{"thread":{"id":"12030","subject":"libcrypto core dump in 64bit","startedAt":"2008-02-11T10:28:22Z","lastAt":"2008-02-13T10:57:38Z","messageCount":7,"participants":["H.Merijn Brand","Johannes Schindelin","Michal Rokos"],"isPatch":false,"patchVersion":null,"patchTotal":null},"messages":[{"id":"68365","messageId":"20080211112822.16b69495@pc09.procura.nl","threadId":"12030","inReplyTo":null,"subject":"libcrypto core dump in 64bit","fromName":"H.Merijn Brand","fromEmail":"h.m.brand@xs4all.nl","sentAt":"2008-02-11T10:28:22Z","receivedAt":"2008-02-11T10:28:22Z","isPatch":false,"sender":{"key":"h.m.brand@xs4all.nl","avatar":"https://gravatar.com/avatar/5b8f83ee35c427a646cbea3b104346e00ab3663b99bbf435cddeb75cd4b3857b?d=mp&s=160"},"body":"Any hints?\n\n/pro/3gl/LINUX/git-1.5.4.rc5/t 154 > sh t5302-pack-index.sh\n* ok 1: setup\n* ok 2: pack-objects with index version 1\n* ok 3: pack-objects with index version 2\n* ok 4: both packs should be identical\n* ok 5: index v1 and index v2 should be different\n* ok 6: index-pack with index version 1\n* ok 7: index-pack with index version 2\n* ok 8: index-pack results should match pack-objects ones\n* ok 9: index v2: force some 64-bit offsets with pack-objects\nt5302-pack-index.sh[67]: 3156 Illegal instruction(coredump)\n* FAIL 10: index v2: verify a pack with some 64-bit offsets\n git verify-pack -v \"test-3-${pack3}.pack\"\nFATAL: Unexpected exit with code 0\n\n\nHP gdb 5.7 for HP Itanium (32 or 64 bit) and target HP-UX 11.2x.\nCopyright 1986 - 2001 Free Software Foundation, Inc.\nHewlett-Packard Wildebeest 5.7 (based on GDB) is covered by the\nGNU General Public License. Type \"show copying\" to see the conditions to\nchange it and/or distribute copies. Type \"show warranty\" for warranty/support.\n..\nCore was generated by `git'.\nProgram terminated with signal 4, Illegal instruction.\nILL_REGNAT - Register NaT Consumption\n#0 0xc0000000033c8940:0 in sha1_block_asm_host_order+0x22e0 ()\n from /usr/local/ssl/lib/libcrypto.so\n(gdb) where\n#0 0xc0000000033c8940:0 in sha1_block_asm_host_order+0x22e0 ()\n from /usr/local/ssl/lib/libcrypto.so\n#1 0xc0000000033cb680:0 in SHA1_Update () at ../md32_common.h:476\n#2 0x40000000000ef560:0 in verify_pack (p=0x6000000000037e40, verbose=1)\n at pack-check.c:168\n#3 0x40000000000d5130:0 in cmd_verify_pack (argc=2, argv=,\n prefix=) at builtin-verify-pack.c:43\n#4 0x40000000000440e0:0 in handle_internal_command (argc=3,\n argv=0x9fffffffffffee88) at git.c:257\n#5 0x40000000000446a0:0 in main (argc=3, argv=0x9fffffffffffee88) at git.c:447\n(gdb)\n\n-- \nH.Merijn Brand Amsterdam Perl Mongers (http://amsterdam.pm.org/)\nusing & porting perl 5.6.2, 5.8.x, 5.10.x on HP-UX 10.20, 11.00, 11.11,\n& 11.23, SuSE 10.1 & 10.2, AIX 5.2, and Cygwin. http://qa.perl.org\nhttp://mirrors.develooper.com/hpux/ http://www.test-smoke.org\n http://www.goldmark.org/jeff/stupid-disclaimers/\n"},{"id":"68458","messageId":"alpine.LSU.1.00.0802112240280.3870@racer.site","threadId":"12030","inReplyTo":"20080211112822.16b69495@pc09.procura.nl","subject":"Re: libcrypto core dump in 64bit","fromName":"Johannes Schindelin","fromEmail":"johannes.schindelin@gmx.de","sentAt":"2008-02-11T22:42:16Z","receivedAt":"2008-02-11T22:42:16Z","isPatch":false,"sender":{"key":"johannes.schindelin@gmx.de","avatar":"https://avatars.githubusercontent.com/u/127790?v=4"},"body":"Hi,\n\nOn Mon, 11 Feb 2008, H.Merijn Brand wrote:\n\n> Any hints?