mirror of
https://github.com/RPCS3/llvm-mirror.git
synced 2024-11-26 04:32:44 +01:00
3be79b1237
This was (partially) reverted in cfe8f8e0 because the conversion from readonly to readnone in Intrinsics.td exposed a couple of problems. This change has been reworked to not need that change (via some explicit checks in client code). This is being done to address the original optimization issue and simplify the testing of the readonly changes. I'm working on that piece under 49607. Original commit message follows: The last two operands to a gc.relocate represent indices into the associated gc.statepoint's gc bundle list. (Effectively, gc.relocates are projections from the gc.statepoints multiple return values.) We can use this to recognize when two gc.relocates are equivalent (and can be CSEd), even when the indices are non-equal. This is particular useful when considering a chain of multiple statepoints as it lets us eliminate all duplicate gc.relocates in a single pass. Differential Revision: https://reviews.llvm.org/D97974 |
||
---|---|---|
.. | ||
AArch64 | ||
AMDGPU | ||
PowerPC | ||
X86 | ||
and_or.ll | ||
atomics.ll | ||
basic.ll | ||
commute.ll | ||
conditional.ll | ||
const-speculation.ll | ||
debug-info-undef.ll | ||
debuginfo-dce.ll | ||
edge.ll | ||
fence.ll | ||
flags.ll | ||
floatingpoint.ll | ||
gc_relocate.ll | ||
getmatchingvalue-crash.ll | ||
globalsaa-memoryssa.ll | ||
guards.ll | ||
instsimplify-dom.ll | ||
int_sideeffect.ll | ||
invariant-loads.ll | ||
invariant.start.ll | ||
masked-intrinsics-unequal-masks.ll | ||
masked-intrinsics.ll | ||
memoryssa.ll | ||
noalias-scope-decl.ll | ||
phi.ll | ||
pr33406.ll | ||
readnone-mayunwind.ll | ||
reuse-preserved-memoryssa.ll | ||
writeonly.ll |