Skip to contentSkip to conversation

greybus: lights: check return of get_channel_from_mode

PATCH In review Rui Miguel Silva 3 participants 7 messages Mar 7, 2024 – Mar 25, 2024

Subsystems (3)
lore
Share & tools
lore
#0 lore raw

PATCH 1/1 greybus: lights: check return of get_channel_from_mode

If channel for the given node is not found we return null from get_channel_from_mode. Make sure we validate the return pointer before using it in two of the missing places.

This was originally reported in [0]:
Found by Linux Verification Center (linuxtesting.org) with SVACE.

[0] https://lore.kernel.org/all/[email protected]

Fixes 2870b52bae4c ("greybus: lights: add lights implementation")
---
 drivers/staging/greybus/light.c | 6 +++++-
 1 file changed, 5 insertions(+), 1 deletion(-)
drivers/staging/greybus/light.c +5 -1
index c6bd86a5335a..6f10b9e2c053 100644
@@ -147,6 +147,9 @@ static int __gb_lights_flash_brightness_set(struct gb_channel *channel)
 		channel = get_channel_from_mode(channel->light,
 						GB_CHANNEL_MODE_TORCH);
 
+	if (!channel)
+		return -EINVAL;
+
 	/* For not flash we need to convert brightness to intensity */
 	intensity = channel->intensity_uA.min +
 			(channel->intensity_uA.step * channel->led->brightness);
@@ -549,7 +552,8 @@ static int gb_lights_light_v4l2_register(struct gb_light *light)
 	}
 
 	channel_flash = get_channel_from_mode(light, GB_CHANNEL_MODE_FLASH);
-	WARN_ON(!channel_flash);
+	if (WARN_ON(!channel_flash))
+		return -EINVAL;
 
 	fled = &channel_flash->fled;
 
Signature · 1 line
2.44.0
#1 lore raw
Quoted 33 lines · Rui Miguel Silva

If channel for the given node is not found we return null from get_channel_from_mode. Make sure we validate the return pointer before using it in two of the missing places.

This was originally reported in [0]:
Found by Linux Verification Center (linuxtesting.org) with SVACE.

[0] https://lore.kernel.org/all/[email protected]

Fixes 2870b52bae4c ("greybus: lights: add lights implementation")
---
 drivers/staging/greybus/light.c | 6 +++++-
 1 file changed, 5 insertions(+), 1 deletion(-)
drivers/staging/greybus/light.c +3 -0
index c6bd86a5335a..6f10b9e2c053 100644
@@ -147,6 +147,9 @@ static int __gb_lights_flash_brightness_set(struct gb_channel *channel)
 		channel = get_channel_from_mode(channel->light,
 						GB_CHANNEL_MODE_TORCH);
 
+	if (!channel)
+		return -EINVAL;
+
 	/* For not flash we need to convert brightness to intensity */
 	intensity = channel->intensity_uA.min +
 			(channel->intensity_uA.step * channel->led->brightness);

drivers/staging/greybus/light.c static int gb_lights_light_v4l2_register(struct gb_light *light) · line 556

Quoted 6 lines
drivers/staging/greybus/light.c +2 -1
@@ -549,7 +552,8 @@ static int gb_lights_light_v4l2_register(struct gb_light *light)
 	}
 
 	channel_flash = get_channel_from_mode(light, GB_CHANNEL_MODE_FLASH);
-	WARN_ON(!channel_flash);
+	if (WARN_ON(!channel_flash))
+		return -EINVAL;

We should NOT crash machines just because of this, the WARN_ON() should be removed and just properly handle the error please.

thanks,

greg k-h

#2 lore raw
Quoted 36 lines · Greg Kroah-Hartman
Quoted 35 lines · Rui Miguel Silva

If channel for the given node is not found we return null from get_channel_from_mode. Make sure we validate the return pointer before using it in two of the missing places.

This was originally reported in [0]:
Found by Linux Verification Center (linuxtesting.org) with SVACE.

[0] https://lore.kernel.org/all/[email protected]

Fixes 2870b52bae4c ("greybus: lights: add lights implementation")
---
  drivers/staging/greybus/light.c | 6 +++++-
  1 file changed, 5 insertions(+), 1 deletion(-)
drivers/staging/greybus/light.c +3 -0
index c6bd86a5335a..6f10b9e2c053 100644
@@ -147,6 +147,9 @@ static int __gb_lights_flash_brightness_set(struct gb_channel *channel)
  		channel = get_channel_from_mode(channel->light,
  						GB_CHANNEL_MODE_TORCH);
 
+	if (!channel)
+		return -EINVAL;
+
  	/* For not flash we need to convert brightness to intensity */
  	intensity = channel->intensity_uA.min +
  			(channel->intensity_uA.step * channel->led->brightness);
@@ -549,7 +552,8 @@ static int gb_lights_light_v4l2_register(struct gb_light *light)
  	}
 
Quoted 7 lines

drivers/staging/greybus/light.c static int gb_lights_light_v4l2_register(struct gb_light *light) · line 556

Quoted 4 lines
drivers/staging/greybus/light.c +2 -1
  	channel_flash = get_channel_from_mode(light, GB_CHANNEL_MODE_FLASH);
-	WARN_ON(!channel_flash);
+	if (WARN_ON(!channel_flash))
+		return -EINVAL;

We should NOT crash machines just because of this, the WARN_ON() should be removed and just properly handle the error please.

