summaryrefslogtreecommitdiffstats
path: root/llvm/lib/Target/ARM
diff options
context:
space:
mode:
authorOwen Anderson <resistor@mac.com>2010-10-27 21:28:09 +0000
committerOwen Anderson <resistor@mac.com>2010-10-27 21:28:09 +0000
commited9652f9593134d2e09690f3baf52a01e6cc4fb2 (patch)
tree3a91781b88c448347f17de851c6ec1fdafe67a93 /llvm/lib/Target/ARM
parentf8537165bd32cf43f53e131e1c4031a4cd7312bf (diff)
downloadbcm5719-llvm-ed9652f9593134d2e09690f3baf52a01e6cc4fb2.tar.gz
bcm5719-llvm-ed9652f9593134d2e09690f3baf52a01e6cc4fb2.zip
Provide correct encodings for the get_lane and set_lane variants of vmov.
llvm-svn: 117495
Diffstat (limited to 'llvm/lib/Target/ARM')
-rw-r--r--llvm/lib/Target/ARM/ARMInstrFormats.td11
-rw-r--r--llvm/lib/Target/ARM/ARMInstrNEON.td94
2 files changed, 64 insertions, 41 deletions
diff --git a/llvm/lib/Target/ARM/ARMInstrFormats.td b/llvm/lib/Target/ARM/ARMInstrFormats.td
index 73efdeb1756..891c2976f5f 100644
--- a/llvm/lib/Target/ARM/ARMInstrFormats.td
+++ b/llvm/lib/Target/ARM/ARMInstrFormats.td
@@ -1868,14 +1868,15 @@ class NVLaneOp<bits<8> opcod1, bits<4> opcod2, bits<2> opcod3,
let Pattern = pattern;
list<Predicate> Predicates = [HasNEON];
- bits<5> Vd;
- bits<4> Rt;
+ bits<5> V;
+ bits<4> R;
bits<4> p;
+ bits<4> lane;
let Inst{31-28} = p{3-0};
- let Inst{7} = Vd{4};
- let Inst{19-16} = Vd{3-0};
- let Inst{15-12} = Rt{3-0};
+ let Inst{7} = V{4};
+ let Inst{19-16} = V{3-0};
+ let Inst{15-12} = R{3-0};
}
class NVGetLane<bits<8> opcod1, bits<4> opcod2, bits<2> opcod3,
dag oops, dag iops, InstrItinClass itin,
diff --git a/llvm/lib/Target/ARM/ARMInstrNEON.td b/llvm/lib/Target/ARM/ARMInstrNEON.td
index 4e81265b19c..a746049a30c 100644
--- a/llvm/lib/Target/ARM/ARMInstrNEON.td
+++ b/llvm/lib/Target/ARM/ARMInstrNEON.td
@@ -3501,30 +3501,44 @@ def VMOVv2i64 : N1ModImm<1, 0b000, 0b1110, 0, 1, 1, 1, (outs QPR:$dst),
// VMOV : Vector Get Lane (move scalar to ARM core register)
def VGETLNs8 : NVGetLane<{1,1,1,0,0,1,?,1}, 0b1011, {?,?},
- (outs GPR:$dst), (ins DPR:$src, nohash_imm:$lane),
- IIC_VMOVSI, "vmov", "s8", "$dst, $src[$lane]",
- [(set GPR:$dst, (NEONvgetlanes (v8i8 DPR:$src),
- imm:$lane))]>;
+ (outs GPR:$R), (ins DPR:$V, nohash_imm:$lane),
+ IIC_VMOVSI, "vmov", "s8", "$R, $V[$lane]",
+ [(set GPR:$R, (NEONvgetlanes (v8i8 DPR:$V),
+ imm:$lane))]> {
+ let Inst{21} = lane{2};
+ let Inst{6-5} = lane{1-0};
+}
def VGETLNs16 : NVGetLane<{1,1,1,0,0,0,?,1}, 0b1011, {?,1},
- (outs GPR:$dst), (ins DPR:$src, nohash_imm:$lane),
- IIC_VMOVSI, "vmov", "s16", "$dst, $src[$lane]",
- [(set GPR:$dst, (NEONvgetlanes (v4i16 DPR:$src),
- imm:$lane))]>;
+ (outs GPR:$R), (ins DPR:$V, nohash_imm:$lane),
+ IIC_VMOVSI, "vmov", "s16", "$R, $V[$lane]",
+ [(set GPR:$R, (NEONvgetlanes (v4i16 DPR:$V),
+ imm:$lane))]> {
+ let Inst{21} = lane{1};
+ let Inst{6} = lane{0};
+}
def VGETLNu8 : NVGetLane<{1,1,1,0,1,1,?,1}, 0b1011, {?,?},
- (outs GPR:$dst), (ins DPR:$src, nohash_imm:$lane),
- IIC_VMOVSI, "vmov", "u8", "$dst, $src[$lane]",
- [(set GPR:$dst, (NEONvgetlaneu (v8i8 DPR:$src),
- imm:$lane))]>;
+ (outs GPR:$R), (ins DPR:$V, nohash_imm:$lane),
+ IIC_VMOVSI, "vmov", "u8", "$R, $V[$lane]",
