From mboxrd@z Thu Jan  1 00:00:00 1970
Message-ID: <54690CC9.9070201@free.fr>
Date: Sun, 16 Nov 2014 21:44:57 +0100
From: Thierry Bultel <tbultel@free.fr>
MIME-Version: 1.0
References: <20141113145115.GD4656@sisyphus.hd.free.fr>
 <37845582.117817940.1415960100507.JavaMail.root@zimbra90-e16.priv.proxad.net>
 <20141114102833.GF4656@sisyphus.hd.free.fr>
In-Reply-To: <20141114102833.GF4656@sisyphus.hd.free.fr>
Content-Type: text/plain; charset="windows-1252"; format="flowed"
Content-Transfer-Encoding: quoted-printable
Subject: Re: [Xenomai] IMX kernel 3.0.35_4.1.0 +
 adeos-ipipe-3.0.43-mx6q-1.18-14 -> very slow boot
List-Id: Discussions about the Xenomai project <xenomai.xenomai.org>
List-Unsubscribe: <http://www.xenomai.org/mailman/options/xenomai>,
 <mailto:xenomai-request@xenomai.org?subject=unsubscribe>
List-Archive: <http://www.xenomai.org/pipermail/xenomai/>
List-Post: <mailto:xenomai@xenomai.org>
List-Help: <mailto:xenomai-request@xenomai.org?subject=help>
List-Subscribe: <http://www.xenomai.org/mailman/listinfo/xenomai>,
 <mailto:xenomai-request@xenomai.org?subject=subscribe>
To: Gilles Chanteperdrix <gilles.chanteperdrix@xenomai.org>
Cc: nicolas Mabire <nicolas.mabire@basystemes.fr>, xenomai@xenomai.org

Le 14/11/2014 11:28, Gilles Chanteperdrix a =E9crit :
> On Fri, Nov 14, 2014 at 11:15:00AM +0100, tbultel@free.fr wrote:
>> Gilles, it has taken time (about 3 hours), but I finally got
>> a backtrace with IPIPE_DEBUG: (I have disabled
>> CONFIG_IPIPE_TRACE_VMALLOC)
>
> Normally CONFIG_IPIPE_TRACE_VMALLOC is required on ARM. kmalloc fails
> to allocate large enough areas, which causes the kernel to fail booting.
>
>>
>> Before the trace went to the serial console, I saw that there were no mo=
re timer irq
>> on 2 CPUs (my ssh session froze). Some 3 or 4 seconds later it came up.
>> Not clear to me why there are only 100 points, because I have CONFIG_IPI=
PE_TRACE_SHIFT=3D14
>
> You have to write the number of back_trace_points you want to
> /proc/ipipe/trace/back_trace_points
>
> Also your trace contains a lot of writing to the console, you should
> move the ipipe_trace_panic_freeze to before any console output.
>
> Also note that the UART being slow, a lot of trace points will take
> a lot of time to print out.
>
> You may want to enable the tracer verbose mode to have the timestamps.
>

FYI, the kernel without IPIPE but with the PREEMPT_RT patch freezes as=20
well, when flooding with ping, and mostly when started after a poweroff.
I have not been able (yet) to freeze it after a reboot.
I will consider looking deeper in the bootloader.