ChanServ changed the topic of #linux-rockchip to: Rockchip development discussion | Wiki at http://linux-rockchip.info | Logs at http://irclog.whitequark.org/linux-rockchip | ML at http://groups.google.com/group/linux-rockchip
FreezingCold has quit [Ping timeout: 260 seconds]
FreezingCold has joined #linux-rockchip
Tony_ has joined #linux-rockchip
<Tony_> Good morning.
<Tony_> I can solder too. see:http://img.blog.csdn.net/20141027092236450
FergusL has quit [Quit: http://quassel-irc.org - Chat comfortably. Anywhere.]
mrcan_ has joined #linux-rockchip
<Tony_> naobsd, are you here?
mrcan has quit [Ping timeout: 258 seconds]
Bludot has quit [Quit: Connection closed for inactivity]
<Tony_> Anybody here ?
naobsd has joined #linux-rockchip
<naobsd> Tony_: from dmesg, you're using rknand.ko. please try to disable(not to load) it and re-enable mmc0 driver in kernel
<Tony_> naobsd, thank you. but I barely find any common about rknand.ko and mmc1.
<naobsd> Tony_: do you know any common about rknand.ko and mmc0?
<Tony_> Nope.
<naobsd> then why you disabled mmc0?
<naobsd> well
Tony_ has quit [Ping timeout: 246 seconds]
Tony_ has joined #linux-rockchip
<Tony_> rk's doc told me.
<naobsd> hm
<Tony_> It told me that I have to disable mmc0 if I want to run system from sdcard.
<Tony_> <SD Card Boot User Guide >
<naobsd> if you cannot try someone's suggestion because of undocumented, why you ask things on here?
<naobsd> do you know "trial and error"?
<Tony_> okay, okay, I'm so sorry.
<Tony_> how to disable rknand.ko ? just del it ?
<naobsd> I don't know about your system, if you can see rknand.ko in somewhere, rename or remove will disable it.
<Tony_> RK3188_ANDROID4.2.2-SDK_V1.00_20130425
<Tony_> that is my system sdk.
<Tony_> from rk.
<naobsd> then rknand.ko will be in ramdisk in boot image/partition
<naobsd> but I think it needs more change to run "mmc0 enabled and rknand.ko disable" system because
<naobsd> partitions such as system may be available differently because of mmc0 and rknand.ko difference
<naobsd> well, it _may_ need,
<Tony_> rmmod rk30xxnand_ko
<Tony_> how about "rmmmod rk30xxnand_ko"
<naobsd> no, it's "unload module from kernel". what I said is "don't load it"
<naobsd> do you know how to modify ramdisk image?
<Tony_> Maybe. I know rk3188 for one years.
<Tony_> build it very day. :P
<naobsd> then please modify ramdisk image, it should have rknand.ko and insmod line in *.rc in ramdisk
<Tony_> I think that will be fine. just del android_tree/out/target/product/rk30sdk/root/rk30xxnand_ko.ko.3.0.36+
<Tony_> do you think so?
<naobsd> yes
FergusL has joined #linux-rockchip
<Tony_> Hi, naobsd.
<Tony_> It hang up at spheniscidae logo.
<naobsd> any log?
<naobsd> I guess partitions cannot be mounted
<naobsd> ah, mmc0 is enabled, right?
<naobsd> your sd card should be detected by mmc0 driver
<naobsd> and if your kernel doesn't support parttitions defined in CMDLINE in parameter on sd card,
<Tony_> wait for while. :)
<naobsd> you need to mount partitions (system, data, etc) differently
<naobsd> mount lines in init.*.rc should be configured for rknand.ko
<naobsd> use MBR partitions (mmcblkXpY), or, add RK partition support to rk mmc driver(related code should be already available to support emmc based system)
<Tony_> I will do it again and again. to prevent other case.
<Tony_> okay, this time, I'm sure that mmc0 is enabled and mv rk30xxnand_ko.ko.3.0.36+ to rk30xxnand_ko.ko.3.0.36+.bak.
<Tony_> there is boot log:https://gist.github.com/kangear/0755731cec3368e25d6f
<Tony_> naobsd,
<Tony_> naobsd, maybe you have to enjoy your lunch now.
<naobsd> Tony_: strange, no mmc(sd card) is detected...
<naobsd> kernel configuration should be ok (there are both mmc0 and mmc1 lines in log)
<naobsd> did you pull & re-insert sd card?
<naobsd> ah, you should do it to update kernel/ramdisk...
<naobsd> hmm...
<Tony_> I think ramdisk in the boot.img
<Tony_> wrong ?
<naobsd> ramdisk is in boot.img
<Tony_> rebuild the boot.img and update it.
<Tony_> That is what I done.
<naobsd> I see, and you pulled SD card to write net boot.img, right?
<Tony_> oh, sorry.
<Tony_> Iboot from main flash.
<Tony_> not sdcard.
<Tony_> I will do what you said.
<Tony_> there is log:https://drive.google.com/file/d/0B6NGoH-iSqrKSFZvUlRLVEIxSUk/view
<Tony_> naobsd,
<naobsd> Tony_: it seems it's "recovery" mode, probably your boot partition on sd card is broken.
cedinu has joined #linux-rockchip
<cedinu> Tony_: [ 2.754763] gp_rknand_info = : 0xc24ef100 you haven't disabled rknand
<naobsd> cedinu: well, it's in recovery mode, what he tried is to update boot.img, not recovery.img
<cedinu> naobsd: this is before entering recovery, it goes recovery when didn't found partition
<naobsd> Tony_: please rewrite boot.img(mmc0 enabled and rknand.ko disabled) to boot partition on SD card properly
<cedinu> naobsd: [ 2.829294] SdmmcInit ret = fffffff0! after detecting rknand
<cedinu> Tony_: naobsd is right
<cedinu> naobsd: do you know if anyone made some progress in booting 3188 from nand with a more recent kernel?
