<naobsd>
JeffyChen: changing direction in rk_gpio_set_pin_level() shouldn't be done
<naobsd>
JeffyChen: I don't have R89 so currently I'm waiting confirmation from person who has it, but current code seems wrong
<JeffyChen>
sure
<naobsd_>
oops
naobsd_ has joined #linux-rockchip
_whitelogger___ has quit [Excess Flood]
naobsd has quit [Ping timeout: 246 seconds]
_whitelogger___ has joined #linux-rockchip
naobsd_ is now known as naobsd
<JeffyChen>
ok, he will check it, thank you~
<naobsd>
JeffyChen: thank you too :)
Avasz has quit [Ping timeout: 240 seconds]
<naobsd>
(btw I don't have RK3288 TRM, GPIO0_A6 reset state is guessed from info from netxeon)
<naobsd>
oh wait, reset state should be described in intro section. there might be info in datasheet pdf...
<naobsd>
^true
<naobsd>
R89 uses GPIO0_A6, default is pull-up
_whitelogger___ has joined #linux-rockchip
<naobsd>
Firefly (beta) uses GPIO0_A1, default is pull-down
<naobsd>
it matches behavior, reasonable.
_whitelogger___ has quit [Excess Flood]
_whitelogger___ has joined #linux-rockchip
<naobsd>
(not yet pushed to github, I'll fix it)
<naobsd>
btw, u-boot uses resource(.img), it should be available if next kernel is not Linux ;)
<JeffyChen>
yes, but i could not figure out a better way to do this...
Avasz has joined #linux-rockchip
_whitelogger___ has quit [Excess Flood]
_whitelogger___ has joined #linux-rockchip
_whitelogger___ has quit [Excess Flood]
_whitelogger___ has joined #linux-rockchip
<naobsd>
probably another kernel such as FreeBSD will have own dtb in somewhere because implementation is different, dts node structure will be different
<naobsd>
ah, it's same for mainline Linux ;)
GriefNorth has joined #linux-rockchip
<naobsd>
at least current u-boot want PWRHLD pin info from dts from 3.10 kernel tree
<naobsd>
^for ACT8848
<naobsd>
or keep power button on til kernel boot and act8846 driver handle it ;)
<JeffyChen>
yes, our kernel & uboot team had talk about uboot & dts, they thought it is ok for now, that uboot can still have a chance to get into rockusb w/o resource...
_whitelogger___ has quit [Excess Flood]
_whitelogger___ has joined #linux-rockchip
<naobsd>
sure
<naobsd>
I guess most board are designed that it just work with default state (w/o act8846 driver)
<JeffyChen>
yes, and we are still tring to figure out this.
<naobsd>
pwrhld gpio has no effect. act8846 linux driver use it for shutdown, but it doesn't power-off, so it needs patch e.g. set 0xc3 register in act8846(netxeon), toggle different gpio(hotack)
<naobsd>
hotack added 3rd gpio to act8846 driver named "pmic_poweroff_gpio"
<naobsd>
I don't know why they just use pmic_hold_gpio ;)
<naobsd>
^not use
<naobsd>
as far as I know, they don't open rk3288 kernel source yet
<naobsd>
ganbold_: if it needs patch for FreeBSD, please send it :)
_whitelogger___ has joined #linux-rockchip
_whitelogger___ has quit [Excess Flood]
_whitelogger___ has joined #linux-rockchip
_whitelogger___ has quit [Excess Flood]
_whitelogger___ has joined #linux-rockchip
_whitelogger___ has joined #linux-rockchip
Avasz has quit [Ping timeout: 250 seconds]
JeffyChen has quit [Ping timeout: 246 seconds]
_whitelogger___ has quit [Excess Flood]
_whitelogger___ has joined #linux-rockchip
Soopaman has quit [Quit: Leaving.]
_whitelogger___ has joined #linux-rockchip
_whitelogger___ has quit [Excess Flood]
_whitelogger___ has joined #linux-rockchip
_whitelogger___ has joined #linux-rockchip
_whitelogger___ has quit [Excess Flood]
_whitelogger___ has joined #linux-rockchip
_whitelogger___ has quit [Excess Flood]
_whitelogger___ has joined #linux-rockchip
_whitelogger___ has quit [Excess Flood]
_whitelogger___ has joined #linux-rockchip
_whitelogger___ has quit [Excess Flood]
_whitelogger___ has joined #linux-rockchip
hipboi has quit [Ping timeout: 272 seconds]
_whitelogger___ has joined #linux-rockchip
hipboi has joined #linux-rockchip
FreezingAlt has joined #linux-rockchip
FreezingCold has quit [Ping timeout: 260 seconds]
FreezingAlt is now known as FreezingCold
_whitelogger___ has joined #linux-rockchip
_whitelogger___ has joined #linux-rockchip
_whitelogger___ has joined #linux-rockchip
<rperier>
mmind00: don't compute something, how do you transform "GPIO3_D6/UART3_RTS" (schematic and kernel 3.0.8) to "RK_GPIO3" with pin 30 (mainline kernel) ?
