[RFC PATCH v1] mm: improve call_controls_lock

[RFC PATCH v1] mm: improve call_controls_lock

[Asier Gutierrez]

This is a minor patch set for a call_controls_lock synchronization improvement.

Spinlocks are faster than mutexes, even when the mutex takes the fast
path. Hence, this patch replaces the mutex call_controls_lock with a spinlock.

Initial benchmarking shows the following results


# bpftrace -e 'kprobe:kdamond_call { @start[tid] = nsecs; }
    kretprobe:kdamond_call /@start[tid]/
    { @ns[comm] = hist(nsecs - @start[tid]); delete(@start[tid]); }'

@ns[kdamond.0]:
[16K, 32K)            22 |@@@@@@@@@@@@                                        |
[32K, 64K)            90 |@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@|
[64K, 128K)            0 |                                                    |
[128K, 256K)           1 |                                                    |

@ns[kdamond.0]:
[16K, 32K)            19 |@@@@@@@@                                            |
[32K, 64K)           118 |@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@|
[64K, 128K)            0 |                                                    |
[128K, 256K)           1 |

Signed-off-by: Asier Gutierrez <gutierrez.asier@huawei-partners.com>
---
 include/linux/damon.h |  2 +-
 mm/damon/core.c       | 17 ++++++++---------
 2 files changed, 9 insertions(+), 10 deletions(-)

diff --git a/include/linux/damon.h b/include/linux/damon.h
index 3813373a9200..43665e63a498 100644
--- a/include/linux/damon.h
+++ b/include/linux/damon.h
@@ -801,7 +801,7 @@ struct damon_ctx {
 
 	/* lists of &struct damon_call_control */
 	struct list_head call_controls;
-	struct mutex call_controls_lock;
+	spinlock_t call_controls_lock;
 
 	struct damos_walk_control *walk_control;
 	struct mutex walk_control_lock;
diff --git a/mm/damon/core.c b/mm/damon/core.c
index f9fc0375890a..a929bdf3bc7b 100644
--- a/mm/damon/core.c
+++ b/mm/damon/core.c
@@ -539,7 +539,6 @@ struct damon_ctx *damon_new_ctx(void)
 
 	mutex_init(&ctx->kdamond_lock);
 	INIT_LIST_HEAD(&ctx->call_controls);
-	mutex_init(&ctx->call_controls_lock);
 	mutex_init(&ctx->walk_control_lock);
 
 	ctx->attrs.min_nr_regions = 10;
@@ -1457,9 +1456,9 @@ int damon_call(struct damon_ctx *ctx, struct damon_call_control *control)
 	control->canceled = false;
 	INIT_LIST_HEAD(&control->list);
 
-	mutex_lock(&ctx->call_controls_lock);
+	spin_lock(&ctx->call_controls_lock);
 	list_add_tail(&control->list, &ctx->call_controls);
-	mutex_unlock(&ctx->call_controls_lock);
+	spin_unlock(&ctx->call_controls_lock);
 	if (!damon_is_running(ctx))
 		return -EINVAL;
 	if (control->repeat)
@@ -2549,10 +2548,10 @@ static void kdamond_call(struct damon_ctx *ctx, bool cancel)
 	int ret = 0;
 
 	while (true) {
-		mutex_lock(&ctx->call_controls_lock);
+		spin_lock(&ctx->call_controls_lock);
 		control = list_first_entry_or_null(&ctx->call_controls,
 				struct damon_call_control, list);
-		mutex_unlock(&ctx->call_controls_lock);
+		spin_unlock(&ctx->call_controls_lock);
 		if (!control)
 			break;
 		if (cancel) {
@@ -2561,9 +2560,9 @@ static void kdamond_call(struct damon_ctx *ctx, bool cancel)
 			ret = control->fn(control->data);
 			control->return_code = ret;
 		}
-		mutex_lock(&ctx->call_controls_lock);
+		spin_lock(&ctx->call_controls_lock);
 		list_del(&control->list);
-		mutex_unlock(&ctx->call_controls_lock);
+		spin_unlock(&ctx->call_controls_lock);
 		if (!control->repeat) {
 			complete(&control->completion);
 		} else if (control->canceled && control->dealloc_on_cancel) {
@@ -2577,9 +2576,9 @@ static void kdamond_call(struct damon_ctx *ctx, bool cancel)
 			struct damon_call_control, list);
 	if (!control || cancel)
 		return;
-	mutex_lock(&ctx->call_controls_lock);
+	spin_lock(&ctx->call_controls_lock);
 	list_add_tail(&control->list, &ctx->call_controls);
-	mutex_unlock(&ctx->call_controls_lock);
+	spin_unlock(&ctx->call_controls_lock);
 }
 
 /* Returns negative error code if it's not activated but should return */
-- 
2.43.0

Re: [RFC PATCH v1] mm: improve call_controls_lock

[SeongJae Park]

Hello Asier,


Thank you for sending this patch!

On Mon, 29 Dec 2025 14:55:32 +0000 Asier Gutierrez <gutierrez.asier@huawei-partners.com> wrote:

> This is a minor patch set for a call_controls_lock synchronization improvement.

Please break description lines to not exceed 75 characters per line.

> 
> Spinlocks are faster than mutexes, even when the mutex takes the fast
> path. Hence, this patch replaces the mutex call_controls_lock with a spinlock.

But call_controls_lock is not being used on performance critical part.
Actually, most of DAMON code is not performance critical.  I really appreciate
your patch, but I have to say I don't think this change is really needed now.
Please let me know if I'm missing something.

> 
> Initial benchmarking shows the following results
> 
> 
> # bpftrace -e 'kprobe:kdamond_call { @start[tid] = nsecs; }

Commit log shouldn't start with '#'.  Please consider indenting the above
command and below outputs of it.


Thanks,
SJ

[...]

Re: [RFC PATCH v1] mm: improve call_controls_lock

[Gutierrez Asier]

On 12/29/2025 6:22 PM, SeongJae Park wrote:
> Hello Asier,

Hi SeongJae,

> 
> Thank you for sending this patch!
> 
> On Mon, 29 Dec 2025 14:55:32 +0000 Asier Gutierrez <gutierrez.asier@huawei-partners.com> wrote:
> 
>> This is a minor patch set for a call_controls_lock synchronization improvement.
> 
> Please break description lines to not exceed 75 characters per line.
> 
>>
>> Spinlocks are faster than mutexes, even when the mutex takes the fast
>> path. Hence, this patch replaces the mutex call_controls_lock with a spinlock.
> 
> But call_controls_lock is not being used on performance critical part.
> Actually, most of DAMON code is not performance critical.  I really appreciate
> your patch, but I have to say I don't think this change is really needed now.
> Please let me know if I'm missing something.

I was just reviewing the code and I noticed this. Yes, you are right, the performance 
is not critical.

>>
>> Initial benchmarking shows the following results
>>
>>
>> # bpftrace -e 'kprobe:kdamond_call { @start[tid] = nsecs; }
> 
> Commit log shouldn't start with '#'.  Please consider indenting the above
> command and below outputs of it.
> 
> 
> Thanks,
> SJ
> 
> [...]
> 

Thanks a lot for the review!

-- 
Asier Gutierrez
Huawei

Re: [RFC PATCH v1] mm: improve call_controls_lock

[SeongJae Park]

On Tue, 30 Dec 2025 12:02:31 +0300 Gutierrez Asier <gutierrez.asier@huawei-partners.com> wrote:

> 
> 
> On 12/29/2025 6:22 PM, SeongJae Park wrote:
> > Hello Asier,
> 
> Hi SeongJae,
> 
> > 
> > Thank you for sending this patch!
> > 
> > On Mon, 29 Dec 2025 14:55:32 +0000 Asier Gutierrez <gutierrez.asier@huawei-partners.com> wrote:
> > 
> >> This is a minor patch set for a call_controls_lock synchronization improvement.
> > 
> > Please break description lines to not exceed 75 characters per line.
> > 
> >>
> >> Spinlocks are faster than mutexes, even when the mutex takes the fast
> >> path. Hence, this patch replaces the mutex call_controls_lock with a spinlock.
> > 
> > But call_controls_lock is not being used on performance critical part.
> > Actually, most of DAMON code is not performance critical.  I really appreciate
> > your patch, but I have to say I don't think this change is really needed now.
> > Please let me know if I'm missing something.
> 
> I was just reviewing the code and I noticed this. Yes, you are right, the performance 
> is not critical.

