Boszormenyi Zoltan (zboszor@mol.hu)
Mon, 21 Sep 1998 07:59:40 +0100
As I was titled as the best parport tester :) and Andrea
asked me to do ;) I tested the latest parport-arca patch (7th)
during the weekend. The testing procedure was as usual:
copying a large arj archive to the ZIP and/or the PHd and
printing a large postscript picture (tiges.ps from the ghostscript
examples). Again, all the combinations were tested, such as:
Single device tests:
- copying from the hd to ZIP, then reading back the file
- copying from the hd to PHd, then reading back the file
- printing
Two device tests:
- copying to the ZIP and the PHd at the same time
- reading back the files at the same time
- copying to the ZIP and reading back the file from the PHd
at the same time and vice-versa
- copying to the ZIP/PHd while printing, reading back the file
while printing
Three device tests:
while printing
- copying to the ZIP and the PHd at the same time,
- reading back the files at the same time
- copying to the ZIP and reading back the files from the PHd
at the same time and vice-versa
After copying the file the ZIP and the PHd were unmounted to
make sure that reading the file comes really from the device
and not from the cache.
The result in one sentence: No warnings, no data loss.
And finally some notes:
1. Now it is sure that neither the ZIP nor the PHd use IRQ.
Before loading the lp driver (and after the first ZIP and PHd tests)
I did a "cat /proc/interrupt" and that showed that no IRQ 7 happened.
2. As I said before the PHd is a very slow device and using it makes
my Linux box slow as well. Even changing VCs are slow.
I use the pd driver always with nice=1. I tried some cluster setting
as well such as the default (64), 32, 16, 1. My machine's responsiveness
weren't better. What can I do to achieve what I want? Maybe put some
schedule(); in the pd driver? Where? Grant, can you suggest some points?
:)
Zoltan Boszormenyi
-- To unsubscribe, send mail to: linux-parport-request@torque.net --
-- with the single word "unsubscribe" in the body of the message. --
This archive was generated by hypermail 2.0b3 on Wed 30 Dec 1998 - 10:18:20 EST