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 target-arm queue, mostly SME preliminaries. In the unlikely event we don't land the rest of SME before freeze for 7.1 we can revert the docs/property changes included here. -- PMM The following changes since commit 097ccbbbaf2681df1e65542e5b7d2b2d0c66e2bc: Merge tag 'qemu-sparc-20220626' of https://github.com/mcayland/qemu into staging (2022-06-27 05:21:05 +0530) are available in the Git repository at: https://git.linaro.org/people/pmaydell/qemu-arm.git tags/pull-target-arm-20220627 for you to fetch changes up to 59e1b8a22ea9f947d038ccac784de1020f266e14: target/arm: Check V7VE as well as LPAE in arm_pamax (2022-06-27 11:18:17 +0100) ---------------------------------------------------------------- target-arm queue: * sphinx: change default language to 'en' * Diagnose attempts to emulate EL3 in hvf as well as kvm * More SME groundwork patches * virt: Fix calculation of physical address space size for v7VE CPUs (eg cortex-a15) ---------------------------------------------------------------- Alexander Graf (2): accel: Introduce current_accel_name() target/arm: Catch invalid kvm state also for hvf Martin Liška (1): sphinx: change default language to 'en' Richard Henderson (22): target/arm: Implement TPIDR2_EL0 target/arm: Add SMEEXC_EL to TB flags target/arm: Add syn_smetrap target/arm: Add ARM_CP_SME target/arm: Add SVCR target/arm: Add SMCR_ELx target/arm: Add SMIDR_EL1, SMPRI_EL1, SMPRIMAP_EL2 target/arm: Add PSTATE.{SM,ZA} to TB flags target/arm: Add the SME ZA storage to CPUARMState target/arm: Implement SMSTART, SMSTOP target/arm: Move error for sve%d property to arm_cpu_sve_finalize target/arm: Create ARMVQMap target/arm: Generalize cpu_arm_{get,set}_vq target/arm: Generalize cpu_arm_{get, set}_default_vec_len target/arm: Move arm_cpu_*_finalize to internals.h target/arm: Unexport aarch64_add_*_properties target/arm: Add cpu properties for SME target/arm: Introduce sve_vqm1_for_el_sm target/arm: Add SVL to TB flags target/arm: Move pred_{full, gvec}_reg_{offset, size} to translate-a64.h target/arm: Extend arm_pamax to more than aarch64 target/arm: Check V7VE as well as LPAE in arm_pamax docs/conf.py | 2 +- docs/system/arm/cpu-features.rst | 56 ++++++++++ include/qemu/accel.h | 1 + target/arm/cpregs.h | 5 + target/arm/cpu.h | 103 ++++++++++++++----- target/arm/helper-sme.h | 21 ++++ target/arm/helper.h | 1 + target/arm/internals.h | 4 + target/arm/syndrome.h | 14 +++ target/arm/translate-a64.h | 38 +++++++ target/arm/translate.h | 6 ++ accel/accel-common.c | 8 ++ hw/arm/virt.c | 10 +- softmmu/vl.c | 3 +- target/arm/cpu.c | 32 ++++-- target/arm/cpu64.c | 205 ++++++++++++++++++++++++++++--------- target/arm/helper.c | 213 +++++++++++++++++++++++++++++++++++++-- target/arm/kvm64.c | 2 +- target/arm/machine.c | 34 +++++++ target/arm/ptw.c | 26 +++-- target/arm/sme_helper.c | 61 +++++++++++ target/arm/translate-a64.c | 46 +++++++++ target/arm/translate-sve.c | 36 ------- target/arm/meson.build | 1 + 24 files changed, 782 insertions(+), 146 deletions(-) create mode 100644 target/arm/helper-sme.h create mode 100644 target/arm/sme_helper.c From: Martin Liška <mliska@suse.cz> Fixes the following Sphinx warning (treated as error) starting with 5.0 release: Warning, treated as error: Invalid configuration value found: 'language = None'. Update your configuration to a valid langauge code. Falling back to 'en' (English). Signed-off-by: Martin Liska <mliska@suse.cz> Message-id: e91e51ee-48ac-437e-6467-98b56ee40042@suse.cz Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- docs/conf.py | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/docs/conf.py b/docs/conf.py index XXXXXXX..XXXXXXX 100644 --- a/docs/conf.py +++ b/docs/conf.py @@ -XXX,XX +XXX,XX @@ # # This is also used if you do content translation via gettext catalogs. # Usually you set "language" from the command line for these cases. -language = None +language = 'en' # List of patterns, relative to source directory, that match files and # directories to ignore when looking for source files. -- 2.25.1 From: Alexander Graf <agraf@csgraf.de> We need to fetch the name of the current accelerator in flexible error messages more going forward. Let's create a helper that gives it to us without casting in the target code. Signed-off-by: Alexander Graf <agraf@csgraf.de> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20220620192242.70573-1-agraf@csgraf.de Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- include/qemu/accel.h | 1 + accel/accel-common.c | 8 ++++++++ softmmu/vl.c | 3 +-- 3 files changed, 10 insertions(+), 2 deletions(-) diff --git a/include/qemu/accel.h b/include/qemu/accel.h index XXXXXXX..XXXXXXX 100644 --- a/include/qemu/accel.h +++ b/include/qemu/accel.h @@ -XXX,XX +XXX,XX @@ typedef struct AccelClass { AccelClass *accel_find(const char *opt_name); AccelState *current_accel(void); +const char *current_accel_name(void); void accel_init_interfaces(AccelClass *ac); diff --git a/accel/accel-common.c b/accel/accel-common.c index XXXXXXX..XXXXXXX 100644 --- a/accel/accel-common.c +++ b/accel/accel-common.c @@ -XXX,XX +XXX,XX @@ AccelClass *accel_find(const char *opt_name) return ac; } +/* Return the name of the current accelerator */ +const char *current_accel_name(void) +{ + AccelClass *ac = ACCEL_GET_CLASS(current_accel()); + + return ac->name; +} + static void accel_init_cpu_int_aux(ObjectClass *klass, void *opaque) { CPUClass *cc = CPU_CLASS(klass); diff --git a/softmmu/vl.c b/softmmu/vl.c index XXXXXXX..XXXXXXX 100644 --- a/softmmu/vl.c +++ b/softmmu/vl.c @@ -XXX,XX +XXX,XX @@ static void configure_accelerators(const char *progname) } if (init_failed && !qtest_chrdev) { - AccelClass *ac = ACCEL_GET_CLASS(current_accel()); - error_report("falling back to %s", ac->name); + error_report("falling back to %s", current_accel_name()); } if (icount_enabled() && !tcg_enabled()) { -- 2.25.1 From: Alexander Graf <agraf@csgraf.de> Some features such as running in EL3 or running M profile code are incompatible with virtualization as QEMU implements it today. To prevent users from picking invalid configurations on other virt solutions like Hvf, let's run the same checks there too. Resolves: https://gitlab.com/qemu-project/qemu/-/issues/1073 Signed-off-by: Alexander Graf <agraf@csgraf.de> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20220620192242.70573-2-agraf@csgraf.de [PMM: Allow qtest accelerator too; tweak comment] Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/cpu.c | 16 ++++++++++++---- 1 file changed, 12 insertions(+), 4 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 @@ #include "hw/boards.h" #endif #include "sysemu/tcg.h" +#include "sysemu/qtest.h" #include "sysemu/hw_accel.h" #include "kvm_arm.h" #include "disas/capstone.h" @@ -XXX,XX +XXX,XX @@ static void arm_cpu_realizefn(DeviceState *dev, Error **errp) } } - if (kvm_enabled()) { + if (!tcg_enabled() && !qtest_enabled()) { /* + * We assume that no accelerator except TCG (and the "not really an + * accelerator" qtest) can handle these features, because Arm hardware + * virtualization can't virtualize them. + * * Catch all the cases which might cause us to create more than one * address space for the CPU (otherwise we will assert() later in * cpu_address_space_init()). */ if (arm_feature(env, ARM_FEATURE_M)) { error_setg(errp, - "Cannot enable KVM when using an M-profile guest CPU"); + "Cannot enable %s when using an M-profile guest CPU", + current_accel_name()); return; } if (cpu->has_el3) { error_setg(errp, - "Cannot enable KVM when guest CPU has EL3 enabled"); + "Cannot enable %s when guest CPU has EL3 enabled", + current_accel_name()); return; } if (cpu->tag_memory) { error_setg(errp, - "Cannot enable KVM when guest CPUs has MTE enabled"); + "Cannot enable %s when guest CPUs has MTE enabled", + current_accel_name()); return; } } -- 2.25.1 From: Richard Henderson <richard.henderson@linaro.org> This register is part of SME, but isn't closely related to the rest of the extension. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20220620175235.60881-2-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/cpu.h | 1 + target/arm/helper.c | 32 ++++++++++++++++++++++++++++++++ 2 files changed, 33 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 CPUArchState { }; uint64_t tpidr_el[4]; }; + uint64_t tpidr2_el0; /* The secure banks of these registers don't map anywhere */ uint64_t tpidrurw_s; uint64_t tpidrprw_s; 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 zcr_reginfo[] = { .writefn = zcr_write, .raw_writefn = raw_write }, }; +#ifdef TARGET_AARCH64 +static CPAccessResult access_tpidr2(CPUARMState *env, const ARMCPRegInfo *ri, + bool isread) +{ + int el = arm_current_el(env); + + if (el == 0) { + uint64_t sctlr = arm_sctlr(env, el); + if (!