Serial Console

Serial Console

[Aman]

Hi All

I have an custom board which has PPC 440GP as its processor. I have build a
kernel image and ramdisk to run on the custom board, which is running fine.
Now I want to disable the serial console in my kernel and use the serial
port for some other purpose. If I have to disable the serial console , what
command should be given for it in the kernel.

Thanking you in advance
Regards
Aman


** Sent via the linuxppc-embedded mail list. See http://lists.linuxppc.org/

Serial Console

[Aman]

 Hi

 In the Redhat 7.0 linux, I was able to enable/disable the serial console
 using the following command in the inittab
 " T0:12345:respawn:/sbin/getty ttyS0 9600 vt100"
 However in the case of the 440 Linux the serial console is enabled already.
 I wanted to know how this enabling is done. Can someone add some points to
 this.

 Thanking you in advance
 Regards
 Aman


** Sent via the linuxppc-embedded mail list. See http://lists.linuxppc.org/

Re: Serial Console

[Mark Hatle]

Aman wrote:
>  Hi
>
>  In the Redhat 7.0 linux, I was able to enable/disable the serial console
>  using the following command in the inittab
>  " T0:12345:respawn:/sbin/getty ttyS0 9600 vt100"

That is actually not a serial "console", it is a serial "login".  (It's a small
distinction but an important one)

On a workstation system (with say VGA), when the system boots the "console"
comes up and displays your boot messages.  Eventually init executes and starts a
login sesssion on the same device as the default console.  Thus your screen has
both a "console" and "login" active.

>  However in the case of the 440 Linux the serial console is enabled already.
>  I wanted to know how this enabling is done. Can someone add some points to
>  this.

Probably with your configuration, the default system console is on a serial
port.  This means all boot and kernel messages will go there.

Again you will want to start a login process via init in order to do multi-user
logins.

(But in the embedded world, init, getty/login and such aren't required.  You can
just as easily boot right into /bin/sh.)

Now back to your original question.. How is it enabled:

The console is enabled either by the system configuration or by the kernel
command line, "console=".  Look in the "Documentation" directory in the kernel,
something in there explains serial consoles and the cmdline semantics.

--Mark

>
>  Thanking you in advance
>  Regards
>  Aman
>
>
>
>
>


** Sent via the linuxppc-embedded mail list. See http://lists.linuxppc.org/

Re: Serial Console

[Ibon Gotxi Garcia]

On Tue, 2003-03-04 at 15:40, Mark Hatle wrote:

>
> The console is enabled either by the system configuration or by the kernel
> command line, "console=".  Look in the "Documentation" directory in the kernel,
> something in there explains serial consoles and the cmdline semantics.
>

newbie question:
is there any chance of having virtual terminals through the serial port?

Cheers.

--
Ibon Gotxi Garcia <igotxi@jazzfree.com>


** Sent via the linuxppc-embedded mail list. See http://lists.linuxppc.org/

Re: Serial Console

[Dave Wolfe]

[ Ibon Gotxi Garcia writes: ]
>
> newbie question:
> is there any chance of having virtual terminals through the serial port?

Check out screen ftp:ftp.uni-erlangen.de/pub/utilities/screen

   Screen is a full-screen window manager that multiplexes a physical ter-
   minal between several processes (typically interactive  shells).  Each
   virtual terminal provides the functions of a DEC VT100 terminal and, in
   addition, several control functions from the ISO 6429  (ECMA  48, ANSI
   X3.64)  and ISO 2022 standards (e.g. insert/delete line and support for
   multiple character sets).  There is a  scrollback  history buffer  for
   each virtual terminal and a copy-and-paste mechanism that allows moving
   text regions between windows.

and much, much more... It's not limited to serial terminals, BTW.

--
 Dave Wolfe

** Sent via the linuxppc-embedded mail list. See http://lists.linuxppc.org/

Serial console

[Daniel Ann]

Hi folks,

Just wondering if anyone could lend a hand with this problem I have
with serial console. I'm trying to boot up my board (very similar to
sandpoint using MPC8245) with kernel 2.6.12.3, and most of it is
working but console will display up to,
[snip]
RAMDISK: Compressed image found at block 0
VFS: Mounted root (ext2 filesystem) readonly.
Freeing unused kernel memory: 112k init                                    =
   =20

I've done series of printk in sys_execve() to see if /sbin/init is
working, and found out it went thru the whole rcS file okay. Mind you,
printk is successfully displaying the output on the console while I'm
still not getting anything from the user processes.