<naobsd> cedinu: please explain "recent kernel" more specific...
<cedinu> naobsd: ex. 3.10 or newer
<naobsd> cedinu: and what you asked is "_using_ NAND with 3.10 or newer kernel", not "boot(load) kernel from NAND", right?
<cedinu> naobsd: if _using_, then booting is easy, i think
<cedinu> naobsd: ok, using
<naobsd> cedinu: as far as I know, no one succeeded to use NAND with 3.10 or newer kernel
<cedinu> naobsd: yet, the code exist for nand controller, i suspect that the problem is in ftl, am i right?
Astralix1 has joined #linux-rockchip
<naobsd> I tried some with 3.10, NAND can be recognized but I couldn't make it work (I didn't have enough time to continue that work, I'm not sure it's really impossible)
<naobsd> cedinu: where is NAND code for 3.10 or newer?
AstralixNB1 has joined #linux-rockchip
<naobsd> cedinu: it's just a glue code
AstralixNB has quit [Ping timeout: 250 seconds]
Astralix has quit [Ping timeout: 250 seconds]
<cedinu> naobsd: ah, ok
<naobsd> cedinu: it just call functions in rknand.ko
<naobsd> there is rknand.ko for 3.10, I think it's possible to use NAND on rk3188 with 3.10, but I cannot make stable 3.10 for rk3188
<cedinu> naobsd: it may be possible to make a rknand.ko from uu encoded blobs found on other rk chips? (just a thought...)
<naobsd> I don't have enough time and "using NAND" priority is low for me
<cedinu> yes, i saw that you have started 3288
<cedinu> but 3188 is still a fine platform, i think that i will try some things
<naobsd> I don't say I'm not working for rk3188...
<naobsd> :(
<cedinu> i have an minix neo x7 with serial console, it should be easier than first board that i bought, with no serial
<cedinu> yes, true, i understand, but nand is so provocative :))
<cedinu> i mena it's there, i should use it
<cedinu> *mena
<cedinu> *mean
<cedinu> ok, i'll let you now, should try some things myself
GriefNorth has joined #linux-rockchip
<Tony_> naobsd, cedinu
<Tony_> run from sdcard will boot from recovery at first.
<naobsd> Tony_: it's not normal, or misc partition on sd card should have command "boot to recovery"
<cedinu> Tony_: it shouldn't, i've booted from sdcard with success
<naobsd> Tony_: anyway, what we're talking is your new boot.img, not recovery
<naobsd> Tony_: https://drive.google.com/file/d/0B6NGoH-iSqrKSFZvUlRLVEIxSUk/view this is not what we're talking
<Tony_> yes, but it will boot from recovery not booting.
<Tony_> You told me that write boot.ing into sdcard.
<Tony_> And I did.
<cedinu> Tony_: have you tried naobsd sdcard image?
<Tony_> no, Everything I used is that from rk.
<naobsd> "run from sdcard will boot from recovery at first." is not right. if it happens, your boot image/partition is broken or there is "boot to recovery" command in misc partition
<cedinu> naobsd: http://androtab.info/radxa_rock/sdboot/ is up to date?
<naobsd> Tony_: you know how to make bootable SD card, and I asked to do it with new boot.img. I didn't ask to do something special.
<Tony_> make sdcard boot, I used SD_Fireware_Tool.
<Tony_> make bootable SD card, I used SD_Fireware_Tool which from rk.
<Tony_> partition will be broken very time.
<Tony_> *every time
<naobsd> Tony_: I know that tool, and if you used misc.img with "boot to recovery" command, then first boot will be recovery
<naobsd> Tony_: but it's _not_ "boot from SD _always_ go to recovery", if happens only when you used misc.img with "boot to recovery"
<Tony_> okay, clear.
<naobsd> if happens -> it happens
<Tony_> actually, I know a little about misc.img how to be created.
<Tony_> so, I thought so.
<cedinu> you should put update.img, not misc.img
<naobsd> talking about that recovery log, you removed rknand.ko from root/ (ramdisk), it is also removed from recovery ramdisk
<naobsd> cedinu: no, what we're talking is which misc.img is used for making update.img. he uses update.img to make bootable SD
<cedinu> sorry, wasn't pay attention, you are right
<Tony_> naobsd, yes, it is. I used update.img to make bootable SD.
<naobsd> Tony_: actual command in misc.img is "boot to recovery and wipe(format) partitions", format is need to boot Android properly
<naobsd> Tony_: probably you can use stock recovery.img
<naobsd> stock -> non-modified
<Tony_> Yeah, got it.
<naobsd> with rknand.ko and mmc0 disabled kernel
<Tony_> I will do that.
<naobsd> (of course you can modify recovery for "mmc0 enabled and rknand.ko disabled" but current goal is sdio wifi, not fully-functional "mmc0 enabled and rknank disabled" system)
<naobsd> btw
<Tony_> Agreed. I will use stock recovery(about mmc0 and rknand.ko)
<naobsd> writing (only) boot.img to sdcard without SD_Firmware_Tool is possible, but you need to know proper location (sector #)
<naobsd> 0x00008000@0x00008000(boot) in this case, you have to write boot.img at (0x2000+0x8000)th sector
<naobsd> cedinu: sorry, about sdboot page on my site, I have newer files, but files on that page works too
<naobsd> cedinu: please tell me what you need
<Tony_> I do NOT know anything without SD_Firmware_Tool.
<naobsd> Tony_: http://androtab.info/radxa_rock/sdboot/ here is another example
<naobsd> I'll rewrite that page (soon if possible, but probably not so soon...)
