Skip to contentSkip to conversation

hang on PPC

Bridge Stephen Hemminger 2 participants 2 messages Jun 28, 2004 – Jun 29, 2004 1 message not in the archive Ethernet bridge

lore
Share & tools
lore

A message that is not in the archive (sent off-list or not yet received): [email protected]

Stephen Hemminger

author
#1 lore raw
Quoted 24 lines · On Sun, 27 Jun 2004 15:43:30 +0700 <[email protected]> wrote:

Hi.

I'm using a powerpc-based board (CPU - Motorolla 82855, 50 Mhz) with 32 Mb of RAM, Big Endian.
Linux kernel - 2.4.22, bridge - statically linked & as a module
brctl - 0.9.6
Ethernet - fec (arch/ppc/8xx_io/fec.c)
All this was compiled with gcc version 3.2.2 20030217 (Yellow Dog Linux 3.0 3.2.2-2a_1)) with math emulation enabled in gcc and in kernel.

When I do

# ifconfig eth0 0.0.0.0
# brctl addbr br0
# brctl addif br0 eth0
eth0: Promiscuous mode enabled.
device eth0 entered promiscuous mode
# brctl show
bridge name     bridge id               STP enabled
interfaces
br0             8000.ce3dfa010010       no              eth0
# ifconfig br0 up
Quoted 6 lines

I get
br0: port 1(eth0) entering learning state
br0: port 1(eth0) entering forwarding state
br0: topology change detected, propagating

Now brctl is sleeping, forever...

Any brctl or which command? The usual way of debugging this is to enable SysRq in kernel and hit SysRq-t to see what is happening. Also SysRq-p to show processes.

Quoted 2 lines

On i386 with 2.4.24 I've tested the same - bridge works fine.

The 2.4 code hasn't changed much in a long time. It could be a PPC issue You could rebuild a kernel with just an update of the the bridge code (net/bridge) from 2.4.27.

Quoted 5 lines

Can anybody advice me a solution ?

I have enough experience in kernel development to fix it myself. The question is where ?

Quoted 8 lines

Please, reply to pavel[AT]parabel.inc.ru
OSDL rejected my mail with:
Failed to deliver to '[email protected]'
SMTP module(domain osdl.org) reports:
message text rejected by smtp.osdl.org:
554 5.7.1 Your server (195.161.113.6) is not listed as an
IP for the domain 'da.ru' -- please have your misconfiguration corrected

OSDL does anti-spam detection by doing a reverse DNS lookup of incoming mail. Since most spammers use bogus addresses, it filters about 10% of spam.

Quoted 87 lines

Thanks in advance.

BTW: My final task is make a patch for kernel to make bridge working with cisco-hdlc devices (hdlc_cisco.c will have type = ARPHRD_ETHER, and do appropriate encapsulation). Maybe someone already made this, but I don't know ?

Kernel boot log:

