Re: [syzbot] possible deadlock in start_this_handle (3)

[Tetsuo Handa]

On 2022/07/15 10:39, Shakeel Butt wrote:
>> I think mem_cgroup_print_oom_meminfo() should use GFP_ATOMIC, for it will fall into
>> infinite loop if kmalloc(GFP_NOFS) under oom_lock reached __alloc_pages_may_oom() path.
> 
> I would prefer GFP_NOWAIT. This is printing info for memcg OOMs and if
> the system is low on memory then memcg OOMs has lower importance than
> the system state.

Since killing a process in some memcg likely helps solving global OOM state,
system OOM condition might not be reported when memory allocation by
mem_cgroup_print_oom_meminfo() caused system OOM condition.

Therefore, we don't need to discard output from memcg OOM condition.