Olaf Titz wrote: >>+ state->tx_tuna[8] =~ 0x40; //1 >>+ state->tx_tuna[8] |= 0x40; // 0 >> >> > >I don't know anything about that device, but this code looks wrong to >me. The "1" line sets the whole register and the "0" line a single >bit. Shouldn't that rather be something like > > > >>+ state->tx_tuna[8] &= ~0x40; //1 >> >> > >If the wanted effect is really setting the whole register to ~0x40 vs. >all ones and the other bits are not affected independently, that >should be reformulated just for clarity: > > > >>+ state->tx_tuna[8] = ~0x40; //1 >>+ state->tx_tuna[8] = ~0x00; // 0 >> >> > >(note where the blanks are...) > >Olaf > > >_______________________________________________ > >linux-dvb@xxxxxxxxxxx >http://www.linuxtv.org/cgi-bin/mailman/listinfo/linux-dvb > > (One of these days I'll figure out why Thunderbird is defaulting to reply-to-author rather than the list) Thanks to Olaf and Edgar for pointing this out. Assuming it's correct (again, it works for me), here's a revised patch: Not sure if this still applies (I'm new here :) but just in case: Signed-off-by: Allan Stirling <Dibblahdvb0042@xxxxxxxxxx> Cheers, Allan. --- ./drivers/media/dvb/bt8xx/dst.c.orig 2005-06-21 07:31:05.190721000 +0100 +++ ./drivers/media/dvb/bt8xx/dst.c 2005-06-21 07:31:10.156051000 +0100 @@ -325,12 +325,12 @@ static int dst_set_polarization(struct d switch (state->voltage) { case SEC_VOLTAGE_13: // vertical printk("%s: Polarization=[Vertical]\n", __FUNCTION__); - state->tx_tuna[8] |= 0x40; //1 + state->tx_tuna[8] &= ~0x40; //0 break; case SEC_VOLTAGE_18: // horizontal printk("%s: Polarization=[Horizontal]\n", __FUNCTION__); - state->tx_tuna[8] =~ 0x40; // 0 + state->tx_tuna[8] |= 0x40; //1 break; case SEC_VOLTAGE_OFF:
