Return-Path: X-Spam-Checker-Version: SpamAssassin 3.1.7 (2006-10-05) on imap.leidinger.net X-Spam-Level: X-Spam-Status: No, score=-2.5 required=5.0 tests=BAYES_00,DK_POLICY_SIGNSOME,FORGED_RCVD_HELO autolearn=disabled version=3.1.7 X-Original-To: Alexander+FreeBSDCommitter@Leidinger.net Delivered-To: alexander+freebsdcommitter@leidinger.net Received: from redbullneu.bpaserver.net by imap.leidinger.net with POP3 (fetchmail-6.3.5) for (single-drop); Thu, 11 Jan 2007 08:37:27 +0100 (CET) Received: from mx2.freebsd.org (mx2.freebsd.org [69.147.83.53]) by redbull.bpaserver.net (Postfix) with ESMTP id 4AF2B2E1A6 for ; Thu, 11 Jan 2007 08:37:08 +0100 (CET) Received: from hub.freebsd.org (hub.freebsd.org [69.147.83.54]) by mx2.freebsd.org (Postfix) with ESMTP id A6AAE558DA for ; Thu, 11 Jan 2007 07:30:58 +0000 (GMT) (envelope-from kris@obsecurity.org) Received: by hub.freebsd.org (Postfix) id A447E16A40F; Thu, 11 Jan 2007 07:30:58 +0000 (UTC) Delivered-To: netchild@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [69.147.83.52]) by hub.freebsd.org (Postfix) with ESMTP id A2CE116A407 for ; Thu, 11 Jan 2007 07:30:58 +0000 (UTC) (envelope-from kris@obsecurity.org) Received: from elvis.mu.org (elvis.mu.org [192.203.228.196]) by mx1.freebsd.org (Postfix) with ESMTP id 931F013C468 for ; Thu, 11 Jan 2007 07:30:58 +0000 (UTC) (envelope-from kris@obsecurity.org) Received: from obsecurity.dyndns.org (elvis.mu.org [192.203.228.196]) by elvis.mu.org (Postfix) with ESMTP id 79A141A3C1A for ; Wed, 10 Jan 2007 23:30:58 -0800 (PST) Received: by obsecurity.dyndns.org (Postfix, from userid 1000) id D78395148C; Thu, 11 Jan 2007 02:30:57 -0500 (EST) Date: Thu, 11 Jan 2007 02:30:57 -0500 From: Kris Kennaway To: netchild@FreeBSD.org Subject: [kris@obsecurity.org: Re: TMPFS I/O Performance Benchmarks] Message-ID: <20070111073057.GE95053@xor.obsecurity.org> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline User-Agent: Mutt/1.4.2.2i X-BPAnet-MailScanner-Information: Please contact the ISP for more information X-BPAnet-MailScanner: Found to be clean X-BPAnet-MailScanner-SpamCheck: not spam, SpamAssassin (not cached, score=-14.864, required 6, autolearn=not spam, BAYES_00 -15.00, DK_POLICY_SIGNSOME 0.00, FORGED_RCVD_HELO 0.14) X-BPAnet-MailScanner-From: kris@obsecurity.org ----- Forwarded message from Kris Kennaway ----- Date: Sat, 25 Mar 2006 18:47:20 -0500 From: Kris Kennaway To: Kris Kennaway Cc: Rohit Jalan Subject: Re: TMPFS I/O Performance Benchmarks In-Reply-To: <20060325233009.GA66568@xor.obsecurity.org> User-Agent: Mutt/1.4.2.1i On Sat, Mar 25, 2006 at 06:30:09PM -0500, Kris Kennaway wrote: > On Sat, Mar 25, 2006 at 10:05:27AM +0530, Rohit Jalan wrote: > > I am yet trying to reproduce the panic. I shall send you a patch as > > soon as I manage to do so. > > > > I am pretty confident that this panic is not due to a programming > > error in the TMPFS module. > > I think it's because the module isn't picking up things like WITNESS, > INVARIANTS, etc from my kernel config (since e.g. the struct mtx > inside