14:40 <chewitt> f_ some progress on getting dvb drivers to compile https://github.com/chewitt/linux/commits/dvb-sucks

14:40 <f_> Hi chewitt

14:40 <chewitt> for kicks I'm going to see what happens when I build the front-end and demux

14:40 <f_> >dvb-sucks

14:41 <f_> chewitt: Sure. :)

14:42 <chewitt> i'm starting to get the hunch that the DVB bits of the kernel being shitty and prehistoric (compared to other kernel bits) might actually be in our favour

14:42 <f_> Nice XD

14:43 <f_> I guess...

14:44 <f_> chewitt: Did the front-end and demux compile?

14:46 <chewitt> all the drivers in that branch apart from CXD2837 compile

14:46 <f_> Great!

14:46 <chewitt> whether they work is something entirely different :)

14:46 <f_> Of course! :)

14:47 <chewitt> my level of build/code knowledge is a bit like a kid riding a bike saying "look mum, no hands!"

14:47 <f_> It could be compiling successfully but then cause a kernel panic :)

14:47 <f_> s/panic/pain/

14:47 <chewitt> the modules all load .. but there's no hardware probing so we didn't hit panics yet

14:47 <f_> YET

14:49 <f_> "look mum, I'm falling!"

14:49 <chewitt> in my part of the world the locals use the phrase insh'allah a lot :)

14:50 <f_> https://forum.libreelec.tv/core/images/avatars/62/1-627c88a26a9f1d3bd50e67ea5924065e67834864.png

14:52 <f_> chewitt: Are the drivers taken from the downstream kernel?

14:53 <chewitt> from the 3.14 vendor kernel that LE used to use

14:53 <f_> Eww that's old.

14:53 <chewitt> fortunately that corner of the media-tree is a bit of a graveyard

14:55 <f_> Isn't there a "newer" kernel (amlogic-ng) for Amlogic TV boxes?

14:56 <f_> At least that's what CoreELEC uses.

14:57 <f_> Although it's based on linux v4.x .... still old.

15:06 <f_> chewitt: Also. Did you send the audio and BT patches (with Tested-by)?

15:12 <chewitt> https://patchwork.kernel.org/project/linux-amlogic/list/?series=720753

15:16 <f_> chewitt: Nice!

15:20 <f_> I see you Cc'd me.

15:20 <f_> Nice.

15:22 <f_> chewitt: Want me to test your DVB changes?

15:22 <chewitt> nothing to test really

15:22 <f_> Or help in any way?

15:22 <chewitt> it's just a bunch of modules

15:23 <chewitt> i'm looking at the demux and front-end drivers here https://github.com/chewitt/dvb-wetek/

15:24 <chewitt> the tuner/demod drivers don't really have dependencies on vendor code

15:24 <chewitt> but the demux and front-end look to be referencing some proprietary bits

15:25 <chewitt> I suspect all the regs data will need translating

15:25 <f_> Eww.

15:25 <f_> Which proprietary bits?

15:25 <chewitt> some of the headers imported are from the bsp codebase

15:26 <chewitt> some will have upstream equivalents these days

15:28 <chewitt> in some cases I can simply omit things and clean up the code a bit, e.g. where they are checking Amlogic cpu types (as we can ignore the 32-bit socs)

15:29 <f_> Makes sense.

15:31 <chewitt> but at some point we'll hit unique vendor code, and then i'm stuck

15:33 <f_> You said the tuner/demod drivers don't seem to have dependencies on vendor code, so that shouldn't be too hard for those.

15:45 <chewitt> they're a bit ugly in places, but generally following the template that seems to be used in the upstream drivers

15:46 <chewitt> but the demux and glue that binds everything together looks more entwined

15:48 <f_> wetek_dvb.c: 55: /* #include "linux/amlogic/cpu_version.h" */

15:49 <f_> I guess you commented that, right?

16:54 <chewitt> Yes, that's me

16:55 <chewitt> in the original code there's lots of ifdef based on 32/64 bit and which SoC is present

17:08 <f_> chewitt: Sure.

20:28 <xdarklight> narmstrong: thanks for starting the HHI/sysctrl conversion to dt-bindings!

21:07 <narmstrong> xdarklight: I kept the harder for the end !

21:17 <xdarklight> narmstrong: once we have it ACKed by the dt maintainers I'll start with the first patches for upstream video output support on the 32-bit SoCs. first on the list is a dedicated driver for the CVBS DAC (PHY) which will also be used on the 64-bit SoCs :-)