Skip to contentSkip to conversation

delay in bridge learning when forward delay is 0

Bridge Uli Luckas 3 participants 5 messages Sep 18, 2008 – Oct 14, 2008 Ethernet bridge

lore
Share & tools
lore

Uli Luckas

author
#0 lore raw

Hi,
In July 2007 Philip Craig reported the following issue with 'forward delay'=0 in great detail without receiving an answer. Has Philip's message got lost or was his analysis wrong?
The problem becomes a real problem if you bridge a fast LAN to a slow port like a bluetooth pan for example.

https://lists.linux-foundation.org/pipermail/bridge/2007-July/005476.html
Philip Craig philipc at snapgear.com

Quoted 22 lines

Hi,

If you set the bridge forward delay to 0 with:
brctl setfd br0 0
then the bridge does not learn addresses for the first 20 seconds, and so it floods everything during this time.

The reason for this is that hold_time() returns 0 after a topology change, br_fdb_update() is a no-op if hold_time() is 0 (so that 'brctl setmaxage br0 0' can be used to disable learning), and the topology change flag isn't cleared for max_age seconds, so nothing is learnt during that time.

It seems that the intent of hold_time() is to expire entries that are older than forward_delay seconds at the time of the topology change, which it does, but then it keeps on checking this expiry again for max_age seconds, and bases these checks on the current time rather than the time of the change.

A quick fix for the forward delay 0 case would be to skip the topology change check if stp is disabled, but if I understand things correctly then the expiry isn't right for non-zero cases either.

#1 lore raw

Is any one willing to look into this problem? Or even aknowledege it's existence?
If the description is not clear, please let me know.

Any response would be appreciated.

regards,
Uli

Quoted 38 lines

Hi,
In July 2007 Philip Craig reported the following issue with 'forward delay'=0 in great detail without receiving an answer. Has Philip's message got lost or was his analysis wrong?
The problem becomes a real problem if you bridge a fast LAN to a slow port like a bluetooth pan for example.

https://lists.linux-foundation.org/pipermail/bridge/2007-July/005476.html
Philip Craig philipc at snapgear.com

Quoted 19 lines

Hi,

If you set the bridge forward delay to 0 with:
brctl setfd br0 0
then the bridge does not learn addresses for the first 20 seconds, and so it floods everything during this time.

The reason for this is that hold_time() returns 0 after a topology change, br_fdb_update() is a no-op if hold_time() is 0 (so that 'brctl setmaxage br0 0' can be used to disable learning), and the topology change flag isn't cleared for max_age seconds, so nothing is learnt during that time.

It seems that the intent of hold_time() is to expire entries that are older than forward_delay seconds at the time of the topology change, which it does, but then it keeps on checking this expiry again for max_age seconds, and bases these checks on the current time rather than the time of the change.

Quoted 3 lines

A quick fix for the forward delay 0 case would be to skip the topology change check if stp is disabled, but if I understand things correctly then the expiry isn't right for non-zero cases either.

_______________________________________________
Bridge mailing list
[email protected]
https://lists.linux-foundation.org/mailman/listinfo/bridge

#2 lore raw
Quoted 48 lines

Is any one willing to look into this problem? Or even aknowledege it's existence?
If the description is not clear, please let me know.

Any response would be appreciated.

regards,
Uli

Quoted 30 lines

Hi,
In July 2007 Philip Craig reported the following issue with 'forward delay'=0 in great detail without receiving an answer. Has Philip's message got lost or was his analysis wrong?
The problem becomes a real problem if you bridge a fast LAN to a slow port like a bluetooth pan for example.

https://lists.linux-foundation.org/pipermail/bridge/2007-July/005476.html
Philip Craig philipc at snapgear.com

Quoted 19 lines

Hi,

If you set the bridge forward delay to 0 with:
brctl setfd br0 0
then the bridge does not learn addresses for the first 20 seconds, and so it floods everything during this time.

The reason for this is that hold_time() returns 0 after a topology change, br_fdb_update() is a no-op if hold_time() is 0 (so that 'brctl setmaxage br0 0' can be used to disable learning), and the topology change flag isn't cleared for max_age seconds, so nothing is learnt during that time.

