration/vmstate.h" #include "qapi/error.h" #include "qapi/visitor.h" @@ -XXX,XX +XXX,XX @@ static void pca955x_update_pin_input(PCA955xState *s) switch (config) { case PCA9552_LED_ON: + qemu_set_irq(s->gpio[i], 1); s->regs[input_reg] |= 1 << input_shift; break; case PCA9552_LED_OFF: + qemu_set_irq(s->gpio[i], 0); s->regs[input_reg] &= ~(1 << input_shift); break; case PCA9552_LED_PWM0: @@ -XXX,XX +XXX,XX @@ static void pca955x_initfn(Object *obj) static void pca955x_realize(DeviceState *dev, Error **errp) { + PCA955xClass *k = PCA955X_GET_CLASS(dev); PCA955xState *s = PCA955X(dev); if (!s->description) { s->description = g_strdup("pca-unspecified"); } + + qdev_init_gpio_out(dev, s->gpio, k->pin_count); } static Property pca955x_properties[] = { -- 2.20.1 From: Richard Henderson <richard.henderson@linaro.org> Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200626033144.790098-2-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/cpu.h | 10 ++++++++++ 1 file changed, 10 insertions(+) diff --git a/target/arm/cpu.h b/target/arm/cpu.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/cpu.h +++ b/target/arm/cpu.h @@ -XXX,XX +XXX,XX @@ static inline bool isar_feature_aa64_bti(const ARMISARegisters *id) return FIELD_EX64(id->id_aa64pfr1, ID_AA64PFR1, BT) != 0; } +static inline bool isar_feature_aa64_mte_insn_reg(const ARMISARegisters *id) +{ + return FIELD_EX64(id->id_aa64pfr1, ID_AA64PFR1, MTE) != 0; +} + +static inline bool isar_feature_aa64_mte(const ARMISARegisters *id) +{ + return FIELD_EX64(id->id_aa64pfr1, ID_AA64PFR1, MTE) >= 2; +} + static inline bool isar_feature_aa64_pmu_8_1(const ARMISARegisters *id) { return FIELD_EX64(id->id_aa64dfr0, ID_AA64DFR0, PMUVER) >= 4 && -- 2.20.1 From: Richard Henderson <richard.henderson@linaro.org> Protect reads of aa64 id registers with ARM_CP_STATE_AA64. Use this as a simpler test than arm_el_is_aa64, since EL3 cannot change mode. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200626033144.790098-3-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/helper.c | 15 ++++++++------- 1 file changed, 8 insertions(+), 7 deletions(-) diff --git a/target/arm/helper.c b/target/arm/helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper.c +++ b/target/arm/helper.c @@ -XXX,XX +XXX,XX @@ static void scr_write(CPUARMState *env, const ARMCPRegInfo *ri, uint64_t value) uint32_t valid_mask = 0x3fff; ARMCPU *cpu = env_archcpu(env); - if (arm_el_is_aa64(env, 3)) { + if (ri->state == ARM_CP_STATE_AA64) { value |= SCR_FW | SCR_AW; /* these two bits are RES1. */ valid_mask &= ~SCR_NET; + + if (cpu_isar_feature(aa64_lor, cpu)) { + valid_mask |= SCR_TLOR; + } + if (cpu_isar_feature(aa64_pauth, cpu)) { + valid_mask |= SCR_API | SCR_APK; + } } else { valid_mask &= ~(SCR_RW | SCR_ST); } @@ -XXX,XX +XXX,XX @@ static void scr_write(CPUARMState *env, const ARMCPRegInfo *ri, uint64_t value) valid_mask &= ~SCR_SMD; } } - if (cpu_isar_feature(aa64_lor, cpu)) { - valid_mask |= SCR_TLOR; - } - if (cpu_isar_feature(aa64_pauth, cpu)) { - valid_mask |= SCR_API | SCR_APK; - } /* Clear all-context RES0 bits. */ value &= valid_mask; -- 2.20.1 From: Richard Henderson <richard.henderson@linaro.org> This does not attempt to rectify all of the res0 bits, but does clear the mte bits when not enabled. Since there is no high-part mapping of SCTLR, aa32 mode cannot write to these bits. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200626033144.790098-4-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/helper.c | 23 +++++++++++++++++------ 1 file changed, 17 insertions(+), 6 deletions(-) diff --git a/target/arm/helper.c b/target/arm/helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper.c +++ b/target/arm/helper.c @@ -XXX,XX +XXX,XX @@ static void sctlr_write(CPUARMState *env, const ARMCPRegInfo *ri, { ARMCPU *cpu = env_archcpu(env); + if (arm_feature(env, ARM_FEATURE_PMSA) && !cpu->has_mpu) { + /* M bit is RAZ/WI for PMSA with no MPU implemented */ + value &= ~SCTLR_M; + } + + /* ??? Lots of these bits are not implemented. */ + + if (ri->state == ARM_CP_STATE_AA64 && !cpu_isar_feature(aa64_mte, cpu)) { + if (ri->opc1 == 6) { /* SCTLR_EL3 */ + value &= ~(SCTLR_ITFSB | SCTLR_TCF | SCTLR_ATA); + } else { + value &= ~(SCTLR_ITFSB | SCTLR_TCF0 | SCTLR_TCF | + SCTLR_ATA0 | SCTLR_ATA); + } + } + if (raw_read(env, ri) == value) { /* Skip the TLB flush if nothing actually changed; Linux likes * to do a lot of pointless SCTLR writes. @@ -XXX,XX +XXX,XX @@ static void sctlr_write(CPUARMState *env, const ARMCPRegInfo *ri, return; } - if (arm_feature(env, ARM_FEATURE_PMSA) && !cpu->has_mpu) { - /* M bit is RAZ/WI for PMSA with no MPU implemented */ - value &= ~SCTLR_M; - } - raw_write(env, ri, value); - /* ??? Lots of these bits are not implemented. */ + /* This may enable/disable the MMU, so do a TLB flush. */ tlb_flush(CPU(cpu)); -- 2.20.1 From: Richard Henderson <richard.henderson@linaro.org> Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200626033144.790098-5-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/helper.c | 14 +++++++++++--- 1 file changed, 11 insertions(+), 3 deletions(-) diff --git a/target/arm/helper.c b/target/arm/helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper.c +++ b/target/arm/helper.c @@ -XXX,XX +XXX,XX @@ static void scr_write(CPUARMState *env, const ARMCPRegInfo *ri, uint64_t value) if (cpu_isar_feature(aa64_pauth, cpu)) { valid_mask |= SCR_API | SCR_APK; } + if (cpu_isar_feature(aa64_mte, cpu)) { + valid_mask |= SCR_ATA; + } } else { valid_mask &= ~(SCR_RW | SCR_ST); } @@ -XXX,XX +XXX,XX @@ static void do_hcr_write(CPUARMState *env, uint64_t value, uint64_t valid_mask) if (cpu_isar_feature(aa64_pauth, cpu)) { valid_mask |= HCR_API | HCR_APK; } + if (cpu_isar_feature(aa64_mte, cpu)) { + valid_mask |= HCR_ATA | HCR_DCT | HCR_TID5; + } } /* Clear RES0 bits. */ value &= valid_mask; - /* These bits change the MMU setup: + /* + * These bits change the MMU setup: * HCR_VM enables stage 2 translation * HCR_PTW forbids certain page-table setups - * HCR_DC Disables stage1 and enables stage2 translation + * HCR_DC disables stage1 and enables stage2 translation + * HCR_DCT enables tagging on (disabled) stage1 translation */ - if ((env->cp15.hcr_el2 ^ value) & (HCR_VM | HCR_PTW | HCR_DC)) { + if ((env->cp15.hcr_el2 ^ value) & (HCR_VM | HCR_PTW | HCR_DC | HCR_DCT)) { tlb_flush(CPU(cpu)); } env->cp15.hcr_el2 = value; -- 2.20.1 From: Richard Henderson <richard.henderson@linaro.org> Emphasize that the is_jmp option exits to the main loop. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200626033144.790098-6-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/translate.h | 14 ++++++++------ target/arm/translate-a64.c | 8 ++++---- target/arm/translate-vfp.inc.c | 4 ++-- target/arm/translate.c | 12 ++++++------ 4 files changed, 20 insertions(+), 18 deletions(-) diff --git a/target/arm/translate.h b/target/arm/translate.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate.h +++ b/target/arm/translate.h @@ -XXX,XX +XXX,XX @@ static inline void disas_set_insn_syndrome(DisasContext *s, uint32_t syn) /* is_jmp field values */ #define DISAS_JUMP DISAS_TARGET_0 /* only pc was modified dynamically */ -#define DISAS_UPDATE DISAS_TARGET_1 /* cpu state was modified dynamically */ +/* CPU state was modified dynamically; exit to main loop for interrupts. */ +#define DISAS_UPDATE_EXIT DISAS_TARGET_1 /* These instructions trap after executing, so the A32/T32 decoder must * defer them until after the conditional execution state has been updated. * WFI also needs special handling when single-stepping. @@ -XXX,XX +XXX,XX @@ static inline void disas_set_insn_syndrome(DisasContext *s, uint32_t syn) * custom end-of-TB code) */ #define DISAS_BX_EXCRET DISAS_TARGET_8 -/* For instructions which want an immediate exit to the main loop, - * as opposed to attempting to use lookup_and_goto_ptr. Unlike - * DISAS_UPDATE this doesn't write the PC on exiting the translation - * loop so you need to ensure something (gen_a64_set_pc_im or runtime - * helper) has done so before we reach return from cpu_tb_exec. +/* + * For instructions which want an immediate exit to the main loop, as opposed + * to attempting to use lookup_and_goto_ptr. Unlike DISAS_UPDATE_EXIT, this + * doesn't write the PC on exiting the translation loop so you need to ensure + * something (gen_a64_set_pc_im or runtime helper) has done so before we reach + * return from cpu_tb_exec. */ #define DISAS_EXIT DISAS_TARGET_9 diff --git a/target/arm/translate-a64.c b/target/arm/translate-a64.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-a64.c +++ b/target/arm/translate-a64.c @@ -XXX,XX +XXX,XX @@ static void handle_msr_i(DisasContext *s, uint32_t insn, gen_helper_msr_i_daifclear(cpu_env, t1); tcg_temp_free_i32(t1); /* For DAIFClear, exit the cpu loop to re-evaluate pending IRQs. */ - s->base.is_jmp = DISAS_UPDATE; + s->base.is_jmp = DISAS_UPDATE_EXIT; break; default: @@ -XXX,XX +XXX,XX @@ static void handle_sys(DisasContext *s, uint32_t insn, bool isread, if ((tb_cflags(s->base.tb) & CF_USE_ICOUNT) && (ri->type & ARM_CP_IO)) { /* I/O operations must end the TB here (whether read or write) */ - s->base.is_jmp = DISAS_UPDATE; + s->base.is_jmp = DISAS_UPDATE_EXIT; } if (!isread && !(ri->type & ARM_CP_SUPPRESS_TB_END)) { /* @@ -XXX,XX +XXX,XX @@ static void handle_sys(DisasContext *s, uint32_t insn, bool isread, * but allow this to be suppressed by the register definition * (usually only necessary to work around guest bugs). */ - s->base.is_jmp = DISAS_UPDATE; + s->base.is_jmp = DISAS_UPDATE_EXIT; } } @@ -XXX,XX +XXX,XX @@ static void aarch64_tr_tb_stop(DisasContextBase *dcbase, CPUState *cpu) gen_goto_tb(dc, 1, dc->base.pc_next); break; default: - case DISAS_UPDATE: + case DISAS_UPDATE_EXIT: gen_a64_set_pc_im(dc->base.pc_next); /* fall through */ case DISAS_EXIT: diff --git a/target/arm/translate-vfp.inc.c b/target/arm/translate-vfp.inc.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-vfp.inc.c +++ b/target/arm/translate-vfp.inc.c @@ -XXX,XX +XXX,XX @@ static bool full_vfp_access_check(DisasContext *s, bool ignore_vfp_enabled) * this to be the last insn in the TB). */ if (tb_cflags(s->base.tb) & CF_USE_ICOUNT) { - s->base.is_jmp = DISAS_UPDATE; + s->base.is_jmp = DISAS_UPDATE_EXIT; gen_io_start(); } gen_helper_v7m_preserve_fp_state(cpu_env); @@ -XXX,XX +XXX,XX @@ static bool trans_VLLDM_VLSTM(DisasContext *s, arg_VLLDM_VLSTM *a) tcg_temp_free_i32(fptr); /* End the TB, because we have updated FP control bits */ - s->base.is_jmp = DISAS_UPDATE; + s->base.is_jmp = DISAS_UPDATE_EXIT; return true; } diff --git a/target/arm/translate.c b/target/arm/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate.c +++ b/target/arm/translate.c @@ -XXX,XX +XXX,XX @@ static void gen_msr_banked(DisasContext *s, int r, int sysm, int rn) tcg_temp_free_i32(tcg_tgtmode); tcg_temp_free_i32(tcg_regno); tcg_temp_free_i32(tcg_reg); - s->base.is_jmp = DISAS_UPDATE; + s->base.is_jmp = DISAS_UPDATE_EXIT; } static void gen_mrs_banked(DisasContext *s, int r, int sysm, int rn) @@ -XXX,XX +XXX,XX @@ static void gen_mrs_banked(DisasContext *s, int r, int sysm, int rn) tcg_temp_free_i32(tcg_tgtmode); tcg_temp_free_i32(tcg_regno); store_reg(s, rn, tcg_reg); - s->base.is_jmp = DISAS_UPDATE; + s->base.is_jmp = DISAS_UPDATE_EXIT; } /* Store value to PC as for an exception return (ie don't @@ -XXX,XX +XXX,XX @@ static void gen_srs(DisasContext *s, tcg_temp_free_i32(tmp); } tcg_temp_free_i32(addr); - s->base.is_jmp = DISAS_UPDATE; + s->base.is_jmp = DISAS_UPDATE_EXIT; } /* Generate a label used for skipping this instruction */ @@ -XXX,XX +XXX,XX @@ static bool trans_SETEND(DisasContext *s, arg_SETEND *a) } if (a->E != (s->be_data == MO_BE)) { gen_helper_setend(cpu_env); - s->base.is_jmp = DISAS_UPDATE; + s->base.is_jmp = DISAS_UPDATE_EXIT; } return true; } @@ -XXX,XX +XXX,XX @@ static void arm_tr_tb_stop(DisasContextBase *dcbase, CPUState *cpu) break; case DISAS_NEXT: case DISAS_TOO_MANY: - case DISAS_UPDATE: + case DISAS_UPDATE_EXIT: gen_set_pc_im(dc, dc->base.pc_next); /* fall through */ default: @@ -XXX,XX +XXX,XX @@ static void arm_tr_tb_stop(DisasContextBase *dcbase, CPUState *cpu) case DISAS_JUMP: gen_goto_ptr(); break; - case DISAS_UPDATE: + case DISAS_UPDATE_EXIT: gen_set_pc_im(dc, dc->base.pc_next); /* fall through */ default: -- 2.20.1 From: Richard Henderson <richard.henderson@linaro.org> Add an option that writes back the PC, like DISAS_UPDATE_EXIT, but does not exit back to the main loop. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200626033144.790098-7-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/translate.h | 2 ++ target/arm/translate-a64.c | 3 +++ target/arm/translate.c | 4 ++++ 3 files changed, 9 insertions(+) diff --git a/target/arm/translate.h b/target/arm/translate.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate.h +++ b/target/arm/translate.h @@ -XXX,XX +XXX,XX @@ static inline void disas_set_insn_syndrome(DisasContext *s, uint32_t syn) * return from cpu_tb_exec. */ #define DISAS_EXIT DISAS_TARGET_9 +/* CPU state was modified dynamically; no need to exit, but do not chain. */ +#define DISAS_UPDATE_NOCHAIN DISAS_TARGET_10 #ifdef TARGET_AARCH64 void a64_translate_init(void); diff --git a/target/arm/translate-a64.c b/target/arm/translate-a64.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-a64.c +++ b/target/arm/translate-a64.c @@ -XXX,XX +XXX,XX @@ static void aarch64_tr_tb_stop(DisasContextBase *dcbase, CPUState *cpu) case DISAS_EXIT: tcg_gen_exit_tb(NULL, 0); break; + case DISAS_UPDATE_NOCHAIN: + gen_a64_set_pc_im(dc->base.pc_next); + /* fall through */ case DISAS_JUMP: tcg_gen_lookup_and_goto_ptr(); break; diff --git a/target/arm/translate.c b/target/arm/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate.c +++ b/target/arm/translate.c @@ -XXX,XX +XXX,XX @@ static void arm_tr_tb_stop(DisasContextBase *dcbase, CPUState *cpu) case DISAS_NEXT: case DISAS_TOO_MANY: case DISAS_UPDATE_EXIT: + case DISAS_UPDATE_NOCHAIN: gen_set_pc_im(dc, dc->base.pc_next); /* fall through */ default: @@ -XXX,XX +XXX,XX @@ static void arm_tr_tb_stop(DisasContextBase *dcbase, CPUState *cpu) case DISAS_TOO_MANY: gen_goto_tb(dc, 1, dc->base.pc_next); break; + case DISAS_UPDATE_NOCHAIN: + gen_set_pc_im(dc, dc->base.pc_next); + /* fall through */ case DISAS_JUMP: gen_goto_ptr(); break; -- 2.20.1 From: Richard Henderson <richard.henderson@linaro.org> This is TFSRE0_EL1, TFSR_EL1, TFSR_EL2, TFSR_EL3, RGSR_EL1, GCR_EL1, GMID_EL1, and PSTATE.TCO. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200626033144.790098-8-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/cpu.h | 4 ++ target/arm/internals.h | 9 ++++ target/arm/helper.c | 94 ++++++++++++++++++++++++++++++++++++++ target/arm/translate-a64.c | 21 +++++++++ 4 files changed, 128 insertions(+) diff --git a/target/arm/cpu.h b/target/arm/cpu.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/cpu.h +++ b/target/arm/cpu.h @@ -XXX,XX +XXX,XX @@ typedef struct CPUARMState { uint64_t pmccfiltr_el0; /* Performance Monitor Filter Register */ uint64_t vpidr_el2; /* Virtualization Processor ID Register */ uint64_t vmpidr_el2; /* Virtualization Multiprocessor ID Register */ + uint64_t tfsr_el[4]; /* tfsre0_el1 is index 0. */ + uint64_t gcr_el1; + uint64_t rgsr_el1; } cp15; struct { @@ -XXX,XX +XXX,XX @@ void pmu_init(ARMCPU *cpu); #define PSTATE_SS (1U << 21) #define PSTATE_PAN (1U << 22) #define PSTATE_UAO (1U << 23) +#define PSTATE_TCO (1U << 25) #define PSTATE_V (1U << 28) #define PSTATE_C (1U << 29) #define PSTATE_Z (1U << 30) diff --git a/target/arm/internals.h b/target/arm/internals.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/internals.h +++ b/target/arm/internals.h @@ -XXX,XX +XXX,XX @@ static inline uint32_t aarch64_pstate_valid_mask(const ARMISARegisters *id) if (isar_feature_aa64_uao(id)) { valid |= PSTATE_UAO; } + if (isar_feature_aa64_mte(id)) { + valid |= PSTATE_TCO; + } return valid; } @@ -XXX,XX +XXX,XX @@ void arm_log_exception(int idx); #endif /* !CONFIG_USER_ONLY */ +/* + * The log2 of the words in the tag block, for GMID_EL1.BS. + * The is the maximum, 256 bytes, which manipulates 64-bits of tags. + */ +#define GMID_EL1_BS 6 + #endif diff --git a/target/arm/helper.c b/target/arm/helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper.c +++ b/target/arm/helper.c @@ -XXX,XX +XXX,XX @@ static void define_arm_vh_e2h_redirects_aliases(ARMCPU *cpu) { K(3, 0, 1, 2, 0), K(3, 4, 1, 2, 0), K(3, 5, 1, 2, 0), "ZCR_EL1", "ZCR_EL2", "ZCR_EL12", isar_feature_aa64_sve }, + { K(3, 0, 5, 6, 0), K(3, 4, 5, 6, 0), K(3, 5, 5, 6, 0), + "TFSR_EL1", "TFSR_EL2", "TFSR_EL12", isar_feature_aa64_mte }, + /* TODO: ARMv8.2-SPE -- PMSCR_EL2 */ /* TODO: ARMv8.4-Trace -- TRFCR_EL2 */ }; @@ -XXX,XX +XXX,XX @@ static const ARMCPRegInfo dcpodp_reg[] = { }; #endif /*CONFIG_USER_ONLY*/ +static CPAccessResult access_aa64_tid5(CPUARMState *env, const ARMCPRegInfo *ri, + bool isread) +{ + if ((arm_current_el(env) < 2) && (arm_hcr_el2_eff(env) & HCR_TID5)) { + return CP_ACCESS_TRAP_EL2; + } + + return CP_ACCESS_OK; +} + +static CPAccessResult access_mte(CPUARMState *env, const ARMCPRegInfo *ri, + bool isread) +{ + int el = arm_current_el(env); + + if (el < 2 && + arm_feature(env, ARM_FEATURE_EL2) && + !(arm_hcr_el2_eff(env) & HCR_ATA)) { + return CP_ACCESS_TRAP_EL2; + } + if (el < 3 && + arm_feature(env, ARM_FEATURE_EL3) && + !(env->cp15.scr_el3 & SCR_ATA)) { + return CP_ACCESS_TRAP_EL3; + } + return CP_ACCESS_OK; +} + +static uint64_t tco_read(CPUARMState *env, const ARMCPRegInfo *ri) +{ + return env->pstate & PSTATE_TCO; +} + +static void tco_write(CPUARMState *env, const ARMCPRegInfo *ri, uint64_t val) +{ + env->pstate = (env->pstate & ~PSTATE_TCO) | (val & PSTATE_TCO); +} + +static const ARMCPRegInfo mte_reginfo[] = { + { .name = "TFSRE0_EL1", .state = ARM_CP_STATE_AA64, + .opc0 = 3, .opc1 = 0, .crn = 5, .crm = 6, .opc2 = 1, + .access = PL1_RW, .accessfn = access_mte, + .fieldoffset = offsetof(CPUARMState, cp15.tfsr_el[0]) }, + { .name = "TFSR_EL1", .state = ARM_CP_STATE_AA64, + .opc0 = 3, .opc1 = 0, .crn = 5, .crm = 6, .opc2 = 0, + .access = PL1_RW, .accessfn = access_mte, + .fieldoffset = offsetof(CPUARMState, cp15.tfsr_el[1]) }, + { .name = "TFSR_EL2", .state = ARM_CP_STATE_AA64, + .opc0 = 3, .opc1 = 4, .crn = 5, .crm = 6, .opc2 = 0, + .access = PL2_RW, .accessfn = access_mte, + .fieldoffset = offsetof(CPUARMState, cp15.tfsr_el[2]) }, + { .name = "TFSR_EL3", .state = ARM_CP_STATE_AA64, + .opc0 = 3, .opc1 = 6, .crn = 5, .crm = 6, .opc2 = 0, + .access = PL3_RW, + .fieldoffset = offsetof(CPUARMState, cp15.tfsr_el[3]) }, + { .name = "RGSR_EL1", .state = ARM_CP_STATE_AA64, + .opc0 = 3, .opc1 = 0, .crn = 1, .crm = 0, .opc2 = 5, + .access = PL1_RW, .accessfn = access_mte, + .fieldoffset = offsetof(CPUARMState, cp15.rgsr_el1) }, + { .name = "GCR_EL1", .state = ARM_CP_STATE_AA64, + .opc0 = 3, .opc1 = 0, .crn = 1, .crm = 0, .opc2 = 6, + .access = PL1_RW, .accessfn = access_mte, + .fieldoffset = offsetof(CPUARMState, cp15.gcr_el1) }, + { .name = "GMID_EL1", .state = ARM_CP_STATE_AA64, + .opc0 = 3, .opc1 = 1, .crn = 0, .crm = 0, .opc2 = 4, + .access = PL1_R, .accessfn = access_aa64_tid5, + .type = ARM_CP_CONST, .resetvalue = GMID_EL1_BS }, + { .name = "TCO", .state = ARM_CP_STATE_AA64, + .opc0 = 3, .opc1 = 3, .crn = 4, .crm = 2, .opc2 = 7, + .type = ARM_CP_NO_RAW, + .access = PL0_RW, .readfn = tco_read, .writefn = tco_write }, + REGINFO_SENTINEL +}; + +static const ARMCPRegInfo mte_tco_ro_reginfo[] = { + { .name = "TCO", .state = ARM_CP_STATE_AA64, + .opc0 = 3, .opc1 = 3, .crn = 4, .crm = 2, .opc2 = 7, + .type = ARM_CP_CONST, .access = PL0_RW, }, + REGINFO_SENTINEL +}; #endif static CPAccessResult access_predinv(CPUARMState *env, const ARMCPRegInfo *ri, @@ -XXX,XX +XXX,XX @@ void register_cp_regs_for_features(ARMCPU *cpu) } } #endif /*CONFIG_USER_ONLY*/ + + /* + * If full MTE is enabled, add all of the system registers. + * If only "instructions available at EL0" are enabled, + * then define only a RAZ/WI version of PSTATE.TCO. + */ + if (cpu_isar_feature(aa64_mte, cpu)) { + define_arm_cp_regs(cpu, mte_reginfo); + } else if (cpu_isar_feature(aa64_mte_insn_reg, cpu)) { + define_arm_cp_regs(cpu, mte_tco_ro_reginfo); + } #endif if (cpu_isar_feature(any_predinv, cpu)) { diff --git a/target/arm/translate-a64.c b/target/arm/translate-a64.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-a64.c +++ b/target/arm/translate-a64.c @@ -XXX,XX +XXX,XX @@ static void handle_msr_i(DisasContext *s, uint32_t insn, s->base.is_jmp = DISAS_UPDATE_EXIT; break; + case 0x1c: /* TCO */ + if (dc_isar_feature(aa64_mte, s)) { + /* Full MTE is enabled -- set the TCO bit as directed. */ + if (crm & 1) { + set_pstate_bits(PSTATE_TCO); + } else { + clear_pstate_bits(PSTATE_TCO); + } + t1 = tcg_const_i32(s->current_el); + gen_helper_rebuild_hflags_a64(cpu_env, t1); + tcg_temp_free_i32(t1); + /* Many factors, including TCO, go into MTE_ACTIVE. */ + s->base.is_jmp = DISAS_UPDATE_NOCHAIN; + } else if (dc_isar_feature(aa64_mte_insn_reg, s)) { + /* Only "instructions accessible at EL0" -- PSTATE.TCO is WI. */ + s->base.is_jmp = DISAS_NEXT; + } else { + goto do_unallocated; + } + break; + default: do_unallocated: unallocated_encoding(s); -- 2.20.1 From: Richard Henderson <richard.henderson@linaro.org> Cache the composite ATA setting. Cache when MTE is fully enabled, i.e. access to tags are enabled and tag checks affect the PE. Do this for both the normal context and the UNPRIV context. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200626033144.790098-9-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/cpu.h | 12 ++++++++---- target/arm/internals.h | 18 +++++++++++++++++ target/arm/translate.h | 5 +++++ target/arm/helper.c | 40 ++++++++++++++++++++++++++++++++++++++ target/arm/translate-a64.c | 4 ++++ 5 files changed, 75 insertions(+), 4 deletions(-) diff --git a/target/arm/cpu.h b/target/arm/cpu.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/cpu.h +++ b/target/arm/cpu.h @@ -XXX,XX +XXX,XX @@ typedef ARMCPU ArchCPU; * | | | TBFLAG_A32 | | * | | +-----+----------+ TBFLAG_AM32 | * | TBFLAG_ANY | |TBFLAG_M32| | - * | | +-+----------+--------------| - * | | | TBFLAG_A64 | - * +--------------+---------+---------------------------+ - * 31 20 15 0 + * | +-----------+----------+--------------| + * | | TBFLAG_A64 | + * +--------------+-------------------------------------+ + * 31 20 0 * * Unless otherwise noted, these bits are cached in env->hflags. */ @@ -XXX,XX +XXX,XX @@ FIELD(TBFLAG_A64, BT, 9, 1) FIELD(TBFLAG_A64, BTYPE, 10, 2) /* Not cached. */ FIELD(TBFLAG_A64, TBID, 12, 2) FIELD(TBFLAG_A64, UNPRIV, 14, 1) +FIELD(TBFLAG_A64, ATA, 15, 1) +FIELD(TBFLAG_A64, TCMA, 16, 2) +FIELD(TBFLAG_A64, MTE_ACTIVE, 18, 1) +FIELD(TBFLAG_A64, MTE0_ACTIVE, 19, 1) /** * cpu_mmu_index: diff --git a/target/arm/internals.h b/target/arm/internals.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/internals.h +++ b/target/arm/internals.h @@ -XXX,XX +XXX,XX @@ static inline int exception_target_el(CPUARMState *env) return target_el; } +/* Determine if allocation tags are available. */ +static inline bool allocation_tag_access_enabled(CPUARMState *env, int el, + uint64_t sctlr) +{ + if (el < 3 + && arm_feature(env, ARM_FEATURE_EL3) + && !(env->cp15.scr_el3 & SCR_ATA)) { + return false; + } + if (el < 2 + && arm_feature(env, ARM_FEATURE_EL2) + && !(arm_hcr_el2_eff(env) & HCR_ATA)) { + return false; + } + sctlr &= (el == 0 ? SCTLR_ATA0 : SCTLR_ATA); + return sctlr != 0; +} + #ifndef CONFIG_USER_ONLY /* Security attributes for an address, as returned by v8m_security_lookup. */ diff --git a/target/arm/translate.h b/target/arm/translate.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate.h +++ b/target/arm/translate.h @@ -XXX,XX +XXX,XX @@ typedef struct DisasContext { ARMMMUIdx mmu_idx; /* MMU index to use for normal loads/stores */ uint8_t tbii; /* TBI1|TBI0 for insns */ uint8_t tbid; /* TBI1|TBI0 for data */ + uint8_t tcma; /* TCMA1|TCMA0 for MTE */ bool ns; /* Use non-secure CPREG bank on access */ int fp_excp_el; /* FP exception EL or 0 if enabled */ int sve_excp_el; /* SVE exception EL or 0 if enabled */ @@ -XXX,XX +XXX,XX @@ typedef struct DisasContext { bool unpriv; /* True if v8.3-PAuth is active. */ bool pauth_active; + /* True if v8.5-MTE access to tags is enabled. */ + bool ata; + /* True if v8.5-MTE tag checks affect the PE; index with is_unpriv. */ + bool mte_active[2]; /* True with v8.5-BTI and SCTLR_ELx.BT* set. */ bool bt; /* True if any CP15 access is trapped by HSTR_EL2 */ diff --git a/target/arm/helper.c b/target/arm/helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper.c +++ b/target/arm/helper.c @@ -XXX,XX +XXX,XX @@ static int aa64_va_parameter_tbid(uint64_t tcr, ARMMMUIdx mmu_idx) } } +static int aa64_va_parameter_tcma(uint64_t tcr, ARMMMUIdx mmu_idx) +{ + if (regime_has_2_ranges(mmu_idx)) { + return extract64(tcr, 57, 2); + } else { + /* Replicate the single TCMA bit so we always have 2 bits. */ + return extract32(tcr, 30, 1) * 3; + } +} + ARMVAParameters aa64_va_parameters(CPUARMState *env, uint64_t va, ARMMMUIdx mmu_idx, bool data) { @@ -XXX,XX +XXX,XX @@ static uint32_t rebuild_hflags_a64(CPUARMState *env, int el, int fp_el, } } + if (cpu_isar_feature(aa64_mte, env_archcpu(env))) { + /* + * Set MTE_ACTIVE if any access may be Checked, and leave clear + * if all accesses must be Unchecked: + * 1) If no TBI, then there are no tags in the address to check, + * 2) If Tag Check Override, then all accesses are Unchecked, + * 3) If Tag Check Fail == 0, then Checked access have no effect, + * 4) If no Allocation Tag Access, then all accesses are Unchecked. + */ + if (allocation_tag_access_enabled(env, el, sctlr)) { + flags = FIELD_DP32(flags, TBFLAG_A64, ATA, 1); + if (tbid + && !(env->pstate & PSTATE_TCO) + && (sctlr & (el == 0 ? SCTLR_TCF0 : SCTLR_TCF))) { + flags = FIELD_DP32(flags, TBFLAG_A64, MTE_ACTIVE, 1); + } + } + /* And again for unprivileged accesses, if required. */ + if (FIELD_EX32(flags, TBFLAG_A64, UNPRIV) + && tbid + && !(env->pstate & PSTATE_TCO) + && (sctlr & SCTLR_TCF0) + && allocation_tag_access_enabled(env, 0, sctlr)) { + flags = FIELD_DP32(flags, TBFLAG_A64, MTE0_ACTIVE, 1); + } + /* Cache TCMA as well as TBI. */ + flags = FIELD_DP32(flags, TBFLAG_A64, TCMA, + aa64_va_parameter_tcma(tcr, mmu_idx)); + } + return rebuild_hflags_common(env, fp_el, mmu_idx, flags); } diff --git a/target/arm/translate-a64.c b/target/arm/translate-a64.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-a64.c +++ b/target/arm/translate-a64.c @@ -XXX,XX +XXX,XX @@ static void aarch64_tr_init_disas_context(DisasContextBase *dcbase, dc->mmu_idx = core_to_aa64_mmu_idx(core_mmu_idx); dc->tbii = FIELD_EX32(tb_flags, TBFLAG_A64, TBII); dc->tbid = FIELD_EX32(tb_flags, TBFLAG_A64, TBID); + dc->tcma = FIELD_EX32(tb_flags, TBFLAG_A64, TCMA); dc->current_el = arm_mmu_idx_to_el(dc->mmu_idx); #if !defined(CONFIG_USER_ONLY) dc->user = (dc->current_el == 0); @@ -XXX,XX +XXX,XX @@ static void aarch64_tr_init_disas_context(DisasContextBase *dcbase, dc->bt = FIELD_EX32(tb_flags, TBFLAG_A64, BT); dc->btype = FIELD_EX32(tb_flags, TBFLAG_A64, BTYPE); dc->unpriv = FIELD_EX32(tb_flags, TBFLAG_A64, UNPRIV); + dc->ata = FIELD_EX32(tb_flags, TBFLAG_A64, ATA); + dc->mte_active[0] = FIELD_EX32(tb_flags, TBFLAG_A64, MTE_ACTIVE); + dc->mte_active[1] = FIELD_EX32(tb_flags, TBFLAG_A64, MTE0_ACTIVE); dc->vec_len = 0; dc->vec_stride = 0; dc->cp_regs = arm_cpu->cp_regs; -- 2.20.1 From: Richard Henderson <richard.henderson@linaro.org> Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200626033144.790098-10-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/helper-a64.h | 2 ++ target/arm/internals.h | 5 +++ target/arm/mte_helper.c | 72 ++++++++++++++++++++++++++++++++++++++ target/arm/translate-a64.c | 18 ++++++++++ target/arm/Makefile.objs | 1 + 5 files changed, 98 insertions(+) create mode 100644 target/arm/mte_helper.c diff --git a/target/arm/helper-a64.h b/target/arm/helper-a64.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper-a64.h +++ b/target/arm/helper-a64.h @@ -XXX,XX +XXX,XX @@ DEF_HELPER_FLAGS_3(autda, TCG_CALL_NO_WG, i64, env, i64, i64) DEF_HELPER_FLAGS_3(autdb, TCG_CALL_NO_WG, i64, env, i64, i64) DEF_HELPER_FLAGS_2(xpaci, TCG_CALL_NO_RWG_SE, i64, env, i64) DEF_HELPER_FLAGS_2(xpacd, TCG_CALL_NO_RWG_SE, i64, env, i64) + +DEF_HELPER_FLAGS_3(irg, TCG_CALL_NO_RWG, i64, env, i64, i64) diff --git a/target/arm/internals.h b/target/arm/internals.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/internals.h +++ b/target/arm/internals.h @@ -XXX,XX +XXX,XX @@ void arm_log_exception(int idx); */ #define GMID_EL1_BS 6 +static inline uint64_t address_with_allocation_tag(uint64_t ptr, int rtag) +{ + return deposit64(ptr, 56, 4, rtag); +} + #endif diff --git a/target/arm/mte_helper.c b/target/arm/mte_helper.c new file mode 100644 index XXXXXXX..XXXXXXX --- /dev/null +++ b/target/arm/mte_helper.c @@ -XXX,XX +XXX,XX @@ +/* + * ARM v8.5-MemTag Operations + * + * Copyright (c) 2020 Linaro, Ltd. + * + * This library is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * This library is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with this library; if not, see <http://www.gnu.org/licenses/>. + */ + +#include "qemu/osdep.h" +#include "cpu.h" +#include "internals.h" +#include "exec/exec-all.h" +#include "exec/cpu_ldst.h" +#include "exec/helper-proto.h" + + +static int choose_nonexcluded_tag(int tag, int offset, uint16_t exclude) +{ + if (exclude == 0xffff) { + return 0; + } + if (offset == 0) { + while (exclude & (1 << tag)) { + tag = (tag + 1) & 15; + } + } else { + do { + do { + tag = (tag + 1) & 15; + } while (exclude & (1 << tag)); + } while (--offset > 0); + } + return tag; +} + +uint64_t HELPER(irg)(CPUARMState *env, uint64_t rn, uint64_t rm) +{ + int rtag; + + /* + * Our IMPDEF choice for GCR_EL1.RRND==1 is to behave as if + * GCR_EL1.RRND==0, always producing deterministic results. + */ + uint16_t exclude = extract32(rm | env->cp15.gcr_el1, 0, 16); + int start = extract32(env->cp15.rgsr_el1, 0, 4); + int seed = extract32(env->cp15.rgsr_el1, 8, 16); + int offset, i; + + /* RandomTag */ + for (i = offset = 0; i < 4; ++i) { + /* NextRandomTagBit */ + int top = (extract32(seed, 5, 1) ^ extract32(seed, 3, 1) ^ + extract32(seed, 2, 1) ^ extract32(seed, 0, 1)); + seed = (top << 15) | (seed >> 1); + offset |= top << i; + } + rtag = choose_nonexcluded_tag(start, offset, exclude); + env->cp15.rgsr_el1 = rtag | (seed << 8); + + return address_with_allocation_tag(rn, rtag); +} diff --git a/target/arm/translate-a64.c b/target/arm/translate-a64.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-a64.c +++ b/target/arm/translate-a64.c @@ -XXX,XX +XXX,XX @@ static TCGv_i64 clean_data_tbi(DisasContext *s, TCGv_i64 addr) return clean; } +/* Insert a zero tag into src, with the result at dst. */ +static void gen_address_with_allocation_tag0(TCGv_i64 dst, TCGv_i64 src) +{ + tcg_gen_andi_i64(dst, src, ~MAKE_64BIT_MASK(56, 4)); +} + typedef struct DisasCompare64 { TCGCond cond; TCGv_i64 value; @@ -XXX,XX +XXX,XX @@ static void disas_data_proc_2src(DisasContext *s, uint32_t insn) case 3: /* SDIV */ handle_div(s, true, sf, rm, rn, rd); break; + case 4: /* IRG */ + if (sf == 0 || !dc_isar_feature(aa64_mte_insn_reg, s)) { + goto do_unallocated; + } + if (s->ata) { + gen_helper_irg(cpu_reg_sp(s, rd), cpu_env, + cpu_reg_sp(s, rn), cpu_reg(s, rm)); + } else { + gen_address_with_allocation_tag0(cpu_reg_sp(s, rd), + cpu_reg_sp(s, rn)); + } + break; case 8: /* LSLV */ handle_shift_reg(s, A64_SHIFT_TYPE_LSL, sf, rm, rn, rd); break; diff --git a/target/arm/Makefile.objs b/target/arm/Makefile.objs index XXXXXXX..XXXXXXX 100644 --- a/target/arm/Makefile.objs +++ b/target/arm/Makefile.objs @@ -XXX,XX +XXX,XX @@ obj-$(CONFIG_SOFTMMU) += psci.o obj-$(TARGET_AARCH64) += translate-a64.o helper-a64.o obj-$(TARGET_AARCH64) += translate-sve.o sve_helper.o obj-$(TARGET_AARCH64) += pauth_helper.o +obj-$(TARGET_AARCH64) += mte_helper.o -- 2.20.1 From: Richard Henderson <richard.henderson@linaro.org> The current Arm ARM has adjusted the official decode of "Add/subtract (immediate)" so that the shift field is only bit 22, and bit 23 is part of the op1 field of the parent category "Data processing - immediate". Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200626033144.790098-11-richard.henderson@linaro.org Suggested-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/translate-a64.c | 23 ++++++++--------------- 1 file changed, 8 insertions(+), 15 deletions(-) diff --git a/target/arm/translate-a64.c b/target/arm/translate-a64.