Greg, WARN_ON() doesn't normally crash the machine. That said, it's reasonable to remove the WARN_ON().

I think the purpose of the warning is that this is a case that should "never happen," so if it does, it's making some noise.

The only caller is gb_lights_light_register(), and it checks for an error and does "properly handle" the\ error (assuming any error meaning "light has no flash" is correct).

Rui, please weigh in.

-Alex

Quoted 4 lines

thanks,

greg k-h

#3 lore raw
Quoted 43 lines · Alex Elder
Quoted 33 lines · Greg Kroah-Hartman
Quoted 32 lines · Rui Miguel Silva

If channel for the given node is not found we return null from get_channel_from_mode. Make sure we validate the return pointer before using it in two of the missing places.

This was originally reported in [0]:
Found by Linux Verification Center (linuxtesting.org) with SVACE.

[0] https://lore.kernel.org/all/[email protected]

Fixes 2870b52bae4c ("greybus: lights: add lights implementation")
---
  drivers/staging/greybus/light.c | 6 +++++-
  1 file changed, 5 insertions(+), 1 deletion(-)
drivers/staging/greybus/light.c +3 -0
index c6bd86a5335a..6f10b9e2c053 100644
@@ -147,6 +147,9 @@ static int __gb_lights_flash_brightness_set(struct gb_channel *channel)
  		channel = get_channel_from_mode(channel->light,
  						GB_CHANNEL_MODE_TORCH);
+	if (!channel)
+		return -EINVAL;
+
  	/* For not flash we need to convert brightness to intensity */
  	intensity = channel->intensity_uA.min +
  			(channel->intensity_uA.step * channel->led->brightness);
@@ -549,7 +552,8 @@ static int gb_lights_light_v4l2_register(struct gb_light *light)
Quoted 8 lines

drivers/staging/greybus/light.c static int gb_lights_light_v4l2_register(struct gb_light *light) · line 556

Quoted 5 lines
drivers/staging/greybus/light.c +2 -1
  	}
  	channel_flash = get_channel_from_mode(light, GB_CHANNEL_MODE_FLASH);
-	WARN_ON(!channel_flash);
+	if (WARN_ON(!channel_flash))
+		return -EINVAL;

We should NOT crash machines just because of this, the WARN_ON() should be removed and just properly handle the error please.

Quoted 2 lines

Greg, WARN_ON() doesn't normally crash the machine. That said, it's reasonable to remove the WARN_ON().

The huge majority of running Linux systems in the world run with panic-on-warn enabled, including the one in your pocket :(

Quoted 2 lines

I think the purpose of the warning is that this is a case that should "never happen," so if it does, it's making some noise.

Making noise by rebooting the box is not good.

thanks,

greg k-h

#4 lore raw
Quoted 44 lines · Greg Kroah-Hartman
Quoted 43 lines · Alex Elder
Quoted 41 lines · Greg Kroah-Hartman
Quoted 32 lines · Rui Miguel Silva

If channel for the given node is not found we return null from get_channel_from_mode. Make sure we validate the return pointer before using it in two of the missing places.

This was originally reported in [0]:
Found by Linux Verification Center (linuxtesting.org) with SVACE.

[0] https://lore.kernel.org/all/[email protected]

Fixes 2870b52bae4c ("greybus: lights: add lights implementation")
---
   drivers/staging/greybus/light.c | 6 +++++-
   1 file changed, 5 insertions(+), 1 deletion(-)
drivers/staging/greybus/light.c +3 -0
index c6bd86a5335a..6f10b9e2c053 100644
@@ -147,6 +147,9 @@ static int __gb_lights_flash_brightness_set(struct gb_channel *channel)
   		channel = get_channel_from_mode(channel->light,
   						GB_CHANNEL_MODE_TORCH);
+	if (!channel)
+		return -EINVAL;
+
   	/* For not flash we need to convert brightness to intensity */
   	intensity = channel->intensity_uA.min +
   			(channel->intensity_uA.step * channel->led->brightness);

drivers/staging/greybus/light.c static int gb_lights_light_v4l2_register(struct gb_light *light) · line 555

Quoted 5 lines
drivers/staging/greybus/light.c +2 -1
@@ -549,7 +552,8 @@ static int gb_lights_light_v4l2_register(struct gb_light *light)
   	}
   	channel_flash = get_channel_from_mode(light, GB_CHANNEL_MODE_FLASH);
-	WARN_ON(!channel_flash);
+	if (WARN_ON(!channel_flash))
+		return -EINVAL;

We should NOT crash machines just because of this, the WARN_ON() should be removed and just properly handle the error please.

Quoted 5 lines
Quoted 2 lines

Greg, WARN_ON() doesn't normally crash the machine. That said, it's reasonable to remove the WARN_ON().

