Re: [PATCH] sunrpc/cache: drop reference when sunrpc_cache_pipe_upcall() detects a race

[NeilBrown <neilb@suse.com>]

[-- Attachment #1: Type: text/plain, Size: 1198 bytes --]

On Wed, Mar 16 2016, J. Bruce Fields wrote:

> On Tue, Mar 15, 2016 at 01:43:37PM +0100, William Dauchy wrote:
>> Hello,
>> 
>> On Mon, Mar 14, 2016 at 9:38 PM, J. Bruce Fields <bfields@fieldses.org> wrote:
>> > Sorry for the delay.  I agree, it seems simple enough; applying for
>> > 4.6....
>> 
>> It it something we can consider for stable?
>
> It's an unlikely-looking race, I haven't seen it reported in the wild,
> and the consequences are just a reference-count leak (though admittedly
> those can sometimes cause crashes later--I haven't checked for that
> here).
>
> So I wasn't planning too, but any additional information's welcomed.

I agree that it isn't needed for stable - the race is tiny and the
consequence of losing the race is that entries get stuck in the cache
and possible an exported filesystem cannot be unmounted.  I don't think
that is a problem severe enough to meet the rules for -stable - though I
know the rules aren't always closely followed.

I wouldn't exactly object to it going to stable if someone wanted to
push, but I would rather it didn't appear in -stable before appearing in
a released (not -rc) kernel from Linus.

NeilBrown

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 818 bytes --]