Having all the kernel boot up log on console means that I've done some
part right. But why am I not getting anything from the user processes
on the console screen ?

Is there anything I need to do on the kernel config ?

--=20
Daniel

Re: Serial console

[Josh Boyer]

On Sat, 2005-07-30 at 10:06 +0900, Daniel Ann wrote:
> Hi folks,
> 
> Just wondering if anyone could lend a hand with this problem I have
> with serial console. I'm trying to boot up my board (very similar to
> sandpoint using MPC8245) with kernel 2.6.12.3, and most of it is
> working but console will display up to,
> [snip]
> RAMDISK: Compressed image found at block 0
> VFS: Mounted root (ext2 filesystem) readonly.
> Freeing unused kernel memory: 112k init                                        
> 
> I've done series of printk in sys_execve() to see if /sbin/init is
> working, and found out it went thru the whole rcS file okay. Mind you,
> printk is successfully displaying the output on the console while I'm
> still not getting anything from the user processes.
> 
> Having all the kernel boot up log on console means that I've done some
> part right. But why am I not getting anything from the user processes
> on the console screen ?
> 
> Is there anything I need to do on the kernel config ?

Do you have a /dev/console device node in your initrd?  If not, that is
one of the reasons you could be seeing that problem.  Make
sure /dev/null is there too.

josh

Re: Serial console

[Ricardo Scop]

Hi Daniel,

On Friday 29 July 2005 22:06, Daniel Ann wrote:
> Hi folks,
>
> Just wondering if anyone could lend a hand with this problem I have
> with serial console. I'm trying to boot up my board (very similar to
> sandpoint using MPC8245) with kernel 2.6.12.3, and most of it is
> working but console will display up to,
> [snip]
> RAMDISK: Compressed image found at block 0
> VFS: Mounted root (ext2 filesystem) readonly.
> Freeing unused kernel memory: 112k init
>
> I've done series of printk in sys_execve() to see if /sbin/init is
> working, and found out it went thru the whole rcS file okay. Mind you,
> printk is successfully displaying the output on the console while I'm
> still not getting anything from the user processes.
>
> Having all the kernel boot up log on console means that I've done some
> part right. But why am I not getting anything from the user processes
> on the console screen ?
>
Please verify that you have a /dev/console in your rootfs, and that it's =
being=20
correctly acessed by the init process.

HTH,

--=20
Ricardo Scop.

        \|/
    ___ -*-
   (@ @)/|\
  /  V  \|  R SCOP Consult.
 /(     )\  Linux-based communications
--^^---^^+------------------------------
rscop@matrix.com.br
+55 51 999-36-777
Porto Alegre, RS - BRazil

Re: Serial console

[Daniel Ann]

Looks like I have /dev/console
0 crw-------  1 root root 5, 1 Sep 26  2002 console
0 crw-rw-rw-  1 root root 4, 64 Jun 17  2004 ttyS0

I havent yet figured out how I should go about seeing if init is
correctly accessing it tho. Let me take a look at busybox init part.
BTW, any more reasons why this problem might exist ?



On 7/30/05, Ricardo Scop <scop@digitel.com.br> wrote:
> Hi Daniel,
>=20
> On Friday 29 July 2005 22:06, Daniel Ann wrote:
> > Hi folks,
> >
> > Just wondering if anyone could lend a hand with this problem I have
> > with serial console. I'm trying to boot up my board (very similar to
> > sandpoint using MPC8245) with kernel 2.6.12.3, and most of it is
> > working but console will display up to,
> > [snip]
> > RAMDISK: Compressed image found at block 0
> > VFS: Mounted root (ext2 filesystem) readonly.
> > Freeing unused kernel memory: 112k init
> >
> > I've done series of printk in sys_execve() to see if /sbin/init is
> > working, and found out it went thru the whole rcS file okay. Mind you,
> > printk is successfully displaying the output on the console while I'm
> > still not getting anything from the user processes.
> >
> > Having all the kernel boot up log on console means that I've done some
> > part right. But why am I not getting anything from the user processes
> > on the console screen ?
> >
> Please verify that you have a /dev/console in your rootfs, and that it's =
being
> correctly acessed by the init process.
>=20
> HTH,
>=20
> --
> Ricardo Scop.
>=20
>        \|/
>    ___ -*-
>   (@ @)/|\
>  /  V  \|  R SCOP Consult.
>  /(     )\  Linux-based communications
> --^^---^^+------------------------------
> rscop@matrix.com.br
> +55 51 999-36-777
> Porto Alegre, RS - BRazil
>=20
>=20


--=20
Daniel

Re: Serial console

[Daniel Ann]

I too have /dev/null
0 crw-rw-rw-  1 root root 1, 3 Aug 31  2001 null

However, Im not using initrd. null, and console devices are in my RAMDISK t=
ho.

