linux-pmac-stable SMP issues

linux-pmac-stable SMP issues

[Tony Mantler]

I just picked up a 9600/200mp, and the first thing I did with it (well,
maybe the second or third thing) was install Debian to play with linux SMP
stuff (and later play with some IDE PCI cards).

Debian apparently ships with a UP kernel, so I rsynced with
linux-pmac-stable and compiled an SMP kernel.

The first thing I noticed is that some EXPORT_SYMBOLs are missing, namely:
global_irq_holder, global_irq_lock, global_irq_count, global_bh_count,
global_bh_lock, kernel_flag and synchronize_bh. These missing symbols
prevent modules like lockd and appletalk from loading. Adding these in
makes everything happy-happy.

The second thing I noticed is that "Processor 1 is stuck.". I seem to
recall this issue has been brought up once or twice before, but a look
through the mailing list archives didn't reveal any clues.

I vaguely recall the issue might be related to the booter used, I'm using
BootX 1.2.2 (from MacOS 9.0.4). Is there a different booter I should be
using to acheive full SMP functionality - hopefully one that allows easy
dual-booting?

merida:~$ uname -a
Linux merida 2.2.17 #1 SMP Sat Oct 14 14:34:02 CDT 2000 ppc unknown
merida:~$ gcc -v
Reading specs from /usr/lib/gcc-lib/powerpc-linux/2.95.2/specs
gcc version 2.95.2 20000220 (Debian GNU/Linux)


Other than that, things work rather well, though 5 MB/s powermac scsi is
anything but spectacular (10MB/s made lots of stuff segfault) - leaves the
system feeling a tad sluggish compared to, say, my SGI Indigo2, but I
suppose that's not really a fair comparison to begin with. :)


Cheers - Tony 'Nicoya' Mantler :)


--
Tony "Nicoya" Mantler - Renaissance Nerd Extraordinaire - nicoya@apia.dhs.org
Winnipeg, Manitoba, Canada           --           http://nicoya.feline.pp.se/


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

Re: linux-pmac-stable SMP issues

[Michel Lanners]

On  16 Oct, this message from Tony Mantler echoed through cyberspace:
> The second thing I noticed is that "Processor 1 is stuck.". I seem to
> recall this issue has been brought up once or twice before, but a look
> through the mailing list archives didn't reveal any clues.
>
> I vaguely recall the issue might be related to the booter used, I'm using
> BootX 1.2.2 (from MacOS 9.0.4). Is there a different booter I should be
> using to acheive full SMP functionality - hopefully one that allows easy
> dual-booting?

It might not work as you want for dual booting, but you might want to
try OF booting with quik. It's rather manual, but it may bring up your
second processor...

> Other than that, things work rather well, though 5 MB/s powermac scsi is
> anything but spectacular (10MB/s made lots of stuff segfault) - leaves the
> system feeling a tad sluggish compared to, say, my SGI Indigo2, but I
> suppose that's not really a fair comparison to begin with. :)

PowerMac SCSI has always been on the slow side.... Nevertheless, I have
no problem running my MESH SCSI (7600) at 10MB/s. Strange... Maybe
that's why the default speed is 5 MB/s. Although, since the external
(53Cxx) also runs at 5 MB/s, and the internal SCSI is meant to be
faster, I'd say MacOS runs it at 10MB/s also?

Cheers

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: linux-pmac-stable SMP issues

[Benjamin Herrenschmidt]

>It might not work as you want for dual booting, but you might want to
>try OF booting with quik. It's rather manual, but it may bring up your
>second processor...

Yup, try miBoot or quik. The problem here is that new versions of MacOS
have a small SMP nanokernel that is taking over the second CPU before
BootX can react.


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

Re: linux-pmac-stable SMP issues

[Tony Mantler]

At 5:38 AM -0500 10/17/2000, Benjamin Herrenschmidt wrote:
>>It might not work as you want for dual booting, but you might want to
>>try OF booting with quik. It's rather manual, but it may bring up your
>>second processor...
>
>Yup, try miBoot or quik. The problem here is that new versions of MacOS
>have a small SMP nanokernel that is taking over the second CPU before
>BootX can react.

Ah, that would appear to be it. I booted with an 8.1 CD + BootX and I get
the full twin-penguins and everything. Spiff.

I would think it should be possible to have BootX hijack the second CPU in
a way that would make it un-stuck-able for when linux starts. I'd do this
myself but BootX doesn't build in MPW and I don't have codewarrior.

As much as I hate codewarrior, anyone care to make a donation? :)


merida:~$ cat /proc/cpuinfo
get_cpuinfo(): ran out of cpu nodes.
processor	: 0
cpu		: 604e
clock		: 200MHz
revision	: 2.2
bogomips	: 398.95

processor	: 1
cpu		: 604e
clock		: 200MHz
revision	: 2.2
bogomips	: 398.95

total bogomips	: 797.90
zero pages	: total 0 (0Kb) current: 0 (0Kb) hits: 0/282 (0%)
machine		: Power Macintosh
motherboard	: AAPL,9500 MacRISC
L2 cache	: 512K unified
memory		: 64MB
pmac-generation	: OldWorld
get_cpuinfo(): ran out of cpu nodes.


Cheers - Tony 'Nicoya' Mantler :)


--
Tony "Nicoya" Mantler - Renaissance Nerd Extraordinaire - nicoya@apia.dhs.org
Winnipeg, Manitoba, Canada           --           http://nicoya.feline.pp.se/


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