07:11 <Omegamoon> any suggestions for a good rk3328 device?

07:31 <stdint> Omegamoon, asus tinker?

07:32 <sjoerd> tha'ts a 3288

07:32 <Omegamoon> tinker is a rk3288 device... I'm looking for a rk3328

07:32 <Omegamoon> :)

08:23 <stdint> Omegamoon, sorry, I hear there is some vendor plan to make one

08:23 <stdint> but I don't know whether there is a dev board available

08:47 <wzyy2> buy rk3328-evb board from rockchip ; )

08:50 <phh> wait, that's possible?

08:52 <Omegamoon> @wzyy2: isn't that crazy expensive? :S

08:52 <wzyy2> just joking, it's crazy expensive

09:05 <Tef62> What kind of stuff do you guys do with Rockchip?

14:31 <kanadenipun> unable to clone repo from git.us.linux-rockchip.org

14:32 <kanadenipun> taking too long

15:00 <ayaka> kanadenipun, why do you do that

17:43 <infiniteloop> Hello! I'm looking for any information on setting up framebuffer console on a tablet with a RK3188-T chip. I'm trying to use the kernel from there: https://github.com/rockchip-linux/kernel/tree/release-4.4 If there's a chance it could be achieved, any help would be appreciated.

17:48 <infiniteloop> Additional info: the tablet has an TFT panel with parallel RGB interface, presumably driven directly by LCDC. Unfortunately, I can't get any data on the CS/WR/EN/RST GPIOs for it. Can someone suggest a way GPIO map of those pins can be identified using the existing Android kernel, that would be great news for me.

17:58 <phh> this kernel isn't really meant for 3188 afaik

17:59 <phh> and I think this kernel doesn't have parallel RGB support

17:59 <infiniteloop> Oh

17:59 <infiniteloop> Can you suggest better one?

17:59 <infiniteloop> I mean, the one that would be easier to adapt.

18:00 <phh> well easiest would be to use rockchip's original 3.10 kernel

18:00 <phh> but erm...

18:00 <phh> no wait, it was 3.0.36+

18:00 <infiniteloop> I also had al ook on this one: https://github.com/bbelos/rk3188-kernel

18:01 <phh> yup, 3.0.36

18:01 <infiniteloop> It has a driver for touch and other stuff as well, at leat on the first sight...

18:01 <infiniteloop> *at least

18:02 <phh> mmm it looks like mainline doesn't have any video output support on 3188

18:02 <phh> so yeah, you're most likely stuck with original 3.0.36

18:03 <infiniteloop> I see. Thank you for the help!

18:04 <infiniteloop> One more question: would it be possible to back-port mali/ump and drm drivers from the first kernel I've mentioned to the older one?

18:06 <phh> well, mali kernel-space driver is meant to support many kernel versions, so it's not technically impossible

18:06 <phh> I mean, the gap to bridge is not 4.4-3.0, it's much smaller, perhaps there is no gap, or perhaps 3.10-3.0

18:07 <infiniteloop> Okay, it makes sense now. Thanks!

18:10 <infiniteloop> I'd like to return to my initial question: is there a way to determine GPIO map with the running stock Android kernel made for the device?

18:11 <infiniteloop> Or should I try and dig the kernel binary for some known structures to find it?

18:30 <infiniteloop> Sorry if it's inappropriate question for this channel.

18:37 <Myy> The thing is, there's no guarantee that everything was done using a DTS/DTB file. Some hacky kernel drivers encode everything directly into the code so... Anyway, if you want to retrieve the DTB file, you'll first need to understand how the system boots and how the boot files/images are constructed. If there's a /proc/cmdline, you might want to have a look at it.

18:40 <infiniteloop> I am sure the stock kernel doesn't use the device tree, because it was introduced somewhere between 3.10 and 3.11

18:41 <infiniteloop> And the stock kernel is 3.0.36, at least that's what uname says =)

18:42 <vagrantc> hrm. still no MMC on odroid-c2 with linux v4.11-rc5

18:43 <vagrantc> wonder if it needs to be initialized by u-boot

18:43 <infiniteloop> The kernel command line has only initrd memory offset and a partition map for the mtd NAND driver

18:46 <Myy> If you come during morning (6H~13H GMT), there might more actual rockchip engineers and reverse engineers able to tell you how to get the GPIO code from such kernels. Aren't the sources available somewhere, though ? If you can compile and run a similar/identical kernel on the tablet, then you'll "just" have to look at the code, which might be easier than trying to RE the tablet.

18:47 <Myy> *there might be actual rockchip engineers

18:49 <Myy> That said, I think that before doing anything, you'll need to : 1) Find a way to backup the current working image 2) Find a way to unbrick and restore your tablet it if anything went to happen 3) Test various kernels on it.

18:51 <Myy> Is there an HDMI output on the tablet ? If there is, you might want to try a mainline kernel, plug the tablet on an HDMI screen and see if you can get an actual image using the HDMI output.

18:53 <Myy> Or at least, you can try a mainline kernel and see if you can connect to it with ADB. Getting any other kernel actually boot might be more difficult than getting the touchscreen working, IMHO.

18:55 <Myy> Also, you might be interested by such documents : http://mail.olimex.com/Products/SOM/RK3188/RK3188-SOM/resources/RK3188_SOM_GPIOs.pdf

18:55 <Myy> https://github.com/davy39/rk3188-gpio-dump

18:56 <Myy> http://rockchip.fr/RK3188%20datasheet%20V1.0.pdf

19:04 <infiniteloop> Nice, that GPIO dump script looks promising. Thanks a lot!

19:06 <infiniteloop> I have a stock update image ready, actually, that's what I was going to tinker with at first. I use SD card boot method to try out the kernels, and keep the embedded NAND untouched. I hope that's safe enough.

19:11 <Myy> Well, if you know how to boot any kernel safely, then you could always try mainline kernels then, see if you can at least get an ADB access to the tablet.

19:14 <Myy> I got a script that compiles mainline kernels for RK3288 boards. You might want to add a "make menuconfig" in the script, just before the "make $DTB_FILES zImage modules -j5", to reconfigure the kernel for your board, and add rk3188 boards DTB files in the DTB_FILES list.

19:14 <Myy> https://github.com/Miouyouyou/MyyQi

19:21 <infiniteloop> All right, I'll check this out, thank you =)

20:14 <infiniteloop> Have a nice evening, guys!