[PATCH 2.6.10-rc3][PPC32] Fix Motorola PReP (PowerstackII Utah) PCI IRQ map

The PCI IRQ map for the old Motorola PowerStackII (Utah) boards was
incorrect, but this breakage wasn't exposed until 2.5, and finally fixed
until recently by Sebastian Heutling [off-list ref].

Signed-off-by: Christian Kujau <redacted>
Signed-off-by: Tom Rini <redacted>

--- 1.32/arch/ppc/platforms/prep_pci.c	2004-10-12 14:29:11 -07:00
+++ edited/arch/ppc/platforms/prep_pci.c	2004-12-06 09:20:52 -07:00

@@ -49,10 +49,10 @@
         0,   /* Slot 1  - unused */
         5,   /* Slot 2  - SCSI - NCR825A  */
         0,   /* Slot 3  - unused */
-        1,   /* Slot 4  - Ethernet - DEC2114x */
+        3,   /* Slot 4  - Ethernet - DEC2114x */
         0,   /* Slot 5  - unused */
-        3,   /* Slot 6  - PCI Card slot #1 */
-        4,   /* Slot 7  - PCI Card slot #2 */
+        2,   /* Slot 6  - PCI Card slot #1 */
+        3,   /* Slot 7  - PCI Card slot #2 */
         5,   /* Slot 8  - PCI Card slot #3 */
         5,   /* Slot 9  - PCI Bridge */
              /* added here in case we ever support PCI bridges */

-- 
Tom Rini
http://gate.crashing.org/~trini/

On Tue, 7 Dec 2004, Linus Torvalds wrote:

Tssk. Total whitespace corruption. Please fix your mailer.

Heh. I take that back. The whitespace "corruption" was in the file itself. 
Seems to have spaces instead of tabs there..

		Linus

On Mon, 6 Dec 2004, Tom Rini wrote:

The PCI IRQ map for the old Motorola PowerStackII (Utah) boards was
incorrect, but this breakage wasn't exposed until 2.5, and finally fixed
until recently by Sebastian Heutling [off-list ref].

Tssk. Total whitespace corruption. Please fix your mailer.

		Linus

The PCI IRQ map for the old Motorola PowerStackII (Utah) boards was
incorrect, but this breakage wasn't exposed until 2.5, and finally fixed
until recently by Sebastian Heutling [off-list ref].

Yesterday I finally got around to testing it. It seems the patch has 
been applied in Linus's tree so I downloaded the latest BK and tried it.

Still does not work for me but this time it's different. Before the 
patch SCSI worked fine but PCI NICs caused hangs. Now I can't test PCI 
NICs because even the onboard 53c825 SCSI hangs - seems it gets no 
interrupts.

It detects the HBA, tries device discovery, gets a timeout, ABORT, 
timeout, TARGET RESET, timeout, BUS RESET, timeout, HOST RESET and there 
it hangs.

Does it work for anyone else on Powerstack II Pro4000 (Utah)?

-- 
Meelis Roos (mroos@linux.ee)

Meelis Roos wrote:

quoted

The PCI IRQ map for the old Motorola PowerStackII (Utah) boards was
incorrect, but this breakage wasn't exposed until 2.5, and finally fixed
until recently by Sebastian Heutling [off-list ref].

Yesterday I finally got around to testing it. It seems the patch has 
been applied in Linus's tree so I downloaded the latest BK and tried it.

Still does not work for me but this time it's different. Before the 
patch SCSI worked fine but PCI NICs caused hangs. Now I can't test PCI 
NICs because even the onboard 53c825 SCSI hangs - seems it gets no 
interrupts.

It detects the HBA, tries device discovery, gets a timeout, ABORT, 
timeout, TARGET RESET, timeout, BUS RESET, timeout, HOST RESET and 
there it hangs.

Does it work for anyone else on Powerstack II Pro4000 (Utah)?

It does work in 2.6.8 using backported patches (e.g. the debian 2.6.8 
kernel). But it doesn't work above that version because of other patches 
in arch/ppc/platforms/prep_pci.c and arch/ppc/platforms/prep_setup.c 
(made by Tom Rini?). I couldn't find out what exactly is causing this 
problem yet (because lack of time and the fact that my Powerstack is 
used as a router).

