Skip to contentSkip to conversation

mfd: abx500-core: Fix compiler warning larger stack frame

PATCH In review Jay Aurabind 2 participants 6 messages May 10, 2014 – May 22, 2014

Subsystems (2)
lore
Share & tools
lore
#0 lore raw

PATCH 1/1 mfd: abx500-core: Fix compiler warning larger stack frame

Quoted 1 line

From 95f6ac368dbe352af0dfa2097b7d58b416447be5 Mon Sep 17 00:00:00 2001

From: Aurabindo J <[email protected]>
Date: Sat, 10 May 2014 15:16:58 +0530
Subject: [PATCH] mfd: abx500-core: Fix compiler warning larger stack frame

On systems with CONFIG_FRAME_WARN=1024, compiler warns the allocation of an object of struct device on stack. Make the allocation dynamically to fix the warning.

drivers/mfd/abx500-core.c: In function ?abx500_dump_all_banks?:
drivers/mfd/abx500-core.c:167:1: warning: the frame size of 1032 bytes is larger than 1024 bytes [-Wframe-larger-than=]

---
 drivers/mfd/abx500-core.c | 9 ++++++---
 1 file changed, 6 insertions(+), 3 deletions(-)
drivers/mfd/abx500-core.c +6 -3
index f3a15aa..87d80c3 100644
@@ -154,16 +154,19 @@ EXPORT_SYMBOL(abx500_startup_irq_enabled);
 void abx500_dump_all_banks(void)
 {
 	struct abx500_ops *ops;
-	struct device dummy_child = {NULL};
+	struct device *dummy_child;
 	struct abx500_device_entry *dev_entry;
 
+	dummy_child = kzalloc(sizeof(struct device), GFP_KERNEL);
+
 	list_for_each_entry(dev_entry, &abx500_list, list) {
-		dummy_child.parent = dev_entry->dev;
+		dummy_child->parent = dev_entry->dev;
 		ops = &dev_entry->ops;
 
 		if ((ops != NULL) && (ops->dump_all_banks != NULL))
-			ops->dump_all_banks(&dummy_child);
+			ops->dump_all_banks(dummy_child);
 	}
+	kfree(dummy_child);
 }
 EXPORT_SYMBOL(abx500_dump_all_banks);
 
Signature · 11 lines
1.9.1



-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 278 bytes
Desc: OpenPGP digital signature
URL: <http://lists.infradead.org/pipermail/linux-arm-kernel/attachments/20140510/f96bad0f/attachment.sig>
#1 lore raw
Quoted 9 lines

From 95f6ac368dbe352af0dfa2097b7d58b416447be5 Mon Sep 17 00:00:00 2001
From: Aurabindo J <[email protected]>
Date: Sat, 10 May 2014 15:16:58 +0530
Subject: [PATCH] mfd: abx500-core: Fix compiler warning larger stack frame

On systems with CONFIG_FRAME_WARN=1024, compiler warns the allocation of an object of struct device on stack. Make the allocation dynamically to fix the warning.

Quoted 8 lines

drivers/mfd/abx500-core.c: In function ?abx500_dump_all_banks?:
drivers/mfd/abx500-core.c:167:1: warning: the frame size of 1032 bytes is larger than 1024 bytes [-Wframe-larger-than=]

---
 drivers/mfd/abx500-core.c | 9 ++++++---
 1 file changed, 6 insertions(+), 3 deletions(-)

Looks good to me. I'll apply it until Linus complains.

Signature · 4 lines
Lee Jones
Linaro STMicroelectronics Landing Team Lead
Linaro.org ? Open source software for ARM SoCs
Follow Linaro: Facebook | Twitter | Blog
#2 lore raw

H1,

Change from V1 -> V2: Fix a potential null pointer dereference as reported by kbuild test robot <[email protected]>.

Inorder to report the kzalloc failure, I have just used a pr_debug statement. If it looks ugly, should I change the function's return value to int and give out an -ENOMEM ?

