Skip to contentSkip to conversation

irqchip/gic: check return value of of_address_to_resource

PATCH In review Bo Yan 2 participants 5 messages Jul 5, 2018 – Jul 7, 2018

Subsystems (2)
lore
Share & tools
lore

Bo Yan

author
#0 lore raw

PATCH 1/1 irqchip/gic: check return value of of_address_to_resource

The of_address_to_resource returns 0 if successful. gic_check_eoimode calls it without checking the return value. This induces Coverity
warning: "Unchecked return value".

Return false from gic_check_eoimode if of_address_to_resource returns non-0 value.

---
 drivers/irqchip/irq-gic.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
drivers/irqchip/irq-gic.c +2 -1
index ced10c4..0bceb10 100644
@@ -1284,7 +1284,8 @@ static bool gic_check_eoimode(struct device_node *node, void __iomem **base)
 {
 	struct resource cpuif_res;
 
-	of_address_to_resource(node, 1, &cpuif_res);
+	if (of_address_to_resource(node, 1, &cpuif_res))
+		return false;
 
 	if (!is_hyp_mode_available())
 		return false;
Signature · 1 line
2.7.4
#1 lore raw

Hi Bo,

Quoted 17 lines · On Thu, 5 Jul 2018 11:20:59 -0700 Bo Yan <[email protected]> wrote:

The of_address_to_resource returns 0 if successful. gic_check_eoimode calls it without checking the return value. This induces Coverity
warning: "Unchecked return value".

Return false from gic_check_eoimode if of_address_to_resource returns non-0 value.

---
 drivers/irqchip/irq-gic.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
drivers/irqchip/irq-gic.c +0 -0
index ced10c4..0bceb10 100644

drivers/irqchip/irq-gic.c static bool gic_check_eoimode(struct device_node *node, void __iomem **base) · line 1288

Quoted 6 lines
drivers/irqchip/irq-gic.c +2 -1
@@ -1284,7 +1284,8 @@ static bool gic_check_eoimode(struct device_node *node, void __iomem **base)
 {
 	struct resource cpuif_res;
 
-	of_address_to_resource(node, 1, &cpuif_res);
+	if (of_address_to_resource(node, 1, &cpuif_res))
+		return false;

We've just done an of_iomap() on this resource, which succeeded. How can the same thing now fail? It would mean that the device tree has been pulled from under our feet...

And if it could happen, why is returning false the right thing to do? Why would we say we want EOImode==0 instead of 1?

drivers/irqchip/irq-gic.c static bool gic_check_eoimode(struct device_node *node, void __iomem **base) · line 1291

Quoted 3 lines
drivers/irqchip/irq-gic.c +0 -0
 	if (!is_hyp_mode_available())
 		return false;

As it stands, I'm not taking such a patch. It either papers over a bigger problem, or just keeps a warning quiet for the sake of it.

Thanks,

	M.
Signature · 1 line
Without deviation from the norm, progress is not possible.
author
#2 lore raw

Marc,

I'm also wondering if of_address_to_resource can really fail in this particular case?

What if we just explicitly discard the return value like this:

(void)of_address_to_resource(node, 1, &cpuif_res);

This suppresses Coverity warning by explicitly stating we are 100% sure the function call will always return success.

Quoted 47 lines · Marc Zyngier

Hi Bo,

Quoted 17 lines · On Thu, 5 Jul 2018 11:20:59 -0700 Bo Yan <[email protected]> wrote:

The of_address_to_resource returns 0 if successful. gic_check_eoimode calls it without checking the return value. This induces Coverity
warning: "Unchecked return value".

Return false from gic_check_eoimode if of_address_to_resource returns non-0 value.

---
  drivers/irqchip/irq-gic.c | 3 ++-
  1 file changed, 2 insertions(+), 1 deletion(-)
drivers/irqchip/irq-gic.c +0 -0
index ced10c4..0bceb10 100644

drivers/irqchip/irq-gic.c static bool gic_check_eoimode(struct device_node *node, void __iomem **base) · line 1288

Quoted 6 lines
drivers/irqchip/irq-gic.c +2 -1
@@ -1284,7 +1284,8 @@ static bool gic_check_eoimode(struct device_node *node, void __iomem **base)
  {
  	struct resource cpuif_res;
 
-	of_address_to_resource(node, 1, &cpuif_res);
+	if (of_address_to_resource(node, 1, &cpuif_res))
+		return false;

We've just done an of_iomap() on this resource, which succeeded. How can the same thing now fail? It would mean that the device tree has been pulled from under our feet...

