Re: [PATCHv4 01/11] videobuf2: add cache management members

[Tomasz Figa]

On Mon, Mar 9, 2020 at 4:28 PM Hans Verkuil <hverkuil@xxxxxxxxx> wrote:
>
> On 3/9/20 8:25 AM, Sergey Senozhatsky wrote:
> > On (20/03/09 08:21), Hans Verkuil wrote:
> >> On 3/9/20 4:27 AM, Sergey Senozhatsky wrote:
> >>> On (20/03/07 12:47), Hans Verkuil wrote:
> >>>>
> >>>> Create those tests in v4l2-compliance: that's where they belong.
> >>>>
> >>>> You need these tests:
> >>>>
> >>>> For non-MMAP modes:
> >>>>
> >>>> 1) test that V4L2_BUF_CAP_SUPPORTS_CACHE_HINTS is never set.
> >>>>
> >>>> If V4L2_BUF_CAP_SUPPORTS_CACHE_HINTS is not set, then:
> >>>>
> >>>> 1) attempting to use V4L2_FLAG_MEMORY_NON_CONSISTENT will clear the flag
> >>>>    upon return (test with both reqbufs and create_bufs).
> >>>> 2) attempting to use V4L2_BUF_FLAG_NO_CACHE_INVALIDATE or V4L2_BUF_FLAG_NO_CACHE_CLEAN
> >>>>    will clear those flags upon return (do we actually do that in the patch series?).
> >
> > [..]
> >
> >>> I'm looking into it. Will it work if I patch the vivid test driver to
> >>> enable/disable ->allow_cache_hints bit per-node and include the patch
> >>> into the series. So then v4l2 tests can create some nodes with
> >>> ->allow_cache_hints.  Something like this:
> >>
> >> I would add a 'cache_hints' module parameter (array of bool) to tell vivid
> >> whether cache hints should be set or not for each instance. It would be useful
> >> to have this in vivid. Don't forget to update Documentation/media/v4l-drivers/vivid.rst
> >> as well.
> >
> > I see. Hmm, how do I do "test that V4L2_BUF_CAP_SUPPORTS_CACHE_HINTS
> > is never set" then?
>
> Not sure I understand your question. When requesting buffers for non-MMAP memory,
> this capability must never be returned. That has nothing to do with a cache_hints
> module option.

Have we decided that we explicitly don't want to support this for
USERPTR memory, even though technically possible and without much
extra code needed?

Best regards,
Tomasz