From: Bijan Tabatabai <bijantabatab@micron.com>
migrate_{hot,cold} are paddr schemes that are used to migrate hot/cold
data to a specified node. However, these schemes are only available when
doing physical address monitoring. This patch adds an implementation for
them virtual address monitoring as well.
Co-developed-by: Ravi Shankar Jonnalagadda <ravis.opensrc@micron.com>
Signed-off-by: Ravi Shankar Jonnalagadda <ravis.opensrc@micron.com>
Signed-off-by: Bijan Tabatabai <bijantabatab@micron.com>
---
mm/damon/vaddr.c | 102 +++++++++++++++++++++++++++++++++++++++++++++++
1 file changed, 102 insertions(+)
diff --git a/mm/damon/vaddr.c b/mm/damon/vaddr.c
index 46554e49a478..5cdfdc47c5ff 100644
--- a/mm/damon/vaddr.c
+++ b/mm/damon/vaddr.c
@@ -15,6 +15,7 @@
#include <linux/pagewalk.h>
#include <linux/sched/mm.h>
+#include "../internal.h"
#include "ops-common.h"
#ifdef CONFIG_DAMON_VADDR_KUNIT_TEST
@@ -610,6 +611,65 @@ static unsigned int damon_va_check_accesses(struct damon_ctx *ctx)
return max_nr_accesses;
}
+#ifdef CONFIG_TRANSPARENT_HUGEPAGE
+static int damos_va_migrate_pmd_entry(pmd_t *pmd, unsigned long addr,
+ unsigned long next, struct mm_walk *walk)
+{
+ struct list_head *migration_list = walk->private;
+ struct folio *folio;
+ spinlock_t *ptl;
+ pmd_t pmde;
+
+ ptl = pmd_lock(walk->mm, pmd);
+ pmde = pmdp_get(pmd);
+
+ if (!pmd_present(pmde) || !pmd_trans_huge(pmde))
+ goto unlock;
+
+ folio = damon_get_folio(pmd_pfn(pmde));
+ if (!folio)
+ goto unlock;
+
+ if (!folio_isolate_lru(folio))
+ goto put_folio;
+
+ list_add(&folio->lru, migration_list);
+
+put_folio:
+ folio_put(folio);
+unlock:
+ spin_unlock(ptl);
+ return 0;
+}
+#else
+#define damos_va_migrate_pmd_entry NULL
+#endif /* CONFIG_TRANSPARENT_HUGEPAGE */
+
+static int damos_va_migrate_pte_entry(pte_t *pte, unsigned long addr,
+ unsigned long enxt, struct mm_walk *walk)
+{
+ struct list_head *migration_list = walk->private;
+ struct folio *folio;
+ pte_t ptent;
+
+ ptent = ptep_get(pte);
+ if (pte_none(*pte) || !pte_present(*pte))
+ return 0;
+
+ folio = damon_get_folio(pte_pfn(ptent));
+ if (!folio)
+ return 0;
+
+ if (!folio_isolate_lru(folio))
+ goto out;
+
+ list_add(&folio->lru, migration_list);
+
+out:
+ folio_put(folio);
+ return 0;
+}
+
/*
* Functions for the target validity check and cleanup
*/
@@ -653,6 +713,41 @@ static unsigned long damos_madvise(struct damon_target *target,
}
#endif /* CONFIG_ADVISE_SYSCALLS */
+static unsigned long damos_va_migrate(struct damon_target *target,
+ struct damon_region *r, struct damos *s,
+ unsigned long *sz_filter_passed)
+{
+ LIST_HEAD(folio_list);
+ struct task_struct *task;
+ struct mm_struct *mm;
+ unsigned long applied = 0;
+ struct mm_walk_ops walk_ops = {
+ .pmd_entry = damos_va_migrate_pmd_entry,
+ .pte_entry = damos_va_migrate_pte_entry,
+ .walk_lock = PGWALK_RDLOCK,
+ };
+
+ task = damon_get_task_struct(target);
+ if (!task)
+ return 0;
+
+ mm = damon_get_mm(target);
+ if (!mm)
+ goto put_task;
+
+ mmap_read_lock(mm);
+ walk_page_range(mm, r->ar.start, r->ar.end, &walk_ops, &folio_list);
+ mmap_read_unlock(mm);
+ mmput(mm);
+
+ applied = damon_migrate_pages(&folio_list, s->target_nid);
+ cond_resched();
+
+put_task:
+ put_task_struct(task);
+ return applied * PAGE_SIZE;
+}
+
static unsigned long damon_va_apply_scheme(struct damon_ctx *ctx,
struct damon_target *t, struct damon_region *r,
struct damos *scheme, unsigned long *sz_filter_passed)
@@ -675,6 +770,9 @@ static unsigned long damon_va_apply_scheme(struct damon_ctx *ctx,
case DAMOS_NOHUGEPAGE:
madv_action = MADV_NOHUGEPAGE;
break;
+ case DAMOS_MIGRATE_HOT:
+ case DAMOS_MIGRATE_COLD:
+ return damos_va_migrate(t, r, scheme, sz_filter_passed);
case DAMOS_STAT:
return 0;
default:
@@ -695,6 +793,10 @@ static int damon_va_scheme_score(struct damon_ctx *context,
switch (scheme->action) {
case DAMOS_PAGEOUT:
return damon_cold_score(context, r, scheme);
+ case DAMOS_MIGRATE_HOT:
+ return damon_hot_score(context, r, scheme);
+ case DAMOS_MIGRATE_COLD:
+ return damon_cold_score(context, r, scheme);
default:
break;
}
--
2.43.5On Wed, 2 Jul 2025 15:13:32 -0500 Bijan Tabatabai <bijan311@gmail.com> wrote:
> From: Bijan Tabatabai <bijantabatab@micron.com>
>
> migrate_{hot,cold} are paddr schemes that are used to migrate hot/cold
> data to a specified node. However, these schemes are only available when
> doing physical address monitoring. This patch adds an implementation for
> them virtual address monitoring as well.