struct mount will have a different size, so the module is > accessing mnt_optnew with the wrong offset). I tried to use > MODULES_OVERRIDE to compile it with my kernel (after moving the module > to sys/modules/) - I think that will make it automatically pick up the > kernel options - but 'make depend' failed. I gave up and just moved it into sys/fs/tmpfs, and got it to compile in statically to the kernel with the attached diff. Now it works! Kris P.S. You'll need to remove the need for -fms-extensions before this can be imported though. Index: sys/conf/files =================================================================== RCS file: /data/ncvs/src/sys/conf/files,v retrieving revision 1.1104 diff -u -r1.1104 files --- sys/conf/files 23 Mar 2006 23:06:14 -0000 1.1104 +++ sys/conf/files 25 Mar 2006 23:32:30 -0000 @@ -1088,6 +1088,12 @@ fs/unionfs/union_subr.c optional unionfs fs/unionfs/union_vfsops.c optional unionfs fs/unionfs/union_vnops.c optional unionfs +fs/tmpfs/tmpfs_vnops.c optional tmpfs +fs/tmpfs/tmpfs_fifoops.c optional tmpfs +fs/tmpfs/tmpfs_vfsops.c optional tmpfs +fs/tmpfs/tmpfs_subr.c optional tmpfs +fs/tmpfs/tmpfs_uma.c optional tmpfs +fs/tmpfs/tmpfs_vacache.c optional tmpfs gdb/gdb_cons.c optional gdb gdb/gdb_main.c optional gdb gdb/gdb_packet.c optional gdb Index: sys/conf/options =================================================================== RCS file: /data/ncvs/src/sys/conf/options,v retrieving revision 1.531 diff -u -r1.531 options --- sys/conf/options 5 Mar 2006 22:52:16 -0000 1.531 +++ sys/conf/options 25 Mar 2006 23:31:23 -0000 @@ -180,6 +180,7 @@ MSDOSFS opt_dontuse.h NTFS opt_dontuse.h NULLFS opt_dontuse.h +TMPFS opt_dontuse.h NWFS opt_dontuse.h PORTALFS opt_dontuse.h PROCFS opt_dontuse.h ----- End forwarded message ----- Return-Path: X-Spam-Checker-Version: SpamAssassin 3.1.7 (2006-10-05) on imap.leidinger.net X-Spam-Level: X-Spam-Status: No, score=-2.5 required=5.0 tests=BAYES_00,DK_POLICY_SIGNSOME,FORGED_RCVD_HELO autolearn=disabled version=3.1.7 X-Original-To: Alexander+FreeBSDCommitter@Leidinger.net Delivered-To: alexander+freebsdcommitter@leidinger.net Received: from redbullneu.bpaserver.net by imap.leidinger.net with POP3 (fetchmail-6.3.5) for (single-drop); Thu, 11 Jan 2007 08:37:24 +0100 (CET) Received: from mx2.freebsd.org (mx2.freebsd.org [69.147.83.53]) by redbull.bpaserver.net (Postfix) with ESMTP id 614372E204 for ; Thu, 11 Jan 2007 08:36:32 +0100 (CET) Received: from hub.freebsd.org (hub.freebsd.org [69.147.83.54]) by mx2.freebsd.org (Postfix) with ESMTP id BCE2355F16 for ; Thu, 11 Jan 2007 07:30:22 +0000 (GMT) (envelope-from kris@obsecurity.org) Received: by hub.freebsd.org (Postfix) id BA3D616A407; Thu, 11 Jan 2007 07:30:22 +0000 (UTC) Delivered-To: netchild@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [69.147.83.52]) by hub.freebsd.org (Postfix) with ESMTP id B8E0316A403 for ; Thu, 11 Jan 2007 07:30:22 +0000 (UTC) (envelope-from kris@obsecurity.org) Received: from elvis.mu.org (elvis.mu.org [192.203.228.196]) by mx1.freebsd.org (Postfix) with