dillon at apollo.backplane.com
Sat Sep 13 08:46:41 PDT 2003
:-On [20030913 10:22], Jeroen Ruigrok/asmodai (asmodai at xxxxxx) wrote:
:>Is it me or is top missing nice and interrupt percentage reporting?
:Also, all userland processes run at priority 50 and niceness 0.
:That would explain some of the hangs I've been having.
:Now to see if either top and ps are just not getting the right
:information or whether we don't have a real scheduler anymore.
:Jeroen Ruigrok van der Werven <asmodai(at)wxs.nl> / asmodai / kita no mono
:PGP fingerprint: 2D92 980E 45FE 2C28 9DB7 9D88 97E6 839B 2EAC 625B
top has to deal with the fact that the userland priority only applies
while running in user mode, so it could be reporting bad values. If
you are getting scheduler related hangs you should still be able to
ctl-alt-esc into ddb> and do a 'ps' and 'trace'. If you can't break
into ddb it probably isn't the scheduler.
If you can reproduce a priority related hang (other then using rtpri to
create a high priority user process and spinning it) post it.
<dillon at xxxxxxxxxxxxx>
More information about the Bugs