[2.5.64-mm1] sysfs oops

[2.5.64-mm1] sysfs oops

[Alistair Strachan]

[-- Attachment #1: Type: text/plain, Size: 370 bytes --]

Hi,

The following was experienced while running Andrew's 2.5.64-mm1 kernel. My 
dialup (ppp) connection was terminated, and when I attempted to re-establish, 
pppd would not come up. I found the following oops in dmesg and have provided 
a decoded version below.

This doesn't appear to be fixed in mainline, yet. Is it mm specific? dcache?

Cheers,
Alistair Strachan.

[-- Attachment #2: ksymoops --]
[-- Type: text/plain, Size: 2951 bytes --]

ksymoops 2.4.8 on i686 2.5.64-mm1.  Options used
     -V (default)
     -k /proc/ksyms (default)
     -l /proc/modules (default)
     -o /lib/modules/2.5.64-mm1/ (default)
     -m /boot/System.map-2.5.64-mm1 (default)

Warning: You did not tell me where to find symbol information.  I will
assume that the log matches the kernel and modules that are running
right now and I'll use the default options above for symbol resolution.
If the current kernel and/or modules do not match the log, you can get
more accurate output by telling me the kernel version and where to find
map, modules, ksyms etc.  ksymoops -h explains the options.

Error (regular_file): read_ksyms stat /proc/ksyms failed
ksymoops: No such file or directory
No modules in ksyms, skipping objects
No ksyms, skipping lsmod
kernel BUG at include/linux/dcache.h:266!
invalid operand: 0000
CPU:    0
EIP:    0060:[<c01759fc>]    Tainted: PF
Using defaults from ksymoops -t elf32-i386 -a i386
EFLAGS: 00010246
eax: 00000000   ebx: dc6d9184   ecx: 00000000   edx: 00000000
esi: d7b9d300   edi: c959a000   ebp: 00000000   esp: c959bed8
ds: 007b   es: 007b   ss: 0068
Stack: c02b0fc1 dc6d9000 c03455f4 dc6d9184 dc6d9000 c959a000 00000000 c01afb73
       dc6d9184 dc6d9184 c01afba3 dc6d9184 00000000 c026b64d dc6d9184 00000006
       dc6d9000 00000000 dc6d9000 dfd2b680 c0334600 c01f0bbd dc6d9000 cd57d440
Call Trace: [<c02b0fc1>]  [<c01afb73>]  [<c01afba3>]  [<c026b64d>]  [<c01f0bbd>]  [<c01ee67c>]  [<c01edb06>]  [<c0149e68>]  [<c0158e45>]  [<c0148700>]  [<c01090c7>]
Code: 0f 0b 0a 01 9e 23 2e c0 ff 06 83 4e 04 08 85 f6 0f 84 02 01


>>EIP; c01759fc <sysfs_remove_dir+1c/13c>   <=====

Trace; c02b0fc1 <addrconf_notify+71/e0>
Trace; c01afb73 <kobject_del+13/30>
Trace; c01afba3 <kobject_unregister+13/30>
Trace; c026b64d <unregister_netdevice+1bd/295>
Trace; c01f0bbd <ppp_shutdown_interface+ad/120>
Trace; c01ee67c <ppp_ioctl+7cc/800>
Trace; c01edb06 <ppp_release+46/70>
Trace; c0149e68 <__fput+98/e0>
Trace; c0158e45 <sys_ioctl+c5/250>
Trace; c0148700 <sys_close+50/60>
Trace; c01090c7 <syscall_call+7/b>

Code;  c01759fc <sysfs_remove_dir+1c/13c>
00000000 <_EIP>:
Code;  c01759fc <sysfs_remove_dir+1c/13c>   <=====
   0:   0f 0b                     ud2a      <=====
Code;  c01759fe <sysfs_remove_dir+1e/13c>
   2:   0a 01                     or     (%ecx),%al
Code;  c0175a00 <sysfs_remove_dir+20/13c>
   4:   9e                        sahf   
Code;  c0175a01 <sysfs_remove_dir+21/13c>
   5:   23 2e                     and    (%esi),%ebp
Code;  c0175a03 <sysfs_remove_dir+23/13c>
   7:   c0 ff 06                  sar    $0x6,%bh
Code;  c0175a06 <sysfs_remove_dir+26/13c>
   a:   83 4e 04 08               orl    $0x8,0x4(%esi)
Code;  c0175a0a <sysfs_remove_dir+2a/13c>
   e:   85 f6                     test   %esi,%esi
Code;  c0175a0c <sysfs_remove_dir+2c/13c>
  10:   0f 84 02 01 00 00         je     118 <_EIP+0x118>