<rperier>
naobsd: where did you get this information ?
_whitelogger___ has joined #linux-rockchip
<rperier>
naobsd: it's noted, thanks
akaizen_ is now known as akazien
anwardz has quit [Ping timeout: 244 seconds]
_whitelogger___ has quit [Excess Flood]
_whitelogger___ has joined #linux-rockchip
Soopaman has joined #linux-rockchip
_whitelogger___ has quit [Excess Flood]
_whitelogger___ has joined #linux-rockchip
_whitelogger___ has joined #linux-rockchip
akazien has joined #linux-rockchip
akazien has quit [Remote host closed the connection]
_whitelogger___ has quit [Excess Flood]
_whitelogger___ has joined #linux-rockchip
akazien has quit [Ping timeout: 260 seconds]
AstralixNB has quit [Read error: Connection reset by peer]
AstralixNB has joined #linux-rockchip
Soopaman has quit [Quit: Leaving.]
<AstralixNB>
naobsd, do you know how to handle Loader 2 and SELinux well?
Soopaman has joined #linux-rockchip
_whitelogger___ has quit [Excess Flood]
_whitelogger___ has joined #linux-rockchip
_whitelogger___ has quit [Excess Flood]
_whitelogger___ has joined #linux-rockchip
_whitelogger___ has joined #linux-rockchip
akazien has joined #linux-rockchip
<karlp>
rperier: I found it in one of the header files in the 3.0 source vs the 3.17 source, it wasn't obvious yo me either at first
FreezingCold has joined #linux-rockchip
_whitelogger___ has quit [Excess Flood]
_whitelogger___ has joined #linux-rockchip
_whitelogger___ has joined #linux-rockchip
_whitelogger___ has quit [Excess Flood]
_whitelogger___ has joined #linux-rockchip
_whitelogger___ has quit [Excess Flood]
_whitelogger___ has joined #linux-rockchip
_whitelogger___ has joined #linux-rockchip
RayFlower has joined #linux-rockchip
RayFlower has quit [Read error: Connection reset by peer]
RayFlower has joined #linux-rockchip
<naobsd>
AstralixNB: there is no RK Loader ver.2 <-> SELinux relation. some times ago some people asked me about it and I confirmed it doesn't exist.
_whitelogger___ has joined #linux-rockchip
<naobsd>
AstralixNB: and I'm very sure 2.10 works on RR and Hotack T010
<naobsd>
with Android
<AstralixNB>
naobsd, I am pretty sure it will work, I just try to find out why it doesn't at my desk... We have made lots of mixed cROMs. And if the original ROM used Loader 2 we kept loader 2 as these often do not boot with a gen1 loader.
<AstralixNB>
But our very own andoid, completely compiled on our machines does not start if loader 2 is involved nor does it with SELinux enabled on a loader 1
<AstralixNB>
So we have to stay at loader 1 and without SEL
<AstralixNB>
but that is a 2nd story
<AstralixNB>
And it looks like using multi-user in combination with NAND is killig performance completely
<karlp>
multi user android?
_whitelogger___ has quit [Excess Flood]
_whitelogger___ has joined #linux-rockchip
_whitelogger___ has joined #linux-rockchip
_whitelogger___ has joined #linux-rockchip
<naobsd>
it seems "SELinux" is used as a magic keyword for "because impossible"
<AstralixNB>
unfortunately it is more... I miss documentation and / or sourcecode to find out what I am missing
<AstralixNB>
no, SELinux is just a more secure linux
<naobsd>
no need to use "SELinux" name for this problem
<AstralixNB>
no, misunderstanding
<naobsd>
someone asked me to "how to fix SELinux issue?", my answer was "there is no issue"
<AstralixNB>
the guys only tried combinations
<AstralixNB>
but we not fully understand the interconnections
<AstralixNB>
so I would like to understand and to correct so it works
<naobsd>
I confirmed that both loader 1.x and 2.x can be used, I made working images
<naobsd>
no need to discuss about non-existing "SELinux" issue
<AstralixNB>
So may be you can help me to get my images to work with loader 2 ?