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-a64.c +++ b/target/arm/translate-a64.c @@ -XXX,XX +XXX,XX @@ static void disas_pc_rel_adr(DisasContext *s, uint32_t insn) /* * Add/subtract (immediate) * - * 31 30 29 28 24 23 22 21 10 9 5 4 0 - * +--+--+--+-----------+-----+-------------+-----+-----+ - * |sf|op| S| 1 0 0 0 1 |shift| imm12 | Rn | Rd | - * +--+--+--+-----------+-----+-------------+-----+-----+ + * 31 30 29 28 23 22 21 10 9 5 4 0 + * +--+--+--+-------------+--+-------------+-----+-----+ + * |sf|op| S| 1 0 0 0 1 0 |sh| imm12 | Rn | Rd | + * +--+--+--+-------------+--+-------------+-----+-----+ * * sf: 0 -> 32bit, 1 -> 64bit * op: 0 -> add , 1 -> sub * S: 1 -> set flags - * shift: 00 -> LSL imm by 0, 01 -> LSL imm by 12 + * sh: 1 -> LSL imm by 12 */ static void disas_add_sub_imm(DisasContext *s, uint32_t insn) { int rd = extract32(insn, 0, 5); int rn = extract32(insn, 5, 5); uint64_t imm = extract32(insn, 10, 12); - int shift = extract32(insn, 22, 2); + bool shift = extract32(insn, 22, 1); bool setflags = extract32(insn, 29, 1); bool sub_op = extract32(insn, 30, 1); bool is_64bit = extract32(insn, 31, 1); @@ -XXX,XX +XXX,XX @@ static void disas_add_sub_imm(DisasContext *s, uint32_t insn) TCGv_i64 tcg_rd = setflags ? cpu_reg(s, rd) : cpu_reg_sp(s, rd); TCGv_i64 tcg_result; - switch (shift) { - case 0x0: - break; - case 0x1: + if (shift) { imm <<= 12; - break; - default: - unallocated_encoding(s); - return; } tcg_result = tcg_temp_new_i64(); @@ -XXX,XX +XXX,XX @@ static void disas_data_proc_imm(DisasContext *s, uint32_t insn) case 0x20: case 0x21: /* PC-rel. addressing */ disas_pc_rel_adr(s, insn); break; - case 0x22: case 0x23: /* Add/subtract (immediate) */ + case 0x22: /* Add/subtract (immediate) */ disas_add_sub_imm(s, insn); break; case 0x24: /* Logical (immediate) */ -- 2.20.1 From: Richard Henderson <richard.henderson@linaro.org> Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200626033144.790098-12-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/helper-a64.h | 1 + target/arm/internals.h | 9 +++++++ target/arm/mte_helper.c | 10 ++++++++ target/arm/translate-a64.c | 51 ++++++++++++++++++++++++++++++++++++++ 4 files changed, 71 insertions(+) diff --git a/target/arm/helper-a64.h b/target/arm/helper-a64.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper-a64.h +++ b/target/arm/helper-a64.h @@ -XXX,XX +XXX,XX @@ DEF_HELPER_FLAGS_2(xpaci, TCG_CALL_NO_RWG_SE, i64, env, i64) DEF_HELPER_FLAGS_2(xpacd, TCG_CALL_NO_RWG_SE, i64, env, i64) DEF_HELPER_FLAGS_3(irg, TCG_CALL_NO_RWG, i64, env, i64, i64) +DEF_HELPER_FLAGS_4(addsubg, TCG_CALL_NO_RWG_SE, i64, env, i64, s32, i32) diff --git a/target/arm/internals.h b/target/arm/internals.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/internals.h +++ b/target/arm/internals.h @@ -XXX,XX +XXX,XX @@ void arm_log_exception(int idx); */ #define GMID_EL1_BS 6 +/* We associate one allocation tag per 16 bytes, the minimum. */ +#define LOG2_TAG_GRANULE 4 +#define TAG_GRANULE (1 << LOG2_TAG_GRANULE) + +static inline int allocation_tag_from_addr(uint64_t ptr) +{ + return extract64(ptr, 56, 4); +} + static inline uint64_t address_with_allocation_tag(uint64_t ptr, int rtag) { return deposit64(ptr, 56, 4, rtag); diff --git a/target/arm/mte_helper.c b/target/arm/mte_helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/mte_helper.c +++ b/target/arm/mte_helper.c @@ -XXX,XX +XXX,XX @@ uint64_t HELPER(irg)(CPUARMState *env, uint64_t rn, uint64_t rm) return address_with_allocation_tag(rn, rtag); } + +uint64_t HELPER(addsubg)(CPUARMState *env, uint64_t ptr, + int32_t offset, uint32_t tag_offset) +{ + int start_tag = allocation_tag_from_addr(ptr); + uint16_t exclude = extract32(env->cp15.gcr_el1, 0, 16); + int rtag = choose_nonexcluded_tag(start_tag, tag_offset, exclude); + + return address_with_allocation_tag(ptr + offset, rtag); +} diff --git a/target/arm/translate-a64.c b/target/arm/translate-a64.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-a64.c +++ b/target/arm/translate-a64.c @@ -XXX,XX +XXX,XX @@ static void disas_add_sub_imm(DisasContext *s, uint32_t insn) tcg_temp_free_i64(tcg_result); } +/* + * Add/subtract (immediate, with tags) + * + * 31 30 29 28 23 22 21 16 14 10 9 5 4 0 + * +--+--+--+-------------+--+---------+--+-------+-----+-----+ + * |sf|op| S| 1 0 0 0 1 1 |o2| uimm6 |o3| uimm4 | Rn | Rd | + * +--+--+--+-------------+--+---------+--+-------+-----+-----+ + * + * op: 0 -> add, 1 -> sub + */ +static void disas_add_sub_imm_with_tags(DisasContext *s, uint32_t insn) +{ + int rd = extract32(insn, 0, 5); + int rn = extract32(insn, 5, 5); + int uimm4 = extract32(insn, 10, 4); + int uimm6 = extract32(insn, 16, 6); + bool sub_op = extract32(insn, 30, 1); + TCGv_i64 tcg_rn, tcg_rd; + int imm; + + /* Test all of sf=1, S=0, o2=0, o3=0. */ + if ((insn & 0xa040c000u) != 0x80000000u || + !dc_isar_feature(aa64_mte_insn_reg, s)) { + unallocated_encoding(s); + return; + } + + imm = uimm6 << LOG2_TAG_GRANULE; + if (sub_op) { + imm = -imm; + } + + tcg_rn = cpu_reg_sp(s, rn); + tcg_rd = cpu_reg_sp(s, rd); + + if (s->ata) { + TCGv_i32 offset = tcg_const_i32(imm); + TCGv_i32 tag_offset = tcg_const_i32(uimm4); + + gen_helper_addsubg(tcg_rd, cpu_env, tcg_rn, offset, tag_offset); + tcg_temp_free_i32(tag_offset); + tcg_temp_free_i32(offset); + } else { + tcg_gen_addi_i64(tcg_rd, tcg_rn, imm); + gen_address_with_allocation_tag0(tcg_rd, tcg_rd); + } +} + /* The input should be a value in the bottom e bits (with higher * bits zero); returns that value replicated into every element * of size e in a 64 bit integer. @@ -XXX,XX +XXX,XX @@ static void disas_data_proc_imm(DisasContext *s, uint32_t insn) case 0x22: /* Add/subtract (immediate) */ disas_add_sub_imm(s, insn); break; + case 0x23: /* Add/subtract (immediate, with tags) */ + disas_add_sub_imm_with_tags(s, insn); + break; case 0x24: /* Logical (immediate) */ disas_logic_imm(s, insn); break; -- 2.20.1 From: Richard Henderson <richard.henderson@linaro.org> Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200626033144.790098-13-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/translate-a64.c | 15 +++++++++++++++ 1 file changed, 15 insertions(+) diff --git a/target/arm/translate-a64.c b/target/arm/translate-a64.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-a64.c +++ b/target/arm/translate-a64.c @@ -XXX,XX +XXX,XX @@ static void disas_data_proc_2src(DisasContext *s, uint32_t insn) cpu_reg_sp(s, rn)); } break; + case 5: /* GMI */ + if (sf == 0 || !dc_isar_feature(aa64_mte_insn_reg, s)) { + goto do_unallocated; + } else { + TCGv_i64 t1 = tcg_const_i64(1); + TCGv_i64 t2 = tcg_temp_new_i64(); + + tcg_gen_extract_i64(t2, cpu_reg_sp(s, rn), 56, 4); + tcg_gen_shl_i64(t1, t1, t2); + tcg_gen_or_i64(cpu_reg(s, rd), cpu_reg(s, rm), t1); + + tcg_temp_free_i64(t1); + tcg_temp_free_i64(t2); + } + break; case 8: /* LSLV */ handle_shift_reg(s, A64_SHIFT_TYPE_LSL, sf, rm, rn, rd); break; -- 2.20.1 From: Richard Henderson <richard.henderson@linaro.org> Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200626033144.790098-14-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/translate-a64.c | 24 ++++++++++++++++++++++-- 1 file changed, 22 insertions(+), 2 deletions(-) diff --git a/target/arm/translate-a64.c b/target/arm/translate-a64.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-a64.c +++ b/target/arm/translate-a64.c @@ -XXX,XX +XXX,XX @@ static void handle_crc32(DisasContext *s, */ static void disas_data_proc_2src(DisasContext *s, uint32_t insn) { - unsigned int sf, rm, opcode, rn, rd; + unsigned int sf, rm, opcode, rn, rd, setflag; sf = extract32(insn, 31, 1); + setflag = extract32(insn, 29, 1); rm = extract32(insn, 16, 5); opcode = extract32(insn, 10, 6); rn = extract32(insn, 5, 5); rd = extract32(insn, 0, 5); - if (extract32(insn, 29, 1)) { + if (setflag && opcode != 0) { unallocated_encoding(s); return; } switch (opcode) { + case 0: /* SUBP(S) */ + if (sf == 0 || !dc_isar_feature(aa64_mte_insn_reg, s)) { + goto do_unallocated; + } else { + TCGv_i64 tcg_n, tcg_m, tcg_d; + + tcg_n = read_cpu_reg_sp(s, rn, true); + tcg_m = read_cpu_reg_sp(s, rm, true); + tcg_gen_sextract_i64(tcg_n, tcg_n, 0, 56); + tcg_gen_sextract_i64(tcg_m, tcg_m, 0, 56); + tcg_d = cpu_reg(s, rd); + + if (setflag) { + gen_sub_CC(true, tcg_d, tcg_n, tcg_m); + } else { + tcg_gen_sub_i64(tcg_d, tcg_n, tcg_m); + } + } + break; case 2: /* UDIV */ handle_div(s, false, sf, rm, rn, rd); break; -- 2.20.1 From: Richard Henderson <richard.henderson@linaro.org> Use the same code as system mode, so that we generate the same exception + syndrome for the unaligned access. For the moment, if MTE is enabled so that this path is reachable, this would generate a SIGSEGV in the user-only cpu_loop. Decoding the syndrome to produce the proper SIGBUS will be done later. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200626033144.790098-15-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/cpu.c | 2 +- target/arm/tlb_helper.c | 4 ++-- 2 files changed, 3 insertions(+), 3 deletions(-) diff --git a/target/arm/cpu.c b/target/arm/cpu.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/cpu.c +++ b/target/arm/cpu.c @@ -XXX,XX +XXX,XX @@ static void arm_cpu_class_init(ObjectClass *oc, void *data) cc->tlb_fill = arm_cpu_tlb_fill; cc->debug_excp_handler = arm_debug_excp_handler; cc->debug_check_watchpoint = arm_debug_check_watchpoint; -#if !defined(CONFIG_USER_ONLY) cc->do_unaligned_access = arm_cpu_do_unaligned_access; +#if !defined(CONFIG_USER_ONLY) cc->do_transaction_failed = arm_cpu_do_transaction_failed; cc->adjust_watchpoint_address = arm_adjust_watchpoint_address; #endif /* CONFIG_TCG && !CONFIG_USER_ONLY */ diff --git a/target/arm/tlb_helper.c b/target/arm/tlb_helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/tlb_helper.c +++ b/target/arm/tlb_helper.c @@ -XXX,XX +XXX,XX @@ #include "internals.h" #include "exec/exec-all.h" -#if !defined(CONFIG_USER_ONLY) - static inline uint32_t merge_syn_data_abort(uint32_t template_syn, unsigned int target_el, bool same_el, bool ea, @@ -XXX,XX +XXX,XX @@ void arm_cpu_do_unaligned_access(CPUState *cs, vaddr vaddr, arm_deliver_fault(cpu, vaddr, access_type, mmu_idx, &fi); } +#if !defined(CONFIG_USER_ONLY) + /* * arm_cpu_do_transaction_failed: handle a memory system error response * (eg "no device/memory present at address") by raising an external abort -- 2.20.1 From: Richard Henderson <richard.henderson@linaro.org> Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200626033144.790098-16-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/helper-a64.h | 7 ++ target/arm/helper.h | 2 + target/arm/mte_helper.c | 194 +++++++++++++++++++++++++++++++++++++ target/arm/op_helper.c | 16 +++ target/arm/translate-a64.c | 172 +++++++++++++++++++++++++++++++- 5 files changed, 386 insertions(+), 5 deletions(-) diff --git a/target/arm/helper-a64.h b/target/arm/helper-a64.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper-a64.h +++ b/target/arm/helper-a64.h @@ -XXX,XX +XXX,XX @@ DEF_HELPER_FLAGS_2(xpacd, TCG_CALL_NO_RWG_SE, i64, env, i64) DEF_HELPER_FLAGS_3(irg, TCG_CALL_NO_RWG, i64, env, i64, i64) DEF_HELPER_FLAGS_4(addsubg, TCG_CALL_NO_RWG_SE, i64, env, i64, s32, i32) +DEF_HELPER_FLAGS_3(ldg, TCG_CALL_NO_WG, i64, env, i64, i64) +DEF_HELPER_FLAGS_3(stg, TCG_CALL_NO_WG, void, env, i64, i64) +DEF_HELPER_FLAGS_3(stg_parallel, TCG_CALL_NO_WG, void, env, i64, i64) +DEF_HELPER_FLAGS_2(stg_stub, TCG_CALL_NO_WG, void, env, i64) +DEF_HELPER_FLAGS_3(st2g, TCG_CALL_NO_WG, void, env, i64, i64) +DEF_HELPER_FLAGS_3(st2g_parallel, TCG_CALL_NO_WG, void, env, i64, i64) +DEF_HELPER_FLAGS_2(st2g_stub, TCG_CALL_NO_WG, void, env, i64) diff --git a/target/arm/helper.h b/target/arm/helper.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper.h +++ b/target/arm/helper.h @@ -XXX,XX +XXX,XX @@ DEF_HELPER_FLAGS_1(rebuild_hflags_a32_newel, TCG_CALL_NO_RWG, void, env) DEF_HELPER_FLAGS_2(rebuild_hflags_a32, TCG_CALL_NO_RWG, void, env, int) DEF_HELPER_FLAGS_2(rebuild_hflags_a64, TCG_CALL_NO_RWG, void, env, int) +DEF_HELPER_FLAGS_5(probe_access, TCG_CALL_NO_WG, void, env, tl, i32, i32, i32) + DEF_HELPER_1(vfp_get_fpscr, i32, env) DEF_HELPER_2(vfp_set_fpscr, void, env, i32) diff --git a/target/arm/mte_helper.c b/target/arm/mte_helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/mte_helper.c +++ b/target/arm/mte_helper.c @@ -XXX,XX +XXX,XX @@ static int choose_nonexcluded_tag(int tag, int offset, uint16_t exclude) return tag; } +/** + * allocation_tag_mem: + * @env: the cpu environment + * @ptr_mmu_idx: the addressing regime to use for the virtual address + * @ptr: the virtual address for which to look up tag memory + * @ptr_access: the access to use for the virtual address + * @ptr_size: the number of bytes in the normal memory access + * @tag_access: the access to use for the tag memory + * @tag_size: the number of bytes in the tag memory access + * @ra: the return address for exception handling + * + * Our tag memory is formatted as a sequence of little-endian nibbles. + * That is, the byte at (addr >> (LOG2_TAG_GRANULE + 1)) contains two + * tags, with the tag at [3:0] for the lower addr and the tag at [7:4] + * for the higher addr. + * + * Here, resolve the physical address from the virtual address, and return + * a pointer to the corresponding tag byte. Exit with exception if the + * virtual address is not accessible for @ptr_access. + * + * The @ptr_size and @tag_size values may not have an obvious relation + * due to the alignment of @ptr, and the number of tag checks required. + * + * If there is no tag storage corresponding to @ptr, return NULL. + */ +static uint8_t *allocation_tag_mem(CPUARMState *env, int ptr_mmu_idx, + uint64_t ptr, MMUAccessType ptr_access, + int ptr_size, MMUAccessType tag_access, + int tag_size, uintptr_t ra) +{ + /* Tag storage not implemented. */ + return NULL; +} + uint64_t HELPER(irg)(CPUARMState *env, uint64_t rn, uint64_t rm) { int rtag; @@ -XXX,XX +XXX,XX @@ uint64_t HELPER(addsubg)(CPUARMState *env, uint64_t ptr, return address_with_allocation_tag(ptr + offset, rtag); } + +static int load_tag1(uint64_t ptr, uint8_t *mem) +{ + int ofs = extract32(ptr, LOG2_TAG_GRANULE, 1) * 4; + return extract32(*mem, ofs, 4); +} + +uint64_t HELPER(ldg)(CPUARMState *env, uint64_t ptr, uint64_t xt) +{ + int mmu_idx = cpu_mmu_index(env, false); + uint8_t *mem; + int rtag = 0; + + /* Trap if accessing an invalid page. */ + mem = allocation_tag_mem(env, mmu_idx, ptr, MMU_DATA_LOAD, 1, + MMU_DATA_LOAD, 1, GETPC()); + + /* Load if page supports tags. */ + if (mem) { + rtag = load_tag1(ptr, mem); + } + + return address_with_allocation_tag(xt, rtag); +} + +static void check_tag_aligned(CPUARMState *env, uint64_t ptr, uintptr_t ra) +{ + if (unlikely(!QEMU_IS_ALIGNED(ptr, TAG_GRANULE))) { + arm_cpu_do_unaligned_access(env_cpu(env), ptr, MMU_DATA_STORE, + cpu_mmu_index(env, false), ra); + g_assert_not_reached(); + } +} + +/* For use in a non-parallel context, store to the given nibble. */ +static void store_tag1(uint64_t ptr, uint8_t *mem, int tag) +{ + int ofs = extract32(ptr, LOG2_TAG_GRANULE, 1) * 4; + *mem = deposit32(*mem, ofs, 4, tag); +} + +/* For use in a parallel context, atomically store to the given nibble. */ +static void store_tag1_parallel(uint64_t ptr, uint8_t *mem, int tag) +{ + int ofs = extract32(ptr, LOG2_TAG_GRANULE, 1) * 4; + uint8_t old = atomic_read(mem); + + while (1) { + uint8_t new = deposit32(old, ofs, 4, tag); + uint8_t cmp = atomic_cmpxchg(mem, old, new); + if (likely(cmp == old)) { + return; + } + old = cmp; + } +} + +typedef void stg_store1(uint64_t, uint8_t *, int); + +static inline void do_stg(CPUARMState *env, uint64_t ptr, uint64_t xt, + uintptr_t ra, stg_store1 store1) +{ + int mmu_idx = cpu_mmu_index(env, false); + uint8_t *mem; + + check_tag_aligned(env, ptr, ra); + + /* Trap if accessing an invalid page. */ + mem = allocation_tag_mem(env, mmu_idx, ptr, MMU_DATA_STORE, TAG_GRANULE, + MMU_DATA_STORE, 1, ra); + + /* Store if page supports tags. */ + if (mem) { + store1(ptr, mem, allocation_tag_from_addr(xt)); + } +} + +void HELPER(stg)(CPUARMState *env, uint64_t ptr, uint64_t xt) +{ + do_stg(env, ptr, xt, GETPC(), store_tag1); +} + +void HELPER(stg_parallel)(CPUARMState *env, uint64_t ptr, uint64_t xt) +{ + do_stg(env, ptr, xt, GETPC(), store_tag1_parallel); +} + +void HELPER(stg_stub)(CPUARMState *env, uint64_t ptr) +{ + int mmu_idx = cpu_mmu_index(env, false); + uintptr_t ra = GETPC(); + + check_tag_aligned(env, ptr, ra); + probe_write(env, ptr, TAG_GRANULE, mmu_idx, ra); +} + +static inline void do_st2g(CPUARMState *env, uint64_t ptr, uint64_t xt, + uintptr_t ra, stg_store1 store1) +{ + int mmu_idx = cpu_mmu_index(env, false); + int tag = allocation_tag_from_addr(xt); + uint8_t *mem1, *mem2; + + check_tag_aligned(env, ptr, ra); + + /* + * Trap if accessing an invalid page(s). + * This takes priority over !allocation_tag_access_enabled. + */ + if (ptr & TAG_GRANULE) { + /* Two stores unaligned mod TAG_GRANULE*2 -- modify two bytes. */ + mem1 = allocation_tag_mem(env, mmu_idx, ptr, MMU_DATA_STORE, + TAG_GRANULE, MMU_DATA_STORE, 1, ra); + mem2 = allocation_tag_mem(env, mmu_idx, ptr + TAG_GRANULE, + MMU_DATA_STORE, TAG_GRANULE, + MMU_DATA_STORE, 1, ra); + + /* Store if page(s) support tags. */ + if (mem1) { + store1(TAG_GRANULE, mem1, tag); + } + if (mem2) { + store1(0, mem2, tag); + } + } else { + /* Two stores aligned mod TAG_GRANULE*2 -- modify one byte. */ + mem1 = allocation_tag_mem(env, mmu_idx, ptr, MMU_DATA_STORE, + 2 * TAG_GRANULE, MMU_DATA_STORE, 1, ra); + if (mem1) { + tag |= tag << 4; + atomic_set(mem1, tag); + } + } +} + +void HELPER(st2g)(CPUARMState *env, uint64_t ptr, uint64_t xt) +{ + do_st2g(env, ptr, xt, GETPC(), store_tag1); +} + +void HELPER(st2g_parallel)(CPUARMState *env, uint64_t ptr, uint64_t xt) +{ + do_st2g(env, ptr, xt, GETPC(), store_tag1_parallel); +} + +void HELPER(st2g_stub)(CPUARMState *env, uint64_t ptr) +{ + int mmu_idx = cpu_mmu_index(env, false); + uintptr_t ra = GETPC(); + int in_page = -(ptr | TARGET_PAGE_MASK); + + check_tag_aligned(env, ptr, ra); + + if (likely(in_page >= 2 * TAG_GRANULE)) { + probe_write(env, ptr, 2 * TAG_GRANULE, mmu_idx, ra); + } else { + probe_write(env, ptr, TAG_GRANULE, mmu_idx, ra); + probe_write(env, ptr + TAG_GRANULE, TAG_GRANULE, mmu_idx, ra); + } +} diff --git a/target/arm/op_helper.c b/target/arm/op_helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/op_helper.c +++ b/target/arm/op_helper.c @@ -XXX,XX +XXX,XX @@ uint32_t HELPER(ror_cc)(CPUARMState *env, uint32_t x, uint32_t i) return ((uint32_t)x >> shift) | (x << (32 - shift)); } } + +void HELPER(probe_access)(CPUARMState *env, target_ulong ptr, + uint32_t access_type, uint32_t mmu_idx, + uint32_t size) +{ + uint32_t in_page = -((uint32_t)ptr | TARGET_PAGE_SIZE); + uintptr_t ra = GETPC(); + + if (likely(size <= in_page)) { + probe_access(env, ptr, size, access_type, mmu_idx, ra); + } else { + probe_access(env, ptr, in_page, access_type, mmu_idx, ra); + probe_access(env, ptr + in_page, size - in_page, + access_type, mmu_idx, ra); + } +} diff --git a/target/arm/translate-a64.c b/target/arm/translate-a64.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-a64.c +++ b/target/arm/translate-a64.c @@ -XXX,XX +XXX,XX @@ static void gen_address_with_allocation_tag0(TCGv_i64 dst, TCGv_i64 src) tcg_gen_andi_i64(dst, src, ~MAKE_64BIT_MASK(56, 4)); } +static void gen_probe_access(DisasContext *s, TCGv_i64 ptr, + MMUAccessType acc, int log2_size) +{ + TCGv_i32 t_acc = tcg_const_i32(acc); + TCGv_i32 t_idx = tcg_const_i32(get_mem_index(s)); + TCGv_i32 t_size = tcg_const_i32(1 << log2_size); + + gen_helper_probe_access(cpu_env, ptr, t_acc, t_idx, t_size); + tcg_temp_free_i32(t_acc); + tcg_temp_free_i32(t_idx); + tcg_temp_free_i32(t_size); +} + typedef struct DisasCompare64 { TCGCond cond; TCGv_i64 value; @@ -XXX,XX +XXX,XX @@ static void disas_ldst_single_struct(DisasContext *s, uint32_t insn) } } +/* + * Load/Store memory tags + * + * 31 30 29 24 22 21 12 10 5 0 + * +-----+-------------+-----+---+------+-----+------+------+ + * | 1 1 | 0 1 1 0 0 1 | op1 | 1 | imm9 | op2 | Rn | Rt | + * +-----+-------------+-----+---+------+-----+------+------+ + */ +static void disas_ldst_tag(DisasContext *s, uint32_t insn) +{ + int rt = extract32(insn, 0, 5); + int rn = extract32(insn, 5, 5); + uint64_t offset = sextract64(insn, 12, 9) << LOG2_TAG_GRANULE; + int op2 = extract32(insn, 10, 2); + int op1 = extract32(insn, 22, 2); + bool is_load = false, is_pair = false, is_zero = false; + int index = 0; + TCGv_i64 addr, clean_addr, tcg_rt; + + /* We checked insn bits [29:24,21] in the caller. */ + if (extract32(insn, 30, 2) != 3) { + goto do_unallocated; + } + + /* + * @index is a tri-state variable which has 3 states: + * < 0 : post-index, writeback + * = 0 : signed offset + * > 0 : pre-index, writeback + */ + switch (op1) { + case 0: + if (op2 != 0) { + /* STG */ + index = op2 - 2; + break; + } + goto do_unallocated; + case 1: + if (op2 != 0) { + /* STZG */ + is_zero = true; + index = op2 - 2; + } else { + /* LDG */ + is_load = true; + } + break; + case 2: + if (op2 != 0) { + /* ST2G */ + is_pair = true; + index = op2 - 2; + break; + } + goto do_unallocated; + case 3: + if (op2 != 0) { + /* STZ2G */ + is_pair = is_zero = true; + index = op2 - 2; + break; + } + goto do_unallocated; + + default: + do_unallocated: + unallocated_encoding(s); + return; + } + + if (!dc_isar_feature(aa64_mte_insn_reg, s)) { + goto do_unallocated; + } + + if (rn == 31) { + gen_check_sp_alignment(s); + } + + addr = read_cpu_reg_sp(s, rn, true); + if (index >= 0) { + /* pre-index or signed offset */ + tcg_gen_addi_i64(addr, addr, offset); + } + + if (is_load) { + tcg_gen_andi_i64(addr, addr, -TAG_GRANULE); + tcg_rt = cpu_reg(s, rt); + if (s->ata) { + gen_helper_ldg(tcg_rt, cpu_env, addr, tcg_rt); + } else { + clean_addr = clean_data_tbi(s, addr); + gen_probe_access(s, clean_addr, MMU_DATA_LOAD, MO_8); + gen_address_with_allocation_tag0(tcg_rt, addr); + } + } else { + tcg_rt = cpu_reg_sp(s, rt); + if (!s->ata) { + /* + * For STG and ST2G, we need to check alignment and probe memory. + * TODO: For STZG and STZ2G, we could rely on the stores below, + * at least for system mode; user-only won't enforce alignment. + */ + if (is_pair) { + gen_helper_st2g_stub(cpu_env, addr); + } else { + gen_helper_stg_stub(cpu_env, addr); + } + } else if (tb_cflags(s->base.tb) & CF_PARALLEL) { + if (is_pair) { + gen_helper_st2g_parallel(cpu_env, addr, tcg_rt); + } else { + gen_helper_stg_parallel(cpu_env, addr, tcg_rt); + } + } else { + if (is_pair) { + gen_helper_st2g(cpu_env, addr, tcg_rt); + } else { + gen_helper_stg(cpu_env, addr, tcg_rt); + } + } + } + + if (is_zero) { + TCGv_i64 clean_addr = clean_data_tbi(s, addr); + TCGv_i64 tcg_zero = tcg_const_i64(0); + int mem_index = get_mem_index(s); + int i, n = (1 + is_pair) << LOG2_TAG_GRANULE; + + tcg_gen_qemu_st_i64(tcg_zero, clean_addr, mem_index, + MO_Q | MO_ALIGN_16); + for (i = 8; i < n; i += 8) { + tcg_gen_addi_i64(clean_addr, clean_addr, 8); + tcg_gen_qemu_st_i64(tcg_zero, clean_addr, mem_index, MO_Q); + } + tcg_temp_free_i64(tcg_zero); + } + + if (index != 0) { + /* pre-index or post-index */ + if (index < 0) { + /* post-index */ + tcg_gen_addi_i64(addr, addr, offset); + } + tcg_gen_mov_i64(cpu_reg_sp(s, rn), addr); + } +} + /* Loads and stores */ static void disas_ldst(DisasContext *s, uint32_t insn) { @@ -XXX,XX +XXX,XX @@ static void disas_ldst(DisasContext *s, uint32_t insn) case 0x0d: /* AdvSIMD load/store single structure */ disas_ldst_single_struct(s, insn); break; - case 0x19: /* LDAPR/STLR (unscaled immediate) */ - if (extract32(insn, 10, 2) != 0 || - extract32(insn, 21, 1) != 0) { + case 0x19: + if (extract32(insn, 21, 1) != 0) { + disas_ldst_tag(s, insn); + } else if (extract32(insn, 10, 2) == 0) { + disas_ldst_ldapr_stlr(s, insn); + } else { unallocated_encoding(s); - break; } - disas_ldst_ldapr_stlr(s, insn); break; default: unallocated_encoding(s); -- 2.20.1 From: Richard Henderson <richard.henderson@linaro.org> Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200626033144.790098-17-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/translate-a64.c | 29 ++++++++++++++++++++++++++--- 1 file changed, 26 insertions(+), 3 deletions(-) diff --git a/target/arm/translate-a64.c b/target/arm/translate-a64.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-a64.c +++ b/target/arm/translate-a64.c @@ -XXX,XX +XXX,XX @@ static void disas_ld_lit(DisasContext *s, uint32_t insn) * +-----+-------+---+---+-------+---+-------+-------+------+------+ * * opc: LDP/STP/LDNP/STNP 00 -> 32 bit, 10 -> 64 bit - * LDPSW 01 + * LDPSW/STGP 01 * LDP/STP/LDNP/STNP (SIMD) 00 -> 32 bit, 01 -> 64 bit, 10 -> 128 bit * V: 0 -> GPR, 1 -> Vector * idx: 00 -> signed offset with non-temporal hint, 01 -> post-index, @@ -XXX,XX +XXX,XX @@ static void disas_ldst_pair(DisasContext *s, uint32_t insn) bool is_signed = false; bool postindex = false; bool wback = false; + bool set_tag = false; TCGv_i64 clean_addr, dirty_addr; @@ -XXX,XX +XXX,XX @@ static void disas_ldst_pair(DisasContext *s, uint32_t insn) if (is_vector) { size = 2 + opc; + } else if (opc == 1 && !is_load) { + /* STGP */ + if (!dc_isar_feature(aa64_mte_insn_reg, s) || index == 0) { + unallocated_encoding(s); + return; + } + size = 3; + set_tag = true; } else { size = 2 + extract32(opc, 1, 1); is_signed = extract32(opc, 0, 1); @@ -XXX,XX +XXX,XX @@ static void disas_ldst_pair(DisasContext *s, uint32_t insn) return; } - offset <<= size; + offset <<= (set_tag ? LOG2_TAG_GRANULE : size); if (rn == 31) { gen_check_sp_alignment(s); @@ -XXX,XX +XXX,XX @@ static void disas_ldst_pair(DisasContext *s, uint32_t insn) if (!postindex) { tcg_gen_addi_i64(dirty_addr, dirty_addr, offset); } - clean_addr = clean_data_tbi(s, dirty_addr); + if (set_tag) { + if (!s->ata) { + /* + * TODO: We could rely on the stores below, at least for + * system mode, if we arrange to add MO_ALIGN_16. + */ + gen_helper_stg_stub(cpu_env, dirty_addr); + } else if (tb_cflags(s->base.tb) & CF_PARALLEL) { + gen_helper_stg_parallel(cpu_env, dirty_addr, dirty_addr); + } else { + gen_helper_stg(cpu_env, dirty_addr, dirty_addr); + } + } + + clean_addr = clean_data_tbi(s, dirty_addr); if (is_vector) { if (is_load) { do_fp_ld(s, rt, clean_addr, size); -- 2.20.1 From: Richard Henderson <richard.henderson@linaro.org> We can simplify our DC_ZVA if we recognize that the largest BS that we actually use in system mode is 64. Let us just assert that it fits within TARGET_PAGE_SIZE. For DC_GVA and STZGM, we want to be able to write whole bytes of tag memory, so assert that BS is >= 2 * TAG_GRANULE, or 32. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200626033144.790098-18-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/cpu.c | 24 ++++++++++++++++++++++++ 1 file changed, 24 insertions(+) diff --git a/target/arm/cpu.c b/target/arm/cpu.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/cpu.c +++ b/target/arm/cpu.c @@ -XXX,XX +XXX,XX @@ static void arm_cpu_realizefn(DeviceState *dev, Error **errp) } #endif + if (tcg_enabled()) { + int dcz_blocklen = 4 << cpu->dcz_blocksize; + + /* + * We only support DCZ blocklen that fits on one page. + * + * Architectually this is always true. However TARGET_PAGE_SIZE + * is variable and, for compatibility with -machine virt-2.7, + * is only 1KiB, as an artifact of legacy ARMv5 subpage support. + * But even then, while the largest architectural DCZ blocklen + * is 2KiB, no cpu actually uses such a large blocklen. + */ + assert(dcz_blocklen <= TARGET_PAGE_SIZE); + + /* + * We only support DCZ blocksize >= 2*TAG_GRANULE, which is to say + * both nibbles of each byte storing tag data may be written at once. + * Since TAG_GRANULE is 16, this means that blocklen must be >= 32. + */ + if (cpu_isar_feature(aa64_mte, cpu)) { + assert(dcz_blocklen >= 2 * TAG_GRANULE); + } + } + qemu_init_vcpu(cs); cpu_reset(cs); -- 2.20.1 From: Richard Henderson <richard.henderson@linaro.org> Now that we know that the operation is on a single page, we need not loop over pages while probing. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200626033144.790098-19-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/helper-a64.c | 94 +++++++++++------------------------------ 1 file changed, 25 insertions(+), 69 deletions(-) diff --git a/target/arm/helper-a64.c b/target/arm/helper-a64.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper-a64.c +++ b/target/arm/helper-a64.c @@ -XXX,XX +XXX,XX @@ void HELPER(dc_zva)(CPUARMState *env, uint64_t vaddr_in) * (which matches the usual QEMU behaviour of not implementing either * alignment faults or any memory attribute handling). */ - - ARMCPU *cpu = env_archcpu(env); - uint64_t blocklen = 4 << cpu->dcz_blocksize; + int blocklen = 4 << env_archcpu(env)->dcz_blocksize; uint64_t vaddr = vaddr_in & ~(blocklen - 1); + int mmu_idx = cpu_mmu_index(env, false); + void *mem; + + /* + * Trapless lookup. In addition to actual invalid page, may + * return NULL for I/O, watchpoints, clean pages, etc. + */ + mem = tlb_vaddr_to_host(env, vaddr, MMU_DATA_STORE, mmu_idx); #ifndef CONFIG_USER_ONLY - { + if (unlikely(!mem)) { + uintptr_t ra = GETPC(); + /* - * Slightly awkwardly, QEMU's TARGET_PAGE_SIZE may be less than - * the block size so we might have to do more than one TLB lookup. - * We know that in fact for any v8 CPU the page size is at least 4K - * and the block size must be 2K or less, but TARGET_PAGE_SIZE is only - * 1K as an artefact of legacy v5 subpage support being present in the - * same QEMU executable. So in practice the hostaddr[] array has - * two entries, given the current setting of TARGET_PAGE_BITS_MIN. + * Trap if accessing an invalid page. DC_ZVA requires that we supply + * the original pointer for an invalid page. But watchpoints require + * that we probe the actual space. So do both. */ - int maxidx = DIV_ROUND_UP(blocklen, TARGET_PAGE_SIZE); - void *hostaddr[DIV_ROUND_UP(2 * KiB, 1 << TARGET_PAGE_BITS_MIN)]; - int try, i; - unsigned mmu_idx = cpu_mmu_index(env, false); - TCGMemOpIdx oi = make_memop_idx(MO_UB, mmu_idx); + (void) probe_write(env, vaddr_in, 1, mmu_idx, ra); + mem = probe_write(env, vaddr, blocklen, mmu_idx, ra); - assert(maxidx <= ARRAY_SIZE(hostaddr)); - - for (try = 0; try < 2; try++) { - - for (i = 0; i < maxidx; i++) { - hostaddr[i] = tlb_vaddr_to_host(env, - vaddr + TARGET_PAGE_SIZE * i, - 1, mmu_idx); - if (!hostaddr[i]) { - break; - } - } - if (i == maxidx) { - /* - * If it's all in the TLB it's fair game for just writing to; - * we know we don't need to update dirty status, etc. - */ - for (i = 0; i < maxidx - 1; i++) { - memset(hostaddr[i], 0, TARGET_PAGE_SIZE); - } - memset(hostaddr[i], 0, blocklen - (i * TARGET_PAGE_SIZE)); - return; - } + if (unlikely(!mem)) { /* - * OK, try a store and see if we can populate the tlb. This - * might cause an exception if the memory isn't writable, - * in which case we will longjmp out of here. We must for - * this purpose use the actual register value passed to us - * so that we get the fault address right. + * The only remaining reason for mem == NULL is I/O. + * Just do a series of byte writes as the architecture demands. */ - helper_ret_stb_mmu(env, vaddr_in, 0, oi, GETPC()); - /* Now we can populate the other TLB entries, if any */ - for (i = 0; i < maxidx; i++) { - uint64_t va = vaddr + TARGET_PAGE_SIZE * i; - if (va != (vaddr_in & TARGET_PAGE_MASK)) { - helper_ret_stb_mmu(env, va, 0, oi, GETPC()); - } + for (int i = 0; i < blocklen; i++) { + cpu_stb_mmuidx_ra(env, vaddr + i, 0, mmu_idx, ra); } - } - - /* - * Slow path (probably attempt to do this to an I/O device or - * similar, or clearing of a block of code we have translations - * cached for). Just do a series of byte writes as the architecture - * demands. It's not worth trying to use a cpu_physical_memory_map(), - * memset(), unmap() sequence here because: - * + we'd need to account for the blocksize being larger than a page - * + the direct-RAM access case is almost always going to be dealt - * with in the fastpath code above, so there's no speed benefit - * + we would have to deal with the map returning NULL because the - * bounce buffer was in use - */ - for (i = 0; i < blocklen; i++) { - helper_ret_stb_mmu(env, vaddr + i, 0, oi, GETPC()); + return; } } -#else - memset(g2h(vaddr), 0, blocklen); #endif + + memset(mem, 0, blocklen); } -- 2.20.1 From: Richard Henderson <richard.henderson@linaro.org> Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200626033144.790098-20-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/helper-a64.h | 3 ++ target/arm/translate.h | 2 + target/arm/mte_helper.c | 84 ++++++++++++++++++++++++++++++++++++++ target/arm/translate-a64.c | 72 ++++++++++++++++++++++++++++---- 4 files changed, 153 insertions(+), 8 deletions(-) diff --git a/target/arm/helper-a64.h b/target/arm/helper-a64.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper-a64.h +++ b/target/arm/helper-a64.h @@ -XXX,XX +XXX,XX @@ DEF_HELPER_FLAGS_2(stg_stub, TCG_CALL_NO_WG, void, env, i64) DEF_HELPER_FLAGS_3(st2g, TCG_CALL_NO_WG, void, env, i64, i64) DEF_HELPER_FLAGS_3(st2g_parallel, TCG_CALL_NO_WG, void, env, i64, i64) DEF_HELPER_FLAGS_2(st2g_stub, TCG_CALL_NO_WG, void, env, i64) +DEF_HELPER_FLAGS_2(ldgm, TCG_CALL_NO_WG, i64, env, i64) +DEF_HELPER_FLAGS_3(stgm, TCG_CALL_NO_WG, void, env, i64, i64) +DEF_HELPER_FLAGS_3(stzgm_tags, TCG_CALL_NO_WG, void, env, i64, i64) diff --git a/target/arm/translate.h b/target/arm/translate.