00:55 <micken> another stupid RK3288 question:

00:56 <micken> Is it possible to slow down timers? Setting something in CRU to make it like 1Mhz

01:10 <micken> never mind

12:39 <mmind00> micken: looks like you solved your issue already ... normally Rockchip timers are sourced directly from the 24MHz oscillator without any dividers between them

13:40 <lvrp16> is rockchip working on rk3229?

13:41 <lvrp16> for their mainline efforts

13:46 <mmind00> lvrp16: you'll need to ask Rockchip what their plans are ... in general I see patches for all socs from them and they did the rk3229 mainline uboot port, so this soc seems definitly to be on their agenda

13:47 <lvrp16> mmind00: thanks, that puts the shelf life in perspective

13:48 <mmind00> lvrp16: and there are things like https://www.cnx-software.com/2018/03/13/respeaker-core-v2-is-a-6-mic-array-audio-development-kit-powered-by-rockchip-rk3229-processor/

14:49 <grw> no serial output from my rk339 sapphire and nothing on usb-c when holding maskrom button

14:49 <grw> is it dead?

14:55 <micken> mmind00: yes I just used a multiplier in my delay

14:56 <micken> 1 tick / microsecond

17:58 <alyssa> General open-ended question for the community: What remains to be done to get Rockchip hardware, particularly based on the RK3288 and RK3399, in a state where it can be used by the non-technical general public with a mainline, blobless stack?

17:59 <alyssa> My personal experience with a RK3288 chromebook has been that it is generally usable, but outstanding pain points include: convoluted booting procedure (with depthcharge), low distribution support, buggy eMMC and USB drivers, blobby internal wifi, non-mainlined VPU, and blobby GPUs

18:01 <alyssa> I'm curious what other people's perspectives are, since I do believe these devices will be able to play a major role in free software... if they didn't have the "blobs or nerds" dichotomy.

18:05 <alyssa> Admittedly even x86 devices aren't really there, but there's -significantly- more inertia. A person of modest technical skills can figure out how to download an ISO and go through the graphical installer, and once initial setup is done, most people can get the hang of GNOME and such no problem. The same can't be said for the RK devices quite yet, but... I can certainly help code the dream :)

18:07 <mmind00> alyssa: speaking of rk3399 chromeos-devices, booting is the same ... I think you may be able to use uboot in some form [there is uboot support for rk3288-chromeos devices at least], but I try not to touch bootloaders if I don't have to

18:08 <alyssa> "I try not to touch bootloaders if I don't have to" I think most people would agree there :)

18:08 <mmind00> alyssa: emmc is a shiny new arasan sdhci controller, we lost dwc2 and got (hopefully slightly) less buggy dwc3 ... rest is the same I think ... with the wifi attached to pcie

18:09 <alyssa> "less buggy dwc3" Do you know if it plays nice with ath9k_htc devices?

18:10 <mmind00> alyssa: what is nice is that suspend/resume is done via the ATF now [sources available], so we actually get working suspend with mainline

18:10 <alyssa> Ooo, I forgot about suspend being broken on RK3288 since boot up / shutdown is so fast :p

18:12 <alyssa> From what you're saying, RK3399 is a 'healthier' target, I suppose.

18:12 <mmind00> I don't really know about specific device compatiblities on usb though ... my kevin is permanently sitting in my boardfarm

18:13 <mmind00> alyssa: slightly but yes ... i.e. I'm not sure if somebody will spend the energie to get working (deep) suspend on rk3288 (ddr-stuff being in the way) and dwc2 is unfixeable ... so one can only try to lower the pain

18:14 <mmind00> rk3288 does have a light suspend mode, without putting the ddr in self-refresh - so that doesn't save to much

18:14 <alyssa> I don't think suspend is critical, given boot is so fast.

18:17 <alyssa> Hmm, is there work to have a normal free sw distribution (i.e. not CrOS :p) package for depthcharge systems, alleviating the need for cgpt magic and manual kernel installation?

18:33 <mmind00> alyssa: I don't really track who works on what and I replaced my stock coreboot with a netboot-only variant for my uses

18:33 <mmind00> alyssa: but in general, something like https://fosdem.org/2018/schedule/event/hwenablement_simplifying_soc_enablement_in_linux/ might be easiest

18:34 <mmind00> or best

18:36 <mmind00> i.e. make coreboot start into some sort of uefi environment and then the general distribution-installers should work out of the box

18:57 <hanetzer> heh. there's work being done to add grub2 as a payload supported for 'veyron' targets; works, enough to get to a grub shell, but drivers are lacking

18:58 <hanetzer> so, no typing on the chromebook keyboard, and it seems it cannot yet see emmc devices or usb

19:52 <adj_> alyssa, for me, booting with the chromebook with developer mode is painful: a very noisy beep, Ctrl+D or Ctrl+U for booting from internal or USB, and specially the high risk of pressing space+enter and wiping the full internal storage

19:52 <adj_> installing a custom coreboot+payload is not for general public

19:52 <alyssa> adj_: yes, I was alluding to that re boot insanity. Despite being a butterfingers, I have not yet had the space+enter issue :p

19:54 <adj_> I installed Linux in my wife's chromebook (she doesn't like chromeos), but she didn't pay attention to the booting message and pressed space

19:54 <adj_> I needed to reinstall everything again

19:55 <adj_> a friend of mine had the same experience with his chromebook and his little nephew

19:56 <alyssa> adj_: I'm partial to libreboot, of course. If I assume the target will need the boot firmware flashed anyway, we might as well just disable this "feature" in the libreboot builds to begin with.

19:56 <adj_> someone opens the lid, presses the autodestruction button and everythings wipes

19:56 <alyssa> that's concerning, yes

19:56 <adj_> opening the device for flashing the firmware is for technical people, not regular one

19:57 <adj_> other than that, the situation is maybe better than on x86-64

19:58 <alyssa> adj_: sure, but it's something generally done once in the lifetime of the machine, as opposed to the distro installation. i.e. libreboot vendors would be able to take care of that with minimum cost premium

19:58 <adj_> yes

19:59 <adj_> I'm thinking about buying an ASUS C101PA

19:59 <adj_> probably a blob is needed for wifi :(

20:01 <alyssa> I believe so. The FSF takes issue with this -- an issue I plan to discuss with them further -- but a USB wi-fi adaptor is usable instead... provided the USB drivers work ;V

20:12 <micken> I will try to get my hand on a RK3288 chromebook , when I am done with the tinker-board port.