Home / Browse / ndiswrapper / Mail / Archive

ndiswrapper

Email Archive: ndiswrapper-general (read-only)

[Ndiswrapper-general] Re: ndiswrapper-snap-200401020157.tar.gz

From: Tomas Szepe <szepe@pi...> - 2004-01-02 00:49

On Jan-02 2004, Fri, 02:33 +0200
Kaj J. Niemi <kaj.niemi@...> wrote:

> Whoops, seems like I missed a log file. Here's what really happened 
> (when the keyboard locked up):

...

> Jan  2 00:44:27 aurora kernel: ndiswrapper version 0.3+CVS loaded

...

> Jan  2 00:45:05 aurora kernel: Unable to handle kernel paging request at virtual address df712f50
> Jan  2 00:45:05 aurora kernel:  printing eip:
> Jan  2 00:45:05 aurora kernel: fcf4c089
> Jan  2 00:45:05 aurora kernel: *pde = 0007d067
> Jan  2 00:45:05 aurora kernel: Oops: 0000 [#1]
> Jan  2 00:45:05 aurora kernel: CPU:    0
> Jan  2 00:45:05 aurora kernel: EIP:    0060:[<fcf4c089>]    Tainted: P
> Jan  2 00:45:05 aurora kernel: EFLAGS: 00010292
> Jan  2 00:45:05 aurora kernel: EIP is at send_one+0x139/0x160 [ndiswrapper]
> Jan  2 00:45:05 aurora kernel: eax: 00000001   ebx: df712ef8   ecx: 00000001   edx: e81e9394
> Jan  2 00:45:05 aurora kernel: esi: e81e91b0   edi: 00000001   ebp: d3471240   esp: f7f4ff24
> Jan  2 00:45:05 aurora kernel: ds: 007b   es: 007b   ss: 0068
> Jan  2 00:45:05 aurora kernel: Process events/0 (pid: 3,  threadinfo=f7f4e000 task=f7fc4960)
> Jan  2 00:45:05 aurora kernel: Stack: df712ef8 00000283 d3471240 00000286 e81e91b0 fcf4c1fc d3471240 e81e91b0
> Jan  2 00:45:05 aurora kernel:        00000001 eedf9960 eedf9980 0000018f 51780f46 d3471630 00000283 d34715e4
> Jan  2 00:45:05 aurora kernel:        d3471240 f7f87ef8 c0141086 d3471240 f7f4ffa0 00000000 5a5a5a5a 5a5a5a5a
> Jan  2 00:45:05 aurora kernel: Call Trace:
> Jan  2 00:45:05 aurora kernel:  [<fcf4c1fc>] xmit_bh+0x14c/0x2b0 [ndiswrapper]
> Jan  2 00:45:05 aurora kernel:  [<c0141086>] worker_thread+0x246/0x4a0
> Jan  2 00:45:05 aurora kernel:  [<fcf4c0b0>] xmit_bh+0x0/0x2b0 [ndiswrapper]
> Jan  2 00:45:05 aurora kernel:  [<c0124d50>] default_wake_function+0x0/0x20
> Jan  2 00:45:05 aurora kernel:  [<c010c662>] ret_from_fork+0x6/0x14
> Jan  2 00:45:05 aurora kernel:  [<c0124d50>] default_wake_function+0x0/0x20
> Jan  2 00:45:05 aurora kernel:  [<c0140e40>] worker_thread+0x0/0x4a0
> Jan  2 00:45:05 aurora kernel:  [<c01092a5>] kernel_thread_helper+0x5/0x10
> Jan  2 00:45:05 aurora kernel:
> Jan  2 00:45:05 aurora kernel: Code: 8b 43 58 eb a3 8b 46 0c 89 04 24 e8 c7 e6 20 c3 89 34 24 e8

Giri, has anything changed in send_one() or xmit_bh() recently?

Kaj, do you have kernel preemption enabled?  Can you retry with a vanilla
kernel.org tree -- 2.6.1-rc1 preferably?  I can't seem to reproduce
this problem (identical hardware, identical gcc, identical ndis driver).

