<karlp>
anyone used usbtest.c or testusb.c and can tell me the differences? I just get unsupporte ioctl with usbtest.c presumably it needs a kernel module to do all the work, which is presumably not in my regular deskto linux kernel
<karlp>
hrm, usbtest.c is a module that you install with a vid/pid to test, testusb is a user script that relies on some module to do the testing.
levd has joined #linux-rockchip
levd1 has quit [Ping timeout: 244 seconds]
dlezcano has quit [Remote host closed the connection]
jey has joined #linux-rockchip
pizthewiz has quit [Quit:
levd1 has joined #linux-rockchip
levd has quit [Ping timeout: 265 seconds]
pizthewiz has joined #linux-rockchip
pizthewiz has quit [Client Quit]
levd has joined #linux-rockchip
levd1 has quit [Ping timeout: 272 seconds]
jey has quit [Remote host closed the connection]
levd1 has joined #linux-rockchip
levd has quit [Ping timeout: 250 seconds]
GTRsdk has joined #linux-rockchip
levd has joined #linux-rockchip
levd1 has quit [Ping timeout: 256 seconds]
<leming>
mmind00, i can get you configs this weekend.. away on a business trip right now
<leming>
but they're just an alsa stored config to enable left/right speaker left and right DACs, and then tweaks to the pulseaudio config to set speakers as the default instead of hdmi
<leming>
nothing particularly special.. been doing the same audio setup for chromebooks since the first one
levd1 has joined #linux-rockchip
jey has joined #linux-rockchip
levd has quit [Ping timeout: 256 seconds]
levd has joined #linux-rockchip
levd1 has quit [Ping timeout: 272 seconds]
JohnDoe_71Rus has joined #linux-rockchip
levd1 has joined #linux-rockchip
levd has quit [Ping timeout: 272 seconds]
levd has joined #linux-rockchip
pizthewiz has joined #linux-rockchip
levd1 has quit [Ping timeout: 244 seconds]
pizthewiz has quit [Remote host closed the connection]
levd1 has joined #linux-rockchip
levd has quit [Ping timeout: 265 seconds]
<rperier>
mrjay, mmind00: does not boot here, it seems to hang in "of_platform_bus_create" for rk3066a-pinctrl
<rperier>
even after applied what heiko suggested
<rperier>
I will investigate this evening
akaizen has quit [Remote host closed the connection]
levd has joined #linux-rockchip
cnxsoft has quit [Ping timeout: 240 seconds]
cnxsoft has joined #linux-rockchip
levd1 has quit [Ping timeout: 240 seconds]
jey has quit [Remote host closed the connection]
levd1 has joined #linux-rockchip
levd has quit [Ping timeout: 250 seconds]
levd1 has quit [Remote host closed the connection]
levd has joined #linux-rockchip
levd has quit [Remote host closed the connection]
levd has joined #linux-rockchip
levd1 has joined #linux-rockchip
levd has quit [Ping timeout: 256 seconds]
rory096 has quit [Ping timeout: 246 seconds]
<Apocx>
is there any version of the linux kernel >= 3.4 where the rockchip NAND stuff still works?
<sjoerd>
naobsd: something you might know, does the rockchip SROM just look at "IDR 0" for a bootloader or will it try a few different places?
<cristian_c>
tony_: hello
<tony_>
cristian_c, hi.
<cristian_c>
tony_: I'm trying to make touchscreen working, cat /dev/input/event1 answers to touch events
<tony_>
cristian_c, you can try this one. getevent
<cristian_c>
mtdev-test provides data when I generate touch events, too
<cristian_c>
tony_: ah, ok
<tony_>
cristian_c, better than "cat"
<cristian_c>
tony_: I've found a way to start ts_calibrate, using tslib_tsdevice=/dev/input/event1 in ts_calibrate command
<tony_>
cristian_c, what os do you use ?
<tony_>
cristian_c, not Android ?
<cristian_c>
tony_: ok, but already mtdev-test provides better data than ascii provided by cat :P
<cristian_c>
tony_: no
<tony_>
cristian_c, for now, everything works fine ?
<cristian_c>
tony_: so, if I launch ts_calibrate either ts_test, application screen appears, I try to touch the cross air, as asked by the app, but nothing happens
<cristian_c>
so, I've to restart the device by ctrl+alt+del, because I can't exit from the app
<cristian_c>
tony_: I've some issues with some devices
<cristian_c>
tony_: I don't know how to test gsensor
<cristian_c>
backlight is good, but I have not a brightness applet
<tony_>
cristian_c, I want to know what os do you use ?
<tony_>
cristian_c, what cpu ?
<cristian_c>
tony_: screen is partially duplicated, if I use fbset, I can change resolution but issue is not fixed. Duplicated screen issue appears also in early kernel boot messages screen
<cristian_c>
tony_: rk3066, ubuntu
<cristian_c>
tony_: front camera is well initialized but it fails when I launch guvcview
<cristian_c>
for strange reasons
<tony_>
cristian_c, what version kernel do you use ? mainline ?
<tony_>
cristian_c, or your vendor ?
<cristian_c>
tony_: 3.0.8+, not mainline
<cristian_c>
tony_: the vendor does not provides spurces for this device
<cristian_c>
tony_: I've looked at git.kernel
<cristian_c>
but I don't understand exactly what I've to do :D
<tony_>
cristian_c, cool, you tried ubuntu on rk3066.
<tony_>
cristian_c, I have tried ubuntu on rk3188, it works well, such as lcd, audio...
<tony_>
cristian_c, but the camera don't work fine.
<cristian_c>
tony_: the difficult/particular thing is booting on tablet
<cristian_c>
tony_: ah, ok, I didn't know anyone had tried rockchip + linux + tablet
<tony_>
cristian_c, I was lucky than you. I can get source.
<tony_>
cristian_c, I think there are more information if you use radax board, it for developer.
<tony_>
cristian_c, a product is for enduser. ;)
von_fritz has joined #linux-rockchip
levd1 has joined #linux-rockchip
levd has quit [Ping timeout: 252 seconds]
levd1 has quit [Ping timeout: 245 seconds]
ssvb has joined #linux-rockchip
<cristian_c>
tony_: ok, getevent is only for android
<tony_>
cristian_c, I modified it for normal linux os. do you want it ?
<cristian_c>
tony_: ok
<tony_>
cristian_c, how to send it to you ?
<cristian_c>
Uhm
<tony_>
cristian_c, I can not use google drive recently.
<tony_>
cristian_c, ;(
<cristian_c>
what service do you use?
<tony_>
maybe I upload it to github.
<cristian_c>
ok
<c0d3z3r0>
rperier: the _same_ kernel version with the _same_ config which worked for days doesn't work anymore (except the cpu problem) and is stuck at "Failed to find PMU node" o.O sometimes I believe that radxarock leads a life of its own
<rperier>
c0d3z3r0: what I found is that rockchip_dt_init (arch/arm/mach-rockchip/rockchip.c) calls of_platform_populate which calls of_platform_bus_create which hangs for rk3066a-pinctrl
<rperier>
which suggests that this is another clk or pinctrl issue
<c0d3z3r0>
I sent in a patch to add pclck_cpu to the critical clocks which solved the problem that the regulator wasn't found anymore
<rperier>
your patch is applied on my kernel
<rperier>
but it does not work :)
<rperier>
it hangs anyway
<c0d3z3r0>
I try again with after reverting the patch
<rperier>
The message "Failed to find PMU node" should not be critical, that's because the kernel don't find a variant for the pmu and suspend part for your soc (see arch/arm/mach-rockchip/pm.c)
<c0d3z3r0>
rperier: yeah. what I don't understad is why this kernel worked for the past 72h and now it hangs
<c0d3z3r0>
that doesn't make any sense
<rperier>
because the problem is not deterministic at all ?
<rperier>
it happens sometime :)
<c0d3z3r0>
it doesn't work at all… even after 20 retries :D
<c0d3z3r0>
ok… let's test reverting the patch
<rperier>
or perhaps that you went through a time warp :P ... "back to the future"
<c0d3z3r0>
maybe :D
<rperier>
:D
<tony_>
cristian_c, I want to know how to set resolution by fbset.
<tony_>
cristian_c, would you help me ?