Thank you for flexibly accepting my humble opinion.

> 
> >>
> >> Initial benchmarking shows the following results
> >>
> >>
> >> # bpftrace -e 'kprobe:kdamond_call { @start[tid] = nsecs; }
> > 
> > Commit log shouldn't start with '#'.  Please consider indenting the above
> > command and below outputs of it.
> > 
> > 
> > Thanks,
> > SJ
> > 
> > [...]
> > 
> 
> Thanks a lot for the review!

My pleasure!  Looking forward to your next patch :)


Thanks,
SJ

[...]

Re: [RFC PATCH v1] mm: improve call_controls_lock

[JaeJoon Jung]

On Tue, 30 Dec 2025 at 00:23, SeongJae Park <sj@kernel.org> wrote:
>
> Hello Asier,
>
>
> Thank you for sending this patch!
>
> On Mon, 29 Dec 2025 14:55:32 +0000 Asier Gutierrez <gutierrez.asier@huawei-partners.com> wrote:
>
> > This is a minor patch set for a call_controls_lock synchronization improvement.
>
> Please break description lines to not exceed 75 characters per line.
>
> >
> > Spinlocks are faster than mutexes, even when the mutex takes the fast
> > path. Hence, this patch replaces the mutex call_controls_lock with a spinlock.
>
> But call_controls_lock is not being used on performance critical part.
> Actually, most of DAMON code is not performance critical.  I really appreciate
> your patch, but I have to say I don't think this change is really needed now.
> Please let me know if I'm missing something.

Paradoxically, when it comes to locking, spin_lock is better than
mutex_lock
because "most of DAMON code is not performance critical."

DAMON code only accesses the ctx belonging to kdamond itself. For
example:
kdamond.0 --> ctx.0
kdamond.1 --> ctx.1
kdamond.2 --> ctx.2
kdamond.# --> ctx.#

There is no cross-approach as shown below:
kdamond.0 --> ctx.1
kdamond.1 --> ctx.2
kdamond.2 --> ctx.0

Only the data belonging to kdamond needs to be resolved for concurrent access.
most DAMON code needs to lock/unlock briefly when add/del linked
lists,
so spin_lock is effective.  If you handle it with a mutex, it becomes
more
complicated because the rescheduling occurs as a context switch occurs
inside the kernel.  Moreover, since the call_controls_lock that is
currently
being raised as a problem only occurs in two places, the kdamon_call()
loop
and the damon_call() function, it is effective to handle it with a
spin_lock
as shown below.

@@ -1502,14 +1501,15 @@ int damon_call(struct damon_ctx *ctx, struct
damon_call_control *control)
        control->canceled = false;
        INIT_LIST_HEAD(&control->list);

-       mutex_lock(&ctx->call_controls_lock);
+       spin_lock(&ctx->call_controls_lock);
+       /* damon_is_running */
        if (ctx->kdamond) {
                list_add_tail(&control->list, &ctx->call_controls);
        } else {
-               mutex_unlock(&ctx->call_controls_lock);
+               spin_unlock(&ctx->call_controls_lock);
                return -EINVAL;
        }
-       mutex_unlock(&ctx->call_controls_lock);
+       spin_unlock(&ctx->call_controls_lock);

        if (control->repeat)
                return 0;

>
> >
> > Initial benchmarking shows the following results
> >
> >
> > # bpftrace -e 'kprobe:kdamond_call { @start[tid] = nsecs; }
>
> Commit log shouldn't start with '#'.  Please consider indenting the above
> command and below outputs of it.
>
>
> Thanks,
> SJ
>
> [...]
>