(sctlr & SCTLR_EnTP2)) { + return CP_ACCESS_TRAP; + } + } + /* TODO: FEAT_FGT */ + if (el < 3 + && arm_feature(env, ARM_FEATURE_EL3) + && !(env->cp15.scr_el3 & SCR_ENTP2)) { + return CP_ACCESS_TRAP_EL3; + } + return CP_ACCESS_OK; +} + +static const ARMCPRegInfo sme_reginfo[] = { + { .name = "TPIDR2_EL0", .state = ARM_CP_STATE_AA64, + .opc0 = 3, .opc1 = 3, .crn = 13, .crm = 0, .opc2 = 5, + .access = PL0_RW, .accessfn = access_tpidr2, + .fieldoffset = offsetof(CPUARMState, cp15.tpidr2_el0) }, +}; +#endif /* TARGET_AARCH64 */ + void hw_watchpoint_update(ARMCPU *cpu, int n) { CPUARMState *env = &cpu->env; @@ -XXX,XX +XXX,XX @@ void register_cp_regs_for_features(ARMCPU *cpu) } #ifdef TARGET_AARCH64 + if (cpu_isar_feature(aa64_sme, cpu)) { + define_arm_cp_regs(cpu, sme_reginfo); + } if (cpu_isar_feature(aa64_pauth, cpu)) { define_arm_cp_regs(cpu, pauth_reginfo); } -- 2.25.1 From: Richard Henderson <richard.henderson@linaro.org> This is CheckSMEAccess, which is the basis for a set of related tests for various SME cpregs and instructions. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20220620175235.60881-3-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/cpu.h | 2 ++ target/arm/translate.h | 1 + target/arm/helper.c | 52 ++++++++++++++++++++++++++++++++++++++ target/arm/translate-a64.c | 1 + 4 files changed, 56 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 @@ void aarch64_sync_64_to_32(CPUARMState *env); int fp_exception_el(CPUARMState *env, int cur_el); int sve_exception_el(CPUARMState *env, int cur_el); +int sme_exception_el(CPUARMState *env, int cur_el); /** * sve_vqm1_for_el: @@ -XXX,XX +XXX,XX @@ 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) +FIELD(TBFLAG_A64, SMEEXC_EL, 20, 2) /* * Helpers for using the above. 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 { 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 */ + int sme_excp_el; /* SME exception EL or 0 if enabled */ int vl; /* current vector length in bytes */ bool vfp_enabled; /* FP enabled via FPSCR.EN */ int vec_len; 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 @@ int sve_exception_el(CPUARMState *env, int el) return 0; } +/* + * Return the exception level to which exceptions should be taken for SME. + * C.f. the ARM pseudocode function CheckSMEAccess. + */ +int sme_exception_el(CPUARMState *env, int el) +{ +#ifndef CONFIG_USER_ONLY + if (el <= 1 && !el_is_in_host(env, el)) { + switch (FIELD_EX64(env->cp15.cpacr_el1, CPACR_EL1, SMEN)) { + case 1: + if (el != 0) { + break; + } + /* fall through */ + case 0: + case 2: + return 1; + } + } + + if (el <= 2 && arm_is_el2_enabled(env)) { + /* CPTR_EL2 changes format with HCR_EL2.E2H (regardless of TGE). */ + if (env->cp15.hcr_el2 & HCR_E2H) { + switch (FIELD_EX64(env->cp15.cptr_el[2], CPTR_EL2, SMEN)) { + case 1: + if (el != 0 || !(env->cp15.hcr_el2 & HCR_TGE)) { + break; + } + /* fall through */ + case 0: + case 2: + return 2; + } + } else { + if (FIELD_EX64(env->cp15.cptr_el[2], CPTR_EL2, TSM)) { + return 2; + } + } + } + + /* CPTR_EL3. Since ESM is negative we must check for EL3. */ + if (arm_feature(env, ARM_FEATURE_EL3) + && !FIELD_EX64(env->cp15.cptr_el[3], CPTR_EL3, ESM)) { + return 3; + } +#endif + return 0; +} + /* * Given that SVE is enabled, return the vector length for EL. */ @@ -XXX,XX +XXX,XX @@ static CPUARMTBFlags rebuild_hflags_a64(CPUARMState *env, int el, int fp_el, } DP_TBFLAG_A64(flags, SVEEXC_EL, sve_el); } + if (cpu_isar_feature(aa64_sme, env_archcpu(env))) { + DP_TBFLAG_A64(flags, SMEEXC_EL, sme_exception_el(env, el)); + } sctlr = regime_sctlr(env, stage1); 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->align_mem = EX_TBFLAG_ANY(tb_flags, ALIGN_MEM); dc->pstate_il = EX_TBFLAG_ANY(tb_flags, PSTATE__IL); dc->sve_excp_el = EX_TBFLAG_A64(tb_flags, SVEEXC_EL); + dc->sme_excp_el = EX_TBFLAG_A64(tb_flags, SMEEXC_EL); dc->vl = (EX_TBFLAG_A64(tb_flags, VL) + 1) * 16; dc->pauth_active = EX_TBFLAG_A64(tb_flags, PAUTH_ACTIVE); dc->bt = EX_TBFLAG_A64(tb_flags, BT); -- 2.25.1 From: Richard Henderson <richard.henderson@linaro.org> This will be used for raising various traps for SME. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20220620175235.60881-4-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/syndrome.h | 14 ++++++++++++++ 1 file changed, 14 insertions(+) diff --git a/target/arm/syndrome.h b/target/arm/syndrome.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/syndrome.h +++ b/target/arm/syndrome.h @@ -XXX,XX +XXX,XX @@ enum arm_exception_class { EC_AA64_SMC = 0x17, EC_SYSTEMREGISTERTRAP = 0x18, EC_SVEACCESSTRAP = 0x19, + EC_SMETRAP = 0x1d, EC_INSNABORT = 0x20, EC_INSNABORT_SAME_EL = 0x21, EC_PCALIGNMENT = 0x22, @@ -XXX,XX +XXX,XX @@ enum arm_exception_class { EC_AA64_BKPT = 0x3c, }; +typedef enum { + SME_ET_AccessTrap, + SME_ET_Streaming, + SME_ET_NotStreaming, + SME_ET_InactiveZA, +} SMEExceptionType; + #define ARM_EL_EC_SHIFT 26 #define ARM_EL_IL_SHIFT 25 #define ARM_EL_ISV_SHIFT 24 @@ -XXX,XX +XXX,XX @@ static inline uint32_t syn_sve_access_trap(void) return EC_SVEACCESSTRAP << ARM_EL_EC_SHIFT; } +static inline uint32_t syn_smetrap(SMEExceptionType etype, bool is_16bit) +{ + return (EC_SMETRAP << ARM_EL_EC_SHIFT) + | (is_16bit ? 0 : ARM_EL_IL) | etype; +} + static inline uint32_t syn_pactrap(void) { return EC_PACTRAP << ARM_EL_EC_SHIFT; -- 2.25.1 From: Richard Henderson <richard.henderson@linaro.org> This will be used for controlling access to SME cpregs. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20220620175235.60881-5-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/cpregs.h | 5 +++++ target/arm/translate-a64.c | 18 ++++++++++++++++++ 2 files changed, 23 insertions(+) diff --git a/target/arm/cpregs.h b/target/arm/cpregs.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/cpregs.h +++ b/target/arm/cpregs.h @@ -XXX,XX +XXX,XX @@ enum { ARM_CP_EL3_NO_EL2_UNDEF = 1 << 16, ARM_CP_EL3_NO_EL2_KEEP = 1 << 17, ARM_CP_EL3_NO_EL2_C_NZ = 1 << 18, + /* + * Flag: Access check for this sysreg is constrained by the + * ARM pseudocode function CheckSMEAccess(). + */ + ARM_CP_SME = 1 << 19, }; /* 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 @@ bool sve_access_check(DisasContext *s) return fp_access_check(s); } +/* + * Check that SME access is enabled, raise an exception if not. + * Note that this function corresponds to CheckSMEAccess and is + * only used directly for cpregs. + */ +static bool sme_access_check(DisasContext *s) +{ + if (s->sme_excp_el) { + gen_exception_insn_el(s, s->pc_curr, EXCP_UDEF, + syn_smetrap(SME_ET_AccessTrap, false), + s->sme_excp_el); + return false; + } + return true; +} + /* * This utility function is for doing register extension with an * optional shift. You will likely want to pass a temporary for the @@ -XXX,XX +XXX,XX @@ static void handle_sys(DisasContext *s, uint32_t insn, bool isread, return; } else if ((ri->type & ARM_CP_SVE) && !sve_access_check(s)) { return; + } else if ((ri->type & ARM_CP_SME) && !sme_access_check(s)) { + return; } if ((tb_cflags(s->base.tb) & CF_USE_ICOUNT) && (ri->type & ARM_CP_IO)) { -- 2.25.1 From: Richard Henderson <richard.henderson@linaro.org> This cpreg is used to access two new bits of PSTATE that are not visible via any other mechanism. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20220620175235.60881-6-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/cpu.h | 6 ++++++ target/arm/helper.c | 13 +++++++++++++ 2 files changed, 19 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 CPUArchState { * nRW (also known as M[4]) is kept, inverted, in env->aarch64 * DAIF (exception masks) are kept in env->daif * BTYPE is kept in env->btype + * SM and ZA are kept in env->svcr * all other bits are stored in their correct places in env->pstate */ uint32_t pstate; @@ -XXX,XX +XXX,XX @@ typedef struct CPUArchState { uint32_t condexec_bits; /* IT bits. cpsr[15:10,26:25]. */ uint32_t btype; /* BTI branch type. spsr[11:10]. */ uint64_t daif; /* exception masks, in the bits they are in PSTATE */ + uint64_t svcr; /* PSTATE.