Nuvoton new board models, and some more minor stuff. I also put in the deprecation patches for unicore32 and lm32. thanks -- PMM The following changes since commit a68694cd1f3e5448cca814ff39b871f9ebd71ed5: Merge remote-tracking branch 'remotes/philmd-gitlab/tags/edk2-next-20200914' into staging (2020-09-14 12:18:58 +0100) are available in the Git repository at: https://git.linaro.org/people/pmaydell/qemu-arm.git tags/pull-target-arm-20200914 for you to fetch changes up to dd44ae00fc5342ed99acb68ec3508f76a71d523a: MAINTAINERS: Add maintainer entry for Xilinx ZynqMP CAN controller (2020-09-14 14:27:08 +0100) ---------------------------------------------------------------- target-arm queue: * hw/misc/a9scu: Do not allow invalid CPU count * hw/misc/a9scu: Minor cleanups * hw/timer/armv7m_systick: assert that board code set system_clock_scale * decodetree: Improve identifier matching * target/arm: Clean up neon fp insn size field decode * target/arm: Remove KVM support for 32-bit Arm hosts * hw/arm/mps2: New board models mps2-an386, mps2-an500 * Deprecate Unicore32 port * Deprecate lm32 port * target/arm: Count PMU events when MDCR.SPME is set * hw/arm: versal-virt: Correct the tx/rx GEM clocks * New Nuvoton iBMC board models npcm750-evb, quanta-gsj * xlnx-zynqmp: implement ZynqMP CAN controllers ---------------------------------------------------------------- Aaron Lindsay (1): target/arm: Count PMU events when MDCR.SPME is set Edgar E. Iglesias (1): hw/arm: versal-virt: Correct the tx/rx GEM clocks Havard Skinnemoen (14): hw/misc: Add NPCM7xx System Global Control Registers device model hw/misc: Add NPCM7xx Clock Controller device model hw/timer: Add NPCM7xx Timer device model hw/arm: Add NPCM730 and NPCM750 SoC models hw/arm: Add two NPCM7xx-based machines roms: Add virtual Boot ROM for NPCM7xx SoCs hw/arm: Load -bios image as a boot ROM for npcm7xx hw/nvram: NPCM7xx OTP device model hw/mem: Stubbed out NPCM7xx Memory Controller model hw/ssi: NPCM7xx Flash Interface Unit device model hw/arm: Wire up BMC boot flash for npcm750-evb and quanta-gsj hw/arm/npcm7xx: add board setup stub for CPU and UART clocks docs/system: Add Nuvoton machine documentation tests/acceptance: console boot tests for quanta-gsj Peter Maydell (11): hw/timer/armv7m_systick: assert that board code set system_clock_scale target/arm: Convert Neon 3-same-fp size field to MO_* in decode target/arm: Convert Neon VCVT fp size field to MO_* in decode target/arm: Convert VCMLA, VCADD size field to MO_* in decode target/arm: Remove KVM support for 32-bit Arm hosts target/arm: Remove no-longer-reachable 32-bit KVM code hw/arm/mps2: New board model mps2-an386 hw/arm/mps2: New board model mps2-an500 docs/system/arm/mps2.rst: Make board list consistent Deprecate Unicore32 port Deprecate lm32 port Philippe Mathieu-Daudé (4): hw/misc/a9scu: Do not allow invalid CPU count hw/misc/a9scu: Simplify setting MemoryRegionOps::valid fields hw/misc/a9scu: Simplify setting MemoryRegionOps::impl fields hw/misc/a9scu: Report unimplemented accesses with qemu_log_mask(UNIMP) Richard Henderson (1): decodetree: Improve identifier matching Vikram Garhwal (4): hw/net/can: Introduce Xilinx ZynqMP CAN controller xlnx-zynqmp: Connect Xilinx ZynqMP CAN controllers tests/qtest: Introduce tests for Xilinx ZynqMP CAN controller MAINTAINERS: Add maintainer entry for Xilinx ZynqMP CAN controller docs/system/arm/mps2.rst | 20 +- docs/system/arm/nuvoton.rst | 92 +++ docs/system/deprecated.rst | 32 +- docs/system/target-arm.rst | 1 + configure | 2 +- default-configs/arm-softmmu.mak | 1 + include/hw/arm/npcm7xx.h | 112 +++ include/hw/arm/xlnx-zynqmp.h | 8 + include/hw/mem/npcm7xx_mc.h | 36 + include/hw/misc/npcm7xx_clk.h | 48 ++ include/hw/misc/npcm7xx_gcr.h | 43 ++ include/hw/net/xlnx-zynqmp-can.h | 78 +++ include/hw/nvram/npcm7xx_otp.h | 79 +++ include/hw/ssi/npcm7xx_fiu.h | 73 ++ include/hw/timer/npcm7xx_timer.h | 78 +++ target/arm/kvm-consts.h | 7 - target/arm/kvm_arm.h | 6 - target/arm/neon-dp.decode | 18 +- target/arm/neon-shared.decode | 18 +- tests/decode/succ_ident1.decode | 7 + hw/arm/mps2.c | 97 ++- hw/arm/npcm7xx.c | 532 +++++++++++++++ hw/arm/npcm7xx_boards.c | 197 ++++++ hw/arm/xlnx-versal-virt.c | 2 +- hw/arm/xlnx-zcu102.c | 20 + hw/arm/xlnx-zynqmp.c | 34 + hw/mem/npcm7xx_mc.c | 84 +++ hw/misc/a9scu.c | 59 +- hw/misc/npcm7xx_clk.c | 266 ++++++++ hw/misc/npcm7xx_gcr.c | 269 ++++++++ hw/net/can/xlnx-zynqmp-can.c | 1165 ++++++++++++++++++++++++++++++++ hw/nvram/npcm7xx_otp.c | 440 ++++++++++++ hw/ssi/npcm7xx_fiu.c | 572 ++++++++++++++++ hw/timer/armv7m_systick.c | 8 + hw/timer/npcm7xx_timer.c | 543 +++++++++++++++ target/arm/cpu.c | 101 ++- target/arm/helper.c | 2 +- target/arm/kvm.c | 7 - target/arm/kvm32.c | 595 ---------------- tests/qtest/xlnx-can-test.c | 359 ++++++++++ .gitmodules | 3 + MAINTAINERS | 18 + hw/arm/Kconfig | 9 + hw/arm/meson.build | 1 + hw/mem/meson.build | 1 + hw/misc/meson.build | 4 + hw/misc/trace-events | 8 + hw/net/can/meson.build | 1 + hw/nvram/meson.build | 1 + hw/ssi/meson.build | 1 + hw/ssi/trace-events | 11 + hw/timer/meson.build | 1 + hw/timer/trace-events | 5 + pc-bios/README | 6 + pc-bios/meson.build | 1 + pc-bios/npcm7xx_bootrom.bin | Bin 0 -> 768 bytes roms/Makefile | 7 + roms/vbootrom | 1 + scripts/decodetree.py | 46 +- target/arm/meson.build | 5 +- target/arm/translate-neon.c.inc | 42 +- tests/acceptance/boot_linux_console.py | 83 +++ tests/qtest/meson.build | 1 + 63 files changed, 5584 insertions(+), 783 deletions(-) create mode 100644 docs/system/arm/nuvoton.rst create mode 100644 include/hw/arm/npcm7xx.h create mode 100644 include/hw/mem/npcm7xx_mc.h create mode 100644 include/hw/misc/npcm7xx_clk.h create mode 100644 include/hw/misc/npcm7xx_gcr.h create mode 100644 include/hw/net/xlnx-zynqmp-can.h create mode 100644 include/hw/nvram/npcm7xx_otp.h create mode 100644 include/hw/ssi/npcm7xx_fiu.h create mode 100644 include/hw/timer/npcm7xx_timer.h create mode 100644 tests/decode/succ_ident1.decode create mode 100644 hw/arm/npcm7xx.c create mode 100644 hw/arm/npcm7xx_boards.c create mode 100644 hw/mem/npcm7xx_mc.c create mode 100644 hw/misc/npcm7xx_clk.c create mode 100644 hw/misc/npcm7xx_gcr.c create mode 100644 hw/net/can/xlnx-zynqmp-can.c create mode 100644 hw/nvram/npcm7xx_otp.c create mode 100644 hw/ssi/npcm7xx_fiu.c create mode 100644 hw/timer/npcm7xx_timer.c delete mode 100644 target/arm/kvm32.c create mode 100644 tests/qtest/xlnx-can-test.c create mode 100644 pc-bios/npcm7xx_bootrom.bin create mode 160000 roms/vbootrom From: Philippe Mathieu-Daudé <f4bug@amsat.org> Per the datasheet (DDI0407 r2p0): "The SCU connects one to four Cortex-A9 processors to the memory system through the AXI interfaces." Change the instance_init() handler to a device_realize() one so we can verify the property is in range, and return an error to the caller if not. Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200901144100.116742-2-f4bug@amsat.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- hw/misc/a9scu.c | 18 +++++++++++++----- 1 file changed, 13 insertions(+), 5 deletions(-) diff --git a/hw/misc/a9scu.c b/hw/misc/a9scu.c index XXXXXXX..XXXXXXX 100644 --- a/hw/misc/a9scu.c +++ b/hw/misc/a9scu.c @@ -XXX,XX +XXX,XX @@ #include "hw/misc/a9scu.h" #include "hw/qdev-properties.h" #include "migration/vmstate.h" +#include "qapi/error.h" #include "qemu/module.h" +#define A9_SCU_CPU_MAX 4 + static uint64_t a9_scu_read(void *opaque, hwaddr offset, unsigned size) { @@ -XXX,XX +XXX,XX @@ static void a9_scu_reset(DeviceState *dev) s->control = 0; } -static void a9_scu_init(Object *obj) +static void a9_scu_realize(DeviceState *dev, Error **errp) { - A9SCUState *s = A9_SCU(obj); - SysBusDevice *sbd = SYS_BUS_DEVICE(obj); + A9SCUState *s = A9_SCU(dev); + SysBusDevice *sbd = SYS_BUS_DEVICE(dev); - memory_region_init_io(&s->iomem, obj, &a9_scu_ops, s, + if (!s->num_cpu || s->num_cpu > A9_SCU_CPU_MAX) { + error_setg(errp, "Illegal CPU count: %u", s->num_cpu); + return; + } + + memory_region_init_io(&s->iomem, OBJECT(s), &a9_scu_ops, s, "a9-scu", 0x100); sysbus_init_mmio(sbd, &s->iomem); } @@ -XXX,XX +XXX,XX @@ static void a9_scu_class_init(ObjectClass *klass, void *data) device_class_set_props(dc, a9_scu_properties); dc->vmsd = &vmstate_a9_scu; dc->reset = a9_scu_reset; + dc->realize = a9_scu_realize; } static const TypeInfo a9_scu_info = { .name = TYPE_A9_SCU, .parent = TYPE_SYS_BUS_DEVICE, .instance_size = sizeof(A9SCUState), - .instance_init = a9_scu_init, .class_init = a9_scu_class_init, }; -- 2.20.1 From: Philippe Mathieu-Daudé <f4bug@amsat.org> Per the datasheet (DDI0407 r2p0): "All SCU registers are byte accessible" and are 32-bit aligned. Set MemoryRegionOps::valid min/max fields and simplify the write() handler. Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200901144100.116742-3-f4bug@amsat.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- hw/misc/a9scu.c | 21 +++++---------------- 1 file changed, 5 insertions(+), 16 deletions(-) diff --git a/hw/misc/a9scu.c b/hw/misc/a9scu.c index XXXXXXX..XXXXXXX 100644 --- a/hw/misc/a9scu.c +++ b/hw/misc/a9scu.c @@ -XXX,XX +XXX,XX @@ static void a9_scu_write(void *opaque, hwaddr offset, uint64_t value, unsigned size) { A9SCUState *s = (A9SCUState *)opaque; - uint32_t mask; + uint32_t mask = MAKE_64BIT_MASK(0, size * 8); uint32_t shift; - switch (size) { - case 1: - mask = 0xff; - break; - case 2: - mask = 0xffff; - break; - case 4: - mask = 0xffffffff; - break; - default: - fprintf(stderr, "Invalid size %u in write to a9 scu register %x\n", - size, (unsigned)offset); - return; - } switch (offset) { case 0x00: /* Control */ @@ -XXX,XX +XXX,XX @@ static void a9_scu_write(void *opaque, hwaddr offset, static const MemoryRegionOps a9_scu_ops = { .read = a9_scu_read, .write = a9_scu_write, + .valid = { + .min_access_size = 1, + .max_access_size = 4, + }, .endianness = DEVICE_NATIVE_ENDIAN, }; -- 2.20.1 From: Philippe Mathieu-Daudé <f4bug@amsat.org> This model implementation is designed for 32-bit accesses. We can simplify setting the MemoryRegionOps::impl min/max fields to 32-bit (memory::access_with_adjusted_size() will take care of the 8/16-bit accesses). Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200901144100.116742-4-f4bug@amsat.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- hw/misc/a9scu.c | 16 +++++----------- 1 file changed, 5 insertions(+), 11 deletions(-) diff --git a/hw/misc/a9scu.c b/hw/misc/a9scu.c index XXXXXXX..XXXXXXX 100644 --- a/hw/misc/a9scu.c +++ b/hw/misc/a9scu.c @@ -XXX,XX +XXX,XX @@ static uint64_t a9_scu_read(void *opaque, hwaddr offset, return (((1 << s->num_cpu) - 1) << 4) | (s->num_cpu - 1); case 0x08: /* CPU Power Status */ return s->status; - case 0x09: /* CPU status. */ - return s->status >> 8; - case 0x0a: /* CPU status. */ - return s->status >> 16; - case 0x0b: /* CPU status. */ - return s->status >> 24; case 0x0c: /* Invalidate All Registers In Secure State */ return 0; case 0x40: /* Filtering Start Address Register */ @@ -XXX,XX +XXX,XX @@ static void a9_scu_write(void *opaque, hwaddr offset, uint64_t value, unsigned size) { A9SCUState *s = (A9SCUState *)opaque; - uint32_t mask = MAKE_64BIT_MASK(0, size * 8); - uint32_t shift; switch (offset) { case 0x00: /* Control */ @@ -XXX,XX +XXX,XX @@ static void a9_scu_write(void *opaque, hwaddr offset, case 0x4: /* Configuration: RO */ break; case 0x08: case 0x09: case 0x0A: case 0x0B: /* Power Control */ - shift = (offset - 0x8) * 8; - s->status &= ~(mask << shift); - s->status |= ((value & mask) << shift); + s->status = value; break; case 0x0c: /* Invalidate All Registers In Secure State */ /* no-op as we do not implement caches */ @@ -XXX,XX +XXX,XX @@ static void a9_scu_write(void *opaque, hwaddr offset, static const MemoryRegionOps a9_scu_ops = { .read = a9_scu_read, .write = a9_scu_write, + .impl = { + .min_access_size = 4, + .max_access_size = 4, + }, .valid = { .min_access_size = 1, .max_access_size = 4, -- 2.20.1 From: Philippe Mathieu-Daudé <f4bug@amsat.org> Report unimplemented register accesses using qemu_log_mask(UNIMP). Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200901144100.116742-5-f4bug@amsat.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- hw/misc/a9scu.c | 6 ++++++ 1 file changed, 6 insertions(+) diff --git a/hw/misc/a9scu.c b/hw/misc/a9scu.c index XXXXXXX..XXXXXXX 100644 --- a/hw/misc/a9scu.c +++ b/hw/misc/a9scu.c @@ -XXX,XX +XXX,XX @@ #include "hw/qdev-properties.h" #include "migration/vmstate.h" #include "qapi/error.h" +#include "qemu/log.h" #include "qemu/module.h" #define A9_SCU_CPU_MAX 4 @@ -XXX,XX +XXX,XX @@ static uint64_t a9_scu_read(void *opaque, hwaddr offset, case 0x54: /* SCU Non-secure Access Control Register */ /* unimplemented, fall through */ default: + qemu_log_mask(LOG_UNIMP, "%s: Unsupported offset 0x%"HWADDR_PRIx"\n", + __func__, offset); return 0; } } @@ -XXX,XX +XXX,XX @@ static void a9_scu_write(void *opaque, hwaddr offset, case 0x54: /* SCU Non-secure Access Control Register */ /* unimplemented, fall through */ default: + qemu_log_mask(LOG_UNIMP, "%s: Unsupported offset 0x%"HWADDR_PRIx + " value 0x%"PRIx64"\n", + __func__, offset, value); break; } } -- 2.20.1 It is the responsibility of board code for an armv7m system to set system_clock_scale appropriately for the CPU speed of the core. If it forgets to do this, then QEMU will hang if the guest tries to use the systick timer in the "tick at the CPU clock frequency" mode. We forgot that in a couple of our boards (see commits ce4f70e81ed23c93f, e7e5a9595ab1136). Add an assertion in the systick reset method so we don't let any new boards in with the same bug. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Message-id: 20200825160847.18091-1-peter.maydell@linaro.org --- hw/timer/armv7m_systick.c | 8 ++++++++ 1 file changed, 8 insertions(+) diff --git a/hw/timer/armv7m_systick.c b/hw/timer/armv7m_systick.c index XXXXXXX..XXXXXXX 100644 --- a/hw/timer/armv7m_systick.c +++ b/hw/timer/armv7m_systick.c @@ -XXX,XX +XXX,XX @@ static void systick_reset(DeviceState *dev) { SysTickState *s = SYSTICK(dev); + /* + * Forgetting to set system_clock_scale is always a board code + * bug. We can't check this earlier because for some boards + * (like stellaris) it is not yet configured at the point where + * the systick device is realized. + */ + assert(system_clock_scale != 0); + s->control = 0; s->reload = 0; s->tick = 0; -- 2.20.1 From: Richard Henderson <richard.henderson@linaro.org> Only argument set members have to be C identifiers, everything else gets prefixed during conversion to C. Some places just checked the leading character, and some places matched a leading character plus a C identifier. Convert everything to match full identifiers, including the [&%@&] prefix, and drop the full C identifier requirement. Reported-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Message-id: 20200903192334.1603773-1-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- tests/decode/succ_ident1.decode | 7 +++++ scripts/decodetree.py | 46 +++++++++++++++++++++------------ 2 files changed, 37 insertions(+), 16 deletions(-) create mode 100644 tests/decode/succ_ident1.decode diff --git a/tests/decode/succ_ident1.decode b/tests/decode/succ_ident1.decode new file mode 100644 index XXXXXXX..XXXXXXX --- /dev/null +++ b/tests/decode/succ_ident1.decode @@ -XXX,XX +XXX,XX @@ +%1f 0:8 +%2f 8:8 +%3f 16:8 + +&3arg a b c +@3arg ........ ........ ........ ........ &3arg a=%1f b=%2f c=%3f +3insn 00000000 ........ ........ ........ @3arg diff --git a/scripts/decodetree.py b/scripts/decodetree.py index XXXXXXX..XXXXXXX 100644 --- a/scripts/decodetree.py +++ b/scripts/decodetree.py @@ -XXX,XX +XXX,XX @@ output_fd = None insntype = 'uint32_t' decode_function = 'decode' -re_ident = '[a-zA-Z][a-zA-Z0-9_]*' +# An identifier for C. +re_C_ident = '[a-zA-Z][a-zA-Z0-9_]*' +# Identifiers for Arguments, Fields, Formats and Patterns. +re_arg_ident = '&[a-zA-Z0-9_]*' +re_fld_ident = '%[a-zA-Z0-9_]*' +re_fmt_ident = '@[a-zA-Z0-9_]*' +re_pat_ident = '[a-zA-Z0-9_]*' def error_with_file(file, lineno, *args): """Print an error message from file:line and args and exit.""" @@ -XXX,XX +XXX,XX @@ class ExcMultiPattern(MultiPattern): def parse_field(lineno, name, toks): """Parse one instruction field from TOKS at LINENO""" global fields - global re_ident global insnwidth # A "simple" field will have only one entry; @@ -XXX,XX +XXX,XX @@ def parse_field(lineno, name, toks): width = 0 func = None for t in toks: - if re.fullmatch('!function=' + re_ident, t): + if re.match('^!function=', t): if func: error(lineno, 'duplicate function') func = t.split('=') @@ -XXX,XX +XXX,XX @@ def parse_field(lineno, name, toks): def parse_arguments(lineno, name, toks): """Parse one argument set from TOKS at LINENO""" global arguments - global re_ident + global re_C_ident global anyextern flds = [] @@ -XXX,XX +XXX,XX @@ def parse_arguments(lineno, name, toks): extern = True anyextern = True continue - if not re.fullmatch(re_ident, t): + if not re.fullmatch(re_C_ident, t): error(lineno, 'invalid argument set token "{0}"'.format(t)) if t in flds: error(lineno, 'duplicate argument "{0}"'.format(t)) @@ -XXX,XX +XXX,XX @@ def parse_generic(lineno, parent_pat, name, toks): global arguments global formats global allpatterns - global re_ident + global re_arg_ident + global re_fld_ident + global re_fmt_ident + global re_C_ident global insnwidth global insnmask global variablewidth @@ -XXX,XX +XXX,XX @@ def parse_generic(lineno, parent_pat, name, toks): fmt = None for t in toks: # '&Foo' gives a format an explcit argument set. - if t[0] == '&': + if re.fullmatch(re_arg_ident, t): tt = t[1:] if arg: error(lineno, 'multiple argument sets') @@ -XXX,XX +XXX,XX @@ def parse_generic(lineno, parent_pat, name, toks): continue # '@Foo' gives a pattern an explicit format. - if t[0] == '@': + if re.fullmatch(re_fmt_ident, t): tt = t[1:] if fmt: error(lineno, 'multiple formats') @@ -XXX,XX +XXX,XX @@ def parse_generic(lineno, parent_pat, name, toks): continue # '%Foo' imports a field. - if t[0] == '%': + if re.fullmatch(re_fld_ident, t): tt = t[1:] flds = add_field_byname(lineno, flds, tt, tt) continue # 'Foo=%Bar' imports a field with a different name. - if re.fullmatch(re_ident + '=%' + re_ident, t): + if re.fullmatch(re_C_ident + '=' + re_fld_ident, t): (fname, iname) = t.split('=%') flds = add_field_byname(lineno, flds, fname, iname) continue # 'Foo=number' sets an argument field to a constant value - if re.fullmatch(re_ident + '=[+-]?[0-9]+', t): + if re.fullmatch(re_C_ident + '=[+-]?[0-9]+', t): (fname, value) = t.split('=') value = int(value) flds = add_field(lineno, flds, fname, ConstField(value)) @@ -XXX,XX +XXX,XX @@ def parse_generic(lineno, parent_pat, name, toks): fixedmask = (fixedmask << shift) | fms undefmask = (undefmask << shift) | ubm # Otherwise, fieldname:fieldwidth - elif re.fullmatch(re_ident + ':s?[0-9]+', t): + elif re.fullmatch(re_C_ident + ':s?[0-9]+', t): (fname, flen) = t.split(':') sign = False if flen[0] == 's': @@ -XXX,XX +XXX,XX @@ def parse_generic(lineno, parent_pat, name, toks): def parse_file(f, parent_pat): """Parse all of the patterns within a file""" + global re_arg_ident + global re_fld_ident + global re_fmt_ident + global re_pat_ident # Read all of the lines of the file. Concatenate lines # ending in backslash; discard empty lines and comments. @@ -XXX,XX +XXX,XX @@ def parse_file(f, parent_pat): continue # Determine the type of object needing to be parsed. - if name[0] == '%': + if re.fullmatch(re_fld_ident, name): parse_field(start_lineno, name[1:], toks) - elif name[0] == '&': + elif re.fullmatch(re_arg_ident, name): parse_arguments(start_lineno, name[1:], toks) - elif name[0] == '@': + elif re.fullmatch(re_fmt_ident, name): parse_generic(start_lineno, None, name[1:], toks) - else: + elif re.fullmatch(re_pat_ident, name): parse_generic(start_lineno, parent_pat, name, toks) + else: + error(lineno, 'invalid token "{0}"'.format(name)) toks = [] if nesting != 0: -- 2.20.1 In the Neon instructions, some instruction formats have a 2-bit size field which corresponds exactly to QEMU's MO_8/16/32/64. However the floating-point insns in the 3-same group have a 1-bit size field which is "0 for 32-bit float and 1 for 16-bit float". Currently we pass these values directly through to trans_ functions, which means that when reading a particular trans_ function you need to know if that insn uses a 2-bit size or a 1-bit size. Move the handling of the 1-bit size to the decodetree file, so that all these insns consistently pass a size to the trans_ function which is an MO_8/16/32/64 value. In this commit we switch over the insns using the 3same_fp and 3same_fp_q0 formats. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200903133209.5141-2-peter.maydell@linaro.org --- target/arm/neon-dp.decode | 15 ++++++++++----- target/arm/translate-neon.c.inc | 16 +++++++++++----- 2 files changed, 21 insertions(+), 10 deletions(-) diff --git a/target/arm/neon-dp.decode b/target/arm/neon-dp.decode index XXXXXXX..XXXXXXX 100644 --- a/target/arm/neon-dp.decode +++ b/target/arm/neon-dp.decode @@ -XXX,XX +XXX,XX @@ @3same_q0 .... ... . . . size:2 .... .... .... . 0 . . .... \ &3same vm=%vm_dp vn=%vn_dp vd=%vd_dp q=0 -# For FP insns the high bit of 'size' is used as part of opcode decode -@3same_fp .... ... . . . . size:1 .... .... .... . q:1 . . .... \ - &3same vm=%vm_dp vn=%vn_dp vd=%vd_dp -@3same_fp_q0 .... ... . . . . size:1 .... .... .... . 0 . . .... \ - &3same vm=%vm_dp vn=%vn_dp vd=%vd_dp q=0 +# For FP insns the high bit of 'size' is used as part of opcode decode, +# and the 'size' bit is 0 for 32-bit float and 1 for 16-bit float. +# This converts this encoding to the same MO_8/16/32/64 values that the +# integer neon insns use. +%3same_fp_size 20:1 !function=neon_3same_fp_size + +@3same_fp .... ... . . . . . .... .... .... . q:1 . . .... \ + &3same vm=%vm_dp vn=%vn_dp vd=%vd_dp size=%3same_fp_size +@3same_fp_q0 .... ... . . . . . .... .... .... . 0 . . .... \ + &3same vm=%vm_dp vn=%vn_dp vd=%vd_dp q=0 size=%3same_fp_size VHADD_S_3s 1111 001 0 0 . .. .... .... 0000 . . . 0 .... @3same VHADD_U_3s 1111 001 1 0 . .. .... .... 0000 . . . 0 .... @3same diff --git a/target/arm/translate-neon.c.inc b/target/arm/translate-neon.c.inc index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-neon.c.inc +++ b/target/arm/translate-neon.c.inc @@ -XXX,XX +XXX,XX @@ static inline int rsub_8(DisasContext *s, int x) return 8 - x; } +static inline int neon_3same_fp_size(DisasContext *s, int x) +{ + /* Convert 0==fp32, 1==fp16 into a MO_* value */ + return MO_32 - x; +} + /* Include the generated Neon decoder */ #include "decode-neon-dp.c.inc" #include "decode-neon-ls.c.inc" @@ -XXX,XX +XXX,XX @@ DO_3SAME_VQDMULH(VQRDMULH, qrdmulh) WRAP_FP_GVEC(gen_##INSN##_fp16_3s, FPST_STD_F16, HFUNC) \ static bool trans_##INSN##_fp_3s(DisasContext *s, arg_3same *a) \ { \ - if (a->size != 0) { \ + if (a->size == MO_16) { \ if (!dc_isar_feature(aa32_fp16_arith, s)) { \ return false; \ } \ @@ -XXX,XX +XXX,XX @@ static bool trans_VMAXNM_fp_3s(DisasContext *s, arg_3same *a) return false; } - if (a->size != 0) { + if (a->size == MO_16) { if (!dc_isar_feature(aa32_fp16_arith, s)) { return false; } @@ -XXX,XX +XXX,XX @@ static bool trans_VMINNM_fp_3s(DisasContext *s, arg_3same *a) return false; } - if (a->size != 0) { + if (a->size == MO_16) { if (!dc_isar_feature(aa32_fp16_arith, s)) { return false; } @@ -XXX,XX +XXX,XX @@ static bool do_3same_fp_pair(DisasContext *s, arg_3same *a, assert(a->q == 0); /* enforced by decode patterns */ - fpstatus = fpstatus_ptr(a->size != 0 ? FPST_STD_F16 : FPST_STD); + fpstatus = fpstatus_ptr(a->size == MO_16 ? FPST_STD_F16 : FPST_STD); tcg_gen_gvec_3_ptr(vfp_reg_offset(1, a->vd), vfp_reg_offset(1, a->vn), vfp_reg_offset(1, a->vm), @@ -XXX,XX +XXX,XX @@ static bool do_3same_fp_pair(DisasContext *s, arg_3same *a, #define DO_3S_FP_PAIR(INSN,FUNC) \ static bool trans_##INSN##_fp_3s(DisasContext *s, arg_3same *a) \ { \ - if (a->size != 0) { \ + if (a->size == MO_16) { \ if (!dc_isar_feature(aa32_fp16_arith, s)) { \ return false; \ } \ -- 2.20.1 Convert the insns using the 2reg_vcvt and 2reg_vcvt_f16 formats to pass the size through to the trans function as a MO_* value rather than the '0==f32, 1==f16' used in the fp 3-same encodings. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200903133209.5141-3-peter.maydell@linaro.org --- target/arm/neon-dp.decode | 3 +-- target/arm/translate-neon.c.inc | 4 ++-- 2 files changed, 3 insertions(+), 4 deletions(-) diff --git a/target/arm/neon-dp.decode b/target/arm/neon-dp.decode index XXXXXXX..XXXXXXX 100644 --- a/target/arm/neon-dp.decode +++ b/target/arm/neon-dp.decode @@ -XXX,XX +XXX,XX @@ VMINNM_fp_3s 1111 001 1 0 . 1 . .... .... 1111 ... 1 .... @3same_fp @2reg_shll_b .... ... . . . 001 shift:3 .... .... 0 . . . .... \ &2reg_shift vm=%vm_dp vd=%vd_dp size=0 q=0 -# We use size=0 for fp32 and size=1 for fp16 to match the 3-same encodings. @2reg_vcvt .... ... . . . 1 ..... .... .... . q:1 . . .... \ - &2reg_shift vm=%vm_dp vd=%vd_dp size=0 shift=%neon_rshift_i5 + &2reg_shift vm=%vm_dp vd=%vd_dp size=2 shift=%neon_rshift_i5 @2reg_vcvt_f16 .... ... . . . 11 .... .... .... . q:1 . . .... \ &2reg_shift vm=%vm_dp vd=%vd_dp size=1 shift=%neon_rshift_i4 diff --git a/target/arm/translate-neon.c.inc b/target/arm/translate-neon.c.inc index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-neon.c.inc +++ b/target/arm/translate-neon.c.inc @@ -XXX,XX +XXX,XX @@ static bool do_fp_2sh(DisasContext *s, arg_2reg_shift *a, return false; } - if (a->size != 0) { + if (a->size == MO_16) { if (!dc_isar_feature(aa32_fp16_arith, s)) { return false; } @@ -XXX,XX +XXX,XX @@ static bool do_fp_2sh(DisasContext *s, arg_2reg_shift *a, return true; } - fpst = fpstatus_ptr(a->size ? FPST_STD_F16 : FPST_STD); + fpst = fpstatus_ptr(a->size == MO_16 ? FPST_STD_F16 : FPST_STD); tcg_gen_gvec_2_ptr(rd_ofs, rm_ofs, fpst, vec_size, vec_size, a->shift, fn); tcg_temp_free_ptr(fpst); return true; -- 2.20.1 The VCMLA and VCADD insns have a size field which is 0 for fp16 and 1 for fp32 (note that this is the reverse of the Neon 3-same encoding!). Convert it to MO_* values in decode for consistency. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200903133209.5141-4-peter.maydell@linaro.org --- target/arm/neon-shared.decode | 18 ++++++++++++------ target/arm/translate-neon.c.inc | 22 ++++++++++++---------- 2 files changed, 24 insertions(+), 16 deletions(-) diff --git a/target/arm/neon-shared.decode b/target/arm/neon-shared.decode index XXXXXXX..XXXXXXX 100644 --- a/target/arm/neon-shared.decode +++ b/target/arm/neon-shared.decode @@ -XXX,XX +XXX,XX @@ %vd_dp 22:1 12:4 %vd_sp 12:4 22:1 -VCMLA 1111 110 rot:2 . 1 size:1 .... .... 1000 . q:1 . 0 .... \ - vm=%vm_dp vn=%vn_dp vd=%vd_dp +# For VCMLA/VCADD insns, convert the single-bit size field +# which is 0 for fp16 and 1 for fp32 into a MO_* constant. +# (Note that this is the reverse of the sense of the 1-bit size +# field in the 3same_fp Neon insns.) +%vcadd_size 20:1 !function=plus1 -VCADD 1111 110 rot:1 1 . 0 size:1 .... .... 1000 . q:1 . 0 .... \ - vm=%vm_dp vn=%vn_dp vd=%vd_dp +VCMLA 1111 110 rot:2 . 1 . .... .... 1000 . q:1 . 0 .... \ + vm=%vm_dp vn=%vn_dp vd=%vd_dp size=%vcadd_size + +VCADD 1111 110 rot:1 1 . 0 . .... .... 1000 . q:1 . 0 .... \ + vm=%vm_dp vn=%vn_dp vd=%vd_dp size=%vcadd_size # VUDOT and VSDOT VDOT 1111 110 00 . 10 .... .... 1101 . q:1 . u:1 .... \ @@ -XXX,XX +XXX,XX @@ VFML 1111 110 0 s:1 . 10 .... .... 1000 . 1 . 1 .... \ vm=%vm_dp vn=%vn_dp vd=%vd_dp q=1 VCMLA_scalar 1111 1110 0 . rot:2 .... .... 1000 . q:1 index:1 0 vm:4 \ - vn=%vn_dp vd=%vd_dp size=0 + vn=%vn_dp vd=%vd_dp size=1 VCMLA_scalar 1111 1110 1 . rot:2 .... .... 1000 . q:1 . 0 .... \ - vm=%vm_dp vn=%vn_dp vd=%vd_dp size=1 index=0 + vm=%vm_dp vn=%vn_dp vd=%vd_dp size=2 index=0 VDOT_scalar 1111 1110 0 . 10 .... .... 1101 . q:1 index:1 u:1 rm:4 \ vm=%vm_dp vn=%vn_dp vd=%vd_dp diff --git a/target/arm/translate-neon.c.inc b/target/arm/translate-neon.c.inc index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-neon.c.inc +++ b/target/arm/translate-neon.c.inc @@ -XXX,XX +XXX,XX @@ static bool trans_VCMLA(DisasContext *s, arg_VCMLA *a) gen_helper_gvec_3_ptr *fn_gvec_ptr; if (!dc_isar_feature(aa32_vcma, s) - || (!a->size && !dc_isar_feature(aa32_fp16_arith, s))) { + || (a->size == MO_16 && !dc_isar_feature(aa32_fp16_arith, s))) { return false; } @@ -XXX,XX +XXX,XX @@ static bool trans_VCMLA(DisasContext *s, arg_VCMLA *a) } opr_sz = (1 + a->q) * 8; - fpst = fpstatus_ptr(a->size == 0 ? FPST_STD_F16 : FPST_STD); - fn_gvec_ptr = a->size ? gen_helper_gvec_fcmlas : gen_helper_gvec_fcmlah; + fpst = fpstatus_ptr(a->size == MO_16 ? FPST_STD_F16 : FPST_STD); + fn_gvec_ptr = (a->size == MO_16) ? + gen_helper_gvec_fcmlah : gen_helper_gvec_fcmlas; tcg_gen_gvec_3_ptr(vfp_reg_offset(1, a->vd), vfp_reg_offset(1, a->vn), vfp_reg_offset(1, a->vm), @@ -XXX,XX +XXX,XX @@ static bool trans_VCADD(DisasContext *s, arg_VCADD *a) gen_helper_gvec_3_ptr *fn_gvec_ptr; if (!dc_isar_feature(aa32_vcma, s) - || (!a->size && !dc_isar_feature(aa32_fp16_arith, s))) { + || (a->size == MO_16 && !dc_isar_feature(aa32_fp16_arith, s))) { return false; } @@ -XXX,XX +XXX,XX @@ static bool trans_VCADD(DisasContext *s, arg_VCADD *a) } opr_sz = (1 + a->q) * 8; - fpst = fpstatus_ptr(a->size == 0 ? FPST_STD_F16 : FPST_STD); - fn_gvec_ptr = a->size ? gen_helper_gvec_fcadds : gen_helper_gvec_fcaddh; + fpst = fpstatus_ptr(a->size == MO_16 ? FPST_STD_F16 : FPST_STD); + fn_gvec_ptr = (a->size == MO_16) ? + gen_helper_gvec_fcaddh : gen_helper_gvec_fcadds; tcg_gen_gvec_3_ptr(vfp_reg_offset(1, a->vd), vfp_reg_offset(1, a->vn), vfp_reg_offset(1, a->vm), @@ -XXX,XX +XXX,XX @@ static bool trans_VCMLA_scalar(DisasContext *s, arg_VCMLA_scalar *a) if (!dc_isar_feature(aa32_vcma, s)) { return false; } - if (a->size == 0 && !dc_isar_feature(aa32_fp16_arith, s)) { + if (a->size == MO_16 && !dc_isar_feature(aa32_fp16_arith, s)) { return false; } @@ -XXX,XX +XXX,XX @@ static bool trans_VCMLA_scalar(DisasContext *s, arg_VCMLA_scalar *a) return true; } - fn_gvec_ptr = (a->size ? gen_helper_gvec_fcmlas_idx - : gen_helper_gvec_fcmlah_idx); + fn_gvec_ptr = (a->size == MO_16) ? + gen_helper_gvec_fcmlah_idx : gen_helper_gvec_fcmlas_idx; opr_sz = (1 + a->q) * 8; - fpst = fpstatus_ptr(a->size == 0 ? FPST_STD_F16 : FPST_STD); + fpst = fpstatus_ptr(a->size == MO_16 ? FPST_STD_F16 : FPST_STD); tcg_gen_gvec_3_ptr(vfp_reg_offset(1, a->vd), vfp_reg_offset(1, a->vn), vfp_reg_offset(1, a->vm), -- 2.20.1 We deprecated the support for KVM on 32-bit Arm hosts in time for release 5.0, which means that our deprecation policy allows us to drop it in release 5.2. Remove the code. To repeat the rationale from the deprecation note: the Linux kernel dropped support for 32-bit Arm KVM hosts in 5.7. Running 32-bit guests on a 64-bit Arm host remains supported. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com> Message-id: 20200904154156.31943-2-peter.maydell@linaro.org --- docs/system/deprecated.rst | 16 +- configure | 2 +- target/arm/kvm32.c | 595 ------------------------------------- target/arm/meson.build | 5 +- 4 files changed, 10 insertions(+), 608 deletions(-) delete mode 100644 target/arm/kvm32.c diff --git a/docs/system/deprecated.rst b/docs/system/deprecated.rst index XXXXXXX..XXXXXXX 100644 --- a/docs/system/deprecated.rst +++ b/docs/system/deprecated.rst @@ -XXX,XX +XXX,XX @@ The ``compat`` property used to set backwards compatibility modes for the processor has been deprecated. The ``max-cpu-compat`` property of the ``pseries`` machine type should be used instead. -KVM guest support on 32-bit Arm hosts (since 5.0) -''''''''''''''''''''''''''''''''''''''''''''''''' - -The Linux kernel has dropped support for allowing 32-bit Arm systems -to host KVM guests as of the 5.7 kernel. Accordingly, QEMU is deprecating -its support for this configuration and will remove it in a future version. -Running 32-bit guests on a 64-bit Arm host remains supported. - System emulator devices ----------------------- @@ -XXX,XX +XXX,XX @@ should be used instead of the 1.09.1 version. System emulator CPUS -------------------- +KVM guest support on 32-bit Arm hosts (removed in 5.2) +'''''''''''''''''''''''''''''''''''''''''''''''''''''' + +The Linux kernel has dropped support for allowing 32-bit Arm systems +to host KVM guests as of the 5.7 kernel. Accordingly, QEMU is deprecating +its support for this configuration and will remove it in a future version. +Running 32-bit guests on a 64-bit Arm host remains supported. + RISC-V ISA Specific CPUs (removed in 5.1) ''''''''''''''''''''''''''''''''''''''''' diff --git a/configure b/configure index XXXXXXX..XXXXXXX 100755 --- a/configure +++ b/configure @@ -XXX,XX +XXX,XX @@ supported_kvm_target() { test "$kvm" = "yes" || return 1 glob "$1" "*-softmmu" || return 1 case "${1%-softmmu}:$cpu" in - arm:arm | aarch64:aarch64 | \ + aarch64:aarch64 | \ i386:i386 | i386:x86_64 | i386:x32 | \ x86_64:i386 | x86_64:x86_64 | x86_64:x32 | \ mips:mips | mipsel:mips | mips64:mips | mips64el:mips | \ diff --git a/target/arm/kvm32.c b/target/arm/kvm32.c deleted file mode 100644 index XXXXXXX..XXXXXXX --- a/target/arm/kvm32.c +++ /dev/null @@ -XXX,XX +XXX,XX @@ -/* - * ARM implementation of KVM hooks, 32 bit specific code. - * - * Copyright Christoffer Dall 2009-2010 - * - * This work is licensed under the terms of the GNU GPL, version 2 or later. - * See the COPYING file in the top-level directory. - * - */ - -#include "qemu/osdep.h" -#include <sys/ioctl.h> - -#include <linux/kvm.h> - -#include "qemu-common.h" -#include "cpu.h" -#include "qemu/timer.h" -#include "sysemu/runstate.h" -#include "sysemu/kvm.h" -#include "kvm_arm.h" -#include "internals.h" -#include "qemu/log.h" - -static int read_sys_reg32(int fd, uint32_t *pret, uint64_t id) -{ - struct kvm_one_reg idreg = { .id = id, .addr = (uintptr_t)pret }; - - assert((id & KVM_REG_SIZE_MASK) == KVM_REG_SIZE_U32); - return ioctl(fd, KVM_GET_ONE_REG, &idreg); -} - -bool kvm_arm_get_host_cpu_features(ARMHostCPUFeatures *ahcf) -{ - /* Identify the feature bits corresponding to the host CPU, and - * fill out the ARMHostCPUClass fields accordingly. To do this - * we have to create a scratch VM, create a single CPU inside it, - * and then query that CPU for the relevant ID registers. - */ - int err = 0, fdarray[3]; - uint32_t midr, id_pfr0; - uint64_t features = 0; - - /* Old kernels may not know about the PREFERRED_TARGET ioctl: however - * we know these will only support creating one kind of guest CPU, - * which is its preferred CPU type. - */ - static const uint32_t cpus_to_try[] = { - QEMU_KVM_ARM_TARGET_CORTEX_A15, - QEMU_KVM_ARM_TARGET_NONE - }; - /* - * target = -1 informs kvm_arm_create_scratch_host_vcpu() - * to use the preferred target - */ - struct kvm_vcpu_init init = { .target = -1, }; - - if (!kvm_arm_create_scratch_host_vcpu(cpus_to_try, fdarray, &init)) { - return false; - } - - ahcf->target = init.target; - - /* This is not strictly blessed by the device tree binding docs yet, - * but in practice the kernel does not care about this string so - * there is no point maintaining an KVM_ARM_TARGET_* -> string table. - */ - ahcf->dtb_compatible = "arm,arm-v7"; - - err |= read_sys_reg32(fdarray[2], &midr, ARM_CP15_REG32(0, 0, 0, 0)); - err |= read_sys_reg32(fdarray[2], &id_pfr0, ARM_CP15_REG32(0, 0, 1, 0)); - - err |= read_sys_reg32(fdarray[2], &ahcf->isar.id_isar0, - ARM_CP15_REG32(0, 0, 2, 0)); - err |= read_sys_reg32(fdarray[2], &ahcf->isar.id_isar1, - ARM_CP15_REG32(0, 0, 2, 1)); - err |= read_sys_reg32(fdarray[2], &ahcf->isar.id_isar2, - ARM_CP15_REG32(0, 0, 2, 2)); - err |= read_sys_reg32(fdarray[2], &ahcf->isar.id_isar3, - ARM_CP15_REG32(0, 0, 2, 3)); - err |= read_sys_reg32(fdarray[2], &ahcf->isar.id_isar4, - ARM_CP15_REG32(0, 0, 2, 4)); - err |= read_sys_reg32(fdarray[2], &ahcf->isar.id_isar5, - ARM_CP15_REG32(0, 0, 2, 5)); - if (read_sys_reg32(fdarray[2], &ahcf->isar.id_isar6, - ARM_CP15_REG32(0, 0, 2, 7))) { - /* - * Older kernels don't support reading ID_ISAR6. This register was - * only introduced in ARMv8, so we can assume that it is zero on a - * CPU that a kernel this old is running on. - */ - ahcf->isar.id_isar6 = 0; - } - - err |= read_sys_reg32(fdarray[2], &ahcf->isar.id_dfr0, - ARM_CP15_REG32(0, 0, 1, 2)); - - err |= read_sys_reg32(fdarray[2], &ahcf->isar.mvfr0, - KVM_REG_ARM | KVM_REG_SIZE_U32 | - KVM_REG_ARM_VFP | KVM_REG_ARM_VFP_MVFR0); - err |= read_sys_reg32(fdarray[2], &ahcf->isar.mvfr1, - KVM_REG_ARM | KVM_REG_SIZE_U32 | - KVM_REG_ARM_VFP | KVM_REG_ARM_VFP_MVFR1); - /* - * FIXME: There is not yet a way to read MVFR2. - * Fortunately there is not yet anything in there that affects migration. - */ - - err |= read_sys_reg32(fdarray[2], &ahcf->isar.id_mmfr0, - ARM_CP15_REG32(0, 0, 1, 4)); - err |= read_sys_reg32(fdarray[2], &ahcf->isar.id_mmfr1, - ARM_CP15_REG32(0, 0, 1, 5)); - err |= read_sys_reg32(fdarray[2], &ahcf->isar.id_mmfr2, - ARM_CP15_REG32(0, 0, 1, 6)); - err |= read_sys_reg32(fdarray[2], &ahcf->isar.id_mmfr3, - ARM_CP15_REG32(0, 0, 1, 7)); - if (read_sys_reg32(fdarray[2], &ahcf->isar.id_mmfr4, - ARM_CP15_REG32(0, 0, 2, 6))) { - /* - * Older kernels don't support reading ID_MMFR4 (a new in v8 - * register); assume it's zero. - */ - ahcf->isar.id_mmfr4 = 0; - } - - /* - * There is no way to read DBGDIDR, because currently 32-bit KVM - * doesn't implement debug at all. Leave it at zero. - */ - - kvm_arm_destroy_scratch_host_vcpu(fdarray); - - if (err < 0) { - return false; - } - - /* Now we've retrieved all the register information we can - * set the feature bits based on the ID register fields. - * We can assume any KVM supporting CPU is at least a v7 - * with VFPv3, virtualization extensions, and the generic - * timers; this in turn implies most of the other feature - * bits, but a few must be tested. - */ - features |= 1ULL << ARM_FEATURE_V7VE; - features |= 1ULL << ARM_FEATURE_GENERIC_TIMER; - - if (extract32(id_pfr0, 12, 4) == 1) { - features |= 1ULL << ARM_FEATURE_THUMB2EE; - } - if (extract32(ahcf->isar.mvfr1, 12, 4) == 1) { - features |= 1ULL << ARM_FEATURE_NEON; - } - - ahcf->features = features; - - return true; -} - -bool kvm_arm_reg_syncs_via_cpreg_list(uint64_t regidx) -{ - /* Return true if the regidx is a register we should synchronize - * via the cpreg_tuples array (ie is not a core reg we sync by - * hand in kvm_arch_get/put_registers()) - */ - switch (regidx & KVM_REG_ARM_COPROC_MASK) { - case KVM_REG_ARM_CORE: - case KVM_REG_ARM_VFP: - return false; - default: - return true; - } -} - -typedef struct CPRegStateLevel { - uint64_t regidx; - int level; -} CPRegStateLevel; - -/* All coprocessor registers not listed in the following table are assumed to - * be of the level KVM_PUT_RUNTIME_STATE. If a register should be written less - * often, you must add it to this table with a state of either - * KVM_PUT_RESET_STATE or KVM_PUT_FULL_STATE. - */ -static const CPRegStateLevel non_runtime_cpregs[] = { - { KVM_REG_ARM_TIMER_CNT, KVM_PUT_FULL_STATE }, -}; - -int kvm_arm_cpreg_level(uint64_t regidx) -{ - int i; - - for (i = 0; i < ARRAY_SIZE(non_runtime_cpregs); i++) { - const CPRegStateLevel *l = &non_runtime_cpregs[i]; - if (l->regidx == regidx) { - return l->level; - } - } - - return KVM_PUT_RUNTIME_STATE; -} - -#define ARM_CPU_ID_MPIDR 0, 0, 0, 5 - -int kvm_arch_init_vcpu(CPUState *cs) -{ - int ret; - uint64_t v; - uint32_t mpidr; - struct kvm_one_reg r; - ARMCPU *cpu = ARM_CPU(cs); - - if (cpu->kvm_target == QEMU_KVM_ARM_TARGET_NONE) { - fprintf(stderr, "KVM is not supported for this guest CPU type\n"); - return -EINVAL; - } - - qemu_add_vm_change_state_handler(kvm_arm_vm_state_change, cs); - - /* Determine init features for this CPU */ - memset(cpu->kvm_init_features, 0, sizeof(cpu->kvm_init_features)); - if (cs->start_powered_off) { - cpu->kvm_init_features[0] |= 1 << KVM_ARM_VCPU_POWER_OFF; - } - if (kvm_check_extension(cs->kvm_state, KVM_CAP_ARM_PSCI_0_2)) { - cpu->psci_version = 2; - cpu->kvm_init_features[0] |= 1 << KVM_ARM_VCPU_PSCI_0_2; - } - - /* Do KVM_ARM_VCPU_INIT ioctl */ - ret = kvm_arm_vcpu_init(cs); - if (ret) { - return ret; - } - - /* Query the kernel to make sure it supports 32 VFP - * registers: QEMU's "cortex-a15" CPU is always a - * VFP-D32 core. The simplest way to do this is just - * to attempt to read register d31. - */ - r.id = KVM_REG_ARM | KVM_REG_SIZE_U64 | KVM_REG_ARM_VFP | 31; - r.addr = (uintptr_t)(&v); - ret = kvm_vcpu_ioctl(cs, KVM_GET_ONE_REG, &r); - if (ret == -ENOENT) { - return -EINVAL; - } - - /* - * When KVM is in use, PSCI is emulated in-kernel and not by qemu. - * Currently KVM has its own idea about MPIDR assignment, so we - * override our defaults with what we get from KVM. - */ - ret = kvm_get_one_reg(cs, ARM_CP15_REG32(ARM_CPU_ID_MPIDR), &mpidr); - if (ret) { - return ret; - } - cpu->mp_affinity = mpidr & ARM32_AFFINITY_MASK; - - /* Check whether userspace can specify guest syndrome value */ - kvm_arm_init_serror_injection(cs); - - return kvm_arm_init_cpreg_list(cpu); -} - -int kvm_arch_destroy_vcpu(CPUState *cs) -{ - return 0; -} - -typedef struct Reg { - uint64_t id; - int offset; -} Reg; - -#define COREREG(KERNELNAME, QEMUFIELD) \ - { \ - KVM_REG_ARM | KVM_REG_SIZE_U32 | \ - KVM_REG_ARM_CORE | KVM_REG_ARM_CORE_REG(KERNELNAME), \ - offsetof(CPUARMState, QEMUFIELD) \ - } - -#define VFPSYSREG(R) \ - { \ - KVM_REG_ARM | KVM_REG_SIZE_U32 | KVM_REG_ARM_VFP | \ - KVM_REG_ARM_VFP_##R, \ - offsetof(CPUARMState, vfp.xregs[ARM_VFP_##R]) \ - } - -/* Like COREREG, but handle fields which are in a uint64_t in CPUARMState. */ -#define COREREG64(KERNELNAME, QEMUFIELD) \ - { \ - KVM_REG_ARM | KVM_REG_SIZE_U32 | \ - KVM_REG_ARM_CORE | KVM_REG_ARM_CORE_REG(KERNELNAME), \ - offsetoflow32(CPUARMState, QEMUFIELD) \ - } - -static const Reg regs[] = { - /* R0_usr .. R14_usr */ - COREREG(usr_regs.uregs[0], regs[0]), - COREREG(usr_regs.uregs[1], regs[1]), - COREREG(usr_regs.uregs[2], regs[2]), - COREREG(usr_regs.uregs[3], regs[3]), - COREREG(usr_regs.uregs[4], regs[4]), - COREREG(usr_regs.uregs[5], regs[5]), - COREREG(usr_regs.uregs[6], regs[6]), - COREREG(usr_regs.uregs[7], regs[7]), - COREREG(usr_regs.uregs[8], usr_regs[0]), - COREREG(usr_regs.uregs[9], usr_regs[1]), - COREREG(usr_regs.uregs[10], usr_regs[2]), - COREREG(usr_regs.uregs[11], usr_regs[3]), - COREREG(usr_regs.uregs[12], usr_regs[4]), - COREREG(usr_regs.uregs[13], banked_r13[BANK_USRSYS]), - COREREG(usr_regs.uregs[14], banked_r14[BANK_USRSYS]), - /* R13, R14, SPSR for SVC, ABT, UND, IRQ banks */ - COREREG(svc_regs[0], banked_r13[BANK_SVC]), - COREREG(svc_regs[1], banked_r14[BANK_SVC]), - COREREG64(svc_regs[2], banked_spsr[BANK_SVC]), - COREREG(abt_regs[0], banked_r13[BANK_ABT]), - COREREG(abt_regs[1], banked_r14[BANK_ABT]), - COREREG64(abt_regs[2], banked_spsr[BANK_ABT]), - COREREG(und_regs[0], banked_r13[BANK_UND]), - COREREG(und_regs[1], banked_r14[BANK_UND]), - COREREG64(und_regs[2], banked_spsr[BANK_UND]), - COREREG(irq_regs[0], banked_r13[BANK_IRQ]), - COREREG(irq_regs[1], banked_r14[BANK_IRQ]), - COREREG64(irq_regs[2], banked_spsr[BANK_IRQ]), - /* R8_fiq .. R14_fiq and SPSR_fiq */ - COREREG(fiq_regs[0], fiq_regs[0]), - COREREG(fiq_regs[1], fiq_regs[1]), - COREREG(fiq_regs[2], fiq_regs[2]), - COREREG(fiq_regs[3], fiq_regs[3]), - COREREG(fiq_regs[4], fiq_regs[4]), - COREREG(fiq_regs[5], banked_r13[BANK_FIQ]), - COREREG(fiq_regs[6], banked_r14[BANK_FIQ]), - COREREG64(fiq_regs[7], banked_spsr[BANK_FIQ]), - /* R15 */ - COREREG(usr_regs.uregs[15], regs[15]), - /* VFP system registers */ - VFPSYSREG(FPSID), - VFPSYSREG(MVFR1), - VFPSYSREG(MVFR0), - VFPSYSREG(FPEXC), - VFPSYSREG(FPINST), - VFPSYSREG(FPINST2), -}; - -int kvm_arch_put_registers(CPUState *cs, int level) -{ - ARMCPU *cpu = ARM_CPU(cs); - CPUARMState *env = &cpu->env; - struct kvm_one_reg r; - int mode, bn; - int ret, i; - uint32_t cpsr, fpscr; - - /* Make sure the banked regs are properly set */ - mode = env->uncached_cpsr & CPSR_M; - bn = bank_number(mode); - if (mode == ARM_CPU_MODE_FIQ) { - memcpy(env->fiq_regs, env->regs + 8, 5 * sizeof(uint32_t)); - } else { - memcpy(env->usr_regs, env->regs + 8, 5 * sizeof(uint32_t)); - } - env->banked_r13[bn] = env->regs[13]; - env->banked_spsr[bn] = env->spsr; - env->banked_r14[r14_bank_number(mode)] = env->regs[14]; - - /* Now we can safely copy stuff down to the kernel */ - for (i = 0; i < ARRAY_SIZE(regs); i++) { - r.id = regs[i].id; - r.addr = (uintptr_t)(env) + regs[i].offset; - ret = kvm_vcpu_ioctl(cs, KVM_SET_ONE_REG, &r); - if (ret) { - return ret; - } - } - - /* Special cases which aren't a single CPUARMState field */ - cpsr = cpsr_read(env); - r.id = KVM_REG_ARM | KVM_REG_SIZE_U32 | - KVM_REG_ARM_CORE | KVM_REG_ARM_CORE_REG(usr_regs.ARM_cpsr); - r.addr = (uintptr_t)(&cpsr); - ret = kvm_vcpu_ioctl(cs, KVM_SET_ONE_REG, &r); - if (ret) { - return ret; - } - - /* VFP registers */ - r.id = KVM_REG_ARM | KVM_REG_SIZE_U64 | KVM_REG_ARM_VFP; - for (i = 0; i < 32; i++) { - r.addr = (uintptr_t)aa32_vfp_dreg(env, i); - ret = kvm_vcpu_ioctl(cs, KVM_SET_ONE_REG, &r); - if (ret) { - return ret; - } - r.id++; - } - - r.id = KVM_REG_ARM | KVM_REG_SIZE_U32 | KVM_REG_ARM_VFP | - KVM_REG_ARM_VFP_FPSCR; - fpscr = vfp_get_fpscr(env); - r.addr = (uintptr_t)&fpscr; - ret = kvm_vcpu_ioctl(cs, KVM_SET_ONE_REG, &r); - if (ret) { - return ret; - } - - write_cpustate_to_list(cpu, true); - - if (!write_list_to_kvmstate(cpu, level)) { - return EINVAL; - } - - /* - * Setting VCPU events should be triggered after syncing the registers - * to avoid overwriting potential changes made by KVM upon calling - * KVM_SET_VCPU_EVENTS ioctl - */ - ret = kvm_put_vcpu_events(cpu); - if (ret) { - return ret; - } - - kvm_arm_sync_mpstate_to_kvm(cpu); - - return ret; -} - -int kvm_arch_get_registers(CPUState *cs) -{ - ARMCPU *cpu = ARM_CPU(cs); - CPUARMState *env = &cpu->env; - struct kvm_one_reg r; - int mode, bn; - int ret, i; - uint32_t cpsr, fpscr; - - for (i = 0; i < ARRAY_SIZE(regs); i++) { - r.id = regs[i].id; - r.addr = (uintptr_t)(env) + regs[i].offset; - ret = kvm_vcpu_ioctl(cs, KVM_GET_ONE_REG, &r); - if (ret) { - return ret; - } - } - - /* Special cases which aren't a single CPUARMState field */ - r.id = KVM_REG_ARM | KVM_REG_SIZE_U32 | - KVM_REG_ARM_CORE | KVM_REG_ARM_CORE_REG(usr_regs.ARM_cpsr); - r.addr = (uintptr_t)(&cpsr); - ret = kvm_vcpu_ioctl(cs, KVM_GET_ONE_REG, &r); - if (ret) { - return ret; - } - cpsr_write(env, cpsr, 0xffffffff, CPSRWriteRaw); - - /* Make sure the current mode regs are properly set */ - mode = env->uncached_cpsr & CPSR_M; - bn = bank_number(mode); - if (mode == ARM_CPU_MODE_FIQ) { - memcpy(env->regs + 8, env->fiq_regs, 5 * sizeof(uint32_t)); - } else { - memcpy(env->regs + 8, env->usr_regs, 5 * sizeof(uint32_t)); - } - env->regs[13] = env->banked_r13[bn]; - env->spsr = env->banked_spsr[bn]; - env->regs[14] = env->banked_r14[r14_bank_number(mode)]; - - /* VFP registers */ - r.id = KVM_REG_ARM | KVM_REG_SIZE_U64 | KVM_REG_ARM_VFP; - for (i = 0; i < 32; i++) { - r.addr = (uintptr_t)aa32_vfp_dreg(env, i); - ret = kvm_vcpu_ioctl(cs, KVM_GET_ONE_REG, &r); - if (ret) { - return ret; - } - r.id++; - } - - r.id = KVM_REG_ARM | KVM_REG_SIZE_U32 | KVM_REG_ARM_VFP | - KVM_REG_ARM_VFP_FPSCR; - r.addr = (uintptr_t)&fpscr; - ret = kvm_vcpu_ioctl(cs, KVM_GET_ONE_REG, &r); - if (ret) { - return ret; - } - vfp_set_fpscr(env, fpscr); - - ret = kvm_get_vcpu_events(cpu); - if (ret) { - return ret; - } - - if (!write_kvmstate_to_list(cpu)) { - return EINVAL; - } - /* Note that it's OK to have registers which aren't in CPUState, - * so we can ignore a failure return here. - */ - write_list_to_cpustate(cpu); - - kvm_arm_sync_mpstate_to_qemu(cpu); - - return 0; -} - -int kvm_arch_insert_sw_breakpoint(CPUState *cs, struct kvm_sw_breakpoint *bp) -{ - qemu_log_mask(LOG_UNIMP, "%s: guest debug not yet implemented\n", __func__); - return -EINVAL; -} - -int kvm_arch_remove_sw_breakpoint(CPUState *cs, struct kvm_sw_breakpoint *bp) -{ - qemu_log_mask(LOG_UNIMP, "%s: guest debug not yet implemented\n", __func__); - return -EINVAL; -} - -bool kvm_arm_handle_debug(CPUState *cs, struct kvm_debug_exit_arch *debug_exit) -{ - qemu_log_mask(LOG_UNIMP, "%s: guest debug not yet implemented\n", __func__); - return false; -} - -int kvm_arch_insert_hw_breakpoint(target_ulong addr, - target_ulong len, int type) -{ - qemu_log_mask(LOG_UNIMP, "%s: not implemented\n", __func__); - return -EINVAL; -} - -int kvm_arch_remove_hw_breakpoint(target_ulong addr, - target_ulong len, int type) -{ - qemu_log_mask(LOG_UNIMP, "%s: not implemented\n", __func__); - return -EINVAL; -} - -void kvm_arch_remove_all_hw_breakpoints(void) -{ - qemu_log_mask(LOG_UNIMP, "%s: not implemented\n", __func__); -} - -void kvm_arm_copy_hw_debug_data(struct kvm_guest_debug_arch *ptr) -{ - qemu_log_mask(LOG_UNIMP, "%s: not implemented\n", __func__); -} - -bool kvm_arm_hw_debug_active(CPUState *cs) -{ - return false; -} - -void kvm_arm_pmu_set_irq(CPUState *cs, int irq) -{ - qemu_log_mask(LOG_UNIMP, "%s: not implemented\n", __func__); -} - -void kvm_arm_pmu_init(CPUState *cs) -{ - qemu_log_mask(LOG_UNIMP, "%s: not implemented\n", __func__); -} - -#define ARM_REG_DFSR ARM_CP15_REG32(0, 5, 0, 0) -#define ARM_REG_TTBCR ARM_CP15_REG32(0, 2, 0, 2) -/* - *DFSR: - * TTBCR.EAE == 0 - * FS[4] - DFSR[10] - * FS[3:0] - DFSR[3:0] - * TTBCR.EAE == 1 - * FS, bits [5:0] - */ -#define DFSR_FSC(lpae, v) \ - ((lpae) ? ((v) & 0x3F) : (((v) >> 6) | ((v) & 0x1F))) - -#define DFSC_EXTABT(lpae) ((lpae) ? 0x10 : 0x08) - -bool kvm_arm_verify_ext_dabt_pending(CPUState *cs) -{ - uint32_t dfsr_val; - - if (!kvm_get_one_reg(cs, ARM_REG_DFSR, &dfsr_val)) { - ARMCPU *cpu = ARM_CPU(cs); - CPUARMState *env = &cpu->env; - uint32_t ttbcr; - int lpae = 0; - - if (!kvm_get_one_reg(cs, ARM_REG_TTBCR, &ttbcr)) { - lpae = arm_feature(env, ARM_FEATURE_LPAE) && (ttbcr & TTBCR_EAE); - } - /* The verification is based on FS filed of the DFSR reg only*/ - return (DFSR_FSC(lpae, dfsr_val) == DFSC_EXTABT(lpae)); - } - return false; -} 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(zlib) arm_ss.add(when: 'CONFIG_TCG', if_true: files('arm-semi.c')) -kvm_ss = ss.source_set() -kvm_ss.add(when: 'TARGET_AARCH64', if_true: files('kvm64.c'), if_false: files('kvm32.c')) -arm_ss.add_all(when: 'CONFIG_KVM', if_true: kvm_ss) -arm_ss.add(when: 'CONFIG_KVM', if_true: files('kvm.c'), if_false: files('kvm-stub.c')) +arm_ss.add(when: 'CONFIG_KVM', if_true: files('kvm.c', 'kvm64.c'), if_false: files('kvm-stub.c')) arm_ss.add(when: 'TARGET_AARCH64', if_true: files( 'cpu64.c', -- 2.20.1 Now that 32-bit KVM host support is gone, KVM can never be enabled unless CONFIG_AARCH64 is true, and some code paths are no longer reachable and can be deleted. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com> Message-id: 20200904154156.31943-3-peter.maydell@linaro.org --- target/arm/kvm-consts.h | 7 --- target/arm/kvm_arm.h | 6 --- target/arm/cpu.c | 101 +++++++++++++++++++--------------------- target/arm/kvm.c | 7 --- 4 files changed, 47 insertions(+), 74 deletions(-) diff --git a/target/arm/kvm-consts.h b/target/arm/kvm-consts.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/kvm-consts.h +++ b/target/arm/kvm-consts.h @@ -XXX,XX +XXX,XX @@ MISMATCH_CHECK(QEMU_PSCI_RET_DISABLED, PSCI_RET_DISABLED); */ #define QEMU_KVM_ARM_TARGET_NONE UINT_MAX -#ifdef TARGET_AARCH64 MISMATCH_CHECK(QEMU_KVM_ARM_TARGET_AEM_V8, KVM_ARM_TARGET_AEM_V8); MISMATCH_CHECK(QEMU_KVM_ARM_TARGET_FOUNDATION_V8, KVM_ARM_TARGET_FOUNDATION_V8); MISMATCH_CHECK(QEMU_KVM_ARM_TARGET_CORTEX_A57, KVM_ARM_TARGET_CORTEX_A57); MISMATCH_CHECK(QEMU_KVM_ARM_TARGET_XGENE_POTENZA, KVM_ARM_TARGET_XGENE_POTENZA); MISMATCH_CHECK(QEMU_KVM_ARM_TARGET_CORTEX_A53, KVM_ARM_TARGET_CORTEX_A53); -#else -MISMATCH_CHECK(QEMU_KVM_ARM_TARGET_CORTEX_A15, KVM_ARM_TARGET_CORTEX_A15); -MISMATCH_CHECK(QEMU_KVM_ARM_TARGET_CORTEX_A7, KVM_ARM_TARGET_CORTEX_A7); -#endif #define CP_REG_ARM64 0x6000000000000000ULL #define CP_REG_ARM_COPROC_MASK 0x000000000FFF0000 @@ -XXX,XX +XXX,XX @@ MISMATCH_CHECK(QEMU_KVM_ARM_TARGET_CORTEX_A7, KVM_ARM_TARGET_CORTEX_A7); /* No kernel define but it's useful to QEMU */ #define CP_REG_ARM64_SYSREG_CP (CP_REG_ARM64_SYSREG >> CP_REG_ARM_COPROC_SHIFT) -#ifdef TARGET_AARCH64 MISMATCH_CHECK(CP_REG_ARM64, KVM_REG_ARM64); MISMATCH_CHECK(CP_REG_ARM_COPROC_MASK, KVM_REG_ARM_COPROC_MASK); MISMATCH_CHECK(CP_REG_ARM_COPROC_SHIFT, KVM_REG_ARM_COPROC_SHIFT); @@ -XXX,XX +XXX,XX @@ MISMATCH_CHECK(CP_REG_ARM64_SYSREG_CRM_MASK, KVM_REG_ARM64_SYSREG_CRM_MASK); MISMATCH_CHECK(CP_REG_ARM64_SYSREG_CRM_SHIFT, KVM_REG_ARM64_SYSREG_CRM_SHIFT); MISMATCH_CHECK(CP_REG_ARM64_SYSREG_OP2_MASK, KVM_REG_ARM64_SYSREG_OP2_MASK); MISMATCH_CHECK(CP_REG_ARM64_SYSREG_OP2_SHIFT, KVM_REG_ARM64_SYSREG_OP2_SHIFT); -#endif #undef MISMATCH_CHECK diff --git a/target/arm/kvm_arm.h b/target/arm/kvm_arm.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/kvm_arm.h +++ b/target/arm/kvm_arm.h @@ -XXX,XX +XXX,XX @@ static inline const char *gic_class_name(void) static inline const char *gicv3_class_name(void) { if (kvm_irqchip_in_kernel()) { -#ifdef TARGET_AARCH64 return "kvm-arm-gicv3"; -#else - error_report("KVM GICv3 acceleration is not supported on this " - "platform"); - exit(1); -#endif } else { if (kvm_enabled()) { error_report("Userspace GICv3 is not supported with KVM"); 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 cortex_a15_initfn(Object *obj) } #ifndef TARGET_AARCH64 -/* -cpu max: if KVM is enabled, like -cpu host (best possible with this host); - * otherwise, a CPU with as many features enabled as our emulation supports. +/* + * -cpu max: a CPU with as many features enabled as our emulation supports. * The version of '-cpu max' for qemu-system-aarch64 is defined in cpu64.c; - * this only needs to handle 32 bits. + * this only needs to handle 32 bits, and need not care about KVM. */ static void arm_max_initfn(Object *obj) { ARMCPU *cpu = ARM_CPU(obj); - if (kvm_enabled()) { - kvm_arm_set_cpu_features_from_host(cpu); - } else { - cortex_a15_initfn(obj); + cortex_a15_initfn(obj); - /* old-style VFP short-vector support */ - cpu->isar.mvfr0 = FIELD_DP32(cpu->isar.mvfr0, MVFR0, FPSHVEC, 1); + /* old-style VFP short-vector support */ + cpu->isar.mvfr0 = FIELD_DP32(cpu->isar.mvfr0, MVFR0, FPSHVEC, 1); #ifdef CONFIG_USER_ONLY - /* We don't set these in system emulation mode for the moment, - * since we don't correctly set (all of) the ID registers to - * advertise them. - */ - set_feature(&cpu->env, ARM_FEATURE_V8); - { - uint32_t t; + /* + * We don't set these in system emulation mode for the moment, + * since we don't correctly set (all of) the ID registers to + * advertise them. + */ + set_feature(&cpu->env, ARM_FEATURE_V8); + { + uint32_t t; - t = cpu->isar.id_isar5; - t = FIELD_DP32(t, ID_ISAR5, AES, 2); - t = FIELD_DP32(t, ID_ISAR5, SHA1, 1); - t = FIELD_DP32(t, ID_ISAR5, SHA2, 1); - t = FIELD_DP32(t, ID_ISAR5, CRC32, 1); - t = FIELD_DP32(t, ID_ISAR5, RDM, 1); - t = FIELD_DP32(t, ID_ISAR5, VCMA, 1); - cpu->isar.id_isar5 = t; + t = cpu->isar.id_isar5; + t = FIELD_DP32(t, ID_ISAR5, AES, 2); + t = FIELD_DP32(t, ID_ISAR5, SHA1, 1); + t = FIELD_DP32(t, ID_ISAR5, SHA2, 1); + t = FIELD_DP32(t, ID_ISAR5, CRC32, 1); + t = FIELD_DP32(t, ID_ISAR5, RDM, 1); + t = FIELD_DP32(t, ID_ISAR5, VCMA, 1); + cpu->isar.id_isar5 = t; - t = cpu->isar.id_isar6; - t = FIELD_DP32(t, ID_ISAR6, JSCVT, 1); - t = FIELD_DP32(t, ID_ISAR6, DP, 1); - t = FIELD_DP32(t, ID_ISAR6, FHM, 1); - t = FIELD_DP32(t, ID_ISAR6, SB, 1); - t = FIELD_DP32(t, ID_ISAR6, SPECRES, 1); - cpu->isar.id_isar6 = t; + t = cpu->isar.id_isar6; + t = FIELD_DP32(t, ID_ISAR6, JSCVT, 1); + t = FIELD_DP32(t, ID_ISAR6, DP, 1); + t = FIELD_DP32(t, ID_ISAR6, FHM, 1); + t = FIELD_DP32(t, ID_ISAR6, SB, 1); + t = FIELD_DP32(t, ID_ISAR6, SPECRES, 1); + cpu->isar.id_isar6 = t; - t = cpu->isar.mvfr1; - t = FIELD_DP32(t, MVFR1, FPHP, 3); /* v8.2-FP16 */ - t = FIELD_DP32(t, MVFR1, SIMDHP, 2); /* v8.2-FP16 */ - cpu->isar.mvfr1 = t; + t = cpu->isar.mvfr1; + t = FIELD_DP32(t, MVFR1, FPHP, 3); /* v8.2-FP16 */ + t = FIELD_DP32(t, MVFR1, SIMDHP, 2); /* v8.2-FP16 */ + cpu->isar.mvfr1 = t; - t = cpu->isar.mvfr2; - t = FIELD_DP32(t, MVFR2, SIMDMISC, 3); /* SIMD MaxNum */ - t = FIELD_DP32(t, MVFR2, FPMISC, 4); /* FP MaxNum */ - cpu->isar.mvfr2 = t; + t = cpu->isar.mvfr2; + t = FIELD_DP32(t, MVFR2, SIMDMISC, 3); /* SIMD MaxNum */ + t = FIELD_DP32(t, MVFR2, FPMISC, 4); /* FP MaxNum */ + cpu->isar.mvfr2 = t; - t = cpu->isar.id_mmfr3; - t = FIELD_DP32(t, ID_MMFR3, PAN, 2); /* ATS1E1 */ - cpu->isar.id_mmfr3 = t; + t = cpu->isar.id_mmfr3; + t = FIELD_DP32(t, ID_MMFR3, PAN, 2); /* ATS1E1 */ + cpu->isar.id_mmfr3 = t; - t = cpu->isar.id_mmfr4; - t = FIELD_DP32(t, ID_MMFR4, HPDS, 1); /* AA32HPD */ - t = FIELD_DP32(t, ID_MMFR4, AC2, 1); /* ACTLR2, HACTLR2 */ - t = FIELD_DP32(t, ID_MMFR4, CNP, 1); /* TTCNP */ - t = FIELD_DP32(t, ID_MMFR4, XNX, 1); /* TTS2UXN */ - cpu->isar.id_mmfr4 = t; - } -#endif + t = cpu->isar.id_mmfr4; + t = FIELD_DP32(t, ID_MMFR4, HPDS, 1); /* AA32HPD */ + t = FIELD_DP32(t, ID_MMFR4, AC2, 1); /* ACTLR2, HACTLR2 */ + t = FIELD_DP32(t, ID_MMFR4, CNP, 1); /* TTCNP */ + t = FIELD_DP32(t, ID_MMFR4, XNX, 1); /* TTS2UXN */ + cpu->isar.id_mmfr4 = t; } +#endif } #endif @@ -XXX,XX +XXX,XX @@ static void arm_host_initfn(Object *obj) static const TypeInfo host_arm_cpu_type_info = { .name = TYPE_ARM_HOST_CPU, -#ifdef TARGET_AARCH64 .parent = TYPE_AARCH64_CPU, -#else - .parent = TYPE_ARM_CPU, -#endif .instance_init = arm_host_initfn, }; diff --git a/target/arm/kvm.c b/target/arm/kvm.