This is a note to let you know that I've just added the patch titled
media: i2c: ccs-core: fix pm_runtime_get_sync() usage count
to the 5.12-stable tree which can be found at:
http://www.kernel.org/git/?p=linux/kernel/git/stable/stable-queue.git;a=summary
The filename of the patch is:
media-i2c-ccs-core-fix-pm_runtime_get_sync-usage-count.patch
and it can be found in the queue-5.12 subdirectory.
If you, or anyone else, feels it should not be added to the stable tree,
please let <stable@xxxxxxxxxxxxxxx> know about it.
>From da3a1858c3a37c09446e1470c48352897d59d11b Mon Sep 17 00:00:00 2001
From: Mauro Carvalho Chehab <mchehab+huawei@xxxxxxxxxx>
Date: Fri, 23 Apr 2021 17:19:11 +0200
Subject: media: i2c: ccs-core: fix pm_runtime_get_sync() usage count
From: Mauro Carvalho Chehab <mchehab+huawei@xxxxxxxxxx>
commit da3a1858c3a37c09446e1470c48352897d59d11b upstream.
The pm_runtime_get_sync() internally increments the
dev->power.usage_count without decrementing it, even on errors.
There is a bug at ccs_pm_get_init(): when this function returns
an error, the stream is not started, and RPM usage_count
should not be incremented. However, if the calls to
v4l2_ctrl_handler_setup() return errors, it will be kept
incremented.
At ccs_suspend() the best is to replace it by the new
pm_runtime_resume_and_get(), introduced by:
commit dd8088d5a896 ("PM: runtime: Add pm_runtime_resume_and_get to deal with usage counter")
in order to properly decrement the usage counter automatically,
in the case of errors.
Fixes: 96e3a6b92f23 ("media: smiapp: Avoid maintaining power state information")
Cc: stable@xxxxxxxxxxxxxxx
Acked-by: Sakari Ailus <sakari.ailus@xxxxxxxxxxxxxxx>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@xxxxxxxxxx>
Signed-off-by: Greg Kroah-Hartman <gregkh@xxxxxxxxxxxxxxxxxxx>
---
drivers/media/i2c/ccs/ccs-core.c | 32 ++++++++++++++++++++++----------
1 file changed, 22 insertions(+), 10 deletions(-)
--- a/drivers/media/i2c/ccs/ccs-core.c
+++ b/drivers/media/i2c/ccs/ccs-core.c
@@ -1880,21 +1880,33 @@ static int ccs_pm_get_init(struct ccs_se
struct i2c_client *client = v4l2_get_subdevdata(&sensor->src->sd);
int rval;
+ /*
+ * It can't use pm_runtime_resume_and_get() here, as the driver
+ * relies at the returned value to detect if the device was already
+ * active or not.
+ */
rval = pm_runtime_get_sync(&client->dev);
- if (rval < 0) {
- pm_runtime_put_noidle(&client->dev);
+ if (rval < 0)
+ goto error;
- return rval;
- } else if (!rval) {
- rval = v4l2_ctrl_handler_setup(&sensor->pixel_array->
- ctrl_handler);
- if (rval)
- return rval;
+ /* Device was already active, so don't set controls */
+ if (rval == 1)
+ return 0;
- return v4l2_ctrl_handler_setup(&sensor->src->ctrl_handler);
- }
+ /* Restore V4L2 controls to the previously suspended device */
+ rval = v4l2_ctrl_handler_setup(&sensor->pixel_array->ctrl_handler);
+ if (rval)
+ goto error;
+ rval = v4l2_ctrl_handler_setup(&sensor->src->ctrl_handler);
+ if (rval)
+ goto error;
+
+ /* Keep PM runtime usage_count incremented on success */
return 0;
+error:
+ pm_runtime_put(&client->dev);
+ return rval;
}
static int ccs_set_stream(struct v4l2_subdev *subdev, int enable)
Patches currently in stable-queue which might be from mchehab+huawei@xxxxxxxxxx are
queue-5.12/media-subdev-disallow-ioctl-for-saa6588-davinci.patch
queue-5.12/media-i2c-ccs-core-fix-pm_runtime_get_sync-usage-count.patch
