06:18 <rperier> hi all

06:19 <rperier> Arf.... David did not apply the right patch to his repo o.O (v6 instead of v7)

09:34 <naobsd> trying USB on rk3288... root hub found, on-board uhub found, but no device detected under on-board hub...

09:34 <naobsd> all 3 ports are under on-board hub...

09:34 <naobsd> guess: power?

09:35 <naobsd> correct

09:35 <naobsd> where is uhub power...

09:36 <naobsd> brute force /sys/class/gpio/*?

09:37 <naobsd> (I tried some unknown gpio config in original dts)

10:15 <naobsd> rk3288 smp patch, great

10:15 <naobsd> "rk3288 is dual-core CPU Soc" oh ;)

10:24 <naobsd> ah, I understand vbus gpio control... now I can see device under hub

10:34 <rperier> naobsd: how did you get a rk3288 board ? throught your work or you just asked to contact@rockchip ?

10:37 <naobsd> rperier: I bought consumer STB named CS819II from online shop https://plus.google.com/u/0/105214450751618183430/posts/FRkAKw7amWM

10:38 <naobsd> mmind00: currently no way to toggle vbus gpio, right? output-high in pinctrl seems not work :(

10:39 <naobsd> I confirmed comment out gpio/pinctrl lines in dts, then export /sys/class/gpio/gpioX and set out/1, then I can see devices under hub

10:41 <naobsd> rperier: HTC-T031-V0.2 seems almost same as Rockchip's reference (RBOX), firefly's RBOX SDK can be used

10:41 <naobsd> rperier: I modified very few lines

10:42 <mmind00> naobsd: which gpio is it? gpio0-14?

10:43 <naobsd> mmind00: gpio0 14 for host, gpio0 12 for otg

10:43 <naobsd> vcc_host: vcc-host-regulator { compatible = "regulator-fixed"; gpio = <&gpio0 14 0>; pinctrl-names = "default"; pinctrl-0 = <&host_vbus_drv>; regulator-name = "vcc_host"; regulator-min-microvolt = <5000000>; regulator-max-microvolt = <5000000>; regulator-always-on; regulator-b

10:43 <naobsd> oops

10:43 <naobsd> vcc_host: vcc-host-regulator {

10:44 <naobsd> compatible = "regulator-fixed";

10:44 <naobsd> gpio = <&gpio0 14 0>;

10:44 <naobsd> pinctrl-names = "default";

10:44 <naobsd> pinctrl-0 = <&host_vbus_drv>;

10:44 <naobsd> (snip)

10:44 <naobsd> regulator-always-on;

10:44 <naobsd> regulator-boot-on;

10:44 <naobsd> };

10:44 <naobsd> and

10:44 <naobsd> pcfg_output_high: pcfg-output-high {

10:44 <naobsd> output-high;

10:44 <naobsd> (snip)

10:45 <naobsd> host_vbus_drv: host-vbus-drv {

10:45 <naobsd> rockchip,pins = <0 14 RK_FUNC_GPIO &pcfg_output_high>;

10:45 <naobsd> this is current dts, not working :(

10:45 <naobsd> I'm sure "echo 1 > /sys/class/gpio/gpio14/value" work (w/o above lines in dts)

10:46 <mmind00> it should be pcfg_pull none ... as the regulator-always-on and regulator-boot-on should already make sure that the regulator is active

10:46 <mmind00> and your vcc_host needs a "enable-active-high;" property

10:46 <naobsd> ah

10:46 <naobsd> I see

10:47 <naobsd> in RK's original dts for android

10:47 <naobsd> host_drv_gpio = <&gpio0 GPIO_B6 GPIO_ACTIVE_LOW>;

10:48 <naobsd> are defined, so I removed enable-active-high; :(

10:49 <naobsd> I should read docs, sorry ;)

10:49 <mmind00> :-D

10:51 <naobsd> I don't understand pcfg_pull_* yet :(

10:51 <naobsd> probably none should be used for output

10:51 <naobsd> up/down depends active_low/high input

10:54 <naobsd> works now

10:54 <naobsd> mmind00: really thanks!

10:54 <naobsd> I'll send some dts to ml

10:55 <naobsd> and rk30/31 and rk3288 defconfig updates to github :)

10:56 <naobsd> ah otg port cannot be used as host port yet, right?

