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:   John Ogness <john.ogness@linutronix.de>
To:     Marek Szyprowski <m.szyprowski@samsung.com>,
        Petr Mladek <pmladek@suse.com>
Cc:     Sergey Senozhatsky <sergey.senozhatsky.work@gmail.com>,
        Sergey Senozhatsky <sergey.senozhatsky@gmail.com>,
        Steven Rostedt <rostedt@goodmis.org>,
        Linus Torvalds <torvalds@linux-foundation.org>,
        Greg Kroah-Hartman <gregkh@linuxfoundation.org>,
        Thomas Gleixner <tglx@linutronix.de>,
        Peter Zijlstra <peterz@infradead.org>,
        Andrea Parri <parri.andrea@gmail.com>,
        Paul McKenney <paulmck@kernel.org>, kexec@lists.infradead.org,
        linux-kernel@vger.kernel.org
Subject: Re: [PATCH printk v5 6/6] printk: reimplement log_cont using record extension
In-Reply-To: <f1651593-3579-5820-6863-5f4973d2bfdc@samsung.com>
References: <20200914123354.832-1-john.ogness@linutronix.de> <20200914123354.832-7-john.ogness@linutronix.de> <CGME20200925190829eucas1p2f35866317389cd5e842e99b465632902@eucas1p2.samsung.com> <f1651593-3579-5820-6863-5f4973d2bfdc@samsung.com>
Date:   Sat, 26 Sep 2020 01:05:51 +0206
Message-ID: <87v9g1bgyg.fsf@jogness.linutronix.de>
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: quoted-printable
Precedence: bulk

On 2020-09-25, Marek Szyprowski <m.szyprowski@samsung.com> wrote:
> This patch landed recently in linux-next as commit f5f022e53b87=20
> ("printk: reimplement log_cont using record extension"). I've noticed=20
> that it causes a regression on my test system (ARM 32bit Samsung Exynos=20
> 4412-based Trats2 board). The messages are printed correctly on the=20
> serial console during boot, but then when I run 'dmesg' command, the log=
=20
> is truncated.
>
> Here is are the last lines of the dmesg log after this patch:
>
> [=C2=A0=C2=A0=C2=A0 6.649018] Waiting 2 sec before mounting root device...
> [=C2=A0=C2=A0=C2=A0 6.766423] dwc2 12480000.hsotg: new device is high-spe=
ed
> [=C2=A0=C2=A0=C2=A0 6.845290] dwc2 12480000.hsotg: new device is high-spe=
ed
> [=C2=A0=C2=A0=C2=A0 6.914217] dwc2 12480000.hsotg: new address 51
> [=C2=A0=C2=A0=C2=A0 8.710351] RAMDISK: squashfs filesystem found at block=
 0
>
> The corresponding dmesg lines before applying this patch:
>
> [=C2=A0=C2=A0=C2=A0 8.864320] RAMDISK: squashfs filesystem found at block=
 0
> [=C2=A0=C2=A0=C2=A0 8.868410] RAMDISK: Loading 37692KiB [1 disk] into ram=
 disk... /
> [=C2=A0=C2=A0=C2=A0 9.071670] /
> [=C2=A0=C2=A0=C2=A0 9.262498] /
> [=C2=A0=C2=A0=C2=A0 9.540711] /
> [=C2=A0=C2=A0=C2=A0 9.818031] done.

Ah. One of the more creative printk users...
init/do_mounts_rd.c:rd_load_image(). This is a set of LOG_CONT messages
that try to display a rotating line, complete with '\b' control
characters. The code is totally broken, but that is no excuse for printk
to break. It should be easy to reproduce on any architecture. I will
investigate it further. Thanks for reporting.

John Ogness