>Number:         36824
>Category:       kern
>Synopsis:       CF card reading creates 97% interrupt storm
>Confidential:   no
>Severity:       serious
>Priority:       medium
>Responsible:    kern-bug-people
>State:          open
>Class:          sw-bug
>Submitter-Id:   net
>Arrival-Date:   Fri Aug 24 01:20:00 +0000 2007
>Originator:     George Michaelson
>Release:        NetBSD 4.99.28
>Organization:
George Michaelson                      email:           ggm@apnic.net
Senior Technical Officer, APNIC          sip:      ggm@voip.apnic.net
http://www.apnic.net                   phone:         +61 7 3858 3100
>Environment:
	
	
System: NetBSD garlique.algebras.org 4.99.28 NetBSD 4.99.28 (GGM_ACPI) #4: Mon Aug 13 09:57:37 EST 2007 ggm@garlique.algebras.org:/data/Build/obj/usr/src/sys/arch/i386/compile/GGM_ACPI i386
Architecture: i386
Machine: i386
>Description:
	I mount a SanDisk UltraII 2.0Gb CF card, formatted as ms-dos
	(camera) and cp -r its directory contents.

	the Card entry causes this dmesg:

wdc2 at pcmcia1 function 0: <SanDisk, SDP, 5/3 0.6>
wdc2: i/o mapped mode
atabus2 at wdc2 channel 0
wd1 at atabus2 drive 0: <SanDisk SDCFH-2048>
wd1: drive supports 4-sector PIO transfers, LBA addressing
wd1: 1953 MB, 3970 cyl, 16 head, 63 sec, 512 bytes/sect x 4001760 sectors
wd1: drive supports PIO mode 4
wd1 detached

	While this is going on, X, xmms (audio) ground to a stuttering
	halt. Ok, nothing unusual there, but it was at the extreme end.

	I had a top up, and noticed the system was 97% interrupt, close
	to zero in system and usertime.

	

>How-To-Repeat:
	well.. its probably not that simple, jobmix and all that
	but anyone with an interactive jobmix laptop and a CF reader
	could try.
	
>Fix:
	nfi. But, it might be worth instrumenting something to tell if
	the amount of interrupt service is getting excessive maybe and
	log it?
	

>Unformatted: