[OpenAFS] Linux 2.6 support

[Horst Birthelmer]

On Wednesday, June 9, 2004, at 09:45  AM, Miroslav Ruda wrote:

> Horst Birthelmer wrote:
>>> /usr/vice/etc/afsd -debug -verbose
>>> afsd: My home cell is 'ics.muni.cz'
>>> ParseCacheInfoFile: Opening cache info file 
>>> '/usr/vice/etc/cacheinfo'...
>>> ParseCacheInfoFile: Cache info file successfully parsed:
>>> cacheMountDir: '/afs'
>>> cacheBaseDir: '/usr/vice/cache'
>>> cacheBlocks: 40000
>>> afsd: 4000 inode_for_V entries at 0x550710, 16000 bytes
>>> SScall(183, 28, 17)=0 afsd: Forking rx listener daemon.
>>> afsd: Forking rx callback listener.
>>> afsd: Forking rxevent daemon.
>>> SScall(183, 28, 19)=0 SScall(183, 28, 0)=0
>>>
>> ... and you're sure there is no other output like "SScall....=..." 
>> after that??
>> What the client does is initializing rx, forking the callback and ... 
>> that's it.
>> It says "afsd: Forking rxevent daemon." but form the syscalls output 
>> that never happened.
>> I've never seen anything like that before.
>
> I'm pretty sure that nothing else was logged - it's still blocking one 
> of my xterms :-)
> Nothing in logs, dmesg ...
>
> I'll send you strace of afsd to your email.
>

That could be almost anything ;-)
If my assumptions are right and it really gets stuck in the syscall for 
initializing the rxevent deamon then it could be a deadlock or 
something similar ...

The syscall is waiting for afs_initState < AFSOP_START_BKG. Before that 
line it just deals with priority and such ...
Now I don't know if it gets there.

That's the most suspicious line to me for those symptoms. The rest of 
it is starting the deamon itself (actually it can get stuck there, too, 
but I don't think that'll be the case).

Could you enter some debug outputs there?? or don't you want to alter 
the code??

Horst