| ↑back Search ←Prev date Next date→ Show only urls | (Click on time to select a line by its url) |
| Who | What | When |
|---|---|---|
| *** | Muzer has quit IRC (Ping timeout: 256 seconds) | [00:58] |
| ................................................................................................... (idle for 8h11mn) | ||
| KindOne has left | [09:09] | |
| ................................................................................. (idle for 6h40mn) | ||
| kbingham | sailus, Hi, I've just refreshed the max9286/rdacm20 gmsl set with an update from jmondi (fixing comments from Hans), and to resolve the build-bot error I got sent.
sailus, We're really hoping to get this series in, would you be able to take a look please and let me know if there is any blocker ? | [15:49] |
| ................................ (idle for 2h35mn) | ||
| *** | aballier has quit IRC (Ping timeout: 240 seconds) | [18:25] |
| .................... (idle for 1h38mn) | ||
| b-rad | hverkuil, assistance desired. Your commit a3fbc2e6bb05a has broken a bunch of usb dvb devices | [20:03] |
| pinchartl | b-rad: do the corresponding drivers fail to initialize the pad flags by any chance ? | [20:06] |
| b-rad | seems mc doesn't like tuner si2157 having a sink pad, so any cx231xx/em28xx device that passes media_dev to the tuner fails later during graph creation and theres no dvb device
all pads appear to have source and sink set the warning is mismatch of sink and source seems like si2157 is the only tuner with mc pads, so nothing to compare to | [20:06] |
| pinchartl | which of the checks fail ? | [20:08] |
| b-rad | if (WARN_ON(!(source->pads[source_pad].flags & MEDIA_PAD_FL_SOURCE)))
that is the line failing | [20:09] |
| *** | Dominionionion has quit IRC (Remote host closed the connection) | [20:11] |
| b-rad | since the em28xx devices which are failing don't even have such analog capability i'm tempted to remove the connection, but the cx231xx which fail do have analog | [20:11] |
| *** | lyakh has quit IRC (Ping timeout: 260 seconds) | [20:12] |
| pinchartl | what is the source entity in that case ? | [20:15] |
| b-rad | well the source would be the rf connector
which seems like it's setup by dvbdev this is the first time i've looked at mc dvb | [20:18] |
| pinchartl | can you print source->name ? | [20:21] |
| b-rad | lemme do some more instrumentation
having 1300+ entities some with 256 pads is one heck of an empty topology dump (unrelated) | [20:22] |
| pinchartl | pinchartl tries very hard not to comment on that
you're really really not making it easy :-) | [20:25] |
| b-rad | just surpised when i did a topology dump to see so much
the error is connecting the rf tuner sink to the demod sink source: si2157, sink: lgdt3306a behaviour is the same when sink is si2168 as well | [20:27] |
| pinchartl | is si2157 the rf tuner ? | [20:32] |
| b-rad | yes | [20:32] |
| pinchartl | did yo mean "the error is connecting the rf tuner *source* to the demod sink" | [20:32] |
| b-rad | seems like it should have a rf tuner source
in the driver it is rf input listed as sink | [20:33] |
| pinchartl | just to make sure I understand, the rf tuner (si2157) has an analog rf input and an analog baseband output, is that correct ? | [20:34] |
| b-rad | yes | [20:35] |
| pinchartl | and the kernel is trying to connect the si2157 *sink* pad to the lgdt3306a *sink* pad ? | [20:36] |
| b-rad | yup | [20:36] |
| pinchartl | so the code that tries to do that is at fault
it should connect the si2157 *source* pad to the lgdt3306a sink pad is it the cx231xx/em28xx drivers that create those links ? | [20:36] |
| b-rad | well in this case it's missing a source pad
the analog video and audio out would not go to the demod | [20:37] |
| pinchartl | the si2157 entity has no source pad ? | [20:38] |
| b-rad | they'd go to the cx25840 decoder
it has two analog source pads the si2157 only has three pads, all analog seems it's missing dtv source which would connect to demod | [20:38] |
| pinchartl | a brief look at the si2157 show that it has two outputs
so two source pads seems right (not that a given source pad can be connected to multiple sink pads if that's what happens at the hardware level) s/not/note/ | [20:39] |
| b-rad | both cx231x and em28xx have nearly identical functions to setup the links | [20:43] |
| pinchartl | I assume they're equally buggy then :-) | [20:47] |
| b-rad | i'd say so | [20:47] |
| pinchartl | you will have "fun" there, as I'm not sure there's any way for them to know which pad of the rf tuner to connect to
they can't assume that the analog video output will always be pad 1 for instance | [20:47] |
| b-rad | well i guess the race is on to figure it out | [20:48] |
| pinchartl | there are flags to report if a pad carries analog or digital data
PAD_SIGNAL_ANALOG for "analog signal", PAD_SIGNAL_DV for "digital video signal" and PAD_SIGNAL_AUDIO for "intermediate frequency analog audio or audio bitstream" | [20:49] |
| b-rad | i'll read more into the mc code setting it all up and see if i can sort it out
thanks for the input when does 5.7 merge window close this is a "little bit" important to fix | [20:57] |
| .............. (idle for 1h6mn) | ||
| sailus | kbingham: Thanks for the ping.
I'll try to look at it tomorrow. | [22:04] |
| ↑back Search ←Prev date Next date→ Show only urls | (Click on time to select a line by its url) |