Messages in this thread | | | Date | Mon, 1 Feb 2016 22:32:36 -0800 | From | Andrew Morton <> | Subject | Re: [PATCH] vmpressure: Fix subtree pressure detection |
| |
On Fri, 29 Jan 2016 11:37:49 +0300 Vladimir Davydov <vdavydov@virtuozzo.com> wrote:
> On Thu, Jan 28, 2016 at 08:24:30PM +0100, Vlastimil Babka wrote: > > On 28.1.2016 16:55, Michal Hocko wrote: > > > On Wed 27-01-16 19:28:57, Vladimir Davydov wrote: > > >> When vmpressure is called for the entire subtree under pressure we > > >> mistakenly use vmpressure->scanned instead of vmpressure->tree_scanned > > >> when checking if vmpressure work is to be scheduled. This results in > > >> suppressing all vmpressure events in the legacy cgroup hierarchy. Fix > > >> it. > > >> > > >> Fixes: 8e8ae645249b ("mm: memcontrol: hook up vmpressure to socket pressure") > > >> Signed-off-by: Vladimir Davydov <vdavydov@virtuozzo.com> > > > > > > a = b += c made me scratch my head for a second but this looks correct > > > > Ugh, it's actually a = b += a > > > > While clever and compact, this will make scratch their head anyone looking at > > the code in the future. Is it worth it? > > I'm just trying to be consistend with the !tree case, where we do > exactly the same.
I stared suspiciously at it for a while, decided to let it go. Possibly we can remove local `scanned' altogether. No matter, someone will clean it all up sometime.
| |