On Wed 27-06-18 19:10:41, Andrew Morton wrote: > On Sat, 23 Jun 2018 22:12:51 +0800 ufo19890...@gmail.com wrote: > > > From: yuzhoujian <yuzhouj...@didichuxing.com> > > > > The current system wide oom report prints information about the victim > > and the allocation context and restrictions. It, however, doesn't > > provide any information about memory cgroup the victim belongs to. This > > information can be interesting for container users because they can find > > the victim's container much more easily. > > > > I follow the advices of David Rientjes and Michal Hocko, and refactor > > part of the oom report. After this patch, users can get the memcg's > > path from the oom report and check the certain container more quickly. > > > > The oom print info after this patch: > > oom-kill:constraint=<constraint>,nodemask=<nodemask>,oom_memcg=<memcg>,task_memcg=<memcg>,task=<commm>,pid=<pid>,uid=<uid> > > > > Signed-off-by: yuzhoujian <yuzhouj...@didichuxing.com> > > --- > > Below is the part of the oom report in the dmesg > > ... > > [ 134.873392] panic invoked oom-killer: > > gfp_mask=0x6280ca(GFP_HIGHUSER_MOVABLE|__GFP_ZERO), nodemask=(null), > > order=0, oom_score_adj=0 > > > > ... > > > > [ 134.873480] > > oom-kill:constraint=CONSTRAINT_NONE,nodemask=(null),task_memcg=/test/test1/test2,task=panic,pid= > > 8669, uid= 0 > > We're displaying nodemask twice there. Avoidable? > > Also, the spaces after pid= and uid= don't seem useful. Why not use > plain old %d?
I've been discussing this with yuzhoujian off-list so please drop the current pile -- Michal Hocko SUSE Labs