>
> Co-developed-by: Ravi Shankar Jonnalagadda <ravis.opensrc@micron.com>
> Signed-off-by: Ravi Shankar Jonnalagadda <ravis.opensrc@micron.com>
> Signed-off-by: Bijan Tabatabai <bijantabatab@micron.com>
> ---
> mm/damon/vaddr.c | 102 +++++++++++++++++++++++++++++++++++++++++++++++
> 1 file changed, 102 insertions(+)
>
> diff --git a/mm/damon/vaddr.c b/mm/damon/vaddr.c
> index 46554e49a478..5cdfdc47c5ff 100644
> --- a/mm/damon/vaddr.c
> +++ b/mm/damon/vaddr.c
> @@ -15,6 +15,7 @@
> #include <linux/pagewalk.h>
> #include <linux/sched/mm.h>
>
> +#include "../internal.h"
> #include "ops-common.h"
>
> #ifdef CONFIG_DAMON_VADDR_KUNIT_TEST
> @@ -610,6 +611,65 @@ static unsigned int damon_va_check_accesses(struct damon_ctx *ctx)
> return max_nr_accesses;
> }
>
> +#ifdef CONFIG_TRANSPARENT_HUGEPAGE
> +static int damos_va_migrate_pmd_entry(pmd_t *pmd, unsigned long addr,
> + unsigned long next, struct mm_walk *walk)
I'd suggest to put CONFIG_TRANSPARENT_HUGEPAGE check into the body of this
function and handle both pmd and pte here, consistent to
damon_young_pmd_entry().
> +{
> + struct list_head *migration_list = walk->private;
> + struct folio *folio;
> + spinlock_t *ptl;
> + pmd_t pmde;
> +
> + ptl = pmd_lock(walk->mm, pmd);
> + pmde = pmdp_get(pmd);
> +
> + if (!pmd_present(pmde) || !pmd_trans_huge(pmde))
> + goto unlock;
> +
> + folio = damon_get_folio(pmd_pfn(pmde));
> + if (!folio)
> + goto unlock;
> +
> + if (!folio_isolate_lru(folio))
> + goto put_folio;
> +
> + list_add(&folio->lru, migration_list);
> +
> +put_folio:
> + folio_put(folio);
> +unlock:
> + spin_unlock(ptl);
> + return 0;
> +}
> +#else
> +#define damos_va_migrate_pmd_entry NULL
> +#endif /* CONFIG_TRANSPARENT_HUGEPAGE */
> +
> +static int damos_va_migrate_pte_entry(pte_t *pte, unsigned long addr,
> + unsigned long enxt, struct mm_walk *walk)
Nit. s/enxt/next/ ?
> +{
> + struct list_head *migration_list = walk->private;
> + struct folio *folio;
> + pte_t ptent;
> +
> + ptent = ptep_get(pte);
> + if (pte_none(*pte) || !pte_present(*pte))
> + return 0;
Shouldn't we use cached pte value (ptent) instad of *pte? I'd suggest merging
this into damos_va_migrate_pmd_entry() consistent to damon_young_pmd_entry().
> +
> + folio = damon_get_folio(pte_pfn(ptent));
> + if (!folio)
> + return 0;
> +
> + if (!folio_isolate_lru(folio))
> + goto out;
> +
> + list_add(&folio->lru, migration_list);
> +
> +out:
> + folio_put(folio);
> + return 0;
> +}
> +
> /*
> * Functions for the target validity check and cleanup
> */
> @@ -653,6 +713,41 @@ static unsigned long damos_madvise(struct damon_target *target,
> }
> #endif /* CONFIG_ADVISE_SYSCALLS */
>
> +static unsigned long damos_va_migrate(struct damon_target *target,
> + struct damon_region *r, struct damos *s,
> + unsigned long *sz_filter_passed)
> +{
> + LIST_HEAD(folio_list);
> + struct task_struct *task;
> + struct mm_struct *mm;
> + unsigned long applied = 0;
> + struct mm_walk_ops walk_ops = {
> + .pmd_entry = damos_va_migrate_pmd_entry,
> + .pte_entry = damos_va_migrate_pte_entry,
> + .walk_lock = PGWALK_RDLOCK,
> + };
> +
> + task = damon_get_task_struct(target);
> + if (!task)
> + return 0;
> +
> + mm = damon_get_mm(target);
> + if (!mm)
> + goto put_task;
> +
> + mmap_read_lock(mm);
> + walk_page_range(mm, r->ar.start, r->ar.end, &walk_ops, &folio_list);
> + mmap_read_unlock(mm);
> + mmput(mm);
> +
> + applied = damon_migrate_pages(&folio_list, s->target_nid);
> + cond_resched();
> +
> +put_task:
> + put_task_struct(task);
Seems task is not being used in real, so this variable and the related code in
this function can be removed? Or, am I missing something?
