Arm queue -- I have more stuff pending but I prefer to push this first lot out and keep the pull below 50 patches. Most of this is Alex's FP16 support work. -- PMM The following changes since commit 6697439794f72b3501ee16bb95d16854f9981421: Merge remote-tracking branch 'remotes/kraxel/tags/usb-20180227-pull-request' into staging (2018-02-27 17:50:46 +0000) are available in the Git repository at: git://git.linaro.org/people/pmaydell/qemu-arm.git tags/pull-target-arm-20180301 for you to fetch changes up to c22e580c2ad1cccef582e1490e732f254d4ac064: MAINTAINERS: Update my email address (2018-03-01 11:13:59 +0000) ---------------------------------------------------------------- target-arm queue: * update MAINTAINERS for Alistair's new email address * add Arm v8.2 FP16 arithmetic extension for linux-user * implement display connector emulation for vexpress board * xilinx_spips: Enable only two slaves when reading/writing with stripe * xilinx_spips: Use 8 dummy cycles with the QIOR/QIOR4 commands * hw: register: Run post_write hook on reset ---------------------------------------------------------------- Alex Bennée (31): include/exec/helper-head.h: support f16 in helper calls target/arm/cpu64: introduce ARM_V8_FP16 feature bit target/arm/cpu.h: update comment for half-precision values target/arm/cpu.h: add additional float_status flags target/arm/helper: pass explicit fpst to set_rmode arm/translate-a64: implement half-precision F(MIN|MAX)(V|NMV) arm/translate-a64: handle_3same_64 comment fix arm/translate-a64: initial decode for simd_three_reg_same_fp16 arm/translate-a64: add FP16 FADD/FABD/FSUB/FMUL/FDIV to simd_three_reg_same_fp16 arm/translate-a64: add FP16 F[A]C[EQ/GE/GT] to simd_three_reg_same_fp16 arm/translate-a64: add FP16 FMULA/X/S to simd_three_reg_same_fp16 arm/translate-a64: add FP16 FR[ECP/SQRT]S to simd_three_reg_same_fp16 arm/translate-a64: add FP16 pairwise ops simd_three_reg_same_fp16 arm/translate-a64: add FP16 FMULX/MLS/FMLA to simd_indexed arm/translate-a64: add FP16 x2 ops for simd_indexed arm/translate-a64: initial decode for simd_two_reg_misc_fp16 arm/translate-a64: add FP16 FPRINTx to simd_two_reg_misc_fp16 arm/translate-a64: add FCVTxx to simd_two_reg_misc_fp16 arm/translate-a64: add FP16 FCMxx (zero) to simd_two_reg_misc_fp16 arm/translate-a64: add FP16 SCVTF/UCVFT to simd_two_reg_misc_fp16 arm/translate-a64: add FP16 FNEG/FABS to simd_two_reg_misc_fp16 arm/helper.c: re-factor recpe and add recepe_f16 arm/translate-a64: add FP16 FRECPE arm/translate-a64: add FP16 FRCPX to simd_two_reg_misc_fp16 arm/translate-a64: add FP16 FSQRT to simd_two_reg_misc_fp16 arm/helper.c: re-factor rsqrte and add rsqrte_f16 arm/translate-a64: add FP16 FRSQRTE to simd_two_reg_misc_fp16 arm/translate-a64: add FP16 FMOV to simd_mod_imm arm/translate-a64: add all FP16 ops in simd_scalar_pairwise arm/translate-a64: implement simd_scalar_three_reg_same_fp16 arm/translate-a64: add all single op FP16 to handle_fp_1src_half Alistair Francis (2): hw: register: Run post_write hook on reset MAINTAINERS: Update my email address Corey Minyard (2): i2c: Fix some brace style issues i2c: Move the bus class to i2c.h Francisco Iglesias (2): xilinx_spips: Enable only two slaves when reading/writing with stripe xilinx_spips: Use 8 dummy cycles with the QIOR/QIOR4 commands Linus Walleij (3): hw/i2c-ddc: Do not fail writes hw/sii9022: Add support for Silicon Image SII9022 arm/vexpress: Add proper display connector emulation Peter Maydell (2): target/arm: Enable ARM_V8_FP16 feature bit for the AArch64 "any" CPU linux-user: Report AArch64 FP16 support via hwcap bits hw/display/Makefile.objs | 1 + include/exec/helper-head.h | 3 + include/fpu/softfloat.h | 18 +- include/hw/i2c/i2c.h | 23 +- include/hw/register.h | 6 +- target/arm/cpu.h | 34 +- target/arm/helper-a64.h | 33 + target/arm/helper.h | 14 +- hw/arm/vexpress.c | 6 +- hw/core/register.c | 8 + hw/display/sii9022.c | 191 ++++++ hw/i2c/core.c | 18 - hw/i2c/i2c-ddc.c | 4 +- hw/ssi/xilinx_spips.c | 43 +- linux-user/elfload.c | 2 + target/arm/cpu64.c | 1 + target/arm/helper-a64.c | 269 +++++++++ target/arm/helper.c | 481 ++++++++------- target/arm/translate-a64.c | 1266 +++++++++++++++++++++++++++++++++------ target/arm/translate.c | 12 +- MAINTAINERS | 12 +- default-configs/arm-softmmu.mak | 2 + hw/display/trace-events | 5 + 23 files changed, 1981 insertions(+), 471 deletions(-) create mode 100644 hw/display/sii9022.c From: Alistair Francis <alistair.francis@xilinx.com> Ensure that the post write hook is called during reset. This allows us to rely on the post write functions instead of having to call them from the reset() function. Signed-off-by: Alistair Francis <alistair.francis@xilinx.com> Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Message-id: d131e24b911653a945e46ca2d8f90f572469e1dd.1517856214.git.alistair.francis@xilinx.com Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- include/hw/register.h | 6 +++--- hw/core/register.c | 8 ++++++++ 2 files changed, 11 insertions(+), 3 deletions(-) diff --git a/include/hw/register.h b/include/hw/register.h index XXXXXXX..XXXXXXX 100644 --- a/include/hw/register.h +++ b/include/hw/register.h @@ -XXX,XX +XXX,XX @@ typedef struct RegisterInfoArray RegisterInfoArray; * immediately before the actual write. The returned value is what is written, * giving the handler a chance to modify the written value. * @post_write: Post write callback. Passed the written value. Most write side - * effects should be implemented here. + * effects should be implemented here. This is called during device reset. * * @post_read: Post read callback. Passes the value that is about to be returned * for a read. The return value from this function is what is ultimately read, @@ -XXX,XX +XXX,XX @@ uint64_t register_read(RegisterInfo *reg, uint64_t re, const char* prefix, bool debug); /** - * reset a register - * @reg: register to reset + * Resets a register. This will also call the post_write hook if it exists. + * @reg: The register to reset. */ void register_reset(RegisterInfo *reg); diff --git a/hw/core/register.c b/hw/core/register.c index XXXXXXX..XXXXXXX 100644 --- a/hw/core/register.c +++ b/hw/core/register.c @@ -XXX,XX +XXX,XX @@ uint64_t register_read(RegisterInfo *reg, uint64_t re, const char* prefix, void register_reset(RegisterInfo *reg) { + const RegisterAccessInfo *ac; + g_assert(reg); if (!reg->data || !reg->access) { return; } + ac = reg->access; + register_write_val(reg, reg->access->reset); + + if (ac->post_write) { + ac->post_write(reg, reg->access->reset); + } } void register_init(RegisterInfo *reg) -- 2.16.2 From: Francisco Iglesias <frasse.iglesias@gmail.com> Assert only the lower cs on bus 0 and upper cs on bus 1 when both buses and chip selects are enabled (e.g reading/writing with stripe). Signed-off-by: Francisco Iglesias <frasse.iglesias@gmail.com> Reviewed-by: Alistair Francis <alistair.francis@xilinx.com> Tested-by: Alistair Francis <alistair.francis@xilinx.com> Message-id: 20180223232233.31482-2-frasse.iglesias@gmail.com Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- hw/ssi/xilinx_spips.c | 41 +++++++++++++++++++++++++++++++++++++---- 1 file changed, 37 insertions(+), 4 deletions(-) diff --git a/hw/ssi/xilinx_spips.c b/hw/ssi/xilinx_spips.c index XXXXXXX..XXXXXXX 100644 --- a/hw/ssi/xilinx_spips.c +++ b/hw/ssi/xilinx_spips.c @@ -XXX,XX +XXX,XX @@ static void xilinx_spips_update_cs(XilinxSPIPS *s, int field) { int i; - for (i = 0; i < s->num_cs; i++) { + for (i = 0; i < s->num_cs * s->num_busses; i++) { bool old_state = s->cs_lines_state[i]; bool new_state = field & (1 << i); @@ -XXX,XX +XXX,XX @@ static void xilinx_spips_update_cs(XilinxSPIPS *s, int field) } qemu_set_irq(s->cs_lines[i], !new_state); } - if (!(field & ((1 << s->num_cs) - 1))) { + if (!(field & ((1 << (s->num_cs * s->num_busses)) - 1))) { s->snoop_state = SNOOP_CHECKING; s->cmd_dummies = 0; s->link_state = 1; @@ -XXX,XX +XXX,XX @@ static void xlnx_zynqmp_qspips_update_cs_lines(XlnxZynqMPQSPIPS *s) { if (s->regs[R_GQSPI_GF_SNAPSHOT]) { int field = ARRAY_FIELD_EX32(s->regs, GQSPI_GF_SNAPSHOT, CHIP_SELECT); - xilinx_spips_update_cs(XILINX_SPIPS(s), field); + bool upper_cs_sel = field & (1 << 1); + bool lower_cs_sel = field & 1; + bool bus0_enabled; + bool bus1_enabled; + uint8_t buses; + int cs = 0; + + buses = ARRAY_FIELD_EX32(s->regs, GQSPI_GF_SNAPSHOT, DATA_BUS_SELECT); + bus0_enabled = buses & 1; + bus1_enabled = buses & (1 << 1); + + if (bus0_enabled && bus1_enabled) { + if (lower_cs_sel) { + cs |= 1; + } + if (upper_cs_sel) { + cs |= 1 << 3; + } + } else if (bus0_enabled) { + if (lower_cs_sel) { + cs |= 1; + } + if (upper_cs_sel) { + cs |= 1 << 1; + } + } else if (bus1_enabled) { + if (lower_cs_sel) { + cs |= 1 << 2; + } + if (upper_cs_sel) { + cs |= 1 << 3; + } + } + xilinx_spips_update_cs(XILINX_SPIPS(s), cs); } } @@ -XXX,XX +XXX,XX @@ static void xilinx_spips_update_cs_lines(XilinxSPIPS *s) if (num_effective_busses(s) == 2) { /* Single bit chip-select for qspi */ field &= 0x1; - field |= field << 1; + field |= field << 3; /* Dual stack U-Page */ } else if (s->regs[R_LQSPI_CFG] & LQSPI_CFG_TWO_MEM && s->regs[R_LQSPI_STS] & LQSPI_CFG_U_PAGE) { -- 2.16.2 From: Francisco Iglesias <frasse.iglesias@gmail.com> Use 8 dummy cycles (4 dummy bytes) with the QIOR/QIOR4 commands in legacy mode for matching what is expected by Micron (Numonyx) flashes (the default target flash type of the QSPI). Signed-off-by: Francisco Iglesias <frasse.iglesias@gmail.com> Tested-by: Alistair Francis <alistair.francis@xilinx.com> Reviewed-by: Alistair Francis <alistair.francis@xilinx.com> Message-id: 20180223232233.31482-3-frasse.iglesias@gmail.com Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- hw/ssi/xilinx_spips.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/hw/ssi/xilinx_spips.c b/hw/ssi/xilinx_spips.c index XXXXXXX..XXXXXXX 100644 --- a/hw/ssi/xilinx_spips.c +++ b/hw/ssi/xilinx_spips.c @@ -XXX,XX +XXX,XX @@ static int xilinx_spips_num_dummies(XilinxQSPIPS *qs, uint8_t command) return 2; case QIOR: case QIOR_4: - return 5; + return 4; default: return -1; } -- 2.16.2 From: Corey Minyard <cminyard@mvista.com> Signed-off-by: Corey Minyard <cminyard@mvista.com> Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Linus Walleij <linus.walleij@linaro.org> Message-id: 20180227104903.21353-2-linus.walleij@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- include/hw/i2c/i2c.h | 6 ++---- hw/i2c/core.c | 3 +-- 2 files changed, 3 insertions(+), 6 deletions(-) diff --git a/include/hw/i2c/i2c.h b/include/hw/i2c/i2c.h index XXXXXXX..XXXXXXX 100644 --- a/include/hw/i2c/i2c.h +++ b/include/hw/i2c/i2c.h @@ -XXX,XX +XXX,XX @@ typedef struct I2CSlave I2CSlave; #define I2C_SLAVE_GET_CLASS(obj) \ OBJECT_GET_CLASS(I2CSlaveClass, (obj), TYPE_I2C_SLAVE) -typedef struct I2CSlaveClass -{ +typedef struct I2CSlaveClass { DeviceClass parent_class; /* Callbacks provided by the device. */ @@ -XXX,XX +XXX,XX @@ typedef struct I2CSlaveClass int (*event)(I2CSlave *s, enum i2c_event event); } I2CSlaveClass; -struct I2CSlave -{ +struct I2CSlave { DeviceState qdev; /* Remaining fields for internal use by the I2C code. */ diff --git a/hw/i2c/core.c b/hw/i2c/core.c index XXXXXXX..XXXXXXX 100644 --- a/hw/i2c/core.c +++ b/hw/i2c/core.c @@ -XXX,XX +XXX,XX @@ struct I2CNode { #define I2C_BROADCAST 0x00 -struct I2CBus -{ +struct I2CBus { BusState qbus; QLIST_HEAD(, I2CNode) current_devs; uint8_t saved_address; -- 2.16.2 From: Corey Minyard <cminyard@mvista.com> Some devices need access to it. Signed-off-by: Corey Minyard <cminyard@mvista.com> Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Linus Walleij <linus.walleij@linaro.org> Message-id: 20180227104903.21353-3-linus.walleij@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- include/hw/i2c/i2c.h | 17 +++++++++++++++++ hw/i2c/core.c | 17 ----------------- 2 files changed, 17 insertions(+), 17 deletions(-) diff --git a/include/hw/i2c/i2c.h b/include/hw/i2c/i2c.h index XXXXXXX..XXXXXXX 100644 --- a/include/hw/i2c/i2c.h +++ b/include/hw/i2c/i2c.h @@ -XXX,XX +XXX,XX @@ struct I2CSlave { uint8_t address; }; +#define TYPE_I2C_BUS "i2c-bus" +#define I2C_BUS(obj) OBJECT_CHECK(I2CBus, (obj), TYPE_I2C_BUS) + +typedef struct I2CNode I2CNode; + +struct I2CNode { + I2CSlave *elt; + QLIST_ENTRY(I2CNode) next; +}; + +struct I2CBus { + BusState qbus; + QLIST_HEAD(, I2CNode) current_devs; + uint8_t saved_address; + bool broadcast; +}; + I2CBus *i2c_init_bus(DeviceState *parent, const char *name); void i2c_set_slave_address(I2CSlave *dev, uint8_t address); int i2c_bus_busy(I2CBus *bus); diff --git a/hw/i2c/core.c b/hw/i2c/core.c index XXXXXXX..XXXXXXX 100644 --- a/hw/i2c/core.c +++ b/hw/i2c/core.c @@ -XXX,XX +XXX,XX @@ #include "qemu/osdep.h" #include "hw/i2c/i2c.h" -typedef struct I2CNode I2CNode; - -struct I2CNode { - I2CSlave *elt; - QLIST_ENTRY(I2CNode) next; -}; - #define I2C_BROADCAST 0x00 -struct I2CBus { - BusState qbus; - QLIST_HEAD(, I2CNode) current_devs; - uint8_t saved_address; - bool broadcast; -}; - static Property i2c_props[] = { DEFINE_PROP_UINT8("address", struct I2CSlave, address, 0), DEFINE_PROP_END_OF_LIST(), }; -#define TYPE_I2C_BUS "i2c-bus" -#define I2C_BUS(obj) OBJECT_CHECK(I2CBus, (obj), TYPE_I2C_BUS) - static const TypeInfo i2c_bus_info = { .name = TYPE_I2C_BUS, .parent = TYPE_BUS, -- 2.16.2 From: Linus Walleij <linus.walleij@linaro.org> The tx function of the DDC I2C slave emulation was returning 1 on all writes resulting in NACK in the I2C bus. Changing it to 0 makes the DDC I2C work fine with bit-banged I2C such as the versatile I2C. I guess it was not affecting whatever I2C controller this was used with until now, but with the Versatile I2C it surely does not work. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Linus Walleij <linus.walleij@linaro.org> Message-id: 20180227104903.21353-4-linus.walleij@linaro.org Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- hw/i2c/i2c-ddc.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/hw/i2c/i2c-ddc.c b/hw/i2c/i2c-ddc.c index XXXXXXX..XXXXXXX 100644 --- a/hw/i2c/i2c-ddc.c +++ b/hw/i2c/i2c-ddc.c @@ -XXX,XX +XXX,XX @@ static int i2c_ddc_tx(I2CSlave *i2c, uint8_t data) s->reg = data; s->firstbyte = false; DPRINTF("[EDID] Written new pointer: %u\n", data); - return 1; + return 0; } /* Ignore all writes */ s->reg++; - return 1; + return 0; } static void i2c_ddc_init(Object *obj) -- 2.16.2 From: Linus Walleij <linus.walleij@linaro.org> This adds support for emulating the Silicon Image SII9022 DVI/HDMI bridge. It's not very clever right now, it just acknowledges the switch into DDC I2C mode and back. Combining this with the existing DDC I2C emulation gives the right behavior on the Versatile Express emulation passing through the QEMU EDID to the emulated platform. Cc: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Linus Walleij <linus.walleij@linaro.org> Message-id: 20180227104903.21353-5-linus.walleij@linaro.org Reviewed-by: Peter Maydell <peter.maydell@linaro.org> [PMM: explictly reset ddc_req/ddc_skip_finish/ddc] Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- hw/display/Makefile.objs | 1 + hw/display/sii9022.c | 191 +++++++++++++++++++++++++++++++++++++++++++++++ hw/display/trace-events | 5 ++ 3 files changed, 197 insertions(+) create mode 100644 hw/display/sii9022.c diff --git a/hw/display/Makefile.objs b/hw/display/Makefile.objs index XXXXXXX..XXXXXXX 100644 --- a/hw/display/Makefile.objs +++ b/hw/display/Makefile.objs @@ -XXX,XX +XXX,XX @@ common-obj-$(CONFIG_VGA_CIRRUS) += cirrus_vga.o common-obj-$(CONFIG_G364FB) += g364fb.o common-obj-$(CONFIG_JAZZ_LED) += jazz_led.o common-obj-$(CONFIG_PL110) += pl110.o +common-obj-$(CONFIG_SII9022) += sii9022.o common-obj-$(CONFIG_SSD0303) += ssd0303.o common-obj-$(CONFIG_SSD0323) += ssd0323.o common-obj-$(CONFIG_XEN) += xenfb.o diff --git a/hw/display/sii9022.c b/hw/display/sii9022.c new file mode 100644 index XXXXXXX..XXXXXXX --- /dev/null +++ b/hw/display/sii9022.c @@ -XXX,XX +XXX,XX @@ +/* + * Silicon Image SiI9022 + * + * This is a pretty hollow emulation: all we do is acknowledge that we + * exist (chip ID) and confirm that we get switched over into DDC mode + * so the emulated host can proceed to read out EDID data. All subsequent + * set-up of connectors etc will be acknowledged and ignored. + * + * Copyright (C) 2018 Linus Walleij + * + * This work is licensed under the terms of the GNU GPL, version 2 or later. + * See the COPYING file in the top-level directory. + * SPDX-License-Identifier: GPL-2.0-or-later + */ + +#include "qemu/osdep.h" +#include "qemu-common.h" +#include "hw/i2c/i2c.h" +#include "hw/i2c/i2c-ddc.h" +#include "trace.h" + +#define SII9022_SYS_CTRL_DATA 0x1a +#define SII9022_SYS_CTRL_PWR_DWN 0x10 +#define SII9022_SYS_CTRL_AV_MUTE 0x08 +#define SII9022_SYS_CTRL_DDC_BUS_REQ 0x04 +#define SII9022_SYS_CTRL_DDC_BUS_GRTD 0x02 +#define SII9022_SYS_CTRL_OUTPUT_MODE 0x01 +#define SII9022_SYS_CTRL_OUTPUT_HDMI 1 +#define SII9022_SYS_CTRL_OUTPUT_DVI 0 +#define SII9022_REG_CHIPID 0x1b +#define SII9022_INT_ENABLE 0x3c +#define SII9022_INT_STATUS 0x3d +#define SII9022_INT_STATUS_HOTPLUG 0x01; +#define SII9022_INT_STATUS_PLUGGED 0x04; + +#define TYPE_SII9022 "sii9022" +#define SII9022(obj) OBJECT_CHECK(sii9022_state, (obj), TYPE_SII9022) + +typedef struct sii9022_state { + I2CSlave parent_obj; + uint8_t ptr; + bool addr_byte; + bool ddc_req; + bool ddc_skip_finish; + bool ddc; +} sii9022_state; + +static const VMStateDescription vmstate_sii9022 = { + .name = "sii9022", + .version_id = 1, + .minimum_version_id = 1, + .fields = (VMStateField[]) { + VMSTATE_I2C_SLAVE(parent_obj, sii9022_state), + VMSTATE_UINT8(ptr, sii9022_state), + VMSTATE_BOOL(addr_byte, sii9022_state), + VMSTATE_BOOL(ddc_req, sii9022_state), + VMSTATE_BOOL(ddc_skip_finish, sii9022_state), + VMSTATE_BOOL(ddc, sii9022_state), + VMSTATE_END_OF_LIST() + } +}; + +static int sii9022_event(I2CSlave *i2c, enum i2c_event event) +{ + sii9022_state *s = SII9022(i2c); + + switch (event) { + case I2C_START_SEND: + s->addr_byte = true; + break; + case I2C_START_RECV: + break; + case I2C_FINISH: + break; + case I2C_NACK: + break; + } + + return 0; +} + +static int sii9022_rx(I2CSlave *i2c) +{ + sii9022_state *s = SII9022(i2c); + uint8_t res = 0x00; + + switch (s->ptr) { + case SII9022_SYS_CTRL_DATA: + if (s->ddc_req) { + /* Acknowledge DDC bus request */ + res = SII9022_SYS_CTRL_DDC_BUS_GRTD | SII9022_SYS_CTRL_DDC_BUS_REQ; + } + break; + case SII9022_REG_CHIPID: + res = 0xb0; + break; + case SII9022_INT_STATUS: + /* Something is cold-plugged in, no interrupts */ + res = SII9022_INT_STATUS_PLUGGED; + break; + default: + break; + } + + trace_sii9022_read_reg(s->ptr, res); + s->ptr++; + + return res; +} + +static int sii9022_tx(I2CSlave *i2c, uint8_t data) +{ + sii9022_state *s = SII9022(i2c); + + if (s->addr_byte) { + s->ptr = data; + s->addr_byte = false; + return 0; + } + + switch (s->ptr) { + case SII9022_SYS_CTRL_DATA: + if (data & SII9022_SYS_CTRL_DDC_BUS_REQ) { + s->ddc_req = true; + if (data & SII9022_SYS_CTRL_DDC_BUS_GRTD) { + s->ddc = true; + /* Skip this finish since we just switched to DDC */ + s->ddc_skip_finish = true; + trace_sii9022_switch_mode("DDC"); + } + } else { + s->ddc_req = false; + s->ddc = false; + trace_sii9022_switch_mode("normal"); + } + break; + default: + break; + } + + trace_sii9022_write_reg(s->ptr, data); + s->ptr++; + + return 0; +} + +static void sii9022_reset(DeviceState *dev) +{ + sii9022_state *s = SII9022(dev); + + s->ptr = 0; + s->addr_byte = false; + s->ddc_req = false; + s->ddc_skip_finish = false; + s->ddc = false; +} + +static void sii9022_realize(DeviceState *dev, Error **errp) +{ + I2CBus *bus; + + bus = I2C_BUS(qdev_get_parent_bus(dev)); + i2c_create_slave(bus, TYPE_I2CDDC, 0x50); +} + +static void sii9022_class_init(ObjectClass *klass, void *data) +{ + DeviceClass *dc = DEVICE_CLASS(klass); + I2CSlaveClass *k = I2C_SLAVE_CLASS(klass); + + k->event = sii9022_event; + k->recv = sii9022_rx; + k->send = sii9022_tx; + dc->reset = sii9022_reset; + dc->realize = sii9022_realize; + dc->vmsd = &vmstate_sii9022; +} + +static const TypeInfo sii9022_info = { + .name = TYPE_SII9022, + .parent = TYPE_I2C_SLAVE, + .instance_size = sizeof(sii9022_state), + .class_init = sii9022_class_init, +}; + +static void sii9022_register_types(void) +{ + type_register_static(&sii9022_info); +} + +type_init(sii9022_register_types) diff --git a/hw/display/trace-events b/hw/display/trace-events index XXXXXXX..XXXXXXX 100644 --- a/hw/display/trace-events +++ b/hw/display/trace-events @@ -XXX,XX +XXX,XX @@ vga_cirrus_read_io(uint32_t addr, uint32_t val) "addr 0x%x, val 0x%x" vga_cirrus_write_io(uint32_t addr, uint32_t val) "addr 0x%x, val 0x%x" vga_cirrus_read_blt(uint32_t offset, uint32_t val) "offset 0x%x, val 0x%x" vga_cirrus_write_blt(uint32_t offset, uint32_t val) "offset 0x%x, val 0x%x" + +# hw/display/sii9022.c +sii9022_read_reg(uint8_t addr, uint8_t val) "addr 0x%02x, val 0x%02x" +sii9022_write_reg(uint8_t addr, uint8_t val) "addr 0x%02x, val 0x%02x" +sii9022_switch_mode(const char *mode) "mode: %s" -- 2.16.2 From: Linus Walleij <linus.walleij@linaro.org> This adds the SiI9022 (and implicitly EDID I2C) device to the ARM Versatile Express machine, and selects the two I2C devices necessary in the arm-softmmu.mak configuration so everything will build smoothly. I am implementing proper handling of the graphics in the Linux kernel and adding proper emulation of SiI9022 and EDID makes the driver probe as nicely as before, retrieving the resolutions supported by the "QEMU monitor" and overall just working nice. Cc: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Linus Walleij <linus.walleij@linaro.org> Message-id: 20180227104903.21353-6-linus.walleij@linaro.org Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- hw/arm/vexpress.c | 6 +++++- default-configs/arm-softmmu.mak | 2 ++ 2 files changed, 7 insertions(+), 1 deletion(-) diff --git a/hw/arm/vexpress.c b/hw/arm/vexpress.c index XXXXXXX..XXXXXXX 100644 --- a/hw/arm/vexpress.c +++ b/hw/arm/vexpress.c @@ -XXX,XX +XXX,XX @@ #include "hw/arm/arm.h" #include "hw/arm/primecell.h" #include "hw/devices.h" +#include "hw/i2c/i2c.h" #include "net/net.h" #include "sysemu/sysemu.h" #include "hw/boards.h" @@ -XXX,XX +XXX,XX @@ static void vexpress_common_init(MachineState *machine) uint32_t sys_id; DriveInfo *dinfo; pflash_t *pflash0; + I2CBus *i2c; ram_addr_t vram_size, sram_size; MemoryRegion *sysmem = get_system_memory(); MemoryRegion *vram = g_new(MemoryRegion, 1); @@ -XXX,XX +XXX,XX @@ static void vexpress_common_init(MachineState *machine) sysbus_create_simple("sp804", map[VE_TIMER01], pic[2]); sysbus_create_simple("sp804", map[VE_TIMER23], pic[3]); - /* VE_SERIALDVI: not modelled */ + dev = sysbus_create_simple("versatile_i2c", map[VE_SERIALDVI], NULL); + i2c = (I2CBus *)qdev_get_child_bus(dev, "i2c"); + i2c_create_slave(i2c, "sii9022", 0x39); sysbus_create_simple("pl031", map[VE_RTC], pic[4]); /* RTC */ diff --git a/default-configs/arm-softmmu.mak b/default-configs/arm-softmmu.mak index XXXXXXX..XXXXXXX 100644 --- a/default-configs/arm-softmmu.mak +++ b/default-configs/arm-softmmu.mak @@ -XXX,XX +XXX,XX @@ CONFIG_STELLARIS_INPUT=y CONFIG_STELLARIS_ENET=y CONFIG_SSD0303=y CONFIG_SSD0323=y +CONFIG_DDC=y +CONFIG_SII9022=y CONFIG_ADS7846=y CONFIG_MAX111X=y CONFIG_SSI=y -- 2.16.2 From: Alex Bennée <alex.bennee@linaro.org> This allows us to explicitly pass float16 to helpers rather than assuming uint32_t and dealing with the result. Of course they will be passed in i32 sized registers by default. Signed-off-by: Alex Bennée <alex.bennee@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20180227143852.11175-2-alex.bennee@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- include/exec/helper-head.h | 3 +++ 1 file changed, 3 insertions(+) diff --git a/include/exec/helper-head.h b/include/exec/helper-head.h index XXXXXXX..XXXXXXX 100644 --- a/include/exec/helper-head.h +++ b/include/exec/helper-head.h @@ -XXX,XX +XXX,XX @@ #define dh_alias_int i32 #define dh_alias_i64 i64 #define dh_alias_s64 i64 +#define dh_alias_f16 i32 #define dh_alias_f32 i32 #define dh_alias_f64 i64 #define dh_alias_ptr ptr @@ -XXX,XX +XXX,XX @@ #define dh_ctype_int int #define dh_ctype_i64 uint64_t #define dh_ctype_s64 int64_t +#define dh_ctype_f16 float16 #define dh_ctype_f32 float32 #define dh_ctype_f64 float64 #define dh_ctype_ptr void * @@ -XXX,XX +XXX,XX @@ #define dh_is_signed_s32 1 #define dh_is_signed_i64 0 #define dh_is_signed_s64 1 +#define dh_is_signed_f16 0 #define dh_is_signed_f32 0 #define dh_is_signed_f64 0 #define dh_is_signed_tl 0 -- 2.16.2 From: Alex Bennée <alex.bennee@linaro.org> Signed-off-by: Alex Bennée <alex.bennee@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20180227143852.11175-3-alex.bennee@linaro.org [PMM: postpone actually enabling feature until end of the patch series] Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/cpu.h | 1 + 1 file changed, 1 insertion(+) diff --git a/target/arm/cpu.h b/target/arm/cpu.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/cpu.h +++ b/target/arm/cpu.h @@ -XXX,XX +XXX,XX @@ enum arm_features { ARM_FEATURE_V8_SHA3, /* implements SHA3 part of v8 Crypto Extensions */ ARM_FEATURE_V8_SM3, /* implements SM3 part of v8 Crypto Extensions */ ARM_FEATURE_V8_SM4, /* implements SM4 part of v8 Crypto Extensions */ + ARM_FEATURE_V8_FP16, /* implements v8.2 half-precision float */ }; static inline int arm_feature(CPUARMState *env, int feature) -- 2.16.2 From: Alex Bennée <alex.bennee@linaro.org> Signed-off-by: Alex Bennée <alex.bennee@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20180227143852.11175-4-alex.bennee@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/cpu.h | 1 + 1 file changed, 1 insertion(+) diff --git a/target/arm/cpu.h b/target/arm/cpu.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/cpu.h +++ b/target/arm/cpu.h @@ -XXX,XX +XXX,XX @@ typedef struct { * Qn = regs[n].d[1]:regs[n].d[0] * Dn = regs[n].d[0] * Sn = regs[n].d[0] bits 31..0 + * Hn = regs[n].d[0] bits 15..0 * * This corresponds to the architecturally defined mapping between * the two execution states, and means we do not need to explicitly -- 2.16.2 From: Alex Bennée <alex.bennee@linaro.org> Half-precision flush to zero behaviour is controlled by a separate FZ16 bit in the FPCR. To handle this we pass a pointer to fp_status_fp16 when working on half-precision operations. The value of the presented FPCR is calculated from an amalgam of the two when read. Signed-off-by: Alex Bennée <alex.bennee@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20180227143852.11175-5-alex.bennee@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/cpu.h | 32 ++++++++++++++++++++++------ target/arm/helper.c | 26 ++++++++++++++++++----- target/arm/translate-a64.c | 53 +++++++++++++++++++++++++--------------------- 3 files changed, 75 insertions(+), 36 deletions(-) diff --git a/target/arm/cpu.h b/target/arm/cpu.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/cpu.h +++ b/target/arm/cpu.h @@ -XXX,XX +XXX,XX @@ typedef struct CPUARMState { /* scratch space when Tn are not sufficient. */ uint32_t scratch[8]; - /* fp_status is the "normal" fp status. standard_fp_status retains - * values corresponding to the ARM "Standard FPSCR Value", ie - * default-NaN, flush-to-zero, round-to-nearest and is used by - * any operations (generally Neon) which the architecture defines - * as controlled by the standard FPSCR value rather than the FPSCR. + /* There are a number of distinct float control structures: + * + * fp_status: is the "normal" fp status. + * fp_status_fp16: used for half-precision calculations + * standard_fp_status : the ARM "Standard FPSCR Value" + * + * Half-precision operations are governed by a separate + * flush-to-zero control bit in FPSCR:FZ16. We pass a separate + * status structure to control this. + * + * The "Standard FPSCR", ie default-NaN, flush-to-zero, + * round-to-nearest and is used by any operations (generally + * Neon) which the architecture defines as controlled by the + * standard FPSCR value rather than the FPSCR. * * To avoid having to transfer exception bits around, we simply * say that the FPSCR cumulative exception flags are the logical - * OR of the flags in the two fp statuses. This relies on the + * OR of the flags in the three fp statuses. This relies on the * only thing which needs to read the exception flags being * an explicit FPSCR read. */ float_status fp_status; + float_status fp_status_f16; float_status standard_fp_status; /* ZCR_EL[1-3] */ @@ -XXX,XX +XXX,XX @@ static inline void xpsr_write(CPUARMState *env, uint32_t val, uint32_t mask) uint32_t vfp_get_fpscr(CPUARMState *env); void vfp_set_fpscr(CPUARMState *env, uint32_t val); -/* For A64 the FPSCR is split into two logically distinct registers, +/* FPCR, Floating Point Control Register + * FPSR, Floating Poiht Status Register + * + * For A64 the FPSCR is split into two logically distinct registers, * FPCR and FPSR. However since they still use non-overlapping bits * we store the underlying state in fpscr and just mask on read/write. */ #define FPSR_MASK 0xf800009f #define FPCR_MASK 0x07f79f00 + +#define FPCR_FZ16 (1 << 19) /* ARMv8.2+, FP16 flush-to-zero */ +#define FPCR_FZ (1 << 24) /* Flush-to-zero enable bit */ +#define FPCR_DN (1 << 25) /* Default NaN enable bit */ + static inline uint32_t vfp_get_fpsr(CPUARMState *env) { return vfp_get_fpscr(env) & FPSR_MASK; diff --git a/target/arm/helper.c b/target/arm/helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper.c +++ b/target/arm/helper.c @@ -XXX,XX +XXX,XX @@ uint32_t HELPER(vfp_get_fpscr)(CPUARMState *env) | (env->vfp.vec_stride << 20); i = get_float_exception_flags(&env->vfp.fp_status); i |= get_float_exception_flags(&env->vfp.standard_fp_status); + i |= get_float_exception_flags(&env->vfp.fp_status_f16); fpscr |= vfp_exceptbits_from_host(i); return fpscr; } @@ -XXX,XX +XXX,XX @@ void HELPER(vfp_set_fpscr)(CPUARMState *env, uint32_t val) break; } set_float_rounding_mode(i, &env->vfp.fp_status); + set_float_rounding_mode(i, &env->vfp.fp_status_f16); } - if (changed & (1 << 24)) { - set_flush_to_zero((val & (1 << 24)) != 0, &env->vfp.fp_status); - set_flush_inputs_to_zero((val & (1 << 24)) != 0, &env->vfp.fp_status); + if (changed & FPCR_FZ16) { + bool ftz_enabled = val & FPCR_FZ16; + set_flush_to_zero(ftz_enabled, &env->vfp.fp_status_f16); + set_flush_inputs_to_zero(ftz_enabled, &env->vfp.fp_status_f16); + } + if (changed & FPCR_FZ) { + bool ftz_enabled = val & FPCR_FZ; + set_flush_to_zero(ftz_enabled, &env->vfp.fp_status); + set_flush_inputs_to_zero(ftz_enabled, &env->vfp.fp_status); + } + if (changed & FPCR_DN) { + bool dnan_enabled = val & FPCR_DN; + set_default_nan_mode(dnan_enabled, &env->vfp.fp_status); + set_default_nan_mode(dnan_enabled, &env->vfp.fp_status_f16); } - if (changed & (1 << 25)) - set_default_nan_mode((val & (1 << 25)) != 0, &env->vfp.fp_status); + /* The exception flags are ORed together when we read fpscr so we + * only need to preserve the current state in one of our + * float_status values. + */ i = vfp_exceptbits_to_host(val); set_float_exception_flags(i, &env->vfp.fp_status); + set_float_exception_flags(0, &env->vfp.fp_status_f16); set_float_exception_flags(0, &env->vfp.standard_fp_status); } 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 write_fp_sreg(DisasContext *s, int reg, TCGv_i32 v) tcg_temp_free_i64(tmp); } -static TCGv_ptr get_fpstatus_ptr(void) +static TCGv_ptr get_fpstatus_ptr(bool is_f16) { TCGv_ptr statusptr = tcg_temp_new_ptr(); int offset; - /* In A64 all instructions (both FP and Neon) use the FPCR; - * there is no equivalent of the A32 Neon "standard FPSCR value" - * and all operations use vfp.fp_status. + /* In A64 all instructions (both FP and Neon) use the FPCR; there + * is no equivalent of the A32 Neon "standard FPSCR value". + * However half-precision operations operate under a different + * FZ16 flag and use vfp.fp_status_f16 instead of vfp.fp_status. */ - offset = offsetof(CPUARMState, vfp.fp_status); + if (is_f16) { + offset = offsetof(CPUARMState, vfp.fp_status_f16); + } else { + offset = offsetof(CPUARMState, vfp.fp_status); + } tcg_gen_addi_ptr(statusptr, cpu_env, offset); return statusptr; } @@ -XXX,XX +XXX,XX @@ static void handle_fp_compare(DisasContext *s, bool is_double, bool cmp_with_zero, bool signal_all_nans) { TCGv_i64 tcg_flags = tcg_temp_new_i64(); - TCGv_ptr fpst = get_fpstatus_ptr(); + TCGv_ptr fpst = get_fpstatus_ptr(false); if (is_double) { TCGv_i64 tcg_vn, tcg_vm; @@ -XXX,XX +XXX,XX @@ static void handle_fp_1src_single(DisasContext *s, int opcode, int rd, int rn) TCGv_i32 tcg_op; TCGv_i32 tcg_res; - fpst = get_fpstatus_ptr(); + fpst = get_fpstatus_ptr(false); tcg_op = read_fp_sreg(s, rn); tcg_res = tcg_temp_new_i32(); @@ -XXX,XX +XXX,XX @@ static void handle_fp_1src_double(DisasContext *s, int opcode, int rd, int rn) return; } - fpst = get_fpstatus_ptr(); + fpst = get_fpstatus_ptr(false); tcg_op = read_fp_dreg(s, rn); tcg_res = tcg_temp_new_i64(); @@ -XXX,XX +XXX,XX @@ static void handle_fp_2src_single(DisasContext *s, int opcode, TCGv_ptr fpst; tcg_res = tcg_temp_new_i32(); - fpst = get_fpstatus_ptr(); + fpst = get_fpstatus_ptr(false); tcg_op1 = read_fp_sreg(s, rn); tcg_op2 = read_fp_sreg(s, rm); @@ -XXX,XX +XXX,XX @@ static void handle_fp_2src_double(DisasContext *s, int opcode, TCGv_ptr fpst; tcg_res = tcg_temp_new_i64(); - fpst = get_fpstatus_ptr(); + fpst = get_fpstatus_ptr(false); tcg_op1 = read_fp_dreg(s, rn); tcg_op2 = read_fp_dreg(s, rm); @@ -XXX,XX +XXX,XX @@ static void handle_fp_3src_single(DisasContext *s, bool o0, bool o1, { TCGv_i32 tcg_op1, tcg_op2, tcg_op3; TCGv_i32 tcg_res = tcg_temp_new_i32(); - TCGv_ptr fpst = get_fpstatus_ptr(); + TCGv_ptr fpst = get_fpstatus_ptr(false); tcg_op1 = read_fp_sreg(s, rn); tcg_op2 = read_fp_sreg(s, rm); @@ -XXX,XX +XXX,XX @@ static void handle_fp_3src_double(DisasContext *s, bool o0, bool o1, { TCGv_i64 tcg_op1, tcg_op2, tcg_op3; TCGv_i64 tcg_res = tcg_temp_new_i64(); - TCGv_ptr fpst = get_fpstatus_ptr(); + TCGv_ptr fpst = get_fpstatus_ptr(false); tcg_op1 = read_fp_dreg(s, rn); tcg_op2 = read_fp_dreg(s, rm); @@ -XXX,XX +XXX,XX @@ static void handle_fpfpcvt(DisasContext *s, int rd, int rn, int opcode, TCGv_ptr tcg_fpstatus; TCGv_i32 tcg_shift; - tcg_fpstatus = get_fpstatus_ptr(); + tcg_fpstatus = get_fpstatus_ptr(false); tcg_shift = tcg_const_i32(64 - scale); @@ -XXX,XX +XXX,XX @@ static void disas_simd_across_lanes(DisasContext *s, uint32_t insn) TCGv_i32 tcg_elt1 = tcg_temp_new_i32(); TCGv_i32 tcg_elt2 = tcg_temp_new_i32(); TCGv_i32 tcg_elt3 = tcg_temp_new_i32(); - TCGv_ptr fpst = get_fpstatus_ptr(); + TCGv_ptr fpst = get_fpstatus_ptr(false); assert(esize == 32); assert(elements == 4); @@ -XXX,XX +XXX,XX @@ static void disas_simd_scalar_pairwise(DisasContext *s, uint32_t insn) } size = extract32(size, 0, 1) ? 