{SM,ZA} in the bits they are in SVCR */ uint64_t elr_el[4]; /* AArch64 exception link regs */ uint64_t sp_el[4]; /* AArch64 banked stack pointers */ @@ -XXX,XX +XXX,XX @@ FIELD(CPTR_EL3, TCPAC, 31, 1) #define PSTATE_MODE_EL1t 4 #define PSTATE_MODE_EL0t 0 +/* PSTATE bits that are accessed via SVCR and not stored in SPSR_ELx. */ +FIELD(SVCR, SM, 0, 1) +FIELD(SVCR, ZA, 1, 1) + /* Write a new value to v7m.exception, thus transitioning into or out * of Handler mode; this may result in a change of active stack pointer. */ 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 CPAccessResult access_tpidr2(CPUARMState *env, const ARMCPRegInfo *ri, return CP_ACCESS_OK; } +static void svcr_write(CPUARMState *env, const ARMCPRegInfo *ri, + uint64_t value) +{ + value &= R_SVCR_SM_MASK | R_SVCR_ZA_MASK; + /* TODO: Side effects. */ + env->svcr = value; +} + static const ARMCPRegInfo sme_reginfo[] = { { .name = "TPIDR2_EL0", .state = ARM_CP_STATE_AA64, .opc0 = 3, .opc1 = 3, .crn = 13, .crm = 0, .opc2 = 5, .access = PL0_RW, .accessfn = access_tpidr2, .fieldoffset = offsetof(CPUARMState, cp15.tpidr2_el0) }, + { .name = "SVCR", .state = ARM_CP_STATE_AA64, + .opc0 = 3, .opc1 = 3, .crn = 4, .crm = 2, .opc2 = 2, + .access = PL0_RW, .type = ARM_CP_SME, + .fieldoffset = offsetof(CPUARMState, svcr), + .writefn = svcr_write, .raw_writefn = raw_write }, }; #endif /* TARGET_AARCH64 */ -- 2.25.1 From: Richard Henderson <richard.henderson@linaro.org> These cpregs control the streaming vector length and whether the full a64 instruction set is allowed while in streaming mode. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20220620175235.60881-7-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/cpu.h | 8 ++++++-- target/arm/helper.c | 41 +++++++++++++++++++++++++++++++++++++++++ 2 files changed, 47 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 @@ typedef struct CPUArchState { float_status standard_fp_status; float_status standard_fp_status_f16; - /* ZCR_EL[1-3] */ - uint64_t zcr_el[4]; + uint64_t zcr_el[4]; /* ZCR_EL[1-3] */ + uint64_t smcr_el[4]; /* SMCR_EL[1-3] */ } vfp; uint64_t exclusive_addr; uint64_t exclusive_val; @@ -XXX,XX +XXX,XX @@ FIELD(CPTR_EL3, TCPAC, 31, 1) FIELD(SVCR, SM, 0, 1) FIELD(SVCR, ZA, 1, 1) +/* Fields for SMCR_ELx. */ +FIELD(SMCR, LEN, 0, 4) +FIELD(SMCR, FA64, 31, 1) + /* Write a new value to v7m.exception, thus transitioning into or out * of Handler mode; this may result in a change of active stack pointer. */ 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, 1, 2, 6), K(3, 4, 1, 2, 6), K(3, 5, 1, 2, 6), + "SMCR_EL1", "SMCR_EL2", "SMCR_EL12", isar_feature_aa64_sme }, { 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 }, @@ -XXX,XX +XXX,XX @@ static void svcr_write(CPUARMState *env, const ARMCPRegInfo *ri, env->svcr = value; } +static void smcr_write(CPUARMState *env, const ARMCPRegInfo *ri, + uint64_t value) +{ + int cur_el = arm_current_el(env); + int old_len = sve_vqm1_for_el(env, cur_el); + int new_len; + + QEMU_BUILD_BUG_ON(ARM_MAX_VQ > R_SMCR_LEN_MASK + 1); + value &= R_SMCR_LEN_MASK | R_SMCR_FA64_MASK; + raw_write(env, ri, value); + + /* + * Note that it is CONSTRAINED UNPREDICTABLE what happens to ZA storage + * when SVL is widened (old values kept, or zeros). Choose to keep the + * current values for simplicity. But for QEMU internals, we must still + * apply the narrower SVL to the Zregs and Pregs -- see the comment + * above aarch64_sve_narrow_vq. + */ + new_len = sve_vqm1_for_el(env, cur_el); + if (new_len < old_len) { + aarch64_sve_narrow_vq(env, new_len + 1); + } +} + static const ARMCPRegInfo sme_reginfo[] = { { .name = "TPIDR2_EL0", .state = ARM_CP_STATE_AA64, .opc0 = 3, .opc1 = 3, .crn = 13, .crm = 0, .opc2 = 5, @@ -XXX,XX +XXX,XX @@ static const ARMCPRegInfo sme_reginfo[] = { .access = PL0_RW, .type = ARM_CP_SME, .fieldoffset = offsetof(CPUARMState, svcr), .writefn = svcr_write, .raw_writefn = raw_write }, + { .name = "SMCR_EL1", .state = ARM_CP_STATE_AA64, + .opc0 = 3, .opc1 = 0, .crn = 1, .crm = 2, .opc2 = 6, + .access = PL1_RW, .type = ARM_CP_SME, + .fieldoffset = offsetof(CPUARMState, vfp.smcr_el[1]), + .writefn = smcr_write, .raw_writefn = raw_write }, + { .name = "SMCR_EL2", .state = ARM_CP_STATE_AA64, + .opc0 = 3, .opc1 = 4, .crn = 1, .crm = 2, .opc2 = 6, + .access = PL2_RW, .type = ARM_CP_SME, + .fieldoffset = offsetof(CPUARMState, vfp.smcr_el[2]), + .writefn = smcr_write, .raw_writefn = raw_write }, + { .name = "SMCR_EL3", .state = ARM_CP_STATE_AA64, + .opc0 = 3, .opc1 = 6, .crn = 1, .crm = 2, .opc2 = 6, + .access = PL3_RW, .type = ARM_CP_SME, + .fieldoffset = offsetof(CPUARMState, vfp.smcr_el[3]), + .writefn = smcr_write, .raw_writefn = raw_write }, }; #endif /* TARGET_AARCH64 */ -- 2.25.1 From: Richard Henderson <richard.henderson@linaro.org> Implement the streaming mode identification register, and the two streaming priority registers. For QEMU, they are all RES0. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20220620175235.60881-8-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/helper.c | 33 +++++++++++++++++++++++++++++++++ 1 file changed, 33 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 CPAccessResult access_tpidr2(CPUARMState *env, const ARMCPRegInfo *ri, return CP_ACCESS_OK; } +static CPAccessResult access_esm(CPUARMState *env, const ARMCPRegInfo *ri, + bool isread) +{ + /* TODO: FEAT_FGT for SMPRI_EL1 but not SMPRIMAP_EL2 */ + if (arm_current_el(env) < 3 + && arm_feature(env, ARM_FEATURE_EL3) + && !FIELD_EX64(env->cp15.cptr_el[3], CPTR_EL3, ESM)) { + return CP_ACCESS_TRAP_EL3; + } + return CP_ACCESS_OK; +} + static void svcr_write(CPUARMState *env, const ARMCPRegInfo *ri, uint64_t value) { @@ -XXX,XX +XXX,XX @@ static const ARMCPRegInfo sme_reginfo[] = { .access = PL3_RW, .type = ARM_CP_SME, .fieldoffset = offsetof(CPUARMState, vfp.smcr_el[3]), .writefn = smcr_write, .raw_writefn = raw_write }, + { .name = "SMIDR_EL1", .state = ARM_CP_STATE_AA64, + .opc0 = 3, .opc1 = 1, .crn = 0, .crm = 0, .opc2 = 6, + .access = PL1_R, .accessfn = access_aa64_tid1, + /* + * IMPLEMENTOR = 0 (software) + * REVISION = 0 (implementation defined) + * SMPS = 0 (no streaming execution priority in QEMU) + * AFFINITY = 0 (streaming sve mode not shared with other PEs) + */ + .type = ARM_CP_CONST, .resetvalue = 0, }, + /* + * Because SMIDR_EL1.SMPS is 0, SMPRI_EL1 and SMPRIMAP_EL2 are RES 0. + */ + { .name = "SMPRI_EL1", .state = ARM_CP_STATE_AA64, + .opc0 = 3, .opc1 = 0, .crn = 1, .crm = 2, .opc2 = 4, + .access = PL1_RW, .accessfn = access_esm, + .type = ARM_CP_CONST, .resetvalue = 0 }, + { .name = "SMPRIMAP_EL2", .state = ARM_CP_STATE_AA64, + .opc0 = 3, .opc1 = 4, .crn = 1, .crm = 2, .opc2 = 5, + .access = PL2_RW, .accessfn = access_esm, + .type = ARM_CP_CONST, .resetvalue = 0 }, }; #endif /* TARGET_AARCH64 */ -- 2.25.1 From: Richard Henderson <richard.henderson@linaro.org> These are required to determine if various insns are allowed to issue. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20220620175235.60881-9-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/cpu.h | 2 ++ target/arm/translate.h | 4 ++++ target/arm/helper.c | 4 ++++ target/arm/translate-a64.c | 2 ++ 4 files changed, 12 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 @@ FIELD(TBFLAG_A64, TCMA, 16, 2) FIELD(TBFLAG_A64, MTE_ACTIVE, 18, 1) FIELD(TBFLAG_A64, MTE0_ACTIVE, 19, 1) FIELD(TBFLAG_A64, SMEEXC_EL, 20, 2) +FIELD(TBFLAG_A64, PSTATE_SM, 22, 1) +FIELD(TBFLAG_A64, PSTATE_ZA, 23, 1) /* * Helpers for using the above. 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 { bool align_mem; /* True if PSTATE.IL is set */ bool pstate_il; + /* True if PSTATE.SM is set. */ + bool pstate_sm; + /* True if PSTATE.ZA is set. */ + bool pstate_za; /* True if MVE insns are definitely not predicated by VPR or LTPSIZE */ bool mve_no_pred; /* 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 CPUARMTBFlags rebuild_hflags_a64(CPUARMState *env, int el, int fp_el, } if (cpu_isar_feature(aa64_sme, env_archcpu(env))) { DP_TBFLAG_A64(flags, SMEEXC_EL, sme_exception_el(env, el)); + if (FIELD_EX64(env->svcr, SVCR, SM)) { + DP_TBFLAG_A64(flags, PSTATE_SM, 1); + } + DP_TBFLAG_A64(flags, PSTATE_ZA, FIELD_EX64(env->svcr, SVCR, ZA)); } sctlr = regime_sctlr(env, stage1); 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->ata = EX_TBFLAG_A64(tb_flags, ATA); dc->mte_active[0] = EX_TBFLAG_A64(tb_flags, MTE_ACTIVE); dc->mte_active[1] = EX_TBFLAG_A64(tb_flags, MTE0_ACTIVE); + dc->pstate_sm = EX_TBFLAG_A64(tb_flags, PSTATE_SM); + dc->pstate_za = EX_TBFLAG_A64(tb_flags, PSTATE_ZA); dc->vec_len = 0; dc->vec_stride = 0; dc->cp_regs = arm_cpu->cp_regs; -- 2.25.1 From: Richard Henderson <richard.henderson@linaro.org> Place this late in the resettable section of the structure, to keep the most common element offsets from being > 64k. