A grab-bag of minor stuff for the end of the year. My to-review queue is not empty, but it it at least in single figures... -- PMM The following changes since commit 5bfbd8170ce7acb98a1834ff49ed7340b0837144: Merge remote-tracking branch 'remotes/vivier2/tags/trivial-branch-for-6.0-pull-request' into staging (2020-12-14 20:32:38 +0000) are available in the Git repository at: https://git.linaro.org/people/pmaydell/qemu-arm.git tags/pull-target-arm-20201215 for you to fetch changes up to 23af268566069183285bebbdf95b1b37cb7c0942: hw/block/m25p80: Fix Numonyx fast read dummy cycle count (2020-12-15 13:39:30 +0000) ---------------------------------------------------------------- target-arm queue: * gdbstub: Correct misparsing of vCont C/S requests * openrisc: Move pic_cpu code into CPU object proper * nios2: Move IIC code into CPU object proper * Improve reporting of ROM overlap errors * xlnx-versal: Add USB support * hw/misc/zynq_slcr: Avoid #DIV/0! error * Numonyx: Fix dummy cycles and check for SPI mode on cmds ---------------------------------------------------------------- Joe Komlodi (4): hw/block/m25p80: Make Numonyx config field names more accurate hw/block/m25p80: Fix when VCFG XIP bit is set for Numonyx hw/block/m25p80: Check SPI mode before running some Numonyx commands hw/block/m25p80: Fix Numonyx fast read dummy cycle count Peter Maydell (11): gdbstub: Correct misparsing of vCont C/S requests hw/openrisc/openrisc_sim: Use IRQ splitter when connecting IRQ to multiple CPUs hw/openrisc/openrisc_sim: Abstract out "get IRQ x of CPU y" target/openrisc: Move pic_cpu code into CPU object proper target/nios2: Move IIC code into CPU object proper target/nios2: Move nios2_check_interrupts() into target/nios2 target/nios2: Use deposit32() to update ipending register hw/core/loader.c: Track last-seen ROM in rom_check_and_register_reset() hw/core/loader.c: Improve reporting of ROM overlap errors elf_ops.h: Don't truncate name of the ROM blobs we create elf_ops.h: Be more verbose with ROM blob names Philippe Mathieu-Daudé (1): hw/misc/zynq_slcr: Avoid #DIV/0! error Sai Pavan Boddu (2): usb: Add versal-usb2-ctrl-regs module usb: xlnx-usb-subsystem: Add xilinx usb subsystem Vikram Garhwal (2): usb: Add DWC3 model arm: xlnx-versal: Connect usb to virt-versal include/hw/arm/xlnx-versal.h | 9 + include/hw/elf_ops.h | 5 +- include/hw/usb/hcd-dwc3.h | 55 +++ include/hw/usb/xlnx-usb-subsystem.h | 45 ++ include/hw/usb/xlnx-versal-usb2-ctrl-regs.h | 45 ++ target/nios2/cpu.h | 3 - target/openrisc/cpu.h | 1 - gdbstub.c | 2 +- hw/arm/xlnx-versal-virt.c | 55 +++ hw/arm/xlnx-versal.c | 26 ++ hw/block/m25p80.c | 158 +++++-- hw/core/loader.c | 67 ++- hw/intc/nios2_iic.c | 95 ---- hw/misc/zynq_slcr.c | 5 + hw/nios2/10m50_devboard.c | 13 +- hw/nios2/cpu_pic.c | 67 --- hw/openrisc/openrisc_sim.c | 46 +- hw/openrisc/pic_cpu.c | 61 --- hw/usb/hcd-dwc3.c | 689 ++++++++++++++++++++++++++++ hw/usb/xlnx-usb-subsystem.c | 94 ++++ hw/usb/xlnx-versal-usb2-ctrl-regs.c | 229 +++++++++ softmmu/vl.c | 1 - target/nios2/cpu.c | 29 ++ target/nios2/op_helper.c | 9 + target/openrisc/cpu.c | 32 ++ MAINTAINERS | 1 - hw/intc/meson.build | 1 - hw/nios2/meson.build | 2 +- hw/openrisc/Kconfig | 1 + hw/openrisc/meson.build | 2 +- hw/usb/Kconfig | 10 + hw/usb/meson.build | 3 + 32 files changed, 1557 insertions(+), 304 deletions(-) create mode 100644 include/hw/usb/hcd-dwc3.h create mode 100644 include/hw/usb/xlnx-usb-subsystem.h create mode 100644 include/hw/usb/xlnx-versal-usb2-ctrl-regs.h delete mode 100644 hw/intc/nios2_iic.c delete mode 100644 hw/nios2/cpu_pic.c delete mode 100644 hw/openrisc/pic_cpu.c create mode 100644 hw/usb/hcd-dwc3.c create mode 100644 hw/usb/xlnx-usb-subsystem.c create mode 100644 hw/usb/xlnx-versal-usb2-ctrl-regs.c In the vCont packet, two of the command actions (C and S) take an argument specifying the signal to be sent to the process/thread, which is sent as an ASCII string of two hex digits which immediately follow the 'C' or 'S' character. Our code for parsing this packet accidentally skipped the first of the two bytes of the signal value, because it started parsing the hex string at 'p + 1' when the preceding code had already moved past the 'C' or 'S' with "cur_action = *p++". This meant that we would only do the right thing for signals below 10, and would misinterpret the rest. For instance, when the debugger wants to send the process a SIGPROF (27 on x86-64) we mangle this into a SIGSEGV (11). Remove the accidental double increment. Fixes: https://bugs.launchpad.net/qemu/+bug/1773743 Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com> Reviewed-by: Alex Bennée <alex.bennee@linaro.org> Message-id: 20201121210342.10089-1-peter.maydell@linaro.org --- gdbstub.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/gdbstub.c b/gdbstub.c index XXXXXXX..XXXXXXX 100644 --- a/gdbstub.c +++ b/gdbstub.c @@ -XXX,XX +XXX,XX @@ static int gdb_handle_vcont(const char *p) cur_action = *p++; if (cur_action == 'C' || cur_action == 'S') { cur_action = qemu_tolower(cur_action); - res = qemu_strtoul(p + 1, &p, 16, &tmp); + res = qemu_strtoul(p, &p, 16, &tmp); if (res) { goto out; } -- 2.20.1 openrisc_sim_net_init() attempts to connect the IRQ line from the ethernet device to both CPUs in an SMP configuration by simply caling sysbus_connect_irq() for it twice. This doesn't work, because the second connection simply overrides the first. Fix this by creating a TYPE_SPLIT_IRQ to split the IRQ in the SMP case. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Stafford Horne <shorne@gmail.com> Message-id: 20201127225127.14770-2-peter.maydell@linaro.org --- hw/openrisc/openrisc_sim.c | 13 +++++++++++-- hw/openrisc/Kconfig | 1 + 2 files changed, 12 insertions(+), 2 deletions(-) diff --git a/hw/openrisc/openrisc_sim.c b/hw/openrisc/openrisc_sim.c index XXXXXXX..XXXXXXX 100644 --- a/hw/openrisc/openrisc_sim.c +++ b/hw/openrisc/openrisc_sim.c @@ -XXX,XX +XXX,XX @@ #include "hw/sysbus.h" #include "sysemu/qtest.h" #include "sysemu/reset.h" +#include "hw/core/split-irq.h" #define KERNEL_LOAD_ADDR 0x100 @@ -XXX,XX +XXX,XX @@ static void openrisc_sim_net_init(hwaddr base, hwaddr descriptors, s = SYS_BUS_DEVICE(dev); sysbus_realize_and_unref(s, &error_fatal); - for (i = 0; i < num_cpus; i++) { - sysbus_connect_irq(s, 0, cpu_irqs[i][irq_pin]); + if (num_cpus > 1) { + DeviceState *splitter = qdev_new(TYPE_SPLIT_IRQ); + qdev_prop_set_uint32(splitter, "num-lines", num_cpus); + qdev_realize_and_unref(splitter, NULL, &error_fatal); + for (i = 0; i < num_cpus; i++) { + qdev_connect_gpio_out(splitter, i, cpu_irqs[i][irq_pin]); + } + sysbus_connect_irq(s, 0, qdev_get_gpio_in(splitter, 0)); + } else { + sysbus_connect_irq(s, 0, cpu_irqs[0][irq_pin]); } sysbus_mmio_map(s, 0, base); sysbus_mmio_map(s, 1, descriptors); diff --git a/hw/openrisc/Kconfig b/hw/openrisc/Kconfig index XXXXXXX..XXXXXXX 100644 --- a/hw/openrisc/Kconfig +++ b/hw/openrisc/Kconfig @@ -XXX,XX +XXX,XX @@ config OR1K_SIM select SERIAL select OPENCORES_ETH select OMPIC + select SPLIT_IRQ -- 2.20.1 We're about to refactor the OpenRISC pic_cpu code in a way that means that just grabbing the whole qemu_irq[] array of inbound IRQs for a CPU won't be possible any more. Abstract out a function for "return the qemu_irq for IRQ x input of CPU y" so we can more easily replace the implementation. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Stafford Horne <shorne@gmail.com> Message-id: 20201127225127.14770-3-peter.maydell@linaro.org --- hw/openrisc/openrisc_sim.c | 38 +++++++++++++++++++++----------------- 1 file changed, 21 insertions(+), 17 deletions(-) diff --git a/hw/openrisc/openrisc_sim.c b/hw/openrisc/openrisc_sim.c index XXXXXXX..XXXXXXX 100644 --- a/hw/openrisc/openrisc_sim.c +++ b/hw/openrisc/openrisc_sim.c @@ -XXX,XX +XXX,XX @@ static void main_cpu_reset(void *opaque) cpu_set_pc(cs, boot_info.bootstrap_pc); } +static qemu_irq get_cpu_irq(OpenRISCCPU *cpus[], int cpunum, int irq_pin) +{ + return cpus[cpunum]->env.irq[irq_pin]; +} + static void openrisc_sim_net_init(hwaddr base, hwaddr descriptors, - int num_cpus, qemu_irq **cpu_irqs, + int num_cpus, OpenRISCCPU *cpus[], int irq_pin, NICInfo *nd) { DeviceState *dev; @@ -XXX,XX +XXX,XX @@ static void openrisc_sim_net_init(hwaddr base, hwaddr descriptors, qdev_prop_set_uint32(splitter, "num-lines", num_cpus); qdev_realize_and_unref(splitter, NULL, &error_fatal); for (i = 0; i < num_cpus; i++) { - qdev_connect_gpio_out(splitter, i, cpu_irqs[i][irq_pin]); + qdev_connect_gpio_out(splitter, i, get_cpu_irq(cpus, i, irq_pin)); } sysbus_connect_irq(s, 0, qdev_get_gpio_in(splitter, 0)); } else { - sysbus_connect_irq(s, 0, cpu_irqs[0][irq_pin]); + sysbus_connect_irq(s, 0, get_cpu_irq(cpus, 0, irq_pin)); } sysbus_mmio_map(s, 0, base); sysbus_mmio_map(s, 1, descriptors); } static void openrisc_sim_ompic_init(hwaddr base, int num_cpus, - qemu_irq **cpu_irqs, int irq_pin) + OpenRISCCPU *cpus[], int irq_pin) { DeviceState *dev; SysBusDevice *s; @@ -XXX,XX +XXX,XX @@ static void openrisc_sim_ompic_init(hwaddr base, int num_cpus, s = SYS_BUS_DEVICE(dev); sysbus_realize_and_unref(s, &error_fatal); for (i = 0; i < num_cpus; i++) { - sysbus_connect_irq(s, i, cpu_irqs[i][irq_pin]); + sysbus_connect_irq(s, i, get_cpu_irq(cpus, i, irq_pin)); } sysbus_mmio_map(s, 0, base); } @@ -XXX,XX +XXX,XX @@ static void openrisc_sim_init(MachineState *machine) { ram_addr_t ram_size = machine->ram_size; const char *kernel_filename = machine->kernel_filename; - OpenRISCCPU *cpu = NULL; + OpenRISCCPU *cpus[2] = {}; MemoryRegion *ram; - qemu_irq *cpu_irqs[2]; qemu_irq serial_irq; int n; unsigned int smp_cpus = machine->smp.cpus; assert(smp_cpus >= 1 && smp_cpus <= 2); for (n = 0; n < smp_cpus; n++) { - cpu = OPENRISC_CPU(cpu_create(machine->cpu_type)); - if (cpu == NULL) { + cpus[n] = OPENRISC_CPU(cpu_create(machine->cpu_type)); + if (cpus[n] == NULL) { fprintf(stderr, "Unable to find CPU definition!\n"); exit(1); } - cpu_openrisc_pic_init(cpu); - cpu_irqs[n] = (qemu_irq *) cpu->env.irq; + cpu_openrisc_pic_init(cpus[n]); - cpu_openrisc_clock_init(cpu); + cpu_openrisc_clock_init(cpus[n]); - qemu_register_reset(main_cpu_reset, cpu); + qemu_register_reset(main_cpu_reset, cpus[n]); } ram = g_malloc(sizeof(*ram)); @@ -XXX,XX +XXX,XX @@ static void openrisc_sim_init(MachineState *machine) if (nd_table[0].used) { openrisc_sim_net_init(0x92000000, 0x92000400, smp_cpus, - cpu_irqs, 4, nd_table); + cpus, 4, nd_table); } if (smp_cpus > 1) { - openrisc_sim_ompic_init(0x98000000, smp_cpus, cpu_irqs, 1); + openrisc_sim_ompic_init(0x98000000, smp_cpus, cpus, 1); - serial_irq = qemu_irq_split(cpu_irqs[0][2], cpu_irqs[1][2]); + serial_irq = qemu_irq_split(get_cpu_irq(cpus, 0, 2), + get_cpu_irq(cpus, 1, 2)); } else { - serial_irq = cpu_irqs[0][2]; + serial_irq = get_cpu_irq(cpus, 0, 2); } serial_mm_init(get_system_memory(), 0x90000000, 0, serial_irq, -- 2.20.1 The openrisc code uses an old style of interrupt handling, where a separate standalone set of qemu_irqs invoke a function openrisc_pic_cpu_handler() which signals the interrupt to the CPU proper by directly calling cpu_interrupt() and cpu_reset_interrupt(). Because CPU objects now inherit (indirectly) from TYPE_DEVICE, they can have GPIO input lines themselves, and the neater modern way to implement this is to simply have the CPU object itself provide the input IRQ lines. Create GPIO inputs to the OpenRISC CPU object, and make the only user of cpu_openrisc_pic_init() wire up directly to those instead. This allows us to delete the hw/openrisc/pic_cpu.c file entirely. This fixes a trivial memory leak reported by Coverity of the IRQs allocated in cpu_openrisc_pic_init(). Fixes: Coverity CID 1421934 Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Stafford Horne <shorne@gmail.com> Message-id: 20201127225127.14770-4-peter.maydell@linaro.org --- target/openrisc/cpu.h | 1 - hw/openrisc/openrisc_sim.c | 3 +- hw/openrisc/pic_cpu.c | 61 -------------------------------------- target/openrisc/cpu.c | 32 ++++++++++++++++++++ hw/openrisc/meson.build | 2 +- 5 files changed, 34 insertions(+), 65 deletions(-) delete mode 100644 hw/openrisc/pic_cpu.c diff --git a/target/openrisc/cpu.h b/target/openrisc/cpu.h index XXXXXXX..XXXXXXX 100644 --- a/target/openrisc/cpu.h +++ b/target/openrisc/cpu.h @@ -XXX,XX +XXX,XX @@ typedef struct CPUOpenRISCState { uint32_t picmr; /* Interrupt mask register */ uint32_t picsr; /* Interrupt contrl register*/ #endif - void *irq[32]; /* Interrupt irq input */ } CPUOpenRISCState; /** diff --git a/hw/openrisc/openrisc_sim.c b/hw/openrisc/openrisc_sim.c index XXXXXXX..XXXXXXX 100644 --- a/hw/openrisc/openrisc_sim.c +++ b/hw/openrisc/openrisc_sim.c @@ -XXX,XX +XXX,XX @@ static void main_cpu_reset(void *opaque) static qemu_irq get_cpu_irq(OpenRISCCPU *cpus[], int cpunum, int irq_pin) { - return cpus[cpunum]->env.irq[irq_pin]; + return qdev_get_gpio_in_named(DEVICE(cpus[cpunum]), "IRQ", irq_pin); } static void openrisc_sim_net_init(hwaddr base, hwaddr descriptors, @@ -XXX,XX +XXX,XX @@ static void openrisc_sim_init(MachineState *machine) fprintf(stderr, "Unable to find CPU definition!