<cedinu> naobsd: first i want to try some of my ideas then i will bother you. thanks for asking. it's premature to have a talk
<naobsd> sdboot_rk3188.img on that page doesn't support u-boot and some explanation should be rewritten...
<cedinu> btw, i have booted succesfully with that sdboot
<naobsd> http://files.androtab.info/radxa/sdboot_rk3188.zip this includes some newer file but it's just a snapshot at some time, I forgot exact version, less document, etc
<cedinu> it was about a month ago
<naobsd> I said it worked ;)
<naobsd> it -> file on that page
<cedinu> yes, true, i was asking to reccomend to Tony
<naobsd> I see
<cedinu> first try something that surely boots, then some steps to boot in his own configuration
<naobsd> cedinu: he already got bootable sd image/card. his problem is sdio wifi
<naobsd> cedinu: and I don't have any experience/answer about sdio wifi problem it happens only when sd boot
<cedinu> i have booted on my board with modifications (ap6330, usb ethernet, sdcard) and worked
<naobsd> I guess you're not using rknand.ko for sd boot, right?
<cedinu> modified sdk from minix, which is almost rk
<cedinu> no, sure not
<cedinu> removed both 3.0.06 and 3.0.8 modules
<cedinu> 3.0.36
<cedinu> not sure if i still have the config... long time :))
<naobsd> I guessed rknand.ko also has problem with mmc1(sdio wifi) in addition to mmc0(sdcard), and probably your system is the answer
<naobsd> not to use rknand.ko :)
<cedinu> wifi is sdio on my board
<naobsd> yes, ap6330 is sdio wifi
<cedinu> that's why i have removed altogether the nand modules
<naobsd> did you tweak some lines to mount partitions on sd card in init.rc?
<cedinu> it seems that when nand modules are present, they hook themselves to mmc (don't know why, saw on Tony log)
<cedinu> yes, i remember that i mount them by name
<naobsd> like mmcblk0pX?
<cedinu> of course, partitions must have names...
<cedinu> no, user, kernel...
<naobsd> hmm
<cedinu> but it should mount like mmcblk also
<cedinu> don't see why not
<naobsd> then I think no need to tweak lines for mount... it should work as like as NAND...
<naobsd> well
<naobsd> RK mmc driver supports emmc system too,
<cedinu> i have done my sdcard by hand, so partitions had names
<naobsd> on that system, partition name such as system, user, etc (defined in parameter) can be used
<cedinu> never used a tool like sd_fw_tool...
<naobsd> well, e.g. standard name for NAND is mtdblockX (X = partition number) and for eMMC/SD is mmcblkXpY (X=card#, Y=part#)
<cedinu> yes, i know
<naobsd> RK add support partitioning by parameter for NAND and eMMC/SD, partitions can be accessed by names
<cedinu> it's safer for Tony to modify init.*.rc
<naobsd> ^for NAND and eMMC(and maybe SD)
<cedinu> to put dev names
<cedinu> instead part names
<naobsd> probably partition name such as "system" will not work when rknand.ko disabled and mmc0 enabled, so init.rc need to be modified
<cedinu> if sd partitions don't have names, it won't boot because the init can't know which is which
<naobsd> ^it depends how kernel/driver is coded. some kernel may support partition names for SD too
<naobsd> cedinu: yes
<cedinu> mount is in init, so if init can't read a name, it's better to use dev names
<naobsd> rknand.ko supports SD card (works as SD card driver) when "boot from SD", so it conflicts in-kernel mmc driver
<cedinu> i don't know if that tool writes part names
<cedinu> true that, from what i saw in that log
<naobsd> sd boot with rknand.ko can use partition name
<naobsd> so same init.rc can be used w/o modification
<naobsd> well, that tool make same bootable SD card. sdboot info on my page was dumped from sd card made by that tool
<cedinu> then Tony's result is weird...
<naobsd> nothing is weird
<naobsd> he modified ramdisk for both boot/recovery, but only kernel for boot is modified
<naobsd> kernel for recovery is not updated
<naobsd> not matched to modified ramdisk
<cedinu> can't really say; it means i was lucky with my procedure
<naobsd> hehe :)
<cedinu> built from sdk after modifications
<cedinu> not trying to modify directly ramdisk
<cedinu> anyway, i don't like android too much, my goal is to run linux on the board
<naobsd> docs in sdk says "disable mmc0 driver (=use rknand.ko)" ;)
<cedinu> so, good that i read only what i want ;)
<naobsd> :)
<naobsd> (oh, I misunderstood recovery log a little, sorry. but anyway what we wanted to update/try is boot, not recovery. there is no reason to use more time to investigate that log...)
<cedinu> i have an rk30xxnand_ko.ko.3.10.0; i will try today some things with it
<Tony_> That so difficult for me.
<Tony_> I forgot disable mmc0 at kernel from recovery.
<cedinu> At least you know now how to make it work.
<naobsd> Tony_: cedinu has working example "boot from SD and sdio wifi", he doesn't use rknand.ko, you should try it
<naobsd> I don't know any other way to do "boot from SD and sdio wifi"
<cedinu> naobsd: it it works for him WITH rknand, let him do that way; I am curious, too
<cedinu> *if it works
<cedinu> sorry, it's 8 am here and i haven't sleep all night
<Tony_> okay, let me cool.
<naobsd> there might be a way to do it, many things are not open by RK...
<naobsd> cedinu: oops, sorry
<cedinu> Tony_: first try with rknand in both krn and recovery
<cedinu> Tony_: if it doesn't work, try without, also in both
<cedinu> naobsd: it seems that i must extract binary blobs for functions
<Tony_> I will do it again. boot.img(diable mmc0 + without rknand) + recovery(disable mmc0 + rknand)
<cedinu> naobsd: a part are already in the sources that i have pointed
<cedinu> Tony_: my opinion is to try similar krn and recovery, not mixed
<naobsd> cedinu: these nand related code in kernel is already used, it's needed to use rknand.ko
<Tony_> but use recovery(disable mmc0 + without rknand) can't work fine.
