* [PATCH 01/10] accel/tcg: Generalize fake_user_interrupt test
2025-04-28 20:10 [PATCH 00/10] accel/tcg: Compile cpu-exec.c twice Richard Henderson
@ 2025-04-28 20:10 ` Richard Henderson
2025-04-28 21:36 ` Pierrick Bouvier
2025-04-29 6:45 ` Philippe Mathieu-Daudé
2025-04-28 20:10 ` [PATCH 02/10] accel/tcg: Unconditionally use CPU_DUMP_CCOP in log_cpu_exec Richard Henderson
` (8 subsequent siblings)
9 siblings, 2 replies; 32+ messages in thread
From: Richard Henderson @ 2025-04-28 20:10 UTC (permalink / raw)
To: qemu-devel; +Cc: pierrick.bouvier, philmd
Test for the hook being present instead of ifdef TARGET_I386.
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
---
accel/tcg/cpu-exec.c | 6 +++---
1 file changed, 3 insertions(+), 3 deletions(-)
diff --git a/accel/tcg/cpu-exec.c b/accel/tcg/cpu-exec.c
index 279df5fae7..8ff4a34509 100644
--- a/accel/tcg/cpu-exec.c
+++ b/accel/tcg/cpu-exec.c
@@ -732,10 +732,10 @@ static inline bool cpu_handle_exception(CPUState *cpu, int *ret)
* If user mode only, we simulate a fake exception which will be
* handled outside the cpu execution loop.
*/
-#if defined(TARGET_I386)
const TCGCPUOps *tcg_ops = cpu->cc->tcg_ops;
- tcg_ops->fake_user_interrupt(cpu);
-#endif /* TARGET_I386 */
+ if (tcg_ops->fake_user_interrupt) {
+ tcg_ops->fake_user_interrupt(cpu);
+ }
*ret = cpu->exception_index;
cpu->exception_index = -1;
return true;
--
2.43.0
^ permalink raw reply related [flat|nested] 32+ messages in thread* Re: [PATCH 01/10] accel/tcg: Generalize fake_user_interrupt test
2025-04-28 20:10 ` [PATCH 01/10] accel/tcg: Generalize fake_user_interrupt test Richard Henderson
@ 2025-04-28 21:36 ` Pierrick Bouvier
2025-04-29 6:45 ` Philippe Mathieu-Daudé
1 sibling, 0 replies; 32+ messages in thread
From: Pierrick Bouvier @ 2025-04-28 21:36 UTC (permalink / raw)
To: Richard Henderson, qemu-devel; +Cc: philmd
On 4/28/25 1:10 PM, Richard Henderson wrote:
> Test for the hook being present instead of ifdef TARGET_I386.
>
> Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
> ---
> accel/tcg/cpu-exec.c | 6 +++---
> 1 file changed, 3 insertions(+), 3 deletions(-)
Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org>
^ permalink raw reply [flat|nested] 32+ messages in thread
* Re: [PATCH 01/10] accel/tcg: Generalize fake_user_interrupt test
2025-04-28 20:10 ` [PATCH 01/10] accel/tcg: Generalize fake_user_interrupt test Richard Henderson
2025-04-28 21:36 ` Pierrick Bouvier
@ 2025-04-29 6:45 ` Philippe Mathieu-Daudé
1 sibling, 0 replies; 32+ messages in thread
From: Philippe Mathieu-Daudé @ 2025-04-29 6:45 UTC (permalink / raw)
To: Richard Henderson, qemu-devel; +Cc: pierrick.bouvier
On 28/4/25 22:10, Richard Henderson wrote:
> Test for the hook being present instead of ifdef TARGET_I386.
>
> Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
> ---
> accel/tcg/cpu-exec.c | 6 +++---
> 1 file changed, 3 insertions(+), 3 deletions(-)
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
^ permalink raw reply [flat|nested] 32+ messages in thread
* [PATCH 02/10] accel/tcg: Unconditionally use CPU_DUMP_CCOP in log_cpu_exec
2025-04-28 20:10 [PATCH 00/10] accel/tcg: Compile cpu-exec.c twice Richard Henderson
2025-04-28 20:10 ` [PATCH 01/10] accel/tcg: Generalize fake_user_interrupt test Richard Henderson
@ 2025-04-28 20:10 ` Richard Henderson
2025-04-28 21:37 ` Pierrick Bouvier
2025-04-28 20:10 ` [PATCH 03/10] accel/tcg: Introduce TCGCPUOps.cpu_exec_reset Richard Henderson
` (7 subsequent siblings)
9 siblings, 1 reply; 32+ messages in thread
From: Richard Henderson @ 2025-04-28 20:10 UTC (permalink / raw)
To: qemu-devel; +Cc: pierrick.bouvier, philmd
This flag is only tested by target/i386, so including this
makes no functional change. This is similar to other places
like cpu-target.c which use CPU_DUMP_CCOP unconditionally.
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
---
We could just as easily remove CPU_DUMP_CCOP entirely, and let
target/i386 dump CCOP along with general regs like other targets.
---
accel/tcg/cpu-exec.c | 5 +----
1 file changed, 1 insertion(+), 4 deletions(-)
diff --git a/accel/tcg/cpu-exec.c b/accel/tcg/cpu-exec.c
index 8ff4a34509..ff979a2c57 100644
--- a/accel/tcg/cpu-exec.c
+++ b/accel/tcg/cpu-exec.c
@@ -285,14 +285,11 @@ static void log_cpu_exec(vaddr pc, CPUState *cpu,
if (qemu_loglevel_mask(CPU_LOG_TB_CPU)) {
FILE *logfile = qemu_log_trylock();
if (logfile) {
- int flags = 0;
+ int flags = CPU_DUMP_CCOP;
if (qemu_loglevel_mask(CPU_LOG_TB_FPU)) {
flags |= CPU_DUMP_FPU;
}
-#if defined(TARGET_I386)
- flags |= CPU_DUMP_CCOP;
-#endif
if (qemu_loglevel_mask(CPU_LOG_TB_VPU)) {
flags |= CPU_DUMP_VPU;
}
--
2.43.0
^ permalink raw reply related [flat|nested] 32+ messages in thread* Re: [PATCH 02/10] accel/tcg: Unconditionally use CPU_DUMP_CCOP in log_cpu_exec
2025-04-28 20:10 ` [PATCH 02/10] accel/tcg: Unconditionally use CPU_DUMP_CCOP in log_cpu_exec Richard Henderson
@ 2025-04-28 21:37 ` Pierrick Bouvier
0 siblings, 0 replies; 32+ messages in thread
From: Pierrick Bouvier @ 2025-04-28 21:37 UTC (permalink / raw)
To: Richard Henderson, qemu-devel; +Cc: philmd
On 4/28/25 1:10 PM, Richard Henderson wrote:
> This flag is only tested by target/i386, so including this
> makes no functional change. This is similar to other places
> like cpu-target.c which use CPU_DUMP_CCOP unconditionally.
>
> Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
> ---
> We could just as easily remove CPU_DUMP_CCOP entirely, and let
> target/i386 dump CCOP along with general regs like other targets.
> ---
> accel/tcg/cpu-exec.c | 5 +----
> 1 file changed, 1 insertion(+), 4 deletions(-)
Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org>
^ permalink raw reply [flat|nested] 32+ messages in thread
* [PATCH 03/10] accel/tcg: Introduce TCGCPUOps.cpu_exec_reset
2025-04-28 20:10 [PATCH 00/10] accel/tcg: Compile cpu-exec.c twice Richard Henderson
2025-04-28 20:10 ` [PATCH 01/10] accel/tcg: Generalize fake_user_interrupt test Richard Henderson
2025-04-28 20:10 ` [PATCH 02/10] accel/tcg: Unconditionally use CPU_DUMP_CCOP in log_cpu_exec Richard Henderson
@ 2025-04-28 20:10 ` Richard Henderson
2025-04-28 21:38 ` Pierrick Bouvier
2025-04-29 6:44 ` Philippe Mathieu-Daudé
2025-04-28 20:10 ` [PATCH 04/10] target/i386: Split out x86_cpu_exec_reset Richard Henderson
` (6 subsequent siblings)
9 siblings, 2 replies; 32+ messages in thread
From: Richard Henderson @ 2025-04-28 20:10 UTC (permalink / raw)
To: qemu-devel; +Cc: pierrick.bouvier, philmd
Initialize all instances with cpu_reset(), so that there
is no functional change.
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
---
include/accel/tcg/cpu-ops.h | 2 ++
accel/tcg/cpu-exec.c | 3 ++-
target/alpha/cpu.c | 1 +
target/arm/cpu.c | 1 +
target/arm/tcg/cpu-v7m.c | 1 +
target/avr/cpu.c | 1 +
target/hppa/cpu.c | 1 +
target/i386/tcg/tcg-cpu.c | 1 +
target/loongarch/cpu.c | 1 +
target/m68k/cpu.c | 1 +
target/microblaze/cpu.c | 1 +
target/mips/cpu.c | 1 +
target/openrisc/cpu.c | 1 +
target/ppc/cpu_init.c | 1 +
target/riscv/tcg/tcg-cpu.c | 1 +
target/rx/cpu.c | 1 +
target/s390x/cpu.c | 1 +
target/sh4/cpu.c | 1 +
target/sparc/cpu.c | 1 +
target/tricore/cpu.c | 1 +
target/xtensa/cpu.c | 1 +
21 files changed, 23 insertions(+), 1 deletion(-)
diff --git a/include/accel/tcg/cpu-ops.h b/include/accel/tcg/cpu-ops.h
index 60b5e97205..3ff72b8d9d 100644
--- a/include/accel/tcg/cpu-ops.h
+++ b/include/accel/tcg/cpu-ops.h
@@ -155,6 +155,8 @@ struct TCGCPUOps {
void (*do_interrupt)(CPUState *cpu);
/** @cpu_exec_interrupt: Callback for processing interrupts in cpu_exec */
bool (*cpu_exec_interrupt)(CPUState *cpu, int interrupt_request);
+ /** @cpu_exec_reset: Callback for reset in cpu_exec. */
+ void (*cpu_exec_reset)(CPUState *cpu);
/**
* @cpu_exec_halt: Callback for handling halt in cpu_exec.
*
diff --git a/accel/tcg/cpu-exec.c b/accel/tcg/cpu-exec.c
index ff979a2c57..010f38edaa 100644
--- a/accel/tcg/cpu-exec.c
+++ b/accel/tcg/cpu-exec.c
@@ -834,7 +834,7 @@ static inline bool cpu_handle_interrupt(CPUState *cpu,
#else
else if (interrupt_request & CPU_INTERRUPT_RESET) {
replay_interrupt();
- cpu_reset(cpu);
+ cpu->cc->tcg_ops->cpu_exec_reset(cpu);
bql_unlock();
return true;
}
@@ -1070,6 +1070,7 @@ bool tcg_exec_realizefn(CPUState *cpu, Error **errp)
#ifndef CONFIG_USER_ONLY
assert(tcg_ops->cpu_exec_halt);
assert(tcg_ops->cpu_exec_interrupt);
+ assert(tcg_ops->cpu_exec_reset);
#endif /* !CONFIG_USER_ONLY */
assert(tcg_ops->translate_code);
assert(tcg_ops->mmu_index);
diff --git a/target/alpha/cpu.c b/target/alpha/cpu.c
index cb3f382dc6..9dc1c7242b 100644
--- a/target/alpha/cpu.c
+++ b/target/alpha/cpu.c
@@ -251,6 +251,7 @@ static const TCGCPUOps alpha_tcg_ops = {
.tlb_fill = alpha_cpu_tlb_fill,
.cpu_exec_interrupt = alpha_cpu_exec_interrupt,
.cpu_exec_halt = alpha_cpu_has_work,
+ .cpu_exec_reset = cpu_reset,
.do_interrupt = alpha_cpu_do_interrupt,
.do_transaction_failed = alpha_cpu_do_transaction_failed,
.do_unaligned_access = alpha_cpu_do_unaligned_access,
diff --git a/target/arm/cpu.c b/target/arm/cpu.c
index e149a46250..9e74d7bc07 100644
--- a/target/arm/cpu.c
+++ b/target/arm/cpu.c
@@ -2688,6 +2688,7 @@ static const TCGCPUOps arm_tcg_ops = {
.tlb_fill_align = arm_cpu_tlb_fill_align,
.cpu_exec_interrupt = arm_cpu_exec_interrupt,
.cpu_exec_halt = arm_cpu_exec_halt,
+ .cpu_exec_reset = cpu_reset,
.do_interrupt = arm_cpu_do_interrupt,
.do_transaction_failed = arm_cpu_do_transaction_failed,
.do_unaligned_access = arm_cpu_do_unaligned_access,
diff --git a/target/arm/tcg/cpu-v7m.c b/target/arm/tcg/cpu-v7m.c
index f71560aa43..57ed3f3a06 100644
--- a/target/arm/tcg/cpu-v7m.c
+++ b/target/arm/tcg/cpu-v7m.c
@@ -250,6 +250,7 @@ static const TCGCPUOps arm_v7m_tcg_ops = {
.tlb_fill_align = arm_cpu_tlb_fill_align,
.cpu_exec_interrupt = arm_v7m_cpu_exec_interrupt,
.cpu_exec_halt = arm_cpu_exec_halt,
+ .cpu_exec_reset = cpu_reset,
.do_interrupt = arm_v7m_cpu_do_interrupt,
.do_transaction_failed = arm_cpu_do_transaction_failed,
.do_unaligned_access = arm_cpu_do_unaligned_access,
diff --git a/target/avr/cpu.c b/target/avr/cpu.c
index 86e53ef9f6..e50ebfd8f8 100644
--- a/target/avr/cpu.c
+++ b/target/avr/cpu.c
@@ -232,6 +232,7 @@ static const TCGCPUOps avr_tcg_ops = {
.mmu_index = avr_cpu_mmu_index,
.cpu_exec_interrupt = avr_cpu_exec_interrupt,
.cpu_exec_halt = avr_cpu_has_work,
+ .cpu_exec_reset = cpu_reset,
.tlb_fill = avr_cpu_tlb_fill,
.do_interrupt = avr_cpu_do_interrupt,
};
diff --git a/target/hppa/cpu.c b/target/hppa/cpu.c
index 6acbf3de27..300c6b809f 100644
--- a/target/hppa/cpu.c
+++ b/target/hppa/cpu.c
@@ -271,6 +271,7 @@ static const TCGCPUOps hppa_tcg_ops = {
.tlb_fill_align = hppa_cpu_tlb_fill_align,
.cpu_exec_interrupt = hppa_cpu_exec_interrupt,
.cpu_exec_halt = hppa_cpu_has_work,
+ .cpu_exec_reset = cpu_reset,
.do_interrupt = hppa_cpu_do_interrupt,
.do_unaligned_access = hppa_cpu_do_unaligned_access,
.do_transaction_failed = hppa_cpu_do_transaction_failed,
diff --git a/target/i386/tcg/tcg-cpu.c b/target/i386/tcg/tcg-cpu.c
index d441c0752e..63ddf87aa8 100644
--- a/target/i386/tcg/tcg-cpu.c
+++ b/target/i386/tcg/tcg-cpu.c
@@ -147,6 +147,7 @@ const TCGCPUOps x86_tcg_ops = {
.do_interrupt = x86_cpu_do_interrupt,
.cpu_exec_halt = x86_cpu_exec_halt,
.cpu_exec_interrupt = x86_cpu_exec_interrupt,
+ .cpu_exec_reset = cpu_reset,
.do_unaligned_access = x86_cpu_do_unaligned_access,
.debug_excp_handler = breakpoint_handler,
.debug_check_breakpoint = x86_debug_check_breakpoint,
diff --git a/target/loongarch/cpu.c b/target/loongarch/cpu.c
index 1fb6633581..3cb7aa6039 100644
--- a/target/loongarch/cpu.c
+++ b/target/loongarch/cpu.c
@@ -877,6 +877,7 @@ static const TCGCPUOps loongarch_tcg_ops = {
.tlb_fill = loongarch_cpu_tlb_fill,
.cpu_exec_interrupt = loongarch_cpu_exec_interrupt,
.cpu_exec_halt = loongarch_cpu_has_work,
+ .cpu_exec_reset = cpu_reset,
.do_interrupt = loongarch_cpu_do_interrupt,
.do_transaction_failed = loongarch_cpu_do_transaction_failed,
#endif
diff --git a/target/m68k/cpu.c b/target/m68k/cpu.c
index 99adc5eb91..c8c1dbeac9 100644
--- a/target/m68k/cpu.c
+++ b/target/m68k/cpu.c
@@ -602,6 +602,7 @@ static const TCGCPUOps m68k_tcg_ops = {
.tlb_fill = m68k_cpu_tlb_fill,
.cpu_exec_interrupt = m68k_cpu_exec_interrupt,
.cpu_exec_halt = m68k_cpu_has_work,
+ .cpu_exec_reset = cpu_reset,
.do_interrupt = m68k_cpu_do_interrupt,
.do_transaction_failed = m68k_cpu_transaction_failed,
#endif /* !CONFIG_USER_ONLY */
diff --git a/target/microblaze/cpu.c b/target/microblaze/cpu.c
index 658d3cb3d7..6b14377819 100644
--- a/target/microblaze/cpu.c
+++ b/target/microblaze/cpu.c
@@ -440,6 +440,7 @@ static const TCGCPUOps mb_tcg_ops = {
.tlb_fill = mb_cpu_tlb_fill,
.cpu_exec_interrupt = mb_cpu_exec_interrupt,
.cpu_exec_halt = mb_cpu_has_work,
+ .cpu_exec_reset = cpu_reset,
.do_interrupt = mb_cpu_do_interrupt,
.do_transaction_failed = mb_cpu_transaction_failed,
.do_unaligned_access = mb_cpu_do_unaligned_access,
diff --git a/target/mips/cpu.c b/target/mips/cpu.c
index 26f4f038cd..9f6441d7af 100644
--- a/target/mips/cpu.c
+++ b/target/mips/cpu.c
@@ -563,6 +563,7 @@ static const TCGCPUOps mips_tcg_ops = {
.tlb_fill = mips_cpu_tlb_fill,
.cpu_exec_interrupt = mips_cpu_exec_interrupt,
.cpu_exec_halt = mips_cpu_has_work,
+ .cpu_exec_reset = cpu_reset,
.do_interrupt = mips_cpu_do_interrupt,
.do_transaction_failed = mips_cpu_do_transaction_failed,
.do_unaligned_access = mips_cpu_do_unaligned_access,
diff --git a/target/openrisc/cpu.c b/target/openrisc/cpu.c
index d3c366dd86..ffdc011a44 100644
--- a/target/openrisc/cpu.c
+++ b/target/openrisc/cpu.c
@@ -255,6 +255,7 @@ static const TCGCPUOps openrisc_tcg_ops = {
.tlb_fill = openrisc_cpu_tlb_fill,
.cpu_exec_interrupt = openrisc_cpu_exec_interrupt,
.cpu_exec_halt = openrisc_cpu_has_work,
+ .cpu_exec_reset = cpu_reset,
.do_interrupt = openrisc_cpu_do_interrupt,
#endif /* !CONFIG_USER_ONLY */
};
diff --git a/target/ppc/cpu_init.c b/target/ppc/cpu_init.c
index fde7d71fc6..5981f607d2 100644
--- a/target/ppc/cpu_init.c
+++ b/target/ppc/cpu_init.c
@@ -7492,6 +7492,7 @@ static const TCGCPUOps ppc_tcg_ops = {
.tlb_fill = ppc_cpu_tlb_fill,
.cpu_exec_interrupt = ppc_cpu_exec_interrupt,
.cpu_exec_halt = ppc_cpu_has_work,
+ .cpu_exec_reset = cpu_reset,
.do_interrupt = ppc_cpu_do_interrupt,
.cpu_exec_enter = ppc_cpu_exec_enter,
.cpu_exec_exit = ppc_cpu_exec_exit,
diff --git a/target/riscv/tcg/tcg-cpu.c b/target/riscv/tcg/tcg-cpu.c
index 916fd6fb43..e3e7fea66a 100644
--- a/target/riscv/tcg/tcg-cpu.c
+++ b/target/riscv/tcg/tcg-cpu.c
@@ -153,6 +153,7 @@ const TCGCPUOps riscv_tcg_ops = {
.tlb_fill = riscv_cpu_tlb_fill,
.cpu_exec_interrupt = riscv_cpu_exec_interrupt,
.cpu_exec_halt = riscv_cpu_has_work,
+ .cpu_exec_reset = cpu_reset,
.do_interrupt = riscv_cpu_do_interrupt,
.do_transaction_failed = riscv_cpu_do_transaction_failed,
.do_unaligned_access = riscv_cpu_do_unaligned_access,
diff --git a/target/rx/cpu.c b/target/rx/cpu.c
index 0a7a2b55b5..c9615f0655 100644
--- a/target/rx/cpu.c
+++ b/target/rx/cpu.c
@@ -217,6 +217,7 @@ static const TCGCPUOps rx_tcg_ops = {
.cpu_exec_interrupt = rx_cpu_exec_interrupt,
.cpu_exec_halt = rx_cpu_has_work,
+ .cpu_exec_reset = cpu_reset,
.do_interrupt = rx_cpu_do_interrupt,
};
diff --git a/target/s390x/cpu.c b/target/s390x/cpu.c
index e3623ad32a..f4289f9857 100644
--- a/target/s390x/cpu.c
+++ b/target/s390x/cpu.c
@@ -365,6 +365,7 @@ static const TCGCPUOps s390_tcg_ops = {
.tlb_fill = s390_cpu_tlb_fill,
.cpu_exec_interrupt = s390_cpu_exec_interrupt,
.cpu_exec_halt = s390_cpu_has_work,
+ .cpu_exec_reset = cpu_reset,
.do_interrupt = s390_cpu_do_interrupt,
.debug_excp_handler = s390x_cpu_debug_excp_handler,
.do_unaligned_access = s390x_cpu_do_unaligned_access,
diff --git a/target/sh4/cpu.c b/target/sh4/cpu.c
index 0a04f149d7..2cafb56a23 100644
--- a/target/sh4/cpu.c
+++ b/target/sh4/cpu.c
@@ -275,6 +275,7 @@ static const TCGCPUOps superh_tcg_ops = {
.tlb_fill = superh_cpu_tlb_fill,
.cpu_exec_interrupt = superh_cpu_exec_interrupt,
.cpu_exec_halt = superh_cpu_has_work,
+ .cpu_exec_reset = cpu_reset,
.do_interrupt = superh_cpu_do_interrupt,
.do_unaligned_access = superh_cpu_do_unaligned_access,
.io_recompile_replay_branch = superh_io_recompile_replay_branch,
diff --git a/target/sparc/cpu.c b/target/sparc/cpu.c
index c864217175..5a1f5b7915 100644
--- a/target/sparc/cpu.c
+++ b/target/sparc/cpu.c
@@ -1034,6 +1034,7 @@ static const TCGCPUOps sparc_tcg_ops = {
.tlb_fill = sparc_cpu_tlb_fill,
.cpu_exec_interrupt = sparc_cpu_exec_interrupt,
.cpu_exec_halt = sparc_cpu_has_work,
+ .cpu_exec_reset = cpu_reset,
.do_interrupt = sparc_cpu_do_interrupt,
.do_transaction_failed = sparc_cpu_do_transaction_failed,
.do_unaligned_access = sparc_cpu_do_unaligned_access,
diff --git a/target/tricore/cpu.c b/target/tricore/cpu.c
index bdbcaf0dfc..36c82888e8 100644
--- a/target/tricore/cpu.c
+++ b/target/tricore/cpu.c
@@ -182,6 +182,7 @@ static const TCGCPUOps tricore_tcg_ops = {
.tlb_fill = tricore_cpu_tlb_fill,
.cpu_exec_interrupt = tricore_cpu_exec_interrupt,
.cpu_exec_halt = tricore_cpu_has_work,
+ .cpu_exec_reset = cpu_reset,
};
static void tricore_cpu_class_init(ObjectClass *c, void *data)
diff --git a/target/xtensa/cpu.c b/target/xtensa/cpu.c
index 971e67ad97..a13f1f950a 100644
--- a/target/xtensa/cpu.c
+++ b/target/xtensa/cpu.c
@@ -246,6 +246,7 @@ static const TCGCPUOps xtensa_tcg_ops = {
.tlb_fill = xtensa_cpu_tlb_fill,
.cpu_exec_interrupt = xtensa_cpu_exec_interrupt,
.cpu_exec_halt = xtensa_cpu_has_work,
+ .cpu_exec_reset = cpu_reset,
.do_interrupt = xtensa_cpu_do_interrupt,
.do_transaction_failed = xtensa_cpu_do_transaction_failed,
.do_unaligned_access = xtensa_cpu_do_unaligned_access,
--
2.43.0
^ permalink raw reply related [flat|nested] 32+ messages in thread* Re: [PATCH 03/10] accel/tcg: Introduce TCGCPUOps.cpu_exec_reset
2025-04-28 20:10 ` [PATCH 03/10] accel/tcg: Introduce TCGCPUOps.cpu_exec_reset Richard Henderson
@ 2025-04-28 21:38 ` Pierrick Bouvier
2025-04-29 6:44 ` Philippe Mathieu-Daudé
1 sibling, 0 replies; 32+ messages in thread
From: Pierrick Bouvier @ 2025-04-28 21:38 UTC (permalink / raw)
To: Richard Henderson, qemu-devel; +Cc: philmd
On 4/28/25 1:10 PM, Richard Henderson wrote:
> Initialize all instances with cpu_reset(), so that there
> is no functional change.
