01:00 <pnill> "he mainboard LVDS output need to turn off NOSSC(Spread Spectrum function) and turn on the FREERUN function. The mainboard does not turn off the NOSSC, the frequency is unstable and the screen will flicker." so I probably won't get an answer as I'd want to do this on a BSP kernel... but does anyone know if the options mentioned in the quote there are available in the LCD driver provided with the BSP kernel?

01:00 <pnill> I tried looking around and I found nossc=1 is available in the intel integrated graphics driver but I don't see anyone else call it that or really understand what it is tbh.

05:58 <gamiee> pnill: only way is to dig in into the source codes of BSP

06:00 <pnill> Yeah, I attempted and failed lol... I don't know anything about the settings they're speaking of.

06:00 <pnill> "NOSSC" / "FREERUN" as nothing in BSP or mainline for the drm driver stuff seems to have anything of that sort.

06:00 <pnill> I just know it's clock related but haven't researched much further than that.

06:35 <libv> spread spectrum is tiny semi-random variations of the display clock

06:35 <libv> this is to reduce electrostatic emissions and to stop it from interfering with other devices

06:36 <libv> i spent 3d hunting that bit on the radeonhd project

06:36 <libv> back in late 2007

06:37 <libv> and as usual, mr bridgman (the ati guy who should've been providing us with information - hah) was no help, and the fork just took the change, removed author info, and went with it without blinking

06:52 <libv> spread spectrum is of course more than that, but this is what it does in the display driver or in many modern clock driver circuitry, basically a means to keep the FCC happy

07:32 <hlauer> Just wondering, if that could be related to i2s failing on A20 without the following patch: https://bitbucket.org/hlauer/kernel/src/master/patches/ccu-pll2-no-tuning.patchl

07:34 <hlauer> At one point in the past the i2s stopped working and at the end if figured out the patch set the pll2 registers again as before, but I never understand that

07:35 <hlauer> s/if/I

07:43 <libv> hlauer: according to wens patch, it allows one of the pll modifiers to be fractional

07:43 <libv> hlauer: instead of it being an integer, it becomes an int like float

07:44 <libv> which could indeed throw off some things if the clock becomes unstable, and audio is very very sensitive to that

07:44 <libv> but good catch, i am sure that took quite some bisecting/debugging

07:46 <libv> so it could be that the audio base pll on a10/a20 does not support it, and then the whole pll divider/multiplier calculation becomes false

07:46 <libv> so your patch can be reduced to just removing that flag for this clock

07:47 <hlauer> Found it from the hardware side: Just read the hardware registers and noticed the difference.

07:47 <libv> nice catch

07:47 <libv> and register dumpers ftw

07:48 <libv> oh, and extra thanks for bringing it up now

07:49 <hlauer> Thanks. What means "register dumpers ftw" ?

07:49 <libv> as i had an i2s issue myself when i tried to verify the fosdem video hw close to a year ago

07:49 <libv> the audio was strangle garbled, but basically working

07:50 <libv> a register dumper is a simple tool that allows one to read from like /dev/mem and dump regname, value

07:51 <libv> ftw is for the win

07:51 <hlauer> Mine based on hifiberry didn't work at all any more - just my private setup

07:54 <hlauer> Doing that from python. Code may be inside https://bitbucket.org/hlauer/bananaio/src/master/i2s.py. I have debugged that back in 2017

07:55 <libv> hlauer: i'm sticking that straight in my fosdem kernel now so i do not forget about it

07:58 <hlauer> Thanks a lot. Will that hit vanilla later?

08:13 <libv> i catch myself not spending time on upstreaming, instead spending time on running towards fosdems goals instead, so while this is a goal, it's not the most immediate

08:13 <libv> hlauer: do send it to the ml though

08:13 <libv> reduce it to just the flag, add an explanation of what it is, and send it to the ml

08:14 <libv> that's the fastest way to mainline

08:23 <hlauer> Ok, fair. Is the fosdem kernel code available somewhere? I'm bound to admin stuff at the moment, so will look into it when time permits

08:23 <libv> https://github.com/libv/fosdem-video-linux

08:31 <libv> we need i2s there to talk to grab the audio from the adv7611 hdmi capture chip

08:40 <libv> ah, heh, that patch is not as simple as i assumed, i seemed to have overlooked the first part...

08:41 <libv> heh, will keep this in mind for when i get back to fosdem stuff regardless, as i do not have the capacity to test this code now

09:28 <hlauer> I'll be glad to test anything you provide on my BananaPro setup, as I'm routinely compiling the rc kernels - it will just take a few days for feedback

21:58 <tuxd3v> hello Guys I have a Olimex OlinuxinoLime2( A20 ), and I was thinking in using it as my personal ntp server

21:58 <tuxd3v> does any one knows if it has a good RTC oscilator?

22:00 <tuxd3v> seems that I found at least one :)

22:01 <tuxd3v> 32 768 Hz (RTC) quartz crystal Q4 is found at pins F1 and F2 of the A20 microcontroller.

23:01 <tuxd3v> hello,

23:02 <tuxd3v> does we in Olinuxino( A20 ) have to create a deamon for RTC?

23:02 <tuxd3v> https://linux-sunxi.org/RTC

23:02 <tuxd3v> ^^