Re: [microblaze-uclinux] Little problem with ML300

[John Williams <jwilliams@xxxxxxxxxxxxxx>]

Hi Roberto,

Strange problem - please see my comments below:

Roberto Farina wrote:
> 
> Xilinx Microprocessor Debug (XMD) Engine
> Xilinx EDK 6.1.1 Build EDK_G.13
> Copyright (c) 1995-2002 Xilinx, Inc.  All rights reserved.
> XMD% mbconnect stub -comm serial -port com1 -baud 57600
> Connecting to XMD stub..
> XMD communication stub initialized. Version No: 2
> Communicating with XMD stub on target board
> UART opened successfully
> Connected to MicroBlaze "stub" target. id = 0
> Starting GDB server for "stub" target (id = 0) at TCP port no 1234
> XMD% con 0x80000000
> USERSTOP at
>    A0000:   BA0C0118  brki    r16, 280
> XMD%

When control returns to xmd here, can you please execute the "rrd" 
command to dump the contents of the registers, then post it back here?

It's strange - at that stage in the boot sequence, interrupts have not 
been switched on (at least, not by the kernel).  So, it's unlikely to be 
an interrupt to a null vector or something like that.

It almost has me wondering if there's a problem with the memory address 
map of your hardware (mbvanilla_net) project.

Here's something to try.  Create a disassembly dump of your kernel 
image, with the command "mb-objdump -S image.elf > image.dump".  Then, 
search this dump file for the addresses of those printk statements.

Finally, from within XMD, check the contents of that address range (with 
mrd), *before and after* you have executed the kernel, to see if there's 
something funny going on there.

Let us know what you find...

Cheers,

John

___________________________
microblaze-uclinux mailing list
microblaze-uclinux@itee.uq.edu.au
Project Home Page : http://www.itee.uq.edu.au/~jwilliams/mblaze-uclinux
Mailing List Archive : http://www.itee.uq.edu.au/~listarch/microblaze-uclinux/