media: uvcvideo: avoid variable shadowing in uvc_ctrl_cleanup_fh
Subsystems (2)
PATCH 1/1 media: uvcvideo: avoid variable shadowing in uvc_ctrl_cleanup_fh
This avoids a variable loop shadowing occurring between the local loop iterating through the uvc_entity's controls and the global one going through the pending async controls of the file handle
--- drivers/media/usb/uvc/uvc_ctrl.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-)
drivers/media/usb/uvc/uvc_ctrl.c +1 -2
@@ -3260,7 +3260,6 @@ int uvc_ctrl_init_device(struct uvc_device *dev)void uvc_ctrl_cleanup_fh (struct uvc_fh *handle) {struct uvc_entity *entity; -int i;guard (mutex)(&handle->chain->ctrl_mutex);
@@ -3278,7 +3277,7 @@ void uvc_ctrl_cleanup_fh(struct uvc_fh *handle)if (!WARN_ON (handle->pending_async_ctrls))return ; -for (i =0 ; i < handle->pending_async_ctrls; i++) +for (unsigned int j =0 ; j < handle->pending_async_ctrls; j++)uvc_pm_put (handle->stream->dev); }
Signature · 1 line
2.49.0
Hi Desnes
Quoted 6 lines · Desnes Nunes
This avoids a variable loop shadowing occurring between the local loop iterating through the uvc_entity's controls and the global one going through the pending async controls of the file handle
Fixes 10acb9101355 ("media: uvcvideo: Increase/decrease the PM counter per IOCTL")
If you add a fixes you need to add
Cc: [email protected]
Quoted 18 lines
Signed-off-by Desnes Nunes <[email protected]>--- drivers/media/usb/uvc/uvc_ctrl.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-)drivers/media/usb/uvc/uvc_ctrl.c +0
-1@@ -3260,7 +3260,6 @@ int uvc_ctrl_init_device(struct uvc_device *dev)void uvc_ctrl_cleanup_fh (struct uvc_fh *handle) {struct uvc_entity *entity; -int i;guard (mutex)(&handle->chain->ctrl_mutex);
drivers/media/usb/uvc/uvc_ctrl.c void uvc_ctrl_cleanup_fh(struct uvc_fh *handle) · line 3281
Quoted 4 lines
drivers/media/usb/uvc/uvc_ctrl.c +0
-1@@ -3278,7 +3277,7 @@ void uvc_ctrl_cleanup_fh(struct uvc_fh *handle)if (!WARN_ON (handle->pending_async_ctrls))return ; -for (i =0 ; i < handle->pending_async_ctrls; i++)
nitpick: I would have called the variable i, not j. For me j usually means nested loop. But up to you
I am also not against your first version with a different commit message.
Thanks!
Quoted 8 lines
drivers/media/usb/uvc/uvc_ctrl.c +1
-1+for (unsigned int j =0 ; j < handle->pending_async_ctrls; j++)uvc_pm_put (handle->stream->dev); } --2.49.0
Signature · 1 line
Ricardo Ribalda
Hello Ricardo,
Quoted 5 lines · Ricardo Ribalda
Hi Desnes
Quoted 1 line · Desnes Nunes
Quoted 7 lines
Quoted 5 lines
This avoids a variable loop shadowing occurring between the local loop iterating through the uvc_entity's controls and the global one going through the pending async controls of the file handle
Fixes 10acb9101355 ("media: uvcvideo: Increase/decrease the PM counter per IOCTL")If you add a fixes you need to add
Cc: [email protected]
Thanks for letting me know
Quoted 19 lines
Reviewed-by Ricardo Ribalda <[email protected]>Quoted 17 lines
Signed-off-by Desnes Nunes <[email protected]>--- drivers/media/usb/uvc/uvc_ctrl.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-)drivers/media/usb/uvc/uvc_ctrl.c +0
-1@@ -3260,7 +3260,6 @@ int uvc_ctrl_init_device(struct uvc_device *dev)void uvc_ctrl_cleanup_fh (struct uvc_fh *handle) {struct uvc_entity *entity; -int i;guard (mutex)(&handle->chain->ctrl_mutex);
Quoted 8 lines
drivers/media/usb/uvc/uvc_ctrl.c void uvc_ctrl_cleanup_fh(struct uvc_fh *handle) · line 3281
Quoted 5 lines
drivers/media/usb/uvc/uvc_ctrl.c +0
-1@@ -3278,7 +3277,7 @@ void uvc_ctrl_cleanup_fh(struct uvc_fh *handle)if (!WARN_ON (handle->pending_async_ctrls))return ; -for (i =0 ; i < handle->pending_async_ctrls; i++)nitpick: I would have called the variable i, not j. For me j usually means nested loop. But up to you
Noted - I used a different variable name because I wanted to differentiate the loops.
Quoted 2 lines
I am also not against your first version with a different commit message.
Third time's a charm then!
Will send a v2 with the first version having this commit message.
