| ↑back Search ←Prev date Next date→ Show only urls | (Click on time to select a line by its url) |
| Who | What | When |
|---|---|---|
| NyB | Hi, has anyone tried rmmod'ing their saa7134_alsa module on recent kernels? I got this on my Mageia 3.18.3 kernel: https://bugs.mageia.org/show_bug.cgi?id=15111 and I am wondering if it is distribution-specific... | [10:41] |
| ..... (idle for 21mn) | ||
| *** | Juan changes his alias to Guest73776 | [11:02] |
| ................. (idle for 1h23mn) | ||
| stormer' changes his alias to stormer | [12:25] | |
| ........... (idle for 53mn) | ||
| pinchartl | hverkuil: are you still fine with making the ctrl type init ops initialize the whole control array instead of a single element ? | [13:18] |
| hverkuil | pinchartl: can you refresh my memory? | [13:20] |
| pinchartl | I have this 3x3 matrix control
with a default value that isn't constant across the whole matrix I thus need to define a custom control type in my driver to override the init operation init is called in a loop over the matrix elements which is pretty inefficient in most cases making the init operation initialize the whole control would simplify implementations | [13:20] |
| hverkuil | It's coming back to me. | [13:24] |
| pinchartl | and ? :-) | [13:25] |
| hverkuil | The main issue was that in user_to_ptr() the init op is called for the 'tail' of the array, but that is old code (since c->size < total array size isn't allowed) so the init call there can be removed.
Leaving only two inits when the control is created, and changing those to init the whole array is fine with me. | [13:26] |
| pinchartl | ok, I'll prepare patches
thanks | [13:27] |
| .......... (idle for 45mn) | ||
| jmleo | hverkuil: ping | [14:12] |
| hverkuil | jmleo: pong | [14:14] |
| jmleo | I have discussed a bit with pinchartl about MC in my i.MX6 case
and right now, I have something like that : http://snag.gy/dI0SD.jpg So, basically, each IPU could be a media entity but, the FPGA I have would be "shared" by both entities | [14:15] |
| so, it seems to be "necessary" to have something like a "multi-master" which would handle this kind of cases
(I hope I explain it correctly... :/) so, the first IPU, when probing, would register to this master if he is first, he creates the media_device if not, it registers to the existing media device | [14:23] | |
| hverkuil | Too be honest, I have no time to dig into this. Sorry. | [14:24] |
| pinchartl | hverkuil: would you have time to discuss that face to face during fosdem ? | [14:25] |
| hverkuil | I think so, certainly the hour before and after my presentations.
I probably have time next week (Monday?) as well to discuss this on irc, just not today. | [14:26] |
| ↑back Search ←Prev date Next date→ Show only urls | (Click on time to select a line by its url) |