3 : 2; - fpst = get_fpstatus_ptr(); + fpst = get_fpstatus_ptr(false); break; default: unallocated_encoding(s); @@ -XXX,XX +XXX,XX @@ static void handle_simd_intfp_conv(DisasContext *s, int rd, int rn, int fracbits, int size) { bool is_double = size == 3 ? true : false; - TCGv_ptr tcg_fpst = get_fpstatus_ptr(); + TCGv_ptr tcg_fpst = get_fpstatus_ptr(false); TCGv_i32 tcg_shift = tcg_const_i32(fracbits); TCGv_i64 tcg_int = tcg_temp_new_i64(); TCGMemOp mop = size | (is_signed ? MO_SIGN : 0); @@ -XXX,XX +XXX,XX @@ static void handle_simd_shift_fpint_conv(DisasContext *s, bool is_scalar, tcg_rmode = tcg_const_i32(arm_rmode_to_sf(FPROUNDING_ZERO)); gen_helper_set_rmode(tcg_rmode, tcg_rmode, cpu_env); - tcg_fpstatus = get_fpstatus_ptr(); + tcg_fpstatus = get_fpstatus_ptr(false); tcg_shift = tcg_const_i32(fracbits); if (is_double) { @@ -XXX,XX +XXX,XX @@ static void handle_3same_float(DisasContext *s, int size, int elements, int fpopcode, int rd, int rn, int rm) { int pass; - TCGv_ptr fpst = get_fpstatus_ptr(); + TCGv_ptr fpst = get_fpstatus_ptr(false); for (pass = 0; pass < elements; pass++) { if (size) { @@ -XXX,XX +XXX,XX @@ static void handle_2misc_fcmp_zero(DisasContext *s, int opcode, return; } - fpst = get_fpstatus_ptr(); + fpst = get_fpstatus_ptr(false); if (is_double) { TCGv_i64 tcg_op = tcg_temp_new_i64(); @@ -XXX,XX +XXX,XX @@ static void handle_2misc_reciprocal(DisasContext *s, int opcode, int size, int rn, int rd) { bool is_double = (size == 3); - TCGv_ptr fpst = get_fpstatus_ptr(); + TCGv_ptr fpst = get_fpstatus_ptr(false); if (is_double) { TCGv_i64 tcg_op = tcg_temp_new_i64(); @@ -XXX,XX +XXX,XX @@ static void disas_simd_scalar_two_reg_misc(DisasContext *s, uint32_t insn) if (is_fcvt) { tcg_rmode = tcg_const_i32(arm_rmode_to_sf(rmode)); gen_helper_set_rmode(tcg_rmode, tcg_rmode, cpu_env); - tcg_fpstatus = get_fpstatus_ptr(); + tcg_fpstatus = get_fpstatus_ptr(false); } else { tcg_rmode = NULL; tcg_fpstatus = NULL; @@ -XXX,XX +XXX,XX @@ static void handle_simd_3same_pair(DisasContext *s, int is_q, int u, int opcode, /* Floating point operations need fpst */ if (opcode >= 0x58) { - fpst = get_fpstatus_ptr(); + fpst = get_fpstatus_ptr(false); } else { fpst = NULL; } @@ -XXX,XX +XXX,XX @@ static void disas_simd_two_reg_misc(DisasContext *s, uint32_t insn) } if (need_fpstatus) { - tcg_fpstatus = get_fpstatus_ptr(); + tcg_fpstatus = get_fpstatus_ptr(false); } else { tcg_fpstatus = NULL; } @@ -XXX,XX +XXX,XX @@ static void disas_simd_indexed(DisasContext *s, uint32_t insn) } if (is_fp) { - fpst = get_fpstatus_ptr(); + fpst = get_fpstatus_ptr(false); } else { fpst = NULL; } -- 2.16.2 From: Alex Bennée <alex.bennee@linaro.org> As the rounding mode is now split between FP16 and the rest of floating point we need to be explicit when tweaking it. Instead of passing the CPU env we now pass the appropriate fpst pointer directly. Signed-off-by: Alex Bennée <alex.bennee@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20180227143852.11175-6-alex.bennee@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/helper.h | 2 +- target/arm/helper.c | 4 ++-- target/arm/translate-a64.c | 26 +++++++++++++------------- target/arm/translate.c | 12 ++++++------ 4 files changed, 22 insertions(+), 22 deletions(-) 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_3(vfp_uhtod, f64, i64, i32, ptr) DEF_HELPER_3(vfp_ultod, f64, i64, i32, ptr) DEF_HELPER_3(vfp_uqtod, f64, i64, i32, ptr) -DEF_HELPER_FLAGS_2(set_rmode, TCG_CALL_NO_RWG, i32, i32, env) +DEF_HELPER_FLAGS_2(set_rmode, TCG_CALL_NO_RWG, i32, i32, ptr) DEF_HELPER_FLAGS_2(set_neon_rmode, TCG_CALL_NO_RWG, i32, i32, env) DEF_HELPER_2(vfp_fcvt_f16_to_f32, f32, i32, env) diff --git a/target/arm/helper.c b/target/arm/helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper.c +++ b/target/arm/helper.c @@ -XXX,XX +XXX,XX @@ VFP_CONV_FIX_A64(uq, s, 32, 64, uint64) /* Set the current fp rounding mode and return the old one. * The argument is a softfloat float_round_ value. */ -uint32_t HELPER(set_rmode)(uint32_t rmode, CPUARMState *env) +uint32_t HELPER(set_rmode)(uint32_t rmode, void *fpstp) { - float_status *fp_status = &env->vfp.fp_status; + float_status *fp_status = fpstp; uint32_t prev_rmode = get_float_rounding_mode(fp_status); set_float_rounding_mode(rmode, fp_status); diff --git a/target/arm/translate-a64.c b/target/arm/translate-a64.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-a64.c +++ b/target/arm/translate-a64.c @@ -XXX,XX +XXX,XX @@ static void handle_fp_1src_single(DisasContext *s, int opcode, int rd, int rn) { TCGv_i32 tcg_rmode = tcg_const_i32(arm_rmode_to_sf(opcode & 7)); - gen_helper_set_rmode(tcg_rmode, tcg_rmode, cpu_env); + gen_helper_set_rmode(tcg_rmode, tcg_rmode, fpst); gen_helper_rints(tcg_res, tcg_op, fpst); - gen_helper_set_rmode(tcg_rmode, tcg_rmode, cpu_env); + gen_helper_set_rmode(tcg_rmode, tcg_rmode, fpst); tcg_temp_free_i32(tcg_rmode); break; } @@ -XXX,XX +XXX,XX @@ static void handle_fp_1src_double(DisasContext *s, int opcode, int rd, int rn) { TCGv_i32 tcg_rmode = tcg_const_i32(arm_rmode_to_sf(opcode & 7)); - gen_helper_set_rmode(tcg_rmode, tcg_rmode, cpu_env); + gen_helper_set_rmode(tcg_rmode, tcg_rmode, fpst); gen_helper_rintd(tcg_res, tcg_op, fpst); - gen_helper_set_rmode(tcg_rmode, tcg_rmode, cpu_env); + gen_helper_set_rmode(tcg_rmode, tcg_rmode, fpst); tcg_temp_free_i32(tcg_rmode); break; } @@ -XXX,XX +XXX,XX @@ static void handle_fpfpcvt(DisasContext *s, int rd, int rn, int opcode, tcg_rmode = tcg_const_i32(arm_rmode_to_sf(rmode)); - gen_helper_set_rmode(tcg_rmode, tcg_rmode, cpu_env); + gen_helper_set_rmode(tcg_rmode, tcg_rmode, tcg_fpstatus); if (is_double) { TCGv_i64 tcg_double = read_fp_dreg(s, rn); @@ -XXX,XX +XXX,XX @@ static void handle_fpfpcvt(DisasContext *s, int rd, int rn, int opcode, tcg_temp_free_i32(tcg_single); } - gen_helper_set_rmode(tcg_rmode, tcg_rmode, cpu_env); + gen_helper_set_rmode(tcg_rmode, tcg_rmode, tcg_fpstatus); tcg_temp_free_i32(tcg_rmode); if (!sf) { @@ -XXX,XX +XXX,XX @@ static void handle_simd_shift_fpint_conv(DisasContext *s, bool is_scalar, assert(!(is_scalar && is_q)); tcg_rmode = tcg_const_i32(arm_rmode_to_sf(FPROUNDING_ZERO)); - gen_helper_set_rmode(tcg_rmode, tcg_rmode, cpu_env); tcg_fpstatus = get_fpstatus_ptr(false); + gen_helper_set_rmode(tcg_rmode, tcg_rmode, tcg_fpstatus); tcg_shift = tcg_const_i32(fracbits); if (is_double) { @@ -XXX,XX +XXX,XX @@ static void handle_simd_shift_fpint_conv(DisasContext *s, bool is_scalar, tcg_temp_free_ptr(tcg_fpstatus); tcg_temp_free_i32(tcg_shift); - gen_helper_set_rmode(tcg_rmode, tcg_rmode, cpu_env); + gen_helper_set_rmode(tcg_rmode, tcg_rmode, tcg_fpstatus); tcg_temp_free_i32(tcg_rmode); } @@ -XXX,XX +XXX,XX @@ static void disas_simd_scalar_two_reg_misc(DisasContext *s, uint32_t insn) if (is_fcvt) { tcg_rmode = tcg_const_i32(arm_rmode_to_sf(rmode)); - gen_helper_set_rmode(tcg_rmode, tcg_rmode, cpu_env); tcg_fpstatus = get_fpstatus_ptr(false); + gen_helper_set_rmode(tcg_rmode, tcg_rmode, tcg_fpstatus); } else { tcg_rmode = NULL; tcg_fpstatus = NULL; @@ -XXX,XX +XXX,XX @@ static void disas_simd_scalar_two_reg_misc(DisasContext *s, uint32_t insn) } if (is_fcvt) { - gen_helper_set_rmode(tcg_rmode, tcg_rmode, cpu_env); + gen_helper_set_rmode(tcg_rmode, tcg_rmode, tcg_fpstatus); tcg_temp_free_i32(tcg_rmode); tcg_temp_free_ptr(tcg_fpstatus); } @@ -XXX,XX +XXX,XX @@ static void disas_simd_two_reg_misc(DisasContext *s, uint32_t insn) return; } - if (need_fpstatus) { + if (need_fpstatus || need_rmode) { tcg_fpstatus = get_fpstatus_ptr(false); } else { tcg_fpstatus = NULL; } if (need_rmode) { tcg_rmode = tcg_const_i32(arm_rmode_to_sf(rmode)); - gen_helper_set_rmode(tcg_rmode, tcg_rmode, cpu_env); + gen_helper_set_rmode(tcg_rmode, tcg_rmode, tcg_fpstatus); } else { tcg_rmode = NULL; } @@ -XXX,XX +XXX,XX @@ static void disas_simd_two_reg_misc(DisasContext *s, uint32_t insn) clear_vec_high(s, is_q, rd); if (need_rmode) { - gen_helper_set_rmode(tcg_rmode, tcg_rmode, cpu_env); + gen_helper_set_rmode(tcg_rmode, tcg_rmode, tcg_fpstatus); tcg_temp_free_i32(tcg_rmode); } if (need_fpstatus) { 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 int handle_vrint(uint32_t insn, uint32_t rd, uint32_t rm, uint32_t dp, TCGv_i32 tcg_rmode; tcg_rmode = tcg_const_i32(arm_rmode_to_sf(rounding)); - gen_helper_set_rmode(tcg_rmode, tcg_rmode, cpu_env); + gen_helper_set_rmode(tcg_rmode, tcg_rmode, fpst); if (dp) { TCGv_i64 tcg_op; @@ -XXX,XX +XXX,XX @@ static int handle_vrint(uint32_t insn, uint32_t rd, uint32_t rm, uint32_t dp, tcg_temp_free_i32(tcg_res); } - gen_helper_set_rmode(tcg_rmode, tcg_rmode, cpu_env); + gen_helper_set_rmode(tcg_rmode, tcg_rmode, fpst); tcg_temp_free_i32(tcg_rmode); tcg_temp_free_ptr(fpst); @@ -XXX,XX +XXX,XX @@ static int handle_vcvt(uint32_t insn, uint32_t rd, uint32_t rm, uint32_t dp, tcg_shift = tcg_const_i32(0); tcg_rmode = tcg_const_i32(arm_rmode_to_sf(rounding)); - gen_helper_set_rmode(tcg_rmode, tcg_rmode, cpu_env); + gen_helper_set_rmode(tcg_rmode, tcg_rmode, fpst); if (dp) { TCGv_i64 tcg_double, tcg_res; @@ -XXX,XX +XXX,XX @@ static int handle_vcvt(uint32_t insn, uint32_t rd, uint32_t rm, uint32_t dp, tcg_temp_free_i32(tcg_single); } - gen_helper_set_rmode(tcg_rmode, tcg_rmode, cpu_env); + gen_helper_set_rmode(tcg_rmode, tcg_rmode, fpst); tcg_temp_free_i32(tcg_rmode); tcg_temp_free_i32(tcg_shift); @@ -XXX,XX +XXX,XX @@ static int disas_vfp_insn(DisasContext *s, uint32_t insn) TCGv_ptr fpst = get_fpstatus_ptr(0); TCGv_i32 tcg_rmode; tcg_rmode = tcg_const_i32(float_round_to_zero); - gen_helper_set_rmode(tcg_rmode, tcg_rmode, cpu_env); + gen_helper_set_rmode(tcg_rmode, tcg_rmode, fpst); if (dp) { gen_helper_rintd(cpu_F0d, cpu_F0d, fpst); } else { gen_helper_rints(cpu_F0s, cpu_F0s, fpst); } - gen_helper_set_rmode(tcg_rmode, tcg_rmode, cpu_env); + gen_helper_set_rmode(tcg_rmode, tcg_rmode, fpst); tcg_temp_free_i32(tcg_rmode); tcg_temp_free_ptr(fpst); break; -- 2.16.2 From: Alex Bennée <alex.bennee@linaro.org> This implements the half-precision variants of the across vector reduction operations. This involves a re-factor of the reduction code which more closely matches the ARM ARM order (and handles 8 element reductions). Signed-off-by: Alex Bennée <alex.bennee@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20180227143852.11175-7-alex.bennee@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/helper-a64.h | 4 ++ target/arm/helper-a64.c | 18 ++++++ target/arm/translate-a64.c | 140 ++++++++++++++++++++++++++++----------------- 3 files changed, 109 insertions(+), 53 deletions(-) diff --git a/target/arm/helper-a64.h b/target/arm/helper-a64.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper-a64.h +++ b/target/arm/helper-a64.h @@ -XXX,XX +XXX,XX @@ DEF_HELPER_FLAGS_4(paired_cmpxchg64_le_parallel, TCG_CALL_NO_WG, DEF_HELPER_FLAGS_4(paired_cmpxchg64_be, TCG_CALL_NO_WG, i64, env, i64, i64, i64) DEF_HELPER_FLAGS_4(paired_cmpxchg64_be_parallel, TCG_CALL_NO_WG, i64, env, i64, i64, i64) +DEF_HELPER_FLAGS_3(advsimd_maxh, TCG_CALL_NO_RWG, f16, f16, f16, ptr) +DEF_HELPER_FLAGS_3(advsimd_minh, TCG_CALL_NO_RWG, f16, f16, f16, ptr) +DEF_HELPER_FLAGS_3(advsimd_maxnumh, TCG_CALL_NO_RWG, f16, f16, f16, ptr) +DEF_HELPER_FLAGS_3(advsimd_minnumh, TCG_CALL_NO_RWG, f16, f16, f16, ptr) diff --git a/target/arm/helper-a64.c b/target/arm/helper-a64.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper-a64.c +++ b/target/arm/helper-a64.c @@ -XXX,XX +XXX,XX @@ uint64_t HELPER(paired_cmpxchg64_be_parallel)(CPUARMState *env, uint64_t addr, { return do_paired_cmpxchg64_be(env, addr, new_lo, new_hi, true, GETPC()); } + +/* + * AdvSIMD half-precision + */ + +#define ADVSIMD_HELPER(name, suffix) HELPER(glue(glue(advsimd_, name), suffix)) + +#define ADVSIMD_HALFOP(name) \ +float16 ADVSIMD_HELPER(name, h)(float16 a, float16 b, void *fpstp) \ +{ \ + float_status *fpst = fpstp; \ + return float16_ ## name(a, b, fpst); \ +} + +ADVSIMD_HALFOP(min) +ADVSIMD_HALFOP(max) +ADVSIMD_HALFOP(minnum) +ADVSIMD_HALFOP(maxnum) diff --git a/target/arm/translate-a64.c b/target/arm/translate-a64.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-a64.c +++ b/target/arm/translate-a64.c @@ -XXX,XX +XXX,XX @@ static void disas_simd_zip_trn(DisasContext *s, uint32_t insn) tcg_temp_free_i64(tcg_resh); } -static void do_minmaxop(DisasContext *s, TCGv_i32 tcg_elt1, TCGv_i32 tcg_elt2, - int opc, bool is_min, TCGv_ptr fpst) +/* + * do_reduction_op helper + * + * This mirrors the Reduce() pseudocode in the ARM ARM. It is + * important for correct NaN propagation that we do these + * operations in exactly the order specified by the pseudocode. + * + * This is a recursive function, TCG temps should be freed by the + * calling function once it is done with the values. + */ +static TCGv_i32 do_reduction_op(DisasContext *s, int fpopcode, int rn, + int esize, int size, int vmap, TCGv_ptr fpst) { - /* Helper function for disas_simd_across_lanes: do a single precision - * min/max operation on the specified two inputs, - * and return the result in tcg_elt1. - */ - if (opc == 0xc) { - if (is_min) { - gen_helper_vfp_minnums(tcg_elt1, tcg_elt1, tcg_elt2, fpst); - } else { - gen_helper_vfp_maxnums(tcg_elt1, tcg_elt1, tcg_elt2, fpst); - } + if (esize == size) { + int element; + TCGMemOp msize = esize == 16 ? MO_16 : MO_32; + TCGv_i32 tcg_elem; + + /* We should have one register left here */ + assert(ctpop8(vmap) == 1); + element = ctz32(vmap); + assert(element < 8); + + tcg_elem = tcg_temp_new_i32(); + read_vec_element_i32(s, tcg_elem, rn, element, msize); + return tcg_elem; } else { - assert(opc == 0xf); - if (is_min) { - gen_helper_vfp_mins(tcg_elt1, tcg_elt1, tcg_elt2, fpst); - } else { - gen_helper_vfp_maxs(tcg_elt1, tcg_elt1, tcg_elt2, fpst); + int bits = size / 2; + int shift = ctpop8(vmap) / 2; + int vmap_lo = (vmap >> shift) & vmap; + int vmap_hi = (vmap & ~vmap_lo); + TCGv_i32 tcg_hi, tcg_lo, tcg_res; + + tcg_hi = do_reduction_op(s, fpopcode, rn, esize, bits, vmap_hi, fpst); + tcg_lo = do_reduction_op(s, fpopcode, rn, esize, bits, vmap_lo, fpst); + tcg_res = tcg_temp_new_i32(); + + switch (fpopcode) { + case 0x0c: /* fmaxnmv half-precision */ + gen_helper_advsimd_maxnumh(tcg_res, tcg_lo, tcg_hi, fpst); + break; + case 0x0f: /* fmaxv half-precision */ + gen_helper_advsimd_maxh(tcg_res, tcg_lo, tcg_hi, fpst); + break; + case 0x1c: /* fminnmv half-precision */ + gen_helper_advsimd_minnumh(tcg_res, tcg_lo, tcg_hi, fpst); + break; + case 0x1f: /* fminv half-precision */ + gen_helper_advsimd_minh(tcg_res, tcg_lo, tcg_hi, fpst); + break; + case 0x2c: /* fmaxnmv */ + gen_helper_vfp_maxnums(tcg_res, tcg_lo, tcg_hi, fpst); + break; + case 0x2f: /* fmaxv */ + gen_helper_vfp_maxs(tcg_res, tcg_lo, tcg_hi, fpst); + break; + case 0x3c: /* fminnmv */ + gen_helper_vfp_minnums(tcg_res, tcg_lo, tcg_hi, fpst); + break; + case 0x3f: /* fminv */ + gen_helper_vfp_mins(tcg_res, tcg_lo, tcg_hi, fpst); + break; + default: + g_assert_not_reached(); } + + tcg_temp_free_i32(tcg_hi); + tcg_temp_free_i32(tcg_lo); + return tcg_res; } } @@ -XXX,XX +XXX,XX @@ static void disas_simd_across_lanes(DisasContext *s, uint32_t insn) break; case 0xc: /* FMAXNMV, FMINNMV */ case 0xf: /* FMAXV, FMINV */ - if (!is_u || !is_q || extract32(size, 0, 1)) { - unallocated_encoding(s); - return; - } - /* Bit 1 of size field encodes min vs max, and actual size is always - * 32 bits: adjust the size variable so following code can rely on it + /* Bit 1 of size field encodes min vs max and the actual size + * depends on the encoding of the U bit. If not set (and FP16 + * enabled) then we do half-precision float instead of single + * precision. */ is_min = extract32(size, 1, 1); is_fp = true; - size = 2; + if (!is_u && arm_dc_feature(s, ARM_FEATURE_V8_FP16)) { + size = 1; + } else if (!is_u || !is_q || extract32(size, 0, 1)) { + unallocated_encoding(s); + return; + } else { + size = 2; + } break; default: unallocated_encoding(s); @@ -XXX,XX +XXX,XX @@ static void disas_simd_across_lanes(DisasContext *s, uint32_t insn) } } else { - /* Floating point ops which work on 32 bit (single) intermediates. + /* Floating point vector reduction ops which work across 32 + * bit (single) or 16 bit (half-precision) intermediates. * Note that correct NaN propagation requires that we do these * operations in exactly the order specified by the pseudocode. */ - TCGv_i32 tcg_elt1 = tcg_temp_new_i32(); - TCGv_i32 tcg_elt2 = tcg_temp_new_i32(); - TCGv_i32 tcg_elt3 = tcg_temp_new_i32(); - TCGv_ptr fpst = get_fpstatus_ptr(false); - - assert(esize == 32); - assert(elements == 4); - - read_vec_element(s, tcg_elt, rn, 0, MO_32); - tcg_gen_extrl_i64_i32(tcg_elt1, tcg_elt); - read_vec_element(s, tcg_elt, rn, 1, MO_32); - tcg_gen_extrl_i64_i32(tcg_elt2, tcg_elt); - - do_minmaxop(s, tcg_elt1, tcg_elt2, opcode, is_min, fpst); - - read_vec_element(s, tcg_elt, rn, 2, MO_32); - tcg_gen_extrl_i64_i32(tcg_elt2, tcg_elt); - read_vec_element(s, tcg_elt, rn, 3, MO_32); - tcg_gen_extrl_i64_i32(tcg_elt3, tcg_elt); - - do_minmaxop(s, tcg_elt2, tcg_elt3, opcode, is_min, fpst); - - do_minmaxop(s, tcg_elt1, tcg_elt2, opcode, is_min, fpst); - - tcg_gen_extu_i32_i64(tcg_res, tcg_elt1); - tcg_temp_free_i32(tcg_elt1); - tcg_temp_free_i32(tcg_elt2); - tcg_temp_free_i32(tcg_elt3); + TCGv_ptr fpst = get_fpstatus_ptr(size == MO_16); + int fpopcode = opcode | is_min << 4 | is_u << 5; + int vmap = (1 << elements) - 1; + TCGv_i32 tcg_res32 = do_reduction_op(s, fpopcode, rn, esize, + (is_q ? 128 : 64), vmap, fpst); + tcg_gen_extu_i32_i64(tcg_res, tcg_res32); + tcg_temp_free_i32(tcg_res32); tcg_temp_free_ptr(fpst); } -- 2.16.2 From: Alex Bennée <alex.bennee@linaro.org> We do implement all the opcodes. Signed-off-by: Alex Bennée <alex.bennee@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20180227143852.11175-8-alex.bennee@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/translate-a64.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) diff --git a/target/arm/translate-a64.c b/target/arm/translate-a64.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-a64.c +++ b/target/arm/translate-a64.c @@ -XXX,XX +XXX,XX @@ static void handle_3same_64(DisasContext *s, int opcode, bool u, /* Handle 64x64->64 opcodes which are shared between the scalar * and vector 3-same groups. We cover every opcode where size == 3 * is valid in either the three-reg-same (integer, not pairwise) - * or scalar-three-reg-same groups. (Some opcodes are not yet - * implemented.) + * or scalar-three-reg-same groups. */ TCGCond cond; -- 2.16.2 From: Alex Bennée <alex.bennee@linaro.org> This is the initial decode skeleton for the Advanced SIMD three same instruction group. The fprintf is purely to aid debugging as the additional instructions are added. It will be removed once the group is complete. Signed-off-by: Alex Bennée <alex.bennee@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20180227143852.11175-9-alex.bennee@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/translate-a64.c | 73 ++++++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 73 insertions(+) diff --git a/target/arm/translate-a64.c b/target/arm/translate-a64.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-a64.c +++ b/target/arm/translate-a64.c @@ -XXX,XX +XXX,XX @@ static void disas_simd_three_reg_same(DisasContext *s, uint32_t insn) } } +/* + * Advanced SIMD three same (ARMv8.2 FP16 variants) + * + * 31 30 29 28 24 23 22 21 20 16 15 14 13 11 10 9 5 4 0 + * +---+---+---+-----------+---------+------+-----+--------+---+------+------+ + * | 0 | Q | U | 0 1 1 1 0 | a | 1 0 | Rm | 0 0 | opcode | 1 | Rn | Rd | + * +---+---+---+-----------+---------+------+-----+--------+---+------+------+ + * + * This includes FMULX, FCMEQ (register), FRECPS, FRSQRTS, FCMGE + * (register), FACGE, FABD, FCMGT (register) and FACGT. + * + */ +static void disas_simd_three_reg_same_fp16(DisasContext *s, uint32_t insn) +{ + int opcode, fpopcode; + int is_q, u, a, rm, rn, rd; + int datasize, elements; + int pass; + TCGv_ptr fpst; + + if (!arm_dc_feature(s, ARM_FEATURE_V8_FP16)) { + unallocated_encoding(s); + return; + } + + if (!fp_access_check(s)) { + return; + } + + /* For these floating point ops, the U, a and opcode bits + * together indicate the operation. + */ + opcode = extract32(insn, 11, 3); + u = extract32(insn, 29, 1); + a = extract32(insn, 23, 1); + is_q = extract32(insn, 30, 1); + rm = extract32(insn, 16, 5); + rn = extract32(insn, 5, 5); + rd = extract32(insn, 0, 5); + + fpopcode = opcode | (a << 3) | (u << 4); + datasize = is_q ? 128 : 64; + elements = datasize / 16; + + fpst = get_fpstatus_ptr(true); + + for (pass = 0; pass < elements; pass++) { + TCGv_i32 tcg_op1 = tcg_temp_new_i32(); + TCGv_i32 tcg_op2 = tcg_temp_new_i32(); + TCGv_i32 tcg_res = tcg_temp_new_i32(); + + read_vec_element_i32(s, tcg_op1, rn, pass, MO_16); + read_vec_element_i32(s, tcg_op2, rm, pass, MO_16); + + switch (fpopcode) { + default: + fprintf(stderr, "%s: insn %#04x, fpop %#2x @ %#" PRIx64 "\n", + __func__, insn, fpopcode, s->pc); + g_assert_not_reached(); + } + + write_vec_element_i32(s, tcg_res, rd, pass, MO_16); + tcg_temp_free_i32(tcg_res); + tcg_temp_free_i32(tcg_op1); + tcg_temp_free_i32(tcg_op2); + } + + tcg_temp_free_ptr(fpst); + + clear_vec_high(s, is_q, rd); +} + static void handle_2misc_widening(DisasContext *s, int opcode, bool is_q, int size, int rn, int rd) { @@ -XXX,XX +XXX,XX @@ static const AArch64DecodeTable data_proc_simd[] = { { 0xce000000, 0xff808000, disas_crypto_four_reg }, { 0xce800000, 0xffe00000, disas_crypto_xar }, { 0xce408000, 0xffe0c000, disas_crypto_three_reg_imm2 }, + { 0x0e400400, 0x9f60c400, disas_simd_three_reg_same_fp16 }, { 0x00000000, 0x00000000, NULL } }; -- 2.16.2 From: Alex Bennée <alex.bennee@linaro.org> The fprintf is only there for debugging as the skeleton is added to, it will be removed once the skeleton is complete. Signed-off-by: Alex Bennée <alex.bennee@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20180227143852.11175-10-alex.bennee@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/helper-a64.h | 4 ++++ target/arm/helper-a64.c | 4 ++++ target/arm/translate-a64.c | 28 ++++++++++++++++++++++++++++ 3 files changed, 36 insertions(+) diff --git a/target/arm/helper-a64.h b/target/arm/helper-a64.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper-a64.h +++ b/target/arm/helper-a64.h @@ -XXX,XX +XXX,XX @@ DEF_HELPER_FLAGS_3(advsimd_maxh, TCG_CALL_NO_RWG, f16, f16, f16, ptr) DEF_HELPER_FLAGS_3(advsimd_minh, TCG_CALL_NO_RWG, f16, f16, f16, ptr) DEF_HELPER_FLAGS_3(advsimd_maxnumh, TCG_CALL_NO_RWG, f16, f16, f16, ptr) DEF_HELPER_FLAGS_3(advsimd_minnumh, TCG_CALL_NO_RWG, f16, f16, f16, ptr) +DEF_HELPER_3(advsimd_addh, f16, f16, f16, ptr) +DEF_HELPER_3(advsimd_subh, f16, f16, f16, ptr) +DEF_HELPER_3(advsimd_mulh, f16, f16, f16, ptr) +DEF_HELPER_3(advsimd_divh, f16, f16, f16, ptr) diff --git a/target/arm/helper-a64.c b/target/arm/helper-a64.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper-a64.c +++ b/target/arm/helper-a64.c @@ -XXX,XX +XXX,XX @@ float16 ADVSIMD_HELPER(name, h)(float16 a, float16 b, void *fpstp) \ return float16_ ## name(a, b, fpst); \ } +ADVSIMD_HALFOP(add) +ADVSIMD_HALFOP(sub) +ADVSIMD_HALFOP(mul) +ADVSIMD_HALFOP(div) ADVSIMD_HALFOP(min) ADVSIMD_HALFOP(max) ADVSIMD_HALFOP(minnum) diff --git a/target/arm/translate-a64.c b/target/arm/translate-a64.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-a64.c +++ b/target/arm/translate-a64.c @@ -XXX,XX +XXX,XX @@ static void disas_simd_three_reg_same_fp16(DisasContext *s, uint32_t insn) read_vec_element_i32(s, tcg_op2, rm, pass, MO_16); switch (fpopcode) { + case 0x0: /* FMAXNM */ + gen_helper_advsimd_maxnumh(tcg_res, tcg_op1, tcg_op2, fpst); + break; + case 0x2: /* FADD */ + gen_helper_advsimd_addh(tcg_res, tcg_op1, tcg_op2, fpst); + break; + case 0x6: /* FMAX */ + gen_helper_advsimd_maxh(tcg_res, tcg_op1, tcg_op2, fpst); + break; + case 0x8: /* FMINNM */ + gen_helper_advsimd_minnumh(tcg_res, tcg_op1, tcg_op2, fpst); + break; + case 0xa: /* FSUB */ + gen_helper_advsimd_subh(tcg_res, tcg_op1, tcg_op2, fpst); + break; + case 0xe: /* FMIN */ + gen_helper_advsimd_minh(tcg_res, tcg_op1, tcg_op2, fpst); + break; + case 0x13: /* FMUL */ + gen_helper_advsimd_mulh(tcg_res, tcg_op1, tcg_op2, fpst); + break; + case 0x17: /* FDIV */ + gen_helper_advsimd_divh(tcg_res, tcg_op1, tcg_op2, fpst); + break; + case 0x1a: /* FABD */ + gen_helper_advsimd_subh(tcg_res, tcg_op1, tcg_op2, fpst); + tcg_gen_andi_i32(tcg_res, tcg_res, 0x7fff); + break; default: fprintf(stderr, "%s: insn %#04x, fpop %#2x @ %#" PRIx64 "\n", __func__, insn, fpopcode, s->pc); -- 2.16.2 From: Alex Bennée <alex.bennee@linaro.org> These use the generic float16_compare functionality which in turn uses the common float_compare code from the softfloat re-factor. Signed-off-by: Alex Bennée <alex.bennee@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20180227143852.11175-11-alex.bennee@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/helper-a64.h | 5 +++++ target/arm/helper-a64.c | 49 ++++++++++++++++++++++++++++++++++++++++++++++ target/arm/translate-a64.c | 15 ++++++++++++++ 3 files changed, 69 insertions(+) diff --git a/target/arm/helper-a64.h b/target/arm/helper-a64.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper-a64.h +++ b/target/arm/helper-a64.h @@ -XXX,XX +XXX,XX @@ DEF_HELPER_3(advsimd_addh, f16, f16, f16, ptr) DEF_HELPER_3(advsimd_subh, f16, f16, f16, ptr) DEF_HELPER_3(advsimd_mulh, f16, f16, f16, ptr) DEF_HELPER_3(advsimd_divh, f16, f16, f16, ptr) +DEF_HELPER_3(advsimd_ceq_f16, i32, f16, f16, ptr) +DEF_HELPER_3(advsimd_cge_f16, i32, f16, f16, ptr) +DEF_HELPER_3(advsimd_cgt_f16, i32, f16, f16, ptr) +DEF_HELPER_3(advsimd_acge_f16, i32, f16, f16, ptr) +DEF_HELPER_3(advsimd_acgt_f16, i32, f16, f16, ptr) diff --git a/target/arm/helper-a64.c b/target/arm/helper-a64.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper-a64.c +++ b/target/arm/helper-a64.c @@ -XXX,XX +XXX,XX @@ ADVSIMD_HALFOP(min) ADVSIMD_HALFOP(max) ADVSIMD_HALFOP(minnum) ADVSIMD_HALFOP(maxnum) + +/* + * Floating point comparisons produce an integer result. Softfloat + * routines return float_relation types which we convert to the 0/-1 + * Neon requires. + */ + +#define ADVSIMD_CMPRES(test) (test) ? 0xffff : 0 + +uint32_t HELPER(advsimd_ceq_f16)(float16 a, float16 b, void *fpstp) +{ + float_status *fpst = fpstp; + int compare = float16_compare_quiet(a, b, fpst); + return ADVSIMD_CMPRES(compare == float_relation_equal); +} + +uint32_t HELPER(advsimd_cge_f16)(float16 a, float16 b, void *fpstp) +{ + float_status *fpst = fpstp; + int compare = float16_compare(a, b, fpst); + return ADVSIMD_CMPRES(compare == float_relation_greater || + compare == float_relation_equal); +} + +uint32_t HELPER(advsimd_cgt_f16)(float16 a, float16 b, void *fpstp) +{ + float_status *fpst = fpstp; + int compare = float16_compare(a, b, fpst); + return ADVSIMD_CMPRES(compare == float_relation_greater); +} + +uint32_t HELPER(advsimd_acge_f16)(float16 a, float16 b, void *fpstp) +{ + float_status *fpst = fpstp; + float16 f0 = float16_abs(a); + float16 f1 = float16_abs(b); + int compare = float16_compare(f0, f1, fpst); + return ADVSIMD_CMPRES(compare == float_relation_greater || + compare == float_relation_equal); +} + +uint32_t HELPER(advsimd_acgt_f16)(float16 a, float16 b, void *fpstp) +{ + float_status *fpst = fpstp; + float16 f0 = float16_abs(a); + float16 f1 = float16_abs(b); + int compare = float16_compare(f0, f1, fpst); + return ADVSIMD_CMPRES(compare == float_relation_greater); +} diff --git a/target/arm/translate-a64.c b/target/arm/translate-a64.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-a64.c +++ b/target/arm/translate-a64.c @@ -XXX,XX +XXX,XX @@ static void disas_simd_three_reg_same_fp16(DisasContext *s, uint32_t insn) case 0x2: /* FADD */ gen_helper_advsimd_addh(tcg_res, tcg_op1, tcg_op2, fpst); break; + case 0x4: /* FCMEQ */ + gen_helper_advsimd_ceq_f16(tcg_res, tcg_op1, tcg_op2, fpst); + break; case 0x6: /* FMAX */ gen_helper_advsimd_maxh(tcg_res, tcg_op1, tcg_op2, fpst); break; @@ -XXX,XX +XXX,XX @@ static void disas_simd_three_reg_same_fp16(DisasContext *s, uint32_t insn) case 0x13: /* FMUL */ gen_helper_advsimd_mulh(tcg_res, tcg_op1, tcg_op2, fpst); break; + case 0x14: /* FCMGE */ + gen_helper_advsimd_cge_f16(tcg_res, tcg_op1, tcg_op2, fpst); + break; + case 0x15: /* FACGE */ + gen_helper_advsimd_acge_f16(tcg_res, tcg_op1, tcg_op2, fpst); + break; case 0x17: /* FDIV */ gen_helper_advsimd_divh(tcg_res, tcg_op1, tcg_op2, fpst); break; @@ -XXX,XX +XXX,XX @@ static void disas_simd_three_reg_same_fp16(DisasContext *s, uint32_t insn) gen_helper_advsimd_subh(tcg_res, tcg_op1, tcg_op2, fpst); tcg_gen_andi_i32(tcg_res, tcg_res, 0x7fff); break; + case 0x1c: /* FCMGT */ + gen_helper_advsimd_cgt_f16(tcg_res, tcg_op1, tcg_op2, fpst); + break; + case 0x1d: /* FACGT */ + gen_helper_advsimd_acgt_f16(tcg_res, tcg_op1, tcg_op2, fpst); + break; default: fprintf(stderr, "%s: insn %#04x, fpop %#2x @ %#" PRIx64 "\n", __func__, insn, fpopcode, s->pc); -- 2.16.2 From: Alex Bennée <alex.bennee@linaro.org> Signed-off-by: Alex Bennée <alex.bennee@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20180227143852.11175-12-alex.bennee@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/helper-a64.h | 2 ++ target/arm/helper-a64.c | 24 ++++++++++++++++++++++++ target/arm/translate-a64.c | 15 +++++++++++++++ 3 files changed, 41 insertions(+) diff --git a/target/arm/helper-a64.h b/target/arm/helper-a64.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper-a64.h +++ b/target/arm/helper-a64.h @@ -XXX,XX +XXX,XX @@ DEF_HELPER_3(advsimd_cge_f16, i32, f16, f16, ptr) DEF_HELPER_3(advsimd_cgt_f16, i32, f16, f16, ptr) DEF_HELPER_3(advsimd_acge_f16, i32, f16, f16, ptr) DEF_HELPER_3(advsimd_acgt_f16, i32, f16, f16, ptr) +DEF_HELPER_3(advsimd_mulxh, f16, f16, f16, ptr) +DEF_HELPER_4(advsimd_muladdh, f16, f16, f16, f16, ptr) diff --git a/target/arm/helper-a64.c b/target/arm/helper-a64.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper-a64.c +++ b/target/arm/helper-a64.c @@ -XXX,XX +XXX,XX @@ ADVSIMD_HALFOP(max) ADVSIMD_HALFOP(minnum) ADVSIMD_HALFOP(maxnum) +/* Data processing - scalar floating-point and advanced SIMD */ +float16 HELPER(advsimd_mulxh)(float16 a, float16 b, void *fpstp) +{ + float_status *fpst = fpstp; + + a = float16_squash_input_denormal(a, fpst); + b = float16_squash_input_denormal(b, fpst); + + if ((float16_is_zero(a) && float16_is_infinity(b)) || + (float16_is_infinity(a) && float16_is_zero(b))) { + /* 2.0 with the sign bit set to sign(A) XOR sign(B) */ + return make_float16((1U << 14) | + ((float16_val(a) ^ float16_val(b)) & (1U << 15))); + } + return float16_mul(a, b, fpst); +} + +/* fused multiply-accumulate */ +float16 HELPER(advsimd_muladdh)(float16 a, float16 b, float16 c, void *fpstp) +{ + float_status *fpst = fpstp; + return float16_muladd(a, b, c, 0, fpst); +} + /* * Floating point comparisons produce an integer result. Softfloat * routines return float_relation types which we convert to the 0/-1 diff --git a/target/arm/translate-a64.c b/target/arm/translate-a64.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-a64.c +++ b/target/arm/translate-a64.c @@ -XXX,XX +XXX,XX @@ static void disas_simd_three_reg_same_fp16(DisasContext *s, uint32_t insn) case 0x0: /* FMAXNM */ gen_helper_advsimd_maxnumh(tcg_res, tcg_op1, tcg_op2, fpst); break; + case 0x1: /* FMLA */ + read_vec_element_i32(s, tcg_res, rd, pass, MO_16); + gen_helper_advsimd_muladdh(tcg_res, tcg_op1, tcg_op2, tcg_res, + fpst); + break; case 0x2: /* FADD */ gen_helper_advsimd_addh(tcg_res, tcg_op1, tcg_op2, fpst); break; + case 0x3: /* FMULX */ + gen_helper_advsimd_mulxh(tcg_res, tcg_op1, tcg_op2, fpst); + break; case 0x4: /* FCMEQ */ gen_helper_advsimd_ceq_f16(tcg_res, tcg_op1, tcg_op2, fpst); break; @@ -XXX,XX +XXX,XX @@ static void disas_simd_three_reg_same_fp16(DisasContext *s, uint32_t insn) case 0x8: /* FMINNM */ gen_helper_advsimd_minnumh(tcg_res, tcg_op1, tcg_op2, fpst); break; + case 0x9: /* FMLS */ + /* As usual for ARM, separate negation for fused multiply-add */ + tcg_gen_xori_i32(tcg_op1, tcg_op1, 0x8000); + read_vec_element_i32(s, tcg_res, rd, pass, MO_16); + gen_helper_advsimd_muladdh(tcg_res, tcg_op1, tcg_op2, tcg_res, + fpst); + break; case 0xa: /* FSUB */ gen_helper_advsimd_subh(tcg_res, tcg_op1, tcg_op2, fpst); break; -- 2.16.2 From: Alex Bennée <alex.bennee@linaro.org> As some of the constants here will also be needed elsewhere (specifically for the upcoming SVE support) we move them out to softfloat.h. Signed-off-by: Alex Bennée <alex.bennee@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20180227143852.11175-13-alex.bennee@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- include/fpu/softfloat.h | 18 +++++++++++++----- target/arm/helper-a64.h | 2 ++ target/arm/helper-a64.c | 34 ++++++++++++++++++++++++++++++++++ target/arm/translate-a64.c | 6 ++++++ 4 files changed, 55 insertions(+), 5 deletions(-) diff --git a/include/fpu/softfloat.h b/include/fpu/softfloat.h index XXXXXXX..XXXXXXX 100644 --- a/include/fpu/softfloat.h +++ b/include/fpu/softfloat.h @@ -XXX,XX +XXX,XX @@ static inline float16 float16_set_sign(float16 a, int sign) } #define float16_zero make_float16(0) -#define float16_one make_float16(0x3c00) #define float16_half make_float16(0x3800) +#define float16_one make_float16(0x3c00) +#define float16_one_point_five make_float16(0x3e00) +#define float16_two make_float16(0x4000) +#define float16_three make_float16(0x4200) #define float16_infinity make_float16(0x7c00) /*---------------------------------------------------------------------------- @@ -XXX,XX +XXX,XX @@ static inline float32 float32_set_sign(float32 a, int sign) } #define float32_zero make_float32(0) -#define float32_one make_float32(0x3f800000) #define float32_half make_float32(0x3f000000) +#define float32_one make_float32(0x3f800000) +#define float32_one_point_five make_float32(0x3fc00000) +#define float32_two make_float32(0x40000000) +#define float32_three make_float32(0x40400000) #define float32_infinity make_float32(0x7f800000) - /*---------------------------------------------------------------------------- | The pattern for a default generated single-precision NaN. *----------------------------------------------------------------------------*/ @@ -XXX,XX +XXX,XX @@ static inline float64 float64_set_sign(float64 a, int sign) } #define float64_zero make_float64(0) -#define float64_one make_float64(0x3ff0000000000000LL) -#define float64_ln2 make_float64(0x3fe62e42fefa39efLL) #define float64_half make_float64(0x3fe0000000000000LL) +#define float64_one make_float64(0x3ff0000000000000LL) +#define float64_one_point_five make_float64(0x3FF8000000000000ULL) +#define float64_two make_float64(0x4000000000000000ULL) +#define float64_three make_float64(0x4008000000000000ULL) +#define float64_ln2 make_float64(0x3fe62e42fefa39efLL) #define float64_infinity make_float64(0x7ff0000000000000LL) /*---------------------------------------------------------------------------- diff --git a/target/arm/helper-a64.h b/target/arm/helper-a64.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper-a64.h +++ b/target/arm/helper-a64.h @@ -XXX,XX +XXX,XX @@ DEF_HELPER_FLAGS_3(vfp_mulxd, TCG_CALL_NO_RWG, f64, f64, f64, ptr) DEF_HELPER_FLAGS_3(neon_ceq_f64, TCG_CALL_NO_RWG, i64, i64, i64, ptr) DEF_HELPER_FLAGS_3(neon_cge_f64, TCG_CALL_NO_RWG, i64, i64, i64, ptr) DEF_HELPER_FLAGS_3(neon_cgt_f64, TCG_CALL_NO_RWG, i64, i64, i64, ptr) +DEF_HELPER_FLAGS_3(recpsf_f16, TCG_CALL_NO_RWG, f16, f16, f16, ptr) DEF_HELPER_FLAGS_3(recpsf_f32, TCG_CALL_NO_RWG, f32, f32, f32, ptr) DEF_HELPER_FLAGS_3(recpsf_f64, TCG_CALL_NO_RWG, f64, f64, f64, ptr) +DEF_HELPER_FLAGS_3(rsqrtsf_f16, TCG_CALL_NO_RWG, f16, f16, f16, ptr) DEF_HELPER_FLAGS_3(rsqrtsf_f32, TCG_CALL_NO_RWG, f32, f32, f32, ptr) DEF_HELPER_FLAGS_3(rsqrtsf_f64, TCG_CALL_NO_RWG, f64, f64, f64, ptr) DEF_HELPER_FLAGS_1(neon_addlp_s8, TCG_CALL_NO_RWG_SE, i64, i64) diff --git a/target/arm/helper-a64.c b/target/arm/helper-a64.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper-a64.c +++ b/target/arm/helper-a64.c @@ -XXX,XX +XXX,XX @@ uint64_t HELPER(neon_cgt_f64)(float64 a, float64 b, void *fpstp) * versions, these do a fully fused multiply-add or * multiply-add-and-halve. */ +#define float16_two make_float16(0x4000) +#define float16_three make_float16(0x4200) +#define float16_one_point_five make_float16(0x3e00) + #define float32_two make_float32(0x40000000) #define float32_three make_float32(0x40400000) #define float32_one_point_five make_float32(0x3fc00000) @@ -XXX,XX +XXX,XX @@ uint64_t HELPER(neon_cgt_f64)(float64 a, float64 b, void *fpstp) #define float64_three make_float64(0x4008000000000000ULL) #define float64_one_point_five make_float64(0x3FF8000000000000ULL) +float16 HELPER(recpsf_f16)(float16 a, float16 b, void *fpstp) +{ + float_status *fpst = fpstp; + + a = float16_squash_input_denormal(a, fpst); + b = float16_squash_input_denormal(b, fpst); + + a = float16_chs(a); + if ((float16_is_infinity(a) && float16_is_zero(b)) || + (float16_is_infinity(b) && float16_is_zero(a))) { + return float16_two; + } + return float16_muladd(a, b, float16_two, 0, fpst); +} + float32 HELPER(recpsf_f32)(float32 a, float32 b, void *fpstp) { float_status *fpst = fpstp; @@ -XXX,XX +XXX,XX @@ float64 HELPER(recpsf_f64)(float64 a, float64 b, void *fpstp) return float64_muladd(a, b, float64_two, 0, fpst); } +float16 HELPER(rsqrtsf_f16)(float16 a, float16 b, void *fpstp) +{ + float_status *fpst = fpstp; + + a = float16_squash_input_denormal(a, fpst); + b = float16_squash_input_denormal(b, fpst); + + a = float16_chs(a); + if ((float16_is_infinity(a) && float16_is_zero(b)) || + (float16_is_infinity(b) && float16_is_zero(a))) { + return float16_one_point_five; + } + return float16_muladd(a, b, float16_three, float_muladd_halve_result, fpst); +} + float32 HELPER(rsqrtsf_f32)(float32 a, float32 b, void *fpstp) { float_status *fpst = fpstp; diff --git a/target/arm/translate-a64.c b/target/arm/translate-a64.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-a64.c +++ b/target/arm/translate-a64.c @@ -XXX,XX +XXX,XX @@ static void disas_simd_three_reg_same_fp16(DisasContext *s, uint32_t insn) case 0x6: /* FMAX */ gen_helper_advsimd_maxh(tcg_res, tcg_op1, tcg_op2, fpst); break; + case 0x7: /* FRECPS */ + gen_helper_recpsf_f16(tcg_res, tcg_op1, tcg_op2, fpst); + break; case 0x8: /* FMINNM */ gen_helper_advsimd_minnumh(tcg_res, tcg_op1, tcg_op2, fpst); break; @@ -XXX,XX +XXX,XX @@ static void disas_simd_three_reg_same_fp16(DisasContext *s, uint32_t insn) case 0xe: /* FMIN */ gen_helper_advsimd_minh(tcg_res, tcg_op1, tcg_op2, fpst); break; + case 0xf: /* FRSQRTS */ + gen_helper_rsqrtsf_f16(tcg_res, tcg_op1, tcg_op2, fpst); + break; case 0x13: /* FMUL */ gen_helper_advsimd_mulh(tcg_res, tcg_op1, tcg_op2, fpst); break; -- 2.16.2 From: Alex Bennée <alex.bennee@linaro.org> This includes FMAXNMP, FADDP, FMAXP, FMINNMP, FMINP. Signed-off-by: Alex Bennée <alex.bennee@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20180227143852.11175-14-alex.bennee@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/translate-a64.c | 208 +++++++++++++++++++++++++++++---------------- 1 file changed, 133 insertions(+), 75 deletions(-) diff --git a/target/arm/translate-a64.c b/target/arm/translate-a64.