LD_ASSUME_KERNEL does not fix the problem ...

Here is what I get when I try to migrate a very simple process :


fdy@fuyutsuki:/opt/openmosix/cluster/utils/eatcpu$ 
LD_ASSUME_KERNEL=2.4.1 ./eatcpu &
[1] 5673
fdy@fuyutsuki:/opt/openmosix/cluster/utils/eatcpu$ echo 192.168.168.11 > 
/proc/5673/om/where

(eatcpu is a dump program that just loops forever doing some square root 
calculations) ...

I get an Oops :

<1>Unable to handle kernel NULL pointer dereference at 0000000000000000 RIP:
<ffffffff802fc67c>{proc_pid_set_where+214}
PGD 13c817067 PUD 13cc8f067 PMD 0
Oops: 0002 [2]

CPU 0
Modules linked in: kcomd
Pid: 5266, comm: bash Not tainted 2.6.17-om #8
RIP: 0010:[<ffffffff802fc67c>] <ffffffff802fc67c>{proc_pid_set_where+214}
RSP: 0018:ffff81013c8bbe58  EFLAGS: 00010246
RAX: 0000000000000000 RBX: ffff81013f4c9140 RCX: 0000000000000088
RDX: 0000000000000068 RSI: ffffffff8045be59 RDI: ffff81013c9d8001
RBP: ffff81013c8bbe78 R08: 0000000000000068 R09: 0000000000000004
R10: 0000000000000031 R11: 0000000000000002 R12: 0000000000000000
R13: ffff81013c9d8000 R14: 000000000000000f R15: ffff81013f4c9140
FS:  00002b49e28ea6d0(0000) GS:ffffffff80562000(0000) knlGS:0000000000000000
CS:  0010 DS: 0000 ES: 0000 CR0: 000000008005003b
CR2: 0000000000000000 CR3: 000000013cd87000 CR4: 00000000000006e0
Process bash (pid: 5266, threadinfo ffff81013c8ba000, task ffff81013ec97890)
Stack: ffffffff804bbcc0 000000000000000f ffff81013ca7665c ffff81013c9d8000
       ffff81013c8bbeb8 ffffffff802fc16c fffffffffffffff4 ffff81013c9d8000
       fffffffffffffff2 000000000000000f
Call Trace: <ffffffff802fc16c>{openmosix_proc_pid_setattr+68}
       <ffffffff802979e1>{proc_pid_openmosix_write+143} 
<ffffffff80263666>{vfs_w rite+171}
       <ffffffff80263b08>{sys_write+71} <ffffffff8020a2ba>{syscall_call+10}

Code: 66 41 c7 04 24 02 00 66 41 c7 44 24 02 0b 55 4c 89 ef e8 2b
RIP <ffffffff802fc67c>{proc_pid_set_where+214} RSP <ffff81013c8bbe58>
CR2: 0000000000000000

Looking at the Oops log, it seems that the crash occurs in the current() 
call (KnlGS seems corrupted == 0x0). I am trying to fix the problem, but 
if you have an idea, I would greatly appreciate.

BTW: Matt, what prevents kcomd to be built into the kernel ? I am 
(trying?) to use kgdb to debug openmosix, but I really need to compile 
every thing in the kernel (kgdb is not really skilled at debugging 
modules yet).



-------------------------------------------------------------------------
Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel and you'll get the chance to share your
opinions on IT & business topics through brief surveys -- and earn cash
http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV