| ↑back Search ←Prev date Next date→ Show only urls | (Click on time to select a line by its url) |
| Who | What | When | |
|---|---|---|---|
| *** | alarumbe has quit IRC (Read error: No route to host)
alarumbe has joined #linux-media | [00:09] | |
| ................................ (idle for 2h38mn) | |||
| wallacer has quit IRC (Ping timeout: 480 seconds)
wallacer has joined #linux-media | [02:48] | ||
| ...... (idle for 25mn) | |||
| cptaffe has joined #linux-media | [03:16] | ||
| ............................. (idle for 2h24mn) | |||
| Mo_ has joined #linux-media | [05:40] | ||
| Mo has quit IRC (Ping timeout: 480 seconds) | [05:48] | ||
| .......................... (idle for 2h6mn) | |||
| jmassot has joined #linux-media
jmassot_ has quit IRC (Ping timeout: 480 seconds) | [07:54] | ||
| ...... (idle for 26mn) | |||
| dcz has joined #linux-media | [08:23] | ||
| dcz | trying to understand how subdevice formats work right now, got a couple questions
in the same place, what's the point of setting `which`? At first I thought that it's because the format is influenced by other pads, but " media bus formats may depend on the current ‘try’ formats at other pads", so it's always dependent on the same values as opposed to if it said " media bus formats may depend on the current ‘__which__’ formats at other pads" | [08:27] | |
| ....... (idle for 31mn) | |||
| "When a try or active format is set on a pad, corresponding formats on other pads of the same sub-device can be modified" - is there a way to enumerate all possibilities then?
there's "Formats should be propagated from sink pads to source pads. Modifying a format on a source pad should not modify the format on any sink pad." but that doesn't make it clear if one source pad can modify another source pad I guess with "should", there's no way to actually find out the capabilities reliably :/ | [09:01] | ||
| .... (idle for 19mn) | |||
| *** | mriesch_ has joined #linux-media
HackerBikepacker_ has joined #linux-media | [09:22] | |
| HackerBikepacker has quit IRC (Ping timeout: 480 seconds)
mriesch has quit IRC (Ping timeout: 480 seconds) | [09:29] | ||
| HackerBikepacker_ has quit IRC (Quit: https://quassel-irc.org - Chat comfortably. Anywhere.)
mriesch_ has quit IRC () mriesch has joined #linux-media HackerBikepacker has joined #linux-media | [09:41] | ||
| .............. (idle for 1h6mn) | |||
| jmassot_ has joined #linux-media
jmassot has quit IRC (Ping timeout: 480 seconds) | [10:47] | ||
| .............. (idle for 1h5mn) | |||
| BrianG61UK has quit IRC (Read error: Connection reset by peer) | [11:55] | ||
| ........................................... (idle for 3h33mn) | |||
| pinchartl | dcz: the stream ID is set by applications, yes. on camera sensor subdevs it will be hardcoded by the kernel though. that's work in progress, patches haven't been merged yet to support streams on source subdevs
as for the "which" argument, think about it as selecting between two separate and completely independent sets of configuration for the whole subdev they both operate in exactly the same way, except for the fact that the ACTIVE configuration is applied to the device and the TRY configuration isn't. the TRY configuration is meant to try configurations without impacting the device operation enumerating formats on a source pad can return different results depending on the formats set on the sink pas and setting a format on a source pad can possibly modify formats on other source pads the details of which formats impact which other formats is device-specific | [15:28] | |
| ............ (idle for 59mn) | |||
| *** | BrianG61UK has joined #linux-media | [16:30] | |
| .... (idle for 15mn) | |||
| dcz | "completely independent sets" - that's what I suspected, thanks. But then I read the docs as contradicting that with " media bus formats may depend on the current ‘try’ formats at other pads" | [16:45] | |
| pinchartl | I'm sure the wording of the documentation can be improved in numerous places | [16:47] | |
| dcz | I guess I'll just pretend that source pads are independent for now. I can't think of an algorithm that can figure out arbitrary dependencies | [16:47] | |
