From: Anton Johansson via <qemu-devel@nongnu.org> Signed-off-by: Anton Johansson <a...@rev.ng> Reviewed-by: Philippe Mathieu-Daudé <phi...@linaro.org> Message-Id: <20230227135202.9710-5-a...@rev.ng> Signed-off-by: Richard Henderson <richard.hender...@linaro.org> --- accel/tcg/internal.h | 10 ++++---- accel/tcg/tb-jmp-cache.h | 48 +++++++++++++++++++-------------------- accel/tcg/cpu-exec.c | 8 +++---- accel/tcg/perf.c | 2 +- accel/tcg/tb-maint.c | 8 +++---- accel/tcg/translate-all.c | 14 ++++++------ 6 files changed, 44 insertions(+), 46 deletions(-)
diff --git a/accel/tcg/internal.h b/accel/tcg/internal.h index 130d7fd564..5daa4cf0d7 100644 --- a/accel/tcg/internal.h +++ b/accel/tcg/internal.h @@ -57,11 +57,11 @@ void cpu_restore_state_from_tb(CPUState *cpu, TranslationBlock *tb, /* Return the current PC from CPU, which may be cached in TB. */ static inline target_ulong log_pc(CPUState *cpu, const TranslationBlock *tb) { -#if TARGET_TB_PCREL - return cpu->cc->get_pc(cpu); -#else - return tb_pc(tb); -#endif + if (tb_cflags(tb) & CF_PCREL) { + return cpu->cc->get_pc(cpu); + } else { + return tb_pc(tb); + } } extern int64_t max_delay; diff --git a/accel/tcg/tb-jmp-cache.h b/accel/tcg/tb-jmp-cache.h index b3f6e78835..083939b302 100644 --- a/accel/tcg/tb-jmp-cache.h +++ b/accel/tcg/tb-jmp-cache.h @@ -14,53 +14,51 @@ /* * Accessed in parallel; all accesses to 'tb' must be atomic. - * For TARGET_TB_PCREL, accesses to 'pc' must be protected by - * a load_acquire/store_release to 'tb'. + * For CF_PCREL, accesses to 'pc' must be protected by a + * load_acquire/store_release to 'tb'. */ struct CPUJumpCache { struct rcu_head rcu; struct { TranslationBlock *tb; -#if TARGET_TB_PCREL target_ulong pc; -#endif } array[TB_JMP_CACHE_SIZE]; }; static inline TranslationBlock * -tb_jmp_cache_get_tb(CPUJumpCache *jc, uint32_t hash) +tb_jmp_cache_get_tb(CPUJumpCache *jc, uint32_t cflags, uint32_t hash) { -#if TARGET_TB_PCREL - /* Use acquire to ensure current load of pc from jc. */ - return qatomic_load_acquire(&jc->array[hash].tb); -#else - /* Use rcu_read to ensure current load of pc from *tb. */ - return qatomic_rcu_read(&jc->array[hash].tb); -#endif + if (cflags & CF_PCREL) { + /* Use acquire to ensure current load of pc from jc. */ + return qatomic_load_acquire(&jc->array[hash].tb); + } else { + /* Use rcu_read to ensure current load of pc from *tb. */ + return qatomic_rcu_read(&jc->array[hash].tb); + } } static inline target_ulong tb_jmp_cache_get_pc(CPUJumpCache *jc, uint32_t hash, TranslationBlock *tb) { -#if TARGET_TB_PCREL - return jc->array[hash].pc; -#else - return tb_pc(tb); -#endif + if (tb_cflags(tb) & CF_PCREL) { + return jc->array[hash].pc; + } else { + return tb_pc(tb); + } } static inline void tb_jmp_cache_set(CPUJumpCache *jc, uint32_t hash, TranslationBlock *tb, target_ulong pc) { -#if TARGET_TB_PCREL - jc->array[hash].pc = pc; - /* Use store_release on tb to ensure pc is written first. */ - qatomic_store_release(&jc->array[hash].tb, tb); -#else - /* Use the pc value already stored in tb->pc. */ - qatomic_set(&jc->array[hash].tb, tb); -#endif + if (tb_cflags(tb) & CF_PCREL) { + jc->array[hash].pc = pc; + /* Use store_release on tb to ensure pc is written first. */ + qatomic_store_release(&jc->array[hash].tb, tb); + } else{ + /* Use the pc value already stored in tb->pc. */ + qatomic_set(&jc->array[hash].tb, tb); + } } #endif /* ACCEL_TCG_TB_JMP_CACHE_H */ diff --git a/accel/tcg/cpu-exec.c b/accel/tcg/cpu-exec.c index ef557e57aa..013856f103 100644 --- a/accel/tcg/cpu-exec.c +++ b/accel/tcg/cpu-exec.c @@ -183,7 +183,7 @@ static bool tb_lookup_cmp(const void *p, const void *d) const TranslationBlock *tb = p; const struct tb_desc *desc = d; - if ((TARGET_TB_PCREL || tb_pc(tb) == desc->pc) && + if ((tb_cflags(tb) & CF_PCREL || tb_pc(tb) == desc->pc) && tb_page_addr0(tb) == desc->page_addr0 && tb->cs_base == desc->cs_base && tb->flags == desc->flags && @@ -235,7 +235,7 @@ static TranslationBlock *tb_htable_lookup(CPUState *cpu, target_ulong pc, return NULL; } desc.page_addr0 = phys_pc; - h = tb_hash_func(phys_pc, (TARGET_TB_PCREL ? 0 : pc), + h = tb_hash_func(phys_pc, (cflags & CF_PCREL ? 0 : pc), flags, cflags, *cpu->trace_dstate); return qht_lookup_custom(&tb_ctx.htable, &desc, h, tb_lookup_cmp); } @@ -254,7 +254,7 @@ static inline TranslationBlock *tb_lookup(CPUState *cpu, target_ulong pc, hash = tb_jmp_cache_hash_func(pc); jc = cpu->tb_jmp_cache; - tb = tb_jmp_cache_get_tb(jc, hash); + tb = tb_jmp_cache_get_tb(jc, cflags, hash); if (likely(tb && tb_jmp_cache_get_pc(jc, hash, tb) == pc && @@ -457,7 +457,7 @@ cpu_tb_exec(CPUState *cpu, TranslationBlock *itb, int *tb_exit) if (cc->tcg_ops->synchronize_from_tb) { cc->tcg_ops->synchronize_from_tb(cpu, last_tb); } else { - assert(!TARGET_TB_PCREL); + tcg_debug_assert(!(tb_cflags(last_tb) & CF_PCREL)); assert(cc->set_pc); cc->set_pc(cpu, tb_pc(last_tb)); } diff --git a/accel/tcg/perf.c b/accel/tcg/perf.c index ae19f6e28f..65e35ea3b9 100644 --- a/accel/tcg/perf.c +++ b/accel/tcg/perf.c @@ -328,7 +328,7 @@ void perf_report_code(uint64_t guest_pc, TranslationBlock *tb, for (insn = 0; insn < tb->icount; insn++) { /* FIXME: This replicates the restore_state_to_opc() logic. */ q[insn].address = tcg_ctx->gen_insn_data[insn][0]; - if (TARGET_TB_PCREL) { + if (tb_cflags(tb) & CF_PCREL) { q[insn].address |= (guest_pc & TARGET_PAGE_MASK); } else { #if defined(TARGET_I386) diff --git a/accel/tcg/tb-maint.c b/accel/tcg/tb-maint.c index b3d6529ae2..2dbc2ce172 100644 --- a/accel/tcg/tb-maint.c +++ b/accel/tcg/tb-maint.c @@ -44,7 +44,7 @@ static bool tb_cmp(const void *ap, const void *bp) const TranslationBlock *a = ap; const TranslationBlock *b = bp; - return ((TARGET_TB_PCREL || tb_pc(a) == tb_pc(b)) && + return ((tb_cflags(a) & CF_PCREL || tb_pc(a) == tb_pc(b)) && a->cs_base == b->cs_base && a->flags == b->flags && (tb_cflags(a) & ~CF_INVALID) == (tb_cflags(b) & ~CF_INVALID) && @@ -847,7 +847,7 @@ static void tb_jmp_cache_inval_tb(TranslationBlock *tb) { CPUState *cpu; - if (TARGET_TB_PCREL) { + if (tb_cflags(tb) & CF_PCREL) { /* A TB may be at any virtual address */ CPU_FOREACH(cpu) { tcg_flush_jmp_cache(cpu); @@ -885,7 +885,7 @@ static void do_tb_phys_invalidate(TranslationBlock *tb, bool rm_from_page_list) /* remove the TB from the hash list */ phys_pc = tb_page_addr0(tb); - h = tb_hash_func(phys_pc, (TARGET_TB_PCREL ? 0 : tb_pc(tb)), + h = tb_hash_func(phys_pc, (orig_cflags & CF_PCREL ? 0 : tb_pc(tb)), tb->flags, orig_cflags, tb->trace_vcpu_dstate); if (!qht_remove(&tb_ctx.htable, tb, h)) { return; @@ -966,7 +966,7 @@ TranslationBlock *tb_link_page(TranslationBlock *tb, tb_page_addr_t phys_pc, tb_record(tb, p, p2); /* add in the hash table */ - h = tb_hash_func(phys_pc, (TARGET_TB_PCREL ? 0 : tb_pc(tb)), + h = tb_hash_func(phys_pc, (tb->cflags & CF_PCREL ? 0 : tb_pc(tb)), tb->flags, tb->cflags, tb->trace_vcpu_dstate); qht_insert(&tb_ctx.htable, tb, h, &existing_tb); diff --git a/accel/tcg/translate-all.c b/accel/tcg/translate-all.c index 84f129337b..7969c1ee4c 100644 --- a/accel/tcg/translate-all.c +++ b/accel/tcg/translate-all.c @@ -134,7 +134,7 @@ static int encode_search(TranslationBlock *tb, uint8_t *block) for (j = 0; j < TARGET_INSN_START_WORDS; ++j) { if (i == 0) { - prev = (!TARGET_TB_PCREL && j == 0 ? tb_pc(tb) : 0); + prev = (!(tb_cflags(tb) & CF_PCREL) && j == 0 ? tb_pc(tb) : 0); } else { prev = tcg_ctx->gen_insn_data[i - 1][j]; } @@ -169,7 +169,7 @@ static int cpu_unwind_data_from_tb(TranslationBlock *tb, uintptr_t host_pc, } memset(data, 0, sizeof(uint64_t) * TARGET_INSN_START_WORDS); - if (!TARGET_TB_PCREL) { + if (!(tb_cflags(tb) & CF_PCREL)) { data[0] = tb_pc(tb); } @@ -340,9 +340,9 @@ TranslationBlock *tb_gen_code(CPUState *cpu, gen_code_buf = tcg_ctx->code_gen_ptr; tb->tc.ptr = tcg_splitwx_to_rx(gen_code_buf); -#if !TARGET_TB_PCREL - tb->pc = pc; -#endif + if (!(cflags & CF_PCREL)) { + tb->pc = pc; + } tb->cs_base = cs_base; tb->flags = flags; tb->cflags = cflags; @@ -407,8 +407,8 @@ TranslationBlock *tb_gen_code(CPUState *cpu, tb->tc.size = gen_code_size; /* - * For TARGET_TB_PCREL, attribute all executions of the generated - * code to its first mapping. + * For CF_PCREL, attribute all executions of the generated code + * to its first mapping. */ perf_report_code(pc, tb, tcg_splitwx_to_rx(gen_code_buf)); -- 2.34.1