kern/114597: System hangs at SCSI bus reset with dual HBAs

Juergen Dankoweit Juergen.Dankoweit at FreeBSD-Onkel.de
Sun Jul 15 08:50:02 UTC 2007


>Number:         114597
>Category:       kern
>Synopsis:       System hangs at SCSI bus reset with dual HBAs
>Confidential:   no
>Severity:       non-critical
>Priority:       high
>Responsible:    freebsd-bugs
>State:          open
>Quarter:        
>Keywords:       
>Date-Required:
>Class:          sw-bug
>Submitter-Id:   current-users
>Arrival-Date:   Sun Jul 15 08:50:02 GMT 2007
>Closed-Date:
>Last-Modified:
>Originator:     Juergen Dankoweit
>Release:        FreeBSD 7.0 current 06/2007
>Organization:
>Environment:
I only can tell you that it is FreeBSD 7.0 current 06/2007 with GENERIC kernel
>Description:
The following tests were made on production machines during a short downtime.
During boot time FreeBSD 7.0 current GENERIC hangs while examine the SCSI bus. Only in "Safe Mode" it boots complete into login prompt.

I have tested the following cases:
normal boot: ACPI enabled and ACPI disabled
-------------------------------------------
Kernel messages
Waiting 5 sec to settle SCSI device
(noperiph: sym0:0:-1:-1) SCSI BUS reset delivered
(noperiph: sym1:0:-1:-1) SCSI BUS reset delivered
After this the machine hangs, no keyboard entries are possible, only hardware reset

Verbose logging:
----------------
Waiting 5 sec to settle SCSI device
(noperiph: sym0:0:-1:-1) SCSI BUS reset delivered
sym0: enabling clock multiplier
sym0: download SCSI SCRIPTS
(noperiph: sym1:0:-1:-1) SCSI BUS reset delivered
sym1: enabling clock multiplier
sym1: download SCSI SCRIPTS
After this the machine hangs, no keyboard entries are possible, only hardware reset

>From FreeBSD 5.5 I know that now the following message should be shown:
(noperiph: sym1:0:-1:-1) change from SE to SE

FreeBSD 4.11, 5.3 realese #3, 5.4 and 5.5 run perfect with this hardware constellation. But FreeBSD 6.x and 7.x hang at the described position. I have tested this on eight other systems with two HBAs and they show the same behavior. All were SMP machines and the CPUs were correctly detected.
>How-To-Repeat:
Restart the system with FreeBSD 6.x or 7.x
>Fix:


>Release-Note:
>Audit-Trail:
>Unformatted:


More information about the freebsd-bugs mailing list