mm/damon: allow NULL or unstarted damon_ctx parameter for damon_call()
Subsystems (2)
SJ Park
maintainercover mm/damon: allow NULL or unstarted damon_ctx parameter for damon_call()
Callers of damon_call() should validate the damon_ctx object parameter. If it is NULL or never damon_start()-ed object, damon_call() could dereference the NULL pointer or indefinitely hang. Ensuring all callers doing the validation correctly has turned out to be difficult. Handle the corner cases inside the core layer and remove callers' validations.
Patches 1 and 2 respectively allow passing NULL and not yet damon_start()-ed ctx parameter to damon_call(). Patches 3 and 4 remove the callers side validations in DMON_RECLAIM and DASMON_LRU_SORT, respectively.
Changes from RFC
- RFC: https://lore.kernel.org/[email protected]
- Drop RFC tag.
- Rebase to lates mm-new.
SJ Park (4):
mm/damon/core: handle NULL ctx parameter in damon_call() mm/damon/core: set ctx->call_controls_obsolete in damon_new_ctx() mm/damon/reclaim: remove unnecessary damon_call() param validation mm/damon/lru_sort: remove unnecessary damon_call() param validation
mm/damon/core.c | 7 +++---- mm/damon/lru_sort.c | 8 -------- mm/damon/reclaim.c | 8 -------- 3 files changed, 3 insertions(+), 20 deletions(-)
base-commit: 7fe9a8a366a72a631bcb23c2b108669ab182f2f2
Signature · 1 line
2.47.3
SJ Park
maintainerPATCH 1/4 mm/damon/core: handle NULL ctx parameter in damon_call()
When NULL damon_ctx pointer parameter is passed, damon_call() could do NULL dereference. The caller is responsible to avoid that. It is easy to forget, and there are many damon_call() callers. Meanwhile, damon_call() is never meant to be performance critical. It uses mutex and completion. Add the NULL pointer check inside damon_call() so that callers can pass the parameter without NULL checks.
--- mm/damon/core.c | 2 ++ 1 file changed, 2 insertions(+)
mm/damon/core.c +2 -0
@@ -2206,6 +2206,8 @@ int damon_kdamond_pid(struct damon_ctx *ctx)*/ int damon_call (struct damon_ctx *ctx,struct damon_call_control *control) { +if (!ctx) +return -EINVAL ;if (!control->repeat)init_completion (&control->completion); control->canceled = false;
Signature · 1 line
2.47.3
SJ Park
maintainerPATCH 2/4 mm/damon/core: set ctx->call_controls_obsolete in damon_new_ctx()
In reply to SJ Parkdamon_ctx->call_controls_obsolete is used to disallow damon_call() requests when the request cannot be served. The field is unset and set when the context execution is started and terminated, respectively. The intention is to allow damon_call() requests only while the context is actively being executed.
damon_ctx constructor, damon_new_ctx() unsets the field, though. As a result, passing the damon_ctx parameter that never successfully damon_start()-ed to damon_call() can indefinitely hang. The callers should ensure to avoid the case. Such parameter validation is not always simple. Actually such bugs in DAMON_RECLAIM and DAMON_LRU_SORT have been found and fixed [1].
Set the field in damon_new_ctx(), so that DAMON API callers can pass the context parameter to damon_call() without the additional check.
[1] https://lore.kernel.org/[email protected]
--- mm/damon/core.c | 5 +---- 1 file changed, 1 insertion(+), 4 deletions(-)
mm/damon/core.c +1 -4
@@ -933,6 +933,7 @@ struct damon_ctx *damon_new_ctx(void)INIT_LIST_HEAD (&ctx->adaptive_targets);INIT_LIST_HEAD (&ctx->schemes); + ctx->call_controls_obsolete = true;prandom_seed_state (&ctx->rnd_state,get_random_u64 ());return ctx;
@@ -2196,10 +2197,6 @@ int damon_kdamond_pid(struct damon_ctx *ctx)* synchronization. The return value of the function will be saved in * &damon_call_control->return_code. * -* Note that this function should be called only after damon_start() with the -* @ctx has succeeded. Otherwise, this function could fall into an indefinite -* wait. -* * When this function is failed, the @ctx is guaranteed to be stopped. * * Return: 0 on success, negative error code otherwise.
Signature · 1 line
2.47.3
SJ Park
maintainerPATCH 3/4 mm/damon/reclaim: remove unnecessary damon_call() param validation
In reply to SJ ParkDAMON_RECLAIM avoids passing NULL or unstarted damon_ctx to damon_call() with its own validation. The validation is no longer needed, because the DAMON core layer now handles the corner cases itself. Remove the unnecessary check.