<cedinu> naobsd: in this module are functions defined as external in source code from rk
<naobsd> cedinu: Tony_: normal boot and recovery is totally independent, no need to try something for both...
<cedinu> naobsd: module also contains ftl
<naobsd> cedinu: rknand.ko has function, kernel (open source) part uses it
<naobsd> cedinu: but I might be wrong, please try
<cedinu> naobsd: functions in this rknand overlaps with functions defoned in source
<Tony_> I will try.
<cedinu> naobsd: i might need only the functions defined as external and their dependencies.
<cedinu> naobsd: i should try and tell the results
<cedinu> now i'm going to sleep for few hours...
<Tony_> good night for cedinu.
<Tony_> Hi naobsd.
<Tony_> System is hang up too. only kernel logo.
<Tony_> naobsd
<naobsd> Tony_: strange, sd card is not detected...
<Tony_> sdcard detected ?
<Tony_> No, It can't be.
<Tony_> disable mmc0, mmc0 is sdcard. That's why is not detectd.
<Tony_> naobsd,
<naobsd> Tony_: what I asked is "mmc0 enabled and rknand.ko disabled", i.e. default setting but rknand.ko disabled
<Tony_> egg on my face.
<Tony_> this time:https://drive.google.com/file/d/0B6NGoH-iSqrKVGpQV3FsNUhvZ0k/view?usp=sharing
<Tony_> naobsd, this time, I‘m sure that enable mmc0 and remove rknand.
<naobsd> sd card is detected, fine
<naobsd> you need to update MBR partition by fdisk or something like that. it seems there is 1 partition but probably it's not proper partition
<Tony_> naobsd, but It's hang up.
<Tony_> Use your tools ?
<naobsd> you have to make partition carefully. offset/size need to be 100% matched to partitions described in parameter
<naobsd> no, I'm just talking general thing, do you know fdisk command?
<naobsd> you need to make partitions (at least) for /system, /data, /cache, and internal sd
<Tony_> It's seem so difficult for me. I haven't do that.
<naobsd> hmm
<Tony_> fdisk I knew.
<naobsd> well
<naobsd> for example
<naobsd> 0x00040000@0x00040000(cache) this is /cache
<Tony_> I just know a little, and "dd".
<naobsd> when you used rknand.ko, partition is created automatically
<Tony_> oh...
<naobsd> but without rknand.ko, you need to make partition (information) yourself
<naobsd> partition info in parameter is ignored
<naobsd> you need to make partition which start at 0x2000+0x40000=0x42000 and size 0x40000 (unit is sector)
<naobsd> ^for /cache
<Tony_> I thought that SD_Fireware_Tool make partition.
<naobsd> to run Android, /system, /data, /cache are required. and RK Android also uses "user" partition as internal SD card, it is also required
<naobsd> on RK device, partition information is just a text in parameter
<Tony_> Can you give me a full doc about how to add partition by hand.
<naobsd> rknand.ko read it and prepare partition
<naobsd> well
<Tony_> Yeah, rk never tell that to me.
<naobsd> it's same as "make partition on SD card/USB HDD/etc on Linux"
<naobsd> if you can modify kernel, you can change mmc driver to read/use parameter instead of "making partition by hand"
<Tony_> Did you do that before ?
<Tony_> :P
<naobsd> actually there should be a code for eMMC based device
<naobsd> well, I did what?
<Tony_> rknand.ko will make partition every booting time ?
<Tony_> for [well, I did what?], modify kernel, change mmc driver to read/use parameter instead of "making partition by hand".
<naobsd> it makes _partition information_, making partition information doesn't do anything for storage
<naobsd> "making partition information" == read/parse information and make it usable
arokux2 has joined #linux-rockchip
<naobsd> it doesn't change anything
<naobsd> no I didn't
<naobsd> "making partition information" is not "formatting SD card"
<Tony_> Yeah.
<naobsd> anyway
<Tony_> mtd of cmdline in the parameter is wrong.
<Tony_> Do you think so ?
<naobsd> fdisk command can make(add) partition information on SD card. you may use any tool (GUI/CLI) which can handle MBR partition
<Tony_> who make partition when use rknand.ko /
<Tony_> ?
<naobsd> Tony_: well? if you use correct parameter file, mtdparts info in parameter is correct. if you use wrong parameter file, mtdparts info in parameter is wrong
<Tony_> you said that rknand.ko just making partition information.
<naobsd> Tony_: you define partition in parameter. rknand.ko just use it
<naobsd> I said "making partition information" == "reading/parsing parameter"
<naobsd> rknand.ko does it, mmc0 driver doesn't do it
<Tony_> I just want to know who make the partition at the first time booting.
<naobsd> you flashed parameter(included in update.img), then your device has that info
<naobsd> well
<naobsd> partition is logical thing
<naobsd> it exists it, AND it doesn't exist
<Tony_> than
<naobsd> in other words
<Tony_> Maybe I should read wikimedia about partition. :P
<naobsd> rknand.ko read/use partition info in parameter, then partitions exist
<naobsd> if no one read/use partition information, then partition doesn't exist
<Tony_> I mean your system should be in a correct position.
<Tony_> system etc
<naobsd> sd tool should put images (system etc) to the place specified by parameter
<naobsd> rknand uses parameter(same info) so partitions should be in correct position
<Tony_> after use SD_Fireware_Tool, the system/data exist a correct position. right ?
<Tony_> Yes, clear.
