Re; 2.4.0-test9 on B&W G3

Re; 2.4.0-test9 on B&W G3

[Benjamin Herrenschmidt]

[-- Attachment #1: Type: text/plain, Size: 618 bytes --]

>I have also been seeing the same old 'lost interrupt' problem when trying
>to boot a B&W G3 using the latest yaboot (0.8) on the latest pmac-devel
>rsync kernel.  I was wondering if this was a know bug with the latest 2.4
>tree kernel's.  I also booted the same kernel on a new G4/450 with no
>problems.

I have a fix for that too. I'll push it as soon as I have access to bk again.
In the meantime, apply enclosed changeset to the current linuxppc_2_3 bk tree.

This changeset revives Geert PCI code, adds Samuel fix (indeed the
correct fix), and
add some more IDE fix for PCI-IDE controllers like the CMD646.

Ben.

[-- Attachment #2: pci_ide_fix.cset.bz2 --]
[-- Type: application/octet-stream, Size: 1929 bytes --]

Re: Re; 2.4.0-test9 on B&W G3

[Martin Costabel]

Benjamin Herrenschmidt wrote:
>
> >I have also been seeing the same old 'lost interrupt' problem when trying
> >to boot a B&W G3 using the latest yaboot (0.8) on the latest pmac-devel
> >rsync kernel.  I was wondering if this was a know bug with the latest 2.4
> >tree kernel's.  I also booted the same kernel on a new G4/450 with no
> >problems.
>
> I have a fix for that too. I'll push it as soon as I have access to bk again.
> In the meantime, apply enclosed changeset to the current linuxppc_2_3 bk tree.

Glad to hear that the linuxppc_2_3 bk tree is not dead. I was just going
to ask about it, because since 2 weeks it has remained at -test8 which
is feared for its terrible mailbox-eating habits.

On the other hand, I am running a kernel obtained from this tree by
applying just the "official" test9-pre1 patch, and I have now an uptime
of 4 days, 22:02. This is the longest uptime I ever had with a 2.3/4
kernel. No spontaneous freezes so far <knocking on wood>.

--
Martin

** Sent via the linuxppc-dev mail list. See http://lists.linuxppc.org/

Re: Re; 2.4.0-test9 on B&W G3

[Martin Costabel]

Martin Costabel wrote:

> Glad to hear that the linuxppc_2_3 bk tree is not dead. I was just going
> to ask about it, because since 2 weeks it has remained at -test8 which
> is feared for its terrible mailbox-eating habits.
>
> On the other hand, I am running a kernel obtained from this tree by
> applying just the "official" test9-pre1 patch, and I have now an uptime
> of 4 days, 22:02. This is the longest uptime I ever had with a 2.3/4
> kernel. No spontaneous freezes so far <knocking on wood>.

OK, I asked for it: As of right now, the bk linux_2_3 bk tree is at
2.4.0-test9-pre4. It is less dangerous than -test8 insofar as it doesn't
boot :-)

It compiles if one upgrades arch/ppc/kernel/time.c to, for example, the
one from Paul's linux-pmac-devel tree.

But it doesn't recognize my SCSI disk. I think it's a feature of
-test9-pre4 that SCSI is broken. This is patched in pre5 or pre6. So
I'll wait or use Paul's drivers/scsi directory...

--
Martin

** Sent via the linuxppc-dev mail list. See http://lists.linuxppc.org/

Re: Re; 2.4.0-test9 on B&W G3

[Tom Rini]

On Sat, Sep 23, 2000 at 07:19:17AM +0200, Martin Costabel wrote:

> But it doesn't recognize my SCSI disk. I think it's a feature of
> -test9-pre4 that SCSI is broken. This is patched in pre5 or pre6. So
> I'll wait or use Paul's drivers/scsi directory...

Yes, test9 is the "Lets play with VM and SCSI" release.  Tho every one has
booted on my x86 SCSI box (bk was up to pre5, or so jack howarth figured, either
way going up to 6 shouldn't be too hard..)

--
Tom Rini (TR1265)
http://gate.crashing.org/~trini/

** Sent via the linuxppc-dev mail list. See http://lists.linuxppc.org/

Re: Re; 2.4.0-test9 on B&W G3

[Martin Costabel]

Tom Rini wrote:
>
> On Sat, Sep 23, 2000 at 07:19:17AM +0200, Martin Costabel wrote:
>
> > But it doesn't recognize my SCSI disk. I think it's a feature of
> > -test9-pre4 that SCSI is broken. This is patched in pre5 or pre6. So
> > I'll wait or use Paul's drivers/scsi directory...
>
> Yes, test9 is the "Lets play with VM and SCSI" release.  Tho every one has
> booted on my x86 SCSI box (bk was up to pre5, or so jack howarth figured, either
> way going up to 6 shouldn't be too hard..)

