Re: The mistery of the dlil_filter_detach call
Re: The mistery of the dlil_filter_detach call
- Subject: Re: The mistery of the dlil_filter_detach call
- From: Stephane Sudre <email@hidden>
- Date: Tue, 21 Feb 2006 20:00:54 +0100
Actually, my question is not related to this particular kernel panic.
It's "ok" it's kernel-panicing since this is a project in dedvelopment.
My question is more related to the fact that, for any 10.4.x version,
it seems that the path of the backtrace makes no sense after the
address of dlil_detach_filter. That is if you look at the source code
of dlil_detach_filter.
On 21 févr. 06, at 18:30, Vincent Lubet wrote:
Stéphane,
xnu-790.1 is a pre-released version of Tiger (may be a seed) for which
we don't have a symbol version of the mach_kernel.
Please try to reproduce on an official release version of Tiger (xnu
version 792 or higher) and we'll be able to help.
Vincent
On Feb 21, 2006, at 8:12 AM, Stephane wrote:
I'm beginning to wonder if I'm not missing one point with the panic
logs.
Let's say we have the following KP log:
Tue Feb 21 16:51:12 2006
panic(cpu 0 caller 0x002917A8): sbappendaddr
Latest stack backtrace for cpu 0:
Backtrace:
0x00095564 0x00095A7C 0x00026838 0x002917A8 0x00291994
0x001614D0 0x0014E6CC 0x0014F3A8
0x00139FFC 0x0013A110 0x0011C2A8 0x000A9654
Proceeding back via exception chain:
Exception state (sv=0x44553280)
PC=0x00000000; MSR=0x0000D030; DAR=0x00000000;
DSISR=0x00000000; LR=0x00000000; R1=0x00000000; XCP=0x00000000
(Unknown)
Kernel version:
Darwin Kernel Version 8.0.0: Thu Mar 24 19:14:36 PST 2005;
root:xnu-790.1.obj~1/RELEASE_PPC
The first two calls are:
Call_continuation and dlil_detach_filter
and then it goes into ip_input, etc...
Problems:
- there's been no interface filter detached as far as I know.
- I don't see ifmaddr_ifnet being called in the source code of
dlil_detach_filter.
Funny part:
This Kernel Panic happens with a kext in its early stage of
development and when double-clicking on a .pkg (Installer.app is
launched and then bang!)
Question:
Is it possible to trust a KP log on 10.4.x? I don't see any reason
why dlil_detach_filter would be called (no interface filter is
detached). So why can I see it in the log? The trace does not look
perfectly correct as usual with the KP I see involving
dlil_detach_filter.
_______________________________________________
Do not post admin requests to the list. They will be ignored.
Darwin-kernel mailing list (email@hidden)
Help/Unsubscribe/Update your Subscription:
This email sent to email@hidden