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/kvm.c +++ b/target/arm/kvm.c @@ -XXX,XX +XXX,XX @@ int kvm_arch_process_async_events(CPUState *cs) return 0; } -/* The #ifdef protections are until 32bit headers are imported and can - * be removed once both 32 and 64 bit reach feature parity. - */ void kvm_arch_update_guest_debug(CPUState *cs, struct kvm_guest_debug *dbg) { -#ifdef KVM_GUESTDBG_USE_SW_BP if (kvm_sw_breakpoints_active(cs)) { dbg->control |= KVM_GUESTDBG_ENABLE | KVM_GUESTDBG_USE_SW_BP; } -#endif -#ifdef KVM_GUESTDBG_USE_HW if (kvm_arm_hw_debug_active(cs)) { dbg->control |= KVM_GUESTDBG_ENABLE | KVM_GUESTDBG_USE_HW; kvm_arm_copy_hw_debug_data(&dbg->arch); } -#endif } void kvm_arch_init_irq_routing(KVMState *s) -- 2.20.1 Implement a model of the MPS2 with the AN386 firmware. This is essentially identical to the AN385 firmware, but it has a Cortex-M4 rather than a Cortex-M3. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Message-id: 20200903202048.15370-2-peter.maydell@linaro.org --- docs/system/arm/mps2.rst | 8 +++++--- hw/arm/mps2.c | 34 +++++++++++++++++++++++++++++----- 2 files changed, 34 insertions(+), 8 deletions(-) diff --git a/docs/system/arm/mps2.rst b/docs/system/arm/mps2.rst index XXXXXXX..XXXXXXX 100644 --- a/docs/system/arm/mps2.rst +++ b/docs/system/arm/mps2.rst @@ -XXX,XX +XXX,XX @@ -Arm MPS2 boards (``mps2-an385``, ``mps2-an505``, ``mps2-an511``, ``mps2-an521``) -================================================================================ +Arm MPS2 boards (``mps2-an385``, ``mps2-an386``, ``mps2-an505``, ``mps2-an511``, ``mps2-an521``) +================================================================================================ These board models all use Arm M-profile CPUs. @@ -XXX,XX +XXX,XX @@ QEMU models the following FPGA images: ``mps2-an385`` Cortex-M3 as documented in ARM Application Note AN385 +``mps2-an386`` + Cortex-M4 as documented in ARM Application Note AN386 ``mps2-an511`` Cortex-M3 'DesignStart' as documented in AN511 ``mps2-an505`` @@ -XXX,XX +XXX,XX @@ QEMU models the following FPGA images: Differences between QEMU and real hardware: -- AN385 remapping of low 16K of memory to either ZBT SSRAM1 or to +- AN385/AN386 remapping of low 16K of memory to either ZBT SSRAM1 or to block RAM is unimplemented (QEMU always maps this to ZBT SSRAM1, as if zbt_boot_ctrl is always zero) - QEMU provides a LAN9118 ethernet rather than LAN9220; the only guest 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 @@ * as seen by the guest depend significantly on the FPGA image. * We model the following FPGA images: * "mps2-an385" -- Cortex-M3 as documented in ARM Application Note AN385 + * "mps2-an386" -- Cortex-M4 as documented in ARM Application Note AN386 * "mps2-an511" -- Cortex-M3 'DesignStart' as documented in AN511 * * Links to the TRM for the board itself and to the various Application @@ -XXX,XX +XXX,XX @@ typedef enum MPS2FPGAType { FPGA_AN385, + FPGA_AN386, FPGA_AN511, } MPS2FPGAType; @@ -XXX,XX +XXX,XX @@ typedef struct MPS2MachineState MPS2MachineState; #define TYPE_MPS2_MACHINE "mps2" #define TYPE_MPS2_AN385_MACHINE MACHINE_TYPE_NAME("mps2-an385") +#define TYPE_MPS2_AN386_MACHINE MACHINE_TYPE_NAME("mps2-an386") #define TYPE_MPS2_AN511_MACHINE MACHINE_TYPE_NAME("mps2-an511") DECLARE_OBJ_CHECKERS(MPS2MachineState, MPS2MachineClass, @@ -XXX,XX +XXX,XX @@ static void mps2_common_init(MachineState *machine) * tradeoffs. For QEMU they're all just RAM, though. We arbitrarily * call the 16MB our "system memory", as it's the largest lump. * - * Common to both boards: - * 0x21000000..0x21ffffff : PSRAM (16MB) - * AN385 only: + * AN385/AN386/AN511: + * 0x21000000 .. 0x21ffffff : PSRAM (16MB) + * AN385/AN386 only: * 0x00000000 .. 0x003fffff : ZBT SSRAM1 * 0x00400000 .. 0x007fffff : mirror of ZBT SSRAM1 * 0x20000000 .. 0x203fffff : ZBT SSRAM 2&3 @@ -XXX,XX +XXX,XX @@ static void mps2_common_init(MachineState *machine) * 0x20000000 .. 0x2001ffff : SRAM * 0x20400000 .. 0x207fffff : ZBT SSRAM 2&3 * - * The AN385 has a feature where the lowest 16K can be mapped + * The AN385/AN386 has a feature where the lowest 16K can be mapped * either to the bottom of the ZBT SSRAM1 or to the block RAM. * This is of no use for QEMU so we don't implement it (as if * zbt_boot_ctrl is always zero). @@ -XXX,XX +XXX,XX @@ static void mps2_common_init(MachineState *machine) switch (mmc->fpga_type) { case FPGA_AN385: + case FPGA_AN386: make_ram(&mms->ssram1, "mps.ssram1", 0x0, 0x400000); make_ram_alias(&mms->ssram1_m, "mps.ssram1_m", &mms->ssram1, 0x400000); make_ram(&mms->ssram23, "mps.ssram23", 0x20000000, 0x400000); @@ -XXX,XX +XXX,XX @@ static void mps2_common_init(MachineState *machine) armv7m = DEVICE(&mms->armv7m); switch (mmc->fpga_type) { case FPGA_AN385: + case FPGA_AN386: qdev_prop_set_uint32(armv7m, "num-irq", 32); break; case FPGA_AN511: @@ -XXX,XX +XXX,XX @@ static void mps2_common_init(MachineState *machine) switch (mmc->fpga_type) { case FPGA_AN385: + case FPGA_AN386: { /* The overflow IRQs for UARTs 0, 1 and 2 are ORed together. * Overflow for UARTs 4 and 5 doesn't trigger any interrupt. @@ -XXX,XX +XXX,XX @@ static void mps2_common_init(MachineState *machine) */ lan9118_init(&nd_table[0], 0x40200000, qdev_get_gpio_in(armv7m, - mmc->fpga_type == FPGA_AN385 ? 13 : 47)); + mmc->fpga_type == FPGA_AN511 ? 47 : 13)); system_clock_scale = NANOSECONDS_PER_SECOND / SYSCLK_FRQ; @@ -XXX,XX +XXX,XX @@ static void mps2_an385_class_init(ObjectClass *oc, void *data) mmc->scc_id = 0x41043850; } +static void mps2_an386_class_init(ObjectClass *oc, void *data) +{ + MachineClass *mc = MACHINE_CLASS(oc); + MPS2MachineClass *mmc = MPS2_MACHINE_CLASS(oc); + + mc->desc = "ARM MPS2 with AN386 FPGA image for Cortex-M4"; + mmc->fpga_type = FPGA_AN386; + mc->default_cpu_type = ARM_CPU_TYPE_NAME("cortex-m4"); + mmc->scc_id = 0x41043860; +} + static void mps2_an511_class_init(ObjectClass *oc, void *data) { MachineClass *mc = MACHINE_CLASS(oc); @@ -XXX,XX +XXX,XX @@ static const TypeInfo mps2_an385_info = { .class_init = mps2_an385_class_init, }; +static const TypeInfo mps2_an386_info = { + .name = TYPE_MPS2_AN386_MACHINE, + .parent = TYPE_MPS2_MACHINE, + .class_init = mps2_an386_class_init, +}; + static const TypeInfo mps2_an511_info = { .name = TYPE_MPS2_AN511_MACHINE, .parent = TYPE_MPS2_MACHINE, @@ -XXX,XX +XXX,XX @@ static void mps2_machine_init(void) { type_register_static(&mps2_info); type_register_static(&mps2_an385_info); + type_register_static(&mps2_an386_info); type_register_static(&mps2_an511_info); } -- 2.20.1 Implement a model of the MPS2 with the AN500 firmware. This is similar to the AN385, with the following differences: * Cortex-M7 CPU * PSRAM is at 0x6000_0000 * Ethernet is at 0xa000_0000 * No zbt_boot_ctrl remapping of the low 16K (but QEMU doesn't implement this anyway) * no "block RAM" at 0x01000000 Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Message-id: 20200903202048.15370-3-peter.maydell@linaro.org --- docs/system/arm/mps2.rst | 6 ++-- hw/arm/mps2.c | 67 +++++++++++++++++++++++++++++++++------- 2 files changed, 60 insertions(+), 13 deletions(-) diff --git a/docs/system/arm/mps2.rst b/docs/system/arm/mps2.rst index XXXXXXX..XXXXXXX 100644 --- a/docs/system/arm/mps2.rst +++ b/docs/system/arm/mps2.rst @@ -XXX,XX +XXX,XX @@ -Arm MPS2 boards (``mps2-an385``, ``mps2-an386``, ``mps2-an505``, ``mps2-an511``, ``mps2-an521``) -================================================================================================ +Arm MPS2 boards (``mps2-an385``, ``mps2-an386``, ``mps2-an500``, ``mps2-an505``, ``mps2-an511``, ``mps2-an521``) +================================================================================================================ These board models all use Arm M-profile CPUs. @@ -XXX,XX +XXX,XX @@ QEMU models the following FPGA images: Cortex-M3 as documented in ARM Application Note AN385 ``mps2-an386`` Cortex-M4 as documented in ARM Application Note AN386 +``mps2-an500`` + Cortex-M7 as documented in ARM Application Note AN500 ``mps2-an511`` Cortex-M3 'DesignStart' as documented in AN511 ``mps2-an505`` 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 @@ * We model the following FPGA images: * "mps2-an385" -- Cortex-M3 as documented in ARM Application Note AN385 * "mps2-an386" -- Cortex-M4 as documented in ARM Application Note AN386 + * "mps2-an500" -- Cortex-M7 as documented in ARM Application Note AN500 * "mps2-an511" -- Cortex-M3 'DesignStart' as documented in AN511 * * Links to the TRM for the board itself and to the various Application @@ -XXX,XX +XXX,XX @@ typedef enum MPS2FPGAType { FPGA_AN385, FPGA_AN386, + FPGA_AN500, FPGA_AN511, } MPS2FPGAType; @@ -XXX,XX +XXX,XX @@ struct MPS2MachineClass { MachineClass parent; MPS2FPGAType fpga_type; uint32_t scc_id; + bool has_block_ram; + hwaddr ethernet_base; + hwaddr psram_base; }; typedef struct MPS2MachineClass MPS2MachineClass; @@ -XXX,XX +XXX,XX @@ typedef struct MPS2MachineState MPS2MachineState; #define TYPE_MPS2_MACHINE "mps2" #define TYPE_MPS2_AN385_MACHINE MACHINE_TYPE_NAME("mps2-an385") #define TYPE_MPS2_AN386_MACHINE MACHINE_TYPE_NAME("mps2-an386") +#define TYPE_MPS2_AN500_MACHINE MACHINE_TYPE_NAME("mps2-an500") #define TYPE_MPS2_AN511_MACHINE MACHINE_TYPE_NAME("mps2-an511") DECLARE_OBJ_CHECKERS(MPS2MachineState, MPS2MachineClass, @@ -XXX,XX +XXX,XX @@ static void mps2_common_init(MachineState *machine) * * AN385/AN386/AN511: * 0x21000000 .. 0x21ffffff : PSRAM (16MB) - * AN385/AN386 only: + * AN385/AN386/AN500: * 0x00000000 .. 0x003fffff : ZBT SSRAM1 * 0x00400000 .. 0x007fffff : mirror of ZBT SSRAM1 * 0x20000000 .. 0x203fffff : ZBT SSRAM 2&3 * 0x20400000 .. 0x207fffff : mirror of ZBT SSRAM 2&3 + * AN385/AN386 only: * 0x01000000 .. 0x01003fff : block RAM (16K) * 0x01004000 .. 0x01007fff : mirror of above * 0x01008000 .. 0x0100bfff : mirror of above @@ -XXX,XX +XXX,XX @@ static void mps2_common_init(MachineState *machine) * 0x00400000 .. 0x007fffff : ZBT SSRAM1 * 0x20000000 .. 0x2001ffff : SRAM * 0x20400000 .. 0x207fffff : ZBT SSRAM 2&3 + * AN500 only: + * 0x60000000 .. 0x60ffffff : PSRAM (16MB) * * The AN385/AN386 has a feature where the lowest 16K can be mapped * either to the bottom of the ZBT SSRAM1 or to the block RAM. * This is of no use for QEMU so we don't implement it (as if * zbt_boot_ctrl is always zero). */ - memory_region_add_subregion(system_memory, 0x21000000, machine->ram); + memory_region_add_subregion(system_memory, mmc->psram_base, machine->ram); - switch (mmc->fpga_type) { - case FPGA_AN385: - case FPGA_AN386: - make_ram(&mms->ssram1, "mps.ssram1", 0x0, 0x400000); - make_ram_alias(&mms->ssram1_m, "mps.ssram1_m", &mms->ssram1, 0x400000); - make_ram(&mms->ssram23, "mps.ssram23", 0x20000000, 0x400000); - make_ram_alias(&mms->ssram23_m, "mps.ssram23_m", - &mms->ssram23, 0x20400000); + if (mmc->has_block_ram) { make_ram(&mms->blockram, "mps.blockram", 0x01000000, 0x4000); make_ram_alias(&mms->blockram_m1, "mps.blockram_m1", &mms->blockram, 0x01004000); @@ -XXX,XX +XXX,XX @@ static void mps2_common_init(MachineState *machine) &mms->blockram, 0x01008000); make_ram_alias(&mms->blockram_m3, "mps.blockram_m3", &mms->blockram, 0x0100c000); + } + + switch (mmc->fpga_type) { + case FPGA_AN385: + case FPGA_AN386: + case FPGA_AN500: + make_ram(&mms->ssram1, "mps.ssram1", 0x0, 0x400000); + make_ram_alias(&mms->ssram1_m, "mps.ssram1_m", &mms->ssram1, 0x400000); + make_ram(&mms->ssram23, "mps.ssram23", 0x20000000, 0x400000); + make_ram_alias(&mms->ssram23_m, "mps.ssram23_m", + &mms->ssram23, 0x20400000); break; case FPGA_AN511: make_ram(&mms->blockram, "mps.blockram", 0x0, 0x40000); @@ -XXX,XX +XXX,XX @@ static void mps2_common_init(MachineState *machine) switch (mmc->fpga_type) { case FPGA_AN385: case FPGA_AN386: + case FPGA_AN500: qdev_prop_set_uint32(armv7m, "num-irq", 32); break; case FPGA_AN511: @@ -XXX,XX +XXX,XX @@ static void mps2_common_init(MachineState *machine) switch (mmc->fpga_type) { case FPGA_AN385: case FPGA_AN386: + case FPGA_AN500: { /* The overflow IRQs for UARTs 0, 1 and 2 are ORed together. * Overflow for UARTs 4 and 5 doesn't trigger any interrupt. @@ -XXX,XX +XXX,XX @@ static void mps2_common_init(MachineState *machine) /* In hardware this is a LAN9220; the LAN9118 is software compatible * except that it doesn't support the checksum-offload feature. */ - lan9118_init(&nd_table[0], 0x40200000, + lan9118_init(&nd_table[0], mmc->ethernet_base, qdev_get_gpio_in(armv7m, mmc->fpga_type == FPGA_AN511 ? 47 : 13)); @@ -XXX,XX +XXX,XX @@ static void mps2_an385_class_init(ObjectClass *oc, void *data) mmc->fpga_type = FPGA_AN385; mc->default_cpu_type = ARM_CPU_TYPE_NAME("cortex-m3"); mmc->scc_id = 0x41043850; + mmc->psram_base = 0x21000000; + mmc->ethernet_base = 0x40200000; + mmc->has_block_ram = true; } static void mps2_an386_class_init(ObjectClass *oc, void *data) @@ -XXX,XX +XXX,XX @@ static void mps2_an386_class_init(ObjectClass *oc, void *data) mmc->fpga_type = FPGA_AN386; mc->default_cpu_type = ARM_CPU_TYPE_NAME("cortex-m4"); mmc->scc_id = 0x41043860; + mmc->psram_base = 0x21000000; + mmc->ethernet_base = 0x40200000; + mmc->has_block_ram = true; +} + +static void mps2_an500_class_init(ObjectClass *oc, void *data) +{ + MachineClass *mc = MACHINE_CLASS(oc); + MPS2MachineClass *mmc = MPS2_MACHINE_CLASS(oc); + + mc->desc = "ARM MPS2 with AN500 FPGA image for Cortex-M7"; + mmc->fpga_type = FPGA_AN500; + mc->default_cpu_type = ARM_CPU_TYPE_NAME("cortex-m7"); + mmc->scc_id = 0x41045000; + mmc->psram_base = 0x60000000; + mmc->ethernet_base = 0xa0000000; + mmc->has_block_ram = false; } static void mps2_an511_class_init(ObjectClass *oc, void *data) @@ -XXX,XX +XXX,XX @@ static void mps2_an511_class_init(ObjectClass *oc, void *data) mmc->fpga_type = FPGA_AN511; mc->default_cpu_type = ARM_CPU_TYPE_NAME("cortex-m3"); mmc->scc_id = 0x41045110; + mmc->psram_base = 0x21000000; + mmc->ethernet_base = 0x40200000; + mmc->has_block_ram = false; } static const TypeInfo mps2_info = { @@ -XXX,XX +XXX,XX @@ static const TypeInfo mps2_an386_info = { .class_init = mps2_an386_class_init, }; +static const TypeInfo mps2_an500_info = { + .name = TYPE_MPS2_AN500_MACHINE, + .parent = TYPE_MPS2_MACHINE, + .class_init = mps2_an500_class_init, +}; + static const TypeInfo mps2_an511_info = { .name = TYPE_MPS2_AN511_MACHINE, .parent = TYPE_MPS2_MACHINE, @@ -XXX,XX +XXX,XX @@ static void mps2_machine_init(void) type_register_static(&mps2_info); type_register_static(&mps2_an385_info); type_register_static(&mps2_an386_info); + type_register_static(&mps2_an500_info); type_register_static(&mps2_an511_info); } -- 2.20.1 Make the list of MPS2 boards consistent in the phrasing of each entry, use the correct casing of "Arm", and move the mps2-an511 entry so the list is in numeric order. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Message-id: 20200903202048.15370-4-peter.maydell@linaro.org --- docs/system/arm/mps2.rst | 14 +++++++------- 1 file changed, 7 insertions(+), 7 deletions(-) diff --git a/docs/system/arm/mps2.rst b/docs/system/arm/mps2.rst index XXXXXXX..XXXXXXX 100644 --- a/docs/system/arm/mps2.rst +++ b/docs/system/arm/mps2.rst @@ -XXX,XX +XXX,XX @@ as seen by the guest depend significantly on the FPGA image. QEMU models the following FPGA images: ``mps2-an385`` - Cortex-M3 as documented in ARM Application Note AN385 + Cortex-M3 as documented in Arm Application Note AN385 ``mps2-an386`` - Cortex-M4 as documented in ARM Application Note AN386 + Cortex-M4 as documented in Arm Application Note AN386 ``mps2-an500`` - Cortex-M7 as documented in ARM Application Note AN500 -``mps2-an511`` - Cortex-M3 'DesignStart' as documented in AN511 + Cortex-M7 as documented in Arm Application Note AN500 ``mps2-an505`` - Cortex-M33 as documented in ARM Application Note AN505 + Cortex-M33 as documented in Arm Application Note AN505 +``mps2-an511`` + Cortex-M3 'DesignStart' as documented in Arm Application Note AN511 ``mps2-an521`` - Dual Cortex-M33 as documented in Application Note AN521 + Dual Cortex-M33 as documented in Arm Application Note AN521 Differences between QEMU and real hardware: -- 2.20.1 Deprecate our Unicore32 target support: * the Linux kernel dropped support for unicore32 in commit 05119217a9bd199c for its 5.9 release (with rationale in the cover letter: https://lkml.org/lkml/2020/8/3/232 ) * there is apparently no upstream toolchain that can create unicore32 binaries * the maintainer doesn't seem to have made any contributions to QEMU since the port first landed in 2012 * nobody else seems to have made changes to the unicore code except for generic cleanups either Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> Message-id: 20200825172719.19422-1-peter.maydell@linaro.org --- docs/system/deprecated.rst | 8 ++++++++ 1 file changed, 8 insertions(+) diff --git a/docs/system/deprecated.rst b/docs/system/deprecated.rst index XXXXXXX..XXXXXXX 100644 --- a/docs/system/deprecated.rst +++ b/docs/system/deprecated.rst @@ -XXX,XX +XXX,XX @@ The above, converted to the current supported format:: linux-user mode CPUs -------------------- +``unicore32`` CPUs (since 5.2.0) +'''''''''''''''''''''''''''''''' + +The ``unicore32`` guest CPU support is deprecated and will be removed in +a future version of QEMU. Support for this CPU was removed from the +upstream Linux kernel, and there is no available upstream toolchain +to build binaries for it. + ``tilegx`` CPUs (since 5.1.0) ''''''''''''''''''''''''''''' -- 2.20.1 Deprecate our lm32 target support. Michael Walle (former lm32 maintainer) suggested that we do this in 2019: https://www.mail-archive.com/qemu-devel@nongnu.org/msg605024.html because the only public user of the architecture is the many-years-dead milkymist project. (The Linux port to lm32 was never merged upstream.) In commit 4b4d96c776f552e (March 2020) we marked it as 'orphan' in the MAINTAINERS file, but didn't officially deprecate it. Mark it deprecated now, with the intention of removing it from QEMU in mid-2021 before the 6.1 release. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Reviewed-by: Thomas Huth <thuth@redhat.com> Acked-by: Michael Walle <michael@walle.cc> Message-id: 20200827113259.25064-1-peter.maydell@linaro.org --- docs/system/deprecated.rst | 8 ++++++++ 1 file changed, 8 insertions(+) diff --git a/docs/system/deprecated.rst b/docs/system/deprecated.rst index XXXXXXX..XXXXXXX 100644 --- a/docs/system/deprecated.rst +++ b/docs/system/deprecated.rst @@ -XXX,XX +XXX,XX @@ The above, converted to the current supported format:: linux-user mode CPUs -------------------- +``lm32`` CPUs (since 5.2.0) +''''''''''''''''''''''''''' + +The ``lm32`` guest CPU support is deprecated and will be removed in +a future version of QEMU. The only public user of this architecture +was the milkymist project, which has been dead for years; there was +never an upstream Linux port. + ``unicore32`` CPUs (since 5.2.0) '''''''''''''''''''''''''''''''' -- 2.20.1 From: Aaron Lindsay <aaron@os.amperecomputing.com> This check was backwards when introduced in commit 033614c47de78409ad3fb39bb7bd1483b71c6789: target/arm: Filter cycle counter based on PMCCFILTR_EL0 Cc: qemu-stable@nongnu.org Signed-off-by: Aaron Lindsay <aaron@os.amperecomputing.com> Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/helper.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/target/arm/helper.c b/target/arm/helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper.c +++ b/target/arm/helper.c @@ -XXX,XX +XXX,XX @@ static bool pmu_counter_enabled(CPUARMState *env, uint8_t counter) } } else { prohibited = arm_feature(env, ARM_FEATURE_EL3) && - (env->cp15.mdcr_el3 & MDCR_SPME); + !(env->cp15.mdcr_el3 & MDCR_SPME); } if (prohibited && counter == 31) { -- 2.20.1 From: "Edgar E. Iglesias" <edgar.iglesias@xilinx.com> Correct the GEMs tx/rx clocks to use the 125Mhz fixed-clock. This matches the setup with the fixed-link 100Mbit PHY. It also avoids the following warnings from the Linux kernel driver: eth0: unable to generate target frequency: 125000000 Hz Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Reviewed-by: Luc Michel <luc.michel@greensocs.com> Signed-off-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com> Message-id: 20200909174647.662864-2-edgar.iglesias@gmail.com Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- hw/arm/xlnx-versal-virt.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/hw/arm/xlnx-versal-virt.c b/hw/arm/xlnx-versal-virt.c index XXXXXXX..XXXXXXX 100644 --- a/hw/arm/xlnx-versal-virt.c +++ b/hw/arm/xlnx-versal-virt.c @@ -XXX,XX +XXX,XX @@ static void fdt_add_gem_nodes(VersalVirt *s) s->phandle.ethernet_phy[i]); qemu_fdt_setprop_cells(s->fdt, name, "clocks", s->phandle.clk_25Mhz, s->phandle.clk_25Mhz, - s->phandle.clk_25Mhz, s->phandle.clk_25Mhz); + s->phandle.clk_125Mhz, s->phandle.clk_125Mhz); qemu_fdt_setprop(s->fdt, name, "clock-names", clocknames, sizeof(clocknames)); qemu_fdt_setprop_cells(s->fdt, name, "interrupts", -- 2.20.1 From: Havard Skinnemoen <hskinnemoen@google.com> Implement a device model for the System Global Control Registers in the NPCM730 and NPCM750 BMC SoCs. This is primarily used to enable SMP boot (the boot ROM spins reading the SCRPAD register) and DDR memory initialization; other registers are best effort for now. The reset values of the MDLR and PWRON registers are determined by the SoC variant (730 vs 750) and board straps respectively. Reviewed-by: Joel Stanley <joel@jms.id.au> Reviewed-by: Cédric Le Goater <clg@kaod.org> Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Tested-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Tested-by: Alexander Bulekov <alxndr@bu.edu> Signed-off-by: Havard Skinnemoen <hskinnemoen@google.com> Message-id: 20200911052101.2602693-2-hskinnemoen@google.com Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- include/hw/misc/npcm7xx_gcr.h | 43 ++++++ hw/misc/npcm7xx_gcr.c | 269 ++++++++++++++++++++++++++++++++++ MAINTAINERS | 8 + hw/arm/Kconfig | 3 + hw/misc/meson.build | 3 + hw/misc/trace-events | 4 + 6 files changed, 330 insertions(+) create mode 100644 include/hw/misc/npcm7xx_gcr.h create mode 100644 hw/misc/npcm7xx_gcr.c diff --git a/include/hw/misc/npcm7xx_gcr.h b/include/hw/misc/npcm7xx_gcr.h new file mode 100644 index XXXXXXX..XXXXXXX --- /dev/null +++ b/include/hw/misc/npcm7xx_gcr.h @@ -XXX,XX +XXX,XX @@ +/* + * Nuvoton NPCM7xx System Global Control Registers. + * + * Copyright 2020 Google LLC + * + * This program is free software; you can redistribute it and/or modify it + * under the terms of the GNU General Public License as published by the + * Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program 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 General Public License + * for more details. + */ +#ifndef NPCM7XX_GCR_H +#define NPCM7XX_GCR_H + +#include "exec/memory.h" +#include "hw/sysbus.h" + +/* + * Number of registers in our device state structure. Don't change this without + * incrementing the version_id in the vmstate. + */ +#define NPCM7XX_GCR_NR_REGS (0x148 / sizeof(uint32_t)) + +typedef struct NPCM7xxGCRState { + SysBusDevice parent; + + MemoryRegion iomem; + + uint32_t regs[NPCM7XX_GCR_NR_REGS]; + + uint32_t reset_pwron; + uint32_t reset_mdlr; + uint32_t reset_intcr3; +} NPCM7xxGCRState; + +#define TYPE_NPCM7XX_GCR "npcm7xx-gcr" +#define NPCM7XX_GCR(obj) OBJECT_CHECK(NPCM7xxGCRState, (obj), TYPE_NPCM7XX_GCR) + +#endif /* NPCM7XX_GCR_H */ diff --git a/hw/misc/npcm7xx_gcr.c b/hw/misc/npcm7xx_gcr.c new file mode 100644 index XXXXXXX..XXXXXXX --- /dev/null +++ b/hw/misc/npcm7xx_gcr.c @@ -XXX,XX +XXX,XX @@ +/* + * Nuvoton NPCM7xx System Global Control Registers. + * + * Copyright 2020 Google LLC + * + * This program is free software; you can redistribute it and/or modify it + * under the terms of the GNU General Public License as published by the + * Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program 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 General Public License + * for more details. + */ + +#include "qemu/osdep.h" + +#include "hw/misc/npcm7xx_gcr.h" +#include "hw/qdev-properties.h" +#include "migration/vmstate.h" +#include "qapi/error.h" +#include "qemu/cutils.h" +#include "qemu/log.h" +#include "qemu/module.h" +#include "qemu/units.h" + +#include "trace.h" + +#define NPCM7XX_GCR_MIN_DRAM_SIZE (128 * MiB) +#define NPCM7XX_GCR_MAX_DRAM_SIZE (2 * GiB) + +enum NPCM7xxGCRRegisters { + NPCM7XX_GCR_PDID, + NPCM7XX_GCR_PWRON, + NPCM7XX_GCR_MFSEL1 = 0x0c / sizeof(uint32_t), + NPCM7XX_GCR_MFSEL2, + NPCM7XX_GCR_MISCPE, + NPCM7XX_GCR_SPSWC = 0x038 / sizeof(uint32_t), + NPCM7XX_GCR_INTCR, + NPCM7XX_GCR_INTSR, + NPCM7XX_GCR_HIFCR = 0x050 / sizeof(uint32_t), + NPCM7XX_GCR_INTCR2 = 0x060 / sizeof(uint32_t), + NPCM7XX_GCR_MFSEL3, + NPCM7XX_GCR_SRCNT, + NPCM7XX_GCR_RESSR, + NPCM7XX_GCR_RLOCKR1, + NPCM7XX_GCR_FLOCKR1, + NPCM7XX_GCR_DSCNT, + NPCM7XX_GCR_MDLR, + NPCM7XX_GCR_SCRPAD3, + NPCM7XX_GCR_SCRPAD2, + NPCM7XX_GCR_DAVCLVLR = 0x098 / sizeof(uint32_t), + NPCM7XX_GCR_INTCR3, + NPCM7XX_GCR_VSINTR = 0x0ac / sizeof(uint32_t), + NPCM7XX_GCR_MFSEL4, + NPCM7XX_GCR_CPBPNTR = 0x0c4 / sizeof(uint32_t), + NPCM7XX_GCR_CPCTL = 0x0d0 / sizeof(uint32_t), + NPCM7XX_GCR_CP2BST, + NPCM7XX_GCR_B2CPNT, + NPCM7XX_GCR_CPPCTL, + NPCM7XX_GCR_I2CSEGSEL, + NPCM7XX_GCR_I2CSEGCTL, + NPCM7XX_GCR_VSRCR, + NPCM7XX_GCR_MLOCKR, + NPCM7XX_GCR_SCRPAD = 0x013c / sizeof(uint32_t), + NPCM7XX_GCR_USB1PHYCTL, + NPCM7XX_GCR_USB2PHYCTL, + NPCM7XX_GCR_REGS_END, +}; + +static const uint32_t cold_reset_values[NPCM7XX_GCR_NR_REGS] = { + [NPCM7XX_GCR_PDID] = 0x04a92750, /* Poleg A1 */ + [NPCM7XX_GCR_MISCPE] = 0x0000ffff, + [NPCM7XX_GCR_SPSWC] = 0x00000003, + [NPCM7XX_GCR_INTCR] = 0x0000035e, + [NPCM7XX_GCR_HIFCR] = 0x0000004e, + [NPCM7XX_GCR_INTCR2] = (1U << 19), /* DDR initialized */ + [NPCM7XX_GCR_RESSR] = 0x80000000, + [NPCM7XX_GCR_DSCNT] = 0x000000c0, + [NPCM7XX_GCR_DAVCLVLR] = 0x5a00f3cf, + [NPCM7XX_GCR_SCRPAD] = 0x00000008, + [NPCM7XX_GCR_USB1PHYCTL] = 0x034730e4, + [NPCM7XX_GCR_USB2PHYCTL] = 0x034730e4, +}; + +static uint64_t npcm7xx_gcr_read(void *opaque, hwaddr offset, unsigned size) +{ + uint32_t reg = offset / sizeof(uint32_t); + NPCM7xxGCRState *s = opaque; + + if (reg >= NPCM7XX_GCR_NR_REGS) { + qemu_log_mask(LOG_GUEST_ERROR, + "%s: offset 0x%04" HWADDR_PRIx " out of range\n", + __func__, offset); + return 0; + } + + trace_npcm7xx_gcr_read(offset, s->regs[reg]); + + return s->regs[reg]; +} + +static void npcm7xx_gcr_write(void *opaque, hwaddr offset, + uint64_t v, unsigned size) +{ + uint32_t reg = offset / sizeof(uint32_t); + NPCM7xxGCRState *s = opaque; + uint32_t value = v; + + trace_npcm7xx_gcr_write(offset, value); + + if (reg >= NPCM7XX_GCR_NR_REGS) { + qemu_log_mask(LOG_GUEST_ERROR, + "%s: offset 0x%04" HWADDR_PRIx " out of range\n", + __func__, offset); + return; + } + + switch (reg) { + case NPCM7XX_GCR_PDID: + case NPCM7XX_GCR_PWRON: + case NPCM7XX_GCR_INTSR: + qemu_log_mask(LOG_GUEST_ERROR, + "%s: register @ 0x%04" HWADDR_PRIx " is read-only\n", + __func__, offset); + return; + + case NPCM7XX_GCR_RESSR: + case NPCM7XX_GCR_CP2BST: + /* Write 1 to clear */ + value = s->regs[reg] & ~value; + break; + + case NPCM7XX_GCR_RLOCKR1: + case NPCM7XX_GCR_MDLR: + /* Write 1 to set */ + value |= s->regs[reg]; + break; + }; + + s->regs[reg] = value; +} + +static const struct MemoryRegionOps npcm7xx_gcr_ops = { + .read = npcm7xx_gcr_read, + .write = npcm7xx_gcr_write, + .endianness = DEVICE_LITTLE_ENDIAN, + .valid = { + .min_access_size = 4, + .max_access_size = 4, + .unaligned = false, + }, +}; + +static void npcm7xx_gcr_enter_reset(Object *obj, ResetType type) +{ + NPCM7xxGCRState *s = NPCM7XX_GCR(obj); + + QEMU_BUILD_BUG_ON(sizeof(s->regs) != sizeof(cold_reset_values)); + + switch (type) { + case RESET_TYPE_COLD: + memcpy(s->regs, cold_reset_values, sizeof(s->regs)); + s->regs[NPCM7XX_GCR_PWRON] = s->reset_pwron; + s->regs[NPCM7XX_GCR_MDLR] = s->reset_mdlr; + s->regs[NPCM7XX_GCR_INTCR3] = s->reset_intcr3; + break; + } +} + +static void npcm7xx_gcr_realize(DeviceState *dev, Error **errp) +{ + ERRP_GUARD(); + NPCM7xxGCRState *s = NPCM7XX_GCR(dev); + uint64_t dram_size; + Object *obj; + + obj = object_property_get_link(OBJECT(dev), "dram-mr", errp); + if (!obj) { + error_prepend(errp, "%s: required dram-mr link not found: ", __func__); + return; + } + dram_size = memory_region_size(MEMORY_REGION(obj)); + if (!is_power_of_2(dram_size) || + dram_size < NPCM7XX_GCR_MIN_DRAM_SIZE || + dram_size > NPCM7XX_GCR_MAX_DRAM_SIZE) { + g_autofree char *sz = size_to_str(dram_size); + g_autofree char *min_sz = size_to_str(NPCM7XX_GCR_MIN_DRAM_SIZE); + g_autofree char *max_sz = size_to_str(NPCM7XX_GCR_MAX_DRAM_SIZE); + error_setg(errp, "%s: unsupported DRAM size %s", __func__, sz); + error_append_hint(errp, + "DRAM size must be a power of two between %s and %s," + " inclusive.\n", min_sz, max_sz); + return; + } + + /* Power-on reset value */ + s->reset_intcr3 = 0x00001002; + + /* + * The GMMAP (Graphics Memory Map) field is used by u-boot to detect the + * DRAM size, and is normally initialized by the boot block as part of DRAM + * training. However, since we don't have a complete emulation of the + * memory controller and try to make it look like it has already been + * initialized, the boot block will skip this initialization, and we need + * to make sure this field is set correctly up front. + * + * WARNING: some versions of u-boot only looks at bits 8 and 9, so 2 GiB of + * DRAM will be interpreted as 128 MiB. + * + * https://github.com/Nuvoton-Israel/u-boot/blob/2aef993bd2aafeb5408dbaad0f3ce099ee40c4aa/board/nuvoton/poleg/poleg.c#L244 + */ + s->reset_intcr3 |= ctz64(dram_size / NPCM7XX_GCR_MIN_DRAM_SIZE) << 8; +} + +static void npcm7xx_gcr_init(Object *obj) +{ + NPCM7xxGCRState *s = NPCM7XX_GCR(obj); + + memory_region_init_io(&s->iomem, obj, &npcm7xx_gcr_ops, s, + TYPE_NPCM7XX_GCR, 4 * KiB); + sysbus_init_mmio(&s->parent, &s->iomem); +} + +static const VMStateDescription vmstate_npcm7xx_gcr = { + .name = "npcm7xx-gcr", + .version_id = 0, + .minimum_version_id = 0, + .fields = (VMStateField[]) { + VMSTATE_UINT32_ARRAY(regs, NPCM7xxGCRState, NPCM7XX_GCR_NR_REGS), + VMSTATE_END_OF_LIST(), + }, +}; + +static Property npcm7xx_gcr_properties[] = { + DEFINE_PROP_UINT32("disabled-modules", NPCM7xxGCRState, reset_mdlr, 0), + DEFINE_PROP_UINT32("power-on-straps", NPCM7xxGCRState, reset_pwron, 0), + DEFINE_PROP_END_OF_LIST(), +}; + +static void npcm7xx_gcr_class_init(ObjectClass *klass, void *data) +{ + ResettableClass *rc = RESETTABLE_CLASS(klass); + DeviceClass *dc = DEVICE_CLASS(klass); + + QEMU_BUILD_BUG_ON(NPCM7XX_GCR_REGS_END > NPCM7XX_GCR_NR_REGS); + + dc->desc = "NPCM7xx System Global Control Registers"; + dc->realize = npcm7xx_gcr_realize; + dc->vmsd = &vmstate_npcm7xx_gcr; + rc->phases.enter = npcm7xx_gcr_enter_reset; + + device_class_set_props(dc, npcm7xx_gcr_properties); +} + +static const TypeInfo npcm7xx_gcr_info = { + .name = TYPE_NPCM7XX_GCR, + .parent = TYPE_SYS_BUS_DEVICE, + .instance_size = sizeof(NPCM7xxGCRState), + .instance_init = npcm7xx_gcr_init, + .class_init = npcm7xx_gcr_class_init, +}; + +static void npcm7xx_gcr_register_type(void) +{ + type_register_static(&npcm7xx_gcr_info); +} +type_init(npcm7xx_gcr_register_type); diff --git a/MAINTAINERS b/MAINTAINERS index XXXXXXX..XXXXXXX 100644 --- a/MAINTAINERS +++ b/MAINTAINERS @@ -XXX,XX +XXX,XX @@ S: Odd Fixes F: hw/arm/musicpal.c F: docs/system/arm/musicpal.rst +Nuvoton NPCM7xx +M: Havard Skinnemoen <hskinnemoen@google.com> +M: Tyrone Ting <kfting@nuvoton.com> +L: qemu-arm@nongnu.org +S: Supported +F: hw/*/npcm7xx* +F: include/hw/*/npcm7xx* + nSeries M: Andrzej Zaborowski <balrogg@gmail.com> M: Peter Maydell <peter.maydell@linaro.org> diff --git a/hw/arm/Kconfig b/hw/arm/Kconfig index XXXXXXX..XXXXXXX 100644 --- a/hw/arm/Kconfig +++ b/hw/arm/Kconfig @@ -XXX,XX +XXX,XX @@ config XLNX_VERSAL select VIRTIO_MMIO select UNIMP +config NPCM7XX + bool + config FSL_IMX25 bool select IMX diff --git a/hw/misc/meson.build b/hw/misc/meson.build index XXXXXXX..XXXXXXX 100644 --- a/hw/misc/meson.build +++ b/hw/misc/meson.build @@ -XXX,XX +XXX,XX @@ softmmu_ss.add(when: 'CONFIG_IMX', if_true: files( )) softmmu_ss.add(when: 'CONFIG_MILKYMIST', if_true: files('milkymist-hpdmc.c', 'milkymist-pfpu.c')) softmmu_ss.add(when: 'CONFIG_MAINSTONE', if_true: files('mst_fpga.c')) +softmmu_ss.add(when: 'CONFIG_NPCM7XX', if_true: files( + 'npcm7xx_gcr.c', +)) softmmu_ss.add(when: 'CONFIG_OMAP', if_true: files( 'omap_clk.c', 'omap_gpmc.c', 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 @@ mos6522_set_sr_int(void) "set sr_int" mos6522_write(uint64_t addr, uint64_t val) "reg=0x%"PRIx64 " val=0x%"PRIx64 mos6522_read(uint64_t addr, unsigned val) "reg=0x%"PRIx64 " val=0x%x" +# npcm7xx_gcr.c +npcm7xx_gcr_read(uint64_t offset, uint32_t value) " offset: 0x%04" PRIx64 " value: 0x%08" PRIx32 +npcm7xx_gcr_write(uint64_t offset, uint32_t value) "offset: 0x%04" PRIx64 " value: 0x%08" PRIx32 + # stm32f4xx_syscfg.c stm32f4xx_syscfg_set_irq(int gpio, int line, int level) "Interupt: GPIO: %d, Line: %d; Level: %d" stm32f4xx_pulse_exti(int irq) "Pulse EXTI: %d" -- 2.20.1 From: Havard Skinnemoen <hskinnemoen@google.com> Enough functionality to boot the Linux kernel has been implemented. This includes: - Correct power-on reset values so the various clock rates can be accurately calculated. - Clock enables stick around when written. In addition, a best effort attempt to implement SECCNT and CNTR25M was made even though I don't think the kernel needs them. Reviewed-by: Tyrone Ting <kfting@nuvoton.com> Reviewed-by: Joel Stanley <joel@jms.id.au> Reviewed-by: Cédric Le Goater <clg@kaod.org> Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Tested-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Tested-by: Alexander Bulekov <alxndr@bu.edu> Signed-off-by: Havard Skinnemoen <hskinnemoen@google.com> Message-id: 20200911052101.2602693-3-hskinnemoen@google.com Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- include/hw/misc/npcm7xx_clk.h | 48 ++++++ hw/misc/npcm7xx_clk.c | 266 ++++++++++++++++++++++++++++++++++ hw/misc/meson.build | 1 + hw/misc/trace-events | 4 + 4 files changed, 319 insertions(+) create mode 100644 include/hw/misc/npcm7xx_clk.h create mode 100644 hw/misc/npcm7xx_clk.c diff --git a/include/hw/misc/npcm7xx_clk.h b/include/hw/misc/npcm7xx_clk.h new file mode 100644 index XXXXXXX..XXXXXXX --- /dev/null +++ b/include/hw/misc/npcm7xx_clk.h @@ -XXX,XX +XXX,XX @@ +/* + * Nuvoton NPCM7xx Clock Control Registers. + * + * Copyright 2020 Google LLC + * + * This program is free software; you can redistribute it and/or modify it + * under the terms of the GNU General Public License as published by the + * Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program 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 General Public License + * for more details. + */ +#ifndef NPCM7XX_CLK_H +#define NPCM7XX_CLK_H + +#include "exec/memory.h" +#include "hw/sysbus.h" + +/* + * The reference clock frequency for the timer modules, and the SECCNT and + * CNTR25M registers in this module, is always 25 MHz. + */ +#define NPCM7XX_TIMER_REF_HZ (25000000) + +/* + * Number of registers in our device state structure. Don't change this without + * incrementing the version_id in the vmstate. + */ +#define NPCM7XX_CLK_NR_REGS (0x70 / sizeof(uint32_t)) + +typedef struct NPCM7xxCLKState { + SysBusDevice parent; + + MemoryRegion iomem; + + uint32_t regs[NPCM7XX_CLK_NR_REGS]; + + /* Time reference for SECCNT and CNTR25M, initialized by power on reset */ + int64_t ref_ns; +} NPCM7xxCLKState; + +#define TYPE_NPCM7XX_CLK "npcm7xx-clk" +#define NPCM7XX_CLK(obj) OBJECT_CHECK(NPCM7xxCLKState, (obj), TYPE_NPCM7XX_CLK) + +#endif /* NPCM7XX_CLK_H */ diff --git a/hw/misc/npcm7xx_clk.c b/hw/misc/npcm7xx_clk.c new file mode 100644 index XXXXXXX..XXXXXXX --- /dev/null +++ b/hw/misc/npcm7xx_clk.c @@ -XXX,XX +XXX,XX @@ +/* + * Nuvoton NPCM7xx Clock Control Registers. + * + * Copyright 2020 Google LLC + * + * This program is free software; you can redistribute it and/or modify it + * under the terms of the GNU General Public License as published by the + * Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program 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 General Public License + * for more details. + */ + +#include "qemu/osdep.h" + +#include "hw/misc/npcm7xx_clk.h" +#include "migration/vmstate.h" +#include "qemu/error-report.h" +#include "qemu/log.h" +#include "qemu/module.h" +#include "qemu/timer.h" +#include "qemu/units.h" +#include "trace.h" + +#define PLLCON_LOKI BIT(31) +#define PLLCON_LOKS BIT(30) +#define PLLCON_PWDEN BIT(12) + +enum NPCM7xxCLKRegisters { + NPCM7XX_CLK_CLKEN1, + NPCM7XX_CLK_CLKSEL, + NPCM7XX_CLK_CLKDIV1, + NPCM7XX_CLK_PLLCON0, + NPCM7XX_CLK_PLLCON1, + NPCM7XX_CLK_SWRSTR, + NPCM7XX_CLK_IPSRST1 = 0x20 / sizeof(uint32_t), + NPCM7XX_CLK_IPSRST2, + NPCM7XX_CLK_CLKEN2, + NPCM7XX_CLK_CLKDIV2, + NPCM7XX_CLK_CLKEN3, + NPCM7XX_CLK_IPSRST3, + NPCM7XX_CLK_WD0RCR, + NPCM7XX_CLK_WD1RCR, + NPCM7XX_CLK_WD2RCR, + NPCM7XX_CLK_SWRSTC1, + NPCM7XX_CLK_SWRSTC2, + NPCM7XX_CLK_SWRSTC3, + NPCM7XX_CLK_SWRSTC4, + NPCM7XX_CLK_PLLCON2, + NPCM7XX_CLK_CLKDIV3, + NPCM7XX_CLK_CORSTC, + NPCM7XX_CLK_PLLCONG, + NPCM7XX_CLK_AHBCKFI, + NPCM7XX_CLK_SECCNT, + NPCM7XX_CLK_CNTR25M, + NPCM7XX_CLK_REGS_END, +}; + +/* + * These reset values were taken from version 0.91 of the NPCM750R data sheet. + * + * All are loaded on power-up reset. CLKENx and SWRSTR should also be loaded on + * core domain reset, but this reset type is not yet supported by QEMU. + */ +static const uint32_t cold_reset_values[NPCM7XX_CLK_NR_REGS] = { + [NPCM7XX_CLK_CLKEN1] = 0xffffffff, + [NPCM7XX_CLK_CLKSEL] = 0x004aaaaa, + [NPCM7XX_CLK_CLKDIV1] = 0x5413f855, + [NPCM7XX_CLK_PLLCON0] = 0x00222101 | PLLCON_LOKI, + [NPCM7XX_CLK_PLLCON1] = 0x00202101 | PLLCON_LOKI, + [NPCM7XX_CLK_IPSRST1] = 0x00001000, + [NPCM7XX_CLK_IPSRST2] = 0x80000000, + [NPCM7XX_CLK_CLKEN2] = 0xffffffff, + [NPCM7XX_CLK_CLKDIV2] = 0xaa4f8f9f, + [NPCM7XX_CLK_CLKEN3] = 0xffffffff, + [NPCM7XX_CLK_IPSRST3] = 0x03000000, + [NPCM7XX_CLK_WD0RCR] = 0xffffffff, + [NPCM7XX_CLK_WD1RCR] = 0xffffffff, + [NPCM7XX_CLK_WD2RCR] = 0xffffffff, + [NPCM7XX_CLK_SWRSTC1] = 0x00000003, + [NPCM7XX_CLK_PLLCON2] = 0x00c02105 | PLLCON_LOKI, + [NPCM7XX_CLK_CORSTC] = 0x04000003, + [NPCM7XX_CLK_PLLCONG] = 0x01228606 | PLLCON_LOKI, + [NPCM7XX_CLK_AHBCKFI] = 0x000000c8, +}; + +static uint64_t npcm7xx_clk_read(void *opaque, hwaddr offset, unsigned size) +{ + uint32_t reg = offset / sizeof(uint32_t); + NPCM7xxCLKState *s = opaque; + int64_t now_ns; + uint32_t value = 0; + + if (reg >= NPCM7XX_CLK_NR_REGS) { + qemu_log_mask(LOG_GUEST_ERROR, + "%s: offset 0x%04" HWADDR_PRIx " out of range\n", + __func__, offset); + return 0; + } + + switch (reg) { + case NPCM7XX_CLK_SWRSTR: + qemu_log_mask(LOG_GUEST_ERROR, + "%s: register @ 0x%04" HWADDR_PRIx " is write-only\n", + __func__, offset); + break; + + case NPCM7XX_CLK_SECCNT: + now_ns = qemu_clock_get_ns(QEMU_CLOCK_VIRTUAL); + value = (now_ns - s->ref_ns) / NANOSECONDS_PER_SECOND; + break; + + case NPCM7XX_CLK_CNTR25M: + now_ns = qemu_clock_get_ns(QEMU_CLOCK_VIRTUAL); + /* + * This register counts 25 MHz cycles, updating every 640 ns. It rolls + * over to zero every second. + * + * The 4 LSBs are always zero: (1e9 / 640) << 4 = 25000000. + */ + value = (((now_ns - s->ref_ns) / 640) << 4) % NPCM7XX_TIMER_REF_HZ; + break; + + default: + value = s->regs[reg]; + break; + }; + + trace_npcm7xx_clk_read(offset, value); + + return value; +} + +static void npcm7xx_clk_write(void *opaque, hwaddr offset, + uint64_t v, unsigned size) +{ + uint32_t reg = offset / sizeof(uint32_t); + NPCM7xxCLKState *s = opaque; + uint32_t value = v; + + trace_npcm7xx_clk_write(offset, value); + + if (reg >= NPCM7XX_CLK_NR_REGS) { + qemu_log_mask(LOG_GUEST_ERROR, + "%s: offset 0x%04" HWADDR_PRIx " out of range\n", + __func__, offset); + return; + } + + switch (reg) { + case NPCM7XX_CLK_SWRSTR: + qemu_log_mask(LOG_UNIMP, "%s: SW reset not implemented: 0x%02x\n", + __func__, value); + value = 0; + break; + + case NPCM7XX_CLK_PLLCON0: + case NPCM7XX_CLK_PLLCON1: + case NPCM7XX_CLK_PLLCON2: + case NPCM7XX_CLK_PLLCONG: + if (value & PLLCON_PWDEN) { + /* Power down -- clear lock and indicate loss of lock */ + value &= ~PLLCON_LOKI; + value |= PLLCON_LOKS; + } else { + /* Normal mode -- assume always locked */ + value |= PLLCON_LOKI; + /* Keep LOKS unchanged unless cleared by writing 1 */ + if (value & PLLCON_LOKS) { + value &= ~PLLCON_LOKS; + } else { + value |= (value & PLLCON_LOKS); + } + } + break; + + case NPCM7XX_CLK_CNTR25M: + qemu_log_mask(LOG_GUEST_ERROR, + "%s: register @ 0x%04" HWADDR_PRIx " is read-only\n", + __func__, offset); + return; + } + + s->regs[reg] = value; +} + +static const struct MemoryRegionOps npcm7xx_clk_ops = { + .read = npcm7xx_clk_read, + .write = npcm7xx_clk_write, + .endianness = DEVICE_LITTLE_ENDIAN, + .valid = { + .min_access_size = 4, + .max_access_size = 4, + .unaligned = false, + }, +}; + +static void npcm7xx_clk_enter_reset(Object *obj, ResetType type) +{ + NPCM7xxCLKState *s = NPCM7XX_CLK(obj); + + QEMU_BUILD_BUG_ON(sizeof(s->regs) != sizeof(cold_reset_values)); + + switch (type) { + case RESET_TYPE_COLD: + memcpy(s->regs, cold_reset_values, sizeof(cold_reset_values)); + s->ref_ns = qemu_clock_get_ns(QEMU_CLOCK_VIRTUAL); + return; + } + + /* + * A small number of registers need to be reset on a core domain reset, + * but no such reset type exists yet. + */ + qemu_log_mask(LOG_UNIMP, "%s: reset type %d not implemented.", + __func__, type); +} + +static void npcm7xx_clk_init(Object *obj) +{ + NPCM7xxCLKState *s = NPCM7XX_CLK(obj); + + memory_region_init_io(&s->iomem, obj, &npcm7xx_clk_ops, s, + TYPE_NPCM7XX_CLK, 4 * KiB); + sysbus_init_mmio(&s->parent, &s->iomem); +} + +static const VMStateDescription vmstate_npcm7xx_clk = { + .name = "npcm7xx-clk", + .version_id = 0, + .minimum_version_id = 0, + .fields = (VMStateField[]) { + VMSTATE_UINT32_ARRAY(regs, NPCM7xxCLKState, NPCM7XX_CLK_NR_REGS), + VMSTATE_INT64(ref_ns, NPCM7xxCLKState), + VMSTATE_END_OF_LIST(), + }, +}; + +static void npcm7xx_clk_class_init(ObjectClass *klass, void *data) +{ + ResettableClass *rc = RESETTABLE_CLASS(klass); + DeviceClass *dc = DEVICE_CLASS(klass); + + QEMU_BUILD_BUG_ON(NPCM7XX_CLK_REGS_END > NPCM7XX_CLK_NR_REGS); + + dc->desc = "NPCM7xx Clock Control Registers"; + dc->vmsd = &vmstate_npcm7xx_clk; + rc->phases.enter = npcm7xx_clk_enter_reset; +} + +static const TypeInfo npcm7xx_clk_info = { + .name = TYPE_NPCM7XX_CLK, + .parent = TYPE_SYS_BUS_DEVICE, + .instance_size = sizeof(NPCM7xxCLKState), + .instance_init = npcm7xx_clk_init, + .class_init = npcm7xx_clk_class_init, +}; + +static void npcm7xx_clk_register_type(void) +{ + type_register_static(&npcm7xx_clk_info); +} +type_init(npcm7xx_clk_register_type); diff --git a/hw/misc/meson.build b/hw/misc/meson.build index XXXXXXX..XXXXXXX 100644 --- a/hw/misc/meson.build +++ b/hw/misc/meson.build @@ -XXX,XX +XXX,XX @@ softmmu_ss.add(when: 'CONFIG_IMX', if_true: files( softmmu_ss.add(when: 'CONFIG_MILKYMIST', if_true: files('milkymist-hpdmc.c', 'milkymist-pfpu.c')) softmmu_ss.add(when: 'CONFIG_MAINSTONE', if_true: files('mst_fpga.c')) softmmu_ss.add(when: 'CONFIG_NPCM7XX', if_true: files( + 'npcm7xx_clk.c', 'npcm7xx_gcr.c', )) softmmu_ss.add(when: 'CONFIG_OMAP', if_true: files( 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 @@ mos6522_set_sr_int(void) "set sr_int" mos6522_write(uint64_t addr, uint64_t val) "reg=0x%"PRIx64 " val=0x%"PRIx64 mos6522_read(uint64_t addr, unsigned val) "reg=0x%"PRIx64 " val=0x%x" +# npcm7xx_clk.c +npcm7xx_clk_read(uint64_t offset, uint32_t value) " offset: 0x%04" PRIx64 " value: 0x%08" PRIx32 +npcm7xx_clk_write(uint64_t offset, uint32_t value) "offset: 0x%04" PRIx64 " value: 0x%08" PRIx32 + # npcm7xx_gcr.c npcm7xx_gcr_read(uint64_t offset, uint32_t value) " offset: 0x%04" PRIx64 " value: 0x%08" PRIx32 npcm7xx_gcr_write(uint64_t offset, uint32_t value) "offset: 0x%04" PRIx64 " value: 0x%08" PRIx32 -- 2.20.1 From: Havard Skinnemoen <hskinnemoen@google.com> The NPCM730 and NPCM750 SoCs have three timer modules each holding five timers and some shared registers (e.g. interrupt status). Each timer runs at 25 MHz divided by a prescaler, and counts down from a configurable initial value to zero. When zero is reached, the interrupt flag for the timer is set, and the timer is disabled (one-shot mode) or reloaded from its initial value (periodic mode). This implementation is sufficient to boot a Linux kernel configured for NPCM750. Note that the kernel does not seem to actually turn on the interrupts. Reviewed-by: Tyrone Ting <kfting@nuvoton.com> Reviewed-by: Joel Stanley <joel@jms.id.au> Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Tested-by: Alexander Bulekov <alxndr@bu.edu> Tested-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Signed-off-by: Havard Skinnemoen <hskinnemoen@google.com> Message-id: 20200911052101.2602693-4-hskinnemoen@google.com Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- include/hw/timer/npcm7xx_timer.h | 78 +++++ hw/timer/npcm7xx_timer.c | 543 +++++++++++++++++++++++++++++++ hw/timer/meson.build | 1 + hw/timer/trace-events | 5 + 4 files changed, 627 insertions(+) create mode 100644 include/hw/timer/npcm7xx_timer.h create mode 100644 hw/timer/npcm7xx_timer.c diff --git a/include/hw/timer/npcm7xx_timer.h b/include/hw/timer/npcm7xx_timer.h new file mode 100644 index XXXXXXX..XXXXXXX --- /dev/null +++ b/include/hw/timer/npcm7xx_timer.h @@ -XXX,XX +XXX,XX @@ +/* + * Nuvoton NPCM7xx Timer Controller + * + * Copyright 2020 Google LLC + * + * This program is free software; you can redistribute it and/or modify it + * under the terms of the GNU General Public License as published by the + * Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program 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 General Public License + * for more details. + */ +#ifndef NPCM7XX_TIMER_H +#define NPCM7XX_TIMER_H + +#include "exec/memory.h" +#include "hw/sysbus.h" +#include "qemu/timer.h" + +/* Each Timer Module (TIM) instance holds five 25 MHz timers. */ +#define NPCM7XX_TIMERS_PER_CTRL (5) + +/* + * Number of registers in our device state structure. Don't change this without + * incrementing the version_id in the vmstate. + */ +#define NPCM7XX_TIMER_NR_REGS (0x54 / sizeof(uint32_t)) + +typedef struct NPCM7xxTimerCtrlState NPCM7xxTimerCtrlState; + +/** + * struct NPCM7xxTimer - Individual timer state. + * @irq: GIC interrupt line to fire on expiration (if enabled). + * @qtimer: QEMU timer that notifies us on expiration. + * @expires_ns: Absolute virtual expiration time. + * @remaining_ns: Remaining time until expiration if timer is paused. + * @tcsr: The Timer Control and Status Register. + * @ticr: The Timer Initial Count Register. + */ +typedef struct NPCM7xxTimer { + NPCM7xxTimerCtrlState *ctrl; + + qemu_irq irq; + QEMUTimer qtimer; + int64_t expires_ns; + int64_t remaining_ns; + + uint32_t tcsr; + uint32_t ticr; +} NPCM7xxTimer; + +/** + * struct NPCM7xxTimerCtrlState - Timer Module device state. + * @parent: System bus device. + * @iomem: Memory region through which registers are accessed. + * @tisr: The Timer Interrupt Status Register. + * @wtcr: The Watchdog Timer Control Register. + * @timer: The five individual timers managed by this module. + */ +struct NPCM7xxTimerCtrlState { + SysBusDevice parent; + + MemoryRegion iomem; + + uint32_t tisr; + uint32_t wtcr; + + NPCM7xxTimer timer[NPCM7XX_TIMERS_PER_CTRL]; +}; + +#define TYPE_NPCM7XX_TIMER "npcm7xx-timer" +#define NPCM7XX_TIMER(obj) \ + OBJECT_CHECK(NPCM7xxTimerCtrlState, (obj), TYPE_NPCM7XX_TIMER) + +#endif /* NPCM7XX_TIMER_H */ diff --git a/hw/timer/npcm7xx_timer.c b/hw/timer/npcm7xx_timer.c new file mode 100644 index XXXXXXX..XXXXXXX --- /dev/null +++ b/hw/timer/npcm7xx_timer.c @@ -XXX,XX +XXX,XX @@ +/* + * Nuvoton NPCM7xx Timer Controller + * + * Copyright 2020 Google LLC + * + * This program is free software; you can redistribute it and/or modify it + * under the terms of the GNU General Public License as published by the + * Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program 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 General Public License + * for more details. + */ + +#include "qemu/osdep.h" + +#include "hw/irq.h" +#include "hw/misc/npcm7xx_clk.h" +#include "hw/timer/npcm7xx_timer.h" +#include "migration/vmstate.h" +#include "qemu/bitops.h" +#include "qemu/error-report.h" +#include "qemu/log.h" +#include "qemu/module.h" +#include "qemu/timer.h" +#include "qemu/units.h" +#include "trace.h" + +/* 32-bit register indices. */ +enum NPCM7xxTimerRegisters { + NPCM7XX_TIMER_TCSR0, + NPCM7XX_TIMER_TCSR1, + NPCM7XX_TIMER_TICR0, + NPCM7XX_TIMER_TICR1, + NPCM7XX_TIMER_TDR0, + NPCM7XX_TIMER_TDR1, + NPCM7XX_TIMER_TISR, + NPCM7XX_TIMER_WTCR, + NPCM7XX_TIMER_TCSR2, + NPCM7XX_TIMER_TCSR3, + NPCM7XX_TIMER_TICR2, + NPCM7XX_TIMER_TICR3, + NPCM7XX_TIMER_TDR2, + NPCM7XX_TIMER_TDR3, + NPCM7XX_TIMER_TCSR4 = 0x0040 / sizeof(uint32_t), + NPCM7XX_TIMER_TICR4 = 0x0048 / sizeof(uint32_t), + NPCM7XX_TIMER_TDR4 = 0x0050 / sizeof(uint32_t), + NPCM7XX_TIMER_REGS_END, +}; + +/* Register field definitions. */ +#define NPCM7XX_TCSR_CEN BIT(30) +#define NPCM7XX_TCSR_IE BIT(29) +#define NPCM7XX_TCSR_PERIODIC BIT(27) +#define NPCM7XX_TCSR_CRST BIT(26) +#define NPCM7XX_TCSR_CACT BIT(25) +#define NPCM7XX_TCSR_RSVD 0x01ffff00 +#define NPCM7XX_TCSR_PRESCALE_START 0 +#define NPCM7XX_TCSR_PRESCALE_LEN 8 + +/* + * Returns the index of timer in the tc->timer array. This can be used to + * locate the registers that belong to this timer. + */ +static int npcm7xx_timer_index(NPCM7xxTimerCtrlState *tc, NPCM7xxTimer *timer) +{ + int index = timer - tc->timer; + + g_assert(index >= 0 && index < NPCM7XX_TIMERS_PER_CTRL); + + return index; +} + +/* Return the value by which to divide the reference clock rate. */ +static uint32_t npcm7xx_tcsr_prescaler(uint32_t tcsr) +{ + return extract32(tcsr, NPCM7XX_TCSR_PRESCALE_START, + NPCM7XX_TCSR_PRESCALE_LEN) + 1; +} + +/* Convert a timer cycle count to a time interval in nanoseconds. */ +static int64_t npcm7xx_timer_count_to_ns(NPCM7xxTimer *t, uint32_t count) +{ + int64_t ns = count; + + ns *= NANOSECONDS_PER_SECOND / NPCM7XX_TIMER_REF_HZ; + ns *= npcm7xx_tcsr_prescaler(t->tcsr); + + return ns; +} + +/* Convert a time interval in nanoseconds to a timer cycle count. */ +static uint32_t npcm7xx_timer_ns_to_count(NPCM7xxTimer *t, int64_t ns) +{ + int64_t count; + + count = ns / (NANOSECONDS_PER_SECOND / NPCM7XX_TIMER_REF_HZ); + count /= npcm7xx_tcsr_prescaler(t->tcsr); + + return count; +} + +/* + * Raise the interrupt line if there's a pending interrupt and interrupts are + * enabled for this timer. If not, lower it. + */ +static void npcm7xx_timer_check_interrupt(NPCM7xxTimer *t) +{ + NPCM7xxTimerCtrlState *tc = t->ctrl; + int index = npcm7xx_timer_index(tc, t); + bool pending = (t->tcsr & NPCM7XX_TCSR_IE) && (tc->tisr & BIT(index)); + + qemu_set_irq(t->irq, pending); + trace_npcm7xx_timer_irq(DEVICE(tc)->canonical_path, index, pending); +} + +/* Start or resume the timer. */ +static void npcm7xx_timer_start(NPCM7xxTimer *t) +{ + int64_t now; + + now = qemu_clock_get_ns(QEMU_CLOCK_VIRTUAL); + t->expires_ns = now + t->remaining_ns; + timer_mod(&t->qtimer, t->expires_ns); +} + +/* + * Called when the counter reaches zero. Sets the interrupt flag, and either + * restarts or disables the timer. + */ +static void npcm7xx_timer_reached_zero(NPCM7xxTimer *t) +{ + NPCM7xxTimerCtrlState *tc = t->ctrl; + int index = npcm7xx_timer_index(tc, t); + + tc->tisr |= BIT(index); + + if (t->tcsr & NPCM7XX_TCSR_PERIODIC) { + t->remaining_ns = npcm7xx_timer_count_to_ns(t, t->ticr); + if (t->tcsr & NPCM7XX_TCSR_CEN) { + npcm7xx_timer_start(t); + } + } else { + t->tcsr &= ~(NPCM7XX_TCSR_CEN | NPCM7XX_TCSR_CACT); + } + + npcm7xx_timer_check_interrupt(t); +} + +/* Stop counting. Record the time remaining so we can continue later. */ +static void npcm7xx_timer_pause(NPCM7xxTimer *t) +{ + int64_t now; + + timer_del(&t->qtimer); + now = qemu_clock_get_ns(QEMU_CLOCK_VIRTUAL); + t->remaining_ns = t->expires_ns - now; + if (t->remaining_ns <= 0) { + npcm7xx_timer_reached_zero(t); + } +} + +/* + * Restart the timer from its initial value. If the timer was enabled and stays + * enabled, adjust the QEMU timer according to the new count. If the timer is + * transitioning from disabled to enabled, the caller is expected to start the + * timer later. + */ +static void npcm7xx_timer_restart(NPCM7xxTimer *t, uint32_t old_tcsr) +{ + t->remaining_ns = npcm7xx_timer_count_to_ns(t, t->ticr); + + if (old_tcsr & t->tcsr & NPCM7XX_TCSR_CEN) { + npcm7xx_timer_start(t); + } +} + +/* Register read and write handlers */ + +static uint32_t npcm7xx_timer_read_tdr(NPCM7xxTimer *t) +{ + if (t->tcsr & NPCM7XX_TCSR_CEN) { + int64_t now = qemu_clock_get_ns(QEMU_CLOCK_VIRTUAL); + + return npcm7xx_timer_ns_to_count(t, t->expires_ns - now); + } + + return npcm7xx_timer_ns_to_count(t, t->remaining_ns); +} + +static void npcm7xx_timer_write_tcsr(NPCM7xxTimer *t, uint32_t new_tcsr) +{ + uint32_t old_tcsr = t->tcsr; + uint32_t tdr; + + if (new_tcsr & NPCM7XX_TCSR_RSVD) { + qemu_log_mask(LOG_GUEST_ERROR, "%s: reserved bits in 0x%08x ignored\n", + __func__, new_tcsr); + new_tcsr &= ~NPCM7XX_TCSR_RSVD; + } + if (new_tcsr & NPCM7XX_TCSR_CACT) { + qemu_log_mask(LOG_GUEST_ERROR, "%s: read-only bits in 0x%08x ignored\n", + __func__, new_tcsr); + new_tcsr &= ~NPCM7XX_TCSR_CACT; + } + if ((new_tcsr & NPCM7XX_TCSR_CRST) && (new_tcsr & NPCM7XX_TCSR_CEN)) { + qemu_log_mask(LOG_GUEST_ERROR, + "%s: both CRST and CEN set; ignoring CEN.