\n"); exit(1); } - cpu_openrisc_pic_init(cpus[n]); cpu_openrisc_clock_init(cpus[n]); diff --git a/hw/openrisc/pic_cpu.c b/hw/openrisc/pic_cpu.c deleted file mode 100644 index XXXXXXX..XXXXXXX --- a/hw/openrisc/pic_cpu.c +++ /dev/null @@ -XXX,XX +XXX,XX @@ -/* - * OpenRISC Programmable Interrupt Controller support. - * - * Copyright (c) 2011-2012 Jia Liu <proljc@gmail.com> - * Feng Gao <gf91597@gmail.com> - * - * 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 "hw/irq.h" -#include "cpu.h" - -/* OpenRISC pic handler */ -static void openrisc_pic_cpu_handler(void *opaque, int irq, int level) -{ - OpenRISCCPU *cpu = (OpenRISCCPU *)opaque; - CPUState *cs = CPU(cpu); - uint32_t irq_bit; - - if (irq > 31 || irq < 0) { - return; - } - - irq_bit = 1U << irq; - - if (level) { - cpu->env.picsr |= irq_bit; - } else { - cpu->env.picsr &= ~irq_bit; - } - - if (cpu->env.picsr & cpu->env.picmr) { - cpu_interrupt(cs, CPU_INTERRUPT_HARD); - } else { - cpu_reset_interrupt(cs, CPU_INTERRUPT_HARD); - cpu->env.picsr = 0; - } -} - -void cpu_openrisc_pic_init(OpenRISCCPU *cpu) -{ - int i; - qemu_irq *qi; - qi = qemu_allocate_irqs(openrisc_pic_cpu_handler, cpu, NR_IRQS); - - for (i = 0; i < NR_IRQS; i++) { - cpu->env.irq[i] = qi[i]; - } -} diff --git a/target/openrisc/cpu.c b/target/openrisc/cpu.c index XXXXXXX..XXXXXXX 100644 --- a/target/openrisc/cpu.c +++ b/target/openrisc/cpu.c @@ -XXX,XX +XXX,XX @@ static void openrisc_cpu_reset(DeviceState *dev) #endif } +#ifndef CONFIG_USER_ONLY +static void openrisc_cpu_set_irq(void *opaque, int irq, int level) +{ + OpenRISCCPU *cpu = (OpenRISCCPU *)opaque; + CPUState *cs = CPU(cpu); + uint32_t irq_bit; + + if (irq > 31 || irq < 0) { + return; + } + + irq_bit = 1U << irq; + + if (level) { + cpu->env.picsr |= irq_bit; + } else { + cpu->env.picsr &= ~irq_bit; + } + + if (cpu->env.picsr & cpu->env.picmr) { + cpu_interrupt(cs, CPU_INTERRUPT_HARD); + } else { + cpu_reset_interrupt(cs, CPU_INTERRUPT_HARD); + cpu->env.picsr = 0; + } +} +#endif + static void openrisc_cpu_realizefn(DeviceState *dev, Error **errp) { CPUState *cs = CPU(dev); @@ -XXX,XX +XXX,XX @@ static void openrisc_cpu_initfn(Object *obj) OpenRISCCPU *cpu = OPENRISC_CPU(obj); cpu_set_cpustate_pointers(cpu); + +#ifndef CONFIG_USER_ONLY + qdev_init_gpio_in_named(DEVICE(cpu), openrisc_cpu_set_irq, "IRQ", NR_IRQS); +#endif } /* CPU models */ diff --git a/hw/openrisc/meson.build b/hw/openrisc/meson.build index XXXXXXX..XXXXXXX 100644 --- a/hw/openrisc/meson.build +++ b/hw/openrisc/meson.build @@ -XXX,XX +XXX,XX @@ openrisc_ss = ss.source_set() -openrisc_ss.add(files('pic_cpu.c', 'cputimer.c')) +openrisc_ss.add(files('cputimer.c')) openrisc_ss.add(when: 'CONFIG_OR1K_SIM', if_true: files('openrisc_sim.c')) hw_arch += {'openrisc': openrisc_ss} -- 2.20.1 The Nios2 architecture supports two different interrupt controller options: * The IIC (Internal Interrupt Controller) is part of the CPU itself; it has 32 IRQ input lines and no NMI support. Interrupt status is queried and controlled via the CPU's ipending and istatus registers. * The EIC (External Interrupt Controller) interface allows the CPU to connect to an external interrupt controller. The interface allows the interrupt controller to present a packet of information containing: - handler address - interrupt level - register set - NMI mode QEMU does not model an EIC currently. We do model the IIC, but its implementation is split across code in hw/nios2/cpu_pic.c and hw/intc/nios2_iic.c. The code in those two files has no state of its own -- the IIC state is in the Nios2CPU state struct. Because CPU objects now inherit (indirectly) from TYPE_DEVICE, they can have GPIO input lines themselves, so we can implement the IIC directly in the CPU object the same way that real hardware does. Create named "IRQ" GPIO inputs to the Nios2 CPU object, and make the only user of the IIC wire up directly to those instead. Note that the old code had an "NMI" concept which was entirely unused and also as far as I can see not architecturally correct, since only the EIC has a concept of an NMI. This fixes a Coverity-reported trivial memory leak of the IRQ array allocated in nios2_cpu_pic_init(). Fixes: Coverity CID 1421916 Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Message-id: 20201129174022.26530-2-peter.maydell@linaro.org Reviewed-by: Wentong Wu <wentong.wu@intel.com> Tested-by: Wentong Wu <wentong.wu@intel.com> --- target/nios2/cpu.h | 1 - hw/intc/nios2_iic.c | 95 --------------------------------------- hw/nios2/10m50_devboard.c | 13 +----- hw/nios2/cpu_pic.c | 31 ------------- target/nios2/cpu.c | 30 +++++++++++++ MAINTAINERS | 1 - hw/intc/meson.build | 1 - 7 files changed, 32 insertions(+), 140 deletions(-) delete mode 100644 hw/intc/nios2_iic.c diff --git a/target/nios2/cpu.h b/target/nios2/cpu.h index XXXXXXX..XXXXXXX 100644 --- a/target/nios2/cpu.h +++ b/target/nios2/cpu.h @@ -XXX,XX +XXX,XX @@ void nios2_cpu_do_unaligned_access(CPUState *cpu, vaddr addr, MMUAccessType access_type, int mmu_idx, uintptr_t retaddr); -qemu_irq *nios2_cpu_pic_init(Nios2CPU *cpu); void nios2_check_interrupts(CPUNios2State *env); void do_nios2_semihosting(CPUNios2State *env); diff --git a/hw/intc/nios2_iic.c b/hw/intc/nios2_iic.c deleted file mode 100644 index XXXXXXX..XXXXXXX --- a/hw/intc/nios2_iic.c +++ /dev/null @@ -XXX,XX +XXX,XX @@ -/* - * QEMU Altera Internal Interrupt Controller. - * - * Copyright (c) 2012 Chris Wulff <crwulff@gmail.com> - * - * 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/lgpl-2.1.html> - */ - -#include "qemu/osdep.h" -#include "qemu/module.h" -#include "qapi/error.h" - -#include "hw/irq.h" -#include "hw/sysbus.h" -#include "cpu.h" -#include "qom/object.h" - -#define TYPE_ALTERA_IIC "altera,iic" -OBJECT_DECLARE_SIMPLE_TYPE(AlteraIIC, ALTERA_IIC) - -struct AlteraIIC { - SysBusDevice parent_obj; - void *cpu; - qemu_irq parent_irq; -}; - -static void update_irq(AlteraIIC *pv) -{ - CPUNios2State *env = &((Nios2CPU *)(pv->cpu))->env; - - qemu_set_irq(pv->parent_irq, - env->regs[CR_IPENDING] & env->regs[CR_IENABLE]); -} - -static void irq_handler(void *opaque, int irq, int level) -{ - AlteraIIC *pv = opaque; - CPUNios2State *env = &((Nios2CPU *)(pv->cpu))->env; - - env->regs[CR_IPENDING] &= ~(1 << irq); - env->regs[CR_IPENDING] |= !!level << irq; - - update_irq(pv); -} - -static void altera_iic_init(Object *obj) -{ - AlteraIIC *pv = ALTERA_IIC(obj); - - qdev_init_gpio_in(DEVICE(pv), irq_handler, 32); - sysbus_init_irq(SYS_BUS_DEVICE(obj), &pv->parent_irq); -} - -static void altera_iic_realize(DeviceState *dev, Error **errp) -{ - struct AlteraIIC *pv = ALTERA_IIC(dev); - - pv->cpu = object_property_get_link(OBJECT(dev), "cpu", &error_abort); -} - -static void altera_iic_class_init(ObjectClass *klass, void *data) -{ - DeviceClass *dc = DEVICE_CLASS(klass); - - /* Reason: needs to be wired up, e.g. by nios2_10m50_ghrd_init() */ - dc->user_creatable = false; - dc->realize = altera_iic_realize; -} - -static TypeInfo altera_iic_info = { - .name = TYPE_ALTERA_IIC, - .parent = TYPE_SYS_BUS_DEVICE, - .instance_size = sizeof(AlteraIIC), - .instance_init = altera_iic_init, - .class_init = altera_iic_class_init, -}; - -static void altera_iic_register(void) -{ - type_register_static(&altera_iic_info); -} - -type_init(altera_iic_register) diff --git a/hw/nios2/10m50_devboard.c b/hw/nios2/10m50_devboard.c index XXXXXXX..XXXXXXX 100644 --- a/hw/nios2/10m50_devboard.c +++ b/hw/nios2/10m50_devboard.c @@ -XXX,XX +XXX,XX @@ static void nios2_10m50_ghrd_init(MachineState *machine) ram_addr_t tcm_size = 0x1000; /* 1 kiB, but QEMU limit is 4 kiB */ ram_addr_t ram_base = 0x08000000; ram_addr_t ram_size = 0x08000000; - qemu_irq *cpu_irq, irq[32]; + qemu_irq irq[32]; int i; /* Physical TCM (tb_ram_1k) with alias at 0xc0000000 */ @@ -XXX,XX +XXX,XX @@ static void nios2_10m50_ghrd_init(MachineState *machine) /* Create CPU -- FIXME */ cpu = NIOS2_CPU(cpu_create(TYPE_NIOS2_CPU)); - - /* Register: CPU interrupt controller (PIC) */ - cpu_irq = nios2_cpu_pic_init(cpu); - - /* Register: Internal Interrupt Controller (IIC) */ - dev = qdev_new("altera,iic"); - object_property_add_const_link(OBJECT(dev), "cpu", OBJECT(cpu)); - sysbus_realize_and_unref(SYS_BUS_DEVICE(dev), &error_fatal); - sysbus_connect_irq(SYS_BUS_DEVICE(dev), 0, cpu_irq[0]); for (i = 0; i < 32; i++) { - irq[i] = qdev_get_gpio_in(dev, i); + irq[i] = qdev_get_gpio_in_named(DEVICE(cpu), "IRQ", i); } /* Register: Altera 16550 UART */ diff --git a/hw/nios2/cpu_pic.c b/hw/nios2/cpu_pic.c index XXXXXXX..XXXXXXX 100644 --- a/hw/nios2/cpu_pic.c +++ b/hw/nios2/cpu_pic.c @@ -XXX,XX +XXX,XX @@ #include "boot.h" -static void nios2_pic_cpu_handler(void *opaque, int irq, int level) -{ - Nios2CPU *cpu = opaque; - CPUNios2State *env = &cpu->env; - CPUState *cs = CPU(cpu); - int type = irq ? CPU_INTERRUPT_NMI : CPU_INTERRUPT_HARD; - - if (type == CPU_INTERRUPT_HARD) { - env->irq_pending = level; - - if (level && (env->regs[CR_STATUS] & CR_STATUS_PIE)) { - env->irq_pending = 0; - cpu_interrupt(cs, type); - } else if (!level) { - env->irq_pending = 0; - cpu_reset_interrupt(cs, type); - } - } else { - if (level) { - cpu_interrupt(cs, type); - } else { - cpu_reset_interrupt(cs, type); - } - } -} - void nios2_check_interrupts(CPUNios2State *env) { if (env->irq_pending && @@ -XXX,XX +XXX,XX @@ void nios2_check_interrupts(CPUNios2State *env) cpu_interrupt(env_cpu(env), CPU_INTERRUPT_HARD); } } - -qemu_irq *nios2_cpu_pic_init(Nios2CPU *cpu) -{ - return qemu_allocate_irqs(nios2_pic_cpu_handler, cpu, 2); -} diff --git a/target/nios2/cpu.c b/target/nios2/cpu.c index XXXXXXX..XXXXXXX 100644 --- a/target/nios2/cpu.c +++ b/target/nios2/cpu.c @@ -XXX,XX +XXX,XX @@ static void nios2_cpu_reset(DeviceState *dev) #endif } +#ifndef CONFIG_USER_ONLY +static void nios2_cpu_set_irq(void *opaque, int irq, int level) +{ + Nios2CPU *cpu = opaque; + CPUNios2State *env = &cpu->env; + CPUState *cs = CPU(cpu); + + env->regs[CR_IPENDING] &= ~(1 << irq); + env->regs[CR_IPENDING] |= !!level << irq; + + env->irq_pending = env->regs[CR_IPENDING] & env->regs[CR_IENABLE]; + + if (env->irq_pending && (env->regs[CR_STATUS] & CR_STATUS_PIE)) { + env->irq_pending = 