>
> Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
> ---
> include/accel/tcg/cpu-ops.h | 2 ++
> accel/tcg/cpu-exec.c | 3 ++-
> target/alpha/cpu.c | 1 +
> target/arm/cpu.c | 1 +
> target/arm/tcg/cpu-v7m.c | 1 +
> target/avr/cpu.c | 1 +
> target/hppa/cpu.c | 1 +
> target/i386/tcg/tcg-cpu.c | 1 +
> target/loongarch/cpu.c | 1 +
> target/m68k/cpu.c | 1 +
> target/microblaze/cpu.c | 1 +
> target/mips/cpu.c | 1 +
> target/openrisc/cpu.c | 1 +
> target/ppc/cpu_init.c | 1 +
> target/riscv/tcg/tcg-cpu.c | 1 +
> target/rx/cpu.c | 1 +
> target/s390x/cpu.c | 1 +
> target/sh4/cpu.c | 1 +
> target/sparc/cpu.c | 1 +
> target/tricore/cpu.c | 1 +
> target/xtensa/cpu.c | 1 +
> 21 files changed, 23 insertions(+), 1 deletion(-)
Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org>
^ permalink raw reply [flat|nested] 32+ messages in thread
* Re: [PATCH 03/10] accel/tcg: Introduce TCGCPUOps.cpu_exec_reset
2025-04-28 20:10 ` [PATCH 03/10] accel/tcg: Introduce TCGCPUOps.cpu_exec_reset Richard Henderson
2025-04-28 21:38 ` Pierrick Bouvier
@ 2025-04-29 6:44 ` Philippe Mathieu-Daudé
2025-04-29 19:53 ` Richard Henderson
1 sibling, 1 reply; 32+ messages in thread
From: Philippe Mathieu-Daudé @ 2025-04-29 6:44 UTC (permalink / raw)
To: Richard Henderson, qemu-devel; +Cc: pierrick.bouvier
On 28/4/25 22:10, Richard Henderson wrote:
> Initialize all instances with cpu_reset(), so that there
> is no functional change.
>
> Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
> ---
> include/accel/tcg/cpu-ops.h | 2 ++
> accel/tcg/cpu-exec.c | 3 ++-
> target/alpha/cpu.c | 1 +
> target/arm/cpu.c | 1 +
> target/arm/tcg/cpu-v7m.c | 1 +
> target/avr/cpu.c | 1 +
> target/hppa/cpu.c | 1 +
> target/i386/tcg/tcg-cpu.c | 1 +
> target/loongarch/cpu.c | 1 +
> target/m68k/cpu.c | 1 +
> target/microblaze/cpu.c | 1 +
> target/mips/cpu.c | 1 +
> target/openrisc/cpu.c | 1 +
> target/ppc/cpu_init.c | 1 +
> target/riscv/tcg/tcg-cpu.c | 1 +
> target/rx/cpu.c | 1 +
> target/s390x/cpu.c | 1 +
> target/sh4/cpu.c | 1 +
> target/sparc/cpu.c | 1 +
> target/tricore/cpu.c | 1 +
> target/xtensa/cpu.c | 1 +
> 21 files changed, 23 insertions(+), 1 deletion(-)
>
> diff --git a/include/accel/tcg/cpu-ops.h b/include/accel/tcg/cpu-ops.h
> index 60b5e97205..3ff72b8d9d 100644
> --- a/include/accel/tcg/cpu-ops.h
> +++ b/include/accel/tcg/cpu-ops.h
> @@ -155,6 +155,8 @@ struct TCGCPUOps {
> void (*do_interrupt)(CPUState *cpu);
> /** @cpu_exec_interrupt: Callback for processing interrupts in cpu_exec */
> bool (*cpu_exec_interrupt)(CPUState *cpu, int interrupt_request);
> + /** @cpu_exec_reset: Callback for reset in cpu_exec. */
> + void (*cpu_exec_reset)(CPUState *cpu);
I'm not sure "cpu_exec" is still relevant these days, maybe we can
directly call it "cpu_reset()". Anyhow,
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
> /**
> * @cpu_exec_halt: Callback for handling halt in cpu_exec.
> *
^ permalink raw reply [flat|nested] 32+ messages in thread* Re: [PATCH 03/10] accel/tcg: Introduce TCGCPUOps.cpu_exec_reset
2025-04-29 6:44 ` Philippe Mathieu-Daudé
@ 2025-04-29 19:53 ` Richard Henderson
0 siblings, 0 replies; 32+ messages in thread
From: Richard Henderson @ 2025-04-29 19:53 UTC (permalink / raw)
To: Philippe Mathieu-Daudé, qemu-devel; +Cc: pierrick.bouvier
On 4/28/25 23:44, Philippe Mathieu-Daudé wrote:
> On 28/4/25 22:10, Richard Henderson wrote:
>> Initialize all instances with cpu_reset(), so that there
>> is no functional change.
>>
>> Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
>> ---
>> include/accel/tcg/cpu-ops.h | 2 ++
>> accel/tcg/cpu-exec.c | 3 ++-
>> target/alpha/cpu.c | 1 +
>> target/arm/cpu.c | 1 +
>> target/arm/tcg/cpu-v7m.c | 1 +
>> target/avr/cpu.c | 1 +
>> target/hppa/cpu.c | 1 +
>> target/i386/tcg/tcg-cpu.c | 1 +
>> target/loongarch/cpu.c | 1 +
>> target/m68k/cpu.c | 1 +
>> target/microblaze/cpu.c | 1 +
>> target/mips/cpu.c | 1 +
>> target/openrisc/cpu.c | 1 +
>> target/ppc/cpu_init.c | 1 +
>> target/riscv/tcg/tcg-cpu.c | 1 +
>> target/rx/cpu.c | 1 +
>> target/s390x/cpu.c | 1 +
>> target/sh4/cpu.c | 1 +
>> target/sparc/cpu.c | 1 +
>> target/tricore/cpu.c | 1 +
>> target/xtensa/cpu.c | 1 +
>> 21 files changed, 23 insertions(+), 1 deletion(-)
>>
>> diff --git a/include/accel/tcg/cpu-ops.h b/include/accel/tcg/cpu-ops.h
>> index 60b5e97205..3ff72b8d9d 100644
>> --- a/include/accel/tcg/cpu-ops.h
>> +++ b/include/accel/tcg/cpu-ops.h
>> @@ -155,6 +155,8 @@ struct TCGCPUOps {
>> void (*do_interrupt)(CPUState *cpu);
>> /** @cpu_exec_interrupt: Callback for processing interrupts in cpu_exec */
>> bool (*cpu_exec_interrupt)(CPUState *cpu, int interrupt_request);
>> + /** @cpu_exec_reset: Callback for reset in cpu_exec. */
>> + void (*cpu_exec_reset)(CPUState *cpu);
>
> I'm not sure "cpu_exec" is still relevant these days, maybe we can
> directly call it "cpu_reset()". Anyhow,
>
> Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
It's called from within the "cpu_exec loop".
r~
^ permalink raw reply [flat|nested] 32+ messages in thread
* [PATCH 04/10] target/i386: Split out x86_cpu_exec_reset
2025-04-28 20:10 [PATCH 00/10] accel/tcg: Compile cpu-exec.c twice Richard Henderson
` (2 preceding siblings ...)
2025-04-28 20:10 ` [PATCH 03/10] accel/tcg: Introduce TCGCPUOps.cpu_exec_reset Richard Henderson
@ 2025-04-28 20:10 ` Richard Henderson
2025-04-28 21:40 ` Pierrick Bouvier
2025-04-28 20:10 ` [PATCH 05/10] accel/tcg: Hoist cpu_get_tb_cpu_state decl to accl/tcg/cpu-ops.h Richard Henderson
` (5 subsequent siblings)
9 siblings, 1 reply; 32+ messages in thread
From: Richard Henderson @ 2025-04-28 20:10 UTC (permalink / raw)
To: qemu-devel; +Cc: pierrick.bouvier, philmd
Note that target/i386/cpu.h defines CPU_INTERRUPT_INIT
as CPU_INTERRUPT_RESET. Therefore we can handle the
#if TARGET_I386 block in cpu_handle_interrupt with the
new TCGCPUOps.cpu_exec_reset hook.
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
---
accel/tcg/cpu-exec.c | 39 ++++++++++++++-------------------------
target/i386/tcg/tcg-cpu.c | 11 ++++++++++-
2 files changed, 24 insertions(+), 26 deletions(-)
diff --git a/accel/tcg/cpu-exec.c b/accel/tcg/cpu-exec.c
index 010f38edaa..c21c5d202d 100644
--- a/accel/tcg/cpu-exec.c
+++ b/accel/tcg/cpu-exec.c
@@ -819,33 +819,22 @@ static inline bool cpu_handle_interrupt(CPUState *cpu,
cpu->exception_index = EXCP_HLT;
bql_unlock();
return true;
- }
-#if defined(TARGET_I386)
- else if (interrupt_request & CPU_INTERRUPT_INIT) {
- X86CPU *x86_cpu = X86_CPU(cpu);
- CPUArchState *env = &x86_cpu->env;
- replay_interrupt();
- cpu_svm_check_intercept_param(env, SVM_EXIT_INIT, 0, 0);
- do_cpu_init(x86_cpu);
- cpu->exception_index = EXCP_HALTED;
- bql_unlock();
- return true;
- }
-#else
- else if (interrupt_request & CPU_INTERRUPT_RESET) {
- replay_interrupt();
- cpu->cc->tcg_ops->cpu_exec_reset(cpu);
- bql_unlock();
- return true;
- }
-#endif /* !TARGET_I386 */
- /* The target hook has 3 exit conditions:
- False when the interrupt isn't processed,
- True when it is, and we should restart on a new TB,
- and via longjmp via cpu_loop_exit. */
- else {
+ } else {
const TCGCPUOps *tcg_ops = cpu->cc->tcg_ops;
+ if (interrupt_request & CPU_INTERRUPT_RESET) {
+ replay_interrupt();
+ tcg_ops->cpu_exec_reset(cpu);
+ bql_unlock();
+ return true;
+ }
+
+ /*
+ * The target hook has 3 exit conditions:
+ * False when the interrupt isn't processed,
+ * True when it is, and we should restart on a new TB,
+ * and via longjmp via cpu_loop_exit.
+ */
if (tcg_ops->cpu_exec_interrupt(cpu, interrupt_request)) {
if (!tcg_ops->need_replay_interrupt ||
tcg_ops->need_replay_interrupt(interrupt_request)) {
diff --git a/target/i386/tcg/tcg-cpu.c b/target/i386/tcg/tcg-cpu.c
index 63ddf87aa8..0184e308e6 100644
--- a/target/i386/tcg/tcg-cpu.c
+++ b/target/i386/tcg/tcg-cpu.c
@@ -120,6 +120,15 @@ static bool x86_debug_check_breakpoint(CPUState *cs)
/* RF disables all architectural breakpoints. */
return !(env->eflags & RF_MASK);
}
+
+static void x86_cpu_exec_reset(CPUState *cs)
+{
+ CPUArchState *env = cpu_env(cs);
+
+ cpu_svm_check_intercept_param(env, SVM_EXIT_INIT, 0, 0);
+ do_cpu_init(env_archcpu(env));
+ cs->exception_index = EXCP_HALTED;
+}
#endif
#include "accel/tcg/cpu-ops.h"
@@ -147,7 +156,7 @@ const TCGCPUOps x86_tcg_ops = {
.do_interrupt = x86_cpu_do_interrupt,
.cpu_exec_halt = x86_cpu_exec_halt,
.cpu_exec_interrupt = x86_cpu_exec_interrupt,
- .cpu_exec_reset = cpu_reset,
+ .cpu_exec_reset = x86_cpu_exec_reset,
.do_unaligned_access = x86_cpu_do_unaligned_access,
.debug_excp_handler = breakpoint_handler,
.debug_check_breakpoint = x86_debug_check_breakpoint,
--
2.43.0
^ permalink raw reply related [flat|nested] 32+ messages in thread* Re: [PATCH 04/10] target/i386: Split out x86_cpu_exec_reset
2025-04-28 20:10 ` [PATCH 04/10] target/i386: Split out x86_cpu_exec_reset Richard Henderson
@ 2025-04-28 21:40 ` Pierrick Bouvier
0 siblings, 0 replies; 32+ messages in thread
From: Pierrick Bouvier @ 2025-04-28 21:40 UTC (permalink / raw)
To: Richard Henderson, qemu-devel; +Cc: philmd
On 4/28/25 1:10 PM, Richard Henderson wrote:
> Note that target/i386/cpu.h defines CPU_INTERRUPT_INIT
> as CPU_INTERRUPT_RESET. Therefore we can handle the
> #if TARGET_I386 block in cpu_handle_interrupt with the
> new TCGCPUOps.cpu_exec_reset hook.
>
> Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
> ---
> accel/tcg/cpu-exec.c | 39 ++++++++++++++-------------------------
> target/i386/tcg/tcg-cpu.c | 11 ++++++++++-
> 2 files changed, 24 insertions(+), 26 deletions(-)
Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org>
^ permalink raw reply [flat|nested] 32+ messages in thread
* [PATCH 05/10] accel/tcg: Hoist cpu_get_tb_cpu_state decl to accl/tcg/cpu-ops.h
2025-04-28 20:10 [PATCH 00/10] accel/tcg: Compile cpu-exec.c twice Richard Henderson
` (3 preceding siblings ...)
2025-04-28 20:10 ` [PATCH 04/10] target/i386: Split out x86_cpu_exec_reset Richard Henderson
@ 2025-04-28 20:10 ` Richard Henderson
2025-04-28 21:43 ` Pierrick Bouvier
2025-04-28 20:10 ` [PATCH 06/10] target/riscv: Move cpu_get_tb_cpu_state to tcg-cpu.c Richard Henderson
` (4 subsequent siblings)
9 siblings, 1 reply; 32+ messages in thread
From: Richard Henderson @ 2025-04-28 20:10 UTC (permalink / raw)
To: qemu-devel; +Cc: pierrick.bouvier, philmd
For some targets, simply remove the local definition.
For other targets, move the inline definition out of line.
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
---
include/accel/tcg/cpu-ops.h | 3 ++
target/alpha/cpu.h | 11 ------
target/arm/cpu.h | 3 --
target/avr/cpu.h | 18 ----------
target/hexagon/cpu.h | 15 --------
target/hppa/cpu.h | 3 --
target/i386/cpu.h | 14 --------
target/loongarch/cpu.h | 12 -------
target/m68k/cpu.h | 16 ---------
target/microblaze/cpu.h | 8 -----
target/mips/cpu.h | 9 -----
target/openrisc/cpu.h | 10 ------
target/ppc/cpu.h | 13 -------
target/riscv/cpu.h | 3 --
target/rx/cpu.h | 9 -----
target/s390x/cpu.h | 9 -----
target/sh4/cpu.h | 15 --------
target/sparc/cpu.h | 3 --
target/tricore/cpu.h | 12 -------
target/xtensa/cpu.h | 68 -----------------------------------
target/alpha/cpu.c | 14 ++++++--
target/arm/helper.c | 1 +
target/avr/cpu.c | 21 +++++++++--
target/hexagon/cpu.c | 18 ++++++++--
target/hppa/cpu.c | 3 +-
target/i386/tcg/tcg-cpu.c | 17 +++++++--
target/loongarch/cpu.c | 15 ++++++--
target/m68k/cpu.c | 19 ++++++++--
target/microblaze/cpu.c | 11 ++++--
target/mips/cpu.c | 9 +++++
target/openrisc/cpu.c | 13 +++++--
target/ppc/helper_regs.c | 16 ++++-----
target/rx/cpu.c | 12 +++++--
target/s390x/cpu.c | 1 +
target/sh4/cpu.c | 18 ++++++++--
target/tricore/cpu.c | 15 ++++++--
target/xtensa/cpu.c | 71 +++++++++++++++++++++++++++++++++++--
37 files changed, 243 insertions(+), 285 deletions(-)
diff --git a/include/accel/tcg/cpu-ops.h b/include/accel/tcg/cpu-ops.h
index 3ff72b8d9d..f5e5746976 100644
--- a/include/accel/tcg/cpu-ops.h
+++ b/include/accel/tcg/cpu-ops.h
@@ -18,6 +18,9 @@
#include "exec/vaddr.h"
#include "tcg/tcg-mo.h"
+void cpu_get_tb_cpu_state(CPUArchState *env, vaddr *pc,
+ uint64_t *cs_base, uint32_t *flags);
+
struct TCGCPUOps {
/**
* mttcg_supported: multi-threaded TCG is supported
diff --git a/target/alpha/cpu.h b/target/alpha/cpu.h
index 849f673489..45944e46b5 100644
--- a/target/alpha/cpu.h
+++ b/target/alpha/cpu.h
@@ -464,17 +464,6 @@ void alpha_cpu_do_transaction_failed(CPUState *cs, hwaddr physaddr,
MemTxResult response, uintptr_t retaddr);
#endif
-static inline void cpu_get_tb_cpu_state(CPUAlphaState *env, vaddr *pc,
- uint64_t *cs_base, uint32_t *pflags)
-{
- *pc = env->pc;
- *cs_base = 0;
- *pflags = env->flags & ENV_FLAG_TB_MASK;
-#ifdef CONFIG_USER_ONLY
- *pflags |= TB_FLAG_UNALIGN * !env_cpu(env)->prctl_unalign_sigbus;
-#endif
-}
-
#ifdef CONFIG_USER_ONLY
/* Copied from linux ieee_swcr_to_fpcr. */
static inline uint64_t alpha_ieee_swcr_to_fpcr(uint64_t swcr)
diff --git a/target/arm/cpu.h b/target/arm/cpu.h
index 3705b34285..8dd7c17533 100644
--- a/target/arm/cpu.h
+++ b/target/arm/cpu.h
@@ -3119,9 +3119,6 @@ static inline bool bswap_code(bool sctlr_b)
#endif
}
-void cpu_get_tb_cpu_state(CPUARMState *env, vaddr *pc,
- uint64_t *cs_base, uint32_t *flags);
-
enum {
QEMU_PSCI_CONDUIT_DISABLED = 0,
QEMU_PSCI_CONDUIT_SMC = 1,
diff --git a/target/avr/cpu.h b/target/avr/cpu.h
index d6666175a9..518e243d81 100644
--- a/target/avr/cpu.h
+++ b/target/avr/cpu.h
@@ -205,24 +205,6 @@ enum {
TB_FLAGS_SKIP = 2,
};
-static inline void cpu_get_tb_cpu_state(CPUAVRState *env, vaddr *pc,
- uint64_t *cs_base, uint32_t *pflags)
-{
- uint32_t flags = 0;
-
- *pc = env->pc_w * 2;
- *cs_base = 0;
-
- if (env->fullacc) {
- flags |= TB_FLAGS_FULL_ACCESS;
- }
- if (env->skip) {
- flags |= TB_FLAGS_SKIP;
- }
-
- *pflags = flags;
-}
-
static inline int cpu_interrupts_enabled(CPUAVRState *env)
{
return env->sregI != 0;
diff --git a/target/hexagon/cpu.h b/target/hexagon/cpu.h
index c065fa8ddc..43a854f517 100644
--- a/target/hexagon/cpu.h
+++ b/target/hexagon/cpu.h
@@ -137,21 +137,6 @@ G_NORETURN void hexagon_raise_exception_err(CPUHexagonState *env,
uint32_t exception,
uintptr_t pc);
-static inline void cpu_get_tb_cpu_state(CPUHexagonState *env, vaddr *pc,
- uint64_t *cs_base, uint32_t *flags)
-{
- uint32_t hex_flags = 0;
- *pc = env->gpr[HEX_REG_PC];
- *cs_base = 0;
- if (*pc == env->gpr[HEX_REG_SA0]) {
- hex_flags = FIELD_DP32(hex_flags, TB_FLAGS, IS_TIGHT_LOOP, 1);
- }
- *flags = hex_flags;
- if (*pc & PCALIGN_MASK) {
- hexagon_raise_exception_err(env, HEX_CAUSE_PC_NOT_ALIGNED, 0);
- }
-}
-
typedef HexagonCPU ArchCPU;
void hexagon_translate_init(void);
diff --git a/target/hppa/cpu.h b/target/hppa/cpu.h
index acc9937240..11d59d11ca 100644
--- a/target/hppa/cpu.h
+++ b/target/hppa/cpu.h
@@ -351,9 +351,6 @@ hwaddr hppa_abs_to_phys_pa2_w1(vaddr addr);
#define CS_BASE_DIFFPAGE (1 << 12)
#define CS_BASE_DIFFSPACE (1 << 13)
-void cpu_get_tb_cpu_state(CPUHPPAState *env, vaddr *pc,
- uint64_t *cs_base, uint32_t *pflags);
-
target_ulong cpu_hppa_get_psw(CPUHPPAState *env);
void cpu_hppa_put_psw(CPUHPPAState *env, target_ulong);
void update_gva_offset_mask(CPUHPPAState *env);
diff --git a/target/i386/cpu.h b/target/i386/cpu.h
index 877df7de1f..a7133c65ab 100644
--- a/target/i386/cpu.h
+++ b/target/i386/cpu.h
@@ -2602,20 +2602,6 @@ static inline bool is_mmu_index_32(int mmu_index)
#include "hw/i386/apic.h"
#endif
-static inline void cpu_get_tb_cpu_state(CPUX86State *env, vaddr *pc,
- uint64_t *cs_base, uint32_t *flags)
-{
- *flags = env->hflags |
- (env->eflags & (IOPL_MASK | TF_MASK | RF_MASK | VM_MASK | AC_MASK));
- if (env->hflags & HF_CS64_MASK) {
- *cs_base = 0;
- *pc = env->eip;
- } else {
- *cs_base = env->segs[R_CS].base;
- *pc = (uint32_t)(*cs_base + env->eip);
- }
-}
-
void do_cpu_init(X86CPU *cpu);
#define MCE_INJECT_BROADCAST 1
diff --git a/target/loongarch/cpu.h b/target/loongarch/cpu.h
index 70ff56e60c..262bf87f7b 100644
--- a/target/loongarch/cpu.h
+++ b/target/loongarch/cpu.h
@@ -492,18 +492,6 @@ static inline void set_pc(CPULoongArchState *env, uint64_t value)
#define HW_FLAGS_VA32 0x20
#define HW_FLAGS_EUEN_ASXE 0x40
-static inline void cpu_get_tb_cpu_state(CPULoongArchState *env, vaddr *pc,
- uint64_t *cs_base, uint32_t *flags)
-{
- *pc = env->pc;
- *cs_base = 0;
- *flags = env->CSR_CRMD & (R_CSR_CRMD_PLV_MASK | R_CSR_CRMD_PG_MASK);
- *flags |= FIELD_EX64(env->CSR_EUEN, CSR_EUEN, FPE) * HW_FLAGS_EUEN_FPE;
- *flags |= FIELD_EX64(env->CSR_EUEN, CSR_EUEN, SXE) * HW_FLAGS_EUEN_SXE;
- *flags |= FIELD_EX64(env->CSR_EUEN, CSR_EUEN, ASXE) * HW_FLAGS_EUEN_ASXE;
- *flags |= is_va32(env) * HW_FLAGS_VA32;
-}
-
#define CPU_RESOLVING_TYPE TYPE_LOONGARCH_CPU
void loongarch_cpu_post_init(Object *obj);
diff --git a/target/m68k/cpu.h b/target/m68k/cpu.h
index 39d0b9d6d7..d9db6a486a 100644
--- a/target/m68k/cpu.h
+++ b/target/m68k/cpu.h
@@ -605,22 +605,6 @@ void m68k_cpu_transaction_failed(CPUState *cs, hwaddr physaddr, vaddr addr,
#define TB_FLAGS_TRACE 16
#define TB_FLAGS_TRACE_BIT (1 << TB_FLAGS_TRACE)
-static inline void cpu_get_tb_cpu_state(CPUM68KState *env, vaddr *pc,
- uint64_t *cs_base, uint32_t *flags)
-{
- *pc = env->pc;
- *cs_base = 0;
- *flags = (env->macsr >> 4) & TB_FLAGS_MACSR;
- if (env->sr & SR_S) {
- *flags |= TB_FLAGS_MSR_S;
- *flags |= (env->sfc << (TB_FLAGS_SFC_S_BIT - 2)) & TB_FLAGS_SFC_S;
- *flags |= (env->dfc << (TB_FLAGS_DFC_S_BIT - 2)) & TB_FLAGS_DFC_S;
- }
- if (M68K_SR_TRACE(env->sr) == M68K_SR_TRACE_ANY_INS) {
- *flags |= TB_FLAGS_TRACE;
- }
-}
-
void dump_mmu(CPUM68KState *env);
#endif
diff --git a/target/microblaze/cpu.h b/target/microblaze/cpu.h
index d511f22a55..6ad8643f2e 100644
--- a/target/microblaze/cpu.h
+++ b/target/microblaze/cpu.h
@@ -419,14 +419,6 @@ static inline bool mb_cpu_is_big_endian(CPUState *cs)
return !cpu->cfg.endi;
}
-static inline void cpu_get_tb_cpu_state(CPUMBState *env, vaddr *pc,
- uint64_t *cs_base, uint32_t *flags)
-{
- *pc = env->pc;
- *flags = (env->iflags & IFLAGS_TB_MASK) | (env->msr & MSR_TB_MASK);
- *cs_base = (*flags & IMM_FLAG ? env->imm : 0);
-}
-
#if !defined(CONFIG_USER_ONLY)
bool mb_cpu_tlb_fill(CPUState *cs, vaddr address, int size,
MMUAccessType access_type, int mmu_idx,
diff --git a/target/mips/cpu.h b/target/mips/cpu.h
index d16f9a7220..5cd4c6c818 100644
--- a/target/mips/cpu.h
+++ b/target/mips/cpu.h
@@ -1366,15 +1366,6 @@ void cpu_mips_clock_init(MIPSCPU *cpu);
/* helper.c */
target_ulong exception_resume_pc(CPUMIPSState *env);
-static inline void cpu_get_tb_cpu_state(CPUMIPSState *env, vaddr *pc,
- uint64_t *cs_base, uint32_t *flags)
-{
- *pc = env->active_tc.PC;
- *cs_base = 0;
- *flags = env->hflags & (MIPS_HFLAG_TMASK | MIPS_HFLAG_BMASK |
- MIPS_HFLAG_HWRENA_ULR);
-}
-
/**
* mips_cpu_create_with_clock:
* @typename: a MIPS CPU type.
