From: "Baek Kyu, Kim" <pocj8ur4in@gmail.com>
To: Andrew Morton <akpm@linux-foundation.org>,
Mike Rapoport <rppt@kernel.org>
Cc: "Baek Kyu, Kim" <pocj8ur4in@gmail.com>,
David Hildenbrand <david@kernel.org>,
Lorenzo Stoakes <ljs@kernel.org>,
linux-mm@kvack.org, linux-kernel@vger.kernel.org
Subject: [PATCH v2] mm/secretmem: Keep locked memory accounting until inode eviction
Date: Sun, 22 Nov 2026 01:40:20 +0900 [thread overview]
Message-ID: <20261121164020.1465498-1-pocj8ur4in@gmail.com> (raw)
In-Reply-To: <20261030152407.113568-1-pocj8ur4in@gmail.com>
A secretmem inode can outlive its original file. An O_PATH descriptor
opened through /proc/self/fd keeps the inode and its unevictable folios
alive after the mapping and original descriptor are closed. The file
release path nevertheless drops the locked-memory charge, allowing the
same user to reuse RLIMIT_MEMLOCK while those folios remain allocated.
Move private-state and locked-memory accounting cleanup to inode eviction,
after truncating the page cache. Keep secretmem_users active over the same
lifetime so that the hibernation guard covers retained secret pages.
Allocate private state before inode creation and release it if the
secure-inode helper fails. Attach it immediately after creation so
later failures can use iput() for cleanup. An LSM denial inside the
helper can call iput() before i_private is set; skip private cleanup
at eviction when i_private is NULL.
On arm64 QEMU with an 8 KiB RLIMIT_MEMLOCK, the original kernel allowed
an additional 4 KiB secretmem fault while O_PATH retained two pages.
v2 rejected that fault with SIGBUS and allowed it after O_PATH closed.
Injected LSM denial and file-allocation failure completed without a
crash; file-allocation cleanup returned secretmem_users to zero.
KASAN, lockdep, hibernation and long-running concurrency were not
tested in this v2 run.
Fixes: 97d34aa65c29 ("mm/secretmem: properly account locked pages")
Signed-off-by: Baek Kyu, Kim <pocj8ur4in@gmail.com>
---
Changes in v2:
- Allocate state before inode creation, as Mike Rapoport suggested,
and free it if the secure-inode helper fails.
- Guard private-state cleanup in secretmem_evict_inode() because an
LSM denial can call iput() before i_private is set.
Link to v1: https://lists.openwall.net/linux-kernel/2026/10/30/1883
mm/secretmem.c | 58 ++++++++++++++++++++++++++++------------------------------
1 file changed, 28 insertions(+), 30 deletions(-)
diff --git a/mm/secretmem.c b/mm/secretmem.c
index b3f79ad56790..5a7378b62bb2 100644
--- a/mm/secretmem.c
+++ b/mm/secretmem.c
@@ -189,14 +189,7 @@ static void secretmem_destroy_inode_priv(struct inode *inode)
free_uid(state->user);
kfree(state);
inode->i_private = NULL;
-}
-
-static int secretmem_release(struct inode *inode, struct file *file)
-{
atomic_dec(&secretmem_users);
- secretmem_destroy_inode_priv(inode);
-
- return 0;
}
static int secretmem_mmap_prepare(struct vm_area_desc *desc)
@@ -216,7 +209,6 @@ bool vma_is_secretmem(struct vm_area_struct *vma)
}
static const struct file_operations secretmem_fops = {
- .release = secretmem_release,
.mmap_prepare = secretmem_mmap_prepare,
};
@@ -264,39 +256,34 @@ static const struct inode_operations secretmem_iops = {
static struct vfsmount *secretmem_mnt;
-static int secretmem_init_inode_priv(struct inode *inode)
-{
- struct secretmem_inode_state *state;
-
- state = kzalloc_obj(*state);
- if (!state)
- return -ENOMEM;
-
- state->user = get_uid(current_user());
- inode->i_private = state;
- return 0;
-}
-
static struct file *secretmem_file_create(unsigned long flags)
{
+ struct secretmem_inode_state *state;
struct file *file;
struct inode *inode;
const char *anon_name = "[secretmem]";
int err;
+ state = kzalloc_obj(*state);
+ if (!state)
+ return ERR_PTR(-ENOMEM);
+ state->user = get_uid(current_user());
+
inode = anon_inode_make_secure_inode(secretmem_mnt->mnt_sb, anon_name, NULL);
- if (IS_ERR(inode))
+ if (IS_ERR(inode)) {
+ free_uid(state->user);
+ kfree(state);
return ERR_CAST(inode);
+ }
- err = secretmem_init_inode_priv(inode);
- if (err)
- goto err_free_inode;
+ inode->i_private = state;
+ atomic_inc(&secretmem_users);
file = alloc_file_pseudo(inode, secretmem_mnt, "secretmem",
O_RDWR | O_LARGEFILE, &secretmem_fops);
if (IS_ERR(file)) {
err = PTR_ERR(file);
- goto err_free_priv;
+ goto err_free_inode;
}
mapping_set_gfp_mask(inode->i_mapping, GFP_USER);
@@ -309,11 +296,7 @@ static struct file *secretmem_file_create(unsigned long flags)
inode->i_mode |= S_IFREG;
inode->i_size = 0;
- atomic_inc(&secretmem_users);
-
return file;
-err_free_priv:
- secretmem_destroy_inode_priv(inode);
err_free_inode:
iput(inode);
return ERR_PTR(err);
@@ -335,6 +318,20 @@ SYSCALL_DEFINE1(memfd_secret, unsigned int, flags)
return FD_ADD(flags & O_CLOEXEC, secretmem_file_create(flags));
}
+static void secretmem_evict_inode(struct inode *inode)
+{
+ truncate_inode_pages_final(&inode->i_data);
+ clear_inode(inode);
+ /* The filesystem root and failed secure inodes have no private state. */
+ if (inode->i_private)
+ secretmem_destroy_inode_priv(inode);
+}
+
+static const struct super_operations secretmem_super_ops = {
+ .statfs = simple_statfs,
+ .evict_inode = secretmem_evict_inode,
+};
+
static int secretmem_init_fs_context(struct fs_context *fc)
{
struct pseudo_fs_context *ctx;
@@ -342,6 +339,7 @@ static int secretmem_init_fs_context(struct fs_context *fc)
ctx = init_pseudo(fc, SECRETMEM_MAGIC);
if (!ctx)
return -ENOMEM;
+ ctx->ops = &secretmem_super_ops;
return 0;
}
prev parent reply other threads:[~2026-11-21 16:40 UTC|newest]
Thread overview: 5+ messages / expand[flat|nested] mbox.gz Atom feed top
2026-10-30 15:24 [PATCH] " Baek Kyu, Kim
2026-11-18 22:59 ` David Hildenbrand (Arm)
2026-11-19 7:36 ` Mike Rapoport
2026-11-20 9:32 ` David Hildenbrand (Arm)
2026-11-21 16:40 ` Baek Kyu, Kim [this message]
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=20261121164020.1465498-1-pocj8ur4in@gmail.com \
--to=pocj8ur4in@gmail.com \
--cc=akpm@linux-foundation.org \
--cc=david@kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-mm@kvack.org \
--cc=ljs@kernel.org \
--cc=rppt@kernel.org \
/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
all inboxes | Powered by JetHome®