aboutsummaryrefslogtreecommitdiffstats
AgeCommit message (Collapse)AuthorFilesLines
2017-01-22Prepare for 1.12.0 releasev4l-utils-1.12.0Gregor Jasny6-1571/+2308
2017-01-19ir-ctl: ensure that device can record and that correct mode is setSean Young1-1/+20
Also ensure the correct recording and receiving mode is set. We might add new modes in the future (e.g. scancodes). Signed-off-by: Sean Young <sean@mess.org>
2017-01-19ir-ctl: uninitialised memory usedSean Young1-0/+1
We might end up with a corrupt rc6. Signed-off-by: Sean Young <sean@mess.org>
2017-01-18v4l2-ctl: add --stream-sleep optionHans Verkuil3-0/+26
Add option to periodically sleep for one second while streaming. This makes it easier to test what happens when an application cannot keep up with streaming. Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
2017-01-16ir-ctl: `strndupa' undefined with --disable-nlsGreg Whiteley1-0/+1
Signed-off-by: Greg Whiteley <greg.whiteley@atomos.com>
2017-01-16ir-ctl: rc5 command 6th bit missingSean Young1-2/+3
This is called extended rc5, not be confused with rc5x_20. Signed-off-by: Sean Young <sean@mess.org>
2017-01-16ir-ctl: rename rc5x to rc5x_20Sean Young3-6/+7
There are many extended rc5 protocols and we can only generate the 20 bit variant. Signed-off-by: Sean Young <sean@mess.org>
2017-01-16ir-keytable: "-p all" or "-p mce-kdb" does not workSean Young1-1/+1
When writing to the sysfs protocols file, use the underscore variant. The kernel does not accept "mce-kdb" and it never did. Signed-off-by: Sean Young <sean@mess.org>
2017-01-16ir-ctl: improve scancode validationSean Young3-2/+21
Ensure rc6 mce is that just that and that nec32 is not necx or nec. Signed-off-by: Sean Young <sean@mess.org>
2017-01-16ir-ctl: 0 is valid scancodeSean Young1-1/+1
Same for 0xffffffff. Signed-off-by: Sean Young <sean@mess.org>
2017-01-16ir-ctl: fix rc5x encodingSean Young1-2/+2
6th command bit was missing. Signed-off-by: Sean Young <sean@mess.org>
2017-01-16Update gitignoreGregor Jasny1-0/+2
2017-01-16v4l2-ctl: Remove vanished v4l2-tpg.h.patchGregor Jasny1-1/+1
2017-01-12v4l2-ctl: fix wrong Vsync in HDMI EDID detailed timingsHans Verkuil1-2/+2
The HDMI entry of the example EDIDs had a wrong Vsync size in the detailed timings for 1080p60: 6 instead of 5. Fix this. Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
2017-01-10cec-ctl: when setting both the PA and LAs, clear the LAs firstHans Verkuil1-8/+14
Transitioning first the PA, then the LAs, will cause double polling since after each change the framework attempts to claim logical addresses. It's better to clear the LAs, then set the PA and finally the LAs. Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
2017-01-09cec-ctl: the PA should be set before configuring the LAsHans Verkuil1-6/+6
cec-ctl -p 0.0.0.0 --tv would set the LA to 14 if the physical address before calling this command was != 0.0.0.0. The problem was that the physical address was set after the adapter was configured to a TV, and the CEC adapter will select LA 14 instead of 0 if the physical address is != 0.0.0.0. So first set the physical address, then claim the logical address(es). Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
2016-12-31dvb-fe-tool: add an option to set maximum number of stat samplesDietmar Spingler1-0/+7
Just like "ping", add a counter for the maximum number of statistics samples. Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
2016-12-31libdvbv5: fix handling of satellite numberMauro Carvalho Chehab3-4/+4
The logic that enforces satellite number to be a valid value is wrong: instead of using a bitmask, it was using a modulo. Also, we don't need to enforce it at dvb-scan/dvb-zap, as the library does that for us. Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
2016-12-28dvb-sat: fix logic that sets the satellite numberMauro Carvalho Chehab1-21/+23
The logic that sets DiSEqC logic is broken: it doesn't work for satellite #0, as the check logic there is wrong. Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
2016-12-28dvb-sat: reinsert legacy fields to avoid API breakagesMauro Carvalho Chehab2-2/+92
The struct dvb_sat_lnb has some fields that got deprecated in favor of a more complete approach. Yet, apps might be expecting those fields, so re-add them, to avoid compilation breakages. Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
2016-12-28dvb-sat: add support for several BrasilSat LNBf modelsMauro Carvalho Chehab1-0/+49