ESMTP id AA31013C459 for ; Thu, 11 Jan 2007 07:30:22 +0000 (UTC) (envelope-from kris@obsecurity.org) Received: from obsecurity.dyndns.org (elvis.mu.org [192.203.228.196]) by elvis.mu.org (Postfix) with ESMTP id 8FA051A3C1A for ; Wed, 10 Jan 2007 23:30:22 -0800 (PST) Received: by obsecurity.dyndns.org (Postfix, from userid 1000) id EB647517C9; Thu, 11 Jan 2007 02:30:21 -0500 (EST) Date: Thu, 11 Jan 2007 02:30:21 -0500 From: Kris Kennaway To: netchild@FreeBSD.org Subject: [kris@obsecurity.org: tmpfs bug?] Message-ID: <20070111073021.GB95053@xor.obsecurity.org> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline User-Agent: Mutt/1.4.2.2i X-BPAnet-MailScanner-Information: Please contact the ISP for more information X-BPAnet-MailScanner: Found to be clean X-BPAnet-MailScanner-SpamCheck: not spam, SpamAssassin (not cached, score=-14.864, required 6, autolearn=not spam, BAYES_00 -15.00, DK_POLICY_SIGNSOME 0.00, FORGED_RCVD_HELO 0.14) X-BPAnet-MailScanner-From: kris@obsecurity.org ----- Forwarded message from Kris Kennaway ----- Date: Thu, 27 Apr 2006 14:58:49 -0400 From: Kris Kennaway To: Rohit Jalan Subject: tmpfs bug? User-Agent: Mutt/1.4.2.1i Hi Rohit, Sorry I didn't get back to you for a while. I have just started using tmpfs on my UP package build machines, since the Giant locking should not be a big deal there. However I am seeing strange behaviour: what should be essentially a rm -rf of large directory trees is not removing all of the files, but leaving many behind. When I do a second rm -rf it indeed removes all files. In this configuration, I have 32GB of swap configured on a system with 1GB of RAM. Thus, at any given time a lot of the filesystem is swapped out, and rm -rf will have to first page it back in. I wonder if something here is broken? I'll keep looking at it to see if I can diagnose the problem further. Kris ----- End forwarded message ----- Return-Path: X-Spam-Checker-Version: SpamAssassin 3.1.7 (2006-10-05) on imap.leidinger.net X-Spam-Level: X-Spam-Status: No, score=-0.7 required=5.0 tests=BAYES_00,DK_POLICY_SIGNSOME,FORGED_RCVD_HELO,IMPRONONCABLE_2,TW_BF,TW_DV,TW_XB autolearn=disabled version=3.1.7 X-Original-To: Alexander+FreeBSDCommitter@Leidinger.net Delivered-To: alexander+freebsdcommitter@leidinger.net Received: from redbullneu.bpaserver.net by imap.leidinger.net with POP3 (fetchmail-6.3.5) for (single-drop); Thu, 11 Jan 2007 08:37:25 +0100 (CET) Received: from mx2.freebsd.org (mx2.freebsd.org [69.147.83.53]) by redbull.bpaserver.net (Postfix) with ESMTP id 5F9F02E1A6 for ; Thu, 11 Jan 2007 08:36:36 +0100 (CET) Received: from hub.freebsd.org (hub.freebsd.org [69.147.83.54]) by mx2.freebsd.org (Postfix) with ESMTP id BF5DB56436 for ; Thu, 11 Jan 2007 07:30:26 +0000 (GMT) (envelope-from kris@obsecurity.org) Received: by hub.freebsd.org (Postfix) id BCD0C16A415; Thu, 11 Jan 2007 07:30:26 +0000 (UTC) Delivered-To: netchild@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [69.147.83.52]) by hub.freebsd.org (Postfix) with ESMTP id BBBEC16A40F for ; Thu, 11 Jan 2007 