| [00:17:48] |
*** |
GPSFan has quit IRC |
| [00:45:47] |
*** |
drmikecrowe_ has quit IRC |
| [09:41:01] |
*** |
GPSFan has joined #zipit |
| [09:57:23] |
*** |
bryan_w has quit IRC |
| [11:58:10] |
*** |
Magon has joined #zipit |
| [17:36:28] |
<Rossimo> |
Backer: i haven't see any others being used |
| [19:04:37] |
*** |
Magon has quit IRC |
| [19:35:25] |
<Backer> |
Rossimo: Ok, just checking. I saw a reference on the wiki to the marvell driver |
| [19:35:40] |
<Backer> |
I have been having problems setting an essid via iwconfig |
| [19:35:46] |
<Backer> |
yet iwlist scanning on the interface works fine |
| [19:36:53] |
<Rossimo> |
have you tried wpa_supplicant? it's been working for me for WPA and WEP networks |
| [19:45:32] |
<Backer> |
i've run it without any luck |
| [19:46:01] |
<Backer> |
so then I tried to see if i could at least configure it manually |
| [19:46:25] |
<Backer> |
I get "SET failed on device eth0 ; Network is unreachable." |
| [19:49:29] |
<Backer> |
I wonder... I tried setting the wpa config inside /etc/network/interfaces ... so on bootup it (unsuccessfully) tries to configure. I'm going to remove that and try to manually configure on a fresh reboot... |
| [19:52:00] |
<Backer> |
are you setting your wpa config in /etc/wpa_supplicant.conf ? |
| [19:56:49] |
<Rossimo> |
yep |
| [19:57:08] |
<Rossimo> |
are you following the instructions on the wiki? |
| [19:57:26] |
<Rossimo> |
wpa_supplicant -Dwext -c /etc/wpa_supplicant.conf -ieth0 |
| [20:03:11] |
<Backer> |
yep, no luck though |
| [20:03:25] |
<Backer> |
wpa_cli status says 'scanning' |
| [20:03:33] |
<Backer> |
iwconfig eth0 doesn't report an essid being set |
| [20:05:58] |
<Rossimo> |
what does your wpa_supplicant look like? |
| [20:06:18] |
<Backer> |
after first running wpa_supplicant, I get "ioctl[SIOCSIWPMKSA]: Invalid argument" |
| [20:07:11] |
<Backer> |
network={ |
| [20:07:18] |
<Backer> |
ssid="xxx" |
| [20:07:22] |
<Backer> |
psk="yyy" |
| [20:07:23] |
<Backer> |
} |
| [20:08:36] |
<Rossimo> |
ctrl_interface=/tmp/wpa_ctrl |
| [20:08:37] |
<Rossimo> |
network={ |
| [20:08:37] |
<Rossimo> |
ssid="xxx" |
| [20:08:37] |
<Rossimo> |
key_mgmt=WPA-PSK #change if needed |
| [20:08:37] |
<Rossimo> |
psk="yyy" |
| [20:08:37] |
<Rossimo> |
} |
| [20:08:41] |
<Rossimo> |
try that |
| [20:09:31] |
<Backer> |
hmmm, ok ctrl_interface was set to /var/run/wpa_supplicant |
| [20:10:45] |
<Backer> |
don't expect that matters as wpa_cli seemed to connect... |
| [20:11:01] |
<Rossimo> |
oh |
| [20:11:16] |
<Rossimo> |
have you tried an older image? |
| [20:11:39] |
<Rossimo> |
such as http://clemsonlinux.org/~rossimo/Angstrom-base-image-glibc-ipk-2008.1-test-20080225-zipit2.rootfs.tar |
| [20:12:45] |
<Backer> |
will try that |
| [20:14:59] |
<Backer> |
I should be able to use the wpa_supplicant and wireless-tools ipkgs from my own build on that image, no? |
| [20:21:06] |
<Rossimo> |
should be, i would think. but, we have had trouble with the later OE builds. i'm not %100 sure |
| [20:22:29] |
<Rossimo> |
i've got a full image at http://clemsonlinux.org/~rossimo/oe-zipit2-image-compressed.tar , if you would like to use it |
| [20:22:30] |
<Backer> |
I took a fresh svn build on the 21st and didn't have the problem with the missing libgcc_s.so.1... everything came up ok |
| [20:22:37] |
<Rossimo> |
you'll have to extract it, first |
| [20:23:06] |
<Backer> |
thanks, I'm grabbing it now |
| [20:25:15] |
<Backer> |
i need to get kbdrate or an equivalent on this thing... key repeat is driving me wild! :o) |
| [20:46:53] |
<Backer> |
looks like the same results with 20080225 image |
| [21:12:00] |
<Backer> |
root password set on that last image ;o) |
| [21:29:55] |
<Backer> |
Bizarre. I'm getting the same behavior even with the full image. Essentially, the ioctl() error on starting wpa_supplicant (running your wireless.sh script) and "SET failed on device eth0 ; Network is unreachable." when trying to set ssid via 'iwconfig eth0 essid <ssid>' (I'm guessing the unreachable part is a generic errno string) |
| [21:30:43] |
<Backer> |
I would guess that, no matter what, setting an ssid manually should not evoke that error. |
| [21:31:15] |
<Backer> |
Yet, I know something with the driver must be working because iwlist scanning eth0 shows the local networks in the area. |
| [21:31:48] |
<Backer> |
I will have to put it away... maybe next time I will try to get strace on it and see what ioctl() iwconfig is dying on. There is nothing showing in the dmesg either. |
| [22:19:06] |
*** |
bryank has joined #zipit |
| [22:22:43] |
<bryank> |
Rossimo: I figured out a way to get libgcc_s in the image, sort of: if i include gcc (not gcc-cross) in the wireless-image.bb file, it works. |
| [22:23:06] |
<bryank> |
i'd like to find a way to modify the gcc_4.2.2.bb file to build *just* libgcc, but i haven't had a lot of time to try that |
