Date: Wed, 19 Jul 2006 09:09:29 -0600 From: hideo <hideo@lastamericanempire.com> To: John Baldwin <jhb@freebsd.org> Cc: amd64@freebsd.org, Alexander Leidinger <Alexander@leidinger.net>, y.pankov@irbis.net.ru, freebsd-amd64@freebsd.org Subject: Re: I need linuxolator testers with an amd64 machine (easy but time consuming) Message-ID: <20060719150929.GA29210@lastamericanempire.com> In-Reply-To: <200607181658.19626.jhb@freebsd.org> References: <20060717170446.22himtda6848w8cw@netchild.homeip.net> <20060717221302.3a0781a8@Magellan.Leidinger.net> <20060718193044.GA16690@lastamericanempire.com> <200607181658.19626.jhb@freebsd.org>
next in thread | previous in thread | raw e-mail | index | archive | help
John Baldwin (Tue 07/18/06 16:58): > On Tuesday 18 July 2006 15:30, hideo wrote: > > Alexander Leidinger (Mon 07/17/06 22:13): > > > Quoting hideo <hideo@lastamericanempire.com> (Mon, 17 Jul 2006 > 13:59:52 -0600): > > > > > > > Alexander Leidinger (Mon 07/17/06 17:04): > > > > > Yuri wrote: > > > > > > > > > > >I can't confirm that *every* linux binary segfaults. I've tried all > > > > > >binaries from fc4 linux base, and only `dd' and `sort' segfaulted, > when > > > > > >I've pressed ctrl+c. -CURRENT/amd64 updated 1 hour ago. > > > > > > > > > > Did you tried acroread? Could you please try acroread if you haven't > > > > > already? > > > > > > > > > > > > > > > > > I can confirm acroread and realplayer both segfault. I also see the same > > > > behaviour with dd/sort above. I haven't had time yet to do any further > > > > investigating. This is on -current @ Thu Jul 13 07:34:24 MDT 2006. > > > > > > Then I need someone to do the binary search as outlined in the first > > > message of the thread. > > > > These two together produce the segfaults here. > > > > $FreeBSD: src/sys/compat/linux/linux_ipc.c,v 1.47 2006/06/27 > > 18:28:49 jhb Exp $ > > It should just be this one then. Can you run a ktrace to try and see which > specific IPC command is failing? You could also maybe add printf's to > linux_semctl() to print out the args->cmd. That would help narrow it down. Here are a couple of the kdumps: bash: 29104 ktrace RET ktrace 0 29104 ktrace CALL execve(0x7fffffffed3a,0x7fffffffeb70,0x7fffffffeb80) 29104 ktrace NAMI "/compat/linux/bin/bash" 29104 ktrace NAMI "/compat/linux/lib/ld-linux.so.2" 29104 bash RET execve 0 29104 bash CALL ktrace(0) 29104 bash RET ktrace 135217152/0x80f4000 29104 bash CALL settimeofday(0xffffd91a) 29104 bash RET settimeofday 0 29104 bash CALL access(0x280fe564,R_OK) 29104 bash NAMI "/compat/linux/etc/ld.so.preload" 29104 bash NAMI "/etc/ld.so.preload" 29104 bash RET access JUSTRETURN 29104 bash CALL open(0x28100603,O_RDONLY,<unused>0x1) 29104 bash NAMI "/compat/linux/etc/ld.so.cache" 29104 bash NAMI "/compat/linux" 29104 bash NAMI "/compat/linux/etc/ld.so.cache" 29104 bash RET open 3 29104 bash CALL mmap(0x3,0xffffd1fc,PROT_EXEC,MAP_FILE,0,0,0,0,0,0,0,0....etc sort: 29168 ktrace RET ktrace 0 29168 ktrace CALL execve(0x7fffffffed3a,0x7fffffffeb70,0x7fffffffeb80) 29168 ktrace NAMI "/compat/linux/bin/sort" 29168 ktrace NAMI "/compat/linux/lib/ld-linux.so.2" 29168 sort RET execve 0 29168 sort CALL ktrace(0) 29168 sort RET ktrace 134565888/0x8055000 29168 sort CALL settimeofday(0xffffd91a) 29168 sort RET settimeofday 0 29168 sort CALL access(0x28069564,R_OK) 29168 sort NAMI "/compat/linux/etc/ld.so.preload" 29168 sort NAMI "/etc/ld.so.preload" 29168 sort RET access JUSTRETURN 29168 sort CALL open(0x2806b603,O_RDONLY,<unused>0x1) 29168 sort NAMI "/compat/linux/etc/ld.so.cache" 29168 sort NAMI "/compat/linux" 29168 sort NAMI "/compat/linux/etc/ld.so.cache" 29168 sort RET open 3 29168 sort CALL mmap(0x3,0xffffd1fc,PROT_EXEC,MAP_FILE,0,0,0,0,0,0,0,0,....etc reaplay, acroread, dd all show the same thing. acroread just bails executing bash in the same manner as above. If you want the full ktrace files let me know. Didn't see anything from printf though it's more than likely that I didn't have them in the right place, look for the output correctly, or all of the above. If you want to give me a diff with the printf's exactly where you need them I can do that. Zach
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20060719150929.GA29210>