From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1946212AbXBII2m (ORCPT ); Fri, 9 Feb 2007 03:28:42 -0500 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1946219AbXBII2m (ORCPT ); Fri, 9 Feb 2007 03:28:42 -0500 Received: from mail.gmx.net ([213.165.64.20]:59662 "HELO mail.gmx.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with SMTP id S1946212AbXBII2l (ORCPT ); Fri, 9 Feb 2007 03:28:41 -0500 X-Provags-ID: V01U2FsdGVkX1+D4S41Qv2DPbcP5e18EwPtNWP04fABtc8xCfmUSq dkBA== From: Frank Hartmann To: Jan Kara Cc: linux-kernel@vger.kernel.org Subject: Re: PROBLEM: 2.6.19.1 Oops while doing Disk IO + playing sound In-Reply-To: <20070208093930.GB10973@duck.suse.cz> (Jan Kara's message of "Thu, 8 Feb 2007 10:39:30 +0100") References: <877ivr6t0k.fsf@fantasio.hh.de> <20070207091841.GE24590@atrey.karlin.mff.cuni.cz> <87wt2thhux.fsf@fantasio.hh.de> <20070208093930.GB10973@duck.suse.cz> Date: Fri, 09 Feb 2007 09:28:47 +0100 Message-ID: <878xf7pwe8.fsf@fantasio.hh.de> User-Agent: Gnus/5.110006 (No Gnus v0.6) Emacs/22.0.93 (gnu/linux) MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii X-Y-GMX-Trusted: 0 Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org Hi Jan, Jan Kara writes: > On Wed 07-02-07 20:45:26, Frank Hartmann wrote: >> you are the first to react on my message. I did not try 2.6.20 uptonow. >> >> What I did in the meantime: >> >> Switched to 2.6.19.2 >> >> I run extensivly memtest+ and started something called smrtctl which >> should give more information on my harddisk status. Everything looks >> ok. >> >> From what I read on linux-kernel I came to the conclusion, >> that it might be benifical not using PREMPT. So I changed kernel configuration. >> >> Yesterday I had: > The oops below is from kernel without preempt? The first report(thread starter) using 2.6.19.1 was with preemp fantasio:~/tmp/linux-2.6.19.1$ fgrep -i preempt .config # CONFIG_PREEMPT_NONE is not set # CONFIG_PREEMPT_VOLUNTARY is not set CONFIG_PREEMPT=y CONFIG_PREEMPT_BKL=y The Oops below has: fantasio:~/tmp/linux-2.6.19.2$ fgrep -i preempt .config CONFIG_PREEMPT_NONE=y # CONFIG_PREEMPT_VOLUNTARY is not set # CONFIG_PREEMPT is not set >> Feb 6 22:35:44 fantasio kernel: BUG: unable to handle kernel NULL pointer dereference at virtual address 00000029 >> Feb 6 22:35:44 fantasio kernel: printing eip: >> Feb 6 22:35:44 fantasio kernel: e0c3b737 >> Feb 6 22:35:44 fantasio kernel: *pde = 00000000 >> Feb 6 22:35:44 fantasio kernel: Oops: 0000 [#1] >> Feb 6 22:35:44 fantasio kernel: Modules linked in: ext2 cbc blkcipher binfmt_misc bnep rfcomm l2cap nfs nfsd exportfs lockd sunrpc ip_nat_irc ip_nat_ftp ip_conntrack_irc ip_conntrack_ftp xt_state ipt_REJECT i >> pt_LOG iptable_mangle iptable_filter iptable_nat ip_nat ip_tables ip_conntrack bridge llc cryptoloop loop sha256 aes it87 hwmon_vid i2c_isa i2c_core md5 ipv6 af_packet radeon drm dm_crypt dm_mod snd_via82xx sn >> d_ac97_codec snd_ac97_bus snd_pcm_oss snd_mixer_oss analog 8250_pnp 8250 serial_core snd_pcm snd_timer snd_page_alloc snd_mpu401_uart gameport snd_rawmidi rtc snd soundcore prism54 via_agp agpgart firmware_cla >> ss evdev sd_mod ext3 jbd mbcache raid1 raid0 multipath linear md_mod ide_cd cdrom ide_disk usbhid usb_storage via_rhine mii via82cxxx ide_core ehci_hcd sym53c8xx scsi_transport_spi uhci_hcd usbcore scsi_mod un >> ix >> Feb 6 22:35:44 fantasio kernel: CPU: 0 >> Feb 6 22:35:44 fantasio kernel: EIP: 0060:[pg0+546215735/1070253056] Not tainted VLI >> Feb 6 22:35:44 fantasio kernel: EFLAGS: 00010213 (2.6.19.2 #1) >> Feb 6 22:35:44 fantasio kernel: EIP is at journal_grab_journal_head+0x2/0x14 [jbd] >> Feb 6 22:35:44 fantasio kernel: eax: 00000029 ebx: 00000029 ecx: 00000029 edx: 00000000 >> Feb 6 22:35:44 fantasio kernel: esi: c10d2580 edi: 00000000 ebp: c1feb6ec esp: df9fde00 >> Feb 6 22:35:44 fantasio kernel: ds: 007b es: 007b ss: 0068 >> Feb 6 22:35:44 fantasio kernel: Process kswapd0 (pid: 132, ti=df9fc000 task=dffca590 task.ti=df9fc000) >> Feb 6 22:35:44 fantasio kernel: Stack: e0c375ca 00000246 c1feb720 c24c532c c1186ac0 00000001 e0c6930d 000000d0 >> Feb 6 22:35:44 fantasio kernel: 00000001 c02aed00 c012f883 c10d2580 c24c532c c01344b1 df9fdf1c c0117e5d >> Feb 6 22:35:44 fantasio kernel: df9fdf80 00000020 00000000 00000000 00000020 00000001 00000001 00000000 >> Feb 6 22:35:44 fantasio kernel: Call Trace: >> Feb 6 22:35:44 fantasio kernel: [pg0+546198986/1070253056] journal_try_to_free_buffers+0x5e/0xf9 [jbd] >> Feb 6 22:35:44 fantasio kernel: [pg0+546403085/1070253056] ext3_releasepage+0x0/0x72 [ext3] >> Feb 6 22:35:44 fantasio kernel: [try_to_release_page+52/70] try_to_release_page+0x34/0x46 >> Feb 6 22:35:44 fantasio kernel: [shrink_inactive_list+1096/1677] shrink_inactive_list+0x448/0x68d >> Feb 6 22:35:44 fantasio kernel: [tasklet_action+55/87] tasklet_action+0x37/0x57 >> Feb 6 22:35:44 fantasio kernel: [__pagevec_release+21/29] __pagevec_release+0x15/0x1d >> Feb 6 22:35:45 fantasio kernel: [shrink_active_list+726/734] shrink_active_list+0x2d6/0x2de >> Feb 6 22:35:45 fantasio kernel: [shrink_zone+180/213] shrink_zone+0xb4/0xd5 >> Feb 6 22:35:45 fantasio kernel: [kswapd+623/886] kswapd+0x26f/0x376 >> Feb 6 22:35:45 fantasio kernel: [autoremove_wake_function+0/53] autoremove_wake_function+0x0/0x35 >> Feb 6 22:35:45 fantasio kernel: [kswapd+0/886] kswapd+0x0/0x376 >> Feb 6 22:35:45 fantasio kernel: [kthread+173/216] kthread+0xad/0xd8 >> Feb 6 22:35:45 fantasio kernel: [kthread+0/216] kthread+0x0/0xd8 >> Feb 6 22:35:45 fantasio kernel: [kernel_thread_helper+7/16] kernel_thread_helper+0x7/0x10 >> Feb 6 22:35:45 fantasio kernel: ======================= >> >> It happened probably while I was running >> >> 408 dd if=/dev/urandom of=crypt.img bs=1M count=20000 >> >> The file seems to be more or less complete and the process modified the file later: >> >> ls -l crypt.img >> -rw-r--r-- 1 frank frank 20970175708 Feb 6 22:46 crypt.img >> .. >> I will start using 2.6.20 soon. > OK, thanks for the news. Please, write me when you try out 2.6.20. Just > if it rings bell to someone on this list: It looks like the page had no > buffers but PagePrivate was set. Strange. > regards Frank