Skip to contentSkip to conversation

iio: Use an early return in iio_device_alloc to simplify code.

PATCH In review [email protected] 2 participants 4 messages Apr 19, 2020 – Apr 26, 2020 IIO

lore
Share & tools
lore

[email protected]

author maintainer
#0 lore raw

PATCH 1/1 iio: Use an early return in iio_device_alloc to simplify code.

From: Jonathan Cameron <[email protected]>

Noticed whilst reviewing Alexandru's patch to the same function. If we simply flip the logic and return NULL immediately after memory allocation failure we reduce the indent of the following block and end up with more 'idiomatic' kernel code.

Cc Alexandru Ardelean <[email protected]>
---
 drivers/iio/industrialio-core.c | 38 ++++++++++++++++-----------------
 1 file changed, 19 insertions(+), 19 deletions(-)
drivers/iio/industrialio-core.c +19 -19
index f4daf19f2a3b..96f6dacb206d 100644
@@ -1504,27 +1504,27 @@ struct iio_dev *iio_device_alloc(int sizeof_priv)
 	alloc_size += IIO_ALIGN - 1;
 
 	dev = kzalloc(alloc_size, GFP_KERNEL);
+	if (!dev)
+		return NULL;
 
-	if (dev) {
-		dev->dev.groups = dev->groups;
-		dev->dev.type = &iio_device_type;
-		dev->dev.bus = &iio_bus_type;
-		device_initialize(&dev->dev);
-		dev_set_drvdata(&dev->dev, (void *)dev);
-		mutex_init(&dev->mlock);
-		mutex_init(&dev->info_exist_lock);
-		INIT_LIST_HEAD(&dev->channel_attr_list);
-
-		dev->id = ida_simple_get(&iio_ida, 0, 0, GFP_KERNEL);
-		if (dev->id < 0) {
-			/* cannot use a dev_err as the name isn't available */
-			pr_err("failed to get device id\n");
-			kfree(dev);
-			return NULL;
-		}
-		dev_set_name(&dev->dev, "iio:device%d", dev->id);
-		INIT_LIST_HEAD(&dev->buffer_list);
+	dev->dev.groups = dev->groups;
+	dev->dev.type = &iio_device_type;
+	dev->dev.bus = &iio_bus_type;
+	device_initialize(&dev->dev);
+	dev_set_drvdata(&dev->dev, (void *)dev);
+	mutex_init(&dev->mlock);
+	mutex_init(&dev->info_exist_lock);
+	INIT_LIST_HEAD(&dev->channel_attr_list);
+
+	dev->id = ida_simple_get(&iio_ida, 0, 0, GFP_KERNEL);
+	if (dev->id < 0) {
+		/* cannot use a dev_err as the name isn't available */
+		pr_err("failed to get device id\n");
+		kfree(dev);
+		return NULL;
 	}
+	dev_set_name(&dev->dev, "iio:device%d", dev->id);
+	INIT_LIST_HEAD(&dev->buffer_list);
 
 	return dev;
 }
Signature · 1 line
2.26.1
Reviewed-by on v1 1/1
#1 lore raw
Quoted 9 lines · On Sun, 2020-04-19 at 16:13 +0100, [email protected] wrote:

[External]

From: Jonathan Cameron <[email protected]>

Noticed whilst reviewing Alexandru's patch to the same function. If we simply flip the logic and return NULL immediately after memory allocation failure we reduce the indent of the following block and end up with more 'idiomatic' kernel code.

I also was tempted to do it, but was tempted [a bit more] by the initial change that I goofed.

A few thoughts on this [can be ignored].
But, since doing this change, should 'dev' be renamed to 'indio_dev'? It shouldn't be a lot more code than the current change [I hope]. When looking through IIO core, I got a minor/slight confusion on this alloc code about the name 'dev' [which is of type 'struct iio_dev' vs 'struct device', as is more customary].

If 'dev' was chosen to fit within any 80 col-width limit, that limit should be less likely to hit now.

1 more inline.

