Bug 147188

Summary: kernel without iptable modules will hang boot
Product: [openSUSE] SUSE Linux 10.1 Reporter: Olaf Hering <ohering>
Component: KernelAssignee: E-mail List <kernel-maintainers>
Status: RESOLVED FIXED QA Contact: E-mail List <qa-bugs>
Severity: Major    
Priority: P5 - None    
Version: Beta 3   
Target Milestone: ---   
Hardware: All   
OS: Linux   
Whiteboard:
Found By: Development Services Priority:
Business Priority: Blocker: ---
Marketing QA Status: --- IT Deployment: ---

Description Olaf Hering 2006-02-01 02:56:12 UTC 
happend already in 10.0-i386, its still broken.

arambola:~ # ps fax | grep -m1 -wA 5 2242 ; lsmod
 2242 ?        Ss     0:00 /bin/bash /etc/init.d/rc 5
 2263 ?        S      0:00  \_ startpar -p4 -t 30 -T 3 -M start -P N -R 5
 4710 ?        Ss     0:00      \_ startpar -f -- xdm
 4817 ?        S      0:00      \_ /bin/sh /etc/init.d/SuSEfirewall2_setup start
 4823 ?        S      0:00          \_ /bin/bash /sbin/SuSEfirewall2 -q start
 4940 ?        S      0:00              \_ /usr/sbin/iptables -A reject_func -p tcp -j REJECT --reject-with tcp-reset
Module                  Size  Used by
carambola:~ # 

SuSEfirewall2 S 000000000fe14de4 11280  4823   4817  4940               (NOTLB)
Call Trace:
[C000000074FF7940] [000000007BAA6196] 0x7baa6196 (unreliable)
[C000000074FF7B10] [C00000000000E70C] .__switch_to+0x100/0x120
[C000000074FF7BA0] [C000000000470F6C] .schedule+0x62c/0x778
[C000000074FF7C80] [C00000000005AE44] .do_wait+0xc44/0xde8
[C000000074FF7DC0] [C0000000000127A4] .compat_sys_waitpid+0x18/0x2c
[C000000074FF7E30] [C0000000000086F8] syscall_exit+0x0/0x40
iptables      S 000004000017b79c 11664  4940   4823                     (NOTLB)
Call Trace:
[C000000076CF37C0] [C000000076CF3880] 0xc000000076cf3880 (unreliable)
[C000000076CF3990] [C00000000000E70C] .__switch_to+0x100/0x120
[C000000076CF3A20] [C000000000470F6C] .schedule+0x62c/0x778
[C000000076CF3B00] [C00000000024FD60] .write_chan+0x3dc/0x464
[C000000076CF3C10] [C00000000024B93C] .tty_write+0x1e4/0x2e0
[C000000076CF3CF0] [C0000000000AEF0C] .vfs_write+0xd0/0x1b0
[C000000076CF3D90] [C0000000000AF624] .sys_write+0x4c/0x8c
[C000000076CF3E30] [C0000000000086F8] syscall_exit+0x0/0x40

arambola:~ #  strace -fp 4940
Process 4940 attached - interrupt to quit
write(2, "les table `filter\': iptables who"..., 58

I can kill -11 4940, then the parent will hang. 
if I kill -9 it, its parent will hang too, until I kill -9 it as well.

makes kernel testing in default config nearly impossible.
Comment 1 Andreas Kleen 2006-02-01 12:25:25 UTC 
Hmm - my x86-64 test kernels all don't have iptables, but I've never seen this with beta2 or 10.0 (haven't tried beta3 yet)
Comment 2 Olaf Kirch 2006-02-02 14:14:00 UTC 
But that's not an iptables problem. Something got stuck in the tty
driver - it hangs in tty_write.
Comment 3 Olaf Kirch 2006-02-02 14:20:40 UTC 
Sounds like a bug in the ppc serial or console driver to me.
Comment 4 Harald Mueller-Ney 2006-04-20 22:11:57 UTC 
Power Team, Status on this please? No reaction for more than 2 month on a major bug! Should we solve it as invalid or wontfix?
Comment 5 Harald Mueller-Ney 2006-04-27 20:05:38 UTC 
Your decission, see my last comment.
Comment 6 Matthias Fruehauf 2006-04-27 20:36:09 UTC 
OK, this want obviously disappeared from our radar. Olaf, would you please be so kind and retest with latest Beta?
Comment 7 Lars Marowsky-Bree 2006-05-16 09:43:22 UTC 
Resolved invalid as no information has been forthcoming for over 3 months - assuming the bug is fixed by the various updates since.
Comment 8 Olaf Hering 2006-09-28 20:01:50 UTC 
still broken.