07:30:26 +0000 (UTC) (envelope-from kris@obsecurity.org) Received: from elvis.mu.org (elvis.mu.org [192.203.228.196]) by mx1.freebsd.org (Postfix) with ESMTP id 9EE4D13C45D for ; Thu, 11 Jan 2007 07:30:26 +0000 (UTC) (envelope-from kris@obsecurity.org) Received: from obsecurity.dyndns.org (elvis.mu.org [192.203.228.196]) by elvis.mu.org (Postfix) with ESMTP id 586511A3C1A for ; Wed, 10 Jan 2007 23:30:26 -0800 (PST) Received: by obsecurity.dyndns.org (Postfix, from userid 1000) id 99C205148C; Thu, 11 Jan 2007 02:30:25 -0500 (EST) Date: Thu, 11 Jan 2007 02:30:25 -0500 From: Kris Kennaway To: netchild@FreeBSD.org Subject: [kris@obsecurity.org: Re: tmpfs bug?] Message-ID: <20070111073025.GC95053@xor.obsecurity.org> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline User-Agent: Mutt/1.4.2.2i X-BPAnet-MailScanner-Information: Please contact the ISP for more information X-BPAnet-MailScanner: Found to be clean X-BPAnet-MailScanner-SpamCheck: not spam, SpamAssassin (not cached, score=-13.133, required 6, BAYES_00 -15.00, DK_POLICY_SIGNSOME 0.00, FORGED_RCVD_HELO 0.14, IMPRONONCABLE_2 1.50, TW_BF 0.08, TW_DV 0.08, TW_XB 0.08) X-BPAnet-MailScanner-From: kris@obsecurity.org ----- Forwarded message from Kris Kennaway ----- Date: Thu, 27 Apr 2006 23:05:58 -0400 From: Kris Kennaway To: Kris Kennaway Cc: Rohit Jalan Subject: Re: tmpfs bug? In-Reply-To: <20060427185849.GA68736@xor.obsecurity.org> User-Agent: Mutt/1.4.2.1i On Thu, Apr 27, 2006 at 02:58:49PM -0400, Kris Kennaway wrote: > Hi Rohit, > > Sorry I didn't get back to you for a while. I have just started using > tmpfs on my UP package build machines, since the Giant locking should > not be a big deal there. However I am seeing strange behaviour: what > should be essentially a rm -rf of large directory trees is not > removing all of the files, but leaving many behind. When I do a > second rm -rf it indeed removes all files. > > In this configuration, I have 32GB of swap configured on a system with > 1GB of RAM. Thus, at any given time a lot of the filesystem is > swapped out, and rm -rf will have to first page it back in. I wonder > if something here is broken? > > I'll keep looking at it to see if I can diagnose the problem further. Got this panic: VNASSERT failed 0xc5553d20: tag none, type VBAD usecount 0, writecount 0, refcount 0 mountedhere 0 flags (VI_DOOMED) VI_LOCKed #0 0xc04b9fa8 at lockmgr+0x510 #1 0xc051617d at vop_stdlock+0x21 #2 0xc0630283 at VOP_LOCK_APV+0x87 #3 0xc051d0f4 at vtryrecycle+0x54 #4 0xc051cbdf at vnlru_free+0x193 #5 0xc051d21f at getnewvnode+0x33 #6 0xc048ed6a at tmpfs_alloc_vp+0x52 #7 0xc048cf52 at tmpfs_lookup+0x1da #8 0xc062e417 at VOP_CACHEDLOOKUP_APV+0x9b #9 0xc0513ebd at vfs_cache_lookup+0xb5 #10 0xc062e343 at VOP_LOOKUP_APV+0x87 #11 0xc051813e at lookup+0x42e #12 0xc0517aae at namei+0x382 #13 0xc05254f3 at kern_lstat+0x47 #14 0xc052548f at lstat+0x1b #15 0xc0622cc6 at syscall+0x27e #16 0xc06109cf at Xint0x80_syscall+0x1f KDB: stack backtrace: kdb_backtrace(edaac860,c0520de1,c0657fe1,c067127c,c5553d20) at kdb_backtrace+0x29 