NetBSD Problem Report #7479

Received: (qmail 26186 invoked from network); 26 Apr 1999 21:34:47 -0000
Message-Id: <199904262134.OAA00411@carl.nas.nasa.gov>
Date: Mon, 26 Apr 1999 14:34:16 -0700 (PDT)
From: mjacob@nas.nasa.gov
Reply-To: mjacob@feral.com
To: gnats-bugs@gnats.netbsd.org
Subject: recent source change reduced FRU information for Rawhide
X-Send-Pr-Version: 3.95

>Number:         7479
>Category:       port-alpha
>Synopsis:       recent source change reduced FRU information for Rawhide
>Confidential:   no
>Severity:       critical
>Priority:       medium
>Responsible:    port-alpha-maintainer
>State:          open
>Class:          sw-bug
>Submitter-Id:   net
>Arrival-Date:   Mon Apr 26 14:35:00 +0000 1999
>Closed-Date:    
>Last-Modified:  Mon Apr 26 21:54:20 +0000 1999
>Originator:     
>Release:        source ca. 4/16/99
>Organization:
	NASA Ames
>Environment:
AlphaServer 4100 Systems

>Description:

A recent change to pci_kn300.c eliminated to printing of related mcpcia
module IDs for spurious interrupts. The lines previously had been a
possibly too chatty and (wrong) irq level:

Apr 26 14:24:01 carl /netbsd: stray mcbus0 mid 4 PCI Slot 2 PCI Interrupt Pin A irq 0
Apr 26 14:24:01 carl /netbsd: stray mcbus0 mid 5 PCI Slot 1 PCI Interrupt Pin A irq 16
Apr 26 14:24:01 carl /netbsd: stray mcbus0 mid 6 PCI Slot 2 PCI Interrupt Pin A irq 0
Apr 26 14:24:01 carl /netbsd: stray mcbus0 mid 7 PCI Slot 2 PCI Interrupt Pin A irq 0
Apr 26 14:24:01 carl /netbsd: stray mcbus0 mid 5 PCI Slot 2 PCI Interrupt Pin A irq 0


The new lines are just a simple:

stray kn300 irq 0
...
stray kn300 irq 16


This loses FRU information as to where the possibly errant PCI card is located.

+The change also disables interrupts after an arbitrary and non-adjustable
+counter is reached. Spurious interrupts are a fact of life for a large number
+of cards. This particular change will render these systems wedged and needing
+rebooting in a fairly short time period.
Mon Apr 26 14:53:48 PDT 1999 mjacob

The above '+' comment is bogus- the interrupts were being disabled before-
it's still wrong but it's a separate issue from this one.

>How-To-Repeat:

>Fix:

>Release-Note:
>Audit-Trail:

From: Ross Harvey <ross@ghs.com>
To: gnats-bugs@gnats.netbsd.org, mjacob@nas.nasa.gov
Cc:  Subject: Re: port-alpha/7479: recent source change reduced FRU information for Rawhide
Date: Mon, 26 Apr 1999 14:39:41 -0700 (PDT)

 You are saying it should print the `mid' numbers?

From: Matthew Jacob <mjacob@nas.nasa.gov>
To: Ross Harvey <ross@ghs.com>
Cc: gnats-bugs@gnats.netbsd.org
Subject: Re: port-alpha/7479: recent source change reduced FRU information
 for Rawhide
Date: Mon, 26 Apr 1999 14:52:40 -0700 (PDT)

 > You are saying it should print the `mid' numbers?

 The MID numbers were being printed for a reason. They have a specific
 physical location. The same for the PCI slot numbers. They have a
 physical location too. If there's a card that's having spurious interrupts
 (e.g., for no driver), having this information seemed useful to have to
 try and locate a problem. In fact, it *is* useful as it isolates (at NAS)
 the HIPPI spurious interrupts from the Qlogic PCI card spurious interrupts
 (and can and is then correlated with Hippi F/W crashes).

 Bridges could possibly foul up the ultimate source of the offending card,
 but at least you can track it to a certain point.

 I'll edit this PR and change the comment on the interrupt disable code-
 it really isn't new- it was there before, and it's too draconian- separate
 problem.


 -matt


>Unformatted:

Home	PR Database Search