Andrea Arcangeli <andrea@suse.de> writes:
> On Wed, 8 Mar 2000, Jamie Lokier wrote:
>
> >Khimenko Victor wrote:
> >> You can limit PROCESS'S cpu but you can not limit KERNEL'S cpu and in
> >> your second example your system was mostly waited for disk to deliver
> >> data in kernel mode (as for first example: 2.3 should help great deal
> >> here but do not install it on production server now :-)
> >
> >Ingo's "low latency" patch should fix the kernel-hogs-CPU shouldn't it?
>
> That's not the case because you never spin while waiting disk I/O
> completation.
Actually some SCSI controllers (especially intelligent ones like Mylex
or GDT) seem to spin a potentially long time queueing packets with
interrupts off to the SCSI adapter. See all the "TLB IPI timeout" problems
reiserfs has with heavy IO load (reiserfs seems to do more aggressive
IO than ext2, resulting in longer queues, resulting in more latency needed
for queue handling)
-Andi
-
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 : Wed Mar 15 2000 - 21:00:15 EST