| T42 | <b100dian> mal: I'm available if you have any insights on what's happening | 12:48 |
|---|---|---|
| mal | @b100dian the crash I see is probably because of they way you have the new code | 13:30 |
| mal | you are calling buffer->m_queue->releaseBufferResources(buffer); in droid_media_buffer_destroy and it does decStrong which probably leads to buffer being destroyed but you are calling that from within the buffer object so I think that causes the crash, afaik you can't call decStrong like that from within buffer | 13:32 |
| mal | I wonder about your comment in the code "instead of clear-ing the entire memory which leads to droid_media_buffer_destroy crashing." maybe that is related somehow | 13:34 |
| T42 | <b100dian> But droid_media_buffer_destroy was already calling decStrong on the buffer before. That comment refers to calling clear on this guy https://github.com/sailfishos/droidmedia/blob/master/droidmediabuffer.h#L32 which has its pointer passed to clients (e.g. the pointer received by the destroy) | 14:34 |
| T42 | <b100dian> So basically the reordering was because crashes in accesing buffer-> after clearing it because `m_cb.buffers_released(m_data)` calls back into the client | 14:36 |
| T42 | <Mister_Magister> @b100dian u doing some progress in video topic? | 14:55 |
| T42 | <b100dian> No, just the patch that worked for the Zenfone and doesn't work for other bases, trying to get it inside droidmedia | 14:56 |
| T42 | <Mister_Magister> huh | 15:01 |
| mal | @b100dian calling decStrong in droid_media_buffer_destroy is not the problem, the problem is that you are calling it now from within buffer object i.e. buffer->m_queue->releaseBufferResources(buffer); if you change the code to be so that in droid_media_buffer_destroy you somehow check whether decStrong should be called, also curious why you now need to call buffer->m_buffer.clear(); which was not | 15:50 |
| mal | called before | 15:51 |
| mal | my checking whether decStrong should be called means that you maybe add a new function which checks if the buffer is unslotted | 15:53 |
| *** phlixi is now known as Guest9305 | 18:32 | |
| *** phlixi_ is now known as phlixi | 18:32 | |
| T42 | <b100dian> I will be able to test more on a 17.1 based port on Sunday, I will have time to reproduce and think then.. Thanks! | 18:34 |
| piggz_ | why the fsck is 01.org/ofono redirecting to some crap intel site now! | 19:51 |
| T42 | <TheVancedGamer> piggz_: had this issue too, apparently they merged stuff to GitHub.com/Intel now | 20:02 |
| T42 | <adampigg> ofono is also in the kernel git :) | 20:03 |
| T42 | <adampigg> ofono is not in that project (re @TheVancedGamer: piggz_: had this iss...) | 20:04 |
| *** Mister_Magister_ is now known as Mister_Magister | 20:25 | |
Generated by irclog2html.py 2.17.1 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!