\n\nYes.\n\n> #0 0xc0000000033c8940:0 in sha1_block_asm_host_order+0x22e0 ()\n> from /usr/local/ssl/lib/libcrypto.so\n\nThis seems to be an OpenSSL issue, probably in its Itanium-optimised code \n(since Itanium is not _all_ that common, it is quite likely that no many \npeople exercise this part of the code).\n\nUnfortunately, I am not at all an expert in Itanium's assembler, otherwise \nI'd try to help...\n\nSorry,\nDscho\n"},{"id":"68519","messageId":"20080212150612.4d28c373@pc09.procura.nl","threadId":"12030","inReplyTo":"alpine.LSU.1.00.0802112240280.3870@racer.site","subject":"Re: libcrypto core dump in 64bit","fromName":"H.Merijn Brand","fromEmail":"h.m.brand@xs4all.nl","sentAt":"2008-02-12T14:06:12Z","receivedAt":"2008-02-12T14:06:12Z","isPatch":false,"sender":{"key":"h.m.brand@xs4all.nl","avatar":"https://gravatar.com/avatar/5b8f83ee35c427a646cbea3b104346e00ab3663b99bbf435cddeb75cd4b3857b?d=mp&s=160"},"body":"On Mon, 11 Feb 2008 22:42:16 +0000 (GMT), Johannes Schindelin\n wrote:\n\n> Hi,\n> \n> On Mon, 11 Feb 2008, H.Merijn Brand wrote:\n> \n> > Any hints?\n> \n> Yes.\n> \n> > #0 0xc0000000033c8940:0 in sha1_block_asm_host_order+0x22e0 ()\n> > from /usr/local/ssl/lib/libcrypto.so\n> \n> This seems to be an OpenSSL issue, probably in its Itanium-optimised code \n> (since Itanium is not _all_ that common, it is quite likely that no many \n> people exercise this part of the code).\n\nNow tried with HP C-ANSI-C -O3, -O2, -O1, and -O0 and with GNU gcc 4.2.1 -O3\nAll give the same failure\n \n> Unfortunately, I am not at all an expert in Itanium's assembler, otherwise \n> I'd try to help...\n\n-- \nH.Merijn Brand Amsterdam Perl Mongers (http://amsterdam.pm.org/)\nusing & porting perl 5.6.2, 5.8.x, 5.10.x on HP-UX 10.20, 11.00, 11.11,\n& 11.23, SuSE 10.1 & 10.2, AIX 5.2, and Cygwin. http://qa.perl.org\nhttp://mirrors.develooper.com/hpux/ http://www.test-smoke.org\n http://www.goldmark.org/jeff/stupid-disclaimers/\n"},{"id":"68524","messageId":"loom.20080212T152138-849@post.gmane.org","threadId":"12030","inReplyTo":"20080212150612.4d28c373@pc09.procura.nl","subject":"Re: libcrypto core dump in 64bit","fromName":"Michal Rokos","fromEmail":"michal@rokos.cz","sentAt":"2008-02-12T15:26:16Z","receivedAt":"2008-02-12T15:26:16Z","isPatch":false,"sender":{"key":"michal@rokos.cz","avatar":null},"body":"Hello,\n\nH.Merijn Brand xs4all.nl> writes:\n> On Mon, 11 Feb 2008 22:42:16 +0000 (GMT), Johannes Schindelin\n> gmx.de> wrote:\n> > > #0 0xc0000000033c8940:0 in sha1_block_asm_host_order+0x22e0 ()\n> > > from /usr/local/ssl/lib/libcrypto.so\n> > \n> > This seems to be an OpenSSL issue, probably in its Itanium-optimised code \n> > (since Itanium is not _all_ that common, it is quite likely that no many \n> > people exercise this part of the code).\n> \n> Now tried with HP C-ANSI-C -O3, -O2, -O1, and -O0 and with GNU gcc 4.2.1 -O3\n> All give the same failure\n> \n> > Unfortunately, I am not at all an expert in Itanium's assembler, otherwise \n> > I'd try to help...