[3583] in linux-net channel archive
Re: Linux 2.0.0 bugs (was Mrouted and multicast tools on ONE LINUX
daemon@ATHENA.MIT.EDU (Malcolm Beattie)
Fri Jul 5 12:30:31 1996
From: Malcolm Beattie <malcolm.beattie@computing-services.oxford.ac.uk>
To: alan@lxorguk.ukuu.org.uk (Alan Cox)
Date: Fri, 5 Jul 1996 10:18:49 +0100 (BST)
Cc: linux-net@vger.rutgers.edu
In-Reply-To: <m0ubY6e-0005FbC@lightning.swansea.linux.org.uk> from "Alan Cox" at Jul 3, 96 09:01:00 pm
With 2.0 + Alan's patch-kit + the skb_reserve() fix (effectively
2.0.1 for the purposes of this posting), I'm getting a new
"kernel aiee" in ip_forward() with mrouted routing.
The slightly good news is that, modulo this, I have an mrouted tunnel
partially working but I'll post about that separately.
The oopses are as follows:
Jul 4 18:44:35 plutonium kernel: Unable to handle kernel paging request at virtual address c18361a8
Jul 4 18:44:35 plutonium kernel: current->tss.cr3 = 00101000, Hr3 = 00101000
Jul 4 18:44:35 plutonium kernel: *pde = 00001067
Jul 4 18:44:35 plutonium kernel: *pte = 00000000
Jul 4 18:44:35 plutonium kernel: Oops: 0000
Jul 4 18:44:35 plutonium kernel: CPU: 0
Jul 4 18:44:35 plutonium kernel: EIP: 0010:[<00146492>]
Jul 4 18:44:35 plutonium kernel: EFLAGS: 00010202
Jul 4 18:44:35 plutonium kernel: eax: 00000001 ebx: 0a303d76 ecx: 01836170 edx: 00000000
Jul 4 18:44:35 plutonium kernel: esi: 00b00364 edi: 00000018 ebp: 001b28d8 esp: 001a62b0
Jul 4 18:44:35 plutonium kernel: ds: 0018 es: 0018 fs: 002b gs: 0018 ss: 0018
Jul 4 18:44:35 plutonium kernel: Process swapper (pid: 0, process nr: 0, stackpage=001a4448)
Jul 4 18:44:35 plutonium kernel: Stack: 00b00364 00000010 00000118 001b28d8 00000014 0000014c 00000000 00b003b4
Jul 4 18:44:35 plutonium kernel: 00000000 00b00364 00b0025c 01836170 001573f0 00b00364 001b28d8 00000018
Jul 4 18:44:35 plutonium kernel: 0e0201a3 00000007 00000002 00431958 00355498 001574d8 00b00364 001babc8
Jul 4 18:44:35 plutonium kernel: Call Trace: [<01836170>] [<001573f0>] [<001574d8>] [<0014558e>] [<0013d66c>] [<0011706b>] [<0010a56b>]
Jul 4 18:44:35 plutonium kernel: [<01b2d00f>] [<001096bc>] [<0010a5e2>] [<001093d0>] [<00109243>] [<00116760>] [<01ae975c>] [<0011049c>]
Jul 4 18:44:35 plutonium kernel: [<01b2d011>]
Jul 4 18:44:35 plutonium kernel: Code: f6 41 38 01 0f 84 94 04 00 00 8b 54 24 10 8b 74 24 34 03 56
Jul 4 18:44:35 plutonium kernel: Aiee, killing interrupt handler
Jul 4 18:44:35 plutonium kernel: kfree of non-kmalloced memory: 001a6490, next= fe2001a3, order=14485504
Jul 4 18:44:35 plutonium kernel: kfree of non-kmalloced memory: 001a6480, next= fe2001a3, order=14485504
Jul 4 18:44:35 plutonium kernel: kfree of non-kmalloced memory: 001a68b4, next= fe2001a3, order=14485504
Jul 4 18:44:35 plutonium kernel: idle task may not sleep
Jul 4 18:44:35 plutonium last message repeated 4 times
Jul 4 18:47:12 plutonium kernel: Unable to handle kernel paging request at virtual address c18361a8
Jul 4 18:47:12 plutonium kernel: current->tss.cr3 = 00101000, Hr3 = 00101000
Jul 4 18:47:12 plutonium kernel: *pde = 00001067
Jul 4 18:47:12 plutonium kernel: *pte = 00000000
Jul 4 18:47:12 plutonium kernel: Oops: 0000
Jul 4 18:47:12 plutonium kernel: CPU: 0
Jul 4 18:47:12 plutonium kernel: EIP: 0010:[<00146492>]
Jul 4 18:47:12 plutonium kernel: EFLAGS: 00010202
Jul 4 18:47:12 plutonium kernel: eax: 00000001 ebx: 0a303d76 ecx: 01836170 edx: 00000000
Jul 4 18:47:12 plutonium kernel: esi: 00b00d50 edi: 00000018 ebp: 001b28d8 esp: 001a60d4
Jul 4 18:47:12 plutonium kernel: ds: 0018 es: 0018 fs: 002b gs: 0000 ss: 0018
Jul 4 18:47:12 plutonium kernel: Process swapper (pid: 0, process nr: 0, stackpage=001a4448)
Jul 4 18:47:12 plutonium kernel: Stack: 00b00d50 00000010 00000118 001b28d8 00000014 0000014c 00000000 00b00da0
Jul 4 18:47:12 plutonium kernel: 00000000 00b00d50 00b00c48 01836170 001573f0 00b00d50 001b28d8 00000018
Jul 4 18:47:12 plutonium kernel: 0e0201a3 00000007 00000002 00431f4c 00355498 001574d8 00b00d50 001babc8
Jul 4 18:47:12 plutonium kernel: Call Trace: [<01836170>] [<001573f0>] [<001574d8>] [<0014558e>] [<0013d66c>] [<0011706b>] [<0010a56b>]
Jul 4 18:47:12 plutonium kernel: [<00115a2d>] [<00115d7a>] [<0010ab92>] [<02000000>] [<01800000>] [<0011012f>] [<0010fe70>] [<0010a77b>]
Jul 4 18:47:12 plutonium kernel: [<01836170>] [<00146492>] [<01836170>] [<001573f0>] [<001574d8>] [<0014558e>] [<0013d66c>] [<0011706b>]
Jul 4 18:47:12 plutonium kernel: [<0010a56b>] [<01b2d00f>] [<001096bc>] [<0010a5e2>] [<001093d0>] [<00109243>] [<00116760>] [<01aed420>]
Jul 4 18:47:12 plutonium kernel: [<0011049c>] [<01b30d9d>]
Jul 4 18:47:12 plutonium kernel: Code: f6 41 38 01 0f 84 94 04 00 00 8b 54 24 10 8b 74 24 34 03 56
Jul 4 18:47:12 plutonium kernel: Aiee, killing interrupt handler
ksymoops shows the stack trace to be:
>>EIP: 146492 <ip_forward+202/700>
Trace: 1836170
Trace: 1573f0 <ipmr_queue_xmit+50/70>
Trace: 1574d8 <ipmr_forward+c8/100>
Trace: 14558e <ip_rcv+45e/590>
Trace: 13d66c <net_bh+ec/120>
Trace: 11706b <do_bottom_half+3b/70>
Trace: 10a56b <handle_bottom_half+b/20>
Trace: 1b2d00f
Trace: 1096bc <sys_idle+6c/80>
Trace: 10a5e2 <system_call+52/80>
Trace: 1093d0 <init>
Trace: 109243 <start_kernel+163/170>
Trace: 116760 <it_real_fn>
Trace: 1ae975c
Trace: 11049c <schedule+24c/2b0>
Trace: 1b2d011
Disassembling ip_forward and matching it up with the assembler, the
code path is:
if(!(is_frag&IPFWD_MULTICASTING))
...
else
{
/*
* Multicast route forward. Routing is already done
*/
dev2=skb->dev;
raddr=skb->raddr;
if(is_frag&IPFWD_MULTITUNNEL) /* VIFF_TUNNEL mode */
encap=20;
rt=NULL;
}
[CONFIG_FIREWALL is *not* set so no code in here]
/*
* We now may allocate a new buffer, and copy the datagram into it.
* If the indicated interface is up and running, kick it.
*/
if (dev2->flags & IFF_UP)
{
0x14648e <ip_forward+510>: movl 0x2c(%esp,1),%ecx
0x146492 <ip_forward+514>: testb $0x1,0x38(%ecx)
and trying to dereference dev2->flags bombs out. So presumably the
dev2=skb->dev;
a little way back didn't produce a valid device after all so maybe the
comment
* Multicast route forward. Routing is already done
is wrong.
--Malcolm
--
Malcolm Beattie <mbeattie@sable.ox.ac.uk>
Unix Systems Programmer
Oxford University Computing Services