Thanks for the review Ricardo,
Signature · 1 line
Desnes Nunes
Laurent Pinchart
maintainerQuoted 35 lines · Desnes Nunes
Quoted 9 lines · Ricardo Ribalda
Quoted 6 lines · Desnes Nunes
This avoids a variable loop shadowing occurring between the local loop iterating through the uvc_entity's controls and the global one going through the pending async controls of the file handle
Fixes 10acb9101355 ("media: uvcvideo: Increase/decrease the PM counter per IOCTL")If you add a fixes you need to add
Cc: [email protected]Thanks for letting me know
Quoted 22 lines
Reviewed-by Ricardo Ribalda <[email protected]>Quoted 21 lines
Signed-off-by Desnes Nunes <[email protected]>--- drivers/media/usb/uvc/uvc_ctrl.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-)drivers/media/usb/uvc/uvc_ctrl.c +0
-1@@ -3260,7 +3260,6 @@ int uvc_ctrl_init_device(struct uvc_device *dev)void uvc_ctrl_cleanup_fh (struct uvc_fh *handle) {struct uvc_entity *entity; -int i;guard (mutex)(&handle->chain->ctrl_mutex);@@ -3278,7 +3277,7 @@ void uvc_ctrl_cleanup_fh(struct uvc_fh *handle)if (!WARN_ON (handle->pending_async_ctrls))return ;
Quoted 7 lines
Quoted 4 lines
drivers/media/usb/uvc/uvc_ctrl.c void uvc_ctrl_cleanup_fh(struct uvc_fh *handle) · line 3281
Quoted 1 line
drivers/media/usb/uvc/uvc_ctrl.c +0
-1-for (i =0 ; i < handle->pending_async_ctrls; i++)nitpick: I would have called the variable i, not j. For me j usually means nested loop. But up to you
Noted - I used a different variable name because I wanted to differentiate the loops.
Variable declaration in the loop statement is relatively new in the kernel, so there's no consensus yet (to my knowledge) on clear coding styles, but I would have simply used the same variable name in both loops, with two separate declarations:
drivers/media/usb/uvc/uvc_ctrl.c +1 -2
@@ -3299,7 +3299,6 @@ int uvc_ctrl_init_device(struct uvc_device *dev)void uvc_ctrl_cleanup_fh (struct uvc_fh *handle) {struct uvc_entity *entity; -int i;guard (mutex)(&handle->chain->ctrl_mutex);
@@ -3317,7 +3316,7 @@ void uvc_ctrl_cleanup_fh(struct uvc_fh *handle)if (!WARN_ON (handle->pending_async_ctrls))return ; -for (i =0 ; i < handle->pending_async_ctrls; i++) +for (unsigned int i =0 ; i < handle->pending_async_ctrls; i++)uvc_pm_put (handle->stream->dev); }
Is there a downside to this ?
Quoted 7 lines
Quoted 1 line
I am also not against your first version with a different commit message.
Third time's a charm then!
Will send a v2 with the first version having this commit message.
Thanks for the review Ricardo,
Signature · 3 lines
Regards, Laurent Pinchart
Quoted 71 lines · Laurent Pinchart
Quoted 35 lines · Desnes Nunes
Quoted 9 lines · Ricardo Ribalda
Quoted 6 lines · Desnes Nunes
This avoids a variable loop shadowing occurring between the local loop iterating through the uvc_entity's controls and the global one going through the pending async controls of the file handle
Fixes 10acb9101355 ("media: uvcvideo: Increase/decrease the PM counter per IOCTL")If you add a fixes you need to add
Cc: [email protected]Thanks for letting me know
Quoted 22 lines
Reviewed-by Ricardo Ribalda <[email protected]>Quoted 21 lines
Signed-off-by Desnes Nunes <[email protected]>--- drivers/media/usb/uvc/uvc_ctrl.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-)drivers/media/usb/uvc/uvc_ctrl.c +0
-1@@ -3260,7 +3260,6 @@ int uvc_ctrl_init_device(struct uvc_device *dev)void uvc_ctrl_cleanup_fh (struct uvc_fh *handle) {struct uvc_entity *entity; -int i;guard (mutex)(&handle->chain->ctrl_mutex);@@ -3278,7 +3277,7 @@ void uvc_ctrl_cleanup_fh(struct uvc_fh *handle)if (!WARN_ON (handle->pending_async_ctrls))return ;Quoted 7 lines
Quoted 4 lines
drivers/media/usb/uvc/uvc_ctrl.c void uvc_ctrl_cleanup_fh(struct uvc_fh *handle) · line 3320
Quoted 1 line
drivers/media/usb/uvc/uvc_ctrl.c +0
-1-for (i =0 ; i < handle->pending_async_ctrls; i++)nitpick: I would have called the variable i, not j. For me j usually means nested loop. But up to you
Noted - I used a different variable name because I wanted to differentiate the loops.
