fix fp16 tininess

Message ID 20180510140934.22855-1-peter.maydell@linaro.org
State Accepted
Headers show
Series
  • fix fp16 tininess
Related show

Commit Message

Peter Maydell May 10, 2018, 2:09 p.m.
In commit d81ce0ef2c4f105 we added an extra float_status field
fp_status_fp16 for Arm, but forgot to initialize it correctly
by setting it to float_tininess_before_rounding. This currently
will only cause problems for the new V8_FP16 feature, since the
float-to-float conversion code doesn't use it yet. The effect
would be that we failed to set the Underflow IEEE exception flag
in all the cases where we should.

Add the missing initialization.

Fixes: d81ce0ef2c4f105
Cc: qemu-stable@nongnu.org
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

---
 target/arm/cpu.c | 2 ++
 1 file changed, 2 insertions(+)

-- 
2.17.0

Comments

Peter Maydell May 10, 2018, 2:11 p.m. | #1
On 10 May 2018 at 15:09, Peter Maydell <peter.maydell@linaro.org> wrote:
> In commit d81ce0ef2c4f105 we added an extra float_status field

> fp_status_fp16 for Arm, but forgot to initialize it correctly

> by setting it to float_tininess_before_rounding. This currently

> will only cause problems for the new V8_FP16 feature, since the

> float-to-float conversion code doesn't use it yet. The effect

> would be that we failed to set the Underflow IEEE exception flag

> in all the cases where we should.


Oops, I forgot to fix the patch summary line from my
quick-hack commit. Something like:

target/arm: Fix tininess detection for fp16 arithmetic

would be a bit clearer.

thanks
-- PMM
Richard Henderson May 10, 2018, 2:34 p.m. | #2
On 05/10/2018 07:09 AM, Peter Maydell wrote:
> In commit d81ce0ef2c4f105 we added an extra float_status field

> fp_status_fp16 for Arm, but forgot to initialize it correctly

> by setting it to float_tininess_before_rounding. This currently

> will only cause problems for the new V8_FP16 feature, since the

> float-to-float conversion code doesn't use it yet. The effect

> would be that we failed to set the Underflow IEEE exception flag

> in all the cases where we should.

> 

> Add the missing initialization.

> 

> Fixes: d81ce0ef2c4f105

> Cc: qemu-stable@nongnu.org

> Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

> ---

>  target/arm/cpu.c | 2 ++

>  1 file changed, 2 insertions(+)


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



r~
Alex Bennée May 10, 2018, 3:07 p.m. | #3
Peter Maydell <peter.maydell@linaro.org> writes:

> In commit d81ce0ef2c4f105 we added an extra float_status field

> fp_status_fp16 for Arm, but forgot to initialize it correctly

> by setting it to float_tininess_before_rounding. This currently

> will only cause problems for the new V8_FP16 feature, since the

> float-to-float conversion code doesn't use it yet. The effect

> would be that we failed to set the Underflow IEEE exception flag

> in all the cases where we should.

>

> Add the missing initialization.

>

> Fixes: d81ce0ef2c4f105

> Cc: qemu-stable@nongnu.org

> Signed-off-by: Peter Maydell <peter.maydell@linaro.org>


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


> ---

>  target/arm/cpu.c | 2 ++

>  1 file changed, 2 insertions(+)

>

> diff --git a/target/arm/cpu.c b/target/arm/cpu.c

> index d175c5e94f..7939c6b8ae 100644

> --- a/target/arm/cpu.c

> +++ b/target/arm/cpu.c

> @@ -324,6 +324,8 @@ static void arm_cpu_reset(CPUState *s)

>                                &env->vfp.fp_status);

>      set_float_detect_tininess(float_tininess_before_rounding,

>                                &env->vfp.standard_fp_status);

> +    set_float_detect_tininess(float_tininess_before_rounding,

> +                              &env->vfp.fp_status_f16);

>  #ifndef CONFIG_USER_ONLY

>      if (kvm_enabled()) {

>          kvm_arm_reset_vcpu(cpu);



--
Alex Bennée

Patch

diff --git a/target/arm/cpu.c b/target/arm/cpu.c
index d175c5e94f..7939c6b8ae 100644
--- a/target/arm/cpu.c
+++ b/target/arm/cpu.c
@@ -324,6 +324,8 @@  static void arm_cpu_reset(CPUState *s)
                               &env->vfp.fp_status);
     set_float_detect_tininess(float_tininess_before_rounding,
                               &env->vfp.standard_fp_status);
+    set_float_detect_tininess(float_tininess_before_rounding,
+                              &env->vfp.fp_status_f16);
 #ifndef CONFIG_USER_ONLY
     if (kvm_enabled()) {
         kvm_arm_reset_vcpu(cpu);