\n", + __func__); + new_tcsr &= ~NPCM7XX_TCSR_CEN; + } + + /* Calculate the value of TDR before potentially changing the prescaler. */ + tdr = npcm7xx_timer_read_tdr(t); + + t->tcsr = (t->tcsr & NPCM7XX_TCSR_CACT) | new_tcsr; + + if (npcm7xx_tcsr_prescaler(old_tcsr) != npcm7xx_tcsr_prescaler(new_tcsr)) { + /* Recalculate time remaining based on the current TDR value. */ + t->remaining_ns = npcm7xx_timer_count_to_ns(t, tdr); + if (old_tcsr & t->tcsr & NPCM7XX_TCSR_CEN) { + npcm7xx_timer_start(t); + } + } + + if ((old_tcsr ^ new_tcsr) & NPCM7XX_TCSR_IE) { + npcm7xx_timer_check_interrupt(t); + } + if (new_tcsr & NPCM7XX_TCSR_CRST) { + npcm7xx_timer_restart(t, old_tcsr); + t->tcsr &= ~NPCM7XX_TCSR_CRST; + } + if ((old_tcsr ^ new_tcsr) & NPCM7XX_TCSR_CEN) { + if (new_tcsr & NPCM7XX_TCSR_CEN) { + t->tcsr |= NPCM7XX_TCSR_CACT; + npcm7xx_timer_start(t); + } else { + t->tcsr &= ~NPCM7XX_TCSR_CACT; + npcm7xx_timer_pause(t); + } + } +} + +static void npcm7xx_timer_write_ticr(NPCM7xxTimer *t, uint32_t new_ticr) +{ + t->ticr = new_ticr; + + npcm7xx_timer_restart(t, t->tcsr); +} + +static void npcm7xx_timer_write_tisr(NPCM7xxTimerCtrlState *s, uint32_t value) +{ + int i; + + s->tisr &= ~value; + for (i = 0; i < ARRAY_SIZE(s->timer); i++) { + if (value & (1U << i)) { + npcm7xx_timer_check_interrupt(&s->timer[i]); + } + } +} + +static hwaddr npcm7xx_tcsr_index(hwaddr reg) +{ + switch (reg) { + case NPCM7XX_TIMER_TCSR0: + return 0; + case NPCM7XX_TIMER_TCSR1: + return 1; + case NPCM7XX_TIMER_TCSR2: + return 2; + case NPCM7XX_TIMER_TCSR3: + return 3; + case NPCM7XX_TIMER_TCSR4: + return 4; + default: + g_assert_not_reached(); + } +} + +static hwaddr npcm7xx_ticr_index(hwaddr reg) +{ + switch (reg) { + case NPCM7XX_TIMER_TICR0: + return 0; + case NPCM7XX_TIMER_TICR1: + return 1; + case NPCM7XX_TIMER_TICR2: + return 2; + case NPCM7XX_TIMER_TICR3: + return 3; + case NPCM7XX_TIMER_TICR4: + return 4; + default: + g_assert_not_reached(); + } +} + +static hwaddr npcm7xx_tdr_index(hwaddr reg) +{ + switch (reg) { + case NPCM7XX_TIMER_TDR0: + return 0; + case NPCM7XX_TIMER_TDR1: + return 1; + case NPCM7XX_TIMER_TDR2: + return 2; + case NPCM7XX_TIMER_TDR3: + return 3; + case NPCM7XX_TIMER_TDR4: + return 4; + default: + g_assert_not_reached(); + } +} + +static uint64_t npcm7xx_timer_read(void *opaque, hwaddr offset, unsigned size) +{ + NPCM7xxTimerCtrlState *s = opaque; + uint64_t value = 0; + hwaddr reg; + + reg = offset / sizeof(uint32_t); + switch (reg) { + case NPCM7XX_TIMER_TCSR0: + case NPCM7XX_TIMER_TCSR1: + case NPCM7XX_TIMER_TCSR2: + case NPCM7XX_TIMER_TCSR3: + case NPCM7XX_TIMER_TCSR4: + value = s->timer[npcm7xx_tcsr_index(reg)].tcsr; + break; + + case NPCM7XX_TIMER_TICR0: + case NPCM7XX_TIMER_TICR1: + case NPCM7XX_TIMER_TICR2: + case NPCM7XX_TIMER_TICR3: + case NPCM7XX_TIMER_TICR4: + value = s->timer[npcm7xx_ticr_index(reg)].ticr; + break; + + case NPCM7XX_TIMER_TDR0: + case NPCM7XX_TIMER_TDR1: + case NPCM7XX_TIMER_TDR2: + case NPCM7XX_TIMER_TDR3: + case NPCM7XX_TIMER_TDR4: + value = npcm7xx_timer_read_tdr(&s->timer[npcm7xx_tdr_index(reg)]); + break; + + case NPCM7XX_TIMER_TISR: + value = s->tisr; + break; + + case NPCM7XX_TIMER_WTCR: + value = s->wtcr; + break; + + default: + qemu_log_mask(LOG_GUEST_ERROR, + "%s: invalid offset 0x%04" HWADDR_PRIx "\n", + __func__, offset); + break; + } + + trace_npcm7xx_timer_read(DEVICE(s)->canonical_path, offset, value); + + return value; +} + +static void npcm7xx_timer_write(void *opaque, hwaddr offset, + uint64_t v, unsigned size) +{ + uint32_t reg = offset / sizeof(uint32_t); + NPCM7xxTimerCtrlState *s = opaque; + uint32_t value = v; + + trace_npcm7xx_timer_write(DEVICE(s)->canonical_path, offset, value); + + switch (reg) { + case NPCM7XX_TIMER_TCSR0: + case NPCM7XX_TIMER_TCSR1: + case NPCM7XX_TIMER_TCSR2: + case NPCM7XX_TIMER_TCSR3: + case NPCM7XX_TIMER_TCSR4: + npcm7xx_timer_write_tcsr(&s->timer[npcm7xx_tcsr_index(reg)], value); + return; + + case NPCM7XX_TIMER_TICR0: + case NPCM7XX_TIMER_TICR1: + case NPCM7XX_TIMER_TICR2: + case NPCM7XX_TIMER_TICR3: + case NPCM7XX_TIMER_TICR4: + npcm7xx_timer_write_ticr(&s->timer[npcm7xx_ticr_index(reg)], value); + return; + + case NPCM7XX_TIMER_TDR0: + case NPCM7XX_TIMER_TDR1: + case NPCM7XX_TIMER_TDR2: + case NPCM7XX_TIMER_TDR3: + case NPCM7XX_TIMER_TDR4: + qemu_log_mask(LOG_GUEST_ERROR, + "%s: register @ 0x%04" HWADDR_PRIx " is read-only\n", + __func__, offset); + return; + + case NPCM7XX_TIMER_TISR: + npcm7xx_timer_write_tisr(s, value); + return; + + case NPCM7XX_TIMER_WTCR: + qemu_log_mask(LOG_UNIMP, "%s: WTCR write not implemented: 0x%08x\n", + __func__, value); + return; + } + + qemu_log_mask(LOG_GUEST_ERROR, + "%s: invalid offset 0x%04" HWADDR_PRIx "\n", + __func__, offset); +} + +static const struct MemoryRegionOps npcm7xx_timer_ops = { + .read = npcm7xx_timer_read, + .write = npcm7xx_timer_write, + .endianness = DEVICE_LITTLE_ENDIAN, + .valid = { + .min_access_size = 4, + .max_access_size = 4, + .unaligned = false, + }, +}; + +/* Called when the QEMU timer expires. */ +static void npcm7xx_timer_expired(void *opaque) +{ + NPCM7xxTimer *t = opaque; + + if (t->tcsr & NPCM7XX_TCSR_CEN) { + npcm7xx_timer_reached_zero(t); + } +} + +static void npcm7xx_timer_enter_reset(Object *obj, ResetType type) +{ + NPCM7xxTimerCtrlState *s = NPCM7XX_TIMER(obj); + int i; + + for (i = 0; i < NPCM7XX_TIMERS_PER_CTRL; i++) { + NPCM7xxTimer *t = &s->timer[i]; + + timer_del(&t->qtimer); + t->expires_ns = 0; + t->remaining_ns = 0; + t->tcsr = 0x00000005; + t->ticr = 0x00000000; + } + + s->tisr = 0x00000000; + s->wtcr = 0x00000400; +} + +static void npcm7xx_timer_hold_reset(Object *obj) +{ + NPCM7xxTimerCtrlState *s = NPCM7XX_TIMER(obj); + int i; + + for (i = 0; i < NPCM7XX_TIMERS_PER_CTRL; i++) { + qemu_irq_lower(s->timer[i].irq); + } +} + +static void npcm7xx_timer_realize(DeviceState *dev, Error **errp) +{ + NPCM7xxTimerCtrlState *s = NPCM7XX_TIMER(dev); + SysBusDevice *sbd = &s->parent; + int i; + + for (i = 0; i < NPCM7XX_TIMERS_PER_CTRL; i++) { + NPCM7xxTimer *t = &s->timer[i]; + t->ctrl = s; + timer_init_ns(&t->qtimer, QEMU_CLOCK_VIRTUAL, npcm7xx_timer_expired, t); + sysbus_init_irq(sbd, &t->irq); + } + + memory_region_init_io(&s->iomem, OBJECT(s), &npcm7xx_timer_ops, s, + TYPE_NPCM7XX_TIMER, 4 * KiB); + sysbus_init_mmio(sbd, &s->iomem); +} + +static const VMStateDescription vmstate_npcm7xx_timer = { + .name = "npcm7xx-timer", + .version_id = 0, + .minimum_version_id = 0, + .fields = (VMStateField[]) { + VMSTATE_TIMER(qtimer, NPCM7xxTimer), + VMSTATE_INT64(expires_ns, NPCM7xxTimer), + VMSTATE_INT64(remaining_ns, NPCM7xxTimer), + VMSTATE_UINT32(tcsr, NPCM7xxTimer), + VMSTATE_UINT32(ticr, NPCM7xxTimer), + VMSTATE_END_OF_LIST(), + }, +}; + +static const VMStateDescription vmstate_npcm7xx_timer_ctrl = { + .name = "npcm7xx-timer-ctrl", + .version_id = 0, + .minimum_version_id = 0, + .fields = (VMStateField[]) { + VMSTATE_UINT32(tisr, NPCM7xxTimerCtrlState), + VMSTATE_UINT32(wtcr, NPCM7xxTimerCtrlState), + VMSTATE_STRUCT_ARRAY(timer, NPCM7xxTimerCtrlState, + NPCM7XX_TIMERS_PER_CTRL, 0, vmstate_npcm7xx_timer, + NPCM7xxTimer), + VMSTATE_END_OF_LIST(), + }, +}; + +static void npcm7xx_timer_class_init(ObjectClass *klass, void *data) +{ + ResettableClass *rc = RESETTABLE_CLASS(klass); + DeviceClass *dc = DEVICE_CLASS(klass); + + QEMU_BUILD_BUG_ON(NPCM7XX_TIMER_REGS_END > NPCM7XX_TIMER_NR_REGS); + + dc->desc = "NPCM7xx Timer Controller"; + dc->realize = npcm7xx_timer_realize; + dc->vmsd = &vmstate_npcm7xx_timer_ctrl; + rc->phases.enter = npcm7xx_timer_enter_reset; + rc->phases.hold = npcm7xx_timer_hold_reset; +} + +static const TypeInfo npcm7xx_timer_info = { + .name = TYPE_NPCM7XX_TIMER, + .parent = TYPE_SYS_BUS_DEVICE, + .instance_size = sizeof(NPCM7xxTimerCtrlState), + .class_init = npcm7xx_timer_class_init, +}; + +static void npcm7xx_timer_register_type(void) +{ + type_register_static(&npcm7xx_timer_info); +} +type_init(npcm7xx_timer_register_type); diff --git a/hw/timer/meson.build b/hw/timer/meson.build index XXXXXXX..XXXXXXX 100644 --- a/hw/timer/meson.build +++ b/hw/timer/meson.build @@ -XXX,XX +XXX,XX @@ softmmu_ss.add(when: 'CONFIG_LM32', if_true: files('lm32_timer.c')) softmmu_ss.add(when: 'CONFIG_MILKYMIST', if_true: files('milkymist-sysctl.c')) softmmu_ss.add(when: 'CONFIG_MIPS_CPS', if_true: files('mips_gictimer.c')) softmmu_ss.add(when: 'CONFIG_MSF2', if_true: files('mss-timer.c')) +softmmu_ss.add(when: 'CONFIG_NPCM7XX', if_true: files('npcm7xx_timer.c')) softmmu_ss.add(when: 'CONFIG_NRF51_SOC', if_true: files('nrf51_timer.c')) softmmu_ss.add(when: 'CONFIG_OMAP', if_true: files('omap_gptimer.c')) softmmu_ss.add(when: 'CONFIG_OMAP', if_true: files('omap_synctimer.c')) 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_dualtimer_read(uint64_t offset, uint64_t data, unsigned size) "CMSDK A 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" +# npcm7xx_timer.c +npcm7xx_timer_read(const char *id, uint64_t offset, uint64_t value) " %s offset: 0x%04" PRIx64 " value 0x%08" PRIx64 +npcm7xx_timer_write(const char *id, uint64_t offset, uint64_t value) "%s offset: 0x%04" PRIx64 " value 0x%08" PRIx64 +npcm7xx_timer_irq(const char *id, int timer, int state) "%s timer %d state %d" + # nrf51_timer.c nrf51_timer_read(uint8_t timer_id, uint64_t addr, uint32_t value, unsigned size) "timer %u read addr 0x%" PRIx64 " data 0x%" PRIx32 " size %u" nrf51_timer_write(uint8_t timer_id, uint64_t addr, uint32_t value, unsigned size) "timer %u write addr 0x%" PRIx64 " data 0x%" PRIx32 " size %u" -- 2.20.1 From: Havard Skinnemoen <hskinnemoen@google.com> The Nuvoton NPCM7xx SoC family are used to implement Baseboard Management Controllers in servers. While the family includes four SoCs, this patch implements limited support for two of them: NPCM730 (targeted for Data Center applications) and NPCM750 (targeted for Enterprise applications). This patch includes little more than the bare minimum needed to boot a Linux kernel built with NPCM7xx support in direct-kernel mode: - Two Cortex-A9 CPU cores with built-in periperhals. - Global Configuration Registers. - Clock Management. - 3 Timer Modules with 5 timers each. - 4 serial ports. The chips themselves have a lot more features, some of which will be added to the model at a later stage. Reviewed-by: Tyrone Ting <kfting@nuvoton.com> Reviewed-by: Joel Stanley <joel@jms.id.au> Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Tested-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Tested-by: Alexander Bulekov <alxndr@bu.edu> Signed-off-by: Havard Skinnemoen <hskinnemoen@google.com> Message-id: 20200911052101.2602693-5-hskinnemoen@google.com Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- include/hw/arm/npcm7xx.h | 85 ++++++++ hw/arm/npcm7xx.c | 407 +++++++++++++++++++++++++++++++++++++++ hw/arm/Kconfig | 5 + hw/arm/meson.build | 1 + 4 files changed, 498 insertions(+) create mode 100644 include/hw/arm/npcm7xx.h create mode 100644 hw/arm/npcm7xx.c diff --git a/include/hw/arm/npcm7xx.h b/include/hw/arm/npcm7xx.h new file mode 100644 index XXXXXXX..XXXXXXX --- /dev/null +++ b/include/hw/arm/npcm7xx.h @@ -XXX,XX +XXX,XX @@ +/* + * Nuvoton NPCM7xx SoC family. + * + * Copyright 2020 Google LLC + * + * This program is free software; you can redistribute it and/or modify it + * under the terms of the GNU General Public License as published by the + * Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program 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 General Public License + * for more details. + */ +#ifndef NPCM7XX_H +#define NPCM7XX_H + +#include "hw/boards.h" +#include "hw/cpu/a9mpcore.h" +#include "hw/misc/npcm7xx_clk.h" +#include "hw/misc/npcm7xx_gcr.h" +#include "hw/timer/npcm7xx_timer.h" +#include "target/arm/cpu.h" + +#define NPCM7XX_MAX_NUM_CPUS (2) + +/* The first half of the address space is reserved for DDR4 DRAM. */ +#define NPCM7XX_DRAM_BA (0x00000000) +#define NPCM7XX_DRAM_SZ (2 * GiB) + +/* Magic addresses for setting up direct kernel booting and SMP boot stubs. */ +#define NPCM7XX_LOADER_START (0x00000000) /* Start of SDRAM */ +#define NPCM7XX_SMP_LOADER_START (0xffff0000) /* Boot ROM */ +#define NPCM7XX_SMP_BOOTREG_ADDR (0xf080013c) /* GCR.SCRPAD */ +#define NPCM7XX_GIC_CPU_IF_ADDR (0xf03fe100) /* GIC within A9 */ + +typedef struct NPCM7xxState { + DeviceState parent; + + ARMCPU cpu[NPCM7XX_MAX_NUM_CPUS]; + A9MPPrivState a9mpcore; + + MemoryRegion sram; + MemoryRegion irom; + MemoryRegion ram3; + MemoryRegion *dram; + + NPCM7xxGCRState gcr; + NPCM7xxCLKState clk; + NPCM7xxTimerCtrlState tim[3]; +} NPCM7xxState; + +#define TYPE_NPCM7XX "npcm7xx" +#define NPCM7XX(obj) OBJECT_CHECK(NPCM7xxState, (obj), TYPE_NPCM7XX) + +#define TYPE_NPCM730 "npcm730" +#define TYPE_NPCM750 "npcm750" + +typedef struct NPCM7xxClass { + DeviceClass parent; + + /* Bitmask of modules that are permanently disabled on this chip. */ + uint32_t disabled_modules; + /* Number of CPU cores enabled in this SoC class (may be 1 or 2). */ + uint32_t num_cpus; +} NPCM7xxClass; + +#define NPCM7XX_CLASS(klass) \ + OBJECT_CLASS_CHECK(NPCM7xxClass, (klass), TYPE_NPCM7XX) +#define NPCM7XX_GET_CLASS(obj) \ + OBJECT_GET_CLASS(NPCM7xxClass, (obj), TYPE_NPCM7XX) + +/** + * npcm7xx_load_kernel - Loads memory with everything needed to boot + * @machine - The machine containing the SoC to be booted. + * @soc - The SoC containing the CPU to be booted. + * + * This will set up the ARM boot info structure for the specific NPCM7xx + * derivative and call arm_load_kernel() to set up loading of the kernel, etc. + * into memory, if requested by the user. + */ +void npcm7xx_load_kernel(MachineState *machine, NPCM7xxState *soc); + +#endif /* NPCM7XX_H */ diff --git a/hw/arm/npcm7xx.c b/hw/arm/npcm7xx.c new file mode 100644 index XXXXXXX..XXXXXXX --- /dev/null +++ b/hw/arm/npcm7xx.c @@ -XXX,XX +XXX,XX @@ +/* + * Nuvoton NPCM7xx SoC family. + * + * Copyright 2020 Google LLC + * + * This program is free software; you can redistribute it and/or modify it + * under the terms of the GNU General Public License as published by the + * Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program 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 General Public License + * for more details. + */ + +#include "qemu/osdep.h" + +#include "exec/address-spaces.h" +#include "hw/arm/boot.h" +#include "hw/arm/npcm7xx.h" +#include "hw/char/serial.h" +#include "hw/loader.h" +#include "hw/misc/unimp.h" +#include "hw/qdev-properties.h" +#include "qapi/error.h" +#include "qemu/units.h" +#include "sysemu/sysemu.h" + +/* + * This covers the whole MMIO space. We'll use this to catch any MMIO accesses + * that aren't handled by any device. + */ +#define NPCM7XX_MMIO_BA (0x80000000) +#define NPCM7XX_MMIO_SZ (0x7ffd0000) + +/* Core system modules. */ +#define NPCM7XX_L2C_BA (0xf03fc000) +#define NPCM7XX_CPUP_BA (0xf03fe000) +#define NPCM7XX_GCR_BA (0xf0800000) +#define NPCM7XX_CLK_BA (0xf0801000) + +/* Internal AHB SRAM */ +#define NPCM7XX_RAM3_BA (0xc0008000) +#define NPCM7XX_RAM3_SZ (4 * KiB) + +/* Memory blocks at the end of the address space */ +#define NPCM7XX_RAM2_BA (0xfffd0000) +#define NPCM7XX_RAM2_SZ (128 * KiB) +#define NPCM7XX_ROM_BA (0xffff0000) +#define NPCM7XX_ROM_SZ (64 * KiB) + +/* + * Interrupt lines going into the GIC. This does not include internal Cortex-A9 + * interrupts. + */ +enum NPCM7xxInterrupt { + NPCM7XX_UART0_IRQ = 2, + NPCM7XX_UART1_IRQ, + NPCM7XX_UART2_IRQ, + NPCM7XX_UART3_IRQ, + NPCM7XX_TIMER0_IRQ = 32, /* Timer Module 0 */ + NPCM7XX_TIMER1_IRQ, + NPCM7XX_TIMER2_IRQ, + NPCM7XX_TIMER3_IRQ, + NPCM7XX_TIMER4_IRQ, + NPCM7XX_TIMER5_IRQ, /* Timer Module 1 */ + NPCM7XX_TIMER6_IRQ, + NPCM7XX_TIMER7_IRQ, + NPCM7XX_TIMER8_IRQ, + NPCM7XX_TIMER9_IRQ, + NPCM7XX_TIMER10_IRQ, /* Timer Module 2 */ + NPCM7XX_TIMER11_IRQ, + NPCM7XX_TIMER12_IRQ, + NPCM7XX_TIMER13_IRQ, + NPCM7XX_TIMER14_IRQ, +}; + +/* Total number of GIC interrupts, including internal Cortex-A9 interrupts. */ +#define NPCM7XX_NUM_IRQ (160) + +/* Register base address for each Timer Module */ +static const hwaddr npcm7xx_tim_addr[] = { + 0xf0008000, + 0xf0009000, + 0xf000a000, +}; + +/* Register base address for each 16550 UART */ +static const hwaddr npcm7xx_uart_addr[] = { + 0xf0001000, + 0xf0002000, + 0xf0003000, + 0xf0004000, +}; + +static void npcm7xx_write_secondary_boot(ARMCPU *cpu, + const struct arm_boot_info *info) +{ + /* + * The default smpboot stub halts the secondary CPU with a 'wfi' + * instruction, but the arch/arm/mach-npcm/platsmp.c in the Linux kernel + * does not send an IPI to wake it up, so the second CPU fails to boot. So + * we need to provide our own smpboot stub that can not use 'wfi', it has + * to spin the secondary CPU until the first CPU writes to the SCRPAD reg. + */ + uint32_t smpboot[] = { + 0xe59f2018, /* ldr r2, bootreg_addr */ + 0xe3a00000, /* mov r0, #0 */ + 0xe5820000, /* str r0, [r2] */ + 0xe320f002, /* wfe */ + 0xe5921000, /* ldr r1, [r2] */ + 0xe1110001, /* tst r1, r1 */ + 0x0afffffb, /* beq <wfe> */ + 0xe12fff11, /* bx r1 */ + NPCM7XX_SMP_BOOTREG_ADDR, + }; + int i; + + for (i = 0; i < ARRAY_SIZE(smpboot); i++) { + smpboot[i] = tswap32(smpboot[i]); + } + + rom_add_blob_fixed("smpboot", smpboot, sizeof(smpboot), + NPCM7XX_SMP_LOADER_START); +} + +static struct arm_boot_info npcm7xx_binfo = { + .loader_start = NPCM7XX_LOADER_START, + .smp_loader_start = NPCM7XX_SMP_LOADER_START, + .smp_bootreg_addr = NPCM7XX_SMP_BOOTREG_ADDR, + .gic_cpu_if_addr = NPCM7XX_GIC_CPU_IF_ADDR, + .write_secondary_boot = npcm7xx_write_secondary_boot, + .board_id = -1, +}; + +void npcm7xx_load_kernel(MachineState *machine, NPCM7xxState *soc) +{ + NPCM7xxClass *sc = NPCM7XX_GET_CLASS(soc); + + npcm7xx_binfo.ram_size = machine->ram_size; + npcm7xx_binfo.nb_cpus = sc->num_cpus; + + arm_load_kernel(&soc->cpu[0], machine, &npcm7xx_binfo); +} + +static qemu_irq npcm7xx_irq(NPCM7xxState *s, int n) +{ + return qdev_get_gpio_in(DEVICE(&s->a9mpcore), n); +} + +static void npcm7xx_init(Object *obj) +{ + NPCM7xxState *s = NPCM7XX(obj); + int i; + + for (i = 0; i < NPCM7XX_MAX_NUM_CPUS; i++) { + object_initialize_child(obj, "cpu[*]", &s->cpu[i], + ARM_CPU_TYPE_NAME("cortex-a9")); + } + + object_initialize_child(obj, "a9mpcore", &s->a9mpcore, TYPE_A9MPCORE_PRIV); + object_initialize_child(obj, "gcr", &s->gcr, TYPE_NPCM7XX_GCR); + object_property_add_alias(obj, "power-on-straps", OBJECT(&s->gcr), + "power-on-straps"); + object_initialize_child(obj, "clk", &s->clk, TYPE_NPCM7XX_CLK); + + for (i = 0; i < ARRAY_SIZE(s->tim); i++) { + object_initialize_child(obj, "tim[*]", &s->tim[i], TYPE_NPCM7XX_TIMER); + } +} + +static void npcm7xx_realize(DeviceState *dev, Error **errp) +{ + NPCM7xxState *s = NPCM7XX(dev); + NPCM7xxClass *nc = NPCM7XX_GET_CLASS(s); + int i; + + if (memory_region_size(s->dram) > NPCM7XX_DRAM_SZ) { + error_setg(errp, "%s: NPCM7xx cannot address more than %" PRIu64 + " MiB of DRAM", __func__, NPCM7XX_DRAM_SZ / MiB); + return; + } + + /* CPUs */ + for (i = 0; i < nc->num_cpus; i++) { + object_property_set_int(OBJECT(&s->cpu[i]), "mp-affinity", + arm_cpu_mp_affinity(i, NPCM7XX_MAX_NUM_CPUS), + &error_abort); + object_property_set_int(OBJECT(&s->cpu[i]), "reset-cbar", + NPCM7XX_GIC_CPU_IF_ADDR, &error_abort); + object_property_set_bool(OBJECT(&s->cpu[i]), "reset-hivecs", true, + &error_abort); + + /* Disable security extensions. */ + object_property_set_bool(OBJECT(&s->cpu[i]), "has_el3", false, + &error_abort); + + if (!qdev_realize(DEVICE(&s->cpu[i]), NULL, errp)) { + return; + } + } + + /* A9MPCORE peripherals. Can only fail if we pass bad parameters here. */ + object_property_set_int(OBJECT(&s->a9mpcore), "num-cpu", nc->num_cpus, + &error_abort); + object_property_set_int(OBJECT(&s->a9mpcore), "num-irq", NPCM7XX_NUM_IRQ, + &error_abort); + sysbus_realize(SYS_BUS_DEVICE(&s->a9mpcore), &error_abort); + sysbus_mmio_map(SYS_BUS_DEVICE(&s->a9mpcore), 0, NPCM7XX_CPUP_BA); + + for (i = 0; i < nc->num_cpus; i++) { + sysbus_connect_irq(SYS_BUS_DEVICE(&s->a9mpcore), i, + qdev_get_gpio_in(DEVICE(&s->cpu[i]), ARM_CPU_IRQ)); + sysbus_connect_irq(SYS_BUS_DEVICE(&s->a9mpcore), i + nc->num_cpus, + qdev_get_gpio_in(DEVICE(&s->cpu[i]), ARM_CPU_FIQ)); + } + + /* L2 cache controller */ + sysbus_create_simple("l2x0", NPCM7XX_L2C_BA, NULL); + + /* System Global Control Registers (GCR). Can fail due to user input. */ + object_property_set_int(OBJECT(&s->gcr), "disabled-modules", + nc->disabled_modules, &error_abort); + object_property_add_const_link(OBJECT(&s->gcr), "dram-mr", OBJECT(s->dram)); + if (!sysbus_realize(SYS_BUS_DEVICE(&s->gcr), errp)) { + return; + } + sysbus_mmio_map(SYS_BUS_DEVICE(&s->gcr), 0, NPCM7XX_GCR_BA); + + /* Clock Control Registers (CLK). Cannot fail. */ + sysbus_realize(SYS_BUS_DEVICE(&s->clk), &error_abort); + sysbus_mmio_map(SYS_BUS_DEVICE(&s->clk), 0, NPCM7XX_CLK_BA); + + /* Timer Modules (TIM). Cannot fail. */ + QEMU_BUILD_BUG_ON(ARRAY_SIZE(npcm7xx_tim_addr) != ARRAY_SIZE(s->tim)); + for (i = 0; i < ARRAY_SIZE(s->tim); i++) { + SysBusDevice *sbd = SYS_BUS_DEVICE(&s->tim[i]); + int first_irq; + int j; + + sysbus_realize(sbd, &error_abort); + sysbus_mmio_map(sbd, 0, npcm7xx_tim_addr[i]); + + first_irq = NPCM7XX_TIMER0_IRQ + i * NPCM7XX_TIMERS_PER_CTRL; + for (j = 0; j < NPCM7XX_TIMERS_PER_CTRL; j++) { + qemu_irq irq = npcm7xx_irq(s, first_irq + j); + sysbus_connect_irq(sbd, j, irq); + } + } + + /* UART0..3 (16550 compatible) */ + for (i = 0; i < ARRAY_SIZE(npcm7xx_uart_addr); i++) { + serial_mm_init(get_system_memory(), npcm7xx_uart_addr[i], 2, + npcm7xx_irq(s, NPCM7XX_UART0_IRQ + i), 115200, + serial_hd(i), DEVICE_LITTLE_ENDIAN); + } + + /* RAM2 (SRAM) */ + memory_region_init_ram(&s->sram, OBJECT(dev), "ram2", + NPCM7XX_RAM2_SZ, &error_abort); + memory_region_add_subregion(get_system_memory(), NPCM7XX_RAM2_BA, &s->sram); + + /* RAM3 (SRAM) */ + memory_region_init_ram(&s->ram3, OBJECT(dev), "ram3", + NPCM7XX_RAM3_SZ, &error_abort); + memory_region_add_subregion(get_system_memory(), NPCM7XX_RAM3_BA, &s->ram3); + + /* Internal ROM */ + memory_region_init_rom(&s->irom, OBJECT(dev), "irom", NPCM7XX_ROM_SZ, + &error_abort); + memory_region_add_subregion(get_system_memory(), NPCM7XX_ROM_BA, &s->irom); + + create_unimplemented_device("npcm7xx.shm", 0xc0001000, 4 * KiB); + create_unimplemented_device("npcm7xx.vdmx", 0xe0800000, 4 * KiB); + create_unimplemented_device("npcm7xx.pcierc", 0xe1000000, 64 * KiB); + create_unimplemented_device("npcm7xx.kcs", 0xf0007000, 4 * KiB); + create_unimplemented_device("npcm7xx.rng", 0xf000b000, 4 * KiB); + create_unimplemented_device("npcm7xx.adc", 0xf000c000, 4 * KiB); + create_unimplemented_device("npcm7xx.gfxi", 0xf000e000, 4 * KiB); + create_unimplemented_device("npcm7xx.gpio[0]", 0xf0010000, 4 * KiB); + create_unimplemented_device("npcm7xx.gpio[1]", 0xf0011000, 4 * KiB); + create_unimplemented_device("npcm7xx.gpio[2]", 0xf0012000, 4 * KiB); + create_unimplemented_device("npcm7xx.gpio[3]", 0xf0013000, 4 * KiB); + create_unimplemented_device("npcm7xx.gpio[4]", 0xf0014000, 4 * KiB); + create_unimplemented_device("npcm7xx.gpio[5]", 0xf0015000, 4 * KiB); + create_unimplemented_device("npcm7xx.gpio[6]", 0xf0016000, 4 * KiB); + create_unimplemented_device("npcm7xx.gpio[7]", 0xf0017000, 4 * KiB); + create_unimplemented_device("npcm7xx.smbus[0]", 0xf0080000, 4 * KiB); + create_unimplemented_device("npcm7xx.smbus[1]", 0xf0081000, 4 * KiB); + create_unimplemented_device("npcm7xx.smbus[2]", 0xf0082000, 4 * KiB); + create_unimplemented_device("npcm7xx.smbus[3]", 0xf0083000, 4 * KiB); + create_unimplemented_device("npcm7xx.smbus[4]", 0xf0084000, 4 * KiB); + create_unimplemented_device("npcm7xx.smbus[5]", 0xf0085000, 4 * KiB); + create_unimplemented_device("npcm7xx.smbus[6]", 0xf0086000, 4 * KiB); + create_unimplemented_device("npcm7xx.smbus[7]", 0xf0087000, 4 * KiB); + create_unimplemented_device("npcm7xx.smbus[8]", 0xf0088000, 4 * KiB); + create_unimplemented_device("npcm7xx.smbus[9]", 0xf0089000, 4 * KiB); + create_unimplemented_device("npcm7xx.smbus[10]", 0xf008a000, 4 * KiB); + create_unimplemented_device("npcm7xx.smbus[11]", 0xf008b000, 4 * KiB); + create_unimplemented_device("npcm7xx.smbus[12]", 0xf008c000, 4 * KiB); + create_unimplemented_device("npcm7xx.smbus[13]", 0xf008d000, 4 * KiB); + create_unimplemented_device("npcm7xx.smbus[14]", 0xf008e000, 4 * KiB); + create_unimplemented_device("npcm7xx.smbus[15]", 0xf008f000, 4 * KiB); + create_unimplemented_device("npcm7xx.espi", 0xf009f000, 4 * KiB); + create_unimplemented_device("npcm7xx.peci", 0xf0100000, 4 * KiB); + create_unimplemented_device("npcm7xx.siox[1]", 0xf0101000, 4 * KiB); + create_unimplemented_device("npcm7xx.siox[2]", 0xf0102000, 4 * KiB); + create_unimplemented_device("npcm7xx.pwm[0]", 0xf0103000, 4 * KiB); + create_unimplemented_device("npcm7xx.pwm[1]", 0xf0104000, 4 * KiB); + create_unimplemented_device("npcm7xx.mft[0]", 0xf0180000, 4 * KiB); + create_unimplemented_device("npcm7xx.mft[1]", 0xf0181000, 4 * KiB); + create_unimplemented_device("npcm7xx.mft[2]", 0xf0182000, 4 * KiB); + create_unimplemented_device("npcm7xx.mft[3]", 0xf0183000, 4 * KiB); + create_unimplemented_device("npcm7xx.mft[4]", 0xf0184000, 4 * KiB); + create_unimplemented_device("npcm7xx.mft[5]", 0xf0185000, 4 * KiB); + create_unimplemented_device("npcm7xx.mft[6]", 0xf0186000, 4 * KiB); + create_unimplemented_device("npcm7xx.mft[7]", 0xf0187000, 4 * KiB); + create_unimplemented_device("npcm7xx.pspi1", 0xf0200000, 4 * KiB); + create_unimplemented_device("npcm7xx.pspi2", 0xf0201000, 4 * KiB); + create_unimplemented_device("npcm7xx.ahbpci", 0xf0400000, 1 * MiB); + create_unimplemented_device("npcm7xx.mcphy", 0xf05f0000, 64 * KiB); + create_unimplemented_device("npcm7xx.gmac1", 0xf0802000, 8 * KiB); + create_unimplemented_device("npcm7xx.gmac2", 0xf0804000, 8 * KiB); + create_unimplemented_device("npcm7xx.ehci", 0xf0806000, 4 * KiB); + create_unimplemented_device("npcm7xx.ohci", 0xf0807000, 4 * KiB); + create_unimplemented_device("npcm7xx.vcd", 0xf0810000, 64 * KiB); + create_unimplemented_device("npcm7xx.ece", 0xf0820000, 8 * KiB); + create_unimplemented_device("npcm7xx.vdma", 0xf0822000, 8 * KiB); + create_unimplemented_device("npcm7xx.emc1", 0xf0825000, 4 * KiB); + create_unimplemented_device("npcm7xx.emc2", 0xf0826000, 4 * KiB); + create_unimplemented_device("npcm7xx.usbd[0]", 0xf0830000, 4 * KiB); + create_unimplemented_device("npcm7xx.usbd[1]", 0xf0831000, 4 * KiB); + create_unimplemented_device("npcm7xx.usbd[2]", 0xf0832000, 4 * KiB); + create_unimplemented_device("npcm7xx.usbd[3]", 0xf0833000, 4 * KiB); + create_unimplemented_device("npcm7xx.usbd[4]", 0xf0834000, 4 * KiB); + create_unimplemented_device("npcm7xx.usbd[5]", 0xf0835000, 4 * KiB); + create_unimplemented_device("npcm7xx.usbd[6]", 0xf0836000, 4 * KiB); + create_unimplemented_device("npcm7xx.usbd[7]", 0xf0837000, 4 * KiB); + create_unimplemented_device("npcm7xx.usbd[8]", 0xf0838000, 4 * KiB); + create_unimplemented_device("npcm7xx.usbd[9]", 0xf0839000, 4 * KiB); + create_unimplemented_device("npcm7xx.sd", 0xf0840000, 8 * KiB); + create_unimplemented_device("npcm7xx.mmc", 0xf0842000, 8 * KiB); + create_unimplemented_device("npcm7xx.pcimbx", 0xf0848000, 512 * KiB); + create_unimplemented_device("npcm7xx.aes", 0xf0858000, 4 * KiB); + create_unimplemented_device("npcm7xx.des", 0xf0859000, 4 * KiB); + create_unimplemented_device("npcm7xx.sha", 0xf085a000, 4 * KiB); + create_unimplemented_device("npcm7xx.secacc", 0xf085b000, 4 * KiB); + create_unimplemented_device("npcm7xx.spixcs0", 0xf8000000, 16 * MiB); + create_unimplemented_device("npcm7xx.spixcs1", 0xf9000000, 16 * MiB); + create_unimplemented_device("npcm7xx.spix", 0xfb001000, 4 * KiB); +} + +static Property npcm7xx_properties[] = { + DEFINE_PROP_LINK("dram-mr", NPCM7xxState, dram, TYPE_MEMORY_REGION, + MemoryRegion *), + DEFINE_PROP_END_OF_LIST(), +}; + +static void npcm7xx_class_init(ObjectClass *oc, void *data) +{ + DeviceClass *dc = DEVICE_CLASS(oc); + + dc->realize = npcm7xx_realize; + dc->user_creatable = false; + device_class_set_props(dc, npcm7xx_properties); +} + +static void npcm730_class_init(ObjectClass *oc, void *data) +{ + NPCM7xxClass *nc = NPCM7XX_CLASS(oc); + + /* NPCM730 is optimized for data center use, so no graphics, etc. */ + nc->disabled_modules = 0x00300395; + nc->num_cpus = 2; +} + +static void npcm750_class_init(ObjectClass *oc, void *data) +{ + NPCM7xxClass *nc = NPCM7XX_CLASS(oc); + + /* NPCM750 has 2 cores and a full set of peripherals */ + nc->disabled_modules = 0x00000000; + nc->num_cpus = 2; +} + +static const TypeInfo npcm7xx_soc_types[] = { + { + .name = TYPE_NPCM7XX, + .parent = TYPE_DEVICE, + .instance_size = sizeof(NPCM7xxState), + .instance_init = npcm7xx_init, + .class_size = sizeof(NPCM7xxClass), + .class_init = npcm7xx_class_init, + .abstract = true, + }, { + .name = TYPE_NPCM730, + .parent = TYPE_NPCM7XX, + .class_init = npcm730_class_init, + }, { + .name = TYPE_NPCM750, + .parent = TYPE_NPCM7XX, + .class_init = npcm750_class_init, + }, +}; + +DEFINE_TYPES(npcm7xx_soc_types); diff --git a/hw/arm/Kconfig b/hw/arm/Kconfig index XXXXXXX..XXXXXXX 100644 --- a/hw/arm/Kconfig +++ b/hw/arm/Kconfig @@ -XXX,XX +XXX,XX @@ config XLNX_VERSAL config NPCM7XX bool + select A9MPCORE + select ARM_GIC + select PL310 # cache controller + select SERIAL + select UNIMP config FSL_IMX25 bool diff --git a/hw/arm/meson.build b/hw/arm/meson.build index XXXXXXX..XXXXXXX 100644 --- a/hw/arm/meson.build +++ b/hw/arm/meson.build @@ -XXX,XX +XXX,XX @@ arm_ss.add(when: 'CONFIG_MICROBIT', if_true: files('microbit.c')) arm_ss.add(when: 'CONFIG_MUSICPAL', if_true: files('musicpal.c')) arm_ss.add(when: 'CONFIG_NETDUINO2', if_true: files('netduino2.c')) arm_ss.add(when: 'CONFIG_NETDUINOPLUS2', if_true: files('netduinoplus2.c')) +arm_ss.add(when: 'CONFIG_NPCM7XX', if_true: files('npcm7xx.c')) arm_ss.add(when: 'CONFIG_NSERIES', if_true: files('nseries.c')) arm_ss.add(when: 'CONFIG_SX1', if_true: files('omap_sx1.c')) arm_ss.add(when: 'CONFIG_CHEETAH', if_true: files('palm.c')) -- 2.20.1 From: Havard Skinnemoen <hskinnemoen@google.com> This adds two new machines, both supported by OpenBMC: - npcm750-evb: Nuvoton NPCM750 Evaluation Board. - quanta-gsj: A board with a NPCM730 chip. They rely on the NPCM7xx SoC device to do the heavy lifting. They are almost completely identical at the moment, apart from the SoC type, which currently only changes the reset contents of one register (GCR.MDLR), but they might grow apart a bit more as more functionality is added. Both machines can boot the Linux kernel into /bin/sh. Reviewed-by: Tyrone Ting <kfting@nuvoton.com> Reviewed-by: Joel Stanley <joel@jms.id.au> Reviewed-by: Cédric Le Goater <clg@kaod.org> Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Tested-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Tested-by: Alexander Bulekov <alxndr@bu.edu> Signed-off-by: Havard Skinnemoen <hskinnemoen@google.com> Message-id: 20200911052101.2602693-6-hskinnemoen@google.com Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- default-configs/arm-softmmu.mak | 1 + include/hw/arm/npcm7xx.h | 19 +++++ hw/arm/npcm7xx_boards.c | 145 ++++++++++++++++++++++++++++++++ hw/arm/meson.build | 2 +- 4 files changed, 166 insertions(+), 1 deletion(-) create mode 100644 hw/arm/npcm7xx_boards.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_GUMSTIX=y CONFIG_SPITZ=y CONFIG_TOSA=y CONFIG_Z2=y +CONFIG_NPCM7XX=y CONFIG_COLLIE=y CONFIG_ASPEED_SOC=y CONFIG_NETDUINO2=y diff --git a/include/hw/arm/npcm7xx.h b/include/hw/arm/npcm7xx.h index XXXXXXX..XXXXXXX 100644 --- a/include/hw/arm/npcm7xx.h +++ b/include/hw/arm/npcm7xx.h @@ -XXX,XX +XXX,XX @@ #define NPCM7XX_SMP_BOOTREG_ADDR (0xf080013c) /* GCR.SCRPAD */ #define NPCM7XX_GIC_CPU_IF_ADDR (0xf03fe100) /* GIC within A9 */ +typedef struct NPCM7xxMachine { + MachineState parent; +} NPCM7xxMachine; + +#define TYPE_NPCM7XX_MACHINE MACHINE_TYPE_NAME("npcm7xx") +#define NPCM7XX_MACHINE(obj) \ + OBJECT_CHECK(NPCM7xxMachine, (obj), TYPE_NPCM7XX_MACHINE) + +typedef struct NPCM7xxMachineClass { + MachineClass parent; + + const char *soc_type; +} NPCM7xxMachineClass; + +#define NPCM7XX_MACHINE_CLASS(klass) \ + OBJECT_CLASS_CHECK(NPCM7xxMachineClass, (klass), TYPE_NPCM7XX_MACHINE) +#define NPCM7XX_MACHINE_GET_CLASS(obj) \ + OBJECT_GET_CLASS(NPCM7xxMachineClass, (obj), TYPE_NPCM7XX_MACHINE) + typedef struct NPCM7xxState { DeviceState parent; diff --git a/hw/arm/npcm7xx_boards.c b/hw/arm/npcm7xx_boards.c new file mode 100644 index XXXXXXX..XXXXXXX --- /dev/null +++ b/hw/arm/npcm7xx_boards.c @@ -XXX,XX +XXX,XX @@ +/* + * Machine definitions for boards featuring an NPCM7xx SoC. + * + * Copyright 2020 Google LLC + * + * This program is free software; you can redistribute it and/or modify it + * under the terms of the GNU General Public License as published by the + * Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program 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 General Public License + * for more details. + */ + +#include "qemu/osdep.h" + +#include "exec/address-spaces.h" +#include "hw/arm/npcm7xx.h" +#include "hw/core/cpu.h" +#include "qapi/error.h" +#include "qemu/units.h" + +#define NPCM750_EVB_POWER_ON_STRAPS 0x00001ff7 +#define QUANTA_GSJ_POWER_ON_STRAPS 0x00001fff + +static void npcm7xx_connect_dram(NPCM7xxState *soc, MemoryRegion *dram) +{ + memory_region_add_subregion(get_system_memory(), NPCM7XX_DRAM_BA, dram); + + object_property_set_link(OBJECT(soc), "dram-mr", OBJECT(dram), + &error_abort); +} + +static NPCM7xxState *npcm7xx_create_soc(MachineState *machine, + uint32_t hw_straps) +{ + NPCM7xxMachineClass *nmc = NPCM7XX_MACHINE_GET_CLASS(machine); + MachineClass *mc = &nmc->parent; + Object *obj; + + if (strcmp(machine->cpu_type, mc->default_cpu_type) != 0) { + error_report("This board can only be used with %s", + mc->default_cpu_type); + exit(1); + } + + obj = object_new_with_props(nmc->soc_type, OBJECT(machine), "soc", + &error_abort, NULL); + object_property_set_uint(obj, "power-on-straps", hw_straps, &error_abort); + + return NPCM7XX(obj); +} + +static void npcm750_evb_init(MachineState *machine) +{ + NPCM7xxState *soc; + + soc = npcm7xx_create_soc(machine, NPCM750_EVB_POWER_ON_STRAPS); + npcm7xx_connect_dram(soc, machine->ram); + qdev_realize(DEVICE(soc), NULL, &error_fatal); + + npcm7xx_load_kernel(machine, soc); +} + +static void quanta_gsj_init(MachineState *machine) +{ + NPCM7xxState *soc; + + soc = npcm7xx_create_soc(machine, QUANTA_GSJ_POWER_ON_STRAPS); + npcm7xx_connect_dram(soc, machine->ram); + qdev_realize(DEVICE(soc), NULL, &error_fatal); + + npcm7xx_load_kernel(machine, soc); +} + +static void npcm7xx_set_soc_type(NPCM7xxMachineClass *nmc, const char *type) +{ + NPCM7xxClass *sc = NPCM7XX_CLASS(object_class_by_name(type)); + MachineClass *mc = MACHINE_CLASS(nmc); + + nmc->soc_type = type; + mc->default_cpus = mc->min_cpus = mc->max_cpus = sc->num_cpus; +} + +static void npcm7xx_machine_class_init(ObjectClass *oc, void *data) +{ + MachineClass *mc = MACHINE_CLASS(oc); + + mc->no_floppy = 1; + mc->no_cdrom = 1; + mc->no_parallel = 1; + mc->default_ram_id = "ram"; + mc->default_cpu_type = ARM_CPU_TYPE_NAME("cortex-a9"); +} + +/* + * Schematics: + * https://github.com/Nuvoton-Israel/nuvoton-info/blob/master/npcm7xx-poleg/evaluation-board/board_deliverables/NPCM750x_EB_ver.A1.1_COMPLETE.pdf + */ +static void npcm750_evb_machine_class_init(ObjectClass *oc, void *data) +{ + NPCM7xxMachineClass *nmc = NPCM7XX_MACHINE_CLASS(oc); + MachineClass *mc = MACHINE_CLASS(oc); + + npcm7xx_set_soc_type(nmc, TYPE_NPCM750); + + mc->desc = "Nuvoton NPCM750 Evaluation Board (Cortex A9)"; + mc->init = npcm750_evb_init; + mc->default_ram_size = 512 * MiB; +}; + +static void gsj_machine_class_init(ObjectClass *oc, void *data) +{ + NPCM7xxMachineClass *nmc = NPCM7XX_MACHINE_CLASS(oc); + MachineClass *mc = MACHINE_CLASS(oc); + + npcm7xx_set_soc_type(nmc, TYPE_NPCM730); + + mc->desc = "Quanta GSJ (Cortex A9)"; + mc->init = quanta_gsj_init; + mc->default_ram_size = 512 * MiB; +}; + +static const TypeInfo npcm7xx_machine_types[] = { + { + .name = TYPE_NPCM7XX_MACHINE, + .parent = TYPE_MACHINE, + .instance_size = sizeof(NPCM7xxMachine), + .class_size = sizeof(NPCM7xxMachineClass), + .class_init = npcm7xx_machine_class_init, + .abstract = true, + }, { + .name = MACHINE_TYPE_NAME("npcm750-evb"), + .parent = TYPE_NPCM7XX_MACHINE, + .class_init = npcm750_evb_machine_class_init, + }, { + .name = MACHINE_TYPE_NAME("quanta-gsj"), + .parent = TYPE_NPCM7XX_MACHINE, + .class_init = gsj_machine_class_init, + }, +}; + +DEFINE_TYPES(npcm7xx_machine_types) diff --git a/hw/arm/meson.build b/hw/arm/meson.build index XXXXXXX..XXXXXXX 100644 --- a/hw/arm/meson.build +++ b/hw/arm/meson.build @@ -XXX,XX +XXX,XX @@ arm_ss.add(when: 'CONFIG_MICROBIT', if_true: files('microbit.c')) arm_ss.add(when: 'CONFIG_MUSICPAL', if_true: files('musicpal.c')) arm_ss.add(when: 'CONFIG_NETDUINO2', if_true: files('netduino2.c')) arm_ss.add(when: 'CONFIG_NETDUINOPLUS2', if_true: files('netduinoplus2.c')) -arm_ss.add(when: 'CONFIG_NPCM7XX', if_true: files('npcm7xx.c')) +arm_ss.add(when: 'CONFIG_NPCM7XX', if_true: files('npcm7xx.c', 'npcm7xx_boards.c')) arm_ss.add(when: 'CONFIG_NSERIES', if_true: files('nseries.c')) arm_ss.add(when: 'CONFIG_SX1', if_true: files('omap_sx1.c')) arm_ss.add(when: 'CONFIG_CHEETAH', if_true: files('palm.c')) -- 2.20.1 From: Havard Skinnemoen <hskinnemoen@google.com> This is a minimalistic boot ROM written specifically for use with QEMU. It supports loading the second-stage loader from SPI flash into RAM, SMP boot, and not much else. Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Tested-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Signed-off-by: Havard Skinnemoen <hskinnemoen@google.com> Message-id: 20200911052101.2602693-7-hskinnemoen@google.com Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- .gitmodules | 3 +++ MAINTAINERS | 2 ++ pc-bios/README | 6 ++++++ pc-bios/meson.build | 1 + pc-bios/npcm7xx_bootrom.bin | Bin 0 -> 768 bytes roms/Makefile | 7 +++++++ roms/vbootrom | 1 + 7 files changed, 20 insertions(+) create mode 100644 pc-bios/npcm7xx_bootrom.bin create mode 160000 roms/vbootrom diff --git a/.gitmodules b/.gitmodules index XXXXXXX..XXXXXXX 100644 --- a/.gitmodules +++ b/.gitmodules @@ -XXX,XX +XXX,XX @@ [submodule "meson"] path = meson url = https://github.com/mesonbuild/meson/ +[submodule "roms/vbootrom"] + path = roms/vbootrom + url = https://github.com/google/vbootrom.git diff --git a/MAINTAINERS b/MAINTAINERS index XXXXXXX..XXXXXXX 100644 --- a/MAINTAINERS +++ b/MAINTAINERS @@ -XXX,XX +XXX,XX @@ L: qemu-arm@nongnu.org S: Supported F: hw/*/npcm7xx* F: include/hw/*/npcm7xx* +F: pc-bios/npcm7xx_bootrom.bin +F: roms/vbootrom nSeries M: Andrzej Zaborowski <balrogg@gmail.com> diff --git a/pc-bios/README b/pc-bios/README index XXXXXXX..XXXXXXX 100644 --- a/pc-bios/README +++ b/pc-bios/README @@ -XXX,XX +XXX,XX @@ ("Simplified BSD License" or "FreeBSD License", SPDX: BSD-2-Clause). OpenSBI source code also contains code reused from other projects desribed here: https://github.com/riscv/opensbi/blob/master/ThirdPartyNotices.md. + +- npcm7xx_bootrom.bin is a simplified, free (Apache 2.0) boot ROM for Nuvoton + NPCM7xx BMC devices. It currently implements the bare minimum to load, parse, + initialize and run boot images stored in SPI flash, but may grow more + features over time as needed. The source code is available at: + https://github.com/google/vbootrom diff --git a/pc-bios/meson.build b/pc-bios/meson.build index XXXXXXX..XXXXXXX 100644 --- a/pc-bios/meson.build +++ b/pc-bios/meson.build @@ -XXX,XX +XXX,XX @@ blobs = files( 'opensbi-riscv64-generic-fw_dynamic.bin', 'opensbi-riscv32-generic-fw_dynamic.elf', 'opensbi-riscv64-generic-fw_dynamic.elf', + 'npcm7xx_bootrom.bin', ) if install_blobs diff --git a/pc-bios/npcm7xx_bootrom.bin b/pc-bios/npcm7xx_bootrom.bin new file mode 100644 index XXXXXXX..XXXXXXX GIT binary patch literal 768 zcmd5)JxClu6n-<aczPbVhZYusb8wKx;7TklHfmuZdYT9pDRLwd1p_t-DFpWpyA+8( zwKtZg3J4a0aCM3_X(ZL&4g;46VVk5e$K;z;L99|b@aE%v^S$rQ8)h(Vm@cB9IYc+2 z2SHd4^NwTIGE%w>9S05p1#kf90Sj5Z(jG8}+)IZIp~iXK=T&)dL`%d-q*8aR#mq{7 z9`=6;Dr(H0ACe72R5x?!)^86Qj-X%{+!K9iZNA@*wkBAV&iZ(l^I9?!Gz=S2I_*1d zr+tTQDHjvyzKnw(hu00yX`u!Fv<!~XVcX?@kr#<B0(gGU?$W{gSsQa}CF^8Cfzp2X z@P}yDV-bci(K9XL$FU!som2C`c)?Uc&294s^}Wzumap{hg1X^jN|V25M5tQZ=<9lN z%(zKz#t-qCwHKb;HygOCpvCNL_4@1tXV1YGf^XUE_$zr{g8zWh-6gz-teI(eibtxo z?0OZI4%rU0741PgUD`2xq@H|*4=+Rs?%N)Ox5G+q>C;DilBe_YlkeSUVHA-crNk+k jtiF_MudA<CB(}8|fqYwCf3re&=&@_s761P#-ID$TwgmBa literal 0 HcmV?d00001 diff --git a/roms/Makefile b/roms/Makefile index XXXXXXX..XXXXXXX 100644 --- a/roms/Makefile +++ b/roms/Makefile @@ -XXX,XX +XXX,XX @@ find-cross-gcc = $(firstword $(wildcard $(patsubst %ld,%gcc,$(call find-cross-ld # finally strip off path + toolname so we get the prefix find-cross-prefix = $(subst gcc,,$(notdir $(call find-cross-gcc,$(1)))) +arm_cross_prefix := $(call find-cross-prefix,arm) powerpc64_cross_prefix := $(call find-cross-prefix,powerpc64) powerpc_cross_prefix := $(call find-cross-prefix,powerpc) x86_64_cross_prefix := $(call find-cross-prefix,x86_64) @@ -XXX,XX +XXX,XX @@ default help: @echo " skiboot -- update skiboot.lid" @echo " u-boot.e500 -- update u-boot.e500" @echo " u-boot.sam460 -- update u-boot.sam460" + @echo " npcm7xx_bootrom -- update vbootrom for npcm7xx" @echo " efi -- update UEFI (edk2) platform firmware" @echo " opensbi32-generic -- update OpenSBI for 32-bit generic machine" @echo " opensbi64-generic -- update OpenSBI for 64-bit generic machine" @@ -XXX,XX +XXX,XX @@ bios-microvm: $(MAKE) -C qboot cp qboot/bios.bin ../pc-bios/bios-microvm.bin +npcm7xx_bootrom: + $(MAKE) -C vbootrom CROSS_COMPILE=$(arm_cross_prefix) + cp vbootrom/npcm7xx_bootrom.bin ../pc-bios/npcm7xx_bootrom.bin + clean: rm -rf seabios/.config seabios/out seabios/builds $(MAKE) -C sgabios clean @@ -XXX,XX +XXX,XX @@ clean: $(MAKE) -f Makefile.edk2 clean $(MAKE) -C opensbi clean $(MAKE) -C qboot clean + $(MAKE) -C vbootrom clean diff --git a/roms/vbootrom b/roms/vbootrom new file mode 160000 index XXXXXXX..XXXXXXX --- /dev/null +++ b/roms/vbootrom @@ -0,0 +1 @@ +Subproject commit 0c37a43527f0ee2b9584e7fb2fdc805e902635ac -- 2.20.1 From: Havard Skinnemoen <hskinnemoen@google.com> If a -bios option is specified on the command line, load the image into the internal ROM memory region, which contains the first instructions run by the CPU after reset. If -bios is not specified, the vbootrom included with qemu is loaded by default. Reviewed-by: Tyrone Ting <kfting@nuvoton.com> Reviewed-by: Cédric Le Goater <clg@kaod.org> Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Tested-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Signed-off-by: Havard Skinnemoen <hskinnemoen@google.com> Message-id: 20200911052101.2602693-8-hskinnemoen@google.com Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- hw/arm/npcm7xx_boards.c | 32 ++++++++++++++++++++++++++++++++ 1 file changed, 32 insertions(+) diff --git a/hw/arm/npcm7xx_boards.c b/hw/arm/npcm7xx_boards.c index XXXXXXX..XXXXXXX 100644 --- a/hw/arm/npcm7xx_boards.c +++ b/hw/arm/npcm7xx_boards.c @@ -XXX,XX +XXX,XX @@ #include "exec/address-spaces.h" #include "hw/arm/npcm7xx.h" #include "hw/core/cpu.h" +#include "hw/loader.h" #include "qapi/error.h" +#include "qemu-common.h" #include "qemu/units.h" +#include "sysemu/sysemu.h" #define NPCM750_EVB_POWER_ON_STRAPS 0x00001ff7 #define QUANTA_GSJ_POWER_ON_STRAPS 0x00001fff +static const char npcm7xx_default_bootrom[] = "npcm7xx_bootrom.bin"; + +static void npcm7xx_load_bootrom(MachineState *machine, NPCM7xxState *soc) +{ + g_autofree char *filename = NULL; + int ret; + + if (!bios_name) { + bios_name = npcm7xx_default_bootrom; + } + + filename = qemu_find_file(QEMU_FILE_TYPE_BIOS, bios_name); + if (!filename) { + error_report("Could not find ROM image '%s'", bios_name); + if (!machine->kernel_filename) { + /* We can't boot without a bootrom or a kernel image. */ + exit(1); + } + return; + } + ret = load_image_mr(filename, &soc->irom); + if (ret < 0) { + error_report("Failed to load ROM image '%s'", filename); + exit(1); + } +} + static void npcm7xx_connect_dram(NPCM7xxState *soc, MemoryRegion *dram) { memory_region_add_subregion(get_system_memory(), NPCM7XX_DRAM_BA, dram); @@ -XXX,XX +XXX,XX @@ static void npcm750_evb_init(MachineState *machine) npcm7xx_connect_dram(soc, machine->ram); qdev_realize(DEVICE(soc), NULL, &error_fatal); + npcm7xx_load_bootrom(machine, soc); npcm7xx_load_kernel(machine, soc); } @@ -XXX,XX +XXX,XX @@ static void quanta_gsj_init(MachineState *machine) npcm7xx_connect_dram(soc, machine->ram); qdev_realize(DEVICE(soc), NULL, &error_fatal); + npcm7xx_load_bootrom(machine, soc); npcm7xx_load_kernel(machine, soc); } -- 2.20.1 From: Havard Skinnemoen <hskinnemoen@google.com> This supports reading and writing OTP fuses and keys. Only fuse reading has been tested. Protection is not implemented. Reviewed-by: Avi Fishman <avi.fishman@nuvoton.com> Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Tested-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Tested-by: Alexander Bulekov <alxndr@bu.edu> Signed-off-by: Havard Skinnemoen <hskinnemoen@google.com> Message-id: 20200911052101.2602693-9-hskinnemoen@google.com Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- include/hw/arm/npcm7xx.h | 3 + include/hw/nvram/npcm7xx_otp.h | 79 ++++++ hw/arm/npcm7xx.c | 29 +++ hw/nvram/npcm7xx_otp.c | 440 +++++++++++++++++++++++++++++++++ hw/nvram/meson.build | 1 + 5 files changed, 552 insertions(+) create mode 100644 include/hw/nvram/npcm7xx_otp.h create mode 100644 hw/nvram/npcm7xx_otp.c diff --git a/include/hw/arm/npcm7xx.h b/include/hw/arm/npcm7xx.h index XXXXXXX..XXXXXXX 100644 --- a/include/hw/arm/npcm7xx.h +++ b/include/hw/arm/npcm7xx.h @@ -XXX,XX +XXX,XX @@ #include "hw/cpu/a9mpcore.h" #include "hw/misc/npcm7xx_clk.h" #include "hw/misc/npcm7xx_gcr.h" +#include "hw/nvram/npcm7xx_otp.h" #include "hw/timer/npcm7xx_timer.h" #include "target/arm/cpu.h" @@ -XXX,XX +XXX,XX @@ typedef struct NPCM7xxState { NPCM7xxGCRState gcr; NPCM7xxCLKState clk; NPCM7xxTimerCtrlState tim[3]; + NPCM7xxOTPState key_storage; + NPCM7xxOTPState fuse_array; } NPCM7xxState; #define TYPE_NPCM7XX "npcm7xx" diff --git a/include/hw/nvram/npcm7xx_otp.h b/include/hw/nvram/npcm7xx_otp.h new file mode 100644 index XXXXXXX..XXXXXXX --- /dev/null +++ b/include/hw/nvram/npcm7xx_otp.h @@ -XXX,XX +XXX,XX @@ +/* + * Nuvoton NPCM7xx OTP (Fuse Array) Interface + * + * Copyright 2020 Google LLC + * + * This program is free software; you can redistribute it and/or modify it + * under the terms of the GNU General Public License as published by the + * Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program 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 General Public License + * for more details. + */ +#ifndef NPCM7XX_OTP_H +#define NPCM7XX_OTP_H + +#include "exec/memory.h" +#include "hw/sysbus.h" + +/* Each OTP module holds 8192 bits of one-time programmable storage */ +#define NPCM7XX_OTP_ARRAY_BITS (8192) +#define NPCM7XX_OTP_ARRAY_BYTES (NPCM7XX_OTP_ARRAY_BITS / BITS_PER_BYTE) + +/* Fuse array offsets */ +#define NPCM7XX_FUSE_FUSTRAP (0) +#define NPCM7XX_FUSE_CP_FUSTRAP (12) +#define NPCM7XX_FUSE_DAC_CALIB (16) +#define NPCM7XX_FUSE_ADC_CALIB (24) +#define NPCM7XX_FUSE_DERIVATIVE (64) +#define NPCM7XX_FUSE_TEST_SIG (72) +#define NPCM7XX_FUSE_DIE_LOCATION (74) +#define NPCM7XX_FUSE_GP1 (80) +#define NPCM7XX_FUSE_GP2 (128) + +/* + * Number of registers in our device state structure. Don't change this without + * incrementing the version_id in the vmstate. + */ +#define NPCM7XX_OTP_NR_REGS (0x18 / sizeof(uint32_t)) + +/** + * struct NPCM7xxOTPState - Device state for one OTP module. + * @parent: System bus device. + * @mmio: Memory region through which registers are accessed. + * @regs: Register contents. + * @array: OTP storage array. + */ +typedef struct NPCM7xxOTPState { + SysBusDevice parent; + + MemoryRegion mmio; + uint32_t regs[NPCM7XX_OTP_NR_REGS]; + uint8_t array[NPCM7XX_OTP_ARRAY_BYTES]; +} NPCM7xxOTPState; + +#define TYPE_NPCM7XX_OTP "npcm7xx-otp" +#define NPCM7XX_OTP(obj) OBJECT_CHECK(NPCM7xxOTPState, (obj), TYPE_NPCM7XX_OTP) + +#define TYPE_NPCM7XX_KEY_STORAGE "npcm7xx-key-storage" +#define TYPE_NPCM7XX_FUSE_ARRAY "npcm7xx-fuse-array" + +typedef struct NPCM7xxOTPClass NPCM7xxOTPClass; + +/** + * npcm7xx_otp_array_write - ECC encode and write data to OTP array. + * @s: OTP module. + * @data: Data to be encoded and written. + * @offset: Offset of first byte to be written in the OTP array. + * @len: Number of bytes before ECC encoding. + * + * Each nibble of data is encoded into a byte, so the number of bytes written + * to the array will be @len * 2. + */ +extern void npcm7xx_otp_array_write(NPCM7xxOTPState *s, const void *data, + unsigned int offset, unsigned int len); + +#endif /* NPCM7XX_OTP_H */ diff --git a/hw/arm/npcm7xx.c b/hw/arm/npcm7xx.c index XXXXXXX..XXXXXXX 100644 --- a/hw/arm/npcm7xx.c +++ b/hw/arm/npcm7xx.c @@ -XXX,XX +XXX,XX @@ #define NPCM7XX_MMIO_BA (0x80000000) #define NPCM7XX_MMIO_SZ (0x7ffd0000) +/* OTP key storage and fuse strap array */ +#define NPCM7XX_OTP1_BA (0xf0189000) +#define NPCM7XX_OTP2_BA (0xf018a000) + /* Core system modules. */ #define NPCM7XX_L2C_BA (0xf03fc000) #define NPCM7XX_CPUP_BA (0xf03fe000) @@ -XXX,XX +XXX,XX @@ void npcm7xx_load_kernel(MachineState *machine, NPCM7xxState *soc) arm_load_kernel(&soc->cpu[0], machine, &npcm7xx_binfo); } +static void npcm7xx_init_fuses(NPCM7xxState *s) +{ + NPCM7xxClass *nc = NPCM7XX_GET_CLASS(s); + uint32_t value; + + /* + * The initial mask of disabled modules indicates the chip derivative (e.g. + * NPCM750 or NPCM730). + */ + value = tswap32(nc->disabled_modules); + npcm7xx_otp_array_write(&s->fuse_array, &value, NPCM7XX_FUSE_DERIVATIVE, + sizeof(value)); +} + static qemu_irq npcm7xx_irq(NPCM7xxState *s, int n) { return qdev_get_gpio_in(DEVICE(&s->a9mpcore), n); @@ -XXX,XX +XXX,XX @@ static void npcm7xx_init(Object *obj) object_property_add_alias(obj, "power-on-straps", OBJECT(&s->gcr), "power-on-straps"); object_initialize_child(obj, "clk", &s->clk, TYPE_NPCM7XX_CLK); + object_initialize_child(obj, "otp1", &s->key_storage, + TYPE_NPCM7XX_KEY_STORAGE); + object_initialize_child(obj, "otp2", &s->fuse_array, + TYPE_NPCM7XX_FUSE_ARRAY); for (i = 0; i < ARRAY_SIZE(s->tim); i++) { object_initialize_child(obj, "tim[*]", &s->tim[i], TYPE_NPCM7XX_TIMER); @@ -XXX,XX +XXX,XX @@ static void npcm7xx_realize(DeviceState *dev, Error **errp) sysbus_realize(SYS_BUS_DEVICE(&s->clk), &error_abort); sysbus_mmio_map(SYS_BUS_DEVICE(&s->clk), 0, NPCM7XX_CLK_BA); + /* OTP key storage and fuse strap array. Cannot fail. */ + sysbus_realize(SYS_BUS_DEVICE(&s->key_storage), &error_abort); + sysbus_mmio_map(SYS_BUS_DEVICE(&s->key_storage), 0, NPCM7XX_OTP1_BA); + sysbus_realize(SYS_BUS_DEVICE(&s->fuse_array), &error_abort); + sysbus_mmio_map(SYS_BUS_DEVICE(&s->fuse_array), 0, NPCM7XX_OTP2_BA); + npcm7xx_init_fuses(s); + /* Timer Modules (TIM). Cannot fail. */ QEMU_BUILD_BUG_ON(ARRAY_SIZE(npcm7xx_tim_addr) != ARRAY_SIZE(s->tim)); for (i = 0; i < ARRAY_SIZE(s->tim); i++) { diff --git a/hw/nvram/npcm7xx_otp.c b/hw/nvram/npcm7xx_otp.c new file mode 100644 index XXXXXXX..XXXXXXX --- /dev/null +++ b/hw/nvram/npcm7xx_otp.c @@ -XXX,XX +XXX,XX @@ +/* + * Nuvoton NPCM7xx OTP (Fuse Array) Interface + * + * Copyright 2020 Google LLC + * + * This program is free software; you can redistribute it and/or modify it + * under the terms of the GNU General Public License as published by the + * Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program 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 General Public License + * for more details. + */ + +#include "qemu/osdep.h" + +#include "hw/nvram/npcm7xx_otp.h" +#include "migration/vmstate.h" +#include "qapi/error.h" +#include "qemu/bitops.h" +#include "qemu/log.h" +#include "qemu/module.h" +#include "qemu/units.h" + +/* Each module has 4 KiB of register space. Only a fraction of it is used. */ +#define NPCM7XX_OTP_REGS_SIZE (4 * KiB) + +/* 32-bit register indices. */ +typedef enum NPCM7xxOTPRegister { + NPCM7XX_OTP_FST, + NPCM7XX_OTP_FADDR, + NPCM7XX_OTP_FDATA, + NPCM7XX_OTP_FCFG, + /* Offset 0x10 is FKEYIND in OTP1, FUSTRAP in OTP2 */ + NPCM7XX_OTP_FKEYIND = 0x0010 / sizeof(uint32_t), + NPCM7XX_OTP_FUSTRAP = 0x0010 / sizeof(uint32_t), + NPCM7XX_OTP_FCTL, + NPCM7XX_OTP_REGS_END, +} NPCM7xxOTPRegister; + +/* Register field definitions. */ +#define FST_RIEN BIT(2) +#define FST_RDST BIT(1) +#define FST_RDY BIT(0) +#define FST_RO_MASK (FST_RDST | FST_RDY) + +#define FADDR_BYTEADDR(rv) extract32((rv), 0, 10) +#define FADDR_BITPOS(rv) extract32((rv), 10, 3) + +#define FDATA_CLEAR 0x00000001 + +#define FCFG_FDIS BIT(31) +#define FCFG_FCFGLK_MASK 0x00ff0000 + +#define FCTL_PROG_CMD1 0x00000001 +#define FCTL_PROG_CMD2 0xbf79e5d0 +#define FCTL_READ_CMD 0x00000002 + +/** + * struct NPCM7xxOTPClass - OTP module class. + * @parent: System bus device class. + * @mmio_ops: MMIO register operations for this type of module. + * + * The two OTP modules (key-storage and fuse-array) have slightly different + * behavior, so we give them different MMIO register operations. + */ +struct NPCM7xxOTPClass { + SysBusDeviceClass parent; + + const MemoryRegionOps *mmio_ops; +}; + +#define NPCM7XX_OTP_CLASS(klass) \ + OBJECT_CLASS_CHECK(NPCM7xxOTPClass, (klass), TYPE_NPCM7XX_OTP) +#define NPCM7XX_OTP_GET_CLASS(obj) \ + OBJECT_GET_CLASS(NPCM7xxOTPClass, (obj), TYPE_NPCM7XX_OTP) + +static uint8_t ecc_encode_nibble(uint8_t n) +{ + uint8_t result = n; + + result |= (((n >> 0) & 1) ^ ((n >> 1) & 1)) << 4; + result |= (((n >> 2) & 1) ^ ((n >> 3) & 1)) << 5; + result |= (((n >> 0) & 1) ^ ((n >> 2) & 1)) << 6; + result |= (((n >> 1) & 1) ^ ((n >> 3) & 1)) << 7; + + return result; +} + +void npcm7xx_otp_array_write(NPCM7xxOTPState *s, const void *data, + unsigned int offset, unsigned int len) +{ + const uint8_t *src = data; + uint8_t *dst = &s->array[offset]; + + while (len-- > 0) { + uint8_t c = *src++; + + *dst++ = ecc_encode_nibble(extract8(c, 0, 4)); + *dst++ = ecc_encode_nibble(extract8(c, 4, 4)); + } +} + +/* Common register read handler for both OTP classes. */ +static uint64_t npcm7xx_otp_read(NPCM7xxOTPState *s, NPCM7xxOTPRegister reg) +{ + uint32_t value = 0; + + switch (reg) { + case NPCM7XX_OTP_FST: + case NPCM7XX_OTP_FADDR: + case NPCM7XX_OTP_FDATA: + case NPCM7XX_OTP_FCFG: + value = s->regs[reg]; + break; + + case NPCM7XX_OTP_FCTL: + qemu_log_mask(LOG_GUEST_ERROR, + "%s: read from write-only FCTL register\n", + DEVICE(s)->canonical_path); + break; + + default: + qemu_log_mask(LOG_GUEST_ERROR, "%s: read from invalid offset 0x%zx\n", + DEVICE(s)->canonical_path, reg * sizeof(uint32_t)); + break; + } + + return value; +} + +/* Read a byte from the OTP array into the data register. */ +static void npcm7xx_otp_read_array(NPCM7xxOTPState *s) +{ + uint32_t faddr = s->regs[NPCM7XX_OTP_FADDR]; + + s->regs[NPCM7XX_OTP_FDATA] = s->array[FADDR_BYTEADDR(faddr)]; + s->regs[NPCM7XX_OTP_FST] |= FST_RDST | FST_RDY; +} + +/* Program a byte from the data register into the OTP array. */ +static void npcm7xx_otp_program_array(NPCM7xxOTPState *s) +{ + uint32_t faddr = s->regs[NPCM7XX_OTP_FADDR]; + + /* Bits can only go 0->1, never 1->0. */ + s->array[FADDR_BYTEADDR(faddr)] |= (1U << FADDR_BITPOS(faddr)); + s->regs[NPCM7XX_OTP_FST] |= FST_RDST | FST_RDY; +} + +/* Compute the next value of the FCFG register. */ +static uint32_t npcm7xx_otp_compute_fcfg(uint32_t cur_value, uint32_t new_value) +{ + uint32_t lock_mask; + uint32_t value; + + /* + * FCFGLK holds sticky bits 16..23, indicating which bits in FPRGLK (8..15) + * and FRDLK (0..7) that are read-only. + */ + lock_mask = (cur_value & FCFG_FCFGLK_MASK) >> 8; + lock_mask |= lock_mask >> 8; + /* FDIS and FCFGLK bits are sticky (write 1 to set; can't clear). */ + value = cur_value & (FCFG_FDIS | FCFG_FCFGLK_MASK); + /* Preserve read-only bits in FPRGLK and FRDLK */ + value |= cur_value & lock_mask; + /* Set all bits that aren't read-only. */ + value |= new_value & ~lock_mask; + + return value; +} + +/* Common register write handler for both OTP classes. */ +static void npcm7xx_otp_write(NPCM7xxOTPState *s, NPCM7xxOTPRegister reg, + uint32_t value) +{ + switch (reg) { + case NPCM7XX_OTP_FST: + /* RDST is cleared by writing 1 to it. */ + if (value & FST_RDST) { + s->regs[NPCM7XX_OTP_FST] &= ~FST_RDST; + } + /* Preserve read-only and write-one-to-clear bits */ + value &= ~FST_RO_MASK; + value |= s->regs[NPCM7XX_OTP_FST] & FST_RO_MASK; + break; + + case NPCM7XX_OTP_FADDR: + break; + + case NPCM7XX_OTP_FDATA: + /* + * This register is cleared by writing a magic value to it; no other + * values can be written. + */ + if (value == FDATA_CLEAR) { + value = 0; + } else { + value = s->regs[NPCM7XX_OTP_FDATA]; + } + break; + + case NPCM7XX_OTP_FCFG: + value = npcm7xx_otp_compute_fcfg(s->regs[NPCM7XX_OTP_FCFG], value); + break; + + case NPCM7XX_OTP_FCTL: + switch (value) { + case FCTL_READ_CMD: + npcm7xx_otp_read_array(s); + break; + + case FCTL_PROG_CMD1: + /* + * Programming requires writing two separate magic values to this + * register; this is the first one. Just store it so it can be + * verified later when the second magic value is received. + */ + break; + + case FCTL_PROG_CMD2: + /* + * Only initiate programming if we received the first half of the + * command immediately before this one. + */ + if (s->regs[NPCM7XX_OTP_FCTL] == FCTL_PROG_CMD1) { + npcm7xx_otp_program_array(s); + } + break; + + default: + qemu_log_mask(LOG_GUEST_ERROR, + "%s: unrecognized FCNTL value 0x%" PRIx32 "\n", + DEVICE(s)->canonical_path, value); + break; + } + if (value != FCTL_PROG_CMD1) { + value = 0; + } + break; + + default: + qemu_log_mask(LOG_GUEST_ERROR, "%s: write to invalid offset 0x%zx\n", + DEVICE(s)->canonical_path, reg * sizeof(uint32_t)); + return; + } + + s->regs[reg] = value; +} + +/* Register read handler specific to the fuse array OTP module. */ +static uint64_t npcm7xx_fuse_array_read(void *opaque, hwaddr addr, + unsigned int size) +{ + NPCM7xxOTPRegister reg = addr / sizeof(uint32_t); + NPCM7xxOTPState *s = opaque; + uint32_t value; + + /* + * Only the Fuse Strap register needs special handling; all other registers + * work the same way for both kinds of OTP modules. + */ + if (reg != NPCM7XX_OTP_FUSTRAP) { + value = npcm7xx_otp_read(s, reg); + } else { + /* FUSTRAP is stored as three copies in the OTP array. */ + uint32_t fustrap[3]; + + memcpy(fustrap, &s->array[0], sizeof(fustrap)); + + /* Determine value by a majority vote on each bit. */ + value = (fustrap[0] & fustrap[1]) | (fustrap[0] & fustrap[2]) | + (fustrap[1] & fustrap[2]); + } + + return value; +} + +/* Register write handler specific to the fuse array OTP module. */ +static void npcm7xx_fuse_array_write(void *opaque, hwaddr addr, uint64_t v, + unsigned int size) +{ + NPCM7xxOTPRegister reg = addr / sizeof(uint32_t); + NPCM7xxOTPState *s = opaque; + + /* + * The Fuse Strap register is read-only. Other registers are handled by + * common code. + */ + if (reg != NPCM7XX_OTP_FUSTRAP) { + npcm7xx_otp_write(s, reg, v); + } +} + +static const MemoryRegionOps npcm7xx_fuse_array_ops = { + .read = npcm7xx_fuse_array_read, + .write = npcm7xx_fuse_array_write, + .endianness = DEVICE_LITTLE_ENDIAN, + .valid = { + .min_access_size = 4, + .max_access_size = 4, + .unaligned = false, + }, +}; + +/* Register read handler specific to the key storage OTP module. */ +static uint64_t npcm7xx_key_storage_read(void *opaque, hwaddr addr, + unsigned int size) +{ + NPCM7xxOTPRegister reg = addr / sizeof(uint32_t); + NPCM7xxOTPState *s = opaque; + + /* + * Only the Fuse Key Index register needs special handling; all other + * registers work the same way for both kinds of OTP modules. + */ + if (reg != NPCM7XX_OTP_FKEYIND) { + return npcm7xx_otp_read(s, reg); + } + + qemu_log_mask(LOG_UNIMP, "%s: FKEYIND is not implemented\n", __func__); + + return s->regs[NPCM7XX_OTP_FKEYIND]; +} + +/* Register write handler specific to the key storage OTP module. */ +static void npcm7xx_key_storage_write(void *opaque, hwaddr addr, uint64_t v, + unsigned int size) +{ + NPCM7xxOTPRegister reg = addr / sizeof(uint32_t); + NPCM7xxOTPState *s = opaque; + + /* + * Only the Fuse Key Index register needs special handling; all other + * registers work the same way for both kinds of OTP modules. + */ + if (reg != NPCM7XX_OTP_FKEYIND) { + npcm7xx_otp_write(s, reg, v); + return; + } + + qemu_log_mask(LOG_UNIMP, "%s: FKEYIND is not implemented\n", __func__); + + s->regs[NPCM7XX_OTP_FKEYIND] = v; +} + +static const MemoryRegionOps npcm7xx_key_storage_ops = { + .read = npcm7xx_key_storage_read, + .write = npcm7xx_key_storage_write, + .endianness = DEVICE_LITTLE_ENDIAN, + .valid = { + .min_access_size = 4, + .max_access_size = 4, + .unaligned = false, + }, +}; + +static void npcm7xx_otp_enter_reset(Object *obj, ResetType type) +{ + NPCM7xxOTPState *s = NPCM7XX_OTP(obj); + + memset(s->regs, 0, sizeof(s->regs)); + + s->regs[NPCM7XX_OTP_FST] = 0x00000001; + s->regs[NPCM7XX_OTP_FCFG] = 0x20000000; +} + +static void npcm7xx_otp_realize(DeviceState *dev, Error **errp) +{ + NPCM7xxOTPClass *oc = NPCM7XX_OTP_GET_CLASS(dev); + NPCM7xxOTPState *s = NPCM7XX_OTP(dev); + SysBusDevice *sbd = &s->parent; + + memset(s->array, 0, sizeof(s->array)); + + memory_region_init_io(&s->mmio, OBJECT(s), oc->mmio_ops, s, "regs", + NPCM7XX_OTP_REGS_SIZE); + sysbus_init_mmio(sbd, &s->mmio); +} + +static const VMStateDescription vmstate_npcm7xx_otp = { + .name = "npcm7xx-otp", + .version_id = 0, + .minimum_version_id = 0, + .fields = (VMStateField[]) { + VMSTATE_UINT32_ARRAY(regs, NPCM7xxOTPState, NPCM7XX_OTP_NR_REGS), + VMSTATE_UINT8_ARRAY(array, NPCM7xxOTPState, NPCM7XX_OTP_ARRAY_BYTES), + VMSTATE_END_OF_LIST(), + }, +}; + +static void npcm7xx_otp_class_init(ObjectClass *klass, void *data) +{ + ResettableClass *rc = RESETTABLE_CLASS(klass); + DeviceClass *dc = DEVICE_CLASS(klass); + + QEMU_BUILD_BUG_ON(NPCM7XX_OTP_REGS_END > NPCM7XX_OTP_NR_REGS); + + dc->realize = npcm7xx_otp_realize; + dc->vmsd = &vmstate_npcm7xx_otp; + rc->phases.enter = npcm7xx_otp_enter_reset; +} + +static void npcm7xx_key_storage_class_init(ObjectClass *klass, void *data) +{ + NPCM7xxOTPClass *oc = NPCM7XX_OTP_CLASS(klass); + + oc->mmio_ops = &npcm7xx_key_storage_ops; +} + +static void npcm7xx_fuse_array_class_init(ObjectClass *klass, void *data) +{ + NPCM7xxOTPClass *oc = NPCM7XX_OTP_CLASS(klass); + + oc->mmio_ops = &npcm7xx_fuse_array_ops; +} + +static const TypeInfo npcm7xx_otp_types[] = { + { + .name = TYPE_NPCM7XX_OTP, + .parent = TYPE_SYS_BUS_DEVICE, + .instance_size = sizeof(NPCM7xxOTPState), + .class_size = sizeof(NPCM7xxOTPClass), + .class_init = npcm7xx_otp_class_init, + .abstract = true, + }, + { + .name = TYPE_NPCM7XX_KEY_STORAGE, + .parent = TYPE_NPCM7XX_OTP, + .class_init = npcm7xx_key_storage_class_init, + }, + { + .name = TYPE_NPCM7XX_FUSE_ARRAY, + .parent = TYPE_NPCM7XX_OTP, + .class_init = npcm7xx_fuse_array_class_init, + }, +}; +DEFINE_TYPES(npcm7xx_otp_types); diff --git a/hw/nvram/meson.build b/hw/nvram/meson.build index XXXXXXX..XXXXXXX 100644 --- a/hw/nvram/meson.build +++ b/hw/nvram/meson.build @@ -XXX,XX +XXX,XX @@ softmmu_ss.add(when: 'CONFIG_DS1225Y', if_true: files('ds1225y.c')) softmmu_ss.add(when: 'CONFIG_NMC93XX_EEPROM', if_true: files('eeprom93xx.c')) softmmu_ss.add(when: 'CONFIG_AT24C', if_true: files('eeprom_at24c.c')) softmmu_ss.add(when: 'CONFIG_MAC_NVRAM', if_true: files('mac_nvram.c')) +softmmu_ss.add(when: 'CONFIG_NPCM7XX', if_true: files('npcm7xx_otp.c')) softmmu_ss.add(when: 'CONFIG_NRF51_SOC', if_true: files('nrf51_nvm.c')) specific_ss.add(when: 'CONFIG_PSERIES', if_true: files('spapr_nvram.c')) -- 2.20.1 From: Havard Skinnemoen <hskinnemoen@google.com> This just implements the bare minimum to cause the boot block to skip memory initialization. Reviewed-by: Tyrone Ting <kfting@nuvoton.com> Reviewed-by: Cédric Le Goater <clg@kaod.org> Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Tested-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Tested-by: Alexander Bulekov <alxndr@bu.edu> Signed-off-by: Havard Skinnemoen <hskinnemoen@google.com> Message-id: 20200911052101.2602693-10-hskinnemoen@google.com Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- include/hw/arm/npcm7xx.h | 2 + include/hw/mem/npcm7xx_mc.h | 36 ++++++++++++++++ hw/arm/npcm7xx.c | 6 +++ hw/mem/npcm7xx_mc.c | 84 +++++++++++++++++++++++++++++++++++++ hw/mem/meson.build | 1 + 5 files changed, 129 insertions(+) create mode 100644 include/hw/mem/npcm7xx_mc.h create mode 100644 hw/mem/npcm7xx_mc.c diff --git a/include/hw/arm/npcm7xx.h b/include/hw/arm/npcm7xx.h index XXXXXXX..XXXXXXX 100644 --- a/include/hw/arm/npcm7xx.h +++ b/include/hw/arm/npcm7xx.h @@ -XXX,XX +XXX,XX @@ #include "hw/boards.h" #include "hw/cpu/a9mpcore.h" +#include "hw/mem/npcm7xx_mc.h" #include "hw/misc/npcm7xx_clk.h" #include "hw/misc/npcm7xx_gcr.h" #include "hw/nvram/npcm7xx_otp.h" @@ -XXX,XX +XXX,XX @@ typedef struct NPCM7xxState { NPCM7xxTimerCtrlState tim[3]; NPCM7xxOTPState key_storage; NPCM7xxOTPState fuse_array; + NPCM7xxMCState mc; } NPCM7xxState; #define TYPE_NPCM7XX "npcm7xx" diff --git a/include/hw/mem/npcm7xx_mc.h b/include/hw/mem/npcm7xx_mc.h new file mode 100644 index XXXXXXX..XXXXXXX --- /dev/null +++ b/include/hw/mem/npcm7xx_mc.h @@ -XXX,XX +XXX,XX @@ +/* + * Nuvoton NPCM7xx Memory Controller stub + * + * Copyright 2020 Google LLC + * + * This program is free software; you can redistribute it and/or modify it + * under the terms of the GNU General Public License as published by the + * Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program 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 General Public License + * for more details. + */ +#ifndef NPCM7XX_MC_H +#define NPCM7XX_MC_H + +#include "exec/memory.h" +#include "hw/sysbus.h" + +/** + * struct NPCM7xxMCState - Device state for the memory controller. + * @parent: System bus device. + * @mmio: Memory region through which registers are accessed. + */ +typedef struct NPCM7xxMCState { + SysBusDevice parent; + + MemoryRegion mmio; +} NPCM7xxMCState; + +#define TYPE_NPCM7XX_MC "npcm7xx-mc" +#define NPCM7XX_MC(obj) OBJECT_CHECK(NPCM7xxMCState, (obj), TYPE_NPCM7XX_MC) + +#endif /* NPCM7XX_MC_H */ diff --git a/hw/arm/npcm7xx.c b/hw/arm/npcm7xx.c index XXXXXXX..XXXXXXX 100644 --- a/hw/arm/npcm7xx.c +++ b/hw/arm/npcm7xx.c @@ -XXX,XX +XXX,XX @@ #define NPCM7XX_CPUP_BA (0xf03fe000) #define NPCM7XX_GCR_BA (0xf0800000) #define NPCM7XX_CLK_BA (0xf0801000) +#define NPCM7XX_MC_BA (0xf0824000) /* Internal AHB SRAM */ #define NPCM7XX_RAM3_BA (0xc0008000) @@ -XXX,XX +XXX,XX @@ static void npcm7xx_init(Object *obj) TYPE_NPCM7XX_KEY_STORAGE); object_initialize_child(obj, "otp2", &s->fuse_array, TYPE_NPCM7XX_FUSE_ARRAY); + object_initialize_child(obj, "mc", &s->mc, TYPE_NPCM7XX_MC); for (i = 0; i < ARRAY_SIZE(s->tim); i++) { object_initialize_child(obj, "tim[*]", &s->tim[i], TYPE_NPCM7XX_TIMER); @@ -XXX,XX +XXX,XX @@ static void npcm7xx_realize(DeviceState *dev, Error **errp) sysbus_mmio_map(SYS_BUS_DEVICE(&s->fuse_array), 0, NPCM7XX_OTP2_BA); npcm7xx_init_fuses(s); + /* Fake Memory Controller (MC). Cannot fail. */ + sysbus_realize(SYS_BUS_DEVICE(&s->mc), &error_abort); + sysbus_mmio_map(SYS_BUS_DEVICE(&s->mc), 0, NPCM7XX_MC_BA); + /* Timer Modules (TIM). Cannot fail. */ QEMU_BUILD_BUG_ON(ARRAY_SIZE(npcm7xx_tim_addr) != ARRAY_SIZE(s->tim)); for (i = 0; i < ARRAY_SIZE(s->tim); i++) { diff --git a/hw/mem/npcm7xx_mc.c b/hw/mem/npcm7xx_mc.c new file mode 100644 index XXXXXXX..XXXXXXX --- /dev/null +++ b/hw/mem/npcm7xx_mc.c @@ -XXX,XX +XXX,XX @@ +/* + * Nuvoton NPCM7xx Memory Controller stub + * + * Copyright 2020 Google LLC + * + * This program is free software; you can redistribute it and/or modify it + * under the terms of the GNU General Public License as published by the + * Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program 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 General Public License + * for more details. + */ + +#include "qemu/osdep.h" + +#include "hw/mem/npcm7xx_mc.h" +#include "qapi/error.h" +#include "qemu/log.h" +#include "qemu/module.h" +#include "qemu/units.h" + +#define NPCM7XX_MC_REGS_SIZE (4 * KiB) + +static uint64_t npcm7xx_mc_read(void *opaque, hwaddr addr, unsigned int size) +{ + /* + * If bits 8..11 @ offset 0 are not zero, the boot block thinks the memory + * controller has already been initialized and will skip DDR training. + */ + if (addr == 0) { + return 0x100; + } + + qemu_log_mask(LOG_UNIMP, "%s: mostly unimplemented\n", __func__); + + return 0; +} + +static void npcm7xx_mc_write(void *opaque, hwaddr addr, uint64_t v, + unsigned int size) +{ + qemu_log_mask(LOG_UNIMP, "%s: mostly unimplemented\n", __func__); +} + +static const MemoryRegionOps npcm7xx_mc_ops = { + .read = npcm7xx_mc_read, + .write = npcm7xx_mc_write, + .endianness = DEVICE_LITTLE_ENDIAN, + .valid = { + .min_access_size = 4, + .max_access_size = 4, + .unaligned = false, + }, +}; + +static void npcm7xx_mc_realize(DeviceState *dev, Error **errp) +{ + NPCM7xxMCState *s = NPCM7XX_MC(dev); + + memory_region_init_io(&s->mmio, OBJECT(s), &npcm7xx_mc_ops, s, "regs", + NPCM7XX_MC_REGS_SIZE); + sysbus_init_mmio(&s->parent, &s->mmio); +} + +static void npcm7xx_mc_class_init(ObjectClass *klass, void *data) +{ + DeviceClass *dc = DEVICE_CLASS(klass); + + dc->desc = "NPCM7xx Memory Controller stub"; + dc->realize = npcm7xx_mc_realize; +} + +static const TypeInfo npcm7xx_mc_types[] = { + { + .name = TYPE_NPCM7XX_MC, + .parent = TYPE_SYS_BUS_DEVICE, + .instance_size = sizeof(NPCM7xxMCState), + .class_init = npcm7xx_mc_class_init, + }, +}; +DEFINE_TYPES(npcm7xx_mc_types); diff --git a/hw/mem/meson.build b/hw/mem/meson.build index XXXXXXX..XXXXXXX 100644 --- a/hw/mem/meson.build +++ b/hw/mem/meson.build @@ -XXX,XX +XXX,XX @@ mem_ss = ss.source_set() mem_ss.add(files('memory-device.c')) mem_ss.add(when: 'CONFIG_DIMM', if_true: files('pc-dimm.c')) +mem_ss.add(when: 'CONFIG_NPCM7XX', if_true: files('npcm7xx_mc.c')) mem_ss.add(when: 'CONFIG_NVDIMM', if_true: files('nvdimm.c')) softmmu_ss.add_all(when: 'CONFIG_MEM_DEVICE', if_true: mem_ss) -- 2.20.1 From: Havard Skinnemoen <hskinnemoen@google.com> This implements a device model for the NPCM7xx SPI flash controller. Direct reads and writes, and user-mode transactions have been tested in various modes. Protection features are not implemented yet. All the FIU instances are available in the SoC's address space, regardless of whether or not they're connected to actual flash chips. Reviewed-by: Tyrone Ting <kfting@nuvoton.com> Reviewed-by: Cédric Le Goater <clg@kaod.org> Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Tested-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Tested-by: Alexander Bulekov <alxndr@bu.edu> Signed-off-by: Havard Skinnemoen <hskinnemoen@google.com> Message-id: 20200911052101.2602693-11-hskinnemoen@google.com Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- include/hw/arm/npcm7xx.h | 2 + include/hw/ssi/npcm7xx_fiu.h | 73 +++++ hw/arm/npcm7xx.c | 58 ++++ hw/ssi/npcm7xx_fiu.c | 572 +++++++++++++++++++++++++++++++++++ hw/arm/Kconfig | 1 + hw/ssi/meson.build | 1 + hw/ssi/trace-events | 11 + 7 files changed, 718 insertions(+) create mode 100644 include/hw/ssi/npcm7xx_fiu.h create mode 100644 hw/ssi/npcm7xx_fiu.c diff --git a/include/hw/arm/npcm7xx.h b/include/hw/arm/npcm7xx.h index XXXXXXX..XXXXXXX 100644 --- a/include/hw/arm/npcm7xx.h +++ b/include/hw/arm/npcm7xx.h @@ -XXX,XX +XXX,XX @@ #include "hw/misc/npcm7xx_gcr.h" #include "hw/nvram/npcm7xx_otp.h" #include "hw/timer/npcm7xx_timer.h" +#include "hw/ssi/npcm7xx_fiu.h" #include "target/arm/cpu.h" #define NPCM7XX_MAX_NUM_CPUS (2) @@ -XXX,XX +XXX,XX @@ typedef struct NPCM7xxState { NPCM7xxOTPState key_storage; NPCM7xxOTPState fuse_array; NPCM7xxMCState mc; + NPCM7xxFIUState fiu[2]; } NPCM7xxState; #define TYPE_NPCM7XX "npcm7xx" diff --git a/include/hw/ssi/npcm7xx_fiu.h b/include/hw/ssi/npcm7xx_fiu.h new file mode 100644 index XXXXXXX..XXXXXXX --- /dev/null +++ b/include/hw/ssi/npcm7xx_fiu.h @@ -XXX,XX +XXX,XX @@ +/* + * Nuvoton NPCM7xx Flash Interface Unit (FIU) + * + * Copyright 2020 Google LLC + * + * This program is free software; you can redistribute it and/or modify it + * under the terms of the GNU General Public License as published by the + * Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program 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 General Public License + * for more details. + */ +#ifndef NPCM7XX_FIU_H +#define NPCM7XX_FIU_H + +#include "hw/ssi/ssi.h" +#include "hw/sysbus.h" + +/* + * Number of registers in our device state structure. Don't change this without + * incrementing the version_id in the vmstate. + */ +#define NPCM7XX_FIU_NR_REGS (0x7c / sizeof(uint32_t)) + +typedef struct NPCM7xxFIUState NPCM7xxFIUState; + +/** + * struct NPCM7xxFIUFlash - Per-chipselect flash controller state. + * @direct_access: Memory region for direct flash access. + * @fiu: Pointer to flash controller shared state. + */ +typedef struct NPCM7xxFIUFlash { + MemoryRegion direct_access; + NPCM7xxFIUState *fiu; +} NPCM7xxFIUFlash; + +/** + * NPCM7xxFIUState - Device state for one Flash Interface Unit. + * @parent: System bus device. + * @mmio: Memory region for register access. + * @cs_count: Number of flash chips that may be connected to this module. + * @active_cs: Currently active chip select, or -1 if no chip is selected. + * @cs_lines: GPIO lines that may be wired to flash chips. + * @flash: Array of @cs_count per-flash-chip state objects. + * @spi: The SPI bus mastered by this controller. + * @regs: Register contents. + * + * Each FIU has a shared bank of registers, and controls up to four chip + * selects. Each chip select has a dedicated memory region which may be used to + * read and write the flash connected to that chip select as if it were memory. + */ +struct NPCM7xxFIUState { + SysBusDevice parent; + + MemoryRegion mmio; + + int32_t cs_count; + int32_t active_cs; + qemu_irq *cs_lines; + NPCM7xxFIUFlash *flash; + + SSIBus *spi; + + uint32_t regs[NPCM7XX_FIU_NR_REGS]; +}; + +#define TYPE_NPCM7XX_FIU "npcm7xx-fiu" +#define NPCM7XX_FIU(obj) OBJECT_CHECK(NPCM7xxFIUState, (obj), TYPE_NPCM7XX_FIU) + +#endif /* NPCM7XX_FIU_H */ diff --git a/hw/arm/npcm7xx.c b/hw/arm/npcm7xx.c index XXXXXXX..XXXXXXX 100644 --- a/hw/arm/npcm7xx.c +++ b/hw/arm/npcm7xx.c @@ -XXX,XX +XXX,XX @@ static const hwaddr npcm7xx_uart_addr[] = { 0xf0004000, }; +/* Direct memory-mapped access to SPI0 CS0-1. */ +static const hwaddr npcm7xx_fiu0_flash_addr[] = { + 0x80000000, /* CS0 */ + 0x88000000, /* CS1 */ +}; + +/* Direct memory-mapped access to SPI3 CS0-3. */ +static const hwaddr npcm7xx_fiu3_flash_addr[] = { + 0xa0000000, /* CS0 */ + 0xa8000000, /* CS1 */ + 0xb0000000, /* CS2 */ + 0xb8000000, /* CS3 */ +}; + +static const struct { + const char *name; + hwaddr regs_addr; + int cs_count; + const hwaddr *flash_addr; +} npcm7xx_fiu[] = { + { + .name = "fiu0", + .regs_addr = 0xfb000000, + .cs_count = ARRAY_SIZE(npcm7xx_fiu0_flash_addr), + .flash_addr = npcm7xx_fiu0_flash_addr, + }, { + .name = "fiu3", + .regs_addr = 0xc0000000, + .cs_count = ARRAY_SIZE(npcm7xx_fiu3_flash_addr), + .flash_addr = npcm7xx_fiu3_flash_addr, + }, +}; + static void npcm7xx_write_secondary_boot(ARMCPU *cpu, const struct arm_boot_info *info) { @@ -XXX,XX +XXX,XX @@ static void npcm7xx_init(Object *obj) for (i = 0; i < ARRAY_SIZE(s->tim); i++) { object_initialize_child(obj, "tim[*]", &s->tim[i], TYPE_NPCM7XX_TIMER); } + + QEMU_BUILD_BUG_ON(ARRAY_SIZE(npcm7xx_fiu) != ARRAY_SIZE(s->fiu)); + for (i = 0; i < ARRAY_SIZE(s->fiu); i++) { + object_initialize_child(obj, npcm7xx_fiu[i].name, &s->fiu[i], + TYPE_NPCM7XX_FIU); + } } static void npcm7xx_realize(DeviceState *dev, Error **errp) @@ -XXX,XX +XXX,XX @@ static void npcm7xx_realize(DeviceState *dev, Error **errp) serial_hd(i), DEVICE_LITTLE_ENDIAN); } + /* + * Flash Interface Unit (FIU). Can fail if incorrect number of chip selects + * specified, but this is a programming error. + */ + QEMU_BUILD_BUG_ON(ARRAY_SIZE(npcm7xx_fiu) != ARRAY_SIZE(s->fiu)); + for (i = 0; i < ARRAY_SIZE(s->fiu); i++) { + SysBusDevice *sbd = SYS_BUS_DEVICE(&s->fiu[i]); + int j; + + object_property_set_int(OBJECT(sbd), "cs-count", + npcm7xx_fiu[i].cs_count, &error_abort); + sysbus_realize(sbd, &error_abort); + + sysbus_mmio_map(sbd, 0, npcm7xx_fiu[i].regs_addr); + for (j = 0; j < npcm7xx_fiu[i].cs_count; j++) { + sysbus_mmio_map(sbd, j + 1, npcm7xx_fiu[i].flash_addr[j]); + } + } + /* RAM2 (SRAM) */ memory_region_init_ram(&s->sram, OBJECT(dev), "ram2", NPCM7XX_RAM2_SZ, &error_abort); diff --git a/hw/ssi/npcm7xx_fiu.c b/hw/ssi/npcm7xx_fiu.c new file mode 100644 index XXXXXXX..XXXXXXX --- /dev/null +++ b/hw/ssi/npcm7xx_fiu.c @@ -XXX,XX +XXX,XX @@ +/* + * Nuvoton NPCM7xx Flash Interface Unit (FIU) + * + * Copyright 2020 Google LLC + * + * This program is free software; you can redistribute it and/or modify it + * under the terms of the GNU General Public License as published by the + * Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program 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 General Public License + * for more details. + */ + +#include "qemu/osdep.h" + +#include "hw/irq.h" +#include "hw/qdev-properties.h" +#include "hw/ssi/npcm7xx_fiu.h" +#include "migration/vmstate.h" +#include "qapi/error.h" +#include "qemu/error-report.h" +#include "qemu/log.h" +#include "qemu/module.h" +#include "qemu/units.h" + +#include "trace.h" + +/* Up to 128 MiB of flash may be accessed directly as memory. */ +#define NPCM7XX_FIU_FLASH_WINDOW_SIZE (128 * MiB) + +/* Each module has 4 KiB of register space. Only a fraction of it is used. */ +#define NPCM7XX_FIU_CTRL_REGS_SIZE (4 * KiB) + +/* 32-bit FIU register indices. */ +enum NPCM7xxFIURegister { + NPCM7XX_FIU_DRD_CFG, + NPCM7XX_FIU_DWR_CFG, + NPCM7XX_FIU_UMA_CFG, + NPCM7XX_FIU_UMA_CTS, + NPCM7XX_FIU_UMA_CMD, + NPCM7XX_FIU_UMA_ADDR, + NPCM7XX_FIU_PRT_CFG, + NPCM7XX_FIU_UMA_DW0 = 0x0020 / sizeof(uint32_t), + NPCM7XX_FIU_UMA_DW1, + NPCM7XX_FIU_UMA_DW2, + NPCM7XX_FIU_UMA_DW3, + NPCM7XX_FIU_UMA_DR0, + NPCM7XX_FIU_UMA_DR1, + NPCM7XX_FIU_UMA_DR2, + NPCM7XX_FIU_UMA_DR3, + NPCM7XX_FIU_PRT_CMD0, + NPCM7XX_FIU_PRT_CMD1, + NPCM7XX_FIU_PRT_CMD2, + NPCM7XX_FIU_PRT_CMD3, + NPCM7XX_FIU_PRT_CMD4, + NPCM7XX_FIU_PRT_CMD5, + NPCM7XX_FIU_PRT_CMD6, + NPCM7XX_FIU_PRT_CMD7, + NPCM7XX_FIU_PRT_CMD8, + NPCM7XX_FIU_PRT_CMD9, + NPCM7XX_FIU_CFG = 0x78 / sizeof(uint32_t), + NPCM7XX_FIU_REGS_END, +}; + +/* FIU_{DRD,DWR,UMA,PTR}_CFG cannot be written when this bit is set. */ +#define NPCM7XX_FIU_CFG_LCK BIT(31) + +/* Direct Read configuration register fields. */ +#define FIU_DRD_CFG_ADDSIZ(rv) extract32(rv, 16, 2) +#define FIU_ADDSIZ_3BYTES 0 +#define FIU_ADDSIZ_4BYTES 1 +#define FIU_DRD_CFG_DBW(rv) extract32(rv, 12, 2) +#define FIU_DRD_CFG_ACCTYPE(rv) extract32(rv, 8, 2) +#define FIU_DRD_CFG_RDCMD(rv) extract32(rv, 0, 8) + +/* Direct Write configuration register fields. */ +#define FIU_DWR_CFG_ADDSIZ(rv) extract32(rv, 16, 2) +#define FIU_DWR_CFG_WRCMD(rv) extract32(rv, 0, 8) + +/* User-Mode Access register fields. */ + +/* Command Mode Lock and the bits protected by it. */ +#define FIU_UMA_CFG_CMMLCK BIT(30) +#define FIU_UMA_CFG_CMMLCK_MASK 0x00000403 + +#define FIU_UMA_CFG_RDATSIZ(rv) extract32(rv, 24, 5) +#define FIU_UMA_CFG_DBSIZ(rv) extract32(rv, 21, 3) +#define FIU_UMA_CFG_WDATSIZ(rv) extract32(rv, 16, 5) +#define FIU_UMA_CFG_ADDSIZ(rv) extract32(rv, 11, 3) +#define FIU_UMA_CFG_CMDSIZ(rv) extract32(rv, 10, 1) +#define FIU_UMA_CFG_DBPCK(rv) extract32(rv, 6, 2) + +#define FIU_UMA_CTS_RDYIE BIT(25) +#define FIU_UMA_CTS_RDYST BIT(24) +#define FIU_UMA_CTS_SW_CS BIT(16) +#define FIU_UMA_CTS_DEV_NUM(rv) extract32(rv, 8, 2) +#define FIU_UMA_CTS_EXEC_DONE BIT(0) + +/* + * Returns the index of flash in the fiu->flash array. This corresponds to the + * chip select ID of the flash. + */ +static int npcm7xx_fiu_cs_index(NPCM7xxFIUState *fiu, NPCM7xxFIUFlash *flash) +{ + int index = flash - fiu->flash; + + g_assert(index >= 0 && index < fiu->cs_count); + + return index; +} + +/* Assert the chip select specified in the UMA Control/Status Register. */ +static void npcm7xx_fiu_select(NPCM7xxFIUState *s, int cs_id) +{ + trace_npcm7xx_fiu_select(DEVICE(s)->canonical_path, cs_id); + + if (cs_id < s->cs_count) { + qemu_irq_lower(s->cs_lines[cs_id]); + } else { + qemu_log_mask(LOG_GUEST_ERROR, + "%s: UMA to CS%d; this module has only %d chip selects", + DEVICE(s)->canonical_path, cs_id, s->cs_count); + cs_id = -1; + } + + s->active_cs = cs_id; +} + +/* Deassert the currently active chip select. */ +static void npcm7xx_fiu_deselect(NPCM7xxFIUState *s) +{ + if (s->active_cs < 0) { + return; + } + + trace_npcm7xx_fiu_deselect(DEVICE(s)->canonical_path, s->active_cs); + + qemu_irq_raise(s->cs_lines[s->active_cs]); + s->active_cs = -1; +} + +/* Direct flash memory read handler. */ +static uint64_t npcm7xx_fiu_flash_read(void *opaque, hwaddr addr, + unsigned int size) +{ + NPCM7xxFIUFlash *f = opaque; + NPCM7xxFIUState *fiu = f->fiu; + uint64_t value = 0; + uint32_t drd_cfg; + int dummy_cycles; + int i; + + if (fiu->active_cs != -1) { + qemu_log_mask(LOG_GUEST_ERROR, + "%s: direct flash read with CS%d already active", + DEVICE(fiu)->canonical_path, fiu->active_cs); + } + + npcm7xx_fiu_select(fiu, npcm7xx_fiu_cs_index(fiu, f)); + + drd_cfg = fiu->regs[NPCM7XX_FIU_DRD_CFG]; + ssi_transfer(fiu->spi, FIU_DRD_CFG_RDCMD(drd_cfg)); + + switch (FIU_DRD_CFG_ADDSIZ(drd_cfg)) { + case FIU_ADDSIZ_4BYTES: + ssi_transfer(fiu->spi, extract32(addr, 24, 8)); + /* fall through */ + case FIU_ADDSIZ_3BYTES: + ssi_transfer(fiu->spi, extract32(addr, 16, 8)); + ssi_transfer(fiu->spi, extract32(addr, 8, 8)); + ssi_transfer(fiu->spi, extract32(addr, 0, 8)); + break; + + default: + qemu_log_mask(LOG_GUEST_ERROR, "%s: bad address size %d\n", + DEVICE(fiu)->canonical_path, FIU_DRD_CFG_ADDSIZ(drd_cfg)); + break; + } + + /* Flash chip model expects one transfer per dummy bit, not byte */ + dummy_cycles = + (FIU_DRD_CFG_DBW(drd_cfg) * 8) >> FIU_DRD_CFG_ACCTYPE(drd_cfg); + for (i = 0; i < dummy_cycles; i++) { + ssi_transfer(fiu->spi, 0); + } + + for (i = 0; i < size; i++) { + value = deposit64(value, 8 * i, 8, ssi_transfer(fiu->spi, 0)); + } + + trace_npcm7xx_fiu_flash_read(DEVICE(fiu)->canonical_path, fiu->active_cs, + addr, size, value); + + npcm7xx_fiu_deselect(fiu); + + return value; +} + +/* Direct flash memory write handler. */ +static void npcm7xx_fiu_flash_write(void *opaque, hwaddr addr, uint64_t v, + unsigned int size) +{ + NPCM7xxFIUFlash *f = opaque; + NPCM7xxFIUState *fiu = f->fiu; + uint32_t dwr_cfg; + int cs_id; + int i; + + if (fiu->active_cs != -1) { + qemu_log_mask(LOG_GUEST_ERROR, + "%s: direct flash write with CS%d already active", + DEVICE(fiu)->canonical_path, fiu->active_cs); + } + + cs_id = npcm7xx_fiu_cs_index(fiu, f); + trace_npcm7xx_fiu_flash_write(DEVICE(fiu)->canonical_path, cs_id, addr, + size, v); + npcm7xx_fiu_select(fiu, cs_id); + + dwr_cfg = fiu->regs[NPCM7XX_FIU_DWR_CFG]; + ssi_transfer(fiu->spi, FIU_DWR_CFG_WRCMD(dwr_cfg)); + + switch (FIU_DWR_CFG_ADDSIZ(dwr_cfg)) { + case FIU_ADDSIZ_4BYTES: + ssi_transfer(fiu->spi, extract32(addr, 24, 8)); + /* fall through */ + case FIU_ADDSIZ_3BYTES: + ssi_transfer(fiu->spi, extract32(addr, 16, 8)); + ssi_transfer(fiu->spi, extract32(addr, 8, 8)); + ssi_transfer(fiu->spi, extract32(addr, 0, 8)); + break; + + default: + qemu_log_mask(LOG_GUEST_ERROR, "%s: bad address size %d\n", + DEVICE(fiu)->canonical_path, FIU_DWR_CFG_ADDSIZ(dwr_cfg)); + break; + } + + for (i = 0; i < size; i++) { + ssi_transfer(fiu->spi, extract64(v, i * 8, 8)); + } + + npcm7xx_fiu_deselect(fiu); +} + +static const MemoryRegionOps npcm7xx_fiu_flash_ops = { + .read = npcm7xx_fiu_flash_read, + .write = npcm7xx_fiu_flash_write, + .endianness = DEVICE_LITTLE_ENDIAN, + .valid = { + .min_access_size = 1, + .max_access_size = 8, + .unaligned = true, + }, +}; + +/* Control register read handler. */ +static uint64_t npcm7xx_fiu_ctrl_read(void *opaque, hwaddr addr, + unsigned int size) +{ + hwaddr reg = addr / sizeof(uint32_t); + NPCM7xxFIUState *s = opaque; + uint32_t value; + + if (reg < NPCM7XX_FIU_NR_REGS) { + value = s->regs[reg]; + } else { + qemu_log_mask(LOG_GUEST_ERROR, + "%s: read from invalid offset 0x%" PRIx64 "\n", + DEVICE(s)->canonical_path, addr); + value = 0; + } + + trace_npcm7xx_fiu_ctrl_read(DEVICE(s)->canonical_path, addr, value); + + return value; +} + +/* Send the specified number of address bytes from the UMA address register. */ +static void send_address(SSIBus *spi, unsigned int addsiz, uint32_t addr) +{ + switch (addsiz) { + case 4: + ssi_transfer(spi, extract32(addr, 24, 8)); + /* fall through */ + case 3: + ssi_transfer(spi, extract32(addr, 16, 8)); + /* fall through */ + case 2: + ssi_transfer(spi, extract32(addr, 8, 8)); + /* fall through */ + case 1: + ssi_transfer(spi, extract32(addr, 0, 8)); + /* fall through */ + case 0: + break; + } +} + +/* Send the number of dummy bits specified in the UMA config register. */ +static void send_dummy_bits(SSIBus *spi, uint32_t uma_cfg, uint32_t uma_cmd) +{ + unsigned int bits_per_clock = 1U << FIU_UMA_CFG_DBPCK(uma_cfg); + unsigned int i; + + for (i = 0; i < FIU_UMA_CFG_DBSIZ(uma_cfg); i++) { + /* Use bytes 0 and 1 first, then keep repeating byte 2 */ + unsigned int field = (i < 2) ? ((i + 1) * 8) : 24; + unsigned int j; + + for (j = 0; j < 8; j += bits_per_clock) { + ssi_transfer(spi, extract32(uma_cmd, field + j, bits_per_clock)); + } + } +} + +/* Perform a User-Mode Access transaction. */ +static void npcm7xx_fiu_uma_transaction(NPCM7xxFIUState *s) +{ + uint32_t uma_cts = s->regs[NPCM7XX_FIU_UMA_CTS]; + uint32_t uma_cfg; + unsigned int i; + + /* SW_CS means the CS is already forced low, so don't touch it. */ + if (uma_cts & FIU_UMA_CTS_SW_CS) { + int cs_id = FIU_UMA_CTS_DEV_NUM(s->regs[NPCM7XX_FIU_UMA_CTS]); + npcm7xx_fiu_select(s, cs_id); + } + + /* Send command, if present. */ + uma_cfg = s->regs[NPCM7XX_FIU_UMA_CFG]; + if (FIU_UMA_CFG_CMDSIZ(uma_cfg) > 0) { + ssi_transfer(s->spi, extract32(s->regs[NPCM7XX_FIU_UMA_CMD], 0, 8)); + } + + /* Send address, if present. */ + send_address(s->spi, FIU_UMA_CFG_ADDSIZ(uma_cfg), + s->regs[NPCM7XX_FIU_UMA_ADDR]); + + /* Write data, if present. */ + for (i = 0; i < FIU_UMA_CFG_WDATSIZ(uma_cfg); i++) { + unsigned int reg = + (i < 16) ? (NPCM7XX_FIU_UMA_DW0 + i / 4) : NPCM7XX_FIU_UMA_DW3; + unsigned int field = (i % 4) * 8; + + ssi_transfer(s->spi, extract32(s->regs[reg], field, 8)); + } + + /* Send dummy bits, if present. */ + send_dummy_bits(s->spi, uma_cfg, s->regs[NPCM7XX_FIU_UMA_CMD]); + + /* Read data, if present. */ + for (i = 0; i < FIU_UMA_CFG_RDATSIZ(uma_cfg); i++) { + unsigned int reg = NPCM7XX_FIU_UMA_DR0 + i / 4; + unsigned int field = (i % 4) * 8; + uint8_t c; + + c = ssi_transfer(s->spi, 0); + if (reg <= NPCM7XX_FIU_UMA_DR3) { + s->regs[reg] = deposit32(s->regs[reg], field, 8, c); + } + } + + /* Again, don't touch CS if the user is forcing it low. */ + if (uma_cts & FIU_UMA_CTS_SW_CS) { + npcm7xx_fiu_deselect(s); + } + + /* RDYST means a command has completed since it was cleared. */ + s->regs[NPCM7XX_FIU_UMA_CTS] |= FIU_UMA_CTS_RDYST; + /* EXEC_DONE means Execute Command / Not Done, so clear it here. */ + s->regs[NPCM7XX_FIU_UMA_CTS] &= ~FIU_UMA_CTS_EXEC_DONE; +} + +/* Control register write handler. */ +static void npcm7xx_fiu_ctrl_write(void *opaque, hwaddr addr, uint64_t v, + unsigned int size) +{ + hwaddr reg = addr / sizeof(uint32_t); + NPCM7xxFIUState *s = opaque; + uint32_t value = v; + + trace_npcm7xx_fiu_ctrl_write(DEVICE(s)->canonical_path, addr, value); + + switch (reg) { + case NPCM7XX_FIU_UMA_CFG: + if (s->regs[reg] & FIU_UMA_CFG_CMMLCK) { + value &= ~FIU_UMA_CFG_CMMLCK_MASK; + value |= (s->regs[reg] & FIU_UMA_CFG_CMMLCK_MASK); + } + /* fall through */ + case NPCM7XX_FIU_DRD_CFG: + case NPCM7XX_FIU_DWR_CFG: + if (s->regs[reg] & NPCM7XX_FIU_CFG_LCK) { + qemu_log_mask(LOG_GUEST_ERROR, + "%s: write to locked register @ 0x%" PRIx64 "\n", + DEVICE(s)->canonical_path, addr); + return; + } + s->regs[reg] = value; + break; + + case NPCM7XX_FIU_UMA_CTS: + if (value & FIU_UMA_CTS_RDYST) { + value &= ~FIU_UMA_CTS_RDYST; + } else { + value |= s->regs[reg] & FIU_UMA_CTS_RDYST; + } + if ((s->regs[reg] ^ value) & FIU_UMA_CTS_SW_CS) { + if (value & FIU_UMA_CTS_SW_CS) { + /* + * Don't drop CS if there's a transfer in progress, or we're + * about to start one. + */ + if (!((value | s->regs[reg]) & FIU_UMA_CTS_EXEC_DONE)) { + npcm7xx_fiu_deselect(s); + } + } else { + int cs_id = FIU_UMA_CTS_DEV_NUM(s->regs[NPCM7XX_FIU_UMA_CTS]); + npcm7xx_fiu_select(s, cs_id); + } + } + s->regs[reg] = value | (s->regs[reg] & FIU_UMA_CTS_EXEC_DONE); + if (value & FIU_UMA_CTS_EXEC_DONE) { + npcm7xx_fiu_uma_transaction(s); + } + break; + + case NPCM7XX_FIU_UMA_DR0 ... NPCM7XX_FIU_UMA_DR3: + qemu_log_mask(LOG_GUEST_ERROR, + "%s: write to read-only register @ 0x%" PRIx64 "\n", + DEVICE(s)->canonical_path, addr); + return; + + case NPCM7XX_FIU_PRT_CFG: + case NPCM7XX_FIU_PRT_CMD0 ... NPCM7XX_FIU_PRT_CMD9: + qemu_log_mask(LOG_UNIMP, "%s: PRT is not implemented\n", __func__); + break; + + case NPCM7XX_FIU_UMA_CMD: + case NPCM7XX_FIU_UMA_ADDR: + case NPCM7XX_FIU_UMA_DW0 ... NPCM7XX_FIU_UMA_DW3: + case NPCM7XX_FIU_CFG: + s->regs[reg] = value; + break; + + default: + qemu_log_mask(LOG_GUEST_ERROR, + "%s: write to invalid offset 0x%" PRIx64 "\n", + DEVICE(s)->canonical_path, addr); + return; + } +} + +static const MemoryRegionOps npcm7xx_fiu_ctrl_ops = { + .read = npcm7xx_fiu_ctrl_read, + .write = npcm7xx_fiu_ctrl_write, + .endianness = DEVICE_LITTLE_ENDIAN, + .valid = { + .min_access_size = 4, + .max_access_size = 4, + .unaligned = false, + }, +}; + +static void npcm7xx_fiu_enter_reset(Object *obj, ResetType type) +{ + NPCM7xxFIUState *s = NPCM7XX_FIU(obj); + + trace_npcm7xx_fiu_enter_reset(DEVICE(obj)->canonical_path, type); + + memset(s->regs, 0, sizeof(s->regs)); + + s->regs[NPCM7XX_FIU_DRD_CFG] = 0x0300100b; + s->regs[NPCM7XX_FIU_DWR_CFG] = 0x03000002; + s->regs[NPCM7XX_FIU_UMA_CFG] = 0x00000400; + s->regs[NPCM7XX_FIU_UMA_CTS] = 0x00010000; + s->regs[NPCM7XX_FIU_UMA_CMD] = 0x0000000b; + s->regs[NPCM7XX_FIU_PRT_CFG] = 0x00000400; + s->regs[NPCM7XX_FIU_CFG] = 0x0000000b; +} + +static void npcm7xx_fiu_hold_reset(Object *obj) +{ + NPCM7xxFIUState *s = NPCM7XX_FIU(obj); + int i; + + trace_npcm7xx_fiu_hold_reset(DEVICE(obj)->canonical_path); + + for (i = 0; i < s->cs_count; i++) { + qemu_irq_raise(s->cs_lines[i]); + } +} + +static void npcm7xx_fiu_realize(DeviceState *dev, Error **errp) +{ + NPCM7xxFIUState *s = NPCM7XX_FIU(dev); + SysBusDevice *sbd = &s->parent; + int i; + + if (s->cs_count <= 0) { + error_setg(errp, "%s: %d chip selects specified, need at least one", + dev->canonical_path, s->cs_count); + return; + } + + s->spi = ssi_create_bus(dev, "spi"); + s->cs_lines = g_new0(qemu_irq, s->cs_count); + qdev_init_gpio_out_named(DEVICE(s), s->cs_lines, "cs", s->cs_count); + s->flash = g_new0(NPCM7xxFIUFlash, s->cs_count); + + /* + * Register the control registers region first. It may be followed by one + * or more direct flash access regions. + */ + memory_region_init_io(&s->mmio, OBJECT(s), &npcm7xx_fiu_ctrl_ops, s, "ctrl", + NPCM7XX_FIU_CTRL_REGS_SIZE); + sysbus_init_mmio(sbd, &s->mmio); + + for (i = 0; i < s->cs_count; i++) { + NPCM7xxFIUFlash *flash = &s->flash[i]; + flash->fiu = s; + memory_region_init_io(&flash->direct_access, OBJECT(s), + &npcm7xx_fiu_flash_ops, &s->flash[i], "flash", + NPCM7XX_FIU_FLASH_WINDOW_SIZE); + sysbus_init_mmio(sbd, &flash->direct_access); + } +} + +static const VMStateDescription vmstate_npcm7xx_fiu = { + .name = "npcm7xx-fiu", + .version_id = 0, + .minimum_version_id = 0, + .fields = (VMStateField[]) { + VMSTATE_INT32(active_cs, NPCM7xxFIUState), + VMSTATE_UINT32_ARRAY(regs, NPCM7xxFIUState, NPCM7XX_FIU_NR_REGS), + VMSTATE_END_OF_LIST(), + }, +}; + +static Property npcm7xx_fiu_properties[] = { + DEFINE_PROP_INT32("cs-count", NPCM7xxFIUState, cs_count, 0), + DEFINE_PROP_END_OF_LIST(), +}; + +static void npcm7xx_fiu_class_init(ObjectClass *klass, void *data) +{ + ResettableClass *rc = RESETTABLE_CLASS(klass); + DeviceClass *dc = DEVICE_CLASS(klass); + + QEMU_BUILD_BUG_ON(NPCM7XX_FIU_REGS_END > NPCM7XX_FIU_NR_REGS); + + dc->desc = "NPCM7xx Flash Interface Unit"; + dc->realize = npcm7xx_fiu_realize; + dc->vmsd = &vmstate_npcm7xx_fiu; + rc->phases.enter = npcm7xx_fiu_enter_reset; + rc->phases.hold = npcm7xx_fiu_hold_reset; + device_class_set_props(dc, npcm7xx_fiu_properties); +} + +static const TypeInfo npcm7xx_fiu_types[] = { + { + .name = TYPE_NPCM7XX_FIU, + .parent = TYPE_SYS_BUS_DEVICE, + .instance_size = sizeof(NPCM7xxFIUState), + .class_init = npcm7xx_fiu_class_init, + }, +}; +DEFINE_TYPES(npcm7xx_fiu_types); diff --git a/hw/arm/Kconfig b/hw/arm/Kconfig index XXXXXXX..XXXXXXX 100644 --- a/hw/arm/Kconfig +++ b/hw/arm/Kconfig @@ -XXX,XX +XXX,XX @@ config NPCM7XX select ARM_GIC select PL310 # cache controller select SERIAL + select SSI select UNIMP config FSL_IMX25 diff --git a/hw/ssi/meson.build b/hw/ssi/meson.build index XXXXXXX..XXXXXXX 100644 --- a/hw/ssi/meson.build +++ b/hw/ssi/meson.build @@ -XXX,XX +XXX,XX @@ softmmu_ss.add(when: 'CONFIG_ASPEED_SOC', if_true: files('aspeed_smc.c')) softmmu_ss.add(when: 'CONFIG_MSF2', if_true: files('mss-spi.c')) +softmmu_ss.add(when: 'CONFIG_NPCM7XX', if_true: files('npcm7xx_fiu.c')) softmmu_ss.add(when: 'CONFIG_PL022', if_true: files('pl022.c')) softmmu_ss.add(when: 'CONFIG_SSI', if_true: files('ssi.c')) softmmu_ss.add(when: 'CONFIG_STM32F2XX_SPI', if_true: files('stm32f2xx_spi.c')) diff --git a/hw/ssi/trace-events b/hw/ssi/trace-events index XXXXXXX..XXXXXXX 100644 --- a/hw/ssi/trace-events +++ b/hw/ssi/trace-events @@ -XXX,XX +XXX,XX @@ aspeed_smc_dma_checksum(uint32_t addr, uint32_t data) "0x%08x: 0x%08x" aspeed_smc_dma_rw(const char *dir, uint32_t flash_addr, uint32_t dram_addr, uint32_t size) "%s flash:@0x%08x dram:@0x%08x size:0x%08x" aspeed_smc_write(uint64_t addr, uint32_t size, uint64_t data) "@0x%" PRIx64 " size %u: 0x%" PRIx64 aspeed_smc_flash_select(int cs, const char *prefix) "CS%d %sselect" + +# npcm7xx_fiu.c + +npcm7xx_fiu_enter_reset(const char *id, int reset_type) "%s reset type: %d" +npcm7xx_fiu_hold_reset(const char *id) "%s" +npcm7xx_fiu_select(const char *id, int cs) "%s select CS%d" +npcm7xx_fiu_deselect(const char *id, int cs) "%s deselect CS%d" +npcm7xx_fiu_ctrl_read(const char *id, uint64_t addr, uint32_t data) "%s offset: 0x%04" PRIx64 " value: 0x%08" PRIx32 +npcm7xx_fiu_ctrl_write(const char *id, uint64_t addr, uint32_t data) "%s offset: 0x%04" PRIx64 " value: 0x%08" PRIx32 +npcm7xx_fiu_flash_read(const char *id, int cs, uint64_t addr, unsigned int size, uint64_t value) "%s[%d] offset: 0x%08" PRIx64 " size: %u value: 0x%" PRIx64 +npcm7xx_fiu_flash_write(const char *id, int cs, uint64_t addr, unsigned int size, uint64_t value) "%s[%d] offset: 0x%08" PRIx64 " size: %u value: 0x%" PRIx64 -- 2.20.1 From: Havard Skinnemoen <hskinnemoen@google.com> This allows these NPCM7xx-based boards to boot from a flash image, e.g. one built with OpenBMC. For example like this: IMAGE=${OPENBMC}/build/tmp/deploy/images/gsj/image-bmc qemu-system-arm -machine quanta-gsj -nographic \ -drive file=${IMAGE},if=mtd,bus=0,unit=0,format=raw,snapshot=on Reviewed-by: Tyrone Ting <kfting@nuvoton.com> Reviewed-by: Cédric Le Goater <clg@kaod.org> Tested-by: Cédric Le Goater <clg@kaod.org> Tested-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Signed-off-by: Havard Skinnemoen <hskinnemoen@google.com> Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Message-id: 20200911052101.2602693-12-hskinnemoen@google.com Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- hw/arm/npcm7xx_boards.c | 20 ++++++++++++++++++++ 1 file changed, 20 insertions(+) diff --git a/hw/arm/npcm7xx_boards.c b/hw/arm/npcm7xx_boards.c index XXXXXXX..XXXXXXX 100644 --- a/hw/arm/npcm7xx_boards.c +++ b/hw/arm/npcm7xx_boards.c @@ -XXX,XX +XXX,XX @@ #include "hw/arm/npcm7xx.h" #include "hw/core/cpu.h" #include "hw/loader.h" +#include "hw/qdev-properties.h" #include "qapi/error.h" #include "qemu-common.h" #include "qemu/units.h" @@ -XXX,XX +XXX,XX @@ static void npcm7xx_load_bootrom(MachineState *machine, NPCM7xxState *soc) } } +static void npcm7xx_connect_flash(NPCM7xxFIUState *fiu, int cs_no, + const char *flash_type, DriveInfo *dinfo) +{ + DeviceState *flash; + qemu_irq flash_cs; + + flash = qdev_new(flash_type); + if (dinfo) { + qdev_prop_set_drive(flash, "drive", blk_by_legacy_dinfo(dinfo)); + } + qdev_realize_and_unref(flash, BUS(fiu->spi), &error_fatal); + + flash_cs = qdev_get_gpio_in_named(flash, SSI_GPIO_CS, 0); + qdev_connect_gpio_out_named(DEVICE(fiu), "cs", cs_no, flash_cs); +} + static void npcm7xx_connect_dram(NPCM7xxState *soc, MemoryRegion *dram) { memory_region_add_subregion(get_system_memory(), NPCM7XX_DRAM_BA, dram); @@ -XXX,XX +XXX,XX @@ static void npcm750_evb_init(MachineState *machine) qdev_realize(DEVICE(soc), NULL, &error_fatal); npcm7xx_load_bootrom(machine, soc); + npcm7xx_connect_flash(&soc->fiu[0], 0, "w25q256", drive_get(IF_MTD, 0, 0)); npcm7xx_load_kernel(machine, soc); } @@ -XXX,XX +XXX,XX @@ static void quanta_gsj_init(MachineState *machine) qdev_realize(DEVICE(soc), NULL, &error_fatal); npcm7xx_load_bootrom(machine, soc); + npcm7xx_connect_flash(&soc->fiu[0], 0, "mx25l25635e", + drive_get(IF_MTD, 0, 0)); npcm7xx_load_kernel(machine, soc); } -- 2.20.1 From: Havard Skinnemoen <hskinnemoen@google.com> When booting directly into a kernel, bypassing the boot loader, the CPU and UART clocks are not set up correctly. This makes the system appear very slow, and causes the initrd boot test to fail when optimization is off. The UART clock must run at 24 MHz. The default 25 MHz reference clock cannot achieve this, so switch to PLL2/2 @ 480 MHz, which works perfectly with the default /20 divider. The CPU clock should run at 800 MHz, so switch it to PLL1/2. PLL1 runs at 800 MHz by default, so we need to double the feedback divider as well to make it run at 1600 MHz (so PLL1/2 runs at 800 MHz). We don't bother checking for PLL lock because we know our emulated PLLs lock instantly. Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Tested-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Signed-off-by: Havard Skinnemoen <hskinnemoen@google.com> Message-id: 20200911052101.2602693-13-hskinnemoen@google.com Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- include/hw/arm/npcm7xx.h | 1 + hw/arm/npcm7xx.c | 32 ++++++++++++++++++++++++++++++++ 2 files changed, 33 insertions(+) diff --git a/include/hw/arm/npcm7xx.h b/include/hw/arm/npcm7xx.h index XXXXXXX..XXXXXXX 100644 --- a/include/hw/arm/npcm7xx.h +++ b/include/hw/arm/npcm7xx.h @@ -XXX,XX +XXX,XX @@ #define NPCM7XX_SMP_LOADER_START (0xffff0000) /* Boot ROM */ #define NPCM7XX_SMP_BOOTREG_ADDR (0xf080013c) /* GCR.SCRPAD */ #define NPCM7XX_GIC_CPU_IF_ADDR (0xf03fe100) /* GIC within A9 */ +#define NPCM7XX_BOARD_SETUP_ADDR (0xffff1000) /* Boot ROM */ typedef struct NPCM7xxMachine { MachineState parent; diff --git a/hw/arm/npcm7xx.c b/hw/arm/npcm7xx.c index XXXXXXX..XXXXXXX 100644 --- a/hw/arm/npcm7xx.c +++ b/hw/arm/npcm7xx.c @@ -XXX,XX +XXX,XX @@ #define NPCM7XX_ROM_BA (0xffff0000) #define NPCM7XX_ROM_SZ (64 * KiB) +/* Clock configuration values to be fixed up when bypassing bootloader */ + +/* Run PLL1 at 1600 MHz */ +#define NPCM7XX_PLLCON1_FIXUP_VAL (0x00402101) +/* Run the CPU from PLL1 and UART from PLL2 */ +#define NPCM7XX_CLKSEL_FIXUP_VAL (0x004aaba9) + /* * Interrupt lines going into the GIC. This does not include internal Cortex-A9 * interrupts. @@ -XXX,XX +XXX,XX @@ static const struct { }, }; +static void npcm7xx_write_board_setup(ARMCPU *cpu, + const struct arm_boot_info *info) +{ + uint32_t board_setup[] = { + 0xe59f0010, /* ldr r0, clk_base_addr */ + 0xe59f1010, /* ldr r1, pllcon1_value */ + 0xe5801010, /* str r1, [r0, #16] */ + 0xe59f100c, /* ldr r1, clksel_value */ + 0xe5801004, /* str r1, [r0, #4] */ + 0xe12fff1e, /* bx lr */ + NPCM7XX_CLK_BA, + NPCM7XX_PLLCON1_FIXUP_VAL, + NPCM7XX_CLKSEL_FIXUP_VAL, + }; + int i; + + for (i = 0; i < ARRAY_SIZE(board_setup); i++) { + board_setup[i] = tswap32(board_setup[i]); + } + rom_add_blob_fixed("board-setup", board_setup, sizeof(board_setup), + info->board_setup_addr); +} + static void npcm7xx_write_secondary_boot(ARMCPU *cpu, const struct arm_boot_info *info) { @@ -XXX,XX +XXX,XX @@ static struct arm_boot_info npcm7xx_binfo = { .gic_cpu_if_addr = NPCM7XX_GIC_CPU_IF_ADDR, .write_secondary_boot = npcm7xx_write_secondary_boot, .board_id = -1, + .board_setup_addr = NPCM7XX_BOARD_SETUP_ADDR, + .write_board_setup = npcm7xx_write_board_setup, }; void npcm7xx_load_kernel(MachineState *machine, NPCM7xxState *soc) -- 2.20.1 From: Havard Skinnemoen <hskinnemoen@google.com> Reviewed-by: Cédric Le Goater <clg@kaod.org> Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Signed-off-by: Havard Skinnemoen <hskinnemoen@google.com> Message-id: 20200911052101.2602693-14-hskinnemoen@google.com Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- docs/system/arm/nuvoton.rst | 92 +++++++++++++++++++++++++++++++++++++ docs/system/target-arm.rst | 1 + 2 files changed, 93 insertions(+) create mode 100644 docs/system/arm/nuvoton.rst diff --git a/docs/system/arm/nuvoton.rst b/docs/system/arm/nuvoton.rst new file mode 100644 index XXXXXXX..XXXXXXX --- /dev/null +++ b/docs/system/arm/nuvoton.rst @@ -XXX,XX +XXX,XX @@ +Nuvoton iBMC boards (``npcm750-evb``, ``quanta-gsj``) +===================================================== + +The `Nuvoton iBMC`_ chips (NPCM7xx) are a family of ARM-based SoCs that are +designed to be used as Baseboard Management Controllers (BMCs) in various +servers. They all feature one or two ARM Cortex A9 CPU cores, as well as an +assortment of peripherals targeted for either Enterprise or Data Center / +Hyperscale applications. The former is a superset of the latter, so NPCM750 has +all the peripherals of NPCM730 and more. + +.. _Nuvoton iBMC: https://www.nuvoton.com/products/cloud-computing/ibmc/ + +The NPCM750 SoC has two Cortex A9 cores and is targeted for the Enterprise +segment. The following machines are based on this chip : + +- ``npcm750-evb`` Nuvoton NPCM750 Evaluation board + +The NPCM730 SoC has two Cortex A9 cores and is targeted for Data Center and +Hyperscale applications. The following machines are based on this chip : + +- ``quanta-gsj`` Quanta GSJ server BMC + +There are also two more SoCs, NPCM710 and NPCM705, which are single-core +variants of NPCM750 and NPCM730, respectively. These are currently not +supported by QEMU. + +Supported devices +----------------- + + * SMP (Dual Core Cortex-A9) + * Cortex-A9MPCore built-in peripherals: SCU, GIC, Global Timer, Private Timer + and Watchdog. + * SRAM, ROM and DRAM mappings + * System Global Control Registers (GCR) + * Clock and reset controller (CLK) + * Timer controller (TIM) + * Serial ports (16550-based) + * DDR4 memory controller (dummy interface indicating memory training is done) + * OTP controllers (no protection features) + * Flash Interface Unit (FIU; no protection features) + +Missing devices +--------------- + + * GPIO controller + * LPC/eSPI host-to-BMC interface, including + + * Keyboard and mouse controller interface (KBCI) + * Keyboard Controller Style (KCS) channels + * BIOS POST code FIFO + * System Wake-up Control (SWC) + * Shared memory (SHM) + * eSPI slave interface + + * Ethernet controllers (GMAC and EMC) + * USB host (USBH) + * USB device (USBD) + * SMBus controller (SMBF) + * Peripheral SPI controller (PSPI) + * Analog to Digital Converter (ADC) + * SD/MMC host + * Random Number Generator (RNG) + * PECI interface + * Pulse Width Modulation (PWM) + * Tachometer + * PCI and PCIe root complex and bridges + * VDM and MCTP support + * Serial I/O expansion + * LPC/eSPI host + * Coprocessor + * Graphics + * Video capture + * Encoding compression engine + * Security features + +Boot options +------------ + +The Nuvoton machines can boot from an OpenBMC firmware image, or directly into +a kernel using the ``-kernel`` option. OpenBMC images for `quanta-gsj` and +possibly others can be downloaded from the OpenPOWER jenkins : + + https://openpower.xyz/ + +The firmware image should be attached as an MTD drive. Example : + +.. code-block:: bash + + $ qemu-system-arm -machine quanta-gsj -nographic \ + -drive file=image-bmc,if=mtd,bus=0,unit=0,format=raw + +The default root password for test images is usually ``0penBmc``. diff --git a/docs/system/target-arm.rst b/docs/system/target-arm.rst index XXXXXXX..XXXXXXX 100644 --- a/docs/system/target-arm.rst +++ b/docs/system/target-arm.rst @@ -XXX,XX +XXX,XX @@ undocumented; you can get a complete list by running arm/musicpal arm/gumstix arm/nseries + arm/nuvoton arm/orangepi arm/palm arm/xscale -- 2.20.1 From: Havard Skinnemoen <hskinnemoen@google.com> This adds two acceptance tests for the quanta-gsj machine. One test downloads a lightly patched openbmc flash image from github and verifies that it boots all the way to the login prompt. The other test downloads a kernel, initrd and dtb built from the same openbmc source and verifies that the kernel detects all CPUs and boots to the point where it can't find the root filesystem (because we have no flash image in this case). Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Tested-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Signed-off-by: Havard Skinnemoen <hskinnemoen@google.com> Message-id: 20200911052101.2602693-15-hskinnemoen@google.com Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- tests/acceptance/boot_linux_console.py | 83 ++++++++++++++++++++++++++ 1 file changed, 83 insertions(+) diff --git a/tests/acceptance/boot_linux_console.py b/tests/acceptance/boot_linux_console.py index XXXXXXX..XXXXXXX 100644 --- a/tests/acceptance/boot_linux_console.py +++ b/tests/acceptance/boot_linux_console.py @@ -XXX,XX +XXX,XX @@ class BootLinuxConsole(LinuxKernelTest): 'sda') # cubieboard's reboot is not functioning; omit reboot test. + def test_arm_quanta_gsj(self): + """ + :avocado: tags=arch:arm + :avocado: tags=machine:quanta-gsj + """ + # 25 MiB compressed, 32 MiB uncompressed. + image_url = ( + 'https://github.com/hskinnemoen/openbmc/releases/download/' + '20200711-gsj-qemu-0/obmc-phosphor-image-gsj.static.mtd.gz') + image_hash = '14895e634923345cb5c8776037ff7876df96f6b1' + image_path_gz = self.fetch_asset(image_url, asset_hash=image_hash) + image_name = 'obmc.mtd' + image_path = os.path.join(self.workdir, image_name) + archive.gzip_uncompress(image_path_gz, image_path) + + self.vm.set_console() + drive_args = 'file=' + image_path + ',if=mtd,bus=0,unit=0' + self.vm.add_args('-drive', drive_args) + self.vm.launch() + + # Disable drivers and services that stall for a long time during boot, + # to avoid running past the 90-second timeout. These may be removed + # as the corresponding device support is added. + kernel_command_line = self.KERNEL_COMMON_COMMAND_LINE + ( + 'console=${console} ' + 'mem=${mem} ' + 'initcall_blacklist=npcm_i2c_bus_driver_init ' + 'systemd.mask=systemd-random-seed.service ' + 'systemd.mask=dropbearkey.service ' + ) + + self.wait_for_console_pattern('> BootBlock by Nuvoton') + self.wait_for_console_pattern('>Device: Poleg BMC NPCM730') + self.wait_for_console_pattern('>Skip DDR init.') + self.wait_for_console_pattern('U-Boot ') + interrupt_interactive_console_until_pattern( + self, 'Hit any key to stop autoboot:', 'U-Boot>') + exec_command_and_wait_for_pattern( + self, "setenv bootargs ${bootargs} " + kernel_command_line, + 'U-Boot>') + exec_command_and_wait_for_pattern( + self, 'run romboot', 'Booting Kernel from flash') + self.wait_for_console_pattern('Booting Linux on physical CPU 0x0') + self.wait_for_console_pattern('CPU1: thread -1, cpu 1, socket 0') + self.wait_for_console_pattern('OpenBMC Project Reference Distro') + self.wait_for_console_pattern('gsj login:') + + def test_arm_quanta_gsj_initrd(self): + """ + :avocado: tags=arch:arm + :avocado: tags=machine:quanta-gsj + """ + initrd_url = ( + 'https://github.com/hskinnemoen/openbmc/releases/download/' + '20200711-gsj-qemu-0/obmc-phosphor-initramfs-gsj.cpio.xz') + initrd_hash = '98fefe5d7e56727b1eb17d5c00311b1b5c945300' + initrd_path = self.fetch_asset(initrd_url, asset_hash=initrd_hash) + kernel_url = ( + 'https://github.com/hskinnemoen/openbmc/releases/download/' + '20200711-gsj-qemu-0/uImage-gsj.bin') + kernel_hash = 'fa67b2f141d56d39b3c54305c0e8a899c99eb2c7' + kernel_path = self.fetch_asset(kernel_url, asset_hash=kernel_hash) + dtb_url = ( + 'https://github.com/hskinnemoen/openbmc/releases/download/' + '20200711-gsj-qemu-0/nuvoton-npcm730-gsj.dtb') + dtb_hash = '18315f7006d7b688d8312d5c727eecd819aa36a4' + dtb_path = self.fetch_asset(dtb_url, asset_hash=dtb_hash) + + self.vm.set_console() + kernel_command_line = (self.KERNEL_COMMON_COMMAND_LINE + + 'console=ttyS0,115200n8 ' + 'earlycon=uart8250,mmio32,0xf0001000') + self.vm.add_args('-kernel', kernel_path, + '-initrd', initrd_path, + '-dtb', dtb_path, + '-append', kernel_command_line) + self.vm.launch() + + self.wait_for_console_pattern('Booting Linux on physical CPU 0x0') + self.wait_for_console_pattern('CPU1: thread -1, cpu 1, socket 0') + self.wait_for_console_pattern( + 'Give root password for system maintenance') + def test_arm_orangepi(self): """ :avocado: tags=arch:arm -- 2.20.1 From: Vikram Garhwal <fnu.vikram@xilinx.com> The Xilinx ZynqMP CAN controller is developed based on SocketCAN, QEMU CAN bus implementation. Bus connection and socketCAN connection for each CAN module can be set through command lines. Example for using single CAN: -object can-bus,id=canbus0 \ -machine xlnx-zcu102.canbus0=canbus0 \ -object can-host-socketcan,id=socketcan0,if=vcan0,canbus=canbus0 Example for connecting both CAN to same virtual CAN on host machine: -object can-bus,id=canbus0 -object can-bus,id=canbus1 \ -machine xlnx-zcu102.canbus0=canbus0 \ -machine xlnx-zcu102.canbus1=canbus1 \ -object can-host-socketcan,id=socketcan0,if=vcan0,canbus=canbus0 \ -object can-host-socketcan,id=socketcan1,if=vcan0,canbus=canbus1 To create virtual CAN on the host machine, please check the QEMU CAN docs: https://github.com/qemu/qemu/blob/master/docs/can.txt Signed-off-by: Vikram Garhwal <fnu.vikram@xilinx.com> Reviewed-by: Francisco Iglesias <francisco.iglesias@xilinx.com> Reviewed-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com> Message-id: 1597278668-339715-2-git-send-email-fnu.vikram@xilinx.com [PMM: updated to meson build system] Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- include/hw/net/xlnx-zynqmp-can.h | 78 ++ hw/net/can/xlnx-zynqmp-can.c | 1165 ++++++++++++++++++++++++++++++ hw/net/can/meson.build | 1 + 3 files changed, 1244 insertions(+) create mode 100644 include/hw/net/xlnx-zynqmp-can.h create mode 100644 hw/net/can/xlnx-zynqmp-can.c diff --git a/include/hw/net/xlnx-zynqmp-can.h b/include/hw/net/xlnx-zynqmp-can.h new file mode 100644 index XXXXXXX..XXXXXXX --- /dev/null +++ b/include/hw/net/xlnx-zynqmp-can.h @@ -XXX,XX +XXX,XX @@ +/* + * QEMU model of the Xilinx ZynqMP CAN controller. + * + * Copyright (c) 2020 Xilinx Inc. + * + * Written-by: Vikram Garhwal<fnu.vikram@xilinx.com> + * + * Based on QEMU CAN Device emulation implemented by Jin Yang, Deniz Eren and + * Pavel Pisa. + * + * Permission is hereby granted, free of charge, to any person obtaining a copy + * of this software and associated documentation files (the "Software"), to deal + * in the Software without restriction, including without limitation the rights + * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell + * copies of the Software, and to permit persons to whom the Software is + * furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in + * all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER + * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, + * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN + * THE SOFTWARE. + */ + +#ifndef XLNX_ZYNQMP_CAN_H +#define XLNX_ZYNQMP_CAN_H + +#include "hw/register.h" +#include "net/can_emu.h" +#include "net/can_host.h" +#include "qemu/fifo32.h" +#include "hw/ptimer.h" +#include "hw/qdev-clock.h" + +#define TYPE_XLNX_ZYNQMP_CAN "xlnx.zynqmp-can" + +#define XLNX_ZYNQMP_CAN(obj) \ + OBJECT_CHECK(XlnxZynqMPCANState, (obj), TYPE_XLNX_ZYNQMP_CAN) + +#define MAX_CAN_CTRLS 2 +#define XLNX_ZYNQMP_CAN_R_MAX (0x84 / 4) +#define MAILBOX_CAPACITY 64 +#define CAN_TIMER_MAX 0XFFFFUL +#define CAN_DEFAULT_CLOCK (24 * 1000 * 1000) + +/* Each CAN_FRAME will have 4 * 32bit size. */ +#define CAN_FRAME_SIZE 4 +#define RXFIFO_SIZE (MAILBOX_CAPACITY * CAN_FRAME_SIZE) + +typedef struct XlnxZynqMPCANState { + SysBusDevice parent_obj; + MemoryRegion iomem; + + qemu_irq irq; + + CanBusClientState bus_client; + CanBusState *canbus; + + struct { + uint32_t ext_clk_freq; + } cfg; + + RegisterInfo reg_info[XLNX_ZYNQMP_CAN_R_MAX]; + uint32_t regs[XLNX_ZYNQMP_CAN_R_MAX]; + + Fifo32 rx_fifo; + Fifo32 tx_fifo; + Fifo32 txhpb_fifo; + + ptimer_state *can_timer; +} XlnxZynqMPCANState; + +#endif diff --git a/hw/net/can/xlnx-zynqmp-can.c b/hw/net/can/xlnx-zynqmp-can.c new file mode 100644 index XXXXXXX..XXXXXXX --- /dev/null +++ b/hw/net/can/xlnx-zynqmp-can.c @@ -XXX,XX +XXX,XX @@ +/* + * QEMU model of the Xilinx ZynqMP CAN controller. + * This implementation is based on the following datasheet: + * https://www.xilinx.com/support/documentation/user_guides/ug1085-zynq-ultrascale-trm.pdf + * + * Copyright (c) 2020 Xilinx Inc. + * + * Written-by: Vikram Garhwal<fnu.vikram@xilinx.com> + * + * Based on QEMU CAN Device emulation implemented by Jin Yang, Deniz Eren and + * Pavel Pisa + * + * Permission is hereby granted, free of charge, to any person obtaining a copy + * of this software and associated documentation files (the "Software"), to deal + * in the Software without restriction, including without limitation the rights + * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell + * copies of the Software, and to permit persons to whom the Software is + * furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in + * all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER + * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, + * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN + * THE SOFTWARE. + */ + +#include "qemu/osdep.h" +#include "hw/sysbus.h" +#include "hw/register.h" +#include "hw/irq.h" +#include "qapi/error.h" +#include "qemu/bitops.h" +#include "qemu/log.h" +#include "qemu/cutils.h" +#include "sysemu/sysemu.h" +#include "migration/vmstate.h" +#include "hw/qdev-properties.h" +#include "net/can_emu.h" +#include "net/can_host.h" +#include "qemu/event_notifier.h" +#include "qom/object_interfaces.h" +#include "hw/net/xlnx-zynqmp-can.h" + +#ifndef XLNX_ZYNQMP_CAN_ERR_DEBUG +#define XLNX_ZYNQMP_CAN_ERR_DEBUG 0 +#endif + +#define DB_PRINT(dev, ...) do { \ + if (XLNX_ZYNQMP_CAN_ERR_DEBUG) { \ + g_autofree char *path = object_get_canonical_path(OBJECT(dev)); \ + qemu_log("%s: %s", path, ## __VA_ARGS__); \ + } \ +} while (0) + +#define MAX_DLC 8 +#undef ERROR + +REG32(SOFTWARE_RESET_REGISTER, 0x0) + FIELD(SOFTWARE_RESET_REGISTER, CEN, 1, 1) + FIELD(SOFTWARE_RESET_REGISTER, SRST, 0, 1) +REG32(MODE_SELECT_REGISTER, 0x4) + FIELD(MODE_SELECT_REGISTER, SNOOP, 2, 1) + FIELD(MODE_SELECT_REGISTER, LBACK, 1, 1) + FIELD(MODE_SELECT_REGISTER, SLEEP, 0, 1) +REG32(ARBITRATION_PHASE_BAUD_RATE_PRESCALER_REGISTER, 0x8) + FIELD(ARBITRATION_PHASE_BAUD_RATE_PRESCALER_REGISTER, BRP, 0, 8) +REG32(ARBITRATION_PHASE_BIT_TIMING_REGISTER, 0xc) + FIELD(ARBITRATION_PHASE_BIT_TIMING_REGISTER, SJW, 7, 2) + FIELD(ARBITRATION_PHASE_BIT_TIMING_REGISTER, TS2, 4, 3) + FIELD(ARBITRATION_PHASE_BIT_TIMING_REGISTER, TS1, 0, 4) +REG32(ERROR_COUNTER_REGISTER, 0x10) + FIELD(ERROR_COUNTER_REGISTER, REC, 8, 8) + FIELD(ERROR_COUNTER_REGISTER, TEC, 0, 8) +REG32(ERROR_STATUS_REGISTER, 0x14) + FIELD(ERROR_STATUS_REGISTER, ACKER, 4, 1) + FIELD(ERROR_STATUS_REGISTER, BERR, 3, 1) + FIELD(ERROR_STATUS_REGISTER, STER, 2, 1) + FIELD(ERROR_STATUS_REGISTER, FMER, 1, 1) + FIELD(ERROR_STATUS_REGISTER, CRCER, 0, 1) +REG32(STATUS_REGISTER, 0x18) + FIELD(STATUS_REGISTER, SNOOP, 12, 1) + FIELD(STATUS_REGISTER, ACFBSY, 11, 1) + FIELD(STATUS_REGISTER, TXFLL, 10, 1) + FIELD(STATUS_REGISTER, TXBFLL, 9, 1) + FIELD(STATUS_REGISTER, ESTAT, 7, 2) + FIELD(STATUS_REGISTER, ERRWRN, 6, 1) + FIELD(STATUS_REGISTER, BBSY, 5, 1) + FIELD(STATUS_REGISTER, BIDLE, 4, 1) + FIELD(STATUS_REGISTER, NORMAL, 3, 1) + FIELD(STATUS_REGISTER, SLEEP, 2, 1) + FIELD(STATUS_REGISTER, LBACK, 1, 1) + FIELD(STATUS_REGISTER, CONFIG, 0, 1) +REG32(INTERRUPT_STATUS_REGISTER, 0x1c) + FIELD(INTERRUPT_STATUS_REGISTER, TXFEMP, 14, 1) + FIELD(INTERRUPT_STATUS_REGISTER, TXFWMEMP, 13, 1) + FIELD(INTERRUPT_STATUS_REGISTER, RXFWMFLL, 12, 1) + FIELD(INTERRUPT_STATUS_REGISTER, WKUP, 11, 1) + FIELD(INTERRUPT_STATUS_REGISTER, SLP, 10, 1) + FIELD(INTERRUPT_STATUS_REGISTER, BSOFF, 9, 1) + FIELD(INTERRUPT_STATUS_REGISTER, ERROR, 8, 1) + FIELD(INTERRUPT_STATUS_REGISTER, RXNEMP, 7, 1) + FIELD(INTERRUPT_STATUS_REGISTER, RXOFLW, 6, 1) + FIELD(INTERRUPT_STATUS_REGISTER, RXUFLW, 5, 1) + FIELD(INTERRUPT_STATUS_REGISTER, RXOK, 4, 1) + FIELD(INTERRUPT_STATUS_REGISTER, TXBFLL, 3, 1) + FIELD(INTERRUPT_STATUS_REGISTER, TXFLL, 2, 1) + FIELD(INTERRUPT_STATUS_REGISTER, TXOK, 1, 1) + FIELD(INTERRUPT_STATUS_REGISTER, ARBLST, 0, 1) +REG32(INTERRUPT_ENABLE_REGISTER, 0x20) + FIELD(INTERRUPT_ENABLE_REGISTER, ETXFEMP, 14, 1) + FIELD(INTERRUPT_ENABLE_REGISTER, ETXFWMEMP, 13, 1) + FIELD(INTERRUPT_ENABLE_REGISTER, ERXFWMFLL, 12, 1) + FIELD(INTERRUPT_ENABLE_REGISTER, EWKUP, 11, 1) + FIELD(INTERRUPT_ENABLE_REGISTER, ESLP, 10, 1) + FIELD(INTERRUPT_ENABLE_REGISTER, EBSOFF, 9, 1) + FIELD(INTERRUPT_ENABLE_REGISTER, EERROR, 8, 1) + FIELD(INTERRUPT_ENABLE_REGISTER, ERXNEMP, 7, 1) + FIELD(INTERRUPT_ENABLE_REGISTER, ERXOFLW, 6, 1) + FIELD(INTERRUPT_ENABLE_REGISTER, ERXUFLW, 5, 1) + FIELD(INTERRUPT_ENABLE_REGISTER, ERXOK, 4, 1) + FIELD(INTERRUPT_ENABLE_REGISTER, ETXBFLL, 3, 1) + FIELD(INTERRUPT_ENABLE_REGISTER, ETXFLL, 2, 1) + FIELD(INTERRUPT_ENABLE_REGISTER, ETXOK, 1, 1) + FIELD(INTERRUPT_ENABLE_REGISTER, EARBLST, 0, 1) +REG32(INTERRUPT_CLEAR_REGISTER, 0x24) + FIELD(INTERRUPT_CLEAR_REGISTER, CTXFEMP, 14, 1) + FIELD(INTERRUPT_CLEAR_REGISTER, CTXFWMEMP, 13, 1) + FIELD(INTERRUPT_CLEAR_REGISTER, CRXFWMFLL, 12, 1) + FIELD(INTERRUPT_CLEAR_REGISTER, CWKUP, 11, 1) + FIELD(INTERRUPT_CLEAR_REGISTER, CSLP, 10, 1) + FIELD(INTERRUPT_CLEAR_REGISTER, CBSOFF, 9, 1) + FIELD(INTERRUPT_CLEAR_REGISTER, CERROR, 8, 1) + FIELD(INTERRUPT_CLEAR_REGISTER, CRXNEMP, 7, 1) + FIELD(INTERRUPT_CLEAR_REGISTER, CRXOFLW, 6, 1) + FIELD(INTERRUPT_CLEAR_REGISTER, CRXUFLW, 5, 1) + FIELD(INTERRUPT_CLEAR_REGISTER, CRXOK, 4, 1) + FIELD(INTERRUPT_CLEAR_REGISTER, CTXBFLL, 3, 1) + FIELD(INTERRUPT_CLEAR_REGISTER, CTXFLL, 2, 1) + FIELD(INTERRUPT_CLEAR_REGISTER, CTXOK, 1, 1) + FIELD(INTERRUPT_CLEAR_REGISTER, CARBLST, 0, 1) +REG32(TIMESTAMP_REGISTER, 0x28) + FIELD(TIMESTAMP_REGISTER, CTS, 0, 1) +REG32(WIR, 0x2c) + FIELD(WIR, EW, 8, 8) + FIELD(WIR, FW, 0, 8) +REG32(TXFIFO_ID, 0x30) + FIELD(TXFIFO_ID, IDH, 21, 11) + FIELD(TXFIFO_ID, SRRRTR, 20, 1) + FIELD(TXFIFO_ID, IDE, 19, 1) + FIELD(TXFIFO_ID, IDL, 1, 18) + FIELD(TXFIFO_ID, RTR, 0, 1) +REG32(TXFIFO_DLC, 0x34) + FIELD(TXFIFO_DLC, DLC, 28, 4) +REG32(TXFIFO_DATA1, 0x38) + FIELD(TXFIFO_DATA1, DB0, 24, 8) + FIELD(TXFIFO_DATA1, DB1, 16, 8) + FIELD(TXFIFO_DATA1, DB2, 8, 8) + FIELD(TXFIFO_DATA1, DB3, 0, 8) +REG32(TXFIFO_DATA2, 0x3c) + FIELD(TXFIFO_DATA2, DB4, 24, 8) + FIELD(TXFIFO_DATA2, DB5, 16, 8) + FIELD(TXFIFO_DATA2, DB6, 8, 8) + FIELD(TXFIFO_DATA2, DB7, 0, 8) +REG32(TXHPB_ID, 0x40) + FIELD(TXHPB_ID, IDH, 21, 11) + FIELD(TXHPB_ID, SRRRTR, 20, 1) + FIELD(TXHPB_ID, IDE, 19, 1) + FIELD(TXHPB_ID, IDL, 1, 18) + FIELD(TXHPB_ID, RTR, 0, 1) +REG32(TXHPB_DLC, 0x44) + FIELD(TXHPB_DLC, DLC, 28, 4) +REG32(TXHPB_DATA1, 0x48) + FIELD(TXHPB_DATA1, DB0, 24, 8) + FIELD(TXHPB_DATA1, DB1, 16, 8) + FIELD(TXHPB_DATA1, DB2, 8, 8) + FIELD(TXHPB_DATA1, DB3, 0, 8) +REG32(TXHPB_DATA2, 0x4c) + FIELD(TXHPB_DATA2, DB4, 24, 8) + FIELD(TXHPB_DATA2, DB5, 16, 8) + FIELD(TXHPB_DATA2, DB6, 8, 8) + FIELD(TXHPB_DATA2, DB7, 0, 8) +REG32(RXFIFO_ID, 0x50) + FIELD(RXFIFO_ID, IDH, 21, 11) + FIELD(RXFIFO_ID, SRRRTR, 20, 1) + FIELD(RXFIFO_ID, IDE, 19, 1) + FIELD(RXFIFO_ID, IDL, 1, 18) + FIELD(RXFIFO_ID, RTR, 0, 1) +REG32(RXFIFO_DLC, 0x54) + FIELD(RXFIFO_DLC, DLC, 28, 4) + FIELD(RXFIFO_DLC, RXT, 0, 16) +REG32(RXFIFO_DATA1, 0x58) + FIELD(RXFIFO_DATA1, DB0, 24, 8) + FIELD(RXFIFO_DATA1, DB1, 16, 8) + FIELD(RXFIFO_DATA1, DB2, 8, 8) + FIELD(RXFIFO_DATA1, DB3, 0, 8) +REG32(RXFIFO_DATA2, 0x5c) + FIELD(RXFIFO_DATA2, DB4, 24, 8) + FIELD(RXFIFO_DATA2, DB5, 16, 8) + FIELD(RXFIFO_DATA2, DB6, 8, 8) + FIELD(RXFIFO_DATA2, DB7, 0, 8) +REG32(AFR, 0x60) + FIELD(AFR, UAF4, 3, 1) + FIELD(AFR, UAF3, 2, 1) + FIELD(AFR, UAF2, 1, 1) + FIELD(AFR, UAF1, 0, 1) +REG32(AFMR1, 0x64) + FIELD(AFMR1, AMIDH, 21, 11) + FIELD(AFMR1, AMSRR, 20, 1) + FIELD(AFMR1, AMIDE, 19, 1) + FIELD(AFMR1, AMIDL, 1, 18) + FIELD(AFMR1, AMRTR, 0, 1) +REG32(AFIR1, 0x68) + FIELD(AFIR1, AIIDH, 21, 11) + FIELD(AFIR1, AISRR, 20, 1) + FIELD(AFIR1, AIIDE, 19, 1) + FIELD(AFIR1, AIIDL, 1, 18) + FIELD(AFIR1, AIRTR, 0, 1) +REG32(AFMR2, 0x6c) + FIELD(AFMR2, AMIDH, 21, 11) + FIELD(AFMR2, AMSRR, 20, 1) + FIELD(AFMR2, AMIDE, 19, 1) + FIELD(AFMR2, AMIDL, 1, 18) + FIELD(AFMR2, AMRTR, 0, 1) +REG32(AFIR2, 0x70) + FIELD(AFIR2, AIIDH, 21, 11) + FIELD(AFIR2, AISRR, 20, 1) + FIELD(AFIR2, AIIDE, 19, 1) + FIELD(AFIR2, AIIDL, 1, 18) + FIELD(AFIR2, AIRTR, 0, 1) +REG32(AFMR3, 0x74) + FIELD(AFMR3, AMIDH, 21, 11) + FIELD(AFMR3, AMSRR, 20, 1) + FIELD(AFMR3, AMIDE, 19, 1) + FIELD(AFMR3, AMIDL, 1, 18) + FIELD(AFMR3, AMRTR, 0, 1) +REG32(AFIR3, 0x78) + FIELD(AFIR3, AIIDH, 21, 11) + FIELD(AFIR3, AISRR, 20, 1) + FIELD(AFIR3, AIIDE, 19, 1) + FIELD(AFIR3, AIIDL, 1, 18) + FIELD(AFIR3, AIRTR, 0, 1) +REG32(AFMR4, 0x7c) + FIELD(AFMR4, AMIDH, 21, 11) + FIELD(AFMR4, AMSRR, 20, 1) + FIELD(AFMR4, AMIDE, 19, 1) + FIELD(AFMR4, AMIDL, 1, 18) + FIELD(AFMR4, AMRTR, 0, 1) +REG32(AFIR4, 0x80) + FIELD(AFIR4, AIIDH, 21, 11) + FIELD(AFIR4, AISRR, 20, 1) + FIELD(AFIR4, AIIDE, 19, 1) + FIELD(AFIR4, AIIDL, 1, 18) + FIELD(AFIR4, AIRTR, 0, 1) + +static void can_update_irq(XlnxZynqMPCANState *s) +{ + uint32_t irq; + + /* Watermark register interrupts. */ + if ((fifo32_num_free(&s->tx_fifo) / CAN_FRAME_SIZE) > + ARRAY_FIELD_EX32(s->regs, WIR, EW)) { + ARRAY_FIELD_DP32(s->regs, INTERRUPT_STATUS_REGISTER, TXFWMEMP, 1); + } + + if ((fifo32_num_used(&s->rx_fifo) / CAN_FRAME_SIZE) > + ARRAY_FIELD_EX32(s->regs, WIR, FW)) { + ARRAY_FIELD_DP32(s->regs, INTERRUPT_STATUS_REGISTER, RXFWMFLL, 1); + } + + /* RX Interrupts. */ + if (fifo32_num_used(&s->rx_fifo) >= CAN_FRAME_SIZE) { + ARRAY_FIELD_DP32(s->regs, INTERRUPT_STATUS_REGISTER, RXNEMP, 1); + } + + /* TX interrupts. */ + if (fifo32_is_empty(&s->tx_fifo)) { + ARRAY_FIELD_DP32(s->regs, INTERRUPT_STATUS_REGISTER, TXFEMP, 1); + } + + if (fifo32_is_full(&s->tx_fifo)) { + ARRAY_FIELD_DP32(s->regs, INTERRUPT_STATUS_REGISTER, TXFLL, 1); + } + + if (fifo32_is_full(&s->txhpb_fifo)) { + ARRAY_FIELD_DP32(s->regs, INTERRUPT_STATUS_REGISTER, TXBFLL, 1); + } + + irq = s->regs[R_INTERRUPT_STATUS_REGISTER]; + irq &= s->regs[R_INTERRUPT_ENABLE_REGISTER]; + + qemu_set_irq(s->irq, irq); +} + +static void can_ier_post_write(RegisterInfo *reg, uint64_t val64) +{ + XlnxZynqMPCANState *s = XLNX_ZYNQMP_CAN(reg->opaque); + + can_update_irq(s); +} + +static uint64_t can_icr_pre_write(RegisterInfo *reg, uint64_t val64) +{ + XlnxZynqMPCANState *s = XLNX_ZYNQMP_CAN(reg->opaque); + uint32_t val = val64; + + s->regs[R_INTERRUPT_STATUS_REGISTER] &= ~val; + can_update_irq(s); + + return 0; +} + +static void can_config_reset(XlnxZynqMPCANState *s) +{ + /* Reset all the configuration registers. */ + register_reset(&s->reg_info[R_SOFTWARE_RESET_REGISTER]); + register_reset(&s->reg_info[R_MODE_SELECT_REGISTER]); + register_reset( + &s->reg_info[R_ARBITRATION_PHASE_BAUD_RATE_PRESCALER_REGISTER]); + register_reset(&s->reg_info[R_ARBITRATION_PHASE_BIT_TIMING_REGISTER]); + register_reset(&s->reg_info[R_STATUS_REGISTER]); + register_reset(&s->reg_info[R_INTERRUPT_STATUS_REGISTER]); + register_reset(&s->reg_info[R_INTERRUPT_ENABLE_REGISTER]); + register_reset(&s->reg_info[R_INTERRUPT_CLEAR_REGISTER]); + register_reset(&s->reg_info[R_WIR]); +} + +static void can_config_mode(XlnxZynqMPCANState *s) +{ + register_reset(&s->reg_info[R_ERROR_COUNTER_REGISTER]); + register_reset(&s->reg_info[R_ERROR_STATUS_REGISTER]); + + /* Put XlnxZynqMPCAN in configuration mode. */ + ARRAY_FIELD_DP32(s->regs, STATUS_REGISTER, CONFIG, 1); + ARRAY_FIELD_DP32(s->regs, INTERRUPT_STATUS_REGISTER, WKUP, 0); + ARRAY_FIELD_DP32(s->regs, INTERRUPT_STATUS_REGISTER, SLP, 0); + ARRAY_FIELD_DP32(s->regs, INTERRUPT_STATUS_REGISTER, BSOFF, 0); + ARRAY_FIELD_DP32(s->regs, INTERRUPT_STATUS_REGISTER, ERROR, 0); + ARRAY_FIELD_DP32(s->regs, INTERRUPT_STATUS_REGISTER, RXOFLW, 0); + ARRAY_FIELD_DP32(s->regs, INTERRUPT_STATUS_REGISTER, RXOK, 0); + ARRAY_FIELD_DP32(s->regs, INTERRUPT_STATUS_REGISTER, TXOK, 0); + ARRAY_FIELD_DP32(s->regs, INTERRUPT_STATUS_REGISTER, ARBLST, 0); + + can_update_irq(s); +} + +static void update_status_register_mode_bits(XlnxZynqMPCANState *s) +{ + bool sleep_status = ARRAY_FIELD_EX32(s->regs, STATUS_REGISTER, SLEEP); + bool sleep_mode = ARRAY_FIELD_EX32(s->regs, MODE_SELECT_REGISTER, SLEEP); + /* Wake up interrupt bit. */ + bool wakeup_irq_val = sleep_status && (sleep_mode == 0); + /* Sleep interrupt bit. */ + bool sleep_irq_val = sleep_mode && (sleep_status == 0); + + /* Clear previous core mode status bits. */ + ARRAY_FIELD_DP32(s->regs, STATUS_REGISTER, LBACK, 0); + ARRAY_FIELD_DP32(s->regs, STATUS_REGISTER, SLEEP, 0); + ARRAY_FIELD_DP32(s->regs, STATUS_REGISTER, SNOOP, 0); + ARRAY_FIELD_DP32(s->regs, STATUS_REGISTER, NORMAL, 0); + + /* set current mode bit and generate irqs accordingly. */ + if (ARRAY_FIELD_EX32(s->regs, MODE_SELECT_REGISTER, LBACK)) { + ARRAY_FIELD_DP32(s->regs, STATUS_REGISTER, LBACK, 1); + } else if (ARRAY_FIELD_EX32(s->regs, MODE_SELECT_REGISTER, SLEEP)) { + ARRAY_FIELD_DP32(s->regs, STATUS_REGISTER, SLEEP, 1); + ARRAY_FIELD_DP32(s->regs, INTERRUPT_STATUS_REGISTER, SLP, + sleep_irq_val); + } else if (ARRAY_FIELD_EX32(s->regs, MODE_SELECT_REGISTER, SNOOP)) { + ARRAY_FIELD_DP32(s->regs, STATUS_REGISTER, SNOOP, 1); + } else { + /* + * If all bits are zero then XlnxZynqMPCAN is set in normal mode. + */ + ARRAY_FIELD_DP32(s->regs, STATUS_REGISTER, NORMAL, 1); + /* Set wakeup interrupt bit. */ + ARRAY_FIELD_DP32(s->regs, INTERRUPT_STATUS_REGISTER, WKUP, + wakeup_irq_val); + } + + can_update_irq(s); +} + +static void can_exit_sleep_mode(XlnxZynqMPCANState *s) +{ + ARRAY_FIELD_DP32(s->regs, MODE_SELECT_REGISTER, SLEEP, 0); + update_status_register_mode_bits(s); +} + +static void generate_frame(qemu_can_frame *frame, uint32_t *data) +{ + frame->can_id = data[0]; + frame->can_dlc = FIELD_EX32(data[1], TXFIFO_DLC, DLC); + + frame->data[0] = FIELD_EX32(data[2], TXFIFO_DATA1, DB3); + frame->data[1] = FIELD_EX32(data[2], TXFIFO_DATA1, DB2); + frame->data[2] = FIELD_EX32(data[2], TXFIFO_DATA1, DB1); + frame->data[3] = FIELD_EX32(data[2], TXFIFO_DATA1, DB0); + + frame->data[4] = FIELD_EX32(data[3], TXFIFO_DATA2, DB7); + frame->data[5] = FIELD_EX32(data[3], TXFIFO_DATA2, DB6); + frame->data[6] = FIELD_EX32(data[3], TXFIFO_DATA2, DB5); + frame->data[7] = FIELD_EX32(data[3], TXFIFO_DATA2, DB4); +} + +static bool tx_ready_check(XlnxZynqMPCANState *s) +{ + if (ARRAY_FIELD_EX32(s->regs, SOFTWARE_RESET_REGISTER, SRST)) { + g_autofree char *path = object_get_canonical_path(OBJECT(s)); + + qemu_log_mask(LOG_GUEST_ERROR, "%s: Attempting to transfer data while" + " data while controller is in reset mode.