Re: [PATCH] cfq: explicitly use 64bit divide operation for 64bit arguments

[Jens Axboe <axboe-tSWWG44O7X1aa/9Udqfwiw@public.gmane.org>]

On Sun, Sep 22 2013, Tejun Heo wrote:
> (cc'ing Jens)
> 
> Hello,
> 
> On Fri, Sep 13, 2013 at 09:20:45AM -0700, Anatol Pomozov wrote:
> > Hi, Tejun
> > 
> > On Fri, Aug 30, 2013 at 10:52 AM, Anatol Pomozov
> > <anatol.pomozov-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org> wrote:
> > > 'samples' is 64bit operant, but do_div() second parameter is 32.
> > > do_div silently truncates high 32 bits and calculated result
> > > is invalid.
> > >
> > > In case if low 32bit of 'samples' are zeros then do_div() produces
> > > kernel crash.
> > >
> > > Signed-off-by: Anatol Pomozov <anatol.pomozov-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>
> > > ---
> > >  block/cfq-iosched.c | 2 +-
> > >  1 file changed, 1 insertion(+), 1 deletion(-)
> > >
> > > diff --git a/block/cfq-iosched.c b/block/cfq-iosched.c
> > > index d5bbdcf..c410752 100644
> > > --- a/block/cfq-iosched.c
> > > +++ b/block/cfq-iosched.c
> > > @@ -1803,7 +1803,7 @@ static u64 cfqg_prfill_avg_queue_size(struct seq_file *sf,
> > >
> > >         if (samples) {
> > >                 v = blkg_stat_read(&cfqg->stats.avg_queue_size_sum);
> > > -               do_div(v, samples);
> > > +               v = div64_u64(v, samples);
> > >         }
> > >         __blkg_prfill_u64(sf, pd, v);
> > >         return 0;
> > 
> > what do you think about this patch?
> 
> Sorry about the delay.  Was traveling.
> 
>   Acked-by: Tejun Heo <tj-DgEjT+Ai2ygdnm+yROfE0A@public.gmane.org>
> 
> blkcg patches are routed through Jens's block tree but I haven't been
> able to get hold of Jens lately.  Jens?

Sorry, been a bit crazy on my end. Will queue it up for 3.12, thanks.

-- 
Jens Axboe