the state of SMP workingness?

James B. Hiller jhiller
Fri Dec 5 20:36:02 PST 2003

> Can anyone comment on how their smp box is performing, or where else I 
> should be looking? When I get some time, and if i remember, i'll do some 
> testing to boot in uniprocessor mode.

I have two different SMP motherboards, a Tyan S1662D that runs two
Pentium Pro 200s, and a Tyan S2450 that runs two AMD Athlon Palomino
1.2s.  Both have worked flawlessly with hostap and three different
Prism 2.5 cards (Zcomax XI-626, Netgear MA 311, and DLink DWL-520,
revs A1 and B1) since I started using hostap, which was 0.0.4.

I think you are probably right that the issue isn't as much to do
with SMP-ness as it is with something else.  Not at all suggesting that
the following could be related, but I offer it as a data point:

I use hpoj and hpijs to make my HP all-in-one device work with linux.
There is an issue that keeps hpoj from working when one uses SMP,
a USB HP all-in-one, and hpoj.  For quite awhile, the blame was being
laid on linux kernel SMP code, but the hpoj developer/maintainer later
determined the bug was actually in the linux kernel USB code, and only
manifested when using SMP AND hpoj.  I will attest to this effect, as my
all-in-one is parallel port, and I've never had any problems.  Also,
I have run other USB devices, and the hang doesn't happen.  So this
would support the idea that the bug is in the USB code, and it only
happens when the three facilities are used together.

You can get more particular details by browsing hpoj-devel on

Hope all this info is of some use.


More information about the Hostap mailing list