On Thu 09-07-15 14:03:53, David Rientjes wrote: > On Thu, 9 Jul 2015, Michal Hocko wrote: > > > > the titles were wrong for patches 2 and 3, but it doesn't mean we need to > > > add hacks around the code before organizing this into struct oom_control > > > > It is much easier to backport _fixes_ into older kernels (and yes I do > > care about that) if they do not depend on other cleanups. So I do not > > understand your point here. Besides that the cleanup really didn't make > > much change to the actuall fix because one way or another you still have > > to add a simple condition to rule out a heuristic/configuration which > > doesn't apply to sysrq+f path. > > > > So I am really lost in your argumentation here. > > > > This isn't a bugfix: sysrq+f has, at least for eight years, been able to > panic the kernel. This is an unwanted behavior and that is why I call it a bug. The mere fact that nobody has noticed because panic_on_oom is not used widely and even less with sysrq+f has nothing to do with it. > We're not fixing a bug, we're changing behavior. It's > quite appropriate to reorganize code before a behavior change to make it > cleaner. > > > > or completely pointless comments and printks that will fill the kernel > > > log. > > > > Could you explain what is so pointless about a comment which clarifies > > the fact which is not obviously visible from the current function? > > > > It states the obvious, a kthread is not going to be oom killed for > oom_kill_allocating_task: Sigh. The comment says that the force_kill path _runs_ from the kthread context which is far from obvious in out_of_memory. [...] > > Also could you explain why the admin shouldn't get an information if > > sysrq+f didn't kill anything because no eligible task has been found? > > The kernel log is the only notification mechanism that we have of the > kernel killing a process, we want to avoid spamming it unnecessarily. The > kernel log is not the appropriate place for your debugging information > that would only specify that yes, out_of_memory() was called, but there > was nothing actionable, especially when that trigger can be constantly > invoked by userspace once panicking is no longer possible. So how would you find out that there is no oom killable task? -- Michal Hocko SUSE Labs -- To unsubscribe, send a message with 'unsubscribe linux-mm' in the body to majordomo@xxxxxxxxx. For more info on Linux MM, see: http://www.linux-mm.org/ . Don't email: <a href=mailto:"dont@xxxxxxxxx";> email@xxxxxxxxx </a>