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

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:   Wed, 12 Feb 2020 15:18:32 +0100
From:   Petr Mladek <pmladek@...e.com>
To:     Sergey Senozhatsky <sergey.senozhatsky.work@...il.com>
Cc:     Peter Zijlstra <peterz@...radead.org>,
        linux-kernel@...r.kernel.org, Steven Rostedt <rostedt@...dmis.org>,
        Sergey Senozhatsky <sergey.senozhatsky@...il.com>,
        Andrew Morton <akpm@...ux-foundation.org>,
        Dmitry Monakhov <dmtrmonakhov@...dex-team.ru>,
        Konstantin Khlebnikov <khlebnikov@...dex-team.ru>
Subject: Re: [PATCH] kernel/watchdog: flush all printk nmi buffers when
 hardlockup detected

On Wed 2020-02-12 10:15:51, Sergey Senozhatsky 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?

There was a possible deadlock when printing backtraces from all CPUs.
The CPUs were serialized via a lock in nmi_cpu_backtrace(). One of
them might have been interrupted under logbuf_lock.

ftrace was needed because it printed too many messages. And it was
safe because the ftrace log was read from a single CPU without
any lock.

Best Regards,
Petr