\n\nyes, this is OpenSSL issue. I had to recompile OpenSSL myself to make it work\n(with no-asm option)\n\nI'm using only HP cc for building git (with all the toolchain). (No gcc here)\n\n# uname -mrs\nHP-UX B.11.23 ia64\n# cc --version\ncc: HP C/aC++ B3910B A.06.14 [Feb 22 2007]\n\nMichal\n"},{"id":"68529","messageId":"20080212173842.0a3704b1@pc09.procura.nl","threadId":"12030","inReplyTo":"loom.20080212T152138-849@post.gmane.org","subject":"Re: libcrypto core dump in 64bit","fromName":"H.Merijn Brand","fromEmail":"h.m.brand@xs4all.nl","sentAt":"2008-02-12T16:38:42Z","receivedAt":"2008-02-12T16:38:42Z","isPatch":false,"sender":{"key":"h.m.brand@xs4all.nl","avatar":"https://gravatar.com/avatar/5b8f83ee35c427a646cbea3b104346e00ab3663b99bbf435cddeb75cd4b3857b?d=mp&s=160"},"body":"On Tue, 12 Feb 2008 15:26:16 +0000 (UTC), Michal Rokos \nwrote:\n\n> Hello,\n> \n> H.Merijn Brand xs4all.nl> writes:\n> > On Mon, 11 Feb 2008 22:42:16 +0000 (GMT), Johannes Schindelin\n> > gmx.de> wrote:\n> > > > #0 0xc0000000033c8940:0 in sha1_block_asm_host_order+0x22e0 ()\n> > > > from /usr/local/ssl/lib/libcrypto.so\n> > > \n> > > This seems to be an OpenSSL issue, probably in its Itanium-optimised code \n> > > (since Itanium is not _all_ that common, it is quite likely that no many \n> > > people exercise this part of the code).\n> > \n> > Now tried with HP C-ANSI-C -O3, -O2, -O1, and -O0 and with GNU gcc 4.2.1 -O3\n> > All give the same failure\n> > \n> > > Unfortunately, I am not at all an expert in Itanium's assembler, otherwise \n> > > I'd try to help...\n> \n> yes, this is OpenSSL issue. I had to recompile OpenSSL myself to make it work\n> (with no-asm option)\n\nOK, I did the same now, and the docs tell me to turn down optimization to\n+O2 when you choose to do so.\n\nApart from that, when you create shared libs, the default build creates a\nconflicting situation, so you have to throw away the files that cause dup\nnames: aes-ia64.o rc4-ia64.o bn-ia64.o from libcrypto.a\n\nt5302-pack-index.sh now passes\n\n> I'm using only HP cc for building git (with all the toolchain). (No gcc here)\n> \n> # uname -mrs\n> HP-UX B.11.23 ia64\n> # cc --version\n> cc: HP C/aC++ B3910B A.06.14 [Feb 22 2007]\n\nHP-UX 11.23/64 U rx1620/64 Itanium 2/1600(2) ia64 2037 Mb\ncc: HP C/aC++ B3910B A.06.15 [May 16 2007]\n\nDid you get the full test suite to pass?\n\n-- \nH.Merijn Brand Amsterdam Perl Mongers (http://amsterdam.pm.org/)\nusing & porting perl 5.6.2, 5.8.x, 5.10.x on HP-UX 10.20, 11.00, 11.11,\n& 11.23, SuSE 10.1 & 10.2, AIX 5.2, and Cygwin. http://qa.perl.org\nhttp://mirrors.develooper.com/hpux/ http://www.test-smoke.org\n http://www.goldmark.org/jeff/stupid-disclaimers/\n"},{"id":"68616","messageId":"bcfb3e870802130104n46d88964uee3eb3841c38807b@mail.gmail.com","threadId":"12030","inReplyTo":"20080212173842.0a3704b1@pc09.procura.nl","subject":"Re: libcrypto core dump in 64bit","fromName":"Michal Rokos","fromEmail":"michal@rokos.cz","sentAt":"2008-02-13T09:04:23Z","receivedAt":"2008-02-13T09:04:23Z","isPatch":false,"sender":{"key":"michal@rokos.cz","avatar":null},"body":"Hello,\n\nOn Feb 12, 2008 5:38 PM, H.Merijn Brand wrote:\n> Did you get the full test suite to pass?