From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from out1-smtp.messagingengine.com ([66.111.4.25]:57145 "EHLO out1-smtp.messagingengine.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1728393AbeHMMFp (ORCPT ); Mon, 13 Aug 2018 08:05:45 -0400 Date: Mon, 13 Aug 2018 11:24:17 +0200 From: Greg KH To: Mike Rapoport Cc: stable@vger.kernel.org, James Bottomley Subject: Re: [PATCH 0/3 stable 4.9] proc/sysctl: prune stale dentries during unregistering Message-ID: <20180813092417.GC6381@kroah.com> References: <1534151298-32258-1-git-send-email-rppt@linux.vnet.ibm.com> <20180813092325.GB6381@kroah.com> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <20180813092325.GB6381@kroah.com> Sender: stable-owner@vger.kernel.org List-ID: On Mon, Aug 13, 2018 at 11:23:25AM +0200, Greg KH wrote: > On Mon, Aug 13, 2018 at 12:08:15PM +0300, Mike Rapoport wrote: > > Hi, > > > > These patches fix a memory leak in per-cpu areas in v4.9.y > > > > The following command sequence > > > > while :; do > > c=$(docker create busybox) > > docker start $c > > docker stop $c > > docker rm $c > > done > > > > causes the number of pcpu_get_vm_area entries in /proc/vmallocinfo to grow. > > > > The commit d6cffbbe9a7e ("proc/sysctl: prune stale dentries during > > unregistering") fixes the issue and the commits ace0c791e6c3 ("proc/sysctl: > > Don't grab i_lock under sysctl_lock.") and 2fd1d2c4ceb2 ("proc: Fix > > proc_sys_prune_dcache to hold a sb reference") are the follow-up fixups. > > > > I've also checked v4.4 and the issue does not appear there. > > > > Eric W. Biederman (2): > > proc/sysctl: Don't grab i_lock under sysctl_lock. > > proc: Fix proc_sys_prune_dcache to hold a sb reference > > > > Konstantin Khlebnikov (1): > > proc/sysctl: prune stale dentries during unregistering > > What is the git commit ids of these patches that are in Linus's tree? I > need those to be able to apply this. Ok, you listed the in the paragraph above, sorry I missed that. Putting it in the patch itself makes it more obvious what is going on here... thanks, greg k-h