<Tony_> Now, how to tell the partition to the kernel or os without rknand.ko ?
<naobsd> use MBR parittion/fdisk, or modify mmc driver
<naobsd> there are many ways
<Tony_> I select the first one.
<Tony_> I think that the first one is normal way.
<Tony_> but the second way is not!
<naobsd> there are other ways
<Tony_> Why SD_Fireware_Tool write the partition info to the MBR at the same time ?
<naobsd> it doesn't write MBR
<Tony_> *Why not
<naobsd> RK kernel(rknand.ko) only use parameter for partition information
<naobsd> please think about NAND
<Tony_> Why don't SD_Fireware_Tool write the partition info to the MBR at the same time ?
<naobsd> MBR is not common in outside of PC world
<naobsd> you may use GPT if RK 3.0 kernel understand it. there are several ways to prepare partition info on storage
arokux2 has quit [Remote host closed the connection]
<naobsd> or you can use losetup to make partition information without MBR or any other partition info... but you need some commands before system want to mount partitions
<Tony_> Which way did you use ?
<naobsd> MBR/fdisk
<Tony_> I want to use the way that you used before.
<naobsd> please see fdisk man page on Linux
<Tony_> Dose kernel can read MBR ?
<Tony_> right ?
akaizen has quit [Remote host closed the connection]
akaizen has joined #linux-rockchip
<naobsd> MBR should be supported by RK 3.0 default configuration
<Tony_> Let me check it up.
<Tony_> my defconfig file has no MBR key word.
<naobsd> it's required to support commonly used SD card
<naobsd> and USB HDD
<naobsd> see your dmesg
<naobsd> [ 1.623508] msdos_partition..530... ==== The sdcard has MBR. [mmc0]
<naobsd> [ 1.623621] mmcblk0: p1
<naobsd> your sd card already has 1 partition (probably default one, incorrect for RK bootable SD), and detected
<Tony_> I see.
dlezcano has joined #linux-rockchip
<Tony_> Can you tell me why rk required disable mmc0 when i want to boot from sdcard.
akaizen has quit [Ping timeout: 272 seconds]
<naobsd> rknand.ko act as sdcard driver
<naobsd> run/use 2 drivers for 1 device (sd card controller, mmc0) is impossible
<naobsd> and it will be the reason that mmc1 (sdio wifi) cannot run properly
<Tony_> significant clear than before. you are so sweet. rk never tell me why do this or that. let me just do it.
<naobsd> well, do you have communication with RK?
<naobsd> (I don't)
<naobsd> you make RK based product?
<rperier> hi all
<Tony_> Hi rperier.
<naobsd> hi
<Tony_> Yes, I do communication with RK.
<Tony_> And if you have some thing need translate from Chinese, I can help you. :P
<naobsd> sounds good even if their explanation is not clear...
<naobsd> I only read english part such as code, command ;)
<naobsd> all I can do is guessing, trial & error...
<naobsd> this is the reason that I don't follow RK docs
<naobsd> I don't have proper information :(
<Tony_> ;P
akaizen has joined #linux-rockchip
<Tony_> I have question still.
<Tony_> mmc0 and mmc1 use the same one driver ?
<naobsd> same driver
<Tony_> I just remove mmc0's driver.
<Tony_> yeah, clear.
<naobsd> but different units, so there is kernel config for enable/disable mmc0 and mmc1 independently
<naobsd> (and mmc2 for some SoC)
<naobsd> probably rknand.ko does something to mmc1 too
<Tony_> one more question.
<Tony_> boot from main flash (not sdcard), rknand.ko is also exsit.
<Tony_> and Why?
<naobsd> there is no NAND driver in kernel, rknand.ko is required to use NAND (system, data, cache, etc)
dlezcano has quit [Ping timeout: 244 seconds]
<Tony_> I want to say, My main flash is not NAND.
irsol has quit [Ping timeout: 245 seconds]
<naobsd> eMMC?
<Tony_> Yes.
<Tony_> same thing ?
<naobsd> rknand.ko can be used as eMMC/SD driver
<naobsd> but you can disable rknand.ko
<naobsd> but you can disable rknand.ko and use mmc driver in kernel
<naobsd> I think ASUS MeMO Pad8 (eMMC) doesn't use rknand
<naobsd> and my RK3288 box and firefly doesn't use rknand too... (if my memory is correct)
irsol has joined #linux-rockchip
<naobsd> ^they are eMMC based
<Tony_> Why rknand.ko can't work fine when booting from sdcard ?
<naobsd> because mmc driver doesn't use "partition information" in parameter... but
<naobsd> but actually there should be "partition info in parameter" support code in mmc driver
<naobsd> because eMMC based device uses it (eMMC based devices supports "partition info in parameter")
<naobsd> I need to check kernel source, but I guess it's enabled only when eMMC is used
<Tony_> Nope, that's anther question.
<Tony_> If I just disable CONFIG_SDMMC0_RK29, It want boot well from sdcard.
<naobsd> it may be easy to share support code for sd boot
akaizen has quit [Remote host closed the connection]
akaizen has joined #linux-rockchip
<Tony_> The only qustion is that the sdio wifi can work fine.
<Tony_> can't
<Tony_> *can't
<naobsd> rknand.ko and mmc driver cannot enable/use same time
<Tony_> I'm confused about now. boot from eMMC.
<naobsd> you can ignore "share support code" if you can prepare partitions by hand
<Tony_> Why everything is work fine?