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-a64.c +++ b/target/arm/translate-a64.c @@ -XXX,XX +XXX,XX @@ static void disas_simd_three_reg_same_fp16(DisasContext *s, uint32_t insn) int datasize, elements; int pass; TCGv_ptr fpst; + bool pairwise = false; if (!arm_dc_feature(s, ARM_FEATURE_V8_FP16)) { unallocated_encoding(s); @@ -XXX,XX +XXX,XX @@ static void disas_simd_three_reg_same_fp16(DisasContext *s, uint32_t insn) datasize = is_q ? 128 : 64; elements = datasize / 16; + switch (fpopcode) { + case 0x10: /* FMAXNMP */ + case 0x12: /* FADDP */ + case 0x16: /* FMAXP */ + case 0x18: /* FMINNMP */ + case 0x1e: /* FMINP */ + pairwise = true; + break; + } + fpst = get_fpstatus_ptr(true); - for (pass = 0; pass < elements; pass++) { + if (pairwise) { + int maxpass = is_q ? 8 : 4; TCGv_i32 tcg_op1 = tcg_temp_new_i32(); TCGv_i32 tcg_op2 = tcg_temp_new_i32(); - TCGv_i32 tcg_res = tcg_temp_new_i32(); + TCGv_i32 tcg_res[8]; - read_vec_element_i32(s, tcg_op1, rn, pass, MO_16); - read_vec_element_i32(s, tcg_op2, rm, pass, MO_16); + for (pass = 0; pass < maxpass; pass++) { + int passreg = pass < (maxpass / 2) ? rn : rm; + int passelt = (pass << 1) & (maxpass - 1); - switch (fpopcode) { - case 0x0: /* FMAXNM */ - gen_helper_advsimd_maxnumh(tcg_res, tcg_op1, tcg_op2, fpst); - break; - case 0x1: /* FMLA */ - read_vec_element_i32(s, tcg_res, rd, pass, MO_16); - gen_helper_advsimd_muladdh(tcg_res, tcg_op1, tcg_op2, tcg_res, - fpst); - break; - case 0x2: /* FADD */ - gen_helper_advsimd_addh(tcg_res, tcg_op1, tcg_op2, fpst); - break; - case 0x3: /* FMULX */ - gen_helper_advsimd_mulxh(tcg_res, tcg_op1, tcg_op2, fpst); - break; - case 0x4: /* FCMEQ */ - gen_helper_advsimd_ceq_f16(tcg_res, tcg_op1, tcg_op2, fpst); - break; - case 0x6: /* FMAX */ - gen_helper_advsimd_maxh(tcg_res, tcg_op1, tcg_op2, fpst); - break; - case 0x7: /* FRECPS */ - gen_helper_recpsf_f16(tcg_res, tcg_op1, tcg_op2, fpst); - break; - case 0x8: /* FMINNM */ - gen_helper_advsimd_minnumh(tcg_res, tcg_op1, tcg_op2, fpst); - break; - case 0x9: /* FMLS */ - /* As usual for ARM, separate negation for fused multiply-add */ - tcg_gen_xori_i32(tcg_op1, tcg_op1, 0x8000); - read_vec_element_i32(s, tcg_res, rd, pass, MO_16); - gen_helper_advsimd_muladdh(tcg_res, tcg_op1, tcg_op2, tcg_res, - fpst); - break; - case 0xa: /* FSUB */ - gen_helper_advsimd_subh(tcg_res, tcg_op1, tcg_op2, fpst); - break; - case 0xe: /* FMIN */ - gen_helper_advsimd_minh(tcg_res, tcg_op1, tcg_op2, fpst); - break; - case 0xf: /* FRSQRTS */ - gen_helper_rsqrtsf_f16(tcg_res, tcg_op1, tcg_op2, fpst); - break; - case 0x13: /* FMUL */ - gen_helper_advsimd_mulh(tcg_res, tcg_op1, tcg_op2, fpst); - break; - case 0x14: /* FCMGE */ - gen_helper_advsimd_cge_f16(tcg_res, tcg_op1, tcg_op2, fpst); - break; - case 0x15: /* FACGE */ - gen_helper_advsimd_acge_f16(tcg_res, tcg_op1, tcg_op2, fpst); - break; - case 0x17: /* FDIV */ - gen_helper_advsimd_divh(tcg_res, tcg_op1, tcg_op2, fpst); - break; - case 0x1a: /* FABD */ - gen_helper_advsimd_subh(tcg_res, tcg_op1, tcg_op2, fpst); - tcg_gen_andi_i32(tcg_res, tcg_res, 0x7fff); - break; - case 0x1c: /* FCMGT */ - gen_helper_advsimd_cgt_f16(tcg_res, tcg_op1, tcg_op2, fpst); - break; - case 0x1d: /* FACGT */ - gen_helper_advsimd_acgt_f16(tcg_res, tcg_op1, tcg_op2, fpst); - break; - default: - fprintf(stderr, "%s: insn %#04x, fpop %#2x @ %#" PRIx64 "\n", - __func__, insn, fpopcode, s->pc); - g_assert_not_reached(); + read_vec_element_i32(s, tcg_op1, passreg, passelt, MO_16); + read_vec_element_i32(s, tcg_op2, passreg, passelt + 1, MO_16); + tcg_res[pass] = tcg_temp_new_i32(); + + switch (fpopcode) { + case 0x10: /* FMAXNMP */ + gen_helper_advsimd_maxnumh(tcg_res[pass], tcg_op1, tcg_op2, + fpst); + break; + case 0x12: /* FADDP */ + gen_helper_advsimd_addh(tcg_res[pass], tcg_op1, tcg_op2, fpst); + break; + case 0x16: /* FMAXP */ + gen_helper_advsimd_maxh(tcg_res[pass], tcg_op1, tcg_op2, fpst); + break; + case 0x18: /* FMINNMP */ + gen_helper_advsimd_minnumh(tcg_res[pass], tcg_op1, tcg_op2, + fpst); + break; + case 0x1e: /* FMINP */ + gen_helper_advsimd_minh(tcg_res[pass], tcg_op1, tcg_op2, fpst); + break; + default: + g_assert_not_reached(); + } + } + + for (pass = 0; pass < maxpass; pass++) { + write_vec_element_i32(s, tcg_res[pass], rd, pass, MO_16); + tcg_temp_free_i32(tcg_res[pass]); } - write_vec_element_i32(s, tcg_res, rd, pass, MO_16); - tcg_temp_free_i32(tcg_res); tcg_temp_free_i32(tcg_op1); tcg_temp_free_i32(tcg_op2); + + } else { + for (pass = 0; pass < elements; pass++) { + TCGv_i32 tcg_op1 = tcg_temp_new_i32(); + TCGv_i32 tcg_op2 = tcg_temp_new_i32(); + TCGv_i32 tcg_res = tcg_temp_new_i32(); + + read_vec_element_i32(s, tcg_op1, rn, pass, MO_16); + read_vec_element_i32(s, tcg_op2, rm, pass, MO_16); + + switch (fpopcode) { + case 0x0: /* FMAXNM */ + gen_helper_advsimd_maxnumh(tcg_res, tcg_op1, tcg_op2, fpst); + break; + case 0x1: /* FMLA */ + read_vec_element_i32(s, tcg_res, rd, pass, MO_16); + gen_helper_advsimd_muladdh(tcg_res, tcg_op1, tcg_op2, tcg_res, + fpst); + break; + case 0x2: /* FADD */ + gen_helper_advsimd_addh(tcg_res, tcg_op1, tcg_op2, fpst); + break; + case 0x3: /* FMULX */ + gen_helper_advsimd_mulxh(tcg_res, tcg_op1, tcg_op2, fpst); + break; + case 0x4: /* FCMEQ */ + gen_helper_advsimd_ceq_f16(tcg_res, tcg_op1, tcg_op2, fpst); + break; + case 0x6: /* FMAX */ + gen_helper_advsimd_maxh(tcg_res, tcg_op1, tcg_op2, fpst); + break; + case 0x7: /* FRECPS */ + gen_helper_recpsf_f16(tcg_res, tcg_op1, tcg_op2, fpst); + break; + case 0x8: /* FMINNM */ + gen_helper_advsimd_minnumh(tcg_res, tcg_op1, tcg_op2, fpst); + break; + case 0x9: /* FMLS */ + /* As usual for ARM, separate negation for fused multiply-add */ + tcg_gen_xori_i32(tcg_op1, tcg_op1, 0x8000); + read_vec_element_i32(s, tcg_res, rd, pass, MO_16); + gen_helper_advsimd_muladdh(tcg_res, tcg_op1, tcg_op2, tcg_res, + fpst); + break; + case 0xa: /* FSUB */ + gen_helper_advsimd_subh(tcg_res, tcg_op1, tcg_op2, fpst); + break; + case 0xe: /* FMIN */ + gen_helper_advsimd_minh(tcg_res, tcg_op1, tcg_op2, fpst); + break; + case 0xf: /* FRSQRTS */ + gen_helper_rsqrtsf_f16(tcg_res, tcg_op1, tcg_op2, fpst); + break; + case 0x13: /* FMUL */ + gen_helper_advsimd_mulh(tcg_res, tcg_op1, tcg_op2, fpst); + break; + case 0x14: /* FCMGE */ + gen_helper_advsimd_cge_f16(tcg_res, tcg_op1, tcg_op2, fpst); + break; + case 0x15: /* FACGE */ + gen_helper_advsimd_acge_f16(tcg_res, tcg_op1, tcg_op2, fpst); + break; + case 0x17: /* FDIV */ + gen_helper_advsimd_divh(tcg_res, tcg_op1, tcg_op2, fpst); + break; + case 0x1a: /* FABD */ + gen_helper_advsimd_subh(tcg_res, tcg_op1, tcg_op2, fpst); + tcg_gen_andi_i32(tcg_res, tcg_res, 0x7fff); + break; + case 0x1c: /* FCMGT */ + gen_helper_advsimd_cgt_f16(tcg_res, tcg_op1, tcg_op2, fpst); + break; + case 0x1d: /* FACGT */ + gen_helper_advsimd_acgt_f16(tcg_res, tcg_op1, tcg_op2, fpst); + break; + default: + fprintf(stderr, "%s: insn %#04x, fpop %#2x @ %#" PRIx64 "\n", + __func__, insn, fpopcode, s->pc); + g_assert_not_reached(); + } + + write_vec_element_i32(s, tcg_res, rd, pass, MO_16); + tcg_temp_free_i32(tcg_res); + tcg_temp_free_i32(tcg_op1); + tcg_temp_free_i32(tcg_op2); + } } tcg_temp_free_ptr(fpst); -- 2.16.2 From: Alex Bennée <alex.bennee@linaro.org> The helpers use the new re-factored muladd support in SoftFloat for the float16 work. Signed-off-by: Alex Bennée <alex.bennee@linaro.org> Message-id: 20180227143852.11175-15-alex.bennee@linaro.org Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/translate-a64.c | 82 +++++++++++++++++++++++++++++++++++++--------- 1 file changed, 66 insertions(+), 16 deletions(-) diff --git a/target/arm/translate-a64.c b/target/arm/translate-a64.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-a64.c +++ b/target/arm/translate-a64.c @@ -XXX,XX +XXX,XX @@ static void disas_simd_indexed(DisasContext *s, uint32_t insn) int rd = extract32(insn, 0, 5); bool is_long = false; bool is_fp = false; + bool is_fp16 = false; int index; TCGv_ptr fpst; @@ -XXX,XX +XXX,XX @@ static void disas_simd_indexed(DisasContext *s, uint32_t insn) } /* fall through */ case 0x9: /* FMUL, FMULX */ - if (!extract32(size, 1, 1)) { + if (size == 1) { unallocated_encoding(s); return; } @@ -XXX,XX +XXX,XX @@ static void disas_simd_indexed(DisasContext *s, uint32_t insn) } if (is_fp) { - /* low bit of size indicates single/double */ - size = extract32(size, 0, 1) ? 3 : 2; - if (size == 2) { + /* convert insn encoded size to TCGMemOp size */ + switch (size) { + case 2: /* single precision */ + size = MO_32; index = h << 1 | l; - } else { + rm |= (m << 4); + break; + case 3: /* double precision */ + size = MO_64; if (l || !is_q) { unallocated_encoding(s); return; } index = h; + rm |= (m << 4); + break; + case 0: /* half precision */ + size = MO_16; + index = h << 2 | l << 1 | m; + is_fp16 = true; + if (arm_dc_feature(s, ARM_FEATURE_V8_FP16)) { + break; + } + /* fallthru */ + default: /* unallocated */ + unallocated_encoding(s); + return; } - rm |= (m << 4); } else { switch (size) { case 1: @@ -XXX,XX +XXX,XX @@ static void disas_simd_indexed(DisasContext *s, uint32_t insn) } if (is_fp) { - fpst = get_fpstatus_ptr(false); + fpst = get_fpstatus_ptr(is_fp16); } else { fpst = NULL; } @@ -XXX,XX +XXX,XX @@ static void disas_simd_indexed(DisasContext *s, uint32_t insn) break; } case 0x5: /* FMLS */ - /* As usual for ARM, separate negation for fused multiply-add */ - gen_helper_vfp_negs(tcg_op, tcg_op); - /* fall through */ case 0x1: /* FMLA */ - read_vec_element_i32(s, tcg_res, rd, pass, MO_32); - gen_helper_vfp_muladds(tcg_res, tcg_op, tcg_idx, tcg_res, fpst); + read_vec_element_i32(s, tcg_res, rd, pass, + is_scalar ? size : MO_32); + switch (size) { + case 1: + if (opcode == 0x5) { + /* As usual for ARM, separate negation for fused + * multiply-add */ + tcg_gen_xori_i32(tcg_op, tcg_op, 0x80008000); + } + gen_helper_advsimd_muladdh(tcg_res, tcg_op, tcg_idx, + tcg_res, fpst); + break; + case 2: + if (opcode == 0x5) { + /* As usual for ARM, separate negation for + * fused multiply-add */ + tcg_gen_xori_i32(tcg_op, tcg_op, 0x80000000); + } + gen_helper_vfp_muladds(tcg_res, tcg_op, tcg_idx, + tcg_res, fpst); + break; + default: + g_assert_not_reached(); + } break; case 0x9: /* FMUL, FMULX */ - if (u) { - gen_helper_vfp_mulxs(tcg_res, tcg_op, tcg_idx, fpst); - } else { - gen_helper_vfp_muls(tcg_res, tcg_op, tcg_idx, fpst); + switch (size) { + case 1: + if (u) { + gen_helper_advsimd_mulxh(tcg_res, tcg_op, tcg_idx, + fpst); + } else { + g_assert_not_reached(); + } + break; + case 2: + if (u) { + gen_helper_vfp_mulxs(tcg_res, tcg_op, tcg_idx, fpst); + } else { + gen_helper_vfp_muls(tcg_res, tcg_op, tcg_idx, fpst); + } + break; + default: + g_assert_not_reached(); } break; case 0xc: /* SQDMULH */ -- 2.16.2 From: Alex Bennée <alex.bennee@linaro.org> A bunch of the vectorised bitwise operations just operate on larger chunks at a time. We can do the same for the new half-precision operations by introducing some TWOHALFOP helpers which work on each half of a pair of half-precision operations at once. Hopefully all this hoop jumping will get simpler once we have generically vectorised helpers here. Signed-off-by: Alex Bennée <alex.bennee@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20180227143852.11175-16-alex.bennee@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/helper-a64.h | 10 ++++++++++ target/arm/helper-a64.c | 46 +++++++++++++++++++++++++++++++++++++++++++++- target/arm/translate-a64.c | 26 +++++++++++++++++++++----- 3 files changed, 76 insertions(+), 6 deletions(-) diff --git a/target/arm/helper-a64.h b/target/arm/helper-a64.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper-a64.h +++ b/target/arm/helper-a64.h @@ -XXX,XX +XXX,XX @@ DEF_HELPER_3(advsimd_acge_f16, i32, f16, f16, ptr) DEF_HELPER_3(advsimd_acgt_f16, i32, f16, f16, ptr) DEF_HELPER_3(advsimd_mulxh, f16, f16, f16, ptr) DEF_HELPER_4(advsimd_muladdh, f16, f16, f16, f16, ptr) +DEF_HELPER_3(advsimd_add2h, i32, i32, i32, ptr) +DEF_HELPER_3(advsimd_sub2h, i32, i32, i32, ptr) +DEF_HELPER_3(advsimd_mul2h, i32, i32, i32, ptr) +DEF_HELPER_3(advsimd_div2h, i32, i32, i32, ptr) +DEF_HELPER_3(advsimd_max2h, i32, i32, i32, ptr) +DEF_HELPER_3(advsimd_min2h, i32, i32, i32, ptr) +DEF_HELPER_3(advsimd_maxnum2h, i32, i32, i32, ptr) +DEF_HELPER_3(advsimd_minnum2h, i32, i32, i32, ptr) +DEF_HELPER_3(advsimd_mulx2h, i32, i32, i32, ptr) +DEF_HELPER_4(advsimd_muladd2h, i32, i32, i32, i32, ptr) diff --git a/target/arm/helper-a64.c b/target/arm/helper-a64.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper-a64.c +++ b/target/arm/helper-a64.c @@ -XXX,XX +XXX,XX @@ ADVSIMD_HALFOP(max) ADVSIMD_HALFOP(minnum) ADVSIMD_HALFOP(maxnum) +#define ADVSIMD_TWOHALFOP(name) \ +uint32_t ADVSIMD_HELPER(name, 2h)(uint32_t two_a, uint32_t two_b, void *fpstp) \ +{ \ + float16 a1, a2, b1, b2; \ + uint32_t r1, r2; \ + float_status *fpst = fpstp; \ + a1 = extract32(two_a, 0, 16); \ + a2 = extract32(two_a, 16, 16); \ + b1 = extract32(two_b, 0, 16); \ + b2 = extract32(two_b, 16, 16); \ + r1 = float16_ ## name(a1, b1, fpst); \ + r2 = float16_ ## name(a2, b2, fpst); \ + return deposit32(r1, 16, 16, r2); \ +} + +ADVSIMD_TWOHALFOP(add) +ADVSIMD_TWOHALFOP(sub) +ADVSIMD_TWOHALFOP(mul) +ADVSIMD_TWOHALFOP(div) +ADVSIMD_TWOHALFOP(min) +ADVSIMD_TWOHALFOP(max) +ADVSIMD_TWOHALFOP(minnum) +ADVSIMD_TWOHALFOP(maxnum) + /* Data processing - scalar floating-point and advanced SIMD */ -float16 HELPER(advsimd_mulxh)(float16 a, float16 b, void *fpstp) +static float16 float16_mulx(float16 a, float16 b, void *fpstp) { float_status *fpst = fpstp; @@ -XXX,XX +XXX,XX @@ float16 HELPER(advsimd_mulxh)(float16 a, float16 b, void *fpstp) return float16_mul(a, b, fpst); } +ADVSIMD_HALFOP(mulx) +ADVSIMD_TWOHALFOP(mulx) + /* fused multiply-accumulate */ float16 HELPER(advsimd_muladdh)(float16 a, float16 b, float16 c, void *fpstp) { @@ -XXX,XX +XXX,XX @@ float16 HELPER(advsimd_muladdh)(float16 a, float16 b, float16 c, void *fpstp) return float16_muladd(a, b, c, 0, fpst); } +uint32_t HELPER(advsimd_muladd2h)(uint32_t two_a, uint32_t two_b, + uint32_t two_c, void *fpstp) +{ + float_status *fpst = fpstp; + float16 a1, a2, b1, b2, c1, c2; + uint32_t r1, r2; + a1 = extract32(two_a, 0, 16); + a2 = extract32(two_a, 16, 16); + b1 = extract32(two_b, 0, 16); + b2 = extract32(two_b, 16, 16); + c1 = extract32(two_c, 0, 16); + c2 = extract32(two_c, 16, 16); + r1 = float16_muladd(a1, b1, c1, 0, fpst); + r2 = float16_muladd(a2, b2, c2, 0, fpst); + return deposit32(r1, 16, 16, r2); +} + /* * Floating point comparisons produce an integer result. Softfloat * routines return float_relation types which we convert to the 0/-1 diff --git a/target/arm/translate-a64.c b/target/arm/translate-a64.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-a64.c +++ b/target/arm/translate-a64.c @@ -XXX,XX +XXX,XX @@ static void disas_simd_indexed(DisasContext *s, uint32_t insn) * multiply-add */ tcg_gen_xori_i32(tcg_op, tcg_op, 0x80008000); } - gen_helper_advsimd_muladdh(tcg_res, tcg_op, tcg_idx, - tcg_res, fpst); + if (is_scalar) { + gen_helper_advsimd_muladdh(tcg_res, tcg_op, tcg_idx, + tcg_res, fpst); + } else { + gen_helper_advsimd_muladd2h(tcg_res, tcg_op, tcg_idx, + tcg_res, fpst); + } break; case 2: if (opcode == 0x5) { @@ -XXX,XX +XXX,XX @@ static void disas_simd_indexed(DisasContext *s, uint32_t insn) switch (size) { case 1: if (u) { - gen_helper_advsimd_mulxh(tcg_res, tcg_op, tcg_idx, - fpst); + if (is_scalar) { + gen_helper_advsimd_mulxh(tcg_res, tcg_op, + tcg_idx, fpst); + } else { + gen_helper_advsimd_mulx2h(tcg_res, tcg_op, + tcg_idx, fpst); + } } else { - g_assert_not_reached(); + if (is_scalar) { + gen_helper_advsimd_mulh(tcg_res, tcg_op, + tcg_idx, fpst); + } else { + gen_helper_advsimd_mul2h(tcg_res, tcg_op, + tcg_idx, fpst); + } } break; case 2: -- 2.16.2 From: Alex Bennée <alex.bennee@linaro.org> This actually covers two different sections of the encoding table: Advanced SIMD scalar two-register miscellaneous FP16 Advanced SIMD two-register miscellaneous (FP16) The difference between the two is covered by a combination of Q (bit 30) and S (bit 28). Notably the FRINTx instructions are only available in the vector form. This is just the decode skeleton which will be filled out by later patches. Signed-off-by: Alex Bennée <alex.bennee@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20180227143852.11175-17-alex.bennee@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/translate-a64.c | 40 ++++++++++++++++++++++++++++++++++++++++ 1 file changed, 40 insertions(+) diff --git a/target/arm/translate-a64.c b/target/arm/translate-a64.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-a64.c +++ b/target/arm/translate-a64.c @@ -XXX,XX +XXX,XX @@ static void disas_simd_two_reg_misc(DisasContext *s, uint32_t insn) } } +/* AdvSIMD [scalar] two register miscellaneous (FP16) + * + * 31 30 29 28 27 24 23 22 21 17 16 12 11 10 9 5 4 0 + * +---+---+---+---+---------+---+-------------+--------+-----+------+------+ + * | 0 | Q | U | S | 1 1 1 0 | a | 1 1 1 1 0 0 | opcode | 1 0 | Rn | Rd | + * +---+---+---+---+---------+---+-------------+--------+-----+------+------+ + * mask: 1000 1111 0111 1110 0000 1100 0000 0000 0x8f7e 0c00 + * val: 0000 1110 0111 1000 0000 1000 0000 0000 0x0e78 0800 + * + * This actually covers two groups where scalar access is governed by + * bit 28. A bunch of the instructions (float to integral) only exist + * in the vector form and are un-allocated for the scalar decode. Also + * in the scalar decode Q is always 1. + */ +static void disas_simd_two_reg_misc_fp16(DisasContext *s, uint32_t insn) +{ + int fpop, opcode, a; + + if (!arm_dc_feature(s, ARM_FEATURE_V8_FP16)) { + unallocated_encoding(s); + return; + } + + if (!fp_access_check(s)) { + return; + } + + opcode = extract32(insn, 12, 4); + a = extract32(insn, 23, 1); + fpop = deposit32(opcode, 5, 1, a); + + switch (fpop) { + default: + fprintf(stderr, "%s: insn %#04x fpop %#2x\n", __func__, insn, fpop); + g_assert_not_reached(); + } + +} + /* AdvSIMD scalar x indexed element * 31 30 29 28 24 23 22 21 20 19 16 15 12 11 10 9 5 4 0 * +-----+---+-----------+------+---+---+------+-----+---+---+------+------+ @@ -XXX,XX +XXX,XX @@ static const AArch64DecodeTable data_proc_simd[] = { { 0xce800000, 0xffe00000, disas_crypto_xar }, { 0xce408000, 0xffe0c000, disas_crypto_three_reg_imm2 }, { 0x0e400400, 0x9f60c400, disas_simd_three_reg_same_fp16 }, + { 0x0e780800, 0x8f7e0c00, disas_simd_two_reg_misc_fp16 }, { 0x00000000, 0x00000000, NULL } }; -- 2.16.2 From: Alex Bennée <alex.bennee@linaro.org> This adds the full range of half-precision floating point to integral instructions. Signed-off-by: Alex Bennée <alex.bennee@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20180227143852.11175-18-alex.bennee@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/helper-a64.h | 2 + target/arm/helper-a64.c | 22 ++++++++ target/arm/translate-a64.c | 123 +++++++++++++++++++++++++++++++++++++++++++-- 3 files changed, 142 insertions(+), 5 deletions(-) diff --git a/target/arm/helper-a64.h b/target/arm/helper-a64.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper-a64.h +++ b/target/arm/helper-a64.h @@ -XXX,XX +XXX,XX @@ DEF_HELPER_3(advsimd_maxnum2h, i32, i32, i32, ptr) DEF_HELPER_3(advsimd_minnum2h, i32, i32, i32, ptr) DEF_HELPER_3(advsimd_mulx2h, i32, i32, i32, ptr) DEF_HELPER_4(advsimd_muladd2h, i32, i32, i32, i32, ptr) +DEF_HELPER_2(advsimd_rinth_exact, f16, f16, ptr) +DEF_HELPER_2(advsimd_rinth, f16, f16, ptr) diff --git a/target/arm/helper-a64.c b/target/arm/helper-a64.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper-a64.c +++ b/target/arm/helper-a64.c @@ -XXX,XX +XXX,XX @@ uint32_t HELPER(advsimd_acgt_f16)(float16 a, float16 b, void *fpstp) int compare = float16_compare(f0, f1, fpst); return ADVSIMD_CMPRES(compare == float_relation_greater); } + +/* round to integral */ +float16 HELPER(advsimd_rinth_exact)(float16 x, void *fp_status) +{ + return float16_round_to_int(x, fp_status); +} + +float16 HELPER(advsimd_rinth)(float16 x, void *fp_status) +{ + int old_flags = get_float_exception_flags(fp_status), new_flags; + float16 ret; + + ret = float16_round_to_int(x, fp_status); + + /* Suppress any inexact exceptions the conversion produced */ + if (!(old_flags & float_flag_inexact)) { + new_flags = get_float_exception_flags(fp_status); + set_float_exception_flags(new_flags & ~float_flag_inexact, fp_status); + } + + return ret; +} diff --git a/target/arm/translate-a64.c b/target/arm/translate-a64.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-a64.c +++ b/target/arm/translate-a64.c @@ -XXX,XX +XXX,XX @@ static void disas_simd_two_reg_misc(DisasContext *s, uint32_t insn) */ static void disas_simd_two_reg_misc_fp16(DisasContext *s, uint32_t insn) { - int fpop, opcode, a; + int fpop, opcode, a, u; + int rn, rd; + bool is_q; + bool is_scalar; + bool only_in_vector = false; + + int pass; + TCGv_i32 tcg_rmode = NULL; + TCGv_ptr tcg_fpstatus = NULL; + bool need_rmode = false; + int rmode; if (!arm_dc_feature(s, ARM_FEATURE_V8_FP16)) { unallocated_encoding(s); return; } - if (!fp_access_check(s)) { - return; - } + rd = extract32(insn, 0, 5); + rn = extract32(insn, 5, 5); - opcode = extract32(insn, 12, 4); a = extract32(insn, 23, 1); + u = extract32(insn, 29, 1); + is_scalar = extract32(insn, 28, 1); + is_q = extract32(insn, 30, 1); + + opcode = extract32(insn, 12, 5); fpop = deposit32(opcode, 5, 1, a); + fpop = deposit32(fpop, 6, 1, u); switch (fpop) { + case 0x18: /* FRINTN */ + need_rmode = true; + only_in_vector = true; + rmode = FPROUNDING_TIEEVEN; + break; + case 0x19: /* FRINTM */ + need_rmode = true; + only_in_vector = true; + rmode = FPROUNDING_NEGINF; + break; + case 0x38: /* FRINTP */ + need_rmode = true; + only_in_vector = true; + rmode = FPROUNDING_POSINF; + break; + case 0x39: /* FRINTZ */ + need_rmode = true; + only_in_vector = true; + rmode = FPROUNDING_ZERO; + break; + case 0x58: /* FRINTA */ + need_rmode = true; + only_in_vector = true; + rmode = FPROUNDING_TIEAWAY; + break; + case 0x59: /* FRINTX */ + case 0x79: /* FRINTI */ + only_in_vector = true; + /* current rounding mode */ + break; default: fprintf(stderr, "%s: insn %#04x fpop %#2x\n", __func__, insn, fpop); g_assert_not_reached(); } + + /* Check additional constraints for the scalar encoding */ + if (is_scalar) { + if (!is_q) { + unallocated_encoding(s); + return; + } + /* FRINTxx is only in the vector form */ + if (only_in_vector) { + unallocated_encoding(s); + return; + } + } + + if (!fp_access_check(s)) { + return; + } + + if (need_rmode) { + tcg_fpstatus = get_fpstatus_ptr(true); + } + + if (need_rmode) { + tcg_rmode = tcg_const_i32(arm_rmode_to_sf(rmode)); + gen_helper_set_rmode(tcg_rmode, tcg_rmode, tcg_fpstatus); + } + + if (is_scalar) { + /* no operations yet */ + } else { + for (pass = 0; pass < (is_q ? 8 : 4); pass++) { + TCGv_i32 tcg_op = tcg_temp_new_i32(); + TCGv_i32 tcg_res = tcg_temp_new_i32(); + + read_vec_element_i32(s, tcg_op, rn, pass, MO_16); + + switch (fpop) { + case 0x18: /* FRINTN */ + case 0x19: /* FRINTM */ + case 0x38: /* FRINTP */ + case 0x39: /* FRINTZ */ + case 0x58: /* FRINTA */ + case 0x79: /* FRINTI */ + gen_helper_advsimd_rinth(tcg_res, tcg_op, tcg_fpstatus); + break; + case 0x59: /* FRINTX */ + gen_helper_advsimd_rinth_exact(tcg_res, tcg_op, tcg_fpstatus); + break; + default: + g_assert_not_reached(); + } + + write_vec_element_i32(s, tcg_res, rd, pass, MO_16); + + tcg_temp_free_i32(tcg_res); + tcg_temp_free_i32(tcg_op); + } + + clear_vec_high(s, is_q, rd); + } + + if (tcg_rmode) { + gen_helper_set_rmode(tcg_rmode, tcg_rmode, tcg_fpstatus); + tcg_temp_free_i32(tcg_rmode); + } + + if (tcg_fpstatus) { + tcg_temp_free_ptr(tcg_fpstatus); + } } /* AdvSIMD scalar x indexed element -- 2.16.2 From: Alex Bennée <alex.bennee@linaro.org> This covers all the floating point convert operations. Signed-off-by: Alex Bennée <alex.bennee@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20180227143852.11175-19-alex.bennee@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/helper-a64.h | 2 ++ target/arm/helper-a64.c | 32 +++++++++++++++++ target/arm/translate-a64.c | 85 +++++++++++++++++++++++++++++++++++++++++++++- 3 files changed, 118 insertions(+), 1 deletion(-) diff --git a/target/arm/helper-a64.h b/target/arm/helper-a64.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper-a64.h +++ b/target/arm/helper-a64.h @@ -XXX,XX +XXX,XX @@ DEF_HELPER_3(advsimd_mulx2h, i32, i32, i32, ptr) DEF_HELPER_4(advsimd_muladd2h, i32, i32, i32, i32, ptr) DEF_HELPER_2(advsimd_rinth_exact, f16, f16, ptr) DEF_HELPER_2(advsimd_rinth, f16, f16, ptr) +DEF_HELPER_2(advsimd_f16tosinth, i32, f16, ptr) +DEF_HELPER_2(advsimd_f16touinth, i32, f16, ptr) diff --git a/target/arm/helper-a64.c b/target/arm/helper-a64.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper-a64.c +++ b/target/arm/helper-a64.c @@ -XXX,XX +XXX,XX @@ float16 HELPER(advsimd_rinth)(float16 x, void *fp_status) return ret; } + +/* + * Half-precision floating point conversion functions + * + * There are a multitude of conversion functions with various + * different rounding modes. This is dealt with by the calling code + * setting the mode appropriately before calling the helper. + */ + +uint32_t HELPER(advsimd_f16tosinth)(float16 a, void *fpstp) +{ + float_status *fpst = fpstp; + + /* Invalid if we are passed a NaN */ + if (float16_is_any_nan(a)) { + float_raise(float_flag_invalid, fpst); + return 0; + } + return float16_to_int16(a, fpst); +} + +uint32_t HELPER(advsimd_f16touinth)(float16 a, void *fpstp) +{ + float_status *fpst = fpstp; + + /* Invalid if we are passed a NaN */ + if (float16_is_any_nan(a)) { + float_raise(float_flag_invalid, fpst); + return 0; + } + return float16_to_uint16(a, fpst); +} diff --git a/target/arm/translate-a64.c b/target/arm/translate-a64.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-a64.c +++ b/target/arm/translate-a64.c @@ -XXX,XX +XXX,XX @@ static void disas_simd_two_reg_misc_fp16(DisasContext *s, uint32_t insn) only_in_vector = true; /* current rounding mode */ break; + case 0x1a: /* FCVTNS */ + need_rmode = true; + rmode = FPROUNDING_TIEEVEN; + break; + case 0x1b: /* FCVTMS */ + need_rmode = true; + rmode = FPROUNDING_NEGINF; + break; + case 0x1c: /* FCVTAS */ + need_rmode = true; + rmode = FPROUNDING_TIEAWAY; + break; + case 0x3a: /* FCVTPS */ + need_rmode = true; + rmode = FPROUNDING_POSINF; + break; + case 0x3b: /* FCVTZS */ + need_rmode = true; + rmode = FPROUNDING_ZERO; + break; + case 0x5a: /* FCVTNU */ + need_rmode = true; + rmode = FPROUNDING_TIEEVEN; + break; + case 0x5b: /* FCVTMU */ + need_rmode = true; + rmode = FPROUNDING_NEGINF; + break; + case 0x5c: /* FCVTAU */ + need_rmode = true; + rmode = FPROUNDING_TIEAWAY; + break; + case 0x7a: /* FCVTPU */ + need_rmode = true; + rmode = FPROUNDING_POSINF; + break; + case 0x7b: /* FCVTZU */ + need_rmode = true; + rmode = FPROUNDING_ZERO; + break; default: fprintf(stderr, "%s: insn %#04x fpop %#2x\n", __func__, insn, fpop); g_assert_not_reached(); @@ -XXX,XX +XXX,XX @@ static void disas_simd_two_reg_misc_fp16(DisasContext *s, uint32_t insn) } if (is_scalar) { - /* no operations yet */ + TCGv_i32 tcg_op = tcg_temp_new_i32(); + TCGv_i32 tcg_res = tcg_temp_new_i32(); + + read_vec_element_i32(s, tcg_op, rn, 0, MO_16); + + switch (fpop) { + case 0x1a: /* FCVTNS */ + case 0x1b: /* FCVTMS */ + case 0x1c: /* FCVTAS */ + case 0x3a: /* FCVTPS */ + case 0x3b: /* FCVTZS */ + gen_helper_advsimd_f16tosinth(tcg_res, tcg_op, tcg_fpstatus); + break; + case 0x5a: /* FCVTNU */ + case 0x5b: /* FCVTMU */ + case 0x5c: /* FCVTAU */ + case 0x7a: /* FCVTPU */ + case 0x7b: /* FCVTZU */ + gen_helper_advsimd_f16touinth(tcg_res, tcg_op, tcg_fpstatus); + break; + default: + g_assert_not_reached(); + } + + /* limit any sign extension going on */ + tcg_gen_andi_i32(tcg_res, tcg_res, 0xffff); + write_fp_sreg(s, rd, tcg_res); + + tcg_temp_free_i32(tcg_res); + tcg_temp_free_i32(tcg_op); } else { for (pass = 0; pass < (is_q ? 8 : 4); pass++) { TCGv_i32 tcg_op = tcg_temp_new_i32(); @@ -XXX,XX +XXX,XX @@ static void disas_simd_two_reg_misc_fp16(DisasContext *s, uint32_t insn) read_vec_element_i32(s, tcg_op, rn, pass, MO_16); switch (fpop) { + case 0x1a: /* FCVTNS */ + case 0x1b: /* FCVTMS */ + case 0x1c: /* FCVTAS */ + case 0x3a: /* FCVTPS */ + case 0x3b: /* FCVTZS */ + gen_helper_advsimd_f16tosinth(tcg_res, tcg_op, tcg_fpstatus); + break; + case 0x5a: /* FCVTNU */ + case 0x5b: /* FCVTMU */ + case 0x5c: /* FCVTAU */ + case 0x7a: /* FCVTPU */ + case 0x7b: /* FCVTZU */ + gen_helper_advsimd_f16touinth(tcg_res, tcg_op, tcg_fpstatus); + break; case 0x18: /* FRINTN */ case 0x19: /* FRINTM */ case 0x38: /* FRINTP */ -- 2.16.2 From: Alex Bennée <alex.bennee@linaro.org> I re-use the existing handle_2misc_fcmp_zero handler and tweak it slightly to deal with the half-precision case. Signed-off-by: Alex Bennée <alex.bennee@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20180227143852.11175-20-alex.bennee@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/translate-a64.c | 80 +++++++++++++++++++++++++++++++++------------- 1 file changed, 57 insertions(+), 23 deletions(-) diff --git a/target/arm/translate-a64.c b/target/arm/translate-a64.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-a64.c +++ b/target/arm/translate-a64.c @@ -XXX,XX +XXX,XX @@ static void handle_2misc_fcmp_zero(DisasContext *s, int opcode, bool is_scalar, bool is_u, bool is_q, int size, int rn, int rd) { - bool is_double = (size == 3); + bool is_double = (size == MO_64); TCGv_ptr fpst; if (!fp_access_check(s)) { return; } - fpst = get_fpstatus_ptr(false); + fpst = get_fpstatus_ptr(size == MO_16); if (is_double) { TCGv_i64 tcg_op = tcg_temp_new_i64(); @@ -XXX,XX +XXX,XX @@ static void handle_2misc_fcmp_zero(DisasContext *s, int opcode, bool swap = false; int pass, maxpasses; - switch (opcode) { - case 0x2e: /* FCMLT (zero) */ - swap = true; - /* fall through */ - case 0x2c: /* FCMGT (zero) */ - genfn = gen_helper_neon_cgt_f32; - break; - case 0x2d: /* FCMEQ (zero) */ - genfn = gen_helper_neon_ceq_f32; - break; - case 0x6d: /* FCMLE (zero) */ - swap = true; - /* fall through */ - case 0x6c: /* FCMGE (zero) */ - genfn = gen_helper_neon_cge_f32; - break; - default: - g_assert_not_reached(); + if (size == MO_16) { + switch (opcode) { + case 0x2e: /* FCMLT (zero) */ + swap = true; + /* fall through */ + case 0x2c: /* FCMGT (zero) */ + genfn = gen_helper_advsimd_cgt_f16; + break; + case 0x2d: /* FCMEQ (zero) */ + genfn = gen_helper_advsimd_ceq_f16; + break; + case 0x6d: /* FCMLE (zero) */ + swap = true; + /* fall through */ + case 0x6c: /* FCMGE (zero) */ + genfn = gen_helper_advsimd_cge_f16; + break; + default: + g_assert_not_reached(); + } + } else { + switch (opcode) { + case 0x2e: /* FCMLT (zero) */ + swap = true; + /* fall through */ + case 0x2c: /* FCMGT (zero) */ + genfn = gen_helper_neon_cgt_f32; + break; + case 0x2d: /* FCMEQ (zero) */ + genfn = gen_helper_neon_ceq_f32; + break; + case 0x6d: /* FCMLE (zero) */ + swap = true; + /* fall through */ + case 0x6c: /* FCMGE (zero) */ + genfn = gen_helper_neon_cge_f32; + break; + default: + g_assert_not_reached(); + } } if (is_scalar) { maxpasses = 1; } else { - maxpasses = is_q ? 4 : 2; + int vector_size = 8 << is_q; + maxpasses = vector_size >> size; } for (pass = 0; pass < maxpasses; pass++) { - read_vec_element_i32(s, tcg_op, rn, pass, MO_32); + read_vec_element_i32(s, tcg_op, rn, pass, size); if (swap) { genfn(tcg_res, tcg_zero, tcg_op, fpst); } else { @@ -XXX,XX +XXX,XX @@ static void handle_2misc_fcmp_zero(DisasContext *s, int opcode, if (is_scalar) { write_fp_sreg(s, rd, tcg_res); } else { - write_vec_element_i32(s, tcg_res, rd, pass, MO_32); + write_vec_element_i32(s, tcg_res, rd, pass, size); } } tcg_temp_free_i32(tcg_res); @@ -XXX,XX +XXX,XX @@ static void disas_simd_two_reg_misc_fp16(DisasContext *s, uint32_t insn) fpop = deposit32(opcode, 5, 1, a); fpop = deposit32(fpop, 6, 1, u); + rd = extract32(insn, 0, 5); + rn = extract32(insn, 5, 5); + switch (fpop) { + break; + case 0x2c: /* FCMGT (zero) */ + case 0x2d: /* FCMEQ (zero) */ + case 0x2e: /* FCMLT (zero) */ + case 0x6c: /* FCMGE (zero) */ + case 0x6d: /* FCMLE (zero) */ + handle_2misc_fcmp_zero(s, fpop, is_scalar, 0, is_q, MO_16, rn, rd); + return; case 0x18: /* FRINTN */ need_rmode = true; only_in_vector = true; -- 2.16.2 From: Alex Bennée <alex.bennee@linaro.org> I've re-factored the handle_simd_intfp_conv helper to properly handle half-precision as well as call plain conversion helpers when we are not doing fixed point conversion. Signed-off-by: Alex Bennée <alex.bennee@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20180227143852.11175-21-alex.bennee@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/helper.h | 10 ++++ target/arm/helper.c | 4 ++ target/arm/translate-a64.c | 122 ++++++++++++++++++++++++++++++++++----------- 3 files changed, 108 insertions(+), 28 deletions(-) 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_3(vfp_cmped, void, f64, f64, env) DEF_HELPER_2(vfp_fcvtds, f64, f32, env) DEF_HELPER_2(vfp_fcvtsd, f32, f64, env) +DEF_HELPER_2(vfp_uitoh, f16, i32, ptr) DEF_HELPER_2(vfp_uitos, f32, i32, ptr) DEF_HELPER_2(vfp_uitod, f64, i32, ptr) +DEF_HELPER_2(vfp_sitoh, f16, i32, ptr) DEF_HELPER_2(vfp_sitos, f32, i32, ptr) DEF_HELPER_2(vfp_sitod, f64, i32, ptr) +DEF_HELPER_2(vfp_touih, i32, f16, ptr) DEF_HELPER_2(vfp_touis, i32, f32, ptr) DEF_HELPER_2(vfp_touid, i32, f64, ptr) +DEF_HELPER_2(vfp_touizh, i32, f16, ptr) DEF_HELPER_2(vfp_touizs, i32, f32, ptr) DEF_HELPER_2(vfp_touizd, i32, f64, ptr) +DEF_HELPER_2(vfp_tosih, i32, f16, ptr) DEF_HELPER_2(vfp_tosis, i32, f32, ptr) DEF_HELPER_2(vfp_tosid, i32, f64, ptr) +DEF_HELPER_2(vfp_tosizh, i32, f16, ptr) DEF_HELPER_2(vfp_tosizs, i32, f32, ptr) DEF_HELPER_2(vfp_tosizd, i32, f64, ptr) @@ -XXX,XX +XXX,XX @@ DEF_HELPER_3(vfp_toshd_round_to_zero, i64, f64, i32, ptr) DEF_HELPER_3(vfp_tosld_round_to_zero, i64, f64, i32, ptr) DEF_HELPER_3(vfp_touhd_round_to_zero, i64, f64, i32, ptr) DEF_HELPER_3(vfp_tould_round_to_zero, i64, f64, i32, ptr) +DEF_HELPER_3(vfp_toulh, i32, f16, i32, ptr) +DEF_HELPER_3(vfp_toslh, i32, f16, i32, ptr) DEF_HELPER_3(vfp_toshs, i32, f32, i32, ptr) DEF_HELPER_3(vfp_tosls, i32, f32, i32, ptr) DEF_HELPER_3(vfp_tosqs, i64, f32, i32, ptr) @@ -XXX,XX +XXX,XX @@ DEF_HELPER_3(vfp_sqtod, f64, i64, i32, ptr) DEF_HELPER_3(vfp_uhtod, f64, i64, i32, ptr) DEF_HELPER_3(vfp_ultod, f64, i64, i32, ptr) DEF_HELPER_3(vfp_uqtod, f64, i64, i32, ptr) +DEF_HELPER_3(vfp_sltoh, f16, i32, i32, ptr) +DEF_HELPER_3(vfp_ultoh, f16, i32, i32, ptr) DEF_HELPER_FLAGS_2(set_rmode, TCG_CALL_NO_RWG, i32, i32, ptr) DEF_HELPER_FLAGS_2(set_neon_rmode, TCG_CALL_NO_RWG, i32, i32, env) diff --git a/target/arm/helper.c b/target/arm/helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper.c +++ b/target/arm/helper.c @@ -XXX,XX +XXX,XX @@ CONV_ITOF(vfp_##name##to##p, fsz, sign) \ CONV_FTOI(vfp_to##name##p, fsz, sign, ) \ CONV_FTOI(vfp_to##name##z##p, fsz, sign, _round_to_zero) +FLOAT_CONVS(si, h, 16, ) FLOAT_CONVS(si, s, 32, ) FLOAT_CONVS(si, d, 64, ) +FLOAT_CONVS(ui, h, 16, u) FLOAT_CONVS(ui, s, 32, u) FLOAT_CONVS(ui, d, 64, u) @@ -XXX,XX +XXX,XX @@ VFP_CONV_FIX_A64(sq, s, 32, 64, int64) VFP_CONV_FIX(uh, s, 32, 32, uint16) VFP_CONV_FIX(ul, s, 32, 32, uint32) VFP_CONV_FIX_A64(uq, s, 32, 64, uint64) +VFP_CONV_FIX_A64(sl, h, 16, 32, int32) +VFP_CONV_FIX_A64(ul, h, 16, 32, uint32) #undef VFP_CONV_FIX #undef VFP_CONV_FIX_FLOAT #undef VFP_CONV_FLOAT_FIX_ROUND diff --git a/target/arm/translate-a64.c b/target/arm/translate-a64.