> >attached is a patch which fixes this. > >It seems to be a timing issue with readaheadthread and the reader and it > >gets stuck in the loop > >since the buffer is too full to continue but the reader isnt reading. > > You can uncomment the section of code that begins with "// HACK > Sometimes the readhead thread gets borked on startup." I hadn't > intended to leave that in there as I thought I'd tracked down the root > cause, I'll do some more digging.
restarting the readaheadthread is not the right way. the problem is that the readahead thread has the buffer full before the reader starts. thus allowing the readaheadthread to set things incl the waitcond even though it can read anymore into the buffer is a good soln IMO. could just delete the commented out HACK now. syncing the readaheadthread start and the reader start would be another way. I also see a burst of AV then a 1sec pause then AV continues when I start most of the time. seems to be a startvation issue due to alloc/start times for various threads in the chain. cheers mark _______________________________________________ mythtv-dev mailing list [email protected] http://mythtv.org/cgi-bin/mailman/listinfo/mythtv-dev