Well, even with/without these changes.

Quoted 41 lines
Cc Alexandru Ardelean <[email protected]>
---
 drivers/iio/industrialio-core.c | 38 ++++++++++++++++-----------------
 1 file changed, 19 insertions(+), 19 deletions(-)
drivers/iio/industrialio-core.c +7 -19
index f4daf19f2a3b..96f6dacb206d 100644
@@ -1504,27 +1504,27 @@ struct iio_dev *iio_device_alloc(int sizeof_priv)
 	alloc_size += IIO_ALIGN - 1;
 
 	dev = kzalloc(alloc_size, GFP_KERNEL);
+	if (!dev)
+		return NULL;
 
-	if (dev) {
-		dev->dev.groups = dev->groups;
-		dev->dev.type = &iio_device_type;
-		dev->dev.bus = &iio_bus_type;
-		device_initialize(&dev->dev);
-		dev_set_drvdata(&dev->dev, (void *)dev);
-		mutex_init(&dev->mlock);
-		mutex_init(&dev->info_exist_lock);
-		INIT_LIST_HEAD(&dev->channel_attr_list);
-
-		dev->id = ida_simple_get(&iio_ida, 0, 0, GFP_KERNEL);
-		if (dev->id < 0) {
-			/* cannot use a dev_err as the name isn't available */
-			pr_err("failed to get device id\n");
-			kfree(dev);
-			return NULL;
-		}
-		dev_set_name(&dev->dev, "iio:device%d", dev->id);
-		INIT_LIST_HEAD(&dev->buffer_list);
+	dev->dev.groups = dev->groups;
+	dev->dev.type = &iio_device_type;
+	dev->dev.bus = &iio_bus_type;
+	device_initialize(&dev->dev);
+	dev_set_drvdata(&dev->dev, (void *)dev);

drivers/iio/industrialio-core.c struct iio_dev *iio_device_alloc(int sizeof_priv) · line 1524

Quoted 10 lines
drivers/iio/industrialio-core.c +10 -0
+	mutex_init(&dev->mlock);
+	mutex_init(&dev->info_exist_lock);
+	INIT_LIST_HEAD(&dev->channel_attr_list);
+
+	dev->id = ida_simple_get(&iio_ida, 0, 0, GFP_KERNEL);
+	if (dev->id < 0) {
+		/* cannot use a dev_err as the name isn't available */
+		pr_err("failed to get device id\n");
+		kfree(dev);
+		return NULL;

would it be too much for this patch to move this right after the kzalloc()? no strong opinion from my side to do it or not;
but it does save some init cycles, and compresses this init block a bit;

Quoted 6 lines
drivers/iio/industrialio-core.c +2 -0
 	}
+	dev_set_name(&dev->dev, "iio:device%d", dev->id);
+	INIT_LIST_HEAD(&dev->buffer_list);
 
 	return dev;
 }
author maintainer
#2 lore raw
Quoted 21 lines · Alexandru
Quoted 9 lines · On Sun, 2020-04-19 at 16:13 +0100, [email protected] wrote:

[External]

From: Jonathan Cameron <[email protected]>

Noticed whilst reviewing Alexandru's patch to the same function. If we simply flip the logic and return NULL immediately after memory allocation failure we reduce the indent of the following block and end up with more 'idiomatic' kernel code.

I also was tempted to do it, but was tempted [a bit more] by the initial change that I goofed.

A few thoughts on this [can be ignored].
But, since doing this change, should 'dev' be renamed to 'indio_dev'? It shouldn't be a lot more code than the current change [I hope]. When looking through IIO core, I got a minor/slight confusion on this alloc code about the name 'dev' [which is of type 'struct iio_dev' vs 'struct device', as is more customary].

Quoted 2 lines

If 'dev' was chosen to fit within any 80 col-width limit, that limit should be less likely to hit now.

A different type of cleanup, so I think worth a separate patch (even though it's messing with the same block of code.)

