Re: [Qemu-devel] [PATCH 08/40] Add vcmpequ{b, h, w} and vcmpgt{s, u}{b, h, w} instructions.

[Nathan Froyd <froydnj@codesourcery.com>]

On Sat, Jan 03, 2009 at 03:02:14PM +0100, Aurelien Jarno wrote:
> On Tue, Dec 30, 2008 at 07:09:50PM -0800, Nathan Froyd wrote:
> > +#define VCMP(suffix, compare, element)                                  \
> > +    uint32_t helper_vcmp##suffix (ppc_avr_t *r, ppc_avr_t *a, ppc_avr_t *b) \
> > +    {                                                                   \
> > +        uint32_t ones = (sizeof (a->element[0]) == 4                    \
> > +                         ? 0xffffffff                                   \
> > +                         : (sizeof (a->element[0]) == 2                 \
> > +                            ? 0xffff                                    \
> > +                            : 0xff));                                   \
> > +        uint32_t all = ones;                                            \
> > +        uint32_t none = 0;                                              \
> > +        int i;                                                          \
> > +        for (i = 0; i < ARRAY_SIZE(r->element); i++) {                  \
> > +            uint32_t result = (a->element[i] compare b->element[i] ? ones : 0x0); \
> > +            switch (sizeof (a->element[0])) {                           \
> > +            case 4: r->u32[i] = result; break;                          \
> > +            case 2: r->u16[i] = result; break;                          \
> > +            case 1: r->u8[i] = result; break;                           \
> > +            }                                                           \
> > +            all &= result;                                              \
> > +            none |= result;                                             \
> > +        }                                                               \
> 
> The part defining the ones looks a bit complicated. You may want to
> define result as int32_t, and put either -1 or 0. Then you can just cast
> the value:
>                case 4: r->u32[i] = (int32_t)result; break;
>                case 2: r->u16[i] = (int16_t)result; break;
>                case 1: r->u8[i] = (int8_t)result; break;

Done (or something close to it) thusly.

-Nathan

Create separate helpers for record and non-recording versions.

Signed-off-by: Nathan Froyd <froydnj@codesourcery.com>
---
 target-ppc/helper.h    |   18 ++++++++++++++++++
 target-ppc/op_helper.c |   36 ++++++++++++++++++++++++++++++++++++
 target-ppc/translate.c |   10 ++++++++++
 3 files changed, 64 insertions(+), 0 deletions(-)

diff --git a/target-ppc/helper.h b/target-ppc/helper.h
index f319fdb..bae137b 100644
--- a/target-ppc/helper.h
+++ b/target-ppc/helper.h
@@ -123,6 +123,24 @@ DEF_HELPER_3(vminuw, void, avr, avr, avr)
 DEF_HELPER_3(vmaxub, void, avr, avr, avr)
 DEF_HELPER_3(vmaxuh, void, avr, avr, avr)
 DEF_HELPER_3(vmaxuw, void, avr, avr, avr)
+DEF_HELPER_3(vcmpequb, void, avr, avr, avr)
+DEF_HELPER_3(vcmpequh, void, avr, avr, avr)
+DEF_HELPER_3(vcmpequw, void, avr, avr, avr)
+DEF_HELPER_3(vcmpgtub, void, avr, avr, avr)
+DEF_HELPER_3(vcmpgtuh, void, avr, avr, avr)
+DEF_HELPER_3(vcmpgtuw, void, avr, avr, avr)
+DEF_HELPER_3(vcmpgtsb, void, avr, avr, avr)
+DEF_HELPER_3(vcmpgtsh, void, avr, avr, avr)
+DEF_HELPER_3(vcmpgtsw, void, avr, avr, avr)
+DEF_HELPER_3(vcmpequb_dot, void, avr, avr, avr)
+DEF_HELPER_3(vcmpequh_dot, void, avr, avr, avr)
+DEF_HELPER_3(vcmpequw_dot, void, avr, avr, avr)
+DEF_HELPER_3(vcmpgtub_dot, void, avr, avr, avr)
+DEF_HELPER_3(vcmpgtuh_dot, void, avr, avr, avr)
+DEF_HELPER_3(vcmpgtuw_dot, void, avr, avr, avr)
+DEF_HELPER_3(vcmpgtsb_dot, void, avr, avr, avr)
+DEF_HELPER_3(vcmpgtsh_dot, void, avr, avr, avr)
+DEF_HELPER_3(vcmpgtsw_dot, void, avr, avr, avr)
 DEF_HELPER_3(vmrglb, void, avr, avr, avr)
 DEF_HELPER_3(vmrglh, void, avr, avr, avr)
 DEF_HELPER_3(vmrglw, void, avr, avr, avr)
diff --git a/target-ppc/op_helper.c b/target-ppc/op_helper.c
index 5e40e42..eb155aa 100644
--- a/target-ppc/op_helper.c
+++ b/target-ppc/op_helper.c
@@ -2101,6 +2101,42 @@ VAVG(w, s32, int64_t, u32, uint64_t)
 #undef VAVG_DO
 #undef VAVG
 
