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;
From:   Sergey Senozhatsky <sergey.senozhatsky@gmail.com>
Date:   Wed, 20 May 2020 13:21:02 +0900
To:     Petr Mladek <pmladek@suse.com>
Cc:     Daniel Thompson <daniel.thompson@linaro.org>,
        Sergey Senozhatsky <sergey.senozhatsky@gmail.com>,
        Sumit Garg <sumit.garg@linaro.org>,
        Jason Wessel <jason.wessel@windriver.com>,
        Douglas Anderson <dianders@chromium.org>,
        kgdb-bugreport@lists.sourceforge.net,
        Linux Kernel Mailing List <linux-kernel@vger.kernel.org>,
        Arnd Bergmann <arnd@arndb.de>,
        Andrew Morton <akpm@linux-foundation.org>,
        Peter Zijlstra <peterz@infradead.org>,
        Steven Rostedt <rostedt@goodmis.org>,
        John Ogness <john.ogness@linutronix.de>
Subject: Re: [PATCH] printk/kdb: Redirect printk messages into kdb in any
 context
Message-ID: <20200520042102.GA938@jagdpanzerIV.localdomain>
References: <1589273314-12060-1-git-send-email-sumit.garg@linaro.org>
 <20200512142533.ta4uejwmq5gchtlx@holly.lan>
 <CAFA6WYOV7oPbYE=9fXueYMacb5wv0r9T6F8tmECt-Eafe-fctw@mail.gmail.com>
 <20200514084230.GO17734@linux-b0ei>
 <CAFA6WYPSsgdAB-wJC0e2YkVkW0XsqQsu5wrn4iB4M-cwvS7z2g@mail.gmail.com>
 <20200515085021.GS17734@linux-b0ei>
 <20200515103308.GD42471@jagdpanzerIV.localdomain>
 <20200515134806.5cw4xxnxw7k3223l@holly.lan>
 <20200518092139.GK7340@linux-b0ei>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <20200518092139.GK7340@linux-b0ei>
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

On (20/05/18 11:21), Petr Mladek wrote:
[..]
> > > Is this guaranteed that we never execute this path from NMI?
> 
> Good question!
> 
> > Absolutely not.
> > 
> > The execution context for kdb is pretty much unique... we are running a
> > debug mode with all CPUs parked in a holding loop with interrupts
> > disabled. One CPU is at an unknown exception state and the others are
> > either handling an IRQ or NMI depending on architecture[1].
> 
> This is similar to the situation in panic() when other CPUs are
> stopped. It is more safe when the CPUs are stopped using IRQ.
> There is higher danger of a deadlock when NMI is used.
> 
> bust_spinlock() is used in panic() to increase the chance to go over
> the deadlock and actually see the messages. It is not enough when
> more locks are used by the console (VT/TTY is good example). And
> it is not guaranteed that the console will still work after
> the hack is disabled by bust_spinlocks(0).

Good point. It's not guaranteed to help, but bust_spinlocks() does
help in general, many serial drivers do check oops_in_progress and
use a deadlock safe approach when locking port lock. I don't see
bust_spinlocks() being used in kdb, so it probably better start
doing so (along with general for_each_console() loop improvements,
like checking if console is enabled/available/etc).

[..]
> > > If so, can this please be added to the commit message? A more
> > > detailed commit message will help a lot.
> 
> What about?
> 
> "KDB has to get messages on consoles even when the system is stopped.
> It uses kdb_printf() internally and calls console drivers on its own.
> 
> It uses a hack to reuse an existing code. It sets "kdb_trap_printk"
> global variable to redirect even the normal printk() into the
> kdb_printf() variant.
> 
> The variable "kdb_trap_printk" is checked in printk_default() and
> it is ignored when printk is redirected to printk_safe in NMI context.
> Solve this by moving the check into printk_func().
> 
> It is obvious that it is not fully safe. But it does not make things
> worse. The console drivers are already called in this context by
> kdb_printf() direct calls."

This looks more informative indeed. Thanks!

	-ss