Quoted 1 line

From 6615c7a995a6bb4dcf4c1e9ac59bca823ec821a5 Mon Sep 17 00:00:00 2001

From: Aurabindo J <[email protected]>
Date: Wed, 21 May 2014 11:18:10 +0530
Subject: [PATCH] mfd: abx500-core: Fix compiler warning larger stack frame

On systems with CONFIG_FRAME_WARN=1024, compiler warns the allocation of an object of struct device on stack. Make the allocation dynamically to fix the warning.

---
 drivers/mfd/abx500-core.c | 19 ++++++++++++-------
 1 file changed, 12 insertions(+), 7 deletions(-)
drivers/mfd/abx500-core.c +12 -7
index f3a15aa..5e5098d 100644
@@ -154,16 +154,21 @@ EXPORT_SYMBOL(abx500_startup_irq_enabled);
 void abx500_dump_all_banks(void)
 {
 	struct abx500_ops *ops;
-	struct device dummy_child = {NULL};
+	struct device *dummy_child;
 	struct abx500_device_entry *dev_entry;
 
-	list_for_each_entry(dev_entry, &abx500_list, list) {
-		dummy_child.parent = dev_entry->dev;
-		ops = &dev_entry->ops;
+	dummy_child = kzalloc(sizeof(struct device), GFP_KERNEL);
+	if (dummy_child != NULL) {
+		list_for_each_entry(dev_entry, &abx500_list, list) {
+			dummy_child->parent = dev_entry->dev;
+			ops = &dev_entry->ops;
 
-		if ((ops != NULL) && (ops->dump_all_banks != NULL))
-			ops->dump_all_banks(&dummy_child);
-	}
+			if ((ops != NULL) && (ops->dump_all_banks != NULL))
+				ops->dump_all_banks(dummy_child);
+		}
+		kfree(dummy_child);
+	} else
+		pr_debug("kzalloc failure in abx500_dump_all_banks\n");
 }
 EXPORT_SYMBOL(abx500_dump_all_banks);
 
Signature · 11 lines
1.9.1



-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 278 bytes
Desc: OpenPGP digital signature
URL: <http://lists.infradead.org/pipermail/linux-arm-kernel/attachments/20140521/8791bd05/attachment.sig>
#3 lore raw
Quoted 3 lines

Change from V1 -> V2: Fix a potential null pointer dereference as reported by kbuild test robot <[email protected]>.

Inorder to report the kzalloc failure, I have just used a pr_debug statement. If it looks ugly, should I change the function's return value to int and give out an -ENOMEM ?

Yes, please do that - and obviously add the checks to the calling code.

Quoted 47 lines

From 6615c7a995a6bb4dcf4c1e9ac59bca823ec821a5 Mon Sep 17 00:00:00 2001
From: Aurabindo J <[email protected]>
Date: Wed, 21 May 2014 11:18:10 +0530
Subject: [PATCH] mfd: abx500-core: Fix compiler warning larger stack frame

On systems with CONFIG_FRAME_WARN=1024, compiler warns the allocation of an object of struct device on stack. Make the allocation dynamically to fix the warning.

---
 drivers/mfd/abx500-core.c | 19 ++++++++++++-------
 1 file changed, 12 insertions(+), 7 deletions(-)
drivers/mfd/abx500-core.c +12 -7
index f3a15aa..5e5098d 100644
@@ -154,16 +154,21 @@ EXPORT_SYMBOL(abx500_startup_irq_enabled);
 void abx500_dump_all_banks(void)
 {
 	struct abx500_ops *ops;
-	struct device dummy_child = {NULL};
+	struct device *dummy_child;
 	struct abx500_device_entry *dev_entry;
 
-	list_for_each_entry(dev_entry, &abx500_list, list) {
-		dummy_child.parent = dev_entry->dev;
-		ops = &dev_entry->ops;
+	dummy_child = kzalloc(sizeof(struct device), GFP_KERNEL);
+	if (dummy_child != NULL) {
+		list_for_each_entry(dev_entry, &abx500_list, list) {
+			dummy_child->parent = dev_entry->dev;
+			ops = &dev_entry->ops;
 
