ad0 READ_DMA TIMEOUT errors on install of 7.0-RELEASE

[Stephen Hurd]

Scott Long wrote:
> I'm betting that it's a driver problem and not
> a hardware problem, though you should probably think about migrating
> your data off to a new drive sometime soon.

Yeah, ordered a replacement drive today.

> I'd like to attack these driver problems.  What I need is to spend a
> couple of days with an affected system that can reliably reproduce the
> problem, instrumenting and testing the driver.  I have a number of
> theories about what might be going wrong, but nothing that I'm
> definitely sure about.  If you are willing to set up your system with
> remote power and remote serial, and if we knew a reliable way to
> reproduce the problem, I could probably have the problem identified and
> fixed pretty quickly.

Hrm... if my ISP allows multiple PPPoE sessions, the remote serial 
shouldn't be a problem.  Remote power though... would my idling in IRC 
and you poking me whenever you want the power state changed work?

I'm not completely certain it's even possible to turn the system in 
question off short of the physical power switch and even if it is, I'm 
positive that it's not possible to turn it back on again since there is 
no power switch header on the motherboard.

If I *can* get multiple PPPoE sessions running, the setup would be 
SSHing to a FreeBSD-sparc64 system with a serial and network connection 
to the affected system.  I could give you root access on one or both 
boxes as needed.

I would want you to make me feel a bit more comfortable about letting 
someone play with the ata driver on a system with a known flakey HD.  
Some kinda of "It's super-duper unlikely that I would thrash the FS" 
thing since I won't be able to put the new HD in for a week or two and I 
have the usual home system backup plan in place (that is, I plan on 
backing up someday...)

I'll try setting up the extra PPPoE session now (since I'm curious about 
it anyways) and get back to you on that detail.