0; + cpu_interrupt(cs, CPU_INTERRUPT_HARD); + } else if (!env->irq_pending) { + cpu_reset_interrupt(cs, CPU_INTERRUPT_HARD); + } +} +#endif + static void nios2_cpu_initfn(Object *obj) { Nios2CPU *cpu = NIOS2_CPU(obj); @@ -XXX,XX +XXX,XX @@ static void nios2_cpu_initfn(Object *obj) #if !defined(CONFIG_USER_ONLY) mmu_init(&cpu->env); + + /* + * These interrupt lines model the IIC (internal interrupt + * controller). QEMU does not currently support the EIC + * (external interrupt controller) -- if we did it would be + * a separate device in hw/intc with a custom interface to + * the CPU, and boards using it would not wire up these IRQ lines. + */ + qdev_init_gpio_in_named(DEVICE(cpu), nios2_cpu_set_irq, "IRQ", 32); #endif } diff --git a/MAINTAINERS b/MAINTAINERS index XXXXXXX..XXXXXXX 100644 --- a/MAINTAINERS +++ b/MAINTAINERS @@ -XXX,XX +XXX,XX @@ M: Marek Vasut <marex@denx.de> S: Maintained F: target/nios2/ F: hw/nios2/ -F: hw/intc/nios2_iic.c F: disas/nios2.c F: default-configs/nios2-softmmu.mak diff --git a/hw/intc/meson.build b/hw/intc/meson.build index XXXXXXX..XXXXXXX 100644 --- a/hw/intc/meson.build +++ b/hw/intc/meson.build @@ -XXX,XX +XXX,XX @@ specific_ss.add(when: 'CONFIG_IBEX', if_true: files('ibex_plic.c')) specific_ss.add(when: 'CONFIG_IOAPIC', if_true: files('ioapic.c')) specific_ss.add(when: 'CONFIG_LOONGSON_LIOINTC', if_true: files('loongson_liointc.c')) specific_ss.add(when: 'CONFIG_MIPS_CPS', if_true: files('mips_gic.c')) -specific_ss.add(when: 'CONFIG_NIOS2', if_true: files('nios2_iic.c')) specific_ss.add(when: 'CONFIG_OMAP', if_true: files('omap_intc.c')) specific_ss.add(when: 'CONFIG_OMPIC', if_true: files('ompic.c')) specific_ss.add(when: 'CONFIG_OPENPIC_KVM', if_true: files('openpic_kvm.c')) -- 2.20.1 The function nios2_check_interrupts)() looks only at CPU-internal state; it belongs in target/nios2, not hw/nios2. Move it into the same file as its only caller, so it can just be local to that file. This removes the only remaining code from cpu_pic.c, so we can delete that file entirely. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Message-id: 20201129174022.26530-3-peter.maydell@linaro.org Reviewed-by: Wentong Wu <wentong.wu@intel.com> Tested-by: Wentong Wu <wentong.wu@intel.com> --- target/nios2/cpu.h | 2 -- hw/nios2/cpu_pic.c | 36 ------------------------------------ target/nios2/op_helper.c | 9 +++++++++ hw/nios2/meson.build | 2 +- 4 files changed, 10 insertions(+), 39 deletions(-) delete mode 100644 hw/nios2/cpu_pic.c diff --git a/target/nios2/cpu.h b/target/nios2/cpu.h index XXXXXXX..XXXXXXX 100644 --- a/target/nios2/cpu.h +++ b/target/nios2/cpu.h @@ -XXX,XX +XXX,XX @@ void nios2_cpu_do_unaligned_access(CPUState *cpu, vaddr addr, MMUAccessType access_type, int mmu_idx, uintptr_t retaddr); -void nios2_check_interrupts(CPUNios2State *env); - void do_nios2_semihosting(CPUNios2State *env); #define CPU_RESOLVING_TYPE TYPE_NIOS2_CPU diff --git a/hw/nios2/cpu_pic.c b/hw/nios2/cpu_pic.c deleted file mode 100644 index XXXXXXX..XXXXXXX --- a/hw/nios2/cpu_pic.c +++ /dev/null @@ -XXX,XX +XXX,XX @@ -/* - * Altera Nios2 CPU PIC - * - * Copyright (c) 2016 Marek Vasut <marek.vasut@gmail.com> - * - * 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/lgpl-2.1.html> - */ - -#include "qemu/osdep.h" -#include "cpu.h" -#include "hw/irq.h" - -#include "qemu/config-file.h" - -#include "boot.h" - -void nios2_check_interrupts(CPUNios2State *env) -{ - if (env->irq_pending && - (env->regs[CR_STATUS] & CR_STATUS_PIE)) { - env->irq_pending = 0; - cpu_interrupt(env_cpu(env), CPU_INTERRUPT_HARD); - } -} diff --git a/target/nios2/op_helper.c b/target/nios2/op_helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/nios2/op_helper.c +++ b/target/nios2/op_helper.c @@ -XXX,XX +XXX,XX @@ void helper_mmu_write(CPUNios2State *env, uint32_t rn, uint32_t v) mmu_write(env, rn, v); } +static void nios2_check_interrupts(CPUNios2State *env) +{ + if (env->irq_pending && + (env->regs[CR_STATUS] & CR_STATUS_PIE)) { + env->irq_pending = 0; + cpu_interrupt(env_cpu(env), CPU_INTERRUPT_HARD); + } +} + void helper_check_interrupts(CPUNios2State *env) { qemu_mutex_lock_iothread(); diff --git a/hw/nios2/meson.build b/hw/nios2/meson.build index XXXXXXX..XXXXXXX 100644 --- a/hw/nios2/meson.build +++ b/hw/nios2/meson.build @@ -XXX,XX +XXX,XX @@ nios2_ss = ss.source_set() -nios2_ss.add(files('boot.c', 'cpu_pic.c')) +nios2_ss.add(files('boot.c')) nios2_ss.add(when: 'CONFIG_NIOS2_10M50', if_true: files('10m50_devboard.c')) nios2_ss.add(when: 'CONFIG_NIOS2_GENERIC_NOMMU', if_true: files('generic_nommu.c')) -- 2.20.1 In nios2_cpu_set_irq(), use deposit32() rather than raw shift-and-mask operations to set the appropriate bit in the ipending register. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Message-id: 20201129174022.26530-4-peter.maydell@linaro.org --- target/nios2/cpu.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) diff --git a/target/nios2/cpu.c b/target/nios2/cpu.c index XXXXXXX..XXXXXXX 100644 --- a/target/nios2/cpu.c +++ b/target/nios2/cpu.c @@ -XXX,XX +XXX,XX @@ static void nios2_cpu_set_irq(void *opaque, int irq, int level) CPUNios2State *env = &cpu->env; CPUState *cs = CPU(cpu); - env->regs[CR_IPENDING] &= ~(1 << irq); - env->regs[CR_IPENDING] |= !!level << irq; + env->regs[CR_IPENDING] = deposit32(env->regs[CR_IPENDING], irq, 1, !!level); env->irq_pending = env->regs[CR_IPENDING] & env->regs[CR_IENABLE]; -- 2.20.1 In rom_check_and_register_reset() we detect overlaps by looking at whether the ROM blob we're currently examining is in the same address space and starts before the previous ROM blob ends. (This works because the ROM list is kept sorted in order by AddressSpace and then by address.) Instead of keeping the AddressSpace and last address of the previous ROM blob in local variables, just keep a pointer to it. This will allow us to print more useful information when we do detect an overlap. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20201129203923.10622-2-peter.maydell@linaro.org --- hw/core/loader.c | 23 +++++++++++++++-------- 1 file changed, 15 insertions(+), 8 deletions(-) diff --git a/hw/core/loader.c b/hw/core/loader.c index XXXXXXX..XXXXXXX 100644 --- a/hw/core/loader.c +++ b/hw/core/loader.c @@ -XXX,XX +XXX,XX @@ static void rom_reset(void *unused) } } +/* Return true if two consecutive ROMs in the ROM list overlap */ +static bool roms_overlap(Rom *last_rom, Rom *this_rom) +{ + if (!last_rom) { + return false; + } + return last_rom->as == this_rom->as && + last_rom->addr + last_rom->romsize > this_rom->addr; +} + int rom_check_and_register_reset(void) { - hwaddr addr = 0; MemoryRegionSection section; - Rom *rom; - AddressSpace *as = NULL; + Rom *rom, *last_rom = NULL; QTAILQ_FOREACH(rom, &roms, next) { if (rom->fw_file) { continue; } if (!rom->mr) { - if ((addr > rom->addr) && (as == rom->as)) { + if (roms_overlap(last_rom, rom)) { fprintf(stderr, "rom: requested regions overlap " "(rom %s. free=0x" TARGET_FMT_plx ", addr=0x" TARGET_FMT_plx ")\n", - rom->name, addr, rom->addr); + rom->name, last_rom->addr + last_rom->romsize, + rom->addr); return -1; } - addr = rom->addr; - addr += rom->romsize; - as = rom->as; + last_rom = rom; } section = memory_region_find(rom->mr ? rom->mr : get_system_memory(), rom->addr, 1); -- 2.20.1 In rom_check_and_register_reset() we report to the user if there is a "ROM region overlap". This has a couple of problems: * the reported information is not very easy to intepret * the function just prints the overlap to stderr (and relies on its single callsite in vl.c to do an error_report() and exit) * only the first overlap encountered is diagnosed Make this function use error_report() and error_printf() and report a more user-friendly report with all the overlaps diagnosed. Sample old output: rom: requested regions overlap (rom dtb. free=0x0000000000008000, addr=0x0000000000000000) qemu-system-aarch64: rom check and register reset failed Sample new output: qemu-system-aarch64: Some ROM regions are overlapping These ROM regions might have been loaded by direct user request or by default. They could be BIOS/firmware images, a guest kernel, initrd or some other file loaded into guest memory. Check whether you intended to load all this guest code, and whether it has been built to load to the correct addresses. The following two regions overlap (in the cpu-memory-0 address space): phdr #0: /home/petmay01/linaro/qemu-misc-tests/ldmia-fault.axf (addresses 0x0000000000000000 - 0x0000000000008000) dtb (addresses 0x0000000000000000 - 0x0000000000100000) The following two regions overlap (in the cpu-memory-0 address space): phdr #1: /home/petmay01/linaro/qemu-misc-tests/bad-psci-call.axf (addresses 0x0000000040000000 - 0x0000000040000010) phdr #0: /home/petmay01/linaro/qemu-misc-tests/bp-test.elf (addresses 0x0000000040000000 - 0x0000000040000020) Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20201129203923.10622-3-peter.maydell@linaro.org --- hw/core/loader.c | 48 ++++++++++++++++++++++++++++++++++++++++++------ softmmu/vl.c | 1 - 2 files changed, 42 insertions(+), 7 deletions(-) diff --git a/hw/core/loader.c b/hw/core/loader.c index XXXXXXX..XXXXXXX 100644 --- a/hw/core/loader.c +++ b/hw/core/loader.c @@ -XXX,XX +XXX,XX @@ static bool roms_overlap(Rom *last_rom, Rom *this_rom) last_rom->addr + last_rom->romsize > this_rom->addr; } +static const char *rom_as_name(Rom *rom) +{ + const char *name = rom->as ? rom->as->name : NULL; + return name ?: "anonymous"; +} + +static void rom_print_overlap_error_header(void) +{ + error_report("Some ROM regions are overlapping"); + error_printf( + "These ROM regions might have been loaded by " + "direct user request or by default.\n" + "They could be BIOS/firmware images, a guest kernel, " + "initrd or some other file loaded into guest memory.\n" + "Check whether you intended to load all this guest code, and " + "whether it has been built to load to the correct addresses.\n"); +} + +static void rom_print_one_overlap_error(Rom *last_rom, Rom *rom) +{ + error_printf( + "\nThe following two regions overlap (in the %s address space):\n", + rom_as_name(rom)); + error_printf( + " %s (addresses 0x" TARGET_FMT_plx " - 0x" TARGET_FMT_plx ")\n", + last_rom->name, last_rom->addr, last_rom->addr + last_rom->romsize); + error_printf( + " %s (addresses 0x" TARGET_FMT_plx " - 0x" TARGET_FMT_plx ")\n", + rom->name, rom->addr, rom->addr + rom->romsize); +} + int rom_check_and_register_reset(void) { MemoryRegionSection section; Rom *rom, *last_rom = NULL; + bool found_overlap = false; QTAILQ_FOREACH(rom, &roms, next) { if (rom->fw_file) { @@ -XXX,XX +XXX,XX @@ int rom_check_and_register_reset(void) } if (!rom->mr) { if (roms_overlap(last_rom, rom)) { - fprintf(stderr, "rom: requested regions overlap " - "(rom %s. free=0x" TARGET_FMT_plx - ", addr=0x" TARGET_FMT_plx ")\n", - rom->name, last_rom->addr + last_rom->romsize, - rom->addr); - return -1; + if (!found_overlap) { + found_overlap = true; + rom_print_overlap_error_header(); + } + rom_print_one_overlap_error(last_rom, rom); + /* Keep going through the list so we report all overlaps */ } last_rom = rom; } @@ -XXX,XX +XXX,XX @@ int rom_check_and_register_reset(void) rom->isrom = int128_nz(section.size) && memory_region_is_rom(section.mr); memory_region_unref(section.mr); } + if (found_overlap) { + return -1; + } + qemu_register_reset(rom_reset, NULL); roms_loaded = 1; return 0; diff --git a/softmmu/vl.c b/softmmu/vl.c index XXXXXXX..XXXXXXX 100644 --- a/softmmu/vl.c +++ b/softmmu/vl.c @@ -XXX,XX +XXX,XX @@ static void qemu_machine_creation_done(void) qemu_run_machine_init_done_notifiers(); if (rom_check_and_register_reset() != 0) { - error_report("rom check and register reset failed"); exit(1); } -- 2.20.1 Currently the load_elf code assembles the ROM blob name into a local 128 byte fixed-size array. Use g_strdup_printf() instead so that we don't truncate the pathname if it happens to be long. (This matters mostly for monitor 'info roms' output and for the error messages if ROM blobs overlap.) Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20201129203923.10622-4-peter.maydell@linaro.org --- include/hw/elf_ops.h | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/include/hw/elf_ops.h b/include/hw/elf_ops.h index XXXXXXX..XXXXXXX 100644 --- a/include/hw/elf_ops.h +++ b/include/hw/elf_ops.h @@ -XXX,XX +XXX,XX @@ static int glue(load_elf, SZ)(const char *name, int fd, uint64_t addr, low = (uint64_t)-1, high = 0; GMappedFile *mapped_file = NULL; uint8_t *data = NULL; - char label[128]; int ret = ELF_LOAD_FAILED; if (read(fd, &ehdr, sizeof(ehdr)) != sizeof(ehdr)) @@ -XXX,XX +XXX,XX @@ static int glue(load_elf, SZ)(const char *name, int fd, */ if (mem_size != 0) { if (load_rom) { - snprintf(label, sizeof(label), "phdr #%d: %s", i, name); + g_autofree char *label = + g_strdup_printf("phdr #%d: %s", i, name); /* * rom_add_elf_program() takes its own reference to -- 2.20.1 Instead of making the ROM blob name something like: phdr #0: /home/petmay01/linaro/qemu-misc-tests/ldmia-fault.axf make it a little more self-explanatory for people who don't know ELF format details: /home/petmay01/linaro/qemu-misc-tests/ldmia-fault.axf ELF program header segment 0 Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20201129203923.10622-5-peter.maydell@linaro.org --- include/hw/elf_ops.h | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/include/hw/elf_ops.h b/include/hw/elf_ops.h index XXXXXXX..XXXXXXX 100644 --- a/include/hw/elf_ops.h +++ b/include/hw/elf_ops.h @@ -XXX,XX +XXX,XX @@ static int glue(load_elf, SZ)(const char *name, int fd, if (mem_size != 0) { if (load_rom) { g_autofree char *label = - g_strdup_printf("phdr #%d: %s", i, name); + g_strdup_printf("%s ELF program header segment %d", + name, i); /* * rom_add_elf_program() takes its own reference to -- 2.20.1 From: Sai Pavan Boddu <sai.pavan.boddu@xilinx.com> This module emulates control registers of versal usb2 controller, this is added just to make guest happy. In general this module would control the phy-reset signal from usb controller, data coherency of the transactions, signals the host system errors received from controller. Signed-off-by: Sai Pavan Boddu <sai.pavan.boddu@xilinx.com> Signed-off-by: Vikram Garhwal <fnu.vikram@xilinx.com> Reviewed-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com> Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Message-id: 1607023357-5096-2-git-send-email-sai.pavan.boddu@xilinx.com Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- include/hw/usb/xlnx-versal-usb2-ctrl-regs.h | 45 ++++ hw/usb/xlnx-versal-usb2-ctrl-regs.c | 229 ++++++++++++++++++++ hw/usb/meson.build | 1 + 3 files changed, 275 insertions(+) create mode 100644 include/hw/usb/xlnx-versal-usb2-ctrl-regs.h create mode 100644 hw/usb/xlnx-versal-usb2-ctrl-regs.c diff --git a/include/hw/usb/xlnx-versal-usb2-ctrl-regs.h b/include/hw/usb/xlnx-versal-usb2-ctrl-regs.h new file mode 100644 index XXXXXXX..XXXXXXX --- /dev/null +++ b/include/hw/usb/xlnx-versal-usb2-ctrl-regs.h @@ -XXX,XX +XXX,XX @@ +/* + * QEMU model of the VersalUsb2CtrlRegs Register control/Status block for + * USB2.0 controller + * + * Copyright (c) 2020 Xilinx Inc. 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. + */ + +#ifndef _XLNX_USB2_REGS_H_ +#define _XLNX_USB2_REGS_H_ + +#define TYPE_XILINX_VERSAL_USB2_CTRL_REGS "xlnx.versal-usb2-ctrl-regs" + +#define XILINX_VERSAL_USB2_CTRL_REGS(obj) \ + OBJECT_CHECK(VersalUsb2CtrlRegs, (obj), TYPE_XILINX_VERSAL_USB2_CTRL_REGS) + +#define USB2_REGS_R_MAX ((0x78 / 4) + 1) + +typedef struct VersalUsb2CtrlRegs { + SysBusDevice parent_obj; + MemoryRegion iomem; + qemu_irq irq_ir; + + uint32_t regs[USB2_REGS_R_MAX]; + RegisterInfo regs_info[USB2_REGS_R_MAX]; +} VersalUsb2CtrlRegs; + +#endif diff --git a/hw/usb/xlnx-versal-usb2-ctrl-regs.c b/hw/usb/xlnx-versal-usb2-ctrl-regs.c new file mode 100644 index XXXXXXX..XXXXXXX --- /dev/null +++ b/hw/usb/xlnx-versal-usb2-ctrl-regs.c @@ -XXX,XX +XXX,XX @@ +/* + * QEMU model of the VersalUsb2CtrlRegs Register control/Status block for + * USB2.0 controller + * + * This module should control phy_reset, permanent device plugs, frame length + * time adjust & setting of coherency paths. None of which are emulated in + * present model. + * + * Copyright (c) 2020 Xilinx Inc. 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 "hw/sysbus.h" +#include "hw/irq.h" +#include "hw/register.h" +#include "qemu/bitops.h" +#include "qemu/log.h" +#include "qom/object.h" +#include "migration/vmstate.h" +#include "hw/usb/xlnx-versal-usb2-ctrl-regs.h" + +#ifndef XILINX_VERSAL_USB2_CTRL_REGS_ERR_DEBUG +#define XILINX_VERSAL_USB2_CTRL_REGS_ERR_DEBUG 0 +#endif + +REG32(BUS_FILTER, 0x30) + FIELD(BUS_FILTER, BYPASS, 0, 4) +REG32(PORT, 0x34) + FIELD(PORT, HOST_SMI_BAR_WR, 4, 1) + FIELD(PORT, HOST_SMI_PCI_CMD_REG_WR, 3, 1) + FIELD(PORT, HOST_MSI_ENABLE, 2, 1) + FIELD(PORT, PWR_CTRL_PRSNT, 1, 1) + FIELD(PORT, HUB_PERM_ATTACH, 0, 1) +REG32(JITTER_ADJUST, 0x38) + FIELD(JITTER_ADJUST, FLADJ, 0, 6) +REG32(BIGENDIAN, 0x40) + FIELD(BIGENDIAN, ENDIAN_GS, 0, 1) +REG32(COHERENCY, 0x44) + FIELD(COHERENCY, USB_COHERENCY, 0, 1) +REG32(XHC_BME, 0x48) + FIELD(XHC_BME, XHC_BME, 0, 1) +REG32(REG_CTRL, 0x60) + FIELD(REG_CTRL, SLVERR_ENABLE, 0, 1) +REG32(IR_STATUS, 0x64) + FIELD(IR_STATUS, HOST_SYS_ERR, 1, 1) + FIELD(IR_STATUS, ADDR_DEC_ERR, 0, 1) +REG32(IR_MASK, 0x68) + FIELD(IR_MASK, HOST_SYS_ERR, 1, 1) + FIELD(IR_MASK, ADDR_DEC_ERR, 0, 1) +REG32(IR_ENABLE, 0x6c) + FIELD(IR_ENABLE, HOST_SYS_ERR, 1, 1) + FIELD(IR_ENABLE, ADDR_DEC_ERR, 0, 1) +REG32(IR_DISABLE, 0x70) + FIELD(IR_DISABLE, HOST_SYS_ERR, 1, 1) + FIELD(IR_DISABLE, ADDR_DEC_ERR, 0, 1) +REG32(USB3, 0x78) + +static void ir_update_irq(VersalUsb2CtrlRegs *s) +{ + bool pending = s->regs[R_IR_STATUS] & ~s->regs[R_IR_MASK]; + qemu_set_irq(s->irq_ir, pending); +} + +static void ir_status_postw(RegisterInfo *reg, uint64_t val64) +{ + VersalUsb2CtrlRegs *s = XILINX_VERSAL_USB2_CTRL_REGS(reg->opaque); + /* + * TODO: This should also clear USBSTS.HSE field in USB XHCI register. + * May be combine both the modules. + */ + ir_update_irq(s); +} + +static uint64_t ir_enable_prew(RegisterInfo *reg, uint64_t val64) +{ + VersalUsb2CtrlRegs *s = XILINX_VERSAL_USB2_CTRL_REGS(reg->opaque); + uint32_t val = val64; + + s->regs[R_IR_MASK] &= ~val; + ir_update_irq(s); + return 0; +} + +static uint64_t ir_disable_prew(RegisterInfo *reg, uint64_t val64) +{ + VersalUsb2CtrlRegs *s = XILINX_VERSAL_USB2_CTRL_REGS(reg->opaque); + uint32_t val = val64; + + s->regs[R_IR_MASK] |= val; + ir_update_irq(s); + return 0; +} + +static const RegisterAccessInfo usb2_ctrl_regs_regs_info[] = { + { .name = "BUS_FILTER", .addr = A_BUS_FILTER, + .rsvd = 0xfffffff0, + },{ .name = "PORT", .addr = A_PORT, + .rsvd = 0xffffffe0, + },{ .name = "JITTER_ADJUST", .addr = A_JITTER_ADJUST, + .reset = 0x20, + .rsvd = 0xffffffc0, + },{ .name = "BIGENDIAN", .addr = A_BIGENDIAN, + .rsvd = 0xfffffffe, + },{ .name = "COHERENCY", .addr = A_COHERENCY, + .rsvd = 0xfffffffe, + },{ .name = "XHC_BME", .addr = A_XHC_BME, + .reset = 0x1, + .rsvd = 0xfffffffe, + },{ .name = "REG_CTRL", .addr = A_REG_CTRL, + .rsvd = 0xfffffffe, + },{ .name = "IR_STATUS", .addr = A_IR_STATUS, + .rsvd = 0xfffffffc, + .w1c = 0x3, + .post_write = ir_status_postw, + },{ .name = "IR_MASK", .addr = A_IR_MASK, + .reset = 0x3, + .rsvd = 0xfffffffc, + .ro = 0x3, + },{ .name = "IR_ENABLE", .addr = A_IR_ENABLE, + .rsvd = 0xfffffffc, + .pre_write = ir_enable_prew, + },{ .name = "IR_DISABLE", .addr = A_IR_DISABLE, + .rsvd = 0xfffffffc, + .pre_write = ir_disable_prew, + },{ .name = "USB3", .addr = A_USB3, + } +}; + +static void usb2_ctrl_regs_reset_init(Object *obj, ResetType type) +{ + VersalUsb2CtrlRegs *s = XILINX_VERSAL_USB2_CTRL_REGS(obj); + unsigned int i; + + for (i = 0; i < ARRAY_SIZE(s->regs_info); ++i) { + register_reset(&s->regs_info[i]); + } +} + +static void usb2_ctrl_regs_reset_hold(Object *obj) +{ + VersalUsb2CtrlRegs *s = XILINX_VERSAL_USB2_CTRL_REGS(obj); + + ir_update_irq(s); +} + +static const MemoryRegionOps usb2_ctrl_regs_ops = { + .read = register_read_memory, + .write = register_write_memory, + .endianness = DEVICE_LITTLE_ENDIAN, + .valid = { + .min_access_size = 4, + .max_access_size = 4, + }, +}; + +static void usb2_ctrl_regs_init(Object *obj) +{ + VersalUsb2CtrlRegs *s = XILINX_VERSAL_USB2_CTRL_REGS(obj); + SysBusDevice *sbd = SYS_BUS_DEVICE(obj); + RegisterInfoArray *reg_array; + + memory_region_init(&s->iomem, obj, TYPE_XILINX_VERSAL_USB2_CTRL_REGS, + USB2_REGS_R_MAX * 4); + reg_array = + register_init_block32(DEVICE(obj), usb2_ctrl_regs_regs_info, + ARRAY_SIZE(usb2_ctrl_regs_regs_info), + s->regs_info, s->regs, + &usb2_ctrl_regs_ops, + XILINX_VERSAL_USB2_CTRL_REGS_ERR_DEBUG, + USB2_REGS_R_MAX * 4); + memory_region_add_subregion(&s->iomem, + 0x0, + &reg_array->mem); + sysbus_init_mmio(sbd, &s->iomem); + sysbus_init_irq(sbd, &s->irq_ir); +} + +static const VMStateDescription vmstate_usb2_ctrl_regs = { + .name = TYPE_XILINX_VERSAL_USB2_CTRL_REGS, + .version_id = 1, + .minimum_version_id = 1, + .fields = (VMStateField[]) { + VMSTATE_UINT32_ARRAY(regs, VersalUsb2CtrlRegs, USB2_REGS_R_MAX), + VMSTATE_END_OF_LIST(), + } +}; + +static void usb2_ctrl_regs_class_init(ObjectClass *klass, void *data) +{ + DeviceClass *dc = DEVICE_CLASS(klass); + ResettableClass *rc = RESETTABLE_CLASS(klass); + + rc->phases.enter = usb2_ctrl_regs_reset_init; + rc->phases.hold = usb2_ctrl_regs_reset_hold; + dc->vmsd = &vmstate_usb2_ctrl_regs; +} + +static const TypeInfo usb2_ctrl_regs_info = { + .name = TYPE_XILINX_VERSAL_USB2_CTRL_REGS, + .parent = TYPE_SYS_BUS_DEVICE, + .instance_size = sizeof(VersalUsb2CtrlRegs), + .class_init = usb2_ctrl_regs_class_init, + .instance_init = usb2_ctrl_regs_init, +}; + +static void usb2_ctrl_regs_register_types(void) +{ + type_register_static(&usb2_ctrl_regs_info); +} + +type_init(usb2_ctrl_regs_register_types) diff --git a/hw/usb/meson.build b/hw/usb/meson.build index XXXXXXX..XXXXXXX 100644 --- a/hw/usb/meson.build +++ b/hw/usb/meson.build @@ -XXX,XX +XXX,XX @@ softmmu_ss.add(when: 'CONFIG_USB_DWC2', if_true: files('hcd-dwc2.c')) softmmu_ss.add(when: 'CONFIG_TUSB6010', if_true: files('tusb6010.c')) softmmu_ss.add(when: 'CONFIG_IMX', if_true: files('chipidea.c')) softmmu_ss.add(when: 'CONFIG_IMX_USBPHY', if_true: files('imx-usb-phy.c')) +specific_ss.add(when: 'CONFIG_XLNX_VERSAL', if_true: files('xlnx-versal-usb2-ctrl-regs.c')) # emulated usb devices softmmu_ss.add(when: 'CONFIG_USB', if_true: files('dev-hub.c')) -- 2.20.1 From: Vikram Garhwal <fnu.vikram@xilinx.com> This patch adds skeleton model of dwc3 usb controller attached to xhci-sysbus device. It defines global register space of DWC3 controller, global registers control the AXI/AHB interfaces properties, external FIFO support and event count support. All of which are unimplemented at present,we are only supporting core reset and read of ID register. Signed-off-by: Vikram Garhwal <fnu.vikram@xilinx.com> Signed-off-by: Sai Pavan Boddu <sai.pavan.boddu@xilinx.com> Reviewed-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com> Message-id: 1607023357-5096-3-git-send-email-sai.pavan.boddu@xilinx.com Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- include/hw/usb/hcd-dwc3.h | 55 +++ hw/usb/hcd-dwc3.c | 689 ++++++++++++++++++++++++++++++++++++++ hw/usb/Kconfig | 5 + hw/usb/meson.build | 1 + 4 files changed, 750 insertions(+) create mode 100644 include/hw/usb/hcd-dwc3.h create mode 100644 hw/usb/hcd-dwc3.c diff --git a/include/hw/usb/hcd-dwc3.h b/include/hw/usb/hcd-dwc3.h new file mode 100644 index XXXXXXX..XXXXXXX --- /dev/null +++ b/include/hw/usb/hcd-dwc3.h @@ -XXX,XX +XXX,XX @@ +/* + * QEMU model of the USB DWC3 host controller emulation. + * + * 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. + */ +#ifndef HCD_DWC3_H +#define HCD_DWC3_H + +#include "hw/usb/hcd-xhci.h" +#include "hw/usb/hcd-xhci-sysbus.h" + +#define TYPE_USB_DWC3 "usb_dwc3" + +#define USB_DWC3(obj) \ + OBJECT_CHECK(USBDWC3, (obj), TYPE_USB_DWC3) + +#define USB_DWC3_R_MAX ((0x530 / 4) + 1) +#define DWC3_SIZE 0x10000 + +typedef struct USBDWC3 { + SysBusDevice parent_obj; + MemoryRegion iomem; + XHCISysbusState sysbus_xhci; + + uint32_t regs[USB_DWC3_R_MAX]; + RegisterInfo regs_info[USB_DWC3_R_MAX]; + + struct { + uint8_t mode; + uint32_t dwc_usb3_user; + } cfg; + +} USBDWC3; + +#endif diff --git a/hw/usb/hcd-dwc3.c b/hw/usb/hcd-dwc3.c new file mode 100644 index XXXXXXX..XXXXXXX --- /dev/null +++ b/hw/usb/hcd-dwc3.c @@ -XXX,XX +XXX,XX @@ +/* + * QEMU model of the USB DWC3 host controller emulation. + * + * This model defines global register space of DWC3 controller. Global + * registers control the AXI/AHB interfaces properties, external FIFO support + * and event count support. All of which are unimplemented at present. We are + * only supporting core reset and read of ID register. + * + * Copyright (c) 2020 Xilinx Inc. 