--- mm/damon/reclaim.c | 8 -------- 1 file changed, 8 deletions(-)
mm/damon/reclaim.c +0 -8
@@ -271,8 +271,6 @@ static int damon_reclaim_commit_inputs_fn(void *arg)return damon_reclaim_apply_parameters (); } -static bool damon_reclaim_damon_has_started; -static int damon_reclaim_commit_inputs_store (const char *val,const struct kernel_param *kp) {
@@ -293,10 +291,6 @@ static int damon_reclaim_commit_inputs_store(const char *val,if (!commit_inputs_request)return 0 ; -/* Skip damon_call() if ctx has not successfully started. */ -if (!damon_reclaim_damon_has_started) -return -EINVAL ; - err =damon_call (ctx, &control);return err ? err : control.return_code;
@@ -343,8 +337,6 @@ static int damon_reclaim_turn(bool on) err =damon_start (&ctx,1 , true);if (err)return err; -if (!damon_reclaim_damon_has_started) - damon_reclaim_damon_has_started = true;return damon_call (ctx, &call_control); }
Signature · 1 line
2.47.3
SJ Park
maintainerPATCH 4/4 mm/damon/lru_sort: remove unnecessary damon_call() param validation
In reply to SJ ParkDAMON_LRU_SORT avoids passing NULL or unstarted damon_ctx to damon_call() with its own validation. The validation is no longer needed, because the DAMON core layer now handles the corner cases itself. Remove the unnecessary check.
--- mm/damon/lru_sort.c | 8 -------- 1 file changed, 8 deletions(-)
mm/damon/lru_sort.c +0 -8
@@ -346,8 +346,6 @@ static int damon_lru_sort_commit_inputs_fn(void *arg)return damon_lru_sort_apply_parameters (); } -static bool damon_lru_sort_damon_has_started; -static int damon_lru_sort_commit_inputs_store (const char *val,const struct kernel_param *kp) {
@@ -368,10 +366,6 @@ static int damon_lru_sort_commit_inputs_store(const char *val,if (!commit_inputs_request)return 0 ; -/* Skip damon_call() if ctx has not successfully started. */ -if (!damon_lru_sort_damon_has_started) -return -EINVAL ; - err =damon_call (ctx, &control);return err ? err : control.return_code;
@@ -422,8 +416,6 @@ static int damon_lru_sort_turn(bool on) err =damon_start (&ctx,1 , true);if (err)return err; -if (!damon_lru_sort_damon_has_started) - damon_lru_sort_damon_has_started = true;return damon_call (ctx, &call_control); }
Signature · 1 line
2.47.3
[email protected]
AI reviewRe: [PATCH 4/4] mm/damon/lru_sort: remove unnecessary damon_call() param validation
Sashiko has reviewed this patch and found no issues. It looks great!
Quoted 6 lines
DAMON_LRU_SORTavoids passing NULL or unstarteddamon_ctxtodamon_call()with its own validation. The validation is no longer needed, because the DAMON core layer now handles the corner cases itself. Remove the unnecessary check.Signed-off-by SJ Park <[email protected]>
Sashiko has reviewed this patch and found no issues. It looks great!
Signature · 1 line
Sashiko AI review · https://sashiko.dev/#/patchset/[email protected]?part=4
[email protected]
AI reviewRe: [PATCH 3/4] mm/damon/reclaim: remove unnecessary damon_call() param validation
In reply to SJ ParkSashiko has reviewed this patch and found no issues. It looks great!
Quoted 6 lines
DAMON_RECLAIMavoids passing NULL or unstarteddamon_ctxtodamon_call()with its own validation. The validation is no longer needed, because the DAMON core layer now handles the corner cases itself. Remove the unnecessary check.Signed-off-by SJ Park <[email protected]>
Sashiko has reviewed this patch and found no issues. It looks great!
Signature · 1 line
Sashiko AI review · https://sashiko.dev/#/patchset/[email protected]?part=3
[email protected]
AI reviewRe: [PATCH 1/4] mm/damon/core: handle NULL ctx parameter in damon_call()
In reply to SJ ParkSashiko has reviewed this patch and found no issues. It looks great!
Quoted 8 lines
When NULL
damon_ctxpointer parameter is passed,damon_call()could do NULL dereference. The caller is responsible to avoid that. It is easy to forget, and there are manydamon_call()callers. Meanwhile,damon_call()is never meant to be performance critical. It uses mutex and completion. Add the NULL pointer check insidedamon_call()so that callers can pass the parameter without NULL checks.Signed-off-by SJ Park <[email protected]>
Sashiko has reviewed this patch and found no issues. It looks great!
