http://bugzilla.novell.com/show_bug.cgi?id=509398 User pbaudis@novell.com added comment http://bugzilla.novell.com/show_bug.cgi?id=509398#c1 Petr Baudis changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |NEEDINFO Info Provider| |wolfgang@rosenauer.org --- Comment #1 from Petr Baudis 2009-06-03 04:39:33 MDT --- (i) Is Mozilla's custom malloc/free override used at this point in the testsuite? (ii) Does the override use glibc malloc/free as a backend? (iii) Can you paste some backtraces? Is there a pattern in them? (Always crashes in the same free(), or so.) Thanks. -- Configure bugmail: http://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

http://bugzilla.novell.com/show_bug.cgi?id=509398 User wolfgang@rosenauer.org added comment http://bugzilla.novell.com/show_bug.cgi?id=509398#c3 Wolfgang Rosenauer changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEEDINFO |NEW Info Provider|wolfgang@rosenauer.org | --- Comment #3 from Wolfgang Rosenauer 2009-06-03 06:57:17 MDT --- Here are some: #0 0x00007f44cde54875 in raise () from /lib64/libc.so.6 #1 0x00007f44cde55d8a in abort () from /lib64/libc.so.6 #2 0x00007f44cde9001d in ?? () from /lib64/libc.so.6 #3 0x00007f44cde95776 in ?? () from /lib64/libc.so.6 #4 0x0000000000407b91 in printSecurityInfo (fd=<value optimized out>) at selfserv.c:404 #5 0x0000000000407fcb in handle_connection (tcp_sock=<value optimized out>, model_sock=<value optimized out>, requestCert=<value optimized out>) at selfserv.c:1062 #6 0x0000000000408904 in jobLoop (a=<value optimized out>, b=<value optimized out>, c=<value optimized out>) at selfserv.c:527 #7 0x0000000000407211 in thread_wrapper (arg=0x6880e0) at selfserv.c:496 #0 0x00007f242ac61875 in raise () from /lib64/libc.so.6 #1 0x00007f242ac62d8a in abort () from /lib64/libc.so.6 #2 0x00007f242ac9d01d in ?? () from /lib64/libc.so.6 #3 0x00007f242aca2776 in ?? () from /lib64/libc.so.6 #4 0x00007f242c193142 in ssl_DestroySocketContents (ss=0x75ce60) at sslsock.c:403 #5 0x00007f242c194902 in ssl_FreeSocket (ss=0x75ce60) at sslsock.c:466 #6 0x00007f242c18bed8 in ssl_DefClose (ss=0x75ce60) at ssldef.c:233 #7 0x0000000000407f08 in handle_connection (tcp_sock=<value optimized out>, model_sock=<value optimized out>, requestCert=<value optimized out>) at selfserv.c:1283 #8 0x0000000000408904 in jobLoop (a=<value optimized out>, b=<value optimized out>, c=<value optimized out>) at selfserv.c:527 #9 0x0000000000407211 in thread_wrapper (arg=0x6a9e00) at selfserv.c:496 #10 0x00007f242b3d5b83 in ?? () from /usr/lib64/libnspr4.so #11 0x00007f242b19564d in start_thread () from /lib64/libpthread.so.0 #12 0x00007f242ad013fd in clone () from /lib64/libc.so.6 #0 0x00007f3765b7e875 in raise () from /lib64/libc.so.6 #1 0x00007f3765b7fd8a in abort () from /lib64/libc.so.6 #2 0x00007f3765bba01d in ?? () from /lib64/libc.so.6 #3 0x00007f3765bbf776 in ?? () from /lib64/libc.so.6 #4 0x00007f37670a8ec5 in ssl_DefClose (ss=0x6da940) at ssldef.c:226 #5 0x0000000000407f08 in handle_connection (tcp_sock=<value optimized out>, model_sock=<value optimized out>, requestCert=<value optimized out>) at selfserv.c:1283 #6 0x0000000000408904 in jobLoop (a=<value optimized out>, b=<value optimized out>, c=<value optimized out>) at selfserv.c:527 #7 0x0000000000407211 in thread_wrapper (arg=0x688110) at selfserv.c:496 #8 0x00007f37662f2b83 in ?? () from /usr/lib64/libnspr4.so #9 0x00007f37660b264d in start_thread () from /lib64/libpthread.so.0 #10 0x00007f3765c1e3fd in clone () from /lib64/libc.so.6 #0 0x00007f692e39f875 in raise () from /lib64/libc.so.6 #1 0x00007f692e3a0d8a in abort () from /lib64/libc.so.6 #2 0x00007f692e3db01d in ?? () from /lib64/libc.so.6 #3 0x00007f692e3e0776 in ?? () from /lib64/libc.so.6 #4 0x00007f692f383b34 in PK11_DestroyContext (context=0x7ed6d0, freeit=1) at pk11cxt.c:100 #5 0x00007f692f384aa4 in PK11_HashBuf (hashAlg=SEC_OID_MD5, out=0x7dfe08 "\223&\266\26>%\310\71\360\t\216f\353~\27\245l\247\251zH&?