vfs_badlock(c0657fe1,c067127c,c5553d20) at vfs_badlock+0x11 assert_vi_unlocked(c5553d20,c067127c) at assert_vi_unlocked+0x31 VOP_PRINT_APV(c067c960,edaac888) at VOP_PRINT_APV+0x89 vn_printf(c5553d20,c0656861,c5553d20,edaac918,c051f5fe) at vn_printf+0x182 vdestroy(c5553d20,c5553d20,edaac938,c051f5b9,c5553d20) at vdestroy+0x71 vdropl(c5553d20,c5553de8,0,c06576b0,878) at vdropl+0x3e vdrop(c5553d20,edaacb70,c6745630,c4ec9870,edaac968) at vdrop+0x25 vnlru_free(1) at vnlru_free+0x1ba getnewvnode(c0647176,c4fd12d4,c067dc00,edaac980,0) at getnewvnode+0x33 tmpfs_alloc_vp(c4fd12d4,c6745630,edaacb5c,c4ec9870,c6725334) at tmpfs_alloc_vp+0x52 tmpfs_lookup(edaaca20,c4c82bd0,edaacb70,edaaca5c,c0513ebd) at tmpfs_lookup+0x1da VOP_CACHEDLOOKUP_APV(c067dc00,edaaca20) at VOP_CACHEDLOOKUP_APV+0x9b vfs_cache_lookup(edaacabc,c4c82bd0,edaacb70,edaacad8,c051813e) at vfs_cache_lookup+0xb5 VOP_LOOKUP_APV(c067dc00,edaacabc) at VOP_LOOKUP_APV+0x87 lookup(edaacb48,6b2,1,c4ec9870,c04eb136) at lookup+0x42e namei(edaacb48,805cca8,0,0,c4c4e2a0) at namei+0x382 kern_lstat(c4ec9870,805cca8,0,edaacc1c) at kern_lstat+0x47 lstat(c4ec9870,edaacd04,114,c4ec9870,edaacd38) at lstat+0x1b syscall(805003b,805003b,bfbf003b,805cc48,805cc00) at syscall+0x27e Xint0x80_syscall() at Xint0x80_syscall+0x1f --- syscall (190, FreeBSD ELF32, lstat), eip = 0x2813b83f, esp = 0xbfbf75fc, ebp = 0xbfbf7698 --- VOP_PRINT: 0xc5553d20 interlock is locked but should not be KDB: enter: lock violation db> show lockedvnods Locked vnodes 0xc4c82bd0: tag tmpfs, type VDIR usecount 3, writecount 0, refcount 4 mountedhere 0 flags () lock type tmpfs: EXCL (count 1) by thread 0xc4ec9870 (pid 10591)#0 0xc04b9fa8 at lockmgr+0x510 #1 0xc051617d at vop_stdlock+0x21 #2 0xc0630283 at VOP_LOCK_APV+0x87 #3 0xc052b160 at vn_lock+0xac #4 0xc0517de2 at lookup+0xd2 #5 0xc0517aae at namei+0x382 #6 0xc05254f3 at kern_lstat+0x47 #7 0xc052548f at lstat+0x1b #8 0xc0622cc6 at syscall+0x27e #9 0xc06109cf at Xint0x80_syscall+0x1f tag VT_TMPFS, tmpfs_node 0xc63be420, flags 0x0, links 2 mode 0755, owner 0, group 0, size 11100, status 0x2 lock type tmpfs: EXCL (count 1) by thread 0xc4ec9870 (pid 10591)#0 0xc04b9fa8 at lockmgr+0x510 #1 0xc051617d at vop_stdlock+0x21 #2 0xc0630283 at VOP_LOCK_APV+0x87 #3 0xc052b160 at vn_lock+0xac #4 0xc0517de2 at lookup+0xd2 #5 0xc0517aae at namei+0x382 #6 0xc05254f3 at kern_lstat+0x47 #7 0xc052548f at lstat+0x1b #8 0xc0622cc6 at syscall+0x27e #9 0xc06109cf at Xint0x80_syscall+0x1f Core available should you need it. Kris ----- End forwarded message ----- Return-Path: X-Spam-Checker-Version: SpamAssassin 3.1.7 (2006-10-05) on imap.leidinger.net X-Spam-Level: X-Spam-Status: No, score=-0.9 required=5.0 tests=BAYES_00,DK_POLICY_SIGNSOME,FORGED_RCVD_HELO,IMPRONONCABLE_2,TW_DV autolearn=disabled version=3.1.7 X-Original-To: Alexander+FreeBSDCommitter@Leidinger.net Delivered-To: alexander+freebsdcommitter@leidinger.net Received: from redbullneu.bpaserver.net by imap.leidinger.net with POP3 (fetchmail-6.3.5) for (single-drop); Thu, 11 Jan 2007 08:37:26 +0100 (CET) Received: from mx2.freebsd.org (mx2.freebsd.org [69.147.83.53]) by redbull.bpaserver.net (Postfix) with ESMTP id 249C22E204 for ; Thu, 11 Jan 2007 08:36:39 +0100 (CET) Received: from hub.freebsd.org (hub.freebsd.org [69.147.83.54]) by mx2.freebsd.org (Postfix) with ESMTP id 815ED564D4 for ; Thu, 11 Jan 2007 07:30:29 +0000 (GMT) (envelope-from kris@obsecurity.org) Received: by hub.freebsd.org (Postfix) id 7EA4516A412; Thu, 11 Jan 2007 07:30:29 +0000 (UTC) Delivered-To: netchild@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [69.147.83.52]) by hub.freebsd.org (Postfix) with ESMTP id 7D60116A40F for ; Thu, 11 Jan 2007 07:30:29 +0000 (UTC) (envelope-from kris@obsecurity.org) Received: from elvis.mu.org (elvis.mu.org [192.203.228.196]) by mx1.freebsd.org (Postfix) with ESMTP id 6EBA913C448 for ; Thu, 11 Jan 2007 07:30:29 +0000 (UTC) (envelope-from kris@obsecurity.org) Received: from obsecurity.dyndns.org (elvis.mu.org [192.203.228.196]) by elvis.mu.org (Postfix) with ESMTP id 55F121A3C1A for ; Wed, 10 Jan 2007 23:30:29 -0800 (PST) Received: by obsecurity.dyndns.org (Postfix, from userid 1000) id B690C519EB; Thu, 11 Jan 2007 02:30:28 -0500 (EST) Date: Thu, 11 Jan 2007 02:30:28 -0500 From: Kris Kennaway To: netchild@FreeBSD.org Subject: [kris@obsecurity.org: Re: tmpfs bug?] Message-ID: <20070111073028.GD95053@xor.obsecurity.org> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline User-Agent: Mutt/1.4.2.2i X-BPAnet-MailScanner-Information: Please contact the ISP for more information X-BPAnet-MailScanner: Found to be clean X-BPAnet-MailScanner-SpamCheck: not spam, SpamAssassin (not cached, score=-13.287, required 6, BAYES_00 -15.00, DK_POLICY_SIGNSOME 0.00, FORGED_RCVD_HELO 0.14, IMPRONONCABLE_2 1.50, TW_DV 0.08) X-BPAnet-MailScanner-From: kris@obsecurity.org ----- Forwarded message from Kris Kennaway ----- Date: Fri, 28 Apr 2006 01:23:24 -0400 From: Kris Kennaway To: Kris Kennaway Cc: Rohit Jalan Subject: Re: tmpfs bug? In-Reply-To: <20060428030558.GA75954@xor.obsecurity.org> User-Agent: Mutt/1.4.2.1i On Thu, Apr 27, 2006 at 11:05:58PM -0400, Kris Kennaway wrote: > On Thu, Apr 27, 2006 at 02:58:49PM -0400, Kris Kennaway wrote: > > Hi Rohit, > > > > Sorry I didn't get back to you for a while. I have just started using > > tmpfs on my UP package build machines, since the Giant locking should > > not be a big deal there. However I am seeing strange behaviour: what > > should be essentially a rm -rf of large directory trees is not > > removing all of the files, but leaving many behind. When I do a > > second rm -rf it indeed removes all files. > > > > In this configuration, I have 32GB of swap configured on a system with > > 1GB of RAM. Thus, at any given time a lot of the filesystem is > > swapped out, and rm -rf will have to first page it back in. I wonder > > if something here is broken? > > > > I'll keep looking at it to see if I can diagnose the problem further. > > Got this panic: Another one: VNASSERT failed 0xc5548bd0: tag none, type VBAD usecount 0, writecount 0, refcount 0 