<naobsd> again
<naobsd> rknand.ko can act as eMMC/SD driver
mrcan has joined #linux-rockchip
<naobsd> you may use rknand.ko
<naobsd> and
<naobsd> mmc driver can be used for eMMC/SD driver
<naobsd> but you cannot use both at same time
<naobsd> I never said eMMC based device uses both
mrcan_ has quit [Ping timeout: 250 seconds]
<naobsd> "disable mmc(0) and enable(keep) rknand.ko" == rknand.ko is used
<naobsd> "enable mmc(0) and disable rknand.ko" == mmc driver is used
akaizen has quit [Ping timeout: 245 seconds]
<naobsd> (actually I never tried rknand.ko for eMMC, but I cannot find reason that it's impossible)
<Tony_> rknand.ko is only used for eMMC when booting from eMMC. right ?
<Tony_> rknand.ko is only used for SD when booting from SD.
<naobsd> rknand.ko is not for _booting_, it's kernel module, it's for mounting file system on flash storage on Linux kernel
<Tony_> I know.
<naobsd> rknand.ko is just a storage driver in Linux. same as mmc driver.
<naobsd> if you used "booting" == "mounting file systems", well
<Tony_> yeah.
<naobsd> rknand.ko can be used for eMMC to mount file systems on eMMC
<naobsd> err
<naobsd> rknand.ko can be used to mount file systems on eMMC when booting from eMMC
<naobsd> rknand.ko can be used to mount file systems on SD when booting from SD
<naobsd> rknand.ko can be used to mount file systems on NAND when booting from NAND
<naobsd> and
<naobsd> mmc driver can be used to mount file systems on eMMC when booting from eMMC
<naobsd> mmc driver can be used to mount file systems on SD when booting from SD but it seems "support partition in parameter" is missing
<naobsd> and
<naobsd> mmc driver can be used to mount file systems on SD when booting from NAND but not booting from eMMC/SD
<naobsd> rknand.ko cannot be used to mount file systems on NAND when booting from SD/eMMC because it tries to act as SD/eMMC driver and break mmc driver
<naobsd> clear?
<Tony_> i'm thinking for a while.
<naobsd> it seems rknand.ko knows "where is boot from"
<naobsd> ^no idea how
akaizen has joined #linux-rockchip
cristian_c has joined #linux-rockchip
<Tony_> mmc driver is mean mmc controller right?
<cristian_c> naobsd, ok, the recovery.fstab restore works
<cristian_c> I can browse into internal memory (sdcard)
<Tony_> naobsd, mmc driver is mean mmc controller driver right?
<cristian_c> and this directory is not empty via adb
<cristian_c> I cal also browse /system via adb after mounting it via cwm option
<naobsd> mmc driver is CONFIG_SDMMC_RK29
<naobsd> mmc controller driver
<naobsd> cristian_c: everything works fine now?
<cristian_c> naobsd, it soeems so
<cristian_c> *seems
<naobsd> cristian_c: congrat
<naobsd> cristian_c: btw, my cwm for rk cannot be used to install kitkat based update.zip and RK official(built from RK SDK) update.zip because it's based on cm-10.x cwm
<naobsd> cristian_c: cwm for rk should be updated to cm-11 based and rk recovery based, but I have no time for now
<cristian_c> ok
<cristian_c> naobsd, I need only to boot one linux distro via external sd
<naobsd> backup/restore and installing basic zip should work
<naobsd> I see
<cristian_c> I was told that I need u-boot, but it seems I've already got u-boot internally in the device
<cristian_c> naobsd, ok
<naobsd> u-boot internally?
<cristian_c> so, I was told I need to configure the existing u-boot to support micro-sd boot
<cristian_c> naobsd, maybe the guy has given me a wrong tip
<naobsd> cristian_c: which bootloader are you using? RK proprietary bootloader can be used for sd boot
<naobsd> cristian_c: I see.
<cristian_c> naobsd, for security reasons I'd like to backup the partitions/pieces that can be lost
<cristian_c> naobsd, ah, ok, I must see
<naobsd> well
<naobsd> cristian_c: ah, well, your device is rk3188?
ganbold_ has quit [Remote host closed the connection]
<cristian_c> naobsd, does it exist a method to know what bootloader I'm using?
<cristian_c> naobsd, soc is rk3188, yes
<naobsd> cristian_c: see serial console output, or guessing from dmesg
<cristian_c> I think I've to dissemble the board to see serial console
<cristian_c> ah, ok, then I can see the dmesg
<naobsd> ah
<naobsd> please tell me exact name of RKxxxLoader.bin in your update.img
<cristian_c> ok
<naobsd> it should tell loader version clearly...
<naobsd> rk3188 based product with u-boot is very very rare
ganbold_ has joined #linux-rockchip
<naobsd> and you can put any loader on SD card for sd boot, loader on NAND is not used for sd boot
<cristian_c> update.img?
<cristian_c> naobsd, ah, ok
<naobsd> update.img == firmware for your device
<cristian_c> ok
<cristian_c> naobsd, I didn't know
<naobsd> http://androtab.info/radxa_rock/sdboot/ you may refer this for sd boot
<cristian_c> ok, I see
<naobsd> little old but it should work on your rk3188 too
<naobsd> you need use your own kernel.img/boot.img
<naobsd> http://files.androtab.info/radxa/sdboot_rk3188.zip you may try newer files which is not linked/documented on that page
<naobsd> Tony_: rknand.ko is multi-role, but only 1 role(driver for the storage booted from) is activated at a time
<naobsd> Tony_: I don't know detail about rknand.ko. maybe role can be changed (e.g. use as NAND-only driver when booting from SD) but no information at all
<naobsd> well, I have to do my work... today's progress is not good... ;)
<cristian_c> ok
<Tony_> naobsd, I just disble CONFIG_SDMMC0_RK29 than the whole mmc driver is disable ?