<JeffyChen>
if nand ko not match loader, then you will not see partition info in console log
<AstralixNB>
naobsd, nice to see that you have done this, but this doesn't heöp me to understand how to do it myself
<naobsd>
^please read my post which include link to CrewRKTablets_Odys_Iron_Omni_.4.4_v1.4.1_RR.zip
<naobsd>
I explained what I did
<AstralixNB>
It is hard to read as the talk is full of html code on my browser too
<naobsd>
replace bootloader and rknand.ko, that's all
<AstralixNB>
hmm... I need to do it myself, cause I told the guys to do exactly that...
<AstralixNB>
Probably they missed the module in the initramfs or boot.img...
<naobsd>
of course, please don't use random files. you need to properly pick working combination
<naobsd>
people assume "I did it properly" so they want some magic word to explain the problem
<AstralixNB>
I got a confirmed-to-be-working set of files directly from RK, I confirmed that set of files to work with radxa and linux
<AstralixNB>
So I'll try that set of files (loader and module) on my tablet tonight
<naobsd>
there is another but may be related issue, your kernel may not be compatible with some rknand.ko version
<naobsd>
SDK contains 2 kind of rknand.ko
<AstralixNB>
what to check to identify that?
<naobsd>
kernel had some change related to nand
bengal has joined #linux-rockchip
<AstralixNB>
ah... ahem... this trial and error thingy was the one causing you jelling on me...
<naobsd>
don't assume, _try and see_, this is my strategy
<naobsd>
try both, and use working one :)
<AstralixNB>
yelling?
_whitelogger___ has joined #linux-rockchip
<naobsd>
2set of rknand.ko (see update/)
<naobsd>
latest kk images for RR uses update/* one
<JeffyChen>
and kernel's version should be end with a '+', like '3.0.36+'...
<naobsd>
ah, it's another "relation between kernel and modules"
<naobsd>
but vermagic thing is not only for RK, very common issue :)
hipboi has quit [Ping timeout: 245 seconds]
<naobsd>
I have no direct support from RK or any vendor
<naobsd>
all information I know is result of "trial & error"
<karlp>
in that rk2918 trm/datasheet earlier, what on earth is the "8/16bit modem interface" ?!
_whitelogger___ has joined #linux-rockchip
<AstralixNB>
I am using 3.0.36+ kernels for 3188 only.
<AstralixNB>
I pinned the post of you naobsd and follow that on my own. Then we see if the tablet will work with gen2 loader.
anwardz has quit [Ping timeout: 260 seconds]
<naobsd>
AstralixNB: I remembered that some kernels on crew seems to be old
<naobsd>
AstralixNB: maybe they are come from stock rom
<AstralixNB>
naobsd, there are two kinds of images... some of them are only stacked, so took several existing images, ripped them apart, chenged things and put it together again.
<naobsd>
AstralixNB: I'm not sure 2.10 loader in SDK can be used for such kernel
<naobsd>
such _old_ kernel
<AstralixNB>
The other images (like Omni for Iron tablet) are completely compiled on our own machines.
<naobsd>
AstralixNB: rknand.ko outside of update/ may work, but I'm not sure
<naobsd>
what I confirmed is loader 1.x and 2.x can be exchanged on same Android image
<AstralixNB>
I told this to hte guys mixing up the images and they tried and failed every single time.
<naobsd>
I didn't confirm like "very old stock kernel supports latest rknand.ko"
<naobsd>
I know there is newer version, probably I tried 2.13 and it worked, but I don't use newer because it may have some problem
<naobsd>
I'm using set of files in same SDK
<naobsd>
it should work
<AstralixNB>
I switched from 2.08 to 2.16 and upgrade_tool version 1.21 as I have too many verify-errors writing large images with the older set of tools.
<naobsd>
I don't stop it but if your rom doesn't work, please try 2.10 loader and rknand.ko from same SDK
<AstralixNB>
With 1.16 and 2.08 or even 1.xx loaders I often had to totally erase the whole flash or write it 4 times to have zero verify errors. With 2.16 / 1.21 i only had this once or twice.
<AstralixNB>
sure
<naobsd>
you may know, switching between 1.x <> 2.x loader, you need to format whole nand
Soopaman has quit [Read error: Connection reset by peer]
<AstralixNB>
of course!
