WLAN: known issues and how to help - uptime test - new record!
Philip Rhoades
phil at pricom.com.au
Wed Feb 11 13:32:15 CET 2009
Werner,
Werner Almesberger wrote:
> Philip Rhoades wrote:
>> I then setup wireless networking and run my script for testing LAN
>> connection - it has now been running for nearly 26 hours! (and is still
>> going) - about 10 hours longer than the previous record!
>
> Cool !
It finally failed at about 30 hours.
>> The question is was this just a lucky test
>
> I'd say, once we see a 10x difference to your average update (that
> would be something like 10 hours ?), then I'll slowly start to
> believe it's not just Murphy being busy to bring mischief to
> someone else ;-)
>
> Is the rest of the system behaving normally ?
When I am running the test I am usually not using the FR for anything
else - maybe just some terminal stuff.
>> that is allowing the much improved uptime?
>
> Hmm, there are three charger changes to WLAN:
>
> - SIOCGIWSCAN no longer returns an error if it doesn't find an AP.
> That may keep user space from unnecessarily entering panic mode.
>
> - the addition of rfkill. Don't see how this one could affect it,
> though.
>
> - the "shot in the dark" for SDIO communication. If this does
> anything, the kernel prints either
>
> READ WHILE BUSY !
> or
> WRITE WHILE BUSY !
>
> Do you see any of these in dmesg ?
Oddly, after it had failed, I restarted the network and it failed again
in about a minute. I then started it again and copied the attached
files over FYI - no "BUSY" in dmesg output . .
Phil.
--
Philip Rhoades
GPO Box 3411
Sydney NSW 2001
Australia
E-mail: phil at pricom.com.au
-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
Name: dmesg.txt
Url: http://lists.openmoko.org/pipermail/devel/attachments/20090211/7c12012e/attachment.txt
-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
Name: tst_wlan.txt
Url: http://lists.openmoko.org/pipermail/devel/attachments/20090211/7c12012e/attachment-0001.txt
More information about the devel
mailing list