<Tony_> *disable
<naobsd> Tony_: with rknand.ko enabled system? I'm not sure rknand.ko can act as sdio driver for wifi
c0d3z3r0 has quit [Ping timeout: 245 seconds]
<naobsd> Tony_: what you want is sdio wifi with sd boot, cedinu told that it's possible with RK mmc controller driver and without rknand.ko. I recommend to try it
c0d3z3r0 has joined #linux-rockchip
<naobsd> Tony_: that is, enable whole mmc driver and mmc0/mmc1 driver, and disable rknand.ko
<Tony_> naobsd, now I just ask question that CONFIG_SDMMC0_RK29 and CONFIG_SDMMC_RK29 are same thing?
<naobsd> no, CONFIG_SDMMC_RK29 enable/disable whole mmc controller driver,
<naobsd> CONFIG_SDMMC0_RK29 enable/disable 1st mmc controller
<naobsd> CONFIG_SDMMC0_RK29 enable/disable 1st mmc controller
<naobsd> CONFIG_SDMMC1_RK29 enable/disable 2nd mmc controller
<Tony_> now, one more question that why just disble CONFIG_SDMMC0_RK29 than the whole mmc driver is disable ?
<naobsd> etc
<naobsd> Tony_: you're talking about RK doc for sd boot?
<Tony_> Nope ?
<Tony_> boot from eMMC.
<naobsd> well
<Tony_> based that CONFIG_SDMMC0_RK29 enable/disable 1st mmc controller and CONFIG_SDMMC1_RK29 enable/disable 2nd mmc controller.
<naobsd> emmc is different controller
<Tony_> I know that.
<naobsd> there is CONFIG_EMMC_RK
<Tony_> Just let me say it.
wildea01 has joined #linux-rockchip
<Tony_> boot from eMMC, every thing is okay on my device.
<naobsd> no need to disable CONFIG_SDMMC_RK29/CONFIG_SDMMC0_RK29 for eMMC system
<naobsd> well
<Tony_> WiFi/SDCARD etc.
<Tony_> I know no need to do that.
<naobsd> are you using rknand.ko on your "boot from eMMC" system?
<naobsd> well
<Tony_> But I think that It should be still fine when I just disable mmc0.
<Tony_> Yes, I use rknand.ko on my "boot from eMMC"
<naobsd> mmc0 is not needed to "boot from eMMC"
<Tony_> I just want to test okay ?
<naobsd> and if you don't disable whole mmc driver and enable mmc1 driver, sdio wifi should work
<Tony_> don't tell me that not needed.
<naobsd> what I said is no need to _disable_ mmc driver for eMMC
<naobsd> I didn't say "mmc driver is not needed for eMMC"
<Tony_> I know that.
<naobsd> and if you need to use SD card on "boot from eMMC" system, you need mmc0
<Tony_> I just want to say that if i just disable mmc0.
<Tony_> What happens will ?
<naobsd> I'm confusing, what you asked was about "boot from SD", this is first time talking about "boot from eMMC" system, and I didn't tell you anything "not needed"
<naobsd> well, now talking about "boot from eMMC" system, if you disabled mmc0, you cannot use SD card (slot). sdio wifi(mmc1) should work
<Tony_> Yes, I think so. but the question is I just disable mmc0 than the sdio wifi(mmc1) can't work.
<naobsd> Tony_: you're talking about "boot from eMMC" system, right?
<Tony_> Yes.
<Tony_> boot from eMMC with rknand.ko.
<naobsd> Tony_: this is first time talking about it, right? I don't have any information about four "boot from eMMC" system. can you post error message?
<naobsd> four -> your
<naobsd> anyway I cannot know exactly what will happen when CONFIG_SDMMC0_RK29 is disabled _in your tree_, if it causes problem, I just recommend to enable it...
<naobsd> there should be no doc which says to do(disable mmc0) it...
<karlp> tony said RK told him he had too
<karlp> he was getting your help to try and work out why
<naobsd> karlp: there is RK doc about "boot from SD", it says "disable mmc0"
<naobsd> karlp: but it's not "boot from eMMC" doc
<naobsd> and I cannot guarantee CONFIG_SDMMC0_RK29 works as expected in any source tree... I still think it's just for mmc0, not mmc1(sdio wifi), but I cannot guarantee anything...
<karlp> rperier: mmind and julien were saying there's no pull_cfg_up/down, only pull_cfg_default on rk3066a, what's your pull_cfg_none meant to be then?
<Tony_> naobsd, after you tell me that about rknand.ko and mmc driver, I know a lot.
<naobsd> Tony_: please keep in mind that most of my knowledge is "guessing" with trial & error ;)
akaizen has quit [Ping timeout: 260 seconds]
<Tony_> Yes, thank you. and because my poor English, I can't describe my real question at first.
<naobsd> my english is not so clear, sorry
<naobsd> karlp: I think _none disables pull up/down, and rk3066 doesn't support any pull
<karlp> naobsd: yes, there was already "pull_default" though, I'm curious where rperier was going with _none to go with it.
<naobsd> ah
Tony_ has quit [Quit: Page closed]
akaizen has joined #linux-rockchip
akaizen has quit [Remote host closed the connection]
cristian_c has quit [Excess Flood]
cristian_c has joined #linux-rockchip
Tony_ has joined #linux-rockchip
<rperier> karlp: I am busy today, but basically all the pins state are set to "down" as default on rk3066, I mean these pins. (which is used when you set pcfg_pull_default), it does not work on the marsboard if you switch everything to default because most of these pins are connected to vcc (look at the schematic, I am not sure if it is specific to the board or the soc)
<Tony_> I can understand why sdio wifi can't run well now, because about rknand.ko.
phh_ is now known as phh
<rperier> karlp: pcfg_pull_default means: use the default pin state (as explained in the rk3066a datasheet), none means: bias is disabled, don't touch the state (at least this is how I understood it)
<karlp> rperier: ok, yeah, was just reading http://lists.infradead.org/pipermail/linux-rockchip/2014-October/000871.html and wanted to clarify.