\n", + path); + return false; + } + + if (ARRAY_FIELD_EX32(s->regs, SOFTWARE_RESET_REGISTER, CEN) == 0) { + g_autofree char *path = object_get_canonical_path(OBJECT(s)); + + qemu_log_mask(LOG_GUEST_ERROR, "%s: Attempting to transfer" + " data while controller is in configuration mode. Reset" + " the core so operations can start fresh.\n", + path); + return false; + } + + if (ARRAY_FIELD_EX32(s->regs, STATUS_REGISTER, SNOOP)) { + g_autofree char *path = object_get_canonical_path(OBJECT(s)); + + qemu_log_mask(LOG_GUEST_ERROR, "%s: Attempting to transfer" + " data while controller is in SNOOP MODE.\n", + path); + return false; + } + + return true; +} + +static void transfer_fifo(XlnxZynqMPCANState *s, Fifo32 *fifo) +{ + qemu_can_frame frame; + uint32_t data[CAN_FRAME_SIZE]; + int i; + bool can_tx = tx_ready_check(s); + + if (can_tx) { + while (!fifo32_is_empty(fifo)) { + for (i = 0; i < CAN_FRAME_SIZE; i++) { + data[i] = fifo32_pop(fifo); + } + + if (ARRAY_FIELD_EX32(s->regs, STATUS_REGISTER, LBACK)) { + /* + * Controller is in loopback. In Loopback mode, the CAN core + * transmits a recessive bitstream on to the XlnxZynqMPCAN Bus. + * Any message transmitted is looped back to the RX line and + * acknowledged. The XlnxZynqMPCAN core receives any message + * that it transmits. + */ + if (fifo32_is_full(&s->rx_fifo)) { + DB_PRINT(s, "Loopback: RX FIFO is full." + " TX FIFO will be flushed.\n"); + + ARRAY_FIELD_DP32(s->regs, INTERRUPT_STATUS_REGISTER, + RXOFLW, 1); + } else { + for (i = 0; i < CAN_FRAME_SIZE; i++) { + fifo32_push(&s->rx_fifo, data[i]); + } + + ARRAY_FIELD_DP32(s->regs, INTERRUPT_STATUS_REGISTER, + RXOK, 1); + } + } else { + /* Normal mode Tx. */ + generate_frame(&frame, data); + + can_bus_client_send(&s->bus_client, &frame, 1); + } + } + + ARRAY_FIELD_DP32(s->regs, INTERRUPT_STATUS_REGISTER, TXOK, 1); + ARRAY_FIELD_DP32(s->regs, STATUS_REGISTER, TXBFLL, 0); + + if (ARRAY_FIELD_EX32(s->regs, STATUS_REGISTER, SLEEP)) { + can_exit_sleep_mode(s); + } + } else { + DB_PRINT(s, "Not enabled for data transfer.\n"); + } + + can_update_irq(s); +} + +static uint64_t can_srr_pre_write(RegisterInfo *reg, uint64_t val64) +{ + XlnxZynqMPCANState *s = XLNX_ZYNQMP_CAN(reg->opaque); + uint32_t val = val64; + + ARRAY_FIELD_DP32(s->regs, SOFTWARE_RESET_REGISTER, CEN, + FIELD_EX32(val, SOFTWARE_RESET_REGISTER, CEN)); + + if (FIELD_EX32(val, SOFTWARE_RESET_REGISTER, SRST)) { + DB_PRINT(s, "Resetting controller.\n"); + + /* First, core will do software reset then will enter in config mode. */ + can_config_reset(s); + } + + if (ARRAY_FIELD_EX32(s->regs, SOFTWARE_RESET_REGISTER, CEN) == 0) { + can_config_mode(s); + } else { + /* + * Leave config mode. Now XlnxZynqMPCAN core will enter normal, + * sleep, snoop or loopback mode depending upon LBACK, SLEEP, SNOOP + * register states. + */ + ARRAY_FIELD_DP32(s->regs, STATUS_REGISTER, CONFIG, 0); + + ptimer_transaction_begin(s->can_timer); + ptimer_set_count(s->can_timer, 0); + ptimer_transaction_commit(s->can_timer); + + /* XlnxZynqMPCAN is out of config mode. It will send pending data. */ + transfer_fifo(s, &s->txhpb_fifo); + transfer_fifo(s, &s->tx_fifo); + } + + update_status_register_mode_bits(s); + + return s->regs[R_SOFTWARE_RESET_REGISTER]; +} + +static uint64_t can_msr_pre_write(RegisterInfo *reg, uint64_t val64) +{ + XlnxZynqMPCANState *s = XLNX_ZYNQMP_CAN(reg->opaque); + uint32_t val = val64; + uint8_t multi_mode; + + /* + * Multiple mode set check. This is done to make sure user doesn't set + * multiple modes. + */ + multi_mode = FIELD_EX32(val, MODE_SELECT_REGISTER, LBACK) + + FIELD_EX32(val, MODE_SELECT_REGISTER, SLEEP) + + FIELD_EX32(val, MODE_SELECT_REGISTER, SNOOP); + + if (multi_mode > 1) { + g_autofree char *path = object_get_canonical_path(OBJECT(s)); + + qemu_log_mask(LOG_GUEST_ERROR, "%s: Attempting to config" + " several modes simultaneously. One mode will be selected" + " according to their priority: LBACK > SLEEP > SNOOP.\n", + path); + } + + if (ARRAY_FIELD_EX32(s->regs, SOFTWARE_RESET_REGISTER, CEN) == 0) { + /* We are in configuration mode, any mode can be selected. */ + s->regs[R_MODE_SELECT_REGISTER] = val; + } else { + bool sleep_mode_bit = FIELD_EX32(val, MODE_SELECT_REGISTER, SLEEP); + + ARRAY_FIELD_DP32(s->regs, MODE_SELECT_REGISTER, SLEEP, sleep_mode_bit); + + if (FIELD_EX32(val, MODE_SELECT_REGISTER, LBACK)) { + g_autofree char *path = object_get_canonical_path(OBJECT(s)); + + qemu_log_mask(LOG_GUEST_ERROR, "%s: Attempting to set" + " LBACK mode without setting CEN bit as 0.\n", + path); + } else if (FIELD_EX32(val, MODE_SELECT_REGISTER, SNOOP)) { + g_autofree char *path = object_get_canonical_path(OBJECT(s)); + + qemu_log_mask(LOG_GUEST_ERROR, "%s: Attempting to set" + " SNOOP mode without setting CEN bit as 0.\n", + path); + } + + update_status_register_mode_bits(s); + } + + return s->regs[R_MODE_SELECT_REGISTER]; +} + +static uint64_t can_brpr_pre_write(RegisterInfo *reg, uint64_t val64) +{ + XlnxZynqMPCANState *s = XLNX_ZYNQMP_CAN(reg->opaque); + uint32_t val = val64; + + /* Only allow writes when in config mode. */ + if (ARRAY_FIELD_EX32(s->regs, SOFTWARE_RESET_REGISTER, CEN)) { + val = s->regs[R_ARBITRATION_PHASE_BAUD_RATE_PRESCALER_REGISTER]; + } + + return val; +} + +static uint64_t can_btr_pre_write(RegisterInfo *reg, uint64_t val64) +{ + XlnxZynqMPCANState *s = XLNX_ZYNQMP_CAN(reg->opaque); + uint32_t val = val64; + + /* Only allow writes when in config mode. */ + if (ARRAY_FIELD_EX32(s->regs, SOFTWARE_RESET_REGISTER, CEN)) { + val = s->regs[R_ARBITRATION_PHASE_BIT_TIMING_REGISTER]; + } + + return val; +} + +static uint64_t can_tcr_pre_write(RegisterInfo *reg, uint64_t val64) +{ + XlnxZynqMPCANState *s = XLNX_ZYNQMP_CAN(reg->opaque); + uint32_t val = val64; + + if (FIELD_EX32(val, TIMESTAMP_REGISTER, CTS)) { + ptimer_transaction_begin(s->can_timer); + ptimer_set_count(s->can_timer, 0); + ptimer_transaction_commit(s->can_timer); + } + + return 0; +} + +static void update_rx_fifo(XlnxZynqMPCANState *s, const qemu_can_frame *frame) +{ + bool filter_pass = false; + uint16_t timestamp = 0; + + /* If no filter is enabled. Message will be stored in FIFO. */ + if (!((ARRAY_FIELD_EX32(s->regs, AFR, UAF1)) | + (ARRAY_FIELD_EX32(s->regs, AFR, UAF2)) | + (ARRAY_FIELD_EX32(s->regs, AFR, UAF3)) | + (ARRAY_FIELD_EX32(s->regs, AFR, UAF4)))) { + filter_pass = true; + } + + /* + * Messages that pass any of the acceptance filters will be stored in + * the RX FIFO. + */ + if (ARRAY_FIELD_EX32(s->regs, AFR, UAF1)) { + uint32_t id_masked = s->regs[R_AFMR1] & frame->can_id; + uint32_t filter_id_masked = s->regs[R_AFMR1] & s->regs[R_AFIR1]; + + if (filter_id_masked == id_masked) { + filter_pass = true; + } + } + + if (ARRAY_FIELD_EX32(s->regs, AFR, UAF2)) { + uint32_t id_masked = s->regs[R_AFMR2] & frame->can_id; + uint32_t filter_id_masked = s->regs[R_AFMR2] & s->regs[R_AFIR2]; + + if (filter_id_masked == id_masked) { + filter_pass = true; + } + } + + if (ARRAY_FIELD_EX32(s->regs, AFR, UAF3)) { + uint32_t id_masked = s->regs[R_AFMR3] & frame->can_id; + uint32_t filter_id_masked = s->regs[R_AFMR3] & s->regs[R_AFIR3]; + + if (filter_id_masked == id_masked) { + filter_pass = true; + } + } + + if (ARRAY_FIELD_EX32(s->regs, AFR, UAF4)) { + uint32_t id_masked = s->regs[R_AFMR4] & frame->can_id; + uint32_t filter_id_masked = s->regs[R_AFMR4] & s->regs[R_AFIR4]; + + if (filter_id_masked == id_masked) { + filter_pass = true; + } + } + + /* Store the message in fifo if it passed through any of the filters. */ + if (filter_pass && frame->can_dlc <= MAX_DLC) { + + if (fifo32_is_full(&s->rx_fifo)) { + DB_PRINT(s, "RX FIFO is full.\n"); + + ARRAY_FIELD_DP32(s->regs, INTERRUPT_STATUS_REGISTER, RXOFLW, 1); + } else { + timestamp = CAN_TIMER_MAX - ptimer_get_count(s->can_timer); + + fifo32_push(&s->rx_fifo, frame->can_id); + + fifo32_push(&s->rx_fifo, deposit32(0, R_RXFIFO_DLC_DLC_SHIFT, + R_RXFIFO_DLC_DLC_LENGTH, + frame->can_dlc) | + deposit32(0, R_RXFIFO_DLC_RXT_SHIFT, + R_RXFIFO_DLC_RXT_LENGTH, + timestamp)); + + /* First 32 bit of the data. */ + fifo32_push(&s->rx_fifo, deposit32(0, R_TXFIFO_DATA1_DB3_SHIFT, + R_TXFIFO_DATA1_DB3_LENGTH, + frame->data[0]) | + deposit32(0, R_TXFIFO_DATA1_DB2_SHIFT, + R_TXFIFO_DATA1_DB2_LENGTH, + frame->data[1]) | + deposit32(0, R_TXFIFO_DATA1_DB1_SHIFT, + R_TXFIFO_DATA1_DB1_LENGTH, + frame->data[2]) | + deposit32(0, R_TXFIFO_DATA1_DB0_SHIFT, + R_TXFIFO_DATA1_DB0_LENGTH, + frame->data[3])); + /* Last 32 bit of the data. */ + fifo32_push(&s->rx_fifo, deposit32(0, R_TXFIFO_DATA2_DB7_SHIFT, + R_TXFIFO_DATA2_DB7_LENGTH, + frame->data[4]) | + deposit32(0, R_TXFIFO_DATA2_DB6_SHIFT, + R_TXFIFO_DATA2_DB6_LENGTH, + frame->data[5]) | + deposit32(0, R_TXFIFO_DATA2_DB5_SHIFT, + R_TXFIFO_DATA2_DB5_LENGTH, + frame->data[6]) | + deposit32(0, R_TXFIFO_DATA2_DB4_SHIFT, + R_TXFIFO_DATA2_DB4_LENGTH, + frame->data[7])); + + ARRAY_FIELD_DP32(s->regs, INTERRUPT_STATUS_REGISTER, RXOK, 1); + } + + can_update_irq(s); + } else { + DB_PRINT(s, "Message didn't pass through any filter or dlc" + " is not in range.\n"); + } +} + +static uint64_t can_rxfifo_pre_read(RegisterInfo *reg, uint64_t val64) +{ + XlnxZynqMPCANState *s = XLNX_ZYNQMP_CAN(reg->opaque); + uint32_t r = 0; + + if (!fifo32_is_empty(&s->rx_fifo)) { + r = fifo32_pop(&s->rx_fifo); + } else { + DB_PRINT(s, "No message in RXFIFO.\n"); + + ARRAY_FIELD_DP32(s->regs, INTERRUPT_STATUS_REGISTER, RXUFLW, 1); + } + + can_update_irq(s); + return r; +} + +static void can_filter_enable_post_write(RegisterInfo *reg, uint64_t val64) +{ + XlnxZynqMPCANState *s = XLNX_ZYNQMP_CAN(reg->opaque); + + if (ARRAY_FIELD_EX32(s->regs, AFR, UAF1) && + ARRAY_FIELD_EX32(s->regs, AFR, UAF2) && + ARRAY_FIELD_EX32(s->regs, AFR, UAF3) && + ARRAY_FIELD_EX32(s->regs, AFR, UAF4)) { + ARRAY_FIELD_DP32(s->regs, STATUS_REGISTER, ACFBSY, 1); + } else { + ARRAY_FIELD_DP32(s->regs, STATUS_REGISTER, ACFBSY, 0); + } +} + +static uint64_t can_filter_mask_pre_write(RegisterInfo *reg, uint64_t val64) +{ + XlnxZynqMPCANState *s = XLNX_ZYNQMP_CAN(reg->opaque); + uint32_t reg_idx = (reg->access->addr) / 4; + uint32_t val = val64; + uint32_t filter_number = (reg_idx - R_AFMR1) / 2; + + /* modify an acceptance filter, the corresponding UAF bit should be '0.' */ + if (!(s->regs[R_AFR] & (1 << filter_number))) { + s->regs[reg_idx] = val; + } else { + g_autofree char *path = object_get_canonical_path(OBJECT(s)); + + qemu_log_mask(LOG_GUEST_ERROR, "%s: Acceptance filter %d" + " mask is not set as corresponding UAF bit is not 0.\n", + path, filter_number + 1); + } + + return s->regs[reg_idx]; +} + +static uint64_t can_filter_id_pre_write(RegisterInfo *reg, uint64_t val64) +{ + XlnxZynqMPCANState *s = XLNX_ZYNQMP_CAN(reg->opaque); + uint32_t reg_idx = (reg->access->addr) / 4; + uint32_t val = val64; + uint32_t filter_number = (reg_idx - R_AFIR1) / 2; + + if (!(s->regs[R_AFR] & (1 << filter_number))) { + s->regs[reg_idx] = val; + } else { + g_autofree char *path = object_get_canonical_path(OBJECT(s)); + + qemu_log_mask(LOG_GUEST_ERROR, "%s: Acceptance filter %d" + " id is not set as corresponding UAF bit is not 0.\n", + path, filter_number + 1); + } + + return s->regs[reg_idx]; +} + +static void can_tx_post_write(RegisterInfo *reg, uint64_t val64) +{ + XlnxZynqMPCANState *s = XLNX_ZYNQMP_CAN(reg->opaque); + uint32_t val = val64; + + bool is_txhpb = reg->access->addr > A_TXFIFO_DATA2; + + bool initiate_transfer = (reg->access->addr == A_TXFIFO_DATA2) || + (reg->access->addr == A_TXHPB_DATA2); + + Fifo32 *f = is_txhpb ? &s->txhpb_fifo : &s->tx_fifo; + + DB_PRINT(s, "TX FIFO write.\n"); + + if (!fifo32_is_full(f)) { + fifo32_push(f, val); + } else { + g_autofree char *path = object_get_canonical_path(OBJECT(s)); + + qemu_log_mask(LOG_GUEST_ERROR, "%s: TX FIFO is full.\n", path); + } + + /* Initiate the message send if TX register is written. */ + if (initiate_transfer && + ARRAY_FIELD_EX32(s->regs, SOFTWARE_RESET_REGISTER, CEN)) { + transfer_fifo(s, f); + } + + can_update_irq(s); +} + +static const RegisterAccessInfo can_regs_info[] = { + { .name = "SOFTWARE_RESET_REGISTER", + .addr = A_SOFTWARE_RESET_REGISTER, + .rsvd = 0xfffffffc, + .pre_write = can_srr_pre_write, + },{ .name = "MODE_SELECT_REGISTER", + .addr = A_MODE_SELECT_REGISTER, + .rsvd = 0xfffffff8, + .pre_write = can_msr_pre_write, + },{ .name = "ARBITRATION_PHASE_BAUD_RATE_PRESCALER_REGISTER", + .addr = A_ARBITRATION_PHASE_BAUD_RATE_PRESCALER_REGISTER, + .rsvd = 0xffffff00, + .pre_write = can_brpr_pre_write, + },{ .name = "ARBITRATION_PHASE_BIT_TIMING_REGISTER", + .addr = A_ARBITRATION_PHASE_BIT_TIMING_REGISTER, + .rsvd = 0xfffffe00, + .pre_write = can_btr_pre_write, + },{ .name = "ERROR_COUNTER_REGISTER", + .addr = A_ERROR_COUNTER_REGISTER, + .rsvd = 0xffff0000, + .ro = 0xffffffff, + },{ .name = "ERROR_STATUS_REGISTER", + .addr = A_ERROR_STATUS_REGISTER, + .rsvd = 0xffffffe0, + .w1c = 0x1f, + },{ .name = "STATUS_REGISTER", .addr = A_STATUS_REGISTER, + .reset = 0x1, + .rsvd = 0xffffe000, + .ro = 0x1fff, + },{ .name = "INTERRUPT_STATUS_REGISTER", + .addr = A_INTERRUPT_STATUS_REGISTER, + .reset = 0x6000, + .rsvd = 0xffff8000, + .ro = 0x7fff, + },{ .name = "INTERRUPT_ENABLE_REGISTER", + .addr = A_INTERRUPT_ENABLE_REGISTER, + .rsvd = 0xffff8000, + .post_write = can_ier_post_write, + },{ .name = "INTERRUPT_CLEAR_REGISTER", + .addr = A_INTERRUPT_CLEAR_REGISTER, + .rsvd = 0xffff8000, + .pre_write = can_icr_pre_write, + },{ .name = "TIMESTAMP_REGISTER", + .addr = A_TIMESTAMP_REGISTER, + .rsvd = 0xfffffffe, + .pre_write = can_tcr_pre_write, + },{ .name = "WIR", .addr = A_WIR, + .reset = 0x3f3f, + .rsvd = 0xffff0000, + },{ .name = "TXFIFO_ID", .addr = A_TXFIFO_ID, + .post_write = can_tx_post_write, + },{ .name = "TXFIFO_DLC", .addr = A_TXFIFO_DLC, + .rsvd = 0xfffffff, + .post_write = can_tx_post_write, + },{ .name = "TXFIFO_DATA1", .addr = A_TXFIFO_DATA1, + .post_write = can_tx_post_write, + },{ .name = "TXFIFO_DATA2", .addr = A_TXFIFO_DATA2, + .post_write = can_tx_post_write, + },{ .name = "TXHPB_ID", .addr = A_TXHPB_ID, + .post_write = can_tx_post_write, + },{ .name = "TXHPB_DLC", .addr = A_TXHPB_DLC, + .rsvd = 0xfffffff, + .post_write = can_tx_post_write, + },{ .name = "TXHPB_DATA1", .addr = A_TXHPB_DATA1, + .post_write = can_tx_post_write, + },{ .name = "TXHPB_DATA2", .addr = A_TXHPB_DATA2, + .post_write = can_tx_post_write, + },{ .name = "RXFIFO_ID", .addr = A_RXFIFO_ID, + .ro = 0xffffffff, + .post_read = can_rxfifo_pre_read, + },{ .name = "RXFIFO_DLC", .addr = A_RXFIFO_DLC, + .rsvd = 0xfff0000, + .post_read = can_rxfifo_pre_read, + },{ .name = "RXFIFO_DATA1", .addr = A_RXFIFO_DATA1, + .post_read = can_rxfifo_pre_read, + },{ .name = "RXFIFO_DATA2", .addr = A_RXFIFO_DATA2, + .post_read = can_rxfifo_pre_read, + },{ .name = "AFR", .addr = A_AFR, + .rsvd = 0xfffffff0, + .post_write = can_filter_enable_post_write, + },{ .name = "AFMR1", .addr = A_AFMR1, + .pre_write = can_filter_mask_pre_write, + },{ .name = "AFIR1", .addr = A_AFIR1, + .pre_write = can_filter_id_pre_write, + },{ .name = "AFMR2", .addr = A_AFMR2, + .pre_write = can_filter_mask_pre_write, + },{ .name = "AFIR2", .addr = A_AFIR2, + .pre_write = can_filter_id_pre_write, + },{ .name = "AFMR3", .addr = A_AFMR3, + .pre_write = can_filter_mask_pre_write, + },{ .name = "AFIR3", .addr = A_AFIR3, + .pre_write = can_filter_id_pre_write, + },{ .name = "AFMR4", .addr = A_AFMR4, + .pre_write = can_filter_mask_pre_write, + },{ .name = "AFIR4", .addr = A_AFIR4, + .pre_write = can_filter_id_pre_write, + } +}; + +static void xlnx_zynqmp_can_ptimer_cb(void *opaque) +{ + /* No action required on the timer rollover. */ +} + +static const MemoryRegionOps can_ops = { + .read = register_read_memory, + .write = register_write_memory, + .endianness = DEVICE_LITTLE_ENDIAN, + .valid = { + .min_access_size = 4, + .max_access_size = 4, + }, +}; + +static void xlnx_zynqmp_can_reset_init(Object *obj, ResetType type) +{ + XlnxZynqMPCANState *s = XLNX_ZYNQMP_CAN(obj); + unsigned int i; + + for (i = R_RXFIFO_ID; i < ARRAY_SIZE(s->reg_info); ++i) { + register_reset(&s->reg_info[i]); + } + + ptimer_transaction_begin(s->can_timer); + ptimer_set_count(s->can_timer, 0); + ptimer_transaction_commit(s->can_timer); +} + +static void xlnx_zynqmp_can_reset_hold(Object *obj) +{ + XlnxZynqMPCANState *s = XLNX_ZYNQMP_CAN(obj); + unsigned int i; + + for (i = 0; i < R_RXFIFO_ID; ++i) { + register_reset(&s->reg_info[i]); + } + + /* + * Reset FIFOs when CAN model is reset. This will clear the fifo writes + * done by post_write which gets called from register_reset function, + * post_write handle will not be able to trigger tx because CAN will be + * disabled when software_reset_register is cleared first. + */ + fifo32_reset(&s->rx_fifo); + fifo32_reset(&s->tx_fifo); + fifo32_reset(&s->txhpb_fifo); +} + +static bool xlnx_zynqmp_can_can_receive(CanBusClientState *client) +{ + XlnxZynqMPCANState *s = container_of(client, XlnxZynqMPCANState, + bus_client); + + if (ARRAY_FIELD_EX32(s->regs, SOFTWARE_RESET_REGISTER, SRST)) { + DB_PRINT(s, "Controller is in reset.\n"); + return false; + } else if ((ARRAY_FIELD_EX32(s->regs, SOFTWARE_RESET_REGISTER, CEN)) == 0) { + DB_PRINT(s, "Controller is disabled. Incoming messages" + " will be discarded.\n"); + return false; + } else { + return true; + } +} + +static ssize_t xlnx_zynqmp_can_receive(CanBusClientState *client, + const qemu_can_frame *buf, size_t buf_size) { + XlnxZynqMPCANState *s = container_of(client, XlnxZynqMPCANState, + bus_client); + const qemu_can_frame *frame = buf; + + DB_PRINT(s, "Incoming data.\n"); + + if (buf_size <= 0) { + DB_PRINT(s, "Junk data received.\n"); + return 0; + } + if (ARRAY_FIELD_EX32(s->regs, STATUS_REGISTER, LBACK)) { + /* + * XlnxZynqMPCAN will not participate in normal bus communication + * and will not receive any messages transmitted by other CAN nodes. + */ + DB_PRINT(s, "Controller is in loopback mode. It will not" + " receive data.\n"); + + } else if (ARRAY_FIELD_EX32(s->regs, STATUS_REGISTER, SNOOP)) { + /* Snoop Mode: Just keep the data. no response back. */ + update_rx_fifo(s, frame); + } else if ((ARRAY_FIELD_EX32(s->regs, STATUS_REGISTER, SLEEP))) { + /* + * XlnxZynqMPCAN is in sleep mode. Any data on bus will bring it to wake + * up state. + */ + can_exit_sleep_mode(s); + update_rx_fifo(s, frame); + } else if ((ARRAY_FIELD_EX32(s->regs, STATUS_REGISTER, SLEEP)) == 0) { + update_rx_fifo(s, frame); + } else { + DB_PRINT(s, "Cannot receive data as controller is not configured" + " correctly.\n"); + } + + return 1; +} + +static CanBusClientInfo can_xilinx_bus_client_info = { + .can_receive = xlnx_zynqmp_can_can_receive, + .receive = xlnx_zynqmp_can_receive, +}; + +static int xlnx_zynqmp_can_connect_to_bus(XlnxZynqMPCANState *s, + CanBusState *bus) +{ + s->bus_client.info = &can_xilinx_bus_client_info; + + if (can_bus_insert_client(bus, &s->bus_client) < 0) { + return -1; + } + return 0; +} + +static void xlnx_zynqmp_can_realize(DeviceState *dev, Error **errp) +{ + XlnxZynqMPCANState *s = XLNX_ZYNQMP_CAN(dev); + + if (s->canbus) { + if (xlnx_zynqmp_can_connect_to_bus(s, s->canbus) < 0) { + g_autofree char *path = object_get_canonical_path(OBJECT(s)); + + error_setg(errp, "%s: xlnx_zynqmp_can_connect_to_bus" + " failed.", path); + return; + } + + } else { + /* If no bus is set. */ + DB_PRINT(s, "Canbus property is not set.\n"); + } + + /* Create RX FIFO, TXFIFO, TXHPB storage. */ + fifo32_create(&s->rx_fifo, RXFIFO_SIZE); + fifo32_create(&s->tx_fifo, RXFIFO_SIZE); + fifo32_create(&s->txhpb_fifo, CAN_FRAME_SIZE); + + /* Allocate a new timer. */ + s->can_timer = ptimer_init(xlnx_zynqmp_can_ptimer_cb, s, + PTIMER_POLICY_DEFAULT); + + ptimer_transaction_begin(s->can_timer); + + ptimer_set_freq(s->can_timer, s->cfg.ext_clk_freq); + ptimer_set_limit(s->can_timer, CAN_TIMER_MAX, 1); + ptimer_run(s->can_timer, 0); + ptimer_transaction_commit(s->can_timer); +} + +static void xlnx_zynqmp_can_init(Object *obj) +{ + XlnxZynqMPCANState *s = XLNX_ZYNQMP_CAN(obj); + SysBusDevice *sbd = SYS_BUS_DEVICE(obj); + + RegisterInfoArray *reg_array; + + memory_region_init(&s->iomem, obj, TYPE_XLNX_ZYNQMP_CAN, + XLNX_ZYNQMP_CAN_R_MAX * 4); + reg_array = register_init_block32(DEVICE(obj), can_regs_info, + ARRAY_SIZE(can_regs_info), + s->reg_info, s->regs, + &can_ops, + XLNX_ZYNQMP_CAN_ERR_DEBUG, + XLNX_ZYNQMP_CAN_R_MAX * 4); + + memory_region_add_subregion(&s->iomem, 0x00, &reg_array->mem); + sysbus_init_mmio(sbd, &s->iomem); + sysbus_init_irq(SYS_BUS_DEVICE(obj), &s->irq); +} + +static const VMStateDescription vmstate_can = { + .name = TYPE_XLNX_ZYNQMP_CAN, + .version_id = 1, + .minimum_version_id = 1, + .fields = (VMStateField[]) { + VMSTATE_FIFO32(rx_fifo, XlnxZynqMPCANState), + VMSTATE_FIFO32(tx_fifo, XlnxZynqMPCANState), + VMSTATE_FIFO32(txhpb_fifo, XlnxZynqMPCANState), + VMSTATE_UINT32_ARRAY(regs, XlnxZynqMPCANState, XLNX_ZYNQMP_CAN_R_MAX), + VMSTATE_PTIMER(can_timer, XlnxZynqMPCANState), + VMSTATE_END_OF_LIST(), + } +}; + +static Property xlnx_zynqmp_can_properties[] = { + DEFINE_PROP_UINT32("ext_clk_freq", XlnxZynqMPCANState, cfg.ext_clk_freq, + CAN_DEFAULT_CLOCK), + DEFINE_PROP_LINK("canbus", XlnxZynqMPCANState, canbus, TYPE_CAN_BUS, + CanBusState *), + DEFINE_PROP_END_OF_LIST(), +}; + +static void xlnx_zynqmp_can_class_init(ObjectClass *klass, void *data) +{ + DeviceClass *dc = DEVICE_CLASS(klass); + ResettableClass *rc = RESETTABLE_CLASS(klass); + + rc->phases.enter = xlnx_zynqmp_can_reset_init; + rc->phases.hold = xlnx_zynqmp_can_reset_hold; + dc->realize = xlnx_zynqmp_can_realize; + device_class_set_props(dc, xlnx_zynqmp_can_properties); + dc->vmsd = &vmstate_can; +} + +static const TypeInfo can_info = { + .name = TYPE_XLNX_ZYNQMP_CAN, + .parent = TYPE_SYS_BUS_DEVICE, + .instance_size = sizeof(XlnxZynqMPCANState), + .class_init = xlnx_zynqmp_can_class_init, + .instance_init = xlnx_zynqmp_can_init, +}; + +static void can_register_types(void) +{ + type_register_static(&can_info); +} + +type_init(can_register_types) diff --git a/hw/net/can/meson.build b/hw/net/can/meson.build index XXXXXXX..XXXXXXX 100644 --- a/hw/net/can/meson.build +++ b/hw/net/can/meson.build @@ -XXX,XX +XXX,XX @@ softmmu_ss.add(when: 'CONFIG_CAN_SJA1000', if_true: files('can_sja1000.c')) softmmu_ss.add(when: 'CONFIG_CAN_PCI', if_true: files('can_kvaser_pci.c')) softmmu_ss.add(when: 'CONFIG_CAN_PCI', if_true: files('can_pcm3680_pci.c')) softmmu_ss.add(when: 'CONFIG_CAN_PCI', if_true: files('can_mioe3680_pci.c')) +softmmu_ss.add(when: 'CONFIG_XLNX_ZYNQMP', if_true: files('xlnx-zynqmp-can.c')) -- 2.20.1 From: Vikram Garhwal <fnu.vikram@xilinx.com> Connect CAN0 and CAN1 on the ZynqMP. Reviewed-by: Francisco Iglesias <francisco.iglesias@xilinx.com> Reviewed-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com> Signed-off-by: Vikram Garhwal <fnu.vikram@xilinx.com> Message-id: 1597278668-339715-3-git-send-email-fnu.vikram@xilinx.com Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- include/hw/arm/xlnx-zynqmp.h | 8 ++++++++ hw/arm/xlnx-zcu102.c | 20 ++++++++++++++++++++ hw/arm/xlnx-zynqmp.c | 34 ++++++++++++++++++++++++++++++++++ 3 files changed, 62 insertions(+) diff --git a/include/hw/arm/xlnx-zynqmp.h b/include/hw/arm/xlnx-zynqmp.h index XXXXXXX..XXXXXXX 100644 --- a/include/hw/arm/xlnx-zynqmp.h +++ b/include/hw/arm/xlnx-zynqmp.h @@ -XXX,XX +XXX,XX @@ #include "hw/intc/arm_gic.h" #include "hw/net/cadence_gem.h" #include "hw/char/cadence_uart.h" +#include "hw/net/xlnx-zynqmp-can.h" #include "hw/ide/ahci.h" #include "hw/sd/sdhci.h" #include "hw/ssi/xilinx_spips.h" @@ -XXX,XX +XXX,XX @@ #include "hw/cpu/cluster.h" #include "target/arm/cpu.h" #include "qom/object.h" +#include "net/can_emu.h" #define TYPE_XLNX_ZYNQMP "xlnx,zynqmp" typedef struct XlnxZynqMPState XlnxZynqMPState; @@ -XXX,XX +XXX,XX @@ DECLARE_INSTANCE_CHECKER(XlnxZynqMPState, XLNX_ZYNQMP, #define XLNX_ZYNQMP_NUM_RPU_CPUS 2 #define XLNX_ZYNQMP_NUM_GEMS 4 #define XLNX_ZYNQMP_NUM_UARTS 2 +#define XLNX_ZYNQMP_NUM_CAN 2 +#define XLNX_ZYNQMP_CAN_REF_CLK (24 * 1000 * 1000) #define XLNX_ZYNQMP_NUM_SDHCI 2 #define XLNX_ZYNQMP_NUM_SPIS 2 #define XLNX_ZYNQMP_NUM_GDMA_CH 8 @@ -XXX,XX +XXX,XX @@ struct XlnxZynqMPState { CadenceGEMState gem[XLNX_ZYNQMP_NUM_GEMS]; CadenceUARTState uart[XLNX_ZYNQMP_NUM_UARTS]; + XlnxZynqMPCANState can[XLNX_ZYNQMP_NUM_CAN]; SysbusAHCIState sata; SDHCIState sdhci[XLNX_ZYNQMP_NUM_SDHCI]; XilinxSPIPS spi[XLNX_ZYNQMP_NUM_SPIS]; @@ -XXX,XX +XXX,XX @@ struct XlnxZynqMPState { bool virt; /* Has the RPU subsystem? */ bool has_rpu; + + /* CAN bus. */ + CanBusState *canbus[XLNX_ZYNQMP_NUM_CAN]; }; #endif diff --git a/hw/arm/xlnx-zcu102.c b/hw/arm/xlnx-zcu102.c index XXXXXXX..XXXXXXX 100644 --- a/hw/arm/xlnx-zcu102.c +++ b/hw/arm/xlnx-zcu102.c @@ -XXX,XX +XXX,XX @@ #include "sysemu/qtest.h" #include "sysemu/device_tree.h" #include "qom/object.h" +#include "net/can_emu.h" struct XlnxZCU102 { MachineState parent_obj; @@ -XXX,XX +XXX,XX @@ struct XlnxZCU102 { bool secure; bool virt; + CanBusState *canbus[XLNX_ZYNQMP_NUM_CAN]; + struct arm_boot_info binfo; }; typedef struct XlnxZCU102 XlnxZCU102; @@ -XXX,XX +XXX,XX @@ static void xlnx_zcu102_init(MachineState *machine) object_property_set_bool(OBJECT(&s->soc), "virtualization", s->virt, &error_fatal); + for (i = 0; i < XLNX_ZYNQMP_NUM_CAN; i++) { + gchar *bus_name = g_strdup_printf("canbus%d", i); + + object_property_set_link(OBJECT(&s->soc), bus_name, + OBJECT(s->canbus[i]), &error_fatal); + g_free(bus_name); + } + qdev_realize(DEVICE(&s->soc), NULL, &error_fatal); /* Create and plug in the SD cards */ @@ -XXX,XX +XXX,XX @@ static void xlnx_zcu102_machine_instance_init(Object *obj) "Set on/off to enable/disable emulating a " "guest CPU which implements the ARM " "Virtualization Extensions"); + object_property_add_link(obj, "xlnx-zcu102.canbus0", TYPE_CAN_BUS, + (Object **)&s->canbus[0], + object_property_allow_set_link, + 0); + + object_property_add_link(obj, "xlnx-zcu102.canbus1", TYPE_CAN_BUS, + (Object **)&s->canbus[1], + object_property_allow_set_link, + 0); } static void xlnx_zcu102_machine_class_init(ObjectClass *oc, void *data) diff --git a/hw/arm/xlnx-zynqmp.c b/hw/arm/xlnx-zynqmp.c index XXXXXXX..XXXXXXX 100644 --- a/hw/arm/xlnx-zynqmp.c +++ b/hw/arm/xlnx-zynqmp.c @@ -XXX,XX +XXX,XX @@ static const int uart_intr[XLNX_ZYNQMP_NUM_UARTS] = { 21, 22, }; +static const uint64_t can_addr[XLNX_ZYNQMP_NUM_CAN] = { + 0xFF060000, 0xFF070000, +}; + +static const int can_intr[XLNX_ZYNQMP_NUM_CAN] = { + 23, 24, +}; + static const uint64_t sdhci_addr[XLNX_ZYNQMP_NUM_SDHCI] = { 0xFF160000, 0xFF170000, }; @@ -XXX,XX +XXX,XX @@ static void xlnx_zynqmp_init(Object *obj) TYPE_CADENCE_UART); } + for (i = 0; i < XLNX_ZYNQMP_NUM_CAN; i++) { + object_initialize_child(obj, "can[*]", &s->can[i], + TYPE_XLNX_ZYNQMP_CAN); + } + object_initialize_child(obj, "sata", &s->sata, TYPE_SYSBUS_AHCI); for (i = 0; i < XLNX_ZYNQMP_NUM_SDHCI; i++) { @@ -XXX,XX +XXX,XX @@ static void xlnx_zynqmp_realize(DeviceState *dev, Error **errp) gic_spi[uart_intr[i]]); } + for (i = 0; i < XLNX_ZYNQMP_NUM_CAN; i++) { + object_property_set_int(OBJECT(&s->can[i]), "ext_clk_freq", + XLNX_ZYNQMP_CAN_REF_CLK, &error_abort); + + object_property_set_link(OBJECT(&s->can[i]), "canbus", + OBJECT(s->canbus[i]), &error_fatal); + + sysbus_realize(SYS_BUS_DEVICE(&s->can[i]), &err); + if (err) { + error_propagate(errp, err); + return; + } + sysbus_mmio_map(SYS_BUS_DEVICE(&s->can[i]), 0, can_addr[i]); + sysbus_connect_irq(SYS_BUS_DEVICE(&s->can[i]), 0, + gic_spi[can_intr[i]]); + } + object_property_set_int(OBJECT(&s->sata), "num-ports", SATA_NUM_PORTS, &error_abort); if (!sysbus_realize(SYS_BUS_DEVICE(&s->sata), errp)) { @@ -XXX,XX +XXX,XX @@ static Property xlnx_zynqmp_props[] = { DEFINE_PROP_BOOL("has_rpu", XlnxZynqMPState, has_rpu, false), DEFINE_PROP_LINK("ddr-ram", XlnxZynqMPState, ddr_ram, TYPE_MEMORY_REGION, MemoryRegion *), + DEFINE_PROP_LINK("canbus0", XlnxZynqMPState, canbus[0], TYPE_CAN_BUS, + CanBusState *), + DEFINE_PROP_LINK("canbus1", XlnxZynqMPState, canbus[1], TYPE_CAN_BUS, + CanBusState *), DEFINE_PROP_END_OF_LIST() }; -- 2.20.1 From: Vikram Garhwal <fnu.vikram@xilinx.com> The QTests perform five tests on the Xilinx ZynqMP CAN controller: Tests the CAN controller in loopback, sleep and snoop mode. Tests filtering of incoming CAN messages. Reviewed-by: Francisco Iglesias <francisco.iglesias@xilinx.com> Signed-off-by: Vikram Garhwal <fnu.vikram@xilinx.com> Message-id: 1597278668-339715-4-git-send-email-fnu.vikram@xilinx.com [PMM: updated to meson build system] Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- tests/qtest/xlnx-can-test.c | 359 ++++++++++++++++++++++++++++++++++++ tests/qtest/meson.build | 1 + 2 files changed, 360 insertions(+) create mode 100644 tests/qtest/xlnx-can-test.c diff --git a/tests/qtest/xlnx-can-test.c b/tests/qtest/xlnx-can-test.c new file mode 100644 index XXXXXXX..XXXXXXX --- /dev/null +++ b/tests/qtest/xlnx-can-test.c @@ -XXX,XX +XXX,XX @@ +/* + * QTests for the Xilinx ZynqMP CAN controller. + * + * Copyright (c) 2020 Xilinx Inc. + * + * Written-by: Vikram Garhwal<fnu.vikram@xilinx.com> + * + * Permission is hereby granted, free of charge, to any person obtaining a copy + * of this software and associated documentation files (the "Software"), to deal + * in the Software without restriction, including without limitation the rights + * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell + * copies of the Software, and to permit persons to whom the Software is + * furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in + * all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER + * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, + * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN + * THE SOFTWARE. + */ + +#include "qemu/osdep.h" +#include "libqos/libqtest.h" + +/* Base address. */ +#define CAN0_BASE_ADDR 0xFF060000 +#define CAN1_BASE_ADDR 0xFF070000 + +/* Register addresses. */ +#define R_SRR_OFFSET 0x00 +#define R_MSR_OFFSET 0x04 +#define R_SR_OFFSET 0x18 +#define R_ISR_OFFSET 0x1C +#define R_ICR_OFFSET 0x24 +#define R_TXID_OFFSET 0x30 +#define R_TXDLC_OFFSET 0x34 +#define R_TXDATA1_OFFSET 0x38 +#define R_TXDATA2_OFFSET 0x3C +#define R_RXID_OFFSET 0x50 +#define R_RXDLC_OFFSET 0x54 +#define R_RXDATA1_OFFSET 0x58 +#define R_RXDATA2_OFFSET 0x5C +#define R_AFR 0x60 +#define R_AFMR1 0x64 +#define R_AFIR1 0x68 +#define R_AFMR2 0x6C +#define R_AFIR2 0x70 +#define R_AFMR3 0x74 +#define R_AFIR3 0x78 +#define R_AFMR4 0x7C +#define R_AFIR4 0x80 + +/* CAN modes. */ +#define CONFIG_MODE 0x00 +#define NORMAL_MODE 0x00 +#define LOOPBACK_MODE 0x02 +#define SNOOP_MODE 0x04 +#define SLEEP_MODE 0x01 +#define ENABLE_CAN (1 << 1) +#define STATUS_NORMAL_MODE (1 << 3) +#define STATUS_LOOPBACK_MODE (1 << 1) +#define STATUS_SNOOP_MODE (1 << 12) +#define STATUS_SLEEP_MODE (1 << 2) +#define ISR_TXOK (1 << 1) +#define ISR_RXOK (1 << 4) + +static void match_rx_tx_data(uint32_t *buf_tx, uint32_t *buf_rx, + uint8_t can_timestamp) +{ + uint16_t size = 0; + uint8_t len = 4; + + while (size < len) { + if (R_RXID_OFFSET + 4 * size == R_RXDLC_OFFSET) { + g_assert_cmpint(buf_rx[size], ==, buf_tx[size] + can_timestamp); + } else { + g_assert_cmpint(buf_rx[size], ==, buf_tx[size]); + } + + size++; + } +} + +static void read_data(QTestState *qts, uint64_t can_base_addr, uint32_t *buf_rx) +{ + uint32_t int_status; + + /* Read the interrupt on CAN rx. */ + int_status = qtest_readl(qts, can_base_addr + R_ISR_OFFSET) & ISR_RXOK; + + g_assert_cmpint(int_status, ==, ISR_RXOK); + + /* Read the RX register data for CAN. */ + buf_rx[0] = qtest_readl(qts, can_base_addr + R_RXID_OFFSET); + buf_rx[1] = qtest_readl(qts, can_base_addr + R_RXDLC_OFFSET); + buf_rx[2] = qtest_readl(qts, can_base_addr + R_RXDATA1_OFFSET); + buf_rx[3] = qtest_readl(qts, can_base_addr + R_RXDATA2_OFFSET); + + /* Clear the RX interrupt. */ + qtest_writel(qts, CAN1_BASE_ADDR + R_ICR_OFFSET, ISR_RXOK); +} + +static void send_data(QTestState *qts, uint64_t can_base_addr, uint32_t *buf_tx) +{ + uint32_t int_status; + + /* Write the TX register data for CAN. */ + qtest_writel(qts, can_base_addr + R_TXID_OFFSET, buf_tx[0]); + qtest_writel(qts, can_base_addr + R_TXDLC_OFFSET, buf_tx[1]); + qtest_writel(qts, can_base_addr + R_TXDATA1_OFFSET, buf_tx[2]); + qtest_writel(qts, can_base_addr + R_TXDATA2_OFFSET, buf_tx[3]); + + /* Read the interrupt on CAN for tx. */ + int_status = qtest_readl(qts, can_base_addr + R_ISR_OFFSET) & ISR_TXOK; + + g_assert_cmpint(int_status, ==, ISR_TXOK); + + /* Clear the interrupt for tx. */ + qtest_writel(qts, CAN0_BASE_ADDR + R_ICR_OFFSET, ISR_TXOK); +} + +/* + * This test will be transferring data from CAN0 and CAN1 through canbus. CAN0 + * initiate the data transfer to can-bus, CAN1 receives the data. Test compares + * the data sent from CAN0 with received on CAN1. + */ +static void test_can_bus(void) +{ + uint32_t buf_tx[4] = { 0xFF, 0x80000000, 0x12345678, 0x87654321 }; + uint32_t buf_rx[4] = { 0x00, 0x00, 0x00, 0x00 }; + uint32_t status = 0; + uint8_t can_timestamp = 1; + + QTestState *qts = qtest_init("-machine xlnx-zcu102" + " -object can-bus,id=canbus0" + " -machine xlnx-zcu102.canbus0=canbus0" + " -machine xlnx-zcu102.canbus1=canbus0" + ); + + /* Configure the CAN0 and CAN1. */ + qtest_writel(qts, CAN0_BASE_ADDR + R_SRR_OFFSET, ENABLE_CAN); + qtest_writel(qts, CAN0_BASE_ADDR + R_MSR_OFFSET, NORMAL_MODE); + qtest_writel(qts, CAN1_BASE_ADDR + R_SRR_OFFSET, ENABLE_CAN); + qtest_writel(qts, CAN1_BASE_ADDR + R_MSR_OFFSET, NORMAL_MODE); + + /* Check here if CAN0 and CAN1 are in normal mode. */ + status = qtest_readl(qts, CAN0_BASE_ADDR + R_SR_OFFSET); + g_assert_cmpint(status, ==, STATUS_NORMAL_MODE); + + status = qtest_readl(qts, CAN1_BASE_ADDR + R_SR_OFFSET); + g_assert_cmpint(status, ==, STATUS_NORMAL_MODE); + + send_data(qts, CAN0_BASE_ADDR, buf_tx); + + read_data(qts, CAN1_BASE_ADDR, buf_rx); + match_rx_tx_data(buf_tx, buf_rx, can_timestamp); + + qtest_quit(qts); +} + +/* + * This test is performing loopback mode on CAN0 and CAN1. Data sent from TX of + * each CAN0 and CAN1 are compared with RX register data for respective CAN. + */ +static void test_can_loopback(void) +{ + uint32_t buf_tx[4] = { 0xFF, 0x80000000, 0x12345678, 0x87654321 }; + uint32_t buf_rx[4] = { 0x00, 0x00, 0x00, 0x00 }; + uint32_t status = 0; + + QTestState *qts = qtest_init("-machine xlnx-zcu102" + " -object can-bus,id=canbus0" + " -machine xlnx-zcu102.canbus0=canbus0" + " -machine xlnx-zcu102.canbus1=canbus0" + ); + + /* Configure the CAN0 in loopback mode. */ + qtest_writel(qts, CAN0_BASE_ADDR + R_SRR_OFFSET, CONFIG_MODE); + qtest_writel(qts, CAN0_BASE_ADDR + R_MSR_OFFSET, LOOPBACK_MODE); + qtest_writel(qts, CAN0_BASE_ADDR + R_SRR_OFFSET, ENABLE_CAN); + + /* Check here if CAN0 is set in loopback mode. */ + status = qtest_readl(qts, CAN0_BASE_ADDR + R_SR_OFFSET); + + g_assert_cmpint(status, ==, STATUS_LOOPBACK_MODE); + + send_data(qts, CAN0_BASE_ADDR, buf_tx); + read_data(qts, CAN0_BASE_ADDR, buf_rx); + match_rx_tx_data(buf_tx, buf_rx, 0); + + /* Configure the CAN1 in loopback mode. */ + qtest_writel(qts, CAN1_BASE_ADDR + R_SRR_OFFSET, CONFIG_MODE); + qtest_writel(qts, CAN1_BASE_ADDR + R_MSR_OFFSET, LOOPBACK_MODE); + qtest_writel(qts, CAN1_BASE_ADDR + R_SRR_OFFSET, ENABLE_CAN); + + /* Check here if CAN1 is set in loopback mode. */ + status = qtest_readl(qts, CAN1_BASE_ADDR + R_SR_OFFSET); + + g_assert_cmpint(status, ==, STATUS_LOOPBACK_MODE); + + send_data(qts, CAN1_BASE_ADDR, buf_tx); + read_data(qts, CAN1_BASE_ADDR, buf_rx); + match_rx_tx_data(buf_tx, buf_rx, 0); + + qtest_quit(qts); +} + +/* + * Enable filters for CAN1. This will filter incoming messages with ID. In this + * test message will pass through filter 2. + */ +static void test_can_filter(void) +{ + uint32_t buf_tx[4] = { 0x14, 0x80000000, 0x12345678, 0x87654321 }; + uint32_t buf_rx[4] = { 0x00, 0x00, 0x00, 0x00 }; + uint32_t status = 0; + uint8_t can_timestamp = 1; + + QTestState *qts = qtest_init("-machine xlnx-zcu102" + " -object can-bus,id=canbus0" + " -machine xlnx-zcu102.canbus0=canbus0" + " -machine xlnx-zcu102.canbus1=canbus0" + ); + + /* Configure the CAN0 and CAN1. */ + qtest_writel(qts, CAN0_BASE_ADDR + R_SRR_OFFSET, ENABLE_CAN); + qtest_writel(qts, CAN0_BASE_ADDR + R_MSR_OFFSET, NORMAL_MODE); + qtest_writel(qts, CAN1_BASE_ADDR + R_SRR_OFFSET, ENABLE_CAN); + qtest_writel(qts, CAN1_BASE_ADDR + R_MSR_OFFSET, NORMAL_MODE); + + /* Check here if CAN0 and CAN1 are in normal mode. */ + status = qtest_readl(qts, CAN0_BASE_ADDR + R_SR_OFFSET); + g_assert_cmpint(status, ==, STATUS_NORMAL_MODE); + + status = qtest_readl(qts, CAN1_BASE_ADDR + R_SR_OFFSET); + g_assert_cmpint(status, ==, STATUS_NORMAL_MODE); + + /* Set filter for CAN1 for incoming messages. */ + qtest_writel(qts, CAN1_BASE_ADDR + R_AFR, 0x0); + qtest_writel(qts, CAN1_BASE_ADDR + R_AFMR1, 0xF7); + qtest_writel(qts, CAN1_BASE_ADDR + R_AFIR1, 0x121F); + qtest_writel(qts, CAN1_BASE_ADDR + R_AFMR2, 0x5431); + qtest_writel(qts, CAN1_BASE_ADDR + R_AFIR2, 0x14); + qtest_writel(qts, CAN1_BASE_ADDR + R_AFMR3, 0x1234); + qtest_writel(qts, CAN1_BASE_ADDR + R_AFIR3, 0x5431); + qtest_writel(qts, CAN1_BASE_ADDR + R_AFMR4, 0xFFF); + qtest_writel(qts, CAN1_BASE_ADDR + R_AFIR4, 0x1234); + + qtest_writel(qts, CAN1_BASE_ADDR + R_AFR, 0xF); + + send_data(qts, CAN0_BASE_ADDR, buf_tx); + + read_data(qts, CAN1_BASE_ADDR, buf_rx); + match_rx_tx_data(buf_tx, buf_rx, can_timestamp); + + qtest_quit(qts); +} + +/* Testing sleep mode on CAN0 while CAN1 is in normal mode. */ +static void test_can_sleepmode(void) +{ + uint32_t buf_tx[4] = { 0x14, 0x80000000, 0x12345678, 0x87654321 }; + uint32_t buf_rx[4] = { 0x00, 0x00, 0x00, 0x00 }; + uint32_t status = 0; + uint8_t can_timestamp = 1; + + QTestState *qts = qtest_init("-machine xlnx-zcu102" + " -object can-bus,id=canbus0" + " -machine xlnx-zcu102.canbus0=canbus0" + " -machine xlnx-zcu102.canbus1=canbus0" + ); + + /* Configure the CAN0. */ + qtest_writel(qts, CAN0_BASE_ADDR + R_SRR_OFFSET, CONFIG_MODE); + qtest_writel(qts, CAN0_BASE_ADDR + R_MSR_OFFSET, SLEEP_MODE); + qtest_writel(qts, CAN0_BASE_ADDR + R_SRR_OFFSET, ENABLE_CAN); + + qtest_writel(qts, CAN1_BASE_ADDR + R_SRR_OFFSET, ENABLE_CAN); + qtest_writel(qts, CAN1_BASE_ADDR + R_MSR_OFFSET, NORMAL_MODE); + + /* Check here if CAN0 is in SLEEP mode and CAN1 in normal mode. */ + status = qtest_readl(qts, CAN0_BASE_ADDR + R_SR_OFFSET); + g_assert_cmpint(status, ==, STATUS_SLEEP_MODE); + + status = qtest_readl(qts, CAN1_BASE_ADDR + R_SR_OFFSET); + g_assert_cmpint(status, ==, STATUS_NORMAL_MODE); + + send_data(qts, CAN1_BASE_ADDR, buf_tx); + + /* + * Once CAN1 sends data on can-bus. CAN0 should exit sleep mode. + * Check the CAN0 status now. It should exit the sleep mode and receive the + * incoming data. + */ + status = qtest_readl(qts, CAN0_BASE_ADDR + R_SR_OFFSET); + g_assert_cmpint(status, ==, STATUS_NORMAL_MODE); + + read_data(qts, CAN0_BASE_ADDR, buf_rx); + + match_rx_tx_data(buf_tx, buf_rx, can_timestamp); + + qtest_quit(qts); +} + +/* Testing Snoop mode on CAN0 while CAN1 is in normal mode. */ +static void test_can_snoopmode(void) +{ + uint32_t buf_tx[4] = { 0x14, 0x80000000, 0x12345678, 0x87654321 }; + uint32_t buf_rx[4] = { 0x00, 0x00, 0x00, 0x00 }; + uint32_t status = 0; + uint8_t can_timestamp = 1; + + QTestState *qts = qtest_init("-machine xlnx-zcu102" + " -object can-bus,id=canbus0" + " -machine xlnx-zcu102.canbus0=canbus0" + " -machine xlnx-zcu102.canbus1=canbus0" + ); + + /* Configure the CAN0. */ + qtest_writel(qts, CAN0_BASE_ADDR + R_SRR_OFFSET, CONFIG_MODE); + qtest_writel(qts, CAN0_BASE_ADDR + R_MSR_OFFSET, SNOOP_MODE); + qtest_writel(qts, CAN0_BASE_ADDR + R_SRR_OFFSET, ENABLE_CAN); + + qtest_writel(qts, CAN1_BASE_ADDR + R_SRR_OFFSET, ENABLE_CAN); + qtest_writel(qts, CAN1_BASE_ADDR + R_MSR_OFFSET, NORMAL_MODE); + + /* Check here if CAN0 is in SNOOP mode and CAN1 in normal mode. */ + status = qtest_readl(qts, CAN0_BASE_ADDR + R_SR_OFFSET); + g_assert_cmpint(status, ==, STATUS_SNOOP_MODE); + + status = qtest_readl(qts, CAN1_BASE_ADDR + R_SR_OFFSET); + g_assert_cmpint(status, ==, STATUS_NORMAL_MODE); + + send_data(qts, CAN1_BASE_ADDR, buf_tx); + + read_data(qts, CAN0_BASE_ADDR, buf_rx); + + match_rx_tx_data(buf_tx, buf_rx, can_timestamp); + + qtest_quit(qts); +} + +int main(int argc, char **argv) +{ + g_test_init(&argc, &argv, NULL); + + qtest_add_func("/net/can/can_bus", test_can_bus); + qtest_add_func("/net/can/can_loopback", test_can_loopback); + qtest_add_func("/net/can/can_filter", test_can_filter); + qtest_add_func("/net/can/can_test_snoopmode", test_can_snoopmode); + qtest_add_func("/net/can/can_test_sleepmode", test_can_sleepmode); + + return g_test_run(); +} diff --git a/tests/qtest/meson.build b/tests/qtest/meson.build index XXXXXXX..XXXXXXX 100644 --- a/tests/qtest/meson.build +++ b/tests/qtest/meson.build @@ -XXX,XX +XXX,XX @@ qtests_aarch64 = \ (config_all_devices.has_key('CONFIG_TPM_TIS_SYSBUS') ? ['tpm-tis-device-swtpm-test'] : []) + \ ['numa-test', 'boot-serial-test', + 'xlnx-can-test', 'migration-test'] qtests_s390x = \ -- 2.20.1 From: Vikram Garhwal <fnu.vikram@xilinx.com> Reviewed-by: Francisco Iglesias <francisco.iglesias@xilinx.com> Reviewed-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com> Signed-off-by: Vikram Garhwal <fnu.vikram@xilinx.com> Message-id: 1597278668-339715-5-git-send-email-fnu.vikram@xilinx.com Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- MAINTAINERS | 8 ++++++++ 1 file changed, 8 insertions(+) diff --git a/MAINTAINERS b/MAINTAINERS index XXXXXXX..XXXXXXX 100644 --- a/MAINTAINERS +++ b/MAINTAINERS @@ -XXX,XX +XXX,XX @@ F: hw/net/opencores_eth.c Devices ------- +Xilinx CAN +M: Vikram Garhwal <fnu.vikram@xilinx.com> +M: Francisco Iglesias <francisco.iglesias@xilinx.com> +S: Maintained +F: hw/net/can/xlnx-* +F: include/hw/net/xlnx-* +F: tests/qtest/xlnx-can-test* + EDU M: Jiri Slaby <jslaby@suse.cz> S: Maintained -- 2.20.1 I don't have anything else queued up at the moment, so this is just Richard's SME patches. -- PMM The following changes since commit 63b38f6c85acd312c2cab68554abf33adf4ee2b3: Merge tag 'pull-target-arm-20220707' of https://git.linaro.org/people/pmaydell/qemu-arm into staging (2022-07-08 06:17:11 +0530) are available in the Git repository at: https://git.linaro.org/people/pmaydell/qemu-arm.git tags/pull-target-arm-20220711 for you to fetch changes up to f9982ceaf26df27d15547a3a7990a95019e9e3a8: linux-user/aarch64: Add SME related hwcap entries (2022-07-11 13:43:52 +0100) ---------------------------------------------------------------- target-arm: * Implement SME emulation, for both system and linux-user ---------------------------------------------------------------- Richard Henderson (45): target/arm: Handle SME in aarch64_cpu_dump_state target/arm: Add infrastructure for disas_sme target/arm: Trap non-streaming usage when Streaming SVE is active target/arm: Mark ADR as non-streaming target/arm: Mark RDFFR, WRFFR, SETFFR as non-streaming target/arm: Mark BDEP, BEXT, BGRP, COMPACT, FEXPA, FTSSEL as non-streaming target/arm: Mark PMULL, FMMLA as non-streaming target/arm: Mark FTSMUL, FTMAD, FADDA as non-streaming target/arm: Mark SMMLA, UMMLA, USMMLA as non-streaming target/arm: Mark string/histo/crypto as non-streaming target/arm: Mark gather/scatter load/store as non-streaming target/arm: Mark gather prefetch as non-streaming target/arm: Mark LDFF1 and LDNF1 as non-streaming target/arm: Mark LD1RO as non-streaming target/arm: Add SME enablement checks target/arm: Handle SME in sve_access_check target/arm: Implement SME RDSVL, ADDSVL, ADDSPL target/arm: Implement SME ZERO target/arm: Implement SME MOVA target/arm: Implement SME LD1, ST1 target/arm: Export unpredicated ld/st from translate-sve.c target/arm: Implement SME LDR, STR target/arm: Implement SME ADDHA, ADDVA target/arm: Implement FMOPA, FMOPS (non-widening) target/arm: Implement BFMOPA, BFMOPS target/arm: Implement FMOPA, FMOPS (widening) target/arm: Implement SME integer outer product target/arm: Implement PSEL target/arm: Implement REVD target/arm: Implement SCLAMP, UCLAMP target/arm: Reset streaming sve state on exception boundaries target/arm: Enable SME for -cpu max linux-user/aarch64: Clear tpidr2_el0 if CLONE_SETTLS linux-user/aarch64: Reset PSTATE.SM on syscalls linux-user/aarch64: Add SM bit to SVE signal context linux-user/aarch64: Tidy target_restore_sigframe error return linux-user/aarch64: Do not allow duplicate or short sve records linux-user/aarch64: Verify extra record lock succeeded linux-user/aarch64: Move sve record checks into restore linux-user/aarch64: Implement SME signal handling linux-user: Rename sve prctls linux-user/aarch64: Implement PR_SME_GET_VL, PR_SME_SET_VL target/arm: Only set ZEN in reset if SVE present target/arm: Enable SME for user-only linux-user/aarch64: Add SME related hwcap entries docs/system/arm/emulation.rst | 4 + linux-user/aarch64/target_cpu.h | 5 +- linux-user/aarch64/target_prctl.h | 62 +- target/arm/cpu.h | 7 + target/arm/helper-sme.h | 126 ++++ target/arm/helper-sve.h | 4 + target/arm/helper.h | 18 + target/arm/translate-a64.h | 45 ++ target/arm/translate.h | 16 + target/arm/sme-fa64.decode | 60 ++ target/arm/sme.decode | 88 +++ target/arm/sve.decode | 41 +- linux-user/aarch64/cpu_loop.c | 9 + linux-user/aarch64/signal.c | 243 ++++++-- linux-user/elfload.c | 20 + linux-user/syscall.c | 28 +- target/arm/cpu.c | 35 +- target/arm/cpu64.c | 11 + target/arm/helper.c | 56 +- target/arm/sme_helper.c | 1140 +++++++++++++++++++++++++++++++++++++ target/arm/sve_helper.c | 28 + target/arm/translate-a64.c | 103 +++- target/arm/translate-sme.c | 373 ++++++++++++ target/arm/translate-sve.c | 393 ++++++++++--- target/arm/translate-vfp.c | 12 + target/arm/translate.c | 2 + target/arm/vec_helper.c | 24 + target/arm/meson.build | 3 + 28 files changed, 2821 insertions(+), 135 deletions(-) create mode 100644 target/arm/sme-fa64.decode create mode 100644 target/arm/sme.decode create mode 100644 target/arm/translate-sme.c From: Richard Henderson <richard.henderson@linaro.org> Dump SVCR, plus use the correct access check for Streaming Mode. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20220708151540.18136-2-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/cpu.c | 17 ++++++++++++++++- 1 file changed, 16 insertions(+), 1 deletion(-) 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 aarch64_cpu_dump_state(CPUState *cs, FILE *f, int flags) int i; int el = arm_current_el(env); const char *ns_status; + bool sve; qemu_fprintf(f, " PC=%016" PRIx64 " ", env->pc); for (i = 0; i < 32; i++) { @@ -XXX,XX +XXX,XX @@ static void aarch64_cpu_dump_state(CPUState *cs, FILE *f, int flags) el, psr & PSTATE_SP ? 'h' : 't'); + if (cpu_isar_feature(aa64_sme, cpu)) { + qemu_fprintf(f, " SVCR=%08" PRIx64 " %c%c", + env->svcr, + (FIELD_EX64(env->svcr, SVCR, ZA) ? 'Z' : '-'), + (FIELD_EX64(env->svcr, SVCR, SM) ? 'S' : '-')); + } if (cpu_isar_feature(aa64_bti, cpu)) { qemu_fprintf(f, " BTYPE=%d", (psr & PSTATE_BTYPE) >> 10); } @@ -XXX,XX +XXX,XX @@ static void aarch64_cpu_dump_state(CPUState *cs, FILE *f, int flags) qemu_fprintf(f, " FPCR=%08x FPSR=%08x\n", vfp_get_fpcr(env), vfp_get_fpsr(env)); - if (cpu_isar_feature(aa64_sve, cpu) && sve_exception_el(env, el) == 0) { + if (cpu_isar_feature(aa64_sme, cpu) && FIELD_EX64(env->svcr, SVCR, SM)) { + sve = sme_exception_el(env, el) == 0; + } else if (cpu_isar_feature(aa64_sve, cpu)) { + sve = sve_exception_el(env, el) == 0; + } else { + sve = false; + } + + if (sve) { int j, zcr_len = sve_vqm1_for_el(env, el); for (i = 0; i <= FFR_PRED_NUM; i++) { -- 2.25.1 From: Richard Henderson <richard.henderson@linaro.org> This includes the build rules for the decoder, and the new file for translation, but excludes any instructions. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20220708151540.18136-3-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/translate-a64.h | 1 + target/arm/sme.decode | 20 ++++++++++++++++++++ target/arm/translate-a64.c | 7 ++++++- target/arm/translate-sme.c | 35 +++++++++++++++++++++++++++++++++++ target/arm/meson.build | 2 ++ 5 files changed, 64 insertions(+), 1 deletion(-) create mode 100644 target/arm/sme.decode create mode 100644 target/arm/translate-sme.c 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 pred_gvec_reg_size(DisasContext *s) } bool disas_sve(DisasContext *, uint32_t); +bool disas_sme(DisasContext *, uint32_t); void gen_gvec_rax1(unsigned vece, uint32_t rd_ofs, uint32_t rn_ofs, uint32_t rm_ofs, uint32_t opr_sz, uint32_t max_sz); diff --git a/target/arm/sme.decode b/target/arm/sme.decode new file mode 100644 index XXXXXXX..XXXXXXX --- /dev/null +++ b/target/arm/sme.decode @@ -XXX,XX +XXX,XX @@ +# AArch64 SME instruction descriptions +# +# 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/>. + +# +# This file is processed by scripts/decodetree.py +# 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_translate_insn(DisasContextBase *dcbase, CPUState *cpu) } switch (extract32(insn, 25, 4)) { - case 0x0: case 0x1: case 0x3: /* UNALLOCATED */ + case 0x0: + if (!extract32(insn, 31, 1) || !disas_sme(s, insn)) { + unallocated_encoding(s); + } + break; + case 0x1: case 0x3: /* UNALLOCATED */ unallocated_encoding(s); break; case 0x2: diff --git a/target/arm/translate-sme.c b/target/arm/translate-sme.c new file mode 100644 index XXXXXXX..XXXXXXX --- /dev/null +++ b/target/arm/translate-sme.c @@ -XXX,XX +XXX,XX @@ +/* + * AArch64 SME translation + * + * 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 "tcg/tcg-op.h" +#include "tcg/tcg-op-gvec.h" +#include "tcg/tcg-gvec-desc.h" +#include "translate.h" +#include "exec/helper-gen.h" +#include "translate-a64.h" +#include "fpu/softfloat.h" + + +/* + * Include the generated decoder. + */ + +#include "decode-sme.c.inc" 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 @@ gen = [ decodetree.process('sve.decode', extra_args: '--decode=disas_sve'), + decodetree.process('sme.decode', extra_args: '--decode=disas_sme'), decodetree.process('neon-shared.decode', extra_args: '--decode=disas_neon_shared'), decodetree.process('neon-dp.decode', extra_args: '--decode=disas_neon_dp'), decodetree.process('neon-ls.decode', extra_args: '--decode=disas_neon_ls'), @@ -XXX,XX +XXX,XX @@ arm_ss.add(when: 'TARGET_AARCH64', if_true: files( 'sme_helper.c', 'translate-a64.c', 'translate-sve.c', + 'translate-sme.c', )) arm_softmmu_ss = ss.source_set() -- 2.25.1 From: Richard Henderson <richard.henderson@linaro.org> This new behaviour is in the ARM pseudocode function AArch64.CheckFPAdvSIMDEnabled, which applies to AArch32 via AArch32.CheckAdvSIMDOrFPEnabled when the EL to which the trap would be delivered is in AArch64 mode. Given that ARMv9 drops support for AArch32 outside EL0, the trap EL detection ought to be trivially true, but the pseudocode still contains a number of conditions, and QEMU has not yet committed to dropping A32 support for EL[12] when v9 features are present. Since the computation of SME_TRAP_NONSTREAMING is necessarily different for the two modes, we might as well preserve bits within TBFLAG_ANY and allocate separate bits within TBFLAG_A32 and TBFLAG_A64 instead. Note that DDI0616A.a has typos for bits [22:21] of LD1RO in the table of instructions illegal in streaming mode. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20220708151540.18136-4-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/cpu.h | 7 +++ target/arm/translate.h | 4 ++ target/arm/sme-fa64.decode | 90 ++++++++++++++++++++++++++++++++++++++ target/arm/helper.c | 41 +++++++++++++++++ target/arm/translate-a64.c | 40 ++++++++++++++++- target/arm/translate-vfp.c | 12 +++++ target/arm/translate.c | 2 + target/arm/meson.build | 1 + 8 files changed, 195 insertions(+), 2 deletions(-) create mode 100644 target/arm/sme-fa64.decode 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_A32, HSTR_ACTIVE, 9, 1) * the same thing as the current security state of the processor! */ FIELD(TBFLAG_A32, NS, 10, 1) +/* + * Indicates that SME Streaming mode is active, and SMCR_ELx.FA64 is not. + * This requires an SME trap from AArch32 mode when using NEON. + */ +FIELD(TBFLAG_A32, SME_TRAP_NONSTREAMING, 11, 1) /* * Bit usage when in AArch32 state, for M-profile only. @@ -XXX,XX +XXX,XX @@ 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) +/* Indicates that SME Streaming mode is active, and SMCR_ELx.FA64 is not. */ +FIELD(TBFLAG_A64, SME_TRAP_NONSTREAMING, 28, 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 pstate_sm; /* True if PSTATE.ZA is set. */ bool pstate_za; + /* True if non-streaming insns should raise an SME Streaming exception. */ + bool sme_trap_nonstreaming; + /* True if the current instruction is non-streaming. */ + bool is_nonstreaming; /* True if MVE insns are definitely not predicated by VPR or LTPSIZE */ bool mve_no_pred; /* diff --git a/target/arm/sme-fa64.decode b/target/arm/sme-fa64.decode new file mode 100644 index XXXXXXX..XXXXXXX --- /dev/null +++ b/target/arm/sme-fa64.decode @@ -XXX,XX +XXX,XX @@ +# AArch64 SME allowed instruction decoding +# +# 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/>. + +# +# This file is processed by scripts/decodetree.py +# + +# These patterns are taken from Appendix E1.1 of DDI0616 A.a, +# Arm Architecture Reference Manual Supplement, +# The Scalable Matrix Extension (SME), for Armv9-A + +{ + [ + OK 0-00 1110 0000 0001 0010 11-- ---- ---- # SMOV W|Xd,Vn.B[0] + OK 0-00 1110 0000 0010 0010 11-- ---- ---- # SMOV W|Xd,Vn.H[0] + OK 0100 1110 0000 0100 0010 11-- ---- ---- # SMOV Xd,Vn.S[0] + OK 0000 1110 0000 0001 0011 11-- ---- ---- # UMOV Wd,Vn.B[0] + OK 0000 1110 0000 0010 0011 11-- ---- ---- # UMOV Wd,Vn.H[0] + OK 0000 1110 0000 0100 0011 11-- ---- ---- # UMOV Wd,Vn.S[0] + OK 0100 1110 0000 1000 0011 11-- ---- ---- # UMOV Xd,Vn.D[0] + ] + FAIL 0--0 111- ---- ---- ---- ---- ---- ---- # Advanced SIMD vector operations +} + +{ + [ + OK 0101 1110 --1- ---- 11-1 11-- ---- ---- # FMULX/FRECPS/FRSQRTS (scalar) + OK 0101 1110 -10- ---- 00-1 11-- ---- ---- # FMULX/FRECPS/FRSQRTS (scalar, FP16) + OK 01-1 1110 1-10 0001 11-1 10-- ---- ---- # FRECPE/FRSQRTE/FRECPX (scalar) + OK 01-1 1110 1111 1001 11-1 10-- ---- ---- # FRECPE/FRSQRTE/FRECPX (scalar, FP16) + ] + FAIL 01-1 111- ---- ---- ---- ---- ---- ---- # Advanced SIMD single-element operations +} + +FAIL 0-00 110- ---- ---- ---- ---- ---- ---- # Advanced SIMD structure load/store +FAIL 1100 1110 ---- ---- ---- ---- ---- ---- # Advanced SIMD cryptography extensions +FAIL 0001 1110 0111 1110 0000 00-- ---- ---- # FJCVTZS + +# These are the "avoidance of doubt" final table of Illegal Advanced SIMD instructions +# We don't actually need to include these, as the default is OK. +# -001 111- ---- ---- ---- ---- ---- ---- # Scalar floating-point operations +# --10 110- ---- ---- ---- ---- ---- ---- # Load/store pair of FP registers +# --01 1100 ---- ---- ---- ---- ---- ---- # Load FP register (PC-relative literal) +# --11 1100 --0- ---- ---- ---- ---- ---- # Load/store FP register (unscaled imm) +# --11 1100 --1- ---- ---- ---- ---- --10 # Load/store FP register (register offset) +# --11 1101 ---- ---- ---- ---- ---- ---- # Load/store FP register (scaled imm) + +FAIL 0000 0100 --1- ---- 1010 ---- ---- ---- # ADR +FAIL 0000 0100 --1- ---- 1011 -0-- ---- ---- # FTSSEL, FEXPA +FAIL 0000 0101 --10 0001 100- ---- ---- ---- # COMPACT +FAIL 0010 0101 --01 100- 1111 000- ---0 ---- # RDFFR, RDFFRS +FAIL 0010 0101 --10 1--- 1001 ---- ---- ---- # WRFFR, SETFFR +FAIL 0100 0101 --0- ---- 1011 ---- ---- ---- # BDEP, BEXT, BGRP +FAIL 0100 0101 000- ---- 0110 1--- ---- ---- # PMULLB, PMULLT (128b result) +FAIL 0110 0100 --1- ---- 1110 01-- ---- ---- # FMMLA, BFMMLA +FAIL 0110 0101 --0- ---- 0000 11-- ---- ---- # FTSMUL +FAIL 0110 0101 --01 0--- 100- ---- ---- ---- # FTMAD +FAIL 0110 0101 --01 1--- 001- ---- ---- ---- # FADDA +FAIL 0100 0101 --0- ---- 1001 10-- ---- ---- # SMMLA, UMMLA, USMMLA +FAIL 0100 0101 --1- ---- 1--- ---- ---- ---- # SVE2 string/histo/crypto instructions +FAIL 1000 010- -00- ---- 10-- ---- ---- ---- # SVE2 32-bit gather NT load (vector+scalar) +FAIL 1000 010- -00- ---- 111- ---- ---- ---- # SVE 32-bit gather prefetch (vector+imm) +FAIL 1000 0100 0-1- ---- 0--- ---- ---- ---- # SVE 32-bit gather prefetch (scalar+vector) +FAIL 1000 010- -01- ---- 1--- ---- ---- ---- # SVE 32-bit gather load (vector+imm) +FAIL 1000 0100 0-0- ---- 0--- ---- ---- ---- # SVE 32-bit gather load byte (scalar+vector) +FAIL 1000 0100 1--- ---- 0--- ---- ---- ---- # SVE 32-bit gather load half (scalar+vector) +FAIL 1000 0101 0--- ---- 0--- ---- ---- ---- # SVE 32-bit gather load word (scalar+vector) +FAIL 1010 010- ---- ---- 011- ---- ---- ---- # SVE contiguous FF load (scalar+scalar) +FAIL 1010 010- ---1 ---- 101- ---- ---- ---- # SVE contiguous NF load (scalar+imm) +FAIL 1010 010- -01- ---- 000- ---- ---- ---- # SVE load & replicate 32 bytes (scalar+scalar) +FAIL 1010 010- -010 ---- 001- ---- ---- ---- # SVE load & replicate 32 bytes (scalar+imm) +FAIL 1100 010- ---- ---- ---- ---- ---- ---- # SVE 64-bit gather load/prefetch +FAIL 1110 010- -00- ---- 001- ---- ---- ---- # SVE2 64-bit scatter NT store (vector+scalar) +FAIL 1110 010- -10- ---- 001- ---- ---- ---- # SVE2 32-bit scatter NT store (vector+scalar) +FAIL 1110 010- ---- ---- 1-0- ---- ---- ---- # SVE scatter store (scalar+32-bit vector) +FAIL 1110 010- ---- ---- 101- ---- ---- ---- # SVE scatter store (misc) 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) return 0; } +/* This corresponds to the ARM pseudocode function IsFullA64Enabled(). */ +static bool sme_fa64(CPUARMState *env, int el) +{ + if (!cpu_isar_feature(aa64_sme_fa64, env_archcpu(env))) { + return false; + } + + if (el <= 1 && !el_is_in_host(env, el)) { + if (!FIELD_EX64(env->vfp.smcr_el[1], SMCR, FA64)) { + return false; + } + } + if (el <= 2 && arm_is_el2_enabled(env)) { + if (!FIELD_EX64(env->vfp.smcr_el[2], SMCR, FA64)) { + return false; + } + } + if (arm_feature(env, ARM_FEATURE_EL3)) { + if (!FIELD_EX64(env->vfp.smcr_el[3], SMCR, FA64)) { + return false; + } + } + + return true; +} + /* * Given that SVE is enabled, return the vector length for EL. */ @@ -XXX,XX +XXX,XX @@ static CPUARMTBFlags rebuild_hflags_a32(CPUARMState *env, int fp_el, DP_TBFLAG_ANY(flags, PSTATE__IL, 1); } + /* + * The SME exception we are testing for is raised via + * AArch64.CheckFPAdvSIMDEnabled(), as called from + * AArch32.CheckAdvSIMDOrFPEnabled(). + */ + if (el == 0 + && FIELD_EX64(env->svcr, SVCR, SM) + && (!arm_is_el2_enabled(env) + || (arm_el_is_aa64(env, 2) && !(env->cp15.hcr_el2 & HCR_TGE))) + && arm_el_is_aa64(env, 1) + && !sme_fa64(env, el)) { + DP_TBFLAG_A32(flags, SME_TRAP_NONSTREAMING, 1); + } + return rebuild_hflags_common_32(env, fp_el, mmu_idx, flags); } @@ -XXX,XX +XXX,XX @@ static CPUARMTBFlags rebuild_hflags_a64(CPUARMState *env, int el, int fp_el, } if (FIELD_EX64(env->svcr, SVCR, SM)) { DP_TBFLAG_A64(flags, PSTATE_SM, 1); + DP_TBFLAG_A64(flags, SME_TRAP_NONSTREAMING, !sme_fa64(env, el)); } DP_TBFLAG_A64(flags, PSTATE_ZA, FIELD_EX64(env->svcr, SVCR, ZA)); } 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 do_vec_ld(DisasContext *s, int destidx, int element, * unallocated-encoding checks (otherwise the syndrome information * for the resulting exception will be incorrect). */ -static bool fp_access_check(DisasContext *s) +static bool fp_access_check_only(DisasContext *s) { if (s->fp_excp_el) { assert(!s->fp_access_checked); @@ -XXX,XX +XXX,XX @@ static bool fp_access_check(DisasContext *s) return true; } +static bool fp_access_check(DisasContext *s) +{ + if (!fp_access_check_only(s)) { + return false; + } + if (s->sme_trap_nonstreaming && s->is_nonstreaming) { + gen_exception_insn(s, s->pc_curr, EXCP_UDEF, + syn_smetrap(SME_ET_Streaming, false)); + return false; + } + return true; +} + /* Check that SVE access is enabled. If it is, return true. * If not, emit code to generate an appropriate exception and return false. */ @@ -XXX,XX +XXX,XX @@ static void handle_sys(DisasContext *s, uint32_t insn, bool isread, default: g_assert_not_reached(); } - if ((ri->type & ARM_CP_FPU) && !fp_access_check(s)) { + if ((ri->type & ARM_CP_FPU) && !fp_access_check_only(s)) { return; } else if ((ri->type & ARM_CP_SVE) && !sve_access_check(s)) { return; @@ -XXX,XX +XXX,XX @@ static void disas_data_proc_simd_fp(DisasContext *s, uint32_t insn) } } +/* + * Include the generated SME FA64 decoder. + */ + +#include "decode-sme-fa64.c.inc" + +static bool trans_OK(DisasContext *s, arg_OK *a) +{ + return true; +} + +static bool trans_FAIL(DisasContext *s, arg_OK *a) +{ + s->is_nonstreaming = true; + return true; +} + /** * is_guarded_page: * @env: The cpu environment @@ -XXX,XX +XXX,XX @@ static void aarch64_tr_init_disas_context(DisasContextBase *dcbase, 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->sme_trap_nonstreaming = EX_TBFLAG_A64(tb_flags, SME_TRAP_NONSTREAMING); dc->vec_len = 0; dc->vec_stride = 0; dc->cp_regs = arm_cpu->cp_regs; @@ -XXX,XX +XXX,XX @@ static void aarch64_tr_translate_insn(DisasContextBase *dcbase, CPUState *cpu) } } + s->is_nonstreaming = false; + if (s->sme_trap_nonstreaming) { + disas_sme_fa64(s, insn); + } + switch (extract32(insn, 25, 4)) { case 0x0: if (!extract32(insn, 31, 1) || !disas_sme(s, insn)) { diff --git a/target/arm/translate-vfp.c b/target/arm/translate-vfp.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-vfp.c +++ b/target/arm/translate-vfp.c @@ -XXX,XX +XXX,XX @@ static bool vfp_access_check_a(DisasContext *s, bool ignore_vfp_enabled) return false; } + /* + * Note that rebuild_hflags_a32 has already accounted for being in EL0 + * and the higher EL in A64 mode, etc. Unlike A64 mode, there do not + * appear to be any insns which touch VFP which are allowed. + */ + if (s->sme_trap_nonstreaming) { + gen_exception_insn(s, s->pc_curr, EXCP_UDEF, + syn_smetrap(SME_ET_Streaming, + s->base.pc_next - s->pc_curr == 2)); + return false; + } + if (!s->vfp_enabled && !ignore_vfp_enabled) { assert(!arm_dc_feature(s, ARM_FEATURE_M)); unallocated_encoding(s); diff --git a/target/arm/translate.c b/target/arm/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate.c +++ b/target/arm/translate.c @@ -XXX,XX +XXX,XX @@ static void arm_tr_init_disas_context(DisasContextBase *dcbase, CPUState *cs) dc->vec_len = EX_TBFLAG_A32(tb_flags, VECLEN); dc->vec_stride = EX_TBFLAG_A32(tb_flags, VECSTRIDE); } + dc->sme_trap_nonstreaming = + EX_TBFLAG_A32(tb_flags, SME_TRAP_NONSTREAMING); } dc->cp_regs = cpu->cp_regs; dc->features = env->features; 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 @@ gen = [ decodetree.process('sve.decode', extra_args: '--decode=disas_sve'), decodetree.process('sme.decode', extra_args: '--decode=disas_sme'), + decodetree.process('sme-fa64.decode', extra_args: '--static-decode=disas_sme_fa64'), decodetree.process('neon-shared.decode', extra_args: '--decode=disas_neon_shared'), decodetree.process('neon-dp.decode', extra_args: '--decode=disas_neon_dp'), decodetree.process('neon-ls.decode', extra_args: '--decode=disas_neon_ls'), -- 2.25.1 From: Richard Henderson <richard.henderson@linaro.org> Mark ADR as a non-streaming instruction, which should trap if full a64 support is not enabled in streaming mode. Removing entries from sme-fa64.decode is an easy way to see what remains to be done. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20220708151540.18136-5-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/translate.h | 7 +++++++ target/arm/sme-fa64.decode | 1 - target/arm/translate-sve.c | 8 ++++---- 3 files changed, 11 insertions(+), 5 deletions(-) diff --git a/target/arm/translate.h b/target/arm/translate.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate.h +++ b/target/arm/translate.h @@ -XXX,XX +XXX,XX @@ uint64_t asimd_imm_const(uint32_t imm, int cmode, int op); static bool trans_##NAME(DisasContext *s, arg_##NAME *a) \ { return dc_isar_feature(FEAT, s) && FUNC(s, __VA_ARGS__); } +#define TRANS_FEAT_NONSTREAMING(NAME, FEAT, FUNC, ...) \ + static bool trans_##NAME(DisasContext *s, arg_##NAME *a) \ + { \ + s->is_nonstreaming = true; \ + return dc_isar_feature(FEAT, s) && FUNC(s, __VA_ARGS__); \ + } + #endif /* TARGET_ARM_TRANSLATE_H */ diff --git a/target/arm/sme-fa64.decode b/target/arm/sme-fa64.decode index XXXXXXX..XXXXXXX 100644 --- a/target/arm/sme-fa64.decode +++ b/target/arm/sme-fa64.decode @@ -XXX,XX +XXX,XX @@ FAIL 0001 1110 0111 1110 0000 00-- ---- ---- # FJCVTZS # --11 1100 --1- ---- ---- ---- ---- --10 # Load/store FP register (register offset) # --11 1101 ---- ---- ---- ---- ---- ---- # Load/store FP register (scaled imm) -FAIL 0000 0100 --1- ---- 1010 ---- ---- ---- # ADR FAIL 0000 0100 --1- ---- 1011 -0-- ---- ---- # FTSSEL, FEXPA FAIL 0000 0101 --10 0001 100- ---- ---- ---- # COMPACT FAIL 0010 0101 --01 100- 1111 000- ---0 ---- # RDFFR, RDFFRS diff --git a/target/arm/translate-sve.c b/target/arm/translate-sve.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-sve.c +++ b/target/arm/translate-sve.c @@ -XXX,XX +XXX,XX @@ static bool do_adr(DisasContext *s, arg_rrri *a, gen_helper_gvec_3 *fn) return gen_gvec_ool_zzz(s, fn, a->rd, a->rn, a->rm, a->imm); } -TRANS_FEAT(ADR_p32, aa64_sve, do_adr, a, gen_helper_sve_adr_p32) -TRANS_FEAT(ADR_p64, aa64_sve, do_adr, a, gen_helper_sve_adr_p64) -TRANS_FEAT(ADR_s32, aa64_sve, do_adr, a, gen_helper_sve_adr_s32) -TRANS_FEAT(ADR_u32, aa64_sve, do_adr, a, gen_helper_sve_adr_u32) +TRANS_FEAT_NONSTREAMING(ADR_p32, aa64_sve, do_adr, a, gen_helper_sve_adr_p32) +TRANS_FEAT_NONSTREAMING(ADR_p64, aa64_sve, do_adr, a, gen_helper_sve_adr_p64) +TRANS_FEAT_NONSTREAMING(ADR_s32, aa64_sve, do_adr, a, gen_helper_sve_adr_s32) +TRANS_FEAT_NONSTREAMING(ADR_u32, aa64_sve, do_adr, a, gen_helper_sve_adr_u32) /* *** SVE Integer Misc - Unpredicated Group -- 2.25.1 From: Richard Henderson <richard.henderson@linaro.org> Mark these as a non-streaming instructions, which should trap if full a64 support is not enabled in streaming mode. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20220708151540.18136-6-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/sme-fa64.decode | 2 -- target/arm/translate-sve.c | 9 ++++++--- 2 files changed, 6 insertions(+), 5 deletions(-) diff --git a/target/arm/sme-fa64.decode b/target/arm/sme-fa64.decode index XXXXXXX..XXXXXXX 100644 --- a/target/arm/sme-fa64.decode +++ b/target/arm/sme-fa64.decode @@ -XXX,XX +XXX,XX @@ FAIL 0001 1110 0111 1110 0000 00-- ---- ---- # FJCVTZS FAIL 0000 0100 --1- ---- 1011 -0-- ---- ---- # FTSSEL, FEXPA FAIL 0000 0101 --10 0001 100- ---- ---- ---- # COMPACT -FAIL 0010 0101 --01 100- 1111 000- ---0 ---- # RDFFR, RDFFRS -FAIL 0010 0101 --10 1--- 1001 ---- ---- ---- # WRFFR, SETFFR FAIL 0100 0101 --0- ---- 1011 ---- ---- ---- # BDEP, BEXT, BGRP FAIL 0100 0101 000- ---- 0110 1--- ---- ---- # PMULLB, PMULLT (128b result) FAIL 0110 0100 --1- ---- 1110 01-- ---- ---- # FMMLA, BFMMLA diff --git a/target/arm/translate-sve.c b/target/arm/translate-sve.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-sve.c +++ b/target/arm/translate-sve.c @@ -XXX,XX +XXX,XX @@ static bool do_predset(DisasContext *s, int esz, int rd, int pat, bool setflag) TRANS_FEAT(PTRUE, aa64_sve, do_predset, a->esz, a->rd, a->pat, a->s) /* Note pat == 31 is #all, to set all elements. */ -TRANS_FEAT(SETFFR, aa64_sve, do_predset, 0, FFR_PRED_NUM, 31, false) +TRANS_FEAT_NONSTREAMING(SETFFR, aa64_sve, + do_predset, 0, FFR_PRED_NUM, 31, false) /* Note pat == 32 is #unimp, to set no elements. */ TRANS_FEAT(PFALSE, aa64_sve, do_predset, 0, a->rd, 32, false) @@ -XXX,XX +XXX,XX @@ static bool trans_RDFFR_p(DisasContext *s, arg_RDFFR_p *a) .rd = a->rd, .pg = a->pg, .s = a->s, .rn = FFR_PRED_NUM, .rm = FFR_PRED_NUM, }; + + s->is_nonstreaming = true; return trans_AND_pppp(s, &alt_a); } -TRANS_FEAT(RDFFR, aa64_sve, do_mov_p, a->rd, FFR_PRED_NUM) -TRANS_FEAT(WRFFR, aa64_sve, do_mov_p, FFR_PRED_NUM, a->rn) +TRANS_FEAT_NONSTREAMING(RDFFR, aa64_sve, do_mov_p, a->rd, FFR_PRED_NUM) +TRANS_FEAT_NONSTREAMING(WRFFR, aa64_sve, do_mov_p, FFR_PRED_NUM, a->rn) static bool do_pfirst_pnext(DisasContext *s, arg_rr_esz *a, void (*gen_fn)(TCGv_i32, TCGv_ptr, -- 2.25.1 From: Richard Henderson <richard.henderson@linaro.org> Mark these as a non-streaming instructions, which should trap if full a64 support is not enabled in streaming mode. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20220708151540.18136-7-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/sme-fa64.decode | 3 --- target/arm/translate-sve.c | 22 ++++++++++++---------- 2 files changed, 12 insertions(+), 13 deletions(-) diff --git a/target/arm/sme-fa64.decode b/target/arm/sme-fa64.decode index XXXXXXX..XXXXXXX 100644 --- a/target/arm/sme-fa64.decode +++ b/target/arm/sme-fa64.decode @@ -XXX,XX +XXX,XX @@ FAIL 0001 1110 0111 1110 0000 00-- ---- ---- # FJCVTZS # --11 1100 --1- ---- ---- ---- ---- --10 # Load/store FP register (register offset) # --11 1101 ---- ---- ---- ---- ---- ---- # Load/store FP register (scaled imm) -FAIL 0000 0100 --1- ---- 1011 -0-- ---- ---- # FTSSEL, FEXPA -FAIL 0000 0101 --10 0001 100- ---- ---- ---- # COMPACT -FAIL 0100 0101 --0- ---- 1011 ---- ---- ---- # BDEP, BEXT, BGRP FAIL 0100 0101 000- ---- 0110 1--- ---- ---- # PMULLB, PMULLT (128b result) FAIL 0110 0100 --1- ---- 1110 01-- ---- ---- # FMMLA, BFMMLA FAIL 0110 0101 --0- ---- 0000 11-- ---- ---- # FTSMUL 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 gen_helper_gvec_2 * const fexpa_fns[4] = { NULL, gen_helper_sve_fexpa_h, gen_helper_sve_fexpa_s, gen_helper_sve_fexpa_d, }; -TRANS_FEAT(FEXPA, aa64_sve, gen_gvec_ool_zz, - fexpa_fns[a->esz], a->rd, a->rn, 0) +TRANS_FEAT_NONSTREAMING(FEXPA, aa64_sve, gen_gvec_ool_zz, + fexpa_fns[a->esz], a->rd, a->rn, 0) static gen_helper_gvec_3 * const ftssel_fns[4] = { NULL, gen_helper_sve_ftssel_h, gen_helper_sve_ftssel_s, gen_helper_sve_ftssel_d, }; -TRANS_FEAT(FTSSEL, aa64_sve, gen_gvec_ool_arg_zzz, ftssel_fns[a->esz], a, 0) +TRANS_FEAT_NONSTREAMING(FTSSEL, aa64_sve, gen_gvec_ool_arg_zzz, + ftssel_fns[a->esz], a, 0) /* *** SVE Predicate Logical Operations Group @@ -XXX,XX +XXX,XX @@ TRANS_FEAT(TRN2_q, aa64_sve_f64mm, gen_gvec_ool_arg_zzz, static gen_helper_gvec_3 * const compact_fns[4] = { NULL, NULL, gen_helper_sve_compact_s, gen_helper_sve_compact_d }; -TRANS_FEAT(COMPACT, aa64_sve, gen_gvec_ool_arg_zpz, compact_fns[a->esz], a, 0) +TRANS_FEAT_NONSTREAMING(COMPACT, aa64_sve, gen_gvec_ool_arg_zpz, + compact_fns[a->esz], a, 0) /* Call the helper that computes the ARM LastActiveElement pseudocode * function, scaled by the element size. This includes the not found @@ -XXX,XX +XXX,XX @@ static gen_helper_gvec_3 * const bext_fns[4] = { gen_helper_sve2_bext_b, gen_helper_sve2_bext_h, gen_helper_sve2_bext_s, gen_helper_sve2_bext_d, }; -TRANS_FEAT(BEXT, aa64_sve2_bitperm, gen_gvec_ool_arg_zzz, - bext_fns[a->esz], a, 0) +TRANS_FEAT_NONSTREAMING(BEXT, aa64_sve2_bitperm, gen_gvec_ool_arg_zzz, + bext_fns[a->esz], a, 0) static gen_helper_gvec_3 * const bdep_fns[4] = { gen_helper_sve2_bdep_b, gen_helper_sve2_bdep_h, gen_helper_sve2_bdep_s, gen_helper_sve2_bdep_d, }; -TRANS_FEAT(BDEP, aa64_sve2_bitperm, gen_gvec_ool_arg_zzz, - bdep_fns[a->esz], a, 0) +TRANS_FEAT_NONSTREAMING(BDEP, aa64_sve2_bitperm, gen_gvec_ool_arg_zzz, + bdep_fns[a->esz], a, 0) static gen_helper_gvec_3 * const bgrp_fns[4] = { gen_helper_sve2_bgrp_b, gen_helper_sve2_bgrp_h, gen_helper_sve2_bgrp_s, gen_helper_sve2_bgrp_d, }; -TRANS_FEAT(BGRP, aa64_sve2_bitperm, gen_gvec_ool_arg_zzz, - bgrp_fns[a->esz], a, 0) +TRANS_FEAT_NONSTREAMING(BGRP, aa64_sve2_bitperm, gen_gvec_ool_arg_zzz, + bgrp_fns[a->esz], a, 0) static gen_helper_gvec_3 * const cadd_fns[4] = { gen_helper_sve2_cadd_b, gen_helper_sve2_cadd_h, -- 2.25.1 From: Richard Henderson <richard.henderson@linaro.org> Mark these as a non-streaming instructions, which should trap if full a64 support is not enabled in streaming mode. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20220708151540.18136-8-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/sme-fa64.decode | 2 -- target/arm/translate-sve.c | 24 +++++++++++++++--------- 2 files changed, 15 insertions(+), 11 deletions(-) diff --git a/target/arm/sme-fa64.decode b/target/arm/sme-fa64.decode index XXXXXXX..XXXXXXX 100644 --- a/target/arm/sme-fa64.decode +++ b/target/arm/sme-fa64.decode @@ -XXX,XX +XXX,XX @@ FAIL 0001 1110 0111 1110 0000 00-- ---- ---- # FJCVTZS # --11 1100 --1- ---- ---- ---- ---- --10 # Load/store FP register (register offset) # --11 1101 ---- ---- ---- ---- ---- ---- # Load/store FP register (scaled imm) -FAIL 0100 0101 000- ---- 0110 1--- ---- ---- # PMULLB, PMULLT (128b result) -FAIL 0110 0100 --1- ---- 1110 01-- ---- ---- # FMMLA, BFMMLA FAIL 0110 0101 --0- ---- 0000 11-- ---- ---- # FTSMUL FAIL 0110 0101 --01 0--- 100- ---- ---- ---- # FTMAD FAIL 0110 0101 --01 1--- 001- ---- ---- ---- # FADDA diff --git a/target/arm/translate-sve.c b/target/arm/translate-sve.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-sve.c +++ b/target/arm/translate-sve.c @@ -XXX,XX +XXX,XX @@ static bool do_trans_pmull(DisasContext *s, arg_rrr_esz *a, bool sel) gen_helper_gvec_pmull_q, gen_helper_sve2_pmull_h, NULL, gen_helper_sve2_pmull_d, }; - if (a->esz == 0 - ? !dc_isar_feature(aa64_sve2_pmull128, s) - : !dc_isar_feature(aa64_sve, s)) { + + if (a->esz == 0) { + if (!dc_isar_feature(aa64_sve2_pmull128, s)) { + return false; + } + s->is_nonstreaming = true; + } else if (!dc_isar_feature(aa64_sve, s)) { return false; } return gen_gvec_ool_arg_zzz(s, fns[a->esz], a, sel); @@ -XXX,XX +XXX,XX @@ DO_ZPZZ_FP(FMINP, aa64_sve2, sve2_fminp_zpzz) * SVE Integer Multiply-Add (unpredicated) */ -TRANS_FEAT(FMMLA_s, aa64_sve_f32mm, gen_gvec_fpst_zzzz, gen_helper_fmmla_s, - a->rd, a->rn, a->rm, a->ra, 0, FPST_FPCR) -TRANS_FEAT(FMMLA_d, aa64_sve_f64mm, gen_gvec_fpst_zzzz, gen_helper_fmmla_d, - a->rd, a->rn, a->rm, a->ra, 0, FPST_FPCR) +TRANS_FEAT_NONSTREAMING(FMMLA_s, aa64_sve_f32mm, gen_gvec_fpst_zzzz, + gen_helper_fmmla_s, a->rd, a->rn, a->rm, a->ra, + 0, FPST_FPCR) +TRANS_FEAT_NONSTREAMING(FMMLA_d, aa64_sve_f64mm, gen_gvec_fpst_zzzz, + gen_helper_fmmla_d, a->rd, a->rn, a->rm, a->ra, + 0, FPST_FPCR) static gen_helper_gvec_4 * const sqdmlal_zzzw_fns[] = { NULL, gen_helper_sve2_sqdmlal_zzzw_h, @@ -XXX,XX +XXX,XX @@ TRANS_FEAT(BFDOT_zzzz, aa64_sve_bf16, gen_gvec_ool_arg_zzzz, TRANS_FEAT(BFDOT_zzxz, aa64_sve_bf16, gen_gvec_ool_arg_zzxz, gen_helper_gvec_bfdot_idx, a) -TRANS_FEAT(BFMMLA, aa64_sve_bf16, gen_gvec_ool_arg_zzzz, - gen_helper_gvec_bfmmla, a, 0) +TRANS_FEAT_NONSTREAMING(BFMMLA, aa64_sve_bf16, gen_gvec_ool_arg_zzzz, + gen_helper_gvec_bfmmla, a, 0) static bool do_BFMLAL_zzzw(DisasContext *s, arg_rrrr_esz *a, bool sel) { -- 2.25.1 From: Richard Henderson <richard.henderson@linaro.org> Mark these as a non-streaming instructions, which should trap if full a64 support is not enabled in streaming mode. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20220708151540.18136-9-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/sme-fa64.decode | 3 --- target/arm/translate-sve.c | 15 +++++++++++---- 2 files changed, 11 insertions(+), 7 deletions(-) diff --git a/target/arm/sme-fa64.decode b/target/arm/sme-fa64.decode index XXXXXXX..XXXXXXX 100644 --- a/target/arm/sme-fa64.decode +++ b/target/arm/sme-fa64.decode @@ -XXX,XX +XXX,XX @@ FAIL 0001 1110 0111 1110 0000 00-- ---- ---- # FJCVTZS # --11 1100 --1- ---- ---- ---- ---- --10 # Load/store FP register (register offset) # --11 1101 ---- ---- ---- ---- ---- ---- # Load/store FP register (scaled imm) -FAIL 0110 0101 --0- ---- 0000 11-- ---- ---- # FTSMUL -FAIL 0110 0101 --01 0--- 100- ---- ---- ---- # FTMAD -FAIL 0110 0101 --01 1--- 001- ---- ---- ---- # FADDA FAIL 0100 0101 --0- ---- 1001 10-- ---- ---- # SMMLA, UMMLA, USMMLA FAIL 0100 0101 --1- ---- 1--- ---- ---- ---- # SVE2 string/histo/crypto instructions FAIL 1000 010- -00- ---- 10-- ---- ---- ---- # SVE2 32-bit gather NT load (vector+scalar) 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 gen_helper_gvec_3_ptr * const ftmad_fns[4] = { NULL, gen_helper_sve_ftmad_h, gen_helper_sve_ftmad_s, gen_helper_sve_ftmad_d, }; -TRANS_FEAT(FTMAD, aa64_sve, gen_gvec_fpst_zzz, - ftmad_fns[a->esz], a->rd, a->rn, a->rm, a->imm, - a->esz == MO_16 ? FPST_FPCR_F16 : FPST_FPCR) +TRANS_FEAT_NONSTREAMING(FTMAD, aa64_sve, gen_gvec_fpst_zzz, + ftmad_fns[a->esz], a->rd, a->rn, a->rm, a->imm, + a->esz == MO_16 ? FPST_FPCR_F16 : FPST_FPCR) /* *** SVE Floating Point Accumulating Reduction Group @@ -XXX,XX +XXX,XX @@ static bool trans_FADDA(DisasContext *s, arg_rprr_esz *a) if (a->esz == 0 || !dc_isar_feature(aa64_sve, s)) { return false; } + s->is_nonstreaming = true; if (!sve_access_check(s)) { return true; } @@ -XXX,XX +XXX,XX @@ static bool trans_FADDA(DisasContext *s, arg_rprr_esz *a) DO_FP3(FADD_zzz, fadd) DO_FP3(FSUB_zzz, fsub) DO_FP3(FMUL_zzz, fmul) -DO_FP3(FTSMUL, ftsmul) DO_FP3(FRECPS, recps) DO_FP3(FRSQRTS, rsqrts) #undef DO_FP3 +static gen_helper_gvec_3_ptr * const ftsmul_fns[4] = { + NULL, gen_helper_gvec_ftsmul_h, + gen_helper_gvec_ftsmul_s, gen_helper_gvec_ftsmul_d +}; +TRANS_FEAT_NONSTREAMING(FTSMUL, aa64_sve, gen_gvec_fpst_arg_zzz, + ftsmul_fns[a->esz], a, 0) + /* *** SVE Floating Point Arithmetic - Predicated Group */ -- 2.25.1 From: Richard Henderson <richard.henderson@linaro.org> Mark these as a non-streaming instructions, which should trap if full a64 support is not enabled in streaming mode. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20220708151540.18136-10-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/sme-fa64.decode | 1 - target/arm/translate-sve.c | 12 ++++++------ 2 files changed, 6 insertions(+), 7 deletions(-) diff --git a/target/arm/sme-fa64.decode b/target/arm/sme-fa64.decode index XXXXXXX..XXXXXXX 100644 --- a/target/arm/sme-fa64.decode +++ b/target/arm/sme-fa64.decode @@ -XXX,XX +XXX,XX @@ FAIL 0001 1110 0111 1110 0000 00-- ---- ---- # FJCVTZS # --11 1100 --1- ---- ---- ---- ---- --10 # Load/store FP register (register offset) # --11 1101 ---- ---- ---- ---- ---- ---- # Load/store FP register (scaled imm) -FAIL 0100 0101 --0- ---- 1001 10-- ---- ---- # SMMLA, UMMLA, USMMLA FAIL 0100 0101 --1- ---- 1--- ---- ---- ---- # SVE2 string/histo/crypto instructions FAIL 1000 010- -00- ---- 10-- ---- ---- ---- # SVE2 32-bit gather NT load (vector+scalar) FAIL 1000 010- -00- ---- 111- ---- ---- ---- # SVE 32-bit gather prefetch (vector+imm) 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 @@ TRANS_FEAT(FMLALT_zzxw, aa64_sve2, do_FMLAL_zzxw, a, false, true) TRANS_FEAT(FMLSLB_zzxw, aa64_sve2, do_FMLAL_zzxw, a, true, false) TRANS_FEAT(FMLSLT_zzxw, aa64_sve2, do_FMLAL_zzxw, a, true, true) -TRANS_FEAT(SMMLA, aa64_sve_i8mm, gen_gvec_ool_arg_zzzz, - gen_helper_gvec_smmla_b, a, 0) -TRANS_FEAT(USMMLA, aa64_sve_i8mm, gen_gvec_ool_arg_zzzz, - gen_helper_gvec_usmmla_b, a, 0) -TRANS_FEAT(UMMLA, aa64_sve_i8mm, gen_gvec_ool_arg_zzzz, - gen_helper_gvec_ummla_b, a, 0) +TRANS_FEAT_NONSTREAMING(SMMLA, aa64_sve_i8mm, gen_gvec_ool_arg_zzzz, + gen_helper_gvec_smmla_b, a, 0) +TRANS_FEAT_NONSTREAMING(USMMLA, aa64_sve_i8mm, gen_gvec_ool_arg_zzzz, + gen_helper_gvec_usmmla_b, a, 0) +TRANS_FEAT_NONSTREAMING(UMMLA, aa64_sve_i8mm, gen_gvec_ool_arg_zzzz, + gen_helper_gvec_ummla_b, a, 0) TRANS_FEAT(BFDOT_zzzz, aa64_sve_bf16, gen_gvec_ool_arg_zzzz, gen_helper_gvec_bfdot, a, 0) -- 2.25.1 From: Richard Henderson <richard.henderson@linaro.org> Mark these as non-streaming instructions, which should trap if full a64 support is not enabled in streaming mode. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20220708151540.18136-11-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/sme-fa64.decode | 1 - target/arm/translate-sve.c | 35 ++++++++++++++++++----------------- 2 files changed, 18 insertions(+), 18 deletions(-) diff --git a/target/arm/sme-fa64.decode b/target/arm/sme-fa64.decode index XXXXXXX..XXXXXXX 100644 --- a/target/arm/sme-fa64.decode +++ b/target/arm/sme-fa64.decode @@ -XXX,XX +XXX,XX @@ FAIL 0001 1110 0111 1110 0000 00-- ---- ---- # FJCVTZS # --11 1100 --1- ---- ---- ---- ---- --10 # Load/store FP register (register offset) # --11 1101 ---- ---- ---- ---- ---- ---- # Load/store FP register (scaled imm) -FAIL 0100 0101 --1- ---- 1--- ---- ---- ---- # SVE2 string/histo/crypto instructions FAIL 1000 010- -00- ---- 10-- ---- ---- ---- # SVE2 32-bit gather NT load (vector+scalar) FAIL 1000 010- -00- ---- 111- ---- ---- ---- # SVE 32-bit gather prefetch (vector+imm) FAIL 1000 0100 0-1- ---- 0--- ---- ---- ---- # SVE 32-bit gather prefetch (scalar+vector) 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 @@ DO_SVE2_ZZZ_NARROW(RSUBHNT, rsubhnt) static gen_helper_gvec_flags_4 * const match_fns[4] = { gen_helper_sve2_match_ppzz_b, gen_helper_sve2_match_ppzz_h, NULL, NULL }; -TRANS_FEAT(MATCH, aa64_sve2, do_ppzz_flags, a, match_fns[a->esz]) +TRANS_FEAT_NONSTREAMING(MATCH, aa64_sve2, do_ppzz_flags, a, match_fns[a->esz]) static gen_helper_gvec_flags_4 * const nmatch_fns[4] = { gen_helper_sve2_nmatch_ppzz_b, gen_helper_sve2_nmatch_ppzz_h, NULL, NULL }; -TRANS_FEAT(NMATCH, aa64_sve2, do_ppzz_flags, a, nmatch_fns[a->esz]) +TRANS_FEAT_NONSTREAMING(NMATCH, aa64_sve2, do_ppzz_flags, a, nmatch_fns[a->esz]) static gen_helper_gvec_4 * const histcnt_fns[4] = { NULL, NULL, gen_helper_sve2_histcnt_s, gen_helper_sve2_histcnt_d }; -TRANS_FEAT(HISTCNT, aa64_sve2, gen_gvec_ool_arg_zpzz, - histcnt_fns[a->esz], a, 0) +TRANS_FEAT_NONSTREAMING(HISTCNT, aa64_sve2, gen_gvec_ool_arg_zpzz, + histcnt_fns[a->esz], a, 0) -TRANS_FEAT(HISTSEG, aa64_sve2, gen_gvec_ool_arg_zzz, - a->esz == 0 ? gen_helper_sve2_histseg : NULL, a, 0) +TRANS_FEAT_NONSTREAMING(HISTSEG, aa64_sve2, gen_gvec_ool_arg_zzz, + a->esz == 0 ? gen_helper_sve2_histseg : NULL, a, 0) DO_ZPZZ_FP(FADDP, aa64_sve2, sve2_faddp_zpzz) DO_ZPZZ_FP(FMAXNMP, aa64_sve2, sve2_fmaxnmp_zpzz) @@ -XXX,XX +XXX,XX @@ TRANS_FEAT(SQRDCMLAH_zzzz, aa64_sve2, gen_gvec_ool_zzzz, TRANS_FEAT(USDOT_zzzz, aa64_sve_i8mm, gen_gvec_ool_arg_zzzz, a->esz == 2 ? gen_helper_gvec_usdot_b : NULL, a, 0) -TRANS_FEAT(AESMC, aa64_sve2_aes, gen_gvec_ool_zz, - gen_helper_crypto_aesmc, a->rd, a->rd, a->decrypt) +TRANS_FEAT_NONSTREAMING(AESMC, aa64_sve2_aes, gen_gvec_ool_zz, + gen_helper_crypto_aesmc, a->rd, a->rd, a->decrypt) -TRANS_FEAT(AESE, aa64_sve2_aes, gen_gvec_ool_arg_zzz, - gen_helper_crypto_aese, a, false) -TRANS_FEAT(AESD, aa64_sve2_aes, gen_gvec_ool_arg_zzz, - gen_helper_crypto_aese, a, true) +TRANS_FEAT_NONSTREAMING(AESE, aa64_sve2_aes, gen_gvec_ool_arg_zzz, + gen_helper_crypto_aese, a, false) +TRANS_FEAT_NONSTREAMING(AESD, aa64_sve2_aes, gen_gvec_ool_arg_zzz, + gen_helper_crypto_aese, a, true) -TRANS_FEAT(SM4E, aa64_sve2_sm4, gen_gvec_ool_arg_zzz, - gen_helper_crypto_sm4e, a, 0) -TRANS_FEAT(SM4EKEY, aa64_sve2_sm4, gen_gvec_ool_arg_zzz, - gen_helper_crypto_sm4ekey, a, 0) +TRANS_FEAT_NONSTREAMING(SM4E, aa64_sve2_sm4, gen_gvec_ool_arg_zzz, + gen_helper_crypto_sm4e, a, 0) +TRANS_FEAT_NONSTREAMING(SM4EKEY, aa64_sve2_sm4, gen_gvec_ool_arg_zzz, + gen_helper_crypto_sm4ekey, a, 0) -TRANS_FEAT(RAX1, aa64_sve2_sha3, gen_gvec_fn_arg_zzz, gen_gvec_rax1, a) +TRANS_FEAT_NONSTREAMING(RAX1, aa64_sve2_sha3, gen_gvec_fn_arg_zzz, + gen_gvec_rax1, a) TRANS_FEAT(FCVTNT_sh, aa64_sve2, gen_gvec_fpst_arg_zpz, gen_helper_sve2_fcvtnt_sh, a, 0, FPST_FPCR) -- 2.25.1 From: Richard Henderson <richard.henderson@linaro.org> Mark these as a non-streaming instructions, which should trap if full a64 support is not enabled in streaming mode. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20220708151540.18136-12-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/sme-fa64.decode | 9 --------- target/arm/translate-sve.c | 6 ++++++ 2 files changed, 6 insertions(+), 9 deletions(-) diff --git a/target/arm/sme-fa64.decode b/target/arm/sme-fa64.decode index XXXXXXX..XXXXXXX 100644 --- a/target/arm/sme-fa64.decode +++ b/target/arm/sme-fa64.decode @@ -XXX,XX +XXX,XX @@ FAIL 0001 1110 0111 1110 0000 00-- ---- ---- # FJCVTZS # --11 1100 --1- ---- ---- ---- ---- --10 # Load/store FP register (register offset) # --11 1101 ---- ---- ---- ---- ---- ---- # Load/store FP register (scaled imm) -FAIL 1000 010- -00- ---- 10-- ---- ---- ---- # SVE2 32-bit gather NT load (vector+scalar) FAIL 1000 010- -00- ---- 111- ---- ---- ---- # SVE 32-bit gather prefetch (vector+imm) FAIL 1000 0100 0-1- ---- 0--- ---- ---- ---- # SVE 32-bit gather prefetch (scalar+vector) -FAIL 1000 010- -01- ---- 1--- ---- ---- ---- # SVE 32-bit gather load (vector+imm) -FAIL 1000 0100 0-0- ---- 0--- ---- ---- ---- # SVE 32-bit gather load byte (scalar+vector) -FAIL 1000 0100 1--- ---- 0--- ---- ---- ---- # SVE 32-bit gather load half (scalar+vector) -FAIL 1000 0101 0--- ---- 0--- ---- ---- ---- # SVE 32-bit gather load word (scalar+vector) FAIL 1010 010- ---- ---- 011- ---- ---- ---- # SVE contiguous FF load (scalar+scalar) FAIL 1010 010- ---1 ---- 101- ---- ---- ---- # SVE contiguous NF load (scalar+imm) FAIL 1010 010- -01- ---- 000- ---- ---- ---- # SVE load & replicate 32 bytes (scalar+scalar) FAIL 1010 010- -010 ---- 001- ---- ---- ---- # SVE load & replicate 32 bytes (scalar+imm) FAIL 1100 010- ---- ---- ---- ---- ---- ---- # SVE 64-bit gather load/prefetch -FAIL 1110 010- -00- ---- 001- ---- ---- ---- # SVE2 64-bit scatter NT store (vector+scalar) -FAIL 1110 010- -10- ---- 001- ---- ---- ---- # SVE2 32-bit scatter NT store (vector+scalar) -FAIL 1110 010- ---- ---- 1-0- ---- ---- ---- # SVE scatter store (scalar+32-bit vector) -FAIL 1110 010- ---- ---- 101- ---- ---- ---- # SVE scatter store (misc) diff --git a/target/arm/translate-sve.c b/target/arm/translate-sve.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-sve.c +++ b/target/arm/translate-sve.c @@ -XXX,XX +XXX,XX @@ static bool trans_LD1_zprz(DisasContext *s, arg_LD1_zprz *a) if (!dc_isar_feature(aa64_sve, s)) { return false; } + s->is_nonstreaming = true; if (!sve_access_check(s)) { return true; } @@ -XXX,XX +XXX,XX @@ static bool trans_LD1_zpiz(DisasContext *s, arg_LD1_zpiz *a) if (!dc_isar_feature(aa64_sve, s)) { return false; } + s->is_nonstreaming = true; if (!sve_access_check(s)) { return true; } @@ -XXX,XX +XXX,XX @@ static bool trans_LDNT1_zprz(DisasContext *s, arg_LD1_zprz *a) if (!dc_isar_feature(aa64_sve2, s)) { return false; } + s->is_nonstreaming = true; if (!sve_access_check(s)) { return true; } @@ -XXX,XX +XXX,XX @@ static bool trans_ST1_zprz(DisasContext *s, arg_ST1_zprz *a) if (!dc_isar_feature(aa64_sve, s)) { return false; } + s->is_nonstreaming = true; if (!sve_access_check(s)) { return true; } @@ -XXX,XX +XXX,XX @@ static bool trans_ST1_zpiz(DisasContext *s, arg_ST1_zpiz *a) if (!dc_isar_feature(aa64_sve, s)) { return false; } + s->is_nonstreaming = true; if (!sve_access_check(s)) { return true; } @@ -XXX,XX +XXX,XX @@ static bool trans_STNT1_zprz(DisasContext *s, arg_ST1_zprz *a) if (!dc_isar_feature(aa64_sve2, s)) { return false; } + s->is_nonstreaming = true; if (!sve_access_check(s)) { return true; } -- 2.25.1 From: Richard Henderson <richard.henderson@linaro.org> Mark these as a non-streaming instructions, which should trap if full a64 support is not enabled in streaming mode. In this case, introduce PRF_ns (prefetch non-streaming) to handle the checks. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20220708151540.18136-13-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/sme-fa64.decode | 3 --- target/arm/sve.decode | 10 +++++----- target/arm/translate-sve.c | 11 +++++++++++ 3 files changed, 16 insertions(+), 8 deletions(-) diff --git a/target/arm/sme-fa64.decode b/target/arm/sme-fa64.decode index XXXXXXX..XXXXXXX 100644 --- a/target/arm/sme-fa64.decode +++ b/target/arm/sme-fa64.decode @@ -XXX,XX +XXX,XX @@ FAIL 0001 1110 0111 1110 0000 00-- ---- ---- # FJCVTZS # --11 1100 --1- ---- ---- ---- ---- --10 # Load/store FP register (register offset) # --11 1101 ---- ---- ---- ---- ---- ---- # Load/store FP register (scaled imm) -FAIL 1000 010- -00- ---- 111- ---- ---- ---- # SVE 32-bit gather prefetch (vector+imm) -FAIL 1000 0100 0-1- ---- 0--- ---- ---- ---- # SVE 32-bit gather prefetch (scalar+vector) FAIL 1010 010- ---- ---- 011- ---- ---- ---- # SVE contiguous FF load (scalar+scalar) FAIL 1010 010- ---1 ---- 101- ---- ---- ---- # SVE contiguous NF load (scalar+imm) FAIL 1010 010- -01- ---- 000- ---- ---- ---- # SVE load & replicate 32 bytes (scalar+scalar) FAIL 1010 010- -010 ---- 001- ---- ---- ---- # SVE load & replicate 32 bytes (scalar+imm) -FAIL 1100 010- ---- ---- ---- ---- ---- ---- # SVE 64-bit gather load/prefetch diff --git a/target/arm/sve.decode b/target/arm/sve.decode index XXXXXXX..XXXXXXX 100644 --- a/target/arm/sve.decode +++ b/target/arm/sve.decode @@ -XXX,XX +XXX,XX @@ LD1RO_zpri 1010010 .. 01 0.... 001 ... ..... ..... \ @rpri_load_msz nreg=0 # SVE 32-bit gather prefetch (scalar plus 32-bit scaled offsets) -PRF 1000010 00 -1 ----- 0-- --- ----- 0 ---- +PRF_ns 1000010 00 -1 ----- 0-- --- ----- 0 ---- # SVE 32-bit gather prefetch (vector plus immediate) -PRF 1000010 -- 00 ----- 111 --- ----- 0 ---- +PRF_ns 1000010 -- 00 ----- 111 --- ----- 0 ---- # SVE contiguous prefetch (scalar plus immediate) PRF 1000010 11 1- ----- 0-- --- ----- 0 ---- @@ -XXX,XX +XXX,XX @@ LD1_zpiz 1100010 .. 01 ..... 1.. ... ..... ..... \ @rpri_g_load esz=3 # SVE 64-bit gather prefetch (scalar plus 64-bit scaled offsets) -PRF 1100010 00 11 ----- 1-- --- ----- 0 ---- +PRF_ns 1100010 00 11 ----- 1-- --- ----- 0 ---- # SVE 64-bit gather prefetch (scalar plus unpacked 32-bit scaled offsets) -PRF 1100010 00 -1 ----- 0-- --- ----- 0 ---- +PRF_ns 1100010 00 -1 ----- 0-- --- ----- 0 ---- # SVE 64-bit gather prefetch (vector plus immediate) -PRF 1100010 -- 00 ----- 111 --- ----- 0 ---- +PRF_ns 1100010 -- 00 ----- 111 --- ----- 0 ---- ### SVE Memory Store Group diff --git a/target/arm/translate-sve.c b/target/arm/translate-sve.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-sve.c +++ b/target/arm/translate-sve.c @@ -XXX,XX +XXX,XX @@ static bool trans_PRF_rr(DisasContext *s, arg_PRF_rr *a) return true; } +static bool trans_PRF_ns(DisasContext *s, arg_PRF_ns *a) +{ + if (!dc_isar_feature(aa64_sve, s)) { + return false; + } + /* Prefetch is a nop within QEMU. */ + s->is_nonstreaming = true; + (void)sve_access_check(s); + return true; +} + /* * Move Prefix * -- 2.25.1 From: Richard Henderson <richard.henderson@linaro.org> Mark these as a non-streaming instructions, which should trap if full a64 support is not enabled in streaming mode. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20220708151540.18136-14-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/sme-fa64.decode | 2 -- target/arm/translate-sve.c | 2 ++ 2 files changed, 2 insertions(+), 2 deletions(-) diff --git a/target/arm/sme-fa64.decode b/target/arm/sme-fa64.decode index XXXXXXX..XXXXXXX 100644 --- a/target/arm/sme-fa64.decode +++ b/target/arm/sme-fa64.decode @@ -XXX,XX +XXX,XX @@ FAIL 0001 1110 0111 1110 0000 00-- ---- ---- # FJCVTZS # --11 1100 --1- ---- ---- ---- ---- --10 # Load/store FP register (register offset) # --11 1101 ---- ---- ---- ---- ---- ---- # Load/store FP register (scaled imm) -FAIL 1010 010- ---- ---- 011- ---- ---- ---- # SVE contiguous FF load (scalar+scalar) -FAIL 1010 010- ---1 ---- 101- ---- ---- ---- # SVE contiguous NF load (scalar+imm) FAIL 1010 010- -01- ---- 000- ---- ---- ---- # SVE load & replicate 32 bytes (scalar+scalar) FAIL 1010 010- -010 ---- 001- ---- ---- ---- # SVE load & replicate 32 bytes (scalar+imm) diff --git a/target/arm/translate-sve.c b/target/arm/translate-sve.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-sve.c +++ b/target/arm/translate-sve.c @@ -XXX,XX +XXX,XX @@ static bool trans_LDFF1_zprr(DisasContext *s, arg_rprr_load *a) if (!dc_isar_feature(aa64_sve, s)) { return false; } + s->is_nonstreaming = true; if (sve_access_check(s)) { TCGv_i64 addr = new_tmp_a64(s); tcg_gen_shli_i64(addr, cpu_reg(s, a->rm), dtype_msz(a->dtype)); @@ -XXX,XX +XXX,XX @@ static bool trans_LDNF1_zpri(DisasContext *s, arg_rpri_load *a) if (!dc_isar_feature(aa64_sve, s)) { return false; } + s->is_nonstreaming = true; if (sve_access_check(s)) { int vsz = vec_full_reg_size(s); int elements = vsz >> dtype_esz[a->dtype]; -- 2.25.1 From: Richard Henderson <richard.henderson@linaro.org> Mark these as a non-streaming instructions, which should trap if full a64 support is not enabled in streaming mode. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20220708151540.18136-15-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/sme-fa64.decode | 3 --- target/arm/translate-sve.c | 2 ++ 2 files changed, 2 insertions(+), 3 deletions(-) diff --git a/target/arm/sme-fa64.decode b/target/arm/sme-fa64.decode index XXXXXXX..XXXXXXX 100644 --- a/target/arm/sme-fa64.decode +++ b/target/arm/sme-fa64.decode @@ -XXX,XX +XXX,XX @@ FAIL 0001 1110 0111 1110 0000 00-- ---- ---- # FJCVTZS # --11 1100 --0- ---- ---- ---- ---- ---- # Load/store FP register (unscaled imm) # --11 1100 --1- ---- ---- ---- ---- --10 # Load/store FP register (register offset) # --11 1101 ---- ---- ---- ---- ---- ---- # Load/store FP register (scaled imm) - -FAIL 1010 010- -01- ---- 000- ---- ---- ---- # SVE load & replicate 32 bytes (scalar+scalar) -FAIL 1010 010- -010 ---- 001- ---- ---- ---- # SVE load & replicate 32 bytes (scalar+imm) diff --git a/target/arm/translate-sve.c b/target/arm/translate-sve.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-sve.c +++ b/target/arm/translate-sve.c @@ -XXX,XX +XXX,XX @@ static bool trans_LD1RO_zprr(DisasContext *s, arg_rprr_load *a) if (a->rm == 31) { return false; } + s->is_nonstreaming = true; if (sve_access_check(s)) { TCGv_i64 addr = new_tmp_a64(s); tcg_gen_shli_i64(addr, cpu_reg(s, a->rm), dtype_msz(a->dtype)); @@ -XXX,XX +XXX,XX @@ static bool trans_LD1RO_zpri(DisasContext *s, arg_rpri_load *a) if (!dc_isar_feature(aa64_sve_f64mm, s)) { return false; } + s->is_nonstreaming = true; if (sve_access_check(s)) { TCGv_i64 addr = new_tmp_a64(s); tcg_gen_addi_i64(addr, cpu_reg_sp(s, a->rn), a->imm * 32); -- 2.25.1 From: Richard Henderson <richard.henderson@linaro.org> These functions will be used to verify that the cpu is in the correct state for a given instruction. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20220708151540.18136-16-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/translate-a64.h | 21 +++++++++++++++++++++ target/arm/translate-a64.c | 34 ++++++++++++++++++++++++++++++++++ 2 files changed, 55 insertions(+) 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 @@ void write_fp_dreg(DisasContext *s, int reg, TCGv_i64 v); bool logic_imm_decode_wmask(uint64_t *result, unsigned int immn, unsigned int imms, unsigned int immr); bool sve_access_check(DisasContext *s); +bool sme_enabled_check(DisasContext *s); +bool sme_enabled_check_with_svcr(DisasContext *s, unsigned); + +/* This function corresponds to CheckStreamingSVEEnabled. */ +static inline bool sme_sm_enabled_check(DisasContext *s) +{ + return sme_enabled_check_with_svcr(s, R_SVCR_SM_MASK); +} + +/* This function corresponds to CheckSMEAndZAEnabled. */ +static inline bool sme_za_enabled_check(DisasContext *s) +{ + return sme_enabled_check_with_svcr(s, R_SVCR_ZA_MASK); +} + +/* Note that this function corresponds to CheckStreamingSVEAndZAEnabled. */ +static inline bool sme_smza_enabled_check(DisasContext *s) +{ + return sme_enabled_check_with_svcr(s, R_SVCR_SM_MASK | R_SVCR_ZA_MASK); +} + TCGv_i64 clean_data_tbi(DisasContext *s, TCGv_i64 addr); TCGv_i64 gen_mte_check1(DisasContext *s, TCGv_i64 addr, bool is_write, bool tag_checked, int log2_size); diff --git a/target/arm/translate-a64.c b/target/arm/translate-a64.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-a64.c +++ b/target/arm/translate-a64.c @@ -XXX,XX +XXX,XX @@ static bool sme_access_check(DisasContext *s) return true; } +/* This function corresponds to CheckSMEEnabled. */ +bool sme_enabled_check(DisasContext *s) +{ + /* + * Note that unlike sve_excp_el, we have not constrained sme_excp_el + * to be zero when fp_excp_el has priority. This is because we need + * sme_excp_el by itself for cpregs access checks. + */ + if (!s->fp_excp_el || s->sme_excp_el < s->fp_excp_el) { + s->fp_access_checked = true; + return sme_access_check(s); + } + return fp_access_check_only(s); +} + +/* Common subroutine for CheckSMEAnd*Enabled. */ +bool sme_enabled_check_with_svcr(DisasContext *s, unsigned req) +{ + if (!sme_enabled_check(s)) { + return false; + } + if (FIELD_EX64(req, SVCR, SM) && !s->pstate_sm) { + gen_exception_insn(s, s->pc_curr, EXCP_UDEF, + syn_smetrap(SME_ET_NotStreaming, false)); + return false; + } + if (FIELD_EX64(req, SVCR, ZA) && !s->pstate_za) { + gen_exception_insn(s, s->pc_curr, EXCP_UDEF, + syn_smetrap(SME_ET_InactiveZA, false)); + 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 -- 2.25.1 From: Richard Henderson <richard.henderson@linaro.org> The pseudocode for CheckSVEEnabled gains a check for Streaming SVE mode, and for SME present but SVE absent. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20220708151540.18136-17-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/translate-a64.c | 22 ++++++++++++++++------ 1 file changed, 16 insertions(+), 6 deletions(-) diff --git a/target/arm/translate-a64.c b/target/arm/translate-a64.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-a64.c +++ b/target/arm/translate-a64.c @@ -XXX,XX +XXX,XX @@ static bool fp_access_check(DisasContext *s) return true; } -/* Check that SVE access is enabled. If it is, return true. +/* + * Check that SVE access is enabled. If it is, return true. * If not, emit code to generate an appropriate exception and return false. + * This function corresponds to CheckSVEEnabled(). */ bool sve_access_check(DisasContext *s) { - if (s->sve_excp_el) { - assert(!s->sve_access_checked); - s->sve_access_checked = true; - + if (s->pstate_sm || !dc_isar_feature(aa64_sve, s)) { + assert(dc_isar_feature(aa64_sme, s)); + if (!sme_sm_enabled_check(s)) { + goto fail_exit; + } + } else if (s->sve_excp_el) { gen_exception_insn_el(s, s->pc_curr, EXCP_UDEF, syn_sve_access_trap(), s->sve_excp_el); - return false; + goto fail_exit; } s->sve_access_checked = true; return fp_access_check(s); + + fail_exit: + /* Assert that we only raise one exception per instruction. */ + assert(!s->sve_access_checked); + s->sve_access_checked = true; + return false; } /* -- 2.25.1 From: Richard Henderson <richard.henderson@linaro.org> These SME instructions are nominally within the SVE decode space, so we add them to sve.decode and translate-sve.c. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20220708151540.18136-18-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/translate-a64.h | 12 ++++++++++++ target/arm/sve.decode | 5 ++++- target/arm/translate-sve.c | 38 ++++++++++++++++++++++++++++++++++++++ 3 files changed, 54 insertions(+), 1 deletion(-) 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 byte size of the vector register, SVL / 8. */ +static inline int streaming_vec_reg_size(DisasContext *s) +{ + return s->svl; +} + /* * Return the offset info CPUARMState of the predicate vector register Pn. * Note for this purpose, FFR is P16. @@ -XXX,XX +XXX,XX @@ static inline int pred_full_reg_size(DisasContext *s) return s->vl >> 3; } +/* Return the byte size of the predicate register, SVL / 64. */ +static inline int streaming_pred_reg_size(DisasContext *s) +{ + return s->svl >> 3; +} + /* * Round up the size of a register to a size allowed by * the tcg vector infrastructure. Any operation which uses this diff --git a/target/arm/sve.decode b/target/arm/sve.decode index XXXXXXX..XXXXXXX 100644 --- a/target/arm/sve.decode +++ b/target/arm/sve.decode @@ -XXX,XX +XXX,XX @@ INDEX_ri 00000100 esz:2 1 imm:s5 010001 rn:5 rd:5 # SVE index generation (register start, register increment) INDEX_rr 00000100 .. 1 ..... 010011 ..... ..... @rd_rn_rm -### SVE Stack Allocation Group +### SVE / Streaming SVE Stack Allocation Group # SVE stack frame adjustment ADDVL 00000100 001 ..... 01010 ...... ..... @rd_rn_i6 +ADDSVL 00000100 001 ..... 01011 ...... ..... @rd_rn_i6 ADDPL 00000100 011 ..... 01010 ...... ..... @rd_rn_i6 +ADDSPL 00000100 011 ..... 01011 ...... ..... @rd_rn_i6 # SVE stack frame size RDVL 00000100 101 11111 01010 imm:s6 rd:5 +RDSVL 00000100 101 11111 01011 imm:s6 rd:5 ### SVE Bitwise Shift - Unpredicated Group diff --git a/target/arm/translate-sve.c b/target/arm/translate-sve.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-sve.c +++ b/target/arm/translate-sve.c @@ -XXX,XX +XXX,XX @@ static bool trans_ADDVL(DisasContext *s, arg_ADDVL *a) return true; } +static bool trans_ADDSVL(DisasContext *s, arg_ADDSVL *a) +{ + if (!dc_isar_feature(aa64_sme, s)) { + return false; + } + if (sme_enabled_check(s)) { + TCGv_i64 rd = cpu_reg_sp(s, a->rd); + TCGv_i64 rn = cpu_reg_sp(s, a->rn); + tcg_gen_addi_i64(rd, rn, a->imm * streaming_vec_reg_size(s)); + } + return true; +} + static bool trans_ADDPL(DisasContext *s, arg_ADDPL *a) { if (!dc_isar_feature(aa64_sve, s)) { @@ -XXX,XX +XXX,XX @@ static bool trans_ADDPL(DisasContext *s, arg_ADDPL *a) return true; } +static bool trans_ADDSPL(DisasContext *s, arg_ADDSPL *a) +{ + if (!dc_isar_feature(aa64_sme, s)) { + return false; + } + if (sme_enabled_check(s)) { + TCGv_i64 rd = cpu_reg_sp(s, a->rd); + TCGv_i64 rn = cpu_reg_sp(s, a->rn); + tcg_gen_addi_i64(rd, rn, a->imm * streaming_pred_reg_size(s)); + } + return true; +} + static bool trans_RDVL(DisasContext *s, arg_RDVL *a) { if (!dc_isar_feature(aa64_sve, s)) { @@ -XXX,XX +XXX,XX @@ static bool trans_RDVL(DisasContext *s, arg_RDVL *a) return true; } +static bool trans_RDSVL(DisasContext *s, arg_RDSVL *a) +{ + if (!dc_isar_feature(aa64_sme, s)) { + return false; + } + if (sme_enabled_check(s)) { + TCGv_i64 reg = cpu_reg(s, a->rd); + tcg_gen_movi_i64(reg, a->imm * streaming_vec_reg_size(s)); + } + return true; +} + /* *** SVE Compute Vector Address Group */ -- 2.25.1 From: Richard Henderson <richard.henderson@linaro.org> Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20220708151540.18136-19-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/helper-sme.h | 2 ++ target/arm/sme.decode | 4 ++++ target/arm/sme_helper.c | 25 +++++++++++++++++++++++++ target/arm/translate-sme.c | 13 +++++++++++++ 4 files changed, 44 insertions(+) diff --git a/target/arm/helper-sme.h b/target/arm/helper-sme.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper-sme.h +++ b/target/arm/helper-sme.h @@ -XXX,XX +XXX,XX @@ 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) + +DEF_HELPER_FLAGS_3(sme_zero, TCG_CALL_NO_RWG, void, env, i32, i32) diff --git a/target/arm/sme.decode b/target/arm/sme.decode index XXXXXXX..XXXXXXX 100644 --- a/target/arm/sme.decode +++ b/target/arm/sme.decode @@ -XXX,XX +XXX,XX @@ # # This file is processed by scripts/decodetree.py # + +### SME Misc + +ZERO 11000000 00 001 00000000000 imm:8 diff --git a/target/arm/sme_helper.c b/target/arm/sme_helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/sme_helper.c +++ b/target/arm/sme_helper.c @@ -XXX,XX +XXX,XX @@ void helper_set_pstate_za(CPUARMState *env, uint32_t i) memset(env->zarray, 0, sizeof(env->zarray)); } } + +void helper_sme_zero(CPUARMState *env, uint32_t imm, uint32_t svl) +{ + uint32_t i; + + /* + * Special case clearing the entire ZA space. + * This falls into the CONSTRAINED UNPREDICTABLE zeroing of any + * parts of the ZA storage outside of SVL. + */ + if (imm == 0xff) { + memset(env->zarray, 0, sizeof(env->zarray)); + return; + } + + /* + * Recall that ZAnH.D[m] is spread across ZA[n+8*m], + * so each row is discontiguous within ZA[]. + */ + for (i = 0; i < svl; i++) { + if (imm & (1 << (i % 8))) { + memset(&env->zarray[i], 0, svl); + } + } +} diff --git a/target/arm/translate-sme.c b/target/arm/translate-sme.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-sme.c +++ b/target/arm/translate-sme.c @@ -XXX,XX +XXX,XX @@ */ #include "decode-sme.c.inc" + + +static bool trans_ZERO(DisasContext *s, arg_ZERO *a) +{ + if (!dc_isar_feature(aa64_sme, s)) { + return false; + } + if (sme_za_enabled_check(s)) { + gen_helper_sme_zero(cpu_env, tcg_constant_i32(a->imm), + tcg_constant_i32(streaming_vec_reg_size(s))); + } + return true; +} -- 2.25.1 From: Richard Henderson <richard.henderson@linaro.org> We can reuse the SVE functions for implementing moves to/from horizontal tile slices, but we need new ones for moves to/from vertical tile slices. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20220708151540.18136-20-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/helper-sme.h | 12 +++ target/arm/helper-sve.h | 2 + target/arm/translate-a64.h | 8 ++ target/arm/translate.h | 5 ++ target/arm/sme.decode | 15 ++++ target/arm/sme_helper.c | 151 ++++++++++++++++++++++++++++++++++++- target/arm/sve_helper.c | 12 +++ target/arm/translate-sme.c | 127 +++++++++++++++++++++++++++++++ 8 files changed, 331 insertions(+), 1 deletion(-) diff --git a/target/arm/helper-sme.h b/target/arm/helper-sme.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper-sme.h +++ b/target/arm/helper-sme.h @@ -XXX,XX +XXX,XX @@ 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) DEF_HELPER_FLAGS_3(sme_zero, TCG_CALL_NO_RWG, void, env, i32, i32) + +/* Move to/from vertical array slices, i.e. columns, so 'c'. */ +DEF_HELPER_FLAGS_4(sme_mova_cz_b, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, i32) +DEF_HELPER_FLAGS_4(sme_mova_zc_b, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, i32) +DEF_HELPER_FLAGS_4(sme_mova_cz_h, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, i32) +DEF_HELPER_FLAGS_4(sme_mova_zc_h, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, i32) +DEF_HELPER_FLAGS_4(sme_mova_cz_s, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, i32) +DEF_HELPER_FLAGS_4(sme_mova_zc_s, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, i32) +DEF_HELPER_FLAGS_4(sme_mova_cz_d, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, i32) +DEF_HELPER_FLAGS_4(sme_mova_zc_d, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, i32) +DEF_HELPER_FLAGS_4(sme_mova_cz_q, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, i32) +DEF_HELPER_FLAGS_4(sme_mova_zc_q, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, i32) diff --git a/target/arm/helper-sve.h b/target/arm/helper-sve.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper-sve.h +++ b/target/arm/helper-sve.h @@ -XXX,XX +XXX,XX @@ DEF_HELPER_FLAGS_5(sve_sel_zpzz_s, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, ptr, i32) DEF_HELPER_FLAGS_5(sve_sel_zpzz_d, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, ptr, i32) +DEF_HELPER_FLAGS_5(sve_sel_zpzz_q, TCG_CALL_NO_RWG, + void, ptr, ptr, ptr, ptr, i32) DEF_HELPER_FLAGS_5(sve2_addp_zpzz_b, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, ptr, i32) 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 pred_gvec_reg_size(DisasContext *s) return size_for_gvec(pred_full_reg_size(s)); } +/* Return a newly allocated pointer to the predicate register. */ +static inline TCGv_ptr pred_full_reg_ptr(DisasContext *s, int regno) +{ + TCGv_ptr ret = tcg_temp_new_ptr(); + tcg_gen_addi_ptr(ret, cpu_env, pred_full_reg_offset(s, regno)); + return ret; +} + bool disas_sve(DisasContext *, uint32_t); bool disas_sme(DisasContext *, uint32_t); diff --git a/target/arm/translate.h b/target/arm/translate.