softfloat: Set 2-NaN propagation rule in float_status, not at compile time

[PATCH 11/21] target/sparc: Move cpu_put_fsr(env, 0) call to reset

Posted by Peter Maydell 4 weeks, 1 day ago

Currently we call cpu_put_fsr(0) in sparc_cpu_realizefn(), which
initializes various fields in the CPU struct:
 * fsr_cexc_ftt
 * fcc[]
 * fsr_qne
 * fsr
It also sets the rounding mode in env->fp_status.

This is largely pointless, because when we later reset the CPU
this will zero out all the fields up until the "end_reset_fields"
label, which includes all of these (but not fp_status!)

Move the cpu_put_fsr(env, 0) call to reset, because that expresses
the logical requirement: we want to reset FSR to 0 on every reset.
This isn't a behaviour change because the fields are all zero anyway.

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
---
 target/sparc/cpu.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/target/sparc/cpu.c b/target/sparc/cpu.c
index 54cb269e0af..e7f4068a162 100644
--- a/target/sparc/cpu.c
+++ b/target/sparc/cpu.c
@@ -76,6 +76,7 @@ static void sparc_cpu_reset_hold(Object *obj, ResetType type)
     env->npc = env->pc + 4;
 #endif
     env->cache_control = 0;
+    cpu_put_fsr(env, 0);
 }
 
 #ifndef CONFIG_USER_ONLY
@@ -805,7 +806,6 @@ static void sparc_cpu_realizefn(DeviceState *dev, Error **errp)
     env->version |= env->def.maxtl << 8;
     env->version |= env->def.nwindows - 1;
 #endif
-    cpu_put_fsr(env, 0);
 
     cpu_exec_realizefn(cs, &local_err);
     if (local_err != NULL) {
-- 
2.34.1

Re: [PATCH 11/21] target/sparc: Move cpu_put_fsr(env, 0) call to reset

Posted by Richard Henderson 3 weeks, 5 days ago

On 10/25/24 15:12, Peter Maydell wrote:
> Currently we call cpu_put_fsr(0) in sparc_cpu_realizefn(), which
> initializes various fields in the CPU struct:
>   * fsr_cexc_ftt
>   * fcc[]
>   * fsr_qne
>   * fsr
> It also sets the rounding mode in env->fp_status.
> 
> This is largely pointless, because when we later reset the CPU
> this will zero out all the fields up until the "end_reset_fields"
> label, which includes all of these (but not fp_status!)
> 
> Move the cpu_put_fsr(env, 0) call to reset, because that expresses
> the logical requirement: we want to reset FSR to 0 on every reset.
> This isn't a behaviour change because the fields are all zero anyway.
> 
> Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
> ---

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


r~

Re: [PATCH 11/21] target/sparc: Move cpu_put_fsr(env, 0) call to reset

Posted by Mark Cave-Ayland 3 weeks, 6 days ago

On 25/10/2024 15:12, Peter Maydell wrote:

> Currently we call cpu_put_fsr(0) in sparc_cpu_realizefn(), which
> initializes various fields in the CPU struct:
>   * fsr_cexc_ftt
>   * fcc[]
>   * fsr_qne
>   * fsr
> It also sets the rounding mode in env->fp_status.
> 
> This is largely pointless, because when we later reset the CPU
> this will zero out all the fields up until the "end_reset_fields"
> label, which includes all of these (but not fp_status!)
> 
> Move the cpu_put_fsr(env, 0) call to reset, because that expresses
> the logical requirement: we want to reset FSR to 0 on every reset.
> This isn't a behaviour change because the fields are all zero anyway.
> 
> Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
> ---
>   target/sparc/cpu.c | 2 +-
>   1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/target/sparc/cpu.c b/target/sparc/cpu.c
> index 54cb269e0af..e7f4068a162 100644
> --- a/target/sparc/cpu.c
> +++ b/target/sparc/cpu.c
> @@ -76,6 +76,7 @@ static void sparc_cpu_reset_hold(Object *obj, ResetType type)
>       env->npc = env->pc + 4;
>   #endif
>       env->cache_control = 0;
> +    cpu_put_fsr(env, 0);
>   }
>   
>   #ifndef CONFIG_USER_ONLY
> @@ -805,7 +806,6 @@ static void sparc_cpu_realizefn(DeviceState *dev, Error **errp)
>       env->version |= env->def.maxtl << 8;
>       env->version |= env->def.nwindows - 1;
>   #endif
> -    cpu_put_fsr(env, 0);
>   
>       cpu_exec_realizefn(cs, &local_err);
>       if (local_err != NULL) {

Reviewed-by: Mark Cave-Ayland <mark.cave-ayland@ilande.co.uk>


ATB,

Mark.

Re: [PATCH 11/21] target/sparc: Move cpu_put_fsr(env, 0) call to reset

Posted by Philippe Mathieu-Daudé 4 weeks ago

On 25/10/24 11:12, Peter Maydell wrote:
> Currently we call cpu_put_fsr(0) in sparc_cpu_realizefn(), which
> initializes various fields in the CPU struct:
>   * fsr_cexc_ftt
>   * fcc[]
>   * fsr_qne
>   * fsr
> It also sets the rounding mode in env->fp_status.
> 
> This is largely pointless, because when we later reset the CPU
> this will zero out all the fields up until the "end_reset_fields"
> label, which includes all of these (but not fp_status!)
> 
> Move the cpu_put_fsr(env, 0) call to reset, because that expresses
> the logical requirement: we want to reset FSR to 0 on every reset.
> This isn't a behaviour change because the fields are all zero anyway.
> 
> Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
> ---
>   target/sparc/cpu.c | 2 +-
>   1 file changed, 1 insertion(+), 1 deletion(-)

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