Got to keep to the rules I pester everyone else into following :)

So I'll apply this as is and might get the dev->indio_dev one out after I've caught up with rest of email queue.

Thanks,

Jonathan

Quoted 54 lines

1 more inline.

Well, even with/without these changes.

Quoted 47 lines
Cc Alexandru Ardelean <[email protected]>
---
 drivers/iio/industrialio-core.c | 38 ++++++++++++++++-----------------
 1 file changed, 19 insertions(+), 19 deletions(-)
drivers/iio/industrialio-core.c +13 -19
index f4daf19f2a3b..96f6dacb206d 100644
@@ -1504,27 +1504,27 @@ struct iio_dev *iio_device_alloc(int sizeof_priv)
 	alloc_size += IIO_ALIGN - 1;
 
 	dev = kzalloc(alloc_size, GFP_KERNEL);
+	if (!dev)
+		return NULL;
 
-	if (dev) {
-		dev->dev.groups = dev->groups;
-		dev->dev.type = &iio_device_type;
-		dev->dev.bus = &iio_bus_type;
-		device_initialize(&dev->dev);
-		dev_set_drvdata(&dev->dev, (void *)dev);
-		mutex_init(&dev->mlock);
-		mutex_init(&dev->info_exist_lock);
-		INIT_LIST_HEAD(&dev->channel_attr_list);
-
-		dev->id = ida_simple_get(&iio_ida, 0, 0, GFP_KERNEL);
-		if (dev->id < 0) {
-			/* cannot use a dev_err as the name isn't available */
-			pr_err("failed to get device id\n");
-			kfree(dev);
-			return NULL;
-		}
-		dev_set_name(&dev->dev, "iio:device%d", dev->id);
-		INIT_LIST_HEAD(&dev->buffer_list);
+	dev->dev.groups = dev->groups;
+	dev->dev.type = &iio_device_type;
+	dev->dev.bus = &iio_bus_type;
+	device_initialize(&dev->dev);
+	dev_set_drvdata(&dev->dev, (void *)dev);
+	mutex_init(&dev->mlock);
+	mutex_init(&dev->info_exist_lock);
+	INIT_LIST_HEAD(&dev->channel_attr_list);
+
+	dev->id = ida_simple_get(&iio_ida, 0, 0, GFP_KERNEL);
+	if (dev->id < 0) {
Quoted 8 lines

drivers/iio/industrialio-core.c struct iio_dev *iio_device_alloc(int sizeof_priv) · line 1524

Quoted 4 lines
drivers/iio/industrialio-core.c +4 -0
+		/* cannot use a dev_err as the name isn't available */
+		pr_err("failed to get device id\n");
+		kfree(dev);
+		return NULL;  

would it be too much for this patch to move this right after the kzalloc()? no strong opinion from my side to do it or not;
but it does save some init cycles, and compresses this init block a bit;

It doesn't really save any cycles because the chance of failure of ID allocation is negligible... Now I'd agree with you if writing from scratch, but as a tidy up patch, it's good to keep things really simple.

Quoted 7 lines
Quoted 7 lines
drivers/iio/industrialio-core.c +2 -0
 	}
+	dev_set_name(&dev->dev, "iio:device%d", dev->id);
+	INIT_LIST_HEAD(&dev->buffer_list);
 
 	return dev;
 }  
#3 lore raw
Quoted 27 lines · Jonathan Cameron
Quoted 24 lines · Alexandru
Quoted 9 lines · On Sun, 2020-04-19 at 16:13 +0100, [email protected] wrote:

[External]

From: Jonathan Cameron <[email protected]>

Noticed whilst reviewing Alexandru's patch to the same function. If we simply flip the logic and return NULL immediately after memory allocation failure we reduce the indent of the following block and end up with more 'idiomatic' kernel code.

I also was tempted to do it, but was tempted [a bit more] by the initial change
that I goofed.

A few thoughts on this [can be ignored].
But, since doing this change, should 'dev' be renamed to 'indio_dev'? It shouldn't be a lot more code than the current change [I hope]. When looking through IIO core, I got a minor/slight confusion on this alloc code
about the name 'dev' [which is of type 'struct iio_dev' vs 'struct device', as
is more customary].