| pinchartl | you'll need device-specific code to solve that problem
because at the end of the day it's a hardware issue there are lots of weird dependencies at the hardware level | [16:47] | |
| dcz | while working on this I was wishing for an interface that just let the driver dump on me the rules that it's following. No idea what that would look like with C ABI, but my thoughts still went in that direction | [16:49] | |
| pinchartl | you can always wish :-)
and then you'll have the next device that has a set of rules that don't match what can be expressed you can't win that fight without hardware standardization, and the industry it not taking that direction so you need device-specific code in userspace the alternative is to move the whole complexity of libcamera pipeline handlers to the kernel, and that won't happen | [16:49] | |
| dcz | if the driver could dump the raw rules, that would solve the problem of unexpected configs. The problem would then be in how complex the rules can be
oh, that's what you wrote. "expressed", not "expected" | [16:51] | |
| ...... (idle for 27mn) | |||
| the video capture device has a sink pad, how do I get its mbus format? v4l2_subdev ioctls complain about wrong device kind
or should I derive it from the fourcc code? | [17:18] | ||
| pinchartl | you can enumerate the pixel formats for a given media bus code using the enum fmt ioctl, assuming the driver implements CAP_IO_MC
there's no 1:1 mapping between media bus codes and pixel formats the mapping is device-dependent | [17:21] | |
| dcz | what "input or output" is meant here? "After switching the input or output the list of enumerated image formats may be different." https://www.kernel.org/doc/html/latest/userspace-api/media/v4l/vidioc-enum-fmt.html#description | [17:35] | |
| *** | cyrinux has quit IRC () | [17:38] | |
| dcz | do I get it right that a capture device is limited to 1 pad and that pad's info is in enum_fmt for the device?
no, that call still doesn't return the current mbus format for that pad | [17:38] | |
| *** | cyrinux has joined #linux-media | [17:40] | |
| dcz | I need to set the formats on both sides of the link, so how do I set the mbus format on the video device pad? Is it through setting a corresponding fourcc? | [17:46] | |
| ........... (idle for 53mn) | |||
| *** | lexano has quit IRC (Remote host closed the connection) | [18:39] | |
| ....... (idle for 33mn) | |||
| lexano has joined #linux-media | [19:12] | ||
| .... (idle for 17mn) | |||
| jmassot has joined #linux-media
jmassot_ has quit IRC (Ping timeout: 480 seconds) | [19:29] | ||
| ......... (idle for 43mn) | |||
| jmassot has quit IRC (Ping timeout: 480 seconds) | [20:14] | ||
| ............ (idle for 56mn) | |||
| jmassot has joined #linux-media
jmassot_ has joined #linux-media | [21:10] | ||
| jmassot has quit IRC (Ping timeout: 480 seconds)
dcz has quit IRC (Quit: Konversation terminated!) | [21:19] | ||
| BrianG61UK has quit IRC (Ping timeout: 480 seconds)
BrianG61UK has joined #linux-media | [21:28] | ||
| BrianG61UK has quit IRC (Read error: Connection reset by peer) | [21:42] | ||
| danitool has joined #linux-media
BrianG61UK has joined #linux-media | [21:54] | ||
| ...... (idle for 25mn) | |||
| BrianG61UK has quit IRC (Ping timeout: 480 seconds)
BrianG61UK has joined #linux-media | [22:19] | ||
| ........ (idle for 35mn) | |||
| BrianG61UK has quit IRC (Read error: Connection reset by peer) | [22:55] | ||
| BrianG61UK has joined #linux-media | [23:00] | ||
| BrianG61UK has quit IRC (Ping timeout: 480 seconds)
BrianG61UK has joined #linux-media ten157237743246305066182150355 has quit IRC (Quit: The Lounge - https://thelounge.chat) ten157237743246305066182150355 has joined #linux-media djrscally has joined #linux-media | [23:12] | ||
| .... (idle for 18mn) | |||
| BrianG61UK has quit IRC (Ping timeout: 480 seconds)
BrianG61UK has joined #linux-media BrianG61UK has quit IRC (Read error: Connection reset by peer) | [23:34] | ||
| ↑back Search ←Prev date Next date→ Show only urls | (Click on time to select a line by its url) |