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20220620175235.60881-10-richard.henderson@linaro.org [PMM: expanded comment on zarray[] format] Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/cpu.h | 22 ++++++++++++++++++++++ target/arm/machine.c | 34 ++++++++++++++++++++++++++++++++++ 2 files changed, 56 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 CPUArchState { } keys; uint64_t scxtnum_el[4]; + + /* + * SME ZA storage -- 256 x 256 byte array, with bytes in host word order, + * as we do with vfp.zregs[]. This corresponds to the architectural ZA + * array, where ZA[N] is in the least-significant bytes of env->zarray[N]. + * When SVL is less than the architectural maximum, the accessible + * storage is restricted, such that if the SVL is X bytes the guest can + * see only the bottom X elements of zarray[], and only the least + * significant X bytes of each element of the array. (In other words, + * the observable part is always square.) + * + * The ZA storage can also be considered as a set of square tiles of + * elements of different sizes. The mapping from tiles to the ZA array + * is architecturally defined, such that for tiles of elements of esz + * bytes, the Nth row (or "horizontal slice") of tile T is in + * ZA[T + N * esz]. Note that this means that each tile is not contiguous + * in the ZA storage, because its rows are striped through the ZA array. + * + * Because this is so large, keep this toward the end of the reset area, + * to keep the offsets into the rest of the structure smaller. + */ + ARMVectorReg zarray[ARM_MAX_VQ * 16]; #endif #if defined(CONFIG_USER_ONLY) diff --git a/target/arm/machine.c b/target/arm/machine.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/machine.c +++ b/target/arm/machine.c @@ -XXX,XX +XXX,XX @@ static const VMStateDescription vmstate_sve = { VMSTATE_END_OF_LIST() } }; + +static const VMStateDescription vmstate_vreg = { + .name = "vreg", + .version_id = 1, + .minimum_version_id = 1, + .fields = (VMStateField[]) { + VMSTATE_UINT64_ARRAY(d, ARMVectorReg, ARM_MAX_VQ * 2), + VMSTATE_END_OF_LIST() + } +}; + +static bool za_needed(void *opaque) +{ + ARMCPU *cpu = opaque; + + /* + * When ZA storage is disabled, its contents are discarded. + * It will be zeroed when ZA storage is re-enabled. + */ + return FIELD_EX64(cpu->env.svcr, SVCR, ZA); +} + +static const VMStateDescription vmstate_za = { + .name = "cpu/sme", + .version_id = 1, + .minimum_version_id = 1, + .needed = za_needed, + .fields = (VMStateField[]) { + VMSTATE_STRUCT_ARRAY(env.zarray, ARMCPU, ARM_MAX_VQ * 16, 0, + vmstate_vreg, ARMVectorReg), + VMSTATE_END_OF_LIST() + } +}; #endif /* AARCH64 */ static bool serror_needed(void *opaque) @@ -XXX,XX +XXX,XX @@ const VMStateDescription vmstate_arm_cpu = { &vmstate_m_security, #ifdef TARGET_AARCH64 &vmstate_sve, + &vmstate_za, #endif &vmstate_serror, &vmstate_irq_line_state, -- 2.25.1 From: Richard Henderson <richard.henderson@linaro.org> These two instructions are aliases of MSR (immediate). Use the two helpers to properly implement svcr_write. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20220620175235.60881-11-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/cpu.h | 1 + target/arm/helper-sme.h | 21 +++++++++++++ target/arm/helper.h | 1 + target/arm/helper.c | 6 ++-- target/arm/sme_helper.c | 61 ++++++++++++++++++++++++++++++++++++++ target/arm/translate-a64.c | 24 +++++++++++++++ target/arm/meson.build | 1 + 7 files changed, 112 insertions(+), 3 deletions(-) create mode 100644 target/arm/helper-sme.h create mode 100644 target/arm/sme_helper.c 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 @@ void aarch64_sve_change_el(CPUARMState *env, int old_el, int new_el, bool el0_a64); void aarch64_add_sve_properties(Object *obj); void aarch64_add_pauth_properties(Object *obj); +void arm_reset_sve_state(CPUARMState *env); /* * SVE registers are encoded in KVM's memory in an endianness-invariant format. diff --git a/target/arm/helper-sme.h b/target/arm/helper-sme.h new file mode 100644 index XXXXXXX..XXXXXXX --- /dev/null +++ b/target/arm/helper-sme.h @@ -XXX,XX +XXX,XX @@ +/* + * AArch64 SME specific helper definitions + * + * Copyright (c) 2022 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/>. + */ + +DEF_HELPER_FLAGS_2(set_pstate_sm, TCG_CALL_NO_RWG, void, env, i32) +DEF_HELPER_FLAGS_2(set_pstate_za, TCG_CALL_NO_RWG, void, env, i32) 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_6(gvec_bfmlal_idx, TCG_CALL_NO_RWG, #ifdef TARGET_AARCH64 #include "helper-a64.h" #include "helper-sve.h" +#include "helper-sme.h" #endif #include "helper-mve.h" 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 CPAccessResult access_esm(CPUARMState *env, const ARMCPRegInfo *ri, static void svcr_write(CPUARMState *env, const ARMCPRegInfo *ri, uint64_t value) { - value &= R_SVCR_SM_MASK | R_SVCR_ZA_MASK; - /* TODO: Side effects. */ - env->svcr = value; + helper_set_pstate_sm(env, FIELD_EX64(value, SVCR, SM)); + helper_set_pstate_za(env, FIELD_EX64(value, SVCR, ZA)); + arm_rebuild_hflags(env); } static void smcr_write(CPUARMState *env, const ARMCPRegInfo *ri, diff --git a/target/arm/sme_helper.c b/target/arm/sme_helper.c new file mode 100644 index XXXXXXX..XXXXXXX --- /dev/null +++ b/target/arm/sme_helper.c @@ -XXX,XX +XXX,XX @@ +/* + * ARM SME Operations + * + * Copyright (c) 2022 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/helper-proto.h" + +/* ResetSVEState */ +void arm_reset_sve_state(CPUARMState *env) +{ + memset(env->vfp.zregs, 0, sizeof(env->vfp.zregs)); + /* Recall that FFR is stored as pregs[16]. */ + memset(env->vfp.pregs, 0, sizeof(env->vfp.pregs)); + vfp_set_fpcr(env, 0x0800009f); +} + +void helper_set_pstate_sm(CPUARMState *env, uint32_t i) +{ + if (i == FIELD_EX64(env->svcr, SVCR, SM)) { + return; + } + env->svcr ^= R_SVCR_SM_MASK; + arm_reset_sve_state(env); +} + +void helper_set_pstate_za(CPUARMState *env, uint32_t i) +{ + if (i == FIELD_EX64(env->svcr, SVCR, ZA)) { + return; + } + env->svcr ^= R_SVCR_ZA_MASK; + + /* + * ResetSMEState. + * + * SetPSTATE_ZA zeros on enable and disable. We can zero this only + * on enable: while disabled, the storage is inaccessible and the + * value does not matter. We're not saving the storage in vmstate + * when disabled either. + */ + if (i) { + memset(env->zarray, 0, sizeof(env->zarray)); + } +} 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, } break; + case 0x1b: /* SVCR* */ + if (!dc_isar_feature(aa64_sme, s) || crm < 2 || crm > 7) { + goto do_unallocated; + } + if (sme_access_check(s)) { + bool i = crm & 1; + bool changed = false; + + if ((crm & 2) && i != s->pstate_sm) { + gen_helper_set_pstate_sm(cpu_env, tcg_constant_i32(i)); + changed = true; + } + if ((crm & 4) && i != s->pstate_za) { + gen_helper_set_pstate_za(cpu_env, tcg_constant_i32(i)); + changed = true; + } + if (changed) { + gen_rebuild_hflags(s); + } else { + s->base.is_jmp = DISAS_NEXT; + } + } + break; + default: do_unallocated: unallocated_encoding(s); diff --git a/target/arm/meson.build b/target/arm/meson.build index XXXXXXX..XXXXXXX 100644 --- a/target/arm/meson.build +++ b/target/arm/meson.build @@ -XXX,XX +XXX,XX @@ arm_ss.add(when: 'TARGET_AARCH64', if_true: files( 'mte_helper.c', 'pauth_helper.c', 'sve_helper.c', + 'sme_helper.c', 'translate-a64.c', 'translate-sve.c', )) -- 2.25.1 From: Richard Henderson <richard.henderson@linaro.org> Keep all of the error messages together. This does mean that when setting many sve length properties we'll only generate one error, but we only really need one. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20220620175235.60881-12-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/cpu64.c | 15 +++++++-------- 1 file changed, 7 insertions(+), 8 deletions(-) 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 @@ void arm_cpu_sve_finalize(ARMCPU *cpu, Error **errp) "using only sve<N> properties.\n"); } else { error_setg(errp, "cannot enable sve%d", vq * 128); - error_append_hint(errp, "This CPU does not support " - "the vector length %d-bits.\n", vq * 128); + if (vq_supported) { + error_append_hint(errp, "This CPU does not support " + "the vector length %d-bits.\n", vq * 128); + } else { + error_append_hint(errp, "SVE not supported by KVM " + "on this host\n"); + } } return; } else { @@ -XXX,XX +XXX,XX @@ static void cpu_arm_set_sve_vq(Object *obj, Visitor *v, const char *name, return; } - if (value && kvm_enabled() && !kvm_arm_sve_supported()) { - error_setg(errp, "cannot enable %s", name); - error_append_hint(errp, "SVE not supported by KVM on this host\n"); - return; - } - cpu->sve_vq_map = deposit32(cpu->sve_vq_map, vq - 1, 1, value); cpu->sve_vq_init |= 1 << (vq - 1); } -- 2.25.1 From: Richard Henderson <richard.henderson@linaro.org> Pull the three sve_vq_* values into a structure. This will be reused for SME. