Bug in 2.0.35 AIC7xxx driver?

Bug in 2.0.35 AIC7xxx driver?

Post by John Steve » Sun, 26 Jul 1998 04:00:00



I just attempted to recompile and install kernel version 2.0.35.

On rebooting, the driver falls into a loop that continually resets
the SCSI bus.

Anybody know what has changed in this part of the kernel since
2.0.32?  (Which was the last "stable" kernel I got to work. . .).

Thanks,
John S.

 
 
 

Bug in 2.0.35 AIC7xxx driver?

Post by Friedemann Baitinge » Mon, 27 Jul 1998 04:00:00



> I just attempted to recompile and install kernel version 2.0.35.

> On rebooting, the driver falls into a loop that continually resets
> the SCSI bus.

> Anybody know what has changed in this part of the kernel since
> 2.0.32?  (Which was the last "stable" kernel I got to work. . .).

2.0.33 works for me. 2.0.34 and 2.0.35 are broken in respect to the
AIC7xxx driver on my machine (Dual PPro-233, SMP mode)

--


 
 
 

Bug in 2.0.35 AIC7xxx driver?

Post by Paul Mart » Mon, 27 Jul 1998 04:00:00



>I just attempted to recompile and install kernel version 2.0.35.

>On rebooting, the driver falls into a loop that continually resets
>the SCSI bus.

>Anybody know what has changed in this part of the kernel since
>2.0.32?  (Which was the last "stable" kernel I got to work. . .).

It might be a problem with the card itself.

If /proc/pci tells you that you have anything other than a (revision 0)
card you will probably have problems. I don't know why they're so
troublesome, but all the (revision 1) cards we've had have given us no end
of problems (corruption, crashes, bus resets). Swapping them for rev 0's
has stopped the problem completely.

Hopefully this will get fixed reasonably soon as the driver developer now
has information directly from Adaptec.

--

at home, swap dash to dot to email.        

 
 
 

Bug in 2.0.35 AIC7xxx driver?

Post by Chris Sai » Mon, 27 Jul 1998 04:00:00



> 2.0.33 works for me. 2.0.34 and 2.0.35 are broken in respect to the
> AIC7xxx driver on my machine (Dual PPro-233, SMP mode)

Interesting.  There seems to be a wide variance in works and doesn't work
for the AIC7XXX crowd.  My particular on-board AIC7880 (for which I don't
have the rev. number handy) has had little to no problems with any of the
Adaptec code.

My understanding is that now Adaptec is going to help out Linux rather
than hinder its driver development.  Who knows?  Maybe we'll be able to
start recommending people buy Adaptec cards again.  :)

=C=

--

PGP Public Key is available at http://www.wtower.com/~csaia/personal/pubkey.asc
| A!JW223 YK+++!/+++^/+++"/+++'/ir WK++ DT+ PI+ BR++^ MM++  PFeather XSire P+ |
|    I++++ Dswiad $+++dap{c} Vrm{jt} T212/Pinky E77 EeH5 EeeH2 H17 Ay75 M     |

 
 
 

Bug in 2.0.35 AIC7xxx driver?

Post by Friedemann Baitinge » Mon, 27 Jul 1998 04:00:00




> >I just attempted to recompile and install kernel version 2.0.35.

> >On rebooting, the driver falls into a loop that continually resets
> >the SCSI bus.

> >Anybody know what has changed in this part of the kernel since
> >2.0.32?  (Which was the last "stable" kernel I got to work. . .).

> It might be a problem with the card itself.

> If /proc/pci tells you that you have anything other than a (revision 0)
> card you will probably have problems. I don't know why they're so
> troublesome, but all the (revision 1) cards we've had have given us no end
> of problems (corruption, crashes, bus resets). Swapping them for rev 0's
> has stopped the problem completely.

I have a revision 0 card and still have the problems:

  Bus  0, device  11, function  0:
    SCSI storage controller: Adaptec AIC-7881U (rev 0).

