Hi, based on the recent discussions I have accumulated the following three patches. I haven't tested them yet but I would like to hear your opinion. The first patch only affects sysrq+f OOM killer. I believe it should be relatively uncontroversial. The patch 2 tweaks how we handle children tasks standing for the parent oom victim. This should help the test case Tetsuo shown [1]. The patch 3 is just a rough idea. I can see objections there but this is mainly to start discussion about ho to deal with small children which basically do not sit on any memory. Maybe we do not need anything like that at all and realy on multiple OOM invocations as a safer option. I dunno but I would like to hear your opinions. --- [1] http://lkml.kernel.org/r/201512292258.ABF87505.OFOSJLHMFVOQFt%40I-love.SAKURA.ne.jp -- 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>