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20220620175235.60881-13-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/cpu.h | 29 ++++++++++++++--------------- target/arm/cpu64.c | 22 +++++++++++----------- target/arm/helper.c | 2 +- target/arm/kvm64.c | 2 +- 4 files changed, 27 insertions(+), 28 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 enum ARMPSCIState { typedef struct ARMISARegisters ARMISARegisters; +/* + * In map, each set bit is a supported vector length of (bit-number + 1) * 16 + * bytes, i.e. each bit number + 1 is the vector length in quadwords. + * + * While processing properties during initialization, corresponding init bits + * are set for bits in sve_vq_map that have been set by properties. + * + * Bits set in supported represent valid vector lengths for the CPU type. + */ +typedef struct { + uint32_t map, init, supported; +} ARMVQMap; + /** * ARMCPU: * @env: #CPUARMState @@ -XXX,XX +XXX,XX @@ struct ArchCPU { uint32_t sve_default_vq; #endif - /* - * In sve_vq_map each set bit is a supported vector length of - * (bit-number + 1) * 16 bytes, i.e. each bit number + 1 is the vector - * length in quadwords. - * - * While processing properties during initialization, corresponding - * sve_vq_init bits are set for bits in sve_vq_map that have been - * set by properties. - * - * Bits set in sve_vq_supported represent valid vector lengths for - * the CPU type. - */ - uint32_t sve_vq_map; - uint32_t sve_vq_init; - uint32_t sve_vq_supported; + ARMVQMap sve_vq; /* Generic timer counter frequency, in Hz */ uint64_t gt_cntfrq_hz; 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 @@ void arm_cpu_sve_finalize(ARMCPU *cpu, Error **errp) * any of the above. Finally, if SVE is not disabled, then at least one * vector length must be enabled. */ - uint32_t vq_map = cpu->sve_vq_map; - uint32_t vq_init = cpu->sve_vq_init; + uint32_t vq_map = cpu->sve_vq.map; + uint32_t vq_init = cpu->sve_vq.init; uint32_t vq_supported; uint32_t vq_mask = 0; uint32_t tmp, vq, max_vq = 0; @@ -XXX,XX +XXX,XX @@ void arm_cpu_sve_finalize(ARMCPU *cpu, Error **errp) */ if (kvm_enabled()) { if (kvm_arm_sve_supported()) { - cpu->sve_vq_supported = kvm_arm_sve_get_vls(CPU(cpu)); - vq_supported = cpu->sve_vq_supported; + cpu->sve_vq.supported = kvm_arm_sve_get_vls(CPU(cpu)); + vq_supported = cpu->sve_vq.supported; } else { assert(!cpu_isar_feature(aa64_sve, cpu)); vq_supported = 0; } } else { - vq_supported = cpu->sve_vq_supported; + vq_supported = cpu->sve_vq.supported; } /* @@ -XXX,XX +XXX,XX @@ void arm_cpu_sve_finalize(ARMCPU *cpu, Error **errp) /* From now on sve_max_vq is the actual maximum supported length. */ cpu->sve_max_vq = max_vq; - cpu->sve_vq_map = vq_map; + cpu->sve_vq.map = vq_map; } static void cpu_max_get_sve_max_vq(Object *obj, Visitor *v, const char *name, @@ -XXX,XX +XXX,XX @@ static void cpu_arm_get_sve_vq(Object *obj, Visitor *v, const char *name, if (!cpu_isar_feature(aa64_sve, cpu)) { value = false; } else { - value = extract32(cpu->sve_vq_map, vq - 1, 1); + value = extract32(cpu->sve_vq.map, vq - 1, 1); } visit_type_bool(v, name, &value, errp); } @@ -XXX,XX +XXX,XX @@ static void cpu_arm_set_sve_vq(Object *obj, Visitor *v, const char *name, return; } - cpu->sve_vq_map = deposit32(cpu->sve_vq_map, vq - 1, 1, value); - cpu->sve_vq_init |= 1 << (vq - 1); + cpu->sve_vq.map = deposit32(cpu->sve_vq.map, vq - 1, 1, value); + cpu->sve_vq.init |= 1 << (vq - 1); } static bool cpu_arm_get_sve(Object *obj, Error **errp) @@ -XXX,XX +XXX,XX @@ static void aarch64_max_initfn(Object *obj) cpu->dcz_blocksize = 7; /* 512 bytes */ #endif - cpu->sve_vq_supported = MAKE_64BIT_MASK(0, ARM_MAX_VQ); + cpu->sve_vq.supported = MAKE_64BIT_MASK(0, ARM_MAX_VQ); aarch64_add_pauth_properties(obj); aarch64_add_sve_properties(obj); @@ -XXX,XX +XXX,XX @@ static void aarch64_a64fx_initfn(Object *obj) /* The A64FX supports only 128, 256 and 512 bit vector lengths */ aarch64_add_sve_properties(obj); - cpu->sve_vq_supported = (1 << 0) /* 128bit */ + cpu->sve_vq.supported = (1 << 0) /* 128bit */ | (1 << 1) /* 256bit */ | (1 << 3); /* 512bit */ 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 @@ uint32_t sve_vqm1_for_el(CPUARMState *env, int el) len = MIN(len, 0xf & (uint32_t)env->vfp.zcr_el[3]); } - len = 31 - clz32(cpu->sve_vq_map & MAKE_64BIT_MASK(0, len + 1)); + len = 31 - clz32(cpu->sve_vq.map & MAKE_64BIT_MASK(0, len + 1)); return len; } diff --git a/target/arm/kvm64.c b/target/arm/kvm64.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/kvm64.c +++ b/target/arm/kvm64.c @@ -XXX,XX +XXX,XX @@ uint32_t kvm_arm_sve_get_vls(CPUState *cs) static int kvm_arm_sve_set_vls(CPUState *cs) { ARMCPU *cpu = ARM_CPU(cs); - uint64_t vls[KVM_ARM64_SVE_VLS_WORDS] = { cpu->sve_vq_map }; + uint64_t vls[KVM_ARM64_SVE_VLS_WORDS] = { cpu->sve_vq.map }; struct kvm_one_reg reg = { .id = KVM_REG_ARM64_SVE_VLS, .addr = (uint64_t)&vls[0], -- 2.25.1 From: Richard Henderson <richard.henderson@linaro.org> Rename from cpu_arm_{get,set}_sve_vq, and take the ARMVQMap as the opaque parameter. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20220620175235.60881-14-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/cpu64.c | 29 +++++++++++++++-------------- 1 file changed, 15 insertions(+), 14 deletions(-) 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 cpu_max_set_sve_max_vq(Object *obj, Visitor *v, const char *name, } /* - * Note that cpu_arm_get/set_sve_vq cannot use the simpler - * object_property_add_bool interface because they make use - * of the contents of "name" to determine which bit on which - * to operate. + * Note that cpu_arm_{get,set}_vq cannot use the simpler + * object_property_add_bool interface because they make use of the + * contents of "name" to determine which bit on which to operate. */ -static void cpu_arm_get_sve_vq(Object *obj, Visitor *v, const char *name, - void *opaque, Error **errp) +static void cpu_arm_get_vq(Object *obj, Visitor *v, const char *name, + void *opaque, Error **errp) { ARMCPU *cpu = ARM_CPU(obj); + ARMVQMap *vq_map = opaque; uint32_t vq = atoi(&name[3]) / 128; bool value; @@ -XXX,XX +XXX,XX @@ static void cpu_arm_get_sve_vq(Object *obj, Visitor *v, const char *name, if (!cpu_isar_feature(aa64_sve, cpu)) { value = false; } else { - value = extract32(cpu->sve_vq.map, vq - 1, 1); + value = extract32(vq_map->map, vq - 1, 1); } visit_type_bool(v, name, &value, errp); } -static void cpu_arm_set_sve_vq(Object *obj, Visitor *v, const char *name, - void *opaque, Error **errp) +static void cpu_arm_set_vq(Object *obj, Visitor *v, const char *name, + void *opaque, Error **errp) { - ARMCPU *cpu = ARM_CPU(obj); + ARMVQMap *vq_map = opaque; uint32_t vq = atoi(&name[3]) / 128; bool value; @@ -XXX,XX +XXX,XX @@ static void cpu_arm_set_sve_vq(Object *obj, Visitor *v, const char *name, return; } - cpu->sve_vq.map = deposit32(cpu->sve_vq.map, vq - 1, 1, value); - cpu->sve_vq.init |= 1 << (vq - 1); + vq_map->map = deposit32(vq_map->map, vq - 1, 1, value); + vq_map->init |= 1 << (vq - 1); } static bool cpu_arm_get_sve(Object *obj, Error **errp) @@ -XXX,XX +XXX,XX @@ static void cpu_arm_get_sve_default_vec_len(Object *obj, Visitor *v, void aarch64_add_sve_properties(Object *obj) { + ARMCPU *cpu = ARM_CPU(obj); uint32_t vq; object_property_add_bool(obj, "sve", cpu_arm_get_sve, cpu_arm_set_sve); @@ -XXX,XX +XXX,XX @@ void aarch64_add_sve_properties(Object *obj) for (vq = 1; vq <= ARM_MAX_VQ; ++vq) { char name[8]; sprintf(name, "sve%d", vq * 128); - object_property_add(obj, name, "bool", cpu_arm_get_sve_vq, - cpu_arm_set_sve_vq, NULL, NULL); + object_property_add(obj, name, "bool", cpu_arm_get_vq, + cpu_arm_set_vq, NULL, &cpu->sve_vq); } #ifdef CONFIG_USER_ONLY -- 2.25.1 From: Richard Henderson <richard.henderson@linaro.org> Rename from cpu_arm_{get,set}_sve_default_vec_len, and take the pointer to default_vq from opaque. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20220620175235.60881-15-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/cpu64.c | 27 ++++++++++++++------------- 1 file changed, 14 insertions(+), 13 deletions(-) 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 cpu_arm_set_sve(Object *obj, bool value, Error **errp) #ifdef CONFIG_USER_ONLY /* Mirror linux /proc/sys/abi/sve_default_vector_length. */ -static void cpu_arm_set_sve_default_vec_len(Object *obj, Visitor *v, - const char *name, void *opaque, - Error **errp) +static void cpu_arm_set_default_vec_len(Object *obj, Visitor *v, + const char *name, void *opaque, + Error **errp) { - ARMCPU *cpu = ARM_CPU(obj); + uint32_t *ptr_default_vq = opaque; int32_t default_len, default_vq, remainder; if (!visit_type_int32(v, name, &default_len, errp)) { @@ -XXX,XX +XXX,XX @@ static void cpu_arm_set_sve_default_vec_len(Object *obj, Visitor *v, /* Undocumented, but the kernel allows -1 to indicate "maximum". */ if (default_len == -1) { - cpu->sve_default_vq = ARM_MAX_VQ; + *ptr_default_vq = ARM_MAX_VQ; return; } @@ -XXX,XX +XXX,XX @@ static void cpu_arm_set_sve_default_vec_len(Object *obj, Visitor *v, return; } - cpu->sve_default_vq = default_vq; + *ptr_default_vq = default_vq; } -static void cpu_arm_get_sve_default_vec_len(Object *obj, Visitor *v, - const char *name, void *opaque, - Error **errp) +static void cpu_arm_get_default_vec_len(Object *obj, Visitor *v, + const char *name, void *opaque, + Error **errp) { - ARMCPU *cpu = ARM_CPU(obj); - int32_t value = cpu->sve_default_vq * 16; + uint32_t *ptr_default_vq = opaque; + int32_t value = *ptr_default_vq * 16; visit_type_int32(v, name, &value, errp); } @@ -XXX,XX +XXX,XX @@ void aarch64_add_sve_properties(Object *obj) #ifdef CONFIG_USER_ONLY /* Mirror linux /proc/sys/abi/sve_default_vector_length. */ object_property_add(obj, "sve-default-vector-length", "int32", - cpu_arm_get_sve_default_vec_len, - cpu_arm_set_sve_default_vec_len, NULL, NULL); + cpu_arm_get_default_vec_len, + cpu_arm_set_default_vec_len, NULL, + &cpu->sve_default_vq); #endif } -- 2.25.1 From: Richard Henderson <richard.henderson@linaro.org> Drop the aa32-only inline fallbacks, and just use a couple of ifdefs. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20220620175235.60881-16-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/cpu.h | 6 ------ target/arm/internals.h | 3 +++ target/arm/cpu.c | 2 ++ 3 files changed, 5 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 @@ typedef struct { #ifdef TARGET_AARCH64 # define ARM_MAX_VQ 16 -void arm_cpu_sve_finalize(ARMCPU *cpu, Error **errp); -void arm_cpu_pauth_finalize(ARMCPU *cpu, Error **errp); -void arm_cpu_lpa2_finalize(ARMCPU *cpu, Error **errp); #else # define ARM_MAX_VQ 1 -static inline void arm_cpu_sve_finalize(ARMCPU *cpu, Error **errp) { } -static inline void arm_cpu_pauth_finalize(ARMCPU *cpu, Error **errp) { } -static inline void arm_cpu_lpa2_finalize(ARMCPU *cpu, Error **errp) { } #endif typedef struct ARMVectorReg { 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 @@ int arm_gdb_get_svereg(CPUARMState *env, GByteArray *buf, int reg); int arm_gdb_set_svereg(CPUARMState *env, uint8_t *buf, int reg); int aarch64_fpu_gdb_get_reg(CPUARMState *env, GByteArray *buf, int reg); int aarch64_fpu_gdb_set_reg(CPUARMState *env, uint8_t *buf, int reg); +void arm_cpu_sve_finalize(ARMCPU *cpu, Error **errp); +void arm_cpu_pauth_finalize(ARMCPU *cpu, Error **errp); +void arm_cpu_lpa2_finalize(ARMCPU *cpu, Error **errp); #endif #ifdef CONFIG_USER_ONLY 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_finalize_features(ARMCPU *cpu, Error **errp) { Error *local_err = NULL; +#ifdef TARGET_AARCH64 if (arm_feature(&cpu->env, ARM_FEATURE_AARCH64)) { arm_cpu_sve_finalize(cpu, &local_err); if (local_err != NULL) { @@ -XXX,XX +XXX,XX @@ void arm_cpu_finalize_features(ARMCPU *cpu, Error **errp) return; } } +#endif if (kvm_enabled()) { kvm_arm_steal_time_finalize(cpu, &local_err); -- 2.25.1 From: Richard Henderson <richard.henderson@linaro.org> These functions are not used outside cpu64.c, so make them static. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20220620175235.60881-17-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/cpu.h | 3 --- target/arm/cpu64.c | 4 ++-- 2 files changed, 2 insertions(+), 5 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 @@ int aarch64_cpu_gdb_write_register(CPUState *cpu, uint8_t *buf, int reg); void aarch64_sve_narrow_vq(CPUARMState *env, unsigned vq); void aarch64_sve_change_el(CPUARMState *env, int old_el, int new_el, bool el0_a64); -void aarch64_add_sve_properties(Object *obj); -void aarch64_add_pauth_properties(Object *obj); void arm_reset_sve_state(CPUARMState *env); /* @@ -XXX,XX +XXX,XX @@ static inline void aarch64_sve_narrow_vq(CPUARMState *env, unsigned vq) { } static inline void aarch64_sve_change_el(CPUARMState *env, int o, int n, bool a) { } -static inline void aarch64_add_sve_properties(Object *obj) { } #endif void aarch64_sync_32_to_64(CPUARMState *env); 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 cpu_arm_get_default_vec_len(Object *obj, Visitor *v, } #endif -void aarch64_add_sve_properties(Object *obj) +static void aarch64_add_sve_properties(Object *obj) { ARMCPU *cpu = ARM_CPU(obj); uint32_t vq; @@ -XXX,XX +XXX,XX @@ static Property arm_cpu_pauth_property = static Property arm_cpu_pauth_impdef_property = DEFINE_PROP_BOOL("pauth-impdef", ARMCPU, prop_pauth_impdef, false); -void aarch64_add_pauth_properties(Object *obj) +static void aarch64_add_pauth_properties(Object *obj) { ARMCPU *cpu = ARM_CPU(obj); -- 2.25.1 From: Richard Henderson <richard.henderson@linaro.org> Mirror the properties for SVE. The main difference is that any arbitrary set of powers of 2 may be supported, and not the stricter constraints that apply to SVE. Include a property to control FEAT_SME_FA64, as failing to restrict the runtime to the proper subset of insns could be a major point for bugs. Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Message-id: 20220620175235.60881-18-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- docs/system/arm/cpu-features.rst | 56 +++++++++++++++ target/arm/cpu.h | 2 + target/arm/internals.h | 1 + target/arm/cpu.c | 14 +++- target/arm/cpu64.c | 114 +++++++++++++++++++++++++++++-- 5 files changed, 180 insertions(+), 7 deletions(-) diff --git a/docs/system/arm/cpu-features.rst b/docs/system/arm/cpu-features.rst index XXXXXXX..XXXXXXX 100644 --- a/docs/system/arm/cpu-features.rst +++ b/docs/system/arm/cpu-features.rst @@ -XXX,XX +XXX,XX @@ verbose command lines. However, the recommended way to select vector lengths is to explicitly enable each desired length. Therefore only example's (1), (4), and (6) exhibit recommended uses of the properties. +SME CPU Property Examples +------------------------- + + 1) Disable SME:: + + $ qemu-system-aarch64 -M virt -cpu max,sme=off + + 2) Implicitly enable all vector lengths for the ``max`` CPU type:: + + $ qemu-system-aarch64 -M virt -cpu max + + 3) Only enable the 256-bit vector length:: + + $ qemu-system-aarch64 -M virt -cpu max,sme256=on + + 3) Enable the 256-bit and 1024-bit vector lengths:: + + $ qemu-system-aarch64 -M virt -cpu max,sme256=on,sme1024=on + + 4) Disable the 512-bit vector length. This results in all the other + lengths supported by ``max`` defaulting to enabled + (128, 256, 1024 and 2048):: + + $ qemu-system-aarch64 -M virt -cpu max,sve512=off + SVE User-mode Default Vector Length Property -------------------------------------------- @@ -XXX,XX +XXX,XX @@ length supported by QEMU is 256. If this property is set to ``-1`` then the default vector length is set to the maximum possible length. + +SME CPU Properties +================== + +The SME CPU properties are much like the SVE properties: ``sme`` is +used to enable or disable the entire SME feature, and ``sme<N>`` is +used to enable or disable specific vector lengths. Finally, +``sme_fa64`` is used to enable or disable ``FEAT_SME_FA64``, which +allows execution of the "full a64" instruction set while Streaming +SVE mode is enabled. + +SME is not supported by KVM at this time. + +At least one vector length must be enabled when ``sme`` is enabled, +and all vector lengths must be powers of 2. The maximum vector +length supported by qemu is 2048 bits. Otherwise, there are no +additional constraints on the set of vector lengths supported by SME. + +SME User-mode Default Vector Length Property +-------------------------------------------- + +For qemu-aarch64, the cpu propery ``sme-default-vector-length=N`` is +defined to mirror the Linux kernel parameter file +``/proc/sys/abi/sme_default_vector_length``. The default length, ``N``, +is in units of bytes and must be between 16 and 8192. +If not specified, the default vector length is 32. + +As with ``sve-default-vector-length``, if the default length is larger +than the maximum vector length enabled, the actual vector length will +be reduced. If this property is set to ``-1`` then the default vector +length is set to the maximum possible length. 