From: David S. <ds...@ja...> - 2006-09-27 18:53:48
|
2.6.17 kernel on SMP box (2 real processors, both hyperthreaded): First one: Sep 27 14:47:40 hostname kernel: BUG: soft lockup detected on CPU#1! Sep 27 14:47:40 hostname kernel: <c0447823> softlockup_tick+0x9f/0xb2 <c042b497> update_process_times+0x39/0x5c Sep 27 14:47:40 hostname kernel: <c041737d> smp_apic_timer_interrupt+0x54/0x5a <c0404800> apic_timer_interrupt+0x1c/0x24 Sep 27 14:47:40 hostname kernel: <c046eed0> generic_fillattr+0x68/0x9f <f8bf5ec4> fuse_getattr+0x76/0x7d [fuse] Sep 27 14:47:40 hostname kernel: <f8bf5e4e> fuse_getattr+0x0/0x7d [fuse] <c046ef45> vfs_getattr+0x3e/0x97 Sep 27 14:47:40 hostname kernel: <f8d59c43> encode_post_op_attr+0x38/0x212 [nfsd] <f8d527df> nfsd_write+0xa3/0xab [nfsd] Sep 27 14:47:40 hostname kernel: <f8d5a345> nfs3svc_encode_writeres+0xe/0x57 [nfsd] <f8d5a337> nfs3svc_encode_writeres+0x0/0x57 [nfsd] Sep 27 14:47:40 hostname kernel: <f8d4f134> nfsd_dispatch+0x124/0x16f [nfsd] <f8c1aa26> svc_process+0x37a/0x5b4 [sunrpc] Sep 27 14:47:40 hostname kernel: <f8d4f5e8> nfsd+0x1d3/0x377 [nfsd] <f8d4f415> nfsd+0x0/0x377 [nfsd] Sep 27 14:47:40 hostname kernel: <c0402005> kernel_thread_helper+0x5/0xb Another one: Sep 27 14:47:56 gw-colo-110 kernel: BUG: soft lockup detected on CPU#2! Sep 27 14:47:56 gw-colo-110 kernel: <c0447823> softlockup_tick+0x9f/0xb2 <c042b497> update_process_times+0x39/0x5c Sep 27 14:47:56 gw-colo-110 kernel: <c041737d> smp_apic_timer_interrupt+0x54/0x5a <c0404800> apic_timer_interrupt+0x1c/0x24 Sep 27 14:47:56 gw-colo-110 kernel: <f8bf7c00> fuse_commit_write+0xf2/0x16e [fuse] <c044a7ca> generic_file_buffered_write+0x3a5/0x56b Sep 27 14:47:56 gw-colo-110 kernel: <c04ba4fc> inode_has_perm+0x4e/0x56 <c044bb79> __generic_file_aio_write_nolock+0x3c5/0x402 Sep 27 14:47:56 gw-colo-110 kernel: <c0600aa3> _spin_unlock_irq+0x5/0x7 <c05fed2d> schedule+0xb17/0xb80 Sep 27 14:47:56 gw-colo-110 kernel: <c044bc3e> __generic_file_write_nolock+0x88/0x9d <c0434024> autoremove_wake_function+0x0/0x2d Sep 27 14:47:56 gw-colo-110 kernel: <c0600160> __mutex_lock_slowpath+0x2f1/0x3d1 <c044bd14> generic_file_write+0x29/0x97 Sep 27 14:47:56 gw-colo-110 kernel: <c044bd28> generic_file_write+0x3d/0x97 <c046730d> do_readv_writev+0x177/0x252 Sep 27 14:47:56 gw-colo-110 kernel: <c044bceb> generic_file_write+0x0/0x97 <f8bf57ee> request_send+0x2ab/0x2b3 [fuse] Sep 27 14:47:57 gw-colo-110 kernel: <f8bf71eb> fuse_finish_open+0x29/0x3e [fuse] <c046741f> vfs_writev+0x37/0x43 Sep 27 14:47:57 gw-colo-110 kernel: <f8d52168> nfsd_vfs_write+0xce/0x288 [nfsd] <f8bf77d9> fuse_open+0x0/0x7 [fuse] Sep 27 14:47:57 gw-colo-110 kernel: <c0466556> __dentry_open+0xe9/0x1aa <c046665c> dentry_open+0x45/0x4b Sep 27 14:47:57 gw-colo-110 kernel: <f8d527d2> nfsd_write+0x96/0xab [nfsd] <f8d5a683> nfs3svc_decode_writeargs+0x0/0x15d [nfsd] Sep 27 14:47:57 gw-colo-110 kernel: <f8d58c9d> nfsd3_proc_write+0xd0/0xea [nfsd] <f8d5a683> nfs3svc_decode_writeargs+0x0/0x15d [nfsd] Sep 27 14:47:57 gw-colo-110 kernel: <f8d4f0cc> nfsd_dispatch+0xbc/0x16f [nfsd] <f8c1aa26> svc_process+0x37a/0x5b4 [sunrpc] Sep 27 14:47:57 gw-colo-110 kernel: <f8d4f5e8> nfsd+0x1d3/0x377 [nfsd] <f8d4f415> nfsd+0x0/0x377 [nfsd] Sep 27 14:47:57 gw-colo-110 kernel: <c0402005> kernel_thread_helper+0x5/0xb David |
From: Miklos S. <mi...@sz...> - 2006-09-27 19:30:45
|
> 2.6.17 kernel on SMP box (2 real processors, both hyperthreaded): Very interesting, I'm almost sure this is the same bug which was reported by Franco Broi some years ago, although there wasn't soft-lockup detection in the kernel at the time. And I'm still not understanding what's going on. Can you press SysRq-P a number of times to see where it is wondering around? Also if you have a chance to compile a kernel with frame pointers and frame unwind info (in 'Kernel hacking' secion), it would help a lot in understanding the stack traces. Thanks, Miklos |