On 7/30/05, Josh Boyer <jwboyer@jdub.homelinux.org> wrote:
> On Sat, 2005-07-30 at 10:06 +0900, Daniel Ann wrote:
> > Hi folks,
> >
> > Just wondering if anyone could lend a hand with this problem I have
> > with serial console. I'm trying to boot up my board (very similar to
> > sandpoint using MPC8245) with kernel 2.6.12.3, and most of it is
> > working but console will display up to,
> > [snip]
> > RAMDISK: Compressed image found at block 0
> > VFS: Mounted root (ext2 filesystem) readonly.
> > Freeing unused kernel memory: 112k init
> >
> > I've done series of printk in sys_execve() to see if /sbin/init is
> > working, and found out it went thru the whole rcS file okay. Mind you,
> > printk is successfully displaying the output on the console while I'm
> > still not getting anything from the user processes.
> >
> > Having all the kernel boot up log on console means that I've done some
> > part right. But why am I not getting anything from the user processes
> > on the console screen ?
> >
> > Is there anything I need to do on the kernel config ?
>=20
> Do you have a /dev/console device node in your initrd?  If not, that is
> one of the reasons you could be seeing that problem.  Make
> sure /dev/null is there too.
>=20
> josh
>=20
>=20


--=20
Daniel

Re: Serial console

[Daniel Ann]

I've got a feeling that this has alot to do with IRQ.
Would I be correct to assume printk doesnt require interrupt to work
but printf does ?

After doing some testing, I've found my network interface ping-able
when I got printf working in 2.4.31. But if I alter the openpic source
to not initialize the interrupt, it also stop displaying at the same
point and network interface isnt pingable.

I guess my next problem would be then, why isnt my openpic working. :(
I've done exactly the same as what I did with 2.4.31 and that works.
Hmmm. Something must have changed in 2.6.12.3.

I dont think this is has anything to do with serial console, so I'm
gonna end the thread here.

Thanks.

On 7/30/05, Daniel Ann <ktdann@gmail.com> wrote:
> I too have /dev/null
> 0 crw-rw-rw-  1 root root 1, 3 Aug 31  2001 null
>=20
> However, Im not using initrd. null, and console devices are in my RAMDISK=
 tho.
>=20
> On 7/30/05, Josh Boyer <jwboyer@jdub.homelinux.org> wrote:
> > On Sat, 2005-07-30 at 10:06 +0900, Daniel Ann wrote:
> > > Hi folks,
> > >
> > > Just wondering if anyone could lend a hand with this problem I have
> > > with serial console. I'm trying to boot up my board (very similar to
> > > sandpoint using MPC8245) with kernel 2.6.12.3, and most of it is
> > > working but console will display up to,
> > > [snip]
> > > RAMDISK: Compressed image found at block 0
> > > VFS: Mounted root (ext2 filesystem) readonly.
> > > Freeing unused kernel memory: 112k init
> > >
> > > I've done series of printk in sys_execve() to see if /sbin/init is
> > > working, and found out it went thru the whole rcS file okay. Mind you=
,
> > > printk is successfully displaying the output on the console while I'm
> > > still not getting anything from the user processes.
> > >
> > > Having all the kernel boot up log on console means that I've done som=
e
> > > part right. But why am I not getting anything from the user processes
> > > on the console screen ?
> > >
> > > Is there anything I need to do on the kernel config ?
> >
> > Do you have a /dev/console device node in your initrd?  If not, that is
> > one of the reasons you could be seeing that problem.  Make
> > sure /dev/null is there too.
> >
> > josh
> >
> >
>=20
>=20
> --
> Daniel
>=20


--=20
Daniel

Re: Serial console

[Anton Wöllert]

Daniel Ann wrote:
> Hi folks,
> 
> Having all the kernel boot up log on console means that I've done some
> part right. But why am I not getting anything from the user processes
> on the console screen ?
> 
> Is there anything I need to do on the kernel config ?
> 

Hi, we've had the same problem some time ago on a tqm850l (mpc850). Our 
problem was, that we had registered two serial consoles in the kernel 
config or something like that. we had to disable all serial-driver stuff 
and enable just the platform-specific serial-console driver. we tracked 
this down with inserting a printk(buf) into the tty_write fs-op of the 
serial-port driver in drivers/char/tty_io.c. the ttyS0 was used but has 
taken another serial-driver than the platform specific.

may this help you :)


anton

Re: Serial console

[Daniel Ann]

Thanks for your input Anton.
I'm not all that certain about sandpoint platform has its own serial
driver so I've worked with 8250.c. Anyway, I've got it working now. It
was problem with interrupt. As soon as I changed interrupt from EGDE
to LEVEL it began working. Altho I'm faced with yet another problem
with openpic but I'm digging thru it at the moment.

