[parisc-linux] 2.5.56-pa0 64bit successful boot

[parisc-linux] 2.5.56-pa0 64bit successful boot

[Thibaut VARENE]

[-- Attachment #1: Type: text/plain, Size: 649 bytes --]

Hi pa-ckers!

I just tested fresh 2.5.56 and successfully booted 64bit UP kernel.
Linux mkhppa3 2.5.56-pa0 #2 Sun Jan 12 01:34:36 CET 2003 parisc64 unknown unknown GNU/Linux

(native build with gcc 3.0.4 unofficial debs)

I had some errors with SCSI and SERIAL (see the attached dmesg and .config).
I'll have a look at them ASAP.

I'm gonna let the box run under load (seti), and see how it behaves.

I guess we can now bump pa nb to 'pa1' since the kernel has been booted ;)

FWIW, SMP kernel with the exact same configuration died just at:
POSIX conformance testing by UNIFIX

HTH,


Thibaut VARENE
The PA/Linux ESIEE Team
http://pateam.esiee.fr/

[-- Attachment #2: config.gz --]
[-- Type: application/octet-stream, Size: 3053 bytes --]

[-- Attachment #3: dmesg-2.5.56-a500.gz --]
[-- Type: application/octet-stream, Size: 2910 bytes --]

Re: [parisc-linux] 2.5.56-pa0 64bit successful boot

[Grant Grundler]

On Sun, Jan 12, 2003 at 02:10:41AM +0100, Thibaut VARENE wrote:
> Hi pa-ckers!
> 
> I just tested fresh 2.5.56 and successfully booted 64bit UP kernel.
> Linux mkhppa3 2.5.56-pa0 #2 Sun Jan 12 01:34:36 CET 2003 parisc64 unknown unknown GNU/Linux

cool!

> I had some errors with SCSI and SERIAL (see the attached dmesg and .config).
> I'll have a look at them ASAP.

you can also post them here...I'll be able to look at them tomorrow.

> FWIW, SMP kernel with the exact same configuration died just at:
> POSIX conformance testing by UNIFIX

What's the UP kernel say after that when it boots?

thanks,
grant

Re: [parisc-linux] 2.5.56-pa0 64bit successful boot

[Grant Grundler]

On Sun, Jan 12, 2003 at 02:10:41AM +0100, Thibaut VARENE wrote:
> FWIW, SMP kernel with the exact same configuration died just at:
> POSIX conformance testing by UNIFIX

I looked at the dmesg you had attached. Please don't compress
them in the future - defeats archive searches later.

The dmesg output reminded me the init sequence has substantially changed.
I'd be very susicious that additional CPUs are brought "online" later than
before.  I would guess that arch/parisc/kernel/smp.c and processor.c
needs to be changed to match. Can you retry CONFIG_SMP with
EARLY_BOOTUP_DEBUG enabled in arch/parisc/kernel/pdc_cons.c?

Oh, and also set kDEBUG in arch/parisc/kernel/smp.c to 100.

thanks,
grant