Skip to contentSkip to conversation

Strange TCP with 2.5.60

DISCUSSION Kostadin Karaivanov 5 participants 6 messages Feb 12, 2003 – Feb 14, 2003

Subsystems (2)
lore
Share & tools
lore
#0 lore raw

Trying to use traceroute i get folowing error:

root@larry:~# uname -a
Linux larry 2.5.60 #20 Tue Feb 11 17:38:42 EET 2003 i686 unknown
root@larry:~# traceroute www.kernel.org
traceroute: findsaddr: Can't find interface

Networking works....... but command route can't give me the status of default route:

root@larry:~# route -n
Kernel IP routing table
Destination     Gateway         Genmask         Flags Metric Ref    Use
Iface
192.168.0.0     0.0.0.0         255.255.240.0   U     0      0        0 eth0
127.0.0.0       0.0.0.0         255.0.0.0       U     0      0        0 lo
root@larry:~# ip route list
192.168.0.0/20 dev eth0  proto kernel  scope link  src 192.168.1.12
127.0.0.0/8 dev lo  scope link
default via 192.168.1.1 dev eth0  metric 1
root@larry:~#

The version of traceroute is 1.4a12
The version of route is net-tools 1.60 / route 1.98 (2001-04-15)

iptables doesn't work as expected too.....

root@larry:~# iptables -A INPUT -i lo -j ACCEPT
root@larry:~# iptables -L -n
Chain INPUT (policy ACCEPT)
target     prot opt source               destination
ACCEPT     all  --  0.0.0.0/0            0.0.0.0/0

Chain FORWARD (policy ACCEPT)
target prot opt source destination

Chain OUTPUT (policy ACCEPT)
target prot opt source destination

everything works fine with 2.5.59 and 2.4.20

I can provide any additional information if needed

Kostadin Karaivanov
Senior System Administrator @ Ministry Of Finance
tel: +359 2 98592062
[email protected]

#1 lore raw
Quoted 10 lines · Kostadin Karaivanov

Trying to use traceroute i get folowing error:
...

Networking works....... but command route can't give me the status of default route:
...
iptables doesn't work as expected too.....
...
everything works fine with 2.5.59 and 2.4.20

Could you pelase retest with this patch, and tell us how many of these problems go away?

net/ipv4/fib_hash.c renamed from net/ipv4/fib_hash.c~a +1 -1
diff -puN net/ipv4/fib_hash.c~a net/ipv4/fib_hash.c
@@ -941,7 +941,7 @@ static __inline__ struct fib_node *fib_g
 
 			if (!iter->zone)
 				goto out;
-			if (iter->zone->fz_next)
+			if (iter->zone->fz_next);
 				break;
 		}
 		

_

#2 lore raw
Quoted 16 lines · Andrew Morton

Could you pelase retest with this patch, and tell us how many of these problems go away?

diff -puN net/ipv4/fib_hash.c~a net/ipv4/fib_hash.c

net/ipv4/fib_hash.c renamed from net/ipv4/fib_hash.c~a +1 -1
@@ -941,7 +941,7 @@ static __inline__ struct fib_node *fib_g
 
 			if (!iter->zone)
 				goto out;
-			if (iter->zone->fz_next)
+			if (iter->zone->fz_next);
 				break;
 		}
 
Quoted 1 line

_

is that patch reversed, I hope? The if doesn't do a great deal.

Signature · 3 lines
bill davidsen <[email protected]>
  CTO, TMR Associates, Inc
Doing interesting things with little computers since 1979.
#3 lore raw
Quoted 19 lines · Bill Davidsen
Quoted 16 lines · Andrew Morton

Could you pelase retest with this patch, and tell us how many of these problems go away?

diff -puN net/ipv4/fib_hash.c~a net/ipv4/fib_hash.c

net/ipv4/fib_hash.c renamed from net/ipv4/fib_hash.c~a +1 -1
@@ -941,7 +941,7 @@ static __inline__ struct fib_node *fib_g
 
 			if (!iter->zone)
 				goto out;
-			if (iter->zone->fz_next)
+			if (iter->zone->fz_next);
 				break;
 		}
 
Quoted 3 lines
Quoted 1 line

_

is that patch reversed, I hope? The if doesn't do a great deal.

It is deliberate. It directly reverts a bugfix which was recently merged.

Apparently the problem is that this combined with recent seq_file conversions has resulted in changed format of some /proc/net/foo files, which is confusing some network admin apps.

It is being worked on.

#4 lore raw
Quoted 20 lines
Quoted 16 lines
Quoted 16 lines

Could you pelase retest with this patch, and tell us how many of these problems go away?

diff -puN net/ipv4/fib_hash.c~a net/ipv4/fib_hash.c

net/ipv4/fib_hash.c renamed from net/ipv4/fib_hash.c~a +1 -1
@@ -941,7 +941,7 @@ static __inline__ struct fib_node *fib_g
 
 			if (!iter->zone)
 				goto out;
-			if (iter->zone->fz_next)
+			if (iter->zone->fz_next);
 				break;
 		}
 
Quoted 3 lines
Quoted 1 line

_

is that patch reversed, I hope? The if doesn't do a great deal.

Quoted 7 lines

It is deliberate. It directly reverts a bugfix which was recently merged.

Apparently the problem is that this combined with recent seq_file conversions has resulted in changed format of some /proc/net/foo files, which is confusing some network admin apps.

It is being worked on.

Can't we do:

- if (iter->zone->fz_next)
+ /* if (iter->zone->fz_next) */

or something similar? Is much less confusing ...

M.

#5 lore raw
Quoted 11 lines
Quoted 5 lines

Apparently the problem is that this combined with recent seq_file conversions has resulted in changed format of some /proc/net/foo files, which is confusing some network admin apps.

It is being worked on.

Can't we do:

(quoted diff) +1 -1
-			if (iter->zone->fz_next)
+			/* if (iter->zone->fz_next) */
Quoted 1 line

or something similar? Is much less confusing ...

See "It is being worked on" 8-)
Better to fix it properly..

Dave

Signature · 1 line
| Dave Jones.        http://www.codemonkey.org.uk