Package: libc6-bin
Version: 2.11.3-3
Severity:important
Using the function getpwuid() in my application, resulted in a memory leak. The
minimal test case is this:
$ cat c.c
#include<pwd.h>
main() {
struct passwd *q = getpwuid(geteuid());
}
$ gcc c.c&& valgrind --leak-check=full ./a.out
==22230== Memcheck, a memory error detector
==22230== Copyright (C) 2002-2010, and GNU GPL'd, by Julian Seward et al.
==22230== Using Valgrind-3.6.0.SVN-Debian and LibVEX; rerun with -h for
copyright info
==22230== Command: ./a.out
==22230==
==22230==
==22230== HEAP SUMMARY:
==22230== in use at exit: 300 bytes in 11 blocks
==22230== total heap usage: 68 allocs, 57 frees, 9,898 bytes allocated
==22230==
==22230== 300 (60 direct, 240 indirect) bytes in 1 blocks are definitely lost
in loss record 11 of 11
==22230== at 0x4C244E8: malloc (vg_replace_malloc.c:236)
==22230== by 0x4F08A3C: nss_parse_service_list (nsswitch.c:622)
==22230== by 0x4F0922D: __nss_database_lookup (nsswitch.c:164)
==22230== by 0x558D34F: ???
==22230== by 0x558DFA4: ???
==22230== by 0x4EC8D8C: getpwuid_r@@GLIBC_2.2.5 (getXXbyYY_r.c:253)
==22230== by 0x4EC867E: getpwuid (getXXbyYY.c:117)
==22230== by 0x40054C: main (in /tmp/a.out)
==22230==
==22230== LEAK SUMMARY:
==22230== definitely lost: 60 bytes in 1 blocks
==22230== indirectly lost: 240 bytes in 10 blocks
==22230== possibly lost: 0 bytes in 0 blocks
==22230== still reachable: 0 bytes in 0 blocks
==22230== suppressed: 0 bytes in 0 blocks
==22230==
==22230== For counts of detected and suppressed errors, rerun with: -v
==22230== ERROR SUMMARY: 1 errors from 1 contexts (suppressed: 4 from 4)
The bug is described here:
http://sourceware.org/bugzilla/show_bug.cgi?id=14122
The fix is here:
http://sourceware.org/git/?p=glibc.git;a=commitdiff;h=d44638b0a7fc1f01c3b2153cfa1bbb849f01f217
Since this is fixed only in glibc v2.16, and debian squeeze is using 2.11 and
wheezy uses 2.13 it would be great if this could be fixed also Debian stable
and testing.
Alois
