<cristian_c>
apritzel: I've checked as you suggested me
levd1 has joined #linux-rockchip
levd has quit [Ping timeout: 264 seconds]
levd has joined #linux-rockchip
levd1 has quit [Ping timeout: 255 seconds]
levd1 has joined #linux-rockchip
levd has quit [Ping timeout: 244 seconds]
<cristian_c>
apritzel: I've looked at arch/arm in the sources
<cristian_c>
arch/arm/boot
<cristian_c>
apritzel: I've found no .dts either .dtb files
<apritzel>
none at all or none for your board?
<apritzel>
cristian_c: the upstream tree has over 100 .dts files in arch/arm/boot/dts
<cristian_c>
apritzel: there are no 'dts' directories in arch/arm/boot for my sources
<apritzel>
cristian_c: what kernel are we talking about? this is in there for ages
<cristian_c>
apritzel: I've checked also my filesystem for /proc/device-tree and /sys/firmware
<cristian_c>
apritzel: I'm using a 3.0.8+ kernel by crewrktablets
<cristian_c>
'rk_30'
<sjoerd>
they didn't use devicetree back then
<sjoerd>
ye olde kernel
<apritzel>
cristian_c: then you are out of luck, I guess
<cristian_c>
sjoerd: ah, ok
<apritzel>
cristian_c: try to find a newer kernel tree, this old crap will be a pain to work with
<mrutland>
mmind00: as soon as the pinctrl for "vcc-host-regulator" gets configured, I get no more UART output (I put a printk after that gets applied, and that doesn't get output). Looks like the required pin setup is different on my board.
<mrutland>
mmind00: I managed to get to rootfs mounting by removing all nodes requiring pinctrl, but haven't tried anything else yet (I got fed up with waiting for each reflashing and went home)
<sjoerd>
mmind00: As you were looking at memory controllers recently, i've got a puzzle for you.. When loading the fdt or initramfs on 3288 above 512Mb the kernel can't read them properly (I didn't check how it gets corrupted thusfar though).. Any cunning ideas about that by chance?
<mmind00>
sjoerd: :-D I never said I fully understood what happens there ... something about highmem perhaps?
<sjoerd>
mmind00: I don't know
<sjoerd>
u-boot normally relocated fdt (and i guess the ramdisk as well) to the end of memory
<sjoerd>
so it's not a "normal" issue
<lautriv_>
sjoerd, do you pass any mem=<somevalue> to the cmdline ?
<mmind00>
mrutland: that's really strange ... the host-regulator pin (gpio0-a4) is gpio-only, so even has no other function
<sjoerd>
mmind00: I havne't checked but i *think* on my rock2 there are 4 512mb chips so there may be something odd with mapping (though afaik both linux and u-boot use a 1:1 mapping between physcial and virtual for these things..)
levd1 has quit [Remote host closed the connection]
<lautriv_>
sjoerd, in that case you would have issues above 1G since that are 2 banks
levd has joined #linux-rockchip
<mmind00>
mrutland: also in the tronsmart sdk they seem to use that same pin ... and you had serial output with the original android *puzzled*
<mrutland>
mmind00: Interesing; that I cannot explain.
levd1 has joined #linux-rockchip
<mmind00>
mrutland: of course there is simply the possibility that I messed up in the pinctrl driver somewhere
<mmind00>
mrutland: but of course, then I should see the issue as well
<cristian_c>
apritzel: I'd like to find a newer kernel, but I need a kernel containing ds975 board configuration in arch/arm/
<mrutland>
mmind00: have you tried a defconfig next-20150908 ?
levd has quit [Ping timeout: 250 seconds]
<mrutland>
It's also possible that some pinctrl framework code is broken in next and I just happened to hit it :/
<cristian_c>
apritzel: I've found old bq sources, they contain ds975 support. They appear using also 3.0.36 kernel, but I'm not sure
<apritzel>
cristian_c: that's the point: the board configuration is now in device trees, not in the kernel source anymore
<cristian_c>
apritzel: ok, but where?
<cristian_c>
:O
<cristian_c>
apritzel: I've tried to look for in mainline
<cristian_c>
git.kernel.org
<cristian_c>
apritzel: btw, thanks for the help
apritzel has left #linux-rockchip [#linux-rockchip]
Apocx has joined #linux-rockchip
apritzel has joined #linux-rockchip
<apritzel>
cristian_c: is that ds975 an old tablet?
<cristian_c>
apritzel: it's a board
<cristian_c>
for tablets
<apritzel>
cristian_c: looks to me as there is only some hacked, old kernel available which supports it? Mainly to get that one Android working?
<apritzel>
cristian_c: and there were never any upstream efforts?
<cristian_c>
apritzel: anyway, crewrktablets sources work better than aloksinha sources for my device
<cristian_c>
they support most of my hardware
Apocx_ has quit [Ping timeout: 260 seconds]
<cristian_c>
apritzel: bq sources seem support ds975
<cristian_c>
but i've to compile them yet
<cristian_c>
apritzel: if I go in android, into Settings->Information
<cristian_c>
I see in 'build_number': 'rk30 android sdk....'
<cristian_c>
'ds975-eng4.1.1 jro03h....'
<cristian_c>
apritzel: android stock rom developers have used rk30sdk and ds975 build, even if tablet is ds973h board
<cristian_c>
but ds975 build is working for android, so ds975 configuration should be used for all similar boards, included ds973
<lautriv_>
apritzel, nah. i'm aware of the SoC , i meant the board.
