Thread (60 messages) 60 messages, 9 authors, 2009-07-10

Re: rib_trie / Fix inflate_threshold_root. Now=15 size=11 bits

From: Jarek Poplawski <hidden>
Date: 2009-06-28 19:56:34

On Sun, Jun 28, 2009 at 05:48:19PM +0200, Paweł Staszewski wrote:
After apply this patch something is wrong

Traffic is not forwarded
no info in dmesg / no info from bgp
and also i can't connect to bgpd process

I revert kernel to past version with first Jarek patch
Thank you very much, Pawel, for trying this. I'm starting to look for
the reason. In the meantime try to get some fib_trie stats for Robert.

Jarek P.

Paweł Staszewski pisze:
quoted

After 18 hours from apply first Jarek patch i have no info about Fix  
inflate_threshold_root
even if i make: "clear ip bgp *" on router
So i change Jarek patch from previous to this new one for test and we  
will see ...

Regards
Pawel Staszewski


Jarek Poplawski pisze:
quoted
On Sun, Jun 28, 2009 at 01:04:51PM +0200, Robert Olsson wrote:
...
 
quoted
 The memory patches and "manual RCU" are also interesting to address
 the case with PREEMTP's.
    
Since 2.6.29 looks like prefered here, and there were a lot of takes
in this thread, I attach below a combined all-in-one patch including:
- 2.6.29 -> 2.6.30 preemption disable patch
- 2 RCU vs. preemption fixes from 2.6.31-rc
- "manual RCU" patch to force vfree/kfree before root's resize (take 3)
- "automatic" inflate_threshold_root fix (take 2)

Thanks,
Jarek P.
--- (for 2.6.29.x or even .28 or .27; any testing appreciated)
diff -Nurp a/net/ipv4/fib_trie.c b/net/ipv4/fib_trie.c
--- a/net/ipv4/fib_trie.c    2009-06-27 20:25:06.000000000 +0200
+++ b/net/ipv4/fib_trie.c    2009-06-28 15:45:02.000000000 +0200
...
Keyboard shortcuts
hback out one level
jnext message in thread
kprevious message in thread
ldrill in
Escclose help / fold thread tree
?toggle this help