Oops 2.6.14-rc1, AMD Sempton

Oops 2.6.14-rc1, AMD Sempton

[Joost Witteveen]

The Oops below happend while swsuspending my Sempron laptop. The system 
freezes afterwards.

On 2.6.12.5 this happened very rarely (but when I had all modules 
loaded, it happend more often, so I wrote a little script around the 
swsuspend, removing most modules before the suspend)

I noticed that the freezes occur somewhat more often on 2.6.13, 
2.6.13.1; but on 2.6.14-rc1 it seems the freeze happens just about 
always, even with most modules unloaded.

.config, System.map and the .jpg of the photo I took of the oops can be 
found at:

http://foko.komputilo.org/~joostje/linux/

Stopping tasks: ------------------------------------------------|
Freeing memory... done (25044 pages freed)
swsusp: Need to copy 29618 pages
Error 12 suspending
-----------[ cut here ]--------------
kernel BUG <bad filename>:8680!
invalid operand: 0000 [#1]
PREEMPT
Modules linked in: powernow_k8 processor
CPU:    0
EIP:    0060:[<c0130361>]    Not tainted VLI
EFLAGS: 00010093   (2.6.14-rc1)
EIP is at swsusp_suspend+073/0x92
eax: 00007427   ebx: fffffff4  ecx: 00000c00  edx: c04d52a0
esi: 00000004   edi: c03c9d89  ebp: c03f4ec4  esp: c0e13f04
ds: 007b   es: 007b   ss: 0068
Process sh (pid: 5324, threadinfo=c0e12000 task=c0ccd030)
Stack: 00000000 c0130c1f fffffff0 c012f1fd 00000005 c9114004 c012f311 
00000004
       00000004 cd19d900 cd19d914 c9f0bbc0 c0e13fa4 c017f7ec c03f4ee0 
c9114000
       00000005 00000005 c017f82c ccea5090 cd19d900 00000005 00000005 
c9f0bbc0
Call Trace:
 [<c0130c1f>] pm_suspend_disk+0x3b/0x7f
 [<c012f1fd>] enter_state+0x27/0x59
 [<c012f311>] state_store+0x7f/0x8b
 [<c017f7ec>] flush_write_buffer+0x24/0x28
 [<c017f82c>] sysfs_write_file+0x3c/0x5b
 [<c014be06>] vfs_write+0xa1/0x138
 [<c014bf3c>] sys_write+0x3b/0x64
 [<c0102ca9>] syscall_call+0x7/0xb
Code: 00 e8 a4 88 1b 00 85 c0 89 c3 74 0d 50 68 f3 25 3a c0 e8 78 76 fe 
ff 58 5a e8 7a 88 1b 00 a1 00 b0 45 c0 39 05 08 b0 4b c0 74 02 <0f> 0b 
e8 21 b4 15 00 fb eb 11 fa 6a 01 e8 9d b1 15 00 89 c3 85
 <7> APIC error on CPU0: 80(08)

---------------------------
The text above is hand-copied from the photo I took, there may be erros 
in it, the real photo is here:
http://foko.komputilo.org/~joostje/linux/swsuspend-2.6.14-rc1-1.jpg

-- 
Groetjes
joostje
47d3fcfe28f2a83497e79d9bc7d5087c-4a1ee1fddab4648175518cb4c1c9edb3ed0e89f0

Re: Oops 2.6.14-rc1, AMD Sempton

[Joost Witteveen]

Joost Witteveen wrote:

> The Oops below happend while swsuspending my Sempron laptop. The 
> system freezes afterwards.

OK, replying to my own email as I haven't had any reply yet (I'm not 
subscribed to the list).

Since I sent the email, I've had two more oopses (they don't happen as 
often as before), which I've added to:

http://www.komputilo.org/~joostje/linux/swsuspend/

I haven't transcribed them, as I haven't had a responce to my first 
email -- if I could help transcribing the other oopses, please ask!

Also, if I can help any other way, I'd be hapy to try whatever you 
suggest doing.

-- 
Groetjes
joostje
47d3fcfe28f2a83497e79d9bc7d5087c-4a1ee1fddab4648175518cb4c1c9edb3ed0e89f0

Re: Re: Oops 2.6.14-rc1, AMD Sempton

[R. J. Wysocki]

Hi,

On Monday, 19 of September 2005 17:44, Joost Witteveen wrote:
> Joost Witteveen wrote:
> 
> > The Oops below happend while swsuspending my Sempron laptop. The 
> > system freezes afterwards.
> 
> OK, replying to my own email as I haven't had any reply yet (I'm not 
> subscribed to the list).
> 
> Since I sent the email, I've had two more oopses (they don't happen as 
> often as before), which I've added to:
> 
> http://www.komputilo.org/~joostje/linux/swsuspend/

There sometimes is not enough memory to suspend on your box, and
the BUG() triggers because nr_copy_pages_check is initialized too late.
The following patch should fix it.  Could you please test?

Greetings,
Rafael 


Index: linux-2.6.14-rc1/kernel/power/swsusp.c
===================================================================
--- linux-2.6.14-rc1.orig/kernel/power/swsusp.c	2005-09-22 22:05:52.000000000 +0200
+++ linux-2.6.14-rc1/kernel/power/swsusp.c	2005-09-22 22:08:14.000000000 +0200
@@ -918,6 +918,7 @@
 
 	pagedir_nosave = NULL;
 	nr_copy_pages = calc_nr(nr_copy_pages);
+	nr_copy_pages_check = nr_copy_pages;
 
 	pr_debug("suspend: (pages needed: %d + %d free: %d)\n",
 		 nr_copy_pages, PAGES_FOR_IO, nr_free_pages());
@@ -940,7 +941,6 @@
 		return error;
 	}
 
-	nr_copy_pages_check = nr_copy_pages;
 	return 0;
 }
 

-- 
Rafael J. Wysocki, Ph.D.
Institute of Theoretical Physics
Faculty of Physics of Warsaw University
ul. Hoza 69, 00-681 Warsaw
[tel: +48 22 55 32 263]
[mob: +48 60 50 53 693]
----------------------------
One should not increase, beyond what is necessary,
the number of entities required to explain anything.
			-- William of Ockham