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-a64.c +++ b/target/arm/translate-a64.c @@ -XXX,XX +XXX,XX @@ static void handle_simd_intfp_conv(DisasContext *s, int rd, int rn, int elements, int is_signed, int fracbits, int size) { - bool is_double = size == 3 ? true : false; - TCGv_ptr tcg_fpst = get_fpstatus_ptr(false); - TCGv_i32 tcg_shift = tcg_const_i32(fracbits); - TCGv_i64 tcg_int = tcg_temp_new_i64(); + TCGv_ptr tcg_fpst = get_fpstatus_ptr(size == MO_16); + TCGv_i32 tcg_shift = NULL; + TCGMemOp mop = size | (is_signed ? MO_SIGN : 0); int pass; - for (pass = 0; pass < elements; pass++) { - read_vec_element(s, tcg_int, rn, pass, mop); + if (fracbits || size == MO_64) { + tcg_shift = tcg_const_i32(fracbits); + } + + if (size == MO_64) { + TCGv_i64 tcg_int64 = tcg_temp_new_i64(); + TCGv_i64 tcg_double = tcg_temp_new_i64(); + + for (pass = 0; pass < elements; pass++) { + read_vec_element(s, tcg_int64, rn, pass, mop); - if (is_double) { - TCGv_i64 tcg_double = tcg_temp_new_i64(); if (is_signed) { - gen_helper_vfp_sqtod(tcg_double, tcg_int, + gen_helper_vfp_sqtod(tcg_double, tcg_int64, tcg_shift, tcg_fpst); } else { - gen_helper_vfp_uqtod(tcg_double, tcg_int, + gen_helper_vfp_uqtod(tcg_double, tcg_int64, tcg_shift, tcg_fpst); } if (elements == 1) { @@ -XXX,XX +XXX,XX @@ static void handle_simd_intfp_conv(DisasContext *s, int rd, int rn, } else { write_vec_element(s, tcg_double, rd, pass, MO_64); } - tcg_temp_free_i64(tcg_double); - } else { - TCGv_i32 tcg_single = tcg_temp_new_i32(); - if (is_signed) { - gen_helper_vfp_sqtos(tcg_single, tcg_int, - tcg_shift, tcg_fpst); - } else { - gen_helper_vfp_uqtos(tcg_single, tcg_int, - tcg_shift, tcg_fpst); - } - if (elements == 1) { - write_fp_sreg(s, rd, tcg_single); - } else { - write_vec_element_i32(s, tcg_single, rd, pass, MO_32); - } - tcg_temp_free_i32(tcg_single); } + + tcg_temp_free_i64(tcg_int64); + tcg_temp_free_i64(tcg_double); + + } else { + TCGv_i32 tcg_int32 = tcg_temp_new_i32(); + TCGv_i32 tcg_float = tcg_temp_new_i32(); + + for (pass = 0; pass < elements; pass++) { + read_vec_element_i32(s, tcg_int32, rn, pass, mop); + + switch (size) { + case MO_32: + if (fracbits) { + if (is_signed) { + gen_helper_vfp_sltos(tcg_float, tcg_int32, + tcg_shift, tcg_fpst); + } else { + gen_helper_vfp_ultos(tcg_float, tcg_int32, + tcg_shift, tcg_fpst); + } + } else { + if (is_signed) { + gen_helper_vfp_sitos(tcg_float, tcg_int32, tcg_fpst); + } else { + gen_helper_vfp_uitos(tcg_float, tcg_int32, tcg_fpst); + } + } + break; + case MO_16: + if (fracbits) { + if (is_signed) { + gen_helper_vfp_sltoh(tcg_float, tcg_int32, + tcg_shift, tcg_fpst); + } else { + gen_helper_vfp_ultoh(tcg_float, tcg_int32, + tcg_shift, tcg_fpst); + } + } else { + if (is_signed) { + gen_helper_vfp_sitoh(tcg_float, tcg_int32, tcg_fpst); + } else { + gen_helper_vfp_uitoh(tcg_float, tcg_int32, tcg_fpst); + } + } + break; + default: + g_assert_not_reached(); + } + + if (elements == 1) { + write_fp_sreg(s, rd, tcg_float); + } else { + write_vec_element_i32(s, tcg_float, rd, pass, size); + } + } + + tcg_temp_free_i32(tcg_int32); + tcg_temp_free_i32(tcg_float); } - tcg_temp_free_i64(tcg_int); tcg_temp_free_ptr(tcg_fpst); - tcg_temp_free_i32(tcg_shift); + if (tcg_shift) { + tcg_temp_free_i32(tcg_shift); + } clear_vec_high(s, elements << size == 16, rd); } @@ -XXX,XX +XXX,XX @@ static void disas_simd_two_reg_misc_fp16(DisasContext *s, uint32_t insn) rn = extract32(insn, 5, 5); switch (fpop) { + case 0x1d: /* SCVTF */ + case 0x5d: /* UCVTF */ + { + int elements; + + if (is_scalar) { + elements = 1; + } else { + elements = (is_q ? 8 : 4); + } + + if (!fp_access_check(s)) { + return; + } + handle_simd_intfp_conv(s, rd, rn, elements, !u, 0, MO_16); + return; + } break; case 0x2c: /* FCMGT (zero) */ case 0x2d: /* FCMEQ (zero) */ -- 2.16.2 From: Alex Bennée <alex.bennee@linaro.org> Neither of these operations alter the floating point status registers so we can do a pure bitwise operation, either squashing any sign bit (ABS) or inverting it (NEG). Signed-off-by: Alex Bennée <alex.bennee@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20180227143852.11175-22-alex.bennee@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/translate-a64.c | 16 +++++++++++++++- 1 file changed, 15 insertions(+), 1 deletion(-) diff --git a/target/arm/translate-a64.c b/target/arm/translate-a64.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-a64.c +++ b/target/arm/translate-a64.c @@ -XXX,XX +XXX,XX @@ static void disas_simd_two_reg_misc_fp16(DisasContext *s, uint32_t insn) TCGv_i32 tcg_rmode = NULL; TCGv_ptr tcg_fpstatus = NULL; bool need_rmode = false; + bool need_fpst = true; int rmode; if (!arm_dc_feature(s, ARM_FEATURE_V8_FP16)) { @@ -XXX,XX +XXX,XX @@ static void disas_simd_two_reg_misc_fp16(DisasContext *s, uint32_t insn) need_rmode = true; rmode = FPROUNDING_ZERO; break; + case 0x2f: /* FABS */ + case 0x6f: /* FNEG */ + need_fpst = false; + break; default: fprintf(stderr, "%s: insn %#04x fpop %#2x\n", __func__, insn, fpop); g_assert_not_reached(); @@ -XXX,XX +XXX,XX @@ static void disas_simd_two_reg_misc_fp16(DisasContext *s, uint32_t insn) return; } - if (need_rmode) { + if (need_rmode || need_fpst) { tcg_fpstatus = get_fpstatus_ptr(true); } @@ -XXX,XX +XXX,XX @@ static void disas_simd_two_reg_misc_fp16(DisasContext *s, uint32_t insn) case 0x7b: /* FCVTZU */ gen_helper_advsimd_f16touinth(tcg_res, tcg_op, tcg_fpstatus); break; + case 0x6f: /* FNEG */ + tcg_gen_xori_i32(tcg_res, tcg_op, 0x8000); + break; default: g_assert_not_reached(); } @@ -XXX,XX +XXX,XX @@ static void disas_simd_two_reg_misc_fp16(DisasContext *s, uint32_t insn) case 0x59: /* FRINTX */ gen_helper_advsimd_rinth_exact(tcg_res, tcg_op, tcg_fpstatus); break; + case 0x2f: /* FABS */ + tcg_gen_andi_i32(tcg_res, tcg_op, 0x7fff); + break; + case 0x6f: /* FNEG */ + tcg_gen_xori_i32(tcg_res, tcg_op, 0x8000); + break; default: g_assert_not_reached(); } -- 2.16.2 From: Alex Bennée <alex.bennee@linaro.org> It looks like the ARM ARM has simplified the pseudo code for the calculation which is done on a fixed point 9 bit integer maths. So while adding f16 we can also clean this up to be a little less heavy on the floating point and just return the fractional part and leave the calle's to do the final packing of the result. Signed-off-by: Alex Bennée <alex.bennee@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20180227143852.11175-23-alex.bennee@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/helper.h | 1 + target/arm/helper.c | 226 +++++++++++++++++++++++++++++----------------------- 2 files changed, 129 insertions(+), 98 deletions(-) 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_4(vfp_muladds, f32, f32, f32, f32, ptr) DEF_HELPER_3(recps_f32, f32, f32, f32, env) DEF_HELPER_3(rsqrts_f32, f32, f32, f32, env) +DEF_HELPER_FLAGS_2(recpe_f16, TCG_CALL_NO_RWG, f16, f16, ptr) DEF_HELPER_FLAGS_2(recpe_f32, TCG_CALL_NO_RWG, f32, f32, ptr) DEF_HELPER_FLAGS_2(recpe_f64, TCG_CALL_NO_RWG, f64, f64, ptr) DEF_HELPER_FLAGS_2(rsqrte_f32, TCG_CALL_NO_RWG, f32, f32, ptr) diff --git a/target/arm/helper.c b/target/arm/helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper.c +++ b/target/arm/helper.c @@ -XXX,XX +XXX,XX @@ float32 HELPER(rsqrts_f32)(float32 a, float32 b, CPUARMState *env) * int->float conversions at run-time. */ #define float64_256 make_float64(0x4070000000000000LL) #define float64_512 make_float64(0x4080000000000000LL) +#define float16_maxnorm make_float16(0x7bff) #define float32_maxnorm make_float32(0x7f7fffff) #define float64_maxnorm make_float64(0x7fefffffffffffffLL) /* Reciprocal functions * * The algorithm that must be used to calculate the estimate - * is specified by the ARM ARM, see FPRecipEstimate() + * is specified by the ARM ARM, see FPRecipEstimate()/RecipEstimate */ -static float64 recip_estimate(float64 a, float_status *real_fp_status) +/* See RecipEstimate() + * + * input is a 9 bit fixed point number + * input range 256 .. 511 for a number from 0.5 <= x < 1.0. + * result range 256 .. 511 for a number from 1.0 to 511/256. + */ + +static int recip_estimate(int input) { - /* These calculations mustn't set any fp exception flags, - * so we use a local copy of the fp_status. - */ - float_status dummy_status = *real_fp_status; - float_status *s = &dummy_status; - /* q = (int)(a * 512.0) */ - float64 q = float64_mul(float64_512, a, s); - int64_t q_int = float64_to_int64_round_to_zero(q, s); - - /* r = 1.0 / (((double)q + 0.5) / 512.0) */ - q = int64_to_float64(q_int, s); - q = float64_add(q, float64_half, s); - q = float64_div(q, float64_512, s); - q = float64_div(float64_one, q, s); - - /* s = (int)(256.0 * r + 0.5) */ - q = float64_mul(q, float64_256, s); - q = float64_add(q, float64_half, s); - q_int = float64_to_int64_round_to_zero(q, s); - - /* return (double)s / 256.0 */ - return float64_div(int64_to_float64(q_int, s), float64_256, s); + int a, b, r; + assert(256 <= input && input < 512); + a = (input * 2) + 1; + b = (1 << 19) / a; + r = (b + 1) >> 1; + assert(256 <= r && r < 512); + return r; } -/* Common wrapper to call recip_estimate */ -static float64 call_recip_estimate(float64 num, int off, float_status *fpst) -{ - uint64_t val64 = float64_val(num); - uint64_t frac = extract64(val64, 0, 52); - int64_t exp = extract64(val64, 52, 11); - uint64_t sbit; - float64 scaled, estimate; +/* + * Common wrapper to call recip_estimate + * + * The parameters are exponent and 64 bit fraction (without implicit + * bit) where the binary point is nominally at bit 52. Returns a + * float64 which can then be rounded to the appropriate size by the + * callee. + */ - /* Generate the scaled number for the estimate function */ - if (exp == 0) { +static uint64_t call_recip_estimate(int *exp, int exp_off, uint64_t frac) +{ + uint32_t scaled, estimate; + uint64_t result_frac; + int result_exp; + + /* Handle sub-normals */ + if (*exp == 0) { if (extract64(frac, 51, 1) == 0) { - exp = -1; - frac = extract64(frac, 0, 50) << 2; + *exp = -1; + frac <<= 2; } else { - frac = extract64(frac, 0, 51) << 1; + frac <<= 1; } } - /* scaled = '0' : '01111111110' : fraction<51:44> : Zeros(44); */ - scaled = make_float64((0x3feULL << 52) - | extract64(frac, 44, 8) << 44); + /* scaled = UInt('1':fraction<51:44>) */ + scaled = deposit32(1 << 8, 0, 8, extract64(frac, 44, 8)); + estimate = recip_estimate(scaled); - estimate = recip_estimate(scaled, fpst); - - /* Build new result */ - val64 = float64_val(estimate); - sbit = 0x8000000000000000ULL & val64; - exp = off - exp; - frac = extract64(val64, 0, 52); - - if (exp == 0) { - frac = 1ULL << 51 | extract64(frac, 1, 51); - } else if (exp == -1) { - frac = 1ULL << 50 | extract64(frac, 2, 50); - exp = 0; + result_exp = exp_off - *exp; + result_frac = deposit64(0, 44, 8, estimate); + if (result_exp == 0) { + result_frac = deposit64(result_frac >> 1, 51, 1, 1); + } else if (result_exp == -1) { + result_frac = deposit64(result_frac >> 2, 50, 2, 1); + result_exp = 0; } - return make_float64(sbit | (exp << 52) | frac); + *exp = result_exp; + + return result_frac; } static bool round_to_inf(float_status *fpst, bool sign_bit) @@ -XXX,XX +XXX,XX @@ static bool round_to_inf(float_status *fpst, bool sign_bit) g_assert_not_reached(); } +float16 HELPER(recpe_f16)(float16 input, void *fpstp) +{ + float_status *fpst = fpstp; + float16 f16 = float16_squash_input_denormal(input, fpst); + uint32_t f16_val = float16_val(f16); + uint32_t f16_sign = float16_is_neg(f16); + int f16_exp = extract32(f16_val, 10, 5); + uint32_t f16_frac = extract32(f16_val, 0, 10); + uint64_t f64_frac; + + if (float16_is_any_nan(f16)) { + float16 nan = f16; + if (float16_is_signaling_nan(f16, fpst)) { + float_raise(float_flag_invalid, fpst); + nan = float16_maybe_silence_nan(f16, fpst); + } + if (fpst->default_nan_mode) { + nan = float16_default_nan(fpst); + } + return nan; + } else if (float16_is_infinity(f16)) { + return float16_set_sign(float16_zero, float16_is_neg(f16)); + } else if (float16_is_zero(f16)) { + float_raise(float_flag_divbyzero, fpst); + return float16_set_sign(float16_infinity, float16_is_neg(f16)); + } else if (float16_abs(f16) < (1 << 8)) { + /* Abs(value) < 2.0^-16 */ + float_raise(float_flag_overflow | float_flag_inexact, fpst); + if (round_to_inf(fpst, f16_sign)) { + return float16_set_sign(float16_infinity, f16_sign); + } else { + return float16_set_sign(float16_maxnorm, f16_sign); + } + } else if (f16_exp >= 29 && fpst->flush_to_zero) { + float_raise(float_flag_underflow, fpst); + return float16_set_sign(float16_zero, float16_is_neg(f16)); + } + + f64_frac = call_recip_estimate(&f16_exp, 29, + ((uint64_t) f16_frac) << (52 - 10)); + + /* result = sign : result_exp<4:0> : fraction<51:42> */ + f16_val = deposit32(0, 15, 1, f16_sign); + f16_val = deposit32(f16_val, 10, 5, f16_exp); + f16_val = deposit32(f16_val, 0, 10, extract64(f64_frac, 52 - 10, 10)); + return make_float16(f16_val); +} + float32 HELPER(recpe_f32)(float32 input, void *fpstp) { float_status *fpst = fpstp; float32 f32 = float32_squash_input_denormal(input, fpst); uint32_t f32_val = float32_val(f32); - uint32_t f32_sbit = 0x80000000ULL & f32_val; - int32_t f32_exp = extract32(f32_val, 23, 8); + bool f32_sign = float32_is_neg(f32); + int f32_exp = extract32(f32_val, 23, 8); uint32_t f32_frac = extract32(f32_val, 0, 23); - float64 f64, r64; - uint64_t r64_val; - int64_t r64_exp; - uint64_t r64_frac; + uint64_t f64_frac; if (float32_is_any_nan(f32)) { float32 nan = f32; @@ -XXX,XX +XXX,XX @@ float32 HELPER(recpe_f32)(float32 input, void *fpstp) } else if (float32_is_zero(f32)) { float_raise(float_flag_divbyzero, fpst); return float32_set_sign(float32_infinity, float32_is_neg(f32)); - } else if ((f32_val & ~(1ULL << 31)) < (1ULL << 21)) { + } else if (float32_abs(f32) < (1ULL << 21)) { /* Abs(value) < 2.0^-128 */ float_raise(float_flag_overflow | float_flag_inexact, fpst); - if (round_to_inf(fpst, f32_sbit)) { - return float32_set_sign(float32_infinity, float32_is_neg(f32)); + if (round_to_inf(fpst, f32_sign)) { + return float32_set_sign(float32_infinity, f32_sign); } else { - return float32_set_sign(float32_maxnorm, float32_is_neg(f32)); + return float32_set_sign(float32_maxnorm, f32_sign); } } else if (f32_exp >= 253 && fpst->flush_to_zero) { float_raise(float_flag_underflow, fpst); return float32_set_sign(float32_zero, float32_is_neg(f32)); } + f64_frac = call_recip_estimate(&f32_exp, 253, + ((uint64_t) f32_frac) << (52 - 23)); - f64 = make_float64(((int64_t)(f32_exp) << 52) | (int64_t)(f32_frac) << 29); - r64 = call_recip_estimate(f64, 253, fpst); - r64_val = float64_val(r64); - r64_exp = extract64(r64_val, 52, 11); - r64_frac = extract64(r64_val, 0, 52); - - /* result = sign : result_exp<7:0> : fraction<51:29>; */ - return make_float32(f32_sbit | - (r64_exp & 0xff) << 23 | - extract64(r64_frac, 29, 24)); + /* result = sign : result_exp<7:0> : fraction<51:29> */ + f32_val = deposit32(0, 31, 1, f32_sign); + f32_val = deposit32(f32_val, 23, 8, f32_exp); + f32_val = deposit32(f32_val, 0, 23, extract64(f64_frac, 52 - 23, 23)); + return make_float32(f32_val); } float64 HELPER(recpe_f64)(float64 input, void *fpstp) @@ -XXX,XX +XXX,XX @@ float64 HELPER(recpe_f64)(float64 input, void *fpstp) float_status *fpst = fpstp; float64 f64 = float64_squash_input_denormal(input, fpst); uint64_t f64_val = float64_val(f64); - uint64_t f64_sbit = 0x8000000000000000ULL & f64_val; - int64_t f64_exp = extract64(f64_val, 52, 11); - float64 r64; - uint64_t r64_val; - int64_t r64_exp; - uint64_t r64_frac; + bool f64_sign = float64_is_neg(f64); + int f64_exp = extract64(f64_val, 52, 11); + uint64_t f64_frac = extract64(f64_val, 0, 52); /* Deal with any special cases */ if (float64_is_any_nan(f64)) { @@ -XXX,XX +XXX,XX @@ float64 HELPER(recpe_f64)(float64 input, void *fpstp) } else if ((f64_val & ~(1ULL << 63)) < (1ULL << 50)) { /* Abs(value) < 2.0^-1024 */ float_raise(float_flag_overflow | float_flag_inexact, fpst); - if (round_to_inf(fpst, f64_sbit)) { - return float64_set_sign(float64_infinity, float64_is_neg(f64)); + if (round_to_inf(fpst, f64_sign)) { + return float64_set_sign(float64_infinity, f64_sign); } else { - return float64_set_sign(float64_maxnorm, float64_is_neg(f64)); + return float64_set_sign(float64_maxnorm, f64_sign); } } else if (f64_exp >= 2045 && fpst->flush_to_zero) { float_raise(float_flag_underflow, fpst); return float64_set_sign(float64_zero, float64_is_neg(f64)); } - r64 = call_recip_estimate(f64, 2045, fpst); - r64_val = float64_val(r64); - r64_exp = extract64(r64_val, 52, 11); - r64_frac = extract64(r64_val, 0, 52); + f64_frac = call_recip_estimate(&f64_exp, 2045, f64_frac); - /* result = sign : result_exp<10:0> : fraction<51:0> */ - return make_float64(f64_sbit | - ((r64_exp & 0x7ff) << 52) | - r64_frac); + /* result = sign : result_exp<10:0> : fraction<51:0>; */ + f64_val = deposit64(0, 63, 1, f64_sign); + f64_val = deposit64(f64_val, 52, 11, f64_exp); + f64_val = deposit64(f64_val, 0, 52, f64_frac); + return make_float64(f64_val); } /* The algorithm that must be used to calculate the estimate @@ -XXX,XX +XXX,XX @@ float64 HELPER(rsqrte_f64)(float64 input, void *fpstp) uint32_t HELPER(recpe_u32)(uint32_t a, void *fpstp) { - float_status *s = fpstp; - float64 f64; + /* float_status *s = fpstp; */ + int input, estimate; if ((a & 0x80000000) == 0) { return 0xffffffff; } - f64 = make_float64((0x3feULL << 52) - | ((int64_t)(a & 0x7fffffff) << 21)); + input = extract32(a, 23, 9); + estimate = recip_estimate(input); - f64 = recip_estimate(f64, s); - - return 0x80000000 | ((float64_val(f64) >> 21) & 0x7fffffff); + return deposit32(0, (32 - 9), 9, estimate); } uint32_t HELPER(rsqrte_u32)(uint32_t a, void *fpstp) -- 2.16.2 From: Alex Bennée <alex.bennee@linaro.org> Now we have added f16 during the re-factoring we can simply call the helper. Signed-off-by: Alex Bennée <alex.bennee@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20180227143852.11175-24-alex.bennee@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/translate-a64.c | 8 ++++++++ 1 file changed, 8 insertions(+) diff --git a/target/arm/translate-a64.c b/target/arm/translate-a64.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-a64.c +++ b/target/arm/translate-a64.c @@ -XXX,XX +XXX,XX @@ static void disas_simd_two_reg_misc_fp16(DisasContext *s, uint32_t insn) case 0x6d: /* FCMLE (zero) */ handle_2misc_fcmp_zero(s, fpop, is_scalar, 0, is_q, MO_16, rn, rd); return; + case 0x3d: /* FRECPE */ + break; case 0x18: /* FRINTN */ need_rmode = true; only_in_vector = true; @@ -XXX,XX +XXX,XX @@ static void disas_simd_two_reg_misc_fp16(DisasContext *s, uint32_t insn) case 0x3b: /* FCVTZS */ gen_helper_advsimd_f16tosinth(tcg_res, tcg_op, tcg_fpstatus); break; + case 0x3d: /* FRECPE */ + gen_helper_recpe_f16(tcg_res, tcg_op, tcg_fpstatus); + break; case 0x5a: /* FCVTNU */ case 0x5b: /* FCVTMU */ case 0x5c: /* FCVTAU */ @@ -XXX,XX +XXX,XX @@ static void disas_simd_two_reg_misc_fp16(DisasContext *s, uint32_t insn) case 0x3b: /* FCVTZS */ gen_helper_advsimd_f16tosinth(tcg_res, tcg_op, tcg_fpstatus); break; + case 0x3d: /* FRECPE */ + gen_helper_recpe_f16(tcg_res, tcg_op, tcg_fpstatus); + break; case 0x5a: /* FCVTNU */ case 0x5b: /* FCVTMU */ case 0x5c: /* FCVTAU */ -- 2.16.2 From: Alex Bennée <alex.bennee@linaro.org> We go with the localised helper. Signed-off-by: Alex Bennée <alex.bennee@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20180227143852.11175-25-alex.bennee@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/helper-a64.h | 1 + target/arm/helper-a64.c | 29 +++++++++++++++++++++++++++++ target/arm/translate-a64.c | 4 ++++ 3 files changed, 34 insertions(+) diff --git a/target/arm/helper-a64.h b/target/arm/helper-a64.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper-a64.h +++ b/target/arm/helper-a64.h @@ -XXX,XX +XXX,XX @@ DEF_HELPER_FLAGS_1(neon_addlp_s16, TCG_CALL_NO_RWG_SE, i64, i64) DEF_HELPER_FLAGS_1(neon_addlp_u16, TCG_CALL_NO_RWG_SE, i64, i64) DEF_HELPER_FLAGS_2(frecpx_f64, TCG_CALL_NO_RWG, f64, f64, ptr) DEF_HELPER_FLAGS_2(frecpx_f32, TCG_CALL_NO_RWG, f32, f32, ptr) +DEF_HELPER_FLAGS_2(frecpx_f16, TCG_CALL_NO_RWG, f16, f16, ptr) DEF_HELPER_FLAGS_2(fcvtx_f64_to_f32, TCG_CALL_NO_RWG, f32, f64, env) DEF_HELPER_FLAGS_3(crc32_64, TCG_CALL_NO_RWG_SE, i64, i64, i64, i32) DEF_HELPER_FLAGS_3(crc32c_64, TCG_CALL_NO_RWG_SE, i64, i64, i64, i32) diff --git a/target/arm/helper-a64.c b/target/arm/helper-a64.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper-a64.c +++ b/target/arm/helper-a64.c @@ -XXX,XX +XXX,XX @@ uint64_t HELPER(neon_addlp_u16)(uint64_t a) } /* Floating-point reciprocal exponent - see FPRecpX in ARM ARM */ +float16 HELPER(frecpx_f16)(float16 a, void *fpstp) +{ + float_status *fpst = fpstp; + uint16_t val16, sbit; + int16_t exp; + + if (float16_is_any_nan(a)) { + float16 nan = a; + if (float16_is_signaling_nan(a, fpst)) { + float_raise(float_flag_invalid, fpst); + nan = float16_maybe_silence_nan(a, fpst); + } + if (fpst->default_nan_mode) { + nan = float16_default_nan(fpst); + } + return nan; + } + + val16 = float16_val(a); + sbit = 0x8000 & val16; + exp = extract32(val16, 10, 5); + + if (exp == 0) { + return make_float16(deposit32(sbit, 10, 5, 0x1e)); + } else { + return make_float16(deposit32(sbit, 10, 5, ~exp)); + } +} + float32 HELPER(frecpx_f32)(float32 a, void *fpstp) { float_status *fpst = fpstp; diff --git a/target/arm/translate-a64.c b/target/arm/translate-a64.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-a64.c +++ b/target/arm/translate-a64.c @@ -XXX,XX +XXX,XX @@ static void disas_simd_two_reg_misc_fp16(DisasContext *s, uint32_t insn) handle_2misc_fcmp_zero(s, fpop, is_scalar, 0, is_q, MO_16, rn, rd); return; case 0x3d: /* FRECPE */ + case 0x3f: /* FRECPX */ break; case 0x18: /* FRINTN */ need_rmode = true; @@ -XXX,XX +XXX,XX @@ static void disas_simd_two_reg_misc_fp16(DisasContext *s, uint32_t insn) case 0x3d: /* FRECPE */ gen_helper_recpe_f16(tcg_res, tcg_op, tcg_fpstatus); break; + case 0x3f: /* FRECPX */ + gen_helper_frecpx_f16(tcg_res, tcg_op, tcg_fpstatus); + break; case 0x5a: /* FCVTNU */ case 0x5b: /* FCVTMU */ case 0x5c: /* FCVTAU */ -- 2.16.2 From: Alex Bennée <alex.bennee@linaro.org> Signed-off-by: Alex Bennée <alex.bennee@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20180227143852.11175-26-alex.bennee@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/helper-a64.h | 1 + target/arm/helper-a64.c | 13 +++++++++++++ target/arm/translate-a64.c | 5 +++++ 3 files changed, 19 insertions(+) diff --git a/target/arm/helper-a64.h b/target/arm/helper-a64.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper-a64.h +++ b/target/arm/helper-a64.h @@ -XXX,XX +XXX,XX @@ DEF_HELPER_2(advsimd_rinth_exact, f16, f16, ptr) DEF_HELPER_2(advsimd_rinth, f16, f16, ptr) DEF_HELPER_2(advsimd_f16tosinth, i32, f16, ptr) DEF_HELPER_2(advsimd_f16touinth, i32, f16, ptr) +DEF_HELPER_2(sqrt_f16, f16, f16, ptr) diff --git a/target/arm/helper-a64.c b/target/arm/helper-a64.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper-a64.c +++ b/target/arm/helper-a64.c @@ -XXX,XX +XXX,XX @@ uint32_t HELPER(advsimd_f16touinth)(float16 a, void *fpstp) } return float16_to_uint16(a, fpst); } + +/* + * Square Root and Reciprocal square root + */ + +float16 HELPER(sqrt_f16)(float16 a, void *fpstp) +{ + float_status *s = fpstp; + + return float16_sqrt(a, s); +} + + diff --git a/target/arm/translate-a64.c b/target/arm/translate-a64.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-a64.c +++ b/target/arm/translate-a64.c @@ -XXX,XX +XXX,XX @@ static void disas_simd_two_reg_misc_fp16(DisasContext *s, uint32_t insn) case 0x6f: /* FNEG */ need_fpst = false; break; + case 0x7f: /* FSQRT (vector) */ + break; default: fprintf(stderr, "%s: insn %#04x fpop %#2x\n", __func__, insn, fpop); g_assert_not_reached(); @@ -XXX,XX +XXX,XX @@ static void disas_simd_two_reg_misc_fp16(DisasContext *s, uint32_t insn) case 0x6f: /* FNEG */ tcg_gen_xori_i32(tcg_res, tcg_op, 0x8000); break; + case 0x7f: /* FSQRT */ + gen_helper_sqrt_f16(tcg_res, tcg_op, tcg_fpstatus); + break; default: g_assert_not_reached(); } -- 2.16.2 From: Alex Bennée <alex.bennee@linaro.org> Much like recpe the ARM ARM has simplified the pseudo code for the calculation which is done on a fixed point 9 bit integer maths. So while adding f16 we can also clean this up to be a little less heavy on the floating point and just return the fractional part and leave the calle's to do the final packing of the result. Signed-off-by: Alex Bennée <alex.bennee@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20180227143852.11175-27-alex.bennee@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/helper.h | 1 + target/arm/helper.c | 221 ++++++++++++++++++++++++---------------------------- 2 files changed, 104 insertions(+), 118 deletions(-) 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_3(rsqrts_f32, f32, f32, f32, env) DEF_HELPER_FLAGS_2(recpe_f16, TCG_CALL_NO_RWG, f16, f16, ptr) DEF_HELPER_FLAGS_2(recpe_f32, TCG_CALL_NO_RWG, f32, f32, ptr) DEF_HELPER_FLAGS_2(recpe_f64, TCG_CALL_NO_RWG, f64, f64, ptr) +DEF_HELPER_FLAGS_2(rsqrte_f16, TCG_CALL_NO_RWG, f16, f16, ptr) DEF_HELPER_FLAGS_2(rsqrte_f32, TCG_CALL_NO_RWG, f32, f32, ptr) DEF_HELPER_FLAGS_2(rsqrte_f64, TCG_CALL_NO_RWG, f64, f64, ptr) DEF_HELPER_2(recpe_u32, i32, i32, ptr) 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 @@ float64 HELPER(recpe_f64)(float64 input, void *fpstp) /* The algorithm that must be used to calculate the estimate * is specified by the ARM ARM. */ -static float64 recip_sqrt_estimate(float64 a, float_status *real_fp_status) + +static int do_recip_sqrt_estimate(int a) { - /* These calculations mustn't set any fp exception flags, - * so we use a local copy of the fp_status. - */ - float_status dummy_status = *real_fp_status; - float_status *s = &dummy_status; - float64 q; - int64_t q_int; + int b, estimate; - if (float64_lt(a, float64_half, s)) { - /* range 0.25 <= a < 0.5 */ - - /* a in units of 1/512 rounded down */ - /* q0 = (int)(a * 512.0); */ - q = float64_mul(float64_512, a, s); - q_int = float64_to_int64_round_to_zero(q, s); - - /* reciprocal root r */ - /* r = 1.0 / sqrt(((double)q0 + 0.5) / 512.0); */ - q = int64_to_float64(q_int, s); - q = float64_add(q, float64_half, s); - q = float64_div(q, float64_512, s); - q = float64_sqrt(q, s); - q = float64_div(float64_one, q, s); + assert(128 <= a && a < 512); + if (a < 256) { + a = a * 2 + 1; } else { - /* range 0.5 <= a < 1.0 */ - - /* a in units of 1/256 rounded down */ - /* q1 = (int)(a * 256.0); */ - q = float64_mul(float64_256, a, s); - int64_t q_int = float64_to_int64_round_to_zero(q, s); - - /* reciprocal root r */ - /* r = 1.0 /sqrt(((double)q1 + 0.5) / 256); */ - q = int64_to_float64(q_int, s); - q = float64_add(q, float64_half, s); - q = float64_div(q, float64_256, s); - q = float64_sqrt(q, s); - q = float64_div(float64_one, q, s); + a = (a >> 1) << 1; + a = (a + 1) * 2; } - /* r in units of 1/256 rounded to nearest */ - /* s = (int)(256.0 * r + 0.5); */ + b = 512; + while (a * (b + 1) * (b + 1) < (1 << 28)) { + b += 1; + } + estimate = (b + 1) / 2; + assert(256 <= estimate && estimate < 512); - q = float64_mul(q, float64_256,s ); - q = float64_add(q, float64_half, s); - q_int = float64_to_int64_round_to_zero(q, s); + return estimate; +} - /* return (double)s / 256.0;*/ - return float64_div(int64_to_float64(q_int, s), float64_256, s); + +static uint64_t recip_sqrt_estimate(int *exp , int exp_off, uint64_t frac) +{ + int estimate; + uint32_t scaled; + + if (*exp == 0) { + while (extract64(frac, 51, 1) == 0) { + frac = frac << 1; + *exp -= 1; + } + frac = extract64(frac, 0, 51) << 1; + } + + if (*exp & 1) { + /* scaled = UInt('01':fraction<51:45>) */ + scaled = deposit32(1 << 7, 0, 7, extract64(frac, 45, 7)); + } else { + /* scaled = UInt('1':fraction<51:44>) */ + scaled = deposit32(1 << 8, 0, 8, extract64(frac, 44, 8)); + } + estimate = do_recip_sqrt_estimate(scaled); + + *exp = (exp_off - *exp) / 2; + return extract64(estimate, 0, 8) << 44; +} + +float16 HELPER(rsqrte_f16)(float16 input, void *fpstp) +{ + float_status *s = fpstp; + float16 f16 = float16_squash_input_denormal(input, s); + uint16_t val = float16_val(f16); + bool f16_sign = float16_is_neg(f16); + int f16_exp = extract32(val, 10, 5); + uint16_t f16_frac = extract32(val, 0, 10); + uint64_t f64_frac; + + if (float16_is_any_nan(f16)) { + float16 nan = f16; + if (float16_is_signaling_nan(f16, s)) { + float_raise(float_flag_invalid, s); + nan = float16_maybe_silence_nan(f16, s); + } + if (s->default_nan_mode) { + nan = float16_default_nan(s); + } + return nan; + } else if (float16_is_zero(f16)) { + float_raise(float_flag_divbyzero, s); + return float16_set_sign(float16_infinity, f16_sign); + } else if (f16_sign) { + float_raise(float_flag_invalid, s); + return float16_default_nan(s); + } else if (float16_is_infinity(f16)) { + return float16_zero; + } + + /* Scale and normalize to a double-precision value between 0.25 and 1.0, + * preserving the parity of the exponent. */ + + f64_frac = ((uint64_t) f16_frac) << (52 - 10); + + f64_frac = recip_sqrt_estimate(&f16_exp, 44, f64_frac); + + /* result = sign : result_exp<4:0> : estimate<7:0> : Zeros(2) */ + val = deposit32(0, 15, 1, f16_sign); + val = deposit32(val, 10, 5, f16_exp); + val = deposit32(val, 2, 8, extract64(f64_frac, 52 - 8, 8)); + return make_float16(val); } float32 HELPER(rsqrte_f32)(float32 input, void *fpstp) @@ -XXX,XX +XXX,XX @@ float32 HELPER(rsqrte_f32)(float32 input, void *fpstp) float_status *s = fpstp; float32 f32 = float32_squash_input_denormal(input, s); uint32_t val = float32_val(f32); - uint32_t f32_sbit = 0x80000000 & val; - int32_t f32_exp = extract32(val, 23, 8); + uint32_t f32_sign = float32_is_neg(f32); + int f32_exp = extract32(val, 23, 8); uint32_t f32_frac = extract32(val, 0, 23); uint64_t f64_frac; - uint64_t val64; - int result_exp; - float64 f64; if (float32_is_any_nan(f32)) { float32 nan = f32; @@ -XXX,XX +XXX,XX @@ float32 HELPER(rsqrte_f32)(float32 input, void *fpstp) * preserving the parity of the exponent. */ f64_frac = ((uint64_t) f32_frac) << 29; - if (f32_exp == 0) { - while (extract64(f64_frac, 51, 1) == 0) { - f64_frac = f64_frac << 1; - f32_exp = f32_exp-1; - } - f64_frac = extract64(f64_frac, 0, 51) << 1; - } - if (extract64(f32_exp, 0, 1) == 0) { - f64 = make_float64(((uint64_t) f32_sbit) << 32 - | (0x3feULL << 52) - | f64_frac); - } else { - f64 = make_float64(((uint64_t) f32_sbit) << 32 - | (0x3fdULL << 52) - | f64_frac); - } + f64_frac = recip_sqrt_estimate(&f32_exp, 380, f64_frac); - result_exp = (380 - f32_exp) / 2; - - f64 = recip_sqrt_estimate(f64, s); - - val64 = float64_val(f64); - - val = ((result_exp & 0xff) << 23) - | ((val64 >> 29) & 0x7fffff); + /* result = sign : result_exp<4:0> : estimate<7:0> : Zeros(15) */ + val = deposit32(0, 31, 1, f32_sign); + val = deposit32(val, 23, 8, f32_exp); + val = deposit32(val, 15, 8, extract64(f64_frac, 52 - 8, 8)); return make_float32(val); } @@ -XXX,XX +XXX,XX @@ float64 HELPER(rsqrte_f64)(float64 input, void *fpstp) float_status *s = fpstp; float64 f64 = float64_squash_input_denormal(input, s); uint64_t val = float64_val(f64); - uint64_t f64_sbit = 0x8000000000000000ULL & val; - int64_t f64_exp = extract64(val, 52, 11); + bool f64_sign = float64_is_neg(f64); + int f64_exp = extract64(val, 52, 11); uint64_t f64_frac = extract64(val, 0, 52); - int64_t result_exp; - uint64_t result_frac; if (float64_is_any_nan(f64)) { float64 nan = f64; @@ -XXX,XX +XXX,XX @@ float64 HELPER(rsqrte_f64)(float64 input, void *fpstp) return float64_zero; } - /* Scale and normalize to a double-precision value between 0.25 and 1.0, - * preserving the parity of the exponent. */ + f64_frac = recip_sqrt_estimate(&f64_exp, 3068, f64_frac); - if (f64_exp == 0) { - while (extract64(f64_frac, 51, 1) == 0) { - f64_frac = f64_frac << 1; - f64_exp = f64_exp - 1; - } - f64_frac = extract64(f64_frac, 0, 51) << 1; - } - - if (extract64(f64_exp, 0, 1) == 0) { - f64 = make_float64(f64_sbit - | (0x3feULL << 52) - | f64_frac); - } else { - f64 = make_float64(f64_sbit - | (0x3fdULL << 52) - | f64_frac); - } - - result_exp = (3068 - f64_exp) / 2; - - f64 = recip_sqrt_estimate(f64, s); - - result_frac = extract64(float64_val(f64), 0, 52); - - return make_float64(f64_sbit | - ((result_exp & 0x7ff) << 52) | - result_frac); + /* result = sign : result_exp<4:0> : estimate<7:0> : Zeros(44) */ + val = deposit64(0, 61, 1, f64_sign); + val = deposit64(val, 52, 11, f64_exp); + val = deposit64(val, 44, 8, extract64(f64_frac, 52 - 8, 8)); + return make_float64(val); } uint32_t HELPER(recpe_u32)(uint32_t a, void *fpstp) @@ -XXX,XX +XXX,XX @@ uint32_t HELPER(recpe_u32)(uint32_t a, void *fpstp) uint32_t HELPER(rsqrte_u32)(uint32_t a, void *fpstp) { - float_status *fpst = fpstp; - float64 f64; + int estimate; if ((a & 0xc0000000) == 0) { return 0xffffffff; } - if (a & 0x80000000) { - f64 = make_float64((0x3feULL << 52) - | ((uint64_t)(a & 0x7fffffff) << 21)); - } else { /* bits 31-30 == '01' */ - f64 = make_float64((0x3fdULL << 52) - | ((uint64_t)(a & 0x3fffffff) << 22)); - } + estimate = do_recip_sqrt_estimate(extract32(a, 23, 9)); - f64 = recip_sqrt_estimate(f64, fpst); - - return 0x80000000 | ((float64_val(f64) >> 21) & 0x7fffffff); + return deposit32(0, 23, 9, estimate); } /* VFPv4 fused multiply-accumulate */ -- 2.16.2 From: Alex Bennée <alex.bennee@linaro.org> Signed-off-by: Alex Bennée <alex.bennee@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20180227143852.11175-28-alex.bennee@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/translate-a64.c | 7 +++++++ 1 file changed, 7 insertions(+) diff --git a/target/arm/translate-a64.c b/target/arm/translate-a64.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-a64.c +++ b/target/arm/translate-a64.c @@ -XXX,XX +XXX,XX @@ static void disas_simd_two_reg_misc_fp16(DisasContext *s, uint32_t insn) case 0x6f: /* FNEG */ need_fpst = false; break; + case 0x7d: /* FRSQRTE */ case 0x7f: /* FSQRT (vector) */ break; default: @@ -XXX,XX +XXX,XX @@ static void disas_simd_two_reg_misc_fp16(DisasContext *s, uint32_t insn) case 0x6f: /* FNEG */ tcg_gen_xori_i32(tcg_res, tcg_op, 0x8000); break; + case 0x7d: /* FRSQRTE */ + gen_helper_rsqrte_f16(tcg_res, tcg_op, tcg_fpstatus); + break; default: g_assert_not_reached(); } @@ -XXX,XX +XXX,XX @@ static void disas_simd_two_reg_misc_fp16(DisasContext *s, uint32_t insn) case 0x6f: /* FNEG */ tcg_gen_xori_i32(tcg_res, tcg_op, 0x8000); break; + case 0x7d: /* FRSQRTE */ + gen_helper_rsqrte_f16(tcg_res, tcg_op, tcg_fpstatus); + break; case 0x7f: /* FSQRT */ gen_helper_sqrt_f16(tcg_res, tcg_op, tcg_fpstatus); break; -- 2.16.2 From: Alex Bennée <alex.bennee@linaro.org> Only one half-precision instruction has been added to this group. Signed-off-by: Alex Bennée <alex.bennee@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20180227143852.11175-29-alex.bennee@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/translate-a64.c | 35 +++++++++++++++++++++++++---------- 1 file changed, 25 insertions(+), 10 deletions(-) diff --git a/target/arm/translate-a64.c b/target/arm/translate-a64.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-a64.c +++ b/target/arm/translate-a64.c @@ -XXX,XX +XXX,XX @@ static void disas_simd_copy(DisasContext *s, uint32_t insn) * MVNI - move inverted (shifted) imm into register * ORR - bitwise OR of (shifted) imm with register * BIC - bitwise clear of (shifted) imm with register + * With ARMv8.2 we also have: + * FMOV half-precision */ static void disas_simd_mod_imm(DisasContext *s, uint32_t insn) { @@ -XXX,XX +XXX,XX @@ static void disas_simd_mod_imm(DisasContext *s, uint32_t insn) uint64_t imm = 0; if (o2 != 0 || ((cmode == 0xf) && is_neg && !is_q)) { - unallocated_encoding(s); - return; + /* Check for FMOV (vector, immediate) - half-precision */ + if (!(arm_dc_feature(s, ARM_FEATURE_V8_FP16) && o2 && cmode == 0xf)) { + unallocated_encoding(s); + return; + } } if (!fp_access_check(s)) { @@ -XXX,XX +XXX,XX @@ static void disas_simd_mod_imm(DisasContext *s, uint32_t insn) imm |= 0x4000000000000000ULL; } } else { - imm = (abcdefgh & 0x3f) << 19; - if (abcdefgh & 0x80) { - imm |= 0x80000000; - } - if (abcdefgh & 0x40) { - imm |= 0x3e000000; + if (o2) { + /* FMOV (vector, immediate) - half-precision */ + imm = vfp_expand_imm(MO_16, abcdefgh); + /* now duplicate across the lanes */ + imm = bitfield_replicate(imm, 16); } else { - imm |= 0x40000000; + imm = (abcdefgh & 0x3f) << 19; + if (abcdefgh & 0x80) { + imm |= 0x80000000; + } + if (abcdefgh & 0x40) { + imm |= 0x3e000000; + } else { + imm |= 0x40000000; + } + imm |= (imm << 32); } - imm |= (imm << 32); } } break; + default: + fprintf(stderr, "%s: cmode_3_1: %x\n", __func__, cmode_3_1); + g_assert_not_reached(); } if (cmode_3_1 != 7 && is_neg) { -- 2.16.2 From: Alex Bennée <alex.bennee@linaro.org> I only needed to do a little light re-factoring to support the half-precision helpers. Signed-off-by: Alex Bennée <alex.bennee@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20180227143852.11175-30-alex.bennee@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/translate-a64.c | 80 +++++++++++++++++++++++++++++++--------------- 1 file changed, 54 insertions(+), 26 deletions(-) diff --git a/target/arm/translate-a64.c b/target/arm/translate-a64.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-a64.c +++ b/target/arm/translate-a64.c @@ -XXX,XX +XXX,XX @@ static void disas_simd_scalar_pairwise(DisasContext *s, uint32_t insn) case 0xf: /* FMAXP */ case 0x2c: /* FMINNMP */ case 0x2f: /* FMINP */ - /* FP op, size[0] is 32 or 64 bit */ + /* FP op, size[0] is 32 or 64 bit*/ if (!u) { - unallocated_encoding(s); - return; + if (!arm_dc_feature(s, ARM_FEATURE_V8_FP16)) { + unallocated_encoding(s); + return; + } else { + size = MO_16; + } + } else { + size = extract32(size, 0, 1) ? MO_64 : MO_32; } + if (!fp_access_check(s)) { return; } - size = extract32(size, 0, 1) ? 