RE: devboard_82 problems with etrax100LX

[Mikael Starvik <starvik@xxxxxxx.com>]

Hi,

ksymoops can be used with -v vmlinux -K -L -O -m System.map

/Mikael

-----Original Message-----
From: owner-dev-etrax@xxxxxxx.com">mailto:owner-dev-etrax@xxxxxxx.com] On Behalf Of Pieter Grimmerink
Sent: Tuesday, July 22, 2003 5:24 PM
To: dev-etrax
Subject: devboard_82 problems with etrax100LX


Hi all,

We're trying to run the devboard_82 environment on our custom board, based on the ETRAX100LX.

Our board has:

4MB flash
16MB SDRAM

With the devboard_lx distribution, and kernel 2.4.14 as well as 2.4.19 it runs fine. I also patched the tools / scripts to use jffs2, still no problem.

But now I decided that devboard_82 was probably the way to go, much less modifications than with the old devboard_lx distro.

I'm using the same kernelconfig as with 2.4.19 for devboard_lx, the board flashes and boots fine, also keeps running when left alone.

Problems occur though, when

-I do a lot of 'ls -la' commands in large directories, from a telnet shell (ls -la from the serial console does not work at all) -I try to copy a large file to tmpfs (cp /bin/busybox /var) Plenty of free RAM though

In both situations, the etrax hangs completely.
Tried to disable/enable the watchdog, no difference.

What could be wrong?

Some other strange thing:
-when I do a lot of http requests (browse through filesystem, open a few
files)
eventually the etrax crashes.

I've seen two different kinds of crashes resulting from this so far:

1. about 20 of the following error messages:
'bad pgdir entry 0x2 at 0x60e66000'

The etrax hangs after this. (no response on ethernet / serial console)

2. an oops occurs, I've attached the info.
I'm not sure what I can do to trace it, tried to use ksymoops with the etrax's System.map, got lots of errors, and eventually a traceback of 'timer_interrupt', but I don't know whether to trust this.


At first I thought this could be hardware related, but all these errors don't occur with devboard_lx and kernel 2.4.14 or 2.4.19, only with devboard_82 and kernel 2.4.20

Has anyone experienced the same kind of behaviour?
Could there be something wrong with my kernelconfig?

Best regards,

Pieter Grimmerink