<rperier>
amstan: perhaps that coffee might help :p
* apritzel
is confused now, I haven't mentioned an 3188 board, I only have a RK3288 dev board
<lautriv_>
apritzel, maybe i mixed your nick o.O
rory096 has quit [Ping timeout: 268 seconds]
rory096 has joined #linux-rockchip
premoboss has joined #linux-rockchip
neyder_ has joined #linux-rockchip
rory096 has quit [Ping timeout: 246 seconds]
GriefNorth has joined #linux-rockchip
GriefNorth has quit [Client Quit]
GriefNorth has joined #linux-rockchip
GriefNorth has quit [Quit: Konversation terminated!]
GriefNorth has joined #linux-rockchip
ssvb has quit [Ping timeout: 264 seconds]
<mrutland>
mmind00: I tried to write code dumping the pinctrl hw state just after probing, but it appears to hang within regmap_read (in the critical section) when 743 static int rockchip_get_pull(struct rockchip_pin_bank *bank, int pin_num)
<mrutland>
aargh
<mrutland>
copy paste error
<mrutland>
mmind00: i tried rockchip_get_pull on pin 0, but this hangs in the critical section of regmap_read, which strikes me as surprising. Perhaps that programming interface requires a clock?
<mrutland>
Or I'm just doing something stupid ;)
<mmind00>
mrutland: yes it does ... but that should stay untouched
<mmind00>
mrutland: aha ... I think I know what might be responsible
<mrutland>
mmind00: That sounds promising :)
JohnDoe_71Rus has joined #linux-rockchip
<mmind00>
mrutland: the rockchip pinctrl/gpio driver got dynamic clocking ... and I'd guess that is one more piece fallout
<mrutland>
mmind00: Ah. Is that in mainline? I was doing this on today's HEAD - a794b4f3292160bb3fd0f1f90ec8df454e3b17b3
<mrutland>
Ah, it is
<mrutland>
qI guess I can try reverting 07a06ae99ef9b8eda3ec0b69c8f477856042a511
<mmind00>
mrutland: yep ... we had some fallout already, but I totally forgot that that have the rk3368 now too :-S
<mrutland>
mmind00: No worries, I'll jsut check that sovles the issue
<mrutland>
(then I'll learn to type)
<mmind00>
:-D
<mrutland>
mmind00: It looks like that's to blame. I'm just enabling drivers and verifying that other things work
<mrutland>
mmind00: ethernet, usb, mmc (I think) all seem to work with that, and the appropriate drivers enabled
<mrutland>
mmind00: I'll see about putting together a proper fix for that patch
<mmind00>
mrutland: great ... I can also track down the critical clock ...
<mrutland>
mmind00: If you're going to do that, should I leave it with you? This is basically guesswork for me
<mmind00>
mrutland: I think so :-)
<mmind00>
need to follow up on the existing clock fixes anyway
neyder_ has quit [Ping timeout: 240 seconds]
<lautriv_>
mmind00, mrutland your GPIO issues on vanilla-kernel ?
<mmind00>
lautriv_: yes ... the top of torvald's tree (and linux-next)
<lautriv_>
mmind00, i don't work on any part of rockchip but there is already a common of-routine for pinctrl which should make the timing less problematic, afair it was switched atomic.
<mmind00>
lautriv_: the problem we had is just that the pinctrl driver gained dynamic clocking (only enabling the gpio pclks when needed) ... that in turn caused clocks higher up in the clock tree to get disabled, that should stay on
<lautriv_>
ah
gb_master has joined #linux-rockchip
<mrutland>
Shame the FW left the CPUs at min freq though :/
<mrutland>
judging by perf measuring a busy loop, 600MHz for CPU{0-3}, 816MHz for CPU{4-7}
ssvb has joined #linux-rockchip
apritzel has quit [Ping timeout: 264 seconds]
<mmind00>
mrutland: the bigLittle cpufreq driver is your friend :-) ... athough somebody needs to teach it to handle the same regulator for two clusters
premoboss has quit [Remote host closed the connection]
<mrutland>
mmind00: Oh, I didn't see clocks for the CPUs in the DT
<mmind00>
mrutland: look in your inbox :-)
<mrutland>
mmind00: :)
<rperier>
ipad pro... 949$... (it's a surface with a bigger screen a "iPencil" and "an apple" behind)
<rperier>
technology...
<rperier>
:/
neyder_ has joined #linux-rockchip
markm has quit [Read error: Connection reset by peer]
rory096 has joined #linux-rockchip
<lautriv_>
does anyone know about an image for 3188T not using selinux ?
nighty^ has quit [Read error: Connection reset by peer]
nighty^ has joined #linux-rockchip
gb_master has quit [Remote host closed the connection]
mrueg has quit [Remote host closed the connection]
mrueg has joined #linux-rockchip
gb_master has joined #linux-rockchip
chrs_ has quit [Ping timeout: 250 seconds]
chrs_ has joined #linux-rockchip
gb_master has quit [Remote host closed the connection]
naobsd has quit [Remote host closed the connection]
<bashintosh>
lautriv_: you can set SELinux in permissive mode by modifying init.rc (ramdisk): s/on property:selinux.reload_policy=1/on property:selinux.reload_policy=0 - if that's not enough, add "setenforce 0" at the end of the file.