02:03 | rmull | Here's some verbose output I get from my cubox when I plug in a USB mouse: http://sprunge.us/CETO |
02:05 | rmull | I interpret this to mean that I have this bit set: QTD_STS_XACT, which is "device gave illegal response" |
02:06 | rmull | There's really no way for me to know whether things are behaving correctly without a USB sniffer device, right? |
02:07 | cbxbiker61 | have you tried a different mouse? |
02:08 | rmull | No, unfortunately I only have this one |
02:09 | cbxbiker61 | did you say it worked on a different linux box? |
02:09 | rmull | Yes, works on my Linux PC and my Windows laptop |
02:09 | rmull | It's a Razer Krait mouse |
02:10 | cbxbiker61 | are you sure it doesn't require a special driver? |
02:10 | rmull | Fairly sure, in my 5-6 years of using this mouse it never has |
02:10 | rmull | It's never not worked before |
02:11 | cbxbiker61 | isn't Razer a "sophisticated" mouse? |
02:11 | rmull | This particular mouse has only two buttons and a wheel/button, nothing fancy |
02:11 | rmull | I don't know if it has any specific protocol extensions but I would guess not |
02:12 | rmull | a search for "razer" in the menuconfig reveals nothing |
02:13 | cbxbiker61 | i'd suggest buying a "cheapy" mouse, to verify if it's specific to the mouse |
02:13 | rmull | That's a fine idea, I can borrow one from work |
02:13 | cbxbiker61 | find -name Kconfig -exec grep -H RAZER {} ';' |
02:13 | cbxbiker61 | in the kernel source |
02:14 | rmull | Yields nothing except when I pass grep the -i flag, but then it's an unrelated product (a sound device) |
02:16 | cbxbiker61 | grep MOUSE .config |
02:17 | rmull | CONFIG_INPUT_MOUSE and CONFIG_MOUSE_PS2 are set |
02:17 | rmull | all others are not set |
02:18 | rmull | This is from my running /proc/config.gz FWIW |
02:18 | cbxbiker61 | check /proc/config.gz on your working linux box and see if it has any other MOUSE options set |
02:20 | rmull | MOUSEDEV is set on my working machine, I'll add that and see if it makes any difference. I was under the impression that it wasn't necessary in a system that uses evdev. |
02:22 | cbxbiker61 | it's set on my cubox kernel builds |
02:23 | rmull | Alright, I'll go quiet until I boot the new kernel and have a new report |
02:24 | rmull | Thanks for the attention, by the way. |
02:28 | rmull | Hmm, no change at all with the new kernel |
02:50 | rmull | So ruling out kernel configuration - how should I try to debug this thing? I can test another mouse to see if it works, but that doesn't really address the problem. The mouse works with all other machine it's been connected to. |
02:51 | cbxbiker61 | yeah, check the other mouse, other than that i'm out of advice |
02:52 | rmull | I'll see if I can get a transaction log with usbmon |
02:56 | rmull | Should there be section mismatches when I build my kernel? |
02:57 | rmull | here is an example, enabled by turning on section mismatch analysis in the menuconfig: http://sprunge.us/XHEd |
02:58 | rmull | It looks really ugly and I don't think it's good |
02:58 | rmull | But I don't know the actual problem |
02:59 | cbxbiker61 | section mismatches are typically non-critical |
02:59 | rmull | Okay, thanks. Good for the peace of mind :P |
02:59 | cbxbiker61 | with that said, it's best to fix them |
02:59 | cbxbiker61 | there is a patch set under rabeeh kernel thread on forum |
03:00 | cbxbiker61 | they look fine to me, i'll push them to rabeeh's kernel after i have a chance to test them out |
03:01 | cbxbiker61 | i've already applied them to my 3.5.0 kernel |
03:02 | rmull | They only fix the mismatch warnings? No functional change, right? |
03:03 | cbxbiker61 | they put code in the proper sections, that's all |
03:03 | rmull | alright. |
03:04 | rmull | Okay, so I took a usbmon capture, but I have to use my other computer which has wireshark to decode it. That'll be for tomorrow. |
20:30 | dv505 | tomlohave, I downloaded the new image but still have the same issue. no network service, xorg, xbmc |
20:31 | dv505 | one thing I noticed (might be unrelated), there is an error when extracting the archive |
20:31 | dv505 | in /usr/lib, there are folders for udisks and terminfo, but also symlink folders. can't have them both |
20:33 | dv505 | maybe there are other problems with symlinks too. e.g. libEGL.so and libEGL.so.1 both exists as files, whereas libEGL.so is originally a symlink |