Re: [PATCH] net/skbuff: silence warnings under memory pressure

On 2019/09/04 17:25, Michal Hocko wrote:

On Wed 04-09-19 16:00:42, Sergey Senozhatsky wrote:

quoted

On (09/04/19 15:41), Sergey Senozhatsky wrote:

quoted

But the thing is different in case of dump_stack() + show_mem() +
some other output. Because now we ratelimit not a single printk() line,
but hundreds of them. The ratelimit becomes - 10 * $$$ lines in 5 seconds
(IOW, now we talk about thousands of lines).

And on devices with slow serial consoles this can be somewhat close to
"no ratelimit". *Suppose* that warn_alloc() adds 700 lines each time.
Within 5 seconds we can call warn_alloc() 10 times, which will add 7000
lines to the logbuf. If printk() can evict only 6000 lines in 5 seconds
then we have a growing number of pending logbuf messages.

Yes, ratelimit is problematic when the ratelimited operation is slow. I
guess that is a well known problem and we would need to rework both the
api and the implementation to make it work in those cases as well.
Essentially we need to make the ratelimit act as a gatekeeper to an
operation section - something like a critical section except you can
tolerate more code executions but not too many. So effectively

	start_throttle(rate, number);
	/* here goes your operation */
	end_throttle();

one operation is not considered done until the whole section ends.
Or something along those lines.

Regarding OOM killer which is serialized by oom_lock mutex, I proposed
"mm, oom: More aggressively ratelimit dump_header()." at
https://lkml.kernel.org/r/1550325895-9291-2-git-send-email-penguin-kernel@I-love.SAKURA.ne.jp
but it was ignored.

In this particular case we can increase the rate limit parameters of
course but I think that longterm we need a better api.

I proposed "printk: Introduce "store now but print later" prefix." at
https://lkml.kernel.org/r/1550896930-12324-1-git-send-email-penguin-kernel@I-love.SAKURA.ne.jp
but it was not accepted.

But I think that more better solution for warn_alloc() is to defer printing
 from allocating context (which may not be allowed to sleep) to WQ context.
I proposed "mm,oom: Defer dump_tasks() output." at
https://lkml.kernel.org/r/7de2310d-afbd-e616-e83a-d75103b986c6@i-love.sakura.ne.jp
and answered to Michal's concerns. I hope we can apply the same thing for
warn_alloc() and show_mem(). Then, we can minimize latency for both
"memory allocation failures" and "OOM killer invocations".