<amstan>
yeah, so leaving it always on is super bad
<amstan>
what happens is once a card enumerates the vqmmc gets set to 1.8V for faster speeds
<amstan>
then the card gets ejected, voltage stays at 1.8
<amstan>
but then a new card comes, till the mmc driver realizes that it has to change the regulator back to 3.3V for enumeration the sd card is actually outputting 3.3V on the io signals to the soc which is powered of 1.8V
<amstan>
so bad stuff could happen there
<amstan>
so you need that patch, and make sure to keep it off when there's no card(that way when it does turn on it goes to 3.3V directly)
<naobsd>
ah, my trial with patch also keep always-on
<naobsd>
I'm not sure vccio_sd on firefly can be off because it supplies sdmmc unit in RK3288
<naobsd>
then
<naobsd>
as a workaround, I'll change min-microvolt to 3.3V
<naobsd>
more accurately, sdmmc on firefly may not be able to work at 1.8V
<naobsd>
because sd slot signal lines are pulled by vcc_io 3.3v
<naobsd>
it should be pulled by vccio_sd
<naobsd>
anyway
<naobsd>
I don't have any uhs card
<naobsd>
I can't test any case with 1.8V
<naobsd>
(I'm not sure why vccio_sd is required, I don't have enough knowledge)
<phaedrus424>
naobsd: yeah, I know, but an improvement over 3.0.x! Trying to get an Ralink RT5392 working, tried to backport driver with no success, so ..
<phaedrus424>
Anyway, thanks, I will give it a whirl...
<naobsd>
higher version number doesn't mean "more functionality for everything"
<naobsd>
anyway you can try and solve your issue freely
rperier_ is now known as rperier
rperier has quit [Changing host]
rperier has joined #linux-rockchip
premoboss has joined #linux-rockchip
enyc_ has joined #linux-rockchip
<phaedrus424>
enyc_: stalker :)
ssvb has joined #linux-rockchip
<paowz>
naobsd: hi !! I'm still trying to mount a rootfs (I wont give up..) for the moment, when booting, it still hangs when trying to mount the rootfs.. http://pastebin.com/Wqyu6mV8
<paowz>
when I flash a valid image (rock_pro_lite_ubuntu_14.04_server_141030_sdcard.img) on sdcard I noticed, this sdcard is being properly read by my host OS (which mounts it on /media)
<paowz>
which tends to confirm the fact that a proper filesystem is created on the sdcard (otherwise it would not be mounted) and thus, shows the content of the rootfs..
<paowz>
when I start my board with this image, /dev contains mmcblk0p1, the partition is properly discovered. But I can't succeed at obtaining the same result with my own rootfs.. any clue/idea ??
paowz has quit [Ping timeout: 264 seconds]
paowz has joined #linux-rockchip
RayFlower_ has joined #linux-rockchip
<phaedrus424>
paowz: might be worth using the debug options in the ubuntu initramfs??
<phaedrus424>
paowz: debug=y on kernel cmdline
<paowz>
phaedrus424: I give it a try right now !
<phaedrus424>
paowz: OK, I'm rebooting my router, will brb
phaedrus424 has quit [Quit: leaving]
ssvb has quit [Ping timeout: 246 seconds]
RayFlower_ has quit [Read error: Connection reset by peer]
phaedrus424 has joined #linux-rockchip
<phaedrus424>
hmmm, that didn't work :(
RayFlower_ has joined #linux-rockchip
<paowz>
I'm almost sure my issue is around here..: [ 3.085360] mmc1: new high speed SDHC card at address aaaa
<phaedrus424>
how have you got the card partitioned?
<paowz>
phaedrus424: nop, actually.. I "dd" binaries (boot.img, rootfs and so on..) right onto the sdcard..
<paowz>
this is raw access.. not filesystem access..
<paowz>
(as long as I understand it..)
<naobsd>
paowz: you need to make partition
<naobsd>
well
<naobsd>
paowz: you need to make partition in MBR on your SD card with fdisk command or something like that
<naobsd>
paowz: if you want to use partition defined in parameter, you need to insmod rknand kernel module in initramfs and disable sdmmc0 driver from kernel config
<paowz>
wooww..
<paowz>
naobsd: ok.. now I understand why parameter file you find in rock_pro_lite_ubuntu_14.04_server_141030_sdcard.img just defines one partition (boot)
<paowz>
ok, I define partition with fdisk, how do I make rootfs address sector match with those defined by fdisk ?
<paowz>
I mean, when "dd"-ing rootfs onto the sdcard..
<naobsd>
paowz: if you make a partition on SD card, you can image with dd of=/path/to/partition without seek
<naobsd>
you can write image...
<paowz>
ah yes, indeed.. sorry, silly question..
RayFlower_ has quit [Read error: Connection reset by peer]
RayFlower_ has joined #linux-rockchip
premoboss has quit [Remote host closed the connection]
<paowz>
naobsd, phaedrus424: thanks to you both, it works like a charm ! I have to carry out some further tests, but I'm reaching the prompt !
<paowz>
..champagne ;)
* phaedrus424
makes ready to break his router again
<phaedrus424>
really should brb this time, hopefully ;-)
phaedrus424 has quit [Quit: leaving]
phaedrus424 has joined #linux-rockchip
FreezingAlt has joined #linux-rockchip
FreezingCold has quit [Ping timeout: 264 seconds]
paowz has quit [Ping timeout: 245 seconds]
ssvb has joined #linux-rockchip
<naobsd>
hmm, sdmmc/emmc on px2 seems slow...?
