The following changes since commit 3d48b6b687c558a042d91370633b91c6e29e0e05: Merge tag 'pull-request-2024-05-14' of https://gitlab.com/thuth/qemu into staging (2024-05-14 17:24:04 +0200) are available in the Git repository at: https://gitlab.com/rth7680/qemu.git tags/pull-tcg-20240515 for you to fetch changes up to c9290dfebfdba5c13baa5e1f10e13a1c876b0643: tcg/loongarch64: Fill out tcg_out_{ld,st} for vector regs (2024-05-15 08:57:39 +0200) ---------------------------------------------------------------- tcg/loongarch64: Fill out tcg_out_{ld,st} for vector regs accel/tcg: Improve disassembly for target and plugin ---------------------------------------------------------------- Philippe Mathieu-Daudé (1): accel/tcg: Remove cpu_ldsb_code / cpu_ldsw_code Richard Henderson (33): accel/tcg: Use vaddr in translator_ld* accel/tcg: Hide in_same_page outside of a target-specific context accel/tcg: Pass DisasContextBase to translator_fake_ldb accel/tcg: Reorg translator_ld* accel/tcg: Cap the translation block when we encounter mmio accel/tcg: Record mmio bytes during translation accel/tcg: Record when translator_fake_ldb is used accel/tcg: Record DisasContextBase in tcg_ctx for plugins plugins: Copy memory in qemu_plugin_insn_data accel/tcg: Implement translator_st plugins: Use translator_st for qemu_plugin_insn_data plugins: Read mem_only directly from TB cflags plugins: Use DisasContextBase for qemu_plugin_insn_haddr plugins: Use DisasContextBase for qemu_plugin_tb_vaddr plugins: Merge alloc_tcg_plugin_context into plugin_gen_tb_start accel/tcg: Provide default implementation of disas_log accel/tcg: Return bool from TranslatorOps.disas_log disas: Split disas.c disas: Use translator_st to get disassembly data accel/tcg: Introduce translator_fake_ld target/s390x: Fix translator_fake_ld length target/s390x: Disassemble EXECUTEd instructions target/hexagon: Use translator_ldl in pkt_crosses_page target/microblaze: Use translator_ldl target/i386: Use translator_ldub for everything target/avr: Use translator_lduw target/cris: Use translator_ld* in cris_fetch target/cris: Use cris_fetch in translate_v10.c.inc target/riscv: Use translator_ld* for everything target/rx: Use translator_ld* target/xtensa: Use translator_ldub in xtensa_insn_len target/s390x: Use translator_lduw in get_next_pc tcg/loongarch64: Fill out tcg_out_{ld,st} for vector regs disas/disas-internal.h | 4 + include/disas/disas.h | 9 +- include/exec/cpu_ldst.h | 10 -- include/exec/plugin-gen.h | 7 +- include/exec/translator.h | 74 ++++++--- include/qemu/plugin.h | 22 +-- include/qemu/qemu-plugin.h | 15 +- include/qemu/typedefs.h | 1 + include/tcg/tcg.h | 1 + accel/tcg/plugin-gen.c | 63 +++----- accel/tcg/translator.c | 331 ++++++++++++++++++++++++-------------- contrib/plugins/execlog.c | 5 +- contrib/plugins/howvec.c | 4 +- disas/disas-common.c | 104 ++++++++++++ disas/disas-host.c | 129 +++++++++++++++ disas/disas-mon.c | 15 ++ disas/disas-target.c | 99 ++++++++++++ disas/disas.c | 338 --------------------------------------- disas/objdump.c | 37 +++++ plugins/api.c | 57 +++++-- target/alpha/translate.c | 9 -- target/arm/tcg/translate-a64.c | 11 -- target/arm/tcg/translate.c | 12 -- target/avr/translate.c | 11 +- target/cris/translate.c | 37 +---- target/hexagon/translate.c | 11 +- target/hppa/translate.c | 21 ++- target/i386/tcg/translate.c | 19 +-- target/loongarch/tcg/translate.c | 8 - target/m68k/translate.c | 9 -- target/microblaze/translate.c | 11 +- target/mips/tcg/translate.c | 9 -- target/openrisc/translate.c | 11 -- target/ppc/translate.c | 9 -- target/riscv/translate.c | 24 +-- target/rx/translate.c | 35 ++-- target/s390x/tcg/translate.c | 26 ++- target/sh4/translate.c | 9 -- target/sparc/translate.c | 9 -- target/tricore/translate.c | 9 -- target/xtensa/translate.c | 12 +- tcg/tcg.c | 12 -- target/cris/translate_v10.c.inc | 30 ++-- tcg/loongarch64/tcg-target.c.inc | 103 +++++++++--- disas/meson.build | 8 +- 45 files changed, 899 insertions(+), 891 deletions(-) create mode 100644 disas/disas-common.c create mode 100644 disas/disas-host.c create mode 100644 disas/disas-target.c delete mode 100644 disas/disas.c create mode 100644 disas/objdump.c Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- include/exec/translator.h | 21 +++++++++------------ accel/tcg/translator.c | 15 ++++++++------- target/hexagon/translate.c | 1 + target/microblaze/translate.c | 1 + 4 files changed, 19 insertions(+), 19 deletions(-) diff --git a/include/exec/translator.h b/include/exec/translator.h index XXXXXXX..XXXXXXX 100644 --- a/include/exec/translator.h +++ b/include/exec/translator.h @@ -XXX,XX +XXX,XX @@ */ #include "qemu/bswap.h" -#include "exec/cpu-common.h" -#include "exec/cpu-defs.h" -#include "exec/abi_ptr.h" -#include "cpu.h" +#include "exec/vaddr.h" /** * gen_intermediate_code @@ -XXX,XX +XXX,XX @@ bool translator_io_start(DisasContextBase *db); * the relevant information at translation time. */ -uint8_t translator_ldub(CPUArchState *env, DisasContextBase *db, abi_ptr pc); -uint16_t translator_lduw(CPUArchState *env, DisasContextBase *db, abi_ptr pc); -uint32_t translator_ldl(CPUArchState *env, DisasContextBase *db, abi_ptr pc); -uint64_t translator_ldq(CPUArchState *env, DisasContextBase *db, abi_ptr pc); +uint8_t translator_ldub(CPUArchState *env, DisasContextBase *db, vaddr pc); +uint16_t translator_lduw(CPUArchState *env, DisasContextBase *db, vaddr pc); +uint32_t translator_ldl(CPUArchState *env, DisasContextBase *db, vaddr pc); +uint64_t translator_ldq(CPUArchState *env, DisasContextBase *db, vaddr pc); static inline uint16_t translator_lduw_swap(CPUArchState *env, DisasContextBase *db, - abi_ptr pc, bool do_swap) + vaddr pc, bool do_swap) { uint16_t ret = translator_lduw(env, db, pc); if (do_swap) { @@ -XXX,XX +XXX,XX @@ translator_lduw_swap(CPUArchState *env, DisasContextBase *db, static inline uint32_t translator_ldl_swap(CPUArchState *env, DisasContextBase *db, - abi_ptr pc, bool do_swap) + vaddr pc, bool do_swap) { uint32_t ret = translator_ldl(env, db, pc); if (do_swap) { @@ -XXX,XX +XXX,XX @@ translator_ldl_swap(CPUArchState *env, DisasContextBase *db, static inline uint64_t translator_ldq_swap(CPUArchState *env, DisasContextBase *db, - abi_ptr pc, bool do_swap) + vaddr pc, bool do_swap) { uint64_t ret = translator_ldq(env, db, pc); if (do_swap) { @@ -XXX,XX +XXX,XX @@ translator_ldq_swap(CPUArchState *env, DisasContextBase *db, * re-synthesised for s390x "ex"). It ensures we update other areas of * the translator with details of the executed instruction. */ -void translator_fake_ldb(uint8_t insn8, abi_ptr pc); +void translator_fake_ldb(uint8_t insn8, vaddr pc); /* * Return whether addr is on the same page as where disassembly started. diff --git a/accel/tcg/translator.c b/accel/tcg/translator.c index XXXXXXX..XXXXXXX 100644 --- a/accel/tcg/translator.c +++ b/accel/tcg/translator.c @@ -XXX,XX +XXX,XX @@ #include "exec/translator.h" #include "exec/cpu_ldst.h" #include "exec/plugin-gen.h" +#include "exec/cpu_ldst.h" #include "tcg/tcg-op-common.h" #include "internal-target.h" @@ -XXX,XX +XXX,XX @@ static void *translator_access(CPUArchState *env, DisasContextBase *db, return host + (pc - base); } -static void plugin_insn_append(abi_ptr pc, const void *from, size_t size) +static void plugin_insn_append(vaddr pc, const void *from, size_t size) { #ifdef CONFIG_PLUGIN struct qemu_plugin_insn *insn = tcg_ctx->plugin_insn; - abi_ptr off; + size_t off; if (insn == NULL) { return; @@ -XXX,XX +XXX,XX @@ static void plugin_insn_append(abi_ptr pc, const void *from, size_t size) #endif } -uint8_t translator_ldub(CPUArchState *env, DisasContextBase *db, abi_ptr pc) +uint8_t translator_ldub(CPUArchState *env, DisasContextBase *db, vaddr pc) { uint8_t ret; void *p = translator_access(env, db, pc, sizeof(ret)); @@ -XXX,XX +XXX,XX @@ uint8_t translator_ldub(CPUArchState *env, DisasContextBase *db, abi_ptr pc) return ret; } -uint16_t translator_lduw(CPUArchState *env, DisasContextBase *db, abi_ptr pc) +uint16_t translator_lduw(CPUArchState *env, DisasContextBase *db, vaddr pc) { uint16_t ret, plug; void *p = translator_access(env, db, pc, sizeof(ret)); @@ -XXX,XX +XXX,XX @@ uint16_t translator_lduw(CPUArchState *env, DisasContextBase *db, abi_ptr pc) return ret; } -uint32_t translator_ldl(CPUArchState *env, DisasContextBase *db, abi_ptr pc) +uint32_t translator_ldl(CPUArchState *env, DisasContextBase *db, vaddr pc) { uint32_t ret, plug; void *p = translator_access(env, db, pc, sizeof(ret)); @@ -XXX,XX +XXX,XX @@ uint32_t translator_ldl(CPUArchState *env, DisasContextBase *db, abi_ptr pc) return ret; } -uint64_t translator_ldq(CPUArchState *env, DisasContextBase *db, abi_ptr pc) +uint64_t translator_ldq(CPUArchState *env, DisasContextBase *db, vaddr pc) { uint64_t ret, plug; void *p = translator_access(env, db, pc, sizeof(ret)); @@ -XXX,XX +XXX,XX @@ uint64_t translator_ldq(CPUArchState *env, DisasContextBase *db, abi_ptr pc) return ret; } -void translator_fake_ldb(uint8_t insn8, abi_ptr pc) +void translator_fake_ldb(uint8_t insn8, vaddr pc) { plugin_insn_append(pc, &insn8, sizeof(insn8)); } diff --git a/target/hexagon/translate.c b/target/hexagon/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/hexagon/translate.c +++ b/target/hexagon/translate.c @@ -XXX,XX +XXX,XX @@ #include "exec/translation-block.h" #include "exec/cpu_ldst.h" #include "exec/log.h" +#include "exec/cpu_ldst.h" #include "internal.h" #include "attribs.h" #include "insn.h" diff --git a/target/microblaze/translate.c b/target/microblaze/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/microblaze/translate.c +++ b/target/microblaze/translate.c @@ -XXX,XX +XXX,XX @@ #include "tcg/tcg-op.h" #include "exec/helper-proto.h" #include "exec/helper-gen.h" +#include "exec/cpu_ldst.h" #include "exec/translator.h" #include "qemu/qemu-print.h" -- 2.34.1 While there are other methods that could be used to replace TARGET_PAGE_MASK, the function is not really required outside the context of target-specific translation. This makes the header usable by target independent code. Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- include/exec/translator.h | 2 ++ 1 file changed, 2 insertions(+) diff --git a/include/exec/translator.h b/include/exec/translator.h index XXXXXXX..XXXXXXX 100644 --- a/include/exec/translator.h +++ b/include/exec/translator.h @@ -XXX,XX +XXX,XX @@ translator_ldq_swap(CPUArchState *env, DisasContextBase *db, */ void translator_fake_ldb(uint8_t insn8, vaddr pc); +#ifdef COMPILING_PER_TARGET /* * Return whether addr is on the same page as where disassembly started. * Translators can use this to enforce the rule that only single-insn @@ -XXX,XX +XXX,XX @@ static inline bool is_same_page(const DisasContextBase *db, vaddr addr) { return ((addr ^ db->pc_first) & TARGET_PAGE_MASK) == 0; } +#endif #endif /* EXEC__TRANSLATOR_H */ -- 2.34.1 Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- include/exec/translator.h | 5 +++-- accel/tcg/translator.c | 2 +- target/s390x/tcg/translate.c | 2 +- 3 files changed, 5 insertions(+), 4 deletions(-) diff --git a/include/exec/translator.h b/include/exec/translator.h index XXXXXXX..XXXXXXX 100644 --- a/include/exec/translator.h +++ b/include/exec/translator.h @@ -XXX,XX +XXX,XX @@ translator_ldq_swap(CPUArchState *env, DisasContextBase *db, /** * translator_fake_ldb - fake instruction load - * @insn8: byte of instruction + * @db: Disassembly context * @pc: program counter of instruction + * @insn8: byte of instruction * * This is a special case helper used where the instruction we are * about to translate comes from somewhere else (e.g. being * re-synthesised for s390x "ex"). It ensures we update other areas of * the translator with details of the executed instruction. */ -void translator_fake_ldb(uint8_t insn8, vaddr pc); +void translator_fake_ldb(DisasContextBase *db, vaddr pc, uint8_t insn8); #ifdef COMPILING_PER_TARGET /* diff --git a/accel/tcg/translator.c b/accel/tcg/translator.c index XXXXXXX..XXXXXXX 100644 --- a/accel/tcg/translator.c +++ b/accel/tcg/translator.c @@ -XXX,XX +XXX,XX @@ uint64_t translator_ldq(CPUArchState *env, DisasContextBase *db, vaddr pc) return ret; } -void translator_fake_ldb(uint8_t insn8, vaddr pc) +void translator_fake_ldb(DisasContextBase *db, vaddr pc, uint8_t insn8) { plugin_insn_append(pc, &insn8, sizeof(insn8)); } diff --git a/target/s390x/tcg/translate.c b/target/s390x/tcg/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/s390x/tcg/translate.c +++ b/target/s390x/tcg/translate.c @@ -XXX,XX +XXX,XX @@ static const DisasInsn *extract_insn(CPUS390XState *env, DisasContext *s) /* Register insn bytes with translator so plugins work. */ for (int i = 0; i < ilen; i++) { uint8_t byte = extract64(insn, 56 - (i * 8), 8); - translator_fake_ldb(byte, pc + i); + translator_fake_ldb(&s->base, pc + i, byte); } op = insn >> 56; } else { -- 2.34.1 Reorg translator_access into translator_ld, with a more memcpy-ish interface. If both pages are in ram, do not go through the caller's slow path. Assert that the access is within the two pages that we are prepared to protect, per TranslationBlock. Allow access prior to pc_first, so long as it is within the first page. Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- accel/tcg/translator.c | 189 ++++++++++++++++++++++------------------- 1 file changed, 101 insertions(+), 88 deletions(-) diff --git a/accel/tcg/translator.c b/accel/tcg/translator.c index XXXXXXX..XXXXXXX 100644 --- a/accel/tcg/translator.c +++ b/accel/tcg/translator.c @@ -XXX,XX +XXX,XX @@ void translator_loop(CPUState *cpu, TranslationBlock *tb, int *max_insns, } } -static void *translator_access(CPUArchState *env, DisasContextBase *db, - vaddr pc, size_t len) +static bool translator_ld(CPUArchState *env, DisasContextBase *db, + void *dest, vaddr pc, size_t len) { + TranslationBlock *tb = db->tb; + vaddr last = pc + len - 1; void *host; - vaddr base, end; - TranslationBlock *tb; - - tb = db->tb; + vaddr base; /* Use slow path if first page is MMIO. */ if (unlikely(tb_page_addr0(tb) == -1)) { - return NULL; + return false; } - end = pc + len - 1; - if (likely(is_same_page(db, end))) { - host = db->host_addr[0]; - base = db->pc_first; - } else { + host = db->host_addr[0]; + base = db->pc_first; + + if (likely(((base ^ last) & TARGET_PAGE_MASK) == 0)) { + /* Entire read is from the first page. */ + memcpy(dest, host + (pc - base), len); + return true; + } + + if (unlikely(((base ^ pc) & TARGET_PAGE_MASK) == 0)) { + /* Read begins on the first page and extends to the second. */ + size_t len0 = -(pc | TARGET_PAGE_MASK); + memcpy(dest, host + (pc - base), len0); + pc += len0; + dest += len0; + len -= len0; + } + + /* + * The read must conclude on the second page and not extend to a third. + * + * TODO: We could allow the two pages to be virtually discontiguous, + * since we already allow the two pages to be physically discontiguous. + * The only reasonable use case would be executing an insn at the end + * of the address space wrapping around to the beginning. For that, + * we would need to know the current width of the address space. + * In the meantime, assert. + */ + base = (base & TARGET_PAGE_MASK) + TARGET_PAGE_SIZE; + assert(((base ^ pc) & TARGET_PAGE_MASK) == 0); + assert(((base ^ last) & TARGET_PAGE_MASK) == 0); + host = db->host_addr[1]; + + if (host == NULL) { + tb_page_addr_t page0, old_page1, new_page1; + + new_page1 = get_page_addr_code_hostp(env, base, &db->host_addr[1]); + + /* + * If the second page is MMIO, treat as if the first page + * was MMIO as well, so that we do not cache the TB. + */ + if (unlikely(new_page1 == -1)) { + tb_unlock_pages(tb); + tb_set_page_addr0(tb, -1); + return false; + } + + /* + * If this is not the first time around, and page1 matches, + * then we already have the page locked. Alternately, we're + * not doing anything to prevent the PTE from changing, so + * we might wind up with a different page, requiring us to + * re-do the locking. + */ + old_page1 = tb_page_addr1(tb); + if (likely(new_page1 != old_page1)) { + page0 = tb_page_addr0(tb); + if (unlikely(old_page1 != -1)) { + tb_unlock_page1(page0, old_page1); + } + tb_set_page_addr1(tb, new_page1); + tb_lock_page1(page0, new_page1); + } host = db->host_addr[1]; - base = TARGET_PAGE_ALIGN(db->pc_first); - if (host == NULL) { - tb_page_addr_t page0, old_page1, new_page1; - - new_page1 = get_page_addr_code_hostp(env, base, &db->host_addr[1]); - - /* - * If the second page is MMIO, treat as if the first page - * was MMIO as well, so that we do not cache the TB. - */ - if (unlikely(new_page1 == -1)) { - tb_unlock_pages(tb); - tb_set_page_addr0(tb, -1); - return NULL; - } - - /* - * If this is not the first time around, and page1 matches, - * then we already have the page locked. Alternately, we're - * not doing anything to prevent the PTE from changing, so - * we might wind up with a different page, requiring us to - * re-do the locking. - */ - old_page1 = tb_page_addr1(tb); - if (likely(new_page1 != old_page1)) { - page0 = tb_page_addr0(tb); - if (unlikely(old_page1 != -1)) { - tb_unlock_page1(page0, old_page1); - } - tb_set_page_addr1(tb, new_page1); - tb_lock_page1(page0, new_page1); - } - host = db->host_addr[1]; - } - - /* Use slow path when crossing pages. */ - if (is_same_page(db, pc)) { - return NULL; - } } - tcg_debug_assert(pc >= base); - return host + (pc - base); + memcpy(dest, host + (pc - base), len); + return true; } static void plugin_insn_append(vaddr pc, const void *from, size_t size) @@ -XXX,XX +XXX,XX @@ static void plugin_insn_append(vaddr pc, const void *from, size_t size) uint8_t translator_ldub(CPUArchState *env, DisasContextBase *db, vaddr pc) { - uint8_t ret; - void *p = translator_access(env, db, pc, sizeof(ret)); + uint8_t raw; - if (p) { - plugin_insn_append(pc, p, sizeof(ret)); - return ldub_p(p); + if (!translator_ld(env, db, &raw, pc, sizeof(raw))) { + raw = cpu_ldub_code(env, pc); } - ret = cpu_ldub_code(env, pc); - plugin_insn_append(pc, &ret, sizeof(ret)); - return ret; + plugin_insn_append(pc, &raw, sizeof(raw)); + return raw; } uint16_t translator_lduw(CPUArchState *env, DisasContextBase *db, vaddr pc) { - uint16_t ret, plug; - void *p = translator_access(env, db, pc, sizeof(ret)); + uint16_t raw, tgt; - if (p) { - plugin_insn_append(pc, p, sizeof(ret)); - return lduw_p(p); + if (translator_ld(env, db, &raw, pc, sizeof(raw))) { + tgt = tswap16(raw); + } else { + tgt = cpu_lduw_code(env, pc); + raw = tswap16(tgt); } - ret = cpu_lduw_code(env, pc); - plug = tswap16(ret); - plugin_insn_append(pc, &plug, sizeof(ret)); - return ret; + plugin_insn_append(pc, &raw, sizeof(raw)); + return tgt; } uint32_t translator_ldl(CPUArchState *env, DisasContextBase *db, vaddr pc) { - uint32_t ret, plug; - void *p = translator_access(env, db, pc, sizeof(ret)); + uint32_t raw, tgt; - if (p) { - plugin_insn_append(pc, p, sizeof(ret)); - return ldl_p(p); + if (translator_ld(env, db, &raw, pc, sizeof(raw))) { + tgt = tswap32(raw); + } else { + tgt = cpu_ldl_code(env, pc); + raw = tswap32(tgt); } - ret = cpu_ldl_code(env, pc); - plug = tswap32(ret); - plugin_insn_append(pc, &plug, sizeof(ret)); - return ret; + plugin_insn_append(pc, &raw, sizeof(raw)); + return tgt; } uint64_t translator_ldq(CPUArchState *env, DisasContextBase *db, vaddr pc) { - uint64_t ret, plug; - void *p = translator_access(env, db, pc, sizeof(ret)); + uint64_t raw, tgt; - if (p) { - plugin_insn_append(pc, p, sizeof(ret)); - return ldq_p(p); + if (translator_ld(env, db, &raw, pc, sizeof(raw))) { + tgt = tswap64(raw); + } else { + tgt = cpu_ldq_code(env, pc); + raw = tswap64(tgt); } - ret = cpu_ldq_code(env, pc); - plug = tswap64(ret); - plugin_insn_append(pc, &plug, sizeof(ret)); - return ret; + plugin_insn_append(pc, &raw, sizeof(raw)); + return tgt; } void translator_fake_ldb(DisasContextBase *db, vaddr pc, uint8_t insn8) -- 2.34.1 Do not allow translation to proceed beyond one insn with mmio, as we will not be caching the TranslationBlock. Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- accel/tcg/translator.c | 4 ++++ 1 file changed, 4 insertions(+) diff --git a/accel/tcg/translator.c b/accel/tcg/translator.c index XXXXXXX..XXXXXXX 100644 --- a/accel/tcg/translator.c +++ b/accel/tcg/translator.c @@ -XXX,XX +XXX,XX @@ static bool translator_ld(CPUArchState *env, DisasContextBase *db, /* Use slow path if first page is MMIO. */ if (unlikely(tb_page_addr0(tb) == -1)) { + /* We capped translation with first page MMIO in tb_gen_code. */ + tcg_debug_assert(db->max_insns == 1); return false; } @@ -XXX,XX +XXX,XX @@ static bool translator_ld(CPUArchState *env, DisasContextBase *db, if (unlikely(new_page1 == -1)) { tb_unlock_pages(tb); tb_set_page_addr0(tb, -1); + /* Require that this be the final insn. */ + db->max_insns = db->num_insns; return false; } -- 2.34.1 This will be able to replace plugin_insn_append, and will be usable for disassembly. Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- include/exec/translator.h | 12 ++++++++++++ accel/tcg/translator.c | 41 +++++++++++++++++++++++++++++++++++++++ 2 files changed, 53 insertions(+) diff --git a/include/exec/translator.h b/include/exec/translator.h index XXXXXXX..XXXXXXX 100644 --- a/include/exec/translator.h +++ b/include/exec/translator.h @@ -XXX,XX +XXX,XX @@ typedef struct DisasContextBase { bool plugin_enabled; struct TCGOp *insn_start; void *host_addr[2]; + + /* + * Record insn data that we cannot read directly from host memory. + * There are only two reasons we cannot use host memory: + * (1) We are executing from I/O, + * (2) We are executing a synthetic instruction (s390x EX). + * In both cases we need record exactly one instruction, + * and thus the maximum amount of data we record is limited. + */ + int record_start; + int record_len; + uint8_t record[32]; } DisasContextBase; /** diff --git a/accel/tcg/translator.c b/accel/tcg/translator.c index XXXXXXX..XXXXXXX 100644 --- a/accel/tcg/translator.c +++ b/accel/tcg/translator.c @@ -XXX,XX +XXX,XX @@ void translator_loop(CPUState *cpu, TranslationBlock *tb, int *max_insns, db->insn_start = NULL; db->host_addr[0] = host_pc; db->host_addr[1] = NULL; + db->record_start = 0; + db->record_len = 0; ops->init_disas_context(db, cpu); tcg_debug_assert(db->is_jmp == DISAS_NEXT); /* no early exit */ @@ -XXX,XX +XXX,XX @@ static bool translator_ld(CPUArchState *env, DisasContextBase *db, return true; } +static void record_save(DisasContextBase *db, vaddr pc, + const void *from, int size) +{ + int offset; + + /* Do not record probes before the start of TB. */ + if (pc < db->pc_first) { + return; + } + + /* + * In translator_access, we verified that pc is within 2 pages + * of pc_first, thus this will never overflow. + */ + offset = pc - db->pc_first; + + /* + * Either the first or second page may be I/O. If it is the second, + * then the first byte we need to record will be at a non-zero offset. + * In either case, we should not need to record but a single insn. + */ + if (db->record_len == 0) { + db->record_start = offset; + db->record_len = size; + } else { + assert(offset == db->record_start + db->record_len); + assert(db->record_len + size <= sizeof(db->record)); + db->record_len += size; + } + + memcpy(db->record + (offset - db->record_start), from, size); +} + static void plugin_insn_append(vaddr pc, const void *from, size_t size) { #ifdef CONFIG_PLUGIN @@ -XXX,XX +XXX,XX @@ uint8_t translator_ldub(CPUArchState *env, DisasContextBase *db, vaddr pc) if (!translator_ld(env, db, &raw, pc, sizeof(raw))) { raw = cpu_ldub_code(env, pc); + record_save(db, pc, &raw, sizeof(raw)); } plugin_insn_append(pc, &raw, sizeof(raw)); return raw; @@ -XXX,XX +XXX,XX @@ uint16_t translator_lduw(CPUArchState *env, DisasContextBase *db, vaddr pc) } else { tgt = cpu_lduw_code(env, pc); raw = tswap16(tgt); + record_save(db, pc, &raw, sizeof(raw)); } plugin_insn_append(pc, &raw, sizeof(raw)); return tgt; @@ -XXX,XX +XXX,XX @@ uint32_t translator_ldl(CPUArchState *env, DisasContextBase *db, vaddr pc) } else { tgt = cpu_ldl_code(env, pc); raw = tswap32(tgt); + record_save(db, pc, &raw, sizeof(raw)); } plugin_insn_append(pc, &raw, sizeof(raw)); return tgt; @@ -XXX,XX +XXX,XX @@ uint64_t translator_ldq(CPUArchState *env, DisasContextBase *db, vaddr pc) } else { tgt = cpu_ldq_code(env, pc); raw = tswap64(tgt); + record_save(db, pc, &raw, sizeof(raw)); } plugin_insn_append(pc, &raw, sizeof(raw)); return tgt; @@ -XXX,XX +XXX,XX @@ uint64_t translator_ldq(CPUArchState *env, DisasContextBase *db, vaddr pc) void translator_fake_ldb(DisasContextBase *db, vaddr pc, uint8_t insn8) { + assert(pc >= db->pc_first); + record_save(db, pc, &insn8, sizeof(insn8)); plugin_insn_append(pc, &insn8, sizeof(insn8)); } -- 2.34.1 Remove left-over comment from commit dcd092a063 ("accel/tcg: Improve can_do_io management"). Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- include/exec/translator.h | 3 ++- accel/tcg/translator.c | 2 ++ 2 files changed, 4 insertions(+), 1 deletion(-) diff --git a/include/exec/translator.h b/include/exec/translator.h index XXXXXXX..XXXXXXX 100644 --- a/include/exec/translator.h +++ b/include/exec/translator.h @@ -XXX,XX +XXX,XX @@ typedef enum DisasJumpType { * @num_insns: Number of translated instructions (including current). * @max_insns: Maximum number of instructions to be translated in this TB. * @singlestep_enabled: "Hardware" single stepping enabled. - * @saved_can_do_io: Known value of cpu->neg.can_do_io, or -1 for unknown. * @plugin_enabled: TCG plugin enabled in this TB. + * @fake_insn: True if translator_fake_ldb used. * @insn_start: The last op emitted by the insn_start hook, * which is expected to be INDEX_op_insn_start. * @@ -XXX,XX +XXX,XX @@ typedef struct DisasContextBase { int max_insns; bool singlestep_enabled; bool plugin_enabled; + bool fake_insn; struct TCGOp *insn_start; void *host_addr[2]; diff --git a/accel/tcg/translator.c b/accel/tcg/translator.c index XXXXXXX..XXXXXXX 100644 --- a/accel/tcg/translator.c +++ b/accel/tcg/translator.c @@ -XXX,XX +XXX,XX @@ void translator_loop(CPUState *cpu, TranslationBlock *tb, int *max_insns, db->max_insns = *max_insns; db->singlestep_enabled = cflags & CF_SINGLE_STEP; db->insn_start = NULL; + db->fake_insn = false; db->host_addr[0] = host_pc; db->host_addr[1] = NULL; db->record_start = 0; @@ -XXX,XX +XXX,XX @@ uint64_t translator_ldq(CPUArchState *env, DisasContextBase *db, vaddr pc) void translator_fake_ldb(DisasContextBase *db, vaddr pc, uint8_t insn8) { assert(pc >= db->pc_first); + db->fake_insn = true; record_save(db, pc, &insn8, sizeof(insn8)); plugin_insn_append(pc, &insn8, sizeof(insn8)); } -- 2.34.1 Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- include/tcg/tcg.h | 1 + accel/tcg/plugin-gen.c | 1 + 2 files changed, 2 insertions(+) diff --git a/include/tcg/tcg.h b/include/tcg/tcg.h index XXXXXXX..XXXXXXX 100644 --- a/include/tcg/tcg.h +++ b/include/tcg/tcg.h @@ -XXX,XX +XXX,XX @@ struct TCGContext { * space for instructions (for variable-instruction-length ISAs). */ struct qemu_plugin_tb *plugin_tb; + const struct DisasContextBase *plugin_db; /* descriptor of the instruction being translated */ struct qemu_plugin_insn *plugin_insn; diff --git a/accel/tcg/plugin-gen.c b/accel/tcg/plugin-gen.c index XXXXXXX..XXXXXXX 100644 --- a/accel/tcg/plugin-gen.c +++ b/accel/tcg/plugin-gen.c @@ -XXX,XX +XXX,XX @@ bool plugin_gen_tb_start(CPUState *cpu, const DisasContextBase *db, tcg_gen_plugin_cb(PLUGIN_GEN_FROM_TB); } + tcg_ctx->plugin_db = db; tcg_ctx->plugin_insn = NULL; return ret; -- 2.34.1 Instead of returning a host pointer, copy the data into storage provided by the caller. Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- include/qemu/qemu-plugin.h | 15 +++++++-------- contrib/plugins/execlog.c | 5 +++-- contrib/plugins/howvec.c | 4 ++-- plugins/api.c | 7 +++++-- 4 files changed, 17 insertions(+), 14 deletions(-) diff --git a/include/qemu/qemu-plugin.h b/include/qemu/qemu-plugin.h index XXXXXXX..XXXXXXX 100644 --- a/include/qemu/qemu-plugin.h +++ b/include/qemu/qemu-plugin.h @@ -XXX,XX +XXX,XX @@ typedef uint64_t qemu_plugin_id_t; extern QEMU_PLUGIN_EXPORT int qemu_plugin_version; -#define QEMU_PLUGIN_VERSION 2 +#define QEMU_PLUGIN_VERSION 3 /** * struct qemu_info_t - system information for plugins @@ -XXX,XX +XXX,XX @@ struct qemu_plugin_insn * qemu_plugin_tb_get_insn(const struct qemu_plugin_tb *tb, size_t idx); /** - * qemu_plugin_insn_data() - return ptr to instruction data + * qemu_plugin_insn_data() - copy instruction data * @insn: opaque instruction handle from qemu_plugin_tb_get_insn() + * @dest: destination into which data is copied + * @len: length of dest * - * Note: data is only valid for duration of callback. See - * qemu_plugin_insn_size() to calculate size of stream. - * - * Returns: pointer to a stream of bytes containing the value of this - * instructions opcode. + * Returns the number of bytes copied, minimum of @len and insn size. */ QEMU_PLUGIN_API -const void *qemu_plugin_insn_data(const struct qemu_plugin_insn *insn); +size_t qemu_plugin_insn_data(const struct qemu_plugin_insn *insn, + void *dest, size_t len); /** * qemu_plugin_insn_size() - return size of instruction diff --git a/contrib/plugins/execlog.c b/contrib/plugins/execlog.c index XXXXXXX..XXXXXXX 100644 --- a/contrib/plugins/execlog.c +++ b/contrib/plugins/execlog.c @@ -XXX,XX +XXX,XX @@ static void vcpu_tb_trans(qemu_plugin_id_t id, struct qemu_plugin_tb *tb) NULL); } } else { - uint32_t insn_opcode; - insn_opcode = *((uint32_t *)qemu_plugin_insn_data(insn)); + uint32_t insn_opcode = 0; + qemu_plugin_insn_data(insn, &insn_opcode, sizeof(insn_opcode)); + char *output = g_strdup_printf("0x%"PRIx64", 0x%"PRIx32", \"%s\"", insn_vaddr, insn_opcode, insn_disas); diff --git a/contrib/plugins/howvec.c b/contrib/plugins/howvec.c index XXXXXXX..XXXXXXX 100644 --- a/contrib/plugins/howvec.c +++ b/contrib/plugins/howvec.c @@ -XXX,XX +XXX,XX @@ static struct qemu_plugin_scoreboard *find_counter( { int i; uint64_t *cnt = NULL; - uint32_t opcode; + uint32_t opcode = 0; InsnClassExecCount *class = NULL; /* @@ -XXX,XX +XXX,XX @@ static struct qemu_plugin_scoreboard *find_counter( * They would probably benefit from a more tailored plugin. * However we can fall back to individual instruction counting. */ - opcode = *((uint32_t *)qemu_plugin_insn_data(insn)); + qemu_plugin_insn_data(insn, &opcode, sizeof(opcode)); for (i = 0; !cnt && i < class_table_sz; i++) { class = &class_table[i]; diff --git a/plugins/api.c b/plugins/api.c index XXXXXXX..XXXXXXX 100644 --- a/plugins/api.c +++ b/plugins/api.c @@ -XXX,XX +XXX,XX @@ qemu_plugin_tb_get_insn(const struct qemu_plugin_tb *tb, size_t idx) * instruction being translated. */ -const void *qemu_plugin_insn_data(const struct qemu_plugin_insn *insn) +size_t qemu_plugin_insn_data(const struct qemu_plugin_insn *insn, + void *dest, size_t len) { - return insn->data->data; + len = MIN(len, insn->data->len); + memcpy(dest, insn->data->data, len); + return len; } size_t qemu_plugin_insn_size(const struct qemu_plugin_insn *insn) -- 2.34.1 Copy data out of a completed translation. This will be used for both plugins and disassembly. Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- include/exec/translator.h | 23 ++++++++++++++++ accel/tcg/translator.c | 55 +++++++++++++++++++++++++++++++++++++++ 2 files changed, 78 insertions(+) diff --git a/include/exec/translator.h b/include/exec/translator.h index XXXXXXX..XXXXXXX 100644 --- a/include/exec/translator.h +++ b/include/exec/translator.h @@ -XXX,XX +XXX,XX @@ translator_ldq_swap(CPUArchState *env, DisasContextBase *db, */ void translator_fake_ldb(DisasContextBase *db, vaddr pc, uint8_t insn8); +/** + * translator_st + * @db: disassembly context + * @dest: address to copy into + * @addr: virtual address within TB + * @len: length + * + * Copy @len bytes from @addr into @dest. + * All bytes must have been read during translation. + * Return true on success or false on failure. + */ +bool translator_st(const DisasContextBase *db, void *dest, + vaddr addr, size_t len); + +/** + * translator_st_len + * @db: disassembly context + * + * Return the number of bytes available to copy from the + * current translation block with translator_st. + */ +size_t translator_st_len(const DisasContextBase *db); + #ifdef COMPILING_PER_TARGET /* * Return whether addr is on the same page as where disassembly started. diff --git a/accel/tcg/translator.c b/accel/tcg/translator.c index XXXXXXX..XXXXXXX 100644 --- a/accel/tcg/translator.c +++ b/accel/tcg/translator.c @@ -XXX,XX +XXX,XX @@ static void record_save(DisasContextBase *db, vaddr pc, memcpy(db->record + (offset - db->record_start), from, size); } +size_t translator_st_len(const DisasContextBase *db) +{ + return db->fake_insn ? db->record_len : db->tb->size; +} + +bool translator_st(const DisasContextBase *db, void *dest, + vaddr addr, size_t len) +{ + size_t offset, offset_end; + + if (addr < db->pc_first) { + return false; + } + offset = addr - db->pc_first; + offset_end = offset + len; + if (offset_end > translator_st_len(db)) { + return false; + } + + if (!db->fake_insn) { + size_t offset_page1 = -(db->pc_first | TARGET_PAGE_MASK); + + /* Get all the bytes from the first page. */ + if (db->host_addr[0]) { + if (offset_end <= offset_page1) { + memcpy(dest, db->host_addr[0] + offset, len); + return true; + } + if (offset < offset_page1) { + size_t len0 = offset_page1 - offset; + memcpy(dest, db->host_addr[0] + offset, len0); + offset += len0; + dest += len0; + } + } + + /* Get any bytes from the second page. */ + if (db->host_addr[1] && offset >= offset_page1) { + memcpy(dest, db->host_addr[1] + (offset - offset_page1), + offset_end - offset); + return true; + } + } + + /* Else get recorded bytes. */ + if (db->record_len != 0 && + offset >= db->record_start && + offset_end <= db->record_start + db->record_len) { + memcpy(dest, db->record + (offset - db->record_start), + offset_end - offset); + return true; + } + return false; +} + static void plugin_insn_append(vaddr pc, const void *from, size_t size) { #ifdef CONFIG_PLUGIN -- 2.34.1 Use the bytes that we record for the entire TB, rather than a per-insn GByteArray. Record the length of the insn in plugin_gen_insn_end rather than infering from the length of the array. Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- include/qemu/plugin.h | 14 +------------- accel/tcg/plugin-gen.c | 7 +++++-- accel/tcg/translator.c | 26 -------------------------- plugins/api.c | 12 +++++++----- tcg/tcg.c | 3 +-- 5 files changed, 14 insertions(+), 48 deletions(-) diff --git a/include/qemu/plugin.h b/include/qemu/plugin.h index XXXXXXX..XXXXXXX 100644 --- a/include/qemu/plugin.h +++ b/include/qemu/plugin.h @@ -XXX,XX +XXX,XX @@ struct qemu_plugin_dyn_cb { /* Internal context for instrumenting an instruction */ struct qemu_plugin_insn { - GByteArray *data; uint64_t vaddr; void *haddr; GArray *insn_cbs; GArray *mem_cbs; + uint8_t len; bool calls_helpers; /* if set, the instruction calls helpers that might access guest memory */ @@ -XXX,XX +XXX,XX @@ struct qemu_plugin_scoreboard { QLIST_ENTRY(qemu_plugin_scoreboard) entry; }; -/* - * qemu_plugin_insn allocate and cleanup functions. We don't expect to - * cleanup many of these structures. They are reused for each fresh - * translation. - */ - -static inline void qemu_plugin_insn_cleanup_fn(gpointer data) -{ - struct qemu_plugin_insn *insn = (struct qemu_plugin_insn *) data; - g_byte_array_free(insn->data, true); -} - /* Internal context for this TranslationBlock */ struct qemu_plugin_tb { GPtrArray *insns; diff --git a/accel/tcg/plugin-gen.c b/accel/tcg/plugin-gen.c index XXXXXXX..XXXXXXX 100644 --- a/accel/tcg/plugin-gen.c +++ b/accel/tcg/plugin-gen.c @@ -XXX,XX +XXX,XX @@ void plugin_gen_insn_start(CPUState *cpu, const DisasContextBase *db) ptb->n = n; if (n <= ptb->insns->len) { insn = g_ptr_array_index(ptb->insns, n - 1); - g_byte_array_set_size(insn->data, 0); } else { assert(n - 1 == ptb->insns->len); insn = g_new0(struct qemu_plugin_insn, 1); - insn->data = g_byte_array_sized_new(4); g_ptr_array_add(ptb->insns, insn); } @@ -XXX,XX +XXX,XX @@ void plugin_gen_insn_start(CPUState *cpu, const DisasContextBase *db) void plugin_gen_insn_end(void) { + const DisasContextBase *db = tcg_ctx->plugin_db; + struct qemu_plugin_insn *pinsn = tcg_ctx->plugin_insn; + + pinsn->len = db->fake_insn ? db->record_len : db->pc_next - pinsn->vaddr; + tcg_gen_plugin_cb(PLUGIN_GEN_AFTER_INSN); } diff --git a/accel/tcg/translator.c b/accel/tcg/translator.c index XXXXXXX..XXXXXXX 100644 --- a/accel/tcg/translator.c +++ b/accel/tcg/translator.c @@ -XXX,XX +XXX,XX @@ bool translator_st(const DisasContextBase *db, void *dest, return false; } -static void plugin_insn_append(vaddr pc, const void *from, size_t size) -{ -#ifdef CONFIG_PLUGIN - struct qemu_plugin_insn *insn = tcg_ctx->plugin_insn; - size_t off; - - if (insn == NULL) { - return; - } - off = pc - insn->vaddr; - if (off < insn->data->len) { - g_byte_array_set_size(insn->data, off); - } else if (off > insn->data->len) { - /* we have an unexpected gap */ - g_assert_not_reached(); - } - - insn->data = g_byte_array_append(insn->data, from, size); -#endif -} - uint8_t translator_ldub(CPUArchState *env, DisasContextBase *db, vaddr pc) { uint8_t raw; @@ -XXX,XX +XXX,XX @@ uint8_t translator_ldub(CPUArchState *env, DisasContextBase *db, vaddr pc) raw = cpu_ldub_code(env, pc); record_save(db, pc, &raw, sizeof(raw)); } - plugin_insn_append(pc, &raw, sizeof(raw)); return raw; } @@ -XXX,XX +XXX,XX @@ uint16_t translator_lduw(CPUArchState *env, DisasContextBase *db, vaddr pc) raw = tswap16(tgt); record_save(db, pc, &raw, sizeof(raw)); } - plugin_insn_append(pc, &raw, sizeof(raw)); return tgt; } @@ -XXX,XX +XXX,XX @@ uint32_t translator_ldl(CPUArchState *env, DisasContextBase *db, vaddr pc) raw = tswap32(tgt); record_save(db, pc, &raw, sizeof(raw)); } - plugin_insn_append(pc, &raw, sizeof(raw)); return tgt; } @@ -XXX,XX +XXX,XX @@ uint64_t translator_ldq(CPUArchState *env, DisasContextBase *db, vaddr pc) raw = tswap64(tgt); record_save(db, pc, &raw, sizeof(raw)); } - plugin_insn_append(pc, &raw, sizeof(raw)); return tgt; } @@ -XXX,XX +XXX,XX @@ void translator_fake_ldb(DisasContextBase *db, vaddr pc, uint8_t insn8) assert(pc >= db->pc_first); db->fake_insn = true; record_save(db, pc, &insn8, sizeof(insn8)); - plugin_insn_append(pc, &insn8, sizeof(insn8)); } diff --git a/plugins/api.c b/plugins/api.c index XXXXXXX..XXXXXXX 100644 --- a/plugins/api.c +++ b/plugins/api.c @@ -XXX,XX +XXX,XX @@ #include "tcg/tcg.h" #include "exec/exec-all.h" #include "exec/gdbstub.h" +#include "exec/translator.h" #include "disas/disas.h" #include "plugin.h" #ifndef CONFIG_USER_ONLY @@ -XXX,XX +XXX,XX @@ qemu_plugin_tb_get_insn(const struct qemu_plugin_tb *tb, size_t idx) size_t qemu_plugin_insn_data(const struct qemu_plugin_insn *insn, void *dest, size_t len) { - len = MIN(len, insn->data->len); - memcpy(dest, insn->data->data, len); - return len; + const DisasContextBase *db = tcg_ctx->plugin_db; + + len = MIN(len, insn->len); + return translator_st(db, dest, insn->vaddr, len) ? len : 0; } size_t qemu_plugin_insn_size(const struct qemu_plugin_insn *insn) { - return insn->data->len; + return insn->len; } uint64_t qemu_plugin_insn_vaddr(const struct qemu_plugin_insn *insn) @@ -XXX,XX +XXX,XX @@ void *qemu_plugin_insn_haddr(const struct qemu_plugin_insn *insn) char *qemu_plugin_insn_disas(const struct qemu_plugin_insn *insn) { CPUState *cpu = current_cpu; - return plugin_disas(cpu, insn->vaddr, insn->data->len); + return plugin_disas(cpu, insn->vaddr, insn->len); } const char *qemu_plugin_insn_symbol(const struct qemu_plugin_insn *insn) diff --git a/tcg/tcg.c b/tcg/tcg.c index XXXXXXX..XXXXXXX 100644 --- a/tcg/tcg.c +++ b/tcg/tcg.c @@ -XXX,XX +XXX,XX @@ static void alloc_tcg_plugin_context(TCGContext *s) { #ifdef CONFIG_PLUGIN s->plugin_tb = g_new0(struct qemu_plugin_tb, 1); - s->plugin_tb->insns = - g_ptr_array_new_with_free_func(qemu_plugin_insn_cleanup_fn); + s->plugin_tb->insns = g_ptr_array_new(); #endif } -- 2.34.1 Do not pass around a boolean between multiple structures, just read it from the TranslationBlock in the TCGContext. Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- include/exec/plugin-gen.h | 7 +++---- include/qemu/plugin.h | 3 --- accel/tcg/plugin-gen.c | 4 +--- accel/tcg/translator.c | 2 +- plugins/api.c | 14 +++++++++----- 5 files changed, 14 insertions(+), 16 deletions(-) diff --git a/include/exec/plugin-gen.h b/include/exec/plugin-gen.h index XXXXXXX..XXXXXXX 100644 --- a/include/exec/plugin-gen.h +++ b/include/exec/plugin-gen.h @@ -XXX,XX +XXX,XX @@ struct DisasContextBase; #ifdef CONFIG_PLUGIN -bool plugin_gen_tb_start(CPUState *cpu, const struct DisasContextBase *db, - bool supress); +bool plugin_gen_tb_start(CPUState *cpu, const struct DisasContextBase *db); void plugin_gen_tb_end(CPUState *cpu, size_t num_insns); void plugin_gen_insn_start(CPUState *cpu, const struct DisasContextBase *db); void plugin_gen_insn_end(void); @@ -XXX,XX +XXX,XX @@ void plugin_gen_disable_mem_helpers(void); #else /* !CONFIG_PLUGIN */ -static inline bool -plugin_gen_tb_start(CPUState *cpu, const struct DisasContextBase *db, bool sup) +static inline +bool plugin_gen_tb_start(CPUState *cpu, const struct DisasContextBase *db) { return false; } diff --git a/include/qemu/plugin.h b/include/qemu/plugin.h index XXXXXXX..XXXXXXX 100644 --- a/include/qemu/plugin.h +++ b/include/qemu/plugin.h @@ -XXX,XX +XXX,XX @@ struct qemu_plugin_insn { /* if set, the instruction calls helpers that might access guest memory */ bool mem_helper; - - bool mem_only; }; /* A scoreboard is an array of values, indexed by vcpu_index */ @@ -XXX,XX +XXX,XX @@ struct qemu_plugin_tb { uint64_t vaddr2; void *haddr1; void *haddr2; - bool mem_only; /* if set, the TB calls helpers that might access guest memory */ bool mem_helper; diff --git a/accel/tcg/plugin-gen.c b/accel/tcg/plugin-gen.c index XXXXXXX..XXXXXXX 100644 --- a/accel/tcg/plugin-gen.c +++ b/accel/tcg/plugin-gen.c @@ -XXX,XX +XXX,XX @@ static void plugin_gen_inject(struct qemu_plugin_tb *plugin_tb) } } -bool plugin_gen_tb_start(CPUState *cpu, const DisasContextBase *db, - bool mem_only) +bool plugin_gen_tb_start(CPUState *cpu, const DisasContextBase *db) { bool ret = false; @@ -XXX,XX +XXX,XX @@ bool plugin_gen_tb_start(CPUState *cpu, const DisasContextBase *db, ptb->vaddr2 = -1; ptb->haddr1 = db->host_addr[0]; ptb->haddr2 = NULL; - ptb->mem_only = mem_only; ptb->mem_helper = false; tcg_gen_plugin_cb(PLUGIN_GEN_FROM_TB); diff --git a/accel/tcg/translator.c b/accel/tcg/translator.c index XXXXXXX..XXXXXXX 100644 --- a/accel/tcg/translator.c +++ b/accel/tcg/translator.c @@ -XXX,XX +XXX,XX @@ void translator_loop(CPUState *cpu, TranslationBlock *tb, int *max_insns, ops->tb_start(db, cpu); tcg_debug_assert(db->is_jmp == DISAS_NEXT); /* no early exit */ - plugin_enabled = plugin_gen_tb_start(cpu, db, cflags & CF_MEMI_ONLY); + plugin_enabled = plugin_gen_tb_start(cpu, db); db->plugin_enabled = plugin_enabled; while (true) { diff --git a/plugins/api.c b/plugins/api.c index XXXXXXX..XXXXXXX 100644 --- a/plugins/api.c +++ b/plugins/api.c @@ -XXX,XX +XXX,XX @@ void qemu_plugin_register_vcpu_exit_cb(qemu_plugin_id_t id, plugin_register_cb(id, QEMU_PLUGIN_EV_VCPU_EXIT, cb); } +static bool tb_is_mem_only(void) +{ + return tb_cflags(tcg_ctx->gen_tb) & CF_MEMI_ONLY; +} + void qemu_plugin_register_vcpu_tb_exec_cb(struct qemu_plugin_tb *tb, qemu_plugin_vcpu_udata_cb_t cb, enum qemu_plugin_cb_flags flags, void *udata) { - if (!tb->mem_only) { + if (!tb_is_mem_only()) { plugin_register_dyn_cb__udata(&tb->cbs, cb, flags, udata); } } @@ -XXX,XX +XXX,XX @@ void qemu_plugin_register_vcpu_tb_exec_inline_per_vcpu( qemu_plugin_u64 entry, uint64_t imm) { - if (!tb->mem_only) { + if (!tb_is_mem_only()) { plugin_register_inline_op_on_entry(&tb->cbs, 0, op, entry, imm); } } @@ -XXX,XX +XXX,XX @@ void qemu_plugin_register_vcpu_insn_exec_cb(struct qemu_plugin_insn *insn, enum qemu_plugin_cb_flags flags, void *udata) { - if (!insn->mem_only) { + if (!tb_is_mem_only()) { plugin_register_dyn_cb__udata(&insn->insn_cbs, cb, flags, udata); } } @@ -XXX,XX +XXX,XX @@ void qemu_plugin_register_vcpu_insn_exec_inline_per_vcpu( qemu_plugin_u64 entry, uint64_t imm) { - if (!insn->mem_only) { + if (!tb_is_mem_only()) { plugin_register_inline_op_on_entry(&insn->insn_cbs, 0, op, entry, imm); } } @@ -XXX,XX +XXX,XX @@ qemu_plugin_tb_get_insn(const struct qemu_plugin_tb *tb, size_t idx) return NULL; } insn = g_ptr_array_index(tb->insns, idx); - insn->mem_only = tb->mem_only; return insn; } -- 2.34.1 We can delay the computation of haddr until the plugin actually requests it. Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- include/qemu/plugin.h | 4 ---- accel/tcg/plugin-gen.c | 20 -------------------- plugins/api.c | 25 ++++++++++++++++++++++++- 3 files changed, 24 insertions(+), 25 deletions(-) diff --git a/include/qemu/plugin.h b/include/qemu/plugin.h index XXXXXXX..XXXXXXX 100644 --- a/include/qemu/plugin.h +++ b/include/qemu/plugin.h @@ -XXX,XX +XXX,XX @@ struct qemu_plugin_dyn_cb { /* Internal context for instrumenting an instruction */ struct qemu_plugin_insn { uint64_t vaddr; - void *haddr; GArray *insn_cbs; GArray *mem_cbs; uint8_t len; @@ -XXX,XX +XXX,XX @@ struct qemu_plugin_tb { GPtrArray *insns; size_t n; uint64_t vaddr; - uint64_t vaddr2; - void *haddr1; - void *haddr2; /* if set, the TB calls helpers that might access guest memory */ bool mem_helper; diff --git a/accel/tcg/plugin-gen.c b/accel/tcg/plugin-gen.c index XXXXXXX..XXXXXXX 100644 --- a/accel/tcg/plugin-gen.c +++ b/accel/tcg/plugin-gen.c @@ -XXX,XX +XXX,XX @@ bool plugin_gen_tb_start(CPUState *cpu, const DisasContextBase *db) ret = true; ptb->vaddr = db->pc_first; - ptb->vaddr2 = -1; - ptb->haddr1 = db->host_addr[0]; - ptb->haddr2 = NULL; ptb->mem_helper = false; tcg_gen_plugin_cb(PLUGIN_GEN_FROM_TB); @@ -XXX,XX +XXX,XX @@ void plugin_gen_insn_start(CPUState *cpu, const DisasContextBase *db) pc = db->pc_next; insn->vaddr = pc; - /* - * Detect page crossing to get the new host address. - * Note that we skip this when haddr1 == NULL, e.g. when we're - * fetching instructions from a region not backed by RAM. - */ - if (ptb->haddr1 == NULL) { - insn->haddr = NULL; - } else if (is_same_page(db, db->pc_next)) { - insn->haddr = ptb->haddr1 + pc - ptb->vaddr; - } else { - if (ptb->vaddr2 == -1) { - ptb->vaddr2 = TARGET_PAGE_ALIGN(db->pc_first); - get_page_addr_code_hostp(cpu_env(cpu), ptb->vaddr2, &ptb->haddr2); - } - insn->haddr = ptb->haddr2 + pc - ptb->vaddr2; - } - tcg_gen_plugin_cb(PLUGIN_GEN_FROM_INSN); } diff --git a/plugins/api.c b/plugins/api.c index XXXXXXX..XXXXXXX 100644 --- a/plugins/api.c +++ b/plugins/api.c @@ -XXX,XX +XXX,XX @@ uint64_t qemu_plugin_insn_vaddr(const struct qemu_plugin_insn *insn) void *qemu_plugin_insn_haddr(const struct qemu_plugin_insn *insn) { - return insn->haddr; + const DisasContextBase *db = tcg_ctx->plugin_db; + vaddr page0_last = db->pc_first | ~TARGET_PAGE_MASK; + + if (db->fake_insn) { + return NULL; + } + + /* + * ??? The return value is not intended for use of host memory, + * but as a proxy for address space and physical address. + * Thus we are only interested in the first byte and do not + * care about spanning pages. + */ + if (insn->vaddr <= page0_last) { + if (db->host_addr[0] == NULL) { + return NULL; + } + return db->host_addr[0] + insn->vaddr - db->pc_first; + } else { + if (db->host_addr[1] == NULL) { + return NULL; + } + return db->host_addr[1] + insn->vaddr - (page0_last + 1); + } } char *qemu_plugin_insn_disas(const struct qemu_plugin_insn *insn) -- 2.34.1 We do not need to separately record the start of the TB. Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- include/qemu/plugin.h | 1 - accel/tcg/plugin-gen.c | 3 +-- plugins/api.c | 3 ++- 3 files changed, 3 insertions(+), 4 deletions(-) diff --git a/include/qemu/plugin.h b/include/qemu/plugin.h index XXXXXXX..XXXXXXX 100644 --- a/include/qemu/plugin.h +++ b/include/qemu/plugin.h @@ -XXX,XX +XXX,XX @@ struct qemu_plugin_scoreboard { struct qemu_plugin_tb { GPtrArray *insns; size_t n; - uint64_t vaddr; /* if set, the TB calls helpers that might access guest memory */ bool mem_helper; diff --git a/accel/tcg/plugin-gen.c b/accel/tcg/plugin-gen.c index XXXXXXX..XXXXXXX 100644 --- a/accel/tcg/plugin-gen.c +++ b/accel/tcg/plugin-gen.c @@ -XXX,XX +XXX,XX @@ static void plugin_gen_inject(struct qemu_plugin_tb *plugin_tb) int insn_idx = -1; if (unlikely(qemu_loglevel_mask(LOG_TB_OP_PLUGIN) - && qemu_log_in_addr_range(plugin_tb->vaddr))) { + && qemu_log_in_addr_range(tcg_ctx->plugin_db->pc_first))) { FILE *logfile = qemu_log_trylock(); if (logfile) { fprintf(logfile, "OP before plugin injection:\n"); @@ -XXX,XX +XXX,XX @@ bool plugin_gen_tb_start(CPUState *cpu, const DisasContextBase *db) ret = true; - ptb->vaddr = db->pc_first; ptb->mem_helper = false; tcg_gen_plugin_cb(PLUGIN_GEN_FROM_TB); diff --git a/plugins/api.c b/plugins/api.c index XXXXXXX..XXXXXXX 100644 --- a/plugins/api.c +++ b/plugins/api.c @@ -XXX,XX +XXX,XX @@ size_t qemu_plugin_tb_n_insns(const struct qemu_plugin_tb *tb) uint64_t qemu_plugin_tb_vaddr(const struct qemu_plugin_tb *tb) { - return tb->vaddr; + const DisasContextBase *db = tcg_ctx->plugin_db; + return db->pc_first; } struct qemu_plugin_insn * -- 2.34.1 We don't need to allocate plugin context at startup, we can wait until we actually use it. Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- accel/tcg/plugin-gen.c | 36 ++++++++++++++++++++---------------- tcg/tcg.c | 11 ----------- 2 files changed, 20 insertions(+), 27 deletions(-) diff --git a/accel/tcg/plugin-gen.c b/accel/tcg/plugin-gen.c index XXXXXXX..XXXXXXX 100644 --- a/accel/tcg/plugin-gen.c +++ b/accel/tcg/plugin-gen.c @@ -XXX,XX +XXX,XX @@ static void plugin_gen_inject(struct qemu_plugin_tb *plugin_tb) bool plugin_gen_tb_start(CPUState *cpu, const DisasContextBase *db) { - bool ret = false; + struct qemu_plugin_tb *ptb; - if (test_bit(QEMU_PLUGIN_EV_VCPU_TB_TRANS, cpu->plugin_state->event_mask)) { - struct qemu_plugin_tb *ptb = tcg_ctx->plugin_tb; - - /* reset callbacks */ - if (ptb->cbs) { - g_array_set_size(ptb->cbs, 0); - } - ptb->n = 0; - - ret = true; - - ptb->mem_helper = false; - - tcg_gen_plugin_cb(PLUGIN_GEN_FROM_TB); + if (!test_bit(QEMU_PLUGIN_EV_VCPU_TB_TRANS, + cpu->plugin_state->event_mask)) { + return false; } tcg_ctx->plugin_db = db; tcg_ctx->plugin_insn = NULL; + ptb = tcg_ctx->plugin_tb; - return ret; + if (ptb) { + /* Reset callbacks */ + if (ptb->cbs) { + g_array_set_size(ptb->cbs, 0); + } + ptb->n = 0; + ptb->mem_helper = false; + } else { + ptb = g_new0(struct qemu_plugin_tb, 1); + tcg_ctx->plugin_tb = ptb; + ptb->insns = g_ptr_array_new(); + } + + tcg_gen_plugin_cb(PLUGIN_GEN_FROM_TB); + return true; } void plugin_gen_insn_start(CPUState *cpu, const DisasContextBase *db) diff --git a/tcg/tcg.c b/tcg/tcg.c index XXXXXXX..XXXXXXX 100644 --- a/tcg/tcg.c +++ b/tcg/tcg.c @@ -XXX,XX +XXX,XX @@ QEMU_BUILD_BUG_ON((int)(offsetof(CPUNegativeOffsetState, tlb.f[0]) - < MIN_TLB_MASK_TABLE_OFS); #endif -static void alloc_tcg_plugin_context(TCGContext *s) -{ -#ifdef CONFIG_PLUGIN - s->plugin_tb = g_new0(struct qemu_plugin_tb, 1); - s->plugin_tb->insns = g_ptr_array_new(); -#endif -} - /* * All TCG threads except the parent (i.e. the one that called tcg_context_init * and registered the target's TCG globals) must register with this function @@ -XXX,XX +XXX,XX @@ void tcg_register_thread(void) qatomic_set(&tcg_ctxs[n], s); if (n > 0) { - alloc_tcg_plugin_context(s); tcg_region_initial_alloc(s); } @@ -XXX,XX +XXX,XX @@ static void tcg_context_init(unsigned max_cpus) indirect_reg_alloc_order[i] = tcg_target_reg_alloc_order[i]; } - alloc_tcg_plugin_context(s); - tcg_ctx = s; /* * In user-mode we simply share the init context among threads, since we -- 2.34.1 Almost all of the disas_log implementations are identical. Unify them within translator_loop. Drop extra Priv/Virt logging from target/riscv. Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- accel/tcg/translator.c | 9 ++++++++- target/alpha/translate.c | 9 --------- target/arm/tcg/translate-a64.c | 11 ----------- target/arm/tcg/translate.c | 12 ------------ target/avr/translate.c | 8 -------- target/cris/translate.c | 11 ----------- target/hexagon/translate.c | 9 --------- target/hppa/translate.c | 6 ++++-- target/i386/tcg/translate.c | 11 ----------- target/loongarch/tcg/translate.c | 8 -------- target/m68k/translate.c | 9 --------- target/microblaze/translate.c | 9 --------- target/mips/tcg/translate.c | 9 --------- target/openrisc/translate.c | 11 ----------- target/ppc/translate.c | 9 --------- target/riscv/translate.c | 18 ------------------ target/rx/translate.c | 8 -------- target/sh4/translate.c | 9 --------- target/sparc/translate.c | 9 --------- target/tricore/translate.c | 9 --------- target/xtensa/translate.c | 9 --------- 21 files changed, 12 insertions(+), 191 deletions(-) diff --git a/accel/tcg/translator.c b/accel/tcg/translator.c index XXXXXXX..XXXXXXX 100644 --- a/accel/tcg/translator.c +++ b/accel/tcg/translator.c @@ -XXX,XX +XXX,XX @@ #include "exec/cpu_ldst.h" #include "tcg/tcg-op-common.h" #include "internal-target.h" +#include "disas/disas.h" static void set_can_do_io(DisasContextBase *db, bool val) { @@ -XXX,XX +XXX,XX @@ void translator_loop(CPUState *cpu, TranslationBlock *tb, int *max_insns, FILE *logfile = qemu_log_trylock(); if (logfile) { fprintf(logfile, "----------------\n"); - ops->disas_log(db, cpu, logfile); + + if (ops->disas_log) { + ops->disas_log(db, cpu, logfile); + } else { + fprintf(logfile, "IN: %s\n", lookup_symbol(db->pc_first)); + target_disas(logfile, cpu, db->pc_first, db->tb->size); + } fprintf(logfile, "\n"); qemu_log_unlock(logfile); } diff --git a/target/alpha/translate.c b/target/alpha/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/alpha/translate.c +++ b/target/alpha/translate.c @@ -XXX,XX +XXX,XX @@ #include "qemu/osdep.h" #include "cpu.h" #include "sysemu/cpus.h" -#include "disas/disas.h" #include "qemu/host-utils.h" #include "exec/exec-all.h" #include "tcg/tcg-op.h" @@ -XXX,XX +XXX,XX @@ static void alpha_tr_tb_stop(DisasContextBase *dcbase, CPUState *cpu) } } -static void alpha_tr_disas_log(const DisasContextBase *dcbase, - CPUState *cpu, FILE *logfile) -{ - fprintf(logfile, "IN: %s\n", lookup_symbol(dcbase->pc_first)); - target_disas(logfile, cpu, dcbase->pc_first, dcbase->tb->size); -} - static const TranslatorOps alpha_tr_ops = { .init_disas_context = alpha_tr_init_disas_context, .tb_start = alpha_tr_tb_start, .insn_start = alpha_tr_insn_start, .translate_insn = alpha_tr_translate_insn, .tb_stop = alpha_tr_tb_stop, - .disas_log = alpha_tr_disas_log, }; void gen_intermediate_code(CPUState *cpu, TranslationBlock *tb, int *max_insns, diff --git a/target/arm/tcg/translate-a64.c b/target/arm/tcg/translate-a64.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/tcg/translate-a64.c +++ b/target/arm/tcg/translate-a64.c @@ -XXX,XX +XXX,XX @@ #include "translate.h" #include "translate-a64.h" #include "qemu/log.h" -#include "disas/disas.h" #include "arm_ldst.h" #include "semihosting/semihost.h" #include "cpregs.h" @@ -XXX,XX +XXX,XX @@ static void aarch64_tr_tb_stop(DisasContextBase *dcbase, CPUState *cpu) } } -static void aarch64_tr_disas_log(const DisasContextBase *dcbase, - CPUState *cpu, FILE *logfile) -{ - DisasContext *dc = container_of(dcbase, DisasContext, base); - - fprintf(logfile, "IN: %s\n", lookup_symbol(dc->base.pc_first)); - target_disas(logfile, cpu, dc->base.pc_first, dc->base.tb->size); -} - const TranslatorOps aarch64_translator_ops = { .init_disas_context = aarch64_tr_init_disas_context, .tb_start = aarch64_tr_tb_start, .insn_start = aarch64_tr_insn_start, .translate_insn = aarch64_tr_translate_insn, .tb_stop = aarch64_tr_tb_stop, - .disas_log = aarch64_tr_disas_log, }; diff --git a/target/arm/tcg/translate.c b/target/arm/tcg/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/tcg/translate.c +++ b/target/arm/tcg/translate.c @@ -XXX,XX +XXX,XX @@ #include "translate.h" #include "translate-a32.h" #include "qemu/log.h" -#include "disas/disas.h" #include "arm_ldst.h" #include "semihosting/semihost.h" #include "cpregs.h" @@ -XXX,XX +XXX,XX @@ static void arm_tr_tb_stop(DisasContextBase *dcbase, CPUState *cpu) } } -static void arm_tr_disas_log(const DisasContextBase *dcbase, - CPUState *cpu, FILE *logfile) -{ - DisasContext *dc = container_of(dcbase, DisasContext, base); - - fprintf(logfile, "IN: %s\n", lookup_symbol(dc->base.pc_first)); - target_disas(logfile, cpu, dc->base.pc_first, dc->base.tb->size); -} - static const TranslatorOps arm_translator_ops = { .init_disas_context = arm_tr_init_disas_context, .tb_start = arm_tr_tb_start, .insn_start = arm_tr_insn_start, .translate_insn = arm_tr_translate_insn, .tb_stop = arm_tr_tb_stop, - .disas_log = arm_tr_disas_log, }; static const TranslatorOps thumb_translator_ops = { @@ -XXX,XX +XXX,XX @@ static const TranslatorOps thumb_translator_ops = { .insn_start = arm_tr_insn_start, .translate_insn = thumb_tr_translate_insn, .tb_stop = arm_tr_tb_stop, - .disas_log = arm_tr_disas_log, }; /* generate intermediate code for basic block 'tb'. */ diff --git a/target/avr/translate.c b/target/avr/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/avr/translate.c +++ b/target/avr/translate.c @@ -XXX,XX +XXX,XX @@ static void avr_tr_tb_stop(DisasContextBase *dcbase, CPUState *cs) } } -static void avr_tr_disas_log(const DisasContextBase *dcbase, - CPUState *cs, FILE *logfile) -{ - fprintf(logfile, "IN: %s\n", lookup_symbol(dcbase->pc_first)); - target_disas(logfile, cs, dcbase->pc_first, dcbase->tb->size); -} - static const TranslatorOps avr_tr_ops = { .init_disas_context = avr_tr_init_disas_context, .tb_start = avr_tr_tb_start, .insn_start = avr_tr_insn_start, .translate_insn = avr_tr_translate_insn, .tb_stop = avr_tr_tb_stop, - .disas_log = avr_tr_disas_log, }; void gen_intermediate_code(CPUState *cs, TranslationBlock *tb, int *max_insns, diff --git a/target/cris/translate.c b/target/cris/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/cris/translate.c +++ b/target/cris/translate.c @@ -XXX,XX +XXX,XX @@ #include "qemu/osdep.h" #include "cpu.h" -#include "disas/disas.h" #include "exec/exec-all.h" #include "tcg/tcg-op.h" #include "exec/helper-proto.h" @@ -XXX,XX +XXX,XX @@ static void cris_tr_tb_stop(DisasContextBase *dcbase, CPUState *cpu) } } -static void cris_tr_disas_log(const DisasContextBase *dcbase, - CPUState *cpu, FILE *logfile) -{ - if (!DISAS_CRIS) { - fprintf(logfile, "IN: %s\n", lookup_symbol(dcbase->pc_first)); - target_disas(logfile, cpu, dcbase->pc_first, dcbase->tb->size); - } -} - static const TranslatorOps cris_tr_ops = { .init_disas_context = cris_tr_init_disas_context, .tb_start = cris_tr_tb_start, .insn_start = cris_tr_insn_start, .translate_insn = cris_tr_translate_insn, .tb_stop = cris_tr_tb_stop, - .disas_log = cris_tr_disas_log, }; void gen_intermediate_code(CPUState *cs, TranslationBlock *tb, int *max_insns, diff --git a/target/hexagon/translate.c b/target/hexagon/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/hexagon/translate.c +++ b/target/hexagon/translate.c @@ -XXX,XX +XXX,XX @@ static void hexagon_tr_tb_stop(DisasContextBase *dcbase, CPUState *cpu) } } -static void hexagon_tr_disas_log(const DisasContextBase *dcbase, - CPUState *cpu, FILE *logfile) -{ - fprintf(logfile, "IN: %s\n", lookup_symbol(dcbase->pc_first)); - target_disas(logfile, cpu, dcbase->pc_first, dcbase->tb->size); -} - - static const TranslatorOps hexagon_tr_ops = { .init_disas_context = hexagon_tr_init_disas_context, .tb_start = hexagon_tr_tb_start, .insn_start = hexagon_tr_insn_start, .translate_insn = hexagon_tr_translate_packet, .tb_stop = hexagon_tr_tb_stop, - .disas_log = hexagon_tr_disas_log, }; void gen_intermediate_code(CPUState *cs, TranslationBlock *tb, int *max_insns, diff --git a/target/hppa/translate.c b/target/hppa/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/hppa/translate.c +++ b/target/hppa/translate.c @@ -XXX,XX +XXX,XX @@ static void hppa_tr_tb_stop(DisasContextBase *dcbase, CPUState *cs) } } +#ifdef CONFIG_USER_ONLY static void hppa_tr_disas_log(const DisasContextBase *dcbase, CPUState *cs, FILE *logfile) { target_ulong pc = dcbase->pc_first; -#ifdef CONFIG_USER_ONLY switch (pc) { case 0x00: fprintf(logfile, "IN:\n0x00000000: (null)\n"); @@ -XXX,XX +XXX,XX @@ static void hppa_tr_disas_log(const DisasContextBase *dcbase, fprintf(logfile, "IN:\n0x00000100: syscall\n"); return; } -#endif fprintf(logfile, "IN: %s\n", lookup_symbol(pc)); target_disas(logfile, cs, pc, dcbase->tb->size); } +#endif static const TranslatorOps hppa_tr_ops = { .init_disas_context = hppa_tr_init_disas_context, @@ -XXX,XX +XXX,XX @@ static const TranslatorOps hppa_tr_ops = { .insn_start = hppa_tr_insn_start, .translate_insn = hppa_tr_translate_insn, .tb_stop = hppa_tr_tb_stop, +#ifdef CONFIG_USER_ONLY .disas_log = hppa_tr_disas_log, +#endif }; void gen_intermediate_code(CPUState *cs, TranslationBlock *tb, int *max_insns, diff --git a/target/i386/tcg/translate.c b/target/i386/tcg/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/i386/tcg/translate.c +++ b/target/i386/tcg/translate.c @@ -XXX,XX +XXX,XX @@ #include "qemu/host-utils.h" #include "cpu.h" -#include "disas/disas.h" #include "exec/exec-all.h" #include "tcg/tcg-op.h" #include "tcg/tcg-op-gvec.h" @@ -XXX,XX +XXX,XX @@ static void i386_tr_tb_stop(DisasContextBase *dcbase, CPUState *cpu) } } -static void i386_tr_disas_log(const DisasContextBase *dcbase, - CPUState *cpu, FILE *logfile) -{ - DisasContext *dc = container_of(dcbase, DisasContext, base); - - fprintf(logfile, "IN: %s\n", lookup_symbol(dc->base.pc_first)); - target_disas(logfile, cpu, dc->base.pc_first, dc->base.tb->size); -} - static const TranslatorOps i386_tr_ops = { .init_disas_context = i386_tr_init_disas_context, .tb_start = i386_tr_tb_start, .insn_start = i386_tr_insn_start, .translate_insn = i386_tr_translate_insn, .tb_stop = i386_tr_tb_stop, - .disas_log = i386_tr_disas_log, }; /* generate intermediate code for basic block 'tb'. */ diff --git a/target/loongarch/tcg/translate.c b/target/loongarch/tcg/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/loongarch/tcg/translate.c +++ b/target/loongarch/tcg/translate.c @@ -XXX,XX +XXX,XX @@ static void loongarch_tr_tb_stop(DisasContextBase *dcbase, CPUState *cs) } } -static void loongarch_tr_disas_log(const DisasContextBase *dcbase, - CPUState *cpu, FILE *logfile) -{ - qemu_log("IN: %s\n", lookup_symbol(dcbase->pc_first)); - target_disas(logfile, cpu, dcbase->pc_first, dcbase->tb->size); -} - static const TranslatorOps loongarch_tr_ops = { .init_disas_context = loongarch_tr_init_disas_context, .tb_start = loongarch_tr_tb_start, .insn_start = loongarch_tr_insn_start, .translate_insn = loongarch_tr_translate_insn, .tb_stop = loongarch_tr_tb_stop, - .disas_log = loongarch_tr_disas_log, }; void gen_intermediate_code(CPUState *cs, TranslationBlock *tb, int *max_insns, diff --git a/target/m68k/translate.c b/target/m68k/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/m68k/translate.c +++ b/target/m68k/translate.c @@ -XXX,XX +XXX,XX @@ #include "qemu/osdep.h" #include "cpu.h" -#include "disas/disas.h" #include "exec/exec-all.h" #include "tcg/tcg-op.h" #include "qemu/log.h" @@ -XXX,XX +XXX,XX @@ static void m68k_tr_tb_stop(DisasContextBase *dcbase, CPUState *cpu) } } -static void m68k_tr_disas_log(const DisasContextBase *dcbase, - CPUState *cpu, FILE *logfile) -{ - fprintf(logfile, "IN: %s\n", lookup_symbol(dcbase->pc_first)); - target_disas(logfile, cpu, dcbase->pc_first, dcbase->tb->size); -} - static const TranslatorOps m68k_tr_ops = { .init_disas_context = m68k_tr_init_disas_context, .tb_start = m68k_tr_tb_start, .insn_start = m68k_tr_insn_start, .translate_insn = m68k_tr_translate_insn, .tb_stop = m68k_tr_tb_stop, - .disas_log = m68k_tr_disas_log, }; void gen_intermediate_code(CPUState *cpu, TranslationBlock *tb, int *max_insns, diff --git a/target/microblaze/translate.c b/target/microblaze/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/microblaze/translate.c +++ b/target/microblaze/translate.c @@ -XXX,XX +XXX,XX @@ #include "qemu/osdep.h" #include "cpu.h" -#include "disas/disas.h" #include "exec/exec-all.h" #include "exec/cpu_ldst.h" #include "tcg/tcg-op.h" @@ -XXX,XX +XXX,XX @@ static void mb_tr_tb_stop(DisasContextBase *dcb, CPUState *cs) } } -static void mb_tr_disas_log(const DisasContextBase *dcb, - CPUState *cs, FILE *logfile) -{ - fprintf(logfile, "IN: %s\n", lookup_symbol(dcb->pc_first)); - target_disas(logfile, cs, dcb->pc_first, dcb->tb->size); -} - static const TranslatorOps mb_tr_ops = { .init_disas_context = mb_tr_init_disas_context, .tb_start = mb_tr_tb_start, .insn_start = mb_tr_insn_start, .translate_insn = mb_tr_translate_insn, .tb_stop = mb_tr_tb_stop, - .disas_log = mb_tr_disas_log, }; void gen_intermediate_code(CPUState *cpu, TranslationBlock *tb, int *max_insns, diff --git a/target/mips/tcg/translate.c b/target/mips/tcg/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/mips/tcg/translate.c +++ b/target/mips/tcg/translate.c @@ -XXX,XX +XXX,XX @@ #include "exec/translation-block.h" #include "semihosting/semihost.h" #include "trace.h" -#include "disas/disas.h" #include "fpu_helper.h" #define HELPER_H "helper.h" @@ -XXX,XX +XXX,XX @@ static void mips_tr_tb_stop(DisasContextBase *dcbase, CPUState *cs) } } -static void mips_tr_disas_log(const DisasContextBase *dcbase, - CPUState *cs, FILE *logfile) -{ - fprintf(logfile, "IN: %s\n", lookup_symbol(dcbase->pc_first)); - target_disas(logfile, cs, dcbase->pc_first, dcbase->tb->size); -} - static const TranslatorOps mips_tr_ops = { .init_disas_context = mips_tr_init_disas_context, .tb_start = mips_tr_tb_start, .insn_start = mips_tr_insn_start, .translate_insn = mips_tr_translate_insn, .tb_stop = mips_tr_tb_stop, - .disas_log = mips_tr_disas_log, }; void gen_intermediate_code(CPUState *cs, TranslationBlock *tb, int *max_insns, diff --git a/target/openrisc/translate.c b/target/openrisc/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/openrisc/translate.c +++ b/target/openrisc/translate.c @@ -XXX,XX +XXX,XX @@ #include "qemu/osdep.h" #include "cpu.h" #include "exec/exec-all.h" -#include "disas/disas.h" #include "tcg/tcg-op.h" #include "qemu/log.h" #include "qemu/bitops.h" @@ -XXX,XX +XXX,XX @@ static void openrisc_tr_tb_stop(DisasContextBase *dcbase, CPUState *cs) } } -static void openrisc_tr_disas_log(const DisasContextBase *dcbase, - CPUState *cs, FILE *logfile) -{ - DisasContext *s = container_of(dcbase, DisasContext, base); - - fprintf(logfile, "IN: %s\n", lookup_symbol(s->base.pc_first)); - target_disas(logfile, cs, s->base.pc_first, s->base.tb->size); -} - static const TranslatorOps openrisc_tr_ops = { .init_disas_context = openrisc_tr_init_disas_context, .tb_start = openrisc_tr_tb_start, .insn_start = openrisc_tr_insn_start, .translate_insn = openrisc_tr_translate_insn, .tb_stop = openrisc_tr_tb_stop, - .disas_log = openrisc_tr_disas_log, }; void gen_intermediate_code(CPUState *cs, TranslationBlock *tb, int *max_insns, diff --git a/target/ppc/translate.c b/target/ppc/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/ppc/translate.c +++ b/target/ppc/translate.c @@ -XXX,XX +XXX,XX @@ #include "qemu/osdep.h" #include "cpu.h" #include "internal.h" -#include "disas/disas.h" #include "exec/exec-all.h" #include "tcg/tcg-op.h" #include "tcg/tcg-op-gvec.h" @@ -XXX,XX +XXX,XX @@ static void ppc_tr_tb_stop(DisasContextBase *dcbase, CPUState *cs) } } -static void ppc_tr_disas_log(const DisasContextBase *dcbase, - CPUState *cs, FILE *logfile) -{ - fprintf(logfile, "IN: %s\n", lookup_symbol(dcbase->pc_first)); - target_disas(logfile, cs, dcbase->pc_first, dcbase->tb->size); -} - static const TranslatorOps ppc_tr_ops = { .init_disas_context = ppc_tr_init_disas_context, .tb_start = ppc_tr_tb_start, .insn_start = ppc_tr_insn_start, .translate_insn = ppc_tr_translate_insn, .tb_stop = ppc_tr_tb_stop, - .disas_log = ppc_tr_disas_log, }; void gen_intermediate_code(CPUState *cs, TranslationBlock *tb, int *max_insns, diff --git a/target/riscv/translate.c b/target/riscv/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/riscv/translate.c +++ b/target/riscv/translate.c @@ -XXX,XX +XXX,XX @@ #include "qemu/log.h" #include "cpu.h" #include "tcg/tcg-op.h" -#include "disas/disas.h" #include "exec/cpu_ldst.h" #include "exec/exec-all.h" #include "exec/helper-proto.h" @@ -XXX,XX +XXX,XX @@ static void riscv_tr_tb_stop(DisasContextBase *dcbase, CPUState *cpu) } } -static void riscv_tr_disas_log(const DisasContextBase *dcbase, - CPUState *cpu, FILE *logfile) -{ -#ifndef CONFIG_USER_ONLY - RISCVCPU *rvcpu = RISCV_CPU(cpu); - CPURISCVState *env = &rvcpu->env; -#endif - - fprintf(logfile, "IN: %s\n", lookup_symbol(dcbase->pc_first)); -#ifndef CONFIG_USER_ONLY - fprintf(logfile, "Priv: "TARGET_FMT_ld"; Virt: %d\n", - env->priv, env->virt_enabled); -#endif - target_disas(logfile, cpu, dcbase->pc_first, dcbase->tb->size); -} - static const TranslatorOps riscv_tr_ops = { .init_disas_context = riscv_tr_init_disas_context, .tb_start = riscv_tr_tb_start, .insn_start = riscv_tr_insn_start, .translate_insn = riscv_tr_translate_insn, .tb_stop = riscv_tr_tb_stop, - .disas_log = riscv_tr_disas_log, }; void gen_intermediate_code(CPUState *cs, TranslationBlock *tb, int *max_insns, diff --git a/target/rx/translate.c b/target/rx/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/rx/translate.c +++ b/target/rx/translate.c @@ -XXX,XX +XXX,XX @@ static void rx_tr_tb_stop(DisasContextBase *dcbase, CPUState *cs) } } -static void rx_tr_disas_log(const DisasContextBase *dcbase, - CPUState *cs, FILE *logfile) -{ - fprintf(logfile, "IN: %s\n", lookup_symbol(dcbase->pc_first)); - target_disas(logfile, cs, dcbase->pc_first, dcbase->tb->size); -} - static const TranslatorOps rx_tr_ops = { .init_disas_context = rx_tr_init_disas_context, .tb_start = rx_tr_tb_start, .insn_start = rx_tr_insn_start, .translate_insn = rx_tr_translate_insn, .tb_stop = rx_tr_tb_stop, - .disas_log = rx_tr_disas_log, }; void gen_intermediate_code(CPUState *cs, TranslationBlock *tb, int *max_insns, diff --git a/target/sh4/translate.c b/target/sh4/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/sh4/translate.c +++ b/target/sh4/translate.c @@ -XXX,XX +XXX,XX @@ #include "qemu/osdep.h" #include "cpu.h" -#include "disas/disas.h" #include "exec/exec-all.h" #include "tcg/tcg-op.h" #include "exec/helper-proto.h" @@ -XXX,XX +XXX,XX @@ static void sh4_tr_tb_stop(DisasContextBase *dcbase, CPUState *cs) } } -static void sh4_tr_disas_log(const DisasContextBase *dcbase, - CPUState *cs, FILE *logfile) -{ - fprintf(logfile, "IN: %s\n", lookup_symbol(dcbase->pc_first)); - target_disas(logfile, cs, dcbase->pc_first, dcbase->tb->size); -} - static const TranslatorOps sh4_tr_ops = { .init_disas_context = sh4_tr_init_disas_context, .tb_start = sh4_tr_tb_start, .insn_start = sh4_tr_insn_start, .translate_insn = sh4_tr_translate_insn, .tb_stop = sh4_tr_tb_stop, - .disas_log = sh4_tr_disas_log, }; void gen_intermediate_code(CPUState *cs, TranslationBlock *tb, int *max_insns, diff --git a/target/sparc/translate.c b/target/sparc/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/sparc/translate.c +++ b/target/sparc/translate.c @@ -XXX,XX +XXX,XX @@ #include "qemu/osdep.h" #include "cpu.h" -#include "disas/disas.h" #include "exec/helper-proto.h" #include "exec/exec-all.h" #include "tcg/tcg-op.h" @@ -XXX,XX +XXX,XX @@ static void sparc_tr_tb_stop(DisasContextBase *dcbase, CPUState *cs) } } -static void sparc_tr_disas_log(const DisasContextBase *dcbase, - CPUState *cpu, FILE *logfile) -{ - fprintf(logfile, "IN: %s\n", lookup_symbol(dcbase->pc_first)); - target_disas(logfile, cpu, dcbase->pc_first, dcbase->tb->size); -} - static const TranslatorOps sparc_tr_ops = { .init_disas_context = sparc_tr_init_disas_context, .tb_start = sparc_tr_tb_start, .insn_start = sparc_tr_insn_start, .translate_insn = sparc_tr_translate_insn, .tb_stop = sparc_tr_tb_stop, - .disas_log = sparc_tr_disas_log, }; void gen_intermediate_code(CPUState *cs, TranslationBlock *tb, int *max_insns, diff --git a/target/tricore/translate.c b/target/tricore/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/tricore/translate.c +++ b/target/tricore/translate.c @@ -XXX,XX +XXX,XX @@ #include "qemu/osdep.h" #include "cpu.h" -#include "disas/disas.h" #include "exec/exec-all.h" #include "tcg/tcg-op.h" #include "exec/cpu_ldst.h" @@ -XXX,XX +XXX,XX @@ static void tricore_tr_tb_stop(DisasContextBase *dcbase, CPUState *cpu) } } -static void tricore_tr_disas_log(const DisasContextBase *dcbase, - CPUState *cpu, FILE *logfile) -{ - fprintf(logfile, "IN: %s\n", lookup_symbol(dcbase->pc_first)); - target_disas(logfile, cpu, dcbase->pc_first, dcbase->tb->size); -} - static const TranslatorOps tricore_tr_ops = { .init_disas_context = tricore_tr_init_disas_context, .tb_start = tricore_tr_tb_start, .insn_start = tricore_tr_insn_start, .translate_insn = tricore_tr_translate_insn, .tb_stop = tricore_tr_tb_stop, - .disas_log = tricore_tr_disas_log, }; diff --git a/target/xtensa/translate.c b/target/xtensa/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/xtensa/translate.c +++ b/target/xtensa/translate.c @@ -XXX,XX +XXX,XX @@ #include "cpu.h" #include "exec/exec-all.h" -#include "disas/disas.h" #include "tcg/tcg-op.h" #include "qemu/log.h" #include "qemu/qemu-print.h" @@ -XXX,XX +XXX,XX @@ static void xtensa_tr_tb_stop(DisasContextBase *dcbase, CPUState *cpu) } } -static void xtensa_tr_disas_log(const DisasContextBase *dcbase, - CPUState *cpu, FILE *logfile) -{ - fprintf(logfile, "IN: %s\n", lookup_symbol(dcbase->pc_first)); - target_disas(logfile, cpu, dcbase->pc_first, dcbase->tb->size); -} - static const TranslatorOps xtensa_translator_ops = { .init_disas_context = xtensa_tr_init_disas_context, .tb_start = xtensa_tr_tb_start, .insn_start = xtensa_tr_insn_start, .translate_insn = xtensa_tr_translate_insn, .tb_stop = xtensa_tr_tb_stop, - .disas_log = xtensa_tr_disas_log, }; void gen_intermediate_code(CPUState *cpu, TranslationBlock *tb, int *max_insns, -- 2.34.1 We have eliminated most uses of this hook. Reduce further by allowing the hook to handle only the special cases, returning false for normal processing. Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- include/exec/translator.h | 2 +- accel/tcg/translator.c | 5 ++--- target/hppa/translate.c | 15 ++++++--------- target/s390x/tcg/translate.c | 8 +++----- 4 files changed, 12 insertions(+), 18 deletions(-) diff --git a/include/exec/translator.h b/include/exec/translator.h index XXXXXXX..XXXXXXX 100644 --- a/include/exec/translator.h +++ b/include/exec/translator.h @@ -XXX,XX +XXX,XX @@ typedef struct TranslatorOps { void (*insn_start)(DisasContextBase *db, CPUState *cpu); void (*translate_insn)(DisasContextBase *db, CPUState *cpu); void (*tb_stop)(DisasContextBase *db, CPUState *cpu); - void (*disas_log)(const DisasContextBase *db, CPUState *cpu, FILE *f); + bool (*disas_log)(const DisasContextBase *db, CPUState *cpu, FILE *f); } TranslatorOps; /** diff --git a/accel/tcg/translator.c b/accel/tcg/translator.c index XXXXXXX..XXXXXXX 100644 --- a/accel/tcg/translator.c +++ b/accel/tcg/translator.c @@ -XXX,XX +XXX,XX @@ void translator_loop(CPUState *cpu, TranslationBlock *tb, int *max_insns, if (logfile) { fprintf(logfile, "----------------\n"); - if (ops->disas_log) { - ops->disas_log(db, cpu, logfile); - } else { + if (!ops->disas_log || + !ops->disas_log(db, cpu, logfile)) { fprintf(logfile, "IN: %s\n", lookup_symbol(db->pc_first)); target_disas(logfile, cpu, db->pc_first, db->tb->size); } diff --git a/target/hppa/translate.c b/target/hppa/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/hppa/translate.c +++ b/target/hppa/translate.c @@ -XXX,XX +XXX,XX @@ #include "qemu/osdep.h" #include "cpu.h" -#include "disas/disas.h" #include "qemu/host-utils.h" #include "exec/exec-all.h" #include "exec/page-protection.h" @@ -XXX,XX +XXX,XX @@ static void hppa_tr_tb_stop(DisasContextBase *dcbase, CPUState *cs) } #ifdef CONFIG_USER_ONLY -static void hppa_tr_disas_log(const DisasContextBase *dcbase, +static bool hppa_tr_disas_log(const DisasContextBase *dcbase, CPUState *cs, FILE *logfile) { target_ulong pc = dcbase->pc_first; @@ -XXX,XX +XXX,XX @@ static void hppa_tr_disas_log(const DisasContextBase *dcbase, switch (pc) { case 0x00: fprintf(logfile, "IN:\n0x00000000: (null)\n"); - return; + return true; case 0xb0: fprintf(logfile, "IN:\n0x000000b0: light-weight-syscall\n"); - return; + return true; case 0xe0: fprintf(logfile, "IN:\n0x000000e0: set-thread-pointer-syscall\n"); - return; + return true; case 0x100: fprintf(logfile, "IN:\n0x00000100: syscall\n"); - return; + return true; } - - fprintf(logfile, "IN: %s\n", lookup_symbol(pc)); - target_disas(logfile, cs, pc, dcbase->tb->size); + return false; } #endif diff --git a/target/s390x/tcg/translate.c b/target/s390x/tcg/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/s390x/tcg/translate.c +++ b/target/s390x/tcg/translate.c @@ -XXX,XX +XXX,XX @@ #include "qemu/osdep.h" #include "cpu.h" #include "s390x-internal.h" -#include "disas/disas.h" #include "exec/exec-all.h" #include "tcg/tcg-op.h" #include "tcg/tcg-op-gvec.h" @@ -XXX,XX +XXX,XX @@ static void s390x_tr_tb_stop(DisasContextBase *dcbase, CPUState *cs) } } -static void s390x_tr_disas_log(const DisasContextBase *dcbase, +static bool s390x_tr_disas_log(const DisasContextBase *dcbase, CPUState *cs, FILE *logfile) { DisasContext *dc = container_of(dcbase, DisasContext, base); @@ -XXX,XX +XXX,XX @@ static void s390x_tr_disas_log(const DisasContextBase *dcbase, if (unlikely(dc->ex_value)) { /* ??? Unfortunately target_disas can't use host memory. */ fprintf(logfile, "IN: EXECUTE %016" PRIx64, dc->ex_value); - } else { - fprintf(logfile, "IN: %s\n", lookup_symbol(dc->base.pc_first)); - target_disas(logfile, cs, dc->base.pc_first, dc->base.tb->size); + return true; } + return false; } static const TranslatorOps s390x_tr_ops = { -- 2.34.1 The routines in disas-common.c are also used from disas-mon.c. Otherwise the rest of disassembly is only used from tcg. While we're at it, put host and target code into separate files. Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- disas/disas-internal.h | 4 + include/disas/disas.h | 4 + disas/disas-common.c | 118 ++++++++++++++ disas/disas-host.c | 129 ++++++++++++++++ disas/disas-target.c | 84 ++++++++++ disas/disas.c | 338 ----------------------------------------- disas/objdump.c | 37 +++++ disas/meson.build | 8 +- 8 files changed, 382 insertions(+), 340 deletions(-) create mode 100644 disas/disas-common.c create mode 100644 disas/disas-host.c create mode 100644 disas/disas-target.c delete mode 100644 disas/disas.c create mode 100644 disas/objdump.c diff --git a/disas/disas-internal.h b/disas/disas-internal.h index XXXXXXX..XXXXXXX 100644 --- a/disas/disas-internal.h +++ b/disas/disas-internal.h @@ -XXX,XX +XXX,XX @@ typedef struct CPUDebug { CPUState *cpu; } CPUDebug; +void disas_initialize_debug(CPUDebug *s); void disas_initialize_debug_target(CPUDebug *s, CPUState *cpu); int disas_gstring_printf(FILE *stream, const char *fmt, ...) G_GNUC_PRINTF(2, 3); +int print_insn_od_host(bfd_vma pc, disassemble_info *info); +int print_insn_od_target(bfd_vma pc, disassemble_info *info); + #endif diff --git a/include/disas/disas.h b/include/disas/disas.h index XXXXXXX..XXXXXXX 100644 --- a/include/disas/disas.h +++ b/include/disas/disas.h @@ -XXX,XX +XXX,XX @@ #define QEMU_DISAS_H /* Disassemble this for me please... (debugging). */ +#ifdef CONFIG_TCG void disas(FILE *out, const void *code, size_t size); void target_disas(FILE *out, CPUState *cpu, uint64_t code, size_t size); +#endif void monitor_disas(Monitor *mon, CPUState *cpu, uint64_t pc, int nb_insn, bool is_physical); +#ifdef CONFIG_PLUGIN char *plugin_disas(CPUState *cpu, uint64_t addr, size_t size); +#endif /* Look up symbol for debugging purpose. Returns "" if unknown. */ const char *lookup_symbol(uint64_t orig_addr); diff --git a/disas/disas-common.c b/disas/disas-common.c new file mode 100644 index XXXXXXX..XXXXXXX --- /dev/null +++ b/disas/disas-common.c @@ -XXX,XX +XXX,XX @@ +/* + * Common routines for disassembly. + * SPDX-License-Identifier: GPL-2.0-or-later + */ + +#include "qemu/osdep.h" +#include "disas/disas.h" +#include "disas/capstone.h" +#include "hw/core/cpu.h" +#include "exec/tswap.h" +#include "exec/memory.h" +#include "disas-internal.h" + + +/* Filled in by elfload.c. Simplistic, but will do for now. */ +struct syminfo *syminfos = NULL; + +/* + * Get LENGTH bytes from info's buffer, at target address memaddr. + * Transfer them to myaddr. + */ +static int target_read_memory(bfd_vma memaddr, bfd_byte *myaddr, int length, + struct disassemble_info *info) +{ + CPUDebug *s = container_of(info, CPUDebug, info); + int r = cpu_memory_rw_debug(s->cpu, memaddr, myaddr, length, 0); + return r ? EIO : 0; +} + +/* + * Print an error message. We can assume that this is in response to + * an error return from {host,target}_read_memory. + */ +static void perror_memory(int status, bfd_vma memaddr, + struct disassemble_info *info) +{ + if (status != EIO) { + /* Can't happen. */ + info->fprintf_func(info->stream, "Unknown error %d\n", status); + } else { + /* Address between memaddr and memaddr + len was out of bounds. */ + info->fprintf_func(info->stream, + "Address 0x%" PRIx64 " is out of bounds.\n", + memaddr); + } +} + +/* Print address in hex. */ +static void print_address(bfd_vma addr, struct disassemble_info *info) +{ + info->fprintf_func(info->stream, "0x%" PRIx64, addr); +} + +/* Stub prevents some fruitless earching in optabs disassemblers. */ +static int symbol_at_address(bfd_vma addr, struct disassemble_info *info) +{ + return 1; +} + +void disas_initialize_debug(CPUDebug *s) +{ + memset(s, 0, sizeof(*s)); + s->info.arch = bfd_arch_unknown; + s->info.cap_arch = -1; + s->info.cap_insn_unit = 4; + s->info.cap_insn_split = 4; + s->info.memory_error_func = perror_memory; + s->info.symbol_at_address_func = symbol_at_address; +} + +void disas_initialize_debug_target(CPUDebug *s, CPUState *cpu) +{ + disas_initialize_debug(s); + + s->cpu = cpu; + s->info.read_memory_func = target_read_memory; + s->info.print_address_func = print_address; + if (target_words_bigendian()) { + s->info.endian = BFD_ENDIAN_BIG; + } else { + s->info.endian = BFD_ENDIAN_LITTLE; + } + + CPUClass *cc = CPU_GET_CLASS(cpu); + if (cc->disas_set_info) { + cc->disas_set_info(cpu, &s->info); + } +} + +int disas_gstring_printf(FILE *stream, const char *fmt, ...) +{ + /* We abuse the FILE parameter to pass a GString. */ + GString *s = (GString *)stream; + int initial_len = s->len; + va_list va; + + va_start(va, fmt); + g_string_append_vprintf(s, fmt, va); + va_end(va); + + return s->len - initial_len; +} + +/* Look up symbol for debugging purpose. Returns "" if unknown. */ +const char *lookup_symbol(uint64_t orig_addr) +{ + const char *symbol = ""; + struct syminfo *s; + + for (s = syminfos; s; s = s->next) { + symbol = s->lookup_symbol(s, orig_addr); + if (symbol[0] != '\0') { + break; + } + } + + return symbol; +} diff --git a/disas/disas-host.c b/disas/disas-host.c new file mode 100644 index XXXXXXX..XXXXXXX --- /dev/null +++ b/disas/disas-host.c @@ -XXX,XX +XXX,XX @@ +/* + * Routines for host instruction disassembly. + * SPDX-License-Identifier: GPL-2.0-or-later + */ + +#include "qemu/osdep.h" +#include "disas/disas.h" +#include "disas/capstone.h" +#include "disas-internal.h" + + +/* + * Get LENGTH bytes from info's buffer, at host address memaddr. + * Transfer them to myaddr. + */ +static int host_read_memory(bfd_vma memaddr, bfd_byte *myaddr, int length, + struct disassemble_info *info) +{ + if (memaddr < info->buffer_vma + || memaddr + length > info->buffer_vma + info->buffer_length) { + /* Out of bounds. Use EIO because GDB uses it. */ + return EIO; + } + memcpy (myaddr, info->buffer + (memaddr - info->buffer_vma), length); + return 0; +} + +/* Print address in hex, truncated to the width of a host virtual address. */ +static void host_print_address(bfd_vma addr, struct disassemble_info *info) +{ + info->fprintf_func(info->stream, "0x%" PRIxPTR, (uintptr_t)addr); +} + +static void initialize_debug_host(CPUDebug *s) +{ + disas_initialize_debug(s); + + s->info.read_memory_func = host_read_memory; + s->info.print_address_func = host_print_address; +#if HOST_BIG_ENDIAN + s->info.endian = BFD_ENDIAN_BIG; +#else + s->info.endian = BFD_ENDIAN_LITTLE; +#endif +#if defined(CONFIG_TCG_INTERPRETER) + s->info.print_insn = print_insn_tci; +#elif defined(__i386__) + s->info.mach = bfd_mach_i386_i386; + s->info.cap_arch = CS_ARCH_X86; + s->info.cap_mode = CS_MODE_32; + s->info.cap_insn_unit = 1; + s->info.cap_insn_split = 8; +#elif defined(__x86_64__) + s->info.mach = bfd_mach_x86_64; + s->info.cap_arch = CS_ARCH_X86; + s->info.cap_mode = CS_MODE_64; + s->info.cap_insn_unit = 1; + s->info.cap_insn_split = 8; +#elif defined(_ARCH_PPC) + s->info.cap_arch = CS_ARCH_PPC; +# ifdef _ARCH_PPC64 + s->info.cap_mode = CS_MODE_64; +# endif +#elif defined(__riscv) +#if defined(_ILP32) || (__riscv_xlen == 32) + s->info.print_insn = print_insn_riscv32; +#elif defined(_LP64) + s->info.print_insn = print_insn_riscv64; +#else +#error unsupported RISC-V ABI +#endif +#elif defined(__aarch64__) + s->info.cap_arch = CS_ARCH_ARM64; +#elif defined(__alpha__) + s->info.print_insn = print_insn_alpha; +#elif defined(__sparc__) + s->info.print_insn = print_insn_sparc; + s->info.mach = bfd_mach_sparc_v9b; +#elif defined(__arm__) + /* TCG only generates code for arm mode. */ + s->info.cap_arch = CS_ARCH_ARM; +#elif defined(__MIPSEB__) + s->info.print_insn = print_insn_big_mips; +#elif defined(__MIPSEL__) + s->info.print_insn = print_insn_little_mips; +#elif defined(__m68k__) + s->info.print_insn = print_insn_m68k; +#elif defined(__s390__) + s->info.cap_arch = CS_ARCH_SYSZ; + s->info.cap_insn_unit = 2; + s->info.cap_insn_split = 6; +#elif defined(__hppa__) + s->info.print_insn = print_insn_hppa; +#elif defined(__loongarch__) + s->info.print_insn = print_insn_loongarch; +#endif +} + +/* Disassemble this for me please... (debugging). */ +void disas(FILE *out, const void *code, size_t size) +{ + uintptr_t pc; + int count; + CPUDebug s; + + initialize_debug_host(&s); + s.info.fprintf_func = fprintf; + s.info.stream = out; + s.info.buffer = code; + s.info.buffer_vma = (uintptr_t)code; + s.info.buffer_length = size; + s.info.show_opcodes = true; + + if (s.info.cap_arch >= 0 && cap_disas_host(&s.info, code, size)) { + return; + } + + if (s.info.print_insn == NULL) { + s.info.print_insn = print_insn_od_host; + } + for (pc = (uintptr_t)code; size > 0; pc += count, size -= count) { + fprintf(out, "0x%08" PRIxPTR ": ", pc); + count = s.info.print_insn(pc, &s.info); + fprintf(out, "\n"); + if (count < 0) { + break; + } + } +} diff --git a/disas/disas-target.c b/disas/disas-target.c new file mode 100644 index XXXXXXX..XXXXXXX --- /dev/null +++ b/disas/disas-target.c @@ -XXX,XX +XXX,XX @@ +/* + * Routines for target instruction disassembly. + * SPDX-License-Identifier: GPL-2.0-or-later + */ + +#include "qemu/osdep.h" +#include "disas/disas.h" +#include "disas/capstone.h" +#include "disas-internal.h" + + +void target_disas(FILE *out, CPUState *cpu, uint64_t code, size_t size) +{ + uint64_t pc; + int count; + CPUDebug s; + + disas_initialize_debug_target(&s, cpu); + s.info.fprintf_func = fprintf; + s.info.stream = out; + s.info.buffer_vma = code; + s.info.buffer_length = size; + s.info.show_opcodes = true; + + if (s.info.cap_arch >= 0 && cap_disas_target(&s.info, code, size)) { + return; + } + + if (s.info.print_insn == NULL) { + s.info.print_insn = print_insn_od_target; + } + + for (pc = code; size > 0; pc += count, size -= count) { + fprintf(out, "0x%08" PRIx64 ": ", pc); + count = s.info.print_insn(pc, &s.info); + fprintf(out, "\n"); + if (count < 0) { + break; + } + if (size < count) { + fprintf(out, + "Disassembler disagrees with translator over instruction " + "decoding\n" + "Please report this to qemu-devel@nongnu.org\n"); + break; + } + } +} + +#ifdef CONFIG_PLUGIN +static void plugin_print_address(bfd_vma addr, struct disassemble_info *info) +{ + /* does nothing */ +} + +/* + * We should only be dissembling one instruction at a time here. If + * there is left over it usually indicates the front end has read more + * bytes than it needed. + */ +char *plugin_disas(CPUState *cpu, uint64_t addr, size_t size) +{ + CPUDebug s; + GString *ds = g_string_new(NULL); + + disas_initialize_debug_target(&s, cpu); + s.info.fprintf_func = disas_gstring_printf; + s.info.stream = (FILE *)ds; /* abuse this slot */ + s.info.buffer_vma = addr; + s.info.buffer_length = size; + s.info.print_address_func = plugin_print_address; + + if (s.info.cap_arch >= 0 && cap_disas_plugin(&s.info, addr, size)) { + ; /* done */ + } else if (s.info.print_insn) { + s.info.print_insn(addr, &s.info); + } else { + ; /* cannot disassemble -- return empty string */ + } + + /* Return the buffer, freeing the GString container. */ + return g_string_free(ds, false); +} +#endif /* CONFIG_PLUGIN */ diff --git a/disas/disas.c b/disas/disas.c deleted file mode 100644 index XXXXXXX..XXXXXXX --- a/disas/disas.c +++ /dev/null @@ -XXX,XX +XXX,XX @@ -/* General "disassemble this chunk" code. Used for debugging. */ -#include "qemu/osdep.h" -#include "disas/disas-internal.h" -#include "elf.h" -#include "qemu/qemu-print.h" -#include "disas/disas.h" -#include "disas/capstone.h" -#include "hw/core/cpu.h" -#include "exec/tswap.h" -#include "exec/memory.h" - -/* Filled in by elfload.c. Simplistic, but will do for now. */ -struct syminfo *syminfos = NULL; - -/* - * Get LENGTH bytes from info's buffer, at host address memaddr. - * Transfer them to myaddr. - */ -static int host_read_memory(bfd_vma memaddr, bfd_byte *myaddr, int length, - struct disassemble_info *info) -{ - if (memaddr < info->buffer_vma - || memaddr + length > info->buffer_vma + info->buffer_length) { - /* Out of bounds. Use EIO because GDB uses it. */ - return EIO; - } - memcpy (myaddr, info->buffer + (memaddr - info->buffer_vma), length); - return 0; -} - -/* - * Get LENGTH bytes from info's buffer, at target address memaddr. - * Transfer them to myaddr. - */ -static int target_read_memory(bfd_vma memaddr, bfd_byte *myaddr, int length, - struct disassemble_info *info) -{ - CPUDebug *s = container_of(info, CPUDebug, info); - int r = cpu_memory_rw_debug(s->cpu, memaddr, myaddr, length, 0); - return r ? EIO : 0; -} - -/* - * Print an error message. We can assume that this is in response to - * an error return from {host,target}_read_memory. - */ -static void perror_memory(int status, bfd_vma memaddr, - struct disassemble_info *info) -{ - if (status != EIO) { - /* Can't happen. */ - info->fprintf_func(info->stream, "Unknown error %d\n", status); - } else { - /* Address between memaddr and memaddr + len was out of bounds. */ - info->fprintf_func(info->stream, - "Address 0x%" PRIx64 " is out of bounds.\n", - memaddr); - } -} - -/* Print address in hex. */ -static void print_address(bfd_vma addr, struct disassemble_info *info) -{ - info->fprintf_func(info->stream, "0x%" PRIx64, addr); -} - -/* Print address in hex, truncated to the width of a host virtual address. */ -static void host_print_address(bfd_vma addr, struct disassemble_info *info) -{ - print_address((uintptr_t)addr, info); -} - -/* Stub prevents some fruitless earching in optabs disassemblers. */ -static int symbol_at_address(bfd_vma addr, struct disassemble_info *info) -{ - return 1; -} - -static int print_insn_objdump(bfd_vma pc, disassemble_info *info, - const char *prefix) -{ - int i, n = info->buffer_length; - g_autofree uint8_t *buf = g_malloc(n); - - if (info->read_memory_func(pc, buf, n, info) == 0) { - for (i = 0; i < n; ++i) { - if (i % 32 == 0) { - info->fprintf_func(info->stream, "\n%s: ", prefix); - } - info->fprintf_func(info->stream, "%02x", buf[i]); - } - } else { - info->fprintf_func(info->stream, "unable to read memory"); - } - return n; -} - -static int print_insn_od_host(bfd_vma pc, disassemble_info *info) -{ - return print_insn_objdump(pc, info, "OBJD-H"); -} - -static int print_insn_od_target(bfd_vma pc, disassemble_info *info) -{ - return print_insn_objdump(pc, info, "OBJD-T"); -} - -static void initialize_debug(CPUDebug *s) -{ - memset(s, 0, sizeof(*s)); - s->info.arch = bfd_arch_unknown; - s->info.cap_arch = -1; - s->info.cap_insn_unit = 4; - s->info.cap_insn_split = 4; - s->info.memory_error_func = perror_memory; - s->info.symbol_at_address_func = symbol_at_address; -} - -void disas_initialize_debug_target(CPUDebug *s, CPUState *cpu) -{ - initialize_debug(s); - - s->cpu = cpu; - s->info.read_memory_func = target_read_memory; - s->info.print_address_func = print_address; - if (target_words_bigendian()) { - s->info.endian = BFD_ENDIAN_BIG; - } else { - s->info.endian = BFD_ENDIAN_LITTLE; - } - - CPUClass *cc = CPU_GET_CLASS(cpu); - if (cc->disas_set_info) { - cc->disas_set_info(cpu, &s->info); - } -} - -static void initialize_debug_host(CPUDebug *s) -{ - initialize_debug(s); - - s->info.read_memory_func = host_read_memory; - s->info.print_address_func = host_print_address; -#if HOST_BIG_ENDIAN - s->info.endian = BFD_ENDIAN_BIG; -#else - s->info.endian = BFD_ENDIAN_LITTLE; -#endif -#if defined(CONFIG_TCG_INTERPRETER) - s->info.print_insn = print_insn_tci; -#elif defined(__i386__) - s->info.mach = bfd_mach_i386_i386; - s->info.cap_arch = CS_ARCH_X86; - s->info.cap_mode = CS_MODE_32; - s->info.cap_insn_unit = 1; - s->info.cap_insn_split = 8; -#elif defined(__x86_64__) - s->info.mach = bfd_mach_x86_64; - s->info.cap_arch = CS_ARCH_X86; - s->info.cap_mode = CS_MODE_64; - s->info.cap_insn_unit = 1; - s->info.cap_insn_split = 8; -#elif defined(_ARCH_PPC) - s->info.cap_arch = CS_ARCH_PPC; -# ifdef _ARCH_PPC64 - s->info.cap_mode = CS_MODE_64; -# endif -#elif defined(__riscv) -#if defined(_ILP32) || (__riscv_xlen == 32) - s->info.print_insn = print_insn_riscv32; -#elif defined(_LP64) - s->info.print_insn = print_insn_riscv64; -#else -#error unsupported RISC-V ABI -#endif -#elif defined(__aarch64__) - s->info.cap_arch = CS_ARCH_ARM64; -#elif defined(__alpha__) - s->info.print_insn = print_insn_alpha; -#elif defined(__sparc__) - s->info.print_insn = print_insn_sparc; - s->info.mach = bfd_mach_sparc_v9b; -#elif defined(__arm__) - /* TCG only generates code for arm mode. */ - s->info.cap_arch = CS_ARCH_ARM; -#elif defined(__MIPSEB__) - s->info.print_insn = print_insn_big_mips; -#elif defined(__MIPSEL__) - s->info.print_insn = print_insn_little_mips; -#elif defined(__m68k__) - s->info.print_insn = print_insn_m68k; -#elif defined(__s390__) - s->info.cap_arch = CS_ARCH_SYSZ; - s->info.cap_insn_unit = 2; - s->info.cap_insn_split = 6; -#elif defined(__hppa__) - s->info.print_insn = print_insn_hppa; -#elif defined(__loongarch__) - s->info.print_insn = print_insn_loongarch; -#endif -} - -/* Disassemble this for me please... (debugging). */ -void target_disas(FILE *out, CPUState *cpu, uint64_t code, size_t size) -{ - uint64_t pc; - int count; - CPUDebug s; - - disas_initialize_debug_target(&s, cpu); - s.info.fprintf_func = fprintf; - s.info.stream = out; - s.info.buffer_vma = code; - s.info.buffer_length = size; - s.info.show_opcodes = true; - - if (s.info.cap_arch >= 0 && cap_disas_target(&s.info, code, size)) { - return; - } - - if (s.info.print_insn == NULL) { - s.info.print_insn = print_insn_od_target; - } - - for (pc = code; size > 0; pc += count, size -= count) { - fprintf(out, "0x%08" PRIx64 ": ", pc); - count = s.info.print_insn(pc, &s.info); - fprintf(out, "\n"); - if (count < 0) { - break; - } - if (size < count) { - fprintf(out, - "Disassembler disagrees with translator over instruction " - "decoding\n" - "Please report this to qemu-devel@nongnu.org\n"); - break; - } - } -} - -int disas_gstring_printf(FILE *stream, const char *fmt, ...) -{ - /* We abuse the FILE parameter to pass a GString. */ - GString *s = (GString *)stream; - int initial_len = s->len; - va_list va; - - va_start(va, fmt); - g_string_append_vprintf(s, fmt, va); - va_end(va); - - return s->len - initial_len; -} - -static void plugin_print_address(bfd_vma addr, struct disassemble_info *info) -{ - /* does nothing */ -} - - -/* - * We should only be dissembling one instruction at a time here. If - * there is left over it usually indicates the front end has read more - * bytes than it needed. - */ -char *plugin_disas(CPUState *cpu, uint64_t addr, size_t size) -{ - CPUDebug s; - GString *ds = g_string_new(NULL); - - disas_initialize_debug_target(&s, cpu); - s.info.fprintf_func = disas_gstring_printf; - s.info.stream = (FILE *)ds; /* abuse this slot */ - s.info.buffer_vma = addr; - s.info.buffer_length = size; - s.info.print_address_func = plugin_print_address; - - if (s.info.cap_arch >= 0 && cap_disas_plugin(&s.info, addr, size)) { - ; /* done */ - } else if (s.info.print_insn) { - s.info.print_insn(addr, &s.info); - } else { - ; /* cannot disassemble -- return empty string */ - } - - /* Return the buffer, freeing the GString container. */ - return g_string_free(ds, false); -} - -/* Disassemble this for me please... (debugging). */ -void disas(FILE *out, const void *code, size_t size) -{ - uintptr_t pc; - int count; - CPUDebug s; - - initialize_debug_host(&s); - s.info.fprintf_func = fprintf; - s.info.stream = out; - s.info.buffer = code; - s.info.buffer_vma = (uintptr_t)code; - s.info.buffer_length = size; - s.info.show_opcodes = true; - - if (s.info.cap_arch >= 0 && cap_disas_host(&s.info, code, size)) { - return; - } - - if (s.info.print_insn == NULL) { - s.info.print_insn = print_insn_od_host; - } - for (pc = (uintptr_t)code; size > 0; pc += count, size -= count) { - fprintf(out, "0x%08" PRIxPTR ": ", pc); - count = s.info.print_insn(pc, &s.info); - fprintf(out, "\n"); - if (count < 0) { - break; - } - } - -} - -/* Look up symbol for debugging purpose. Returns "" if unknown. */ -const char *lookup_symbol(uint64_t orig_addr) -{ - const char *symbol = ""; - struct syminfo *s; - - for (s = syminfos; s; s = s->next) { - symbol = s->lookup_symbol(s, orig_addr); - if (symbol[0] != '\0') { - break; - } - } - - return symbol; -} diff --git a/disas/objdump.c b/disas/objdump.c new file mode 100644 index XXXXXXX..XXXXXXX --- /dev/null +++ b/disas/objdump.c @@ -XXX,XX +XXX,XX @@ +/* + * Dump disassembly as text, for processing by scripts/disas-objdump.pl. + * SPDX-License-Identifier: GPL-2.0-or-later + */ + +#include "qemu/osdep.h" +#include "disas-internal.h" + + +static int print_insn_objdump(bfd_vma pc, disassemble_info *info, + const char *prefix) +{ + int i, n = info->buffer_length; + g_autofree uint8_t *buf = g_malloc(n); + + if (info->read_memory_func(pc, buf, n, info) == 0) { + for (i = 0; i < n; ++i) { + if (i % 32 == 0) { + info->fprintf_func(info->stream, "\n%s: ", prefix); + } + info->fprintf_func(info->stream, "%02x", buf[i]); + } + } else { + info->fprintf_func(info->stream, "unable to read memory"); + } + return n; +} + +int print_insn_od_host(bfd_vma pc, disassemble_info *info) +{ + return print_insn_objdump(pc, info, "OBJD-H"); +} + +int print_insn_od_target(bfd_vma pc, disassemble_info *info) +{ + return print_insn_objdump(pc, info, "OBJD-T"); +} diff --git a/disas/meson.build b/disas/meson.build index XXXXXXX..XXXXXXX 100644 --- a/disas/meson.build +++ b/disas/meson.build @@ -XXX,XX +XXX,XX @@ common_ss.add(when: 'CONFIG_SH4_DIS', if_true: files('sh4.c')) common_ss.add(when: 'CONFIG_SPARC_DIS', if_true: files('sparc.c')) common_ss.add(when: 'CONFIG_XTENSA_DIS', if_true: files('xtensa.c')) common_ss.add(when: capstone, if_true: [files('capstone.c'), capstone]) -common_ss.add(files('disas.c')) - +common_ss.add(when: 'CONFIG_TCG', if_true: files( + 'disas-host.c', + 'disas-target.c', + 'objdump.c' +)) +common_ss.add(files('disas-common.c')) system_ss.add(files('disas-mon.c')) specific_ss.add(capstone) -- 2.34.1 Read from already translated pages, or saved mmio data. Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- include/disas/disas.h | 5 +++-- include/exec/translator.h | 4 ++-- include/qemu/typedefs.h | 1 + accel/tcg/translator.c | 2 +- disas/disas-common.c | 14 -------------- disas/disas-mon.c | 15 +++++++++++++++ disas/disas-target.c | 19 +++++++++++++++++-- plugins/api.c | 4 ++-- 8 files changed, 41 insertions(+), 23 deletions(-) diff --git a/include/disas/disas.h b/include/disas/disas.h index XXXXXXX..XXXXXXX 100644 --- a/include/disas/disas.h +++ b/include/disas/disas.h @@ -XXX,XX +XXX,XX @@ /* Disassemble this for me please... (debugging). */ #ifdef CONFIG_TCG void disas(FILE *out, const void *code, size_t size); -void target_disas(FILE *out, CPUState *cpu, uint64_t code, size_t size); +void target_disas(FILE *out, CPUState *cpu, const DisasContextBase *db); #endif void monitor_disas(Monitor *mon, CPUState *cpu, uint64_t pc, int nb_insn, bool is_physical); #ifdef CONFIG_PLUGIN -char *plugin_disas(CPUState *cpu, uint64_t addr, size_t size); +char *plugin_disas(CPUState *cpu, const DisasContextBase *db, + uint64_t addr, size_t size); #endif /* Look up symbol for debugging purpose. Returns "" if unknown. */ diff --git a/include/exec/translator.h b/include/exec/translator.h index XXXXXXX..XXXXXXX 100644 --- a/include/exec/translator.h +++ b/include/exec/translator.h @@ -XXX,XX +XXX,XX @@ typedef enum DisasJumpType { * * Architecture-agnostic disassembly context. */ -typedef struct DisasContextBase { +struct DisasContextBase { TranslationBlock *tb; vaddr pc_first; vaddr pc_next; @@ -XXX,XX +XXX,XX @@ typedef struct DisasContextBase { int record_start; int record_len; uint8_t record[32]; -} DisasContextBase; +}; /** * TranslatorOps: diff --git a/include/qemu/typedefs.h b/include/qemu/typedefs.h index XXXXXXX..XXXXXXX 100644 --- a/include/qemu/typedefs.h +++ b/include/qemu/typedefs.h @@ -XXX,XX +XXX,XX @@ typedef struct CPUPluginState CPUPluginState; typedef struct CPUState CPUState; typedef struct DeviceState DeviceState; typedef struct DirtyBitmapSnapshot DirtyBitmapSnapshot; +typedef struct DisasContextBase DisasContextBase; typedef struct DisplayChangeListener DisplayChangeListener; typedef struct DriveInfo DriveInfo; typedef struct DumpState DumpState; diff --git a/accel/tcg/translator.c b/accel/tcg/translator.c index XXXXXXX..XXXXXXX 100644 --- a/accel/tcg/translator.c +++ b/accel/tcg/translator.c @@ -XXX,XX +XXX,XX @@ void translator_loop(CPUState *cpu, TranslationBlock *tb, int *max_insns, if (!ops->disas_log || !ops->disas_log(db, cpu, logfile)) { fprintf(logfile, "IN: %s\n", lookup_symbol(db->pc_first)); - target_disas(logfile, cpu, db->pc_first, db->tb->size); + target_disas(logfile, cpu, db); } fprintf(logfile, "\n"); qemu_log_unlock(logfile); diff --git a/disas/disas-common.c b/disas/disas-common.c index XXXXXXX..XXXXXXX 100644 --- a/disas/disas-common.c +++ b/disas/disas-common.c @@ -XXX,XX +XXX,XX @@ #include "disas/capstone.h" #include "hw/core/cpu.h" #include "exec/tswap.h" -#include "exec/memory.h" #include "disas-internal.h" /* Filled in by elfload.c. Simplistic, but will do for now. */ struct syminfo *syminfos = NULL; -/* - * Get LENGTH bytes from info's buffer, at target address memaddr. - * Transfer them to myaddr. - */ -static int target_read_memory(bfd_vma memaddr, bfd_byte *myaddr, int length, - struct disassemble_info *info) -{ - CPUDebug *s = container_of(info, CPUDebug, info); - int r = cpu_memory_rw_debug(s->cpu, memaddr, myaddr, length, 0); - return r ? EIO : 0; -} - /* * Print an error message. We can assume that this is in response to * an error return from {host,target}_read_memory. @@ -XXX,XX +XXX,XX @@ void disas_initialize_debug_target(CPUDebug *s, CPUState *cpu) disas_initialize_debug(s); s->cpu = cpu; - s->info.read_memory_func = target_read_memory; s->info.print_address_func = print_address; if (target_words_bigendian()) { s->info.endian = BFD_ENDIAN_BIG; diff --git a/disas/disas-mon.c b/disas/disas-mon.c index XXXXXXX..XXXXXXX 100644 --- a/disas/disas-mon.c +++ b/disas/disas-mon.c @@ -XXX,XX +XXX,XX @@ #include "hw/core/cpu.h" #include "monitor/monitor.h" +/* + * Get LENGTH bytes from info's buffer, at target address memaddr. + * Transfer them to myaddr. + */ +static int +virtual_read_memory(bfd_vma memaddr, bfd_byte *myaddr, int length, + struct disassemble_info *info) +{ + CPUDebug *s = container_of(info, CPUDebug, info); + int r = cpu_memory_rw_debug(s->cpu, memaddr, myaddr, length, 0); + return r ? EIO : 0; +} + static int physical_read_memory(bfd_vma memaddr, bfd_byte *myaddr, int length, struct disassemble_info *info) @@ -XXX,XX +XXX,XX @@ void monitor_disas(Monitor *mon, CPUState *cpu, uint64_t pc, if (is_physical) { s.info.read_memory_func = physical_read_memory; + } else { + s.info.read_memory_func = virtual_read_memory; } s.info.buffer_vma = pc; diff --git a/disas/disas-target.c b/disas/disas-target.c index XXXXXXX..XXXXXXX 100644 --- a/disas/disas-target.c +++ b/disas/disas-target.c @@ -XXX,XX +XXX,XX @@ #include "qemu/osdep.h" #include "disas/disas.h" #include "disas/capstone.h" +#include "exec/translator.h" #include "disas-internal.h" -void target_disas(FILE *out, CPUState *cpu, uint64_t code, size_t size) +static int translator_read_memory(bfd_vma memaddr, bfd_byte *myaddr, + int length, struct disassemble_info *info) { + const DisasContextBase *db = info->application_data; + return translator_st(db, myaddr, memaddr, length) ? 0 : EIO; +} + +void target_disas(FILE *out, CPUState *cpu, const struct DisasContextBase *db) +{ + uint64_t code = db->pc_first; + size_t size = translator_st_len(db); uint64_t pc; int count; CPUDebug s; disas_initialize_debug_target(&s, cpu); + s.info.read_memory_func = translator_read_memory; + s.info.application_data = (void *)db; s.info.fprintf_func = fprintf; s.info.stream = out; s.info.buffer_vma = code; @@ -XXX,XX +XXX,XX @@ static void plugin_print_address(bfd_vma addr, struct disassemble_info *info) * there is left over it usually indicates the front end has read more * bytes than it needed. */ -char *plugin_disas(CPUState *cpu, uint64_t addr, size_t size) +char *plugin_disas(CPUState *cpu, const DisasContextBase *db, + uint64_t addr, size_t size) { CPUDebug s; GString *ds = g_string_new(NULL); disas_initialize_debug_target(&s, cpu); + s.info.read_memory_func = translator_read_memory; + s.info.application_data = (void *)db; s.info.fprintf_func = disas_gstring_printf; s.info.stream = (FILE *)ds; /* abuse this slot */ s.info.buffer_vma = addr; diff --git a/plugins/api.c b/plugins/api.c index XXXXXXX..XXXXXXX 100644 --- a/plugins/api.c +++ b/plugins/api.c @@ -XXX,XX +XXX,XX @@ void *qemu_plugin_insn_haddr(const struct qemu_plugin_insn *insn) char *qemu_plugin_insn_disas(const struct qemu_plugin_insn *insn) { - CPUState *cpu = current_cpu; - return plugin_disas(cpu, insn->vaddr, insn->len); + return plugin_disas(tcg_ctx->cpu, tcg_ctx->plugin_db, + insn->vaddr, insn->len); } const char *qemu_plugin_insn_symbol(const struct qemu_plugin_insn *insn) -- 2.34.1 Replace translator_fake_ldb, which required multiple calls, with translator_fake_ld, which can take all data at once. Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- include/exec/translator.h | 8 ++++---- accel/tcg/translator.c | 5 ++--- target/s390x/tcg/translate.c | 8 ++++---- 3 files changed, 10 insertions(+), 11 deletions(-) diff --git a/include/exec/translator.h b/include/exec/translator.h index XXXXXXX..XXXXXXX 100644 --- a/include/exec/translator.h +++ b/include/exec/translator.h @@ -XXX,XX +XXX,XX @@ translator_ldq_swap(CPUArchState *env, DisasContextBase *db, } /** - * translator_fake_ldb - fake instruction load + * translator_fake_ld - fake instruction load * @db: Disassembly context - * @pc: program counter of instruction - * @insn8: byte of instruction + * @data: bytes of instruction + * @len: number of bytes * * This is a special case helper used where the instruction we are * about to translate comes from somewhere else (e.g. being * re-synthesised for s390x "ex"). It ensures we update other areas of * the translator with details of the executed instruction. */ -void translator_fake_ldb(DisasContextBase *db, vaddr pc, uint8_t insn8); +void translator_fake_ld(DisasContextBase *db, const void *data, size_t len); /** * translator_st diff --git a/accel/tcg/translator.c b/accel/tcg/translator.c index XXXXXXX..XXXXXXX 100644 --- a/accel/tcg/translator.c +++ b/accel/tcg/translator.c @@ -XXX,XX +XXX,XX @@ uint64_t translator_ldq(CPUArchState *env, DisasContextBase *db, vaddr pc) return tgt; } -void translator_fake_ldb(DisasContextBase *db, vaddr pc, uint8_t insn8) +void translator_fake_ld(DisasContextBase *db, const void *data, size_t len) { - assert(pc >= db->pc_first); db->fake_insn = true; - record_save(db, pc, &insn8, sizeof(insn8)); + record_save(db, db->pc_first, data, len); } diff --git a/target/s390x/tcg/translate.c b/target/s390x/tcg/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/s390x/tcg/translate.c +++ b/target/s390x/tcg/translate.c @@ -XXX,XX +XXX,XX @@ static const DisasInsn *extract_insn(CPUS390XState *env, DisasContext *s) const DisasInsn *info; if (unlikely(s->ex_value)) { + uint64_t be_insn; + /* Drop the EX data now, so that it's clear on exception paths. */ tcg_gen_st_i64(tcg_constant_i64(0), tcg_env, offsetof(CPUS390XState, ex_value)); @@ -XXX,XX +XXX,XX @@ static const DisasInsn *extract_insn(CPUS390XState *env, DisasContext *s) ilen = s->ex_value & 0xf; /* Register insn bytes with translator so plugins work. */ - for (int i = 0; i < ilen; i++) { - uint8_t byte = extract64(insn, 56 - (i * 8), 8); - translator_fake_ldb(&s->base, pc + i, byte); - } + be_insn = cpu_to_be64(insn); + translator_fake_ld(&s->base, &be_insn, ilen); op = insn >> 56; } else { insn = ld_code2(env, s, pc); -- 2.34.1 The ilen value extracted from ex_value is the length of the EXECUTE instruction itself, and so is the increment to the pc. However, the length of the synthetic insn is located in the opcode like all other instructions. Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- target/s390x/tcg/translate.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/target/s390x/tcg/translate.c b/target/s390x/tcg/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/s390x/tcg/translate.c +++ b/target/s390x/tcg/translate.c @@ -XXX,XX +XXX,XX @@ static const DisasInsn *extract_insn(CPUS390XState *env, DisasContext *s) /* Extract the values saved by EXECUTE. */ insn = s->ex_value & 0xffffffffffff0000ull; ilen = s->ex_value & 0xf; + op = insn >> 56; /* Register insn bytes with translator so plugins work. */ be_insn = cpu_to_be64(insn); - translator_fake_ld(&s->base, &be_insn, ilen); - op = insn >> 56; + translator_fake_ld(&s->base, &be_insn, get_ilen(op)); } else { insn = ld_code2(env, s, pc); op = (insn >> 8) & 0xff; -- 2.34.1 Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- target/s390x/tcg/translate.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-) diff --git a/target/s390x/tcg/translate.c b/target/s390x/tcg/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/s390x/tcg/translate.c +++ b/target/s390x/tcg/translate.c @@ -XXX,XX +XXX,XX @@ static bool s390x_tr_disas_log(const DisasContextBase *dcbase, DisasContext *dc = container_of(dcbase, DisasContext, base); if (unlikely(dc->ex_value)) { - /* ??? Unfortunately target_disas can't use host memory. */ - fprintf(logfile, "IN: EXECUTE %016" PRIx64, dc->ex_value); + /* The ex_value has been recorded with translator_fake_ld. */ + fprintf(logfile, "IN: EXECUTE\n"); + target_disas(logfile, cs, &dc->base); return true; } return false; -- 2.34.1 Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- target/hexagon/translate.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) diff --git a/target/hexagon/translate.c b/target/hexagon/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/hexagon/translate.c +++ b/target/hexagon/translate.c @@ -XXX,XX +XXX,XX @@ #include "exec/translation-block.h" #include "exec/cpu_ldst.h" #include "exec/log.h" -#include "exec/cpu_ldst.h" #include "internal.h" #include "attribs.h" #include "insn.h" @@ -XXX,XX +XXX,XX @@ static bool pkt_crosses_page(CPUHexagonState *env, DisasContext *ctx) int nwords; for (nwords = 0; !found_end && nwords < PACKET_WORDS_MAX; nwords++) { - uint32_t word = cpu_ldl_code(env, + uint32_t word = translator_ldl(env, &ctx->base, ctx->base.pc_next + nwords * sizeof(uint32_t)); found_end = is_packet_end(word); } -- 2.34.1 Reviewed-by: Edgar E. Iglesias <edgar.iglesias@amd.com> Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- target/microblaze/translate.