Re: [PATCH] SMP initialization: detection and enumeration

From:

Almudena Garcia

Subject:

Date:

Wed, 12 Aug 2020 16:03:06 +0200

I attach a new set of patches, fixing some errors

El mar., 11 ago. 2020 a las 23:44, Almudena Garcia (<liberamenso10000@gmail.com>) escribió:

I fixed the problem manually. I expect the patch works fine.

El mar., 11 ago. 2020 a las 23:24, Almudena Garcia (<liberamenso10000@gmail.com>) escribió:
oops!! I forgot add i386/i386/smp.h and i386/i386/smp.c to i386/Makefrag.am

I will try to fix the patch to add this

El mar., 11 ago. 2020 a las 23:03, Almudena Garcia (<liberamenso10000@gmail.com>) escribió:
I attach a new set of patches. I wait this time will be better than last time

El mar., 11 ago. 2020 a las 2:27, Almudena Garcia (<liberamenso10000@gmail.com>) escribió:
> So don't bring them online then? The user asked for 4 CPUs, so bring up
> 3 APs alongside the BSP and that's that.
It's not a bad idea. Then I will fix my cpu enumeration loop to add this new condition.

El mar., 11 ago. 2020 a las 2:23, Jessica Clarke (<jrtc27@jrtc27.com>) escribió:
On 11 Aug 2020, at 01:17, Almudena Garcia <liberamenso10000@gmail.com> wrote:
>
> > That being said, instead of hardcoding the maximum number of CPUs to
> > be 256, you can just let the user choose whatever value is preferred.
> > That's what Linux does.
>
> But this could cause coherency problems.

Coherency is a very specific thing in operating systems about the
properties of the underlying memory subsystem. It's not the word you're
looking for.

> By example, if the user sets mach_ncpus=4 , and the processor has 8 cores, It can produce an out-of-index error in the access to the arrays which store the info about the cpus.

So don't bring them online then? The user asked for 4 CPUs, so bring up
3 APs alongside the BSP and that's that.

Jess