| ↑back Search ←Prev date Next date→ Show only urls | (Click on time to select a line by its url) |
| Who | What | When |
|---|---|---|
| *** | shoragan has quit IRC (Read error: Network is unreachable) | [00:04] |
| shoragan has joined #linux-media
mtretter has quit IRC (Read error: Connection reset by peer) mtretter has joined #linux-media | [00:10] | |
| shoragan has quit IRC (Quit: quit) | [00:21] | |
| shoragan has joined #linux-media | [00:26] | |
| ................ (idle for 1h15mn) | ||
| lexano has quit IRC (Ping timeout: 480 seconds) | [01:41] | |
| .............. (idle for 1h6mn) | ||
| danitool has quit IRC (Quit: Cubum autem in duos cubos, aut quadratoquadratum in duos quadratoquadratos) | [02:47] | |
| ............ (idle for 59mn) | ||
| xecutercmd has joined #linux-media | [03:46] | |
| xecuter has quit IRC (Ping timeout: 480 seconds) | [03:52] | |
| ....................... (idle for 1h50mn) | ||
| mvaittin has joined #linux-media | [05:42] | |
| ........ (idle for 38mn) | ||
| djrscally has joined #linux-media
djrscally has quit IRC () mvaittin has quit IRC (Ping timeout: 480 seconds) djrscally has joined #linux-media | [06:20] | |
| jm_h | Hi there ! I am working on adding imx708 support up to the (next to come) unicam upstream driver. I started from https://github.com/raspberrypi/linux/commits/rpi-6.6.y/drivers/media/i2c/imx708.c on top of pinchartl 's branch pinchartl/rpi/v6.9/overlays
After a few modifications unrelated to unicam, I tested it on my board with a imx708_noir and it fails in v4l2_subdev_enable_streams() when calling v4l2_subdev_call(sd, pad, enable_streams...). At the very least, it seems that the .{enable,disable}_streams() need to be implemented (even if in v4l2_subdev_enable_streams() there is a test v4l2_subdev_enable_streams_fallback()) | [06:27] |
| It does not seem to be enough though as it (badly) fails in v4l2_subdev_enable_streams: https://paste.debian.net/1316242/ | [06:35] | |
| *** | mvaittin has joined #linux-media | [06:36] |
| pinchartl | I don't recall a requirement for sensor drivers to use .enable_streams() to be compatible | [06:37] |
| but if your sensor driver exposes multiple pads, it won't be compatible with the helpers | [06:46] | |
| jm_h | So, if the driver does not implement streams yet, it must have only one pad to be compatible | [06:52] |
| pinchartl | yes
and drivers shouldn't have multiple pads anyway that's a downstream hack the upstream driver isn't meant to be compatible with downstream drivers | [06:53] |
| jm_h | AFAICT the imx219 driver used to test your branch has multiple pads or did I miss something (which is highly likely -_-) ?
Oh, when you say mutiple pads, you mean multiple source pads ? | [07:01] |
| *** | frieder has joined #linux-media | [07:04] |
| pinchartl | yes. internal pads are fine | [07:05] |
| *** | GBenji has joined #linux-media | [07:15] |
| ao2 has joined #linux-media | [07:24] | |
| mripard_ has joined #linux-media | [07:37] | |
| jm_h | pinchartl, thanks, removing the second source pad makes it work fine ;-) | [07:41] |
| *** | mripard has quit IRC (Ping timeout: 480 seconds) | [07:42] |
| .... (idle for 18mn) | ||
| mripard has joined #linux-media
tmerciai has quit IRC (Remote host closed the connection) mripard_ has quit IRC (Ping timeout: 480 seconds) sigmaris_ has joined #linux-media sigmaris has quit IRC (Ping timeout: 480 seconds) | [08:00] | |
| tmerciai has joined #linux-media
sigmaris has joined #linux-media | [08:10] | |
| sigmaris_ has quit IRC (Read error: Connection reset by peer) | [08:15] | |
| ........ (idle for 39mn) | ||
| hansg has joined #linux-media | [08:54] | |
| ............ (idle for 57mn) | ||
| jm_h | sailus, I have a camera pi 3 and the imx708 is working nicely with upstream unicam. *But* there is a dw9817 VCM on it (AFAIK) and I have I2C issues when enabling it. I can't find the datasheet but basically the issue is: https://paste.debian.net/1316258/
Sounds like an issue in writing to the STATUS register, but I can't really see why. Any idea/pointers ? | [09:51] |
| .... (idle for 17mn) | ||
| sailus | jm_h: Does the VCM share regulators or GPIOs with the sensor? | [10:08] |
| jm_h | sailus: yes, the regulator is the same | [10:14] |
| ..... (idle for 21mn) | ||
| kbingham | The fact that RPi cameras and the VCM share the regulator seems to be a pain ;-( | [10:35] |
| sailus | I had patches to add modules to DT, with the power sequence defined for the module, but they need some work. Neither there is an agreement to go this way. There are so many modules (so a large number of drivers would likely be needed) and even the module isn't always even known. | [10:43] |
| ....... (idle for 34mn) | ||
| jm_h | kbingham: sailus: an idea to easily solve this in a short notice ? | [11:17] |
| kbingham | jm_h, You said you can't write to the status register. Is the camera streaming at that time? The RPi drivers tie in the VCM power events to the camera lifetime. If you're trying to talk to the VCM when the camera isn't available it might simply be powered down.