BTW, I have no idea why EDGE didnt work and LEVEL does.

Cheers,
Daniel

On 7/30/05, Anton W=F6llert <a.woellert@gmail.com> wrote:
> Daniel Ann wrote:
> > Hi folks,
> >
> > Having all the kernel boot up log on console means that I've done some
> > part right. But why am I not getting anything from the user processes
> > on the console screen ?
> >
> > Is there anything I need to do on the kernel config ?
> >
>=20
> Hi, we've had the same problem some time ago on a tqm850l (mpc850). Our
> problem was, that we had registered two serial consoles in the kernel
> config or something like that. we had to disable all serial-driver stuff
> and enable just the platform-specific serial-console driver. we tracked
> this down with inserting a printk(buf) into the tty_write fs-op of the
> serial-port driver in drivers/char/tty_io.c. the ttyS0 was used but has
> taken another serial-driver than the platform specific.
>=20
> may this help you :)
>=20
>=20
> anton
>=20


--=20
Daniel

RE: Serial console

[JohnsonCheng]

Dear Daniel,

I also meet this problem on MPC8245 with linux-2.6.12.3, and I found it =
uses
LEVEL for interrupt in sandpoint.c. Do you have any idea for it?

Thanks,
Johnson Cheng

-----Original Message-----
From: linuxppc-embedded-bounces@ozlabs.org
[mailto:linuxppc-embedded-bounces@ozlabs.org] On Behalf Of Daniel Ann
Sent: Monday, August 01, 2005 9:27 AM
To: Anton W=F6llert
Cc: linuxppc-embedded@ozlabs.org
Subject: Re: Serial console

Thanks for your input Anton.
I'm not all that certain about sandpoint platform has its own serial
driver so I've worked with 8250.c. Anyway, I've got it working now. It
was problem with interrupt. As soon as I changed interrupt from EGDE
to LEVEL it began working. Altho I'm faced with yet another problem
with openpic but I'm digging thru it at the moment.

BTW, I have no idea why EDGE didnt work and LEVEL does.

Cheers,
Daniel

On 7/30/05, Anton W=F6llert <a.woellert@gmail.com> wrote:
> Daniel Ann wrote:
> > Hi folks,
> >
> > Having all the kernel boot up log on console means that I've done =
some
> > part right. But why am I not getting anything from the user =
processes
> > on the console screen ?
> >
> > Is there anything I need to do on the kernel config ?
> >
>=20
> Hi, we've had the same problem some time ago on a tqm850l (mpc850). =
Our
> problem was, that we had registered two serial consoles in the kernel
> config or something like that. we had to disable all serial-driver =
stuff
> and enable just the platform-specific serial-console driver. we =
tracked
> this down with inserting a printk(buf) into the tty_write fs-op of the
> serial-port driver in drivers/char/tty_io.c. the ttyS0 was used but =
has
> taken another serial-driver than the platform specific.
>=20
> may this help you :)
>=20
>=20
> anton
>=20


--=20
Daniel
_______________________________________________
Linuxppc-embedded mailing list
Linuxppc-embedded@ozlabs.org
https://ozlabs.org/mailman/listinfo/linuxppc-embedded

Re: Serial console

[Kumar Gala]

On Aug 1, 2005, at 10:38 PM, JohnsonCheng wrote:

> Dear Daniel,
>
> I also meet this problem on MPC8245 with linux-2.6.12.3, and I =20
> found it uses
> LEVEL for interrupt in sandpoint.c. Do you have any idea for it?

All internal interrupts on the MPC8245/1 should be LEVEL interrupts.