+#define VCMP_DO(suffix, compare, element, record)                       \
+    void helper_vcmp##suffix (ppc_avr_t *r, ppc_avr_t *a, ppc_avr_t *b) \
+    {                                                                   \
+        uint32_t ones = (uint32_t)-1;                                   \
+        uint32_t all = ones;                                            \
+        uint32_t none = 0;                                              \
+        int i;                                                          \
+        for (i = 0; i < ARRAY_SIZE(r->element); i++) {                  \
+            uint32_t result = (a->element[i] compare b->element[i] ? ones : 0x0); \
+            switch (sizeof (a->element[0])) {                           \
+            case 4: r->u32[i] = result; break;                          \
+            case 2: r->u16[i] = result; break;                          \
+            case 1: r->u8[i] = result; break;                           \
+            }                                                           \
+            all &= result;                                              \
+            none |= result;                                             \
+        }                                                               \
+        if (record) {                                                   \
+            env->crf[6] = ((all != 0) << 3) | ((none == 0) << 1);       \
+        }                                                               \
+    }
+#define VCMP(suffix, compare, element)          \
+    VCMP_DO(suffix, compare, element, 0)        \
+    VCMP_DO(suffix##_dot, compare, element, 1)
+VCMP(equb, ==, u8)
+VCMP(equh, ==, u16)
+VCMP(equw, ==, u32)
+VCMP(gtub, >, u8)
+VCMP(gtuh, >, u16)
+VCMP(gtuw, >, u32)
+VCMP(gtsb, >, s8)
+VCMP(gtsh, >, s16)
+VCMP(gtsw, >, s32)
+#undef VCMP_DO
+#undef VCMP
+
 void helper_vmhaddshs (ppc_avr_t *r, ppc_avr_t *a, ppc_avr_t *b, ppc_avr_t *c)
 {
     int sat = 0;
diff --git a/target-ppc/translate.c b/target-ppc/translate.c
index ee3c747..20e9e0c 100644
--- a/target-ppc/translate.c
+++ b/target-ppc/translate.c
@@ -6380,6 +6380,16 @@ GEN_VXFORM(vsumsws, 4, 30);
     GEN_VXRFORM1(name, name, #name, opc2, opc3)                      \
     GEN_VXRFORM1(name##_dot, name##_, #name ".", opc2, (opc3 | (0x1 << 4)))
 
+GEN_VXRFORM(vcmpequb, 3, 0)
+GEN_VXRFORM(vcmpequh, 3, 1)
+GEN_VXRFORM(vcmpequw, 3, 2)
+GEN_VXRFORM(vcmpgtsb, 3, 12)
+GEN_VXRFORM(vcmpgtsh, 3, 13)
+GEN_VXRFORM(vcmpgtsw, 3, 14)
+GEN_VXRFORM(vcmpgtub, 3, 8)
+GEN_VXRFORM(vcmpgtuh, 3, 9)
+GEN_VXRFORM(vcmpgtuw, 3, 10)
+
 #define GEN_VXFORM_NOA(name, opc2, opc3)                                \
     GEN_HANDLER(name, 0x04, opc2, opc3, 0x001f0000, PPC_ALTIVEC)        \
     {                                                                   \
-- 
1.6.0.5