On bootup, I get the following warning logged: [ 12.192109] ------------[ cut here ]------------ [ 12.192118] WARNING: at fs/fs-writeback.c:972 __mark_inode_dirty+0x1d4/0x1e0() [ 12.192122] Hardware name: HP Pavilion dv2700 Notebook PC [ 12.192124] bdi-block not registered [ 12.192126] Modules linked in: loop(+) dm_mod ide_cd_mod cdrom snd_hda_codec_conexant ide_pci_generic snd_hda_intel snd_hda_codec r8712u(C) amd74xx snd_pcm ide_core snd_timer snd k8temp sg joydev snd_page_alloc forcedeth hwmon i2c_nforce2 serio_raw i2c_core battery ac button ext3 jbd mbcache ohci_hcd sd_mod ehci_hcd usbcore fan processor ahci libahci libata scsi_mod thermal [ 12.192164] Pid: 2338, comm: udisks-part-id Tainted: G C 2.6.36-rc3-realtek+ #7 [ 12.192167] Call Trace: [ 12.192174] [<ffffffff8104614a>] warn_slowpath_common+0x7a/0xb0 [ 12.192178] [<ffffffff81046221>] warn_slowpath_fmt+0x41/0x50 [ 12.192183] [<ffffffff81325584>] ? _raw_spin_lock+0x34/0x40 [ 12.192186] [<ffffffff8112ded5>] ? __mark_inode_dirty+0x65/0x1e0 [ 12.192190] [<ffffffff8112e044>] __mark_inode_dirty+0x1d4/0x1e0 [ 12.192195] [<ffffffff811210bd>] touch_atime+0x12d/0x170 [ 12.192201] [<ffffffff810cbb61>] generic_file_aio_read+0x5c1/0x720 [ 12.192205] [<ffffffff810e7fbd>] ? handle_mm_fault+0x75d/0x950 [ 12.192211] [<ffffffff8110a862>] do_sync_read+0xd2/0x110 [ 12.192214] [<ffffffff8107a4a0>] ? mark_held_locks+0x70/0xa0 [ 12.192218] [<ffffffff8132436f>] ? __mutex_unlock_slowpath+0xbf/0x150 [ 12.192222] [<ffffffff8107a6cd>] ? trace_hardirqs_on+0xd/0x10 [ 12.192226] [<ffffffff8110af93>] vfs_read+0xc3/0x170 [ 12.192229] [<ffffffff8110b08c>] sys_read+0x4c/0x80 [ 12.192234] [<ffffffff81002d2b>] system_call_fastpath+0x16/0x1b [ 12.192237] ---[ end trace 9b3df7d4a57af8dc ]--- [ 12.198091] loop: module loaded The kernel is a fresh pull from linux-2.6.git with some code added to staging for an RTL8192 USB device.
Jan, is this related to the block bdi thing we talked about before? Thanks.
Yes, this seems exactly like the problem you were fixing...
PS: I'm sorry, I've got sidetracked with other work so I didn't try pushing your fix yet.
Ah... good. I completely forgot what the original issue was and can't find the thread now. It was related to block layer bdi not ready for writeback or some such, right? Can you please give me a pointer? Thanks.
Larry, this seems like a duplicate of bug 16312. So are you running with Tejun's fix from that bug? From your initial comment it seems you don't...
Tejun's patch entitles "sd-delay-uevent" does not fix the problem. With it, I warm rebooted 3 times without the problem, but got the warning with a cold boot. As it is intermittent, I do not know if that is significant. I will be going off-line today for about 60 days. Perhaps someone else will turn up with this warning and they will be able to test. If not, I'll be back in November.
*** This bug has been marked as a duplicate of bug 16312 ***