On Thu 03-08-17 14:17:26, Paul Moore wrote: > On Thu, Aug 3, 2017 at 7:05 AM, Michal Hocko <mhocko@xxxxxxxxxx> wrote: > > On Thu 03-08-17 19:44:46, Tetsuo Handa wrote: [...] > >> When allocating thread is selected as an OOM victim, it gets TIF_MEMDIE. > >> Since that function might be called from !in_interrupt() context, it is > >> possible that gfp_pfmemalloc_allowed() returns true due to TIF_MEMDIE and > >> the OOM victim will dip into memory reserves even when allocation failure > >> is not a problem. > > > > Yes this is possible but I do not see any major problem with that. > > I wouldn't add __GFP_NOMEMALLOC unless there is a real runaway of some > > sort that could be abused. > > Adding __GFP_NOMEMALLOC would not hurt anything would it? I is not harmfull but I fail to see how it would be useful either and as such it just adds a pointless gfp flag and confusion to whoever tries to modify the code in future. Really the main purpose of __GFP_NOMEMALLOC is to override the process scope PF_MEMALLOC. As such it is quite a hack and the fewer users we have the better. Btw. Should I resend the patch or somebody will take it from this email thread? -- Michal Hocko SUSE Labs
