| ↑back Search ←Prev date Next date→ Show only urls | (Click on time to select a line by its url) |
| Who | What | When |
|---|---|---|
| *** | svarbanov has quit IRC (Ping timeout: 480 seconds) | [01:07] |
| ......... (idle for 44mn) | ||
| camus has joined #linux-media | [01:51] | |
| ...... (idle for 25mn) | ||
| camus has quit IRC (Remote host closed the connection)
camus has joined #linux-media | [02:16] | |
| camus1 has joined #linux-media
camus has quit IRC (Ping timeout: 480 seconds) | [02:25] | |
| .................. (idle for 1h25mn) | ||
| NiksDev has joined #linux-media | [03:54] | |
| ............ (idle for 55mn) | ||
| camus has joined #linux-media
camus1 has quit IRC (Remote host closed the connection) | [04:49] | |
| NiksDev has quit IRC (Ping timeout: 480 seconds)
NiksDev has joined #linux-media svarbanov has joined #linux-media | [04:58] | |
| ........... (idle for 52mn) | ||
| jm_h has joined #linux-media | [05:56] | |
| .......... (idle for 47mn) | ||
| hverkuil | pinchartl: I think you are right. Since QBUF returns an error you expect that it isn't queued yet, but in this case it actually still is. | [06:43] |
| *** | GBenji has joined #linux-media | [06:52] |
| hverkuil | pinchartl: I have a patch (untested). Does libcamera work with vivid? You can inject a start streaming error in vivid that would reproduce this issue. | [06:57] |
| ..... (idle for 20mn) | ||
| *** | ao2 has joined #linux-media | [07:17] |
| ........ (idle for 36mn) | ||
| pinchartl | hverkuil: we have out-of-tree vivid support (and in-tree vimc support), so it could be possible to try that, yes | [07:53] |
| hverkuil | I'll mail you the patch. | [07:53] |
| done
note: only compile tested. | [07:58] | |
| pinchartl | thanks | [08:05] |
| *** | ao2 has quit IRC (Remote host closed the connection) | [08:15] |
| ......... (idle for 40mn) | ||
| NoGuest17 has quit IRC (Remote host closed the connection)
NoGuest17 has joined #linux-media | [08:55] | |
| ........ (idle for 36mn) | ||
| kbingham | hverkuil, thanks!
hverkuil, I have a out of tree support for vivid, so I can test with that - and manually set that erorr inject. Haha, it might be harder to actually enable vivid in the ChromeOS kernel though ;-) | [09:32] |
| hverkuil | kbingham: OK, let me know if you need help. The patch *looks* good, but I'm not sure if it really is correct. Also, this is a nice corner case for v4l2-compliance.to test when it detect vivid.
Adding a test to v4l2-compliance would be an alternative for testing this. ...testing this with libcamera. | [09:34] |
| kbingham | hverkuil, indeed, the error case is hard to reproduce for me thoguh as it requires a 'random' event to cause the sensor to fail - so I am keen to test in libcamera with a replicated event to show first that the error is the same, and then it's fixed ;-)
(Of course, then we'll find some arbitrary secondary failure after that ... hehe) | [09:36] |
| hverkuil | Typically applications just give up if qbuf fails (stop streaming altogether), so this hasn't been seen before.
It might give 'unbalanced' messages in the kernel log, though. Never really checked this. kbingham: hmm, testing this error injection with qv4l2 didn't give the result I expected. Let me test a bit more... Ah, qv4l2 will queue the buffers first before calling streamon, so it is streamon that will fail in qv4l2, not qbuf. To test this properly you need libcamera or v4l2-compliance in combination with vivid. Ah, v4l2-compliance already tests this error injection, but only with streamon (i.e. same situation as qv4l2). | [09:37] |
| kbingham | Aha I see. Yes, I guess we differ because we can't give up and close down ;-) and I guess the async buffer event handling is active so it then receives the buffer.... and we didn't expect that. | [09:50] |
| .... (idle for 16mn) | ||
| hverkuil | kbingham: I added a test to v4l2-compliance and with the videobuf2-core.c patch it passes. So it looks good from my side. | [10:06] |
| kbingham | hverkuil, excellent. I've now got a CrOS kernel rebuilt with the patch and vivid built as a module - so I'm ... hopefully not far away from testing here. | [10:06] |
| hverkuil | Hmm, I think I need a similar check in v4l2-compliance when using the request api. | [10:08] |
| There is a bigger problem with the request api as the req_queue callback which in turn calls qbuf assumes that that cannot fail. I need to dig deeper into that. | [10:16] | |
| The combination of min_buffers_needed > 0, Request API, and a failing start_streaming call is a non-trivial corner case. I need to look at that when I have more time. I added this to my TODO list. | [10:28] | |
| ............ (idle for 56mn) | ||
| *** | svarbanov has quit IRC (Ping timeout: 480 seconds) | [11:24] |
| ....... (idle for 32mn) | ||
| NiksDev has quit IRC (Ping timeout: 480 seconds) | [11:56] | |
| ....... (idle for 34mn) | ||
| mriesch has quit IRC (Remote host closed the connection)
NiksDev has joined #linux-media | [12:30] | |
| ...... (idle for 26mn) | ||
| svarbanov has joined #linux-media | [12:59] | |
| ..... (idle for 24mn) | ||
| NiksDev2 has joined #linux-media
NiksDev has quit IRC (Remote host closed the connection) | [13:23] | |
| ....... (idle for 31mn) | ||
| GBenji has left | [13:54] | |
| ................ (idle for 1h16mn) | ||
| camus1 has joined #linux-media
camus has quit IRC (Remote host closed the connection) | [15:10] | |
| .............. (idle for 1h8mn) | ||
| Ares has quit IRC ()
NiksDev2 has quit IRC (Ping timeout: 480 seconds) | [16:18] | |
| ...... (idle for 27mn) | ||
| svarbanov has quit IRC (Ping timeout: 480 seconds)
gouchi has joined #linux-media | [16:47] | |
| svarbanov has joined #linux-media | [16:56] | |
| ......... (idle for 42mn) | ||
| NiksDev has joined #linux-media | [17:38] | |
| ............. (idle for 1h4mn) | ||
| mriesch has joined #linux-media | [18:42] | |
| ................... (idle for 1h33mn) | ||
| jm_h has quit IRC (Remote host closed the connection) | [20:15] | |
| ....... (idle for 32mn) | ||
| gouchi has quit IRC (Remote host closed the connection) | [20:47] | |
| .................. (idle for 1h29mn) | ||
| svarbanov has quit IRC (Ping timeout: 480 seconds) | [22:16] | |
| ........ (idle for 37mn) | ||
| svarbanov has joined #linux-media | [22:53] | |
| ............. (idle for 1h2mn) | ||
| camus has joined #linux-media
camus1 has quit IRC (Ping timeout: 480 seconds) | [23:55] | |
| ↑back Search ←Prev date Next date→ Show only urls | (Click on time to select a line by its url) |