* FWD: RE: excessive interrupts on network cards
@ 2001-09-25 23:04 Chris Newton
0 siblings, 0 replies; 4+ messages in thread
From: Chris Newton @ 2001-09-25 23:04 UTC (permalink / raw)
To: linux-kernel
Francois Romieu got me to run 'lspci -x', and came to the conclusion that the
net card and scsi card are sharing IRQs...
Someone just told me tha the had had a sound card and a network card sharing
IRQs, and that caused the network card to generate 'oodles of interrupts for
no apparent reason'.
This on the right track?
Thanks
Chris
>===== Original Message From Francois Romieu <romieu@cogenit.fr> =====
Chris Newton <newton@unb.ca> :
[...]
> 00:02.0 Ethernet controller: Intel Corporation 82557 [Ethernet Pro 100] (rev
> 08)
> 00: 86 80 29 12 17 01 90 02 08 00 00 02 08 20 00 00
> 10: 00 20 10 fe c1 ec 00 00 00 00 00 fe 00 00 00 00
> 20: 00 00 00 00 00 00 00 00 00 00 00 00 28 10 9b 00
> 30: 00 00 00 fd dc 00 00 00 00 00 00 00 0b 01 08 38
^
[...]
> 01:02.0 SCSI storage controller: Adaptec 7899P (rev 01)
> 00: 05 90 cf 00 16 01 b0 02 01 00 00 01 08 20 80 80
> 10: 01 dc 00 00 04 10 00 f9 00 00 00 00 00 00 00 00
> 20: 00 00 00 00 00 00 00 00 00 00 00 00 28 10 ce 00
> 30: 00 00 00 f8 dc 00 00 00 00 00 00 00 05 01 28 19
^
[...]
> 01:02.1 SCSI storage controller: Adaptec 7899P (rev 01)
> 00: 05 90 cf 00 16 01 b0 02 01 00 00 01 08 20 80 80
> 10: 01 d8 00 00 04 00 00 f9 00 00 00 00 00 00 00 00
> 20: 00 00 00 00 00 00 00 00 00 00 00 00 28 10 ce 00
> 30: 00 00 00 f8 dc 00 00 00 00 00 00 00 0b 02 28 19
^
[...]
> 01:06.0 Ethernet controller: 3Com Corporation 3c980-TX 10/100baseTX NIC
> [Python-T] (rev 78)
> 00: b7 10 05 98 17 01 10 02 78 00 00 02 08 20 00 00
> 10: 81 d4 00 00 00 24 00 f9 00 00 00 00 00 00 00 00
> 20: 00 00 00 00 00 00 00 00 00 00 00 00 b7 10 00 10
> 30: 00 00 00 f8 dc 00 00 00 00 00 00 00 05 01 0a 0a
^
[...]
> 01:08.0 Ethernet controller: 3Com Corporation 3c980-TX 10/100baseTX NIC
> [Python-T] (rev 78)
> 00: b7 10 05 98 17 01 10 02 78 00 00 02 08 20 00 00
> 10: 01 d4 00 00 00 20 00 f9 00 00 00 00 00 00 00 00
> 20: 00 00 00 00 00 00 00 00 00 00 00 00 b7 10 00 10
> 30: 00 00 00 f8 dc 00 00 00 00 00 00 00 0b 01 0a 0a
^
Each of your ethernet adapter shares an irq with a scsi controller.
I had some results pulling some cards from the PCI slots and moving
the network adapter around until its irq differs but I won't claim
it's the cure for your problem (it was on HP Netserver motherboards).
--
Ueimor
^ permalink raw reply [flat|nested] 4+ messages in thread
* FWD: RE: excessive interrupts on network cards
@ 2001-09-25 23:19 Chris Newton
2001-09-25 23:58 ` Tim Moore
0 siblings, 1 reply; 4+ messages in thread
From: Chris Newton @ 2001-09-25 23:19 UTC (permalink / raw)
To: linux-kernel
Ok, I checked... and yup, lots of sharing:
one scsi channel on irq 5, one net card on 5
one scsi channel on irq 11, one net card (the busy one) on 11
but..
I disabled all the serial, parallel ports, and USB, rebooted, and the irq
sharing cleaned all up..
Everything now has it's own IRQ...