Basti

Sebastian Heutling said:

Meelis Roos wrote:

quoted

quoted

The PCI IRQ map for the old Motorola PowerStackII (Utah) boards was
incorrect, but this breakage wasn't exposed until 2.5, and finally
fixed
until recently by Sebastian Heutling [off-list ref].

Yesterday I finally got around to testing it. It seems the patch has
been applied in Linus's tree so I downloaded the latest BK and tried it.

Still does not work for me but this time it's different. Before the
patch SCSI worked fine but PCI NICs caused hangs. Now I can't test PCI
NICs because even the onboard 53c825 SCSI hangs - seems it gets no
interrupts.

It detects the HBA, tries device discovery, gets a timeout, ABORT,
timeout, TARGET RESET, timeout, BUS RESET, timeout, HOST RESET and
there it hangs.

Does it work for anyone else on Powerstack II Pro4000 (Utah)?

It does work in 2.6.8 using backported patches (e.g. the debian 2.6.8
kernel). But it doesn't work above that version because of other patches
in arch/ppc/platforms/prep_pci.c and arch/ppc/platforms/prep_setup.c
(made by Tom Rini?). I couldn't find out what exactly is causing this
problem yet (because lack of time and the fact that my Powerstack is
used as a router).

Ah, this could well be my fault.  Those patches were to improve support
of IBM RS/6000 PReP boxes.  Do those machines have residual data?  If
so, could anyone who has one send me the contents of /proc/residual?

Also, a full boot log when working and failing would be cool.

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Leigh Brown wrote:

quoted

quoted

It detects the HBA, tries device discovery, gets a timeout, ABORT,
timeout, TARGET RESET, timeout, BUS RESET, timeout, HOST RESET and
there it hangs.

it does not really hang, it just tries to initialize every target of the
HBA (here: from sym0:0:0: to sym0:15:0, see [1] for more info) and it is
so busy with it that the bootprocess seems to hang. after failing with the
last target, booting continues just fine. (i have no disks attached,
booting via nfsroot)

Ah, this could well be my fault.  Those patches were to improve support
of IBM RS/6000 PReP boxes.  Do those machines have residual data?  If
so, could anyone who has one send me the contents of /proc/residual?

Also, a full boot log when working and failing would be cool.

[1] it's all here: http://nerdbynature.de/bits/hal/2.6.11-rc3/

(yes, they are from different dates, but the setup is the same. the
kernelversion from messages is 2.6.11-rc2, the rest is all 2.6.11-rc3,
from vanilla (-BK) sources)

thanks,
Christian.
- --
BOFH excuse #67:

descramble code needed from software company
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.0 (GNU/Linux)
Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org

iD8DBQFCG11y+A7rjkF8z0wRAvMoAKCWliE97XWNmFv+xf7d3yU5vN3tDQCffMCj
Y8hf0xXrOsCA6WkZUPKkUa0=
=ECSk
-----END PGP SIGNATURE-----

Leigh Brown wrote:

Sebastian Heutling said:
 

quoted

Meelis Roos wrote:

   

quoted

quoted

The PCI IRQ map for the old Motorola PowerStackII (Utah) boards was
incorrect, but this breakage wasn't exposed until 2.5, and finally
fixed
until recently by Sebastian Heutling [off-list ref].
       

Yesterday I finally got around to testing it. It seems the patch has
been applied in Linus's tree so I downloaded the latest BK and tried it.

Still does not work for me but this time it's different. Before the
patch SCSI worked fine but PCI NICs caused hangs. Now I can't test PCI
NICs because even the onboard 53c825 SCSI hangs - seems it gets no
interrupts.

It detects the HBA, tries device discovery, gets a timeout, ABORT,
timeout, TARGET RESET, timeout, BUS RESET, timeout, HOST RESET and
there it hangs.

Does it work for anyone else on Powerstack II Pro4000 (Utah)?

     

It does work in 2.6.8 using backported patches (e.g. the debian 2.6.8
kernel). But it doesn't work above that version because of other patches
in arch/ppc/platforms/prep_pci.c and arch/ppc/platforms/prep_setup.c
(made by Tom Rini?). I couldn't find out what exactly is causing this
problem yet (because lack of time and the fact that my Powerstack is
used as a router).
   

