From: Nhat Pham <nphamcs@gmail.com>
To: linux-mm@kvack.org
Cc: akpm@linux-foundation.org, hannes@cmpxchg.org, hughd@google.com,
yosry.ahmed@linux.dev, mhocko@kernel.org,
roman.gushchin@linux.dev, shakeel.butt@linux.dev,
muchun.song@linux.dev, len.brown@intel.com,
chengming.zhou@linux.dev, kasong@tencent.com, chrisl@kernel.org,
huang.ying.caritas@gmail.com, ryan.roberts@arm.com,
viro@zeniv.linux.org.uk, baohua@kernel.org, osalvador@suse.de,
lorenzo.stoakes@oracle.com, christophe.leroy@csgroup.eu,
pavel@kernel.org, kernel-team@meta.com,
linux-kernel@vger.kernel.org, cgroups@vger.kernel.org,
linux-pm@vger.kernel.org
Subject: [RFC PATCH 09/14] swap: implement locking out swapoff using virtual swap slot
Date: Mon, 7 Apr 2025 16:42:10 -0700 [thread overview]
Message-ID: <20250407234223.1059191-10-nphamcs@gmail.com> (raw)
In-Reply-To: <20250407234223.1059191-1-nphamcs@gmail.com>
In the old design, we obtain a reference to the swap device to maintain
the validity of the device's metadata struct (i.e swap_info_struct), as
well as the swap entry itself, before various operations.
In the new virtual swap space design, however, this is no longer
necessary - we can simply acquire a reference to the virtual swap slot
itself to ensure it remains valid.
Furthermore, once we decouple virtual swap slots from their backing,
obtaining a reference to the backing swap device itself is not
sufficient or even possible anyway, as the backing of a virtual swap
slot can change under it.
Signed-off-by: Nhat Pham <nphamcs@gmail.com>
---
include/linux/swap.h | 24 +++++++++++++++++++++++-
mm/vswap.c | 36 ++++++++++++++++++++++++++++++++++++
2 files changed, 59 insertions(+), 1 deletion(-)
diff --git a/include/linux/swap.h b/include/linux/swap.h
index 1d8679bd57f3..7f6200f1db33 100644
--- a/include/linux/swap.h
+++ b/include/linux/swap.h
@@ -730,11 +730,33 @@ int vswap_init(void);
void vswap_exit(void);
swp_slot_t swp_entry_to_swp_slot(swp_entry_t entry);
swp_entry_t swp_slot_to_swp_entry(swp_slot_t slot);
+bool vswap_tryget(swp_entry_t entry);
+void vswap_put(swp_entry_t entry);
bool folio_swapped(struct folio *folio);
bool vswap_swapcache_only(swp_entry_t entry, int nr);
int non_swapcache_batch(swp_entry_t entry, int nr);
bool swap_free_nr_any_cache_only(swp_entry_t entry, int nr);
void put_swap_folio(struct folio *folio, swp_entry_t entry);
+
+static inline bool trylock_swapoff(swp_entry_t entry,
+ struct swap_info_struct **si)
+{
+ if (!vswap_tryget(entry))
+ return false;
+
+ /*
+ * No need to hold a reference to the swap device. The virtual swap slot pins
+ * the physical swap slot, which in turns pin the swap device.
+ */
+ *si = swap_slot_swap_info(swp_entry_to_swp_slot(entry));
+ return true;
+}
+
+static inline void unlock_swapoff(swp_entry_t entry,
+ struct swap_info_struct *si)
+{
+ vswap_put(entry);
+}
#else
static inline int vswap_init(void)
{
@@ -773,7 +795,6 @@ static inline void put_swap_folio(struct folio *folio, swp_entry_t entry)
{
swap_slot_put_folio(swp_entry_to_swp_slot(entry), folio);
}
-#endif
static inline bool trylock_swapoff(swp_entry_t entry,
struct swap_info_struct **si)
@@ -789,6 +810,7 @@ static inline void unlock_swapoff(swp_entry_t entry,
{
swap_slot_put_swap_info(si);
}
+#endif
#if defined(CONFIG_SWAP) && !defined(CONFIG_VIRTUAL_SWAP)
int add_swap_count_continuation(swp_entry_t, gfp_t);
diff --git a/mm/vswap.c b/mm/vswap.c
index 1b8cf894390c..8a518ebd20e4 100644
--- a/mm/vswap.c
+++ b/mm/vswap.c
@@ -425,6 +425,42 @@ swp_entry_t swp_slot_to_swp_entry(swp_slot_t slot)
return ret;
}
+/**
+ * vswap_tryget - try to obtain an ephemeral reference to a virtual swap slot.
+ *
+ * @entry: the virtual swap slot.
+ *
+ * Return: true if the reference was obtained.
+ */
+bool vswap_tryget(swp_entry_t entry)
+{
+ struct swp_desc *desc;
+ bool ret;
+
+ rcu_read_lock();
+ desc = xa_load(&vswap_map, entry.val);
+ if (!desc) {
+ rcu_read_unlock();
+ return false;
+ }
+
+ ret = kref_get_unless_zero(&desc->refcnt);
+ rcu_read_unlock();
+ return ret;
+}
+
+/**
+ * vswap_put - release an ephemeral reference to the virtual swap slot.
