From: John Heffner <[EMAIL PROTECTED]>
Date: Thu, 23 Mar 2006 10:33:09 -0500
> How about something like max(64k, min(4M, tcp_mem[2]*PAGE_SIZE/100))?
I think you mean "SK_STREAM_MEM_QUANTUM"? :-)
Care to cook up a patch with proper changelog and signed-off-by
lines? I'll seriously consider it.
On Thursday 23 March 2006 12:58, Rick Jones wrote:
> > For a sender, defense is more difficult because you can't throw away
> > unacknowledged data. An attacker can consume 2*mss kernel memory per ack
> > it
>
> WIth ABC in place, isn't that "up to" or "no more than?" And that is
> only if the co
For a sender, defense is more difficult because you can't throw away
unacknowledged data. An attacker can consume 2*mss kernel memory per ack it
WIth ABC in place, isn't that "up to" or "no more than?" And that is
only if the connection is still in the slow-start phase rather than
bandwidth
On Thursday 23 March 2006 03:31, David S. Miller wrote:
> From: John Heffner <[EMAIL PROTECTED]>
> Date: Wed, 22 Mar 2006 23:38:25 -0500
>
> > Given the relatively widespread availability of 100 Mbps or greater
> > connectivity on college campuses and larger companies, and the increasing
> > availa
On Thursday 23 March 2006 09:31, David S. Miller wrote:
> The key point is to keep the per-socket limits far enough away from
> the global pool limits such that it is not easy for a single entity
> to maliciously put the allocator into conservative mode and penalize
> the legitimate users.
It's p
From: John Heffner <[EMAIL PROTECTED]>
Date: Wed, 22 Mar 2006 23:38:25 -0500
> Given the relatively widespread availability of 100 Mbps or greater
> connectivity on college campuses and larger companies, and the increasing
> availability of fiber to the home (especially in places like S. Korea and
A while ago I sent in an incomplete patch to the TCP memory allocation code
to help it behave better when under memory stress. I sort of never had
enough time to follow up and finish it, and it has grown very stale by this
point. I was working on it in order to make a strong case for the default
