Date: Sun, 13 Jan 2002 23:08:27 -0500 (EST)
From: Ian Morgan <imorgan@webcon.net>
To: andre@linuxdiskcert.org
cc: Linux Kernel Mailing List <linux-kernel@vger.kernel.org>
Subject: ide.2.4.16.12102001 chokes on HPT366
Message-ID: <Pine.LNX.4.40.0201132231560.1753-100000@light.webcon.net>
Organization: "Webcon, Inc."
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Sender: linux-kernel-owner@vger.kernel.org

Andre (and the rest of the crowd),

I really hate to be one to provide some negative news of your IDE patch,
but here goes:

I have been hearing very good things about your patch lately, so after
having not tried it since the 2.4.1[0-3] days, I figured it was time for
another round of testing. Specifically, some people, including Alan,
mentioned the stability of the Highpoint drivers. Since I've had dismal luck
with my HPT366 in the past, I was excited at the prospect of being able to
use it without trashing my disks (again).

No such luck, I'm afraid. While my disks have survived seemingly unharmed,
the system [1] barfed on speed tests.

I booted into multi-user (no X), but dhcpd hadn't started properly. After
some investigation, I found it to be core-dumping as a result of a corrupted
binary. Re-installing the RPM got it back up in a jiffy. While this could
have been an unrelated anomaly, my suspicions were on the Highpoint driver.

Next, I set the drive for 32-bit IO...

	hdparm -c1 /dev/hde

... then switched to X (exec init 5).

The first real test, now, was to speed-test the drive:

	hdparm -tT /dev/hde

This yielded practically no improvement over the PIIX4 (both averaged
~25 MiB/s). Of course the usual routine is to run the test 3 or 4 times and
take an average. Well, lo and behold, the 2nd run of the speed test went as
follows:

	/dev/hda:
	Timing buffer-cache reads:   128 MB in  ~1.4 seconds = ~90 MB/s
	Timing buffered disk reads:

... and nothing! The hdparm process went into D state, and everything else
that tried to access the drive also locked.

I rebooted, into single-user mode this time, and tried the speed test again
(after enabling 32-bit IO). This time it locked on the first test in the
same way as above. At this point I gave up and returned the drive to the
PIIX4 and prayed it was not completely fubared.

So, any ideas? Is you patch really supposed to work well on the HPT366, or
am I dreaming in Technicolour? Interestingly, most everything on the system
seemed to work OK. File access, loading apps, etc all seemed OK, but the
lockup during speed test really worries me.


[1] Abit BP6, 2x 400MHz Celeron, 384MB RAM
    HPT 366 w/ Quantum Fireball LM30 (PM by CS)
    PIIX4 w/ WDC AC33100H (SM), Mitsumi FX4820T (SS)
    Linux 2.4.17 w/ ide.2.4.16.12102001 & freeswan-1.91
    hdparm v4.4

Regards,
Ian Morgan
-- 
-------------------------------------------------------------------
 Ian E. Morgan        Vice President & C.O.O.         Webcon, Inc.
 imorgan@webcon.net         PGP: #2DA40D07          www.webcon.net
-------------------------------------------------------------------

-
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/