[v1,5/5] target-ppc: add vector permute right indexed instruction

Commit Message

Rajalakshmi Srinivasaraghavan Aug. 4, 2016, 1:03 p.m. UTC

Add vpermr instruction from ISA 3.0.

Signed-off-by: Rajalakshmi Srinivasaraghavan <[e-mail]>
---
 target-ppc/helper.h             |    1 +
 target-ppc/int_helper.c         |   23 +++++++++++++++++++++++
 target-ppc/translate/vmx-impl.c |   18 ++++++++++++++++++
 target-ppc/translate/vmx-ops.c  |    1 +
 4 files changed, 43 insertions(+), 0 deletions(-)

Comments

Richard Henderson Aug. 4, 2016, 4:53 p.m. UTC | #1

On 08/04/2016 06:33 PM, Rajalakshmi Srinivasaraghavan wrote:
> +        if (s & 0x10) {
> +            result.u8[i] = a->u8[15 - index];
> +        } else {
> +            result.u8[i] = b->u8[15 - index];
> +        }

Fold these "15 -" into the computation of index above?


r~

Patch

diff --git a/target-ppc/helper.h b/target-ppc/helper.h
index d1d9418..3c476c9 100644
--- a/target-ppc/helper.h
+++ b/target-ppc/helper.h
@@ -270,6 +270,7 @@  DEF_HELPER_5(vmsumubm, void, env, avr, avr, avr, avr)
 DEF_HELPER_5(vmsummbm, void, env, avr, avr, avr, avr)
 DEF_HELPER_5(vsel, void, env, avr, avr, avr, avr)
 DEF_HELPER_5(vperm, void, env, avr, avr, avr, avr)
+DEF_HELPER_5(vpermr, void, env, avr, avr, avr, avr)
 DEF_HELPER_4(vpkshss, void, env, avr, avr, avr)
 DEF_HELPER_4(vpkshus, void, env, avr, avr, avr)
 DEF_HELPER_4(vpkswss, void, env, avr, avr, avr)
diff --git a/target-ppc/int_helper.c b/target-ppc/int_helper.c
index d8ad56f..79e4273 100644
--- a/target-ppc/int_helper.c
+++ b/target-ppc/int_helper.c
@@ -1126,6 +1126,29 @@  void helper_vperm(CPUPPCState *env, ppc_avr_t *r, ppc_avr_t *a, ppc_avr_t *b,
     *r = result;
 }
 
+void helper_vpermr(CPUPPCState *env, ppc_avr_t *r, ppc_avr_t *a, ppc_avr_t *b,
+                  ppc_avr_t *c)
+{
+    ppc_avr_t result;
+    int i;
+
+    VECTOR_FOR_INORDER_I(i, u8) {
+        int s = c->u8[i] & 0x1f;
+#if defined(HOST_WORDS_BIGENDIAN)
+        int index = s & 0xf;
+#else
+        int index = 15 - (s & 0xf);
+#endif
+
+        if (s & 0x10) {
+            result.u8[i] = a->u8[15 - index];
+        } else {
+            result.u8[i] = b->u8[15 - index];
+        }
+    }
+    *r = result;
+}
+
 #if defined(HOST_WORDS_BIGENDIAN)
 #define VBPERMQ_INDEX(avr, i) ((avr)->u8[(i)])
 #define VBPERMQ_DW(index) (((index) & 0x40) != 0)
diff --git a/target-ppc/translate/vmx-impl.c b/target-ppc/translate/vmx-impl.c
index 5ddff58..d13640f 100644
--- a/target-ppc/translate/vmx-impl.c
+++ b/target-ppc/translate/vmx-impl.c
@@ -728,6 +728,24 @@  static void gen_vmladduhm(DisasContext *ctx)
     tcg_temp_free_ptr(rd);
 }
 
+static void gen_vpermr(DisasContext *ctx)
+{
+    TCGv_ptr ra, rb, rc, rd;
+    if (unlikely(!ctx->altivec_enabled)) {
+        gen_exception(ctx, POWERPC_EXCP_VPU);
+        return;
+    }
+    ra = gen_avr_ptr(rA(ctx->opcode));
+    rb = gen_avr_ptr(rB(ctx->opcode));
+    rc = gen_avr_ptr(rC(ctx->opcode));
+    rd = gen_avr_ptr(rD(ctx->opcode));
+    gen_helper_vpermr(cpu_env, rd, ra, rb, rc);
+    tcg_temp_free_ptr(ra);
+    tcg_temp_free_ptr(rb);
+    tcg_temp_free_ptr(rc);
+    tcg_temp_free_ptr(rd);
+}
+
 GEN_VAFORM_PAIRED(vmsumubm, vmsummbm, 18)
 GEN_VAFORM_PAIRED(vmsumuhm, vmsumuhs, 19)
 GEN_VAFORM_PAIRED(vmsumshm, vmsumshs, 20)
diff --git a/target-ppc/translate/vmx-ops.c b/target-ppc/translate/vmx-ops.c
index 32bd533..ad72db5 100644
--- a/target-ppc/translate/vmx-ops.c
+++ b/target-ppc/translate/vmx-ops.c
@@ -219,6 +219,7 @@  GEN_VXFORM_300_EO(vctzb, 0x01, 0x18, 0x1C),
 GEN_VXFORM_300_EO(vctzh, 0x01, 0x18, 0x1D),
 GEN_VXFORM_300_EO(vctzw, 0x01, 0x18, 0x1E),
 GEN_VXFORM_300_EO(vctzd, 0x01, 0x18, 0x1F),
+GEN_VXFORM_300(vpermr, 0x1D, 0xFF),
 
 #define GEN_VXFORM_NOA(name, opc2, opc3)                                \
     GEN_HANDLER(name, 0x04, opc2, opc3, 0x001f0000, PPC_ALTIVEC)