Ah, this could well be my fault.  Those patches were to improve support
of IBM RS/6000 PReP boxes.  Do those machines have residual data?  If
so, could anyone who has one send me the contents of /proc/residual?

Also, a full boot log when working and failing would be cool.
 

No, the PowerstackII Pro4000 (Utah) has no residual. I couldn't see 
anything unusal in the bootlogs except that
neither IDE nor SCSI interrupts occour (timeouts for SCSI and lost 
interrupts for IDE). I assume other PCI-devices have a similar problem. 
I will try to extract some bootlogs in the next days but I think they 
won't help.

Basti

On Tue, Feb 22, 2005 at 10:36:36AM +0200, Meelis Roos wrote:

quoted

The PCI IRQ map for the old Motorola PowerStackII (Utah) boards was
incorrect, but this breakage wasn't exposed until 2.5, and finally fixed
until recently by Sebastian Heutling [off-list ref].

Yesterday I finally got around to testing it. It seems the patch has 
been applied in Linus's tree so I downloaded the latest BK and tried it.

Still does not work for me but this time it's different. Before the 
patch SCSI worked fine but PCI NICs caused hangs. Now I can't test PCI 
NICs because even the onboard 53c825 SCSI hangs - seems it gets no 
interrupts.

It detects the HBA, tries device discovery, gets a timeout, ABORT, 
timeout, TARGET RESET, timeout, BUS RESET, timeout, HOST RESET and there 
it hangs.

Does it work for anyone else on Powerstack II Pro4000 (Utah)?

Can you try : 

  http://people.debian.org/~luther/d-i/images/daily/powerpc/netboot/vmlinuz-prep.initrd

It works for me, and the kernel (2.6.8) has the irqs patched, but not the scsi
stuff touched, i think.

Friendly,

Sven Luther

Can you try :

 http://people.debian.org/~luther/d-i/images/daily/powerpc/netboot/vmlinuz-prep.initrd

Unfortunately there are only floppy and floppy-2.4 dirs under powerpc.

-- 
Meelis Roos (mroos@linux.ee)

On Thu, Feb 24, 2005 at 05:47:15PM +0200, Meelis Roos wrote:

quoted

Can you try :

http://people.debian.org/~luther/d-i/images/daily/powerpc/netboot/vmlinuz-prep.initrd

Unfortunately there are only floppy and floppy-2.4 dirs under powerpc.

Oh, damn, need to fix my daily builder, should be ok for tomorrow. IN the
meanwhile, you can try : 

  http://people.debian.org/~luther/d-i/images/2005-02-23/powerpc/netboot/vmlinuz-prep.initrd

This is a zImage.prep kernel with builtin initrd, you just put it somewhere
where you can boot it from, usually a tftp server.

Friendly,

Sven Luther

Oh, damn, need to fix my daily builder, should be ok for tomorrow. IN the
meanwhile, you can try :

 http://people.debian.org/~luther/d-i/images/2005-02-23/powerpc/netboot/vmlinuz-prep.initrd

This seems to work fine: onboard scsi is OK, pci nic with de4x5 is OK 
too. Haven't got more PCI cards in there currently.

-- 
Meelis Roos (mroos@linux.ee)

On Thu, Feb 24, 2005 at 06:34:03PM +0200, Meelis Roos wrote:

quoted

Oh, damn, need to fix my daily builder, should be ok for tomorrow. IN the
meanwhile, you can try :

http://people.debian.org/~luther/d-i/images/2005-02-23/powerpc/netboot/vmlinuz-prep.initrd

This seems to work fine: onboard scsi is OK, pci nic with de4x5 is OK 
too. Haven't got more PCI cards in there currently.

Ok, can you now install the 2.6.10-3 package from : 

  http://people.debian.org/~luther/powerpc/2.6.10-3 

(you need kernel-image-2.6.10-powerpc and mkvmlinuz), call mkvmlinuz on it (or
add it to kernel-img.conf as postinst_hook=/usr/sbin/mkvmlinuz).

You may probably want also to modify /etc/mkinitrd/mkinitrd:MODULES_DEP to dep
instead of MOST, or you may hit size problems with your initrd, i would be
interested in knowing that.