And if it could happen, why is returning false the right thing to do? Why would we say we want EOImode==0 instead of 1?

drivers/irqchip/irq-gic.c static bool gic_check_eoimode(struct device_node *node, void __iomem **base) · line 1291

Quoted 3 lines
drivers/irqchip/irq-gic.c +0 -0
  	if (!is_hyp_mode_available())
  		return false;

As it stands, I'm not taking such a patch. It either papers over a bigger problem, or just keeps a warning quiet for the sake of it.

Thanks,

	M.
author
#3 lore raw

Marc,

Sorry for the previous reply. My email settings were not correct, so it inserted those confidentiality text, which was not what I intended.

This is what I think:

drivers/irqchip/irq-gic.c +1 -1
index ced10c4..0b60bb0 100644
@@ -1284,7 +1284,7 @@ static bool gic_check_eoimode(struct device_node *node, void __iomem **base)
 {
        struct resource cpuif_res;
 
-       of_address_to_resource(node, 1, &cpuif_res);
+       (void)of_address_to_resource(node, 1, &cpuif_res);
 
        if (!is_hyp_mode_available())
                return false;

We are 100% sure of_address_to_resource will succeed in this particular case, so "(void)" will help suppress Coverity warning.

Quoted 57 lines · Bo Yan

Marc,

I'm also wondering if of_address_to_resource can really fail in this particular case?

What if we just explicitly discard the return value like this:

(void)of_address_to_resource(node, 1, &cpuif_res);

This suppresses Coverity warning by explicitly stating we are 100% sure the function call will always return success.

Quoted 46 lines · Marc Zyngier

Hi Bo,

Quoted 17 lines · On Thu, 5 Jul 2018 11:20:59 -0700 Bo Yan <[email protected]> wrote:

The of_address_to_resource returns 0 if successful. gic_check_eoimode calls it without checking the return value. This induces Coverity
warning: "Unchecked return value".

Return false from gic_check_eoimode if of_address_to_resource returns non-0 value.

---
  drivers/irqchip/irq-gic.c | 3 ++-

  1 file changed, 2 insertions(+), 1 deletion(-)
drivers/irqchip/irq-gic.c +0 -0
index ced10c4..0bceb10 100644

drivers/irqchip/irq-gic.c static bool gic_check_eoimode(struct · line 1288

Quoted 6 lines
drivers/irqchip/irq-gic.c +2 -1
@@ -1284,7 +1284,8 @@ static bool gic_check_eoimode(struct 
 device_node *node, void __iomem **base)
   {
       struct resource cpuif_res;
-    of_address_to_resource(node, 1, &cpuif_res);
+    if (of_address_to_resource(node, 1, &cpuif_res))
+        return false;

We've just done an of_iomap() on this resource, which succeeded. How can the same thing now fail? It would mean that the device tree has been pulled from under our feet...

And if it could happen, why is returning false the right thing to do? Why would we say we want EOImode==0 instead of 1?

drivers/irqchip/irq-gic.c static bool gic_check_eoimode(struct · line 1290

Quoted 2 lines
drivers/irqchip/irq-gic.c +0 -0
       if (!is_hyp_mode_available())
           return false;

As it stands, I'm not taking such a patch. It either papers over a bigger problem, or just keeps a warning quiet for the sake of it.

Thanks,

    M.

#4 lore raw
Quoted 20 lines · On Thu, 5 Jul 2018 12:32:22 -0700 Bo Yan <[email protected]> wrote:

Marc,

Sorry for the previous reply. My email settings were not correct, so it inserted those confidentiality text, which was not what I intended.

This is what I think:

drivers/irqchip/irq-gic.c +1 -1
index ced10c4..0b60bb0 100644
@@ -1284,7 +1284,7 @@ static bool gic_check_eoimode(struct device_node *node, void __iomem **base)
  {
         struct resource cpuif_res;
 
-       of_address_to_resource(node, 1, &cpuif_res);
+       (void)of_address_to_resource(node, 1, &cpuif_res);
 
         if (!is_hyp_mode_available())
                 return false;
Quoted 1 line

We are 100% sure of_address_to_resource will succeed in this particular case, so "(void)" will help suppress Coverity warning.

In all honesty. I don't see the point of patching the kernel to silence a warning when we know that this is a false positive. I'm sure you can flag that one as "false positive" in Coverity.

Thanks,

	M.
Signature · 1 line
Without deviation from the norm, progress is not possible.