Greg- You will probably get better results if you re-send this patch as inline text. It must be viewable within the email so that people on the list can make comments on the patch if they so choose. Only after then, will somebody be able to apply it to cvs. frosts1@xxxxxxxxxxxxx wrote: >I posted previously to this mailing list about the problems with the >definition of "struct dvb_pll_desc dvb_pll_lg_z201" in dvb-pll.c, but >got no response. What do I need to do to get this fix included in cvs? >It appears the the DVICO Fusion HDTV DVB-T1 is the only card that uses >this tuner definition, and the definition is wrong, so I can't see any >problem in just fixing it. > >Can someone with cvs update access please patch the file and commit it >to CVS. I have attached the patch. > >Thanks (hopefully) in advance, > >Greg Frost > > > >>There have been a number of people reporting success with the DVICO cards and >>the 2.6.12 kernel. Unfortunately, these are people with the Plus card and I >>have the T1 card. The differences are highlighted on this page: >>http://www.fusionhdtv.co.kr/eng/Products/DVBT.aspx in the table at the bottom. >>The principal difference is the tuner. T1=LG-Z201 Plus=Thomson 7579. >> >>So I went hunting in the kernel code to try to see what was different in the >>tuner area for my patched 2.6.11.7 (using Chris Pascoes patches from >>www.itee.uq.edu.au/~chrisp/Linux-DVB/DVICO/) which works and 2.6.12 which >>doesnt. >> >>This was a lot harder than I thought it would be because the structure of the >>code has been completely changed, but here's what I found: >> >>This is the bit of code that controls the lg-z201 tuner in my patched 2.6.11.7 >>kernel: >>Code: >>static int lg_z201_pll_set(struct dvb_frontend* fe, >> struct dvb_frontend_parameters* params, u8* pllbuf) >>{ >> u32 div; >> unsigned char cp = 0; >> unsigned char bs = 0; >> >> printk ("lg_z201_pll_set %dn", params->frequency); >> >> div = (((params->frequency + 83333) * 3) / 500000) + IF_FREQUENCYx6; >> >> if (params->frequency < 542000000) cp = 0xbc; >> else if (params->frequency < 830000000) cp = 0xf4; >> else cp = 0xfc; >> >> if (params->frequency == 0) bs = 0x03; >> else if (params->frequency < 157500000) bs = 0x01; >> else if (params->frequency < 443250000) bs = 0x02; >> else bs = 0x04; >> >> pllbuf[0] = 0xC2; /* Note: non-linux standard PLL I2C address */ >> pllbuf[1] = div >> 8; >> pllbuf[2] = div & 0xff; >> pllbuf[3] = cp; >> pllbuf[4] = bs; >> >> return 0; >>} >>Notice in particular the cp and bs values and how they vary with frequency. >> >>I then looked in the 2.6.12 code and found the following in >>drivers/media/dvb/fontends/dvb-pll.c: >>Code: >>struct dvb_pll_desc dvb_pll_lg_z201 = { >> .name = "LG z201", >> .min = 174000000, >> .max = 862000000, >> .count = 5, >> .entries = { >> { 0, 36166667, 166666, 0xbc, 0x03 }, >> { 443250000, 36166667, 166666, 0xbc, 0x01 }, >> { 542000000, 36166667, 166666, 0xbc, 0x02 }, >> { 830000000, 36166667, 166666, 0xf4, 0x02 }, >> { 999999999, 36166667, 166666, 0xfc, 0x02 }, >> }, >>}; >>EXPORT_SYMBOL(dvb_pll_lg_z201); >> >> >>So it looks like the same stuff is represented in a table and there must be >>some >>generic bit of code that uses the info in the table to tune the card. It looks >>like the last two fields are the cp and bs values. Looking at some of the >>other >>tuners, I established that the values are to be used for frequencies below the >>frequency in the first field. >> >>Now compare this with the code for the lg_z201 from my working kernel and you >>will see that the numbers don't match. The bs value should only be 0x01 below >>157500000 and 0x02 below 443250000 and then 0x04 for all other frequencies. I >>modified the 2.6.12 (which I have check is the same as cvs at the moment) code >>as shown below: >>Code: >>struct dvb_pll_desc dvb_pll_lg_z201 = { >> .name = "LG z201", >> .min = 174000000, >> .max = 862000000, >> .count = 6, >> .entries = { >> { 0, 36166667, 166666, 0xbc, 0x03 }, >> { 157500000, 36166667, 166666, 0xbc, 0x01 }, >> { 443250000, 36166667, 166666, 0xbc, 0x02 }, >> { 542000000, 36166667, 166666, 0xbc, 0x04 }, >> { 830000000, 36166667, 166666, 0xf4, 0x04 }, >> { 999999999, 36166667, 166666, 0xfc, 0x04 }, >> }, >>}; >>EXPORT_SYMBOL(dvb_pll_lg_z201); >>Rebuilt the kernel and everything worked! >> >>Was this just an oversight from the person who re-structured the code, or is >>there a card using an LG ZX201 that needs to be controlled this way. If it is >>the former, it would be nice to have the change incorporated. >> >> -- Michael Krufky
