on 2004-02-07 at 12:57, I wrote: > > The rule seems to be: > > If I wait for more than 8 seconds after a successful tuning, the > tuning will fail. > > An immediate re-tuning attempt after a failed or successful tuning > will succeed.
OK, here's a dirty modification which makes the no-lock problem go away. This is from someone who doesn't understand shit about the driver and has never seen any hardware specifications, but was just poking around in the dark... :) I just hope it helps others figure out what the real problem is. In dvb_frontend.c function dvb_frontend_set_parameters(), add an extra call and a delay before the dvb_frontend_internal_ioctl call: dvb_frontend_internal_ioctl (frontend, FE_SET_FRONTEND, param); /* wake up! */ dvb_delay(100); /* clear your head before action */ err = dvb_frontend_internal_ioctl (frontend, FE_SET_FRONTEND, param); Somehow the extra ioctl call wakes the card up and properly prepares it for the real action. I'm kind of simulating the two-tunings-in-a-row-succeed phenomenon I've encountered before. With this ugly fix dvbtune now works correctly, i.e. tunes successfully also after a longer pause than 8 seconds. I haven't experimented too much with the delay in between, I just noticed that if I remove it, I still get some no-locks. -- Info: To unsubscribe send a mail to [EMAIL PROTECTED] with "unsubscribe linux-dvb" as subject.