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 ArchCPU { #ifdef CONFIG_USER_ONLY /* Used to set the default vector length at process start. */ uint32_t sve_default_vq; + uint32_t sme_default_vq; #endif ARMVQMap sve_vq; + ARMVQMap sme_vq; /* Generic timer counter frequency, in Hz */ uint64_t gt_cntfrq_hz; 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 @@ int arm_gdb_set_svereg(CPUARMState *env, uint8_t *buf, int reg); int aarch64_fpu_gdb_get_reg(CPUARMState *env, GByteArray *buf, int reg); int aarch64_fpu_gdb_set_reg(CPUARMState *env, uint8_t *buf, int reg); void arm_cpu_sve_finalize(ARMCPU *cpu, Error **errp); +void arm_cpu_sme_finalize(ARMCPU *cpu, Error **errp); void arm_cpu_pauth_finalize(ARMCPU *cpu, Error **errp); void arm_cpu_lpa2_finalize(ARMCPU *cpu, Error **errp); #endif 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_initfn(Object *obj) #ifdef CONFIG_USER_ONLY # ifdef TARGET_AARCH64 /* - * The linux kernel defaults to 512-bit vectors, when sve is supported. - * See documentation for /proc/sys/abi/sve_default_vector_length, and - * our corresponding sve-default-vector-length cpu property. + * The linux kernel defaults to 512-bit for SVE, and 256-bit for SME. + * These values were chosen to fit within the default signal frame. + * See documentation for /proc/sys/abi/{sve,sme}_default_vector_length, + * and our corresponding cpu property. */ cpu->sve_default_vq = 4; + cpu->sme_default_vq = 2; # endif #else /* Our inbound IRQ and FIQ lines */ @@ -XXX,XX +XXX,XX @@ void arm_cpu_finalize_features(ARMCPU *cpu, Error **errp) return; } + arm_cpu_sme_finalize(cpu, &local_err); + if (local_err != NULL) { + error_propagate(errp, local_err); + return; + } + arm_cpu_pauth_finalize(cpu, &local_err); if (local_err != NULL) { error_propagate(errp, local_err); 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 cpu_arm_get_vq(Object *obj, Visitor *v, const char *name, ARMCPU *cpu = ARM_CPU(obj); ARMVQMap *vq_map = opaque; uint32_t vq = atoi(&name[3]) / 128; + bool sve = vq_map == &cpu->sve_vq; bool value; - /* All vector lengths are disabled when SVE is off. */ - if (!cpu_isar_feature(aa64_sve, cpu)) { + /* All vector lengths are disabled when feature is off. */ + if (sve + ? !cpu_isar_feature(aa64_sve, cpu) + : !cpu_isar_feature(aa64_sme, cpu)) { value = false; } else { value = extract32(vq_map->map, vq - 1, 1); @@ -XXX,XX +XXX,XX @@ static void cpu_arm_set_sve(Object *obj, bool value, Error **errp) cpu->isar.id_aa64pfr0 = t; } +void arm_cpu_sme_finalize(ARMCPU *cpu, Error **errp) +{ + uint32_t vq_map = cpu->sme_vq.map; + uint32_t vq_init = cpu->sme_vq.init; + uint32_t vq_supported = cpu->sme_vq.supported; + uint32_t vq; + + if (vq_map == 0) { + if (!cpu_isar_feature(aa64_sme, cpu)) { + cpu->isar.id_aa64smfr0 = 0; + return; + } + + /* TODO: KVM will require limitations via SMCR_EL2. */ + vq_map = vq_supported & ~vq_init; + + if (vq_map == 0) { + vq = ctz32(vq_supported) + 1; + error_setg(errp, "cannot disable sme%d", vq * 128); + error_append_hint(errp, "All SME vector lengths are disabled.\n"); + error_append_hint(errp, "With SME enabled, at least one " + "vector length must be enabled.\n"); + return; + } + } else { + if (!cpu_isar_feature(aa64_sme, cpu)) { + vq = 32 - clz32(vq_map); + error_setg(errp, "cannot enable sme%d", vq * 128); + error_append_hint(errp, "SME must be enabled to enable " + "vector lengths.\n"); + error_append_hint(errp, "Add sme=on to the CPU property list.\n"); + return; + } + /* TODO: KVM will require limitations via SMCR_EL2. */ + } + + cpu->sme_vq.map = vq_map; +} + +static bool cpu_arm_get_sme(Object *obj, Error **errp) +{ + ARMCPU *cpu = ARM_CPU(obj); + return cpu_isar_feature(aa64_sme, cpu); +} + +static void cpu_arm_set_sme(Object *obj, bool value, Error **errp) +{ + ARMCPU *cpu = ARM_CPU(obj); + uint64_t t; + + t = cpu->isar.id_aa64pfr1; + t = FIELD_DP64(t, ID_AA64PFR1, SME, value); + cpu->isar.id_aa64pfr1 = t; +} + +static bool cpu_arm_get_sme_fa64(Object *obj, Error **errp) +{ + ARMCPU *cpu = ARM_CPU(obj); + return cpu_isar_feature(aa64_sme, cpu) && + cpu_isar_feature(aa64_sme_fa64, cpu); +} + +static void cpu_arm_set_sme_fa64(Object *obj, bool value, Error **errp) +{ + ARMCPU *cpu = ARM_CPU(obj); + uint64_t t; + + t = cpu->isar.id_aa64smfr0; + t = FIELD_DP64(t, ID_AA64SMFR0, FA64, value); + cpu->isar.id_aa64smfr0 = t; +} + #ifdef CONFIG_USER_ONLY -/* Mirror linux /proc/sys/abi/sve_default_vector_length. */ +/* Mirror linux /proc/sys/abi/{sve,sme}_default_vector_length. */ static void cpu_arm_set_default_vec_len(Object *obj, Visitor *v, const char *name, void *opaque, Error **errp) @@ -XXX,XX +XXX,XX @@ static void cpu_arm_set_default_vec_len(Object *obj, Visitor *v, * and is the maximum architectural width of ZCR_ELx.LEN. */ if (remainder || default_vq < 1 || default_vq > 512) { - error_setg(errp, "cannot set sve-default-vector-length"); + ARMCPU *cpu = ARM_CPU(obj); + const char *which = + (ptr_default_vq == &cpu->sve_default_vq ? "sve" : "sme"); + + error_setg(errp, "cannot set %s-default-vector-length", which); if (remainder) { error_append_hint(errp, "Vector length not a multiple of 16\n"); } else if (default_vq < 1) { @@ -XXX,XX +XXX,XX @@ static void aarch64_add_sve_properties(Object *obj) #endif } +static void aarch64_add_sme_properties(Object *obj) +{ + ARMCPU *cpu = ARM_CPU(obj); + uint32_t vq; + + object_property_add_bool(obj, "sme", cpu_arm_get_sme, cpu_arm_set_sme); + object_property_add_bool(obj, "sme_fa64", cpu_arm_get_sme_fa64, + cpu_arm_set_sme_fa64); + + for (vq = 1; vq <= ARM_MAX_VQ; vq <<= 1) { + char name[8]; + sprintf(name, "sme%d", vq * 128); + object_property_add(obj, name, "bool", cpu_arm_get_vq, + cpu_arm_set_vq, NULL, &cpu->sme_vq); + } + +#ifdef CONFIG_USER_ONLY + /* Mirror linux /proc/sys/abi/sme_default_vector_length. */ + object_property_add(obj, "sme-default-vector-length", "int32", + cpu_arm_get_default_vec_len, + cpu_arm_set_default_vec_len, NULL, + &cpu->sme_default_vq); +#endif +} + void arm_cpu_pauth_finalize(ARMCPU *cpu, Error **errp) { int arch_val = 0, impdef_val = 0; @@ -XXX,XX +XXX,XX @@ static void aarch64_max_initfn(Object *obj) #endif cpu->sve_vq.supported = MAKE_64BIT_MASK(0, ARM_MAX_VQ); + cpu->sme_vq.supported = SVE_VQ_POW2_MAP; aarch64_add_pauth_properties(obj); aarch64_add_sve_properties(obj); + aarch64_add_sme_properties(obj); object_property_add(obj, "sve-max-vq", "uint32", cpu_max_get_sve_max_vq, cpu_max_set_sve_max_vq, NULL, NULL); qdev_property_add_static(DEVICE(obj), &arm_cpu_lpa2_property); -- 2.25.1 From: Richard Henderson <richard.henderson@linaro.org> When Streaming SVE mode is enabled, the size is taken from SMCR_ELx instead of ZCR_ELx. The format is shared, but the set of vector lengths is not. Further, Streaming SVE does not require any particular length to be supported. Adjust sve_vqm1_for_el to pass the current value of PSTATE.SM to the new function. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20220620175235.60881-19-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/cpu.h | 9 +++++++-- target/arm/helper.c | 32 +++++++++++++++++++++++++------- 2 files changed, 32 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 @@ int sve_exception_el(CPUARMState *env, int cur_el); int sme_exception_el(CPUARMState *env, int cur_el); /** - * sve_vqm1_for_el: + * sve_vqm1_for_el_sm: * @env: CPUARMState * @el: exception level + * @sm: streaming mode * - * Compute the current SVE vector length for @el, in units of + * Compute the current vector length for @el & @sm, in units of * Quadwords Minus 1 -- the same scale used for ZCR_ELx.LEN. + * If @sm, compute for SVL, otherwise NVL. */ +uint32_t sve_vqm1_for_el_sm(CPUARMState *env, int el, bool sm); + +/* Likewise, but using @sm = PSTATE.SM. */ uint32_t sve_vqm1_for_el(CPUARMState *env, int el); static inline bool is_a64(CPUARMState *env) 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 @@ int sme_exception_el(CPUARMState *env, int el) /* * Given that SVE is enabled, return the vector length for EL. */ -uint32_t sve_vqm1_for_el(CPUARMState *env, int el) +uint32_t sve_vqm1_for_el_sm(CPUARMState *env, int el, bool sm) { ARMCPU *cpu = env_archcpu(env); - uint32_t len = cpu->sve_max_vq - 1; + uint64_t *cr = env->vfp.zcr_el; + uint32_t map = cpu->sve_vq.map; + uint32_t len = ARM_MAX_VQ - 1; + + if (sm) { + cr = env->vfp.smcr_el; + map = cpu->sme_vq.map; + } if (el <= 1 && !el_is_in_host(env, el)) { - len = MIN(len, 0xf & (uint32_t)env->vfp.zcr_el[1]); + len = MIN(len, 0xf & (uint32_t)cr[1]); } if (el <= 2 && arm_feature(env, ARM_FEATURE_EL2)) { - len = MIN(len, 0xf & (uint32_t)env->vfp.zcr_el[2]); + len = MIN(len, 0xf & (uint32_t)cr[2]); } if (arm_feature(env, ARM_FEATURE_EL3)) { - len = MIN(len, 0xf & (uint32_t)env->vfp.zcr_el[3]); + len = MIN(len, 0xf & (uint32_t)cr[3]); } - len = 31 - clz32(cpu->sve_vq.map & MAKE_64BIT_MASK(0, len + 1)); - return len; + map &= MAKE_64BIT_MASK(0, len + 1); + if (map != 0) { + return 31 - clz32(map); + } + + /* Bit 0 is always set for Normal SVE -- not so for Streaming SVE. */ + assert(sm); + return ctz32(cpu->sme_vq.map); +} + +uint32_t sve_vqm1_for_el(CPUARMState *env, int el) +{ + return sve_vqm1_for_el_sm(env, el, FIELD_EX64(env->svcr, SVCR, SM)); } static void zcr_write(CPUARMState *env, const ARMCPRegInfo *ri, -- 2.25.1 From: Richard Henderson <richard.henderson@linaro.org> We need SVL separate from VL for RDSVL et al, as well as ZA storage loads and stores, which do not require PSTATE.SM. