diff options
| author | Nemanja Ivanovic <nemanja.i.ibm@gmail.com> | 2021-01-27 23:08:39 -0600 |
|---|---|---|
| committer | Nemanja Ivanovic <nemanja.i.ibm@gmail.com> | 2021-01-28 04:16:48 -0600 |
| commit | 54e570d94af995ff58287a8288389641910a8239 (patch) | |
| tree | 57935c571c59b995bb871b47a8cec56e98c6c9c5 /llvm/lib/Support/DAGDeltaAlgorithm.cpp | |
| parent | fc2f27ccf3c5982a8f89d2c85d83578a713c95c7 (diff) | |
| download | llvm-54e570d94af995ff58287a8288389641910a8239.zip llvm-54e570d94af995ff58287a8288389641910a8239.tar.gz llvm-54e570d94af995ff58287a8288389641910a8239.tar.bz2 | |
[PowerPC] Do not emit XXSPLTI32DX for sub 64-bit constants
If the APInt returned by BuildVectorSDNode::isConstantSplat() is narrower than
64 bits, the result produced by XXSPLTI32DX is incorrect. The result returned
by the function appears to be incorrect and we'll investigate/fix it in a
follow-up commit. However, since this causes miscompiles, we must
temporarily disable emitting this instruction for such values.
Diffstat (limited to 'llvm/lib/Support/DAGDeltaAlgorithm.cpp')
0 files changed, 0 insertions, 0 deletions