Neal Katz wrote:
>
> The entry.S file looks okay from what you described, I've appended mine at
> the end of the message anyhow. Double checked linux/.config and AIO is
> selected, even dumped symbols from fs/read_write.o and verified that
> sys_aio was compiled. Everything looked good. I'm including output from
> "strace ./a.out", maybe that will help.
>
> I did join the mailing list, today. Didn't see any similar problems
> mentioned in the archive.
Hmm. I don't see anything wrong with the entry.S either.
Ok, here's another thing to check. Run gdb on your vmlinux
(btw, I hope you're running the right kernel!):
----------
% gdb ./vmlinux
(gdb) p /x &sys_call_table
$6 = 0x802edeac
(gdb) p /x 0x802edeac + (201 * 4)
$7 = 0x802ee1d0
(gdb) x /x 0x802ee1d0
0x802ee1d0 <sys_call_table+804>: 0x8012bdb8
(gdb) disas 0x8012bdb8
Dump of assembler code for function sys_aio:
0x8012bdb8 <sys_aio>: push %ebp
0x8012bdb9 <sys_aio+1>: mov %esp,%ebp
[ ... ]
----------
Of course, the absolute hex numbers above are for my kernel
and won't be the same in your vmlinux. But basically, I'm
trying to see what's in the sys_call_table at 201'st entry.
thanks,
ananth.
PS: I'm sending this to kaio just in case someone else ran
into the problem, but didn't report it on the list.
|