Hello,
El 14/04/15 a les 8.00, Gustau Pérez ha escrit:
> this morning I was able to the patch, and apparently it fixes the
> reboots with PV domU guests. Right now I'm installing debian as PV with
> no problems. Thank you.
>
> As I'm using 4.5 from the ports tree on this laptop, I'd li
On 14/04/2015 09:10, Roger Pau Monné wrote:
> Hello,
>
> El 14/04/15 a les 8.00, Gustau Pérez ha escrit:
>> this morning I was able to the patch, and apparently it fixes the
>> reboots with PV domU guests. Right now I'm installing debian as PV with
>> no problems. Thank you.
>>
>> As
> The interrupt timeouts are probably related to the IOMMU errata, the
> ACPI poweroff problem is because we need to negotiate the shutdown with
> the hypervisor, which we currently don't do. I have a very hacky patch
> for FreeBSD that should fix it, but I haven't found time to properly
> integr
Hello,
El 14/04/15 a les 9.40, Gustau Pérez ha escrit:
>
>
>> The interrupt timeouts are probably related to the IOMMU errata, the
>> ACPI poweroff problem is because we need to negotiate the shutdown with
>> the hypervisor, which we currently don't do. I have a very hacky patch
>> for FreeBSD t
On 14/04/2015 10:36, Roger Pau Monné wrote:
>>
>>Hello,
>>
>>I've tried the IN/OUT patch (I haven't tested the acpi shutdown patch
>> yet) but the problem remains the same. In the serial console I can see
>> different interruption problems (USB_ERR_TIMEOUTS, mpt0, bce0, etc).
> Can you
Hello,
Re-adding freebsd-xen ;).
El 14/04/15 a les 11.33, Gustau Pérez ha escrit:
>
>
> On 14/04/2015 10:36, Roger Pau Monné wrote:
>>>
>>>Hello,
>>>
>>>I've tried the IN/OUT patch (I haven't tested the acpi shutdown patch
>>> yet) but the problem remains the same. In the serial console
On 14/04/2015 11:51, Roger Pau Monné wrote:
> Thanks, could you add ioapic_ack=old to your xen_cmdline and try again?
Adding ioapic_ack=old seems to prevent the problem with the
interruptions. Anything else you want me to test?
Best
___
free
Hi,
the box I'm doing testing with has all its interfaces set with mtu
9000. That is a problem at domain creation time, because taps are set to
1500, and bridge [1] is created with the mtu of the first interface added.
So if you're using an mtu different than 1500 (it could also be les
