From mboxrd@z Thu Jan 1 00:00:00 1970 From: Stefan Seyfried Subject: Re: 3.17-rc6: bcache_gc: BUG: soft lockup - CPU#2 stuck for 23s! Date: Wed, 03 Dec 2014 10:32:17 +0100 Message-ID: <547ED8A1.5050706@message-id.googlemail.com> References: <20141101204447.GB22219@kmo-pixel> <546FC2B4.5020201@message-id.invalid.com> <1416615728.4084.325.camel@geekdesk.ewheeler.net> <5470858B.9090900@message-id.googlemail.com> Mime-Version: 1.0 Content-Type: text/plain; charset=utf-8 Content-Transfer-Encoding: QUOTED-PRINTABLE Return-path: Received: from mail-wg0-f43.google.com ([74.125.82.43]:36652 "EHLO mail-wg0-f43.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752255AbaLCJcW (ORCPT ); Wed, 3 Dec 2014 04:32:22 -0500 Received: by mail-wg0-f43.google.com with SMTP id l18so19506749wgh.2 for ; Wed, 03 Dec 2014 01:32:20 -0800 (PST) In-Reply-To: Sender: linux-bcache-owner@vger.kernel.org List-Id: linux-bcache@vger.kernel.org To: Eric Wheeler Cc: Kent Overstreet , "linux-bcache@vger.kernel.org" , Ross Anderson , Stefan Priebe Hi all, Am 24.11.2014 um 19:52 schrieb Eric Wheeler: >>>> Before the patch goes upstream I need it tested so I know if it fi= xes >>>> the actual issue or not. >>> >>> We have been using the rcu_sched patch and the cond_resched patch >>> together >>> (both attached) since November 3rd on 3.17.2 without any bcache >>> backtraces. bcache is running in writeback mode. The server is >>> predominantly write-only with relatively few reads. >> >> Is the rcu_sched patch supposed to help the same or a totally differ= ent >> problem? Means: should I also apply it, rebuild the module and reboo= t >> (resetting the test time to zero :-) Ok, I am also running both patches, my uptime is 10 days and I'm right now doing a bit of the usual stress-testing on the bcached partitions (find /m1 /m2 /m3 > /dev/null which does read-mostly, and a rebuild of all my yocto-project embedded board targets, which does read-write) I have not seen any problems yet, so the patches don't seem to hurt. >> I had the "deadly" soft lockup once per week since updating to >> 3.16/3.17, so I can only tell if it might help in two weeks earliest= =2E It's not two weeks yet, but getting close. Load average never goes below 1 (with 3 bcache-backed mounts). Best regards, Stefan --=20 Stefan Seyfried Linux Consultant & Developer -- GPG Key: 0x731B665B B1 Systems GmbH Osterfeldstra=C3=9Fe 7 / 85088 Vohburg / http://www.b1-systems.de GF: Ralph Dehner / Unternehmenssitz: Vohburg / AG: Ingolstadt,HRB 3537