diff --git a/target/openrisc/cpu.h b/target/openrisc/cpu.h
index 569819bfb0..f4bcf00b07 100644
--- a/target/openrisc/cpu.h
+++ b/target/openrisc/cpu.h
@@ -349,16 +349,6 @@ static inline void cpu_set_gpr(CPUOpenRISCState *env, int i, uint32_t val)
env->shadow_gpr[0][i] = val;
}
-static inline void cpu_get_tb_cpu_state(CPUOpenRISCState *env, vaddr *pc,
- uint64_t *cs_base, uint32_t *flags)
-{
- *pc = env->pc;
- *cs_base = 0;
- *flags = (env->dflag ? TB_FLAGS_DFLAG : 0)
- | (cpu_get_gpr(env, 0) ? 0 : TB_FLAGS_R0_0)
- | (env->sr & (SR_SM | SR_DME | SR_IME | SR_OVE));
-}
-
static inline uint32_t cpu_get_sr(const CPUOpenRISCState *env)
{
return (env->sr
diff --git a/target/ppc/cpu.h b/target/ppc/cpu.h
index 3c02f7f7d4..13f4b7f8e0 100644
--- a/target/ppc/cpu.h
+++ b/target/ppc/cpu.h
@@ -2755,19 +2755,6 @@ void cpu_write_xer(CPUPPCState *env, target_ulong xer);
*/
#define is_book3s_arch2x(ctx) (!!((ctx)->insns_flags & PPC_SEGMENT_64B))
-#ifdef CONFIG_DEBUG_TCG
-void cpu_get_tb_cpu_state(CPUPPCState *env, vaddr *pc,
- uint64_t *cs_base, uint32_t *flags);
-#else
-static inline void cpu_get_tb_cpu_state(CPUPPCState *env, vaddr *pc,
- uint64_t *cs_base, uint32_t *flags)
-{
- *pc = env->nip;
- *cs_base = 0;
- *flags = env->hflags;
-}
-#endif
-
G_NORETURN void raise_exception_err_ra(CPUPPCState *env, uint32_t exception,
uint32_t error_code, uintptr_t raddr);
diff --git a/target/riscv/cpu.h b/target/riscv/cpu.h
index 167909c89b..c66ac3bc27 100644
--- a/target/riscv/cpu.h
+++ b/target/riscv/cpu.h
@@ -802,9 +802,6 @@ static inline uint32_t vext_get_vlmax(uint32_t vlenb, uint32_t vsew,
return vlen >> (vsew + 3 - lmul);
}
-void cpu_get_tb_cpu_state(CPURISCVState *env, vaddr *pc,
- uint64_t *cs_base, uint32_t *pflags);
-
bool riscv_cpu_is_32bit(RISCVCPU *cpu);
bool riscv_cpu_virt_mem_enabled(CPURISCVState *env);
diff --git a/target/rx/cpu.h b/target/rx/cpu.h
index 5c19c83219..ba5761b647 100644
--- a/target/rx/cpu.h
+++ b/target/rx/cpu.h
@@ -153,15 +153,6 @@ void rx_cpu_unpack_psw(CPURXState *env, uint32_t psw, int rte);
#define RX_CPU_IRQ 0
#define RX_CPU_FIR 1
-static inline void cpu_get_tb_cpu_state(CPURXState *env, vaddr *pc,
- uint64_t *cs_base, uint32_t *flags)
-{
- *pc = env->pc;
- *cs_base = 0;
- *flags = FIELD_DP32(0, PSW, PM, env->psw_pm);
- *flags = FIELD_DP32(*flags, PSW, U, env->psw_u);
-}
-
static inline uint32_t rx_cpu_pack_psw(CPURXState *env)
{
uint32_t psw = 0;
diff --git a/target/s390x/cpu.h b/target/s390x/cpu.h
index ee59039879..f97d024b74 100644
--- a/target/s390x/cpu.h
+++ b/target/s390x/cpu.h
@@ -411,15 +411,6 @@ static inline int s390x_env_mmu_index(CPUS390XState *env, bool ifetch)
#endif
}
-#ifdef CONFIG_TCG
-
-#include "tcg/tcg_s390x.h"
-
-void cpu_get_tb_cpu_state(CPUS390XState *env, vaddr *pc,
- uint64_t *cs_base, uint32_t *flags);
-
-#endif /* CONFIG_TCG */
-
/* PER bits from control register 9 */
#define PER_CR9_EVENT_BRANCH 0x80000000
#define PER_CR9_EVENT_IFETCH 0x40000000
diff --git a/target/sh4/cpu.h b/target/sh4/cpu.h
index 906f99ddf0..c41ab70dd7 100644
--- a/target/sh4/cpu.h
+++ b/target/sh4/cpu.h
@@ -380,19 +380,4 @@ static inline void cpu_write_sr(CPUSH4State *env, target_ulong sr)
env->sr = sr & ~((1u << SR_M) | (1u << SR_Q) | (1u << SR_T));
}
-static inline void cpu_get_tb_cpu_state(CPUSH4State *env, vaddr *pc,
- uint64_t *cs_base, uint32_t *flags)
-{
- *pc = env->pc;
- /* For a gUSA region, notice the end of the region. */
- *cs_base = env->flags & TB_FLAG_GUSA_MASK ? env->gregs[0] : 0;
- *flags = env->flags
- | (env->fpscr & TB_FLAG_FPSCR_MASK)
- | (env->sr & TB_FLAG_SR_MASK)
- | (env->movcal_backup ? TB_FLAG_PENDING_MOVCA : 0); /* Bit 3 */
-#ifdef CONFIG_USER_ONLY
- *flags |= TB_FLAG_UNALIGN * !env_cpu(env)->prctl_unalign_sigbus;
-#endif
-}
-
#endif /* SH4_CPU_H */
diff --git a/target/sparc/cpu.h b/target/sparc/cpu.h
index 83ac818933..5958f5671b 100644
--- a/target/sparc/cpu.h
+++ b/target/sparc/cpu.h
@@ -744,9 +744,6 @@ trap_state* cpu_tsptr(CPUSPARCState* env);
#define TB_FLAG_FSR_QNE (1 << 8)
#define TB_FLAG_ASI_SHIFT 24
-void cpu_get_tb_cpu_state(CPUSPARCState *env, vaddr *pc,
- uint64_t *cs_base, uint32_t *pflags);
-
static inline bool tb_fpu_enabled(int tb_flags)
{
#if defined(CONFIG_USER_ONLY)
diff --git a/target/tricore/cpu.h b/target/tricore/cpu.h
index c76e65f818..82085fbc32 100644
--- a/target/tricore/cpu.h
+++ b/target/tricore/cpu.h
@@ -258,18 +258,6 @@ void tricore_tcg_init(void);
void tricore_translate_code(CPUState *cs, TranslationBlock *tb,
int *max_insns, vaddr pc, void *host_pc);
-static inline void cpu_get_tb_cpu_state(CPUTriCoreState *env, vaddr *pc,
- uint64_t *cs_base, uint32_t *flags)
-{
- uint32_t new_flags = 0;
- *pc = env->PC;
- *cs_base = 0;
-
- new_flags |= FIELD_DP32(new_flags, TB_FLAGS, PRIV,
- extract32(env->PSW, 10, 2));
- *flags = new_flags;
-}
-
#define CPU_RESOLVING_TYPE TYPE_TRICORE_CPU
/* helpers.c */
diff --git a/target/xtensa/cpu.h b/target/xtensa/cpu.h
index c03ed71c94..74122ebe15 100644
--- a/target/xtensa/cpu.h
+++ b/target/xtensa/cpu.h
@@ -733,74 +733,6 @@ static inline uint32_t xtensa_replicate_windowstart(CPUXtensaState *env)
#define XTENSA_CSBASE_LBEG_OFF_MASK 0x00ff0000
#define XTENSA_CSBASE_LBEG_OFF_SHIFT 16
-static inline void cpu_get_tb_cpu_state(CPUXtensaState *env, vaddr *pc,
- uint64_t *cs_base, uint32_t *flags)
-{
- *pc = env->pc;
- *cs_base = 0;
- *flags = 0;
- *flags |= xtensa_get_ring(env);
- if (env->sregs[PS] & PS_EXCM) {
- *flags |= XTENSA_TBFLAG_EXCM;
- } else if (xtensa_option_enabled(env->config, XTENSA_OPTION_LOOP)) {
- target_ulong lend_dist =
- env->sregs[LEND] - (env->pc & -(1u << TARGET_PAGE_BITS));
-
- /*
- * 0 in the csbase_lend field means that there may not be a loopback
- * for any instruction that starts inside this page. Any other value
- * means that an instruction that ends at this offset from the page
- * start may loop back and will need loopback code to be generated.
- *
- * lend_dist is 0 when LEND points to the start of the page, but
- * no instruction that starts inside this page may end at offset 0,
- * so it's still correct.
- *
- * When an instruction ends at a page boundary it may only start in
- * the previous page. lend_dist will be encoded as TARGET_PAGE_SIZE
- * for the TB that contains this instruction.
- */
- if (lend_dist < (1u << TARGET_PAGE_BITS) + env->config->max_insn_size) {
- target_ulong lbeg_off = env->sregs[LEND] - env->sregs[LBEG];
-
- *cs_base = lend_dist;
- if (lbeg_off < 256) {
- *cs_base |= lbeg_off << XTENSA_CSBASE_LBEG_OFF_SHIFT;
- }
- }
- }
- if (xtensa_option_enabled(env->config, XTENSA_OPTION_EXTENDED_L32R) &&
- (env->sregs[LITBASE] & 1)) {
- *flags |= XTENSA_TBFLAG_LITBASE;
- }
- if (xtensa_option_enabled(env->config, XTENSA_OPTION_DEBUG)) {
- if (xtensa_get_cintlevel(env) < env->config->debug_level) {
- *flags |= XTENSA_TBFLAG_DEBUG;
- }
- if (xtensa_get_cintlevel(env) < env->sregs[ICOUNTLEVEL]) {
- *flags |= XTENSA_TBFLAG_ICOUNT;
- }
- }
- if (xtensa_option_enabled(env->config, XTENSA_OPTION_COPROCESSOR)) {
- *flags |= env->sregs[CPENABLE] << XTENSA_TBFLAG_CPENABLE_SHIFT;
- }
- if (xtensa_option_enabled(env->config, XTENSA_OPTION_WINDOWED_REGISTER) &&
- (env->sregs[PS] & (PS_WOE | PS_EXCM)) == PS_WOE) {
- uint32_t windowstart = xtensa_replicate_windowstart(env) >>
- (env->sregs[WINDOW_BASE] + 1);
- uint32_t w = ctz32(windowstart | 0x8);
-
- *flags |= (w << XTENSA_TBFLAG_WINDOW_SHIFT) | XTENSA_TBFLAG_CWOE;
- *flags |= extract32(env->sregs[PS], PS_CALLINC_SHIFT,
- PS_CALLINC_LEN) << XTENSA_TBFLAG_CALLINC_SHIFT;
- } else {
- *flags |= 3 << XTENSA_TBFLAG_WINDOW_SHIFT;
- }
- if (env->yield_needed) {
- *flags |= XTENSA_TBFLAG_YIELD;
- }
-}
-
XtensaCPU *xtensa_cpu_create_with_clock(const char *cpu_type,
Clock *cpu_refclk);
diff --git a/target/alpha/cpu.c b/target/alpha/cpu.c
index 9dc1c7242b..d1fddef1f4 100644
--- a/target/alpha/cpu.c
+++ b/target/alpha/cpu.c
@@ -25,6 +25,7 @@
#include "cpu.h"
#include "exec/translation-block.h"
#include "exec/target_page.h"
+#include "accel/tcg/cpu-ops.h"
#include "fpu/softfloat.h"
@@ -40,6 +41,17 @@ static vaddr alpha_cpu_get_pc(CPUState *cs)
return env->pc;
}
+void cpu_get_tb_cpu_state(CPUAlphaState *env, vaddr *pc,
+ uint64_t *cs_base, uint32_t *pflags)
+{
+ *pc = env->pc;
+ *cs_base = 0;
+ *pflags = env->flags & ENV_FLAG_TB_MASK;
+#ifdef CONFIG_USER_ONLY
+ *pflags |= TB_FLAG_UNALIGN * !env_cpu(env)->prctl_unalign_sigbus;
+#endif
+}
+
static void alpha_cpu_synchronize_from_tb(CPUState *cs,
const TranslationBlock *tb)
{
@@ -231,8 +243,6 @@ static const struct SysemuCPUOps alpha_sysemu_ops = {
};
#endif
-#include "accel/tcg/cpu-ops.h"
-
static const TCGCPUOps alpha_tcg_ops = {
/* Alpha processors have a weak memory model */
.guest_default_memory_order = 0,
diff --git a/target/arm/helper.c b/target/arm/helper.c
index 8de4eb2c1f..98adeb7086 100644
--- a/target/arm/helper.c
+++ b/target/arm/helper.c
@@ -30,6 +30,7 @@
#include "qemu/guest-random.h"
#ifdef CONFIG_TCG
#include "accel/tcg/probe.h"
+#include "accel/tcg/cpu-ops.h"
#include "semihosting/common-semi.h"
#endif
#include "cpregs.h"
diff --git a/target/avr/cpu.c b/target/avr/cpu.c
index e50ebfd8f8..fac8954c39 100644
--- a/target/avr/cpu.c
+++ b/target/avr/cpu.c
@@ -27,6 +27,7 @@
#include "disas/dis-asm.h"
#include "tcg/debug-assert.h"
#include "hw/qdev-properties.h"
+#include "accel/tcg/cpu-ops.h"
static void avr_cpu_set_pc(CPUState *cs, vaddr value)
{
@@ -53,6 +54,24 @@ static int avr_cpu_mmu_index(CPUState *cs, bool ifetch)
return ifetch ? MMU_CODE_IDX : MMU_DATA_IDX;
}
+void cpu_get_tb_cpu_state(CPUAVRState *env, vaddr *pc,
+ uint64_t *cs_base, uint32_t *pflags)
+{
+ uint32_t flags = 0;
+
+ *pc = env->pc_w * 2;
+ *cs_base = 0;
+
+ if (env->fullacc) {
+ flags |= TB_FLAGS_FULL_ACCESS;
+ }
+ if (env->skip) {
+ flags |= TB_FLAGS_SKIP;
+ }
+
+ *pflags = flags;
+}
+
static void avr_cpu_synchronize_from_tb(CPUState *cs,
const TranslationBlock *tb)
{
@@ -220,8 +239,6 @@ static const struct SysemuCPUOps avr_sysemu_ops = {
.get_phys_page_debug = avr_cpu_get_phys_page_debug,
};
-#include "accel/tcg/cpu-ops.h"
-
static const TCGCPUOps avr_tcg_ops = {
.guest_default_memory_order = 0,
.mttcg_supported = false,
diff --git a/target/hexagon/cpu.c b/target/hexagon/cpu.c
index 9f93c17009..e1a93ce24f 100644
--- a/target/hexagon/cpu.c
+++ b/target/hexagon/cpu.c
@@ -25,6 +25,7 @@
#include "fpu/softfloat-helpers.h"
#include "tcg/tcg.h"
#include "exec/gdbstub.h"
+#include "accel/tcg/cpu-ops.h"
static void hexagon_v66_cpu_init(Object *obj) { }
static void hexagon_v67_cpu_init(Object *obj) { }
@@ -254,6 +255,21 @@ static vaddr hexagon_cpu_get_pc(CPUState *cs)
return cpu_env(cs)->gpr[HEX_REG_PC];
}
+void cpu_get_tb_cpu_state(CPUHexagonState *env, vaddr *pc,
+ uint64_t *cs_base, uint32_t *flags)
+{
+ uint32_t hex_flags = 0;
+ *pc = env->gpr[HEX_REG_PC];
+ *cs_base = 0;
+ if (*pc == env->gpr[HEX_REG_SA0]) {
+ hex_flags = FIELD_DP32(hex_flags, TB_FLAGS, IS_TIGHT_LOOP, 1);
+ }
+ *flags = hex_flags;
+ if (*pc & PCALIGN_MASK) {
+ hexagon_raise_exception_err(env, HEX_CAUSE_PC_NOT_ALIGNED, 0);
+ }
+}
+
static void hexagon_cpu_synchronize_from_tb(CPUState *cs,
const TranslationBlock *tb)
{
@@ -321,8 +337,6 @@ static void hexagon_cpu_init(Object *obj)
{
}
-#include "accel/tcg/cpu-ops.h"
-
static const TCGCPUOps hexagon_tcg_ops = {
/* MTTCG not yet supported: require strict ordering */
.guest_default_memory_order = TCG_MO_ALL,
diff --git a/target/hppa/cpu.c b/target/hppa/cpu.c
index 300c6b809f..24ca2d7175 100644
--- a/target/hppa/cpu.c
+++ b/target/hppa/cpu.c
@@ -29,6 +29,7 @@
#include "fpu/softfloat.h"
#include "tcg/tcg.h"
#include "hw/hppa/hppa_hardware.h"
+#include "accel/tcg/cpu-ops.h"
static void hppa_cpu_set_pc(CPUState *cs, vaddr value)
{
@@ -249,8 +250,6 @@ static const struct SysemuCPUOps hppa_sysemu_ops = {
};
#endif
-#include "accel/tcg/cpu-ops.h"
-
static const TCGCPUOps hppa_tcg_ops = {
/* PA-RISC 1.x processors have a strong memory model. */
/*
diff --git a/target/i386/tcg/tcg-cpu.c b/target/i386/tcg/tcg-cpu.c
index 0184e308e6..cf9ce70139 100644
--- a/target/i386/tcg/tcg-cpu.c
+++ b/target/i386/tcg/tcg-cpu.c
@@ -24,6 +24,7 @@
#include "accel/accel-cpu-target.h"
#include "exec/translation-block.h"
#include "exec/target_page.h"
+#include "accel/tcg/cpu-ops.h"
#include "tcg-cpu.h"
/* Frob eflags into and out of the CPU temporary format. */
@@ -47,6 +48,20 @@ static void x86_cpu_exec_exit(CPUState *cs)
env->eflags = cpu_compute_eflags(env);
}
+void cpu_get_tb_cpu_state(CPUX86State *env, vaddr *pc,
+ uint64_t *cs_base, uint32_t *flags)
+{
+ *flags = env->hflags |
+ (env->eflags & (IOPL_MASK | TF_MASK | RF_MASK | VM_MASK | AC_MASK));
+ if (env->hflags & HF_CS64_MASK) {
+ *cs_base = 0;
+ *pc = env->eip;
+ } else {
+ *cs_base = env->segs[R_CS].base;
+ *pc = (uint32_t)(*cs_base + env->eip);
+ }
+}
+
static void x86_cpu_synchronize_from_tb(CPUState *cs,
const TranslationBlock *tb)
{
@@ -131,8 +146,6 @@ static void x86_cpu_exec_reset(CPUState *cs)
}
#endif
-#include "accel/tcg/cpu-ops.h"
-
const TCGCPUOps x86_tcg_ops = {
.mttcg_supported = true,
.precise_smc = true,
diff --git a/target/loongarch/cpu.c b/target/loongarch/cpu.c
index 3cb7aa6039..db66a6bdeb 100644
--- a/target/loongarch/cpu.c
+++ b/target/loongarch/cpu.c
@@ -29,6 +29,7 @@
#endif
#ifdef CONFIG_TCG
#include "accel/tcg/cpu-ldst.h"
+#include "accel/tcg/cpu-ops.h"
#include "tcg/tcg.h"
#endif
#include "tcg/tcg_loongarch.h"
@@ -335,6 +336,18 @@ static bool loongarch_cpu_exec_interrupt(CPUState *cs, int interrupt_request)
}
#endif
+void cpu_get_tb_cpu_state(CPULoongArchState *env, vaddr *pc,
+ uint64_t *cs_base, uint32_t *flags)
+{
+ *pc = env->pc;
+ *cs_base = 0;
+ *flags = env->CSR_CRMD & (R_CSR_CRMD_PLV_MASK | R_CSR_CRMD_PG_MASK);
+ *flags |= FIELD_EX64(env->CSR_EUEN, CSR_EUEN, FPE) * HW_FLAGS_EUEN_FPE;
+ *flags |= FIELD_EX64(env->CSR_EUEN, CSR_EUEN, SXE) * HW_FLAGS_EUEN_SXE;
+ *flags |= FIELD_EX64(env->CSR_EUEN, CSR_EUEN, ASXE) * HW_FLAGS_EUEN_ASXE;
+ *flags |= is_va32(env) * HW_FLAGS_VA32;
+}
+
static void loongarch_cpu_synchronize_from_tb(CPUState *cs,
const TranslationBlock *tb)
{
@@ -861,8 +874,6 @@ static void loongarch_cpu_dump_state(CPUState *cs, FILE *f, int flags)
}
#ifdef CONFIG_TCG
-#include "accel/tcg/cpu-ops.h"
-
static const TCGCPUOps loongarch_tcg_ops = {
.guest_default_memory_order = 0,
.mttcg_supported = true,
diff --git a/target/m68k/cpu.c b/target/m68k/cpu.c
index c8c1dbeac9..0fc3ed316d 100644
--- a/target/m68k/cpu.c
+++ b/target/m68k/cpu.c
@@ -23,6 +23,7 @@
#include "cpu.h"
#include "migration/vmstate.h"
#include "fpu/softfloat.h"
+#include "accel/tcg/cpu-ops.h"
static void m68k_cpu_set_pc(CPUState *cs, vaddr value)
{
@@ -38,6 +39,22 @@ static vaddr m68k_cpu_get_pc(CPUState *cs)
return cpu->env.pc;
}
+void cpu_get_tb_cpu_state(CPUM68KState *env, vaddr *pc,
+ uint64_t *cs_base, uint32_t *flags)
+{
+ *pc = env->pc;
+ *cs_base = 0;
+ *flags = (env->macsr >> 4) & TB_FLAGS_MACSR;
+ if (env->sr & SR_S) {
+ *flags |= TB_FLAGS_MSR_S;
+ *flags |= (env->sfc << (TB_FLAGS_SFC_S_BIT - 2)) & TB_FLAGS_SFC_S;
+ *flags |= (env->dfc << (TB_FLAGS_DFC_S_BIT - 2)) & TB_FLAGS_DFC_S;
+ }
+ if (M68K_SR_TRACE(env->sr) == M68K_SR_TRACE_ANY_INS) {
+ *flags |= TB_FLAGS_TRACE;
+ }
+}
+
static void m68k_restore_state_to_opc(CPUState *cs,
const TranslationBlock *tb,
const uint64_t *data)
@@ -586,8 +603,6 @@ static const struct SysemuCPUOps m68k_sysemu_ops = {
};
#endif /* !CONFIG_USER_ONLY */
-#include "accel/tcg/cpu-ops.h"
-
static const TCGCPUOps m68k_tcg_ops = {
/* MTTCG not yet supported: require strict ordering */
.guest_default_memory_order = TCG_MO_ALL,
diff --git a/target/microblaze/cpu.c b/target/microblaze/cpu.c
index 6b14377819..a91844e4bf 100644
--- a/target/microblaze/cpu.c
+++ b/target/microblaze/cpu.c
@@ -31,6 +31,7 @@
#include "exec/gdbstub.h"
#include "exec/translation-block.h"
#include "fpu/softfloat-helpers.h"
+#include "accel/tcg/cpu-ops.h"
#include "tcg/tcg.h"
static const struct {
@@ -94,6 +95,14 @@ static vaddr mb_cpu_get_pc(CPUState *cs)
return cpu->env.pc;
}
+void cpu_get_tb_cpu_state(CPUMBState *env, vaddr *pc,
+ uint64_t *cs_base, uint32_t *flags)
+{
+ *pc = env->pc;
+ *flags = (env->iflags & IFLAGS_TB_MASK) | (env->msr & MSR_TB_MASK);
+ *cs_base = (*flags & IMM_FLAG ? env->imm : 0);
+}
+
static void mb_cpu_synchronize_from_tb(CPUState *cs,
const TranslationBlock *tb)
{
@@ -423,8 +432,6 @@ static const struct SysemuCPUOps mb_sysemu_ops = {
};
#endif
-#include "accel/tcg/cpu-ops.h"
-
static const TCGCPUOps mb_tcg_ops = {
/* MicroBlaze is always in-order. */
.guest_default_memory_order = TCG_MO_ALL,
diff --git a/target/mips/cpu.c b/target/mips/cpu.c
index 9f6441d7af..066dadc71b 100644
--- a/target/mips/cpu.c
+++ b/target/mips/cpu.c
@@ -549,6 +549,15 @@ static int mips_cpu_mmu_index(CPUState *cs, bool ifunc)
return mips_env_mmu_index(cpu_env(cs));
}
+void cpu_get_tb_cpu_state(CPUMIPSState *env, vaddr *pc,
+ uint64_t *cs_base, uint32_t *flags)
+{
+ *pc = env->active_tc.PC;
+ *cs_base = 0;
+ *flags = env->hflags & (MIPS_HFLAG_TMASK | MIPS_HFLAG_BMASK |
+ MIPS_HFLAG_HWRENA_ULR);
+}
+
static const TCGCPUOps mips_tcg_ops = {
.mttcg_supported = TARGET_LONG_BITS == 32,
.guest_default_memory_order = 0,
diff --git a/target/openrisc/cpu.c b/target/openrisc/cpu.c
index ffdc011a44..cfb3f62663 100644
--- a/target/openrisc/cpu.c
+++ b/target/openrisc/cpu.c
@@ -23,6 +23,7 @@
#include "cpu.h"
#include "exec/translation-block.h"
#include "fpu/softfloat-helpers.h"
+#include "accel/tcg/cpu-ops.h"
#include "tcg/tcg.h"
static void openrisc_cpu_set_pc(CPUState *cs, vaddr value)
@@ -40,6 +41,16 @@ static vaddr openrisc_cpu_get_pc(CPUState *cs)
return cpu->env.pc;
}
+void cpu_get_tb_cpu_state(CPUOpenRISCState *env, vaddr *pc,
+ uint64_t *cs_base, uint32_t *flags)
+{
+ *pc = env->pc;
+ *cs_base = 0;
+ *flags = (env->dflag ? TB_FLAGS_DFLAG : 0)
