Return-Path: Date: Fri, 12 Jan 2018 12:39:28 -0500 From: Scott Mayhew To: Trond Myklebust Cc: "anna.schumaker@netapp.com" , bcodding redhat , "fsorenso@redhat.com" , "linux-nfs@vger.kernel.org" , "dwysocha@redhat.com" Subject: Re: [PATCH v2] nfs/pnfs: fix nfs_direct_req ref leak when i/o falls back to the mds Message-ID: <20180112173928.wjefygafbw5fe6w4@tonberry.usersys.redhat.com> References: <20171213182220.lrfbvkaj65fg6abk@tonberry.usersys.redhat.com> <20171215211232.18942-1-smayhew@redhat.com> <20180111131038.vi25s2rvmjgqr3l6@tonberry.usersys.redhat.com> <1515676811.3463.0.camel@primarydata.com> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii In-Reply-To: <1515676811.3463.0.camel@primarydata.com> List-ID: On Thu, 11 Jan 2018, Trond Myklebust wrote: > On Thu, 2018-01-11 at 08:10 -0500, Scott Mayhew wrote: > > Trond, Anna, > > > > Any concerns with this patch? > > > > We left this run on multiple machines through the holidays... all > > doing > > direct I/O to a Netapp filer while doing storage failovers on an > > hourly > > basis on the filer. No more issues with processes getting stuck in > > inode_dio_wait. > > I'm fine with this version. Should it be a stable fix? If so, do we > have a "Fixes:" commit for it? Yeah it should probably be a stable fix. The hdr->release() call was added in commit 1ca018d28d ("pNFS: Fix a memory leak when attempted pnfs fails")... but I can reproduce this on earlier kernels than that (I tested 4.0) so it looks like this goes way back... -Scott > > -- > Trond Myklebust > Linux NFS client maintainer, PrimaryData > trond.myklebust@primarydata.com