linux-kernel.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Petr Mladek <pmladek@suse.com>
To: John Ogness <john.ogness@linutronix.de>
Cc: linux-kernel@vger.kernel.org,
	Peter Zijlstra <peterz@infradead.org>,
	Sergey Senozhatsky <sergey.senozhatsky.work@gmail.com>,
	Steven Rostedt <rostedt@goodmis.org>,
	Daniel Wang <wonderfly@google.com>,
	Andrew Morton <akpm@linux-foundation.org>,
	Linus Torvalds <torvalds@linux-foundation.org>,
	Greg Kroah-Hartman <gregkh@linuxfoundation.org>,
	Alan Cox <gnomes@lxorguk.ukuu.org.uk>,
	Jiri Slaby <jslaby@suse.com>, Peter Feiner <pfeiner@google.com>,
	linux-serial@vger.kernel.org,
	Sergey Senozhatsky <sergey.senozhatsky@gmail.com>
Subject: Re: [RFC PATCH v1 20/25] serial: 8250: implement write_atomic
Date: Wed, 27 Feb 2019 10:46:55 +0100	[thread overview]
Message-ID: <20190227094655.ecdwhsc2bf5spkqx@pathway.suse.cz> (raw)
In-Reply-To: <20190212143003.48446-21-john.ogness@linutronix.de>

On Tue 2019-02-12 15:29:58, John Ogness wrote:
> Implement a non-sleeping NMI-safe write_atomic console function in
> order to support emergency printk messages.

It uses console_atomic_lock() added in 18th patch. That one uses
prb_lock() added by 2nd patch.

Now, prb_lock() allows recursion on the same CPU. But it still needs
to wait until it is released on another CPU.

It means that it is not completely save when NMIs happen on more CPUs in
parallel, for example, when calling nmi_trigger_cpumask_backtrace().

OK, it would be safe when prb_lock() is the only lock taken
in the NMI handler. But printk() should not make such limitation
to the rest of the system. Not to say, that we would most
likely need to add a lock back into nmi_cpu_backtrace()
to keep the output sane.


Peter Zijlstra several times talked about fully lockless
consoles. He is using the early console for debugging, see
the patchset
https://lkml.kernel.org/r/20170928121823.430053219@infradead.org

I am not sure if it is always possible. I personally see
the following way:

   1. Make the printk ring buffer fully lockless. Then we reduce
      the problem only to console locking. And we could
      have a per-console-driver lock (no the big lock like
      prb_lock()).

   2. I am afraid that we need to add some locking between CPUs
      to avoid mixing characters from directly printed messages.
      This would be safe everywhere expect in NMI. Then we could
      either risk ignoring the lock in NMI (there should be few
      messages anyway, the backtraces would get synchronized
      another way). Or we might need a compromise between
      handling console using the current owner and offload.


Best Regards,
Petr

  reply	other threads:[~2019-02-27  9:47 UTC|newest]