- kumar

> -----Original Message-----
> From: linuxppc-embedded-bounces@ozlabs.org
> [mailto:linuxppc-embedded-bounces@ozlabs.org] On Behalf Of Daniel Ann
> Sent: Monday, August 01, 2005 9:27 AM
> To: Anton W=F6llert
> Cc: linuxppc-embedded@ozlabs.org
> Subject: Re: Serial console
>
> Thanks for your input Anton.
> I'm not all that certain about sandpoint platform has its own serial
> driver so I've worked with 8250.c. Anyway, I've got it working now. It
> was problem with interrupt. As soon as I changed interrupt from EGDE
> to LEVEL it began working. Altho I'm faced with yet another problem
> with openpic but I'm digging thru it at the moment.
>
> BTW, I have no idea why EDGE didnt work and LEVEL does.
>
> Cheers,
> Daniel
>
> On 7/30/05, Anton W=F6llert <a.woellert@gmail.com> wrote:
>
>> Daniel Ann wrote:
>>
>>> Hi folks,
>>>
>>> Having all the kernel boot up log on console means that I've done =20=

>>> some
>>> part right. But why am I not getting anything from the user =20
>>> processes
>>> on the console screen ?
>>>
>>> Is there anything I need to do on the kernel config ?
>>>
>>>
>>
>> Hi, we've had the same problem some time ago on a tqm850l =20
>> (mpc850). Our
>> problem was, that we had registered two serial consoles in the kernel
>> config or something like that. we had to disable all serial-driver =20=

>> stuff
>> and enable just the platform-specific serial-console driver. we =20
>> tracked
>> this down with inserting a printk(buf) into the tty_write fs-op of =20=

>> the
>> serial-port driver in drivers/char/tty_io.c. the ttyS0 was used =20
>> but has
>> taken another serial-driver than the platform specific.
>>
>> may this help you :)
>>
>>
>> anton
>>
>>
>
>
> --=20
> Daniel
> _______________________________________________
> Linuxppc-embedded mailing list
> Linuxppc-embedded@ozlabs.org
> https://ozlabs.org/mailman/listinfo/linuxppc-embedded
>
> _______________________________________________
> Linuxppc-embedded mailing list
> Linuxppc-embedded@ozlabs.org
> https://ozlabs.org/mailman/listinfo/linuxppc-embedded
>

Re: Serial console

[Daniel Ann]

Right. Guess I found it hard way :)
After using LEVEL interrupt, it all works now.

On 8/2/05, Kumar Gala <kumar.gala@freescale.com> wrote:
>=20
> On Aug 1, 2005, at 10:38 PM, JohnsonCheng wrote:
>=20
> > Dear Daniel,
> >
> > I also meet this problem on MPC8245 with linux-2.6.12.3, and I
> > found it uses
> > LEVEL for interrupt in sandpoint.c. Do you have any idea for it?
>=20
> All internal interrupts on the MPC8245/1 should be LEVEL interrupts.
>=20
> - kumar
>=20
> > -----Original Message-----
> > From: linuxppc-embedded-bounces@ozlabs.org
> > [mailto:linuxppc-embedded-bounces@ozlabs.org] On Behalf Of Daniel Ann
> > Sent: Monday, August 01, 2005 9:27 AM
> > To: Anton W=F6llert
> > Cc: linuxppc-embedded@ozlabs.org
> > Subject: Re: Serial console
> >
> > Thanks for your input Anton.
> > I'm not all that certain about sandpoint platform has its own serial
> > driver so I've worked with 8250.c. Anyway, I've got it working now. It
> > was problem with interrupt. As soon as I changed interrupt from EGDE
> > to LEVEL it began working. Altho I'm faced with yet another problem
> > with openpic but I'm digging thru it at the moment.
> >
> > BTW, I have no idea why EDGE didnt work and LEVEL does.
> >
> > Cheers,
> > Daniel
> >
> > On 7/30/05, Anton W=F6llert <a.woellert@gmail.com> wrote:
> >
> >> Daniel Ann wrote:
> >>
> >>> Hi folks,
> >>>
> >>> Having all the kernel boot up log on console means that I've done
> >>> some
> >>> part right. But why am I not getting anything from the user
> >>> processes
> >>> on the console screen ?
> >>>
> >>> Is there anything I need to do on the kernel config ?
> >>>
> >>>
> >>
> >> Hi, we've had the same problem some time ago on a tqm850l
> >> (mpc850). Our
> >> problem was, that we had registered two serial consoles in the kernel
> >> config or something like that. we had to disable all serial-driver
> >> stuff
> >> and enable just the platform-specific serial-console driver. we
> >> tracked
> >> this down with inserting a printk(buf) into the tty_write fs-op of
> >> the
> >> serial-port driver in drivers/char/tty_io.c. the ttyS0 was used
> >> but has
> >> taken another serial-driver than the platform specific.
> >>
> >> may this help you :)
> >>
> >>
> >> anton
> >>
> >>
> >
> >
> > --
> > Daniel
> > _______________________________________________
> > Linuxppc-embedded mailing list
> > Linuxppc-embedded@ozlabs.org
> > https://ozlabs.org/mailman/listinfo/linuxppc-embedded
> >
> > _______________________________________________
> > Linuxppc-embedded mailing list
> > Linuxppc-embedded@ozlabs.org
> > https://ozlabs.org/mailman/listinfo/linuxppc-embedded
> >
>=20
>=20


