0001122: ifupdown on non-existent interface segfaults, busybox-1.3

(0001877)
vda
12-18-06 11:57

Show your /etc/network/interfaces please.

(0001878)
vda
12-18-06 13:00

Try attached c.patch and let us know whether that works

(0001879)
vda
12-18-06 13:02

c.patch applied in rev 16994, waiting for feedback

(0001886)
espakman
12-19-06 11:52

The segfault is gone, but now the following message is displayed on every interface:
"ifup: don't see to have all variables for ethx/inet"
(x is 0, 1, ..)

I think it's because of the "[[ ip ....%gateway% ]]". Removed extra spaces w/o checking" change. But haven't tested it yet.

(0001887)
espakman
12-19-06 12:37

Just tested and the "don't seem to have all variables" problem is not gone with reverting the "Removed extra spaces w/o checking" part (although I'm not sure why you did that).

(0001888)
vda
12-19-06 12:42

>(although I'm not sure why you did that)

Because there is no reason why we should issue " ip ... " with leading space, I think.

Show your /etc/network/interfaces. Please.

(0001889)
espakman
12-19-06 12:46

My /etc/network/interfaces file:

# Loopback interface.
auto lo
iface lo inet loopback

# dhcp
auto eth0
iface eth0 inet dhcp

# fixed
auto eth1
iface eth1 inet static
    address 192.168.1.254
    netmask 255.255.255.0
    broadcast 192.168.1.255

(0001890)
bernhardf
12-19-06 13:29

1) add -a in the commands below, if you specify it please.
Just curious as to what libc are you using?
$ ifdown -v
$ ifup -v

(0001891)
espakman
12-19-06 14:05

> Just curious as to what libc are you using?

Using uClibc-0.9.28

> $ ifdown -va
> $ ifup -v

I have to run it in Qemu, running it on my router will bring my connection down and hard to repair (have to switch busybox too..). I will give the differences in output between busybox-1.2.2.1 and 1.3.0 + your ifupdown patch.

ifdown -va
with 1.3.0:
....
ip addr flush dev eth1
Nothing to flush.
ip link set eth1 down

with 1.2.2.1:
ip addr flush dev eth1
ip link set eth1 down

ifup -va
with 1.3.0:
......
ifup: don't seem to have all the variables for eth1/inet

with 1.2.2.1:
......
run-parts /etc/network/if-pre-up.d
ip addr add 192.168.1.254/24 broadcast 192.168.1.255 dev eth1 label eth1
ip link set eth1 up

run-parts /etc/network/if-up.d

The /etc/network/interfaces file is exactly the same in both cases.

(0001892)
vda
12-19-06 14:06

Hmmm.... indeed. Looks like parse() returning "" and returning NULL actually means different things! Attached (possibly fixed) ifup comes further on my system:

# ./ifup eth1
run-parts: cannot open '/etc/network/if-pre-up.d': No such file or directory

Please test-and-report and/or send me your entire /etc/network/*.

(0001893)
espakman
12-19-06 14:24

Still the same problem:

ifup: don't seem to have all the variables for eth1/inet

The contents of my /etc/network/ directory:
if-down.d/
if-post-down.d/
if-pre-up.d/
if-up.d/
interfaces

For testing the if-* directories contains no files.

(0001894)
vda
12-19-06 15:00

I created empty directories and renamed eth1 to if (my real ethernet iface name). Then:

# ifup -vn if
run-parts /etc/network/if-pre-up.d
ip addr add 192.168.1.254/24 dev if label if
ip link set if up

run-parts /etc/network/if-up.d

# ifup -v if
run-parts /etc/network/if-pre-up.d
ip addr add 192.168.1.254/24 dev if label if
ip link set if up

run-parts /etc/network/if-up.d

# ip a
1: lo: <LOOPBACK,UP,10000> mtu 16436 qdisc noqueue
    link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
    inet 127.0.0.1/8 scope host lo
2: if: <BROADCAST,MULTICAST,UP,10000> mtu 1500 qdisc pfifo_fast qlen 1000
    link/ether 00:16:17:6d:77:db brd ff:ff:ff:ff:ff:ff
    inet 192.168.1.111/24 brd 192.168.1.255 scope global if
    inet 192.168.1.254/24 scope global secondary if

So it works for me with /etc/network exactly as yours. I also double-checked that I really use ifupdown.c like one attached on this page.

