Changes between Version 1 and Version 2 of Ticket #11497, comment 12


Ignore:
Timestamp:
Dec 12, 2014, 8:30:04 PM (10 years ago)
Author:
ttcoder

Legend:

Unmodified
Added
Removed
Modified

  • Ticket #11497, comment 12

    v1 v2  
    2727
    2828EDIT:
    29 reason being, we're attacking the problem from two opposite ends (under heavy load froma CPU-intensive app, and me it's right at the start when the BBuffers start being created at all), and maybe on this end we can find good hints too
     29reason being, we're attacking the problem from two opposite ends (under heavy load froma CPU-intensive app, and me it's right at the start when the BBuffers start being created at all), and maybe on this end we can find good hints too .. What I'm seeing here is that a NULL Data() is correlated to, or  even a precursor (so maybe also the cause?) of the "oops" assert/debugger drop.
    3030
    31 Still, I'd like to follow your lead too so I'm going to study the code some more :-p.. So far for me, who is sending what buffer in what thread to whom is still foggy
     31Still, I'd like to /also/ follow your lead too so I'm going to study the code some more :-p.. So far for me, who is sending what buffer in what thread to whom is still foggy.. I see that BufferCache is only referenced by [http://grok.bikemonkey.org/source/search?q=BufferCache&defs=&refs=&path=&hist=&project=haiku BufferConsumer] (and it also owns a [http://grok.bikemonkey.org/source/xref/haiku/headers/os/media/BufferGroup.h BBufferGroup]).. and there are at least a couple threads based on BufferConsumer in media_addon_server:  [http://grok.bikemonkey.org/source/xref/haiku/src/add-ons/media/media-add-ons/mixer/AudioMixer.cpp#105 AudioMixer], [http://grok.bikemonkey.org/source/xref/haiku/src/add-ons/media/media-add-ons/multi_audio/MultiAudioNode.cpp#171 MultiAudioNode] ..