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate.h +++ b/target/arm/translate.h @@ -XXX,XX +XXX,XX @@ typedef struct DisasContext { * < 0, set by the current instruction. */ int8_t btype; + /* A copy of cpu->dcz_blocksize. */ + uint8_t dcz_blocksize; /* True if this page is guarded. */ bool guarded_page; /* Bottom two bits of XScale c15_cpar coprocessor access control reg */ diff --git a/target/arm/mte_helper.c b/target/arm/mte_helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/mte_helper.c +++ b/target/arm/mte_helper.c @@ -XXX,XX +XXX,XX @@ void HELPER(st2g_stub)(CPUARMState *env, uint64_t ptr) probe_write(env, ptr + TAG_GRANULE, TAG_GRANULE, mmu_idx, ra); } } + +#define LDGM_STGM_SIZE (4 << GMID_EL1_BS) + +uint64_t HELPER(ldgm)(CPUARMState *env, uint64_t ptr) +{ + int mmu_idx = cpu_mmu_index(env, false); + uintptr_t ra = GETPC(); + void *tag_mem; + + ptr = QEMU_ALIGN_DOWN(ptr, LDGM_STGM_SIZE); + + /* Trap if accessing an invalid page. */ + tag_mem = allocation_tag_mem(env, mmu_idx, ptr, MMU_DATA_LOAD, + LDGM_STGM_SIZE, MMU_DATA_LOAD, + LDGM_STGM_SIZE / (2 * TAG_GRANULE), ra); + + /* The tag is squashed to zero if the page does not support tags. */ + if (!tag_mem) { + return 0; + } + + QEMU_BUILD_BUG_ON(GMID_EL1_BS != 6); + /* + * We are loading 64-bits worth of tags. The ordering of elements + * within the word corresponds to a 64-bit little-endian operation. + */ + return ldq_le_p(tag_mem); +} + +void HELPER(stgm)(CPUARMState *env, uint64_t ptr, uint64_t val) +{ + int mmu_idx = cpu_mmu_index(env, false); + uintptr_t ra = GETPC(); + void *tag_mem; + + ptr = QEMU_ALIGN_DOWN(ptr, LDGM_STGM_SIZE); + + /* Trap if accessing an invalid page. */ + tag_mem = allocation_tag_mem(env, mmu_idx, ptr, MMU_DATA_STORE, + LDGM_STGM_SIZE, MMU_DATA_LOAD, + LDGM_STGM_SIZE / (2 * TAG_GRANULE), ra); + + /* + * Tag store only happens if the page support tags, + * and if the OS has enabled access to the tags. + */ + if (!tag_mem) { + return; + } + + QEMU_BUILD_BUG_ON(GMID_EL1_BS != 6); + /* + * We are storing 64-bits worth of tags. The ordering of elements + * within the word corresponds to a 64-bit little-endian operation. + */ + stq_le_p(tag_mem, val); +} + +void HELPER(stzgm_tags)(CPUARMState *env, uint64_t ptr, uint64_t val) +{ + uintptr_t ra = GETPC(); + int mmu_idx = cpu_mmu_index(env, false); + int log2_dcz_bytes, log2_tag_bytes; + intptr_t dcz_bytes, tag_bytes; + uint8_t *mem; + + /* + * In arm_cpu_realizefn, we assert that dcz > LOG2_TAG_GRANULE+1, + * i.e. 32 bytes, which is an unreasonably small dcz anyway, + * to make sure that we can access one complete tag byte here. + */ + log2_dcz_bytes = env_archcpu(env)->dcz_blocksize + 2; + log2_tag_bytes = log2_dcz_bytes - (LOG2_TAG_GRANULE + 1); + dcz_bytes = (intptr_t)1 << log2_dcz_bytes; + tag_bytes = (intptr_t)1 << log2_tag_bytes; + ptr &= -dcz_bytes; + + mem = allocation_tag_mem(env, mmu_idx, ptr, MMU_DATA_STORE, dcz_bytes, + MMU_DATA_STORE, tag_bytes, ra); + if (mem) { + int tag_pair = (val & 0xf) * 0x11; + memset(mem, tag_pair, tag_bytes); + } +} diff --git a/target/arm/translate-a64.c b/target/arm/translate-a64.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-a64.c +++ b/target/arm/translate-a64.c @@ -XXX,XX +XXX,XX @@ static void disas_ldst_tag(DisasContext *s, uint32_t insn) uint64_t offset = sextract64(insn, 12, 9) << LOG2_TAG_GRANULE; int op2 = extract32(insn, 10, 2); int op1 = extract32(insn, 22, 2); - bool is_load = false, is_pair = false, is_zero = false; + bool is_load = false, is_pair = false, is_zero = false, is_mult = false; int index = 0; TCGv_i64 addr, clean_addr, tcg_rt; @@ -XXX,XX +XXX,XX @@ static void disas_ldst_tag(DisasContext *s, uint32_t insn) if (op2 != 0) { /* STG */ index = op2 - 2; - break; + } else { + /* STZGM */ + if (s->current_el == 0 || offset != 0) { + goto do_unallocated; + } + is_mult = is_zero = true; } - goto do_unallocated; + break; case 1: if (op2 != 0) { /* STZG */ @@ -XXX,XX +XXX,XX @@ static void disas_ldst_tag(DisasContext *s, uint32_t insn) /* ST2G */ is_pair = true; index = op2 - 2; - break; + } else { + /* STGM */ + if (s->current_el == 0 || offset != 0) { + goto do_unallocated; + } + is_mult = true; } - goto do_unallocated; + break; case 3: if (op2 != 0) { /* STZ2G */ is_pair = is_zero = true; index = op2 - 2; - break; + } else { + /* LDGM */ + if (s->current_el == 0 || offset != 0) { + goto do_unallocated; + } + is_mult = is_load = true; } - goto do_unallocated; + break; default: do_unallocated: @@ -XXX,XX +XXX,XX @@ static void disas_ldst_tag(DisasContext *s, uint32_t insn) return; } - if (!dc_isar_feature(aa64_mte_insn_reg, s)) { + if (is_mult + ? !dc_isar_feature(aa64_mte, s) + : !dc_isar_feature(aa64_mte_insn_reg, s)) { goto do_unallocated; } @@ -XXX,XX +XXX,XX @@ static void disas_ldst_tag(DisasContext *s, uint32_t insn) tcg_gen_addi_i64(addr, addr, offset); } + if (is_mult) { + tcg_rt = cpu_reg(s, rt); + + if (is_zero) { + int size = 4 << s->dcz_blocksize; + + if (s->ata) { + gen_helper_stzgm_tags(cpu_env, addr, tcg_rt); + } + /* + * The non-tags portion of STZGM is mostly like DC_ZVA, + * except the alignment happens before the access. + */ + clean_addr = clean_data_tbi(s, addr); + tcg_gen_andi_i64(clean_addr, clean_addr, -size); + gen_helper_dc_zva(cpu_env, clean_addr); + } else if (s->ata) { + if (is_load) { + gen_helper_ldgm(tcg_rt, cpu_env, addr); + } else { + gen_helper_stgm(cpu_env, addr, tcg_rt); + } + } else { + MMUAccessType acc = is_load ? MMU_DATA_LOAD : MMU_DATA_STORE; + int size = 4 << GMID_EL1_BS; + + clean_addr = clean_data_tbi(s, addr); + tcg_gen_andi_i64(clean_addr, clean_addr, -size); + gen_probe_access(s, clean_addr, acc, size); + + if (is_load) { + /* The result tags are zeros. */ + tcg_gen_movi_i64(tcg_rt, 0); + } + } + return; + } + if (is_load) { tcg_gen_andi_i64(addr, addr, -TAG_GRANULE); tcg_rt = cpu_reg(s, rt); @@ -XXX,XX +XXX,XX @@ static void aarch64_tr_init_disas_context(DisasContextBase *dcbase, dc->vec_stride = 0; dc->cp_regs = arm_cpu->cp_regs; dc->features = env->features; + dc->dcz_blocksize = arm_cpu->dcz_blocksize; /* Single step state. The code-generation logic here is: * SS_ACTIVE == 0: -- 2.20.1 From: Richard Henderson <richard.henderson@linaro.org> Like the regular data cache flushes, these are nops within qemu. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200626033144.790098-21-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/helper.c | 65 +++++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 65 insertions(+) diff --git a/target/arm/helper.c b/target/arm/helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper.c +++ b/target/arm/helper.c @@ -XXX,XX +XXX,XX @@ static const ARMCPRegInfo mte_reginfo[] = { .opc0 = 3, .opc1 = 3, .crn = 4, .crm = 2, .opc2 = 7, .type = ARM_CP_NO_RAW, .access = PL0_RW, .readfn = tco_read, .writefn = tco_write }, + { .name = "DC_IGVAC", .state = ARM_CP_STATE_AA64, + .opc0 = 1, .opc1 = 0, .crn = 7, .crm = 6, .opc2 = 3, + .type = ARM_CP_NOP, .access = PL1_W, + .accessfn = aa64_cacheop_poc_access }, + { .name = "DC_IGSW", .state = ARM_CP_STATE_AA64, + .opc0 = 1, .opc1 = 0, .crn = 7, .crm = 6, .opc2 = 4, + .type = ARM_CP_NOP, .access = PL1_W, .accessfn = access_tsw }, + { .name = "DC_IGDVAC", .state = ARM_CP_STATE_AA64, + .opc0 = 1, .opc1 = 0, .crn = 7, .crm = 6, .opc2 = 5, + .type = ARM_CP_NOP, .access = PL1_W, + .accessfn = aa64_cacheop_poc_access }, + { .name = "DC_IGDSW", .state = ARM_CP_STATE_AA64, + .opc0 = 1, .opc1 = 0, .crn = 7, .crm = 6, .opc2 = 6, + .type = ARM_CP_NOP, .access = PL1_W, .accessfn = access_tsw }, + { .name = "DC_CGSW", .state = ARM_CP_STATE_AA64, + .opc0 = 1, .opc1 = 0, .crn = 7, .crm = 10, .opc2 = 4, + .type = ARM_CP_NOP, .access = PL1_W, .accessfn = access_tsw }, + { .name = "DC_CGDSW", .state = ARM_CP_STATE_AA64, + .opc0 = 1, .opc1 = 0, .crn = 7, .crm = 10, .opc2 = 6, + .type = ARM_CP_NOP, .access = PL1_W, .accessfn = access_tsw }, + { .name = "DC_CIGSW", .state = ARM_CP_STATE_AA64, + .opc0 = 1, .opc1 = 0, .crn = 7, .crm = 14, .opc2 = 4, + .type = ARM_CP_NOP, .access = PL1_W, .accessfn = access_tsw }, + { .name = "DC_CIGDSW", .state = ARM_CP_STATE_AA64, + .opc0 = 1, .opc1 = 0, .crn = 7, .crm = 14, .opc2 = 6, + .type = ARM_CP_NOP, .access = PL1_W, .accessfn = access_tsw }, REGINFO_SENTINEL }; @@ -XXX,XX +XXX,XX @@ static const ARMCPRegInfo mte_tco_ro_reginfo[] = { .type = ARM_CP_CONST, .access = PL0_RW, }, REGINFO_SENTINEL }; + +static const ARMCPRegInfo mte_el0_cacheop_reginfo[] = { + { .name = "DC_CGVAC", .state = ARM_CP_STATE_AA64, + .opc0 = 1, .opc1 = 3, .crn = 7, .crm = 10, .opc2 = 3, + .type = ARM_CP_NOP, .access = PL0_W, + .accessfn = aa64_cacheop_poc_access }, + { .name = "DC_CGDVAC", .state = ARM_CP_STATE_AA64, + .opc0 = 1, .opc1 = 3, .crn = 7, .crm = 10, .opc2 = 5, + .type = ARM_CP_NOP, .access = PL0_W, + .accessfn = aa64_cacheop_poc_access }, + { .name = "DC_CGVAP", .state = ARM_CP_STATE_AA64, + .opc0 = 1, .opc1 = 3, .crn = 7, .crm = 12, .opc2 = 3, + .type = ARM_CP_NOP, .access = PL0_W, + .accessfn = aa64_cacheop_poc_access }, + { .name = "DC_CGDVAP", .state = ARM_CP_STATE_AA64, + .opc0 = 1, .opc1 = 3, .crn = 7, .crm = 12, .opc2 = 5, + .type = ARM_CP_NOP, .access = PL0_W, + .accessfn = aa64_cacheop_poc_access }, + { .name = "DC_CGVADP", .state = ARM_CP_STATE_AA64, + .opc0 = 1, .opc1 = 3, .crn = 7, .crm = 13, .opc2 = 3, + .type = ARM_CP_NOP, .access = PL0_W, + .accessfn = aa64_cacheop_poc_access }, + { .name = "DC_CGDVADP", .state = ARM_CP_STATE_AA64, + .opc0 = 1, .opc1 = 3, .crn = 7, .crm = 13, .opc2 = 5, + .type = ARM_CP_NOP, .access = PL0_W, + .accessfn = aa64_cacheop_poc_access }, + { .name = "DC_CIGVAC", .state = ARM_CP_STATE_AA64, + .opc0 = 1, .opc1 = 3, .crn = 7, .crm = 14, .opc2 = 3, + .type = ARM_CP_NOP, .access = PL0_W, + .accessfn = aa64_cacheop_poc_access }, + { .name = "DC_CIGDVAC", .state = ARM_CP_STATE_AA64, + .opc0 = 1, .opc1 = 3, .crn = 7, .crm = 14, .opc2 = 5, + .type = ARM_CP_NOP, .access = PL0_W, + .accessfn = aa64_cacheop_poc_access }, + REGINFO_SENTINEL +}; + #endif static CPAccessResult access_predinv(CPUARMState *env, const ARMCPRegInfo *ri, @@ -XXX,XX +XXX,XX @@ void register_cp_regs_for_features(ARMCPU *cpu) */ if (cpu_isar_feature(aa64_mte, cpu)) { define_arm_cp_regs(cpu, mte_reginfo); + define_arm_cp_regs(cpu, mte_el0_cacheop_reginfo); } else if (cpu_isar_feature(aa64_mte_insn_reg, cpu)) { define_arm_cp_regs(cpu, mte_tco_ro_reginfo); + define_arm_cp_regs(cpu, mte_el0_cacheop_reginfo); } #endif -- 2.20.1 From: Richard Henderson <richard.henderson@linaro.org> We will shortly need this in mte_helper.c as well. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200626033144.790098-22-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/internals.h | 36 ++++++++++++++++++++++++++++++++++++ target/arm/helper.c | 36 ------------------------------------ 2 files changed, 36 insertions(+), 36 deletions(-) diff --git a/target/arm/internals.h b/target/arm/internals.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/internals.h +++ b/target/arm/internals.h @@ -XXX,XX +XXX,XX @@ static inline bool regime_is_pan(CPUARMState *env, ARMMMUIdx mmu_idx) } } +/* Return the exception level which controls this address translation regime */ +static inline uint32_t regime_el(CPUARMState *env, ARMMMUIdx mmu_idx) +{ + switch (mmu_idx) { + case ARMMMUIdx_E20_0: + case ARMMMUIdx_E20_2: + case ARMMMUIdx_E20_2_PAN: + case ARMMMUIdx_Stage2: + case ARMMMUIdx_E2: + return 2; + case ARMMMUIdx_SE3: + return 3; + case ARMMMUIdx_SE10_0: + return arm_el_is_aa64(env, 3) ? 1 : 3; + case ARMMMUIdx_SE10_1: + case ARMMMUIdx_SE10_1_PAN: + case ARMMMUIdx_Stage1_E0: + case ARMMMUIdx_Stage1_E1: + case ARMMMUIdx_Stage1_E1_PAN: + case ARMMMUIdx_E10_0: + case ARMMMUIdx_E10_1: + case ARMMMUIdx_E10_1_PAN: + case ARMMMUIdx_MPrivNegPri: + case ARMMMUIdx_MUserNegPri: + case ARMMMUIdx_MPriv: + case ARMMMUIdx_MUser: + case ARMMMUIdx_MSPrivNegPri: + case ARMMMUIdx_MSUserNegPri: + case ARMMMUIdx_MSPriv: + case ARMMMUIdx_MSUser: + return 1; + default: + g_assert_not_reached(); + } +} + /* Return the FSR value for a debug exception (watchpoint, hardware * breakpoint or BKPT insn) targeting the specified exception level. */ diff --git a/target/arm/helper.c b/target/arm/helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper.c +++ b/target/arm/helper.c @@ -XXX,XX +XXX,XX @@ void arm_cpu_do_interrupt(CPUState *cs) } #endif /* !CONFIG_USER_ONLY */ -/* Return the exception level which controls this address translation regime */ -static uint32_t regime_el(CPUARMState *env, ARMMMUIdx mmu_idx) -{ - switch (mmu_idx) { - case ARMMMUIdx_E20_0: - case ARMMMUIdx_E20_2: - case ARMMMUIdx_E20_2_PAN: - case ARMMMUIdx_Stage2: - case ARMMMUIdx_E2: - return 2; - case ARMMMUIdx_SE3: - return 3; - case ARMMMUIdx_SE10_0: - return arm_el_is_aa64(env, 3) ? 1 : 3; - case ARMMMUIdx_SE10_1: - case ARMMMUIdx_SE10_1_PAN: - case ARMMMUIdx_Stage1_E0: - case ARMMMUIdx_Stage1_E1: - case ARMMMUIdx_Stage1_E1_PAN: - case ARMMMUIdx_E10_0: - case ARMMMUIdx_E10_1: - case ARMMMUIdx_E10_1_PAN: - case ARMMMUIdx_MPrivNegPri: - case ARMMMUIdx_MUserNegPri: - case ARMMMUIdx_MPriv: - case ARMMMUIdx_MUser: - case ARMMMUIdx_MSPrivNegPri: - case ARMMMUIdx_MSUserNegPri: - case ARMMMUIdx_MSPriv: - case ARMMMUIdx_MSUser: - return 1; - default: - g_assert_not_reached(); - } -} - uint64_t arm_sctlr(CPUARMState *env, int el) { /* Only EL0 needs to be adjusted for EL1&0 or EL2&0. */ -- 2.20.1 From: Richard Henderson <richard.henderson@linaro.org> We will shortly need this in mte_helper.c as well. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200626033144.790098-23-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/internals.h | 9 +++++++++ target/arm/helper.c | 9 --------- 2 files changed, 9 insertions(+), 9 deletions(-) diff --git a/target/arm/internals.h b/target/arm/internals.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/internals.h +++ b/target/arm/internals.h @@ -XXX,XX +XXX,XX @@ static inline uint32_t regime_el(CPUARMState *env, ARMMMUIdx mmu_idx) } } +/* Return the TCR controlling this translation regime */ +static inline TCR *regime_tcr(CPUARMState *env, ARMMMUIdx mmu_idx) +{ + if (mmu_idx == ARMMMUIdx_Stage2) { + return &env->cp15.vtcr_el2; + } + return &env->cp15.tcr_el[regime_el(env, mmu_idx)]; +} + /* Return the FSR value for a debug exception (watchpoint, hardware * breakpoint or BKPT insn) targeting the specified exception level. */ diff --git a/target/arm/helper.c b/target/arm/helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper.c +++ b/target/arm/helper.c @@ -XXX,XX +XXX,XX @@ static inline uint64_t regime_ttbr(CPUARMState *env, ARMMMUIdx mmu_idx, #endif /* !CONFIG_USER_ONLY */ -/* Return the TCR controlling this translation regime */ -static inline TCR *regime_tcr(CPUARMState *env, ARMMMUIdx mmu_idx) -{ - if (mmu_idx == ARMMMUIdx_Stage2) { - return &env->cp15.vtcr_el2; - } - return &env->cp15.tcr_el[regime_el(env, mmu_idx)]; -} - /* Convert a possible stage1+2 MMU index into the appropriate * stage 1 MMU index */ -- 2.20.1 From: Richard Henderson <richard.henderson@linaro.org> Replace existing uses of check_data_tbi in translate-a64.c that perform a single logical memory access. Leave the helper blank for now to reduce the patch size. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200626033144.790098-24-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/helper-a64.h | 1 + target/arm/internals.h | 8 +++ target/arm/translate-a64.h | 2 + target/arm/mte_helper.c | 8 +++ target/arm/translate-a64.c | 100 ++++++++++++++++++++++++++++--------- 5 files changed, 95 insertions(+), 24 deletions(-) diff --git a/target/arm/helper-a64.h b/target/arm/helper-a64.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper-a64.h +++ b/target/arm/helper-a64.h @@ -XXX,XX +XXX,XX @@ DEF_HELPER_FLAGS_3(autdb, TCG_CALL_NO_WG, i64, env, i64, i64) DEF_HELPER_FLAGS_2(xpaci, TCG_CALL_NO_RWG_SE, i64, env, i64) DEF_HELPER_FLAGS_2(xpacd, TCG_CALL_NO_RWG_SE, i64, env, i64) +DEF_HELPER_FLAGS_3(mte_check1, TCG_CALL_NO_WG, i64, env, i32, i64) DEF_HELPER_FLAGS_3(irg, TCG_CALL_NO_RWG, i64, env, i64, i64) DEF_HELPER_FLAGS_4(addsubg, TCG_CALL_NO_RWG_SE, i64, env, i64, s32, i32) DEF_HELPER_FLAGS_3(ldg, TCG_CALL_NO_WG, i64, env, i64, i64) diff --git a/target/arm/internals.h b/target/arm/internals.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/internals.h +++ b/target/arm/internals.h @@ -XXX,XX +XXX,XX @@ void arm_log_exception(int idx); #define LOG2_TAG_GRANULE 4 #define TAG_GRANULE (1 << LOG2_TAG_GRANULE) +/* Bits within a descriptor passed to the helper_mte_check* functions. */ +FIELD(MTEDESC, MIDX, 0, 4) +FIELD(MTEDESC, TBI, 4, 2) +FIELD(MTEDESC, TCMA, 6, 2) +FIELD(MTEDESC, WRITE, 8, 1) +FIELD(MTEDESC, ESIZE, 9, 5) +FIELD(MTEDESC, TSIZE, 14, 10) /* mte_checkN only */ + static inline int allocation_tag_from_addr(uint64_t ptr) { return extract64(ptr, 56, 4); diff --git a/target/arm/translate-a64.h b/target/arm/translate-a64.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-a64.h +++ b/target/arm/translate-a64.h @@ -XXX,XX +XXX,XX @@ TCGv_ptr get_fpstatus_ptr(bool); bool logic_imm_decode_wmask(uint64_t *result, unsigned int immn, unsigned int imms, unsigned int immr); bool sve_access_check(DisasContext *s); +TCGv_i64 gen_mte_check1(DisasContext *s, TCGv_i64 addr, bool is_write, + bool tag_checked, int log2_size); /* We should have at some point before trying to access an FP register * done the necessary access check, so assert that diff --git a/target/arm/mte_helper.c b/target/arm/mte_helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/mte_helper.c +++ b/target/arm/mte_helper.c @@ -XXX,XX +XXX,XX @@ void HELPER(stzgm_tags)(CPUARMState *env, uint64_t ptr, uint64_t val) memset(mem, tag_pair, tag_bytes); } } + +/* + * Perform an MTE checked access for a single logical or atomic access. + */ +uint64_t HELPER(mte_check1)(CPUARMState *env, uint32_t desc, uint64_t ptr) +{ + return ptr; +} diff --git a/target/arm/translate-a64.c b/target/arm/translate-a64.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-a64.c +++ b/target/arm/translate-a64.c @@ -XXX,XX +XXX,XX @@ static void gen_a64_set_pc(DisasContext *s, TCGv_i64 src) } /* - * Return a "clean" address for ADDR according to TBID. - * This is always a fresh temporary, as we need to be able to - * increment this independently of a dirty write-back address. + * Handle MTE and/or TBI. + * + * For TBI, ideally, we would do nothing. Proper behaviour on fault is + * for the tag to be present in the FAR_ELx register. But for user-only + * mode we do not have a TLB with which to implement this, so we must + * remove the top byte now. + * + * Always return a fresh temporary that we can increment independently + * of the write-back address. */ + static TCGv_i64 clean_data_tbi(DisasContext *s, TCGv_i64 addr) { TCGv_i64 clean = new_tmp_a64(s); - /* - * In order to get the correct value in the FAR_ELx register, - * we must present the memory subsystem with the "dirty" address - * including the TBI. In system mode we can make this work via - * the TLB, dropping the TBI during translation. But for user-only - * mode we don't have that option, and must remove the top byte now. - */ #ifdef CONFIG_USER_ONLY gen_top_byte_ignore(s, clean, addr, s->tbid); #else @@ -XXX,XX +XXX,XX @@ static void gen_probe_access(DisasContext *s, TCGv_i64 ptr, tcg_temp_free_i32(t_size); } +/* + * For MTE, check a single logical or atomic access. This probes a single + * address, the exact one specified. The size and alignment of the access + * is not relevant to MTE, per se, but watchpoints do require the size, + * and we want to recognize those before making any other changes to state. + */ +static TCGv_i64 gen_mte_check1_mmuidx(DisasContext *s, TCGv_i64 addr, + bool is_write, bool tag_checked, + int log2_size, bool is_unpriv, + int core_idx) +{ + if (tag_checked && s->mte_active[is_unpriv]) { + TCGv_i32 tcg_desc; + TCGv_i64 ret; + int desc = 0; + + desc = FIELD_DP32(desc, MTEDESC, MIDX, core_idx); + desc = FIELD_DP32(desc, MTEDESC, TBI, s->tbid); + desc = FIELD_DP32(desc, MTEDESC, TCMA, s->tcma); + desc = FIELD_DP32(desc, MTEDESC, WRITE, is_write); + desc = FIELD_DP32(desc, MTEDESC, ESIZE, 1 << log2_size); + tcg_desc = tcg_const_i32(desc); + + ret = new_tmp_a64(s); + gen_helper_mte_check1(ret, cpu_env, tcg_desc, addr); + tcg_temp_free_i32(tcg_desc); + + return ret; + } + return clean_data_tbi(s, addr); +} + +TCGv_i64 gen_mte_check1(DisasContext *s, TCGv_i64 addr, bool is_write, + bool tag_checked, int log2_size) +{ + return gen_mte_check1_mmuidx(s, addr, is_write, tag_checked, log2_size, + false, get_mem_index(s)); +} + typedef struct DisasCompare64 { TCGCond cond; TCGv_i64 value; @@ -XXX,XX +XXX,XX @@ static void gen_compare_and_swap(DisasContext *s, int rs, int rt, if (rn == 31) { gen_check_sp_alignment(s); } - clean_addr = clean_data_tbi(s, cpu_reg_sp(s, rn)); + clean_addr = gen_mte_check1(s, cpu_reg_sp(s, rn), true, rn != 31, size); tcg_gen_atomic_cmpxchg_i64(tcg_rs, clean_addr, tcg_rs, tcg_rt, memidx, size | MO_ALIGN | s->be_data); } @@ -XXX,XX +XXX,XX @@ static void gen_compare_and_swap_pair(DisasContext *s, int rs, int rt, if (rn == 31) { gen_check_sp_alignment(s); } - clean_addr = clean_data_tbi(s, cpu_reg_sp(s, rn)); + + /* This is a single atomic access, despite the "pair". */ + clean_addr = gen_mte_check1(s, cpu_reg_sp(s, rn), true, rn != 31, size + 1); if (size == 2) { TCGv_i64 cmp = tcg_temp_new_i64(); @@ -XXX,XX +XXX,XX @@ static void disas_ldst_excl(DisasContext *s, uint32_t insn) if (is_lasr) { tcg_gen_mb(TCG_MO_ALL | TCG_BAR_STRL); } - clean_addr = clean_data_tbi(s, cpu_reg_sp(s, rn)); + clean_addr = gen_mte_check1(s, cpu_reg_sp(s, rn), + true, rn != 31, size); gen_store_exclusive(s, rs, rt, rt2, clean_addr, size, false); return; @@ -XXX,XX +XXX,XX @@ static void disas_ldst_excl(DisasContext *s, uint32_t insn) if (rn == 31) { gen_check_sp_alignment(s); } - clean_addr = clean_data_tbi(s, cpu_reg_sp(s, rn)); + clean_addr = gen_mte_check1(s, cpu_reg_sp(s, rn), + false, rn != 31, size); s->is_ldex = true; gen_load_exclusive(s, rt, rt2, clean_addr, size, false); if (is_lasr) { @@ -XXX,XX +XXX,XX @@ static void disas_ldst_excl(DisasContext *s, uint32_t insn) gen_check_sp_alignment(s); } tcg_gen_mb(TCG_MO_ALL | TCG_BAR_STRL); - clean_addr = clean_data_tbi(s, cpu_reg_sp(s, rn)); + clean_addr = gen_mte_check1(s, cpu_reg_sp(s, rn), + true, rn != 31, size); do_gpr_st(s, cpu_reg(s, rt), clean_addr, size, true, rt, disas_ldst_compute_iss_sf(size, false, 0), is_lasr); return; @@ -XXX,XX +XXX,XX @@ static void disas_ldst_excl(DisasContext *s, uint32_t insn) if (rn == 31) { gen_check_sp_alignment(s); } - clean_addr = clean_data_tbi(s, cpu_reg_sp(s, rn)); + clean_addr = gen_mte_check1(s, cpu_reg_sp(s, rn), + false, rn != 31, size); do_gpr_ld(s, cpu_reg(s, rt), clean_addr, size, false, false, true, rt, disas_ldst_compute_iss_sf(size, false, 0), is_lasr); tcg_gen_mb(TCG_MO_ALL | TCG_BAR_LDAQ); @@ -XXX,XX +XXX,XX @@ static void disas_ldst_excl(DisasContext *s, uint32_t insn) if (is_lasr) { tcg_gen_mb(TCG_MO_ALL | TCG_BAR_STRL); } - clean_addr = clean_data_tbi(s, cpu_reg_sp(s, rn)); + clean_addr = gen_mte_check1(s, cpu_reg_sp(s, rn), + true, rn != 31, size); gen_store_exclusive(s, rs, rt, rt2, clean_addr, size, true); return; } @@ -XXX,XX +XXX,XX @@ static void disas_ldst_excl(DisasContext *s, uint32_t insn) if (rn == 31) { gen_check_sp_alignment(s); } - clean_addr = clean_data_tbi(s, cpu_reg_sp(s, rn)); + clean_addr = gen_mte_check1(s, cpu_reg_sp(s, rn), + false, rn != 31, size); s->is_ldex = true; gen_load_exclusive(s, rt, rt2, clean_addr, size, true); if (is_lasr) { @@ -XXX,XX +XXX,XX @@ static void disas_ldst_reg_imm9(DisasContext *s, uint32_t insn, bool iss_valid = !is_vector; bool post_index; bool writeback; + int memidx; TCGv_i64 clean_addr, dirty_addr; @@ -XXX,XX +XXX,XX @@ static void disas_ldst_reg_imm9(DisasContext *s, uint32_t insn, if (!post_index) { tcg_gen_addi_i64(dirty_addr, dirty_addr, imm9); } - clean_addr = clean_data_tbi(s, dirty_addr); + + memidx = is_unpriv ? get_a64_user_mem_index(s) : get_mem_index(s); + clean_addr = gen_mte_check1_mmuidx(s, dirty_addr, is_store, + writeback || rn != 31, + size, is_unpriv, memidx); if (is_vector) { if (is_store) { @@ -XXX,XX +XXX,XX @@ static void disas_ldst_reg_imm9(DisasContext *s, uint32_t insn, } } else { TCGv_i64 tcg_rt = cpu_reg(s, rt); - int memidx = is_unpriv ? get_a64_user_mem_index(s) : get_mem_index(s); bool iss_sf = disas_ldst_compute_iss_sf(size, is_signed, opc); if (is_store) { @@ -XXX,XX +XXX,XX @@ static void disas_ldst_reg_roffset(DisasContext *s, uint32_t insn, ext_and_shift_reg(tcg_rm, tcg_rm, opt, shift ? size : 0); tcg_gen_add_i64(dirty_addr, dirty_addr, tcg_rm); - clean_addr = clean_data_tbi(s, dirty_addr); + clean_addr = gen_mte_check1(s, dirty_addr, is_store, true, size); if (is_vector) { if (is_store) { @@ -XXX,XX +XXX,XX @@ static void disas_ldst_reg_unsigned_imm(DisasContext *s, uint32_t insn, dirty_addr = read_cpu_reg_sp(s, rn, 1); offset = imm12 << size; tcg_gen_addi_i64(dirty_addr, dirty_addr, offset); - clean_addr = clean_data_tbi(s, dirty_addr); + clean_addr = gen_mte_check1(s, dirty_addr, is_store, rn != 31, size); if (is_vector) { if (is_store) { @@ -XXX,XX +XXX,XX @@ static void disas_ldst_atomic(DisasContext *s, uint32_t insn, if (rn == 31) { gen_check_sp_alignment(s); } - clean_addr = clean_data_tbi(s, cpu_reg_sp(s, rn)); + clean_addr = gen_mte_check1(s, cpu_reg_sp(s, rn), false, rn != 31, size); if (o3_opc == 014) { /* @@ -XXX,XX +XXX,XX @@ static void disas_ldst_pac(DisasContext *s, uint32_t insn, tcg_gen_addi_i64(dirty_addr, dirty_addr, offset); /* Note that "clean" and "dirty" here refer to TBI not PAC. */ - clean_addr = clean_data_tbi(s, dirty_addr); + clean_addr = gen_mte_check1(s, dirty_addr, false, + is_wback || rn != 31, size); tcg_rt = cpu_reg(s, rt); do_gpr_ld(s, tcg_rt, clean_addr, size, /* is_signed */ false, -- 2.20.1 From: Richard Henderson <richard.henderson@linaro.org> Replace existing uses of check_data_tbi in translate-a64.c that perform multiple logical memory access. Leave the helper blank for now to reduce the patch size. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200626033144.790098-25-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/helper-a64.h | 1 + target/arm/translate-a64.h | 2 ++ target/arm/mte_helper.c | 8 +++++ target/arm/translate-a64.c | 71 +++++++++++++++++++++++++++++--------- 4 files changed, 66 insertions(+), 16 deletions(-) diff --git a/target/arm/helper-a64.h b/target/arm/helper-a64.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper-a64.h +++ b/target/arm/helper-a64.h @@ -XXX,XX +XXX,XX @@ DEF_HELPER_FLAGS_2(xpaci, TCG_CALL_NO_RWG_SE, i64, env, i64) DEF_HELPER_FLAGS_2(xpacd, TCG_CALL_NO_RWG_SE, i64, env, i64) DEF_HELPER_FLAGS_3(mte_check1, TCG_CALL_NO_WG, i64, env, i32, i64) +DEF_HELPER_FLAGS_3(mte_checkN, TCG_CALL_NO_WG, i64, env, i32, i64) DEF_HELPER_FLAGS_3(irg, TCG_CALL_NO_RWG, i64, env, i64, i64) DEF_HELPER_FLAGS_4(addsubg, TCG_CALL_NO_RWG_SE, i64, env, i64, s32, i32) DEF_HELPER_FLAGS_3(ldg, TCG_CALL_NO_WG, i64, env, i64, i64) diff --git a/target/arm/translate-a64.h b/target/arm/translate-a64.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-a64.h +++ b/target/arm/translate-a64.h @@ -XXX,XX +XXX,XX @@ bool logic_imm_decode_wmask(uint64_t *result, unsigned int immn, bool sve_access_check(DisasContext *s); TCGv_i64 gen_mte_check1(DisasContext *s, TCGv_i64 addr, bool is_write, bool tag_checked, int log2_size); +TCGv_i64 gen_mte_checkN(DisasContext *s, TCGv_i64 addr, bool is_write, + bool tag_checked, int count, int log2_esize); /* We should have at some point before trying to access an FP register * done the necessary access check, so assert that diff --git a/target/arm/mte_helper.c b/target/arm/mte_helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/mte_helper.c +++ b/target/arm/mte_helper.c @@ -XXX,XX +XXX,XX @@ uint64_t HELPER(mte_check1)(CPUARMState *env, uint32_t desc, uint64_t ptr) { return ptr; } + +/* + * Perform an MTE checked access for multiple logical accesses. + */ +uint64_t HELPER(mte_checkN)(CPUARMState *env, uint32_t desc, uint64_t ptr) +{ + return ptr; +} diff --git a/target/arm/translate-a64.c b/target/arm/translate-a64.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-a64.c +++ b/target/arm/translate-a64.c @@ -XXX,XX +XXX,XX @@ TCGv_i64 gen_mte_check1(DisasContext *s, TCGv_i64 addr, bool is_write, false, get_mem_index(s)); } +/* + * For MTE, check multiple logical sequential accesses. + */ +TCGv_i64 gen_mte_checkN(DisasContext *s, TCGv_i64 addr, bool is_write, + bool tag_checked, int log2_esize, int total_size) +{ + if (tag_checked && s->mte_active[0] && total_size != (1 << log2_esize)) { + TCGv_i32 tcg_desc; + TCGv_i64 ret; + int desc = 0; + + desc = FIELD_DP32(desc, MTEDESC, MIDX, get_mem_index(s)); + desc = FIELD_DP32(desc, MTEDESC, TBI, s->tbid); + desc = FIELD_DP32(desc, MTEDESC, TCMA, s->tcma); + desc = FIELD_DP32(desc, MTEDESC, WRITE, is_write); + desc = FIELD_DP32(desc, MTEDESC, ESIZE, 1 << log2_esize); + desc = FIELD_DP32(desc, MTEDESC, TSIZE, total_size); + tcg_desc = tcg_const_i32(desc); + + ret = new_tmp_a64(s); + gen_helper_mte_checkN(ret, cpu_env, tcg_desc, addr); + tcg_temp_free_i32(tcg_desc); + + return ret; + } + return gen_mte_check1(s, addr, is_write, tag_checked, log2_esize); +} + typedef struct DisasCompare64 { TCGCond cond; TCGv_i64 value; @@ -XXX,XX +XXX,XX @@ static void disas_ldst_pair(DisasContext *s, uint32_t insn) } } - clean_addr = clean_data_tbi(s, dirty_addr); + clean_addr = gen_mte_checkN(s, dirty_addr, !is_load, + (wback || rn != 31) && !set_tag, + size, 2 << size); + if (is_vector) { if (is_load) { do_fp_ld(s, rt, clean_addr, size); @@ -XXX,XX +XXX,XX @@ static void disas_ldst_multiple_struct(DisasContext *s, uint32_t insn) TCGv_i64 clean_addr, tcg_rn, tcg_ebytes; MemOp endian = s->be_data; - int ebytes; /* bytes per element */ + int total; /* total bytes */ int elements; /* elements per vector */ int rpt; /* num iterations */ int selem; /* structure elements */ @@ -XXX,XX +XXX,XX @@ static void disas_ldst_multiple_struct(DisasContext *s, uint32_t insn) endian = MO_LE; } - /* Consecutive little-endian elements from a single register + total = rpt * selem * (is_q ? 16 : 8); + tcg_rn = cpu_reg_sp(s, rn); + + /* + * Issue the MTE check vs the logical repeat count, before we + * promote consecutive little-endian elements below. + */ + clean_addr = gen_mte_checkN(s, tcg_rn, is_store, is_postidx || rn != 31, + size, total); + + /* + * Consecutive little-endian elements from a single register * can be promoted to a larger little-endian operation. */ if (selem == 1 && endian == MO_LE) { size = 3; } - ebytes = 1 << size; - elements = (is_q ? 16 : 8) / ebytes; - - tcg_rn = cpu_reg_sp(s, rn); - clean_addr = clean_data_tbi(s, tcg_rn); - tcg_ebytes = tcg_const_i64(ebytes); + elements = (is_q ? 16 : 8) >> size; + tcg_ebytes = tcg_const_i64(1 << size); for (r = 0; r < rpt; r++) { int e; for (e = 0; e < elements; e++) { @@ -XXX,XX +XXX,XX @@ static void disas_ldst_multiple_struct(DisasContext *s, uint32_t insn) if (is_postidx) { if (rm == 31) { - tcg_gen_addi_i64(tcg_rn, tcg_rn, rpt * elements * selem * ebytes); + tcg_gen_addi_i64(tcg_rn, tcg_rn, total); } else { tcg_gen_add_i64(tcg_rn, tcg_rn, cpu_reg(s, rm)); } @@ -XXX,XX +XXX,XX @@ static void disas_ldst_single_struct(DisasContext *s, uint32_t insn) int selem = (extract32(opc, 0, 1) << 1 | R) + 1; bool replicate = false; int index = is_q << 3 | S << 2 | size; - int ebytes, xs; + int xs, total; TCGv_i64 clean_addr, tcg_rn, tcg_ebytes; if (extract32(insn, 31, 1)) { @@ -XXX,XX +XXX,XX @@ static void disas_ldst_single_struct(DisasContext *s, uint32_t insn) return; } - ebytes = 1 << scale; - if (rn == 31) { gen_check_sp_alignment(s); } + total = selem << scale; tcg_rn = cpu_reg_sp(s, rn); - clean_addr = clean_data_tbi(s, tcg_rn); - tcg_ebytes = tcg_const_i64(ebytes); + clean_addr = gen_mte_checkN(s, tcg_rn, !is_load, is_postidx || rn != 31, + scale, total); + + tcg_ebytes = tcg_const_i64(1 << scale); for (xs = 0; xs < selem; xs++) { if (replicate) { /* Load and replicate to all elements */ @@ -XXX,XX +XXX,XX @@ static void disas_ldst_single_struct(DisasContext *s, uint32_t insn) if (is_postidx) { if (rm == 31) { - tcg_gen_addi_i64(tcg_rn, tcg_rn, selem * ebytes); + tcg_gen_addi_i64(tcg_rn, tcg_rn, total); } else { tcg_gen_add_i64(tcg_rn, tcg_rn, cpu_reg(s, rm)); } -- 2.20.1 From: Richard Henderson <richard.henderson@linaro.org> Fill out the stub that was added earlier. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200626033144.790098-26-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/internals.h | 48 +++++++++++++++ target/arm/mte_helper.c | 132 +++++++++++++++++++++++++++++++++++++++- 2 files changed, 179 insertions(+), 1 deletion(-) diff --git a/target/arm/internals.h b/target/arm/internals.