maxjhandsome
/
llvm-project
forked from OSchip/llvm-project
1
0
Fork 0
Code Issues Pull Requests Packages Releases Wiki Activity

[AMDGPU][MC] Fix for Bug 28167 + LIT tests 

Corrected src0 for v_writelane_b32:
- Enabled inline constants and literals for SI/CI (VOP2)
- Enabled inline constants for VI (VOP3)

Reviewers: vpykhtin, arsenm

https://reviews.llvm.org/D31463

llvm-svn: 299555
This commit is contained in:
Dmitry Preobrazhensky 2017-04-05 16:08:21 +00:00
parent 92a5ba6da5
commit 45db65037f
2 changed files with 12 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

5
llvm/lib/Target/AMDGPU/VOP2Instructions.td
View File

@ -283,7 +283,7 @@ def VOP_READLANE : VOPProfile<[i32, i32, i32]> {
def VOP_WRITELANE : VOPProfile<[i32, i32, i32]> {
let Outs32 = (outs VGPR_32:$vdst);
let Outs64 = Outs32;
let Ins32 = (ins SReg_32:$src0, SCSrc_b32:$src1);
let Ins32 = (ins SCSrc_b32:$src0, SCSrc_b32:$src1);
let Ins64 = Ins32;
let Asm32 = " $vdst, $src0, $src1";
let Asm64 = Asm32;
@ -577,7 +577,10 @@ defm V_SUBB_U32 : VOP2be_Real_e32e64_si <0x29>;
defm V_SUBBREV_U32 : VOP2be_Real_e32e64_si <0x2a>;
defm V_READLANE_B32 : VOP2_Real_si <0x01>;
let InOperandList = (ins SSrc_b32:$src0, SCSrc_b32:$src1) in {
defm V_WRITELANE_B32 : VOP2_Real_si <0x02>;
}
defm V_MAC_LEGACY_F32 : VOP2_Real_e32e64_si <0x6>;
defm V_MIN_LEGACY_F32 : VOP2_Real_e32e64_si <0xd>;

8
llvm/test/MC/AMDGPU/vop2.s
View File

@ -120,6 +120,14 @@ v_readlane_b32 s1, v2, s3
// VI: v_writelane_b32 v1, s2, 4 ; encoding: [0x01,0x00,0x8a,0xd2,0x02,0x08,0x01,0x00]
v_writelane_b32 v1, s2, 4
// SICI: v_writelane_b32 v2, 1, s4 ; encoding: [0x81,0x08,0x04,0x04]
// VI: v_writelane_b32 v2, 1, s4 ; encoding: [0x02,0x00,0x8a,0xd2,0x81,0x08,0x00,0x00]
v_writelane_b32 v2, 1, s4
// SICI: v_writelane_b32 v255, 0xaf123456, 2 ; encoding: [0xff,0x04,0xff,0x05,0x56,0x34,0x12,0xaf]
// NOVI: error: instruction not supported on this GPU
v_writelane_b32 v255, 0xaf123456, 2
// SICI: v_add_f32_e32 v1, v2, v3 ; encoding: [0x02,0x07,0x02,0x06]
// VI: v_add_f32_e32 v1, v2, v3 ; encoding: [0x02,0x07,0x02,0x02]
v_add_f32 v1, v2, v3