<cristian_c>
tony_: I'm not an expert, but I can try to help :)
<rperier>
c0d3z3r0: I does not make sense.... I mean... in the commit I posted few minutes ago
<rperier>
bank->clk is "PCLK_GPIO0"
<tony_>
cristian_c, I want to set resolution by fbset on rk3188 (3.0 version kernel).
<tony_>
cristian_c, but I get error.
<tony_>
cristian_c, the xres_virtual and yres_virtual can not be set.
<c0d3z3r0>
rperier: it boots fine with reverted pinctrl commit; the cpu problem stays
<rperier>
c0d3z3r0: pinctrl-rockchip.c load bank->clk from device tree, the node is the one with the compatible string "rockchip,gpio-bank" (or rockchip,rk3188-gpio-bank for you)
<tony_>
cristian_c, I'm sure I give vxres agr, but I don't know why ...
<rperier>
the only clock I found here is PCLK_GPIO0
<rperier>
which has pclk_cpu as parent
<rperier>
so...
<cristian_c>
but your xres and yres have not to go beyond max allowed values for ypur display
<cristian_c>
else you can get a fbset error message
<rperier>
c0d3z3r0: what do you think ?
<naobsd>
sjoerd: as far as I know, there is no backup/alternative location for IDB blocks.
<c0d3z3r0>
rperier: act8846 is connected to i2c1 which is supplied by pclk_cpu
<tony_>
cristian_c, really ?
<cristian_c>
tony_: I've fbset without busybox
<tony_>
cristian_c, oh, my god.
<cristian_c>
tony_: do you talk about android
<cristian_c>
?
<c0d3z3r0>
rperier: adding pclk_cpu solved the i2c / regulator issue
<cristian_c>
tony_: i've tried only in linux distro
<tony_>
cristian_c, yes. I'm talk about android, but i think the fbset is a cmd on linux.
<cristian_c>
and a years ago in cwm recovery
<cristian_c>
year
<rperier>
yeah but our freeze seems to be caused by the commit in pinctrl-rockchip.c which changes how bank->clk is used...
<cristian_c>
tony_: sometimes there are some differences between busybox and distro tools
<rperier>
what I mean is that this clock is the gate PCLK_GPIO0
akaizen has joined #linux-rockchip
<cristian_c>
tony_: so, I can't say anymore if I don't try in android directly
<cristian_c>
until
<tony_>
cristian_c, okay, I will try the little reslution than before.
naobsd has quit [Quit: naobsd]
<cristian_c>
tony_: what lcd display are you using? :)
<c0d3z3r0>
rperier: PCLK_GPIO0 is also supplied by pclk_cpu
<cristian_c>
tony_: two days ago,cI've tried to change xres higher than 1024, and yres higher than 768
<tony_>
cristian_c, I think the vxres cab be also adjusted.
<tony_>
cristian_c, then ...
<cristian_c>
but I've got a fbset error. I've not saved output , anyway
<rperier>
c0d3z3r0: I just did read what is changed by commit 07a06ae99ef9b8eda3ec0b69c8f477856042a511 , the only part changed is clk handling for "bank->clk" (see the diff)
<cristian_c>
I remember there was a vidioc error, maybe
<cristian_c>
tony_: ah, ok
<rperier>
this clk is PCLK_GPIO0 for rk3066 and rk3188 so I have just checked that it was not another clk issue
<cristian_c>
but I'm not an expert
<cristian_c>
tony_: thanks for the link
<rperier>
however, as you said this clock is supplied by pclk_cpu, so it seems to be good. the freeze seems to be caused by something else
<tony_>
cristian_c, after all, little resolution won't got error. ;P
<tony_>
cristian_c, but after that, I get current info, it is old one.
<cristian_c>
uhm
<tony_>
cristian_c, maybe I should get answer from code.
<c0d3z3r0>
rperier: interesting.. applied the pinctrl patch again. now it boots
<rperier>
are you sure that your kernel has been correctly built... ? (I mean no ccache or something)
<rperier>
:/
<c0d3z3r0>
I completely disabled ccache because I had such problems…
<c0d3z3r0>
time to reboot my build machin
<c0d3z3r0>
e
<rperier>
interesting...
