linux-kernel - Re: [PATCH v2 1/6] sysfs: do not create empty directories if no attributes are present

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:   Wed, 24 Aug 2022 17:17:44 +0200
From:   Pierre-Louis Bossart <pierre-louis.bossart@...ux.intel.com>
To:     Greg Kroah-Hartman <gregkh@...uxfoundation.org>,
        alsa-devel@...a-project.org
Cc:     Vinod Koul <vkoul@...nel.org>,
        Bard Liao <yung-chuan.liao@...ux.intel.com>,
        Sanyog Kale <sanyog.r.kale@...el.com>,
        linux-kernel@...r.kernel.org
Subject: Re: [PATCH v2 1/6] sysfs: do not create empty directories if no
 attributes are present



On 8/24/22 15:59, Greg Kroah-Hartman wrote:
> When creating an attribute group, if it is named a subdirectory is
> created and the sysfs files are placed into that subdirectory.  If no
> files are created, normally the directory would still be present, but it
> would be empty.  Clean this up by removing the directory if no files
> were successfully created in the group at all.
> 
> Cc: Vinod Koul <vkoul@...nel.org>
> Cc: Bard Liao <yung-chuan.liao@...ux.intel.com>
> Cc: Pierre-Louis Bossart <pierre-louis.bossart@...ux.intel.com>
> Cc: Sanyog Kale <sanyog.r.kale@...el.com>
> Cc: alsa-devel@...a-project.org
> Cc: linux-kernel@...r.kernel.org
> Signed-off-by: Greg Kroah-Hartman <gregkh@...uxfoundation.org>
> ---
> v2: new patch
> 
> Note, totally untested!  The following soundwire patches will need this,
> if a soundwire developer could test this out, it would be most
> apreciated.

Not able to see the kernel boot with this first patch. The device is
stuck with the cursor not even blinking. It seems our CI test devices
are also stuck.

This is completely beyond my comfort zone but I can run more tests to
root cause this.


> 
> fs/sysfs/group.c | 19 ++++++++++++++++---
>  1 file changed, 16 insertions(+), 3 deletions(-)
> 
> diff --git a/fs/sysfs/group.c b/fs/sysfs/group.c
> index eeb0e3099421..9fe0b47db47f 100644
> --- a/fs/sysfs/group.c
> +++ b/fs/sysfs/group.c
> @@ -31,12 +31,14 @@ static void remove_files(struct kernfs_node *parent,
>  			kernfs_remove_by_name(parent, (*bin_attr)->attr.name);
>  }
>  
> +/* returns -ERROR if error, or >= 0 for number of files actually created */
>  static int create_files(struct kernfs_node *parent, struct kobject *kobj,
>  			kuid_t uid, kgid_t gid,
>  			const struct attribute_group *grp, int update)
>  {
>  	struct attribute *const *attr;
>  	struct bin_attribute *const *bin_attr;
> +	int files_created = 0;
>  	int error = 0, i;
>  
>  	if (grp->attrs) {
> @@ -65,6 +67,8 @@ static int create_files(struct kernfs_node *parent, struct kobject *kobj,
>  						       gid, NULL);
>  			if (unlikely(error))
>  				break;
> +
> +			files_created++;
>  		}
>  		if (error) {
>  			remove_files(parent, grp);
> @@ -95,12 +99,15 @@ static int create_files(struct kernfs_node *parent, struct kobject *kobj,
>  							   NULL);
>  			if (error)
>  				break;
> +			files_created++;
>  		}
>  		if (error)
>  			remove_files(parent, grp);
>  	}
>  exit:
> -	return error;
> +	if (error)
> +		return error;
> +	return files_created;
>  }
>  
>  
> @@ -146,10 +153,16 @@ static int internal_create_group(struct kobject *kobj, int update,
>  		kn = kobj->sd;
>  	kernfs_get(kn);
>  	error = create_files(kn, kobj, uid, gid, grp, update);
> -	if (error) {
> +	if (error <= 0) {
> +		/*
> +		 * If an error happened _OR_ if no files were created in the
> +		 * attribute group, and we have a name for this group, delete
> +		 * the name so there's not an empty directory.
> +		 */
>  		if (grp->name)
>  			kernfs_remove(kn);
> -	}
> +	} else
> +		error = 0;
>  	kernfs_put(kn);
>  
>  	if (grp->name && update)