06:02 < lja> it is just me or the media_build is broken for kernels <= 4.6
06:02 < lja> due to the i2c_mux_priv func
06:20 < lja> well just checked the "media_tree daily build: ERRORS" so it's not me...
06:34 < crope> i2c-mux is totally rewritten on 4.7
06:34 < crope> and I suspect changes are so big that backporting is not option - those drivers
will be disabled
06:35 < lja> they aren't much but will force everyone to start using 4.7 asap
06:36 < lja> I count 5
06:36 < crope> feel free to backport new i2c-mux to media-build :)
06:37 < lja> :)
06:37 < lja> as soon as more drivers migrate to the new i2c_mux, more will be broken
06:38 < crope> yeah, that is how it works. media-build backport is just backport, if backporting
is too hard then drivers will be disabled
06:38 < lja> anyway, not having the option to backport is really bad for those who use media_tree
on arm boxes (which usually have older kernels)
06:39 < lja> most of those ARM boards won't get a kernel update to 4.7 anytime soon
06:39 < crope> you can look how to backport new i2c-mux. it is not trivial, but it is not mission
impossible
06:39 < lja> I know, I already took a look
06:41 < crope> it is that patch set https://lkml.org/lkml/2016/5/4/769
06:41 < lja> I'm just saying that I feel that a lot of people will get unhappy
06:43 < crope> and that patch set also https://lkml.org/lkml/2016/4/20/467
06:44 < crope> but it makes developers happy as many hacks could be removed
06:45 < lja> true, but developers develop for users :)
06:48 < lja> in fact I just noticed this because I was about to build media_tree for a rpi
06:48 < lja> rpi3
06:52 < crope> yeah. and surely in next weeks we will see many this kind of reports, until someone
backports i2c-mux
tl;dr: <lunixfags> hey lets go changing shit around again, its that time of the month. who cares if it breaks everything
No comments:
Post a Comment