| > | Please let me know in case that you want me to send you the configs and strace log files.
| > Yes, please do -- on the computers I have tried (32/64 bit) it works.
| >
| > Gerrit
| >
|
| Hello Gerrit,
| here they are!
Thanks very much.
In the error log the EUNATCH appears within the call to socket, before any
other call (connect, listen). This means that DCCP is not able to load the
CCID modules at initialisation time.
I have compared the two .config files, below are a few ideas. I tested
by adding/disabling configuration options, then unloading all modules and
starting 'ipdf -sd'. If this loads the CCID modules, then everything is ok.
1. the Pentium4 has no CONFIG_HIBERNATION set, while the K8 (which has
the module load error) has it set; adding this feature did not cause
an error to appear;
2. the K8 has no IPv6 support enabled -- I built a kernel without IPv6
and it still worked ok;
3. could it have to do with security ?:- the Pentium (which works)
has no security enabled, while the K8 (which doesn't work) has
CONFIG_SECURITY=y;
4. a thing which I could not test was CONFIG_USB_NET, which was enabled
on the K8.
Unfortunately, on the Centrino-Duo I could not provoke a faulty behaviour.
Since 2.6.27 just came out, would it be a lot of effort to try again
with the more recent kernel and see if the problem persists?
Incidentally the new netdev-2.6 has a commit called "Remove CONFIG_KMOD from
net/ (towards removing KMOD entirely)", so perhaps there is something in
progress.
I have uploaded the new test tree, after changing the patch "Auto-load
(when supported) CCID plugins for negotiation"
http://eden-feed.erg.abdn.ac.uk/cgi-bin/gitweb.cgi?p=dccp_exp.git;a=commitdiff;h=56bcfc75b874d27039c1a26055669dd8c69c57b7
Gerrit
--
To unsubscribe from this list: send the line "unsubscribe dccp" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