The huge majority of running Linux systems in the world run with panic-on-warn enabled, including the one in your pocket :(

I did not know that. Then WARN_ON() is no better than BUG_ON(). I'm still learning. Thank you.

-Alex

Quoted 8 lines
Quoted 2 lines

I think the purpose of the warning is that this is a case that should "never happen," so if it does, it's making some noise.

Making noise by rebooting the box is not good.

thanks,

greg k-h

#5 lore raw

Hey Alex,

Quoted 45 lines · Alex Elder
Quoted 44 lines · Greg Kroah-Hartman
Quoted 43 lines · Alex Elder
Quoted 41 lines · Greg Kroah-Hartman
Quoted 32 lines · Rui Miguel Silva

If channel for the given node is not found we return null from get_channel_from_mode. Make sure we validate the return pointer before using it in two of the missing places.

This was originally reported in [0]:
Found by Linux Verification Center (linuxtesting.org) with SVACE.

[0] https://lore.kernel.org/all/[email protected]

Fixes 2870b52bae4c ("greybus: lights: add lights implementation")
---
   drivers/staging/greybus/light.c | 6 +++++-
   1 file changed, 5 insertions(+), 1 deletion(-)
drivers/staging/greybus/light.c +3 -0
index c6bd86a5335a..6f10b9e2c053 100644
@@ -147,6 +147,9 @@ static int __gb_lights_flash_brightness_set(struct gb_channel *channel)
   		channel = get_channel_from_mode(channel->light,
   						GB_CHANNEL_MODE_TORCH);
+	if (!channel)
+		return -EINVAL;
+
   	/* For not flash we need to convert brightness to intensity */
   	intensity = channel->intensity_uA.min +
   			(channel->intensity_uA.step * channel->led->brightness);

drivers/staging/greybus/light.c static int gb_lights_light_v4l2_register(struct gb_light *light) · line 555

Quoted 5 lines
drivers/staging/greybus/light.c +2 -1
@@ -549,7 +552,8 @@ static int gb_lights_light_v4l2_register(struct gb_light *light)
   	}
   	channel_flash = get_channel_from_mode(light, GB_CHANNEL_MODE_FLASH);
-	WARN_ON(!channel_flash);
+	if (WARN_ON(!channel_flash))
+		return -EINVAL;

We should NOT crash machines just because of this, the WARN_ON() should be removed and just properly handle the error please.

Quoted 8 lines
Quoted 5 lines
Quoted 2 lines

Greg, WARN_ON() doesn't normally crash the machine. That said, it's reasonable to remove the WARN_ON().

The huge majority of running Linux systems in the world run with panic-on-warn enabled, including the one in your pocket :(

I did not know that. Then WARN_ON() is no better than BUG_ON(). I'm still learning. Thank you.

I also lost track of all this failure cascade options that normally take the all system down.

Thanks anyway for the comments,
Cheers,
Rui

Quoted 11 lines

-Alex

Quoted 8 lines
Quoted 2 lines

I think the purpose of the warning is that this is a case that should "never happen," so if it does, it's making some noise.

Making noise by rebooting the box is not good.

thanks,

greg k-h

#6 lore raw
In reply to Greg Kroah-Hartman

Hi Greg,

Quoted 36 lines · Greg Kroah-Hartman
Quoted 35 lines · Rui Miguel Silva

If channel for the given node is not found we return null from get_channel_from_mode. Make sure we validate the return pointer before using it in two of the missing places.

This was originally reported in [0]:
Found by Linux Verification Center (linuxtesting.org) with SVACE.

[0] https://lore.kernel.org/all/[email protected]

Fixes 2870b52bae4c ("greybus: lights: add lights implementation")
---
 drivers/staging/greybus/light.c | 6 +++++-
 1 file changed, 5 insertions(+), 1 deletion(-)
drivers/staging/greybus/light.c +3 -0
index c6bd86a5335a..6f10b9e2c053 100644
@@ -147,6 +147,9 @@ static int __gb_lights_flash_brightness_set(struct gb_channel *channel)
 		channel = get_channel_from_mode(channel->light,
 						GB_CHANNEL_MODE_TORCH);
 
+	if (!channel)
+		return -EINVAL;
+
 	/* For not flash we need to convert brightness to intensity */
 	intensity = channel->intensity_uA.min +
 			(channel->intensity_uA.step * channel->led->brightness);
@@ -549,7 +552,8 @@ static int gb_lights_light_v4l2_register(struct gb_light *light)
 	}
 
Quoted 7 lines

drivers/staging/greybus/light.c static int gb_lights_light_v4l2_register(struct gb_light *light) · line 556

Quoted 4 lines
drivers/staging/greybus/light.c +2 -1
 	channel_flash = get_channel_from_mode(light, GB_CHANNEL_MODE_FLASH);
-	WARN_ON(!channel_flash);
+	if (WARN_ON(!channel_flash))
+		return -EINVAL;

We should NOT crash machines just because of this, the WARN_ON() should be removed and just properly handle the error please.

Yeah, will move this to a less severe option (dev_err) to make some noise about this "this should never happen" issue.

Cheers,
Rui

Quoted 3 lines

thanks,

greg k-h