+ | (cpu_get_gpr(env, 0) ? 0 : TB_FLAGS_R0_0)
+ | (env->sr & (SR_SM | SR_DME | SR_IME | SR_OVE));
+}
+
static void openrisc_cpu_synchronize_from_tb(CPUState *cs,
const TranslationBlock *tb)
{
@@ -239,8 +250,6 @@ static const struct SysemuCPUOps openrisc_sysemu_ops = {
};
#endif
-#include "accel/tcg/cpu-ops.h"
-
static const TCGCPUOps openrisc_tcg_ops = {
.guest_default_memory_order = 0,
.mttcg_supported = true,
diff --git a/target/ppc/helper_regs.c b/target/ppc/helper_regs.c
index f211bc9830..8d248bcbb9 100644
--- a/target/ppc/helper_regs.c
+++ b/target/ppc/helper_regs.c
@@ -27,6 +27,7 @@
#include "power8-pmu.h"
#include "cpu-models.h"
#include "spr_common.h"
+#include "accel/tcg/cpu-ops.h"
/* Swap temporary saved registers with GPRs */
void hreg_swap_gpr_tgpr(CPUPPCState *env)
@@ -255,26 +256,25 @@ void hreg_update_pmu_hflags(CPUPPCState *env)
env->hflags |= hreg_compute_pmu_hflags_value(env);
}
-#ifdef CONFIG_DEBUG_TCG
void cpu_get_tb_cpu_state(CPUPPCState *env, vaddr *pc,
uint64_t *cs_base, uint32_t *flags)
{
uint32_t hflags_current = env->hflags;
- uint32_t hflags_rebuilt;
- *pc = env->nip;
- *cs_base = 0;
- *flags = hflags_current;
-
- hflags_rebuilt = hreg_compute_hflags_value(env);
+#ifdef CONFIG_DEBUG_TCG
+ uint32_t hflags_rebuilt = hreg_compute_hflags_value(env);
if (unlikely(hflags_current != hflags_rebuilt)) {
cpu_abort(env_cpu(env),
"TCG hflags mismatch (current:0x%08x rebuilt:0x%08x)\n",
hflags_current, hflags_rebuilt);
}
-}
#endif
+ *pc = env->nip;
+ *cs_base = 0;
+ *flags = hflags_current;
+}
+
void cpu_interrupt_exittb(CPUState *cs)
{
/*
diff --git a/target/rx/cpu.c b/target/rx/cpu.c
index c9615f0655..f22737865e 100644
--- a/target/rx/cpu.c
+++ b/target/rx/cpu.c
@@ -28,6 +28,7 @@
#include "hw/loader.h"
#include "fpu/softfloat.h"
#include "tcg/debug-assert.h"
+#include "accel/tcg/cpu-ops.h"
static void rx_cpu_set_pc(CPUState *cs, vaddr value)
{
@@ -43,6 +44,15 @@ static vaddr rx_cpu_get_pc(CPUState *cs)
return cpu->env.pc;
}
+void cpu_get_tb_cpu_state(CPURXState *env, vaddr *pc,
+ uint64_t *cs_base, uint32_t *flags)
+{
+ *pc = env->pc;
+ *cs_base = 0;
+ *flags = FIELD_DP32(0, PSW, PM, env->psw_pm);
+ *flags = FIELD_DP32(*flags, PSW, U, env->psw_u);
+}
+
static void rx_cpu_synchronize_from_tb(CPUState *cs,
const TranslationBlock *tb)
{
@@ -201,8 +211,6 @@ static const struct SysemuCPUOps rx_sysemu_ops = {
.get_phys_page_debug = rx_cpu_get_phys_page_debug,
};
-#include "accel/tcg/cpu-ops.h"
-
static const TCGCPUOps rx_tcg_ops = {
/* MTTCG not yet supported: require strict ordering */
.guest_default_memory_order = TCG_MO_ALL,
diff --git a/target/s390x/cpu.c b/target/s390x/cpu.c
index f4289f9857..3ca3f880a7 100644
--- a/target/s390x/cpu.c
+++ b/target/s390x/cpu.c
@@ -302,6 +302,7 @@ static const Property s390x_cpu_properties[] = {
#ifdef CONFIG_TCG
#include "accel/tcg/cpu-ops.h"
+#include "tcg/tcg_s390x.h"
static int s390x_cpu_mmu_index(CPUState *cs, bool ifetch)
{
diff --git a/target/sh4/cpu.c b/target/sh4/cpu.c
index 2cafb56a23..adbd59cd68 100644
--- a/target/sh4/cpu.c
+++ b/target/sh4/cpu.c
@@ -26,6 +26,7 @@
#include "migration/vmstate.h"
#include "exec/translation-block.h"
#include "fpu/softfloat-helpers.h"
+#include "accel/tcg/cpu-ops.h"
#include "tcg/tcg.h"
static void superh_cpu_set_pc(CPUState *cs, vaddr value)
@@ -42,6 +43,21 @@ static vaddr superh_cpu_get_pc(CPUState *cs)
return cpu->env.pc;
}
+void cpu_get_tb_cpu_state(CPUSH4State *env, vaddr *pc,
+ uint64_t *cs_base, uint32_t *flags)
+{
+ *pc = env->pc;
+ /* For a gUSA region, notice the end of the region. */
+ *cs_base = env->flags & TB_FLAG_GUSA_MASK ? env->gregs[0] : 0;
+ *flags = env->flags
+ | (env->fpscr & TB_FLAG_FPSCR_MASK)
+ | (env->sr & TB_FLAG_SR_MASK)
+ | (env->movcal_backup ? TB_FLAG_PENDING_MOVCA : 0); /* Bit 3 */
+#ifdef CONFIG_USER_ONLY
+ *flags |= TB_FLAG_UNALIGN * !env_cpu(env)->prctl_unalign_sigbus;
+#endif
+}
+
static void superh_cpu_synchronize_from_tb(CPUState *cs,
const TranslationBlock *tb)
{
@@ -258,8 +274,6 @@ static const struct SysemuCPUOps sh4_sysemu_ops = {
};
#endif
-#include "accel/tcg/cpu-ops.h"
-
static const TCGCPUOps superh_tcg_ops = {
/* MTTCG not yet supported: require strict ordering */
.guest_default_memory_order = TCG_MO_ALL,
diff --git a/target/tricore/cpu.c b/target/tricore/cpu.c
index 36c82888e8..8538066114 100644
--- a/target/tricore/cpu.c
+++ b/target/tricore/cpu.c
@@ -23,6 +23,7 @@
#include "exec/translation-block.h"
#include "qemu/error-report.h"
#include "tcg/debug-assert.h"
+#include "accel/tcg/cpu-ops.h"
static inline void set_feature(CPUTriCoreState *env, int feature)
{
@@ -44,6 +45,18 @@ static vaddr tricore_cpu_get_pc(CPUState *cs)
return cpu_env(cs)->PC;
}
+void cpu_get_tb_cpu_state(CPUTriCoreState *env, vaddr *pc,
+ uint64_t *cs_base, uint32_t *flags)
+{
+ uint32_t new_flags = 0;
+ *pc = env->PC;
+ *cs_base = 0;
+
+ new_flags |= FIELD_DP32(new_flags, TB_FLAGS, PRIV,
+ extract32(env->PSW, 10, 2));
+ *flags = new_flags;
+}
+
static void tricore_cpu_synchronize_from_tb(CPUState *cs,
const TranslationBlock *tb)
{
@@ -168,8 +181,6 @@ static const struct SysemuCPUOps tricore_sysemu_ops = {
.get_phys_page_debug = tricore_cpu_get_phys_page_debug,
};
-#include "accel/tcg/cpu-ops.h"
-
static const TCGCPUOps tricore_tcg_ops = {
/* MTTCG not yet supported: require strict ordering */
.guest_default_memory_order = TCG_MO_ALL,
diff --git a/target/xtensa/cpu.c b/target/xtensa/cpu.c
index a13f1f950a..fe974e90f7 100644
--- a/target/xtensa/cpu.c
+++ b/target/xtensa/cpu.c
@@ -35,6 +35,7 @@
#include "qemu/module.h"
#include "migration/vmstate.h"
#include "hw/qdev-clock.h"
+#include "accel/tcg/cpu-ops.h"
#ifndef CONFIG_USER_ONLY
#include "system/memory.h"
#endif
@@ -54,6 +55,74 @@ static vaddr xtensa_cpu_get_pc(CPUState *cs)
return cpu->env.pc;
}
+void cpu_get_tb_cpu_state(CPUXtensaState *env, vaddr *pc,
+ uint64_t *cs_base, uint32_t *flags)
+{
+ *pc = env->pc;
+ *cs_base = 0;
+ *flags = 0;
+ *flags |= xtensa_get_ring(env);
+ if (env->sregs[PS] & PS_EXCM) {
+ *flags |= XTENSA_TBFLAG_EXCM;
+ } else if (xtensa_option_enabled(env->config, XTENSA_OPTION_LOOP)) {
+ target_ulong lend_dist =
+ env->sregs[LEND] - (env->pc & -(1u << TARGET_PAGE_BITS));
+
+ /*
+ * 0 in the csbase_lend field means that there may not be a loopback
+ * for any instruction that starts inside this page. Any other value
+ * means that an instruction that ends at this offset from the page
+ * start may loop back and will need loopback code to be generated.
+ *
+ * lend_dist is 0 when LEND points to the start of the page, but
+ * no instruction that starts inside this page may end at offset 0,
+ * so it's still correct.
+ *
+ * When an instruction ends at a page boundary it may only start in
+ * the previous page. lend_dist will be encoded as TARGET_PAGE_SIZE
+ * for the TB that contains this instruction.
+ */
+ if (lend_dist < (1u << TARGET_PAGE_BITS) + env->config->max_insn_size) {
+ target_ulong lbeg_off = env->sregs[LEND] - env->sregs[LBEG];
+
+ *cs_base = lend_dist;
+ if (lbeg_off < 256) {
+ *cs_base |= lbeg_off << XTENSA_CSBASE_LBEG_OFF_SHIFT;
+ }
+ }
+ }
+ if (xtensa_option_enabled(env->config, XTENSA_OPTION_EXTENDED_L32R) &&
+ (env->sregs[LITBASE] & 1)) {
+ *flags |= XTENSA_TBFLAG_LITBASE;
+ }
+ if (xtensa_option_enabled(env->config, XTENSA_OPTION_DEBUG)) {
+ if (xtensa_get_cintlevel(env) < env->config->debug_level) {
+ *flags |= XTENSA_TBFLAG_DEBUG;
+ }
+ if (xtensa_get_cintlevel(env) < env->sregs[ICOUNTLEVEL]) {
+ *flags |= XTENSA_TBFLAG_ICOUNT;
+ }
+ }
+ if (xtensa_option_enabled(env->config, XTENSA_OPTION_COPROCESSOR)) {
+ *flags |= env->sregs[CPENABLE] << XTENSA_TBFLAG_CPENABLE_SHIFT;
+ }
+ if (xtensa_option_enabled(env->config, XTENSA_OPTION_WINDOWED_REGISTER) &&
+ (env->sregs[PS] & (PS_WOE | PS_EXCM)) == PS_WOE) {
+ uint32_t windowstart = xtensa_replicate_windowstart(env) >>
+ (env->sregs[WINDOW_BASE] + 1);
+ uint32_t w = ctz32(windowstart | 0x8);
+
+ *flags |= (w << XTENSA_TBFLAG_WINDOW_SHIFT) | XTENSA_TBFLAG_CWOE;
+ *flags |= extract32(env->sregs[PS], PS_CALLINC_SHIFT,
+ PS_CALLINC_LEN) << XTENSA_TBFLAG_CALLINC_SHIFT;
+ } else {
+ *flags |= 3 << XTENSA_TBFLAG_WINDOW_SHIFT;
+ }
+ if (env->yield_needed) {
+ *flags |= XTENSA_TBFLAG_YIELD;
+ }
+}
+
static void xtensa_restore_state_to_opc(CPUState *cs,
const TranslationBlock *tb,
const uint64_t *data)
@@ -229,8 +298,6 @@ static const struct SysemuCPUOps xtensa_sysemu_ops = {
};
#endif
-#include "accel/tcg/cpu-ops.h"
-
static const TCGCPUOps xtensa_tcg_ops = {
/* Xtensa processors have a weak memory model */
.guest_default_memory_order = 0,
--
2.43.0
^ permalink raw reply related [flat|nested] 32+ messages in thread* Re: [PATCH 05/10] accel/tcg: Hoist cpu_get_tb_cpu_state decl to accl/tcg/cpu-ops.h
2025-04-28 20:10 ` [PATCH 05/10] accel/tcg: Hoist cpu_get_tb_cpu_state decl to accl/tcg/cpu-ops.h Richard Henderson
@ 2025-04-28 21:43 ` Pierrick Bouvier
0 siblings, 0 replies; 32+ messages in thread
From: Pierrick Bouvier @ 2025-04-28 21:43 UTC (permalink / raw)
To: Richard Henderson, qemu-devel; +Cc: philmd
On 4/28/25 1:10 PM, Richard Henderson wrote:
> For some targets, simply remove the local definition.
> For other targets, move the inline definition out of line.
>
> Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
> ---
> include/accel/tcg/cpu-ops.h | 3 ++
> target/alpha/cpu.h | 11 ------
> target/arm/cpu.h | 3 --
> target/avr/cpu.h | 18 ----------
> target/hexagon/cpu.h | 15 --------
> target/hppa/cpu.h | 3 --
> target/i386/cpu.h | 14 --------
> target/loongarch/cpu.h | 12 -------
> target/m68k/cpu.h | 16 ---------
> target/microblaze/cpu.h | 8 -----
> target/mips/cpu.h | 9 -----
> target/openrisc/cpu.h | 10 ------
> target/ppc/cpu.h | 13 -------
> target/riscv/cpu.h | 3 --
> target/rx/cpu.h | 9 -----
> target/s390x/cpu.h | 9 -----
> target/sh4/cpu.h | 15 --------
> target/sparc/cpu.h | 3 --
> target/tricore/cpu.h | 12 -------
> target/xtensa/cpu.h | 68 -----------------------------------
> target/alpha/cpu.c | 14 ++++++--
> target/arm/helper.c | 1 +
> target/avr/cpu.c | 21 +++++++++--
> target/hexagon/cpu.c | 18 ++++++++--
> target/hppa/cpu.c | 3 +-
> target/i386/tcg/tcg-cpu.c | 17 +++++++--
> target/loongarch/cpu.c | 15 ++++++--
> target/m68k/cpu.c | 19 ++++++++--
> target/microblaze/cpu.c | 11 ++++--
> target/mips/cpu.c | 9 +++++
> target/openrisc/cpu.c | 13 +++++--
> target/ppc/helper_regs.c | 16 ++++-----
> target/rx/cpu.c | 12 +++++--
> target/s390x/cpu.c | 1 +
> target/sh4/cpu.c | 18 ++++++++--
> target/tricore/cpu.c | 15 ++++++--
> target/xtensa/cpu.c | 71 +++++++++++++++++++++++++++++++++++--
> 37 files changed, 243 insertions(+), 285 deletions(-)
Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org>
^ permalink raw reply [flat|nested] 32+ messages in thread
* [PATCH 06/10] target/riscv: Move cpu_get_tb_cpu_state to tcg-cpu.c
2025-04-28 20:10 [PATCH 00/10] accel/tcg: Compile cpu-exec.c twice Richard Henderson
` (4 preceding siblings ...)
2025-04-28 20:10 ` [PATCH 05/10] accel/tcg: Hoist cpu_get_tb_cpu_state decl to accl/tcg/cpu-ops.h Richard Henderson
@ 2025-04-28 20:10 ` Richard Henderson
2025-04-28 21:41 ` Pierrick Bouvier
2025-04-29 6:39 ` Philippe Mathieu-Daudé
2025-04-28 20:10 ` [PATCH 07/10] accel/tcg: Return TCGTBCPUState from cpu_get_tb_cpu_state Richard Henderson
` (3 subsequent siblings)
9 siblings, 2 replies; 32+ messages in thread
From: Richard Henderson @ 2025-04-28 20:10 UTC (permalink / raw)
To: qemu-devel; +Cc: pierrick.bouvier, philmd
This function is only relevant to tcg.
Move it to a tcg-specific file.
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
---
target/riscv/cpu_helper.c | 97 -------------------------------------
target/riscv/tcg/tcg-cpu.c | 98 ++++++++++++++++++++++++++++++++++++++
2 files changed, 98 insertions(+), 97 deletions(-)
diff --git a/target/riscv/cpu_helper.c b/target/riscv/cpu_helper.c
index f2e90a9889..d5039f69a9 100644
--- a/target/riscv/cpu_helper.c
+++ b/target/riscv/cpu_helper.c
@@ -135,103 +135,6 @@ bool riscv_env_smode_dbltrp_enabled(CPURISCVState *env, bool virt)
#endif
}
-void cpu_get_tb_cpu_state(CPURISCVState *env, vaddr *pc,
- uint64_t *cs_base, uint32_t *pflags)
-{
- RISCVCPU *cpu = env_archcpu(env);
- RISCVExtStatus fs, vs;
- uint32_t flags = 0;
- bool pm_signext = riscv_cpu_virt_mem_enabled(env);
-
- *pc = env->xl == MXL_RV32 ? env->pc & UINT32_MAX : env->pc;
- *cs_base = 0;
-
- if (cpu->cfg.ext_zve32x) {
- /*
- * If env->vl equals to VLMAX, we can use generic vector operation
- * expanders (GVEC) to accerlate the vector operations.
- * However, as LMUL could be a fractional number. The maximum
- * vector size can be operated might be less than 8 bytes,
- * which is not supported by GVEC. So we set vl_eq_vlmax flag to true
- * only when maxsz >= 8 bytes.
- */
-
- /* lmul encoded as in DisasContext::lmul */
- int8_t lmul = sextract32(FIELD_EX64(env->vtype, VTYPE, VLMUL), 0, 3);
- uint32_t vsew = FIELD_EX64(env->vtype, VTYPE, VSEW);
- uint32_t vlmax = vext_get_vlmax(cpu->cfg.vlenb, vsew, lmul);
- uint32_t maxsz = vlmax << vsew;
- bool vl_eq_vlmax = (env->vstart == 0) && (vlmax == env->vl) &&
- (maxsz >= 8);
- flags = FIELD_DP32(flags, TB_FLAGS, VILL, env->vill);
- flags = FIELD_DP32(flags, TB_FLAGS, SEW, vsew);
- flags = FIELD_DP32(flags, TB_FLAGS, LMUL,
- FIELD_EX64(env->vtype, VTYPE, VLMUL));
- flags = FIELD_DP32(flags, TB_FLAGS, VL_EQ_VLMAX, vl_eq_vlmax);
- flags = FIELD_DP32(flags, TB_FLAGS, VTA,
- FIELD_EX64(env->vtype, VTYPE, VTA));
- flags = FIELD_DP32(flags, TB_FLAGS, VMA,
- FIELD_EX64(env->vtype, VTYPE, VMA));
- flags = FIELD_DP32(flags, TB_FLAGS, VSTART_EQ_ZERO, env->vstart == 0);
- } else {
- flags = FIELD_DP32(flags, TB_FLAGS, VILL, 1);
- }
-
- if (cpu_get_fcfien(env)) {
- /*
- * For Forward CFI, only the expectation of a lpad at
- * the start of the block is tracked via env->elp. env->elp
- * is turned on during jalr translation.
- */
- flags = FIELD_DP32(flags, TB_FLAGS, FCFI_LP_EXPECTED, env->elp);
- flags = FIELD_DP32(flags, TB_FLAGS, FCFI_ENABLED, 1);
- }
-
- if (cpu_get_bcfien(env)) {
- flags = FIELD_DP32(flags, TB_FLAGS, BCFI_ENABLED, 1);
- }
-
-#ifdef CONFIG_USER_ONLY
- fs = EXT_STATUS_DIRTY;
- vs = EXT_STATUS_DIRTY;
-#else
- flags = FIELD_DP32(flags, TB_FLAGS, PRIV, env->priv);
-
- flags |= riscv_env_mmu_index(env, 0);
- fs = get_field(env->mstatus, MSTATUS_FS);
- vs = get_field(env->mstatus, MSTATUS_VS);
-
- if (env->virt_enabled) {
- flags = FIELD_DP32(flags, TB_FLAGS, VIRT_ENABLED, 1);
- /*
- * Merge DISABLED and !DIRTY states using MIN.
- * We will set both fields when dirtying.
- */
- fs = MIN(fs, get_field(env->mstatus_hs, MSTATUS_FS));
- vs = MIN(vs, get_field(env->mstatus_hs, MSTATUS_VS));
- }
-
- /* With Zfinx, floating point is enabled/disabled by Smstateen. */
- if (!riscv_has_ext(env, RVF)) {
- fs = (smstateen_acc_ok(env, 0, SMSTATEEN0_FCSR) == RISCV_EXCP_NONE)
- ? EXT_STATUS_DIRTY : EXT_STATUS_DISABLED;
- }
-
- if (cpu->cfg.debug && !icount_enabled()) {
- flags = FIELD_DP32(flags, TB_FLAGS, ITRIGGER, env->itrigger_enabled);
- }
-#endif
-
- flags = FIELD_DP32(flags, TB_FLAGS, FS, fs);
- flags = FIELD_DP32(flags, TB_FLAGS, VS, vs);
- flags = FIELD_DP32(flags, TB_FLAGS, XL, env->xl);
- flags = FIELD_DP32(flags, TB_FLAGS, AXL, cpu_address_xl(env));
- flags = FIELD_DP32(flags, TB_FLAGS, PM_PMM, riscv_pm_get_pmm(env));
- flags = FIELD_DP32(flags, TB_FLAGS, PM_SIGNEXTEND, pm_signext);
-
- *pflags = flags;
-}
-
RISCVPmPmm riscv_pm_get_pmm(CPURISCVState *env)
{
#ifndef CONFIG_USER_ONLY
diff --git a/target/riscv/tcg/tcg-cpu.c b/target/riscv/tcg/tcg-cpu.c
index e3e7fea66a..f2cacef5e5 100644
--- a/target/riscv/tcg/tcg-cpu.c
+++ b/target/riscv/tcg/tcg-cpu.c
@@ -36,6 +36,7 @@
#ifndef CONFIG_USER_ONLY
#include "hw/boards.h"
#include "system/tcg.h"
+#include "exec/icount.h"
#endif
/* Hash that stores user set extensions */
@@ -97,6 +98,103 @@ static int riscv_cpu_mmu_index(CPUState *cs, bool ifetch)
return riscv_env_mmu_index(cpu_env(cs), ifetch);
}
+void cpu_get_tb_cpu_state(CPURISCVState *env, vaddr *pc,
+ uint64_t *cs_base, uint32_t *pflags)
+{
+ RISCVCPU *cpu = env_archcpu(env);
+ RISCVExtStatus fs, vs;
+ uint32_t flags = 0;
+ bool pm_signext = riscv_cpu_virt_mem_enabled(env);
+
+ *pc = env->xl == MXL_RV32 ? env->pc & UINT32_MAX : env->pc;
+ *cs_base = 0;
+
+ if (cpu->cfg.ext_zve32x) {
+ /*
+ * If env->vl equals to VLMAX, we can use generic vector operation
+ * expanders (GVEC) to accerlate the vector operations.
+ * However, as LMUL could be a fractional number. The maximum
+ * vector size can be operated might be less than 8 bytes,
+ * which is not supported by GVEC. So we set vl_eq_vlmax flag to true
+ * only when maxsz >= 8 bytes.
+ */
+
+ /* lmul encoded as in DisasContext::lmul */
+ int8_t lmul = sextract32(FIELD_EX64(env->vtype, VTYPE, VLMUL), 0, 3);
+ uint32_t vsew = FIELD_EX64(env->vtype, VTYPE, VSEW);
+ uint32_t vlmax = vext_get_vlmax(cpu->cfg.vlenb, vsew, lmul);
+ uint32_t maxsz = vlmax << vsew;
+ bool vl_eq_vlmax = (env->vstart == 0) && (vlmax == env->vl) &&
+ (maxsz >= 8);
+ flags = FIELD_DP32(flags, TB_FLAGS, VILL, env->vill);
+ flags = FIELD_DP32(flags, TB_FLAGS, SEW, vsew);
+ flags = FIELD_DP32(flags, TB_FLAGS, LMUL,
+ FIELD_EX64(env->vtype, VTYPE, VLMUL));
+ flags = FIELD_DP32(flags, TB_FLAGS, VL_EQ_VLMAX, vl_eq_vlmax);
+ flags = FIELD_DP32(flags, TB_FLAGS, VTA,
+ FIELD_EX64(env->vtype, VTYPE, VTA));
+ flags = FIELD_DP32(flags, TB_FLAGS, VMA,
+ FIELD_EX64(env->vtype, VTYPE, VMA));
+ flags = FIELD_DP32(flags, TB_FLAGS, VSTART_EQ_ZERO, env->vstart == 0);
+ } else {
+ flags = FIELD_DP32(flags, TB_FLAGS, VILL, 1);
+ }
+
+ if (cpu_get_fcfien(env)) {
+ /*
+ * For Forward CFI, only the expectation of a lpad at
+ * the start of the block is tracked via env->elp. env->elp
+ * is turned on during jalr translation.
