Re: [PATCH v3] mm: introduce copy_mc_highpages() and batch copy in folio_mc_copy()
From: Andrew Morton
Date: Sun Nov 22 2026 - 02:30:10 EST
On Fri, 13 Nov 2026 17:09:21 +0000 Shivank Garg <shivankg@xxxxxxx> wrote:
> folio_mc_copy() copies a folio one page at a time with copy_mc_highpage(),
> so a 2MB folio makes 512 separate copy_mc_to_kernel() calls. Each page
> pays the startup cost of the string instruction ("REP; MOVSB" on x86
> CPUs with ERMS) plus a cond_resched().
>
> Introduce copy_mc_highpages(), the page range variant of
> copy_mc_highpage(). When copy_mc_to_kernel() is available and the pages
> are in the linear map, copy the whole range with a single
> copy_mc_to_kernel() call, so the CPU can optimize for the larger extent.
> Otherwise, fall back to copy_mc_highpage() one page at a time.
>
> On a machine check, queue the poisoned source page for memory_failure(),
> as copy_mc_highpage() does, and return -EHWPOISON, as folio_mc_copy()
> does today.
>
> copy_mc_to_kernel() cannot reschedule, so only take the single-call path
> when CONFIG_PREEMPTION is set and the copy can be preempted. Since commit
> 7dadeaa6e851 ("sched: Further restrict the preemption modes"), x86 and
> powerpc, the only architectures that select ARCH_HAS_COPY_MC, only offer
> the full, lazy and RT preemption models, also with PREEMPT_DYNAMIC, so
> they always take this path. The per-page fallback calls cond_resched()
> before each page, as clear_contig_highpages() does. Other architectures
> keep copying one page at a time.
>
> Switch folio_mc_copy() to copy the whole folio with copy_mc_highpages().
OK.
> ---
> On a Zen 6 system with CXL memory, move_pages() gets ~1.4x faster for
> 2MB THPs and up to ~2.8x faster for 1GB hugetlb pages (numbers below).
> When demoting to CXL, kswapd uses ~31% less CPU per GB demoted.
But that's the most important part! It's the whole point of the patch!
Yet here it is below the ---, which means "throw this away". It
should be the first paragraph, in all caps.
<edit edit fixed>
OK, thanks, queued,
nit: does it really need cond_resched()? That's not appropriate for
current mainline kernels. I guess "yes", because this patch looks like
something which downstream people will eagerly backport into older
kernels, even though we didn't cc:stable.
And we care about downstream people. Globally: "Roughly 60% are on 6.x
and 30% are on 5.x (with the remaining ~10% split between 7.x and 4.x
or older)".
I'll take a look at that Sashiko thing.