neon_vget_high and neon_vget_low extract one half of a vector.
The patterns look like:
(define_insn "neon_vget_highv16qi"
[(set (match_operand:V8QI 0 "s_register_operand" "=w")
(vec_select:V8QI (match_operand:V16QI 1 "s_register_operand" "w")
(parallel [(const_int 8) (const_int 9)
(const_int 10) (const_int 11)
(const_int 12) (const_int 13)
(const_int 14) (const_int 15)])))]
"TARGET_NEON"
{
int dest = REGNO (operands[0]);
int src = REGNO (operands[1]);
if (dest != src + 2)
return "vmov\t%P0, %f1";
else
return "";
}
[(set_attr "neon_type" "neon_bp_simple")]
)
But there's nothing here to tell the register allocator what's expected
of it, so we do often get the move. The patch below makes the patterns
expand to normal subreg moves instead.
Unfortunately, when I first tried this, I ran across some bugs in
simplify-rtx.c. They should be fixed now. Of course, I can't
guarantee that there are other similar bugs elsewhere, but I'll
try to fix any that crop up.
The new patterns preserve the current treatment on big-endian targets.
Namely, the "low" half is always in the lower-numbered registers
(subreg byte offset 0).
Tested on arm-linux-gnueabi. OK to install?
Richard
gcc/
* config/arm/neon.md (neon_vget_highv16qi, neon_vget_highv8hi)
(neon_vget_highv4si, neon_vget_highv4sf, neon_vget_highv2di)
(neon_vget_lowv16qi, neon_vget_lowv8hi, neon_vget_lowv4si)
(neon_vget_lowv4sf, neon_vget_lowv2di): Turn into define_expands
that produce subreg moves. Define using VQX iterators.
Index: gcc/config/arm/neon.md
===================================================================
--- gcc/config/arm/neon.md 2011-09-13 13:33:29.000000000 +0100
+++ gcc/config/arm/neon.md 2011-09-13 16:21:23.189304498 +0100
@@ -2969,183 +2969,27 @@ (define_insn "neon_vcombine<mode>"
(set_attr "neon_type" "neon_bp_simple")]
)
-(define_insn "neon_vget_highv16qi"
- [(set (match_operand:V8QI 0 "s_register_operand" "=w")
- (vec_select:V8QI (match_operand:V16QI 1 "s_register_operand" "w")
- (parallel [(const_int 8) (const_int 9)
- (const_int 10) (const_int 11)
- (const_int 12) (const_int 13)
- (const_int 14) (const_int 15)])))]
- "TARGET_NEON"
-{
- int dest = REGNO (operands[0]);
- int src = REGNO (operands[1]);
-
- if (dest != src + 2)
- return "vmov\t%P0, %f1";
- else
- return "";
-}
- [(set_attr "neon_type" "neon_bp_simple")]
-)
-
-(define_insn "neon_vget_highv8hi"
- [(set (match_operand:V4HI 0 "s_register_operand" "=w")
- (vec_select:V4HI (match_operand:V8HI 1 "s_register_operand" "w")
- (parallel [(const_int 4) (const_int 5)
- (const_int 6) (const_int 7)])))]
- "TARGET_NEON"
-{
- int dest = REGNO (operands[0]);
- int src = REGNO (operands[1]);
-
- if (dest != src + 2)
- return "vmov\t%P0, %f1";
- else
- return "";
-}
- [(set_attr "neon_type" "neon_bp_simple")]
-)
-
-(define_insn "neon_vget_highv4si"
- [(set (match_operand:V2SI 0 "s_register_operand" "=w")
- (vec_select:V2SI (match_operand:V4SI 1 "s_register_operand" "w")
- (parallel [(const_int 2) (const_int 3)])))]
- "TARGET_NEON"
-{
- int dest = REGNO (operands[0]);
- int src = REGNO (operands[1]);
-
- if (dest != src + 2)
- return "vmov\t%P0, %f1";
- else
- return "";
-}
- [(set_attr "neon_type" "neon_bp_simple")]
-)
-
-(define_insn "neon_vget_highv4sf"
- [(set (match_operand:V2SF 0 "s_register_operand" "=w")
- (vec_select:V2SF (match_operand:V4SF 1 "s_register_operand" "w")
- (parallel [(const_int 2) (const_int 3)])))]
