[ECOS] IXP425 RedBoot reset Ethernet problem
Mark Salter
msalter@redhat.com
Thu Mar 11 16:47:00 GMT 2004
>>>>> Casey Anderson writes:
> Hi,
> We are developing a product based on the IXDP425 Richfield board from
> Intel. RedBoot has been working just fine for our development activities
> on the IXDP425 board, but we are having some problems with the Ethernets
> on our own board.
> First off, we are using RedBoot 1.92 (custom built). It includes the
> Intel code for brining up the built-in Ethernet modules on the IXP425.
> We have noticed that if we:
> 1. power on the board
> 2. wait for the RedBoot prompt
> 3. type "reset"
> 4. then at the RedBoot prompt again we load the kernel and ramdisk
> the Ethernets come up in Linux every time. However if we just load Linux
> from a power-up, without doing a reset, the Ethernets won't come up
> (Linux THINKS they are up, but no pings go through). If we ifconfig down
> and ifconfig up the Ethernets a few times, they usually start passing
> traffic, but after about 100 packets the kernel panics.
> On the Richfield, the Ethernets come up every time (whether power-cycled
> or software "reset") and we can ping flood the Ethernet interfaces and
> all is well.
> I'm not very familiar with the RedBoot code so I'm not sure where to
> start looking to solve this problem, or to figure out what is going on.
> It could be a hardware difference between our boards and the Richfield,
> but the hardware engineer and I have gone over the design with a fine
> tooth comb and can't see anything that would cause this.
Seeing as the eth port[s] work for RedBoot, it might be best to approach
the debugging from the linux side. Hmm, but then you might not have the
sources to do that. :(
The reset command on Richfield simply jumps to the reset vector, so I don't
know why linux would behave the way you say. The board initialization after
a reset command should be exactly the same as a power-on reset.
--Mark
--
Before posting, please read the FAQ: http://ecos.sourceware.org/fom/ecos
and search the list archive: http://ecos.sourceware.org/ml/ecos-discuss
More information about the Ecos-discuss
mailing list