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/internals.h +++ b/target/arm/internals.h @@ -XXX,XX +XXX,XX @@ FIELD(MTEDESC, WRITE, 8, 1) FIELD(MTEDESC, ESIZE, 9, 5) FIELD(MTEDESC, TSIZE, 14, 10) /* mte_checkN only */ +bool mte_probe1(CPUARMState *env, uint32_t desc, uint64_t ptr); +uint64_t mte_check1(CPUARMState *env, uint32_t desc, + uint64_t ptr, uintptr_t ra); + static inline int allocation_tag_from_addr(uint64_t ptr) { return extract64(ptr, 56, 4); @@ -XXX,XX +XXX,XX @@ static inline uint64_t address_with_allocation_tag(uint64_t ptr, int rtag) return deposit64(ptr, 56, 4, rtag); } +/* Return true if tbi bits mean that the access is checked. */ +static inline bool tbi_check(uint32_t desc, int bit55) +{ + return (desc >> (R_MTEDESC_TBI_SHIFT + bit55)) & 1; +} + +/* Return true if tcma bits mean that the access is unchecked. */ +static inline bool tcma_check(uint32_t desc, int bit55, int ptr_tag) +{ + /* + * We had extracted bit55 and ptr_tag for other reasons, so fold + * (ptr<59:55> == 00000 || ptr<59:55> == 11111) into a single test. + */ + bool match = ((ptr_tag + bit55) & 0xf) == 0; + bool tcma = (desc >> (R_MTEDESC_TCMA_SHIFT + bit55)) & 1; + return tcma && match; +} + +/* + * For TBI, ideally, we would do nothing. Proper behaviour on fault is + * for the tag to be present in the FAR_ELx register. But for user-only + * mode, we do not have a TLB with which to implement this, so we must + * remove the top byte. + */ +static inline uint64_t useronly_clean_ptr(uint64_t ptr) +{ + /* TBI is known to be enabled. */ +#ifdef CONFIG_USER_ONLY + ptr = sextract64(ptr, 0, 56); +#endif + return ptr; +} + +static inline uint64_t useronly_maybe_clean_ptr(uint32_t desc, uint64_t ptr) +{ +#ifdef CONFIG_USER_ONLY + int64_t clean_ptr = sextract64(ptr, 0, 56); + if (tbi_check(desc, clean_ptr < 0)) { + ptr = clean_ptr; + } +#endif + return ptr; +} + #endif diff --git a/target/arm/mte_helper.c b/target/arm/mte_helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/mte_helper.c +++ b/target/arm/mte_helper.c @@ -XXX,XX +XXX,XX @@ void HELPER(stzgm_tags)(CPUARMState *env, uint64_t ptr, uint64_t val) } } +/* Record a tag check failure. */ +static void mte_check_fail(CPUARMState *env, int mmu_idx, + uint64_t dirty_ptr, uintptr_t ra) +{ + ARMMMUIdx arm_mmu_idx = core_to_aa64_mmu_idx(mmu_idx); + int el, reg_el, tcf, select; + uint64_t sctlr; + + reg_el = regime_el(env, arm_mmu_idx); + sctlr = env->cp15.sctlr_el[reg_el]; + + switch (arm_mmu_idx) { + case ARMMMUIdx_E10_0: + case ARMMMUIdx_E20_0: + el = 0; + tcf = extract64(sctlr, 38, 2); + break; + default: + el = reg_el; + tcf = extract64(sctlr, 40, 2); + } + + switch (tcf) { + case 1: + /* + * Tag check fail causes a synchronous exception. + * + * In restore_state_to_opc, we set the exception syndrome + * for the load or store operation. Unwind first so we + * may overwrite that with the syndrome for the tag check. + */ + cpu_restore_state(env_cpu(env), ra, true); + env->exception.vaddress = dirty_ptr; + raise_exception(env, EXCP_DATA_ABORT, + syn_data_abort_no_iss(el != 0, 0, 0, 0, 0, 0, 0x11), + exception_target_el(env)); + /* noreturn, but fall through to the assert anyway */ + + case 0: + /* + * Tag check fail does not affect the PE. + * We eliminate this case by not setting MTE_ACTIVE + * in tb_flags, so that we never make this runtime call. + */ + g_assert_not_reached(); + + case 2: + /* Tag check fail causes asynchronous flag set. */ + mmu_idx = arm_mmu_idx_el(env, el); + if (regime_has_2_ranges(mmu_idx)) { + select = extract64(dirty_ptr, 55, 1); + } else { + select = 0; + } + env->cp15.tfsr_el[el] |= 1 << select; + break; + + default: + /* Case 3: Reserved. */ + qemu_log_mask(LOG_GUEST_ERROR, + "Tag check failure with SCTLR_EL%d.TCF%s " + "set to reserved value %d\n", + reg_el, el ? "" : "0", tcf); + break; + } +} + /* * Perform an MTE checked access for a single logical or atomic access. */ +static bool mte_probe1_int(CPUARMState *env, uint32_t desc, uint64_t ptr, + uintptr_t ra, int bit55) +{ + int mem_tag, mmu_idx, ptr_tag, size; + MMUAccessType type; + uint8_t *mem; + + ptr_tag = allocation_tag_from_addr(ptr); + + if (tcma_check(desc, bit55, ptr_tag)) { + return true; + } + + mmu_idx = FIELD_EX32(desc, MTEDESC, MIDX); + type = FIELD_EX32(desc, MTEDESC, WRITE) ? MMU_DATA_STORE : MMU_DATA_LOAD; + size = FIELD_EX32(desc, MTEDESC, ESIZE); + + mem = allocation_tag_mem(env, mmu_idx, ptr, type, size, + MMU_DATA_LOAD, 1, ra); + if (!mem) { + return true; + } + + mem_tag = load_tag1(ptr, mem); + return ptr_tag == mem_tag; +} + +/* + * No-fault version of mte_check1, to be used by SVE for MemSingleNF. + * Returns false if the access is Checked and the check failed. This + * is only intended to probe the tag -- the validity of the page must + * be checked beforehand. + */ +bool mte_probe1(CPUARMState *env, uint32_t desc, uint64_t ptr) +{ + int bit55 = extract64(ptr, 55, 1); + + /* If TBI is disabled, the access is unchecked. */ + if (unlikely(!tbi_check(desc, bit55))) { + return true; + } + + return mte_probe1_int(env, desc, ptr, 0, bit55); +} + +uint64_t mte_check1(CPUARMState *env, uint32_t desc, + uint64_t ptr, uintptr_t ra) +{ + int bit55 = extract64(ptr, 55, 1); + + /* If TBI is disabled, the access is unchecked, and ptr is not dirty. */ + if (unlikely(!tbi_check(desc, bit55))) { + return ptr; + } + + if (unlikely(!mte_probe1_int(env, desc, ptr, ra, bit55))) { + int mmu_idx = FIELD_EX32(desc, MTEDESC, MIDX); + mte_check_fail(env, mmu_idx, ptr, ra); + } + + return useronly_clean_ptr(ptr); +} + uint64_t HELPER(mte_check1)(CPUARMState *env, uint32_t desc, uint64_t ptr) { - return ptr; + return mte_check1(env, desc, ptr, GETPC()); } /* -- 2.20.1 From: Richard Henderson <richard.henderson@linaro.org> Fill out the stub that was added earlier. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200626033144.790098-27-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/internals.h | 2 + target/arm/mte_helper.c | 165 +++++++++++++++++++++++++++++++++++++++- 2 files changed, 166 insertions(+), 1 deletion(-) diff --git a/target/arm/internals.h b/target/arm/internals.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/internals.h +++ b/target/arm/internals.h @@ -XXX,XX +XXX,XX @@ FIELD(MTEDESC, TSIZE, 14, 10) /* mte_checkN only */ bool mte_probe1(CPUARMState *env, uint32_t desc, uint64_t ptr); uint64_t mte_check1(CPUARMState *env, uint32_t desc, uint64_t ptr, uintptr_t ra); +uint64_t mte_checkN(CPUARMState *env, uint32_t desc, + uint64_t ptr, uintptr_t ra); static inline int allocation_tag_from_addr(uint64_t ptr) { diff --git a/target/arm/mte_helper.c b/target/arm/mte_helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/mte_helper.c +++ b/target/arm/mte_helper.c @@ -XXX,XX +XXX,XX @@ uint64_t HELPER(mte_check1)(CPUARMState *env, uint32_t desc, uint64_t ptr) /* * Perform an MTE checked access for multiple logical accesses. */ + +/** + * checkN: + * @tag: tag memory to test + * @odd: true to begin testing at tags at odd nibble + * @cmp: the tag to compare against + * @count: number of tags to test + * + * Return the number of successful tests. + * Thus a return value < @count indicates a failure. + * + * A note about sizes: count is expected to be small. + * + * The most common use will be LDP/STP of two integer registers, + * which means 16 bytes of memory touching at most 2 tags, but + * often the access is aligned and thus just 1 tag. + * + * Using AdvSIMD LD/ST (multiple), one can access 64 bytes of memory, + * touching at most 5 tags. SVE LDR/STR (vector) with the default + * vector length is also 64 bytes; the maximum architectural length + * is 256 bytes touching at most 9 tags. + * + * The loop below uses 7 logical operations and 1 memory operation + * per tag pair. An implementation that loads an aligned word and + * uses masking to ignore adjacent tags requires 18 logical operations + * and thus does not begin to pay off until 6 tags. + * Which, according to the survey above, is unlikely to be common. + */ +static int checkN(uint8_t *mem, int odd, int cmp, int count) +{ + int n = 0, diff; + + /* Replicate the test tag and compare. */ + cmp *= 0x11; + diff = *mem++ ^ cmp; + + if (odd) { + goto start_odd; + } + + while (1) { + /* Test even tag. */ + if (unlikely((diff) & 0x0f)) { + break; + } + if (++n == count) { + break; + } + + start_odd: + /* Test odd tag. */ + if (unlikely((diff) & 0xf0)) { + break; + } + if (++n == count) { + break; + } + + diff = *mem++ ^ cmp; + } + return n; +} + +uint64_t mte_checkN(CPUARMState *env, uint32_t desc, + uint64_t ptr, uintptr_t ra) +{ + int mmu_idx, ptr_tag, bit55; + uint64_t ptr_last, ptr_end, prev_page, next_page; + uint64_t tag_first, tag_end; + uint64_t tag_byte_first, tag_byte_end; + uint32_t esize, total, tag_count, tag_size, n, c; + uint8_t *mem1, *mem2; + MMUAccessType type; + + bit55 = extract64(ptr, 55, 1); + + /* If TBI is disabled, the access is unchecked, and ptr is not dirty. */ + if (unlikely(!tbi_check(desc, bit55))) { + return ptr; + } + + ptr_tag = allocation_tag_from_addr(ptr); + + if (tcma_check(desc, bit55, ptr_tag)) { + goto done; + } + + mmu_idx = FIELD_EX32(desc, MTEDESC, MIDX); + type = FIELD_EX32(desc, MTEDESC, WRITE) ? MMU_DATA_STORE : MMU_DATA_LOAD; + esize = FIELD_EX32(desc, MTEDESC, ESIZE); + total = FIELD_EX32(desc, MTEDESC, TSIZE); + + /* Find the addr of the end of the access, and of the last element. */ + ptr_end = ptr + total; + ptr_last = ptr_end - esize; + + /* Round the bounds to the tag granule, and compute the number of tags. */ + tag_first = QEMU_ALIGN_DOWN(ptr, TAG_GRANULE); + tag_end = QEMU_ALIGN_UP(ptr_last, TAG_GRANULE); + tag_count = (tag_end - tag_first) / TAG_GRANULE; + + /* Round the bounds to twice the tag granule, and compute the bytes. */ + tag_byte_first = QEMU_ALIGN_DOWN(ptr, 2 * TAG_GRANULE); + tag_byte_end = QEMU_ALIGN_UP(ptr_last, 2 * TAG_GRANULE); + + /* Locate the page boundaries. */ + prev_page = ptr & TARGET_PAGE_MASK; + next_page = prev_page + TARGET_PAGE_SIZE; + + if (likely(tag_end - prev_page <= TARGET_PAGE_SIZE)) { + /* Memory access stays on one page. */ + tag_size = (tag_byte_end - tag_byte_first) / (2 * TAG_GRANULE); + mem1 = allocation_tag_mem(env, mmu_idx, ptr, type, total, + MMU_DATA_LOAD, tag_size, ra); + if (!mem1) { + goto done; + } + /* Perform all of the comparisons. */ + n = checkN(mem1, ptr & TAG_GRANULE, ptr_tag, tag_count); + } else { + /* Memory access crosses to next page. */ + tag_size = (next_page - tag_byte_first) / (2 * TAG_GRANULE); + mem1 = allocation_tag_mem(env, mmu_idx, ptr, type, next_page - ptr, + MMU_DATA_LOAD, tag_size, ra); + + tag_size = (tag_byte_end - next_page) / (2 * TAG_GRANULE); + mem2 = allocation_tag_mem(env, mmu_idx, next_page, type, + ptr_end - next_page, + MMU_DATA_LOAD, tag_size, ra); + + /* + * Perform all of the comparisons. + * Note the possible but unlikely case of the operation spanning + * two pages that do not both have tagging enabled. + */ + n = c = (next_page - tag_first) / TAG_GRANULE; + if (mem1) { + n = checkN(mem1, ptr & TAG_GRANULE, ptr_tag, c); + } + if (n == c) { + if (!mem2) { + goto done; + } + n += checkN(mem2, 0, ptr_tag, tag_count - c); + } + } + + /* + * If we failed, we know which granule. Compute the element that + * is first in that granule, and signal failure on that element. + */ + if (unlikely(n < tag_count)) { + uint64_t fail_ofs; + + fail_ofs = tag_first + n * TAG_GRANULE - ptr; + fail_ofs = ROUND_UP(fail_ofs, esize); + mte_check_fail(env, mmu_idx, ptr + fail_ofs, ra); + } + + done: + return useronly_clean_ptr(ptr); +} + uint64_t HELPER(mte_checkN)(CPUARMState *env, uint32_t desc, uint64_t ptr) { - return ptr; + return mte_checkN(env, desc, ptr, GETPC()); } -- 2.20.1 From: Richard Henderson <richard.henderson@linaro.org> Use a special helper for DC_ZVA, rather than the more general mte_checkN. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200626033144.790098-28-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/helper-a64.h | 1 + target/arm/mte_helper.c | 106 +++++++++++++++++++++++++++++++++++++ target/arm/translate-a64.c | 16 +++++- 3 files changed, 122 insertions(+), 1 deletion(-) diff --git a/target/arm/helper-a64.h b/target/arm/helper-a64.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper-a64.h +++ b/target/arm/helper-a64.h @@ -XXX,XX +XXX,XX @@ DEF_HELPER_FLAGS_2(xpacd, TCG_CALL_NO_RWG_SE, i64, env, i64) DEF_HELPER_FLAGS_3(mte_check1, TCG_CALL_NO_WG, i64, env, i32, i64) DEF_HELPER_FLAGS_3(mte_checkN, TCG_CALL_NO_WG, i64, env, i32, i64) +DEF_HELPER_FLAGS_3(mte_check_zva, TCG_CALL_NO_WG, i64, env, i32, i64) DEF_HELPER_FLAGS_3(irg, TCG_CALL_NO_RWG, i64, env, i64, i64) DEF_HELPER_FLAGS_4(addsubg, TCG_CALL_NO_RWG_SE, i64, env, i64, s32, i32) DEF_HELPER_FLAGS_3(ldg, TCG_CALL_NO_WG, i64, env, i64, i64) diff --git a/target/arm/mte_helper.c b/target/arm/mte_helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/mte_helper.c +++ b/target/arm/mte_helper.c @@ -XXX,XX +XXX,XX @@ uint64_t HELPER(mte_checkN)(CPUARMState *env, uint32_t desc, uint64_t ptr) { return mte_checkN(env, desc, ptr, GETPC()); } + +/* + * Perform an MTE checked access for DC_ZVA. + */ +uint64_t HELPER(mte_check_zva)(CPUARMState *env, uint32_t desc, uint64_t ptr) +{ + uintptr_t ra = GETPC(); + int log2_dcz_bytes, log2_tag_bytes; + int mmu_idx, bit55; + intptr_t dcz_bytes, tag_bytes, i; + void *mem; + uint64_t ptr_tag, mem_tag, align_ptr; + + bit55 = extract64(ptr, 55, 1); + + /* If TBI is disabled, the access is unchecked, and ptr is not dirty. */ + if (unlikely(!tbi_check(desc, bit55))) { + return ptr; + } + + ptr_tag = allocation_tag_from_addr(ptr); + + if (tcma_check(desc, bit55, ptr_tag)) { + goto done; + } + + /* + * In arm_cpu_realizefn, we asserted that dcz > LOG2_TAG_GRANULE+1, + * i.e. 32 bytes, which is an unreasonably small dcz anyway, to make + * sure that we can access one complete tag byte here. + */ + log2_dcz_bytes = env_archcpu(env)->dcz_blocksize + 2; + log2_tag_bytes = log2_dcz_bytes - (LOG2_TAG_GRANULE + 1); + dcz_bytes = (intptr_t)1 << log2_dcz_bytes; + tag_bytes = (intptr_t)1 << log2_tag_bytes; + align_ptr = ptr & -dcz_bytes; + + /* + * Trap if accessing an invalid page. DC_ZVA requires that we supply + * the original pointer for an invalid page. But watchpoints require + * that we probe the actual space. So do both. + */ + mmu_idx = FIELD_EX32(desc, MTEDESC, MIDX); + (void) probe_write(env, ptr, 1, mmu_idx, ra); + mem = allocation_tag_mem(env, mmu_idx, align_ptr, MMU_DATA_STORE, + dcz_bytes, MMU_DATA_LOAD, tag_bytes, ra); + if (!mem) { + goto done; + } + + /* + * Unlike the reasoning for checkN, DC_ZVA is always aligned, and thus + * it is quite easy to perform all of the comparisons at once without + * any extra masking. + * + * The most common zva block size is 64; some of the thunderx cpus use + * a block size of 128. For user-only, aarch64_max_initfn will set the + * block size to 512. Fill out the other cases for future-proofing. + * + * In order to be able to find the first miscompare later, we want the + * tag bytes to be in little-endian order. + */ + switch (log2_tag_bytes) { + case 0: /* zva_blocksize 32 */ + mem_tag = *(uint8_t *)mem; + ptr_tag *= 0x11u; + break; + case 1: /* zva_blocksize 64 */ + mem_tag = cpu_to_le16(*(uint16_t *)mem); + ptr_tag *= 0x1111u; + break; + case 2: /* zva_blocksize 128 */ + mem_tag = cpu_to_le32(*(uint32_t *)mem); + ptr_tag *= 0x11111111u; + break; + case 3: /* zva_blocksize 256 */ + mem_tag = cpu_to_le64(*(uint64_t *)mem); + ptr_tag *= 0x1111111111111111ull; + break; + + default: /* zva_blocksize 512, 1024, 2048 */ + ptr_tag *= 0x1111111111111111ull; + i = 0; + do { + mem_tag = cpu_to_le64(*(uint64_t *)(mem + i)); + if (unlikely(mem_tag != ptr_tag)) { + goto fail; + } + i += 8; + align_ptr += 16 * TAG_GRANULE; + } while (i < tag_bytes); + goto done; + } + + if (likely(mem_tag == ptr_tag)) { + goto done; + } + + fail: + /* Locate the first nibble that differs. */ + i = ctz64(mem_tag ^ ptr_tag) >> 4; + mte_check_fail(env, mmu_idx, align_ptr + i * TAG_GRANULE, ra); + + done: + return useronly_clean_ptr(ptr); +} diff --git a/target/arm/translate-a64.c b/target/arm/translate-a64.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-a64.c +++ b/target/arm/translate-a64.c @@ -XXX,XX +XXX,XX @@ static void handle_sys(DisasContext *s, uint32_t insn, bool isread, return; case ARM_CP_DC_ZVA: /* Writes clear the aligned block of memory which rt points into. */ - tcg_rt = clean_data_tbi(s, cpu_reg(s, rt)); + if (s->mte_active[0]) { + TCGv_i32 t_desc; + int desc = 0; + + desc = FIELD_DP32(desc, MTEDESC, MIDX, get_mem_index(s)); + desc = FIELD_DP32(desc, MTEDESC, TBI, s->tbid); + desc = FIELD_DP32(desc, MTEDESC, TCMA, s->tcma); + t_desc = tcg_const_i32(desc); + + tcg_rt = new_tmp_a64(s); + gen_helper_mte_check_zva(tcg_rt, cpu_env, t_desc, cpu_reg(s, rt)); + tcg_temp_free_i32(t_desc); + } else { + tcg_rt = clean_data_tbi(s, cpu_reg(s, rt)); + } gen_helper_dc_zva(cpu_env, tcg_rt); return; default: -- 2.20.1 From: Richard Henderson <richard.henderson@linaro.org> Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200626033144.790098-29-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/translate-sve.c | 61 +++++++++++++++++++++----------------- 1 file changed, 33 insertions(+), 28 deletions(-) diff --git a/target/arm/translate-sve.c b/target/arm/translate-sve.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-sve.c +++ b/target/arm/translate-sve.c @@ -XXX,XX +XXX,XX @@ static void do_ldr(DisasContext *s, uint32_t vofs, int len, int rn, int imm) int len_remain = len % 8; int nparts = len / 8 + ctpop8(len_remain); int midx = get_mem_index(s); - TCGv_i64 addr, t0, t1; + TCGv_i64 dirty_addr, clean_addr, t0, t1; - addr = tcg_temp_new_i64(); - t0 = tcg_temp_new_i64(); + dirty_addr = tcg_temp_new_i64(); + tcg_gen_addi_i64(dirty_addr, cpu_reg_sp(s, rn), imm); + clean_addr = gen_mte_checkN(s, dirty_addr, false, rn != 31, len, MO_8); + tcg_temp_free_i64(dirty_addr); - /* Note that unpredicated load/store of vector/predicate registers + /* + * Note that unpredicated load/store of vector/predicate registers * are defined as a stream of bytes, which equates to little-endian - * operations on larger quantities. There is no nice way to force - * a little-endian load for aarch64_be-linux-user out of line. - * + * operations on larger quantities. * Attempt to keep code expansion to a minimum by limiting the * amount of unrolling done. */ if (nparts <= 4) { int i; + t0 = tcg_temp_new_i64(); for (i = 0; i < len_align; i += 8) { - tcg_gen_addi_i64(addr, cpu_reg_sp(s, rn), imm + i); - tcg_gen_qemu_ld_i64(t0, addr, midx, MO_LEQ); + tcg_gen_qemu_ld_i64(t0, clean_addr, midx, MO_LEQ); tcg_gen_st_i64(t0, cpu_env, vofs + i); + tcg_gen_addi_i64(clean_addr, cpu_reg_sp(s, rn), 8); } + tcg_temp_free_i64(t0); } else { TCGLabel *loop = gen_new_label(); TCGv_ptr tp, i = tcg_const_local_ptr(0); + /* Copy the clean address into a local temp, live across the loop. */ + t0 = clean_addr; + clean_addr = tcg_temp_local_new_i64(); + tcg_gen_mov_i64(clean_addr, t0); + tcg_temp_free_i64(t0); + gen_set_label(loop); - /* Minimize the number of local temps that must be re-read from - * the stack each iteration. Instead, re-compute values other - * than the loop counter. - */ + t0 = tcg_temp_new_i64(); + tcg_gen_qemu_ld_i64(t0, clean_addr, midx, MO_LEQ); + tcg_gen_addi_i64(clean_addr, clean_addr, 8); + tp = tcg_temp_new_ptr(); - tcg_gen_addi_ptr(tp, i, imm); - tcg_gen_extu_ptr_i64(addr, tp); - tcg_gen_add_i64(addr, addr, cpu_reg_sp(s, rn)); - - tcg_gen_qemu_ld_i64(t0, addr, midx, MO_LEQ); - tcg_gen_add_ptr(tp, cpu_env, i); tcg_gen_addi_ptr(i, i, 8); tcg_gen_st_i64(t0, tp, vofs); tcg_temp_free_ptr(tp); + tcg_temp_free_i64(t0); tcg_gen_brcondi_ptr(TCG_COND_LTU, i, len_align, loop); tcg_temp_free_ptr(i); } - /* Predicate register loads can be any multiple of 2. + /* + * Predicate register loads can be any multiple of 2. * Note that we still store the entire 64-bit unit into cpu_env. */ if (len_remain) { - tcg_gen_addi_i64(addr, cpu_reg_sp(s, rn), imm + len_align); - + t0 = tcg_temp_new_i64(); switch (len_remain) { case 2: case 4: case 8: - tcg_gen_qemu_ld_i64(t0, addr, midx, MO_LE | ctz32(len_remain)); + tcg_gen_qemu_ld_i64(t0, clean_addr, midx, + MO_LE | ctz32(len_remain)); break; case 6: t1 = tcg_temp_new_i64(); - tcg_gen_qemu_ld_i64(t0, addr, midx, MO_LEUL); - tcg_gen_addi_i64(addr, addr, 4); - tcg_gen_qemu_ld_i64(t1, addr, midx, MO_LEUW); + tcg_gen_qemu_ld_i64(t0, clean_addr, midx, MO_LEUL); + tcg_gen_addi_i64(clean_addr, clean_addr, 4); + tcg_gen_qemu_ld_i64(t1, clean_addr, midx, MO_LEUW); tcg_gen_deposit_i64(t0, t0, t1, 32, 32); tcg_temp_free_i64(t1); break; @@ -XXX,XX +XXX,XX @@ static void do_ldr(DisasContext *s, uint32_t vofs, int len, int rn, int imm) g_assert_not_reached(); } tcg_gen_st_i64(t0, cpu_env, vofs + len_align); + tcg_temp_free_i64(t0); } - tcg_temp_free_i64(addr); - tcg_temp_free_i64(t0); + tcg_temp_free_i64(clean_addr); } /* Similarly for stores. */ -- 2.20.1 From: Richard Henderson <richard.henderson@linaro.org> Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200626033144.790098-30-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/translate-sve.c | 61 +++++++++++++++++++++----------------- 1 file changed, 33 insertions(+), 28 deletions(-) diff --git a/target/arm/translate-sve.c b/target/arm/translate-sve.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-sve.c +++ b/target/arm/translate-sve.c @@ -XXX,XX +XXX,XX @@ static void do_str(DisasContext *s, uint32_t vofs, int len, int rn, int imm) int len_remain = len % 8; int nparts = len / 8 + ctpop8(len_remain); int midx = get_mem_index(s); - TCGv_i64 addr, t0; + TCGv_i64 dirty_addr, clean_addr, t0; - addr = tcg_temp_new_i64(); - t0 = tcg_temp_new_i64(); + dirty_addr = tcg_temp_new_i64(); + tcg_gen_addi_i64(dirty_addr, cpu_reg_sp(s, rn), imm); + clean_addr = gen_mte_checkN(s, dirty_addr, false, rn != 31, len, MO_8); + tcg_temp_free_i64(dirty_addr); /* Note that unpredicated load/store of vector/predicate registers * are defined as a stream of bytes, which equates to little-endian @@ -XXX,XX +XXX,XX @@ static void do_str(DisasContext *s, uint32_t vofs, int len, int rn, int imm) if (nparts <= 4) { int i; + t0 = tcg_temp_new_i64(); for (i = 0; i < len_align; i += 8) { tcg_gen_ld_i64(t0, cpu_env, vofs + i); - tcg_gen_addi_i64(addr, cpu_reg_sp(s, rn), imm + i); - tcg_gen_qemu_st_i64(t0, addr, midx, MO_LEQ); + tcg_gen_qemu_st_i64(t0, clean_addr, midx, MO_LEQ); + tcg_gen_addi_i64(clean_addr, cpu_reg_sp(s, rn), 8); } + tcg_temp_free_i64(t0); } else { TCGLabel *loop = gen_new_label(); - TCGv_ptr t2, i = tcg_const_local_ptr(0); + TCGv_ptr tp, i = tcg_const_local_ptr(0); + + /* Copy the clean address into a local temp, live across the loop. */ + t0 = clean_addr; + clean_addr = tcg_temp_local_new_i64(); + tcg_gen_mov_i64(clean_addr, t0); + tcg_temp_free_i64(t0); gen_set_label(loop); - t2 = tcg_temp_new_ptr(); - tcg_gen_add_ptr(t2, cpu_env, i); - tcg_gen_ld_i64(t0, t2, vofs); - - /* Minimize the number of local temps that must be re-read from - * the stack each iteration. Instead, re-compute values other - * than the loop counter. - */ - tcg_gen_addi_ptr(t2, i, imm); - tcg_gen_extu_ptr_i64(addr, t2); - tcg_gen_add_i64(addr, addr, cpu_reg_sp(s, rn)); - tcg_temp_free_ptr(t2); - - tcg_gen_qemu_st_i64(t0, addr, midx, MO_LEQ); - + t0 = tcg_temp_new_i64(); + tp = tcg_temp_new_ptr(); + tcg_gen_add_ptr(tp, cpu_env, i); + tcg_gen_ld_i64(t0, tp, vofs); tcg_gen_addi_ptr(i, i, 8); + tcg_temp_free_ptr(tp); + + tcg_gen_qemu_st_i64(t0, clean_addr, midx, MO_LEQ); + tcg_gen_addi_i64(clean_addr, clean_addr, 8); + tcg_temp_free_i64(t0); tcg_gen_brcondi_ptr(TCG_COND_LTU, i, len_align, loop); tcg_temp_free_ptr(i); @@ -XXX,XX +XXX,XX @@ static void do_str(DisasContext *s, uint32_t vofs, int len, int rn, int imm) /* Predicate register stores can be any multiple of 2. */ if (len_remain) { + t0 = tcg_temp_new_i64(); tcg_gen_ld_i64(t0, cpu_env, vofs + len_align); - tcg_gen_addi_i64(addr, cpu_reg_sp(s, rn), imm + len_align); switch (len_remain) { case 2: case 4: case 8: - tcg_gen_qemu_st_i64(t0, addr, midx, MO_LE | ctz32(len_remain)); + tcg_gen_qemu_st_i64(t0, clean_addr, midx, + MO_LE | ctz32(len_remain)); break; case 6: - tcg_gen_qemu_st_i64(t0, addr, midx, MO_LEUL); - tcg_gen_addi_i64(addr, addr, 4); + tcg_gen_qemu_st_i64(t0, clean_addr, midx, MO_LEUL); + tcg_gen_addi_i64(clean_addr, clean_addr, 4); tcg_gen_shri_i64(t0, t0, 32); - tcg_gen_qemu_st_i64(t0, addr, midx, MO_LEUW); + tcg_gen_qemu_st_i64(t0, clean_addr, midx, MO_LEUW); break; default: g_assert_not_reached(); } + tcg_temp_free_i64(t0); } - tcg_temp_free_i64(addr); - tcg_temp_free_i64(t0); + tcg_temp_free_i64(clean_addr); } static bool trans_LDR_zri(DisasContext *s, arg_rri *a) -- 2.20.1 From: Richard Henderson <richard.henderson@linaro.org> Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200626033144.790098-31-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/translate-sve.c | 6 ++++-- 1 file changed, 4 insertions(+), 2 deletions(-) diff --git a/target/arm/translate-sve.c b/target/arm/translate-sve.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-sve.c +++ b/target/arm/translate-sve.c @@ -XXX,XX +XXX,XX @@ static bool trans_LD1R_zpri(DisasContext *s, arg_rpri_load *a) unsigned esz = dtype_esz[a->dtype]; unsigned msz = dtype_msz(a->dtype); TCGLabel *over = gen_new_label(); - TCGv_i64 temp; + TCGv_i64 temp, clean_addr; /* If the guarding predicate has no bits set, no load occurs. */ if (psz <= 8) { @@ -XXX,XX +XXX,XX @@ static bool trans_LD1R_zpri(DisasContext *s, arg_rpri_load *a) /* Load the data. */ temp = tcg_temp_new_i64(); tcg_gen_addi_i64(temp, cpu_reg_sp(s, a->rn), a->imm << msz); - tcg_gen_qemu_ld_i64(temp, temp, get_mem_index(s), + clean_addr = gen_mte_check1(s, temp, false, true, msz); + + tcg_gen_qemu_ld_i64(temp, clean_addr, get_mem_index(s), s->be_data | dtype_mop[a->dtype]); /* Broadcast to *all* elements. */ -- 2.20.1 From: Richard Henderson <richard.henderson@linaro.org> Move the variable declarations to the top of the function, but do not create a new label before sve_access_check. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200626033144.790098-32-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/translate-sve.c | 12 +++++++----- 1 file changed, 7 insertions(+), 5 deletions(-) diff --git a/target/arm/translate-sve.c b/target/arm/translate-sve.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-sve.c +++ b/target/arm/translate-sve.c @@ -XXX,XX +XXX,XX @@ static bool trans_LD1RQ_zpri(DisasContext *s, arg_rpri_load *a) /* Load and broadcast element. */ static bool trans_LD1R_zpri(DisasContext *s, arg_rpri_load *a) { - if (!sve_access_check(s)) { - return true; - } - unsigned vsz = vec_full_reg_size(s); unsigned psz = pred_full_reg_size(s); unsigned esz = dtype_esz[a->dtype]; unsigned msz = dtype_msz(a->dtype); - TCGLabel *over = gen_new_label(); + TCGLabel *over; TCGv_i64 temp, clean_addr; + if (!sve_access_check(s)) { + return true; + } + + over = gen_new_label(); + /* If the guarding predicate has no bits set, no load occurs. */ if (psz <= 8) { /* Reduce the pred_esz_masks value simply to reduce the -- 2.20.1 From: Richard Henderson <richard.henderson@linaro.org> Introduce an lvalue macro to wrap target_tlb_bit0. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200626033144.790098-33-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/cpu.h | 13 +++++++++++++ target/arm/helper.c | 2 +- target/arm/translate-a64.c | 2 +- 3 files changed, 15 insertions(+), 2 deletions(-) diff --git a/target/arm/cpu.h b/target/arm/cpu.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/cpu.h +++ b/target/arm/cpu.h @@ -XXX,XX +XXX,XX @@ static inline uint64_t *aa64_vfp_qreg(CPUARMState *env, unsigned regno) /* Shared between translate-sve.c and sve_helper.c. */ extern const uint64_t pred_esz_masks[4]; +/* Helper for the macros below, validating the argument type. */ +static inline MemTxAttrs *typecheck_memtxattrs(MemTxAttrs *x) +{ + return x; +} + +/* + * Lvalue macros for ARM TLB bits that we must cache in the TCG TLB. + * Using these should be a bit more self-documenting than using the + * generic target bits directly. + */ +#define arm_tlb_bti_gp(x) (typecheck_memtxattrs(x)->target_tlb_bit0) + /* * Naming convention for isar_feature functions: * Functions which test 32-bit ID registers should have _aa32_ in diff --git a/target/arm/helper.c b/target/arm/helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper.c +++ b/target/arm/helper.c @@ -XXX,XX +XXX,XX @@ static bool get_phys_addr_lpae(CPUARMState *env, target_ulong address, } /* When in aarch64 mode, and BTI is enabled, remember GP in the IOTLB. */ if (aarch64 && guarded && cpu_isar_feature(aa64_bti, cpu)) { - txattrs->target_tlb_bit0 = true; + arm_tlb_bti_gp(txattrs) = true; } if (cacheattrs != NULL) { diff --git a/target/arm/translate-a64.c b/target/arm/translate-a64.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-a64.c +++ b/target/arm/translate-a64.c @@ -XXX,XX +XXX,XX @@ static bool is_guarded_page(CPUARMState *env, DisasContext *s) * table entry even for that case. */ return (tlb_hit(entry->addr_code, addr) && - env_tlb(env)->d[mmu_idx].iotlb[index].attrs.target_tlb_bit0); + arm_tlb_bti_gp(&env_tlb(env)->d[mmu_idx].iotlb[index].attrs)); #endif } -- 2.20.1 From: Richard Henderson <richard.henderson@linaro.org> Because the elements are sequential, we can eliminate many tests all at once when the tag hits TCMA, or if the page(s) are not Tagged. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200626033144.790098-34-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/cpu.h | 1 + target/arm/helper-sve.h | 58 ++++++++++ target/arm/internals.h | 6 + target/arm/sve_helper.c | 218 ++++++++++++++++++++++++++++++------- target/arm/translate-sve.c | 186 ++++++++++++++++++++++--------- 5 files changed, 378 insertions(+), 91 deletions(-) diff --git a/target/arm/cpu.h b/target/arm/cpu.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/cpu.h +++ b/target/arm/cpu.h @@ -XXX,XX +XXX,XX @@ static inline MemTxAttrs *typecheck_memtxattrs(MemTxAttrs *x) * generic target bits directly. */ #define arm_tlb_bti_gp(x) (typecheck_memtxattrs(x)->target_tlb_bit0) +#define arm_tlb_mte_tagged(x) (typecheck_memtxattrs(x)->target_tlb_bit1) /* * Naming convention for isar_feature functions: diff --git a/target/arm/helper-sve.h b/target/arm/helper-sve.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper-sve.h +++ b/target/arm/helper-sve.h @@ -XXX,XX +XXX,XX @@ DEF_HELPER_FLAGS_4(sve_ld1sds_le_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) DEF_HELPER_FLAGS_4(sve_ld1sdu_be_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) DEF_HELPER_FLAGS_4(sve_ld1sds_be_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ld1bb_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ld2bb_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ld3bb_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ld4bb_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) + +DEF_HELPER_FLAGS_4(sve_ld1hh_le_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ld2hh_le_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ld3hh_le_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ld4hh_le_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) + +DEF_HELPER_FLAGS_4(sve_ld1hh_be_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ld2hh_be_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ld3hh_be_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ld4hh_be_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) + +DEF_HELPER_FLAGS_4(sve_ld1ss_le_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ld2ss_le_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ld3ss_le_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ld4ss_le_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) + +DEF_HELPER_FLAGS_4(sve_ld1ss_be_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ld2ss_be_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ld3ss_be_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ld4ss_be_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) + +DEF_HELPER_FLAGS_4(sve_ld1dd_le_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ld2dd_le_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ld3dd_le_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ld4dd_le_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) + +DEF_HELPER_FLAGS_4(sve_ld1dd_be_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ld2dd_be_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ld3dd_be_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ld4dd_be_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) + +DEF_HELPER_FLAGS_4(sve_ld1bhu_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ld1bsu_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ld1bdu_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ld1bhs_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ld1bss_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ld1bds_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) + +DEF_HELPER_FLAGS_4(sve_ld1hsu_le_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ld1hdu_le_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ld1hss_le_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ld1hds_le_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) + +DEF_HELPER_FLAGS_4(sve_ld1hsu_be_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ld1hdu_be_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ld1hss_be_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ld1hds_be_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) + +DEF_HELPER_FLAGS_4(sve_ld1sdu_le_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ld1sds_le_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) + +DEF_HELPER_FLAGS_4(sve_ld1sdu_be_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ld1sds_be_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) + DEF_HELPER_FLAGS_4(sve_ldff1bb_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) DEF_HELPER_FLAGS_4(sve_ldff1bhu_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) DEF_HELPER_FLAGS_4(sve_ldff1bsu_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) diff --git a/target/arm/internals.h b/target/arm/internals.