| [22:23:36] |
<bryank> |
actually, getting it to install from the gcc-cross sysroot build (like it worked before sysroot) should work too. |
| [22:27:31] |
<bryank> |
Backer: i assume you tried to set the essid more than once? i've seen issues sometimes the first time i try to set anything, but it clears up for me eventually. |
| [22:27:56] |
<Backer> |
I tried repeatedly but no luck |
| [22:33:08] |
<Backer> |
bryank: Are you having the missing libgcc_s problem with the latest in svn? |
| [22:34:36] |
<bryank> |
sort of, yeah. i haven't tried from a completely clean directory though. |
| [22:35:50] |
<Backer> |
I pulled a tree (mainline anyway) two days ago, after I heard about the problem, but it's build turned out to run fine and libgcc_s was in place. |
| [22:36:31] |
<bryank> |
weird. what rev of oe did you get? (look in the org.openembedded.dev/_MTN/revision file) |
| [22:37:36] |
<Backer> |
old_revision [5744cdb7ad43a125c73b2dd57d51ded57a140c60] |
| [22:37:38] |
<Backer> |
that field? |
| [22:37:41] |
<bryank> |
yep |
| [22:38:03] |
<bryank> |
the last rev that i built was... hmm... f7df80732af56b59315ef203e69f092313ec2ddd, on march 20, so maybe they fixed it. |
| [22:38:49] |
<Backer> |
maybe, I don't know if something about my local environment would have anything to do with it. |
| [22:39:39] |
<bryank> |
that's possible, but i think you were using newer code. i'll try another build of current head (as of right now) tomorrow. |
| [22:41:05] |
<Backer> |
Can the built-in flash on the unit be accessed, via mtd or some such? I'd like to get at the gspi modules |
| [22:41:50] |
<Backer> |
Hm, three mtd entries |
| [22:41:58] |
<bryank> |
i know T0mW was doing something to overwrite the built-in flash kernel file, but he bricked several devices doing it |
| [22:42:18] |
<bryank> |
i believe the mtd modules themselves are in the initramfs inside the kernel image |
| [22:42:23] |
<bryank> |
er, wireless modules |
| [22:42:24] |
<Backer> |
I would be happy to mount ro ;o) |
| [22:42:45] |
<Backer> |
ah |
| [22:42:53] |
<Backer> |
ok, that's mounted I think.. |
| [22:44:38] |
<bryank> |
if it helps, i think there are some details about the mtd devices in dmesg (names and such) |
| [22:46:34] |
<bryank> |
actually, are you getting a hostname of "(none)" with your build, or "zipit"? |
| [22:48:15] |
<bryank> |
oh, wait, never mind: that rev isn't current oe: you have current zipit2 svn, but the current svn makefile pulls a known-working rev of oe |
| [22:49:33] |
<Backer> |
Ah, I see |
| [22:50:13] |
<Backer> |
Sorry for the red herring. |
| [22:50:41] |
<bryank> |
on mtd: 0 is the bootloader, 1 is the kernel (presumably plus initramfs), and 2 is the (minimal) fs containing the standard zipit firmware files (program, etc. |
| [22:50:44] |
<bryank> |
np |
| [22:50:55] |
<bryank> |
(i'll still try the build, just in case ;-) ) |
| [22:52:52] |
<Backer> |
gladly found kbd_mode :o) |
| [22:53:25] |
<bryank> |
does it help? i've been thinking about redoing the zipit keyboard helper module to autorepeat on its own, but if kbd_mode works, maybe i'll just use it. |
| [22:54:32] |
<Backer> |
well kbdmode -d 500 helps |
| [22:54:44] |
<Backer> |
I am getting duplicate keystrokes |
| [22:54:57] |
<Backer> |
for a second I thought maybe it wasn't debounced but then it seemed like it's just a short time |
| [22:55:28] |
<Backer> |
combined with keypresses not registering cleanly... at least mine |
| [22:55:41] |
<Backer> |
now if I could turn on this kbd light ;o) |
| [22:55:44] |
<bryank> |
the driver doesn't get interrupts when the keys get pressed, so i don't *think* it's a debounce thing |
| [22:55:59] |
<bryank> |
yeah, the kb backlight is something that (afaik) we don't have the interface to yet |
| [22:57:14] |
<bryank> |
i think there's some code in the kernel patch (z2kernel project) to turn it on and off, but it's disabled, and i'm not sure if it'll actually work or not. |
| [22:58:46] |
<bryank> |
wow, kbd_mode seems to really help. |
| [22:59:08] |
<Backer> |
oh, that would explain things some! |
| [23:02:06] |
<Backer> |
and yet pxa27x-keyboard is getting interrupts on irq 4 |
| [23:02:13] |
<bryank> |
i *think* that's a timer |
| [23:02:23] |
<bryank> |
i know it sets up a kernel timer to poll the keyboard matrix |
| [23:02:25] |
<Backer> |
I see |
| [23:03:07] |
<bryank> |
anyway, i should go. good luck with the wireless -- and if you find a way to extract the initramfs from the kernel, i'd be interested. :-) |
| [23:03:26] |
<Backer> |
thanks for your help! |
| [23:03:38] |
<bryank> |
you're welcome! |
| [23:03:41] |
*** |
bryank has quit IRC |
| [23:19:08] |
*** |
GPSFan has quit IRC |
| [23:51:56] |
*** |
Rossimo has quit IRC |
| [23:57:10] |
*** |
Rossimo has joined #zipit |