+ */
+ flags = FIELD_DP32(flags, TB_FLAGS, FCFI_LP_EXPECTED, env->elp);
+ flags = FIELD_DP32(flags, TB_FLAGS, FCFI_ENABLED, 1);
+ }
+
+ if (cpu_get_bcfien(env)) {
+ flags = FIELD_DP32(flags, TB_FLAGS, BCFI_ENABLED, 1);
+ }
+
+#ifdef CONFIG_USER_ONLY
+ fs = EXT_STATUS_DIRTY;
+ vs = EXT_STATUS_DIRTY;
+#else
+ flags = FIELD_DP32(flags, TB_FLAGS, PRIV, env->priv);
+
+ flags |= riscv_env_mmu_index(env, 0);
+ fs = get_field(env->mstatus, MSTATUS_FS);
+ vs = get_field(env->mstatus, MSTATUS_VS);
+
+ if (env->virt_enabled) {
+ flags = FIELD_DP32(flags, TB_FLAGS, VIRT_ENABLED, 1);
+ /*
+ * Merge DISABLED and !DIRTY states using MIN.
+ * We will set both fields when dirtying.
+ */
+ fs = MIN(fs, get_field(env->mstatus_hs, MSTATUS_FS));
+ vs = MIN(vs, get_field(env->mstatus_hs, MSTATUS_VS));
+ }
+
+ /* With Zfinx, floating point is enabled/disabled by Smstateen. */
+ if (!riscv_has_ext(env, RVF)) {
+ fs = (smstateen_acc_ok(env, 0, SMSTATEEN0_FCSR) == RISCV_EXCP_NONE)
+ ? EXT_STATUS_DIRTY : EXT_STATUS_DISABLED;
+ }
+
+ if (cpu->cfg.debug && !icount_enabled()) {
+ flags = FIELD_DP32(flags, TB_FLAGS, ITRIGGER, env->itrigger_enabled);
+ }
+#endif
+
+ flags = FIELD_DP32(flags, TB_FLAGS, FS, fs);
+ flags = FIELD_DP32(flags, TB_FLAGS, VS, vs);
+ flags = FIELD_DP32(flags, TB_FLAGS, XL, env->xl);
+ flags = FIELD_DP32(flags, TB_FLAGS, AXL, cpu_address_xl(env));
+ flags = FIELD_DP32(flags, TB_FLAGS, PM_PMM, riscv_pm_get_pmm(env));
+ flags = FIELD_DP32(flags, TB_FLAGS, PM_SIGNEXTEND, pm_signext);
+
+ *pflags = flags;
+}
+
static void riscv_cpu_synchronize_from_tb(CPUState *cs,
const TranslationBlock *tb)
{
--
2.43.0
^ permalink raw reply related [flat|nested] 32+ messages in thread* Re: [PATCH 06/10] target/riscv: Move cpu_get_tb_cpu_state to tcg-cpu.c
2025-04-28 20:10 ` [PATCH 06/10] target/riscv: Move cpu_get_tb_cpu_state to tcg-cpu.c Richard Henderson
@ 2025-04-28 21:41 ` Pierrick Bouvier
2025-04-29 6:39 ` Philippe Mathieu-Daudé
1 sibling, 0 replies; 32+ messages in thread
From: Pierrick Bouvier @ 2025-04-28 21:41 UTC (permalink / raw)
To: Richard Henderson, qemu-devel; +Cc: philmd
On 4/28/25 1:10 PM, Richard Henderson wrote:
> This function is only relevant to tcg.
> Move it to a tcg-specific file.
>
> Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
> ---
> target/riscv/cpu_helper.c | 97 -------------------------------------
> target/riscv/tcg/tcg-cpu.c | 98 ++++++++++++++++++++++++++++++++++++++
> 2 files changed, 98 insertions(+), 97 deletions(-)
Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org>
^ permalink raw reply [flat|nested] 32+ messages in thread
* Re: [PATCH 06/10] target/riscv: Move cpu_get_tb_cpu_state to tcg-cpu.c
2025-04-28 20:10 ` [PATCH 06/10] target/riscv: Move cpu_get_tb_cpu_state to tcg-cpu.c Richard Henderson
2025-04-28 21:41 ` Pierrick Bouvier
@ 2025-04-29 6:39 ` Philippe Mathieu-Daudé
2025-04-29 20:00 ` Richard Henderson
1 sibling, 1 reply; 32+ messages in thread
From: Philippe Mathieu-Daudé @ 2025-04-29 6:39 UTC (permalink / raw)
To: Richard Henderson, qemu-devel; +Cc: pierrick.bouvier
On 28/4/25 22:10, Richard Henderson wrote:
> This function is only relevant to tcg.
> Move it to a tcg-specific file.
>
> Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
> ---
> target/riscv/cpu_helper.c | 97 -------------------------------------
> target/riscv/tcg/tcg-cpu.c | 98 ++++++++++++++++++++++++++++++++++++++
> 2 files changed, 98 insertions(+), 97 deletions(-)
https://lore.kernel.org/qemu-devel/20230703183145.24779-9-philmd@linaro.org/
;)
^ permalink raw reply [flat|nested] 32+ messages in thread
* [PATCH 07/10] accel/tcg: Return TCGTBCPUState from cpu_get_tb_cpu_state
2025-04-28 20:10 [PATCH 00/10] accel/tcg: Compile cpu-exec.c twice Richard Henderson
` (5 preceding siblings ...)
2025-04-28 20:10 ` [PATCH 06/10] target/riscv: Move cpu_get_tb_cpu_state to tcg-cpu.c Richard Henderson
@ 2025-04-28 20:10 ` Richard Henderson
2025-04-28 22:00 ` Pierrick Bouvier
2025-04-28 20:10 ` [PATCH 08/10] accel/tcg: Move cpu_get_tb_cpu_state to TCGCPUOps Richard Henderson
` (2 subsequent siblings)
9 siblings, 1 reply; 32+ messages in thread
From: Richard Henderson @ 2025-04-28 20:10 UTC (permalink / raw)
To: qemu-devel; +Cc: pierrick.bouvier, philmd
Combine 3 different pointer returns into one structure return.
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
---
accel/tcg/internal-common.h | 6 +++++
include/accel/tcg/cpu-ops.h | 9 +++++--
include/qemu/typedefs.h | 1 +
accel/tcg/cpu-exec.c | 48 +++++++++++++++----------------------
accel/tcg/translate-all.c | 8 ++-----
target/alpha/cpu.c | 13 +++++-----
target/arm/helper.c | 16 ++++++++-----
target/avr/cpu.c | 9 +++----
target/hexagon/cpu.c | 15 ++++++------
target/hppa/cpu.c | 10 ++++----
target/i386/tcg/tcg-cpu.c | 19 +++++++++------
target/loongarch/cpu.c | 20 +++++++++-------
target/m68k/cpu.c | 20 +++++++++-------
target/microblaze/cpu.c | 13 ++++++----
target/mips/cpu.c | 14 ++++++-----
target/openrisc/cpu.c | 16 +++++++------
target/ppc/helper_regs.c | 8 +++----
target/riscv/tcg/tcg-cpu.c | 12 +++++-----
target/rx/cpu.c | 14 ++++++-----
target/s390x/cpu.c | 14 ++++++-----
target/sh4/cpu.c | 22 +++++++++++------
target/sparc/cpu.c | 17 +++++++------
target/tricore/cpu.c | 14 +++++------
target/xtensa/cpu.c | 36 +++++++++++++++-------------
24 files changed, 202 insertions(+), 172 deletions(-)
diff --git a/accel/tcg/internal-common.h b/accel/tcg/internal-common.h
index 98c702422f..3ff4c1cb54 100644
--- a/accel/tcg/internal-common.h
+++ b/accel/tcg/internal-common.h
@@ -12,6 +12,7 @@
#include "exec/cpu-common.h"
#include "exec/translation-block.h"
#include "exec/mmap-lock.h"
+#include "accel/tcg/cpu-ops.h"
extern int64_t max_delay;
extern int64_t max_advance;
@@ -56,6 +57,11 @@ TranslationBlock *tb_link_page(TranslationBlock *tb);
void cpu_restore_state_from_tb(CPUState *cpu, TranslationBlock *tb,
uintptr_t host_pc);
+static inline TCGTBCPUState get_tb_cpu_state(CPUState *cs)
+{
+ return cpu_get_tb_cpu_state(cs);
+}
+
/**
* tlb_init - initialize a CPU's TLB
* @cpu: CPU whose TLB should be initialized
diff --git a/include/accel/tcg/cpu-ops.h b/include/accel/tcg/cpu-ops.h
index f5e5746976..8dbe79ea7c 100644
--- a/include/accel/tcg/cpu-ops.h
+++ b/include/accel/tcg/cpu-ops.h
@@ -18,8 +18,13 @@
#include "exec/vaddr.h"
#include "tcg/tcg-mo.h"
-void cpu_get_tb_cpu_state(CPUArchState *env, vaddr *pc,
- uint64_t *cs_base, uint32_t *flags);
+struct TCGTBCPUState {
+ vaddr pc;
+ uint32_t flags;
+ uint64_t flags2;
+};
+
+TCGTBCPUState cpu_get_tb_cpu_state(CPUState *cs);
struct TCGCPUOps {
/**
diff --git a/include/qemu/typedefs.h b/include/qemu/typedefs.h
index 507f0814d5..551239ca61 100644
--- a/include/qemu/typedefs.h
+++ b/include/qemu/typedefs.h
@@ -114,6 +114,7 @@ typedef struct SHPCDevice SHPCDevice;
typedef struct SSIBus SSIBus;
typedef struct TCGCPUOps TCGCPUOps;
typedef struct TCGHelperInfo TCGHelperInfo;
+typedef struct TCGTBCPUState TCGTBCPUState;
typedef struct TaskState TaskState;
typedef struct TranslationBlock TranslationBlock;
typedef struct VirtIODevice VirtIODevice;
diff --git a/accel/tcg/cpu-exec.c b/accel/tcg/cpu-exec.c
index c21c5d202d..a48df5f291 100644
--- a/accel/tcg/cpu-exec.c
+++ b/accel/tcg/cpu-exec.c
@@ -385,9 +385,6 @@ const void *HELPER(lookup_tb_ptr)(CPUArchState *env)
{
CPUState *cpu = env_cpu(env);
TranslationBlock *tb;
- vaddr pc;
- uint64_t cs_base;
- uint32_t flags, cflags;
/*
* By definition we've just finished a TB, so I/O is OK.
@@ -397,20 +394,21 @@ const void *HELPER(lookup_tb_ptr)(CPUArchState *env)
* The next TB, if we chain to it, will clear the flag again.
*/
cpu->neg.can_do_io = true;
- cpu_get_tb_cpu_state(env, &pc, &cs_base, &flags);
- cflags = curr_cflags(cpu);
- if (check_for_breakpoints(cpu, pc, &cflags)) {
+ TCGTBCPUState s = get_tb_cpu_state(cpu);
+ uint32_t cflags = curr_cflags(cpu);
+
+ if (check_for_breakpoints(cpu, s.pc, &cflags)) {
cpu_loop_exit(cpu);
}
- tb = tb_lookup(cpu, pc, cs_base, flags, cflags);
+ tb = tb_lookup(cpu, s.pc, s.flags2, s.flags, cflags);
if (tb == NULL) {
return tcg_code_gen_epilogue;
}
if (qemu_loglevel_mask(CPU_LOG_TB_CPU | CPU_LOG_EXEC)) {
- log_cpu_exec(pc, cpu, tb);
+ log_cpu_exec(s.pc, cpu, tb);
}
return tb->tc.ptr;
@@ -560,11 +558,7 @@ static void cpu_exec_longjmp_cleanup(CPUState *cpu)
void cpu_exec_step_atomic(CPUState *cpu)
{
- CPUArchState *env = cpu_env(cpu);
TranslationBlock *tb;
- vaddr pc;
- uint64_t cs_base;
- uint32_t flags, cflags;
int tb_exit;
if (sigsetjmp(cpu->jmp_env, 0) == 0) {
@@ -573,9 +567,9 @@ void cpu_exec_step_atomic(CPUState *cpu)
g_assert(!cpu->running);
cpu->running = true;
- cpu_get_tb_cpu_state(env, &pc, &cs_base, &flags);
+ TCGTBCPUState s = get_tb_cpu_state(cpu);
+ uint32_t cflags = curr_cflags(cpu);
- cflags = curr_cflags(cpu);
/* Execute in a serial context. */
cflags &= ~CF_PARALLEL;
/* After 1 insn, return and release the exclusive lock. */
@@ -587,16 +581,16 @@ void cpu_exec_step_atomic(CPUState *cpu)
* Any breakpoint for this insn will have been recognized earlier.
*/
- tb = tb_lookup(cpu, pc, cs_base, flags, cflags);
+ tb = tb_lookup(cpu, s.pc, s.flags2, s.flags, cflags);
if (tb == NULL) {
mmap_lock();
- tb = tb_gen_code(cpu, pc, cs_base, flags, cflags);
+ tb = tb_gen_code(cpu, s.pc, s.flags2, s.flags, cflags);
mmap_unlock();
}
cpu_exec_enter(cpu);
/* execute the generated code */
- trace_exec_tb(tb, pc);
+ trace_exec_tb(tb, s.pc);
cpu_tb_exec(cpu, tb, &tb_exit);
cpu_exec_exit(cpu);
} else {
@@ -941,11 +935,8 @@ cpu_exec_loop(CPUState *cpu, SyncClocks *sc)
while (!cpu_handle_interrupt(cpu, &last_tb)) {
TranslationBlock *tb;
- vaddr pc;
- uint64_t cs_base;
- uint32_t flags, cflags;
-
- cpu_get_tb_cpu_state(cpu_env(cpu), &pc, &cs_base, &flags);
+ TCGTBCPUState s = get_tb_cpu_state(cpu);
+ uint32_t cflags = cpu->cflags_next_tb;
/*
* When requested, use an exact setting for cflags for the next
@@ -954,33 +945,32 @@ cpu_exec_loop(CPUState *cpu, SyncClocks *sc)
* have CF_INVALID set, -1 is a convenient invalid value that
* does not require tcg headers for cpu_common_reset.
*/
- cflags = cpu->cflags_next_tb;
if (cflags == -1) {
cflags = curr_cflags(cpu);
} else {
cpu->cflags_next_tb = -1;
}
- if (check_for_breakpoints(cpu, pc, &cflags)) {
+ if (check_for_breakpoints(cpu, s.pc, &cflags)) {
break;
}
- tb = tb_lookup(cpu, pc, cs_base, flags, cflags);
+ tb = tb_lookup(cpu, s.pc, s.flags2, s.flags, cflags);
if (tb == NULL) {
CPUJumpCache *jc;
uint32_t h;
mmap_lock();
- tb = tb_gen_code(cpu, pc, cs_base, flags, cflags);
+ tb = tb_gen_code(cpu, s.pc, s.flags2, s.flags, cflags);
mmap_unlock();
/*
* We add the TB in the virtual pc hash table
* for the fast lookup
*/
- h = tb_jmp_cache_hash_func(pc);
+ h = tb_jmp_cache_hash_func(s.pc);
jc = cpu->tb_jmp_cache;
- jc->array[h].pc = pc;
+ jc->array[h].pc = s.pc;
qatomic_set(&jc->array[h].tb, tb);
}
@@ -1000,7 +990,7 @@ cpu_exec_loop(CPUState *cpu, SyncClocks *sc)
tb_add_jump(last_tb, tb_exit, tb);
}
- cpu_loop_exec_tb(cpu, tb, pc, &last_tb, &tb_exit);
+ cpu_loop_exec_tb(cpu, tb, s.pc, &last_tb, &tb_exit);
/* Try to align the host and virtual clocks
if the guest is in advance */
diff --git a/accel/tcg/translate-all.c b/accel/tcg/translate-all.c
index 31c7f9927f..759ae77559 100644
--- a/accel/tcg/translate-all.c
+++ b/accel/tcg/translate-all.c
@@ -590,13 +590,9 @@ void tb_check_watchpoint(CPUState *cpu, uintptr_t retaddr)
/* The exception probably happened in a helper. The CPU state should
have been saved before calling it. Fetch the PC from there. */
CPUArchState *env = cpu_env(cpu);
- vaddr pc;
- uint64_t cs_base;
- tb_page_addr_t addr;
- uint32_t flags;
+ TCGTBCPUState s = get_tb_cpu_state(cpu);
+ tb_page_addr_t addr = get_page_addr_code(env, s.pc);
- cpu_get_tb_cpu_state(env, &pc, &cs_base, &flags);
- addr = get_page_addr_code(env, pc);
if (addr != -1) {
tb_invalidate_phys_range(cpu, addr, addr);
}
diff --git a/target/alpha/cpu.c b/target/alpha/cpu.c
index d1fddef1f4..6efaa71543 100644
--- a/target/alpha/cpu.c
+++ b/target/alpha/cpu.c
@@ -41,15 +41,16 @@ static vaddr alpha_cpu_get_pc(CPUState *cs)
return env->pc;
}
-void cpu_get_tb_cpu_state(CPUAlphaState *env, vaddr *pc,
- uint64_t *cs_base, uint32_t *pflags)
+TCGTBCPUState cpu_get_tb_cpu_state(CPUState *cs)
{
- *pc = env->pc;
- *cs_base = 0;
- *pflags = env->flags & ENV_FLAG_TB_MASK;
+ CPUAlphaState *env = cpu_env(cs);
+ uint32_t flags = env->flags & ENV_FLAG_TB_MASK;
+
#ifdef CONFIG_USER_ONLY
- *pflags |= TB_FLAG_UNALIGN * !env_cpu(env)->prctl_unalign_sigbus;
+ flags |= TB_FLAG_UNALIGN * !cs->prctl_unalign_sigbus;
#endif
+
+ return (TCGTBCPUState){ .pc = env->pc, .flags = flags };
}
static void alpha_cpu_synchronize_from_tb(CPUState *cs,
diff --git a/target/arm/helper.c b/target/arm/helper.c
index 98adeb7086..aa887af50f 100644
--- a/target/arm/helper.c
+++ b/target/arm/helper.c
@@ -11453,21 +11453,22 @@ static bool mve_no_pred(CPUARMState *env)
return true;
}
-void cpu_get_tb_cpu_state(CPUARMState *env, vaddr *pc,
- uint64_t *cs_base, uint32_t *pflags)
+TCGTBCPUState cpu_get_tb_cpu_state(CPUState *cs)
{
+ CPUARMState *env = cpu_env(cs);
CPUARMTBFlags flags;
+ vaddr pc;
assert_hflags_rebuild_correctly(env);
flags = env->hflags;
if (EX_TBFLAG_ANY(flags, AARCH64_STATE)) {
- *pc = env->pc;
+ pc = env->pc;
if (cpu_isar_feature(aa64_bti, env_archcpu(env))) {
DP_TBFLAG_A64(flags, BTYPE, env->btype);
}
} else {
- *pc = env->regs[15];
+ pc = env->regs[15];
if (arm_feature(env, ARM_FEATURE_M)) {
if (arm_feature(env, ARM_FEATURE_M_SECURITY) &&
@@ -11529,8 +11530,11 @@ void cpu_get_tb_cpu_state(CPUARMState *env, vaddr *pc,
DP_TBFLAG_ANY(flags, PSTATE__SS, 1);
}
- *pflags = flags.flags;
- *cs_base = flags.flags2;
+ return (TCGTBCPUState){
+ .pc = pc,
+ .flags = flags.flags,
+ .flags2 = flags.flags2,
+ };
}
#ifdef TARGET_AARCH64
diff --git a/target/avr/cpu.c b/target/avr/cpu.c
index fac8954c39..88a4b970df 100644
--- a/target/avr/cpu.c
+++ b/target/avr/cpu.c
@@ -54,14 +54,11 @@ static int avr_cpu_mmu_index(CPUState *cs, bool ifetch)
return ifetch ? MMU_CODE_IDX : MMU_DATA_IDX;
}
-void cpu_get_tb_cpu_state(CPUAVRState *env, vaddr *pc,
- uint64_t *cs_base, uint32_t *pflags)
+TCGTBCPUState cpu_get_tb_cpu_state(CPUState *cs)
{
+ CPUAVRState *env = cpu_env(cs);
uint32_t flags = 0;
- *pc = env->pc_w * 2;
- *cs_base = 0;
-
if (env->fullacc) {
flags |= TB_FLAGS_FULL_ACCESS;
}
@@ -69,7 +66,7 @@ void cpu_get_tb_cpu_state(CPUAVRState *env, vaddr *pc,
flags |= TB_FLAGS_SKIP;
}
- *pflags = flags;
+ return (TCGTBCPUState){ .pc = env->pc_w * 2, .flags = flags };
}
static void avr_cpu_synchronize_from_tb(CPUState *cs,
diff --git a/target/hexagon/cpu.c b/target/hexagon/cpu.c
index e1a93ce24f..08c9aef55e 100644
--- a/target/hexagon/cpu.c
+++ b/target/hexagon/cpu.c
@@ -255,19 +255,20 @@ static vaddr hexagon_cpu_get_pc(CPUState *cs)
return cpu_env(cs)->gpr[HEX_REG_PC];
}
-void cpu_get_tb_cpu_state(CPUHexagonState *env, vaddr *pc,
- uint64_t *cs_base, uint32_t *flags)
+TCGTBCPUState cpu_get_tb_cpu_state(CPUState *cs)
{
+ CPUHexagonState *env = cpu_env(cs);
+ vaddr pc = env->gpr[HEX_REG_PC];
uint32_t hex_flags = 0;
- *pc = env->gpr[HEX_REG_PC];
- *cs_base = 0;
- if (*pc == env->gpr[HEX_REG_SA0]) {
+
+ if (pc == env->gpr[HEX_REG_SA0]) {
hex_flags = FIELD_DP32(hex_flags, TB_FLAGS, IS_TIGHT_LOOP, 1);
}
- *flags = hex_flags;
- if (*pc & PCALIGN_MASK) {
+ if (pc & PCALIGN_MASK) {
hexagon_raise_exception_err(env, HEX_CAUSE_PC_NOT_ALIGNED, 0);
}
+
+ return (TCGTBCPUState){ .pc = pc, .flags = hex_flags };
}
static void hexagon_cpu_synchronize_from_tb(CPUState *cs,
diff --git a/target/hppa/cpu.c b/target/hppa/cpu.c
index 24ca2d7175..7270b0a2a7 100644
--- a/target/hppa/cpu.c
+++ b/target/hppa/cpu.c
@@ -51,11 +51,12 @@ static vaddr hppa_cpu_get_pc(CPUState *cs)
env->iaoq_f & -4);
}
-void cpu_get_tb_cpu_state(CPUHPPAState *env, vaddr *pc,
- uint64_t *pcsbase, uint32_t *pflags)
+TCGTBCPUState cpu_get_tb_cpu_state(CPUState *cs)
{
+ CPUHPPAState *env = cpu_env(cs);
uint32_t flags = 0;
uint64_t cs_base = 0;
+ vaddr pc;
/*
* TB lookup assumes that PC contains the complete virtual address.
@@ -63,7 +64,7 @@ void cpu_get_tb_cpu_state(CPUHPPAState *env, vaddr *pc,
* incomplete virtual address. This also means that we must separate
* out current cpu privilege from the low bits of IAOQ_F.
