interrupt storm on irq 10

2009-05-27 Thread Andrew Gould
I purchased a NetGear WPN511 cardbus wireless adapter (atheros chipset)
yesterday.  The card uses irq 10, as does the firewire port and ethernet
port (fxp0) on my Dell Inspiron 8100.  The laptop is running FreeBSD 7.2
Release (generic kernel).

When I bootup the laptop with the wireless adapter in the cardbus slot, I
see messages regarding interrupt storm detected throttling interrupt
source.  This did not occur prior to adding the wireless adapter.  If I
insert the wireless adapter after bootup, I don't see the messages on the
console or as dmesg output.  Otherwise, the adapter works fine.

Are interrupt storms a problem?  Do I need to worry about them?  If so, is
there anything I can do about them?

Thanks,

Andrew
___
freebsd-questions@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-questions
To unsubscribe, send any mail to freebsd-questions-unsubscr...@freebsd.org


Re: interrupt storm on irq 10

2009-05-27 Thread David Kelly
On Wed, May 27, 2009 at 12:51:45PM -0500, Andrew Gould wrote:
[...]

 Are interrupt storms a problem?  Do I need to worry about them?  If
 so, is there anything I can do about them?

Have run across interrupt storms for the first time myself last night.
Am thinking they are from interrupt sources that interrupt handlers do
not fully support. So the interrupt is not being serviced and is
repeatedly being invoked.

Probably PCI doesn't behave the same as much simpler embedded hardware
that I am used to, but the above is what an interrupt storm looks like
on simple embedded hardware.

My source of interrupt storms was caused by a bad SATA cable. Installed
a new VIA 6421-based SATA card (selected because it was only $15) and
two new hard drives for the purpose of copying files off two older
drives. New drives were detected but ad4 did not work when ad6 did.
Swapped drives and the other drive on ad6 worked. Thought the card was
bad but decided to try swapping cables which fixed ad4 and broke ad6.
Ergo, bad cable.

-- 
David Kelly N4HHE, dke...@hiwaay.net

Whom computers would destroy, they must first drive mad.
___
freebsd-questions@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-questions
To unsubscribe, send any mail to freebsd-questions-unsubscr...@freebsd.org


Re: interrupt storm on irq 10

2009-05-27 Thread Wojciech Puchar

My source of interrupt storms was caused by a bad SATA cable. Installed
a new VIA 6421-based SATA card (selected because it was only $15) and
two new hard drives for the purpose of copying files off two older
drives. New drives were detected but ad4 did not work when ad6 did.
Swapped drives and the other drive on ad6 worked. Thought the card was
bad but decided to try swapping cables which fixed ad4 and broke ad6.
Ergo, bad cable.

anyway very strange controller reaction to that case. i can hardly believe 
the way cable have to be broken to produce interrupt storm.


maybe this way:

controller sends message to drive, bad cable causes CRC errors, hard 
drive reacts with some message for that, interrupt is generated, driver in 
case of detected transmission problem instantly resends last commands, 
situation repeats.


But it's contrary to what you said that interrupt storm are lots of 
interrupts that are not serviced by any driver.


I'm not telling that you are wrong that cable produced this, but i can't 
find any explanation for that.


any idea?
___
freebsd-questions@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-questions
To unsubscribe, send any mail to freebsd-questions-unsubscr...@freebsd.org


Re: interrupt storm on irq 10

2009-05-27 Thread David Kelly
On Wed, May 27, 2009 at 09:30:19PM +0200, Wojciech Puchar wrote:
 
 I'm not telling that you are wrong that cable produced this, but i
 can't find any explanation for that.

One two-port controller card, two drives, two cables. Interrupt storms
move from one port to the other with the suspect cable no matter which
drive is connected to that cable, no matter which port it is connected
to.

Two supposedly identical SATA cables purchased together. Will purchase
new cables to try tonight.

-- 
David Kelly N4HHE, dke...@hiwaay.net

Whom computers would destroy, they must first drive mad.
___
freebsd-questions@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-questions
To unsubscribe, send any mail to freebsd-questions-unsubscr...@freebsd.org