diff options
author | Lasse Collin <lasse.collin@tukaani.org> | 2015-01-26 20:39:28 +0200 |
---|---|---|
committer | Lasse Collin <lasse.collin@tukaani.org> | 2015-01-26 20:39:28 +0200 |
commit | fec88d41e672d9e197c9442aecf02bd0dfa6d516 (patch) | |
tree | 513cd95bd616cfe19b86ba3faba35246d8466ae4 /src/liblzma/lz/lz_encoder.c | |
parent | xz: Fix comments. (diff) | |
download | xz-fec88d41e672d9e197c9442aecf02bd0dfa6d516.tar.xz |
liblzma: Silence harmless Valgrind errors.
Thanks to Torsten Rupp for reporting this. I had
forgotten to run Valgrind before the 5.2.0 release.
Diffstat (limited to '')
-rw-r--r-- | src/liblzma/lz/lz_encoder.c | 6 |
1 files changed, 6 insertions, 0 deletions
diff --git a/src/liblzma/lz/lz_encoder.c b/src/liblzma/lz/lz_encoder.c index 2033844f..01dfc06f 100644 --- a/src/liblzma/lz/lz_encoder.c +++ b/src/liblzma/lz/lz_encoder.c @@ -110,6 +110,12 @@ fill_window(lzma_coder *coder, const lzma_allocator *allocator, coder->mf.write_pos = write_pos; + // Silence Valgrind. lzma_memcmplen() can read extra bytes + // and Valgrind will give warnings if those bytes are uninitialized + // because Valgrind cannot see that the values of the uninitialized + // bytes are eventually ignored. + memzero(coder->mf.buffer + write_pos, LZMA_MEMCMPLEN_EXTRA); + // If end of stream has been reached or flushing completed, we allow // the encoder to process all the input (that is, read_pos is allowed // to reach write_pos). Otherwise we keep keep_size_after bytes |