On Fri, Mar 20, 2026 at 10:20:37AM -0400, Mathieu Desnoyers wrote:
> On 2026-03-20 09:31, Mathieu Desnoyers wrote:
> > On 2026-03-20 09:21, Harry Yoo (Oracle) wrote:
> > > On Fri, Mar 20, 2026 at 08:35:46AM -0400, Mathieu Desnoyers wrote:
> > > > On 2026-03-20 00:17, Harry Yoo wrote:
> > > > [...]
> > > > > > [1]: https://lore.kernel.org/20260227153730.1556542-4-
> > > > > > [email protected]/
> > > > > 
> > > > > @Mathieu: In patch 1/3 description,
> > > > > > Changes since v7:
> > > > > > - Explicitly initialize the subsystem from start_kernel() right
> > > > > >     after mm_core_init() so it is up and running before
> > > > > > the creation of
> > > > > >     the first mm at boot.
> > > > > 
> > > > > But how does this work when someone calls mm_cpumask() on
> > > > > init_mm early?
> > > > > Looks like it will behave incorrectly because 
> > > > > get_rss_stat_items_size()
> > > > > returns zero?
> > > > 
> > > > It doesn't work as expected at all. I missed that all users of
> > > > mm_cpumask()
> > > > end up relying on get_rss_stat_items_size(), which now calls
> > > > percpu_counter_tree_items_size(), which depends on initialization from
> > > > percpu_counter_tree_subsystem_init().
> > > > 
> > > > If you add a call to percpu_counter_tree_subsystem_init in
> > > > arch/powerpc/kernel/setup_arch() just before:
> 
> [...]
> 
> One thing we could do to catch this kind of init sequence issue
> is to add a WARN_ON_ONCE in percpu_counter_tree_items_size:
> 
> size_t percpu_counter_tree_items_size(void)
> {
>         if (WARN_ON_ONCE(!nr_cpus_order))
>                 return 0;
>         return counter_config->nr_items * sizeof(struct 
> percpu_counter_tree_level_item);

Looks good!

-- 
Cheers,
Harry / Hyeonggon

Reply via email to