-- 
Tomas Szepe <szepe@...>

Thread View

Thread	Author	Date
[Ndiswrapper-general] Re: ndiswrapper-snap-200401020157.tar.gz	Tomas Szepe <szepe@pi...>

[Ndiswrapper-general] Re: ndiswrapper-snap-200401020157.tar.gz

From: Kaj J. Niemi <kaj.niemi@ba...> - 2004-01-02 01:04

Tomas Szepe wrote:

> Kaj, do you have kernel preemption enabled?  Can you retry with a vanilla
> kernel.org tree -- 2.6.1-rc1 preferably?  I can't seem to reproduce
> this problem (identical hardware, identical gcc, identical ndis driver).

According to 
<http://www.redhat.com/archives/fedora-devel-list/2003-December/msg00543.html> 
the Rawhide 2.6.0 kernels so far do *not* have pre-emption enabled, just 
patches for low latency scheduling.

I run vendor provided kernels 'cause I'm a lazy bum (and they happen to 
work) ;-) It's a pain in the ass to package a vanilla kernel for 
installation.

I'll try 2.6.1 rc1 in a second.

// kaj

[Ndiswrapper-general] Re: ndiswrapper-snap-200401020157.tar.gz

From: Kaj J. Niemi <kaj.niemi@ba...> - 2004-01-02 01:26

Kaj J. Niemi wrote:

> I'll try 2.6.1 rc1 in a second.

CONFIG_REEMPT is not set in either one (2.6.1 rc1 or 2.6.0)

Okay, not much luck with 2.6.1 rc1 either. The trace does look different 
now, though.

Jan  2 03:12:33 aurora kernel: ndiswrapper: no version for 
"struct_module" found: kernel tainted.
Jan  2 03:12:33 aurora kernel: ndiswrapper version 0.3+CVS loaded
Jan  2 03:12:47 aurora kernel: Unable to handle kernel paging request at 
virtual address 293c8000
Jan  2 03:12:47 aurora kernel:  printing eip:
Jan  2 03:12:47 aurora kernel: 46e36792
Jan  2 03:12:47 aurora kernel: *pde = 00099063
Jan  2 03:12:47 aurora kernel: Debug: sleeping function called from 
invalid context at include/linux/rwsem.h:43
Jan  2 03:12:47 aurora kernel: in_atomic():0, irqs_disabled():1
Jan  2 03:12:47 aurora kernel: Call Trace:
Jan  2 03:12:47 aurora kernel:  [<0211f1dc>] __might_sleep+0x7e/0x88
Jan  2 03:12:47 aurora kernel:  [<0215ccdc>] rw_vm+0x1bc/0x376
Jan  2 03:12:47 aurora kernel:  [<46e36792>] my_memset+0x8/0xf [ndiswrapper]
Jan  2 03:12:47 aurora kernel:  [<46e36792>] my_memset+0x8/0xf [ndiswrapper]
Jan  2 03:12:47 aurora kernel:  [<0215d0c6>] get_user_size+0x30/0x57
Jan  2 03:12:47 aurora kernel:  [<46e36792>] my_memset+0x8/0xf [ndiswrapper]
Jan  2 03:12:47 aurora kernel:  [<0210d03d>] handle_BUG+0x32/0xdf
Jan  2 03:12:47 aurora kernel:  [<021229f2>] printk+0x25c/0x2d5
Jan  2 03:12:47 aurora kernel:  [<0210d1ab>] die+0xc1/0x1b1
Jan  2 03:12:47 aurora kernel:  [<0211a653>] do_page_fault+0x2f8/0x447
Jan  2 03:12:47 aurora kernel:  [<46e36792>] my_memset+0x8/0xf [ndiswrapper]
Jan  2 03:12:47 aurora kernel:  [<0211d6aa>] __wake_up_common+0x33/0x57
Jan  2 03:12:47 aurora kernel:  [<0211d758>] __wake_up+0x8a/0xee
Jan  2 03:12:47 aurora kernel:  [<0211d6aa>] __wake_up_common+0x33/0x57
Jan  2 03:12:48 aurora kernel:  [<0211a35b>] do_page_fault+0x0/0x447
Jan  2 03:12:48 aurora kernel:  [<46e36792>] my_memset+0x8/0xf [ndiswrapper]
Jan  2 03:12:48 aurora kernel:
Jan  2 03:12:48 aurora kernel: Oops: 0002 [#1]
Jan  2 03:12:48 aurora kernel: CPU:    0
Jan  2 03:12:48 aurora kernel: EIP:    0060:[<46e36792>]    Tainted: P
Jan  2 03:12:48 aurora kernel: EFLAGS: 00010202
Jan  2 03:12:48 aurora kernel: EIP is at my_memset+0x8/0xf [ndiswrapper]
Jan  2 03:12:48 aurora kernel: eax: 293c5f14   ebx: 293c5e84   ecx: 
c0042c86   edx: 293c5f14
Jan  2 03:12:48 aurora kernel: esi: 29572df8   edi: 293c8000   ebp: 
293c5f0c   esp: 293c5e64
Jan  2 03:12:48 aurora kernel: ds: 007b   es: 007b   ss: 0068
Jan  2 03:12:48 aurora kernel: Process loadndisdriver (pid: 3521, 
threadinfo=293c4000 task=26830950)
Jan  2 03:12:48 aurora kernel: Stack: 29572df8 28f4df60 46e8747a 
293c5e84 00000000 0000007c 293c5f24 00000000
Jan  2 03:12:48 aurora kernel:        14141414 14141414 14141414 
14141414 14141414 14141414 14141414 14141414
Jan  2 03:12:48 aurora kernel:        14141414 14141414 14141414 
14141414 14141414 14141414 14141414 14141414
Jan  2 03:12:48 aurora kernel: Call Trace:
Jan  2 03:12:48 aurora kernel:
Jan  2 03:12:48 aurora kernel: Code: f3 aa 89 d8 5b 5f c3 89 10 c3 0f b7 
d2 66 89 10 c3 88 10 c3