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/internals.h +++ b/target/arm/internals.h @@ -XXX,XX +XXX,XX @@ void arm_log_exception(int idx); #define LOG2_TAG_GRANULE 4 #define TAG_GRANULE (1 << LOG2_TAG_GRANULE) +/* + * The SVE simd_data field, for memory ops, contains either + * rd (5 bits) or a shift count (2 bits). + */ +#define SVE_MTEDESC_SHIFT 5 + /* Bits within a descriptor passed to the helper_mte_check* functions. */ FIELD(MTEDESC, MIDX, 0, 4) FIELD(MTEDESC, TBI, 4, 2) diff --git a/target/arm/sve_helper.c b/target/arm/sve_helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/sve_helper.c +++ b/target/arm/sve_helper.c @@ -XXX,XX +XXX,XX @@ static void sve_cont_ldst_watchpoints(SVEContLdSt *info, CPUARMState *env, #endif } +typedef uint64_t mte_check_fn(CPUARMState *, uint32_t, uint64_t, uintptr_t); + +static inline QEMU_ALWAYS_INLINE +void sve_cont_ldst_mte_check_int(SVEContLdSt *info, CPUARMState *env, + uint64_t *vg, target_ulong addr, int esize, + int msize, uint32_t mtedesc, uintptr_t ra, + mte_check_fn *check) +{ + intptr_t mem_off, reg_off, reg_last; + + /* Process the page only if MemAttr == Tagged. */ + if (arm_tlb_mte_tagged(&info->page[0].attrs)) { + mem_off = info->mem_off_first[0]; + reg_off = info->reg_off_first[0]; + reg_last = info->reg_off_split; + if (reg_last < 0) { + reg_last = info->reg_off_last[0]; + } + + do { + uint64_t pg = vg[reg_off >> 6]; + do { + if ((pg >> (reg_off & 63)) & 1) { + check(env, mtedesc, addr, ra); + } + reg_off += esize; + mem_off += msize; + } while (reg_off <= reg_last && (reg_off & 63)); + } while (reg_off <= reg_last); + } + + mem_off = info->mem_off_first[1]; + if (mem_off >= 0 && arm_tlb_mte_tagged(&info->page[1].attrs)) { + reg_off = info->reg_off_first[1]; + reg_last = info->reg_off_last[1]; + + do { + uint64_t pg = vg[reg_off >> 6]; + do { + if ((pg >> (reg_off & 63)) & 1) { + check(env, mtedesc, addr, ra); + } + reg_off += esize; + mem_off += msize; + } while (reg_off & 63); + } while (reg_off <= reg_last); + } +} + +typedef void sve_cont_ldst_mte_check_fn(SVEContLdSt *info, CPUARMState *env, + uint64_t *vg, target_ulong addr, + int esize, int msize, uint32_t mtedesc, + uintptr_t ra); + +static void sve_cont_ldst_mte_check1(SVEContLdSt *info, CPUARMState *env, + uint64_t *vg, target_ulong addr, + int esize, int msize, uint32_t mtedesc, + uintptr_t ra) +{ + sve_cont_ldst_mte_check_int(info, env, vg, addr, esize, msize, + mtedesc, ra, mte_check1); +} + +static void sve_cont_ldst_mte_checkN(SVEContLdSt *info, CPUARMState *env, + uint64_t *vg, target_ulong addr, + int esize, int msize, uint32_t mtedesc, + uintptr_t ra) +{ + sve_cont_ldst_mte_check_int(info, env, vg, addr, esize, msize, + mtedesc, ra, mte_checkN); +} + + /* * Common helper for all contiguous 1,2,3,4-register predicated stores. */ static inline QEMU_ALWAYS_INLINE void sve_ldN_r(CPUARMState *env, uint64_t *vg, const target_ulong addr, uint32_t desc, const uintptr_t retaddr, - const int esz, const int msz, const int N, + const int esz, const int msz, const int N, uint32_t mtedesc, sve_ldst1_host_fn *host_fn, - sve_ldst1_tlb_fn *tlb_fn) + sve_ldst1_tlb_fn *tlb_fn, + sve_cont_ldst_mte_check_fn *mte_check_fn) { const unsigned rd = simd_data(desc); const intptr_t reg_max = simd_oprsz(desc); @@ -XXX,XX +XXX,XX @@ void sve_ldN_r(CPUARMState *env, uint64_t *vg, const target_ulong addr, sve_cont_ldst_watchpoints(&info, env, vg, addr, 1 << esz, N << msz, BP_MEM_READ, retaddr); - /* TODO: MTE check. */ + /* + * Handle mte checks for all active elements. + * Since TBI must be set for MTE, !mtedesc => !mte_active. + */ + if (mte_check_fn && mtedesc) { + mte_check_fn(&info, env, vg, addr, 1 << esz, N << msz, + mtedesc, retaddr); + } flags = info.page[0].flags | info.page[1].flags; if (unlikely(flags != 0)) { @@ -XXX,XX +XXX,XX @@ void sve_ldN_r(CPUARMState *env, uint64_t *vg, const target_ulong addr, } } -#define DO_LD1_1(NAME, ESZ) \ -void HELPER(sve_##NAME##_r)(CPUARMState *env, void *vg, \ - target_ulong addr, uint32_t desc) \ -{ \ - sve_ldN_r(env, vg, addr, desc, GETPC(), ESZ, MO_8, 1, \ - sve_##NAME##_host, sve_##NAME##_tlb); \ +static inline QEMU_ALWAYS_INLINE +void sve_ldN_r_mte(CPUARMState *env, uint64_t *vg, target_ulong addr, + uint32_t desc, const uintptr_t ra, + const int esz, const int msz, const int N, + sve_ldst1_host_fn *host_fn, + sve_ldst1_tlb_fn *tlb_fn) +{ + uint32_t mtedesc = desc >> (SIMD_DATA_SHIFT + SVE_MTEDESC_SHIFT); + int bit55 = extract64(addr, 55, 1); + + /* Remove mtedesc from the normal sve descriptor. */ + desc = extract32(desc, 0, SIMD_DATA_SHIFT + SVE_MTEDESC_SHIFT); + + /* Perform gross MTE suppression early. */ + if (!tbi_check(desc, bit55) || + tcma_check(desc, bit55, allocation_tag_from_addr(addr))) { + mtedesc = 0; + } + + sve_ldN_r(env, vg, addr, desc, ra, esz, msz, N, mtedesc, host_fn, tlb_fn, + N == 1 ? sve_cont_ldst_mte_check1 : sve_cont_ldst_mte_checkN); } -#define DO_LD1_2(NAME, ESZ, MSZ) \ -void HELPER(sve_##NAME##_le_r)(CPUARMState *env, void *vg, \ - target_ulong addr, uint32_t desc) \ -{ \ - sve_ldN_r(env, vg, addr, desc, GETPC(), ESZ, MSZ, 1, \ - sve_##NAME##_le_host, sve_##NAME##_le_tlb); \ -} \ -void HELPER(sve_##NAME##_be_r)(CPUARMState *env, void *vg, \ - target_ulong addr, uint32_t desc) \ -{ \ - sve_ldN_r(env, vg, addr, desc, GETPC(), ESZ, MSZ, 1, \ - sve_##NAME##_be_host, sve_##NAME##_be_tlb); \ +#define DO_LD1_1(NAME, ESZ) \ +void HELPER(sve_##NAME##_r)(CPUARMState *env, void *vg, \ + target_ulong addr, uint32_t desc) \ +{ \ + sve_ldN_r(env, vg, addr, desc, GETPC(), ESZ, MO_8, 1, 0, \ + sve_##NAME##_host, sve_##NAME##_tlb, NULL); \ +} \ +void HELPER(sve_##NAME##_r_mte)(CPUARMState *env, void *vg, \ + target_ulong addr, uint32_t desc) \ +{ \ + sve_ldN_r_mte(env, vg, addr, desc, GETPC(), ESZ, MO_8, 1, \ + sve_##NAME##_host, sve_##NAME##_tlb); \ +} + +#define DO_LD1_2(NAME, ESZ, MSZ) \ +void HELPER(sve_##NAME##_le_r)(CPUARMState *env, void *vg, \ + target_ulong addr, uint32_t desc) \ +{ \ + sve_ldN_r(env, vg, addr, desc, GETPC(), ESZ, MSZ, 1, 0, \ + sve_##NAME##_le_host, sve_##NAME##_le_tlb, NULL); \ +} \ +void HELPER(sve_##NAME##_be_r)(CPUARMState *env, void *vg, \ + target_ulong addr, uint32_t desc) \ +{ \ + sve_ldN_r(env, vg, addr, desc, GETPC(), ESZ, MSZ, 1, 0, \ + sve_##NAME##_be_host, sve_##NAME##_be_tlb, NULL); \ +} \ +void HELPER(sve_##NAME##_le_r_mte)(CPUARMState *env, void *vg, \ + target_ulong addr, uint32_t desc) \ +{ \ + sve_ldN_r_mte(env, vg, addr, desc, GETPC(), ESZ, MSZ, 1, \ + sve_##NAME##_le_host, sve_##NAME##_le_tlb); \ +} \ +void HELPER(sve_##NAME##_be_r_mte)(CPUARMState *env, void *vg, \ + target_ulong addr, uint32_t desc) \ +{ \ + sve_ldN_r_mte(env, vg, addr, desc, GETPC(), ESZ, MSZ, 1, \ + sve_##NAME##_be_host, sve_##NAME##_be_tlb); \ } DO_LD1_1(ld1bb, MO_8) @@ -XXX,XX +XXX,XX @@ DO_LD1_2(ld1dd, MO_64, MO_64) #undef DO_LD1_1 #undef DO_LD1_2 -#define DO_LDN_1(N) \ -void HELPER(sve_ld##N##bb_r)(CPUARMState *env, void *vg, \ - target_ulong addr, uint32_t desc) \ -{ \ - sve_ldN_r(env, vg, addr, desc, GETPC(), MO_8, MO_8, N, \ - sve_ld1bb_host, sve_ld1bb_tlb); \ +#define DO_LDN_1(N) \ +void HELPER(sve_ld##N##bb_r)(CPUARMState *env, void *vg, \ + target_ulong addr, uint32_t desc) \ +{ \ + sve_ldN_r(env, vg, addr, desc, GETPC(), MO_8, MO_8, N, 0, \ + sve_ld1bb_host, sve_ld1bb_tlb, NULL); \ +} \ +void HELPER(sve_ld##N##bb_r_mte)(CPUARMState *env, void *vg, \ + target_ulong addr, uint32_t desc) \ +{ \ + sve_ldN_r_mte(env, vg, addr, desc, GETPC(), MO_8, MO_8, N, \ + sve_ld1bb_host, sve_ld1bb_tlb); \ } -#define DO_LDN_2(N, SUFF, ESZ) \ -void HELPER(sve_ld##N##SUFF##_le_r)(CPUARMState *env, void *vg, \ - target_ulong addr, uint32_t desc) \ -{ \ - sve_ldN_r(env, vg, addr, desc, GETPC(), ESZ, ESZ, N, \ - sve_ld1##SUFF##_le_host, sve_ld1##SUFF##_le_tlb); \ -} \ -void HELPER(sve_ld##N##SUFF##_be_r)(CPUARMState *env, void *vg, \ - target_ulong addr, uint32_t desc) \ -{ \ - sve_ldN_r(env, vg, addr, desc, GETPC(), ESZ, ESZ, N, \ - sve_ld1##SUFF##_be_host, sve_ld1##SUFF##_be_tlb); \ +#define DO_LDN_2(N, SUFF, ESZ) \ +void HELPER(sve_ld##N##SUFF##_le_r)(CPUARMState *env, void *vg, \ + target_ulong addr, uint32_t desc) \ +{ \ + sve_ldN_r(env, vg, addr, desc, GETPC(), ESZ, ESZ, N, 0, \ + sve_ld1##SUFF##_le_host, sve_ld1##SUFF##_le_tlb, NULL); \ +} \ +void HELPER(sve_ld##N##SUFF##_be_r)(CPUARMState *env, void *vg, \ + target_ulong addr, uint32_t desc) \ +{ \ + sve_ldN_r(env, vg, addr, desc, GETPC(), ESZ, ESZ, N, 0, \ + sve_ld1##SUFF##_be_host, sve_ld1##SUFF##_be_tlb, NULL); \ +} \ +void HELPER(sve_ld##N##SUFF##_le_r_mte)(CPUARMState *env, void *vg, \ + target_ulong addr, uint32_t desc) \ +{ \ + sve_ldN_r_mte(env, vg, addr, desc, GETPC(), ESZ, ESZ, N, \ + sve_ld1##SUFF##_le_host, sve_ld1##SUFF##_le_tlb); \ +} \ +void HELPER(sve_ld##N##SUFF##_be_r_mte)(CPUARMState *env, void *vg, \ + target_ulong addr, uint32_t desc) \ +{ \ + sve_ldN_r_mte(env, vg, addr, desc, GETPC(), ESZ, ESZ, N, \ + sve_ld1##SUFF##_be_host, sve_ld1##SUFF##_be_tlb); \ } DO_LDN_1(2) diff --git a/target/arm/translate-sve.c b/target/arm/translate-sve.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-sve.c +++ b/target/arm/translate-sve.c @@ -XXX,XX +XXX,XX @@ static const uint8_t dtype_esz[16] = { }; static void do_mem_zpa(DisasContext *s, int zt, int pg, TCGv_i64 addr, - int dtype, gen_helper_gvec_mem *fn) + int dtype, uint32_t mte_n, bool is_write, + gen_helper_gvec_mem *fn) { unsigned vsz = vec_full_reg_size(s); TCGv_ptr t_pg; TCGv_i32 t_desc; - int desc; + int desc = 0; - /* For e.g. LD4, there are not enough arguments to pass all 4 + /* + * For e.g. LD4, there are not enough arguments to pass all 4 * registers as pointers, so encode the regno into the data field. * For consistency, do this even for LD1. + * TODO: mte_n check here while callers are updated. */ - desc = simd_desc(vsz, vsz, zt); + if (mte_n && s->mte_active[0]) { + int msz = dtype_msz(dtype); + + desc = FIELD_DP32(desc, MTEDESC, MIDX, get_mem_index(s)); + desc = FIELD_DP32(desc, MTEDESC, TBI, s->tbid); + desc = FIELD_DP32(desc, MTEDESC, TCMA, s->tcma); + desc = FIELD_DP32(desc, MTEDESC, WRITE, is_write); + desc = FIELD_DP32(desc, MTEDESC, ESIZE, 1 << msz); + desc = FIELD_DP32(desc, MTEDESC, TSIZE, mte_n << msz); + desc <<= SVE_MTEDESC_SHIFT; + } + desc = simd_desc(vsz, vsz, zt | desc); t_desc = tcg_const_i32(desc); t_pg = tcg_temp_new_ptr(); @@ -XXX,XX +XXX,XX @@ static void do_mem_zpa(DisasContext *s, int zt, int pg, TCGv_i64 addr, static void do_ld_zpa(DisasContext *s, int zt, int pg, TCGv_i64 addr, int dtype, int nreg) { - static gen_helper_gvec_mem * const fns[2][16][4] = { - /* Little-endian */ - { { gen_helper_sve_ld1bb_r, gen_helper_sve_ld2bb_r, + static gen_helper_gvec_mem * const fns[2][2][16][4] = { + { /* mte inactive, little-endian */ + { { gen_helper_sve_ld1bb_r, gen_helper_sve_ld2bb_r, gen_helper_sve_ld3bb_r, gen_helper_sve_ld4bb_r }, - { gen_helper_sve_ld1bhu_r, NULL, NULL, NULL }, - { gen_helper_sve_ld1bsu_r, NULL, NULL, NULL }, - { gen_helper_sve_ld1bdu_r, NULL, NULL, NULL }, + { gen_helper_sve_ld1bhu_r, NULL, NULL, NULL }, + { gen_helper_sve_ld1bsu_r, NULL, NULL, NULL }, + { gen_helper_sve_ld1bdu_r, NULL, NULL, NULL }, - { gen_helper_sve_ld1sds_le_r, NULL, NULL, NULL }, - { gen_helper_sve_ld1hh_le_r, gen_helper_sve_ld2hh_le_r, - gen_helper_sve_ld3hh_le_r, gen_helper_sve_ld4hh_le_r }, - { gen_helper_sve_ld1hsu_le_r, NULL, NULL, NULL }, - { gen_helper_sve_ld1hdu_le_r, NULL, NULL, NULL }, + { gen_helper_sve_ld1sds_le_r, NULL, NULL, NULL }, + { gen_helper_sve_ld1hh_le_r, gen_helper_sve_ld2hh_le_r, + gen_helper_sve_ld3hh_le_r, gen_helper_sve_ld4hh_le_r }, + { gen_helper_sve_ld1hsu_le_r, NULL, NULL, NULL }, + { gen_helper_sve_ld1hdu_le_r, NULL, NULL, NULL }, - { gen_helper_sve_ld1hds_le_r, NULL, NULL, NULL }, - { gen_helper_sve_ld1hss_le_r, NULL, NULL, NULL }, - { gen_helper_sve_ld1ss_le_r, gen_helper_sve_ld2ss_le_r, - gen_helper_sve_ld3ss_le_r, gen_helper_sve_ld4ss_le_r }, - { gen_helper_sve_ld1sdu_le_r, NULL, NULL, NULL }, + { gen_helper_sve_ld1hds_le_r, NULL, NULL, NULL }, + { gen_helper_sve_ld1hss_le_r, NULL, NULL, NULL }, + { gen_helper_sve_ld1ss_le_r, gen_helper_sve_ld2ss_le_r, + gen_helper_sve_ld3ss_le_r, gen_helper_sve_ld4ss_le_r }, + { gen_helper_sve_ld1sdu_le_r, NULL, NULL, NULL }, - { gen_helper_sve_ld1bds_r, NULL, NULL, NULL }, - { gen_helper_sve_ld1bss_r, NULL, NULL, NULL }, - { gen_helper_sve_ld1bhs_r, NULL, NULL, NULL }, - { gen_helper_sve_ld1dd_le_r, gen_helper_sve_ld2dd_le_r, - gen_helper_sve_ld3dd_le_r, gen_helper_sve_ld4dd_le_r } }, + { gen_helper_sve_ld1bds_r, NULL, NULL, NULL }, + { gen_helper_sve_ld1bss_r, NULL, NULL, NULL }, + { gen_helper_sve_ld1bhs_r, NULL, NULL, NULL }, + { gen_helper_sve_ld1dd_le_r, gen_helper_sve_ld2dd_le_r, + gen_helper_sve_ld3dd_le_r, gen_helper_sve_ld4dd_le_r } }, - /* Big-endian */ - { { gen_helper_sve_ld1bb_r, gen_helper_sve_ld2bb_r, - gen_helper_sve_ld3bb_r, gen_helper_sve_ld4bb_r }, - { gen_helper_sve_ld1bhu_r, NULL, NULL, NULL }, - { gen_helper_sve_ld1bsu_r, NULL, NULL, NULL }, - { gen_helper_sve_ld1bdu_r, NULL, NULL, NULL }, + /* mte inactive, big-endian */ + { { gen_helper_sve_ld1bb_r, gen_helper_sve_ld2bb_r, + gen_helper_sve_ld3bb_r, gen_helper_sve_ld4bb_r }, + { gen_helper_sve_ld1bhu_r, NULL, NULL, NULL }, + { gen_helper_sve_ld1bsu_r, NULL, NULL, NULL }, + { gen_helper_sve_ld1bdu_r, NULL, NULL, NULL }, - { gen_helper_sve_ld1sds_be_r, NULL, NULL, NULL }, - { gen_helper_sve_ld1hh_be_r, gen_helper_sve_ld2hh_be_r, - gen_helper_sve_ld3hh_be_r, gen_helper_sve_ld4hh_be_r }, - { gen_helper_sve_ld1hsu_be_r, NULL, NULL, NULL }, - { gen_helper_sve_ld1hdu_be_r, NULL, NULL, NULL }, + { gen_helper_sve_ld1sds_be_r, NULL, NULL, NULL }, + { gen_helper_sve_ld1hh_be_r, gen_helper_sve_ld2hh_be_r, + gen_helper_sve_ld3hh_be_r, gen_helper_sve_ld4hh_be_r }, + { gen_helper_sve_ld1hsu_be_r, NULL, NULL, NULL }, + { gen_helper_sve_ld1hdu_be_r, NULL, NULL, NULL }, - { gen_helper_sve_ld1hds_be_r, NULL, NULL, NULL }, - { gen_helper_sve_ld1hss_be_r, NULL, NULL, NULL }, - { gen_helper_sve_ld1ss_be_r, gen_helper_sve_ld2ss_be_r, - gen_helper_sve_ld3ss_be_r, gen_helper_sve_ld4ss_be_r }, - { gen_helper_sve_ld1sdu_be_r, NULL, NULL, NULL }, + { gen_helper_sve_ld1hds_be_r, NULL, NULL, NULL }, + { gen_helper_sve_ld1hss_be_r, NULL, NULL, NULL }, + { gen_helper_sve_ld1ss_be_r, gen_helper_sve_ld2ss_be_r, + gen_helper_sve_ld3ss_be_r, gen_helper_sve_ld4ss_be_r }, + { gen_helper_sve_ld1sdu_be_r, NULL, NULL, NULL }, - { gen_helper_sve_ld1bds_r, NULL, NULL, NULL }, - { gen_helper_sve_ld1bss_r, NULL, NULL, NULL }, - { gen_helper_sve_ld1bhs_r, NULL, NULL, NULL }, - { gen_helper_sve_ld1dd_be_r, gen_helper_sve_ld2dd_be_r, - gen_helper_sve_ld3dd_be_r, gen_helper_sve_ld4dd_be_r } } + { gen_helper_sve_ld1bds_r, NULL, NULL, NULL }, + { gen_helper_sve_ld1bss_r, NULL, NULL, NULL }, + { gen_helper_sve_ld1bhs_r, NULL, NULL, NULL }, + { gen_helper_sve_ld1dd_be_r, gen_helper_sve_ld2dd_be_r, + gen_helper_sve_ld3dd_be_r, gen_helper_sve_ld4dd_be_r } } }, + + { /* mte active, little-endian */ + { { gen_helper_sve_ld1bb_r_mte, + gen_helper_sve_ld2bb_r_mte, + gen_helper_sve_ld3bb_r_mte, + gen_helper_sve_ld4bb_r_mte }, + { gen_helper_sve_ld1bhu_r_mte, NULL, NULL, NULL }, + { gen_helper_sve_ld1bsu_r_mte, NULL, NULL, NULL }, + { gen_helper_sve_ld1bdu_r_mte, NULL, NULL, NULL }, + + { gen_helper_sve_ld1sds_le_r_mte, NULL, NULL, NULL }, + { gen_helper_sve_ld1hh_le_r_mte, + gen_helper_sve_ld2hh_le_r_mte, + gen_helper_sve_ld3hh_le_r_mte, + gen_helper_sve_ld4hh_le_r_mte }, + { gen_helper_sve_ld1hsu_le_r_mte, NULL, NULL, NULL }, + { gen_helper_sve_ld1hdu_le_r_mte, NULL, NULL, NULL }, + + { gen_helper_sve_ld1hds_le_r_mte, NULL, NULL, NULL }, + { gen_helper_sve_ld1hss_le_r_mte, NULL, NULL, NULL }, + { gen_helper_sve_ld1ss_le_r_mte, + gen_helper_sve_ld2ss_le_r_mte, + gen_helper_sve_ld3ss_le_r_mte, + gen_helper_sve_ld4ss_le_r_mte }, + { gen_helper_sve_ld1sdu_le_r_mte, NULL, NULL, NULL }, + + { gen_helper_sve_ld1bds_r_mte, NULL, NULL, NULL }, + { gen_helper_sve_ld1bss_r_mte, NULL, NULL, NULL }, + { gen_helper_sve_ld1bhs_r_mte, NULL, NULL, NULL }, + { gen_helper_sve_ld1dd_le_r_mte, + gen_helper_sve_ld2dd_le_r_mte, + gen_helper_sve_ld3dd_le_r_mte, + gen_helper_sve_ld4dd_le_r_mte } }, + + /* mte active, big-endian */ + { { gen_helper_sve_ld1bb_r_mte, + gen_helper_sve_ld2bb_r_mte, + gen_helper_sve_ld3bb_r_mte, + gen_helper_sve_ld4bb_r_mte }, + { gen_helper_sve_ld1bhu_r_mte, NULL, NULL, NULL }, + { gen_helper_sve_ld1bsu_r_mte, NULL, NULL, NULL }, + { gen_helper_sve_ld1bdu_r_mte, NULL, NULL, NULL }, + + { gen_helper_sve_ld1sds_be_r_mte, NULL, NULL, NULL }, + { gen_helper_sve_ld1hh_be_r_mte, + gen_helper_sve_ld2hh_be_r_mte, + gen_helper_sve_ld3hh_be_r_mte, + gen_helper_sve_ld4hh_be_r_mte }, + { gen_helper_sve_ld1hsu_be_r_mte, NULL, NULL, NULL }, + { gen_helper_sve_ld1hdu_be_r_mte, NULL, NULL, NULL }, + + { gen_helper_sve_ld1hds_be_r_mte, NULL, NULL, NULL }, + { gen_helper_sve_ld1hss_be_r_mte, NULL, NULL, NULL }, + { gen_helper_sve_ld1ss_be_r_mte, + gen_helper_sve_ld2ss_be_r_mte, + gen_helper_sve_ld3ss_be_r_mte, + gen_helper_sve_ld4ss_be_r_mte }, + { gen_helper_sve_ld1sdu_be_r_mte, NULL, NULL, NULL }, + + { gen_helper_sve_ld1bds_r_mte, NULL, NULL, NULL }, + { gen_helper_sve_ld1bss_r_mte, NULL, NULL, NULL }, + { gen_helper_sve_ld1bhs_r_mte, NULL, NULL, NULL }, + { gen_helper_sve_ld1dd_be_r_mte, + gen_helper_sve_ld2dd_be_r_mte, + gen_helper_sve_ld3dd_be_r_mte, + gen_helper_sve_ld4dd_be_r_mte } } }, }; - gen_helper_gvec_mem *fn = fns[s->be_data == MO_BE][dtype][nreg]; + gen_helper_gvec_mem *fn + = fns[s->mte_active[0]][s->be_data == MO_BE][dtype][nreg]; - /* While there are holes in the table, they are not + /* + * While there are holes in the table, they are not * accessible via the instruction encoding. */ assert(fn != NULL); - do_mem_zpa(s, zt, pg, addr, dtype, fn); + do_mem_zpa(s, zt, pg, addr, dtype, nreg, false, fn); } static bool trans_LD_zprr(DisasContext *s, arg_rprr_load *a) @@ -XXX,XX +XXX,XX @@ static bool trans_LDFF1_zprr(DisasContext *s, arg_rprr_load *a) TCGv_i64 addr = new_tmp_a64(s); tcg_gen_shli_i64(addr, cpu_reg(s, a->rm), dtype_msz(a->dtype)); tcg_gen_add_i64(addr, addr, cpu_reg_sp(s, a->rn)); - do_mem_zpa(s, a->rd, a->pg, addr, a->dtype, + do_mem_zpa(s, a->rd, a->pg, addr, a->dtype, 0, false, fns[s->be_data == MO_BE][a->dtype]); } return true; @@ -XXX,XX +XXX,XX @@ static bool trans_LDNF1_zpri(DisasContext *s, arg_rpri_load *a) TCGv_i64 addr = new_tmp_a64(s); tcg_gen_addi_i64(addr, cpu_reg_sp(s, a->rn), off); - do_mem_zpa(s, a->rd, a->pg, addr, a->dtype, + do_mem_zpa(s, a->rd, a->pg, addr, a->dtype, 0, false, fns[s->be_data == MO_BE][a->dtype]); } return true; @@ -XXX,XX +XXX,XX @@ static void do_st_zpa(DisasContext *s, int zt, int pg, TCGv_i64 addr, fn = fn_multiple[be][nreg - 1][msz]; } assert(fn != NULL); - do_mem_zpa(s, zt, pg, addr, msz_dtype(s, msz), fn); + do_mem_zpa(s, zt, pg, addr, msz_dtype(s, msz), 0, true, fn); } static bool trans_ST_zprr(DisasContext *s, arg_rprr_store *a) -- 2.20.1 From: Richard Henderson <richard.henderson@linaro.org> Because the elements are sequential, we can eliminate many tests all at once when the tag hits TCMA, or if the page(s) are not Tagged. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200626033144.790098-35-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/helper-sve.h | 47 +++++++++++ target/arm/sve_helper.c | 95 ++++++++++++++++------ target/arm/translate-sve.c | 162 ++++++++++++++++++++++++------------- 3 files changed, 226 insertions(+), 78 deletions(-) diff --git a/target/arm/helper-sve.h b/target/arm/helper-sve.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper-sve.h +++ b/target/arm/helper-sve.h @@ -XXX,XX +XXX,XX @@ DEF_HELPER_FLAGS_4(sve_st1hd_be_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) DEF_HELPER_FLAGS_4(sve_st1sd_le_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) DEF_HELPER_FLAGS_4(sve_st1sd_be_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_st1bb_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_st2bb_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_st3bb_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_st4bb_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) + +DEF_HELPER_FLAGS_4(sve_st1hh_le_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_st2hh_le_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_st3hh_le_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_st4hh_le_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) + +DEF_HELPER_FLAGS_4(sve_st1hh_be_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_st2hh_be_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_st3hh_be_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_st4hh_be_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) + +DEF_HELPER_FLAGS_4(sve_st1ss_le_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_st2ss_le_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_st3ss_le_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_st4ss_le_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) + +DEF_HELPER_FLAGS_4(sve_st1ss_be_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_st2ss_be_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_st3ss_be_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_st4ss_be_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) + +DEF_HELPER_FLAGS_4(sve_st1dd_le_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_st2dd_le_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_st3dd_le_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_st4dd_le_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) + +DEF_HELPER_FLAGS_4(sve_st1dd_be_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_st2dd_be_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_st3dd_be_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_st4dd_be_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) + +DEF_HELPER_FLAGS_4(sve_st1bh_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_st1bs_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_st1bd_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) + +DEF_HELPER_FLAGS_4(sve_st1hs_le_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_st1hd_le_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_st1hs_be_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_st1hd_be_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) + +DEF_HELPER_FLAGS_4(sve_st1sd_le_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_st1sd_be_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) + DEF_HELPER_FLAGS_6(sve_ldbsu_zsu, TCG_CALL_NO_WG, void, env, ptr, ptr, ptr, tl, i32) DEF_HELPER_FLAGS_6(sve_ldhsu_le_zsu, TCG_CALL_NO_WG, diff --git a/target/arm/sve_helper.c b/target/arm/sve_helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/sve_helper.c +++ b/target/arm/sve_helper.c @@ -XXX,XX +XXX,XX @@ DO_LDFF1_LDNF1_2(dd, MO_64, MO_64) */ static inline QEMU_ALWAYS_INLINE -void sve_stN_r(CPUARMState *env, uint64_t *vg, target_ulong addr, uint32_t desc, - const uintptr_t retaddr, const int esz, - const int msz, const int N, +void sve_stN_r(CPUARMState *env, uint64_t *vg, target_ulong addr, + uint32_t desc, const uintptr_t retaddr, + const int esz, const int msz, const int N, uint32_t mtedesc, sve_ldst1_host_fn *host_fn, - sve_ldst1_tlb_fn *tlb_fn) + sve_ldst1_tlb_fn *tlb_fn, + sve_cont_ldst_mte_check_fn *mte_check_fn) { const unsigned rd = simd_data(desc); const intptr_t reg_max = simd_oprsz(desc); @@ -XXX,XX +XXX,XX @@ void sve_stN_r(CPUARMState *env, uint64_t *vg, target_ulong addr, uint32_t desc, sve_cont_ldst_watchpoints(&info, env, vg, addr, 1 << esz, N << msz, BP_MEM_WRITE, retaddr); - /* TODO: MTE check. */ + /* + * Handle mte checks for all active elements. + * Since TBI must be set for MTE, !mtedesc => !mte_active. + */ + if (mte_check_fn && mtedesc) { + mte_check_fn(&info, env, vg, addr, 1 << esz, N << msz, + mtedesc, retaddr); + } flags = info.page[0].flags | info.page[1].flags; if (unlikely(flags != 0)) { @@ -XXX,XX +XXX,XX @@ void sve_stN_r(CPUARMState *env, uint64_t *vg, target_ulong addr, uint32_t desc, } } -#define DO_STN_1(N, NAME, ESZ) \ -void HELPER(sve_st##N##NAME##_r)(CPUARMState *env, void *vg, \ - target_ulong addr, uint32_t desc) \ -{ \ - sve_stN_r(env, vg, addr, desc, GETPC(), ESZ, MO_8, N, \ - sve_st1##NAME##_host, sve_st1##NAME##_tlb); \ +static inline QEMU_ALWAYS_INLINE +void sve_stN_r_mte(CPUARMState *env, uint64_t *vg, target_ulong addr, + uint32_t desc, const uintptr_t ra, + const int esz, const int msz, const int N, + sve_ldst1_host_fn *host_fn, + sve_ldst1_tlb_fn *tlb_fn) +{ + uint32_t mtedesc = desc >> (SIMD_DATA_SHIFT + SVE_MTEDESC_SHIFT); + int bit55 = extract64(addr, 55, 1); + + /* Remove mtedesc from the normal sve descriptor. */ + desc = extract32(desc, 0, SIMD_DATA_SHIFT + SVE_MTEDESC_SHIFT); + + /* Perform gross MTE suppression early. */ + if (!tbi_check(desc, bit55) || + tcma_check(desc, bit55, allocation_tag_from_addr(addr))) { + mtedesc = 0; + } + + sve_stN_r(env, vg, addr, desc, ra, esz, msz, N, mtedesc, host_fn, tlb_fn, + N == 1 ? sve_cont_ldst_mte_check1 : sve_cont_ldst_mte_checkN); } -#define DO_STN_2(N, NAME, ESZ, MSZ) \ -void HELPER(sve_st##N##NAME##_le_r)(CPUARMState *env, void *vg, \ - target_ulong addr, uint32_t desc) \ -{ \ - sve_stN_r(env, vg, addr, desc, GETPC(), ESZ, MSZ, N, \ - sve_st1##NAME##_le_host, sve_st1##NAME##_le_tlb); \ -} \ -void HELPER(sve_st##N##NAME##_be_r)(CPUARMState *env, void *vg, \ - target_ulong addr, uint32_t desc) \ -{ \ - sve_stN_r(env, vg, addr, desc, GETPC(), ESZ, MSZ, N, \ - sve_st1##NAME##_be_host, sve_st1##NAME##_be_tlb); \ +#define DO_STN_1(N, NAME, ESZ) \ +void HELPER(sve_st##N##NAME##_r)(CPUARMState *env, void *vg, \ + target_ulong addr, uint32_t desc) \ +{ \ + sve_stN_r(env, vg, addr, desc, GETPC(), ESZ, MO_8, N, 0, \ + sve_st1##NAME##_host, sve_st1##NAME##_tlb, NULL); \ +} \ +void HELPER(sve_st##N##NAME##_r_mte)(CPUARMState *env, void *vg, \ + target_ulong addr, uint32_t desc) \ +{ \ + sve_stN_r_mte(env, vg, addr, desc, GETPC(), ESZ, MO_8, N, \ + sve_st1##NAME##_host, sve_st1##NAME##_tlb); \ +} + +#define DO_STN_2(N, NAME, ESZ, MSZ) \ +void HELPER(sve_st##N##NAME##_le_r)(CPUARMState *env, void *vg, \ + target_ulong addr, uint32_t desc) \ +{ \ + sve_stN_r(env, vg, addr, desc, GETPC(), ESZ, MSZ, N, 0, \ + sve_st1##NAME##_le_host, sve_st1##NAME##_le_tlb, NULL); \ +} \ +void HELPER(sve_st##N##NAME##_be_r)(CPUARMState *env, void *vg, \ + target_ulong addr, uint32_t desc) \ +{ \ + sve_stN_r(env, vg, addr, desc, GETPC(), ESZ, MSZ, N, 0, \ + sve_st1##NAME##_be_host, sve_st1##NAME##_be_tlb, NULL); \ +} \ +void HELPER(sve_st##N##NAME##_le_r_mte)(CPUARMState *env, void *vg, \ + target_ulong addr, uint32_t desc) \ +{ \ + sve_stN_r_mte(env, vg, addr, desc, GETPC(), ESZ, MSZ, N, \ + sve_st1##NAME##_le_host, sve_st1##NAME##_le_tlb); \ +} \ +void HELPER(sve_st##N##NAME##_be_r_mte)(CPUARMState *env, void *vg, \ + target_ulong addr, uint32_t desc) \ +{ \ + sve_stN_r_mte(env, vg, addr, desc, GETPC(), ESZ, MSZ, N, \ + sve_st1##NAME##_be_host, sve_st1##NAME##_be_tlb); \ } DO_STN_1(1, bb, MO_8) diff --git a/target/arm/translate-sve.c b/target/arm/translate-sve.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-sve.c +++ b/target/arm/translate-sve.c @@ -XXX,XX +XXX,XX @@ static bool trans_LD1R_zpri(DisasContext *s, arg_rpri_load *a) static void do_st_zpa(DisasContext *s, int zt, int pg, TCGv_i64 addr, int msz, int esz, int nreg) { - static gen_helper_gvec_mem * const fn_single[2][4][4] = { - { { gen_helper_sve_st1bb_r, - gen_helper_sve_st1bh_r, - gen_helper_sve_st1bs_r, - gen_helper_sve_st1bd_r }, - { NULL, - gen_helper_sve_st1hh_le_r, - gen_helper_sve_st1hs_le_r, - gen_helper_sve_st1hd_le_r }, - { NULL, NULL, - gen_helper_sve_st1ss_le_r, - gen_helper_sve_st1sd_le_r }, - { NULL, NULL, NULL, - gen_helper_sve_st1dd_le_r } }, - { { gen_helper_sve_st1bb_r, - gen_helper_sve_st1bh_r, - gen_helper_sve_st1bs_r, - gen_helper_sve_st1bd_r }, - { NULL, - gen_helper_sve_st1hh_be_r, - gen_helper_sve_st1hs_be_r, - gen_helper_sve_st1hd_be_r }, - { NULL, NULL, - gen_helper_sve_st1ss_be_r, - gen_helper_sve_st1sd_be_r }, - { NULL, NULL, NULL, - gen_helper_sve_st1dd_be_r } }, + static gen_helper_gvec_mem * const fn_single[2][2][4][4] = { + { { { gen_helper_sve_st1bb_r, + gen_helper_sve_st1bh_r, + gen_helper_sve_st1bs_r, + gen_helper_sve_st1bd_r }, + { NULL, + gen_helper_sve_st1hh_le_r, + gen_helper_sve_st1hs_le_r, + gen_helper_sve_st1hd_le_r }, + { NULL, NULL, + gen_helper_sve_st1ss_le_r, + gen_helper_sve_st1sd_le_r }, + { NULL, NULL, NULL, + gen_helper_sve_st1dd_le_r } }, + { { gen_helper_sve_st1bb_r, + gen_helper_sve_st1bh_r, + gen_helper_sve_st1bs_r, + gen_helper_sve_st1bd_r }, + { NULL, + gen_helper_sve_st1hh_be_r, + gen_helper_sve_st1hs_be_r, + gen_helper_sve_st1hd_be_r }, + { NULL, NULL, + gen_helper_sve_st1ss_be_r, + gen_helper_sve_st1sd_be_r }, + { NULL, NULL, NULL, + gen_helper_sve_st1dd_be_r } } }, + + { { { gen_helper_sve_st1bb_r_mte, + gen_helper_sve_st1bh_r_mte, + gen_helper_sve_st1bs_r_mte, + gen_helper_sve_st1bd_r_mte }, + { NULL, + gen_helper_sve_st1hh_le_r_mte, + gen_helper_sve_st1hs_le_r_mte, + gen_helper_sve_st1hd_le_r_mte }, + { NULL, NULL, + gen_helper_sve_st1ss_le_r_mte, + gen_helper_sve_st1sd_le_r_mte }, + { NULL, NULL, NULL, + gen_helper_sve_st1dd_le_r_mte } }, + { { gen_helper_sve_st1bb_r_mte, + gen_helper_sve_st1bh_r_mte, + gen_helper_sve_st1bs_r_mte, + gen_helper_sve_st1bd_r_mte }, + { NULL, + gen_helper_sve_st1hh_be_r_mte, + gen_helper_sve_st1hs_be_r_mte, + gen_helper_sve_st1hd_be_r_mte }, + { NULL, NULL, + gen_helper_sve_st1ss_be_r_mte, + gen_helper_sve_st1sd_be_r_mte }, + { NULL, NULL, NULL, + gen_helper_sve_st1dd_be_r_mte } } }, }; - static gen_helper_gvec_mem * const fn_multiple[2][3][4] = { - { { gen_helper_sve_st2bb_r, - gen_helper_sve_st2hh_le_r, - gen_helper_sve_st2ss_le_r, - gen_helper_sve_st2dd_le_r }, - { gen_helper_sve_st3bb_r, - gen_helper_sve_st3hh_le_r, - gen_helper_sve_st3ss_le_r, - gen_helper_sve_st3dd_le_r }, - { gen_helper_sve_st4bb_r, - gen_helper_sve_st4hh_le_r, - gen_helper_sve_st4ss_le_r, - gen_helper_sve_st4dd_le_r } }, - { { gen_helper_sve_st2bb_r, - gen_helper_sve_st2hh_be_r, - gen_helper_sve_st2ss_be_r, - gen_helper_sve_st2dd_be_r }, - { gen_helper_sve_st3bb_r, - gen_helper_sve_st3hh_be_r, - gen_helper_sve_st3ss_be_r, - gen_helper_sve_st3dd_be_r }, - { gen_helper_sve_st4bb_r, - gen_helper_sve_st4hh_be_r, - gen_helper_sve_st4ss_be_r, - gen_helper_sve_st4dd_be_r } }, + static gen_helper_gvec_mem * const fn_multiple[2][2][3][4] = { + { { { gen_helper_sve_st2bb_r, + gen_helper_sve_st2hh_le_r, + gen_helper_sve_st2ss_le_r, + gen_helper_sve_st2dd_le_r }, + { gen_helper_sve_st3bb_r, + gen_helper_sve_st3hh_le_r, + gen_helper_sve_st3ss_le_r, + gen_helper_sve_st3dd_le_r }, + { gen_helper_sve_st4bb_r, + gen_helper_sve_st4hh_le_r, + gen_helper_sve_st4ss_le_r, + gen_helper_sve_st4dd_le_r } }, + { { gen_helper_sve_st2bb_r, + gen_helper_sve_st2hh_be_r, + gen_helper_sve_st2ss_be_r, + gen_helper_sve_st2dd_be_r }, + { gen_helper_sve_st3bb_r, + gen_helper_sve_st3hh_be_r, + gen_helper_sve_st3ss_be_r, + gen_helper_sve_st3dd_be_r }, + { gen_helper_sve_st4bb_r, + gen_helper_sve_st4hh_be_r, + gen_helper_sve_st4ss_be_r, + gen_helper_sve_st4dd_be_r } } }, + { { { gen_helper_sve_st2bb_r_mte, + gen_helper_sve_st2hh_le_r_mte, + gen_helper_sve_st2ss_le_r_mte, + gen_helper_sve_st2dd_le_r_mte }, + { gen_helper_sve_st3bb_r_mte, + gen_helper_sve_st3hh_le_r_mte, + gen_helper_sve_st3ss_le_r_mte, + gen_helper_sve_st3dd_le_r_mte }, + { gen_helper_sve_st4bb_r_mte, + gen_helper_sve_st4hh_le_r_mte, + gen_helper_sve_st4ss_le_r_mte, + gen_helper_sve_st4dd_le_r_mte } }, + { { gen_helper_sve_st2bb_r_mte, + gen_helper_sve_st2hh_be_r_mte, + gen_helper_sve_st2ss_be_r_mte, + gen_helper_sve_st2dd_be_r_mte }, + { gen_helper_sve_st3bb_r_mte, + gen_helper_sve_st3hh_be_r_mte, + gen_helper_sve_st3ss_be_r_mte, + gen_helper_sve_st3dd_be_r_mte }, + { gen_helper_sve_st4bb_r_mte, + gen_helper_sve_st4hh_be_r_mte, + gen_helper_sve_st4ss_be_r_mte, + gen_helper_sve_st4dd_be_r_mte } } }, }; gen_helper_gvec_mem *fn; int be = s->be_data == MO_BE; if (nreg == 0) { /* ST1 */ - fn = fn_single[be][msz][esz]; + fn = fn_single[s->mte_active[0]][be][msz][esz]; + nreg = 1; } else { /* ST2, ST3, ST4 -- msz == esz, enforced by encoding */ assert(msz == esz); - fn = fn_multiple[be][nreg - 1][msz]; + fn = fn_multiple[s->mte_active[0]][be][nreg - 1][msz]; } assert(fn != NULL); - do_mem_zpa(s, zt, pg, addr, msz_dtype(s, msz), 0, true, fn); + do_mem_zpa(s, zt, pg, addr, msz_dtype(s, msz), nreg, true, fn); } static bool trans_ST_zprr(DisasContext *s, arg_rprr_store *a) -- 2.20.1 From: Richard Henderson <richard.henderson@linaro.org> Because the elements are sequential, we can eliminate many tests all at once when the tag hits TCMA, or if the page(s) are not Tagged. