Messages in this thread | | | Date | Tue, 30 Jun 2015 15:50:24 -0700 (PDT) | From | David Rientjes <> | Subject | Re: [patch 2/3] mm, oom: pass an oom order of -1 when triggered by sysrq |
| |
On Fri, 19 Jun 2015, Michal Hocko wrote:
> > The force_kill member of struct oom_context isn't needed if an order of > > -1 is used instead. > > But this doesn't make much sense to me. It is not like we would _have_ > to spare few bytes here. The meaning of force_kill is clear while order > with a weird value is a hack. It is harder to follow without any good > reason. >
To me, this is the same as treating order == -1 as special in struct compact_control meaning that it was triggered from the command line and we really want to fully compact memory. It seems to have a nice symmetry.
| |