Re: [RFC PATCH v1] mm: improve call_controls_lock

[SeongJae Park]

On Wed, 31 Dec 2025 11:15:00 +0900 JaeJoon Jung <rgbi3307@gmail.com> wrote:

> On Tue, 30 Dec 2025 at 00:23, SeongJae Park <sj@kernel.org> wrote:
> >
> > Hello Asier,
> >
> >
> > Thank you for sending this patch!
> >
> > On Mon, 29 Dec 2025 14:55:32 +0000 Asier Gutierrez <gutierrez.asier@huawei-partners.com> wrote:
> >
> > > This is a minor patch set for a call_controls_lock synchronization improvement.
> >
> > Please break description lines to not exceed 75 characters per line.
> >
> > >
> > > Spinlocks are faster than mutexes, even when the mutex takes the fast
> > > path. Hence, this patch replaces the mutex call_controls_lock with a spinlock.
> >
> > But call_controls_lock is not being used on performance critical part.
> > Actually, most of DAMON code is not performance critical.  I really appreciate
> > your patch, but I have to say I don't think this change is really needed now.
> > Please let me know if I'm missing something.
> 
> Paradoxically, when it comes to locking, spin_lock is better than
> mutex_lock
> because "most of DAMON code is not performance critical."
> 
> DAMON code only accesses the ctx belonging to kdamond itself. For
> example:
> kdamond.0 --> ctx.0
> kdamond.1 --> ctx.1
> kdamond.2 --> ctx.2
> kdamond.# --> ctx.#
> 
> There is no cross-approach as shown below:
> kdamond.0 --> ctx.1
> kdamond.1 --> ctx.2
> kdamond.2 --> ctx.0
> 
> Only the data belonging to kdamond needs to be resolved for concurrent access.
> most DAMON code needs to lock/unlock briefly when add/del linked
> lists,
> so spin_lock is effective.

I don't disagree this.  Both spinlock and mutex effectively work for DAMON's
locking usages.

> If you handle it with a mutex, it becomes
> more
> complicated because the rescheduling occurs as a context switch occurs
> inside the kernel.

Can you please elaborate what kind of complexities you are saying about?
Adding some examples would be nice.

> Moreover, since the call_controls_lock that is
> currently
> being raised as a problem only occurs in two places, the kdamon_call()
> loop
> and the damon_call() function, it is effective to handle it with a
> spin_lock
> as shown below.
> 
> @@ -1502,14 +1501,15 @@ int damon_call(struct damon_ctx *ctx, struct
> damon_call_control *control)
>         control->canceled = false;
>         INIT_LIST_HEAD(&control->list);
> 
> -       mutex_lock(&ctx->call_controls_lock);
> +       spin_lock(&ctx->call_controls_lock);
> +       /* damon_is_running */
>         if (ctx->kdamond) {
>                 list_add_tail(&control->list, &ctx->call_controls);
>         } else {
> -               mutex_unlock(&ctx->call_controls_lock);
> +               spin_unlock(&ctx->call_controls_lock);
>                 return -EINVAL;
>         }
> -       mutex_unlock(&ctx->call_controls_lock);
> +       spin_unlock(&ctx->call_controls_lock);
> 
>         if (control->repeat)
>                 return 0;

Are you saying the above diff can fix the damon_call() use-after-free bug [1]?
Can you please elaborate why you think so?

[1] https://lore.kernel.org/20251231012315.75835-1-sj@kernel.org


Thanks,
SJ

[...]

