misc/71603: "systat -v" enters infinite loop

David Kelly dkelly at hiwaay.net
Sat Sep 11 11:00:49 PDT 2004


>Number:         71603
>Category:       misc
>Synopsis:       "systat -v" enters infinite loop
>Confidential:   no
>Severity:       serious
>Priority:       medium
>Responsible:    freebsd-bugs
>State:          open
>Quarter:        
>Keywords:       
>Date-Required:
>Class:          sw-bug
>Submitter-Id:   current-users
>Arrival-Date:   Sat Sep 11 18:00:47 GMT 2004
>Closed-Date:
>Last-Modified:
>Originator:     David Kelly
>Release:        5.2.1-p9
>Organization:
>Environment:
FreeBSD Opus.home 5.2.1-RELEASE-p9 FreeBSD 5.2.1-RELEASE-p9 #2: Fri Aug 27 22:13:02 CDT 2004     dkelly at Opus.home:/usr/obj/usr/src/sys/OPUS  i386

>Description:
Apparently triggered by some sort of load condition, "systat -v" will enter an infinite loop and consume all available CPU cycles. The application accepts key input but does not act on commands. Control-C cleanly kills the runaway systat.

My Athlon 800 MHz system with same FreeBSD exhibits this problem more than my new Dell PE400SC P4-2.8G w/HT.

Hyperthreading is disabled on the P4 in the BIOS.

It sounds odd and unrelated but the P4 did not have this problem prior to upgrading to KDE 3.3.0 this past week. The Athlon still has kde-3.2.3. I usually run "systat -v" in a KDE Terminal.

I don't believe it is related to swap activity as the 1GB machine is is using very little swap.

>How-To-Repeat:
 I have not found a reliable means of creating the situation but it seems to happen under heavy disk activity with multiple processes using single drive, and each process is able to run the drive at its limits. Such as md5, par2, or urar.

Does not seem to matter if the drive is on parallel ATA or serial ATA.
>Fix:
      
>Release-Note:
>Audit-Trail:
>Unformatted:


More information about the freebsd-bugs mailing list