From: Rick <rbo...@gm...> - 2009-12-22 00:29:04
|
Jarod Wilson wrote: > On Dec 20, 2009, at 12:15 PM, Rick wrote: > >> Jarod Wilson wrote: >>> On Dec 19, 2009, at 4:20 PM, Dale Pontius wrote: >>>> Hopping off-thread for a related question... >>>> >>>> For a long time, lirc has been out-of-kernel. Beyond that, lirc_zilog >>>> (or lirc_pvr150) has been out-of-lirc. I've gotten one explanation of >>>> this, that lirc_zilog has a binary blob of code for the embedded Z8, and >>>> the copyright status of that blob keeps it out of the mainline. >>>> >>>> In and around this time, the kernel portion of nouveau is also getting >>>> merged, and it has (or had) a binary blob. For that matter, lirc_zilog >>>> also works with the hvr-1600, and it has 3 binary blobs. >>>> >>>> The kernel has a firmware loader mechanism to handle this, everything is >>>> already architected, with plenty of examples. Can lirc_zilog be made to >>>> use this mechanism, and get into the mainstream? Obviously this binary >>>> blob is extracted, not provided, so its status is still questionable. >>>> Still, the pure code should be able to be merged, and the binary blob >>>> should be distributable under whatever terms lirc_zilog/lirc_pvr150 is >>>> being distributed now. >>>> >>>> I guess according to kernel protocols, I shouldn't ask such questions, >>>> because the best way is to submit a patch. I'm not terribly qualified, >>>> but I guess if that's what I need to do, I should start looking at code. >>>> In the meantime, if it's a cake-walk for someone else... >>> Don't worry, the lirc_zilog driver is definitely on the list to be submitted for upstream kernel inclusion after lirc_dev gets merged. >> Jarod, >> >> Is there a utility or a command that can scan for two zilogs on two PVR 150's? I'm finding that sometimes after a boot my system register /dev/lirc0 and /dev/lirc1 (good), but other times it registers just one /dev/lirc0 (bad). I have not idea why. > > No clue, sorry. The one and only lirc_zilog device I have is an HD-PVR. The lirc_zilog driver simply looks for i2c parts exposed on devices with the right device name though (assuming 2.6.31 and later). So I'd look at the dmesg output from your PVR-150's, particularly with debugging for its modules cranked up, and see if there are any errors there at initialization time. > Thank you, Jarod. I'll see what I discover. |