*/
- *pc = hppa_cpu_get_pc(env_cpu(env));
+ pc = hppa_cpu_get_pc(env_cpu(env));
flags |= (env->iaoq_f & 3) << TB_FLAG_PRIV_SHIFT;
/*
@@ -99,8 +100,7 @@ void cpu_get_tb_cpu_state(CPUHPPAState *env, vaddr *pc,
}
#endif
- *pcsbase = cs_base;
- *pflags = flags;
+ return (TCGTBCPUState){ .pc = pc, .flags = flags, .flags2 = cs_base };
}
static void hppa_cpu_synchronize_from_tb(CPUState *cs,
diff --git a/target/i386/tcg/tcg-cpu.c b/target/i386/tcg/tcg-cpu.c
index cf9ce70139..950b3bfa76 100644
--- a/target/i386/tcg/tcg-cpu.c
+++ b/target/i386/tcg/tcg-cpu.c
@@ -48,18 +48,23 @@ static void x86_cpu_exec_exit(CPUState *cs)
env->eflags = cpu_compute_eflags(env);
}
-void cpu_get_tb_cpu_state(CPUX86State *env, vaddr *pc,
- uint64_t *cs_base, uint32_t *flags)
+TCGTBCPUState cpu_get_tb_cpu_state(CPUState *cs)
{
- *flags = env->hflags |
+ CPUX86State *env = cpu_env(cs);
+ uint32_t flags, cs_base;
+ vaddr pc;
+
+ flags = env->hflags |
(env->eflags & (IOPL_MASK | TF_MASK | RF_MASK | VM_MASK | AC_MASK));
if (env->hflags & HF_CS64_MASK) {
- *cs_base = 0;
- *pc = env->eip;
+ cs_base = 0;
+ pc = env->eip;
} else {
- *cs_base = env->segs[R_CS].base;
- *pc = (uint32_t)(*cs_base + env->eip);
+ cs_base = env->segs[R_CS].base;
+ pc = (uint32_t)(cs_base + env->eip);
}
+
+ return (TCGTBCPUState){ .pc = pc, .flags = flags, .flags2 = cs_base };
}
static void x86_cpu_synchronize_from_tb(CPUState *cs,
diff --git a/target/loongarch/cpu.c b/target/loongarch/cpu.c
index db66a6bdeb..1691fe1d9f 100644
--- a/target/loongarch/cpu.c
+++ b/target/loongarch/cpu.c
@@ -336,16 +336,18 @@ static bool loongarch_cpu_exec_interrupt(CPUState *cs, int interrupt_request)
}
#endif
-void cpu_get_tb_cpu_state(CPULoongArchState *env, vaddr *pc,
- uint64_t *cs_base, uint32_t *flags)
+TCGTBCPUState cpu_get_tb_cpu_state(CPUState *cs)
{
- *pc = env->pc;
- *cs_base = 0;
- *flags = env->CSR_CRMD & (R_CSR_CRMD_PLV_MASK | R_CSR_CRMD_PG_MASK);
- *flags |= FIELD_EX64(env->CSR_EUEN, CSR_EUEN, FPE) * HW_FLAGS_EUEN_FPE;
- *flags |= FIELD_EX64(env->CSR_EUEN, CSR_EUEN, SXE) * HW_FLAGS_EUEN_SXE;
- *flags |= FIELD_EX64(env->CSR_EUEN, CSR_EUEN, ASXE) * HW_FLAGS_EUEN_ASXE;
- *flags |= is_va32(env) * HW_FLAGS_VA32;
+ CPULoongArchState *env = cpu_env(cs);
+ uint32_t flags;
+
+ flags = env->CSR_CRMD & (R_CSR_CRMD_PLV_MASK | R_CSR_CRMD_PG_MASK);
+ flags |= FIELD_EX64(env->CSR_EUEN, CSR_EUEN, FPE) * HW_FLAGS_EUEN_FPE;
+ flags |= FIELD_EX64(env->CSR_EUEN, CSR_EUEN, SXE) * HW_FLAGS_EUEN_SXE;
+ flags |= FIELD_EX64(env->CSR_EUEN, CSR_EUEN, ASXE) * HW_FLAGS_EUEN_ASXE;
+ flags |= is_va32(env) * HW_FLAGS_VA32;
+
+ return (TCGTBCPUState){ .pc = env->pc, .flags = flags };
}
static void loongarch_cpu_synchronize_from_tb(CPUState *cs,
diff --git a/target/m68k/cpu.c b/target/m68k/cpu.c
index 0fc3ed316d..1786d9e0ce 100644
--- a/target/m68k/cpu.c
+++ b/target/m68k/cpu.c
@@ -39,20 +39,22 @@ static vaddr m68k_cpu_get_pc(CPUState *cs)
return cpu->env.pc;
}
-void cpu_get_tb_cpu_state(CPUM68KState *env, vaddr *pc,
- uint64_t *cs_base, uint32_t *flags)
+TCGTBCPUState cpu_get_tb_cpu_state(CPUState *cs)
{
- *pc = env->pc;
- *cs_base = 0;
- *flags = (env->macsr >> 4) & TB_FLAGS_MACSR;
+ CPUM68KState *env = cpu_env(cs);
+ uint32_t flags;
+
+ flags = (env->macsr >> 4) & TB_FLAGS_MACSR;
if (env->sr & SR_S) {
- *flags |= TB_FLAGS_MSR_S;
- *flags |= (env->sfc << (TB_FLAGS_SFC_S_BIT - 2)) & TB_FLAGS_SFC_S;
- *flags |= (env->dfc << (TB_FLAGS_DFC_S_BIT - 2)) & TB_FLAGS_DFC_S;
+ flags |= TB_FLAGS_MSR_S;
+ flags |= (env->sfc << (TB_FLAGS_SFC_S_BIT - 2)) & TB_FLAGS_SFC_S;
+ flags |= (env->dfc << (TB_FLAGS_DFC_S_BIT - 2)) & TB_FLAGS_DFC_S;
}
if (M68K_SR_TRACE(env->sr) == M68K_SR_TRACE_ANY_INS) {
- *flags |= TB_FLAGS_TRACE;
+ flags |= TB_FLAGS_TRACE;
}
+
+ return (TCGTBCPUState){ .pc = env->pc, .flags = flags };
}
static void m68k_restore_state_to_opc(CPUState *cs,
diff --git a/target/microblaze/cpu.c b/target/microblaze/cpu.c
index a91844e4bf..91e77a4e53 100644
--- a/target/microblaze/cpu.c
+++ b/target/microblaze/cpu.c
@@ -95,12 +95,15 @@ static vaddr mb_cpu_get_pc(CPUState *cs)
return cpu->env.pc;
}
-void cpu_get_tb_cpu_state(CPUMBState *env, vaddr *pc,
- uint64_t *cs_base, uint32_t *flags)
+TCGTBCPUState cpu_get_tb_cpu_state(CPUState *cs)
{
- *pc = env->pc;
- *flags = (env->iflags & IFLAGS_TB_MASK) | (env->msr & MSR_TB_MASK);
- *cs_base = (*flags & IMM_FLAG ? env->imm : 0);
+ CPUMBState *env = cpu_env(cs);
+
+ return (TCGTBCPUState){
+ .pc = env->pc,
+ .flags = (env->iflags & IFLAGS_TB_MASK) | (env->msr & MSR_TB_MASK),
+ .flags2 = (env->iflags & IMM_FLAG ? env->imm : 0),
+ };
}
static void mb_cpu_synchronize_from_tb(CPUState *cs,
diff --git a/target/mips/cpu.c b/target/mips/cpu.c
index 066dadc71b..c514202fbc 100644
--- a/target/mips/cpu.c
+++ b/target/mips/cpu.c
@@ -549,13 +549,15 @@ static int mips_cpu_mmu_index(CPUState *cs, bool ifunc)
return mips_env_mmu_index(cpu_env(cs));
}
-void cpu_get_tb_cpu_state(CPUMIPSState *env, vaddr *pc,
- uint64_t *cs_base, uint32_t *flags)
+TCGTBCPUState cpu_get_tb_cpu_state(CPUState *cs)
{
- *pc = env->active_tc.PC;
- *cs_base = 0;
- *flags = env->hflags & (MIPS_HFLAG_TMASK | MIPS_HFLAG_BMASK |
- MIPS_HFLAG_HWRENA_ULR);
+ CPUMIPSState *env = cpu_env(cs);
+
+ return (TCGTBCPUState){
+ .pc = env->active_tc.PC,
+ .flags = env->hflags & (MIPS_HFLAG_TMASK | MIPS_HFLAG_BMASK |
+ MIPS_HFLAG_HWRENA_ULR),
+ };
}
static const TCGCPUOps mips_tcg_ops = {
diff --git a/target/openrisc/cpu.c b/target/openrisc/cpu.c
index cfb3f62663..6b31f9c80e 100644
--- a/target/openrisc/cpu.c
+++ b/target/openrisc/cpu.c
@@ -41,14 +41,16 @@ static vaddr openrisc_cpu_get_pc(CPUState *cs)
return cpu->env.pc;
}
-void cpu_get_tb_cpu_state(CPUOpenRISCState *env, vaddr *pc,
- uint64_t *cs_base, uint32_t *flags)
+TCGTBCPUState cpu_get_tb_cpu_state(CPUState *cs)
{
- *pc = env->pc;
- *cs_base = 0;
- *flags = (env->dflag ? TB_FLAGS_DFLAG : 0)
- | (cpu_get_gpr(env, 0) ? 0 : TB_FLAGS_R0_0)
- | (env->sr & (SR_SM | SR_DME | SR_IME | SR_OVE));
+ CPUOpenRISCState *env = cpu_env(cs);
+
+ return (TCGTBCPUState){
+ .pc = env->pc,
+ .flags = ((env->dflag ? TB_FLAGS_DFLAG : 0)
+ | (cpu_get_gpr(env, 0) ? 0 : TB_FLAGS_R0_0)
+ | (env->sr & (SR_SM | SR_DME | SR_IME | SR_OVE))),
+ };
}
static void openrisc_cpu_synchronize_from_tb(CPUState *cs,
diff --git a/target/ppc/helper_regs.c b/target/ppc/helper_regs.c
index 8d248bcbb9..ccaf2b0343 100644
--- a/target/ppc/helper_regs.c
+++ b/target/ppc/helper_regs.c
@@ -256,9 +256,9 @@ void hreg_update_pmu_hflags(CPUPPCState *env)
env->hflags |= hreg_compute_pmu_hflags_value(env);
}
-void cpu_get_tb_cpu_state(CPUPPCState *env, vaddr *pc,
- uint64_t *cs_base, uint32_t *flags)
+TCGTBCPUState cpu_get_tb_cpu_state(CPUState *cs)
{
+ CPUPPCState *env = cpu_env(cs);
uint32_t hflags_current = env->hflags;
#ifdef CONFIG_DEBUG_TCG
@@ -270,9 +270,7 @@ void cpu_get_tb_cpu_state(CPUPPCState *env, vaddr *pc,
}
#endif
- *pc = env->nip;
- *cs_base = 0;
- *flags = hflags_current;
+ return (TCGTBCPUState){ .pc = env->nip, .flags = hflags_current };
}
void cpu_interrupt_exittb(CPUState *cs)
diff --git a/target/riscv/tcg/tcg-cpu.c b/target/riscv/tcg/tcg-cpu.c
index f2cacef5e5..461b8c97b4 100644
--- a/target/riscv/tcg/tcg-cpu.c
+++ b/target/riscv/tcg/tcg-cpu.c
@@ -98,17 +98,14 @@ static int riscv_cpu_mmu_index(CPUState *cs, bool ifetch)
return riscv_env_mmu_index(cpu_env(cs), ifetch);
}
-void cpu_get_tb_cpu_state(CPURISCVState *env, vaddr *pc,
- uint64_t *cs_base, uint32_t *pflags)
+TCGTBCPUState cpu_get_tb_cpu_state(CPUState *cs)
{
+ CPURISCVState *env = cpu_env(cs);
RISCVCPU *cpu = env_archcpu(env);
RISCVExtStatus fs, vs;
uint32_t flags = 0;
bool pm_signext = riscv_cpu_virt_mem_enabled(env);
- *pc = env->xl == MXL_RV32 ? env->pc & UINT32_MAX : env->pc;
- *cs_base = 0;
-
if (cpu->cfg.ext_zve32x) {
/*
* If env->vl equals to VLMAX, we can use generic vector operation
@@ -192,7 +189,10 @@ void cpu_get_tb_cpu_state(CPURISCVState *env, vaddr *pc,
flags = FIELD_DP32(flags, TB_FLAGS, PM_PMM, riscv_pm_get_pmm(env));
flags = FIELD_DP32(flags, TB_FLAGS, PM_SIGNEXTEND, pm_signext);
- *pflags = flags;
+ return (TCGTBCPUState){
+ .pc = env->xl == MXL_RV32 ? env->pc & UINT32_MAX : env->pc,
+ .flags = flags
+ };
}
static void riscv_cpu_synchronize_from_tb(CPUState *cs,
diff --git a/target/rx/cpu.c b/target/rx/cpu.c
index f22737865e..b8bba1a0c7 100644
--- a/target/rx/cpu.c
+++ b/target/rx/cpu.c
@@ -44,13 +44,15 @@ static vaddr rx_cpu_get_pc(CPUState *cs)
return cpu->env.pc;
}
-void cpu_get_tb_cpu_state(CPURXState *env, vaddr *pc,
- uint64_t *cs_base, uint32_t *flags)
+TCGTBCPUState cpu_get_tb_cpu_state(CPUState *cs)
{
- *pc = env->pc;
- *cs_base = 0;
- *flags = FIELD_DP32(0, PSW, PM, env->psw_pm);
- *flags = FIELD_DP32(*flags, PSW, U, env->psw_u);
+ CPURXState *env = cpu_env(cs);
+ uint32_t flags = 0;
+
+ flags = FIELD_DP32(flags, PSW, PM, env->psw_pm);
+ flags = FIELD_DP32(flags, PSW, U, env->psw_u);
+
+ return (TCGTBCPUState){ .pc = env->pc, .flags = flags };
}
static void rx_cpu_synchronize_from_tb(CPUState *cs,
diff --git a/target/s390x/cpu.c b/target/s390x/cpu.c
index 3ca3f880a7..127574c085 100644
--- a/target/s390x/cpu.c
+++ b/target/s390x/cpu.c
@@ -309,9 +309,9 @@ static int s390x_cpu_mmu_index(CPUState *cs, bool ifetch)
return s390x_env_mmu_index(cpu_env(cs), ifetch);
}
-void cpu_get_tb_cpu_state(CPUS390XState *env, vaddr *pc,
- uint64_t *cs_base, uint32_t *pflags)
+TCGTBCPUState cpu_get_tb_cpu_state(CPUState *cs)
{
+ CPUS390XState *env = cpu_env(cs);
uint32_t flags;
if (env->psw.addr & 1) {
@@ -323,9 +323,6 @@ void cpu_get_tb_cpu_state(CPUS390XState *env, vaddr *pc,
tcg_s390_program_interrupt(env, PGM_SPECIFICATION, 0);
}
- *pc = env->psw.addr;
- *cs_base = env->ex_value;
-
flags = (env->psw.mask >> FLAG_MASK_PSW_SHIFT) & FLAG_MASK_PSW;
if (env->psw.mask & PSW_MASK_PER) {
flags |= env->cregs[9] & (FLAG_MASK_PER_BRANCH |
@@ -342,7 +339,12 @@ void cpu_get_tb_cpu_state(CPUS390XState *env, vaddr *pc,
if (env->cregs[0] & CR0_VECTOR) {
flags |= FLAG_MASK_VECTOR;
}
- *pflags = flags;
+
+ return (TCGTBCPUState){
+ .pc = env->psw.addr,
+ .flags = flags,
+ .flags2 = env->ex_value
+ };
}
static const TCGCPUOps s390_tcg_ops = {
diff --git a/target/sh4/cpu.c b/target/sh4/cpu.c
index adbd59cd68..320cb8ddcc 100644
--- a/target/sh4/cpu.c
+++ b/target/sh4/cpu.c
@@ -43,19 +43,27 @@ static vaddr superh_cpu_get_pc(CPUState *cs)
return cpu->env.pc;
}
-void cpu_get_tb_cpu_state(CPUSH4State *env, vaddr *pc,
- uint64_t *cs_base, uint32_t *flags)
+TCGTBCPUState cpu_get_tb_cpu_state(CPUState *cs)
{
- *pc = env->pc;
- /* For a gUSA region, notice the end of the region. */
- *cs_base = env->flags & TB_FLAG_GUSA_MASK ? env->gregs[0] : 0;
- *flags = env->flags
+ CPUSH4State *env = cpu_env(cs);
+ uint32_t flags;
+
+ flags = env->flags
| (env->fpscr & TB_FLAG_FPSCR_MASK)
| (env->sr & TB_FLAG_SR_MASK)
| (env->movcal_backup ? TB_FLAG_PENDING_MOVCA : 0); /* Bit 3 */
#ifdef CONFIG_USER_ONLY
- *flags |= TB_FLAG_UNALIGN * !env_cpu(env)->prctl_unalign_sigbus;
+ flags |= TB_FLAG_UNALIGN * !cs->prctl_unalign_sigbus;
#endif
+
+ return (TCGTBCPUState){
+ .pc = env->pc,
+ .flags = flags,
+#ifdef CONFIG_USER_ONLY
+ /* For a gUSA region, notice the end of the region. */
+ .flags2 = flags & TB_FLAG_GUSA_MASK ? env->gregs[0] : 0,
+#endif
+ };
}
static void superh_cpu_synchronize_from_tb(CPUState *cs,
diff --git a/target/sparc/cpu.c b/target/sparc/cpu.c
index 5a1f5b7915..77591e84ba 100644
--- a/target/sparc/cpu.c
+++ b/target/sparc/cpu.c
@@ -716,13 +716,11 @@ static void sparc_cpu_synchronize_from_tb(CPUState *cs,
cpu->env.npc = tb->cs_base;
}
-void cpu_get_tb_cpu_state(CPUSPARCState *env, vaddr *pc,
- uint64_t *cs_base, uint32_t *pflags)
+TCGTBCPUState cpu_get_tb_cpu_state(CPUState *cs)
{
- uint32_t flags;
- *pc = env->pc;
- *cs_base = env->npc;
- flags = cpu_mmu_index(env_cpu(env), false);
+ CPUSPARCState *env = cpu_env(cs);
+ uint32_t flags = cpu_mmu_index(cs, false);
+
#ifndef CONFIG_USER_ONLY
if (cpu_supervisor_mode(env)) {
flags |= TB_FLAG_SUPER;
@@ -751,7 +749,12 @@ void cpu_get_tb_cpu_state(CPUSPARCState *env, vaddr *pc,
}
#endif /* !CONFIG_USER_ONLY */
#endif /* TARGET_SPARC64 */
- *pflags = flags;
+
+ return (TCGTBCPUState){
+ .pc = env->pc,
+ .flags = flags,
+ .flags2 = env->npc,
+ };
}
static void sparc_restore_state_to_opc(CPUState *cs,
diff --git a/target/tricore/cpu.c b/target/tricore/cpu.c
index 8538066114..fdc2012d16 100644
--- a/target/tricore/cpu.c
+++ b/target/tricore/cpu.c
@@ -45,16 +45,14 @@ static vaddr tricore_cpu_get_pc(CPUState *cs)
return cpu_env(cs)->PC;
}
-void cpu_get_tb_cpu_state(CPUTriCoreState *env, vaddr *pc,
- uint64_t *cs_base, uint32_t *flags)
+TCGTBCPUState cpu_get_tb_cpu_state(CPUState *cs)
{
- uint32_t new_flags = 0;
- *pc = env->PC;
- *cs_base = 0;
+ CPUTriCoreState *env = cpu_env(cs);
- new_flags |= FIELD_DP32(new_flags, TB_FLAGS, PRIV,
- extract32(env->PSW, 10, 2));
- *flags = new_flags;
+ return (TCGTBCPUState){
+ .pc = env->PC,
+ .flags = FIELD_DP32(0, TB_FLAGS, PRIV, extract32(env->PSW, 10, 2)),
+ };
}
static void tricore_cpu_synchronize_from_tb(CPUState *cs,
diff --git a/target/xtensa/cpu.c b/target/xtensa/cpu.c
index fe974e90f7..4a8de9f42c 100644
--- a/target/xtensa/cpu.c
+++ b/target/xtensa/cpu.c
@@ -55,15 +55,15 @@ static vaddr xtensa_cpu_get_pc(CPUState *cs)
return cpu->env.pc;
}
-void cpu_get_tb_cpu_state(CPUXtensaState *env, vaddr *pc,
- uint64_t *cs_base, uint32_t *flags)
+TCGTBCPUState cpu_get_tb_cpu_state(CPUState *cs)
{
- *pc = env->pc;
- *cs_base = 0;
- *flags = 0;
- *flags |= xtensa_get_ring(env);
+ CPUXtensaState *env = cpu_env(cs);
+ uint32_t flags = 0;
+ target_ulong cs_base = 0;
+
+ flags |= xtensa_get_ring(env);
if (env->sregs[PS] & PS_EXCM) {
- *flags |= XTENSA_TBFLAG_EXCM;
+ flags |= XTENSA_TBFLAG_EXCM;
} else if (xtensa_option_enabled(env->config, XTENSA_OPTION_LOOP)) {
target_ulong lend_dist =
env->sregs[LEND] - (env->pc & -(1u << TARGET_PAGE_BITS));
@@ -85,26 +85,26 @@ void cpu_get_tb_cpu_state(CPUXtensaState *env, vaddr *pc,
if (lend_dist < (1u << TARGET_PAGE_BITS) + env->config->max_insn_size) {
target_ulong lbeg_off = env->sregs[LEND] - env->sregs[LBEG];
- *cs_base = lend_dist;
+ cs_base = lend_dist;
if (lbeg_off < 256) {
- *cs_base |= lbeg_off << XTENSA_CSBASE_LBEG_OFF_SHIFT;
+ cs_base |= lbeg_off << XTENSA_CSBASE_LBEG_OFF_SHIFT;
}
}
}
if (xtensa_option_enabled(env->config, XTENSA_OPTION_EXTENDED_L32R) &&
(env->sregs[LITBASE] & 1)) {
- *flags |= XTENSA_TBFLAG_LITBASE;
+ flags |= XTENSA_TBFLAG_LITBASE;
}
if (xtensa_option_enabled(env->config, XTENSA_OPTION_DEBUG)) {
if (xtensa_get_cintlevel(env) < env->config->debug_level) {
- *flags |= XTENSA_TBFLAG_DEBUG;
+ flags |= XTENSA_TBFLAG_DEBUG;
}
if (xtensa_get_cintlevel(env) < env->sregs[ICOUNTLEVEL]) {
- *flags |= XTENSA_TBFLAG_ICOUNT;
+ flags |= XTENSA_TBFLAG_ICOUNT;
}
}
if (xtensa_option_enabled(env->config, XTENSA_OPTION_COPROCESSOR)) {
- *flags |= env->sregs[CPENABLE] << XTENSA_TBFLAG_CPENABLE_SHIFT;
+ flags |= env->sregs[CPENABLE] << XTENSA_TBFLAG_CPENABLE_SHIFT;
}
if (xtensa_option_enabled(env->config, XTENSA_OPTION_WINDOWED_REGISTER) &&
(env->sregs[PS] & (PS_WOE | PS_EXCM)) == PS_WOE) {
@@ -112,15 +112,17 @@ void cpu_get_tb_cpu_state(CPUXtensaState *env, vaddr *pc,
(env->sregs[WINDOW_BASE] + 1);
uint32_t w = ctz32(windowstart | 0x8);
- *flags |= (w << XTENSA_TBFLAG_WINDOW_SHIFT) | XTENSA_TBFLAG_CWOE;
- *flags |= extract32(env->sregs[PS], PS_CALLINC_SHIFT,
+ flags |= (w << XTENSA_TBFLAG_WINDOW_SHIFT) | XTENSA_TBFLAG_CWOE;
+ flags |= extract32(env->sregs[PS], PS_CALLINC_SHIFT,
PS_CALLINC_LEN) << XTENSA_TBFLAG_CALLINC_SHIFT;
} else {
- *flags |= 3 << XTENSA_TBFLAG_WINDOW_SHIFT;
+ flags |= 3 << XTENSA_TBFLAG_WINDOW_SHIFT;
}
if (env->yield_needed) {
- *flags |= XTENSA_TBFLAG_YIELD;
+ flags |= XTENSA_TBFLAG_YIELD;
}
+
+ return (TCGTBCPUState){ .pc = env->pc, .flags = flags, .flags2 = cs_base };
}
static void xtensa_restore_state_to_opc(CPUState *cs,
--
2.43.0
^ permalink raw reply related [flat|nested] 32+ messages in thread* Re: [PATCH 07/10] accel/tcg: Return TCGTBCPUState from cpu_get_tb_cpu_state
2025-04-28 20:10 ` [PATCH 07/10] accel/tcg: Return TCGTBCPUState from cpu_get_tb_cpu_state Richard Henderson
@ 2025-04-28 22:00 ` Pierrick Bouvier
2025-04-28 22:07 ` Richard Henderson
0 siblings, 1 reply; 32+ messages in thread
From: Pierrick Bouvier @ 2025-04-28 22:00 UTC (permalink / raw)
To: Richard Henderson, qemu-devel; +Cc: philmd
On 4/28/25 1:10 PM, Richard Henderson wrote:
> diff --git a/include/accel/tcg/cpu-ops.h b/include/accel/tcg/cpu-ops.h
> index f5e5746976..8dbe79ea7c 100644
> --- a/include/accel/tcg/cpu-ops.h
> +++ b/include/accel/tcg/cpu-ops.h
> @@ -18,8 +18,13 @@
> #include "exec/vaddr.h"
> #include "tcg/tcg-mo.h"
>
> -void cpu_get_tb_cpu_state(CPUArchState *env, vaddr *pc,
> - uint64_t *cs_base, uint32_t *flags);
> +struct TCGTBCPUState {
> + vaddr pc;
> + uint32_t flags;
> + uint64_t flags2;
Could it be named cs_base instead?
flags2 is a little bit generic.
> +};
Else,
Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org>
^ permalink raw reply [flat|nested] 32+ messages in thread* Re: [PATCH 07/10] accel/tcg: Return TCGTBCPUState from cpu_get_tb_cpu_state
2025-04-28 22:00 ` Pierrick Bouvier
@ 2025-04-28 22:07 ` Richard Henderson
2025-04-28 22:26 ` Pierrick Bouvier
0 siblings, 1 reply; 32+ messages in thread
From: Richard Henderson @ 2025-04-28 22:07 UTC (permalink / raw)
To: Pierrick Bouvier, qemu-devel; +Cc: philmd
On 4/28/25 15:00, Pierrick Bouvier wrote:
> On 4/28/25 1:10 PM, Richard Henderson wrote:
>> diff --git a/include/accel/tcg/cpu-ops.h b/include/accel/tcg/cpu-ops.h
>> index f5e5746976..8dbe79ea7c 100644
>> --- a/include/accel/tcg/cpu-ops.h
>> +++ b/include/accel/tcg/cpu-ops.h
>> @@ -18,8 +18,13 @@
>> #include "exec/vaddr.h"
>> #include "tcg/tcg-mo.h"
>> -void cpu_get_tb_cpu_state(CPUArchState *env, vaddr *pc,
>> - uint64_t *cs_base, uint32_t *flags);
>> +struct TCGTBCPUState {
>> + vaddr pc;
>> + uint32_t flags;
>> + uint64_t flags2;
>
> Could it be named cs_base instead?
> flags2 is a little bit generic.
Of course flags2 is generic -- it's only cs_base for x86.
r~
^ permalink raw reply [flat|nested] 32+ messages in thread* Re: [PATCH 07/10] accel/tcg: Return TCGTBCPUState from cpu_get_tb_cpu_state
2025-04-28 22:07 ` Richard Henderson
@ 2025-04-28 22:26 ` Pierrick Bouvier
2025-04-28 22:37 ` Richard Henderson
0 siblings, 1 reply; 32+ messages in thread
From: Pierrick Bouvier @ 2025-04-28 22:26 UTC (permalink / raw)
To: Richard Henderson, qemu-devel; +Cc: philmd
On 4/28/25 3:07 PM, Richard Henderson wrote:
> On 4/28/25 15:00, Pierrick Bouvier wrote:
>> On 4/28/25 1:10 PM, Richard Henderson wrote:
>>> diff --git a/include/accel/tcg/cpu-ops.h b/include/accel/tcg/cpu-ops.h
>>> index f5e5746976..8dbe79ea7c 100644
>>> --- a/include/accel/tcg/cpu-ops.h
>>> +++ b/include/accel/tcg/cpu-ops.h
>>> @@ -18,8 +18,13 @@
>>> #include "exec/vaddr.h"
>>> #include "tcg/tcg-mo.h"
>>> -void cpu_get_tb_cpu_state(CPUArchState *env, vaddr *pc,
>>> - uint64_t *cs_base, uint32_t *flags);
>>> +struct TCGTBCPUState {
>>> + vaddr pc;
>>> + uint32_t flags;
>>> + uint64_t flags2;
>>
>> Could it be named cs_base instead?