+ *
+ * @entry: the virtual swap slot.
+ */
+void vswap_put(swp_entry_t entry)
+{
+ struct swp_desc *desc = xa_load(&vswap_map, entry.val);
+
+ kref_put(&desc->refcnt, vswap_ref_release);
+}
+
/**
* swap_free_nr_any_cache_only - decrease the swap count of nr contiguous swap
* entries by 1 (when the swap entries are removed
--
2.47.1
next prev parent reply other threads:[~2025-04-07 23:42 UTC|newest]
Thread overview: 35+ messages / expand[flat|nested] mbox.gz Atom feed top
2025-04-07 23:42 [RFC PATCH 00/14] Virtual Swap Space Nhat Pham
2025-04-07 23:42 ` [RFC PATCH 01/14] swapfile: rearrange functions Nhat Pham
2025-04-07 23:42 ` [RFC PATCH 02/14] mm: swap: add an abstract API for locking out swapoff Nhat Pham
2025-04-07 23:42 ` [RFC PATCH 03/14] mm: swap: add a separate type for physical swap slots Nhat Pham
2025-04-08 14:15 ` Johannes Weiner
2025-04-08 15:11 ` Nhat Pham
2025-04-22 14:41 ` Yosry Ahmed
[not found] ` <6807ab09.670a0220.152ca3.502fSMTPIN_ADDED_BROKEN@mx.google.com>
2025-04-22 15:50 ` Nhat Pham
2025-04-22 18:50 ` Kairui Song
2025-04-07 23:42 ` [RFC PATCH 04/14] mm: swap: swap cache support for virtualized swap Nhat Pham
2025-04-08 15:00 ` Johannes Weiner
2025-04-08 15:34 ` Nhat Pham
2025-04-08 15:43 ` Nhat Pham
2025-04-07 23:42 ` [RFC PATCH 05/14] zswap: unify zswap tree " Nhat Pham
2025-04-07 23:42 ` [RFC PATCH 06/14] mm: swap: allocate a virtual swap slot for each swapped out page Nhat Pham
2025-04-07 23:42 ` [RFC PATCH 07/14] swap: implement the swap_cgroup API using virtual swap Nhat Pham
2025-04-07 23:42 ` [RFC PATCH 08/14] swap: manage swap entry lifetime at the virtual swap layer Nhat Pham
2025-04-07 23:42 ` Nhat Pham [this message]
2025-04-07 23:42 ` [RFC PATCH 10/14] mm: swap: decouple virtual swap slot from backing store Nhat Pham
2025-04-07 23:42 ` [RFC PATCH 11/14] memcg: swap: only charge physical swap slots Nhat Pham
2025-04-07 23:42 ` [RFC PATCH 12/14] vswap: support THP swapin and batch free_swap_and_cache Nhat Pham
2025-04-07 23:42 ` [RFC PATCH 13/14] swap: simplify swapoff using virtual swap Nhat Pham
2025-04-07 23:42 ` [RFC PATCH 14/14] zswap: do not start zswap shrinker if there is no physical swap slots Nhat Pham
2025-04-08 13:04 ` [RFC PATCH 00/14] Virtual Swap Space Usama Arif
2025-04-08 15:20 ` Nhat Pham
2025-04-08 15:45 ` Johannes Weiner
2025-04-08 16:25 ` Nhat Pham
2025-04-08 16:27 ` Nhat Pham
2025-04-08 16:22 ` Kairui Song
2025-04-08 16:47 ` Nhat Pham
2025-04-08 16:59 ` Kairui Song
2025-04-22 14:43 ` Yosry Ahmed
2025-04-22 14:56 ` Yosry Ahmed
[not found] ` <6807afd0.a70a0220.2ae8b9.e07cSMTPIN_ADDED_BROKEN@mx.google.com>
2025-04-22 17:15 ` Nhat Pham
2025-04-22 19:29 ` Nhat Pham
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20250407234223.1059191-10-nphamcs@gmail.com \
--to=nphamcs@gmail.com \
--cc=akpm@linux-foundation.org \
--cc=baohua@kernel.org \
--cc=cgroups@vger.kernel.org \
--cc=chengming.zhou@linux.dev \
--cc=chrisl@kernel.org \
--cc=christophe.leroy@csgroup.eu \
--cc=hannes@cmpxchg.org \
--cc=huang.ying.caritas@gmail.com \
--cc=hughd@google.com \
--cc=kasong@tencent.com \
--cc=kernel-team@meta.com \
--cc=len.brown@intel.com \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-mm@kvack.org \
--cc=linux-pm@vger.kernel.org \
--cc=lorenzo.stoakes@oracle.com \
--cc=mhocko@kernel.org \
--cc=muchun.song@linux.dev \
--cc=osalvador@suse.de \
--cc=pavel@kernel.org \
--cc=roman.gushchin@linux.dev \
--cc=ryan.roberts@arm.com \
--cc=shakeel.butt@linux.dev \
--cc=viro@zeniv.linux.org.uk \
--cc=yosry.ahmed@linux.dev \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox