Subject: kern/17491: pppoe0 panic 1.6_BETA3/i386
To: None <>
From: None <>
List: netbsd-bugs
Date: 07/06/2002 09:21:41
>Number:         17491
>Category:       kern
>Synopsis:       pppoe0 panic 1.6_BETA3/i386
>Confidential:   no
>Severity:       critical
>Priority:       high
>Responsible:    kern-bug-people
>State:          open
>Class:          sw-bug
>Submitter-Id:   net
>Arrival-Date:   Sat Jul 06 06:22:00 PDT 2002
>Originator:     Ty Sarna
>Release:        NetBSD 1.6_BETA3 20020626
System: NetBSD 1.6_BETA3 NetBSD 1.6_BETA3 (TALYN) #10: Wed Jun 26 21:03:14 EDT 2002 i386
Architecture: i386
Machine: i386
For the second morning in a row, I woke up to find one of my
machines "off the air" wrt pppoe0, and the other paniced.
Both have independant pppoe connections to the same provider over the
same ethernet.

The first machine is running -current (1.6B/i386) without gif, has a
number of "fxp1: device timeout"s logged. 

The second machine (the one described in Environment) had these
panic tracebacks:

(missed panic string for first one)
db> trace
sppp_output(c0d3f000,c0fa1f00,c0d385d0,c0c98c60,c0fa1f00) at sppp_output+0x170
ip_output(c0fa1f00,0,c0d47f68,0,0) at ip_output+0x789
in_gif_output(c0d47e00,18,c0fa1f00,1000000,c0d381c0) at in_gif_output+0x2e3
gifintr(c0d47e00,c0fa1f00,c029a169,e35ede00) at gifintr+0xd7
softintr_dispatch(1) at softintr_dispatch+0x99
Xsoftnet() at Xsoftnet+0x40
--- interrupt ---
idle(e34cd738,e35edf78,bfbf001f,bfbf001f) at idle+0x20
bpendtsleep(e35d0618,128,c02d96a7,0,0,e35edf78,e35edf40,c01b22d7) at bpendtsleep

sigsuspend1(e34cd738,e35edf30,10,283,e35edf80) at sigsuspend1+0x106
sys___sigsuspend14(e34cd738,e35edf80,e35edf78,c0c3b038) at sys___sigsuspend14+0x3c
syscall_plain(bfbf001f,bfbf001f,1f,bfbf001f,bfbfdc1c) at syscall_plain+0xa7

Second one:
uvm_fault(0xc0322e40, 0xc0e3d000, 0, 1) -> e

sppp_output(c0d49000,c0e3cf00,c0d434f0,c0c98c60,c0e3cf00) at
ip_output(c0e3cf00,0,c0d51f68,0,0) at ip_output+0x789
in_gif_output(c0d51e00,18,c0e3cf00,e39d0df0,c0d43200) at
gifintr(c0d51e00,6,c0cd5c00,42ea) at gifintr+0xd7
softintr_dispatch(1) at softintr_dispatch+0x99
Xsoftnet() at Xsoftnet+0x40
--- interrupt ---

An additional problem, when the second machine was rebooted it got to
the point at which it tries to start pppoe0, and then continuously spewed:

pppoe0: ipcp illegal conf-req in state initial

at a very high rate until I pulled the ethernet out, at which point I
could get control of the machine.  Seems like this error handling could
be improved so that it doesn't turn into a DoS like this :-)

	Wait 24 hours?