Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
Date:   Thu, 24 Sep 2020 10:40:10 +0900
From:   Sergey Senozhatsky <sergey.senozhatsky@gmail.com>
To:     Petr Mladek <pmladek@suse.com>
Cc:     Sergey Senozhatsky <sergey.senozhatsky@gmail.com>,
        Steven Rostedt <rostedt@goodmis.org>,
        John Ogness <john.ogness@linutronix.de>,
        Linus Torvalds <torvalds@linux-foundation.org>,
        Thomas Gleixner <tglx@linutronix.de>,
        Prarit Bhargava <prarit@redhat.com>,
        Mark Salyzyn <salyzyn@android.com>,
        Chunyan Zhang <zhang.lyra@gmail.com>,
        Orson Zhai <orsonzhai@gmail.com>,
        Changki Kim <changki.kim@samsung.com>,
        Sergey Senozhatsky <sergey.senozhatsky.work@gmail.com>,
        linux-kernel@vger.kernel.org
Subject: Re: [RFC 2/2] printk: Add more information about the printk caller
Message-ID: <20200924014010.GB577@jagdpanzerIV.localdomain>
References: <20200923135617.27149-1-pmladek@suse.com>
 <20200923135617.27149-3-pmladek@suse.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <20200923135617.27149-3-pmladek@suse.com>
Precedence: bulk

On (20/09/23 15:56), Petr Mladek wrote:
> The information about the printk caller has been added by the commit
> 15ff2069cb7f967da ("printk: Add caller information to printk() output.").
> The main motivation was to reconstruct original messages when they
> longer output from different CPUs got mixed.
> 
> But there are more usecases. The number of CPUs and tasks running on
> a single system is growing. It brings bigger challenges when debugging
> problems. The context in which kernel reports its state is sometimes
> pretty useful.
> 
> In particular, people suggest to remember the task id, cpu id, more details
> about the interrupt context, and even the task name [0][1].
> 
> Prepare for extending the caller information by replacing caller_id
> with struct printk_caller. And always store task id, cpu id, and
> exact interrupt context.

Wild idea:

Currently, we have two sides to the problem, I think. There are tasks
that store messages, but then there are tasks that print those messages
on the consoles. And those tasks and contexts are not always the same.
What I found helpful in the past was not only the caller_id, but the
emitter_id (especially preemption count and irqs state of the context
that prints messages on the slow consoles).

	-ss