6502/llvm-6502
1
0
Fork 0
mirror of https://github.com/c64scene-ar/llvm-6502.git synced 2024-08-22 10:29:35 +00:00
Code Issues Projects Releases Wiki Activity

R600: Take alignment into account when calculating the stack offset

Browse Source 
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@199826 91177308-0d34-0410-b5e6-96231b3b80d8
This commit is contained in:
Tom Stellard 2014-01-22 19:24:23 +00:00
parent 9c3e0ede1d
commit e7d4e83702
2 changed files with 42 additions and 4 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

15
lib/Target/R600/AMDGPUFrameLowering.cpp
View File

@ -77,14 +77,21 @@ int AMDGPUFrameLowering::getFrameIndexOffset(const MachineFunction &MF,
// Start the offset at 2 so we don't overwrite work group information.
// XXX: We should only do this when the shader actually uses this
// information.
unsigned Offset = 2;
unsigned OffsetBytes = 2 * (getStackWidth(MF) * 4);
int UpperBound = FI == -1 ? MFI->getNumObjects() : FI;
for (int i = MFI->getObjectIndexBegin(); i < UpperBound; ++i) {
unsigned Size = MFI->getObjectSize(i);
Offset += (Size / (getStackWidth(MF) * 4));
OffsetBytes = RoundUpToAlignment(OffsetBytes, MFI->getObjectAlignment(i));
OffsetBytes += MFI->getObjectSize(i);
// Each regiter holds 4 bytes, so we must always align the offset to at
// least 4 bytes, so that 2 frame objects won't share the same register.
OffsetBytes = RoundUpToAlignment(OffsetBytes, 4);
}
return Offset;
if (FI != -1)
OffsetBytes = RoundUpToAlignment(OffsetBytes, MFI->getObjectAlignment(FI));
return OffsetBytes / (getStackWidth(MF) * 4);
}
const TargetFrameLowering::SpillSlot *

31
test/CodeGen/R600/private-memory.ll
View File

@ -180,4 +180,35 @@ entry:
ret void
}
; Test that two stack objects are not stored in the same register
; The second stack object should be in T3.X
; FUNC-LABEL: @no_overlap
; R600-CHECK: MOV {{\** *}}T3.X
; SI-CHECK: V_MOV_B32_e32 v3
define void @no_overlap(i32 addrspace(1)* %out, i32 %in) {
entry:
%0 = alloca [3 x i8], align 1
%1 = alloca [2 x i8], align 1
%2 = getelementptr [3 x i8]* %0, i32 0, i32 0
%3 = getelementptr [3 x i8]* %0, i32 0, i32 1
%4 = getelementptr [3 x i8]* %0, i32 0, i32 2
%5 = getelementptr [2 x i8]* %1, i32 0, i32 0
%6 = getelementptr [2 x i8]* %1, i32 0, i32 1
store i8 0, i8* %2
store i8 1, i8* %3
store i8 2, i8* %4
store i8 1, i8* %5
store i8 0, i8* %6
%7 = getelementptr [3 x i8]* %0, i32 0, i32 %in
%8 = getelementptr [2 x i8]* %1, i32 0, i32 %in
%9 = load i8* %7
%10 = load i8* %8
%11 = add i8 %9, %10
%12 = sext i8 %11 to i32
store i32 %12, i32 addrspace(1)* %out
ret void
}
declare i32 @llvm.r600.read.tidig.x() nounwind readnone