get rid of console messages!
Carsten Haitzler (The Rasterman)
raster at openmoko.org
Fri May 16 05:14:47 CEST 2008
On Thu, 15 May 2008 21:44:08 -0500 "Mike (mwester)" <mwester at dls.net> babbled:
> Andy Green wrote:
> ...
> > | So clearly this will not work for the GTA01. Someone needs to verify
> > | that on the GTA02, this output will be directed to the tty associated
> > | with ttySAC2 (which is harmless on the GTA02), and not ttySAC0 (where it
> > | will be sent to the GSM just like the GTA01).
> >
> > Its probably OK to set loglevel to something low (from
> > Documentation/kernel-parameters.txt)
> >
>
> That will fix the kernel messages... but that's not the issue I'm seeing.
>
> Here's the funny stuff I'm seeing; I don't know why it does this. Note the
> command line - there's only the ttSAC0 specified as the console, just as
> raster suggests:
actually i suggested ttySAC2 :) and thus stdout gets its fd inherited from its
parent - which is init, which inherits the console fd - same as for kernel
output normally... thus... it all goes there UNLESS you tell it to go somewhere
else. i told it to live on ttySAC2 - all nicely vanished :)
> root at om-gta01:~# cat /proc/cmdline
> rootfstype=jffs2 root=/dev/mtdblock4 console=ttySAC0,115200 loglevel=8
> psplash=false mtdparts=neo1973-nand:0x00040000(u-boot),0x00004000(u-boot_env),
> 0x00200000(kernel),0x000a0000(splash),0x03d1c000(rootfs)
>
> Now here's a dump of the chatter on serial port 0. Note the output being
> sent to the port -- that's *NOT* from the kernel; rather it's gsmd's comms
> intermingled with stdout from the init process -- why did init decide that it
> could commandeer /dev/ttySAC0 for that purpose? So basically the absence of
> console=ttyS0 on the command line means that both gsmd and init seem to have
> /dev/ttySAC0 open, and are actively using it. Ouch.
>
> My point in bringing this up is that we need to verify on the GTA02 if it is
> the case that init will use the first tty device available (which would be
> bad, that would be ttySAC0 which is the GSM, and the GTA02 would have this
> same funky trash on the port that you see below!). Or, if it merely takes
> over the other console device on the command line, which in the case of the
> GTA02 would be ttySAC2, and would be harmless.
>
> (The GTA01 will just not be able to use the new boot splash, unless we come
> up with some other solution.)
>
> root at om-gta01:~# ./dump_serial_log
> 5007.065: -> Starting GSM daemon: gsmd - (C) 2006-2007 by OpenMoko, Inc.
> 5007.145: -> and contributors..This pro
> 5007.150: S M
> 5007.150: <- AT-Comma
> 5007.150: -> gram is FREE SOF
> 5007.150: <- nd Interpreter r
> 5007.150: -> TWARE under the
> 5007.150: <- eady..ating GSM
> 5007.155: -> terms of GNU GPL
> 5007.155: <- daemon: gsmd - (
> 5007.155: -> ....AT-Command I
> 5007.155: <- C) 2006-2007 by
> 5007.155: -> nterpreter ready
> 5007.155: <- OpenMoko, Inc. a
> 5007.155: -> ....ating GSM da
> 5007.155: <- nd contribut
> 5007.160: -> emon: gsmd - (C)
> 5007.160: <- ors.
> 5007.160: -> 2006-2007 by OpenMoko, Inc. and contributors..
> 5007.165: <- ..ERROR..
> 5007.170: -> ....ERROR....
> 5007.170: <- er the terms of
> 5007.170: -> .
> 5007.170: <- GNU GPL.
> 5007.170: S M
> 5007.175: <- ..OK.....OK..utors...OK..
> 5007.195: -> .
> 5007.200: <- ...OK..
> 5007.210: -> ATZ.
> 5007.215: <- ATZ...OK..
> 5007.290: -> gsmd.. .
> 5007.320: <- ...OK..
> 5007.335: -> ATE0V1.
> 5007.340: <- ATE0V1...OK..
> 5007.350: -> .
> 5007.355: <- ..OK..
> 5007.360: -> AT+CRC=1.
> 5007.370: <- ..OK..
> 5007.375: -> .
> 5007.380: <- ..OK..
> 5007.385: -> AT+CREG=2.
> 5007.395: <- ..OK..
> 5007.400: -> .Starting audio server:
> 5007.405: <- ..OK..
> 5007.410: -> AT+CMEE=1.
> 5007.420: <- ..ERROR..
> 5007.425: -> .
>
>
> Mike (mwester)
--
Carsten Haitzler (The Rasterman) <raster at openmoko.org>
More information about the distro-devel
mailing list