[U-Boot-Users] Netconsole

[Daren Hayward]

Does netconsole work from boot?? I have implemented it on my board, but if I
saveenv and reset, it comes up with stdxxx=serial in the environment!?!?!

netconsole works fine after implementing Travis' patch for the 440gx_enet,
but evey time I reset I have to use "setenv stdxxx nc"

Cheers

Daren

[Wolfgang Denk]

In message <C18BA5DDB58DD511BD0700C0DF0DD4502941B9 at NTSERVER4> you wrote:
> Does netconsole work from boot?? I have implemented it on my board, but if I

Not immediately after boot; only after most of the initializtion is
through, i. e. it starts right before entring the command main loop.

> saveenv and reset, it comes up with stdxxx=serial in the environment!?!?!

Feel free to add whatever you like to your preboot command.

> netconsole works fine after implementing Travis' patch for the 440gx_enet,
> but evey time I reset I have to use "setenv stdxxx nc"

Use CONFIG_PREBOOT then.

Best regards,

Wolfgang Denk

--
Software Engineering: Embedded and Realtime Systems, Embedded Linux
Phone: [phone] Fax: [phone] Email: wd at denx.de
Backed up the system lately?

[llandre]

> > Does netconsole work from boot?? I have implemented it on my board, but
> if I
>
>Not immediately after boot; only after most of the initializtion is
>through, i. e. it starts right before entring the command main loop.
>
> > saveenv and reset, it comes up with stdxxx=serial in the environment!?!?!
>
>Feel free to add whatever you like to your preboot command.
>
> > netconsole works fine after implementing Travis' patch for the 440gx_enet,
> > but evey time I reset I have to use "setenv stdxxx nc"
>
>Use CONFIG_PREBOOT then.

Hi Wolfgang,

what about netconsole on ARM architecture?
In my understanding, as Ethernet interface management is different from ppc,
it does not work out of the box. Do you confirm this?


Regards,

llandre

DAVE Electronics System House - R&D Department
web: http://www.dave-tech.it
email: r&d2 at dave-tech.it

[Wolfgang Denk]

Dear Andrea,

in message <6.0.1.1.0.20040907130301.01eaff18 at 192.168.2.1> you wrote:
>
> what about netconsole on ARM architecture?

netconsole does not depend on a specific processor architecture.

> In my understanding, as Ethernet interface management is different from ppc,
> it does not work out of the box. Do you confirm this?

There should be no relevant differences. If you check which files
contain NC specific code you will see that there is no hardware
dependend code involved:

-> find * -type f | xargs egrep -l CONFIG_NETCONSOLE
common/devices.c
drivers/netconsole.c
include/devices.h
include/configs/TQM860L.h
net/net.c

But I haven't run any tests on ARM myself (I think it was tested on
the "trab" board, but I may be wrong).

Best regards,

Wolfgang Denk

--
Software Engineering: Embedded and Realtime Systems, Embedded Linux
Phone: [phone] Fax: [phone] Email: wd at denx.de
It would be illogical to kill without reason
-- Spock, "Journey to Babel", stardate 3842.4

[llandre]

> > what about netconsole on ARM architecture?
>
>netconsole does not depend on a specific processor architecture.
>
> > In my understanding, as Ethernet interface management is different from
> ppc,
> > it does not work out of the box. Do you confirm this?
>
>There should be no relevant differences. If you check which files
>contain NC specific code you will see that there is no hardware
>dependend code involved:
>
> -> find * -type f | xargs egrep -l CONFIG_NETCONSOLE
> common/devices.c
> drivers/netconsole.c
> include/devices.h
> include/configs/TQM860L.h
> net/net.c
>
>But I haven't run any tests on ARM myself (I think it was tested on
>the "trab" board, but I may be wrong).

I'm tried to enable netconsole support on Cirrus EDB93xx (ARM920T). I
defined the two symbols:

CONFIG_NETCONSOLE
CONFIG_NET_MULTI

in order to include the necessary code (they are not mentioned in the
file doc/README.NetConsole; I think it is a good idea to add a couple
of lines to recall this).
After rebuilding U-Boot, ethernet interface stopped working and a new
message appeared ("Using unknown device") when trying to perform a TFTP
download.
It is not clear if it is somehow due to the variables ethprime and ethact that
are not defined or there are other ARM-related problems.
However it seems that the problems are caused by the symbol CONFIG_NET_MULTI.
Do you agree?



Best regards,

llandre

DAVE Electronics System House - R&D Department
web: http://www.dave-tech.it
email: r&d2 at dave-tech.it