|
Message-ID: <4E7A3687.7010806@bredband.net> Date: Wed, 21 Sep 2011 21:09:59 +0200 From: magnum <rawsmooth@...dband.net> To: john-dev@...ts.openwall.com Subject: Re: 1.7.8-jumbo-6-RC* on Mac OS X We've had a valgrind warning for a while. Jim and I have looked at it some time ago and regarded it as false alarm (ptrcheck is experimental and I've seen more false problems than real) but just in case someone gets an idea, here is the exact warning for jumbo-6 on a linux-x86-64-32 build: magnum@...ge:~/src/john/john/src$ valgrind --tool=exp-ptrcheck ../run/john --test=0 ==27591== exp-ptrcheck, a heap, stack and global array overrun detector ==27591== NOTE: This is an Experimental-Class Valgrind Tool ==27591== Copyright (C) 2003-2010, and GNU GPL'd, by OpenWorks Ltd et al. ==27591== Using Valgrind-3.6.1 and LibVEX; rerun with -h for copyright info ==27591== Command: ../run/john --test=0 ==27591== ==27591== Invalid read of size 1 ==27591== at 0x78A5A5F: vfprintf (vfprintf.c:1620) ==27591== by 0x78C540B: vsprintf (iovsprintf.c:43) ==27591== by 0x78AE28A: sprintf (sprintf.c:34) ==27591== by 0x80659A6: md5_gen_LOAD_PARSER_FUNCTIONS_LoadLINE (md5_gen_parser.c:399) ==27591== by 0x8065F3C: md5_gen_LOAD_PARSER_FUNCTIONS (md5_gen_parser.c:535) ==27591== by 0x8064584: LoadOneFormat (md5_gen_fmt.c:6626) ==27591== by 0x8064841: md5_gen_Register_formats (md5_gen_fmt.c:6682) ==27591== by 0x8095367: john_register_all (john.c:125) ==27591== by 0x8095F55: john_init (john.c:438) ==27591== by 0x80964C8: main (john.c:627) ==27591== Address 0x8a5cff3 expected vs actual: ==27591== Expected: global array "SetupName" in object with soname "NONE" ==27591== Actual: unknown ==27591== Benchmarking: Traditional DES [128/128 BS SSE2]... DONE ... Could this by any chance be related? Normal memcheck (which has much fewer false warnings) shows NO problems. magnum On 2011-09-21 14:29, Solar Designer wrote: > Jim, magnum, all - > > I am testing the current code on Mac OS X 10.5. macosx-x86-64 builds > and passes tests fine, but -sse2 fails at initialization: > > (gdb) r > Starting program: /Users/user/john/john-1.7.8-jumbo-6-RC6/run/john > Reading symbols for shared libraries +++++. done > > Program received signal EXC_BAD_ACCESS, Could not access memory. > Reason: KERN_PROTECTION_FAILURE at address: 0x00000000 > 0x0001e1b1 in md5_gen_SETUP () > (gdb) disass 0x0001e1a0 0x0001e1c0 > Dump of assembler code from 0x1e1a0 to 0x1e1c0: > 0x0001e1a0<md5_gen_SETUP+528>: stos %eax,%es:(%edi) > 0x0001e1a1<md5_gen_SETUP+529>: mov %esi,%ecx > 0x0001e1a3<md5_gen_SETUP+531>: mov 0x70(%esp),%esi > 0x0001e1a7<md5_gen_SETUP+535>: lea 0x64(%ebp),%edi > 0x0001e1aa<md5_gen_SETUP+538>: rep stos %eax,%es:(%edi) > 0x0001e1ac<md5_gen_SETUP+540>: mov 0xc(%esi),%eax > 0x0001e1af<md5_gen_SETUP+543>: xor %esi,%esi > 0x0001e1b1<md5_gen_SETUP+545>: mov (%eax),%eax > 0x0001e1b3<md5_gen_SETUP+547>: test %eax,%eax > 0x0001e1b5<md5_gen_SETUP+549>: je 0x1e202<md5_gen_SETUP+626> > 0x0001e1b7<md5_gen_SETUP+551>: nopw 0x0(%eax,%eax,1) > End of assembler dump. > (gdb) p $eax > $1 = 0 > > I'll investigate this. > > Also, there's this warning: > > md5-mmx.S:36:Alignment too large: 15. assumed. > > I think we should define/use DO_ALIGN(), like we do in other files. But > I might not bother fixing this now. > > Alexander
Powered by blists - more mailing lists
Confused about mailing lists and their use? Read about mailing lists on Wikipedia and check out these guidelines on proper formatting of your messages.