diff options
| author | Florian Hahn <flo@fhahn.com> | 2025-10-28 19:03:17 -0700 |
|---|---|---|
| committer | GitHub <noreply@github.com> | 2025-10-28 19:03:17 -0700 |
| commit | 53785846aa11bdecbbb683e1ebec9461539f1e97 (patch) | |
| tree | 6239e330ec352272861dbcd590745b581ab32a58 /lldb/unittests/ScriptInterpreter/Python/PythonTestSuite.cpp | |
| parent | eb40a0150251c724d5c51bef85ef8d9d59d5c462 (diff) | |
| download | llvm-53785846aa11bdecbbb683e1ebec9461539f1e97.zip llvm-53785846aa11bdecbbb683e1ebec9461539f1e97.tar.gz llvm-53785846aa11bdecbbb683e1ebec9461539f1e97.tar.bz2 | |
[Clang] Freeze padded vectors before storing. (#164821)
Currently Clang usually leaves padding bits uninitialized, which means
they are undef at the moment.
When expanding stores of vector types to include padding, the padding
lanes will be poison, hence the padding bits will be poison.
This interacts badly with coercion of arguments and return values, where
3 x float vectors will be loaded as i128 integer; poisoning the padding
bits will make the whole value poison.
Not sure if there's a better way, but I think we have a number of places
that currently rely on the padding being undef, not poison.
PR: https://github.com/llvm/llvm-project/pull/164821
Diffstat (limited to 'lldb/unittests/ScriptInterpreter/Python/PythonTestSuite.cpp')
0 files changed, 0 insertions, 0 deletions
