Dear Simone,

it is a very interesting discussion and I am deeply looking into your patch.

On Wed, Mar 09, 2011 at 04:29:05PM +0100, Simone Abbakus wrote:
> In the patch in attachment you can find a "proof of concept" about the
> improvements for vde_switch packets queue management I did (the
> interesting part is in packetq.c).
There are other changes in your patch not related to the dinamic timeout
of the packet queue.
- you added an empty element at the end of struct option arrays.
this change is wrong because optcpy concatenates all the option arrays
of the submodules and adds one empty element (optail) at the end.
> The basic idea is to change the TIMEOUT dynamically, based on the
> current queue status to avoid queue fullness. "If the queue is filling
> up, try to manage it more frequently".
It is an idea. I have some doubts: if the queue is filling up for a congestion
somewhere in the network, managing the queue more frequently could 
lead to a worse situation....
> As I stated somewhere else, the microsecond resolution of poll is not
> enough, please consider to switch to ppoll coupled with clock_gettime
> instead of gettimeofday.
Yes. I think I wrote the vde packetq before ppoll entered the kernel
as a syscall (2.6.16).
....I'll try to change packetq to permit different timeouts and queuing 
disciplines so that we'll be able to test multiple scenarios.

thank you.

renzo davoli

Colocation vs. Managed Hosting
A question and answer guide to determining the best fit
for your organization - today and in the future.
vde-users mailing list

Reply via email to