Re: [RFC PATCH v1] mm: improve call_controls_lock

[JaeJoon Jung]

On Wed, 31 Dec 2025 at 13:59, SeongJae Park <sj@kernel.org> wrote:
>
> On Wed, 31 Dec 2025 11:15:00 +0900 JaeJoon Jung <rgbi3307@gmail.com> wrote:
>
> > On Tue, 30 Dec 2025 at 00:23, SeongJae Park <sj@kernel.org> wrote:
> > >
> > > Hello Asier,
> > >
> > >
> > > Thank you for sending this patch!
> > >
> > > On Mon, 29 Dec 2025 14:55:32 +0000 Asier Gutierrez <gutierrez.asier@huawei-partners.com> wrote:
> > >
> > > > This is a minor patch set for a call_controls_lock synchronization improvement.
> > >
> > > Please break description lines to not exceed 75 characters per line.
> > >
> > > >
> > > > Spinlocks are faster than mutexes, even when the mutex takes the fast
> > > > path. Hence, this patch replaces the mutex call_controls_lock with a spinlock.
> > >
> > > But call_controls_lock is not being used on performance critical part.
> > > Actually, most of DAMON code is not performance critical.  I really appreciate
> > > your patch, but I have to say I don't think this change is really needed now.
> > > Please let me know if I'm missing something.
> >
> > Paradoxically, when it comes to locking, spin_lock is better than
> > mutex_lock
> > because "most of DAMON code is not performance critical."
> >
> > DAMON code only accesses the ctx belonging to kdamond itself. For
> > example:
> > kdamond.0 --> ctx.0
> > kdamond.1 --> ctx.1
> > kdamond.2 --> ctx.2
> > kdamond.# --> ctx.#
> >
> > There is no cross-approach as shown below:
> > kdamond.0 --> ctx.1
> > kdamond.1 --> ctx.2
> > kdamond.2 --> ctx.0
> >
> > Only the data belonging to kdamond needs to be resolved for concurrent access.
> > most DAMON code needs to lock/unlock briefly when add/del linked
> > lists,
> > so spin_lock is effective.
>
> I don't disagree this.  Both spinlock and mutex effectively work for DAMON's
> locking usages.
>
> > If you handle it with a mutex, it becomes
> > more
> > complicated because the rescheduling occurs as a context switch occurs
> > inside the kernel.
>
> Can you please elaborate what kind of complexities you are saying about?
> Adding some examples would be nice.
>
> > Moreover, since the call_controls_lock that is
> > currently
> > being raised as a problem only occurs in two places, the kdamon_call()
> > loop
> > and the damon_call() function, it is effective to handle it with a
> > spin_lock
> > as shown below.
> >
> > @@ -1502,14 +1501,15 @@ int damon_call(struct damon_ctx *ctx, struct
> > damon_call_control *control)
> >         control->canceled = false;
> >         INIT_LIST_HEAD(&control->list);
> >
> > -       mutex_lock(&ctx->call_controls_lock);
> > +       spin_lock(&ctx->call_controls_lock);
> > +       /* damon_is_running */
> >         if (ctx->kdamond) {
> >                 list_add_tail(&control->list, &ctx->call_controls);
> >         } else {
> > -               mutex_unlock(&ctx->call_controls_lock);
> > +               spin_unlock(&ctx->call_controls_lock);
> >                 return -EINVAL;
> >         }
> > -       mutex_unlock(&ctx->call_controls_lock);
> > +       spin_unlock(&ctx->call_controls_lock);
> >
> >         if (control->repeat)
> >                 return 0;
>
> Are you saying the above diff can fix the damon_call() use-after-free bug [1]?
> Can you please elaborate why you think so?
>
> [1] https://lore.kernel.org/20251231012315.75835-1-sj@kernel.org
>

The above code works fine with spin_lock.  However, when booting the kernel,
the spin_lock call trace from damon_call() is output as follows:
If you have any experience with the following, please share it.