\n\nNo - vanilla git's master won't pass (since HPUX vsnprintf is broken\n(it returns -1 when n is exceeded).\n\nWith strbuf workarounds test suite passes up to t4200-rerere.sh.\n\nI did not try to investigate further or continue with the rest of the tests.\n\nMR\n"},{"id":"68629","messageId":"20080213115738.75301520@pc09.procura.nl","threadId":"12030","inReplyTo":"bcfb3e870802130104n46d88964uee3eb3841c38807b@mail.gmail.com","subject":"Re: libcrypto core dump in 64bit","fromName":"H.Merijn Brand","fromEmail":"h.m.brand@xs4all.nl","sentAt":"2008-02-13T10:57:38Z","receivedAt":"2008-02-13T10:57:38Z","isPatch":false,"sender":{"key":"h.m.brand@xs4all.nl","avatar":"https://gravatar.com/avatar/5b8f83ee35c427a646cbea3b104346e00ab3663b99bbf435cddeb75cd4b3857b?d=mp&s=160"},"body":"On Wed, 13 Feb 2008 10:04:23 +0100, \"Michal Rokos\" wrote:\n\n> Hello,\n> \n> On Feb 12, 2008 5:38 PM, H.Merijn Brand wrote:\n> > Did you get the full test suite to pass?\n> \n> No - vanilla git's master won't pass (since HPUX vsnprintf is broken\n> (it returns -1 when n is exceeded).\n\ndiff -purN git-1.5.4/strbuf.c git-1.5.4p/strbuf.c\n--- git-1.5.4/strbuf.c 2008-02-02 05:09:01 +0100\n+++ git-1.5.4p/strbuf.c 2008-02-13 11:26:56 +0100\n@@ -127,7 +127,7 @@ void strbuf_addf(struct strbuf *sb, cons\n int len;\n va_list ap;\n\n- if (!strbuf_avail(sb))\n+ if (strbuf_avail(sb) < 64)\n strbuf_grow(sb, 64);\n va_start(ap, fmt);\n len = vsnprintf(sb->buf + sb->len, sb->alloc - sb->len, fmt, ap);\n\n\n> With strbuf workarounds test suite passes up to t4200-rerere.sh.\n> \n> I did not try to investigate further or continue with the rest of the tests.\n\nI did, and got here\nhttp://www.xs4all.nl/~procura/git-1.5.4-hpux1123ipf64.diff\n\nNote, ntar = GNU tar, and npatch is the 'patch' as the rest of the world\nknows it. Our perl is in /pro. We do not want to interfere with the system\nperl (and mine is always newer anyway).\n\ngit-clone is a shell script, and it's return value is ignored somewhere\n\n/pro/3gl/LINUX/git-1.5.4p/t 122 > cat do-tests\n#!/bin/sh\n\nexport TAR=ntar\nrm -f *.err\nfor t in t[0-9]*.sh ; do\n echo $t\n sh $t > test.err 2>&1 || mv test.err $t.err\n rm -f test.err\n done\n/pro/3gl/LINUX/git-1.5.4p/t 123 > ll *.err\n204744 -rw-rw-rw- 1 merijn softwr 344 Feb 13 11:49 t5600-clone-fail-cleanup.sh.err\n204788 -rw-rw-rw- 1 merijn softwr 458 Feb 13 11:49 t5701-clone-local.sh.err\n205057 -rw-rw-rw- 1 merijn softwr 3039 Feb 13 11:50 t6002-rev-list-bisect.sh.err\n204643 -rw-rw-rw- 1 merijn softwr 3980 Feb 13 11:50 t6003-rev-list-topo-order.sh.err\n204808 -rw-rw-rw- 1 merijn softwr 899 Feb 13 11:50 t6022-merge-rename.sh.err\n204697 -rw-rw-rw- 1 merijn softwr 1340 Feb 13 11:52 t7201-co.sh.err\n204747 -rw-rw-rw- 1 merijn softwr 149 Feb 13 11:53 t9300-fast-import.sh.err\n204699 -rw-rw-rw- 1 merijn softwr 1651 Feb 13 11:53 t9301-fast-export.sh.err\n/pro/3gl/LINUX/git-1.5.4p/t 124 >\n\n-- \nH.Merijn Brand Amsterdam Perl Mongers (http://amsterdam.pm.org/)\nusing & porting perl 5.6.2, 5.8.x, 5.10.x on HP-UX 10.20, 11.00, 11.11,\n& 11.23, SuSE 10.1 & 10.2, AIX 5.2, and Cygwin. http://qa.perl.org\nhttp://mirrors.develooper.com/hpux/ http://www.test-smoke.org\n http://www.goldmark.org/jeff/stupid-disclaimers/\n"}]}