It is a Ultra Wide card. The machine is a Dual PPro-233MHz running in
SMP mode. I have tried to apply the patch:

    aic7xxx-5.1.0-pre3-2.0.35

but this doesn't fix it either.

--

 
 
 

Bug in 2.0.35 AIC7xxx driver?

Post by Paul Mart » Mon, 27 Jul 1998 04:00:00



>I have a revision 0 card and still have the problems:

>  Bus  0, device  11, function  0:
>    SCSI storage controller: Adaptec AIC-7881U (rev 0).

>It is a Ultra Wide card. The machine is a Dual PPro-233MHz running in
>SMP mode. I have tried to apply the patch:

Interesting. What's the BIOS revision on this card? Is it 1.23, 1.25 or
1.32?

Our news server has two AIC-7881U (rev 0, BIOS 1.23) on a uniprocessor
PPro 200MHz.

When it had two (rev 1, BIOS 1.32) it fell over at least once a week due
to SCSI errors.

--

at home, swap dash to dot to email.        

 
 
 

Bug in 2.0.35 AIC7xxx driver?

Post by Steven A. Falc » Tue, 28 Jul 1998 04:00:00


> I have a revision 0 card and still have the problems:

>   Bus  0, device  11, function  0:
>     SCSI storage controller: Adaptec AIC-7881U (rev 0).

> It is a Ultra Wide card. The machine is a Dual PPro-233MHz running in
> SMP mode. I have tried to apply the patch:

>     aic7xxx-5.1.0-pre3-2.0.35

> but this doesn't fix it either.

> --


I had problems with my AHA-2940UW card when I upgraded to Linux 2.0.34
and 2.0.35.

I just patched in the 5.1.0-pre5 version of the Adaptec driver, and my
card now works fine again.  A pre6 driver (to clean up some remaining
problems) is in the works, I believe.

Try ftp://ftp.dialnet.net/pub/linux/aic7xxx/ for the latest patches.

        Steve Falco

 
 
 

Bug in 2.0.35 AIC7xxx driver?

Post by Steven A. Falc » Tue, 28 Jul 1998 04:00:00


> I have a revision 0 card and still have the problems:

>   Bus  0, device  11, function  0:
>     SCSI storage controller: Adaptec AIC-7881U (rev 0).

> It is a Ultra Wide card. The machine is a Dual PPro-233MHz running in
> SMP mode. I have tried to apply the patch:

>     aic7xxx-5.1.0-pre3-2.0.35

> but this doesn't fix it either.

> --


I had problems with my AHA-2940UW card when I upgraded to Linux 2.0.34
and 2.0.35.

I just patched in the 5.1.0-pre5 version of the Adaptec driver, and my
card now works fine again.  A pre6 driver (to clean up some remaining
problems) is in the works, I believe.

Try ftp://ftp.dialnet.net/pub/linux/aic7xxx/ for the latest patches.

        Steve Falco

 
 
 

Bug in 2.0.35 AIC7xxx driver?

Post by Friedemann Baitinge » Tue, 28 Jul 1998 04:00:00



> Just an unposted hint, I have been VERY happy with 2.1.101!

What type of Adaptec do you have? Are you running an SMP kernel?
If I'd switch from a 2.0.x kernel to a 2.1.1xx kernel, what things other
than the kernel would I have to change? (I heard that the layout of the
/proc filesystem has changed such that tools like xsysinfo and others
which parse the /proc data do not work anymore)

--

 
 
 

Bug in 2.0.35 AIC7xxx driver?

Post by Friedemann Baitinge » Tue, 28 Jul 1998 04:00:00



> I had problems with my AHA-2940UW card when I upgraded to Linux 2.0.34
> and 2.0.35.

> I just patched in the 5.1.0-pre5 version of the Adaptec driver, and my
> card now works fine again.  A pre6 driver (to clean up some remaining
> problems) is in the works, I believe.

