diff options
| author | Pavel Labath <pavel@labath.sk> | 2020-04-14 17:06:04 +0200 |
|---|---|---|
| committer | Pavel Labath <pavel@labath.sk> | 2020-04-15 12:47:57 +0200 |
| commit | 122c50d5efc2d5bc8d6716a532ba0cc6b773ab3a (patch) | |
| tree | a336fb66c26726932419d2b05bed407c8ee1e894 /llvm/lib/Transforms/Utils/LoopUtils.cpp | |
| parent | cf9ee49b4d7f2e1341b84f941ad55cce1b16722d (diff) | |
| download | llvm-122c50d5efc2d5bc8d6716a532ba0cc6b773ab3a.zip llvm-122c50d5efc2d5bc8d6716a532ba0cc6b773ab3a.tar.gz llvm-122c50d5efc2d5bc8d6716a532ba0cc6b773ab3a.tar.bz2 | |
Fix DWARFDataExtractor::getRelocatedValue near EOF
Summary:
If we have an (invalid) relocation which relocates bytes which partially
lie outside the range of the relocated section, the getRelocatedValue
would return confusing results. It would first read zero (because that's
what the underlying DataExtractor api does for out-of-bounds reads), and
then relocate that zero anyway.
A more appropriate behavior is to return zero straight away. This is
what this patch does.
Reviewers: dblaikie, jhenderson
Subscribers: hiraditya, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D78113
Diffstat (limited to 'llvm/lib/Transforms/Utils/LoopUtils.cpp')
0 files changed, 0 insertions, 0 deletions