(0001897)
espakman
12-19-06 23:50

My latest tests also indicate that, when using 1 interface. When using two interfaces eth0 and eth1 (and loopback) the "don't seem to have all variables" message showed up on the second physical interface (busybox 1.3.0, not with 1.2.2.1).

I don't have the tools at the moment to doublecheck, will do that this evening (UTC).

(0001898)
espakman
12-20-06 11:59
edited on: 12-22-06 00:00

I doublechecked and recompiled and it's indeed working correct now! Thanks!

This bug can be closed.

(0001991)
jonlar
01-15-07 00:53

Please apply the fix for this bug in busybox_1_3_stable branch and provide a patch in http://busybox.net/downloads/fixes-1.3.1/ [^]

I had to use ifupdown.c from the trunk (r17011) to get ifup to work with busybox 1.3.1.

Notes
(0001877) vda 12-18-06 11:57	Show your /etc/network/interfaces please.

(0001878) vda 12-18-06 13:00	Try attached c.patch and let us know whether that works

(0001879) vda 12-18-06 13:02	c.patch applied in rev 16994, waiting for feedback

(0001886) espakman 12-19-06 11:52	The segfault is gone, but now the following message is displayed on every interface: "ifup: don't see to have all variables for ethx/inet" (x is 0, 1, ..) I think it's because of the "[[ ip ....%gateway% ]]". Removed extra spaces w/o checking" change. But haven't tested it yet.

(0001887) espakman 12-19-06 12:37	Just tested and the "don't seem to have all variables" problem is not gone with reverting the "Removed extra spaces w/o checking" part (although I'm not sure why you did that).

(0001888) vda 12-19-06 12:42	>(although I'm not sure why you did that) Because there is no reason why we should issue " ip ... " with leading space, I think. Show your /etc/network/interfaces. Please.

(0001889) espakman 12-19-06 12:46	My /etc/network/interfaces file: # Loopback interface. auto lo iface lo inet loopback # dhcp auto eth0 iface eth0 inet dhcp # fixed auto eth1 iface eth1 inet static address 192.168.1.254 netmask 255.255.255.0 broadcast 192.168.1.255

(0001890) bernhardf 12-19-06 13:29	1) add -a in the commands below, if you specify it please. Just curious as to what libc are you using? $ ifdown -v $ ifup -v

(0001891) espakman 12-19-06 14:05	> Just curious as to what libc are you using? Using uClibc-0.9.28 > $ ifdown -va > $ ifup -v I have to run it in Qemu, running it on my router will bring my connection down and hard to repair (have to switch busybox too..). I will give the differences in output between busybox-1.2.2.1 and 1.3.0 + your ifupdown patch. ifdown -va with 1.3.0: .... ip addr flush dev eth1 Nothing to flush. ip link set eth1 down with 1.2.2.1: ip addr flush dev eth1 ip link set eth1 down ifup -va with 1.3.0: ...... ifup: don't seem to have all the variables for eth1/inet with 1.2.2.1: ...... run-parts /etc/network/if-pre-up.d ip addr add 192.168.1.254/24 broadcast 192.168.1.255 dev eth1 label eth1 ip link set eth1 up run-parts /etc/network/if-up.d The /etc/network/interfaces file is exactly the same in both cases.

