Hi,
this time I've attached a complete ksymoops output
for the error i am encountering again and again. The
error is in a small hack for the linux kernel's nfs
module which was working just fine before i added the
following code which works till a point.
static void nfs_print_path(struct dentry *d) {
struct dentry *parent;
struct qstr *qs;
char name[64];
struct inode *inode_parent, *inode;
void *p, *me;
if(!d) {
return;
}
parent = d->d_parent;
qs = &d->d_name;
if(parent) {
inode_parent = parent->d_inode;
inode = d->d_inode;
p = (void *)inode_parent;
me = (void *)inode;
// Till here things work just fine. I am DEAD SURE of
that as i put printk()
// followed by return here and there and checked.
// My analysis with printk's and return's shows that
the next statement, or somewhere
// after that is the problem. ksymoops identifies the
fault triggering instrxn as
// 'cmp' (see output below) which (??) could be for
this statement.
if( p - me != 0 ) {
printk("return 3\n");
return;
}
KSYMOOPS OUTPUT:
<1>Unable to handle kernel NULL pointer dereference at
virtual address 0000000f
c88b6956
*pde = 00000000
Oops: 0000
CPU: 0
EIP: 0010:[<c88b6956>]
Not tainted
Using defaults from
ksymoops -t elf32-i386 -a i386
EFLAGS: 00010202
eax: 00000006 ebx:
c6675024 ecx: 00000001 edx:
00000007
esi: 429d3663 edi:
0000005d ebp: c6598cdc esp:
c5b21f08
ds: 0018 es: 0018 ss:
0018
Process ls (pid: 1996,
stackpage=c5b21000)
Stack: 00000246 ffffffd2
c011b91b 0001ea92 0001ea96
00000282 0001ea96
0001ea8f
00000001 00000282
00000001 c033c964 00000000
00000004 c6598cdc
c011bb41
c6675044 429d3663
c6675044 c88b80e2 c6675024
bffff930 c014f67e
c71bf000
Call Trace: [<c011b91b>]
[<c011bb41>] [<c88b80e2>]
[<c014f67e>] [<c0151111>]
[<c014cd04>]
[<c010775c>] [<c010766b>]
Code: 39 42 08 74 0d 83
ec 0c 68 2c 81 8c c8 eb 0b 8d
76 00 83 ec
>>EIP; c88b6956
<[nfs]nfs_print_path+2e/58> <=====
Trace; c011b91b
<call_console_drivers+eb/100>
Trace; c011bb41
<printk+1a1/1f0>
Trace; c88b80e2
<[nfs]nfs_revalidate+116/1b0>
Trace; c014f67e
<getname+5e/a0>
Trace; c0151111
<__user_walk+41/50>
Trace; c014cd04
<sys_lstat64+34/70>
Trace; c010775c
<error_code+34/3c>
Trace; c010766b
<system_call+33/38>
Code; c88b6956
<[nfs]nfs_print_path+2e/58>
00000000 <_EIP>:
Code; c88b6956
<[nfs]nfs_print_path+2e/58> <=====
0: 39 42 08
cmp
%eax,0x8(%edx) <=====
Code; c88b6959
<[nfs]nfs_print_path+31/58>
3: 74 0d
je 12 <_EIP+0x12>
c88b6968
<[nfs]nfs_print_path+40/58>
Code; c88b695b
<[nfs]nfs_print_path+33/58>
5: 83 ec 0c
sub $0xc,%esp
Code; c88b695e
<[nfs]nfs_print_path+36/58>
8: 68 2c 81 8c c8
push $0xc88c812c
Code; c88b6963
<[nfs]nfs_print_path+3b/58>
d: eb 0b
jmp 1a <_EIP+0x1a>
c88b6970
<[nfs]nfs_print_path+48/58>
Code; c88b6965
<[nfs]nfs_print_path+3d/58>
f: 8d 76 00
lea 0x0(%esi),%esi
Code; c88b6968
<[nfs]nfs_print_path+40/58>
12: 83 ec 00
sub $0x0,%esp
so whats so wrong. as per what ksymoops says, the
final error is triggered by a 'cmp' instruction: a
mere register operation.
if this is found insufficient could u tell me some
better kernel debugger(i use only ksymoops).
regards
abhishek
__________________________________________________
Do you Yahoo!?
The New Yahoo! Search - Faster. Easier. Bingo
http://search.yahoo.com
-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
http://thinkgeek.com/sf
_______________________________________________
NFS maillist - [email protected]
https://lists.sourceforge.net/lists/listinfo/nfs
On Wed, 23 Apr 2003, Abhishek Rai wrote:
> virtual address 0000000f
is exactly...
> c6675024 ecx: 00000001 edx: 00000007
... EDX + 8 which is obtained here:
> 0: 39 42 08 cmp %eax,0x8(%edx) <=====
> ... a 'cmp' instruction: a mere register operation.
This statement means that you do not understand asm syntax properly.
Placing a register between paranthesis ( "(" and ")" ) is like
dereferencing a pointer in C - access the data that is at the address
found in the register; the 0x8(%edx) notation means 0x8 bytes offset
from the address in EDX.
If you want live debugging, you could try kdbg...
--
Bogdan Costescu
IWR - Interdisziplinaeres Zentrum fuer Wissenschaftliches Rechnen
Universitaet Heidelberg, INF 368, D-69120 Heidelberg, GERMANY
Telephone: +49 6221 54 8869, Telefax: +49 6221 54 8868
E-mail: [email protected]
-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
http://thinkgeek.com/sf
_______________________________________________
NFS maillist - [email protected]
https://lists.sourceforge.net/lists/listinfo/nfs