ChanServ changed the topic of #linux-rockchip to: Rockchip development discussion | Wiki at | Logs at
Celia has joined #linux-rockchip
rperier has joined #linux-rockchip
dianders_ has joined #linux-rockchip
ChiaSmurf has quit [*.net *.split]
dianders has quit [*.net *.split]
rperier_ has quit [*.net *.split]
FergusL has quit [*.net *.split]
hramrach has quit [*.net *.split]
dianders_ is now known as dianders
ChiaSmurf has joined #linux-rockchip
FreezingCold has quit [Ping timeout: 245 seconds]
FreezingCold has joined #linux-rockchip
hipboi has quit [Ping timeout: 272 seconds]
hipboi has joined #linux-rockchip
FightingCat has joined #linux-rockchip
FightingCat has quit [Ping timeout: 272 seconds]
hipboi has quit [Ping timeout: 260 seconds]
hipboi has joined #linux-rockchip
FightingCat has joined #linux-rockchip
ChiaSmurf has quit [Ping timeout: 258 seconds]
<naobsd> karlp: 1: use gcc arm-eabi-4.6 from android prebuilts 2: compare dmesg from stock rom?
<naobsd> karlp: and gpio dump? lcdc configuration may assert reset on your board
GriefNorth has joined #linux-rockchip
ChiaSmurf has joined #linux-rockchip
ferric_ has joined #linux-rockchip
ferric has quit [Ping timeout: 250 seconds]
FreezingCold has quit [Ping timeout: 272 seconds]
Celia is now known as FergusL
FreezingCold has joined #linux-rockchip
GriefNorth has quit [Ping timeout: 258 seconds]
GriefNorth has joined #linux-rockchip
hipboi has quit [Ping timeout: 255 seconds]
<rperier> hi all
hipboi has joined #linux-rockchip
phh__ is now known as phh
hramrach has joined #linux-rockchip
dianders has quit [Changing host]
dianders has joined #linux-rockchip
viric_ has joined #linux-rockchip
viric has quit [Ping timeout: 272 seconds]
viric_ is now known as viric
<karlp> naobsd: what's the best way of getting a gpio dump from the factory rom?
<karlp> will try that tonight, it's not very comforting when I can't get a single 3.0.8+ kernel to boot, from any source
<naobsd> there is no generic source for any device...
<mmind00> naobsd: at least not yet ;-)
* karlp grins.
<karlp> the tedious bit is that there's sources labelled as "this works on device x" that don't work on device x for me.
<karlp> but it is very comforting that linux mainline works as advertised :)
bengal has joined #linux-rockchip
<naobsd> mmind00: oh, you have R89 :)
<mmind00> yep
<naobsd> I'll submit dts for my rk3288 box :)
bengal has quit [Ping timeout: 260 seconds]
viric has quit [Ping timeout: 272 seconds]
viric has joined #linux-rockchip
nighty^ has joined #linux-rockchip
FightingCat has quit [Ping timeout: 246 seconds]
GriefNorth has quit [Ping timeout: 260 seconds]
viric has quit [Read error: Connection reset by peer]
FightingCat has joined #linux-rockchip
viric has joined #linux-rockchip
FightingCat has quit [Ping timeout: 246 seconds]
<rperier> in your opinion, it would be interesting to submit defconfig for Rockchip SoCs and/or Rockchip boards to the mainline kernel ?
<karlp> why not? there's already bqcurie, which is just a generic tablet.
<mmind00> I think the suggested way nowadays is to have everything necessary in the multi_v7_defconfig ... it is supposed to be a multiplatform-kernel afterall
VargaD_ has joined #linux-rockchip
<mmind00> the maximum would be a generic rockchip_defconfig for all socs, without debug_ll etc
VargaD has quit [Ping timeout: 250 seconds]
VargaD_ is now known as VargaD
<karlp> oh, yeah, he meant defconfig, not dts.