<cristian_c>
-vxres <value>
<cristian_c>
set virtual horizontal resolution (in pixels)
<cristian_c>
tony_: interesting
<karlp>
rperier: do you still have ccache problems, or see them anywhere? or is this old suspicions fromt he past?
<tony_>
cristian_c, maybe there is bug in rk's code. ;P
<cristian_c>
tony_: I'll use this option for trying to fix my issue
<karlp>
c0d3z3r0: that'
<cristian_c>
tony_: possible :D
<tony_>
cristian_c, I will try to fix it .
<karlp>
s for you too :)
<cristian_c>
tony_: ok
<rperier>
karlp: no I don't use ccache (I build everything via yocto which does not use ccache), but some people here do
<cristian_c>
tony_: let me know :)
<rperier>
I asked just in case
<tony_>
cristian_c, no problem. ;P
<karlp>
I know some build environments have issues, but those are normally fixed in the build environment, I wouldn't have expected anything like yocto to have ccache problems that's all.
<c0d3z3r0>
karlp: there was one problem some days ago… I added a critical clock that is only used by rk3288 but it worked for rk3188.. after cleaning the ccache and git workdir everything was fine again
<karlp>
yeah, that doesn't sound like a ccache problem...
<c0d3z3r0>
karlp: I use the linaro toolchain 4.9-2015.05
naobsd has joined #linux-rockchip
cosm has joined #linux-rockchip
naobsd has quit [Quit: naobsd]
cosm has quit [Ping timeout: 244 seconds]
tony_ has quit [Remote host closed the connection]
cristian_c has quit [Quit: Bye]
<c0d3z3r0>
rperier: after some more testing I found out that this is another kexec problem. the kernel boots fine normally but not with kexec
<c0d3z3r0>
but that does not explain why it worked until this morning
<rperier>
you mean... if you boot your kernel with all lastest fixes (pclk_cpu+pll_mux from heiko) from... NAND for example (or sdcard) it boots just fine ?
<rperier>
but when you boot it via kexec , ie you load your new kernel in memory and you reboot to this new kernel on the fly (without touching the old one in flash) it does not boot at all... right?
<rperier>
or you just talk about CONFIG_KEXEC ?
<c0d3z3r0>
rperier: yes, that is the same problem like the cpus… booting without kexec works but loading the new kernel with kexec reveals those problems
<rperier>
I will definitely work on this issue, I use kexec everytime on all my devices... (usually I only flash stable kernel to nand and use petitboot to reboot a new one via kexec)
<rperier>
a simple bash script to boot over the network using kexec
<rperier>
it is very small
<rperier>
your project might be interesting for development tasks
<rperier>
(because net-kexec does nothing by itself...)
<rperier>
the nice stuff with petitboot is that you can boot from all locations , the media or the location just need to be supported by the linux kernel
<rperier>
so you can boot from usb, sdcard... nand... emmc or ethernet or even wifi
<rperier>
for example
<c0d3z3r0>
wow that is really cool :)
<rperier>
yes
<rperier>
it is bigger but cool
<rperier>
until we get all these features in uboot, that's a good alternative
<rperier>
(because booting from the network via uboot is definitely faster)
<rperier>
c0d3z3r0: the support is still a bit buggy in yocto, so I think that for your own development tasks rockboot is fine
<sjoerd>
nice
<rperier>
wait an example
<sjoerd>
fwiw i've been slowly poking u-boot to bring up the network on 3288, but it's slowgoing due to mimimal amount of free time and getting to learn a lot more low level bits of rockchip socs ;)
* sjoerd
hopes he's got something working sometime next week
<rperier>
this is an example of what you can do with petitboot
<c0d3z3r0>
rperier: very nice :D
<rperier>
hehe ;D
<c0d3z3r0>
I'll try this out :D
<rperier>
are you familiar with yocto ?
<c0d3z3r0>
no, I use the linaro toolchain
<c0d3z3r0>
or.. is this more than a toolchain?
<rperier>
yes
<rperier>
that's a project with tools (like toolchain and many other stuffs) which can build an embedded distribution for you
<rperier>
the distro is "poky" or "oe-core" usually
<rperier>
including kernel, rootfs... everything
<rperier>
like buildroot
<c0d3z3r0>
rperier: I tried an ealier version of the toolchain just to be sure that it's not the toolchain that is bad… but no change
<c0d3z3r0>
how does u-boot load the kernel? is that kexec, too?
