>
>
> On Wed, 11 Jun 1997, Richard B. Johnson wrote:
>
> > On Wed, 11 Jun 1997, yuri mironoff wrote:
> >
> > I submitted it a few days ago. It fixes a problem with old '386SXes and
> > their interrupt controllers. It ACKs the cascade with a SPECIFIC
> > EOL so other pending interrupt(s) still stay pending and are not lost.
> >
>
> Think you could work that into a command line option? Or would that be a
> 'Bad Idea' without using selfmodifing code or something so hackish?
>
> If you can then I would think it should be included into the kernel right
> away at least with it defaulting to off.. If is solves the problems you
> think it does then we should identify what it breaks (If I recall OLD
> linux kernels used specfic. EOIs so there must be someplace they dont
> work) If it's only a certian stupid chip then we can detect it and work
> around it... And as long as its a commandline option then it can be set
> accordingly..
It could certainly be a "compile" option. A command-line option might
not work because, if it doesn't work, you never get to read the
command-line!
I have a great deal more faith in specific EOIs than otherwise and,
frankly, I don't see how the current ACKs work as well as they do!
There may have been another problem that was fixed by changing something
in addition to changing the interrupt ACKs. The non-specific EOIs
may very well do nothing helpful and, sometimes, hurt things. It's
hard to know.
Cheers,
DJ
-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
Richard B. Johnson
Analogic Corporation
Email : rjohnson@analogic.com, johnson@analogic.com
Penguin : Linux version 2.1.42 on an i586 machine (66.15 BogoMips).
Warning : It's hard to stay on the trailing edge of technology.
-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-