Skip site navigation (1)Skip section navigation (2)
Date:      Thu, 23 Feb 2006 11:37:26 -0500
From:      John Baldwin <jhb@freebsd.org>
To:        freebsd-hardware@freebsd.org
Cc:        "Gustavo A. Baratto" <gbaratto@superb.net>
Subject:   Re: interrupts for ATA and BGE
Message-ID:  <200602231137.28409.jhb@freebsd.org>
In-Reply-To: <016e01c6380c$c98b7240$6500a8c0@garrincha>
References:  <016e01c6380c$c98b7240$6500a8c0@garrincha>

next in thread | previous in thread | raw e-mail | index | archive | help
On Wednesday 22 February 2006 19:04, Gustavo A. Baratto wrote:
> Two of our servers have an onboard broadcom NIC, and I noticed that in both
> of them, the number of interrupts for the ata device is always the same as
> the onboard bge device. I don't really understand what an ATA driver has to
> do with the bge0, and worse, do these numbers below mean that there are two
> interrupts going on where it should be just one?
> 
> Thanks for any input...
> 
> # vmstat -i
> interrupt                          total       rate
> irq1: atkbd0                          38          0
> irq6: fdc0                            12          0
> irq13: npx0                            1          0
> irq14: ata0                           58          0
> irq15: ata1                    192549752        165
> irq16: ahc0                      1795054          1
> irq17: ahc1                           15          0
> irq30: fxp0                    545717199        468
> irq31: bge0                    192987168        165
> cpu0: timer                   2330016902       1999
> Total                         3263066199       2800
> 
> 
> # Interrupt output from "systat -vmstat". The numbers for ata on icq 15 and
> bge on irq31, are always the same here:
> 2648 total
>          1: atkb
>          6: fdc0
>          13: npx
>          14: ata
> 212   15: ata
>          16: ahc
>          17: ahc
> 222   30: fxp
> 212   31: bge
> 2002 cpu0: time

This is probably an instance of the interrupt aliasing problem with certain
PXH bridges in Intel server chipsets.  Unfortunately there is not a workaround
as it the result of brain damage in the PXH design itself.

-- 
John Baldwin <jhb@FreeBSD.org>  <><  http://www.FreeBSD.org/~jhb/
"Power Users Use the Power to Serve"  =  http://www.FreeBSD.org



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?200602231137.28409.jhb>