3 : 2; - fpst = get_fpstatus_ptr(false); + fpst = get_fpstatus_ptr(size == MO_16); break; default: unallocated_encoding(s); return; } - if (size == 3) { + if (size == MO_64) { TCGv_i64 tcg_op1 = tcg_temp_new_i64(); TCGv_i64 tcg_op2 = tcg_temp_new_i64(); TCGv_i64 tcg_res = tcg_temp_new_i64(); @@ -XXX,XX +XXX,XX @@ static void disas_simd_scalar_pairwise(DisasContext *s, uint32_t insn) TCGv_i32 tcg_op2 = tcg_temp_new_i32(); TCGv_i32 tcg_res = tcg_temp_new_i32(); - read_vec_element_i32(s, tcg_op1, rn, 0, MO_32); - read_vec_element_i32(s, tcg_op2, rn, 1, MO_32); + read_vec_element_i32(s, tcg_op1, rn, 0, size); + read_vec_element_i32(s, tcg_op2, rn, 1, size); - switch (opcode) { - case 0xc: /* FMAXNMP */ - gen_helper_vfp_maxnums(tcg_res, tcg_op1, tcg_op2, fpst); - break; - case 0xd: /* FADDP */ - gen_helper_vfp_adds(tcg_res, tcg_op1, tcg_op2, fpst); - break; - case 0xf: /* FMAXP */ - gen_helper_vfp_maxs(tcg_res, tcg_op1, tcg_op2, fpst); - break; - case 0x2c: /* FMINNMP */ - gen_helper_vfp_minnums(tcg_res, tcg_op1, tcg_op2, fpst); - break; - case 0x2f: /* FMINP */ - gen_helper_vfp_mins(tcg_res, tcg_op1, tcg_op2, fpst); - break; - default: - g_assert_not_reached(); + if (size == MO_16) { + switch (opcode) { + case 0xc: /* FMAXNMP */ + gen_helper_advsimd_maxnumh(tcg_res, tcg_op1, tcg_op2, fpst); + break; + case 0xd: /* FADDP */ + gen_helper_advsimd_addh(tcg_res, tcg_op1, tcg_op2, fpst); + break; + case 0xf: /* FMAXP */ + gen_helper_advsimd_maxh(tcg_res, tcg_op1, tcg_op2, fpst); + break; + case 0x2c: /* FMINNMP */ + gen_helper_advsimd_minnumh(tcg_res, tcg_op1, tcg_op2, fpst); + break; + case 0x2f: /* FMINP */ + gen_helper_advsimd_minh(tcg_res, tcg_op1, tcg_op2, fpst); + break; + default: + g_assert_not_reached(); + } + } else { + switch (opcode) { + case 0xc: /* FMAXNMP */ + gen_helper_vfp_maxnums(tcg_res, tcg_op1, tcg_op2, fpst); + break; + case 0xd: /* FADDP */ + gen_helper_vfp_adds(tcg_res, tcg_op1, tcg_op2, fpst); + break; + case 0xf: /* FMAXP */ + gen_helper_vfp_maxs(tcg_res, tcg_op1, tcg_op2, fpst); + break; + case 0x2c: /* FMINNMP */ + gen_helper_vfp_minnums(tcg_res, tcg_op1, tcg_op2, fpst); + break; + case 0x2f: /* FMINP */ + gen_helper_vfp_mins(tcg_res, tcg_op1, tcg_op2, fpst); + break; + default: + g_assert_not_reached(); + } } write_fp_sreg(s, rd, tcg_res); -- 2.16.2 From: Alex Bennée <alex.bennee@linaro.org> This covers the encoding group: Advanced SIMD scalar three same FP16 As all the helpers are already there it is simply a case of calling the existing helpers in the scalar context. Signed-off-by: Alex Bennée <alex.bennee@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20180227143852.11175-31-alex.bennee@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/translate-a64.c | 99 ++++++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 99 insertions(+) diff --git a/target/arm/translate-a64.c b/target/arm/translate-a64.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-a64.c +++ b/target/arm/translate-a64.c @@ -XXX,XX +XXX,XX @@ static void disas_simd_scalar_three_reg_same(DisasContext *s, uint32_t insn) tcg_temp_free_i64(tcg_rd); } +/* AdvSIMD scalar three same FP16 + * 31 30 29 28 24 23 22 21 20 16 15 14 13 11 10 9 5 4 0 + * +-----+---+-----------+---+-----+------+-----+--------+---+----+----+ + * | 0 1 | U | 1 1 1 1 0 | a | 1 0 | Rm | 0 0 | opcode | 1 | Rn | Rd | + * +-----+---+-----------+---+-----+------+-----+--------+---+----+----+ + * v: 0101 1110 0100 0000 0000 0100 0000 0000 => 5e400400 + * m: 1101 1111 0110 0000 1100 0100 0000 0000 => df60c400 + */ +static void disas_simd_scalar_three_reg_same_fp16(DisasContext *s, + uint32_t insn) +{ + int rd = extract32(insn, 0, 5); + int rn = extract32(insn, 5, 5); + int opcode = extract32(insn, 11, 3); + int rm = extract32(insn, 16, 5); + bool u = extract32(insn, 29, 1); + bool a = extract32(insn, 23, 1); + int fpopcode = opcode | (a << 3) | (u << 4); + TCGv_ptr fpst; + TCGv_i32 tcg_op1; + TCGv_i32 tcg_op2; + TCGv_i32 tcg_res; + + switch (fpopcode) { + case 0x03: /* FMULX */ + case 0x04: /* FCMEQ (reg) */ + case 0x07: /* FRECPS */ + case 0x0f: /* FRSQRTS */ + case 0x14: /* FCMGE (reg) */ + case 0x15: /* FACGE */ + case 0x1a: /* FABD */ + case 0x1c: /* FCMGT (reg) */ + case 0x1d: /* FACGT */ + break; + default: + unallocated_encoding(s); + return; + } + + if (!arm_dc_feature(s, ARM_FEATURE_V8_FP16)) { + unallocated_encoding(s); + } + + if (!fp_access_check(s)) { + return; + } + + fpst = get_fpstatus_ptr(true); + + tcg_op1 = tcg_temp_new_i32(); + tcg_op2 = tcg_temp_new_i32(); + tcg_res = tcg_temp_new_i32(); + + read_vec_element_i32(s, tcg_op1, rn, 0, MO_16); + read_vec_element_i32(s, tcg_op2, rm, 0, MO_16); + + switch (fpopcode) { + case 0x03: /* FMULX */ + gen_helper_advsimd_mulxh(tcg_res, tcg_op1, tcg_op2, fpst); + break; + case 0x04: /* FCMEQ (reg) */ + gen_helper_advsimd_ceq_f16(tcg_res, tcg_op1, tcg_op2, fpst); + break; + case 0x07: /* FRECPS */ + gen_helper_recpsf_f16(tcg_res, tcg_op1, tcg_op2, fpst); + break; + case 0x0f: /* FRSQRTS */ + gen_helper_rsqrtsf_f16(tcg_res, tcg_op1, tcg_op2, fpst); + break; + case 0x14: /* FCMGE (reg) */ + gen_helper_advsimd_cge_f16(tcg_res, tcg_op1, tcg_op2, fpst); + break; + case 0x15: /* FACGE */ + gen_helper_advsimd_acge_f16(tcg_res, tcg_op1, tcg_op2, fpst); + break; + case 0x1a: /* FABD */ + gen_helper_advsimd_subh(tcg_res, tcg_op1, tcg_op2, fpst); + tcg_gen_andi_i32(tcg_res, tcg_res, 0x7fff); + break; + case 0x1c: /* FCMGT (reg) */ + gen_helper_advsimd_cgt_f16(tcg_res, tcg_op1, tcg_op2, fpst); + break; + case 0x1d: /* FACGT */ + gen_helper_advsimd_acgt_f16(tcg_res, tcg_op1, tcg_op2, fpst); + break; + default: + g_assert_not_reached(); + } + + write_fp_sreg(s, rd, tcg_res); + + + tcg_temp_free_i32(tcg_res); + tcg_temp_free_i32(tcg_op1); + tcg_temp_free_i32(tcg_op2); + tcg_temp_free_ptr(fpst); +} + static void handle_2misc_64(DisasContext *s, int opcode, bool u, TCGv_i64 tcg_rd, TCGv_i64 tcg_rn, TCGv_i32 tcg_rmode, TCGv_ptr tcg_fpstatus) @@ -XXX,XX +XXX,XX @@ static const AArch64DecodeTable data_proc_simd[] = { { 0xce408000, 0xffe0c000, disas_crypto_three_reg_imm2 }, { 0x0e400400, 0x9f60c400, disas_simd_three_reg_same_fp16 }, { 0x0e780800, 0x8f7e0c00, disas_simd_two_reg_misc_fp16 }, + { 0x5e400400, 0xdf60c400, disas_simd_scalar_three_reg_same_fp16 }, { 0x00000000, 0x00000000, NULL } }; -- 2.16.2 From: Alex Bennée <alex.bennee@linaro.org> This includes FMOV, FABS, FNEG, FSQRT and FRINT[NPMZAXI]. We re-use existing helpers to achieve this. Signed-off-by: Alex Bennée <alex.bennee@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20180227143852.11175-32-alex.bennee@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/translate-a64.c | 71 ++++++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 71 insertions(+) diff --git a/target/arm/translate-a64.c b/target/arm/translate-a64.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-a64.c +++ b/target/arm/translate-a64.c @@ -XXX,XX +XXX,XX @@ static void disas_fp_csel(DisasContext *s, uint32_t insn) tcg_temp_free_i64(t_true); } +/* Floating-point data-processing (1 source) - half precision */ +static void handle_fp_1src_half(DisasContext *s, int opcode, int rd, int rn) +{ + TCGv_ptr fpst = NULL; + TCGv_i32 tcg_op = tcg_temp_new_i32(); + TCGv_i32 tcg_res = tcg_temp_new_i32(); + + read_vec_element_i32(s, tcg_op, rn, 0, MO_16); + + switch (opcode) { + case 0x0: /* FMOV */ + tcg_gen_mov_i32(tcg_res, tcg_op); + break; + case 0x1: /* FABS */ + tcg_gen_andi_i32(tcg_res, tcg_op, 0x7fff); + break; + case 0x2: /* FNEG */ + tcg_gen_xori_i32(tcg_res, tcg_op, 0x8000); + break; + case 0x3: /* FSQRT */ + gen_helper_sqrt_f16(tcg_res, tcg_op, cpu_env); + break; + case 0x8: /* FRINTN */ + case 0x9: /* FRINTP */ + case 0xa: /* FRINTM */ + case 0xb: /* FRINTZ */ + case 0xc: /* FRINTA */ + { + TCGv_i32 tcg_rmode = tcg_const_i32(arm_rmode_to_sf(opcode & 7)); + fpst = get_fpstatus_ptr(true); + + gen_helper_set_rmode(tcg_rmode, tcg_rmode, fpst); + gen_helper_advsimd_rinth(tcg_res, tcg_op, fpst); + + gen_helper_set_rmode(tcg_rmode, tcg_rmode, fpst); + tcg_temp_free_i32(tcg_rmode); + break; + } + case 0xe: /* FRINTX */ + fpst = get_fpstatus_ptr(true); + gen_helper_advsimd_rinth_exact(tcg_res, tcg_op, fpst); + break; + case 0xf: /* FRINTI */ + fpst = get_fpstatus_ptr(true); + gen_helper_advsimd_rinth(tcg_res, tcg_op, fpst); + break; + default: + abort(); + } + + write_fp_sreg(s, rd, tcg_res); + + if (fpst) { + tcg_temp_free_ptr(fpst); + } + tcg_temp_free_i32(tcg_op); + tcg_temp_free_i32(tcg_res); +} + /* Floating-point data-processing (1 source) - single precision */ static void handle_fp_1src_single(DisasContext *s, int opcode, int rd, int rn) { @@ -XXX,XX +XXX,XX @@ static void disas_fp_1src(DisasContext *s, uint32_t insn) handle_fp_1src_double(s, opcode, rd, rn); break; + case 3: + if (!arm_dc_feature(s, ARM_FEATURE_V8_FP16)) { + unallocated_encoding(s); + return; + } + + if (!fp_access_check(s)) { + return; + } + + handle_fp_1src_half(s, opcode, rd, rn); + break; default: unallocated_encoding(s); } -- 2.16.2 Now we have implemented FP16 we can enable it for the "any" CPU. Signed-off-by: Alex Bennée <alex.bennee@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> [PMM: split out from an earlier patch in the series] Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/cpu64.c | 1 + 1 file changed, 1 insertion(+) 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_any_initfn(Object *obj) set_feature(&cpu->env, ARM_FEATURE_V8_SM4); set_feature(&cpu->env, ARM_FEATURE_V8_PMULL); set_feature(&cpu->env, ARM_FEATURE_CRC); + set_feature(&cpu->env, ARM_FEATURE_V8_FP16); cpu->ctr = 0x80038003; /* 32 byte I and D cacheline size, VIPT icache */ cpu->dcz_blocksize = 7; /* 512 bytes */ } -- 2.16.2 Set the appropriate Linux hwcap bits to tell the guest binary if we have implemented half-precision floating point support. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> --- linux-user/elfload.c | 2 ++ 1 file changed, 2 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 @@ static uint32_t get_elf_hwcap(void) GET_FEATURE(ARM_FEATURE_V8_SM3, ARM_HWCAP_A64_SM3); GET_FEATURE(ARM_FEATURE_V8_SM4, ARM_HWCAP_A64_SM4); GET_FEATURE(ARM_FEATURE_V8_SHA512, ARM_HWCAP_A64_SHA512); + GET_FEATURE(ARM_FEATURE_V8_FP16, + ARM_HWCAP_A64_FPHP | ARM_HWCAP_A64_ASIMDHP); #undef GET_FEATURE return hwcaps; -- 2.16.2 From: Alistair Francis <alistair.francis@xilinx.com> I am leaving Xilinx, so to avoid having an email address that bounces update my maintainer address to point to my personal email address. Signed-off-by: Alistair Francis <alistair.francis@xilinx.com> Signed-off-by: Alistair Francis <alistair@alistair23.me> Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Message-id: 7bb690382e3370aa1c1e047a84e36603c787ec0e.1519749987.git.alistair.francis@xilinx.com Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- MAINTAINERS | 12 ++++++------ 1 file changed, 6 insertions(+), 6 deletions(-) diff --git a/MAINTAINERS b/MAINTAINERS index XXXXXXX..XXXXXXX 100644 --- a/MAINTAINERS +++ b/MAINTAINERS @@ -XXX,XX +XXX,XX @@ F: hw/misc/arm_sysctl.c Xilinx Zynq M: Edgar E. Iglesias <edgar.iglesias@gmail.com> -M: Alistair Francis <alistair.francis@xilinx.com> +M: Alistair Francis <alistair@alistair23.me> L: qemu-arm@nongnu.org S: Maintained F: hw/*/xilinx_* @@ -XXX,XX +XXX,XX @@ F: include/hw/misc/zynq* X: hw/ssi/xilinx_* Xilinx ZynqMP -M: Alistair Francis <alistair.francis@xilinx.com> +M: Alistair Francis <alistair@alistair23.me> M: Edgar E. Iglesias <edgar.iglesias@gmail.com> L: qemu-arm@nongnu.org S: Maintained @@ -XXX,XX +XXX,XX @@ T: git git://github.com/bonzini/qemu.git scsi-next SSI M: Peter Crosthwaite <crosthwaite.peter@gmail.com> -M: Alistair Francis <alistair.francis@xilinx.com> +M: Alistair Francis <alistair@alistair23.me> S: Maintained F: hw/ssi/* F: hw/block/m25p80.c @@ -XXX,XX +XXX,XX @@ X: hw/ssi/xilinx_* F: tests/m25p80-test.c Xilinx SPI -M: Alistair Francis <alistair.francis@xilinx.com> +M: Alistair Francis <alistair@alistair23.me> M: Peter Crosthwaite <crosthwaite.peter@gmail.com> S: Maintained F: hw/ssi/xilinx_* @@ -XXX,XX +XXX,XX @@ S: Maintained F: hw/net/eepro100.c Generic Loader -M: Alistair Francis <alistair.francis@xilinx.com> +M: Alistair Francis <alistair@alistair23.me> S: Maintained F: hw/core/generic-loader.c F: include/hw/core/generic-loader.h @@ -XXX,XX +XXX,XX @@ F: tests/qmp-test.c T: git git://repo.or.cz/qemu/armbru.git qapi-next Register API -M: Alistair Francis <alistair.francis@xilinx.com> +M: Alistair Francis <alistair@alistair23.me> S: Maintained F: hw/core/register.c F: include/hw/register.h -- 2.16.2 Most of this is the Neon decodetree patches, followed by Edgar's versal cleanups. thanks -- PMM The following changes since commit 2ef486e76d64436be90f7359a3071fb2a56ce835: Merge remote-tracking branch 'remotes/marcel/tags/rdma-pull-request' into staging (2020-05-03 14:12:56 +0100) are available in the Git repository at: https://git.linaro.org/people/pmaydell/qemu-arm.git tags/pull-target-arm-20200504 for you to fetch changes up to 9aefc6cf9b73f66062d2f914a0136756e7a28211: target/arm: Move gen_ function typedefs to translate.h (2020-05-04 12:59:26 +0100) ---------------------------------------------------------------- target-arm queue: * Start of conversion of Neon insns to decodetree * versal board: support SD and RTC * Implement ARMv8.2-TTS2UXN * Make VQDMULL undefined when U=1 * Some minor code cleanups ---------------------------------------------------------------- Edgar E. Iglesias (11): hw/arm: versal: Remove inclusion of arm_gicv3_common.h hw/arm: versal: Move misplaced comment hw/arm: versal-virt: Fix typo xlnx-ve -> xlnx-versal hw/arm: versal: Embed the UARTs into the SoC type hw/arm: versal: Embed the GEMs into the SoC type hw/arm: versal: Embed the ADMAs into the SoC type hw/arm: versal: Embed the APUs into the SoC type hw/arm: versal: Add support for SD hw/arm: versal: Add support for the RTC hw/arm: versal-virt: Add support for SD hw/arm: versal-virt: Add support for the RTC Fredrik Strupe (1): target/arm: Make VQDMULL undefined when U=1 Peter Maydell (25): target/arm: Don't use a TLB for ARMMMUIdx_Stage2 target/arm: Use enum constant in get_phys_addr_lpae() call target/arm: Add new 's1_is_el0' argument to get_phys_addr_lpae() target/arm: Implement ARMv8.2-TTS2UXN target/arm: Use correct variable for setting 'max' cpu's ID_AA64DFR0 target/arm/translate-vfp.inc.c: Remove duplicate simd_r32 check target/arm: Don't allow Thumb Neon insns without FEATURE_NEON target/arm: Add stubs for AArch32 Neon decodetree target/arm: Convert VCMLA (vector) to decodetree target/arm: Convert VCADD (vector) to decodetree target/arm: Convert V[US]DOT (vector) to decodetree target/arm: Convert VFM[AS]L (vector) to decodetree target/arm: Convert VCMLA (scalar) to decodetree target/arm: Convert V[US]DOT (scalar) to decodetree target/arm: Convert VFM[AS]L (scalar) to decodetree target/arm: Convert Neon load/store multiple structures to decodetree target/arm: Convert Neon 'load single structure to all lanes' to decodetree target/arm: Convert Neon 'load/store single structure' to decodetree target/arm: Convert Neon 3-reg-same VADD/VSUB to decodetree target/arm: Convert Neon 3-reg-same logic ops to decodetree target/arm: Convert Neon 3-reg-same VMAX/VMIN to decodetree target/arm: Convert Neon 3-reg-same comparisons to decodetree target/arm: Convert Neon 3-reg-same VQADD/VQSUB to decodetree target/arm: Convert Neon 3-reg-same VMUL, VMLA, VMLS, VSHL to decodetree target/arm: Move gen_ function typedefs to translate.h Philippe Mathieu-Daudé (2): hw/arm/mps2-tz: Use TYPE_IOTKIT instead of hardcoded string target/arm: Use uint64_t for midr field in CPU state struct include/hw/arm/xlnx-versal.h | 31 +- target/arm/cpu-param.h | 2 +- target/arm/cpu.h | 38 ++- target/arm/translate-a64.h | 9 - target/arm/translate.h | 26 ++ target/arm/neon-dp.decode | 86 +++++ target/arm/neon-ls.decode | 52 +++ target/arm/neon-shared.decode | 66 ++++ hw/arm/mps2-tz.c | 2 +- hw/arm/xlnx-versal-virt.c | 74 ++++- hw/arm/xlnx-versal.c | 115 +++++-- target/arm/cpu.c | 3 +- target/arm/cpu64.c | 8 +- target/arm/helper.c | 183 ++++------ target/arm/translate-a64.c | 17 - target/arm/translate-neon.inc.c | 714 +++++++++++++++++++++++++++++++++++++++ target/arm/translate-vfp.inc.c | 6 - target/arm/translate.c | 716 +++------------------------------------- target/arm/Makefile.objs | 18 + 19 files changed, 1302 insertions(+), 864 deletions(-) create mode 100644 target/arm/neon-dp.decode create mode 100644 target/arm/neon-ls.decode create mode 100644 target/arm/neon-shared.decode create mode 100644 target/arm/translate-neon.inc.c From: Fredrik Strupe <fredrik@strupe.net> According to Arm ARM, VQDMULL is only valid when U=0, while having U=1 is unallocated. Signed-off-by: Fredrik Strupe <fredrik@strupe.net> Fixes: 695272dcb976 ("target-arm: Handle UNDEF cases for Neon 3-regs-different-widths") Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/translate.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) 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 int disas_neon_data_insn(DisasContext *s, uint32_t insn) {0, 0, 0, 0}, /* VMLSL */ {0, 0, 0, 9}, /* VQDMLSL */ {0, 0, 0, 0}, /* Integer VMULL */ - {0, 0, 0, 1}, /* VQDMULL */ + {0, 0, 0, 9}, /* VQDMULL */ {0, 0, 0, 0xa}, /* Polynomial VMULL */ {0, 0, 0, 7}, /* Reserved: always UNDEF */ }; -- 2.20.1 From: Philippe Mathieu-Daudé <f4bug@amsat.org> By using the TYPE_* definitions for devices, we can: - quickly find where devices are used with 'git-grep' - easily rename a device (one-line change). Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Message-id: 20200428154650.21991-1-f4bug@amsat.org Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- hw/arm/mps2-tz.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/hw/arm/mps2-tz.c b/hw/arm/mps2-tz.c index XXXXXXX..XXXXXXX 100644 --- a/hw/arm/mps2-tz.c +++ b/hw/arm/mps2-tz.c @@ -XXX,XX +XXX,XX @@ static void mps2tz_common_init(MachineState *machine) exit(EXIT_FAILURE); } - sysbus_init_child_obj(OBJECT(machine), "iotkit", &mms->iotkit, + sysbus_init_child_obj(OBJECT(machine), TYPE_IOTKIT, &mms->iotkit, sizeof(mms->iotkit), mmc->armsse_type); iotkitdev = DEVICE(&mms->iotkit); object_property_set_link(OBJECT(&mms->iotkit), OBJECT(system_memory), -- 2.20.1 We define ARMMMUIdx_Stage2 as being an MMU index which uses a QEMU TLB. However we never actually use the TLB -- all stage 2 lookups are done by direct calls to get_phys_addr_lpae() followed by a physical address load via address_space_ld*(). Remove Stage2 from the list of ARM MMU indexes which correspond to real core MMU indexes, and instead put it in the set of "NOTLB" ARM MMU indexes. This allows us to drop NB_MMU_MODES to 11. It also means we can safely add support for the ARMv8.3-TTS2UXN extension, which adds permission bits to the stage 2 descriptors which define execute permission separatel for EL0 and EL1; supporting that while keeping Stage2 in a QEMU TLB would require us to use separate TLBs for "Stage2 for an EL0 access" and "Stage2 for an EL1 access", which is a lot of extra complication given we aren't even using the QEMU TLB. In the process of updating the comment on our MMU index use, fix a couple of other minor errors: * NS EL2 EL2&0 was missing from the list in the comment * some text hadn't been updated from when we bumped NB_MMU_MODES above 8 Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200330210400.11724-2-peter.maydell@linaro.org --- target/arm/cpu-param.h | 2 +- target/arm/cpu.h | 21 +++++--- target/arm/helper.c | 112 ++++------------------------------------- 3 files changed, 27 insertions(+), 108 deletions(-) diff --git a/target/arm/cpu-param.h b/target/arm/cpu-param.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/cpu-param.h +++ b/target/arm/cpu-param.h @@ -XXX,XX +XXX,XX @@ # define TARGET_PAGE_BITS_MIN 10 #endif -#define NB_MMU_MODES 12 +#define NB_MMU_MODES 11 #endif 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 @@ bool write_cpustate_to_list(ARMCPU *cpu, bool kvm_sync); * handling via the TLB. The only way to do a stage 1 translation without * the immediate stage 2 translation is via the ATS or AT system insns, * which can be slow-pathed and always do a page table walk. + * The only use of stage 2 translations is either as part of an s1+2 + * lookup or when loading the descriptors during a stage 1 page table walk, + * and in both those cases we don't use the TLB. * 4. we can also safely fold together the "32 bit EL3" and "64 bit EL3" * translation regimes, because they map reasonably well to each other * and they can't both be active at the same time. @@ -XXX,XX +XXX,XX @@ bool write_cpustate_to_list(ARMCPU *cpu, bool kvm_sync); * NS EL1 EL1&0 stage 1+2 (aka NS PL1) * NS EL1 EL1&0 stage 1+2 +PAN * NS EL0 EL2&0 + * NS EL2 EL2&0 * NS EL2 EL2&0 +PAN * NS EL2 (aka NS PL2) * S EL0 EL1&0 (aka S PL0) * S EL1 EL1&0 (not used if EL3 is 32 bit) * S EL1 EL1&0 +PAN * S EL3 (aka S PL1) - * NS EL1&0 stage 2 * - * for a total of 12 different mmu_idx. + * for a total of 11 different mmu_idx. * * R profile CPUs have an MPU, but can use the same set of MMU indexes * as A profile. They only need to distinguish NS EL0 and NS EL1 (and @@ -XXX,XX +XXX,XX @@ bool write_cpustate_to_list(ARMCPU *cpu, bool kvm_sync); * are not quite the same -- different CPU types (most notably M profile * vs A/R profile) would like to use MMU indexes with different semantics, * but since we don't ever need to use all of those in a single CPU we - * can avoid setting NB_MMU_MODES to more than 8. The lower bits of + * can avoid having to set NB_MMU_MODES to "total number of A profile MMU + * modes + total number of M profile MMU modes". The lower bits of * ARMMMUIdx are the core TLB mmu index, and the higher bits are always * the same for any particular CPU. * Variables of type ARMMUIdx are always full values, and the core @@ -XXX,XX +XXX,XX @@ typedef enum ARMMMUIdx { ARMMMUIdx_SE10_1_PAN = 9 | ARM_MMU_IDX_A, ARMMMUIdx_SE3 = 10 | ARM_MMU_IDX_A, - ARMMMUIdx_Stage2 = 11 | ARM_MMU_IDX_A, - /* * These are not allocated TLBs and are used only for AT system * instructions or for the first stage of an S12 page table walk. @@ -XXX,XX +XXX,XX @@ typedef enum ARMMMUIdx { ARMMMUIdx_Stage1_E0 = 0 | ARM_MMU_IDX_NOTLB, ARMMMUIdx_Stage1_E1 = 1 | ARM_MMU_IDX_NOTLB, ARMMMUIdx_Stage1_E1_PAN = 2 | ARM_MMU_IDX_NOTLB, + /* + * Not allocated a TLB: used only for second stage of an S12 page + * table walk, or for descriptor loads during first stage of an S1 + * page table walk. Note that if we ever want to have a TLB for this + * then various TLB flush insns which currently are no-ops or flush + * only stage 1 MMU indexes will need to change to flush stage 2. + */ + ARMMMUIdx_Stage2 = 3 | ARM_MMU_IDX_NOTLB, /* * M-profile. @@ -XXX,XX +XXX,XX @@ typedef enum ARMMMUIdxBit { TO_CORE_BIT(SE10_1), TO_CORE_BIT(SE10_1_PAN), TO_CORE_BIT(SE3), - TO_CORE_BIT(Stage2), TO_CORE_BIT(MUser), TO_CORE_BIT(MPriv), diff --git a/target/arm/helper.c b/target/arm/helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper.c +++ b/target/arm/helper.c @@ -XXX,XX +XXX,XX @@ static void tlbiall_nsnh_write(CPUARMState *env, const ARMCPRegInfo *ri, tlb_flush_by_mmuidx(cs, ARMMMUIdxBit_E10_1 | ARMMMUIdxBit_E10_1_PAN | - ARMMMUIdxBit_E10_0 | - ARMMMUIdxBit_Stage2); + ARMMMUIdxBit_E10_0); } static void tlbiall_nsnh_is_write(CPUARMState *env, const ARMCPRegInfo *ri, @@ -XXX,XX +XXX,XX @@ static void tlbiall_nsnh_is_write(CPUARMState *env, const ARMCPRegInfo *ri, tlb_flush_by_mmuidx_all_cpus_synced(cs, ARMMMUIdxBit_E10_1 | ARMMMUIdxBit_E10_1_PAN | - ARMMMUIdxBit_E10_0 | - ARMMMUIdxBit_Stage2); + ARMMMUIdxBit_E10_0); } -static void tlbiipas2_write(CPUARMState *env, const ARMCPRegInfo *ri, - uint64_t value) -{ - /* Invalidate by IPA. This has to invalidate any structures that - * contain only stage 2 translation information, but does not need - * to apply to structures that contain combined stage 1 and stage 2 - * translation information. - * This must NOP if EL2 isn't implemented or SCR_EL3.NS is zero. - */ - CPUState *cs = env_cpu(env); - uint64_t pageaddr; - - if (!arm_feature(env, ARM_FEATURE_EL2) || !(env->cp15.scr_el3 & SCR_NS)) { - return; - } - - pageaddr = sextract64(value << 12, 0, 40); - - tlb_flush_page_by_mmuidx(cs, pageaddr, ARMMMUIdxBit_Stage2); -} - -static void tlbiipas2_is_write(CPUARMState *env, const ARMCPRegInfo *ri, - uint64_t value) -{ - CPUState *cs = env_cpu(env); - uint64_t pageaddr; - - if (!arm_feature(env, ARM_FEATURE_EL2) || !(env->cp15.scr_el3 & SCR_NS)) { - return; - } - - pageaddr = sextract64(value << 12, 0, 40); - - tlb_flush_page_by_mmuidx_all_cpus_synced(cs, pageaddr, - ARMMMUIdxBit_Stage2); -} static void tlbiall_hyp_write(CPUARMState *env, const ARMCPRegInfo *ri, uint64_t value) @@ -XXX,XX +XXX,XX @@ static void vttbr_write(CPUARMState *env, const ARMCPRegInfo *ri, tlb_flush_by_mmuidx(cs, ARMMMUIdxBit_E10_1 | ARMMMUIdxBit_E10_1_PAN | - ARMMMUIdxBit_E10_0 | - ARMMMUIdxBit_Stage2); + ARMMMUIdxBit_E10_0); raw_write(env, ri, value); } } @@ -XXX,XX +XXX,XX @@ static int alle1_tlbmask(CPUARMState *env) return ARMMMUIdxBit_SE10_1 | ARMMMUIdxBit_SE10_1_PAN | ARMMMUIdxBit_SE10_0; - } else if (arm_feature(env, ARM_FEATURE_EL2)) { - return ARMMMUIdxBit_E10_1 | - ARMMMUIdxBit_E10_1_PAN | - ARMMMUIdxBit_E10_0 | - ARMMMUIdxBit_Stage2; } else { return ARMMMUIdxBit_E10_1 | ARMMMUIdxBit_E10_1_PAN | @@ -XXX,XX +XXX,XX @@ static void tlbi_aa64_vae3is_write(CPUARMState *env, const ARMCPRegInfo *ri, ARMMMUIdxBit_SE3); } -static void tlbi_aa64_ipas2e1_write(CPUARMState *env, const ARMCPRegInfo *ri, - uint64_t value) -{ - /* Invalidate by IPA. This has to invalidate any structures that - * contain only stage 2 translation information, but does not need - * to apply to structures that contain combined stage 1 and stage 2 - * translation information. - * This must NOP if EL2 isn't implemented or SCR_EL3.NS is zero. - */ - ARMCPU *cpu = env_archcpu(env); - CPUState *cs = CPU(cpu); - uint64_t pageaddr; - - if (!arm_feature(env, ARM_FEATURE_EL2) || !(env->cp15.scr_el3 & SCR_NS)) { - return; - } - - pageaddr = sextract64(value << 12, 0, 48); - - tlb_flush_page_by_mmuidx(cs, pageaddr, ARMMMUIdxBit_Stage2); -} - -static void tlbi_aa64_ipas2e1is_write(CPUARMState *env, const ARMCPRegInfo *ri, - uint64_t value) -{ - CPUState *cs = env_cpu(env); - uint64_t pageaddr; - - if (!arm_feature(env, ARM_FEATURE_EL2) || !(env->cp15.scr_el3 & SCR_NS)) { - return; - } - - pageaddr = sextract64(value << 12, 0, 48); - - tlb_flush_page_by_mmuidx_all_cpus_synced(cs, pageaddr, - ARMMMUIdxBit_Stage2); -} - static CPAccessResult aa64_zva_access(CPUARMState *env, const ARMCPRegInfo *ri, bool isread) { @@ -XXX,XX +XXX,XX @@ static const ARMCPRegInfo v8_cp_reginfo[] = { .writefn = tlbi_aa64_vae1_write }, { .name = "TLBI_IPAS2E1IS", .state = ARM_CP_STATE_AA64, .opc0 = 1, .opc1 = 4, .crn = 8, .crm = 0, .opc2 = 1, - .access = PL2_W, .type = ARM_CP_NO_RAW, - .writefn = tlbi_aa64_ipas2e1is_write }, + .access = PL2_W, .type = ARM_CP_NOP }, { .name = "TLBI_IPAS2LE1IS", .state = ARM_CP_STATE_AA64, .opc0 = 1, .opc1 = 4, .crn = 8, .crm = 0, .opc2 = 5, - .access = PL2_W, .type = ARM_CP_NO_RAW, - .writefn = tlbi_aa64_ipas2e1is_write }, + .access = PL2_W, .type = ARM_CP_NOP }, { .name = "TLBI_ALLE1IS", .state = ARM_CP_STATE_AA64, .opc0 = 1, .opc1 = 4, .crn = 8, .crm = 3, .opc2 = 4, .access = PL2_W, .type = ARM_CP_NO_RAW, @@ -XXX,XX +XXX,XX @@ static const ARMCPRegInfo v8_cp_reginfo[] = { .writefn = tlbi_aa64_alle1is_write }, { .name = "TLBI_IPAS2E1", .state = ARM_CP_STATE_AA64, .opc0 = 1, .opc1 = 4, .crn = 8, .crm = 4, .opc2 = 1, - .access = PL2_W, .type = ARM_CP_NO_RAW, - .writefn = tlbi_aa64_ipas2e1_write }, + .access = PL2_W, .type = ARM_CP_NOP }, { .name = "TLBI_IPAS2LE1", .state = ARM_CP_STATE_AA64, .opc0 = 1, .opc1 = 4, .crn = 8, .crm = 4, .opc2 = 5, - .access = PL2_W, .type = ARM_CP_NO_RAW, - .writefn = tlbi_aa64_ipas2e1_write }, + .access = PL2_W, .type = ARM_CP_NOP }, { .name = "TLBI_ALLE1", .state = ARM_CP_STATE_AA64, .opc0 = 1, .opc1 = 4, .crn = 8, .crm = 7, .opc2 = 4, .access = PL2_W, .type = ARM_CP_NO_RAW, @@ -XXX,XX +XXX,XX @@ static const ARMCPRegInfo v8_cp_reginfo[] = { .writefn = tlbimva_hyp_is_write }, { .name = "TLBIIPAS2", .cp = 15, .opc1 = 4, .crn = 8, .crm = 4, .opc2 = 1, - .type = ARM_CP_NO_RAW, .access = PL2_W, - .writefn = tlbiipas2_write }, + .type = ARM_CP_NOP, .access = PL2_W }, { .name = "TLBIIPAS2IS", .cp = 15, .opc1 = 4, .crn = 8, .crm = 0, .opc2 = 1, - .type = ARM_CP_NO_RAW, .access = PL2_W, - .writefn = tlbiipas2_is_write }, + .type = ARM_CP_NOP, .access = PL2_W }, { .name = "TLBIIPAS2L", .cp = 15, .opc1 = 4, .crn = 8, .crm = 4, .opc2 = 5, - .type = ARM_CP_NO_RAW, .access = PL2_W, - .writefn = tlbiipas2_write }, + .type = ARM_CP_NOP, .access = PL2_W }, { .name = "TLBIIPAS2LIS", .cp = 15, .opc1 = 4, .crn = 8, .crm = 0, .opc2 = 5, - .type = ARM_CP_NO_RAW, .access = PL2_W, - .writefn = tlbiipas2_is_write }, + .type = ARM_CP_NOP, .access = PL2_W }, /* 32 bit cache operations */ { .name = "ICIALLUIS", .cp = 15, .opc1 = 0, .crn = 7, .crm = 1, .opc2 = 0, .type = ARM_CP_NOP, .access = PL1_W, .accessfn = aa64_cacheop_pou_access }, -- 2.20.1 The access_type argument to get_phys_addr_lpae() is an MMUAccessType; use the enum constant MMU_DATA_LOAD rather than a literal 0 when we call it in S1_ptw_translate(). Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200330210400.11724-3-peter.maydell@linaro.org --- target/arm/helper.c | 5 +++-- 1 file changed, 3 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 @@ static hwaddr S1_ptw_translate(CPUARMState *env, ARMMMUIdx mmu_idx, pcacheattrs = &cacheattrs; } - ret = get_phys_addr_lpae(env, addr, 0, ARMMMUIdx_Stage2, &s2pa, - &txattrs, &s2prot, &s2size, fi, pcacheattrs); + ret = get_phys_addr_lpae(env, addr, MMU_DATA_LOAD, ARMMMUIdx_Stage2, + &s2pa, &txattrs, &s2prot, &s2size, fi, + pcacheattrs); if (ret) { assert(fi->type != ARMFault_None); fi->s2addr = addr; -- 2.20.1 For ARMv8.2-TTS2UXN, the stage 2 page table walk wants to know whether the stage 1 access is for EL0 or not, because whether exec permission is given can depend on whether this is an EL0 or EL1 access. Add a new argument to get_phys_addr_lpae() so the call sites can pass this information in. Since get_phys_addr_lpae() doesn't already have a doc comment, add one so we have a place to put the documentation of the semantics of the new s1_is_el0 argument. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200330210400.11724-4-peter.maydell@linaro.org --- target/arm/helper.c | 29 ++++++++++++++++++++++++++++- 1 file changed, 28 insertions(+), 1 deletion(-) diff --git a/target/arm/helper.c b/target/arm/helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper.c +++ b/target/arm/helper.c @@ -XXX,XX +XXX,XX @@ static bool get_phys_addr_lpae(CPUARMState *env, target_ulong address, MMUAccessType access_type, ARMMMUIdx mmu_idx, + bool s1_is_el0, hwaddr *phys_ptr, MemTxAttrs *txattrs, int *prot, target_ulong *page_size_ptr, ARMMMUFaultInfo *fi, ARMCacheAttrs *cacheattrs); @@ -XXX,XX +XXX,XX @@ static hwaddr S1_ptw_translate(CPUARMState *env, ARMMMUIdx mmu_idx, } ret = get_phys_addr_lpae(env, addr, MMU_DATA_LOAD, ARMMMUIdx_Stage2, + false, &s2pa, &txattrs, &s2prot, &s2size, fi, pcacheattrs); if (ret) { @@ -XXX,XX +XXX,XX @@ static ARMVAParameters aa32_va_parameters(CPUARMState *env, uint32_t va, }; } +/** + * get_phys_addr_lpae: perform one stage of page table walk, LPAE format + * + * Returns false if the translation was successful. Otherwise, phys_ptr, attrs, + * prot and page_size may not be filled in, and the populated fsr value provides + * information on why the translation aborted, in the format of a long-format + * DFSR/IFSR fault register, with the following caveats: + * * the WnR bit is never set (the caller must do this). + * + * @env: CPUARMState + * @address: virtual address to get physical address for + * @access_type: MMU_DATA_LOAD, MMU_DATA_STORE or MMU_INST_FETCH + * @mmu_idx: MMU index indicating required translation regime + * @s1_is_el0: if @mmu_idx is ARMMMUIdx_Stage2 (so this is a stage 2 page table + * walk), must be true if this is stage 2 of a stage 1+2 walk for an + * EL0 access). If @mmu_idx is anything else, @s1_is_el0 is ignored. + * @phys_ptr: set to the physical address corresponding to the virtual address + * @attrs: set to the memory transaction attributes to use + * @prot: set to the permissions for the page containing phys_ptr + * @page_size_ptr: set to the size of the page containing phys_ptr + * @fi: set to fault info if the translation fails + * @cacheattrs: (if non-NULL) set to the cacheability/shareability attributes + */ static bool get_phys_addr_lpae(CPUARMState *env, target_ulong address, MMUAccessType access_type, ARMMMUIdx mmu_idx, + bool s1_is_el0, hwaddr *phys_ptr, MemTxAttrs *txattrs, int *prot, target_ulong *page_size_ptr, ARMMMUFaultInfo *fi, ARMCacheAttrs *cacheattrs) @@ -XXX,XX +XXX,XX @@ bool get_phys_addr(CPUARMState *env, target_ulong address, /* S1 is done. Now do S2 translation. */ ret = get_phys_addr_lpae(env, ipa, access_type, ARMMMUIdx_Stage2, + mmu_idx == ARMMMUIdx_E10_0, phys_ptr, attrs, &s2_prot, page_size, fi, cacheattrs != NULL ? &cacheattrs2 : NULL); @@ -XXX,XX +XXX,XX @@ bool get_phys_addr(CPUARMState *env, target_ulong address, } if (regime_using_lpae_format(env, mmu_idx)) { - return get_phys_addr_lpae(env, address, access_type, mmu_idx, + return get_phys_addr_lpae(env, address, access_type, mmu_idx, false, phys_ptr, attrs, prot, page_size, fi, cacheattrs); } else if (regime_sctlr(env, mmu_idx) & SCTLR_XP) { -- 2.20.1 The ARMv8.2-TTS2UXN feature extends the XN field in stage 2 translation table descriptors from just bit [54] to bits [54:53], allowing stage 2 to control execution permissions separately for EL0 and EL1. Implement the new semantics of the XN field and enable the feature for our 'max' CPU. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200330210400.11724-5-peter.maydell@linaro.org --- target/arm/cpu.h | 15 +++++++++++++++ target/arm/cpu.c | 1 + target/arm/cpu64.c | 2 ++ target/arm/helper.c | 37 +++++++++++++++++++++++++++++++------ 4 files changed, 49 insertions(+), 6 deletions(-) diff --git a/target/arm/cpu.h b/target/arm/cpu.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/cpu.h +++ b/target/arm/cpu.h @@ -XXX,XX +XXX,XX @@ static inline bool isar_feature_aa32_ccidx(const ARMISARegisters *id) return FIELD_EX32(id->id_mmfr4, ID_MMFR4, CCIDX) != 0; } +static inline bool isar_feature_aa32_tts2uxn(const ARMISARegisters *id) +{ + return FIELD_EX32(id->id_mmfr4, ID_MMFR4, XNX) != 0; +} + /* * 64-bit feature tests via id registers. */ @@ -XXX,XX +XXX,XX @@ static inline bool isar_feature_aa64_ccidx(const ARMISARegisters *id) return FIELD_EX64(id->id_aa64mmfr2, ID_AA64MMFR2, CCIDX) != 0; } +static inline bool isar_feature_aa64_tts2uxn(const ARMISARegisters *id) +{ + return FIELD_EX64(id->id_aa64mmfr1, ID_AA64MMFR1, XNX) != 0; +} + /* * Feature tests for "does this exist in either 32-bit or 64-bit?" */ @@ -XXX,XX +XXX,XX @@ static inline bool isar_feature_any_ccidx(const ARMISARegisters *id) return isar_feature_aa64_ccidx(id) || isar_feature_aa32_ccidx(id); } +static inline bool isar_feature_any_tts2uxn(const ARMISARegisters *id) +{ + return isar_feature_aa64_tts2uxn(id) || isar_feature_aa32_tts2uxn(id); +} + /* * Forward to the above feature tests given an ARMCPU pointer. */ 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_max_initfn(Object *obj) t = FIELD_DP32(t, ID_MMFR4, HPDS, 1); /* AA32HPD */ t = FIELD_DP32(t, ID_MMFR4, AC2, 1); /* ACTLR2, HACTLR2 */ t = FIELD_DP32(t, ID_MMFR4, CNP, 1); /* TTCNP */ + t = FIELD_DP32(t, ID_MMFR4, XNX, 1); /* TTS2UXN */ cpu->isar.id_mmfr4 = t; } #endif 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_AA64MMFR1, VH, 1); t = FIELD_DP64(t, ID_AA64MMFR1, PAN, 2); /* ATS1E1 */ t = FIELD_DP64(t, ID_AA64MMFR1, VMIDBITS, 2); /* VMID16 */ + t = FIELD_DP64(t, ID_AA64MMFR1, XNX, 1); /* TTS2UXN */ cpu->isar.id_aa64mmfr1 = t; t = cpu->isar.id_aa64mmfr2; @@ -XXX,XX +XXX,XX @@ static void aarch64_max_initfn(Object *obj) u = FIELD_DP32(u, ID_MMFR4, HPDS, 1); /* AA32HPD */ u = FIELD_DP32(u, ID_MMFR4, AC2, 1); /* ACTLR2, HACTLR2 */ u = FIELD_DP32(u, ID_MMFR4, CNP, 1); /* TTCNP */ + u = FIELD_DP32(u, ID_MMFR4, XNX, 1); /* TTS2UXN */ cpu->isar.id_mmfr4 = u; u = cpu->isar.id_aa64dfr0; 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 @@ simple_ap_to_rw_prot(CPUARMState *env, ARMMMUIdx mmu_idx, int ap) * * @env: CPUARMState * @s2ap: The 2-bit stage2 access permissions (S2AP) - * @xn: XN (execute-never) bit + * @xn: XN (execute-never) bits + * @s1_is_el0: true if this is S2 of an S1+2 walk for EL0 */ -static int get_S2prot(CPUARMState *env, int s2ap, int xn) +static int get_S2prot(CPUARMState *env, int s2ap, int xn, bool s1_is_el0) { int prot = 0; @@ -XXX,XX +XXX,XX @@ static int get_S2prot(CPUARMState *env, int s2ap, int xn) if (s2ap & 2) { prot |= PAGE_WRITE; } - if (!xn) { - if (arm_el_is_aa64(env, 2) || prot & PAGE_READ) { + + if (cpu_isar_feature(any_tts2uxn, env_archcpu(env))) { + switch (xn) { + case 0: prot |= PAGE_EXEC; + break; + case 1: + if (s1_is_el0) { + prot |= PAGE_EXEC; + } + break; + case 2: + break; + case 3: + if (!s1_is_el0) { + prot |= PAGE_EXEC; + } + break; + default: + g_assert_not_reached(); + } + } else { + if (!extract32(xn, 1, 1)) { + if (arm_el_is_aa64(env, 2) || prot & PAGE_READ) { + prot |= PAGE_EXEC; + } } } return prot; @@ -XXX,XX +XXX,XX @@ static bool get_phys_addr_lpae(CPUARMState *env, target_ulong address, } ap = extract32(attrs, 4, 2); - xn = extract32(attrs, 12, 1); if (mmu_idx == ARMMMUIdx_Stage2) { ns = true; - *prot = get_S2prot(env, ap, xn); + xn = extract32(attrs, 11, 2); + *prot = get_S2prot(env, ap, xn, s1_is_el0); } else { ns = extract32(attrs, 3, 1); + xn = extract32(attrs, 12, 1); pxn = extract32(attrs, 11, 1); *prot = get_S1prot(env, mmu_idx, aarch64, ap, ns, xn, pxn); } -- 2.20.1 In aarch64_max_initfn() we update both 32-bit and 64-bit ID registers. The intended pattern is that for 64-bit ID registers we use FIELD_DP64 and the uint64_t 't' register, while 32-bit ID registers use FIELD_DP32 and the uint32_t 'u' register. For ID_AA64DFR0 we accidentally used 'u', meaning that the top 32 bits of this 64-bit ID register would end up always zero. Luckily at the moment that's what they should be anyway, so this bug has no visible effects. Use the right-sized variable. Fixes: 3bec78447a958d481991 Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Laurent Desnogues <laurent.desnogues@gmail.com> Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Message-id: 20200423110915.10527-1-peter.maydell@linaro.org --- target/arm/cpu64.