Re: Re: pv_ops dom0 USB fixed

[Jeremy Fitzhardinge <jeremy@goop.org>]

Todd Deshane wrote:
>> That's because your initramfs only contains modules for
>> 2.6.27-7-generic but your Xen kernel is 2.6.28-rc8-tip so the ramdisk
>> fails to load the modules it needs.
>>     
>
> Yeah, I figured that out this morning as well (Sorry for the silly mistake
> and noise previously).
>
> I have 3 new logs though, that may give some more information
> into the detection of the USB disk as well as others.
>
> Still not booting, but it looks like it *should* work.
>
> The first shows that it clearly detects the usb disk, but as
> sda{1}, which makes some sense since it doesn't detect
> the other two SCSI disks. Also it detects it late...
>
> So, I then changed the root to sda1 and added a rootdelay
> option (twice I tried to add rootdelay, the second one I
> put up to 90 seconds even.
>
> I don't know if they shed any light or if there is something
> obvious that I could do to fix the situation.

So I've been getting things going on my test box here.  What I've observed:

When I set the SATA controller to either "Legacy" mode or "Native / IDE" 
mode in the BIOS, it works fine.

When I set it to AHCI, I see the same timeouts on IDENTIFY, but the DVD 
drive is detected properly.

USB seems to be working fine.  If I plug in a memory stick,  I can mount 
the filesystem and read things from it.


In these logs, it looks like you've overlooked the patch I added on 
Friday (vmap-allow-unflush-tlb.patch), and your udevd processes have 
been caught up in a deadlock which is preventing them from doing their 
thing.  I suspect that's the root cause of your problems.

    J