--=20
Daniel

serial console

[KokHow Teh]

Hi;
      I have debian linux running on my PQ2FADS-ZU with devfs and seria=
l
driver compiled in. I have read Documentation/serial-console.txt and I =
have
tried both passing "console=3D" option as well as without passing that =
to the
kernel and both don't give me the expected results.
      When I pass "console=3Dtts/0,115200n8" to the kernel, I don't get=
 any
console output until the login prompt which I set to 115200 in
/etc/inittab. When I don't pass "console=3D" option to the kernel, I ge=
t
console output all the way until "Freeing unused kernel memory:" and th=
e
console output is garbled and I don't see any login prompt. Here is the=

/dev/* content:


(none) login:
Debian GNU/Linux 3.1 (none) tts/0
Password:
Last login: Thu Jan  1 02:29:06 1970 on tts/0

The programs included with the Debian GNU/Linux system are free softwar=
e;
the exact distribution terms for each program are described in the
individual files in /usr/share/doc/*/copyright.

Debian GNU/Linux comes with ABSOLUTELY NO WARRANTY, to the extent
permitted by applicable law.
(none):~# l /dev
total 0
crw-------  1 root root 8, 0 Jan  1 02:00 .devfsd
lr-xr-xr-x  1 root root   13 Jan  1 02:37 MAKEDEV -> /sbin/MAKEDEV
crw-------  1 root root 5, 1 Jan  1 02:00 console
drwxr-xr-x  1 root root    0 Jan  1 02:00 discs
crw-rw-rw-  1 root root 1, 7 Jan  1 02:00 full
drwxr-xr-x  1 root root    0 Jan  1 02:00 ide
prw-------  1 root root    0 Jan  1 02:37 initctl
crw-r-----  1 root root 1, 2 Jan  1 02:00 kmem
srw-rw-rw-  1 root root    0 Jan  1 02:37 log
crw-r-----  1 root root 1, 1 Jan  1 02:00 mem
drwxr-xr-x  1 root root    0 Jan  1 02:00 misc
drwxr-xr-x  1 root root    0 Jan  1 02:00 mtd
drwxr-xr-x  1 root root    0 Jan  1 02:00 mtdblock
crw-rw-rw-  1 root root 1, 3 Jan  1 02:00 null
crw-r-----  1 root root 1, 4 Jan  1 02:00 port
crw-rw-rw-  1 root root 5, 2 Jan  1 02:00 ptmx
drwxr-xr-x  1 root root    0 Jan  1 02:00 pts
drwxr-xr-x  1 root root    0 Jan  1 02:00 pty
crw-r--r--  1 root root 1, 8 Jan  1 02:00 random
drwxr-xr-x  1 root root    0 Jan  1 02:00 rd
lr-xr-xr-x  1 root root   33 Jan  1 02:37 root -> ide/host0/bus0/target=
0/lun0/part2
drwxrwxrwt  3 root root   60 Jan  1 02:37 shm
drwxr-xr-x  1 root root    0 Jan  1 02:00 soc
drwxr-xr-x  1 root root    0 Jan  1 02:00 tts
crw-rw-rw-  1 root root 5, 0 Jan  1 02:00 tty
crw-r--r--  1 root root 1, 9 Jan  1 02:37 urandom
prw-r-----  1 root adm     0 Jan  1 02:37 xconsole
crw-rw-rw-  1 root root 1, 5 Jan  1 02:00 zero
(none):~#

      This is what I end up without "console=3D" option passed to the k=
ernel:

<snip>
Mounted devfs on /dev
Freeing unused kernel memory: 232k init 4k prep
..=EF=BF=BD=EF=BF=BD=EF=BF=BD.=EF=BF=BD...=EF=BF=BD=EF=BF=BD=EF=BF=BD=EF=
=BF=BD=EF=BF=BD=EF=BF=BD=EF=BF=BD=EF=BF=BD.=EF=BF=BD=EF=BF=BD.=EF=BF=BD=
...=EF=BF=BD...=EF=BF=BD=EF=BF=BD=EF=BF=BD......=EF=BF=BD..=EF=BF=BD...=
=EF=BF=BD=EF=BF=BD=EF=BF=BD..=EF=BF=BD.=EF=BF=BD.=EF=BF=BD=EF=BF=BD.=EF=
=BF=BD=EF=BF=BD.....=EF=BF=BD=EF=BF=BD[42;124H.
 <no login prompt>

      My question is, how can I get a complete serial console output al=
l
the way from bootup to the login prompt? I am missing something. It is
either the kernel configuration or I have to tweak the codes somewhere.=
 Any
input and insight is appreciated.

