[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Xen-devel] Keeping PCI devices in low power states while bound to pciback module

Monday, October 27, 2014, 2:51:44 PM, you wrote:

>>>> On 27.10.14 at 14:28, <linux@xxxxxxxxxxxxxx> wrote:

>> Monday, October 27, 2014, 1:49:13 PM, you wrote:
>>>>>> On 26.10.14 at 02:09, <martin@xxxxxxxxxxxxxxxxx> wrote:
>>>> A device could be put i D3 state (if all other functions are owned by
>>>> pciback) when we transfer ownership to pciback. The patches I tried (links
>>>> sent by Konrad) toggled the power states. Is there a reason for not keeping
>>>> the power state set to D3 and bringing it back to D0 immediately before
>>>> assigning the function to a domU?
>> Please also keep in consideration the situation where the device is unbound 
>> from 
>> pciback in dom0 (to be bound to a driver/module in dom0).

> Not sure what you mean here - when a (real) driver attaches to a
> device, it shouldn't make assumptions on its power state anyway;
> I would assume the core PCI subsystem to take care of powering
> up devices in that process.

> Jan

Ok if it works that way that's fine, just wasn't sure and it is a use case that 
could easily be forgotten when testing if things still work as expected.


Xen-devel mailing list



Lists.xenproject.org is hosted with RackSpace, monitoring our
servers 24x7x365 and backed by RackSpace's Fanatical Support®.