// kaj

[Ndiswrapper-general] Re: ndiswrapper-snap-200401020157.tar.gz

From: Giridhar Pemmasani <giri@lm...> - 2004-01-02 06:00

I did clean up send_one a bit yesterday, but I am sure I didn't change
the functionality.

Kaj, by any chance are you enabling any debug options in the kernel?
Can you try CVS of couple of days ago and confirm if that does/doesn't
work for you? If you can isolate the problem to a commit, it would be
easier to fix.

-- 
Giri

Re: [Ndiswrapper-general] Re: ndiswrapper-snap-200401020157.tar.gz

From: Tomas Szepe <szepe@pi...> - 2004-01-02 13:21

On Jan-02 2004, Fri, 03:25 +0200
Kaj J. Niemi <kaj.niemi@...> wrote:

> Kaj J. Niemi wrote:
> 
> >I'll try 2.6.1 rc1 in a second.
> 
> CONFIG_REEMPT is not set in either one (2.6.1 rc1 or 2.6.0)
> 
> Okay, not much luck with 2.6.1 rc1 either. The trace does look different 
> now, though.

Ok, I'm out of ideas as to what could be causing this --
do you have an older version of ndiswrapper that doesn't
exhibit this problem?

-- 
Tomas Szepe <szepe@...>

Re: [Ndiswrapper-general] Re: ndiswrapper-snap-200401020157.tar.gz

From: Kaj J. Niemi <kaj.niemi@ba...> - 2004-01-02 14:45

Tomas Szepe wrote:

> Ok, I'm out of ideas as to what could be causing this --
> do you have an older version of ndiswrapper that doesn't
> exhibit this problem?

Getting some sleep must have been the key component... Now ndiswrapper 
loads fine under 2.6.1 rc 1 again. No panics so far on load/unload. :)

// kaj