On Monday 06 June 2005 15:44, Johannes Stezenbach wrote: > On Sat, Jun 04, 2005 at 09:48:13PM +0200, Peter Beutner wrote: > > i recently stumbled over these two little bugs in the dvb_core code: > > 1. > > As the FE_SET_FRONTEND ioctl is asynchronous it is not guaranteed that > > the tuner is already reconfigured when the next ioctl occurs. In case of > > the FE_READ_STATUS ioctl this would return the signal status of the > > previous,now obsolete tuning to the user. Which is kind of bad for apps > > which do something like: > > tune(); > > while(!has_lock()) > > sleep(500); > > read_data(); > > Prevent this by automatically returning zero if frontend state is > > FE_RETUNE. > > Hm, nice catch. Did you actually manage to hit this race? I've definitely seen this occasionally, but didn't have time to investigate :(