\317\66\324t(\332\4\aD)\204\261\207\354\260&Y\216Ϻ\244\36QGʠ\344", , in=0x854983 "0p1\v0\t\6\3U\4\6\23\2US1\23\60\21\6\3U\4\b\23\nCalifornia1\24\60\22\6\3U\4\a\23\vSanta Clara1\22\60\20\6\3U\4\n\23\tBOGUS NSS1\"0 \6\3U\4\3\23\31NSS Chain2 Client Test CA0\201\237\60\r\6\t*\206H\206\367\r\1\1\1\5", len=114) at pk11cxt.c:661 #6 0x00007f692f36b8d5 in ocsp_DigestValue (arena=0x800710, digestAlg=SEC_OID_MD5, fill=0x7dfd60, src=0x7f692b790f50) at ocsp.c:1594 #7 0x00007f692f36d630 in cert_GetSubjectNameDigest (fill=<value optimized out>, digestAlg=<value optimized out>, cert=<value optimized out>, arena=<value optimized out>) at ocsp.c:1658 #8 ocsp_CreateCertID (fill=<value optimized out>, digestAlg=<value optimized out>, cert=<value optimized out>, arena=<value optimized out>) at ocsp.c:1704 #0 0x00007f8ac97d3875 in raise () from /lib64/libc.so.6 #1 0x00007f8ac97d4d8a in abort () from /lib64/libc.so.6 #2 0x00007f8ac980f01d in ?? () from /lib64/libc.so.6 #3 0x00007f8ac9814776 in ?? () from /lib64/libc.so.6 #4 0x00007f8ac95826d3 in sftk_DestroySession (session=<value optimized out>) at pkcs11u.c:1844 #5 sftk_FreeSession (session=<value optimized out>) at pkcs11u.c:1893 #6 0x00007f8ac9570c5b in NSC_CloseSession (hSession=0) at pkcs11.c:3463 #7 0x00007f8aca7c9cdc in pk11_CloseSession (slot=0x670000, session=18267, owner=<value optimized out>) at pk11obj.c:386 #8 0x00007f8aca7b7aeb in PK11_DestroyContext (context=0x81bc90, freeit=1) at pk11cxt.c:94 #9 0x00007f8acacee3fa in ssl3_DestroySSL3Info (ss=0x7f5bc0) at ssl3con.c:8745 #10 0x00007f8acad0512a in ssl_DestroySocketContents (ss=0x7f5bc0) at sslsock.c:400 #11 0x00007f8acad06902 in ssl_FreeSocket (ss=0x7f5bc0) at sslsock.c:466 #12 0x00007f8acacfded8 in ssl_DefClose (ss=0x7f5bc0) at ssldef.c:233 #13 0x0000000000406e37 in do_connects (a=<value optimized out>, b=0x67d1f0, tid=2) at strsclnt.c:880 #14 0x0000000000405c84 in thread_wrapper (arg=<value optimized out>) at strsclnt.c:426 #15 0x00007f8ac9f47b83 in ?? () from /usr/lib64/libnspr4.so #16 0x00007f8ac9d0764d in start_thread () from /lib64/libpthread.so.0 #17 0x00007f8ac98733fd in clone () from /lib64/libc.so.6 #0 0x00007fefaec7d875 in raise () from /lib64/libc.so.6 #1 0x00007fefaec7ed8a in abort () from /lib64/libc.so.6 #2 0x00007fefaecb901d in ?? () from /lib64/libc.so.6 #3 0x00007fefaecbe776 in ?? () from /lib64/libc.so.6 #4 0x00007fefae51f68c in HMAC_Destroy (cx=0x6d1920, freeit=1) at alghmac.c:65 #5 0x00007fefae5346ae in sftk_P_hash (hashType=<value optimized out>, secret=<value optimized out>, label=0x7fefb01b90e8 "client finished", seed=0x7fefadd0d8e0, result=<value optimized out>, isFIPS=<value optimized out>) at tlsprfalg.c:118 #6 0x00007fefae5347f1 in TLS_PRF (secret=<value optimized out>, label=0x7fefb01b90e8 "client finished", seed=0x7fefadd0d8e0, result=0x7fefadd0d8c0, isFIPS=0) at tlsprfalg.c:155 #7 0x00007fefb019b1cf in ssl3_ComputeTLSFinished (spec=<value optimized out>, isServer=<value optimized out>, hashes=0x7fefadd0d950, tlsFinished=0x7fefadd0d980) at ssl3con.c:7540 #8 0x00007fefb019b3bb in ssl3_SendFinished (ss=0x6ce820, flags=1073741824) at ssl3con.c:7572 #9 0x00007fefb01a0b65 in ssl3_HandleFinished (ss=0x6ce820, b=<value optimized out>, length=<value optimized out>, hashes=<value optimized out>) at ssl3con.c:7780 -- Configure bugmail: http://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