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) diff --git a/target/microblaze/translate.c b/target/microblaze/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/microblaze/translate.c +++ b/target/microblaze/translate.c @@ -XXX,XX +XXX,XX @@ #include "tcg/tcg-op.h" #include "exec/helper-proto.h" #include "exec/helper-gen.h" -#include "exec/cpu_ldst.h" #include "exec/translator.h" #include "qemu/qemu-print.h" @@ -XXX,XX +XXX,XX @@ static void mb_tr_translate_insn(DisasContextBase *dcb, CPUState *cs) dc->tb_flags_to_set = 0; - ir = cpu_ldl_code(cpu_env(cs), dc->base.pc_next); + ir = translator_ldl(cpu_env(cs), &dc->base, dc->base.pc_next); if (!decode(dc, ir)) { trap_illegal(dc, true); } -- 2.34.1 Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- target/i386/tcg/translate.c | 8 +++----- 1 file changed, 3 insertions(+), 5 deletions(-) diff --git a/target/i386/tcg/translate.c b/target/i386/tcg/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/i386/tcg/translate.c +++ b/target/i386/tcg/translate.c @@ -XXX,XX +XXX,XX @@ #include "exec/exec-all.h" #include "tcg/tcg-op.h" #include "tcg/tcg-op-gvec.h" -#include "exec/cpu_ldst.h" #include "exec/translator.h" #include "fpu/softfloat.h" @@ -XXX,XX +XXX,XX @@ static uint64_t advance_pc(CPUX86State *env, DisasContext *s, int num_bytes) * This can happen even if the operand is only one byte long! */ if (((s->pc - 1) ^ (pc - 1)) & TARGET_PAGE_MASK) { - volatile uint8_t unused = - cpu_ldub_code(env, (s->pc - 1) & TARGET_PAGE_MASK); - (void) unused; + (void)translator_ldub(env, &s->base, + (s->pc - 1) & TARGET_PAGE_MASK); } siglongjmp(s->jmpbuf, 1); } @@ -XXX,XX +XXX,XX @@ static void gen_unknown_opcode(CPUX86State *env, DisasContext *s) fprintf(logfile, "ILLOPC: " TARGET_FMT_lx ":", pc); for (; pc < end; ++pc) { - fprintf(logfile, " %02x", cpu_ldub_code(env, pc)); + fprintf(logfile, " %02x", translator_ldub(env, &s->base, pc)); } fprintf(logfile, "\n"); qemu_log_unlock(logfile); -- 2.34.1 Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- target/avr/translate.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) diff --git a/target/avr/translate.c b/target/avr/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/avr/translate.c +++ b/target/avr/translate.c @@ -XXX,XX +XXX,XX @@ #include "cpu.h" #include "exec/exec-all.h" #include "tcg/tcg-op.h" -#include "exec/cpu_ldst.h" #include "exec/helper-proto.h" #include "exec/helper-gen.h" #include "exec/log.h" @@ -XXX,XX +XXX,XX @@ static int to_regs_00_30_by_two(DisasContext *ctx, int indx) static uint16_t next_word(DisasContext *ctx) { - return cpu_lduw_code(ctx->env, ctx->npc++ * 2); + return translator_lduw(ctx->env, &ctx->base, ctx->npc++ * 2); } static int append_16(DisasContext *ctx, int x) -- 2.34.1 Reviewed-by: Edgar E. Iglesias <edgar.iglesias@amd.com> Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- target/cris/translate.c | 25 ++++++++----------------- 1 file changed, 8 insertions(+), 17 deletions(-) diff --git a/target/cris/translate.c b/target/cris/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/cris/translate.c +++ b/target/cris/translate.c @@ -XXX,XX +XXX,XX @@ static int sign_extend(unsigned int val, unsigned int width) } static int cris_fetch(CPUCRISState *env, DisasContext *dc, uint32_t addr, - unsigned int size, unsigned int sign) + unsigned int size, bool sign) { int r; switch (size) { case 4: - { - r = cpu_ldl_code(env, addr); + r = translator_ldl(env, &dc->base, addr); break; - } case 2: - { + r = translator_lduw(env, &dc->base, addr); if (sign) { - r = cpu_ldsw_code(env, addr); - } else { - r = cpu_lduw_code(env, addr); + r = (int16_t)r; } break; - } case 1: - { + r = translator_ldub(env, &dc->base, addr); if (sign) { - r = cpu_ldsb_code(env, addr); - } else { - r = cpu_ldub_code(env, addr); + r = (int8_t)r; } break; - } default: - cpu_abort(CPU(dc->cpu), "Invalid fetch size %d\n", size); - break; + g_assert_not_reached(); } return r; } @@ -XXX,XX +XXX,XX @@ static unsigned int crisv32_decoder(CPUCRISState *env, DisasContext *dc) int i; /* Load a halfword onto the instruction register. */ - dc->ir = cris_fetch(env, dc, dc->pc, 2, 0); + dc->ir = cris_fetch(env, dc, dc->pc, 2, 0); /* Now decode it. */ dc->opcode = EXTRACT_FIELD(dc->ir, 4, 11); -- 2.34.1 Reviewed-by: Edgar E. Iglesias <edgar.iglesias@amd.com> Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- target/cris/translate.c | 1 - target/cris/translate_v10.c.inc | 30 +++++++++--------------------- 2 files changed, 9 insertions(+), 22 deletions(-) diff --git a/target/cris/translate.c b/target/cris/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/cris/translate.c +++ b/target/cris/translate.c @@ -XXX,XX +XXX,XX @@ #include "tcg/tcg-op.h" #include "exec/helper-proto.h" #include "mmu.h" -#include "exec/cpu_ldst.h" #include "exec/translator.h" #include "crisv32-decode.h" #include "qemu/qemu-print.h" diff --git a/target/cris/translate_v10.c.inc b/target/cris/translate_v10.c.inc index XXXXXXX..XXXXXXX 100644 --- a/target/cris/translate_v10.c.inc +++ b/target/cris/translate_v10.c.inc @@ -XXX,XX +XXX,XX @@ static int dec10_prep_move_m(CPUCRISState *env, DisasContext *dc, /* Load [$rs] onto T1. */ if (is_imm) { - if (memsize != 4) { - if (s_ext) { - if (memsize == 1) - imm = cpu_ldsb_code(env, dc->pc + 2); - else - imm = cpu_ldsw_code(env, dc->pc + 2); - } else { - if (memsize == 1) - imm = cpu_ldub_code(env, dc->pc + 2); - else - imm = cpu_lduw_code(env, dc->pc + 2); - } - } else - imm = cpu_ldl_code(env, dc->pc + 2); + imm = cris_fetch(env, dc, dc->pc + 2, memsize, s_ext); tcg_gen_movi_tl(dst, imm); @@ -XXX,XX +XXX,XX @@ static int dec10_dip(CPUCRISState *env, DisasContext *dc) LOG_DIS("dip pc=%x opcode=%d r%d r%d\n", dc->pc, dc->opcode, dc->src, dc->dst); if (dc->src == 15) { - imm = cpu_ldl_code(env, dc->pc + 2); + imm = cris_fetch(env, dc, dc->pc + 2, 4, 0); tcg_gen_movi_tl(cpu_PR[PR_PREFIX], imm); - if (dc->postinc) + if (dc->postinc) { insn_len += 4; + } tcg_gen_addi_tl(cpu_R[15], cpu_R[15], insn_len - 2); } else { gen_load(dc, cpu_PR[PR_PREFIX], cpu_R[dc->src], 4, 0); @@ -XXX,XX +XXX,XX @@ static unsigned int dec10_ind(CPUCRISState *env, DisasContext *dc) if (dc->src == 15) { LOG_DIS("jump.%d %d r%d r%d direct\n", size, dc->opcode, dc->src, dc->dst); - imm = cpu_ldl_code(env, dc->pc + 2); - if (dc->mode == CRISV10_MODE_AUTOINC) + imm = cris_fetch(env, dc, dc->pc + 2, size, 0); + if (dc->mode == CRISV10_MODE_AUTOINC) { insn_len += size; - + } c = tcg_constant_tl(dc->pc + insn_len); t_gen_mov_preg_TN(dc, dc->dst, c); dc->jmp_pc = imm; @@ -XXX,XX +XXX,XX @@ static unsigned int dec10_ind(CPUCRISState *env, DisasContext *dc) case CRISV10_IND_BCC_M: cris_cc_mask(dc, 0); - simm = cpu_ldsw_code(env, dc->pc + 2); + simm = cris_fetch(env, dc, dc->pc + 2, 2, 1); simm += 4; LOG_DIS("bcc_m: b%s %x\n", cc_name(dc->cond), dc->pc + simm); @@ -XXX,XX +XXX,XX @@ static unsigned int crisv10_decoder(CPUCRISState *env, DisasContext *dc) unsigned int insn_len = 2; /* Load a halfword onto the instruction register. */ - dc->ir = cpu_lduw_code(env, dc->pc); + dc->ir = cris_fetch(env, dc, dc->pc, 2, 0); /* Now decode it. */ dc->opcode = EXTRACT_FIELD(dc->ir, 6, 9); -- 2.34.1 Reviewed-by: Alistair Francis <alistair.francis@wdc.com> Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- target/riscv/translate.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) diff --git a/target/riscv/translate.c b/target/riscv/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/riscv/translate.c +++ b/target/riscv/translate.c @@ -XXX,XX +XXX,XX @@ #include "qemu/log.h" #include "cpu.h" #include "tcg/tcg-op.h" -#include "exec/cpu_ldst.h" #include "exec/exec-all.h" #include "exec/helper-proto.h" #include "exec/helper-gen.h" @@ -XXX,XX +XXX,XX @@ static uint32_t opcode_at(DisasContextBase *dcbase, target_ulong pc) CPUState *cpu = ctx->cs; CPURISCVState *env = cpu_env(cpu); - return cpu_ldl_code(env, pc); + return translator_ldl(env, &ctx->base, pc); } /* Include insn module translation function */ @@ -XXX,XX +XXX,XX @@ static void riscv_tr_translate_insn(DisasContextBase *dcbase, CPUState *cpu) unsigned page_ofs = ctx->base.pc_next & ~TARGET_PAGE_MASK; if (page_ofs > TARGET_PAGE_SIZE - MAX_INSN_LEN) { - uint16_t next_insn = cpu_lduw_code(env, ctx->base.pc_next); + uint16_t next_insn = + translator_lduw(env, &ctx->base, ctx->base.pc_next); int len = insn_len(next_insn); if (!is_same_page(&ctx->base, ctx->base.pc_next + len - 1)) { -- 2.34.1 Reviewed-by: Yoshinori Sato <ysato@users.sourceforge.jp> Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- target/rx/translate.c | 27 ++++++++++++++------------- 1 file changed, 14 insertions(+), 13 deletions(-) diff --git a/target/rx/translate.c b/target/rx/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/rx/translate.c +++ b/target/rx/translate.c @@ -XXX,XX +XXX,XX @@ #include "cpu.h" #include "exec/exec-all.h" #include "tcg/tcg-op.h" -#include "exec/cpu_ldst.h" #include "exec/helper-proto.h" #include "exec/helper-gen.h" #include "exec/translator.h" @@ -XXX,XX +XXX,XX @@ static TCGv_i64 cpu_acc; /* decoder helper */ static uint32_t decode_load_bytes(DisasContext *ctx, uint32_t insn, - int i, int n) + int i, int n) { while (++i <= n) { - uint8_t b = cpu_ldub_code(ctx->env, ctx->base.pc_next++); + uint8_t b = translator_ldub(ctx->env, &ctx->base, ctx->base.pc_next++); insn |= b << (32 - i * 8); } return insn; @@ -XXX,XX +XXX,XX @@ static uint32_t li(DisasContext *ctx, int sz) CPURXState *env = ctx->env; addr = ctx->base.pc_next; - tcg_debug_assert(sz < 4); switch (sz) { case 1: ctx->base.pc_next += 1; - return cpu_ldsb_code(env, addr); + return (int8_t)translator_ldub(env, &ctx->base, addr); case 2: ctx->base.pc_next += 2; - return cpu_ldsw_code(env, addr); + return (int16_t)translator_lduw(env, &ctx->base, addr); case 3: ctx->base.pc_next += 3; - tmp = cpu_ldsb_code(env, addr + 2) << 16; - tmp |= cpu_lduw_code(env, addr) & 0xffff; + tmp = (int8_t)translator_ldub(env, &ctx->base, addr + 2); + tmp <<= 16; + tmp |= translator_lduw(env, &ctx->base, addr); return tmp; case 0: ctx->base.pc_next += 4; - return cpu_ldl_code(env, addr); + return translator_ldl(env, &ctx->base, addr); + default: + g_assert_not_reached(); } return 0; } @@ -XXX,XX +XXX,XX @@ static inline TCGv rx_index_addr(DisasContext *ctx, TCGv mem, { uint32_t dsp; - tcg_debug_assert(ld < 3); switch (ld) { case 0: return cpu_regs[reg]; case 1: - dsp = cpu_ldub_code(ctx->env, ctx->base.pc_next) << size; + dsp = translator_ldub(ctx->env, &ctx->base, ctx->base.pc_next) << size; tcg_gen_addi_i32(mem, cpu_regs[reg], dsp); ctx->base.pc_next += 1; return mem; case 2: - dsp = cpu_lduw_code(ctx->env, ctx->base.pc_next) << size; + dsp = translator_lduw(ctx->env, &ctx->base, ctx->base.pc_next) << size; tcg_gen_addi_i32(mem, cpu_regs[reg], dsp); ctx->base.pc_next += 2; return mem; + default: + g_assert_not_reached(); } - return NULL; } static inline MemOp mi_to_mop(unsigned mi) -- 2.34.1 Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- target/xtensa/translate.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) diff --git a/target/xtensa/translate.c b/target/xtensa/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/xtensa/translate.c +++ b/target/xtensa/translate.c @@ -XXX,XX +XXX,XX @@ #include "tcg/tcg-op.h" #include "qemu/log.h" #include "qemu/qemu-print.h" -#include "exec/cpu_ldst.h" #include "semihosting/semihost.h" #include "exec/translator.h" @@ -XXX,XX +XXX,XX @@ static void disas_xtensa_insn(CPUXtensaState *env, DisasContext *dc) static inline unsigned xtensa_insn_len(CPUXtensaState *env, DisasContext *dc) { - uint8_t b0 = cpu_ldub_code(env, dc->pc); + uint8_t b0 = translator_ldub(env, &dc->base, dc->pc); return xtensa_op0_insn_len(dc, b0); } -- 2.34.1 Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- target/s390x/tcg/translate.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) diff --git a/target/s390x/tcg/translate.c b/target/s390x/tcg/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/s390x/tcg/translate.c +++ b/target/s390x/tcg/translate.c @@ -XXX,XX +XXX,XX @@ #include "tcg/tcg-op-gvec.h" #include "qemu/log.h" #include "qemu/host-utils.h" -#include "exec/cpu_ldst.h" #include "exec/helper-proto.h" #include "exec/helper-gen.h" @@ -XXX,XX +XXX,XX @@ static void s390x_tr_insn_start(DisasContextBase *dcbase, CPUState *cs) static target_ulong get_next_pc(CPUS390XState *env, DisasContext *s, uint64_t pc) { - uint64_t insn = cpu_lduw_code(env, pc); + uint64_t insn = translator_lduw(env, &s->base, pc); return pc + get_ilen((insn >> 8) & 0xff); } -- 2.34.1 From: Philippe Mathieu-Daudé <philmd@linaro.org> Previous commits replaced them by translator_ld* calls. Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org> Message-Id: <20240405131532.40913-1-philmd@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- include/exec/cpu_ldst.h | 10 ---------- 1 file changed, 10 deletions(-) diff --git a/include/exec/cpu_ldst.h b/include/exec/cpu_ldst.h index XXXXXXX..XXXXXXX 100644 --- a/include/exec/cpu_ldst.h +++ b/include/exec/cpu_ldst.h @@ -XXX,XX +XXX,XX @@ uint32_t cpu_lduw_code(CPUArchState *env, abi_ptr addr); uint32_t cpu_ldl_code(CPUArchState *env, abi_ptr addr); uint64_t cpu_ldq_code(CPUArchState *env, abi_ptr addr); -static inline int cpu_ldsb_code(CPUArchState *env, abi_ptr addr) -{ - return (int8_t)cpu_ldub_code(env, addr); -} - -static inline int cpu_ldsw_code(CPUArchState *env, abi_ptr addr) -{ - return (int16_t)cpu_lduw_code(env, addr); -} - /** * tlb_vaddr_to_host: * @env: CPUArchState -- 2.34.1 TCG register spill/fill uses tcg_out_ld/st with all types, not necessarily going through INDEX_op_{ld,st}_vec. Cc: qemu-stable@nongnu.org Fixes: 16288ded944 ("tcg/loongarch64: Lower basic tcg vec ops to LSX") Resolves: https://gitlab.com/qemu-project/qemu/-/issues/2336 Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Reviewed-by: Song Gao <gaosong@loongson.cn> Tested-by: Song Gao <gaosong@loongson.cn> --- tcg/loongarch64/tcg-target.c.inc | 103 ++++++++++++++++++++++++------- 1 file changed, 80 insertions(+), 23 deletions(-) diff --git a/tcg/loongarch64/tcg-target.c.inc b/tcg/loongarch64/tcg-target.c.inc index XXXXXXX..XXXXXXX 100644 --- a/tcg/loongarch64/tcg-target.c.inc +++ b/tcg/loongarch64/tcg-target.c.inc @@ -XXX,XX +XXX,XX @@ static void tcg_out_ldst(TCGContext *s, LoongArchInsn opc, TCGReg data, } } -static void tcg_out_ld(TCGContext *s, TCGType type, TCGReg arg, - TCGReg arg1, intptr_t arg2) +static void tcg_out_ld(TCGContext *s, TCGType type, TCGReg dest, + TCGReg base, intptr_t offset) { - bool is_32bit = type == TCG_TYPE_I32; - tcg_out_ldst(s, is_32bit ? OPC_LD_W : OPC_LD_D, arg, arg1, arg2); + switch (type) { + case TCG_TYPE_I32: + if (dest < TCG_REG_V0) { + tcg_out_ldst(s, OPC_LD_W, dest, base, offset); + } else { + tcg_out_dupm_vec(s, TCG_TYPE_I128, MO_32, dest, base, offset); + } + break; + case TCG_TYPE_I64: + if (dest < TCG_REG_V0) { + tcg_out_ldst(s, OPC_LD_D, dest, base, offset); + } else { + tcg_out_dupm_vec(s, TCG_TYPE_I128, MO_64, dest, base, offset); + } + break; + case TCG_TYPE_V128: + if (-0x800 <= offset && offset <= 0x7ff) { + tcg_out_opc_vld(s, dest, base, offset); + } else { + tcg_out_movi(s, TCG_TYPE_PTR, TCG_REG_TMP0, offset); + tcg_out_opc_vldx(s, dest, base, TCG_REG_TMP0); + } + break; + default: + g_assert_not_reached(); + } } -static void tcg_out_st(TCGContext *s, TCGType type, TCGReg arg, - TCGReg arg1, intptr_t arg2) +static void tcg_out_st(TCGContext *s, TCGType type, TCGReg src, + TCGReg base, intptr_t offset) { - bool is_32bit = type == TCG_TYPE_I32; - tcg_out_ldst(s, is_32bit ? OPC_ST_W : OPC_ST_D, arg, arg1, arg2); + switch (type) { + case TCG_TYPE_I32: + if (src < TCG_REG_V0) { + tcg_out_ldst(s, OPC_ST_W, src, base, offset); + } else { + /* TODO: Could use fst_s, fstx_s */ + if (offset < -0x100 || offset > 0xff || (offset & 3)) { + if (-0x800 <= offset && offset <= 0x7ff) { + tcg_out_opc_addi_d(s, TCG_REG_TMP0, base, offset); + } else { + tcg_out_movi(s, TCG_TYPE_PTR, TCG_REG_TMP0, offset); + tcg_out_opc_add_d(s, TCG_REG_TMP0, TCG_REG_TMP0, base); + } + base = TCG_REG_TMP0; + offset = 0; + } + tcg_out_opc_vstelm_w(s, src, base, offset, 0); + } + break; + case TCG_TYPE_I64: + if (src < TCG_REG_V0) { + tcg_out_ldst(s, OPC_ST_D, src, base, offset); + } else { + /* TODO: Could use fst_d, fstx_d */ + if (offset < -0x100 || offset > 0xff || (offset & 7)) { + if (-0x800 <= offset && offset <= 0x7ff) { + tcg_out_opc_addi_d(s, TCG_REG_TMP0, base, offset); + } else { + tcg_out_movi(s, TCG_TYPE_PTR, TCG_REG_TMP0, offset); + tcg_out_opc_add_d(s, TCG_REG_TMP0, TCG_REG_TMP0, base); + } + base = TCG_REG_TMP0; + offset = 0; + } + tcg_out_opc_vstelm_d(s, src, base, offset, 0); + } + break; + case TCG_TYPE_V128: + if (-0x800 <= offset && offset <= 0x7ff) { + tcg_out_opc_vst(s, src, base, offset); + } else { + tcg_out_movi(s, TCG_TYPE_PTR, TCG_REG_TMP0, offset); + tcg_out_opc_vstx(s, src, base, TCG_REG_TMP0); + } + break; + default: + g_assert_not_reached(); + } } static bool tcg_out_sti(TCGContext *s, TCGType type, TCGArg val, @@ -XXX,XX +XXX,XX @@ static void tcg_out_vec_op(TCGContext *s, TCGOpcode opc, { TCGType type = vecl + TCG_TYPE_V64; TCGArg a0, a1, a2, a3; - TCGReg temp = TCG_REG_TMP0; TCGReg temp_vec = TCG_VEC_TMP0; static const LoongArchInsn cmp_vec_insn[16][4] = { @@ -XXX,XX +XXX,XX @@ static void tcg_out_vec_op(TCGContext *s, TCGOpcode opc, switch (opc) { case INDEX_op_st_vec: - /* Try to fit vst imm */ - if (-0x800 <= a2 && a2 <= 0x7ff) { - tcg_out_opc_vst(s, a0, a1, a2); - } else { - tcg_out_movi(s, TCG_TYPE_I64, temp, a2); - tcg_out_opc_vstx(s, a0, a1, temp); - } + tcg_out_st(s, type, a0, a1, a2); break; case INDEX_op_ld_vec: - /* Try to fit vld imm */ - if (-0x800 <= a2 && a2 <= 0x7ff) { - tcg_out_opc_vld(s, a0, a1, a2); - } else { - tcg_out_movi(s, TCG_TYPE_I64, temp, a2); - tcg_out_opc_vldx(s, a0, a1, temp); - } + tcg_out_ld(s, type, a0, a1, a2); break; case INDEX_op_and_vec: tcg_out_opc_vand_v(s, a0, a1, a2); -- 2.34.1 The following changes since commit aa3a285b5bc56a4208b3b57d4a55291e9c260107: Merge tag 'mem-2024-12-21' of https://github.com/davidhildenbrand/qemu into staging (2024-12-22 14:33:27 -0500) are available in the Git repository at: https://gitlab.com/rth7680/qemu.git tags/pull-tcg-20241224 for you to fetch changes up to e4a8e093dc74be049f4829831dce76e5edab0003: accel/tcg: Move gen_intermediate_code to TCGCPUOps.translate_core (2024-12-24 08:32:15 -0800) ---------------------------------------------------------------- tcg/optimize: Remove in-flight mask data from OptContext fpu: Add float*_muladd_scalbn fpu: Remove float_muladd_halve_result fpu: Add float_round_nearest_even_max fpu: Add float_muladd_suppress_add_product_zero target/hexagon: Use float32_muladd accel/tcg: Move gen_intermediate_code to TCGCPUOps.translate_core ---------------------------------------------------------------- Ilya Leoshkevich (1): tests/tcg: Do not use inttypes.h in multiarch/system/memory.c Pierrick Bouvier (1): plugins: optimize cpu_index code generation Richard Henderson (70): tcg/optimize: Split out finish_bb, finish_ebb tcg/optimize: Split out fold_affected_mask tcg/optimize: Copy mask writeback to fold_masks tcg/optimize: Split out fold_masks_zs tcg/optimize: Augment s_mask from z_mask in fold_masks_zs tcg/optimize: Change representation of s_mask tcg/optimize: Use finish_folding in fold_add, fold_add_vec, fold_addsub2 tcg/optimize: Introduce const value accessors for TempOptInfo tcg/optimize: Use fold_masks_zs in fold_and tcg/optimize: Use fold_masks_zs in fold_andc tcg/optimize: Use fold_masks_zs in fold_bswap tcg/optimize: Use fold_masks_zs in fold_count_zeros tcg/optimize: Use fold_masks_z in fold_ctpop tcg/optimize: Use fold_and and fold_masks_z in fold_deposit tcg/optimize: Compute sign mask in fold_deposit tcg/optimize: Use finish_folding in fold_divide tcg/optimize: Use finish_folding in fold_dup, fold_dup2 tcg/optimize: Use fold_masks_s in fold_eqv tcg/optimize: Use fold_masks_z in fold_extract tcg/optimize: Use finish_folding in fold_extract2 tcg/optimize: Use fold_masks_zs in fold_exts tcg/optimize: Use fold_masks_z in fold_extu tcg/optimize: Use fold_masks_zs in fold_movcond tcg/optimize: Use finish_folding in fold_mul* tcg/optimize: Use fold_masks_s in fold_nand tcg/optimize: Use fold_masks_z in fold_neg_no_const tcg/optimize: Use fold_masks_s in fold_nor tcg/optimize: Use fold_masks_s in fold_not tcg/optimize: Use fold_masks_zs in fold_or tcg/optimize: Use fold_masks_zs in fold_orc tcg/optimize: Use fold_masks_zs in fold_qemu_ld tcg/optimize: Return true from fold_qemu_st, fold_tcg_st tcg/optimize: Use finish_folding in fold_remainder tcg/optimize: Distinguish simplification in fold_setcond_zmask tcg/optimize: Use fold_masks_z in fold_setcond tcg/optimize: Use fold_masks_s in fold_negsetcond tcg/optimize: Use fold_masks_z in fold_setcond2 tcg/optimize: Use finish_folding in fold_cmp_vec tcg/optimize: Use finish_folding in fold_cmpsel_vec tcg/optimize: Use fold_masks_zs in fold_sextract tcg/optimize: Use fold_masks_zs, fold_masks_s in fold_shift tcg/optimize: Simplify sign bit test in fold_shift tcg/optimize: Use finish_folding in fold_sub, fold_sub_vec tcg/optimize: Use fold_masks_zs in fold_tcg_ld tcg/optimize: Use finish_folding in fold_tcg_ld_memcopy tcg/optimize: Use fold_masks_zs in fold_xor tcg/optimize: Use finish_folding in fold_bitsel_vec tcg/optimize: Use finish_folding as default in tcg_optimize tcg/optimize: Remove z_mask, s_mask from OptContext tcg/optimize: Re-enable sign-mask optimizations tcg/optimize: Move fold_bitsel_vec into alphabetic sort tcg/optimize: Move fold_cmp_vec, fold_cmpsel_vec into alphabetic sort softfloat: Add float{16,32,64}_muladd_scalbn target/arm: Use float*_muladd_scalbn target/sparc: Use float*_muladd_scalbn softfloat: Remove float_muladd_halve_result softfloat: Add float_round_nearest_even_max softfloat: Add float_muladd_suppress_add_product_zero target/hexagon: Use float32_mul in helper_sfmpy target/hexagon: Use float32_muladd for helper_sffma target/hexagon: Use float32_muladd for helper_sffms target/hexagon: Use float32_muladd_scalbn for helper_sffma_sc target/hexagon: Use float32_muladd for helper_sffm[as]_lib target/hexagon: Remove internal_fmafx target/hexagon: Expand GEN_XF_ROUND target/hexagon: Remove Float target/hexagon: Remove Double target/hexagon: Use mulu64 for int128_mul_6464 target/hexagon: Simplify internal_mpyhh setup accel/tcg: Move gen_intermediate_code to TCGCPUOps.translate_core include/exec/translator.h | 14 - include/fpu/softfloat-types.h | 2 + include/fpu/softfloat.h | 14 +- include/hw/core/tcg-cpu-ops.h | 13 + target/alpha/cpu.h | 2 + target/arm/internals.h | 2 + target/avr/cpu.h | 2 + target/hexagon/cpu.h | 2 + target/hexagon/fma_emu.h | 3 - target/hppa/cpu.h | 2 + target/i386/tcg/helper-tcg.h | 2 + target/loongarch/internals.h | 2 + target/m68k/cpu.h | 2 + target/microblaze/cpu.h | 2 + target/mips/tcg/tcg-internal.h | 2 + target/openrisc/cpu.h | 2 + target/ppc/cpu.h | 2 + target/riscv/cpu.h | 3 + target/rx/cpu.h | 2 + target/s390x/s390x-internal.h | 2 + target/sh4/cpu.h | 2 + target/sparc/cpu.h | 2 + target/sparc/helper.h | 4 +- target/tricore/cpu.h | 2 + target/xtensa/cpu.h | 2 + accel/tcg/cpu-exec.c | 8 +- accel/tcg/plugin-gen.c | 9 + accel/tcg/translate-all.c | 8 +- fpu/softfloat.c | 63 +-- target/alpha/cpu.c | 1 + target/alpha/translate.c | 4 +- target/arm/cpu.c | 1 + target/arm/tcg/cpu-v7m.c | 1 + target/arm/tcg/helper-a64.c | 6 +- target/arm/tcg/translate.c | 5 +- target/avr/cpu.c | 1 + target/avr/translate.c | 6 +- target/hexagon/cpu.c | 1 + target/hexagon/fma_emu.c | 496 ++++++--------------- target/hexagon/op_helper.c | 125 ++---- target/hexagon/translate.c | 4 +- target/hppa/cpu.c | 1 + target/hppa/translate.c | 4 +- target/i386/tcg/tcg-cpu.c | 1 + target/i386/tcg/translate.c | 5 +- target/loongarch/cpu.c | 1 + target/loongarch/tcg/translate.c | 4 +- target/m68k/cpu.c | 1 + target/m68k/translate.c | 4 +- target/microblaze/cpu.c | 1 + target/microblaze/translate.c | 4 +- target/mips/cpu.c | 1 + target/mips/tcg/translate.c | 4 +- target/openrisc/cpu.c | 1 + target/openrisc/translate.c | 4 +- target/ppc/cpu_init.c | 1 + target/ppc/translate.c | 4 +- target/riscv/tcg/tcg-cpu.c | 1 + target/riscv/translate.c | 4 +- target/rx/cpu.c | 1 + target/rx/translate.c | 4 +- target/s390x/cpu.c | 1 + target/s390x/tcg/translate.c | 4 +- target/sh4/cpu.c | 1 + target/sh4/translate.c | 4 +- target/sparc/cpu.c | 1 + target/sparc/fop_helper.c | 8 +- target/sparc/translate.c | 84 ++-- target/tricore/cpu.c | 1 + target/tricore/translate.c | 5 +- target/xtensa/cpu.c | 1 + target/xtensa/translate.c | 4 +- tcg/optimize.c | 857 +++++++++++++++++++----------------- tests/tcg/multiarch/system/memory.c | 9 +- fpu/softfloat-parts.c.inc | 16 +- 75 files changed, 866 insertions(+), 1009 deletions(-) From: Ilya Leoshkevich <iii@linux.ibm.com> make check-tcg fails on Fedora with the following error message: alpha-linux-gnu-gcc [...] qemu/tests/tcg/multiarch/system/memory.c -o memory [...] qemu/tests/tcg/multiarch/system/memory.c:17:10: fatal error: inttypes.h: No such file or directory 17 | #include <inttypes.h> | ^~~~~~~~~~~~ compilation terminated. The reason is that Fedora has cross-compilers, but no cross-glibc headers. Fix by hardcoding the format specifiers and dropping the include. An alternative fix would be to introduce a configure check for inttypes.h. But this would make it impossible to use Fedora cross-compilers for softmmu tests, which used to work so far. Fixes: ecbcc9ead2f8 ("tests/tcg: add a system test to check memory instrumentation") Signed-off-by: Ilya Leoshkevich <iii@linux.ibm.com> Reviewed-by: Paolo Bonzini <pbonzini@redhat.com> Message-ID: <20241010085906.226249-1-iii@linux.ibm.com> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- tests/tcg/multiarch/system/memory.c | 9 ++++----- 1 file changed, 4 insertions(+), 5 deletions(-) diff --git a/tests/tcg/multiarch/system/memory.c b/tests/tcg/multiarch/system/memory.c index XXXXXXX..XXXXXXX 100644 --- a/tests/tcg/multiarch/system/memory.c +++ b/tests/tcg/multiarch/system/memory.c @@ -XXX,XX +XXX,XX @@ #include <stdint.h> #include <stdbool.h> -#include <inttypes.h> #include <minilib.h> #ifndef CHECK_UNALIGNED @@ -XXX,XX +XXX,XX @@ int main(void) int i; bool ok = true; - ml_printf("Test data start: 0x%"PRIxPTR"\n", &test_data[0]); - ml_printf("Test data end: 0x%"PRIxPTR"\n", &test_data[TEST_SIZE]); + ml_printf("Test data start: 0x%lx\n", (unsigned long)&test_data[0]); + ml_printf("Test data end: 0x%lx\n", (unsigned long)&test_data[TEST_SIZE]); /* Run through the unsigned tests first */ for (i = 0; i < ARRAY_SIZE(init_ufns) && ok; i++) { @@ -XXX,XX +XXX,XX @@ int main(void) ok = do_signed_reads(true); } - ml_printf("Test data read: %"PRId32"\n", test_read_count); - ml_printf("Test data write: %"PRId32"\n", test_write_count); + ml_printf("Test data read: %lu\n", (unsigned long)test_read_count); + ml_printf("Test data write: %lu\n", (unsigned long)test_write_count); ml_printf("Test complete: %s\n", ok ? "PASSED" : "FAILED"); return ok ? 0 : -1; } -- 2.43.0 From: Pierrick Bouvier <pierrick.bouvier@linaro.org> When running with a single vcpu, we can return a constant instead of a load when accessing cpu_index. A side effect is that all tcg operations using it are optimized, most notably scoreboard access. When running a simple loop in user-mode, the speedup is around 20%. Signed-off-by: Pierrick Bouvier <pierrick.bouvier@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-ID: <20241128213843.1023080-1-pierrick.bouvier@linaro.org> --- accel/tcg/plugin-gen.c | 9 +++++++++ 1 file changed, 9 insertions(+) diff --git a/accel/tcg/plugin-gen.c b/accel/tcg/plugin-gen.c index XXXXXXX..XXXXXXX 100644 --- a/accel/tcg/plugin-gen.c +++ b/accel/tcg/plugin-gen.c @@ -XXX,XX +XXX,XX @@ static void gen_disable_mem_helper(void) static TCGv_i32 gen_cpu_index(void) { + /* + * Optimize when we run with a single vcpu. All values using cpu_index, + * including scoreboard index, will be optimized out. + * User-mode calls tb_flush when setting this flag. In system-mode, all + * vcpus are created before generating code. + */ + if (!tcg_cflags_has(current_cpu, CF_PARALLEL)) { + return tcg_constant_i32(current_cpu->cpu_index); + } TCGv_i32 cpu_index = tcg_temp_ebb_new_i32(); tcg_gen_ld_i32(cpu_index, tcg_env, -offsetof(ArchCPU, env) + offsetof(CPUState, cpu_index)); -- 2.43.0 Call them directly from the opcode switch statement in tcg_optimize, rather than in finish_folding based on opcode flags. Adjust folding of conditional branches to match. Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- tcg/optimize.c | 47 +++++++++++++++++++++++++++++++---------------- 1 file changed, 31 insertions(+), 16 deletions(-) diff --git a/tcg/optimize.c b/tcg/optimize.c index XXXXXXX..XXXXXXX 100644 --- a/tcg/optimize.c +++ b/tcg/optimize.c @@ -XXX,XX +XXX,XX @@ static void copy_propagate(OptContext *ctx, TCGOp *op, } } +static void finish_bb(OptContext *ctx) +{ + /* We only optimize memory barriers across basic blocks. */ + ctx->prev_mb = NULL; +} + +static void finish_ebb(OptContext *ctx) +{ + finish_bb(ctx); + /* We only optimize across extended basic blocks. */ + memset(&ctx->temps_used, 0, sizeof(ctx->temps_used)); + remove_mem_copy_all(ctx); +} + static void finish_folding(OptContext *ctx, TCGOp *op) { const TCGOpDef *def = &tcg_op_defs[op->opc]; int i, nb_oargs; - /* - * We only optimize extended basic blocks. If the opcode ends a BB - * and is not a conditional branch, reset all temp data. - */ - if (def->flags & TCG_OPF_BB_END) { - ctx->prev_mb = NULL; - if (!(def->flags & TCG_OPF_COND_BRANCH)) { - memset(&ctx->temps_used, 0, sizeof(ctx->temps_used)); - remove_mem_copy_all(ctx); - } - return; - } - nb_oargs = def->nb_oargs; for (i = 0; i < nb_oargs; i++) { TCGTemp *ts = arg_temp(op->args[i]); @@ -XXX,XX +XXX,XX @@ static bool fold_brcond(OptContext *ctx, TCGOp *op) if (i > 0) { op->opc = INDEX_op_br; op->args[0] = op->args[3]; + finish_ebb(ctx); + } else { + finish_bb(ctx); } - return false; + return true; } static bool fold_brcond2(OptContext *ctx, TCGOp *op) @@ -XXX,XX +XXX,XX @@ static bool fold_brcond2(OptContext *ctx, TCGOp *op) } op->opc = INDEX_op_br; op->args[0] = label; - break; + finish_ebb(ctx); + return true; } - return false; + + finish_bb(ctx); + return true; } static bool fold_bswap(OptContext *ctx, TCGOp *op) @@ -XXX,XX +XXX,XX @@ void tcg_optimize(TCGContext *s) CASE_OP_32_64_VEC(xor): done = fold_xor(&ctx, op); break; + case INDEX_op_set_label: + case INDEX_op_br: + case INDEX_op_exit_tb: + case INDEX_op_goto_tb: + case INDEX_op_goto_ptr: + finish_ebb(&ctx); + done = true; + break; default: break; } -- 2.43.0 There are only a few logical operations which can compute an "affected" mask. Split out handling of this optimization to a separate function, only to be called when applicable. Remove the a_mask field from OptContext, as the mask is no longer stored anywhere. Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- tcg/optimize.c | 42 +++++++++++++++++++++++++++--------------- 1 file changed, 27 insertions(+), 15 deletions(-) diff --git a/tcg/optimize.c b/tcg/optimize.c index XXXXXXX..XXXXXXX 100644 --- a/tcg/optimize.c +++ b/tcg/optimize.c @@ -XXX,XX +XXX,XX @@ typedef struct OptContext { QSIMPLEQ_HEAD(, MemCopyInfo) mem_free; /* In flight values from optimization. */ - uint64_t a_mask; /* mask bit is 0 iff value identical to first input */ uint64_t z_mask; /* mask bit is 0 iff value bit is 0 */ uint64_t s_mask; /* mask of clrsb(value) bits */ TCGType type; @@ -XXX,XX +XXX,XX @@ static bool fold_const2_commutative(OptContext *ctx, TCGOp *op) static bool fold_masks(OptContext *ctx, TCGOp *op) { - uint64_t a_mask = ctx->a_mask; uint64_t z_mask = ctx->z_mask; uint64_t s_mask = ctx->s_mask; @@ -XXX,XX +XXX,XX @@ static bool fold_masks(OptContext *ctx, TCGOp *op) * type changing opcodes. */ if (ctx->type == TCG_TYPE_I32) { - a_mask = (int32_t)a_mask; z_mask = (int32_t)z_mask; s_mask |= MAKE_64BIT_MASK(32, 32); ctx->z_mask = z_mask; @@ -XXX,XX +XXX,XX @@ static bool fold_masks(OptContext *ctx, TCGOp *op) if (z_mask == 0) { return tcg_opt_gen_movi(ctx, op, op->args[0], 0); } + return false; +} + +/* + * An "affected" mask bit is 0 if and only if the result is identical + * to the first input. Thus if the entire mask is 0, the operation + * is equivalent to a copy. + */ +static bool fold_affected_mask(OptContext *ctx, TCGOp *op, uint64_t a_mask) +{ + if (ctx->type == TCG_TYPE_I32) { + a_mask = (uint32_t)a_mask; + } if (a_mask == 0) { return tcg_opt_gen_mov(ctx, op, op->args[0], op->args[1]); } @@ -XXX,XX +XXX,XX @@ static bool fold_and(OptContext *ctx, TCGOp *op) * Known-zeros does not imply known-ones. Therefore unless * arg2 is constant, we can't infer affected bits from it. */ - if (arg_is_const(op->args[2])) { - ctx->a_mask = z1 & ~z2; + if (arg_is_const(op->args[2]) && + fold_affected_mask(ctx, op, z1 & ~z2)) { + return true; } return fold_masks(ctx, op); @@ -XXX,XX +XXX,XX @@ static bool fold_andc(OptContext *ctx, TCGOp *op) */ if (arg_is_const(op->args[2])) { uint64_t z2 = ~arg_info(op->args[2])->z_mask; - ctx->a_mask = z1 & ~z2; + if (fold_affected_mask(ctx, op, z1 & ~z2)) { + return true; + } z1 &= z2; } ctx->z_mask = z1; @@ -XXX,XX +XXX,XX @@ static bool fold_extract(OptContext *ctx, TCGOp *op) z_mask_old = arg_info(op->args[1])->z_mask; z_mask = extract64(z_mask_old, pos, len); - if (pos == 0) { - ctx->a_mask = z_mask_old ^ z_mask; + if (pos == 0 && fold_affected_mask(ctx, op, z_mask_old ^ z_mask)) { + return true; } ctx->z_mask = z_mask; ctx->s_mask = smask_from_zmask(z_mask); @@ -XXX,XX +XXX,XX @@ static bool fold_exts(OptContext *ctx, TCGOp *op) ctx->z_mask = z_mask; ctx->s_mask = s_mask; - if (!type_change) { - ctx->a_mask = s_mask & ~s_mask_old; + if (!type_change && fold_affected_mask(ctx, op, s_mask & ~s_mask_old)) { + return true; } return fold_masks(ctx, op); @@ -XXX,XX +XXX,XX @@ static bool fold_extu(OptContext *ctx, TCGOp *op) ctx->z_mask = z_mask; ctx->s_mask = smask_from_zmask(z_mask); - if (!type_change) { - ctx->a_mask = z_mask_old ^ z_mask; + if (!type_change && fold_affected_mask(ctx, op, z_mask_old ^ z_mask)) { + return true; } return fold_masks(ctx, op); } @@ -XXX,XX +XXX,XX @@ static bool fold_sextract(OptContext *ctx, TCGOp *op) s_mask |= MAKE_64BIT_MASK(len, 64 - len); ctx->s_mask = s_mask; - if (pos == 0) { - ctx->a_mask = s_mask & ~s_mask_old; + if (pos == 0 && fold_affected_mask(ctx, op, s_mask & ~s_mask_old)) { + return true; } return fold_masks(ctx, op); @@ -XXX,XX +XXX,XX @@ void tcg_optimize(TCGContext *s) } /* Assume all bits affected, no bits known zero, no sign reps. */ - ctx.a_mask = -1; ctx.z_mask = -1; ctx.s_mask = 0; -- 2.43.0 Use of fold_masks should be restricted to those opcodes that can reliably make use of it -- those with a single output, and from higher-level folders that set up the masks. Prepare for conversion of each folder in turn. Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- tcg/optimize.c | 17 ++++++++++++++--- 1 file changed, 14 insertions(+), 3 deletions(-) diff --git a/tcg/optimize.c b/tcg/optimize.c index XXXXXXX..XXXXXXX 100644 --- a/tcg/optimize.c +++ b/tcg/optimize.c @@ -XXX,XX +XXX,XX @@ static bool fold_masks(OptContext *ctx, TCGOp *op) { uint64_t z_mask = ctx->z_mask; uint64_t s_mask = ctx->s_mask; + const TCGOpDef *def = &tcg_op_defs[op->opc]; + TCGTemp *ts; + TempOptInfo *ti; + + /* Only single-output opcodes are supported here. */ + tcg_debug_assert(def->nb_oargs == 1); /* * 32-bit ops generate 32-bit results, which for the purpose of @@ -XXX,XX +XXX,XX @@ static bool fold_masks(OptContext *ctx, TCGOp *op) if (ctx->type == TCG_TYPE_I32) { z_mask = (int32_t)z_mask; s_mask |= MAKE_64BIT_MASK(32, 32); - ctx->z_mask = z_mask; - ctx->s_mask = s_mask; } if (z_mask == 0) { return tcg_opt_gen_movi(ctx, op, op->args[0], 0); } - return false; + + ts = arg_temp(op->args[0]); + reset_ts(ctx, ts); + + ti = ts_info(ts); + ti->z_mask = z_mask; + ti->s_mask = s_mask; + return true; } /* -- 2.43.0 Add a routine to which masks can be passed directly, rather than storing them into OptContext. To be used in upcoming patches. Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- tcg/optimize.c | 15 ++++++++++++--- 1 file changed, 12 insertions(+), 3 deletions(-) diff --git a/tcg/optimize.c b/tcg/optimize.c index XXXXXXX..XXXXXXX 100644 --- a/tcg/optimize.c +++ b/tcg/optimize.c @@ -XXX,XX +XXX,XX @@ static bool fold_const2_commutative(OptContext *ctx, TCGOp *op) return fold_const2(ctx, op); } -static bool fold_masks(OptContext *ctx, TCGOp *op) +/* + * Record "zero" and "sign" masks for the single output of @op. + * See TempOptInfo definition of z_mask and s_mask. + * If z_mask allows, fold the output to constant zero. + */ +static bool fold_masks_zs(OptContext *ctx, TCGOp *op, + uint64_t z_mask, uint64_t s_mask) { - uint64_t z_mask = ctx->z_mask; - uint64_t s_mask = ctx->s_mask; const TCGOpDef *def = &tcg_op_defs[op->opc]; TCGTemp *ts; TempOptInfo *ti; @@ -XXX,XX +XXX,XX @@ static bool fold_masks(OptContext *ctx, TCGOp *op) return true; } +static bool fold_masks(OptContext *ctx, TCGOp *op) +{ + return fold_masks_zs(ctx, op, ctx->z_mask, ctx->s_mask); +} + /* * An "affected" mask bit is 0 if and only if the result is identical * to the first input. Thus if the entire mask is 0, the operation -- 2.43.0 Consider the passed s_mask to be a minimum deduced from either existing s_mask or from a sign-extension operation. We may be able to deduce more from the set of known zeros. Remove identical logic from several opcode folders. Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- tcg/optimize.c | 21 ++++++--------------- 1 file changed, 6 insertions(+), 15 deletions(-) diff --git a/tcg/optimize.c b/tcg/optimize.c index XXXXXXX..XXXXXXX 100644 --- a/tcg/optimize.c +++ b/tcg/optimize.c @@ -XXX,XX +XXX,XX @@ static bool fold_const2_commutative(OptContext *ctx, TCGOp *op) * Record "zero" and "sign" masks for the single output of @op. * See TempOptInfo definition of z_mask and s_mask. * If z_mask allows, fold the output to constant zero. + * The passed s_mask may be augmented by z_mask. */ static bool fold_masks_zs(OptContext *ctx, TCGOp *op, uint64_t z_mask, uint64_t s_mask) @@ -XXX,XX +XXX,XX @@ static bool fold_masks_zs(OptContext *ctx, TCGOp *op, ti = ts_info(ts); ti->z_mask = z_mask; - ti->s_mask = s_mask; + ti->s_mask = s_mask | smask_from_zmask(z_mask); return true; } @@ -XXX,XX +XXX,XX @@ static bool fold_bswap(OptContext *ctx, TCGOp *op) default: g_assert_not_reached(); } - s_mask = smask_from_zmask(z_mask); + s_mask = 0; switch (op->args[2] & (TCG_BSWAP_OZ | TCG_BSWAP_OS)) { case TCG_BSWAP_OZ: break; @@ -XXX,XX +XXX,XX @@ static bool fold_bswap(OptContext *ctx, TCGOp *op) default: /* The high bits are undefined: force all bits above the sign to 1. */ z_mask |= sign << 1; - s_mask = 0; break; } ctx->z_mask = z_mask; @@ -XXX,XX +XXX,XX @@ static bool fold_count_zeros(OptContext *ctx, TCGOp *op) g_assert_not_reached(); } ctx->z_mask = arg_info(op->args[2])->z_mask | z_mask; - ctx->s_mask = smask_from_zmask(ctx->z_mask); return false; } @@ -XXX,XX +XXX,XX @@ static bool fold_ctpop(OptContext *ctx, TCGOp *op) default: g_assert_not_reached(); } - ctx->s_mask = smask_from_zmask(ctx->z_mask); return false; } @@ -XXX,XX +XXX,XX @@ static bool fold_extract(OptContext *ctx, TCGOp *op) return true; } ctx->z_mask = z_mask; - ctx->s_mask = smask_from_zmask(z_mask); return fold_masks(ctx, op); } @@ -XXX,XX +XXX,XX @@ static bool fold_extu(OptContext *ctx, TCGOp *op) } ctx->z_mask = z_mask; - ctx->s_mask = smask_from_zmask(z_mask); if (!type_change && fold_affected_mask(ctx, op, z_mask_old ^ z_mask)) { return true; } @@ -XXX,XX +XXX,XX @@ static bool fold_qemu_ld(OptContext *ctx, TCGOp *op) int width = 8 * memop_size(mop); if (width < 64) { - ctx->s_mask = MAKE_64BIT_MASK(width, 64 - width); - if (!(mop & MO_SIGN)) { + if (mop & MO_SIGN) { + ctx->s_mask = MAKE_64BIT_MASK(width, 64 - width); + } else { ctx->z_mask = MAKE_64BIT_MASK(0, width); - ctx->s_mask <<= 1; } } @@ -XXX,XX +XXX,XX @@ static bool fold_setcond(OptContext *ctx, TCGOp *op) fold_setcond_tst_pow2(ctx, op, false); ctx->z_mask = 1; - ctx->s_mask = smask_from_zmask(1); return false; } @@ -XXX,XX +XXX,XX @@ static bool fold_setcond2(OptContext *ctx, TCGOp *op) } ctx->z_mask = 1; - ctx->s_mask = smask_from_zmask(1); return false; do_setcond_const: @@ -XXX,XX +XXX,XX @@ static bool fold_tcg_ld(OptContext *ctx, TCGOp *op) break; CASE_OP_32_64(ld8u): ctx->z_mask = MAKE_64BIT_MASK(0, 8); - ctx->s_mask = MAKE_64BIT_MASK(9, 55); break; CASE_OP_32_64(ld16s): ctx->s_mask = MAKE_64BIT_MASK(16, 48); break; CASE_OP_32_64(ld16u): ctx->z_mask = MAKE_64BIT_MASK(0, 16); - ctx->s_mask = MAKE_64BIT_MASK(17, 47); break; case INDEX_op_ld32s_i64: ctx->s_mask = MAKE_64BIT_MASK(32, 32); break; case INDEX_op_ld32u_i64: ctx->z_mask = MAKE_64BIT_MASK(0, 32); - ctx->s_mask = MAKE_64BIT_MASK(33, 31); break; default: g_assert_not_reached(); -- 2.43.0 Change the representation from sign bit repetitions to all bits equal to the sign bit, including the sign bit itself. The previous format has a problem in that it is difficult to recreate a valid sign mask after a shift operation: the "repetitions" part of the previous format meant that applying the same shift as for the value lead to an off-by-one value. The new format, including the sign bit itself, means that the sign mask can be manipulated in exactly the same way as the value, canonicalization is easier. Canonicalize the s_mask in fold_masks_zs, rather than requiring callers to do so. Treat 0 as a non-canonical but typeless input for no sign information, which will be reset as appropriate for the data type. We can easily fold in the data from z_mask while canonicalizing. Temporarily disable optimizations using s_mask while each operation is converted to use fold_masks_zs and to the new form. Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- tcg/optimize.c | 64 ++++++++++++-------------------------------------- 1 file changed, 15 insertions(+), 49 deletions(-) diff --git a/tcg/optimize.c b/tcg/optimize.c index XXXXXXX..XXXXXXX 100644 --- a/tcg/optimize.c +++ b/tcg/optimize.c @@ -XXX,XX +XXX,XX @@ typedef struct TempOptInfo { QSIMPLEQ_HEAD(, MemCopyInfo) mem_copy; uint64_t val; uint64_t z_mask; /* mask bit is 0 if and only if value bit is 0 */ - uint64_t s_mask; /* a left-aligned mask of clrsb(value) bits. */ + uint64_t s_mask; /* mask bit is 1 if value bit matches msb */ } TempOptInfo; typedef struct OptContext { @@ -XXX,XX +XXX,XX @@ typedef struct OptContext { /* In flight values from optimization. */ uint64_t z_mask; /* mask bit is 0 iff value bit is 0 */ - uint64_t s_mask; /* mask of clrsb(value) bits */ + uint64_t s_mask; /* mask bit is 1 if value bit matches msb */ TCGType type; } OptContext; -/* Calculate the smask for a specific value. */ -static uint64_t smask_from_value(uint64_t value) -{ - int rep = clrsb64(value); - return ~(~0ull >> rep); -} - -/* - * Calculate the smask for a given set of known-zeros. - * If there are lots of zeros on the left, we can consider the remainder - * an unsigned field, and thus the corresponding signed field is one bit - * larger. - */ -static uint64_t smask_from_zmask(uint64_t zmask) -{ - /* - * Only the 0 bits are significant for zmask, thus the msb itself - * must be zero, else we have no sign information. - */ - int rep = clz64(zmask); - if (rep == 0) { - return 0; - } - rep -= 1; - return ~(~0ull >> rep); -} - -/* - * Recreate a properly left-aligned smask after manipulation. - * Some bit-shuffling, particularly shifts and rotates, may - * retain sign bits on the left, but may scatter disconnected - * sign bits on the right. Retain only what remains to the left. - */ -static uint64_t smask_from_smask(int64_t smask) -{ - /* Only the 1 bits are significant for smask */ - return smask_from_zmask(~smask); -} - static inline TempOptInfo *ts_info(TCGTemp *ts) { return ts->state_ptr; @@ -XXX,XX +XXX,XX @@ static void init_ts_info(OptContext *ctx, TCGTemp *ts) ti->is_const = true; ti->val = ts->val; ti->z_mask = ts->val; - ti->s_mask = smask_from_value(ts->val); + ti->s_mask = INT64_MIN >> clrsb64(ts->val); } else { ti->is_const = false; ti->z_mask = -1; @@ -XXX,XX +XXX,XX @@ static void finish_folding(OptContext *ctx, TCGOp *op) */ if (i == 0) { ts_info(ts)->z_mask = ctx->z_mask; - ts_info(ts)->s_mask = ctx->s_mask; } } } @@ -XXX,XX +XXX,XX @@ static bool fold_const2_commutative(OptContext *ctx, TCGOp *op) * The passed s_mask may be augmented by z_mask. */ static bool fold_masks_zs(OptContext *ctx, TCGOp *op, - uint64_t z_mask, uint64_t s_mask) + uint64_t z_mask, int64_t s_mask) { const TCGOpDef *def = &tcg_op_defs[op->opc]; TCGTemp *ts; TempOptInfo *ti; + int rep; /* Only single-output opcodes are supported here. */ tcg_debug_assert(def->nb_oargs == 1); @@ -XXX,XX +XXX,XX @@ static bool fold_masks_zs(OptContext *ctx, TCGOp *op, */ if (ctx->type == TCG_TYPE_I32) { z_mask = (int32_t)z_mask; - s_mask |= MAKE_64BIT_MASK(32, 32); + s_mask |= INT32_MIN; } if (z_mask == 0) { @@ -XXX,XX +XXX,XX @@ static bool fold_masks_zs(OptContext *ctx, TCGOp *op, ti = ts_info(ts); ti->z_mask = z_mask; - ti->s_mask = s_mask | smask_from_zmask(z_mask); + + /* Canonicalize s_mask and incorporate data from z_mask. */ + rep = clz64(~s_mask); + rep = MAX(rep, clz64(z_mask)); + rep = MAX(rep - 1, 0); + ti->s_mask = INT64_MIN >> rep; + return true; } @@ -XXX,XX +XXX,XX @@ static bool fold_exts(OptContext *ctx, TCGOp *op) ctx->z_mask = z_mask; ctx->s_mask = s_mask; - if (!type_change && fold_affected_mask(ctx, op, s_mask & ~s_mask_old)) { + if (0 && !type_change && fold_affected_mask(ctx, op, s_mask & ~s_mask_old)) { return true; } @@ -XXX,XX +XXX,XX @@ static bool fold_sextract(OptContext *ctx, TCGOp *op) s_mask |= MAKE_64BIT_MASK(len, 64 - len); ctx->s_mask = s_mask; - if (pos == 0 && fold_affected_mask(ctx, op, s_mask & ~s_mask_old)) { + if (0 && pos == 0 && fold_affected_mask(ctx, op, s_mask & ~s_mask_old)) { return true; } @@ -XXX,XX +XXX,XX @@ static bool fold_shift(OptContext *ctx, TCGOp *op) ctx->z_mask = do_constant_folding(op->opc, ctx->type, z_mask, sh); s_mask = do_constant_folding(op->opc, ctx->type, s_mask, sh); - ctx->s_mask = smask_from_smask(s_mask); return fold_masks(ctx, op); } -- 2.43.0 Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- tcg/optimize.c | 9 +++++---- 1 file changed, 5 insertions(+), 4 deletions(-) diff --git a/tcg/optimize.c b/tcg/optimize.c index XXXXXXX..XXXXXXX 100644 --- a/tcg/optimize.c +++ b/tcg/optimize.c @@ -XXX,XX +XXX,XX @@ static void finish_ebb(OptContext *ctx) remove_mem_copy_all(ctx); } -static void finish_folding(OptContext *ctx, TCGOp *op) +static bool finish_folding(OptContext *ctx, TCGOp *op) { const TCGOpDef *def = &tcg_op_defs[op->opc]; int i, nb_oargs; @@ -XXX,XX +XXX,XX @@ static void finish_folding(OptContext *ctx, TCGOp *op) ts_info(ts)->z_mask = ctx->z_mask; } } + return true; } /* @@ -XXX,XX +XXX,XX @@ static bool fold_add(OptContext *ctx, TCGOp *op) fold_xi_to_x(ctx, op, 0)) { return true; } - return false; + return finish_folding(ctx, op); } /* We cannot as yet do_constant_folding with vectors. */ @@ -XXX,XX +XXX,XX @@ static bool fold_add_vec(OptContext *ctx, TCGOp *op) fold_xi_to_x(ctx, op, 0)) { return true; } - return false; + return finish_folding(ctx, op); } static bool fold_addsub2(OptContext *ctx, TCGOp *op, bool add) @@ -XXX,XX +XXX,XX @@ static bool fold_addsub2(OptContext *ctx, TCGOp *op, bool add) op->args[4] = arg_new_constant(ctx, bl); op->args[5] = arg_new_constant(ctx, bh); } - return false; + return finish_folding(ctx, op); } static bool fold_add2(OptContext *ctx, TCGOp *op) -- 2.43.0 Introduce ti_is_const, ti_const_val, ti_is_const_val. Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- tcg/optimize.c | 20 +++++++++++++++++--- 1 file changed, 17 insertions(+), 3 deletions(-) diff --git a/tcg/optimize.c b/tcg/optimize.c index XXXXXXX..XXXXXXX 100644 --- a/tcg/optimize.c +++ b/tcg/optimize.c @@ -XXX,XX +XXX,XX @@ static inline TempOptInfo *arg_info(TCGArg arg) return ts_info(arg_temp(arg)); } +static inline bool ti_is_const(TempOptInfo *ti) +{ + return ti->is_const; +} + +static inline uint64_t ti_const_val(TempOptInfo *ti) +{ + return ti->val; +} + +static inline bool ti_is_const_val(TempOptInfo *ti, uint64_t val) +{ + return ti_is_const(ti) && ti_const_val(ti) == val; +} + static inline bool ts_is_const(TCGTemp *ts) { - return ts_info(ts)->is_const; + return ti_is_const(ts_info(ts)); } static inline bool ts_is_const_val(TCGTemp *ts, uint64_t val) { - TempOptInfo *ti = ts_info(ts); - return ti->is_const && ti->val == val; + return ti_is_const_val(ts_info(ts), val); } static inline bool arg_is_const(TCGArg arg) -- 2.43.0 Avoid the use of the OptContext slots. Find TempOptInfo once. Sink mask computation below fold_affected_mask early exit. Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- tcg/optimize.c | 30 ++++++++++++++++-------------- 1 file changed, 16 insertions(+), 14 deletions(-) diff --git a/tcg/optimize.c b/tcg/optimize.c index XXXXXXX..XXXXXXX 100644 --- a/tcg/optimize.c +++ b/tcg/optimize.c @@ -XXX,XX +XXX,XX @@ static bool fold_add2(OptContext *ctx, TCGOp *op) static bool fold_and(OptContext *ctx, TCGOp *op) { - uint64_t z1, z2; + uint64_t z1, z2, z_mask, s_mask; + TempOptInfo *t1, *t2; if (fold_const2_commutative(ctx, op) || fold_xi_to_i(ctx, op, 0) || @@ -XXX,XX +XXX,XX @@ static bool fold_and(OptContext *ctx, TCGOp *op) return true; } - z1 = arg_info(op->args[1])->z_mask; - z2 = arg_info(op->args[2])->z_mask; - ctx->z_mask = z1 & z2; - - /* - * Sign repetitions are perforce all identical, whether they are 1 or 0. - * Bitwise operations preserve the relative quantity of the repetitions. - */ - ctx->s_mask = arg_info(op->args[1])->s_mask - & arg_info(op->args[2])->s_mask; + t1 = arg_info(op->args[1]); + t2 = arg_info(op->args[2]); + z1 = t1->z_mask; + z2 = t2->z_mask; /* * Known-zeros does not imply known-ones. Therefore unless * arg2 is constant, we can't infer affected bits from it. */ - if (arg_is_const(op->args[2]) && - fold_affected_mask(ctx, op, z1 & ~z2)) { + if (ti_is_const(t2) && fold_affected_mask(ctx, op, z1 & ~z2)) { return true; } - return fold_masks(ctx, op); + z_mask = z1 & z2; + + /* + * Sign repetitions are perforce all identical, whether they are 1 or 0. + * Bitwise operations preserve the relative quantity of the repetitions. + */ + s_mask = t1->s_mask & t2->s_mask; + + return fold_masks_zs(ctx, op, z_mask, s_mask); } static bool fold_andc(OptContext *ctx, TCGOp *op) -- 2.43.0 Avoid the use of the OptContext slots. Find TempOptInfo once. Avoid double inversion of the value of second const operand. Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- tcg/optimize.c | 21 +++++++++++---------- 1 file changed, 11 insertions(+), 10 deletions(-) diff --git a/tcg/optimize.c b/tcg/optimize.c index XXXXXXX..XXXXXXX 100644 --- a/tcg/optimize.c +++ b/tcg/optimize.c @@ -XXX,XX +XXX,XX @@ static bool fold_and(OptContext *ctx, TCGOp *op) static bool fold_andc(OptContext *ctx, TCGOp *op) { - uint64_t z1; + uint64_t z_mask, s_mask; + TempOptInfo *t1, *t2; if (fold_const2(ctx, op) || fold_xx_to_i(ctx, op, 0) || @@ -XXX,XX +XXX,XX @@ static bool fold_andc(OptContext *ctx, TCGOp *op) return true; } - z1 = arg_info(op->args[1])->z_mask; + t1 = arg_info(op->args[1]); + t2 = arg_info(op->args[2]); + z_mask = t1->z_mask; /* * Known-zeros does not imply known-ones. Therefore unless * arg2 is constant, we can't infer anything from it. */ - if (arg_is_const(op->args[2])) { - uint64_t z2 = ~arg_info(op->args[2])->z_mask; - if (fold_affected_mask(ctx, op, z1 & ~z2)) { + if (ti_is_const(t2)) { + uint64_t v2 = ti_const_val(t2); + if (fold_affected_mask(ctx, op, z_mask & v2)) { return true; } - z1 &= z2; + z_mask &= ~v2; } - ctx->z_mask = z1; - ctx->s_mask = arg_info(op->args[1])->s_mask - & arg_info(op->args[2])->s_mask; - return fold_masks(ctx, op); + s_mask = t1->s_mask & t2->s_mask; + return fold_masks_zs(ctx, op, z_mask, s_mask); } static bool fold_brcond(OptContext *ctx, TCGOp *op) -- 2.43.0 Avoid the use of the OptContext slots. Find TempOptInfo once. Always set s_mask along the BSWAP_OS path, since the result is being explicitly sign-extended. Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- tcg/optimize.c | 21 ++++++++++----------- 1 file changed, 10 insertions(+), 11 deletions(-) diff --git a/tcg/optimize.c b/tcg/optimize.c index XXXXXXX..XXXXXXX 100644 --- a/tcg/optimize.c +++ b/tcg/optimize.c @@ -XXX,XX +XXX,XX @@ static bool fold_brcond2(OptContext *ctx, TCGOp *op) static bool fold_bswap(OptContext *ctx, TCGOp *op) { uint64_t z_mask, s_mask, sign; + TempOptInfo *t1 = arg_info(op->args[1]); - if (arg_is_const(op->args[1])) { - uint64_t t = arg_info(op->args[1])->val; - - t = do_constant_folding(op->opc, ctx->type, t, op->args[2]); - return tcg_opt_gen_movi(ctx, op, op->args[0], t); + if (ti_is_const(t1)) { + return tcg_opt_gen_movi(ctx, op, op->args[0], + do_constant_folding(op->opc, ctx->type, + ti_const_val(t1), + op->args[2])); } - z_mask = arg_info(op->args[1])->z_mask; - + z_mask = t1->z_mask; switch (op->opc) { case INDEX_op_bswap16_i32: case INDEX_op_bswap16_i64: @@ -XXX,XX +XXX,XX @@ static bool fold_bswap(OptContext *ctx, TCGOp *op) /* If the sign bit may be 1, force all the bits above to 1. */ if (z_mask & sign) { z_mask |= sign; - s_mask = sign << 1; } + /* The value and therefore s_mask is explicitly sign-extended. */ + s_mask = sign; break; default: /* The high bits are undefined: force all bits above the sign to 1. */ z_mask |= sign << 1; break; } - ctx->z_mask = z_mask; - ctx->s_mask = s_mask; - return fold_masks(ctx, op); + return fold_masks_zs(ctx, op, z_mask, s_mask); } static bool fold_call(OptContext *ctx, TCGOp *op) -- 2.43.0 Avoid the use of the OptContext slots. Find TempOptInfo once. Compute s_mask from the union of the maximum count and the op2 fallback for op1 being zero. Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- tcg/optimize.c | 15 ++++++++++----- 1 file changed, 10 insertions(+), 5 deletions(-) diff --git a/tcg/optimize.c b/tcg/optimize.c index XXXXXXX..XXXXXXX 100644 --- a/tcg/optimize.c +++ b/tcg/optimize.c @@ -XXX,XX +XXX,XX @@ static bool fold_call(OptContext *ctx, TCGOp *op) static bool fold_count_zeros(OptContext *ctx, TCGOp *op) { - uint64_t z_mask; + uint64_t z_mask, s_mask; + TempOptInfo *t1 = arg_info(op->args[1]); + TempOptInfo *t2 = arg_info(op->args[2]); - if (arg_is_const(op->args[1])) { - uint64_t t = arg_info(op->args[1])->val; + if (ti_is_const(t1)) { + uint64_t t = ti_const_val(t1); if (t != 0) { t = do_constant_folding(op->opc, ctx->type, t, 0); @@ -XXX,XX +XXX,XX @@ static bool fold_count_zeros(OptContext *ctx, TCGOp *op) default: g_assert_not_reached(); } - ctx->z_mask = arg_info(op->args[2])->z_mask | z_mask; - return false; + s_mask = ~z_mask; + z_mask |= t2->z_mask; + s_mask &= t2->s_mask; + + return fold_masks_zs(ctx, op, z_mask, s_mask); } static bool fold_ctpop(OptContext *ctx, TCGOp *op) -- 2.43.0 Add fold_masks_z as a trivial wrapper around fold_masks_zs. Avoid the use of the OptContext slots. Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- tcg/optimize.c | 13 ++++++++++--- 1 file changed, 10 insertions(+), 3 deletions(-) diff --git a/tcg/optimize.c b/tcg/optimize.c index XXXXXXX..XXXXXXX 100644 --- a/tcg/optimize.c +++ b/tcg/optimize.c @@ -XXX,XX +XXX,XX @@ static bool fold_masks_zs(OptContext *ctx, TCGOp *op, return true; } +static bool fold_masks_z(OptContext *ctx, TCGOp *op, uint64_t z_mask) +{ + return fold_masks_zs(ctx, op, z_mask, 0); +} + static bool fold_masks(OptContext *ctx, TCGOp *op) { return fold_masks_zs(ctx, op, ctx->z_mask, ctx->s_mask); @@ -XXX,XX +XXX,XX @@ static bool fold_count_zeros(OptContext *ctx, TCGOp *op) static bool fold_ctpop(OptContext *ctx, TCGOp *op) { + uint64_t z_mask; + if (fold_const1(ctx, op)) { return true; } switch (ctx->type) { case TCG_TYPE_I32: - ctx->z_mask = 32 | 31; + z_mask = 32 | 31; break; case TCG_TYPE_I64: - ctx->z_mask = 64 | 63; + z_mask = 64 | 63; break; default: g_assert_not_reached(); } - return false; + return fold_masks_z(ctx, op, z_mask); } static bool fold_deposit(OptContext *ctx, TCGOp *op) -- 2.43.0 Avoid the use of the OptContext slots. Find TempOptInfo once. When we fold to and, use fold_and. Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- tcg/optimize.c | 35 +++++++++++++++++------------------ 1 file changed, 17 insertions(+), 18 deletions(-) diff --git a/tcg/optimize.c b/tcg/optimize.c index XXXXXXX..XXXXXXX 100644 --- a/tcg/optimize.c +++ b/tcg/optimize.c @@ -XXX,XX +XXX,XX @@ static bool fold_ctpop(OptContext *ctx, TCGOp *op) static bool fold_deposit(OptContext *ctx, TCGOp *op) { + TempOptInfo *t1 = arg_info(op->args[1]); + TempOptInfo *t2 = arg_info(op->args[2]); + int ofs = op->args[3]; + int len = op->args[4]; TCGOpcode and_opc; + uint64_t z_mask; - if (arg_is_const(op->args[1]) && arg_is_const(op->args[2])) { - uint64_t t1 = arg_info(op->args[1])->val; - uint64_t t2 = arg_info(op->args[2])->val; - - t1 = deposit64(t1, op->args[3], op->args[4], t2); - return tcg_opt_gen_movi(ctx, op, op->args[0], t1); + if (ti_is_const(t1) && ti_is_const(t2)) { + return tcg_opt_gen_movi(ctx, op, op->args[0], + deposit64(ti_const_val(t1), ofs, len, + ti_const_val(t2))); } switch (ctx->type) { @@ -XXX,XX +XXX,XX @@ static bool fold_deposit(OptContext *ctx, TCGOp *op) } /* Inserting a value into zero at offset 0. */ - if (arg_is_const_val(op->args[1], 0) && op->args[3] == 0) { - uint64_t mask = MAKE_64BIT_MASK(0, op->args[4]); + if (ti_is_const_val(t1, 0) && ofs == 0) { + uint64_t mask = MAKE_64BIT_MASK(0, len); op->opc = and_opc; op->args[1] = op->args[2]; op->args[2] = arg_new_constant(ctx, mask); - ctx->z_mask = mask & arg_info(op->args[1])->z_mask; - return false; + return fold_and(ctx, op); } /* Inserting zero into a value. */ - if (arg_is_const_val(op->args[2], 0)) { - uint64_t mask = deposit64(-1, op->args[3], op->args[4], 0); + if (ti_is_const_val(t2, 0)) { + uint64_t mask = deposit64(-1, ofs, len, 0); op->opc = and_opc; op->args[2] = arg_new_constant(ctx, mask); - ctx->z_mask = mask & arg_info(op->args[1])->z_mask; - return false; + return fold_and(ctx, op); } - ctx->z_mask = deposit64(arg_info(op->args[1])->z_mask, - op->args[3], op->args[4], - arg_info(op->args[2])->z_mask); - return false; + z_mask = deposit64(t1->z_mask, ofs, len, t2->z_mask); + return fold_masks_z(ctx, op, z_mask); } static bool fold_divide(OptContext *ctx, TCGOp *op) -- 2.43.0 The input which overlaps the sign bit of the output can have its input s_mask propagated to the output s_mask. Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- tcg/optimize.c | 14 ++++++++++++-- 1 file changed, 12 insertions(+), 2 deletions(-) diff --git a/tcg/optimize.c b/tcg/optimize.c index XXXXXXX..XXXXXXX 100644 --- a/tcg/optimize.c +++ b/tcg/optimize.c @@ -XXX,XX +XXX,XX @@ static bool fold_deposit(OptContext *ctx, TCGOp *op) TempOptInfo *t2 = arg_info(op->args[2]); int ofs = op->args[3]; int len = op->args[4]; + int width; TCGOpcode and_opc; - uint64_t z_mask; + uint64_t z_mask, s_mask; if (ti_is_const(t1) && ti_is_const(t2)) { return tcg_opt_gen_movi(ctx, op, op->args[0], @@ -XXX,XX +XXX,XX @@ static bool fold_deposit(OptContext *ctx, TCGOp *op) switch (ctx->type) { case TCG_TYPE_I32: and_opc = INDEX_op_and_i32; + width = 32; break; case TCG_TYPE_I64: and_opc = INDEX_op_and_i64; + width = 64; break; default: g_assert_not_reached(); @@ -XXX,XX +XXX,XX @@ static bool fold_deposit(OptContext *ctx, TCGOp *op) return fold_and(ctx, op); } + /* The s_mask from the top portion of the deposit is still valid. */ + if (ofs + len == width) { + s_mask = t2->s_mask << ofs; + } else { + s_mask = t1->s_mask & ~MAKE_64BIT_MASK(0, ofs + len); + } + z_mask = deposit64(t1->z_mask, ofs, len, t2->z_mask); - return fold_masks_z(ctx, op, z_mask); + return fold_masks_zs(ctx, op, z_mask, s_mask); } static bool fold_divide(OptContext *ctx, TCGOp *op) -- 2.43.0 Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- tcg/optimize.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/tcg/optimize.c b/tcg/optimize.c index XXXXXXX..XXXXXXX 100644 --- a/tcg/optimize.c +++ b/tcg/optimize.c @@ -XXX,XX +XXX,XX @@ static bool fold_divide(OptContext *ctx, TCGOp *op) fold_xi_to_x(ctx, op, 1)) { return true; } - return false; + return finish_folding(ctx, op); } static bool fold_dup(OptContext *ctx, TCGOp *op) -- 2.43.0 Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- tcg/optimize.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/tcg/optimize.c b/tcg/optimize.c index XXXXXXX..XXXXXXX 100644 --- a/tcg/optimize.c +++ b/tcg/optimize.c @@ -XXX,XX +XXX,XX @@ static bool fold_dup(OptContext *ctx, TCGOp *op) t = dup_const(TCGOP_VECE(op), t); return tcg_opt_gen_movi(ctx, op, op->args[0], t); } - return false; + return finish_folding(ctx, op); } static bool fold_dup2(OptContext *ctx, TCGOp *op) @@ -XXX,XX +XXX,XX @@ static bool fold_dup2(OptContext *ctx, TCGOp *op) op->opc = INDEX_op_dup_vec; TCGOP_VECE(op) = MO_32; } - return false; + return finish_folding(ctx, op); } static bool fold_eqv(OptContext *ctx, TCGOp *op) -- 2.43.0 Add fold_masks_s as a trivial wrapper around fold_masks_zs. Avoid the use of the OptContext slots. Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- tcg/optimize.c | 13 ++++++++++--- 1 file changed, 10 insertions(+), 3 deletions(-) diff --git a/tcg/optimize.c b/tcg/optimize.c index XXXXXXX..XXXXXXX 100644 --- a/tcg/optimize.c +++ b/tcg/optimize.c @@ -XXX,XX +XXX,XX @@ static bool fold_masks_z(OptContext *ctx, TCGOp *op, uint64_t z_mask) return fold_masks_zs(ctx, op, z_mask, 0); } +static bool fold_masks_s(OptContext *ctx, TCGOp *op, uint64_t s_mask) +{ + return fold_masks_zs(ctx, op, -1, s_mask); +} + static bool fold_masks(OptContext *ctx, TCGOp *op) { return fold_masks_zs(ctx, op, ctx->z_mask, ctx->s_mask); @@ -XXX,XX +XXX,XX @@ static bool fold_dup2(OptContext *ctx, TCGOp *op) static bool fold_eqv(OptContext *ctx, TCGOp *op) { + uint64_t s_mask; + if (fold_const2_commutative(ctx, op) || fold_xi_to_x(ctx, op, -1) || fold_xi_to_not(ctx, op, 0)) { return true; } - ctx->s_mask = arg_info(op->args[1])->s_mask - & arg_info(op->args[2])->s_mask; - return false; + s_mask = arg_info(op->args[1])->s_mask + & arg_info(op->args[2])->s_mask; + return fold_masks_s(ctx, op, s_mask); } static bool fold_extract(OptContext *ctx, TCGOp *op) -- 2.43.0 Avoid the use of the OptContext slots. Find TempOptInfo once. Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- tcg/optimize.c | 15 ++++++--------- 1 file changed, 6 insertions(+), 9 deletions(-) diff --git a/tcg/optimize.c b/tcg/optimize.c index XXXXXXX..XXXXXXX 100644 --- a/tcg/optimize.c +++ b/tcg/optimize.c @@ -XXX,XX +XXX,XX @@ static bool fold_eqv(OptContext *ctx, TCGOp *op) static bool fold_extract(OptContext *ctx, TCGOp *op) { uint64_t z_mask_old, z_mask; + TempOptInfo *t1 = arg_info(op->args[1]); int pos = op->args[2]; int len = op->args[3]; - if (arg_is_const(op->args[1])) { - uint64_t t; - - t = arg_info(op->args[1])->val; - t = extract64(t, pos, len); - return tcg_opt_gen_movi(ctx, op, op->args[0], t); + if (ti_is_const(t1)) { + return tcg_opt_gen_movi(ctx, op, op->args[0], + extract64(ti_const_val(t1), pos, len)); } - z_mask_old = arg_info(op->args[1])->z_mask; + z_mask_old = t1->z_mask; z_mask = extract64(z_mask_old, pos, len); if (pos == 0 && fold_affected_mask(ctx, op, z_mask_old ^ z_mask)) { return true; } - ctx->z_mask = z_mask; - return fold_masks(ctx, op); + return fold_masks_z(ctx, op, z_mask); } static bool fold_extract2(OptContext *ctx, TCGOp *op) -- 2.43.0 Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- tcg/optimize.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/tcg/optimize.c b/tcg/optimize.c index XXXXXXX..XXXXXXX 100644 --- a/tcg/optimize.c +++ b/tcg/optimize.c @@ -XXX,XX +XXX,XX @@ static bool fold_extract2(OptContext *ctx, TCGOp *op) } return tcg_opt_gen_movi(ctx, op, op->args[0], v1 | v2); } - return false; + return finish_folding(ctx, op); } static bool fold_exts(OptContext *ctx, TCGOp *op) -- 2.43.0 Avoid the use of the OptContext slots. Find TempOptInfo once. Explicitly sign-extend z_mask instead of doing that manually. Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- tcg/optimize.c | 29 ++++++++++++----------------- 1 file changed, 12 insertions(+), 17 deletions(-) diff --git a/tcg/optimize.c b/tcg/optimize.c index XXXXXXX..XXXXXXX 100644 --- a/tcg/optimize.c +++ b/tcg/optimize.c @@ -XXX,XX +XXX,XX @@ static bool fold_extract2(OptContext *ctx, TCGOp *op) static bool fold_exts(OptContext *ctx, TCGOp *op) { - uint64_t s_mask_old, s_mask, z_mask, sign; + uint64_t s_mask_old, s_mask, z_mask; bool type_change = false; + TempOptInfo *t1; if (fold_const1(ctx, op)) { return true; } - z_mask = arg_info(op->args[1])->z_mask; - s_mask = arg_info(op->args[1])->s_mask; + t1 = arg_info(op->args[1]); + z_mask = t1->z_mask; + s_mask = t1->s_mask; s_mask_old = s_mask; switch (op->opc) { CASE_OP_32_64(ext8s): - sign = INT8_MIN; - z_mask = (uint8_t)z_mask; + s_mask |= INT8_MIN; + z_mask = (int8_t)z_mask; break; CASE_OP_32_64(ext16s): - sign = INT16_MIN; - z_mask = (uint16_t)z_mask; + s_mask |= INT16_MIN; + z_mask = (int16_t)z_mask; break; case INDEX_op_ext_i32_i64: type_change = true; QEMU_FALLTHROUGH; case INDEX_op_ext32s_i64: - sign = INT32_MIN; - z_mask = (uint32_t)z_mask; + s_mask |= INT32_MIN; + z_mask = (int32_t)z_mask; break; default: g_assert_not_reached(); } - if (z_mask & sign) { - z_mask |= sign; - } - s_mask |= sign << 1; - - ctx->z_mask = z_mask; - ctx->s_mask = s_mask; if (0 && !type_change && fold_affected_mask(ctx, op, s_mask & ~s_mask_old)) { return true; } - return fold_masks(ctx, op); + return fold_masks_zs(ctx, op, z_mask, s_mask); } static bool fold_extu(OptContext *ctx, TCGOp *op) -- 2.43.0 Avoid the use of the OptContext slots. Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- tcg/optimize.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/tcg/optimize.c b/tcg/optimize.c index XXXXXXX..XXXXXXX 100644 --- a/tcg/optimize.c +++ b/tcg/optimize.c @@ -XXX,XX +XXX,XX @@ static bool fold_extu(OptContext *ctx, TCGOp *op) g_assert_not_reached(); } - ctx->z_mask = z_mask; if (!type_change && fold_affected_mask(ctx, op, z_mask_old ^ z_mask)) { return true; } - return fold_masks(ctx, op); + + return fold_masks_z(ctx, op, z_mask); } static bool fold_mb(OptContext *ctx, TCGOp *op) -- 2.43.0 Avoid the use of the OptContext slots. Find TempOptInfo once. Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- tcg/optimize.c | 19 +++++++++++-------- 1 file changed, 11 insertions(+), 8 deletions(-) diff --git a/tcg/optimize.c b/tcg/optimize.c index XXXXXXX..XXXXXXX 100644 --- a/tcg/optimize.c +++ b/tcg/optimize.c @@ -XXX,XX +XXX,XX @@ static bool fold_mov(OptContext *ctx, TCGOp *op) static bool fold_movcond(OptContext *ctx, TCGOp *op) { + uint64_t z_mask, s_mask; + TempOptInfo *tt, *ft; int i; /* If true and false values are the same, eliminate the cmp. */ @@ -XXX,XX +XXX,XX @@ static bool fold_movcond(OptContext *ctx, TCGOp *op) return tcg_opt_gen_mov(ctx, op, op->args[0], op->args[4 - i]); } - ctx->z_mask = arg_info(op->args[3])->z_mask - | arg_info(op->args[4])->z_mask; - ctx->s_mask = arg_info(op->args[3])->s_mask - & arg_info(op->args[4])->s_mask; + tt = arg_info(op->args[3]); + ft = arg_info(op->args[4]); + z_mask = tt->z_mask | ft->z_mask; + s_mask = tt->s_mask & ft->s_mask; - if (arg_is_const(op->args[3]) && arg_is_const(op->args[4])) { - uint64_t tv = arg_info(op->args[3])->val; - uint64_t fv = arg_info(op->args[4])->val; + if (ti_is_const(tt) && ti_is_const(ft)) { + uint64_t tv = ti_const_val(tt); + uint64_t fv = ti_const_val(ft); TCGOpcode opc, negopc = 0; TCGCond cond = op->args[5]; @@ -XXX,XX +XXX,XX @@ static bool fold_movcond(OptContext *ctx, TCGOp *op) } } } - return false; + + return fold_masks_zs(ctx, op, z_mask, s_mask); } static bool fold_mul(OptContext *ctx, TCGOp *op) -- 2.43.0 Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- tcg/optimize.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) diff --git a/tcg/optimize.c b/tcg/optimize.c index XXXXXXX..XXXXXXX 100644 --- a/tcg/optimize.c +++ b/tcg/optimize.c @@ -XXX,XX +XXX,XX @@ static bool fold_mul(OptContext *ctx, TCGOp *op) fold_xi_to_x(ctx, op, 1)) { return true; } - return false; + return finish_folding(ctx, op); } static bool fold_mul_highpart(OptContext *ctx, TCGOp *op) @@ -XXX,XX +XXX,XX @@ static bool fold_mul_highpart(OptContext *ctx, TCGOp *op) fold_xi_to_i(ctx, op, 0)) { return true; } - return false; + return finish_folding(ctx, op); } static bool fold_multiply2(OptContext *ctx, TCGOp *op) @@ -XXX,XX +XXX,XX @@ static bool fold_multiply2(OptContext *ctx, TCGOp *op) tcg_opt_gen_movi(ctx, op2, rh, h); return true; } - return false; + return finish_folding(ctx, op); } static bool fold_nand(OptContext *ctx, TCGOp *op) -- 2.43.0 Avoid the use of the OptContext slots. Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- tcg/optimize.c | 8 +++++--- 1 file changed, 5 insertions(+), 3 deletions(-) diff --git a/tcg/optimize.c b/tcg/optimize.c index XXXXXXX..XXXXXXX 100644 --- a/tcg/optimize.c +++ b/tcg/optimize.c @@ -XXX,XX +XXX,XX @@ static bool fold_multiply2(OptContext *ctx, TCGOp *op) static bool fold_nand(OptContext *ctx, TCGOp *op) { + uint64_t s_mask; + if (fold_const2_commutative(ctx, op) || fold_xi_to_not(ctx, op, -1)) { return true; } - ctx->s_mask = arg_info(op->args[1])->s_mask - & arg_info(op->args[2])->s_mask; - return false; + s_mask = arg_info(op->args[1])->s_mask + & arg_info(op->args[2])->s_mask; + return fold_masks_s(ctx, op, s_mask); } static bool fold_neg_no_const(OptContext *ctx, TCGOp *op) -- 2.43.0 Avoid the use of the OptContext slots. Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- tcg/optimize.c | 9 ++------- 1 file changed, 2 insertions(+), 7 deletions(-) diff --git a/tcg/optimize.c b/tcg/optimize.c index XXXXXXX..XXXXXXX 100644 --- a/tcg/optimize.c +++ b/tcg/optimize.c @@ -XXX,XX +XXX,XX @@ static bool fold_neg_no_const(OptContext *ctx, TCGOp *op) { /* Set to 1 all bits to the left of the rightmost. */ uint64_t z_mask = arg_info(op->args[1])->z_mask; - ctx->z_mask = -(z_mask & -z_mask); + z_mask = -(z_mask & -z_mask); - /* - * Because of fold_sub_to_neg, we want to always return true, - * via finish_folding. - */ - finish_folding(ctx, op); - return true; + return fold_masks_z(ctx, op, z_mask); } static bool fold_neg(OptContext *ctx, TCGOp *op) -- 2.43.0 Avoid the use of the OptContext slots. Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- tcg/optimize.c | 8 +++++--- 1 file changed, 5 insertions(+), 3 deletions(-) diff --git a/tcg/optimize.c b/tcg/optimize.c index XXXXXXX..XXXXXXX 100644 --- a/tcg/optimize.c +++ b/tcg/optimize.c @@ -XXX,XX +XXX,XX @@ static bool fold_neg(OptContext *ctx, TCGOp *op) static bool fold_nor(OptContext *ctx, TCGOp *op) { + uint64_t s_mask; + if (fold_const2_commutative(ctx, op) || fold_xi_to_not(ctx, op, 0)) { return true; } - ctx->s_mask = arg_info(op->args[1])->s_mask - & arg_info(op->args[2])->s_mask; - return false; + s_mask = arg_info(op->args[1])->s_mask + & arg_info(op->args[2])->s_mask; + return fold_masks_s(ctx, op, s_mask); } static bool fold_not(OptContext *ctx, TCGOp *op) -- 2.43.0 Avoid the use of the OptContext slots. Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- tcg/optimize.c | 7 +------ 1 file changed, 1 insertion(+), 6 deletions(-) diff --git a/tcg/optimize.c b/tcg/optimize.c index XXXXXXX..XXXXXXX 100644 --- a/tcg/optimize.c +++ b/tcg/optimize.c @@ -XXX,XX +XXX,XX @@ static bool fold_not(OptContext *ctx, TCGOp *op) if (fold_const1(ctx, op)) { return true; } - - ctx->s_mask = arg_info(op->args[1])->s_mask; - - /* Because of fold_to_not, we want to always return true, via finish. */ - finish_folding(ctx, op); - return true; + return fold_masks_s(ctx, op, arg_info(op->args[1])->s_mask); } static bool fold_or(OptContext *ctx, TCGOp *op) -- 2.43.0 Avoid the use of the OptContext slots. Find TempOptInfo once. Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- tcg/optimize.c | 13 ++++++++----- 1 file changed, 8 insertions(+), 5 deletions(-) diff --git a/tcg/optimize.c b/tcg/optimize.c index XXXXXXX..XXXXXXX 100644 --- a/tcg/optimize.c +++ b/tcg/optimize.c @@ -XXX,XX +XXX,XX @@ static bool fold_not(OptContext *ctx, TCGOp *op) static bool fold_or(OptContext *ctx, TCGOp *op) { + uint64_t z_mask, s_mask; + TempOptInfo *t1, *t2; + if (fold_const2_commutative(ctx, op) || fold_xi_to_x(ctx, op, 0) || fold_xx_to_x(ctx, op)) { return true; } - ctx->z_mask = arg_info(op->args[1])->z_mask - | arg_info(op->args[2])->z_mask; - ctx->s_mask = arg_info(op->args[1])->s_mask - & arg_info(op->args[2])->s_mask; - return fold_masks(ctx, op); + t1 = arg_info(op->args[1]); + t2 = arg_info(op->args[2]); + z_mask = t1->z_mask | t2->z_mask; + s_mask = t1->s_mask & t2->s_mask; + return fold_masks_zs(ctx, op, z_mask, s_mask); } static bool fold_orc(OptContext *ctx, TCGOp *op) -- 2.43.0 Avoid the use of the OptContext slots. Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- tcg/optimize.c | 8 +++++--- 1 file changed, 5 insertions(+), 3 deletions(-) diff --git a/tcg/optimize.c b/tcg/optimize.c index XXXXXXX..XXXXXXX 100644 --- a/tcg/optimize.c +++ b/tcg/optimize.c @@ -XXX,XX +XXX,XX @@ static bool fold_or(OptContext *ctx, TCGOp *op) static bool fold_orc(OptContext *ctx, TCGOp *op) { + uint64_t s_mask; + if (fold_const2(ctx, op) || fold_xx_to_i(ctx, op, -1) || fold_xi_to_x(ctx, op, -1) || @@ -XXX,XX +XXX,XX @@ static bool fold_orc(OptContext *ctx, TCGOp *op) return true; } - ctx->s_mask = arg_info(op->args[1])->s_mask - & arg_info(op->args[2])->s_mask; - return false; + s_mask = arg_info(op->args[1])->s_mask + & arg_info(op->args[2])->s_mask; + return fold_masks_s(ctx, op, s_mask); } static bool fold_qemu_ld(OptContext *ctx, TCGOp *op) -- 2.43.0 Avoid the use of the OptContext slots. Be careful not to call fold_masks_zs when the memory operation is wide enough to require multiple outputs, so split into two functions: fold_qemu_ld_1reg and fold_qemu_ld_2reg. Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- tcg/optimize.c | 26 +++++++++++++++++++++----- 1 file changed, 21 insertions(+), 5 deletions(-) diff --git a/tcg/optimize.c b/tcg/optimize.c index XXXXXXX..XXXXXXX 100644 --- a/tcg/optimize.c +++ b/tcg/optimize.c @@ -XXX,XX +XXX,XX @@ static bool fold_orc(OptContext *ctx, TCGOp *op) return fold_masks_s(ctx, op, s_mask); } -static bool fold_qemu_ld(OptContext *ctx, TCGOp *op) +static bool fold_qemu_ld_1reg(OptContext *ctx, TCGOp *op) { const TCGOpDef *def = &tcg_op_defs[op->opc]; MemOpIdx oi = op->args[def->nb_oargs + def->nb_iargs]; MemOp mop = get_memop(oi); int width = 8 * memop_size(mop); + uint64_t z_mask = -1, s_mask = 0; if (width < 64) { if (mop & MO_SIGN) { - ctx->s_mask = MAKE_64BIT_MASK(width, 64 - width); + s_mask = MAKE_64BIT_MASK(width - 1, 64 - (width - 1)); } else { - ctx->z_mask = MAKE_64BIT_MASK(0, width); + z_mask = MAKE_64BIT_MASK(0, width); } } /* Opcodes that touch guest memory stop the mb optimization. */ ctx->prev_mb = NULL; - return false; + + return fold_masks_zs(ctx, op, z_mask, s_mask); +} + +static bool fold_qemu_ld_2reg(OptContext *ctx, TCGOp *op) +{ + /* Opcodes that touch guest memory stop the mb optimization. */ + ctx->prev_mb = NULL; + return finish_folding(ctx, op); } static bool fold_qemu_st(OptContext *ctx, TCGOp *op) @@ -XXX,XX +XXX,XX @@ void tcg_optimize(TCGContext *s) break; case INDEX_op_qemu_ld_a32_i32: case INDEX_op_qemu_ld_a64_i32: + done = fold_qemu_ld_1reg(&ctx, op); + break; case INDEX_op_qemu_ld_a32_i64: case INDEX_op_qemu_ld_a64_i64: + if (TCG_TARGET_REG_BITS == 64) { + done = fold_qemu_ld_1reg(&ctx, op); + break; + } + QEMU_FALLTHROUGH; case INDEX_op_qemu_ld_a32_i128: case INDEX_op_qemu_ld_a64_i128: - done = fold_qemu_ld(&ctx, op); + done = fold_qemu_ld_2reg(&ctx, op); break; case INDEX_op_qemu_st8_a32_i32: case INDEX_op_qemu_st8_a64_i32: -- 2.43.0 Stores have no output operands, and so need no further work. Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- tcg/optimize.c | 11 +++++------ 1 file changed, 5 insertions(+), 6 deletions(-) diff --git a/tcg/optimize.c b/tcg/optimize.c index XXXXXXX..XXXXXXX 100644 --- a/tcg/optimize.c +++ b/tcg/optimize.c @@ -XXX,XX +XXX,XX @@ static bool fold_qemu_st(OptContext *ctx, TCGOp *op) { /* Opcodes that touch guest memory stop the mb optimization. */ ctx->prev_mb = NULL; - return false; + return true; } static bool fold_remainder(OptContext *ctx, TCGOp *op) @@ -XXX,XX +XXX,XX @@ static bool fold_tcg_st(OptContext *ctx, TCGOp *op) if (op->args[1] != tcgv_ptr_arg(tcg_env)) { remove_mem_copy_all(ctx); - return false; + return true; } switch (op->opc) { @@ -XXX,XX +XXX,XX @@ static bool fold_tcg_st(OptContext *ctx, TCGOp *op) g_assert_not_reached(); } remove_mem_copy_in(ctx, ofs, ofs + lm1); - return false; + return true; } static bool fold_tcg_st_memcopy(OptContext *ctx, TCGOp *op) @@ -XXX,XX +XXX,XX @@ static bool fold_tcg_st_memcopy(OptContext *ctx, TCGOp *op) TCGType type; if (op->args[1] != tcgv_ptr_arg(tcg_env)) { - fold_tcg_st(ctx, op); - return false; + return fold_tcg_st(ctx, op); } src = arg_temp(op->args[0]); @@ -XXX,XX +XXX,XX @@ static bool fold_tcg_st_memcopy(OptContext *ctx, TCGOp *op) last = ofs + tcg_type_size(type) - 1; remove_mem_copy_in(ctx, ofs, last); record_mem_copy(ctx, type, src, ofs, last); - return false; + return true; } static bool fold_xor(OptContext *ctx, TCGOp *op) -- 2.43.0 Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- tcg/optimize.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/tcg/optimize.c b/tcg/optimize.c index XXXXXXX..XXXXXXX 100644 --- a/tcg/optimize.c +++ b/tcg/optimize.c @@ -XXX,XX +XXX,XX @@ static bool fold_remainder(OptContext *ctx, TCGOp *op) fold_xx_to_i(ctx, op, 0)) { return true; } - return false; + return finish_folding(ctx, op); } static bool fold_setcond_zmask(OptContext *ctx, TCGOp *op, bool neg) -- 2.43.0 Change return from bool to int; distinguish between complete folding, simplification, and no change. Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- tcg/optimize.c | 22 ++++++++++++++-------- 1 file changed, 14 insertions(+), 8 deletions(-) diff --git a/tcg/optimize.c b/tcg/optimize.c index XXXXXXX..XXXXXXX 100644 --- a/tcg/optimize.c +++ b/tcg/optimize.c @@ -XXX,XX +XXX,XX @@ static bool fold_remainder(OptContext *ctx, TCGOp *op) return finish_folding(ctx, op); } -static bool fold_setcond_zmask(OptContext *ctx, TCGOp *op, bool neg) +/* Return 1 if finished, -1 if simplified, 0 if unchanged. */ +static int fold_setcond_zmask(OptContext *ctx, TCGOp *op, bool neg) { uint64_t a_zmask, b_val; TCGCond cond; @@ -XXX,XX +XXX,XX @@ static bool fold_setcond_zmask(OptContext *ctx, TCGOp *op, bool neg) op->opc = xor_opc; op->args[2] = arg_new_constant(ctx, 1); } - return false; + return -1; } } - - return false; + return 0; } static void fold_setcond_tst_pow2(OptContext *ctx, TCGOp *op, bool neg) @@ -XXX,XX +XXX,XX @@ static bool fold_setcond(OptContext *ctx, TCGOp *op) return tcg_opt_gen_movi(ctx, op, op->args[0], i); } - if (fold_setcond_zmask(ctx, op, false)) { + i = fold_setcond_zmask(ctx, op, false); + if (i > 0) { return true; } - fold_setcond_tst_pow2(ctx, op, false); + if (i == 0) { + fold_setcond_tst_pow2(ctx, op, false); + } ctx->z_mask = 1; return false; @@ -XXX,XX +XXX,XX @@ static bool fold_negsetcond(OptContext *ctx, TCGOp *op) return tcg_opt_gen_movi(ctx, op, op->args[0], -i); } - if (fold_setcond_zmask(ctx, op, true)) { + i = fold_setcond_zmask(ctx, op, true); + if (i > 0) { return true; } - fold_setcond_tst_pow2(ctx, op, true); + if (i == 0) { + fold_setcond_tst_pow2(ctx, op, true); + } /* Value is {0,-1} so all bits are repetitions of the sign. */ ctx->s_mask = -1; -- 2.43.0 Avoid the use of the OptContext slots. Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- tcg/optimize.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) diff --git a/tcg/optimize.c b/tcg/optimize.c index XXXXXXX..XXXXXXX 100644 --- a/tcg/optimize.c +++ b/tcg/optimize.c @@ -XXX,XX +XXX,XX @@ static bool fold_setcond(OptContext *ctx, TCGOp *op) fold_setcond_tst_pow2(ctx, op, false); } - ctx->z_mask = 1; - return false; + return fold_masks_z(ctx, op, 1); } static bool fold_negsetcond(OptContext *ctx, TCGOp *op) -- 2.43.0 Avoid the use of the OptContext slots. Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- tcg/optimize.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) diff --git a/tcg/optimize.c b/tcg/optimize.c index XXXXXXX..XXXXXXX 100644 --- a/tcg/optimize.c +++ b/tcg/optimize.c @@ -XXX,XX +XXX,XX @@ static bool fold_negsetcond(OptContext *ctx, TCGOp *op) } /* Value is {0,-1} so all bits are repetitions of the sign. */ - ctx->s_mask = -1; - return false; + return fold_masks_s(ctx, op, -1); } static bool fold_setcond2(OptContext *ctx, TCGOp *op) -- 2.43.0 Avoid the use of the OptContext slots. Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- tcg/optimize.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) diff --git a/tcg/optimize.c b/tcg/optimize.c index XXXXXXX..XXXXXXX 100644 --- a/tcg/optimize.c +++ b/tcg/optimize.c @@ -XXX,XX +XXX,XX @@ static bool fold_setcond2(OptContext *ctx, TCGOp *op) return fold_setcond(ctx, op); } - ctx->z_mask = 1; - return false; + return fold_masks_z(ctx, op, 1); do_setcond_const: return tcg_opt_gen_movi(ctx, op, op->args[0], i); -- 2.43.0 Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- tcg/optimize.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/tcg/optimize.c b/tcg/optimize.c index XXXXXXX..XXXXXXX 100644 --- a/tcg/optimize.c +++ b/tcg/optimize.c @@ -XXX,XX +XXX,XX @@ static bool fold_cmp_vec(OptContext *ctx, TCGOp *op) if (swap_commutative(NO_DEST, &op->args[1], &op->args[2])) { op->args[3] = tcg_swap_cond(op->args[3]); } - return false; + return finish_folding(ctx, op); } static bool fold_cmpsel_vec(OptContext *ctx, TCGOp *op) -- 2.43.0 Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- tcg/optimize.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/tcg/optimize.c b/tcg/optimize.c index XXXXXXX..XXXXXXX 100644 --- a/tcg/optimize.c +++ b/tcg/optimize.c @@ -XXX,XX +XXX,XX @@ static bool fold_cmpsel_vec(OptContext *ctx, TCGOp *op) if (swap_commutative(op->args[0], &op->args[4], &op->args[3])) { op->args[5] = tcg_invert_cond(op->args[5]); } - return false; + return finish_folding(ctx, op); } static bool fold_sextract(OptContext *ctx, TCGOp *op) -- 2.43.0 Avoid the use of the OptContext slots. Find TempOptInfo once. Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- tcg/optimize.c | 24 +++++++++--------------- 1 file changed, 9 insertions(+), 15 deletions(-) diff --git a/tcg/optimize.c b/tcg/optimize.c index XXXXXXX..XXXXXXX 100644 --- a/tcg/optimize.c +++ b/tcg/optimize.c @@ -XXX,XX +XXX,XX @@ static bool fold_cmpsel_vec(OptContext *ctx, TCGOp *op) static bool fold_sextract(OptContext *ctx, TCGOp *op) { uint64_t z_mask, s_mask, s_mask_old; + TempOptInfo *t1 = arg_info(op->args[1]); int pos = op->args[2]; int len = op->args[3]; - if (arg_is_const(op->args[1])) { - uint64_t t; - - t = arg_info(op->args[1])->val; - t = sextract64(t, pos, len); - return tcg_opt_gen_movi(ctx, op, op->args[0], t); + if (ti_is_const(t1)) { + return tcg_opt_gen_movi(ctx, op, op->args[0], + sextract64(ti_const_val(t1), pos, len)); } - z_mask = arg_info(op->args[1])->z_mask; - z_mask = sextract64(z_mask, pos, len); - ctx->z_mask = z_mask; - - s_mask_old = arg_info(op->args[1])->s_mask; - s_mask = sextract64(s_mask_old, pos, len); - s_mask |= MAKE_64BIT_MASK(len, 64 - len); - ctx->s_mask = s_mask; + s_mask_old = t1->s_mask; + s_mask = s_mask_old >> pos; + s_mask |= -1ull << (len - 1); if (0 && pos == 0 && fold_affected_mask(ctx, op, s_mask & ~s_mask_old)) { return true; } - return fold_masks(ctx, op); + z_mask = sextract64(t1->z_mask, pos, len); + return fold_masks_zs(ctx, op, z_mask, s_mask); } static bool fold_shift(OptContext *ctx, TCGOp *op) -- 2.43.0 Avoid the use of the OptContext slots. Find TempOptInfo once. Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- tcg/optimize.c | 27 ++++++++++++++------------- 1 file changed, 14 insertions(+), 13 deletions(-) diff --git a/tcg/optimize.c b/tcg/optimize.c index XXXXXXX..XXXXXXX 100644 --- a/tcg/optimize.c +++ b/tcg/optimize.c @@ -XXX,XX +XXX,XX @@ static bool fold_sextract(OptContext *ctx, TCGOp *op) static bool fold_shift(OptContext *ctx, TCGOp *op) { uint64_t s_mask, z_mask, sign; + TempOptInfo *t1, *t2; if (fold_const2(ctx, op) || fold_ix_to_i(ctx, op, 0) || @@ -XXX,XX +XXX,XX @@ static bool fold_shift(OptContext *ctx, TCGOp *op) return true; } - s_mask = arg_info(op->args[1])->s_mask; - z_mask = arg_info(op->args[1])->z_mask; + t1 = arg_info(op->args[1]); + t2 = arg_info(op->args[2]); + s_mask = t1->s_mask; + z_mask = t1->z_mask; - if (arg_is_const(op->args[2])) { - int sh = arg_info(op->args[2])->val; - - ctx->z_mask = do_constant_folding(op->opc, ctx->type, z_mask, sh); + if (ti_is_const(t2)) { + int sh = ti_const_val(t2); + z_mask = do_constant_folding(op->opc, ctx->type, z_mask, sh); s_mask = do_constant_folding(op->opc, ctx->type, s_mask, sh); - return fold_masks(ctx, op); + return fold_masks_zs(ctx, op, z_mask, s_mask); } switch (op->opc) { @@ -XXX,XX +XXX,XX @@ static bool fold_shift(OptContext *ctx, TCGOp *op) * Arithmetic right shift will not reduce the number of * input sign repetitions. */ - ctx->s_mask = s_mask; - break; + return fold_masks_s(ctx, op, s_mask); CASE_OP_32_64(shr): /* * If the sign bit is known zero, then logical right shift - * will not reduced the number of input sign repetitions. + * will not reduce the number of input sign repetitions. */ - sign = (s_mask & -s_mask) >> 1; + sign = -s_mask; if (sign && !(z_mask & sign)) { - ctx->s_mask = s_mask; + return fold_masks_s(ctx, op, s_mask); } break; default: break; } - return false; + return finish_folding(ctx, op); } static bool fold_sub_to_neg(OptContext *ctx, TCGOp *op) -- 2.43.0 Merge the two conditions, sign != 0 && !(z_mask & sign), by testing ~z_mask & sign. If sign == 0, the logical and will produce false. Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- tcg/optimize.c | 5 ++--- 1 file changed, 2 insertions(+), 3 deletions(-) diff --git a/tcg/optimize.c b/tcg/optimize.c index XXXXXXX..XXXXXXX 100644 --- a/tcg/optimize.c +++ b/tcg/optimize.c @@ -XXX,XX +XXX,XX @@ static bool fold_sextract(OptContext *ctx, TCGOp *op) static bool fold_shift(OptContext *ctx, TCGOp *op) { - uint64_t s_mask, z_mask, sign; + uint64_t s_mask, z_mask; TempOptInfo *t1, *t2; if (fold_const2(ctx, op) || @@ -XXX,XX +XXX,XX @@ static bool fold_shift(OptContext *ctx, TCGOp *op) * If the sign bit is known zero, then logical right shift * will not reduce the number of input sign repetitions. */ - sign = -s_mask; - if (sign && !(z_mask & sign)) { + if (~z_mask & -s_mask) { return fold_masks_s(ctx, op, s_mask); } break; -- 2.43.0 Duplicate fold_sub_vec into fold_sub instead of calling it, now that fold_sub_vec always returns true. Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- tcg/optimize.c | 9 ++++++--- 1 file changed, 6 insertions(+), 3 deletions(-) diff --git a/tcg/optimize.c b/tcg/optimize.c index XXXXXXX..XXXXXXX 100644 --- a/tcg/optimize.c +++ b/tcg/optimize.c @@ -XXX,XX +XXX,XX @@ static bool fold_sub_vec(OptContext *ctx, TCGOp *op) fold_sub_to_neg(ctx, op)) { return true; } - return false; + return finish_folding(ctx, op); } static bool fold_sub(OptContext *ctx, TCGOp *op) { - if (fold_const2(ctx, op) || fold_sub_vec(ctx, op)) { + if (fold_const2(ctx, op) || + fold_xx_to_i(ctx, op, 0) || + fold_xi_to_x(ctx, op, 0) || + fold_sub_to_neg(ctx, op)) { return true; } @@ -XXX,XX +XXX,XX @@ static bool fold_sub(OptContext *ctx, TCGOp *op) ? INDEX_op_add_i32 : INDEX_op_add_i64); op->args[2] = arg_new_constant(ctx, -val); } - return false; + return finish_folding(ctx, op); } static bool fold_sub2(OptContext *ctx, TCGOp *op) -- 2.43.0 Avoid the use of the OptContext slots. Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- tcg/optimize.c | 16 +++++++++------- 1 file changed, 9 insertions(+), 7 deletions(-) diff --git a/tcg/optimize.c b/tcg/optimize.c index XXXXXXX..XXXXXXX 100644 --- a/tcg/optimize.c +++ b/tcg/optimize.c @@ -XXX,XX +XXX,XX @@ static bool fold_sub2(OptContext *ctx, TCGOp *op) static bool fold_tcg_ld(OptContext *ctx, TCGOp *op) { + uint64_t z_mask = -1, s_mask = 0; + /* We can't do any folding with a load, but we can record bits. */ switch (op->opc) { CASE_OP_32_64(ld8s): - ctx->s_mask = MAKE_64BIT_MASK(8, 56); + s_mask = INT8_MIN; break; CASE_OP_32_64(ld8u): - ctx->z_mask = MAKE_64BIT_MASK(0, 8); + z_mask = MAKE_64BIT_MASK(0, 8); break; CASE_OP_32_64(ld16s): - ctx->s_mask = MAKE_64BIT_MASK(16, 48); + s_mask = INT16_MIN; break; CASE_OP_32_64(ld16u): - ctx->z_mask = MAKE_64BIT_MASK(0, 16); + z_mask = MAKE_64BIT_MASK(0, 16); break; case INDEX_op_ld32s_i64: - ctx->s_mask = MAKE_64BIT_MASK(32, 32); + s_mask = INT32_MIN; break; case INDEX_op_ld32u_i64: - ctx->z_mask = MAKE_64BIT_MASK(0, 32); + z_mask = MAKE_64BIT_MASK(0, 32); break; default: g_assert_not_reached(); } - return false; + return fold_masks_zs(ctx, op, z_mask, s_mask); } static bool fold_tcg_ld_memcopy(OptContext *ctx, TCGOp *op) -- 2.43.0 Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- tcg/optimize.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/tcg/optimize.c b/tcg/optimize.c index XXXXXXX..XXXXXXX 100644 --- a/tcg/optimize.c +++ b/tcg/optimize.c @@ -XXX,XX +XXX,XX @@ static bool fold_tcg_ld_memcopy(OptContext *ctx, TCGOp *op) TCGType type; if (op->args[1] != tcgv_ptr_arg(tcg_env)) { - return false; + return finish_folding(ctx, op); } type = ctx->type; -- 2.43.0 Avoid the use of the OptContext slots. Find TempOptInfo once. Remove fold_masks as the function becomes unused. Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- tcg/optimize.c | 18 ++++++++---------- 1 file changed, 8 insertions(+), 10 deletions(-) diff --git a/tcg/optimize.c b/tcg/optimize.c index XXXXXXX..XXXXXXX 100644 --- a/tcg/optimize.c +++ b/tcg/optimize.c @@ -XXX,XX +XXX,XX @@ static bool fold_masks_s(OptContext *ctx, TCGOp *op, uint64_t s_mask) return fold_masks_zs(ctx, op, -1, s_mask); } -static bool fold_masks(OptContext *ctx, TCGOp *op) -{ - return fold_masks_zs(ctx, op, ctx->z_mask, ctx->s_mask); -} - /* * An "affected" mask bit is 0 if and only if the result is identical * to the first input. Thus if the entire mask is 0, the operation @@ -XXX,XX +XXX,XX @@ static bool fold_tcg_st_memcopy(OptContext *ctx, TCGOp *op) static bool fold_xor(OptContext *ctx, TCGOp *op) { + uint64_t z_mask, s_mask; + TempOptInfo *t1, *t2; + if (fold_const2_commutative(ctx, op) || fold_xx_to_i(ctx, op, 0) || fold_xi_to_x(ctx, op, 0) || @@ -XXX,XX +XXX,XX @@ static bool fold_xor(OptContext *ctx, TCGOp *op) return true; } - ctx->z_mask = arg_info(op->args[1])->z_mask - | arg_info(op->args[2])->z_mask; - ctx->s_mask = arg_info(op->args[1])->s_mask - & arg_info(op->args[2])->s_mask; - return fold_masks(ctx, op); + t1 = arg_info(op->args[1]); + t2 = arg_info(op->args[2]); + z_mask = t1->z_mask | t2->z_mask; + s_mask = t1->s_mask & t2->s_mask; + return fold_masks_zs(ctx, op, z_mask, s_mask); } static bool fold_bitsel_vec(OptContext *ctx, TCGOp *op) -- 2.43.0 Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- tcg/optimize.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/tcg/optimize.c b/tcg/optimize.c index XXXXXXX..XXXXXXX 100644 --- a/tcg/optimize.c +++ b/tcg/optimize.c @@ -XXX,XX +XXX,XX @@ static bool fold_bitsel_vec(OptContext *ctx, TCGOp *op) return fold_orc(ctx, op); } } - return false; + return finish_folding(ctx, op); } /* Propagate constants and copies, fold constant expressions. */ -- 2.43.0 All non-default cases now finish folding within each function. Do the same with the default case and assert it is done after. Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- tcg/optimize.c | 6 ++---- 1 file changed, 2 insertions(+), 4 deletions(-) diff --git a/tcg/optimize.c b/tcg/optimize.c index XXXXXXX..XXXXXXX 100644 --- a/tcg/optimize.c +++ b/tcg/optimize.c @@ -XXX,XX +XXX,XX @@ void tcg_optimize(TCGContext *s) done = true; break; default: + done = finish_folding(&ctx, op); break; } - - if (!done) { - finish_folding(&ctx, op); - } + tcg_debug_assert(done); } } -- 2.43.0 All mask setting is now done with parameters via fold_masks_*. Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- tcg/optimize.c | 13 ------------- 1 file changed, 13 deletions(-) diff --git a/tcg/optimize.c b/tcg/optimize.c index XXXXXXX..XXXXXXX 100644 --- a/tcg/optimize.c +++ b/tcg/optimize.c @@ -XXX,XX +XXX,XX @@ typedef struct OptContext { QSIMPLEQ_HEAD(, MemCopyInfo) mem_free; /* In flight values from optimization. */ - uint64_t z_mask; /* mask bit is 0 iff value bit is 0 */ - uint64_t s_mask; /* mask bit is 1 if value bit matches msb */ TCGType type; } OptContext; @@ -XXX,XX +XXX,XX @@ static bool finish_folding(OptContext *ctx, TCGOp *op) for (i = 0; i < nb_oargs; i++) { TCGTemp *ts = arg_temp(op->args[i]); reset_ts(ctx, ts); - /* - * Save the corresponding known-zero/sign bits mask for the - * first output argument (only one supported so far). - */ - if (i == 0) { - ts_info(ts)->z_mask = ctx->z_mask; - } } return true; } @@ -XXX,XX +XXX,XX @@ void tcg_optimize(TCGContext *s) ctx.type = TCG_TYPE_I32; } - /* Assume all bits affected, no bits known zero, no sign reps. */ - ctx.z_mask = -1; - ctx.s_mask = 0; - /* * Process each opcode. * Sorted alphabetically by opcode as much as possible. -- 2.43.0 All instances of s_mask have been converted to the new representation. We can now re-enable usage. Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- tcg/optimize.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/tcg/optimize.c b/tcg/optimize.c index XXXXXXX..XXXXXXX 100644 --- a/tcg/optimize.c +++ b/tcg/optimize.c @@ -XXX,XX +XXX,XX @@ static bool fold_exts(OptContext *ctx, TCGOp *op) g_assert_not_reached(); } - if (0 && !type_change && fold_affected_mask(ctx, op, s_mask & ~s_mask_old)) { + if (!type_change && fold_affected_mask(ctx, op, s_mask & ~s_mask_old)) { return true; } @@ -XXX,XX +XXX,XX @@ static bool fold_sextract(OptContext *ctx, TCGOp *op) s_mask = s_mask_old >> pos; s_mask |= -1ull << (len - 1); - if (0 && pos == 0 && fold_affected_mask(ctx, op, s_mask & ~s_mask_old)) { + if (pos == 0 && fold_affected_mask(ctx, op, s_mask & ~s_mask_old)) { return true; } -- 2.43.0 The big comment just above says functions should be sorted. Add forward declarations as needed. Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- tcg/optimize.c | 114 +++++++++++++++++++++++++------------------------ 1 file changed, 59 insertions(+), 55 deletions(-) diff --git a/tcg/optimize.c b/tcg/optimize.c index XXXXXXX..XXXXXXX 100644 --- a/tcg/optimize.c +++ b/tcg/optimize.c @@ -XXX,XX +XXX,XX @@ static bool fold_xx_to_x(OptContext *ctx, TCGOp *op) * 3) those that produce information about the result value. */ +static bool fold_or(OptContext *ctx, TCGOp *op); +static bool fold_orc(OptContext *ctx, TCGOp *op); +static bool fold_xor(OptContext *ctx, TCGOp *op); + static bool fold_add(OptContext *ctx, TCGOp *op) { if (fold_const2_commutative(ctx, op) || @@ -XXX,XX +XXX,XX @@ static bool fold_andc(OptContext *ctx, TCGOp *op) return fold_masks_zs(ctx, op, z_mask, s_mask); } +static bool fold_bitsel_vec(OptContext *ctx, TCGOp *op) +{ + /* If true and false values are the same, eliminate the cmp. */ + if (args_are_copies(op->args[2], op->args[3])) { + return tcg_opt_gen_mov(ctx, op, op->args[0], op->args[2]); + } + + if (arg_is_const(op->args[2]) && arg_is_const(op->args[3])) { + uint64_t tv = arg_info(op->args[2])->val; + uint64_t fv = arg_info(op->args[3])->val; + + if (tv == -1 && fv == 0) { + return tcg_opt_gen_mov(ctx, op, op->args[0], op->args[1]); + } + if (tv == 0 && fv == -1) { + if (TCG_TARGET_HAS_not_vec) { + op->opc = INDEX_op_not_vec; + return fold_not(ctx, op); + } else { + op->opc = INDEX_op_xor_vec; + op->args[2] = arg_new_constant(ctx, -1); + return fold_xor(ctx, op); + } + } + } + if (arg_is_const(op->args[2])) { + uint64_t tv = arg_info(op->args[2])->val; + if (tv == -1) { + op->opc = INDEX_op_or_vec; + op->args[2] = op->args[3]; + return fold_or(ctx, op); + } + if (tv == 0 && TCG_TARGET_HAS_andc_vec) { + op->opc = INDEX_op_andc_vec; + op->args[2] = op->args[1]; + op->args[1] = op->args[3]; + return fold_andc(ctx, op); + } + } + if (arg_is_const(op->args[3])) { + uint64_t fv = arg_info(op->args[3])->val; + if (fv == 0) { + op->opc = INDEX_op_and_vec; + return fold_and(ctx, op); + } + if (fv == -1 && TCG_TARGET_HAS_orc_vec) { + op->opc = INDEX_op_orc_vec; + op->args[2] = op->args[1]; + op->args[1] = op->args[3]; + return fold_orc(ctx, op); + } + } + return finish_folding(ctx, op); +} + static bool fold_brcond(OptContext *ctx, TCGOp *op) { int i = do_constant_folding_cond1(ctx, op, NO_DEST, &op->args[0], @@ -XXX,XX +XXX,XX @@ static bool fold_xor(OptContext *ctx, TCGOp *op) return fold_masks_zs(ctx, op, z_mask, s_mask); } -static bool fold_bitsel_vec(OptContext *ctx, TCGOp *op) -{ - /* If true and false values are the same, eliminate the cmp. */ - if (args_are_copies(op->args[2], op->args[3])) { - return tcg_opt_gen_mov(ctx, op, op->args[0], op->args[2]); - } - - if (arg_is_const(op->args[2]) && arg_is_const(op->args[3])) { - uint64_t tv = arg_info(op->args[2])->val; - uint64_t fv = arg_info(op->args[3])->val; - - if (tv == -1 && fv == 0) { - return tcg_opt_gen_mov(ctx, op, op->args[0], op->args[1]); - } - if (tv == 0 && fv == -1) { - if (TCG_TARGET_HAS_not_vec) { - op->opc = INDEX_op_not_vec; - return fold_not(ctx, op); - } else { - op->opc = INDEX_op_xor_vec; - op->args[2] = arg_new_constant(ctx, -1); - return fold_xor(ctx, op); - } - } - } - if (arg_is_const(op->args[2])) { - uint64_t tv = arg_info(op->args[2])->val; - if (tv == -1) { - op->opc = INDEX_op_or_vec; - op->args[2] = op->args[3]; - return fold_or(ctx, op); - } - if (tv == 0 && TCG_TARGET_HAS_andc_vec) { - op->opc = INDEX_op_andc_vec; - op->args[2] = op->args[1]; - op->args[1] = op->args[3]; - return fold_andc(ctx, op); - } - } - if (arg_is_const(op->args[3])) { - uint64_t fv = arg_info(op->args[3])->val; - if (fv == 0) { - op->opc = INDEX_op_and_vec; - return fold_and(ctx, op); - } - if (fv == -1 && TCG_TARGET_HAS_orc_vec) { - op->opc = INDEX_op_orc_vec; - op->args[2] = op->args[1]; - op->args[1] = op->args[3]; - return fold_orc(ctx, op); - } - } - return finish_folding(ctx, op); -} - /* Propagate constants and copies, fold constant expressions. */ void tcg_optimize(TCGContext *s) { -- 2.43.0 The big comment just above says functions should be sorted. Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- tcg/optimize.c | 60 +++++++++++++++++++++++++------------------------- 1 file changed, 30 insertions(+), 30 deletions(-) diff --git a/tcg/optimize.c b/tcg/optimize.c index XXXXXXX..XXXXXXX 100644 --- a/tcg/optimize.c +++ b/tcg/optimize.c @@ -XXX,XX +XXX,XX @@ static bool fold_call(OptContext *ctx, TCGOp *op) return true; } +static bool fold_cmp_vec(OptContext *ctx, TCGOp *op) +{ + /* Canonicalize the comparison to put immediate second. */ + if (swap_commutative(NO_DEST, &op->args[1], &op->args[2])) { + op->args[3] = tcg_swap_cond(op->args[3]); + } + return finish_folding(ctx, op); +} + +static bool fold_cmpsel_vec(OptContext *ctx, TCGOp *op) +{ + /* If true and false values are the same, eliminate the cmp. */ + if (args_are_copies(op->args[3], op->args[4])) { + return tcg_opt_gen_mov(ctx, op, op->args[0], op->args[3]); + } + + /* Canonicalize the comparison to put immediate second. */ + if (swap_commutative(NO_DEST, &op->args[1], &op->args[2])) { + op->args[5] = tcg_swap_cond(op->args[5]); + } + /* + * Canonicalize the "false" input reg to match the destination, + * so that the tcg backend can implement "move if true". + */ + if (swap_commutative(op->args[0], &op->args[4], &op->args[3])) { + op->args[5] = tcg_invert_cond(op->args[5]); + } + return finish_folding(ctx, op); +} + static bool fold_count_zeros(OptContext *ctx, TCGOp *op) { uint64_t z_mask, s_mask; @@ -XXX,XX +XXX,XX @@ static bool fold_setcond2(OptContext *ctx, TCGOp *op) return tcg_opt_gen_movi(ctx, op, op->args[0], i); } -static bool fold_cmp_vec(OptContext *ctx, TCGOp *op) -{ - /* Canonicalize the comparison to put immediate second. */ - if (swap_commutative(NO_DEST, &op->args[1], &op->args[2])) { - op->args[3] = tcg_swap_cond(op->args[3]); - } - return finish_folding(ctx, op); -} - -static bool fold_cmpsel_vec(OptContext *ctx, TCGOp *op) -{ - /* If true and false values are the same, eliminate the cmp. */ - if (args_are_copies(op->args[3], op->args[4])) { - return tcg_opt_gen_mov(ctx, op, op->args[0], op->args[3]); - } - - /* Canonicalize the comparison to put immediate second. */ - if (swap_commutative(NO_DEST, &op->args[1], &op->args[2])) { - op->args[5] = tcg_swap_cond(op->args[5]); - } - /* - * Canonicalize the "false" input reg to match the destination, - * so that the tcg backend can implement "move if true". - */ - if (swap_commutative(op->args[0], &op->args[4], &op->args[3])) { - op->args[5] = tcg_invert_cond(op->args[5]); - } - return finish_folding(ctx, op); -} - static bool fold_sextract(OptContext *ctx, TCGOp *op) { uint64_t z_mask, s_mask, s_mask_old; -- 2.43.0 We currently have a flag, float_muladd_halve_result, to scale the result by 2**-1. Extend this to handle arbitrary scaling. Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- include/fpu/softfloat.h | 6 ++++ fpu/softfloat.c | 58 ++++++++++++++++++++++----------------- fpu/softfloat-parts.c.inc | 7 +++-- 3 files changed, 44 insertions(+), 27 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 @@ float16 float16_add(float16, float16, float_status *status); float16 float16_sub(float16, float16, float_status *status); float16 float16_mul(float16, float16, float_status *status); float16 float16_muladd(float16, float16, float16, int, float_status *status); +float16 float16_muladd_scalbn(float16, float16, float16, + int, int, float_status *status); float16 float16_div(float16, float16, float_status *status); float16 float16_scalbn(float16, int, float_status *status); float16 float16_min(float16, float16, float_status *status); @@ -XXX,XX +XXX,XX @@ float32 float32_mul(float32, float32, float_status *status); float32 float32_div(float32, float32, float_status *status); float32 float32_rem(float32, float32, float_status *status); float32 float32_muladd(float32, float32, float32, int, float_status *status); +float32 float32_muladd_scalbn(float32, float32, float32, + int, int, float_status *status); float32 float32_sqrt(float32, float_status *status); float32 float32_exp2(float32, float_status *status); float32 float32_log2(float32, float_status *status); @@ -XXX,XX +XXX,XX @@ float64 float64_mul(float64, float64, float_status *status); float64 float64_div(float64, float64, float_status *status); float64 float64_rem(float64, float64, float_status *status); float64 float64_muladd(float64, float64, float64, int, float_status *status); +float64 float64_muladd_scalbn(float64, float64, float64, + int, int, float_status *status); float64 float64_sqrt(float64, float_status *status); float64 float64_log2(float64, float_status *status); FloatRelation float64_compare(float64, float64, float_status *status); diff --git a/fpu/softfloat.c b/fpu/softfloat.c index XXXXXXX..XXXXXXX 100644 --- a/fpu/softfloat.c +++ b/fpu/softfloat.c @@ -XXX,XX +XXX,XX @@ static FloatParts128 *parts128_mul(FloatParts128 *a, FloatParts128 *b, #define parts_mul(A, B, S) \ PARTS_GENERIC_64_128(mul, A)(A, B, S) -static FloatParts64 *parts64_muladd(FloatParts64 *a, FloatParts64 *b, - FloatParts64 *c, int flags, - float_status *s); -static FloatParts128 *parts128_muladd(FloatParts128 *a, FloatParts128 *b, - FloatParts128 *c, int flags, - float_status *s); +static FloatParts64 *parts64_muladd_scalbn(FloatParts64 *a, FloatParts64 *b, + FloatParts64 *c, int scale, + int flags, float_status *s); +static FloatParts128 *parts128_muladd_scalbn(FloatParts128 *a, FloatParts128 *b, + FloatParts128 *c, int scale, + int flags, float_status *s); -#define parts_muladd(A, B, C, Z, S) \ - PARTS_GENERIC_64_128(muladd, A)(A, B, C, Z, S) +#define parts_muladd_scalbn(A, B, C, Z, Y, S) \ + PARTS_GENERIC_64_128(muladd_scalbn, A)(A, B, C, Z, Y, S) static FloatParts64 *parts64_div(FloatParts64 *a, FloatParts64 *b, float_status *s); @@ -XXX,XX +XXX,XX @@ floatx80_mul(floatx80 a, floatx80 b, float_status *status) * Fused multiply-add */ -float16 QEMU_FLATTEN float16_muladd(float16 a, float16 b, float16 c, - int flags, float_status *status) +float16 QEMU_FLATTEN +float16_muladd_scalbn(float16 a, float16 b, float16 c, + int scale, int flags, float_status *status) { FloatParts64 pa, pb, pc, *pr; float16_unpack_canonical(&pa, a, status); float16_unpack_canonical(&pb, b, status); float16_unpack_canonical(&pc, c, status); - pr = parts_muladd(&pa, &pb, &pc, flags, status); + pr = parts_muladd_scalbn(&pa, &pb, &pc, scale, flags, status); return float16_round_pack_canonical(pr, status); } -static float32 QEMU_SOFTFLOAT_ATTR -soft_f32_muladd(float32 a, float32 b, float32 c, int flags, - float_status *status) +float16 float16_muladd(float16 a, float16 b, float16 c, + int flags, float_status *status) +{ + return float16_muladd_scalbn(a, b, c, 0, flags, status); +} + +float32 QEMU_SOFTFLOAT_ATTR +float32_muladd_scalbn(float32 a, float32 b, float32 c, + int scale, int flags, float_status *status) { FloatParts64 pa, pb, pc, *pr; float32_unpack_canonical(&pa, a, status); float32_unpack_canonical(&pb, b, status); float32_unpack_canonical(&pc, c, status); - pr = parts_muladd(&pa, &pb, &pc, flags, status); + pr = parts_muladd_scalbn(&pa, &pb, &pc, scale, flags, status); return float32_round_pack_canonical(pr, status); } -static float64 QEMU_SOFTFLOAT_ATTR -soft_f64_muladd(float64 a, float64 b, float64 c, int flags, - float_status *status) +float64 QEMU_SOFTFLOAT_ATTR +float64_muladd_scalbn(float64 a, float64 b, float64 c, + int scale, int flags, float_status *status) { FloatParts64 pa, pb, pc, *pr; float64_unpack_canonical(&pa, a, status); float64_unpack_canonical(&pb, b, status); float64_unpack_canonical(&pc, c, status); - pr = parts_muladd(&pa, &pb, &pc, flags, status); + pr = parts_muladd_scalbn(&pa, &pb, &pc, scale, flags, status); return float64_round_pack_canonical(pr, status); } @@ -XXX,XX +XXX,XX @@ float32_muladd(float32 xa, float32 xb, float32 xc, int flags, float_status *s) return ur.s; soft: - return soft_f32_muladd(ua.s, ub.s, uc.s, flags, s); + return float32_muladd_scalbn(ua.s, ub.s, uc.s, 0, flags, s); } float64 QEMU_FLATTEN @@ -XXX,XX +XXX,XX @@ float64_muladd(float64 xa, float64 xb, float64 xc, int flags, float_status *s) return ur.s; soft: - return soft_f64_muladd(ua.s, ub.s, uc.s, flags, s); + return float64_muladd_scalbn(ua.s, ub.s, uc.s, 0, flags, s); } float64 float64r32_muladd(float64 a, float64 b, float64 c, @@ -XXX,XX +XXX,XX @@ float64 float64r32_muladd(float64 a, float64 b, float64 c, float64_unpack_canonical(&pa, a, status); float64_unpack_canonical(&pb, b, status); float64_unpack_canonical(&pc, c, status); - pr = parts_muladd(&pa, &pb, &pc, flags, status); + pr = parts_muladd_scalbn(&pa, &pb, &pc, 0, flags, status); return float64r32_round_pack_canonical(pr, status); } @@ -XXX,XX +XXX,XX @@ bfloat16 QEMU_FLATTEN bfloat16_muladd(bfloat16 a, bfloat16 b, bfloat16 c, bfloat16_unpack_canonical(&pa, a, status); bfloat16_unpack_canonical(&pb, b, status); bfloat16_unpack_canonical(&pc, c, status); - pr = parts_muladd(&pa, &pb, &pc, flags, status); + pr = parts_muladd_scalbn(&pa, &pb, &pc, 0, flags, status); return bfloat16_round_pack_canonical(pr, status); } @@ -XXX,XX +XXX,XX @@ float128 QEMU_FLATTEN float128_muladd(float128 a, float128 b, float128 c, float128_unpack_canonical(&pa, a, status); float128_unpack_canonical(&pb, b, status); float128_unpack_canonical(&pc, c, status); - pr = parts_muladd(&pa, &pb, &pc, flags, status); + pr = parts_muladd_scalbn(&pa, &pb, &pc, 0, flags, status); return float128_round_pack_canonical(pr, status); } @@ -XXX,XX +XXX,XX @@ float32 float32_exp2(float32 a, float_status *status) float64_unpack_canonical(&rp, float64_one, status); for (i = 0 ; i < 15 ; i++) { + float64_unpack_canonical(&tp, float32_exp2_coefficients[i], status); - rp = *parts_muladd(&tp, &xnp, &rp, 0, status); + rp = *parts_muladd_scalbn(&tp, &xnp, &rp, 0, 0, status); xnp = *parts_mul(&xnp, &xp, status); } diff --git a/fpu/softfloat-parts.c.inc b/fpu/softfloat-parts.c.inc index XXXXXXX..XXXXXXX 100644 --- a/fpu/softfloat-parts.c.inc +++ b/fpu/softfloat-parts.c.inc @@ -XXX,XX +XXX,XX @@ static FloatPartsN *partsN(mul)(FloatPartsN *a, FloatPartsN *b, * Requires A and C extracted into a double-sized structure to provide the * extra space for the widening multiply. */ -static FloatPartsN *partsN(muladd)(FloatPartsN *a, FloatPartsN *b, - FloatPartsN *c, int flags, float_status *s) +static FloatPartsN *partsN(muladd_scalbn)(FloatPartsN *a, FloatPartsN *b, + FloatPartsN *c, int scale, + int flags, float_status *s) { int ab_mask, abc_mask; FloatPartsW p_widen, c_widen; @@ -XXX,XX +XXX,XX @@ static FloatPartsN *partsN(muladd)(FloatPartsN *a, FloatPartsN *b, a->exp = p_widen.exp; return_normal: + /* TODO: Replace all use of float_muladd_halve_result with scale. */ if (flags & float_muladd_halve_result) { a->exp -= 1; } + a->exp += scale; finish_sign: if (flags & float_muladd_negate_result) { a->sign ^= 1; -- 2.43.0 Use the scalbn interface instead of float_muladd_halve_result. Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- target/arm/tcg/helper-a64.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) diff --git a/target/arm/tcg/helper-a64.c b/target/arm/tcg/helper-a64.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/tcg/helper-a64.c +++ b/target/arm/tcg/helper-a64.c @@ -XXX,XX +XXX,XX @@ uint32_t HELPER(rsqrtsf_f16)(uint32_t a, uint32_t b, float_status *fpst) (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); + return float16_muladd_scalbn(a, b, float16_three, -1, 0, fpst); } float32 HELPER(rsqrtsf_f32)(float32 a, float32 b, float_status *fpst) @@ -XXX,XX +XXX,XX @@ float32 HELPER(rsqrtsf_f32)(float32 a, float32 b, float_status *fpst) (float32_is_infinity(b) && float32_is_zero(a))) { return float32_one_point_five; } - return float32_muladd(a, b, float32_three, float_muladd_halve_result, fpst); + return float32_muladd_scalbn(a, b, float32_three, -1, 0, fpst); } float64 HELPER(rsqrtsf_f64)(float64 a, float64 b, float_status *fpst) @@ -XXX,XX +XXX,XX @@ float64 HELPER(rsqrtsf_f64)(float64 a, float64 b, float_status *fpst) (float64_is_infinity(b) && float64_is_zero(a))) { return float64_one_point_five; } - return float64_muladd(a, b, float64_three, float_muladd_halve_result, fpst); + return float64_muladd_scalbn(a, b, float64_three, -1, 0, fpst); } /* Floating-point reciprocal exponent - see FPRecpX in ARM ARM */ -- 2.43.0 Use the scalbn interface instead of float_muladd_halve_result. Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- target/sparc/helper.h | 4 +- target/sparc/fop_helper.c | 8 ++-- target/sparc/translate.c | 80 +++++++++++++++++++++++---------------- 3 files changed, 54 insertions(+), 38 deletions(-) diff --git a/target/sparc/helper.h b/target/sparc/helper.h index XXXXXXX..XXXXXXX 100644 --- a/target/sparc/helper.h +++ b/target/sparc/helper.h @@ -XXX,XX +XXX,XX @@ DEF_HELPER_FLAGS_3(faddd, TCG_CALL_NO_WG, f64, env, f64, f64) DEF_HELPER_FLAGS_3(fsubd, TCG_CALL_NO_WG, f64, env, f64, f64) DEF_HELPER_FLAGS_3(fmuld, TCG_CALL_NO_WG, f64, env, f64, f64) DEF_HELPER_FLAGS_3(fdivd, TCG_CALL_NO_WG, f64, env, f64, f64) -DEF_HELPER_FLAGS_5(fmaddd, TCG_CALL_NO_WG, f64, env, f64, f64, f64, i32) +DEF_HELPER_FLAGS_6(fmaddd, TCG_CALL_NO_WG, f64, env, f64, f64, f64, s32, i32) DEF_HELPER_FLAGS_3(fnaddd, TCG_CALL_NO_WG, f64, env, f64, f64) DEF_HELPER_FLAGS_3(fnmuld, TCG_CALL_NO_WG, f64, env, f64, f64) @@ -XXX,XX +XXX,XX @@ DEF_HELPER_FLAGS_3(fadds, TCG_CALL_NO_WG, f32, env, f32, f32) DEF_HELPER_FLAGS_3(fsubs, TCG_CALL_NO_WG, f32, env, f32, f32) DEF_HELPER_FLAGS_3(fmuls, TCG_CALL_NO_WG, f32, env, f32, f32) DEF_HELPER_FLAGS_3(fdivs, TCG_CALL_NO_WG, f32, env, f32, f32) -DEF_HELPER_FLAGS_5(fmadds, TCG_CALL_NO_WG, f32, env, f32, f32, f32, i32) +DEF_HELPER_FLAGS_6(fmadds, TCG_CALL_NO_WG, f32, env, f32, f32, f32, s32, i32) DEF_HELPER_FLAGS_3(fnadds, TCG_CALL_NO_WG, f32, env, f32, f32) DEF_HELPER_FLAGS_3(fnmuls, TCG_CALL_NO_WG, f32, env, f32, f32) diff --git a/target/sparc/fop_helper.c b/target/sparc/fop_helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/sparc/fop_helper.c +++ b/target/sparc/fop_helper.c @@ -XXX,XX +XXX,XX @@ Int128 helper_fsqrtq(CPUSPARCState *env, Int128 src) } float32 helper_fmadds(CPUSPARCState *env, float32 s1, - float32 s2, float32 s3, uint32_t op) + float32 s2, float32 s3, int32_t sc, uint32_t op) { - float32 ret = float32_muladd(s1, s2, s3, op, &env->fp_status); + float32 ret = float32_muladd_scalbn(s1, s2, s3, sc, op, &env->fp_status); check_ieee_exceptions(env, GETPC()); return ret; } float64 helper_fmaddd(CPUSPARCState *env, float64 s1, - float64 s2, float64 s3, uint32_t op) + float64 s2, float64 s3, int32_t sc, uint32_t op) { - float64 ret = float64_muladd(s1, s2, s3, op, &env->fp_status); + float64 ret = float64_muladd_scalbn(s1, s2, s3, sc, op, &env->fp_status); check_ieee_exceptions(env, GETPC()); return ret; } diff --git a/target/sparc/translate.c b/target/sparc/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/sparc/translate.c +++ b/target/sparc/translate.c @@ -XXX,XX +XXX,XX @@ static void gen_op_fabsq(TCGv_i128 dst, TCGv_i128 src) static void gen_op_fmadds(TCGv_i32 d, TCGv_i32 s1, TCGv_i32 s2, TCGv_i32 s3) { - gen_helper_fmadds(d, tcg_env, s1, s2, s3, tcg_constant_i32(0)); + TCGv_i32 z = tcg_constant_i32(0); + gen_helper_fmadds(d, tcg_env, s1, s2, s3, z, z); } static void gen_op_fmaddd(TCGv_i64 d, TCGv_i64 s1, TCGv_i64 s2, TCGv_i64 s3) { - gen_helper_fmaddd(d, tcg_env, s1, s2, s3, tcg_constant_i32(0)); + TCGv_i32 z = tcg_constant_i32(0); + gen_helper_fmaddd(d, tcg_env, s1, s2, s3, z, z); } static void gen_op_fmsubs(TCGv_i32 d, TCGv_i32 s1, TCGv_i32 s2, TCGv_i32 s3) { - int op = float_muladd_negate_c; - gen_helper_fmadds(d, tcg_env, s1, s2, s3, tcg_constant_i32(op)); + TCGv_i32 z = tcg_constant_i32(0); + TCGv_i32 op = tcg_constant_i32(float_muladd_negate_c); + gen_helper_fmadds(d, tcg_env, s1, s2, s3, z, op); } static void gen_op_fmsubd(TCGv_i64 d, TCGv_i64 s1, TCGv_i64 s2, TCGv_i64 s3) { - int op = float_muladd_negate_c; - gen_helper_fmaddd(d, tcg_env, s1, s2, s3, tcg_constant_i32(op)); + TCGv_i32 z = tcg_constant_i32(0); + TCGv_i32 op = tcg_constant_i32(float_muladd_negate_c); + gen_helper_fmaddd(d, tcg_env, s1, s2, s3, z, op); } static void gen_op_fnmsubs(TCGv_i32 d, TCGv_i32 s1, TCGv_i32 s2, TCGv_i32 s3) { - int op = float_muladd_negate_c | float_muladd_negate_result; - gen_helper_fmadds(d, tcg_env, s1, s2, s3, tcg_constant_i32(op)); + TCGv_i32 z = tcg_constant_i32(0); + TCGv_i32 op = tcg_constant_i32(float_muladd_negate_c | + float_muladd_negate_result); + gen_helper_fmadds(d, tcg_env, s1, s2, s3, z, op); } static void gen_op_fnmsubd(TCGv_i64 d, TCGv_i64 s1, TCGv_i64 s2, TCGv_i64 s3) { - int op = float_muladd_negate_c | float_muladd_negate_result; - gen_helper_fmaddd(d, tcg_env, s1, s2, s3, tcg_constant_i32(op)); + TCGv_i32 z = tcg_constant_i32(0); + TCGv_i32 op = tcg_constant_i32(float_muladd_negate_c | + float_muladd_negate_result); + gen_helper_fmaddd(d, tcg_env, s1, s2, s3, z, op); } static void gen_op_fnmadds(TCGv_i32 d, TCGv_i32 s1, TCGv_i32 s2, TCGv_i32 s3) { - int op = float_muladd_negate_result; - gen_helper_fmadds(d, tcg_env, s1, s2, s3, tcg_constant_i32(op)); + TCGv_i32 z = tcg_constant_i32(0); + TCGv_i32 op = tcg_constant_i32(float_muladd_negate_result); + gen_helper_fmadds(d, tcg_env, s1, s2, s3, z, op); } static void gen_op_fnmaddd(TCGv_i64 d, TCGv_i64 s1, TCGv_i64 s2, TCGv_i64 s3) { - int op = float_muladd_negate_result; - gen_helper_fmaddd(d, tcg_env, s1, s2, s3, tcg_constant_i32(op)); + TCGv_i32 z = tcg_constant_i32(0); + TCGv_i32 op = tcg_constant_i32(float_muladd_negate_result); + gen_helper_fmaddd(d, tcg_env, s1, s2, s3, z, op); } /* Use muladd to compute (1 * src1) + src2 / 2 with one rounding. */ static void gen_op_fhadds(TCGv_i32 d, TCGv_i32 s1, TCGv_i32 s2) { - TCGv_i32 one = tcg_constant_i32(float32_one); - int op = float_muladd_halve_result; - gen_helper_fmadds(d, tcg_env, one, s1, s2, tcg_constant_i32(op)); + TCGv_i32 fone = tcg_constant_i32(float32_one); + TCGv_i32 mone = tcg_constant_i32(-1); + TCGv_i32 op = tcg_constant_i32(0); + gen_helper_fmadds(d, tcg_env, fone, s1, s2, mone, op); } static void gen_op_fhaddd(TCGv_i64 d, TCGv_i64 s1, TCGv_i64 s2) { - TCGv_i64 one = tcg_constant_i64(float64_one); - int op = float_muladd_halve_result; - gen_helper_fmaddd(d, tcg_env, one, s1, s2, tcg_constant_i32(op)); + TCGv_i64 fone = tcg_constant_i64(float64_one); + TCGv_i32 mone = tcg_constant_i32(-1); + TCGv_i32 op = tcg_constant_i32(0); + gen_helper_fmaddd(d, tcg_env, fone, s1, s2, mone, op); } /* Use muladd to compute (1 * src1) - src2 / 2 with one rounding. */ static void gen_op_fhsubs(TCGv_i32 d, TCGv_i32 s1, TCGv_i32 s2) { - TCGv_i32 one = tcg_constant_i32(float32_one); - int op = float_muladd_negate_c | float_muladd_halve_result; - gen_helper_fmadds(d, tcg_env, one, s1, s2, tcg_constant_i32(op)); + TCGv_i32 fone = tcg_constant_i32(float32_one); + TCGv_i32 mone = tcg_constant_i32(-1); + TCGv_i32 op = tcg_constant_i32(float_muladd_negate_c); + gen_helper_fmadds(d, tcg_env, fone, s1, s2, mone, op); } static void gen_op_fhsubd(TCGv_i64 d, TCGv_i64 s1, TCGv_i64 s2) { - TCGv_i64 one = tcg_constant_i64(float64_one); - int op = float_muladd_negate_c | float_muladd_halve_result; - gen_helper_fmaddd(d, tcg_env, one, s1, s2, tcg_constant_i32(op)); + TCGv_i64 fone = tcg_constant_i64(float64_one); + TCGv_i32 mone = tcg_constant_i32(-1); + TCGv_i32 op = tcg_constant_i32(float_muladd_negate_c); + gen_helper_fmaddd(d, tcg_env, fone, s1, s2, mone, op); } /* Use muladd to compute -((1 * src1) + src2 / 2) with one rounding. */ static void gen_op_fnhadds(TCGv_i32 d, TCGv_i32 s1, TCGv_i32 s2) { - TCGv_i32 one = tcg_constant_i32(float32_one); - int op = float_muladd_negate_result | float_muladd_halve_result; - gen_helper_fmadds(d, tcg_env, one, s1, s2, tcg_constant_i32(op)); + TCGv_i32 fone = tcg_constant_i32(float32_one); + TCGv_i32 mone = tcg_constant_i32(-1); + TCGv_i32 op = tcg_constant_i32(float_muladd_negate_result); + gen_helper_fmadds(d, tcg_env, fone, s1, s2, mone, op); } static void gen_op_fnhaddd(TCGv_i64 d, TCGv_i64 s1, TCGv_i64 s2) { - TCGv_i64 one = tcg_constant_i64(float64_one); - int op = float_muladd_negate_result | float_muladd_halve_result; - gen_helper_fmaddd(d, tcg_env, one, s1, s2, tcg_constant_i32(op)); + TCGv_i64 fone = tcg_constant_i64(float64_one); + TCGv_i32 mone = tcg_constant_i32(-1); + TCGv_i32 op = tcg_constant_i32(float_muladd_negate_result); + gen_helper_fmaddd(d, tcg_env, fone, s1, s2, mone, op); } static void gen_op_fpexception_im(DisasContext *dc, int ftt) -- 2.43.0 All uses have been convered to float*_muladd_scalbn. Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- include/fpu/softfloat.h | 3 --- fpu/softfloat.c | 6 ------ fpu/softfloat-parts.c.inc | 4 ---- 3 files changed, 13 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 @@ bfloat16 bfloat16_squash_input_denormal(bfloat16 a, float_status *status); | Using these differs from negating an input or output before calling | the muladd function in that this means that a NaN doesn't have its | sign bit inverted before it is propagated. -| We also support halving the result before rounding, as a special -| case to support the ARM fused-sqrt-step instruction FRSQRTS. *----------------------------------------------------------------------------*/ enum { float_muladd_negate_c = 1, float_muladd_negate_product = 2, float_muladd_negate_result = 4, - float_muladd_halve_result = 8, }; /*---------------------------------------------------------------------------- diff --git a/fpu/softfloat.c b/fpu/softfloat.c index XXXXXXX..XXXXXXX 100644 --- a/fpu/softfloat.c +++ b/fpu/softfloat.c @@ -XXX,XX +XXX,XX @@ float32_muladd(float32 xa, float32 xb, float32 xc, int flags, float_status *s) if (unlikely(!can_use_fpu(s))) { goto soft; } - if (unlikely(flags & float_muladd_halve_result)) { - goto soft; - } float32_input_flush3(&ua.s, &ub.s, &uc.s, s); if (unlikely(!f32_is_zon3(ua, ub, uc))) { @@ -XXX,XX +XXX,XX @@ float64_muladd(float64 xa, float64 xb, float64 xc, int flags, float_status *s) if (unlikely(!can_use_fpu(s))) { goto soft; } - if (unlikely(flags & float_muladd_halve_result)) { - goto soft; - } float64_input_flush3(&ua.s, &ub.s, &uc.s, s); if (unlikely(!f64_is_zon3(ua, ub, uc))) { diff --git a/fpu/softfloat-parts.c.inc b/fpu/softfloat-parts.c.inc index XXXXXXX..XXXXXXX 100644 --- a/fpu/softfloat-parts.c.inc +++ b/fpu/softfloat-parts.c.inc @@ -XXX,XX +XXX,XX @@ static FloatPartsN *partsN(muladd_scalbn)(FloatPartsN *a, FloatPartsN *b, a->exp = p_widen.exp; return_normal: - /* TODO: Replace all use of float_muladd_halve_result with scale. */ - if (flags & float_muladd_halve_result) { - a->exp -= 1; - } a->exp += scale; finish_sign: if (flags & float_muladd_negate_result) { -- 2.43.0 This rounding mode is used by Hexagon. Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- include/fpu/softfloat-types.h | 2 ++ fpu/softfloat-parts.c.inc | 3 +++ 2 files changed, 5 insertions(+) diff --git a/include/fpu/softfloat-types.h b/include/fpu/softfloat-types.h index XXXXXXX..XXXXXXX 100644 --- a/include/fpu/softfloat-types.h +++ b/include/fpu/softfloat-types.h @@ -XXX,XX +XXX,XX @@ typedef enum __attribute__((__packed__)) { float_round_to_odd = 5, /* Not an IEEE rounding mode: round to closest odd, overflow to inf */ float_round_to_odd_inf = 6, + /* Not an IEEE rounding mode: round to nearest even, overflow to max */ + float_round_nearest_even_max = 7, } FloatRoundMode; /* diff --git a/fpu/softfloat-parts.c.inc b/fpu/softfloat-parts.c.inc index XXXXXXX..XXXXXXX 100644 --- a/fpu/softfloat-parts.c.inc +++ b/fpu/softfloat-parts.c.inc @@ -XXX,XX +XXX,XX @@ static void partsN(uncanon_normal)(FloatPartsN *p, float_status *s, int exp, flags = 0; switch (s->float_rounding_mode) { + case float_round_nearest_even_max: + overflow_norm = true; + /* fall through */ case float_round_nearest_even: if (N > 64 && frac_lsb == 0) { inc = ((p->frac_hi & 1) || (p->frac_lo & round_mask) != frac_lsbm1 -- 2.43.0 Certain Hexagon instructions suppress changes to the result when the product of fma() is a true zero. Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- include/fpu/softfloat.h | 5 +++++ fpu/softfloat.c | 3 +++ fpu/softfloat-parts.c.inc | 4 +++- 3 files changed, 11 insertions(+), 1 deletion(-) 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 @@ bfloat16 bfloat16_squash_input_denormal(bfloat16 a, float_status *status); | Using these differs from negating an input or output before calling | the muladd function in that this means that a NaN doesn't have its | sign bit inverted before it is propagated. +| +| With float_muladd_suppress_add_product_zero, if A or B is zero +| such that the product is a true zero, then return C without addition. +| This preserves the sign of C when C is +/- 0. Used for Hexagon. *----------------------------------------------------------------------------*/ enum { float_muladd_negate_c = 1, float_muladd_negate_product = 2, float_muladd_negate_result = 4, + float_muladd_suppress_add_product_zero = 8, }; /*---------------------------------------------------------------------------- diff --git a/fpu/softfloat.c b/fpu/softfloat.c index XXXXXXX..XXXXXXX 100644 --- a/fpu/softfloat.c +++ b/fpu/softfloat.c @@ -XXX,XX +XXX,XX @@ float32_muladd(float32 xa, float32 xb, float32 xc, int flags, float_status *s) if (unlikely(!can_use_fpu(s))) { goto soft; } + if (unlikely(flags & float_muladd_suppress_add_product_zero)) { + goto soft; + } float32_input_flush3(&ua.s, &ub.s, &uc.s, s); if (unlikely(!f32_is_zon3(ua, ub, uc))) { diff --git a/fpu/softfloat-parts.c.inc b/fpu/softfloat-parts.c.inc index XXXXXXX..XXXXXXX 100644 --- a/fpu/softfloat-parts.c.inc +++ b/fpu/softfloat-parts.c.inc @@ -XXX,XX +XXX,XX @@ static FloatPartsN *partsN(muladd_scalbn)(FloatPartsN *a, FloatPartsN *b, goto return_normal; } if (c->cls == float_class_zero) { - if (a->sign != c->sign) { + if (flags & float_muladd_suppress_add_product_zero) { + a->sign = c->sign; + } else if (a->sign != c->sign) { goto return_sub_zero; } goto return_zero; -- 2.43.0 There are no special cases for this instruction. Remove internal_mpyf as unused. Reviewed-by: Brian Cain <brian.cain@oss.qualcomm.com> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- target/hexagon/fma_emu.h | 1 - target/hexagon/fma_emu.c | 8 -------- target/hexagon/op_helper.c | 2 +- 3 files changed, 1 insertion(+), 10 deletions(-) diff --git a/target/hexagon/fma_emu.h b/target/hexagon/fma_emu.h index XXXXXXX..XXXXXXX 100644 --- a/target/hexagon/fma_emu.h +++ b/target/hexagon/fma_emu.h @@ -XXX,XX +XXX,XX @@ int32_t float32_getexp(float32 f32); float32 infinite_float32(uint8_t sign); float32 internal_fmafx(float32 a, float32 b, float32 c, int scale, float_status *fp_status); -float32 internal_mpyf(float32 a, float32 b, float_status *fp_status); float64 internal_mpyhh(float64 a, float64 b, unsigned long long int accumulated, float_status *fp_status); diff --git a/target/hexagon/fma_emu.c b/target/hexagon/fma_emu.c index XXXXXXX..XXXXXXX 100644 --- a/target/hexagon/fma_emu.c +++ b/target/hexagon/fma_emu.c @@ -XXX,XX +XXX,XX @@ float32 internal_fmafx(float32 a, float32 b, float32 c, int scale, return accum_round_float32(result, fp_status); } -float32 internal_mpyf(float32 a, float32 b, float_status *fp_status) -{ - if (float32_is_zero(a) || float32_is_zero(b)) { - return float32_mul(a, b, fp_status); - } - return internal_fmafx(a, b, float32_zero, 0, fp_status); -} - float64 internal_mpyhh(float64 a, float64 b, unsigned long long int accumulated, float_status *fp_status) diff --git a/target/hexagon/op_helper.c b/target/hexagon/op_helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/hexagon/op_helper.c +++ b/target/hexagon/op_helper.c @@ -XXX,XX +XXX,XX @@ float32 HELPER(sfmpy)(CPUHexagonState *env, float32 RsV, float32 RtV) { float32 RdV; arch_fpop_start(env); - RdV = internal_mpyf(RsV, RtV, &env->fp_status); + RdV = float32_mul(RsV, RtV, &env->fp_status); arch_fpop_end(env); return RdV; } -- 2.43.0 There are no special cases for this instruction. Reviewed-by: Brian Cain <brian.cain@oss.qualcomm.com> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- target/hexagon/op_helper.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/target/hexagon/op_helper.c b/target/hexagon/op_helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/hexagon/op_helper.c +++ b/target/hexagon/op_helper.c @@ -XXX,XX +XXX,XX @@ float32 HELPER(sffma)(CPUHexagonState *env, float32 RxV, float32 RsV, float32 RtV) { arch_fpop_start(env); - RxV = internal_fmafx(RsV, RtV, RxV, 0, &env->fp_status); + RxV = float32_muladd(RsV, RtV, RxV, 0, &env->fp_status); arch_fpop_end(env); return RxV; } -- 2.43.0 There are no special cases for this instruction. Since hexagon always uses default-nan mode, explicitly negating the first input is unnecessary. Use float_muladd_negate_product instead. Reviewed-by: Brian Cain <brian.cain@oss.qualcomm.com> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- target/hexagon/op_helper.c | 5 ++--- 1 file changed, 2 insertions(+), 3 deletions(-) diff --git a/target/hexagon/op_helper.c b/target/hexagon/op_helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/hexagon/op_helper.c +++ b/target/hexagon/op_helper.c @@ -XXX,XX +XXX,XX @@ float32 HELPER(sffma_sc)(CPUHexagonState *env, float32 RxV, float32 HELPER(sffms)(CPUHexagonState *env, float32 RxV, float32 RsV, float32 RtV) { - float32 neg_RsV; arch_fpop_start(env); - neg_RsV = float32_set_sign(RsV, float32_is_neg(RsV) ? 0 : 1); - RxV = internal_fmafx(neg_RsV, RtV, RxV, 0, &env->fp_status); + RxV = float32_muladd(RsV, RtV, RxV, float_muladd_negate_product, + &env->fp_status); arch_fpop_end(env); return RxV; } -- 2.43.0 This instruction has a special case that 0 * x + c returns c without the normal sign folding that comes with 0 + -0. Use the new float_muladd_suppress_add_product_zero to describe this. Reviewed-by: Brian Cain <brian.cain@oss.qualcomm.com> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- target/hexagon/op_helper.c | 11 +++-------- 1 file changed, 3 insertions(+), 8 deletions(-) diff --git a/target/hexagon/op_helper.c b/target/hexagon/op_helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/hexagon/op_helper.c +++ b/target/hexagon/op_helper.c @@ -XXX,XX +XXX,XX @@ static float32 check_nan(float32 dst, float32 x, float_status *fp_status) float32 HELPER(sffma_sc)(CPUHexagonState *env, float32 RxV, float32 RsV, float32 RtV, float32 PuV) { - size4s_t tmp; arch_fpop_start(env); - RxV = check_nan(RxV, RxV, &env->fp_status); - RxV = check_nan(RxV, RsV, &env->fp_status); - RxV = check_nan(RxV, RtV, &env->fp_status); - tmp = internal_fmafx(RsV, RtV, RxV, fSXTN(8, 64, PuV), &env->fp_status); - if (!(float32_is_zero(RxV) && is_zero_prod(RsV, RtV))) { - RxV = tmp; - } + RxV = float32_muladd_scalbn(RsV, RtV, RxV, fSXTN(8, 64, PuV), + float_muladd_suppress_add_product_zero, + &env->fp_status); arch_fpop_end(env); return RxV; } -- 2.43.0 There are multiple special cases for this instruction. (1) The saturate to normal maximum instead of overflow to infinity is handled by the new float_round_nearest_even_max rounding mode. (2) The 0 * n + c special case is handled by the new float_muladd_suppress_add_product_zero flag. (3) The Inf - Inf -> 0 special case can be detected after the fact by examining float_flag_invalid_isi. Reviewed-by: Brian Cain <brian.cain@oss.qualcomm.com> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- target/hexagon/op_helper.c | 105 +++++++++---------------------------- 1 file changed, 26 insertions(+), 79 deletions(-) diff --git a/target/hexagon/op_helper.c b/target/hexagon/op_helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/hexagon/op_helper.c +++ b/target/hexagon/op_helper.c @@ -XXX,XX +XXX,XX @@ float32 HELPER(sffma)(CPUHexagonState *env, float32 RxV, return RxV; } -static bool is_zero_prod(float32 a, float32 b) -{ - return ((float32_is_zero(a) && is_finite(b)) || - (float32_is_zero(b) && is_finite(a))); -} - -static float32 check_nan(float32 dst, float32 x, float_status *fp_status) -{ - float32 ret = dst; - if (float32_is_any_nan(x)) { - if (extract32(x, 22, 1) == 0) { - float_raise(float_flag_invalid, fp_status); - } - ret = make_float32(0xffffffff); /* nan */ - } - return ret; -} - float32 HELPER(sffma_sc)(CPUHexagonState *env, float32 RxV, float32 RsV, float32 RtV, float32 PuV) { @@ -XXX,XX +XXX,XX @@ float32 HELPER(sffms)(CPUHexagonState *env, float32 RxV, return RxV; } -static bool is_inf_prod(int32_t a, int32_t b) +static float32 do_sffma_lib(CPUHexagonState *env, float32 RxV, + float32 RsV, float32 RtV, int negate) { - return (float32_is_infinity(a) && float32_is_infinity(b)) || - (float32_is_infinity(a) && is_finite(b) && !float32_is_zero(b)) || - (float32_is_infinity(b) && is_finite(a) && !float32_is_zero(a)); + int flags; + + arch_fpop_start(env); + + set_float_rounding_mode(float_round_nearest_even_max, &env->fp_status); + RxV = float32_muladd(RsV, RtV, RxV, + negate | float_muladd_suppress_add_product_zero, + &env->fp_status); + + flags = get_float_exception_flags(&env->fp_status); + if (flags) { + /* Flags are suppressed by this instruction. */ + set_float_exception_flags(0, &env->fp_status); + + /* Return 0 for Inf - Inf. */ + if (flags & float_flag_invalid_isi) { + RxV = 0; + } + } + + arch_fpop_end(env); + return RxV; } float32 HELPER(sffma_lib)(CPUHexagonState *env, float32 RxV, float32 RsV, float32 RtV) { - bool infinp; - bool infminusinf; - float32 tmp; - - arch_fpop_start(env); - set_float_rounding_mode(float_round_nearest_even, &env->fp_status); - infminusinf = float32_is_infinity(RxV) && - is_inf_prod(RsV, RtV) && - (fGETBIT(31, RsV ^ RxV ^ RtV) != 0); - infinp = float32_is_infinity(RxV) || - float32_is_infinity(RtV) || - float32_is_infinity(RsV); - RxV = check_nan(RxV, RxV, &env->fp_status); - RxV = check_nan(RxV, RsV, &env->fp_status); - RxV = check_nan(RxV, RtV, &env->fp_status); - tmp = internal_fmafx(RsV, RtV, RxV, 0, &env->fp_status); - if (!(float32_is_zero(RxV) && is_zero_prod(RsV, RtV))) { - RxV = tmp; - } - set_float_exception_flags(0, &env->fp_status); - if (float32_is_infinity(RxV) && !infinp) { - RxV = RxV - 1; - } - if (infminusinf) { - RxV = 0; - } - arch_fpop_end(env); - return RxV; + return do_sffma_lib(env, RxV, RsV, RtV, 0); } float32 HELPER(sffms_lib)(CPUHexagonState *env, float32 RxV, float32 RsV, float32 RtV) { - bool infinp; - bool infminusinf; - float32 tmp; - - arch_fpop_start(env); - set_float_rounding_mode(float_round_nearest_even, &env->fp_status); - infminusinf = float32_is_infinity(RxV) && - is_inf_prod(RsV, RtV) && - (fGETBIT(31, RsV ^ RxV ^ RtV) == 0); - infinp = float32_is_infinity(RxV) || - float32_is_infinity(RtV) || - float32_is_infinity(RsV); - RxV = check_nan(RxV, RxV, &env->fp_status); - RxV = check_nan(RxV, RsV, &env->fp_status); - RxV = check_nan(RxV, RtV, &env->fp_status); - float32 minus_RsV = float32_sub(float32_zero, RsV, &env->fp_status); - tmp = internal_fmafx(minus_RsV, RtV, RxV, 0, &env->fp_status); - if (!(float32_is_zero(RxV) && is_zero_prod(RsV, RtV))) { - RxV = tmp; - } - set_float_exception_flags(0, &env->fp_status); - if (float32_is_infinity(RxV) && !infinp) { - RxV = RxV - 1; - } - if (infminusinf) { - RxV = 0; - } - arch_fpop_end(env); - return RxV; + return do_sffma_lib(env, RxV, RsV, RtV, float_muladd_negate_product); } float64 HELPER(dfmpyfix)(CPUHexagonState *env, float64 RssV, float64 RttV) -- 2.43.0 The function is now unused. Reviewed-by: Brian Cain <brian.cain@oss.qualcomm.com> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- target/hexagon/fma_emu.h | 2 - target/hexagon/fma_emu.c | 171 --------------------------------------- 2 files changed, 173 deletions(-) diff --git a/target/hexagon/fma_emu.h b/target/hexagon/fma_emu.h index XXXXXXX..XXXXXXX 100644 --- a/target/hexagon/fma_emu.h +++ b/target/hexagon/fma_emu.h @@ -XXX,XX +XXX,XX @@ static inline uint32_t float32_getexp_raw(float32 f32) } int32_t float32_getexp(float32 f32); float32 infinite_float32(uint8_t sign); -float32 internal_fmafx(float32 a, float32 b, float32 c, - int scale, float_status *fp_status); float64 internal_mpyhh(float64 a, float64 b, unsigned long long int accumulated, float_status *fp_status); diff --git a/target/hexagon/fma_emu.c b/target/hexagon/fma_emu.c index XXXXXXX..XXXXXXX 100644 --- a/target/hexagon/fma_emu.c +++ b/target/hexagon/fma_emu.c @@ -XXX,XX +XXX,XX @@ int32_t float64_getexp(float64 f64) return -1; } -static uint64_t float32_getmant(float32 f32) -{ - Float a = { .i = f32 }; - if (float32_is_normal(f32)) { - return a.mant | 1ULL << 23; - } - if (float32_is_zero(f32)) { - return 0; - } - if (float32_is_denormal(f32)) { - return a.mant; - } - return ~0ULL; -} - int32_t float32_getexp(float32 f32) { Float a = { .i = f32 }; @@ -XXX,XX +XXX,XX @@ float32 infinite_float32(uint8_t sign) } /* Return a maximum finite value with the requested sign */ -static float32 maxfinite_float32(uint8_t sign) -{ - if (sign) { - return make_float32(SF_MINUS_MAXF); - } else { - return make_float32(SF_MAXF); - } -} - -/* Return a zero value with requested sign */ -static float32 zero_float32(uint8_t sign) -{ - if (sign) { - return make_float32(0x80000000); - } else { - return float32_zero; - } -} - #define GEN_XF_ROUND(SUFFIX, MANTBITS, INF_EXP, INTERNAL_TYPE) \ static SUFFIX accum_round_##SUFFIX(Accum a, float_status * fp_status) \ { \ @@ -XXX,XX +XXX,XX @@ static SUFFIX accum_round_##SUFFIX(Accum a, float_status * fp_status) \ } GEN_XF_ROUND(float64, DF_MANTBITS, DF_INF_EXP, Double) -GEN_XF_ROUND(float32, SF_MANTBITS, SF_INF_EXP, Float) - -static bool is_inf_prod(float64 a, float64 b) -{ - return ((float64_is_infinity(a) && float64_is_infinity(b)) || - (float64_is_infinity(a) && is_finite(b) && (!float64_is_zero(b))) || - (float64_is_infinity(b) && is_finite(a) && (!float64_is_zero(a)))); -} - -static float64 special_fma(float64 a, float64 b, float64 c, - float_status *fp_status) -{ - float64 ret = make_float64(0); - - /* - * If A multiplied by B is an exact infinity and C is also an infinity - * but with the opposite sign, FMA returns NaN and raises invalid. - */ - uint8_t a_sign = float64_is_neg(a); - uint8_t b_sign = float64_is_neg(b); - uint8_t c_sign = float64_is_neg(c); - if (is_inf_prod(a, b) && float64_is_infinity(c)) { - if ((a_sign ^ b_sign) != c_sign) { - ret = make_float64(DF_NAN); - float_raise(float_flag_invalid, fp_status); - return ret; - } - } - if ((float64_is_infinity(a) && float64_is_zero(b)) || - (float64_is_zero(a) && float64_is_infinity(b))) { - ret = make_float64(DF_NAN); - float_raise(float_flag_invalid, fp_status); - return ret; - } - /* - * If none of the above checks are true and C is a NaN, - * a NaN shall be returned - * If A or B are NaN, a NAN shall be returned. - */ - if (float64_is_any_nan(a) || - float64_is_any_nan(b) || - float64_is_any_nan(c)) { - if (float64_is_any_nan(a) && (fGETBIT(51, a) == 0)) { - float_raise(float_flag_invalid, fp_status); - } - if (float64_is_any_nan(b) && (fGETBIT(51, b) == 0)) { - float_raise(float_flag_invalid, fp_status); - } - if (float64_is_any_nan(c) && (fGETBIT(51, c) == 0)) { - float_raise(float_flag_invalid, fp_status); - } - ret = make_float64(DF_NAN); - return ret; - } - /* - * We have checked for adding opposite-signed infinities. - * Other infinities return infinity with the correct sign - */ - if (float64_is_infinity(c)) { - ret = infinite_float64(c_sign); - return ret; - } - if (float64_is_infinity(a) || float64_is_infinity(b)) { - ret = infinite_float64(a_sign ^ b_sign); - return ret; - } - g_assert_not_reached(); -} - -static float32 special_fmaf(float32 a, float32 b, float32 c, - float_status *fp_status) -{ - float64 aa, bb, cc; - aa = float32_to_float64(a, fp_status); - bb = float32_to_float64(b, fp_status); - cc = float32_to_float64(c, fp_status); - return float64_to_float32(special_fma(aa, bb, cc, fp_status), fp_status); -} - -float32 internal_fmafx(float32 a, float32 b, float32 c, int scale, - float_status *fp_status) -{ - Accum prod; - Accum acc; - Accum result; - accum_init(&prod); - accum_init(&acc); - accum_init(&result); - - uint8_t a_sign = float32_is_neg(a); - uint8_t b_sign = float32_is_neg(b); - uint8_t c_sign = float32_is_neg(c); - if (float32_is_infinity(a) || - float32_is_infinity(b) || - float32_is_infinity(c)) { - return special_fmaf(a, b, c, fp_status); - } - if (float32_is_any_nan(a) || - float32_is_any_nan(b) || - float32_is_any_nan(c)) { - return special_fmaf(a, b, c, fp_status); - } - if ((scale == 0) && (float32_is_zero(a) || float32_is_zero(b))) { - float32 tmp = float32_mul(a, b, fp_status); - tmp = float32_add(tmp, c, fp_status); - return tmp; - } - - /* (a * 2**b) * (c * 2**d) == a*c * 2**(b+d) */ - prod.mant = int128_mul_6464(float32_getmant(a), float32_getmant(b)); - - /* - * Note: extracting the mantissa into an int is multiplying by - * 2**23, so adjust here - */ - prod.exp = float32_getexp(a) + float32_getexp(b) - SF_BIAS - 23; - prod.sign = a_sign ^ b_sign; - if (float32_is_zero(a) || float32_is_zero(b)) { - prod.exp = -2 * WAY_BIG_EXP; - } - if ((scale > 0) && float32_is_denormal(c)) { - acc.mant = int128_mul_6464(0, 0); - acc.exp = -WAY_BIG_EXP; - acc.sign = c_sign; - acc.sticky = 1; - result = accum_add(prod, acc); - } else if (!float32_is_zero(c)) { - acc.mant = int128_mul_6464(float32_getmant(c), 1); - acc.exp = float32_getexp(c); - acc.sign = c_sign; - result = accum_add(prod, acc); - } else { - result = prod; - } - result.exp += scale; - return accum_round_float32(result, fp_status); -} float64 internal_mpyhh(float64 a, float64 b, unsigned long long int accumulated, -- 2.43.0 This massive macro is now only used once. Expand it for use only by float64. Reviewed-by: Brian Cain <brian.cain@oss.qualcomm.com> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- target/hexagon/fma_emu.c | 255 +++++++++++++++++++-------------------- 1 file changed, 127 insertions(+), 128 deletions(-) diff --git a/target/hexagon/fma_emu.c b/target/hexagon/fma_emu.c index XXXXXXX..XXXXXXX 100644 --- a/target/hexagon/fma_emu.c +++ b/target/hexagon/fma_emu.c @@ -XXX,XX +XXX,XX @@ float32 infinite_float32(uint8_t sign) } /* Return a maximum finite value with the requested sign */ -#define GEN_XF_ROUND(SUFFIX, MANTBITS, INF_EXP, INTERNAL_TYPE) \ -static SUFFIX accum_round_##SUFFIX(Accum a, float_status * fp_status) \ -{ \ - if ((int128_gethi(a.mant) == 0) && (int128_getlo(a.mant) == 0) \ - && ((a.guard | a.round | a.sticky) == 0)) { \ - /* result zero */ \ - switch (fp_status->float_rounding_mode) { \ - case float_round_down: \ - return zero_##SUFFIX(1); \ - default: \ - return zero_##SUFFIX(0); \ - } \ - } \ - /* Normalize right */ \ - /* We want MANTBITS bits of mantissa plus the leading one. */ \ - /* That means that we want MANTBITS+1 bits, or 0x000000000000FF_FFFF */ \ - /* So we need to normalize right while the high word is non-zero and \ - * while the low word is nonzero when masked with 0xffe0_0000_0000_0000 */ \ - while ((int128_gethi(a.mant) != 0) || \ - ((int128_getlo(a.mant) >> (MANTBITS + 1)) != 0)) { \ - a = accum_norm_right(a, 1); \ - } \ - /* \ - * OK, now normalize left \ - * We want to normalize left until we have a leading one in bit 24 \ - * Theoretically, we only need to shift a maximum of one to the left if we \ - * shifted out lots of bits from B, or if we had no shift / 1 shift sticky \ - * should be 0 \ - */ \ - while ((int128_getlo(a.mant) & (1ULL << MANTBITS)) == 0) { \ - a = accum_norm_left(a); \ - } \ - /* \ - * OK, now we might need to denormalize because of potential underflow. \ - * We need to do this before rounding, and rounding might make us normal \ - * again \ - */ \ - while (a.exp <= 0) { \ - a = accum_norm_right(a, 1 - a.exp); \ - /* \ - * Do we have underflow? \ - * That's when we get an inexact answer because we ran out of bits \ - * in a denormal. \ - */ \ - if (a.guard || a.round || a.sticky) { \ - float_raise(float_flag_underflow, fp_status); \ - } \ - } \ - /* OK, we're relatively canonical... now we need to round */ \ - if (a.guard || a.round || a.sticky) { \ - float_raise(float_flag_inexact, fp_status); \ - switch (fp_status->float_rounding_mode) { \ - case float_round_to_zero: \ - /* Chop and we're done */ \ - break; \ - case float_round_up: \ - if (a.sign == 0) { \ - a.mant = int128_add(a.mant, int128_one()); \ - } \ - break; \ - case float_round_down: \ - if (a.sign != 0) { \ - a.mant = int128_add(a.mant, int128_one()); \ - } \ - break; \ - default: \ - if (a.round || a.sticky) { \ - /* round up if guard is 1, down if guard is zero */ \ - a.mant = int128_add(a.mant, int128_make64(a.guard)); \ - } else if (a.guard) { \ - /* exactly .5, round up if odd */ \ - a.mant = int128_add(a.mant, int128_and(a.mant, int128_one())); \ - } \ - break; \ - } \ - } \ - /* \ - * OK, now we might have carried all the way up. \ - * So we might need to shr once \ - * at least we know that the lsb should be zero if we rounded and \ - * got a carry out... \ - */ \ - if ((int128_getlo(a.mant) >> (MANTBITS + 1)) != 0) { \ - a = accum_norm_right(a, 1); \ - } \ - /* Overflow? */ \ - if (a.exp >= INF_EXP) { \ - /* Yep, inf result */ \ - float_raise(float_flag_overflow, fp_status); \ - float_raise(float_flag_inexact, fp_status); \ - switch (fp_status->float_rounding_mode) { \ - case float_round_to_zero: \ - return maxfinite_##SUFFIX(a.sign); \ - case float_round_up: \ - if (a.sign == 0) { \ - return infinite_##SUFFIX(a.sign); \ - } else { \ - return maxfinite_##SUFFIX(a.sign); \ - } \ - case float_round_down: \ - if (a.sign != 0) { \ - return infinite_##SUFFIX(a.sign); \ - } else { \ - return maxfinite_##SUFFIX(a.sign); \ - } \ - default: \ - return infinite_##SUFFIX(a.sign); \ - } \ - } \ - /* Underflow? */ \ - if (int128_getlo(a.mant) & (1ULL << MANTBITS)) { \ - /* Leading one means: No, we're normal. So, we should be done... */ \ - INTERNAL_TYPE ret; \ - ret.i = 0; \ - ret.sign = a.sign; \ - ret.exp = a.exp; \ - ret.mant = int128_getlo(a.mant); \ - return ret.i; \ - } \ - assert(a.exp == 1); \ - INTERNAL_TYPE ret; \ - ret.i = 0; \ - ret.sign = a.sign; \ - ret.exp = 0; \ - ret.mant = int128_getlo(a.mant); \ - return ret.i; \ +static float64 accum_round_float64(Accum a, float_status *fp_status) +{ + if ((int128_gethi(a.mant) == 0) && (int128_getlo(a.mant) == 0) + && ((a.guard | a.round | a.sticky) == 0)) { + /* result zero */ + switch (fp_status->float_rounding_mode) { + case float_round_down: + return zero_float64(1); + default: + return zero_float64(0); + } + } + /* + * Normalize right + * We want DF_MANTBITS bits of mantissa plus the leading one. + * That means that we want DF_MANTBITS+1 bits, or 0x000000000000FF_FFFF + * So we need to normalize right while the high word is non-zero and + * while the low word is nonzero when masked with 0xffe0_0000_0000_0000 + */ + while ((int128_gethi(a.mant) != 0) || + ((int128_getlo(a.mant) >> (DF_MANTBITS + 1)) != 0)) { + a = accum_norm_right(a, 1); + } + /* + * OK, now normalize left + * We want to normalize left until we have a leading one in bit 24 + * Theoretically, we only need to shift a maximum of one to the left if we + * shifted out lots of bits from B, or if we had no shift / 1 shift sticky + * should be 0 + */ + while ((int128_getlo(a.mant) & (1ULL << DF_MANTBITS)) == 0) { + a = accum_norm_left(a); + } + /* + * OK, now we might need to denormalize because of potential underflow. + * We need to do this before rounding, and rounding might make us normal + * again + */ + while (a.exp <= 0) { + a = accum_norm_right(a, 1 - a.exp); + /* + * Do we have underflow? + * That's when we get an inexact answer because we ran out of bits + * in a denormal. + */ + if (a.guard || a.round || a.sticky) { + float_raise(float_flag_underflow, fp_status); + } + } + /* OK, we're relatively canonical... now we need to round */ + if (a.guard || a.round || a.sticky) { + float_raise(float_flag_inexact, fp_status); + switch (fp_status->float_rounding_mode) { + case float_round_to_zero: + /* Chop and we're done */ + break; + case float_round_up: + if (a.sign == 0) { + a.mant = int128_add(a.mant, int128_one()); + } + break; + case float_round_down: + if (a.sign != 0) { + a.mant = int128_add(a.mant, int128_one()); + } + break; + default: + if (a.round || a.sticky) { + /* round up if guard is 1, down if guard is zero */ + a.mant = int128_add(a.mant, int128_make64(a.guard)); + } else if (a.guard) { + /* exactly .5, round up if odd */ + a.mant = int128_add(a.mant, int128_and(a.mant, int128_one())); + } + break; + } + } + /* + * OK, now we might have carried all the way up. + * So we might need to shr once + * at least we know that the lsb should be zero if we rounded and + * got a carry out... + */ + if ((int128_getlo(a.mant) >> (DF_MANTBITS + 1)) != 0) { + a = accum_norm_right(a, 1); + } + /* Overflow? */ + if (a.exp >= DF_INF_EXP) { + /* Yep, inf result */ + float_raise(float_flag_overflow, fp_status); + float_raise(float_flag_inexact, fp_status); + switch (fp_status->float_rounding_mode) { + case float_round_to_zero: + return maxfinite_float64(a.sign); + case float_round_up: + if (a.sign == 0) { + return infinite_float64(a.sign); + } else { + return maxfinite_float64(a.sign); + } + case float_round_down: + if (a.sign != 0) { + return infinite_float64(a.sign); + } else { + return maxfinite_float64(a.sign); + } + default: + return infinite_float64(a.sign); + } + } + /* Underflow? */ + if (int128_getlo(a.mant) & (1ULL << DF_MANTBITS)) { + /* Leading one means: No, we're normal. So, we should be done... */ + Double ret; + ret.i = 0; + ret.sign = a.sign; + ret.exp = a.exp; + ret.mant = int128_getlo(a.mant); + return ret.i; + } + assert(a.exp == 1); + Double ret; + ret.i = 0; + ret.sign = a.sign; + ret.exp = 0; + ret.mant = int128_getlo(a.mant); + return ret.i; } -GEN_XF_ROUND(float64, DF_MANTBITS, DF_INF_EXP, Double) - float64 internal_mpyhh(float64 a, float64 b, unsigned long long int accumulated, float_status *fp_status) -- 2.43.0 This structure, with bitfields, is incorrect for big-endian. Use the existing float32_getexp_raw which uses extract32. Reviewed-by: Brian Cain <brian.cain@oss.qualcomm.com> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- target/hexagon/fma_emu.c | 16 +++------------- 1 file changed, 3 insertions(+), 13 deletions(-) diff --git a/target/hexagon/fma_emu.c b/target/hexagon/fma_emu.c index XXXXXXX..XXXXXXX 100644 --- a/target/hexagon/fma_emu.c +++ b/target/hexagon/fma_emu.c @@ -XXX,XX +XXX,XX @@ typedef union { }; } Double; -typedef union { - float f; - uint32_t i; - struct { - uint32_t mant:23; - uint32_t exp:8; - uint32_t sign:1; - }; -} Float; - static uint64_t float64_getmant(float64 f64) { Double a = { .i = f64 }; @@ -XXX,XX +XXX,XX @@ int32_t float64_getexp(float64 f64) int32_t float32_getexp(float32 f32) { - Float a = { .i = f32 }; + int exp = float32_getexp_raw(f32); if (float32_is_normal(f32)) { - return a.exp; + return exp; } if (float32_is_denormal(f32)) { - return a.exp + 1; + return exp + 1; } return -1; } -- 2.43.0 This structure, with bitfields, is incorrect for big-endian. Use extract64 and deposit64 instead. Reviewed-by: Brian Cain <brian.cain@oss.qualcomm.com> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- target/hexagon/fma_emu.c | 46 ++++++++++++++-------------------------- 1 file changed, 16 insertions(+), 30 deletions(-) diff --git a/target/hexagon/fma_emu.c b/target/hexagon/fma_emu.c index XXXXXXX..XXXXXXX 100644 --- a/target/hexagon/fma_emu.c +++ b/target/hexagon/fma_emu.c @@ -XXX,XX +XXX,XX @@ #define WAY_BIG_EXP 4096 -typedef union { - double f; - uint64_t i; - struct { - uint64_t mant:52; - uint64_t exp:11; - uint64_t sign:1; - }; -} Double; - static uint64_t float64_getmant(float64 f64) { - Double a = { .i = f64 }; + uint64_t mant = extract64(f64, 0, 52); if (float64_is_normal(f64)) { - return a.mant | 1ULL << 52; + return mant | 1ULL << 52; } if (float64_is_zero(f64)) { return 0; } if (float64_is_denormal(f64)) { - return a.mant; + return mant; } return ~0ULL; } int32_t float64_getexp(float64 f64) { - Double a = { .i = f64 }; + int exp = extract64(f64, 52, 11); if (float64_is_normal(f64)) { - return a.exp; + return exp; } if (float64_is_denormal(f64)) { - return a.exp + 1; + return exp + 1; } return -1; } @@ -XXX,XX +XXX,XX @@ float32 infinite_float32(uint8_t sign) /* Return a maximum finite value with the requested sign */ static float64 accum_round_float64(Accum a, float_status *fp_status) { + uint64_t ret; + if ((int128_gethi(a.mant) == 0) && (int128_getlo(a.mant) == 0) && ((a.guard | a.round | a.sticky) == 0)) { /* result zero */ @@ -XXX,XX +XXX,XX @@ static float64 accum_round_float64(Accum a, float_status *fp_status) } } /* Underflow? */ - if (int128_getlo(a.mant) & (1ULL << DF_MANTBITS)) { + ret = int128_getlo(a.mant); + if (ret & (1ULL << DF_MANTBITS)) { /* Leading one means: No, we're normal. So, we should be done... */ - Double ret; - ret.i = 0; - ret.sign = a.sign; - ret.exp = a.exp; - ret.mant = int128_getlo(a.mant); - return ret.i; + ret = deposit64(ret, 52, 11, a.exp); + } else { + assert(a.exp == 1); + ret = deposit64(ret, 52, 11, 0); } - assert(a.exp == 1); - Double ret; - ret.i = 0; - ret.sign = a.sign; - ret.exp = 0; - ret.mant = int128_getlo(a.mant); - return ret.i; + ret = deposit64(ret, 63, 1, a.sign); + return ret; } float64 internal_mpyhh(float64 a, float64 b, -- 2.43.0 No need to open-code 64x64->128-bit multiplication. Reviewed-by: Brian Cain <brian.cain@oss.qualcomm.com> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- target/hexagon/fma_emu.c | 32 +++----------------------------- 1 file changed, 3 insertions(+), 29 deletions(-) diff --git a/target/hexagon/fma_emu.c b/target/hexagon/fma_emu.c index XXXXXXX..XXXXXXX 100644 --- a/target/hexagon/fma_emu.c +++ b/target/hexagon/fma_emu.c @@ -XXX,XX +XXX,XX @@ int32_t float32_getexp(float32 f32) return -1; } -static uint32_t int128_getw0(Int128 x) -{ - return int128_getlo(x); -} - -static uint32_t int128_getw1(Int128 x) -{ - return int128_getlo(x) >> 32; -} - static Int128 int128_mul_6464(uint64_t ai, uint64_t bi) { - Int128 a, b; - uint64_t pp0, pp1a, pp1b, pp1s, pp2; + uint64_t l, h; - a = int128_make64(ai); - b = int128_make64(bi); - pp0 = (uint64_t)int128_getw0(a) * (uint64_t)int128_getw0(b); - pp1a = (uint64_t)int128_getw1(a) * (uint64_t)int128_getw0(b); - pp1b = (uint64_t)int128_getw1(b) * (uint64_t)int128_getw0(a); - pp2 = (uint64_t)int128_getw1(a) * (uint64_t)int128_getw1(b); - - pp1s = pp1a + pp1b; - if ((pp1s < pp1a) || (pp1s < pp1b)) { - pp2 += (1ULL << 32); - } - uint64_t ret_low = pp0 + (pp1s << 32); - if ((ret_low < pp0) || (ret_low < (pp1s << 32))) { - pp2 += 1; - } - - return int128_make128(ret_low, pp2 + (pp1s >> 32)); + mulu64(&l, &h, ai, bi); + return int128_make128(l, h); } static Int128 int128_sub_borrow(Int128 a, Int128 b, int borrow) -- 2.43.0 Initialize x with accumulated via direct assignment, rather than multiplying by 1. Reviewed-by: Brian Cain <brian.cain@oss.qualcomm.com> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- target/hexagon/fma_emu.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/target/hexagon/fma_emu.c b/target/hexagon/fma_emu.c index XXXXXXX..XXXXXXX 100644 --- a/target/hexagon/fma_emu.c +++ b/target/hexagon/fma_emu.c @@ -XXX,XX +XXX,XX @@ float64 internal_mpyhh(float64 a, float64 b, float64_is_infinity(b)) { return float64_mul(a, b, fp_status); } - x.mant = int128_mul_6464(accumulated, 1); + x.mant = int128_make64(accumulated); x.sticky = sticky; prod = fGETUWORD(1, float64_getmant(a)) * fGETUWORD(1, float64_getmant(b)); x.mant = int128_add(x.mant, int128_mul_6464(prod, 0x100000000ULL)); -- 2.43.0 Convert all targets simultaneously, as the gen_intermediate_code function disappears from the target. While there are possible workarounds, they're larger than simply performing the conversion. Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- include/exec/translator.h | 14 -------------- include/hw/core/tcg-cpu-ops.h | 13 +++++++++++++ target/alpha/cpu.h | 2 ++ target/arm/internals.h | 2 ++ target/avr/cpu.h | 2 ++ target/hexagon/cpu.h | 2 ++ target/hppa/cpu.h | 2 ++ target/i386/tcg/helper-tcg.h | 2 ++ target/loongarch/internals.h | 2 ++ target/m68k/cpu.h | 2 ++ target/microblaze/cpu.h | 2 ++ target/mips/tcg/tcg-internal.h | 2 ++ target/openrisc/cpu.h | 2 ++ target/ppc/cpu.h | 2 ++ target/riscv/cpu.h | 3 +++ target/rx/cpu.h | 2 ++ target/s390x/s390x-internal.h | 2 ++ target/sh4/cpu.h | 2 ++ target/sparc/cpu.h | 2 ++ target/tricore/cpu.h | 2 ++ target/xtensa/cpu.h | 2 ++ accel/tcg/cpu-exec.c | 8 +++++--- accel/tcg/translate-all.c | 8 +++++--- target/alpha/cpu.c | 1 + target/alpha/translate.c | 4 ++-- target/arm/cpu.c | 1 + target/arm/tcg/cpu-v7m.c | 1 + target/arm/tcg/translate.c | 5 ++--- target/avr/cpu.c | 1 + target/avr/translate.c | 6 +++--- target/hexagon/cpu.c | 1 + target/hexagon/translate.c | 4 ++-- target/hppa/cpu.c | 1 + target/hppa/translate.c | 4 ++-- target/i386/tcg/tcg-cpu.c | 1 + target/i386/tcg/translate.c | 5 ++--- target/loongarch/cpu.c | 1 + target/loongarch/tcg/translate.c | 4 ++-- target/m68k/cpu.c | 1 + target/m68k/translate.c | 4 ++-- target/microblaze/cpu.c | 1 + target/microblaze/translate.c | 4 ++-- target/mips/cpu.c | 1 + target/mips/tcg/translate.c | 4 ++-- target/openrisc/cpu.c | 1 + target/openrisc/translate.c | 4 ++-- target/ppc/cpu_init.c | 1 + target/ppc/translate.c | 4 ++-- target/riscv/tcg/tcg-cpu.c | 1 + target/riscv/translate.c | 4 ++-- target/rx/cpu.c | 1 + target/rx/translate.c | 4 ++-- target/s390x/cpu.c | 1 + target/s390x/tcg/translate.c | 4 ++-- target/sh4/cpu.c | 1 + target/sh4/translate.c | 4 ++-- target/sparc/cpu.c | 1 + target/sparc/translate.c | 4 ++-- target/tricore/cpu.c | 1 + target/tricore/translate.c | 5 ++--- target/xtensa/cpu.c | 1 + target/xtensa/translate.c | 4 ++-- 62 files changed, 121 insertions(+), 62 deletions(-) diff --git a/include/exec/translator.h b/include/exec/translator.h index XXXXXXX..XXXXXXX 100644 --- a/include/exec/translator.h +++ b/include/exec/translator.h @@ -XXX,XX +XXX,XX @@ #include "qemu/bswap.h" #include "exec/vaddr.h" -/** - * gen_intermediate_code - * @cpu: cpu context - * @tb: translation block - * @max_insns: max number of instructions to translate - * @pc: guest virtual program counter address - * @host_pc: host physical program counter address - * - * This function must be provided by the target, which should create - * the target-specific DisasContext, and then invoke translator_loop. - */ -void gen_intermediate_code(CPUState *cpu, TranslationBlock *tb, int *max_insns, - vaddr pc, void *host_pc); - /** * DisasJumpType: * @DISAS_NEXT: Next instruction in program order. diff --git a/include/hw/core/tcg-cpu-ops.h b/include/hw/core/tcg-cpu-ops.h index XXXXXXX..XXXXXXX 100644 --- a/include/hw/core/tcg-cpu-ops.h +++ b/include/hw/core/tcg-cpu-ops.h @@ -XXX,XX +XXX,XX @@ struct TCGCPUOps { * Called when the first CPU is realized. */ void (*initialize)(void); + /** + * @translate_code: Translate guest instructions to TCGOps + * @cpu: cpu context + * @tb: translation block + * @max_insns: max number of instructions to translate + * @pc: guest virtual program counter address + * @host_pc: host physical program counter address + * + * This function must be provided by the target, which should create + * the target-specific DisasContext, and then invoke translator_loop. + */ + void (*translate_code)(CPUState *cpu, TranslationBlock *tb, + int *max_insns, vaddr pc, void *host_pc); /** * @synchronize_from_tb: Synchronize state from a TCG #TranslationBlock * diff --git a/target/alpha/cpu.h b/target/alpha/cpu.h index XXXXXXX..XXXXXXX 100644 --- a/target/alpha/cpu.h +++ b/target/alpha/cpu.h @@ -XXX,XX +XXX,XX @@ enum { }; void alpha_translate_init(void); +void alpha_translate_code(CPUState *cs, TranslationBlock *tb, + int *max_insns, vaddr pc, void *host_pc); #define CPU_RESOLVING_TYPE TYPE_ALPHA_CPU diff --git a/target/arm/internals.h b/target/arm/internals.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/internals.h +++ b/target/arm/internals.h @@ -XXX,XX +XXX,XX @@ void init_cpreg_list(ARMCPU *cpu); void arm_cpu_register_gdb_regs_for_features(ARMCPU *cpu); void arm_translate_init(void); +void arm_translate_code(CPUState *cs, TranslationBlock *tb, + int *max_insns, vaddr pc, void *host_pc); void arm_cpu_register_gdb_commands(ARMCPU *cpu); void aarch64_cpu_register_gdb_commands(ARMCPU *cpu, GString *, diff --git a/target/avr/cpu.h b/target/avr/cpu.h index XXXXXXX..XXXXXXX 100644 --- a/target/avr/cpu.h +++ b/target/avr/cpu.h @@ -XXX,XX +XXX,XX @@ static inline void set_avr_feature(CPUAVRState *env, int feature) } void avr_cpu_tcg_init(void); +void avr_cpu_translate_code(CPUState *cs, TranslationBlock *tb, + int *max_insns, vaddr pc, void *host_pc); int cpu_avr_exec(CPUState *cpu); diff --git a/target/hexagon/cpu.h b/target/hexagon/cpu.h index XXXXXXX..XXXXXXX 100644 --- a/target/hexagon/cpu.h +++ b/target/hexagon/cpu.h @@ -XXX,XX +XXX,XX @@ static inline void cpu_get_tb_cpu_state(CPUHexagonState *env, vaddr *pc, typedef HexagonCPU ArchCPU; void hexagon_translate_init(void); +void hexagon_translate_code(CPUState *cs, TranslationBlock *tb, + int *max_insns, vaddr pc, void *host_pc); #include "exec/cpu-all.h" diff --git a/target/hppa/cpu.h b/target/hppa/cpu.h index XXXXXXX..XXXXXXX 100644 --- a/target/hppa/cpu.h +++ b/target/hppa/cpu.h @@ -XXX,XX +XXX,XX @@ static inline int HPPA_BTLB_ENTRIES(CPUHPPAState *env) } void hppa_translate_init(void); +void hppa_translate_code(CPUState *cs, TranslationBlock *tb, + int *max_insns, vaddr pc, void *host_pc); #define CPU_RESOLVING_TYPE TYPE_HPPA_CPU diff --git a/target/i386/tcg/helper-tcg.h b/target/i386/tcg/helper-tcg.h index XXXXXXX..XXXXXXX 100644 --- a/target/i386/tcg/helper-tcg.h +++ b/target/i386/tcg/helper-tcg.h @@ -XXX,XX +XXX,XX @@ static inline target_long lshift(target_long x, int n) /* translate.c */ void tcg_x86_init(void); +void x86_translate_code(CPUState *cs, TranslationBlock *tb, + int *max_insns, vaddr pc, void *host_pc); /* excp_helper.c */ G_NORETURN void raise_exception(CPUX86State *env, int exception_index); diff --git a/target/loongarch/internals.h b/target/loongarch/internals.h index XXXXXXX..XXXXXXX 100644 --- a/target/loongarch/internals.h +++ b/target/loongarch/internals.h @@ -XXX,XX +XXX,XX @@ #define TARGET_VIRT_MASK MAKE_64BIT_MASK(0, TARGET_VIRT_ADDR_SPACE_BITS) void loongarch_translate_init(void); +void loongarch_translate_code(CPUState *cs, TranslationBlock *tb, + int *max_insns, vaddr pc, void *host_pc); void G_NORETURN do_raise_exception(CPULoongArchState *env, uint32_t exception, diff --git a/target/m68k/cpu.h b/target/m68k/cpu.h index XXXXXXX..XXXXXXX 100644 --- a/target/m68k/cpu.h +++ b/target/m68k/cpu.h @@ -XXX,XX +XXX,XX @@ int m68k_cpu_gdb_read_register(CPUState *cpu, GByteArray *buf, int reg); int m68k_cpu_gdb_write_register(CPUState *cpu, uint8_t *buf, int reg); void m68k_tcg_init(void); +void m68k_translate_code(CPUState *cs, TranslationBlock *tb, + int *max_insns, vaddr pc, void *host_pc); void m68k_cpu_init_gdb(M68kCPU *cpu); uint32_t cpu_m68k_get_ccr(CPUM68KState *env); void cpu_m68k_set_ccr(CPUM68KState *env, uint32_t); diff --git a/target/microblaze/cpu.h b/target/microblaze/cpu.h index XXXXXXX..XXXXXXX 100644 --- a/target/microblaze/cpu.h +++ b/target/microblaze/cpu.h @@ -XXX,XX +XXX,XX @@ static inline void mb_cpu_write_msr(CPUMBState *env, uint32_t val) } void mb_tcg_init(void); +void mb_translate_code(CPUState *cs, TranslationBlock *tb, + int *max_insns, vaddr pc, void *host_pc); #define CPU_RESOLVING_TYPE TYPE_MICROBLAZE_CPU diff --git a/target/mips/tcg/tcg-internal.h b/target/mips/tcg/tcg-internal.h index XXXXXXX..XXXXXXX 100644 --- a/target/mips/tcg/tcg-internal.h +++ b/target/mips/tcg/tcg-internal.h @@ -XXX,XX +XXX,XX @@ #include "cpu.h" void mips_tcg_init(void); +void mips_translate_code(CPUState *cs, TranslationBlock *tb, + int *max_insns, vaddr pc, void *host_pc); void mips_cpu_synchronize_from_tb(CPUState *cs, const TranslationBlock *tb); G_NORETURN void mips_cpu_do_unaligned_access(CPUState *cpu, vaddr addr, diff --git a/target/openrisc/cpu.h b/target/openrisc/cpu.h index XXXXXXX..XXXXXXX 100644 --- a/target/openrisc/cpu.h +++ b/target/openrisc/cpu.h @@ -XXX,XX +XXX,XX @@ void openrisc_cpu_dump_state(CPUState *cpu, FILE *f, int flags); int openrisc_cpu_gdb_read_register(CPUState *cpu, GByteArray *buf, int reg); int openrisc_cpu_gdb_write_register(CPUState *cpu, uint8_t *buf, int reg); void openrisc_translate_init(void); +void openrisc_translate_code(CPUState *cs, TranslationBlock *tb, + int *max_insns, vaddr pc, void *host_pc); int print_insn_or1k(bfd_vma addr, disassemble_info *info); #ifndef CONFIG_USER_ONLY diff --git a/target/ppc/cpu.h b/target/ppc/cpu.h index XXXXXXX..XXXXXXX 100644 --- a/target/ppc/cpu.h +++ b/target/ppc/cpu.h @@ -XXX,XX +XXX,XX @@ extern const VMStateDescription vmstate_ppc_cpu; /*****************************************************************************/ void ppc_translate_init(void); +void ppc_translate_code(CPUState *cs, TranslationBlock *tb, + int *max_insns, vaddr pc, void *host_pc); #if !defined(CONFIG_USER_ONLY) void ppc_store_sdr1(CPUPPCState *env, target_ulong value); diff --git a/target/riscv/cpu.h b/target/riscv/cpu.h index XXXXXXX..XXXXXXX 100644 --- a/target/riscv/cpu.h +++ b/target/riscv/cpu.h @@ -XXX,XX +XXX,XX @@ RISCVException smstateen_acc_ok(CPURISCVState *env, int index, uint64_t bit); void riscv_cpu_set_mode(CPURISCVState *env, target_ulong newpriv, bool virt_en); void riscv_translate_init(void); +void riscv_translate_code(CPUState *cs, TranslationBlock *tb, + int *max_insns, vaddr pc, void *host_pc); + G_NORETURN void riscv_raise_exception(CPURISCVState *env, uint32_t exception, uintptr_t pc); diff --git a/target/rx/cpu.h b/target/rx/cpu.h index XXXXXXX..XXXXXXX 100644 --- a/target/rx/cpu.h +++ b/target/rx/cpu.h @@ -XXX,XX +XXX,XX @@ int rx_cpu_gdb_read_register(CPUState *cpu, GByteArray *buf, int reg); int rx_cpu_gdb_write_register(CPUState *cpu, uint8_t *buf, int reg); void rx_translate_init(void); +void rx_translate_code(CPUState *cs, TranslationBlock *tb, + int *max_insns, vaddr pc, void *host_pc); void rx_cpu_unpack_psw(CPURXState *env, uint32_t psw, int rte); #include "exec/cpu-all.h" diff --git a/target/s390x/s390x-internal.h b/target/s390x/s390x-internal.h index XXXXXXX..XXXXXXX 100644 --- a/target/s390x/s390x-internal.h +++ b/target/s390x/s390x-internal.h @@ -XXX,XX +XXX,XX @@ void handle_diag_308(CPUS390XState *env, uint64_t r1, uint64_t r3, /* translate.c */ void s390x_translate_init(void); +void s390x_translate_code(CPUState *cs, TranslationBlock *tb, + int *max_insns, vaddr pc, void *host_pc); void s390x_restore_state_to_opc(CPUState *cs, const TranslationBlock *tb, const uint64_t *data); diff --git a/target/sh4/cpu.h b/target/sh4/cpu.h index XXXXXXX..XXXXXXX 100644 --- a/target/sh4/cpu.h +++ b/target/sh4/cpu.h @@ -XXX,XX +XXX,XX @@ G_NORETURN void superh_cpu_do_unaligned_access(CPUState *cpu, vaddr addr, uintptr_t retaddr); void sh4_translate_init(void); +void sh4_translate_code(CPUState *cs, TranslationBlock *tb, + int *max_insns, vaddr pc, void *host_pc); #if !defined(CONFIG_USER_ONLY) hwaddr superh_cpu_get_phys_page_debug(CPUState *cpu, vaddr addr); diff --git a/target/sparc/cpu.h b/target/sparc/cpu.h index XXXXXXX..XXXXXXX 100644 --- a/target/sparc/cpu.h +++ b/target/sparc/cpu.h @@ -XXX,XX +XXX,XX @@ int sparc_cpu_memory_rw_debug(CPUState *cpu, vaddr addr, /* translate.c */ void sparc_tcg_init(void); +void sparc_translate_code(CPUState *cs, TranslationBlock *tb, + int *max_insns, vaddr pc, void *host_pc); /* fop_helper.c */ target_ulong cpu_get_fsr(CPUSPARCState *); diff --git a/target/tricore/cpu.h b/target/tricore/cpu.h index XXXXXXX..XXXXXXX 100644 --- a/target/tricore/cpu.h +++ b/target/tricore/cpu.h @@ -XXX,XX +XXX,XX @@ FIELD(TB_FLAGS, PRIV, 0, 2) void cpu_state_reset(CPUTriCoreState *s); void tricore_tcg_init(void); +void tricore_translate_code(CPUState *cs, TranslationBlock *tb, + int *max_insns, vaddr pc, void *host_pc); static inline void cpu_get_tb_cpu_state(CPUTriCoreState *env, vaddr *pc, uint64_t *cs_base, uint32_t *flags) diff --git a/target/xtensa/cpu.h b/target/xtensa/cpu.h index XXXXXXX..XXXXXXX 100644 --- a/target/xtensa/cpu.h +++ b/target/xtensa/cpu.h @@ -XXX,XX +XXX,XX @@ G_NORETURN void xtensa_cpu_do_unaligned_access(CPUState *cpu, vaddr addr, void xtensa_collect_sr_names(const XtensaConfig *config); void xtensa_translate_init(void); +void xtensa_translate_code(CPUState *cs, TranslationBlock *tb, + int *max_insns, vaddr pc, void *host_pc); void **xtensa_get_regfile_by_name(const char *name, int entries, int bits); void xtensa_breakpoint_handler(CPUState *cs); void xtensa_register_core(XtensaConfigList *node); diff --git a/accel/tcg/cpu-exec.c b/accel/tcg/cpu-exec.c index XXXXXXX..XXXXXXX 100644 --- a/accel/tcg/cpu-exec.c +++ b/accel/tcg/cpu-exec.c @@ -XXX,XX +XXX,XX @@ bool tcg_exec_realizefn(CPUState *cpu, Error **errp) if (!tcg_target_initialized) { /* Check mandatory TCGCPUOps handlers */ + const TCGCPUOps *tcg_ops = cpu->cc->tcg_ops; #ifndef CONFIG_USER_ONLY - assert(cpu->cc->tcg_ops->cpu_exec_halt); - assert(cpu->cc->tcg_ops->cpu_exec_interrupt); + assert(tcg_ops->cpu_exec_halt); + assert(tcg_ops->cpu_exec_interrupt); #endif /* !CONFIG_USER_ONLY */ - cpu->cc->tcg_ops->initialize(); + assert(tcg_ops->translate_code); + tcg_ops->initialize(); tcg_target_initialized = true; } diff --git a/accel/tcg/translate-all.c b/accel/tcg/translate-all.c index XXXXXXX..XXXXXXX 100644 --- a/accel/tcg/translate-all.c +++ b/accel/tcg/translate-all.c @@ -XXX,XX +XXX,XX @@ static int setjmp_gen_code(CPUArchState *env, TranslationBlock *tb, tcg_func_start(tcg_ctx); - tcg_ctx->cpu = env_cpu(env); - gen_intermediate_code(env_cpu(env), tb, max_insns, pc, host_pc); + CPUState *cs = env_cpu(env); + tcg_ctx->cpu = cs; + cs->cc->tcg_ops->translate_code(cs, tb, max_insns, pc, host_pc); + assert(tb->size != 0); tcg_ctx->cpu = NULL; *max_insns = tb->icount; @@ -XXX,XX +XXX,XX @@ TranslationBlock *tb_gen_code(CPUState *cpu, /* * Overflow of code_gen_buffer, or the current slice of it. * - * TODO: We don't need to re-do gen_intermediate_code, nor + * TODO: We don't need to re-do tcg_ops->translate_code, nor * should we re-do the tcg optimization currently hidden * inside tcg_gen_code. All that should be required is to * flush the TBs, allocate a new TB, re-initialize it per diff --git a/target/alpha/cpu.c b/target/alpha/cpu.c index XXXXXXX..XXXXXXX 100644 --- a/target/alpha/cpu.c +++ b/target/alpha/cpu.c @@ -XXX,XX +XXX,XX @@ static const struct SysemuCPUOps alpha_sysemu_ops = { static const TCGCPUOps alpha_tcg_ops = { .initialize = alpha_translate_init, + .translate_code = alpha_translate_code, .synchronize_from_tb = alpha_cpu_synchronize_from_tb, .restore_state_to_opc = alpha_restore_state_to_opc, diff --git a/target/alpha/translate.c b/target/alpha/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/alpha/translate.c +++ b/target/alpha/translate.c @@ -XXX,XX +XXX,XX @@ static const TranslatorOps alpha_tr_ops = { .tb_stop = alpha_tr_tb_stop, }; -void gen_intermediate_code(CPUState *cpu, TranslationBlock *tb, int *max_insns, - vaddr pc, void *host_pc) +void alpha_translate_code(CPUState *cpu, TranslationBlock *tb, + int *max_insns, vaddr pc, void *host_pc) { DisasContext dc; translator_loop(cpu, tb, max_insns, pc, host_pc, &alpha_tr_ops, &dc.base); 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 struct SysemuCPUOps arm_sysemu_ops = { #ifdef CONFIG_TCG static const TCGCPUOps arm_tcg_ops = { .initialize = arm_translate_init, + .translate_code = arm_translate_code, .synchronize_from_tb = arm_cpu_synchronize_from_tb, .debug_excp_handler = arm_debug_excp_handler, .restore_state_to_opc = arm_restore_state_to_opc, diff --git a/target/arm/tcg/cpu-v7m.c b/target/arm/tcg/cpu-v7m.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/tcg/cpu-v7m.c +++ b/target/arm/tcg/cpu-v7m.c @@ -XXX,XX +XXX,XX @@ static void cortex_m55_initfn(Object *obj) static const TCGCPUOps arm_v7m_tcg_ops = { .initialize = arm_translate_init, + .translate_code = arm_translate_code, .synchronize_from_tb = arm_cpu_synchronize_from_tb, .debug_excp_handler = arm_debug_excp_handler, .restore_state_to_opc = arm_restore_state_to_opc, diff --git a/target/arm/tcg/translate.c b/target/arm/tcg/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/tcg/translate.c +++ b/target/arm/tcg/translate.c @@ -XXX,XX +XXX,XX @@ static const TranslatorOps thumb_translator_ops = { .tb_stop = arm_tr_tb_stop, }; -/* generate intermediate code for basic block 'tb'. */ -void gen_intermediate_code(CPUState *cpu, TranslationBlock *tb, int *max_insns, - vaddr pc, void *host_pc) +void arm_translate_code(CPUState *cpu, TranslationBlock *tb, + int *max_insns, vaddr pc, void *host_pc) { DisasContext dc = { }; const TranslatorOps *ops = &arm_translator_ops; diff --git a/target/avr/cpu.c b/target/avr/cpu.c index XXXXXXX..XXXXXXX 100644 --- a/target/avr/cpu.c +++ b/target/avr/cpu.c @@ -XXX,XX +XXX,XX @@ static const struct SysemuCPUOps avr_sysemu_ops = { static const TCGCPUOps avr_tcg_ops = { .initialize = avr_cpu_tcg_init, + .translate_code = avr_cpu_translate_code, .synchronize_from_tb = avr_cpu_synchronize_from_tb, .restore_state_to_opc = avr_restore_state_to_opc, .cpu_exec_interrupt = avr_cpu_exec_interrupt, diff --git a/target/avr/translate.c b/target/avr/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/avr/translate.c +++ b/target/avr/translate.c @@ -XXX,XX +XXX,XX @@ static bool trans_WDR(DisasContext *ctx, arg_WDR *a) * * - translate() * - canonicalize_skip() - * - gen_intermediate_code() + * - translate_code() * - restore_state_to_opc() * */ @@ -XXX,XX +XXX,XX @@ static const TranslatorOps avr_tr_ops = { .tb_stop = avr_tr_tb_stop, }; -void gen_intermediate_code(CPUState *cs, TranslationBlock *tb, int *max_insns, - vaddr pc, void *host_pc) +void avr_cpu_translate_code(CPUState *cs, TranslationBlock *tb, + int *max_insns, vaddr pc, void *host_pc) { DisasContext dc = { }; translator_loop(cs, tb, max_insns, pc, host_pc, &avr_tr_ops, &dc.base); diff --git a/target/hexagon/cpu.c b/target/hexagon/cpu.c index XXXXXXX..XXXXXXX 100644 --- a/target/hexagon/cpu.c +++ b/target/hexagon/cpu.c @@ -XXX,XX +XXX,XX @@ static void hexagon_cpu_init(Object *obj) static const TCGCPUOps hexagon_tcg_ops = { .initialize = hexagon_translate_init, + .translate_code = hexagon_translate_code, .synchronize_from_tb = hexagon_cpu_synchronize_from_tb, .restore_state_to_opc = hexagon_restore_state_to_opc, }; diff --git a/target/hexagon/translate.c b/target/hexagon/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/hexagon/translate.c +++ b/target/hexagon/translate.c @@ -XXX,XX +XXX,XX @@ static const TranslatorOps hexagon_tr_ops = { .tb_stop = hexagon_tr_tb_stop, }; -void gen_intermediate_code(CPUState *cs, TranslationBlock *tb, int *max_insns, - vaddr pc, void *host_pc) +void hexagon_translate_code(CPUState *cs, TranslationBlock *tb, + int *max_insns, vaddr pc, void *host_pc) { DisasContext ctx; diff --git a/target/hppa/cpu.c b/target/hppa/cpu.c index XXXXXXX..XXXXXXX 100644 --- a/target/hppa/cpu.c +++ b/target/hppa/cpu.c @@ -XXX,XX +XXX,XX @@ static const struct SysemuCPUOps hppa_sysemu_ops = { static const TCGCPUOps hppa_tcg_ops = { .initialize = hppa_translate_init, + .translate_code = hppa_translate_code, .synchronize_from_tb = hppa_cpu_synchronize_from_tb, .restore_state_to_opc = hppa_restore_state_to_opc, diff --git a/target/hppa/translate.c b/target/hppa/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/hppa/translate.c +++ b/target/hppa/translate.c @@ -XXX,XX +XXX,XX @@ static const TranslatorOps hppa_tr_ops = { #endif }; -void gen_intermediate_code(CPUState *cs, TranslationBlock *tb, int *max_insns, - vaddr pc, void *host_pc) +void hppa_translate_code(CPUState *cs, TranslationBlock *tb, + int *max_insns, vaddr pc, void *host_pc) { DisasContext ctx = { }; translator_loop(cs, tb, max_insns, pc, host_pc, &hppa_tr_ops, &ctx.base); diff --git a/target/i386/tcg/tcg-cpu.c b/target/i386/tcg/tcg-cpu.c index XXXXXXX..XXXXXXX 100644 --- a/target/i386/tcg/tcg-cpu.c +++ b/target/i386/tcg/tcg-cpu.c @@ -XXX,XX +XXX,XX @@ static bool x86_debug_check_breakpoint(CPUState *cs) static const TCGCPUOps x86_tcg_ops = { .initialize = tcg_x86_init, + .translate_code = x86_translate_code, .synchronize_from_tb = x86_cpu_synchronize_from_tb, .restore_state_to_opc = x86_restore_state_to_opc, .cpu_exec_enter = x86_cpu_exec_enter, diff --git a/target/i386/tcg/translate.c b/target/i386/tcg/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/i386/tcg/translate.c +++ b/target/i386/tcg/translate.c @@ -XXX,XX +XXX,XX @@ static const TranslatorOps i386_tr_ops = { .tb_stop = i386_tr_tb_stop, }; -/* generate intermediate code for basic block 'tb'. */ -void gen_intermediate_code(CPUState *cpu, TranslationBlock *tb, int *max_insns, - vaddr pc, void *host_pc) +void x86_translate_code(CPUState *cpu, TranslationBlock *tb, + int *max_insns, vaddr pc, void *host_pc) { DisasContext dc; diff --git a/target/loongarch/cpu.c b/target/loongarch/cpu.c index XXXXXXX..XXXXXXX 100644 --- a/target/loongarch/cpu.c +++ b/target/loongarch/cpu.c @@ -XXX,XX +XXX,XX @@ static void loongarch_cpu_dump_state(CPUState *cs, FILE *f, int flags) static const TCGCPUOps loongarch_tcg_ops = { .initialize = loongarch_translate_init, + .translate_code = loongarch_translate_code, .synchronize_from_tb = loongarch_cpu_synchronize_from_tb, .restore_state_to_opc = loongarch_restore_state_to_opc, diff --git a/target/loongarch/tcg/translate.c b/target/loongarch/tcg/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/loongarch/tcg/translate.c +++ b/target/loongarch/tcg/translate.c @@ -XXX,XX +XXX,XX @@ static const TranslatorOps loongarch_tr_ops = { .tb_stop = loongarch_tr_tb_stop, }; -void gen_intermediate_code(CPUState *cs, TranslationBlock *tb, int *max_insns, - vaddr pc, void *host_pc) +void loongarch_translate_code(CPUState *cs, TranslationBlock *tb, + int *max_insns, vaddr pc, void *host_pc) { DisasContext ctx; diff --git a/target/m68k/cpu.c b/target/m68k/cpu.c index XXXXXXX..XXXXXXX 100644 --- a/target/m68k/cpu.c +++ b/target/m68k/cpu.c @@ -XXX,XX +XXX,XX @@ static const struct SysemuCPUOps m68k_sysemu_ops = { static const TCGCPUOps m68k_tcg_ops = { .initialize = m68k_tcg_init, + .translate_code = m68k_translate_code, .restore_state_to_opc = m68k_restore_state_to_opc, #ifndef CONFIG_USER_ONLY diff --git a/target/m68k/translate.c b/target/m68k/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/m68k/translate.c +++ b/target/m68k/translate.c @@ -XXX,XX +XXX,XX @@ static const TranslatorOps m68k_tr_ops = { .tb_stop = m68k_tr_tb_stop, }; -void gen_intermediate_code(CPUState *cpu, TranslationBlock *tb, int *max_insns, - vaddr pc, void *host_pc) +void m68k_translate_code(CPUState *cpu, TranslationBlock *tb, + int *max_insns, vaddr pc, void *host_pc) { DisasContext dc; translator_loop(cpu, tb, max_insns, pc, host_pc, &m68k_tr_ops, &dc.base); diff --git a/target/microblaze/cpu.c b/target/microblaze/cpu.c index XXXXXXX..XXXXXXX 100644 --- a/target/microblaze/cpu.c +++ b/target/microblaze/cpu.c @@ -XXX,XX +XXX,XX @@ static const struct SysemuCPUOps mb_sysemu_ops = { static const TCGCPUOps mb_tcg_ops = { .initialize = mb_tcg_init, + .translate_code = mb_translate_code, .synchronize_from_tb = mb_cpu_synchronize_from_tb, .restore_state_to_opc = mb_restore_state_to_opc, diff --git a/target/microblaze/translate.c b/target/microblaze/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/microblaze/translate.c +++ b/target/microblaze/translate.c @@ -XXX,XX +XXX,XX @@ static const TranslatorOps mb_tr_ops = { .tb_stop = mb_tr_tb_stop, }; -void gen_intermediate_code(CPUState *cpu, TranslationBlock *tb, int *max_insns, - vaddr pc, void *host_pc) +void mb_translate_code(CPUState *cpu, TranslationBlock *tb, + int *max_insns, vaddr pc, void *host_pc) { DisasContext dc; translator_loop(cpu, tb, max_insns, pc, host_pc, &mb_tr_ops, &dc.base); diff --git a/target/mips/cpu.c b/target/mips/cpu.c index XXXXXXX..XXXXXXX 100644 --- a/target/mips/cpu.c +++ b/target/mips/cpu.c @@ -XXX,XX +XXX,XX @@ static const Property mips_cpu_properties[] = { #include "hw/core/tcg-cpu-ops.h" static const TCGCPUOps mips_tcg_ops = { .initialize = mips_tcg_init, + .translate_code = mips_translate_code, .synchronize_from_tb = mips_cpu_synchronize_from_tb, .restore_state_to_opc = mips_restore_state_to_opc, diff --git a/target/mips/tcg/translate.c b/target/mips/tcg/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/mips/tcg/translate.c +++ b/target/mips/tcg/translate.c @@ -XXX,XX +XXX,XX @@ static const TranslatorOps mips_tr_ops = { .tb_stop = mips_tr_tb_stop, }; -void gen_intermediate_code(CPUState *cs, TranslationBlock *tb, int *max_insns, - vaddr pc, void *host_pc) +void mips_translate_code(CPUState *cs, TranslationBlock *tb, + int *max_insns, vaddr pc, void *host_pc) { DisasContext ctx; diff --git a/target/openrisc/cpu.c b/target/openrisc/cpu.c index XXXXXXX..XXXXXXX 100644 --- a/target/openrisc/cpu.c +++ b/target/openrisc/cpu.c @@ -XXX,XX +XXX,XX @@ static const struct SysemuCPUOps openrisc_sysemu_ops = { static const TCGCPUOps openrisc_tcg_ops = { .initialize = openrisc_translate_init, + .translate_code = openrisc_translate_code, .synchronize_from_tb = openrisc_cpu_synchronize_from_tb, .restore_state_to_opc = openrisc_restore_state_to_opc, diff --git a/target/openrisc/translate.c b/target/openrisc/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/openrisc/translate.c +++ b/target/openrisc/translate.c @@ -XXX,XX +XXX,XX @@ static const TranslatorOps openrisc_tr_ops = { .tb_stop = openrisc_tr_tb_stop, }; -void gen_intermediate_code(CPUState *cs, TranslationBlock *tb, int *max_insns, - vaddr pc, void *host_pc) +void openrisc_translate_code(CPUState *cs, TranslationBlock *tb, + int *max_insns, vaddr pc, void *host_pc) { DisasContext ctx; diff --git a/target/ppc/cpu_init.c b/target/ppc/cpu_init.c index XXXXXXX..XXXXXXX 100644 --- a/target/ppc/cpu_init.c +++ b/target/ppc/cpu_init.c @@ -XXX,XX +XXX,XX @@ static const struct SysemuCPUOps ppc_sysemu_ops = { static const TCGCPUOps ppc_tcg_ops = { .initialize = ppc_translate_init, + .translate_code = ppc_translate_code, .restore_state_to_opc = ppc_restore_state_to_opc, #ifdef CONFIG_USER_ONLY diff --git a/target/ppc/translate.c b/target/ppc/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/ppc/translate.c +++ b/target/ppc/translate.c @@ -XXX,XX +XXX,XX @@ static const TranslatorOps ppc_tr_ops = { .tb_stop = ppc_tr_tb_stop, }; -void gen_intermediate_code(CPUState *cs, TranslationBlock *tb, int *max_insns, - vaddr pc, void *host_pc) +void ppc_translate_code(CPUState *cs, TranslationBlock *tb, + int *max_insns, vaddr pc, void *host_pc) { DisasContext ctx; diff --git a/target/riscv/tcg/tcg-cpu.c b/target/riscv/tcg/tcg-cpu.c index XXXXXXX..XXXXXXX 100644 --- a/target/riscv/tcg/tcg-cpu.c +++ b/target/riscv/tcg/tcg-cpu.c @@ -XXX,XX +XXX,XX @@ static void riscv_restore_state_to_opc(CPUState *cs, static const TCGCPUOps riscv_tcg_ops = { .initialize = riscv_translate_init, + .translate_code = riscv_translate_code, .synchronize_from_tb = riscv_cpu_synchronize_from_tb, .restore_state_to_opc = riscv_restore_state_to_opc, diff --git a/target/riscv/translate.c b/target/riscv/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/riscv/translate.c +++ b/target/riscv/translate.c @@ -XXX,XX +XXX,XX @@ static const TranslatorOps riscv_tr_ops = { .tb_stop = riscv_tr_tb_stop, }; -void gen_intermediate_code(CPUState *cs, TranslationBlock *tb, int *max_insns, - vaddr pc, void *host_pc) +void riscv_translate_code(CPUState *cs, TranslationBlock *tb, + int *max_insns, vaddr pc, void *host_pc) { DisasContext ctx; diff --git a/target/rx/cpu.c b/target/rx/cpu.c index XXXXXXX..XXXXXXX 100644 --- a/target/rx/cpu.c +++ b/target/rx/cpu.c @@ -XXX,XX +XXX,XX @@ static const struct SysemuCPUOps rx_sysemu_ops = { static const TCGCPUOps rx_tcg_ops = { .initialize = rx_translate_init, + .translate_code = rx_translate_code, .synchronize_from_tb = rx_cpu_synchronize_from_tb, .restore_state_to_opc = rx_restore_state_to_opc, .tlb_fill = rx_cpu_tlb_fill, diff --git a/target/rx/translate.c b/target/rx/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/rx/translate.c +++ b/target/rx/translate.c @@ -XXX,XX +XXX,XX @@ static const TranslatorOps rx_tr_ops = { .tb_stop = rx_tr_tb_stop, }; -void gen_intermediate_code(CPUState *cs, TranslationBlock *tb, int *max_insns, - vaddr pc, void *host_pc) +void rx_translate_code(CPUState *cs, TranslationBlock *tb, + int *max_insns, vaddr pc, void *host_pc) { DisasContext dc; diff --git a/target/s390x/cpu.c b/target/s390x/cpu.c index XXXXXXX..XXXXXXX 100644 --- a/target/s390x/cpu.c +++ b/target/s390x/cpu.c @@ -XXX,XX +XXX,XX @@ void cpu_get_tb_cpu_state(CPUS390XState *env, vaddr *pc, static const TCGCPUOps s390_tcg_ops = { .initialize = s390x_translate_init, + .translate_code = s390x_translate_code, .restore_state_to_opc = s390x_restore_state_to_opc, #ifdef CONFIG_USER_ONLY diff --git a/target/s390x/tcg/translate.c b/target/s390x/tcg/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/s390x/tcg/translate.c +++ b/target/s390x/tcg/translate.c @@ -XXX,XX +XXX,XX @@ static const TranslatorOps s390x_tr_ops = { .disas_log = s390x_tr_disas_log, }; -void gen_intermediate_code(CPUState *cs, TranslationBlock *tb, int *max_insns, - vaddr pc, void *host_pc) +void s390x_translate_code(CPUState *cs, TranslationBlock *tb, + int *max_insns, vaddr pc, void *host_pc) { DisasContext dc; diff --git a/target/sh4/cpu.c b/target/sh4/cpu.c index XXXXXXX..XXXXXXX 100644 --- a/target/sh4/cpu.c +++ b/target/sh4/cpu.c @@ -XXX,XX +XXX,XX @@ static const struct SysemuCPUOps sh4_sysemu_ops = { static const TCGCPUOps superh_tcg_ops = { .initialize = sh4_translate_init, + .translate_code = sh4_translate_code, .synchronize_from_tb = superh_cpu_synchronize_from_tb, .restore_state_to_opc = superh_restore_state_to_opc, diff --git a/target/sh4/translate.c b/target/sh4/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/sh4/translate.c +++ b/target/sh4/translate.c @@ -XXX,XX +XXX,XX @@ static const TranslatorOps sh4_tr_ops = { .tb_stop = sh4_tr_tb_stop, }; -void gen_intermediate_code(CPUState *cs, TranslationBlock *tb, int *max_insns, - vaddr pc, void *host_pc) +void sh4_translate_code(CPUState *cs, TranslationBlock *tb, + int *max_insns, vaddr pc, void *host_pc) { DisasContext ctx; diff --git a/target/sparc/cpu.c b/target/sparc/cpu.c index XXXXXXX..XXXXXXX 100644 --- a/target/sparc/cpu.c +++ b/target/sparc/cpu.c @@ -XXX,XX +XXX,XX @@ static const struct SysemuCPUOps sparc_sysemu_ops = { static const TCGCPUOps sparc_tcg_ops = { .initialize = sparc_tcg_init, + .translate_code = sparc_translate_code, .synchronize_from_tb = sparc_cpu_synchronize_from_tb, .restore_state_to_opc = sparc_restore_state_to_opc, diff --git a/target/sparc/translate.c b/target/sparc/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/sparc/translate.c +++ b/target/sparc/translate.c @@ -XXX,XX +XXX,XX @@ static const TranslatorOps sparc_tr_ops = { .tb_stop = sparc_tr_tb_stop, }; -void gen_intermediate_code(CPUState *cs, TranslationBlock *tb, int *max_insns, - vaddr pc, void *host_pc) +void sparc_translate_code(CPUState *cs, TranslationBlock *tb, + int *max_insns, vaddr pc, void *host_pc) { DisasContext dc = {}; diff --git a/target/tricore/cpu.c b/target/tricore/cpu.c index XXXXXXX..XXXXXXX 100644 --- a/target/tricore/cpu.c +++ b/target/tricore/cpu.c @@ -XXX,XX +XXX,XX @@ static const struct SysemuCPUOps tricore_sysemu_ops = { static const TCGCPUOps tricore_tcg_ops = { .initialize = tricore_tcg_init, + .translate_code = tricore_translate_code, .synchronize_from_tb = tricore_cpu_synchronize_from_tb, .restore_state_to_opc = tricore_restore_state_to_opc, .tlb_fill = tricore_cpu_tlb_fill, diff --git a/target/tricore/translate.c b/target/tricore/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/tricore/translate.c +++ b/target/tricore/translate.c @@ -XXX,XX +XXX,XX @@ static const TranslatorOps tricore_tr_ops = { .tb_stop = tricore_tr_tb_stop, }; - -void gen_intermediate_code(CPUState *cs, TranslationBlock *tb, int *max_insns, - vaddr pc, void *host_pc) +void tricore_translate_code(CPUState *cs, TranslationBlock *tb, + int *max_insns, vaddr pc, void *host_pc) { DisasContext ctx; translator_loop(cs, tb, max_insns, pc, host_pc, diff --git a/target/xtensa/cpu.c b/target/xtensa/cpu.c index XXXXXXX..XXXXXXX 100644 --- a/target/xtensa/cpu.c +++ b/target/xtensa/cpu.c @@ -XXX,XX +XXX,XX @@ static const struct SysemuCPUOps xtensa_sysemu_ops = { static const TCGCPUOps xtensa_tcg_ops = { .initialize = xtensa_translate_init, + .translate_code = xtensa_translate_code, .debug_excp_handler = xtensa_breakpoint_handler, .restore_state_to_opc = xtensa_restore_state_to_opc, diff --git a/target/xtensa/translate.c b/target/xtensa/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/xtensa/translate.c +++ b/target/xtensa/translate.c @@ -XXX,XX +XXX,XX @@ static const TranslatorOps xtensa_translator_ops = { .tb_stop = xtensa_tr_tb_stop, }; -void gen_intermediate_code(CPUState *cpu, TranslationBlock *tb, int *max_insns, - vaddr pc, void *host_pc) +void xtensa_translate_code(CPUState *cpu, TranslationBlock *tb, + int *max_insns, vaddr pc, void *host_pc) { DisasContext dc = {}; translator_loop(cpu, tb, max_insns, pc, host_pc, -- 2.43.0