The following changes since commit ff56877e911782dedc9a424233fd3f62369c258c: Merge remote-tracking branch 'remotes/kraxel/tags/vga-20181015-pull-request' into staging (2018-10-15 15:03:45 +0100) are available in the Git repository at: https://github.com/rth7680/qemu.git tags/pull-tcg-20181016 for you to fetch changes up to e3e9d1ea20c75718ce7c528c588a0a497f12f750: cputlb: read CPUTLBEntry.addr_write atomically (2018-10-16 10:04:27 -0700) ---------------------------------------------------------------- Queued tcg patches ---------------------------------------------------------------- Emilio G. Cota (10): tcg: access cpu->icount_decr.u16.high with atomics tcg: fix use of uninitialized variable under CONFIG_PROFILER tcg: plug holes in struct TCGProfile tcg: distribute tcg_time into TCG contexts target/alpha: remove tlb_flush from alpha_cpu_initfn target/unicore32: remove tlb_flush from uc32_init_fn exec: introduce tlb_init cputlb: fix assert_cpu_is_self macro cputlb: serialize tlb updates with env->tlb_lock cputlb: read CPUTLBEntry.addr_write atomically Richard Henderson (11): tcg: Implement CPU_LOG_TB_NOCHAIN during expansion tcg: Add tlb_index and tlb_entry helpers tcg: Split CONFIG_ATOMIC128 target/i386: Convert to HAVE_CMPXCHG128 target/arm: Convert to HAVE_CMPXCHG128 target/arm: Check HAVE_CMPXCHG128 at translate time target/ppc: Convert to HAVE_CMPXCHG128 and HAVE_ATOMIC128 target/s390x: Convert to HAVE_CMPXCHG128 and HAVE_ATOMIC128 target/s390x: Split do_cdsg, do_lpq, do_stpq target/s390x: Skip wout, cout helpers if op helper does not return target/s390x: Check HAVE_ATOMIC128 and HAVE_CMPXCHG128 at translate accel/tcg/atomic_template.h | 20 +++- accel/tcg/softmmu_template.h | 64 +++++----- include/exec/cpu-defs.h | 3 + include/exec/cpu_ldst.h | 30 ++++- include/exec/cpu_ldst_template.h | 25 ++-- include/exec/exec-all.h | 8 ++ include/qemu/atomic128.h | 155 ++++++++++++++++++++++++ include/qemu/timer.h | 1 - target/ppc/helper.h | 2 +- tcg/tcg.h | 20 ++-- accel/tcg/cpu-exec.c | 2 +- accel/tcg/cputlb.c | 235 +++++++++++++++++++----------------- accel/tcg/tcg-all.c | 2 +- accel/tcg/translate-all.c | 2 +- accel/tcg/user-exec.c | 5 +- cpus.c | 3 +- exec.c | 1 + monitor.c | 13 +- qom/cpu.c | 2 +- target/alpha/cpu.c | 1 - target/arm/helper-a64.c | 251 +++++++++++++++++++-------------------- target/arm/translate-a64.c | 38 +++--- target/i386/mem_helper.c | 9 +- target/ppc/mem_helper.c | 33 ++++- target/ppc/translate.c | 115 +++++++++--------- target/s390x/mem_helper.c | 202 +++++++++++++++---------------- target/s390x/translate.c | 45 +++++-- target/unicore32/cpu.c | 2 - tcg/tcg-op.c | 9 +- tcg/tcg.c | 25 +++- configure | 19 +++ 31 files changed, 830 insertions(+), 512 deletions(-) create mode 100644 include/qemu/atomic128.h Rather than test NOCHAIN before linking, do not emit the goto_tb opcode at all. We already do this for goto_ptr. Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- accel/tcg/cpu-exec.c | 2 +- tcg/tcg-op.c | 9 ++++++++- 2 files changed, 9 insertions(+), 2 deletions(-) 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 @@ static inline TranslationBlock *tb_find(CPUState *cpu, } #endif /* See if we can patch the calling TB. */ - if (last_tb && !qemu_loglevel_mask(CPU_LOG_TB_NOCHAIN)) { + if (last_tb) { tb_add_jump(last_tb, tb_exit, tb); } return tb; diff --git a/tcg/tcg-op.c b/tcg/tcg-op.c index XXXXXXX..XXXXXXX 100644 --- a/tcg/tcg-op.c +++ b/tcg/tcg-op.c @@ -XXX,XX +XXX,XX @@ void tcg_gen_exit_tb(TranslationBlock *tb, unsigned idx) seen this numbered exit before, via tcg_gen_goto_tb. */ tcg_debug_assert(tcg_ctx->goto_tb_issue_mask & (1 << idx)); #endif + /* When not chaining, exit without indicating a link. */ + if (qemu_loglevel_mask(CPU_LOG_TB_NOCHAIN)) { + val = 0; + } } else { /* This is an exit via the exitreq label. */ tcg_debug_assert(idx == TB_EXIT_REQUESTED); @@ -XXX,XX +XXX,XX @@ void tcg_gen_goto_tb(unsigned idx) tcg_debug_assert((tcg_ctx->goto_tb_issue_mask & (1 << idx)) == 0); tcg_ctx->goto_tb_issue_mask |= 1 << idx; #endif - tcg_gen_op1i(INDEX_op_goto_tb, idx); + /* When not chaining, we simply fall through to the "fallback" exit. */ + if (!qemu_loglevel_mask(CPU_LOG_TB_NOCHAIN)) { + tcg_gen_op1i(INDEX_op_goto_tb, idx); + } } void tcg_gen_lookup_and_goto_ptr(void) -- 2.17.2 From: "Emilio G. Cota" <cota@braap.org> Consistently access u16.high with atomics to avoid undefined behaviour in MTTCG. Note that icount_decr.u16.low is only used in icount mode, so regular accesses to it are OK. Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Signed-off-by: Emilio G. Cota <cota@braap.org> Message-Id: <20181010144853.13005-2-cota@braap.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- accel/tcg/tcg-all.c | 2 +- accel/tcg/translate-all.c | 2 +- qom/cpu.c | 2 +- 3 files changed, 3 insertions(+), 3 deletions(-) diff --git a/accel/tcg/tcg-all.c b/accel/tcg/tcg-all.c index XXXXXXX..XXXXXXX 100644 --- a/accel/tcg/tcg-all.c +++ b/accel/tcg/tcg-all.c @@ -XXX,XX +XXX,XX @@ static void tcg_handle_interrupt(CPUState *cpu, int mask) if (!qemu_cpu_is_self(cpu)) { qemu_cpu_kick(cpu); } else { - cpu->icount_decr.u16.high = -1; + atomic_set(&cpu->icount_decr.u16.high, -1); if (use_icount && !cpu->can_do_io && (mask & ~old_mask) != 0) { 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 @@ void cpu_interrupt(CPUState *cpu, int mask) { g_assert(qemu_mutex_iothread_locked()); cpu->interrupt_request |= mask; - cpu->icount_decr.u16.high = -1; + atomic_set(&cpu->icount_decr.u16.high, -1); } /* diff --git a/qom/cpu.c b/qom/cpu.c index XXXXXXX..XXXXXXX 100644 --- a/qom/cpu.c +++ b/qom/cpu.c @@ -XXX,XX +XXX,XX @@ static void cpu_common_reset(CPUState *cpu) cpu->mem_io_pc = 0; cpu->mem_io_vaddr = 0; cpu->icount_extra = 0; - cpu->icount_decr.u32 = 0; + atomic_set(&cpu->icount_decr.u32, 0); cpu->can_do_io = 1; cpu->exception_index = -1; cpu->crash_occurred = false; -- 2.17.2 From: "Emilio G. Cota" <cota@braap.org> We forgot to initialize n in commit 15fa08f845 ("tcg: Dynamically allocate TCGOps", 2017-12-29). Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com> Signed-off-by: Emilio G. Cota <cota@braap.org> Message-Id: <20181010144853.13005-3-cota@braap.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- tcg/tcg.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) 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 @@ int tcg_gen_code(TCGContext *s, TranslationBlock *tb) #ifdef CONFIG_PROFILER { - int n; + int n = 0; QTAILQ_FOREACH(op, &s->ops, link) { n++; -- 2.17.2 From: "Emilio G. Cota" <cota@braap.org> This plugs two 4-byte holes in 64-bit. Signed-off-by: Emilio G. Cota <cota@braap.org> Message-Id: <20181010144853.13005-4-cota@braap.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- tcg/tcg.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/tcg/tcg.h b/tcg/tcg.h index XXXXXXX..XXXXXXX 100644 --- a/tcg/tcg.h +++ b/tcg/tcg.h @@ -XXX,XX +XXX,XX @@ typedef struct TCGProfile { int64_t tb_count; int64_t op_count; /* total insn count */ int op_count_max; /* max insn per TB */ - int64_t temp_count; int temp_count_max; + int64_t temp_count; int64_t del_op_count; int64_t code_in_len; int64_t code_out_len; -- 2.17.2 From: "Emilio G. Cota" <cota@braap.org> When we implemented per-vCPU TCG contexts, we forgot to also distribute the tcg_time counter, which has remained as a global accessed without any serialization, leading to potentially missed counts. Fix it by distributing the field over the TCG contexts, embedding it into TCGProfile with a field called "cpu_exec_time", which is more descriptive than "tcg_time". Add a function to query this value directly, and for completeness, fill in the field in tcg_profile_snapshot, even though its callers do not use it. Signed-off-by: Emilio G. Cota <cota@braap.org> Message-Id: <20181010144853.13005-5-cota@braap.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- include/qemu/timer.h | 1 - tcg/tcg.h | 2 ++ cpus.c | 3 ++- monitor.c | 13 ++++++++++--- tcg/tcg.c | 23 +++++++++++++++++++++++ 5 files changed, 37 insertions(+), 5 deletions(-) diff --git a/include/qemu/timer.h b/include/qemu/timer.h index XXXXXXX..XXXXXXX 100644 --- a/include/qemu/timer.h +++ b/include/qemu/timer.h @@ -XXX,XX +XXX,XX @@ static inline int64_t profile_getclock(void) return get_clock(); } -extern int64_t tcg_time; extern int64_t dev_time; #endif diff --git a/tcg/tcg.h b/tcg/tcg.h index XXXXXXX..XXXXXXX 100644 --- a/tcg/tcg.h +++ b/tcg/tcg.h @@ -XXX,XX +XXX,XX @@ typedef struct TCGOp { QEMU_BUILD_BUG_ON(NB_OPS > (1 << 8)); typedef struct TCGProfile { + int64_t cpu_exec_time; int64_t tb_count1; int64_t tb_count; int64_t op_count; /* total insn count */ @@ -XXX,XX +XXX,XX @@ int tcg_check_temp_count(void); #define tcg_check_temp_count() 0 #endif +int64_t tcg_cpu_exec_time(void); void tcg_dump_info(FILE *f, fprintf_function cpu_fprintf); void tcg_dump_op_count(FILE *f, fprintf_function cpu_fprintf); diff --git a/cpus.c b/cpus.c index XXXXXXX..XXXXXXX 100644 --- a/cpus.c +++ b/cpus.c @@ -XXX,XX +XXX,XX @@ static int tcg_cpu_exec(CPUState *cpu) ret = cpu_exec(cpu); cpu_exec_end(cpu); #ifdef CONFIG_PROFILER - tcg_time += profile_getclock() - ti; + atomic_set(&tcg_ctx->prof.cpu_exec_time, + tcg_ctx->prof.cpu_exec_time + profile_getclock() - ti); #endif return ret; } diff --git a/monitor.c b/monitor.c index XXXXXXX..XXXXXXX 100644 --- a/monitor.c +++ b/monitor.c @@ -XXX,XX +XXX,XX @@ #include "sysemu/cpus.h" #include "sysemu/iothread.h" #include "qemu/cutils.h" +#include "tcg/tcg.h" #if defined(TARGET_S390X) #include "hw/s390x/storage-keys.h" @@ -XXX,XX +XXX,XX @@ static void hmp_info_numa(Monitor *mon, const QDict *qdict) #ifdef CONFIG_PROFILER -int64_t tcg_time; int64_t dev_time; static void hmp_info_profile(Monitor *mon, const QDict *qdict) { + static int64_t last_cpu_exec_time; + int64_t cpu_exec_time; + int64_t delta; + + cpu_exec_time = tcg_cpu_exec_time(); + delta = cpu_exec_time - last_cpu_exec_time; + monitor_printf(mon, "async time %" PRId64 " (%0.3f)\n", dev_time, dev_time / (double)NANOSECONDS_PER_SECOND); monitor_printf(mon, "qemu time %" PRId64 " (%0.3f)\n", - tcg_time, tcg_time / (double)NANOSECONDS_PER_SECOND); - tcg_time = 0; + delta, delta / (double)NANOSECONDS_PER_SECOND); + last_cpu_exec_time = cpu_exec_time; dev_time = 0; } #else 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 @@ /* Define to jump the ELF file used to communicate with GDB. */ #undef DEBUG_JIT +#include "qemu/error-report.h" #include "qemu/cutils.h" #include "qemu/host-utils.h" #include "qemu/timer.h" @@ -XXX,XX +XXX,XX @@ void tcg_profile_snapshot(TCGProfile *prof, bool counters, bool table) const TCGProfile *orig = &s->prof; if (counters) { + PROF_ADD(prof, orig, cpu_exec_time); PROF_ADD(prof, orig, tb_count1); PROF_ADD(prof, orig, tb_count); PROF_ADD(prof, orig, op_count); @@ -XXX,XX +XXX,XX @@ void tcg_dump_op_count(FILE *f, fprintf_function cpu_fprintf) prof.table_op_count[i]); } } + +int64_t tcg_cpu_exec_time(void) +{ + unsigned int n_ctxs = atomic_read(&n_tcg_ctxs); + unsigned int i; + int64_t ret = 0; + + for (i = 0; i < n_ctxs; i++) { + const TCGContext *s = atomic_read(&tcg_ctxs[i]); + const TCGProfile *prof = &s->prof; + + ret += atomic_read(&prof->cpu_exec_time); + } + return ret; +} #else void tcg_dump_op_count(FILE *f, fprintf_function cpu_fprintf) { cpu_fprintf(f, "[TCG profiler not compiled]\n"); } + +int64_t tcg_cpu_exec_time(void) +{ + error_report("%s: TCG profiler not compiled", __func__); + exit(EXIT_FAILURE); +} #endif -- 2.17.2 From: "Emilio G. Cota" <cota@braap.org> As far as I can tell tlb_flush does not need to be called this early. tlb_flush is eventually called after the CPU has been realized. This change paves the way to the introduction of tlb_init, which will be called from cpu_exec_realizefn. Reviewed-by: Alex Bennée <alex.bennee@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Signed-off-by: Emilio G. Cota <cota@braap.org> Message-Id: <20181009174557.16125-2-cota@braap.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- target/alpha/cpu.c | 1 - 1 file changed, 1 deletion(-) 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 void alpha_cpu_initfn(Object *obj) CPUAlphaState *env = &cpu->env; cs->env_ptr = env; - tlb_flush(cs); env->lock_addr = -1; #if defined(CONFIG_USER_ONLY) -- 2.17.2 From: "Emilio G. Cota" <cota@braap.org> As far as I can tell tlb_flush does not need to be called this early. tlb_flush is eventually called after the CPU has been realized. This change paves the way to the introduction of tlb_init, which will be called from cpu_exec_realizefn. Cc: Guan Xuetao <gxt@mprc.pku.edu.cn> Reviewed-by: Alex Bennée <alex.bennee@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Signed-off-by: Emilio G. Cota <cota@braap.org> Message-Id: <20181009174557.16125-3-cota@braap.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- target/unicore32/cpu.c | 2 -- 1 file changed, 2 deletions(-) diff --git a/target/unicore32/cpu.c b/target/unicore32/cpu.c index XXXXXXX..XXXXXXX 100644 --- a/target/unicore32/cpu.c +++ b/target/unicore32/cpu.c @@ -XXX,XX +XXX,XX @@ static void uc32_cpu_initfn(Object *obj) env->uncached_asr = ASR_MODE_PRIV; env->regs[31] = 0x03000000; #endif - - tlb_flush(cs); } static const VMStateDescription vmstate_uc32_cpu = { -- 2.17.2 From: "Emilio G. Cota" <cota@braap.org> Paves the way for the addition of a per-TLB lock. Reviewed-by: Alex Bennée <alex.bennee@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Signed-off-by: Emilio G. Cota <cota@braap.org> Message-Id: <20181009174557.16125-4-cota@braap.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- include/exec/exec-all.h | 8 ++++++++ accel/tcg/cputlb.c | 4 ++++ exec.c | 1 + 3 files changed, 13 insertions(+) diff --git a/include/exec/exec-all.h b/include/exec/exec-all.h index XXXXXXX..XXXXXXX 100644 --- a/include/exec/exec-all.h +++ b/include/exec/exec-all.h @@ -XXX,XX +XXX,XX @@ void cpu_address_space_init(CPUState *cpu, int asidx, #if !defined(CONFIG_USER_ONLY) && defined(CONFIG_TCG) /* cputlb.c */ +/** + * tlb_init - initialize a CPU's TLB + * @cpu: CPU whose TLB should be initialized + */ +void tlb_init(CPUState *cpu); /** * tlb_flush_page: * @cpu: CPU whose TLB should be flushed @@ -XXX,XX +XXX,XX @@ void tlb_set_page(CPUState *cpu, target_ulong vaddr, void probe_write(CPUArchState *env, target_ulong addr, int size, int mmu_idx, uintptr_t retaddr); #else +static inline void tlb_init(CPUState *cpu) +{ +} static inline void tlb_flush_page(CPUState *cpu, target_ulong addr) { } diff --git a/accel/tcg/cputlb.c b/accel/tcg/cputlb.c index XXXXXXX..XXXXXXX 100644 --- a/accel/tcg/cputlb.c +++ b/accel/tcg/cputlb.c @@ -XXX,XX +XXX,XX @@ QEMU_BUILD_BUG_ON(sizeof(target_ulong) > sizeof(run_on_cpu_data)); QEMU_BUILD_BUG_ON(NB_MMU_MODES > 16); #define ALL_MMUIDX_BITS ((1 << NB_MMU_MODES) - 1) +void tlb_init(CPUState *cpu) +{ +} + /* flush_all_helper: run fn across all cpus * * If the wait flag is set then the src cpu's helper will be queued as diff --git a/exec.c b/exec.c index XXXXXXX..XXXXXXX 100644 --- a/exec.c +++ b/exec.c @@ -XXX,XX +XXX,XX @@ void cpu_exec_realizefn(CPUState *cpu, Error **errp) tcg_target_initialized = true; cc->tcg_initialize(); } + tlb_init(cpu); #ifndef CONFIG_USER_ONLY if (qdev_get_vmsd(DEVICE(cpu)) == NULL) { -- 2.17.2 From: "Emilio G. Cota" <cota@braap.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Reviewed-by: Alex Bennée <alex.bennee@linaro.org> Signed-off-by: Emilio G. Cota <cota@braap.org> Message-Id: <20181009174557.16125-5-cota@braap.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- accel/tcg/cputlb.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/accel/tcg/cputlb.c b/accel/tcg/cputlb.c index XXXXXXX..XXXXXXX 100644 --- a/accel/tcg/cputlb.c +++ b/accel/tcg/cputlb.c @@ -XXX,XX +XXX,XX @@ } \ } while (0) -#define assert_cpu_is_self(this_cpu) do { \ +#define assert_cpu_is_self(cpu) do { \ if (DEBUG_TLB_GATE) { \ - g_assert(!cpu->created || qemu_cpu_is_self(cpu)); \ + g_assert(!(cpu)->created || qemu_cpu_is_self(cpu)); \ } \ } while (0) -- 2.17.2 From: "Emilio G. Cota" <cota@braap.org> Currently we rely on atomic operations for cross-CPU invalidations. There are two cases that these atomics miss: cross-CPU invalidations can race with either (1) vCPU threads flushing their TLB, which happens via memset, or (2) vCPUs calling tlb_reset_dirty on their TLB, which updates .addr_write with a regular store. This results in undefined behaviour, since we're mixing regular and atomic ops on concurrent accesses. Fix it by using tlb_lock, a per-vCPU lock. All updaters of tlb_table and the corresponding victim cache now hold the lock. The readers that do not hold tlb_lock must use atomic reads when reading .addr_write, since this field can be updated by other threads; the conversion to atomic reads is done in the next patch. Note that an alternative fix would be to expand the use of atomic ops. However, in the case of TLB flushes this would have a huge performance impact, since (1) TLB flushes can happen very frequently and (2) we currently use a full memory barrier to flush each TLB entry, and a TLB has many entries. Instead, acquiring the lock is barely slower than a full memory barrier since it is uncontended, and with a single lock acquisition we can flush the entire TLB. Tested-by: Alex Bennée <alex.bennee@linaro.org> Reviewed-by: Alex Bennée <alex.bennee@linaro.org> Signed-off-by: Emilio G. Cota <cota@braap.org> Message-Id: <20181009174557.16125-6-cota@braap.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- include/exec/cpu-defs.h | 3 + accel/tcg/cputlb.c | 155 ++++++++++++++++++++++------------------ 2 files changed, 87 insertions(+), 71 deletions(-) diff --git a/include/exec/cpu-defs.h b/include/exec/cpu-defs.h index XXXXXXX..XXXXXXX 100644 --- a/include/exec/cpu-defs.h +++ b/include/exec/cpu-defs.h @@ -XXX,XX +XXX,XX @@ #endif #include "qemu/host-utils.h" +#include "qemu/thread.h" #include "qemu/queue.h" #ifdef CONFIG_TCG #include "tcg-target.h" @@ -XXX,XX +XXX,XX @@ typedef struct CPUIOTLBEntry { #define CPU_COMMON_TLB \ /* The meaning of the MMU modes is defined in the target code. */ \ + /* tlb_lock serializes updates to tlb_table and tlb_v_table */ \ + QemuSpin tlb_lock; \ CPUTLBEntry tlb_table[NB_MMU_MODES][CPU_TLB_SIZE]; \ CPUTLBEntry tlb_v_table[NB_MMU_MODES][CPU_VTLB_SIZE]; \ CPUIOTLBEntry iotlb[NB_MMU_MODES][CPU_TLB_SIZE]; \ diff --git a/accel/tcg/cputlb.c b/accel/tcg/cputlb.c index XXXXXXX..XXXXXXX 100644 --- a/accel/tcg/cputlb.c +++ b/accel/tcg/cputlb.c @@ -XXX,XX +XXX,XX @@ QEMU_BUILD_BUG_ON(NB_MMU_MODES > 16); void tlb_init(CPUState *cpu) { + CPUArchState *env = cpu->env_ptr; + + qemu_spin_init(&env->tlb_lock); } /* flush_all_helper: run fn across all cpus @@ -XXX,XX +XXX,XX @@ static void tlb_flush_nocheck(CPUState *cpu) atomic_set(&env->tlb_flush_count, env->tlb_flush_count + 1); tlb_debug("(count: %zu)\n", tlb_flush_count()); + /* + * tlb_table/tlb_v_table updates from any thread must hold tlb_lock. + * However, updates from the owner thread (as is the case here; see the + * above assert_cpu_is_self) do not need atomic_set because all reads + * that do not hold the lock are performed by the same owner thread. + */ + qemu_spin_lock(&env->tlb_lock); memset(env->tlb_table, -1, sizeof(env->tlb_table)); memset(env->tlb_v_table, -1, sizeof(env->tlb_v_table)); + qemu_spin_unlock(&env->tlb_lock); + cpu_tb_jmp_cache_clear(cpu); env->vtlb_index = 0; @@ -XXX,XX +XXX,XX @@ static void tlb_flush_by_mmuidx_async_work(CPUState *cpu, run_on_cpu_data data) tlb_debug("start: mmu_idx:0x%04lx\n", mmu_idx_bitmask); + qemu_spin_lock(&env->tlb_lock); for (mmu_idx = 0; mmu_idx < NB_MMU_MODES; mmu_idx++) { if (test_bit(mmu_idx, &mmu_idx_bitmask)) { @@ -XXX,XX +XXX,XX @@ static void tlb_flush_by_mmuidx_async_work(CPUState *cpu, run_on_cpu_data data) memset(env->tlb_v_table[mmu_idx], -1, sizeof(env->tlb_v_table[0])); } } + qemu_spin_unlock(&env->tlb_lock); cpu_tb_jmp_cache_clear(cpu); @@ -XXX,XX +XXX,XX @@ static inline bool tlb_hit_page_anyprot(CPUTLBEntry *tlb_entry, tlb_hit_page(tlb_entry->addr_code, page); } -static inline void tlb_flush_entry(CPUTLBEntry *tlb_entry, target_ulong page) +/* Called with tlb_lock held */ +static inline void tlb_flush_entry_locked(CPUTLBEntry *tlb_entry, + target_ulong page) { if (tlb_hit_page_anyprot(tlb_entry, page)) { memset(tlb_entry, -1, sizeof(*tlb_entry)); } } -static inline void tlb_flush_vtlb_page(CPUArchState *env, int mmu_idx, - target_ulong page) +/* Called with tlb_lock held */ +static inline void tlb_flush_vtlb_page_locked(CPUArchState *env, int mmu_idx, + target_ulong page) { int k; + + assert_cpu_is_self(ENV_GET_CPU(env)); for (k = 0; k < CPU_VTLB_SIZE; k++) { - tlb_flush_entry(&env->tlb_v_table[mmu_idx][k], page); + tlb_flush_entry_locked(&env->tlb_v_table[mmu_idx][k], page); } } @@ -XXX,XX +XXX,XX @@ static void tlb_flush_page_async_work(CPUState *cpu, run_on_cpu_data data) addr &= TARGET_PAGE_MASK; i = (addr >> TARGET_PAGE_BITS) & (CPU_TLB_SIZE - 1); + qemu_spin_lock(&env->tlb_lock); for (mmu_idx = 0; mmu_idx < NB_MMU_MODES; mmu_idx++) { - tlb_flush_entry(&env->tlb_table[mmu_idx][i], addr); - tlb_flush_vtlb_page(env, mmu_idx, addr); + tlb_flush_entry_locked(&env->tlb_table[mmu_idx][i], addr); + tlb_flush_vtlb_page_locked(env, mmu_idx, addr); } + qemu_spin_unlock(&env->tlb_lock); tb_flush_jmp_cache(cpu, addr); } @@ -XXX,XX +XXX,XX @@ static void tlb_flush_page_by_mmuidx_async_work(CPUState *cpu, tlb_debug("page:%d addr:"TARGET_FMT_lx" mmu_idx:0x%lx\n", page, addr, mmu_idx_bitmap); + qemu_spin_lock(&env->tlb_lock); for (mmu_idx = 0; mmu_idx < NB_MMU_MODES; mmu_idx++) { if (test_bit(mmu_idx, &mmu_idx_bitmap)) { - tlb_flush_entry(&env->tlb_table[mmu_idx][page], addr); - tlb_flush_vtlb_page(env, mmu_idx, addr); + tlb_flush_entry_locked(&env->tlb_table[mmu_idx][page], addr); + tlb_flush_vtlb_page_locked(env, mmu_idx, addr); } } + qemu_spin_unlock(&env->tlb_lock); tb_flush_jmp_cache(cpu, addr); } @@ -XXX,XX +XXX,XX @@ void tlb_unprotect_code(ram_addr_t ram_addr) * most usual is detecting writes to code regions which may invalidate * generated code. * - * Because we want other vCPUs to respond to changes straight away we - * update the te->addr_write field atomically. If the TLB entry has - * been changed by the vCPU in the mean time we skip the update. + * Other vCPUs might be reading their TLBs during guest execution, so we update + * te->addr_write with atomic_set. We don't need to worry about this for + * oversized guests as MTTCG is disabled for them. * - * As this function uses atomic accesses we also need to ensure - * updates to tlb_entries follow the same access rules. We don't need - * to worry about this for oversized guests as MTTCG is disabled for - * them. + * Called with tlb_lock held. */ - -static void tlb_reset_dirty_range(CPUTLBEntry *tlb_entry, uintptr_t start, - uintptr_t length) +static void tlb_reset_dirty_range_locked(CPUTLBEntry *tlb_entry, + uintptr_t start, uintptr_t length) { -#if TCG_OVERSIZED_GUEST uintptr_t addr = tlb_entry->addr_write; if ((addr & (TLB_INVALID_MASK | TLB_MMIO | TLB_NOTDIRTY)) == 0) { addr &= TARGET_PAGE_MASK; addr += tlb_entry->addend; if ((addr - start) < length) { +#if TCG_OVERSIZED_GUEST tlb_entry->addr_write |= TLB_NOTDIRTY; - } - } #else - /* paired with atomic_mb_set in tlb_set_page_with_attrs */ - uintptr_t orig_addr = atomic_mb_read(&tlb_entry->addr_write); - uintptr_t addr = orig_addr; - - if ((addr & (TLB_INVALID_MASK | TLB_MMIO | TLB_NOTDIRTY)) == 0) { - addr &= TARGET_PAGE_MASK; - addr += atomic_read(&tlb_entry->addend); - if ((addr - start) < length) { - uintptr_t notdirty_addr = orig_addr | TLB_NOTDIRTY; - atomic_cmpxchg(&tlb_entry->addr_write, orig_addr, notdirty_addr); + atomic_set(&tlb_entry->addr_write, + tlb_entry->addr_write | TLB_NOTDIRTY); +#endif } } -#endif } -/* For atomic correctness when running MTTCG we need to use the right - * primitives when copying entries */ -static inline void copy_tlb_helper(CPUTLBEntry *d, CPUTLBEntry *s, - bool atomic_set) +/* + * Called with tlb_lock held. + * Called only from the vCPU context, i.e. the TLB's owner thread. + */ +static inline void copy_tlb_helper_locked(CPUTLBEntry *d, const CPUTLBEntry *s) { -#if TCG_OVERSIZED_GUEST *d = *s; -#else - if (atomic_set) { - d->addr_read = s->addr_read; - d->addr_code = s->addr_code; - atomic_set(&d->addend, atomic_read(&s->addend)); - /* Pairs with flag setting in tlb_reset_dirty_range */ - atomic_mb_set(&d->addr_write, atomic_read(&s->addr_write)); - } else { - d->addr_read = s->addr_read; - d->addr_write = atomic_read(&s->addr_write); - d->addr_code = s->addr_code; - d->addend = atomic_read(&s->addend); - } -#endif } /* This is a cross vCPU call (i.e. another vCPU resetting the flags of - * the target vCPU). As such care needs to be taken that we don't - * dangerously race with another vCPU update. The only thing actually - * updated is the target TLB entry ->addr_write flags. + * the target vCPU). + * We must take tlb_lock to avoid racing with another vCPU update. The only + * thing actually updated is the target TLB entry ->addr_write flags. */ void tlb_reset_dirty(CPUState *cpu, ram_addr_t start1, ram_addr_t length) { @@ -XXX,XX +XXX,XX @@ void tlb_reset_dirty(CPUState *cpu, ram_addr_t start1, ram_addr_t length) int mmu_idx; env = cpu->env_ptr; + qemu_spin_lock(&env->tlb_lock); for (mmu_idx = 0; mmu_idx < NB_MMU_MODES; mmu_idx++) { unsigned int i; for (i = 0; i < CPU_TLB_SIZE; i++) { - tlb_reset_dirty_range(&env->tlb_table[mmu_idx][i], - start1, length); + tlb_reset_dirty_range_locked(&env->tlb_table[mmu_idx][i], start1, + length); } for (i = 0; i < CPU_VTLB_SIZE; i++) { - tlb_reset_dirty_range(&env->tlb_v_table[mmu_idx][i], - start1, length); + tlb_reset_dirty_range_locked(&env->tlb_v_table[mmu_idx][i], start1, + length); } } + qemu_spin_unlock(&env->tlb_lock); } -static inline void tlb_set_dirty1(CPUTLBEntry *tlb_entry, target_ulong vaddr) +/* Called with tlb_lock held */ +static inline void tlb_set_dirty1_locked(CPUTLBEntry *tlb_entry, + target_ulong vaddr) { if (tlb_entry->addr_write == (vaddr | TLB_NOTDIRTY)) { tlb_entry->addr_write = vaddr; @@ -XXX,XX +XXX,XX @@ void tlb_set_dirty(CPUState *cpu, target_ulong vaddr) vaddr &= TARGET_PAGE_MASK; i = (vaddr >> TARGET_PAGE_BITS) & (CPU_TLB_SIZE - 1); + qemu_spin_lock(&env->tlb_lock); for (mmu_idx = 0; mmu_idx < NB_MMU_MODES; mmu_idx++) { - tlb_set_dirty1(&env->tlb_table[mmu_idx][i], vaddr); + tlb_set_dirty1_locked(&env->tlb_table[mmu_idx][i], vaddr); } for (mmu_idx = 0; mmu_idx < NB_MMU_MODES; mmu_idx++) { int k; for (k = 0; k < CPU_VTLB_SIZE; k++) { - tlb_set_dirty1(&env->tlb_v_table[mmu_idx][k], vaddr); + tlb_set_dirty1_locked(&env->tlb_v_table[mmu_idx][k], vaddr); } } + qemu_spin_unlock(&env->tlb_lock); } /* Our TLB does not support large pages, so remember the area covered by @@ -XXX,XX +XXX,XX @@ void tlb_set_page_with_attrs(CPUState *cpu, target_ulong vaddr, addend = (uintptr_t)memory_region_get_ram_ptr(section->mr) + xlat; } - /* Make sure there's no cached translation for the new page. */ - tlb_flush_vtlb_page(env, mmu_idx, vaddr_page); - code_address = address; iotlb = memory_region_section_get_iotlb(cpu, section, vaddr_page, paddr_page, xlat, prot, &address); @@ -XXX,XX +XXX,XX @@ void tlb_set_page_with_attrs(CPUState *cpu, target_ulong vaddr, index = (vaddr_page >> TARGET_PAGE_BITS) & (CPU_TLB_SIZE - 1); te = &env->tlb_table[mmu_idx][index]; + /* + * Hold the TLB lock for the rest of the function. We could acquire/release + * the lock several times in the function, but it is faster to amortize the + * acquisition cost by acquiring it just once. Note that this leads to + * a longer critical section, but this is not a concern since the TLB lock + * is unlikely to be contended. + */ + qemu_spin_lock(&env->tlb_lock); + + /* Make sure there's no cached translation for the new page. */ + tlb_flush_vtlb_page_locked(env, mmu_idx, vaddr_page); + /* * Only evict the old entry to the victim tlb if it's for a * different page; otherwise just overwrite the stale data. @@ -XXX,XX +XXX,XX @@ void tlb_set_page_with_attrs(CPUState *cpu, target_ulong vaddr, CPUTLBEntry *tv = &env->tlb_v_table[mmu_idx][vidx]; /* Evict the old entry into the victim tlb. */ - copy_tlb_helper(tv, te, true); + copy_tlb_helper_locked(tv, te); env->iotlb_v[mmu_idx][vidx] = env->iotlb[mmu_idx][index]; } @@ -XXX,XX +XXX,XX @@ void tlb_set_page_with_attrs(CPUState *cpu, target_ulong vaddr, } } - /* Pairs with flag setting in tlb_reset_dirty_range */ - copy_tlb_helper(te, &tn, true); - /* atomic_mb_set(&te->addr_write, write_address); */ + copy_tlb_helper_locked(te, &tn); + qemu_spin_unlock(&env->tlb_lock); } /* Add a new TLB entry, but without specifying the memory @@ -XXX,XX +XXX,XX @@ static bool victim_tlb_hit(CPUArchState *env, size_t mmu_idx, size_t index, size_t elt_ofs, target_ulong page) { size_t vidx; + + assert_cpu_is_self(ENV_GET_CPU(env)); for (vidx = 0; vidx < CPU_VTLB_SIZE; ++vidx) { CPUTLBEntry *vtlb = &env->tlb_v_table[mmu_idx][vidx]; target_ulong cmp = *(target_ulong *)((uintptr_t)vtlb + elt_ofs); @@ -XXX,XX +XXX,XX @@ static bool victim_tlb_hit(CPUArchState *env, size_t mmu_idx, size_t index, /* Found entry in victim tlb, swap tlb and iotlb. */ CPUTLBEntry tmptlb, *tlb = &env->tlb_table[mmu_idx][index]; - copy_tlb_helper(&tmptlb, tlb, false); - copy_tlb_helper(tlb, vtlb, true); - copy_tlb_helper(vtlb, &tmptlb, true); + qemu_spin_lock(&env->tlb_lock); + copy_tlb_helper_locked(&tmptlb, tlb); + copy_tlb_helper_locked(tlb, vtlb); + copy_tlb_helper_locked(vtlb, &tmptlb); + qemu_spin_unlock(&env->tlb_lock); CPUIOTLBEntry tmpio, *io = &env->iotlb[mmu_idx][index]; CPUIOTLBEntry *vio = &env->iotlb_v[mmu_idx][vidx]; -- 2.17.2 Isolate the computation of an index from an address into a helper before we change that function. Reviewed-by: Alex Bennée <alex.bennee@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> [ cota: convert tlb_vaddr_to_host; use atomic_read on addr_write ] Signed-off-by: Emilio G. Cota <cota@braap.org> Message-Id: <20181009175129.17888-2-cota@braap.org> --- accel/tcg/softmmu_template.h | 64 +++++++++++++++++--------------- include/exec/cpu_ldst.h | 19 ++++++++-- include/exec/cpu_ldst_template.h | 25 +++++++------ accel/tcg/cputlb.c | 60 ++++++++++++++---------------- 4 files changed, 90 insertions(+), 78 deletions(-) diff --git a/accel/tcg/softmmu_template.h b/accel/tcg/softmmu_template.h index XXXXXXX..XXXXXXX 100644 --- a/accel/tcg/softmmu_template.h +++ b/accel/tcg/softmmu_template.h @@ -XXX,XX +XXX,XX @@ static inline DATA_TYPE glue(io_read, SUFFIX)(CPUArchState *env, WORD_TYPE helper_le_ld_name(CPUArchState *env, target_ulong addr, TCGMemOpIdx oi, uintptr_t retaddr) { - unsigned mmu_idx = get_mmuidx(oi); - int index = (addr >> TARGET_PAGE_BITS) & (CPU_TLB_SIZE - 1); - target_ulong tlb_addr = env->tlb_table[mmu_idx][index].ADDR_READ; + uintptr_t mmu_idx = get_mmuidx(oi); + uintptr_t index = tlb_index(env, mmu_idx, addr); + CPUTLBEntry *entry = tlb_entry(env, mmu_idx, addr); + target_ulong tlb_addr = entry->ADDR_READ; unsigned a_bits = get_alignment_bits(get_memop(oi)); uintptr_t haddr; DATA_TYPE res; @@ -XXX,XX +XXX,XX @@ WORD_TYPE helper_le_ld_name(CPUArchState *env, target_ulong addr, tlb_fill(ENV_GET_CPU(env), addr, DATA_SIZE, READ_ACCESS_TYPE, mmu_idx, retaddr); } - tlb_addr = env->tlb_table[mmu_idx][index].ADDR_READ; + tlb_addr = entry->ADDR_READ; } /* Handle an IO access. */ @@ -XXX,XX +XXX,XX @@ WORD_TYPE helper_le_ld_name(CPUArchState *env, target_ulong addr, return res; } - haddr = addr + env->tlb_table[mmu_idx][index].addend; + haddr = addr + entry->addend; #if DATA_SIZE == 1 res = glue(glue(ld, LSUFFIX), _p)((uint8_t *)haddr); #else @@ -XXX,XX +XXX,XX @@ WORD_TYPE helper_le_ld_name(CPUArchState *env, target_ulong addr, WORD_TYPE helper_be_ld_name(CPUArchState *env, target_ulong addr, TCGMemOpIdx oi, uintptr_t retaddr) { - unsigned mmu_idx = get_mmuidx(oi); - int index = (addr >> TARGET_PAGE_BITS) & (CPU_TLB_SIZE - 1); - target_ulong tlb_addr = env->tlb_table[mmu_idx][index].ADDR_READ; + uintptr_t mmu_idx = get_mmuidx(oi); + uintptr_t index = tlb_index(env, mmu_idx, addr); + CPUTLBEntry *entry = tlb_entry(env, mmu_idx, addr); + target_ulong tlb_addr = entry->ADDR_READ; unsigned a_bits = get_alignment_bits(get_memop(oi)); uintptr_t haddr; DATA_TYPE res; @@ -XXX,XX +XXX,XX @@ WORD_TYPE helper_be_ld_name(CPUArchState *env, target_ulong addr, tlb_fill(ENV_GET_CPU(env), addr, DATA_SIZE, READ_ACCESS_TYPE, mmu_idx, retaddr); } - tlb_addr = env->tlb_table[mmu_idx][index].ADDR_READ; + tlb_addr = entry->ADDR_READ; } /* Handle an IO access. */ @@ -XXX,XX +XXX,XX @@ WORD_TYPE helper_be_ld_name(CPUArchState *env, target_ulong addr, return res; } - haddr = addr + env->tlb_table[mmu_idx][index].addend; + haddr = addr + entry->addend; res = glue(glue(ld, LSUFFIX), _be_p)((uint8_t *)haddr); return res; } @@ -XXX,XX +XXX,XX @@ static inline void glue(io_write, SUFFIX)(CPUArchState *env, void helper_le_st_name(CPUArchState *env, target_ulong addr, DATA_TYPE val, TCGMemOpIdx oi, uintptr_t retaddr) { - unsigned mmu_idx = get_mmuidx(oi); - int index = (addr >> TARGET_PAGE_BITS) & (CPU_TLB_SIZE - 1); - target_ulong tlb_addr = env->tlb_table[mmu_idx][index].addr_write; + uintptr_t mmu_idx = get_mmuidx(oi); + uintptr_t index = tlb_index(env, mmu_idx, addr); + CPUTLBEntry *entry = tlb_entry(env, mmu_idx, addr); + target_ulong tlb_addr = entry->addr_write; unsigned a_bits = get_alignment_bits(get_memop(oi)); uintptr_t haddr; @@ -XXX,XX +XXX,XX @@ void helper_le_st_name(CPUArchState *env, target_ulong addr, DATA_TYPE val, tlb_fill(ENV_GET_CPU(env), addr, DATA_SIZE, MMU_DATA_STORE, mmu_idx, retaddr); } - tlb_addr = env->tlb_table[mmu_idx][index].addr_write & ~TLB_INVALID_MASK; + tlb_addr = entry->addr_write & ~TLB_INVALID_MASK; } /* Handle an IO access. */ @@ -XXX,XX +XXX,XX @@ void helper_le_st_name(CPUArchState *env, target_ulong addr, DATA_TYPE val, if (DATA_SIZE > 1 && unlikely((addr & ~TARGET_PAGE_MASK) + DATA_SIZE - 1 >= TARGET_PAGE_SIZE)) { - int i, index2; - target_ulong page2, tlb_addr2; + int i; + target_ulong page2; + CPUTLBEntry *entry2; do_unaligned_access: /* Ensure the second page is in the TLB. Note that the first page is already guaranteed to be filled, and that the second page cannot evict the first. */ page2 = (addr + DATA_SIZE) & TARGET_PAGE_MASK; - index2 = (page2 >> TARGET_PAGE_BITS) & (CPU_TLB_SIZE - 1); - tlb_addr2 = env->tlb_table[mmu_idx][index2].addr_write; - if (!tlb_hit_page(tlb_addr2, page2) + entry2 = tlb_entry(env, mmu_idx, page2); + if (!tlb_hit_page(entry2->addr_write, page2) && !VICTIM_TLB_HIT(addr_write, page2)) { tlb_fill(ENV_GET_CPU(env), page2, DATA_SIZE, MMU_DATA_STORE, mmu_idx, retaddr); @@ -XXX,XX +XXX,XX @@ void helper_le_st_name(CPUArchState *env, target_ulong addr, DATA_TYPE val, return; } - haddr = addr + env->tlb_table[mmu_idx][index].addend; + haddr = addr + entry->addend; #if DATA_SIZE == 1 glue(glue(st, SUFFIX), _p)((uint8_t *)haddr, val); #else @@ -XXX,XX +XXX,XX @@ void helper_le_st_name(CPUArchState *env, target_ulong addr, DATA_TYPE val, void helper_be_st_name(CPUArchState *env, target_ulong addr, DATA_TYPE val, TCGMemOpIdx oi, uintptr_t retaddr) { - unsigned mmu_idx = get_mmuidx(oi); - int index = (addr >> TARGET_PAGE_BITS) & (CPU_TLB_SIZE - 1); - target_ulong tlb_addr = env->tlb_table[mmu_idx][index].addr_write; + uintptr_t mmu_idx = get_mmuidx(oi); + uintptr_t index = tlb_index(env, mmu_idx, addr); + CPUTLBEntry *entry = tlb_entry(env, mmu_idx, addr); + target_ulong tlb_addr = entry->addr_write; unsigned a_bits = get_alignment_bits(get_memop(oi)); uintptr_t haddr; @@ -XXX,XX +XXX,XX @@ void helper_be_st_name(CPUArchState *env, target_ulong addr, DATA_TYPE val, tlb_fill(ENV_GET_CPU(env), addr, DATA_SIZE, MMU_DATA_STORE, mmu_idx, retaddr); } - tlb_addr = env->tlb_table[mmu_idx][index].addr_write & ~TLB_INVALID_MASK; + tlb_addr = entry->addr_write & ~TLB_INVALID_MASK; } /* Handle an IO access. */ @@ -XXX,XX +XXX,XX @@ void helper_be_st_name(CPUArchState *env, target_ulong addr, DATA_TYPE val, if (DATA_SIZE > 1 && unlikely((addr & ~TARGET_PAGE_MASK) + DATA_SIZE - 1 >= TARGET_PAGE_SIZE)) { - int i, index2; - target_ulong page2, tlb_addr2; + int i; + target_ulong page2; + CPUTLBEntry *entry2; do_unaligned_access: /* Ensure the second page is in the TLB. Note that the first page is already guaranteed to be filled, and that the second page cannot evict the first. */ page2 = (addr + DATA_SIZE) & TARGET_PAGE_MASK; - index2 = (page2 >> TARGET_PAGE_BITS) & (CPU_TLB_SIZE - 1); - tlb_addr2 = env->tlb_table[mmu_idx][index2].addr_write; - if (!tlb_hit_page(tlb_addr2, page2) + entry2 = tlb_entry(env, mmu_idx, page2); + if (!tlb_hit_page(entry2->addr_write, page2) && !VICTIM_TLB_HIT(addr_write, page2)) { tlb_fill(ENV_GET_CPU(env), page2, DATA_SIZE, MMU_DATA_STORE, mmu_idx, retaddr); @@ -XXX,XX +XXX,XX @@ void helper_be_st_name(CPUArchState *env, target_ulong addr, DATA_TYPE val, return; } - haddr = addr + env->tlb_table[mmu_idx][index].addend; + haddr = addr + entry->addend; glue(glue(st, SUFFIX), _be_p)((uint8_t *)haddr, val); } #endif /* DATA_SIZE > 1 */ 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 @@ extern __thread uintptr_t helper_retaddr; /* The memory helpers for tcg-generated code need tcg_target_long etc. */ #include "tcg.h" +/* Find the TLB index corresponding to the mmu_idx + address pair. */ +static inline uintptr_t tlb_index(CPUArchState *env, uintptr_t mmu_idx, + target_ulong addr) +{ + return (addr >> TARGET_PAGE_BITS) & (CPU_TLB_SIZE - 1); +} + +/* Find the TLB entry corresponding to the mmu_idx + address pair. */ +static inline CPUTLBEntry *tlb_entry(CPUArchState *env, uintptr_t mmu_idx, + target_ulong addr) +{ + return &env->tlb_table[mmu_idx][tlb_index(env, mmu_idx, addr)]; +} + #ifdef MMU_MODE0_SUFFIX #define CPU_MMU_INDEX 0 #define MEMSUFFIX MMU_MODE0_SUFFIX @@ -XXX,XX +XXX,XX @@ static inline void *tlb_vaddr_to_host(CPUArchState *env, abi_ptr addr, #if defined(CONFIG_USER_ONLY) return g2h(addr); #else - int index = (addr >> TARGET_PAGE_BITS) & (CPU_TLB_SIZE - 1); - CPUTLBEntry *tlbentry = &env->tlb_table[mmu_idx][index]; + CPUTLBEntry *tlbentry = tlb_entry(env, mmu_idx, addr); abi_ptr tlb_addr; uintptr_t haddr; @@ -XXX,XX +XXX,XX @@ static inline void *tlb_vaddr_to_host(CPUArchState *env, abi_ptr addr, return NULL; } - haddr = addr + env->tlb_table[mmu_idx][index].addend; + haddr = addr + tlbentry->addend; return (void *)haddr; #endif /* defined(CONFIG_USER_ONLY) */ } diff --git a/include/exec/cpu_ldst_template.h b/include/exec/cpu_ldst_template.h index XXXXXXX..XXXXXXX 100644 --- a/include/exec/cpu_ldst_template.h +++ b/include/exec/cpu_ldst_template.h @@ -XXX,XX +XXX,XX @@ glue(glue(glue(cpu_ld, USUFFIX), MEMSUFFIX), _ra)(CPUArchState *env, target_ulong ptr, uintptr_t retaddr) { - int page_index; + CPUTLBEntry *entry; RES_TYPE res; target_ulong addr; int mmu_idx; @@ -XXX,XX +XXX,XX @@ glue(glue(glue(cpu_ld, USUFFIX), MEMSUFFIX), _ra)(CPUArchState *env, #endif addr = ptr; - page_index = (addr >> TARGET_PAGE_BITS) & (CPU_TLB_SIZE - 1); mmu_idx = CPU_MMU_INDEX; - if (unlikely(env->tlb_table[mmu_idx][page_index].ADDR_READ != + entry = tlb_entry(env, mmu_idx, addr); + if (unlikely(entry->ADDR_READ != (addr & (TARGET_PAGE_MASK | (DATA_SIZE - 1))))) { oi = make_memop_idx(SHIFT, mmu_idx); res = glue(glue(helper_ret_ld, URETSUFFIX), MMUSUFFIX)(env, addr, oi, retaddr); } else { - uintptr_t hostaddr = addr + env->tlb_table[mmu_idx][page_index].addend; + uintptr_t hostaddr = addr + entry->addend; res = glue(glue(ld, USUFFIX), _p)((uint8_t *)hostaddr); } return res; @@ -XXX,XX +XXX,XX @@ glue(glue(glue(cpu_lds, SUFFIX), MEMSUFFIX), _ra)(CPUArchState *env, target_ulong ptr, uintptr_t retaddr) { - int res, page_index; + CPUTLBEntry *entry; + int res; target_ulong addr; int mmu_idx; TCGMemOpIdx oi; @@ -XXX,XX +XXX,XX @@ glue(glue(glue(cpu_lds, SUFFIX), MEMSUFFIX), _ra)(CPUArchState *env, #endif addr = ptr; - page_index = (addr >> TARGET_PAGE_BITS) & (CPU_TLB_SIZE - 1); mmu_idx = CPU_MMU_INDEX; - if (unlikely(env->tlb_table[mmu_idx][page_index].ADDR_READ != + entry = tlb_entry(env, mmu_idx, addr); + if (unlikely(entry->ADDR_READ != (addr & (TARGET_PAGE_MASK | (DATA_SIZE - 1))))) { oi = make_memop_idx(SHIFT, mmu_idx); res = (DATA_STYPE)glue(glue(helper_ret_ld, SRETSUFFIX), MMUSUFFIX)(env, addr, oi, retaddr); } else { - uintptr_t hostaddr = addr + env->tlb_table[mmu_idx][page_index].addend; + uintptr_t hostaddr = addr + entry->addend; res = glue(glue(lds, SUFFIX), _p)((uint8_t *)hostaddr); } return res; @@ -XXX,XX +XXX,XX @@ glue(glue(glue(cpu_st, SUFFIX), MEMSUFFIX), _ra)(CPUArchState *env, target_ulong ptr, RES_TYPE v, uintptr_t retaddr) { - int page_index; + CPUTLBEntry *entry; target_ulong addr; int mmu_idx; TCGMemOpIdx oi; @@ -XXX,XX +XXX,XX @@ glue(glue(glue(cpu_st, SUFFIX), MEMSUFFIX), _ra)(CPUArchState *env, #endif addr = ptr; - page_index = (addr >> TARGET_PAGE_BITS) & (CPU_TLB_SIZE - 1); mmu_idx = CPU_MMU_INDEX; - if (unlikely(env->tlb_table[mmu_idx][page_index].addr_write != + entry = tlb_entry(env, mmu_idx, addr); + if (unlikely(entry->addr_write != (addr & (TARGET_PAGE_MASK | (DATA_SIZE - 1))))) { oi = make_memop_idx(SHIFT, mmu_idx); glue(glue(helper_ret_st, SUFFIX), MMUSUFFIX)(env, addr, v, oi, retaddr); } else { - uintptr_t hostaddr = addr + env->tlb_table[mmu_idx][page_index].addend; + uintptr_t hostaddr = addr + entry->addend; glue(glue(st, SUFFIX), _p)((uint8_t *)hostaddr, v); } } diff --git a/accel/tcg/cputlb.c b/accel/tcg/cputlb.c index XXXXXXX..XXXXXXX 100644 --- a/accel/tcg/cputlb.c +++ b/accel/tcg/cputlb.c @@ -XXX,XX +XXX,XX @@ static void tlb_flush_page_async_work(CPUState *cpu, run_on_cpu_data data) { CPUArchState *env = cpu->env_ptr; target_ulong addr = (target_ulong) data.target_ptr; - int i; int mmu_idx; assert_cpu_is_self(cpu); @@ -XXX,XX +XXX,XX @@ static void tlb_flush_page_async_work(CPUState *cpu, run_on_cpu_data data) } addr &= TARGET_PAGE_MASK; - i = (addr >> TARGET_PAGE_BITS) & (CPU_TLB_SIZE - 1); qemu_spin_lock(&env->tlb_lock); for (mmu_idx = 0; mmu_idx < NB_MMU_MODES; mmu_idx++) { - tlb_flush_entry_locked(&env->tlb_table[mmu_idx][i], addr); + tlb_flush_entry_locked(tlb_entry(env, mmu_idx, addr), addr); tlb_flush_vtlb_page_locked(env, mmu_idx, addr); } qemu_spin_unlock(&env->tlb_lock); @@ -XXX,XX +XXX,XX @@ static void tlb_flush_page_by_mmuidx_async_work(CPUState *cpu, target_ulong addr_and_mmuidx = (target_ulong) data.target_ptr; target_ulong addr = addr_and_mmuidx & TARGET_PAGE_MASK; unsigned long mmu_idx_bitmap = addr_and_mmuidx & ALL_MMUIDX_BITS; - int page = (addr >> TARGET_PAGE_BITS) & (CPU_TLB_SIZE - 1); int mmu_idx; assert_cpu_is_self(cpu); - tlb_debug("page:%d addr:"TARGET_FMT_lx" mmu_idx:0x%lx\n", - page, addr, mmu_idx_bitmap); + tlb_debug("flush page addr:"TARGET_FMT_lx" mmu_idx:0x%lx\n", + addr, mmu_idx_bitmap); qemu_spin_lock(&env->tlb_lock); for (mmu_idx = 0; mmu_idx < NB_MMU_MODES; mmu_idx++) { if (test_bit(mmu_idx, &mmu_idx_bitmap)) { - tlb_flush_entry_locked(&env->tlb_table[mmu_idx][page], addr); + tlb_flush_entry_locked(tlb_entry(env, mmu_idx, addr), addr); tlb_flush_vtlb_page_locked(env, mmu_idx, addr); } } @@ -XXX,XX +XXX,XX @@ static inline void tlb_set_dirty1_locked(CPUTLBEntry *tlb_entry, void tlb_set_dirty(CPUState *cpu, target_ulong vaddr) { CPUArchState *env = cpu->env_ptr; - int i; int mmu_idx; assert_cpu_is_self(cpu); vaddr &= TARGET_PAGE_MASK; - i = (vaddr >> TARGET_PAGE_BITS) & (CPU_TLB_SIZE - 1); qemu_spin_lock(&env->tlb_lock); for (mmu_idx = 0; mmu_idx < NB_MMU_MODES; mmu_idx++) { - tlb_set_dirty1_locked(&env->tlb_table[mmu_idx][i], vaddr); + tlb_set_dirty1_locked(tlb_entry(env, mmu_idx, vaddr), vaddr); } for (mmu_idx = 0; mmu_idx < NB_MMU_MODES; mmu_idx++) { @@ -XXX,XX +XXX,XX @@ void tlb_set_page_with_attrs(CPUState *cpu, target_ulong vaddr, iotlb = memory_region_section_get_iotlb(cpu, section, vaddr_page, paddr_page, xlat, prot, &address); - index = (vaddr_page >> TARGET_PAGE_BITS) & (CPU_TLB_SIZE - 1); - te = &env->tlb_table[mmu_idx][index]; + index = tlb_index(env, mmu_idx, vaddr_page); + te = tlb_entry(env, mmu_idx, vaddr_page); /* * Hold the TLB lock for the rest of the function. We could acquire/release @@ -XXX,XX +XXX,XX @@ static uint64_t io_readx(CPUArchState *env, CPUIOTLBEntry *iotlbentry, * repeat the MMU check here. This tlb_fill() call might * longjump out if this access should cause a guest exception. */ - int index; + CPUTLBEntry *entry; target_ulong tlb_addr; tlb_fill(cpu, addr, size, MMU_DATA_LOAD, mmu_idx, retaddr); - index = (addr >> TARGET_PAGE_BITS) & (CPU_TLB_SIZE - 1); - tlb_addr = env->tlb_table[mmu_idx][index].addr_read; + entry = tlb_entry(env, mmu_idx, addr); + tlb_addr = entry->addr_read; if (!(tlb_addr & ~(TARGET_PAGE_MASK | TLB_RECHECK))) { /* RAM access */ - uintptr_t haddr = addr + env->tlb_table[mmu_idx][index].addend; + uintptr_t haddr = addr + entry->addend; return ldn_p((void *)haddr, size); } @@ -XXX,XX +XXX,XX @@ static void io_writex(CPUArchState *env, CPUIOTLBEntry *iotlbentry, * repeat the MMU check here. This tlb_fill() call might * longjump out if this access should cause a guest exception. */ - int index; + CPUTLBEntry *entry; target_ulong tlb_addr; tlb_fill(cpu, addr, size, MMU_DATA_STORE, mmu_idx, retaddr); - index = (addr >> TARGET_PAGE_BITS) & (CPU_TLB_SIZE - 1); - tlb_addr = env->tlb_table[mmu_idx][index].addr_write; + entry = tlb_entry(env, mmu_idx, addr); + tlb_addr = entry->addr_write; if (!(tlb_addr & ~(TARGET_PAGE_MASK | TLB_RECHECK))) { /* RAM access */ - uintptr_t haddr = addr + env->tlb_table[mmu_idx][index].addend; + uintptr_t haddr = addr + entry->addend; stn_p((void *)haddr, size, val); return; @@ -XXX,XX +XXX,XX @@ static bool victim_tlb_hit(CPUArchState *env, size_t mmu_idx, size_t index, */ tb_page_addr_t get_page_addr_code(CPUArchState *env, target_ulong addr) { - int mmu_idx, index; + uintptr_t mmu_idx = cpu_mmu_index(env, true); + uintptr_t index = tlb_index(env, mmu_idx, addr); + CPUTLBEntry *entry = tlb_entry(env, mmu_idx, addr); void *p; - index = (addr >> TARGET_PAGE_BITS) & (CPU_TLB_SIZE - 1); - mmu_idx = cpu_mmu_index(env, true); - if (unlikely(!tlb_hit(env->tlb_table[mmu_idx][index].addr_code, addr))) { + if (unlikely(!tlb_hit(entry->addr_code, addr))) { if (!VICTIM_TLB_HIT(addr_code, addr)) { tlb_fill(ENV_GET_CPU(env), addr, 0, MMU_INST_FETCH, mmu_idx, 0); } - assert(tlb_hit(env->tlb_table[mmu_idx][index].addr_code, addr)); + assert(tlb_hit(entry->addr_code, addr)); } - if (unlikely(env->tlb_table[mmu_idx][index].addr_code & - (TLB_RECHECK | TLB_MMIO))) { + if (unlikely(entry->addr_code & (TLB_RECHECK | TLB_MMIO))) { /* * Return -1 if we can't translate and execute from an entire * page of RAM here, which will cause us to execute by loading @@ -XXX,XX +XXX,XX @@ tb_page_addr_t get_page_addr_code(CPUArchState *env, target_ulong addr) return -1; } - p = (void *)((uintptr_t)addr + env->tlb_table[mmu_idx][index].addend); + p = (void *)((uintptr_t)addr + entry->addend); return qemu_ram_addr_from_host_nofail(p); } @@ -XXX,XX +XXX,XX @@ tb_page_addr_t get_page_addr_code(CPUArchState *env, target_ulong addr) void probe_write(CPUArchState *env, target_ulong addr, int size, int mmu_idx, uintptr_t retaddr) { - int index = (addr >> TARGET_PAGE_BITS) & (CPU_TLB_SIZE - 1); - target_ulong tlb_addr = env->tlb_table[mmu_idx][index].addr_write; + uintptr_t index = tlb_index(env, mmu_idx, addr); + CPUTLBEntry *entry = tlb_entry(env, mmu_idx, addr); - if (!tlb_hit(tlb_addr, addr)) { + if (!tlb_hit(entry->addr_write, addr)) { /* TLB entry is for a different page */ if (!VICTIM_TLB_HIT(addr_write, addr)) { tlb_fill(ENV_GET_CPU(env), addr, size, MMU_DATA_STORE, @@ -XXX,XX +XXX,XX @@ static void *atomic_mmu_lookup(CPUArchState *env, target_ulong addr, NotDirtyInfo *ndi) { size_t mmu_idx = get_mmuidx(oi); - size_t index = (addr >> TARGET_PAGE_BITS) & (CPU_TLB_SIZE - 1); - CPUTLBEntry *tlbe = &env->tlb_table[mmu_idx][index]; + uintptr_t index = tlb_index(env, mmu_idx, addr); + CPUTLBEntry *tlbe = tlb_entry(env, mmu_idx, addr); target_ulong tlb_addr = tlbe->addr_write; TCGMemOp mop = get_memop(oi); int a_bits = get_alignment_bits(mop); -- 2.17.2 GCC7+ will no longer advertise support for 16-byte __atomic operations if only cmpxchg is supported, as for x86_64. Fortunately, x86_64 still has support for __sync_compare_and_swap_16 and we can make use of that. AArch64 does not have, nor ever has had such support, so open-code it. Reviewed-by: Emilio G. Cota <cota@braap.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- accel/tcg/atomic_template.h | 20 ++++- include/qemu/atomic128.h | 155 ++++++++++++++++++++++++++++++++++++ tcg/tcg.h | 16 ++-- accel/tcg/cputlb.c | 3 +- accel/tcg/user-exec.c | 5 +- configure | 19 +++++ 6 files changed, 204 insertions(+), 14 deletions(-) create mode 100644 include/qemu/atomic128.h diff --git a/accel/tcg/atomic_template.h b/accel/tcg/atomic_template.h index XXXXXXX..XXXXXXX 100644 --- a/accel/tcg/atomic_template.h +++ b/accel/tcg/atomic_template.h @@ -XXX,XX +XXX,XX @@ ABI_TYPE ATOMIC_NAME(cmpxchg)(CPUArchState *env, target_ulong addr, DATA_TYPE ret; ATOMIC_TRACE_RMW; +#if DATA_SIZE == 16 + ret = atomic16_cmpxchg(haddr, cmpv, newv); +#else ret = atomic_cmpxchg__nocheck(haddr, cmpv, newv); +#endif ATOMIC_MMU_CLEANUP; return ret; } #if DATA_SIZE >= 16 +#if HAVE_ATOMIC128 ABI_TYPE ATOMIC_NAME(ld)(CPUArchState *env, target_ulong addr EXTRA_ARGS) { ATOMIC_MMU_DECLS; DATA_TYPE val, *haddr = ATOMIC_MMU_LOOKUP; ATOMIC_TRACE_LD; - __atomic_load(haddr, &val, __ATOMIC_RELAXED); + val = atomic16_read(haddr); ATOMIC_MMU_CLEANUP; return val; } @@ -XXX,XX +XXX,XX @@ void ATOMIC_NAME(st)(CPUArchState *env, target_ulong addr, DATA_TYPE *haddr = ATOMIC_MMU_LOOKUP; ATOMIC_TRACE_ST; - __atomic_store(haddr, &val, __ATOMIC_RELAXED); + atomic16_set(haddr, val); ATOMIC_MMU_CLEANUP; } +#endif #else ABI_TYPE ATOMIC_NAME(xchg)(CPUArchState *env, target_ulong addr, ABI_TYPE val EXTRA_ARGS) @@ -XXX,XX +XXX,XX @@ ABI_TYPE ATOMIC_NAME(cmpxchg)(CPUArchState *env, target_ulong addr, DATA_TYPE ret; ATOMIC_TRACE_RMW; +#if DATA_SIZE == 16 + ret = atomic16_cmpxchg(haddr, BSWAP(cmpv), BSWAP(newv)); +#else ret = atomic_cmpxchg__nocheck(haddr, BSWAP(cmpv), BSWAP(newv)); +#endif ATOMIC_MMU_CLEANUP; return BSWAP(ret); } #if DATA_SIZE >= 16 +#if HAVE_ATOMIC128 ABI_TYPE ATOMIC_NAME(ld)(CPUArchState *env, target_ulong addr EXTRA_ARGS) { ATOMIC_MMU_DECLS; DATA_TYPE val, *haddr = ATOMIC_MMU_LOOKUP; ATOMIC_TRACE_LD; - __atomic_load(haddr, &val, __ATOMIC_RELAXED); + val = atomic16_read(haddr); ATOMIC_MMU_CLEANUP; return BSWAP(val); } @@ -XXX,XX +XXX,XX @@ void ATOMIC_NAME(st)(CPUArchState *env, target_ulong addr, ATOMIC_TRACE_ST; val = BSWAP(val); - __atomic_store(haddr, &val, __ATOMIC_RELAXED); + atomic16_set(haddr, val); ATOMIC_MMU_CLEANUP; } +#endif #else ABI_TYPE ATOMIC_NAME(xchg)(CPUArchState *env, target_ulong addr, ABI_TYPE val EXTRA_ARGS) diff --git a/include/qemu/atomic128.h b/include/qemu/atomic128.h new file mode 100644 index XXXXXXX..XXXXXXX --- /dev/null +++ b/include/qemu/atomic128.h @@ -XXX,XX +XXX,XX @@ +/* + * Simple interface for 128-bit atomic operations. + * + * Copyright (C) 2018 Linaro, Ltd. + * + * This work is licensed under the terms of the GNU GPL, version 2 or later. + * See the COPYING file in the top-level directory. + * + * See docs/devel/atomics.txt for discussion about the guarantees each + * atomic primitive is meant to provide. + */ + +#ifndef QEMU_ATOMIC128_H +#define QEMU_ATOMIC128_H + +/* + * GCC is a house divided about supporting large atomic operations. + * + * For hosts that only have large compare-and-swap, a legalistic reading + * of the C++ standard means that one cannot implement __atomic_read on + * read-only memory, and thus all atomic operations must synchronize + * through libatomic. + * + * See https://gcc.gnu.org/bugzilla/show_bug.cgi?id=80878 + * + * This interpretation is not especially helpful for QEMU. + * For softmmu, all RAM is always read/write from the hypervisor. + * For user-only, if the guest doesn't implement such an __atomic_read + * then the host need not worry about it either. + * + * Moreover, using libatomic is not an option, because its interface is + * built for std::atomic<T>, and requires that *all* accesses to such an + * object go through the library. In our case we do not have an object + * in the C/C++ sense, but a view of memory as seen by the guest. + * The guest may issue a large atomic operation and then access those + * pieces using word-sized accesses. From the hypervisor, we have no + * way to connect those two actions. + * + * Therefore, special case each platform. + */ + +#if defined(CONFIG_ATOMIC128) +static inline Int128 atomic16_cmpxchg(Int128 *ptr, Int128 cmp, Int128 new) +{ + return atomic_cmpxchg__nocheck(ptr, cmp, new); +} +# define HAVE_CMPXCHG128 1 +#elif defined(CONFIG_CMPXCHG128) +static inline Int128 atomic16_cmpxchg(Int128 *ptr, Int128 cmp, Int128 new) +{ + return __sync_val_compare_and_swap_16(ptr, cmp, new); +} +# define HAVE_CMPXCHG128 1 +#elif defined(__aarch64__) +/* Through gcc 8, aarch64 has no support for 128-bit at all. */ +static inline Int128 atomic16_cmpxchg(Int128 *ptr, Int128 cmp, Int128 new) +{ + uint64_t cmpl = int128_getlo(cmp), cmph = int128_gethi(cmp); + uint64_t newl = int128_getlo(new), newh = int128_gethi(new); + uint64_t oldl, oldh; + uint32_t tmp; + + asm("0: ldaxp %[oldl], %[oldh], %[mem]\n\t" + "cmp %[oldl], %[cmpl]\n\t" + "ccmp %[oldh], %[cmph], #0, eq\n\t" + "b.ne 1f\n\t" + "stlxp %w[tmp], %[newl], %[newh], %[mem]\n\t" + "cbnz %w[tmp], 0b\n" + "1:" + : [mem] "+m"(*ptr), [tmp] "=&r"(tmp), + [oldl] "=&r"(oldl), [oldh] "=r"(oldh) + : [cmpl] "r"(cmpl), [cmph] "r"(cmph), + [newl] "r"(newl), [newh] "r"(newh) + : "memory", "cc"); + + return int128_make128(oldl, oldh); +} +# define HAVE_CMPXCHG128 1 +#else +/* Fallback definition that must be optimized away, or error. */ +Int128 __attribute__((error("unsupported atomic"))) + atomic16_cmpxchg(Int128 *ptr, Int128 cmp, Int128 new); +# define HAVE_CMPXCHG128 0 +#endif /* Some definition for HAVE_CMPXCHG128 */ + + +#if defined(CONFIG_ATOMIC128) +static inline Int128 atomic16_read(Int128 *ptr) +{ + return atomic_read__nocheck(ptr); +} + +static inline void atomic16_set(Int128 *ptr, Int128 val) +{ + atomic_set__nocheck(ptr, val); +} + +# define HAVE_ATOMIC128 1 +#elif !defined(CONFIG_USER_ONLY) && defined(__aarch64__) +/* We can do better than cmpxchg for AArch64. */ +static inline Int128 atomic16_read(Int128 *ptr) +{ + uint64_t l, h; + uint32_t tmp; + + /* The load must be paired with the store to guarantee not tearing. */ + asm("0: ldxp %[l], %[h], %[mem]\n\t" + "stxp %w[tmp], %[l], %[h], %[mem]\n\t" + "cbnz %w[tmp], 0b" + : [mem] "+m"(*ptr), [tmp] "=r"(tmp), [l] "=r"(l), [h] "=r"(h)); + + return int128_make128(l, h); +} + +static inline void atomic16_set(Int128 *ptr, Int128 val) +{ + uint64_t l = int128_getlo(val), h = int128_gethi(val); + uint64_t t1, t2; + + /* Load into temporaries to acquire the exclusive access lock. */ + asm("0: ldxp %[t1], %[t2], %[mem]\n\t" + "stxp %w[t1], %[l], %[h], %[mem]\n\t" + "cbnz %w[t1], 0b" + : [mem] "+m"(*ptr), [t1] "=&r"(t1), [t2] "=&r"(t2) + : [l] "r"(l), [h] "r"(h)); +} + +# define HAVE_ATOMIC128 1 +#elif !defined(CONFIG_USER_ONLY) && HAVE_CMPXCHG128 +static inline Int128 atomic16_read(Int128 *ptr) +{ + /* Maybe replace 0 with 0, returning the old value. */ + return atomic16_cmpxchg(ptr, 0, 0); +} + +static inline void atomic16_set(Int128 *ptr, Int128 val) +{ + Int128 old = *ptr, cmp; + do { + cmp = old; + old = atomic16_cmpxchg(ptr, cmp, val); + } while (old != cmp); +} + +# define HAVE_ATOMIC128 1 +#else +/* Fallback definitions that must be optimized away, or error. */ +Int128 __attribute__((error("unsupported atomic"))) + atomic16_read(Int128 *ptr); +void __attribute__((error("unsupported atomic"))) + atomic16_set(Int128 *ptr, Int128 val); +# define HAVE_ATOMIC128 0 +#endif /* Some definition for HAVE_ATOMIC128 */ + +#endif /* QEMU_ATOMIC128_H */ diff --git a/tcg/tcg.h b/tcg/tcg.h index XXXXXXX..XXXXXXX 100644 --- a/tcg/tcg.h +++ b/tcg/tcg.h @@ -XXX,XX +XXX,XX @@ #include "qemu/queue.h" #include "tcg-mo.h" #include "tcg-target.h" +#include "qemu/int128.h" /* XXX: make safe guess about sizes */ #define MAX_OP_PER_INSTR 266 @@ -XXX,XX +XXX,XX @@ GEN_ATOMIC_HELPER_ALL(xchg) #undef GEN_ATOMIC_HELPER #endif /* CONFIG_SOFTMMU */ -#ifdef CONFIG_ATOMIC128 -#include "qemu/int128.h" - -/* These aren't really a "proper" helpers because TCG cannot manage Int128. - However, use the same format as the others, for use by the backends. */ +/* + * These aren't really a "proper" helpers because TCG cannot manage Int128. + * However, use the same format as the others, for use by the backends. + * + * The cmpxchg functions are only defined if HAVE_CMPXCHG128; + * the ld/st functions are only defined if HAVE_ATOMIC128, + * as defined by <qemu/atomic128.h>. + */ Int128 helper_atomic_cmpxchgo_le_mmu(CPUArchState *env, target_ulong addr, Int128 cmpv, Int128 newv, TCGMemOpIdx oi, uintptr_t retaddr); @@ -XXX,XX +XXX,XX @@ void helper_atomic_sto_le_mmu(CPUArchState *env, target_ulong addr, Int128 val, void helper_atomic_sto_be_mmu(CPUArchState *env, target_ulong addr, Int128 val, TCGMemOpIdx oi, uintptr_t retaddr); -#endif /* CONFIG_ATOMIC128 */ - #endif /* TCG_H */ diff --git a/accel/tcg/cputlb.c b/accel/tcg/cputlb.c index XXXXXXX..XXXXXXX 100644 --- a/accel/tcg/cputlb.c +++ b/accel/tcg/cputlb.c @@ -XXX,XX +XXX,XX @@ #include "exec/log.h" #include "exec/helper-proto.h" #include "qemu/atomic.h" +#include "qemu/atomic128.h" /* DEBUG defines, enable DEBUG_TLB_LOG to log to the CPU_LOG_MMU target */ /* #define DEBUG_TLB */ @@ -XXX,XX +XXX,XX @@ static void *atomic_mmu_lookup(CPUArchState *env, target_ulong addr, #include "atomic_template.h" #endif -#ifdef CONFIG_ATOMIC128 +#if HAVE_CMPXCHG128 || HAVE_ATOMIC128 #define DATA_SIZE 16 #include "atomic_template.h" #endif diff --git a/accel/tcg/user-exec.c b/accel/tcg/user-exec.c index XXXXXXX..XXXXXXX 100644 --- a/accel/tcg/user-exec.c +++ b/accel/tcg/user-exec.c @@ -XXX,XX +XXX,XX @@ #include "exec/cpu_ldst.h" #include "translate-all.h" #include "exec/helper-proto.h" +#include "qemu/atomic128.h" #undef EAX #undef ECX @@ -XXX,XX +XXX,XX @@ static void *atomic_mmu_lookup(CPUArchState *env, target_ulong addr, /* The following is only callable from other helpers, and matches up with the softmmu version. */ -#ifdef CONFIG_ATOMIC128 +#if HAVE_ATOMIC128 || HAVE_CMPXCHG128 #undef EXTRA_ARGS #undef ATOMIC_NAME @@ -XXX,XX +XXX,XX @@ static void *atomic_mmu_lookup(CPUArchState *env, target_ulong addr, #define DATA_SIZE 16 #include "atomic_template.h" -#endif /* CONFIG_ATOMIC128 */ +#endif diff --git a/configure b/configure index XXXXXXX..XXXXXXX 100755 --- a/configure +++ b/configure @@ -XXX,XX +XXX,XX @@ EOF fi fi +cmpxchg128=no +if test "$int128" = yes -a "$atomic128" = no; then + cat > $TMPC << EOF +int main(void) +{ + unsigned __int128 x = 0, y = 0; + __sync_val_compare_and_swap_16(&x, y, x); + return 0; +} +EOF + if compile_prog "" "" ; then + cmpxchg128=yes + fi +fi + ######################################### # See if 64-bit atomic operations are supported. # Note that without __atomic builtins, we can only @@ -XXX,XX +XXX,XX @@ if test "$atomic128" = "yes" ; then echo "CONFIG_ATOMIC128=y" >> $config_host_mak fi +if test "$cmpxchg128" = "yes" ; then + echo "CONFIG_CMPXCHG128=y" >> $config_host_mak +fi + if test "$atomic64" = "yes" ; then echo "CONFIG_ATOMIC64=y" >> $config_host_mak fi -- 2.17.2 Reviewed-by: Emilio G. Cota <cota@braap.org> Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- target/i386/mem_helper.c | 9 ++++----- 1 file changed, 4 insertions(+), 5 deletions(-) diff --git a/target/i386/mem_helper.c b/target/i386/mem_helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/i386/mem_helper.c +++ b/target/i386/mem_helper.c @@ -XXX,XX +XXX,XX @@ #include "exec/exec-all.h" #include "exec/cpu_ldst.h" #include "qemu/int128.h" +#include "qemu/atomic128.h" #include "tcg.h" void helper_cmpxchg8b_unlocked(CPUX86State *env, target_ulong a0) @@ -XXX,XX +XXX,XX @@ void helper_cmpxchg16b(CPUX86State *env, target_ulong a0) if ((a0 & 0xf) != 0) { raise_exception_ra(env, EXCP0D_GPF, ra); - } else { -#ifndef CONFIG_ATOMIC128 - cpu_loop_exit_atomic(ENV_GET_CPU(env), ra); -#else + } else if (HAVE_CMPXCHG128) { int eflags = cpu_cc_compute_all(env, CC_OP); Int128 cmpv = int128_make128(env->regs[R_EAX], env->regs[R_EDX]); @@ -XXX,XX +XXX,XX @@ void helper_cmpxchg16b(CPUX86State *env, target_ulong a0) eflags &= ~CC_Z; } CC_SRC = eflags; -#endif + } else { + cpu_loop_exit_atomic(ENV_GET_CPU(env), ra); } } #endif -- 2.17.2 Reviewed-by: Emilio G. Cota <cota@braap.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- target/arm/helper-a64.c | 259 +++++++++++++++++++++------------------- 1 file changed, 133 insertions(+), 126 deletions(-) diff --git a/target/arm/helper-a64.c b/target/arm/helper-a64.