<rperier>
I don't think so, as I have the same kind of issue with a different one
<sjoerd>
c0d3z3r0: the problem for the kernel is likely that it has setup more stuff which it needs to set back in a proper state
<sjoerd>
c0d3z3r0: u-boot linux loading is relatively simple as it it won't have poked the hardware lots, so it cna just turn off the mmu, disable caches and jump into the kernel that's in memory
<c0d3z3r0>
sjoerd: ok, that maybe explains why u-boot works and kexec doesn't
<rperier>
exactly, the old kernel is supposed to power off and uninitiliaze the hw but it might also leave the system in an unstable state... uboot is just executed after the first stage bootloader... (which does nothing except configure ddr and some pins)
<rperier>
a lot of things are different
<c0d3z3r0>
u-boot's fatload doesn't work on rk3188, right?
<rperier>
I am wondering something
<rperier>
when you reboot via kexec
<rperier>
which kernel do you use in NAND ?
<c0d3z3r0>
I have built a more or less minimal UP kernel that is booted by u-boot on sdcard
<c0d3z3r0>
that loads an initramfs which finally executes kexec
<rperier>
no I mean, which version
<c0d3z3r0>
next
nighty^ has joined #linux-rockchip
<c0d3z3r0>
next-20150817
tizbac has quit [Ping timeout: 246 seconds]
<rperier>
and the new one too ? (loaded via kexec)
<c0d3z3r0>
yes
<rperier>
if the problem is related to kexec... I think that it is only when you start the kernel not when you turn off the old one
<rperier>
because I have a kernel 4.0.0 in NAND
<c0d3z3r0>
so I should try to load next-20150817 with 4.0.0?
<rperier>
I don't think that a 4.0 leaves the system in an unstable state...
<c0d3z3r0>
I'll try that later brb
tizbac has joined #linux-rockchip
JohnDoe_71Rus has joined #linux-rockchip
ssvb has quit [Ping timeout: 252 seconds]
von_fritz has quit [Quit: Sto andando via]
<c0d3z3r0>
rperier: 4.0 crashes without panic after dwmmc being detected
topi`_ is now known as topi`
<c0d3z3r0>
rperier: 4.1.6 boots but kexec is not working like with next
naobsd has joined #linux-rockchip
<rperier>
c0d3z3r0: same issue ? right ? (it hangs at "PMU not found")
<c0d3z3r0>
yes
akaizen has joined #linux-rockchip
ssvb has joined #linux-rockchip
rory096 has joined #linux-rockchip
akaizen has quit [Remote host closed the connection]
akaizen has joined #linux-rockchip
cosm has joined #linux-rockchip
akaizen has quit [Remote host closed the connection]
akaizen has joined #linux-rockchip
mrjay has joined #linux-rockchip
<mrjay>
Is there a way to enable idmac mode for dwmmc on rk3066? Do i need some extra dts config?
<mrjay>
I'm asking because i'm having some instability with mmc devices (sdcard and sdio wifi)
<mrjay>
sdcard hangs the system on for examle copying big file to dwmmc
<mrjay>
wifi randomly disconnects
<mrjay>
anybody tried that?
<rperier>
there is a patches serie for that on the ML, if I remeber correctly
<rperier>
"[RFC PATCH v5 0/9] Add external dma support for Synopsys MSHC"
<rperier>
with devicetree bindings doc and patches for the devicetree
<c0d3z3r0>
mrjay: I had the sdcard issue on rk3188 and also rpi2… but enabling idmc crashed the kernel…
<mrjay>
rperier: thanks ... i'll try them
<rperier>
so you can test these patch and enable the support in the dts
<rperier>
not sure it will be 100% stable but you should be able to do what you want with these
<mrjay>
c0d3z3r0: yeap mine is crashing also on boot if i change uuid of the partition on the sdcar to boot partitioon uid, after changing uuid to difrent one mk808 booting properly
<c0d3z3r0>
mrjay: with linux-next?