[    0.834450] Call Trace:
[    0.834456] [<ffffffff8001b376>] dump_backtrace+0x1c/0x24
[    0.834471] [<ffffffff800024e0>] show_stack+0x28/0x34
[    0.834480] [<ffffffff80014f4c>] dump_stack_lvl+0x48/0x66
[    0.834493] [<ffffffff80014f7e>] dump_stack+0x14/0x1c
[    0.834503] [<ffffffff800032c6>] spin_dump+0x62/0x6e
[    0.834511] [<ffffffff80087376>] do_raw_spin_lock+0xd0/0x128
[    0.834523] [<ffffffff80de9378>] _raw_spin_lock+0x1a/0x22
[    0.834538] [<ffffffff80255c0c>] damon_call+0x38/0x100
[    0.834548] [<ffffffff8025f022>] damon_stat_start+0x10e/0x168
[    0.834558] [<ffffffff80e21ab4>] damon_stat_init+0x2a/0x44
[    0.834568] [<ffffffff800157c0>] do_one_initcall+0x40/0x202
[    0.834579] [<ffffffff80e015f6>] kernel_init_freeable+0x1fc/0x27e
[    0.834588] [<ffffffff80de0a9e>] kernel_init+0x1e/0x13c
[    0.834599] [<ffffffff8001716a>] ret_from_fork_kernel+0x10/0xf8
[    0.834607] [<ffffffff80deab22>] ret_from_fork_kernel_asm+0x16/0x18
[    0.943407] NFS: Registering the id_resolver key type
[    0.948996] Key type id_resolver registered
[    0.953614] Key type id_legacy registered

Thanks,
JaeJoon

>
> Thanks,
> SJ
>
> [...]

Re: [RFC PATCH v1] mm: improve call_controls_lock

[JaeJoon Jung]

