acenic breakage in 2.4.13-pre

acenic breakage in 2.4.13-pre

[Jan Rekorajski]

On Tue, 23 Oct 2001, Jeff Garzik wrote:

> If there are no complaints nor better suggestions, I would prefer to use
> the code in acenic.c / 8139cp.c as a base, since that code has been
> stable for a little while.

Speaking of acenic - it's broken in 2.4.13-pre. I have 3c985 and all I
get with 2.4.13-pre is "Firmware NOT running!". After I backed the
changes from -pre patch it started and works fine. Maybe the problem is
I have it in 32bit PCI slot?

Jan
-- 
Jan Rękorajski            |  ALL SUSPECTS ARE GUILTY. PERIOD!
baggins<at>mimuw.edu.pl   |  OTHERWISE THEY WOULDN'T BE SUSPECTS, WOULD THEY?
BOFH, MANIAC              |                   -- TROOPS by Kevin Rubio

Re: acenic breakage in 2.4.13-pre

[Jeff Garzik]

Jan Rekorajski wrote:
> 
> On Tue, 23 Oct 2001, Jeff Garzik wrote:
> 
> > If there are no complaints nor better suggestions, I would prefer to use
> > the code in acenic.c / 8139cp.c as a base, since that code has been
> > stable for a little while.
> 
> Speaking of acenic - it's broken in 2.4.13-pre. I have 3c985 and all I
> get with 2.4.13-pre is "Firmware NOT running!". After I backed the
> changes from -pre patch it started and works fine. Maybe the problem is
> I have it in 32bit PCI slot?

Several people have reported this bug.

Alexey, are the 2.4.13 acenic changes yours?  You had mentioned hacking
on it...  Jes, the maintainer, is CC'd too.

	Jeff


-- 
Jeff Garzik      | Only so many songs can be sung
Building 1024    | with two lips, two lungs, and one tongue.
MandrakeSoft     |         - nomeansno

Re: acenic breakage in 2.4.13-pre

[David S. Miller]

   From: Jeff Garzik <jgarzik@mandrakesoft.com>
   Date: Wed, 24 Oct 2001 10:56:44 -0400
   
   Several people have reported this bug.
   
   Alexey, are the 2.4.13 acenic changes yours?  You had mentioned hacking
   on it...  Jes, the maintainer, is CC'd too.

It's not Jes's changes, it's the pci64 bits I did.

Franks a lot,
David S. Miller
davem@redhat.com

Re: acenic breakage in 2.4.13-pre

[David S. Miller]

   From: Jan Rekorajski <baggins@sith.mimuw.edu.pl>
   Date: Wed, 24 Oct 2001 16:45:33 +0200
   
   Speaking of acenic - it's broken in 2.4.13-pre. I have 3c985 and all I
   get with 2.4.13-pre is "Firmware NOT running!". After I backed the
   changes from -pre patch it started and works fine. Maybe the problem is
   I have it in 32bit PCI slot?

Do you have CONFIG_HIGHMEM enabled?  If so, please try with
it turned off.

Franks a lot,
David S. Miller
davem@redhat.com

Re: acenic breakage in 2.4.13-pre

[Jan Rekorajski]

On Wed, 24 Oct 2001, David S. Miller wrote:

>    From: Jan Rekorajski <baggins@sith.mimuw.edu.pl>
>    Date: Wed, 24 Oct 2001 16:45:33 +0200
>    
>    Speaking of acenic - it's broken in 2.4.13-pre. I have 3c985 and all I
>    get with 2.4.13-pre is "Firmware NOT running!". After I backed the
>    changes from -pre patch it started and works fine. Maybe the problem is
>    I have it in 32bit PCI slot?
> 
> Do you have CONFIG_HIGHMEM enabled?  If so, please try with
> it turned off.

Nope. No HIGHMEM here.

Jan
-- 
Jan Rękorajski            |  ALL SUSPECTS ARE GUILTY. PERIOD!
baggins<at>mimuw.edu.pl   |  OTHERWISE THEY WOULDN'T BE SUSPECTS, WOULD THEY?
BOFH, MANIAC              |                   -- TROOPS by Kevin Rubio

Re: acenic breakage in 2.4.13-pre

[David S. Miller]

   From: Jan Rekorajski <baggins@sith.mimuw.edu.pl>
   Date: Wed, 24 Oct 2001 18:04:14 +0200

   On Wed, 24 Oct 2001, David S. Miller wrote:
   
   > Do you have CONFIG_HIGHMEM enabled?  If so, please try with
   > it turned off.
   