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 "hw/sysbus.h" +#include "hw/register.h" +#include "qemu/bitops.h" +#include "qemu/log.h" +#include "qom/object.h" +#include "migration/vmstate.h" +#include "hw/qdev-properties.h" +#include "hw/usb/hcd-dwc3.h" +#include "qapi/error.h" + +#ifndef USB_DWC3_ERR_DEBUG +#define USB_DWC3_ERR_DEBUG 0 +#endif + +#define HOST_MODE 1 +#define FIFO_LEN 0x1000 + +REG32(GSBUSCFG0, 0x00) + FIELD(GSBUSCFG0, DATRDREQINFO, 28, 4) + FIELD(GSBUSCFG0, DESRDREQINFO, 24, 4) + FIELD(GSBUSCFG0, DATWRREQINFO, 20, 4) + FIELD(GSBUSCFG0, DESWRREQINFO, 16, 4) + FIELD(GSBUSCFG0, RESERVED_15_12, 12, 4) + FIELD(GSBUSCFG0, DATBIGEND, 11, 1) + FIELD(GSBUSCFG0, DESBIGEND, 10, 1) + FIELD(GSBUSCFG0, RESERVED_9_8, 8, 2) + FIELD(GSBUSCFG0, INCR256BRSTENA, 7, 1) + FIELD(GSBUSCFG0, INCR128BRSTENA, 6, 1) + FIELD(GSBUSCFG0, INCR64BRSTENA, 5, 1) + FIELD(GSBUSCFG0, INCR32BRSTENA, 4, 1) + FIELD(GSBUSCFG0, INCR16BRSTENA, 3, 1) + FIELD(GSBUSCFG0, INCR8BRSTENA, 2, 1) + FIELD(GSBUSCFG0, INCR4BRSTENA, 1, 1) + FIELD(GSBUSCFG0, INCRBRSTENA, 0, 1) +REG32(GSBUSCFG1, 0x04) + FIELD(GSBUSCFG1, RESERVED_31_13, 13, 19) + FIELD(GSBUSCFG1, EN1KPAGE, 12, 1) + FIELD(GSBUSCFG1, PIPETRANSLIMIT, 8, 4) + FIELD(GSBUSCFG1, RESERVED_7_0, 0, 8) +REG32(GTXTHRCFG, 0x08) + FIELD(GTXTHRCFG, RESERVED_31, 31, 1) + FIELD(GTXTHRCFG, RESERVED_30, 30, 1) + FIELD(GTXTHRCFG, USBTXPKTCNTSEL, 29, 1) + FIELD(GTXTHRCFG, RESERVED_28, 28, 1) + FIELD(GTXTHRCFG, USBTXPKTCNT, 24, 4) + FIELD(GTXTHRCFG, USBMAXTXBURSTSIZE, 16, 8) + FIELD(GTXTHRCFG, RESERVED_15, 15, 1) + FIELD(GTXTHRCFG, RESERVED_14, 14, 1) + FIELD(GTXTHRCFG, RESERVED_13_11, 11, 3) + FIELD(GTXTHRCFG, RESERVED_10_0, 0, 11) +REG32(GRXTHRCFG, 0x0c) + FIELD(GRXTHRCFG, RESERVED_31_30, 30, 2) + FIELD(GRXTHRCFG, USBRXPKTCNTSEL, 29, 1) + FIELD(GRXTHRCFG, RESERVED_28, 28, 1) + FIELD(GRXTHRCFG, USBRXPKTCNT, 24, 4) + FIELD(GRXTHRCFG, USBMAXRXBURSTSIZE, 19, 5) + FIELD(GRXTHRCFG, RESERVED_18_16, 16, 3) + FIELD(GRXTHRCFG, RESERVED_15, 15, 1) + FIELD(GRXTHRCFG, RESERVED_14_13, 13, 2) + FIELD(GRXTHRCFG, RESVISOCOUTSPC, 0, 13) +REG32(GCTL, 0x10) + FIELD(GCTL, PWRDNSCALE, 19, 13) + FIELD(GCTL, MASTERFILTBYPASS, 18, 1) + FIELD(GCTL, BYPSSETADDR, 17, 1) + FIELD(GCTL, U2RSTECN, 16, 1) + FIELD(GCTL, FRMSCLDWN, 14, 2) + FIELD(GCTL, PRTCAPDIR, 12, 2) + FIELD(GCTL, CORESOFTRESET, 11, 1) + FIELD(GCTL, U1U2TIMERSCALE, 9, 1) + FIELD(GCTL, DEBUGATTACH, 8, 1) + FIELD(GCTL, RAMCLKSEL, 6, 2) + FIELD(GCTL, SCALEDOWN, 4, 2) + FIELD(GCTL, DISSCRAMBLE, 3, 1) + FIELD(GCTL, U2EXIT_LFPS, 2, 1) + FIELD(GCTL, GBLHIBERNATIONEN, 1, 1) + FIELD(GCTL, DSBLCLKGTNG, 0, 1) +REG32(GPMSTS, 0x14) +REG32(GSTS, 0x18) + FIELD(GSTS, CBELT, 20, 12) + FIELD(GSTS, RESERVED_19_12, 12, 8) + FIELD(GSTS, SSIC_IP, 11, 1) + FIELD(GSTS, OTG_IP, 10, 1) + FIELD(GSTS, BC_IP, 9, 1) + FIELD(GSTS, ADP_IP, 8, 1) + FIELD(GSTS, HOST_IP, 7, 1) + FIELD(GSTS, DEVICE_IP, 6, 1) + FIELD(GSTS, CSRTIMEOUT, 5, 1) + FIELD(GSTS, BUSERRADDRVLD, 4, 1) + FIELD(GSTS, RESERVED_3_2, 2, 2) + FIELD(GSTS, CURMOD, 0, 2) +REG32(GUCTL1, 0x1c) + FIELD(GUCTL1, RESUME_OPMODE_HS_HOST, 10, 1) +REG32(GSNPSID, 0x20) +REG32(GGPIO, 0x24) + FIELD(GGPIO, GPO, 16, 16) + FIELD(GGPIO, GPI, 0, 16) +REG32(GUID, 0x28) +REG32(GUCTL, 0x2c) + FIELD(GUCTL, REFCLKPER, 22, 10) + FIELD(GUCTL, NOEXTRDL, 21, 1) + FIELD(GUCTL, RESERVED_20_18, 18, 3) + FIELD(GUCTL, SPRSCTRLTRANSEN, 17, 1) + FIELD(GUCTL, RESBWHSEPS, 16, 1) + FIELD(GUCTL, RESERVED_15, 15, 1) + FIELD(GUCTL, USBHSTINAUTORETRYEN, 14, 1) + FIELD(GUCTL, ENOVERLAPCHK, 13, 1) + FIELD(GUCTL, EXTCAPSUPPTEN, 12, 1) + FIELD(GUCTL, INSRTEXTRFSBODI, 11, 1) + FIELD(GUCTL, DTCT, 9, 2) + FIELD(GUCTL, DTFT, 0, 9) +REG32(GBUSERRADDRLO, 0x30) +REG32(GBUSERRADDRHI, 0x34) +REG32(GHWPARAMS0, 0x40) + FIELD(GHWPARAMS0, GHWPARAMS0_31_24, 24, 8) + FIELD(GHWPARAMS0, GHWPARAMS0_23_16, 16, 8) + FIELD(GHWPARAMS0, GHWPARAMS0_15_8, 8, 8) + FIELD(GHWPARAMS0, GHWPARAMS0_7_6, 6, 2) + FIELD(GHWPARAMS0, GHWPARAMS0_5_3, 3, 3) + FIELD(GHWPARAMS0, GHWPARAMS0_2_0, 0, 3) +REG32(GHWPARAMS1, 0x44) + FIELD(GHWPARAMS1, GHWPARAMS1_31, 31, 1) + FIELD(GHWPARAMS1, GHWPARAMS1_30, 30, 1) + FIELD(GHWPARAMS1, GHWPARAMS1_29, 29, 1) + FIELD(GHWPARAMS1, GHWPARAMS1_28, 28, 1) + FIELD(GHWPARAMS1, GHWPARAMS1_27, 27, 1) + FIELD(GHWPARAMS1, GHWPARAMS1_26, 26, 1) + FIELD(GHWPARAMS1, GHWPARAMS1_25_24, 24, 2) + FIELD(GHWPARAMS1, GHWPARAMS1_23, 23, 1) + FIELD(GHWPARAMS1, GHWPARAMS1_22_21, 21, 2) + FIELD(GHWPARAMS1, GHWPARAMS1_20_15, 15, 6) + FIELD(GHWPARAMS1, GHWPARAMS1_14_12, 12, 3) + FIELD(GHWPARAMS1, GHWPARAMS1_11_9, 9, 3) + FIELD(GHWPARAMS1, GHWPARAMS1_8_6, 6, 3) + FIELD(GHWPARAMS1, GHWPARAMS1_5_3, 3, 3) + FIELD(GHWPARAMS1, GHWPARAMS1_2_0, 0, 3) +REG32(GHWPARAMS2, 0x48) +REG32(GHWPARAMS3, 0x4c) + FIELD(GHWPARAMS3, GHWPARAMS3_31, 31, 1) + FIELD(GHWPARAMS3, GHWPARAMS3_30_23, 23, 8) + FIELD(GHWPARAMS3, GHWPARAMS3_22_18, 18, 5) + FIELD(GHWPARAMS3, GHWPARAMS3_17_12, 12, 6) + FIELD(GHWPARAMS3, GHWPARAMS3_11, 11, 1) + FIELD(GHWPARAMS3, GHWPARAMS3_10, 10, 1) + FIELD(GHWPARAMS3, GHWPARAMS3_9_8, 8, 2) + FIELD(GHWPARAMS3, GHWPARAMS3_7_6, 6, 2) + FIELD(GHWPARAMS3, GHWPARAMS3_5_4, 4, 2) + FIELD(GHWPARAMS3, GHWPARAMS3_3_2, 2, 2) + FIELD(GHWPARAMS3, GHWPARAMS3_1_0, 0, 2) +REG32(GHWPARAMS4, 0x50) + FIELD(GHWPARAMS4, GHWPARAMS4_31_28, 28, 4) + FIELD(GHWPARAMS4, GHWPARAMS4_27_24, 24, 4) + FIELD(GHWPARAMS4, GHWPARAMS4_23, 23, 1) + FIELD(GHWPARAMS4, GHWPARAMS4_22, 22, 1) + FIELD(GHWPARAMS4, GHWPARAMS4_21, 21, 1) + FIELD(GHWPARAMS4, GHWPARAMS4_20_17, 17, 4) + FIELD(GHWPARAMS4, GHWPARAMS4_16_13, 13, 4) + FIELD(GHWPARAMS4, GHWPARAMS4_12, 12, 1) + FIELD(GHWPARAMS4, GHWPARAMS4_11, 11, 1) + FIELD(GHWPARAMS4, GHWPARAMS4_10_9, 9, 2) + FIELD(GHWPARAMS4, GHWPARAMS4_8_7, 7, 2) + FIELD(GHWPARAMS4, GHWPARAMS4_6, 6, 1) + FIELD(GHWPARAMS4, GHWPARAMS4_5_0, 0, 6) +REG32(GHWPARAMS5, 0x54) + FIELD(GHWPARAMS5, GHWPARAMS5_31_28, 28, 4) + FIELD(GHWPARAMS5, GHWPARAMS5_27_22, 22, 6) + FIELD(GHWPARAMS5, GHWPARAMS5_21_16, 16, 6) + FIELD(GHWPARAMS5, GHWPARAMS5_15_10, 10, 6) + FIELD(GHWPARAMS5, GHWPARAMS5_9_4, 4, 6) + FIELD(GHWPARAMS5, GHWPARAMS5_3_0, 0, 4) +REG32(GHWPARAMS6, 0x58) + FIELD(GHWPARAMS6, GHWPARAMS6_31_16, 16, 16) + FIELD(GHWPARAMS6, BUSFLTRSSUPPORT, 15, 1) + FIELD(GHWPARAMS6, BCSUPPORT, 14, 1) + FIELD(GHWPARAMS6, OTG_SS_SUPPORT, 13, 1) + FIELD(GHWPARAMS6, ADPSUPPORT, 12, 1) + FIELD(GHWPARAMS6, HNPSUPPORT, 11, 1) + FIELD(GHWPARAMS6, SRPSUPPORT, 10, 1) + FIELD(GHWPARAMS6, GHWPARAMS6_9_8, 8, 2) + FIELD(GHWPARAMS6, GHWPARAMS6_7, 7, 1) + FIELD(GHWPARAMS6, GHWPARAMS6_6, 6, 1) + FIELD(GHWPARAMS6, GHWPARAMS6_5_0, 0, 6) +REG32(GHWPARAMS7, 0x5c) + FIELD(GHWPARAMS7, GHWPARAMS7_31_16, 16, 16) + FIELD(GHWPARAMS7, GHWPARAMS7_15_0, 0, 16) +REG32(GDBGFIFOSPACE, 0x60) + FIELD(GDBGFIFOSPACE, SPACE_AVAILABLE, 16, 16) + FIELD(GDBGFIFOSPACE, RESERVED_15_9, 9, 7) + FIELD(GDBGFIFOSPACE, FIFO_QUEUE_SELECT, 0, 9) +REG32(GUCTL2, 0x9c) + FIELD(GUCTL2, RESERVED_31_26, 26, 6) + FIELD(GUCTL2, EN_HP_PM_TIMER, 19, 7) + FIELD(GUCTL2, NOLOWPWRDUR, 15, 4) + FIELD(GUCTL2, RST_ACTBITLATER, 14, 1) + FIELD(GUCTL2, RESERVED_13, 13, 1) + FIELD(GUCTL2, DISABLECFC, 11, 1) +REG32(GUSB2PHYCFG, 0x100) + FIELD(GUSB2PHYCFG, U2_FREECLK_EXISTS, 30, 1) + FIELD(GUSB2PHYCFG, ULPI_LPM_WITH_OPMODE_CHK, 29, 1) + FIELD(GUSB2PHYCFG, RESERVED_25, 25, 1) + FIELD(GUSB2PHYCFG, LSTRD, 22, 3) + FIELD(GUSB2PHYCFG, LSIPD, 19, 3) + FIELD(GUSB2PHYCFG, ULPIEXTVBUSINDIACTOR, 18, 1) + FIELD(GUSB2PHYCFG, ULPIEXTVBUSDRV, 17, 1) + FIELD(GUSB2PHYCFG, RESERVED_16, 16, 1) + FIELD(GUSB2PHYCFG, ULPIAUTORES, 15, 1) + FIELD(GUSB2PHYCFG, RESERVED_14, 14, 1) + FIELD(GUSB2PHYCFG, USBTRDTIM, 10, 4) + FIELD(GUSB2PHYCFG, XCVRDLY, 9, 1) + FIELD(GUSB2PHYCFG, ENBLSLPM, 8, 1) + FIELD(GUSB2PHYCFG, PHYSEL, 7, 1) + FIELD(GUSB2PHYCFG, SUSPENDUSB20, 6, 1) + FIELD(GUSB2PHYCFG, FSINTF, 5, 1) + FIELD(GUSB2PHYCFG, ULPI_UTMI_SEL, 4, 1) + FIELD(GUSB2PHYCFG, PHYIF, 3, 1) + FIELD(GUSB2PHYCFG, TOUTCAL, 0, 3) +REG32(GUSB2I2CCTL, 0x140) +REG32(GUSB2PHYACC_ULPI, 0x180) + FIELD(GUSB2PHYACC_ULPI, RESERVED_31_27, 27, 5) + FIELD(GUSB2PHYACC_ULPI, DISUIPIDRVR, 26, 1) + FIELD(GUSB2PHYACC_ULPI, NEWREGREQ, 25, 1) + FIELD(GUSB2PHYACC_ULPI, VSTSDONE, 24, 1) + FIELD(GUSB2PHYACC_ULPI, VSTSBSY, 23, 1) + FIELD(GUSB2PHYACC_ULPI, REGWR, 22, 1) + FIELD(GUSB2PHYACC_ULPI, REGADDR, 16, 6) + FIELD(GUSB2PHYACC_ULPI, EXTREGADDR, 8, 8) + FIELD(GUSB2PHYACC_ULPI, REGDATA, 0, 8) +REG32(GTXFIFOSIZ0, 0x200) + FIELD(GTXFIFOSIZ0, TXFSTADDR_N, 16, 16) + FIELD(GTXFIFOSIZ0, TXFDEP_N, 0, 16) +REG32(GTXFIFOSIZ1, 0x204) + FIELD(GTXFIFOSIZ1, TXFSTADDR_N, 16, 16) + FIELD(GTXFIFOSIZ1, TXFDEP_N, 0, 16) +REG32(GTXFIFOSIZ2, 0x208) + FIELD(GTXFIFOSIZ2, TXFSTADDR_N, 16, 16) + FIELD(GTXFIFOSIZ2, TXFDEP_N, 0, 16) +REG32(GTXFIFOSIZ3, 0x20c) + FIELD(GTXFIFOSIZ3, TXFSTADDR_N, 16, 16) + FIELD(GTXFIFOSIZ3, TXFDEP_N, 0, 16) +REG32(GTXFIFOSIZ4, 0x210) + FIELD(GTXFIFOSIZ4, TXFSTADDR_N, 16, 16) + FIELD(GTXFIFOSIZ4, TXFDEP_N, 0, 16) +REG32(GTXFIFOSIZ5, 0x214) + FIELD(GTXFIFOSIZ5, TXFSTADDR_N, 16, 16) + FIELD(GTXFIFOSIZ5, TXFDEP_N, 0, 16) +REG32(GRXFIFOSIZ0, 0x280) + FIELD(GRXFIFOSIZ0, RXFSTADDR_N, 16, 16) + FIELD(GRXFIFOSIZ0, RXFDEP_N, 0, 16) +REG32(GRXFIFOSIZ1, 0x284) + FIELD(GRXFIFOSIZ1, RXFSTADDR_N, 16, 16) + FIELD(GRXFIFOSIZ1, RXFDEP_N, 0, 16) +REG32(GRXFIFOSIZ2, 0x288) + FIELD(GRXFIFOSIZ2, RXFSTADDR_N, 16, 16) + FIELD(GRXFIFOSIZ2, RXFDEP_N, 0, 16) +REG32(GEVNTADRLO_0, 0x300) +REG32(GEVNTADRHI_0, 0x304) +REG32(GEVNTSIZ_0, 0x308) + FIELD(GEVNTSIZ_0, EVNTINTRPTMASK, 31, 1) + FIELD(GEVNTSIZ_0, RESERVED_30_16, 16, 15) + FIELD(GEVNTSIZ_0, EVENTSIZ, 0, 16) +REG32(GEVNTCOUNT_0, 0x30c) + FIELD(GEVNTCOUNT_0, EVNT_HANDLER_BUSY, 31, 1) + FIELD(GEVNTCOUNT_0, RESERVED_30_16, 16, 15) + FIELD(GEVNTCOUNT_0, EVNTCOUNT, 0, 16) +REG32(GEVNTADRLO_1, 0x310) +REG32(GEVNTADRHI_1, 0x314) +REG32(GEVNTSIZ_1, 0x318) + FIELD(GEVNTSIZ_1, EVNTINTRPTMASK, 31, 1) + FIELD(GEVNTSIZ_1, RESERVED_30_16, 16, 15) + FIELD(GEVNTSIZ_1, EVENTSIZ, 0, 16) +REG32(GEVNTCOUNT_1, 0x31c) + FIELD(GEVNTCOUNT_1, EVNT_HANDLER_BUSY, 31, 1) + FIELD(GEVNTCOUNT_1, RESERVED_30_16, 16, 15) + FIELD(GEVNTCOUNT_1, EVNTCOUNT, 0, 16) +REG32(GEVNTADRLO_2, 0x320) +REG32(GEVNTADRHI_2, 0x324) +REG32(GEVNTSIZ_2, 