Then just dd your /boot/vmlinuz-2.6.10-powerpc to your prep partition, or
better yet to a tftp server, and try it out. If the scsi problems are there,
can you fill a bug report against kernel-source-2.6.10 ? 

Friendly,

Sven Luther

Then just dd your /boot/vmlinuz-2.6.10-powerpc to your prep partition, or
better yet to a tftp server, and try it out. If the scsi problems are there,
can you fill a bug report against kernel-source-2.6.10 ?

Thanks for the new kernel. Just filed a bug report on 
kernel-source-2.6.10.

You may probably want also to modify /etc/mkinitrd/mkinitrd:MODULES_DEP to dep
instead of MOST, or you may hit size problems with your initrd, i would be
interested in knowing that.

It worked without changing the module list, with 5.2M resulting vmlinuz.

-- 
Meelis Roos (mroos@linux.ee)

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Sven Luther wrote:

Oh, damn, need to fix my daily builder, should be ok for tomorrow. IN the
meanwhile, you can try : 

  http://people.debian.org/~luther/d-i/images/2005-02-23/powerpc/netboot/vmlinuz-prep.initrd

oh, what fun - it's booting! de4x5 is loading (although i build my kernels
with a (compiled in) tulip driver). sym53c8xx gets loaded and initializing
the scsi bus is *not* blocking all other activities as usual.

here are the logs:

http://nerdbynature.de/bits/hal/d-i-2005.02.23/  (working 2.6.8 from Sven)
http://nerdbynature.de/bits/hal/2.6.11-rc3/      (scsi errors)

(note: i still have no disks attached)

thank you,
Christian.
- --
BOFH excuse #286:

Telecommunications is downgrading.
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.0 (GNU/Linux)
Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org

iD8DBQFCHnAz+A7rjkF8z0wRAsxtAJ9DPKaUg9nCWGZKGxqd6sCOtVqu1QCfRyjp
934bVyFtXvuTFYLCgQcAhrI=
=hSxn
-----END PGP SIGNATURE-----

On Fri, Feb 25, 2005 at 01:24:19AM +0100, Christian Kujau wrote:

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Sven Luther wrote:

quoted

Oh, damn, need to fix my daily builder, should be ok for tomorrow. IN the
meanwhile, you can try : 

  http://people.debian.org/~luther/d-i/images/2005-02-23/powerpc/netboot/vmlinuz-prep.initrd

oh, what fun - it's booting! de4x5 is loading (although i build my kernels
with a (compiled in) tulip driver). sym53c8xx gets loaded and initializing
the scsi bus is *not* blocking all other activities as usual.

here are the logs:

http://nerdbynature.de/bits/hal/d-i-2005.02.23/  (working 2.6.8 from Sven)
http://nerdbynature.de/bits/hal/2.6.11-rc3/      (scsi errors)

(note: i still have no disks attached)

So, now, we need to find out what the problems where, i think it is something
that went in between 2.6.8 and 2.6.10, and leigh said he had some ideas. Leigh
can you elaborate on those ? 

Friendly,

Sven Luther

On Fri, February 25, 2005 7:36, Sven Luther said:

So, now, we need to find out what the problems where, i think it is
something that went in between 2.6.8 and 2.6.10, and leigh said he had
some ideas.

may i ask what patches were applied to a vanilla 2.6.8 kernel to build the
2.6.8-d-i then?

thanks,
Christian.
-- 
make bzImage, not war

On Fri, Feb 25, 2005 at 12:59:04PM +0100, Christian wrote:

On Fri, February 25, 2005 7:36, Sven Luther said:

quoted

So, now, we need to find out what the problems where, i think it is
something that went in between 2.6.8 and 2.6.10, and leigh said he had
some ideas.

may i ask what patches were applied to a vanilla 2.6.8 kernel to build the
2.6.8-d-i then?