Quoted 8 lines
Quoted 3 lines

If 'dev' was chosen to fit within any 80 col-width limit, that limit should be
less likely to hit now.

A different type of cleanup, so I think worth a separate patch (even though it's messing with the same block of code.)

Got to keep to the rules I pester everyone else into following :)

Fair enough.
I was assuming there would be a preference to keep the patch semantic as simple as possible.

I thought I'd point them out, since I also looked over this bit a while back.

Quoted 83 lines

So I'll apply this as is and might get the dev->indio_dev one out after I've caught up with rest of email queue.

Thanks,

Jonathan

Quoted 54 lines

1 more inline.

Well, even with/without these changes.

Quoted 48 lines
Cc Alexandru Ardelean <[email protected]>
---
 drivers/iio/industrialio-core.c | 38 ++++++++++++++++-----------------
 1 file changed, 19 insertions(+), 19 deletions(-)
drivers/iio/industrialio- renamed from drivers/iio/industrialio-core.c +0 -0

core.c
index f4daf19f2a3b..96f6dacb206d 100644

drivers/iio/industrialio-core.c +13 -19
@@ -1504,27 +1504,27 @@ struct iio_dev *iio_device_alloc(int sizeof_priv)
 	alloc_size += IIO_ALIGN - 1;
 
 	dev = kzalloc(alloc_size, GFP_KERNEL);
+	if (!dev)
+		return NULL;
 
-	if (dev) {
-		dev->dev.groups = dev->groups;
-		dev->dev.type = &iio_device_type;
-		dev->dev.bus = &iio_bus_type;
-		device_initialize(&dev->dev);
-		dev_set_drvdata(&dev->dev, (void *)dev);
-		mutex_init(&dev->mlock);
-		mutex_init(&dev->info_exist_lock);
-		INIT_LIST_HEAD(&dev->channel_attr_list);
-
-		dev->id = ida_simple_get(&iio_ida, 0, 0, GFP_KERNEL);
-		if (dev->id < 0) {
-			/* cannot use a dev_err as the name isn't available */
-			pr_err("failed to get device id\n");
-			kfree(dev);
-			return NULL;
-		}
-		dev_set_name(&dev->dev, "iio:device%d", dev->id);
-		INIT_LIST_HEAD(&dev->buffer_list);
+	dev->dev.groups = dev->groups;
+	dev->dev.type = &iio_device_type;
+	dev->dev.bus = &iio_bus_type;
+	device_initialize(&dev->dev);
+	dev_set_drvdata(&dev->dev, (void *)dev);
+	mutex_init(&dev->mlock);
+	mutex_init(&dev->info_exist_lock);
+	INIT_LIST_HEAD(&dev->channel_attr_list);
+
+	dev->id = ida_simple_get(&iio_ida, 0, 0, GFP_KERNEL);
+	if (dev->id < 0) {
Quoted 8 lines

drivers/iio/industrialio-core.c struct iio_dev *iio_device_alloc(int sizeof_priv) · line 1524

Quoted 4 lines
drivers/iio/industrialio-core.c +4 -0
+		/* cannot use a dev_err as the name isn't available */
+		pr_err("failed to get device id\n");
+		kfree(dev);
+		return NULL;  

would it be too much for this patch to move this right after the kzalloc()? no strong opinion from my side to do it or not;
but it does save some init cycles, and compresses this init block a bit;

It doesn't really save any cycles because the chance of failure of ID allocation
is negligible... Now I'd agree with you if writing from scratch, but as a tidy up patch, it's good to keep things really simple.

Quoted 6 lines
Quoted 6 lines
drivers/iio/industrialio-core.c +2 -0
 	}
+	dev_set_name(&dev->dev, "iio:device%d", dev->id);
+	INIT_LIST_HEAD(&dev->buffer_list);
 
 	return dev;
 }