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) diff --git a/target/arm/cpu64.c b/target/arm/cpu64.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/cpu64.c +++ b/target/arm/cpu64.c @@ -XXX,XX +XXX,XX @@ static void aarch64_max_initfn(Object *obj) u = FIELD_DP32(u, ID_MMFR4, XNX, 1); /* TTS2UXN */ cpu->isar.id_mmfr4 = u; - u = cpu->isar.id_aa64dfr0; - u = FIELD_DP64(u, ID_AA64DFR0, PMUVER, 5); /* v8.4-PMU */ - cpu->isar.id_aa64dfr0 = u; + t = cpu->isar.id_aa64dfr0; + t = FIELD_DP64(t, ID_AA64DFR0, PMUVER, 5); /* v8.4-PMU */ + cpu->isar.id_aa64dfr0 = t; u = cpu->isar.id_dfr0; u = FIELD_DP32(u, ID_DFR0, PERFMON, 5); /* v8.4-PMU */ -- 2.20.1 From: Philippe Mathieu-Daudé <f4bug@amsat.org> MIDR_EL1 is a 64-bit system register with the top 32-bit being RES0. Represent it in QEMU's ARMCPU struct with a uint64_t, not a uint32_t. This fixes an error when compiling with -Werror=conversion because we were manipulating the register value using a local uint64_t variable: target/arm/cpu64.c: In function ‘aarch64_max_initfn’: target/arm/cpu64.c:628:21: error: conversion from ‘uint64_t’ {aka ‘long unsigned int’} to ‘uint32_t’ {aka ‘unsigned int’} may change value [-Werror=conversion] 628 | cpu->midr = t; | ^ and future-proofs us against a possible future architecture change using some of the top 32 bits. Suggested-by: Laurent Desnogues <laurent.desnogues@gmail.com> Suggested-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Reviewed-by: Laurent Desnogues <laurent.desnogues@gmail.com> Message-id: 20200428172634.29707-1-f4bug@amsat.org Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/cpu.h | 2 +- target/arm/cpu.c | 2 +- 2 files changed, 2 insertions(+), 2 deletions(-) diff --git a/target/arm/cpu.h b/target/arm/cpu.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/cpu.h +++ b/target/arm/cpu.h @@ -XXX,XX +XXX,XX @@ struct ARMCPU { uint64_t id_aa64dfr0; uint64_t id_aa64dfr1; } isar; - uint32_t midr; + uint64_t midr; uint32_t revidr; uint32_t reset_fpsid; uint32_t ctr; 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 const ARMCPUInfo arm_cpus[] = { static Property arm_cpu_properties[] = { DEFINE_PROP_BOOL("start-powered-off", ARMCPU, start_powered_off, false), DEFINE_PROP_UINT32("psci-conduit", ARMCPU, psci_conduit, 0), - DEFINE_PROP_UINT32("midr", ARMCPU, midr, 0), + DEFINE_PROP_UINT64("midr", ARMCPU, midr, 0), DEFINE_PROP_UINT64("mp-affinity", ARMCPU, mp_affinity, ARM64_AFFINITY_INVALID), DEFINE_PROP_INT32("node-id", ARMCPU, node_id, CPU_UNSET_NUMA_NODE_ID), -- 2.20.1 From: "Edgar E. Iglesias" <edgar.iglesias@xilinx.com> Remove inclusion of arm_gicv3_common.h, this already gets included via xlnx-versal.h. Signed-off-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com> Reviewed-by: Alistair Francis <alistair.francis@wdc.com> Reviewed-by: Luc Michel <luc.michel@greensocs.com> Message-id: 20200427181649.26851-2-edgar.iglesias@gmail.com Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- hw/arm/xlnx-versal.c | 1 - 1 file changed, 1 deletion(-) 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 @@ #include "hw/arm/boot.h" #include "kvm_arm.h" #include "hw/misc/unimp.h" -#include "hw/intc/arm_gicv3_common.h" #include "hw/arm/xlnx-versal.h" #include "hw/char/pl011.h" -- 2.20.1 From: "Edgar E. Iglesias" <edgar.iglesias@xilinx.com> Move misplaced comment. Signed-off-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com> Reviewed-by: Alistair Francis <alistair.francis@wdc.com> Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Reviewed-by: Luc Michel <luc.michel@greensocs.com> Message-id: 20200427181649.26851-3-edgar.iglesias@gmail.com Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- hw/arm/xlnx-versal.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) 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_apu_cpus(Versal *s) obj = object_new(XLNX_VERSAL_ACPU_TYPE); if (!obj) { - /* Secondary CPUs start in PSCI powered-down state */ error_report("Unable to create apu.cpu[%d] of type %s", i, XLNX_VERSAL_ACPU_TYPE); exit(EXIT_FAILURE); @@ -XXX,XX +XXX,XX @@ static void versal_create_apu_cpus(Versal *s) object_property_set_int(obj, s->cfg.psci_conduit, "psci-conduit", &error_abort); if (i) { + /* Secondary CPUs start in PSCI powered-down state */ object_property_set_bool(obj, true, "start-powered-off", &error_abort); } -- 2.20.1 From: "Edgar E. Iglesias" <edgar.iglesias@xilinx.com> Fix typo xlnx-ve -> xlnx-versal. Signed-off-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com> Reviewed-by: Alistair Francis <alistair.francis@wdc.com> Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Reviewed-by: Luc Michel <luc.michel@greensocs.com> Message-id: 20200427181649.26851-4-edgar.iglesias@gmail.com Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- hw/arm/xlnx-versal-virt.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/hw/arm/xlnx-versal-virt.c b/hw/arm/xlnx-versal-virt.c index XXXXXXX..XXXXXXX 100644 --- a/hw/arm/xlnx-versal-virt.c +++ b/hw/arm/xlnx-versal-virt.c @@ -XXX,XX +XXX,XX @@ static void versal_virt_init(MachineState *machine) psci_conduit = QEMU_PSCI_CONDUIT_SMC; } - sysbus_init_child_obj(OBJECT(machine), "xlnx-ve", &s->soc, + sysbus_init_child_obj(OBJECT(machine), "xlnx-versal", &s->soc, sizeof(s->soc), TYPE_XLNX_VERSAL); object_property_set_link(OBJECT(&s->soc), OBJECT(machine->ram), "ddr", &error_abort); -- 2.20.1 From: "Edgar E. Iglesias" <edgar.iglesias@xilinx.com> Embed the UARTs into the SoC type. Suggested-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com> Reviewed-by: Alistair Francis <alistair.francis@wdc.com> Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Reviewed-by: Luc Michel <luc.michel@greensocs.com> Message-id: 20200427181649.26851-5-edgar.iglesias@gmail.com Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- include/hw/arm/xlnx-versal.h | 3 ++- hw/arm/xlnx-versal.c | 12 ++++++------ 2 files changed, 8 insertions(+), 7 deletions(-) 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/sysbus.h" #include "hw/arm/boot.h" #include "hw/intc/arm_gicv3.h" +#include "hw/char/pl011.h" #define TYPE_XLNX_VERSAL "xlnx-versal" #define XLNX_VERSAL(obj) OBJECT_CHECK(Versal, (obj), TYPE_XLNX_VERSAL) @@ -XXX,XX +XXX,XX @@ typedef struct Versal { MemoryRegion mr_ocm; struct { - SysBusDevice *uart[XLNX_VERSAL_NR_UARTS]; + PL011State uart[XLNX_VERSAL_NR_UARTS]; SysBusDevice *gem[XLNX_VERSAL_NR_GEMS]; SysBusDevice *adma[XLNX_VERSAL_NR_ADMAS]; } iou; 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 @@ #include "kvm_arm.h" #include "hw/misc/unimp.h" #include "hw/arm/xlnx-versal.h" -#include "hw/char/pl011.h" #define XLNX_VERSAL_ACPU_TYPE ARM_CPU_TYPE_NAME("cortex-a72") #define GEM_REVISION 0x40070106 @@ -XXX,XX +XXX,XX @@ static void versal_create_uarts(Versal *s, qemu_irq *pic) DeviceState *dev; MemoryRegion *mr; - dev = qdev_create(NULL, TYPE_PL011); - s->lpd.iou.uart[i] = SYS_BUS_DEVICE(dev); + sysbus_init_child_obj(OBJECT(s), name, + &s->lpd.iou.uart[i], sizeof(s->lpd.iou.uart[i]), + TYPE_PL011); + dev = DEVICE(&s->lpd.iou.uart[i]); qdev_prop_set_chr(dev, "chardev", serial_hd(i)); - object_property_add_child(OBJECT(s), name, OBJECT(dev), &error_fatal); qdev_init_nofail(dev); - mr = sysbus_mmio_get_region(s->lpd.iou.uart[i], 0); + mr = sysbus_mmio_get_region(SYS_BUS_DEVICE(dev), 0); memory_region_add_subregion(&s->mr_ps, addrs[i], mr); - sysbus_connect_irq(s->lpd.iou.uart[i], 0, pic[irqs[i]]); + sysbus_connect_irq(SYS_BUS_DEVICE(dev), 0, pic[irqs[i]]); g_free(name); } } -- 2.20.1 From: "Edgar E. Iglesias" <edgar.iglesias@xilinx.com> Embed the GEMs into the SoC type. Suggested-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com> Reviewed-by: Alistair Francis <alistair.francis@wdc.com> Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Reviewed-by: Luc Michel <luc.michel@greensocs.com> Message-id: 20200427181649.26851-6-edgar.iglesias@gmail.com Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- include/hw/arm/xlnx-versal.h | 3 ++- hw/arm/xlnx-versal.c | 15 ++++++++------- 2 files changed, 10 insertions(+), 8 deletions(-) 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/arm/boot.h" #include "hw/intc/arm_gicv3.h" #include "hw/char/pl011.h" +#include "hw/net/cadence_gem.h" #define TYPE_XLNX_VERSAL "xlnx-versal" #define XLNX_VERSAL(obj) OBJECT_CHECK(Versal, (obj), TYPE_XLNX_VERSAL) @@ -XXX,XX +XXX,XX @@ typedef struct Versal { struct { PL011State uart[XLNX_VERSAL_NR_UARTS]; - SysBusDevice *gem[XLNX_VERSAL_NR_GEMS]; + CadenceGEMState gem[XLNX_VERSAL_NR_GEMS]; SysBusDevice *adma[XLNX_VERSAL_NR_ADMAS]; } iou; } lpd; 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_gems(Versal *s, qemu_irq *pic) DeviceState *dev; MemoryRegion *mr; - dev = qdev_create(NULL, "cadence_gem"); - s->lpd.iou.gem[i] = SYS_BUS_DEVICE(dev); - object_property_add_child(OBJECT(s), name, OBJECT(dev), &error_fatal); + sysbus_init_child_obj(OBJECT(s), name, + &s->lpd.iou.gem[i], sizeof(s->lpd.iou.gem[i]), + TYPE_CADENCE_GEM); + dev = DEVICE(&s->lpd.iou.gem[i]); if (nd->used) { qemu_check_nic_model(nd, "cadence_gem"); qdev_set_nic_properties(dev, nd); } - object_property_set_int(OBJECT(s->lpd.iou.gem[i]), + object_property_set_int(OBJECT(dev), 2, "num-priority-queues", &error_abort); - object_property_set_link(OBJECT(s->lpd.iou.gem[i]), + object_property_set_link(OBJECT(dev), OBJECT(&s->mr_ps), "dma", &error_abort); qdev_init_nofail(dev); - mr = sysbus_mmio_get_region(s->lpd.iou.gem[i], 0); + mr = sysbus_mmio_get_region(SYS_BUS_DEVICE(dev), 0); memory_region_add_subregion(&s->mr_ps, addrs[i], mr); - sysbus_connect_irq(s->lpd.iou.gem[i], 0, pic[irqs[i]]); + sysbus_connect_irq(SYS_BUS_DEVICE(dev), 0, pic[irqs[i]]); g_free(name); } } -- 2.20.1 From: "Edgar E. Iglesias" <edgar.iglesias@xilinx.com> Embed the ADMAs into the SoC type. Suggested-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com> Reviewed-by: Alistair Francis <alistair.francis@wdc.com> Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Reviewed-by: Luc Michel <luc.michel@greensocs.com> Message-id: 20200427181649.26851-7-edgar.iglesias@gmail.com Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- include/hw/arm/xlnx-versal.h | 3 ++- hw/arm/xlnx-versal.c | 14 +++++++------- 2 files changed, 9 insertions(+), 8 deletions(-) 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/arm/boot.h" #include "hw/intc/arm_gicv3.h" #include "hw/char/pl011.h" +#include "hw/dma/xlnx-zdma.h" #include "hw/net/cadence_gem.h" #define TYPE_XLNX_VERSAL "xlnx-versal" @@ -XXX,XX +XXX,XX @@ typedef struct Versal { struct { PL011State uart[XLNX_VERSAL_NR_UARTS]; CadenceGEMState gem[XLNX_VERSAL_NR_GEMS]; - SysBusDevice *adma[XLNX_VERSAL_NR_ADMAS]; + XlnxZDMA adma[XLNX_VERSAL_NR_ADMAS]; } iou; } lpd; 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_admas(Versal *s, qemu_irq *pic) DeviceState *dev; MemoryRegion *mr; - dev = qdev_create(NULL, "xlnx.zdma"); - s->lpd.iou.adma[i] = SYS_BUS_DEVICE(dev); - object_property_set_int(OBJECT(s->lpd.iou.adma[i]), 128, "bus-width", - &error_abort); - object_property_add_child(OBJECT(s), name, OBJECT(dev), &error_fatal); + sysbus_init_child_obj(OBJECT(s), name, + &s->lpd.iou.adma[i], sizeof(s->lpd.iou.adma[i]), + TYPE_XLNX_ZDMA); + dev = DEVICE(&s->lpd.iou.adma[i]); + object_property_set_int(OBJECT(dev), 128, "bus-width", &error_abort); qdev_init_nofail(dev); - mr = sysbus_mmio_get_region(s->lpd.iou.adma[i], 0); + mr = sysbus_mmio_get_region(SYS_BUS_DEVICE(dev), 0); memory_region_add_subregion(&s->mr_ps, MM_ADMA_CH0 + i * MM_ADMA_CH0_SIZE, mr); - sysbus_connect_irq(s->lpd.iou.adma[i], 0, pic[VERSAL_ADMA_IRQ_0 + i]); + sysbus_connect_irq(SYS_BUS_DEVICE(dev), 0, pic[VERSAL_ADMA_IRQ_0 + i]); g_free(name); } } -- 2.20.1 From: "Edgar E. Iglesias" <edgar.iglesias@xilinx.com> Embed the APUs into the SoC type. Suggested-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com> Reviewed-by: Alistair Francis <alistair.francis@wdc.com> Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Reviewed-by: Luc Michel <luc.michel@greensocs.com> Message-id: 20200427181649.26851-8-edgar.iglesias@gmail.com Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- include/hw/arm/xlnx-versal.h | 2 +- hw/arm/xlnx-versal-virt.c | 4 ++-- hw/arm/xlnx-versal.c | 19 +++++-------------- 3 files changed, 8 insertions(+), 17 deletions(-) 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 @@ typedef struct Versal { struct { struct { MemoryRegion mr; - ARMCPU *cpu[XLNX_VERSAL_NR_ACPUS]; + ARMCPU cpu[XLNX_VERSAL_NR_ACPUS]; GICv3State gic; } apu; } fpd; diff --git a/hw/arm/xlnx-versal-virt.c b/hw/arm/xlnx-versal-virt.c index XXXXXXX..XXXXXXX 100644 --- a/hw/arm/xlnx-versal-virt.c +++ b/hw/arm/xlnx-versal-virt.c @@ -XXX,XX +XXX,XX @@ static void versal_virt_init(MachineState *machine) s->binfo.get_dtb = versal_virt_get_dtb; s->binfo.modify_dtb = versal_virt_modify_dtb; if (machine->kernel_filename) { - arm_load_kernel(s->soc.fpd.apu.cpu[0], machine, &s->binfo); + arm_load_kernel(&s->soc.fpd.apu.cpu[0], machine, &s->binfo); } else { - AddressSpace *as = arm_boot_address_space(s->soc.fpd.apu.cpu[0], + AddressSpace *as = arm_boot_address_space(&s->soc.fpd.apu.cpu[0], &s->binfo); /* Some boot-loaders (e.g u-boot) don't like blobs at address 0 (NULL). * Offset things by 4K. */ 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_apu_cpus(Versal *s) for (i = 0; i < ARRAY_SIZE(s->fpd.apu.cpu); i++) { Object *obj; - char *name; - - obj = object_new(XLNX_VERSAL_ACPU_TYPE); - if (!obj) { - error_report("Unable to create apu.cpu[%d] of type %s", - i, XLNX_VERSAL_ACPU_TYPE); - exit(EXIT_FAILURE); - } - - name = g_strdup_printf("apu-cpu[%d]", i); - object_property_add_child(OBJECT(s), name, obj, &error_fatal); - g_free(name); + object_initialize_child(OBJECT(s), "apu-cpu[*]", + &s->fpd.apu.cpu[i], sizeof(s->fpd.apu.cpu[i]), + XLNX_VERSAL_ACPU_TYPE, &error_abort, NULL); + obj = OBJECT(&s->fpd.apu.cpu[i]); object_property_set_int(obj, s->cfg.psci_conduit, "psci-conduit", &error_abort); if (i) { @@ -XXX,XX +XXX,XX @@ static void versal_create_apu_cpus(Versal *s) object_property_set_link(obj, OBJECT(&s->fpd.apu.mr), "memory", &error_abort); object_property_set_bool(obj, true, "realized", &error_fatal); - s->fpd.apu.cpu[i] = ARM_CPU(obj); } } @@ -XXX,XX +XXX,XX @@ static void versal_create_apu_gic(Versal *s, qemu_irq *pic) } for (i = 0; i < nr_apu_cpus; i++) { - DeviceState *cpudev = DEVICE(s->fpd.apu.cpu[i]); + DeviceState *cpudev = DEVICE(&s->fpd.apu.cpu[i]); int ppibase = XLNX_VERSAL_NR_IRQS + i * GIC_INTERNAL + GIC_NR_SGIS; qemu_irq maint_irq; int ti; -- 2.20.1 From: "Edgar E. Iglesias" <edgar.iglesias@xilinx.com> Add support for SD. Signed-off-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com> Reviewed-by: Alistair Francis <alistair.francis@wdc.com> Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Reviewed-by: Luc Michel <luc.michel@greensocs.com> Message-id: 20200427181649.26851-9-edgar.iglesias@gmail.com Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- include/hw/arm/xlnx-versal.h | 12 ++++++++++++ hw/arm/xlnx-versal.c | 31 +++++++++++++++++++++++++++++++ 2 files changed, 43 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/sysbus.h" #include "hw/arm/boot.h" +#include "hw/sd/sdhci.h" #include "hw/intc/arm_gicv3.h" #include "hw/char/pl011.h" #include "hw/dma/xlnx-zdma.h" @@ -XXX,XX +XXX,XX @@ #define XLNX_VERSAL_NR_UARTS 2 #define XLNX_VERSAL_NR_GEMS 2 #define XLNX_VERSAL_NR_ADMAS 8 +#define XLNX_VERSAL_NR_SDS 2 #define XLNX_VERSAL_NR_IRQS 192 typedef struct Versal { @@ -XXX,XX +XXX,XX @@ typedef struct Versal { } iou; } lpd; + /* The Platform Management Controller subsystem. */ + struct { + struct { + SDHCIState sd[XLNX_VERSAL_NR_SDS]; + } iou; + } pmc; + struct { MemoryRegion *mr_ddr; uint32_t psci_conduit; @@ -XXX,XX +XXX,XX @@ typedef struct Versal { #define VERSAL_GEM1_IRQ_0 58 #define VERSAL_GEM1_WAKE_IRQ_0 59 #define VERSAL_ADMA_IRQ_0 60 +#define VERSAL_SD0_IRQ_0 126 /* Architecturally reserved IRQs suitable for virtualization. */ #define VERSAL_RSVD_IRQ_FIRST 111 @@ -XXX,XX +XXX,XX @@ typedef struct Versal { #define MM_FPD_CRF 0xfd1a0000U #define MM_FPD_CRF_SIZE 0x140000 +#define MM_PMC_SD0 0xf1040000U +#define MM_PMC_SD0_SIZE 0x10000 #define MM_PMC_CRP 0xf1260000U #define MM_PMC_CRP_SIZE 0x10000 #endif 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_admas(Versal *s, qemu_irq *pic) } } +#define SDHCI_CAPABILITIES 0x280737ec6481 /* Same as on ZynqMP. */ +static void versal_create_sds(Versal *s, qemu_irq *pic) +{ + int i; + + for (i = 0; i < ARRAY_SIZE(s->pmc.iou.sd); i++) { + DeviceState *dev; + MemoryRegion *mr; + + sysbus_init_child_obj(OBJECT(s), "sd[*]", + &s->pmc.iou.sd[i], sizeof(s->pmc.iou.sd[i]), + TYPE_SYSBUS_SDHCI); + dev = DEVICE(&s->pmc.iou.sd[i]); + + object_property_set_uint(OBJECT(dev), + 3, "sd-spec-version", &error_fatal); + object_property_set_uint(OBJECT(dev), SDHCI_CAPABILITIES, "capareg", + &error_fatal); + object_property_set_uint(OBJECT(dev), UHS_I, "uhs", &error_fatal); + qdev_init_nofail(dev); + + mr = sysbus_mmio_get_region(SYS_BUS_DEVICE(dev), 0); + memory_region_add_subregion(&s->mr_ps, + MM_PMC_SD0 + i * MM_PMC_SD0_SIZE, mr); + + sysbus_connect_irq(SYS_BUS_DEVICE(dev), 0, + pic[VERSAL_SD0_IRQ_0 + i * 2]); + } +} + /* This takes the board allocated linear DDR memory and creates aliases * for each split DDR range/aperture on the Versal address map. */ @@ -XXX,XX +XXX,XX @@ static void versal_realize(DeviceState *dev, Error **errp) versal_create_uarts(s, pic); versal_create_gems(s, pic); versal_create_admas(s, pic); + versal_create_sds(s, pic); versal_map_ddr(s); versal_unimp(s); -- 2.20.1 From: "Edgar E. Iglesias" <edgar.iglesias@xilinx.com> hw/arm: versal: Add support for the RTC. Signed-off-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com> Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Reviewed-by: Alistair Francis <alistair.francis@wdc.com> Reviewed-by: Luc Michel <luc.michel@greensocs.com> Message-id: 20200427181649.26851-10-edgar.iglesias@gmail.com Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- include/hw/arm/xlnx-versal.h | 8 ++++++++ hw/arm/xlnx-versal.c | 21 +++++++++++++++++++++ 2 files changed, 29 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/char/pl011.h" #include "hw/dma/xlnx-zdma.h" #include "hw/net/cadence_gem.h" +#include "hw/rtc/xlnx-zynqmp-rtc.h" #define TYPE_XLNX_VERSAL "xlnx-versal" #define XLNX_VERSAL(obj) OBJECT_CHECK(Versal, (obj), TYPE_XLNX_VERSAL) @@ -XXX,XX +XXX,XX @@ typedef struct Versal { struct { SDHCIState sd[XLNX_VERSAL_NR_SDS]; } iou; + + XlnxZynqMPRTC rtc; } pmc; struct { @@ -XXX,XX +XXX,XX @@ typedef struct Versal { #define VERSAL_GEM1_IRQ_0 58 #define VERSAL_GEM1_WAKE_IRQ_0 59 #define VERSAL_ADMA_IRQ_0 60 +#define VERSAL_RTC_APB_ERR_IRQ 121 #define VERSAL_SD0_IRQ_0 126 +#define VERSAL_RTC_ALARM_IRQ 142 +#define VERSAL_RTC_SECONDS_IRQ 143 /* Architecturally reserved IRQs suitable for virtualization. */ #define VERSAL_RSVD_IRQ_FIRST 111 @@ -XXX,XX +XXX,XX @@ typedef struct Versal { #define MM_PMC_SD0_SIZE 0x10000 #define MM_PMC_CRP 0xf1260000U #define MM_PMC_CRP_SIZE 0x10000 +#define MM_PMC_RTC 0xf12a0000 +#define MM_PMC_RTC_SIZE 0x10000 #endif 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_sds(Versal *s, qemu_irq *pic) } } +static void versal_create_rtc(Versal *s, qemu_irq *pic) +{ + SysBusDevice *sbd; + MemoryRegion *mr; + + sysbus_init_child_obj(OBJECT(s), "rtc", &s->pmc.rtc, sizeof(s->pmc.rtc), + TYPE_XLNX_ZYNQMP_RTC); + sbd = SYS_BUS_DEVICE(&s->pmc.rtc); + qdev_init_nofail(DEVICE(sbd)); + + mr = sysbus_mmio_get_region(sbd, 0); + memory_region_add_subregion(&s->mr_ps, MM_PMC_RTC, mr); + + /* + * TODO: Connect the ALARM and SECONDS interrupts once our RTC model + * supports them. + */ + sysbus_connect_irq(sbd, 1, pic[VERSAL_RTC_APB_ERR_IRQ]); +} + /* This takes the board allocated linear DDR memory and creates aliases * for each split DDR range/aperture on the Versal address map. */ @@ -XXX,XX +XXX,XX @@ static void versal_realize(DeviceState *dev, Error **errp) versal_create_gems(s, pic); versal_create_admas(s, pic); versal_create_sds(s, pic); + versal_create_rtc(s, pic); versal_map_ddr(s); versal_unimp(s); -- 2.20.1 From: "Edgar E. Iglesias" <edgar.iglesias@xilinx.com> Add support for SD. Signed-off-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com> Reviewed-by: Alistair Francis <alistair.francis@wdc.com> Reviewed-by: Luc Michel <luc.michel@greensocs.com> Message-id: 20200427181649.26851-11-edgar.iglesias@gmail.com Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- hw/arm/xlnx-versal-virt.c | 46 +++++++++++++++++++++++++++++++++++++++ 1 file changed, 46 insertions(+) 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 @@ #include "hw/arm/sysbus-fdt.h" #include "hw/arm/fdt.h" #include "cpu.h" +#include "hw/qdev-properties.h" #include "hw/arm/xlnx-versal.h" #define TYPE_XLNX_VERSAL_VIRT_MACHINE MACHINE_TYPE_NAME("xlnx-versal-virt") @@ -XXX,XX +XXX,XX @@ static void fdt_add_zdma_nodes(VersalVirt *s) } } +static void fdt_add_sd_nodes(VersalVirt *s) +{ + const char clocknames[] = "clk_xin\0clk_ahb"; + const char compat[] = "arasan,sdhci-8.9a"; + int i; + + for (i = ARRAY_SIZE(s->soc.pmc.iou.sd) - 1; i >= 0; i--) { + uint64_t addr = MM_PMC_SD0 + MM_PMC_SD0_SIZE * i; + char *name = g_strdup_printf("/sdhci@%" PRIx64, addr); + + qemu_fdt_add_subnode(s->fdt, name); + + qemu_fdt_setprop_cells(s->fdt, name, "clocks", + s->phandle.clk_25Mhz, s->phandle.clk_25Mhz); + qemu_fdt_setprop(s->fdt, name, "clock-names", + clocknames, sizeof(clocknames)); + qemu_fdt_setprop_cells(s->fdt, name, "interrupts", + GIC_FDT_IRQ_TYPE_SPI, VERSAL_SD0_IRQ_0 + i * 2, + GIC_FDT_IRQ_FLAGS_LEVEL_HI); + qemu_fdt_setprop_sized_cells(s->fdt, name, "reg", + 2, addr, 2, MM_PMC_SD0_SIZE); + qemu_fdt_setprop(s->fdt, name, "compatible", compat, sizeof(compat)); + g_free(name); + } +} + static void fdt_nop_memory_nodes(void *fdt, Error **errp) { Error *err = NULL; @@ -XXX,XX +XXX,XX @@ static void create_virtio_regions(VersalVirt *s) } } +static void sd_plugin_card(SDHCIState *sd, DriveInfo *di) +{ + BlockBackend *blk = di ? blk_by_legacy_dinfo(di) : NULL; + DeviceState *card; + + card = qdev_create(qdev_get_child_bus(DEVICE(sd), "sd-bus"), TYPE_SD_CARD); + object_property_add_child(OBJECT(sd), "card[*]", OBJECT(card), + &error_fatal); + qdev_prop_set_drive(card, "drive", blk, &error_fatal); + object_property_set_bool(OBJECT(card), true, "realized", &error_fatal); +} + static void versal_virt_init(MachineState *machine) { VersalVirt *s = XLNX_VERSAL_VIRT_MACHINE(machine); int psci_conduit = QEMU_PSCI_CONDUIT_DISABLED; + int i; /* * If the user provides an Operating System to be loaded, we expect them @@ -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_sd_nodes(s); fdt_add_cpu_nodes(s, psci_conduit); fdt_add_clk_node(s, "/clk125", 125000000, s->phandle.clk_125Mhz); fdt_add_clk_node(s, "/clk25", 25000000, s->phandle.clk_25Mhz); @@ -XXX,XX +XXX,XX @@ static void versal_virt_init(MachineState *machine) memory_region_add_subregion_overlap(get_system_memory(), 0, &s->soc.fpd.apu.mr, 0); + /* Plugin SD cards. */ + for (i = 0; i < ARRAY_SIZE(s->soc.pmc.iou.sd); i++) { + sd_plugin_card(&s->soc.pmc.iou.sd[i], drive_get_next(IF_SD)); + } + s->binfo.ram_size = machine->ram_size; s->binfo.loader_start = 0x0; s->binfo.get_dtb = versal_virt_get_dtb; -- 2.20.1 From: "Edgar E. Iglesias" <edgar.iglesias@xilinx.com> Add support for the RTC. Signed-off-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com> Reviewed-by: Alistair Francis <alistair.francis@wdc.com> Reviewed-by: Luc Michel <luc.michel@greensocs.com> Message-id: 20200427181649.26851-12-edgar.iglesias@gmail.com Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- hw/arm/xlnx-versal-virt.c | 22 ++++++++++++++++++++++ 1 file changed, 22 insertions(+) diff --git a/hw/arm/xlnx-versal-virt.c b/hw/arm/xlnx-versal-virt.c index XXXXXXX..XXXXXXX 100644 --- a/hw/arm/xlnx-versal-virt.c +++ b/hw/arm/xlnx-versal-virt.c @@ -XXX,XX +XXX,XX @@ static void fdt_add_sd_nodes(VersalVirt *s) } } +static void fdt_add_rtc_node(VersalVirt *s) +{ + const char compat[] = "xlnx,zynqmp-rtc"; + const char interrupt_names[] = "alarm\0sec"; + char *name = g_strdup_printf("/rtc@%x", MM_PMC_RTC); + + qemu_fdt_add_subnode(s->fdt, name); + + qemu_fdt_setprop_cells(s->fdt, name, "interrupts", + GIC_FDT_IRQ_TYPE_SPI, VERSAL_RTC_ALARM_IRQ, + GIC_FDT_IRQ_FLAGS_LEVEL_HI, + GIC_FDT_IRQ_TYPE_SPI, VERSAL_RTC_SECONDS_IRQ, + GIC_FDT_IRQ_FLAGS_LEVEL_HI); + qemu_fdt_setprop(s->fdt, name, "interrupt-names", + interrupt_names, sizeof(interrupt_names)); + qemu_fdt_setprop_sized_cells(s->fdt, name, "reg", + 2, MM_PMC_RTC, 2, MM_PMC_RTC_SIZE); + qemu_fdt_setprop(s->fdt, name, "compatible", compat, sizeof(compat)); + g_free(name); +} + static void fdt_nop_memory_nodes(void *fdt, Error **errp) { Error *err = NULL; @@ -XXX,XX +XXX,XX @@ static void versal_virt_init(MachineState *machine) fdt_add_timer_nodes(s); fdt_add_zdma_nodes(s); fdt_add_sd_nodes(s); + fdt_add_rtc_node(s); fdt_add_cpu_nodes(s, psci_conduit); fdt_add_clk_node(s, "/clk125", 125000000, s->phandle.clk_125Mhz); fdt_add_clk_node(s, "/clk25", 25000000, s->phandle.clk_25Mhz); -- 2.20.1 Somewhere along theline we accidentally added a duplicate "using D16-D31 when they don't exist" check to do_vfm_dp() (probably an artifact of a patchseries rebase). Remove it. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Message-id: 20200430181003.21682-2-peter.maydell@linaro.org --- target/arm/translate-vfp.inc.c | 6 ------ 1 file changed, 6 deletions(-) diff --git a/target/arm/translate-vfp.inc.c b/target/arm/translate-vfp.inc.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-vfp.inc.c +++ b/target/arm/translate-vfp.inc.c @@ -XXX,XX +XXX,XX @@ static bool do_vfm_dp(DisasContext *s, arg_VFMA_dp *a, bool neg_n, bool neg_d) return false; } - /* UNDEF accesses to D16-D31 if they don't exist. */ - if (!dc_isar_feature(aa32_simd_r32, s) && - ((a->vd | a->vn | a->vm) & 0x10)) { - return false; - } - if (!vfp_access_check(s)) { return true; } -- 2.20.1 We were accidentally permitting decode of Thumb Neon insns even if the CPU didn't have the FEATURE_NEON bit set, because the feature check was being done before the call to disas_neon_data_insn() and disas_neon_ls_insn() in the Arm decoder but was omitted from the Thumb decoder. Push the feature bit check down into the called functions so it is done for both Arm and Thumb encodings. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Message-id: 20200430181003.21682-3-peter.maydell@linaro.org --- target/arm/translate.c | 16 ++++++++-------- 1 file changed, 8 insertions(+), 8 deletions(-) diff --git a/target/arm/translate.c b/target/arm/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate.c +++ b/target/arm/translate.c @@ -XXX,XX +XXX,XX @@ static int disas_neon_ls_insn(DisasContext *s, uint32_t insn) TCGv_i32 tmp2; TCGv_i64 tmp64; + if (!arm_dc_feature(s, ARM_FEATURE_NEON)) { + return 1; + } + /* FIXME: this access check should not take precedence over UNDEF * for invalid encodings; we will generate incorrect syndrome information * for attempts to execute invalid vfp/neon encodings with FP disabled. @@ -XXX,XX +XXX,XX @@ static int disas_neon_data_insn(DisasContext *s, uint32_t insn) TCGv_ptr ptr1, ptr2, ptr3; TCGv_i64 tmp64; + if (!arm_dc_feature(s, ARM_FEATURE_NEON)) { + return 1; + } + /* FIXME: this access check should not take precedence over UNDEF * for invalid encodings; we will generate incorrect syndrome information * for attempts to execute invalid vfp/neon encodings with FP disabled. @@ -XXX,XX +XXX,XX @@ static void disas_arm_insn(DisasContext *s, unsigned int insn) if (((insn >> 25) & 7) == 1) { /* NEON Data processing. */ - if (!arm_dc_feature(s, ARM_FEATURE_NEON)) { - goto illegal_op; - } - if (disas_neon_data_insn(s, insn)) { goto illegal_op; } @@ -XXX,XX +XXX,XX @@ static void disas_arm_insn(DisasContext *s, unsigned int insn) } if ((insn & 0x0f100000) == 0x04000000) { /* NEON load/store. */ - if (!arm_dc_feature(s, ARM_FEATURE_NEON)) { - goto illegal_op; - } - if (disas_neon_ls_insn(s, insn)) { goto illegal_op; } -- 2.20.1 Add the infrastructure for building and invoking a decodetree decoder for the AArch32 Neon encodings. At the moment the new decoder covers nothing, so we always fall back to the existing hand-written decode. We follow the same pattern we did for the VFP decodetree conversion (commit 78e138bc1f672c145ef6ace74617d and following): code that deals with Neon will be moving gradually out to translate-neon.vfp.inc, which we #include into translate.c. In order to share the decode files between A32 and T32, we split Neon into 3 parts: * data-processing * load-store * 'shared' encodings The first two groups of instructions have similar but not identical A32 and T32 encodings, so we need to manually transform the T32 encoding into the A32 one before calling the decoder; the third group covers the Neon instructions which are identical in A32 and T32. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200430181003.21682-4-peter.maydell@linaro.org --- target/arm/neon-dp.decode | 29 ++++++++++++++++++++++++++ target/arm/neon-ls.decode | 29 ++++++++++++++++++++++++++ target/arm/neon-shared.decode | 27 +++++++++++++++++++++++++ target/arm/translate-neon.inc.c | 32 +++++++++++++++++++++++++++++ target/arm/translate.c | 36 +++++++++++++++++++++++++++++++-- target/arm/Makefile.objs | 18 +++++++++++++++++ 6 files changed, 169 insertions(+), 2 deletions(-) create mode 100644 target/arm/neon-dp.decode create mode 100644 target/arm/neon-ls.decode create mode 100644 target/arm/neon-shared.decode create mode 100644 target/arm/translate-neon.inc.c diff --git a/target/arm/neon-dp.decode b/target/arm/neon-dp.decode new file mode 100644 index XXXXXXX..XXXXXXX --- /dev/null +++ b/target/arm/neon-dp.decode @@ -XXX,XX +XXX,XX @@ +# AArch32 Neon data-processing instruction descriptions +# +# Copyright (c) 2020 Linaro, Ltd +# +# This library is free software; you can redistribute it and/or +# modify it under the terms of the GNU Lesser General Public +# License as published by the Free Software Foundation; either +# version 2 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 +# + +# Encodings for Neon data processing instructions where the T32 encoding +# is a simple transformation of the A32 encoding. +# More specifically, this file covers instructions where the A32 encoding is +# 0b1111_001p_qqqq_qqqq_qqqq_qqqq_qqqq_qqqq +# and the T32 encoding is +# 0b111p_1111_qqqq_qqqq_qqqq_qqqq_qqqq_qqqq +# This file works on the A32 encoding only; calling code for T32 has to +# transform the insn into the A32 version first. diff --git a/target/arm/neon-ls.decode b/target/arm/neon-ls.decode new file mode 100644 index XXXXXXX..XXXXXXX --- /dev/null +++ b/target/arm/neon-ls.decode @@ -XXX,XX +XXX,XX @@ +# AArch32 Neon load/store instruction descriptions +# +# Copyright (c) 2020 Linaro, Ltd +# +# This library is free software; you can redistribute it and/or +# modify it under the terms of the GNU Lesser General Public +# License as published by the Free Software Foundation; either +# version 2 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 +# + +# Encodings for Neon load/store instructions where the T32 encoding +# is a simple transformation of the A32 encoding. +# More specifically, this file covers instructions where the A32 encoding is +# 0b1111_0100_xxx0_xxxx_xxxx_xxxx_xxxx_xxxx +# and the T32 encoding is +# 0b1111_1001_xxx0_xxxx_xxxx_xxxx_xxxx_xxxx +# This file works on the A32 encoding only; calling code for T32 has to +# transform the insn into the A32 version first. diff --git a/target/arm/neon-shared.decode b/target/arm/neon-shared.decode new file mode 100644 index XXXXXXX..XXXXXXX --- /dev/null +++ b/target/arm/neon-shared.decode @@ -XXX,XX +XXX,XX @@ +# AArch32 Neon instruction descriptions +# +# Copyright (c) 2020 Linaro, Ltd +# +# This library is free software; you can redistribute it and/or +# modify it under the terms of the GNU Lesser General Public +# License as published by the Free Software Foundation; either +# version 2 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 +# + +# Encodings for Neon instructions whose encoding is the same for +# both A32 and T32. + +# More specifically, this covers: +# 2reg scalar ext: 0b1111_1110_xxxx_xxxx_xxxx_1x0x_xxxx_xxxx +# 3same ext: 0b1111_110x_xxxx_xxxx_xxxx_1x0x_xxxx_xxxx diff --git a/target/arm/translate-neon.inc.c b/target/arm/translate-neon.inc.c new file mode 100644 index XXXXXXX..XXXXXXX --- /dev/null +++ b/target/arm/translate-neon.inc.c @@ -XXX,XX +XXX,XX @@ +/* + * ARM translation: AArch32 Neon instructions + * + * Copyright (c) 2003 Fabrice Bellard + * Copyright (c) 2005-2007 CodeSourcery + * Copyright (c) 2007 OpenedHand, Ltd. + * Copyright (c) 2020 Linaro, Ltd. + * + * This library is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2 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 intended to be included from translate.c; it uses + * some macros and definitions provided by that file. + * It might be possible to convert it to a standalone .c file eventually. + */ + +/* Include the generated Neon decoder */ +#include "decode-neon-dp.inc.c" +#include "decode-neon-ls.inc.c" +#include "decode-neon-shared.inc.c" 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 TCGv_ptr vfp_reg_ptr(bool dp, int reg) #define ARM_CP_RW_BIT (1 << 20) -/* Include the VFP decoder */ +/* Include the VFP and Neon decoders */ #include "translate-vfp.inc.c" +#include "translate-neon.inc.c" static inline void iwmmxt_load_reg(TCGv_i64 var, int reg) { @@ -XXX,XX +XXX,XX @@ static void disas_arm_insn(DisasContext *s, unsigned int insn) /* Unconditional instructions. */ /* TODO: Perhaps merge these into one decodetree output file. */ if (disas_a32_uncond(s, insn) || - disas_vfp_uncond(s, insn)) { + disas_vfp_uncond(s, insn) || + disas_neon_dp(s, insn) || + disas_neon_ls(s, insn) || + disas_neon_shared(s, insn)) { return; } /* fall back to legacy decoder */ @@ -XXX,XX +XXX,XX @@ static void disas_thumb2_insn(DisasContext *s, uint32_t insn) ARCH(6T2); } + if ((insn & 0xef000000) == 0xef000000) { + /* + * T32 encodings 0b111p_1111_qqqq_qqqq_qqqq_qqqq_qqqq_qqqq + * transform into + * A32 encodings 0b1111_001p_qqqq_qqqq_qqqq_qqqq_qqqq_qqqq + */ + uint32_t a32_insn = (insn & 0xe2ffffff) | + ((insn & (1 << 28)) >> 4) | (1 << 28); + + if (disas_neon_dp(s, a32_insn)) { + return; + } + } + + if ((insn & 0xff100000) == 0xf9000000) { + /* + * T32 encodings 0b1111_1001_ppp0_qqqq_qqqq_qqqq_qqqq_qqqq + * transform into + * A32 encodings 0b1111_0100_ppp0_qqqq_qqqq_qqqq_qqqq_qqqq + */ + uint32_t a32_insn = (insn & 0x00ffffff) | 0xf4000000; + + if (disas_neon_ls(s, a32_insn)) { + return; + } + } + /* * TODO: Perhaps merge these into one decodetree output file. * Note disas_vfp is written for a32 with cond field in the @@ -XXX,XX +XXX,XX @@ static void disas_thumb2_insn(DisasContext *s, uint32_t insn) */ if (disas_t32(s, insn) || disas_vfp_uncond(s, insn) || + disas_neon_shared(s, insn) || ((insn >> 28) == 0xe && disas_vfp(s, insn))) { return; } diff --git a/target/arm/Makefile.objs b/target/arm/Makefile.objs index XXXXXXX..XXXXXXX 100644 --- a/target/arm/Makefile.objs +++ b/target/arm/Makefile.objs @@ -XXX,XX +XXX,XX @@ target/arm/decode-sve.inc.c: $(SRC_PATH)/target/arm/sve.decode $(DECODETREE) $(PYTHON) $(DECODETREE) --decode disas_sve -o $@ $<,\ "GEN", $(TARGET_DIR)$@) +target/arm/decode-neon-shared.inc.c: $(SRC_PATH)/target/arm/neon-shared.decode $(DECODETREE) + $(call quiet-command,\ + $(PYTHON) $(DECODETREE) --static-decode disas_neon_shared -o $@ $<,\ + "GEN", $(TARGET_DIR)$@) + +target/arm/decode-neon-dp.inc.c: $(SRC_PATH)/target/arm/neon-dp.decode $(DECODETREE) + $(call quiet-command,\ + $(PYTHON) $(DECODETREE) --static-decode disas_neon_dp -o $@ $<,\ + "GEN", $(TARGET_DIR)$@) + +target/arm/decode-neon-ls.inc.c: $(SRC_PATH)/target/arm/neon-ls.decode $(DECODETREE) + $(call quiet-command,\ + $(PYTHON) $(DECODETREE) --static-decode disas_neon_ls -o $@ $<,\ + "GEN", $(TARGET_DIR)$@) + target/arm/decode-vfp.inc.c: $(SRC_PATH)/target/arm/vfp.decode $(DECODETREE) $(call quiet-command,\ $(PYTHON) $(DECODETREE) --static-decode disas_vfp -o $@ $<,\ @@ -XXX,XX +XXX,XX @@ target/arm/decode-t16.inc.c: $(SRC_PATH)/target/arm/t16.decode $(DECODETREE) "GEN", $(TARGET_DIR)$@) target/arm/translate-sve.o: target/arm/decode-sve.inc.c +target/arm/translate.o: target/arm/decode-neon-shared.inc.c +target/arm/translate.o: target/arm/decode-neon-dp.inc.c +target/arm/translate.o: target/arm/decode-neon-ls.inc.c target/arm/translate.o: target/arm/decode-vfp.inc.c target/arm/translate.o: target/arm/decode-vfp-uncond.inc.c target/arm/translate.o: target/arm/decode-a32.inc.c -- 2.20.1 Convert the VCMLA (vector) insns in the 3same extension group to decodetree. