khym@azeotrope.org wrote:
> Right, I understand... but what I meant was that the presence or absence
> of video was unlikely to affect whether the ethernet driver could allocate
> its memory or not. If you didn't see the "receive DMA buffer not
> contiguous!" message when booting with a monitor, the problem you're
> having when booting without a monitor is probably unrelated to the
> ethernet.

I wasn't implying that video did affect ethernet.  I did not check for 
DMA error messages.  What I know is that I couldn't log in remotely, 
which _happened_ to coincide with being headless.  I think it unlikely 
that being headless is the problem, as that's how the machine has run 
for most of the past year under 2.1, and I hadn't heard of anyone else 
with that problem in 3.x either.  But I could be wrong...

> Hmm, how do you know it worked fine with a monitor attached if you didn't
> test network connectivity? And how do you know that only the network is
> the only thing that's the problem when booting without a monitor? Maybe

Again, as my assumption was simply an intermittent booting problem, 
"worked fine" meant that it did in fact boot.  Beyond that I hadn't had 
time to diagnose.

> Anyways, when you get a chance, see if you can try serial
> console... getting some more facts about what's happening will be more
> useful than speculating about the cause :)

This is true.  The problem is finding the time, and then seeing if I 
remember how to do a serial console, and then hoping I can find the 
cable to do it with.  :)


Tim

-- 
Tim & Alethea
christtrek.org