Skip to contentSkip to conversation

replace BUG_ON to WARN_ON

PATCH Superseded Hyunmin Lee 3 participants 6 messages Jan 27, 2023 – Feb 1, 2023

Subsystems (2)

v1 of 2 Superseded Latest is v2 →

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 replace BUG_ON to WARN_ON

Replace unnacessary BUG_ON to WARN_ON. These BUG_ONs verify aruguments of a function. Thus, the WARN_ONs return an EINVAL error when their condition is true.

---
 mm/vmalloc.c | 11 ++++++++---
 1 file changed, 8 insertions(+), 3 deletions(-)
mm/vmalloc.c +8 -3
index 74afa2208558..9f9dba3132c5 100644
@@ -1587,9 +1587,14 @@ static struct vmap_area *alloc_vmap_area(unsigned long size,
 	int purged = 0;
 	int ret;
 
-	BUG_ON(!size);
-	BUG_ON(offset_in_page(size));
-	BUG_ON(!is_power_of_2(align));
+	if (WARN_ON(!size))
+		return ERR_PTR(-EINVAL);
+
+	if (WARN_ON(offset_in_page(size)))
+		return ERR_PTR(-EINVAL);
+
+	if (WARN_ON(!is_power_of_2(align)))
+		return ERR_PTR(-EINVAL);
 
 	if (unlikely(!vmap_initialized))
 		return ERR_PTR(-EBUSY);
Signature · 1 line
2.25.1
#1 lore raw

Hello Hyunmin.

the subject line could be "mm/vmalloc: replace BUG_ON() with WARN_ON()".

Quoted 10 lines · Hyunmin Lee

Replace unnacessary BUG_ON to WARN_ON. These BUG_ONs verify aruguments of a function. Thus, the WARN_ONs return an EINVAL error when their condition is true.

could be rephrased a little, like:

"As per the coding standards, in the event of an abnormal condition that should not occur under normal circumstances, the kernel should attempt recovery and proceed with execution, rather than halting the machine.

Specifically, in the alloc_vmap_area() function, use WARN_ON() and fail the request instead of using BUG_ON() to halt the machine."

Quoted 28 lines
---
 mm/vmalloc.c | 11 ++++++++---
 1 file changed, 8 insertions(+), 3 deletions(-)
mm/vmalloc.c +8 -3
index 74afa2208558..9f9dba3132c5 100644
@@ -1587,9 +1587,14 @@ static struct vmap_area *alloc_vmap_area(unsigned long size,
 	int purged = 0;
 	int ret;
 
-	BUG_ON(!size);
-	BUG_ON(offset_in_page(size));
-	BUG_ON(!is_power_of_2(align));
+	if (WARN_ON(!size))
+		return ERR_PTR(-EINVAL);
+
+	if (WARN_ON(offset_in_page(size)))
+		return ERR_PTR(-EINVAL);
+
+	if (WARN_ON(!is_power_of_2(align)))
+		return ERR_PTR(-EINVAL);
 
 	if (unlikely(!vmap_initialized))
 		return ERR_PTR(-EBUSY);
Signature · 1 line
2.25.1

The code change itself looks fine to me.

Even if BUG*() -> WARN*() conversion may not be a high priority task, I see no reason to reject such changes.

--
Thanks,
Hyeonggon

#2 lore raw
In reply to Hyunmin Lee

Hi,

Quoted 1 line · Hyunmin Lee

Replace unnacessary BUG_ON to WARN_ON. These BUG_ONs verify aruguments of a function. Thus, the WARN_ONs return an EINVAL error when their condition is true.

Some users enable panic_on_warn, so for them WARN_ON will still crash a machine.

I think a simple if() will be sufficient.

Quoted 38 lines
---
 mm/vmalloc.c | 11 ++++++++---
 1 file changed, 8 insertions(+), 3 deletions(-)
mm/vmalloc.c +8 -3
index 74afa2208558..9f9dba3132c5 100644
@@ -1587,9 +1587,14 @@ static struct vmap_area *alloc_vmap_area(unsigned long size,
 	int purged = 0;
 	int ret;
 
-	BUG_ON(!size);
-	BUG_ON(offset_in_page(size));
-	BUG_ON(!is_power_of_2(align));
+	if (WARN_ON(!size))
+		return ERR_PTR(-EINVAL);
+
+	if (WARN_ON(offset_in_page(size)))
+		return ERR_PTR(-EINVAL);
+
+	if (WARN_ON(!is_power_of_2(align)))
+		return ERR_PTR(-EINVAL);
 
 	if (unlikely(!vmap_initialized))
 		return ERR_PTR(-EBUSY);
Signature · 1 line
2.25.1
Signature · 2 lines
Sincerely yours,
Mike.
#3 lore raw
Quoted 42 lines · Mike Rapoport

Hi,

Quoted 1 line · Hyunmin Lee

Replace unnacessary BUG_ON to WARN_ON. These BUG_ONs verify aruguments of a function. Thus, the WARN_ONs return an EINVAL error when their condition is true.

