[PATCH v2 3/9] accel/tcg: Return the TranslationBlock from cpu_unwind_state_data

Richard Henderson posted 9 patches 5 months, 3 weeks ago
Maintainers: Richard Henderson <richard.henderson@linaro.org>, Paolo Bonzini <pbonzini@redhat.com>, Eduardo Habkost <eduardo@habkost.net>, Marcel Apfelbaum <marcel.apfelbaum@gmail.com>, "Philippe Mathieu-Daudé" <philmd@linaro.org>, Yanan Wang <wangyanan55@huawei.com>, "Alex Bennée" <alex.bennee@linaro.org>, Alexandre Iooss <erdnaxe@crans.org>, Mahmoud Mandour <ma.mandourr@gmail.com>, Pierrick Bouvier <pierrick.bouvier@linaro.org>, Peter Maydell <peter.maydell@linaro.org>
[PATCH v2 3/9] accel/tcg: Return the TranslationBlock from cpu_unwind_state_data
Posted by Richard Henderson 5 months, 3 weeks ago
Adjust the i386 get_memio_eip function to use tb->cflags instead
of tcg_cflags_has, which is technically more correct.

Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
---
 include/exec/cpu-common.h | 9 +++++----
 accel/tcg/translate-all.c | 9 +++++----
 target/i386/helper.c      | 6 ++++--
 3 files changed, 14 insertions(+), 10 deletions(-)

diff --git a/include/exec/cpu-common.h b/include/exec/cpu-common.h
index 815342d043..c1887462e6 100644
--- a/include/exec/cpu-common.h
+++ b/include/exec/cpu-common.h
@@ -189,12 +189,13 @@ uint32_t curr_cflags(CPUState *cpu);
  * @host_pc: the host pc within the translation
  * @data: output data
  *
- * Attempt to load the the unwind state for a host pc occurring in
- * translated code.  If @host_pc is not in translated code, the
- * function returns false; otherwise @data is loaded.
+ * Attempt to load the the unwind state for a host pc occurring in translated
+ * code.  If @host_pc is not in translated code, the function returns NULL;
+ * otherwise @data is loaded and the TranslationBlock is returned.
  * This is the same unwind info as given to restore_state_to_opc.
  */
-bool cpu_unwind_state_data(CPUState *cpu, uintptr_t host_pc, uint64_t *data);
+const TranslationBlock *cpu_unwind_state_data(CPUState *cpu, uintptr_t host_pc,
+                                              uint64_t *data);
 
 /**
  * cpu_restore_state:
diff --git a/accel/tcg/translate-all.c b/accel/tcg/translate-all.c
index fdf6d8ac19..45a1cf57bc 100644
--- a/accel/tcg/translate-all.c
+++ b/accel/tcg/translate-all.c
@@ -243,15 +243,16 @@ bool cpu_restore_state(CPUState *cpu, uintptr_t host_pc)
     return false;
 }
 
-bool cpu_unwind_state_data(CPUState *cpu, uintptr_t host_pc, uint64_t *data)
+const TranslationBlock *
+cpu_unwind_state_data(CPUState *cpu, uintptr_t host_pc, uint64_t *data)
 {
     if (in_code_gen_buffer((const void *)(host_pc - tcg_splitwx_diff))) {
         TranslationBlock *tb = tcg_tb_lookup(host_pc);
-        if (tb) {
-            return cpu_unwind_data_from_tb(tb, host_pc, data) >= 0;
+        if (tb && cpu_unwind_data_from_tb(tb, host_pc, data) >= 0) {
+            return tb;
         }
     }
-    return false;
+    return NULL;
 }
 
 void page_init(void)
diff --git a/target/i386/helper.c b/target/i386/helper.c
index f9d1381f90..565e01a3a9 100644
--- a/target/i386/helper.c
+++ b/target/i386/helper.c
@@ -521,13 +521,15 @@ static inline target_ulong get_memio_eip(CPUX86State *env)
 #ifdef CONFIG_TCG
     uint64_t data[TARGET_INSN_START_WORDS];
     CPUState *cs = env_cpu(env);
+    const TranslationBlock *tb;
 
-    if (!cpu_unwind_state_data(cs, cs->mem_io_pc, data)) {
+    tb = cpu_unwind_state_data(cs, cs->mem_io_pc, data);
+    if (!tb) {
         return env->eip;
     }
 
     /* Per x86_restore_state_to_opc. */
-    if (tcg_cflags_has(cs, CF_PCREL)) {
+    if (tb->cflags & CF_PCREL) {
         return (env->eip & TARGET_PAGE_MASK) | data[0];
     } else {
         return data[0] - env->segs[R_CS].base;
-- 
2.34.1
Re: [PATCH v2 3/9] accel/tcg: Return the TranslationBlock from cpu_unwind_state_data
Posted by Alex Bennée 5 months, 3 weeks ago
Richard Henderson <richard.henderson@linaro.org> writes:

> Adjust the i386 get_memio_eip function to use tb->cflags instead
> of tcg_cflags_has, which is technically more correct.
>
> Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org>
> Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

Reviewed-by: Alex Bennée <alex.bennee@linaro.org>

-- 
Alex Bennée
Virtualisation Tech Lead @ Linaro