<rperier> multi_v7_defconfig a lot of optionnal support for "nothing", right ? (even if everything is built as module I guess, but it increases the roots size and the build time)
<rperier> contains *
<rperier> both might be fine
<rperier> (multi_v7 and rockchip)
<mmind00> you might get problems with the bootargs u-boot provides, as overriding it like the ones in my devel-tree is out of the question
<mmind00> and on this note: ... adding "linux,stdout-path" handling to common serial code might be nice to do
FightingCat has joined #linux-rockchip
FightingCat has quit [Remote host closed the connection]
BorgCuba has joined #linux-rockchip
<BorgCuba> karlp, for the wiki: rk-makebootable from rkboottools can build the "magic header"
<BorgCuba> basically the first 16 bytes look like this: 55 aa f0 0f 00 00 00 00 00 00 00 00 04 00 04 00
<BorgCuba> and the last two dwords of the 512chunk give the size of the two following chunks in 512byte units (iirc)
<naobsd> last two dwords?
<karlp> I had a bit of a skim through the irc logs, I saw mention that rk3066 doesn't actually support sd boot? anyone have any facts for that? (I'm not itching to try out making that magic header on an sd card if it doesn't work anyway)
<BorgCuba> naobsd, you are right 16 bits and the last is zero, and the the two before this are the lengths
<BorgCuba> karlp, I haven tried
<BorgCuba> but let me see
<BorgCuba> karlp, I tried my rk3188 boot sd on a rk3066 device and it just boots the kernel from nand
<BorgCuba> well, on the rk3188 it does not but instead loads my unfinished u-boot code
<BorgCuba> karlp, you want to test this on a rk3066 device?
<BorgCuba> btw, maybe I can help you out on the boot-order section
<BorgCuba> I downloaded the bootrom for my device (rk3188-t) and the is some table for the boot devices
<naobsd> BorgCuba: what's unfinished u-boot code? based on upstream u-boot?
<BorgCuba> no, older 2013 code from rockchip
<BorgCuba> I am trying to remove all this fastboot bs
<naobsd> why remove fastboot?
<BorgCuba> I never used fastboot
<BorgCuba> can you use fastboot to execute a kernel from dram?
<naobsd> not tried yet.
<naobsd> but if "fastboot boot xxx.img" works, it's useful
<BorgCuba> I want to get the cdc_ether gadget running
<BorgCuba> I think it worked almost out of the box, but I had to comment to calls to some pll setup code
<BorgCuba> btw, this is the brom version string of my RK3188-T "B01321020311001V"
<BorgCuba> is this the same as for a regular rk3188?
<rperier> oh
<rperier> kernel BUG at net/core/skbuff.c:100!
<rperier> :D
<naobsd> probably it's same as RK3188. you have to see a bit in efuse to identify T or not
<rperier> well, apparently it's just a buffer overflow...
<BorgCuba> naobsd, I think you are right. I can see some one year old irc-logs mentioning this exact version (I think the rk3188-t appeared later)
<naobsd> that info can be get from USB
<BorgCuba> nice, from brom or the sd/nand loader? and is there a tool for this?
<naobsd> I modified rkflashtool, but I cannot find it now ;)
<naobsd> one another command in loader
<BorgCuba> so probably the sd/nand loader
<naobsd> I guess RK3188 and -T are physically same... some RK3188 which is not stable at 1.6GHz is marked as -T ;)
<karlp> do you know the difference between rk3066a and 3066b?
<BorgCuba> probably, also I had problems with pll code from this u-boot zip I downloaded from radxa wiki
<naobsd> u-boot zip? not on github?
<BorgCuba> karlp, I dont know. I have a "RK3066" "ANF0895 1251"
<BorgCuba> naobsd, yes
<naobsd> this is latest one I know
<BorgCuba> sorry, actually from git
<naobsd> I only tried boot from NAND/SD as like as RK loader
<BorgCuba> but I have another one: ubootrk.tar.gz
<naobsd> ubootrk.tar.gz should be old
<BorgCuba> its from:
<BorgCuba> naobsd, do you recommend cloning your git?
<BorgCuba> naobsd, does it "boot"?
<naobsd> my fork boots from NAND/SD, i.e. it can load parameter/kernel/etc from NAND/SD and can go further
<BorgCuba> cool!
<BorgCuba> I just cloned it
<naobsd> actually NAND boot just work, I just tweaked for SD
<naobsd> I have no idea about ubootrk on crew
<naobsd> I guess it's from Radxa
<BorgCuba> yeah, maybe. I forgot where I downloaded it from
<naobsd> crew's git is protected?