> Try ftp://ftp.dialnet.net/pub/linux/aic7xxx/ for the latest patches.

>         Steve Falco

just installed the 5.1.0-pre5 as you suggested and it indeed works!

--

 
 
 

Bug in 2.0.35 AIC7xxx driver?

Post by John Steve » Wed, 29 Jul 1998 04:00:00




>>I just attempted to recompile and install kernel version 2.0.35.

>>On rebooting, the driver falls into a loop that continually resets
>>the SCSI bus.

>>Anybody know what has changed in this part of the kernel since
>>2.0.32?  (Which was the last "stable" kernel I got to work. . .).

>It might be a problem with the card itself.

>If /proc/pci tells you that you have anything other than a (revision 0)
>card you will probably have problems.

It's an AHA-2940UW, based on chipset AIC-7881U, rev. 0.

I have a wide SCSI drive at 0 on the 16 bit bus, an HP CD-ROM drive
at id 2 on the 8 bit bus, and finally a Jaz drive on id 4, also on
the 8 bit bus.

The driver identifies and accepts the HDD (the only wide device), but
gives a variety of failures when attempting to identify/start up the
devices on the 8 bit bus.

Quote:>Hopefully this will get fixed reasonably soon as the driver developer now
>has information directly from Adaptec.

That would be nice. . . I'd like to know *why* the AIC-7xxx driver
was changed in the "stable" kernel series. . . isn't that contrary
to the express purpose of the stable series?  The AIC-7xxx driver
was working just fine for me in 2.0.32, why did it need to be
modified?  Anybody know?

John S.

 
 
 

Bug in 2.0.35 AIC7xxx driver?

Post by Doug Ledfor » Wed, 29 Jul 1998 04:00:00



> That would be nice. . . I'd like to know *why* the AIC-7xxx driver
> was changed in the "stable" kernel series. . . isn't that contrary
> to the express purpose of the stable series?  The AIC-7xxx driver
> was working just fine for me in 2.0.32, why did it need to be
> modified?  Anybody know?

For the same reason I'm currently working on the 5.1.0 driver version.  It
had nothing to do with the old driver not working on the cards it was
intended to work on, it had to do with newer cards on the market.
Specifically, the 5.0.x driver versions, as are in 2.0.34 and 2.0.35 include
support for the aic7895 chipset as used on many motherboards and as the
newer 3940AUW cards use.  The 5.1.0 driver is adding support for the
7890/1/6/7 chipsets as well (all of the Ultra2 chipsets).

--
--------------------------------------

  Opinions expressed are my own, but
     they should be everybody's.
--------------------------------------

 
 
 

Bug in 2.0.35 AIC7xxx driver?

Post by spaem lur » Wed, 05 Aug 1998 04:00:00




> > That would be nice. . . I'd like to know *why* the AIC-7xxx driver
> > was changed in the "stable" kernel series. . . isn't that contrary
> > to the express purpose of the stable series?  The AIC-7xxx driver
> > was working just fine for me in 2.0.32, why did it need to be
> > modified?  Anybody know?

> For the same reason I'm currently working on the 5.1.0 driver version.  It
> had nothing to do with the old driver not working on the cards it was
> intended to work on, it had to do with newer cards on the market.
> Specifically, the 5.0.x driver versions, as are in 2.0.34 and 2.0.35 include
> support for the aic7895 chipset as used on many motherboards and as the
> newer 3940AUW cards use.  The 5.1.0 driver is adding support for the
> 7890/1/6/7 chipsets as well (all of the Ultra2 chipsets).

> --
> --------------------------------------

>   Opinions expressed are my own, but
>      they should be everybody's.
> --------------------------------------

I do agree with John, WHY change it in the STABLE series, you don't want toknow
the harm it already has done to my own system. I test the latest greatest
(in this case stable!!) kernel on my own system (and leave our servers alone
until
I am sure it keeps working), both 2.0.34 and 2.0.35 will hardlock my system
under
any heavy disk usage, so for me 2.0.34/35 is NOT stable and can't be
implemented
on our production systems. A * little detail: Red Hat  has gone to
2.0.34/35 for
all releases now, so in effect I am stuck at the current 2.0.33 now if I want
to keep
the production systems RH pure :-((

Will 2.0.36 be a STABLE kernel WRT aic7xxx?

Ron Offerman
IT Staff

 
 
 

Bug in 2.0.35 AIC7xxx driver?

Post by Tristan Wibberle » Sat, 08 Aug 1998 04:00:00



> I test the latest greatest
> (in this case stable!!) kernel on my own system (and leave our servers alone
> until
> I am sure it keeps working), both 2.0.34 and 2.0.35 will hardlock my system
> under
> any heavy disk usage,

What compiler are you using? Make sure you are using gcc 2.7.2.3, not
egcs or gcc 2.8.x with the ioperm.c fix - I heard of this being the
cause of locked systems under heavy disk usage for many people.

--
Tristan

 
 
 

Bug in 2.0.35 AIC7xxx driver?

Post by spaem lur » Sun, 09 Aug 1998 04:00:00




> > I test the latest greatest
> > (in this case stable!!) kernel on my own system (and leave our servers alone
> > until
> > I am sure it keeps working), both 2.0.34 and 2.0.35 will hardlock my system
> > under
> > any heavy disk usage,

> What compiler are you using? Make sure you are using gcc 2.7.2.3, not
> egcs or gcc 2.8.x with the ioperm.c fix - I heard of this being the
> cause of locked systems under heavy disk usage for many people.

> --
> Tristan

Both gcc 2.7.2.3 and egcs 1.03a compiled kernels fail hardlocking the system.
Red Hat has updated both 5.0 and 5.1 to 2.0.35, older rpms are no longer
available. The 2.0.35 rpms supplied by RH also fail and as I previously stated, I
would like to keep the production servers as pure as possible.

My biggest 'problem' is however that a kernel in the STABLE series shouldn't
break a system that has been running stable kernels. My bleeding edge
workstation/server I can fix myself and I don't mind living with the 2.0.33
kernekl; the production servers do need a couple of the 2.0.34/35 enhancements
(mainly DoS attack stuff)

BTW, this has happened before with the Adaptec driver somewhere in the 1.2.x
kernels when the driver was switched to the more generic aic7xxx.

Ron Offerman
IT Staff

 
 
 

1. AIC7xxx and 2.0.35: SCSI (version 1.x) fails!

Hello!
I am runnig the 2.0.35 Kernel and it is fine with my AHA2940 at HDD and
CD-ROM!
Just by booting with my SCSI-version 1 Scanner (Mustek 800 SP) the
driver hangs detecting the scanner-device.
With 2.0.34 the scanner is detected well!
Is this a bug or are there some conflicts which appear just with the new
driver?

Thanks!
Tim

Please reply by email, too!
--
-----------------------------------  Tim Schulz
----------------------------------  Gifhorn,Germany
---------------------------------  http://www.wolfsburg.de/~schulz/Tim/

2. network setup with multiple ifaces

3. aic7xxx support on patches 2.0.35

4. Fun with openldap 2.2.25 in a SS20/Sol9

5. Linux 2.0.35 and AIC7XXX

6. CMOS utilities?

7. Partition problems - 2.0.35 + aic7xxx + 8+GByte SCSI disk

8. PLEASE Can Any body HAVE SOLUTION using shell Script and awk???????Please

9. Kernel 2.0.35 memory bug ??

10. SLIP: bug in kernel 2.0.35 ?

11. Weird bug in MS-DOS filesystem handling in 2.0.35

12. Multiple ethernet inerfaces BUG in 2.0.35 ?

13. Possible Bug in aic7xxx driver / Ultra2Disk ?