-		if ((ops != NULL) && (ops->dump_all_banks != NULL))
-			ops->dump_all_banks(&dummy_child);
-	}
+			if ((ops != NULL) && (ops->dump_all_banks != NULL))
+				ops->dump_all_banks(dummy_child);
+		}
+		kfree(dummy_child);
+	} else
+		pr_debug("kzalloc failure in abx500_dump_all_banks\n");
 }
 EXPORT_SYMBOL(abx500_dump_all_banks);
 
Signature · 4 lines
Lee Jones
Linaro STMicroelectronics Landing Team Lead
Linaro.org ? Open source software for ARM SoCs
Follow Linaro: Facebook | Twitter | Blog
#4 lore raw
Quoted 5 lines · Lee Jones
Quoted 3 lines

Change from V1 -> V2: Fix a potential null pointer dereference as reported by kbuild test robot <[email protected]>.

Inorder to report the kzalloc failure, I have just used a pr_debug statement. If it looks ugly, should I change the function's return value to int and give out an -ENOMEM ?

Yes, please do that - and obviously add the checks to the calling code.

But I cannot find any calling code. At least thats what linux cross reference has to say.

Change from V2 -> V3:
Change abx500_dump_all_banks to return int so as to account for error handling.

Change from V1 -> V2:
Fix a potential null pointer dereference as reported by kbuild test robot <[email protected]>.

Quoted 1 line

From 19d34af7240498b2e624e94b2d003f3d9928dee8 Mon Sep 17 00:00:00 2001

From: Aurabindo J <[email protected]>
Date: Wed, 21 May 2014 22:27:30 +0530
Subject: [PATCH] mfd: abx500-core: Fix compiler warning larger stack frame
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

On systems with CONFIG_FRAME_WARN=1024, compiler warns the allocation of an object of struct device on stack. Make the allocation dynamically to fix the warning. Also change the caller's return type to int so as to account for error handling.

drivers/mfd/abx500-core.c: In function ?abx500_dump_all_banks?:
drivers/mfd/abx500-core.c:167:1: warning: the frame size of 1032 bytes is larger than 1024 bytes [-Wframe-larger-than=]

---
 drivers/mfd/abx500-core.c  | 12 ++++++++----
 include/linux/mfd/abx500.h |  2 +-
 2 files changed, 9 insertions(+), 5 deletions(-)
drivers/mfd/abx500-core.c +8 -4
index f3a15aa..d6d0ec4 100644
@@ -151,19 +151,23 @@ int abx500_startup_irq_enabled(struct device *dev, unsigned int irq)
 }
 EXPORT_SYMBOL(abx500_startup_irq_enabled);
 
-void abx500_dump_all_banks(void)
+int abx500_dump_all_banks(void)
 {
 	struct abx500_ops *ops;
-	struct device dummy_child = {NULL};
+	struct device *dummy_child;
 	struct abx500_device_entry *dev_entry;
 
+	dummy_child = kzalloc(sizeof(struct device), GFP_KERNEL);
+	if (!dummy_child)
+		return -ENOMEM;
 	list_for_each_entry(dev_entry, &abx500_list, list) {
-		dummy_child.parent = dev_entry->dev;
+		dummy_child->parent = dev_entry->dev;
 		ops = &dev_entry->ops;
 
 		if ((ops != NULL) && (ops->dump_all_banks != NULL))
-			ops->dump_all_banks(&dummy_child);
+			ops->dump_all_banks(dummy_child);
 	}
+	kfree(dummy_child);
 }
 EXPORT_SYMBOL(abx500_dump_all_banks);
 
include/linux/mfd/abx500.h +1 -1
index 3301b20..df2508f 100644
@@ -330,7 +330,7 @@ int abx500_mask_and_set_register_interruptible(struct device *dev, u8 bank,
 int abx500_get_chip_id(struct device *dev);
 int abx500_event_registers_startup_state_get(struct device *dev, u8 *event);
 int abx500_startup_irq_enabled(struct device *dev, unsigned int irq);
-void abx500_dump_all_banks(void);
+int abx500_dump_all_banks(void);
 
 struct abx500_ops {
 	int (*get_chip_id) (struct device *);
Signature · 11 lines
1.9.1



