linux 2.4.8 vs 2.4.7 AHA2840A/42A "aic7xxx_abort return 8194"

[Mike Isely]

On Tue, 14 Aug 2001, Karl O . Pinc wrote:

> On 2001.08.14 17:07 Karl O . Pinc wrote:
> Hi, I'm a newbie.  Spank me.  :-)  Please let me know what further info is
> needed, etc.
>
> I've got a scsi board labeled AHA2840A/42A that AFIK worked fine under
> linux 2.0.34 for years.  My first question is this:
>
> Is the AHA284xA supported?
>
> The README.aic7xxx says that the AHA284x is supported, but makes no mention
> of AHA284xA.  AFIK, the difference is that the 284x has jumpers, the 284xA
> has dip switches.  The 284xA dip switches control: I/O port address, floppy
> enable/disable, BIOS address, and BIOS enable/disable.  I recall the
> jumpers on the 284x can at least also en/disable scsi target disconnect,
> sync negoiation and probably some other stuff.  I _think_ all this is now
> set in the card's bios cmos.
>
> Assuming the card's supported, here's what's happening:
>
> I'm re-incarnating the box and decided to load test it (copying large files
> from drive to drive. 3 drives, 4 processes, file sizes of {2,3,5,7} *
> 10^7B, 64MB ram).  Running linux-2.4.7, I get a rare (every few hours)
> "aic7xxx_abort return 8194", I've seen it while attempting to send ABORT to
> a drive.  (Sorry, I'm running 2.4.7 on top of the old 2.0.34 system, just
> to test, and syslogd seems to be down, so I don't have a written log to
> work from here.  Better info collected below.)
>
> I just switched to 2.4.8 and I get frequent (most of the time, there's no
> disk activity while the driver issues errors) "aic7xxx_abort return 8194"
> messages.  Mostly on ABORT but sometimes on TARGET_RESET.  There are so
> many the driver goes on to do "Recovery SCB completes" type messages.
>
> Where do I go from here?
>

   [...]

I have a system with an AHA2842; it uses a dipswitch.  The last time I
tried the aic7xxx driver with it, things behaved very badly (infinite
resets at boot).  I think this was around version 5.1.30 of the driver.
After that, Doug did a number of fixes and restablized it.  However a lot
has changed since then.  I hardly use the machine in question anymore so I
haven't updated it recently.  In addition, when I pointed this out on the
list back then, the silence from other users was, well, deafening.  I
think I found one other person who still had one of these adapters in use.
Given the apparently small number of people using this adapter I wouldn't
be at all surprised if things were broken again.

The 486 PC I have with this adapter is still working, stable, and running
Linux of course.  The old Micropolis 2217S hard drive there has trouble
spinning up after being off for more than a week or so ("stiction"
problem), but usually if I just slap the CPU case a bit it unsticks and
comes up :-)  (And to think I spent $1000 on that POS drive back in 1992.)
Perhaps I should try some tests on it and report what happens, as an extra
datapoint or two?

  -Mike


-- 
                        |         Mike Isely          |     PGP fingerprint
    POSITIVELY NO       |                             | 03 54 43 4D 75 E5 CC 92
 UNSOLICITED JUNK MAIL! |   isely @ pobox (dot) com   | 71 16 01 E2 B5 F5 C1 E8
                        |   (spam-foiling  address)   |


To Unsubscribe: send mail to majordomo at FreeBSD.org
with "unsubscribe aic7xxx" in the body of the message