From mboxrd@z Thu Jan 1 00:00:00 1970 From: Will Smith Subject: Re: Kernel panic when using repacker with reverse order Date: Mon, 20 Sep 2004 15:46:31 +0800 Message-ID: <414E8AD7.4020000@willsmith.org> References: <1986966896.20040918020309@tnonline.net> <20040918152442.GS5137@backtop.namesys.com> <7810725217.20040918174019@tnonline.net> <20040920072419.GU5137@backtop.namesys.com> Mime-Version: 1.0 Content-Transfer-Encoding: 7bit Return-path: list-help: list-unsubscribe: list-post: Errors-To: flx@namesys.com In-Reply-To: <20040920072419.GU5137@backtop.namesys.com> List-Id: Content-Type: text/plain; charset="us-ascii"; format="flowed" To: Alex Zarochentsev Cc: Spam , reiserfs-list@namesys.com I'm getting a similar thing - a stack trace at least, resulting in a filesystem that cannot be unmounted. It only ever comes from running in reverse order. And it's triggered, for me, *after* the repacker has finished, but during the large flush of dirty data that the repacker produces. A workaround, that works for me, seems to be to run a while true ; do sync sleep 10 done as soon as I start the repacker. My guess is that the repacker triggers some unusually complex work for the flush() code, showing up an error. The sync's produce many small flushes rather than 1 large complex one. Will Smith Alex Zarochentsev wrote: > On Sat, Sep 18, 2004 at 05:40:19PM +0200, Spam wrote: > >> Debug options were not turned on. Shall I recompile the kernel with >> these on? > > > yes. it gives much more info usually. > > >> I have upgraded to rc2-mm1. I will try the repacker again later >> today and see if there is any change. >> >> ~S >> >> >>>Hi, >> >>>On Sat, Sep 18, 2004 at 02:03:09AM +0200, Spam wrote: >>> >>>> I got a severe crash when tried the repacker using reverse (1) >>>> order. I did first a normal run: >> >>>Thanks for the resport. >> >>>Whether reiser4 debug options (assertions) were turned on or off? >> >> >> >>>> echo 0 > direction >>>> echo 1 > start >>>> >>>> Then I did the reverse: >>>> >>>> echo 1 > direction >>>> echo 1 > start >>>> >>>> It crashed rather quickly. >>>> Unfortunately I could not capture the whole debug output: >>>> >>>>EFLAGS: 00010246 (2.6.9-rc1-mm4) >>>>EIP is at jnode_build_key+0x76/0xc0 >>>>eax: c3b3bae0 ebx: 00000000 ecx: 00000000 edx: 0028e000 >>>>esi: 0028e000 edi: 00000000 ebp: c3b3bae0 esp: d7581c1c >>>>ds: 007b es: 007b ss: 0068 >>>>Process ent:loop0! (pid: 27667, threadinfo=d7580000 task=d758c000) >>>>Stack: d7581c58 d7581c58 00000000 c0c77380 d7581dbc 00000002 c01cb943 d7581e00 >>>>c0389ce8 00000000 c01cc401 00000002 c05a7da0 00000000 d7581dd8 c0b25120 >>>>00000000 00000105 0000001c 00000000 00000000 00000000 d7581c74 d7581d80 >>>>Call Trace: >>>>[] jnode_lock_parent_coord+0x1b3/0x310 >>>>[] scan_by_coord+0x281/0x320 >>>>[] scan_unformatted+0xcd/0x1d0 >>>>[] buffered_rmqueue+0x103/0x210 >>>>[] scan_common+0x3f/0x50 >>>>[] scan_left+0x59/0x90 >>>>[] jnode_flush+0xed/0x3d0 >>>>[] tasklet_action+0x37/0x60 >>>>[] find_first_dirty_jnode+0x2a/0x50 >>>>[] flush_current_atom+0x220/0x2b0 >>>>[] flush_some_atom+0x19c/0x2a0 >>>>[] entd_flush+0x69/0xf0 >>>>[] schedule+0x237/0x4e0 >>>>[] entd+0x13c/0x2a0 >>>>[] entd+0x0/0x2a0 >>>>[] kernel_thread_helper+0x5/0x18 >>>>Code: 14 8b 83 84 00 00 00 85 c0 75 3b 8d 76 00 8d bc 27 00 00 >>>>00 00 89 e8 e8 b9 b5 01 00 89 c3 8b 44 24 04 89 f2 89 f9 89 04 24 >>>>89 e8 53 54 8b 44 24 04 8b 5c 24 08 8b 74 24 0c 8b 7c 24 10 8b >>>>6c >>>><6>note: ent:loop0![27667] exited with preempt_count 1 >>>> >>>> >>>> >> >> > -- 1400 high resolution textures - http://www.mayang.com/textures Panoramic travel photos - http://www.willsmith.org/pano