linux-kernel - Re: [PATCH v3 06/10] sysctl: drop away useless label

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]

Message-Id: <20160405151752.de82f221f1c70ff901b2b958@linux-foundation.org>
Date:	Tue, 5 Apr 2016 15:17:52 -0700
From:	Andrew Morton <akpm@...ux-foundation.org>
To:	Andy Shevchenko <andriy.shevchenko@...ux.intel.com>
Cc:	Dmitry Kasatkin <dmitry.kasatkin@...il.com>,
	Mimi Zohar <zohar@...ux.vnet.ibm.com>,
	linux-kernel@...r.kernel.org, linux-api@...r.kernel.org,
	Rasmus Villemoes <linux@...musvillemoes.dk>,
	linux-efi@...r.kernel.org, Matt Fleming <matt@...eblueprint.co.uk>,
	Arnd Bergmann <arnd@...db.de>, "Theodore Ts'o" <tytso@....edu>
Subject: Re: [PATCH v3 06/10] sysctl: drop away useless label

On Tue,  5 Apr 2016 16:56:15 +0300 Andy Shevchenko <andriy.shevchenko@...ux.intel.com> wrote:

> We have no locking in bin_uuid(). Thus, we may remove the out label and use
> return statements directly.
> 
> ...
>
> --- a/kernel/sysctl_binary.c
> +++ b/kernel/sysctl_binary.c
> @@ -1123,15 +1123,14 @@ static ssize_t bin_uuid(struct file *file,
>  
>  		result = kernel_read(file, 0, buf, sizeof(buf) - 1);
>  		if (result < 0)
> -			goto out;
> +			return result;
>  
>  		buf[result] = '\0';
>  
>  		/* Convert the uuid to from a string to binary */
>  		for (i = 0; i < 16; i++) {
> -			result = -EIO;
>  			if (!isxdigit(str[0]) || !isxdigit(str[1]))
> -				goto out;
> +				return -EIO;
>  
>  			uuid[i] = (hex_to_bin(str[0]) << 4) |
>  					hex_to_bin(str[1]);
> @@ -1143,15 +1142,12 @@ static ssize_t bin_uuid(struct file *file,
>  		if (oldlen > 16)
>  			oldlen = 16;
>  
> -		result = -EFAULT;
>  		if (copy_to_user(oldval, uuid, oldlen))
> -			goto out;
> +			return -EFAULT;
>  
>  		copied = oldlen;
>  	}
> -	result = copied;
> -out:
> -	return result;
> +	return copied;
>  }

Sure, but we may add locking or resource allocation in the future, in
which case this change will need to be undone.  I think it's better to
leave the code as-is.  It's presently quite typical kernel code.