http://bugzilla.novell.com/show_bug.cgi?id=509398 User wolfgang@rosenauer.org added comment http://bugzilla.novell.com/show_bug.cgi?id=509398#c5 Wolfgang Rosenauer changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEEDINFO |NEW Info Provider|wolfgang@rosenauer.org | --- Comment #5 from Wolfgang Rosenauer 2009-06-04 16:20:50 MDT --- I've changed the testscripts to start selfserv and strsclnt using prefix "valgrind" but the output I get doesn't really show an issue. The crashes of strsclnt (if it's running in valgrind) are gone now and the valgrind output shows nothing to me. Here is the output for a strsclnt execution which crashes if not running in valgrind: strsclnt started at Thu Jun 4 21:20:12 UTC 2009 ==28447== Memcheck, a memory error detector. ==28447== Copyright (C) 2002-2008, and GNU GPL'd, by Julian Seward et al. ==28447== Using LibVEX rev 1884, a library for dynamic binary translation. ==28447== Copyright (C) 2004-2008, and GNU GPL'd, by OpenWorks LLP. ==28447== Using valgrind-3.4.1, a dynamic binary instrumentation framework. ==28447== Copyright (C) 2000-2008, and GNU GPL'd, by Julian Seward et al. ==28447== For more details, rerun with: -v ==28447== strsclnt: -- SSL: Server Certificate Validated. strsclnt: 0 cache hits; 1 cache misses, 0 cache not reusable 0 stateless resumes strsclnt: 999 cache hits; 1 cache misses, 0 cache not reusable 999 stateless resumes ==28447== ==28447== ERROR SUMMARY: 0 errors from 0 contexts (suppressed: 43 from 2) ==28447== malloc/free: in use at exit: 27,382 bytes in 30 blocks. ==28447== malloc/free: 36,814 allocs, 36,784 frees, 170,050,981 bytes allocated. ==28447== For counts of detected errors, rerun with: -v ==28447== searching for pointers to 30 not-freed blocks. ==28447== checked 206,168 bytes. ==28447== ==28447== LEAK SUMMARY: ==28447== definitely lost: 0 bytes in 0 blocks. ==28447== possibly lost: 0 bytes in 0 blocks. ==28447== still reachable: 27,382 bytes in 30 blocks. ==28447== suppressed: 0 bytes in 0 blocks. ==28447== Rerun with --leak-check=full to see details of leaked memory. strsclnt completed at Thu Jun 4 21:20:23 UTC 2009 Some crashes in selfserv are still happening but I don't get the valgrind output in the log. Probably because it's a daemon process. Will try further. -- Configure bugmail: http://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

http://bugzilla.novell.com/show_bug.cgi?id=509398 User wolfgang@rosenauer.org added comment http://bugzilla.novell.com/show_bug.cgi?id=509398#c7 Wolfgang Rosenauer changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEEDINFO |NEW Info Provider|wolfgang@rosenauer.org | --- Comment #7 from Wolfgang Rosenauer 2009-06-05 02:06:51 MDT --- MALLOC_CHECK_ unset: I get a lot of "Aborted" for selfserv and strsclnt (around 50) MALLOC_CHECK_=0 I only get two SIGSEGV for selfserv instead of around 50 aborts (these are probably real issues?) MALLOC_CHECK_=1 The two SIGSEVs in selfserv turn into aborts again and the other almost 50 failures are SIGSEVs now Does that make sense? -- Configure bugmail: http://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

http://bugzilla.novell.com/show_bug.cgi?id=509398 User coolo@novell.com added comment http://bugzilla.novell.com/show_bug.cgi?id=509398#c9 Stephan Kulow changed: What |Removed |Added ---------------------------------------------------------------------------- Priority|P5 - None |P1 - Urgent CC| |coolo@novell.com Severity|Normal |Critical --- Comment #9 from Stephan Kulow 2009-06-05 02:33:14 MDT --- if you didn't change it, it's =3. This bug appears in tons of build logs - as valgrind fixes it, it sounds a bit too much like a problem with the malloc code - as valgrind overwrites it. -- Configure bugmail: http://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

http://bugzilla.novell.com/show_bug.cgi?id=509398 User wolfgang@rosenauer.org added comment http://bugzilla.novell.com/show_bug.cgi?id=509398#c11 Wolfgang Rosenauer changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEEDINFO |NEW Info Provider|wolfgang@rosenauer.org | --- Comment #11 from Wolfgang Rosenauer 2009-06-06 00:52:01 MDT --- I have no small testcase, sorry. But unsetting MALLOC_CHECK_ makes the errors go away entirely. -- Configure bugmail: http://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

http://bugzilla.novell.com/show_bug.cgi?id=509398 User pbaudis@novell.com added comment http://bugzilla.novell.com/show_bug.cgi?id=509398#c13 Petr Baudis changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |NEEDINFO Info Provider| |wolfgang@rosenauer.org --- Comment #13 from Petr Baudis 2009-06-08 11:45:34 MDT --- This seems to have been a race in the free() mcheck hook, introduced in glibc-2.10. salieri-pbaudis-23 fixes the issue for me (will be soon at http://www.suse.de/~pbaudis/bug-509398/). -- Configure bugmail: http://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.