Soopaman has joined #linux-rockchip
<AstralixNB>
they have different FTL implementation
<naobsd>
ah, "no direct support from any vendor" was wrong, hipboi helped me a lot :)
<naobsd>
anyway, I always do "trial & error" even if someone said it's impossible.
_whitelogger___ has quit [Excess Flood]
_whitelogger___ has joined #linux-rockchip
bengal has quit [Quit: Leaving]
_whitelogger___ has joined #linux-rockchip
_whitelogger___ has quit [Excess Flood]
_whitelogger___ has joined #linux-rockchip
_whitelogger___ has quit [Excess Flood]
_whitelogger___ has joined #linux-rockchip
_whitelogger___ has quit [Excess Flood]
_whitelogger___ has joined #linux-rockchip
anwardz has joined #linux-rockchip
_whitelogger___ has quit [Excess Flood]
_whitelogger___ has joined #linux-rockchip
dlezcano has joined #linux-rockchip
_whitelogger___ has quit [Excess Flood]
_whitelogger___ has joined #linux-rockchip
<AstralixNB>
so back again
<AstralixNB>
naobsd, as a result of trial and error you mostl times onyl get a binary report... i.e. works / not works.
<AstralixNB>
but I love to understand why and how it works and what to care about if it doesn't
<naobsd>
AstralixNB: I don't object it, but it's better than people who assume "it must not work because (any non-logical/non-technical explanation here)" w/o try/confirm
<naobsd>
same for assuming "it must work because (xxx)" but actually it's technically/logically impossible.
<naobsd>
honestly speaking I surprised that you believed "because SELinux". you are expert, have a lot of experience, you will not believe something w/o technical/logical explanation and/or w/o confirmation
<naobsd>
I have basic knowledge about bootloader/kernel/SELinux. but I cannot explain "bootloader can do something to SELinux" "SELinux can do something to bootloader", so I cannot believe "there is relation between bootloader and SELinux"
<naobsd>
I think when 2.x loader was available, many people tried it and reported error around NAND
<naobsd>
so I can imagine "there is relation between loader and rknand.ko"
<naobsd>
and I tried loader/rknand.ko combination to confirm
Soopaman has joined #linux-rockchip
Soopaman has quit [Read error: Connection reset by peer]
<naobsd>
I think what I did is logical/technical thing... or, am I wrong?
<karlp>
naobsd: lots of things don't make sense when selinux is involved, and even if it turns out to have a logical explanation, given how many problems selinux causes, and how gross it is to try and debug it, the usual answer for _many_ people is, "turn off selinux"
<karlp>
it's just too big a pit of despair for many people to try and resolve.
_whitelogger___ has joined #linux-rockchip
<naobsd>
I know many people do "turn off selinux to solve problem", but it's Linux world story. bootloader doesn't live in Linux world.
<naobsd>
it sounds like "run rm /dev/mtd/mtdblock0, then your firmware is removed, your device will not boot anymore"
RayFlower has quit [Read error: Connection reset by peer]
RayFlower has joined #linux-rockchip
<naobsd>
I know there are many religious FAQ(?) on the net...
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
nighty^ has joined #linux-rockchip
dlezcano has quit [Ping timeout: 260 seconds]
_whitelogger___ has quit [Excess Flood]
_whitelogger___ has joined #linux-rockchip
_whitelogger___ has quit [Excess Flood]
_whitelogger___ has joined #linux-rockchip
RayFlower has joined #linux-rockchip
RayFlower has quit [Read error: Connection reset by peer]
_whitelogger___ has quit [Excess Flood]
_whitelogger___ has joined #linux-rockchip
_whitelogger___ has quit [Excess Flood]
_whitelogger___ has joined #linux-rockchip
RayFlower has joined #linux-rockchip
GriefNorth has quit [Quit: WeeChat 1.0.1]
_whitelogger___ has quit [Excess Flood]
_whitelogger___ has joined #linux-rockchip
_whitelogger___ has quit [Excess Flood]
_whitelogger___ has joined #linux-rockchip
_whitelogger___ has joined #linux-rockchip
benja` is now known as benja
Avasz has joined #linux-rockchip
Avasz has joined #linux-rockchip
naobsd has quit [Quit: Page closed]
_whitelogger___ has quit [Excess Flood]
_whitelogger___ has joined #linux-rockchip
_whitelogger___ has quit [Ping timeout: 260 seconds]
_whitelogger___ has joined #linux-rockchip
<AstralixNB>
naobsd, I never stated that it cannot work, I always stated that we did not have any success with it. I also stated that I want to know why.