Some users enable panic_on_warn, so for them WARN_ON will still crash a machine.

I think a simple if() will be sufficient.

Quoted 32 lines
---
 mm/vmalloc.c | 11 ++++++++---
 1 file changed, 8 insertions(+), 3 deletions(-)
mm/vmalloc.c +8 -3
index 74afa2208558..9f9dba3132c5 100644
@@ -1587,9 +1587,14 @@ static struct vmap_area *alloc_vmap_area(unsigned long size,
 	int purged = 0;
 	int ret;
 
-	BUG_ON(!size);
-	BUG_ON(offset_in_page(size));
-	BUG_ON(!is_power_of_2(align));
+	if (WARN_ON(!size))
+		return ERR_PTR(-EINVAL);
+
+	if (WARN_ON(offset_in_page(size)))
+		return ERR_PTR(-EINVAL);
+
+	if (WARN_ON(!is_power_of_2(align)))
+		return ERR_PTR(-EINVAL);
 
Quoted 10 lines
Quoted 6 lines
mm/vmalloc.c +0 -0
 	if (unlikely(!vmap_initialized))
 		return ERR_PTR(-EBUSY);
Signature · 1 line
2.25.1
Signature · 2 lines
Sincerely yours,
Mike.

Hi Mike,

Thank you for your advice.
Would you please give feedback about the below opinion?
- Printing warning messages is helpful to inform what happened in the system to the users.
- When a simple if() is used instead of WARN_ON, the if() should print a warning message.
- The condition of the simple if() should also have unlikely() for optimization of system performance.
- WARN_ON is a macro doing like thoes easily. It has a notifying function and unlikely optimization.
- Eventhough WARN_ON will still crash like BUG_ON by some users who enable panic_on_warn, it is their intention. They should accept the crash by WARN_ON.
- Therefore, using WARN_ON looks like natural and efficient.

Best,
Hyunmin

#4 lore raw
Quoted 13 lines · Hyunmin Lee
Quoted 10 lines · Mike Rapoport

Hi,

Quoted 1 line · Hyunmin Lee

Replace unnacessary BUG_ON to WARN_ON. These BUG_ONs verify aruguments of a function. Thus, the WARN_ONs return an EINVAL error when their condition is true.

Some users enable panic_on_warn, so for them WARN_ON will still crash a machine.

I think a simple if() will be sufficient.

Hi Mike,

Quoted 8 lines

Thank you for your advice.
Would you please give feedback about the below opinion?
- Printing warning messages is helpful to inform what happened in the system to the users.
- When a simple if() is used instead of WARN_ON, the if() should print a warning message.
- The condition of the simple if() should also have unlikely() for optimization of system performance.
- WARN_ON is a macro doing like thoes easily. It has a notifying function and unlikely optimization.
- Eventhough WARN_ON will still crash like BUG_ON by some users who enable panic_on_warn, it is their intention. They should accept the crash by WARN_ON.
- Therefore, using WARN_ON looks like natural and efficient.

As this is a validation of the function parameters, there is no need in warning messages and if(unlikely()) will do. There is really no point in WARN_ON() for something that's totally recoverable and very unlikely to happen.

Quoted 2 lines

Best,
Hyunmin

Signature · 2 lines
Sincerely yours,
Mike.
#5 lore raw
Quoted 23 lines · Mike Rapoport
Quoted 13 lines · Hyunmin Lee
Quoted 10 lines · Mike Rapoport

Hi,

Quoted 1 line · Hyunmin Lee

Replace unnacessary BUG_ON to WARN_ON. These BUG_ONs verify aruguments of a function. Thus, the WARN_ONs return an EINVAL error when their condition is true.

Some users enable panic_on_warn, so for them WARN_ON will still crash a machine.

I think a simple if() will be sufficient.

Hi Mike,

Quoted 8 lines

Thank you for your advice.
Would you please give feedback about the below opinion?
- Printing warning messages is helpful to inform what happened in the system to the users.
- When a simple if() is used instead of WARN_ON, the if() should print a warning message.
- The condition of the simple if() should also have unlikely() for optimization of system performance.
- WARN_ON is a macro doing like thoes easily. It has a notifying function and unlikely optimization.
- Eventhough WARN_ON will still crash like BUG_ON by some users who enable panic_on_warn, it is their intention. They should accept the crash by WARN_ON.
- Therefore, using WARN_ON looks like natural and efficient.

Quoted 11 lines

As this is a validation of the function parameters, there is no need in warning messages and if(unlikely()) will do. There is really no point in WARN_ON() for something that's totally recoverable and very unlikely to happen.

Quoted 2 lines

Best,
Hyunmin

Signature · 2 lines
Sincerely yours,
Mike.

Hi Mike,

According to your guidance, I will send a v3 patch. Thanks a lot.

Best,
Min