On Mon, 23 Jun 2003, Holger Waechtler wrote: > Holger Kiehl wrote: > > On Tue, 17 Jun 2003, Holger Waechtler wrote: > > > > > >>Holger Kiehl wrote: > >> > >>>If I poweroff the machine and start it again I no longer get this error. > >>>With a soft reboot I always get 'No such device' error. Someone else > >>>also reported the same behaviour. > >> > >>good, please let us know when you can find out more what register > >>setting exactly causes this problem. > >> > > > > I will need some help here since the three drivers just differ to much. > > > > > >>>In the meantime I have also tried the driver from the KNC1, which is > >>>is a branch from the linuxtv driver and they call it DVB-ANDI. This does > >>>not have the problem with soft reboots and I do manage to receive some > >>>multicast packages, but I loose a lot. This time there are lots of > >>>the following messages in /var/log/messages: > >> > >>Can you please analyze the register settings of this driver a bit more > >>in detail if this one is working reliably? > >> > > > > Attached you will find the register settings for all three drivers, > > DVB-ANDI, Metzler and linuxtv. Please give me some hints which registers > > I should change or focus on. What I already tried before I put in the > > printk() messages, to set init_tab of the linuxtv the same as DVB-ANDI, > > but there was no change. > > (well, that's exactly the challenge: find out which setting exactly > causes the problem - ;) > > Hmm, I'd start searching at the PLL bytes (the 4- or 5-byte-string > written in pll_write() or tuner.c or whereever - ), the tuner/syntesizer > is usually pretty sensitive for misconfiguration. > > Maybe that's exactly the problem: your log shows > stv0299.c: setup for tuner BSRU6, TDQB-S00x > DVB: registering frontend 0:0 (STV0299/TSA5059/SL1935 based)... > > but probe_tuner() should dump something like > stv0299.c: setup for tuner SU1278/SH > DVB: registering frontend 0:0 (STV0299/TSA5059/SL1935 based)... > > for the KNC-1 (you have a SU1278SH soldered on your board, right?). > Correct.
> Maybe just the probing code is broken -- you can test this by hardcoding > the return value of probe_tuner() to PHILIPS_SU1278SH. If so we'll > probably have to check the name string to detect the tuner. > It's now hardcoded, but szap does not get a lock and kdvb is eating a lot of cpu time. Between, i2c->xfer() returns a -121. Holger (Kiehl) -- Info: To unsubscribe send a mail to [EMAIL PROTECTED] with "unsubscribe linux-dvb" as subject.
