On Sat, Jul 16, 2005 at 07:49:12PM +0200, Johannes Stezenbach wrote: > Adam Szalkowski wrote: > > I'm using a Pinnacle PCTV SAT on a DiSEqC 2.0 multiswitch. > > Since 2.6.12 these lines have been added to cx24110_send_diseqc_msg() in > > cx24110.c: > > rv = cx24110_readreg(state, 0x77); > > cx24110_writereg(state, 0x77, rv|0x04); > > > > These lines also occur in cx24110_diseqc_send_burst(). When tuning with szap > > (using the -r option, of course) the FE reports to have a lock but I still do > > not receive any data on some channels. Before 2.6.12 this worked. > > > > When I remove those lines in both functions everything works fine as before > > 2.6.12. What is this code supposed to do? Can it be removed again or fixed so > > that my card works again with a stock kernel? > > According to CVS these lines were added with this change: > > revision 1.22 > date: 2005-03-08 12:42:00 +0000; author: kenneth; state: Exp; lines: +27 -0 > - cx24110 diseqc_send_burst (Peter Hettkamp, Uwe Bugla) > > Maybe one of you guys could comment on this? > > Johannes Comment: - When I modified cx24110.c to work again with the refactored drivers around Kernel 2.6.9, I did not include a diseqc_send_burst function, therefore the ioctl of the same name was not available with this frontend. - This broke the usage of szap, because this program uses the ioctl. - This was brought to my attention, and I created a patch for it. I had one problem when creating the patch: The fact that diseqc_send_burst is a separate ioctl from diseqc_send_message obviously means that an application can choose to call none, either one or both ioctls in any order. The cx24110 chip, however, knows only a "start_diseqc" sequence, which includes voltage selection, tone burst (aka Mini-diseqc), DiSEqC message and continuous 22kHz signal all in one. - So, I decided to try: when send_burst is called, do a start_diseqc sequence /without/ the DiSEqC message (set bit 2 of register 0x77). - when send_message is called, do a start_diseqc sequence /with/ message but without burst (clear bit 2 of 0x77) - I had no way of testing whether this approach was working, so I gave the patch to a few people (those that were complaining over the missing send_burst ioctl in the first place). When I had no negative responses, Uwe reformatted the patch and mailed it in. - for some reason or other, the line cx24110_writereg(state, 0x77, rv&~0x04); in cx24110_send_diseqc_msg seems to have mutated into cx24110_writereg(state, 0x77, rv|0x04); in the process, which suppresses the sending of DiSEqC messages altogether. - The best solution might be to only store the requested burst command on a call to send_burst, and include the correct burst bit in the start_diseqc sequence in send_msg. This would, however, require applications that want to send bursts do do so before calling send_msg, and also require them to call send_msg even if they have no message to send (which may break again since the hardware always sends at least 3 bytes of diseqc message unless the message is suppressed completely via bit 2 of 0x77.) LNBDC and ContinuousTone are easier in this respect, just set the correct bits in 0x76, and all is well, i.e. according to my reseach they seem to take immediate effect with or without a start_diseqc sequence. - If anybody has a better solution for this, feel free to mail me about this. If anyone feels the need to donate DiSEqC equipment with known behaviour (does it switch both on Mini-DiSEqC and DiSEqC commands, or, better, on only one of them), contact me, we can work something out ;) - As an aside, my email address changed, the old t-online address will cease to exist one of those days. Could somebody please update my email address within the source? Best regards, Peter -- "Only wimps use tape backup: _real_ men just upload their important stuff on ftp, and let the rest of the world mirror it ;)" (Linus Torvalds, about his failing hard drive on linux.cs.helsinki.fi) -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 189 bytes Desc: not available Url : http://www.linuxtv.org/pipermail/linux-dvb/attachments/20050717/0b4677d0/attachment.pgp
