Subject: Re: IBM drives vs Adaptec controllers/driver
To: NetBSD-current Discussion List <>
From: Greg A. Woods <>
List: current-users
Date: 01/23/1999 16:51:28
(BTW, apologies for not noticing the "ide" in the subject line of the
original thread....)

[ On , January 23, 1999 at 13:28:26 (+0100), Sean Doran wrote: ]
> Subject: Re: IBM drives vs Adaptec controllers/driver [was Re: 10G IBM ide hard disk]
> (Greg A. Woods) writes:
> > I enabled both AHC_TAGENABLE and AHC_SCBPAGING_ENABLE with much more
> > appreciable results.
> AHC_TAGENABLE where it works is great.
> I have yet to have my controller not puke when AHC_SCBPAGING_ENABLE is on.

I had one "burp" from something while running bonnie simultaneously on
each of four partitions on two spindles:

A Bonnie process said:  "drastic I/O error (getc(3)): Input/output
error", and the system logged:

Jan 23 16:23:26 admin /netbsd: sd1(ahc0:1:0): timed out in dataout phase, SCSISIGI == 0x18
Jan 23 16:23:27 admin /netbsd: Ordered Tag queued
Jan 23 16:23:27 admin /netbsd: Ordered Tag sent
Jan 23 16:23:27 admin /netbsd: Ordered Tag sent

I suspect I'll be OK with SCBPAGING, though I will keep a close eye out
for future errors such as this.  I don't expect the system to ever
approach the I/O load it was under during this test (a steady 4-9MB/s on
both drives for over 35 minutes).

One pleasant surprise was that my idle process (rc5des ;-) was still
getting lots of CPU after the initial "writing" stage.  I had been
worried that this older AIC driver would require a lot of CPU cycles,
but it seems to be relatively low overhead.  I was kinda surprised that
the load average went so high even though these processes should have
been waiting for I/O most of the time -- or maybe the load average does
weigh in such processes, not just those waiting for CPU as I'd always
thought it did.

> > Maybe there's some difference in the aic7880 firmware on your adaptec
> > card that's not in my on-board version....
> Maybe.  A colleague of mine was having pretty much the
> same problem (10MHz instead of 20) with his on-motherboard
> controller, whereas my PCI card controller negotiates
> 20MHz just fine.

Hmm...  That's good to know.  I wonder if your colleague's system will
negotiate 20MHz or more with FreeBSD-3.0....

							Greg A. Woods

+1 416 218-0098      VE3TCP      <>      <robohack!woods>
Planix, Inc. <>; Secrets of the Weird <>