<naobsd>
mmc_host mmc2: Bus speed (slot 0) = 4000000Hz (slot req 25000000Hz, actual 4000000HZ div = 0)
<naobsd>
mmc_host mmc1: Bus speed (slot 0) = 32000000Hz (slot req 52000000Hz, actual 32000000HZ div = 0)
<naobsd>
are these clocks normal?
* phaedrus424
wonders why he can't get the MK902 to boot without flashing
<phaedrus424>
Do I need to write the bootloader to SDRAM as well?
viric has left #linux-rockchip [#linux-rockchip]
<naobsd>
phaedrus424: if you flashed bootloader to on-board flash storage, CPU will load it into RAM
<phaedrus424>
~/bin/rkflashtool M 0x62000000 0x1000000 <linuxium-boot.img
<phaedrus424>
~/bin/rkflashtool M 0x63000000 0x1000000 < linuxium-kernel.img
<phaedrus424>
rkflashtool M 0x64000000 0x1000000 < linuxium-parameter
<phaedrus424>
~/bin/rkflashtool B 0x63000000 0x64000000
<phaedrus424>
should have worked then?
<naobsd>
phaedrus424: I know that works with my images on my device, but I have no idea about your images. please try, then you'll get answer very easily
<phaedrus424>
Not working :(
<phaedrus424>
Do I need THE "KRNL" header if I'm writing straight to RAM?
<naobsd>
phaedrus424: both with header and without header work for me. please read description on that page.
<phaedrus424>
Odd. The images I'm writing to RAM are the ones that it already boots quite happily from nand ...
<phaedrus424>
Does it need a specific bootloader version to work?
RayFlower_ has quit [Read error: Connection reset by peer]
RayFlower_ has joined #linux-rockchip
<naobsd>
phaedrus424: are you using u-boot on that page, right?
<phaedrus424>
Nope, standard Rockchip bootloader...
<phaedrus424>
but isn't the RAM booting a MASKROM function?
<phaedrus424>
Is u-boot backward compatible with with the rockchip images and tools?
* phaedrus424
wonders if he should be brave
premoboss has joined #linux-rockchip
<naobsd>
phaedrus424: why do you think description about u-boot can be applied to other boot loader...
<naobsd>
phaedrus424: u-boot for RK on that page supports RK image
bengal has joined #linux-rockchip
<phaedrus424>
naobsd: OK, I will summon up the courage to install u-boot later :)
<phaedrus424>
thanks for having patience with me!
RayFlower_ has quit [Read error: Connection reset by peer]
RayFlower_ has joined #linux-rockchip
RayFlower_ has quit [Read error: Connection reset by peer]
RayFlower_ has joined #linux-rockchip
zombu2 has joined #linux-rockchip
dlezcano has joined #linux-rockchip
<dlezcano>
mmind00: when you have time, please review the new timer version, I would like to merge it for 3.20 and I am not so far to send the PR to Thomas
RayFlower_ has quit [Read error: Connection reset by peer]
RayFlower_ has joined #linux-rockchip
RayFlower_ has quit [Read error: Connection reset by peer]
RayFlower_ has joined #linux-rockchip
<mmind00>
dlezcano: review done ... I hope was not to confusing :-)
<mmind00>
dlezcano: in short drop the dtsi change, and apply the result ... the separate dtsi patch is already in my devicetree branch here and I'm going to push it once you tell me you applied the patch :-)
jas-hacks has left #linux-rockchip [#linux-rockchip]
RayFlower_ has quit [Quit: RayFlower_]
paowz has quit [Remote host closed the connection]
<naobsd>
jas-hacks: nice work!
<naobsd>
btw chromium os repo has video codec support for rk3288 too
naobsd has quit [Quit: naobsd]
phaedrus424 has quit [Quit: leaving]
dan64B has joined #linux-rockchip
FreezingAlt has quit [Remote host closed the connection]
FreezingAlt has joined #linux-rockchip
<dan64B>
Hi Everyone!
<dan64B>
anyone knows why there are so many GitHub repositories of Linux for Rockchip? (e.g. Galland , Ian, linux-rockchip, omegamoon Rockchip-GPL-kernel, etc).
<dan64B>
any good reason?
nighty-_ has quit [Quit: Disappears in a puff of smoke]
bengal has quit [Quit: Leaving]
<karlp>
because they all came from slightly different vendor sdks, with a few hacked up patches, a long time ago
<karlp>
and yes, it's gross
<dan64B>
I see. Would you say that "linux-rockchip" is the mainline then? And the recommended tree to continue developing on?
<karlp>
well, are you hoping to head towards mainline kernels, or do you just need something that works on the device you have, and don't care much about whether it's a pile of blobs of hacked old linux 3.0?
tonikasch has joined #linux-rockchip
<dan64B>
heading towards mainline kernels.
<mmind00>
dan64B: maybe also important, which hardware and what functionality do you expect :-)
<dan64B>
right now I'm testing with Minix neo x7
<mmind00>
I was more fishing for the type of Rockchip soc in it
<dan64B>
rk3188 for now.
<tonikasch>
hi-fi
<karlp>
rk3188 is just the soc, there's a pile of hardware around it that maybe of relevance :)
tonikasch has quit [Remote host closed the connection]
<dan64B>
You mean the hardware that's mapped by GPIO? Those I can work with later and build appropriate modules for.