>> flags2 is a little bit generic.
>
> Of course flags2 is generic -- it's only cs_base for x86.
>
It seems to be named the same for other architectures as well,
so it's wrong for all other arch too?
My point is that:
tb_gen_code(cpu, s.pc, s.flags2, s.flags, cflags);
is error prone, and it's easy to swap flags and flags2 by mistake.
I don't mind the name, but something more distinct would help.
>
> r~
^ permalink raw reply [flat|nested] 32+ messages in thread* Re: [PATCH 07/10] accel/tcg: Return TCGTBCPUState from cpu_get_tb_cpu_state
2025-04-28 22:26 ` Pierrick Bouvier
@ 2025-04-28 22:37 ` Richard Henderson
2025-04-29 18:55 ` Richard Henderson
0 siblings, 1 reply; 32+ messages in thread
From: Richard Henderson @ 2025-04-28 22:37 UTC (permalink / raw)
To: Pierrick Bouvier, qemu-devel; +Cc: philmd
On 4/28/25 15:26, Pierrick Bouvier wrote:
> On 4/28/25 3:07 PM, Richard Henderson wrote:
>> On 4/28/25 15:00, Pierrick Bouvier wrote:
>>> On 4/28/25 1:10 PM, Richard Henderson wrote:
>>>> diff --git a/include/accel/tcg/cpu-ops.h b/include/accel/tcg/cpu-ops.h
>>>> index f5e5746976..8dbe79ea7c 100644
>>>> --- a/include/accel/tcg/cpu-ops.h
>>>> +++ b/include/accel/tcg/cpu-ops.h
>>>> @@ -18,8 +18,13 @@
>>>> #include "exec/vaddr.h"
>>>> #include "tcg/tcg-mo.h"
>>>> -void cpu_get_tb_cpu_state(CPUArchState *env, vaddr *pc,
>>>> - uint64_t *cs_base, uint32_t *flags);
>>>> +struct TCGTBCPUState {
>>>> + vaddr pc;
>>>> + uint32_t flags;
>>>> + uint64_t flags2;
>>>
>>> Could it be named cs_base instead?
>>> flags2 is a little bit generic.
>>
>> Of course flags2 is generic -- it's only cs_base for x86.
>>
>
> It seems to be named the same for other architectures as well,
> so it's wrong for all other arch too?
>
> My point is that:
> tb_gen_code(cpu, s.pc, s.flags2, s.flags, cflags);
>
> is error prone, and it's easy to swap flags and flags2 by mistake.
> I don't mind the name, but something more distinct would help.
I fully intended to go on and pass in TCGTBCPUState, rather than 3 separate args. I guess
I stopped early with the meson.build change for build twice.
r~
^ permalink raw reply [flat|nested] 32+ messages in thread* Re: [PATCH 07/10] accel/tcg: Return TCGTBCPUState from cpu_get_tb_cpu_state
2025-04-28 22:37 ` Richard Henderson
@ 2025-04-29 18:55 ` Richard Henderson
2025-04-29 19:04 ` Pierrick Bouvier
0 siblings, 1 reply; 32+ messages in thread
From: Richard Henderson @ 2025-04-29 18:55 UTC (permalink / raw)
To: Pierrick Bouvier, qemu-devel; +Cc: philmd
On 4/28/25 15:37, Richard Henderson wrote:
> On 4/28/25 15:26, Pierrick Bouvier wrote:
>> On 4/28/25 3:07 PM, Richard Henderson wrote:
>>> On 4/28/25 15:00, Pierrick Bouvier wrote:
>>>> On 4/28/25 1:10 PM, Richard Henderson wrote:
>>>>> diff --git a/include/accel/tcg/cpu-ops.h b/include/accel/tcg/cpu-ops.h
>>>>> index f5e5746976..8dbe79ea7c 100644
>>>>> --- a/include/accel/tcg/cpu-ops.h
>>>>> +++ b/include/accel/tcg/cpu-ops.h
>>>>> @@ -18,8 +18,13 @@
>>>>> #include "exec/vaddr.h"
>>>>> #include "tcg/tcg-mo.h"
>>>>> -void cpu_get_tb_cpu_state(CPUArchState *env, vaddr *pc,
>>>>> - uint64_t *cs_base, uint32_t *flags);
>>>>> +struct TCGTBCPUState {
>>>>> + vaddr pc;
>>>>> + uint32_t flags;
>>>>> + uint64_t flags2;
>>>>
>>>> Could it be named cs_base instead?
>>>> flags2 is a little bit generic.
>>>
>>> Of course flags2 is generic -- it's only cs_base for x86.
>>>
>>
>> It seems to be named the same for other architectures as well,
>> so it's wrong for all other arch too?
>>
>> My point is that:
>> tb_gen_code(cpu, s.pc, s.flags2, s.flags, cflags);
>>
>> is error prone, and it's easy to swap flags and flags2 by mistake.
>> I don't mind the name, but something more distinct would help.
>
> I fully intended to go on and pass in TCGTBCPUState, rather than 3 separate args. I guess
> I stopped early with the meson.build change for build twice.
You do have a point that it's still named "cs_base" within TranslationBlock. I'll keep
the cs_base name for now and rename it all at once at a later date.
r~
^ permalink raw reply [flat|nested] 32+ messages in thread* Re: [PATCH 07/10] accel/tcg: Return TCGTBCPUState from cpu_get_tb_cpu_state
2025-04-29 18:55 ` Richard Henderson
@ 2025-04-29 19:04 ` Pierrick Bouvier
0 siblings, 0 replies; 32+ messages in thread
From: Pierrick Bouvier @ 2025-04-29 19:04 UTC (permalink / raw)
To: Richard Henderson, qemu-devel; +Cc: philmd
On 4/29/25 11:55 AM, Richard Henderson wrote:
> On 4/28/25 15:37, Richard Henderson wrote:
>> On 4/28/25 15:26, Pierrick Bouvier wrote:
>>> On 4/28/25 3:07 PM, Richard Henderson wrote:
>>>> On 4/28/25 15:00, Pierrick Bouvier wrote:
>>>>> On 4/28/25 1:10 PM, Richard Henderson wrote:
>>>>>> diff --git a/include/accel/tcg/cpu-ops.h b/include/accel/tcg/cpu-ops.h
>>>>>> index f5e5746976..8dbe79ea7c 100644
>>>>>> --- a/include/accel/tcg/cpu-ops.h
>>>>>> +++ b/include/accel/tcg/cpu-ops.h
>>>>>> @@ -18,8 +18,13 @@
>>>>>> #include "exec/vaddr.h"
>>>>>> #include "tcg/tcg-mo.h"
>>>>>> -void cpu_get_tb_cpu_state(CPUArchState *env, vaddr *pc,
>>>>>> - uint64_t *cs_base, uint32_t *flags);
>>>>>> +struct TCGTBCPUState {
>>>>>> + vaddr pc;
>>>>>> + uint32_t flags;
>>>>>> + uint64_t flags2;
>>>>>
>>>>> Could it be named cs_base instead?
>>>>> flags2 is a little bit generic.
>>>>
>>>> Of course flags2 is generic -- it's only cs_base for x86.
>>>>
>>>
>>> It seems to be named the same for other architectures as well,
>>> so it's wrong for all other arch too?
>>>
>>> My point is that:
>>> tb_gen_code(cpu, s.pc, s.flags2, s.flags, cflags);
>>>
>>> is error prone, and it's easy to swap flags and flags2 by mistake.
>>> I don't mind the name, but something more distinct would help.
>>
>> I fully intended to go on and pass in TCGTBCPUState, rather than 3 separate args. I guess
>> I stopped early with the meson.build change for build twice.
>
> You do have a point that it's still named "cs_base" within TranslationBlock. I'll keep
> the cs_base name for now and rename it all at once at a later date.
>
Sounds good.
>
> r~
^ permalink raw reply [flat|nested] 32+ messages in thread
* [PATCH 08/10] accel/tcg: Move cpu_get_tb_cpu_state to TCGCPUOps
2025-04-28 20:10 [PATCH 00/10] accel/tcg: Compile cpu-exec.c twice Richard Henderson
` (6 preceding siblings ...)
2025-04-28 20:10 ` [PATCH 07/10] accel/tcg: Return TCGTBCPUState from cpu_get_tb_cpu_state Richard Henderson
@ 2025-04-28 20:10 ` Richard Henderson
2025-04-28 21:49 ` Pierrick Bouvier
2025-04-28 20:10 ` [PATCH 09/10] accel/tcg: Split out accel/tcg/helper-retaddr.h Richard Henderson
2025-04-28 20:10 ` [PATCH 10/10] accel/tcg: Compile cpu-exec.c twice Richard Henderson
9 siblings, 1 reply; 32+ messages in thread
From: Richard Henderson @ 2025-04-28 20:10 UTC (permalink / raw)
To: qemu-devel; +Cc: pierrick.bouvier, philmd
Move the global function name to a hook on TCGCPUOps.
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
---
accel/tcg/internal-common.h | 2 +-
include/accel/tcg/cpu-ops.h | 8 ++++++--
target/arm/internals.h | 1 +
target/ppc/internal.h | 2 ++
accel/tcg/cpu-exec.c | 1 +
target/alpha/cpu.c | 3 ++-
target/arm/cpu.c | 1 +
target/arm/helper.c | 2 +-
target/arm/tcg/cpu-v7m.c | 1 +
target/avr/cpu.c | 3 ++-
target/hexagon/cpu.c | 3 ++-
target/hppa/cpu.c | 3 ++-
target/i386/tcg/tcg-cpu.c | 3 ++-
target/loongarch/cpu.c | 3 ++-
target/m68k/cpu.c | 3 ++-
target/microblaze/cpu.c | 3 ++-
target/mips/cpu.c | 3 ++-
target/openrisc/cpu.c | 3 ++-
target/ppc/cpu_init.c | 2 +-
target/ppc/helper_regs.c | 3 ++-
target/riscv/tcg/tcg-cpu.c | 3 ++-
target/rx/cpu.c | 3 ++-
target/s390x/cpu.c | 3 ++-
target/sh4/cpu.c | 3 ++-
target/sparc/cpu.c | 3 ++-
target/tricore/cpu.c | 3 ++-
target/xtensa/cpu.c | 3 ++-
27 files changed, 51 insertions(+), 23 deletions(-)
diff --git a/accel/tcg/internal-common.h b/accel/tcg/internal-common.h
index 3ff4c1cb54..130af3098c 100644
--- a/accel/tcg/internal-common.h
+++ b/accel/tcg/internal-common.h
@@ -59,7 +59,7 @@ void cpu_restore_state_from_tb(CPUState *cpu, TranslationBlock *tb,
static inline TCGTBCPUState get_tb_cpu_state(CPUState *cs)
{
- return cpu_get_tb_cpu_state(cs);
+ return cs->cc->tcg_ops->get_tb_cpu_state(cs);
}
/**
diff --git a/include/accel/tcg/cpu-ops.h b/include/accel/tcg/cpu-ops.h
index 8dbe79ea7c..25986a351a 100644
--- a/include/accel/tcg/cpu-ops.h
+++ b/include/accel/tcg/cpu-ops.h
@@ -24,8 +24,6 @@ struct TCGTBCPUState {
uint64_t flags2;
};
-TCGTBCPUState cpu_get_tb_cpu_state(CPUState *cs);
-
struct TCGCPUOps {
/**
* mttcg_supported: multi-threaded TCG is supported
@@ -68,6 +66,12 @@ struct TCGCPUOps {
*/
void (*translate_code)(CPUState *cpu, TranslationBlock *tb,
int *max_insns, vaddr pc, void *host_pc);
+ /**
+ * @get_tb_cpu_state: Extract CPU state for a TCG #TranslationBlock
+ *
+ * Fill in all data required to select or compile a TranslationBlock.
+ */
+ TCGTBCPUState (*get_tb_cpu_state)(CPUState *cs);
/**
* @synchronize_from_tb: Synchronize state from a TCG #TranslationBlock
*
diff --git a/target/arm/internals.h b/target/arm/internals.h
index d24acdd672..10f523873a 100644
--- a/target/arm/internals.h
+++ b/target/arm/internals.h
@@ -370,6 +370,7 @@ void arm_restore_state_to_opc(CPUState *cs,
const uint64_t *data);
#ifdef CONFIG_TCG
+TCGTBCPUState arm_get_tb_cpu_state(CPUState *cs);
void arm_cpu_synchronize_from_tb(CPUState *cs, const TranslationBlock *tb);
/* Our implementation of TCGCPUOps::cpu_exec_halt */
diff --git a/target/ppc/internal.h b/target/ppc/internal.h
index 9012d3809c..681d522c2e 100644
--- a/target/ppc/internal.h
+++ b/target/ppc/internal.h
@@ -308,4 +308,6 @@ static inline int ger_pack_masks(int pmsk, int ymsk, int xmsk)
return msk;
}
+TCGTBCPUState ppc_get_tb_cpu_state(CPUState *cs);
+
#endif /* PPC_INTERNAL_H */
diff --git a/accel/tcg/cpu-exec.c b/accel/tcg/cpu-exec.c
index a48df5f291..a10a2c5f29 100644
--- a/accel/tcg/cpu-exec.c
+++ b/accel/tcg/cpu-exec.c
@@ -1052,6 +1052,7 @@ bool tcg_exec_realizefn(CPUState *cpu, Error **errp)
assert(tcg_ops->cpu_exec_reset);
#endif /* !CONFIG_USER_ONLY */
assert(tcg_ops->translate_code);
+ assert(tcg_ops->get_tb_cpu_state);
assert(tcg_ops->mmu_index);
tcg_ops->initialize();
tcg_target_initialized = true;
diff --git a/target/alpha/cpu.c b/target/alpha/cpu.c
index 6efaa71543..695e6f42a6 100644
--- a/target/alpha/cpu.c
+++ b/target/alpha/cpu.c
@@ -41,7 +41,7 @@ static vaddr alpha_cpu_get_pc(CPUState *cs)
return env->pc;
}
-TCGTBCPUState cpu_get_tb_cpu_state(CPUState *cs)
+static TCGTBCPUState alpha_get_tb_cpu_state(CPUState *cs)
{
CPUAlphaState *env = cpu_env(cs);
uint32_t flags = env->flags & ENV_FLAG_TB_MASK;
@@ -251,6 +251,7 @@ static const TCGCPUOps alpha_tcg_ops = {
.initialize = alpha_translate_init,
.translate_code = alpha_translate_code,
+ .get_tb_cpu_state = alpha_get_tb_cpu_state,
.synchronize_from_tb = alpha_cpu_synchronize_from_tb,
.restore_state_to_opc = alpha_restore_state_to_opc,
.mmu_index = alpha_cpu_mmu_index,
diff --git a/target/arm/cpu.c b/target/arm/cpu.c
index 9e74d7bc07..c401f0455b 100644
--- a/target/arm/cpu.c
+++ b/target/arm/cpu.c
@@ -2676,6 +2676,7 @@ static const TCGCPUOps arm_tcg_ops = {
.initialize = arm_translate_init,
.translate_code = arm_translate_code,
+ .get_tb_cpu_state = arm_get_tb_cpu_state,
.synchronize_from_tb = arm_cpu_synchronize_from_tb,
.debug_excp_handler = arm_debug_excp_handler,
.restore_state_to_opc = arm_restore_state_to_opc,
diff --git a/target/arm/helper.c b/target/arm/helper.c
index aa887af50f..84f1d963c2 100644
--- a/target/arm/helper.c
+++ b/target/arm/helper.c
@@ -11453,7 +11453,7 @@ static bool mve_no_pred(CPUARMState *env)
return true;
}
-TCGTBCPUState cpu_get_tb_cpu_state(CPUState *cs)
+TCGTBCPUState arm_get_tb_cpu_state(CPUState *cs)
{
CPUARMState *env = cpu_env(cs);
CPUARMTBFlags flags;
diff --git a/target/arm/tcg/cpu-v7m.c b/target/arm/tcg/cpu-v7m.c
index 57ed3f3a06..ef7e16dfcd 100644
--- a/target/arm/tcg/cpu-v7m.c
+++ b/target/arm/tcg/cpu-v7m.c
@@ -238,6 +238,7 @@ static const TCGCPUOps arm_v7m_tcg_ops = {
.initialize = arm_translate_init,
.translate_code = arm_translate_code,
+ .get_tb_cpu_state = arm_get_tb_cpu_state,
.synchronize_from_tb = arm_cpu_synchronize_from_tb,
.debug_excp_handler = arm_debug_excp_handler,
.restore_state_to_opc = arm_restore_state_to_opc,
diff --git a/target/avr/cpu.c b/target/avr/cpu.c
index 88a4b970df..47b5334906 100644
--- a/target/avr/cpu.c
+++ b/target/avr/cpu.c
@@ -54,7 +54,7 @@ static int avr_cpu_mmu_index(CPUState *cs, bool ifetch)
return ifetch ? MMU_CODE_IDX : MMU_DATA_IDX;
}
-TCGTBCPUState cpu_get_tb_cpu_state(CPUState *cs)
+static TCGTBCPUState avr_get_tb_cpu_state(CPUState *cs)
{
CPUAVRState *env = cpu_env(cs);
uint32_t flags = 0;
@@ -241,6 +241,7 @@ static const TCGCPUOps avr_tcg_ops = {
.mttcg_supported = false,
.initialize = avr_cpu_tcg_init,
.translate_code = avr_cpu_translate_code,
+ .get_tb_cpu_state = avr_get_tb_cpu_state,
.synchronize_from_tb = avr_cpu_synchronize_from_tb,
.restore_state_to_opc = avr_restore_state_to_opc,
.mmu_index = avr_cpu_mmu_index,
diff --git a/target/hexagon/cpu.c b/target/hexagon/cpu.c
index 08c9aef55e..9beb6a2dc2 100644
--- a/target/hexagon/cpu.c
+++ b/target/hexagon/cpu.c
@@ -255,7 +255,7 @@ static vaddr hexagon_cpu_get_pc(CPUState *cs)
return cpu_env(cs)->gpr[HEX_REG_PC];
}
-TCGTBCPUState cpu_get_tb_cpu_state(CPUState *cs)
+static TCGTBCPUState hexagon_get_tb_cpu_state(CPUState *cs)
{
CPUHexagonState *env = cpu_env(cs);
vaddr pc = env->gpr[HEX_REG_PC];
@@ -344,6 +344,7 @@ static const TCGCPUOps hexagon_tcg_ops = {
.mttcg_supported = false,
.initialize = hexagon_translate_init,
.translate_code = hexagon_translate_code,
+ .get_tb_cpu_state = hexagon_get_tb_cpu_state,
.synchronize_from_tb = hexagon_cpu_synchronize_from_tb,
.restore_state_to_opc = hexagon_restore_state_to_opc,
.mmu_index = hexagon_cpu_mmu_index,
diff --git a/target/hppa/cpu.c b/target/hppa/cpu.c
index 7270b0a2a7..58566d9897 100644
--- a/target/hppa/cpu.c
+++ b/target/hppa/cpu.c
@@ -51,7 +51,7 @@ static vaddr hppa_cpu_get_pc(CPUState *cs)
env->iaoq_f & -4);
}
-TCGTBCPUState cpu_get_tb_cpu_state(CPUState *cs)
+static TCGTBCPUState hppa_get_tb_cpu_state(CPUState *cs)
{
CPUHPPAState *env = cpu_env(cs);
uint32_t flags = 0;
@@ -262,6 +262,7 @@ static const TCGCPUOps hppa_tcg_ops = {
.initialize = hppa_translate_init,
.translate_code = hppa_translate_code,
+ .get_tb_cpu_state = hppa_get_tb_cpu_state,
.synchronize_from_tb = hppa_cpu_synchronize_from_tb,
.restore_state_to_opc = hppa_restore_state_to_opc,
.mmu_index = hppa_cpu_mmu_index,
diff --git a/target/i386/tcg/tcg-cpu.c b/target/i386/tcg/tcg-cpu.c
index 950b3bfa76..2845e5b0cc 100644
--- a/target/i386/tcg/tcg-cpu.c
+++ b/target/i386/tcg/tcg-cpu.c
@@ -48,7 +48,7 @@ static void x86_cpu_exec_exit(CPUState *cs)
env->eflags = cpu_compute_eflags(env);
}
-TCGTBCPUState cpu_get_tb_cpu_state(CPUState *cs)
+static TCGTBCPUState x86_get_tb_cpu_state(CPUState *cs)
{
CPUX86State *env = cpu_env(cs);
uint32_t flags, cs_base;
@@ -160,6 +160,7 @@ const TCGCPUOps x86_tcg_ops = {
.guest_default_memory_order = TCG_MO_ALL & ~TCG_MO_ST_LD,
.initialize = tcg_x86_init,
.translate_code = x86_translate_code,
+ .get_tb_cpu_state = x86_get_tb_cpu_state,
.synchronize_from_tb = x86_cpu_synchronize_from_tb,
.restore_state_to_opc = x86_restore_state_to_opc,
.mmu_index = x86_cpu_mmu_index,
diff --git a/target/loongarch/cpu.c b/target/loongarch/cpu.c
index 1691fe1d9f..4ea17a288b 100644
--- a/target/loongarch/cpu.c
+++ b/target/loongarch/cpu.c
@@ -336,7 +336,7 @@ static bool loongarch_cpu_exec_interrupt(CPUState *cs, int interrupt_request)
}
#endif
-TCGTBCPUState cpu_get_tb_cpu_state(CPUState *cs)
+static TCGTBCPUState loongarch_get_tb_cpu_state(CPUState *cs)
{
CPULoongArchState *env = cpu_env(cs);
uint32_t flags;
@@ -882,6 +882,7 @@ static const TCGCPUOps loongarch_tcg_ops = {
.initialize = loongarch_translate_init,
.translate_code = loongarch_translate_code,
+ .get_tb_cpu_state = loongarch_get_tb_cpu_state,
.synchronize_from_tb = loongarch_cpu_synchronize_from_tb,
.restore_state_to_opc = loongarch_restore_state_to_opc,
.mmu_index = loongarch_cpu_mmu_index,
diff --git a/target/m68k/cpu.c b/target/m68k/cpu.c
index 1786d9e0ce..75a860cf6f 100644
--- a/target/m68k/cpu.c
+++ b/target/m68k/cpu.c
@@ -39,7 +39,7 @@ static vaddr m68k_cpu_get_pc(CPUState *cs)
return cpu->env.pc;
}
-TCGTBCPUState cpu_get_tb_cpu_state(CPUState *cs)
+static TCGTBCPUState m68k_get_tb_cpu_state(CPUState *cs)
{
CPUM68KState *env = cpu_env(cs);
uint32_t flags;
@@ -612,6 +612,7 @@ static const TCGCPUOps m68k_tcg_ops = {
.initialize = m68k_tcg_init,
.translate_code = m68k_translate_code,
+ .get_tb_cpu_state = m68k_get_tb_cpu_state,
.restore_state_to_opc = m68k_restore_state_to_opc,
.mmu_index = m68k_cpu_mmu_index,
diff --git a/target/microblaze/cpu.c b/target/microblaze/cpu.c
index 91e77a4e53..6a4f93e99c 100644
--- a/target/microblaze/cpu.c
+++ b/target/microblaze/cpu.c
@@ -95,7 +95,7 @@ static vaddr mb_cpu_get_pc(CPUState *cs)
return cpu->env.pc;
}
-TCGTBCPUState cpu_get_tb_cpu_state(CPUState *cs)
+static TCGTBCPUState mb_get_tb_cpu_state(CPUState *cs)
{
CPUMBState *env = cpu_env(cs);
@@ -442,6 +442,7 @@ static const TCGCPUOps mb_tcg_ops = {
.initialize = mb_tcg_init,
.translate_code = mb_translate_code,
+ .get_tb_cpu_state = mb_get_tb_cpu_state,
.synchronize_from_tb = mb_cpu_synchronize_from_tb,
.restore_state_to_opc = mb_restore_state_to_opc,
.mmu_index = mb_cpu_mmu_index,
diff --git a/target/mips/cpu.c b/target/mips/cpu.c
index c514202fbc..a47a96f060 100644
--- a/target/mips/cpu.c
+++ b/target/mips/cpu.c
@@ -549,7 +549,7 @@ static int mips_cpu_mmu_index(CPUState *cs, bool ifunc)
return mips_env_mmu_index(cpu_env(cs));
}
-TCGTBCPUState cpu_get_tb_cpu_state(CPUState *cs)
+static TCGTBCPUState mips_get_tb_cpu_state(CPUState *cs)
{
CPUMIPSState *env = cpu_env(cs);
@@ -566,6 +566,7 @@ static const TCGCPUOps mips_tcg_ops = {
.initialize = mips_tcg_init,
.translate_code = mips_translate_code,
+ .get_tb_cpu_state = mips_get_tb_cpu_state,
.synchronize_from_tb = mips_cpu_synchronize_from_tb,
.restore_state_to_opc = mips_restore_state_to_opc,
.mmu_index = mips_cpu_mmu_index,
diff --git a/target/openrisc/cpu.c b/target/openrisc/cpu.c
index 6b31f9c80e..c0cd279b1e 100644
--- a/target/openrisc/cpu.c
+++ b/target/openrisc/cpu.c
@@ -41,7 +41,7 @@ static vaddr openrisc_cpu_get_pc(CPUState *cs)
return cpu->env.pc;
}
-TCGTBCPUState cpu_get_tb_cpu_state(CPUState *cs)
+static TCGTBCPUState openrisc_get_tb_cpu_state(CPUState *cs)
{
CPUOpenRISCState *env = cpu_env(cs);
@@ -258,6 +258,7 @@ static const TCGCPUOps openrisc_tcg_ops = {
.initialize = openrisc_translate_init,
.translate_code = openrisc_translate_code,
+ .get_tb_cpu_state = openrisc_get_tb_cpu_state,
.synchronize_from_tb = openrisc_cpu_synchronize_from_tb,
.restore_state_to_opc = openrisc_restore_state_to_opc,
.mmu_index = openrisc_cpu_mmu_index,
diff --git a/target/ppc/cpu_init.c b/target/ppc/cpu_init.c
