Re: [PATCH 1/2] mm/damon/core: cover discrete System RAM areas with per-range regions

[SJ Park]

On Tue, 28 Jul 2026 23:17:01 +0800 Jiayuan Chen <jiayuan.chen@xxxxxxxxx> wrote:

> 
> On 7/28/26 10:30 PM, SJ Park wrote:
> > On Tue, 28 Jul 2026 18:06:48 +0800 Jiayuan Chen <jiayuan.chen@xxxxxxxxx> wrote:
> >
> >> On 7/27/26 10:26 PM, SJ Park wrote:
> >>> Hello Jiayuan,
> >>>
> >>> On Mon, 27 Jul 2026 17:54:22 +0800 Jiayuan Chen <jiayuan.chen@xxxxxxxxx> wrote:
> >>>
> >>>> From: Jiayuan Chen <jiayuan.chen@xxxxxxxxxx>
[...]
> > Actually I was also thinking about this problem for in the past.  One of my
> > idea at that time was, handle only a few largest holes that practically being
> > problems.  That is, while reading the system ram layout, find the holes, sort
> > those by size, and do make holes in DAMON regions layout for the biggest N
> > (say, 2) holes.  This may handle most cases including your 5 TiB hole.
> > Actually vaddr is doing this, so we may be able to reuse some of the code.
> >
> > If it makes sense to you, I will try to implement this.
> 
> One worry with a fixed N is that it's a magic number. If a machine has more
> than N big holes (more sockets / NUMA nodes, or several CXL devices), the
> extra ones are silently not excluded and get walked again.
> 
> And making N configurable just moves the per-machine tuning back to the
> user, which is exactly what I'm trying to avoid.
> 
> vaddr over-includes the gaps into its regions too, but it can skip them
> cheaply: it walks the VMA tree (find_vma / maple tree), it's cheap.
> paddr has no such structure, so an included hole is walked in full.
> 
> That's why a fixed N is safe for vaddr but risky for paddr.

I agree it has the problem of magic number.  That said, I expect common systems
wouldn't have that many problematic large holes.  If it is not clearly wrong,
I'd consider starting with magic number 2.  I feel magic number 2 is appealing
probably because we could reuse vaddr's code in the case.  But it would already
be much better than nothing in the problematic setups.  At least it will cover
your 5 TiB hole.

Depending on what we learn from the fields, we could change the magic number or
add more optimizations in future.  Setting N as a fraction of max_nr_regions
and/or handling holes based on their size (relative to real System RAMs?) are a
few ideas off the top of my head.

> 
> 
> >>> user-visible parameters and internal state is also a concern.
> >>
> >> I think monitor_region_start/end is meant to expose the overall range, and
> >> skipping the holes inside it is an implementation detail. Even if we didn't
> >> skip them, the adaptive region count is never 1 anyway — DAMON already
> >> splits [first,last] into many regions. The holes just add a few more, so
> >> the param never matched the internal state exactly to begin with.
> > It is arguable, but I believe this also makes sense in my opinion.
> >
> >>
> >>> A quick workaround would be adjusting the memory layout in BIOS, using DAMON
> >>> sysfs interface instead, or setting the monitor_region_{start,end} to cover
> >>> only the single area.  Have you considered such workarounds?
> >>>
> >>> Let's complete this high level discussion first.
> >>
> >> Right, this is doable today — the DAMON sysfs interface can set multiple
> >>
> >> regions manually. This patch is only about convenience.
> > Are you actually running DAMON_LRU_SORT or DAMON_RECLAIM in a production system
> > having 5 TiB hole?  Or, planning to do?  If the above idea makes sense to you,
> > I could prioritize implementation of it depending on this.
> 
> 
> I'm not actually running DAMON_LRU_SORT or DAMON_RECLAIM on such a machine.

Thank you for clarifying!  Let's keep this brainstorm level discussion, then.

> I ran into this while working on per-cgroup hot/cold page tracking, where I
> was comparing performance and accuracy across setups — that's where these
> numbers came from.

Sounds interesting.  Btw, there are multiple ways to do per-cgroup monitoring.
Using DAMOS_STAT with memcg DAMOS filter is one of the ways.  Data
access-as-a-data-attribute patch series [1] will allow doing that in a much
more lightweight and flexible way.

> 
> 
> >
> > Thanks,
> > SJ
> >
> > [...]
> 
> 
> 
> And to be clear, I'm not trying to land a patch here — I just noticed this
> as a possible optimization. If you have a better idea, I'm happy to go with
> it :).

No worry!  FYI, using subject tag 'RFC' could be a good way to make the
intention more explicit, if you want.

[1] https://lore.kernel.org/20260728041242.187380-1-sj@xxxxxxxxxx


Thanks,
SJ