Messages in this thread | | | From | Bryan -TheBS- Smith <> | Subject | Re: Kernel 2.2.14, dirty buffers, stalls in interactivity of system/NFS-clients ... | Date | Tue, 28 Mar 2000 16:25:23 -0500 |
| |
Re: Kernel 2.2.14, dirty buffers, stalls in interactivity of system/NFS-clients ...
On Tue, 28 Mar 2000, Mark Hahn wrote: > I've never seen irq priority effect a system in any meaninful way. > do you really see the vortex starving the nic?
Personally, I think the Vortex is starving the whole PCI bus.
> besides, isn't irq7 usually the parport, and wouldn't that > starve most IO?
Nope, this is how the interrupt controller is mapped:
0 1 2 <---------------- 9 3 | 10 4 | 11 5 | 12 6 \ 13 7 \ 14 15
The interrupts 9-15 are mapped through IRQ 2. The history behind this is the old Intel controllers. For the PC/AT, a 2nd one was added. All current mainboard chipsets emulate this scheme.
I've had problems with any IDE-integrated chipsets and use of IRQs 14/15 regardless of their enabling/disabling on mainboard. I sure wish this wasn't such a production system or I'd give this a quick "try."
> I have a system that at the momemnt doesn't think its parport > has a fifo, so does 50-80K int/second during printing, with zero > noticable problem. besides, nics are not especially sensitive > to interrupt-handler latency...
I really have no idea what it is.
I have run into one other person complaining to "too lazy" dirty buffer flushes with the 2.2.14 kernel. I upgraded to the 2.2.14 kernel with numerous NFS patches to fix stability issues with my Solaris clients.
Frankly, I didn't like the setup when I first got here. I don't like using Linux to serve out Solaris clients. Linux, FreeBSD, Irix, yes, Solaris, NO. I have a bone to pick with the consultant on the hardware selection too.
-- TheBS
-- Bryan "TheBS" Smith -- Engineer, IT Professional and Hacker E-mail: mailto:thebs@theseus.com,b.j.smith@ieee.org Disclaimer: http://www.SmithConcepts.com/legal.html ************************************************************* TheBS ... Serving E-mail filters to /dev/null since 1989
- 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/
| |