Re: [PULL 11/13] virtio-gpu: Handle resource blob commands

[BALATON Zoltan]

On Mon, 4 Nov 2024, Alex Bennée wrote:
> Dmitry Osipenko <dmitry.osipenko@collabora.com> writes:
>
> > On 11/1/24 20:16, Alex Bennée wrote:
> > > Also what is the subtlety behind using both stride and bytes_pp in the
> > > calculation. My naive thought would be:
> > >
> > >   fb.bytes_pp * ss.r.width == fb.stride
> > >
> > > Can anyone enlighten me?
> >
> > GPUs want image line size to be aligned to a power of 2 value, like 64
> > bytes for example. This aligned size of the line is called stride.
> >
> > GPU's DMA engine operates with a predefined granularity when it accesses
> > memory, it reads/writes memory chunks that are multiple of a stride.
> > GPUs almost never support memory accesses at a granularity of one byte,
> > like CPUs do it.
>
> Ok that seems worth covering in a comment. Also what is going on with:
>
>    fb->offset = ss->offsets[0] + ss->r.x * fb->bytes_pp + ss->r.y * fb->stride;
>
> and then calculating fbend from offset + the calculation? Is this
> because we need 2 full frames of storage? Can that ever change?

Stride is commonly used for storing data describing 2D blocks such as 
bitmaps so maybe not comment worthy but could be mentioned if you think it 
makes it clearer. Here's some documentation with a figure that may answer 
some of your questions:

https://learn.microsoft.com/en-us/windows/win32/medfound/image-stride

(I don't know what this thread is about, just stumbled upon this message 
and thought this may help. If not then sorry for the noise.)

Regards,
BALATON Zoltan