0x328) + FIELD(GEVNTSIZ_2, EVNTINTRPTMASK, 31, 1) + FIELD(GEVNTSIZ_2, RESERVED_30_16, 16, 15) + FIELD(GEVNTSIZ_2, EVENTSIZ, 0, 16) +REG32(GEVNTCOUNT_2, 0x32c) + FIELD(GEVNTCOUNT_2, EVNT_HANDLER_BUSY, 31, 1) + FIELD(GEVNTCOUNT_2, RESERVED_30_16, 16, 15) + FIELD(GEVNTCOUNT_2, EVNTCOUNT, 0, 16) +REG32(GEVNTADRLO_3, 0x330) +REG32(GEVNTADRHI_3, 0x334) +REG32(GEVNTSIZ_3, 0x338) + FIELD(GEVNTSIZ_3, EVNTINTRPTMASK, 31, 1) + FIELD(GEVNTSIZ_3, RESERVED_30_16, 16, 15) + FIELD(GEVNTSIZ_3, EVENTSIZ, 0, 16) +REG32(GEVNTCOUNT_3, 0x33c) + FIELD(GEVNTCOUNT_3, EVNT_HANDLER_BUSY, 31, 1) + FIELD(GEVNTCOUNT_3, RESERVED_30_16, 16, 15) + FIELD(GEVNTCOUNT_3, EVNTCOUNT, 0, 16) +REG32(GHWPARAMS8, 0x500) +REG32(GTXFIFOPRIDEV, 0x510) + FIELD(GTXFIFOPRIDEV, RESERVED_31_N, 6, 26) + FIELD(GTXFIFOPRIDEV, GTXFIFOPRIDEV, 0, 6) +REG32(GTXFIFOPRIHST, 0x518) + FIELD(GTXFIFOPRIHST, RESERVED_31_16, 3, 29) + FIELD(GTXFIFOPRIHST, GTXFIFOPRIHST, 0, 3) +REG32(GRXFIFOPRIHST, 0x51c) + FIELD(GRXFIFOPRIHST, RESERVED_31_16, 3, 29) + FIELD(GRXFIFOPRIHST, GRXFIFOPRIHST, 0, 3) +REG32(GDMAHLRATIO, 0x524) + FIELD(GDMAHLRATIO, RESERVED_31_13, 13, 19) + FIELD(GDMAHLRATIO, HSTRXFIFO, 8, 5) + FIELD(GDMAHLRATIO, RESERVED_7_5, 5, 3) + FIELD(GDMAHLRATIO, HSTTXFIFO, 0, 5) +REG32(GFLADJ, 0x530) + FIELD(GFLADJ, GFLADJ_REFCLK_240MHZDECR_PLS1, 31, 1) + FIELD(GFLADJ, GFLADJ_REFCLK_240MHZ_DECR, 24, 7) + FIELD(GFLADJ, GFLADJ_REFCLK_LPM_SEL, 23, 1) + FIELD(GFLADJ, RESERVED_22, 22, 1) + FIELD(GFLADJ, GFLADJ_REFCLK_FLADJ, 8, 14) + FIELD(GFLADJ, GFLADJ_30MHZ_SDBND_SEL, 7, 1) + FIELD(GFLADJ, GFLADJ_30MHZ, 0, 6) + +#define DWC3_GLOBAL_OFFSET 0xC100 +static void reset_csr(USBDWC3 * s) +{ + int i = 0; + /* + * We reset all CSR regs except GCTL, GUCTL, GSTS, GSNPSID, GGPIO, GUID, + * GUSB2PHYCFGn registers and GUSB3PIPECTLn registers. We will skip PHY + * register as we don't implement them. + */ + for (i = 0; i < USB_DWC3_R_MAX; i++) { + switch (i) { + case R_GCTL: + break; + case R_GSTS: + break; + case R_GSNPSID: + break; + case R_GGPIO: + break; + case R_GUID: + break; + case R_GUCTL: + break; + case R_GHWPARAMS0...R_GHWPARAMS7: + break; + case R_GHWPARAMS8: + break; + default: + register_reset(&s->regs_info[i]); + break; + } + } + + xhci_sysbus_reset(DEVICE(&s->sysbus_xhci)); +} + +static void usb_dwc3_gctl_postw(RegisterInfo *reg, uint64_t val64) +{ + USBDWC3 *s = USB_DWC3(reg->opaque); + + if (ARRAY_FIELD_EX32(s->regs, GCTL, CORESOFTRESET)) { + reset_csr(s); + } +} + +static void usb_dwc3_guid_postw(RegisterInfo *reg, uint64_t val64) +{ + USBDWC3 *s = USB_DWC3(reg->opaque); + + s->regs[R_GUID] = s->cfg.dwc_usb3_user; +} + +static const RegisterAccessInfo usb_dwc3_regs_info[] = { + { .name = "GSBUSCFG0", .addr = A_GSBUSCFG0, + .ro = 0xf300, + .unimp = 0xffffffff, + },{ .name = "GSBUSCFG1", .addr = A_GSBUSCFG1, + .reset = 0x300, + .ro = 0xffffe0ff, + .unimp = 0xffffffff, + },{ .name = "GTXTHRCFG", .addr = A_GTXTHRCFG, + .ro = 0xd000ffff, + .unimp = 0xffffffff, + },{ .name = "GRXTHRCFG", .addr = A_GRXTHRCFG, + .ro = 0xd007e000, + .unimp = 0xffffffff, + },{ .name = "GCTL", .addr = A_GCTL, + .reset = 0x30c13004, .post_write = usb_dwc3_gctl_postw, + },{ .name = "GPMSTS", .addr = A_GPMSTS, + .ro = 0xfffffff, + .unimp = 0xffffffff, + },{ .name = "GSTS", .addr = A_GSTS, + .reset = 0x7e800000, + .ro = 0xffffffcf, + .w1c = 0x30, + .unimp = 0xffffffff, + },{ .name = "GUCTL1", .addr = A_GUCTL1, + .reset = 0x198a, + .ro = 0x7800, + .unimp = 0xffffffff, + },{ .name = "GSNPSID", .addr = A_GSNPSID, + .reset = 0x5533330a, + .ro = 0xffffffff, + },{ .name = "GGPIO", .addr = A_GGPIO, + .ro = 0xffff, + .unimp = 0xffffffff, + },{ .name = "GUID", .addr = A_GUID, + .reset = 0x12345678, .post_write = usb_dwc3_guid_postw, + },{ .name = "GUCTL", .addr = A_GUCTL, + .reset = 0x0c808010, + .ro = 0x1c8000, + .unimp = 0xffffffff, + },{ .name = "GBUSERRADDRLO", .addr = A_GBUSERRADDRLO, + .ro = 0xffffffff, + },{ .name = "GBUSERRADDRHI", .addr = A_GBUSERRADDRHI, + .ro = 0xffffffff, + },{ .name = "GHWPARAMS0", .addr = A_GHWPARAMS0, + .ro = 0xffffffff, + },{ .name = "GHWPARAMS1", .addr = A_GHWPARAMS1, + .ro = 0xffffffff, + },{ .name = "GHWPARAMS2", .addr = A_GHWPARAMS2, + .ro = 0xffffffff, + },{ .name = "GHWPARAMS3", .addr = A_GHWPARAMS3, + .ro = 0xffffffff, + },{ .name = "GHWPARAMS4", .addr = A_GHWPARAMS4, + .ro = 0xffffffff, + },{ .name = "GHWPARAMS5", .addr = A_GHWPARAMS5, + .ro = 0xffffffff, + },{ .name = "GHWPARAMS6", .addr = A_GHWPARAMS6, + .ro = 0xffffffff, + },{ .name = "GHWPARAMS7", .addr = A_GHWPARAMS7, + .ro = 0xffffffff, + },{ .name = "GDBGFIFOSPACE", .addr = A_GDBGFIFOSPACE, + .reset = 0xa0000, + .ro = 0xfffffe00, + .unimp = 0xffffffff, + },{ .name = "GUCTL2", .addr = A_GUCTL2, + .reset = 0x40d, + .ro = 0x2000, + .unimp = 0xffffffff, + },{ .name = "GUSB2PHYCFG", .addr = A_GUSB2PHYCFG, + .reset = 0x40102410, + .ro = 0x1e014030, + .unimp = 0xffffffff, + },{ .name = "GUSB2I2CCTL", .addr = A_GUSB2I2CCTL, + .ro = 0xffffffff, + .unimp = 0xffffffff, + },{ .name = "GUSB2PHYACC_ULPI", .addr = A_GUSB2PHYACC_ULPI, + .ro = 0xfd000000, + .unimp = 0xffffffff, + },{ .name = "GTXFIFOSIZ0", .addr = A_GTXFIFOSIZ0, + .reset = 0x2c7000a, + .unimp = 0xffffffff, + },{ .name = "GTXFIFOSIZ1", .addr = A_GTXFIFOSIZ1, + .reset = 0x2d10103, + .unimp = 0xffffffff, + },{ .name = "GTXFIFOSIZ2", .addr = A_GTXFIFOSIZ2, + .reset = 0x3d40103, + .unimp = 0xffffffff, + },{ .name = "GTXFIFOSIZ3", .addr = A_GTXFIFOSIZ3, + .reset = 0x4d70083, + .unimp = 0xffffffff, + },{ .name = "GTXFIFOSIZ4", .addr = A_GTXFIFOSIZ4, + .reset = 0x55a0083, + .unimp = 0xffffffff, + },{ .name = "GTXFIFOSIZ5", .addr = A_GTXFIFOSIZ5, + .reset = 0x5dd0083, + .unimp = 0xffffffff, + },{ .name = "GRXFIFOSIZ0", .addr = A_GRXFIFOSIZ0, + .reset = 0x1c20105, + .unimp = 0xffffffff, + },{ .name = "GRXFIFOSIZ1", .addr = A_GRXFIFOSIZ1, + .reset = 0x2c70000, + .unimp = 0xffffffff, + },{ .name = "GRXFIFOSIZ2", .addr = A_GRXFIFOSIZ2, + .reset = 0x2c70000, + .unimp = 0xffffffff, + },{ .name = "GEVNTADRLO_0", .addr = A_GEVNTADRLO_0, + .unimp = 0xffffffff, + },{ .name = "GEVNTADRHI_0", .addr = A_GEVNTADRHI_0, + .unimp = 0xffffffff, + },{ .name = "GEVNTSIZ_0", .addr = A_GEVNTSIZ_0, + .ro = 0x7fff0000, + .unimp = 0xffffffff, + },{ .name = "GEVNTCOUNT_0", .addr = A_GEVNTCOUNT_0, + .ro = 0x7fff0000, + .unimp = 0xffffffff, + },{ .name = "GEVNTADRLO_1", .addr = A_GEVNTADRLO_1, + .unimp = 0xffffffff, + },{ .name = "GEVNTADRHI_1", .addr = A_GEVNTADRHI_1, + .unimp = 0xffffffff, + },{ .name = "GEVNTSIZ_1", .addr = A_GEVNTSIZ_1, + .ro = 0x7fff0000, + .unimp = 0xffffffff, + },{ .name = "GEVNTCOUNT_1", .addr = A_GEVNTCOUNT_1, + .ro = 0x7fff0000, + .unimp = 0xffffffff, + },{ .name = "GEVNTADRLO_2", .addr = A_GEVNTADRLO_2, + .unimp = 0xffffffff, + },{ .name = "GEVNTADRHI_2", .addr = A_GEVNTADRHI_2, + .unimp = 0xffffffff, + },{ .name = "GEVNTSIZ_2", .addr = A_GEVNTSIZ_2, + .ro = 0x7fff0000, + .unimp = 0xffffffff, + },{ .name = "GEVNTCOUNT_2", .addr = A_GEVNTCOUNT_2, + .ro = 0x7fff0000, + .unimp = 0xffffffff, + },{ .name = "GEVNTADRLO_3", .addr = A_GEVNTADRLO_3, + .unimp = 0xffffffff, + },{ .name = "GEVNTADRHI_3", .addr = A_GEVNTADRHI_3, + .unimp = 0xffffffff, + },{ .name = "GEVNTSIZ_3", .addr = A_GEVNTSIZ_3, + .ro = 0x7fff0000, + .unimp = 0xffffffff, + },{ .name = "GEVNTCOUNT_3", .addr = A_GEVNTCOUNT_3, + .ro = 0x7fff0000, + .unimp = 0xffffffff, + },{ .name = "GHWPARAMS8", .addr = A_GHWPARAMS8, + .ro = 0xffffffff, + },{ .name = "GTXFIFOPRIDEV", .addr = A_GTXFIFOPRIDEV, + .ro = 0xffffffc0, + .unimp = 0xffffffff, + },{ .name = "GTXFIFOPRIHST", .addr = A_GTXFIFOPRIHST, + .ro = 0xfffffff8, + .unimp = 0xffffffff, + },{ .name = "GRXFIFOPRIHST", .addr = A_GRXFIFOPRIHST, + .ro = 0xfffffff8, + .unimp = 0xffffffff, + },{ .name = "GDMAHLRATIO", .addr = A_GDMAHLRATIO, + .ro = 0xffffe0e0, + .unimp = 0xffffffff, + },{ .name = "GFLADJ", .addr = A_GFLADJ, + .reset = 0xc83f020, + .rsvd = 0x40, + .ro = 0x400040, + .unimp = 0xffffffff, + } +}; + +static void usb_dwc3_reset(DeviceState *dev) +{ + USBDWC3 *s = USB_DWC3(dev); + unsigned int i; + + for (i = 0; i < ARRAY_SIZE(s->regs_info); ++i) { + switch (i) { + case R_GHWPARAMS0...R_GHWPARAMS7: + break; + case R_GHWPARAMS8: + break; + default: + register_reset(&s->regs_info[i]); + }; + } + + xhci_sysbus_reset(DEVICE(&s->sysbus_xhci)); +} + +static const MemoryRegionOps usb_dwc3_ops = { + .read = register_read_memory, + .write = register_write_memory, + .endianness = DEVICE_LITTLE_ENDIAN, + .valid = { + .min_access_size = 4, + .max_access_size = 4, + }, +}; + +static void usb_dwc3_realize(DeviceState *dev, Error **errp) +{ + USBDWC3 *s = USB_DWC3(dev); + SysBusDevice *sbd = SYS_BUS_DEVICE(dev); + Error *err = NULL; + + sysbus_realize(SYS_BUS_DEVICE(&s->sysbus_xhci), &err); + if (err) { + error_propagate(errp, err); + return; + } + + memory_region_add_subregion(&s->iomem, 0, + sysbus_mmio_get_region(SYS_BUS_DEVICE(&s->sysbus_xhci), 0)); + sysbus_init_mmio(sbd, &s->iomem); + + /* + * Device Configuration + */ + s->regs[R_GHWPARAMS0] = 0x40204048 | s->cfg.mode; + s->regs[R_GHWPARAMS1] = 0x222493b; + s->regs[R_GHWPARAMS2] = 0x12345678; + s->regs[R_GHWPARAMS3] = 0x618c088; + s->regs[R_GHWPARAMS4] = 0x47822004; + s->regs[R_GHWPARAMS5] = 0x4202088; + s->regs[R_GHWPARAMS6] = 0x7850c20; + s->regs[R_GHWPARAMS7] = 0x0; + s->regs[R_GHWPARAMS8] = 0x478; +} + +static void usb_dwc3_init(Object *obj) +{ + USBDWC3 *s = USB_DWC3(obj); + RegisterInfoArray *reg_array; + + memory_region_init(&s->iomem, obj, TYPE_USB_DWC3, DWC3_SIZE); + reg_array = + register_init_block32(DEVICE(obj), usb_dwc3_regs_info, + ARRAY_SIZE(usb_dwc3_regs_info), + s->regs_info, s->regs, + &usb_dwc3_ops, + USB_DWC3_ERR_DEBUG, + USB_DWC3_R_MAX * 4); + memory_region_add_subregion(&s->iomem, + DWC3_GLOBAL_OFFSET, + &reg_array->mem); + object_initialize_child(obj, "dwc3-xhci", &s->sysbus_xhci, + TYPE_XHCI_SYSBUS); + qdev_alias_all_properties(DEVICE(&s->sysbus_xhci), obj); + + s->cfg.mode = HOST_MODE; +} + +static const VMStateDescription vmstate_usb_dwc3 = { + .name = "usb-dwc3", + .version_id = 1, + .fields = (VMStateField[]) { + VMSTATE_UINT32_ARRAY(regs, USBDWC3, USB_DWC3_R_MAX), + VMSTATE_UINT8(cfg.mode, USBDWC3), + VMSTATE_UINT32(cfg.dwc_usb3_user, USBDWC3), + VMSTATE_END_OF_LIST() + } +}; + +static Property usb_dwc3_properties[] = { + DEFINE_PROP_UINT32("DWC_USB3_USERID", USBDWC3, cfg.dwc_usb3_user, + 0x12345678), + DEFINE_PROP_END_OF_LIST(), +}; + +static void usb_dwc3_class_init(ObjectClass *klass, void *data) +{ + DeviceClass *dc = DEVICE_CLASS(klass); + + dc->reset = usb_dwc3_reset; + dc->realize = usb_dwc3_realize; + dc->vmsd = &vmstate_usb_dwc3; + device_class_set_props(dc, usb_dwc3_properties); +} + +static const TypeInfo usb_dwc3_info = { + .name = TYPE_USB_DWC3, + .parent = TYPE_SYS_BUS_DEVICE, + .instance_size = sizeof(USBDWC3), + .class_init = usb_dwc3_class_init, + .instance_init = usb_dwc3_init, +}; + +static void usb_dwc3_register_types(void) +{ + type_register_static(&usb_dwc3_info); +} + +type_init(usb_dwc3_register_types) diff --git a/hw/usb/Kconfig b/hw/usb/Kconfig index XXXXXXX..XXXXXXX 100644 --- a/hw/usb/Kconfig +++ b/hw/usb/Kconfig @@ -XXX,XX +XXX,XX @@ config IMX_USBPHY bool default y depends on USB + +config USB_DWC3 + bool + select USB_XHCI_SYSBUS + select REGISTER diff --git a/hw/usb/meson.build b/hw/usb/meson.build index XXXXXXX..XXXXXXX 100644 --- a/hw/usb/meson.build +++ b/hw/usb/meson.build @@ -XXX,XX +XXX,XX @@ softmmu_ss.add(when: 'CONFIG_USB_XHCI_SYSBUS', if_true: files('hcd-xhci-sysbus.c softmmu_ss.add(when: 'CONFIG_USB_XHCI_NEC', if_true: files('hcd-xhci-nec.c')) softmmu_ss.add(when: 'CONFIG_USB_MUSB', if_true: files('hcd-musb.c')) softmmu_ss.add(when: 'CONFIG_USB_DWC2', if_true: files('hcd-dwc2.c')) +softmmu_ss.add(when: 