-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 278 bytes
Desc: OpenPGP digital signature
URL: <http://lists.infradead.org/pipermail/linux-arm-kernel/attachments/20140521/3bba61a9/attachment.sig>
#5 lore raw
Quoted 5 lines
Quoted 3 lines
Quoted 1 line

Inorder to report the kzalloc failure, I have just used a pr_debug statement. If it looks ugly, should I change the function's return value to int and give out an -ENOMEM ?

Yes, please do that - and obviously add the checks to the calling code.

But I cannot find any calling code. At least thats what linux cross reference has to say.

Okay, thanks.

I've applied this patch and removed the other.

Quoted 74 lines

Change from V2 -> V3:
Change abx500_dump_all_banks to return int so as to account for error handling.

Change from V1 -> V2:
Fix a potential null pointer dereference as reported by kbuild test robot <[email protected]>.

From 19d34af7240498b2e624e94b2d003f3d9928dee8 Mon Sep 17 00:00:00 2001
From: Aurabindo J <[email protected]>
Date: Wed, 21 May 2014 22:27:30 +0530
Subject: [PATCH] mfd: abx500-core: Fix compiler warning larger stack frame
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

On systems with CONFIG_FRAME_WARN=1024, compiler warns the allocation of an object of struct device on stack. Make the allocation dynamically to fix the warning. Also change the caller's return type to int so as to account for error handling.

drivers/mfd/abx500-core.c: In function ?abx500_dump_all_banks?:
drivers/mfd/abx500-core.c:167:1: warning: the frame size of 1032 bytes is larger than 1024 bytes [-Wframe-larger-than=]

---
 drivers/mfd/abx500-core.c  | 12 ++++++++----
 include/linux/mfd/abx500.h |  2 +-
 2 files changed, 9 insertions(+), 5 deletions(-)
drivers/mfd/abx500-core.c +8 -4
index f3a15aa..d6d0ec4 100644
@@ -151,19 +151,23 @@ int abx500_startup_irq_enabled(struct device *dev, unsigned int irq)
 }
 EXPORT_SYMBOL(abx500_startup_irq_enabled);
 
-void abx500_dump_all_banks(void)
+int abx500_dump_all_banks(void)
 {
 	struct abx500_ops *ops;
-	struct device dummy_child = {NULL};
+	struct device *dummy_child;
 	struct abx500_device_entry *dev_entry;
 
+	dummy_child = kzalloc(sizeof(struct device), GFP_KERNEL);
+	if (!dummy_child)
+		return -ENOMEM;
 	list_for_each_entry(dev_entry, &abx500_list, list) {
-		dummy_child.parent = dev_entry->dev;
+		dummy_child->parent = dev_entry->dev;
 		ops = &dev_entry->ops;
 
 		if ((ops != NULL) && (ops->dump_all_banks != NULL))
-			ops->dump_all_banks(&dummy_child);
+			ops->dump_all_banks(dummy_child);
 	}
+	kfree(dummy_child);
 }
 EXPORT_SYMBOL(abx500_dump_all_banks);
 
include/linux/mfd/abx500.h +1 -1
index 3301b20..df2508f 100644
@@ -330,7 +330,7 @@ int abx500_mask_and_set_register_interruptible(struct device *dev, u8 bank,
 int abx500_get_chip_id(struct device *dev);
 int abx500_event_registers_startup_state_get(struct device *dev, u8 *event);
 int abx500_startup_irq_enabled(struct device *dev, unsigned int irq);
-void abx500_dump_all_banks(void);
+int abx500_dump_all_banks(void);
 
 struct abx500_ops {
 	int (*get_chip_id) (struct device *);
Signature · 4 lines
Lee Jones
Linaro STMicroelectronics Landing Team Lead
Linaro.org ? Open source software for ARM SoCs
Follow Linaro: Facebook | Twitter | Blog