Regards,
TEH
=

Re: serial console

[David Jander]

On Tuesday 04 October 2005 08:54, KokHow Teh wrote:
> Hi;
>       I have debian linux running on my PQ2FADS-ZU with devfs and serial
> driver compiled in. I have read Documentation/serial-console.txt and I have
> tried both passing "console=" option as well as without passing that to the
> kernel and both don't give me the expected results.
>       When I pass "console=tts/0,115200n8" to the kernel, I don't get any
> console output until the login prompt which I set to 115200 in
> /etc/inittab. When I don't pass "console=" option to the kernel, I get
> console output all the way until "Freeing unused kernel memory:" and the
> console output is garbled and I don't see any login prompt. Here is the
> /dev/* content:

I don't know exactly what I am talking about right now, but the string "tts/0" 
in the console-line sounds suspicious to me. AFAIK when using devfs still 
device names on the kernel-commandline have to be old-style.
Just a guess: try using "console=ttyS0,115200n8" and see if this helps.

Greetings,

-- 
David Jander

Re: serial console

[Vitaly Bordug]

> On Tuesday 04 October 2005 08:54, KokHow Teh wrote:
> 
>>Hi;
>>      I have debian linux running on my PQ2FADS-ZU with devfs and serial
>>driver compiled in. I have read Documentation/serial-console.txt and I have
>>tried both passing "console=" option as well as without passing that to the
>>kernel and both don't give me the expected results.
>>      When I pass "console=tts/0,115200n8" to the kernel, I don't get any
>>console output until the login prompt which I set to 115200 in
>>/etc/inittab. When I don't pass "console=" option to the kernel, I get
>>console output all the way until "Freeing unused kernel memory:" and the
>>console output is garbled and I don't see any login prompt. Here is the
>>/dev/* content:
> 
> 
> I don't know exactly what I am talking about right now, but the string "tts/0" 
> in the console-line sounds suspicious to me. AFAIK when using devfs still 
> device names on the kernel-commandline have to be old-style.
> Just a guess: try using "console=ttyS0,115200n8" and see if this helps.
> 
Yes, but PQ2FADS use cpm_uart and hence the commandline should be
console=ttyCPM0,115200

> Greetings,
> 


-- 
Sincerely,
Vitaly

Re: serial console

[KokHow Teh]

Hi;
      I have tried with 4 types of command lines with my linux kernel with
devfs and serial device compiled in and none gives me the desired results:

(1)   console=tts/0,115200n8 OR console=ttyCPM0,115200n8

      no serial console output at all during boot time but it gives me
login prompt at the end.

(2)   without "console=" OR console=ttyS0,115200n8

      serial console output fine during bootup but only until the point
when it prints out "Freeing unused kernel memory: 232k init 4k prep". After
that, the serial console output is garbled.

      So this has not solved my problem. I need more advice. Here are some
more questions and observations:

(1)   David Jander's comment:

I don't know exactly what I am talking about right now, but the string
"tts/0"
in the console-line sounds suspicious to me. AFAIK when using devfs still
device names on the kernel-commandline have to be old-style.
Just a guess: try using "console=ttyS0,115200n8" and see if this helps.

      Is it confirmed that kernel-commandline has to be old-style even
though the kernel is using devfs?

(2)   Vitaly Bordug's suggestion of using "console=ttyCPM0". May I know
where is ttyCPM0 defined and how is it bound to the serial console device
driver in the source?

      Thanks for any input and pointer.

