On Thursday 09 February 2006 00:07, David Stevens wrote:
> From: Stephen Hemminger <[EMAIL PROTECTED]>
>
> > IMHO converting skb->dev to skb->devindex and using ifindex sounds best.
> > It gets rid of the need to refcount as much but keeps the safety from
> > buggy protocols. Ipv6 could probably
From: Stephen Hemminger <[EMAIL PROTECTED]>
> IMHO converting skb->dev to skb->devindex and using ifindex sounds best.
> It gets rid of the need to refcount as much but keeps the safety from
> buggy protocols. Ipv6 could probably use ifindex as well.
A couple years ago, we identified a p
From: Andi Kleen <[EMAIL PROTECTED]>
Date: Wed, 8 Feb 2006 23:20:38 +0100
> But are you sure the hash table used for that would be strong enough
> to handle the load?
With RCU locking I think it is.
> And what happens when a ifindex is reused? Then packets could end up
> on the wrong interface.
On Wednesday 08 February 2006 20:12, Stephen Hemminger wrote:
> On Tue, 07 Feb 2006 16:26:01 -0800 (PST)
> "David S. Miller" <[EMAIL PROTECTED]> wrote:
>
> > From: Stephen Hemminger <[EMAIL PROTECTED]>
> > Date: Tue, 7 Feb 2006 16:19:42 -0800
> >
> > > Also, isn't a lot of the problem reduced if
Stephen Hemminger wrote:
On Wed, 08 Feb 2006 11:24:24 -0800
Ben Greear <[EMAIL PROTECTED]> wrote:
Stephen Hemminger wrote:
On Tue, 07 Feb 2006 16:26:01 -0800 (PST)
"David S. Miller" <[EMAIL PROTECTED]> wrote:
From: Stephen Hemminger <[EMAIL PROTECTED]>
Date: Tue, 7 Feb 2006 16:19:42 -080
On 2/8/06, Stephen Hemminger <[EMAIL PROTECTED]> wrote:
> On Wed, 08 Feb 2006 11:24:24 -0800
> Ben Greear <[EMAIL PROTECTED]> wrote:
>
> > Stephen Hemminger wrote:
> > > On Tue, 07 Feb 2006 16:26:01 -0800 (PST)
> > > "David S. Miller" <[EMAIL PROTECTED]> wrote:
> > >
> > >
> > >>From: Stephen Hemmi
On Wed, 08 Feb 2006 11:24:24 -0800
Ben Greear <[EMAIL PROTECTED]> wrote:
> Stephen Hemminger wrote:
> > On Tue, 07 Feb 2006 16:26:01 -0800 (PST)
> > "David S. Miller" <[EMAIL PROTECTED]> wrote:
> >
> >
> >>From: Stephen Hemminger <[EMAIL PROTECTED]>
> >>Date: Tue, 7 Feb 2006 16:19:42 -0800
> >>
Stephen Hemminger wrote:
On Tue, 07 Feb 2006 16:26:01 -0800 (PST)
"David S. Miller" <[EMAIL PROTECTED]> wrote:
From: Stephen Hemminger <[EMAIL PROTECTED]>
Date: Tue, 7 Feb 2006 16:19:42 -0800
Also, isn't a lot of the problem reduced if network devices
are affinitied?
Not for routing/firew
On Tue, 07 Feb 2006 16:26:01 -0800 (PST)
"David S. Miller" <[EMAIL PROTECTED]> wrote:
> From: Stephen Hemminger <[EMAIL PROTECTED]>
> Date: Tue, 7 Feb 2006 16:19:42 -0800
>
> > Also, isn't a lot of the problem reduced if network devices
> > are affinitied?
>
> Not for routing/firewalling, we tou
On Wednesday 08 February 2006 11:34, Eric Dumazet wrote:
> 1) Instead of storing a 2-uple {pointer,generation} (and using 12 or 16 bytes
> on 64 bits platforms), we could just use a 32 bit quantity
> [(ifindex<<8)+(gen_number)]
That would add an 2^24 netdevice limit. Someone will sooner or late
Andi Kleen a écrit :
On Wednesday 08 February 2006 01:44, David S. Miller wrote:
From: Ben Greear <[EMAIL PROTECTED]>
Date: Tue, 07 Feb 2006 16:39:52 -0800
Rick Jones wrote:
In the realm of straw ideas, how often are netdevs added and removed,
and would leaving a tombstone behind consume too
On Wednesday 08 February 2006 01:44, David S. Miller wrote:
> From: Ben Greear <[EMAIL PROTECTED]>
> Date: Tue, 07 Feb 2006 16:39:52 -0800
>
> > Rick Jones wrote:
> > > In the realm of straw ideas, how often are netdevs added and removed,
> > > and would leaving a tombstone behind consume too muc
Rick Jones wrote:
Ben Greear wrote:
Rick Jones wrote:
In the realm of straw ideas, how often are netdevs added and removed,
and would leaving a tombstone behind consume too much memory?