It seems that the intent of hold_time() is to expire entries that are older than forward_delay seconds at the time of the topology change, which it does, but then it keeps on checking this expiry again for max_age seconds, and bases these checks on the current time rather than the time of the change.

Quoted 8 lines
Quoted 3 lines

A quick fix for the forward delay 0 case would be to skip the topology change check if stp is disabled, but if I understand things correctly then the expiry isn't right for non-zero cases either.

_______________________________________________
Bridge mailing list
[email protected]
https://lists.linux-foundation.org/mailman/listinfo/bridge

Quoted 4 lines

_______________________________________________
Bridge mailing list
[email protected]
https://lists.linux-foundation.org/mailman/listinfo/bridge

Hello Stephen,
you are mentioned in the MAINTAINERS file in the linux kernel tree for "ETHERNET BRIDGE". Could you at least, please acknowledge the reception of this mail? And maybe let us know what you think of the bug report?

regards,
Uli

#3 lore raw
Quoted 52 lines · On Mon, 13 Oct 2008 18:04:37 +0200 Uli Luckas <[email protected]> wrote:
Quoted 48 lines

Is any one willing to look into this problem? Or even aknowledege it's existence?
If the description is not clear, please let me know.

Any response would be appreciated.

regards,
Uli

Quoted 30 lines

Hi,
In July 2007 Philip Craig reported the following issue with 'forward delay'=0 in great detail without receiving an answer. Has Philip's message got lost or was his analysis wrong?
The problem becomes a real problem if you bridge a fast LAN to a slow port like a bluetooth pan for example.

https://lists.linux-foundation.org/pipermail/bridge/2007-July/005476.html
Philip Craig philipc at snapgear.com

Quoted 19 lines

Hi,

If you set the bridge forward delay to 0 with:
brctl setfd br0 0
then the bridge does not learn addresses for the first 20 seconds, and so it floods everything during this time.

The reason for this is that hold_time() returns 0 after a topology change, br_fdb_update() is a no-op if hold_time() is 0 (so that 'brctl setmaxage br0 0' can be used to disable learning), and the topology change flag isn't cleared for max_age seconds, so nothing is learnt during that time.

It seems that the intent of hold_time() is to expire entries that are older than forward_delay seconds at the time of the topology change, which it does, but then it keeps on checking this expiry again for max_age seconds, and bases these checks on the current time rather than the time of the change.

Quoted 8 lines
Quoted 3 lines

A quick fix for the forward delay 0 case would be to skip the topology change check if stp is disabled, but if I understand things correctly then the expiry isn't right for non-zero cases either.

_______________________________________________
Bridge mailing list
[email protected]
https://lists.linux-foundation.org/mailman/listinfo/bridge

Quoted 4 lines

_______________________________________________
Bridge mailing list
[email protected]
https://lists.linux-foundation.org/mailman/listinfo/bridge

Quoted 9 lines

Hello Stephen,
you are mentioned in the MAINTAINERS file in the linux kernel tree for "ETHERNET BRIDGE". Could you at least, please acknowledge the reception of this mail? And maybe let us know what you think of the bug report?

regards,
Uli

Yes, I saw it but the resulting patches were too messy and was more concerned about breaking other things. It is on my todo list to go back and fix the hold and forwarding timers in general for the non STP case. For example, the hold timer keeps running but is unused when STP is disabled.

#4 lore raw
Quoted 7 lines
Quoted 5 lines · On Mon, 13 Oct 2008 18:04:37 +0200 Uli Luckas <[email protected]> wrote:

Hello Stephen,
you are mentioned in the MAINTAINERS file in the linux kernel tree for "ETHERNET BRIDGE". Could you at least, please acknowledge the reception of this mail? And maybe let us know what you think of the bug report?

Quoted 8 lines
Quoted 2 lines

regards,
Uli

Yes, I saw it but the resulting patches were too messy and was more concerned about breaking other things. It is on my todo list to go back and fix the hold and forwarding timers in general for the non STP case. For example, the hold timer keeps running but is unused when STP is disabled.

Hi Stephen,
thanks for your reply.
If there is anything I can do to help in getting this fixed, please let me know.

regards,
Uli