(0001892) vda 12-19-06 14:06	Hmmm.... indeed. Looks like parse() returning "" and returning NULL actually means different things! Attached (possibly fixed) ifup comes further on my system: # ./ifup eth1 run-parts: cannot open '/etc/network/if-pre-up.d': No such file or directory Please test-and-report and/or send me your entire /etc/network/*.

(0001893) espakman 12-19-06 14:24	Still the same problem: ifup: don't seem to have all the variables for eth1/inet The contents of my /etc/network/ directory: if-down.d/ if-post-down.d/ if-pre-up.d/ if-up.d/ interfaces For testing the if-* directories contains no files.

(0001894) vda 12-19-06 15:00	I created empty directories and renamed eth1 to if (my real ethernet iface name). Then: # ifup -vn if run-parts /etc/network/if-pre-up.d ip addr add 192.168.1.254/24 dev if label if ip link set if up run-parts /etc/network/if-up.d # ifup -v if run-parts /etc/network/if-pre-up.d ip addr add 192.168.1.254/24 dev if label if ip link set if up run-parts /etc/network/if-up.d # ip a 1: lo: <LOOPBACK,UP,10000> mtu 16436 qdisc noqueue link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00 inet 127.0.0.1/8 scope host lo 2: if: <BROADCAST,MULTICAST,UP,10000> mtu 1500 qdisc pfifo_fast qlen 1000 link/ether 00:16:17:6d:77:db brd ff:ff:ff:ff:ff:ff inet 192.168.1.111/24 brd 192.168.1.255 scope global if inet 192.168.1.254/24 scope global secondary if So it works for me with /etc/network exactly as yours. I also double-checked that I really use ifupdown.c like one attached on this page.

(0001897) espakman 12-19-06 23:50	My latest tests also indicate that, when using 1 interface. When using two interfaces eth0 and eth1 (and loopback) the "don't seem to have all variables" message showed up on the second physical interface (busybox 1.3.0, not with 1.2.2.1). I don't have the tools at the moment to doublecheck, will do that this evening (UTC).

(0001898) espakman 12-20-06 11:59 edited on: 12-22-06 00:00	I doublechecked and recompiled and it's indeed working correct now! Thanks! This bug can be closed.

(0001991) jonlar 01-15-07 00:53	Please apply the fix for this bug in busybox_1_3_stable branch and provide a patch in http://busybox.net/downloads/fixes-1.3.1/ [^] I had to use ifupdown.c from the trunk (r17011) to get ifup to work with busybox 1.3.1.

Issue History
Date Modified	Username	Field	Change
12-17-06 12:10	espakman	New Issue
12-17-06 12:10	espakman	Status	new => assigned
12-17-06 12:10	espakman	Assigned To	=> BusyBox
12-18-06 11:57	vda	Note Added: 0001877
12-18-06 13:00	vda	File Added: c.patch
12-18-06 13:00	vda	Note Added: 0001878
12-18-06 13:02	vda	Note Added: 0001879
12-18-06 14:52	vda	Status	assigned => feedback
12-19-06 11:52	espakman	Note Added: 0001886
12-19-06 12:37	espakman	Note Added: 0001887
12-19-06 12:42	vda	Note Added: 0001888
12-19-06 12:46	espakman	Note Added: 0001889
12-19-06 13:29	bernhardf	Note Added: 0001890
12-19-06 14:05	espakman	Note Added: 0001891
12-19-06 14:06	vda	Note Added: 0001892
12-19-06 14:06	vda	File Added: ifupdown.c
12-19-06 14:24	espakman	Note Added: 0001893
12-19-06 15:00	vda	Note Added: 0001894
12-19-06 23:50	espakman	Note Added: 0001897
12-20-06 11:59	espakman	Note Added: 0001898
12-22-06 00:00	espakman	Note Edited: 0001898
01-15-07 00:53	jonlar	Note Added: 0001991
01-17-07 15:47	vda	Status	feedback => closed
01-17-07 15:47	vda	Resolution	open => fixed

Relationships