Thread overview: 147+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2019-02-12 14:29 [RFC PATCH v1 00/25] printk: new implementation John Ogness
2019-02-12 14:29 ` [RFC PATCH v1 01/25] printk-rb: add printk ring buffer documentation John Ogness
2019-02-12 14:45   ` Greg Kroah-Hartman
2019-02-12 14:29 ` [RFC PATCH v1 02/25] printk-rb: add prb locking functions John Ogness
2019-02-13 15:45   ` Petr Mladek
2019-02-13 21:39     ` John Ogness
2019-02-14 10:33       ` Petr Mladek
2019-02-14 12:10         ` John Ogness
2019-02-15 10:26           ` Petr Mladek
2019-02-15 10:56             ` John Ogness
2019-03-07  2:12   ` Sergey Senozhatsky
2019-02-12 14:29 ` [RFC PATCH v1 03/25] printk-rb: define ring buffer struct and initializer John Ogness
2019-02-12 14:46   ` Greg Kroah-Hartman
2019-02-14 12:46     ` Petr Mladek
2019-02-12 14:29 ` [RFC PATCH v1 04/25] printk-rb: add writer interface John Ogness
2019-02-14 15:16   ` Petr Mladek
2019-02-14 23:36     ` John Ogness
2019-02-15  1:19       ` John Ogness
2019-02-15 13:47       ` Petr Mladek
2019-02-17  1:32         ` John Ogness
2019-02-21 13:51           ` Petr Mladek
2019-02-12 14:29 ` [RFC PATCH v1 05/25] printk-rb: add basic non-blocking reading interface John Ogness
2019-02-18 12:54   ` Petr Mladek
2019-02-19 21:44     ` John Ogness
2019-02-21 16:22       ` Petr Mladek
2019-02-12 14:29 ` [RFC PATCH v1 06/25] printk-rb: add blocking reader support John Ogness
2019-02-18 14:05   ` Petr Mladek
2019-02-19 21:47     ` John Ogness
2019-02-12 14:29 ` [RFC PATCH v1 07/25] printk-rb: add functionality required by printk John Ogness
2019-02-12 17:15   ` Linus Torvalds
2019-02-13  9:20     ` John Ogness
2019-02-18 15:59   ` Petr Mladek
2019-02-19 22:08     ` John Ogness
2019-02-22  9:58       ` Petr Mladek
2019-02-12 14:29 ` [RFC PATCH v1 08/25] printk: add ring buffer and kthread John Ogness
2019-02-12 15:47   ` Sergey Senozhatsky
2019-02-19 13:54   ` Petr Mladek
2019-03-04  7:38   ` Sergey Senozhatsky
2019-03-04 10:00     ` Sergey Senozhatsky
2019-03-04 11:07       ` Sergey Senozhatsky
2019-03-05 21:00         ` John Ogness
2019-03-06 15:57           ` Petr Mladek
2019-03-06 21:17             ` John Ogness
2019-03-06 22:22               ` John Ogness
2019-03-07  6:41                 ` Sergey Senozhatsky
2019-03-07  6:51                   ` Sergey Senozhatsky
2019-03-07 12:50               ` Petr Mladek
2019-03-07  5:15           ` Sergey Senozhatsky
2019-03-11 10:51             ` John Ogness
2019-03-12  9:58               ` Sergey Senozhatsky
2019-03-12 10:30               ` Petr Mladek
2019-03-07 12:06     ` John Ogness
2019-03-08  1:31       ` Sergey Senozhatsky
2019-03-08 10:04         ` Petr Mladek
2019-02-12 14:29 ` [RFC PATCH v1 09/25] printk: remove exclusive console hack John Ogness
2019-02-19 14:03   ` Petr Mladek
2019-02-12 14:29 ` [RFC PATCH v1 10/25] printk: redirect emit/store to new ringbuffer John Ogness
2019-02-20  9:01   ` Petr Mladek
2019-02-20 21:25     ` John Ogness
2019-02-22 14:43       ` Petr Mladek
2019-02-22 15:06         ` John Ogness
2019-02-22 15:25           ` Petr Mladek
2019-02-25 12:11       ` Petr Mladek
2019-02-25 16:41         ` John Ogness
2019-02-26  9:45           ` Petr Mladek
2019-02-12 14:29 ` [RFC PATCH v1 11/25] printk_safe: remove printk safe code John Ogness
2019-02-22 10:37   ` Petr Mladek
2019-02-22 13:38     ` John Ogness
2019-02-22 15:15       ` Petr Mladek
2019-02-12 14:29 ` [RFC PATCH v1 12/25] printk: minimize console locking implementation John Ogness
2019-02-25 13:44   ` Petr Mladek
2019-02-12 14:29 ` [RFC PATCH v1 13/25] printk: track seq per console John Ogness
2019-02-25 14:59   ` Petr Mladek
2019-02-26  8:45     ` John Ogness
2019-02-26 13:11       ` Petr Mladek
2019-02-12 14:29 ` [RFC PATCH v1 14/25] printk: do boot_delay_msec inside printk_delay John Ogness
2019-02-12 14:29 ` [RFC PATCH v1 15/25] printk: print history for new consoles John Ogness
2019-02-26 14:58   ` Petr Mladek
2019-02-26 15:22     ` John Ogness
2019-02-27  9:02       ` Petr Mladek
