os/doc
3
0
Fork 0
Code Issues Pull requests Projects Releases Wiki Activity

Clarify the faq entry about witness lock order reversals.

Browse source 
Robert's description Witness includes mentions of "false-positives",
explain that "false-positives" from witness indicate a more serious
bug elsewhere such as a page-fault, corruption or error in mutex naming.
This commit is contained in:
Alfred Perlstein 2007-03-28 19:21:38 +00:00
parent 686317ff74
commit 6a7eedb282
Notes: svn2git 2020-12-08 03:00:23 +00:00 
svn path=/head/; revision=29877
1 changed files with 6 additions and 0 deletions
Download patch file Download diff file Expand all files Collapse all files

6
en_US.ISO8859-1/books/faq/book.sgml
View file

@ -4045,6 +4045,12 @@ kern.timecounter.hardware: TSC -&gt; i8254</screen>
warnings :-).</para>
</blockquote>
<note>
<para>A so-called "false positive" is actually Witness tripping
over a much more serious bug. Such bugs are typically a page
fault or memory corruption experienced inside of the kernel, or
a name clash with mutexes.</para>
</note>
<note>
<para>See <ulink
url="http://sources.zabbadoz.net/freebsd/lor.html">Bjoern