Tiny Core Linux
Tiny Core Extensions => TCE Bugs => Topic started by: tinypoodle on January 01, 2011, 12:00:07 PM
-
Plugging in a USB WLAN adapter after having loaded wireless-2.6.29.1-tinycore_mod.tcz for the first time ever caused a kernel Oops here, resulting in leaving all USB ports in a totally unresponsive state.
As this is a module extension for 2.x it might not be worth to further look into it, but IMHO even a reported one time occurrence of a bug of such seriousness regarding consequences may be worth a warning in the .info file - already marked as "TESTING" - as long as the extension is still part of repo - even if that is of 2.x
relevant syslog output attached
-
Reproducible!
1. Boot with 'base' 'syslog'
2. 'tail -f /var/log/messages'
2. 'tce-load -i wireless-2.6.29.1-tinycore_mod.tcz'
3. 'tce-load -i wireless'_tools.tcz'
4. Plugging in a USB WLAN adaptor
--> instant Oops
- no way to rmmod the respective modules
- 'ifconfig -a' and 'iwconfig' not showing any wireless interface
- USB ports in a totally unresponsive state.
-
What exact version are you using?
There was a library change from 2.9 to 2.10
-
2.10 with extension as current in 2.x repo.
-
Despite searching, I somehow was unable to find the source for wireless-2.6.29.1-tinycore_mod.tcz.
Any pointers please?
-
I suggest you pm jpeters
-
I suggest you pm jpeters
Last Active: January 05, 2010, 02:16:35 AM
::)
-
Well, the info file does mention it's the 2.6.30 wireless stack backported. But instead of trying to fix that in that version, how about migrating to 3.x? It's very likely your bug has been solved since.
-
curaga,
do you mean to imply there is a high probability that the Oops is caused by being a backport per se?
As the info refers to patches and particularly to atheros, I had in mind examining to which degree the patches have effect to specific modules loaded - as can be seen in syslog attached to earlier post - in this case here, that's why I would like to get a hold of the source (at least the patches used).
As implied in my first post my aim is not to fix this version, but rather to prevent it's use in cases where there is more potential risk than expectable benefit.
One approach could be to remove at least the rt2500 module from the extension, or add a warning to the .info file.
BTW, overall I am better off personally with 2.x regarding wireless, as rt2500usb module in wireless-2.6.29.1-tinycore.tcz has never ever caused any such issues, and AFAIK modules for rtl8187SE are only available for 2.x, as opposed to 3.x.
-
No, I meant that there's a bug in that version, and since it seems to be an easily reproduced oops, it's likely fixed in a later version.
-
No, I meant that there's a bug in that version, and since it seems to be an easily reproduced oops, it's likely fixed in a later version.
Oh, that is a very plausible estimation and had crossed my mind too, and was what prevented me from looking into filing a bug report upstream.
Making such a (likely) assumption would IMHO even more justify to remove the concerned driver from the extension, or at least adding a specific warning to the .info, so that users of that driver (rt2500usb), so they would use the other available wireless extension which doesn't create such an issue.
(I have PMed jpeters, but not much hope of getting any reaction soon)
-
About my only contribution to this extension (as I recall) was to tidy up the module locations to it matched the tc norm. I don't have the hardware to test, so if anybody wants to update the extension that's fine with me.
-
I suggest you pm jpeters
Last Active: January 05, 2010, 02:16:35 AM
::)
Hi Juanito,
The extension worked fine when it was uploaded..I used it myself. I've moved on, so perhaps you can check into current compatibity issues? Thanks
-
First of all, thank you jpeters for the - not so expected by me - swift reply, despite having "moved on", highly appreciated.
The only opportunity I could find to test the extension with different hardware was with an orinoco chipset PCMCIA wifi card using the orinoco_cs driver.
Indeed with some limited testing everyting appeared to work fully as expected in this case.
A suggestion would be to add something along the lines of this to the .info file:
Warning: This extension has been reported to cause a reproducible kernel oops while loaded upon plugging in a WLAN adaptor making use of the rt2500usb driver module (Ralink USB chipset); while working well with other chipsets/drivers.
For affected wireless chipsets please prefer the wireless-2.6.29.1-tinycore.tcz extension instead.