On Wed, 31 Dec 2025 at 15:10, JaeJoon Jung <rgbi3307@gmail.com> wrote:
>
> On Wed, 31 Dec 2025 at 13:59, SeongJae Park <sj@kernel.org> wrote:
> >
> > On Wed, 31 Dec 2025 11:15:00 +0900 JaeJoon Jung <rgbi3307@gmail.com> wrote:
> >
> > > On Tue, 30 Dec 2025 at 00:23, SeongJae Park <sj@kernel.org> wrote:
> > > >
> > > > Hello Asier,
> > > >
> > > >
> > > > Thank you for sending this patch!
> > > >
> > > > On Mon, 29 Dec 2025 14:55:32 +0000 Asier Gutierrez <gutierrez.asier@huawei-partners.com> wrote:
> > > >
> > > > > This is a minor patch set for a call_controls_lock synchronization improvement.
> > > >
> > > > Please break description lines to not exceed 75 characters per line.
> > > >
> > > > >
> > > > > Spinlocks are faster than mutexes, even when the mutex takes the fast
> > > > > path. Hence, this patch replaces the mutex call_controls_lock with a spinlock.
> > > >
> > > > But call_controls_lock is not being used on performance critical part.
> > > > Actually, most of DAMON code is not performance critical.  I really appreciate
> > > > your patch, but I have to say I don't think this change is really needed now.
> > > > Please let me know if I'm missing something.
> > >
> > > Paradoxically, when it comes to locking, spin_lock is better than
> > > mutex_lock
> > > because "most of DAMON code is not performance critical."
> > >
> > > DAMON code only accesses the ctx belonging to kdamond itself. For
> > > example:
> > > kdamond.0 --> ctx.0
> > > kdamond.1 --> ctx.1
> > > kdamond.2 --> ctx.2
> > > kdamond.# --> ctx.#
> > >
> > > There is no cross-approach as shown below:
> > > kdamond.0 --> ctx.1
> > > kdamond.1 --> ctx.2
> > > kdamond.2 --> ctx.0
> > >
> > > Only the data belonging to kdamond needs to be resolved for concurrent access.
> > > most DAMON code needs to lock/unlock briefly when add/del linked
> > > lists,
> > > so spin_lock is effective.
> >
> > I don't disagree this.  Both spinlock and mutex effectively work for DAMON's
> > locking usages.
> >
> > > If you handle it with a mutex, it becomes
> > > more
> > > complicated because the rescheduling occurs as a context switch occurs
> > > inside the kernel.
> >
> > Can you please elaborate what kind of complexities you are saying about?
> > Adding some examples would be nice.
> >
> > > Moreover, since the call_controls_lock that is
> > > currently
> > > being raised as a problem only occurs in two places, the kdamon_call()
> > > loop
> > > and the damon_call() function, it is effective to handle it with a
> > > spin_lock
> > > as shown below.
> > >
> > > @@ -1502,14 +1501,15 @@ int damon_call(struct damon_ctx *ctx, struct
> > > damon_call_control *control)
> > >         control->canceled = false;
> > >         INIT_LIST_HEAD(&control->list);
> > >
> > > -       mutex_lock(&ctx->call_controls_lock);
> > > +       spin_lock(&ctx->call_controls_lock);
> > > +       /* damon_is_running */
> > >         if (ctx->kdamond) {
> > >                 list_add_tail(&control->list, &ctx->call_controls);
> > >         } else {
> > > -               mutex_unlock(&ctx->call_controls_lock);
> > > +               spin_unlock(&ctx->call_controls_lock);
> > >                 return -EINVAL;
> > >         }
> > > -       mutex_unlock(&ctx->call_controls_lock);
> > > +       spin_unlock(&ctx->call_controls_lock);
> > >
> > >         if (control->repeat)
> > >                 return 0;
> >
> > Are you saying the above diff can fix the damon_call() use-after-free bug [1]?
> > Can you please elaborate why you think so?
> >
> > [1] https://lore.kernel.org/20251231012315.75835-1-sj@kernel.org
> >
>
> The above code works fine with spin_lock.  However, when booting the kernel,
> the spin_lock call trace from damon_call() is output as follows:
> If you have any experience with the following, please share it.
>
> [    0.834450] Call Trace:
> [    0.834456] [<ffffffff8001b376>] dump_backtrace+0x1c/0x24
> [    0.834471] [<ffffffff800024e0>] show_stack+0x28/0x34
> [    0.834480] [<ffffffff80014f4c>] dump_stack_lvl+0x48/0x66
> [    0.834493] [<ffffffff80014f7e>] dump_stack+0x14/0x1c
> [    0.834503] [<ffffffff800032c6>] spin_dump+0x62/0x6e
> [    0.834511] [<ffffffff80087376>] do_raw_spin_lock+0xd0/0x128
> [    0.834523] [<ffffffff80de9378>] _raw_spin_lock+0x1a/0x22
> [    0.834538] [<ffffffff80255c0c>] damon_call+0x38/0x100
> [    0.834548] [<ffffffff8025f022>] damon_stat_start+0x10e/0x168
> [    0.834558] [<ffffffff80e21ab4>] damon_stat_init+0x2a/0x44
> [    0.834568] [<ffffffff800157c0>] do_one_initcall+0x40/0x202
> [    0.834579] [<ffffffff80e015f6>] kernel_init_freeable+0x1fc/0x27e
> [    0.834588] [<ffffffff80de0a9e>] kernel_init+0x1e/0x13c
> [    0.834599] [<ffffffff8001716a>] ret_from_fork_kernel+0x10/0xf8
> [    0.834607] [<ffffffff80deab22>] ret_from_fork_kernel_asm+0x16/0x18
> [    0.943407] NFS: Registering the id_resolver key type
> [    0.948996] Key type id_resolver registered
> [    0.953614] Key type id_legacy registered

The above occurred because spin_lock_init() was not performed.  The problem
is that spin_lock_init() was not added while deleting mutex_init().
Please refer to the contents below.

@@ -539,6 +539,7 @@ struct damon_ctx *damon_new_ctx(void)

        mutex_init(&ctx->kdamond_lock);
        INIT_LIST_HEAD(&ctx->call_controls);
-       mutex_init(&ctx->call_controls_lock);
+       spin_lock_init(&ctx->call_controls_lock);
        mutex_init(&ctx->walk_control_lock);

        ctx->attrs.min_nr_regions = 10;

>
> Thanks,
> JaeJoon
>
> >
> > Thanks,
> > SJ
> >
> > [...]

