state_show() reads kdamond->damon_ctx without holding damon_sysfs_lock.
This allows a use-after-free race:

CPU 0                         CPU 1
-----                         -----
state_show()                  damon_sysfs_turn_damon_on()
ctx = kdamond->damon_ctx;     mutex_lock(&damon_sysfs_lock);
                              damon_destroy_ctx(kdamond->damon_ctx);
                              kdamond->damon_ctx = NULL;
                              mutex_unlock(&damon_sysfs_lock);
damon_is_running(ctx);        /* ctx is freed */
mutex_lock(&ctx->kdamond_lock); /* UAF */

(The race can also occur with damon_sysfs_kdamonds_rm_dirs() and
damon_sysfs_kdamond_release(), which free or replace the context under
damon_sysfs_lock.)

Fix by taking damon_sysfs_lock before dereferencing the context,
mirroring the locking used in pid_show().

The bug has existed since state_show() first accessed kdamond->damon_ctx.

Fixes: a61ea561c871 ("mm/damon/sysfs: link DAMON for virtual address spaces monitoring")
Reported-by: Stanislav Fort <disclosure@aisle.com>
Closes: N/A # non-publicly reported
Signed-off-by: Stanislav Fort <disclosure@aisle.com>
---
 mm/damon/sysfs.c | 14 +++++++++-----
 1 file changed, 9 insertions(+), 5 deletions(-)

diff --git a/mm/damon/sysfs.c b/mm/damon/sysfs.c
index 6d2b0dab50cb..7b9254cadd5f 100644
--- a/mm/damon/sysfs.c
+++ b/mm/damon/sysfs.c
@@ -1260,14 +1260,18 @@ static ssize_t state_show(struct kobject *kobj, struct kobj_attribute *attr,
 {
 	struct damon_sysfs_kdamond *kdamond = container_of(kobj,
 			struct damon_sysfs_kdamond, kobj);
-	struct damon_ctx *ctx = kdamond->damon_ctx;
-	bool running;
+	struct damon_ctx *ctx;
+	bool running = false;
 
-	if (!ctx)
-		running = false;
-	else
+	if (!mutex_trylock(&damon_sysfs_lock))
+		return -EBUSY;
+
+	ctx = kdamond->damon_ctx;
+	if (ctx)
 		running = damon_is_running(ctx);
 
+	mutex_unlock(&damon_sysfs_lock);
+
 	return sysfs_emit(buf, "%s\n", running ?
 			damon_sysfs_cmd_strs[DAMON_SYSFS_CMD_ON] :
 			damon_sysfs_cmd_strs[DAMON_SYSFS_CMD_OFF]);
-- 
2.39.3 (Apple Git-146)

+ Andrew

On Fri,  5 Sep 2025 13:10:46 +0300 Stanislav Fort <stanislav.fort@aisle.com> wrote:

> state_show() reads kdamond->damon_ctx without holding damon_sysfs_lock.
> This allows a use-after-free race:
> 
> CPU 0                         CPU 1
> -----                         -----
> state_show()                  damon_sysfs_turn_damon_on()
> ctx = kdamond->damon_ctx;     mutex_lock(&damon_sysfs_lock);
>                               damon_destroy_ctx(kdamond->damon_ctx);
>                               kdamond->damon_ctx = NULL;
>                               mutex_unlock(&damon_sysfs_lock);
> damon_is_running(ctx);        /* ctx is freed */
> mutex_lock(&ctx->kdamond_lock); /* UAF */
> 
> (The race can also occur with damon_sysfs_kdamonds_rm_dirs() and
> damon_sysfs_kdamond_release(), which free or replace the context under
> damon_sysfs_lock.)
> 
> Fix by taking damon_sysfs_lock before dereferencing the context,
> mirroring the locking used in pid_show().
> 
> The bug has existed since state_show() first accessed kdamond->damon_ctx.
> 
> Fixes: a61ea561c871 ("mm/damon/sysfs: link DAMON for virtual address spaces monitoring")
> Reported-by: Stanislav Fort <disclosure@aisle.com>
> Closes: N/A # non-publicly reported
> Signed-off-by: Stanislav Fort <disclosure@aisle.com>

Reviewed-by: SeongJae Park <sj@kernel.org>


Thanks,
SJ

[...]