'CONFIG_USB_DWC3', if_true: files('hcd-dwc3.c')) softmmu_ss.add(when: 'CONFIG_TUSB6010', if_true: files('tusb6010.c')) softmmu_ss.add(when: 'CONFIG_IMX', if_true: files('chipidea.c')) -- 2.20.1 From: Sai Pavan Boddu <sai.pavan.boddu@xilinx.com> This model is a top level integration wrapper for hcd-dwc3 and versal-usb2-ctrl-regs modules, this is used by xilinx versal soc's and future xilinx usb subsystems would also be part of it. Signed-off-by: Sai Pavan Boddu <sai.pavan.boddu@xilinx.com> Reviewed-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com> Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Message-id: 1607023357-5096-4-git-send-email-sai.pavan.boddu@xilinx.com Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- include/hw/usb/xlnx-usb-subsystem.h | 45 ++++++++++++++ hw/usb/xlnx-usb-subsystem.c | 94 +++++++++++++++++++++++++++++ hw/usb/Kconfig | 5 ++ hw/usb/meson.build | 1 + 4 files changed, 145 insertions(+) create mode 100644 include/hw/usb/xlnx-usb-subsystem.h create mode 100644 hw/usb/xlnx-usb-subsystem.c diff --git a/include/hw/usb/xlnx-usb-subsystem.h b/include/hw/usb/xlnx-usb-subsystem.h new file mode 100644 index XXXXXXX..XXXXXXX --- /dev/null +++ b/include/hw/usb/xlnx-usb-subsystem.h @@ -XXX,XX +XXX,XX @@ +/* + * QEMU model of the Xilinx usb subsystem + * + * Copyright (c) 2020 Xilinx Inc. Sai Pavan Boddu <sai.pavan.boddu@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. + */ + +#ifndef _XLNX_VERSAL_USB_SUBSYSTEM_H_ +#define _XLNX_VERSAL_USB_SUBSYSTEM_H_ + +#include "hw/usb/xlnx-versal-usb2-ctrl-regs.h" +#include "hw/usb/hcd-dwc3.h" + +#define TYPE_XILINX_VERSAL_USB2 "xlnx.versal-usb2" + +#define VERSAL_USB2(obj) \ + OBJECT_CHECK(VersalUsb2, (obj), TYPE_XILINX_VERSAL_USB2) + +typedef struct VersalUsb2 { + SysBusDevice parent_obj; + MemoryRegion dwc3_mr; + MemoryRegion usb2Ctrl_mr; + + VersalUsb2CtrlRegs usb2Ctrl; + USBDWC3 dwc3; +} VersalUsb2; + +#endif diff --git a/hw/usb/xlnx-usb-subsystem.c b/hw/usb/xlnx-usb-subsystem.c new file mode 100644 index XXXXXXX..XXXXXXX --- /dev/null +++ b/hw/usb/xlnx-usb-subsystem.c @@ -XXX,XX +XXX,XX @@ +/* + * QEMU model of the Xilinx usb subsystem + * + * Copyright (c) 2020 Xilinx Inc. Sai Pavan Boddu <sai.pava.boddu@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 "hw/sysbus.h" +#include "hw/irq.h" +#include "hw/register.h" +#include "qemu/bitops.h" +#include "qemu/log.h" +#include "qom/object.h" +#include "qapi/error.h" +#include "hw/qdev-properties.h" +#include "hw/usb/xlnx-usb-subsystem.h" + +static void versal_usb2_realize(DeviceState *dev, Error **errp) +{ + VersalUsb2 *s = VERSAL_USB2(dev); + SysBusDevice *sbd = SYS_BUS_DEVICE(dev); + Error *err = NULL; + + sysbus_realize(SYS_BUS_DEVICE(&s->dwc3), &err); + if (err) { + error_propagate(errp, err); + return; + } + sysbus_realize(SYS_BUS_DEVICE(&s->usb2Ctrl), &err); + if (err) { + error_propagate(errp, err); + return; + } + sysbus_init_mmio(sbd, &s->dwc3_mr); + sysbus_init_mmio(sbd, &s->usb2Ctrl_mr); + qdev_pass_gpios(DEVICE(&s->dwc3.sysbus_xhci), dev, SYSBUS_DEVICE_GPIO_IRQ); +} + +static void versal_usb2_init(Object *obj) +{ + VersalUsb2 *s = VERSAL_USB2(obj); + + object_initialize_child(obj, "versal.dwc3", &s->dwc3, + TYPE_USB_DWC3); + object_initialize_child(obj, "versal.usb2-ctrl", &s->usb2Ctrl, + TYPE_XILINX_VERSAL_USB2_CTRL_REGS); + memory_region_init_alias(&s->dwc3_mr, obj, "versal.dwc3_alias", + &s->dwc3.iomem, 0, DWC3_SIZE); + memory_region_init_alias(&s->usb2Ctrl_mr, obj, "versal.usb2Ctrl_alias", + &s->usb2Ctrl.iomem, 0, USB2_REGS_R_MAX * 4); + qdev_alias_all_properties(DEVICE(&s->dwc3), obj); + qdev_alias_all_properties(DEVICE(&s->dwc3.sysbus_xhci), obj); + object_property_add_alias(obj, "dma", OBJECT(&s->dwc3.sysbus_xhci), "dma"); +} + +static void versal_usb2_class_init(ObjectClass *klass, void *data) +{ + DeviceClass *dc = DEVICE_CLASS(klass); + + dc->realize = versal_usb2_realize; +} + +static const TypeInfo versal_usb2_info = { + .name = TYPE_XILINX_VERSAL_USB2, + .parent = TYPE_SYS_BUS_DEVICE, + .instance_size = sizeof(VersalUsb2), + .class_init = versal_usb2_class_init, + .instance_init = versal_usb2_init, +}; + +static void versal_usb_types(void) +{ + type_register_static(&versal_usb2_info); +} + +type_init(versal_usb_types) diff --git a/hw/usb/Kconfig b/hw/usb/Kconfig index XXXXXXX..XXXXXXX 100644 --- a/hw/usb/Kconfig +++ b/hw/usb/Kconfig @@ -XXX,XX +XXX,XX @@ config USB_DWC3 bool select USB_XHCI_SYSBUS select REGISTER + +config XLNX_USB_SUBSYS + bool + default y if XLNX_VERSAL + select USB_DWC3 diff --git a/hw/usb/meson.build b/hw/usb/meson.build index XXXXXXX..XXXXXXX 100644 --- a/hw/usb/meson.build +++ b/hw/usb/meson.build @@ -XXX,XX +XXX,XX @@ softmmu_ss.add(when: 'CONFIG_TUSB6010', if_true: files('tusb6010.c')) softmmu_ss.add(when: 'CONFIG_IMX', if_true: files('chipidea.c')) softmmu_ss.add(when: 'CONFIG_IMX_USBPHY', if_true: files('imx-usb-phy.c')) specific_ss.add(when: 'CONFIG_XLNX_VERSAL', if_true: files('xlnx-versal-usb2-ctrl-regs.c')) +specific_ss.add(when: 'CONFIG_XLNX_USB_SUBSYS', if_true: files('xlnx-usb-subsystem.c')) # emulated usb devices softmmu_ss.add(when: 'CONFIG_USB', if_true: files('dev-hub.c')) -- 2.20.1 From: Vikram Garhwal <fnu.vikram@xilinx.com> Connect VersalUsb2 subsystem to xlnx-versal SOC, its placed in iou of lpd domain and configure it as dual port host controller. Add the respective guest dts nodes for "xlnx-versal-virt" machine. Signed-off-by: Vikram Garhwal <fnu.vikram@xilinx.com> Signed-off-by: Sai Pavan Boddu <sai.pavan.boddu@xilinx.com> Reviewed-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com> Message-id: 1607023357-5096-5-git-send-email-sai.pavan.boddu@xilinx.com Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- include/hw/arm/xlnx-versal.h | 9 ++++++ hw/arm/xlnx-versal-virt.c | 55 ++++++++++++++++++++++++++++++++++++ hw/arm/xlnx-versal.c | 26 +++++++++++++++++ 3 files changed, 90 insertions(+) diff --git a/include/hw/arm/xlnx-versal.h b/include/hw/arm/xlnx-versal.h index XXXXXXX..XXXXXXX 100644 --- a/include/hw/arm/xlnx-versal.h +++ b/include/hw/arm/xlnx-versal.h @@ -XXX,XX +XXX,XX @@ #include "hw/net/cadence_gem.h" #include "hw/rtc/xlnx-zynqmp-rtc.h" #include "qom/object.h" +#include "hw/usb/xlnx-usb-subsystem.h" #define TYPE_XLNX_VERSAL "xlnx-versal" OBJECT_DECLARE_SIMPLE_TYPE(Versal, XLNX_VERSAL) @@ -XXX,XX +XXX,XX @@ struct Versal { PL011State uart[XLNX_VERSAL_NR_UARTS]; CadenceGEMState gem[XLNX_VERSAL_NR_GEMS]; XlnxZDMA adma[XLNX_VERSAL_NR_ADMAS]; + VersalUsb2 usb; } iou; } lpd; @@ -XXX,XX +XXX,XX @@ struct Versal { #define VERSAL_UART0_IRQ_0 18 #define VERSAL_UART1_IRQ_0 19 +#define VERSAL_USB0_IRQ_0 22 #define VERSAL_GEM0_IRQ_0 56 #define VERSAL_GEM0_WAKE_IRQ_0 57 #define VERSAL_GEM1_IRQ_0 58 @@ -XXX,XX +XXX,XX @@ struct Versal { #define MM_OCM 0xfffc0000U #define MM_OCM_SIZE 0x40000 +#define MM_USB2_CTRL_REGS 0xFF9D0000 +#define MM_USB2_CTRL_REGS_SIZE 0x10000 + +#define MM_USB_0 0xFE200000 +#define MM_USB_0_SIZE 0x10000 + #define MM_TOP_DDR 0x0 #define MM_TOP_DDR_SIZE 0x80000000U #define MM_TOP_DDR_2 0x800000000ULL 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 @@ struct VersalVirt { uint32_t ethernet_phy[2]; uint32_t clk_125Mhz; uint32_t clk_25Mhz; + uint32_t usb; + uint32_t dwc; } phandle; struct arm_boot_info binfo; @@ -XXX,XX +XXX,XX @@ static void fdt_create(VersalVirt *s) s->phandle.clk_25Mhz = qemu_fdt_alloc_phandle(s->fdt); s->phandle.clk_125Mhz = qemu_fdt_alloc_phandle(s->fdt); + s->phandle.usb = qemu_fdt_alloc_phandle(s->fdt); + s->phandle.dwc = qemu_fdt_alloc_phandle(s->fdt); /* Create /chosen node for load_dtb. */ qemu_fdt_add_subnode(s->fdt, "/chosen"); @@ -XXX,XX +XXX,XX @@ static void fdt_add_timer_nodes(VersalVirt *s) compat, sizeof(compat)); } +static void fdt_add_usb_xhci_nodes(VersalVirt *s) +{ + const char clocknames[] = "bus_clk\0ref_clk"; + const char irq_name[] = "dwc_usb3"; + const char compatVersalDWC3[] = "xlnx,versal-dwc3"; + const char compatDWC3[] = "snps,dwc3"; + char *name = g_strdup_printf("/usb@%" PRIx32, MM_USB2_CTRL_REGS); + + qemu_fdt_add_subnode(s->fdt, name); + qemu_fdt_setprop(s->fdt, name, "compatible", + compatVersalDWC3, sizeof(compatVersalDWC3)); + qemu_fdt_setprop_sized_cells(s->fdt, name, "reg", + 2, MM_USB2_CTRL_REGS, + 2, MM_USB2_CTRL_REGS_SIZE); + qemu_fdt_setprop(s->fdt, name, "clock-names", + clocknames, sizeof(clocknames)); + qemu_fdt_setprop_cells(s->fdt, name, "clocks", + s->phandle.clk_25Mhz, s->phandle.clk_125Mhz); + qemu_fdt_setprop(s->fdt, name, "ranges", NULL, 0); + qemu_fdt_setprop_cell(s->fdt, name, "#address-cells", 2); + qemu_fdt_setprop_cell(s->fdt, name, "#size-cells", 2); + qemu_fdt_setprop_cell(s->fdt, name, "phandle", s->phandle.usb); + g_free(name); + + name = g_strdup_printf("/usb@%" PRIx32 "/dwc3@%" PRIx32, + MM_USB2_CTRL_REGS, MM_USB_0); + qemu_fdt_add_subnode(s->fdt, name); + qemu_fdt_setprop(s->fdt, name, "compatible", + compatDWC3, sizeof(compatDWC3)); + qemu_fdt_setprop_sized_cells(s->fdt, name, "reg", + 2, MM_USB_0, 2, MM_USB_0_SIZE); + qemu_fdt_setprop(s->fdt, name, "interrupt-names", + irq_name, sizeof(irq_name)); + qemu_fdt_setprop_cells(s->fdt, name, "interrupts", + GIC_FDT_IRQ_TYPE_SPI, VERSAL_USB0_IRQ_0, + GIC_FDT_IRQ_FLAGS_LEVEL_HI); + qemu_fdt_setprop_cell(s->fdt, name, + "snps,quirk-frame-length-adjustment", 0x20); + qemu_fdt_setprop_cells(s->fdt, name, "#stream-id-cells", 1); + qemu_fdt_setprop_string(s->fdt, name, "dr_mode", "host"); + qemu_fdt_setprop_string(s->fdt, name, "phy-names", "usb3-phy"); + qemu_fdt_setprop(s->fdt, name, "snps,dis_u2_susphy_quirk", NULL, 0); + qemu_fdt_setprop(s->fdt, name, "snps,dis_u3_susphy_quirk", NULL, 0); + qemu_fdt_setprop(s->fdt, name, "snps,refclk_fladj", NULL, 0); + qemu_fdt_setprop(s->fdt, name, "snps,mask_phy_reset", NULL, 0); + qemu_fdt_setprop_cell(s->fdt, name, "phandle", s->phandle.dwc); + qemu_fdt_setprop_string(s->fdt, name, "maximum-speed", "high-speed"); + g_free(name); +} + static void fdt_add_uart_nodes(VersalVirt *s) { uint64_t addrs[] = { MM_UART1, MM_UART0 }; @@ -XXX,XX +XXX,XX @@ static void versal_virt_init(MachineState *machine) fdt_add_gic_nodes(s); fdt_add_timer_nodes(s); fdt_add_zdma_nodes(s); + fdt_add_usb_xhci_nodes(s); fdt_add_sd_nodes(s); fdt_add_rtc_node(s); fdt_add_cpu_nodes(s, psci_conduit); diff --git a/hw/arm/xlnx-versal.c b/hw/arm/xlnx-versal.c index XXXXXXX..XXXXXXX 100644 --- a/hw/arm/xlnx-versal.c +++ b/hw/arm/xlnx-versal.c @@ -XXX,XX +XXX,XX @@ static void versal_create_uarts(Versal *s, qemu_irq *pic) } } +static void versal_create_usbs(Versal *s, qemu_irq *pic) +{ + DeviceState *dev; + MemoryRegion *mr; + + object_initialize_child(OBJECT(s), "usb2", &s->lpd.iou.usb, + TYPE_XILINX_VERSAL_USB2); + dev = DEVICE(&s->lpd.iou.usb); + + object_property_set_link(OBJECT(dev), "dma", OBJECT(&s->mr_ps), + &error_abort); + qdev_prop_set_uint32(dev, "intrs", 1); + qdev_prop_set_uint32(dev, "slots", 2); + + sysbus_realize(SYS_BUS_DEVICE(dev), &error_fatal); + + mr = sysbus_mmio_get_region(SYS_BUS_DEVICE(dev), 0); + memory_region_add_subregion(&s->mr_ps, MM_USB_0, mr); + + sysbus_connect_irq(SYS_BUS_DEVICE(dev), 0, pic[VERSAL_USB0_IRQ_0]); + + mr = sysbus_mmio_get_region(SYS_BUS_DEVICE(dev), 1); + memory_region_add_subregion(&s->mr_ps, MM_USB2_CTRL_REGS, mr); +} + static void versal_create_gems(Versal *s, qemu_irq *pic) { int i; @@ -XXX,XX +XXX,XX @@ static void versal_realize(DeviceState *dev, Error **errp) versal_create_apu_cpus(s); versal_create_apu_gic(s, pic); versal_create_uarts(s, pic); + versal_create_usbs(s, pic); versal_create_gems(s, pic); versal_create_admas(s, pic); versal_create_sds(s, pic); -- 2.20.1 From: Philippe Mathieu-Daudé <f4bug@amsat.org> Malicious user can set the feedback divisor for the PLLs to zero, triggering a floating-point exception (SIGFPE). As the datasheet [*] is not clear how hardware behaves when these bits are zeroes, use the maximum divisor possible (128) to avoid the software FPE. [*] Zynq-7000 TRM, UG585 (v1.12.2) B.28 System Level Control Registers (slcr) -> "Register (slcr) ARM_PLL_CTRL" 25.10.4 PLLs -> "Software-Controlled PLL Update" Fixes: 38867cb7ec9 ("hw/misc/zynq_slcr: add clock generation for uarts") Reported-by: Gaoning Pan <pgn@zju.edu.cn> Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Reviewed-by: Alistair Francis <alistair.francis@wdc.com> Reviewed-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com> Reviewed-by: Damien Hedde <damien.hedde@greensocs.com> Message-id: 20201210141610.884600-1-f4bug@amsat.