mountedhere 0 flags (VI_DOOMED) VI_LOCKed #0 0xc04b9fa8 at lockmgr+0x510 #1 0xc051617d at vop_stdlock+0x21 #2 0xc0630283 at VOP_LOCK_APV+0x87 #3 0xc051d0f4 at vtryrecycle+0x54 #4 0xc051cbdf at vnlru_free+0x193 #5 0xc051ccb2 at vnlru_proc+0x82 #6 0xc04b0eb4 at fork_exit+0xa4 #7 0xc06109dc at fork_trampoline+0x8 KDB: stack backtrace: kdb_backtrace(e32cbbfc,c0520de1,c0657fe1,c067127c,c5548bd0) at kdb_backtrace+0x29 vfs_badlock(c0657fe1,c067127c,c5548bd0) at vfs_badlock+0x11 assert_vi_unlocked(c5548bd0,c067127c) at assert_vi_unlocked+0x31 VOP_PRINT_APV(c067c960,e32cbc24) at VOP_PRINT_APV+0x89 vn_printf(c5548bd0,c0656861,c5548bd0,e32cbcb4,c051f5fe) at vn_printf+0x182 vdestroy(c5548bd0,c5548bd0,e32cbcd4,c051f5b9,c5548bd0) at vdestroy+0x71 vdropl(c5548bd0,c5548c98,0,c06576b0,878) at vdropl+0x3e vdrop(c5548bd0,9a569,c051cc30,0,e32cbd0c) at vdrop+0x25 vnlru_free(2f) at vnlru_free+0x1ba vnlru_proc(0,e32cbd38) at vnlru_proc+0x82 fork_exit(c051cc30,0,e32cbd38) at fork_exit+0xa4 fork_trampoline() at fork_trampoline+0x8 --- trap 0x1, eip = 0, esp = 0xe32cbd6c, ebp = 0 --- VOP_PRINT: 0xc5548bd0 interlock is locked but should not be KDB: enter: lock violation db> show lockedvnods Locked vnodes > Core available should you need it. Kris ----- End forwarded message ----- Return-Path: X-Spam-Checker-Version: SpamAssassin 3.1.7 (2006-10-05) on imap.leidinger.net X-Spam-Level: X-Spam-Status: No, score=-2.5 required=5.0 tests=BAYES_00,DK_POLICY_SIGNSOME,FORGED_RCVD_HELO autolearn=disabled version=3.1.7 X-Original-To: Alexander+FreeBSDCommitter@Leidinger.net Delivered-To: alexander+freebsdcommitter@leidinger.net Received: from redbullneu.bpaserver.net by imap.leidinger.net with POP3 (fetchmail-6.3.5) for (single-drop); Thu, 11 Jan 2007 08:37:27 +0100 (CET) Received: from mx2.freebsd.org (mx2.freebsd.org [69.147.83.53]) by redbull.bpaserver.net (Postfix) with ESMTP id 4AF2B2E1A6 for ; Thu, 11 Jan 2007 08:37:08 +0100 (CET) Received: from hub.freebsd.org (hub.freebsd.org [69.147.83.54]) by mx2.freebsd.org (Postfix) with ESMTP id A6AAE558DA for ; Thu, 11 Jan 2007 07:30:58 +0000 (GMT) (envelope-from kris@obsecurity.org) Received: by hub.freebsd.org (Postfix) id A447E16A40F; Thu, 11 Jan 2007 07:30:58 +0000 (UTC) Delivered-To: netchild@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [69.147.83.52]) by hub.freebsd.org (Postfix) with ESMTP id A2CE116A407 for ; Thu, 11 Jan 2007 07:30:58 +0000 (UTC) (envelope-from kris@obsecurity.org) Received: from elvis.mu.org (elvis.mu.org [192.203.228.196]) by mx1.freebsd.org (Postfix) with ESMTP id 931F013C468 for ; Thu, 11 Jan 2007 07:30:58 +0000 (UTC) (envelope-from kris@obsecurity.org) Received: from obsecurity.dyndns.org (elvis.mu.org [192.203.228.196]) by elvis.mu.org (Postfix) with ESMTP id 79A141A3C1A for ; Wed, 10 Jan 2007 23:30:58 -0800 (PST) Received: by obsecurity.dyndns.org (Postfix, from userid 1000) id D78395148C; Thu, 11 Jan 2007 02:30:57 -0500 (EST) Date: Thu, 11 Jan 2007 02:30:57 -0500 From: Kris Kennaway To: netchild@FreeBSD.org Subject: [kris@obsecurity.org: Re: TMPFS