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200430181003.21682-5-peter.maydell@linaro.org --- target/arm/neon-shared.decode | 11 ++++++++++ target/arm/translate-neon.inc.c | 37 +++++++++++++++++++++++++++++++++ target/arm/translate.c | 11 +--------- 3 files changed, 49 insertions(+), 10 deletions(-) diff --git a/target/arm/neon-shared.decode b/target/arm/neon-shared.decode index XXXXXXX..XXXXXXX 100644 --- a/target/arm/neon-shared.decode +++ b/target/arm/neon-shared.decode @@ -XXX,XX +XXX,XX @@ # More specifically, this covers: # 2reg scalar ext: 0b1111_1110_xxxx_xxxx_xxxx_1x0x_xxxx_xxxx # 3same ext: 0b1111_110x_xxxx_xxxx_xxxx_1x0x_xxxx_xxxx + +# VFP/Neon register fields; same as vfp.decode +%vm_dp 5:1 0:4 +%vm_sp 0:4 5:1 +%vn_dp 7:1 16:4 +%vn_sp 16:4 7:1 +%vd_dp 22:1 12:4 +%vd_sp 12:4 22:1 + +VCMLA 1111 110 rot:2 . 1 size:1 .... .... 1000 . q:1 . 0 .... \ + vm=%vm_dp vn=%vn_dp vd=%vd_dp diff --git a/target/arm/translate-neon.inc.c b/target/arm/translate-neon.inc.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-neon.inc.c +++ b/target/arm/translate-neon.inc.c @@ -XXX,XX +XXX,XX @@ #include "decode-neon-dp.inc.c" #include "decode-neon-ls.inc.c" #include "decode-neon-shared.inc.c" + +static bool trans_VCMLA(DisasContext *s, arg_VCMLA *a) +{ + int opr_sz; + TCGv_ptr fpst; + gen_helper_gvec_3_ptr *fn_gvec_ptr; + + if (!dc_isar_feature(aa32_vcma, s) + || (!a->size && !dc_isar_feature(aa32_fp16_arith, s))) { + return false; + } + + /* UNDEF accesses to D16-D31 if they don't exist. */ + if (!dc_isar_feature(aa32_simd_r32, s) && + ((a->vd | a->vn | a->vm) & 0x10)) { + return false; + } + + if ((a->vn | a->vm | a->vd) & a->q) { + return false; + } + + if (!vfp_access_check(s)) { + return true; + } + + opr_sz = (1 + a->q) * 8; + fpst = get_fpstatus_ptr(1); + fn_gvec_ptr = a->size ? gen_helper_gvec_fcmlas : gen_helper_gvec_fcmlah; + tcg_gen_gvec_3_ptr(vfp_reg_offset(1, a->vd), + vfp_reg_offset(1, a->vn), + vfp_reg_offset(1, a->vm), + fpst, opr_sz, opr_sz, a->rot, + fn_gvec_ptr); + tcg_temp_free_ptr(fpst); + return true; +} diff --git a/target/arm/translate.c b/target/arm/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate.c +++ b/target/arm/translate.c @@ -XXX,XX +XXX,XX @@ static int disas_neon_insn_3same_ext(DisasContext *s, uint32_t insn) bool is_long = false, q = extract32(insn, 6, 1); bool ptr_is_env = false; - if ((insn & 0xfe200f10) == 0xfc200800) { - /* VCMLA -- 1111 110R R.1S .... .... 1000 ...0 .... */ - int size = extract32(insn, 20, 1); - data = extract32(insn, 23, 2); /* rot */ - if (!dc_isar_feature(aa32_vcma, s) - || (!size && !dc_isar_feature(aa32_fp16_arith, s))) { - return 1; - } - fn_gvec_ptr = size ? gen_helper_gvec_fcmlas : gen_helper_gvec_fcmlah; - } else if ((insn & 0xfea00f10) == 0xfc800800) { + if ((insn & 0xfea00f10) == 0xfc800800) { /* VCADD -- 1111 110R 1.0S .... .... 1000 ...0 .... */ int size = extract32(insn, 20, 1); data = extract32(insn, 24, 1); /* rot */ -- 2.20.1 Convert the VCADD (vector) insns to decodetree. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200430181003.21682-6-peter.maydell@linaro.org --- target/arm/neon-shared.decode | 3 +++ target/arm/translate-neon.inc.c | 37 +++++++++++++++++++++++++++++++++ target/arm/translate.c | 11 +--------- 3 files changed, 41 insertions(+), 10 deletions(-) diff --git a/target/arm/neon-shared.decode b/target/arm/neon-shared.decode index XXXXXXX..XXXXXXX 100644 --- a/target/arm/neon-shared.decode +++ b/target/arm/neon-shared.decode @@ -XXX,XX +XXX,XX @@ VCMLA 1111 110 rot:2 . 1 size:1 .... .... 1000 . q:1 . 0 .... \ vm=%vm_dp vn=%vn_dp vd=%vd_dp + +VCADD 1111 110 rot:1 1 . 0 size:1 .... .... 1000 . q:1 . 0 .... \ + vm=%vm_dp vn=%vn_dp vd=%vd_dp diff --git a/target/arm/translate-neon.inc.c b/target/arm/translate-neon.inc.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-neon.inc.c +++ b/target/arm/translate-neon.inc.c @@ -XXX,XX +XXX,XX @@ static bool trans_VCMLA(DisasContext *s, arg_VCMLA *a) tcg_temp_free_ptr(fpst); return true; } + +static bool trans_VCADD(DisasContext *s, arg_VCADD *a) +{ + int opr_sz; + TCGv_ptr fpst; + gen_helper_gvec_3_ptr *fn_gvec_ptr; + + if (!dc_isar_feature(aa32_vcma, s) + || (!a->size && !dc_isar_feature(aa32_fp16_arith, s))) { + return false; + } + + /* UNDEF accesses to D16-D31 if they don't exist. */ + if (!dc_isar_feature(aa32_simd_r32, s) && + ((a->vd | a->vn | a->vm) & 0x10)) { + return false; + } + + if ((a->vn | a->vm | a->vd) & a->q) { + return false; + } + + if (!vfp_access_check(s)) { + return true; + } + + opr_sz = (1 + a->q) * 8; + fpst = get_fpstatus_ptr(1); + fn_gvec_ptr = a->size ? gen_helper_gvec_fcadds : gen_helper_gvec_fcaddh; + tcg_gen_gvec_3_ptr(vfp_reg_offset(1, a->vd), + vfp_reg_offset(1, a->vn), + vfp_reg_offset(1, a->vm), + fpst, opr_sz, opr_sz, a->rot, + fn_gvec_ptr); + tcg_temp_free_ptr(fpst); + return true; +} diff --git a/target/arm/translate.c b/target/arm/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate.c +++ b/target/arm/translate.c @@ -XXX,XX +XXX,XX @@ static int disas_neon_insn_3same_ext(DisasContext *s, uint32_t insn) bool is_long = false, q = extract32(insn, 6, 1); bool ptr_is_env = false; - if ((insn & 0xfea00f10) == 0xfc800800) { - /* VCADD -- 1111 110R 1.0S .... .... 1000 ...0 .... */ - int size = extract32(insn, 20, 1); - data = extract32(insn, 24, 1); /* rot */ - if (!dc_isar_feature(aa32_vcma, s) - || (!size && !dc_isar_feature(aa32_fp16_arith, s))) { - return 1; - } - fn_gvec_ptr = size ? gen_helper_gvec_fcadds : gen_helper_gvec_fcaddh; - } else if ((insn & 0xfeb00f00) == 0xfc200d00) { + if ((insn & 0xfeb00f00) == 0xfc200d00) { /* V[US]DOT -- 1111 1100 0.10 .... .... 1101 .Q.U .... */ bool u = extract32(insn, 4, 1); if (!dc_isar_feature(aa32_dp, s)) { -- 2.20.1 Convert the V[US]DOT (vector) insns to decodetree. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200430181003.21682-7-peter.maydell@linaro.org --- target/arm/neon-shared.decode | 4 ++++ target/arm/translate-neon.inc.c | 32 ++++++++++++++++++++++++++++++++ target/arm/translate.c | 9 +-------- 3 files changed, 37 insertions(+), 8 deletions(-) diff --git a/target/arm/neon-shared.decode b/target/arm/neon-shared.decode index XXXXXXX..XXXXXXX 100644 --- a/target/arm/neon-shared.decode +++ b/target/arm/neon-shared.decode @@ -XXX,XX +XXX,XX @@ VCMLA 1111 110 rot:2 . 1 size:1 .... .... 1000 . q:1 . 0 .... \ VCADD 1111 110 rot:1 1 . 0 size:1 .... .... 1000 . q:1 . 0 .... \ vm=%vm_dp vn=%vn_dp vd=%vd_dp + +# VUDOT and VSDOT +VDOT 1111 110 00 . 10 .... .... 1101 . q:1 . u:1 .... \ + vm=%vm_dp vn=%vn_dp vd=%vd_dp diff --git a/target/arm/translate-neon.inc.c b/target/arm/translate-neon.inc.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-neon.inc.c +++ b/target/arm/translate-neon.inc.c @@ -XXX,XX +XXX,XX @@ static bool trans_VCADD(DisasContext *s, arg_VCADD *a) tcg_temp_free_ptr(fpst); return true; } + +static bool trans_VDOT(DisasContext *s, arg_VDOT *a) +{ + int opr_sz; + gen_helper_gvec_3 *fn_gvec; + + if (!dc_isar_feature(aa32_dp, s)) { + return false; + } + + /* UNDEF accesses to D16-D31 if they don't exist. */ + if (!dc_isar_feature(aa32_simd_r32, s) && + ((a->vd | a->vn | a->vm) & 0x10)) { + return false; + } + + if ((a->vn | a->vm | a->vd) & a->q) { + return false; + } + + if (!vfp_access_check(s)) { + return true; + } + + opr_sz = (1 + a->q) * 8; + fn_gvec = a->u ? gen_helper_gvec_udot_b : gen_helper_gvec_sdot_b; + tcg_gen_gvec_3_ool(vfp_reg_offset(1, a->vd), + vfp_reg_offset(1, a->vn), + vfp_reg_offset(1, a->vm), + opr_sz, opr_sz, 0, fn_gvec); + return true; +} diff --git a/target/arm/translate.c b/target/arm/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate.c +++ b/target/arm/translate.c @@ -XXX,XX +XXX,XX @@ static int disas_neon_insn_3same_ext(DisasContext *s, uint32_t insn) bool is_long = false, q = extract32(insn, 6, 1); bool ptr_is_env = false; - if ((insn & 0xfeb00f00) == 0xfc200d00) { - /* V[US]DOT -- 1111 1100 0.10 .... .... 1101 .Q.U .... */ - bool u = extract32(insn, 4, 1); - if (!dc_isar_feature(aa32_dp, s)) { - return 1; - } - fn_gvec = u ? gen_helper_gvec_udot_b : gen_helper_gvec_sdot_b; - } else if ((insn & 0xff300f10) == 0xfc200810) { + if ((insn & 0xff300f10) == 0xfc200810) { /* VFM[AS]L -- 1111 1100 S.10 .... .... 1000 .Q.1 .... */ int is_s = extract32(insn, 23, 1); if (!dc_isar_feature(aa32_fhm, s)) { -- 2.20.1 Convert the VFM[AS]L (vector) insns to decodetree. This is the last insn in the legacy decoder for the 3same_ext group, so we can delete the legacy decoder function for the group entirely. Note that in disas_thumb2_insn() the parts of this encoding space where the decodetree decoder returns false will correctly be directed to illegal_op by the "(insn & (1 << 28))" check so they won't fall into disas_coproc_insn() by mistake. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200430181003.21682-8-peter.maydell@linaro.org --- target/arm/neon-shared.decode | 6 +++ target/arm/translate-neon.inc.c | 31 +++++++++++ target/arm/translate.c | 92 +-------------------------------- 3 files changed, 38 insertions(+), 91 deletions(-) diff --git a/target/arm/neon-shared.decode b/target/arm/neon-shared.decode index XXXXXXX..XXXXXXX 100644 --- a/target/arm/neon-shared.decode +++ b/target/arm/neon-shared.decode @@ -XXX,XX +XXX,XX @@ VCADD 1111 110 rot:1 1 . 0 size:1 .... .... 1000 . q:1 . 0 .... \ # VUDOT and VSDOT VDOT 1111 110 00 . 10 .... .... 1101 . q:1 . u:1 .... \ vm=%vm_dp vn=%vn_dp vd=%vd_dp + +# VFM[AS]L +VFML 1111 110 0 s:1 . 10 .... .... 1000 . 0 . 1 .... \ + vm=%vm_sp vn=%vn_sp vd=%vd_dp q=0 +VFML 1111 110 0 s:1 . 10 .... .... 1000 . 1 . 1 .... \ + vm=%vm_dp vn=%vn_dp vd=%vd_dp q=1 diff --git a/target/arm/translate-neon.inc.c b/target/arm/translate-neon.inc.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-neon.inc.c +++ b/target/arm/translate-neon.inc.c @@ -XXX,XX +XXX,XX @@ static bool trans_VDOT(DisasContext *s, arg_VDOT *a) opr_sz, opr_sz, 0, fn_gvec); return true; } + +static bool trans_VFML(DisasContext *s, arg_VFML *a) +{ + int opr_sz; + + if (!dc_isar_feature(aa32_fhm, s)) { + return false; + } + + /* UNDEF accesses to D16-D31 if they don't exist. */ + if (!dc_isar_feature(aa32_simd_r32, s) && + (a->vd & 0x10)) { + return false; + } + + if (a->vd & a->q) { + return false; + } + + if (!vfp_access_check(s)) { + return true; + } + + opr_sz = (1 + a->q) * 8; + tcg_gen_gvec_3_ptr(vfp_reg_offset(1, a->vd), + vfp_reg_offset(a->q, a->vn), + vfp_reg_offset(a->q, a->vm), + cpu_env, opr_sz, opr_sz, a->s, /* is_2 == 0 */ + gen_helper_gvec_fmlal_a32); + return true; +} diff --git a/target/arm/translate.c b/target/arm/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate.c +++ b/target/arm/translate.c @@ -XXX,XX +XXX,XX @@ static int disas_neon_data_insn(DisasContext *s, uint32_t insn) return 0; } -/* Advanced SIMD three registers of the same length extension. - * 31 25 23 22 20 16 12 11 10 9 8 3 0 - * +---------------+-----+---+-----+----+----+---+----+---+----+---------+----+ - * | 1 1 1 1 1 1 0 | op1 | D | op2 | Vn | Vd | 1 | o3 | 0 | o4 | N Q M U | Vm | - * +---------------+-----+---+-----+----+----+---+----+---+----+---------+----+ - */ -static int disas_neon_insn_3same_ext(DisasContext *s, uint32_t insn) -{ - gen_helper_gvec_3 *fn_gvec = NULL; - gen_helper_gvec_3_ptr *fn_gvec_ptr = NULL; - int rd, rn, rm, opr_sz; - int data = 0; - int off_rn, off_rm; - bool is_long = false, q = extract32(insn, 6, 1); - bool ptr_is_env = false; - - if ((insn & 0xff300f10) == 0xfc200810) { - /* VFM[AS]L -- 1111 1100 S.10 .... .... 1000 .Q.1 .... */ - int is_s = extract32(insn, 23, 1); - if (!dc_isar_feature(aa32_fhm, s)) { - return 1; - } - is_long = true; - data = is_s; /* is_2 == 0 */ - fn_gvec_ptr = gen_helper_gvec_fmlal_a32; - ptr_is_env = true; - } else { - return 1; - } - - VFP_DREG_D(rd, insn); - if (rd & q) { - return 1; - } - if (q || !is_long) { - VFP_DREG_N(rn, insn); - VFP_DREG_M(rm, insn); - if ((rn | rm) & q & !is_long) { - return 1; - } - off_rn = vfp_reg_offset(1, rn); - off_rm = vfp_reg_offset(1, rm); - } else { - rn = VFP_SREG_N(insn); - rm = VFP_SREG_M(insn); - off_rn = vfp_reg_offset(0, rn); - off_rm = vfp_reg_offset(0, rm); - } - - if (s->fp_excp_el) { - gen_exception_insn(s, s->pc_curr, EXCP_UDEF, - syn_simd_access_trap(1, 0xe, false), s->fp_excp_el); - return 0; - } - if (!s->vfp_enabled) { - return 1; - } - - opr_sz = (1 + q) * 8; - if (fn_gvec_ptr) { - TCGv_ptr ptr; - if (ptr_is_env) { - ptr = cpu_env; - } else { - ptr = get_fpstatus_ptr(1); - } - tcg_gen_gvec_3_ptr(vfp_reg_offset(1, rd), off_rn, off_rm, ptr, - opr_sz, opr_sz, data, fn_gvec_ptr); - if (!ptr_is_env) { - tcg_temp_free_ptr(ptr); - } - } else { - tcg_gen_gvec_3_ool(vfp_reg_offset(1, rd), off_rn, off_rm, - opr_sz, opr_sz, data, fn_gvec); - } - return 0; -} - /* Advanced SIMD two registers and a scalar extension. * 31 24 23 22 20 16 12 11 10 9 8 3 0 * +-----------------+----+---+----+----+----+---+----+---+----+---------+----+ @@ -XXX,XX +XXX,XX @@ static void disas_arm_insn(DisasContext *s, unsigned int insn) } } } - } else if ((insn & 0x0e000a00) == 0x0c000800 - && arm_dc_feature(s, ARM_FEATURE_V8)) { - if (disas_neon_insn_3same_ext(s, insn)) { - goto illegal_op; - } - return; } else if ((insn & 0x0f000a00) == 0x0e000800 && arm_dc_feature(s, ARM_FEATURE_V8)) { if (disas_neon_insn_2reg_scalar_ext(s, insn)) { @@ -XXX,XX +XXX,XX @@ static void disas_thumb2_insn(DisasContext *s, uint32_t insn) } break; } - if ((insn & 0xfe000a00) == 0xfc000800 + if ((insn & 0xff000a00) == 0xfe000800 && arm_dc_feature(s, ARM_FEATURE_V8)) { /* The Thumb2 and ARM encodings are identical. */ - if (disas_neon_insn_3same_ext(s, insn)) { - goto illegal_op; - } - } else if ((insn & 0xff000a00) == 0xfe000800 - && arm_dc_feature(s, ARM_FEATURE_V8)) { - /* The Thumb2 and ARM encodings are identical. */ if (disas_neon_insn_2reg_scalar_ext(s, insn)) { goto illegal_op; } -- 2.20.1 Convert VCMLA (scalar) in the 2reg-scalar-ext group to decodetree. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200430181003.21682-9-peter.maydell@linaro.org --- target/arm/neon-shared.decode | 5 +++++ target/arm/translate-neon.inc.c | 40 +++++++++++++++++++++++++++++++++ target/arm/translate.c | 26 +-------------------- 3 files changed, 46 insertions(+), 25 deletions(-) diff --git a/target/arm/neon-shared.decode b/target/arm/neon-shared.decode index XXXXXXX..XXXXXXX 100644 --- a/target/arm/neon-shared.decode +++ b/target/arm/neon-shared.decode @@ -XXX,XX +XXX,XX @@ VFML 1111 110 0 s:1 . 10 .... .... 1000 . 0 . 1 .... \ vm=%vm_sp vn=%vn_sp vd=%vd_dp q=0 VFML 1111 110 0 s:1 . 10 .... .... 1000 . 1 . 1 .... \ vm=%vm_dp vn=%vn_dp vd=%vd_dp q=1 + +VCMLA_scalar 1111 1110 0 . rot:2 .... .... 1000 . q:1 index:1 0 vm:4 \ + vn=%vn_dp vd=%vd_dp size=0 +VCMLA_scalar 1111 1110 1 . rot:2 .... .... 1000 . q:1 . 0 .... \ + vm=%vm_dp vn=%vn_dp vd=%vd_dp size=1 index=0 diff --git a/target/arm/translate-neon.inc.c b/target/arm/translate-neon.inc.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-neon.inc.c +++ b/target/arm/translate-neon.inc.c @@ -XXX,XX +XXX,XX @@ static bool trans_VFML(DisasContext *s, arg_VFML *a) gen_helper_gvec_fmlal_a32); return true; } + +static bool trans_VCMLA_scalar(DisasContext *s, arg_VCMLA_scalar *a) +{ + gen_helper_gvec_3_ptr *fn_gvec_ptr; + int opr_sz; + TCGv_ptr fpst; + + if (!dc_isar_feature(aa32_vcma, s)) { + return false; + } + if (a->size == 0 && !dc_isar_feature(aa32_fp16_arith, s)) { + return false; + } + + /* UNDEF accesses to D16-D31 if they don't exist. */ + if (!dc_isar_feature(aa32_simd_r32, s) && + ((a->vd | a->vn | a->vm) & 0x10)) { + return false; + } + + if ((a->vd | a->vn) & a->q) { + return false; + } + + if (!vfp_access_check(s)) { + return true; + } + + fn_gvec_ptr = (a->size ? gen_helper_gvec_fcmlas_idx + : gen_helper_gvec_fcmlah_idx); + opr_sz = (1 + a->q) * 8; + fpst = get_fpstatus_ptr(1); + tcg_gen_gvec_3_ptr(vfp_reg_offset(1, a->vd), + vfp_reg_offset(1, a->vn), + vfp_reg_offset(1, a->vm), + fpst, opr_sz, opr_sz, + (a->index << 2) | a->rot, fn_gvec_ptr); + tcg_temp_free_ptr(fpst); + return true; +} diff --git a/target/arm/translate.c b/target/arm/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate.c +++ b/target/arm/translate.c @@ -XXX,XX +XXX,XX @@ static int disas_neon_insn_2reg_scalar_ext(DisasContext *s, uint32_t insn) bool is_long = false, q = extract32(insn, 6, 1); bool ptr_is_env = false; - if ((insn & 0xff000f10) == 0xfe000800) { - /* VCMLA (indexed) -- 1111 1110 S.RR .... .... 1000 ...0 .... */ - int rot = extract32(insn, 20, 2); - int size = extract32(insn, 23, 1); - int index; - - if (!dc_isar_feature(aa32_vcma, s)) { - return 1; - } - if (size == 0) { - if (!dc_isar_feature(aa32_fp16_arith, s)) { - return 1; - } - /* For fp16, rm is just Vm, and index is M. */ - rm = extract32(insn, 0, 4); - index = extract32(insn, 5, 1); - } else { - /* For fp32, rm is the usual M:Vm, and index is 0. */ - VFP_DREG_M(rm, insn); - index = 0; - } - data = (index << 2) | rot; - fn_gvec_ptr = (size ? gen_helper_gvec_fcmlas_idx - : gen_helper_gvec_fcmlah_idx); - } else if ((insn & 0xffb00f00) == 0xfe200d00) { + if ((insn & 0xffb00f00) == 0xfe200d00) { /* V[US]DOT -- 1111 1110 0.10 .... .... 1101 .Q.U .... */ int u = extract32(insn, 4, 1); -- 2.20.1 Convert the V[US]DOT (scalar) insns in the 2reg-scalar-ext group to decodetree. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200430181003.21682-10-peter.maydell@linaro.org --- target/arm/neon-shared.decode | 3 +++ target/arm/translate-neon.inc.c | 35 +++++++++++++++++++++++++++++++++ target/arm/translate.c | 13 +----------- 3 files changed, 39 insertions(+), 12 deletions(-) diff --git a/target/arm/neon-shared.decode b/target/arm/neon-shared.decode index XXXXXXX..XXXXXXX 100644 --- a/target/arm/neon-shared.decode +++ b/target/arm/neon-shared.decode @@ -XXX,XX +XXX,XX @@ VCMLA_scalar 1111 1110 0 . rot:2 .... .... 1000 . q:1 index:1 0 vm:4 \ vn=%vn_dp vd=%vd_dp size=0 VCMLA_scalar 1111 1110 1 . rot:2 .... .... 1000 . q:1 . 0 .... \ vm=%vm_dp vn=%vn_dp vd=%vd_dp size=1 index=0 + +VDOT_scalar 1111 1110 0 . 10 .... .... 1101 . q:1 index:1 u:1 rm:4 \ + vm=%vm_dp vn=%vn_dp vd=%vd_dp diff --git a/target/arm/translate-neon.inc.c b/target/arm/translate-neon.inc.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-neon.inc.c +++ b/target/arm/translate-neon.inc.c @@ -XXX,XX +XXX,XX @@ static bool trans_VCMLA_scalar(DisasContext *s, arg_VCMLA_scalar *a) tcg_temp_free_ptr(fpst); return true; } + +static bool trans_VDOT_scalar(DisasContext *s, arg_VDOT_scalar *a) +{ + gen_helper_gvec_3 *fn_gvec; + int opr_sz; + TCGv_ptr fpst; + + if (!dc_isar_feature(aa32_dp, s)) { + return false; + } + + /* UNDEF accesses to D16-D31 if they don't exist. */ + if (!dc_isar_feature(aa32_simd_r32, s) && + ((a->vd | a->vn) & 0x10)) { + return false; + } + + if ((a->vd | a->vn) & a->q) { + return false; + } + + if (!vfp_access_check(s)) { + return true; + } + + fn_gvec = a->u ? gen_helper_gvec_udot_idx_b : gen_helper_gvec_sdot_idx_b; + opr_sz = (1 + a->q) * 8; + fpst = get_fpstatus_ptr(1); + tcg_gen_gvec_3_ool(vfp_reg_offset(1, a->vd), + vfp_reg_offset(1, a->vn), + vfp_reg_offset(1, a->rm), + opr_sz, opr_sz, a->index, fn_gvec); + tcg_temp_free_ptr(fpst); + return true; +} diff --git a/target/arm/translate.c b/target/arm/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate.c +++ b/target/arm/translate.c @@ -XXX,XX +XXX,XX @@ static int disas_neon_insn_2reg_scalar_ext(DisasContext *s, uint32_t insn) bool is_long = false, q = extract32(insn, 6, 1); bool ptr_is_env = false; - if ((insn & 0xffb00f00) == 0xfe200d00) { - /* V[US]DOT -- 1111 1110 0.10 .... .... 1101 .Q.U .... */ - int u = extract32(insn, 4, 1); - - if (!dc_isar_feature(aa32_dp, s)) { - return 1; - } - fn_gvec = u ? gen_helper_gvec_udot_idx_b : gen_helper_gvec_sdot_idx_b; - /* rm is just Vm, and index is M. */ - data = extract32(insn, 5, 1); /* index */ - rm = extract32(insn, 0, 4); - } else if ((insn & 0xffa00f10) == 0xfe000810) { + if ((insn & 0xffa00f10) == 0xfe000810) { /* VFM[AS]L -- 1111 1110 0.0S .... .... 1000 .Q.1 .... */ int is_s = extract32(insn, 20, 1); int vm20 = extract32(insn, 0, 3); -- 2.20.1 Convert the VFM[AS]L (scalar) insns in the 2reg-scalar-ext group to decodetree. These are the last ones in the group so we can remove all the legacy decode for the group. Note that in disas_thumb2_insn() the parts of this encoding space where the decodetree decoder returns false will correctly be directed to illegal_op by the "(insn & (1 << 28))" check so they won't fall into disas_coproc_insn() by mistake. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200430181003.21682-11-peter.maydell@linaro.org --- target/arm/neon-shared.decode | 7 +++ target/arm/translate-neon.inc.c | 32 ++++++++++ target/arm/translate.c | 107 +------------------------------- 3 files changed, 40 insertions(+), 106 deletions(-) diff --git a/target/arm/neon-shared.decode b/target/arm/neon-shared.decode index XXXXXXX..XXXXXXX 100644 --- a/target/arm/neon-shared.decode +++ b/target/arm/neon-shared.decode @@ -XXX,XX +XXX,XX @@ VCMLA_scalar 1111 1110 1 . rot:2 .... .... 1000 . q:1 . 0 .... \ VDOT_scalar 1111 1110 0 . 10 .... .... 1101 . q:1 index:1 u:1 rm:4 \ vm=%vm_dp vn=%vn_dp vd=%vd_dp + +%vfml_scalar_q0_rm 0:3 5:1 +%vfml_scalar_q1_index 5:1 3:1 +VFML_scalar 1111 1110 0 . 0 s:1 .... .... 1000 . 0 . 1 index:1 ... \ + rm=%vfml_scalar_q0_rm vn=%vn_sp vd=%vd_dp q=0 +VFML_scalar 1111 1110 0 . 0 s:1 .... .... 1000 . 1 . 1 . rm:3 \ + index=%vfml_scalar_q1_index vn=%vn_dp vd=%vd_dp q=1 diff --git a/target/arm/translate-neon.inc.c b/target/arm/translate-neon.inc.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-neon.inc.c +++ b/target/arm/translate-neon.inc.c @@ -XXX,XX +XXX,XX @@ static bool trans_VDOT_scalar(DisasContext *s, arg_VDOT_scalar *a) tcg_temp_free_ptr(fpst); return true; } + +static bool trans_VFML_scalar(DisasContext *s, arg_VFML_scalar *a) +{ + int opr_sz; + + if (!dc_isar_feature(aa32_fhm, s)) { + return false; + } + + /* UNDEF accesses to D16-D31 if they don't exist. */ + if (!dc_isar_feature(aa32_simd_r32, s) && + ((a->vd & 0x10) || (a->q && (a->vn & 0x10)))) { + return false; + } + + if (a->vd & a->q) { + return false; + } + + if (!vfp_access_check(s)) { + return true; + } + + opr_sz = (1 + a->q) * 8; + tcg_gen_gvec_3_ptr(vfp_reg_offset(1, a->vd), + vfp_reg_offset(a->q, a->vn), + vfp_reg_offset(a->q, a->rm), + cpu_env, opr_sz, opr_sz, + (a->index << 2) | a->s, /* is_2 == 0 */ + gen_helper_gvec_fmlal_idx_a32); + return true; +} diff --git a/target/arm/translate.c b/target/arm/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate.c +++ b/target/arm/translate.c @@ -XXX,XX +XXX,XX @@ static int disas_dsp_insn(DisasContext *s, uint32_t insn) } #define VFP_REG_SHR(x, n) (((n) > 0) ? (x) >> (n) : (x) << -(n)) -#define VFP_SREG(insn, bigbit, smallbit) \ - ((VFP_REG_SHR(insn, bigbit - 1) & 0x1e) | (((insn) >> (smallbit)) & 1)) #define VFP_DREG(reg, insn, bigbit, smallbit) do { \ if (dc_isar_feature(aa32_simd_r32, s)) { \ reg = (((insn) >> (bigbit)) & 0x0f) \ @@ -XXX,XX +XXX,XX @@ static int disas_dsp_insn(DisasContext *s, uint32_t insn) reg = ((insn) >> (bigbit)) & 0x0f; \ }} while (0) -#define VFP_SREG_D(insn) VFP_SREG(insn, 12, 22) #define VFP_DREG_D(reg, insn) VFP_DREG(reg, insn, 12, 22) -#define VFP_SREG_N(insn) VFP_SREG(insn, 16, 7) #define VFP_DREG_N(reg, insn) VFP_DREG(reg, insn, 16, 7) -#define VFP_SREG_M(insn) VFP_SREG(insn, 0, 5) #define VFP_DREG_M(reg, insn) VFP_DREG(reg, insn, 0, 5) static void gen_neon_dup_low16(TCGv_i32 var) @@ -XXX,XX +XXX,XX @@ static int disas_neon_data_insn(DisasContext *s, uint32_t insn) return 0; } -/* Advanced SIMD two registers and a scalar extension. - * 31 24 23 22 20 16 12 11 10 9 8 3 0 - * +-----------------+----+---+----+----+----+---+----+---+----+---------+----+ - * | 1 1 1 1 1 1 1 0 | o1 | D | o2 | Vn | Vd | 1 | o3 | 0 | o4 | N Q M U | Vm | - * +-----------------+----+---+----+----+----+---+----+---+----+---------+----+ - * - */ - -static int disas_neon_insn_2reg_scalar_ext(DisasContext *s, uint32_t insn) -{ - gen_helper_gvec_3 *fn_gvec = NULL; - gen_helper_gvec_3_ptr *fn_gvec_ptr = NULL; - int rd, rn, rm, opr_sz, data; - int off_rn, off_rm; - bool is_long = false, q = extract32(insn, 6, 1); - bool ptr_is_env = false; - - if ((insn & 0xffa00f10) == 0xfe000810) { - /* VFM[AS]L -- 1111 1110 0.0S .... .... 1000 .Q.1 .... */ - int is_s = extract32(insn, 20, 1); - int vm20 = extract32(insn, 0, 3); - int vm3 = extract32(insn, 3, 1); - int m = extract32(insn, 5, 1); - int index; - - if (!dc_isar_feature(aa32_fhm, s)) { - return 1; - } - if (q) { - rm = vm20; - index = m * 2 + vm3; - } else { - rm = vm20 * 2 + m; - index = vm3; - } - is_long = true; - data = (index << 2) | is_s; /* is_2 == 0 */ - fn_gvec_ptr = gen_helper_gvec_fmlal_idx_a32; - ptr_is_env = true; - } else { - return 1; - } - - VFP_DREG_D(rd, insn); - if (rd & q) { - return 1; - } - if (q || !is_long) { - VFP_DREG_N(rn, insn); - if (rn & q & !is_long) { - return 1; - } - off_rn = vfp_reg_offset(1, rn); - off_rm = vfp_reg_offset(1, rm); - } else { - rn = VFP_SREG_N(insn); - off_rn = vfp_reg_offset(0, rn); - off_rm = vfp_reg_offset(0, rm); - } - if (s->fp_excp_el) { - gen_exception_insn(s, s->pc_curr, EXCP_UDEF, - syn_simd_access_trap(1, 0xe, false), s->fp_excp_el); - return 0; - } - if (!s->vfp_enabled) { - return 1; - } - - opr_sz = (1 + q) * 8; - if (fn_gvec_ptr) { - TCGv_ptr ptr; - if (ptr_is_env) { - ptr = cpu_env; - } else { - ptr = get_fpstatus_ptr(1); - } - tcg_gen_gvec_3_ptr(vfp_reg_offset(1, rd), off_rn, off_rm, ptr, - opr_sz, opr_sz, data, fn_gvec_ptr); - if (!ptr_is_env) { - tcg_temp_free_ptr(ptr); - } - } else { - tcg_gen_gvec_3_ool(vfp_reg_offset(1, rd), off_rn, off_rm, - opr_sz, opr_sz, data, fn_gvec); - } - return 0; -} - static int disas_coproc_insn(DisasContext *s, uint32_t insn) { int cpnum, is64, crn, crm, opc1, opc2, isread, rt, rt2; @@ -XXX,XX +XXX,XX @@ static void disas_arm_insn(DisasContext *s, unsigned int insn) } } } - } else if ((insn & 0x0f000a00) == 0x0e000800 - && arm_dc_feature(s, ARM_FEATURE_V8)) { - if (disas_neon_insn_2reg_scalar_ext(s, insn)) { - goto illegal_op; - } - return; } goto illegal_op; } @@ -XXX,XX +XXX,XX @@ static void disas_thumb2_insn(DisasContext *s, uint32_t insn) } break; } - if ((insn & 0xff000a00) == 0xfe000800 - && arm_dc_feature(s, ARM_FEATURE_V8)) { - /* The Thumb2 and ARM encodings are identical. */ - if (disas_neon_insn_2reg_scalar_ext(s, insn)) { - goto illegal_op; - } - } else if (((insn >> 24) & 3) == 3) { + if (((insn >> 24) & 3) == 3) { /* Translate into the equivalent ARM encoding. */ insn = (insn & 0xe2ffffff) | ((insn & (1 << 28)) >> 4) | (1 << 28); if (disas_neon_data_insn(s, insn)) { -- 2.20.1 Convert the Neon "load/store multiple structures" insns to decodetree. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200430181003.21682-12-peter.maydell@linaro.org --- target/arm/neon-ls.decode | 7 ++ target/arm/translate-neon.inc.c | 124 ++++++++++++++++++++++++++++++++ target/arm/translate.c | 91 +---------------------- 3 files changed, 133 insertions(+), 89 deletions(-) diff --git a/target/arm/neon-ls.decode b/target/arm/neon-ls.decode index XXXXXXX..XXXXXXX 100644 --- a/target/arm/neon-ls.decode +++ b/target/arm/neon-ls.decode @@ -XXX,XX +XXX,XX @@ # 0b1111_1001_xxx0_xxxx_xxxx_xxxx_xxxx_xxxx # This file works on the A32 encoding only; calling code for T32 has to # transform the insn into the A32 version first. + +%vd_dp 22:1 12:4 + +# Neon load/store multiple structures + +VLDST_multiple 1111 0100 0 . l:1 0 rn:4 .... itype:4 size:2 align:2 rm:4 \ + vd=%vd_dp diff --git a/target/arm/translate-neon.inc.c b/target/arm/translate-neon.inc.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-neon.inc.c +++ b/target/arm/translate-neon.inc.c @@ -XXX,XX +XXX,XX @@ static bool trans_VFML_scalar(DisasContext *s, arg_VFML_scalar *a) gen_helper_gvec_fmlal_idx_a32); return true; } + +static struct { + int nregs; + int interleave; + int spacing; +} const neon_ls_element_type[11] = { + {1, 4, 1}, + {1, 4, 2}, + {4, 1, 1}, + {2, 2, 2}, + {1, 3, 1}, + {1, 3, 2}, + {3, 1, 1}, + {1, 1, 1}, + {1, 2, 1}, + {1, 2, 2}, + {2, 1, 1} +}; + +static void gen_neon_ldst_base_update(DisasContext *s, int rm, int rn, + int stride) +{ + if (rm != 15) { + TCGv_i32 base; + + base = load_reg(s, rn); + if (rm == 13) { + tcg_gen_addi_i32(base, base, stride); + } else { + TCGv_i32 index; + index = load_reg(s, rm); + tcg_gen_add_i32(base, base, index); + tcg_temp_free_i32(index); + } + store_reg(s, rn, base); + } +} + +static bool trans_VLDST_multiple(DisasContext *s, arg_VLDST_multiple *a) +{ + /* Neon load/store multiple structures */ + int nregs, interleave, spacing, reg, n; + MemOp endian = s->be_data; + int mmu_idx = get_mem_index(s); + int size = a->size; + TCGv_i64 tmp64; + TCGv_i32 addr, tmp; + + if (!arm_dc_feature(s, ARM_FEATURE_NEON)) { + return false; + } + + /* UNDEF accesses to D16-D31 if they don't exist */ + if (!dc_isar_feature(aa32_simd_r32, s) && (a->vd & 0x10)) { + return false; + } + if (a->itype > 10) { + return false; + } + /* Catch UNDEF cases for bad values of align field */ + switch (a->itype & 0xc) { + case 4: + if (a->align >= 2) { + return false; + } + break; + case 8: + if (a->align == 3) { + return false; + } + break; + default: + break; + } + nregs = neon_ls_element_type[a->itype].nregs; + interleave = neon_ls_element_type[a->itype].interleave; + spacing = neon_ls_element_type[a->itype].spacing; + if (size == 3 && (interleave | spacing) != 1) { + return false; + } + + if (!vfp_access_check(s)) { + return true; + } + + /* For our purposes, bytes are always little-endian. */ + if (size == 0) { + endian = MO_LE; + } + /* + * Consecutive little-endian elements from a single register + * can be promoted to a larger little-endian operation. + */ + if (interleave == 1 && endian == MO_LE) { + size = 3; + } + tmp64 = tcg_temp_new_i64(); + addr = tcg_temp_new_i32(); + tmp = tcg_const_i32(1 << size); + load_reg_var(s, addr, a->rn); + for (reg = 0; reg < nregs; reg++) { + for (n = 0; n < 8 >> size; n++) { + int xs; + for (xs = 0; xs < interleave; xs++) { + int tt = a->vd + reg + spacing * xs; + + if (a->l) { + gen_aa32_ld_i64(s, tmp64, addr, mmu_idx, endian | size); + neon_store_element64(tt, n, size, tmp64); + } else { + neon_load_element64(tmp64, tt, n, size); + gen_aa32_st_i64(s, tmp64, addr, mmu_idx, endian | size); + } + tcg_gen_add_i32(addr, addr, tmp); + } + } + } + tcg_temp_free_i32(addr); + tcg_temp_free_i32(tmp); + tcg_temp_free_i64(tmp64); + + gen_neon_ldst_base_update(s, a->rm, a->rn, nregs * interleave * 8); + return true; +} diff --git a/target/arm/translate.c b/target/arm/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate.c +++ b/target/arm/translate.c @@ -XXX,XX +XXX,XX @@ static void gen_neon_trn_u16(TCGv_i32 t0, TCGv_i32 t1) } -static struct { - int nregs; - int interleave; - int spacing; -} const neon_ls_element_type[11] = { - {1, 4, 1}, - {1, 4, 2}, - {4, 1, 1}, - {2, 2, 2}, - {1, 3, 1}, - {1, 3, 2}, - {3, 1, 1}, - {1, 1, 1}, - {1, 2, 1}, - {1, 2, 2}, - {2, 1, 1} -}; - /* Translate a NEON load/store element instruction. Return nonzero if the instruction is invalid. */ static int disas_neon_ls_insn(DisasContext *s, uint32_t insn) { int rd, rn, rm; - int op; int nregs; - int interleave; - int spacing; int stride; int size; int reg; int load; - int n; int vec_size; - int mmu_idx; - MemOp endian; TCGv_i32 addr; TCGv_i32 tmp; - TCGv_i32 tmp2; - TCGv_i64 tmp64; if (!arm_dc_feature(s, ARM_FEATURE_NEON)) { return 1; @@ -XXX,XX +XXX,XX @@ static int disas_neon_ls_insn(DisasContext *s, uint32_t insn) rn = (insn >> 16) & 0xf; rm = insn & 0xf; load = (insn & (1 << 21)) != 0; - endian = s->be_data; - mmu_idx = get_mem_index(s); if ((insn & (1 << 23)) == 0) { - /* Load store all elements. */ - op = (insn >> 8) & 0xf; - size = (insn >> 6) & 3; - if (op > 10) - return 1; - /* Catch UNDEF cases for bad values of align field */ - switch (op & 0xc) { - case 4: - if (((insn >> 5) & 1) == 1) { - return 1; - } - break; - case 8: - if (((insn >> 4) & 3) == 3) { - return 1; - } - break; - default: - break; - } - nregs = neon_ls_element_type[op].nregs; - interleave = neon_ls_element_type[op].interleave; - spacing = neon_ls_element_type[op].spacing; - if (size == 3 && (interleave | spacing) != 1) { - return 1; - } - /* For our purposes, bytes are always little-endian. */ - if (size == 0) { - endian = MO_LE; - } - /* Consecutive little-endian elements from a single register - * can be promoted to a larger little-endian operation. - */ - if (interleave == 1 && endian == MO_LE) { - size = 3; - } - tmp64 = tcg_temp_new_i64(); - addr = tcg_temp_new_i32(); - tmp2 = tcg_const_i32(1 << size); - load_reg_var(s, addr, rn); - for (reg = 0; reg < nregs; reg++) { - for (n = 0; n < 8 >> size; n++) { - int xs; - for (xs = 0; xs < interleave; xs++) { - int tt = rd + reg + spacing * xs; - - if (load) { - gen_aa32_ld_i64(s, tmp64, addr, mmu_idx, endian | size); - neon_store_element64(tt, n, size, tmp64); - } else { - neon_load_element64(tmp64, tt, n, size); - gen_aa32_st_i64(s, tmp64, addr, mmu_idx, endian | size); - } - tcg_gen_add_i32(addr, addr, tmp2); - } - } - } - tcg_temp_free_i32(addr); - tcg_temp_free_i32(tmp2); - tcg_temp_free_i64(tmp64); - stride = nregs * interleave * 8; + /* Load store all elements -- handled already by decodetree */ + return 1; } else { size = (insn >> 10) & 3; if (size == 3) { -- 2.20.1 Convert the Neon "load single structure to all lanes" insns to decodetree. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200430181003.21682-13-peter.maydell@linaro.org --- target/arm/neon-ls.decode | 5 +++ target/arm/translate-neon.inc.c | 73 +++++++++++++++++++++++++++++++++ target/arm/translate.c | 55 +------------------------ 3 files changed, 80 insertions(+), 53 deletions(-) diff --git a/target/arm/neon-ls.decode b/target/arm/neon-ls.decode index XXXXXXX..XXXXXXX 100644 --- a/target/arm/neon-ls.decode +++ b/target/arm/neon-ls.decode @@ -XXX,XX +XXX,XX @@ VLDST_multiple 1111 0100 0 . l:1 0 rn:4 .... itype:4 size:2 align:2 rm:4 \ vd=%vd_dp + +# Neon load single element to all lanes + +VLD_all_lanes 1111 0100 1 . 1 0 rn:4 .... 11 n:2 size:2 t:1 a:1 rm:4 \ + vd=%vd_dp diff --git a/target/arm/translate-neon.inc.c b/target/arm/translate-neon.inc.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-neon.inc.c +++ b/target/arm/translate-neon.inc.c @@ -XXX,XX +XXX,XX @@ static bool trans_VLDST_multiple(DisasContext *s, arg_VLDST_multiple *a) gen_neon_ldst_base_update(s, a->rm, a->rn, nregs * interleave * 8); return true; } + +static bool trans_VLD_all_lanes(DisasContext *s, arg_VLD_all_lanes *a) +{ + /* Neon load single structure to all lanes */ + int reg, stride, vec_size; + int vd = a->vd; + int size = a->size; + int nregs = a->n + 1; + TCGv_i32 addr, tmp; + + if (!arm_dc_feature(s, ARM_FEATURE_NEON)) { + return false; + } + + /* UNDEF accesses to D16-D31 if they don't exist */ + if (!dc_isar_feature(aa32_simd_r32, s) && (a->vd & 0x10)) { + return false; + } + + if (size == 3) { + if (nregs != 4 || a->a == 0) { + return false; + } + /* For VLD4 size == 3 a == 1 means 32 bits at 16 byte alignment */ + size = 2; + } + if (nregs == 1 && a->a == 1 && size == 0) { + return false; + } + if (nregs == 3 && a->a == 1) { + return false; + } + + if (!vfp_access_check(s)) { + return true; + } + + /* + * VLD1 to all lanes: T bit indicates how many Dregs to write. + * VLD2/3/4 to all lanes: T bit indicates register stride. + */ + stride = a->t ? 