<div>Hi All,</div><div><br></div><div>I've been slowly working at reverse engineering the libvecore.a cedarx library based on the resources we have available. </div><div><br></div><div>As a start I've created a working (although really not finished) memory allocator (libcedarxalloc.a) which is used by libvecore.a to allocate and manage memory reserved for use for the cedarx a10 decoder chip.</div>
<div><br></div><div>I've built and tested this against the code available here: <a href="https://github.com/amery/allwinner-a10-video">https://github.com/amery/allwinner-a10-video</a>. My code is available here <a href="https://github.com/iainb/open_cdxalloc">https://github.com/iainb/open_cdxalloc</a> </div>
<div><br></div><div>I'm fairly certain the original code from amery's repo is broken in some way. Even unmodified code causes random kernel panics, probably caused by some kind of memory corruption. The libvecore.a library manages low level access to memory and the video decoder core so its possible there could be problems there - though that's just a guess, at least its very easy to reproduce.</div>
<div><br></div><div>If anyone can provide me with more better resources such as a working libvecore.a (unstripped for development and testing would be nice!) or more documentation then I would be very happy (it doesn't hurt to ask right?).</div>
<div><br></div><div>Thanks,</div><div><br></div><div>Iain.</div><div><br></div><div><br></div><div><br></div><div><br></div><div><br></div><br>