X server segfault when using BT keyboard

[dimag0g]

Hello all,

I have recently purchased a BT keyboard (Gembird KB-316BT) and successfully connected it to Pandora. However, the X server doesn't seem to like it at all, crashing after a few minutes of keyboard useage. The error is always the same:

0: /usr/bin/X (xorg_backtrace+0x2c) [0xa2d78]

Segmentation fault at address 0x1c

 

The crash seems to happen when at least one key is pushed on the keyboard. After the crash, the keyboard remains fully functional - by pressing Ctrl+Alt+F1 I could go to console,  log in, and I kept using the keyboard for at least 1 hour without any issues.

 

I attach the X log file in case someone is eager to help me.

 

 

[Klumpen]

Same here

 

[Neelix]

I've had this happen to me too.  I stopped using the Bluetooth keyboard on the pandora directly as a result.

- Neelix

 

[ptitSeb]

I would be interested if you could test bt keyboard with my new build of Xlib, using latest MonoRT (or upcoming wxPython).

 

[notaz]

Doesn't look like this can be debugged without having actual BT keyboard..

 

[dimag0g]

I would be interested if you could test bt keyboard with my new build of Xlib, using latest MonoRT (or upcoming wxPython).

I don't quite see how MonoRT could affect the X server which is already running. Anyway, I'm ready to test my keyboard if you explain me what to do.

 

[ptitSeb]

I would be interested if you could test bt keyboard with my new build of Xlib, using latest MonoRT (or upcoming wxPython).

I don't quite see how MonoRT could affect the X server which is already running. Anyway, I'm ready to test my keyboard if you explain me what to do.

I have many libs inside MonoRT that doesn't come from Firmware and X11 is one of them. It's just a test. Just open MonoRT Command Line and try to use your BT Keyboard inside the command line (or in MonoDevelop or anything else spawned from monort). It's a just a test, it may just crash the same way...

 

[ExquisiteCorpse]

That's a shame - I was looking to purchase a BT keyboard for my pandy

 

[dimag0g]

ptitSeb, I tried to use my BT keyboard with MonoRT like you suggested, and it crashes all the same. I'd love to try something else, but I don't have the knowledge to guess what could make it work. I wonder if someone have tried Bluetooth mouse yet.

 

[_wb_]

Doesn't look like this can be debugged without having actual BT keyboard..

If you have two Pandoras, you can try to reproduce the issue by using one of them as a BT keyboard for the other one (cf. http://repo.openpandora.org/?page=detail&app=package.mastercontrol).

 

[ExquisiteCorpse]

Here's the crash page I get.

For whatever reason it seems to set the CPU clock to 60000

....Lol.

 

[levi]

I don't see any error there. To my eyes it looks like the standard boot sequence. Can you get to a terminal there to switch to another tty and do get some proper debug output?

 

[ExquisiteCorpse]

I'll certainly give it a go.

I'll recreate the fault. Once it falls back to terminal what do you want me to punch in?

Edit: Sods law, just had keyboard hooked up for 40 mins typing gibberish and it never crashed yet earlier in the coffeee shop it crashed after 2 minutes.

I am going out for the evening now but will attempt to recreate the crash tomorrow.

 

[levi]

I'm not 100% sure - hopefully someone more experienced with Xorg will pop in here, but for a start the tail end of dmesg should be useful ( type 'dmesg|tail -n 20' or something like that), and if there's anything in /var/log/Xorg.*.log (I have /var/log/Xorg.0.log there) the end of that could be interesting too.

 

[dimag0g]

I'm not 100% sure - hopefully someone more experienced with Xorg will pop in here, but for a start the tail end of dmesg should be useful ( type 'dmesg|tail -n 20' or something like that), and if there's anything in /var/log/Xorg.*.log (I have /var/log/Xorg.0.log there) the end of that could be interesting too.

Yes, well, I have attached Xorg.0.log to the post, haven't I? And I haven't seen anything in dmesg, which is not much of a surprise because the crash doesn't seem to be driver-related (the keyboard functions well after the crash)

 

[levi]

Sorry, you indeed have (I wasn't looking at your post when I last posted). It does indeed reveal the seg fault shortly after recognising the new keyboard input, and it appears to die after '[Xi] Too many valuators reported for device 'Virtual core pointer'. Ignoring event.'. If anyone's familiar with the source code, hopefully that's a useful insight into the code in error.

 

[Klumpen]

Hopefully this gets fixed.

I bought my foldable BT-Keyboard only for Pandora and it's still completely useless, currently thinking about giving it to my GF, since she has a smartphone, but it's one of the crappiest Huawei ever produced and may not even have working BT.

 

[ExquisiteCorpse]

hey guys anything we can do to get this fixed?

Its driving me mad as it turns my pandy from a gaming device into a full fledged netbook.

Perhaps I can purchase the same keyboard for a dev willing to fix it? That way they can debug it and keep the BT keyboard as a small gift?

 

[notaz]

I've tried it with 2 pandoras paired over BT (one running Master Control) and it worked fine (didn't reproduce).

 

[Neelix]

It's one of those nasty bugs that strikes somewhat at random.   You sometimes need to keep a keyboard connected via bluetooth for several hours before it will present itself.

It took me a while before I twigged that it was the keyboard causing it,  because I it didn't seem to have a direct causal relationship on the surface,   but It didn't happen at all before I used the BT keyboard, and it stopped happening when I stopped using the BT keyboard.  (with the Pandora that is...  I'm using the same BT keyboard itself for extended periods on another machine (running Ubuntu Lucid 10.04 LTS)  without any issues)

- Neelix