Skip site navigation (1)Skip section navigation (2)
Date:      Wed, 19 Feb 1997 15:43:12 +0100 (MET)
From:      J Wunsch <j@ida.interface-business.de>
To:        FreeBSD-gnats-submit@freebsd.org
Subject:   kern/2768: ktrace(1) -i dumps corrupted trace data
Message-ID:  <199702191443.PAA01363@ida.interface-business.de>
Resent-Message-ID: <199702191450.GAA26193@freefall.freebsd.org>

next in thread | raw e-mail | index | archive | help

>Number:         2768
>Category:       kern
>Synopsis:       ktrace(1) -i dumps corrupted trace data
>Confidential:   no
>Severity:       serious
>Priority:       low
>Responsible:    freebsd-bugs
>State:          open
>Class:          sw-bug
>Submitter-Id:   current-users
>Arrival-Date:   Wed Feb 19 06:50:02 PST 1997
>Last-Modified:
>Originator:     J Wunsch
>Organization:
interface business GmbH, Dresden
>Release:        FreeBSD 2.2-GAMMA i386
>Environment:

FreeBSD 2.2-GAMMA, probably any other version as well.

>Description:

Running ktrace -i on a process that becomes a daemon and then heavily
forks off children causes parts of the ktrace.out file being
corrupted.  In my case, the ktrace file started with 25160 NUL bytes
before the first usable entry could be found, and the behaviour to
sprinkle the trace file with odd numbers of NUL bytes continued
throughout the entire trace.  No useful pattern could be guessed where
it happens, but i suspect it is related to process switching.  By
manually deleting the junk, i've been able to reveal some usable
entries, but judging from the data contents of the dump it seems the
NUL bytes have been clobbering valid entries (as opposed to being put
in there in excess of the other dump data).


>How-To-Repeat:

Run ktrace -i on something that often forks off children.  In my case,
it's been syslogd with the (yet inofficial) log-to-pipe extensions,
but i assume something similar would happen e.g. for inetd.

A typical symptom is:

j@ida 142% kdump -f /tmp/syslogd.trace 
1937075797 ed
kdump: Cannot allocate memory.
at-7            207/udp    #AppleTal@  UNKNOWN(27745)  j@ida 143% 


>Fix:

Not known.  Maybe a VM problem?  However, 25160 ain't a mulitple of
the page size either.

>Audit-Trail:
>Unformatted:



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199702191443.PAA01363>