Bill Davidsen <davidsen@tmr.com> wrote:
>
> On Fri, 17 Jan 2003, Andrew Morton wrote:
>
> > http://www.zip.com.au/~akpm/linux/patches/2.5/2.5.59/2.5.59-mm1/
>
> > -rcf.patch
> >
> > run-child-first didn't seem to help anything, and an alarming number of
> > cleanups and fixes were needed to get it working right. Later.
>
> I don't know about right, it seems to make threaded applications
> originally developed on BSD work better (much lower context switching).
> Anyone know if BSD does rcf? This may be an artifact of...
"seems to make"? This is too vague for me to comment on, unfortunately.
What applications? What measurements have been made?
It can only affect creation of new threads, not the switching between extant
ones.
> > +ext3-scheduling-storm.patch
> >
> > Fix the bug wherein ext3 sometimes shows blips of 100k context
> > switches/sec.
>
> Is this a 2.5 bug only? Does this need to be back ported to 2.4? Perhaps
> this is why I have ctx rate problems and some other sites don't with a
> certain application. Very commercial, unfortunately.
>
The problem has existed in 2.4 since 2.4.20-pre5. The context switch
problem will only exhibit for small periods of time (say, 10's to 100's of
milliseconds) when the filesystem is under heavy write load.
A patch for 2.4 is at
http://www.zip.com.au/~akpm/linux/patches/2.4/2.4.20/ext3-scheduling-storm.patch
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
This archive was generated by hypermail 2b29 : Thu Jan 23 2003 - 22:00:27 EST