Re: [PATCH] kernel/watchdog: flush all printk nmi buffers when hardlockup detected

[Sergey Senozhatsky]

On (20/02/11 21:49), Steven Rostedt wrote:
> On Wed, 12 Feb 2020 10:15:51 +0900
> Sergey Senozhatsky <sergey.senozhatsky.work@xxxxxxxxx> wrote:
> 
> > On (20/02/10 12:48), Konstantin Khlebnikov wrote:
> > > 
> > > In NMI context printk() could save messages into per-cpu buffers and
> > > schedule flush by irq_work when IRQ are unblocked. This means message
> > > about hardlockup appears in kernel log only when/if lockup is gone.
> > > 
> > > Comment in irq_work_queue_on() states that remote IPI aren't NMI safe
> > > thus printk() cannot schedule flush work to another cpu.
> > > 
> > > This patch adds simple atomic counter of detected hardlockups and
> > > flushes all per-cpu printk buffers in context softlockup watchdog
> > > at any other cpu when it sees changes of this counter.
> > 
> > Petr, could you remind me, why do we do PRINTK_NMI_DIRECT_CONTEXT_MASK
> > only from ftrace?
> 
> Could it be because its from ftrace_dump() which can spit out millions
> of lines from NMI context?

Oh, yes, ftrace printks a lot. But I sort of forgot why don't we do
the same for "regular" NMIs. So NMIs use per-cpu buffers, expect for
NMIs which involve ftrace dump. I'm missing something here.

	-ss