<wpwrak>
(the mmc0 errors are probably because i had a non-mmc card inserted)
<wpwrak>
the UBIFS problems look as if the kernel had troubles finding the right partition offset
<wpwrak>
one thing about the spi errors is that there's no ili8960 in the kernel. i saw a patch for it in the owrt patch collection, though
<wpwrak>
let's try this with a bit more verbosity ...
porchaso0 has joined #qi-hardware
porchao has quit [Ping timeout: 255 seconds]
<wpwrak>
what's odd is that there's no word about NAND detection
jekhor has quit [Ping timeout: 245 seconds]
pcercuei has joined #qi-hardware
xiangfu has quit [Ping timeout: 245 seconds]
<wpwrak>
the spi complaints came from missing 0008-qi_lb60-Don-t-use-3-wire-spi-mode-for-the-display-fo.patch
Calyp has quit [Quit: gone working on freeconomy =o)]
xiangfu has joined #qi-hardware
<wpwrak>
phew. debug output at last. something seems to have changed with loglevel.
<wpwrak>
the ubifs errors seem to be caused by 0020-qi_lb60-NAND-add-data-partition.patch missing
<wpwrak>
one step closer to sanity :)
<wpwrak>
let's see if the stock kernel could still boot on this ben ...
<wpwrak>
naw, dies too. very good.
<wpwrak>
reflash_ben time again ...
pcercuei has quit [Quit: Bye]
xiangfu has quit [Ping timeout: 255 seconds]
xiangfu has joined #qi-hardware
wolfspraul has quit [Quit: leaving]
<wpwrak>
interesting ... after a reinstall with the "official" distro, most printks are gone, too. ah well, life is full of small mysteries :)
<wpwrak>
now the kernel boots into busybox. not making it to a login/shell, though. let's see what else it needs ...
<wpwrak>
aah, a shell. victory at last ! :) devtmpfs did the trick
<wpwrak>
fails to mount /proc, though the kernel has it. hmm.
<wpwrak>
larsc (or maybe xiangfu): do you remember whether 0004-NAND-Add-support-for-subpage-reads-for-NAND_ECC_HW_O.patch is required or whether it's merely an optimization ?
<wpwrak>
seems that there may be trouble in the future with SPI: spi_gpio spi_gpio.1: master is unqueued, this is deprecated
<wpwrak>
ubifs is also quite cheerful: UBI warning: print_rsvd_warning: cannot reserve enough PEBs for bad PEB handling, reserved 10, need 80
<wpwrak>
luckily, there are neither bad nor corrupt PEBs :)
<wpwrak>
i merged 0006-MTD-NAND-JZ4740-Multi-bank-support-with-autodetectio.patch (with whitespace cleanup) and 0008-qi_lb60-Don-t-use-3-wire-spi-mode-for-the-display-fo.patch
<wpwrak>
0004-NAND-Add-support-for-subpage-reads-for-NAND_ECC_HW_O.patch needed some small modifications
<wpwrak>
oh, and why do we have 0009-dev-mem-Add-kernel-config-option-to-omit-this-device.patch ? please don't take /dev/mem away ! anything using ubb from user space (libubb, ubb-vga, etc.) and the ben-wpan userspace tools with atben need it
<kyak>
we had to provide our own libnl-1 for older lowpan-tools, because openwrt shipped libnl-2. Now we have to provide our own libnl-3 for the latest lowpan-tools, because openwrt is still using libnl-2 :)
jekhor has joined #qi-hardware
kilae has joined #qi-hardware
xiangfu has quit [Ping timeout: 260 seconds]
lekernel has joined #qi-hardware
xiangfu has joined #qi-hardware
apelete has joined #qi-hardware
jurting has joined #qi-hardware
LunaVorax has joined #qi-hardware
unclouded has quit [Ping timeout: 246 seconds]
qwebirc19838 has joined #qi-hardware
qwebirc19838 has quit [Client Quit]
security has joined #qi-hardware
xiangfu has quit [Ping timeout: 272 seconds]
megha has quit [Ping timeout: 252 seconds]
Calyp has joined #qi-hardware
wolfspraul has joined #qi-hardware
Calyp has quit [Ping timeout: 255 seconds]
wolfspraul has quit [Ping timeout: 240 seconds]
viric_ has joined #qi-hardware
viric has quit [Ping timeout: 255 seconds]
<wpwrak>
kyak: thanks ! yes, the libnl dependency is a bit messy. the mailing list is full of people wrestling with it ;-)
<viric_>
have you seen this kernel build error? 3.8.4 and 3.8.5
<viric_>
kernel/signal.c: error: 'struct sigaction' has no member named 'sa_restorer'
<wpwrak>
the new version identifies itself as 0.3, like yours. not sure if this means they're the same and "set" is missing for a more obscure reason ?
