Re: [PATCH V5 net-next 1/8] debugfs: Add debugfs_create_devm_dir() helper

[Jijie Shao <shaojijie@huawei.com>]

on 2024/12/9 14:31, Greg KH wrote:
> On Mon, Dec 09, 2024 at 09:02:10AM +0800, Jijie Shao wrote:
>> on 2024/12/6 19:40, Greg KH wrote:
>>> On Fri, Dec 06, 2024 at 07:16:22PM +0800, Jijie Shao wrote:
>>>> Add debugfs_create_devm_dir() helper
>>>>
>>>> Signed-off-by: Jijie Shao <shaojijie@huawei.com>
>>>> ---
>>>>    fs/debugfs/inode.c      | 36 ++++++++++++++++++++++++++++++++++++
>>>>    include/linux/debugfs.h | 10 ++++++++++
>>>>    2 files changed, 46 insertions(+)
>>>>
>>>> diff --git a/fs/debugfs/inode.c b/fs/debugfs/inode.c
>>>> index 38a9c7eb97e6..f682c4952a27 100644
>>>> --- a/fs/debugfs/inode.c
>>>> +++ b/fs/debugfs/inode.c
>>>> @@ -610,6 +610,42 @@ struct dentry *debugfs_create_dir(const char *name, struct dentry *parent)
>>>>    }
>>>>    EXPORT_SYMBOL_GPL(debugfs_create_dir);
>>>> +static void debugfs_remove_devm(void *dentry_rwa)
>>>> +{
>>>> +	struct dentry *dentry = dentry_rwa;
>>>> +
>>>> +	debugfs_remove(dentry);
>>>> +}
>>>> +
>>>> +/**
>>>> + * debugfs_create_devm_dir - Managed debugfs_create_dir()
>>>> + * @dev: Device that owns the action
>>>> + * @name: a pointer to a string containing the name of the directory to
>>>> + *        create.
>>>> + * @parent: a pointer to the parent dentry for this file.  This should be a
>>>> + *          directory dentry if set.  If this parameter is NULL, then the
>>>> + *          directory will be created in the root of the debugfs filesystem.
>>>> + * Managed debugfs_create_dir(). dentry will automatically be remove on
>>>> + * driver detach.
>>>> + */
>>>> +struct dentry *debugfs_create_devm_dir(struct device *dev, const char *name,
>>>> +				       struct dentry *parent)
>>>> +{
>>>> +	struct dentry *dentry;
>>>> +	int ret;
>>>> +
>>>> +	dentry = debugfs_create_dir(name, parent);
>>>> +	if (IS_ERR(dentry))
>>>> +		return dentry;
>>>> +
>>>> +	ret = devm_add_action_or_reset(dev, debugfs_remove_devm, dentry);
>>>> +	if (ret)
>>>> +		ERR_PTR(ret);
>>> You don't clean up the directory you created if this failed?  Why not?
>> Don't need to clean up.
>> in devm_add_action_or_reset(), if failed, will call action: debugfs_remove_devm(),
>> So, not clean up again.
>>
>> #define devm_add_action_or_reset(dev, action, data) \
>> 	__devm_add_action_or_reset(dev, action, data, #action)
>>
>> static inline int __devm_add_action_or_reset(struct device *dev, void (*action)(void *),
>> 					     void *data, const char *name)
>> {
>> 	int ret;
>>
>> 	ret = __devm_add_action(dev, action, data, name);
>> 	if (ret)
>> 		action(data);
>>
>> 	return ret;
>> }
>>
>> But there's a problem with this, I missed return.
>> I will add return in v6.
> As this did not even compile, how did you test any of this?
>
> I'm now loath to add this at all, please let's just keep this "open
> coded" in your driver for now until there are multiple users that need
> this and then convert them all to use the function when you add it.
>
> thanks,
>
> greg k-h

I've tested and it's ok,

But, there is a warning(warn_unused_result)..
My test command does not check warning other than the hibmcge driver.
This was caused by my carelessness. I'm very sorry.

I've looked at other driver codes, and quite a few others have similar codes.
As you suggest, this patchset still uses open code in next version.
In the future, I will try to send a new independent patchset to introduce it.

Thanks,
Jijie Shao