KonstantinSeurer
/
mesa
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

nv50/ir: pre-compute BFE arg when both bits and offset are imm 

Due to a quirk in how the nv50 opt passes run, the algebraic
optimization that looks for these BFE's happens before the constant
folding pass. Rearranging these passes isn't a great idea, but this is
easy enough to fix. Allows a following cvt to eliminate the bfe in
certain situations.

Signed-off-by: Ilia Mirkin <imirkin@alum.mit.edu>
This commit is contained in:
Ilia Mirkin 2015-08-20 22:13:48 -04:00
parent ecebd3dbfc
commit 8483577f6b
1 changed files with 9 additions and 3 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

12
src/gallium/drivers/nouveau/codegen/nv50_ir_from_tgsi.cpp
View File

@ -2990,9 +2990,15 @@ Converter::handleInstruction(const struct tgsi_full_instruction *insn)
case TGSI_OPCODE_UBFE:
FOR_EACH_DST_ENABLED_CHANNEL(0, c, tgsi) {
src0 = fetchSrc(0, c);
src1 = fetchSrc(1, c);
src2 = fetchSrc(2, c);
mkOp3(OP_INSBF, TYPE_U32, src1, src2, mkImm(0x808), src1);
if (tgsi.getSrc(1).getFile() == TGSI_FILE_IMMEDIATE &&
tgsi.getSrc(2).getFile() == TGSI_FILE_IMMEDIATE) {
src1 = loadImm(NULL, tgsi.getSrc(2).getValueU32(c, info) << 8 |
tgsi.getSrc(1).getValueU32(c, info));
} else {
src1 = fetchSrc(1, c);
src2 = fetchSrc(2, c);
mkOp3(OP_INSBF, TYPE_U32, src1, src2, mkImm(0x808), src1);
}
mkOp2(OP_EXTBF, dstTy, dst0[c], src0, src1);
}
break;