In certain cases...say, with vlans, you could very often create and
destroy net devices. I think that
David S. Miller wrote:
From: Ben Greear <[EMAIL PROTECTED]>
Date: Tue, 07 Feb 2006 16:39:52 -0800
Rick Jones wrote:
In the realm of straw ideas, how often are netdevs added and removed,
and would leaving a tombstone behind consume too much memory?
In certain cases...say, with vlans, you co
Ben Greear wrote:
Rick Jones wrote:
In the realm of straw ideas, how often are netdevs added and removed,
and would leaving a tombstone behind consume too much memory?
In certain cases...say, with vlans, you could very often create and
destroy net devices. I think that giving up and leaking
David S. Miller wrote:
From: Ben Greear <[EMAIL PROTECTED]>
Date: Tue, 07 Feb 2006 15:54:06 -0800
What do you think about having no ref counting, and upon removal of
a network device, we notify each logic unit that deals with skbs
or other things that link to the netdev and ask it to clean all
David S. Miller wrote:
From: Rick Jones <[EMAIL PROTECTED]>
Date: Tue, 07 Feb 2006 16:29:34 -0800
In the realm of straw ideas, how often are netdevs added and
removed, and would leaving a tombstone behind consume too much
memory?
That could work.
Another idea is to revisit the scheme of st
From: Ben Greear <[EMAIL PROTECTED]>
Date: Tue, 07 Feb 2006 16:39:52 -0800
> Rick Jones wrote:
> > In the realm of straw ideas, how often are netdevs added and removed,
> > and would leaving a tombstone behind consume too much memory?
>
> In certain cases...say, with vlans, you could very often
Rick Jones wrote:
In the realm of straw ideas, how often are netdevs added and removed,
and would leaving a tombstone behind consume too much memory?
In certain cases...say, with vlans, you could very often create and
destroy net devices. I think that giving up and leaking the memory
is not a
From: Rick Jones <[EMAIL PROTECTED]>
Date: Tue, 07 Feb 2006 16:29:34 -0800
> In the realm of straw ideas, how often are netdevs added and
> removed, and would leaving a tombstone behind consume too much
> memory?
That could work.
Another idea is to revisit the scheme of storing just the
ifindex
In the realm of straw ideas, how often are netdevs added and removed, and would
leaving a tombstone behind consume too much memory?
rick jones
-
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to [EMAIL PROTECTED]
More majordomo info at http://vger.ker
From: Stephen Hemminger <[EMAIL PROTECTED]>
Date: Tue, 7 Feb 2006 16:19:42 -0800
> Also, isn't a lot of the problem reduced if network devices
> are affinitied?
Not for routing/firewalling, we touch the destination device's
counters on input softing of the source device.
-
To unsubscribe from thi
On Tue, 07 Feb 2006 16:11:51 -0800 (PST)
"David S. Miller" <[EMAIL PROTECTED]> wrote:
> From: Ben Greear <[EMAIL PROTECTED]>
> Date: Tue, 07 Feb 2006 15:54:06 -0800
>
> > What do you think about having no ref counting, and upon removal of
> > a network device, we notify each logic unit that deals
From: Ben Greear <[EMAIL PROTECTED]>
Date: Tue, 07 Feb 2006 15:54:06 -0800
> What do you think about having no ref counting, and upon removal of
> a network device, we notify each logic unit that deals with skbs
> or other things that link to the netdev and ask it to clean all
> references to the
Eric Dumazet wrote:
David S. Miller a écrit :
From: Eric Dumazet <[EMAIL PROTECTED]>
Date: Wed, 08 Feb 2006 00:23:45 +0100
Some devices are *never* unregistered : loopback, or statically
linked drivers, thus we are refcounting them for nothing.
Statically linked drivers can have netdev's t
David S. Miller a écrit :
From: Eric Dumazet <[EMAIL PROTECTED]>
Date: Wed, 08 Feb 2006 00:23:45 +0100
Some devices are *never* unregistered : loopback, or statically linked
drivers, thus we are refcounting them for nothing.
Statically linked drivers can have netdev's that get unregistered
an
From: Eric Dumazet <[EMAIL PROTECTED]>
Date: Wed, 08 Feb 2006 00:23:45 +0100
> Some devices are *never* unregistered : loopback, or statically linked
> drivers, thus we are refcounting them for nothing.
Statically linked drivers can have netdev's that get unregistered
and free'd up. For example
Struct net_device's atomic refcnt are probably one of the hotest memory spots
in a SMP/NUMA network router or network server.
This counter is constantly incremented/decremented each time a network packet
is handled, or a IP route is added/deleted in route cache. This is *not* SMP
nor NUMA frie
28 matches
Mail list logo
