3621 Oystein Grovlen 2011-11-17
Bug#12876523: VALGRIND FAILURE IN COMPRESSION TESTS
The valgrind issue happens in the zlib package,
and according to the zlib FAQ, it is intentional for
performance reasons (http://www.zlib.net/zlib_faq.html#faq36).
The current valgrind suppression file contains several call
stacks for valgrind issues with deflate(). However, all the
existing stacks differ from the one reported here at the
4th level.
I assume the failing stack trace must be due to inlining
in optimized builds, but I have not been able to reproduce
it on my available platforms. However, I see no reason that
the suppression rules should distinguish on which MySQL
function call deflate() when there is a known problem with
how deflate() works. By limiting the stack in the suppression
file to 3 levels, we can also reduce the number of suppression
rules for deflate().
@ mysql-test/valgrind.supp
Make a general suppression rule for deflate() so that
it also covers some call stacks that only appear in optimized builds on some platforms.
modified:
mysql-test/valgrind.supp
3620 Rafal Somla 2011-11-17 [merge]
Merge of bug#13101974
modified:
libmysql/CMakeLists.txt
sql/CMakeLists.txt
=== modified file 'mysql-test/valgrind.supp'
--- a/mysql-test/valgrind.supp 2011-11-08 14:48:33 +0000
+++ b/mysql-test/valgrind.supp 2011-11-17 14:07:41 +0000
@@ -187,70 +187,6 @@
fun:longest_match
fun:deflate_slow
fun:deflate
- fun:do_flush
-}
-
-{
- libz longest_match called from btr_store_big_rec_extern_fields
- Memcheck:Cond
- fun:longest_match
- fun:deflate_slow
- fun:deflate
- fun:btr_store_big_rec_extern_fields
-}
-
-{
- libz longest_match called from page_zip_compress
- Memcheck:Cond
- fun:longest_match
- fun:deflate_slow
- fun:deflate
- fun:page_zip_compress
-}
-
-{
- libz longest_match2
- Memcheck:Cond
- fun:longest_match
- fun:deflate_slow
- fun:deflate
- fun:compress2
-}
-
-{
- libz longest_match 3
- Memcheck:Cond
- fun:longest_match
- fun:deflate_slow
- fun:deflate
- fun:gzclose
-}
-
-{
- libz longest_match 4
- Memcheck:Cond
- fun:longest_match
- fun:deflate_slow
- fun:deflate
- fun:gzflush
-}
-
-{
- libz longest_match3
- Memcheck:Cond
- fun:longest_match
- fun:deflate_slow
- fun:deflate
- fun:azflush
-}
-
-{
- libz longest_match3
- Memcheck:Cond
- fun:longest_match
- fun:deflate_slow
- fun:deflate
- fun:azclose
}
{
No bundle (reason: useless for push emails).
| Thread |
|---|
| • bzr push into mysql-trunk branch (oystein.grovlen:3620 to 3621) Bug#12876523 | Oystein Grovlen | 18 Nov |