Date: Thu, 13 Oct 2016 14:08:41 -0700 From: Mark Johnston <markj@freebsd.org> To: Conrad Meyer <cem@freebsd.org> Cc: src-committers <src-committers@freebsd.org>, svn-src-all@freebsd.org, svn-src-head@freebsd.org Subject: Re: svn commit: r307236 - head/sys/vm Message-ID: <20161013210841.GB52539@wkstn-mjohnston.west.isilon.com> In-Reply-To: <CAG6CVpWNOyOoeYFfMEKZy4qXju6ykRgRXBQDA3E2jLtwxhsFnw@mail.gmail.com> References: <201610132039.u9DKdY0u016480@repo.freebsd.org> <CAG6CVpWNOyOoeYFfMEKZy4qXju6ykRgRXBQDA3E2jLtwxhsFnw@mail.gmail.com>
next in thread | previous in thread | raw e-mail | index | archive | help
On Thu, Oct 13, 2016 at 01:49:29PM -0700, Conrad Meyer wrote: > On Thu, Oct 13, 2016 at 1:39 PM, Mark Johnston <markj@freebsd.org> wrote: > > Author: markj > > Date: Thu Oct 13 20:39:34 2016 > > New Revision: 307236 > > URL: https://svnweb.freebsd.org/changeset/base/307236 > > > > Log: > > Plug a potential vnode lock leak in vm_fault_hold(). > > This isn't just potential, right? I believe we hit this one > empirically with an assert in place of this if/vput. Sorry, right, that was bad wording. I meant that the lock is not always leaked in this error path - the map lookup needs to succeed initially, after which we need to block for the vnode lock, restart the fault handler, and then get an error from vm_map_lookup(). If the initial lookup fails, there's no leak.
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20161013210841.GB52539>