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper-a64.c +++ b/target/arm/helper-a64.c @@ -XXX,XX +XXX,XX @@ #include "exec/exec-all.h" #include "exec/cpu_ldst.h" #include "qemu/int128.h" +#include "qemu/atomic128.h" #include "tcg.h" #include "fpu/softfloat.h" #include <zlib.h> /* For crc32 */ @@ -XXX,XX +XXX,XX @@ uint64_t HELPER(crc32c_64)(uint64_t acc, uint64_t val, uint32_t bytes) return crc32c(acc, buf, bytes) ^ 0xffffffff; } -/* Returns 0 on success; 1 otherwise. */ -static uint64_t do_paired_cmpxchg64_le(CPUARMState *env, uint64_t addr, - uint64_t new_lo, uint64_t new_hi, - bool parallel, uintptr_t ra) +uint64_t HELPER(paired_cmpxchg64_le)(CPUARMState *env, uint64_t addr, + uint64_t new_lo, uint64_t new_hi) { - Int128 oldv, cmpv, newv; + Int128 cmpv = int128_make128(env->exclusive_val, env->exclusive_high); + Int128 newv = int128_make128(new_lo, new_hi); + Int128 oldv; + uintptr_t ra = GETPC(); + uint64_t o0, o1; bool success; - cmpv = int128_make128(env->exclusive_val, env->exclusive_high); - newv = int128_make128(new_lo, new_hi); - - if (parallel) { -#ifndef CONFIG_ATOMIC128 - cpu_loop_exit_atomic(ENV_GET_CPU(env), ra); -#else - int mem_idx = cpu_mmu_index(env, false); - TCGMemOpIdx oi = make_memop_idx(MO_LEQ | MO_ALIGN_16, mem_idx); - oldv = helper_atomic_cmpxchgo_le_mmu(env, addr, cmpv, newv, oi, ra); - success = int128_eq(oldv, cmpv); -#endif - } else { - uint64_t o0, o1; - #ifdef CONFIG_USER_ONLY - /* ??? Enforce alignment. */ - uint64_t *haddr = g2h(addr); + /* ??? Enforce alignment. */ + uint64_t *haddr = g2h(addr); - helper_retaddr = ra; - o0 = ldq_le_p(haddr + 0); - o1 = ldq_le_p(haddr + 1); - oldv = int128_make128(o0, o1); + helper_retaddr = ra; + o0 = ldq_le_p(haddr + 0); + o1 = ldq_le_p(haddr + 1); + oldv = int128_make128(o0, o1); - success = int128_eq(oldv, cmpv); - if (success) { - stq_le_p(haddr + 0, int128_getlo(newv)); - stq_le_p(haddr + 1, int128_gethi(newv)); - } - helper_retaddr = 0; -#else - int mem_idx = cpu_mmu_index(env, false); - TCGMemOpIdx oi0 = make_memop_idx(MO_LEQ | MO_ALIGN_16, mem_idx); - TCGMemOpIdx oi1 = make_memop_idx(MO_LEQ, mem_idx); - - o0 = helper_le_ldq_mmu(env, addr + 0, oi0, ra); - o1 = helper_le_ldq_mmu(env, addr + 8, oi1, ra); - oldv = int128_make128(o0, o1); - - success = int128_eq(oldv, cmpv); - if (success) { - helper_le_stq_mmu(env, addr + 0, int128_getlo(newv), oi1, ra); - helper_le_stq_mmu(env, addr + 8, int128_gethi(newv), oi1, ra); - } -#endif + success = int128_eq(oldv, cmpv); + if (success) { + stq_le_p(haddr + 0, int128_getlo(newv)); + stq_le_p(haddr + 1, int128_gethi(newv)); } + helper_retaddr = 0; +#else + int mem_idx = cpu_mmu_index(env, false); + TCGMemOpIdx oi0 = make_memop_idx(MO_LEQ | MO_ALIGN_16, mem_idx); + TCGMemOpIdx oi1 = make_memop_idx(MO_LEQ, mem_idx); + + o0 = helper_le_ldq_mmu(env, addr + 0, oi0, ra); + o1 = helper_le_ldq_mmu(env, addr + 8, oi1, ra); + oldv = int128_make128(o0, o1); + + success = int128_eq(oldv, cmpv); + if (success) { + helper_le_stq_mmu(env, addr + 0, int128_getlo(newv), oi1, ra); + helper_le_stq_mmu(env, addr + 8, int128_gethi(newv), oi1, ra); + } +#endif return !success; } -uint64_t HELPER(paired_cmpxchg64_le)(CPUARMState *env, uint64_t addr, - uint64_t new_lo, uint64_t new_hi) -{ - return do_paired_cmpxchg64_le(env, addr, new_lo, new_hi, false, GETPC()); -} - uint64_t HELPER(paired_cmpxchg64_le_parallel)(CPUARMState *env, uint64_t addr, uint64_t new_lo, uint64_t new_hi) -{ - return do_paired_cmpxchg64_le(env, addr, new_lo, new_hi, true, GETPC()); -} - -static uint64_t do_paired_cmpxchg64_be(CPUARMState *env, uint64_t addr, - uint64_t new_lo, uint64_t new_hi, - bool parallel, uintptr_t ra) { Int128 oldv, cmpv, newv; + uintptr_t ra = GETPC(); bool success; + int mem_idx; + TCGMemOpIdx oi; - /* high and low need to be switched here because this is not actually a - * 128bit store but two doublewords stored consecutively - */ - cmpv = int128_make128(env->exclusive_high, env->exclusive_val); - newv = int128_make128(new_hi, new_lo); - - if (parallel) { -#ifndef CONFIG_ATOMIC128 + if (!HAVE_CMPXCHG128) { cpu_loop_exit_atomic(ENV_GET_CPU(env), ra); -#else - int mem_idx = cpu_mmu_index(env, false); - TCGMemOpIdx oi = make_memop_idx(MO_BEQ | MO_ALIGN_16, mem_idx); - oldv = helper_atomic_cmpxchgo_be_mmu(env, addr, cmpv, newv, oi, ra); - success = int128_eq(oldv, cmpv); -#endif - } else { - uint64_t o0, o1; - -#ifdef CONFIG_USER_ONLY - /* ??? Enforce alignment. */ - uint64_t *haddr = g2h(addr); - - helper_retaddr = ra; - o1 = ldq_be_p(haddr + 0); - o0 = ldq_be_p(haddr + 1); - oldv = int128_make128(o0, o1); - - success = int128_eq(oldv, cmpv); - if (success) { - stq_be_p(haddr + 0, int128_gethi(newv)); - stq_be_p(haddr + 1, int128_getlo(newv)); - } - helper_retaddr = 0; -#else - int mem_idx = cpu_mmu_index(env, false); - TCGMemOpIdx oi0 = make_memop_idx(MO_BEQ | MO_ALIGN_16, mem_idx); - TCGMemOpIdx oi1 = make_memop_idx(MO_BEQ, mem_idx); - - o1 = helper_be_ldq_mmu(env, addr + 0, oi0, ra); - o0 = helper_be_ldq_mmu(env, addr + 8, oi1, ra); - oldv = int128_make128(o0, o1); - - success = int128_eq(oldv, cmpv); - if (success) { - helper_be_stq_mmu(env, addr + 0, int128_gethi(newv), oi1, ra); - helper_be_stq_mmu(env, addr + 8, int128_getlo(newv), oi1, ra); - } -#endif } + mem_idx = cpu_mmu_index(env, false); + oi = make_memop_idx(MO_LEQ | MO_ALIGN_16, mem_idx); + + cmpv = int128_make128(env->exclusive_val, env->exclusive_high); + newv = int128_make128(new_lo, new_hi); + oldv = helper_atomic_cmpxchgo_le_mmu(env, addr, cmpv, newv, oi, ra); + + success = int128_eq(oldv, cmpv); return !success; } uint64_t HELPER(paired_cmpxchg64_be)(CPUARMState *env, uint64_t addr, uint64_t new_lo, uint64_t new_hi) { - return do_paired_cmpxchg64_be(env, addr, new_lo, new_hi, false, GETPC()); + /* + * High and low need to be switched here because this is not actually a + * 128bit store but two doublewords stored consecutively + */ + Int128 cmpv = int128_make128(env->exclusive_val, env->exclusive_high); + Int128 newv = int128_make128(new_lo, new_hi); + Int128 oldv; + uintptr_t ra = GETPC(); + uint64_t o0, o1; + bool success; + +#ifdef CONFIG_USER_ONLY + /* ??? Enforce alignment. */ + uint64_t *haddr = g2h(addr); + + helper_retaddr = ra; + o1 = ldq_be_p(haddr + 0); + o0 = ldq_be_p(haddr + 1); + oldv = int128_make128(o0, o1); + + success = int128_eq(oldv, cmpv); + if (success) { + stq_be_p(haddr + 0, int128_gethi(newv)); + stq_be_p(haddr + 1, int128_getlo(newv)); + } + helper_retaddr = 0; +#else + int mem_idx = cpu_mmu_index(env, false); + TCGMemOpIdx oi0 = make_memop_idx(MO_BEQ | MO_ALIGN_16, mem_idx); + TCGMemOpIdx oi1 = make_memop_idx(MO_BEQ, mem_idx); + + o1 = helper_be_ldq_mmu(env, addr + 0, oi0, ra); + o0 = helper_be_ldq_mmu(env, addr + 8, oi1, ra); + oldv = int128_make128(o0, o1); + + success = int128_eq(oldv, cmpv); + if (success) { + helper_be_stq_mmu(env, addr + 0, int128_gethi(newv), oi1, ra); + helper_be_stq_mmu(env, addr + 8, int128_getlo(newv), oi1, ra); + } +#endif + + return !success; } uint64_t HELPER(paired_cmpxchg64_be_parallel)(CPUARMState *env, uint64_t addr, - uint64_t new_lo, uint64_t new_hi) + uint64_t new_lo, uint64_t new_hi) { - return do_paired_cmpxchg64_be(env, addr, new_lo, new_hi, true, GETPC()); + Int128 oldv, cmpv, newv; + uintptr_t ra = GETPC(); + bool success; + int mem_idx; + TCGMemOpIdx oi; + + if (!HAVE_CMPXCHG128) { + cpu_loop_exit_atomic(ENV_GET_CPU(env), ra); + } + + mem_idx = cpu_mmu_index(env, false); + oi = make_memop_idx(MO_BEQ | MO_ALIGN_16, mem_idx); + + /* + * High and low need to be switched here because this is not actually a + * 128bit store but two doublewords stored consecutively + */ + cmpv = int128_make128(env->exclusive_high, env->exclusive_val); + newv = int128_make128(new_hi, new_lo); + oldv = helper_atomic_cmpxchgo_be_mmu(env, addr, cmpv, newv, oi, ra); + + success = int128_eq(oldv, cmpv); + return !success; } /* Writes back the old data into Rs. */ void HELPER(casp_le_parallel)(CPUARMState *env, uint32_t rs, uint64_t addr, uint64_t new_lo, uint64_t new_hi) { - uintptr_t ra = GETPC(); -#ifndef CONFIG_ATOMIC128 - cpu_loop_exit_atomic(ENV_GET_CPU(env), ra); -#else Int128 oldv, cmpv, newv; + uintptr_t ra = GETPC(); + int mem_idx; + TCGMemOpIdx oi; + + if (!HAVE_CMPXCHG128) { + cpu_loop_exit_atomic(ENV_GET_CPU(env), ra); + } + + mem_idx = cpu_mmu_index(env, false); + oi = make_memop_idx(MO_LEQ | MO_ALIGN_16, mem_idx); cmpv = int128_make128(env->xregs[rs], env->xregs[rs + 1]); newv = int128_make128(new_lo, new_hi); - - int mem_idx = cpu_mmu_index(env, false); - TCGMemOpIdx oi = make_memop_idx(MO_LEQ | MO_ALIGN_16, mem_idx); oldv = helper_atomic_cmpxchgo_le_mmu(env, addr, cmpv, newv, oi, ra); env->xregs[rs] = int128_getlo(oldv); env->xregs[rs + 1] = int128_gethi(oldv); -#endif } void HELPER(casp_be_parallel)(CPUARMState *env, uint32_t rs, uint64_t addr, uint64_t new_hi, uint64_t new_lo) { - uintptr_t ra = GETPC(); -#ifndef CONFIG_ATOMIC128 - cpu_loop_exit_atomic(ENV_GET_CPU(env), ra); -#else Int128 oldv, cmpv, newv; + uintptr_t ra = GETPC(); + int mem_idx; + TCGMemOpIdx oi; + + if (!HAVE_CMPXCHG128) { + cpu_loop_exit_atomic(ENV_GET_CPU(env), ra); + } + + mem_idx = cpu_mmu_index(env, false); + oi = make_memop_idx(MO_LEQ | MO_ALIGN_16, mem_idx); cmpv = int128_make128(env->xregs[rs + 1], env->xregs[rs]); newv = int128_make128(new_lo, new_hi); - - int mem_idx = cpu_mmu_index(env, false); - TCGMemOpIdx oi = make_memop_idx(MO_LEQ | MO_ALIGN_16, mem_idx); oldv = helper_atomic_cmpxchgo_be_mmu(env, addr, cmpv, newv, oi, ra); env->xregs[rs + 1] = int128_getlo(oldv); env->xregs[rs] = int128_gethi(oldv); -#endif } /* -- 2.17.2 Reviewed-by: Emilio G. Cota <cota@braap.org> Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- target/arm/helper-a64.c | 16 ++++------------ target/arm/translate-a64.c | 38 ++++++++++++++++++++++---------------- 2 files changed, 26 insertions(+), 28 deletions(-) diff --git a/target/arm/helper-a64.c b/target/arm/helper-a64.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper-a64.c +++ b/target/arm/helper-a64.c @@ -XXX,XX +XXX,XX @@ uint64_t HELPER(paired_cmpxchg64_le_parallel)(CPUARMState *env, uint64_t addr, int mem_idx; TCGMemOpIdx oi; - if (!HAVE_CMPXCHG128) { - cpu_loop_exit_atomic(ENV_GET_CPU(env), ra); - } + assert(HAVE_CMPXCHG128); mem_idx = cpu_mmu_index(env, false); oi = make_memop_idx(MO_LEQ | MO_ALIGN_16, mem_idx); @@ -XXX,XX +XXX,XX @@ uint64_t HELPER(paired_cmpxchg64_be_parallel)(CPUARMState *env, uint64_t addr, int mem_idx; TCGMemOpIdx oi; - if (!HAVE_CMPXCHG128) { - cpu_loop_exit_atomic(ENV_GET_CPU(env), ra); - } + assert(HAVE_CMPXCHG128); mem_idx = cpu_mmu_index(env, false); oi = make_memop_idx(MO_BEQ | MO_ALIGN_16, mem_idx); @@ -XXX,XX +XXX,XX @@ void HELPER(casp_le_parallel)(CPUARMState *env, uint32_t rs, uint64_t addr, int mem_idx; TCGMemOpIdx oi; - if (!HAVE_CMPXCHG128) { - cpu_loop_exit_atomic(ENV_GET_CPU(env), ra); - } + assert(HAVE_CMPXCHG128); mem_idx = cpu_mmu_index(env, false); oi = make_memop_idx(MO_LEQ | MO_ALIGN_16, mem_idx); @@ -XXX,XX +XXX,XX @@ void HELPER(casp_be_parallel)(CPUARMState *env, uint32_t rs, uint64_t addr, int mem_idx; TCGMemOpIdx oi; - if (!HAVE_CMPXCHG128) { - cpu_loop_exit_atomic(ENV_GET_CPU(env), ra); - } + assert(HAVE_CMPXCHG128); mem_idx = cpu_mmu_index(env, false); oi = make_memop_idx(MO_LEQ | MO_ALIGN_16, mem_idx); diff --git a/target/arm/translate-a64.c b/target/arm/translate-a64.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-a64.c +++ b/target/arm/translate-a64.c @@ -XXX,XX +XXX,XX @@ #include "trace-tcg.h" #include "translate-a64.h" +#include "qemu/atomic128.h" static TCGv_i64 cpu_X[32]; static TCGv_i64 cpu_pc; @@ -XXX,XX +XXX,XX @@ static void gen_store_exclusive(DisasContext *s, int rd, int rt, int rt2, get_mem_index(s), MO_64 | MO_ALIGN | s->be_data); tcg_gen_setcond_i64(TCG_COND_NE, tmp, tmp, cpu_exclusive_val); - } else if (s->be_data == MO_LE) { - if (tb_cflags(s->base.tb) & CF_PARALLEL) { + } else if (tb_cflags(s->base.tb) & CF_PARALLEL) { + if (!HAVE_CMPXCHG128) { + gen_helper_exit_atomic(cpu_env); + s->base.is_jmp = DISAS_NORETURN; + } else if (s->be_data == MO_LE) { gen_helper_paired_cmpxchg64_le_parallel(tmp, cpu_env, cpu_exclusive_addr, cpu_reg(s, rt), cpu_reg(s, rt2)); } else { - gen_helper_paired_cmpxchg64_le(tmp, cpu_env, cpu_exclusive_addr, - cpu_reg(s, rt), cpu_reg(s, rt2)); - } - } else { - if (tb_cflags(s->base.tb) & CF_PARALLEL) { gen_helper_paired_cmpxchg64_be_parallel(tmp, cpu_env, cpu_exclusive_addr, cpu_reg(s, rt), cpu_reg(s, rt2)); - } else { - gen_helper_paired_cmpxchg64_be(tmp, cpu_env, cpu_exclusive_addr, - cpu_reg(s, rt), cpu_reg(s, rt2)); } + } else if (s->be_data == MO_LE) { + gen_helper_paired_cmpxchg64_le(tmp, cpu_env, cpu_exclusive_addr, + cpu_reg(s, rt), cpu_reg(s, rt2)); + } else { + gen_helper_paired_cmpxchg64_be(tmp, cpu_env, cpu_exclusive_addr, + cpu_reg(s, rt), cpu_reg(s, rt2)); } } else { tcg_gen_atomic_cmpxchg_i64(tmp, cpu_exclusive_addr, cpu_exclusive_val, @@ -XXX,XX +XXX,XX @@ static void gen_compare_and_swap_pair(DisasContext *s, int rs, int rt, } tcg_temp_free_i64(cmp); } else if (tb_cflags(s->base.tb) & CF_PARALLEL) { - TCGv_i32 tcg_rs = tcg_const_i32(rs); - - if (s->be_data == MO_LE) { - gen_helper_casp_le_parallel(cpu_env, tcg_rs, addr, t1, t2); + if (HAVE_CMPXCHG128) { + TCGv_i32 tcg_rs = tcg_const_i32(rs); + if (s->be_data == MO_LE) { + gen_helper_casp_le_parallel(cpu_env, tcg_rs, addr, t1, t2); + } else { + gen_helper_casp_be_parallel(cpu_env, tcg_rs, addr, t1, t2); + } + tcg_temp_free_i32(tcg_rs); } else { - gen_helper_casp_be_parallel(cpu_env, tcg_rs, addr, t1, t2); + gen_helper_exit_atomic(cpu_env); + s->base.is_jmp = DISAS_NORETURN; } - tcg_temp_free_i32(tcg_rs); } else { TCGv_i64 d1 = tcg_temp_new_i64(); TCGv_i64 d2 = tcg_temp_new_i64(); -- 2.17.2 Reviewed-by: Emilio G. Cota <cota@braap.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- target/ppc/helper.h | 2 +- target/ppc/mem_helper.c | 33 ++++++++++-- target/ppc/translate.c | 115 +++++++++++++++++++++------------------- 3 files changed, 88 insertions(+), 62 deletions(-) diff --git a/target/ppc/helper.h b/target/ppc/helper.h index XXXXXXX..XXXXXXX 100644 --- a/target/ppc/helper.h +++ b/target/ppc/helper.h @@ -XXX,XX +XXX,XX @@ DEF_HELPER_4(dscliq, void, env, fprp, fprp, i32) DEF_HELPER_1(tbegin, void, env) DEF_HELPER_FLAGS_1(fixup_thrm, TCG_CALL_NO_RWG, void, env) -#if defined(TARGET_PPC64) && defined(CONFIG_ATOMIC128) +#ifdef TARGET_PPC64 DEF_HELPER_FLAGS_3(lq_le_parallel, TCG_CALL_NO_WG, i64, env, tl, i32) DEF_HELPER_FLAGS_3(lq_be_parallel, TCG_CALL_NO_WG, i64, env, tl, i32) DEF_HELPER_FLAGS_5(stq_le_parallel, TCG_CALL_NO_WG, diff --git a/target/ppc/mem_helper.c b/target/ppc/mem_helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/ppc/mem_helper.c +++ b/target/ppc/mem_helper.c @@ -XXX,XX +XXX,XX @@ #include "exec/cpu_ldst.h" #include "tcg.h" #include "internal.h" +#include "qemu/atomic128.h" //#define DEBUG_OP @@ -XXX,XX +XXX,XX @@ target_ulong helper_lscbx(CPUPPCState *env, target_ulong addr, uint32_t reg, return i; } -#if defined(TARGET_PPC64) && defined(CONFIG_ATOMIC128) +#ifdef TARGET_PPC64 uint64_t helper_lq_le_parallel(CPUPPCState *env, target_ulong addr, uint32_t opidx) { - Int128 ret = helper_atomic_ldo_le_mmu(env, addr, opidx, GETPC()); + Int128 ret; + + /* We will have raised EXCP_ATOMIC from the translator. */ + assert(HAVE_ATOMIC128); + ret = helper_atomic_ldo_le_mmu(env, addr, opidx, GETPC()); env->retxh = int128_gethi(ret); return int128_getlo(ret); } @@ -XXX,XX +XXX,XX @@ uint64_t helper_lq_le_parallel(CPUPPCState *env, target_ulong addr, uint64_t helper_lq_be_parallel(CPUPPCState *env, target_ulong addr, uint32_t opidx) { - Int128 ret = helper_atomic_ldo_be_mmu(env, addr, opidx, GETPC()); + Int128 ret; + + /* We will have raised EXCP_ATOMIC from the translator. */ + assert(HAVE_ATOMIC128); + ret = helper_atomic_ldo_be_mmu(env, addr, opidx, GETPC()); env->retxh = int128_gethi(ret); return int128_getlo(ret); } @@ -XXX,XX +XXX,XX @@ uint64_t helper_lq_be_parallel(CPUPPCState *env, target_ulong addr, void helper_stq_le_parallel(CPUPPCState *env, target_ulong addr, uint64_t lo, uint64_t hi, uint32_t opidx) { - Int128 val = int128_make128(lo, hi); + Int128 val; + + /* We will have raised EXCP_ATOMIC from the translator. */ + assert(HAVE_ATOMIC128); + val = int128_make128(lo, hi); helper_atomic_sto_le_mmu(env, addr, val, opidx, GETPC()); } void helper_stq_be_parallel(CPUPPCState *env, target_ulong addr, uint64_t lo, uint64_t hi, uint32_t opidx) { - Int128 val = int128_make128(lo, hi); + Int128 val; + + /* We will have raised EXCP_ATOMIC from the translator. */ + assert(HAVE_ATOMIC128); + val = int128_make128(lo, hi); helper_atomic_sto_be_mmu(env, addr, val, opidx, GETPC()); } @@ -XXX,XX +XXX,XX @@ uint32_t helper_stqcx_le_parallel(CPUPPCState *env, target_ulong addr, { bool success = false; + /* We will have raised EXCP_ATOMIC from the translator. */ + assert(HAVE_CMPXCHG128); + if (likely(addr == env->reserve_addr)) { Int128 oldv, cmpv, newv; @@ -XXX,XX +XXX,XX @@ uint32_t helper_stqcx_be_parallel(CPUPPCState *env, target_ulong addr, { bool success = false; + /* We will have raised EXCP_ATOMIC from the translator. */ + assert(HAVE_CMPXCHG128); + if (likely(addr == env->reserve_addr)) { Int128 oldv, cmpv, newv; 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 "trace-tcg.h" #include "exec/translator.h" #include "exec/log.h" +#include "qemu/atomic128.h" #define CPU_SINGLE_STEP 0x1 @@ -XXX,XX +XXX,XX @@ static void gen_lq(DisasContext *ctx) hi = cpu_gpr[rd]; if (tb_cflags(ctx->base.tb) & CF_PARALLEL) { -#ifdef CONFIG_ATOMIC128 - TCGv_i32 oi = tcg_temp_new_i32(); - if (ctx->le_mode) { - tcg_gen_movi_i32(oi, make_memop_idx(MO_LEQ, ctx->mem_idx)); - gen_helper_lq_le_parallel(lo, cpu_env, EA, oi); + if (HAVE_ATOMIC128) { + TCGv_i32 oi = tcg_temp_new_i32(); + if (ctx->le_mode) { + tcg_gen_movi_i32(oi, make_memop_idx(MO_LEQ, ctx->mem_idx)); + gen_helper_lq_le_parallel(lo, cpu_env, EA, oi); + } else { + tcg_gen_movi_i32(oi, make_memop_idx(MO_BEQ, ctx->mem_idx)); + gen_helper_lq_be_parallel(lo, cpu_env, EA, oi); + } + tcg_temp_free_i32(oi); + tcg_gen_ld_i64(hi, cpu_env, offsetof(CPUPPCState, retxh)); } else { - tcg_gen_movi_i32(oi, make_memop_idx(MO_BEQ, ctx->mem_idx)); - gen_helper_lq_be_parallel(lo, cpu_env, EA, oi); + /* Restart with exclusive lock. */ + gen_helper_exit_atomic(cpu_env); + ctx->base.is_jmp = DISAS_NORETURN; } - tcg_temp_free_i32(oi); - tcg_gen_ld_i64(hi, cpu_env, offsetof(CPUPPCState, retxh)); -#else - /* Restart with exclusive lock. */ - gen_helper_exit_atomic(cpu_env); - ctx->base.is_jmp = DISAS_NORETURN; -#endif } else if (ctx->le_mode) { tcg_gen_qemu_ld_i64(lo, EA, ctx->mem_idx, MO_LEQ); gen_addr_add(ctx, EA, EA, 8); @@ -XXX,XX +XXX,XX @@ static void gen_std(DisasContext *ctx) hi = cpu_gpr[rs]; if (tb_cflags(ctx->base.tb) & CF_PARALLEL) { -#ifdef CONFIG_ATOMIC128 - TCGv_i32 oi = tcg_temp_new_i32(); - if (ctx->le_mode) { - tcg_gen_movi_i32(oi, make_memop_idx(MO_LEQ, ctx->mem_idx)); - gen_helper_stq_le_parallel(cpu_env, EA, lo, hi, oi); + if (HAVE_ATOMIC128) { + TCGv_i32 oi = tcg_temp_new_i32(); + if (ctx->le_mode) { + tcg_gen_movi_i32(oi, make_memop_idx(MO_LEQ, ctx->mem_idx)); + gen_helper_stq_le_parallel(cpu_env, EA, lo, hi, oi); + } else { + tcg_gen_movi_i32(oi, make_memop_idx(MO_BEQ, ctx->mem_idx)); + gen_helper_stq_be_parallel(cpu_env, EA, lo, hi, oi); + } + tcg_temp_free_i32(oi); } else { - tcg_gen_movi_i32(oi, make_memop_idx(MO_BEQ, ctx->mem_idx)); - gen_helper_stq_be_parallel(cpu_env, EA, lo, hi, oi); + /* Restart with exclusive lock. */ + gen_helper_exit_atomic(cpu_env); + ctx->base.is_jmp = DISAS_NORETURN; } - tcg_temp_free_i32(oi); -#else - /* Restart with exclusive lock. */ - gen_helper_exit_atomic(cpu_env); - ctx->base.is_jmp = DISAS_NORETURN; -#endif } else if (ctx->le_mode) { tcg_gen_qemu_st_i64(lo, EA, ctx->mem_idx, MO_LEQ); gen_addr_add(ctx, EA, EA, 8); @@ -XXX,XX +XXX,XX @@ static void gen_lqarx(DisasContext *ctx) hi = cpu_gpr[rd]; if (tb_cflags(ctx->base.tb) & CF_PARALLEL) { -#ifdef CONFIG_ATOMIC128 - TCGv_i32 oi = tcg_temp_new_i32(); - if (ctx->le_mode) { - tcg_gen_movi_i32(oi, make_memop_idx(MO_LEQ | MO_ALIGN_16, - ctx->mem_idx)); - gen_helper_lq_le_parallel(lo, cpu_env, EA, oi); + if (HAVE_ATOMIC128) { + TCGv_i32 oi = tcg_temp_new_i32(); + if (ctx->le_mode) { + tcg_gen_movi_i32(oi, make_memop_idx(MO_LEQ | MO_ALIGN_16, + ctx->mem_idx)); + gen_helper_lq_le_parallel(lo, cpu_env, EA, oi); + } else { + tcg_gen_movi_i32(oi, make_memop_idx(MO_BEQ | MO_ALIGN_16, + ctx->mem_idx)); + gen_helper_lq_be_parallel(lo, cpu_env, EA, oi); + } + tcg_temp_free_i32(oi); + tcg_gen_ld_i64(hi, cpu_env, offsetof(CPUPPCState, retxh)); } else { - tcg_gen_movi_i32(oi, make_memop_idx(MO_BEQ | MO_ALIGN_16, - ctx->mem_idx)); - gen_helper_lq_be_parallel(lo, cpu_env, EA, oi); + /* Restart with exclusive lock. */ + gen_helper_exit_atomic(cpu_env); + ctx->base.is_jmp = DISAS_NORETURN; + tcg_temp_free(EA); + return; } - tcg_temp_free_i32(oi); - tcg_gen_ld_i64(hi, cpu_env, offsetof(CPUPPCState, retxh)); -#else - /* Restart with exclusive lock. */ - gen_helper_exit_atomic(cpu_env); - ctx->base.is_jmp = DISAS_NORETURN; - tcg_temp_free(EA); - return; -#endif } else if (ctx->le_mode) { tcg_gen_qemu_ld_i64(lo, EA, ctx->mem_idx, MO_LEQ | MO_ALIGN_16); tcg_gen_mov_tl(cpu_reserve, EA); @@ -XXX,XX +XXX,XX @@ static void gen_stqcx_(DisasContext *ctx) hi = cpu_gpr[rs]; if (tb_cflags(ctx->base.tb) & CF_PARALLEL) { - TCGv_i32 oi = tcg_const_i32(DEF_MEMOP(MO_Q) | MO_ALIGN_16); -#ifdef CONFIG_ATOMIC128 - if (ctx->le_mode) { - gen_helper_stqcx_le_parallel(cpu_crf[0], cpu_env, EA, lo, hi, oi); + if (HAVE_CMPXCHG128) { + TCGv_i32 oi = tcg_const_i32(DEF_MEMOP(MO_Q) | MO_ALIGN_16); + if (ctx->le_mode) { + gen_helper_stqcx_le_parallel(cpu_crf[0], cpu_env, + EA, lo, hi, oi); + } else { + gen_helper_stqcx_be_parallel(cpu_crf[0], cpu_env, + EA, lo, hi, oi); + } + tcg_temp_free_i32(oi); } else { - gen_helper_stqcx_le_parallel(cpu_crf[0], cpu_env, EA, lo, hi, oi); + /* Restart with exclusive lock. */ + gen_helper_exit_atomic(cpu_env); + ctx->base.is_jmp = DISAS_NORETURN; } -#else - /* Restart with exclusive lock. */ - gen_helper_exit_atomic(cpu_env); - ctx->base.is_jmp = DISAS_NORETURN; -#endif tcg_temp_free(EA); - tcg_temp_free_i32(oi); } else { TCGLabel *lab_fail = gen_new_label(); TCGLabel *lab_over = gen_new_label(); -- 2.17.2 Reviewed-by: David Hildenbrand <david@redhat.com> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- target/s390x/mem_helper.c | 92 +++++++++++++++++---------------------- 1 file changed, 41 insertions(+), 51 deletions(-) diff --git a/target/s390x/mem_helper.c b/target/s390x/mem_helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/s390x/mem_helper.c +++ b/target/s390x/mem_helper.c @@ -XXX,XX +XXX,XX @@ #include "exec/exec-all.h" #include "exec/cpu_ldst.h" #include "qemu/int128.h" +#include "qemu/atomic128.h" #if !defined(CONFIG_USER_ONLY) #include "hw/s390x/storage-keys.h" @@ -XXX,XX +XXX,XX @@ static void do_cdsg(CPUS390XState *env, uint64_t addr, bool fail; if (parallel) { -#ifndef CONFIG_ATOMIC128 +#if !HAVE_CMPXCHG128 cpu_loop_exit_atomic(ENV_GET_CPU(env), ra); #else int mem_idx = cpu_mmu_index(env, false); @@ -XXX,XX +XXX,XX @@ void HELPER(cdsg_parallel)(CPUS390XState *env, uint64_t addr, static uint32_t do_csst(CPUS390XState *env, uint32_t r3, uint64_t a1, uint64_t a2, bool parallel) { -#if !defined(CONFIG_USER_ONLY) || defined(CONFIG_ATOMIC128) uint32_t mem_idx = cpu_mmu_index(env, false); -#endif uintptr_t ra = GETPC(); uint32_t fc = extract32(env->regs[0], 0, 8); uint32_t sc = extract32(env->regs[0], 8, 8); @@ -XXX,XX +XXX,XX @@ static uint32_t do_csst(CPUS390XState *env, uint32_t r3, uint64_t a1, probe_write(env, a2, 0, mem_idx, ra); #endif - /* Note that the compare-and-swap is atomic, and the store is atomic, but - the complete operation is not. Therefore we do not need to assert serial - context in order to implement this. That said, restart early if we can't - support either operation that is supposed to be atomic. */ + /* + * Note that the compare-and-swap is atomic, and the store is atomic, + * but the complete operation is not. Therefore we do not need to + * assert serial context in order to implement this. That said, + * restart early if we can't support either operation that is supposed + * to be atomic. + */ if (parallel) { - int mask = 0; -#if !defined(CONFIG_ATOMIC64) - mask = -8; -#elif !defined(CONFIG_ATOMIC128) - mask = -16; + uint32_t max = 2; +#ifdef CONFIG_ATOMIC64 + max = 3; #endif - if (((4 << fc) | (1 << sc)) & mask) { + if ((HAVE_CMPXCHG128 ? 0 : fc + 2 > max) || + (HAVE_ATOMIC128 ? 0 : sc > max)) { cpu_loop_exit_atomic(ENV_GET_CPU(env), ra); } } @@ -XXX,XX +XXX,XX @@ static uint32_t do_csst(CPUS390XState *env, uint32_t r3, uint64_t a1, Int128 cv = int128_make128(env->regs[r3 + 1], env->regs[r3]); Int128 ov; - if (parallel) { -#ifdef CONFIG_ATOMIC128 - TCGMemOpIdx oi = make_memop_idx(MO_TEQ | MO_ALIGN_16, mem_idx); - ov = helper_atomic_cmpxchgo_be_mmu(env, a1, cv, nv, oi, ra); - cc = !int128_eq(ov, cv); -#else - /* Note that we asserted !parallel above. */ - g_assert_not_reached(); -#endif - } else { + if (!parallel) { uint64_t oh = cpu_ldq_data_ra(env, a1 + 0, ra); uint64_t ol = cpu_ldq_data_ra(env, a1 + 8, ra); @@ -XXX,XX +XXX,XX @@ static uint32_t do_csst(CPUS390XState *env, uint32_t r3, uint64_t a1, cpu_stq_data_ra(env, a1 + 0, int128_gethi(nv), ra); cpu_stq_data_ra(env, a1 + 8, int128_getlo(nv), ra); + } else if (HAVE_CMPXCHG128) { + TCGMemOpIdx oi = make_memop_idx(MO_TEQ | MO_ALIGN_16, mem_idx); + ov = helper_atomic_cmpxchgo_be_mmu(env, a1, cv, nv, oi, ra); + cc = !int128_eq(ov, cv); + } else { + /* Note that we asserted !parallel above. */ + g_assert_not_reached(); } env->regs[r3 + 0] = int128_gethi(ov); @@ -XXX,XX +XXX,XX @@ static uint32_t do_csst(CPUS390XState *env, uint32_t r3, uint64_t a1, cpu_stq_data_ra(env, a2, svh, ra); break; case 4: - if (parallel) { -#ifdef CONFIG_ATOMIC128 + if (!parallel) { + cpu_stq_data_ra(env, a2 + 0, svh, ra); + cpu_stq_data_ra(env, a2 + 8, svl, ra); + } else if (HAVE_ATOMIC128) { TCGMemOpIdx oi = make_memop_idx(MO_TEQ | MO_ALIGN_16, mem_idx); Int128 sv = int128_make128(svl, svh); helper_atomic_sto_be_mmu(env, a2, sv, oi, ra); -#else + } else { /* Note that we asserted !parallel above. */ g_assert_not_reached(); -#endif - } else { - cpu_stq_data_ra(env, a2 + 0, svh, ra); - cpu_stq_data_ra(env, a2 + 8, svl, ra); } break; default: @@ -XXX,XX +XXX,XX @@ static uint64_t do_lpq(CPUS390XState *env, uint64_t addr, bool parallel) uintptr_t ra = GETPC(); uint64_t hi, lo; - if (parallel) { -#ifndef CONFIG_ATOMIC128 - cpu_loop_exit_atomic(ENV_GET_CPU(env), ra); -#else + if (!parallel) { + check_alignment(env, addr, 16, ra); + hi = cpu_ldq_data_ra(env, addr + 0, ra); + lo = cpu_ldq_data_ra(env, addr + 8, ra); + } else if (HAVE_ATOMIC128) { int mem_idx = cpu_mmu_index(env, false); TCGMemOpIdx oi = make_memop_idx(MO_TEQ | MO_ALIGN_16, mem_idx); Int128 v = helper_atomic_ldo_be_mmu(env, addr, oi, ra); hi = int128_gethi(v); lo = int128_getlo(v); -#endif } else { - check_alignment(env, addr, 16, ra); - - hi = cpu_ldq_data_ra(env, addr + 0, ra); - lo = cpu_ldq_data_ra(env, addr + 8, ra); + cpu_loop_exit_atomic(ENV_GET_CPU(env), ra); } env->retxl = lo; @@ -XXX,XX +XXX,XX @@ static void do_stpq(CPUS390XState *env, uint64_t addr, { uintptr_t ra = GETPC(); - if (parallel) { -#ifndef CONFIG_ATOMIC128 - cpu_loop_exit_atomic(ENV_GET_CPU(env), ra); -#else - int mem_idx = cpu_mmu_index(env, false); - TCGMemOpIdx oi = make_memop_idx(MO_TEQ | MO_ALIGN_16, mem_idx); - - Int128 v = int128_make128(low, high); - helper_atomic_sto_be_mmu(env, addr, v, oi, ra); -#endif - } else { + if (!parallel) { check_alignment(env, addr, 16, ra); - cpu_stq_data_ra(env, addr + 0, high, ra); cpu_stq_data_ra(env, addr + 8, low, ra); + } else if (HAVE_ATOMIC128) { + int mem_idx = cpu_mmu_index(env, false); + TCGMemOpIdx oi = make_memop_idx(MO_TEQ | MO_ALIGN_16, mem_idx); + Int128 v = int128_make128(low, high); + helper_atomic_sto_be_mmu(env, addr, v, oi, ra); + } else { + cpu_loop_exit_atomic(ENV_GET_CPU(env), ra); } } -- 2.17.2 Reviewed-by: David Hildenbrand <david@redhat.com> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- target/s390x/mem_helper.c | 128 ++++++++++++++++++-------------------- 1 file changed, 61 insertions(+), 67 deletions(-) diff --git a/target/s390x/mem_helper.c b/target/s390x/mem_helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/s390x/mem_helper.c +++ b/target/s390x/mem_helper.c @@ -XXX,XX +XXX,XX @@ uint32_t HELPER(trXX)(CPUS390XState *env, uint32_t r1, uint32_t r2, return cc; } -static void do_cdsg(CPUS390XState *env, uint64_t addr, - uint32_t r1, uint32_t r3, bool parallel) +void HELPER(cdsg)(CPUS390XState *env, uint64_t addr, + uint32_t r1, uint32_t r3) { uintptr_t ra = GETPC(); Int128 cmpv = int128_make128(env->regs[r1 + 1], env->regs[r1]); Int128 newv = int128_make128(env->regs[r3 + 1], env->regs[r3]); Int128 oldv; + uint64_t oldh, oldl; bool fail; - if (parallel) { -#if !HAVE_CMPXCHG128 - cpu_loop_exit_atomic(ENV_GET_CPU(env), ra); -#else - int mem_idx = cpu_mmu_index(env, false); - TCGMemOpIdx oi = make_memop_idx(MO_TEQ | MO_ALIGN_16, mem_idx); - oldv = helper_atomic_cmpxchgo_be_mmu(env, addr, cmpv, newv, oi, ra); - fail = !int128_eq(oldv, cmpv); -#endif - } else { - uint64_t oldh, oldl; + check_alignment(env, addr, 16, ra); - check_alignment(env, addr, 16, ra); + oldh = cpu_ldq_data_ra(env, addr + 0, ra); + oldl = cpu_ldq_data_ra(env, addr + 8, ra); - oldh = cpu_ldq_data_ra(env, addr + 0, ra); - oldl = cpu_ldq_data_ra(env, addr + 8, ra); - - oldv = int128_make128(oldl, oldh); - fail = !int128_eq(oldv, cmpv); - if (fail) { - newv = oldv; - } - - cpu_stq_data_ra(env, addr + 0, int128_gethi(newv), ra); - cpu_stq_data_ra(env, addr + 8, int128_getlo(newv), ra); + oldv = int128_make128(oldl, oldh); + fail = !int128_eq(oldv, cmpv); + if (fail) { + newv = oldv; } + cpu_stq_data_ra(env, addr + 0, int128_gethi(newv), ra); + cpu_stq_data_ra(env, addr + 8, int128_getlo(newv), ra); + env->cc_op = fail; env->regs[r1] = int128_gethi(oldv); env->regs[r1 + 1] = int128_getlo(oldv); } -void HELPER(cdsg)(CPUS390XState *env, uint64_t addr, - uint32_t r1, uint32_t r3) -{ - do_cdsg(env, addr, r1, r3, false); -} - void HELPER(cdsg_parallel)(CPUS390XState *env, uint64_t addr, uint32_t r1, uint32_t r3) { - do_cdsg(env, addr, r1, r3, true); + uintptr_t ra = GETPC(); + Int128 cmpv = int128_make128(env->regs[r1 + 1], env->regs[r1]); + Int128 newv = int128_make128(env->regs[r3 + 1], env->regs[r3]); + int mem_idx; + TCGMemOpIdx oi; + Int128 oldv; + bool fail; + + if (!HAVE_CMPXCHG128) { + cpu_loop_exit_atomic(ENV_GET_CPU(env), ra); + } + + mem_idx = cpu_mmu_index(env, false); + oi = make_memop_idx(MO_TEQ | MO_ALIGN_16, mem_idx); + oldv = helper_atomic_cmpxchgo_be_mmu(env, addr, cmpv, newv, oi, ra); + fail = !int128_eq(oldv, cmpv); + + env->cc_op = fail; + env->regs[r1] = int128_gethi(oldv); + env->regs[r1 + 1] = int128_getlo(oldv); } static uint32_t do_csst(CPUS390XState *env, uint32_t r3, uint64_t a1, @@ -XXX,XX +XXX,XX @@ uint64_t HELPER(lra)(CPUS390XState *env, uint64_t addr) #endif /* load pair from quadword */ -static uint64_t do_lpq(CPUS390XState *env, uint64_t addr, bool parallel) +uint64_t HELPER(lpq)(CPUS390XState *env, uint64_t addr) { uintptr_t ra = GETPC(); uint64_t hi, lo; - if (!parallel) { - check_alignment(env, addr, 16, ra); - hi = cpu_ldq_data_ra(env, addr + 0, ra); - lo = cpu_ldq_data_ra(env, addr + 8, ra); - } else if (HAVE_ATOMIC128) { + check_alignment(env, addr, 16, ra); + hi = cpu_ldq_data_ra(env, addr + 0, ra); + lo = cpu_ldq_data_ra(env, addr + 8, ra); + + env->retxl = lo; + return hi; +} + +uint64_t HELPER(lpq_parallel)(CPUS390XState *env, uint64_t addr) +{ + uintptr_t ra = GETPC(); + uint64_t hi, lo; + + if (HAVE_ATOMIC128) { int mem_idx = cpu_mmu_index(env, false); TCGMemOpIdx oi = make_memop_idx(MO_TEQ | MO_ALIGN_16, mem_idx); Int128 v = helper_atomic_ldo_be_mmu(env, addr, oi, ra); @@ -XXX,XX +XXX,XX @@ static uint64_t do_lpq(CPUS390XState *env, uint64_t addr, bool parallel) return hi; } -uint64_t HELPER(lpq)(CPUS390XState *env, uint64_t addr) -{ - return do_lpq(env, addr, false); -} - -uint64_t HELPER(lpq_parallel)(CPUS390XState *env, uint64_t addr) -{ - return do_lpq(env, addr, true); -} - /* store pair to quadword */ -static void do_stpq(CPUS390XState *env, uint64_t addr, - uint64_t low, uint64_t high, bool parallel) +void HELPER(stpq)(CPUS390XState *env, uint64_t addr, + uint64_t low, uint64_t high) { uintptr_t ra = GETPC(); - if (!parallel) { - check_alignment(env, addr, 16, ra); - cpu_stq_data_ra(env, addr + 0, high, ra); - cpu_stq_data_ra(env, addr + 8, low, ra); - } else if (HAVE_ATOMIC128) { + check_alignment(env, addr, 16, ra); + cpu_stq_data_ra(env, addr + 0, high, ra); + cpu_stq_data_ra(env, addr + 8, low, ra); +} + +void HELPER(stpq_parallel)(CPUS390XState *env, uint64_t addr, + uint64_t low, uint64_t high) +{ + uintptr_t ra = GETPC(); + + if (HAVE_ATOMIC128) { int mem_idx = cpu_mmu_index(env, false); TCGMemOpIdx oi = make_memop_idx(MO_TEQ | MO_ALIGN_16, mem_idx); Int128 v = int128_make128(low, high); @@ -XXX,XX +XXX,XX @@ static void do_stpq(CPUS390XState *env, uint64_t addr, } } -void HELPER(stpq)(CPUS390XState *env, uint64_t addr, - uint64_t low, uint64_t high) -{ - do_stpq(env, addr, low, high, false); -} - -void HELPER(stpq_parallel)(CPUS390XState *env, uint64_t addr, - uint64_t low, uint64_t high) -{ - do_stpq(env, addr, low, high, true); -} - /* Execute instruction. This instruction executes an insn modified with the contents of r1. It does not change the executed instruction in memory; it does not change the program counter. -- 2.17.2 When op raises an exception, it may not have initialized the output temps that would be written back by wout or cout. Reviewed-by: David Hildenbrand <david@redhat.com> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- target/s390x/translate.c | 20 +++++++++++++++----- 1 file changed, 15 insertions(+), 5 deletions(-) diff --git a/target/s390x/translate.c b/target/s390x/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/s390x/translate.c +++ b/target/s390x/translate.c @@ -XXX,XX +XXX,XX @@ struct DisasInsn { const char *name; + /* Pre-process arguments before HELP_OP. */ void (*help_in1)(DisasContext *, DisasFields *, DisasOps *); void (*help_in2)(DisasContext *, DisasFields *, DisasOps *); void (*help_prep)(DisasContext *, DisasFields *, DisasOps *); + + /* + * Post-process output after HELP_OP. + * Note that these are not called if HELP_OP returns DISAS_NORETURN. + */ void (*help_wout)(DisasContext *, DisasFields *, DisasOps *); void (*help_cout)(DisasContext *, DisasOps *); + + /* Implement the operation itself. */ DisasJumpType (*help_op)(DisasContext *, DisasOps *); uint64_t data; @@ -XXX,XX +XXX,XX @@ static DisasJumpType translate_one(CPUS390XState *env, DisasContext *s) if (insn->help_op) { ret = insn->help_op(s, &o); } - if (insn->help_wout) { - insn->help_wout(s, &f, &o); - } - if (insn->help_cout) { - insn->help_cout(s, &o); + if (ret != DISAS_NORETURN) { + if (insn->help_wout) { + insn->help_wout(s, &f, &o); + } + if (insn->help_cout) { + insn->help_cout(s, &o); + } } /* Free any temporaries created by the helpers. */ -- 2.17.2 Reviewed-by: David Hildenbrand <david@redhat.com> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- target/s390x/mem_helper.c | 40 +++++++++++++++++++-------------------- target/s390x/translate.c | 25 +++++++++++++++++------- 2 files changed, 38 insertions(+), 27 deletions(-) diff --git a/target/s390x/mem_helper.c b/target/s390x/mem_helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/s390x/mem_helper.c +++ b/target/s390x/mem_helper.c @@ -XXX,XX +XXX,XX @@ void HELPER(cdsg_parallel)(CPUS390XState *env, uint64_t addr, Int128 oldv; bool fail; - if (!HAVE_CMPXCHG128) { - cpu_loop_exit_atomic(ENV_GET_CPU(env), ra); - } + assert(HAVE_CMPXCHG128); mem_idx = cpu_mmu_index(env, false); oi = make_memop_idx(MO_TEQ | MO_ALIGN_16, mem_idx); @@ -XXX,XX +XXX,XX @@ uint64_t HELPER(lpq_parallel)(CPUS390XState *env, uint64_t addr) { uintptr_t ra = GETPC(); uint64_t hi, lo; + int mem_idx; + TCGMemOpIdx oi; + Int128 v; - if (HAVE_ATOMIC128) { - int mem_idx = cpu_mmu_index(env, false); - TCGMemOpIdx oi = make_memop_idx(MO_TEQ | MO_ALIGN_16, mem_idx); - Int128 v = helper_atomic_ldo_be_mmu(env, addr, oi, ra); - hi = int128_gethi(v); - lo = int128_getlo(v); - } else { - cpu_loop_exit_atomic(ENV_GET_CPU(env), ra); - } + assert(HAVE_ATOMIC128); + + mem_idx = cpu_mmu_index(env, false); + oi = make_memop_idx(MO_TEQ | MO_ALIGN_16, mem_idx); + v = helper_atomic_ldo_be_mmu(env, addr, oi, ra); + hi = int128_gethi(v); + lo = int128_getlo(v); env->retxl = lo; return hi; @@ -XXX,XX +XXX,XX @@ void HELPER(stpq_parallel)(CPUS390XState *env, uint64_t addr, uint64_t low, uint64_t high) { uintptr_t ra = GETPC(); + int mem_idx; + TCGMemOpIdx oi; + Int128 v; - if (HAVE_ATOMIC128) { - int mem_idx = cpu_mmu_index(env, false); - TCGMemOpIdx oi = make_memop_idx(MO_TEQ | MO_ALIGN_16, mem_idx); - Int128 v = int128_make128(low, high); - helper_atomic_sto_be_mmu(env, addr, v, oi, ra); - } else { - cpu_loop_exit_atomic(ENV_GET_CPU(env), ra); - } + assert(HAVE_ATOMIC128); + + mem_idx = cpu_mmu_index(env, false); + oi = make_memop_idx(MO_TEQ | MO_ALIGN_16, mem_idx); + v = int128_make128(low, high); + helper_atomic_sto_be_mmu(env, addr, v, oi, ra); } /* Execute instruction. This instruction executes an insn modified with diff --git a/target/s390x/translate.c b/target/s390x/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/s390x/translate.c +++ b/target/s390x/translate.c @@ -XXX,XX +XXX,XX @@ #include "trace-tcg.h" #include "exec/translator.h" #include "exec/log.h" +#include "qemu/atomic128.h" /* Information that (most) every instruction needs to manipulate. */ @@ -XXX,XX +XXX,XX @@ static DisasJumpType op_cdsg(DisasContext *s, DisasOps *o) int r3 = get_field(s->fields, r3); int d2 = get_field(s->fields, d2); int b2 = get_field(s->fields, b2); + DisasJumpType ret = DISAS_NEXT; TCGv_i64 addr; TCGv_i32 t_r1, t_r3; @@ -XXX,XX +XXX,XX @@ static DisasJumpType op_cdsg(DisasContext *s, DisasOps *o) addr = get_address(s, 0, b2, d2); t_r1 = tcg_const_i32(r1); t_r3 = tcg_const_i32(r3); - if (tb_cflags(s->base.tb) & CF_PARALLEL) { + if (!(tb_cflags(s->base.tb) & CF_PARALLEL)) { + gen_helper_cdsg(cpu_env, addr, t_r1, t_r3); + } else if (HAVE_CMPXCHG128) { gen_helper_cdsg_parallel(cpu_env, addr, t_r1, t_r3); } else { - gen_helper_cdsg(cpu_env, addr, t_r1, t_r3); + gen_helper_exit_atomic(cpu_env); + ret = DISAS_NORETURN; } tcg_temp_free_i64(addr); tcg_temp_free_i32(t_r1); tcg_temp_free_i32(t_r3); set_cc_static(s); - return DISAS_NEXT; + return ret; } static DisasJumpType op_csst(DisasContext *s, DisasOps *o) @@ -XXX,XX +XXX,XX @@ static DisasJumpType op_lpd(DisasContext *s, DisasOps *o) static DisasJumpType op_lpq(DisasContext *s, DisasOps *o) { - if (tb_cflags(s->base.tb) & CF_PARALLEL) { + if (!(tb_cflags(s->base.tb) & CF_PARALLEL)) { + gen_helper_lpq(o->out, cpu_env, o->in2); + } else if (HAVE_ATOMIC128) { gen_helper_lpq_parallel(o->out, cpu_env, o->in2); } else { - gen_helper_lpq(o->out, cpu_env, o->in2); + gen_helper_exit_atomic(cpu_env); + return DISAS_NORETURN; } return_low128(o->out2); return DISAS_NEXT; @@ -XXX,XX +XXX,XX @@ static DisasJumpType op_stmh(DisasContext *s, DisasOps *o) static DisasJumpType op_stpq(DisasContext *s, DisasOps *o) { - if (tb_cflags(s->base.tb) & CF_PARALLEL) { + if (!(tb_cflags(s->base.tb) & CF_PARALLEL)) { + gen_helper_stpq(cpu_env, o->in2, o->out2, o->out); + } else if (HAVE_ATOMIC128) { gen_helper_stpq_parallel(cpu_env, o->in2, o->out2, o->out); } else { - gen_helper_stpq(cpu_env, o->in2, o->out2, o->out); + gen_helper_exit_atomic(cpu_env); + return DISAS_NORETURN; } return DISAS_NEXT; } -- 2.17.2 From: "Emilio G. Cota" <cota@braap.org> Updates can come from other threads, so readers that do not take tlb_lock must use atomic_read to avoid undefined behaviour (UB). This completes the conversion to tlb_lock. This conversion results on average in no performance loss, as the following experiments (run on an Intel i7-6700K CPU @ 4.00GHz) show. 1. aarch64 bootup+shutdown test: - Before: Performance counter stats for 'taskset -c 0 ../img/aarch64/die.sh' (10 runs): 7487.087786 task-clock (msec) # 0.998 CPUs utilized ( +- 0.12% ) 31,574,905,303 cycles # 4.217 GHz ( +- 0.12% ) 57,097,908,812 instructions # 1.81 insns per cycle ( +- 0.08% ) 10,255,415,367 branches # 1369.747 M/sec ( +- 0.08% ) 173,278,962 branch-misses # 1.69% of all branches ( +- 0.18% ) 7.504481349 seconds time elapsed ( +- 0.14% ) - After: Performance counter stats for 'taskset -c 0 ../img/aarch64/die.sh' (10 runs): 7462.441328 task-clock (msec) # 0.998 CPUs utilized ( +- 0.07% ) 31,478,476,520 cycles # 4.218 GHz ( +- 0.07% ) 57,017,330,084 instructions # 1.81 insns per cycle ( +- 0.05% ) 10,251,929,667 branches # 1373.804 M/sec ( +- 0.05% ) 173,023,787 branch-misses # 1.69% of all branches ( +- 0.11% ) 7.474970463 seconds time elapsed ( +- 0.07% ) 2. SPEC06int: SPEC06int (test set) [Y axis: Speedup over master] 1.15 +-+----+------+------+------+------+------+-------+------+------+------+------+------+------+----+-+ | | 1.1 +-+.................................+++.............................+ tlb-lock-v2 (m+++x) +-+ | +++ | +++ tlb-lock-v3 (spinl|ck) | | +++ | | +++ +++ | | | 1.05 +-+....+++...........####.........|####.+++.|......|.....###....+++...........+++....###.........+-+ | ### ++#| # |# |# ***### +++### +++#+# | +++ | #|# ### | 1 +-+++***+#++++####+++#++#++++++++++#++#+*+*++#++++#+#+****+#++++###++++###++++###++++#+#++++#+#+++-+ | *+* # #++# *** # #### *** # * *++# ****+# *| * # ****|# |# # #|# #+# # # | 0.95 +-+..*.*.#....#..#.*|*..#...#..#.*|*..#.*.*..#.*|.*.#.*++*.#.*++*+#.****.#....#+#....#.#..++#.#..+-+ | * * # # # *|* # # # *|* # * * # *++* # * * # * * # * |* # ++# # # # *** # | | * * # ++# # *+* # # # *|* # * * # * * # * * # * * # *++* # **** # ++# # * * # | 0.9 +-+..*.*.#...|#..#.*.*..#.++#..#.*|*..#.*.*..#.*..*.#.*..*.#.*..*.#.*..*.#.*.|*.#...|#.#..*.*.#..+-+ | * * # *** # * * # |# # *+* # * * # * * # * * # * * # * * # *++* # |# # * * # | 0.85 +-+..*.*.#..*|*..#.*.*..#.***..#.*.*..#.*.*..#.*..*.#.*..*.#.*..*.#.*..*.#.*..*.#.****.#..*.*.#..+-+ | * * # *+* # * * # *|* # * * # * * # * * # * * # * * # * * # * * # * |* # * * # | | * * # * * # * * # *+* # * * # * * # * * # * * # * * # * * # * * # * |* # * * # | 0.8 +-+..*.*.#..*.*..#.*.*..#.*.*..#.*.*..#.*.*..#.*..*.#.*..*.#.*..*.#.*..*.#.*..*.#.*++*.#..*.*.#..+-+ | * * # * * # * * # * * # * * # * * # * * # * * # * * # * * # * * # * * # * * # | 0.75 +-+--***##--***###-***###-***###-***###-***###-****##-****##-****##-****##-****##-****##--***##--+-+ 400.perlben401.bzip2403.gcc429.m445.gob456.hmme45462.libqua464.h26471.omnet473483.xalancbmkgeomean png: https://imgur.com/a/BHzpPTW Notes: - tlb-lock-v2 corresponds to an implementation with a mutex. - tlb-lock-v3 corresponds to the current implementation, i.e. a spinlock and a single lock acquisition in tlb_set_page_with_attrs. Signed-off-by: Emilio G. Cota <cota@braap.org> Message-Id: <20181016153840.25877-1-cota@braap.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- accel/tcg/softmmu_template.h | 12 ++++++------ include/exec/cpu_ldst.h | 11 ++++++++++- include/exec/cpu_ldst_template.h | 2 +- accel/tcg/cputlb.c | 19 +++++++++++++------ 4 files changed, 30 insertions(+), 14 deletions(-) diff --git a/accel/tcg/softmmu_template.h b/accel/tcg/softmmu_template.h index XXXXXXX..XXXXXXX 100644 --- a/accel/tcg/softmmu_template.h +++ b/accel/tcg/softmmu_template.h @@ -XXX,XX +XXX,XX @@ void helper_le_st_name(CPUArchState *env, target_ulong addr, DATA_TYPE val, uintptr_t mmu_idx = get_mmuidx(oi); uintptr_t index = tlb_index(env, mmu_idx, addr); CPUTLBEntry *entry = tlb_entry(env, mmu_idx, addr); - target_ulong tlb_addr = entry->addr_write; + target_ulong tlb_addr = tlb_addr_write(entry); unsigned a_bits = get_alignment_bits(get_memop(oi)); uintptr_t haddr; @@ -XXX,XX +XXX,XX @@ void helper_le_st_name(CPUArchState *env, target_ulong addr, DATA_TYPE val, tlb_fill(ENV_GET_CPU(env), addr, DATA_SIZE, MMU_DATA_STORE, mmu_idx, retaddr); } - tlb_addr = entry->addr_write & ~TLB_INVALID_MASK; + tlb_addr = tlb_addr_write(entry) & ~TLB_INVALID_MASK; } /* Handle an IO access. */ @@ -XXX,XX +XXX,XX @@ void helper_le_st_name(CPUArchState *env, target_ulong addr, DATA_TYPE val, cannot evict the first. */ page2 = (addr + DATA_SIZE) & TARGET_PAGE_MASK; entry2 = tlb_entry(env, mmu_idx, page2); - if (!tlb_hit_page(entry2->addr_write, page2) + if (!tlb_hit_page(tlb_addr_write(entry2), page2) && !VICTIM_TLB_HIT(addr_write, page2)) { tlb_fill(ENV_GET_CPU(env), page2, DATA_SIZE, MMU_DATA_STORE, mmu_idx, retaddr); @@ -XXX,XX +XXX,XX @@ void helper_be_st_name(CPUArchState *env, target_ulong addr, DATA_TYPE val, uintptr_t mmu_idx = get_mmuidx(oi); uintptr_t index = tlb_index(env, mmu_idx, addr); CPUTLBEntry *entry = tlb_entry(env, mmu_idx, addr); - target_ulong tlb_addr = entry->addr_write; + target_ulong tlb_addr = tlb_addr_write(entry); unsigned a_bits = get_alignment_bits(get_memop(oi)); uintptr_t haddr; @@ -XXX,XX +XXX,XX @@ void helper_be_st_name(CPUArchState *env, target_ulong addr, DATA_TYPE val, tlb_fill(ENV_GET_CPU(env), addr, DATA_SIZE, MMU_DATA_STORE, mmu_idx, retaddr); } - tlb_addr = entry->addr_write & ~TLB_INVALID_MASK; + tlb_addr = tlb_addr_write(entry) & ~TLB_INVALID_MASK; } /* Handle an IO access. */ @@ -XXX,XX +XXX,XX @@ void helper_be_st_name(CPUArchState *env, target_ulong addr, DATA_TYPE val, cannot evict the first. */ page2 = (addr + DATA_SIZE) & TARGET_PAGE_MASK; entry2 = tlb_entry(env, mmu_idx, page2); - if (!tlb_hit_page(entry2->addr_write, page2) + if (!tlb_hit_page(tlb_addr_write(entry2), page2) && !VICTIM_TLB_HIT(addr_write, page2)) { tlb_fill(ENV_GET_CPU(env), page2, DATA_SIZE, MMU_DATA_STORE, mmu_idx, retaddr); 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 @@ extern __thread uintptr_t helper_retaddr; /* The memory helpers for tcg-generated code need tcg_target_long etc. */ #include "tcg.h" +static inline target_ulong tlb_addr_write(const CPUTLBEntry *entry) +{ +#if TCG_OVERSIZED_GUEST + return entry->addr_write; +#else + return atomic_read(&entry->addr_write); +#endif +} + /* Find the TLB index corresponding to the mmu_idx + address pair. */ static inline uintptr_t tlb_index(CPUArchState *env, uintptr_t mmu_idx, target_ulong addr) @@ -XXX,XX +XXX,XX @@ static inline void *tlb_vaddr_to_host(CPUArchState *env, abi_ptr addr, tlb_addr = tlbentry->addr_read; break; case 1: - tlb_addr = tlbentry->addr_write; + tlb_addr = tlb_addr_write(tlbentry); break; case 2: tlb_addr = tlbentry->addr_code; diff --git a/include/exec/cpu_ldst_template.h b/include/exec/cpu_ldst_template.h index XXXXXXX..XXXXXXX 100644 --- a/include/exec/cpu_ldst_template.h +++ b/include/exec/cpu_ldst_template.h @@ -XXX,XX +XXX,XX @@ glue(glue(glue(cpu_st, SUFFIX), MEMSUFFIX), _ra)(CPUArchState *env, addr = ptr; mmu_idx = CPU_MMU_INDEX; entry = tlb_entry(env, mmu_idx, addr); - if (unlikely(entry->addr_write != + if (unlikely(tlb_addr_write(entry) != (addr & (TARGET_PAGE_MASK | (DATA_SIZE - 1))))) { oi = make_memop_idx(SHIFT, mmu_idx); glue(glue(helper_ret_st, SUFFIX), MMUSUFFIX)(env, addr, v, oi, diff --git a/accel/tcg/cputlb.c b/accel/tcg/cputlb.c index XXXXXXX..XXXXXXX 100644 --- a/accel/tcg/cputlb.c +++ b/accel/tcg/cputlb.c @@ -XXX,XX +XXX,XX @@ static inline bool tlb_hit_page_anyprot(CPUTLBEntry *tlb_entry, target_ulong page) { return tlb_hit_page(tlb_entry->addr_read, page) || - tlb_hit_page(tlb_entry->addr_write, page) || + tlb_hit_page(tlb_addr_write(tlb_entry), page) || tlb_hit_page(tlb_entry->addr_code, page); } @@ -XXX,XX +XXX,XX @@ static void io_writex(CPUArchState *env, CPUIOTLBEntry *iotlbentry, tlb_fill(cpu, addr, size, MMU_DATA_STORE, mmu_idx, retaddr); entry = tlb_entry(env, mmu_idx, addr); - tlb_addr = entry->addr_write; + tlb_addr = tlb_addr_write(entry); if (!(tlb_addr & ~(TARGET_PAGE_MASK | TLB_RECHECK))) { /* RAM access */ uintptr_t haddr = addr + entry->addend; @@ -XXX,XX +XXX,XX @@ static bool victim_tlb_hit(CPUArchState *env, size_t mmu_idx, size_t index, assert_cpu_is_self(ENV_GET_CPU(env)); for (vidx = 0; vidx < CPU_VTLB_SIZE; ++vidx) { CPUTLBEntry *vtlb = &env->tlb_v_table[mmu_idx][vidx]; - target_ulong cmp = *(target_ulong *)((uintptr_t)vtlb + elt_ofs); + target_ulong cmp; + + /* elt_ofs might correspond to .addr_write, so use atomic_read */ +#if TCG_OVERSIZED_GUEST + cmp = *(target_ulong *)((uintptr_t)vtlb + elt_ofs); +#else + cmp = atomic_read((target_ulong *)((uintptr_t)vtlb + elt_ofs)); +#endif if (cmp == page) { /* Found entry in victim tlb, swap tlb and iotlb. */ @@ -XXX,XX +XXX,XX @@ void probe_write(CPUArchState *env, target_ulong addr, int size, int mmu_idx, uintptr_t index = tlb_index(env, mmu_idx, addr); CPUTLBEntry *entry = tlb_entry(env, mmu_idx, addr); - if (!tlb_hit(entry->addr_write, addr)) { + if (!tlb_hit(tlb_addr_write(entry), addr)) { /* TLB entry is for a different page */ if (!VICTIM_TLB_HIT(addr_write, addr)) { tlb_fill(ENV_GET_CPU(env), addr, size, MMU_DATA_STORE, @@ -XXX,XX +XXX,XX @@ static void *atomic_mmu_lookup(CPUArchState *env, target_ulong addr, size_t mmu_idx = get_mmuidx(oi); uintptr_t index = tlb_index(env, mmu_idx, addr); CPUTLBEntry *tlbe = tlb_entry(env, mmu_idx, addr); - target_ulong tlb_addr = tlbe->addr_write; + target_ulong tlb_addr = tlb_addr_write(tlbe); TCGMemOp mop = get_memop(oi); int a_bits = get_alignment_bits(mop); int s_bits = mop & MO_SIZE; @@ -XXX,XX +XXX,XX @@ static void *atomic_mmu_lookup(CPUArchState *env, target_ulong addr, tlb_fill(ENV_GET_CPU(env), addr, 1 << s_bits, MMU_DATA_STORE, mmu_idx, retaddr); } - tlb_addr = tlbe->addr_write & ~TLB_INVALID_MASK; + tlb_addr = tlb_addr_write(tlbe) & ~TLB_INVALID_MASK; } /* Notice an IO access or a needs-MMU-lookup access */ -- 2.17.2 With a couple of linux-user and target/sparc patches thrown in for good measure. r~ The following changes since commit 495de0fd82d8bb2d7035f82d9869cfeb48de2f9e: Merge tag 'pull-trivial-patches' of https://gitlab.com/mjt0k/qemu into staging (2025-02-14 08:19:05 -0500) are available in the Git repository at: https://gitlab.com/rth7680/qemu.git tags/pull-tcg-20250215 for you to fetch changes up to 2132751069134114814c7e1609e9cf644f077aad: target/sparc: fake UltraSPARC T1 PCR and PIC registers (2025-02-15 12:04:13 -0800) ---------------------------------------------------------------- tcg: Remove last traces of TCG_TARGET_NEED_POOL_LABELS tcg: Cleanups after disallowing 64-on-32 tcg: Introduce constraint for zero register linux-user: Move TARGET_SA_RESTORER out of generic/signal.h linux-user: Fix alignment when unmapping excess reservation target/sparc: Fix register selection for all F*TOx and FxTO* instructions target/sparc: Fix gdbstub incorrectly handling registers f32-f62 target/sparc: fake UltraSPARC T1 PCR and PIC registers ---------------------------------------------------------------- Andreas Schwab (1): linux-user: Move TARGET_SA_RESTORER out of generic/signal.h Artyom Tarasenko (1): target/sparc: fake UltraSPARC T1 PCR and PIC registers Fabiano Rosas (1): elfload: Fix alignment when unmapping excess reservation Mikael Szreder (2): target/sparc: Fix register selection for all F*TOx and FxTO* instructions target/sparc: Fix gdbstub incorrectly handling registers f32-f62 Richard Henderson (19): tcg: Remove last traces of TCG_TARGET_NEED_POOL_LABELS tcg: Remove TCG_OVERSIZED_GUEST tcg: Drop support for two address registers in gen_ldst tcg: Merge INDEX_op_qemu_*_{a32,a64}_* tcg/arm: Drop addrhi from prepare_host_addr tcg/i386: Drop addrhi from prepare_host_addr tcg/mips: Drop addrhi from prepare_host_addr tcg/ppc: Drop addrhi from prepare_host_addr tcg: Replace addr{lo,hi}_reg with addr_reg in TCGLabelQemuLdst plugins: Fix qemu_plugin_read_memory_vaddr parameters accel/tcg: Fix tlb_set_page_with_attrs, tlb_set_page include/exec: Change vaddr to uintptr_t include/exec: Use uintptr_t in CPUTLBEntry tcg: Introduce the 'z' constraint for a hardware zero register tcg/aarch64: Use 'z' constraint tcg/loongarch64: Use 'z' constraint tcg/mips: Use 'z' constraint tcg/riscv: Use 'z' constraint tcg/sparc64: Use 'z' constraint include/exec/tlb-common.h | 10 +- include/exec/vaddr.h | 16 ++-- include/qemu/atomic.h | 18 +--- include/tcg/oversized-guest.h | 23 ----- include/tcg/tcg-opc.h | 28 ++---- include/tcg/tcg.h | 3 +- linux-user/aarch64/target_signal.h | 2 + linux-user/arm/target_signal.h | 2 + linux-user/generic/signal.h | 1 - linux-user/i386/target_signal.h | 2 + linux-user/m68k/target_signal.h | 1 + linux-user/microblaze/target_signal.h | 2 + linux-user/ppc/target_signal.h | 2 + linux-user/s390x/target_signal.h | 2 + linux-user/sh4/target_signal.h | 2 + linux-user/x86_64/target_signal.h | 2 + linux-user/xtensa/target_signal.h | 2 + tcg/aarch64/tcg-target-con-set.h | 12 +-- tcg/aarch64/tcg-target.h | 2 + tcg/loongarch64/tcg-target-con-set.h | 15 ++- tcg/loongarch64/tcg-target-con-str.h | 1 - tcg/loongarch64/tcg-target.h | 2 + tcg/mips/tcg-target-con-set.h | 26 +++--- tcg/mips/tcg-target-con-str.h | 1 - tcg/mips/tcg-target.h | 2 + tcg/riscv/tcg-target-con-set.h | 10 +- tcg/riscv/tcg-target-con-str.h | 1 - tcg/riscv/tcg-target.h | 2 + tcg/sparc64/tcg-target-con-set.h | 12 +-- tcg/sparc64/tcg-target-con-str.h | 1 - tcg/sparc64/tcg-target.h | 3 +- tcg/tci/tcg-target.h | 1 - accel/tcg/cputlb.c | 32 ++----- accel/tcg/tcg-all.c | 9 +- linux-user/elfload.c | 4 +- plugins/api.c | 2 +- target/arm/ptw.c | 34 ------- target/riscv/cpu_helper.c | 13 +-- target/sparc/gdbstub.c | 18 +++- target/sparc/translate.c | 19 ++++ tcg/optimize.c | 21 ++--- tcg/tcg-op-ldst.c | 103 +++++---------------- tcg/tcg.c | 97 +++++++++---------- tcg/tci.c | 119 +++++------------------- docs/devel/multi-thread-tcg.rst | 1 - docs/devel/tcg-ops.rst | 4 +- target/sparc/insns.decode | 19 ++-- tcg/aarch64/tcg-target.c.inc | 86 +++++++---------- tcg/arm/tcg-target.c.inc | 104 ++++++--------------- tcg/i386/tcg-target.c.inc | 125 +++++++------------------ tcg/loongarch64/tcg-target.c.inc | 72 ++++++--------- tcg/mips/tcg-target.c.inc | 169 +++++++++++----------------------- tcg/ppc/tcg-target.c.inc | 164 ++++++++------------------------- tcg/riscv/tcg-target.c.inc | 56 +++++------ tcg/s390x/tcg-target.c.inc | 40 +++----- tcg/sparc64/tcg-target.c.inc | 45 ++++----- tcg/tci/tcg-target.c.inc | 60 +++--------- 57 files changed, 536 insertions(+), 1089 deletions(-) delete mode 100644 include/tcg/oversized-guest.h These should have been removed with the rest. There are a couple of hosts which can emit guest_base into the constant pool: aarch64, mips64, ppc64, riscv64. Fixes: a417ef835058 ("tcg: Remove TCG_TARGET_NEED_LDST_LABELS and TCG_TARGET_NEED_POOL_LABELS") Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- tcg/tci/tcg-target.h | 1 - tcg/tcg.c | 4 ---- 2 files changed, 5 deletions(-) diff --git a/tcg/tci/tcg-target.h b/tcg/tci/tcg-target.h index XXXXXXX..XXXXXXX 100644 --- a/tcg/tci/tcg-target.h +++ b/tcg/tci/tcg-target.h @@ -XXX,XX +XXX,XX @@ typedef enum { } TCGReg; #define HAVE_TCG_QEMU_TB_EXEC -#define TCG_TARGET_NEED_POOL_LABELS #endif /* TCG_TARGET_H */ 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 @@ void tcg_prologue_init(void) tcg_qemu_tb_exec = (tcg_prologue_fn *)tcg_splitwx_to_rx(s->code_ptr); #endif -#ifdef TCG_TARGET_NEED_POOL_LABELS s->pool_labels = NULL; -#endif qemu_thread_jit_write(); /* Generate the prologue. */ tcg_target_qemu_prologue(s); -#ifdef TCG_TARGET_NEED_POOL_LABELS /* Allow the prologue to put e.g. guest_base into a pool entry. */ { int result = tcg_out_pool_finalize(s); tcg_debug_assert(result == 0); } -#endif prologue_size = tcg_current_code_size(s); perf_report_prologue(s->code_gen_ptr, prologue_size); -- 2.43.0 This is now prohibited in configuration. Reviewed-by: Alex Bennée <alex.bennee@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- include/qemu/atomic.h | 18 +++-------------- include/tcg/oversized-guest.h | 23 ---------------------- accel/tcg/cputlb.c | 7 ------- accel/tcg/tcg-all.c | 9 ++++----- target/arm/ptw.c | 34 --------------------------------- target/riscv/cpu_helper.c | 13 +------------ docs/devel/multi-thread-tcg.rst | 1 - 7 files changed, 8 insertions(+), 97 deletions(-) delete mode 100644 include/tcg/oversized-guest.h diff --git a/include/qemu/atomic.h b/include/qemu/atomic.h index XXXXXXX..XXXXXXX 100644 --- a/include/qemu/atomic.h +++ b/include/qemu/atomic.h @@ -XXX,XX +XXX,XX @@ */ #define signal_barrier() __atomic_signal_fence(__ATOMIC_SEQ_CST) -/* Sanity check that the size of an atomic operation isn't "overly large". +/* + * Sanity check that the size of an atomic operation isn't "overly large". * Despite the fact that e.g. i686 has 64-bit atomic operations, we do not * want to use them because we ought not need them, and this lets us do a * bit of sanity checking that other 32-bit hosts might build. - * - * That said, we have a problem on 64-bit ILP32 hosts in that in order to - * sync with TCG_OVERSIZED_GUEST, this must match TCG_TARGET_REG_BITS. - * We'd prefer not want to pull in everything else TCG related, so handle - * those few cases by hand. - * - * Note that x32 is fully detected with __x86_64__ + _ILP32, and that for - * Sparc we always force the use of sparcv9 in configure. MIPS n32 (ILP32) & - * n64 (LP64) ABIs are both detected using __mips64. */ -#if defined(__x86_64__) || defined(__sparc__) || defined(__mips64) -# define ATOMIC_REG_SIZE 8 -#else -# define ATOMIC_REG_SIZE sizeof(void *) -#endif +#define ATOMIC_REG_SIZE sizeof(void *) /* Weak atomic operations prevent the compiler moving other * loads/stores past the atomic operation load/store. However there is diff --git a/include/tcg/oversized-guest.h b/include/tcg/oversized-guest.h deleted file mode 100644 index XXXXXXX..XXXXXXX --- a/include/tcg/oversized-guest.h +++ /dev/null @@ -XXX,XX +XXX,XX @@ -/* SPDX-License-Identifier: MIT */ -/* - * Define TCG_OVERSIZED_GUEST - * Copyright (c) 2008 Fabrice Bellard - */ - -#ifndef EXEC_TCG_OVERSIZED_GUEST_H -#define EXEC_TCG_OVERSIZED_GUEST_H - -#include "tcg-target-reg-bits.h" -#include "cpu-param.h" - -/* - * Oversized TCG guests make things like MTTCG hard - * as we can't use atomics for cputlb updates. - */ -#if TARGET_LONG_BITS > TCG_TARGET_REG_BITS -#define TCG_OVERSIZED_GUEST 1 -#else -#define TCG_OVERSIZED_GUEST 0 -#endif - -#endif diff --git a/accel/tcg/cputlb.c b/accel/tcg/cputlb.c index XXXXXXX..XXXXXXX 100644 --- a/accel/tcg/cputlb.c +++ b/accel/tcg/cputlb.c @@ -XXX,XX +XXX,XX @@ #include "qemu/plugin-memory.h" #endif #include "tcg/tcg-ldst.h" -#include "tcg/oversized-guest.h" /* DEBUG defines, enable DEBUG_TLB_LOG to log to the CPU_LOG_MMU target */ /* #define DEBUG_TLB */ @@ -XXX,XX +XXX,XX @@ static inline uint64_t tlb_read_idx(const CPUTLBEntry *entry, return qatomic_read(ptr); #else const uint64_t *ptr = &entry->addr_idx[access_type]; -# if TCG_OVERSIZED_GUEST - return *ptr; -# else /* ofs might correspond to .addr_write, so use qatomic_read */ return qatomic_read(ptr); -# endif #endif } @@ -XXX,XX +XXX,XX @@ static void tlb_reset_dirty_range_locked(CPUTLBEntry *tlb_entry, uint32_t *ptr_write = (uint32_t *)&tlb_entry->addr_write; ptr_write += HOST_BIG_ENDIAN; qatomic_set(ptr_write, *ptr_write | TLB_NOTDIRTY); -#elif TCG_OVERSIZED_GUEST - tlb_entry->addr_write |= TLB_NOTDIRTY; #else qatomic_set(&tlb_entry->addr_write, tlb_entry->addr_write | TLB_NOTDIRTY); diff --git a/accel/tcg/tcg-all.c b/accel/tcg/tcg-all.c index XXXXXXX..XXXXXXX 100644 --- a/accel/tcg/tcg-all.c +++ b/accel/tcg/tcg-all.c @@ -XXX,XX +XXX,XX @@ #include "exec/replay-core.h" #include "system/cpu-timers.h" #include "tcg/startup.h" -#include "tcg/oversized-guest.h" #include "qapi/error.h" #include "qemu/error-report.h" #include "qemu/accel.h" @@ -XXX,XX +XXX,XX @@ #include "hw/boards.h" #endif #include "internal-common.h" +#include "cpu-param.h" + struct TCGState { AccelState parent_obj; @@ -XXX,XX +XXX,XX @@ DECLARE_INSTANCE_CHECKER(TCGState, TCG_STATE, static bool default_mttcg_enabled(void) { - if (icount_enabled() || TCG_OVERSIZED_GUEST) { + if (icount_enabled()) { return false; } #ifdef TARGET_SUPPORTS_MTTCG @@ -XXX,XX +XXX,XX @@ static void tcg_set_thread(Object *obj, const char *value, Error **errp) TCGState *s = TCG_STATE(obj); if (strcmp(value, "multi") == 0) { - if (TCG_OVERSIZED_GUEST) { - error_setg(errp, "No MTTCG when guest word size > hosts"); - } else if (icount_enabled()) { + if (icount_enabled()) { error_setg(errp, "No MTTCG when icount is enabled"); } else { #ifndef TARGET_SUPPORTS_MTTCG diff --git a/target/arm/ptw.c b/target/arm/ptw.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/ptw.c +++ b/target/arm/ptw.c @@ -XXX,XX +XXX,XX @@ #include "internals.h" #include "cpu-features.h" #include "idau.h" -#ifdef CONFIG_TCG -# include "tcg/oversized-guest.h" -#endif typedef struct S1Translate { /* @@ -XXX,XX +XXX,XX @@ static uint64_t arm_casq_ptw(CPUARMState *env, uint64_t old_val, ptw->out_rw = true; } -#ifdef CONFIG_ATOMIC64 if (ptw->out_be) { old_val = cpu_to_be64(old_val); new_val = cpu_to_be64(new_val); @@ -XXX,XX +XXX,XX @@ static uint64_t arm_casq_ptw(CPUARMState *env, uint64_t old_val, cur_val = qatomic_cmpxchg__nocheck((uint64_t *)host, old_val, new_val); cur_val = le64_to_cpu(cur_val); } -#else - /* - * We can't support the full 64-bit atomic cmpxchg on the host. - * Because this is only used for FEAT_HAFDBS, which is only for AA64, - * we know that TCG_OVERSIZED_GUEST is set, which means that we are - * running in round-robin mode and could only race with dma i/o. - */ -#if !TCG_OVERSIZED_GUEST -# error "Unexpected configuration" -#endif - bool locked = bql_locked(); - if (!locked) { - bql_lock(); - } - if (ptw->out_be) { - cur_val = ldq_be_p(host); - if (cur_val == old_val) { - stq_be_p(host, new_val); - } - } else { - cur_val = ldq_le_p(host); - if (cur_val == old_val) { - stq_le_p(host, new_val); - } - } - if (!locked) { - bql_unlock(); - } -#endif - return cur_val; #else /* AArch32 does not have FEAT_HADFS; non-TCG guests only use debug-mode. */ diff --git a/target/riscv/cpu_helper.c b/target/riscv/cpu_helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/riscv/cpu_helper.c +++ b/target/riscv/cpu_helper.c @@ -XXX,XX +XXX,XX @@ #include "system/cpu-timers.h" #include "cpu_bits.h" #include "debug.h" -#include "tcg/oversized-guest.h" #include "pmp.h" int riscv_env_mmu_index(CPURISCVState *env, bool ifetch) @@ -XXX,XX +XXX,XX @@ static int get_physical_address(CPURISCVState *env, hwaddr *physical, hwaddr pte_addr; int i; -#if !TCG_OVERSIZED_GUEST -restart: -#endif + restart: for (i = 0; i < levels; i++, ptshift -= ptidxbits) { target_ulong idx; if (i == 0) { @@ -XXX,XX +XXX,XX @@ restart: false, MEMTXATTRS_UNSPECIFIED); if (memory_region_is_ram(mr)) { target_ulong *pte_pa = qemu_map_ram_ptr(mr->ram_block, addr1); -#if TCG_OVERSIZED_GUEST - /* - * MTTCG is not enabled on oversized TCG guests so - * page table updates do not need to be atomic - */ - *pte_pa = pte = updated_pte; -#else target_ulong old_pte; if (riscv_cpu_sxl(env) == MXL_RV32) { old_pte = qatomic_cmpxchg((uint32_t *)pte_pa, pte, updated_pte); @@ -XXX,XX +XXX,XX @@ restart: goto restart; } pte = updated_pte; -#endif } else { /* * Misconfigured PTE in ROM (AD bits are not preset) or diff --git a/docs/devel/multi-thread-tcg.rst b/docs/devel/multi-thread-tcg.rst index XXXXXXX..XXXXXXX 100644 --- a/docs/devel/multi-thread-tcg.rst +++ b/docs/devel/multi-thread-tcg.rst @@ -XXX,XX +XXX,XX @@ if: * forced by --accel tcg,thread=single * enabling --icount mode -* 64 bit guests on 32 bit hosts (TCG_OVERSIZED_GUEST) In the general case of running translated code there should be no inter-vCPU dependencies and all vCPUs should be able to run at full -- 2.43.0 Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- tcg/tcg-op-ldst.c | 21 +++------------------ tcg/tcg.c | 4 +--- 2 files changed, 4 insertions(+), 21 deletions(-) diff --git a/tcg/tcg-op-ldst.c b/tcg/tcg-op-ldst.c index XXXXXXX..XXXXXXX 100644 --- a/tcg/tcg-op-ldst.c +++ b/tcg/tcg-op-ldst.c @@ -XXX,XX +XXX,XX @@ static MemOp tcg_canonicalize_memop(MemOp op, bool is64, bool st) static void gen_ldst(TCGOpcode opc, TCGType type, TCGTemp *vl, TCGTemp *vh, TCGTemp *addr, MemOpIdx oi) { - if (TCG_TARGET_REG_BITS == 64 || tcg_ctx->addr_type == TCG_TYPE_I32) { - if (vh) { - tcg_gen_op4(opc, type, temp_arg(vl), temp_arg(vh), - temp_arg(addr), oi); - } else { - tcg_gen_op3(opc, type, temp_arg(vl), temp_arg(addr), oi); - } + if (vh) { + tcg_gen_op4(opc, type, temp_arg(vl), temp_arg(vh), temp_arg(addr), oi); } else { - /* See TCGV_LOW/HIGH. */ - TCGTemp *al = addr + HOST_BIG_ENDIAN; - TCGTemp *ah = addr + !HOST_BIG_ENDIAN; - - if (vh) { - tcg_gen_op5(opc, type, temp_arg(vl), temp_arg(vh), - temp_arg(al), temp_arg(ah), oi); - } else { - tcg_gen_op4(opc, type, temp_arg(vl), - temp_arg(al), temp_arg(ah), oi); - } + tcg_gen_op3(opc, type, temp_arg(vl), temp_arg(addr), oi); } } 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 @@ void tcg_func_start(TCGContext *s) s->emit_before_op = NULL; QSIMPLEQ_INIT(&s->labels); - tcg_debug_assert(s->addr_type == TCG_TYPE_I32 || - s->addr_type == TCG_TYPE_I64); - + tcg_debug_assert(s->addr_type <= TCG_TYPE_REG); tcg_debug_assert(s->insn_start_words > 0); } -- 2.43.0 Since 64-on-32 is now unsupported, guest addresses always fit in one host register. Drop the replication of opcodes. Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- include/tcg/tcg-opc.h | 28 ++------ tcg/optimize.c | 21 ++---- tcg/tcg-op-ldst.c | 82 +++++---------------- tcg/tcg.c | 42 ++++------- tcg/tci.c | 119 ++++++------------------------- tcg/aarch64/tcg-target.c.inc | 36 ++++------ tcg/arm/tcg-target.c.inc | 40 +++-------- tcg/i386/tcg-target.c.inc | 69 ++++-------------- tcg/loongarch64/tcg-target.c.inc | 36 ++++------ tcg/mips/tcg-target.c.inc | 51 +++---------- tcg/ppc/tcg-target.c.inc | 68 ++++-------------- tcg/riscv/tcg-target.c.inc | 24 +++---- tcg/s390x/tcg-target.c.inc | 36 ++++------ tcg/sparc64/tcg-target.c.inc | 24 +++---- tcg/tci/tcg-target.c.inc | 60 ++++------------ 15 files changed, 177 insertions(+), 559 deletions(-) diff --git a/include/tcg/tcg-opc.h b/include/tcg/tcg-opc.h index XXXXXXX..XXXXXXX 100644 --- a/include/tcg/tcg-opc.h +++ b/include/tcg/tcg-opc.h @@ -XXX,XX +XXX,XX @@ DEF(goto_ptr, 0, 1, 0, TCG_OPF_BB_EXIT | TCG_OPF_BB_END) DEF(plugin_cb, 0, 0, 1, TCG_OPF_NOT_PRESENT) DEF(plugin_mem_cb, 0, 1, 1, TCG_OPF_NOT_PRESENT) -/* Replicate ld/st ops for 32 and 64-bit guest addresses. */ -DEF(qemu_ld_a32_i32, 1, 1, 1, +DEF(qemu_ld_i32, 1, 1, 1, TCG_OPF_CALL_CLOBBER | TCG_OPF_SIDE_EFFECTS) -DEF(qemu_st_a32_i32, 0, 1 + 1, 1, +DEF(qemu_st_i32, 0, 1 + 1, 1, TCG_OPF_CALL_CLOBBER | TCG_OPF_SIDE_EFFECTS) -DEF(qemu_ld_a32_i64, DATA64_ARGS, 1, 1, +DEF(qemu_ld_i64, DATA64_ARGS, 1, 1, TCG_OPF_CALL_CLOBBER | TCG_OPF_SIDE_EFFECTS) -DEF(qemu_st_a32_i64, 0, DATA64_ARGS + 1, 1, - TCG_OPF_CALL_CLOBBER | TCG_OPF_SIDE_EFFECTS) - -DEF(qemu_ld_a64_i32, 1, DATA64_ARGS, 1, - TCG_OPF_CALL_CLOBBER | TCG_OPF_SIDE_EFFECTS) -DEF(qemu_st_a64_i32, 0, 1 + DATA64_ARGS, 1, - TCG_OPF_CALL_CLOBBER | TCG_OPF_SIDE_EFFECTS) -DEF(qemu_ld_a64_i64, DATA64_ARGS, DATA64_ARGS, 1, - TCG_OPF_CALL_CLOBBER | TCG_OPF_SIDE_EFFECTS) -DEF(qemu_st_a64_i64, 0, DATA64_ARGS + DATA64_ARGS, 1, +DEF(qemu_st_i64, 0, DATA64_ARGS + 1, 1, TCG_OPF_CALL_CLOBBER | TCG_OPF_SIDE_EFFECTS) /* Only used by i386 to cope with stupid register constraints. */ -DEF(qemu_st8_a32_i32, 0, 1 + 1, 1, - TCG_OPF_CALL_CLOBBER | TCG_OPF_SIDE_EFFECTS) -DEF(qemu_st8_a64_i32, 0, 1 + DATA64_ARGS, 1, +DEF(qemu_st8_i32, 0, 1 + 1, 1, TCG_OPF_CALL_CLOBBER | TCG_OPF_SIDE_EFFECTS) /* Only for 64-bit hosts at the moment. */ -DEF(qemu_ld_a32_i128, 2, 1, 1, TCG_OPF_CALL_CLOBBER | TCG_OPF_SIDE_EFFECTS) -DEF(qemu_ld_a64_i128, 2, 1, 1, TCG_OPF_CALL_CLOBBER | TCG_OPF_SIDE_EFFECTS) -DEF(qemu_st_a32_i128, 0, 3, 1, TCG_OPF_CALL_CLOBBER | TCG_OPF_SIDE_EFFECTS) -DEF(qemu_st_a64_i128, 0, 3, 1, TCG_OPF_CALL_CLOBBER | TCG_OPF_SIDE_EFFECTS) +DEF(qemu_ld_i128, 2, 1, 1, TCG_OPF_CALL_CLOBBER | TCG_OPF_SIDE_EFFECTS) +DEF(qemu_st_i128, 0, 3, 1, TCG_OPF_CALL_CLOBBER | TCG_OPF_SIDE_EFFECTS) /* Host vector support. */ 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) CASE_OP_32_64_VEC(orc): done = fold_orc(&ctx, op); break; - case INDEX_op_qemu_ld_a32_i32: - case INDEX_op_qemu_ld_a64_i32: + case INDEX_op_qemu_ld_i32: done = fold_qemu_ld_1reg(&ctx, op); break; - case INDEX_op_qemu_ld_a32_i64: - case INDEX_op_qemu_ld_a64_i64: + case INDEX_op_qemu_ld_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: + case INDEX_op_qemu_ld_i128: done = fold_qemu_ld_2reg(&ctx, op); break; - case INDEX_op_qemu_st8_a32_i32: - case INDEX_op_qemu_st8_a64_i32: - case INDEX_op_qemu_st_a32_i32: - case INDEX_op_qemu_st_a64_i32: - case INDEX_op_qemu_st_a32_i64: - case INDEX_op_qemu_st_a64_i64: - case INDEX_op_qemu_st_a32_i128: - case INDEX_op_qemu_st_a64_i128: + case INDEX_op_qemu_st8_i32: + case INDEX_op_qemu_st_i32: + case INDEX_op_qemu_st_i64: + case INDEX_op_qemu_st_i128: done = fold_qemu_st(&ctx, op); break; CASE_OP_32_64(rem): diff --git a/tcg/tcg-op-ldst.c b/tcg/tcg-op-ldst.c index XXXXXXX..XXXXXXX 100644 --- a/tcg/tcg-op-ldst.c +++ b/tcg/tcg-op-ldst.c @@ -XXX,XX +XXX,XX @@ static void tcg_gen_qemu_ld_i32_int(TCGv_i32 val, TCGTemp *addr, MemOp orig_memop; MemOpIdx orig_oi, oi; TCGv_i64 copy_addr; - TCGOpcode opc; tcg_gen_req_mo(TCG_MO_LD_LD | TCG_MO_ST_LD); orig_memop = memop = tcg_canonicalize_memop(memop, 0, 0); @@ -XXX,XX +XXX,XX @@ static void tcg_gen_qemu_ld_i32_int(TCGv_i32 val, TCGTemp *addr, } copy_addr = plugin_maybe_preserve_addr(addr); - if (tcg_ctx->addr_type == TCG_TYPE_I32) { - opc = INDEX_op_qemu_ld_a32_i32; - } else { - opc = INDEX_op_qemu_ld_a64_i32; - } - gen_ldst(opc, TCG_TYPE_I32, tcgv_i32_temp(val), NULL, addr, oi); + gen_ldst(INDEX_op_qemu_ld_i32, TCG_TYPE_I32, + tcgv_i32_temp(val), NULL, addr, oi); plugin_gen_mem_callbacks_i32(val, copy_addr, addr, orig_oi, QEMU_PLUGIN_MEM_R); @@ -XXX,XX +XXX,XX @@ static void tcg_gen_qemu_st_i32_int(TCGv_i32 val, TCGTemp *addr, } if (TCG_TARGET_HAS_qemu_st8_i32 && (memop & MO_SIZE) == MO_8) { - if (tcg_ctx->addr_type == TCG_TYPE_I32) { - opc = INDEX_op_qemu_st8_a32_i32; - } else { - opc = INDEX_op_qemu_st8_a64_i32; - } + opc = INDEX_op_qemu_st8_i32; } else { - if (tcg_ctx->addr_type == TCG_TYPE_I32) { - opc = INDEX_op_qemu_st_a32_i32; - } else { - opc = INDEX_op_qemu_st_a64_i32; - } + opc = INDEX_op_qemu_st_i32; } gen_ldst(opc, TCG_TYPE_I32, tcgv_i32_temp(val), NULL, addr, oi); plugin_gen_mem_callbacks_i32(val, NULL, addr, orig_oi, QEMU_PLUGIN_MEM_W); @@ -XXX,XX +XXX,XX @@ static void tcg_gen_qemu_ld_i64_int(TCGv_i64 val, TCGTemp *addr, MemOp orig_memop; MemOpIdx orig_oi, oi; TCGv_i64 copy_addr; - TCGOpcode opc; if (TCG_TARGET_REG_BITS == 32 && (memop & MO_SIZE) < MO_64) { tcg_gen_qemu_ld_i32_int(TCGV_LOW(val), addr, idx, memop); @@ -XXX,XX +XXX,XX @@ static void tcg_gen_qemu_ld_i64_int(TCGv_i64 val, TCGTemp *addr, } copy_addr = plugin_maybe_preserve_addr(addr); - if (tcg_ctx->addr_type == TCG_TYPE_I32) { - opc = INDEX_op_qemu_ld_a32_i64; - } else { - opc = INDEX_op_qemu_ld_a64_i64; - } - gen_ldst_i64(opc, val, addr, oi); + gen_ldst_i64(INDEX_op_qemu_ld_i64, val, addr, oi); plugin_gen_mem_callbacks_i64(val, copy_addr, addr, orig_oi, QEMU_PLUGIN_MEM_R); @@ -XXX,XX +XXX,XX @@ static void tcg_gen_qemu_st_i64_int(TCGv_i64 val, TCGTemp *addr, { TCGv_i64 swap = NULL; MemOpIdx orig_oi, oi; - TCGOpcode opc; if (TCG_TARGET_REG_BITS == 32 && (memop & MO_SIZE) < MO_64) { tcg_gen_qemu_st_i32_int(TCGV_LOW(val), addr, idx, memop); @@ -XXX,XX +XXX,XX @@ static void tcg_gen_qemu_st_i64_int(TCGv_i64 val, TCGTemp *addr, oi = make_memop_idx(memop, idx); } - if (tcg_ctx->addr_type == TCG_TYPE_I32) { - opc = INDEX_op_qemu_st_a32_i64; - } else { - opc = INDEX_op_qemu_st_a64_i64; - } - gen_ldst_i64(opc, val, addr, oi); + gen_ldst_i64(INDEX_op_qemu_st_i64, val, addr, oi); plugin_gen_mem_callbacks_i64(val, NULL, addr, orig_oi, QEMU_PLUGIN_MEM_W); if (swap) { @@ -XXX,XX +XXX,XX @@ static void tcg_gen_qemu_ld_i128_int(TCGv_i128 val, TCGTemp *addr, { MemOpIdx orig_oi; TCGv_i64 ext_addr = NULL; - TCGOpcode opc; check_max_alignment(memop_alignment_bits(memop)); tcg_gen_req_mo(TCG_MO_LD_LD | TCG_MO_ST_LD); @@ -XXX,XX +XXX,XX @@ static void tcg_gen_qemu_ld_i128_int(TCGv_i128 val, TCGTemp *addr, hi = TCGV128_HIGH(val); } - if (tcg_ctx->addr_type == TCG_TYPE_I32) { - opc = INDEX_op_qemu_ld_a32_i128; - } else { - opc = INDEX_op_qemu_ld_a64_i128; - } - gen_ldst(opc, TCG_TYPE_I128, tcgv_i64_temp(lo), + gen_ldst(INDEX_op_qemu_ld_i128, TCG_TYPE_I128, tcgv_i64_temp(lo), tcgv_i64_temp(hi), addr, oi); if (need_bswap) { @@ -XXX,XX +XXX,XX @@ static void tcg_gen_qemu_ld_i128_int(TCGv_i128 val, TCGTemp *addr, canonicalize_memop_i128_as_i64(mop, memop); need_bswap = (mop[0] ^ memop) & MO_BSWAP; - if (tcg_ctx->addr_type == TCG_TYPE_I32) { - opc = INDEX_op_qemu_ld_a32_i64; - } else { - opc = INDEX_op_qemu_ld_a64_i64; - } - /* * Since there are no global TCGv_i128, there is no visible state * changed if the second load faults. Load directly into the two @@ -XXX,XX +XXX,XX @@ static void tcg_gen_qemu_ld_i128_int(TCGv_i128 val, TCGTemp *addr, y = TCGV128_LOW(val); } - gen_ldst_i64(opc, x, addr, make_memop_idx(mop[0], idx)); + gen_ldst_i64(INDEX_op_qemu_ld_i64, x, addr, + make_memop_idx(mop[0], idx)); if (need_bswap) { tcg_gen_bswap64_i64(x, x); @@ -XXX,XX +XXX,XX @@ static void tcg_gen_qemu_ld_i128_int(TCGv_i128 val, TCGTemp *addr, addr_p8 = tcgv_i64_temp(t); } - gen_ldst_i64(opc, y, addr_p8, make_memop_idx(mop[1], idx)); + gen_ldst_i64(INDEX_op_qemu_ld_i64, y, addr_p8, + make_memop_idx(mop[1], idx)); tcg_temp_free_internal(addr_p8); if (need_bswap) { @@ -XXX,XX +XXX,XX @@ static void tcg_gen_qemu_st_i128_int(TCGv_i128 val, TCGTemp *addr, { MemOpIdx orig_oi; TCGv_i64 ext_addr = NULL; - TCGOpcode opc; check_max_alignment(memop_alignment_bits(memop)); tcg_gen_req_mo(TCG_MO_ST_LD | TCG_MO_ST_ST); @@ -XXX,XX +XXX,XX @@ static void tcg_gen_qemu_st_i128_int(TCGv_i128 val, TCGTemp *addr, hi = TCGV128_HIGH(val); } - if (tcg_ctx->addr_type == TCG_TYPE_I32) { - opc = INDEX_op_qemu_st_a32_i128; - } else { - opc = INDEX_op_qemu_st_a64_i128; - } - gen_ldst(opc, TCG_TYPE_I128, tcgv_i64_temp(lo), - tcgv_i64_temp(hi), addr, oi); + gen_ldst(INDEX_op_qemu_st_i128, TCG_TYPE_I128, + tcgv_i64_temp(lo), tcgv_i64_temp(hi), addr, oi); if (need_bswap) { tcg_temp_free_i64(lo); @@ -XXX,XX +XXX,XX @@ static void tcg_gen_qemu_st_i128_int(TCGv_i128 val, TCGTemp *addr, canonicalize_memop_i128_as_i64(mop, memop); - if (tcg_ctx->addr_type == TCG_TYPE_I32) { - opc = INDEX_op_qemu_st_a32_i64; - } else { - opc = INDEX_op_qemu_st_a64_i64; - } - if ((memop & MO_BSWAP) == MO_LE) { x = TCGV128_LOW(val); y = TCGV128_HIGH(val); @@ -XXX,XX +XXX,XX @@ static void tcg_gen_qemu_st_i128_int(TCGv_i128 val, TCGTemp *addr, x = b; } - gen_ldst_i64(opc, x, addr, make_memop_idx(mop[0], idx)); + gen_ldst_i64(INDEX_op_qemu_st_i64, x, addr, + make_memop_idx(mop[0], idx)); if (tcg_ctx->addr_type == TCG_TYPE_I32) { TCGv_i32 t = tcg_temp_ebb_new_i32(); @@ -XXX,XX +XXX,XX @@ static void tcg_gen_qemu_st_i128_int(TCGv_i128 val, TCGTemp *addr, if (b) { tcg_gen_bswap64_i64(b, y); - gen_ldst_i64(opc, b, addr_p8, make_memop_idx(mop[1], idx)); + gen_ldst_i64(INDEX_op_qemu_st_i64, b, addr_p8, + make_memop_idx(mop[1], idx)); tcg_temp_free_i64(b); } else { - gen_ldst_i64(opc, y, addr_p8, make_memop_idx(mop[1], idx)); + gen_ldst_i64(INDEX_op_qemu_st_i64, y, addr_p8, + make_memop_idx(mop[1], idx)); } tcg_temp_free_internal(addr_p8); } else { 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 @@ bool tcg_op_supported(TCGOpcode op, TCGType type, unsigned flags) case INDEX_op_exit_tb: case INDEX_op_goto_tb: case INDEX_op_goto_ptr: - case INDEX_op_qemu_ld_a32_i32: - case INDEX_op_qemu_ld_a64_i32: - case INDEX_op_qemu_st_a32_i32: - case INDEX_op_qemu_st_a64_i32: - case INDEX_op_qemu_ld_a32_i64: - case INDEX_op_qemu_ld_a64_i64: - case INDEX_op_qemu_st_a32_i64: - case INDEX_op_qemu_st_a64_i64: + case INDEX_op_qemu_ld_i32: + case INDEX_op_qemu_st_i32: + case INDEX_op_qemu_ld_i64: + case INDEX_op_qemu_st_i64: return true; - case INDEX_op_qemu_st8_a32_i32: - case INDEX_op_qemu_st8_a64_i32: + case INDEX_op_qemu_st8_i32: return TCG_TARGET_HAS_qemu_st8_i32; - case INDEX_op_qemu_ld_a32_i128: - case INDEX_op_qemu_ld_a64_i128: - case INDEX_op_qemu_st_a32_i128: - case INDEX_op_qemu_st_a64_i128: + case INDEX_op_qemu_ld_i128: + case INDEX_op_qemu_st_i128: return TCG_TARGET_HAS_qemu_ldst_i128; case INDEX_op_mov_i32: @@ -XXX,XX +XXX,XX @@ void tcg_dump_ops(TCGContext *s, FILE *f, bool have_prefs) } i = 1; break; - case INDEX_op_qemu_ld_a32_i32: - case INDEX_op_qemu_ld_a64_i32: - case INDEX_op_qemu_st_a32_i32: - case INDEX_op_qemu_st_a64_i32: - case INDEX_op_qemu_st8_a32_i32: - case INDEX_op_qemu_st8_a64_i32: - case INDEX_op_qemu_ld_a32_i64: - case INDEX_op_qemu_ld_a64_i64: - case INDEX_op_qemu_st_a32_i64: - case INDEX_op_qemu_st_a64_i64: - case INDEX_op_qemu_ld_a32_i128: - case INDEX_op_qemu_ld_a64_i128: - case INDEX_op_qemu_st_a32_i128: - case INDEX_op_qemu_st_a64_i128: + case INDEX_op_qemu_ld_i32: + case INDEX_op_qemu_st_i32: + case INDEX_op_qemu_st8_i32: + case INDEX_op_qemu_ld_i64: + case INDEX_op_qemu_st_i64: + case INDEX_op_qemu_ld_i128: + case INDEX_op_qemu_st_i128: { const char *s_al, *s_op, *s_at; MemOpIdx oi = op->args[k++]; diff --git a/tcg/tci.c b/tcg/tci.c index XXXXXXX..XXXXXXX 100644 --- a/tcg/tci.c +++ b/tcg/tci.c @@ -XXX,XX +XXX,XX @@ static void tci_args_rrrbb(uint32_t insn, TCGReg *r0, TCGReg *r1, *i4 = extract32(insn, 26, 6); } -static void tci_args_rrrrr(uint32_t insn, TCGReg *r0, TCGReg *r1, - TCGReg *r2, TCGReg *r3, TCGReg *r4) -{ - *r0 = extract32(insn, 8, 4); - *r1 = extract32(insn, 12, 4); - *r2 = extract32(insn, 16, 4); - *r3 = extract32(insn, 20, 4); - *r4 = extract32(insn, 24, 4); -} - static void tci_args_rrrr(uint32_t insn, TCGReg *r0, TCGReg *r1, TCGReg *r2, TCGReg *r3) { @@ -XXX,XX +XXX,XX @@ uintptr_t QEMU_DISABLE_CFI tcg_qemu_tb_exec(CPUArchState *env, tb_ptr = ptr; break; - case INDEX_op_qemu_ld_a32_i32: + case INDEX_op_qemu_ld_i32: tci_args_rrm(insn, &r0, &r1, &oi); - taddr = (uint32_t)regs[r1]; - goto do_ld_i32; - case INDEX_op_qemu_ld_a64_i32: - if (TCG_TARGET_REG_BITS == 64) { - tci_args_rrm(insn, &r0, &r1, &oi); - taddr = regs[r1]; - } else { - tci_args_rrrr(insn, &r0, &r1, &r2, &r3); - taddr = tci_uint64(regs[r2], regs[r1]); - oi = regs[r3]; - } - do_ld_i32: + taddr = regs[r1]; regs[r0] = tci_qemu_ld(env, taddr, oi, tb_ptr); break; - case INDEX_op_qemu_ld_a32_i64: - if (TCG_TARGET_REG_BITS == 64) { - tci_args_rrm(insn, &r0, &r1, &oi); - taddr = (uint32_t)regs[r1]; - } else { - tci_args_rrrr(insn, &r0, &r1, &r2, &r3); - taddr = (uint32_t)regs[r2]; - oi = regs[r3]; - } - goto do_ld_i64; - case INDEX_op_qemu_ld_a64_i64: + case INDEX_op_qemu_ld_i64: if (TCG_TARGET_REG_BITS == 64) { tci_args_rrm(insn, &r0, &r1, &oi); taddr = regs[r1]; } else { - tci_args_rrrrr(insn, &r0, &r1, &r2, &r3, &r4); - taddr = tci_uint64(regs[r3], regs[r2]); - oi = regs[r4]; + tci_args_rrrr(insn, &r0, &r1, &r2, &r3); + taddr = regs[r2]; + oi = regs[r3]; } - do_ld_i64: tmp64 = tci_qemu_ld(env, taddr, oi, tb_ptr); if (TCG_TARGET_REG_BITS == 32) { tci_write_reg64(regs, r1, r0, tmp64); @@ -XXX,XX +XXX,XX @@ uintptr_t QEMU_DISABLE_CFI tcg_qemu_tb_exec(CPUArchState *env, } break; - case INDEX_op_qemu_st_a32_i32: + case INDEX_op_qemu_st_i32: tci_args_rrm(insn, &r0, &r1, &oi); - taddr = (uint32_t)regs[r1]; - goto do_st_i32; - case INDEX_op_qemu_st_a64_i32: - if (TCG_TARGET_REG_BITS == 64) { - tci_args_rrm(insn, &r0, &r1, &oi); - taddr = regs[r1]; - } else { - tci_args_rrrr(insn, &r0, &r1, &r2, &r3); - taddr = tci_uint64(regs[r2], regs[r1]); - oi = regs[r3]; - } - do_st_i32: + taddr = regs[r1]; tci_qemu_st(env, taddr, regs[r0], oi, tb_ptr); break; - case INDEX_op_qemu_st_a32_i64: - if (TCG_TARGET_REG_BITS == 64) { - tci_args_rrm(insn, &r0, &r1, &oi); - tmp64 = regs[r0]; - taddr = (uint32_t)regs[r1]; - } else { - tci_args_rrrr(insn, &r0, &r1, &r2, &r3); - tmp64 = tci_uint64(regs[r1], regs[r0]); - taddr = (uint32_t)regs[r2]; - oi = regs[r3]; - } - goto do_st_i64; - case INDEX_op_qemu_st_a64_i64: + case INDEX_op_qemu_st_i64: if (TCG_TARGET_REG_BITS == 64) { tci_args_rrm(insn, &r0, &r1, &oi); tmp64 = regs[r0]; taddr = regs[r1]; } else { - tci_args_rrrrr(insn, &r0, &r1, &r2, &r3, &r4); + tci_args_rrrr(insn, &r0, &r1, &r2, &r3); tmp64 = tci_uint64(regs[r1], regs[r0]); - taddr = tci_uint64(regs[r3], regs[r2]); - oi = regs[r4]; + taddr = regs[r2]; + oi = regs[r3]; } - do_st_i64: tci_qemu_st(env, taddr, tmp64, oi, tb_ptr); break; @@ -XXX,XX +XXX,XX @@ int print_insn_tci(bfd_vma addr, disassemble_info *info) str_r(r3), str_r(r4), str_r(r5)); break; - case INDEX_op_qemu_ld_a32_i32: - case INDEX_op_qemu_st_a32_i32: - len = 1 + 1; - goto do_qemu_ldst; - case INDEX_op_qemu_ld_a32_i64: - case INDEX_op_qemu_st_a32_i64: - case INDEX_op_qemu_ld_a64_i32: - case INDEX_op_qemu_st_a64_i32: - len = 1 + DIV_ROUND_UP(64, TCG_TARGET_REG_BITS); - goto do_qemu_ldst; - case INDEX_op_qemu_ld_a64_i64: - case INDEX_op_qemu_st_a64_i64: - len = 2 * DIV_ROUND_UP(64, TCG_TARGET_REG_BITS); - goto do_qemu_ldst; - do_qemu_ldst: - switch (len) { - case 2: - tci_args_rrm(insn, &r0, &r1, &oi); - info->fprintf_func(info->stream, "%-12s %s, %s, %x", - op_name, str_r(r0), str_r(r1), oi); - break; - case 3: + case INDEX_op_qemu_ld_i64: + case INDEX_op_qemu_st_i64: + if (TCG_TARGET_REG_BITS == 32) { tci_args_rrrr(insn, &r0, &r1, &r2, &r3); info->fprintf_func(info->stream, "%-12s %s, %s, %s, %s", op_name, str_r(r0), str_r(r1), str_r(r2), str_r(r3)); break; - case 4: - tci_args_rrrrr(insn, &r0, &r1, &r2, &r3, &r4); - info->fprintf_func(info->stream, "%-12s %s, %s, %s, %s, %s", - op_name, str_r(r0), str_r(r1), - str_r(r2), str_r(r3), str_r(r4)); - break; - default: - g_assert_not_reached(); } + /* fall through */ + case INDEX_op_qemu_ld_i32: + case INDEX_op_qemu_st_i32: + tci_args_rrm(insn, &r0, &r1, &oi); + info->fprintf_func(info->stream, "%-12s %s, %s, %x", + op_name, str_r(r0), str_r(r1), oi); break; case 0: diff --git a/tcg/aarch64/tcg-target.c.inc b/tcg/aarch64/tcg-target.c.inc index XXXXXXX..XXXXXXX 100644 --- a/tcg/aarch64/tcg-target.c.inc +++ b/tcg/aarch64/tcg-target.c.inc @@ -XXX,XX +XXX,XX @@ static void tcg_out_op(TCGContext *s, TCGOpcode opc, TCGType ext, tcg_out_insn(s, 3506, CSEL, ext, a0, REG0(3), REG0(4), args[5]); break; - case INDEX_op_qemu_ld_a32_i32: - case INDEX_op_qemu_ld_a64_i32: - case INDEX_op_qemu_ld_a32_i64: - case INDEX_op_qemu_ld_a64_i64: + case INDEX_op_qemu_ld_i32: + case INDEX_op_qemu_ld_i64: tcg_out_qemu_ld(s, a0, a1, a2, ext); break; - case INDEX_op_qemu_st_a32_i32: - case INDEX_op_qemu_st_a64_i32: - case INDEX_op_qemu_st_a32_i64: - case INDEX_op_qemu_st_a64_i64: + case INDEX_op_qemu_st_i32: + case INDEX_op_qemu_st_i64: tcg_out_qemu_st(s, REG0(0), a1, a2, ext); break; - case INDEX_op_qemu_ld_a32_i128: - case INDEX_op_qemu_ld_a64_i128: + case INDEX_op_qemu_ld_i128: tcg_out_qemu_ldst_i128(s, a0, a1, a2, args[3], true); break; - case INDEX_op_qemu_st_a32_i128: - case INDEX_op_qemu_st_a64_i128: + case INDEX_op_qemu_st_i128: tcg_out_qemu_ldst_i128(s, REG0(0), REG0(1), a2, args[3], false); break; @@ -XXX,XX +XXX,XX @@ tcg_target_op_def(TCGOpcode op, TCGType type, unsigned flags) case INDEX_op_movcond_i64: return C_O1_I4(r, r, rC, rZ, rZ); - case INDEX_op_qemu_ld_a32_i32: - case INDEX_op_qemu_ld_a64_i32: - case INDEX_op_qemu_ld_a32_i64: - case INDEX_op_qemu_ld_a64_i64: + case INDEX_op_qemu_ld_i32: + case INDEX_op_qemu_ld_i64: return C_O1_I1(r, r); - case INDEX_op_qemu_ld_a32_i128: - case INDEX_op_qemu_ld_a64_i128: + case INDEX_op_qemu_ld_i128: return C_O2_I1(r, r, r); - case INDEX_op_qemu_st_a32_i32: - case INDEX_op_qemu_st_a64_i32: - case INDEX_op_qemu_st_a32_i64: - case INDEX_op_qemu_st_a64_i64: + case INDEX_op_qemu_st_i32: + case INDEX_op_qemu_st_i64: return C_O0_I2(rZ, r); - case INDEX_op_qemu_st_a32_i128: - case INDEX_op_qemu_st_a64_i128: + case INDEX_op_qemu_st_i128: return C_O0_I3(rZ, rZ, r); case INDEX_op_deposit_i32: diff --git a/tcg/arm/tcg-target.c.inc b/tcg/arm/tcg-target.c.inc index XXXXXXX..XXXXXXX 100644 --- a/tcg/arm/tcg-target.c.inc +++ b/tcg/arm/tcg-target.c.inc @@ -XXX,XX +XXX,XX @@ static void tcg_out_op(TCGContext *s, TCGOpcode opc, TCGType type, ARITH_MOV, args[0], 0, 0); break; - case INDEX_op_qemu_ld_a32_i32: + case INDEX_op_qemu_ld_i32: tcg_out_qemu_ld(s, args[0], -1, args[1], -1, args[2], TCG_TYPE_I32); break; - case INDEX_op_qemu_ld_a64_i32: - tcg_out_qemu_ld(s, args[0], -1, args[1], args[2], - args[3], TCG_TYPE_I32); - break; - case INDEX_op_qemu_ld_a32_i64: + case INDEX_op_qemu_ld_i64: tcg_out_qemu_ld(s, args[0], args[1], args[2], -1, args[3], TCG_TYPE_I64); break; - case INDEX_op_qemu_ld_a64_i64: - tcg_out_qemu_ld(s, args[0], args[1], args[2], args[3], - args[4], TCG_TYPE_I64); - break; - case INDEX_op_qemu_st_a32_i32: + case INDEX_op_qemu_st_i32: tcg_out_qemu_st(s, args[0], -1, args[1], -1, args[2], TCG_TYPE_I32); break; - case INDEX_op_qemu_st_a64_i32: - tcg_out_qemu_st(s, args[0], -1, args[1], args[2], - args[3], TCG_TYPE_I32); - break; - case INDEX_op_qemu_st_a32_i64: + case INDEX_op_qemu_st_i64: tcg_out_qemu_st(s, args[0], args[1], args[2], -1, args[3], TCG_TYPE_I64); break; - case INDEX_op_qemu_st_a64_i64: - tcg_out_qemu_st(s, args[0], args[1], args[2], args[3], - args[4], TCG_TYPE_I64); - break; case INDEX_op_bswap16_i32: tcg_out_bswap16(s, COND_AL, args[0], args[1], args[2]); @@ -XXX,XX +XXX,XX @@ tcg_target_op_def(TCGOpcode op, TCGType type, unsigned flags) case INDEX_op_setcond2_i32: return C_O1_I4(r, r, r, rI, rI); - case INDEX_op_qemu_ld_a32_i32: + case INDEX_op_qemu_ld_i32: return C_O1_I1(r, q); - case INDEX_op_qemu_ld_a64_i32: - return C_O1_I2(r, q, q); - case INDEX_op_qemu_ld_a32_i64: + case INDEX_op_qemu_ld_i64: return C_O2_I1(e, p, q); - case INDEX_op_qemu_ld_a64_i64: - return C_O2_I2(e, p, q, q); - case INDEX_op_qemu_st_a32_i32: + case INDEX_op_qemu_st_i32: return C_O0_I2(q, q); - case INDEX_op_qemu_st_a64_i32: - return C_O0_I3(q, q, q); - case INDEX_op_qemu_st_a32_i64: + case INDEX_op_qemu_st_i64: return C_O0_I3(Q, p, q); - case INDEX_op_qemu_st_a64_i64: - return C_O0_I4(Q, p, q, q); case INDEX_op_st_vec: return C_O0_I2(w, r); diff --git a/tcg/i386/tcg-target.c.inc b/tcg/i386/tcg-target.c.inc index XXXXXXX..XXXXXXX 100644 --- a/tcg/i386/tcg-target.c.inc +++ b/tcg/i386/tcg-target.c.inc @@ -XXX,XX +XXX,XX @@ static void tcg_out_op(TCGContext *s, TCGOpcode opc, TCGType type, tcg_out_modrm(s, OPC_GRP3_Ev + rexw, EXT3_NOT, a0); break; - case INDEX_op_qemu_ld_a64_i32: - if (TCG_TARGET_REG_BITS == 32) { - tcg_out_qemu_ld(s, a0, -1, a1, a2, args[3], TCG_TYPE_I32); - break; - } - /* fall through */ - case INDEX_op_qemu_ld_a32_i32: + case INDEX_op_qemu_ld_i32: tcg_out_qemu_ld(s, a0, -1, a1, -1, a2, TCG_TYPE_I32); break; - case INDEX_op_qemu_ld_a32_i64: + case INDEX_op_qemu_ld_i64: if (TCG_TARGET_REG_BITS == 64) { tcg_out_qemu_ld(s, a0, -1, a1, -1, a2, TCG_TYPE_I64); } else { tcg_out_qemu_ld(s, a0, a1, a2, -1, args[3], TCG_TYPE_I64); } break; - case INDEX_op_qemu_ld_a64_i64: - if (TCG_TARGET_REG_BITS == 64) { - tcg_out_qemu_ld(s, a0, -1, a1, -1, a2, TCG_TYPE_I64); - } else { - tcg_out_qemu_ld(s, a0, a1, a2, args[3], args[4], TCG_TYPE_I64); - } - break; - case INDEX_op_qemu_ld_a32_i128: - case INDEX_op_qemu_ld_a64_i128: + case INDEX_op_qemu_ld_i128: tcg_debug_assert(TCG_TARGET_REG_BITS == 64); tcg_out_qemu_ld(s, a0, a1, a2, -1, args[3], TCG_TYPE_I128); break; - case INDEX_op_qemu_st_a64_i32: - case INDEX_op_qemu_st8_a64_i32: - if (TCG_TARGET_REG_BITS == 32) { - tcg_out_qemu_st(s, a0, -1, a1, a2, args[3], TCG_TYPE_I32); - break; - } - /* fall through */ - case INDEX_op_qemu_st_a32_i32: - case INDEX_op_qemu_st8_a32_i32: + case INDEX_op_qemu_st_i32: + case INDEX_op_qemu_st8_i32: tcg_out_qemu_st(s, a0, -1, a1, -1, a2, TCG_TYPE_I32); break; - case INDEX_op_qemu_st_a32_i64: + case INDEX_op_qemu_st_i64: if (TCG_TARGET_REG_BITS == 64) { tcg_out_qemu_st(s, a0, -1, a1, -1, a2, TCG_TYPE_I64); } else { tcg_out_qemu_st(s, a0, a1, a2, -1, args[3], TCG_TYPE_I64); } break; - case INDEX_op_qemu_st_a64_i64: - if (TCG_TARGET_REG_BITS == 64) { - tcg_out_qemu_st(s, a0, -1, a1, -1, a2, TCG_TYPE_I64); - } else { - tcg_out_qemu_st(s, a0, a1, a2, args[3], args[4], TCG_TYPE_I64); - } - break; - case INDEX_op_qemu_st_a32_i128: - case INDEX_op_qemu_st_a64_i128: + case INDEX_op_qemu_st_i128: tcg_debug_assert(TCG_TARGET_REG_BITS == 64); tcg_out_qemu_st(s, a0, a1, a2, -1, args[3], TCG_TYPE_I128); break; @@ -XXX,XX +XXX,XX @@ tcg_target_op_def(TCGOpcode op, TCGType type, unsigned flags) case INDEX_op_clz_i64: return have_lzcnt ? C_N1_I2(r, r, rW) : C_N1_I2(r, r, r); - case INDEX_op_qemu_ld_a32_i32: + case INDEX_op_qemu_ld_i32: return C_O1_I1(r, L); - case INDEX_op_qemu_ld_a64_i32: - return TCG_TARGET_REG_BITS == 64 ? C_O1_I1(r, L) : C_O1_I2(r, L, L); - case INDEX_op_qemu_st_a32_i32: + case INDEX_op_qemu_st_i32: return C_O0_I2(L, L); - case INDEX_op_qemu_st_a64_i32: - return TCG_TARGET_REG_BITS == 64 ? C_O0_I2(L, L) : C_O0_I3(L, L, L); - case INDEX_op_qemu_st8_a32_i32: + case INDEX_op_qemu_st8_i32: return C_O0_I2(s, L); - case INDEX_op_qemu_st8_a64_i32: - return TCG_TARGET_REG_BITS == 64 ? C_O0_I2(s, L) : C_O0_I3(s, L, L); - case INDEX_op_qemu_ld_a32_i64: + case INDEX_op_qemu_ld_i64: return TCG_TARGET_REG_BITS == 64 ? C_O1_I1(r, L) : C_O2_I1(r, r, L); - case INDEX_op_qemu_ld_a64_i64: - return TCG_TARGET_REG_BITS == 64 ? C_O1_I1(r, L) : C_O2_I2(r, r, L, L); - case INDEX_op_qemu_st_a32_i64: + case INDEX_op_qemu_st_i64: return TCG_TARGET_REG_BITS == 64 ? C_O0_I2(L, L) : C_O0_I3(L, L, L); - case INDEX_op_qemu_st_a64_i64: - return TCG_TARGET_REG_BITS == 64 ? C_O0_I2(L, L) : C_O0_I4(L, L, L, L); - case INDEX_op_qemu_ld_a32_i128: - case INDEX_op_qemu_ld_a64_i128: + case INDEX_op_qemu_ld_i128: tcg_debug_assert(TCG_TARGET_REG_BITS == 64); return C_O2_I1(r, r, L); - case INDEX_op_qemu_st_a32_i128: - case INDEX_op_qemu_st_a64_i128: + case INDEX_op_qemu_st_i128: tcg_debug_assert(TCG_TARGET_REG_BITS == 64); return C_O0_I3(L, L, L); 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_op(TCGContext *s, TCGOpcode opc, TCGType type, tcg_out_ldst(s, OPC_ST_D, a0, a1, a2); break; - case INDEX_op_qemu_ld_a32_i32: - case INDEX_op_qemu_ld_a64_i32: + case INDEX_op_qemu_ld_i32: tcg_out_qemu_ld(s, a0, a1, a2, TCG_TYPE_I32); break; - case INDEX_op_qemu_ld_a32_i64: - case INDEX_op_qemu_ld_a64_i64: + case INDEX_op_qemu_ld_i64: tcg_out_qemu_ld(s, a0, a1, a2, TCG_TYPE_I64); break; - case INDEX_op_qemu_ld_a32_i128: - case INDEX_op_qemu_ld_a64_i128: + case INDEX_op_qemu_ld_i128: tcg_out_qemu_ldst_i128(s, a0, a1, a2, a3, true); break; - case INDEX_op_qemu_st_a32_i32: - case INDEX_op_qemu_st_a64_i32: + case INDEX_op_qemu_st_i32: tcg_out_qemu_st(s, a0, a1, a2, TCG_TYPE_I32); break; - case INDEX_op_qemu_st_a32_i64: - case INDEX_op_qemu_st_a64_i64: + case INDEX_op_qemu_st_i64: tcg_out_qemu_st(s, a0, a1, a2, TCG_TYPE_I64); break; - case INDEX_op_qemu_st_a32_i128: - case INDEX_op_qemu_st_a64_i128: + case INDEX_op_qemu_st_i128: tcg_out_qemu_ldst_i128(s, a0, a1, a2, a3, false); break; @@ -XXX,XX +XXX,XX @@ tcg_target_op_def(TCGOpcode op, TCGType type, unsigned flags) case INDEX_op_st32_i64: case INDEX_op_st_i32: case INDEX_op_st_i64: - case INDEX_op_qemu_st_a32_i32: - case INDEX_op_qemu_st_a64_i32: - case INDEX_op_qemu_st_a32_i64: - case INDEX_op_qemu_st_a64_i64: + case INDEX_op_qemu_st_i32: + case INDEX_op_qemu_st_i64: return C_O0_I2(rZ, r); - case INDEX_op_qemu_ld_a32_i128: - case INDEX_op_qemu_ld_a64_i128: + case INDEX_op_qemu_ld_i128: return C_N2_I1(r, r, r); - case INDEX_op_qemu_st_a32_i128: - case INDEX_op_qemu_st_a64_i128: + case INDEX_op_qemu_st_i128: return C_O0_I3(r, r, r); case INDEX_op_brcond_i32: @@ -XXX,XX +XXX,XX @@ tcg_target_op_def(TCGOpcode op, TCGType type, unsigned flags) case INDEX_op_ld32u_i64: case INDEX_op_ld_i32: case INDEX_op_ld_i64: - case INDEX_op_qemu_ld_a32_i32: - case INDEX_op_qemu_ld_a64_i32: - case INDEX_op_qemu_ld_a32_i64: - case INDEX_op_qemu_ld_a64_i64: + case INDEX_op_qemu_ld_i32: + case INDEX_op_qemu_ld_i64: return C_O1_I1(r, r); case INDEX_op_andc_i32: diff --git a/tcg/mips/tcg-target.c.inc b/tcg/mips/tcg-target.c.inc index XXXXXXX..XXXXXXX 100644 --- a/tcg/mips/tcg-target.c.inc +++ b/tcg/mips/tcg-target.c.inc @@ -XXX,XX +XXX,XX @@ static void tcg_out_op(TCGContext *s, TCGOpcode opc, TCGType type, tcg_out_setcond2(s, args[5], a0, a1, a2, args[3], args[4]); break; - case INDEX_op_qemu_ld_a64_i32: - if (TCG_TARGET_REG_BITS == 32) { - tcg_out_qemu_ld(s, a0, 0, a1, a2, args[3], TCG_TYPE_I32); - break; - } - /* fall through */ - case INDEX_op_qemu_ld_a32_i32: + case INDEX_op_qemu_ld_i32: tcg_out_qemu_ld(s, a0, 0, a1, 0, a2, TCG_TYPE_I32); break; - case INDEX_op_qemu_ld_a32_i64: + case INDEX_op_qemu_ld_i64: if (TCG_TARGET_REG_BITS == 64) { tcg_out_qemu_ld(s, a0, 0, a1, 0, a2, TCG_TYPE_I64); } else { tcg_out_qemu_ld(s, a0, a1, a2, 0, args[3], TCG_TYPE_I64); } break; - case INDEX_op_qemu_ld_a64_i64: - if (TCG_TARGET_REG_BITS == 64) { - tcg_out_qemu_ld(s, a0, 0, a1, 0, a2, TCG_TYPE_I64); - } else { - tcg_out_qemu_ld(s, a0, a1, a2, args[3], args[4], TCG_TYPE_I64); - } - break; - case INDEX_op_qemu_st_a64_i32: - if (TCG_TARGET_REG_BITS == 32) { - tcg_out_qemu_st(s, a0, 0, a1, a2, args[3], TCG_TYPE_I32); - break; - } - /* fall through */ - case INDEX_op_qemu_st_a32_i32: + case INDEX_op_qemu_st_i32: tcg_out_qemu_st(s, a0, 0, a1, 0, a2, TCG_TYPE_I32); break; - case INDEX_op_qemu_st_a32_i64: + case INDEX_op_qemu_st_i64: if (TCG_TARGET_REG_BITS == 64) { tcg_out_qemu_st(s, a0, 0, a1, 0, a2, TCG_TYPE_I64); } else { tcg_out_qemu_st(s, a0, a1, a2, 0, args[3], TCG_TYPE_I64); } break; - case INDEX_op_qemu_st_a64_i64: - if (TCG_TARGET_REG_BITS == 64) { - tcg_out_qemu_st(s, a0, 0, a1, 0, a2, TCG_TYPE_I64); - } else { - tcg_out_qemu_st(s, a0, a1, a2, args[3], args[4], TCG_TYPE_I64); - } - break; case INDEX_op_add2_i32: tcg_out_addsub2(s, a0, a1, a2, args[3], args[4], args[5], @@ -XXX,XX +XXX,XX @@ tcg_target_op_def(TCGOpcode op, TCGType type, unsigned flags) case INDEX_op_brcond2_i32: return C_O0_I4(rZ, rZ, rZ, rZ); - case INDEX_op_qemu_ld_a32_i32: + case INDEX_op_qemu_ld_i32: return C_O1_I1(r, r); - case INDEX_op_qemu_ld_a64_i32: - return TCG_TARGET_REG_BITS == 64 ? C_O1_I1(r, r) : C_O1_I2(r, r, r); - case INDEX_op_qemu_st_a32_i32: + case INDEX_op_qemu_st_i32: return C_O0_I2(rZ, r); - case INDEX_op_qemu_st_a64_i32: - return TCG_TARGET_REG_BITS == 64 ? C_O0_I2(rZ, r) : C_O0_I3(rZ, r, r); - case INDEX_op_qemu_ld_a32_i64: + case INDEX_op_qemu_ld_i64: return TCG_TARGET_REG_BITS == 64 ? C_O1_I1(r, r) : C_O2_I1(r, r, r); - case INDEX_op_qemu_ld_a64_i64: - return TCG_TARGET_REG_BITS == 64 ? C_O1_I1(r, r) : C_O2_I2(r, r, r, r); - case INDEX_op_qemu_st_a32_i64: + case INDEX_op_qemu_st_i64: return TCG_TARGET_REG_BITS == 64 ? C_O0_I2(rZ, r) : C_O0_I3(rZ, rZ, r); - case INDEX_op_qemu_st_a64_i64: - return (TCG_TARGET_REG_BITS == 64 ? C_O0_I2(rZ, r) - : C_O0_I4(rZ, rZ, r, r)); default: return C_NotImplemented; diff --git a/tcg/ppc/tcg-target.c.inc b/tcg/ppc/tcg-target.c.inc index XXXXXXX..XXXXXXX 100644 --- a/tcg/ppc/tcg-target.c.inc +++ b/tcg/ppc/tcg-target.c.inc @@ -XXX,XX +XXX,XX @@ static void tcg_out_op(TCGContext *s, TCGOpcode opc, TCGType type, tcg_out32(s, MODUD | TAB(args[0], args[1], args[2])); break; - case INDEX_op_qemu_ld_a64_i32: - if (TCG_TARGET_REG_BITS == 32) { - tcg_out_qemu_ld(s, args[0], -1, args[1], args[2], - args[3], TCG_TYPE_I32); - break; - } - /* fall through */ - case INDEX_op_qemu_ld_a32_i32: + case INDEX_op_qemu_ld_i32: tcg_out_qemu_ld(s, args[0], -1, args[1], -1, args[2], TCG_TYPE_I32); break; - case INDEX_op_qemu_ld_a32_i64: + case INDEX_op_qemu_ld_i64: if (TCG_TARGET_REG_BITS == 