Signature · 1 line
Sashiko AI review · https://sashiko.dev/#/patchset/[email protected]?part=1
[email protected]
AI reviewRe: [PATCH 2/4] mm/damon/core: set ctx->call_controls_obsolete in damon_new_ctx()
In reply to SJ ParkThank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
Pre-existing issues:
- [High] The patch correctly fixes an indefinite wait vulnerability in damon_call() by initializing ctx->call_controls_obsolete to true, but fails to initialize the identical ctx->walk_control_obsolete flag. This leaves damos_walk() vulnerable to the exact same indefinite hang if called on an unstarted context. --
commit 55412502962e64df1de83b2c9805595450e93fc7
Author: SJ Park <[email protected]>
mm/damon/core: set ctx->call_controls_obsolete in damon_new_ctx()
This patch sets ctx->call_controls_obsolete to true in damon_new_ctx() to prevent indefinite hangs when damon_call() is used on an unstarted context.
mm/damon/core.c struct damon_ctx *damon_new_ctx(void) · line 936
Quoted 9 lines
mm/damon/core.c +1
-0@@ -933,6 +933,7 @@ struct damon_ctx *damon_new_ctx(void)INIT_LIST_HEAD (&ctx->adaptive_targets);INIT_LIST_HEAD (&ctx->schemes); + ctx->call_controls_obsolete = true;
[Severity: High]
This is a pre-existing issue, but does the same fix need to be applied to ctx->walk_control_obsolete?
Because walk_control_obsolete is zero-initialized (false) by kzalloc_obj(), calling damos_walk() on an unstarted context will bypass the obsolete check:
mm/damon/core.c:damos_walk() {
if (ctx->walk_control_obsolete) {
mutex_unlock(&ctx->walk_control_lock);
return -ECANCELED;
}
...
wait_for_completion(&control->completion);
}This could lead to an indefinite hang in wait_for_completion(), since the DAMON worker thread is not running to signal the completion. Should ctx->walk_control_obsolete also be initialized to true here?
Quoted 3 lines
mm/damon/core.c +0
-0prandom_seed_state (&ctx->rnd_state,get_random_u64 ());return ctx;
[ ... ]
Signature · 1 line
Sashiko AI review · https://sashiko.dev/#/patchset/[email protected]?part=2
SJ Park
maintainerRe: [PATCH 2/4] mm/damon/core: set ctx->call_controls_obsolete in damon_new_ctx()
Quoted 42 lines · On Thu, 03 Sep 2026 01:16:20 +0000 [email protected] wrote:
Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
Pre-existing issues:
- [High] The patch correctly fixes an indefinite wait vulnerability indamon_call()by initializingctx->call_controls_obsoleteto true, but fails to initialize the identicalctx->walk_control_obsoleteflag. This leavesdamos_walk()vulnerable to the exact same indefinite hang if called on an unstarted context. --commit
55412502962e64df1de83b2c9805595450e93fc7
Author: SJ Park <[email protected]>
mm/damon/core: setctx->call_controls_obsoleteindamon_new_ctx()This patch sets
ctx->call_controls_obsoleteto true indamon_new_ctx()to prevent indefinite hangs whendamon_call()is used on an unstarted context.mm/damon/core.c struct damon_ctx *damon_new_ctx(void) · line 936
Quoted 9 lines
mm/damon/core.c +1
-0@@ -933,6 +933,7 @@ struct damon_ctx *damon_new_ctx(void)INIT_LIST_HEAD (&ctx->adaptive_targets);INIT_LIST_HEAD (&ctx->schemes); + ctx->call_controls_obsolete = true;[Severity: High]
This is a pre-existing issue, but does the same fix need to be applied toctx->walk_control_obsolete?Because
walk_control_obsoleteis zero-initialized (false) bykzalloc_obj(), callingdamos_walk()on an unstarted context will bypass the obsolete check:mm/damon/core.c:damos_walk() { if (ctx->walk_control_obsolete) { mutex_unlock(&ctx->walk_control_lock); return -ECANCELED; } ... wait_for_completion(&control->completion); }
Quoted 3 lines
This could lead to an indefinite hang in
wait_for_completion(), since the DAMON worker thread is not running to signal the completion. Shouldctx->walk_control_obsoletealso be initialized to true here?
The issue in damon_call() could happen because it has repeat mode. damos_walk() doesn't have the repeat mode, so I don't think same issue exists here. I will consider doing so in future, for consistency. But not something need to be done now.
Quoted 9 lines
Quoted 4 lines
mm/damon/core.c +0
-0prandom_seed_state (&ctx->rnd_state,get_random_u64 ());return ctx;[ ... ]
Signature · 1 line
Sashiko AI review · https://sashiko.dev/#/patchset/[email protected]?part=2
Thanks,
SJ
SJ Park
maintainerQuoted 5 lines · On Wed, 2 Sep 2026 18:03:29 -0700 SJ Park <[email protected]> wrote:
Callers of
damon_call()should validate thedamon_ctxobject parameter. If it is NULL or never damon_start()-ed object,damon_call()could dereference the NULL pointer or indefinitely hang. Ensuring all callers doing the validation correctly has turned out to be difficult. Handle the corner cases inside the core layer and remove callers' validations.
Sashiko found no blocker for this series. Sashiko sent its findings to damon@ mailing list [1], and I replied to all the comments having findings other than the absence of issues. Please read those for details.
[1] https://lore.kernel.org/damon/
Thanks,
SJ
[...]