FreezingCold has quit [Ping timeout: 255 seconds]
<BorgCuba> I have no idea?
<naobsd> I couldn't clone it...anyway, it's ok, it's not important for me...
<BorgCuba> naobsd, you u-boot code doesnt work for me
<BorgCuba> maybe if I remove the that pll code ...
<naobsd> how it doesn't work :(
<BorgCuba> no output on the serial, the last thing I see is "OUT" from the ddr code
<BorgCuba> I commented out the following lines in arch/arm/cpu/armv7/rk30xx/clock.c:
<BorgCuba> rkclk_pll_clk_set_rate(GPLL_ID, CONFIG_RKCLK_GPLL_FREQ, rkclk_gpll_cb);
<BorgCuba> rkclk_pll_clk_set_rate(CPLL_ID, CONFIG_RKCLK_CPLL_FREQ, NULL);
<naobsd> did you try u-boot w/o mod?
<BorgCuba> and now it looks like this:
<BorgCuba> I put u-boot right after the ddr code
<BorgCuba> have you seen these ftl_memcpy, ftl_xxx functions?
<BorgCuba> naobsd, what do you mean by "w/o mod" ?
<naobsd> without modification
<BorgCuba> only the modification I mentioned
<naobsd> I asked what happen if you don't modify at all...
<naobsd> this is my snapshot image
<naobsd> btw I don't have RK3188-T
<BorgCuba> 1) git clone ... 2) modification of clock.c 3) make ARCH=arm CROSS_COMPILE=arm-linux-gnueabihf- rk30xx 4) make bootable image 5) boot
<BorgCuba> naobsd, it hangs at the pll code
<BorgCuba> I encountered this before
<naobsd> it = non-modified code?
<BorgCuba> yes
<BorgCuba> I have some notes
<BorgCuba> for GPLL it hangs here: g_cruReg->CRU_MODE_CON = (0x3<<((pll_id*4) + 16)) | (0x0<<(pll_id*4));
<BorgCuba> for CPLL it stops here: rkclk_pll_wait_lock(pll_id);
<naobsd> all I know is it works on RR...
<BorgCuba> but I did not encounter the unaligned access exception with the code I used before
<BorgCuba> I think the clocks a relatively low speed, so the -t should be able to handle those
<naobsd> did you try my image?
<BorgCuba> your sd image?
<naobsd> yes
<BorgCuba> it uses the proprietary loader, doesnt it?
<BorgCuba> I think it works
<naobsd> [03:00] <@naobsd> this is my snapshot image
GriefNorth has joined #linux-rockchip
<naobsd> what I'm talking is about u-boot :(
<BorgCuba> okay, let me read the README ;-)
<BorgCuba> I guess I have to decode FlashBoot_uboot.img and the use the decrypted file instead of my u-boot.bin?
<naobsd> well?
<naobsd> all you need is some "dd" commands...
<naobsd> dd if=sd_header_64.enc of=/dev/XXX conv=sync,fsync seek=64
<naobsd> dd if=FlashData.bin of=/dev/XXX conv=sync,fsync seek=68
<naobsd> dd if=FlashBoot_uboot.bin of=/dev/XXX conv=sync,fsync seek=92
<BorgCuba> what I did 0...8000 zeros and then rk-makebootable ddrinit.bin FlashBoot_uboot.dec sd_uboot.tmp
<naobsd> what I asked is trying my image
<BorgCuba> you have a slightly newer ddr load version I think
<BorgCuba> Ill follow you dd instructions
<naobsd> I never tried rk-makebootable, probably your command will work, but I'm not talking about it
<BorgCuba> it works
<naobsd> ddr init blob in u-boot is newer a little. but probably another ddr init (in RK loader 2.10) worked too...
<naobsd> now you know it really works, good
<BorgCuba> same result, last line is "OUT" from the ddr code
<naobsd> ?
<naobsd> "it works"?
<BorgCuba> okay, by "it works" i meant that "rk-makebootable" works
<BorgCuba> sorry :-)
GriefNorth has quit [Remote host closed the connection]
<naobsd> you tried my image with dd, and it stops at "OUT"
<naobsd> right?