However, the problem still exists. The network card is generating 3 times
as many interrupts as there are packets recieved on this wire.
[root@phantom bin]# ifconfig eth0;sleep 1;ifconfig eth0
eth0 Link encap:Ethernet HWaddr 00:01:02:CC:8C:88
inet addr:192.168.1.2 Bcast:192.168.1.255 Mask:255.255.255.0
UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1
RX packets:2024138 errors:0 dropped:0 overruns:1 frame:0
TX packets:0 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:100
Interrupt:20 Base address:0xd480
eth0 Link encap:Ethernet HWaddr 00:01:02:CC:8C:88
inet addr:192.168.1.2 Bcast:192.168.1.255 Mask:255.255.255.0
UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1
RX packets:2031162 errors:0 dropped:0 overruns:1 frame:0
TX packets:0 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:100
Interrupt:20 Base address:0xd480
--------------------------
procinfo -D
Memory: Total Used Free Shared Buffers Cached
Mem: 900988 432032 468956 0 6756 345412
Swap: 449812 0 449812
Bootup: Tue Sep 25 20:00:55 2001 Load average: 0.61 0.67 0.32 1/31 839
user : 0:00:00.96 9.6% page in : 756
nice : 0:00:00.00 0.0% page out: 1452
system: 0:00:01.29 12.9% swap in : 0
idle : 0:00:07.75 77.5% swap out: 0
uptime: 0:07:54.17 context : 43253
irq 0: 500 timer irq 16: 131 eth2
irq 1: 0 keyboard irq 20: 22266 eth0
irq 2: 0 cascade [4] irq 21: 0 eth1
irq 6: 0 irq 30: 0 aic7xxx
irq 12: 0 irq 31: 121 aic7xxx
irq 14: 0 ide0
------------
Francois Romieu got me to run 'lspci -x', and came to the conclusion that the
net card and scsi card are sharing IRQs...
Someone just told me tha the had had a sound card and a network card sharing
IRQs, and that caused the network card to generate 'oodles of interrupts for
no apparent reason'.
This on the right track?
Thanks
Chris
>===== Original Message From Francois Romieu <romieu@cogenit.fr> =====
Chris Newton <newton@unb.ca> :
[...]
> 00:02.0 Ethernet controller: Intel Corporation 82557 [Ethernet Pro 100] (rev
> 08)
> 00: 86 80 29 12 17 01 90 02 08 00 00 02 08 20 00 00
> 10: 00 20 10 fe c1 ec 00 00 00 00 00 fe 00 00 00 00
> 20: 00 00 00 00 00 00 00 00 00 00 00 00 28 10 9b 00
> 30: 00 00 00 fd dc 00 00 00 00 00 00 00 0b 01 08 38
^
[...]
> 01:02.0 SCSI storage controller: Adaptec 7899P (rev 01)
> 00: 05 90 cf 00 16 01 b0 02 01 00 00 01 08 20 80 80
> 10: 01 dc 00 00 04 10 00 f9 00 00 00 00 00 00 00 00
> 20: 00 00 00 00 00 00 00 00 00 00 00 00 28 10 ce 00
> 30: 00 00 00 f8 dc 00 00 00 00 00 00 00 05 01 28 19
^
[...]
> 01:02.1 SCSI storage controller: Adaptec 7899P (rev 01)
> 00: 05 90 cf 00 16 01 b0 02 01 00 00 01 08 20 80 80
> 10: 01 d8 00 00 04 00 00 f9 00 00 00 00 00 00 00 00
> 20: 00 00 00 00 00 00 00 00 00 00 00 00 28 10 ce 00
> 30: 00 00 00 f8 dc 00 00 00 00 00 00 00 0b 02 28 19
^
[...]
> 01:06.0 Ethernet controller: 3Com Corporation 3c980-TX 10/100baseTX NIC
> [Python-T] (rev 78)
> 00: b7 10 05 98 17 01 10 02 78 00 00 02 08 20 00 00
> 10: 81 d4 00 00 00 24 00 f9 00 00 00 00 00 00 00 00
> 20: 00 00 00 00 00 00 00 00 00 00 00 00 b7 10 00 10
> 30: 00 00 00 f8 dc 00 00 00 00 00 00 00 05 01 0a 0a
^
[...]
