Most recent kernel where this bug did *NOT* occur: 2.6.20.1 Distribution: Debian "Etch" 4.0 Hardware Environment: Asus P5B, Core2 Duo System Software Environment: Problem Description: Kernel bug, crashes Steps to reproduce: Start a download in KTorrent, it is reproducable. Hello, It doesn't crash outright but this message pops up and then it slowly grinds to a halt. I've attached config and dmesg output. Asus P5B (intel 965) Core 2 Duo System. Stable on 2.6.20.1. I use JFS for all my disks. Starting a download in KTorrent is what seems to cause it. I'm unsure if this is JFS only related, but didn't know where else to put it. Mar 5 14:51:57 disher kernel: ------------[ cut here ]------------ Mar 5 14:51:57 disher kernel: kernel BUG at fs/mpage.c:535! Mar 5 14:51:57 disher kernel: invalid opcode: 0000 [#1] Mar 5 14:51:57 disher kernel: SMP Mar 5 14:51:57 disher kernel: Modules linked in: af_packet ipt_LOG xt_tcpudp xt_state ppp_synctty iptable_nat nf_nat nf_conntrack_ipv4 iptable_filter ip_tables x_tables ppp_async crc_ccitt w83627ehf hwmon i2c_isa eeprom loop usbhid hid snd_hda_intel snd_hda_codec snd_pcm_oss snd_mixer_oss snd_pcm snd_timer snd sg soundcore 8139too intel_agp i2c_i801 psmouse sr_mod snd_page_alloc ehci_hcd e1000 uhci_hcd agpgart usbcore rtc i2c_core cdrom Mar 5 14:51:57 disher kernel: CPU: 1 Mar 5 14:51:57 disher kernel: EIP: 0060:[<c0174733>] Not tainted VLI Mar 5 14:51:57 disher kernel: EFLAGS: 00010246 (2.6.21-rc2-git3 #1) Mar 5 14:51:57 disher kernel: EIP is at __mpage_writepage+0x13a/0x48c Mar 5 14:51:57 disher kernel: eax: 72010021 ebx: c1e6b4c0 ecx: 00000000 edx: 00000000 Mar 5 14:51:57 disher kernel: esi: c1e6b4c0 edi: 00000000 ebp: f721de6c esp: c2273db8 Mar 5 14:51:57 disher kernel: ds: 007b es: 007b fs: 00d8 gs: 0000 ss: 0068 Mar 5 14:51:57 disher kernel: Process pdflush (pid: 192, ti=c2272000 task=c2258ab0 task.ti=c2272000) Mar 5 14:51:57 disher kernel: Stack: c01932f3 c019f852 c2273eac c0189266 00000000 0000000c f721df10 00000001 Mar 5 14:51:57 disher kernel: f5de4000 00000080 00000001 00000000 f88e1990 00001f03 00000000 c2273e98 Mar 5 14:51:57 disher kernel: 0000003f 00000001 00000001 00000000 00000000 c2273e98 0000000e 00000000 Mar 5 14:51:57 disher kernel: Call Trace: Mar 5 14:51:57 disher kernel: [<c01932f3>] diWrite+0x4ab/0x4b5 Mar 5 14:51:57 disher kernel: [<c019f852>] lmWriteRecord+0x277/0x313 Mar 5 14:51:57 disher kernel: [<c0189266>] jfs_get_block+0x0/0x26f Mar 5 14:51:57 disher kernel: [<c013b778>] find_get_pages_tag+0x38/0x77 Mar 5 14:51:57 disher kernel: [<c0175239>] mpage_writepages+0x1e2/0x30b Mar 5 14:51:57 disher kernel: [<c0188ecc>] jfs_writepage+0x0/0xc Mar 5 14:51:57 disher kernel: [<c0189266>] jfs_get_block+0x0/0x26f Mar 5 14:51:57 disher kernel: [<c01118de>] __wake_up+0x32/0x43 Mar 5 14:51:57 disher kernel: [<c01a2646>] txEnd+0x56/0x11d Mar 5 14:51:57 disher kernel: [<c0188eb6>] jfs_writepages+0x0/0xa Mar 5 14:51:57 disher kernel: [<c01404cf>] do_writepages+0x20/0x30 Mar 5 14:51:57 disher kernel: [<c016b8e5>] __writeback_single_inode+0x198/0x308 Mar 5 14:51:57 disher kernel: [<c016bd30>] sync_sb_inodes+0x168/0x211 Mar 5 14:51:57 disher kernel: [<c016c0fe>] writeback_inodes+0x63/0xa5 Mar 5 14:51:57 disher kernel: [<c0140936>] wb_kupdate+0x7b/0xdf Mar 5 14:51:57 disher kernel: [<c0140c0e>] pdflush+0x0/0x19d Mar 5 14:51:57 disher kernel: [<c0140d17>] pdflush+0x109/0x19d Mar 5 14:51:57 disher kernel: [<c01408bb>] wb_kupdate+0x0/0xdf Mar 5 14:51:57 disher kernel: [<c012693e>] kthread+0xb2/0xdc Mar 5 14:51:57 disher kernel: [<c012688c>] kthread+0x0/0xdc Mar 5 14:51:57 disher kernel: [<c01031af>] kernel_thread_helper+0x7/0x10 Mar 5 14:51:57 disher kernel: ======================= Mar 5 14:51:57 disher kernel: Code: 70 43 8b 4a 18 8b 6c 24 1c 89 4c 24 24 8b 52 04 39 fa 0f 85 75 ff ff ff 85 ed 0f 85 1f 01 00 00 e9 dd 02 00 00 8b 06 a8 08 75 04 <0f> 0b eb fe 8b 7e 14 b9 0c 00 00 00 2b 4c 24 14 8b 44 24 34 8b Mar 5 14:51:57 disher kernel: EIP: [<c0174733>] __mpage_writepage+0x13a/0x48c SS:ESP 0068:c2273db8
Created attachment 10612 [details] Kernel config Config of the kernel
Created attachment 10613 [details] Dmesg output Dmesg Output
Created attachment 10616 [details] nobh_truncate_page() must set the page uptodate Can you give this patch a try?
Yes that works thanks Dave. It's going falling over anymore.
Err going=not.
Submitted the patch to Linus, who has some questions. Resubmitted with a couple of comments added.
Fix is included in 2.6.21-rc3. Thanks for reporting and testing this!