Re: HVM disc corruption (was Re: Xen won't start its device backend?)

[Roger Pau Monné <roger.pau%citrix.com@localhost>]

El 28/05/15 a les 12.05, Reinoud Zandijk ha escrit:
> On Wed, May 27, 2015 at 09:39:19PM +0100, Roger Pau Monn? wrote:
>> El 27/05/15 a les 21.29, Reinoud Zandijk ha escrit:
>>> It then reads the disklabel fine. Inspecting the code it seems to open the
>>> disc with O_BINARY, with a possible O_DIRECT / O_DSYNC. Reading it seems to do
>>> on a 512 block boundary for files opened with O_DIRECT and it uses lseek() in
>>> combination with the qemu_read(fd, buf, cnt) function.
>>>
>>> Is there any special behaviour on raw devices for such flags / functions that
>>> could trigger the bug?
>>
>> FWIW I had to craft the following patch for FreeBSD:
>>
>> http://git.qemu.org/?p=qemu.git;a=patch;h=3cad83075c7b847fe0eb6e61316fdf50984d4570
>>
>> Note that this is for Qemu upstream, not the qemu-xen fork that's used
>> by NetBSD, but I guess the issue could also be there.
> 
> Thanks for the patch. I haven't applied it yet but i did a search for the
> lseek()s in the ktrace i made of the qemu-dm and i havent't found a seek yet
> that is not a multiple of 0x200 so apparently it does do the sector size
> alignment.

IIRC NetBSD doesn't use Qemu upstream and instead relies on the old
qemu-xen fork, so it might not have this issue.

> Do you think it might be usefull to just try out a NetBSD variant of the
> patch?

Somebody should check if this issue is also present in qemu-xen. Since
FreeBSD Dom0 support is very new I've decided to only support Qemu
upstream there, so I have no idea what kind of issues one could face
when trying to use qemu-xen instead on BSD systems.

Roger.