index 5981f607d2..fcd96e369e 100644
--- a/target/ppc/cpu_init.c
+++ b/target/ppc/cpu_init.c
@@ -45,7 +45,6 @@
#include "internal.h"
#include "spr_common.h"
#include "power8-pmu.h"
-
#ifndef CONFIG_USER_ONLY
#include "hw/boards.h"
#include "hw/intc/intc.h"
@@ -7483,6 +7482,7 @@ static const TCGCPUOps ppc_tcg_ops = {
.guest_default_memory_order = 0,
.initialize = ppc_translate_init,
.translate_code = ppc_translate_code,
+ .get_tb_cpu_state = ppc_get_tb_cpu_state,
.restore_state_to_opc = ppc_restore_state_to_opc,
.mmu_index = ppc_cpu_mmu_index,
diff --git a/target/ppc/helper_regs.c b/target/ppc/helper_regs.c
index ccaf2b0343..7e5726871e 100644
--- a/target/ppc/helper_regs.c
+++ b/target/ppc/helper_regs.c
@@ -28,6 +28,7 @@
#include "cpu-models.h"
#include "spr_common.h"
#include "accel/tcg/cpu-ops.h"
+#include "internal.h"
/* Swap temporary saved registers with GPRs */
void hreg_swap_gpr_tgpr(CPUPPCState *env)
@@ -256,7 +257,7 @@ void hreg_update_pmu_hflags(CPUPPCState *env)
env->hflags |= hreg_compute_pmu_hflags_value(env);
}
-TCGTBCPUState cpu_get_tb_cpu_state(CPUState *cs)
+TCGTBCPUState ppc_get_tb_cpu_state(CPUState *cs)
{
CPUPPCState *env = cpu_env(cs);
uint32_t hflags_current = env->hflags;
diff --git a/target/riscv/tcg/tcg-cpu.c b/target/riscv/tcg/tcg-cpu.c
index 461b8c97b4..4ab8f021eb 100644
--- a/target/riscv/tcg/tcg-cpu.c
+++ b/target/riscv/tcg/tcg-cpu.c
@@ -98,7 +98,7 @@ static int riscv_cpu_mmu_index(CPUState *cs, bool ifetch)
return riscv_env_mmu_index(cpu_env(cs), ifetch);
}
-TCGTBCPUState cpu_get_tb_cpu_state(CPUState *cs)
+static TCGTBCPUState riscv_get_tb_cpu_state(CPUState *cs)
{
CPURISCVState *env = cpu_env(cs);
RISCVCPU *cpu = env_archcpu(env);
@@ -243,6 +243,7 @@ const TCGCPUOps riscv_tcg_ops = {
.initialize = riscv_translate_init,
.translate_code = riscv_translate_code,
+ .get_tb_cpu_state = riscv_get_tb_cpu_state,
.synchronize_from_tb = riscv_cpu_synchronize_from_tb,
.restore_state_to_opc = riscv_restore_state_to_opc,
.mmu_index = riscv_cpu_mmu_index,
diff --git a/target/rx/cpu.c b/target/rx/cpu.c
index b8bba1a0c7..786c9d3031 100644
--- a/target/rx/cpu.c
+++ b/target/rx/cpu.c
@@ -44,7 +44,7 @@ static vaddr rx_cpu_get_pc(CPUState *cs)
return cpu->env.pc;
}
-TCGTBCPUState cpu_get_tb_cpu_state(CPUState *cs)
+static TCGTBCPUState rx_get_tb_cpu_state(CPUState *cs)
{
CPURXState *env = cpu_env(cs);
uint32_t flags = 0;
@@ -220,6 +220,7 @@ static const TCGCPUOps rx_tcg_ops = {
.initialize = rx_translate_init,
.translate_code = rx_translate_code,
+ .get_tb_cpu_state = rx_get_tb_cpu_state,
.synchronize_from_tb = rx_cpu_synchronize_from_tb,
.restore_state_to_opc = rx_restore_state_to_opc,
.mmu_index = rx_cpu_mmu_index,
diff --git a/target/s390x/cpu.c b/target/s390x/cpu.c
index 127574c085..65330d8ae5 100644
--- a/target/s390x/cpu.c
+++ b/target/s390x/cpu.c
@@ -309,7 +309,7 @@ static int s390x_cpu_mmu_index(CPUState *cs, bool ifetch)
return s390x_env_mmu_index(cpu_env(cs), ifetch);
}
-TCGTBCPUState cpu_get_tb_cpu_state(CPUState *cs)
+static TCGTBCPUState s390x_get_tb_cpu_state(CPUState *cs)
{
CPUS390XState *env = cpu_env(cs);
uint32_t flags;
@@ -358,6 +358,7 @@ static const TCGCPUOps s390_tcg_ops = {
.initialize = s390x_translate_init,
.translate_code = s390x_translate_code,
+ .get_tb_cpu_state = s390x_get_tb_cpu_state,
.restore_state_to_opc = s390x_restore_state_to_opc,
.mmu_index = s390x_cpu_mmu_index,
diff --git a/target/sh4/cpu.c b/target/sh4/cpu.c
index 320cb8ddcc..20969037d1 100644
--- a/target/sh4/cpu.c
+++ b/target/sh4/cpu.c
@@ -43,7 +43,7 @@ static vaddr superh_cpu_get_pc(CPUState *cs)
return cpu->env.pc;
}
-TCGTBCPUState cpu_get_tb_cpu_state(CPUState *cs)
+static TCGTBCPUState superh_get_tb_cpu_state(CPUState *cs)
{
CPUSH4State *env = cpu_env(cs);
uint32_t flags;
@@ -289,6 +289,7 @@ static const TCGCPUOps superh_tcg_ops = {
.initialize = sh4_translate_init,
.translate_code = sh4_translate_code,
+ .get_tb_cpu_state = superh_get_tb_cpu_state,
.synchronize_from_tb = superh_cpu_synchronize_from_tb,
.restore_state_to_opc = superh_restore_state_to_opc,
.mmu_index = sh4_cpu_mmu_index,
diff --git a/target/sparc/cpu.c b/target/sparc/cpu.c
index 77591e84ba..05805ee127 100644
--- a/target/sparc/cpu.c
+++ b/target/sparc/cpu.c
@@ -716,7 +716,7 @@ static void sparc_cpu_synchronize_from_tb(CPUState *cs,
cpu->env.npc = tb->cs_base;
}
-TCGTBCPUState cpu_get_tb_cpu_state(CPUState *cs)
+static TCGTBCPUState sparc_get_tb_cpu_state(CPUState *cs)
{
CPUSPARCState *env = cpu_env(cs);
uint32_t flags = cpu_mmu_index(cs, false);
@@ -1029,6 +1029,7 @@ static const TCGCPUOps sparc_tcg_ops = {
.initialize = sparc_tcg_init,
.translate_code = sparc_translate_code,
+ .get_tb_cpu_state = sparc_get_tb_cpu_state,
.synchronize_from_tb = sparc_cpu_synchronize_from_tb,
.restore_state_to_opc = sparc_restore_state_to_opc,
.mmu_index = sparc_cpu_mmu_index,
diff --git a/target/tricore/cpu.c b/target/tricore/cpu.c
index fdc2012d16..15e9d0eb1a 100644
--- a/target/tricore/cpu.c
+++ b/target/tricore/cpu.c
@@ -45,7 +45,7 @@ static vaddr tricore_cpu_get_pc(CPUState *cs)
return cpu_env(cs)->PC;
}
-TCGTBCPUState cpu_get_tb_cpu_state(CPUState *cs)
+static TCGTBCPUState tricore_get_tb_cpu_state(CPUState *cs)
{
CPUTriCoreState *env = cpu_env(cs);
@@ -185,6 +185,7 @@ static const TCGCPUOps tricore_tcg_ops = {
.mttcg_supported = false,
.initialize = tricore_tcg_init,
.translate_code = tricore_translate_code,
+ .get_tb_cpu_state = tricore_get_tb_cpu_state,
.synchronize_from_tb = tricore_cpu_synchronize_from_tb,
.restore_state_to_opc = tricore_restore_state_to_opc,
.mmu_index = tricore_cpu_mmu_index,
diff --git a/target/xtensa/cpu.c b/target/xtensa/cpu.c
index 4a8de9f42c..6850a0e508 100644
--- a/target/xtensa/cpu.c
+++ b/target/xtensa/cpu.c
@@ -55,7 +55,7 @@ static vaddr xtensa_cpu_get_pc(CPUState *cs)
return cpu->env.pc;
}
-TCGTBCPUState cpu_get_tb_cpu_state(CPUState *cs)
+static TCGTBCPUState xtensa_get_tb_cpu_state(CPUState *cs)
{
CPUXtensaState *env = cpu_env(cs);
uint32_t flags = 0;
@@ -308,6 +308,7 @@ static const TCGCPUOps xtensa_tcg_ops = {
.initialize = xtensa_translate_init,
.translate_code = xtensa_translate_code,
.debug_excp_handler = xtensa_breakpoint_handler,
+ .get_tb_cpu_state = xtensa_get_tb_cpu_state,
.restore_state_to_opc = xtensa_restore_state_to_opc,
.mmu_index = xtensa_cpu_mmu_index,
--
2.43.0
^ permalink raw reply related [flat|nested] 32+ messages in thread* Re: [PATCH 08/10] accel/tcg: Move cpu_get_tb_cpu_state to TCGCPUOps
2025-04-28 20:10 ` [PATCH 08/10] accel/tcg: Move cpu_get_tb_cpu_state to TCGCPUOps Richard Henderson
@ 2025-04-28 21:49 ` Pierrick Bouvier
0 siblings, 0 replies; 32+ messages in thread
From: Pierrick Bouvier @ 2025-04-28 21:49 UTC (permalink / raw)
To: Richard Henderson, qemu-devel; +Cc: philmd
On 4/28/25 1:10 PM, Richard Henderson wrote:
> Move the global function name to a hook on TCGCPUOps.
>
> Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
> ---
> accel/tcg/internal-common.h | 2 +-
> include/accel/tcg/cpu-ops.h | 8 ++++++--
> target/arm/internals.h | 1 +
> target/ppc/internal.h | 2 ++
> accel/tcg/cpu-exec.c | 1 +
> target/alpha/cpu.c | 3 ++-
> target/arm/cpu.c | 1 +
> target/arm/helper.c | 2 +-
> target/arm/tcg/cpu-v7m.c | 1 +
> target/avr/cpu.c | 3 ++-
> target/hexagon/cpu.c | 3 ++-
> target/hppa/cpu.c | 3 ++-
> target/i386/tcg/tcg-cpu.c | 3 ++-
> target/loongarch/cpu.c | 3 ++-
> target/m68k/cpu.c | 3 ++-
> target/microblaze/cpu.c | 3 ++-
> target/mips/cpu.c | 3 ++-
> target/openrisc/cpu.c | 3 ++-
> target/ppc/cpu_init.c | 2 +-
> target/ppc/helper_regs.c | 3 ++-
> target/riscv/tcg/tcg-cpu.c | 3 ++-
> target/rx/cpu.c | 3 ++-
> target/s390x/cpu.c | 3 ++-
> target/sh4/cpu.c | 3 ++-
> target/sparc/cpu.c | 3 ++-
> target/tricore/cpu.c | 3 ++-
> target/xtensa/cpu.c | 3 ++-
> 27 files changed, 51 insertions(+), 23 deletions(-)
Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org>
^ permalink raw reply [flat|nested] 32+ messages in thread
* [PATCH 09/10] accel/tcg: Split out accel/tcg/helper-retaddr.h
2025-04-28 20:10 [PATCH 00/10] accel/tcg: Compile cpu-exec.c twice Richard Henderson
` (7 preceding siblings ...)
2025-04-28 20:10 ` [PATCH 08/10] accel/tcg: Move cpu_get_tb_cpu_state to TCGCPUOps Richard Henderson
@ 2025-04-28 20:10 ` Richard Henderson
2025-04-28 21:48 ` Pierrick Bouvier
2025-04-29 6:41 ` Philippe Mathieu-Daudé
2025-04-28 20:10 ` [PATCH 10/10] accel/tcg: Compile cpu-exec.c twice Richard Henderson
9 siblings, 2 replies; 32+ messages in thread
From: Richard Henderson @ 2025-04-28 20:10 UTC (permalink / raw)
To: qemu-devel; +Cc: pierrick.bouvier, philmd
Move set_helper_retaddr and clear_helper_retaddr
to a new header file.
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
---
include/accel/tcg/cpu-ldst.h | 34 -----------------------
include/accel/tcg/helper-retaddr.h | 43 ++++++++++++++++++++++++++++++
accel/tcg/cpu-exec.c | 1 +
accel/tcg/user-exec.c | 1 +
target/arm/tcg/helper-a64.c | 1 +
target/arm/tcg/sme_helper.c | 1 +
target/arm/tcg/sve_helper.c | 1 +
target/ppc/mem_helper.c | 1 +
target/s390x/tcg/mem_helper.c | 1 +
9 files changed, 50 insertions(+), 34 deletions(-)
create mode 100644 include/accel/tcg/helper-retaddr.h
diff --git a/include/accel/tcg/cpu-ldst.h b/include/accel/tcg/cpu-ldst.h
index f97a730703..44a62b54da 100644
--- a/include/accel/tcg/cpu-ldst.h
+++ b/include/accel/tcg/cpu-ldst.h
@@ -526,38 +526,4 @@ void *tlb_vaddr_to_host(CPUArchState *env, vaddr addr,
MMUAccessType access_type, int mmu_idx);
#endif
-/*
- * For user-only, helpers that use guest to host address translation
- * must protect the actual host memory access by recording 'retaddr'
- * for the signal handler. This is required for a race condition in
- * which another thread unmaps the page between a probe and the
- * actual access.
- */
-#ifdef CONFIG_USER_ONLY
-extern __thread uintptr_t helper_retaddr;
-
-static inline void set_helper_retaddr(uintptr_t ra)
-{
- helper_retaddr = ra;
- /*
- * Ensure that this write is visible to the SIGSEGV handler that
- * may be invoked due to a subsequent invalid memory operation.
- */
- signal_barrier();
-}
-
-static inline void clear_helper_retaddr(void)
-{
- /*
- * Ensure that previous memory operations have succeeded before
- * removing the data visible to the signal handler.
- */
- signal_barrier();
- helper_retaddr = 0;
-}
-#else
-#define set_helper_retaddr(ra) do { } while (0)
-#define clear_helper_retaddr() do { } while (0)
-#endif
-
#endif /* ACCEL_TCG_CPU_LDST_H */
diff --git a/include/accel/tcg/helper-retaddr.h b/include/accel/tcg/helper-retaddr.h
new file mode 100644
index 0000000000..037fda2b83
--- /dev/null
+++ b/include/accel/tcg/helper-retaddr.h
@@ -0,0 +1,43 @@
+/*
+ * Get user helper pc for memory unwinding.
+ * SPDX-License-Identifier: LGPL-2.1-or-later
+ */
+
+#ifndef ACCEL_TCG_HELPER_RETADDR_H
+#define ACCEL_TCG_HELPER_RETADDR_H
+
+/*
+ * For user-only, helpers that use guest to host address translation
+ * must protect the actual host memory access by recording 'retaddr'
+ * for the signal handler. This is required for a race condition in
+ * which another thread unmaps the page between a probe and the
+ * actual access.
+ */
+#ifdef CONFIG_USER_ONLY
+extern __thread uintptr_t helper_retaddr;
+
+static inline void set_helper_retaddr(uintptr_t ra)
+{
+ helper_retaddr = ra;
+ /*
+ * Ensure that this write is visible to the SIGSEGV handler that
+ * may be invoked due to a subsequent invalid memory operation.
+ */
+ signal_barrier();
+}
+
+static inline void clear_helper_retaddr(void)
+{
+ /*
+ * Ensure that previous memory operations have succeeded before
+ * removing the data visible to the signal handler.
+ */
+ signal_barrier();
+ helper_retaddr = 0;
+}
+#else
+#define set_helper_retaddr(ra) do { } while (0)
+#define clear_helper_retaddr() do { } while (0)
+#endif
+
+#endif /* ACCEL_TCG_HELPER_RETADDR_H */
diff --git a/accel/tcg/cpu-exec.c b/accel/tcg/cpu-exec.c
index a10a2c5f29..985abb1269 100644
--- a/accel/tcg/cpu-exec.c
+++ b/accel/tcg/cpu-exec.c
@@ -24,6 +24,7 @@
#include "hw/core/cpu.h"
#include "accel/tcg/cpu-ldst.h"
#include "accel/tcg/cpu-ops.h"
+#include "accel/tcg/helper-retaddr.h"
#include "trace.h"
#include "disas/disas.h"
#include "exec/cpu-common.h"
diff --git a/accel/tcg/user-exec.c b/accel/tcg/user-exec.c
index 70feee8df9..68e01fc584 100644
--- a/accel/tcg/user-exec.c
+++ b/accel/tcg/user-exec.c
@@ -26,6 +26,7 @@
#include "qemu/bitops.h"
#include "qemu/rcu.h"
#include "accel/tcg/cpu-ldst.h"
+#include "accel/tcg/helper-retaddr.h"
#include "accel/tcg/probe.h"
#include "user/cpu_loop.h"
#include "qemu/main-loop.h"
diff --git a/target/arm/tcg/helper-a64.c b/target/arm/tcg/helper-a64.c
index 9cffda07cd..4f618ae390 100644
--- a/target/arm/tcg/helper-a64.c
+++ b/target/arm/tcg/helper-a64.c
@@ -30,6 +30,7 @@
#include "qemu/crc32c.h"
#include "exec/cpu-common.h"
#include "accel/tcg/cpu-ldst.h"
+#include "accel/tcg/helper-retaddr.h"
#include "accel/tcg/probe.h"
#include "exec/target_page.h"
#include "exec/tlb-flags.h"
diff --git a/target/arm/tcg/sme_helper.c b/target/arm/tcg/sme_helper.c
index 3226895cae..de0c6e54d4 100644
--- a/target/arm/tcg/sme_helper.c
+++ b/target/arm/tcg/sme_helper.c
@@ -23,6 +23,7 @@
#include "tcg/tcg-gvec-desc.h"
#include "exec/helper-proto.h"
#include "accel/tcg/cpu-ldst.h"
+#include "accel/tcg/helper-retaddr.h"
#include "qemu/int128.h"
#include "fpu/softfloat.h"
#include "vec_internal.h"
diff --git a/target/arm/tcg/sve_helper.c b/target/arm/tcg/sve_helper.c
index 9f20ecb51d..a2c363a4e1 100644
--- a/target/arm/tcg/sve_helper.c
+++ b/target/arm/tcg/sve_helper.c
@@ -30,6 +30,7 @@
#include "vec_internal.h"
#include "sve_ldst_internal.h"
#include "accel/tcg/cpu-ldst.h"
+#include "accel/tcg/helper-retaddr.h"
#include "accel/tcg/cpu-ops.h"
#include "accel/tcg/probe.h"
#ifdef CONFIG_USER_ONLY
diff --git a/target/ppc/mem_helper.c b/target/ppc/mem_helper.c
index aa1af44d22..6ab71a6fcb 100644
--- a/target/ppc/mem_helper.c
+++ b/target/ppc/mem_helper.c
@@ -24,6 +24,7 @@
#include "exec/helper-proto.h"
#include "helper_regs.h"
#include "accel/tcg/cpu-ldst.h"
+#include "accel/tcg/helper-retaddr.h"
#include "accel/tcg/probe.h"
#include "internal.h"
#include "qemu/atomic128.h"
diff --git a/target/s390x/tcg/mem_helper.c b/target/s390x/tcg/mem_helper.c
index 857005b120..a03609a140 100644
--- a/target/s390x/tcg/mem_helper.c
+++ b/target/s390x/tcg/mem_helper.c
@@ -32,6 +32,7 @@
#include "exec/target_page.h"
#include "exec/tlb-flags.h"
#include "accel/tcg/cpu-ops.h"
+#include "accel/tcg/helper-retaddr.h"
#include "qemu/int128.h"
#include "qemu/atomic128.h"
--
2.43.0
^ permalink raw reply related [flat|nested] 32+ messages in thread* Re: [PATCH 09/10] accel/tcg: Split out accel/tcg/helper-retaddr.h
2025-04-28 20:10 ` [PATCH 09/10] accel/tcg: Split out accel/tcg/helper-retaddr.h Richard Henderson
@ 2025-04-28 21:48 ` Pierrick Bouvier
2025-04-29 6:41 ` Philippe Mathieu-Daudé
1 sibling, 0 replies; 32+ messages in thread
From: Pierrick Bouvier @ 2025-04-28 21:48 UTC (permalink / raw)
To: Richard Henderson, qemu-devel; +Cc: philmd
On 4/28/25 1:10 PM, Richard Henderson wrote:
> Move set_helper_retaddr and clear_helper_retaddr
> to a new header file.
>
> Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
> ---
> include/accel/tcg/cpu-ldst.h | 34 -----------------------
> include/accel/tcg/helper-retaddr.h | 43 ++++++++++++++++++++++++++++++
> accel/tcg/cpu-exec.c | 1 +
> accel/tcg/user-exec.c | 1 +
> target/arm/tcg/helper-a64.c | 1 +
> target/arm/tcg/sme_helper.c | 1 +
> target/arm/tcg/sve_helper.c | 1 +
> target/ppc/mem_helper.c | 1 +
> target/s390x/tcg/mem_helper.c | 1 +
> 9 files changed, 50 insertions(+), 34 deletions(-)
> create mode 100644 include/accel/tcg/helper-retaddr.h
Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org>
^ permalink raw reply [flat|nested] 32+ messages in thread
* Re: [PATCH 09/10] accel/tcg: Split out accel/tcg/helper-retaddr.h
2025-04-28 20:10 ` [PATCH 09/10] accel/tcg: Split out accel/tcg/helper-retaddr.h Richard Henderson
2025-04-28 21:48 ` Pierrick Bouvier
@ 2025-04-29 6:41 ` Philippe Mathieu-Daudé
1 sibling, 0 replies; 32+ messages in thread
From: Philippe Mathieu-Daudé @ 2025-04-29 6:41 UTC (permalink / raw)
To: Richard Henderson, qemu-devel; +Cc: pierrick.bouvier
On 28/4/25 22:10, Richard Henderson wrote:
> Move set_helper_retaddr and clear_helper_retaddr
> to a new header file.
>
> Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
> ---
> include/accel/tcg/cpu-ldst.h | 34 -----------------------
> include/accel/tcg/helper-retaddr.h | 43 ++++++++++++++++++++++++++++++
> accel/tcg/cpu-exec.c | 1 +
> accel/tcg/user-exec.c | 1 +
> target/arm/tcg/helper-a64.c | 1 +
> target/arm/tcg/sme_helper.c | 1 +
> target/arm/tcg/sve_helper.c | 1 +
> target/ppc/mem_helper.c | 1 +
> target/s390x/tcg/mem_helper.c | 1 +
> 9 files changed, 50 insertions(+), 34 deletions(-)
> create mode 100644 include/accel/tcg/helper-retaddr.h
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
^ permalink raw reply [flat|nested] 32+ messages in thread
* [PATCH 10/10] accel/tcg: Compile cpu-exec.c twice
2025-04-28 20:10 [PATCH 00/10] accel/tcg: Compile cpu-exec.c twice Richard Henderson
` (8 preceding siblings ...)
2025-04-28 20:10 ` [PATCH 09/10] accel/tcg: Split out accel/tcg/helper-retaddr.h Richard Henderson
@ 2025-04-28 20:10 ` Richard Henderson
2025-04-28 21:41 ` Pierrick Bouvier
9 siblings, 1 reply; 32+ messages in thread
From: Richard Henderson @ 2025-04-28 20:10 UTC (permalink / raw)
To: qemu-devel; +Cc: pierrick.bouvier, philmd
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
---
accel/tcg/cpu-exec.c | 2 --
accel/tcg/meson.build | 2 +-
2 files changed, 1 insertion(+), 3 deletions(-)
diff --git a/accel/tcg/cpu-exec.c b/accel/tcg/cpu-exec.c
index 985abb1269..47c34cdbd6 100644
--- a/accel/tcg/cpu-exec.c
+++ b/accel/tcg/cpu-exec.c
@@ -22,7 +22,6 @@
#include "qapi/error.h"
#include "qapi/type-helpers.h"
#include "hw/core/cpu.h"
-#include "accel/tcg/cpu-ldst.h"
#include "accel/tcg/cpu-ops.h"
#include "accel/tcg/helper-retaddr.h"
#include "trace.h"
@@ -37,7 +36,6 @@
#include "qemu/rcu.h"
#include "exec/log.h"
#include "qemu/main-loop.h"
-#include "cpu.h"
#include "exec/icount.h"
#include "exec/replay-core.h"
#include "system/tcg.h"
diff --git a/accel/tcg/meson.build b/accel/tcg/meson.build
index 3f7b127130..0bb089299b 100644
--- a/accel/tcg/meson.build
+++ b/accel/tcg/meson.build
@@ -5,6 +5,7 @@ endif
tcg_ss = ss.source_set()
tcg_ss.add(files(
+ 'cpu-exec.c',
'cpu-exec-common.c',
'tcg-runtime.c',
'tcg-runtime-gvec.c',
@@ -21,7 +22,6 @@ libsystem_ss.add_all(tcg_ss)
tcg_specific_ss = ss.source_set()
tcg_specific_ss.add(files(
'tcg-all.c',
- 'cpu-exec.c',
'translate-all.c',
))
tcg_specific_ss.add(when: 'CONFIG_USER_ONLY', if_true: files('user-exec.c'))
--
2.43.0
^ permalink raw reply related [flat|nested] 32+ messages in thread