[RFH] ACPI testers needed!

[Timothy Pearson]

All,

I have recently updated the coreboot ACPI autogeneration code to 
generate valid processor _PSS objects when more than 10 cores are 
installed in one system.

Unfortunately this required a rather large update to multiple boards in 
order to change the _PSS object name from CPUn to CPnn.  We would like 
some feedback to make sure this update didn't break anything before 
committing.

The patch is available at: http://review.coreboot.org/8422

Thanks!

[Paul Menzel]

Dear Timothy,

Am Mittwoch, den 11.02.2015, 11:41 -0600 schrieb Timothy Pearson:
> I have recently updated the coreboot ACPI autogeneration code to 
> generate valid processor _PSS objects when more than 10 cores are 
> installed in one system.
>
> Unfortunately this required a rather large update to multiple boards in 
> order to change the _PSS object name from CPUn to CPnn.  We would like 
> some feedback to make sure this update didn't break anything before 
> committing.
>
> The patch is available at: http://review.coreboot.org/8422

I tried to test this with an image built for QEMU/KVM as you can specify
the number of CPUs there with the switch `-smp cpus=n`. But I have not
looked yet, if the changed code is even used by the QEMU image and I did
not have time yet, to get an OS image to actually verify that everything
works.

Do you have a suggestion? Just `/proc/cpuinfo`? Does that depend on
correct ACPI entries?

Thanks,

Paul

[Timothy Pearson]

On 02/11/2015 05:51 PM, Paul Menzel wrote:
> Dear Timothy,
>
> Am Mittwoch, den 11.02.2015, 11:41 -0600 schrieb Timothy Pearson:
> > I have recently updated the coreboot ACPI autogeneration code to
> > generate valid processor _PSS objects when more than 10 cores are
> > installed in one system.
> >
> > Unfortunately this required a rather large update to multiple boards in
> > order to change the _PSS object name from CPUn to CPnn.  We would like
> > some feedback to make sure this update didn't break anything before
> > committing.
> >
> > The patch is available at: http://review.coreboot.org/8422
>
> I tried to test this with an image built for QEMU/KVM as you can specify
> the number of CPUs there with the switch `-smp cpus=n`. But I have not
> looked yet, if the changed code is even used by the QEMU image and I did
> not have time yet, to get an OS image to actually verify that everything
> works.
>
> Do you have a suggestion? Just `/proc/cpuinfo`? Does that depend on
> correct ACPI entries?
>
> Thanks,
>
> Paul

Sorry, I should have mentioned a test procedure with my request. :-)

Boot Linux and execute:
dmesg | grep ACPI

Look for any failures in that output; they will be very obvious (e.g. 
AE_NOT_FOUND, etc.).

If the dmesg output looks good and your CPU supports frequency scaling 
you can also check this:
ls /sys/devices/system/cpu/cpu*/cpufreq/cpuinfo_cur_freq | wc -l

If all the processor objects were created correctly the output of that 
command should match the number of physical CPU cores active on your system.

Note that if your mainboard does not support ACPI processor objects 
these checks won't work, but at the same time it should not have been 
affected by the patch in question.

[RFH] ACPI testers needed! -- free food*

[Alex G.]

On 02/11/2015 11:41 AM, Timothy Pearson wrote:
> All,
>
> I have recently updated the coreboot ACPI autogeneration code to
> generate valid processor _PSS objects when more than 10 cores are
> installed in one system.
>
> Unfortunately this required a rather large update to multiple boards in
> order to change the _PSS object name from CPUn to CPnn.  We would like
> some feedback to make sure this update didn't break anything before
> committing.
>
> The patch is available at: http://review.coreboot.org/8422

We're planning to merge that patch, but I'd like to give people a little
more time to make sure there aren't any outstanding or obvious issues.

Of course, the idea of free food usually gets people interested.

Alex

(*) free food for thought