2023-10-28 21:39:59

by SeongJae Park

Subject: [PATCH RESEND 6.5.y] mm/damon/sysfs: check DAMOS regions update progress from before_terminate()

DAMON_SYSFS can receive DAMOS tried regions update request while kdamond
is already out of the main loop and before_terminate callback
(damon_sysfs_before_terminate() in this case) is not yet called. And
damon_sysfs_handle_cmd() can further be finished before the callback is
invoked. Then, damon_sysfs_before_terminate() unlocks damon_sysfs_lock,
which is not locked by anyone. This happens because the callback function
assumes damon_sysfs_cmd_request_callback() should be called before it.
Check if the assumption was true before doing the unlock, to avoid this
problem.

Link: https://lkml.kernel.org/r/[email protected]
Fixes: f1d13cacabe1 ("mm/damon/sysfs: implement DAMOS tried regions update command")
Signed-off-by: SeongJae Park <[email protected]>
Cc: <[email protected]> [6.2.x]
Signed-off-by: Andrew Morton <[email protected]>
(cherry picked from commit 76b7069bcc89dec33f03eb08abee165d0306b754)
---

I sent this patch yesterday, but seems not archived in lore.kernel.org.
Sending again Cc-ing more mailing lists, to ensure it is sent and others can
find it, to avoid any duplicated works.

mm/damon/sysfs.c | 7 +++++--
1 file changed, 5 insertions(+), 2 deletions(-)

diff --git a/mm/damon/sysfs.c b/mm/damon/sysfs.c
index 33e1d5c9cb54..df165820c605 100644
--- a/mm/damon/sysfs.c
+++ b/mm/damon/sysfs.c
@@ -1202,6 +1202,8 @@ static int damon_sysfs_set_targets(struct damon_ctx *ctx,
return 0;
}

+static bool damon_sysfs_schemes_regions_updating;
+
static void damon_sysfs_before_terminate(struct damon_ctx *ctx)
{
struct damon_target *t, *next;
@@ -1209,10 +1211,12 @@ static void damon_sysfs_before_terminate(struct damon_ctx *ctx)

/* damon_sysfs_schemes_update_regions_stop() might not yet called */
kdamond = damon_sysfs_cmd_request.kdamond;
- if (kdamond && damon_sysfs_cmd_request.cmd ==
+ if (kdamond && (damon_sysfs_cmd_request.cmd ==
DAMON_SYSFS_CMD_UPDATE_SCHEMES_TRIED_REGIONS &&
+ damon_sysfs_schemes_regions_updating) &&
ctx == kdamond->damon_ctx) {
damon_sysfs_schemes_update_regions_stop(ctx);
+ damon_sysfs_schemes_regions_updating = false;
mutex_unlock(&damon_sysfs_lock);
}

@@ -1331,7 +1335,6 @@ static int damon_sysfs_commit_input(struct damon_sysfs_kdamond *kdamond)
static int damon_sysfs_cmd_request_callback(struct damon_ctx *c)
{
struct damon_sysfs_kdamond *kdamond;
- static bool damon_sysfs_schemes_regions_updating;
int err = 0;

/* avoid deadlock due to concurrent state_store('off') */
--
2.34.1

2023-10-31 11:16:27

by Greg Kroah-Hartman

Subject: Re: [PATCH RESEND 6.5.y] mm/damon/sysfs: check DAMOS regions update progress from before_terminate()

On Sat, Oct 28, 2023 at 09:39:36PM +0000, SeongJae Park wrote:
> DAMON_SYSFS can receive DAMOS tried regions update request while kdamond
> is already out of the main loop and before_terminate callback
> (damon_sysfs_before_terminate() in this case) is not yet called. And
> damon_sysfs_handle_cmd() can further be finished before the callback is
> invoked. Then, damon_sysfs_before_terminate() unlocks damon_sysfs_lock,
> which is not locked by anyone. This happens because the callback function
> assumes damon_sysfs_cmd_request_callback() should be called before it.
> Check if the assumption was true before doing the unlock, to avoid this
> problem.
>
> Link: https://lkml.kernel.org/r/[email protected]
> Fixes: f1d13cacabe1 ("mm/damon/sysfs: implement DAMOS tried regions update command")
> Signed-off-by: SeongJae Park <[email protected]>
> Cc: <[email protected]> [6.2.x]
> Signed-off-by: Andrew Morton <[email protected]>
> (cherry picked from commit 76b7069bcc89dec33f03eb08abee165d0306b754)
> ---
>
> I sent this patch yesterday, but seems not archived in lore.kernel.org.
> Sending again Cc-ing more mailing lists, to ensure it is sent and others can
> find it, to avoid any duplicated works.

Now queued up, thanks.

greg k-h