Re: [RFC] Unconditionally lock folios when calling rmap_walk()

[Lokesh Gidra <lokeshgidra@google.com>]

On Fri, Aug 22, 2025 at 3:50 AM Lorenzo Stoakes
<lorenzo.stoakes@oracle.com> wrote:
>
> Lokesh, this is a mess :/
>
> I've been a reverse mapping co-maintainer for 4 months now, please check
> MAINTAINERS before sending this stuff out. It's not really impressive to
> learn about this 2nd hand...
>
> Also I cannot get lei to get this mail to me, no matter how hard I try. So
> it's _really hard_ for me to respond to this.
>
> And you label this '[RFC]' but I can't find any code (unless it's lost in
> the thread somehow). It should be '[DISCUSSION]'.
>
> rmap locking is _extremely_ sensitive and the discussion needs careful
> attention.
>
> It's hard for me to even track what's going on here or join the discussion,
> could you just please resend, correctly cc'ing the maintainers/reviewers of
> rmap, and prefix with '[DISCUSSION]'?

Sincere apologies. I'll resend correcting both the mistakes.
>
> You've already got responses here so we're inevitably going to fork the
> conversation, but unfortunately I don't see any way around this, because
> I'm going to miss all the conversation that I'm not cc'd on.
>
> Anwyay I simply can't engage on this as-is, and I really _want to_, because
> rmap and the locking around it are issues of PRIMARY importance to me.
>
> Please try to make my life easier :)
>
> Thanks, Lorenzo