<mrjay>
c0d3z3r0: but not booting from microsd ;)
<mrjay>
c0d3z3r0: yeap
<mrjay>
c0d3z3r0: with linux-next + mmind patches from yesterday
cnxsoft has quit [Remote host closed the connection]
<mrjay>
c0d3z3r0: i'm talking about idmac disabled at boot ... sorry for the confusion
<c0d3z3r0>
idmac disabled crashes?
<mrjay>
c0d3z3r0: yes with uuid of the sdcard partition changed to boot partition uuid (root=UUID=.... in kernel cmdline)
cosm has quit [Ping timeout: 246 seconds]
<c0d3z3r0>
mrjay: ah okay, I never used UUIDs ;)
<c0d3z3r0>
always /dev/mmcblkXX
<mrjay>
c0d3z3r0: if idmac enabled in config kernel crashes always on mk808 :(
<c0d3z3r0>
it crashed for me when idmac was enabled but that was some time ago. currently it's disabled
<Apocx>
I have a feeling this command (find . ! -path "./.git*" | cpio -H newc -ov > initrd.img) should not be copying .c files, and the initrd.img should not be as large as 800mb. Am I wrong in this assumption? Following the mainline linux guide here: http://radxa.com/Rock/Linux_Mainline
<Apocx>
aha, that would make more sense. ok thanks
akaizen has quit [Remote host closed the connection]
gb_master has joined #linux-rockchip
JohnDoe5 has joined #linux-rockchip
JohnDoe_71Rus has quit [Ping timeout: 252 seconds]
<mmind00>
mrjay: rperier: generally, idmac support needs to be build into the dwmmc ip ... rk3288 seems to be the first one to include it
<mmind00>
mrjay: rperier: for extdma the general dma issue might still bite you
<rperier>
arrff, I see
<rperier>
pinctrl on rk3066 seems to hang in "rockchip_interrupts_register"
<mmind00>
rperier: you could narrow it down some more ... I don't have access to anything rk3066/rk3188 based for the next 2.5 weeks ... otherwise I'll take a look after the merge window
<rperier>
I am still investigating, I did some progress
<Apocx>
I followed the Linux_Mainline guide and generated a boot.img file, but when writing it to the device I get "premature end of file reached" and the device no longer boots. so I had to restore the backup. any ideas what could cause that?
<rperier>
mmind00: yeah I know, don't worry I am working on it ;)
<rperier>
mrjay: yeah but it does not include "pclk_peri"
<rperier>
mmind00: if you have time I think that you can send a patch for https://bpaste.net/show/71e37e17ae27 (as this is your fix...) and I will send a patch to add pclk_peri to the criticial clocks
<rperier>
if that is possible of course..
<mmind00>
rperier: yep ... am a bit short on time today, but it is on my list already ;-)
gb_master has quit [Quit: Leaving]
<rperier>
:)
<rperier>
c0d3z3r0: if you're interested I can upload a petitboot image somewhere for radxarock :)
<rperier>
if you want to test...
pjr has joined #linux-rockchip
cristian_c has joined #linux-rockchip
<Apocx>
well trying to get the 3.18 linux kernel on the Rock Pro was a bust. anyone know a different SoC board that has a wifi antenna connector, builtin wifi/bluetooth, and NAND flash storage? maybe one that can actually run a somewhat recent version of linux with minimal headaches?
<Apocx>
cause I followed radxa's guide perfectly, but now the rock pro just blinks the green led and doesn't boot with the new boot image
<mrjay>
Apocx: try jetson tk1. Mainline support is not complete but ok. It has minipcie for wifi/bt, has usb 3.0, sata 3 with port replicator support, opengl 4, cuda ... and good nvidia support. But its not cheap :).
<Apocx>
yeah $192 is a bit steep, trying to keep it $100 or less. looks interesting though
<Apocx>
probably overkill for what we need though
rperier has quit [Ping timeout: 244 seconds]
architekt has quit [Ping timeout: 244 seconds]
rperier has joined #linux-rockchip
<c0d3z3r0>
rperier: thanks! :) can test tommorrow
architekt has joined #linux-rockchip
<Apocx>
when generating a boot.img file, is it necessary to build the kernel modules as well (i.e. make modules modules_install)