2019-02-27 10:02         ` John Ogness
2019-02-27 13:12           ` Petr Mladek
2019-03-04  9:24       ` Sergey Senozhatsky
2019-02-12 14:29 ` [RFC PATCH v1 16/25] printk: implement CON_PRINTBUFFER John Ogness
2019-02-26 15:38   ` Petr Mladek
2019-02-12 14:29 ` [RFC PATCH v1 17/25] printk: add processor number to output John Ogness
2019-02-13 22:29   ` John Ogness
2019-02-12 14:29 ` [RFC PATCH v1 18/25] console: add write_atomic interface John Ogness
2019-02-12 14:29 ` [RFC PATCH v1 19/25] printk: introduce emergency messages John Ogness
2019-03-07  7:30   ` Sergey Senozhatsky
2019-03-08 10:31     ` Petr Mladek
2019-03-11 12:04       ` John Ogness
2019-03-12  2:51         ` Sergey Senozhatsky
2019-03-12  2:58       ` Sergey Senozhatsky
2019-02-12 14:29 ` [RFC PATCH v1 20/25] serial: 8250: implement write_atomic John Ogness
2019-02-27  9:46   ` Petr Mladek [this message]
2019-02-27 10:32     ` John Ogness
2019-02-27 13:55       ` Petr Mladek
2019-03-08  4:05         ` John Ogness
2019-03-08  4:17           ` John Ogness
2019-03-08 10:28           ` Petr Mladek
2019-02-12 14:29 ` [RFC PATCH v1 21/25] printk: implement KERN_CONT John Ogness
2019-02-12 14:30 ` [RFC PATCH v1 22/25] printk: implement /dev/kmsg John Ogness
2019-02-12 14:30 ` [RFC PATCH v1 23/25] printk: implement syslog John Ogness
2019-02-12 14:30 ` [RFC PATCH v1 24/25] printk: implement kmsg_dump John Ogness
2019-02-12 14:30 ` [RFC PATCH v1 25/25] printk: remove unused code John Ogness
2019-03-08 14:02   ` Sebastian Andrzej Siewior
2019-03-11  2:46     ` Sergey Senozhatsky
2019-03-11  8:18       ` Sebastian Andrzej Siewior
2019-03-12  9:38         ` Petr Mladek
2019-02-13  1:31 ` [RFC PATCH v1 00/25] printk: new implementation Sergey Senozhatsky
2019-02-13 13:43   ` John Ogness
2019-03-04  6:39     ` Sergey Senozhatsky
2019-02-13  1:41 ` Sergey Senozhatsky
2019-02-13 14:15   ` John Ogness
2019-03-04  5:31     ` Sergey Senozhatsky
2019-02-13  2:55 ` Sergey Senozhatsky
2019-02-13 14:43   ` John Ogness
2019-03-04  5:23     ` Sergey Senozhatsky
2019-03-07  9:53       ` John Ogness
2019-03-08 10:00         ` Petr Mladek
2019-03-11 10:54         ` Sergey Senozhatsky
2019-03-12 12:38           ` Petr Mladek
2019-03-12 15:15             ` John Ogness
2019-03-13  2:15               ` Sergey Senozhatsky
2019-03-13  8:19                 ` John Ogness
2019-03-13  8:40                   ` Sebastian Siewior
2019-03-13  9:27                     ` Sergey Senozhatsky
2019-03-13 10:06                       ` Sergey Senozhatsky
2019-03-14  9:27                       ` Petr Mladek
2019-03-13  8:46                   ` Sergey Senozhatsky
2019-03-14  9:14               ` Petr Mladek
2019-03-14  9:35                 ` John Ogness
2019-03-13  2:00             ` Sergey Senozhatsky
2019-02-13 16:54 ` David Laight
2019-02-13 22:20   ` John Ogness
2020-01-20 23:05 ` Eugeniu Rosca
2020-01-21 23:56   ` John Ogness
2020-01-22  2:34     ` Eugeniu Rosca
2020-01-22  7:31       ` Geert Uytterhoeven
2020-01-22 16:58         ` Eugeniu Rosca
2020-01-22 19:48           ` Geert Uytterhoeven
2020-01-24 16:09             ` Eugeniu Rosca
2020-01-27 12:32               ` Petr Mladek
2020-01-27 13:45                 ` Eugeniu Rosca
2020-01-22 10:33       ` John Ogness
2020-01-24 12:13         ` Eugeniu Rosca

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20190227094655.ecdwhsc2bf5spkqx@pathway.suse.cz \
    --to=pmladek@suse.com \
    --cc=akpm@linux-foundation.org \
    --cc=gnomes@lxorguk.ukuu.org.uk \
    --cc=gregkh@linuxfoundation.org \
    --cc=john.ogness@linutronix.de \
    --cc=jslaby@suse.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-serial@vger.kernel.org \
    --cc=peterz@infradead.org \
    --cc=pfeiner@google.com \
    --cc=rostedt@goodmis.org \
    --cc=sergey.senozhatsky.work@gmail.com \
    --cc=sergey.senozhatsky@gmail.com \
    --cc=torvalds@linux-foundation.org \
    --cc=wonderfly@google.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body. 
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).