<wpwrak>
scratch the "?".
<wpwrak>
viric_: accidental editing ? and no, haven't seen this yet
<viric_>
hm seems to be a known issue on mipslinux
<kyak>
wpwrak: what's so special about those instructions?
<kyak>
had no problems with libnl dependency.. apart from the fact that i had to package it
<wpwrak>
"iz set" is needed to set addresses "cleanly" with net-next, since the izcoordinator functionality has been lost (at least temporarily)
<wpwrak>
you have a plan B, abusing izcoordinator for this task, but that's a little messy
<kyak>
wpwrak: do you see the version of iz as 0.3?
<wpwrak>
yes. yours (without "iz set") and the one i build from sources for x86_64 (with "iz set") both identify as 0.3
<kyak>
then i have no idea where the "set" option is :)
<kyak>
i could expect a special option to enable it in build instructions, but there is nothing like that
<wpwrak>
src/iz-mac.c line 571 should be .name = "set", that would indicate that you have the right version
<kyak>
so the fact that it grabbed the 0.3 tarball doesn't indicate that? :)
<wpwrak>
and i found that all the bitching from owrt depends on the kernel, even if it looks as if it was entirely user space problems
<wpwrak>
i.e., if i install the owrt kernel, the complaints go away
<wpwrak>
well, this is now lower priority, given that i can use atben with the owrt kernel
<viric>
wpwrak: owrt starts with patching for the path of the 'init' elf
<wpwrak>
hmm, but how would that depend on the kernel ?
xiangfu has joined #qi-hardware
<viric>
?
<viric>
the code that finds the init elf is in the kernel
<wpwrak>
i had to enable devtmpfs to get to the shell prompt at all, which is something the owrt kernel doesn't even have. so there's something fishy.
<viric>
:)
<wpwrak>
aaah !
<viric>
because you start with a too-void fs
<wpwrak>
no i remember what you mean
<wpwrak>
s/no/now
<qi-bot>
wpwrak meant: "now i remember what you mean"
<wpwrak>
right. owrt's "we introduce a random change that guarantees that we'll never work with a vanilla kernel" policy
<viric>
maybe there is an insightful rationale behind that, but I don't know it
<wpwrak>
where's a psychopathologist when you need one ? ;-)
<kyak>
wpwrak: as far as i remember, you can use vanilla kernel (nanonote_defconfig or whatever it is called) and the only patch you actually need is for partition layout
<wpwrak>
that's what i almost got to boot. but then busybox didn't find things.after tossing in devtmpfs, i comes up, but with complaints
<wpwrak>
s/ i / it /
<qi-bot>
wpwrak meant: "that's what it almost got to boot. but then busybox didn't find things.after tossing in devtmpfs, it comes up, but with complaints"
<wpwrak>
s/ needs an index option. s/ i / it /[2]
<wpwrak>
or [1], if we make it 0-based
<viric>
:)
<viric>
the partition layout can be set in the kernel config
<viric>
even in uboot
<wpwrak>
i passed a compile-time command line, just to make sure. also need it for loglevel. the kernel is a bit too silent by default.
<wpwrak>
if anyone wants to play: this is how owrt comes up with the net-next kernel: http://pastebin.com/gibQYdG7
<viric>
wpwrak: I used a ftdi cable with the next-to-battery serial port
<viric>
wpwrak: ha, you git the init name patch thing? :)
<viric>
*hit*
<lindi->
wpwrak: are you using openocd with that?
<viric>
lindi-: does the ingenic support jtag?
<wpwrak>
viric: yeah. fixed in my tree now. one evil critter down :) i think now all that's missing is usb. (the rest are audio, nand optimizations, and cpufreq, which i think i don't need)
<lindi->
viric: what is ingenic?
<viric>
lindi-: I mean the cpu in the nanonote
<wpwrak>
lindi-: yes, i have a (very slow and unreliable) openocd driver for it
<viric>
ah I didn't know
<viric>
I used only kgdb
<lindi->
wpwrak: do you normally use that or something else?
<viric>
which works impressively well
<wpwrak>
lindi-: haven't touched the phones for a long time, though. so i never improved that driver. idng for the ben doesn't have jtag
<lindi->
wpwrak: ok
<whitequark>
hahaha Embedded Air
<whitequark>
excellent
<wpwrak>
lindi-: most of the time, i just use printk debugging