01:18 <dw> phh: useful information, will do some more diffing, thanks

01:18 <dw> *digging. probably also diffing =)

12:02 <dw> i've been trying without success to identify the last vanilla android tree commit that either the 3.0.8 or 3.0.36 kernels are based on.. for lack of a better place to ask, any tips? current thinking is a giant overnight for loop to just brute force diff every commit in android kernel/common repo and identify the smallest

13:11 <dw> seems i misunderstood the android branch names :) "android-3.0" means "the android 3.x kernel", which includes both my source verions. no clue where android keeps tags

13:55 <phh> dw: i'm pretty sure you won't find, because they must have backported some patches

13:55 <phh> i don't think android 4.4 can run on vanilla android 3.0 kernel

14:14 <jbaiter> has anybody else experienced reboots when saturating the ethernet connection?

14:15 <jbaiter> https://groups.google.com/forum/#!topic/radxa/StBCamD1_Lo

14:22 <m1r> jbaiter: can you explain about your reboots ?

14:23 <FergusL> phh: haha, you, here ? vous, ici ?! (i'm in #quassel-irssi)

14:23 <phh> irc is little ?:p

14:23 <FergusL> freenode is !

14:23 <jbaiter> m1r: whenever I try to send a lot of data over netcat to a remote server, the machine reboots after ~1sec

14:24 <FergusL> actually no it's not, randomness just makes things happen sometimes

14:24 <jbaiter> this onyl happens when the radxa is the one sending the data, receiving is not a problem

14:24 <m1r> jbaiter: pvt msg ?

14:30 <FergusL> is there a python library for GPIO on the cb2 ?

14:31 <FergusL> hu, sorry

15:14 <jbaiter> is there a bugtracker somewhere?

15:48 <dw> phh: in this case, 'android-3.0' means 'android kernel based on upstream linux 3.0.x', not 'android kernel for android 3.x' :)

22:57 <dw> so, having dug through every rk3xx kernel i can find, as far as I can tell the rknand driver is utterly incomplete, unless my C skills are failing me

22:58 <dw> add_rknand_device() never gets called, cannot be called since nothing seems to reference it except to stash a funcptr in a struct that is never read, and even if it were called, accesses uninitialized global variables to calculate the size of the flash device

22:58 <dw> is it entirely possible this really is broken? come back to it 3 times tonight, and convinced the driver couldn't possibly work

23:00 <dw> perhaps a better question would be, has anyone ever seen /proc/mtd populated on an open-source built kernel?

23:01 <dw> hmm maybe it's buried in asm somewhere

23:03 <dw> nope :(

23:07 <dw> """So clearly this is the NAND driver for this platform. And quite clearly, the source is not included in the kernel sources we have right now."""

23:07 <dw> doh :) well at least i know i'm not mad, now :P

23:07 <phh> dw: when i said "android 3.0" i shoud have said 3.0.36

23:08 <phh> dw: and yes, nand drivers are available only as kos

23:08 <phh> but what i've said still applies (the two different bootloaders/two different rk30nand_ko.ko)

23:09 <dw> tried to find info on how the bootloaders are swapped (if they can, at all). there is a rom mask that reads secondary stage loader from the flash i guess? but couldn't find any references to what address that loader lives at

23:10 <dw> also tried string'ing the loader .bin, loaded it into IDA and tried to find some regions containing (though did so very lazily), no luck in either case

23:10 <dw> is it encrypted?

23:10 <phh> no clue

23:26 <dw> thanks phh, good pointers :)

23:36 <jbaiter> dw: /proc/mtd lists my partitions

23:36 <jbaiter> using 3.0.36 on a radxa rock, compiled this morning

23:37 <jbaiter> though i'm not entirely sure if there aren't any proprietary drivers in the initrd

23:52 <dw> jbaiter: very useful, thanks :)

23:52 <dw> i have 3gb radxa sdk downloading since earlier, will check its rknand when don

23:52 <dw> ah nope, its the minix neo

23:53 <dw> jbaiter: https://github.com/radxa/initrd/blob/master/rk30xxnand-3.0.36%2B.ko :(