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200626033144.790098-36-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/helper-sve.h | 98 ++++++++++++++++ target/arm/sve_helper.c | 99 ++++++++++++++-- target/arm/translate-sve.c | 232 +++++++++++++++++++++++++------------ 3 files changed, 343 insertions(+), 86 deletions(-) diff --git a/target/arm/helper-sve.h b/target/arm/helper-sve.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper-sve.h +++ b/target/arm/helper-sve.h @@ -XXX,XX +XXX,XX @@ DEF_HELPER_FLAGS_4(sve_ldff1sds_be_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) DEF_HELPER_FLAGS_4(sve_ldff1dd_le_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) DEF_HELPER_FLAGS_4(sve_ldff1dd_be_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ldff1bb_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ldff1bhu_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ldff1bsu_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ldff1bdu_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ldff1bhs_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ldff1bss_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ldff1bds_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) + +DEF_HELPER_FLAGS_4(sve_ldff1hh_le_r_mte, TCG_CALL_NO_WG, + void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ldff1hsu_le_r_mte, TCG_CALL_NO_WG, + void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ldff1hdu_le_r_mte, TCG_CALL_NO_WG, + void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ldff1hss_le_r_mte, TCG_CALL_NO_WG, + void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ldff1hds_le_r_mte, TCG_CALL_NO_WG, + void, env, ptr, tl, i32) + +DEF_HELPER_FLAGS_4(sve_ldff1hh_be_r_mte, TCG_CALL_NO_WG, + void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ldff1hsu_be_r_mte, TCG_CALL_NO_WG, + void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ldff1hdu_be_r_mte, TCG_CALL_NO_WG, + void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ldff1hss_be_r_mte, TCG_CALL_NO_WG, + void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ldff1hds_be_r_mte, TCG_CALL_NO_WG, + void, env, ptr, tl, i32) + +DEF_HELPER_FLAGS_4(sve_ldff1ss_le_r_mte, TCG_CALL_NO_WG, + void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ldff1sdu_le_r_mte, TCG_CALL_NO_WG, + void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ldff1sds_le_r_mte, TCG_CALL_NO_WG, + void, env, ptr, tl, i32) + +DEF_HELPER_FLAGS_4(sve_ldff1ss_be_r_mte, TCG_CALL_NO_WG, + void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ldff1sdu_be_r_mte, TCG_CALL_NO_WG, + void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ldff1sds_be_r_mte, TCG_CALL_NO_WG, + void, env, ptr, tl, i32) + +DEF_HELPER_FLAGS_4(sve_ldff1dd_le_r_mte, TCG_CALL_NO_WG, + void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ldff1dd_be_r_mte, TCG_CALL_NO_WG, + void, env, ptr, tl, i32) + DEF_HELPER_FLAGS_4(sve_ldnf1bb_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) DEF_HELPER_FLAGS_4(sve_ldnf1bhu_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) DEF_HELPER_FLAGS_4(sve_ldnf1bsu_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) @@ -XXX,XX +XXX,XX @@ DEF_HELPER_FLAGS_4(sve_ldnf1sds_be_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) DEF_HELPER_FLAGS_4(sve_ldnf1dd_le_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) DEF_HELPER_FLAGS_4(sve_ldnf1dd_be_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ldnf1bb_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ldnf1bhu_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ldnf1bsu_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ldnf1bdu_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ldnf1bhs_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ldnf1bss_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ldnf1bds_r_mte, TCG_CALL_NO_WG, void, env, ptr, tl, i32) + +DEF_HELPER_FLAGS_4(sve_ldnf1hh_le_r_mte, TCG_CALL_NO_WG, + void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ldnf1hsu_le_r_mte, TCG_CALL_NO_WG, + void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ldnf1hdu_le_r_mte, TCG_CALL_NO_WG, + void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ldnf1hss_le_r_mte, TCG_CALL_NO_WG, + void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ldnf1hds_le_r_mte, TCG_CALL_NO_WG, + void, env, ptr, tl, i32) + +DEF_HELPER_FLAGS_4(sve_ldnf1hh_be_r_mte, TCG_CALL_NO_WG, + void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ldnf1hsu_be_r_mte, TCG_CALL_NO_WG, + void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ldnf1hdu_be_r_mte, TCG_CALL_NO_WG, + void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ldnf1hss_be_r_mte, TCG_CALL_NO_WG, + void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ldnf1hds_be_r_mte, TCG_CALL_NO_WG, + void, env, ptr, tl, i32) + +DEF_HELPER_FLAGS_4(sve_ldnf1ss_le_r_mte, TCG_CALL_NO_WG, + void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ldnf1sdu_le_r_mte, TCG_CALL_NO_WG, + void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ldnf1sds_le_r_mte, TCG_CALL_NO_WG, + void, env, ptr, tl, i32) + +DEF_HELPER_FLAGS_4(sve_ldnf1ss_be_r_mte, TCG_CALL_NO_WG, + void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ldnf1sdu_be_r_mte, TCG_CALL_NO_WG, + void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ldnf1sds_be_r_mte, TCG_CALL_NO_WG, + void, env, ptr, tl, i32) + +DEF_HELPER_FLAGS_4(sve_ldnf1dd_le_r_mte, TCG_CALL_NO_WG, + void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ldnf1dd_be_r_mte, TCG_CALL_NO_WG, + void, env, ptr, tl, i32) + DEF_HELPER_FLAGS_4(sve_st1bb_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) DEF_HELPER_FLAGS_4(sve_st2bb_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) DEF_HELPER_FLAGS_4(sve_st3bb_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) diff --git a/target/arm/sve_helper.c b/target/arm/sve_helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/sve_helper.c +++ b/target/arm/sve_helper.c @@ -XXX,XX +XXX,XX @@ static void record_fault(CPUARMState *env, uintptr_t i, uintptr_t oprsz) */ static inline QEMU_ALWAYS_INLINE void sve_ldnfff1_r(CPUARMState *env, void *vg, const target_ulong addr, - uint32_t desc, const uintptr_t retaddr, + uint32_t desc, const uintptr_t retaddr, uint32_t mtedesc, const int esz, const int msz, const SVEContFault fault, sve_ldst1_host_fn *host_fn, sve_ldst1_tlb_fn *tlb_fn) @@ -XXX,XX +XXX,XX @@ void sve_ldnfff1_r(CPUARMState *env, void *vg, const target_ulong addr, mem_off = info.mem_off_first[0]; flags = info.page[0].flags; + /* + * Disable MTE checking if the Tagged bit is not set. Since TBI must + * be set within MTEDESC for MTE, !mtedesc => !mte_active. + */ + if (arm_tlb_mte_tagged(&info.page[0].attrs)) { + mtedesc = 0; + } + if (fault == FAULT_FIRST) { + /* Trapping mte check for the first-fault element. */ + if (mtedesc) { + mte_check1(env, mtedesc, addr + mem_off, retaddr); + } + /* * Special handling of the first active element, * if it crosses a page boundary or is MMIO. */ bool is_split = mem_off == info.mem_off_split; - /* TODO: MTE check. */ if (unlikely(flags != 0) || unlikely(is_split)) { /* * Use the slow path for cross-page handling. @@ -XXX,XX +XXX,XX @@ void sve_ldnfff1_r(CPUARMState *env, void *vg, const target_ulong addr, /* Watchpoint hit, see below. */ goto do_fault; } - /* TODO: MTE check. */ + if (mtedesc && !mte_probe1(env, mtedesc, addr + mem_off)) { + goto do_fault; + } /* * Use the slow path for cross-page handling. * This is RAM, without a watchpoint, and will not trap. @@ -XXX,XX +XXX,XX @@ void sve_ldnfff1_r(CPUARMState *env, void *vg, const target_ulong addr, & BP_MEM_READ)) { goto do_fault; } - /* TODO: MTE check. */ + if (mtedesc && !mte_probe1(env, mtedesc, addr + mem_off)) { + goto do_fault; + } host_fn(vd, reg_off, host + mem_off); } reg_off += 1 << esz; @@ -XXX,XX +XXX,XX @@ void sve_ldnfff1_r(CPUARMState *env, void *vg, const target_ulong addr, record_fault(env, reg_off, reg_max); } -#define DO_LDFF1_LDNF1_1(PART, ESZ) \ +static inline QEMU_ALWAYS_INLINE +void sve_ldnfff1_r_mte(CPUARMState *env, void *vg, target_ulong addr, + uint32_t desc, const uintptr_t retaddr, + const int esz, const int msz, const SVEContFault fault, + sve_ldst1_host_fn *host_fn, + sve_ldst1_tlb_fn *tlb_fn) +{ + uint32_t mtedesc = desc >> (SIMD_DATA_SHIFT + SVE_MTEDESC_SHIFT); + int bit55 = extract64(addr, 55, 1); + + /* Remove mtedesc from the normal sve descriptor. */ + desc = extract32(desc, 0, SIMD_DATA_SHIFT + SVE_MTEDESC_SHIFT); + + /* Perform gross MTE suppression early. */ + if (!tbi_check(desc, bit55) || + tcma_check(desc, bit55, allocation_tag_from_addr(addr))) { + mtedesc = 0; + } + + sve_ldnfff1_r(env, vg, addr, desc, retaddr, mtedesc, + esz, msz, fault, host_fn, tlb_fn); +} + +#define DO_LDFF1_LDNF1_1(PART, ESZ) \ void HELPER(sve_ldff1##PART##_r)(CPUARMState *env, void *vg, \ target_ulong addr, uint32_t desc) \ { \ - sve_ldnfff1_r(env, vg, addr, desc, GETPC(), ESZ, MO_8, FAULT_FIRST, \ + sve_ldnfff1_r(env, vg, addr, desc, GETPC(), 0, ESZ, MO_8, FAULT_FIRST, \ sve_ld1##PART##_host, sve_ld1##PART##_tlb); \ } \ void HELPER(sve_ldnf1##PART##_r)(CPUARMState *env, void *vg, \ target_ulong addr, uint32_t desc) \ { \ - sve_ldnfff1_r(env, vg, addr, desc, GETPC(), ESZ, MO_8, FAULT_NO, \ + sve_ldnfff1_r(env, vg, addr, desc, GETPC(), 0, ESZ, MO_8, FAULT_NO, \ + sve_ld1##PART##_host, sve_ld1##PART##_tlb); \ +} \ +void HELPER(sve_ldff1##PART##_r_mte)(CPUARMState *env, void *vg, \ + target_ulong addr, uint32_t desc) \ +{ \ + sve_ldnfff1_r_mte(env, vg, addr, desc, GETPC(), ESZ, MO_8, FAULT_FIRST, \ + sve_ld1##PART##_host, sve_ld1##PART##_tlb); \ +} \ +void HELPER(sve_ldnf1##PART##_r_mte)(CPUARMState *env, void *vg, \ + target_ulong addr, uint32_t desc) \ +{ \ + sve_ldnfff1_r_mte(env, vg, addr, desc, GETPC(), ESZ, MO_8, FAULT_NO, \ sve_ld1##PART##_host, sve_ld1##PART##_tlb); \ } -#define DO_LDFF1_LDNF1_2(PART, ESZ, MSZ) \ +#define DO_LDFF1_LDNF1_2(PART, ESZ, MSZ) \ void HELPER(sve_ldff1##PART##_le_r)(CPUARMState *env, void *vg, \ target_ulong addr, uint32_t desc) \ { \ - sve_ldnfff1_r(env, vg, addr, desc, GETPC(), ESZ, MSZ, FAULT_FIRST, \ + sve_ldnfff1_r(env, vg, addr, desc, GETPC(), 0, ESZ, MSZ, FAULT_FIRST, \ sve_ld1##PART##_le_host, sve_ld1##PART##_le_tlb); \ } \ void HELPER(sve_ldnf1##PART##_le_r)(CPUARMState *env, void *vg, \ target_ulong addr, uint32_t desc) \ { \ - sve_ldnfff1_r(env, vg, addr, desc, GETPC(), ESZ, MSZ, FAULT_NO, \ + sve_ldnfff1_r(env, vg, addr, desc, GETPC(), 0, ESZ, MSZ, FAULT_NO, \ sve_ld1##PART##_le_host, sve_ld1##PART##_le_tlb); \ } \ void HELPER(sve_ldff1##PART##_be_r)(CPUARMState *env, void *vg, \ target_ulong addr, uint32_t desc) \ { \ - sve_ldnfff1_r(env, vg, addr, desc, GETPC(), ESZ, MSZ, FAULT_FIRST, \ + sve_ldnfff1_r(env, vg, addr, desc, GETPC(), 0, ESZ, MSZ, FAULT_FIRST, \ sve_ld1##PART##_be_host, sve_ld1##PART##_be_tlb); \ } \ void HELPER(sve_ldnf1##PART##_be_r)(CPUARMState *env, void *vg, \ target_ulong addr, uint32_t desc) \ { \ - sve_ldnfff1_r(env, vg, addr, desc, GETPC(), ESZ, MSZ, FAULT_NO, \ + sve_ldnfff1_r(env, vg, addr, desc, GETPC(), 0, ESZ, MSZ, FAULT_NO, \ sve_ld1##PART##_be_host, sve_ld1##PART##_be_tlb); \ +} \ +void HELPER(sve_ldff1##PART##_le_r_mte)(CPUARMState *env, void *vg, \ + target_ulong addr, uint32_t desc) \ +{ \ + sve_ldnfff1_r_mte(env, vg, addr, desc, GETPC(), ESZ, MSZ, FAULT_FIRST, \ + sve_ld1##PART##_le_host, sve_ld1##PART##_le_tlb); \ +} \ +void HELPER(sve_ldnf1##PART##_le_r_mte)(CPUARMState *env, void *vg, \ + target_ulong addr, uint32_t desc) \ +{ \ + sve_ldnfff1_r_mte(env, vg, addr, desc, GETPC(), ESZ, MSZ, FAULT_NO, \ + sve_ld1##PART##_le_host, sve_ld1##PART##_le_tlb); \ +} \ +void HELPER(sve_ldff1##PART##_be_r_mte)(CPUARMState *env, void *vg, \ + target_ulong addr, uint32_t desc) \ +{ \ + sve_ldnfff1_r_mte(env, vg, addr, desc, GETPC(), ESZ, MSZ, FAULT_FIRST, \ + sve_ld1##PART##_be_host, sve_ld1##PART##_be_tlb); \ +} \ +void HELPER(sve_ldnf1##PART##_be_r_mte)(CPUARMState *env, void *vg, \ + target_ulong addr, uint32_t desc) \ +{ \ + sve_ldnfff1_r_mte(env, vg, addr, desc, GETPC(), ESZ, MSZ, FAULT_NO, \ + sve_ld1##PART##_be_host, sve_ld1##PART##_be_tlb); \ } DO_LDFF1_LDNF1_1(bb, MO_8) diff --git a/target/arm/translate-sve.c b/target/arm/translate-sve.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-sve.c +++ b/target/arm/translate-sve.c @@ -XXX,XX +XXX,XX @@ static bool trans_LD_zpri(DisasContext *s, arg_rpri_load *a) static bool trans_LDFF1_zprr(DisasContext *s, arg_rprr_load *a) { - static gen_helper_gvec_mem * const fns[2][16] = { - /* Little-endian */ - { gen_helper_sve_ldff1bb_r, - gen_helper_sve_ldff1bhu_r, - gen_helper_sve_ldff1bsu_r, - gen_helper_sve_ldff1bdu_r, + static gen_helper_gvec_mem * const fns[2][2][16] = { + { /* mte inactive, little-endian */ + { gen_helper_sve_ldff1bb_r, + gen_helper_sve_ldff1bhu_r, + gen_helper_sve_ldff1bsu_r, + gen_helper_sve_ldff1bdu_r, - gen_helper_sve_ldff1sds_le_r, - gen_helper_sve_ldff1hh_le_r, - gen_helper_sve_ldff1hsu_le_r, - gen_helper_sve_ldff1hdu_le_r, + gen_helper_sve_ldff1sds_le_r, + gen_helper_sve_ldff1hh_le_r, + gen_helper_sve_ldff1hsu_le_r, + gen_helper_sve_ldff1hdu_le_r, - gen_helper_sve_ldff1hds_le_r, - gen_helper_sve_ldff1hss_le_r, - gen_helper_sve_ldff1ss_le_r, - gen_helper_sve_ldff1sdu_le_r, + gen_helper_sve_ldff1hds_le_r, + gen_helper_sve_ldff1hss_le_r, + gen_helper_sve_ldff1ss_le_r, + gen_helper_sve_ldff1sdu_le_r, - gen_helper_sve_ldff1bds_r, - gen_helper_sve_ldff1bss_r, - gen_helper_sve_ldff1bhs_r, - gen_helper_sve_ldff1dd_le_r }, + gen_helper_sve_ldff1bds_r, + gen_helper_sve_ldff1bss_r, + gen_helper_sve_ldff1bhs_r, + gen_helper_sve_ldff1dd_le_r }, - /* Big-endian */ - { gen_helper_sve_ldff1bb_r, - gen_helper_sve_ldff1bhu_r, - gen_helper_sve_ldff1bsu_r, - gen_helper_sve_ldff1bdu_r, + /* mte inactive, big-endian */ + { gen_helper_sve_ldff1bb_r, + gen_helper_sve_ldff1bhu_r, + gen_helper_sve_ldff1bsu_r, + gen_helper_sve_ldff1bdu_r, - gen_helper_sve_ldff1sds_be_r, - gen_helper_sve_ldff1hh_be_r, - gen_helper_sve_ldff1hsu_be_r, - gen_helper_sve_ldff1hdu_be_r, + gen_helper_sve_ldff1sds_be_r, + gen_helper_sve_ldff1hh_be_r, + gen_helper_sve_ldff1hsu_be_r, + gen_helper_sve_ldff1hdu_be_r, - gen_helper_sve_ldff1hds_be_r, - gen_helper_sve_ldff1hss_be_r, - gen_helper_sve_ldff1ss_be_r, - gen_helper_sve_ldff1sdu_be_r, + gen_helper_sve_ldff1hds_be_r, + gen_helper_sve_ldff1hss_be_r, + gen_helper_sve_ldff1ss_be_r, + gen_helper_sve_ldff1sdu_be_r, - gen_helper_sve_ldff1bds_r, - gen_helper_sve_ldff1bss_r, - gen_helper_sve_ldff1bhs_r, - gen_helper_sve_ldff1dd_be_r }, + gen_helper_sve_ldff1bds_r, + gen_helper_sve_ldff1bss_r, + gen_helper_sve_ldff1bhs_r, + gen_helper_sve_ldff1dd_be_r } }, + + { /* mte active, little-endian */ + { gen_helper_sve_ldff1bb_r_mte, + gen_helper_sve_ldff1bhu_r_mte, + gen_helper_sve_ldff1bsu_r_mte, + gen_helper_sve_ldff1bdu_r_mte, + + gen_helper_sve_ldff1sds_le_r_mte, + gen_helper_sve_ldff1hh_le_r_mte, + gen_helper_sve_ldff1hsu_le_r_mte, + gen_helper_sve_ldff1hdu_le_r_mte, + + gen_helper_sve_ldff1hds_le_r_mte, + gen_helper_sve_ldff1hss_le_r_mte, + gen_helper_sve_ldff1ss_le_r_mte, + gen_helper_sve_ldff1sdu_le_r_mte, + + gen_helper_sve_ldff1bds_r_mte, + gen_helper_sve_ldff1bss_r_mte, + gen_helper_sve_ldff1bhs_r_mte, + gen_helper_sve_ldff1dd_le_r_mte }, + + /* mte active, big-endian */ + { gen_helper_sve_ldff1bb_r_mte, + gen_helper_sve_ldff1bhu_r_mte, + gen_helper_sve_ldff1bsu_r_mte, + gen_helper_sve_ldff1bdu_r_mte, + + gen_helper_sve_ldff1sds_be_r_mte, + gen_helper_sve_ldff1hh_be_r_mte, + gen_helper_sve_ldff1hsu_be_r_mte, + gen_helper_sve_ldff1hdu_be_r_mte, + + gen_helper_sve_ldff1hds_be_r_mte, + gen_helper_sve_ldff1hss_be_r_mte, + gen_helper_sve_ldff1ss_be_r_mte, + gen_helper_sve_ldff1sdu_be_r_mte, + + gen_helper_sve_ldff1bds_r_mte, + gen_helper_sve_ldff1bss_r_mte, + gen_helper_sve_ldff1bhs_r_mte, + gen_helper_sve_ldff1dd_be_r_mte } }, }; if (sve_access_check(s)) { TCGv_i64 addr = new_tmp_a64(s); tcg_gen_shli_i64(addr, cpu_reg(s, a->rm), dtype_msz(a->dtype)); tcg_gen_add_i64(addr, addr, cpu_reg_sp(s, a->rn)); - do_mem_zpa(s, a->rd, a->pg, addr, a->dtype, 0, false, - fns[s->be_data == MO_BE][a->dtype]); + do_mem_zpa(s, a->rd, a->pg, addr, a->dtype, 1, false, + fns[s->mte_active[0]][s->be_data == MO_BE][a->dtype]); } return true; } static bool trans_LDNF1_zpri(DisasContext *s, arg_rpri_load *a) { - static gen_helper_gvec_mem * const fns[2][16] = { - /* Little-endian */ - { gen_helper_sve_ldnf1bb_r, - gen_helper_sve_ldnf1bhu_r, - gen_helper_sve_ldnf1bsu_r, - gen_helper_sve_ldnf1bdu_r, + static gen_helper_gvec_mem * const fns[2][2][16] = { + { /* mte inactive, little-endian */ + { gen_helper_sve_ldnf1bb_r, + gen_helper_sve_ldnf1bhu_r, + gen_helper_sve_ldnf1bsu_r, + gen_helper_sve_ldnf1bdu_r, - gen_helper_sve_ldnf1sds_le_r, - gen_helper_sve_ldnf1hh_le_r, - gen_helper_sve_ldnf1hsu_le_r, - gen_helper_sve_ldnf1hdu_le_r, + gen_helper_sve_ldnf1sds_le_r, + gen_helper_sve_ldnf1hh_le_r, + gen_helper_sve_ldnf1hsu_le_r, + gen_helper_sve_ldnf1hdu_le_r, - gen_helper_sve_ldnf1hds_le_r, - gen_helper_sve_ldnf1hss_le_r, - gen_helper_sve_ldnf1ss_le_r, - gen_helper_sve_ldnf1sdu_le_r, + gen_helper_sve_ldnf1hds_le_r, + gen_helper_sve_ldnf1hss_le_r, + gen_helper_sve_ldnf1ss_le_r, + gen_helper_sve_ldnf1sdu_le_r, - gen_helper_sve_ldnf1bds_r, - gen_helper_sve_ldnf1bss_r, - gen_helper_sve_ldnf1bhs_r, - gen_helper_sve_ldnf1dd_le_r }, + gen_helper_sve_ldnf1bds_r, + gen_helper_sve_ldnf1bss_r, + gen_helper_sve_ldnf1bhs_r, + gen_helper_sve_ldnf1dd_le_r }, - /* Big-endian */ - { gen_helper_sve_ldnf1bb_r, - gen_helper_sve_ldnf1bhu_r, - gen_helper_sve_ldnf1bsu_r, - gen_helper_sve_ldnf1bdu_r, + /* mte inactive, big-endian */ + { gen_helper_sve_ldnf1bb_r, + gen_helper_sve_ldnf1bhu_r, + gen_helper_sve_ldnf1bsu_r, + gen_helper_sve_ldnf1bdu_r, - gen_helper_sve_ldnf1sds_be_r, - gen_helper_sve_ldnf1hh_be_r, - gen_helper_sve_ldnf1hsu_be_r, - gen_helper_sve_ldnf1hdu_be_r, + gen_helper_sve_ldnf1sds_be_r, + gen_helper_sve_ldnf1hh_be_r, + gen_helper_sve_ldnf1hsu_be_r, + gen_helper_sve_ldnf1hdu_be_r, - gen_helper_sve_ldnf1hds_be_r, - gen_helper_sve_ldnf1hss_be_r, - gen_helper_sve_ldnf1ss_be_r, - gen_helper_sve_ldnf1sdu_be_r, + gen_helper_sve_ldnf1hds_be_r, + gen_helper_sve_ldnf1hss_be_r, + gen_helper_sve_ldnf1ss_be_r, + gen_helper_sve_ldnf1sdu_be_r, - gen_helper_sve_ldnf1bds_r, - gen_helper_sve_ldnf1bss_r, - gen_helper_sve_ldnf1bhs_r, - gen_helper_sve_ldnf1dd_be_r }, + gen_helper_sve_ldnf1bds_r, + gen_helper_sve_ldnf1bss_r, + gen_helper_sve_ldnf1bhs_r, + gen_helper_sve_ldnf1dd_be_r } }, + + { /* mte inactive, little-endian */ + { gen_helper_sve_ldnf1bb_r_mte, + gen_helper_sve_ldnf1bhu_r_mte, + gen_helper_sve_ldnf1bsu_r_mte, + gen_helper_sve_ldnf1bdu_r_mte, + + gen_helper_sve_ldnf1sds_le_r_mte, + gen_helper_sve_ldnf1hh_le_r_mte, + gen_helper_sve_ldnf1hsu_le_r_mte, + gen_helper_sve_ldnf1hdu_le_r_mte, + + gen_helper_sve_ldnf1hds_le_r_mte, + gen_helper_sve_ldnf1hss_le_r_mte, + gen_helper_sve_ldnf1ss_le_r_mte, + gen_helper_sve_ldnf1sdu_le_r_mte, + + gen_helper_sve_ldnf1bds_r_mte, + gen_helper_sve_ldnf1bss_r_mte, + gen_helper_sve_ldnf1bhs_r_mte, + gen_helper_sve_ldnf1dd_le_r_mte }, + + /* mte inactive, big-endian */ + { gen_helper_sve_ldnf1bb_r_mte, + gen_helper_sve_ldnf1bhu_r_mte, + gen_helper_sve_ldnf1bsu_r_mte, + gen_helper_sve_ldnf1bdu_r_mte, + + gen_helper_sve_ldnf1sds_be_r_mte, + gen_helper_sve_ldnf1hh_be_r_mte, + gen_helper_sve_ldnf1hsu_be_r_mte, + gen_helper_sve_ldnf1hdu_be_r_mte, + + gen_helper_sve_ldnf1hds_be_r_mte, + gen_helper_sve_ldnf1hss_be_r_mte, + gen_helper_sve_ldnf1ss_be_r_mte, + gen_helper_sve_ldnf1sdu_be_r_mte, + + gen_helper_sve_ldnf1bds_r_mte, + gen_helper_sve_ldnf1bss_r_mte, + gen_helper_sve_ldnf1bhs_r_mte, + gen_helper_sve_ldnf1dd_be_r_mte } }, }; if (sve_access_check(s)) { @@ -XXX,XX +XXX,XX @@ static bool trans_LDNF1_zpri(DisasContext *s, arg_rpri_load *a) TCGv_i64 addr = new_tmp_a64(s); tcg_gen_addi_i64(addr, cpu_reg_sp(s, a->rn), off); - do_mem_zpa(s, a->rd, a->pg, addr, a->dtype, 0, false, - fns[s->be_data == MO_BE][a->dtype]); + do_mem_zpa(s, a->rd, a->pg, addr, a->dtype, 1, false, + fns[s->mte_active[0]][s->be_data == MO_BE][a->dtype]); } return true; } -- 2.20.1 From: Richard Henderson <richard.henderson@linaro.org> We still need to handle tbi for user-only when mte is inactive. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200626033144.790098-37-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/translate-a64.h | 1 + target/arm/translate-a64.c | 2 +- target/arm/translate-sve.c | 6 ++++-- 3 files changed, 6 insertions(+), 3 deletions(-) diff --git a/target/arm/translate-a64.h b/target/arm/translate-a64.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-a64.h +++ b/target/arm/translate-a64.h @@ -XXX,XX +XXX,XX @@ TCGv_ptr get_fpstatus_ptr(bool); bool logic_imm_decode_wmask(uint64_t *result, unsigned int immn, unsigned int imms, unsigned int immr); bool sve_access_check(DisasContext *s); +TCGv_i64 clean_data_tbi(DisasContext *s, TCGv_i64 addr); TCGv_i64 gen_mte_check1(DisasContext *s, TCGv_i64 addr, bool is_write, bool tag_checked, int log2_size); TCGv_i64 gen_mte_checkN(DisasContext *s, TCGv_i64 addr, bool is_write, diff --git a/target/arm/translate-a64.c b/target/arm/translate-a64.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-a64.c +++ b/target/arm/translate-a64.c @@ -XXX,XX +XXX,XX @@ static void gen_a64_set_pc(DisasContext *s, TCGv_i64 src) * of the write-back address. */ -static TCGv_i64 clean_data_tbi(DisasContext *s, TCGv_i64 addr) +TCGv_i64 clean_data_tbi(DisasContext *s, TCGv_i64 addr) { TCGv_i64 clean = new_tmp_a64(s); #ifdef CONFIG_USER_ONLY diff --git a/target/arm/translate-sve.c b/target/arm/translate-sve.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-sve.c +++ b/target/arm/translate-sve.c @@ -XXX,XX +XXX,XX @@ static void do_mem_zpa(DisasContext *s, int zt, int pg, TCGv_i64 addr, * For e.g. LD4, there are not enough arguments to pass all 4 * registers as pointers, so encode the regno into the data field. * For consistency, do this even for LD1. - * TODO: mte_n check here while callers are updated. */ - if (mte_n && s->mte_active[0]) { + if (s->mte_active[0]) { int msz = dtype_msz(dtype); desc = FIELD_DP32(desc, MTEDESC, MIDX, get_mem_index(s)); @@ -XXX,XX +XXX,XX @@ static void do_mem_zpa(DisasContext *s, int zt, int pg, TCGv_i64 addr, desc = FIELD_DP32(desc, MTEDESC, ESIZE, 1 << msz); desc = FIELD_DP32(desc, MTEDESC, TSIZE, mte_n << msz); desc <<= SVE_MTEDESC_SHIFT; + } else { + addr = clean_data_tbi(s, addr); } + desc = simd_desc(vsz, vsz, zt | desc); t_desc = tcg_const_i32(desc); t_pg = tcg_temp_new_ptr(); -- 2.20.1 From: Richard Henderson <richard.henderson@linaro.org> Because the elements are non-sequential, we cannot eliminate many tests straight away like we can for sequential operations. But we often have the PTE details handy, so we can test for Tagged. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200626033144.790098-38-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/helper-sve.h | 285 ++++++++++++++++ target/arm/sve_helper.c | 185 +++++++++-- target/arm/translate-sve.c | 650 +++++++++++++++++++++++++------------ 3 files changed, 872 insertions(+), 248 deletions(-) diff --git a/target/arm/helper-sve.h b/target/arm/helper-sve.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper-sve.h +++ b/target/arm/helper-sve.h @@ -XXX,XX +XXX,XX @@ DEF_HELPER_FLAGS_6(sve_ldsds_le_zd, TCG_CALL_NO_WG, DEF_HELPER_FLAGS_6(sve_ldsds_be_zd, TCG_CALL_NO_WG, void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldbsu_zsu_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldhsu_le_zsu_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldhsu_be_zsu_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldss_le_zsu_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldss_be_zsu_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldbss_zsu_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldhss_le_zsu_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldhss_be_zsu_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) + +DEF_HELPER_FLAGS_6(sve_ldbsu_zss_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldhsu_le_zss_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldhsu_be_zss_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldss_le_zss_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldss_be_zss_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldbss_zss_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldhss_le_zss_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldhss_be_zss_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) + +DEF_HELPER_FLAGS_6(sve_ldbdu_zsu_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldhdu_le_zsu_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldhdu_be_zsu_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldsdu_le_zsu_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldsdu_be_zsu_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_lddd_le_zsu_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_lddd_be_zsu_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldbds_zsu_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldhds_le_zsu_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldhds_be_zsu_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldsds_le_zsu_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldsds_be_zsu_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) + +DEF_HELPER_FLAGS_6(sve_ldbdu_zss_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldhdu_le_zss_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldhdu_be_zss_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldsdu_le_zss_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldsdu_be_zss_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_lddd_le_zss_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_lddd_be_zss_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldbds_zss_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldhds_le_zss_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldhds_be_zss_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldsds_le_zss_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldsds_be_zss_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) + +DEF_HELPER_FLAGS_6(sve_ldbdu_zd_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldhdu_le_zd_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldhdu_be_zd_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldsdu_le_zd_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldsdu_be_zd_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_lddd_le_zd_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_lddd_be_zd_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldbds_zd_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldhds_le_zd_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldhds_be_zd_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldsds_le_zd_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldsds_be_zd_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) + DEF_HELPER_FLAGS_6(sve_ldffbsu_zsu, TCG_CALL_NO_WG, void, env, ptr, ptr, ptr, tl, i32) DEF_HELPER_FLAGS_6(sve_ldffhsu_le_zsu, TCG_CALL_NO_WG, @@ -XXX,XX +XXX,XX @@ DEF_HELPER_FLAGS_6(sve_ldffsds_le_zd, TCG_CALL_NO_WG, DEF_HELPER_FLAGS_6(sve_ldffsds_be_zd, TCG_CALL_NO_WG, void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldffbsu_zsu_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldffhsu_le_zsu_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldffhsu_be_zsu_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldffss_le_zsu_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldffss_be_zsu_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldffbss_zsu_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldffhss_le_zsu_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldffhss_be_zsu_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) + +DEF_HELPER_FLAGS_6(sve_ldffbsu_zss_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldffhsu_le_zss_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldffhsu_be_zss_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldffss_le_zss_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldffss_be_zss_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldffbss_zss_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldffhss_le_zss_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldffhss_be_zss_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) + +DEF_HELPER_FLAGS_6(sve_ldffbdu_zsu_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldffhdu_le_zsu_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldffhdu_be_zsu_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldffsdu_le_zsu_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldffsdu_be_zsu_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldffdd_le_zsu_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldffdd_be_zsu_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldffbds_zsu_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldffhds_le_zsu_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldffhds_be_zsu_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldffsds_le_zsu_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldffsds_be_zsu_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) + +DEF_HELPER_FLAGS_6(sve_ldffbdu_zss_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldffhdu_le_zss_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldffhdu_be_zss_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldffsdu_le_zss_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldffsdu_be_zss_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldffdd_le_zss_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldffdd_be_zss_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldffbds_zss_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldffhds_le_zss_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldffhds_be_zss_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldffsds_le_zss_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldffsds_be_zss_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) + +DEF_HELPER_FLAGS_6(sve_ldffbdu_zd_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldffhdu_le_zd_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldffhdu_be_zd_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldffsdu_le_zd_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldffsdu_be_zd_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldffdd_le_zd_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldffdd_be_zd_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldffbds_zd_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldffhds_le_zd_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldffhds_be_zd_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldffsds_le_zd_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldffsds_be_zd_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) + DEF_HELPER_FLAGS_6(sve_stbs_zsu, TCG_CALL_NO_WG, void, env, ptr, ptr, ptr, tl, i32) DEF_HELPER_FLAGS_6(sve_sths_le_zsu, TCG_CALL_NO_WG, @@ -XXX,XX +XXX,XX @@ DEF_HELPER_FLAGS_6(sve_stdd_le_zd, TCG_CALL_NO_WG, DEF_HELPER_FLAGS_6(sve_stdd_be_zd, TCG_CALL_NO_WG, void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_stbs_zsu_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_sths_le_zsu_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_sths_be_zsu_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_stss_le_zsu_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_stss_be_zsu_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) + +DEF_HELPER_FLAGS_6(sve_stbs_zss_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_sths_le_zss_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_sths_be_zss_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_stss_le_zss_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_stss_be_zss_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) + +DEF_HELPER_FLAGS_6(sve_stbd_zsu_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_sthd_le_zsu_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_sthd_be_zsu_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_stsd_le_zsu_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_stsd_be_zsu_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_stdd_le_zsu_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_stdd_be_zsu_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) + +DEF_HELPER_FLAGS_6(sve_stbd_zss_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_sthd_le_zss_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_sthd_be_zss_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_stsd_le_zss_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_stsd_be_zss_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_stdd_le_zss_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_stdd_be_zss_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) + +DEF_HELPER_FLAGS_6(sve_stbd_zd_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_sthd_le_zd_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_sthd_be_zd_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_stsd_le_zd_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_stsd_be_zd_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_stdd_le_zd_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_stdd_be_zd_mte, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) + DEF_HELPER_FLAGS_4(sve2_pmull_h, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, i32) diff --git a/target/arm/sve_helper.c b/target/arm/sve_helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/sve_helper.c +++ b/target/arm/sve_helper.c @@ -XXX,XX +XXX,XX @@ static target_ulong off_zd_d(void *reg, intptr_t reg_ofs) static inline QEMU_ALWAYS_INLINE void sve_ld1_z(CPUARMState *env, void *vd, uint64_t *vg, void *vm, target_ulong base, uint32_t desc, uintptr_t retaddr, - int esize, int msize, zreg_off_fn *off_fn, + uint32_t mtedesc, int esize, int msize, + zreg_off_fn *off_fn, sve_ldst1_host_fn *host_fn, sve_ldst1_tlb_fn *tlb_fn) { @@ -XXX,XX +XXX,XX @@ void sve_ld1_z(CPUARMState *env, void *vd, uint64_t *vg, void *vm, cpu_check_watchpoint(env_cpu(env), addr, msize, info.attrs, BP_MEM_READ, retaddr); } - /* TODO: MTE check */ + if (mtedesc && arm_tlb_mte_tagged(&info.attrs)) { + mte_check1(env, mtedesc, addr, retaddr); + } host_fn(&scratch, reg_off, info.host); } else { /* Element crosses the page boundary. */ @@ -XXX,XX +XXX,XX @@ void sve_ld1_z(CPUARMState *env, void *vd, uint64_t *vg, void *vm, msize, info.attrs, BP_MEM_READ, retaddr); } - /* TODO: MTE check */ + if (mtedesc && arm_tlb_mte_tagged(&info.attrs)) { + mte_check1(env, mtedesc, addr, retaddr); + } tlb_fn(env, &scratch, reg_off, addr, retaddr); } } @@ -XXX,XX +XXX,XX @@ void sve_ld1_z(CPUARMState *env, void *vd, uint64_t *vg, void *vm, memcpy(vd, &scratch, reg_max); } +static inline QEMU_ALWAYS_INLINE +void sve_ld1_z_mte(CPUARMState *env, void *vd, uint64_t *vg, void *vm, + target_ulong base, uint32_t desc, uintptr_t retaddr, + int esize, int msize, zreg_off_fn *off_fn, + sve_ldst1_host_fn *host_fn, + sve_ldst1_tlb_fn *tlb_fn) +{ + uint32_t mtedesc = desc >> (SIMD_DATA_SHIFT + SVE_MTEDESC_SHIFT); + /* Remove mtedesc from the normal sve descriptor. */ + desc = extract32(desc, 0, SIMD_DATA_SHIFT + SVE_MTEDESC_SHIFT); + + /* + * ??? TODO: For the 32-bit offset extractions, base + ofs cannot + * offset base entirely over the address space hole to change the + * pointer tag, or change the bit55 selector. So we could here + * examine TBI + TCMA like we do for sve_ldN_r_mte(). + */ + sve_ld1_z(env, vd, vg, vm, base, desc, retaddr, mtedesc, + esize, msize, off_fn, host_fn, tlb_fn); +} + #define DO_LD1_ZPZ_S(MEM, OFS, MSZ) \ void HELPER(sve_ld##MEM##_##OFS)(CPUARMState *env, void *vd, void *vg, \ void *vm, target_ulong base, uint32_t desc) \ { \ - sve_ld1_z(env, vd, vg, vm, base, desc, GETPC(), 4, 1 << MSZ, \ + sve_ld1_z(env, vd, vg, vm, base, desc, GETPC(), 0, 4, 1 << MSZ, \ off_##OFS##_s, sve_ld1##MEM##_host, sve_ld1##MEM##_tlb); \ +} \ +void HELPER(sve_ld##MEM##_##OFS##_mte)(CPUARMState *env, void *vd, void *vg, \ + void *vm, target_ulong base, uint32_t desc) \ +{ \ + sve_ld1_z_mte(env, vd, vg, vm, base, desc, GETPC(), 4, 1 << MSZ, \ + off_##OFS##_s, sve_ld1##MEM##_host, sve_ld1##MEM##_tlb); \ } #define DO_LD1_ZPZ_D(MEM, OFS, MSZ) \ void HELPER(sve_ld##MEM##_##OFS)(CPUARMState *env, void *vd, void *vg, \ void *vm, target_ulong base, uint32_t desc) \ { \ - sve_ld1_z(env, vd, vg, vm, base, desc, GETPC(), 8, 1 << MSZ, \ + sve_ld1_z(env, vd, vg, vm, base, desc, GETPC(), 0, 8, 1 << MSZ, \ off_##OFS##_d, sve_ld1##MEM##_host, sve_ld1##MEM##_tlb); \ +} \ +void HELPER(sve_ld##MEM##_##OFS##_mte)(CPUARMState *env, void *vd, void *vg, \ + void *vm, target_ulong base, uint32_t desc) \ +{ \ + sve_ld1_z_mte(env, vd, vg, vm, base, desc, GETPC(), 8, 1 << MSZ, \ + off_##OFS##_d, sve_ld1##MEM##_host, sve_ld1##MEM##_tlb); \ } DO_LD1_ZPZ_S(bsu, zsu, MO_8) @@ -XXX,XX +XXX,XX @@ DO_LD1_ZPZ_D(dd_be, zd, MO_64) static inline QEMU_ALWAYS_INLINE void sve_ldff1_z(CPUARMState *env, void *vd, uint64_t *vg, void *vm, target_ulong base, uint32_t desc, uintptr_t retaddr, - const int esz, const int msz, zreg_off_fn *off_fn, + uint32_t mtedesc, const int esz, const int msz, + zreg_off_fn *off_fn, sve_ldst1_host_fn *host_fn, sve_ldst1_tlb_fn *tlb_fn) { @@ -XXX,XX +XXX,XX @@ void sve_ldff1_z(CPUARMState *env, void *vd, uint64_t *vg, void *vm, * Probe the first element, allowing faults. */ addr = base + (off_fn(vm, reg_off) << scale); + if (mtedesc) { + mte_check1(env, mtedesc, addr, retaddr); + } tlb_fn(env, vd, reg_off, addr, retaddr); /* After any fault, zero the other elements. */ @@ -XXX,XX +XXX,XX @@ void sve_ldff1_z(CPUARMState *env, void *vd, uint64_t *vg, void *vm, (env_cpu(env), addr, msize) & BP_MEM_READ)) { goto fault; } - /* TODO: MTE check. */ + if (mtedesc && + arm_tlb_mte_tagged(&info.attrs) && + !mte_probe1(env, mtedesc, addr)) { + goto fault; + } host_fn(vd, reg_off, info.host); } @@ -XXX,XX +XXX,XX @@ void sve_ldff1_z(CPUARMState *env, void *vd, uint64_t *vg, void *vm, record_fault(env, reg_off, reg_max); } -#define DO_LDFF1_ZPZ_S(MEM, OFS, MSZ) \ -void HELPER(sve_ldff##MEM##_##OFS)(CPUARMState *env, void *vd, void *vg, \ - void *vm, target_ulong base, uint32_t desc) \ -{ \ - sve_ldff1_z(env, vd, vg, vm, base, desc, GETPC(), MO_32, MSZ, \ - off_##OFS##_s, sve_ld1##MEM##_host, sve_ld1##MEM##_tlb); \ +static inline QEMU_ALWAYS_INLINE +void sve_ldff1_z_mte(CPUARMState *env, void *vd, uint64_t *vg, void *vm, + target_ulong base, uint32_t desc, uintptr_t retaddr, + const int esz, const int msz, + zreg_off_fn *off_fn, + sve_ldst1_host_fn *host_fn, + sve_ldst1_tlb_fn *tlb_fn) +{ + uint32_t mtedesc = desc >> (SIMD_DATA_SHIFT + SVE_MTEDESC_SHIFT); + /* Remove mtedesc from the normal sve descriptor. */ + desc = extract32(desc, 0, SIMD_DATA_SHIFT + SVE_MTEDESC_SHIFT); + + /* + * ??? TODO: For the 32-bit offset extractions, base + ofs cannot + * offset base entirely over the address space hole to change the + * pointer tag, or change the bit55 selector. So we could here + * examine TBI + TCMA like we do for sve_ldN_r_mte(). + */ + sve_ldff1_z(env, vd, vg, vm, base, desc, retaddr, mtedesc, + esz, msz, off_fn, host_fn, tlb_fn); } -#define DO_LDFF1_ZPZ_D(MEM, OFS, MSZ) \ -void HELPER(sve_ldff##MEM##_##OFS)(CPUARMState *env, void *vd, void *vg, \ - void *vm, target_ulong base, uint32_t desc) \ -{ \ - sve_ldff1_z(env, vd, vg, vm, base, desc, GETPC(), MO_64, MSZ, \ - off_##OFS##_d, sve_ld1##MEM##_host, sve_ld1##MEM##_tlb); \ +#define DO_LDFF1_ZPZ_S(MEM, OFS, MSZ) \ +void HELPER(sve_ldff##MEM##_##OFS) \ + (CPUARMState *env, void *vd, void *vg, \ + void *vm, target_ulong base, uint32_t desc) \ +{ \ + sve_ldff1_z(env, vd, vg, vm, base, desc, GETPC(), 0, MO_32, MSZ, \ + off_##OFS##_s, sve_ld1##MEM##_host, sve_ld1##MEM##_tlb); \ +} \ +void HELPER(sve_ldff##MEM##_##OFS##_mte) \ + (CPUARMState *env, void *vd, void *vg, \ + void *vm, target_ulong base, uint32_t desc) \ +{ \ + sve_ldff1_z_mte(env, vd, vg, vm, base, desc, GETPC(), MO_32, MSZ, \ + off_##OFS##_s, sve_ld1##MEM##_host, sve_ld1##MEM##_tlb); \ +} + +#define DO_LDFF1_ZPZ_D(MEM, OFS, MSZ) \ +void HELPER(sve_ldff##MEM##_##OFS) \ + (CPUARMState *env, void *vd, void *vg, \ + void *vm, target_ulong base, uint32_t desc) \ +{ \ + sve_ldff1_z(env, vd, vg, vm, base, desc, GETPC(), 0, MO_64, MSZ, \ + off_##OFS##_d, sve_ld1##MEM##_host, sve_ld1##MEM##_tlb); \ +} \ +void HELPER(sve_ldff##MEM##_##OFS##_mte) \ + (CPUARMState *env, void *vd, void *vg, \ + void *vm, target_ulong base, uint32_t desc) \ +{ \ + sve_ldff1_z_mte(env, vd, vg, vm, base, desc, GETPC(), MO_64, MSZ, \ + off_##OFS##_d, sve_ld1##MEM##_host, sve_ld1##MEM##_tlb); \ } DO_LDFF1_ZPZ_S(bsu, zsu, MO_8) @@ -XXX,XX +XXX,XX @@ DO_LDFF1_ZPZ_D(dd_be, zd, MO_64) static inline QEMU_ALWAYS_INLINE void sve_st1_z(CPUARMState *env, void *vd, uint64_t *vg, void *vm, target_ulong base, uint32_t desc, uintptr_t retaddr, - int esize, int msize, zreg_off_fn *off_fn, + uint32_t mtedesc, int esize, int msize, + zreg_off_fn *off_fn, sve_ldst1_host_fn *host_fn, sve_ldst1_tlb_fn *tlb_fn) { @@ -XXX,XX +XXX,XX @@ void sve_st1_z(CPUARMState *env, void *vd, uint64_t *vg, void *vm, cpu_check_watchpoint(env_cpu(env), addr, msize, info.attrs, BP_MEM_WRITE, retaddr); } - /* TODO: MTE check. */ + + if (mtedesc && arm_tlb_mte_tagged(&info.attrs)) { + mte_check1(env, mtedesc, addr, retaddr); + } } i += 1; reg_off += esize; @@ -XXX,XX +XXX,XX @@ void sve_st1_z(CPUARMState *env, void *vd, uint64_t *vg, void *vm, } while (reg_off < reg_max); } -#define DO_ST1_ZPZ_S(MEM, OFS, MSZ) \ -void HELPER(sve_st##MEM##_##OFS)(CPUARMState *env, void *vd, void *vg, \ - void *vm, target_ulong base, uint32_t desc) \ -{ \ - sve_st1_z(env, vd, vg, vm, base, desc, GETPC(), 4, 1 << MSZ, \ - off_##OFS##_s, sve_st1##MEM##_host, sve_st1##MEM##_tlb); \ +static inline QEMU_ALWAYS_INLINE +void sve_st1_z_mte(CPUARMState *env, void *vd, uint64_t *vg, void *vm, + target_ulong base, uint32_t desc, uintptr_t retaddr, + int esize, int msize, zreg_off_fn *off_fn, + sve_ldst1_host_fn *host_fn, + sve_ldst1_tlb_fn *tlb_fn) +{ + uint32_t mtedesc = desc >> (SIMD_DATA_SHIFT + SVE_MTEDESC_SHIFT); + /* Remove mtedesc from the normal sve descriptor. */ + desc = extract32(desc, 0, SIMD_DATA_SHIFT + SVE_MTEDESC_SHIFT); + + /* + * ??? TODO: For the 32-bit offset extractions, base + ofs cannot + * offset base entirely over the address space hole to change the + * pointer tag, or change the bit55 selector. So we could here + * examine TBI + TCMA like we do for sve_ldN_r_mte(). + */ + sve_st1_z(env, vd, vg, vm, base, desc, retaddr, mtedesc, + esize, msize, off_fn, host_fn, tlb_fn); } -#define DO_ST1_ZPZ_D(MEM, OFS, MSZ) \ -void HELPER(sve_st##MEM##_##OFS)(CPUARMState *env, void *vd, void *vg, \ +#define DO_ST1_ZPZ_S(MEM, OFS, MSZ) \ +void HELPER(sve_st##MEM##_##OFS)(CPUARMState *env, void *vd, void *vg, \ void *vm, target_ulong base, uint32_t desc) \ -{ \ - sve_st1_z(env, vd, vg, vm, base, desc, GETPC(), 8, 1 << MSZ, \ - off_##OFS##_d, sve_st1##MEM##_host, sve_st1##MEM##_tlb); \ +{ \ + sve_st1_z(env, vd, vg, vm, base, desc, GETPC(), 0, 4, 1 << MSZ, \ + off_##OFS##_s, sve_st1##MEM##_host, sve_st1##MEM##_tlb); \ +} \ +void HELPER(sve_st##MEM##_##OFS##_mte)(CPUARMState *env, void *vd, void *vg, \ + void *vm, target_ulong base, uint32_t desc) \ +{ \ + sve_st1_z_mte(env, vd, vg, vm, base, desc, GETPC(), 4, 1 << MSZ, \ + off_##OFS##_s, sve_st1##MEM##_host, sve_st1##MEM##_tlb); \ +} + +#define DO_ST1_ZPZ_D(MEM, OFS, MSZ) \ +void HELPER(sve_st##MEM##_##OFS)(CPUARMState *env, void *vd, void *vg, \ + void *vm, target_ulong base, uint32_t desc) \ +{ \ + sve_st1_z(env, vd, vg, vm, base, desc, GETPC(), 0, 8, 1 << MSZ, \ + off_##OFS##_d, sve_st1##MEM##_host, sve_st1##MEM##_tlb); \ +} \ +void HELPER(sve_st##MEM##_##OFS##_mte)(CPUARMState *env, void *vd, void *vg, \ + void *vm, target_ulong base, uint32_t desc) \ +{ \ + sve_st1_z_mte(env, vd, vg, vm, base, desc, GETPC(), 8, 1 << MSZ, \ + off_##OFS##_d, sve_st1##MEM##_host, sve_st1##MEM##_tlb); \ } DO_ST1_ZPZ_S(bs, zsu, MO_8) diff --git a/target/arm/translate-sve.c b/target/arm/translate-sve.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-sve.c +++ b/target/arm/translate-sve.c @@ -XXX,XX +XXX,XX @@ static bool trans_ST_zpri(DisasContext *s, arg_rpri_store *a) */ static void do_mem_zpz(DisasContext *s, int zt, int pg, int zm, - int scale, TCGv_i64 scalar, int msz, + int scale, TCGv_i64 scalar, int msz, bool is_write, gen_helper_gvec_mem_scatter *fn) { unsigned vsz = vec_full_reg_size(s); @@ -XXX,XX +XXX,XX @@ static void do_mem_zpz(DisasContext *s, int zt, int pg, int zm, TCGv_ptr t_pg = tcg_temp_new_ptr(); TCGv_ptr t_zt = tcg_temp_new_ptr(); TCGv_i32 t_desc; - int desc; + int desc = 0; + if (s->mte_active[0]) { + desc = FIELD_DP32(desc, MTEDESC, MIDX, get_mem_index(s)); + desc = FIELD_DP32(desc, MTEDESC, TBI, s->tbid); + desc = FIELD_DP32(desc, MTEDESC, TCMA, s->tcma); + desc = FIELD_DP32(desc, MTEDESC, WRITE, is_write); + desc = FIELD_DP32(desc, MTEDESC, ESIZE, 1 << msz); + desc <<= SVE_MTEDESC_SHIFT; + } desc = simd_desc(vsz, vsz, scale); t_desc = tcg_const_i32(desc); @@ -XXX,XX +XXX,XX @@ static void do_mem_zpz(DisasContext *s, int zt, int pg, int zm, tcg_temp_free_i32(t_desc); } -/* Indexed by [be][ff][xs][u][msz]. */ -static gen_helper_gvec_mem_scatter * const gather_load_fn32[2][2][2][2][3] = { - /* Little-endian */ - { { { { gen_helper_sve_ldbss_zsu, - gen_helper_sve_ldhss_le_zsu, - NULL, }, - { gen_helper_sve_ldbsu_zsu, - gen_helper_sve_ldhsu_le_zsu, - gen_helper_sve_ldss_le_zsu, } }, - { { gen_helper_sve_ldbss_zss, - gen_helper_sve_ldhss_le_zss, - NULL, }, - { gen_helper_sve_ldbsu_zss, - gen_helper_sve_ldhsu_le_zss, - gen_helper_sve_ldss_le_zss, } } }, +/* Indexed by [mte][be][ff][xs][u][msz]. */ +static gen_helper_gvec_mem_scatter * const +gather_load_fn32[2][2][2][2][2][3] = { + { /* MTE Inactive */ + { /* Little-endian */ + { { { gen_helper_sve_ldbss_zsu, + gen_helper_sve_ldhss_le_zsu, + NULL, }, + { gen_helper_sve_ldbsu_zsu, + gen_helper_sve_ldhsu_le_zsu, + gen_helper_sve_ldss_le_zsu, } }, + { { gen_helper_sve_ldbss_zss, + gen_helper_sve_ldhss_le_zss, + NULL, }, + { gen_helper_sve_ldbsu_zss, + gen_helper_sve_ldhsu_le_zss, + gen_helper_sve_ldss_le_zss, } } }, - /* First-fault */ - { { { gen_helper_sve_ldffbss_zsu, - gen_helper_sve_ldffhss_le_zsu, - NULL, }, - { gen_helper_sve_ldffbsu_zsu, - gen_helper_sve_ldffhsu_le_zsu, - gen_helper_sve_ldffss_le_zsu, } }, - { { gen_helper_sve_ldffbss_zss, - gen_helper_sve_ldffhss_le_zss, - NULL, }, - { gen_helper_sve_ldffbsu_zss, - gen_helper_sve_ldffhsu_le_zss, - gen_helper_sve_ldffss_le_zss, } } } }, + /* First-fault */ + { { { gen_helper_sve_ldffbss_zsu, + gen_helper_sve_ldffhss_le_zsu, + NULL, }, + { gen_helper_sve_ldffbsu_zsu, + gen_helper_sve_ldffhsu_le_zsu, + gen_helper_sve_ldffss_le_zsu, } }, + { { gen_helper_sve_ldffbss_zss, + gen_helper_sve_ldffhss_le_zss, + NULL, }, + { gen_helper_sve_ldffbsu_zss, + gen_helper_sve_ldffhsu_le_zss, + gen_helper_sve_ldffss_le_zss, } } } }, - /* Big-endian */ - { { { { gen_helper_sve_ldbss_zsu, - gen_helper_sve_ldhss_be_zsu, - NULL, }, - { gen_helper_sve_ldbsu_zsu, - gen_helper_sve_ldhsu_be_zsu, - gen_helper_sve_ldss_be_zsu, } }, - { { gen_helper_sve_ldbss_zss, - gen_helper_sve_ldhss_be_zss, - NULL, }, - { gen_helper_sve_ldbsu_zss, - gen_helper_sve_ldhsu_be_zss, - gen_helper_sve_ldss_be_zss, } } }, + { /* Big-endian */ + { { { gen_helper_sve_ldbss_zsu, + gen_helper_sve_ldhss_be_zsu, + NULL, }, + { gen_helper_sve_ldbsu_zsu, + gen_helper_sve_ldhsu_be_zsu, + gen_helper_sve_ldss_be_zsu, } }, + { { gen_helper_sve_ldbss_zss, + gen_helper_sve_ldhss_be_zss, + NULL, }, + { gen_helper_sve_ldbsu_zss, + gen_helper_sve_ldhsu_be_zss, + gen_helper_sve_ldss_be_zss, } } }, - /* First-fault */ - { { { gen_helper_sve_ldffbss_zsu, - gen_helper_sve_ldffhss_be_zsu, - NULL, }, - { gen_helper_sve_ldffbsu_zsu, - gen_helper_sve_ldffhsu_be_zsu, - gen_helper_sve_ldffss_be_zsu, } }, - { { gen_helper_sve_ldffbss_zss, - gen_helper_sve_ldffhss_be_zss, - NULL, }, - { gen_helper_sve_ldffbsu_zss, - gen_helper_sve_ldffhsu_be_zss, - gen_helper_sve_ldffss_be_zss, } } } }, + /* First-fault */ + { { { gen_helper_sve_ldffbss_zsu, + gen_helper_sve_ldffhss_be_zsu, + NULL, }, + { gen_helper_sve_ldffbsu_zsu, + gen_helper_sve_ldffhsu_be_zsu, + gen_helper_sve_ldffss_be_zsu, } }, + { { gen_helper_sve_ldffbss_zss, + gen_helper_sve_ldffhss_be_zss, + NULL, }, + { gen_helper_sve_ldffbsu_zss, + gen_helper_sve_ldffhsu_be_zss, + gen_helper_sve_ldffss_be_zss, } } } } }, + { /* MTE Active */ + { /* Little-endian */ + { { { gen_helper_sve_ldbss_zsu_mte, + gen_helper_sve_ldhss_le_zsu_mte, + NULL, }, + { gen_helper_sve_ldbsu_zsu_mte, + gen_helper_sve_ldhsu_le_zsu_mte, + gen_helper_sve_ldss_le_zsu_mte, } }, + { { gen_helper_sve_ldbss_zss_mte, + gen_helper_sve_ldhss_le_zss_mte, + NULL, }, + { gen_helper_sve_ldbsu_zss_mte, + gen_helper_sve_ldhsu_le_zss_mte, + gen_helper_sve_ldss_le_zss_mte, } } }, + + /* First-fault */ + { { { gen_helper_sve_ldffbss_zsu_mte, + gen_helper_sve_ldffhss_le_zsu_mte, + NULL, }, + { gen_helper_sve_ldffbsu_zsu_mte, + gen_helper_sve_ldffhsu_le_zsu_mte, + gen_helper_sve_ldffss_le_zsu_mte, } }, + { { gen_helper_sve_ldffbss_zss_mte, + gen_helper_sve_ldffhss_le_zss_mte, + NULL, }, + { gen_helper_sve_ldffbsu_zss_mte, + gen_helper_sve_ldffhsu_le_zss_mte, + gen_helper_sve_ldffss_le_zss_mte, } } } }, + + { /* Big-endian */ + { { { gen_helper_sve_ldbss_zsu_mte, + gen_helper_sve_ldhss_be_zsu_mte, + NULL, }, + { gen_helper_sve_ldbsu_zsu_mte, + gen_helper_sve_ldhsu_be_zsu_mte, + gen_helper_sve_ldss_be_zsu_mte, } }, + { { gen_helper_sve_ldbss_zss_mte, + gen_helper_sve_ldhss_be_zss_mte, + NULL, }, + { gen_helper_sve_ldbsu_zss_mte, + gen_helper_sve_ldhsu_be_zss_mte, + gen_helper_sve_ldss_be_zss_mte, } } }, + + /* First-fault */ + { { { gen_helper_sve_ldffbss_zsu_mte, + gen_helper_sve_ldffhss_be_zsu_mte, + NULL, }, + { gen_helper_sve_ldffbsu_zsu_mte, + gen_helper_sve_ldffhsu_be_zsu_mte, + gen_helper_sve_ldffss_be_zsu_mte, } }, + { { gen_helper_sve_ldffbss_zss_mte, + gen_helper_sve_ldffhss_be_zss_mte, + NULL, }, + { gen_helper_sve_ldffbsu_zss_mte, + gen_helper_sve_ldffhsu_be_zss_mte, + gen_helper_sve_ldffss_be_zss_mte, } } } } }, }; /* Note that we overload xs=2 to indicate 64-bit offset. */ -static gen_helper_gvec_mem_scatter * const gather_load_fn64[2][2][3][2][4] = { - /* Little-endian */ - { { { { gen_helper_sve_ldbds_zsu, - gen_helper_sve_ldhds_le_zsu, - gen_helper_sve_ldsds_le_zsu, - NULL, }, - { gen_helper_sve_ldbdu_zsu, - gen_helper_sve_ldhdu_le_zsu, - gen_helper_sve_ldsdu_le_zsu, - gen_helper_sve_lddd_le_zsu, } }, - { { gen_helper_sve_ldbds_zss, - gen_helper_sve_ldhds_le_zss, - gen_helper_sve_ldsds_le_zss, - NULL, }, - { gen_helper_sve_ldbdu_zss, - gen_helper_sve_ldhdu_le_zss, - gen_helper_sve_ldsdu_le_zss, - gen_helper_sve_lddd_le_zss, } }, - { { gen_helper_sve_ldbds_zd, - gen_helper_sve_ldhds_le_zd, - gen_helper_sve_ldsds_le_zd, - NULL, }, - { gen_helper_sve_ldbdu_zd, - gen_helper_sve_ldhdu_le_zd, - gen_helper_sve_ldsdu_le_zd, - gen_helper_sve_lddd_le_zd, } } }, +static gen_helper_gvec_mem_scatter * const +gather_load_fn64[2][2][2][3][2][4] = { + { /* MTE Inactive */ + { /* Little-endian */ + { { { gen_helper_sve_ldbds_zsu, + gen_helper_sve_ldhds_le_zsu, + gen_helper_sve_ldsds_le_zsu, + NULL, }, + { gen_helper_sve_ldbdu_zsu, + gen_helper_sve_ldhdu_le_zsu, + gen_helper_sve_ldsdu_le_zsu, + gen_helper_sve_lddd_le_zsu, } }, + { { gen_helper_sve_ldbds_zss, + gen_helper_sve_ldhds_le_zss, + gen_helper_sve_ldsds_le_zss, + NULL, }, + { gen_helper_sve_ldbdu_zss, + gen_helper_sve_ldhdu_le_zss, + gen_helper_sve_ldsdu_le_zss, + gen_helper_sve_lddd_le_zss, } }, + { { gen_helper_sve_ldbds_zd, + gen_helper_sve_ldhds_le_zd, + gen_helper_sve_ldsds_le_zd, + NULL, }, + { gen_helper_sve_ldbdu_zd, + gen_helper_sve_ldhdu_le_zd, + gen_helper_sve_ldsdu_le_zd, + gen_helper_sve_lddd_le_zd, } } }, - /* First-fault */ - { { { gen_helper_sve_ldffbds_zsu, - gen_helper_sve_ldffhds_le_zsu, - gen_helper_sve_ldffsds_le_zsu, - NULL, }, - { gen_helper_sve_ldffbdu_zsu, - gen_helper_sve_ldffhdu_le_zsu, - gen_helper_sve_ldffsdu_le_zsu, - gen_helper_sve_ldffdd_le_zsu, } }, - { { gen_helper_sve_ldffbds_zss, - gen_helper_sve_ldffhds_le_zss, - gen_helper_sve_ldffsds_le_zss, - NULL, }, - { gen_helper_sve_ldffbdu_zss, - gen_helper_sve_ldffhdu_le_zss, - gen_helper_sve_ldffsdu_le_zss, - gen_helper_sve_ldffdd_le_zss, } }, - { { gen_helper_sve_ldffbds_zd, - gen_helper_sve_ldffhds_le_zd, - gen_helper_sve_ldffsds_le_zd, - NULL, }, - { gen_helper_sve_ldffbdu_zd, - gen_helper_sve_ldffhdu_le_zd, - gen_helper_sve_ldffsdu_le_zd, - gen_helper_sve_ldffdd_le_zd, } } } }, + /* First-fault */ + { { { gen_helper_sve_ldffbds_zsu, + gen_helper_sve_ldffhds_le_zsu, + gen_helper_sve_ldffsds_le_zsu, + NULL, }, + { gen_helper_sve_ldffbdu_zsu, + gen_helper_sve_ldffhdu_le_zsu, + gen_helper_sve_ldffsdu_le_zsu, + gen_helper_sve_ldffdd_le_zsu, } }, + { { gen_helper_sve_ldffbds_zss, + gen_helper_sve_ldffhds_le_zss, + gen_helper_sve_ldffsds_le_zss, + NULL, }, + { gen_helper_sve_ldffbdu_zss, + gen_helper_sve_ldffhdu_le_zss, + gen_helper_sve_ldffsdu_le_zss, + gen_helper_sve_ldffdd_le_zss, } }, + { { gen_helper_sve_ldffbds_zd, + gen_helper_sve_ldffhds_le_zd, + gen_helper_sve_ldffsds_le_zd, + NULL, }, + { gen_helper_sve_ldffbdu_zd, + gen_helper_sve_ldffhdu_le_zd, + gen_helper_sve_ldffsdu_le_zd, + gen_helper_sve_ldffdd_le_zd, } } } }, + { /* Big-endian */ + { { { gen_helper_sve_ldbds_zsu, + gen_helper_sve_ldhds_be_zsu, + gen_helper_sve_ldsds_be_zsu, + NULL, }, + { gen_helper_sve_ldbdu_zsu, + gen_helper_sve_ldhdu_be_zsu, + gen_helper_sve_ldsdu_be_zsu, + gen_helper_sve_lddd_be_zsu, } }, + { { gen_helper_sve_ldbds_zss, + gen_helper_sve_ldhds_be_zss, + gen_helper_sve_ldsds_be_zss, + NULL, }, + { gen_helper_sve_ldbdu_zss, + gen_helper_sve_ldhdu_be_zss, + gen_helper_sve_ldsdu_be_zss, + gen_helper_sve_lddd_be_zss, } }, + { { gen_helper_sve_ldbds_zd, + gen_helper_sve_ldhds_be_zd, + gen_helper_sve_ldsds_be_zd, + NULL, }, + { gen_helper_sve_ldbdu_zd, + gen_helper_sve_ldhdu_be_zd, + gen_helper_sve_ldsdu_be_zd, + gen_helper_sve_lddd_be_zd, } } }, - /* Big-endian */ - { { { { gen_helper_sve_ldbds_zsu, - gen_helper_sve_ldhds_be_zsu, - gen_helper_sve_ldsds_be_zsu, - NULL, }, - { gen_helper_sve_ldbdu_zsu, - gen_helper_sve_ldhdu_be_zsu, - gen_helper_sve_ldsdu_be_zsu, - gen_helper_sve_lddd_be_zsu, } }, - { { gen_helper_sve_ldbds_zss, - gen_helper_sve_ldhds_be_zss, - gen_helper_sve_ldsds_be_zss, - NULL, }, - { gen_helper_sve_ldbdu_zss, - gen_helper_sve_ldhdu_be_zss, - gen_helper_sve_ldsdu_be_zss, - gen_helper_sve_lddd_be_zss, } }, - { { gen_helper_sve_ldbds_zd, - gen_helper_sve_ldhds_be_zd, - gen_helper_sve_ldsds_be_zd, - NULL, }, - { gen_helper_sve_ldbdu_zd, - gen_helper_sve_ldhdu_be_zd, - gen_helper_sve_ldsdu_be_zd, - gen_helper_sve_lddd_be_zd, } } }, + /* First-fault */ + { { { gen_helper_sve_ldffbds_zsu, + gen_helper_sve_ldffhds_be_zsu, + gen_helper_sve_ldffsds_be_zsu, + NULL, }, + { gen_helper_sve_ldffbdu_zsu, + gen_helper_sve_ldffhdu_be_zsu, + gen_helper_sve_ldffsdu_be_zsu, + gen_helper_sve_ldffdd_be_zsu, } }, + { { gen_helper_sve_ldffbds_zss, + gen_helper_sve_ldffhds_be_zss, + gen_helper_sve_ldffsds_be_zss, + NULL, }, + { gen_helper_sve_ldffbdu_zss, + gen_helper_sve_ldffhdu_be_zss, + gen_helper_sve_ldffsdu_be_zss, + gen_helper_sve_ldffdd_be_zss, } }, + { { gen_helper_sve_ldffbds_zd, + gen_helper_sve_ldffhds_be_zd, + gen_helper_sve_ldffsds_be_zd, + NULL, }, + { gen_helper_sve_ldffbdu_zd, + gen_helper_sve_ldffhdu_be_zd, + gen_helper_sve_ldffsdu_be_zd, + gen_helper_sve_ldffdd_be_zd, } } } } }, + { /* MTE Active */ + { /* Little-endian */ + { { { gen_helper_sve_ldbds_zsu_mte, + gen_helper_sve_ldhds_le_zsu_mte, + gen_helper_sve_ldsds_le_zsu_mte, + NULL, }, + { gen_helper_sve_ldbdu_zsu_mte, + gen_helper_sve_ldhdu_le_zsu_mte, + gen_helper_sve_ldsdu_le_zsu_mte, + gen_helper_sve_lddd_le_zsu_mte, } }, + { { gen_helper_sve_ldbds_zss_mte, + gen_helper_sve_ldhds_le_zss_mte, + gen_helper_sve_ldsds_le_zss_mte, + NULL, }, + { gen_helper_sve_ldbdu_zss_mte, + gen_helper_sve_ldhdu_le_zss_mte, + gen_helper_sve_ldsdu_le_zss_mte, + gen_helper_sve_lddd_le_zss_mte, } }, + { { gen_helper_sve_ldbds_zd_mte, + gen_helper_sve_ldhds_le_zd_mte, + gen_helper_sve_ldsds_le_zd_mte, + NULL, }, + { gen_helper_sve_ldbdu_zd_mte, + gen_helper_sve_ldhdu_le_zd_mte, + gen_helper_sve_ldsdu_le_zd_mte, + gen_helper_sve_lddd_le_zd_mte, } } }, - /* First-fault */ - { { { gen_helper_sve_ldffbds_zsu, - gen_helper_sve_ldffhds_be_zsu, - gen_helper_sve_ldffsds_be_zsu, - NULL, }, - { gen_helper_sve_ldffbdu_zsu, - gen_helper_sve_ldffhdu_be_zsu, - gen_helper_sve_ldffsdu_be_zsu, - gen_helper_sve_ldffdd_be_zsu, } }, - { { gen_helper_sve_ldffbds_zss, - gen_helper_sve_ldffhds_be_zss, - gen_helper_sve_ldffsds_be_zss, - NULL, }, - { gen_helper_sve_ldffbdu_zss, - gen_helper_sve_ldffhdu_be_zss, - gen_helper_sve_ldffsdu_be_zss, - gen_helper_sve_ldffdd_be_zss, } }, - { { gen_helper_sve_ldffbds_zd, - gen_helper_sve_ldffhds_be_zd, - gen_helper_sve_ldffsds_be_zd, - NULL, }, - { gen_helper_sve_ldffbdu_zd, - gen_helper_sve_ldffhdu_be_zd, - gen_helper_sve_ldffsdu_be_zd, - gen_helper_sve_ldffdd_be_zd, } } } }, + /* First-fault */ + { { { gen_helper_sve_ldffbds_zsu_mte, + gen_helper_sve_ldffhds_le_zsu_mte, + gen_helper_sve_ldffsds_le_zsu_mte, + NULL, }, + { gen_helper_sve_ldffbdu_zsu_mte, + gen_helper_sve_ldffhdu_le_zsu_mte, + gen_helper_sve_ldffsdu_le_zsu_mte, + gen_helper_sve_ldffdd_le_zsu_mte, } }, + { { gen_helper_sve_ldffbds_zss_mte, + gen_helper_sve_ldffhds_le_zss_mte, + gen_helper_sve_ldffsds_le_zss_mte, + NULL, }, + { gen_helper_sve_ldffbdu_zss_mte, + gen_helper_sve_ldffhdu_le_zss_mte, + gen_helper_sve_ldffsdu_le_zss_mte, + gen_helper_sve_ldffdd_le_zss_mte, } }, + { { gen_helper_sve_ldffbds_zd_mte, + gen_helper_sve_ldffhds_le_zd_mte, + gen_helper_sve_ldffsds_le_zd_mte, + NULL, }, + { gen_helper_sve_ldffbdu_zd_mte, + gen_helper_sve_ldffhdu_le_zd_mte, + gen_helper_sve_ldffsdu_le_zd_mte, + gen_helper_sve_ldffdd_le_zd_mte, } } } }, + { /* Big-endian */ + { { { gen_helper_sve_ldbds_zsu_mte, + gen_helper_sve_ldhds_be_zsu_mte, + gen_helper_sve_ldsds_be_zsu_mte, + NULL, }, + { gen_helper_sve_ldbdu_zsu_mte, + gen_helper_sve_ldhdu_be_zsu_mte, + gen_helper_sve_ldsdu_be_zsu_mte, + gen_helper_sve_lddd_be_zsu_mte, } }, + { { gen_helper_sve_ldbds_zss_mte, + gen_helper_sve_ldhds_be_zss_mte, + gen_helper_sve_ldsds_be_zss_mte, + NULL, }, + { gen_helper_sve_ldbdu_zss_mte, + gen_helper_sve_ldhdu_be_zss_mte, + gen_helper_sve_ldsdu_be_zss_mte, + gen_helper_sve_lddd_be_zss_mte, } }, + { { gen_helper_sve_ldbds_zd_mte, + gen_helper_sve_ldhds_be_zd_mte, + gen_helper_sve_ldsds_be_zd_mte, + NULL, }, + { gen_helper_sve_ldbdu_zd_mte, + gen_helper_sve_ldhdu_be_zd_mte, + gen_helper_sve_ldsdu_be_zd_mte, + gen_helper_sve_lddd_be_zd_mte, } } }, + + /* First-fault */ + { { { gen_helper_sve_ldffbds_zsu_mte, + gen_helper_sve_ldffhds_be_zsu_mte, + gen_helper_sve_ldffsds_be_zsu_mte, + NULL, }, + { gen_helper_sve_ldffbdu_zsu_mte, + gen_helper_sve_ldffhdu_be_zsu_mte, + gen_helper_sve_ldffsdu_be_zsu_mte, + gen_helper_sve_ldffdd_be_zsu_mte, } }, + { { gen_helper_sve_ldffbds_zss_mte, + gen_helper_sve_ldffhds_be_zss_mte, + gen_helper_sve_ldffsds_be_zss_mte, + NULL, }, + { gen_helper_sve_ldffbdu_zss_mte, + gen_helper_sve_ldffhdu_be_zss_mte, + gen_helper_sve_ldffsdu_be_zss_mte, + gen_helper_sve_ldffdd_be_zss_mte, } }, + { { gen_helper_sve_ldffbds_zd_mte, + gen_helper_sve_ldffhds_be_zd_mte, + gen_helper_sve_ldffsds_be_zd_mte, + NULL, }, + { gen_helper_sve_ldffbdu_zd_mte, + gen_helper_sve_ldffhdu_be_zd_mte, + gen_helper_sve_ldffsdu_be_zd_mte, + gen_helper_sve_ldffdd_be_zd_mte, } } } } }, }; static bool trans_LD1_zprz(DisasContext *s, arg_LD1_zprz *a) { gen_helper_gvec_mem_scatter *fn = NULL; - int be = s->be_data == MO_BE; + bool be = s->be_data == MO_BE; + bool mte = s->mte_active[0]; if (!sve_access_check(s)) { return true; @@ -XXX,XX +XXX,XX @@ static bool trans_LD1_zprz(DisasContext *s, arg_LD1_zprz *a) switch (a->esz) { case MO_32: - fn = gather_load_fn32[be][a->ff][a->xs][a->u][a->msz]; + fn = gather_load_fn32[mte][be][a->ff][a->xs][a->u][a->msz]; break; case MO_64: - fn = gather_load_fn64[be][a->ff][a->xs][a->u][a->msz]; + fn = gather_load_fn64[mte][be][a->ff][a->xs][a->u][a->msz]; break; } assert(fn != NULL); do_mem_zpz(s, a->rd, a->pg, a->rm, a->scale * a->msz, - cpu_reg_sp(s, a->rn), a->msz, fn); + cpu_reg_sp(s, a->rn), a->msz, false, fn); return true; } static bool trans_LD1_zpiz(DisasContext *s, arg_LD1_zpiz *a) { gen_helper_gvec_mem_scatter *fn = NULL; - int be = s->be_data == MO_BE; + bool be = s->be_data == MO_BE; + bool mte = s->mte_active[0]; TCGv_i64 imm; if (a->esz < a->msz || (a->esz == a->msz && !a->u)) { @@ -XXX,XX +XXX,XX @@ static bool trans_LD1_zpiz(DisasContext *s, arg_LD1_zpiz *a) switch (a->esz) { case MO_32: - fn = gather_load_fn32[be][a->ff][0][a->u][a->msz]; + fn = gather_load_fn32[mte][be][a->ff][0][a->u][a->msz]; break; case MO_64: - fn = gather_load_fn64[be][a->ff][2][a->u][a->msz]; + fn = gather_load_fn64[mte][be][a->ff][2][a->u][a->msz]; break; } assert(fn != NULL); @@ -XXX,XX +XXX,XX @@ static bool trans_LD1_zpiz(DisasContext *s, arg_LD1_zpiz *a) * by loading the immediate into the scalar parameter. */ imm = tcg_const_i64(a->imm << a->msz); - do_mem_zpz(s, a->rd, a->pg, a->rn, 0, imm, a->msz, fn); + do_mem_zpz(s, a->rd, a->pg, a->rn, 0, imm, a->msz, false, fn); tcg_temp_free_i64(imm); return true; } -/* Indexed by [be][xs][msz]. */ -static gen_helper_gvec_mem_scatter * const scatter_store_fn32[2][2][3] = { - /* Little-endian */ - { { gen_helper_sve_stbs_zsu, - gen_helper_sve_sths_le_zsu, - gen_helper_sve_stss_le_zsu, }, - { gen_helper_sve_stbs_zss, - gen_helper_sve_sths_le_zss, - gen_helper_sve_stss_le_zss, } }, - /* Big-endian */ - { { gen_helper_sve_stbs_zsu, - gen_helper_sve_sths_be_zsu, - gen_helper_sve_stss_be_zsu, }, - { gen_helper_sve_stbs_zss, - gen_helper_sve_sths_be_zss, - gen_helper_sve_stss_be_zss, } }, +/* Indexed by [mte][be][xs][msz]. */ +static gen_helper_gvec_mem_scatter * const scatter_store_fn32[2][2][2][3] = { + { /* MTE Inactive */ + { /* Little-endian */ + { gen_helper_sve_stbs_zsu, + gen_helper_sve_sths_le_zsu, + gen_helper_sve_stss_le_zsu, }, + { gen_helper_sve_stbs_zss, + gen_helper_sve_sths_le_zss, + gen_helper_sve_stss_le_zss, } }, + { /* Big-endian */ + { gen_helper_sve_stbs_zsu, + gen_helper_sve_sths_be_zsu, + gen_helper_sve_stss_be_zsu, }, + { gen_helper_sve_stbs_zss, + gen_helper_sve_sths_be_zss, + gen_helper_sve_stss_be_zss, } } }, + { /* MTE Active */ + { /* Little-endian */ + { gen_helper_sve_stbs_zsu_mte, + gen_helper_sve_sths_le_zsu_mte, + gen_helper_sve_stss_le_zsu_mte, }, + { gen_helper_sve_stbs_zss_mte, + gen_helper_sve_sths_le_zss_mte, + gen_helper_sve_stss_le_zss_mte, } }, + { /* Big-endian */ + { gen_helper_sve_stbs_zsu_mte, + gen_helper_sve_sths_be_zsu_mte, + gen_helper_sve_stss_be_zsu_mte, }, + { gen_helper_sve_stbs_zss_mte, + gen_helper_sve_sths_be_zss_mte, + gen_helper_sve_stss_be_zss_mte, } } }, }; /* Note that we overload xs=2 to indicate 64-bit offset. */ -static gen_helper_gvec_mem_scatter * const scatter_store_fn64[2][3][4] = { - /* Little-endian */ - { { gen_helper_sve_stbd_zsu, - gen_helper_sve_sthd_le_zsu, - gen_helper_sve_stsd_le_zsu, - gen_helper_sve_stdd_le_zsu, }, - { gen_helper_sve_stbd_zss, - gen_helper_sve_sthd_le_zss, - gen_helper_sve_stsd_le_zss, - gen_helper_sve_stdd_le_zss, }, - { gen_helper_sve_stbd_zd, - gen_helper_sve_sthd_le_zd, - gen_helper_sve_stsd_le_zd, - gen_helper_sve_stdd_le_zd, } }, - /* Big-endian */ - { { gen_helper_sve_stbd_zsu, - gen_helper_sve_sthd_be_zsu, - gen_helper_sve_stsd_be_zsu, - gen_helper_sve_stdd_be_zsu, }, - { gen_helper_sve_stbd_zss, - gen_helper_sve_sthd_be_zss, - gen_helper_sve_stsd_be_zss, - gen_helper_sve_stdd_be_zss, }, - { gen_helper_sve_stbd_zd, - gen_helper_sve_sthd_be_zd, - gen_helper_sve_stsd_be_zd, - gen_helper_sve_stdd_be_zd, } }, +static gen_helper_gvec_mem_scatter * const scatter_store_fn64[2][2][3][4] = { + { /* MTE Inactive */ + { /* Little-endian */ + { gen_helper_sve_stbd_zsu, + gen_helper_sve_sthd_le_zsu, + gen_helper_sve_stsd_le_zsu, + gen_helper_sve_stdd_le_zsu, }, + { gen_helper_sve_stbd_zss, + gen_helper_sve_sthd_le_zss, + gen_helper_sve_stsd_le_zss, + gen_helper_sve_stdd_le_zss, }, + { gen_helper_sve_stbd_zd, + gen_helper_sve_sthd_le_zd, + gen_helper_sve_stsd_le_zd, + gen_helper_sve_stdd_le_zd, } }, + { /* Big-endian */ + { gen_helper_sve_stbd_zsu, + gen_helper_sve_sthd_be_zsu, + gen_helper_sve_stsd_be_zsu, + gen_helper_sve_stdd_be_zsu, }, + { gen_helper_sve_stbd_zss, + gen_helper_sve_sthd_be_zss, + gen_helper_sve_stsd_be_zss, + gen_helper_sve_stdd_be_zss, }, + { gen_helper_sve_stbd_zd, + gen_helper_sve_sthd_be_zd, + gen_helper_sve_stsd_be_zd, + gen_helper_sve_stdd_be_zd, } } }, + { /* MTE Inactive */ + { /* Little-endian */ + { gen_helper_sve_stbd_zsu_mte, + gen_helper_sve_sthd_le_zsu_mte, + gen_helper_sve_stsd_le_zsu_mte, + gen_helper_sve_stdd_le_zsu_mte, }, + { gen_helper_sve_stbd_zss_mte, + gen_helper_sve_sthd_le_zss_mte, + gen_helper_sve_stsd_le_zss_mte, + gen_helper_sve_stdd_le_zss_mte, }, + { gen_helper_sve_stbd_zd_mte, + gen_helper_sve_sthd_le_zd_mte, + gen_helper_sve_stsd_le_zd_mte, + gen_helper_sve_stdd_le_zd_mte, } }, + { /* Big-endian */ + { gen_helper_sve_stbd_zsu_mte, + gen_helper_sve_sthd_be_zsu_mte, + gen_helper_sve_stsd_be_zsu_mte, + gen_helper_sve_stdd_be_zsu_mte, }, + { gen_helper_sve_stbd_zss_mte, + gen_helper_sve_sthd_be_zss_mte, + gen_helper_sve_stsd_be_zss_mte, + gen_helper_sve_stdd_be_zss_mte, }, + { gen_helper_sve_stbd_zd_mte, + gen_helper_sve_sthd_be_zd_mte, + gen_helper_sve_stsd_be_zd_mte, + gen_helper_sve_stdd_be_zd_mte, } } }, }; static bool trans_ST1_zprz(DisasContext *s, arg_ST1_zprz *a) { gen_helper_gvec_mem_scatter *fn; - int be = s->be_data == MO_BE; + bool be = s->be_data == MO_BE; + bool mte = s->mte_active[0]; if (a->esz < a->msz || (a->msz == 0 && a->scale)) { return false; @@ -XXX,XX +XXX,XX @@ static bool trans_ST1_zprz(DisasContext *s, arg_ST1_zprz *a) } switch (a->esz) { case MO_32: - fn = scatter_store_fn32[be][a->xs][a->msz]; + fn = scatter_store_fn32[mte][be][a->xs][a->msz]; break; case MO_64: - fn = scatter_store_fn64[be][a->xs][a->msz]; + fn = scatter_store_fn64[mte][be][a->xs][a->msz]; break; default: g_assert_not_reached(); } do_mem_zpz(s, a->rd, a->pg, a->rm, a->scale * a->msz, - cpu_reg_sp(s, a->rn), a->msz, fn); + cpu_reg_sp(s, a->rn), a->msz, true, fn); return true; } static bool trans_ST1_zpiz(DisasContext *s, arg_ST1_zpiz *a) { gen_helper_gvec_mem_scatter *fn = NULL; - int be = s->be_data == MO_BE; + bool be = s->be_data == MO_BE; + bool mte = s->mte_active[0]; TCGv_i64 imm; if (a->esz < a->msz) { @@ -XXX,XX +XXX,XX @@ static bool trans_ST1_zpiz(DisasContext *s, arg_ST1_zpiz *a) switch (a->esz) { case MO_32: - fn = scatter_store_fn32[be][0][a->msz]; + fn = scatter_store_fn32[mte][be][0][a->msz]; break; case MO_64: - fn = scatter_store_fn64[be][2][a->msz]; + fn = scatter_store_fn64[mte][be][2][a->msz]; break; } assert(fn != NULL); @@ -XXX,XX +XXX,XX @@ static bool trans_ST1_zpiz(DisasContext *s, arg_ST1_zpiz *a) * by loading the immediate into the scalar parameter. */ imm = tcg_const_i64(a->imm << a->msz); - do_mem_zpz(s, a->rd, a->pg, a->rn, 0, imm, a->msz, fn); + do_mem_zpz(s, a->rd, a->pg, a->rn, 0, imm, a->msz, true, fn); tcg_temp_free_i64(imm); return true; } -- 2.20.1 From: Richard Henderson <richard.henderson@linaro.org> There are a number of paths by which the TBI is still intact for user-only in the SVE helpers. Because we currently always set TBI for user-only, we do not need to pass down the actual TBI setting from above, and we can remove the top byte in the inner-most primitives, so that none are forgotten. Moreover, this keeps the "dirty" pointer around at the higher levels, where we need it for any MTE checking. Since the normal case, especially for user-only, goes through RAM, this clearing merely adds two insns per page lookup, which will be completely in the noise. