Re: Linux-2.2.0 bad VM behaviour "dd if=/dev/zero of=/dev/hdc bs=256k"

Heinz Mauelshagen (mauelsha@ez-darmstadt.telekom.de)
Fri, 05 Feb 1999 17:28:27 MET

After thinking about refile_buffer and wakeup conditions for bdflush
for a while i came to the following solution.

It takes care of a fixed minimum of free buffers in the buffer cache.

I hard coded 5% just for testing purposes.

if (nr_buffers_type[BUF_DIRTY] > too_many ||
nr_buffers - nr_buffers_type[BUF_DIRTY] -
nr_buffers_type[BUF_LOCKED]
< nr_buffers_type[BUF_DIRTY] / 20)
wakeup_bdflush(1);

Performance is even better now while doing a
"dd if=/dev/zero of=/dev/v/l bs=1024k" on a 13G device i'm able to perform
a compile in 25s vs 1:20m!

Sure, if we have a reserved amount of buffers, it should be
dynamically dimensioned based on load scenarios.

This is just a quick hack to show the basic idea...

Opinions?

Regards,
Heinz

--


=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-


Systemmanagement C/S                             Deutsche Telekom AG
                                                 Entwicklungszentrum Darmstadt
Heinz Mauelshagen                                Otto-Roehm-Strasse 71c
Senior Systems Engineer                          Postfach 10 05 41
                                                 64205 Darmstadt
mge@ez-darmstadt.telekom.de                      Germany
                                                 +49 6151 886-425
                                                          FAX-386
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-


-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/