1 warning and 1 error issued.  Results may not be reliable.

Re: [2.5.64-mm1] sysfs oops

[Patrick Mochel]

On Thu, 6 Mar 2003, Alistair Strachan wrote:

> Hi,
> 
> The following was experienced while running Andrew's 2.5.64-mm1 kernel. My 
> dialup (ppp) connection was terminated, and when I attempted to re-establish, 
> pppd would not come up. I found the following oops in dmesg and have provided 
> a decoded version below.
> 
> This doesn't appear to be fixed in mainline, yet. Is it mm specific? dcache?

Argh. A bk merge ate my change!

This has been around for a while, and I merged a patch to fix it. However, 
a leter merge with previous work to split fs/sysfs/inode.c up reverted it 
back to the old code. 

Could you (both) try the following patch and let me know if it fixes it?

Thanks,

	-pat

===== fs/sysfs/dir.c 1.2 vs edited =====
--- 1.2/fs/sysfs/dir.c	Thu Jan 30 17:10:06 2003
+++ edited/fs/sysfs/dir.c	Thu Mar  6 14:34:01 2003
@@ -71,7 +71,7 @@
 
 void sysfs_remove_dir(struct kobject * kobj)
 {
-	struct list_head * node, * next;
+	struct list_head * node;
 	struct dentry * dentry = dget(kobj->dentry);
 	struct dentry * parent;
 
@@ -83,31 +83,28 @@
 	down(&parent->d_inode->i_sem);
 	down(&dentry->d_inode->i_sem);
 
-	list_for_each_safe(node,next,&dentry->d_subdirs) {
-		struct dentry * d = dget(list_entry(node,struct dentry,d_child));
-		/** 
-		 * Make sure dentry is still there 
-		 */
-		pr_debug(" o %s: ",d->d_name.name);
-		if (d->d_inode) {
+	spin_lock(&dcache_lock);
+	node = dentry->d_subdirs.next;
+	while (node != &dentry->d_subdirs) {
+		struct dentry * d = list_entry(node,struct dentry,d_child);
+		list_del_init(node);
 
+		pr_debug(" o %s (%d): ",d->d_name.name,atomic_read(&d->d_count));
+		if (d->d_inode) {
+			d = dget_locked(d);
 			pr_debug("removing");
+
 			/**
 			 * Unlink and unhash.
 			 */
-			simple_unlink(dentry->d_inode,d);
+			spin_unlock(&dcache_lock);
 			d_delete(d);
-
-			/**
-			 * Drop reference from initial sysfs_get_dentry().
-			 */
+			simple_unlink(dentry->d_inode,d);
 			dput(d);
+			spin_lock(&dcache_lock);
 		}
-		pr_debug(" done (%d)\n",atomic_read(&d->d_count));
-		/**
-		 * drop reference from dget() above.
-		 */
-		dput(d);
+		pr_debug(" done\n");
+		node = dentry->d_subdirs.next;
 	}
 
 	up(&dentry->d_inode->i_sem);

Re: [2.5.64-mm1] sysfs oops

[Alistair Strachan]

On Thursday 06 March 2003 20:38, Patrick Mochel wrote:
>
> Argh. A bk merge ate my change!
>

Hmm, I recall reporting this before, yes.

> This has been around for a while, and I merged a patch to fix it. However,
> a leter merge with previous work to split fs/sysfs/inode.c up reverted it
> back to the old code.
>
> Could you (both) try the following patch and let me know if it fixes it?

This fixes it. Thanks for the reply.

Cheers,
Alistair.