Re: [PATCH] soft interrupts

[Daniel Wagner]

Joachim Nilsson <joachim.nilsson@vmlinux.org> writes:

> In comparison to what else is missing in the Hurd right now it is
> surely a bit unimportant. And I don't mean to belittle your work
> in any way!

No offense taken.

> I don't think your patch is wrong - it looks quite good actually.
> It might just be something small like calling the SOFTINTs more/less
> often. When we have it on the table someone else might come up
> with the optimization.

Softints are handled at the softclock ipl.  So the only priority level
left is spl0 to call the handler.  You can not handle them on higher
priority, as I understand this.  

I guess the reason for less performance is that we do not include the
Linux drivers directly like in gnumach1 but using more abstracted
interface from OSKit to access the drivers, in short more overhead. 
To be sure, you have to some more serious investigation of course. 

daniel