<karlp> sounds all good though :)
<rperier> karlp: feel free to propose an idea, if you think that something is wrong. I am still improving my skills on hw (as I am a sw/system/embedded system guy and not hw)
<rperier> ;)
cristian_c has quit [Quit: Sto andando via]
hipboi has quit [Ping timeout: 245 seconds]
<naobsd> Tony_: rknand.ko breaks sdio wifi when "boot from eMMC"?
<karlp> rperier: no, just trying to make sure I understood what your changes were :)
<rperier> :)
<Tony_> I'm not sure, I have communicate with RK. I'm waiting. :)
<Tony_> naobsd, Maybe RK can let the rknand.ko support SDIO WiFi, I guess. :)
<mmind00> rperier karlp: pcfg_pull_none is bias disable ... so it _will_ touch the bias state
<naobsd> Tony_: I see
<naobsd> time to return home....
<Tony_> I'm home now. :P
<Tony_> naobsd, I will share you the reason when I'm sure.
hipboi has joined #linux-rockchip
naobsd has quit [Quit: Page closed]
bengal has joined #linux-rockchip
Tony_ has quit [Quit: Page closed]
FreezingCold has quit [Ping timeout: 272 seconds]
hipboi has quit [Ping timeout: 244 seconds]
FreezingCold has joined #linux-rockchip
hipboi has joined #linux-rockchip
rz2k|2 has joined #linux-rockchip
rz2k|3 has joined #linux-rockchip
rz2k_ has quit [Read error: Connection reset by peer]
rz2k|2 has quit [Read error: Connection reset by peer]
rz2k|3 has quit [Client Quit]
<rperier> mmind00: biasing is the only part which is not completly trivial to understand for pins description, imho
<rperier> with practice and time, I will improve myself ;)
<mmind00> rperier: it was only meant to clarify :-)=
<mmind00> (without the "=")
<rperier> yeah I know, don't worry. ;)
<rperier> advices or comments are always welcome
hramrach_ has quit [Remote host closed the connection]
hramrach_ has joined #linux-rockchip
mrcan_ has joined #linux-rockchip
mrcan has quit [Ping timeout: 250 seconds]
GriefNorth has quit [Ping timeout: 256 seconds]
FergusL has quit [Quit: http://quassel-irc.org - Chat comfortably. Anywhere.]
Bludot has joined #linux-rockchip
tonikasch has joined #linux-rockchip
tonikasch has quit [Client Quit]
tonikasch has joined #linux-rockchip
tonikasch has joined #linux-rockchip
mrcan_ has quit [Ping timeout: 240 seconds]
mrcan has joined #linux-rockchip
khilman has left #linux-rockchip [#linux-rockchip]
nighty^ has joined #linux-rockchip
tonikasch has quit [Remote host closed the connection]
naobsd has joined #linux-rockchip
mrcan_ has joined #linux-rockchip
mrcan has quit [Ping timeout: 256 seconds]
c0d3z3r0 has quit [Ping timeout: 250 seconds]
c0d3z3r0 has joined #linux-rockchip
<mmind00> rperier: don't worry about the current flood of mail ... once they read patch1 more carefully they'll understand :-D
<mmind00> rperier: at first I fell into the same trap when looking at patch6 before patch1
<rperier> I did not see yet... oh god... :|
<rperier> "but it doesn't handle poweroff-source which, one way or another, will be merged into Linus' tree creating a bisection point where things won't work."
<rperier> is he serious ?
<mmind00> rperier: I'm thinking Felipe is taking this a bit to far [I'm currently writing a reply about that]
<rperier> if my patches are merged onto the branch "poweroff" in Mark's regulator tree... everything will be merged into the Linus's tree as a whole ?
<rperier> ok thanks :)
<mmind00> rperier: but he has some valid points
<rperier> yeah, each commit is supposed to work, right ?
<mmind00> correct
<rperier> Felipe really needs to relax ^^ , like Marks said that's not the end of the world and not in a release... it is not even used yet...
arokux2 has joined #linux-rockchip
dlezcano has joined #linux-rockchip
gb_master has joined #linux-rockchip
theskilledworker has joined #linux-rockchip
theskilledworker has joined #linux-rockchip
wildea01 has quit [Quit: leaving]
dlezcano has quit [Ping timeout: 255 seconds]
dlezcano has joined #linux-rockchip
<karlp> whcih list are you guys on? linux-arm I guesS? something I'm not on I can only presume :)
theskilledworker has quit [Quit: Leaving]
<rperier> karlp: it is on devicetree but the discussion was fowarded to linux-arm too
<karlp> linux-usb and linux-rockchip are currently enough for me :)
Bludot has quit [Quit: Connection closed for inactivity]
FergusL has joined #linux-rockchip
khilman has joined #linux-rockchip
dlezcano has quit [Ping timeout: 256 seconds]
cedinu has quit [Quit: Page closed]
dlezcano has joined #linux-rockchip
FergusL has quit [Read error: Connection reset by peer]
gb_master has quit [Remote host closed the connection]
dlezcano has quit [Ping timeout: 265 seconds]
hramrach_ has quit [Ping timeout: 246 seconds]
hramrach_ has joined #linux-rockchip
FreezingCold has quit [Ping timeout: 255 seconds]
FreezingCold has joined #linux-rockchip
arokux2 has quit [Remote host closed the connection]
bengal has quit [Quit: Leaving]
prop81 has joined #linux-rockchip
<naobsd> too many unread email in linux-rockchip mbox...
<naobsd> need more time to follow mainline ;)
FergusL has joined #linux-rockchip