Variable declaration in the loop statement is relatively new in the kernel, so there's no consensus yet (to my knowledge) on clear coding styles, but I would have simply used the same variable name in both loops, with two separate declarations:
drivers/media/usb/uvc/uvc_ctrl.c +1
-2@@ -3299,7 +3299,6 @@ int uvc_ctrl_init_device(struct uvc_device *dev)void uvc_ctrl_cleanup_fh (struct uvc_fh *handle) {struct uvc_entity *entity; -int i;guard (mutex)(&handle->chain->ctrl_mutex);@@ -3317,7 +3316,7 @@ void uvc_ctrl_cleanup_fh(struct uvc_fh *handle)if (!WARN_ON (handle->pending_async_ctrls))return ; -for (i =0 ; i < handle->pending_async_ctrls; i++) +for (unsigned int i =0 ; i < handle->pending_async_ctrls; i++)uvc_pm_put (handle->stream->dev); }
Quoted 1 line
Is there a downside to this ?
The toolchain that they are using does not seem to like it, and there is no benefit for having two initialization vs one.
(also makes the lines shorter... but that is just a matter of taste).
I don't really have a preference.
Quoted 13 lines
Quoted 8 lines
Quoted 2 lines
I am also not against your first version with a different commit message.
Third time's a charm then!
Will send a v2 with the first version having this commit message.
Thanks for the review Ricardo,
--
Regards,Laurent Pinchart
Signature · 1 line
Ricardo Ribalda
Laurent Pinchart
maintainerQuoted 68 lines · Ricardo Ribalda
Quoted 66 lines · Laurent Pinchart
Quoted 35 lines · Desnes Nunes
Quoted 9 lines · Ricardo Ribalda
Quoted 6 lines · Desnes Nunes
This avoids a variable loop shadowing occurring between the local loop iterating through the uvc_entity's controls and the global one going through the pending async controls of the file handle
Fixes 10acb9101355 ("media: uvcvideo: Increase/decrease the PM counter per IOCTL")If you add a fixes you need to add
Cc: [email protected]Thanks for letting me know
Quoted 22 lines
Reviewed-by Ricardo Ribalda <[email protected]>Quoted 21 lines
Signed-off-by Desnes Nunes <[email protected]>--- drivers/media/usb/uvc/uvc_ctrl.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-)drivers/media/usb/uvc/uvc_ctrl.c +0
-1@@ -3260,7 +3260,6 @@ int uvc_ctrl_init_device(struct uvc_device *dev)void uvc_ctrl_cleanup_fh (struct uvc_fh *handle) {struct uvc_entity *entity; -int i;guard (mutex)(&handle->chain->ctrl_mutex);@@ -3278,7 +3277,7 @@ void uvc_ctrl_cleanup_fh(struct uvc_fh *handle)if (!WARN_ON (handle->pending_async_ctrls))return ;Quoted 7 lines
Quoted 4 lines
drivers/media/usb/uvc/uvc_ctrl.c void uvc_ctrl_cleanup_fh(struct uvc_fh *handle) · line 3320
Quoted 1 line
drivers/media/usb/uvc/uvc_ctrl.c +0
-1-for (i =0 ; i < handle->pending_async_ctrls; i++)nitpick: I would have called the variable i, not j. For me j usually means nested loop. But up to you
Noted - I used a different variable name because I wanted to differentiate the loops.
Variable declaration in the loop statement is relatively new in the kernel, so there's no consensus yet (to my knowledge) on clear coding styles, but I would have simply used the same variable name in both loops, with two separate declarations:
drivers/media/usb/uvc/uvc_ctrl.c +0
-1@@ -3299,7 +3299,6 @@ int uvc_ctrl_init_device(struct uvc_device *dev)void uvc_ctrl_cleanup_fh (struct uvc_fh *handle) {struct uvc_entity *entity; -int i;guard (mutex)(&handle->chain->ctrl_mutex);@@ -3317,7 +3316,7 @@ void uvc_ctrl_cleanup_fh(struct uvc_fh *handle)if (!WARN_ON (handle->pending_async_ctrls))return ;
Quoted 8 lines
Quoted 6 lines
drivers/media/usb/uvc/uvc_ctrl.c +1
-1-for (i =0 ; i < handle->pending_async_ctrls; i++) +for (unsigned int i =0 ; i < handle->pending_async_ctrls; i++)uvc_pm_put (handle->stream->dev); }Is there a downside to this ?
The toolchain that they are using does not seem to like it,
Well, this is mainline :-)
Quoted 6 lines
and there
is no benefit for having two initialization vs one.(also makes the lines shorter... but that is just a matter of taste).
I don't really have a preference.
Declaring the variable in the loop statement removes the possibility of using it outside of the loop. I think that's an improvement in general, it makes reading the code simpler, and avoids accidental misuse.
Quoted 7 lines
Quoted 7 lines
Quoted 7 lines
Quoted 1 line
I am also not against your first version with a different commit message.
Third time's a charm then!
Will send a v2 with the first version having this commit message.
Thanks for the review Ricardo,
Signature · 3 lines
Regards, Laurent Pinchart