There are some LNBf manufactured by BrasilSat that are widely used in Brazil. Add support for them, based on what's described on this document: http://www.brasilsat.com.br/arquivo/DataSheet1.pdf I tested myself the custom GVT model. Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
2016-12-28dvb-sat: change the LNBf logic to make it more genericMauro Carvalho Chehab3-95/+137
There are some new LNBf models with more than two frequency ranges. Change the logic there to allow adding those new LNBf types. Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
2016-12-28dvb-sat: embeed most stuff internally at struct LNBfMauro Carvalho Chehab2-53/+66
We'll need to change the way dvb-sat.c handle LNBfs, in order to be able to support more advanced LNBf with 4 oscilators. So, let's first use an obscure most stuff for userspace. Please notice that, unfortunately, this will break the API. Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
2016-12-27cec-follower: improve two confusing warningsHans Verkuil1-4/+4
The second line of each warning started with a capital ('Feature Abort') which was confusing. Move the last word of the first line to the second line so that it now starts with 'replying Feature Abort'. This clarifies that the second line is a continuation from the first. Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
2016-12-20cec-compliance: warn instead of fail when unresponsiveHans Verkuil1-8/+8
If during power transitions the CEC adapter becomes unresponsive, then warn about that instead of failing. Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
2016-12-19cec-ctl: print vendor string if knownHans Verkuil1-3/+71
Map the numerical vendor ID to a string for known vendor IDs. Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
2016-12-19cec-ctl/cec-compliance: the transmit time for a byte is 24 ms, not 25Hans Verkuil2-4/+4
The time it takes for a byte to be transmitted on the CEC bus is 24 ms, not 25. Update the response time calculations accordingly. Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
2016-12-17cec-ctl: report approximate response timeHans Verkuil1-2/+14
Useful to know how long it takes to get the reply. Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
2016-12-16cec-compliance: add option to set reply time thresholdHans Verkuil2-2/+17
Make this programmable (default is 1 second). Also improve the calculation by subtracting the time it takes (approximately) to receive the message. That gives a more accurate response time. Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
2016-12-16cec-compliance: report which message took too longHans Verkuil1-1/+2
When a reply to a message takes more than 1s, then a warning is issued. But it didn't say which message was transmitted. Show the opcode to help debugging such situations. Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
2016-12-16cec-compliance: fail if initiate-arc times out, but is supportedHans Verkuil1-0/+1
If the remote CEC 2.0 device told us that it supports ARC-TX and the INITIATE_ARC message times out, then that's a fail. Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
2016-12-16cec-compliance: fix recognized/unrecognized_op logicHans Verkuil1-31/+21
First of all cec_msg_status_is_abort() can't be used with CEC_RECEIVE, instead check if the opcode is FEATURE_ABORT. Secondly, don't assume that when a transmitted message times out, it is not recognized. The reality is that the remote CEC device might be slow or temporarily unresponsive (happens frequently when going in or out of standby). So only mark an op as (un)recoginized if it is unambiguous. Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
2016-12-16cec-compliance: remove bogus Menu Status testHans Verkuil1-19/+0
This is not a valid test. It's only sent as reply to --menu-request, so sending this message to a remote CEC adapter is meaningless. This only makes sense in the follower, not in the compliance test. Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
2016-12-16cec-follower: double the time_to_transient/stable valuesHans Verkuil1-2/+2
These values where very short, which can cause race conditions where the compliance test missed the intermediate power status values (I think). It certainly doesn't hurt to wait longer before transitioning states. Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
2016-12-16cec-ctl: fix copy-and-paste error for --timeoutHans Verkuil1-1/+1