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate.h +++ b/target/arm/translate.h @@ -XXX,XX +XXX,XX @@ static inline int plus_2(DisasContext *s, int x) return x + 2; } +static inline int plus_12(DisasContext *s, int x) +{ + return x + 12; +} + static inline int times_2(DisasContext *s, int x) { return x * 2; diff --git a/target/arm/sme.decode b/target/arm/sme.decode index XXXXXXX..XXXXXXX 100644 --- a/target/arm/sme.decode +++ b/target/arm/sme.decode @@ -XXX,XX +XXX,XX @@ ### SME Misc ZERO 11000000 00 001 00000000000 imm:8 + +### SME Move into/from Array + +%mova_rs 13:2 !function=plus_12 +&mova esz rs pg zr za_imm v:bool to_vec:bool + +MOVA 11000000 esz:2 00000 0 v:1 .. pg:3 zr:5 0 za_imm:4 \ + &mova to_vec=0 rs=%mova_rs +MOVA 11000000 11 00000 1 v:1 .. pg:3 zr:5 0 za_imm:4 \ + &mova to_vec=0 rs=%mova_rs esz=4 + +MOVA 11000000 esz:2 00001 0 v:1 .. pg:3 0 za_imm:4 zr:5 \ + &mova to_vec=1 rs=%mova_rs +MOVA 11000000 11 00001 1 v:1 .. pg:3 0 za_imm:4 zr:5 \ + &mova to_vec=1 rs=%mova_rs esz=4 diff --git a/target/arm/sme_helper.c b/target/arm/sme_helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/sme_helper.c +++ b/target/arm/sme_helper.c @@ -XXX,XX +XXX,XX @@ #include "qemu/osdep.h" #include "cpu.h" -#include "internals.h" +#include "tcg/tcg-gvec-desc.h" #include "exec/helper-proto.h" +#include "qemu/int128.h" +#include "vec_internal.h" /* ResetSVEState */ void arm_reset_sve_state(CPUARMState *env) @@ -XXX,XX +XXX,XX @@ void helper_sme_zero(CPUARMState *env, uint32_t imm, uint32_t svl) } } } + + +/* + * When considering the ZA storage as an array of elements of + * type T, the index within that array of the Nth element of + * a vertical slice of a tile can be calculated like this, + * regardless of the size of type T. This is because the tiles + * are interleaved, so if type T is size N bytes then row 1 of + * the tile is N rows away from row 0. The division by N to + * convert a byte offset into an array index and the multiplication + * by N to convert from vslice-index-within-the-tile to + * the index within the ZA storage cancel out. + */ +#define tile_vslice_index(i) ((i) * sizeof(ARMVectorReg)) + +/* + * When doing byte arithmetic on the ZA storage, the element + * byteoff bytes away in a tile vertical slice is always this + * many bytes away in the ZA storage, regardless of the + * size of the tile element, assuming that byteoff is a multiple + * of the element size. Again this is because of the interleaving + * of the tiles. For instance if we have 1 byte per element then + * each row of the ZA storage has one byte of the vslice data, + * and (counting from 0) byte 8 goes in row 8 of the storage + * at offset (8 * row-size-in-bytes). + * If we have 8 bytes per element then each row of the ZA storage + * has 8 bytes of the data, but there are 8 interleaved tiles and + * so byte 8 of the data goes into row 1 of the tile, + * which is again row 8 of the storage, so the offset is still + * (8 * row-size-in-bytes). Similarly for other element sizes. + */ +#define tile_vslice_offset(byteoff) ((byteoff) * sizeof(ARMVectorReg)) + + +/* + * Move Zreg vector to ZArray column. + */ +#define DO_MOVA_C(NAME, TYPE, H) \ +void HELPER(NAME)(void *za, void *vn, void *vg, uint32_t desc) \ +{ \ + int i, oprsz = simd_oprsz(desc); \ + for (i = 0; i < oprsz; ) { \ + uint16_t pg = *(uint16_t *)(vg + H1_2(i >> 3)); \ + do { \ + if (pg & 1) { \ + *(TYPE *)(za + tile_vslice_offset(i)) = *(TYPE *)(vn + H(i)); \ + } \ + i += sizeof(TYPE); \ + pg >>= sizeof(TYPE); \ + } while (i & 15); \ + } \ +} + +DO_MOVA_C(sme_mova_cz_b, uint8_t, H1) +DO_MOVA_C(sme_mova_cz_h, uint16_t, H1_2) +DO_MOVA_C(sme_mova_cz_s, uint32_t, H1_4) + +void HELPER(sme_mova_cz_d)(void *za, void *vn, void *vg, uint32_t desc) +{ + int i, oprsz = simd_oprsz(desc) / 8; + uint8_t *pg = vg; + uint64_t *n = vn; + uint64_t *a = za; + + for (i = 0; i < oprsz; i++) { + if (pg[H1(i)] & 1) { + a[tile_vslice_index(i)] = n[i]; + } + } +} + +void HELPER(sme_mova_cz_q)(void *za, void *vn, void *vg, uint32_t desc) +{ + int i, oprsz = simd_oprsz(desc) / 16; + uint16_t *pg = vg; + Int128 *n = vn; + Int128 *a = za; + + /* + * Int128 is used here simply to copy 16 bytes, and to simplify + * the address arithmetic. + */ + for (i = 0; i < oprsz; i++) { + if (pg[H2(i)] & 1) { + a[tile_vslice_index(i)] = n[i]; + } + } +} + +#undef DO_MOVA_C + +/* + * Move ZArray column to Zreg vector. + */ +#define DO_MOVA_Z(NAME, TYPE, H) \ +void HELPER(NAME)(void *vd, void *za, void *vg, uint32_t desc) \ +{ \ + int i, oprsz = simd_oprsz(desc); \ + for (i = 0; i < oprsz; ) { \ + uint16_t pg = *(uint16_t *)(vg + H1_2(i >> 3)); \ + do { \ + if (pg & 1) { \ + *(TYPE *)(vd + H(i)) = *(TYPE *)(za + tile_vslice_offset(i)); \ + } \ + i += sizeof(TYPE); \ + pg >>= sizeof(TYPE); \ + } while (i & 15); \ + } \ +} + +DO_MOVA_Z(sme_mova_zc_b, uint8_t, H1) +DO_MOVA_Z(sme_mova_zc_h, uint16_t, H1_2) +DO_MOVA_Z(sme_mova_zc_s, uint32_t, H1_4) + +void HELPER(sme_mova_zc_d)(void *vd, void *za, void *vg, uint32_t desc) +{ + int i, oprsz = simd_oprsz(desc) / 8; + uint8_t *pg = vg; + uint64_t *d = vd; + uint64_t *a = za; + + for (i = 0; i < oprsz; i++) { + if (pg[H1(i)] & 1) { + d[i] = a[tile_vslice_index(i)]; + } + } +} + +void HELPER(sme_mova_zc_q)(void *vd, void *za, void *vg, uint32_t desc) +{ + int i, oprsz = simd_oprsz(desc) / 16; + uint16_t *pg = vg; + Int128 *d = vd; + Int128 *a = za; + + /* + * Int128 is used here simply to copy 16 bytes, and to simplify + * the address arithmetic. + */ + for (i = 0; i < oprsz; i++, za += sizeof(ARMVectorReg)) { + if (pg[H2(i)] & 1) { + d[i] = a[tile_vslice_index(i)]; + } + } +} + +#undef DO_MOVA_Z diff --git a/target/arm/sve_helper.c b/target/arm/sve_helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/sve_helper.c +++ b/target/arm/sve_helper.c @@ -XXX,XX +XXX,XX @@ void HELPER(sve_sel_zpzz_d)(void *vd, void *vn, void *vm, } } +void HELPER(sve_sel_zpzz_q)(void *vd, void *vn, void *vm, + void *vg, uint32_t desc) +{ + intptr_t i, opr_sz = simd_oprsz(desc) / 16; + Int128 *d = vd, *n = vn, *m = vm; + uint16_t *pg = vg; + + for (i = 0; i < opr_sz; i += 1) { + d[i] = (pg[H2(i)] & 1 ? n : m)[i]; + } +} + /* Two operand comparison controlled by a predicate. * ??? It is very tempting to want to be able to expand this inline * with x86 instructions, e.g. diff --git a/target/arm/translate-sme.c b/target/arm/translate-sme.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-sme.c +++ b/target/arm/translate-sme.c @@ -XXX,XX +XXX,XX @@ #include "decode-sme.c.inc" +/* + * Resolve tile.size[index] to a host pointer, where tile and index + * are always decoded together, dependent on the element size. + */ +static TCGv_ptr get_tile_rowcol(DisasContext *s, int esz, int rs, + int tile_index, bool vertical) +{ + int tile = tile_index >> (4 - esz); + int index = esz == MO_128 ? 0 : extract32(tile_index, 0, 4 - esz); + int pos, len, offset; + TCGv_i32 tmp; + TCGv_ptr addr; + + /* Compute the final index, which is Rs+imm. */ + tmp = tcg_temp_new_i32(); + tcg_gen_trunc_tl_i32(tmp, cpu_reg(s, rs)); + tcg_gen_addi_i32(tmp, tmp, index); + + /* Prepare a power-of-two modulo via extraction of @len bits. */ + len = ctz32(streaming_vec_reg_size(s)) - esz; + + if (vertical) { + /* + * Compute the byte offset of the index within the tile: + * (index % (svl / size)) * size + * = (index % (svl >> esz)) << esz + * Perform the power-of-two modulo via extraction of the low @len bits. + * Perform the multiply by shifting left by @pos bits. + * Perform these operations simultaneously via deposit into zero. + */ + pos = esz; + tcg_gen_deposit_z_i32(tmp, tmp, pos, len); + + /* + * For big-endian, adjust the indexed column byte offset within + * the uint64_t host words that make up env->zarray[]. + */ + if (HOST_BIG_ENDIAN && esz < MO_64) { + tcg_gen_xori_i32(tmp, tmp, 8 - (1 << esz)); + } + } else { + /* + * Compute the byte offset of the index within the tile: + * (index % (svl / size)) * (size * sizeof(row)) + * = (index % (svl >> esz)) << (esz + log2(sizeof(row))) + */ + pos = esz + ctz32(sizeof(ARMVectorReg)); + tcg_gen_deposit_z_i32(tmp, tmp, pos, len); + + /* Row slices are always aligned and need no endian adjustment. */ + } + + /* The tile byte offset within env->zarray is the row. */ + offset = tile * sizeof(ARMVectorReg); + + /* Include the byte offset of zarray to make this relative to env. */ + offset += offsetof(CPUARMState, zarray); + tcg_gen_addi_i32(tmp, tmp, offset); + + /* Add the byte offset to env to produce the final pointer. */ + addr = tcg_temp_new_ptr(); + tcg_gen_ext_i32_ptr(addr, tmp); + tcg_temp_free_i32(tmp); + tcg_gen_add_ptr(addr, addr, cpu_env); + + return addr; +} + static bool trans_ZERO(DisasContext *s, arg_ZERO *a) { if (!dc_isar_feature(aa64_sme, s)) { @@ -XXX,XX +XXX,XX @@ static bool trans_ZERO(DisasContext *s, arg_ZERO *a) } return true; } + +static bool trans_MOVA(DisasContext *s, arg_MOVA *a) +{ + static gen_helper_gvec_4 * const h_fns[5] = { + gen_helper_sve_sel_zpzz_b, gen_helper_sve_sel_zpzz_h, + gen_helper_sve_sel_zpzz_s, gen_helper_sve_sel_zpzz_d, + gen_helper_sve_sel_zpzz_q + }; + static gen_helper_gvec_3 * const cz_fns[5] = { + gen_helper_sme_mova_cz_b, gen_helper_sme_mova_cz_h, + gen_helper_sme_mova_cz_s, gen_helper_sme_mova_cz_d, + gen_helper_sme_mova_cz_q, + }; + static gen_helper_gvec_3 * const zc_fns[5] = { + gen_helper_sme_mova_zc_b, gen_helper_sme_mova_zc_h, + gen_helper_sme_mova_zc_s, gen_helper_sme_mova_zc_d, + gen_helper_sme_mova_zc_q, + }; + + TCGv_ptr t_za, t_zr, t_pg; + TCGv_i32 t_desc; + int svl; + + if (!dc_isar_feature(aa64_sme, s)) { + return false; + } + if (!sme_smza_enabled_check(s)) { + return true; + } + + t_za = get_tile_rowcol(s, a->esz, a->rs, a->za_imm, a->v); + t_zr = vec_full_reg_ptr(s, a->zr); + t_pg = pred_full_reg_ptr(s, a->pg); + + svl = streaming_vec_reg_size(s); + t_desc = tcg_constant_i32(simd_desc(svl, svl, 0)); + + if (a->v) { + /* Vertical slice -- use sme mova helpers. */ + if (a->to_vec) { + zc_fns[a->esz](t_zr, t_za, t_pg, t_desc); + } else { + cz_fns[a->esz](t_za, t_zr, t_pg, t_desc); + } + } else { + /* Horizontal slice -- reuse sve sel helpers. */ + if (a->to_vec) { + h_fns[a->esz](t_zr, t_za, t_zr, t_pg, t_desc); + } else { + h_fns[a->esz](t_za, t_zr, t_za, t_pg, t_desc); + } + } + + tcg_temp_free_ptr(t_za); + tcg_temp_free_ptr(t_zr); + tcg_temp_free_ptr(t_pg); + + return true; +} -- 2.25.1 From: Richard Henderson <richard.henderson@linaro.org> We cannot reuse the SVE functions for LD[1-4] and ST[1-4], because those functions accept only a Zreg register number. For SME, we want to pass a pointer into ZA storage. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20220708151540.18136-21-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/helper-sme.h | 82 +++++ target/arm/sme.decode | 9 + target/arm/sme_helper.c | 595 +++++++++++++++++++++++++++++++++++++ target/arm/translate-sme.c | 70 +++++ 4 files changed, 756 insertions(+) diff --git a/target/arm/helper-sme.h b/target/arm/helper-sme.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper-sme.h +++ b/target/arm/helper-sme.h @@ -XXX,XX +XXX,XX @@ DEF_HELPER_FLAGS_4(sme_mova_cz_d, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, i32) DEF_HELPER_FLAGS_4(sme_mova_zc_d, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, i32) DEF_HELPER_FLAGS_4(sme_mova_cz_q, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, i32) DEF_HELPER_FLAGS_4(sme_mova_zc_q, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, i32) + +DEF_HELPER_FLAGS_5(sme_ld1b_h, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_5(sme_ld1b_v, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_5(sme_ld1b_h_mte, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_5(sme_ld1b_v_mte, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) + +DEF_HELPER_FLAGS_5(sme_ld1h_be_h, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_5(sme_ld1h_le_h, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_5(sme_ld1h_be_v, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_5(sme_ld1h_le_v, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_5(sme_ld1h_be_h_mte, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_5(sme_ld1h_le_h_mte, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_5(sme_ld1h_be_v_mte, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_5(sme_ld1h_le_v_mte, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) + +DEF_HELPER_FLAGS_5(sme_ld1s_be_h, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_5(sme_ld1s_le_h, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_5(sme_ld1s_be_v, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_5(sme_ld1s_le_v, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_5(sme_ld1s_be_h_mte, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_5(sme_ld1s_le_h_mte, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_5(sme_ld1s_be_v_mte, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_5(sme_ld1s_le_v_mte, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) + +DEF_HELPER_FLAGS_5(sme_ld1d_be_h, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_5(sme_ld1d_le_h, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_5(sme_ld1d_be_v, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_5(sme_ld1d_le_v, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_5(sme_ld1d_be_h_mte, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_5(sme_ld1d_le_h_mte, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_5(sme_ld1d_be_v_mte, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_5(sme_ld1d_le_v_mte, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) + +DEF_HELPER_FLAGS_5(sme_ld1q_be_h, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_5(sme_ld1q_le_h, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_5(sme_ld1q_be_v, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_5(sme_ld1q_le_v, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_5(sme_ld1q_be_h_mte, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_5(sme_ld1q_le_h_mte, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_5(sme_ld1q_be_v_mte, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_5(sme_ld1q_le_v_mte, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) + +DEF_HELPER_FLAGS_5(sme_st1b_h, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_5(sme_st1b_v, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_5(sme_st1b_h_mte, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_5(sme_st1b_v_mte, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) + +DEF_HELPER_FLAGS_5(sme_st1h_be_h, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_5(sme_st1h_le_h, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_5(sme_st1h_be_v, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_5(sme_st1h_le_v, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_5(sme_st1h_be_h_mte, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_5(sme_st1h_le_h_mte, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_5(sme_st1h_be_v_mte, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_5(sme_st1h_le_v_mte, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) + +DEF_HELPER_FLAGS_5(sme_st1s_be_h, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_5(sme_st1s_le_h, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_5(sme_st1s_be_v, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_5(sme_st1s_le_v, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_5(sme_st1s_be_h_mte, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_5(sme_st1s_le_h_mte, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_5(sme_st1s_be_v_mte, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_5(sme_st1s_le_v_mte, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) + +DEF_HELPER_FLAGS_5(sme_st1d_be_h, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_5(sme_st1d_le_h, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_5(sme_st1d_be_v, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_5(sme_st1d_le_v, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_5(sme_st1d_be_h_mte, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_5(sme_st1d_le_h_mte, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_5(sme_st1d_be_v_mte, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_5(sme_st1d_le_v_mte, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) + +DEF_HELPER_FLAGS_5(sme_st1q_be_h, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_5(sme_st1q_le_h, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_5(sme_st1q_be_v, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_5(sme_st1q_le_v, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_5(sme_st1q_be_h_mte, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_5(sme_st1q_le_h_mte, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_5(sme_st1q_be_v_mte, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_5(sme_st1q_le_v_mte, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) diff --git a/target/arm/sme.decode b/target/arm/sme.decode index XXXXXXX..XXXXXXX 100644 --- a/target/arm/sme.decode +++ b/target/arm/sme.decode @@ -XXX,XX +XXX,XX @@ MOVA 11000000 esz:2 00001 0 v:1 .. pg:3 0 za_imm:4 zr:5 \ &mova to_vec=1 rs=%mova_rs MOVA 11000000 11 00001 1 v:1 .. pg:3 0 za_imm:4 zr:5 \ &mova to_vec=1 rs=%mova_rs esz=4 + +### SME Memory + +&ldst esz rs pg rn rm za_imm v:bool st:bool + +LDST1 1110000 0 esz:2 st:1 rm:5 v:1 .. pg:3 rn:5 0 za_imm:4 \ + &ldst rs=%mova_rs +LDST1 1110000 111 st:1 rm:5 v:1 .. pg:3 rn:5 0 za_imm:4 \ + &ldst esz=4 rs=%mova_rs diff --git a/target/arm/sme_helper.c b/target/arm/sme_helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/sme_helper.c +++ b/target/arm/sme_helper.c @@ -XXX,XX +XXX,XX @@ #include "qemu/osdep.h" #include "cpu.h" +#include "internals.h" #include "tcg/tcg-gvec-desc.h" #include "exec/helper-proto.h" +#include "exec/cpu_ldst.h" +#include "exec/exec-all.h" #include "qemu/int128.h" #include "vec_internal.h" +#include "sve_ldst_internal.h" /* ResetSVEState */ void arm_reset_sve_state(CPUARMState *env) @@ -XXX,XX +XXX,XX @@ void HELPER(sme_mova_zc_q)(void *vd, void *za, void *vg, uint32_t desc) } #undef DO_MOVA_Z + +/* + * Clear elements in a tile slice comprising len bytes. + */ + +typedef void ClearFn(void *ptr, size_t off, size_t len); + +static void clear_horizontal(void *ptr, size_t off, size_t len) +{ + memset(ptr + off, 0, len); +} + +static void clear_vertical_b(void *vptr, size_t off, size_t len) +{ + for (size_t i = 0; i < len; ++i) { + *(uint8_t *)(vptr + tile_vslice_offset(i + off)) = 0; + } +} + +static void clear_vertical_h(void *vptr, size_t off, size_t len) +{ + for (size_t i = 0; i < len; i += 2) { + *(uint16_t *)(vptr + tile_vslice_offset(i + off)) = 0; + } +} + +static void clear_vertical_s(void *vptr, size_t off, size_t len) +{ + for (size_t i = 0; i < len; i += 4) { + *(uint32_t *)(vptr + tile_vslice_offset(i + off)) = 0; + } +} + +static void clear_vertical_d(void *vptr, size_t off, size_t len) +{ + for (size_t i = 0; i < len; i += 8) { + *(uint64_t *)(vptr + tile_vslice_offset(i + off)) = 0; + } +} + +static void clear_vertical_q(void *vptr, size_t off, size_t len) +{ + for (size_t i = 0; i < len; i += 16) { + memset(vptr + tile_vslice_offset(i + off), 0, 16); + } +} + +/* + * Copy elements from an array into a tile slice comprising len bytes. + */ + +typedef void CopyFn(void *dst, const void *src, size_t len); + +static void copy_horizontal(void *dst, const void *src, size_t len) +{ + memcpy(dst, src, len); +} + +static void copy_vertical_b(void *vdst, const void *vsrc, size_t len) +{ + const uint8_t *src = vsrc; + uint8_t *dst = vdst; + size_t i; + + for (i = 0; i < len; ++i) { + dst[tile_vslice_index(i)] = src[i]; + } +} + +static void copy_vertical_h(void *vdst, const void *vsrc, size_t len) +{ + const uint16_t *src = vsrc; + uint16_t *dst = vdst; + size_t i; + + for (i = 0; i < len / 2; ++i) { + dst[tile_vslice_index(i)] = src[i]; + } +} + +static void copy_vertical_s(void *vdst, const void *vsrc, size_t len) +{ + const uint32_t *src = vsrc; + uint32_t *dst = vdst; + size_t i; + + for (i = 0; i < len / 4; ++i) { + dst[tile_vslice_index(i)] = src[i]; + } +} + +static void copy_vertical_d(void *vdst, const void *vsrc, size_t len) +{ + const uint64_t *src = vsrc; + uint64_t *dst = vdst; + size_t i; + + for (i = 0; i < len / 8; ++i) { + dst[tile_vslice_index(i)] = src[i]; + } +} + +static void copy_vertical_q(void *vdst, const void *vsrc, size_t len) +{ + for (size_t i = 0; i < len; i += 16) { + memcpy(vdst + tile_vslice_offset(i), vsrc + i, 16); + } +} + +/* + * Host and TLB primitives for vertical tile slice addressing. + */ + +#define DO_LD(NAME, TYPE, HOST, TLB) \ +static inline void sme_##NAME##_v_host(void *za, intptr_t off, void *host) \ +{ \ + TYPE val = HOST(host); \ + *(TYPE *)(za + tile_vslice_offset(off)) = val; \ +} \ +static inline void sme_##NAME##_v_tlb(CPUARMState *env, void *za, \ + intptr_t off, target_ulong addr, uintptr_t ra) \ +{ \ + TYPE val = TLB(env, useronly_clean_ptr(addr), ra); \ + *(TYPE *)(za + tile_vslice_offset(off)) = val; \ +} + +#define DO_ST(NAME, TYPE, HOST, TLB) \ +static inline void sme_##NAME##_v_host(void *za, intptr_t off, void *host) \ +{ \ + TYPE val = *(TYPE *)(za + tile_vslice_offset(off)); \ + HOST(host, val); \ +} \ +static inline void sme_##NAME##_v_tlb(CPUARMState *env, void *za, \ + intptr_t off, target_ulong addr, uintptr_t ra) \ +{ \ + TYPE val = *(TYPE *)(za + tile_vslice_offset(off)); \ + TLB(env, useronly_clean_ptr(addr), val, ra); \ +} + +/* + * The ARMVectorReg elements are stored in host-endian 64-bit units. + * For 128-bit quantities, the sequence defined by the Elem[] pseudocode + * corresponds to storing the two 64-bit pieces in little-endian order. + */ +#define DO_LDQ(HNAME, VNAME, BE, HOST, TLB) \ +static inline void HNAME##_host(void *za, intptr_t off, void *host) \ +{ \ + uint64_t val0 = HOST(host), val1 = HOST(host + 8); \ + uint64_t *ptr = za + off; \ + ptr[0] = BE ? val1 : val0, ptr[1] = BE ? val0 : val1; \ +} \ +static inline void VNAME##_v_host(void *za, intptr_t off, void *host) \ +{ \ + HNAME##_host(za, tile_vslice_offset(off), host); \ +} \ +static inline void HNAME##_tlb(CPUARMState *env, void *za, intptr_t off, \ + target_ulong addr, uintptr_t ra) \ +{ \ + uint64_t val0 = TLB(env, useronly_clean_ptr(addr), ra); \ + uint64_t val1 = TLB(env, useronly_clean_ptr(addr + 8), ra); \ + uint64_t *ptr = za + off; \ + ptr[0] = BE ? val1 : val0, ptr[1] = BE ? val0 : val1; \ +} \ +static inline void VNAME##_v_tlb(CPUARMState *env, void *za, intptr_t off, \ + target_ulong addr, uintptr_t ra) \ +{ \ + HNAME##_tlb(env, za, tile_vslice_offset(off), addr, ra); \ +} + +#define DO_STQ(HNAME, VNAME, BE, HOST, TLB) \ +static inline void HNAME##_host(void *za, intptr_t off, void *host) \ +{ \ + uint64_t *ptr = za + off; \ + HOST(host, ptr[BE]); \ + HOST(host + 1, ptr[!BE]); \ +} \ +static inline void VNAME##_v_host(void *za, intptr_t off, void *host) \ +{ \ + HNAME##_host(za, tile_vslice_offset(off), host); \ +} \ +static inline void HNAME##_tlb(CPUARMState *env, void *za, intptr_t off, \ + target_ulong addr, uintptr_t ra) \ +{ \ + uint64_t *ptr = za + off; \ + TLB(env, useronly_clean_ptr(addr), ptr[BE], ra); \ + TLB(env, useronly_clean_ptr(addr + 8), ptr[!BE], ra); \ +} \ +static inline void VNAME##_v_tlb(CPUARMState *env, void *za, intptr_t off, \ + target_ulong addr, uintptr_t ra) \ +{ \ + HNAME##_tlb(env, za, tile_vslice_offset(off), addr, ra); \ +} + +DO_LD(ld1b, uint8_t, ldub_p, cpu_ldub_data_ra) +DO_LD(ld1h_be, uint16_t, lduw_be_p, cpu_lduw_be_data_ra) +DO_LD(ld1h_le, uint16_t, lduw_le_p, cpu_lduw_le_data_ra) +DO_LD(ld1s_be, uint32_t, ldl_be_p, cpu_ldl_be_data_ra) +DO_LD(ld1s_le, uint32_t, ldl_le_p, cpu_ldl_le_data_ra) +DO_LD(ld1d_be, uint64_t, ldq_be_p, cpu_ldq_be_data_ra) +DO_LD(ld1d_le, uint64_t, ldq_le_p, cpu_ldq_le_data_ra) + +DO_LDQ(sve_ld1qq_be, sme_ld1q_be, 1, ldq_be_p, cpu_ldq_be_data_ra) +DO_LDQ(sve_ld1qq_le, sme_ld1q_le, 0, ldq_le_p, cpu_ldq_le_data_ra) + +DO_ST(st1b, uint8_t, stb_p, cpu_stb_data_ra) +DO_ST(st1h_be, uint16_t, stw_be_p, cpu_stw_be_data_ra) +DO_ST(st1h_le, uint16_t, stw_le_p, cpu_stw_le_data_ra) +DO_ST(st1s_be, uint32_t, stl_be_p, cpu_stl_be_data_ra) +DO_ST(st1s_le, uint32_t, stl_le_p, cpu_stl_le_data_ra) +DO_ST(st1d_be, uint64_t, stq_be_p, cpu_stq_be_data_ra) +DO_ST(st1d_le, uint64_t, stq_le_p, cpu_stq_le_data_ra) + +DO_STQ(sve_st1qq_be, sme_st1q_be, 1, stq_be_p, cpu_stq_be_data_ra) +DO_STQ(sve_st1qq_le, sme_st1q_le, 0, stq_le_p, cpu_stq_le_data_ra) + +#undef DO_LD +#undef DO_ST +#undef DO_LDQ +#undef DO_STQ + +/* + * Common helper for all contiguous predicated loads. + */ + +static inline QEMU_ALWAYS_INLINE +void sme_ld1(CPUARMState *env, void *za, uint64_t *vg, + const target_ulong addr, uint32_t desc, const uintptr_t ra, + const int esz, uint32_t mtedesc, bool vertical, + sve_ldst1_host_fn *host_fn, + sve_ldst1_tlb_fn *tlb_fn, + ClearFn *clr_fn, + CopyFn *cpy_fn) +{ + const intptr_t reg_max = simd_oprsz(desc); + const intptr_t esize = 1 << esz; + intptr_t reg_off, reg_last; + SVEContLdSt info; + void *host; + int flags; + + /* Find the active elements. */ + if (!sve_cont_ldst_elements(&info, addr, vg, reg_max, esz, esize)) { + /* The entire predicate was false; no load occurs. */ + clr_fn(za, 0, reg_max); + return; + } + + /* Probe the page(s). Exit with exception for any invalid page. */ + sve_cont_ldst_pages(&info, FAULT_ALL, env, addr, MMU_DATA_LOAD, ra); + + /* Handle watchpoints for all active elements. */ + sve_cont_ldst_watchpoints(&info, env, vg, addr, esize, esize, + BP_MEM_READ, ra); + + /* + * Handle mte checks for all active elements. + * Since TBI must be set for MTE, !mtedesc => !mte_active. + */ + if (mtedesc) { + sve_cont_ldst_mte_check(&info, env, vg, addr, esize, esize, + mtedesc, ra); + } + + flags = info.page[0].flags | info.page[1].flags; + if (unlikely(flags != 0)) { +#ifdef CONFIG_USER_ONLY + g_assert_not_reached(); +#else + /* + * At least one page includes MMIO. + * Any bus operation can fail with cpu_transaction_failed, + * which for ARM will raise SyncExternal. Perform the load + * into scratch memory to preserve register state until the end. + */ + ARMVectorReg scratch = { }; + + reg_off = info.reg_off_first[0]; + reg_last = info.reg_off_last[1]; + if (reg_last < 0) { + reg_last = info.reg_off_split; + if (reg_last < 0) { + reg_last = info.reg_off_last[0]; + } + } + + do { + uint64_t pg = vg[reg_off >> 6]; + do { + if ((pg >> (reg_off & 63)) & 1) { + tlb_fn(env, &scratch, reg_off, addr + reg_off, ra); + } + reg_off += esize; + } while (reg_off & 63); + } while (reg_off <= reg_last); + + cpy_fn(za, &scratch, reg_max); + return; +#endif + } + + /* The entire operation is in RAM, on valid pages. */ + + reg_off = info.reg_off_first[0]; + reg_last = info.reg_off_last[0]; + host = info.page[0].host; + + if (!vertical) { + memset(za, 0, reg_max); + } else if (reg_off) { + clr_fn(za, 0, reg_off); + } + + while (reg_off <= reg_last) { + uint64_t pg = vg[reg_off >> 6]; + do { + if ((pg >> (reg_off & 63)) & 1) { + host_fn(za, reg_off, host + reg_off); + } else if (vertical) { + clr_fn(za, reg_off, esize); + } + reg_off += esize; + } while (reg_off <= reg_last && (reg_off & 63)); + } + + /* + * Use the slow path to manage the cross-page misalignment. + * But we know this is RAM and cannot trap. + */ + reg_off = info.reg_off_split; + if (unlikely(reg_off >= 0)) { + tlb_fn(env, za, reg_off, addr + reg_off, ra); + } + + reg_off = info.reg_off_first[1]; + if (unlikely(reg_off >= 0)) { + reg_last = info.reg_off_last[1]; + host = info.page[1].host; + + do { + uint64_t pg = vg[reg_off >> 6]; + do { + if ((pg >> (reg_off & 63)) & 1) { + host_fn(za, reg_off, host + reg_off); + } else if (vertical) { + clr_fn(za, reg_off, esize); + } + reg_off += esize; + } while (reg_off & 63); + } while (reg_off <= reg_last); + } +} + +static inline QEMU_ALWAYS_INLINE +void sme_ld1_mte(CPUARMState *env, void *za, uint64_t *vg, + target_ulong addr, uint32_t desc, uintptr_t ra, + const int esz, bool vertical, + sve_ldst1_host_fn *host_fn, + sve_ldst1_tlb_fn *tlb_fn, + ClearFn *clr_fn, + CopyFn *cpy_fn) +{ + uint32_t mtedesc = desc >> (SIMD_DATA_SHIFT + SVE_MTEDESC_SHIFT); + int bit55 = extract64(addr, 55, 1); + + /* Remove mtedesc from the normal sve descriptor. */ + desc = extract32(desc, 0, SIMD_DATA_SHIFT + SVE_MTEDESC_SHIFT); + + /* Perform gross MTE suppression early. */ + if (!tbi_check(desc, bit55) || + tcma_check(desc, bit55, allocation_tag_from_addr(addr))) { + mtedesc = 0; + } + + sme_ld1(env, za, vg, addr, desc, ra, esz, mtedesc, vertical, + host_fn, tlb_fn, clr_fn, cpy_fn); +} + +#define DO_LD(L, END, ESZ) \ +void HELPER(sme_ld1##L##END##_h)(CPUARMState *env, void *za, void *vg, \ + target_ulong addr, uint32_t desc) \ +{ \ + sme_ld1(env, za, vg, addr, desc, GETPC(), ESZ, 0, false, \ + sve_ld1##L##L##END##_host, sve_ld1##L##L##END##_tlb, \ + clear_horizontal, copy_horizontal); \ +} \ +void HELPER(sme_ld1##L##END##_v)(CPUARMState *env, void *za, void *vg, \ + target_ulong addr, uint32_t desc) \ +{ \ + sme_ld1(env, za, vg, addr, desc, GETPC(), ESZ, 0, true, \ + sme_ld1##L##END##_v_host, sme_ld1##L##END##_v_tlb, \ + clear_vertical_##L, copy_vertical_##L); \ +} \ +void HELPER(sme_ld1##L##END##_h_mte)(CPUARMState *env, void *za, void *vg, \ + target_ulong addr, uint32_t desc) \ +{ \ + sme_ld1_mte(env, za, vg, addr, desc, GETPC(), ESZ, false, \ + sve_ld1##L##L##END##_host, sve_ld1##L##L##END##_tlb, \ + clear_horizontal, copy_horizontal); \ +} \ +void HELPER(sme_ld1##L##END##_v_mte)(CPUARMState *env, void *za, void *vg, \ + target_ulong addr, uint32_t desc) \ +{ \ + sme_ld1_mte(env, za, vg, addr, desc, GETPC(), ESZ, true, \ + sme_ld1##L##END##_v_host, sme_ld1##L##END##_v_tlb, \ + clear_vertical_##L, copy_vertical_##L); \ +} + +DO_LD(b, , MO_8) +DO_LD(h, _be, MO_16) +DO_LD(h, _le, MO_16) +DO_LD(s, _be, MO_32) +DO_LD(s, _le, MO_32) +DO_LD(d, _be, MO_64) +DO_LD(d, _le, MO_64) +DO_LD(q, _be, MO_128) +DO_LD(q, _le, MO_128) + +#undef DO_LD + +/* + * Common helper for all contiguous predicated stores. + */ + +static inline QEMU_ALWAYS_INLINE +void sme_st1(CPUARMState *env, void *za, uint64_t *vg, + const target_ulong addr, uint32_t desc, const uintptr_t ra, + const int esz, uint32_t mtedesc, bool vertical, + sve_ldst1_host_fn *host_fn, + sve_ldst1_tlb_fn *tlb_fn) +{ + const intptr_t reg_max = simd_oprsz(desc); + const intptr_t esize = 1 << esz; + intptr_t reg_off, reg_last; + SVEContLdSt info; + void *host; + int flags; + + /* Find the active elements. */ + if (!sve_cont_ldst_elements(&info, addr, vg, reg_max, esz, esize)) { + /* The entire predicate was false; no store occurs. */ + return; + } + + /* Probe the page(s). Exit with exception for any invalid page. */ + sve_cont_ldst_pages(&info, FAULT_ALL, env, addr, MMU_DATA_STORE, ra); + + /* Handle watchpoints for all active elements. */ + sve_cont_ldst_watchpoints(&info, env, vg, addr, esize, esize, + BP_MEM_WRITE, ra); + + /* + * Handle mte checks for all active elements. + * Since TBI must be set for MTE, !mtedesc => !mte_active. + */ + if (mtedesc) { + sve_cont_ldst_mte_check(&info, env, vg, addr, esize, esize, + mtedesc, ra); + } + + flags = info.page[0].flags | info.page[1].flags; + if (unlikely(flags != 0)) { +#ifdef CONFIG_USER_ONLY + g_assert_not_reached(); +#else + /* + * At least one page includes MMIO. + * Any bus operation can fail with cpu_transaction_failed, + * which for ARM will raise SyncExternal. We cannot avoid + * this fault and will leave with the store incomplete. + */ + reg_off = info.reg_off_first[0]; + reg_last = info.reg_off_last[1]; + if (reg_last < 0) { + reg_last = info.reg_off_split; + if (reg_last < 0) { + reg_last = info.reg_off_last[0]; + } + } + + do { + uint64_t pg = vg[reg_off >> 6]; + do { + if ((pg >> (reg_off & 63)) & 1) { + tlb_fn(env, za, reg_off, addr + reg_off, ra); + } + reg_off += esize; + } while (reg_off & 63); + } while (reg_off <= reg_last); + return; +#endif + } + + reg_off = info.reg_off_first[0]; + reg_last = info.reg_off_last[0]; + host = info.page[0].host; + + while (reg_off <= reg_last) { + uint64_t pg = vg[reg_off >> 6]; + do { + if ((pg >> (reg_off & 63)) & 1) { + host_fn(za, reg_off, host + reg_off); + } + reg_off += 1 << esz; + } while (reg_off <= reg_last && (reg_off & 63)); + } + + /* + * Use the slow path to manage the cross-page misalignment. + * But we know this is RAM and cannot trap. + */ + reg_off = info.reg_off_split; + if (unlikely(reg_off >= 0)) { + tlb_fn(env, za, reg_off, addr + reg_off, ra); + } + + reg_off = info.reg_off_first[1]; + if (unlikely(reg_off >= 0)) { + reg_last = info.reg_off_last[1]; + host = info.page[1].host; + + do { + uint64_t pg = vg[reg_off >> 6]; + do { + if ((pg >> (reg_off & 63)) & 1) { + host_fn(za, reg_off, host + reg_off); + } + reg_off += 1 << esz; + } while (reg_off & 63); + } while (reg_off <= reg_last); + } +} + +static inline QEMU_ALWAYS_INLINE +void sme_st1_mte(CPUARMState *env, void *za, uint64_t *vg, target_ulong addr, + uint32_t desc, uintptr_t ra, int esz, bool vertical, + sve_ldst1_host_fn *host_fn, + sve_ldst1_tlb_fn *tlb_fn) +{ + uint32_t mtedesc = desc >> (SIMD_DATA_SHIFT + SVE_MTEDESC_SHIFT); + int bit55 = extract64(addr, 55, 1); + + /* Remove mtedesc from the normal sve descriptor. */ + desc = extract32(desc, 0, SIMD_DATA_SHIFT + SVE_MTEDESC_SHIFT); + + /* Perform gross MTE suppression early. */ + if (!tbi_check(desc, bit55) || + tcma_check(desc, bit55, allocation_tag_from_addr(addr))) { + mtedesc = 0; + } + + sme_st1(env, za, vg, addr, desc, ra, esz, mtedesc, + vertical, host_fn, tlb_fn); +} + +#define DO_ST(L, END, ESZ) \ +void HELPER(sme_st1##L##END##_h)(CPUARMState *env, void *za, void *vg, \ + target_ulong addr, uint32_t desc) \ +{ \ + sme_st1(env, za, vg, addr, desc, GETPC(), ESZ, 0, false, \ + sve_st1##L##L##END##_host, sve_st1##L##L##END##_tlb); \ +} \ +void HELPER(sme_st1##L##END##_v)(CPUARMState *env, void *za, void *vg, \ + target_ulong addr, uint32_t desc) \ +{ \ + sme_st1(env, za, vg, addr, desc, GETPC(), ESZ, 0, true, \ + sme_st1##L##END##_v_host, sme_st1##L##END##_v_tlb); \ +} \ +void HELPER(sme_st1##L##END##_h_mte)(CPUARMState *env, void *za, void *vg, \ + target_ulong addr, uint32_t desc) \ +{ \ + sme_st1_mte(env, za, vg, addr, desc, GETPC(), ESZ, false, \ + sve_st1##L##L##END##_host, sve_st1##L##L##END##_tlb); \ +} \ +void HELPER(sme_st1##L##END##_v_mte)(CPUARMState *env, void *za, void *vg, \ + target_ulong addr, uint32_t desc) \ +{ \ + sme_st1_mte(env, za, vg, addr, desc, GETPC(), ESZ, true, \ + sme_st1##L##END##_v_host, sme_st1##L##END##_v_tlb); \ +} + +DO_ST(b, , MO_8) +DO_ST(h, _be, MO_16) +DO_ST(h, _le, MO_16) +DO_ST(s, _be, MO_32) +DO_ST(s, _le, MO_32) +DO_ST(d, _be, MO_64) +DO_ST(d, _le, MO_64) +DO_ST(q, _be, MO_128) +DO_ST(q, _le, MO_128) + +#undef DO_ST diff --git a/target/arm/translate-sme.c b/target/arm/translate-sme.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-sme.c +++ b/target/arm/translate-sme.c @@ -XXX,XX +XXX,XX @@ static bool trans_MOVA(DisasContext *s, arg_MOVA *a) return true; } + +static bool trans_LDST1(DisasContext *s, arg_LDST1 *a) +{ + typedef void GenLdSt1(TCGv_env, TCGv_ptr, TCGv_ptr, TCGv, TCGv_i32); + + /* + * Indexed by [esz][be][v][mte][st], which is (except for load/store) + * also the order in which the elements appear in the function names, + * and so how we must concatenate the pieces. + */ + +#define FN_LS(F) { gen_helper_sme_ld1##F, gen_helper_sme_st1##F } +#define FN_MTE(F) { FN_LS(F), FN_LS(F##_mte) } +#define FN_HV(F) { FN_MTE(F##_h), FN_MTE(F##_v) } +#define FN_END(L, B) { FN_HV(L), FN_HV(B) } + + static GenLdSt1 * const fns[5][2][2][2][2] = { + FN_END(b, b), + FN_END(h_le, h_be), + FN_END(s_le, s_be), + FN_END(d_le, d_be), + FN_END(q_le, q_be), + }; + +#undef FN_LS +#undef FN_MTE +#undef FN_HV +#undef FN_END + + TCGv_ptr t_za, t_pg; + TCGv_i64 addr; + int svl, desc = 0; + bool be = s->be_data == MO_BE; + bool mte = s->mte_active[0]; + + if (!dc_isar_feature(aa64_sme, s)) { + return false; + } + if (!sme_smza_enabled_check(s)) { + return true; + } + + t_za = get_tile_rowcol(s, a->esz, a->rs, a->za_imm, a->v); + t_pg = pred_full_reg_ptr(s, a->pg); + addr = tcg_temp_new_i64(); + + tcg_gen_shli_i64(addr, cpu_reg(s, a->rm), a->esz); + tcg_gen_add_i64(addr, addr, cpu_reg_sp(s, a->rn)); + + if (mte) { + desc = FIELD_DP32(desc, MTEDESC, MIDX, get_mem_index(s)); + desc = FIELD_DP32(desc, MTEDESC, TBI, s->tbid); + desc = FIELD_DP32(desc, MTEDESC, TCMA, s->tcma); + desc = FIELD_DP32(desc, MTEDESC, WRITE, a->st); + desc = FIELD_DP32(desc, MTEDESC, SIZEM1, (1 << a->esz) - 1); + desc <<= SVE_MTEDESC_SHIFT; + } else { + addr = clean_data_tbi(s, addr); + } + svl = streaming_vec_reg_size(s); + desc = simd_desc(svl, svl, desc); + + fns[a->esz][be][a->v][mte][a->st](cpu_env, t_za, t_pg, addr, + tcg_constant_i32(desc)); + + tcg_temp_free_ptr(t_za); + tcg_temp_free_ptr(t_pg); + tcg_temp_free_i64(addr); + return true; +} -- 2.25.1 From: Richard Henderson <richard.henderson@linaro.org> Add a TCGv_ptr base argument, which will be cpu_env for SVE. We will reuse this for SME save and restore array insns. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20220708151540.18136-22-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/translate-a64.h | 3 +++ target/arm/translate-sve.c | 48 ++++++++++++++++++++++++++++---------- 2 files changed, 39 insertions(+), 12 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 @@ void gen_gvec_xar(unsigned vece, uint32_t rd_ofs, uint32_t rn_ofs, uint32_t rm_ofs, int64_t shift, uint32_t opr_sz, uint32_t max_sz); +void gen_sve_ldr(DisasContext *s, TCGv_ptr, int vofs, int len, int rn, int imm); +void gen_sve_str(DisasContext *s, TCGv_ptr, int vofs, int len, int rn, int imm); + #endif /* TARGET_ARM_TRANSLATE_A64_H */ 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 @@ TRANS_FEAT(UCVTF_dd, aa64_sve, gen_gvec_fpst_arg_zpz, * The load should begin at the address Rn + IMM. */ -static void do_ldr(DisasContext *s, uint32_t vofs, int len, int rn, int imm) +void gen_sve_ldr(DisasContext *s, TCGv_ptr base, int vofs, + int len, int rn, int imm) { int len_align = QEMU_ALIGN_DOWN(len, 8); int len_remain = len % 8; @@ -XXX,XX +XXX,XX @@ static void do_ldr(DisasContext *s, uint32_t vofs, int len, int rn, int imm) t0 = tcg_temp_new_i64(); for (i = 0; i < len_align; i += 8) { tcg_gen_qemu_ld_i64(t0, clean_addr, midx, MO_LEUQ); - tcg_gen_st_i64(t0, cpu_env, vofs + i); + tcg_gen_st_i64(t0, base, vofs + i); tcg_gen_addi_i64(clean_addr, clean_addr, 8); } tcg_temp_free_i64(t0); @@ -XXX,XX +XXX,XX @@ static void do_ldr(DisasContext *s, uint32_t vofs, int len, int rn, int imm) clean_addr = new_tmp_a64_local(s); tcg_gen_mov_i64(clean_addr, t0); + if (base != cpu_env) { + TCGv_ptr b = tcg_temp_local_new_ptr(); + tcg_gen_mov_ptr(b, base); + base = b; + } + gen_set_label(loop); t0 = tcg_temp_new_i64(); @@ -XXX,XX +XXX,XX @@ static void do_ldr(DisasContext *s, uint32_t vofs, int len, int rn, int imm) tcg_gen_addi_i64(clean_addr, clean_addr, 8); tp = tcg_temp_new_ptr(); - tcg_gen_add_ptr(tp, cpu_env, i); + tcg_gen_add_ptr(tp, base, i); tcg_gen_addi_ptr(i, i, 8); tcg_gen_st_i64(t0, tp, vofs); tcg_temp_free_ptr(tp); @@ -XXX,XX +XXX,XX @@ static void do_ldr(DisasContext *s, uint32_t vofs, int len, int rn, int imm) tcg_gen_brcondi_ptr(TCG_COND_LTU, i, len_align, loop); tcg_temp_free_ptr(i); + + if (base != cpu_env) { + tcg_temp_free_ptr(base); + assert(len_remain == 0); + } } /* @@ -XXX,XX +XXX,XX @@ static void do_ldr(DisasContext *s, uint32_t vofs, int len, int rn, int imm) default: g_assert_not_reached(); } - tcg_gen_st_i64(t0, cpu_env, vofs + len_align); + tcg_gen_st_i64(t0, base, vofs + len_align); tcg_temp_free_i64(t0); } } /* Similarly for stores. */ -static void do_str(DisasContext *s, uint32_t vofs, int len, int rn, int imm) +void gen_sve_str(DisasContext *s, TCGv_ptr base, int vofs, + int len, int rn, int imm) { int len_align = QEMU_ALIGN_DOWN(len, 8); int len_remain = len % 8; @@ -XXX,XX +XXX,XX @@ static void do_str(DisasContext *s, uint32_t vofs, int len, int rn, int imm) t0 = tcg_temp_new_i64(); for (i = 0; i < len_align; i += 8) { - tcg_gen_ld_i64(t0, cpu_env, vofs + i); + tcg_gen_ld_i64(t0, base, vofs + i); tcg_gen_qemu_st_i64(t0, clean_addr, midx, MO_LEUQ); tcg_gen_addi_i64(clean_addr, clean_addr, 8); } @@ -XXX,XX +XXX,XX @@ static void do_str(DisasContext *s, uint32_t vofs, int len, int rn, int imm) clean_addr = new_tmp_a64_local(s); tcg_gen_mov_i64(clean_addr, t0); + if (base != cpu_env) { + TCGv_ptr b = tcg_temp_local_new_ptr(); + tcg_gen_mov_ptr(b, base); + base = b; + } + gen_set_label(loop); t0 = tcg_temp_new_i64(); tp = tcg_temp_new_ptr(); - tcg_gen_add_ptr(tp, cpu_env, i); + tcg_gen_add_ptr(tp, base, i); tcg_gen_ld_i64(t0, tp, vofs); tcg_gen_addi_ptr(i, i, 8); tcg_temp_free_ptr(tp); @@ -XXX,XX +XXX,XX @@ static void do_str(DisasContext *s, uint32_t vofs, int len, int rn, int imm) tcg_gen_brcondi_ptr(TCG_COND_LTU, i, len_align, loop); tcg_temp_free_ptr(i); + + if (base != cpu_env) { + tcg_temp_free_ptr(base); + assert(len_remain == 0); + } } /* Predicate register stores can be any multiple of 2. */ if (len_remain) { t0 = tcg_temp_new_i64(); - tcg_gen_ld_i64(t0, cpu_env, vofs + len_align); + tcg_gen_ld_i64(t0, base, vofs + len_align); switch (len_remain) { case 2: @@ -XXX,XX +XXX,XX @@ static bool trans_LDR_zri(DisasContext *s, arg_rri *a) if (sve_access_check(s)) { int size = vec_full_reg_size(s); int off = vec_full_reg_offset(s, a->rd); - do_ldr(s, off, size, a->rn, a->imm * size); + gen_sve_ldr(s, cpu_env, off, size, a->rn, a->imm * size); } return true; } @@ -XXX,XX +XXX,XX @@ static bool trans_LDR_pri(DisasContext *s, arg_rri *a) if (sve_access_check(s)) { int size = pred_full_reg_size(s); int off = pred_full_reg_offset(s, a->rd); - do_ldr(s, off, size, a->rn, a->imm * size); + gen_sve_ldr(s, cpu_env, off, size, a->rn, a->imm * size); } return true; } @@ -XXX,XX +XXX,XX @@ static bool trans_STR_zri(DisasContext *s, arg_rri *a) if (sve_access_check(s)) { int size = vec_full_reg_size(s); int off = vec_full_reg_offset(s, a->rd); - do_str(s, off, size, a->rn, a->imm * size); + gen_sve_str(s, cpu_env, off, size, a->rn, a->imm * size); } return true; } @@ -XXX,XX +XXX,XX @@ static bool trans_STR_pri(DisasContext *s, arg_rri *a) if (sve_access_check(s)) { int size = pred_full_reg_size(s); int off = pred_full_reg_offset(s, a->rd); - do_str(s, off, size, a->rn, a->imm * size); + gen_sve_str(s, cpu_env, off, size, a->rn, a->imm * size); } return true; } -- 2.25.1 From: Richard Henderson <richard.henderson@linaro.org> We can reuse the SVE functions for LDR and STR, passing in the base of the ZA vector and a zero offset. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20220708151540.18136-23-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/sme.decode | 7 +++++++ target/arm/translate-sme.c | 24 ++++++++++++++++++++++++ 2 files changed, 31 insertions(+) diff --git a/target/arm/sme.decode b/target/arm/sme.decode index XXXXXXX..XXXXXXX 100644 --- a/target/arm/sme.decode +++ b/target/arm/sme.decode @@ -XXX,XX +XXX,XX @@ LDST1 1110000 0 esz:2 st:1 rm:5 v:1 .. pg:3 rn:5 0 za_imm:4 \ &ldst rs=%mova_rs LDST1 1110000 111 st:1 rm:5 v:1 .. pg:3 rn:5 0 za_imm:4 \ &ldst esz=4 rs=%mova_rs + +&ldstr rv rn imm +@ldstr ....... ... . ...... .. ... rn:5 . imm:4 \ + &ldstr rv=%mova_rs + +LDR 1110000 100 0 000000 .. 000 ..... 0 .... @ldstr +STR 1110000 100 1 000000 .. 000 ..... 0 .... @ldstr diff --git a/target/arm/translate-sme.c b/target/arm/translate-sme.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-sme.c +++ b/target/arm/translate-sme.c @@ -XXX,XX +XXX,XX @@ static bool trans_LDST1(DisasContext *s, arg_LDST1 *a) tcg_temp_free_i64(addr); return true; } + +typedef void GenLdStR(DisasContext *, TCGv_ptr, int, int, int, int); + +static bool do_ldst_r(DisasContext *s, arg_ldstr *a, GenLdStR *fn) +{ + int svl = streaming_vec_reg_size(s); + int imm = a->imm; + TCGv_ptr base; + + if (!sme_za_enabled_check(s)) { + return true; + } + + /* ZA[n] equates to ZA0H.B[n]. */ + base = get_tile_rowcol(s, MO_8, a->rv, imm, false); + + fn(s, base, 0, svl, a->rn, imm * svl); + + tcg_temp_free_ptr(base); + return true; +} + +TRANS_FEAT(LDR, aa64_sme, do_ldst_r, a, gen_sve_ldr) +TRANS_FEAT(STR, aa64_sme, do_ldst_r, a, gen_sve_str) -- 2.25.1 From: Richard Henderson <richard.henderson@linaro.org> Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20220708151540.18136-24-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/helper-sme.h | 5 +++ target/arm/sme.decode | 11 +++++ target/arm/sme_helper.c | 90 ++++++++++++++++++++++++++++++++++++++ target/arm/translate-sme.c | 31 +++++++++++++ 4 files changed, 137 insertions(+) diff --git a/target/arm/helper-sme.h b/target/arm/helper-sme.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper-sme.h +++ b/target/arm/helper-sme.h @@ -XXX,XX +XXX,XX @@ DEF_HELPER_FLAGS_5(sme_st1q_be_h_mte, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i DEF_HELPER_FLAGS_5(sme_st1q_le_h_mte, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) DEF_HELPER_FLAGS_5(sme_st1q_be_v_mte, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) DEF_HELPER_FLAGS_5(sme_st1q_le_v_mte, TCG_CALL_NO_WG, void, env, ptr, ptr, tl, i32) + +DEF_HELPER_FLAGS_5(sme_addha_s, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, ptr, i32) +DEF_HELPER_FLAGS_5(sme_addva_s, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, ptr, i32) +DEF_HELPER_FLAGS_5(sme_addha_d, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, ptr, i32) +DEF_HELPER_FLAGS_5(sme_addva_d, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, ptr, i32) diff --git a/target/arm/sme.decode b/target/arm/sme.decode index XXXXXXX..XXXXXXX 100644 --- a/target/arm/sme.decode +++ b/target/arm/sme.decode @@ -XXX,XX +XXX,XX @@ LDST1 1110000 111 st:1 rm:5 v:1 .. pg:3 rn:5 0 za_imm:4 \ LDR 1110000 100 0 000000 .. 000 ..... 0 .... @ldstr STR 1110000 100 1 000000 .. 000 ..... 0 .... @ldstr + +### SME Add Vector to Array + +&adda zad zn pm pn +@adda_32 ........ .. ..... . pm:3 pn:3 zn:5 ... zad:2 &adda +@adda_64 ........ .. ..... . pm:3 pn:3 zn:5 .. zad:3 &adda + +ADDHA_s 11000000 10 01000 0 ... ... ..... 000 .. @adda_32 +ADDVA_s 11000000 10 01000 1 ... ... ..... 000 .. @adda_32 +ADDHA_d 11000000 11 01000 0 ... ... ..... 00 ... @adda_64 +ADDVA_d 11000000 11 01000 1 ... ... ..... 00 ... @adda_64 diff --git a/target/arm/sme_helper.c b/target/arm/sme_helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/sme_helper.c +++ b/target/arm/sme_helper.c @@ -XXX,XX +XXX,XX @@ DO_ST(q, _be, MO_128) DO_ST(q, _le, MO_128) #undef DO_ST + +void HELPER(sme_addha_s)(void *vzda, void *vzn, void *vpn, + void *vpm, uint32_t desc) +{ + intptr_t row, col, oprsz = simd_oprsz(desc) / 4; + uint64_t *pn = vpn, *pm = vpm; + uint32_t *zda = vzda, *zn = vzn; + + for (row = 0; row < oprsz; ) { + uint64_t pa = pn[row >> 4]; + do { + if (pa & 1) { + for (col = 0; col < oprsz; ) { + uint64_t pb = pm[col >> 4]; + do { + if (pb & 1) { + zda[tile_vslice_index(row) + H4(col)] += zn[H4(col)]; + } + pb >>= 4; + } while (++col & 15); + } + } + pa >>= 4; + } while (++row & 15); + } +} + +void HELPER(sme_addha_d)(void *vzda, void *vzn, void *vpn, + void *vpm, uint32_t desc) +{ + intptr_t row, col, oprsz = simd_oprsz(desc) / 8; + uint8_t *pn = vpn, *pm = vpm; + uint64_t *zda = vzda, *zn = vzn; + + for (row = 0; row < oprsz; ++row) { + if (pn[H1(row)] & 1) { + for (col = 0; col < oprsz; ++col) { + if (pm[H1(col)] & 1) { + zda[tile_vslice_index(row) + col] += zn[col]; + } + } + } + } +} + +void HELPER(sme_addva_s)(void *vzda, void *vzn, void *vpn, + void *vpm, uint32_t desc) +{ + intptr_t row, col, oprsz = simd_oprsz(desc) / 4; + uint64_t *pn = vpn, *pm = vpm; + uint32_t *zda = vzda, *zn = vzn; + + for (row = 0; row < oprsz; ) { + uint64_t pa = pn[row >> 4]; + do { + if (pa & 1) { + uint32_t zn_row = zn[H4(row)]; + for (col = 0; col < oprsz; ) { + uint64_t pb = pm[col >> 4]; + do { + if (pb & 1) { + zda[tile_vslice_index(row) + H4(col)] += zn_row; + } + pb >>= 4; + } while (++col & 15); + } + } + pa >>= 4; + } while (++row & 15); + } +} + +void HELPER(sme_addva_d)(void *vzda, void *vzn, void *vpn, + void *vpm, uint32_t desc) +{ + intptr_t row, col, oprsz = simd_oprsz(desc) / 8; + uint8_t *pn = vpn, *pm = vpm; + uint64_t *zda = vzda, *zn = vzn; + + for (row = 0; row < oprsz; ++row) { + if (pn[H1(row)] & 1) { + uint64_t zn_row = zn[row]; + for (col = 0; col < oprsz; ++col) { + if (pm[H1(col)] & 1) { + zda[tile_vslice_index(row) + col] += zn_row; + } + } + } + } +} diff --git a/target/arm/translate-sme.c b/target/arm/translate-sme.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-sme.c +++ b/target/arm/translate-sme.c @@ -XXX,XX +XXX,XX @@ static bool do_ldst_r(DisasContext *s, arg_ldstr *a, GenLdStR *fn) TRANS_FEAT(LDR, aa64_sme, do_ldst_r, a, gen_sve_ldr) TRANS_FEAT(STR, aa64_sme, do_ldst_r, a, gen_sve_str) + +static bool do_adda(DisasContext *s, arg_adda *a, MemOp esz, + gen_helper_gvec_4 *fn) +{ + int svl = streaming_vec_reg_size(s); + uint32_t desc = simd_desc(svl, svl, 0); + TCGv_ptr za, zn, pn, pm; + + if (!sme_smza_enabled_check(s)) { + return true; + } + + /* Sum XZR+zad to find ZAd. */ + za = get_tile_rowcol(s, esz, 31, a->zad, false); + zn = vec_full_reg_ptr(s, a->zn); + pn = pred_full_reg_ptr(s, a->pn); + pm = pred_full_reg_ptr(s, a->pm); + + fn(za, zn, pn, pm, tcg_constant_i32(desc)); + + tcg_temp_free_ptr(za); + tcg_temp_free_ptr(zn); + tcg_temp_free_ptr(pn); + tcg_temp_free_ptr(pm); + return true; +} + +TRANS_FEAT(ADDHA_s, aa64_sme, do_adda, a, MO_32, gen_helper_sme_addha_s) +TRANS_FEAT(ADDVA_s, aa64_sme, do_adda, a, MO_32, gen_helper_sme_addva_s) +TRANS_FEAT(ADDHA_d, aa64_sme_i16i64, do_adda, a, MO_64, gen_helper_sme_addha_d) +TRANS_FEAT(ADDVA_d, aa64_sme_i16i64, do_adda, a, MO_64, gen_helper_sme_addva_d) -- 2.25.1 From: Richard Henderson <richard.henderson@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20220708151540.18136-25-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/helper-sme.h | 5 +++ target/arm/sme.decode | 9 +++++ target/arm/sme_helper.c | 69 ++++++++++++++++++++++++++++++++++++++ target/arm/translate-sme.c | 32 ++++++++++++++++++ 4 files changed, 115 insertions(+) diff --git a/target/arm/helper-sme.h b/target/arm/helper-sme.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper-sme.h +++ b/target/arm/helper-sme.h @@ -XXX,XX +XXX,XX @@ DEF_HELPER_FLAGS_5(sme_addha_s, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, ptr, i32) DEF_HELPER_FLAGS_5(sme_addva_s, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, ptr, i32) DEF_HELPER_FLAGS_5(sme_addha_d, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, ptr, i32) DEF_HELPER_FLAGS_5(sme_addva_d, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, ptr, i32) + +DEF_HELPER_FLAGS_7(sme_fmopa_s, TCG_CALL_NO_RWG, + void, ptr, ptr, ptr, ptr, ptr, ptr, i32) +DEF_HELPER_FLAGS_7(sme_fmopa_d, TCG_CALL_NO_RWG, + void, ptr, ptr, ptr, ptr, ptr, ptr, i32) diff --git a/target/arm/sme.decode b/target/arm/sme.decode index XXXXXXX..XXXXXXX 100644 --- a/target/arm/sme.decode +++ b/target/arm/sme.decode @@ -XXX,XX +XXX,XX @@ ADDHA_s 11000000 10 01000 0 ... ... ..... 000 .. @adda_32 ADDVA_s 11000000 10 01000 1 ... ... ..... 000 .. @adda_32 ADDHA_d 11000000 11 01000 0 ... ... ..... 00 ... @adda_64 ADDVA_d 11000000 11 01000 1 ... ... ..... 00 ... @adda_64 + +### SME Outer Product + +&op zad zn zm pm pn sub:bool +@op_32 ........ ... zm:5 pm:3 pn:3 zn:5 sub:1 .. zad:2 &op +@op_64 ........ ... zm:5 pm:3 pn:3 zn:5 sub:1 . zad:3 &op + +FMOPA_s 10000000 100 ..... ... ... ..... . 00 .. @op_32 +FMOPA_d 10000000 110 ..... ... ... ..... . 0 ... @op_64 diff --git a/target/arm/sme_helper.c b/target/arm/sme_helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/sme_helper.c +++ b/target/arm/sme_helper.c @@ -XXX,XX +XXX,XX @@ #include "exec/cpu_ldst.h" #include "exec/exec-all.h" #include "qemu/int128.h" +#include "fpu/softfloat.h" #include "vec_internal.h" #include "sve_ldst_internal.h" @@ -XXX,XX +XXX,XX @@ void HELPER(sme_addva_d)(void *vzda, void *vzn, void *vpn, } } } + +void HELPER(sme_fmopa_s)(void *vza, void *vzn, void *vzm, void *vpn, + void *vpm, void *vst, uint32_t desc) +{ + intptr_t row, col, oprsz = simd_maxsz(desc); + uint32_t neg = simd_data(desc) << 31; + uint16_t *pn = vpn, *pm = vpm; + float_status fpst; + + /* + * Make a copy of float_status because this operation does not + * update the cumulative fp exception status. It also produces + * default nans. + */ + fpst = *(float_status *)vst; + set_default_nan_mode(true, &fpst); + + for (row = 0; row < oprsz; ) { + uint16_t pa = pn[H2(row >> 4)]; + do { + if (pa & 1) { + void *vza_row = vza + tile_vslice_offset(row); + uint32_t n = *(uint32_t *)(vzn + H1_4(row)) ^ neg; + + for (col = 0; col < oprsz; ) { + uint16_t pb = pm[H2(col >> 4)]; + do { + if (pb & 1) { + uint32_t *a = vza_row + H1_4(col); + uint32_t *m = vzm + H1_4(col); + *a = float32_muladd(n, *m, *a, 0, vst); + } + col += 4; + pb >>= 4; + } while (col & 15); + } + } + row += 4; + pa >>= 4; + } while (row & 15); + } +} + +void HELPER(sme_fmopa_d)(void *vza, void *vzn, void *vzm, void *vpn, + void *vpm, void *vst, uint32_t desc) +{ + intptr_t row, col, oprsz = simd_oprsz(desc) / 8; + uint64_t neg = (uint64_t)simd_data(desc) << 63; + uint64_t *za = vza, *zn = vzn, *zm = vzm; + uint8_t *pn = vpn, *pm = vpm; + float_status fpst = *(float_status *)vst; + + set_default_nan_mode(true, &fpst); + + for (row = 0; row < oprsz; ++row) { + if (pn[H1(row)] & 1) { + uint64_t *za_row = &za[tile_vslice_index(row)]; + uint64_t n = zn[row] ^ neg; + + for (col = 0; col < oprsz; ++col) { + if (pm[H1(col)] & 1) { + uint64_t *a = &za_row[col]; + *a = float64_muladd(n, zm[col], *a, 0, &fpst); + } + } + } + } +} diff --git a/target/arm/translate-sme.c b/target/arm/translate-sme.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-sme.c +++ b/target/arm/translate-sme.c @@ -XXX,XX +XXX,XX @@ TRANS_FEAT(ADDHA_s, aa64_sme, do_adda, a, MO_32, gen_helper_sme_addha_s) TRANS_FEAT(ADDVA_s, aa64_sme, do_adda, a, MO_32, gen_helper_sme_addva_s) TRANS_FEAT(ADDHA_d, aa64_sme_i16i64, do_adda, a, MO_64, gen_helper_sme_addha_d) TRANS_FEAT(ADDVA_d, aa64_sme_i16i64, do_adda, a, MO_64, gen_helper_sme_addva_d) + +static bool do_outprod_fpst(DisasContext *s, arg_op *a, MemOp esz, + gen_helper_gvec_5_ptr *fn) +{ + int svl = streaming_vec_reg_size(s); + uint32_t desc = simd_desc(svl, svl, a->sub); + TCGv_ptr za, zn, zm, pn, pm, fpst; + + if (!sme_smza_enabled_check(s)) { + return true; + } + + /* Sum XZR+zad to find ZAd. */ + za = get_tile_rowcol(s, esz, 31, a->zad, false); + zn = vec_full_reg_ptr(s, a->zn); + zm = vec_full_reg_ptr(s, a->zm); + pn = pred_full_reg_ptr(s, a->pn); + pm = pred_full_reg_ptr(s, a->pm); + fpst = fpstatus_ptr(FPST_FPCR); + + fn(za, zn, zm, pn, pm, fpst, tcg_constant_i32(desc)); + + tcg_temp_free_ptr(za); + tcg_temp_free_ptr(zn); + tcg_temp_free_ptr(pn); + tcg_temp_free_ptr(pm); + tcg_temp_free_ptr(fpst); + return true; +} + +TRANS_FEAT(FMOPA_s, aa64_sme, do_outprod_fpst, a, MO_32, gen_helper_sme_fmopa_s) +TRANS_FEAT(FMOPA_d, aa64_sme_f64f64, do_outprod_fpst, a, MO_64, gen_helper_sme_fmopa_d) -- 2.25.1 From: Richard Henderson <richard.henderson@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20220708151540.18136-26-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/helper-sme.h | 2 ++ target/arm/sme.decode | 2 ++ target/arm/sme_helper.c | 56 ++++++++++++++++++++++++++++++++++++++ target/arm/translate-sme.c | 30 ++++++++++++++++++++ 4 files changed, 90 insertions(+) diff --git a/target/arm/helper-sme.h b/target/arm/helper-sme.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper-sme.h +++ b/target/arm/helper-sme.h @@ -XXX,XX +XXX,XX @@ DEF_HELPER_FLAGS_7(sme_fmopa_s, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, ptr, ptr, ptr, i32) DEF_HELPER_FLAGS_7(sme_fmopa_d, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, ptr, ptr, ptr, i32) +DEF_HELPER_FLAGS_6(sme_bfmopa, TCG_CALL_NO_RWG, + void, ptr, ptr, ptr, ptr, ptr, i32) diff --git a/target/arm/sme.decode b/target/arm/sme.decode index XXXXXXX..XXXXXXX 100644 --- a/target/arm/sme.decode +++ b/target/arm/sme.decode @@ -XXX,XX +XXX,XX @@ ADDVA_d 11000000 11 01000 1 ... ... ..... 00 ... @adda_64 FMOPA_s 10000000 100 ..... ... ... ..... . 00 .. @op_32 FMOPA_d 10000000 110 ..... ... ... ..... . 0 ... @op_64 + +BFMOPA 10000001 100 ..... ... ... ..... . 00 .. @op_32 diff --git a/target/arm/sme_helper.c b/target/arm/sme_helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/sme_helper.c +++ b/target/arm/sme_helper.c @@ -XXX,XX +XXX,XX @@ void HELPER(sme_fmopa_d)(void *vza, void *vzn, void *vzm, void *vpn, } } } + +/* + * Alter PAIR as needed for controlling predicates being false, + * and for NEG on an enabled row element. + */ +static inline uint32_t f16mop_adj_pair(uint32_t pair, uint32_t pg, uint32_t neg) +{ + /* + * The pseudocode uses a conditional negate after the conditional zero. + * It is simpler here to unconditionally negate before conditional zero. + */ + pair ^= neg; + if (!(pg & 1)) { + pair &= 0xffff0000u; + } + if (!(pg & 4)) { + pair &= 0x0000ffffu; + } + return pair; +} + +void HELPER(sme_bfmopa)(void *vza, void *vzn, void *vzm, void *vpn, + void *vpm, uint32_t desc) +{ + intptr_t row, col, oprsz = simd_maxsz(desc); + uint32_t neg = simd_data(desc) * 0x80008000u; + uint16_t *pn = vpn, *pm = vpm; + + for (row = 0; row < oprsz; ) { + uint16_t prow = pn[H2(row >> 4)]; + do { + void *vza_row = vza + tile_vslice_offset(row); + uint32_t n = *(uint32_t *)(vzn + H1_4(row)); + + n = f16mop_adj_pair(n, prow, neg); + + for (col = 0; col < oprsz; ) { + uint16_t pcol = pm[H2(col >> 4)]; + do { + if (prow & pcol & 0b0101) { + uint32_t *a = vza_row + H1_4(col); + uint32_t m = *(uint32_t *)(vzm + H1_4(col)); + + m = f16mop_adj_pair(m, pcol, 0); + *a = bfdotadd(*a, n, m); + + col += 4; + pcol >>= 4; + } + } while (col & 15); + } + row += 4; + prow >>= 4; + } while (row & 15); + } +} diff --git a/target/arm/translate-sme.c b/target/arm/translate-sme.