[Qemu-devel] [PATCH 0/2] Fix VRECPS edge cases handling

[Qemu-devel] [PATCH 0/2] Fix VRECPS edge cases handling

[Peter Maydell]

This patchset fixes the edge case handling of VRECPS. Patch 2/2
is just a bit of cleanup of the neighbouring vrsqrts helper which
can then use the float32_two introduced by 1/1.

Tested in the usual random-insn-generation way and also with the
"neon64" test program from the valgrind ARM testsuite.

Peter Maydell (2):
  target-arm: Fix VRECPS edge cases handling
  target-arm: use make_float32() to make constant floats for VRSQRTS

 target-arm/helper.c |   22 +++++++++++++---------
 1 files changed, 13 insertions(+), 9 deletions(-)

[Qemu-devel] [PATCH 1/2] target-arm: Fix VRECPS edge cases handling

[Peter Maydell]

Correct the handling of edge cases for the VRECPS instruction:
 * this is a Neon instruction so uses the "standard FPSCR value"
 * (zero, inf) is a special case which returns 2.0

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
---
 target-arm/helper.c |   11 ++++++++---
 1 files changed, 8 insertions(+), 3 deletions(-)

diff --git a/target-arm/helper.c b/target-arm/helper.c
index d360121..c01a5a2 100644
--- a/target-arm/helper.c
+++ b/target-arm/helper.c
@@ -2707,11 +2707,16 @@ uint32_t HELPER(vfp_fcvt_f32_to_f16)(float32 a, CPUState *env)
     return do_fcvt_f32_to_f16(a, env, &env->vfp.fp_status);
 }
 
+#define float32_two make_float32(0x40000000)
+
 float32 HELPER(recps_f32)(float32 a, float32 b, CPUState *env)
 {
-    float_status *s = &env->vfp.fp_status;
-    float32 two = int32_to_float32(2, s);
-    return float32_sub(two, float32_mul(a, b, s), s);
+    float_status *s = &env->vfp.standard_fp_status;
+    if ((float32_is_infinity(a) && float32_is_zero_or_denormal(b)) ||
+        (float32_is_infinity(b) && float32_is_zero_or_denormal(a))) {
+        return float32_two;
+    }
+    return float32_sub(float32_two, float32_mul(a, b, s), s);
 }
 
 float32 HELPER(rsqrts_f32)(float32 a, float32 b, CPUState *env)
-- 
1.7.1

[Qemu-devel] [PATCH 2/2] target-arm: use make_float32() to make constant floats for VRSQRTS

[Peter Maydell]

The preferred way to create a constant floating point value is to use
make_float32() rather than doing a runtime int32_to_float32().
Convert the code in the VRSQRTS helper to work this way.

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
---
 target-arm/helper.c |   11 +++++------
 1 files changed, 5 insertions(+), 6 deletions(-)

diff --git a/target-arm/helper.c b/target-arm/helper.c
index c01a5a2..c3238d3 100644
--- a/target-arm/helper.c
+++ b/target-arm/helper.c
@@ -2708,6 +2708,8 @@ uint32_t HELPER(vfp_fcvt_f32_to_f16)(float32 a, CPUState *env)
 }
 
 #define float32_two make_float32(0x40000000)
+#define float32_three make_float32(0x40400000)
+#define float32_one_point_five makefloat32(0x3fc00000)
 
 float32 HELPER(recps_f32)(float32 a, float32 b, CPUState *env)
 {
@@ -2722,16 +2724,13 @@ float32 HELPER(recps_f32)(float32 a, float32 b, CPUState *env)
 float32 HELPER(rsqrts_f32)(float32 a, float32 b, CPUState *env)
 {
     float_status *s = &env->vfp.standard_fp_status;
-    float32 two = int32_to_float32(2, s);
-    float32 three = int32_to_float32(3, s);
     float32 product;
     if ((float32_is_infinity(a) && float32_is_zero_or_denormal(b)) ||
         (float32_is_infinity(b) && float32_is_zero_or_denormal(a))) {
-        product = float32_zero;
-    } else {
-        product = float32_mul(a, b, s);
+        return float32_one_point_five;
     }
-    return float32_div(float32_sub(three, product, s), two, s);
+    product = float32_mul(a, b, s);
+    return float32_div(float32_sub(float32_three, product, s), float32_two, s);
 }
 
 /* NEON helpers.  */
-- 
1.7.1