The timeout was ANDed with 0xf, that's not what you want. So a timeout of 2000 became 0... Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
2016-12-07cec-compliance: only test for has_deck_ctl for playback and record devicesHans Verkuil2-12/+23
The CEC 2.0 has_deck_ctl tests should only be done for playback and record devices. Other devices do not set that feature. Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
2016-12-07cec-compliance: add missing newlinesHans Verkuil2-7/+7
Some messages were missing newlines. Add them. Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
2016-12-06cec-ctl: show the CEC 2.0 featuresHans Verkuil1-17/+68
The cec-ctl -S option didn't show the features that the detected devices support. Add this. Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
2016-12-06cec-compliance: fix typo in usage messageHans Verkuil1-1/+1
It's dynamic-auto-lipsync, not audio-lipsync. Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
2016-12-06cec-compliance: become temporarily a follower in the lost-msgs testHans Verkuil1-0/+10
This test issues a lot of get-cec-version messages, and the remote device will reply with cec-version messages. If we are not a follower, and no other follower is running, then all these cec-version messages are feature aborted by the CEC core, which messes up this test (too much CEC traffic that can cause timeouts). So become a follower, just for this test, so that all these replies are just ignored. Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
2016-12-06cec-ctl: only show the reply if there was a replyHans Verkuil1-1/+1
The test whether or not to show the reply was not correct. Only if there actually was a reply (RX_STATUS_OK or RX_STATUS_FEATURE_ABORT) should the reply be shown. Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
2016-12-06cec-ctl: add --timeout option to override the default timeout.Hans Verkuil1-0/+8
The default 1000 ms timeout can now be overridden if desired. There is generally no reason to do this other than debugging since CEC devices are supposed to reply within a second. Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
2016-12-01cec-compliance: --test-adapter needs a valid PAHans Verkuil1-0/+5
If the --test-adapter test is run, and there is no valid physical address, then bail out with a clear message instead of producing a bunch of weird failures. Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
2016-12-01v4l2-compliance: allow S_SELECTION to return ENOTTYHans Verkuil1-11/+20
It is possible that G_SELECTION is implemented, but S_SELECTION isn't. So update the test to allow for ENOTTY return values from S_SELECTION. Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
2016-11-30v4l-utils: sync with latest kernelHans Verkuil3-0/+7
Main change is a TPG fix for the HSV formats. Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
2016-11-25v4l2-ctl: Show HSV encodings namesRicardo Ribalda1-1/+5
Add HSV encoding names to v4l2-ctl. I.e. Format Video Capture: Width/Height : 640/360 Pixel Format : 'HSV3' Field : None Bytes per Line : 1920 Size Image : 691200 Colorspace : sRGB Transfer Function : Default YCbCr/HSV Encoding: Hue 0 - 179 Quantization : Default Flags : Signed-off-by: Ricardo Ribalda Delgado <ricardo.ribalda@gmail.com> [hans: made some minor string changes] Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
2016-11-25qv4l2: Support for HSV encodingsRicardo Ribalda3-3/+10
Support set/get and override of HSV encodings. Signed-off-by: Ricardo Ribalda Delgado <ricardo.ribalda@gmail.com> [hans: made some minor string changes] Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
2016-11-25qv4l: support for HSV formats via OpenGL.Ricardo Ribalda1-1/+31
Support for HSV32 and HSV24. Signed-off-by: Ricardo Ribalda Delgado <ricardo.ribalda@gmail.com> Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
2016-11-25libv4lconvert: Add support for V4L2_PIX_FMT_{HSV24, HSV32}Ricardo Ribalda3-0/+140
HSV32 and HSV24 are single plane interleaved Hue Saturation Value formats. Signed-off-by: Ricardo Ribalda Delgado <ricardo.ribalda@gmail.com> Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
2016-11-24Drop audio.h/video.h: they are no longer needed.Hans Verkuil3-411/+0
Do not copy these headers anymore: they are no longer needed after the previous ivtv-ctl patch that removed their use. Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>

Privacy Policy