> + return applied * PAGE_SIZE;
> +}
> +
> static unsigned long damon_va_apply_scheme(struct damon_ctx *ctx,
> struct damon_target *t, struct damon_region *r,
> struct damos *scheme, unsigned long *sz_filter_passed)
> @@ -675,6 +770,9 @@ static unsigned long damon_va_apply_scheme(struct damon_ctx *ctx,
> case DAMOS_NOHUGEPAGE:
> madv_action = MADV_NOHUGEPAGE;
> break;
> + case DAMOS_MIGRATE_HOT:
> + case DAMOS_MIGRATE_COLD:
> + return damos_va_migrate(t, r, scheme, sz_filter_passed);
> case DAMOS_STAT:
> return 0;
> default:
> @@ -695,6 +793,10 @@ static int damon_va_scheme_score(struct damon_ctx *context,
> switch (scheme->action) {
> case DAMOS_PAGEOUT:
> return damon_cold_score(context, r, scheme);
> + case DAMOS_MIGRATE_HOT:
> + return damon_hot_score(context, r, scheme);
> + case DAMOS_MIGRATE_COLD:
> + return damon_cold_score(context, r, scheme);
> default:
> break;
> }
> --
> 2.43.5
Thanks,
SJ
On Wed, 2 Jul 2025 16:51:38 -0700 SeongJae Park <sj@kernel.org> wrote:
> On Wed, 2 Jul 2025 15:13:32 -0500 Bijan Tabatabai <bijan311@gmail.com> wrote:
>
> > From: Bijan Tabatabai <bijantabatab@micron.com>
> >
> > migrate_{hot,cold} are paddr schemes that are used to migrate hot/cold
> > data to a specified node. However, these schemes are only available when
> > doing physical address monitoring. This patch adds an implementation for
> > them virtual address monitoring as well.
> >
> > Co-developed-by: Ravi Shankar Jonnalagadda <ravis.opensrc@micron.com>
> > Signed-off-by: Ravi Shankar Jonnalagadda <ravis.opensrc@micron.com>
> > Signed-off-by: Bijan Tabatabai <bijantabatab@micron.com>
> > ---
> > mm/damon/vaddr.c | 102 +++++++++++++++++++++++++++++++++++++++++++++++
> > 1 file changed, 102 insertions(+)
> >
> > diff --git a/mm/damon/vaddr.c b/mm/damon/vaddr.c
> > index 46554e49a478..5cdfdc47c5ff 100644
> > --- a/mm/damon/vaddr.c
> > +++ b/mm/damon/vaddr.c
> > @@ -15,6 +15,7 @@
> > #include <linux/pagewalk.h>
> > #include <linux/sched/mm.h>
> >
> > +#include "../internal.h"
> > #include "ops-common.h"
> >
> > #ifdef CONFIG_DAMON_VADDR_KUNIT_TEST
> > @@ -610,6 +611,65 @@ static unsigned int damon_va_check_accesses(struct damon_ctx *ctx)
> > return max_nr_accesses;
> > }
> >
> > +#ifdef CONFIG_TRANSPARENT_HUGEPAGE
> > +static int damos_va_migrate_pmd_entry(pmd_t *pmd, unsigned long addr,
> > + unsigned long next, struct mm_walk *walk)
>
> I'd suggest to put CONFIG_TRANSPARENT_HUGEPAGE check into the body of this
> function and handle both pmd and pte here, consistent to
> damon_young_pmd_entry().
Ah, unlike damon_young_pmd_entry() which is for a single address, this is for
walking the range of a given DAMON region, and hence should have a separate pte
entry function. Please ignore the above comment.
[...]
> > +static int damos_va_migrate_pte_entry(pte_t *pte, unsigned long addr,
> > + unsigned long enxt, struct mm_walk *walk)
>
> Nit. s/enxt/next/ ?
>
> > +{
> > + struct list_head *migration_list = walk->private;
> > + struct folio *folio;
> > + pte_t ptent;
> > +
> > + ptent = ptep_get(pte);
> > + if (pte_none(*pte) || !pte_present(*pte))
> > + return 0;
>
> Shouldn't we use cached pte value (ptent) instad of *pte? I'd suggest merging
> this into damos_va_migrate_pmd_entry() consistent to damon_young_pmd_entry().
Again, I overlooked the fact that this is for walking not only single address
point but a range. Please ignore the latter suggestion.
Thanks,
SJ
[...]
© 2016 - 2025 Red Hat, Inc.