   Nope. No HIGHMEM here.

Thanks, one more question :-)  What compiler is on your
machine where this driver was built?  Are you using RH7.1
or some variant of gcc-3.x by chance?

Franks a lot,
David S. Miller
davem@redhat.com

Re: acenic breakage in 2.4.13-pre

[Jan Rekorajski]

On Wed, 24 Oct 2001, David S. Miller wrote:

>    From: Jan Rekorajski <baggins@sith.mimuw.edu.pl>
>    Date: Wed, 24 Oct 2001 18:04:14 +0200
> 
>    On Wed, 24 Oct 2001, David S. Miller wrote:
>    
>    > Do you have CONFIG_HIGHMEM enabled?  If so, please try with
>    > it turned off.
>    
>    Nope. No HIGHMEM here.
> 
> Thanks, one more question :-)  What compiler is on your
> machine where this driver was built?  Are you using RH7.1
> or some variant of gcc-3.x by chance?

No. It's plain old egcs 1.1.2 (gcc 2.91.66). I can try with gcc 2.95.x
but 2.96 or 3.x are no-no for me :)

Jan
-- 
Jan Rękorajski            |  ALL SUSPECTS ARE GUILTY. PERIOD!
baggins<at>mimuw.edu.pl   |  OTHERWISE THEY WOULDN'T BE SUSPECTS, WOULD THEY?
BOFH, MANIAC              |                   -- TROOPS by Kevin Rubio

Re: acenic breakage in 2.4.13-pre

[J Sloan]

Jan Rekorajski wrote:

> No. It's plain old egcs 1.1.2 (gcc 2.91.66). I can try with gcc 2.95.x
> but 2.96 or 3.x are no-no for me :)

That's bizzare, 2.96 has been working nicely
on all of my 30-40 Red Hat 7.x boxes. All sorts
of configs, from single P5-166 to Quad PPRO
with highmem, various vendors -

web, mail, dns, firewall, vpn, X workstations,
database, and all running kernels compiled
with gcc-2.96 , and everything running like
a top.

cu

jjs

[PATCH] Acenic fix (was Re: acenic breakage in 2.4.13-pre)

[David S. Miller]

   From: Jan Rekorajski <baggins@sith.mimuw.edu.pl>
   Date: Wed, 24 Oct 2001 18:04:14 +0200

   On Wed, 24 Oct 2001, David S. Miller wrote:
   
   >    From: Jan Rekorajski <baggins@sith.mimuw.edu.pl>
   >    Date: Wed, 24 Oct 2001 16:45:33 +0200
   >    
   >    Speaking of acenic - it's broken in 2.4.13-pre. I have 3c985 and all I
   >    get with 2.4.13-pre is "Firmware NOT running!". After I backed the
   >    changes from -pre patch it started and works fine. Maybe the problem is
   >    I have it in 32bit PCI slot?
   > 
   > Do you have CONFIG_HIGHMEM enabled?  If so, please try with
   > it turned off.
   
   Nope. No HIGHMEM here.

The patch below should cure the problem.

Linus, please apply.

diff -u --recursive --new-file --exclude=CVS --exclude=.cvsignore vanilla/linux/drivers/net/acenic.c linux/drivers/net/acenic.c
--- vanilla/linux/drivers/net/acenic.c	Fri Oct 12 15:35:53 2001
+++ linux/drivers/net/acenic.c	Wed Oct 24 08:32:43 2001
@@ -1051,7 +1051,8 @@
 	struct ace_private *ap;
 	struct ace_regs *regs;
 	struct ace_info *info = NULL;
-	unsigned long tmp_ptr, myjif;
+	u64 tmp_ptr;
+	unsigned long myjif;
 	u32 tig_ver, mac1, mac2, tmp, pci_state;
 	int board_idx, ecode = 0;
 	short i;

Re: acenic breakage in 2.4.13-pre

[David S. Miller]

   From: Jan Rekorajski <baggins@sith.mimuw.edu.pl>
   Date: Wed, 24 Oct 2001 16:45:33 +0200

Czesc Jan,

   Speaking of acenic - it's broken in 2.4.13-pre.

As a side note, it's really unfortunate that such reports surface
several hours after Linus releases the real 2.4.13 :(

Franks a lot,
David S. Miller
davem@redhat.com