64) { tcg_out_qemu_ld(s, args[0], -1, args[1], -1, args[2], TCG_TYPE_I64); @@ -XXX,XX +XXX,XX @@ static void tcg_out_op(TCGContext *s, TCGOpcode opc, TCGType type, args[3], TCG_TYPE_I64); } break; - case INDEX_op_qemu_ld_a64_i64: - if (TCG_TARGET_REG_BITS == 64) { - tcg_out_qemu_ld(s, args[0], -1, args[1], -1, - args[2], TCG_TYPE_I64); - } else { - tcg_out_qemu_ld(s, args[0], args[1], args[2], args[3], - args[4], TCG_TYPE_I64); - } - break; - case INDEX_op_qemu_ld_a32_i128: - case INDEX_op_qemu_ld_a64_i128: + case INDEX_op_qemu_ld_i128: tcg_debug_assert(TCG_TARGET_REG_BITS == 64); tcg_out_qemu_ldst_i128(s, args[0], args[1], args[2], args[3], true); break; - case INDEX_op_qemu_st_a64_i32: - if (TCG_TARGET_REG_BITS == 32) { - tcg_out_qemu_st(s, args[0], -1, args[1], args[2], - args[3], TCG_TYPE_I32); - break; - } - /* fall through */ - case INDEX_op_qemu_st_a32_i32: + case INDEX_op_qemu_st_i32: tcg_out_qemu_st(s, args[0], -1, args[1], -1, args[2], TCG_TYPE_I32); break; - case INDEX_op_qemu_st_a32_i64: + case INDEX_op_qemu_st_i64: if (TCG_TARGET_REG_BITS == 64) { tcg_out_qemu_st(s, args[0], -1, args[1], -1, args[2], TCG_TYPE_I64); @@ -XXX,XX +XXX,XX @@ static void tcg_out_op(TCGContext *s, TCGOpcode opc, TCGType type, args[3], TCG_TYPE_I64); } break; - case INDEX_op_qemu_st_a64_i64: - if (TCG_TARGET_REG_BITS == 64) { - tcg_out_qemu_st(s, args[0], -1, args[1], -1, - args[2], TCG_TYPE_I64); - } else { - tcg_out_qemu_st(s, args[0], args[1], args[2], args[3], - args[4], TCG_TYPE_I64); - } - break; - case INDEX_op_qemu_st_a32_i128: - case INDEX_op_qemu_st_a64_i128: + case INDEX_op_qemu_st_i128: tcg_debug_assert(TCG_TARGET_REG_BITS == 64); tcg_out_qemu_ldst_i128(s, args[0], args[1], args[2], args[3], false); break; @@ -XXX,XX +XXX,XX @@ tcg_target_op_def(TCGOpcode op, TCGType type, unsigned flags) case INDEX_op_sub2_i32: return C_O2_I4(r, r, rI, rZM, r, r); - case INDEX_op_qemu_ld_a32_i32: + case INDEX_op_qemu_ld_i32: return C_O1_I1(r, r); - case INDEX_op_qemu_ld_a64_i32: - return TCG_TARGET_REG_BITS == 64 ? C_O1_I1(r, r) : C_O1_I2(r, r, r); - case INDEX_op_qemu_ld_a32_i64: + case INDEX_op_qemu_ld_i64: return TCG_TARGET_REG_BITS == 64 ? C_O1_I1(r, r) : C_O2_I1(r, r, r); - case INDEX_op_qemu_ld_a64_i64: - return TCG_TARGET_REG_BITS == 64 ? C_O1_I1(r, r) : C_O2_I2(r, r, r, r); - case INDEX_op_qemu_st_a32_i32: + case INDEX_op_qemu_st_i32: return C_O0_I2(r, r); - case INDEX_op_qemu_st_a64_i32: + case INDEX_op_qemu_st_i64: return TCG_TARGET_REG_BITS == 64 ? C_O0_I2(r, r) : C_O0_I3(r, r, r); - case INDEX_op_qemu_st_a32_i64: - return TCG_TARGET_REG_BITS == 64 ? C_O0_I2(r, r) : C_O0_I3(r, r, r); - case INDEX_op_qemu_st_a64_i64: - return TCG_TARGET_REG_BITS == 64 ? C_O0_I2(r, r) : C_O0_I4(r, r, r, r); - case INDEX_op_qemu_ld_a32_i128: - case INDEX_op_qemu_ld_a64_i128: + case INDEX_op_qemu_ld_i128: return C_N1O1_I1(o, m, r); - case INDEX_op_qemu_st_a32_i128: - case INDEX_op_qemu_st_a64_i128: + case INDEX_op_qemu_st_i128: return C_O0_I3(o, m, r); case INDEX_op_add_vec: diff --git a/tcg/riscv/tcg-target.c.inc b/tcg/riscv/tcg-target.c.inc index XXXXXXX..XXXXXXX 100644 --- a/tcg/riscv/tcg-target.c.inc +++ b/tcg/riscv/tcg-target.c.inc @@ -XXX,XX +XXX,XX @@ static void tcg_out_op(TCGContext *s, TCGOpcode opc, TCGType type, args[3], const_args[3], args[4], const_args[4]); break; - case INDEX_op_qemu_ld_a32_i32: - case INDEX_op_qemu_ld_a64_i32: + case INDEX_op_qemu_ld_i32: tcg_out_qemu_ld(s, a0, a1, a2, TCG_TYPE_I32); break; - case INDEX_op_qemu_ld_a32_i64: - case INDEX_op_qemu_ld_a64_i64: + case INDEX_op_qemu_ld_i64: tcg_out_qemu_ld(s, a0, a1, a2, TCG_TYPE_I64); break; - case INDEX_op_qemu_st_a32_i32: - case INDEX_op_qemu_st_a64_i32: + case INDEX_op_qemu_st_i32: tcg_out_qemu_st(s, a0, a1, a2, TCG_TYPE_I32); break; - case INDEX_op_qemu_st_a32_i64: - case INDEX_op_qemu_st_a64_i64: + case INDEX_op_qemu_st_i64: tcg_out_qemu_st(s, a0, a1, a2, TCG_TYPE_I64); break; @@ -XXX,XX +XXX,XX @@ tcg_target_op_def(TCGOpcode op, TCGType type, unsigned flags) case INDEX_op_sub2_i64: return C_O2_I4(r, r, rZ, rZ, rM, rM); - case INDEX_op_qemu_ld_a32_i32: - case INDEX_op_qemu_ld_a64_i32: - case INDEX_op_qemu_ld_a32_i64: - case INDEX_op_qemu_ld_a64_i64: + case INDEX_op_qemu_ld_i32: + case INDEX_op_qemu_ld_i64: return C_O1_I1(r, r); - case INDEX_op_qemu_st_a32_i32: - case INDEX_op_qemu_st_a64_i32: - case INDEX_op_qemu_st_a32_i64: - case INDEX_op_qemu_st_a64_i64: + case INDEX_op_qemu_st_i32: + case INDEX_op_qemu_st_i64: return C_O0_I2(rZ, r); case INDEX_op_st_vec: diff --git a/tcg/s390x/tcg-target.c.inc b/tcg/s390x/tcg-target.c.inc index XXXXXXX..XXXXXXX 100644 --- a/tcg/s390x/tcg-target.c.inc +++ b/tcg/s390x/tcg-target.c.inc @@ -XXX,XX +XXX,XX @@ static void tcg_out_op(TCGContext *s, TCGOpcode opc, TCGType type, args[2], const_args[2], args[3], const_args[3], args[4]); break; - case INDEX_op_qemu_ld_a32_i32: - case INDEX_op_qemu_ld_a64_i32: + case INDEX_op_qemu_ld_i32: tcg_out_qemu_ld(s, args[0], args[1], args[2], TCG_TYPE_I32); break; - case INDEX_op_qemu_ld_a32_i64: - case INDEX_op_qemu_ld_a64_i64: + case INDEX_op_qemu_ld_i64: tcg_out_qemu_ld(s, args[0], args[1], args[2], TCG_TYPE_I64); break; - case INDEX_op_qemu_st_a32_i32: - case INDEX_op_qemu_st_a64_i32: + case INDEX_op_qemu_st_i32: tcg_out_qemu_st(s, args[0], args[1], args[2], TCG_TYPE_I32); break; - case INDEX_op_qemu_st_a32_i64: - case INDEX_op_qemu_st_a64_i64: + case INDEX_op_qemu_st_i64: tcg_out_qemu_st(s, args[0], args[1], args[2], TCG_TYPE_I64); break; - case INDEX_op_qemu_ld_a32_i128: - case INDEX_op_qemu_ld_a64_i128: + case INDEX_op_qemu_ld_i128: tcg_out_qemu_ldst_i128(s, args[0], args[1], args[2], args[3], true); break; - case INDEX_op_qemu_st_a32_i128: - case INDEX_op_qemu_st_a64_i128: + case INDEX_op_qemu_st_i128: tcg_out_qemu_ldst_i128(s, args[0], args[1], args[2], args[3], false); break; @@ -XXX,XX +XXX,XX @@ tcg_target_op_def(TCGOpcode op, TCGType type, unsigned flags) case INDEX_op_ctpop_i64: return C_O1_I1(r, r); - case INDEX_op_qemu_ld_a32_i32: - case INDEX_op_qemu_ld_a64_i32: - case INDEX_op_qemu_ld_a32_i64: - case INDEX_op_qemu_ld_a64_i64: + case INDEX_op_qemu_ld_i32: + case INDEX_op_qemu_ld_i64: return C_O1_I1(r, r); - case INDEX_op_qemu_st_a32_i64: - case INDEX_op_qemu_st_a64_i64: - case INDEX_op_qemu_st_a32_i32: - case INDEX_op_qemu_st_a64_i32: + case INDEX_op_qemu_st_i64: + case INDEX_op_qemu_st_i32: return C_O0_I2(r, r); - case INDEX_op_qemu_ld_a32_i128: - case INDEX_op_qemu_ld_a64_i128: + case INDEX_op_qemu_ld_i128: return C_O2_I1(o, m, r); - case INDEX_op_qemu_st_a32_i128: - case INDEX_op_qemu_st_a64_i128: + case INDEX_op_qemu_st_i128: return C_O0_I3(o, m, r); case INDEX_op_deposit_i32: diff --git a/tcg/sparc64/tcg-target.c.inc b/tcg/sparc64/tcg-target.c.inc index XXXXXXX..XXXXXXX 100644 --- a/tcg/sparc64/tcg-target.c.inc +++ b/tcg/sparc64/tcg-target.c.inc @@ -XXX,XX +XXX,XX @@ static void tcg_out_op(TCGContext *s, TCGOpcode opc, TCGType type, tcg_out_arithi(s, a1, a0, 32, SHIFT_SRLX); break; - case INDEX_op_qemu_ld_a32_i32: - case INDEX_op_qemu_ld_a64_i32: + case INDEX_op_qemu_ld_i32: tcg_out_qemu_ld(s, a0, a1, a2, TCG_TYPE_I32); break; - case INDEX_op_qemu_ld_a32_i64: - case INDEX_op_qemu_ld_a64_i64: + case INDEX_op_qemu_ld_i64: tcg_out_qemu_ld(s, a0, a1, a2, TCG_TYPE_I64); break; - case INDEX_op_qemu_st_a32_i32: - case INDEX_op_qemu_st_a64_i32: + case INDEX_op_qemu_st_i32: tcg_out_qemu_st(s, a0, a1, a2, TCG_TYPE_I32); break; - case INDEX_op_qemu_st_a32_i64: - case INDEX_op_qemu_st_a64_i64: + case INDEX_op_qemu_st_i64: tcg_out_qemu_st(s, a0, a1, a2, TCG_TYPE_I64); break; @@ -XXX,XX +XXX,XX @@ tcg_target_op_def(TCGOpcode op, TCGType type, unsigned flags) case INDEX_op_extu_i32_i64: case INDEX_op_extract_i64: case INDEX_op_sextract_i64: - case INDEX_op_qemu_ld_a32_i32: - case INDEX_op_qemu_ld_a64_i32: - case INDEX_op_qemu_ld_a32_i64: - case INDEX_op_qemu_ld_a64_i64: + case INDEX_op_qemu_ld_i32: + case INDEX_op_qemu_ld_i64: return C_O1_I1(r, r); case INDEX_op_st8_i32: @@ -XXX,XX +XXX,XX @@ tcg_target_op_def(TCGOpcode op, TCGType type, unsigned flags) case INDEX_op_st_i32: case INDEX_op_st32_i64: case INDEX_op_st_i64: - case INDEX_op_qemu_st_a32_i32: - case INDEX_op_qemu_st_a64_i32: - case INDEX_op_qemu_st_a32_i64: - case INDEX_op_qemu_st_a64_i64: + case INDEX_op_qemu_st_i32: + case INDEX_op_qemu_st_i64: return C_O0_I2(rZ, r); case INDEX_op_add_i32: diff --git a/tcg/tci/tcg-target.c.inc b/tcg/tci/tcg-target.c.inc index XXXXXXX..XXXXXXX 100644 --- a/tcg/tci/tcg-target.c.inc +++ b/tcg/tci/tcg-target.c.inc @@ -XXX,XX +XXX,XX @@ tcg_target_op_def(TCGOpcode op, TCGType type, unsigned flags) case INDEX_op_setcond2_i32: return C_O1_I4(r, r, r, r, r); - case INDEX_op_qemu_ld_a32_i32: + case INDEX_op_qemu_ld_i32: return C_O1_I1(r, r); - case INDEX_op_qemu_ld_a64_i32: - return TCG_TARGET_REG_BITS == 64 ? C_O1_I1(r, r) : C_O1_I2(r, r, r); - case INDEX_op_qemu_ld_a32_i64: + case INDEX_op_qemu_ld_i64: return TCG_TARGET_REG_BITS == 64 ? C_O1_I1(r, r) : C_O2_I1(r, r, r); - case INDEX_op_qemu_ld_a64_i64: - return TCG_TARGET_REG_BITS == 64 ? C_O1_I1(r, r) : C_O2_I2(r, r, r, r); - case INDEX_op_qemu_st_a32_i32: + case INDEX_op_qemu_st_i32: return C_O0_I2(r, r); - case INDEX_op_qemu_st_a64_i32: + case INDEX_op_qemu_st_i64: return TCG_TARGET_REG_BITS == 64 ? C_O0_I2(r, r) : C_O0_I3(r, r, r); - case INDEX_op_qemu_st_a32_i64: - return TCG_TARGET_REG_BITS == 64 ? C_O0_I2(r, r) : C_O0_I3(r, r, r); - case INDEX_op_qemu_st_a64_i64: - return TCG_TARGET_REG_BITS == 64 ? C_O0_I2(r, r) : C_O0_I4(r, r, r, r); default: return C_NotImplemented; @@ -XXX,XX +XXX,XX @@ static void tcg_out_op_rrrbb(TCGContext *s, TCGOpcode op, TCGReg r0, tcg_out32(s, insn); } -static void tcg_out_op_rrrrr(TCGContext *s, TCGOpcode op, TCGReg r0, - TCGReg r1, TCGReg r2, TCGReg r3, TCGReg r4) -{ - tcg_insn_unit insn = 0; - - insn = deposit32(insn, 0, 8, op); - insn = deposit32(insn, 8, 4, r0); - insn = deposit32(insn, 12, 4, r1); - insn = deposit32(insn, 16, 4, r2); - insn = deposit32(insn, 20, 4, r3); - insn = deposit32(insn, 24, 4, r4); - tcg_out32(s, insn); -} - static void tcg_out_op_rrrr(TCGContext *s, TCGOpcode op, TCGReg r0, TCGReg r1, TCGReg r2, TCGReg r3) { @@ -XXX,XX +XXX,XX @@ static void tcg_out_op(TCGContext *s, TCGOpcode opc, TCGType type, tcg_out_op_rrrr(s, opc, args[0], args[1], args[2], args[3]); break; - case INDEX_op_qemu_ld_a32_i32: - case INDEX_op_qemu_st_a32_i32: - tcg_out_op_rrm(s, opc, args[0], args[1], args[2]); - break; - case INDEX_op_qemu_ld_a64_i32: - case INDEX_op_qemu_st_a64_i32: - case INDEX_op_qemu_ld_a32_i64: - case INDEX_op_qemu_st_a32_i64: - if (TCG_TARGET_REG_BITS == 64) { - tcg_out_op_rrm(s, opc, args[0], args[1], args[2]); - } else { + case INDEX_op_qemu_ld_i64: + case INDEX_op_qemu_st_i64: + if (TCG_TARGET_REG_BITS == 32) { tcg_out_movi(s, TCG_TYPE_I32, TCG_REG_TMP, args[3]); tcg_out_op_rrrr(s, opc, args[0], args[1], args[2], TCG_REG_TMP); + break; } - break; - case INDEX_op_qemu_ld_a64_i64: - case INDEX_op_qemu_st_a64_i64: - if (TCG_TARGET_REG_BITS == 64) { - tcg_out_op_rrm(s, opc, args[0], args[1], args[2]); + /* fall through */ + case INDEX_op_qemu_ld_i32: + case INDEX_op_qemu_st_i32: + if (TCG_TARGET_REG_BITS == 64 && s->addr_type == TCG_TYPE_I32) { + tcg_out_ext32u(s, TCG_REG_TMP, args[1]); + tcg_out_op_rrm(s, opc, args[0], TCG_REG_TMP, args[2]); } else { - tcg_out_movi(s, TCG_TYPE_I32, TCG_REG_TMP, args[4]); - tcg_out_op_rrrrr(s, opc, args[0], args[1], - args[2], args[3], TCG_REG_TMP); + tcg_out_op_rrm(s, opc, args[0], args[1], args[2]); } break; -- 2.43.0 The guest address will now always be TCG_TYPE_I32. Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- tcg/arm/tcg-target.c.inc | 63 ++++++++++++++-------------------------- 1 file changed, 21 insertions(+), 42 deletions(-) diff --git a/tcg/arm/tcg-target.c.inc b/tcg/arm/tcg-target.c.inc index XXXXXXX..XXXXXXX 100644 --- a/tcg/arm/tcg-target.c.inc +++ b/tcg/arm/tcg-target.c.inc @@ -XXX,XX +XXX,XX @@ static bool tcg_out_qemu_st_slow_path(TCGContext *s, TCGLabelQemuLdst *lb) #define MIN_TLB_MASK_TABLE_OFS -256 static TCGLabelQemuLdst *prepare_host_addr(TCGContext *s, HostAddress *h, - TCGReg addrlo, TCGReg addrhi, - MemOpIdx oi, bool is_ld) + TCGReg addr, MemOpIdx oi, bool is_ld) { TCGLabelQemuLdst *ldst = NULL; MemOp opc = get_memop(oi); @@ -XXX,XX +XXX,XX @@ static TCGLabelQemuLdst *prepare_host_addr(TCGContext *s, HostAddress *h, if (tcg_use_softmmu) { *h = (HostAddress){ .cond = COND_AL, - .base = addrlo, + .base = addr, .index = TCG_REG_R1, .index_scratch = true, }; } else { *h = (HostAddress){ .cond = COND_AL, - .base = addrlo, + .base = addr, .index = guest_base ? TCG_REG_GUEST_BASE : -1, .index_scratch = false, }; @@ -XXX,XX +XXX,XX @@ static TCGLabelQemuLdst *prepare_host_addr(TCGContext *s, HostAddress *h, ldst = new_ldst_label(s); ldst->is_ld = is_ld; ldst->oi = oi; - ldst->addrlo_reg = addrlo; - ldst->addrhi_reg = addrhi; + ldst->addrlo_reg = addr; /* Load cpu->neg.tlb.f[mmu_idx].{mask,table} into {r0,r1}. */ QEMU_BUILD_BUG_ON(offsetof(CPUTLBDescFast, mask) != 0); @@ -XXX,XX +XXX,XX @@ static TCGLabelQemuLdst *prepare_host_addr(TCGContext *s, HostAddress *h, tcg_out_ldrd_8(s, COND_AL, TCG_REG_R0, TCG_AREG0, fast_off); /* Extract the tlb index from the address into R0. */ - tcg_out_dat_reg(s, COND_AL, ARITH_AND, TCG_REG_R0, TCG_REG_R0, addrlo, + tcg_out_dat_reg(s, COND_AL, ARITH_AND, TCG_REG_R0, TCG_REG_R0, addr, SHIFT_IMM_LSR(s->page_bits - CPU_TLB_ENTRY_BITS)); /* * Add the tlb_table pointer, creating the CPUTLBEntry address in R1. - * Load the tlb comparator into R2/R3 and the fast path addend into R1. + * Load the tlb comparator into R2 and the fast path addend into R1. */ QEMU_BUILD_BUG_ON(HOST_BIG_ENDIAN); if (cmp_off == 0) { - if (s->addr_type == TCG_TYPE_I32) { - tcg_out_ld32_rwb(s, COND_AL, TCG_REG_R2, - TCG_REG_R1, TCG_REG_R0); - } else { - tcg_out_ldrd_rwb(s, COND_AL, TCG_REG_R2, - TCG_REG_R1, TCG_REG_R0); - } + tcg_out_ld32_rwb(s, COND_AL, TCG_REG_R2, TCG_REG_R1, TCG_REG_R0); } else { tcg_out_dat_reg(s, COND_AL, ARITH_ADD, TCG_REG_R1, TCG_REG_R1, TCG_REG_R0, 0); - if (s->addr_type == TCG_TYPE_I32) { - tcg_out_ld32_12(s, COND_AL, TCG_REG_R2, TCG_REG_R1, cmp_off); - } else { - tcg_out_ldrd_8(s, COND_AL, TCG_REG_R2, TCG_REG_R1, cmp_off); - } + tcg_out_ld32_12(s, COND_AL, TCG_REG_R2, TCG_REG_R1, cmp_off); } /* Load the tlb addend. */ @@ -XXX,XX +XXX,XX @@ static TCGLabelQemuLdst *prepare_host_addr(TCGContext *s, HostAddress *h, * This leaves the least significant alignment bits unchanged, and of * course must be zero. */ - t_addr = addrlo; + t_addr = addr; if (a_mask < s_mask) { t_addr = TCG_REG_R0; tcg_out_dat_imm(s, COND_AL, ARITH_ADD, t_addr, - addrlo, s_mask - a_mask); + addr, s_mask - a_mask); } if (use_armv7_instructions && s->page_bits <= 16) { tcg_out_movi32(s, COND_AL, TCG_REG_TMP, ~(s->page_mask | a_mask)); @@ -XXX,XX +XXX,XX @@ static TCGLabelQemuLdst *prepare_host_addr(TCGContext *s, HostAddress *h, } else { if (a_mask) { tcg_debug_assert(a_mask <= 0xff); - tcg_out_dat_imm(s, COND_AL, ARITH_TST, 0, addrlo, a_mask); + tcg_out_dat_imm(s, COND_AL, ARITH_TST, 0, addr, a_mask); } tcg_out_dat_reg(s, COND_AL, ARITH_MOV, TCG_REG_TMP, 0, t_addr, SHIFT_IMM_LSR(s->page_bits)); @@ -XXX,XX +XXX,XX @@ static TCGLabelQemuLdst *prepare_host_addr(TCGContext *s, HostAddress *h, 0, TCG_REG_R2, TCG_REG_TMP, SHIFT_IMM_LSL(s->page_bits)); } - - if (s->addr_type != TCG_TYPE_I32) { - tcg_out_dat_reg(s, COND_EQ, ARITH_CMP, 0, TCG_REG_R3, addrhi, 0); - } } else if (a_mask) { ldst = new_ldst_label(s); ldst->is_ld = is_ld; ldst->oi = oi; - ldst->addrlo_reg = addrlo; - ldst->addrhi_reg = addrhi; + ldst->addrlo_reg = addr; /* We are expecting alignment to max out at 7 */ tcg_debug_assert(a_mask <= 0xff); /* tst addr, #mask */ - tcg_out_dat_imm(s, COND_AL, ARITH_TST, 0, addrlo, a_mask); + tcg_out_dat_imm(s, COND_AL, ARITH_TST, 0, addr, a_mask); } return ldst; @@ -XXX,XX +XXX,XX @@ static void tcg_out_qemu_ld_direct(TCGContext *s, MemOp opc, TCGReg datalo, } static void tcg_out_qemu_ld(TCGContext *s, TCGReg datalo, TCGReg datahi, - TCGReg addrlo, TCGReg addrhi, - MemOpIdx oi, TCGType data_type) + TCGReg addr, MemOpIdx oi, TCGType data_type) { MemOp opc = get_memop(oi); TCGLabelQemuLdst *ldst; HostAddress h; - ldst = prepare_host_addr(s, &h, addrlo, addrhi, oi, true); + ldst = prepare_host_addr(s, &h, addr, oi, true); if (ldst) { ldst->type = data_type; ldst->datalo_reg = datalo; @@ -XXX,XX +XXX,XX @@ static void tcg_out_qemu_st_direct(TCGContext *s, MemOp opc, TCGReg datalo, } static void tcg_out_qemu_st(TCGContext *s, TCGReg datalo, TCGReg datahi, - TCGReg addrlo, TCGReg addrhi, - MemOpIdx oi, TCGType data_type) + TCGReg addr, MemOpIdx oi, TCGType data_type) { MemOp opc = get_memop(oi); TCGLabelQemuLdst *ldst; HostAddress h; - ldst = prepare_host_addr(s, &h, addrlo, addrhi, oi, false); + ldst = prepare_host_addr(s, &h, addr, oi, false); if (ldst) { ldst->type = data_type; ldst->datalo_reg = datalo; @@ -XXX,XX +XXX,XX @@ static void tcg_out_op(TCGContext *s, TCGOpcode opc, TCGType type, break; case INDEX_op_qemu_ld_i32: - tcg_out_qemu_ld(s, args[0], -1, args[1], -1, args[2], TCG_TYPE_I32); + tcg_out_qemu_ld(s, args[0], -1, args[1], args[2], TCG_TYPE_I32); break; case INDEX_op_qemu_ld_i64: - tcg_out_qemu_ld(s, args[0], args[1], args[2], -1, - args[3], TCG_TYPE_I64); + tcg_out_qemu_ld(s, args[0], args[1], args[2], args[3], TCG_TYPE_I64); break; case INDEX_op_qemu_st_i32: - tcg_out_qemu_st(s, args[0], -1, args[1], -1, args[2], TCG_TYPE_I32); + tcg_out_qemu_st(s, args[0], -1, args[1], args[2], TCG_TYPE_I32); break; case INDEX_op_qemu_st_i64: - tcg_out_qemu_st(s, args[0], args[1], args[2], -1, - args[3], TCG_TYPE_I64); + tcg_out_qemu_st(s, args[0], args[1], args[2], args[3], TCG_TYPE_I64); break; case INDEX_op_bswap16_i32: -- 2.43.0 The guest address will now always fit in one register. Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- tcg/i386/tcg-target.c.inc | 56 ++++++++++++++------------------------- 1 file changed, 20 insertions(+), 36 deletions(-) diff --git a/tcg/i386/tcg-target.c.inc b/tcg/i386/tcg-target.c.inc index XXXXXXX..XXXXXXX 100644 --- a/tcg/i386/tcg-target.c.inc +++ b/tcg/i386/tcg-target.c.inc @@ -XXX,XX +XXX,XX @@ static inline int setup_guest_base_seg(void) * is required and fill in @h with the host address for the fast path. */ static TCGLabelQemuLdst *prepare_host_addr(TCGContext *s, HostAddress *h, - TCGReg addrlo, TCGReg addrhi, - MemOpIdx oi, bool is_ld) + TCGReg addr, MemOpIdx oi, bool is_ld) { TCGLabelQemuLdst *ldst = NULL; MemOp opc = get_memop(oi); @@ -XXX,XX +XXX,XX @@ static TCGLabelQemuLdst *prepare_host_addr(TCGContext *s, HostAddress *h, } else { *h = x86_guest_base; } - h->base = addrlo; + h->base = addr; h->aa = atom_and_align_for_opc(s, opc, MO_ATOM_IFALIGN, s_bits == MO_128); a_mask = (1 << h->aa.align) - 1; @@ -XXX,XX +XXX,XX @@ static TCGLabelQemuLdst *prepare_host_addr(TCGContext *s, HostAddress *h, ldst = new_ldst_label(s); ldst->is_ld = is_ld; ldst->oi = oi; - ldst->addrlo_reg = addrlo; - ldst->addrhi_reg = addrhi; + ldst->addrlo_reg = addr; if (TCG_TARGET_REG_BITS == 64) { ttype = s->addr_type; @@ -XXX,XX +XXX,XX @@ static TCGLabelQemuLdst *prepare_host_addr(TCGContext *s, HostAddress *h, } } - tcg_out_mov(s, tlbtype, TCG_REG_L0, addrlo); + tcg_out_mov(s, tlbtype, TCG_REG_L0, addr); tcg_out_shifti(s, SHIFT_SHR + tlbrexw, TCG_REG_L0, s->page_bits - CPU_TLB_ENTRY_BITS); @@ -XXX,XX +XXX,XX @@ static TCGLabelQemuLdst *prepare_host_addr(TCGContext *s, HostAddress *h, * check that we don't cross pages for the complete access. */ if (a_mask >= s_mask) { - tcg_out_mov(s, ttype, TCG_REG_L1, addrlo); + tcg_out_mov(s, ttype, TCG_REG_L1, addr); } else { tcg_out_modrm_offset(s, OPC_LEA + trexw, TCG_REG_L1, - addrlo, s_mask - a_mask); + addr, s_mask - a_mask); } tlb_mask = s->page_mask | a_mask; tgen_arithi(s, ARITH_AND + trexw, TCG_REG_L1, tlb_mask, 0); @@ -XXX,XX +XXX,XX @@ static TCGLabelQemuLdst *prepare_host_addr(TCGContext *s, HostAddress *h, ldst->label_ptr[0] = s->code_ptr; s->code_ptr += 4; - if (TCG_TARGET_REG_BITS == 32 && s->addr_type == TCG_TYPE_I64) { - /* cmp 4(TCG_REG_L0), addrhi */ - tcg_out_modrm_offset(s, OPC_CMP_GvEv, addrhi, - TCG_REG_L0, cmp_ofs + 4); - - /* jne slow_path */ - tcg_out_opc(s, OPC_JCC_long + JCC_JNE, 0, 0, 0); - ldst->label_ptr[1] = s->code_ptr; - s->code_ptr += 4; - } - /* TLB Hit. */ tcg_out_ld(s, TCG_TYPE_PTR, TCG_REG_L0, TCG_REG_L0, offsetof(CPUTLBEntry, addend)); @@ -XXX,XX +XXX,XX @@ static TCGLabelQemuLdst *prepare_host_addr(TCGContext *s, HostAddress *h, ldst = new_ldst_label(s); ldst->is_ld = is_ld; ldst->oi = oi; - ldst->addrlo_reg = addrlo; - ldst->addrhi_reg = addrhi; + ldst->addrlo_reg = addr; /* jne slow_path */ - jcc = tcg_out_cmp(s, TCG_COND_TSTNE, addrlo, a_mask, true, false); + jcc = tcg_out_cmp(s, TCG_COND_TSTNE, addr, a_mask, true, false); tcg_out_opc(s, OPC_JCC_long + jcc, 0, 0, 0); ldst->label_ptr[0] = s->code_ptr; s->code_ptr += 4; @@ -XXX,XX +XXX,XX @@ static void tcg_out_qemu_ld_direct(TCGContext *s, TCGReg datalo, TCGReg datahi, } static void tcg_out_qemu_ld(TCGContext *s, TCGReg datalo, TCGReg datahi, - TCGReg addrlo, TCGReg addrhi, - MemOpIdx oi, TCGType data_type) + TCGReg addr, MemOpIdx oi, TCGType data_type) { TCGLabelQemuLdst *ldst; HostAddress h; - ldst = prepare_host_addr(s, &h, addrlo, addrhi, oi, true); + ldst = prepare_host_addr(s, &h, addr, oi, true); tcg_out_qemu_ld_direct(s, datalo, datahi, h, data_type, get_memop(oi)); if (ldst) { @@ -XXX,XX +XXX,XX @@ static void tcg_out_qemu_st_direct(TCGContext *s, TCGReg datalo, TCGReg datahi, } static void tcg_out_qemu_st(TCGContext *s, TCGReg datalo, TCGReg datahi, - TCGReg addrlo, TCGReg addrhi, - MemOpIdx oi, TCGType data_type) + TCGReg addr, MemOpIdx oi, TCGType data_type) { TCGLabelQemuLdst *ldst; HostAddress h; - ldst = prepare_host_addr(s, &h, addrlo, addrhi, oi, false); + ldst = prepare_host_addr(s, &h, addr, oi, false); tcg_out_qemu_st_direct(s, datalo, datahi, h, get_memop(oi)); if (ldst) { @@ -XXX,XX +XXX,XX @@ static void tcg_out_op(TCGContext *s, TCGOpcode opc, TCGType type, break; case INDEX_op_qemu_ld_i32: - tcg_out_qemu_ld(s, a0, -1, a1, -1, a2, TCG_TYPE_I32); + tcg_out_qemu_ld(s, a0, -1, a1, a2, TCG_TYPE_I32); break; case INDEX_op_qemu_ld_i64: if (TCG_TARGET_REG_BITS == 64) { - tcg_out_qemu_ld(s, a0, -1, a1, -1, a2, TCG_TYPE_I64); + tcg_out_qemu_ld(s, a0, -1, a1, a2, TCG_TYPE_I64); } else { - tcg_out_qemu_ld(s, a0, a1, a2, -1, args[3], TCG_TYPE_I64); + tcg_out_qemu_ld(s, a0, a1, a2, args[3], TCG_TYPE_I64); } break; case INDEX_op_qemu_ld_i128: tcg_debug_assert(TCG_TARGET_REG_BITS == 64); - tcg_out_qemu_ld(s, a0, a1, a2, -1, args[3], TCG_TYPE_I128); + tcg_out_qemu_ld(s, a0, a1, a2, args[3], TCG_TYPE_I128); break; case INDEX_op_qemu_st_i32: case INDEX_op_qemu_st8_i32: - tcg_out_qemu_st(s, a0, -1, a1, -1, a2, TCG_TYPE_I32); + tcg_out_qemu_st(s, a0, -1, a1, a2, TCG_TYPE_I32); break; case INDEX_op_qemu_st_i64: if (TCG_TARGET_REG_BITS == 64) { - tcg_out_qemu_st(s, a0, -1, a1, -1, a2, TCG_TYPE_I64); + tcg_out_qemu_st(s, a0, -1, a1, a2, TCG_TYPE_I64); } else { - tcg_out_qemu_st(s, a0, a1, a2, -1, args[3], TCG_TYPE_I64); + tcg_out_qemu_st(s, a0, a1, a2, args[3], TCG_TYPE_I64); } break; case INDEX_op_qemu_st_i128: tcg_debug_assert(TCG_TARGET_REG_BITS == 64); - tcg_out_qemu_st(s, a0, a1, a2, -1, args[3], TCG_TYPE_I128); + tcg_out_qemu_st(s, a0, a1, a2, args[3], TCG_TYPE_I128); break; OP_32_64(mulu2): -- 2.43.0 The guest address will now always fit in one register. Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- tcg/mips/tcg-target.c.inc | 62 ++++++++++++++------------------------- 1 file changed, 22 insertions(+), 40 deletions(-) diff --git a/tcg/mips/tcg-target.c.inc b/tcg/mips/tcg-target.c.inc index XXXXXXX..XXXXXXX 100644 --- a/tcg/mips/tcg-target.c.inc +++ b/tcg/mips/tcg-target.c.inc @@ -XXX,XX +XXX,XX @@ bool tcg_target_has_memory_bswap(MemOp memop) * is required and fill in @h with the host address for the fast path. */ static TCGLabelQemuLdst *prepare_host_addr(TCGContext *s, HostAddress *h, - TCGReg addrlo, TCGReg addrhi, - MemOpIdx oi, bool is_ld) + TCGReg addr, MemOpIdx oi, bool is_ld) { TCGType addr_type = s->addr_type; TCGLabelQemuLdst *ldst = NULL; @@ -XXX,XX +XXX,XX @@ static TCGLabelQemuLdst *prepare_host_addr(TCGContext *s, HostAddress *h, ldst = new_ldst_label(s); ldst->is_ld = is_ld; ldst->oi = oi; - ldst->addrlo_reg = addrlo; - ldst->addrhi_reg = addrhi; + ldst->addrlo_reg = addr; /* Load tlb_mask[mmu_idx] and tlb_table[mmu_idx]. */ tcg_out_ld(s, TCG_TYPE_PTR, TCG_TMP0, TCG_AREG0, mask_off); @@ -XXX,XX +XXX,XX @@ static TCGLabelQemuLdst *prepare_host_addr(TCGContext *s, HostAddress *h, /* Extract the TLB index from the address into TMP3. */ if (TCG_TARGET_REG_BITS == 32 || addr_type == TCG_TYPE_I32) { - tcg_out_opc_sa(s, OPC_SRL, TCG_TMP3, addrlo, + tcg_out_opc_sa(s, OPC_SRL, TCG_TMP3, addr, s->page_bits - CPU_TLB_ENTRY_BITS); } else { - tcg_out_dsrl(s, TCG_TMP3, addrlo, - s->page_bits - CPU_TLB_ENTRY_BITS); + tcg_out_dsrl(s, TCG_TMP3, addr, s->page_bits - CPU_TLB_ENTRY_BITS); } tcg_out_opc_reg(s, OPC_AND, TCG_TMP3, TCG_TMP3, TCG_TMP0); @@ -XXX,XX +XXX,XX @@ static TCGLabelQemuLdst *prepare_host_addr(TCGContext *s, HostAddress *h, tcg_out_opc_imm(s, (TCG_TARGET_REG_BITS == 32 || addr_type == TCG_TYPE_I32 ? OPC_ADDIU : OPC_DADDIU), - TCG_TMP2, addrlo, s_mask - a_mask); + TCG_TMP2, addr, s_mask - a_mask); tcg_out_opc_reg(s, OPC_AND, TCG_TMP1, TCG_TMP1, TCG_TMP2); } else { - tcg_out_opc_reg(s, OPC_AND, TCG_TMP1, TCG_TMP1, addrlo); + tcg_out_opc_reg(s, OPC_AND, TCG_TMP1, TCG_TMP1, addr); } /* Zero extend a 32-bit guest address for a 64-bit host. */ if (TCG_TARGET_REG_BITS == 64 && addr_type == TCG_TYPE_I32) { - tcg_out_ext32u(s, TCG_TMP2, addrlo); - addrlo = TCG_TMP2; + tcg_out_ext32u(s, TCG_TMP2, addr); + addr = TCG_TMP2; } ldst->label_ptr[0] = s->code_ptr; tcg_out_opc_br(s, OPC_BNE, TCG_TMP1, TCG_TMP0); - /* Load and test the high half tlb comparator. */ - if (TCG_TARGET_REG_BITS == 32 && addr_type != TCG_TYPE_I32) { - /* delay slot */ - tcg_out_ldst(s, OPC_LW, TCG_TMP0, TCG_TMP3, cmp_off + HI_OFF); - - /* Load the tlb addend for the fast path. */ - tcg_out_ld(s, TCG_TYPE_PTR, TCG_TMP3, TCG_TMP3, add_off); - - ldst->label_ptr[1] = s->code_ptr; - tcg_out_opc_br(s, OPC_BNE, addrhi, TCG_TMP0); - } - /* delay slot */ base = TCG_TMP3; - tcg_out_opc_reg(s, ALIAS_PADD, base, TCG_TMP3, addrlo); + tcg_out_opc_reg(s, ALIAS_PADD, base, TCG_TMP3, addr); } else { if (a_mask && (use_mips32r6_instructions || a_bits != s_bits)) { ldst = new_ldst_label(s); ldst->is_ld = is_ld; ldst->oi = oi; - ldst->addrlo_reg = addrlo; - ldst->addrhi_reg = addrhi; + ldst->addrlo_reg = addr; /* We are expecting a_bits to max out at 7, much lower than ANDI. */ tcg_debug_assert(a_bits < 16); - tcg_out_opc_imm(s, OPC_ANDI, TCG_TMP0, addrlo, a_mask); + tcg_out_opc_imm(s, OPC_ANDI, TCG_TMP0, addr, a_mask); ldst->label_ptr[0] = s->code_ptr; if (use_mips32r6_instructions) { @@ -XXX,XX +XXX,XX @@ static TCGLabelQemuLdst *prepare_host_addr(TCGContext *s, HostAddress *h, } } - base = addrlo; + base = addr; if (TCG_TARGET_REG_BITS == 64 && addr_type == TCG_TYPE_I32) { tcg_out_ext32u(s, TCG_REG_A0, base); base = TCG_REG_A0; @@ -XXX,XX +XXX,XX @@ static void tcg_out_qemu_ld_unalign(TCGContext *s, TCGReg lo, TCGReg hi, } static void tcg_out_qemu_ld(TCGContext *s, TCGReg datalo, TCGReg datahi, - TCGReg addrlo, TCGReg addrhi, - MemOpIdx oi, TCGType data_type) + TCGReg addr, MemOpIdx oi, TCGType data_type) { MemOp opc = get_memop(oi); TCGLabelQemuLdst *ldst; HostAddress h; - ldst = prepare_host_addr(s, &h, addrlo, addrhi, oi, true); + ldst = prepare_host_addr(s, &h, addr, oi, true); if (use_mips32r6_instructions || h.aa.align >= (opc & MO_SIZE)) { tcg_out_qemu_ld_direct(s, datalo, datahi, h.base, opc, data_type); @@ -XXX,XX +XXX,XX @@ static void tcg_out_qemu_st_unalign(TCGContext *s, TCGReg lo, TCGReg hi, } static void tcg_out_qemu_st(TCGContext *s, TCGReg datalo, TCGReg datahi, - TCGReg addrlo, TCGReg addrhi, - MemOpIdx oi, TCGType data_type) + TCGReg addr, MemOpIdx oi, TCGType data_type) { MemOp opc = get_memop(oi); TCGLabelQemuLdst *ldst; HostAddress h; - ldst = prepare_host_addr(s, &h, addrlo, addrhi, oi, false); + ldst = prepare_host_addr(s, &h, addr, oi, false); if (use_mips32r6_instructions || h.aa.align >= (opc & MO_SIZE)) { tcg_out_qemu_st_direct(s, datalo, datahi, h.base, opc); @@ -XXX,XX +XXX,XX @@ static void tcg_out_op(TCGContext *s, TCGOpcode opc, TCGType type, break; case INDEX_op_qemu_ld_i32: - tcg_out_qemu_ld(s, a0, 0, a1, 0, a2, TCG_TYPE_I32); + tcg_out_qemu_ld(s, a0, 0, a1, a2, TCG_TYPE_I32); break; case INDEX_op_qemu_ld_i64: if (TCG_TARGET_REG_BITS == 64) { - tcg_out_qemu_ld(s, a0, 0, a1, 0, a2, TCG_TYPE_I64); + tcg_out_qemu_ld(s, a0, 0, a1, a2, TCG_TYPE_I64); } else { - tcg_out_qemu_ld(s, a0, a1, a2, 0, args[3], TCG_TYPE_I64); + tcg_out_qemu_ld(s, a0, a1, a2, args[3], TCG_TYPE_I64); } break; case INDEX_op_qemu_st_i32: - tcg_out_qemu_st(s, a0, 0, a1, 0, a2, TCG_TYPE_I32); + tcg_out_qemu_st(s, a0, 0, a1, a2, TCG_TYPE_I32); break; case INDEX_op_qemu_st_i64: if (TCG_TARGET_REG_BITS == 64) { - tcg_out_qemu_st(s, a0, 0, a1, 0, a2, TCG_TYPE_I64); + tcg_out_qemu_st(s, a0, 0, a1, a2, TCG_TYPE_I64); } else { - tcg_out_qemu_st(s, a0, a1, a2, 0, args[3], TCG_TYPE_I64); + tcg_out_qemu_st(s, a0, a1, a2, args[3], TCG_TYPE_I64); } break; -- 2.43.0 The guest address will now always fit in one register. Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- tcg/ppc/tcg-target.c.inc | 75 ++++++++++++---------------------------- 1 file changed, 23 insertions(+), 52 deletions(-) diff --git a/tcg/ppc/tcg-target.c.inc b/tcg/ppc/tcg-target.c.inc index XXXXXXX..XXXXXXX 100644 --- a/tcg/ppc/tcg-target.c.inc +++ b/tcg/ppc/tcg-target.c.inc @@ -XXX,XX +XXX,XX @@ bool tcg_target_has_memory_bswap(MemOp memop) * is required and fill in @h with the host address for the fast path. */ static TCGLabelQemuLdst *prepare_host_addr(TCGContext *s, HostAddress *h, - TCGReg addrlo, TCGReg addrhi, - MemOpIdx oi, bool is_ld) + TCGReg addr, MemOpIdx oi, bool is_ld) { TCGType addr_type = s->addr_type; TCGLabelQemuLdst *ldst = NULL; @@ -XXX,XX +XXX,XX @@ static TCGLabelQemuLdst *prepare_host_addr(TCGContext *s, HostAddress *h, ldst = new_ldst_label(s); ldst->is_ld = is_ld; ldst->oi = oi; - ldst->addrlo_reg = addrlo; - ldst->addrhi_reg = addrhi; + ldst->addrlo_reg = addr; /* Load tlb_mask[mmu_idx] and tlb_table[mmu_idx]. */ tcg_out_ld(s, TCG_TYPE_PTR, TCG_REG_TMP1, TCG_AREG0, mask_off); @@ -XXX,XX +XXX,XX @@ static TCGLabelQemuLdst *prepare_host_addr(TCGContext *s, HostAddress *h, /* Extract the page index, shifted into place for tlb index. */ if (TCG_TARGET_REG_BITS == 32) { - tcg_out_shri32(s, TCG_REG_R0, addrlo, + tcg_out_shri32(s, TCG_REG_R0, addr, s->page_bits - CPU_TLB_ENTRY_BITS); } else { - tcg_out_shri64(s, TCG_REG_R0, addrlo, + tcg_out_shri64(s, TCG_REG_R0, addr, s->page_bits - CPU_TLB_ENTRY_BITS); } tcg_out32(s, AND | SAB(TCG_REG_TMP1, TCG_REG_TMP1, TCG_REG_R0)); @@ -XXX,XX +XXX,XX @@ static TCGLabelQemuLdst *prepare_host_addr(TCGContext *s, HostAddress *h, if (a_bits < s_bits) { a_bits = s_bits; } - tcg_out_rlw(s, RLWINM, TCG_REG_R0, addrlo, 0, + tcg_out_rlw(s, RLWINM, TCG_REG_R0, addr, 0, (32 - a_bits) & 31, 31 - s->page_bits); } else { - TCGReg t = addrlo; + TCGReg t = addr; /* * If the access is unaligned, we need to make sure we fail if we @@ -XXX,XX +XXX,XX @@ static TCGLabelQemuLdst *prepare_host_addr(TCGContext *s, HostAddress *h, } } - if (TCG_TARGET_REG_BITS == 32 && addr_type != TCG_TYPE_I32) { - /* Low part comparison into cr7. */ - tcg_out_cmp(s, TCG_COND_EQ, TCG_REG_R0, TCG_REG_TMP2, - 0, 7, TCG_TYPE_I32); - - /* Load the high part TLB comparator into TMP2. */ - tcg_out_ld(s, TCG_TYPE_I32, TCG_REG_TMP2, TCG_REG_TMP1, - cmp_off + 4 * !HOST_BIG_ENDIAN); - - /* Load addend, deferred for this case. */ - tcg_out_ld(s, TCG_TYPE_PTR, TCG_REG_TMP1, TCG_REG_TMP1, - offsetof(CPUTLBEntry, addend)); - - /* High part comparison into cr6. */ - tcg_out_cmp(s, TCG_COND_EQ, addrhi, TCG_REG_TMP2, - 0, 6, TCG_TYPE_I32); - - /* Combine comparisons into cr0. */ - tcg_out32(s, CRAND | BT(0, CR_EQ) | BA(6, CR_EQ) | BB(7, CR_EQ)); - } else { - /* Full comparison into cr0. */ - tcg_out_cmp(s, TCG_COND_EQ, TCG_REG_R0, TCG_REG_TMP2, - 0, 0, addr_type); - } + /* Full comparison into cr0. */ + tcg_out_cmp(s, TCG_COND_EQ, TCG_REG_R0, TCG_REG_TMP2, 0, 0, addr_type); /* Load a pointer into the current opcode w/conditional branch-link. */ ldst->label_ptr[0] = s->code_ptr; @@ -XXX,XX +XXX,XX @@ static TCGLabelQemuLdst *prepare_host_addr(TCGContext *s, HostAddress *h, ldst = new_ldst_label(s); ldst->is_ld = is_ld; ldst->oi = oi; - ldst->addrlo_reg = addrlo; - ldst->addrhi_reg = addrhi; + ldst->addrlo_reg = addr; /* We are expecting a_bits to max out at 7, much lower than ANDI. */ tcg_debug_assert(a_bits < 16); - tcg_out32(s, ANDI | SAI(addrlo, TCG_REG_R0, (1 << a_bits) - 1)); + tcg_out32(s, ANDI | SAI(addr, TCG_REG_R0, (1 << a_bits) - 1)); ldst->label_ptr[0] = s->code_ptr; tcg_out32(s, BC | BI(0, CR_EQ) | BO_COND_FALSE | LK); @@ -XXX,XX +XXX,XX @@ static TCGLabelQemuLdst *prepare_host_addr(TCGContext *s, HostAddress *h, if (TCG_TARGET_REG_BITS == 64 && addr_type == TCG_TYPE_I32) { /* Zero-extend the guest address for use in the host address. */ - tcg_out_ext32u(s, TCG_REG_TMP2, addrlo); + tcg_out_ext32u(s, TCG_REG_TMP2, addr); h->index = TCG_REG_TMP2; } else { - h->index = addrlo; + h->index = addr; } return ldst; } static void tcg_out_qemu_ld(TCGContext *s, TCGReg datalo, TCGReg datahi, - TCGReg addrlo, TCGReg addrhi, - MemOpIdx oi, TCGType data_type) + TCGReg addr, MemOpIdx oi, TCGType data_type) { MemOp opc = get_memop(oi); TCGLabelQemuLdst *ldst; HostAddress h; - ldst = prepare_host_addr(s, &h, addrlo, addrhi, oi, true); + ldst = prepare_host_addr(s, &h, addr, oi, true); if (TCG_TARGET_REG_BITS == 32 && (opc & MO_SIZE) == MO_64) { if (opc & MO_BSWAP) { @@ -XXX,XX +XXX,XX @@ static void tcg_out_qemu_ld(TCGContext *s, TCGReg datalo, TCGReg datahi, } static void tcg_out_qemu_st(TCGContext *s, TCGReg datalo, TCGReg datahi, - TCGReg addrlo, TCGReg addrhi, - MemOpIdx oi, TCGType data_type) + TCGReg addr, MemOpIdx oi, TCGType data_type) { MemOp opc = get_memop(oi); TCGLabelQemuLdst *ldst; HostAddress h; - ldst = prepare_host_addr(s, &h, addrlo, addrhi, oi, false); + ldst = prepare_host_addr(s, &h, addr, oi, false); if (TCG_TARGET_REG_BITS == 32 && (opc & MO_SIZE) == MO_64) { if (opc & MO_BSWAP) { @@ -XXX,XX +XXX,XX @@ static void tcg_out_qemu_ldst_i128(TCGContext *s, TCGReg datalo, TCGReg datahi, uint32_t insn; TCGReg index; - ldst = prepare_host_addr(s, &h, addr_reg, -1, oi, is_ld); + ldst = prepare_host_addr(s, &h, addr_reg, oi, is_ld); /* Compose the final address, as LQ/STQ have no indexing. */ index = h.index; @@ -XXX,XX +XXX,XX @@ static void tcg_out_op(TCGContext *s, TCGOpcode opc, TCGType type, break; case INDEX_op_qemu_ld_i32: - tcg_out_qemu_ld(s, args[0], -1, args[1], -1, args[2], TCG_TYPE_I32); + tcg_out_qemu_ld(s, args[0], -1, args[1], args[2], TCG_TYPE_I32); break; case INDEX_op_qemu_ld_i64: if (TCG_TARGET_REG_BITS == 64) { - tcg_out_qemu_ld(s, args[0], -1, args[1], -1, - args[2], TCG_TYPE_I64); + tcg_out_qemu_ld(s, args[0], -1, args[1], args[2], TCG_TYPE_I64); } else { - tcg_out_qemu_ld(s, args[0], args[1], args[2], -1, + tcg_out_qemu_ld(s, args[0], args[1], args[2], args[3], TCG_TYPE_I64); } break; @@ -XXX,XX +XXX,XX @@ static void tcg_out_op(TCGContext *s, TCGOpcode opc, TCGType type, break; case INDEX_op_qemu_st_i32: - tcg_out_qemu_st(s, args[0], -1, args[1], -1, args[2], TCG_TYPE_I32); + tcg_out_qemu_st(s, args[0], -1, args[1], args[2], TCG_TYPE_I32); break; case INDEX_op_qemu_st_i64: if (TCG_TARGET_REG_BITS == 64) { - tcg_out_qemu_st(s, args[0], -1, args[1], -1, - args[2], TCG_TYPE_I64); + tcg_out_qemu_st(s, args[0], -1, args[1], args[2], TCG_TYPE_I64); } else { - tcg_out_qemu_st(s, args[0], args[1], args[2], -1, + tcg_out_qemu_st(s, args[0], args[1], args[2], args[3], TCG_TYPE_I64); } break; -- 2.43.0 There is now always only one guest address register. Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- tcg/tcg.c | 18 +++++++++--------- tcg/aarch64/tcg-target.c.inc | 4 ++-- tcg/arm/tcg-target.c.inc | 4 ++-- tcg/i386/tcg-target.c.inc | 4 ++-- tcg/loongarch64/tcg-target.c.inc | 4 ++-- tcg/mips/tcg-target.c.inc | 4 ++-- tcg/ppc/tcg-target.c.inc | 4 ++-- tcg/riscv/tcg-target.c.inc | 4 ++-- tcg/s390x/tcg-target.c.inc | 4 ++-- tcg/sparc64/tcg-target.c.inc | 4 ++-- 10 files changed, 27 insertions(+), 27 deletions(-) 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 @@ struct TCGLabelQemuLdst { bool is_ld; /* qemu_ld: true, qemu_st: false */ MemOpIdx oi; TCGType type; /* result type of a load */ - TCGReg addrlo_reg; /* reg index for low word of guest virtual addr */ - TCGReg addrhi_reg; /* reg index for high word of guest virtual addr */ + TCGReg addr_reg; /* reg index for guest virtual addr */ TCGReg datalo_reg; /* reg index for low word to be loaded or stored */ TCGReg datahi_reg; /* reg index for high word to be loaded or stored */ const tcg_insn_unit *raddr; /* addr of the next IR of qemu_ld/st IR */ @@ -XXX,XX +XXX,XX @@ static void tcg_out_ld_helper_args(TCGContext *s, const TCGLabelQemuLdst *ldst, */ tcg_out_helper_add_mov(mov, loc + HOST_BIG_ENDIAN, TCG_TYPE_I32, TCG_TYPE_I32, - ldst->addrlo_reg, -1); + ldst->addr_reg, -1); tcg_out_helper_load_slots(s, 1, mov, parm); tcg_out_helper_load_imm(s, loc[!HOST_BIG_ENDIAN].arg_slot, @@ -XXX,XX +XXX,XX @@ static void tcg_out_ld_helper_args(TCGContext *s, const TCGLabelQemuLdst *ldst, next_arg += 2; } else { nmov = tcg_out_helper_add_mov(mov, loc, TCG_TYPE_I64, s->addr_type, - ldst->addrlo_reg, ldst->addrhi_reg); + ldst->addr_reg, -1); tcg_out_helper_load_slots(s, nmov, mov, parm); next_arg += nmov; } @@ -XXX,XX +XXX,XX @@ static void tcg_out_st_helper_args(TCGContext *s, const TCGLabelQemuLdst *ldst, /* Handle addr argument. */ loc = &info->in[next_arg]; - if (TCG_TARGET_REG_BITS == 32 && s->addr_type == TCG_TYPE_I32) { + tcg_debug_assert(s->addr_type <= TCG_TYPE_REG); + if (TCG_TARGET_REG_BITS == 32) { /* - * 32-bit host with 32-bit guest: zero-extend the guest address + * 32-bit host (and thus 32-bit guest): zero-extend the guest address * to 64-bits for the helper by storing the low part. Later, * after we have processed the register inputs, we will load a * zero for the high part. */ tcg_out_helper_add_mov(mov, loc + HOST_BIG_ENDIAN, TCG_TYPE_I32, TCG_TYPE_I32, - ldst->addrlo_reg, -1); + ldst->addr_reg, -1); next_arg += 2; nmov += 1; } else { n = tcg_out_helper_add_mov(mov, loc, TCG_TYPE_I64, s->addr_type, - ldst->addrlo_reg, ldst->addrhi_reg); + ldst->addr_reg, -1); next_arg += n; nmov += n; } @@ -XXX,XX +XXX,XX @@ static void tcg_out_st_helper_args(TCGContext *s, const TCGLabelQemuLdst *ldst, g_assert_not_reached(); } - if (TCG_TARGET_REG_BITS == 32 && s->addr_type == TCG_TYPE_I32) { + if (TCG_TARGET_REG_BITS == 32) { /* Zero extend the address by loading a zero for the high part. */ loc = &info->in[1 + !HOST_BIG_ENDIAN]; tcg_out_helper_load_imm(s, loc->arg_slot, TCG_TYPE_I32, 0, parm); diff --git a/tcg/aarch64/tcg-target.c.inc b/tcg/aarch64/tcg-target.c.inc index XXXXXXX..XXXXXXX 100644 --- a/tcg/aarch64/tcg-target.c.inc +++ b/tcg/aarch64/tcg-target.c.inc @@ -XXX,XX +XXX,XX @@ static TCGLabelQemuLdst *prepare_host_addr(TCGContext *s, HostAddress *h, ldst = new_ldst_label(s); ldst->is_ld = is_ld; ldst->oi = oi; - ldst->addrlo_reg = addr_reg; + ldst->addr_reg = addr_reg; mask_type = (s->page_bits + s->tlb_dyn_max_bits > 32 ? TCG_TYPE_I64 : TCG_TYPE_I32); @@ -XXX,XX +XXX,XX @@ static TCGLabelQemuLdst *prepare_host_addr(TCGContext *s, HostAddress *h, ldst->is_ld = is_ld; ldst->oi = oi; - ldst->addrlo_reg = addr_reg; + ldst->addr_reg = addr_reg; /* tst addr, #mask */ tcg_out_logicali(s, I3404_ANDSI, 0, TCG_REG_XZR, addr_reg, a_mask); diff --git a/tcg/arm/tcg-target.c.inc b/tcg/arm/tcg-target.c.inc index XXXXXXX..XXXXXXX 100644 --- a/tcg/arm/tcg-target.c.inc +++ b/tcg/arm/tcg-target.c.inc @@ -XXX,XX +XXX,XX @@ static TCGLabelQemuLdst *prepare_host_addr(TCGContext *s, HostAddress *h, ldst = new_ldst_label(s); ldst->is_ld = is_ld; ldst->oi = oi; - ldst->addrlo_reg = addr; + ldst->addr_reg = addr; /* Load cpu->neg.tlb.f[mmu_idx].{mask,table} into {r0,r1}. */ QEMU_BUILD_BUG_ON(offsetof(CPUTLBDescFast, mask) != 0); @@ -XXX,XX +XXX,XX @@ static TCGLabelQemuLdst *prepare_host_addr(TCGContext *s, HostAddress *h, ldst = new_ldst_label(s); ldst->is_ld = is_ld; ldst->oi = oi; - ldst->addrlo_reg = addr; + ldst->addr_reg = addr; /* We are expecting alignment to max out at 7 */ tcg_debug_assert(a_mask <= 0xff); diff --git a/tcg/i386/tcg-target.c.inc b/tcg/i386/tcg-target.c.inc index XXXXXXX..XXXXXXX 100644 --- a/tcg/i386/tcg-target.c.inc +++ b/tcg/i386/tcg-target.c.inc @@ -XXX,XX +XXX,XX @@ static TCGLabelQemuLdst *prepare_host_addr(TCGContext *s, HostAddress *h, ldst = new_ldst_label(s); ldst->is_ld = is_ld; ldst->oi = oi; - ldst->addrlo_reg = addr; + ldst->addr_reg = addr; if (TCG_TARGET_REG_BITS == 64) { ttype = s->addr_type; @@ -XXX,XX +XXX,XX @@ static TCGLabelQemuLdst *prepare_host_addr(TCGContext *s, HostAddress *h, ldst = new_ldst_label(s); ldst->is_ld = is_ld; ldst->oi = oi; - ldst->addrlo_reg = addr; + ldst->addr_reg = addr; /* jne slow_path */ jcc = tcg_out_cmp(s, TCG_COND_TSTNE, addr, a_mask, true, false); 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 TCGLabelQemuLdst *prepare_host_addr(TCGContext *s, HostAddress *h, ldst = new_ldst_label(s); ldst->is_ld = is_ld; ldst->oi = oi; - ldst->addrlo_reg = addr_reg; + ldst->addr_reg = addr_reg; tcg_out_ld(s, TCG_TYPE_PTR, TCG_REG_TMP0, TCG_AREG0, mask_ofs); tcg_out_ld(s, TCG_TYPE_PTR, TCG_REG_TMP1, TCG_AREG0, table_ofs); @@ -XXX,XX +XXX,XX @@ static TCGLabelQemuLdst *prepare_host_addr(TCGContext *s, HostAddress *h, ldst->is_ld = is_ld; ldst->oi = oi; - ldst->addrlo_reg = addr_reg; + ldst->addr_reg = addr_reg; /* * Without micro-architecture details, we don't know which of diff --git a/tcg/mips/tcg-target.c.inc b/tcg/mips/tcg-target.c.inc index XXXXXXX..XXXXXXX 100644 --- a/tcg/mips/tcg-target.c.inc +++ b/tcg/mips/tcg-target.c.inc @@ -XXX,XX +XXX,XX @@ static TCGLabelQemuLdst *prepare_host_addr(TCGContext *s, HostAddress *h, ldst = new_ldst_label(s); ldst->is_ld = is_ld; ldst->oi = oi; - ldst->addrlo_reg = addr; + ldst->addr_reg = addr; /* Load tlb_mask[mmu_idx] and tlb_table[mmu_idx]. */ tcg_out_ld(s, TCG_TYPE_PTR, TCG_TMP0, TCG_AREG0, mask_off); @@ -XXX,XX +XXX,XX @@ static TCGLabelQemuLdst *prepare_host_addr(TCGContext *s, HostAddress *h, ldst->is_ld = is_ld; ldst->oi = oi; - ldst->addrlo_reg = addr; + ldst->addr_reg = addr; /* We are expecting a_bits to max out at 7, much lower than ANDI. */ tcg_debug_assert(a_bits < 16); diff --git a/tcg/ppc/tcg-target.c.inc b/tcg/ppc/tcg-target.c.inc index XXXXXXX..XXXXXXX 100644 --- a/tcg/ppc/tcg-target.c.inc +++ b/tcg/ppc/tcg-target.c.inc @@ -XXX,XX +XXX,XX @@ static TCGLabelQemuLdst *prepare_host_addr(TCGContext *s, HostAddress *h, ldst = new_ldst_label(s); ldst->is_ld = is_ld; ldst->oi = oi; - ldst->addrlo_reg = addr; + ldst->addr_reg = addr; /* Load tlb_mask[mmu_idx] and tlb_table[mmu_idx]. */ tcg_out_ld(s, TCG_TYPE_PTR, TCG_REG_TMP1, TCG_AREG0, mask_off); @@ -XXX,XX +XXX,XX @@ static TCGLabelQemuLdst *prepare_host_addr(TCGContext *s, HostAddress *h, ldst = new_ldst_label(s); ldst->is_ld = is_ld; ldst->oi = oi; - ldst->addrlo_reg = addr; + ldst->addr_reg = addr; /* We are expecting a_bits to max out at 7, much lower than ANDI. */ tcg_debug_assert(a_bits < 16); diff --git a/tcg/riscv/tcg-target.c.inc b/tcg/riscv/tcg-target.c.inc index XXXXXXX..XXXXXXX 100644 --- a/tcg/riscv/tcg-target.c.inc +++ b/tcg/riscv/tcg-target.c.inc @@ -XXX,XX +XXX,XX @@ static TCGLabelQemuLdst *prepare_host_addr(TCGContext *s, TCGReg *pbase, ldst = new_ldst_label(s); ldst->is_ld = is_ld; ldst->oi = oi; - ldst->addrlo_reg = addr_reg; + ldst->addr_reg = addr_reg; init_setting_vtype(s); @@ -XXX,XX +XXX,XX @@ static TCGLabelQemuLdst *prepare_host_addr(TCGContext *s, TCGReg *pbase, ldst = new_ldst_label(s); ldst->is_ld = is_ld; ldst->oi = oi; - ldst->addrlo_reg = addr_reg; + ldst->addr_reg = addr_reg; init_setting_vtype(s); diff --git a/tcg/s390x/tcg-target.c.inc b/tcg/s390x/tcg-target.c.inc index XXXXXXX..XXXXXXX 100644 --- a/tcg/s390x/tcg-target.c.inc +++ b/tcg/s390x/tcg-target.c.inc @@ -XXX,XX +XXX,XX @@ static TCGLabelQemuLdst *prepare_host_addr(TCGContext *s, HostAddress *h, ldst = new_ldst_label(s); ldst->is_ld = is_ld; ldst->oi = oi; - ldst->addrlo_reg = addr_reg; + ldst->addr_reg = addr_reg; tcg_out_sh64(s, RSY_SRLG, TCG_TMP0, addr_reg, TCG_REG_NONE, s->page_bits - CPU_TLB_ENTRY_BITS); @@ -XXX,XX +XXX,XX @@ static TCGLabelQemuLdst *prepare_host_addr(TCGContext *s, HostAddress *h, ldst = new_ldst_label(s); ldst->is_ld = is_ld; ldst->oi = oi; - ldst->addrlo_reg = addr_reg; + ldst->addr_reg = addr_reg; tcg_debug_assert(a_mask <= 0xffff); tcg_out_insn(s, RI, TMLL, addr_reg, a_mask); diff --git a/tcg/sparc64/tcg-target.c.inc b/tcg/sparc64/tcg-target.c.inc index XXXXXXX..XXXXXXX 100644 --- a/tcg/sparc64/tcg-target.c.inc +++ b/tcg/sparc64/tcg-target.c.inc @@ -XXX,XX +XXX,XX @@ static TCGLabelQemuLdst *prepare_host_addr(TCGContext *s, HostAddress *h, ldst = new_ldst_label(s); ldst->is_ld = is_ld; ldst->oi = oi; - ldst->addrlo_reg = addr_reg; + ldst->addr_reg = addr_reg; ldst->label_ptr[0] = s->code_ptr; /* bne,pn %[xi]cc, label0 */ @@ -XXX,XX +XXX,XX @@ static TCGLabelQemuLdst *prepare_host_addr(TCGContext *s, HostAddress *h, ldst = new_ldst_label(s); ldst->is_ld = is_ld; ldst->oi = oi; - ldst->addrlo_reg = addr_reg; + ldst->addr_reg = addr_reg; ldst->label_ptr[0] = s->code_ptr; /* bne,pn %icc, label0 */ -- 2.43.0 The declaration uses uint64_t for addr. Fixes: 595cd9ce2ec ("plugins: add plugin API to read guest memory") Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- plugins/api.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) 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 @@ GArray *qemu_plugin_get_registers(void) return create_register_handles(regs); } -bool qemu_plugin_read_memory_vaddr(vaddr addr, GByteArray *data, size_t len) +bool qemu_plugin_read_memory_vaddr(uint64_t addr, GByteArray *data, size_t len) { g_assert(current_cpu); -- 2.43.0 The declarations use vaddr for size. Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- accel/tcg/cputlb.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/accel/tcg/cputlb.c b/accel/tcg/cputlb.c index XXXXXXX..XXXXXXX 100644 --- a/accel/tcg/cputlb.c +++ b/accel/tcg/cputlb.c @@ -XXX,XX +XXX,XX @@ void tlb_set_page_full(CPUState *cpu, int mmu_idx, void tlb_set_page_with_attrs(CPUState *cpu, vaddr addr, hwaddr paddr, MemTxAttrs attrs, int prot, - int mmu_idx, uint64_t size) + int mmu_idx, vaddr size) { CPUTLBEntryFull full = { .phys_addr = paddr, @@ -XXX,XX +XXX,XX @@ void tlb_set_page_with_attrs(CPUState *cpu, vaddr addr, void tlb_set_page(CPUState *cpu, vaddr addr, hwaddr paddr, int prot, - int mmu_idx, uint64_t size) + int mmu_idx, vaddr size) { tlb_set_page_with_attrs(cpu, addr, paddr, MEMTXATTRS_UNSPECIFIED, prot, mmu_idx, size); -- 2.43.0 Since we no longer support 64-bit guests on 32-bit hosts, we can use a 32-bit type on a 32-bit host. Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- include/exec/vaddr.h | 16 +++++++++------- 1 file changed, 9 insertions(+), 7 deletions(-) diff --git a/include/exec/vaddr.h b/include/exec/vaddr.h index XXXXXXX..XXXXXXX 100644 --- a/include/exec/vaddr.h +++ b/include/exec/vaddr.h @@ -XXX,XX +XXX,XX @@ /** * vaddr: * Type wide enough to contain any #target_ulong virtual address. + * We do not support 64-bit guest on 32-host and detect at configure time. + * Therefore, a host pointer width will always fit a guest pointer. */ -typedef uint64_t vaddr; -#define VADDR_PRId PRId64 -#define VADDR_PRIu PRIu64 -#define VADDR_PRIo PRIo64 -#define VADDR_PRIx PRIx64 -#define VADDR_PRIX PRIX64 -#define VADDR_MAX UINT64_MAX +typedef uintptr_t vaddr; +#define VADDR_PRId PRIdPTR +#define VADDR_PRIu PRIuPTR +#define VADDR_PRIo PRIoPTR +#define VADDR_PRIx PRIxPTR +#define VADDR_PRIX PRIXPTR +#define VADDR_MAX UINTPTR_MAX #endif -- 2.43.0 Since we no longer support 64-bit guests on 32-bit hosts, we can use a 32-bit type on a 32-bit host. This shrinks the size of the structure to 16 bytes on a 32-bit host. Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- include/exec/tlb-common.h | 10 +++++----- accel/tcg/cputlb.c | 21 ++++----------------- tcg/arm/tcg-target.c.inc | 1 - tcg/mips/tcg-target.c.inc | 12 +++++------- tcg/ppc/tcg-target.c.inc | 21 +++++---------------- 5 files changed, 19 insertions(+), 46 deletions(-) diff --git a/include/exec/tlb-common.h b/include/exec/tlb-common.h index XXXXXXX..XXXXXXX 100644 --- a/include/exec/tlb-common.h +++ b/include/exec/tlb-common.h @@ -XXX,XX +XXX,XX @@ #ifndef EXEC_TLB_COMMON_H #define EXEC_TLB_COMMON_H 1 -#define CPU_TLB_ENTRY_BITS 5 +#define CPU_TLB_ENTRY_BITS (HOST_LONG_BITS == 32 ? 4 : 5) /* Minimalized TLB entry for use by TCG fast path. */ typedef union CPUTLBEntry { struct { - uint64_t addr_read; - uint64_t addr_write; - uint64_t addr_code; + uintptr_t addr_read; + uintptr_t addr_write; + uintptr_t addr_code; /* * Addend to virtual address to get host address. IO accesses * use the corresponding iotlb value. @@ -XXX,XX +XXX,XX @@ typedef union CPUTLBEntry { * Padding to get a power of two size, as well as index * access to addr_{read,write,code}. */ - uint64_t addr_idx[(1 << CPU_TLB_ENTRY_BITS) / sizeof(uint64_t)]; + uintptr_t addr_idx[(1 << CPU_TLB_ENTRY_BITS) / sizeof(uintptr_t)]; } CPUTLBEntry; QEMU_BUILD_BUG_ON(sizeof(CPUTLBEntry) != (1 << CPU_TLB_ENTRY_BITS)); diff --git a/accel/tcg/cputlb.c b/accel/tcg/cputlb.c index XXXXXXX..XXXXXXX 100644 --- a/accel/tcg/cputlb.c +++ b/accel/tcg/cputlb.c @@ -XXX,XX +XXX,XX @@ static inline uint64_t tlb_read_idx(const CPUTLBEntry *entry, { /* Do not rearrange the CPUTLBEntry structure members. */ QEMU_BUILD_BUG_ON(offsetof(CPUTLBEntry, addr_read) != - MMU_DATA_LOAD * sizeof(uint64_t)); + MMU_DATA_LOAD * sizeof(uintptr_t)); QEMU_BUILD_BUG_ON(offsetof(CPUTLBEntry, addr_write) != - MMU_DATA_STORE * sizeof(uint64_t)); + MMU_DATA_STORE * sizeof(uintptr_t)); QEMU_BUILD_BUG_ON(offsetof(CPUTLBEntry, addr_code) != - MMU_INST_FETCH * sizeof(uint64_t)); + MMU_INST_FETCH * sizeof(uintptr_t)); -#if TARGET_LONG_BITS == 32 - /* Use qatomic_read, in case of addr_write; only care about low bits. */ - const uint32_t *ptr = (uint32_t *)&entry->addr_idx[access_type]; - ptr += HOST_BIG_ENDIAN; - return qatomic_read(ptr); -#else - const uint64_t *ptr = &entry->addr_idx[access_type]; + const uintptr_t *ptr = &entry->addr_idx[access_type]; /* ofs might correspond to .addr_write, so use qatomic_read */ return qatomic_read(ptr); -#endif } static inline uint64_t tlb_addr_write(const CPUTLBEntry *entry) @@ -XXX,XX +XXX,XX @@ static void tlb_reset_dirty_range_locked(CPUTLBEntry *tlb_entry, addr &= TARGET_PAGE_MASK; addr += tlb_entry->addend; if ((addr - start) < length) { -#if TARGET_LONG_BITS == 32 - uint32_t *ptr_write = (uint32_t *)&tlb_entry->addr_write; - ptr_write += HOST_BIG_ENDIAN; - qatomic_set(ptr_write, *ptr_write | TLB_NOTDIRTY); -#else qatomic_set(&tlb_entry->addr_write, tlb_entry->addr_write | TLB_NOTDIRTY); -#endif } } } diff --git a/tcg/arm/tcg-target.c.inc b/tcg/arm/tcg-target.c.inc index XXXXXXX..XXXXXXX 100644 --- a/tcg/arm/tcg-target.c.inc +++ b/tcg/arm/tcg-target.c.inc @@ -XXX,XX +XXX,XX @@ static TCGLabelQemuLdst *prepare_host_addr(TCGContext *s, HostAddress *h, * Add the tlb_table pointer, creating the CPUTLBEntry address in R1. * Load the tlb comparator into R2 and the fast path addend into R1. */ - QEMU_BUILD_BUG_ON(HOST_BIG_ENDIAN); if (cmp_off == 0) { tcg_out_ld32_rwb(s, COND_AL, TCG_REG_R2, TCG_REG_R1, TCG_REG_R0); } else { diff --git a/tcg/mips/tcg-target.c.inc b/tcg/mips/tcg-target.c.inc index XXXXXXX..XXXXXXX 100644 --- a/tcg/mips/tcg-target.c.inc +++ b/tcg/mips/tcg-target.c.inc @@ -XXX,XX +XXX,XX @@ static TCGLabelQemuLdst *prepare_host_addr(TCGContext *s, HostAddress *h, /* Add the tlb_table pointer, creating the CPUTLBEntry address. */ tcg_out_opc_reg(s, ALIAS_PADD, TCG_TMP3, TCG_TMP3, TCG_TMP1); - if (TCG_TARGET_REG_BITS == 32 || addr_type == TCG_TYPE_I32) { - /* Load the (low half) tlb comparator. */ + /* Load the tlb comparator. */ + if (TCG_TARGET_REG_BITS == 64 && addr_type == TCG_TYPE_I32) { tcg_out_ld(s, TCG_TYPE_I32, TCG_TMP0, TCG_TMP3, cmp_off + HOST_BIG_ENDIAN * 4); } else { - tcg_out_ld(s, TCG_TYPE_I64, TCG_TMP0, TCG_TMP3, cmp_off); + tcg_out_ld(s, TCG_TYPE_REG, TCG_TMP0, TCG_TMP3, cmp_off); } - if (TCG_TARGET_REG_BITS == 64 || addr_type == TCG_TYPE_I32) { - /* Load the tlb addend for the fast path. */ - tcg_out_ld(s, TCG_TYPE_PTR, TCG_TMP3, TCG_TMP3, add_off); - } + /* Load the tlb addend for the fast path. */ + tcg_out_ld(s, TCG_TYPE_PTR, TCG_TMP3, TCG_TMP3, add_off); /* * Mask the page bits, keeping the alignment bits to compare against. diff --git a/tcg/ppc/tcg-target.c.inc b/tcg/ppc/tcg-target.c.inc index XXXXXXX..XXXXXXX 100644 --- a/tcg/ppc/tcg-target.c.inc +++ b/tcg/ppc/tcg-target.c.inc @@ -XXX,XX +XXX,XX @@ static TCGLabelQemuLdst *prepare_host_addr(TCGContext *s, HostAddress *h, tcg_out32(s, AND | SAB(TCG_REG_TMP1, TCG_REG_TMP1, TCG_REG_R0)); /* - * Load the (low part) TLB comparator into TMP2. + * Load the TLB comparator into TMP2. * For 64-bit host, always load the entire 64-bit slot for simplicity. * We will ignore the high bits with tcg_out_cmp(..., addr_type). */ - if (TCG_TARGET_REG_BITS == 64) { - if (cmp_off == 0) { - tcg_out32(s, LDUX | TAB(TCG_REG_TMP2, - TCG_REG_TMP1, TCG_REG_TMP2)); - } else { - tcg_out32(s, ADD | TAB(TCG_REG_TMP1, - TCG_REG_TMP1, TCG_REG_TMP2)); - tcg_out_ld(s, TCG_TYPE_I64, TCG_REG_TMP2, - TCG_REG_TMP1, cmp_off); - } - } else if (cmp_off == 0 && !HOST_BIG_ENDIAN) { - tcg_out32(s, LWZUX | TAB(TCG_REG_TMP2, - TCG_REG_TMP1, TCG_REG_TMP2)); + if (cmp_off == 0) { + tcg_out32(s, (TCG_TARGET_REG_BITS == 64 ? LDUX : LWZUX) + | TAB(TCG_REG_TMP2, TCG_REG_TMP1, TCG_REG_TMP2)); } else { tcg_out32(s, ADD | TAB(TCG_REG_TMP1, TCG_REG_TMP1, TCG_REG_TMP2)); - tcg_out_ld(s, TCG_TYPE_I32, TCG_REG_TMP2, TCG_REG_TMP1, - cmp_off + 4 * HOST_BIG_ENDIAN); + tcg_out_ld(s, TCG_TYPE_PTR, TCG_REG_TMP2, TCG_REG_TMP1, cmp_off); } /* -- 2.43.0 For loongarch, mips, riscv and sparc, a zero register is available all the time. For aarch64, register index 31 depends on context: sometimes it is the stack pointer, and sometimes it is the zero register. Introduce a new general-purpose constraint which maps 0 to TCG_REG_ZERO, if defined. This differs from existing constant constraints in that const_arg[*] is recorded as false, indicating that the value is in a register. Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- include/tcg/tcg.h | 3 ++- tcg/aarch64/tcg-target.h | 2 ++ tcg/loongarch64/tcg-target.h | 2 ++ tcg/mips/tcg-target.h | 2 ++ tcg/riscv/tcg-target.h | 2 ++ tcg/sparc64/tcg-target.h | 3 ++- tcg/tcg.c | 29 ++++++++++++++++++++++------- docs/devel/tcg-ops.rst | 4 +++- 8 files changed, 37 insertions(+), 10 deletions(-) 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 @@ void tb_target_set_jmp_target(const TranslationBlock *, int, void tcg_set_frame(TCGContext *s, TCGReg reg, intptr_t start, intptr_t size); -#define TCG_CT_CONST 1 /* any constant of register size */ +#define TCG_CT_CONST 1 /* any constant of register size */ +#define TCG_CT_REG_ZERO 2 /* zero, in TCG_REG_ZERO */ typedef struct TCGArgConstraint { unsigned ct : 16; diff --git a/tcg/aarch64/tcg-target.h b/tcg/aarch64/tcg-target.h index XXXXXXX..XXXXXXX 100644 --- a/tcg/aarch64/tcg-target.h +++ b/tcg/aarch64/tcg-target.h @@ -XXX,XX +XXX,XX @@ typedef enum { TCG_AREG0 = TCG_REG_X19, } TCGReg; +#define TCG_REG_ZERO TCG_REG_XZR + #define TCG_TARGET_NB_REGS 64 #endif /* AARCH64_TCG_TARGET_H */ diff --git a/tcg/loongarch64/tcg-target.h b/tcg/loongarch64/tcg-target.h index XXXXXXX..XXXXXXX 100644 --- a/tcg/loongarch64/tcg-target.h +++ b/tcg/loongarch64/tcg-target.h @@ -XXX,XX +XXX,XX @@ typedef enum { TCG_VEC_TMP0 = TCG_REG_V23, } TCGReg; +#define TCG_REG_ZERO TCG_REG_ZERO + #endif /* LOONGARCH_TCG_TARGET_H */ diff --git a/tcg/mips/tcg-target.h b/tcg/mips/tcg-target.h index XXXXXXX..XXXXXXX 100644 --- a/tcg/mips/tcg-target.h +++ b/tcg/mips/tcg-target.h @@ -XXX,XX +XXX,XX @@ typedef enum { TCG_AREG0 = TCG_REG_S8, } TCGReg; +#define TCG_REG_ZERO TCG_REG_ZERO + #endif diff --git a/tcg/riscv/tcg-target.h b/tcg/riscv/tcg-target.h index XXXXXXX..XXXXXXX 100644 --- a/tcg/riscv/tcg-target.h +++ b/tcg/riscv/tcg-target.h @@ -XXX,XX +XXX,XX @@ typedef enum { TCG_REG_TMP2 = TCG_REG_T4, } TCGReg; +#define TCG_REG_ZERO TCG_REG_ZERO + #endif diff --git a/tcg/sparc64/tcg-target.h b/tcg/sparc64/tcg-target.h index XXXXXXX..XXXXXXX 100644 --- a/tcg/sparc64/tcg-target.h +++ b/tcg/sparc64/tcg-target.h @@ -XXX,XX +XXX,XX @@ typedef enum { TCG_REG_I7, } TCGReg; -#define TCG_AREG0 TCG_REG_I0 +#define TCG_AREG0 TCG_REG_I0 +#define TCG_REG_ZERO TCG_REG_G0 #endif 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 process_constraint_sets(void) case 'i': args_ct[i].ct |= TCG_CT_CONST; break; +#ifdef TCG_REG_ZERO + case 'z': + args_ct[i].ct |= TCG_CT_REG_ZERO; + break; +#endif /* Include all of the target-specific constraints. */ @@ -XXX,XX +XXX,XX @@ static void tcg_reg_alloc_op(TCGContext *s, const TCGOp *op) arg_ct = &args_ct[i]; ts = arg_temp(arg); - if (ts->val_type == TEMP_VAL_CONST - && tcg_target_const_match(ts->val, arg_ct->ct, ts->type, - op_cond, TCGOP_VECE(op))) { - /* constant is OK for instruction */ - const_args[i] = 1; - new_args[i] = ts->val; - continue; + if (ts->val_type == TEMP_VAL_CONST) { +#ifdef TCG_REG_ZERO + if (ts->val == 0 && (arg_ct->ct & TCG_CT_REG_ZERO)) { + /* Hardware zero register: indicate register via non-const. */ + const_args[i] = 0; + new_args[i] = TCG_REG_ZERO; + continue; + } +#endif + + if (tcg_target_const_match(ts->val, arg_ct->ct, ts->type, + op_cond, TCGOP_VECE(op))) { + /* constant is OK for instruction */ + const_args[i] = 1; + new_args[i] = ts->val; + continue; + } } reg = ts->reg; diff --git a/docs/devel/tcg-ops.rst b/docs/devel/tcg-ops.rst index XXXXXXX..XXXXXXX 100644 --- a/docs/devel/tcg-ops.rst +++ b/docs/devel/tcg-ops.rst @@ -XXX,XX +XXX,XX @@ operation uses a constant input constraint which does not allow all constants, it must also accept registers in order to have a fallback. The constraint '``i``' is defined generically to accept any constant. The constraint '``r``' is not defined generically, but is consistently -used by each backend to indicate all registers. +used by each backend to indicate all registers. If ``TCG_REG_ZERO`` +is defined by the backend, the constraint '``z``' is defined generically +to map constant 0 to the hardware zero register. The movi_i32 and movi_i64 operations must accept any constants. -- 2.43.0 Note that 'Z' is still used for addsub2. Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- tcg/aarch64/tcg-target-con-set.h | 12 ++++----- tcg/aarch64/tcg-target.c.inc | 46 ++++++++++++++------------------ 2 files changed, 26 insertions(+), 32 deletions(-) diff --git a/tcg/aarch64/tcg-target-con-set.h b/tcg/aarch64/tcg-target-con-set.h index XXXXXXX..XXXXXXX 100644 --- a/tcg/aarch64/tcg-target-con-set.h +++ b/tcg/aarch64/tcg-target-con-set.h @@ -XXX,XX +XXX,XX @@ */ C_O0_I1(r) C_O0_I2(r, rC) -C_O0_I2(rZ, r) +C_O0_I2(rz, r) C_O0_I2(w, r) -C_O0_I3(rZ, rZ, r) +C_O0_I3(rz, rz, r) C_O1_I1(r, r) C_O1_I1(w, r) C_O1_I1(w, w) C_O1_I1(w, wr) -C_O1_I2(r, 0, rZ) +C_O1_I2(r, 0, rz) C_O1_I2(r, r, r) C_O1_I2(r, r, rA) C_O1_I2(r, r, rAL) C_O1_I2(r, r, rC) C_O1_I2(r, r, ri) C_O1_I2(r, r, rL) -C_O1_I2(r, rZ, rZ) +C_O1_I2(r, rz, rz) C_O1_I2(w, 0, w) C_O1_I2(w, w, w) C_O1_I2(w, w, wN) C_O1_I2(w, w, wO) C_O1_I2(w, w, wZ) C_O1_I3(w, w, w, w) -C_O1_I4(r, r, rC, rZ, rZ) +C_O1_I4(r, r, rC, rz, rz) C_O2_I1(r, r, r) -C_O2_I4(r, r, rZ, rZ, rA, rMZ) +C_O2_I4(r, r, rz, rz, rA, rMZ) diff --git a/tcg/aarch64/tcg-target.c.inc b/tcg/aarch64/tcg-target.c.inc index XXXXXXX..XXXXXXX 100644 --- a/tcg/aarch64/tcg-target.c.inc +++ b/tcg/aarch64/tcg-target.c.inc @@ -XXX,XX +XXX,XX @@ static void tcg_out_op(TCGContext *s, TCGOpcode opc, TCGType ext, TCGArg a2 = args[2]; int c2 = const_args[2]; - /* Some operands are defined with "rZ" constraint, a register or - the zero register. These need not actually test args[I] == 0. */ -#define REG0(I) (const_args[I] ? TCG_REG_XZR : (TCGReg)args[I]) - switch (opc) { case INDEX_op_goto_ptr: tcg_out_insn(s, 3207, BR, a0); @@ -XXX,XX +XXX,XX @@ static void tcg_out_op(TCGContext *s, TCGOpcode opc, TCGType ext, case INDEX_op_st8_i32: case INDEX_op_st8_i64: - tcg_out_ldst(s, I3312_STRB, REG0(0), a1, a2, 0); + tcg_out_ldst(s, I3312_STRB, a0, a1, a2, 0); break; case INDEX_op_st16_i32: case INDEX_op_st16_i64: - tcg_out_ldst(s, I3312_STRH, REG0(0), a1, a2, 1); + tcg_out_ldst(s, I3312_STRH, a0, a1, a2, 1); break; case INDEX_op_st_i32: case INDEX_op_st32_i64: - tcg_out_ldst(s, I3312_STRW, REG0(0), a1, a2, 2); + tcg_out_ldst(s, I3312_STRW, a0, a1, a2, 2); break; case INDEX_op_st_i64: - tcg_out_ldst(s, I3312_STRX, REG0(0), a1, a2, 3); + tcg_out_ldst(s, I3312_STRX, a0, a1, a2, 3); break; case INDEX_op_add_i32: @@ -XXX,XX +XXX,XX @@ static void tcg_out_op(TCGContext *s, TCGOpcode opc, TCGType ext, /* FALLTHRU */ case INDEX_op_movcond_i64: tcg_out_cmp(s, ext, args[5], a1, a2, c2); - tcg_out_insn(s, 3506, CSEL, ext, a0, REG0(3), REG0(4), args[5]); + tcg_out_insn(s, 3506, CSEL, ext, a0, args[3], args[4], args[5]); break; case INDEX_op_qemu_ld_i32: @@ -XXX,XX +XXX,XX @@ static void tcg_out_op(TCGContext *s, TCGOpcode opc, TCGType ext, break; case INDEX_op_qemu_st_i32: case INDEX_op_qemu_st_i64: - tcg_out_qemu_st(s, REG0(0), a1, a2, ext); + tcg_out_qemu_st(s, a0, a1, a2, ext); break; case INDEX_op_qemu_ld_i128: tcg_out_qemu_ldst_i128(s, a0, a1, a2, args[3], true); break; case INDEX_op_qemu_st_i128: - tcg_out_qemu_ldst_i128(s, REG0(0), REG0(1), a2, args[3], false); + tcg_out_qemu_ldst_i128(s, a0, a1, a2, args[3], false); break; case INDEX_op_bswap64_i64: @@ -XXX,XX +XXX,XX @@ static void tcg_out_op(TCGContext *s, TCGOpcode opc, TCGType ext, case INDEX_op_deposit_i64: case INDEX_op_deposit_i32: - tcg_out_dep(s, ext, a0, REG0(2), args[3], args[4]); + tcg_out_dep(s, ext, a0, a2, args[3], args[4]); break; case INDEX_op_extract_i64: @@ -XXX,XX +XXX,XX @@ static void tcg_out_op(TCGContext *s, TCGOpcode opc, TCGType ext, case INDEX_op_extract2_i64: case INDEX_op_extract2_i32: - tcg_out_extr(s, ext, a0, REG0(2), REG0(1), args[3]); + tcg_out_extr(s, ext, a0, a2, a1, args[3]); break; case INDEX_op_add2_i32: - tcg_out_addsub2(s, TCG_TYPE_I32, a0, a1, REG0(2), REG0(3), + tcg_out_addsub2(s, TCG_TYPE_I32, a0, a1, a2, args[3], (int32_t)args[4], args[5], const_args[4], const_args[5], false); break; case INDEX_op_add2_i64: - tcg_out_addsub2(s, TCG_TYPE_I64, a0, a1, REG0(2), REG0(3), args[4], + tcg_out_addsub2(s, TCG_TYPE_I64, a0, a1, a2, args[3], args[4], args[5], const_args[4], const_args[5], false); break; case INDEX_op_sub2_i32: - tcg_out_addsub2(s, TCG_TYPE_I32, a0, a1, REG0(2), REG0(3), + tcg_out_addsub2(s, TCG_TYPE_I32, a0, a1, a2, args[3], (int32_t)args[4], args[5], const_args[4], const_args[5], true); break; case INDEX_op_sub2_i64: - tcg_out_addsub2(s, TCG_TYPE_I64, a0, a1, REG0(2), REG0(3), args[4], + tcg_out_addsub2(s, TCG_TYPE_I64, a0, a1, a2, args[3], args[4], args[5], const_args[4], const_args[5], true); break; @@ -XXX,XX +XXX,XX @@ static void tcg_out_op(TCGContext *s, TCGOpcode opc, TCGType ext, default: g_assert_not_reached(); } - -#undef REG0 } static void tcg_out_vec_op(TCGContext *s, TCGOpcode opc, @@ -XXX,XX +XXX,XX @@ tcg_target_op_def(TCGOpcode op, TCGType type, unsigned flags) case INDEX_op_st16_i64: case INDEX_op_st32_i64: case INDEX_op_st_i64: - return C_O0_I2(rZ, r); + return C_O0_I2(rz, r); case INDEX_op_add_i32: case INDEX_op_add_i64: @@ -XXX,XX +XXX,XX @@ tcg_target_op_def(TCGOpcode op, TCGType type, unsigned flags) case INDEX_op_movcond_i32: case INDEX_op_movcond_i64: - return C_O1_I4(r, r, rC, rZ, rZ); + return C_O1_I4(r, r, rC, rz, rz); case INDEX_op_qemu_ld_i32: case INDEX_op_qemu_ld_i64: @@ -XXX,XX +XXX,XX @@ tcg_target_op_def(TCGOpcode op, TCGType type, unsigned flags) return C_O2_I1(r, r, r); case INDEX_op_qemu_st_i32: case INDEX_op_qemu_st_i64: - return C_O0_I2(rZ, r); + return C_O0_I2(rz, r); case INDEX_op_qemu_st_i128: - return C_O0_I3(rZ, rZ, r); + return C_O0_I3(rz, rz, r); case INDEX_op_deposit_i32: case INDEX_op_deposit_i64: - return C_O1_I2(r, 0, rZ); + return C_O1_I2(r, 0, rz); case INDEX_op_extract2_i32: case INDEX_op_extract2_i64: - return C_O1_I2(r, rZ, rZ); + return C_O1_I2(r, rz, rz); case INDEX_op_add2_i32: case INDEX_op_add2_i64: case INDEX_op_sub2_i32: case INDEX_op_sub2_i64: - return C_O2_I4(r, r, rZ, rZ, rA, rMZ); + return C_O2_I4(r, r, rz, rz, rA, rMZ); case INDEX_op_add_vec: case INDEX_op_sub_vec: -- 2.43.0 Replace target-specific 'Z' with generic 'z'. Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- tcg/loongarch64/tcg-target-con-set.h | 15 ++++++------- tcg/loongarch64/tcg-target-con-str.h | 1 - tcg/loongarch64/tcg-target.c.inc | 32 ++++++++++++---------------- 3 files changed, 21 insertions(+), 27 deletions(-) diff --git a/tcg/loongarch64/tcg-target-con-set.h b/tcg/loongarch64/tcg-target-con-set.h index XXXXXXX..XXXXXXX 100644 --- a/tcg/loongarch64/tcg-target-con-set.h +++ b/tcg/loongarch64/tcg-target-con-set.h @@ -XXX,XX +XXX,XX @@ * tcg-target-con-str.h; the constraint combination is inclusive or. */ C_O0_I1(r) -C_O0_I2(rZ, r) -C_O0_I2(rZ, rZ) +C_O0_I2(rz, r) +C_O0_I2(rz, rz) C_O0_I2(w, r) C_O0_I3(r, r, r) C_O1_I1(r, r) @@ -XXX,XX +XXX,XX @@ C_O1_I2(r, r, rI) C_O1_I2(r, r, rJ) C_O1_I2(r, r, rU) C_O1_I2(r, r, rW) -C_O1_I2(r, r, rZ) -C_O1_I2(r, 0, rZ) -C_O1_I2(r, rZ, ri) -C_O1_I2(r, rZ, rJ) -C_O1_I2(r, rZ, rZ) +C_O1_I2(r, 0, rz) +C_O1_I2(r, rz, ri) +C_O1_I2(r, rz, rJ) +C_O1_I2(r, rz, rz) C_O1_I2(w, w, w) C_O1_I2(w, w, wM) C_O1_I2(w, w, wA) C_O1_I3(w, w, w, w) -C_O1_I4(r, rZ, rJ, rZ, rZ) +C_O1_I4(r, rz, rJ, rz, rz) C_N2_I1(r, r, r) diff --git a/tcg/loongarch64/tcg-target-con-str.h b/tcg/loongarch64/tcg-target-con-str.h index XXXXXXX..XXXXXXX 100644 --- a/tcg/loongarch64/tcg-target-con-str.h +++ b/tcg/loongarch64/tcg-target-con-str.h @@ -XXX,XX +XXX,XX @@ REGS('w', ALL_VECTOR_REGS) CONST('I', TCG_CT_CONST_S12) CONST('J', TCG_CT_CONST_S32) CONST('U', TCG_CT_CONST_U12) -CONST('Z', TCG_CT_CONST_ZERO) CONST('C', TCG_CT_CONST_C12) CONST('W', TCG_CT_CONST_WSZ) CONST('M', TCG_CT_CONST_VCMP) 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 TCGReg tcg_target_call_oarg_reg(TCGCallReturnKind kind, int slot) #define TCG_GUEST_BASE_REG TCG_REG_S1 -#define TCG_CT_CONST_ZERO 0x100 -#define TCG_CT_CONST_S12 0x200 -#define TCG_CT_CONST_S32 0x400 -#define TCG_CT_CONST_U12 0x800 -#define TCG_CT_CONST_C12 0x1000 -#define TCG_CT_CONST_WSZ 0x2000 -#define TCG_CT_CONST_VCMP 0x4000 -#define TCG_CT_CONST_VADD 0x8000 +#define TCG_CT_CONST_S12 0x100 +#define TCG_CT_CONST_S32 0x200 +#define TCG_CT_CONST_U12 0x400 +#define TCG_CT_CONST_C12 0x800 +#define TCG_CT_CONST_WSZ 0x1000 +#define TCG_CT_CONST_VCMP 0x2000 +#define TCG_CT_CONST_VADD 0x4000 #define ALL_GENERAL_REGS MAKE_64BIT_MASK(0, 32) #define ALL_VECTOR_REGS MAKE_64BIT_MASK(32, 32) @@ -XXX,XX +XXX,XX @@ static bool tcg_target_const_match(int64_t val, int ct, if (ct & TCG_CT_CONST) { return true; } - if ((ct & TCG_CT_CONST_ZERO) && val == 0) { - return true; - } if ((ct & TCG_CT_CONST_S12) && val == sextreg(val, 0, 12)) { return true; } @@ -XXX,XX +XXX,XX @@ tcg_target_op_def(TCGOpcode op, TCGType type, unsigned flags) case INDEX_op_st_i64: case INDEX_op_qemu_st_i32: case INDEX_op_qemu_st_i64: - return C_O0_I2(rZ, r); + return C_O0_I2(rz, r); case INDEX_op_qemu_ld_i128: return C_N2_I1(r, r, r); @@ -XXX,XX +XXX,XX @@ tcg_target_op_def(TCGOpcode op, TCGType type, unsigned flags) case INDEX_op_brcond_i32: case INDEX_op_brcond_i64: - return C_O0_I2(rZ, rZ); + return C_O0_I2(rz, rz); case INDEX_op_ext8s_i32: case INDEX_op_ext8s_i64: @@ -XXX,XX +XXX,XX @@ tcg_target_op_def(TCGOpcode op, TCGType type, unsigned flags) case INDEX_op_deposit_i32: case INDEX_op_deposit_i64: /* Must deposit into the same register as input */ - return C_O1_I2(r, 0, rZ); + return C_O1_I2(r, 0, rz); case INDEX_op_sub_i32: case INDEX_op_setcond_i32: - return C_O1_I2(r, rZ, ri); + return C_O1_I2(r, rz, ri); case INDEX_op_sub_i64: case INDEX_op_setcond_i64: - return C_O1_I2(r, rZ, rJ); + return C_O1_I2(r, rz, rJ); case INDEX_op_mul_i32: case INDEX_op_mul_i64: @@ -XXX,XX +XXX,XX @@ tcg_target_op_def(TCGOpcode op, TCGType type, unsigned flags) case INDEX_op_rem_i64: case INDEX_op_remu_i32: case INDEX_op_remu_i64: - return C_O1_I2(r, rZ, rZ); + return C_O1_I2(r, rz, rz); case INDEX_op_movcond_i32: case INDEX_op_movcond_i64: - return C_O1_I4(r, rZ, rJ, rZ, rZ); + return C_O1_I4(r, rz, rJ, rz, rz); case INDEX_op_ld_vec: case INDEX_op_dupm_vec: -- 2.43.0 Replace target-specific 'Z' with generic 'z'. Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- tcg/mips/tcg-target-con-set.h | 26 ++++++++++----------- tcg/mips/tcg-target-con-str.h | 1 - tcg/mips/tcg-target.c.inc | 44 ++++++++++++++--------------------- 3 files changed, 31 insertions(+), 40 deletions(-) diff --git a/tcg/mips/tcg-target-con-set.h b/tcg/mips/tcg-target-con-set.h index XXXXXXX..XXXXXXX 100644 --- a/tcg/mips/tcg-target-con-set.h +++ b/tcg/mips/tcg-target-con-set.h @@ -XXX,XX +XXX,XX @@ * tcg-target-con-str.h; the constraint combination is inclusive or. */ C_O0_I1(r) -C_O0_I2(rZ, r) -C_O0_I2(rZ, rZ) -C_O0_I3(rZ, r, r) -C_O0_I3(rZ, rZ, r) -C_O0_I4(rZ, rZ, rZ, rZ) -C_O0_I4(rZ, rZ, r, r) +C_O0_I2(rz, r) +C_O0_I2(rz, rz) +C_O0_I3(rz, r, r) +C_O0_I3(rz, rz, r) +C_O0_I4(rz, rz, rz, rz) +C_O0_I4(rz, rz, r, r) C_O1_I1(r, r) -C_O1_I2(r, 0, rZ) +C_O1_I2(r, 0, rz) C_O1_I2(r, r, r) C_O1_I2(r, r, ri) C_O1_I2(r, r, rI) C_O1_I2(r, r, rIK) C_O1_I2(r, r, rJ) -C_O1_I2(r, r, rWZ) -C_O1_I2(r, rZ, rN) -C_O1_I2(r, rZ, rZ) -C_O1_I4(r, rZ, rZ, rZ, 0) -C_O1_I4(r, rZ, rZ, rZ, rZ) +C_O1_I2(r, r, rzW) +C_O1_I2(r, rz, rN) +C_O1_I2(r, rz, rz) +C_O1_I4(r, rz, rz, rz, 0) +C_O1_I4(r, rz, rz, rz, rz) C_O2_I1(r, r, r) C_O2_I2(r, r, r, r) -C_O2_I4(r, r, rZ, rZ, rN, rN) +C_O2_I4(r, r, rz, rz, rN, rN) diff --git a/tcg/mips/tcg-target-con-str.h b/tcg/mips/tcg-target-con-str.h index XXXXXXX..XXXXXXX 100644 --- a/tcg/mips/tcg-target-con-str.h +++ b/tcg/mips/tcg-target-con-str.h @@ -XXX,XX +XXX,XX @@ CONST('J', TCG_CT_CONST_S16) CONST('K', TCG_CT_CONST_P2M1) CONST('N', TCG_CT_CONST_N16) CONST('W', TCG_CT_CONST_WSZ) -CONST('Z', TCG_CT_CONST_ZERO) diff --git a/tcg/mips/tcg-target.c.inc b/tcg/mips/tcg-target.c.inc index XXXXXXX..XXXXXXX 100644 --- a/tcg/mips/tcg-target.c.inc +++ b/tcg/mips/tcg-target.c.inc @@ -XXX,XX +XXX,XX @@ static bool patch_reloc(tcg_insn_unit *code_ptr, int type, g_assert_not_reached(); } -#define TCG_CT_CONST_ZERO 0x100 -#define TCG_CT_CONST_U16 0x200 /* Unsigned 16-bit: 0 - 0xffff. */ -#define TCG_CT_CONST_S16 0x400 /* Signed 16-bit: -32768 - 32767 */ -#define TCG_CT_CONST_P2M1 0x800 /* Power of 2 minus 1. */ -#define TCG_CT_CONST_N16 0x1000 /* "Negatable" 16-bit: -32767 - 32767 */ -#define TCG_CT_CONST_WSZ 0x2000 /* word size */ +#define TCG_CT_CONST_U16 0x100 /* Unsigned 16-bit: 0 - 0xffff. */ +#define TCG_CT_CONST_S16 0x200 /* Signed 16-bit: -32768 - 32767 */ +#define TCG_CT_CONST_P2M1 0x400 /* Power of 2 minus 1. */ +#define TCG_CT_CONST_N16 0x800 /* "Negatable" 16-bit: -32767 - 32767 */ +#define TCG_CT_CONST_WSZ 0x1000 /* word size */ #define ALL_GENERAL_REGS 0xffffffffu @@ -XXX,XX +XXX,XX @@ static bool tcg_target_const_match(int64_t val, int ct, { if (ct & TCG_CT_CONST) { return 1; - } else if ((ct & TCG_CT_CONST_ZERO) && val == 0) { - return 1; } else if ((ct & TCG_CT_CONST_U16) && val == (uint16_t)val) { return 1; } else if ((ct & TCG_CT_CONST_S16) && val == (int16_t)val) { @@ -XXX,XX +XXX,XX @@ static void tcg_out_op(TCGContext *s, TCGOpcode opc, TCGType type, TCGArg a0, a1, a2; int c2; - /* - * Note that many operands use the constraint set "rZ". - * We make use of the fact that 0 is the ZERO register, - * and hence such cases need not check for const_args. - */ a0 = args[0]; a1 = args[1]; a2 = args[2]; @@ -XXX,XX +XXX,XX @@ tcg_target_op_def(TCGOpcode op, TCGType type, unsigned flags) case INDEX_op_st16_i64: case INDEX_op_st32_i64: case INDEX_op_st_i64: - return C_O0_I2(rZ, r); + return C_O0_I2(rz, r); case INDEX_op_add_i32: case INDEX_op_add_i64: return C_O1_I2(r, r, rJ); case INDEX_op_sub_i32: case INDEX_op_sub_i64: - return C_O1_I2(r, rZ, rN); + return C_O1_I2(r, rz, rN); case INDEX_op_mul_i32: case INDEX_op_mulsh_i32: case INDEX_op_muluh_i32: @@ -XXX,XX +XXX,XX @@ tcg_target_op_def(TCGOpcode op, TCGType type, unsigned flags) case INDEX_op_remu_i64: case INDEX_op_nor_i64: case INDEX_op_setcond_i64: - return C_O1_I2(r, rZ, rZ); + return C_O1_I2(r, rz, rz); case INDEX_op_muls2_i32: case INDEX_op_mulu2_i32: case INDEX_op_muls2_i64: @@ -XXX,XX +XXX,XX @@ tcg_target_op_def(TCGOpcode op, TCGType type, unsigned flags) return C_O1_I2(r, r, ri); case INDEX_op_clz_i32: case INDEX_op_clz_i64: - return C_O1_I2(r, r, rWZ); + return C_O1_I2(r, r, rzW); case INDEX_op_deposit_i32: case INDEX_op_deposit_i64: - return C_O1_I2(r, 0, rZ); + return C_O1_I2(r, 0, rz); case INDEX_op_brcond_i32: case INDEX_op_brcond_i64: - return C_O0_I2(rZ, rZ); + return C_O0_I2(rz, rz); case INDEX_op_movcond_i32: case INDEX_op_movcond_i64: return (use_mips32r6_instructions - ? C_O1_I4(r, rZ, rZ, rZ, rZ) - : C_O1_I4(r, rZ, rZ, rZ, 0)); + ? C_O1_I4(r, rz, rz, rz, rz) + : C_O1_I4(r, rz, rz, rz, 0)); case INDEX_op_add2_i32: case INDEX_op_sub2_i32: - return C_O2_I4(r, r, rZ, rZ, rN, rN); + return C_O2_I4(r, r, rz, rz, rN, rN); case INDEX_op_setcond2_i32: - return C_O1_I4(r, rZ, rZ, rZ, rZ); + return C_O1_I4(r, rz, rz, rz, rz); case INDEX_op_brcond2_i32: - return C_O0_I4(rZ, rZ, rZ, rZ); + return C_O0_I4(rz, rz, rz, rz); case INDEX_op_qemu_ld_i32: return C_O1_I1(r, r); case INDEX_op_qemu_st_i32: - return C_O0_I2(rZ, r); + return C_O0_I2(rz, r); case INDEX_op_qemu_ld_i64: return TCG_TARGET_REG_BITS == 64 ? C_O1_I1(r, r) : C_O2_I1(r, r, r); case INDEX_op_qemu_st_i64: - return TCG_TARGET_REG_BITS == 64 ? C_O0_I2(rZ, r) : C_O0_I3(rZ, rZ, r); + return TCG_TARGET_REG_BITS == 64 ? C_O0_I2(rz, r) : C_O0_I3(rz, rz, r); default: return C_NotImplemented; -- 2.43.0 Replace target-specific 'Z' with generic 'z'. Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- tcg/riscv/tcg-target-con-set.h | 10 +++++----- tcg/riscv/tcg-target-con-str.h | 1 - tcg/riscv/tcg-target.c.inc | 28 ++++++++++++---------------- 3 files changed, 17 insertions(+), 22 deletions(-) diff --git a/tcg/riscv/tcg-target-con-set.h b/tcg/riscv/tcg-target-con-set.h index XXXXXXX..XXXXXXX 100644 --- a/tcg/riscv/tcg-target-con-set.h +++ b/tcg/riscv/tcg-target-con-set.h @@ -XXX,XX +XXX,XX @@ * tcg-target-con-str.h; the constraint combination is inclusive or. */ C_O0_I1(r) -C_O0_I2(rZ, r) -C_O0_I2(rZ, rZ) +C_O0_I2(rz, r) +C_O0_I2(rz, rz) C_O1_I1(r, r) C_O1_I2(r, r, ri) C_O1_I2(r, r, rI) C_O1_I2(r, r, rJ) -C_O1_I2(r, rZ, rN) -C_O1_I2(r, rZ, rZ) +C_O1_I2(r, rz, rN) +C_O1_I2(r, rz, rz) C_N1_I2(r, r, rM) C_O1_I4(r, r, rI, rM, rM) -C_O2_I4(r, r, rZ, rZ, rM, rM) +C_O2_I4(r, r, rz, rz, rM, rM) C_O0_I2(v, r) C_O1_I1(v, r) C_O1_I1(v, v) diff --git a/tcg/riscv/tcg-target-con-str.h b/tcg/riscv/tcg-target-con-str.h index XXXXXXX..XXXXXXX 100644 --- a/tcg/riscv/tcg-target-con-str.h +++ b/tcg/riscv/tcg-target-con-str.h @@ -XXX,XX +XXX,XX @@ CONST('K', TCG_CT_CONST_S5) CONST('L', TCG_CT_CONST_CMP_VI) CONST('N', TCG_CT_CONST_N12) CONST('M', TCG_CT_CONST_M12) -CONST('Z', TCG_CT_CONST_ZERO) diff --git a/tcg/riscv/tcg-target.c.inc b/tcg/riscv/tcg-target.c.inc index XXXXXXX..XXXXXXX 100644 --- a/tcg/riscv/tcg-target.c.inc +++ b/tcg/riscv/tcg-target.c.inc @@ -XXX,XX +XXX,XX @@ static TCGReg tcg_target_call_oarg_reg(TCGCallReturnKind kind, int slot) return TCG_REG_A0 + slot; } -#define TCG_CT_CONST_ZERO 0x100 -#define TCG_CT_CONST_S12 0x200 -#define TCG_CT_CONST_N12 0x400 -#define TCG_CT_CONST_M12 0x800 -#define TCG_CT_CONST_J12 0x1000 -#define TCG_CT_CONST_S5 0x2000 -#define TCG_CT_CONST_CMP_VI 0x4000 +#define TCG_CT_CONST_S12 0x100 +#define TCG_CT_CONST_N12 0x200 +#define TCG_CT_CONST_M12 0x400 +#define TCG_CT_CONST_J12 0x800 +#define TCG_CT_CONST_S5 0x1000 +#define TCG_CT_CONST_CMP_VI 0x2000 #define ALL_GENERAL_REGS MAKE_64BIT_MASK(0, 32) #define ALL_VECTOR_REGS MAKE_64BIT_MASK(32, 32) @@ -XXX,XX +XXX,XX @@ static bool tcg_target_const_match(int64_t val, int ct, if (ct & TCG_CT_CONST) { return 1; } - if ((ct & TCG_CT_CONST_ZERO) && val == 0) { - return 1; - } if (type >= TCG_TYPE_V64) { /* Val is replicated by VECE; extract the highest element. */ val >>= (-8 << vece) & 63; @@ -XXX,XX +XXX,XX @@ tcg_target_op_def(TCGOpcode op, TCGType type, unsigned flags) case INDEX_op_st16_i64: case INDEX_op_st32_i64: case INDEX_op_st_i64: - return C_O0_I2(rZ, r); + return C_O0_I2(rz, r); case INDEX_op_add_i32: case INDEX_op_and_i32: @@ -XXX,XX +XXX,XX @@ tcg_target_op_def(TCGOpcode op, TCGType type, unsigned flags) case INDEX_op_sub_i32: case INDEX_op_sub_i64: - return C_O1_I2(r, rZ, rN); + return C_O1_I2(r, rz, rN); case INDEX_op_mul_i32: case INDEX_op_mulsh_i32: @@ -XXX,XX +XXX,XX @@ tcg_target_op_def(TCGOpcode op, TCGType type, unsigned flags) case INDEX_op_divu_i64: case INDEX_op_rem_i64: case INDEX_op_remu_i64: - return C_O1_I2(r, rZ, rZ); + return C_O1_I2(r, rz, rz); case INDEX_op_shl_i32: case INDEX_op_shr_i32: @@ -XXX,XX +XXX,XX @@ tcg_target_op_def(TCGOpcode op, TCGType type, unsigned flags) case INDEX_op_brcond_i32: case INDEX_op_brcond_i64: - return C_O0_I2(rZ, rZ); + return C_O0_I2(rz, rz); case INDEX_op_movcond_i32: case INDEX_op_movcond_i64: @@ -XXX,XX +XXX,XX @@ tcg_target_op_def(TCGOpcode op, TCGType type, unsigned flags) case INDEX_op_add2_i64: case INDEX_op_sub2_i32: case INDEX_op_sub2_i64: - return C_O2_I4(r, r, rZ, rZ, rM, rM); + return C_O2_I4(r, r, rz, rz, rM, rM); case INDEX_op_qemu_ld_i32: case INDEX_op_qemu_ld_i64: return C_O1_I1(r, r); case INDEX_op_qemu_st_i32: case INDEX_op_qemu_st_i64: - return C_O0_I2(rZ, r); + return C_O0_I2(rz, r); case INDEX_op_st_vec: return C_O0_I2(v, r); -- 2.43.0 Replace target-specific 'Z' with generic 'z'. Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- tcg/sparc64/tcg-target-con-set.h | 12 ++++++------ tcg/sparc64/tcg-target-con-str.h | 1 - tcg/sparc64/tcg-target.c.inc | 17 +++++++---------- 3 files changed, 13 insertions(+), 17 deletions(-) diff --git a/tcg/sparc64/tcg-target-con-set.h b/tcg/sparc64/tcg-target-con-set.h index XXXXXXX..XXXXXXX 100644 --- a/tcg/sparc64/tcg-target-con-set.h +++ b/tcg/sparc64/tcg-target-con-set.h @@ -XXX,XX +XXX,XX @@ * tcg-target-con-str.h; the constraint combination is inclusive or. */ C_O0_I1(r) -C_O0_I2(rZ, r) -C_O0_I2(rZ, rJ) +C_O0_I2(rz, r) +C_O0_I2(rz, rJ) C_O1_I1(r, r) C_O1_I2(r, r, r) -C_O1_I2(r, rZ, rJ) -C_O1_I4(r, rZ, rJ, rI, 0) -C_O2_I2(r, r, rZ, rJ) -C_O2_I4(r, r, rZ, rZ, rJ, rJ) +C_O1_I2(r, rz, rJ) +C_O1_I4(r, rz, rJ, rI, 0) +C_O2_I2(r, r, rz, rJ) +C_O2_I4(r, r, rz, rz, rJ, rJ) diff --git a/tcg/sparc64/tcg-target-con-str.h b/tcg/sparc64/tcg-target-con-str.h index XXXXXXX..XXXXXXX 100644 --- a/tcg/sparc64/tcg-target-con-str.h +++ b/tcg/sparc64/tcg-target-con-str.h @@ -XXX,XX +XXX,XX @@ REGS('r', ALL_GENERAL_REGS) */ CONST('I', TCG_CT_CONST_S11) CONST('J', TCG_CT_CONST_S13) -CONST('Z', TCG_CT_CONST_ZERO) diff --git a/tcg/sparc64/tcg-target.c.inc b/tcg/sparc64/tcg-target.c.inc index XXXXXXX..XXXXXXX 100644 --- a/tcg/sparc64/tcg-target.c.inc +++ b/tcg/sparc64/tcg-target.c.inc @@ -XXX,XX +XXX,XX @@ static const char * const tcg_target_reg_names[TCG_TARGET_NB_REGS] = { #define TCG_CT_CONST_S11 0x100 #define TCG_CT_CONST_S13 0x200 -#define TCG_CT_CONST_ZERO 0x400 #define ALL_GENERAL_REGS MAKE_64BIT_MASK(0, 32) @@ -XXX,XX +XXX,XX @@ static bool tcg_target_const_match(int64_t val, int ct, val = (int32_t)val; } - if ((ct & TCG_CT_CONST_ZERO) && val == 0) { - return 1; - } else if ((ct & TCG_CT_CONST_S11) && check_fit_tl(val, 11)) { + if ((ct & TCG_CT_CONST_S11) && check_fit_tl(val, 11)) { return 1; } else if ((ct & TCG_CT_CONST_S13) && check_fit_tl(val, 13)) { return 1; @@ -XXX,XX +XXX,XX @@ tcg_target_op_def(TCGOpcode op, TCGType type, unsigned flags) case INDEX_op_st_i64: case INDEX_op_qemu_st_i32: case INDEX_op_qemu_st_i64: - return C_O0_I2(rZ, r); + return C_O0_I2(rz, r); case INDEX_op_add_i32: case INDEX_op_add_i64: @@ -XXX,XX +XXX,XX @@ tcg_target_op_def(TCGOpcode op, TCGType type, unsigned flags) case INDEX_op_setcond_i64: case INDEX_op_negsetcond_i32: case INDEX_op_negsetcond_i64: - return C_O1_I2(r, rZ, rJ); + return C_O1_I2(r, rz, rJ); case INDEX_op_brcond_i32: case INDEX_op_brcond_i64: - return C_O0_I2(rZ, rJ); + return C_O0_I2(rz, rJ); case INDEX_op_movcond_i32: case INDEX_op_movcond_i64: - return C_O1_I4(r, rZ, rJ, rI, 0); + return C_O1_I4(r, rz, rJ, rI, 0); case INDEX_op_add2_i32: case INDEX_op_add2_i64: case INDEX_op_sub2_i32: case INDEX_op_sub2_i64: - return C_O2_I4(r, r, rZ, rZ, rJ, rJ); + return C_O2_I4(r, r, rz, rz, rJ, rJ); case INDEX_op_mulu2_i32: case INDEX_op_muls2_i32: - return C_O2_I2(r, r, rZ, rJ); + return C_O2_I2(r, r, rz, rJ); case INDEX_op_muluh_i64: return C_O1_I2(r, r, r); -- 2.43.0 From: Fabiano Rosas <farosas@suse.de> When complying with the alignment requested in the ELF and unmapping the excess reservation, having align_end not aligned to the guest page causes the unmap to be rejected by the alignment check at target_munmap and later brk adjustments hit an EEXIST. Fix by aligning the start of region to be unmapped. Fixes: c81d1fafa6 ("linux-user: Honor elf alignment when placing images") Resolves: https://gitlab.com/qemu-project/qemu/-/issues/1913 Signed-off-by: Fabiano Rosas <farosas@suse.de> [rth: Align load_end as well.] Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-ID: <20250213143558.10504-1-farosas@suse.de> --- linux-user/elfload.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/linux-user/elfload.c b/linux-user/elfload.c index XXXXXXX..XXXXXXX 100644 --- a/linux-user/elfload.c +++ b/linux-user/elfload.c @@ -XXX,XX +XXX,XX @@ static void load_elf_image(const char *image_name, const ImageSource *src, if (align_size != reserve_size) { abi_ulong align_addr = ROUND_UP(load_addr, align); - abi_ulong align_end = align_addr + reserve_size; - abi_ulong load_end = load_addr + align_size; + abi_ulong align_end = TARGET_PAGE_ALIGN(align_addr + reserve_size); + abi_ulong load_end = TARGET_PAGE_ALIGN(load_addr + align_size); if (align_addr != load_addr) { target_munmap(load_addr, align_addr - load_addr); -- 2.43.0 From: Andreas Schwab <schwab@suse.de> SA_RESTORER and the associated sa_restorer field of struct sigaction are an obsolete feature, not expected to be used by future architectures. They are also absent on RISC-V, LoongArch, Hexagon and OpenRISC, but defined due to their use of generic/signal.h. This leads to corrupted data and out-of-bounds accesses. Move the definition of TARGET_SA_RESTORER out of generic/signal.h into the target_signal.h files that need it. Note that m68k has the sa_restorer field, but does not use it and does not define SA_RESTORER. Reported-by: Thomas Weißschuh <thomas@t-8ch.de> Signed-off-by: Andreas Schwab <schwab@suse.de> Reviewed-by: Thomas Weißschuh <thomas@t-8ch.de> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-ID: <mvmed060xc9.fsf@suse.de> --- linux-user/aarch64/target_signal.h | 2 ++ linux-user/arm/target_signal.h | 2 ++ linux-user/generic/signal.h | 1 - linux-user/i386/target_signal.h | 2 ++ linux-user/m68k/target_signal.h | 1 + linux-user/microblaze/target_signal.h | 2 ++ linux-user/ppc/target_signal.h | 2 ++ linux-user/s390x/target_signal.h | 2 ++ linux-user/sh4/target_signal.h | 2 ++ linux-user/x86_64/target_signal.h | 2 ++ linux-user/xtensa/target_signal.h | 2 ++ 11 files changed, 19 insertions(+), 1 deletion(-) diff --git a/linux-user/aarch64/target_signal.h b/linux-user/aarch64/target_signal.h index XXXXXXX..XXXXXXX 100644 --- a/linux-user/aarch64/target_signal.h +++ b/linux-user/aarch64/target_signal.h @@ -XXX,XX +XXX,XX @@ #include "../generic/signal.h" +#define TARGET_SA_RESTORER 0x04000000 + #define TARGET_SEGV_MTEAERR 8 /* Asynchronous ARM MTE error */ #define TARGET_SEGV_MTESERR 9 /* Synchronous ARM MTE exception */ diff --git a/linux-user/arm/target_signal.h b/linux-user/arm/target_signal.h index XXXXXXX..XXXXXXX 100644 --- a/linux-user/arm/target_signal.h +++ b/linux-user/arm/target_signal.h @@ -XXX,XX +XXX,XX @@ #include "../generic/signal.h" +#define TARGET_SA_RESTORER 0x04000000 + #define TARGET_ARCH_HAS_SETUP_FRAME #define TARGET_ARCH_HAS_SIGTRAMP_PAGE 1 diff --git a/linux-user/generic/signal.h b/linux-user/generic/signal.h index XXXXXXX..XXXXXXX 100644 --- a/linux-user/generic/signal.h +++ b/linux-user/generic/signal.h @@ -XXX,XX +XXX,XX @@ #define TARGET_SA_RESTART 0x10000000 #define TARGET_SA_NODEFER 0x40000000 #define TARGET_SA_RESETHAND 0x80000000 -#define TARGET_SA_RESTORER 0x04000000 #define TARGET_SIGHUP 1 #define TARGET_SIGINT 2 diff --git a/linux-user/i386/target_signal.h b/linux-user/i386/target_signal.h index XXXXXXX..XXXXXXX 100644 --- a/linux-user/i386/target_signal.h +++ b/linux-user/i386/target_signal.h @@ -XXX,XX +XXX,XX @@ #include "../generic/signal.h" +#define TARGET_SA_RESTORER 0x04000000 + #define TARGET_ARCH_HAS_SETUP_FRAME #define TARGET_ARCH_HAS_SIGTRAMP_PAGE 1 diff --git a/linux-user/m68k/target_signal.h b/linux-user/m68k/target_signal.h index XXXXXXX..XXXXXXX 100644 --- a/linux-user/m68k/target_signal.h +++ b/linux-user/m68k/target_signal.h @@ -XXX,XX +XXX,XX @@ #include "../generic/signal.h" +#define TARGET_ARCH_HAS_SA_RESTORER 1 #define TARGET_ARCH_HAS_SETUP_FRAME #define TARGET_ARCH_HAS_SIGTRAMP_PAGE 1 diff --git a/linux-user/microblaze/target_signal.h b/linux-user/microblaze/target_signal.h index XXXXXXX..XXXXXXX 100644 --- a/linux-user/microblaze/target_signal.h +++ b/linux-user/microblaze/target_signal.h @@ -XXX,XX +XXX,XX @@ #include "../generic/signal.h" +#define TARGET_SA_RESTORER 0x04000000 + #define TARGET_ARCH_HAS_SIGTRAMP_PAGE 1 #endif /* MICROBLAZE_TARGET_SIGNAL_H */ diff --git a/linux-user/ppc/target_signal.h b/linux-user/ppc/target_signal.h index XXXXXXX..XXXXXXX 100644 --- a/linux-user/ppc/target_signal.h +++ b/linux-user/ppc/target_signal.h @@ -XXX,XX +XXX,XX @@ #include "../generic/signal.h" +#define TARGET_SA_RESTORER 0x04000000 + #if !defined(TARGET_PPC64) #define TARGET_ARCH_HAS_SETUP_FRAME #endif diff --git a/linux-user/s390x/target_signal.h b/linux-user/s390x/target_signal.h index XXXXXXX..XXXXXXX 100644 --- a/linux-user/s390x/target_signal.h +++ b/linux-user/s390x/target_signal.h @@ -XXX,XX +XXX,XX @@ #include "../generic/signal.h" +#define TARGET_SA_RESTORER 0x04000000 + #define TARGET_ARCH_HAS_SETUP_FRAME #define TARGET_ARCH_HAS_SIGTRAMP_PAGE 1 diff --git a/linux-user/sh4/target_signal.h b/linux-user/sh4/target_signal.h index XXXXXXX..XXXXXXX 100644 --- a/linux-user/sh4/target_signal.h +++ b/linux-user/sh4/target_signal.h @@ -XXX,XX +XXX,XX @@ #include "../generic/signal.h" +#define TARGET_SA_RESTORER 0x04000000 + #define TARGET_ARCH_HAS_SETUP_FRAME #define TARGET_ARCH_HAS_SIGTRAMP_PAGE 1 diff --git a/linux-user/x86_64/target_signal.h b/linux-user/x86_64/target_signal.h index XXXXXXX..XXXXXXX 100644 --- a/linux-user/x86_64/target_signal.h +++ b/linux-user/x86_64/target_signal.h @@ -XXX,XX +XXX,XX @@ #include "../generic/signal.h" +#define TARGET_SA_RESTORER 0x04000000 + /* For x86_64, use of SA_RESTORER is mandatory. */ #define TARGET_ARCH_HAS_SIGTRAMP_PAGE 0 diff --git a/linux-user/xtensa/target_signal.h b/linux-user/xtensa/target_signal.h index XXXXXXX..XXXXXXX 100644 --- a/linux-user/xtensa/target_signal.h +++ b/linux-user/xtensa/target_signal.h @@ -XXX,XX +XXX,XX @@ #include "../generic/signal.h" +#define TARGET_SA_RESTORER 0x04000000 + #define TARGET_ARCH_HAS_SIGTRAMP_PAGE 1 #endif -- 2.43.0 From: Mikael Szreder <git@miszr.win> A bug was introduced in commit 0bba7572d40d which causes the fdtox and fqtox instructions to incorrectly select the destination registers. More information and a test program can be found in issue #2802. Fixes: 0bba7572d40d ("target/sparc: Perform DFPREG/QFPREG in decodetree") Resolves: https://gitlab.com/qemu-project/qemu/-/issues/2802 Signed-off-by: Mikael Szreder <git@miszr.win> Acked-by: Artyom Tarasenko <atar4qemu@gmail.com> [rth: Squash patches together, since the second fixes a typo in the first.] Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-ID: <20250205090333.19626-3-git@miszr.win> --- target/sparc/insns.decode | 12 ++++++------ 1 file changed, 6 insertions(+), 6 deletions(-) diff --git a/target/sparc/insns.decode b/target/sparc/insns.decode index XXXXXXX..XXXXXXX 100644 --- a/target/sparc/insns.decode +++ b/target/sparc/insns.decode @@ -XXX,XX +XXX,XX @@ FdMULq 10 ..... 110100 ..... 0 0110 1110 ..... @q_d_d FNHADDs 10 ..... 110100 ..... 0 0111 0001 ..... @r_r_r FNHADDd 10 ..... 110100 ..... 0 0111 0010 ..... @d_d_d FNsMULd 10 ..... 110100 ..... 0 0111 1001 ..... @d_r_r -FsTOx 10 ..... 110100 00000 0 1000 0001 ..... @r_r2 -FdTOx 10 ..... 110100 00000 0 1000 0010 ..... @r_d2 -FqTOx 10 ..... 110100 00000 0 1000 0011 ..... @r_q2 -FxTOs 10 ..... 110100 00000 0 1000 0100 ..... @r_r2 -FxTOd 10 ..... 110100 00000 0 1000 1000 ..... @d_r2 -FxTOq 10 ..... 110100 00000 0 1000 1100 ..... @q_r2 +FsTOx 10 ..... 110100 00000 0 1000 0001 ..... @d_r2 +FdTOx 10 ..... 110100 00000 0 1000 0010 ..... @d_d2 +FqTOx 10 ..... 110100 00000 0 1000 0011 ..... @d_q2 +FxTOs 10 ..... 110100 00000 0 1000 0100 ..... @r_d2 +FxTOd 10 ..... 110100 00000 0 1000 1000 ..... @d_d2 +FxTOq 10 ..... 110100 00000 0 1000 1100 ..... @q_d2 FiTOs 10 ..... 110100 00000 0 1100 0100 ..... @r_r2 FdTOs 10 ..... 110100 00000 0 1100 0110 ..... @r_d2 FqTOs 10 ..... 110100 00000 0 1100 0111 ..... @r_q2 -- 2.43.0 From: Mikael Szreder <git@miszr.win> The gdbstub implementation for the Sparc architecture would incorrectly calculate the the floating point register offset. This resulted in, for example, registers f32 and f34 to point to the same value. The issue was caused by the confusion between even register numbers and even register indexes. For example, the register index of f32 is 64 and f34 is 65. Fixes: 30038fd81808 ("target-sparc: Change fpr representation to doubles.") Signed-off-by: Mikael Szreder <git@miszr.win> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-ID: <20250214070343.11501-1-git@miszr.win> --- target/sparc/gdbstub.c | 18 ++++++++++++++---- 1 file changed, 14 insertions(+), 4 deletions(-) diff --git a/target/sparc/gdbstub.c b/target/sparc/gdbstub.c index XXXXXXX..XXXXXXX 100644 --- a/target/sparc/gdbstub.c +++ b/target/sparc/gdbstub.c @@ -XXX,XX +XXX,XX @@ int sparc_cpu_gdb_read_register(CPUState *cs, GByteArray *mem_buf, int n) } } if (n < 80) { - /* f32-f62 (double width, even numbers only) */ - return gdb_get_reg64(mem_buf, env->fpr[(n - 32) / 2].ll); + /* f32-f62 (16 double width registers, even register numbers only) + * n == 64: f32 : env->fpr[16] + * n == 65: f34 : env->fpr[17] + * etc... + * n == 79: f62 : env->fpr[31] + */ + return gdb_get_reg64(mem_buf, env->fpr[(n - 64) + 16].ll); } switch (n) { case 80: @@ -XXX,XX +XXX,XX @@ int sparc_cpu_gdb_write_register(CPUState *cs, uint8_t *mem_buf, int n) } return 4; } else if (n < 80) { - /* f32-f62 (double width, even numbers only) */ - env->fpr[(n - 32) / 2].ll = tmp; + /* f32-f62 (16 double width registers, even register numbers only) + * n == 64: f32 : env->fpr[16] + * n == 65: f34 : env->fpr[17] + * etc... + * n == 79: f62 : env->fpr[31] + */ + env->fpr[(n - 64) + 16].ll = tmp; } else { switch (n) { case 80: -- 2.43.0 From: Artyom Tarasenko <atar4qemu@gmail.com> Fake access to PCR Performance Control Register and PIC Performance Instrumentation Counter. Ignore writes in privileged mode, and return 0 on reads. This allows booting Tribblix, MilaX and v9os under Niagara target. Signed-off-by: Artyom Tarasenko <atar4qemu@gmail.com> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-ID: <20250209211248.50383-1-atar4qemu@gmail.com> --- target/sparc/translate.c | 19 +++++++++++++++++++ target/sparc/insns.decode | 7 ++++++- 2 files changed, 25 insertions(+), 1 deletion(-) 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 TCGv do_rd_leon3_config(DisasContext *dc, TCGv dst) TRANS(RDASR17, ASR17, do_rd_special, true, a->rd, do_rd_leon3_config) +static TCGv do_rdpic(DisasContext *dc, TCGv dst) +{ + return tcg_constant_tl(0); +} + +TRANS(RDPIC, HYPV, do_rd_special, supervisor(dc), a->rd, do_rdpic) + + static TCGv do_rdccr(DisasContext *dc, TCGv dst) { gen_helper_rdccr(dst, tcg_env); @@ -XXX,XX +XXX,XX @@ static void do_wrfprs(DisasContext *dc, TCGv src) TRANS(WRFPRS, 64, do_wr_special, a, true, do_wrfprs) +static bool do_priv_nop(DisasContext *dc, bool priv) +{ + if (!priv) { + return raise_priv(dc); + } + return advance_pc(dc); +} + +TRANS(WRPCR, HYPV, do_priv_nop, supervisor(dc)) +TRANS(WRPIC, HYPV, do_priv_nop, supervisor(dc)) + static void do_wrgsr(DisasContext *dc, TCGv src) { gen_trap_ifnofpu(dc); diff --git a/target/sparc/insns.decode b/target/sparc/insns.decode index XXXXXXX..XXXXXXX 100644 --- a/target/sparc/insns.decode +++ b/target/sparc/insns.decode @@ -XXX,XX +XXX,XX @@ CALL 01 i:s30 RDTICK 10 rd:5 101000 00100 0 0000000000000 RDPC 10 rd:5 101000 00101 0 0000000000000 RDFPRS 10 rd:5 101000 00110 0 0000000000000 - RDASR17 10 rd:5 101000 10001 0 0000000000000 + { + RDASR17 10 rd:5 101000 10001 0 0000000000000 + RDPIC 10 rd:5 101000 10001 0 0000000000000 + } RDGSR 10 rd:5 101000 10011 0 0000000000000 RDSOFTINT 10 rd:5 101000 10110 0 0000000000000 RDTICK_CMPR 10 rd:5 101000 10111 0 0000000000000 @@ -XXX,XX +XXX,XX @@ CALL 01 i:s30 WRCCR 10 00010 110000 ..... . ............. @n_r_ri WRASI 10 00011 110000 ..... . ............. @n_r_ri WRFPRS 10 00110 110000 ..... . ............. @n_r_ri + WRPCR 10 10000 110000 01000 0 0000000000000 + WRPIC 10 10001 110000 01000 0 0000000000000 { WRGSR 10 10011 110000 ..... . ............. @n_r_ri WRPOWERDOWN 10 10011 110000 ..... . ............. @n_r_ri -- 2.43.0