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- hw/misc/zynq_slcr.c | 5 +++++ 1 file changed, 5 insertions(+) diff --git a/hw/misc/zynq_slcr.c b/hw/misc/zynq_slcr.c index XXXXXXX..XXXXXXX 100644 --- a/hw/misc/zynq_slcr.c +++ b/hw/misc/zynq_slcr.c @@ -XXX,XX +XXX,XX @@ static uint64_t zynq_slcr_compute_pll(uint64_t input, uint32_t ctrl_reg) return 0; } + /* Consider zero feedback as maximum divide ratio possible */ + if (!mult) { + mult = 1 << R_xxx_PLL_CTRL_PLL_FPDIV_LENGTH; + } + /* frequency multiplier -> period division */ return input / mult; } -- 2.20.1 From: Joe Komlodi <joe.komlodi@xilinx.com> The previous naming of the configuration registers made it sound like that if the bits were set the settings would be enabled, while the opposite is true. Signed-off-by: Joe Komlodi <komlodi@xilinx.com> Reviewed-by: Francisco Iglesias <francisco.iglesias@xilinx.com> Message-id: 1605568264-26376-2-git-send-email-komlodi@xilinx.com Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- hw/block/m25p80.c | 12 ++++++------ 1 file changed, 6 insertions(+), 6 deletions(-) diff --git a/hw/block/m25p80.c b/hw/block/m25p80.c index XXXXXXX..XXXXXXX 100644 --- a/hw/block/m25p80.c +++ b/hw/block/m25p80.c @@ -XXX,XX +XXX,XX @@ typedef struct FlashPartInfo { #define VCFG_WRAP_SEQUENTIAL 0x2 #define NVCFG_XIP_MODE_DISABLED (7 << 9) #define NVCFG_XIP_MODE_MASK (7 << 9) -#define VCFG_XIP_MODE_ENABLED (1 << 3) +#define VCFG_XIP_MODE_DISABLED (1 << 3) #define CFG_DUMMY_CLK_LEN 4 #define NVCFG_DUMMY_CLK_POS 12 #define VCFG_DUMMY_CLK_POS 4 @@ -XXX,XX +XXX,XX @@ typedef struct FlashPartInfo { #define EVCFG_VPP_ACCELERATOR (1 << 3) #define EVCFG_RESET_HOLD_ENABLED (1 << 4) #define NVCFG_DUAL_IO_MASK (1 << 2) -#define EVCFG_DUAL_IO_ENABLED (1 << 6) +#define EVCFG_DUAL_IO_DISABLED (1 << 6) #define NVCFG_QUAD_IO_MASK (1 << 3) -#define EVCFG_QUAD_IO_ENABLED (1 << 7) +#define EVCFG_QUAD_IO_DISABLED (1 << 7) #define NVCFG_4BYTE_ADDR_MASK (1 << 0) #define NVCFG_LOWER_SEGMENT_MASK (1 << 1) @@ -XXX,XX +XXX,XX @@ static void reset_memory(Flash *s) s->volatile_cfg |= VCFG_WRAP_SEQUENTIAL; if ((s->nonvolatile_cfg & NVCFG_XIP_MODE_MASK) != NVCFG_XIP_MODE_DISABLED) { - s->volatile_cfg |= VCFG_XIP_MODE_ENABLED; + s->volatile_cfg |= VCFG_XIP_MODE_DISABLED; } s->volatile_cfg |= deposit32(s->volatile_cfg, VCFG_DUMMY_CLK_POS, @@ -XXX,XX +XXX,XX @@ static void reset_memory(Flash *s) s->enh_volatile_cfg |= EVCFG_VPP_ACCELERATOR; s->enh_volatile_cfg |= EVCFG_RESET_HOLD_ENABLED; if (s->nonvolatile_cfg & NVCFG_DUAL_IO_MASK) { - s->enh_volatile_cfg |= EVCFG_DUAL_IO_ENABLED; + s->enh_volatile_cfg |= EVCFG_DUAL_IO_DISABLED; } if (s->nonvolatile_cfg & NVCFG_QUAD_IO_MASK) { - s->enh_volatile_cfg |= EVCFG_QUAD_IO_ENABLED; + s->enh_volatile_cfg |= EVCFG_QUAD_IO_DISABLED; } if (!(s->nonvolatile_cfg & NVCFG_4BYTE_ADDR_MASK)) { s->four_bytes_address_mode = true; -- 2.20.1 From: Joe Komlodi <joe.komlodi@xilinx.com> VCFG XIP is set (disabled) when the NVCFG XIP bits are all set (disabled). Signed-off-by: Joe Komlodi <komlodi@xilinx.com> Reviewed-by: Francisco Iglesias <francisco.iglesias@xilinx.com> Message-id: 1605568264-26376-3-git-send-email-komlodi@xilinx.com Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- hw/block/m25p80.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/hw/block/m25p80.c b/hw/block/m25p80.c index XXXXXXX..XXXXXXX 100644 --- a/hw/block/m25p80.c +++ b/hw/block/m25p80.c @@ -XXX,XX +XXX,XX @@ static void reset_memory(Flash *s) s->volatile_cfg |= VCFG_DUMMY; s->volatile_cfg |= VCFG_WRAP_SEQUENTIAL; if ((s->nonvolatile_cfg & NVCFG_XIP_MODE_MASK) - != NVCFG_XIP_MODE_DISABLED) { + == NVCFG_XIP_MODE_DISABLED) { s->volatile_cfg |= VCFG_XIP_MODE_DISABLED; } s->volatile_cfg |= deposit32(s->volatile_cfg, -- 2.20.1 From: Joe Komlodi <joe.komlodi@xilinx.com> Some Numonyx flash commands cannot be executed in DIO and QIO mode, such as trying to do DPP or DOR when in QIO mode. Signed-off-by: Joe Komlodi <komlodi@xilinx.com> Reviewed-by: Francisco Iglesias <francisco.iglesias@xilinx.com> Message-id: 1605568264-26376-4-git-send-email-komlodi@xilinx.com Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- hw/block/m25p80.c | 114 ++++++++++++++++++++++++++++++++++++++-------- 1 file changed, 95 insertions(+), 19 deletions(-) diff --git a/hw/block/m25p80.c b/hw/block/m25p80.c index XXXXXXX..XXXXXXX 100644 --- a/hw/block/m25p80.c +++ b/hw/block/m25p80.c @@ -XXX,XX +XXX,XX @@ typedef enum { MAN_GENERIC, } Manufacturer; +typedef enum { + MODE_STD = 0, + MODE_DIO = 1, + MODE_QIO = 2 +} SPIMode; + #define M25P80_INTERNAL_DATA_BUFFER_SZ 16 struct Flash { @@ -XXX,XX +XXX,XX @@ static void reset_memory(Flash *s) trace_m25p80_reset_done(s); } +static uint8_t numonyx_mode(Flash *s) +{ + if (!(s->enh_volatile_cfg & EVCFG_QUAD_IO_DISABLED)) { + return MODE_QIO; + } else if (!(s->enh_volatile_cfg & EVCFG_DUAL_IO_DISABLED)) { + return MODE_DIO; + } else { + return MODE_STD; + } +} + static void decode_fast_read_cmd(Flash *s) { s->needed_bytes = get_addr_length(s); @@ -XXX,XX +XXX,XX @@ static void decode_new_cmd(Flash *s, uint32_t value) case ERASE4_32K: case ERASE_SECTOR: case ERASE4_SECTOR: - case READ: - case READ4: - case DPP: - case QPP: - case QPP_4: case PP: case PP4: - case PP4_4: case DIE_ERASE: case RDID_90: case RDID_AB: @@ -XXX,XX +XXX,XX @@ static void decode_new_cmd(Flash *s, uint32_t value) s->len = 0; s->state = STATE_COLLECTING_DATA; break; + case READ: + case READ4: + if (get_man(s) != MAN_NUMONYX || numonyx_mode(s) == MODE_STD) { + s->needed_bytes = get_addr_length(s); + s->pos = 0; + s->len = 0; + s->state = STATE_COLLECTING_DATA; + } else { + qemu_log_mask(LOG_GUEST_ERROR, "M25P80: Cannot execute cmd %x in " + "DIO or QIO mode\n", s->cmd_in_progress); + } + break; + case DPP: + if (get_man(s) != MAN_NUMONYX || numonyx_mode(s) != MODE_QIO) { + s->needed_bytes = get_addr_length(s); + s->pos = 0; + s->len = 0; + s->state = STATE_COLLECTING_DATA; + } else { + qemu_log_mask(LOG_GUEST_ERROR, "M25P80: Cannot execute cmd %x in " + "QIO mode\n", s->cmd_in_progress); + } + break; + case QPP: + case QPP_4: + case PP4_4: + if (get_man(s) != MAN_NUMONYX || numonyx_mode(s) != MODE_DIO) { + s->needed_bytes = get_addr_length(s); + s->pos = 0; + s->len = 0; + s->state = STATE_COLLECTING_DATA; + } else { + qemu_log_mask(LOG_GUEST_ERROR, "M25P80: Cannot execute cmd %x in " + "DIO mode\n", s->cmd_in_progress); + } + break; case FAST_READ: case FAST_READ4: + decode_fast_read_cmd(s); + break; case DOR: case DOR4: + if (get_man(s) != MAN_NUMONYX || numonyx_mode(s) != MODE_QIO) { + decode_fast_read_cmd(s); + } else { + qemu_log_mask(LOG_GUEST_ERROR, "M25P80: Cannot execute cmd %x in " + "QIO mode\n", s->cmd_in_progress); + } + break; case QOR: case QOR4: - decode_fast_read_cmd(s); + if (get_man(s) != MAN_NUMONYX || numonyx_mode(s) != MODE_DIO) { + decode_fast_read_cmd(s); + } else { + qemu_log_mask(LOG_GUEST_ERROR, "M25P80: Cannot execute cmd %x in " + "DIO mode\n", s->cmd_in_progress); + } break; case DIOR: case DIOR4: - decode_dio_read_cmd(s); + if (get_man(s) != MAN_NUMONYX || numonyx_mode(s) != MODE_QIO) { + decode_dio_read_cmd(s); + } else { + qemu_log_mask(LOG_GUEST_ERROR, "M25P80: Cannot execute cmd %x in " + "QIO mode\n", s->cmd_in_progress); + } break; case QIOR: case QIOR4: - decode_qio_read_cmd(s); + if (get_man(s) != MAN_NUMONYX || numonyx_mode(s) != MODE_DIO) { + decode_qio_read_cmd(s); + } else { + qemu_log_mask(LOG_GUEST_ERROR, "M25P80: Cannot execute cmd %x in " + "DIO mode\n", s->cmd_in_progress); + } break; case WRSR: @@ -XXX,XX +XXX,XX @@ static void decode_new_cmd(Flash *s, uint32_t value) break; case JEDEC_READ: - trace_m25p80_populated_jedec(s); - for (i = 0; i < s->pi->id_len; i++) { - s->data[i] = s->pi->id[i]; - } - for (; i < SPI_NOR_MAX_ID_LEN; i++) { - s->data[i] = 0; - } + if (get_man(s) != MAN_NUMONYX || numonyx_mode(s) == MODE_STD) { + trace_m25p80_populated_jedec(s); + for (i = 0; i < s->pi->id_len; i++) { + s->data[i] = s->pi->id[i]; + } + for (; i < SPI_NOR_MAX_ID_LEN; i++) { + s->data[i] = 0; + } - s->len = SPI_NOR_MAX_ID_LEN; - s->pos = 0; - s->state = STATE_READING_DATA; + s->len = SPI_NOR_MAX_ID_LEN; + s->pos = 0; + s->state = STATE_READING_DATA; + } else { + qemu_log_mask(LOG_GUEST_ERROR, "M25P80: Cannot execute JEDEC read " + "in DIO or QIO mode\n"); + } break; case RDCR: -- 2.20.1 From: Joe Komlodi <joe.komlodi@xilinx.com> Numonyx chips determine the number of cycles to wait based on bits 7:4 in the volatile configuration register. However, if these bits are 0x0 or 0xF, the number of dummy cycles to wait is 10 for QIOR and QIOR4 commands or when in QIO mode, and otherwise 8 for the currently supported fast read commands. [1] [1] https://www.micron.com/-/media/client/global/documents/products/data-sheet/nor-flash/serial-nor/mt25q/die-rev-b/mt25q_qlkt_u_02g_cbb_0.pdf?rev=9b167fbf2b3645efba6385949a72e453 Signed-off-by: Joe Komlodi <komlodi@xilinx.com> Reviewed-by: Francisco Iglesias <francisco.iglesias@xilinx.com> Message-id: 1605568264-26376-5-git-send-email-komlodi@xilinx.com Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- hw/block/m25p80.c | 30 +++++++++++++++++++++++++++--- 1 file changed, 27 insertions(+), 3 deletions(-) diff --git a/hw/block/m25p80.c b/hw/block/m25p80.c index XXXXXXX..XXXXXXX 100644 --- a/hw/block/m25p80.c +++ b/hw/block/m25p80.c @@ -XXX,XX +XXX,XX @@ static uint8_t numonyx_mode(Flash *s) } } +static uint8_t numonyx_extract_cfg_num_dummies(Flash *s) +{ + uint8_t num_dummies; + uint8_t mode; + assert(get_man(s) == MAN_NUMONYX); + + mode = numonyx_mode(s); + num_dummies = extract32(s->volatile_cfg, 4, 4); + + if (num_dummies == 0x0 || num_dummies == 0xf) { + switch (s->cmd_in_progress) { + case QIOR: + case QIOR4: + num_dummies = 10; + break; + default: + num_dummies = (mode == MODE_QIO) ? 10 : 8; + break; + } + } + + return num_dummies; +} + static void decode_fast_read_cmd(Flash *s) { s->needed_bytes = get_addr_length(s); @@ -XXX,XX +XXX,XX @@ static void decode_fast_read_cmd(Flash *s) s->needed_bytes += 8; break; case MAN_NUMONYX: - s->needed_bytes += extract32(s->volatile_cfg, 4, 4); + s->needed_bytes += numonyx_extract_cfg_num_dummies(s); break; case MAN_MACRONIX: if (extract32(s->volatile_cfg, 6, 2) == 1) { @@ -XXX,XX +XXX,XX @@ static void decode_dio_read_cmd(Flash *s) ); break; case MAN_NUMONYX: - s->needed_bytes += extract32(s->volatile_cfg, 4, 4); + s->needed_bytes += numonyx_extract_cfg_num_dummies(s); break; case MAN_MACRONIX: switch (extract32(s->volatile_cfg, 6, 2)) { @@ -XXX,XX +XXX,XX @@ static void decode_qio_read_cmd(Flash *s) ); break; case MAN_NUMONYX: - s->needed_bytes += extract32(s->volatile_cfg, 4, 4); + s->needed_bytes += numonyx_extract_cfg_num_dummies(s); break; case MAN_MACRONIX: switch (extract32(s->volatile_cfg, 6, 2)) { -- 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