10:57 <naobsd> (it's ok for now)

11:09 <naobsd> btw, on 3.10 kernel, I tried "openssl speed -multi 4" and some few benchmark

11:09 <naobsd> I saw 93 degree C from tsadc ;)

11:09 <naobsd> (with heatsink)

11:10 <naobsd> I guess, 4core full load + GPU load will go 100- degree C ;)

11:23 <rperier> naobsd: looking thanks

11:27 <naobsd> rperier: I bought it from dealsprime. that shop says it's "in stock" but actually it wasn't. I needed to wait 1.5month to ship

11:28 <rperier> in the worth case I can wait for the firefly dev board

11:29 <naobsd> any RK3288 box was not ready, of course :)

11:29 <naobsd> yes, I recommend firefly and rock2

11:30 <rperier> firefly would be an open hardware, so firefly I think

11:30 <mmind00> rperier: I think that would be the best option ... you don't get schematics/docs for these android boxes and while all use somehow similar layouts, the small differences are hard to get without schematics

11:31 <rperier> I agree

11:31 <rperier> that's why I don't want to buy rk3288 box or tablets

11:31 <naobsd> I agree too

11:32 <naobsd> this is first and last buy of non-dev rk3288 product ;)

11:33 <naobsd> I just wanted to know what I can do w/o information ;)

11:36 <naobsd> or should I try to contact Hotack to get information? :)

11:36 <naobsd> I like their boards (HTC-Txxx)

11:39 <rperier> mmind00: are there critical tasks for RK3188 SoCs or I can work on adding shutdown support to act8865 (including the mail to devicetree ML about unifying the property) ?

11:43 <naobsd> on my rk3288 board, there are 3 LEDs, all of them are same "blue" color, and not blinking.

11:43 <naobsd> just a shiny light...

11:43 <mmind00> rperier: not sure, there isn't much hardware left :-D ... a big task might be the rk1000 (av-out, audio) ... taking the rockchip driver and cleaning it up so that it could go mainline

11:44 <mmind00> otherwise the shutdown issue, is a bit easier to handle I guess [less code, more politics so to speak ;-) ]

11:47 <rperier> I am not convinced about rk1000... well for now I can work the shutdown issue. and then, what about the usb driver ? does it work fine ?

11:47 <naobsd> I bought RGB->VGA converter for Radxa Rock, I want to try it :)

11:47 <naobsd> probably it's same as firefly's VGA

12:03 <mmind00> rperier: basic usb itself works fine, as the usb-phys seem to be in a sane state on bringup, but nevertheless correct usb-phy-drivers would be nice [only warning: it looks like the register layout of the usb-phys in the GRF is somehow different on each soc]

12:12 <rperier> which part would need to be fixed exactly ? (you said "correct the usb-phy-drivers")

12:19 <mmind00> no "correct usb-phy drivers" ... drivers/phy/*

12:21 <mmind00> i.e. there currently are none for the Rockchip phys ... it works without for the moment, but for example the phy driver should handle the regulator supplies when it gets turned on/off ... not how we enable them all the time currently

12:23 <mmind00> in my devel/workbench branch is a stub sitting (drivers/phy/phy-rk3066-usb.c) that at least contains the register map for rk3066 and rk3188

12:24 <rperier> ohhh, it would be an interesting contribution indeed (after shutdown in pmic)

12:25 <rperier> I add it to my todolist :)

16:53 <Raen> Hi guys

16:53 <Raen> mind if I ask you some questions?)

16:57 <karlp> always best to just ask

17:01 <Raen> sry, first time here, don't know your ground rules.. So here is the question: I'm trying to compile 3.0.36 kernel for rk3066 tablet with rk610 chip. Everything's fine, kernel boots, dmesg good, backlight and panels lights up, but everything i get is just black screen.

17:02 <Raen> Lo there is some problem in feeding rk610 with data from lcdc, i suppose

17:05 <Raen> So what can I do with this? I've seen that rk610 driver has some changes between different sources, maybe it's broken in latest rk3188 4.4.2 box?

17:26 <naobsd> primary LCDC is selected properly? (I don't know about RK610)

17:31 <Raen> afaik yes. It's defined in board.c, right? But here's some strange finding - I've have working self-compiled 3.0.8 kernel, where lcdc0 set as primary, but in rk610 chineese manual (judging only by pictures) rk610 is using lcdc0 for hdmi output, and 1 for lvds

17:32 <Raen> and strange fact 2 - even though it has lvds output the panel selected in 8 kernel is rgb type

17:33 <Raen> so can it be, that easipix sourcel are heavily modified in some wierd way, where lcdc are switched and panel selection are hardcoded somewhere?

17:34 <naobsd> no idea, sorry

17:34 <naobsd> which 3.0.36 kernel are you using?

17:36 <Raen> well, i've tried rk3188 4.4.2 basic, modified from your git repo (stable branch) and old ones from rk3188 (4.2I suppose). bq curie are hard to use, since they are too different, so i decided not to adapt to them at this point

17:39 <Raen> in theory - if i pull everything from 3.0.8 regarding video drivers (except rk_fb) - will it work (I mean kernel in general, not in my case)? And since I want it for kitkat, that fb changes are made in generic fb, or rk_fb?

17:43 <Raen> Oh, GriefNorth. I've been using your surf1011 repo at the beggining. Do you know anything abour rk610 in rk3066?