<AstralixNB>
I never claimed that there is anything impossible or it cannot work cause I think it cannot work
<AstralixNB>
Normally things do not listen to what I think, but I listen to things that teach me something.
<AstralixNB>
But I think this discussion is onyl as we both aren't native english speaking people.
Avasz has quit [Quit: No Ping reply in 180 seconds.]
_whitelogger___ has quit [Excess Flood]
_whitelogger___ has joined #linux-rockchip
_whitelogger___ has quit [Excess Flood]
_whitelogger___ has joined #linux-rockchip
Soopaman has quit [Quit: Leaving.]
Avasz has joined #linux-rockchip
Avasz has quit [Changing host]
Avasz has joined #linux-rockchip
naobsd has joined #linux-rockchip
<naobsd>
sorry, I try to use words properly such as you, someone/some people, everyone/people, but I'm sure my words is not clear on many cases
_whitelogger___ has joined #linux-rockchip
<AstralixNB>
ah, never mind, I am pretty sure my english isn't perfect too.
<AstralixNB>
I do have a pretty precise idea of what might have happened, when we tried to use loader 2.xx
<AstralixNB>
I guess you need to have the matching kernel module in three places: initramfs, boot.img and rootfs. I guess that the first, if not the first two places where not updated with the matching kernel module.
<AstralixNB>
A second point of failure might be the fact, that images where mixed and a non matching module was taken to test
naobsd_ has joined #linux-rockchip
naobsd has quit [Ping timeout: 246 seconds]
<naobsd_>
well, disconnected. log is fine?
<naobsd_>
anyway
<naobsd_>
only 1 place is enough
<naobsd_>
only 1 module can be loaded
<naobsd_>
=about rknand.ko
<naobsd_>
^about rknand.ko
<naobsd_>
and for android
<naobsd_>
and for android on rk, boot.img is initramfs & rootfs
<naobsd_>
bootloader and rknand.ko need to be matched
<naobsd_>
kernel and rknand.ko need to be matched too, but this is another story
<naobsd_>
I can guess someone replaced only loader and got error
<naobsd_>
possibly, someone made linux envrionment on nand, there might be 2 rknand.ko, one in boot(initramfs) and another in linuxroot(root file system, e.g. /lib/modules/xxx)
<naobsd_>
and he do insmod in initramfs but try to replace only /lib/modules/xxx
Avasz has quit [Ping timeout: 260 seconds]
<naobsd_>
(putting rknand.ko on some partition(not boot/recovery) on NAND is useless ;)
_whitelogger___ has quit [Excess Flood]
_whitelogger___ has joined #linux-rockchip
_whitelogger___ has joined #linux-rockchip
_whitelogger___ has joined #linux-rockchip
naobsd_ has quit [Quit: Page closed]
_whitelogger____ has joined #linux-rockchip
<hramrach_>
karlp: but to boot you had to load rknand.ko to access the root filesystem so it's already loaded and in use so irreplacable
_whitelogger____ has joined #linux-rockchip
<karlp>
ah, thanks, that makes sense.
Soopaman has joined #linux-rockchip
RayFlower has quit [Read error: Connection reset by peer]
RayFlower has joined #linux-rockchip
_whitelogger____ has joined #linux-rockchip
_whitelogger____ has joined #linux-rockchip
_whitelogger____ has joined #linux-rockchip
_whitelogger____ has quit [Excess Flood]
_whitelogger____ has joined #linux-rockchip
_whitelogger has joined #linux-rockchip
Soopaman1 has joined #linux-rockchip
_whitelogger has joined #linux-rockchip
AstralixNB has quit [Ping timeout: 250 seconds]
naobsd has joined #linux-rockchip
<naobsd>
hm, log is useless...
bengal has joined #linux-rockchip
dlezcano has joined #linux-rockchip
anwardz has quit [Ping timeout: 258 seconds]
Soopaman has joined #linux-rockchip
Soopaman1 has quit [Ping timeout: 250 seconds]
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
Soopaman has quit [Quit: Leaving.]
benja is now known as benja`
JohnDoe_71Rus has joined #linux-rockchip
JohnDoe_71Rus has joined #linux-rockchip
RayFlower has quit [Read error: Connection reset by peer]
RayFlower has joined #linux-rockchip
RayFlower has quit [Read error: Connection reset by peer]