Re: [Qemu-devel] [PATCH] FMULX should flushes operators to zero when FZ is set.

[Xiangyu Hu]

Thanks Peter. 

I have sent another patch with updated commit message.

- xiangyu

On 28 Jan, 2015, at 10:54 pm, Peter Maydell <address@hidden> wrote:

On 28 January 2015 at 14:51, Xiangyu Hu <address@hidden> wrote:

The difference between FMULX and FMUL is that FMULX will return 2.0f when one operator is
FPInfinity and the other one is FPZero, whilst FMUL will return a Default NaN. Without
this patch, the emulation would result in inconsistency.

Signed-off-by: Xiangyu Hu <address@hidden>
---
target-arm/helper-a64.c | 6 ++++++
1 file changed, 6 insertions(+)

diff --git a/target-arm/helper-a64.c b/target-arm/helper-a64.c
index 81066ca..ebd9247 100644
--- a/target-arm/helper-a64.c
+++ b/target-arm/helper-a64.c
@@ -135,6 +135,9 @@ float32 HELPER(vfp_mulxs)(float32 a, float32 b, void *fpstp)
{
    float_status *fpst = fpstp;

+    a = float32_squash_input_denormal(a, fpst);
+    b = float32_squash_input_denormal(b, fpst);
+
    if ((float32_is_zero(a) && float32_is_infinity(b)) ||
        (float32_is_infinity(a) && float32_is_zero(b))) {
        /* 2.0 with the sign bit set to sign(A) XOR sign(B) */
@@ -148,6 +151,9 @@ float64 HELPER(vfp_mulxd)(float64 a, float64 b, void *fpstp)
{
    float_status *fpst = fpstp;

+    a = float64_squash_input_denormal(a, fpst);
+    b = float64_squash_input_denormal(b, fpst);
+
    if ((float64_is_zero(a) && float64_is_infinity(b)) ||
        (float64_is_infinity(a) && float64_is_zero(b))) {
        /* 2.0 with the sign bit set to sign(A) XOR sign(B) */

I think this code change is correct but the commit message doesn't
do a very good job of describing what the change is or why it is
needed... (in particular the difference between FMULX and FMUL is
completely irrelevant to the need for this fix, but it is pretty
much the only thing the commit message talks about).

thanks
-- PMM