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-sme.c +++ b/target/arm/translate-sme.c @@ -XXX,XX +XXX,XX @@ TRANS_FEAT(ADDVA_s, aa64_sme, do_adda, a, MO_32, gen_helper_sme_addva_s) TRANS_FEAT(ADDHA_d, aa64_sme_i16i64, do_adda, a, MO_64, gen_helper_sme_addha_d) TRANS_FEAT(ADDVA_d, aa64_sme_i16i64, do_adda, a, MO_64, gen_helper_sme_addva_d) +static bool do_outprod(DisasContext *s, arg_op *a, MemOp esz, + gen_helper_gvec_5 *fn) +{ + int svl = streaming_vec_reg_size(s); + uint32_t desc = simd_desc(svl, svl, a->sub); + TCGv_ptr za, zn, zm, pn, pm; + + if (!sme_smza_enabled_check(s)) { + return true; + } + + /* Sum XZR+zad to find ZAd. */ + za = get_tile_rowcol(s, esz, 31, a->zad, false); + zn = vec_full_reg_ptr(s, a->zn); + zm = vec_full_reg_ptr(s, a->zm); + pn = pred_full_reg_ptr(s, a->pn); + pm = pred_full_reg_ptr(s, a->pm); + + fn(za, zn, zm, pn, pm, tcg_constant_i32(desc)); + + tcg_temp_free_ptr(za); + tcg_temp_free_ptr(zn); + tcg_temp_free_ptr(pn); + tcg_temp_free_ptr(pm); + return true; +} + static bool do_outprod_fpst(DisasContext *s, arg_op *a, MemOp esz, gen_helper_gvec_5_ptr *fn) { @@ -XXX,XX +XXX,XX @@ static bool do_outprod_fpst(DisasContext *s, arg_op *a, MemOp esz, TRANS_FEAT(FMOPA_s, aa64_sme, do_outprod_fpst, a, MO_32, gen_helper_sme_fmopa_s) TRANS_FEAT(FMOPA_d, aa64_sme_f64f64, do_outprod_fpst, a, MO_64, gen_helper_sme_fmopa_d) + +/* TODO: FEAT_EBF16 */ +TRANS_FEAT(BFMOPA, aa64_sme, do_outprod, a, MO_32, gen_helper_sme_bfmopa) -- 2.25.1 From: Richard Henderson <richard.henderson@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20220708151540.18136-27-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/helper-sme.h | 2 ++ target/arm/sme.decode | 1 + target/arm/sme_helper.c | 74 ++++++++++++++++++++++++++++++++++++++ target/arm/translate-sme.c | 1 + 4 files changed, 78 insertions(+) diff --git a/target/arm/helper-sme.h b/target/arm/helper-sme.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper-sme.h +++ b/target/arm/helper-sme.h @@ -XXX,XX +XXX,XX @@ DEF_HELPER_FLAGS_5(sme_addva_s, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, ptr, i32) DEF_HELPER_FLAGS_5(sme_addha_d, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, ptr, i32) DEF_HELPER_FLAGS_5(sme_addva_d, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, ptr, i32) +DEF_HELPER_FLAGS_7(sme_fmopa_h, TCG_CALL_NO_RWG, + void, ptr, ptr, ptr, ptr, ptr, ptr, i32) DEF_HELPER_FLAGS_7(sme_fmopa_s, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, ptr, ptr, ptr, i32) DEF_HELPER_FLAGS_7(sme_fmopa_d, TCG_CALL_NO_RWG, diff --git a/target/arm/sme.decode b/target/arm/sme.decode index XXXXXXX..XXXXXXX 100644 --- a/target/arm/sme.decode +++ b/target/arm/sme.decode @@ -XXX,XX +XXX,XX @@ FMOPA_s 10000000 100 ..... ... ... ..... . 00 .. @op_32 FMOPA_d 10000000 110 ..... ... ... ..... . 0 ... @op_64 BFMOPA 10000001 100 ..... ... ... ..... . 00 .. @op_32 +FMOPA_h 10000001 101 ..... ... ... ..... . 00 .. @op_32 diff --git a/target/arm/sme_helper.c b/target/arm/sme_helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/sme_helper.c +++ b/target/arm/sme_helper.c @@ -XXX,XX +XXX,XX @@ static inline uint32_t f16mop_adj_pair(uint32_t pair, uint32_t pg, uint32_t neg) return pair; } +static float32 f16_dotadd(float32 sum, uint32_t e1, uint32_t e2, + float_status *s_std, float_status *s_odd) +{ + float64 e1r = float16_to_float64(e1 & 0xffff, true, s_std); + float64 e1c = float16_to_float64(e1 >> 16, true, s_std); + float64 e2r = float16_to_float64(e2 & 0xffff, true, s_std); + float64 e2c = float16_to_float64(e2 >> 16, true, s_std); + float64 t64; + float32 t32; + + /* + * The ARM pseudocode function FPDot performs both multiplies + * and the add with a single rounding operation. Emulate this + * by performing the first multiply in round-to-odd, then doing + * the second multiply as fused multiply-add, and rounding to + * float32 all in one step. + */ + t64 = float64_mul(e1r, e2r, s_odd); + t64 = float64r32_muladd(e1c, e2c, t64, 0, s_std); + + /* This conversion is exact, because we've already rounded. */ + t32 = float64_to_float32(t64, s_std); + + /* The final accumulation step is not fused. */ + return float32_add(sum, t32, s_std); +} + +void HELPER(sme_fmopa_h)(void *vza, void *vzn, void *vzm, void *vpn, + void *vpm, void *vst, uint32_t desc) +{ + intptr_t row, col, oprsz = simd_maxsz(desc); + uint32_t neg = simd_data(desc) * 0x80008000u; + uint16_t *pn = vpn, *pm = vpm; + float_status fpst_odd, fpst_std; + + /* + * Make a copy of float_status because this operation does not + * update the cumulative fp exception status. It also produces + * default nans. Make a second copy with round-to-odd -- see above. + */ + fpst_std = *(float_status *)vst; + set_default_nan_mode(true, &fpst_std); + fpst_odd = fpst_std; + set_float_rounding_mode(float_round_to_odd, &fpst_odd); + + for (row = 0; row < oprsz; ) { + uint16_t prow = pn[H2(row >> 4)]; + do { + void *vza_row = vza + tile_vslice_offset(row); + uint32_t n = *(uint32_t *)(vzn + H1_4(row)); + + n = f16mop_adj_pair(n, prow, neg); + + for (col = 0; col < oprsz; ) { + uint16_t pcol = pm[H2(col >> 4)]; + do { + if (prow & pcol & 0b0101) { + uint32_t *a = vza_row + H1_4(col); + uint32_t m = *(uint32_t *)(vzm + H1_4(col)); + + m = f16mop_adj_pair(m, pcol, 0); + *a = f16_dotadd(*a, n, m, &fpst_std, &fpst_odd); + + col += 4; + pcol >>= 4; + } + } while (col & 15); + } + row += 4; + prow >>= 4; + } while (row & 15); + } +} + void HELPER(sme_bfmopa)(void *vza, void *vzn, void *vzm, void *vpn, void *vpm, uint32_t desc) { diff --git a/target/arm/translate-sme.c b/target/arm/translate-sme.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-sme.c +++ b/target/arm/translate-sme.c @@ -XXX,XX +XXX,XX @@ static bool do_outprod_fpst(DisasContext *s, arg_op *a, MemOp esz, return true; } +TRANS_FEAT(FMOPA_h, aa64_sme, do_outprod_fpst, a, MO_32, gen_helper_sme_fmopa_h) TRANS_FEAT(FMOPA_s, aa64_sme, do_outprod_fpst, a, MO_32, gen_helper_sme_fmopa_s) TRANS_FEAT(FMOPA_d, aa64_sme_f64f64, do_outprod_fpst, a, MO_64, gen_helper_sme_fmopa_d) -- 2.25.1 From: Richard Henderson <richard.henderson@linaro.org> This is SMOPA, SUMOPA, USMOPA_s, UMOPA, for both Int8 and Int16. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20220708151540.18136-28-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/helper-sme.h | 16 ++++++++ target/arm/sme.decode | 10 +++++ target/arm/sme_helper.c | 82 ++++++++++++++++++++++++++++++++++++++ target/arm/translate-sme.c | 10 +++++ 4 files changed, 118 insertions(+) diff --git a/target/arm/helper-sme.h b/target/arm/helper-sme.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper-sme.h +++ b/target/arm/helper-sme.h @@ -XXX,XX +XXX,XX @@ DEF_HELPER_FLAGS_7(sme_fmopa_d, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, ptr, ptr, ptr, i32) DEF_HELPER_FLAGS_6(sme_bfmopa, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, ptr, ptr, i32) +DEF_HELPER_FLAGS_6(sme_smopa_s, TCG_CALL_NO_RWG, + void, ptr, ptr, ptr, ptr, ptr, i32) +DEF_HELPER_FLAGS_6(sme_umopa_s, TCG_CALL_NO_RWG, + void, ptr, ptr, ptr, ptr, ptr, i32) +DEF_HELPER_FLAGS_6(sme_sumopa_s, TCG_CALL_NO_RWG, + void, ptr, ptr, ptr, ptr, ptr, i32) +DEF_HELPER_FLAGS_6(sme_usmopa_s, TCG_CALL_NO_RWG, + void, ptr, ptr, ptr, ptr, ptr, i32) +DEF_HELPER_FLAGS_6(sme_smopa_d, TCG_CALL_NO_RWG, + void, ptr, ptr, ptr, ptr, ptr, i32) +DEF_HELPER_FLAGS_6(sme_umopa_d, TCG_CALL_NO_RWG, + void, ptr, ptr, ptr, ptr, ptr, i32) +DEF_HELPER_FLAGS_6(sme_sumopa_d, TCG_CALL_NO_RWG, + void, ptr, ptr, ptr, ptr, ptr, i32) +DEF_HELPER_FLAGS_6(sme_usmopa_d, TCG_CALL_NO_RWG, + void, ptr, ptr, ptr, ptr, ptr, i32) diff --git a/target/arm/sme.decode b/target/arm/sme.decode index XXXXXXX..XXXXXXX 100644 --- a/target/arm/sme.decode +++ b/target/arm/sme.decode @@ -XXX,XX +XXX,XX @@ FMOPA_d 10000000 110 ..... ... ... ..... . 0 ... @op_64 BFMOPA 10000001 100 ..... ... ... ..... . 00 .. @op_32 FMOPA_h 10000001 101 ..... ... ... ..... . 00 .. @op_32 + +SMOPA_s 1010000 0 10 0 ..... ... ... ..... . 00 .. @op_32 +SUMOPA_s 1010000 0 10 1 ..... ... ... ..... . 00 .. @op_32 +USMOPA_s 1010000 1 10 0 ..... ... ... ..... . 00 .. @op_32 +UMOPA_s 1010000 1 10 1 ..... ... ... ..... . 00 .. @op_32 + +SMOPA_d 1010000 0 11 0 ..... ... ... ..... . 0 ... @op_64 +SUMOPA_d 1010000 0 11 1 ..... ... ... ..... . 0 ... @op_64 +USMOPA_d 1010000 1 11 0 ..... ... ... ..... . 0 ... @op_64 +UMOPA_d 1010000 1 11 1 ..... ... ... ..... . 0 ... @op_64 diff --git a/target/arm/sme_helper.c b/target/arm/sme_helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/sme_helper.c +++ b/target/arm/sme_helper.c @@ -XXX,XX +XXX,XX @@ void HELPER(sme_bfmopa)(void *vza, void *vzn, void *vzm, void *vpn, } while (row & 15); } } + +typedef uint64_t IMOPFn(uint64_t, uint64_t, uint64_t, uint8_t, bool); + +static inline void do_imopa(uint64_t *za, uint64_t *zn, uint64_t *zm, + uint8_t *pn, uint8_t *pm, + uint32_t desc, IMOPFn *fn) +{ + intptr_t row, col, oprsz = simd_oprsz(desc) / 8; + bool neg = simd_data(desc); + + for (row = 0; row < oprsz; ++row) { + uint8_t pa = pn[H1(row)]; + uint64_t *za_row = &za[tile_vslice_index(row)]; + uint64_t n = zn[row]; + + for (col = 0; col < oprsz; ++col) { + uint8_t pb = pm[H1(col)]; + uint64_t *a = &za_row[col]; + + *a = fn(n, zm[col], *a, pa & pb, neg); + } + } +} + +#define DEF_IMOP_32(NAME, NTYPE, MTYPE) \ +static uint64_t NAME(uint64_t n, uint64_t m, uint64_t a, uint8_t p, bool neg) \ +{ \ + uint32_t sum0 = 0, sum1 = 0; \ + /* Apply P to N as a mask, making the inactive elements 0. */ \ + n &= expand_pred_b(p); \ + sum0 += (NTYPE)(n >> 0) * (MTYPE)(m >> 0); \ + sum0 += (NTYPE)(n >> 8) * (MTYPE)(m >> 8); \ + sum0 += (NTYPE)(n >> 16) * (MTYPE)(m >> 16); \ + sum0 += (NTYPE)(n >> 24) * (MTYPE)(m >> 24); \ + sum1 += (NTYPE)(n >> 32) * (MTYPE)(m >> 32); \ + sum1 += (NTYPE)(n >> 40) * (MTYPE)(m >> 40); \ + sum1 += (NTYPE)(n >> 48) * (MTYPE)(m >> 48); \ + sum1 += (NTYPE)(n >> 56) * (MTYPE)(m >> 56); \ + if (neg) { \ + sum0 = (uint32_t)a - sum0, sum1 = (uint32_t)(a >> 32) - sum1; \ + } else { \ + sum0 = (uint32_t)a + sum0, sum1 = (uint32_t)(a >> 32) + sum1; \ + } \ + return ((uint64_t)sum1 << 32) | sum0; \ +} + +#define DEF_IMOP_64(NAME, NTYPE, MTYPE) \ +static uint64_t NAME(uint64_t n, uint64_t m, uint64_t a, uint8_t p, bool neg) \ +{ \ + uint64_t sum = 0; \ + /* Apply P to N as a mask, making the inactive elements 0. */ \ + n &= expand_pred_h(p); \ + sum += (NTYPE)(n >> 0) * (MTYPE)(m >> 0); \ + sum += (NTYPE)(n >> 16) * (MTYPE)(m >> 16); \ + sum += (NTYPE)(n >> 32) * (MTYPE)(m >> 32); \ + sum += (NTYPE)(n >> 48) * (MTYPE)(m >> 48); \ + return neg ? a - sum : a + sum; \ +} + +DEF_IMOP_32(smopa_s, int8_t, int8_t) +DEF_IMOP_32(umopa_s, uint8_t, uint8_t) +DEF_IMOP_32(sumopa_s, int8_t, uint8_t) +DEF_IMOP_32(usmopa_s, uint8_t, int8_t) + +DEF_IMOP_64(smopa_d, int16_t, int16_t) +DEF_IMOP_64(umopa_d, uint16_t, uint16_t) +DEF_IMOP_64(sumopa_d, int16_t, uint16_t) +DEF_IMOP_64(usmopa_d, uint16_t, int16_t) + +#define DEF_IMOPH(NAME) \ + void HELPER(sme_##NAME)(void *vza, void *vzn, void *vzm, void *vpn, \ + void *vpm, uint32_t desc) \ + { do_imopa(vza, vzn, vzm, vpn, vpm, desc, NAME); } + +DEF_IMOPH(smopa_s) +DEF_IMOPH(umopa_s) +DEF_IMOPH(sumopa_s) +DEF_IMOPH(usmopa_s) +DEF_IMOPH(smopa_d) +DEF_IMOPH(umopa_d) +DEF_IMOPH(sumopa_d) +DEF_IMOPH(usmopa_d) diff --git a/target/arm/translate-sme.c b/target/arm/translate-sme.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-sme.c +++ b/target/arm/translate-sme.c @@ -XXX,XX +XXX,XX @@ TRANS_FEAT(FMOPA_d, aa64_sme_f64f64, do_outprod_fpst, a, MO_64, gen_helper_sme_f /* TODO: FEAT_EBF16 */ TRANS_FEAT(BFMOPA, aa64_sme, do_outprod, a, MO_32, gen_helper_sme_bfmopa) + +TRANS_FEAT(SMOPA_s, aa64_sme, do_outprod, a, MO_32, gen_helper_sme_smopa_s) +TRANS_FEAT(UMOPA_s, aa64_sme, do_outprod, a, MO_32, gen_helper_sme_umopa_s) +TRANS_FEAT(SUMOPA_s, aa64_sme, do_outprod, a, MO_32, gen_helper_sme_sumopa_s) +TRANS_FEAT(USMOPA_s, aa64_sme, do_outprod, a, MO_32, gen_helper_sme_usmopa_s) + +TRANS_FEAT(SMOPA_d, aa64_sme_i16i64, do_outprod, a, MO_64, gen_helper_sme_smopa_d) +TRANS_FEAT(UMOPA_d, aa64_sme_i16i64, do_outprod, a, MO_64, gen_helper_sme_umopa_d) +TRANS_FEAT(SUMOPA_d, aa64_sme_i16i64, do_outprod, a, MO_64, gen_helper_sme_sumopa_d) +TRANS_FEAT(USMOPA_d, aa64_sme_i16i64, do_outprod, a, MO_64, gen_helper_sme_usmopa_d) -- 2.25.1 From: Richard Henderson <richard.henderson@linaro.org> This is an SVE instruction that operates using the SVE vector length but that it is present only if SME is implemented. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20220708151540.18136-29-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/sve.decode | 20 +++++++++++++ target/arm/translate-sve.c | 57 ++++++++++++++++++++++++++++++++++++++ 2 files changed, 77 insertions(+) diff --git a/target/arm/sve.decode b/target/arm/sve.decode index XXXXXXX..XXXXXXX 100644 --- a/target/arm/sve.decode +++ b/target/arm/sve.decode @@ -XXX,XX +XXX,XX @@ BFMLALT_zzxw 01100100 11 1 ..... 0100.1 ..... ..... @rrxr_3a esz=2 ### SVE2 floating-point bfloat16 dot-product (indexed) BFDOT_zzxz 01100100 01 1 ..... 010000 ..... ..... @rrxr_2 esz=2 + +### SVE broadcast predicate element + +&psel esz pd pn pm rv imm +%psel_rv 16:2 !function=plus_12 +%psel_imm_b 22:2 19:2 +%psel_imm_h 22:2 20:1 +%psel_imm_s 22:2 +%psel_imm_d 23:1 +@psel ........ .. . ... .. .. pn:4 . pm:4 . pd:4 \ + &psel rv=%psel_rv + +PSEL 00100101 .. 1 ..1 .. 01 .... 0 .... 0 .... \ + @psel esz=0 imm=%psel_imm_b +PSEL 00100101 .. 1 .10 .. 01 .... 0 .... 0 .... \ + @psel esz=1 imm=%psel_imm_h +PSEL 00100101 .. 1 100 .. 01 .... 0 .... 0 .... \ + @psel esz=2 imm=%psel_imm_s +PSEL 00100101 .1 1 000 .. 01 .... 0 .... 0 .... \ + @psel esz=3 imm=%psel_imm_d diff --git a/target/arm/translate-sve.c b/target/arm/translate-sve.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-sve.c +++ b/target/arm/translate-sve.c @@ -XXX,XX +XXX,XX @@ static bool do_BFMLAL_zzxw(DisasContext *s, arg_rrxr_esz *a, bool sel) TRANS_FEAT(BFMLALB_zzxw, aa64_sve_bf16, do_BFMLAL_zzxw, a, false) TRANS_FEAT(BFMLALT_zzxw, aa64_sve_bf16, do_BFMLAL_zzxw, a, true) + +static bool trans_PSEL(DisasContext *s, arg_psel *a) +{ + int vl = vec_full_reg_size(s); + int pl = pred_gvec_reg_size(s); + int elements = vl >> a->esz; + TCGv_i64 tmp, didx, dbit; + TCGv_ptr ptr; + + if (!dc_isar_feature(aa64_sme, s)) { + return false; + } + if (!sve_access_check(s)) { + return true; + } + + tmp = tcg_temp_new_i64(); + dbit = tcg_temp_new_i64(); + didx = tcg_temp_new_i64(); + ptr = tcg_temp_new_ptr(); + + /* Compute the predicate element. */ + tcg_gen_addi_i64(tmp, cpu_reg(s, a->rv), a->imm); + if (is_power_of_2(elements)) { + tcg_gen_andi_i64(tmp, tmp, elements - 1); + } else { + tcg_gen_remu_i64(tmp, tmp, tcg_constant_i64(elements)); + } + + /* Extract the predicate byte and bit indices. */ + tcg_gen_shli_i64(tmp, tmp, a->esz); + tcg_gen_andi_i64(dbit, tmp, 7); + tcg_gen_shri_i64(didx, tmp, 3); + if (HOST_BIG_ENDIAN) { + tcg_gen_xori_i64(didx, didx, 7); + } + + /* Load the predicate word. */ + tcg_gen_trunc_i64_ptr(ptr, didx); + tcg_gen_add_ptr(ptr, ptr, cpu_env); + tcg_gen_ld8u_i64(tmp, ptr, pred_full_reg_offset(s, a->pm)); + + /* Extract the predicate bit and replicate to MO_64. */ + tcg_gen_shr_i64(tmp, tmp, dbit); + tcg_gen_andi_i64(tmp, tmp, 1); + tcg_gen_neg_i64(tmp, tmp); + + /* Apply to either copy the source, or write zeros. */ + tcg_gen_gvec_ands(MO_64, pred_full_reg_offset(s, a->pd), + pred_full_reg_offset(s, a->pn), tmp, pl, pl); + + tcg_temp_free_i64(tmp); + tcg_temp_free_i64(dbit); + tcg_temp_free_i64(didx); + tcg_temp_free_ptr(ptr); + return true; +} -- 2.25.1 From: Richard Henderson <richard.henderson@linaro.org> This is an SVE instruction that operates using the SVE vector length but that it is present only if SME is implemented. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20220708151540.18136-30-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/helper-sve.h | 2 ++ target/arm/sve.decode | 1 + target/arm/sve_helper.c | 16 ++++++++++++++++ target/arm/translate-sve.c | 2 ++ 4 files changed, 21 insertions(+) diff --git a/target/arm/helper-sve.h b/target/arm/helper-sve.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper-sve.h +++ b/target/arm/helper-sve.h @@ -XXX,XX +XXX,XX @@ DEF_HELPER_FLAGS_4(sve_revh_d, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, i32) DEF_HELPER_FLAGS_4(sve_revw_d, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, i32) +DEF_HELPER_FLAGS_4(sme_revd_q, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, i32) + DEF_HELPER_FLAGS_4(sve_rbit_b, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, i32) DEF_HELPER_FLAGS_4(sve_rbit_h, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, i32) DEF_HELPER_FLAGS_4(sve_rbit_s, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, i32) diff --git a/target/arm/sve.decode b/target/arm/sve.decode index XXXXXXX..XXXXXXX 100644 --- a/target/arm/sve.decode +++ b/target/arm/sve.decode @@ -XXX,XX +XXX,XX @@ REVB 00000101 .. 1001 00 100 ... ..... ..... @rd_pg_rn REVH 00000101 .. 1001 01 100 ... ..... ..... @rd_pg_rn REVW 00000101 .. 1001 10 100 ... ..... ..... @rd_pg_rn RBIT 00000101 .. 1001 11 100 ... ..... ..... @rd_pg_rn +REVD 00000101 00 1011 10 100 ... ..... ..... @rd_pg_rn_e0 # SVE vector splice (predicated, destructive) SPLICE 00000101 .. 101 100 100 ... ..... ..... @rdn_pg_rm diff --git a/target/arm/sve_helper.c b/target/arm/sve_helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/sve_helper.c +++ b/target/arm/sve_helper.c @@ -XXX,XX +XXX,XX @@ DO_ZPZ_D(sve_revh_d, uint64_t, hswap64) DO_ZPZ_D(sve_revw_d, uint64_t, wswap64) +void HELPER(sme_revd_q)(void *vd, void *vn, void *vg, uint32_t desc) +{ + intptr_t i, opr_sz = simd_oprsz(desc) / 8; + uint64_t *d = vd, *n = vn; + uint8_t *pg = vg; + + for (i = 0; i < opr_sz; i += 2) { + if (pg[H1(i)] & 1) { + uint64_t n0 = n[i + 0]; + uint64_t n1 = n[i + 1]; + d[i + 0] = n1; + d[i + 1] = n0; + } + } +} + DO_ZPZ(sve_rbit_b, uint8_t, H1, revbit8) DO_ZPZ(sve_rbit_h, uint16_t, H1_2, revbit16) DO_ZPZ(sve_rbit_s, uint32_t, H1_4, revbit32) 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 @@ TRANS_FEAT(REVH, aa64_sve, gen_gvec_ool_arg_zpz, revh_fns[a->esz], a, 0) TRANS_FEAT(REVW, aa64_sve, gen_gvec_ool_arg_zpz, a->esz == 3 ? gen_helper_sve_revw_d : NULL, a, 0) +TRANS_FEAT(REVD, aa64_sme, gen_gvec_ool_arg_zpz, gen_helper_sme_revd_q, a, 0) + TRANS_FEAT(SPLICE, aa64_sve, gen_gvec_ool_arg_zpzz, gen_helper_sve_splice, a, a->esz) -- 2.25.1 From: Richard Henderson <richard.henderson@linaro.org> This is an SVE instruction that operates using the SVE vector length but that it is present only if SME is implemented. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20220708151540.18136-31-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/helper.h | 18 +++++++ target/arm/sve.decode | 5 ++ target/arm/translate-sve.c | 102 +++++++++++++++++++++++++++++++++++++ target/arm/vec_helper.c | 24 +++++++++ 4 files changed, 149 insertions(+) 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, TCG_CALL_NO_RWG, DEF_HELPER_FLAGS_6(gvec_bfmlal_idx, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, ptr, ptr, i32) +DEF_HELPER_FLAGS_5(gvec_sclamp_b, TCG_CALL_NO_RWG, + void, ptr, ptr, ptr, ptr, i32) +DEF_HELPER_FLAGS_5(gvec_sclamp_h, TCG_CALL_NO_RWG, + void, ptr, ptr, ptr, ptr, i32) +DEF_HELPER_FLAGS_5(gvec_sclamp_s, TCG_CALL_NO_RWG, + void, ptr, ptr, ptr, ptr, i32) +DEF_HELPER_FLAGS_5(gvec_sclamp_d, TCG_CALL_NO_RWG, + void, ptr, ptr, ptr, ptr, i32) + +DEF_HELPER_FLAGS_5(gvec_uclamp_b, TCG_CALL_NO_RWG, + void, ptr, ptr, ptr, ptr, i32) +DEF_HELPER_FLAGS_5(gvec_uclamp_h, TCG_CALL_NO_RWG, + void, ptr, ptr, ptr, ptr, i32) +DEF_HELPER_FLAGS_5(gvec_uclamp_s, TCG_CALL_NO_RWG, + void, ptr, ptr, ptr, ptr, i32) +DEF_HELPER_FLAGS_5(gvec_uclamp_d, TCG_CALL_NO_RWG, + void, ptr, ptr, ptr, ptr, i32) + #ifdef TARGET_AARCH64 #include "helper-a64.h" #include "helper-sve.h" diff --git a/target/arm/sve.decode b/target/arm/sve.decode index XXXXXXX..XXXXXXX 100644 --- a/target/arm/sve.decode +++ b/target/arm/sve.decode @@ -XXX,XX +XXX,XX @@ PSEL 00100101 .. 1 100 .. 01 .... 0 .... 0 .... \ @psel esz=2 imm=%psel_imm_s PSEL 00100101 .1 1 000 .. 01 .... 0 .... 0 .... \ @psel esz=3 imm=%psel_imm_d + +### SVE clamp + +SCLAMP 01000100 .. 0 ..... 110000 ..... ..... @rda_rn_rm +UCLAMP 01000100 .. 0 ..... 110001 ..... ..... @rda_rn_rm diff --git a/target/arm/translate-sve.c b/target/arm/translate-sve.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-sve.c +++ b/target/arm/translate-sve.c @@ -XXX,XX +XXX,XX @@ static bool trans_PSEL(DisasContext *s, arg_psel *a) tcg_temp_free_ptr(ptr); return true; } + +static void gen_sclamp_i32(TCGv_i32 d, TCGv_i32 n, TCGv_i32 m, TCGv_i32 a) +{ + tcg_gen_smax_i32(d, a, n); + tcg_gen_smin_i32(d, d, m); +} + +static void gen_sclamp_i64(TCGv_i64 d, TCGv_i64 n, TCGv_i64 m, TCGv_i64 a) +{ + tcg_gen_smax_i64(d, a, n); + tcg_gen_smin_i64(d, d, m); +} + +static void gen_sclamp_vec(unsigned vece, TCGv_vec d, TCGv_vec n, + TCGv_vec m, TCGv_vec a) +{ + tcg_gen_smax_vec(vece, d, a, n); + tcg_gen_smin_vec(vece, d, d, m); +} + +static void gen_sclamp(unsigned vece, uint32_t d, uint32_t n, uint32_t m, + uint32_t a, uint32_t oprsz, uint32_t maxsz) +{ + static const TCGOpcode vecop[] = { + INDEX_op_smin_vec, INDEX_op_smax_vec, 0 + }; + static const GVecGen4 ops[4] = { + { .fniv = gen_sclamp_vec, + .fno = gen_helper_gvec_sclamp_b, + .opt_opc = vecop, + .vece = MO_8 }, + { .fniv = gen_sclamp_vec, + .fno = gen_helper_gvec_sclamp_h, + .opt_opc = vecop, + .vece = MO_16 }, + { .fni4 = gen_sclamp_i32, + .fniv = gen_sclamp_vec, + .fno = gen_helper_gvec_sclamp_s, + .opt_opc = vecop, + .vece = MO_32 }, + { .fni8 = gen_sclamp_i64, + .fniv = gen_sclamp_vec, + .fno = gen_helper_gvec_sclamp_d, + .opt_opc = vecop, + .vece = MO_64, + .prefer_i64 = TCG_TARGET_REG_BITS == 64 } + }; + tcg_gen_gvec_4(d, n, m, a, oprsz, maxsz, &ops[vece]); +} + +TRANS_FEAT(SCLAMP, aa64_sme, gen_gvec_fn_arg_zzzz, gen_sclamp, a) + +static void gen_uclamp_i32(TCGv_i32 d, TCGv_i32 n, TCGv_i32 m, TCGv_i32 a) +{ + tcg_gen_umax_i32(d, a, n); + tcg_gen_umin_i32(d, d, m); +} + +static void gen_uclamp_i64(TCGv_i64 d, TCGv_i64 n, TCGv_i64 m, TCGv_i64 a) +{ + tcg_gen_umax_i64(d, a, n); + tcg_gen_umin_i64(d, d, m); +} + +static void gen_uclamp_vec(unsigned vece, TCGv_vec d, TCGv_vec n, + TCGv_vec m, TCGv_vec a) +{ + tcg_gen_umax_vec(vece, d, a, n); + tcg_gen_umin_vec(vece, d, d, m); +} + +static void gen_uclamp(unsigned vece, uint32_t d, uint32_t n, uint32_t m, + uint32_t a, uint32_t oprsz, uint32_t maxsz) +{ + static const TCGOpcode vecop[] = { + INDEX_op_umin_vec, INDEX_op_umax_vec, 0 + }; + static const GVecGen4 ops[4] = { + { .fniv = gen_uclamp_vec, + .fno = gen_helper_gvec_uclamp_b, + .opt_opc = vecop, + .vece = MO_8 }, + { .fniv = gen_uclamp_vec, + .fno = gen_helper_gvec_uclamp_h, + .opt_opc = vecop, + .vece = MO_16 }, + { .fni4 = gen_uclamp_i32, + .fniv = gen_uclamp_vec, + .fno = gen_helper_gvec_uclamp_s, + .opt_opc = vecop, + .vece = MO_32 }, + { .fni8 = gen_uclamp_i64, + .fniv = gen_uclamp_vec, + .fno = gen_helper_gvec_uclamp_d, + .opt_opc = vecop, + .vece = MO_64, + .prefer_i64 = TCG_TARGET_REG_BITS == 64 } + }; + tcg_gen_gvec_4(d, n, m, a, oprsz, maxsz, &ops[vece]); +} + +TRANS_FEAT(UCLAMP, aa64_sme, gen_gvec_fn_arg_zzzz, gen_uclamp, a) diff --git a/target/arm/vec_helper.c b/target/arm/vec_helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/vec_helper.c +++ b/target/arm/vec_helper.c @@ -XXX,XX +XXX,XX @@ void HELPER(gvec_bfmlal_idx)(void *vd, void *vn, void *vm, } clear_tail(d, opr_sz, simd_maxsz(desc)); } + +#define DO_CLAMP(NAME, TYPE) \ +void HELPER(NAME)(void *d, void *n, void *m, void *a, uint32_t desc) \ +{ \ + intptr_t i, opr_sz = simd_oprsz(desc); \ + for (i = 0; i < opr_sz; i += sizeof(TYPE)) { \ + TYPE aa = *(TYPE *)(a + i); \ + TYPE nn = *(TYPE *)(n + i); \ + TYPE mm = *(TYPE *)(m + i); \ + TYPE dd = MIN(MAX(aa, nn), mm); \ + *(TYPE *)(d + i) = dd; \ + } \ + clear_tail(d, opr_sz, simd_maxsz(desc)); \ +} + +DO_CLAMP(gvec_sclamp_b, int8_t) +DO_CLAMP(gvec_sclamp_h, int16_t) +DO_CLAMP(gvec_sclamp_s, int32_t) +DO_CLAMP(gvec_sclamp_d, int64_t) + +DO_CLAMP(gvec_uclamp_b, uint8_t) +DO_CLAMP(gvec_uclamp_h, uint16_t) +DO_CLAMP(gvec_uclamp_s, uint32_t) +DO_CLAMP(gvec_uclamp_d, uint64_t) -- 2.25.1 From: Richard Henderson <richard.henderson@linaro.org> We can handle both exception entry and exception return by hooking into aarch64_sve_change_el. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20220708151540.18136-32-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/helper.c | 15 +++++++++++++-- 1 file changed, 13 insertions(+), 2 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_sve_change_el(CPUARMState *env, int old_el, return; } + old_a64 = old_el ? arm_el_is_aa64(env, old_el) : el0_a64; + new_a64 = new_el ? arm_el_is_aa64(env, new_el) : el0_a64; + + /* + * Both AArch64.TakeException and AArch64.ExceptionReturn + * invoke ResetSVEState when taking an exception from, or + * returning to, AArch32 state when PSTATE.SM is enabled. + */ + if (old_a64 != new_a64 && FIELD_EX64(env->svcr, SVCR, SM)) { + arm_reset_sve_state(env); + return; + } + /* * DDI0584A.d sec 3.2: "If SVE instructions are disabled or trapped * at ELx, or not available because the EL is in AArch32 state, then @@ -XXX,XX +XXX,XX @@ void aarch64_sve_change_el(CPUARMState *env, int old_el, * we already have the correct register contents when encountering the * vq0->vq0 transition between EL0->EL1. */ - old_a64 = old_el ? arm_el_is_aa64(env, old_el) : el0_a64; old_len = (old_a64 && !sve_exception_el(env, old_el) ? sve_vqm1_for_el(env, old_el) : 0); - new_a64 = new_el ? arm_el_is_aa64(env, new_el) : el0_a64; new_len = (new_a64 && !sve_exception_el(env, new_el) ? sve_vqm1_for_el(env, new_el) : 0); -- 2.25.1 From: Richard Henderson <richard.henderson@linaro.org> Note that SME remains effectively disabled for user-only, because we do not yet set CPACR_EL1.SMEN. This needs to wait until the kernel ABI is implemented. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20220708151540.18136-33-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- docs/system/arm/emulation.rst | 4 ++++ target/arm/cpu64.c | 11 +++++++++++ 2 files changed, 15 insertions(+) diff --git a/docs/system/arm/emulation.rst b/docs/system/arm/emulation.rst index XXXXXXX..XXXXXXX 100644 --- a/docs/system/arm/emulation.rst +++ b/docs/system/arm/emulation.rst @@ -XXX,XX +XXX,XX @@ the following architecture extensions: - FEAT_SHA512 (Advanced SIMD SHA512 instructions) - FEAT_SM3 (Advanced SIMD SM3 instructions) - FEAT_SM4 (Advanced SIMD SM4 instructions) +- FEAT_SME (Scalable Matrix Extension) +- FEAT_SME_FA64 (Full A64 instruction set in Streaming SVE mode) +- FEAT_SME_F64F64 (Double-precision floating-point outer product instructions) +- FEAT_SME_I16I64 (16-bit to 64-bit integer widening outer product instructions) - FEAT_SPECRES (Speculation restriction instructions) - FEAT_SSBS (Speculative Store Bypass Safe) - FEAT_TLBIOS (TLB invalidate instructions in Outer Shareable domain) diff --git a/target/arm/cpu64.c b/target/arm/cpu64.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/cpu64.c +++ b/target/arm/cpu64.c @@ -XXX,XX +XXX,XX @@ static void aarch64_max_initfn(Object *obj) */ t = FIELD_DP64(t, ID_AA64PFR1, MTE, 3); /* FEAT_MTE3 */ t = FIELD_DP64(t, ID_AA64PFR1, RAS_FRAC, 0); /* FEAT_RASv1p1 + FEAT_DoubleFault */ + t = FIELD_DP64(t, ID_AA64PFR1, SME, 1); /* FEAT_SME */ t = FIELD_DP64(t, ID_AA64PFR1, CSV2_FRAC, 0); /* FEAT_CSV2_2 */ cpu->isar.id_aa64pfr1 = t; @@ -XXX,XX +XXX,XX @@ static void aarch64_max_initfn(Object *obj) t = FIELD_DP64(t, ID_AA64DFR0, PMUVER, 5); /* FEAT_PMUv3p4 */ cpu->isar.id_aa64dfr0 = t; + t = cpu->isar.id_aa64smfr0; + t = FIELD_DP64(t, ID_AA64SMFR0, F32F32, 1); /* FEAT_SME */ + t = FIELD_DP64(t, ID_AA64SMFR0, B16F32, 1); /* FEAT_SME */ + t = FIELD_DP64(t, ID_AA64SMFR0, F16F32, 1); /* FEAT_SME */ + t = FIELD_DP64(t, ID_AA64SMFR0, I8I32, 0xf); /* FEAT_SME */ + t = FIELD_DP64(t, ID_AA64SMFR0, F64F64, 1); /* FEAT_SME_F64F64 */ + t = FIELD_DP64(t, ID_AA64SMFR0, I16I64, 0xf); /* FEAT_SME_I16I64 */ + t = FIELD_DP64(t, ID_AA64SMFR0, FA64, 1); /* FEAT_SME_FA64 */ + cpu->isar.id_aa64smfr0 = t; + /* Replicate the same data to the 32-bit id registers. */ aa32_max_features(cpu); -- 2.25.1 From: Richard Henderson <richard.henderson@linaro.org> Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20220708151540.18136-34-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- linux-user/aarch64/target_cpu.h | 5 ++++- 1 file changed, 4 insertions(+), 1 deletion(-) diff --git a/linux-user/aarch64/target_cpu.h b/linux-user/aarch64/target_cpu.h index XXXXXXX..XXXXXXX 100644 --- a/linux-user/aarch64/target_cpu.h +++ b/linux-user/aarch64/target_cpu.h @@ -XXX,XX +XXX,XX @@ static inline void cpu_clone_regs_parent(CPUARMState *env, unsigned flags) static inline void cpu_set_tls(CPUARMState *env, target_ulong newtls) { - /* Note that AArch64 Linux keeps the TLS pointer in TPIDR; this is + /* + * Note that AArch64 Linux keeps the TLS pointer in TPIDR; this is * different from AArch32 Linux, which uses TPIDRRO. */ env->cp15.tpidr_el[0] = newtls; + /* TPIDR2_EL0 is cleared with CLONE_SETTLS. */ + env->cp15.tpidr2_el0 = 0; } static inline abi_ulong get_sp_from_cpustate(CPUARMState *state) -- 2.25.1 From: Richard Henderson <richard.henderson@linaro.org> Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20220708151540.18136-35-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- linux-user/aarch64/cpu_loop.c | 9 +++++++++ 1 file changed, 9 insertions(+) diff --git a/linux-user/aarch64/cpu_loop.c b/linux-user/aarch64/cpu_loop.c index XXXXXXX..XXXXXXX 100644 --- a/linux-user/aarch64/cpu_loop.c +++ b/linux-user/aarch64/cpu_loop.c @@ -XXX,XX +XXX,XX @@ void cpu_loop(CPUARMState *env) switch (trapnr) { case EXCP_SWI: + /* + * On syscall, PSTATE.ZA is preserved, along with the ZA matrix. + * PSTATE.SM is cleared, per SMSTOP, which does ResetSVEState. + */ + if (FIELD_EX64(env->svcr, SVCR, SM)) { + env->svcr = FIELD_DP64(env->svcr, SVCR, SM, 0); + arm_rebuild_hflags(env); + arm_reset_sve_state(env); + } ret = do_syscall(env, env->xregs[8], env->xregs[0], -- 2.25.1 From: Richard Henderson <richard.henderson@linaro.org> Make sure to zero the currently reserved fields. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20220708151540.18136-36-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- linux-user/aarch64/signal.c | 9 ++++++++- 1 file changed, 8 insertions(+), 1 deletion(-) diff --git a/linux-user/aarch64/signal.c b/linux-user/aarch64/signal.c index XXXXXXX..XXXXXXX 100644 --- a/linux-user/aarch64/signal.c +++ b/linux-user/aarch64/signal.c @@ -XXX,XX +XXX,XX @@ struct target_extra_context { struct target_sve_context { struct target_aarch64_ctx head; uint16_t vl; - uint16_t reserved[3]; + uint16_t flags; + uint16_t reserved[2]; /* The actual SVE data immediately follows. It is laid out * according to TARGET_SVE_SIG_{Z,P}REG_OFFSET, based off of * the original struct pointer. @@ -XXX,XX +XXX,XX @@ struct target_sve_context { #define TARGET_SVE_SIG_CONTEXT_SIZE(VQ) \ (TARGET_SVE_SIG_PREG_OFFSET(VQ, 17)) +#define TARGET_SVE_SIG_FLAG_SM 1 + struct target_rt_sigframe { struct target_siginfo info; struct target_ucontext uc; @@ -XXX,XX +XXX,XX @@ static void target_setup_sve_record(struct target_sve_context *sve, { int i, j; + memset(sve, 0, sizeof(*sve)); __put_user(TARGET_SVE_MAGIC, &sve->head.magic); __put_user(size, &sve->head.size); __put_user(vq * TARGET_SVE_VQ_BYTES, &sve->vl); + if (FIELD_EX64(env->svcr, SVCR, SM)) { + __put_user(TARGET_SVE_SIG_FLAG_SM, &sve->flags); + } /* Note that SVE regs are stored as a byte stream, with each byte element * at a subsequent address. This corresponds to a little-endian store -- 2.25.1 From: Richard Henderson <richard.henderson@linaro.org> Fold the return value setting into the goto, so each point of failure need not do both. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20220708151540.18136-37-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- linux-user/aarch64/signal.c | 26 +++++++++++--------------- 1 file changed, 11 insertions(+), 15 deletions(-) diff --git a/linux-user/aarch64/signal.c b/linux-user/aarch64/signal.c index XXXXXXX..XXXXXXX 100644 --- a/linux-user/aarch64/signal.c +++ b/linux-user/aarch64/signal.c @@ -XXX,XX +XXX,XX @@ static int target_restore_sigframe(CPUARMState *env, struct target_sve_context *sve = NULL; uint64_t extra_datap = 0; bool used_extra = false; - bool err = false; int vq = 0, sve_size = 0; target_restore_general_frame(env, sf); @@ -XXX,XX +XXX,XX @@ static int target_restore_sigframe(CPUARMState *env, switch (magic) { case 0: if (size != 0) { - err = true; - goto exit; + goto err; } if (used_extra) { ctx = NULL; @@ -XXX,XX +XXX,XX @@ static int target_restore_sigframe(CPUARMState *env, case TARGET_FPSIMD_MAGIC: if (fpsimd || size != sizeof(struct target_fpsimd_context)) { - err = true; - goto exit; + goto err; } fpsimd = (struct target_fpsimd_context *)ctx; break; @@ -XXX,XX +XXX,XX @@ static int target_restore_sigframe(CPUARMState *env, break; } } - err = true; - goto exit; + goto err; case TARGET_EXTRA_MAGIC: if (extra || size != sizeof(struct target_extra_context)) { - err = true; - goto exit; + goto err; } __get_user(extra_datap, &((struct target_extra_context *)ctx)->datap); @@ -XXX,XX +XXX,XX @@ static int target_restore_sigframe(CPUARMState *env, /* Unknown record -- we certainly didn't generate it. * Did we in fact get out of sync? */ - err = true; - goto exit; + goto err; } ctx = (void *)ctx + size; } @@ -XXX,XX +XXX,XX @@ static int target_restore_sigframe(CPUARMState *env, if (fpsimd) { target_restore_fpsimd_record(env, fpsimd); } else { - err = true; + goto err; } /* SVE data, if present, overwrites FPSIMD data. */ if (sve) { target_restore_sve_record(env, sve, vq); } - - exit: unlock_user(extra, extra_datap, 0); - return err; + return 0; + + err: + unlock_user(extra, extra_datap, 0); + return 1; } static abi_ulong get_sigframe(struct target_sigaction *ka, -- 2.25.1 From: Richard Henderson <richard.henderson@linaro.org> In parse_user_sigframe, the kernel rejects duplicate sve records, or records that are smaller than the header. We were silently allowing these cases to pass, dropping the record. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20220708151540.18136-38-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- linux-user/aarch64/signal.c | 5 ++++- 1 file changed, 4 insertions(+), 1 deletion(-) diff --git a/linux-user/aarch64/signal.c b/linux-user/aarch64/signal.c index XXXXXXX..XXXXXXX 100644 --- a/linux-user/aarch64/signal.c +++ b/linux-user/aarch64/signal.c @@ -XXX,XX +XXX,XX @@ static int target_restore_sigframe(CPUARMState *env, break; case TARGET_SVE_MAGIC: + if (sve || size < sizeof(struct target_sve_context)) { + goto err; + } if (cpu_isar_feature(aa64_sve, env_archcpu(env))) { vq = sve_vq(env); sve_size = QEMU_ALIGN_UP(TARGET_SVE_SIG_CONTEXT_SIZE(vq), 16); - if (!sve && size == sve_size) { + if (size == sve_size) { sve = (struct target_sve_context *)ctx; break; } -- 2.25.1 From: Richard Henderson <richard.henderson@linaro.org> Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20220708151540.18136-39-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- linux-user/aarch64/signal.c | 3 +++ 1 file changed, 3 insertions(+) diff --git a/linux-user/aarch64/signal.c b/linux-user/aarch64/signal.c index XXXXXXX..XXXXXXX 100644 --- a/linux-user/aarch64/signal.c +++ b/linux-user/aarch64/signal.c @@ -XXX,XX +XXX,XX @@ static int target_restore_sigframe(CPUARMState *env, __get_user(extra_size, &((struct target_extra_context *)ctx)->size); extra = lock_user(VERIFY_READ, extra_datap, extra_size, 0); + if (!extra) { + return 1; + } break; default: -- 2.25.1 From: Richard Henderson <richard.henderson@linaro.org> Move the checks out of the parsing loop and into the restore function. This more closely mirrors the code structure in the kernel, and is slightly clearer. Reject rather than silently skip incorrect VL and SVE record sizes, bringing our checks in to line with those the kernel does. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20220708151540.18136-40-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- linux-user/aarch64/signal.c | 51 +++++++++++++++++++++++++------------ 1 file changed, 35 insertions(+), 16 deletions(-) diff --git a/linux-user/aarch64/signal.c b/linux-user/aarch64/signal.c index XXXXXXX..XXXXXXX 100644 --- a/linux-user/aarch64/signal.c +++ b/linux-user/aarch64/signal.c @@ -XXX,XX +XXX,XX @@ static void target_restore_fpsimd_record(CPUARMState *env, } } -static void target_restore_sve_record(CPUARMState *env, - struct target_sve_context *sve, int vq) +static bool target_restore_sve_record(CPUARMState *env, + struct target_sve_context *sve, + int size) { - int i, j; + int i, j, vl, vq; - /* Note that SVE regs are stored as a byte stream, with each byte element + if (!cpu_isar_feature(aa64_sve, env_archcpu(env))) { + return false; + } + + __get_user(vl, &sve->vl); + vq = sve_vq(env); + + /* Reject mismatched VL. */ + if (vl != vq * TARGET_SVE_VQ_BYTES) { + return false; + } + + /* Accept empty record -- used to clear PSTATE.SM. */ + if (size <= sizeof(*sve)) { + return true; + } + + /* Reject non-empty but incomplete record. */ + if (size < TARGET_SVE_SIG_CONTEXT_SIZE(vq)) { + return false; + } + + /* + * Note that SVE regs are stored as a byte stream, with each byte element * at a subsequent address. This corresponds to a little-endian load * of our 64-bit hunks. */ @@ -XXX,XX +XXX,XX @@ static void target_restore_sve_record(CPUARMState *env, } } } + return true; } static int target_restore_sigframe(CPUARMState *env, @@ -XXX,XX +XXX,XX @@ static int target_restore_sigframe(CPUARMState *env, struct target_sve_context *sve = NULL; uint64_t extra_datap = 0; bool used_extra = false; - int vq = 0, sve_size = 0; + int sve_size = 0; target_restore_general_frame(env, sf); @@ -XXX,XX +XXX,XX @@ static int target_restore_sigframe(CPUARMState *env, if (sve || size < sizeof(struct target_sve_context)) { goto err; } - if (cpu_isar_feature(aa64_sve, env_archcpu(env))) { - vq = sve_vq(env); - sve_size = QEMU_ALIGN_UP(TARGET_SVE_SIG_CONTEXT_SIZE(vq), 16); - if (size == sve_size) { - sve = (struct target_sve_context *)ctx; - break; - } - } - goto err; + sve = (struct target_sve_context *)ctx; + sve_size = size; + break; case TARGET_EXTRA_MAGIC: if (extra || size != sizeof(struct target_extra_context)) { @@ -XXX,XX +XXX,XX @@ static int target_restore_sigframe(CPUARMState *env, } /* SVE data, if present, overwrites FPSIMD data. */ - if (sve) { - target_restore_sve_record(env, sve, vq); + if (sve && !target_restore_sve_record(env, sve, sve_size)) { + goto err; } unlock_user(extra, extra_datap, 0); return 0; -- 2.25.1 From: Richard Henderson <richard.henderson@linaro.org> Set the SM bit in the SVE record on signal delivery, create the ZA record. Restore SM and ZA state according to the records present on return. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20220708151540.18136-41-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- linux-user/aarch64/signal.c | 167 +++++++++++++++++++++++++++++++++--- 1 file changed, 154 insertions(+), 13 deletions(-) diff --git a/linux-user/aarch64/signal.c b/linux-user/aarch64/signal.c index XXXXXXX..XXXXXXX 100644 --- a/linux-user/aarch64/signal.c +++ b/linux-user/aarch64/signal.c @@ -XXX,XX +XXX,XX @@ struct target_sve_context { #define TARGET_SVE_SIG_FLAG_SM 1 +#define TARGET_ZA_MAGIC 0x54366345 + +struct target_za_context { + struct target_aarch64_ctx head; + uint16_t vl; + uint16_t reserved[3]; + /* The actual ZA data immediately follows. */ +}; + +#define TARGET_ZA_SIG_REGS_OFFSET \ + QEMU_ALIGN_UP(sizeof(struct target_za_context), TARGET_SVE_VQ_BYTES) +#define TARGET_ZA_SIG_ZAV_OFFSET(VQ, N) \ + (TARGET_ZA_SIG_REGS_OFFSET + (VQ) * TARGET_SVE_VQ_BYTES * (N)) +#define TARGET_ZA_SIG_CONTEXT_SIZE(VQ) \ + TARGET_ZA_SIG_ZAV_OFFSET(VQ, VQ * TARGET_SVE_VQ_BYTES) + struct target_rt_sigframe { struct target_siginfo info; struct target_ucontext uc; @@ -XXX,XX +XXX,XX @@ static void target_setup_end_record(struct target_aarch64_ctx *end) } static void target_setup_sve_record(struct target_sve_context *sve, - CPUARMState *env, int vq, int size) + CPUARMState *env, int size) { - int i, j; + int i, j, vq = sve_vq(env); memset(sve, 0, sizeof(*sve)); __put_user(TARGET_SVE_MAGIC, &sve->head.magic); @@ -XXX,XX +XXX,XX @@ static void target_setup_sve_record(struct target_sve_context *sve, } } +static void target_setup_za_record(struct target_za_context *za, + CPUARMState *env, int size) +{ + int vq = sme_vq(env); + int vl = vq * TARGET_SVE_VQ_BYTES; + int i, j; + + memset(za, 0, sizeof(*za)); + __put_user(TARGET_ZA_MAGIC, &za->head.magic); + __put_user(size, &za->head.size); + __put_user(vl, &za->vl); + + if (size == TARGET_ZA_SIG_CONTEXT_SIZE(0)) { + return; + } + assert(size == TARGET_ZA_SIG_CONTEXT_SIZE(vq)); + + /* + * Note that ZA vectors are stored as a byte stream, + * with each byte element at a subsequent address. + */ + for (i = 0; i < vl; ++i) { + uint64_t *z = (void *)za + TARGET_ZA_SIG_ZAV_OFFSET(vq, i); + for (j = 0; j < vq * 2; ++j) { + __put_user_e(env->zarray[i].d[j], z + j, le); + } + } +} + static void target_restore_general_frame(CPUARMState *env, struct target_rt_sigframe *sf) { @@ -XXX,XX +XXX,XX @@ static void target_restore_fpsimd_record(CPUARMState *env, static bool target_restore_sve_record(CPUARMState *env, struct target_sve_context *sve, - int size) + int size, int *svcr) { - int i, j, vl, vq; + int i, j, vl, vq, flags; + bool sm; - if (!cpu_isar_feature(aa64_sve, env_archcpu(env))) { + __get_user(vl, &sve->vl); + __get_user(flags, &sve->flags); + + sm = flags & TARGET_SVE_SIG_FLAG_SM; + + /* The cpu must support Streaming or Non-streaming SVE. */ + if (sm + ? !cpu_isar_feature(aa64_sme, env_archcpu(env)) + : !cpu_isar_feature(aa64_sve, env_archcpu(env))) { return false; } - __get_user(vl, &sve->vl); - vq = sve_vq(env); + /* + * Note that we cannot use sve_vq() because that depends on the + * current setting of PSTATE.SM, not the state to be restored. + */ + vq = sve_vqm1_for_el_sm(env, 0, sm) + 1; /* Reject mismatched VL. */ if (vl != vq * TARGET_SVE_VQ_BYTES) { @@ -XXX,XX +XXX,XX @@ static bool target_restore_sve_record(CPUARMState *env, return false; } + *svcr = FIELD_DP64(*svcr, SVCR, SM, sm); + /* * Note that SVE regs are stored as a byte stream, with each byte element * at a subsequent address. This corresponds to a little-endian load @@ -XXX,XX +XXX,XX @@ static bool target_restore_sve_record(CPUARMState *env, return true; } +static bool target_restore_za_record(CPUARMState *env, + struct target_za_context *za, + int size, int *svcr) +{ + int i, j, vl, vq; + + if (!cpu_isar_feature(aa64_sme, env_archcpu(env))) { + return false; + } + + __get_user(vl, &za->vl); + vq = sme_vq(env); + + /* Reject mismatched VL. */ + if (vl != vq * TARGET_SVE_VQ_BYTES) { + return false; + } + + /* Accept empty record -- used to clear PSTATE.ZA. */ + if (size <= TARGET_ZA_SIG_CONTEXT_SIZE(0)) { + return true; + } + + /* Reject non-empty but incomplete record. */ + if (size < TARGET_ZA_SIG_CONTEXT_SIZE(vq)) { + return false; + } + + *svcr = FIELD_DP64(*svcr, SVCR, ZA, 1); + + for (i = 0; i < vl; ++i) { + uint64_t *z = (void *)za + TARGET_ZA_SIG_ZAV_OFFSET(vq, i); + for (j = 0; j < vq * 2; ++j) { + __get_user_e(env->zarray[i].d[j], z + j, le); + } + } + return true; +} + static int target_restore_sigframe(CPUARMState *env, struct target_rt_sigframe *sf) { struct target_aarch64_ctx *ctx, *extra = NULL; struct target_fpsimd_context *fpsimd = NULL; struct target_sve_context *sve = NULL; + struct target_za_context *za = NULL; uint64_t extra_datap = 0; bool used_extra = false; int sve_size = 0; + int za_size = 0; + int svcr = 0; target_restore_general_frame(env, sf); @@ -XXX,XX +XXX,XX @@ static int target_restore_sigframe(CPUARMState *env, sve_size = size; break; + case TARGET_ZA_MAGIC: + if (za || size < sizeof(struct target_za_context)) { + goto err; + } + za = (struct target_za_context *)ctx; + za_size = size; + break; + case TARGET_EXTRA_MAGIC: if (extra || size != sizeof(struct target_extra_context)) { goto err; @@ -XXX,XX +XXX,XX @@ static int target_restore_sigframe(CPUARMState *env, } /* SVE data, if present, overwrites FPSIMD data. */ - if (sve && !target_restore_sve_record(env, sve, sve_size)) { + if (sve && !target_restore_sve_record(env, sve, sve_size, &svcr)) { goto err; } + if (za && !target_restore_za_record(env, za, za_size, &svcr)) { + goto err; + } + if (env->svcr != svcr) { + env->svcr = svcr; + arm_rebuild_hflags(env); + } unlock_user(extra, extra_datap, 0); return 0; @@ -XXX,XX +XXX,XX @@ static void target_setup_frame(int usig, struct target_sigaction *ka, .total_size = offsetof(struct target_rt_sigframe, uc.tuc_mcontext.__reserved), }; - int fpsimd_ofs, fr_ofs, sve_ofs = 0, vq = 0, sve_size = 0; + int fpsimd_ofs, fr_ofs, sve_ofs = 0, za_ofs = 0; + int sve_size = 0, za_size = 0; struct target_rt_sigframe *frame; struct target_rt_frame_record *fr; abi_ulong frame_addr, return_addr; @@ -XXX,XX +XXX,XX @@ static void target_setup_frame(int usig, struct target_sigaction *ka, &layout); /* SVE state needs saving only if it exists. */ - if (cpu_isar_feature(aa64_sve, env_archcpu(env))) { - vq = sve_vq(env); - sve_size = QEMU_ALIGN_UP(TARGET_SVE_SIG_CONTEXT_SIZE(vq), 16); + if (cpu_isar_feature(aa64_sve, env_archcpu(env)) || + cpu_isar_feature(aa64_sme, env_archcpu(env))) { + sve_size = QEMU_ALIGN_UP(TARGET_SVE_SIG_CONTEXT_SIZE(sve_vq(env)), 16); sve_ofs = alloc_sigframe_space(sve_size, &layout); } + if (cpu_isar_feature(aa64_sme, env_archcpu(env))) { + /* ZA state needs saving only if it is enabled. */ + if (FIELD_EX64(env->svcr, SVCR, ZA)) { + za_size = TARGET_ZA_SIG_CONTEXT_SIZE(sme_vq(env)); + } else { + za_size = TARGET_ZA_SIG_CONTEXT_SIZE(0); + } + za_ofs = alloc_sigframe_space(za_size, &layout); + } if (layout.extra_ofs) { /* Reserve space for the extra end marker. The standard end marker @@ -XXX,XX +XXX,XX @@ static void target_setup_frame(int usig, struct target_sigaction *ka, target_setup_end_record((void *)frame + layout.extra_end_ofs); } if (sve_ofs) { - target_setup_sve_record((void *)frame + sve_ofs, env, vq, sve_size); + target_setup_sve_record((void *)frame + sve_ofs, env, sve_size); + } + if (za_ofs) { + target_setup_za_record((void *)frame + za_ofs, env, za_size); } /* Set up the stack frame for unwinding. */ @@ -XXX,XX +XXX,XX @@ static void target_setup_frame(int usig, struct target_sigaction *ka, env->btype = 2; } + /* + * Invoke the signal handler with both SM and ZA disabled. + * When clearing SM, ResetSVEState, per SMSTOP. + */ + if (FIELD_EX64(env->svcr, SVCR, SM)) { + arm_reset_sve_state(env); + } + if (env->svcr) { + env->svcr = 0; + arm_rebuild_hflags(env); + } + if (info) { tswap_siginfo(&frame->info, info); env->xregs[1] = frame_addr + offsetof(struct target_rt_sigframe, info); -- 2.25.1 From: Richard Henderson <richard.henderson@linaro.org> Add "sve" to the sve prctl functions, to distinguish them from the coming "sme" prctls with similar names. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20220708151540.18136-42-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- linux-user/aarch64/target_prctl.h | 8 ++++---- linux-user/syscall.c | 12 ++++++------ 2 files changed, 10 insertions(+), 10 deletions(-) diff --git a/linux-user/aarch64/target_prctl.h b/linux-user/aarch64/target_prctl.h index XXXXXXX..XXXXXXX 100644 --- a/linux-user/aarch64/target_prctl.h +++ b/linux-user/aarch64/target_prctl.h @@ -XXX,XX +XXX,XX @@ #ifndef AARCH64_TARGET_PRCTL_H #define AARCH64_TARGET_PRCTL_H -static abi_long do_prctl_get_vl(CPUArchState *env) +static abi_long do_prctl_sve_get_vl(CPUArchState *env) { ARMCPU *cpu = env_archcpu(env); if (cpu_isar_feature(aa64_sve, cpu)) { @@ -XXX,XX +XXX,XX @@ static abi_long do_prctl_get_vl(CPUArchState *env) } return -TARGET_EINVAL; } -#define do_prctl_get_vl do_prctl_get_vl +#define do_prctl_sve_get_vl do_prctl_sve_get_vl -static abi_long do_prctl_set_vl(CPUArchState *env, abi_long arg2) +static abi_long do_prctl_sve_set_vl(CPUArchState *env, abi_long arg2) { /* * We cannot support either PR_SVE_SET_VL_ONEXEC or PR_SVE_VL_INHERIT. @@ -XXX,XX +XXX,XX @@ static abi_long do_prctl_set_vl(CPUArchState *env, abi_long arg2) } return -TARGET_EINVAL; } -#define do_prctl_set_vl do_prctl_set_vl +#define do_prctl_sve_set_vl do_prctl_sve_set_vl static abi_long do_prctl_reset_keys(CPUArchState *env, abi_long arg2) { diff --git a/linux-user/syscall.c b/linux-user/syscall.c index XXXXXXX..XXXXXXX 100644 --- a/linux-user/syscall.c +++ b/linux-user/syscall.c @@ -XXX,XX +XXX,XX @@ static abi_long do_prctl_inval1(CPUArchState *env, abi_long arg2) #ifndef do_prctl_set_fp_mode #define do_prctl_set_fp_mode do_prctl_inval1 #endif -#ifndef do_prctl_get_vl -#define do_prctl_get_vl do_prctl_inval0 +#ifndef do_prctl_sve_get_vl +#define do_prctl_sve_get_vl do_prctl_inval0 #endif -#ifndef do_prctl_set_vl -#define do_prctl_set_vl do_prctl_inval1 +#ifndef do_prctl_sve_set_vl +#define do_prctl_sve_set_vl do_prctl_inval1 #endif #ifndef do_prctl_reset_keys #define do_prctl_reset_keys do_prctl_inval1 @@ -XXX,XX +XXX,XX @@ static abi_long do_prctl(CPUArchState *env, abi_long option, abi_long arg2, case PR_SET_FP_MODE: return do_prctl_set_fp_mode(env, arg2); case PR_SVE_GET_VL: - return do_prctl_get_vl(env); + return do_prctl_sve_get_vl(env); case PR_SVE_SET_VL: - return do_prctl_set_vl(env, arg2); + return do_prctl_sve_set_vl(env, arg2); case PR_PAC_RESET_KEYS: if (arg3 || arg4 || arg5) { return -TARGET_EINVAL; -- 2.25.1 From: Richard Henderson <richard.henderson@linaro.org> These prctl set the Streaming SVE vector length, which may be completely different from the Normal SVE vector length. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20220708151540.18136-43-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- linux-user/aarch64/target_prctl.h | 54 +++++++++++++++++++++++++++++++ linux-user/syscall.c | 16 +++++++++ 2 files changed, 70 insertions(+) diff --git a/linux-user/aarch64/target_prctl.h b/linux-user/aarch64/target_prctl.h index XXXXXXX..XXXXXXX 100644 --- a/linux-user/aarch64/target_prctl.h +++ b/linux-user/aarch64/target_prctl.h @@ -XXX,XX +XXX,XX @@ static abi_long do_prctl_sve_get_vl(CPUArchState *env) { ARMCPU *cpu = env_archcpu(env); if (cpu_isar_feature(aa64_sve, cpu)) { + /* PSTATE.SM is always unset on syscall entry. */ return sve_vq(env) * 16; } return -TARGET_EINVAL; @@ -XXX,XX +XXX,XX @@ static abi_long do_prctl_sve_set_vl(CPUArchState *env, abi_long arg2) && arg2 >= 0 && arg2 <= 512 * 16 && !(arg2 & 15)) { uint32_t vq, old_vq; + /* PSTATE.SM is always unset on syscall entry. */ old_vq = sve_vq(env); /* @@ -XXX,XX +XXX,XX @@ static abi_long do_prctl_sve_set_vl(CPUArchState *env, abi_long arg2) } #define do_prctl_sve_set_vl do_prctl_sve_set_vl +static abi_long do_prctl_sme_get_vl(CPUArchState *env) +{ + ARMCPU *cpu = env_archcpu(env); + if (cpu_isar_feature(aa64_sme, cpu)) { + return sme_vq(env) * 16; + } + return -TARGET_EINVAL; +} +#define do_prctl_sme_get_vl do_prctl_sme_get_vl + +static abi_long do_prctl_sme_set_vl(CPUArchState *env, abi_long arg2) +{ + /* + * We cannot support either PR_SME_SET_VL_ONEXEC or PR_SME_VL_INHERIT. + * Note the kernel definition of sve_vl_valid allows for VQ=512, + * i.e. VL=8192, even though the architectural maximum is VQ=16. + */ + if (cpu_isar_feature(aa64_sme, env_archcpu(env)) + && arg2 >= 0 && arg2 <= 512 * 16 && !(arg2 & 15)) { + int vq, old_vq; + + old_vq = sme_vq(env); + + /* + * Bound the value of vq, so that we know that it fits into + * the 4-bit field in SMCR_EL1. Because PSTATE.SM is cleared + * on syscall entry, we are not modifying the current SVE + * vector length. + */ + vq = MAX(arg2 / 16, 1); + vq = MIN(vq, 16); + env->vfp.smcr_el[1] = + FIELD_DP64(env->vfp.smcr_el[1], SMCR, LEN, vq - 1); + + /* Delay rebuilding hflags until we know if ZA must change. */ + vq = sve_vqm1_for_el_sm(env, 0, true) + 1; + + if (vq != old_vq) { + /* + * PSTATE.ZA state is cleared on any change to SVL. + * We need not call arm_rebuild_hflags because PSTATE.SM was + * cleared on syscall entry, so this hasn't changed VL. + */ + env->svcr = FIELD_DP64(env->svcr, SVCR, ZA, 0); + arm_rebuild_hflags(env); + } + return vq * 16; + } + return -TARGET_EINVAL; +} +#define do_prctl_sme_set_vl do_prctl_sme_set_vl + static abi_long do_prctl_reset_keys(CPUArchState *env, abi_long arg2) { ARMCPU *cpu = env_archcpu(env); diff --git a/linux-user/syscall.c b/linux-user/syscall.c index XXXXXXX..XXXXXXX 100644 --- a/linux-user/syscall.c +++ b/linux-user/syscall.c @@ -XXX,XX +XXX,XX @@ abi_long do_arch_prctl(CPUX86State *env, int code, abi_ulong addr) #ifndef PR_SET_SYSCALL_USER_DISPATCH # define PR_SET_SYSCALL_USER_DISPATCH 59 #endif +#ifndef PR_SME_SET_VL +# define PR_SME_SET_VL 63 +# define PR_SME_GET_VL 64 +# define PR_SME_VL_LEN_MASK 0xffff +# define PR_SME_VL_INHERIT (1 << 17) +#endif #include "target_prctl.h" @@ -XXX,XX +XXX,XX @@ static abi_long do_prctl_inval1(CPUArchState *env, abi_long arg2) #ifndef do_prctl_set_unalign #define do_prctl_set_unalign do_prctl_inval1 #endif +#ifndef do_prctl_sme_get_vl +#define do_prctl_sme_get_vl do_prctl_inval0 +#endif +#ifndef do_prctl_sme_set_vl +#define do_prctl_sme_set_vl do_prctl_inval1 +#endif static abi_long do_prctl(CPUArchState *env, abi_long option, abi_long arg2, abi_long arg3, abi_long arg4, abi_long arg5) @@ -XXX,XX +XXX,XX @@ static abi_long do_prctl(CPUArchState *env, abi_long option, abi_long arg2, return do_prctl_sve_get_vl(env); case PR_SVE_SET_VL: return do_prctl_sve_set_vl(env, arg2); + case PR_SME_GET_VL: + return do_prctl_sme_get_vl(env); + case PR_SME_SET_VL: + return do_prctl_sme_set_vl(env, arg2); case PR_PAC_RESET_KEYS: if (arg3 || arg4 || arg5) { return -TARGET_EINVAL; -- 2.25.1 From: Richard Henderson <richard.henderson@linaro.org> There's no reason to set CPACR_EL1.ZEN if SVE disabled. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20220708151540.18136-44-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/cpu.c | 7 +++---- 1 file changed, 3 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 @@ static void arm_cpu_reset(DeviceState *dev) /* and to the FP/Neon instructions */ env->cp15.cpacr_el1 = FIELD_DP64(env->cp15.cpacr_el1, CPACR_EL1, FPEN, 3); - /* and to the SVE instructions */ - env->cp15.cpacr_el1 = FIELD_DP64(env->cp15.cpacr_el1, - CPACR_EL1, ZEN, 3); - /* with reasonable vector length */ + /* and to the SVE instructions, with default vector length */ if (cpu_isar_feature(aa64_sve, cpu)) { + env->cp15.cpacr_el1 = FIELD_DP64(env->cp15.cpacr_el1, + CPACR_EL1, ZEN, 3); env->vfp.zcr_el[1] = cpu->sve_default_vq - 1; } /* -- 2.25.1 From: Richard Henderson <richard.henderson@linaro.org> Enable SME, TPIDR2_EL0, and FA64 if supported by the cpu. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20220708151540.18136-45-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/cpu.c | 11 +++++++++++ 1 file changed, 11 insertions(+) diff --git a/target/arm/cpu.c b/target/arm/cpu.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/cpu.c +++ b/target/arm/cpu.c @@ -XXX,XX +XXX,XX @@ static void arm_cpu_reset(DeviceState *dev) CPACR_EL1, ZEN, 3); env->vfp.zcr_el[1] = cpu->sve_default_vq - 1; } + /* and for SME instructions, with default vector length, and TPIDR2 */ + if (cpu_isar_feature(aa64_sme, cpu)) { + env->cp15.sctlr_el[1] |= SCTLR_EnTP2; + env->cp15.cpacr_el1 = FIELD_DP64(env->cp15.cpacr_el1, + CPACR_EL1, SMEN, 3); + env->vfp.smcr_el[1] = cpu->sme_default_vq - 1; + if (cpu_isar_feature(aa64_sme_fa64, cpu)) { + env->vfp.smcr_el[1] = FIELD_DP64(env->vfp.smcr_el[1], + SMCR, FA64, 1); + } + } /* * Enable 48-bit address space (TODO: take reserved_va into account). * Enable TBI0 but not TBI1. -- 2.25.1 From: Richard Henderson <richard.henderson@linaro.org> Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20220708151540.18136-46-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- linux-user/elfload.c | 20 ++++++++++++++++++++ 1 file changed, 20 insertions(+) diff --git a/linux-user/elfload.c b/linux-user/elfload.c index XXXXXXX..XXXXXXX 100644 --- a/linux-user/elfload.c +++ b/linux-user/elfload.c @@ -XXX,XX +XXX,XX @@ enum { ARM_HWCAP2_A64_RNG = 1 << 16, ARM_HWCAP2_A64_BTI = 1 << 17, ARM_HWCAP2_A64_MTE = 1 << 18, + ARM_HWCAP2_A64_ECV = 1 << 19, + ARM_HWCAP2_A64_AFP = 1 << 20, + ARM_HWCAP2_A64_RPRES = 1 << 21, + ARM_HWCAP2_A64_MTE3 = 1 << 22, + ARM_HWCAP2_A64_SME = 1 << 23, + ARM_HWCAP2_A64_SME_I16I64 = 1 << 24, + ARM_HWCAP2_A64_SME_F64F64 = 1 << 25, + ARM_HWCAP2_A64_SME_I8I32 = 1 << 26, + ARM_HWCAP2_A64_SME_F16F32 = 1 << 27, + ARM_HWCAP2_A64_SME_B16F32 = 1 << 28, + ARM_HWCAP2_A64_SME_F32F32 = 1 << 29, + ARM_HWCAP2_A64_SME_FA64 = 1 << 30, }; #define ELF_HWCAP get_elf_hwcap() @@ -XXX,XX +XXX,XX @@ static uint32_t get_elf_hwcap2(void) GET_FEATURE_ID(aa64_rndr, ARM_HWCAP2_A64_RNG); GET_FEATURE_ID(aa64_bti, ARM_HWCAP2_A64_BTI); GET_FEATURE_ID(aa64_mte, ARM_HWCAP2_A64_MTE); + GET_FEATURE_ID(aa64_sme, (ARM_HWCAP2_A64_SME | + ARM_HWCAP2_A64_SME_F32F32 | + ARM_HWCAP2_A64_SME_B16F32 | + ARM_HWCAP2_A64_SME_F16F32 | + ARM_HWCAP2_A64_SME_I8I32)); + GET_FEATURE_ID(aa64_sme_f64f64, ARM_HWCAP2_A64_SME_F64F64); + GET_FEATURE_ID(aa64_sme_i16i64, ARM_HWCAP2_A64_SME_I16I64); + GET_FEATURE_ID(aa64_sme_fa64, ARM_HWCAP2_A64_SME_FA64); return hwcaps; } -- 2.25.1