On Mon, 2006-11-20 at 13:42 -0800, Trent Piepho wrote: > On Sat, 18 Nov 2006, Rusty Scott wrote: > > The attached patch makes changes to the or51211 frontend to match the SNR and > > signal strength methods in the lgdt330x frontend. It removes an embedded log > > function in favor of using the functions in dvb_math. Do not confuse this patch > > with the patch just posted for the or51132 frontend. I'm not sure why 'make > > I've fixed up the patches a bit and put them into my repository, > http://linuxtv.org/hg/~tap/v4l-dvb Does everything still look ok? > > I've tested the or51132 patch and it seems to work ok with 8-vsb. I > haven't been able to test the or51211 patch. I'm not sure how much testing > it got before, as it didn't compile. > Odd, it compiled and worked fine on the pcHDTV HD2000 card I used to test it. I will grab your repo and test it to make sure but it probably won't be until the weekend. (Thanksgiving and all that.) Do you remember what the compile issue was? > I'd like to verify a couple things. For the or51211, the error register is > only 8 bits? In the or51132 it's 16 bits. > Yes, the or51211 is only an 8 bit value where the or51132 is 16 bits > The or51211 patch removed a msleep(3) that appears between most of the i2c > transactions. Is this ok? > > The or51132 patch removed some msleep(30) calls too. I added those back in > since the rest of the code had them. Is there anything in the datasheet to > suggest that these are necessary? i.e. after selecting a register you must > wait 30 ms before trying to read it? I wonder if all those msleeps() can > be removed. I couldn't find anything in either datasheet suggesting they were necessary. I removed them as a test and when everything continued to work I left them out of the patch. I don't see any reason to leave them in, but that should probably be a separate patch. Rusty _______________________________________________ linux-dvb mailing list linux-dvb@xxxxxxxxxxx http://www.linuxtv.org/cgi-bin/mailman/listinfo/linux-dvb