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200626033144.790098-39-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/cpu.c | 3 +++ target/arm/sve_helper.c | 19 +++++++++++++++++-- target/arm/translate-a64.c | 5 +++++ 3 files changed, 25 insertions(+), 2 deletions(-) diff --git a/target/arm/cpu.c b/target/arm/cpu.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/cpu.c +++ b/target/arm/cpu.c @@ -XXX,XX +XXX,XX @@ static void arm_cpu_reset(DeviceState *dev) * Enable TBI0 and TBI1. While the real kernel only enables TBI0, * turning on both here will produce smaller code and otherwise * make no difference to the user-level emulation. + * + * In sve_probe_page, we assume that this is set. + * Do not modify this without other changes. */ env->cp15.tcr_el[1].raw_tcr = (3ULL << 37); #else diff --git a/target/arm/sve_helper.c b/target/arm/sve_helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/sve_helper.c +++ b/target/arm/sve_helper.c @@ -XXX,XX +XXX,XX @@ static void sve_##NAME##_host(void *vd, intptr_t reg_off, void *host) \ static void sve_##NAME##_tlb(CPUARMState *env, void *vd, intptr_t reg_off, \ target_ulong addr, uintptr_t ra) \ { \ - *(TYPEE *)(vd + H(reg_off)) = (TYPEM)TLB(env, addr, ra); \ + *(TYPEE *)(vd + H(reg_off)) = \ + (TYPEM)TLB(env, useronly_clean_ptr(addr), ra); \ } #define DO_ST_TLB(NAME, H, TYPEE, TYPEM, TLB) \ static void sve_##NAME##_tlb(CPUARMState *env, void *vd, intptr_t reg_off, \ target_ulong addr, uintptr_t ra) \ { \ - TLB(env, addr, (TYPEM)*(TYPEE *)(vd + H(reg_off)), ra); \ + TLB(env, useronly_clean_ptr(addr), \ + (TYPEM)*(TYPEE *)(vd + H(reg_off)), ra); \ } #define DO_LD_PRIM_1(NAME, H, TE, TM) \ @@ -XXX,XX +XXX,XX @@ static bool sve_probe_page(SVEHostPage *info, bool nofault, int flags; addr += mem_off; + + /* + * User-only currently always issues with TBI. See the comment + * above useronly_clean_ptr. Usually we clean this top byte away + * during translation, but we can't do that for e.g. vector + imm + * addressing modes. + * + * We currently always enable TBI for user-only, and do not provide + * a way to turn it off. So clean the pointer unconditionally here, + * rather than look it up here, or pass it down from above. + */ + addr = useronly_clean_ptr(addr); + flags = probe_access_flags(env, addr, access_type, mmu_idx, nofault, &info->host, retaddr); info->flags = flags; diff --git a/target/arm/translate-a64.c b/target/arm/translate-a64.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-a64.c +++ b/target/arm/translate-a64.c @@ -XXX,XX +XXX,XX @@ static void aarch64_tr_init_disas_context(DisasContextBase *dcbase, dc->features = env->features; dc->dcz_blocksize = arm_cpu->dcz_blocksize; +#ifdef CONFIG_USER_ONLY + /* In sve_probe_page, we assume TBI is enabled. */ + tcg_debug_assert(dc->tbid & 1); +#endif + /* Single step state. The code-generation logic here is: * SS_ACTIVE == 0: * generate code with no special handling for single-stepping (except -- 2.20.1 From: Richard Henderson <richard.henderson@linaro.org> This is DC GVA and DC GZVA, and the tag check for DC ZVA. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200626033144.790098-40-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/cpu.h | 4 +++- target/arm/helper.c | 16 ++++++++++++++++ target/arm/translate-a64.c | 39 ++++++++++++++++++++++++++++++++++++++ 3 files changed, 58 insertions(+), 1 deletion(-) diff --git a/target/arm/cpu.h b/target/arm/cpu.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/cpu.h +++ b/target/arm/cpu.h @@ -XXX,XX +XXX,XX @@ static inline uint64_t cpreg_to_kvm_id(uint32_t cpregid) #define ARM_CP_NZCV (ARM_CP_SPECIAL | 0x0300) #define ARM_CP_CURRENTEL (ARM_CP_SPECIAL | 0x0400) #define ARM_CP_DC_ZVA (ARM_CP_SPECIAL | 0x0500) -#define ARM_LAST_SPECIAL ARM_CP_DC_ZVA +#define ARM_CP_DC_GVA (ARM_CP_SPECIAL | 0x0600) +#define ARM_CP_DC_GZVA (ARM_CP_SPECIAL | 0x0700) +#define ARM_LAST_SPECIAL ARM_CP_DC_GZVA #define ARM_CP_FPU 0x1000 #define ARM_CP_SVE 0x2000 #define ARM_CP_NO_GDB 0x4000 diff --git a/target/arm/helper.c b/target/arm/helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper.c +++ b/target/arm/helper.c @@ -XXX,XX +XXX,XX @@ static const ARMCPRegInfo mte_el0_cacheop_reginfo[] = { .opc0 = 1, .opc1 = 3, .crn = 7, .crm = 14, .opc2 = 5, .type = ARM_CP_NOP, .access = PL0_W, .accessfn = aa64_cacheop_poc_access }, + { .name = "DC_GVA", .state = ARM_CP_STATE_AA64, + .opc0 = 1, .opc1 = 3, .crn = 7, .crm = 4, .opc2 = 3, + .access = PL0_W, .type = ARM_CP_DC_GVA, +#ifndef CONFIG_USER_ONLY + /* Avoid overhead of an access check that always passes in user-mode */ + .accessfn = aa64_zva_access, +#endif + }, + { .name = "DC_GZVA", .state = ARM_CP_STATE_AA64, + .opc0 = 1, .opc1 = 3, .crn = 7, .crm = 4, .opc2 = 4, + .access = PL0_W, .type = ARM_CP_DC_GZVA, +#ifndef CONFIG_USER_ONLY + /* Avoid overhead of an access check that always passes in user-mode */ + .accessfn = aa64_zva_access, +#endif + }, REGINFO_SENTINEL }; diff --git a/target/arm/translate-a64.c b/target/arm/translate-a64.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-a64.c +++ b/target/arm/translate-a64.c @@ -XXX,XX +XXX,XX @@ static void handle_sys(DisasContext *s, uint32_t insn, bool isread, } gen_helper_dc_zva(cpu_env, tcg_rt); return; + case ARM_CP_DC_GVA: + { + TCGv_i64 clean_addr, tag; + + /* + * DC_GVA, like DC_ZVA, requires that we supply the original + * pointer for an invalid page. Probe that address first. + */ + tcg_rt = cpu_reg(s, rt); + clean_addr = clean_data_tbi(s, tcg_rt); + gen_probe_access(s, clean_addr, MMU_DATA_STORE, MO_8); + + if (s->ata) { + /* Extract the tag from the register to match STZGM. */ + tag = tcg_temp_new_i64(); + tcg_gen_shri_i64(tag, tcg_rt, 56); + gen_helper_stzgm_tags(cpu_env, clean_addr, tag); + tcg_temp_free_i64(tag); + } + } + return; + case ARM_CP_DC_GZVA: + { + TCGv_i64 clean_addr, tag; + + /* For DC_GZVA, we can rely on DC_ZVA for the proper fault. */ + tcg_rt = cpu_reg(s, rt); + clean_addr = clean_data_tbi(s, tcg_rt); + gen_helper_dc_zva(cpu_env, clean_addr); + + if (s->ata) { + /* Extract the tag from the register to match STZGM. */ + tag = tcg_temp_new_i64(); + tcg_gen_shri_i64(tag, tcg_rt, 56); + gen_helper_stzgm_tags(cpu_env, clean_addr, tag); + tcg_temp_free_i64(tag); + } + } + return; default: break; } -- 2.20.1 From: Richard Henderson <richard.henderson@linaro.org> D1.10 specifies that exception handlers begin with tag checks overridden. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200626033144.790098-41-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/helper.c | 3 +++ 1 file changed, 3 insertions(+) diff --git a/target/arm/helper.c b/target/arm/helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper.c +++ b/target/arm/helper.c @@ -XXX,XX +XXX,XX @@ static void arm_cpu_do_interrupt_aarch64(CPUState *cs) break; } } + if (cpu_isar_feature(aa64_mte, cpu)) { + new_mode |= PSTATE_TCO; + } pstate_write(env, PSTATE_DAIF | new_mode); env->aarch64 = 1; -- 2.20.1 From: Richard Henderson <richard.henderson@linaro.org> We need to check the memattr of a page in order to determine whether it is Tagged for MTE. Between Stage1 and Stage2, this becomes simpler if we always collect this data, instead of occasionally being presented with NULL. Use the nonnull attribute to allow the compiler to check that all pointer arguments are non-null. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200626033144.790098-42-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/internals.h | 3 ++- target/arm/helper.c | 60 ++++++++++++++++++++--------------------- target/arm/m_helper.c | 11 +++++--- target/arm/tlb_helper.c | 4 ++- 4 files changed, 42 insertions(+), 36 deletions(-) diff --git a/target/arm/internals.h b/target/arm/internals.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/internals.h +++ b/target/arm/internals.h @@ -XXX,XX +XXX,XX @@ bool get_phys_addr(CPUARMState *env, target_ulong address, MMUAccessType access_type, ARMMMUIdx mmu_idx, hwaddr *phys_ptr, MemTxAttrs *attrs, int *prot, target_ulong *page_size, - ARMMMUFaultInfo *fi, ARMCacheAttrs *cacheattrs); + ARMMMUFaultInfo *fi, ARMCacheAttrs *cacheattrs) + __attribute__((nonnull)); void arm_log_exception(int idx); diff --git a/target/arm/helper.c b/target/arm/helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper.c +++ b/target/arm/helper.c @@ -XXX,XX +XXX,XX @@ static bool get_phys_addr_lpae(CPUARMState *env, target_ulong address, bool s1_is_el0, hwaddr *phys_ptr, MemTxAttrs *txattrs, int *prot, target_ulong *page_size_ptr, - ARMMMUFaultInfo *fi, ARMCacheAttrs *cacheattrs); + ARMMMUFaultInfo *fi, ARMCacheAttrs *cacheattrs) + __attribute__((nonnull)); #endif static void switch_mode(CPUARMState *env, int mode); @@ -XXX,XX +XXX,XX @@ static bool get_phys_addr_lpae(CPUARMState *env, target_ulong address, arm_tlb_bti_gp(txattrs) = true; } - if (cacheattrs != NULL) { - if (mmu_idx == ARMMMUIdx_Stage2) { - cacheattrs->attrs = convert_stage2_attrs(env, - extract32(attrs, 0, 4)); - } else { - /* Index into MAIR registers for cache attributes */ - uint8_t attrindx = extract32(attrs, 0, 3); - uint64_t mair = env->cp15.mair_el[regime_el(env, mmu_idx)]; - assert(attrindx <= 7); - cacheattrs->attrs = extract64(mair, attrindx * 8, 8); - } - cacheattrs->shareability = extract32(attrs, 6, 2); + if (mmu_idx == ARMMMUIdx_Stage2) { + cacheattrs->attrs = convert_stage2_attrs(env, extract32(attrs, 0, 4)); + } else { + /* Index into MAIR registers for cache attributes */ + uint8_t attrindx = extract32(attrs, 0, 3); + uint64_t mair = env->cp15.mair_el[regime_el(env, mmu_idx)]; + assert(attrindx <= 7); + cacheattrs->attrs = extract64(mair, attrindx * 8, 8); } + cacheattrs->shareability = extract32(attrs, 6, 2); *phys_ptr = descaddr; *page_size_ptr = page_size; @@ -XXX,XX +XXX,XX @@ bool get_phys_addr(CPUARMState *env, target_ulong address, ret = get_phys_addr_lpae(env, ipa, access_type, ARMMMUIdx_Stage2, mmu_idx == ARMMMUIdx_E10_0, phys_ptr, attrs, &s2_prot, - page_size, fi, - cacheattrs != NULL ? &cacheattrs2 : NULL); + page_size, fi, &cacheattrs2); fi->s2addr = ipa; /* Combine the S1 and S2 perms. */ *prot &= s2_prot; - /* Combine the S1 and S2 cache attributes, if needed */ - if (!ret && cacheattrs != NULL) { - if (env->cp15.hcr_el2 & HCR_DC) { - /* - * HCR.DC forces the first stage attributes to - * Normal Non-Shareable, - * Inner Write-Back Read-Allocate Write-Allocate, - * Outer Write-Back Read-Allocate Write-Allocate. - */ - cacheattrs->attrs = 0xff; - cacheattrs->shareability = 0; - } - *cacheattrs = combine_cacheattrs(*cacheattrs, cacheattrs2); + /* If S2 fails, return early. */ + if (ret) { + return ret; } - return ret; + /* Combine the S1 and S2 cache attributes. */ + if (env->cp15.hcr_el2 & HCR_DC) { + /* + * HCR.DC forces the first stage attributes to + * Normal Non-Shareable, + * Inner Write-Back Read-Allocate Write-Allocate, + * Outer Write-Back Read-Allocate Write-Allocate. + */ + cacheattrs->attrs = 0xff; + cacheattrs->shareability = 0; + } + *cacheattrs = combine_cacheattrs(*cacheattrs, cacheattrs2); + return 0; } else { /* * For non-EL2 CPUs a stage1+stage2 translation is just stage 1. @@ -XXX,XX +XXX,XX @@ hwaddr arm_cpu_get_phys_page_attrs_debug(CPUState *cs, vaddr addr, bool ret; ARMMMUFaultInfo fi = {}; ARMMMUIdx mmu_idx = arm_mmu_idx(env); + ARMCacheAttrs cacheattrs = {}; *attrs = (MemTxAttrs) {}; ret = get_phys_addr(env, addr, 0, mmu_idx, &phys_addr, - attrs, &prot, &page_size, &fi, NULL); + attrs, &prot, &page_size, &fi, &cacheattrs); if (ret) { return -1; diff --git a/target/arm/m_helper.c b/target/arm/m_helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/m_helper.c +++ b/target/arm/m_helper.c @@ -XXX,XX +XXX,XX @@ static bool v7m_stack_write(ARMCPU *cpu, uint32_t addr, uint32_t value, hwaddr physaddr; int prot; ARMMMUFaultInfo fi = {}; + ARMCacheAttrs cacheattrs = {}; bool secure = mmu_idx & ARM_MMU_IDX_M_S; int exc; bool exc_secure; if (get_phys_addr(env, addr, MMU_DATA_STORE, mmu_idx, &physaddr, - &attrs, &prot, &page_size, &fi, NULL)) { + &attrs, &prot, &page_size, &fi, &cacheattrs)) { /* MPU/SAU lookup failed */ if (fi.type == ARMFault_QEMU_SFault) { if (mode == STACK_LAZYFP) { @@ -XXX,XX +XXX,XX @@ static bool v7m_stack_read(ARMCPU *cpu, uint32_t *dest, uint32_t addr, hwaddr physaddr; int prot; ARMMMUFaultInfo fi = {}; + ARMCacheAttrs cacheattrs = {}; bool secure = mmu_idx & ARM_MMU_IDX_M_S; int exc; bool exc_secure; uint32_t value; if (get_phys_addr(env, addr, MMU_DATA_LOAD, mmu_idx, &physaddr, - &attrs, &prot, &page_size, &fi, NULL)) { + &attrs, &prot, &page_size, &fi, &cacheattrs)) { /* MPU/SAU lookup failed */ if (fi.type == ARMFault_QEMU_SFault) { qemu_log_mask(CPU_LOG_INT, @@ -XXX,XX +XXX,XX @@ static bool v7m_read_half_insn(ARMCPU *cpu, ARMMMUIdx mmu_idx, V8M_SAttributes sattrs = {}; MemTxAttrs attrs = {}; ARMMMUFaultInfo fi = {}; + ARMCacheAttrs cacheattrs = {}; MemTxResult txres; target_ulong page_size; hwaddr physaddr; @@ -XXX,XX +XXX,XX @@ static bool v7m_read_half_insn(ARMCPU *cpu, ARMMMUIdx mmu_idx, "...really SecureFault with SFSR.INVEP\n"); return false; } - if (get_phys_addr(env, addr, MMU_INST_FETCH, mmu_idx, - &physaddr, &attrs, &prot, &page_size, &fi, NULL)) { + if (get_phys_addr(env, addr, MMU_INST_FETCH, mmu_idx, &physaddr, + &attrs, &prot, &page_size, &fi, &cacheattrs)) { /* the MPU lookup failed */ env->v7m.cfsr[env->v7m.secure] |= R_V7M_CFSR_IACCVIOL_MASK; armv7m_nvic_set_pending(env->nvic, ARMV7M_EXCP_MEM, env->v7m.secure); diff --git a/target/arm/tlb_helper.c b/target/arm/tlb_helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/tlb_helper.c +++ b/target/arm/tlb_helper.c @@ -XXX,XX +XXX,XX @@ bool arm_cpu_tlb_fill(CPUState *cs, vaddr address, int size, int prot, ret; MemTxAttrs attrs = {}; ARMMMUFaultInfo fi = {}; + ARMCacheAttrs cacheattrs = {}; /* * Walk the page table and (if the mapping exists) add the page @@ -XXX,XX +XXX,XX @@ bool arm_cpu_tlb_fill(CPUState *cs, vaddr address, int size, */ ret = get_phys_addr(&cpu->env, address, access_type, core_to_arm_mmu_idx(&cpu->env, mmu_idx), - &phys_addr, &attrs, &prot, &page_size, &fi, NULL); + &phys_addr, &attrs, &prot, &page_size, + &fi, &cacheattrs); if (likely(!ret)) { /* * Map a single [sub]page. Regions smaller than our declared -- 2.20.1 From: Richard Henderson <richard.henderson@linaro.org> This "bit" is a particular value of the page's MemAttr. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200626033144.790098-43-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/helper.c | 48 ++++++++++++++++++++++++++++++++++++++--- target/arm/tlb_helper.c | 5 +++++ 2 files changed, 50 insertions(+), 3 deletions(-) diff --git a/target/arm/helper.c b/target/arm/helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper.c +++ b/target/arm/helper.c @@ -XXX,XX +XXX,XX @@ static uint8_t combine_cacheattr_nibble(uint8_t s1, uint8_t s2) */ static ARMCacheAttrs combine_cacheattrs(ARMCacheAttrs s1, ARMCacheAttrs s2) { - uint8_t s1lo = extract32(s1.attrs, 0, 4), s2lo = extract32(s2.attrs, 0, 4); - uint8_t s1hi = extract32(s1.attrs, 4, 4), s2hi = extract32(s2.attrs, 4, 4); + uint8_t s1lo, s2lo, s1hi, s2hi; ARMCacheAttrs ret; + bool tagged = false; + + if (s1.attrs == 0xf0) { + tagged = true; + s1.attrs = 0xff; + } + + s1lo = extract32(s1.attrs, 0, 4); + s2lo = extract32(s2.attrs, 0, 4); + s1hi = extract32(s1.attrs, 4, 4); + s2hi = extract32(s2.attrs, 4, 4); /* Combine shareability attributes (table D4-43) */ if (s1.shareability == 2 || s2.shareability == 2) { @@ -XXX,XX +XXX,XX @@ static ARMCacheAttrs combine_cacheattrs(ARMCacheAttrs s1, ARMCacheAttrs s2) } } + /* TODO: CombineS1S2Desc does not consider transient, only WB, RWA. */ + if (tagged && ret.attrs == 0xff) { + ret.attrs = 0xf0; + } + return ret; } @@ -XXX,XX +XXX,XX @@ bool get_phys_addr(CPUARMState *env, target_ulong address, * Normal Non-Shareable, * Inner Write-Back Read-Allocate Write-Allocate, * Outer Write-Back Read-Allocate Write-Allocate. + * Do not overwrite Tagged within attrs. */ - cacheattrs->attrs = 0xff; + if (cacheattrs->attrs != 0xf0) { + cacheattrs->attrs = 0xff; + } cacheattrs->shareability = 0; } *cacheattrs = combine_cacheattrs(*cacheattrs, cacheattrs2); @@ -XXX,XX +XXX,XX @@ bool get_phys_addr(CPUARMState *env, target_ulong address, /* Definitely a real MMU, not an MPU */ if (regime_translation_disabled(env, mmu_idx)) { + uint64_t hcr; + uint8_t memattr; + /* * MMU disabled. S1 addresses within aa64 translation regimes are * still checked for bounds -- see AArch64.TranslateAddressS1Off. @@ -XXX,XX +XXX,XX @@ bool get_phys_addr(CPUARMState *env, target_ulong address, *phys_ptr = address; *prot = PAGE_READ | PAGE_WRITE | PAGE_EXEC; *page_size = TARGET_PAGE_SIZE; + + /* Fill in cacheattr a-la AArch64.TranslateAddressS1Off. */ + hcr = arm_hcr_el2_eff(env); + cacheattrs->shareability = 0; + if (hcr & HCR_DC) { + if (hcr & HCR_DCT) { + memattr = 0xf0; /* Tagged, Normal, WB, RWA */ + } else { + memattr = 0xff; /* Normal, WB, RWA */ + } + } else if (access_type == MMU_INST_FETCH) { + if (regime_sctlr(env, mmu_idx) & SCTLR_I) { + memattr = 0xee; /* Normal, WT, RA, NT */ + } else { + memattr = 0x44; /* Normal, NC, No */ + } + cacheattrs->shareability = 2; /* outer sharable */ + } else { + memattr = 0x00; /* Device, nGnRnE */ + } + cacheattrs->attrs = memattr; return 0; } diff --git a/target/arm/tlb_helper.c b/target/arm/tlb_helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/tlb_helper.c +++ b/target/arm/tlb_helper.c @@ -XXX,XX +XXX,XX @@ bool arm_cpu_tlb_fill(CPUState *cs, vaddr address, int size, phys_addr &= TARGET_PAGE_MASK; address &= TARGET_PAGE_MASK; } + /* Notice and record tagged memory. */ + if (cpu_isar_feature(aa64_mte, cpu) && cacheattrs.attrs == 0xf0) { + arm_tlb_mte_tagged(&attrs) = true; + } + tlb_set_page_with_attrs(cs, address, phys_addr, attrs, prot, mmu_idx, page_size); return true; -- 2.20.1 From: Richard Henderson <richard.henderson@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Message-id: 20200626033144.790098-44-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/cpu.h | 6 ++++++ hw/arm/virt.c | 55 ++++++++++++++++++++++++++++++++++++++++++++++-- target/arm/cpu.c | 52 +++++++++++++++++++++++++++++++++++++++++---- 3 files changed, 107 insertions(+), 6 deletions(-) diff --git a/target/arm/cpu.h b/target/arm/cpu.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/cpu.h +++ b/target/arm/cpu.h @@ -XXX,XX +XXX,XX @@ struct ARMCPU { /* MemoryRegion to use for secure physical accesses */ MemoryRegion *secure_memory; + /* MemoryRegion to use for allocation tag accesses */ + MemoryRegion *tag_memory; + MemoryRegion *secure_tag_memory; + /* For v8M, pointer to the IDAU interface provided by board/SoC */ Object *idau; @@ -XXX,XX +XXX,XX @@ typedef enum ARMMMUIdxBit { typedef enum ARMASIdx { ARMASIdx_NS = 0, ARMASIdx_S = 1, + ARMASIdx_TagNS = 2, + ARMASIdx_TagS = 3, } ARMASIdx; /* Return the Exception Level targeted by debug exceptions. */ diff --git a/hw/arm/virt.c b/hw/arm/virt.c index XXXXXXX..XXXXXXX 100644 --- a/hw/arm/virt.c +++ b/hw/arm/virt.c @@ -XXX,XX +XXX,XX @@ static void create_platform_bus(VirtMachineState *vms) sysbus_mmio_get_region(s, 0)); } +static void create_tag_ram(MemoryRegion *tag_sysmem, + hwaddr base, hwaddr size, + const char *name) +{ + MemoryRegion *tagram = g_new(MemoryRegion, 1); + + memory_region_init_ram(tagram, NULL, name, size / 32, &error_fatal); + memory_region_add_subregion(tag_sysmem, base / 32, tagram); +} + static void create_secure_ram(VirtMachineState *vms, - MemoryRegion *secure_sysmem) + MemoryRegion *secure_sysmem, + MemoryRegion *secure_tag_sysmem) { MemoryRegion *secram = g_new(MemoryRegion, 1); char *nodename; @@ -XXX,XX +XXX,XX @@ static void create_secure_ram(VirtMachineState *vms, qemu_fdt_setprop_string(vms->fdt, nodename, "status", "disabled"); qemu_fdt_setprop_string(vms->fdt, nodename, "secure-status", "okay"); + if (secure_tag_sysmem) { + create_tag_ram(secure_tag_sysmem, base, size, "mach-virt.secure-tag"); + } + g_free(nodename); } @@ -XXX,XX +XXX,XX @@ static void machvirt_init(MachineState *machine) const CPUArchIdList *possible_cpus; MemoryRegion *sysmem = get_system_memory(); MemoryRegion *secure_sysmem = NULL; + MemoryRegion *tag_sysmem = NULL; + MemoryRegion *secure_tag_sysmem = NULL; int n, virt_max_cpus; bool firmware_loaded; bool aarch64 = true; @@ -XXX,XX +XXX,XX @@ static void machvirt_init(MachineState *machine) "secure-memory", &error_abort); } + /* + * The cpu adds the property if and only if MemTag is supported. + * If it is, we must allocate the ram to back that up. + */ + if (object_property_find(cpuobj, "tag-memory", NULL)) { + if (!tag_sysmem) { + tag_sysmem = g_new(MemoryRegion, 1); + memory_region_init(tag_sysmem, OBJECT(machine), + "tag-memory", UINT64_MAX / 32); + + if (vms->secure) { + secure_tag_sysmem = g_new(MemoryRegion, 1); + memory_region_init(secure_tag_sysmem, OBJECT(machine), + "secure-tag-memory", UINT64_MAX / 32); + + /* As with ram, secure-tag takes precedence over tag. */ + memory_region_add_subregion_overlap(secure_tag_sysmem, 0, + tag_sysmem, -1); + } + } + + object_property_set_link(cpuobj, OBJECT(tag_sysmem), + "tag-memory", &error_abort); + if (vms->secure) { + object_property_set_link(cpuobj, OBJECT(secure_tag_sysmem), + "secure-tag-memory", &error_abort); + } + } + qdev_realize(DEVICE(cpuobj), NULL, &error_fatal); object_unref(cpuobj); } @@ -XXX,XX +XXX,XX @@ static void machvirt_init(MachineState *machine) create_uart(vms, VIRT_UART, sysmem, serial_hd(0)); if (vms->secure) { - create_secure_ram(vms, secure_sysmem); + create_secure_ram(vms, secure_sysmem, secure_tag_sysmem); create_uart(vms, VIRT_SECURE_UART, secure_sysmem, serial_hd(1)); } + if (tag_sysmem) { + create_tag_ram(tag_sysmem, vms->memmap[VIRT_MEM].base, + machine->ram_size, "mach-virt.tag"); + } + vms->highmem_ecam &= vms->highmem && (!firmware_loaded || aarch64); create_rtc(vms); diff --git a/target/arm/cpu.c b/target/arm/cpu.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/cpu.c +++ b/target/arm/cpu.c @@ -XXX,XX +XXX,XX @@ void arm_cpu_post_init(Object *obj) if (kvm_enabled()) { kvm_arm_add_vcpu_properties(obj); } + +#ifndef CONFIG_USER_ONLY + if (arm_feature(&cpu->env, ARM_FEATURE_AARCH64) && + cpu_isar_feature(aa64_mte, cpu)) { + object_property_add_link(obj, "tag-memory", + TYPE_MEMORY_REGION, + (Object **)&cpu->tag_memory, + qdev_prop_allow_set_link_before_realize, + OBJ_PROP_LINK_STRONG); + + if (arm_feature(&cpu->env, ARM_FEATURE_EL3)) { + object_property_add_link(obj, "secure-tag-memory", + TYPE_MEMORY_REGION, + (Object **)&cpu->secure_tag_memory, + qdev_prop_allow_set_link_before_realize, + OBJ_PROP_LINK_STRONG); + } + } +#endif } static void arm_cpu_finalizefn(Object *obj) @@ -XXX,XX +XXX,XX @@ static void arm_cpu_realizefn(DeviceState *dev, Error **errp) #ifndef CONFIG_USER_ONLY MachineState *ms = MACHINE(qdev_get_machine()); unsigned int smp_cpus = ms->smp.cpus; + bool has_secure = cpu->has_el3 || arm_feature(env, ARM_FEATURE_M_SECURITY); - if (cpu->has_el3 || arm_feature(env, ARM_FEATURE_M_SECURITY)) { - cs->num_ases = 2; + /* + * We must set cs->num_ases to the final value before + * the first call to cpu_address_space_init. + */ + if (cpu->tag_memory != NULL) { + cs->num_ases = 3 + has_secure; + } else { + cs->num_ases = 1 + has_secure; + } + if (has_secure) { if (!cpu->secure_memory) { cpu->secure_memory = cs->memory; } cpu_address_space_init(cs, ARMASIdx_S, "cpu-secure-memory", cpu->secure_memory); - } else { - cs->num_ases = 1; } + + if (cpu->tag_memory != NULL) { + cpu_address_space_init(cs, ARMASIdx_TagNS, "cpu-tag-memory", + cpu->tag_memory); + if (has_secure) { + cpu_address_space_init(cs, ARMASIdx_TagS, "cpu-tag-memory", + cpu->secure_tag_memory); + } + } else if (cpu_isar_feature(aa64_mte, cpu)) { + /* + * Since there is no tag memory, we can't meaningfully support MTE + * to its fullest. To avoid problems later, when we would come to + * use the tag memory, downgrade support to insns only. + */ + cpu->isar.id_aa64pfr1 = + FIELD_DP64(cpu->isar.id_aa64pfr1, ID_AA64PFR1, MTE, 1); + } + cpu_address_space_init(cs, ARMASIdx_NS, "cpu-memory", cs->memory); /* No core_count specified, default to smp_cpus. */ -- 2.20.1 From: Richard Henderson <richard.henderson@linaro.org> Look up the physical address for the given virtual address, convert that to a tag physical address, and finally return the host address that backs it. Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200626033144.790098-45-richard.henderson@linaro.org Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/mte_helper.c | 131 ++++++++++++++++++++++++++++++++++++++++ 1 file changed, 131 insertions(+) diff --git a/target/arm/mte_helper.c b/target/arm/mte_helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/mte_helper.c +++ b/target/arm/mte_helper.c @@ -XXX,XX +XXX,XX @@ #include "cpu.h" #include "internals.h" #include "exec/exec-all.h" +#include "exec/ram_addr.h" #include "exec/cpu_ldst.h" #include "exec/helper-proto.h" @@ -XXX,XX +XXX,XX @@ static uint8_t *allocation_tag_mem(CPUARMState *env, int ptr_mmu_idx, int ptr_size, MMUAccessType tag_access, int tag_size, uintptr_t ra) { +#ifdef CONFIG_USER_ONLY /* Tag storage not implemented. */ return NULL; +#else + uintptr_t index; + CPUIOTLBEntry *iotlbentry; + int in_page, flags; + ram_addr_t ptr_ra; + hwaddr ptr_paddr, tag_paddr, xlat; + MemoryRegion *mr; + ARMASIdx tag_asi; + AddressSpace *tag_as; + void *host; + + /* + * Probe the first byte of the virtual address. This raises an + * exception for inaccessible pages, and resolves the virtual address + * into the softmmu tlb. + * + * When RA == 0, this is for mte_probe1. The page is expected to be + * valid. Indicate to probe_access_flags no-fault, then assert that + * we received a valid page. + */ + flags = probe_access_flags(env, ptr, ptr_access, ptr_mmu_idx, + ra == 0, &host, ra); + assert(!(flags & TLB_INVALID_MASK)); + + /* + * Find the iotlbentry for ptr. This *must* be present in the TLB + * because we just found the mapping. + * TODO: Perhaps there should be a cputlb helper that returns a + * matching tlb entry + iotlb entry. + */ + index = tlb_index(env, ptr_mmu_idx, ptr); +# ifdef CONFIG_DEBUG_TCG + { + CPUTLBEntry *entry = tlb_entry(env, ptr_mmu_idx, ptr); + target_ulong comparator = (ptr_access == MMU_DATA_LOAD + ? entry->addr_read + : tlb_addr_write(entry)); + g_assert(tlb_hit(comparator, ptr)); + } +# endif + iotlbentry = &env_tlb(env)->d[ptr_mmu_idx].iotlb[index]; + + /* If the virtual page MemAttr != Tagged, access unchecked. */ + if (!arm_tlb_mte_tagged(&iotlbentry->attrs)) { + return NULL; + } + + /* + * If not backed by host ram, there is no tag storage: access unchecked. + * This is probably a guest os bug though, so log it. + */ + if (unlikely(flags & TLB_MMIO)) { + qemu_log_mask(LOG_GUEST_ERROR, + "Page @ 0x%" PRIx64 " indicates Tagged Normal memory " + "but is not backed by host ram\n", ptr); + return NULL; + } + + /* + * The Normal memory access can extend to the next page. E.g. a single + * 8-byte access to the last byte of a page will check only the last + * tag on the first page. + * Any page access exception has priority over tag check exception. + */ + in_page = -(ptr | TARGET_PAGE_MASK); + if (unlikely(ptr_size > in_page)) { + void *ignore; + flags |= probe_access_flags(env, ptr + in_page, ptr_access, + ptr_mmu_idx, ra == 0, &ignore, ra); + assert(!(flags & TLB_INVALID_MASK)); + } + + /* Any debug exception has priority over a tag check exception. */ + if (unlikely(flags & TLB_WATCHPOINT)) { + int wp = ptr_access == MMU_DATA_LOAD ? BP_MEM_READ : BP_MEM_WRITE; + assert(ra != 0); + cpu_check_watchpoint(env_cpu(env), ptr, ptr_size, + iotlbentry->attrs, wp, ra); + } + + /* + * Find the physical address within the normal mem space. + * The memory region lookup must succeed because TLB_MMIO was + * not set in the cputlb lookup above. + */ + mr = memory_region_from_host(host, &ptr_ra); + tcg_debug_assert(mr != NULL); + tcg_debug_assert(memory_region_is_ram(mr)); + ptr_paddr = ptr_ra; + do { + ptr_paddr += mr->addr; + mr = mr->container; + } while (mr); + + /* Convert to the physical address in tag space. */ + tag_paddr = ptr_paddr >> (LOG2_TAG_GRANULE + 1); + + /* Look up the address in tag space. */ + tag_asi = iotlbentry->attrs.secure ? ARMASIdx_TagS : ARMASIdx_TagNS; + tag_as = cpu_get_address_space(env_cpu(env), tag_asi); + mr = address_space_translate(tag_as, tag_paddr, &xlat, NULL, + tag_access == MMU_DATA_STORE, + iotlbentry->attrs); + + /* + * Note that @mr will never be NULL. If there is nothing in the address + * space at @tag_paddr, the translation will return the unallocated memory + * region. For our purposes, the result must be ram. + */ + if (unlikely(!memory_region_is_ram(mr))) { + /* ??? Failure is a board configuration error. */ + qemu_log_mask(LOG_UNIMP, + "Tag Memory @ 0x%" HWADDR_PRIx " not found for " + "Normal Memory @ 0x%" HWADDR_PRIx "\n", + tag_paddr, ptr_paddr); + return NULL; + } + + /* + * Ensure the tag memory is dirty on write, for migration. + * Tag memory can never contain code or display memory (vga). + */ + if (tag_access == MMU_DATA_STORE) { + ram_addr_t tag_ra = memory_region_get_ram_addr(mr) + xlat; + cpu_physical_memory_set_dirty_flag(tag_ra, DIRTY_MEMORY_MIGRATION); + } + + return memory_region_get_ram_ptr(mr) + xlat; +#endif } uint64_t HELPER(irg)(CPUARMState *env, uint64_t rn, uint64_t rm) -- 2.20.1 From: Richard Henderson <richard.henderson@linaro.org> We now implement all of the components of MTE, without actually supporting any tagged memory. All MTE instructions will work, trivially, so we can enable support. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200626033144.790098-46-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/cpu64.c | 5 +++++ 1 file changed, 5 insertions(+) diff --git a/target/arm/cpu64.c b/target/arm/cpu64.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/cpu64.c +++ b/target/arm/cpu64.c @@ -XXX,XX +XXX,XX @@ static void aarch64_max_initfn(Object *obj) t = cpu->isar.id_aa64pfr1; t = FIELD_DP64(t, ID_AA64PFR1, BT, 1); + /* + * Begin with full support for MTE; will be downgraded to MTE=1 + * during realize if the board provides no tag memory. + */ + t = FIELD_DP64(t, ID_AA64PFR1, MTE, 2); cpu->isar.id_aa64pfr1 = t; t = cpu->isar.id_aa64mmfr1; -- 2.20.1