01:46 <tehchubbz0r> i'm trying to figure out if i'm on the right path, a project i'm involved with is looking for a device (stick or box form factor) that we can build a custom android firmware for (signing our own system apps, etc), we've bought some off the shelf devices to "test the waters" with, and so far all have been based around the RK3188 or 3188T processor. anyway, instead of using a rom kitchen to modify someone elses firmware ima

01:46 <tehchubbz0r> i should start looking for a way to build our own kernel and system from scratch, or is there a better way?

01:48 <tehchubbz0r> building for the rockchip-based devices i mean, it'd be even better to find a way to deal directly with the people making these boards instead of buying off-the-shelf devices in bulk and modifying them

02:58 <naobsd> hmm

06:38 <naobsd> hmm.... nfs is extremely slow and hangs very easily... :(

18:35 <ruslan> hello

18:35 <ruslan> may I ask a question?

18:37 <ruslan> I have a device running Android 5.1.1 on Rockchip 3288. Is there SDK like rk3288_r-box_android4.4.2_sdk but for Android 5.1.1?

21:21 <bryanp> Anyone ever look into rknand and write barriers? Is the nand actually flushing?

22:21 <bryanp> More generic question. Has anyone dealt with ext4 corruption with rknand? Data integrity and FS integrity.

23:31 <naobsd> bryanp: nand implementation is in binary blob, no idea about detail

23:32 <naobsd> bryanp: if you are using eMMC device with rknand module (as mtd device), you may try open source mmc driver instead of rknand module

23:33 <bryanp> naobsd: minix neo x5 type board. Nand + driver. No emmc I am afraid

23:34 <naobsd> bryanp: you may try up-to-date kernel/driver/blob

23:34 <naobsd> 3.0.8 must be old

23:35 <bryanp> I think 3.0.36 is newest? It may be filesystem level though. Data=ordered + noauto_da_alloc. Not sure though

23:35 <bryanp> I could definitely test with newer kernel/blob pairs but it would be a large undertaking as that is a move from 4.1.1 -> 4.2.2 or 4.4+.

23:37 <naobsd> I cannot understand neo x5 _type_ board. if it's not x5, it's not x5. "it's very similar spec" is not useful because almost of all rk3066 device is very similar, but different board is different.

23:38 <naobsd> then, if you're sure what is different about your board, I think you can prepare new kernel with 4.4 sdk based source tree

23:38 <bryanp> It is based on neo x5 but is somewhat different including TI PM IC instead of WM

23:39 <naobsd> is it neo x5?

23:39 <naobsd> I cannot understand about "based" part

23:39 <bryanp> except power IC

23:39 <naobsd> why PMIC is different?

23:39 <bryanp> Different requirements that weren't supported by the other PMIC

23:40 <naobsd> are you really sure other parts are same? e.g. you ordered x5 board with different PMIC?

23:41 <bryanp> Yes

23:41 <naobsd> I see

23:41 <naobsd> (sorry, I saw a lot of people assume "it's same as XXX" without any technical reason)

23:41 <bryanp> Yeah no worries.

23:43 <naobsd> I think minix released 4.2 SDK

23:44 <bryanp> no Kconfig tho

23:45 <bryanp> It could just be delayed allocation corruption. I haven't see any corruption on kernel,boot,recovery areas nor system which is readonly most of the time

23:58 <naobsd> bryanp: do you have gmail account?

23:58 <naobsd> well,