On Wed, 17 May 2000, Stephen Frost wrote:
> On Wed, 17 May 2000, Ivan Passos wrote:
>
> > I'm testing the Cyclades-PC300 card (our syncrhonous card) in several
> > different kernels (2.0.38, 2.2.15 and 2.3.99-pre8) and I've seen an
> > interesting behavior that I'd like you guys to comment on.
> >
> > In 2.2 and 2.3, if I do 'ping -f' through 3 different PC300 interfaces at
> > 4Mbps each, I get packet loss rates of around 6%. If I do just one
> > interface at a time, I get packet loss rates of 1-2%.
> >
> > In 2.0, however, it's completely different!! If I do 'ping -f' through 3
> > different PC300 interfaces at 4Mbps each, I get packet loss rates of
> > around 70%!! If I do just one interface at a time, I get packet loss rates
> > of 0%!!
> >
> > Can anyone tell me the reason for that?? Is this a known behavior /
> > problem?? What could I do to make the 2.0 behavior closer to the 2.[23]
> > behavior (which is much more scalable ...)??
>
> Uhh, backport all the 2.2.x kernel changes? Or just rename '2.2'
> to '2.0' and have yourself a ball? Is there a reason you don't just upgrade to
> a stable 2.2.x kernel?
Actually I'm not a user, I'm developing a synchronous card driver, and I
wanted to support the 2.0.x, 2.2.x and 2.3.x (and, thus, future 2.4.x)
kernels. However, after I found this problem, I'd like to know if this is
something possible to solve in driver level or if it's inherent to the 2.0
kernel series.
Later,
Ivan
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/
This archive was generated by hypermail 2b29 : Tue May 23 2000 - 21:00:13 EST