[PATCH v2] mm/khugepaged: cap min_free_kbytes recommendation at 1 GiB

[Nimrod Oren <noren@nvidia.com>]

When THP is enabled, set_recommended_min_free_kbytes() may raise
min_free_kbytes using a heuristic that scales with pageblock_nr_pages.
With MIGRATE_PCPTYPES equal to 3, the formula accounts for 11
pageblocks for every eligible populated zone before capping the result
at 5% of low memory.

This is reasonable when a pageblock is 2 MiB, as on common 4 KiB page
configurations, but scales poorly with larger base page sizes. With
the default arm64 pageblock sizes, the contribution per eligible zone
before the 5% cap is:

  4 KiB pages:   2 MiB pageblock,   22 MiB per zone
 16 KiB pages:  32 MiB pageblock,  352 MiB per zone
 64 KiB pages: 512 MiB pageblock,  5.5 GiB per zone

Consequently, min_free_kbytes can reach excessive and unwanted levels.

Add an absolute 1 GiB cap to the recommendation, in addition to the
existing percentage cap. This bounds the automatic recommendation to a
sane value on systems with large pageblocks while preserving existing
behavior for typical systems with 2 MiB pageblocks.

Link: https://lore.kernel.org/r/20260716173504.760369-1-noren@nvidia.com/
Suggested-by: Lorenzo Stoakes (ARM) <ljs@kernel.org>
Signed-off-by: Nimrod Oren <noren@nvidia.com>
---
v2:
* Use a named constant for the cap.
* Drop explicit linux/sizes.h include.
* Update min_free_kbytes documentation.

v1:
* Cap the final recommendation at 1 GiB as suggested by Lorenzo.
https://lore.kernel.org/r/20260728202014.2517142-1-noren@nvidia.com/

RFC v1:
https://lore.kernel.org/r/20260716173504.760369-1-noren@nvidia.com/
---
 Documentation/admin-guide/sysctl/vm.rst | 8 ++++++++
 mm/khugepaged.c                         | 7 ++++++-
 2 files changed, 14 insertions(+), 1 deletion(-)

diff --git a/Documentation/admin-guide/sysctl/vm.rst b/Documentation/admin-guide/sysctl/vm.rst
index 5b318d17aa4b..c0c925e6f504 100644
--- a/Documentation/admin-guide/sysctl/vm.rst
+++ b/Documentation/admin-guide/sysctl/vm.rst
@@ -539,6 +539,14 @@ watermark[WMARK_MIN] value for each lowmem zone in the system.
 Each lowmem zone gets a number of reserved free pages based
 proportionally on its size.
 
+When Transparent Hugepage (THP) support is enabled through global or
+per-size controls, the kernel may raise this value automatically to
+help keep pageblocks free and reduce fragmentation for THP allocations.
+The automatic recommendation cannot exceed 5% of low memory or 1 GiB,
+whichever is lower.  This limit applies only to the automatic
+recommendation and does not constrain a higher value written by
+userspace.
+
 Some minimal amount of memory is needed to satisfy PF_MEMALLOC
 allocations; if you set this to lower than 1024KB, your system will
 become subtly broken, and prone to deadlock under high loads.
diff --git a/mm/khugepaged.c b/mm/khugepaged.c
index f49a6710933b..a8fc061d2488 100644
--- a/mm/khugepaged.c
+++ b/mm/khugepaged.c
@@ -101,6 +101,7 @@ static DEFINE_READ_MOSTLY_HASHTABLE(mm_slots_hash, MM_SLOTS_HASH_BITS);
 
 static struct kmem_cache *mm_slot_cache __ro_after_init;
 
+#define THP_MIN_FREE_MAX_PAGES		(SZ_1G / PAGE_SIZE)
 #define KHUGEPAGED_MIN_MTHP_ORDER	2
 
 struct collapse_control {
@@ -3120,9 +3121,13 @@ void set_recommended_min_free_kbytes(void)
 	recommended_min += pageblock_nr_pages * nr_zones *
 			   MIGRATE_PCPTYPES * MIGRATE_PCPTYPES;
 
-	/* don't ever allow to reserve more than 5% of the lowmem */
+	/*
+	 * Don't allow the THP recommendation to exceed 5% of lowmem or an
+	 * absolute cap, whichever is smaller.
+	 */
 	recommended_min = min(recommended_min,
 			      (unsigned long) nr_free_buffer_pages() / 20);
+	recommended_min = min(recommended_min, THP_MIN_FREE_MAX_PAGES);
 	recommended_min <<= (PAGE_SHIFT-10);
 
 	if (recommended_min > min_free_kbytes) {
-- 
2.45.0