On Mon, 11 Feb 2002, John P. Looney wrote:
> Would that drag performance down below a 1MB/sec though, when that's the
> only traffic on the cards ?
not pci bus contention if only one card is active.. no. but it could
become an issue if all cards are active @ 100Mbit/s, definitely an
issue if they're all FD.
if the interrupts are shared then upon an interrupt each interrupt
handler has to be run. perhaps driver for common type of card is
clever and only needs to be run once for that type of card, i dont
know.
(at a guess see how many times drivers are registered in
/proc/interrupts and i guess that's how many handlers have to be run.
eg if 3c59x appears twice i /guess/ it has to be run twice. eh..
complete guess though.)
> > if so, (esp if interrupt contention) you could try the NAPI driver
> > architecture, which switches to polled io under high load (iirc).
>> Do tell where I could find info on this...
afaik: ftp://robur.slu.se/pub/Linux/net-development/NAPI/
afaik converted drivers benefit. old drivers dont, but arent harmed.
> Kate
--paulj
Maintained by the ILUG website team. The aim of Linux.ie is to
support and help commercial and private users of Linux in Ireland. You can
display ILUG news in your own webpages, read backend
information to find out how. Networking services kindly provided by HEAnet, server kindly donated by
Dell. Linux is a trademark of Linus Torvalds,
used with permission. No penguins were harmed in the production or maintenance
of this highly praised website. Looking for the
Indian Linux Users' Group? Try here. If you've read all this and aren't a lawyer: you should be!
