Re: [PATCH v3] ARM64: Dynamically allocate cpumasks and increase supported CPUs to 512

[Catalin Marinas <catalin.marinas@arm.com>]

On Fri, Mar 08, 2024 at 03:01:28PM +0100, Marek Szyprowski wrote:
> On 07.03.2024 02:45, Christoph Lameter (Ampere) wrote:
> > Currently defconfig selects NR_CPUS=256, but some vendors (e.g. Ampere
> > Computing) are planning to ship systems with 512 CPUs. So that all CPUs on
> > these systems can be used with defconfig, we'd like to bump NR_CPUS to 512.
> > Therefore this patch increases the default NR_CPUS from 256 to 512.
> >
> > As increasing NR_CPUS will increase the size of cpumasks, there's a fear that
> > this might have a significant impact on stack usage due to code which places
> > cpumasks on the stack. To mitigate that concern, we can select
> > CPUMASK_OFFSTACK. As that doesn't seem to be a problem today with
> > NR_CPUS=256, we only select this when NR_CPUS > 256.
> >
> > CPUMASK_OFFSTACK configures the cpumasks in the kernel to be
> > dynamically allocated. This was used in the X86 architecture in the
> > past to enable support for larger CPU configurations up to 8k cpus.
[...]
> This patch landed in today's linux-next as commit 0499a78369ad ("ARM64: 
> Dynamically allocate cpumasks and increase supported CPUs to 512"). 
> Unfortunately it triggers the following warning during boot on most of 
> my ARM64-based test boards. Here is an example from Odroid-N2 board:

I spent a big part of this afternoon going through the code paths but
there's nothing obvious that triggered this problem. My suspicion is
some memory corruption, algorithmically I can't see anything that could
go wrong with CPUMASK_OFFSTACK. Unfortunately I could not reproduce it
yet to be able to add some debug info.

So I decided to revert this patch. If we get to the bottom of it during
the merging window, I can still revive it. Otherwise we'll add it to
linux-next post -rc1.

Thanks for reporting it and subsequent debugging.

-- 
Catalin