Re: [PATCH v3] mm: page_reporting: allow driver to set batch capacity

["Michael S. Tsirkin" <mst@redhat.com>]

On Sun, Jul 05, 2026 at 01:51:16AM -0700, Andrew Morton wrote:
> On Sun, 5 Jul 2026 04:42:38 -0400 "Michael S. Tsirkin" <mst@redhat.com> wrote:
> 
> > On Sun, Jul 05, 2026 at 01:39:56AM -0700, Andrew Morton wrote:
> > > On Sun, 5 Jul 2026 04:25:53 -0400 "Michael S. Tsirkin" <mst@redhat.com> wrote:
> > > 
> > > > > > If capacity is 0 or exceeds PAGE_REPORTING_CAPACITY, it defaults
> > > > > > to PAGE_REPORTING_CAPACITY.  The 0 check and the clamping is done in
> > > > > > page_reporting_register(), before the reporting work is scheduled,
> > > > > > so we never get division by 0.
> > > > > 
> > > > > Thanks.  What's the priority here?  Should we fix 7.2?  Earlier?
> > > > 
> > > > If possible, I'd like it in 7.2, have a setup like this (small
> > > > queue is useful for perf testing).
> > > > It's very early in the cycle and the code seems straight forward...
> > > > I judge the chances of breaking anything is tiny because
> > > > it just failed probe previously.
> > > 
> > > No probs, I moved this to the the mm-hotfixes-unstable queue.  No cc:stable.
> > 
> > I suspect the Fixes tag will make their AI pick it anyway. But hey.
> 
> -stable maintainers have been asked not to do that - only backport
> patches which have an explicit cc:stable.
> 
> So should I add one?

The "scratch my own itch, I want my hardware to work ASAP" part of me wants
to say yes. But the "decades of experience taught me I'm fallible" part
says let's have this out in the field for a while first. I had time
for coffee, and I'm dr Jekyll now. So no, thanks!

-- 
MST