Re: [PATCH] mm/sparsemem: fix race in accessing memory_section->usage

[David Hildenbrand]

On 25.10.23 23:35, Andrew Morton wrote:
> On Tue, 17 Oct 2023 19:40:15 +0530 Charan Teja Kalla <quic_charante@xxxxxxxxxxx> wrote:
>
> > > >   		 * was allocated during boot.
> > > >   		 */
> > > >   		if (!PageReserved(virt_to_page(ms->usage))) {
> > > > +			synchronize_rcu();
> > > >   			kfree(ms->usage);
> > > >   			ms->usage = NULL;
> > > >   		}
> > > If we add NULL checks around ms->usage, this becomes
> > >
> > > tmp = rcu_replace_pointer(ms->usage, NULL, hotplug_locked());
> > > syncrhonize_rcu();
> > > kfree(tmp);
> > Per David input, I am working on using kfree_rcu().
>
> How's it coming along?
>
> Given that we're at 6.6-rc7 and given that this issue is causing
> daily crashes in your device farm, I'm thinking that we use the current
> version of your patch for 6.6 and for -stable.  We can look at the
> kfree_rcu() optimization for later kernel releases?

Any particular reason we have to rush this in? It's been seen by one 
company in a testing farm; there were no other reports, especially not 
from production systems. ... and the issue seems to be quite old.

--
Cheers,

David / dhildenb