Linux package found at 200000
Kernel size 777k (compressed)
Ramdisk size 2913k (compressed)
Decompressing kernel...ok
Linux version 2.4.22 (root@knight) (gcc version 3.2.2 20030217 (Yellow Dog Linux
3.0 3.2.2-2a_1)) #30 Sun Jun 27 13:31:17 NOVST 2004 On node 0 totalpages: 8192
zone(0): 8192 pages.
zone(1): 0 pages.
zone(2): 0 pages.
Kernel command line: root=/dev/ram rw ramdisk_size=8000 init=/sbin/fl
Decrementer Frequency = 184320000/60
Calibrating delay loop... 48.84 BogoMIPS
Memory: 27560k available (1372k kernel code, 440k data, 84k init, 0k highmem)
Dentry cache hash table entries: 4096 (order: 3, 32768 bytes)
Inode cache hash table entries: 2048 (order: 2, 16384 bytes) Mount cache hash table entries: 512 (order: 0, 4096 bytes) Buffer cache hash table entries: 1024 (order: 0, 4096 bytes) Page-cache hash table entries: 8192 (order: 3, 32768 bytes) POSIX conformance testing by UNIFIX
Linux NET4.0 for Linux 2.4
Based upon Swansea University Computer Society NET3.039 Initializing RT netlink socket
Starting kswapd
CPM UART driver version 0.03
ttyS00 at 0x0280 is a SMC
pty: 256 Unix98 ptys configured
Generic RTC Driver v1.07
DLCI driver v0.35, 4 Jan 1997, [email protected].
eth0: FEC ENET Version 0.2, FEC irq 9, MII irq 10, addr ce:3d:fa:01:00:10
RAMDISK driver initialized: 16 RAM disks of 8000K size 1024 blocksize
loop: loaded (max 8 devices)
PPP generic driver version 2.4.2
PPP Deflate Compression module registered
Cronyx Ltd, Synchronous PPP and CISCO HDLC (c) 1994 Linux port (c) 1998 Building Number Three Ltd & Jan "Yenya" Kasprzak.
LAPB Ethernet driver version 0.01
NET4: Linux TCP/IP 1.0 for NET4.0
IP Protocols: ICMP, UDP, TCP, IGMP
IP: routing cache hash table of 512 buckets, 4Kbytes
TCP: Hash tables configured (established 2048 bind 4096) IPv4 over IPv4 tunneling driver
ip_conntrack version 2.1 (256 buckets, 2048 max) - 292 bytes per conntrack
ip_tables: (C) 2000-2002 Netfilter core team
NET4: Unix domain sockets 1.0/SMP for Linux NET4.0.
NET4: Ethernet Bridge 008 for NET4.0
X.25 for Linux. Version 0.2 for Linux 2.1.15
NET4: LAPB for Linux. Version 0.01 for NET4.0 802.1Q VLAN Support v1.8 Ben Greear <[email protected]>
All bugs added by David S. Miller <[email protected]>
RAMDISK: Compressed image found at block 0
Freeing initrd memory: 2913k freed
VFS: Mounted root (ext2 filesystem).
Freeing unused kernel memory: 84k init

Signature · 9 lines
Regards
Pavel

______ "______________"
web: http://parabel.inc.ru/
mail: pavel[AT]parabel.inc.ru
phone: +7 (3832) 13-87-07
---------------------------------------------------------
http://relax.ngs.ru - ______ __ ____________ __ ________________________ __ ________________________!
#2 lore raw
Quoted 11 lines · On Mon, 28 Jun 2004 09:38:46 -0700 Stephen Hemminger <[email protected]>…
Quoted 6 lines

I get
br0: port 1(eth0) entering learning state
br0: port 1(eth0) entering forwarding state
br0: topology change detected, propagating

Now brctl is sleeping, forever...

Any brctl or which command? The usual way of debugging this is
to enable SysRq in kernel and hit SysRq-t to see what is happening.
Also SysRq-p to show processes.

The process is locked(slept) by bridge module.

I've found a reason, the reason is in "LAPB over Ethernet driver (CONFIG_LAPBETHER)" conflict.
I think it is in register_netdevice_notifier's handlers lock between lapbether & bridge modules.

I've made a modification for hdlc_cisco.c, now it can report cisco-hdlc device as Ethernet (with cisco-compatible bridge encapsulation).
But I have one problem:
hdlc_cisco must add 4-byte header in hard_start_xmit (as long as no one call hard_header. But I have no space, as long as bridge's interface have hard_header_len = 0.... So' I have to make a COPY os skbuff, instead of cloning if I'm the only user....

When I add hard_header_len = 4, it stopped work even on Ethernet.
Is it hard to make bridge working with hard_header_len != 0 (It's necessary for Ethernet in (cisco-hdlc, frame-relay, ppp) encapsulation) ?

Signature · 8 lines
Regards
Pavel

web: http://parabel.inc.ru/
mail: pavel[AT]parabel.inc.ru
phone: +7 (3832) 13-87-07
---------------------------------------------------------
http://relax.ngs.ru - ÷ÓÅ Ï ÄÏÓÕÇÅ É ÒÁÚ×ÌÅÞÅÎÉÑÈ × îÏ×ÏÓÉÂÉÒÓËÅ!