Hi, > The problem isn't the sleep itself, the problem is that in the CVS code > (at least), the FE_GET_TUNE_SETTINGS internal IOCTL falls through and puts
Shit, the classical "switch"-problem :(. Thanks for hunting that bug. > The patch also stops the frontend thread from hunting around for a signal > - the MT352 chip has a good capture range and shouldn't really need it. > Wolfram, is there any particular reason you think we want that enabled? No, I just didn't change it. > In any case, the chip can take >0.5s to re-acquire all parameters so the > 50ms min_delay there is a little short... You're right. I can't do many tuning tests because I can only receive one channel here. That's why I left these code-parts/parameters untouched. I think if no one else needs frequency-hunting the patch should be commited to the cvs. Wolfram.