Re: [RFC PATCH v1] mm: improve call_controls_lock

[SeongJae Park]

On Wed, 31 Dec 2025 15:10:12 +0900 JaeJoon Jung <rgbi3307@gmail.com> wrote:

> On Wed, 31 Dec 2025 at 13:59, SeongJae Park <sj@kernel.org> wrote:
> >
> > On Wed, 31 Dec 2025 11:15:00 +0900 JaeJoon Jung <rgbi3307@gmail.com> wrote:
> >
> > > On Tue, 30 Dec 2025 at 00:23, SeongJae Park <sj@kernel.org> wrote:
> > > >
> > > > Hello Asier,
> > > >
> > > >
> > > > Thank you for sending this patch!
> > > >
> > > > On Mon, 29 Dec 2025 14:55:32 +0000 Asier Gutierrez <gutierrez.asier@huawei-partners.com> wrote:
> > > >
> > > > > This is a minor patch set for a call_controls_lock synchronization improvement.
> > > >
> > > > Please break description lines to not exceed 75 characters per line.
> > > >
> > > > >
> > > > > Spinlocks are faster than mutexes, even when the mutex takes the fast
> > > > > path. Hence, this patch replaces the mutex call_controls_lock with a spinlock.
> > > >
> > > > But call_controls_lock is not being used on performance critical part.
> > > > Actually, most of DAMON code is not performance critical.  I really appreciate
> > > > your patch, but I have to say I don't think this change is really needed now.
> > > > Please let me know if I'm missing something.
> > >
> > > Paradoxically, when it comes to locking, spin_lock is better than
> > > mutex_lock
> > > because "most of DAMON code is not performance critical."
> > >
> > > DAMON code only accesses the ctx belonging to kdamond itself. For
> > > example:
> > > kdamond.0 --> ctx.0
> > > kdamond.1 --> ctx.1
> > > kdamond.2 --> ctx.2
> > > kdamond.# --> ctx.#
> > >
> > > There is no cross-approach as shown below:
> > > kdamond.0 --> ctx.1
> > > kdamond.1 --> ctx.2
> > > kdamond.2 --> ctx.0
> > >
> > > Only the data belonging to kdamond needs to be resolved for concurrent access.
> > > most DAMON code needs to lock/unlock briefly when add/del linked
> > > lists,
> > > so spin_lock is effective.
> >
> > I don't disagree this.  Both spinlock and mutex effectively work for DAMON's
> > locking usages.
> >
> > > If you handle it with a mutex, it becomes
> > > more
> > > complicated because the rescheduling occurs as a context switch occurs
> > > inside the kernel.
> >
> > Can you please elaborate what kind of complexities you are saying about?
> > Adding some examples would be nice.
> >
> > > Moreover, since the call_controls_lock that is
> > > currently
> > > being raised as a problem only occurs in two places, the kdamon_call()
> > > loop
> > > and the damon_call() function, it is effective to handle it with a
> > > spin_lock
> > > as shown below.
> > >
> > > @@ -1502,14 +1501,15 @@ int damon_call(struct damon_ctx *ctx, struct
> > > damon_call_control *control)
> > >         control->canceled = false;
> > >         INIT_LIST_HEAD(&control->list);
> > >
> > > -       mutex_lock(&ctx->call_controls_lock);
> > > +       spin_lock(&ctx->call_controls_lock);
> > > +       /* damon_is_running */
> > >         if (ctx->kdamond) {
> > >                 list_add_tail(&control->list, &ctx->call_controls);
> > >         } else {
> > > -               mutex_unlock(&ctx->call_controls_lock);
> > > +               spin_unlock(&ctx->call_controls_lock);
> > >                 return -EINVAL;
> > >         }
> > > -       mutex_unlock(&ctx->call_controls_lock);
> > > +       spin_unlock(&ctx->call_controls_lock);
> > >
> > >         if (control->repeat)
> > >                 return 0;
> >
> > Are you saying the above diff can fix the damon_call() use-after-free bug [1]?
> > Can you please elaborate why you think so?
> >
> > [1] https://lore.kernel.org/20251231012315.75835-1-sj@kernel.org
> >
> 
> The above code works fine with spin_lock.  However, when booting the kernel,
> the spin_lock call trace from damon_call() is output as follows:
> If you have any experience with the following, please share it.

Can you please reply to my questions above, first?


Thanks,
SJ

[...]