I tried it with Paul's drivers/scsi, and it works now. The bk tree
didn't have the mesh.c update from pre5. Paul has Linus' ear, or so it
seems.

BTW, now that even the MOL patches have sneaked into Linus' tree, has
anyone got MOL to work on a 603 with the 2.4.0 kernel? Or what is
missing?

--
Martin

** Sent via the linuxppc-dev mail list. See http://lists.linuxppc.org/

Re: MOL on 2.4/603

[Samuel Rydh]

On Sat, Sep 23, 2000 at 07:57:22AM +0200, Martin Costabel wrote:
>
> Tom Rini wrote:
>
> BTW, now that even the MOL patches have sneaked into Linus' tree, has
> anyone got MOL to work on a 603 with the 2.4.0 kernel? Or what is
> missing?
>

Well, the MOL code in 2.4 has support for 603. However, the MOL
binary does not yet support the 603. (In 2.4, the PTE hash
is gone so MOL must now support the 603 MMU explicitly).

Fixing this is certainly on my todo-list (but so is a lot of other
things - if someone with MMU knowledge and a 603 wants to fix it,
I can give some assistance).

Btw, the lastest released mol version is *not* compatible with
the 2.4 kernel. To run MOL on the 2.4 kernel, pull the latest
version of MOL from the bitkeeper server:

	bk://waltari.theophys.kth.se:5000

Cheers,

/Samuel


----------------------------------------------------------
 E-mail <samuel@ibrium.se>  WWW: <http://www.ibrium.se>
  Phone/fax: (home) +46 8 4418431, (work) +46 8 7908470
----------------------------------------------------------

** Sent via the linuxppc-dev mail list. See http://lists.linuxppc.org/

Re: Re; 2.4.0-test9 on B&W G3

[Geert Uytterhoeven]

On Fri, 22 Sep 2000, Martin Costabel wrote:
> On the other hand, I am running a kernel obtained from this tree by
> applying just the "official" test9-pre1 patch, and I have now an uptime
> of 4 days, 22:02. This is the longest uptime I ever had with a 2.3/4
> kernel. No spontaneous freezes so far <knocking on wood>.

Looks like I should be happy I wasn't able to upgrade since test1-ac10, which
reached 54 days before I hit the power switch (and the reason I hit that switch
was not computer-related at all :-)...

Gr{oetje,eeting}s,

						Geert

--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
							    -- Linus Torvalds


** Sent via the linuxppc-dev mail list. See http://lists.linuxppc.org/

Re: Re; 2.4.0-test9 on B&W G3

[Michel Lanners]

Ben,

On  22 Sep, this message from Benjamin Herrenschmidt echoed through cyberspace:
> This changeset revives Geert PCI code, adds Samuel fix (indeed the
> correct fix), and
> add some more IDE fix for PCI-IDE controllers like the CMD646.

The pci_enable_device() in your changeset is not enough, unfortunately.
My Promise Ultra66 (which others, including Andre Hedrick also use)
doesn't identify as an IDE storage controller:

[mlan@piglet ~]$ lspci
00:0b.0 Host bridge: Apple Computer Inc. Bandit PowerPC host bridge (rev 03)
00:0d.0 Unknown mass storage controller: Promise Technology, Inc. 20262 (rev 01)

Would it be possible to select on mass storage in general? If that is an
'official' classification, that is...

Michel

-------------------------------------------------------------------------
Michel Lanners                 |  " Read Philosophy.  Study Art.
23, Rue Paul Henkes            |    Ask Questions.  Make Mistakes.
L-1710 Luxembourg              |
email   mlan@cpu.lu            |
http://www.cpu.lu/~mlan        |                     Learn Always. "


** Sent via the linuxppc-dev mail list. See http://lists.linuxppc.org/

Re: Re; 2.4.0-test9 on B&W G3

[Benjamin Herrenschmidt]

>
>The pci_enable_device() in your changeset is not enough, unfortunately.
>My Promise Ultra66 (which others, including Andre Hedrick also use)
>doesn't identify as an IDE storage controller:
>
>[mlan@piglet ~]$ lspci
>00:0b.0 Host bridge: Apple Computer Inc. Bandit PowerPC host bridge (rev 03)
>00:0d.0 Unknown mass storage controller: Promise Technology, Inc. 20262
>(rev 01)
>
>Would it be possible to select on mass storage in general? If that is an
>'official' classification, that is...

Well, we could. That is supposed to remain a temporary hack until we find
a better way to tell the IDE driver to stop avoiding disabled cards if
told so by the arch.

Ben.

** Sent via the linuxppc-dev mail list. See http://lists.linuxppc.org/