Regards,
TEH



Vitaly Bordug <vbordug@ru.mvista.com> on 10/04/2005 11:16:10 PM

To:    KokHow.Teh@marconi.com
cc:    linuxppc-embedded list <linuxppc-embedded@ozlabs.org>

Subject:    Re: serial console



> On Tuesday 04 October 2005 08:54, KokHow Teh wrote:
>
>>Hi;
>>      I have debian linux running on my PQ2FADS-ZU with devfs and serial
>>driver compiled in. I have read Documentation/serial-console.txt and I
have
>>tried both passing "console=" option as well as without passing that to
the
>>kernel and both don't give me the expected results.
>>      When I pass "console=tts/0,115200n8" to the kernel, I don't get any
>>console output until the login prompt which I set to 115200 in
>>/etc/inittab. When I don't pass "console=" option to the kernel, I get
>>console output all the way until "Freeing unused kernel memory:" and the
>>console output is garbled and I don't see any login prompt. Here is the
>>/dev/* content:
>
>
> I don't know exactly what I am talking about right now, but the string
"tts/0"
> in the console-line sounds suspicious to me. AFAIK when using devfs still
> device names on the kernel-commandline have to be old-style.
> Just a guess: try using "console=ttyS0,115200n8" and see if this helps.
>
Yes, but PQ2FADS use cpm_uart and hence the commandline should be
console=ttyCPM0,115200

> Greetings,
>


--
Sincerely,
Vitaly

Re: serial console

[KokHow Teh]

Hi;
      I have tried with 4 types of command lines with my linux kernel with
devfs and serial device compiled in and none gives me the desired results:

(1)   console=tts/0,115200n8 OR console=ttyCPM0,115200n8

      no serial console output at all during boot time but it gives me
login prompt at the end.

(2)   without "console=" OR console=ttyS0,115200n8

      serial console output fine during bootup but only until the point
when it prints out "Freeing unused kernel memory: 232k init 4k prep". After
that, the serial console output is garbled. No login prompt.

      So this has not solved my problem. I need more advice. Here are some
more questions and observations:

(1)   David Jander's comment:

I don't know exactly what I am talking about right now, but the string
"tts/0"
in the console-line sounds suspicious to me. AFAIK when using devfs still
device names on the kernel-commandline have to be old-style.
Just a guess: try using "console=ttyS0,115200n8" and see if this helps.

      Is it confirmed that kernel-commandline has to be old-style even
though the kernel is using devfs?

(2)   Vitaly Bordug's suggestion of using "console=ttyCPM0". May I know
where is ttyCPM0 defined and how is it bound to the serial console device
driver in the source?

      Thanks for any input and pointer.

Regards,
TEH

Re: serial console

[David Jander]

On Tuesday 04 October 2005 17:16, Vitaly Bordug wrote:
> Yes, but PQ2FADS use cpm_uart and hence the commandline should be
> console=ttyCPM0,115200

You are totally right. I did that same mistake once in the past and again now. 
Shame on me ;-)
Now that I remeber, this might give trouble if you happen to use a normal uart 
(16550 or similar) also, and devfs. I had to slightly hack serial.c in my 
version of the kernel for this combination to work. I don't remeber right now 
exactly what I did, but it had to do with "SERIAL_DEV_OFFSET" and major/minor 
numbers ;-)
Anyway, I don't think your board has this combination.

Greetings,

-- 
David Jander

Re: serial console

[Vitaly Bordug]

KokHow Teh wrote:
> Hi;
>       I have tried with 4 types of command lines with my linux kernel with
> devfs and serial device compiled in and none gives me the desired results:
> 
> (1)   console=tts/0,115200n8 OR console=ttyCPM0,115200n8
> 
>       no serial console output at all during boot time but it gives me
> login prompt at the end.
> 
> (2)   without "console=" OR console=ttyS0,115200n8
> 
>       serial console output fine during bootup but only until the point
> when it prints out "Freeing unused kernel memory: 232k init 4k prep". After
> that, the serial console output is garbled. No login prompt.
> 
>       So this has not solved my problem. I need more advice. Here are some
> more questions and observations:
> 
> (1)   David Jander's comment:
> 
> I don't know exactly what I am talking about right now, but the string
> "tts/0"
> in the console-line sounds suspicious to me. AFAIK when using devfs still
> device names on the kernel-commandline have to be old-style.
> Just a guess: try using "console=ttyS0,115200n8" and see if this helps.
> 
>       Is it confirmed that kernel-commandline has to be old-style even
> though the kernel is using devfs?
> 
> (2)   Vitaly Bordug's suggestion of using "console=ttyCPM0". May I know
> where is ttyCPM0 defined and how is it bound to the serial console device
> driver in the source?
> 
Ummm.. The driver itself is in drivers/serial/cpm_uart. It does have the 
different major than ordinary ttyS... This relates to the 2.6.x kernel 
of course. I could suggest you to try the latest kernel.org release, in 
which you board is known to work fine.

>       Thanks for any input and pointer.
> 
> Regards,
> TEH
> 
> 
> 


-- 
Sincerely,
Vitaly