| Message ID | 20220404163533.707508-6-p.zabel@pengutronix.de (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
| Series | [1/7] media: coda: set output buffer bytesused to appease v4l2-compliance | expand |
Le lundi 04 avril 2022 à 18:35 +0200, Philipp Zabel a écrit : > Allow to call G_PARM with type == V4L2_BUF_TYPE_VIDEO_CAPTURE, > to fix the following v4l2-compliance test failure: > > fail: v4l2-test-formats.cpp(1344): ret && node->has_frmintervals > test VIDIOC_G/S_PARM: FAIL So basically the rate written in the bitstream (if any) will be the same as the target real-time rate, which matches my reading of the new spec as what default behaviour we should have. Reviewed-by: Nicolas Dufresne <nicolas.dufresne@collabora.com> > > Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de> > --- > drivers/media/platform/chips-media/coda-common.c | 3 --- > 1 file changed, 3 deletions(-) > > diff --git a/drivers/media/platform/chips-media/coda-common.c b/drivers/media/platform/chips-media/coda-common.c > index c068c16d1eb4..33fcd8c7d72b 100644 > --- a/drivers/media/platform/chips-media/coda-common.c > +++ b/drivers/media/platform/chips-media/coda-common.c > @@ -1341,9 +1341,6 @@ static int coda_g_parm(struct file *file, void *fh, struct v4l2_streamparm *a) > struct coda_ctx *ctx = fh_to_ctx(fh); > struct v4l2_fract *tpf; > > - if (a->type != V4L2_BUF_TYPE_VIDEO_OUTPUT) > - return -EINVAL; > - > a->parm.output.capability = V4L2_CAP_TIMEPERFRAME; > tpf = &a->parm.output.timeperframe; > tpf->denominator = ctx->params.framerate & CODA_FRATE_RES_MASK;
Hi Philipp, On 04/04/2022 18:35, Philipp Zabel wrote: > Allow to call G_PARM with type == V4L2_BUF_TYPE_VIDEO_CAPTURE, > to fix the following v4l2-compliance test failure: > > fail: v4l2-test-formats.cpp(1344): ret && node->has_frmintervals > test VIDIOC_G/S_PARM: FAIL > > Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de> > --- > drivers/media/platform/chips-media/coda-common.c | 3 --- > 1 file changed, 3 deletions(-) > > diff --git a/drivers/media/platform/chips-media/coda-common.c b/drivers/media/platform/chips-media/coda-common.c > index c068c16d1eb4..33fcd8c7d72b 100644 > --- a/drivers/media/platform/chips-media/coda-common.c > +++ b/drivers/media/platform/chips-media/coda-common.c > @@ -1341,9 +1341,6 @@ static int coda_g_parm(struct file *file, void *fh, struct v4l2_streamparm *a) > struct coda_ctx *ctx = fh_to_ctx(fh); > struct v4l2_fract *tpf; > > - if (a->type != V4L2_BUF_TYPE_VIDEO_OUTPUT) > - return -EINVAL; > - > a->parm.output.capability = V4L2_CAP_TIMEPERFRAME; > tpf = &a->parm.output.timeperframe; > tpf->denominator = ctx->params.framerate & CODA_FRATE_RES_MASK; I think this is actually a v4l2-compliance bug, not a driver bug. G/S_PARM doesn't make sense for the capture queue of a stateful encoder, unless V4L2_FMT_FLAG_ENC_CAP_FRAME_INTERVAL is set to reserve HW resources. See https://hverkuil.home.xs4all.nl/spec/userspace-api/v4l/vidioc-enum-fmt.html#fmtdesc-flags That flags isn't used, so v4l2-compliance shouldn't complain. Try this v4l2-compliance patch to see if it resolved the fails for this patch and the next patch (7/7). v4l2-compliance patch: Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> --- diff --git a/utils/v4l2-compliance/v4l2-test-formats.cpp b/utils/v4l2-compliance/v4l2-test-formats.cpp index 3761b1fa..269a3832 100644 --- a/utils/v4l2-compliance/v4l2-test-formats.cpp +++ b/utils/v4l2-compliance/v4l2-test-formats.cpp @@ -1341,8 +1341,16 @@ static int testParmType(struct node *node, unsigned type) case V4L2_BUF_TYPE_VIDEO_OUTPUT: case V4L2_BUF_TYPE_VIDEO_OUTPUT_MPLANE: if (node->g_caps() & buftype2cap[type]) { - fail_on_test(ret && node->has_frmintervals); - fail_on_test(ret && node->has_enc_cap_frame_interval); + if (is_stateful_enc) { + if (V4L2_TYPE_IS_OUTPUT(type)) + fail_on_test(ret && node->has_frmintervals); + else if (node->has_enc_cap_frame_interval) + fail_on_test(ret); + else + fail_on_test(!ret); + } else { + fail_on_test(ret && node->has_frmintervals); + } } break; default:
