From: Samuel Just Subject: __dentry_open warning Date: Mon, 12 Sep 2011 17:00:02 -0700 Message-ID: <4E6E9D02.3010702@hq.newdream.net> Reply-To: samuel.just@dreamhost.com Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit Cc: josef@redhat.com, Sage Weil To: linux-fsdevel@vger.kernel.org Return-path: Received: from mail.hq.newdream.net ([66.33.206.127]:34423 "EHLO mail.hq.newdream.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751089Ab1ILX7p (ORCPT ); Mon, 12 Sep 2011 19:59:45 -0400 Sender: linux-fsdevel-owner@vger.kernel.org List-ID: We ran into this backtrace on v3.1 rc6 + the patch below. Sage says that this might be related to the dentry caching that Josef recently worked on. Processes accessing btrfs volumes became unresponsive at the same time. Any hints or patches? [ 192.134317] WARNING: at kernel/printk.c:322 do_syslog+0x82/0x409() [ 192.140548] Hardware name: PowerEdge R510 [ 192.144626] Attempt to access syslog with CAP_SYS_ADMIN but no CAP_SYSLOG (deprecated). [ 192.152698] Modules linked in: [ 192.155087] igb: eth2 NIC Link is Up 1000 Mbps Full Duplex, Flow Control: RX [ 192.162963] [ 192.164514] Pid: 10414, comm: syslog-ng Not tainted 3.0.0-dho-00059-geb29251-dirty #1 [ 192.172416] Call Trace: [ 192.174923] [] ? warn_slowpath_common+0x78/0x8c [ 192.181165] [] ? warn_slowpath_fmt+0x45/0x4a [ 192.187148] [] ? bit_waitqueue+0x14/0xa1 [ 192.192780] [] ? do_syslog+0x82/0x409 [ 192.198150] [] ? d_instantiate+0x39/0x46 [ 192.203783] [] ? d_rehash+0x15/0x1a [ 192.208986] [] ? kclist_add_private+0x100/0x100 [ 192.215220] [] ? proc_reg_open+0x88/0x11b [ 192.218483] bonding: bond0: link status definitely up for interface eth5, 1000 Mbps full duplex. [ 192.229810] [] ? proc_fill_super+0xa1/0xa1 [ 192.235623] [] ? __dentry_open+0x167/0x265 [ 192.238914] igb: eth3 NIC Link is Up 1000 Mbps Full Duplex, Flow Control: RX [ 192.248533] [] ? path_get+0x1e/0x26 [ 192.253735] [] ? do_last+0x5b1/0x6ed [ 192.254493] bonding: bond1: link status up for interface eth2, enabling it in 0 ms. [ 192.254498] bonding: bond1: link status up for interface eth3, enabling it in 200 ms. [ 192.254504] bonding: bond1: link status definitely up for interface eth2, 1000 Mbps full duplex. [ 192.256764] ADDRCONF(NETDEV_CHANGE): bond1: link becomes ready [ 192.294707] [] ? path_openat+0xc6/0x350 [ 192.300253] [] ? bit_waitqueue+0x14/0xa1 [ 192.305886] [] ? do_filp_open+0x2c/0x72 [ 192.311435] [] ? alloc_fd+0x69/0x10a [ 192.316724] [] ? do_sys_open+0x103/0x18a [ 192.322362] [] ? system_call_fastpath+0x16/0x1b [ 192.328600] ---[ end trace 3b357ca377e9af55 ]--- commit 5a88f351544cfd0039ac05140bc8ebd84813e346 Author: Sage Weil Date: Tue Aug 9 12:52:18 2011 -0700 btrfs fix clone ioctl res diff --git a/fs/btrfs/ioctl.c b/fs/btrfs/ioctl.c index 3351b1b..4997289 100644 --- a/fs/btrfs/ioctl.c +++ b/fs/btrfs/ioctl.c @@ -2323,7 +2323,7 @@ static noinline long btrfs_ioctl_clone(struct file *file, unsigned long srcfd, else new_key.offset = destoff; - trans = btrfs_start_transaction(root, 1); + trans = btrfs_start_transaction(root, 3); if (IS_ERR(trans)) { ret = PTR_ERR(trans); goto out;