Skip to contentSkip to conversation

drivers: edac: Drop unnecessary error check for debugfs_create_dir

PATCH In review Atul Kumar Pant 2 participants 5 messages Aug 15, 2023 – Aug 31, 2023

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

PATCH 1/1 drivers: edac: Drop unnecessary error check for debugfs_create_dir

This patch removes the error checking for debugfs_create_dir. Even if we get an error from this function, other debugfs APIs will handle the error value and doesn't crash in that case. Hence caller can safely ignore the errors that occur during the creation of debugfs nodes.

---
 drivers/edac/zynqmp_edac.c | 2 --
 1 file changed, 2 deletions(-)
drivers/edac/zynqmp_edac.c +0 -2
index cefbbafb945e..a8449f0711d7 100644
@@ -351,8 +351,6 @@ static void setup_debugfs(struct edac_device_ctl_info *edac_dev)
 	struct edac_priv *priv = edac_dev->pvt_info;
 
 	priv->debugfs_dir = edac_debugfs_create_dir("ocm");
-	if (IS_ERR(priv->debugfs_dir))
-		return;
 
 	edac_debugfs_create_x32("inject_fault_count", 0644, priv->debugfs_dir,
 				&priv->fault_injection_cnt);
Signature · 7 lines
2.25.1


_______________________________________________
linux-arm-kernel mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
maintainer
#1 lore raw
Quoted 1 line · Atul Kumar Pant

This patch removes the error checking for debugfs_create_dir.

Avoid using "This patch".

Quoted 3 lines

Even if we get an error from this function, other debugfs APIs will handle the error value and doesn't crash in that case. Hence caller can safely ignore the errors that occur during the creation of debugfs nodes.

First of all which issue do you have? Did you see that folder is not created?

I am not quite sure if this is the right behavior. In the code there is
135 if (!parent)
136 parent = edac_debugfs;

It means you are right that if creating ocm folder can fail and properties will be still created under edac_debugfs but is this the right behavior?

altera_edac/armada_xp_edac/i10nm/i5100/igen6/others are checking return value that's why I can't see any reason to remove this checking from one driver.

If you want to fix all please send patch for all but I don't think it will improve situation and it will just hide different issue if creating folder fails. And debugfs will be disabled in production system anyway.

Thanks,
Michal

_______________________________________________
linux-arm-kernel mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

#2 lore raw
Quoted 6 lines · Michal Simek
Quoted 1 line · Atul Kumar Pant

This patch removes the error checking for debugfs_create_dir.

Avoid using "This patch".

Thanks for pointing this out. I'll remember this.

Quoted 6 lines
Quoted 4 lines

Even if we get an error from this function, other debugfs APIs will handle the error value and doesn't crash in that case. Hence caller can safely ignore the errors that occur during the creation of debugfs nodes.

First of all which issue do you have? Did you see that folder is not created?

I have not seen any issue as such. But going by the comments before
the debugfs_create_dir API (https://elixir.bootlin.com/linux/latest/source/fs/debugfs/inode.c#L583),
we can ignore safely ignore the return value from this API.

Quoted 9 lines

I am not quite sure if this is the right behavior. In the code there is
135 if (!parent)
136 parent = edac_debugfs;

It means you are right that if creating ocm folder can fail and properties will be still created under edac_debugfs but is this the right behavior?

Quoted 7 lines

altera_edac/armada_xp_edac/i10nm/i5100/igen6/others are checking return value that's why I can't see any reason to remove this checking from one driver.

If you want to fix all please send patch for all but I don't think it will improve situation and it will just hide different issue if creating folder fails.

Understood your point. Are you suggesting that we should keep these checks as it is, or should I fix for all the drivers and upload the patch ?

Quoted 8 lines

And debugfs will be disabled in production system anyway.

Thanks,
Michal

_______________________________________________
linux-arm-kernel mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

maintainer
#3 lore raw
Quoted 34 lines · Atul Kumar Pant
Quoted 6 lines · Michal Simek
Quoted 1 line · Atul Kumar Pant

This patch removes the error checking for debugfs_create_dir.

Avoid using "This patch".

Thanks for pointing this out. I'll remember this.

Quoted 6 lines
Quoted 4 lines

Even if we get an error from this function, other debugfs APIs will handle the error value and doesn't crash in that case. Hence caller can safely ignore the errors that occur during the creation of debugfs nodes.

First of all which issue do you have? Did you see that folder is not created?

I have not seen any issue as such. But going by the comments before
the debugfs_create_dir API (https://elixir.bootlin.com/linux/latest/source/fs/debugfs/inode.c#L583),
we can ignore safely ignore the return value from this API.

Quoted 13 lines

I am not quite sure if this is the right behavior. In the code there is
135 if (!parent)
136 parent = edac_debugfs;

It means you are right that if creating ocm folder can fail and properties will be still created under edac_debugfs but is this the right behavior?

altera_edac/armada_xp_edac/i10nm/i5100/igen6/others are checking return value that's why I can't see any reason to remove this checking from one driver.

Quoted 7 lines
Quoted 3 lines

If you want to fix all please send patch for all but I don't think it will improve situation and it will just hide different issue if creating folder fails.

Understood your point. Are you suggesting that we should keep these checks as it is, or should I fix for all the drivers and upload the patch ?

Up to Boris to decide but I would say keep it as is. Even debugfs is not stable interface I would like to be informed if something fails. But just 2c.

Thanks,
Michal

_______________________________________________
linux-arm-kernel mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

#4 lore raw
Quoted 41 lines · Michal Simek
Quoted 38 lines · Atul Kumar Pant
Quoted 6 lines · Michal Simek
Quoted 1 line · Atul Kumar Pant

This patch removes the error checking for debugfs_create_dir.

Avoid using "This patch".

Thanks for pointing this out. I'll remember this.

Quoted 6 lines
Quoted 4 lines

Even if we get an error from this function, other debugfs APIs will handle the error value and doesn't crash in that case. Hence caller can safely ignore the errors that occur during the creation of debugfs nodes.

First of all which issue do you have? Did you see that folder is not created?

I have not seen any issue as such. But going by the comments before
the debugfs_create_dir API (https://elixir.bootlin.com/linux/latest/source/fs/debugfs/inode.c#L583),
we can ignore safely ignore the return value from this API.

Quoted 16 lines

I am not quite sure if this is the right behavior. In the code there is
135 if (!parent)
136 parent = edac_debugfs;

It means you are right that if creating ocm folder can fail and properties will be still created under edac_debugfs but is this the right behavior?

altera_edac/armada_xp_edac/i10nm/i5100/igen6/others are checking return value that's why I can't see any reason to remove this checking from one driver.

If you want to fix all please send patch for all but I don't think it will improve situation and it will just hide different issue if creating folder fails.

Quoted 7 lines
Quoted 3 lines

Understood your point. Are you suggesting that we should keep these checks as it is, or should I fix for all the drivers and upload the patch ?

Up to Boris to decide but I would say keep it as is. Even debugfs is not stable interface I would like to be informed if something fails. But just 2c.

Thanks you Michal. I'll wait for the reply from Boris.

Quoted 6 lines

Thanks,
Michal