From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1752366AbZG0Ohi (ORCPT ); Mon, 27 Jul 2009 10:37:38 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1752019AbZG0Ohh (ORCPT ); Mon, 27 Jul 2009 10:37:37 -0400 Received: from mx2.redhat.com ([66.187.237.31]:41586 "EHLO mx2.redhat.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751596AbZG0Ohg (ORCPT ); Mon, 27 Jul 2009 10:37:36 -0400 Organization: Red Hat UK Ltd. Registered Address: Red Hat UK Ltd, Amberley Place, 107-111 Peascod Street, Windsor, Berkshire, SI4 1TE, United Kingdom. Registered in England and Wales under Company Registration No. 3798903 From: David Howells In-Reply-To: References: <6950.1245661098@redhat.com> To: Takashi Iwai , Ingo Molnar Cc: dhowells@redhat.com, Linux filesystem caching discussion list , LKML Subject: Incorrect circular locking dependency? Date: Mon, 27 Jul 2009 15:37:10 +0100 Message-ID: <24075.1248705430@redhat.com> Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Takashi Iwai wrote: > ======================================================= > [ INFO: possible circular locking dependency detected ] > 2.6.30-test #7 > ------------------------------------------------------- > swapper/0 is trying to acquire lock: > (&cwq->lock){-.-...}, at: [] __queue_work+0x1f/0x4e > > but task is already holding lock: > (&q->lock){-.-.-.}, at: [] __wake_up+0x26/0x5c > > which lock already depends on the new lock. Okay. I think I understand this: (1) cachefiles_read_waiter() intercepts wake up events, and, as such, is run inside the waitqueue spinlock for the page bit waitqueue. (2) cachefiles_read_waiter() calls fscache_enqueue_retrieval() which calls fscache_enqueue_operation() which calls schedule_work() for fast operations, thus taking a per-CPU workqueue spinlock. (3) queue_work(), which is called by many things, calls __queue_work(), which takes the per-CPU workqueue spinlock. (4) __queue_work() then calls insert_work(), which calls wake_up(), which takes the waitqueue spinlock for the per-CPU workqueue waitqueue. Even though the two waitqueues are separate, I think lockdep sees them as having the same lock. Ingo: Is there any way around this? David --- > the existing dependency chain (in reverse order) is: > > -> #1 (&q->lock){-.-.-.}: > [] __lock_acquire+0xfd6/0x12d5 > [] lock_acquire+0xb7/0xeb > [] _spin_lock_irqsave+0x3d/0x5e > [] __wake_up+0x26/0x5c > [] insert_work+0x7b/0x95 > [] __queue_work+0x2e/0x4e > [] delayed_work_timer_fn+0x3c/0x4f > [] run_timer_softirq+0x180/0x206 > [] __do_softirq+0xc3/0x18d > [] do_softirq+0x44/0x7a > [] irq_exit+0x43/0x87 > [] smp_apic_timer_interrupt+0x7c/0x9b > [] apic_timer_interrupt+0x36/0x40 > [] cpu_idle+0xa2/0xbe > [] rest_init+0x66/0x79 > [] start_kernel+0x396/0x3ae > [] __init_begin+0x7f/0x98 > [] 0xffffffff > > -> #0 (&cwq->lock){-.-...}: > [] __lock_acquire+0xd26/0x12d5 > [] lock_acquire+0xb7/0xeb > [] _spin_lock_irqsave+0x3d/0x5e > [] __queue_work+0x1f/0x4e > [] queue_work_on+0x48/0x63 > [] queue_work+0x23/0x38 > [] schedule_work+0x1e/0x31 > [] fscache_enqueue_operation+0xc5/0x102 [fscache] > [] cachefiles_read_waiter+0xb3/0xcd [cachefiles] > [] __wake_up_common+0x4c/0x85 > [] __wake_up+0x38/0x5c > [] __wake_up_bit+0x34/0x4b > [] unlock_page+0x55/0x6a > [] mpage_end_io_read+0x4e/0x71 > [] bio_endio+0x31/0x44 > [] req_bio_endio+0xab/0xde > [] blk_update_request+0x17d/0x321 > [] blk_update_bidi_request+0x22/0x62 > [] blk_end_bidi_request+0x25/0x6e > [] blk_end_request+0x1a/0x30 > [] scsi_io_completion+0x193/0x3bb [scsi_mod] > [] scsi_finish_command+0xd9/0xf2 [scsi_mod] > [] scsi_softirq_done+0xf4/0x10d [scsi_mod] > [] blk_done_softirq+0x6f/0x8e > [] __do_softirq+0xc3/0x18d > [] do_softirq+0x44/0x7a > [] irq_exit+0x43/0x87 > [] do_IRQ+0x8d/0xb2 > [] common_interrupt+0x35/0x40 > [] cpu_idle+0xa2/0xbe > [] rest_init+0x66/0x79 > [] start_kernel+0x396/0x3ae > [] __init_begin+0x7f/0x98 > [] 0xffffffff > > other info that might help us debug this: > > 1 lock held by swapper/0: > #0: (&q->lock){-.-.-.}, at: [] __wake_up+0x26/0x5c > > stack backtrace: > Pid: 0, comm: swapper Not tainted 2.6.30-test #7 > Call Trace: > [] ? printk+0x1d/0x33 > [] print_circular_bug_tail+0xaf/0xcb > [] __lock_acquire+0xd26/0x12d5 > [] ? __queue_work+0x1f/0x4e > [] lock_acquire+0xb7/0xeb > [] ? __queue_work+0x1f/0x4e > [] ? __queue_work+0x1f/0x4e > [] _spin_lock_irqsave+0x3d/0x5e > [] ? __queue_work+0x1f/0x4e > [] __queue_work+0x1f/0x4e > [] queue_work_on+0x48/0x63 > [] queue_work+0x23/0x38 > [] schedule_work+0x1e/0x31 > [] fscache_enqueue_operation+0xc5/0x102 [fscache] > [] cachefiles_read_waiter+0xb3/0xcd [cachefiles] > [] __wake_up_common+0x4c/0x85 > [] __wake_up+0x38/0x5c > [] __wake_up_bit+0x34/0x4b > [] unlock_page+0x55/0x6a > [] mpage_end_io_read+0x4e/0x71 > [] bio_endio+0x31/0x44 > [] req_bio_endio+0xab/0xde > [] blk_update_request+0x17d/0x321 > [] blk_update_bidi_request+0x22/0x62 > [] blk_end_bidi_request+0x25/0x6e > [] blk_end_request+0x1a/0x30 > [] scsi_io_completion+0x193/0x3bb [scsi_mod] > [] ? trace_hardirqs_on+0x19/0x2c > [] ? scsi_device_unbusy+0x92/0xaa [scsi_mod] > [] scsi_finish_command+0xd9/0xf2 [scsi_mod] > [] scsi_softirq_done+0xf4/0x10d [scsi_mod] > [] blk_done_softirq+0x6f/0x8e > [] __do_softirq+0xc3/0x18d > [] do_softirq+0x44/0x7a > [] irq_exit+0x43/0x87 > [] do_IRQ+0x8d/0xb2 > [] common_interrupt+0x35/0x40 > [] ? mwait_idle+0x98/0xec > [] cpu_idle+0xa2/0xbe > [] rest_init+0x66/0x79 > [] start_kernel+0x396/0x3ae > [] __init_begin+0x7f/0x98