[Intel-gfx] [kraxel at redhat.com: Re: [PATCH v9 5/7] vfio: Define vfio based dma-buf operations]

[Zhi Wang]

Hi Gred and Alex:
     Thanks for the reply! It would be better that kernel only provides 
functions instead of maintaining states from my point of view. If there 
is any existing async notification channel in vfio device fd? like 
reporting device events from vfio to QEMU? If so, It would be nice if we 
can extend that channel and notify the userspace application that guest 
framebuffer changed via that channel. So userspace application even 
doesn't need to call the ioctl periodically. :) If VM runs a 
single-buffered application.

Thanks,
Zhi.

On Fri, 2017-06-23 at 15:49 +0800, Zhi Wang wrote:
>> Hi:
>>       Thanks for the discussions! If the userspace application has
>> already maintained a LRU list, it looks like we don't need
>> generation
>> anymore,
> generation isn't required, things are working just fine without that.
> It is just a small optimization, userspace can skip the LRU lookup
> altogether if the generation didn't change.
>
> But of couse that only pays off if the kernel doesn't has to put much
> effort into maintaining the generation id.  Something simple like
> increasing it each time the guest writes a register which affects
> plane_info.
>
> If you think it doesn't make sense from the driver point of view we can
> drop the generation.
>
> cheers,
>    Gerd