[Orca-users] Updated patches to add kstat(1M) syntax to kstat parser in SymbEL

Alex Kiernan alex.kiernan at gmail.com
Thu Jan 18 07:31:23 PST 2007


I mentioned this a while back, but we've only just gotten around to
implementing it. I had someone here add support for kstat(1M) syntax
to SymbEL so that the new statistics introduced as part of the GLDv3
changes could be expressed - this is the diff from USERS.MANUAL:

      +) ":module:instance:name:statistic" - This notation is inspired by
         that used by the kstat command line utility.  Each of the module,
         instance, name, or statistic specifiers may be a shell glob pattern
         or a regular expression enclosed by '/' characters.  It is
         possible to use both specifier types within a single operand.
         Leaving a specifier empty is equivalent to using the '*' glob
         pattern for that specifier.  This format differs from that in the
         kstat command line utility in that the leading colon is required to
         distinguish this form from those above, and the regular expressions
         are Extended Regular Expressions, regex(5), not Perl.  As is the
         case with specifying a module name, the number$ and name$ members
         must be present.

The patch is here (incorporating the changes we'd previously made to
support Solaris x64):

http://www.alexk1.demon.co.uk/setoolkit.diff

Hopefully if you run with this you shouldn't see core dumps as a
result of MAX_IF being incorrectly sized, at least for dmfe, bge, ipge
and e1000g interfaces - there's doubtless others which need kstat.se
updating along the same lines, i.e.:

-kstat struct "*bge"  ks_bge_network {
+kstat struct ":bge::bge[0-9]*:"  ks_bge_network {

I've not run it in anger in production yet, but it all seems okay so far.

-- 
Alex Kiernan



More information about the Orca-users mailing list