On Sat, Apr 20, 2019 at 7:08 PM Zebediah Figura <z.figura12@gmail.com> wrote:

On 4/20/19 11:55 AM, Damjan Jovanovic wrote:
>
>
> On Sat, Apr 20, 2019 at 6:07 PM Zebediah Figura <z.figura12@gmail.com
> <mailto:z.figura12@gmail.com>> wrote:
>
>�� On 4/20/19 10:56 AM, Damjan Jovanovic wrote:
>�� >
>�� >
>�� > On Sat, Apr 20, 2019 at 5:44 PM Zebediah Figura
>�� <z.figura12@gmail.com <mailto:z.figura12@gmail.com>
>�� > <mailto:z.figura12@gmail.com <mailto:z.figura12@gmail.com>>> wrote:
>�� >
>�� >�� On 4/20/19 10:31 AM, Damjan Jovanovic wrote:
>�� >�� >
>�� >�� >
>�� >�� > On Sat, Apr 20, 2019 at 4:52 PM Zebediah Figura
>�� >�� <z.figura12@gmail.com <mailto:z.figura12@gmail.com>
>�� <mailto:z.figura12@gmail.com <mailto:z.figura12@gmail.com>>
>�� >�� > <mailto:z.figura12@gmail.com <mailto:z.figura12@gmail.com>
>�� <mailto:z.figura12@gmail.com <mailto:z.figura12@gmail.com>>>> wrote:
>�� >�� >
>�� >�� >�� On 4/20/19 8:15 AM, Damjan Jovanovic wrote:
>�� >�� >�� > Recent changes to qcap/v4l.c resulted in the
>�� removal of mmap()
>�� >�� >�� > for v4l2 devices, but read() is optional for device
>�� >�� drivers and
>�� >�� >�� > some don't support it. This isn't a major problem
>�� with the
>�� >�� >�� > presense of libv4l2 though, as it can emulate
>�� read() on top of
>�� >�� >�� > mmap(). However the code checks whether the device
>�� can read(),
>�� >�� >�� > and if not, doesn't even try to use it, even though
>�� it can.
>�� >�� >�� >
>�� >�� >�� > Fix this by only warning that read() is being
>�� emulated, and
>�� >�� >�� > continuing with libv4l2 if available.
>�� >�� >�� >
>�� >�� >�� > Also clarifies the logging.
>�� >�� >�� >
>�� >�� >�� > Signed-off-by: Damjan Jovanovic
>�� <damjan.jov@gmail.com <mailto:damjan.jov@gmail.com>
>�� >�� <mailto:damjan.jov@gmail.com <mailto:damjan.jov@gmail.com>>
>�� >�� >�� <mailto:damjan.jov@gmail.com
>�� <mailto:damjan.jov@gmail.com> <mailto:damjan.jov@gmail.com
>�� <mailto:damjan.jov@gmail.com>>>>
>�� >�� >�� > ---
>�� >�� >�� >�� dlls/qcap/v4l.c | 13 ++++++++-----
>�� >�� >�� >�� 1 file changed, 8 insertions(+), 5 deletions(-)
>�� >�� >�� >
>�� >�� >�� >
>�� >�� >�� >
>�� >�� >
>�� >�� >�� This doesn't seem right; libv4l2 should massage the
>�� result of
>�� >�� >�� VIDIOC_QUERYCAP to include V4L2_CAP_READWRITE. See
>�� >�� >
>�� >��
>�� <https://git.linuxtv.org/v4l-utils.git/tree/lib/libv4l2/libv4l2.c#n1210>.
>�� >�� >
>�� >�� >
>�� >�� > On FreeBSD 11.2 with libv4l-1.6.3_4, that is certainly NOT the
>�� >�� case. It
>�� >�� > could be a newer or not yet released feature.
>�� >�� >
>�� >�� >
>�� >�� >
>�� >
>�� >�� It's been the case since libv4l2 0.5:
>�� >��
>�� <https://git.linuxtv.org/v4l-utils.git/commit/lib/libv4l2/libv4l2.c?id=0215f2ac08c3ad0dc66ad8036f4e186a5e8f56d6>
>�� >
>�� >�� If FreeBSD is using 1.6.3, then I think something else is
>�� wrong. Are
>�� >�� you
>�� >�� sure that libv4l2 is actually being loaded?
>�� >
>�� >
>�� > int v4l2_ioctl(int fd, unsigned long int request, ...)
>�� > {
>�� > ...
>�� >�� if (devices[index].convert == NULL)
>�� >�� goto no_capture_request;
>�� >
>�� > <your code nippet from above>
>�� >
>�� > no_capture_request:
>�� >
>�� >
>�� >
>�� > So the code that would set that flag, is skipped if the .convert
>�� field
>�� > is NULL.
>�� >
>�� > When is it NULL?
>�� >
>�� >
>�� >
>
>�� If my reading of the code is correct, only when v4l2_fd_open() is
>�� called
>�� manually with V4L2_DISABLE_CONVERSION. Obviously we don't do that.
>
>�� Perhaps setting LIBV4L2_LOG_FILENAME might give some hints as to what's
>�� going wrong?
>
>
>
> int v4l2_fd_open(int fd, int v4l2_flags)
> {
> ...
> <The actual ioctl() returns cap.capabilities = 0x84a00001,
> cap.device_caps = 0x04a00000>
>
>�� if (cap.capabilities & V4L2_CAP_DEVICE_CAPS)
>�� cap.capabilities = cap.device_caps;
>�� if (!(cap.capabilities & V4L2_CAP_VIDEO_CAPTURE) ||
>�� !(cap.capabilities & (V4L2_CAP_STREAMING |
> V4L2_CAP_READWRITE))) {
>�� goto no_capture;
>�� }
>
> 0x000000001 = V4L2_CAP_VIDEO_CAPTURE
> 0x00200000 = V4L2_CAP_EXT_PIX_FORMAT
> 0x00800000 = V4L2_CAP_META_CAPTURE
> 0x040000000 = V4L2_CAP_STREAMING
> 0x800000000 = V4L2_CAP_DEVICE_CAPS
>
> So cap.capabilities have all those flags set, but cap.device_caps lacks
> V4L2_CAP_VIDEO_CAPTURE.
> cap.capabilities is overwritten by cap.device_caps, thus
> V4L2_CAP_VIDEO_CAPTURE is cleared.
> Since libv4l2 only manages devices with V4L2_CAP_VIDEO_CAPTURE set, and
> that flag is now missing, the first condition in the next "if" takes the
> goto, and .convert remains NULL.
> With .convert NULL, libv4l2 no longer manages the device, so
> communication is directly between Wine and kernel.
> Therefore the V4L2_CAP_READWRITE flag is not going to be set in
> v4l2_ioctl(), and libv4l2 cannot be made to do any conversions for us.
> But then again we can't use a device without V4L2_CAP_VIDEO_CAPTURE
> either, libv4l2 or not.
>
> I think the fundamental problem is that Wine is trying to use
> /dev/video1 which is broken, due to a bug in devenum that overwrites the
> working /dev/video0 registry entry with it, because their descriptions
> are the same.
>
>
>

Thanks, I've just sent patches to resolve both of these bugs.

Wow, so quickly!

Thank you!