On Do, 2022-04-21 at 12:30 +0200, Hans Verkuil wrote: [...] > I think this is actually a v4l2-compliance bug, not a driver bug. > > G/S_PARM doesn't make sense for the capture queue of a stateful encoder, unless > V4L2_FMT_FLAG_ENC_CAP_FRAME_INTERVAL is set to reserve HW resources. > > See https://hverkuil.home.xs4all.nl/spec/userspace-api/v4l/vidioc-enum-fmt.html#fmtdesc-flags > > That flags isn't used, so v4l2-compliance shouldn't complain. > > Try this v4l2-compliance patch to see if it resolved the fails for this patch > and the next patch (7/7). > > v4l2-compliance patch: > > Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> > --- > diff --git a/utils/v4l2-compliance/v4l2-test-formats.cpp b/utils/v4l2-compliance/v4l2-test-formats.cpp > index 3761b1fa..269a3832 100644 > --- a/utils/v4l2-compliance/v4l2-test-formats.cpp > +++ b/utils/v4l2-compliance/v4l2-test-formats.cpp > @@ -1341,8 +1341,16 @@ static int testParmType(struct node *node, unsigned type) > case V4L2_BUF_TYPE_VIDEO_OUTPUT: > case V4L2_BUF_TYPE_VIDEO_OUTPUT_MPLANE: > if (node->g_caps() & buftype2cap[type]) { > - fail_on_test(ret && node->has_frmintervals); > - fail_on_test(ret && node->has_enc_cap_frame_interval); > + if (is_stateful_enc) { > + if (V4L2_TYPE_IS_OUTPUT(type)) > + fail_on_test(ret && node->has_frmintervals); > + else if (node->has_enc_cap_frame_interval) > + fail_on_test(ret); > + else > + fail_on_test(!ret); > + } else { > + fail_on_test(ret && node->has_frmintervals); > + } > } > break; > default: You are right, this patch resolves the compliance failures addressed by patches 6 and 7. regards Philipp
On 21/04/2022 16:58, Philipp Zabel wrote: > On Do, 2022-04-21 at 12:30 +0200, Hans Verkuil wrote: > [...] >> I think this is actually a v4l2-compliance bug, not a driver bug. >> >> G/S_PARM doesn't make sense for the capture queue of a stateful encoder, unless >> V4L2_FMT_FLAG_ENC_CAP_FRAME_INTERVAL is set to reserve HW resources. >> >> See https://hverkuil.home.xs4all.nl/spec/userspace-api/v4l/vidioc-enum-fmt.html#fmtdesc-flags >> >> That flags isn't used, so v4l2-compliance shouldn't complain. >> >> Try this v4l2-compliance patch to see if it resolved the fails for this patch >> and the next patch (7/7). >> >> v4l2-compliance patch: >> >> Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> >> --- >> diff --git a/utils/v4l2-compliance/v4l2-test-formats.cpp b/utils/v4l2-compliance/v4l2-test-formats.cpp >> index 3761b1fa..269a3832 100644 >> --- a/utils/v4l2-compliance/v4l2-test-formats.cpp >> +++ b/utils/v4l2-compliance/v4l2-test-formats.cpp >> @@ -1341,8 +1341,16 @@ static int testParmType(struct node *node, unsigned type) >> case V4L2_BUF_TYPE_VIDEO_OUTPUT: >> case V4L2_BUF_TYPE_VIDEO_OUTPUT_MPLANE: >> if (node->g_caps() & buftype2cap[type]) { >> - fail_on_test(ret && node->has_frmintervals); >> - fail_on_test(ret && node->has_enc_cap_frame_interval); >> + if (is_stateful_enc) { >> + if (V4L2_TYPE_IS_OUTPUT(type)) >> + fail_on_test(ret && node->has_frmintervals); >> + else if (node->has_enc_cap_frame_interval) >> + fail_on_test(ret); >> + else >> + fail_on_test(!ret); >> + } else { >> + fail_on_test(ret && node->has_frmintervals); >> + } >> } >> break; >> default: > > You are right, this patch resolves the compliance failures addressed by > patches 6 and 7. Great! I've committed this change to v4l-utils. Regards, Hans
diff --git a/drivers/media/platform/chips-media/coda-common.c b/drivers/media/platform/chips-media/coda-common.c index c068c16d1eb4..33fcd8c7d72b 100644 --- a/drivers/media/platform/chips-media/coda-common.c +++ b/drivers/media/platform/chips-media/coda-common.c @@ -1341,9 +1341,6 @@ static int coda_g_parm(struct file *file, void *fh, struct v4l2_streamparm *a) struct coda_ctx *ctx = fh_to_ctx(fh); struct v4l2_fract *tpf; - if (a->type != V4L2_BUF_TYPE_VIDEO_OUTPUT) - return -EINVAL; - a->parm.output.capability = V4L2_CAP_TIMEPERFRAME; tpf = &a->parm.output.timeperframe; tpf->denominator = ctx->params.framerate & CODA_FRATE_RES_MASK;
Allow to call G_PARM with type == V4L2_BUF_TYPE_VIDEO_CAPTURE, to fix the following v4l2-compliance test failure: fail: v4l2-test-formats.cpp(1344): ret && node->has_frmintervals test VIDIOC_G/S_PARM: FAIL Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de> --- drivers/media/platform/chips-media/coda-common.c | 3 --- 1 file changed, 3 deletions(-)