2 : 1; + vec_size = nregs == 1 ? stride * 8 : 8; + + tmp = tcg_temp_new_i32(); + addr = tcg_temp_new_i32(); + load_reg_var(s, addr, a->rn); + for (reg = 0; reg < nregs; reg++) { + gen_aa32_ld_i32(s, tmp, addr, get_mem_index(s), + s->be_data | size); + if ((vd & 1) && vec_size == 16) { + /* + * We cannot write 16 bytes at once because the + * destination is unaligned. + */ + tcg_gen_gvec_dup_i32(size, neon_reg_offset(vd, 0), + 8, 8, tmp); + tcg_gen_gvec_mov(0, neon_reg_offset(vd + 1, 0), + neon_reg_offset(vd, 0), 8, 8); + } else { + tcg_gen_gvec_dup_i32(size, neon_reg_offset(vd, 0), + vec_size, vec_size, tmp); + } + tcg_gen_addi_i32(addr, addr, 1 << size); + vd += stride; + } + tcg_temp_free_i32(tmp); + tcg_temp_free_i32(addr); + + gen_neon_ldst_base_update(s, a->rm, a->rn, (1 << size) * nregs); + + return true; +} diff --git a/target/arm/translate.c b/target/arm/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate.c +++ b/target/arm/translate.c @@ -XXX,XX +XXX,XX @@ static int disas_neon_ls_insn(DisasContext *s, uint32_t insn) int size; int reg; int load; - int vec_size; TCGv_i32 addr; TCGv_i32 tmp; @@ -XXX,XX +XXX,XX @@ static int disas_neon_ls_insn(DisasContext *s, uint32_t insn) } else { size = (insn >> 10) & 3; if (size == 3) { - /* Load single element to all lanes. */ - int a = (insn >> 4) & 1; - if (!load) { - return 1; - } - size = (insn >> 6) & 3; - nregs = ((insn >> 8) & 3) + 1; - - if (size == 3) { - if (nregs != 4 || a == 0) { - return 1; - } - /* For VLD4 size==3 a == 1 means 32 bits at 16 byte alignment */ - size = 2; - } - if (nregs == 1 && a == 1 && size == 0) { - return 1; - } - if (nregs == 3 && a == 1) { - return 1; - } - addr = tcg_temp_new_i32(); - load_reg_var(s, addr, rn); - - /* VLD1 to all lanes: bit 5 indicates how many Dregs to write. - * VLD2/3/4 to all lanes: bit 5 indicates register stride. - */ - stride = (insn & (1 << 5)) ? 2 : 1; - vec_size = nregs == 1 ? stride * 8 : 8; - - tmp = tcg_temp_new_i32(); - for (reg = 0; reg < nregs; reg++) { - gen_aa32_ld_i32(s, tmp, addr, get_mem_index(s), - s->be_data | size); - if ((rd & 1) && vec_size == 16) { - /* We cannot write 16 bytes at once because the - * destination is unaligned. - */ - tcg_gen_gvec_dup_i32(size, neon_reg_offset(rd, 0), - 8, 8, tmp); - tcg_gen_gvec_mov(0, neon_reg_offset(rd + 1, 0), - neon_reg_offset(rd, 0), 8, 8); - } else { - tcg_gen_gvec_dup_i32(size, neon_reg_offset(rd, 0), - vec_size, vec_size, tmp); - } - tcg_gen_addi_i32(addr, addr, 1 << size); - rd += stride; - } - tcg_temp_free_i32(tmp); - tcg_temp_free_i32(addr); - stride = (1 << size) * nregs; + /* Load single element to all lanes -- handled by decodetree */ + return 1; } else { /* Single element. */ int idx = (insn >> 4) & 0xf; -- 2.20.1 Convert the Neon "load/store single structure to one lane" insns to decodetree. As this is the last set of insns in the neon load/store group, we can remove the whole disas_neon_ls_insn() function. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200430181003.21682-14-peter.maydell@linaro.org --- target/arm/neon-ls.decode | 11 +++ target/arm/translate-neon.inc.c | 89 +++++++++++++++++++ target/arm/translate.c | 147 -------------------------------- 3 files changed, 100 insertions(+), 147 deletions(-) diff --git a/target/arm/neon-ls.decode b/target/arm/neon-ls.decode index XXXXXXX..XXXXXXX 100644 --- a/target/arm/neon-ls.decode +++ b/target/arm/neon-ls.decode @@ -XXX,XX +XXX,XX @@ VLDST_multiple 1111 0100 0 . l:1 0 rn:4 .... itype:4 size:2 align:2 rm:4 \ VLD_all_lanes 1111 0100 1 . 1 0 rn:4 .... 11 n:2 size:2 t:1 a:1 rm:4 \ vd=%vd_dp + +# Neon load/store single structure to one lane +%imm1_5_p1 5:1 !function=plus1 +%imm1_6_p1 6:1 !function=plus1 + +VLDST_single 1111 0100 1 . l:1 0 rn:4 .... 00 n:2 reg_idx:3 align:1 rm:4 \ + vd=%vd_dp size=0 stride=1 +VLDST_single 1111 0100 1 . l:1 0 rn:4 .... 01 n:2 reg_idx:2 align:2 rm:4 \ + vd=%vd_dp size=1 stride=%imm1_5_p1 +VLDST_single 1111 0100 1 . l:1 0 rn:4 .... 10 n:2 reg_idx:1 align:3 rm:4 \ + vd=%vd_dp size=2 stride=%imm1_6_p1 diff --git a/target/arm/translate-neon.inc.c b/target/arm/translate-neon.inc.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-neon.inc.c +++ b/target/arm/translate-neon.inc.c @@ -XXX,XX +XXX,XX @@ * It might be possible to convert it to a standalone .c file eventually. */ +static inline int plus1(DisasContext *s, int x) +{ + return x + 1; +} + /* Include the generated Neon decoder */ #include "decode-neon-dp.inc.c" #include "decode-neon-ls.inc.c" @@ -XXX,XX +XXX,XX @@ static bool trans_VLD_all_lanes(DisasContext *s, arg_VLD_all_lanes *a) return true; } + +static bool trans_VLDST_single(DisasContext *s, arg_VLDST_single *a) +{ + /* Neon load/store single structure to one lane */ + int reg; + int nregs = a->n + 1; + int vd = a->vd; + TCGv_i32 addr, tmp; + + if (!arm_dc_feature(s, ARM_FEATURE_NEON)) { + return false; + } + + /* UNDEF accesses to D16-D31 if they don't exist */ + if (!dc_isar_feature(aa32_simd_r32, s) && (a->vd & 0x10)) { + return false; + } + + /* Catch the UNDEF cases. This is unavoidably a bit messy. */ + switch (nregs) { + case 1: + if (((a->align & (1 << a->size)) != 0) || + (a->size == 2 && ((a->align & 3) == 1 || (a->align & 3) == 2))) { + return false; + } + break; + case 3: + if ((a->align & 1) != 0) { + return false; + } + /* fall through */ + case 2: + if (a->size == 2 && (a->align & 2) != 0) { + return false; + } + break; + case 4: + if ((a->size == 2) && ((a->align & 3) == 3)) { + return false; + } + break; + default: + abort(); + } + if ((vd + a->stride * (nregs - 1)) > 31) { + /* + * Attempts to write off the end of the register file are + * UNPREDICTABLE; we choose to UNDEF because otherwise we would + * access off the end of the array that holds the register data. + */ + return false; + } + + if (!vfp_access_check(s)) { + return true; + } + + tmp = tcg_temp_new_i32(); + addr = tcg_temp_new_i32(); + load_reg_var(s, addr, a->rn); + /* + * TODO: if we implemented alignment exceptions, we should check + * addr against the alignment encoded in a->align here. + */ + for (reg = 0; reg < nregs; reg++) { + if (a->l) { + gen_aa32_ld_i32(s, tmp, addr, get_mem_index(s), + s->be_data | a->size); + neon_store_element(vd, a->reg_idx, a->size, tmp); + } else { /* Store */ + neon_load_element(tmp, vd, a->reg_idx, a->size); + gen_aa32_st_i32(s, tmp, addr, get_mem_index(s), + s->be_data | a->size); + } + vd += a->stride; + tcg_gen_addi_i32(addr, addr, 1 << a->size); + } + tcg_temp_free_i32(addr); + tcg_temp_free_i32(tmp); + + gen_neon_ldst_base_update(s, a->rm, a->rn, (1 << a->size) * nregs); + + return true; +} diff --git a/target/arm/translate.c b/target/arm/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate.c +++ b/target/arm/translate.c @@ -XXX,XX +XXX,XX @@ static void gen_neon_trn_u16(TCGv_i32 t0, TCGv_i32 t1) tcg_temp_free_i32(rd); } - -/* Translate a NEON load/store element instruction. Return nonzero if the - instruction is invalid. */ -static int disas_neon_ls_insn(DisasContext *s, uint32_t insn) -{ - int rd, rn, rm; - int nregs; - int stride; - int size; - int reg; - int load; - TCGv_i32 addr; - TCGv_i32 tmp; - - if (!arm_dc_feature(s, ARM_FEATURE_NEON)) { - return 1; - } - - /* FIXME: this access check should not take precedence over UNDEF - * for invalid encodings; we will generate incorrect syndrome information - * for attempts to execute invalid vfp/neon encodings with FP disabled. - */ - if (s->fp_excp_el) { - gen_exception_insn(s, s->pc_curr, EXCP_UDEF, - syn_simd_access_trap(1, 0xe, false), s->fp_excp_el); - return 0; - } - - if (!s->vfp_enabled) - return 1; - VFP_DREG_D(rd, insn); - rn = (insn >> 16) & 0xf; - rm = insn & 0xf; - load = (insn & (1 << 21)) != 0; - if ((insn & (1 << 23)) == 0) { - /* Load store all elements -- handled already by decodetree */ - return 1; - } else { - size = (insn >> 10) & 3; - if (size == 3) { - /* Load single element to all lanes -- handled by decodetree */ - return 1; - } else { - /* Single element. */ - int idx = (insn >> 4) & 0xf; - int reg_idx; - switch (size) { - case 0: - reg_idx = (insn >> 5) & 7; - stride = 1; - break; - case 1: - reg_idx = (insn >> 6) & 3; - stride = (insn & (1 << 5)) ? 2 : 1; - break; - case 2: - reg_idx = (insn >> 7) & 1; - stride = (insn & (1 << 6)) ? 2 : 1; - break; - default: - abort(); - } - nregs = ((insn >> 8) & 3) + 1; - /* Catch the UNDEF cases. This is unavoidably a bit messy. */ - switch (nregs) { - case 1: - if (((idx & (1 << size)) != 0) || - (size == 2 && ((idx & 3) == 1 || (idx & 3) == 2))) { - return 1; - } - break; - case 3: - if ((idx & 1) != 0) { - return 1; - } - /* fall through */ - case 2: - if (size == 2 && (idx & 2) != 0) { - return 1; - } - break; - case 4: - if ((size == 2) && ((idx & 3) == 3)) { - return 1; - } - break; - default: - abort(); - } - if ((rd + stride * (nregs - 1)) > 31) { - /* Attempts to write off the end of the register file - * are UNPREDICTABLE; we choose to UNDEF because otherwise - * the neon_load_reg() would write off the end of the array. - */ - return 1; - } - tmp = tcg_temp_new_i32(); - addr = tcg_temp_new_i32(); - load_reg_var(s, addr, rn); - for (reg = 0; reg < nregs; reg++) { - if (load) { - gen_aa32_ld_i32(s, tmp, addr, get_mem_index(s), - s->be_data | size); - neon_store_element(rd, reg_idx, size, tmp); - } else { /* Store */ - neon_load_element(tmp, rd, reg_idx, size); - gen_aa32_st_i32(s, tmp, addr, get_mem_index(s), - s->be_data | size); - } - rd += stride; - tcg_gen_addi_i32(addr, addr, 1 << size); - } - tcg_temp_free_i32(addr); - tcg_temp_free_i32(tmp); - stride = nregs * (1 << size); - } - } - if (rm != 15) { - TCGv_i32 base; - - base = load_reg(s, rn); - if (rm == 13) { - tcg_gen_addi_i32(base, base, stride); - } else { - TCGv_i32 index; - index = load_reg(s, rm); - tcg_gen_add_i32(base, base, index); - tcg_temp_free_i32(index); - } - store_reg(s, rn, base); - } - return 0; -} - static inline void gen_neon_narrow(int size, TCGv_i32 dest, TCGv_i64 src) { switch (size) { @@ -XXX,XX +XXX,XX @@ static void disas_arm_insn(DisasContext *s, unsigned int insn) } return; } - if ((insn & 0x0f100000) == 0x04000000) { - /* NEON load/store. */ - if (disas_neon_ls_insn(s, insn)) { - goto illegal_op; - } - return; - } if ((insn & 0x0e000f00) == 0x0c000100) { if (arm_dc_feature(s, ARM_FEATURE_IWMMXT)) { /* iWMMXt register transfer. */ @@ -XXX,XX +XXX,XX @@ static void disas_thumb2_insn(DisasContext *s, uint32_t insn) } break; case 12: - if ((insn & 0x01100000) == 0x01000000) { - if (disas_neon_ls_insn(s, insn)) { - goto illegal_op; - } - break; - } goto illegal_op; default: illegal_op: -- 2.20.1 Convert the Neon 3-reg-same VADD and VSUB insns to decodetree. Note that we don't need the neon_3r_sizes[op] check here because all size values are OK for VADD and VSUB; we'll add this when we convert the first insn that has size restrictions. For this we need one of the GVecGen*Fn typedefs currently in translate-a64.h; move them all to translate.h as a block so they are visible to the 32-bit decoder. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200430181003.21682-15-peter.maydell@linaro.org --- target/arm/translate-a64.h | 9 -------- target/arm/translate.h | 9 ++++++++ target/arm/neon-dp.decode | 17 +++++++++++++++ target/arm/translate-neon.inc.c | 38 +++++++++++++++++++++++++++++++++ target/arm/translate.c | 14 ++++-------- 5 files changed, 68 insertions(+), 19 deletions(-) diff --git a/target/arm/translate-a64.h b/target/arm/translate-a64.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-a64.h +++ b/target/arm/translate-a64.h @@ -XXX,XX +XXX,XX @@ static inline int vec_full_reg_size(DisasContext *s) bool disas_sve(DisasContext *, uint32_t); -/* Note that the gvec expanders operate on offsets + sizes. */ -typedef void GVecGen2Fn(unsigned, uint32_t, uint32_t, uint32_t, uint32_t); -typedef void GVecGen2iFn(unsigned, uint32_t, uint32_t, int64_t, - uint32_t, uint32_t); -typedef void GVecGen3Fn(unsigned, uint32_t, uint32_t, - uint32_t, uint32_t, uint32_t); -typedef void GVecGen4Fn(unsigned, uint32_t, uint32_t, uint32_t, - uint32_t, uint32_t, uint32_t); - #endif /* TARGET_ARM_TRANSLATE_A64_H */ 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 @@ void gen_sshl_i64(TCGv_i64 d, TCGv_i64 a, TCGv_i64 b); #define dc_isar_feature(name, ctx) \ ({ DisasContext *ctx_ = (ctx); isar_feature_##name(ctx_->isar); }) +/* Note that the gvec expanders operate on offsets + sizes. */ +typedef void GVecGen2Fn(unsigned, uint32_t, uint32_t, uint32_t, uint32_t); +typedef void GVecGen2iFn(unsigned, uint32_t, uint32_t, int64_t, + uint32_t, uint32_t); +typedef void GVecGen3Fn(unsigned, uint32_t, uint32_t, + uint32_t, uint32_t, uint32_t); +typedef void GVecGen4Fn(unsigned, uint32_t, uint32_t, uint32_t, + uint32_t, uint32_t, uint32_t); + #endif /* TARGET_ARM_TRANSLATE_H */ diff --git a/target/arm/neon-dp.decode b/target/arm/neon-dp.decode index XXXXXXX..XXXXXXX 100644 --- a/target/arm/neon-dp.decode +++ b/target/arm/neon-dp.decode @@ -XXX,XX +XXX,XX @@ # # This file is processed by scripts/decodetree.py # +# VFP/Neon register fields; same as vfp.decode +%vm_dp 5:1 0:4 +%vn_dp 7:1 16:4 +%vd_dp 22:1 12:4 # Encodings for Neon data processing instructions where the T32 encoding # is a simple transformation of the A32 encoding. @@ -XXX,XX +XXX,XX @@ # 0b111p_1111_qqqq_qqqq_qqqq_qqqq_qqqq_qqqq # This file works on the A32 encoding only; calling code for T32 has to # transform the insn into the A32 version first. + +###################################################################### +# 3-reg-same grouping: +# 1111 001 U 0 D sz:2 Vn:4 Vd:4 opc:4 N Q M op Vm:4 +###################################################################### + +&3same vm vn vd q size + +@3same .... ... . . . size:2 .... .... .... . q:1 . . .... \ + &3same vm=%vm_dp vn=%vn_dp vd=%vd_dp + +VADD_3s 1111 001 0 0 . .. .... .... 1000 . . . 0 .... @3same +VSUB_3s 1111 001 1 0 . .. .... .... 1000 . . . 0 .... @3same diff --git a/target/arm/translate-neon.inc.c b/target/arm/translate-neon.inc.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-neon.inc.c +++ b/target/arm/translate-neon.inc.c @@ -XXX,XX +XXX,XX @@ static bool trans_VLDST_single(DisasContext *s, arg_VLDST_single *a) return true; } + +static bool do_3same(DisasContext *s, arg_3same *a, GVecGen3Fn fn) +{ + int vec_size = a->q ? 16 : 8; + int rd_ofs = neon_reg_offset(a->vd, 0); + int rn_ofs = neon_reg_offset(a->vn, 0); + int rm_ofs = neon_reg_offset(a->vm, 0); + + if (!arm_dc_feature(s, ARM_FEATURE_NEON)) { + return false; + } + + /* UNDEF accesses to D16-D31 if they don't exist. */ + if (!dc_isar_feature(aa32_simd_r32, s) && + ((a->vd | a->vn | a->vm) & 0x10)) { + return false; + } + + if ((a->vn | a->vm | a->vd) & a->q) { + return false; + } + + if (!vfp_access_check(s)) { + return true; + } + + fn(a->size, rd_ofs, rn_ofs, rm_ofs, vec_size, vec_size); + return true; +} + +#define DO_3SAME(INSN, FUNC) \ + static bool trans_##INSN##_3s(DisasContext *s, arg_3same *a) \ + { \ + return do_3same(s, a, FUNC); \ + } + +DO_3SAME(VADD, tcg_gen_gvec_add) +DO_3SAME(VSUB, tcg_gen_gvec_sub) 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 int disas_neon_data_insn(DisasContext *s, uint32_t insn) } return 0; - case NEON_3R_VADD_VSUB: - if (u) { - tcg_gen_gvec_sub(size, rd_ofs, rn_ofs, rm_ofs, - vec_size, vec_size); - } else { - tcg_gen_gvec_add(size, rd_ofs, rn_ofs, rm_ofs, - vec_size, vec_size); - } - return 0; - case NEON_3R_VQADD: tcg_gen_gvec_4(rd_ofs, offsetof(CPUARMState, vfp.qc), rn_ofs, rm_ofs, vec_size, vec_size, @@ -XXX,XX +XXX,XX @@ static int disas_neon_data_insn(DisasContext *s, uint32_t insn) tcg_gen_gvec_3(rd_ofs, rm_ofs, rn_ofs, vec_size, vec_size, u ? &ushl_op[size] : &sshl_op[size]); return 0; + + case NEON_3R_VADD_VSUB: + /* Already handled by decodetree */ + return 1; } if (size == 3) { -- 2.20.1 Convert the Neon logic ops in the 3-reg-same grouping to decodetree. Note that for the logic ops the 'size' field forms part of their decode and the actual operations are always bitwise. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200430181003.21682-16-peter.maydell@linaro.org --- target/arm/neon-dp.decode | 12 +++++++++++ target/arm/translate-neon.inc.c | 19 +++++++++++++++++ target/arm/translate.c | 38 +-------------------------------- 3 files changed, 32 insertions(+), 37 deletions(-) diff --git a/target/arm/neon-dp.decode b/target/arm/neon-dp.decode index XXXXXXX..XXXXXXX 100644 --- a/target/arm/neon-dp.decode +++ b/target/arm/neon-dp.decode @@ -XXX,XX +XXX,XX @@ @3same .... ... . . . size:2 .... .... .... . q:1 . . .... \ &3same vm=%vm_dp vn=%vn_dp vd=%vd_dp +@3same_logic .... ... . . . .. .... .... .... . q:1 .. .... \ + &3same vm=%vm_dp vn=%vn_dp vd=%vd_dp size=0 + +VAND_3s 1111 001 0 0 . 00 .... .... 0001 ... 1 .... @3same_logic +VBIC_3s 1111 001 0 0 . 01 .... .... 0001 ... 1 .... @3same_logic +VORR_3s 1111 001 0 0 . 10 .... .... 0001 ... 1 .... @3same_logic +VORN_3s 1111 001 0 0 . 11 .... .... 0001 ... 1 .... @3same_logic +VEOR_3s 1111 001 1 0 . 00 .... .... 0001 ... 1 .... @3same_logic +VBSL_3s 1111 001 1 0 . 01 .... .... 0001 ... 1 .... @3same_logic +VBIT_3s 1111 001 1 0 . 10 .... .... 0001 ... 1 .... @3same_logic +VBIF_3s 1111 001 1 0 . 11 .... .... 0001 ... 1 .... @3same_logic + VADD_3s 1111 001 0 0 . .. .... .... 1000 . . . 0 .... @3same VSUB_3s 1111 001 1 0 . .. .... .... 1000 . . . 0 .... @3same diff --git a/target/arm/translate-neon.inc.c b/target/arm/translate-neon.inc.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-neon.inc.c +++ b/target/arm/translate-neon.inc.c @@ -XXX,XX +XXX,XX @@ static bool do_3same(DisasContext *s, arg_3same *a, GVecGen3Fn fn) DO_3SAME(VADD, tcg_gen_gvec_add) DO_3SAME(VSUB, tcg_gen_gvec_sub) +DO_3SAME(VAND, tcg_gen_gvec_and) +DO_3SAME(VBIC, tcg_gen_gvec_andc) +DO_3SAME(VORR, tcg_gen_gvec_or) +DO_3SAME(VORN, tcg_gen_gvec_orc) +DO_3SAME(VEOR, tcg_gen_gvec_xor) + +/* These insns are all gvec_bitsel but with the inputs in various orders. */ +#define DO_3SAME_BITSEL(INSN, O1, O2, O3) \ + static void gen_##INSN##_3s(unsigned vece, uint32_t rd_ofs, \ + uint32_t rn_ofs, uint32_t rm_ofs, \ + uint32_t oprsz, uint32_t maxsz) \ + { \ + tcg_gen_gvec_bitsel(vece, rd_ofs, O1, O2, O3, oprsz, maxsz); \ + } \ + DO_3SAME(INSN, gen_##INSN##_3s) + +DO_3SAME_BITSEL(VBSL, rd_ofs, rn_ofs, rm_ofs) +DO_3SAME_BITSEL(VBIT, rm_ofs, rn_ofs, rd_ofs) +DO_3SAME_BITSEL(VBIF, rm_ofs, rd_ofs, rn_ofs) 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 int disas_neon_data_insn(DisasContext *s, uint32_t insn) } return 1; - case NEON_3R_LOGIC: /* Logic ops. */ - switch ((u << 2) | size) { - case 0: /* VAND */ - tcg_gen_gvec_and(0, rd_ofs, rn_ofs, rm_ofs, - vec_size, vec_size); - break; - case 1: /* VBIC */ - tcg_gen_gvec_andc(0, rd_ofs, rn_ofs, rm_ofs, - vec_size, vec_size); - break; - case 2: /* VORR */ - tcg_gen_gvec_or(0, rd_ofs, rn_ofs, rm_ofs, - vec_size, vec_size); - break; - case 3: /* VORN */ - tcg_gen_gvec_orc(0, rd_ofs, rn_ofs, rm_ofs, - vec_size, vec_size); - break; - case 4: /* VEOR */ - tcg_gen_gvec_xor(0, rd_ofs, rn_ofs, rm_ofs, - vec_size, vec_size); - break; - case 5: /* VBSL */ - tcg_gen_gvec_bitsel(MO_8, rd_ofs, rd_ofs, rn_ofs, rm_ofs, - vec_size, vec_size); - break; - case 6: /* VBIT */ - tcg_gen_gvec_bitsel(MO_8, rd_ofs, rm_ofs, rn_ofs, rd_ofs, - vec_size, vec_size); - break; - case 7: /* VBIF */ - tcg_gen_gvec_bitsel(MO_8, rd_ofs, rm_ofs, rd_ofs, rn_ofs, - vec_size, vec_size); - break; - } - return 0; - case NEON_3R_VQADD: tcg_gen_gvec_4(rd_ofs, offsetof(CPUARMState, vfp.qc), rn_ofs, rm_ofs, vec_size, vec_size, @@ -XXX,XX +XXX,XX @@ static int disas_neon_data_insn(DisasContext *s, uint32_t insn) return 0; case NEON_3R_VADD_VSUB: + case NEON_3R_LOGIC: /* Already handled by decodetree */ return 1; } -- 2.20.1 Convert the Neon 3-reg-same VMAX and VMIN insns to decodetree. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200430181003.21682-17-peter.maydell@linaro.org --- target/arm/neon-dp.decode | 5 +++++ target/arm/translate-neon.inc.c | 14 ++++++++++++++ target/arm/translate.c | 21 ++------------------- 3 files changed, 21 insertions(+), 19 deletions(-) diff --git a/target/arm/neon-dp.decode b/target/arm/neon-dp.decode index XXXXXXX..XXXXXXX 100644 --- a/target/arm/neon-dp.decode +++ b/target/arm/neon-dp.decode @@ -XXX,XX +XXX,XX @@ VBSL_3s 1111 001 1 0 . 01 .... .... 0001 ... 1 .... @3same_logic VBIT_3s 1111 001 1 0 . 10 .... .... 0001 ... 1 .... @3same_logic VBIF_3s 1111 001 1 0 . 11 .... .... 0001 ... 1 .... @3same_logic +VMAX_S_3s 1111 001 0 0 . .. .... .... 0110 . . . 0 .... @3same +VMAX_U_3s 1111 001 1 0 . .. .... .... 0110 . . . 0 .... @3same +VMIN_S_3s 1111 001 0 0 . .. .... .... 0110 . . . 1 .... @3same +VMIN_U_3s 1111 001 1 0 . .. .... .... 0110 . . . 1 .... @3same + VADD_3s 1111 001 0 0 . .. .... .... 1000 . . . 0 .... @3same VSUB_3s 1111 001 1 0 . .. .... .... 1000 . . . 0 .... @3same diff --git a/target/arm/translate-neon.inc.c b/target/arm/translate-neon.inc.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-neon.inc.c +++ b/target/arm/translate-neon.inc.c @@ -XXX,XX +XXX,XX @@ DO_3SAME(VEOR, tcg_gen_gvec_xor) DO_3SAME_BITSEL(VBSL, rd_ofs, rn_ofs, rm_ofs) DO_3SAME_BITSEL(VBIT, rm_ofs, rn_ofs, rd_ofs) DO_3SAME_BITSEL(VBIF, rm_ofs, rd_ofs, rn_ofs) + +#define DO_3SAME_NO_SZ_3(INSN, FUNC) \ + static bool trans_##INSN##_3s(DisasContext *s, arg_3same *a) \ + { \ + if (a->size == 3) { \ + return false; \ + } \ + return do_3same(s, a, FUNC); \ + } + +DO_3SAME_NO_SZ_3(VMAX_S, tcg_gen_gvec_smax) +DO_3SAME_NO_SZ_3(VMAX_U, tcg_gen_gvec_umax) +DO_3SAME_NO_SZ_3(VMIN_S, tcg_gen_gvec_smin) +DO_3SAME_NO_SZ_3(VMIN_U, tcg_gen_gvec_umin) 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 int disas_neon_data_insn(DisasContext *s, uint32_t insn) rd_ofs, rn_ofs, rm_ofs, vec_size, vec_size); return 0; - case NEON_3R_VMAX: - if (u) { - tcg_gen_gvec_umax(size, rd_ofs, rn_ofs, rm_ofs, - vec_size, vec_size); - } else { - tcg_gen_gvec_smax(size, rd_ofs, rn_ofs, rm_ofs, - vec_size, vec_size); - } - return 0; - case NEON_3R_VMIN: - if (u) { - tcg_gen_gvec_umin(size, rd_ofs, rn_ofs, rm_ofs, - vec_size, vec_size); - } else { - tcg_gen_gvec_smin(size, rd_ofs, rn_ofs, rm_ofs, - vec_size, vec_size); - } - return 0; - case NEON_3R_VSHL: /* Note the operation is vshl vd,vm,vn */ tcg_gen_gvec_3(rd_ofs, rm_ofs, rn_ofs, vec_size, vec_size, @@ -XXX,XX +XXX,XX @@ static int disas_neon_data_insn(DisasContext *s, uint32_t insn) case NEON_3R_VADD_VSUB: case NEON_3R_LOGIC: + case NEON_3R_VMAX: + case NEON_3R_VMIN: /* Already handled by decodetree */ return 1; } -- 2.20.1 Convert the Neon comparison ops in the 3-reg-same grouping to decodetree. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200430181003.21682-18-peter.maydell@linaro.org --- target/arm/neon-dp.decode | 8 ++++++++ target/arm/translate-neon.inc.c | 22 ++++++++++++++++++++++ target/arm/translate.c | 23 +++-------------------- 3 files changed, 33 insertions(+), 20 deletions(-) diff --git a/target/arm/neon-dp.decode b/target/arm/neon-dp.decode index XXXXXXX..XXXXXXX 100644 --- a/target/arm/neon-dp.decode +++ b/target/arm/neon-dp.decode @@ -XXX,XX +XXX,XX @@ VBSL_3s 1111 001 1 0 . 01 .... .... 0001 ... 1 .... @3same_logic VBIT_3s 1111 001 1 0 . 10 .... .... 0001 ... 1 .... @3same_logic VBIF_3s 1111 001 1 0 . 11 .... .... 0001 ... 1 .... @3same_logic +VCGT_S_3s 1111 001 0 0 . .. .... .... 0011 . . . 0 .... @3same +VCGT_U_3s 1111 001 1 0 . .. .... .... 0011 . . . 0 .... @3same +VCGE_S_3s 1111 001 0 0 . .. .... .... 0011 . . . 1 .... @3same +VCGE_U_3s 1111 001 1 0 . .. .... .... 0011 . . . 1 .... @3same + VMAX_S_3s 1111 001 0 0 . .. .... .... 0110 . . . 0 .... @3same VMAX_U_3s 1111 001 1 0 . .. .... .... 0110 . . . 0 .... @3same VMIN_S_3s 1111 001 0 0 . .. .... .... 0110 . . . 1 .... @3same @@ -XXX,XX +XXX,XX @@ VMIN_U_3s 1111 001 1 0 . .. .... .... 0110 . . . 1 .... @3same VADD_3s 1111 001 0 0 . .. .... .... 1000 . . . 0 .... @3same VSUB_3s 1111 001 1 0 . .. .... .... 1000 . . . 0 .... @3same + +VTST_3s 1111 001 0 0 . .. .... .... 1000 . . . 1 .... @3same +VCEQ_3s 1111 001 1 0 . .. .... .... 1000 . . . 1 .... @3same diff --git a/target/arm/translate-neon.inc.c b/target/arm/translate-neon.inc.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-neon.inc.c +++ b/target/arm/translate-neon.inc.c @@ -XXX,XX +XXX,XX @@ DO_3SAME_NO_SZ_3(VMAX_S, tcg_gen_gvec_smax) DO_3SAME_NO_SZ_3(VMAX_U, tcg_gen_gvec_umax) DO_3SAME_NO_SZ_3(VMIN_S, tcg_gen_gvec_smin) DO_3SAME_NO_SZ_3(VMIN_U, tcg_gen_gvec_umin) + +#define DO_3SAME_CMP(INSN, COND) \ + static void gen_##INSN##_3s(unsigned vece, uint32_t rd_ofs, \ + uint32_t rn_ofs, uint32_t rm_ofs, \ + uint32_t oprsz, uint32_t maxsz) \ + { \ + tcg_gen_gvec_cmp(COND, vece, rd_ofs, rn_ofs, rm_ofs, oprsz, maxsz); \ + } \ + DO_3SAME_NO_SZ_3(INSN, gen_##INSN##_3s) + +DO_3SAME_CMP(VCGT_S, TCG_COND_GT) +DO_3SAME_CMP(VCGT_U, TCG_COND_GTU) +DO_3SAME_CMP(VCGE_S, TCG_COND_GE) +DO_3SAME_CMP(VCGE_U, TCG_COND_GEU) +DO_3SAME_CMP(VCEQ, TCG_COND_EQ) + +static void gen_VTST_3s(unsigned vece, uint32_t rd_ofs, uint32_t rn_ofs, + uint32_t rm_ofs, uint32_t oprsz, uint32_t maxsz) +{ + tcg_gen_gvec_3(rd_ofs, rn_ofs, rm_ofs, oprsz, maxsz, &cmtst_op[vece]); +} +DO_3SAME_NO_SZ_3(VTST, gen_VTST_3s) 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 int disas_neon_data_insn(DisasContext *s, uint32_t insn) u ? &mls_op[size] : &mla_op[size]); return 0; - case NEON_3R_VTST_VCEQ: - if (u) { /* VCEQ */ - tcg_gen_gvec_cmp(TCG_COND_EQ, size, rd_ofs, rn_ofs, rm_ofs, - vec_size, vec_size); - } else { /* VTST */ - tcg_gen_gvec_3(rd_ofs, rn_ofs, rm_ofs, - vec_size, vec_size, &cmtst_op[size]); - } - return 0; - - case NEON_3R_VCGT: - tcg_gen_gvec_cmp(u ? TCG_COND_GTU : TCG_COND_GT, size, - rd_ofs, rn_ofs, rm_ofs, vec_size, vec_size); - return 0; - - case NEON_3R_VCGE: - tcg_gen_gvec_cmp(u ? TCG_COND_GEU : TCG_COND_GE, size, - rd_ofs, rn_ofs, rm_ofs, vec_size, vec_size); - return 0; - case NEON_3R_VSHL: /* Note the operation is vshl vd,vm,vn */ tcg_gen_gvec_3(rd_ofs, rm_ofs, rn_ofs, vec_size, vec_size, @@ -XXX,XX +XXX,XX @@ static int disas_neon_data_insn(DisasContext *s, uint32_t insn) case NEON_3R_LOGIC: case NEON_3R_VMAX: case NEON_3R_VMIN: + case NEON_3R_VTST_VCEQ: + case NEON_3R_VCGT: + case NEON_3R_VCGE: /* Already handled by decodetree */ return 1; } -- 2.20.1 Convert the Neon VQADD/VQSUB insns in the 3-reg-same grouping to decodetree. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200430181003.21682-19-peter.maydell@linaro.org --- target/arm/neon-dp.decode | 6 ++++++ target/arm/translate-neon.inc.c | 15 +++++++++++++++ target/arm/translate.c | 14 ++------------ 3 files changed, 23 insertions(+), 12 deletions(-) diff --git a/target/arm/neon-dp.decode b/target/arm/neon-dp.decode index XXXXXXX..XXXXXXX 100644 --- a/target/arm/neon-dp.decode +++ b/target/arm/neon-dp.decode @@ -XXX,XX +XXX,XX @@ @3same .... ... . . . size:2 .... .... .... . q:1 . . .... \ &3same vm=%vm_dp vn=%vn_dp vd=%vd_dp +VQADD_S_3s 1111 001 0 0 . .. .... .... 0000 . . . 1 .... @3same +VQADD_U_3s 1111 001 1 0 . .. .... .... 0000 . . . 1 .... @3same + @3same_logic .... ... . . . .. .... .... .... . q:1 .. .... \ &3same vm=%vm_dp vn=%vn_dp vd=%vd_dp size=0 @@ -XXX,XX +XXX,XX @@ VBSL_3s 1111 001 1 0 . 01 .... .... 0001 ... 1 .... @3same_logic VBIT_3s 1111 001 1 0 . 10 .... .... 0001 ... 1 .... @3same_logic VBIF_3s 1111 001 1 0 . 11 .... .... 0001 ... 1 .... @3same_logic +VQSUB_S_3s 1111 001 0 0 . .. .... .... 0010 . . . 1 .... @3same +VQSUB_U_3s 1111 001 1 0 . .. .... .... 0010 . . . 1 .... @3same + VCGT_S_3s 1111 001 0 0 . .. .... .... 0011 . . . 0 .... @3same VCGT_U_3s 1111 001 1 0 . .. .... .... 0011 . . . 0 .... @3same VCGE_S_3s 1111 001 0 0 . .. .... .... 0011 . . . 1 .... @3same diff --git a/target/arm/translate-neon.inc.c b/target/arm/translate-neon.inc.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-neon.inc.c +++ b/target/arm/translate-neon.inc.c @@ -XXX,XX +XXX,XX @@ static void gen_VTST_3s(unsigned vece, uint32_t rd_ofs, uint32_t rn_ofs, tcg_gen_gvec_3(rd_ofs, rn_ofs, rm_ofs, oprsz, maxsz, &cmtst_op[vece]); } DO_3SAME_NO_SZ_3(VTST, gen_VTST_3s) + +#define DO_3SAME_GVEC4(INSN, OPARRAY) \ + static void gen_##INSN##_3s(unsigned vece, uint32_t rd_ofs, \ + uint32_t rn_ofs, uint32_t rm_ofs, \ + uint32_t oprsz, uint32_t maxsz) \ + { \ + tcg_gen_gvec_4(rd_ofs, offsetof(CPUARMState, vfp.qc), \ + rn_ofs, rm_ofs, oprsz, maxsz, &OPARRAY[vece]); \ + } \ + DO_3SAME(INSN, gen_##INSN##_3s) + +DO_3SAME_GVEC4(VQADD_S, sqadd_op) +DO_3SAME_GVEC4(VQADD_U, uqadd_op) +DO_3SAME_GVEC4(VQSUB_S, sqsub_op) +DO_3SAME_GVEC4(VQSUB_U, uqsub_op) 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 int disas_neon_data_insn(DisasContext *s, uint32_t insn) } return 1; - case NEON_3R_VQADD: - tcg_gen_gvec_4(rd_ofs, offsetof(CPUARMState, vfp.qc), - rn_ofs, rm_ofs, vec_size, vec_size, - (u ? uqadd_op : sqadd_op) + size); - return 0; - - case NEON_3R_VQSUB: - tcg_gen_gvec_4(rd_ofs, offsetof(CPUARMState, vfp.qc), - rn_ofs, rm_ofs, vec_size, vec_size, - (u ? uqsub_op : sqsub_op) + size); - return 0; - case NEON_3R_VMUL: /* VMUL */ if (u) { /* Polynomial case allows only P8. */ @@ -XXX,XX +XXX,XX @@ static int disas_neon_data_insn(DisasContext *s, uint32_t insn) case NEON_3R_VTST_VCEQ: case NEON_3R_VCGT: case NEON_3R_VCGE: + case NEON_3R_VQADD: + case NEON_3R_VQSUB: /* Already handled by decodetree */ return 1; } -- 2.20.1 Convert the Neon VMUL, VMLA, VMLS and VSHL insns in the 3-reg-same grouping to decodetree. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200430181003.21682-20-peter.maydell@linaro.org --- target/arm/neon-dp.decode | 9 +++++++ target/arm/translate-neon.inc.c | 44 +++++++++++++++++++++++++++++++++ target/arm/translate.c | 28 +++------------------ 3 files changed, 56 insertions(+), 25 deletions(-) diff --git a/target/arm/neon-dp.decode b/target/arm/neon-dp.decode index XXXXXXX..XXXXXXX 100644 --- a/target/arm/neon-dp.decode +++ b/target/arm/neon-dp.decode @@ -XXX,XX +XXX,XX @@ VCGT_U_3s 1111 001 1 0 . .. .... .... 0011 . . . 0 .... @3same VCGE_S_3s 1111 001 0 0 . .. .... .... 0011 . . . 1 .... @3same VCGE_U_3s 1111 001 1 0 . .. .... .... 0011 . . . 1 .... @3same +VSHL_S_3s 1111 001 0 0 . .. .... .... 0100 . . . 0 .... @3same +VSHL_U_3s 1111 001 1 0 . .. .... .... 0100 . . . 0 .... @3same + VMAX_S_3s 1111 001 0 0 . .. .... .... 0110 . . . 0 .... @3same VMAX_U_3s 1111 001 1 0 . .. .... .... 0110 . . . 0 .... @3same VMIN_S_3s 1111 001 0 0 . .. .... .... 0110 . . . 1 .... @3same @@ -XXX,XX +XXX,XX @@ VSUB_3s 1111 001 1 0 . .. .... .... 1000 . . . 0 .... @3same VTST_3s 1111 001 0 0 . .. .... .... 1000 . . . 1 .... @3same VCEQ_3s 1111 001 1 0 . .. .... .... 1000 . . . 1 .... @3same + +VMLA_3s 1111 001 0 0 . .. .... .... 1001 . . . 0 .... @3same +VMLS_3s 1111 001 1 0 . .. .... .... 1001 . . . 0 .... @3same + +VMUL_3s 1111 001 0 0 . .. .... .... 1001 . . . 1 .... @3same +VMUL_p_3s 1111 001 1 0 . .. .... .... 1001 . . . 1 .... @3same diff --git a/target/arm/translate-neon.inc.c b/target/arm/translate-neon.inc.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-neon.inc.c +++ b/target/arm/translate-neon.inc.c @@ -XXX,XX +XXX,XX @@ DO_3SAME_NO_SZ_3(VMAX_S, tcg_gen_gvec_smax) DO_3SAME_NO_SZ_3(VMAX_U, tcg_gen_gvec_umax) DO_3SAME_NO_SZ_3(VMIN_S, tcg_gen_gvec_smin) DO_3SAME_NO_SZ_3(VMIN_U, tcg_gen_gvec_umin) +DO_3SAME_NO_SZ_3(VMUL, tcg_gen_gvec_mul) #define DO_3SAME_CMP(INSN, COND) \ static void gen_##INSN##_3s(unsigned vece, uint32_t rd_ofs, \ @@ -XXX,XX +XXX,XX @@ DO_3SAME_GVEC4(VQADD_S, sqadd_op) DO_3SAME_GVEC4(VQADD_U, uqadd_op) DO_3SAME_GVEC4(VQSUB_S, sqsub_op) DO_3SAME_GVEC4(VQSUB_U, uqsub_op) + +static void gen_VMUL_p_3s(unsigned vece, uint32_t rd_ofs, uint32_t rn_ofs, + uint32_t rm_ofs, uint32_t oprsz, uint32_t maxsz) +{ + tcg_gen_gvec_3_ool(rd_ofs, rn_ofs, rm_ofs, oprsz, maxsz, + 0, gen_helper_gvec_pmul_b); +} + +static bool trans_VMUL_p_3s(DisasContext *s, arg_3same *a) +{ + if (a->size != 0) { + return false; + } + return do_3same(s, a, gen_VMUL_p_3s); +} + +#define DO_3SAME_GVEC3_NO_SZ_3(INSN, OPARRAY) \ + static void gen_##INSN##_3s(unsigned vece, uint32_t rd_ofs, \ + uint32_t rn_ofs, uint32_t rm_ofs, \ + uint32_t oprsz, uint32_t maxsz) \ + { \ + tcg_gen_gvec_3(rd_ofs, rn_ofs, rm_ofs, \ + oprsz, maxsz, &OPARRAY[vece]); \ + } \ + DO_3SAME_NO_SZ_3(INSN, gen_##INSN##_3s) + + +DO_3SAME_GVEC3_NO_SZ_3(VMLA, mla_op) +DO_3SAME_GVEC3_NO_SZ_3(VMLS, mls_op) + +#define DO_3SAME_GVEC3_SHIFT(INSN, OPARRAY) \ + static void gen_##INSN##_3s(unsigned vece, uint32_t rd_ofs, \ + uint32_t rn_ofs, uint32_t rm_ofs, \ + uint32_t oprsz, uint32_t maxsz) \ + { \ + /* Note the operation is vshl vd,vm,vn */ \ + tcg_gen_gvec_3(rd_ofs, rm_ofs, rn_ofs, \ + oprsz, maxsz, &OPARRAY[vece]); \ + } \ + DO_3SAME(INSN, gen_##INSN##_3s) + +DO_3SAME_GVEC3_SHIFT(VSHL_S, sshl_op) +DO_3SAME_GVEC3_SHIFT(VSHL_U, ushl_op) 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 int disas_neon_data_insn(DisasContext *s, uint32_t insn) } return 1; - case NEON_3R_VMUL: /* VMUL */ - if (u) { - /* Polynomial case allows only P8. */ - if (size != 0) { - return 1; - } - tcg_gen_gvec_3_ool(rd_ofs, rn_ofs, rm_ofs, vec_size, vec_size, - 0, gen_helper_gvec_pmul_b); - } else { - tcg_gen_gvec_mul(size, rd_ofs, rn_ofs, rm_ofs, - vec_size, vec_size); - } - return 0; - - case NEON_3R_VML: /* VMLA, VMLS */ - tcg_gen_gvec_3(rd_ofs, rn_ofs, rm_ofs, vec_size, vec_size, - u ? &mls_op[size] : &mla_op[size]); - return 0; - - case NEON_3R_VSHL: - /* Note the operation is vshl vd,vm,vn */ - tcg_gen_gvec_3(rd_ofs, rm_ofs, rn_ofs, vec_size, vec_size, - u ? &ushl_op[size] : &sshl_op[size]); - return 0; - case NEON_3R_VADD_VSUB: case NEON_3R_LOGIC: case NEON_3R_VMAX: @@ -XXX,XX +XXX,XX @@ static int disas_neon_data_insn(DisasContext *s, uint32_t insn) case NEON_3R_VCGE: case NEON_3R_VQADD: case NEON_3R_VQSUB: + case NEON_3R_VMUL: + case NEON_3R_VML: + case NEON_3R_VSHL: /* Already handled by decodetree */ return 1; } -- 2.20.1 We're going to want at least some of the NeonGen* typedefs for the refactored 32-bit Neon decoder, so move them all to translate.h since it makes more sense to keep them in one group. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200430181003.21682-23-peter.maydell@linaro.org --- target/arm/translate.h | 17 +++++++++++++++++ target/arm/translate-a64.c | 17 ----------------- 2 files changed, 17 insertions(+), 17 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 @@ typedef void GVecGen3Fn(unsigned, uint32_t, uint32_t, typedef void GVecGen4Fn(unsigned, uint32_t, uint32_t, uint32_t, uint32_t, uint32_t, uint32_t); +/* Function prototype for gen_ functions for calling Neon helpers */ +typedef void NeonGenOneOpEnvFn(TCGv_i32, TCGv_ptr, TCGv_i32); +typedef void NeonGenTwoOpFn(TCGv_i32, TCGv_i32, TCGv_i32); +typedef void NeonGenTwoOpEnvFn(TCGv_i32, TCGv_ptr, TCGv_i32, TCGv_i32); +typedef void NeonGenTwo64OpFn(TCGv_i64, TCGv_i64, TCGv_i64); +typedef void NeonGenTwo64OpEnvFn(TCGv_i64, TCGv_ptr, TCGv_i64, TCGv_i64); +typedef void NeonGenNarrowFn(TCGv_i32, TCGv_i64); +typedef void NeonGenNarrowEnvFn(TCGv_i32, TCGv_ptr, TCGv_i64); +typedef void NeonGenWidenFn(TCGv_i64, TCGv_i32); +typedef void NeonGenTwoSingleOPFn(TCGv_i32, TCGv_i32, TCGv_i32, TCGv_ptr); +typedef void NeonGenTwoDoubleOPFn(TCGv_i64, TCGv_i64, TCGv_i64, TCGv_ptr); +typedef void NeonGenOneOpFn(TCGv_i64, TCGv_i64); +typedef void CryptoTwoOpFn(TCGv_ptr, TCGv_ptr); +typedef void CryptoThreeOpIntFn(TCGv_ptr, TCGv_ptr, TCGv_i32); +typedef void CryptoThreeOpFn(TCGv_ptr, TCGv_ptr, TCGv_ptr); +typedef void AtomicThreeOpFn(TCGv_i64, TCGv_i64, TCGv_i64, TCGArg, MemOp); + #endif /* TARGET_ARM_TRANSLATE_H */ 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 @@ typedef struct AArch64DecodeTable { AArch64DecodeFn *disas_fn; } AArch64DecodeTable; -/* Function prototype for gen_ functions for calling Neon helpers */ -typedef void NeonGenOneOpEnvFn(TCGv_i32, TCGv_ptr, TCGv_i32); -typedef void NeonGenTwoOpFn(TCGv_i32, TCGv_i32, TCGv_i32); -typedef void NeonGenTwoOpEnvFn(TCGv_i32, TCGv_ptr, TCGv_i32, TCGv_i32); -typedef void NeonGenTwo64OpFn(TCGv_i64, TCGv_i64, TCGv_i64); -typedef void NeonGenTwo64OpEnvFn(TCGv_i64, TCGv_ptr, TCGv_i64, TCGv_i64); -typedef void NeonGenNarrowFn(TCGv_i32, TCGv_i64); -typedef void NeonGenNarrowEnvFn(TCGv_i32, TCGv_ptr, TCGv_i64); -typedef void NeonGenWidenFn(TCGv_i64, TCGv_i32); -typedef void NeonGenTwoSingleOPFn(TCGv_i32, TCGv_i32, TCGv_i32, TCGv_ptr); -typedef void NeonGenTwoDoubleOPFn(TCGv_i64, TCGv_i64, TCGv_i64, TCGv_ptr); -typedef void NeonGenOneOpFn(TCGv_i64, TCGv_i64); -typedef void CryptoTwoOpFn(TCGv_ptr, TCGv_ptr); -typedef void CryptoThreeOpIntFn(TCGv_ptr, TCGv_ptr, TCGv_i32); -typedef void CryptoThreeOpFn(TCGv_ptr, TCGv_ptr, TCGv_ptr); -typedef void AtomicThreeOpFn(TCGv_i64, TCGv_i64, TCGv_i64, TCGArg, MemOp); - /* initialize TCG globals. */ void a64_translate_init(void) { -- 2.20.1