Timur Tabi <ttabi@interactivesi.com> writes:
> ** Reply to message from Petko Manolov <petkan@spct.net> on Fri, 09 Jun 2000
> 16:56:32 +0300
>
>
> > Moving all drivers in user level is *bad* idea.
>
> The PowerPC version of OS/2 (aka "WorkPlace OS") has drivers in user space.
> That by itself isn't bad. The problem is that, like everything else in user
> space, drivers have priorities and are scheduled. So for instance, under heavy
> disk I/O, the mouse pointer would jump because the mouse driver couldn't get
> enough CPU.
I don't see much difference for the mouse driver irq handler being unable to get enough
CPU (or, rather, to get the CPU on time) because a monolithic kernel
services disk interrupts and the mouse driver *task* being unable to get CPU
because a micro-kernel schedules the disk driver *task*.
I.e. this can happen regardless of the kernel architecture.
The problem is not that the drivers have priorities and are scheduled.
There would be a problem if the drivers had *improper* priorities and
were *improperly" scheduled.
I'd expect the major problem to be the increased TLB miss ratio due
to frequently (compared to a monolithic kernel) changing address
spaces. Especially on the brain damaged Intel MMU.
Regards,
-velco
-
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 : Thu Jun 15 2000 - 21:00:19 EST