Some backports that i got from the list. The complete list of patches is at :

  http://svn.debian.org/wsvn/kernel/trunk/kernel/source/kernel-source-2.6.8-2.6.8/debian/patches/?rev=0&sc=0

And i guess the one at hand here is : 

  http://svn.debian.org/wsvn/kernel/trunk/kernel/source/kernel-source-2.6.8-2.6.8/debian/patches/powerpc-prep-powerstack-irq.dpatch?op=file&rev=0&sc=0

--- kernel-source-2.6.8.orig/arch/ppc/platforms/prep_pci.c	2004-12-28

08:24:07.000000000 +0100

+++ kernel-source-2.6.8/arch/ppc/platforms/prep_pci.c	2005-01-03

11:15:30.604274816 +0100ll lines beginning with `## DP:' are a description of
the patch.
## DP: Description: Fix PReP - motorola powerstack II utah pci irq mapping.
## DP: Patch author: Tom Rini [off-list ref]
## DP: Upstream status: backport

. $(dirname $0)/DPATCH

@DPATCH@

@@ -115,13 +115,13 @@
 static char Utah_pci_IRQ_map[23] __prepdata =
 {
         0,   /* Slot 0  - unused */
-        0,   /* Slot 1  - unused */
+        4,   /* Slot 1  - IDE - SL82C105 */
         5,   /* Slot 2  - SCSI - NCR825A  */
         0,   /* Slot 3  - unused */
-        1,   /* Slot 4  - Ethernet - DEC2114x */
+        3,   /* Slot 4  - Ethernet - DEC2114x */
         0,   /* Slot 5  - unused */
-        3,   /* Slot 6  - PCI Card slot #1 */
-        4,   /* Slot 7  - PCI Card slot #2 */
+        2,   /* Slot 6  - PCI Card slot #1 */
+        3,   /* Slot 7  - PCI Card slot #2 */
         5,   /* Slot 8  - PCI Card slot #3 */
         5,   /* Slot 9  - PCI Bridge */
              /* added here in case we ever support PCI bridges */

Friendly,

Sven Luther

Sven Luther wrote:

Some backports that i got from the list. The complete list of patches is at :

  http://svn.debian.org/wsvn/kernel/trunk/kernel/source/kernel-source-2.6.8-2.6.8/debian/patches/?rev=0&sc=0

dooh, these websvn patches are giving me a headache.... will have to 
learn /usr/bin/svn first :-\

yes, the prep_pci.c and its irq-mappings. the PowerStackII lines were 
changed back and forth, and a current 2.6-BK is only different in one 
line to the patch you mentioned:

http://nerdbynature.de/bits/hal/2.6.11-rc5.patched/powerpc-prep-powerstack-irq_2.6.11-rc5.patch

unfortunately it did not help either and i'll switch back to vanilla 
2.6.8 again and hopefully find out exactly when scsi stopped working.

http://nerdbynature.de/bits/hal/2.6.11-rc5/
http://nerdbynature.de/bits/hal/2.6.11-rc5.patched/


thank you for your concern,
Christian.

On Sat, Feb 26, 2005 at 04:39:45AM +0100, Christian wrote:

Sven Luther wrote:

quoted

Some backports that i got from the list. The complete list of patches is 
at :

 http://svn.debian.org/wsvn/kernel/trunk/kernel/source/kernel-source-2.6.8-2.6.8/debian/patches/?rev=0&sc=0

dooh, these websvn patches are giving me a headache.... will have to 
learn /usr/bin/svn first :-\

Well : 

  svn co svn://svn.debian.org/kernel/trunk/kernel/source/kernel-source-2.6.8-2.6.8/debian/patches

quoted

--- kernel-source-2.6.8.orig/arch/ppc/platforms/prep_pci.c	2004-12-28

yes, the prep_pci.c and its irq-mappings. the PowerStackII lines were 
changed back and forth, and a current 2.6-BK is only different in one 
line to the patch you mentioned:

I guess the one line is the one for the IDE device, ..., indeed. The one line
in question is to enable the onboard IDE controller, which exist but is
probably not used, since the place on the board where it should be has no
connector soldered. I hear that there is an IDE powerstack II model though, so
...

http://nerdbynature.de/bits/hal/2.6.11-rc5.patched/powerpc-prep-powerstack-irq_2.6.11-rc5.patch

unfortunately it did not help either and i'll switch back to vanilla 
2.6.8 again and hopefully find out exactly when scsi stopped working.

As i understand leigh's and other post, i believe that this is the fix, but
that other stuff went in too which did break.

http://nerdbynature.de/bits/hal/2.6.11-rc5/
http://nerdbynature.de/bits/hal/2.6.11-rc5.patched/

Friendly,

Sven Luther