On Sat, 4 Apr 2009 14:24:27 +0200 Jean Delvare <khali@xxxxxxxxxxxx> wrote: > Hi all, > > Here finally comes my conversion of ir-kbd-i2c to the new i2c binding > model. I've split it into 6 pieces for easier review. Firstly there are > 2 preliminary patches: > > media-video-01-cx18-fix-i2c-error-handling.patch > media-video-02-ir-kbd-i2c-dont-abuse-client-name.patch > > Then 2 patches doing the actual conversion: > > media-video-03-ir-kbd-i2c-convert-to-new-style.patch > media-video-04-configure-ir-receiver.patch > > And lastly 2 patches cleaning up saa7134-input thanks to the new > possibilities offered by the conversion: > > media-video-05-saa7134-input-cleanup-msi-ir.patch > media-video-06-saa7134-input-cleanup-avermedia-cardbus.patch > > This patch set is against the v4l-dvb repository, but I didn't pay > attention to the compatibility issues. I simply build-tested it on > 2.6.27 and 2.6.29. > > This patch set touches many different drivers and I can't test any of > them. My only TV card with an IR receiver doesn't make use of > ir-kbd-i2c. So I would warmly welcome testers. The more testing my > changes can get, the better. > > And of course I welcome reviews and comments as well. I had to touch > many drivers I don't know anything about so it is possible that I > missed something. > > I'll post all 6 patches as replies to this post. They can also be > temporarily downloaded from: > http://jdelvare.pck.nerim.net/linux/ir-kbd-i2c/ > Additionally I've put a combined patch there, to make testing easier: > http://jdelvare.pck.nerim.net/linux/ir-kbd-i2c/ir-kbd-i2c-conversion-ALL-IN-ONE.patch > But for review the individual patches are much better. > > Thanks, >From the discussions we already have, I noticed some points to take care of: 1) about the lirc support, I don't think we should change a kernel driver due to an out-of-tree kernel driver. As I've commented on PATCH 3/6 discussion, we need to better address this with lirc developers; 2) the way Mike is proposing to solve the issue with pvrusb2 will break userspace usage for people that have those devices whose IR work with the in-kernel IR i2c driver. This means that we'll cause a kernel regression due to an out-of-tree driver; 3) So far, nobody gave us any positive return that the new IR model is working with any of the touched drivers. So, more tests are needed. I'm expecting to have a positive reply for each of the touched drivers. People, please test! Since the merge window is almost finished, IMO, we should postpone those changes to 2.6.31, to better address the lirc issue and to give people more time for testing, applying the changesets after the end of the merge window at the v4l/dvb development tree. This will help people to test, review and propose changes if needed. Cheers, Mauro -- To unsubscribe from this list: send the line "unsubscribe linux-media" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