> 01:08.0 Ethernet controller: 3Com Corporation 3c980-TX 10/100baseTX NIC
> [Python-T] (rev 78)
> 00: b7 10 05 98 17 01 10 02 78 00 00 02 08 20 00 00
> 10: 01 d4 00 00 00 20 00 f9 00 00 00 00 00 00 00 00
> 20: 00 00 00 00 00 00 00 00 00 00 00 00 b7 10 00 10
> 30: 00 00 00 f8 dc 00 00 00 00 00 00 00 0b 01 0a 0a
^
Each of your ethernet adapter shares an irq with a scsi controller.
I had some results pulling some cards from the PCI slots and moving
the network adapter around until its irq differs but I won't claim
it's the cure for your problem (it was on HP Netserver motherboards).
--
Ueimor
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
^ permalink raw reply [flat|nested] 4+ messages in thread
* Re: FWD: RE: excessive interrupts on network cards
2001-09-25 23:19 Chris Newton
@ 2001-09-25 23:58 ` Tim Moore
0 siblings, 0 replies; 4+ messages in thread
From: Tim Moore @ 2001-09-25 23:58 UTC (permalink / raw)
To: Chris Newton; +Cc: linux-kernel
Just for grins what does 'procinfo -DSn2' say?
Chris Newton wrote:
> ...
> uptime: 0:07:54.17 context : 43253
>
> irq 0: 500 timer irq 16: 131 eth2
> irq 1: 0 keyboard irq 20: 22266 eth0
> irq 2: 0 cascade [4] irq 21: 0 eth1
> irq 6: 0 irq 30: 0 aic7xxx
> irq 12: 0 irq 31: 121 aic7xxx
> irq 14: 0 ide0
--
^ permalink raw reply [flat|nested] 4+ messages in thread
* RE: FWD: RE: excessive interrupts on network cards
@ 2001-09-26 15:41 Chris Newton
0 siblings, 0 replies; 4+ messages in thread
From: Chris Newton @ 2001-09-26 15:41 UTC (permalink / raw)
To: Tim Moore; +Cc: linux-kernel
Tim sent me this... and, yes, the count is much lower... looking around, I
found 'mpstat', which returns processor statistics.
From this output, I would say that procinfo has a bug in the -d and -D
options, or in the man page.
Thanks, for all the help guys, but this seems to be right, according to
mpstat anyways.
Thanks a BUNCH.
Chris
[root@phantom /root]# mpstat 1
Linux 2.4.10 (phantom.csd.unb.ca) 09/26/2001
12:29:48 PM CPU %user %nice %system %idle intr/s
12:29:49 PM all 3.00 0.00 9.50 87.50 5711.00
12:29:50 PM all 5.50 0.00 8.00 86.50 6139.00
12:29:51 PM all 2.00 0.00 11.00 87.00 5976.00
12:29:52 PM all 3.50 0.00 10.00 86.50 5744.00
12:29:53 PM all 5.50 0.00 9.00 85.50 5986.00
12:29:54 PM all 7.00 0.00 10.00 83.00 5904.00
12:29:55 PM all 5.00 0.00 6.50 88.50 5771.00
>===== Original Message From Tim Moore <timothymoore@bigfoot.com> =====
>Just for grins what does 'procinfo -DSn2' say?
>
>Chris Newton wrote:
>> ...
>> uptime: 0:07:54.17 context : 43253
>>
>> irq 0: 500 timer irq 16: 131 eth2
>> irq 1: 0 keyboard irq 20: 22266 eth0
>> irq 2: 0 cascade [4] irq 21: 0 eth1
>> irq 6: 0 irq 30: 0 aic7xxx
>> irq 12: 0 irq 31: 121 aic7xxx
>> irq 14: 0 ide0
>--
^ permalink raw reply [flat|nested] 4+ messages in thread
end of thread, other threads:[~2001-09-26 15:32 UTC | newest]
Thread overview: 4+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2001-09-26 15:41 FWD: RE: excessive interrupts on network cards Chris Newton
-- strict thread matches above, loose matches on Subject: below --
2001-09-25 23:19 Chris Newton
2001-09-25 23:58 ` Tim Moore
2001-09-25 23:04 Chris Newton
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox