diff mbox series

[v2] target/hppa: Speed up hppa_is_pa20()

Message ID Z3BouGUFFP87PJyx@p100 (mailing list archive)
State New
Headers show
Series [v2] target/hppa: Speed up hppa_is_pa20() | expand

Commit Message

Helge Deller Dec. 28, 2024, 9:08 p.m. UTC
Although the hppa_is_pa20() helper is costly due to string comparisms in
object_dynamic_cast(), it is called quite often during memory lookups
and at each start of a block of instruction translations.
Speed hppa_is_pa20() up by calling object_dynamic_cast() only once at
CPU creation and store the result in the is_pa20 of struct CPUArchState.

Signed-off-by: Helge Deller <deller@gmx.de>

v2:
- moved init to hppa_cpu_initfn() and is_pa20 to end of CPUArchState struct
  (feedback by Richard)

Comments

Richard Henderson Dec. 28, 2024, 10:54 p.m. UTC | #1
On 12/28/24 13:08, Helge Deller wrote:
> Although the hppa_is_pa20() helper is costly due to string comparisms in
> object_dynamic_cast(), it is called quite often during memory lookups
> and at each start of a block of instruction translations.
> Speed hppa_is_pa20() up by calling object_dynamic_cast() only once at
> CPU creation and store the result in the is_pa20 of struct CPUArchState.
> 
> Signed-off-by: Helge Deller <deller@gmx.de>
> 
> v2:
> - moved init to hppa_cpu_initfn() and is_pa20 to end of CPUArchState struct
>    (feedback by Richard)
> 

Reviewed-by: Richard Henderson <richard.henderson@linaro.org>

r~
Philippe Mathieu-Daudé Dec. 29, 2024, 5:24 p.m. UTC | #2
On 28/12/24 22:08, Helge Deller wrote:
> Although the hppa_is_pa20() helper is costly due to string comparisms in

"comparisms" -> "comparison"?

> object_dynamic_cast(), it is called quite often during memory lookups
> and at each start of a block of instruction translations.
> Speed hppa_is_pa20() up by calling object_dynamic_cast() only once at
> CPU creation and store the result in the is_pa20 of struct CPUArchState.
> 
> Signed-off-by: Helge Deller <deller@gmx.de>
> 
> v2:
> - moved init to hppa_cpu_initfn() and is_pa20 to end of CPUArchState struct
>    (feedback by Richard)
> 
> diff --git a/target/hppa/cpu.c b/target/hppa/cpu.c
> index b908cf65c6..05de952a87 100644
> --- a/target/hppa/cpu.c
> +++ b/target/hppa/cpu.c
> @@ -199,6 +199,7 @@ static void hppa_cpu_initfn(Object *obj)
>       CPUHPPAState *env = &cpu->env;
>   
>       cs->exception_index = -1;
> +    env->is_pa20 = object_dynamic_cast(obj, TYPE_HPPA64_CPU);

Personally I find explicit casts to boolean clearer:

        env->is_pa20 = !!object_dynamic_cast(obj, TYPE_HPPA64_CPU);

>       cpu_hppa_loaded_fr0(env);
>       cpu_hppa_put_psw(env, PSW_W);
>   }

Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
diff mbox series

Patch

diff --git a/target/hppa/cpu.c b/target/hppa/cpu.c
index b908cf65c6..05de952a87 100644
--- a/target/hppa/cpu.c
+++ b/target/hppa/cpu.c
@@ -199,6 +199,7 @@  static void hppa_cpu_initfn(Object *obj)
     CPUHPPAState *env = &cpu->env;
 
     cs->exception_index = -1;
+    env->is_pa20 = object_dynamic_cast(obj, TYPE_HPPA64_CPU);
     cpu_hppa_loaded_fr0(env);
     cpu_hppa_put_psw(env, PSW_W);
 }
diff --git a/target/hppa/cpu.h b/target/hppa/cpu.h
index 32a674a8b8..288ce6d98a 100644
--- a/target/hppa/cpu.h
+++ b/target/hppa/cpu.h
@@ -266,6 +266,8 @@  typedef struct CPUArchState {
 
     /* Fields up to this point are cleared by a CPU reset */
     struct {} end_reset_fields;
+
+    bool is_pa20;
 } CPUHPPAState;
 
 /**
@@ -299,7 +301,7 @@  struct HPPACPUClass {
 
 static inline bool hppa_is_pa20(CPUHPPAState *env)
 {
-    return object_dynamic_cast(OBJECT(env_cpu(env)), TYPE_HPPA64_CPU) != NULL;
+    return env->is_pa20;
 }
 
 static inline int HPPA_BTLB_ENTRIES(CPUHPPAState *env)