Boot hangs on ips0: resetting adapter, this may take up to 5 minutes

[Oleg Sharoiko]

Scott, just a short question about ips driver (man page states you're 
one of it's authors): is it possible to use non disk devices with ips? 
I suppose not, but maybe I've missed something.

On Mon, 10 Apr 2006, Scott Long wrote:

SL>John Baldwin wrote:
SL>> On Thursday 06 April 2006 15:07, Oleg Sharoiko wrote:
SL>> 
SL>> > Hi, that's me again.
SL>> > 
SL>> > John, I've got more information on my problem:
SL>> > 
SL>> > It looks like the mis-routed interrupt is the one from ips. In my kernel
SL>> > ips is on vector 49 and bge is on vector 60. I've added 
SL>> > 	if (vector == 60)
SL>> > 		vector = 49;
SL>> > 
SL>> > to sys/amd64/amd64/local_apic.c and I have no more interrupt storm until
SL>> > bge really generates interrupt. Am I right with my conclusion about ips
SL>> > interrupt being mis-directed to bge?
SL>> 
SL>> 
SL>> Well, the vectors is the wrong thing to mess with as vector's are IDT
SL>> entries.
SL>> 
SL>> 
SL>> > There's also another interesting point: it looks like ips triggers
SL>> > interrupt on both vectors (49 and 60 - irq 28 and irq 16). Why do I think
SL>> > so?
SL>> 
SL>> 
SL>> This happens in several machines with Intel server chipsets due to a bug
SL>> in the PXH host bridges with no real workaround.
SL>> 
SL>
SL>Well, the work around is to not mask the APIC and instead let the driver
SL>handle masking or ACKing the interrupt =-)  It sucks that there are
SL>undocumented gotchas like this in the PC platform, but it is important
SL>to acknowledge them.
SL>
SL>Scott
SL>
SL>
SL>

-- 
Oleg Sharoiko.
Software and Network Engineer
Computer Center of Rostov State University.