The trail seems to have cooled off here, but it's pretty urgent. Having re-read the threads I find it notable that James hit a kswapd softlockup with "non-PREEMPT CGROUP but disabled GROUP_MEM_RES_CTLR". This suggests that the problem isn't with memcg. Or at least, we should fix this kswapd lockup before worrying about memcg. And I'm not sure that we should be assuming that there's something wrong in shrink_slab(). We know that kswapd has gone berserk, and that it will frequently call shrink_slab() when in that mode. But this may be because the top-level balance_pgdat() loop isn't terminating for reasons unrelated to shrink_slab(). -- To unsubscribe from this list: send the line "unsubscribe linux-fsdevel" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