- "TARGET_NEON"
-{
- int dest = REGNO (operands[0]);
- int src = REGNO (operands[1]);
-
- if (dest != src + 2)
- return "vmov\t%P0, %f1";
- else
- return "";
-}
- [(set_attr "neon_type" "neon_bp_simple")]
-)
-
-(define_insn "neon_vget_highv2di"
- [(set (match_operand:DI 0 "s_register_operand" "=w")
- (vec_select:DI (match_operand:V2DI 1 "s_register_operand" "w")
- (parallel [(const_int 1)])))]
- "TARGET_NEON"
-{
- int dest = REGNO (operands[0]);
- int src = REGNO (operands[1]);
-
- if (dest != src + 2)
- return "vmov\t%P0, %f1";
- else
- return "";
-}
- [(set_attr "neon_type" "neon_bp_simple")]
-)
-
-(define_insn "neon_vget_lowv16qi"
- [(set (match_operand:V8QI 0 "s_register_operand" "=w")
- (vec_select:V8QI (match_operand:V16QI 1 "s_register_operand" "w")
- (parallel [(const_int 0) (const_int 1)
- (const_int 2) (const_int 3)
- (const_int 4) (const_int 5)
- (const_int 6) (const_int 7)])))]
- "TARGET_NEON"
-{
- int dest = REGNO (operands[0]);
- int src = REGNO (operands[1]);
-
- if (dest != src)
- return "vmov\t%P0, %e1";
- else
- return "";
-}
- [(set_attr "neon_type" "neon_bp_simple")]
-)
-
-(define_insn "neon_vget_lowv8hi"
- [(set (match_operand:V4HI 0 "s_register_operand" "=w")
- (vec_select:V4HI (match_operand:V8HI 1 "s_register_operand" "w")
- (parallel [(const_int 0) (const_int 1)
- (const_int 2) (const_int 3)])))]
- "TARGET_NEON"
-{
- int dest = REGNO (operands[0]);
- int src = REGNO (operands[1]);
-
- if (dest != src)
- return "vmov\t%P0, %e1";
- else
- return "";
-}
- [(set_attr "neon_type" "neon_bp_simple")]
-)
-
-(define_insn "neon_vget_lowv4si"
- [(set (match_operand:V2SI 0 "s_register_operand" "=w")
- (vec_select:V2SI (match_operand:V4SI 1 "s_register_operand" "w")
- (parallel [(const_int 0) (const_int 1)])))]
- "TARGET_NEON"
-{
- int dest = REGNO (operands[0]);
- int src = REGNO (operands[1]);
-
- if (dest != src)
- return "vmov\t%P0, %e1";
- else
- return "";
-}
- [(set_attr "neon_type" "neon_bp_simple")]
-)
-
-(define_insn "neon_vget_lowv4sf"
- [(set (match_operand:V2SF 0 "s_register_operand" "=w")
- (vec_select:V2SF (match_operand:V4SF 1 "s_register_operand" "w")
- (parallel [(const_int 0) (const_int 1)])))]
+(define_expand "neon_vget_high<mode>"
+ [(match_operand:<V_HALF> 0 "s_register_operand")
+ (match_operand:VQX 1 "s_register_operand")]
"TARGET_NEON"
{
- int dest = REGNO (operands[0]);
- int src = REGNO (operands[1]);
-
- if (dest != src)
- return "vmov\t%P0, %e1";
- else
- return "";
-}
- [(set_attr "neon_type" "neon_bp_simple")]
-)
+ emit_move_insn (operands[0],
+ simplify_gen_subreg (<V_HALF>mode, operands[1], <MODE>mode,
+ GET_MODE_SIZE (<V_HALF>mode)));
+ DONE;
+})
-(define_insn "neon_vget_lowv2di"
- [(set (match_operand:DI 0 "s_register_operand" "=w")
- (vec_select:DI (match_operand:V2DI 1 "s_register_operand" "w")
- (parallel [(const_int 0)])))]
+(define_expand "neon_vget_low<mode>"
+ [(match_operand:<V_HALF> 0 "s_register_operand")
+ (match_operand:VQX 1 "s_register_operand")]
"TARGET_NEON"
{
- int dest = REGNO (operands[0]);
- int src = REGNO (operands[1]);
-
- if (dest != src)
- return "vmov\t%P0, %e1";
- else
- return "";
-}
- [(set_attr "neon_type" "neon_bp_simple")]
-)
+ emit_move_insn (operands[0],
+ simplify_gen_subreg (<V_HALF>mode, operands[1],
+ <MODE>mode, 0));
+ DONE;
+})
(define_insn "neon_vcvt<mode>"
[(set (match_operand:<V_CVTTO> 0 "s_register_operand" "=w")
