Received-SPF: pass (google.com: best guess record for domain of linux-kernel-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67;
Subject: Re: [PATCH] Staging: exfat: Avoid use of strcpy
To:     Dan Carpenter <dan.carpenter@oracle.com>,
        Sandro Volery <sandro@volery.com>
Cc:     valdis.kletnieks@vt.edu, gregkh@linuxfoundation.org,
        devel@driverdev.osuosl.org, linux-kernel@vger.kernel.org
References: <20190911055749.GA10786@volery> <20190911084111.GG15977@kadam>
From:   Rasmus Villemoes <linux@rasmusvillemoes.dk>
Message-ID: <214f3b20-fac4-df5e-6b3c-a08a19a8ea42@rasmusvillemoes.dk>
Date:   Wed, 11 Sep 2019 10:56:15 +0200
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:60.0) Gecko/20100101
 Thunderbird/60.8.0
MIME-Version: 1.0
In-Reply-To: <20190911084111.GG15977@kadam>
Content-Type: text/plain; charset=windows-1252
Content-Language: en-US
Content-Transfer-Encoding: 7bit
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

On 11/09/2019 10.41, Dan Carpenter wrote:
> On Wed, Sep 11, 2019 at 07:57:49AM +0200, Sandro Volery wrote:
>> Replaced strcpy with strscpy in exfat_core.c.
>>
>> Signed-off-by: Sandro Volery <sandro@volery.com>
>> ---
>>  drivers/staging/exfat/exfat_core.c | 2 +-
>>  1 file changed, 1 insertion(+), 1 deletion(-)
>>
>> diff --git a/drivers/staging/exfat/exfat_core.c b/drivers/staging/exfat/exfat_core.c
>> index da8c58149c35..c71b145e8a24 100644
>> --- a/drivers/staging/exfat/exfat_core.c
>> +++ b/drivers/staging/exfat/exfat_core.c
>> @@ -2964,7 +2964,7 @@ s32 resolve_path(struct inode *inode, char *path, struct chain_t *p_dir,
>>  	if (strlen(path) >= (MAX_NAME_LENGTH * MAX_CHARSET_SIZE))
>>  		return FFS_INVALIDPATH;
>>  
>> -	strcpy(name_buf, path);
>> +	strscpy(name_buf, path, sizeof(name_buf));
> 
> It checked strlen() earlier so we know that it can't overflow but, oh
> wow, the "name_buf" is a shared buffer.  wow wow wow.  This seems very
> racy.

Yeah, and note that the callers of resolve_path do seem to take a lock,
but that's not a global lock but a per-superblock (or
per-something-else) one... Obviously exfat should not rely on such a
single static buffer, but in the meantime, perhaps one should add a
name_buf_lock (though I don't know how long name_buf is actually in use,
so that needs checking).

Apart from the broken/lack of locking, it would be better to combine the
copying and length checking into a single check - i.e. do

  if (strscpy(name_buf, path, sizeof(name_buf)) < 0)
    return FFS_INVALIDPATH;

That's both more efficient and fixes the open-coding of sizeof(name_buf)
that is currently used in the if().

Rasmus