Re: [PATCH] printk/console: Do not suppress information about dropped messages

From: Sergey Senozhatsky
Date: Wed Feb 27 2019 - 03:44:45 EST


On (02/27/19 09:12), John Ogness wrote:
> >>
> >> My only objection to this is that the "messages dropped" only comes if a
> >> non-supressed message comes. So information about dropped information
> >> may never get printed unless some task prints something non-supressed.
> >>
> >> Imagine a situation where I am expecting a message to come, but don't
> >> see it because it was dropped. But if no more non-supressed messages
> >> come, I see neither the expected message nor the dropped message.
> >
> > I think this is exactly the problem (and thus the patch) we discussed
> > some 3 years ago.
>
> I guess you are referring to this [0] thread.

Right.

[..]
> I would agree with the proposed solution from 2016. My experience is
> that the dropped messages are very important. Yes, printing them could
> lead to the loss of even more messages.

Yes, printing out messages does take time. But I think it's easier to
start losing messages due to preemption under console_sem than due to
call_console_drivers() latencies.

> But still, it is important information that needs to get out.

I'd agree. A summary "you lost %d messages somewhere between current
and previous messages" is surely better than what we have now, but is
still a bit less informative than "you lost %d messages just now".

-ss