On Thu, Mar 19, 2020 at 07:27:56PM +0000, Ben Hutchings wrote:
> On Thu, 2020-03-19 at 07:37 +0100, Greg Kroah-Hartman wrote:
> > On Wed, Mar 18, 2020 at 06:29:06PM -0400, Sasha Levin wrote:
> > > On Wed, Mar 18, 2020 at 10:09:20PM +0000, Ben Hutchings wrote:
> > > > This commit (included in 5.6-rc5) seems to be needed for 5.4 and 5.5
> > > > branches:
> > > >
> > > > commit 6d390e4b5d48ec03bb87e63cf0a2bff5f4e116da
> > > > Author: yangerkun <yangerkun@xxxxxxxxxx>
> > > > Date: Wed Mar 4 15:25:56 2020 +0800
> > > >
> > > > locks: fix a potential use-after-free problem when wakeup a waiter
> > >
> > > I've queued it up for 5.5 and 5.4, thanks!
> > >
> > > > I'm a bit surprised that it hasn't yet been applied, while some fixes
> > > > from 5.6-rc6 have.
> > >
> > > Greg, I wonder if it makes sense to have you push a "Greg is here
> > > --->" "bookmark" in the form of a tag/branch on linux-stable-rc.git? at
> > > the very least it'll make it easy to see if something was missed or
> > > still waiting in the queue.
> >
> > To quote Jeff Layton:
> >
> > Hi Greg, there is a performance regression with this patch. We're
> > sorting through potential ways to address it at the moment, but you may
> > want to hold off until we have a fix for that merged.
> >
> > Sorry for the hassle!
> >
> > Which is why I dropped it for now.
> >
> > I'll go drop it again :)
>
> I didn't see any mention of this on the stable list though.
> I also don't think that a performance regression outweighs the
> seriousness of the bug being fixed.
>
> Ben.
>
Looks like a fix for the performance regression was committed yesterday
to mainline.
dcf23ac3e846c ("locks: reinstate locks_delete_block optimization")
Cheers,
Nathan