I/O Performance Benchmarks] Message-ID: <20070111073057.GE95053@xor.obsecurity.org> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline User-Agent: Mutt/1.4.2.2i X-BPAnet-MailScanner-Information: Please contact the ISP for more information X-BPAnet-MailScanner: Found to be clean X-BPAnet-MailScanner-SpamCheck: not spam, SpamAssassin (not cached, score=-14.864, required 6, autolearn=not spam, BAYES_00 -15.00, DK_POLICY_SIGNSOME 0.00, FORGED_RCVD_HELO 0.14) X-BPAnet-MailScanner-From: kris@obsecurity.org ----- Forwarded message from Kris Kennaway ----- Date: Sat, 25 Mar 2006 18:47:20 -0500 From: Kris Kennaway To: Kris Kennaway Cc: Rohit Jalan Subject: Re: TMPFS I/O Performance Benchmarks In-Reply-To: <20060325233009.GA66568@xor.obsecurity.org> User-Agent: Mutt/1.4.2.1i On Sat, Mar 25, 2006 at 06:30:09PM -0500, Kris Kennaway wrote: > On Sat, Mar 25, 2006 at 10:05:27AM +0530, Rohit Jalan wrote: > > I am yet trying to reproduce the panic. I shall send you a patch as > > soon as I manage to do so. > > > > I am pretty confident that this panic is not due to a programming > > error in the TMPFS module. > > I think it's because the module isn't picking up things like WITNESS, > INVARIANTS, etc from my kernel config (since e.g. the struct mtx > inside struct mount will have a different size, so the module is > accessing mnt_optnew with the wrong offset). I tried to use > MODULES_OVERRIDE to compile it with my kernel (after moving the module > to sys/modules/) - I think that will make it automatically pick up the > kernel options - but 'make depend' failed. I gave up and just moved it into sys/fs/tmpfs, and got it to compile in statically to the kernel with the attached diff. Now it works! Kris P.S. You'll need to remove the need for -fms-extensions before this can be imported though. Index: sys/conf/files =================================================================== RCS file: /data/ncvs/src/sys/conf/files,v retrieving revision 1.1104 diff -u -r1.1104 files --- sys/conf/files 23 Mar 2006 23:06:14 -0000 1.1104 +++ sys/conf/files 25 Mar 2006 23:32:30 -0000 @@ -1088,6 +1088,12 @@ fs/unionfs/union_subr.c optional unionfs fs/unionfs/union_vfsops.c optional unionfs fs/unionfs/union_vnops.c optional unionfs +fs/tmpfs/tmpfs_vnops.c optional tmpfs +fs/tmpfs/tmpfs_fifoops.c optional tmpfs +fs/tmpfs/tmpfs_vfsops.c optional tmpfs +fs/tmpfs/tmpfs_subr.c optional tmpfs +fs/tmpfs/tmpfs_uma.c optional tmpfs +fs/tmpfs/tmpfs_vacache.c optional tmpfs gdb/gdb_cons.c optional gdb gdb/gdb_main.c optional gdb gdb/gdb_packet.c optional gdb Index: sys/conf/options =================================================================== RCS file: /data/ncvs/src/sys/conf/options,v retrieving revision 1.531 diff -u -r1.531 options --- sys/conf/options 5 Mar 2006 22:52:16 -0000 1.531 +++ sys/conf/options 25 Mar 2006 23:31:23 -0000 @@ -180,6 +180,7 @@ MSDOSFS opt_dontuse.h NTFS opt_dontuse.h NULLFS opt_dontuse.h +TMPFS opt_dontuse.h NWFS opt_dontuse.h PORTALFS opt_dontuse.h PROCFS opt_dontuse.h ----- End forwarded message -----