+ [(set GPR:$R, (NEONvgetlaneu (v8i8 DPR:$V),
+ imm:$lane))]> {
+ let Inst{21} = lane{2};
+ let Inst{6-5} = lane{1-0};
+}
def VGETLNu16 : NVGetLane<{1,1,1,0,1,0,?,1}, 0b1011, {?,1},
- (outs GPR:$dst), (ins DPR:$src, nohash_imm:$lane),
- IIC_VMOVSI, "vmov", "u16", "$dst, $src[$lane]",
- [(set GPR:$dst, (NEONvgetlaneu (v4i16 DPR:$src),
- imm:$lane))]>;
+ (outs GPR:$R), (ins DPR:$V, nohash_imm:$lane),
+ IIC_VMOVSI, "vmov", "u16", "$R, $V[$lane]",
+ [(set GPR:$R, (NEONvgetlaneu (v4i16 DPR:$V),
+ imm:$lane))]> {
+ let Inst{21} = lane{1};
+ let Inst{6} = lane{0};
+}
def VGETLNi32 : NVGetLane<{1,1,1,0,0,0,?,1}, 0b1011, 0b00,
- (outs GPR:$dst), (ins DPR:$src, nohash_imm:$lane),
- IIC_VMOVSI, "vmov", "32", "$dst, $src[$lane]",
- [(set GPR:$dst, (extractelt (v2i32 DPR:$src),
- imm:$lane))]>;
+ (outs GPR:$R), (ins DPR:$V, nohash_imm:$lane),
+ IIC_VMOVSI, "vmov", "32", "$R, $V[$lane]",
+ [(set GPR:$R, (extractelt (v2i32 DPR:$V),
+ imm:$lane))]> {
+ let Inst{21} = lane{0};
+}
// def VGETLNf32: see FMRDH and FMRDL in ARMInstrVFP.td
def : Pat<(NEONvgetlanes (v16i8 QPR:$src), imm:$lane),
(VGETLNs8 (v8i8 (EXTRACT_SUBREG QPR:$src,
@@ -3560,22 +3574,30 @@ def : Pat<(extractelt (v2f64 QPR:$src1), imm:$src2),
// VMOV : Vector Set Lane (move ARM core register to scalar)
-let Constraints = "$src1 = $dst" in {
-def VSETLNi8 : NVSetLane<{1,1,1,0,0,1,?,0}, 0b1011, {?,?}, (outs DPR:$dst),
- (ins DPR:$src1, GPR:$src2, nohash_imm:$lane),
- IIC_VMOVISL, "vmov", "8", "$dst[$lane], $src2",
- [(set DPR:$dst, (vector_insert (v8i8 DPR:$src1),
- GPR:$src2, imm:$lane))]>;
-def VSETLNi16 : NVSetLane<{1,1,1,0,0,0,?,0}, 0b1011, {?,1}, (outs DPR:$dst),
- (ins DPR:$src1, GPR:$src2, nohash_imm:$lane),
- IIC_VMOVISL, "vmov", "16", "$dst[$lane], $src2",
- [(set DPR:$dst, (vector_insert (v4i16 DPR:$src1),
- GPR:$src2, imm:$lane))]>;
-def VSETLNi32 : NVSetLane<{1,1,1,0,0,0,?,0}, 0b1011, 0b00, (outs DPR:$dst),
- (ins DPR:$src1, GPR:$src2, nohash_imm:$lane),
- IIC_VMOVISL, "vmov", "32", "$dst[$lane], $src2",
- [(set DPR:$dst, (insertelt (v2i32 DPR:$src1),
- GPR:$src2, imm:$lane))]>;
+let Constraints = "$src1 = $V" in {
+def VSETLNi8 : NVSetLane<{1,1,1,0,0,1,?,0}, 0b1011, {?,?}, (outs DPR:$V),
+ (ins DPR:$src1, GPR:$R, nohash_imm:$lane),
+ IIC_VMOVISL, "vmov", "8", "$V[$lane], $R",
+ [(set DPR:$V, (vector_insert (v8i8 DPR:$src1),
+ GPR:$R, imm:$lane))]> {
+ let Inst{21} = lane{2};
+ let Inst{6-5} = lane{1-0};
+}
+def VSETLNi16 : NVSetLane<{1,1,1,0,0,0,?,0}, 0b1011, {?,1}, (outs DPR:$V),
+ (ins DPR:$src1, GPR:$R, nohash_imm:$lane),
+ IIC_VMOVISL, "vmov", "16", "$V[$lane], $R",
+ [(set DPR:$V, (vector_insert (v4i16 DPR:$src1),
+ GPR:$R, imm:$lane))]> {
+ let Inst{21} = lane{1};
+ let Inst{6} = lane{0};
+}
+def VSETLNi32 : NVSetLane<{1,1,1,0,0,0,?,0}, 0b1011, 0b00, (outs DPR:$V),
+ (ins DPR:$src1, GPR:$R, nohash_imm:$lane),
+ IIC_VMOVISL, "vmov", "32", "$V[$lane], $R",
+ [(set DPR:$V, (insertelt (v2i32 DPR:$src1),
+ GPR:$R, imm:$lane))]> {
+ let Inst{21} = lane{0};
+}
}
def : Pat<(vector_insert (v16i8 QPR:$src1), GPR:$src2, imm:$lane),
(v16i8 (INSERT_SUBREG QPR:$src1,
OpenPOWER on IntegriCloud