fdo-mirrors/mesa
1
0
Fork 0
mirror of https://gitlab.freedesktop.org/mesa/mesa.git synced 2025-12-24 15:20:10 +01:00
Code Issues Projects Releases Packages Wiki Activity Actions 1

nv50/ir: add intermediate conversion for f2{i,u}{8,16}

Browse source 
Directly converting from a float to an 8 bit integer and from a 64 bit
float to an integer smaller than 32 bit is not supported, therefore add
an intermediate conversion to an 32 bit integer.

Reviewed-by: Karol Herbst <kherbst@redhat.com>
Signed-off-by: Danilo Krummrich <dakr@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18109>
This commit is contained in:
Danilo Krummrich 2022-08-16 02:03:02 +02:00
parent 6a9825bc1b
commit 8ccba4ea5c
1 changed files with 16 additions and 1 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

17
src/nouveau/codegen/nv50_ir_lowering_helper.cpp
View file

@ -85,8 +85,23 @@ LoweringHelper::handleCVT(Instruction *insn)
DataType dTy = insn->dType;
DataType sTy = insn->sType;
if (typeSizeof(dTy) <= 4 && typeSizeof(sTy) <= 4)
bld.setPosition(insn, true);
/* We can't convert from 32bit floating point to 8bit integer and from 64bit
* floating point to any integer smaller than 32bit, hence add an instruction
* to convert to a 32bit integer first.
*/
if (((typeSizeof(dTy) == 1) && isFloatType(sTy)) ||
((typeSizeof(dTy) <= 2) && sTy == TYPE_F64)) {
Value *tmp = insn->getDef(0);
DataType tmpTy = (isSignedIntType(dTy)) ? TYPE_S32 : TYPE_U32;
insn->setType(tmpTy, sTy);
insn->setDef(0, bld.getSSA());
bld.mkCvt(OP_CVT, dTy, tmp, tmpTy, insn->getDef(0))->saturate = 1;
return true;
}
bld.setPosition(insn, false);