<BorgCuba> yes! thats right
<BorgCuba> must be the pll code
<naobsd> I see, now I understand my image doesn't work on some hardware
<BorgCuba> I once put lots of "putc" calls into uboot to find this out
<naobsd> "it hangs at pll code" = no output after "OUT"(without modification)
<naobsd> right?
<BorgCuba> yes
<BorgCuba> actually second and third call to rkclk_pll_clk_set_rate in rkclk_set_pll
<BorgCuba> thats GPLL and CPLL
<BorgCuba> I mentioned that earlier
<naobsd> I'll try u-boot on another RK3188
<BorgCuba> okay
<BorgCuba> and I guess you dont have this unaligned access issue, right?
<BorgCuba> I saw that there are specials memcpy versions like "ftl_memcpy" that explicitly check if the pointers are aligned
<naobsd> btw, what is your board?
<BorgCuba> beelink a9 /netxeon ...
<BorgCuba> H86_V20_20131116
<BorgCuba> ALC5631 (audio codec), RTL8152 (usb ethernet), fe1.1s (usb hub) AP6630 (wifi+bt) and this standard hdmi chip
<BorgCuba> act8846 pmu
<BorgCuba> no ap6330
bengal has joined #linux-rockchip
<naobsd> usb ether?
<BorgCuba> yes :-)
<BorgCuba> btw, the CRU part is not in the TRM 2.0, do you have any infos on this?
<BorgCuba> i.e. on the Clock and Reset Unit
<naobsd> no, I know only mainline code
<BorgCuba> btw, do you keep pages for different devices on the wiki?
<BorgCuba> like they have on linux-sunxi?
<BorgCuba> I saw you kind of started a "new wiki"
<BorgCuba> naobsd, its seems like you can boot a kernel using fast boot: "boot <kernel> [ <ramdisk> ] download and boot kernel" (
<BorgCuba> "[...] It is very useful while developing a kernel or modifying the rootfs. " :-)
<RaYmAn> sadly, while fastboot is "standard" it's far from all devices that support all features of it (in particular fastboot boot is often horribly broken, even on qualcomm devices)
<BorgCuba> In 10 minutes or so I can tell you if it works for rockchip uboot
<BorgCuba> hmm, fastboot doesnt seem to recognize the device (2207:310b)
<karlp> BorgCuba: the devices section of the wiki I want tobring back :)
<BorgCuba> I could contribute something for the "beelink a9" (rk3188-t)
<BorgCuba> and I have a rk3066 hdmi dongle
<BorgCuba> karlp, do you have a template?
<RaYmAn> try adding -i 0x2107
<BorgCuba> no, this doesnt work
<BorgCuba> maybe its not using fastboot but "rockusb" (although in rk30xx.h its configured with CONFIG_CMD_FASTBOOT) ->
<RaYmAn> ah,yeah. rockchip has their own thing too, but fastboot should also be there - just have to find ut
<karlp> BorgCuba: hah! not a chance, I'm only just starting with bits like this :)
<naobsd> you need to modify u-boot to go fastboot mode
<BorgCuba> karlp, maybe you can "steal" something from linux-sunxi for the template ;-)
<BorgCuba> anyway, I have to leave. I'll inform you if I found a way to easily upload/test kernels
<BorgCuba> bye
BorgCuba has quit [Quit: leaving]
<karlp> will have alook, they do have a better organized community it seems for that sort of stuff
<naobsd> I'll prepare wiki on new server by several reason
<naobsd> I only have admin rights on wiki, it only does things on wiki. no way to tweak things under wiki
<naobsd> e.g. update, configure, add extentions, etc
<naobsd> I have to sleep...
FreezingCold has joined #linux-rockchip
npcomp_ is now known as npcomp
ferric_ has quit [Ping timeout: 272 seconds]
nighty^ has quit [Quit: Disappears in a puff of smoke]
ferric has joined #linux-rockchip
FreezingAlt has joined #linux-rockchip
FreezingCold has quit [Ping timeout: 250 seconds]
FreezingCold has joined #linux-rockchip
ferric has quit [Ping timeout: 255 seconds]
FreezingAlt has quit [Ping timeout: 272 seconds]
niik has joined #linux-rockchip
niik has quit [Client Quit]
bengal has quit [Quit: Leaving]
ChiaSmurf has quit [*.net *.split]
ferric has joined #linux-rockchip