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20220620175235.60881-20-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/cpu.h | 12 ++++++++++++ target/arm/translate.h | 1 + target/arm/helper.c | 8 +++++++- target/arm/translate-a64.c | 1 + 4 files changed, 21 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 @@ FIELD(TBFLAG_A64, MTE0_ACTIVE, 19, 1) FIELD(TBFLAG_A64, SMEEXC_EL, 20, 2) FIELD(TBFLAG_A64, PSTATE_SM, 22, 1) FIELD(TBFLAG_A64, PSTATE_ZA, 23, 1) +FIELD(TBFLAG_A64, SVL, 24, 4) /* * Helpers for using the above. @@ -XXX,XX +XXX,XX @@ static inline int sve_vq(CPUARMState *env) return EX_TBFLAG_A64(env->hflags, VL) + 1; } +/** + * sme_vq + * @env: the cpu context + * + * Return the SVL cached within env->hflags, in units of quadwords. + */ +static inline int sme_vq(CPUARMState *env) +{ + return EX_TBFLAG_A64(env->hflags, SVL) + 1; +} + static inline bool bswap_code(bool sctlr_b) { #ifdef CONFIG_USER_ONLY 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 { int sve_excp_el; /* SVE exception EL or 0 if enabled */ int sme_excp_el; /* SME exception EL or 0 if enabled */ int vl; /* current vector length in bytes */ + int svl; /* current streaming vector length in bytes */ bool vfp_enabled; /* FP enabled via FPSCR.EN */ int vec_len; int vec_stride; 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 CPUARMTBFlags rebuild_hflags_a64(CPUARMState *env, int el, int fp_el, DP_TBFLAG_A64(flags, SVEEXC_EL, sve_el); } if (cpu_isar_feature(aa64_sme, env_archcpu(env))) { - DP_TBFLAG_A64(flags, SMEEXC_EL, sme_exception_el(env, el)); + int sme_el = sme_exception_el(env, el); + + DP_TBFLAG_A64(flags, SMEEXC_EL, sme_el); + if (sme_el == 0) { + /* Similarly, do not compute SVL if SME is disabled. */ + DP_TBFLAG_A64(flags, SVL, sve_vqm1_for_el_sm(env, el, true)); + } if (FIELD_EX64(env->svcr, SVCR, SM)) { DP_TBFLAG_A64(flags, PSTATE_SM, 1); } 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->sve_excp_el = EX_TBFLAG_A64(tb_flags, SVEEXC_EL); dc->sme_excp_el = EX_TBFLAG_A64(tb_flags, SMEEXC_EL); dc->vl = (EX_TBFLAG_A64(tb_flags, VL) + 1) * 16; + dc->svl = (EX_TBFLAG_A64(tb_flags, SVL) + 1) * 16; dc->pauth_active = EX_TBFLAG_A64(tb_flags, PAUTH_ACTIVE); dc->bt = EX_TBFLAG_A64(tb_flags, BT); dc->btype = EX_TBFLAG_A64(tb_flags, BTYPE); -- 2.25.1 From: Richard Henderson <richard.henderson@linaro.org> We will need these functions in translate-sme.c. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20220620175235.60881-21-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/translate-a64.h | 38 ++++++++++++++++++++++++++++++++++++++ target/arm/translate-sve.c | 36 ------------------------------------ 2 files changed, 38 insertions(+), 36 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 @@ static inline int vec_full_reg_size(DisasContext *s) return s->vl; } +/* + * Return the offset info CPUARMState of the predicate vector register Pn. + * Note for this purpose, FFR is P16. + */ +static inline int pred_full_reg_offset(DisasContext *s, int regno) +{ + return offsetof(CPUARMState, vfp.pregs[regno]); +} + +/* Return the byte size of the whole predicate register, VL / 64. */ +static inline int pred_full_reg_size(DisasContext *s) +{ + return s->vl >> 3; +} + +/* + * Round up the size of a register to a size allowed by + * the tcg vector infrastructure. Any operation which uses this + * size may assume that the bits above pred_full_reg_size are zero, + * and must leave them the same way. + * + * Note that this is not needed for the vector registers as they + * are always properly sized for tcg vectors. + */ +static inline int size_for_gvec(int size) +{ + if (size <= 8) { + return 8; + } else { + return QEMU_ALIGN_UP(size, 16); + } +} + +static inline int pred_gvec_reg_size(DisasContext *s) +{ + return size_for_gvec(pred_full_reg_size(s)); +} + bool disas_sve(DisasContext *, uint32_t); void gen_gvec_rax1(unsigned vece, uint32_t rd_ofs, uint32_t rn_ofs, 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 inline int msz_dtype(DisasContext *s, int msz) * Implement all of the translator functions referenced by the decoder. */ -/* Return the offset info CPUARMState of the predicate vector register Pn. - * Note for this purpose, FFR is P16. - */ -static inline int pred_full_reg_offset(DisasContext *s, int regno) -{ - return offsetof(CPUARMState, vfp.pregs[regno]); -} - -/* Return the byte size of the whole predicate register, VL / 64. */ -static inline int pred_full_reg_size(DisasContext *s) -{ - return s->vl >> 3; -} - -/* Round up the size of a register to a size allowed by - * the tcg vector infrastructure. Any operation which uses this - * size may assume that the bits above pred_full_reg_size are zero, - * and must leave them the same way. - * - * Note that this is not needed for the vector registers as they - * are always properly sized for tcg vectors. - */ -static int size_for_gvec(int size) -{ - if (size <= 8) { - return 8; - } else { - return QEMU_ALIGN_UP(size, 16); - } -} - -static int pred_gvec_reg_size(DisasContext *s) -{ - return size_for_gvec(pred_full_reg_size(s)); -} - /* Invoke an out-of-line helper on 2 Zregs. */ static bool gen_gvec_ool_zz(DisasContext *s, gen_helper_gvec_2 *fn, int rd, int rn, int data) -- 2.25.1 From: Richard Henderson <richard.henderson@linaro.org> Move the code from hw/arm/virt.c that is supposed to handle v7 into the one function. Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Reported-by: He Zhe <zhe.he@windriver.com> Message-id: 20220619001541.131672-2-richard.henderson@linaro.org Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- hw/arm/virt.c | 10 +--------- target/arm/ptw.c | 24 ++++++++++++++++-------- 2 files changed, 17 insertions(+), 17 deletions(-) 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 machvirt_init(MachineState *machine) cpuobj = object_new(possible_cpus->cpus[0].type); armcpu = ARM_CPU(cpuobj); - if (object_property_get_bool(cpuobj, "aarch64", NULL)) { - pa_bits = arm_pamax(armcpu); - } else if (arm_feature(&armcpu->env, ARM_FEATURE_LPAE)) { - /* v7 with LPAE */ - pa_bits = 40; - } else { - /* Anything else */ - pa_bits = 32; - } + pa_bits = arm_pamax(armcpu); object_unref(cpuobj); diff --git a/target/arm/ptw.c b/target/arm/ptw.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/ptw.c +++ b/target/arm/ptw.c @@ -XXX,XX +XXX,XX @@ static const uint8_t pamax_map[] = { /* The cpu-specific constant value of PAMax; also used by hw/arm/virt. */ unsigned int arm_pamax(ARMCPU *cpu) { - unsigned int parange = - FIELD_EX64(cpu->isar.id_aa64mmfr0, ID_AA64MMFR0, PARANGE); + if (arm_feature(&cpu->env, ARM_FEATURE_AARCH64)) { + unsigned int parange = + FIELD_EX64(cpu->isar.id_aa64mmfr0, ID_AA64MMFR0, PARANGE); - /* - * id_aa64mmfr0 is a read-only register so values outside of the - * supported mappings can be considered an implementation error. - */ - assert(parange < ARRAY_SIZE(pamax_map)); - return pamax_map[parange]; + /* + * id_aa64mmfr0 is a read-only register so values outside of the + * supported mappings can be considered an implementation error. + */ + assert(parange < ARRAY_SIZE(pamax_map)); + return pamax_map[parange]; + } + if (arm_feature(&cpu->env, ARM_FEATURE_LPAE)) { + /* v7 with LPAE */ + return 40; + } + /* Anything else */ + return 32; } /* -- 2.25.1 From: Richard Henderson <richard.henderson@linaro.org> In machvirt_init we create a cpu but do not fully initialize it. Thus the propagation of V7VE to LPAE has not been done, and we compute the wrong value for some v7 cpus, e.g. cortex-a15. Resolves: https://gitlab.com/qemu-project/qemu/-/issues/1078 Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Reported-by: He Zhe <zhe.he@windriver.com> Message-id: 20220619001541.131672-3-richard.henderson@linaro.org Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/ptw.c | 8 +++++++- 1 file changed, 7 insertions(+), 1 deletion(-) diff --git a/target/arm/ptw.c b/target/arm/ptw.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/ptw.c +++ b/target/arm/ptw.c @@ -XXX,XX +XXX,XX @@ unsigned int arm_pamax(ARMCPU *cpu) assert(parange < ARRAY_SIZE(pamax_map)); return pamax_map[parange]; } - if (arm_feature(&cpu->env, ARM_FEATURE_LPAE)) { + + /* + * In machvirt_init, we call arm_pamax on a cpu that is not fully + * initialized, so we can't rely on the propagation done in realize. + */ + if (arm_feature(&cpu->env, ARM_FEATURE_LPAE) || + arm_feature(&cpu->env, ARM_FEATURE_V7VE)) { /* v7 with LPAE */ return 40; } -- 2.25.1