so in that pastebin, I bet the VCM only gets powered on /after the last line in your log. jm_h, Trace through your (rpi's?) dw9807 driver and you'll probably see something like an event that only handles power on when the regulator is enabled - by the camera ... rather than by the vcm driver. | [11:29] |
| .... (idle for 15mn) | ||
| *** | lexano has joined #linux-media | [11:46] |
| jm_h | The camera is not streaming at that time, I will trace and try to get the full event log in mind thx | [11:51] |
| .......... (idle for 47mn) | ||
| *** | hansg has quit IRC (Quit: Leaving) | [12:38] |
| ..... (idle for 22mn) | ||
| eelstrebor has joined #linux-media | [13:00] | |
| GBenji has quit IRC (Ping timeout: 480 seconds) | [13:11] | |
| mvaittin has quit IRC (Ping timeout: 480 seconds)
GBenji has joined #linux-media | [13:16] | |
| sailus | I guess adding device links for runtime PM could go quite far solving this albeit this looks like a hack more or less, how much exactly is up to discussion.
Sometimes the firmware people see the VCM as part of the sensor and the VCM has no use unless the sensor is used, too. | [13:30] |
| *** | tyalie has quit IRC ()
tyalie has joined #linux-media | [13:38] |
| ...... (idle for 26mn) | ||
| mvaittin has joined #linux-media | [14:05] | |
| kbingham | Part of the problem RPi are trying to solve there is that when the camera is powered up - the VCM /will/ be powered up, and it has to be configured correctly or it will be potentially holding voltage in the wrong position, or trying to move the lens to an invalid position. So they tie the power up and configuration directly to the camera rather than the VCM... It's definitely more complex than just "Oh it's a separate device so it
should just be separate". Perhaps with device links we could say the VCM depends on the sensor ... It certainly doesn't have much use in RPi case unless the sensor is enabled ... I'm not sure a device link covers all aspects though ... And it would certainly be 'module' specific - not sensor specific! | [14:09] |
| *** | BrianG61UK has quit IRC (Read error: Connection reset by peer) | [14:19] |
| mripard has quit IRC (Remote host closed the connection)
GBenji has left BrianG61UK has joined #linux-media mvaittin has quit IRC (Ping timeout: 480 seconds) | [14:27] | |
| ........... (idle for 52mn) | ||
| xecutertool has joined #linux-media
hansg has joined #linux-media xecutercmd has quit IRC (Ping timeout: 480 seconds) | [15:26] | |
| jarthur has joined #linux-media | [15:44] | |
| xecutertool has quit IRC (Remote host closed the connection)
xecuter has joined #linux-media | [15:51] | |
| ..... (idle for 23mn) | ||
| frieder has quit IRC (Remote host closed the connection) | [16:16] | |
| ...... (idle for 27mn) | ||
| Whoopie has quit IRC (Quit: ZNC - http://znc.in) | [16:43] | |
| _0andriy_ | hansg: fun fact: Changing MAX(MAX(MAX(MAX( in AtomISP driver to max(max(max(max( (or two max3() calls) kills the compilation: it consumed 128*G*byte+ memory, and I killed it, no difference if it's GCC or clang. | [16:54] |
| hansg | _0andriy_, that is both hilarious and sad at the same time | [16:55] |
| _0andriy_ | hansg: when kernel plays too smart | [16:55] |
| ..... (idle for 20mn) | ||
| *** | charrus_ has joined #linux-media
charrus has quit IRC (Ping timeout: 480 seconds) | [17:15] |
| .......... (idle for 45mn) | ||
| robertfoss has quit IRC (Ping timeout: 480 seconds) | [18:03] | |
| ........... (idle for 50mn) | ||
| robertfoss has joined #linux-media | [18:53] | |
| ........ (idle for 38mn) | ||
| hansg has quit IRC (Quit: Leaving) | [19:31] | |
| ........... (idle for 51mn) | ||
| darkapex has joined #linux-media | [20:22] | |
| .... (idle for 17mn) | ||
| ten1572377432463050 has quit IRC (Remote host closed the connection)
ten1572377432463050 has joined #linux-media | [20:39] | |
| ............. (idle for 1h2mn) | ||
| ao2 has quit IRC (Quit: Leaving) | [21:41] | |
| ..................... (idle for 1h44mn) | ||
| djrscally has quit IRC (Ping timeout: 480 seconds)
ten15723774324630502 has joined #linux-media ten1572377432463050 has quit IRC (Remote host closed the connection) ten15723774324630502 is now known as ten1572377432463050 | [23:25] | |
| bingbu has quit IRC (Read error: Connection reset by peer) | [23:37] | |
| ↑back Search ←Prev date Next date→ Show only urls | (Click on time to select a line by its url) |