[PATCH] Fix for possible null pointer dereference in auth.c

[PATCH] Fix for possible null pointer dereference in auth.c

[Rickard Strandqvist]

There is otherwise a risk of a possible null pointer dereference.

Was largely found by using a static code analysis program called cppcheck.

Signed-off-by: Rickard Strandqvist <rickard_strandqvist@spectrumdigital.se>
---
 net/sunrpc/auth.c |   10 +++++++---
 1 fil ändrad, 7 tillägg(+), 3 borttagningar(-)

diff --git a/net/sunrpc/auth.c b/net/sunrpc/auth.c
index 5285ead..3a55698 100644
--- a/net/sunrpc/auth.c
+++ b/net/sunrpc/auth.c
@@ -801,10 +801,14 @@ rpcauth_invalcred(struct rpc_task *task)
 {
 	struct rpc_cred *cred = task->tk_rqstp->rq_cred;
 
-	dprintk("RPC: %5u invalidating %s cred %p\n",
-		task->tk_pid, cred->cr_auth->au_ops->au_name, cred);
-	if (cred)
+	if (cred) {
+		dprintk("RPC: %5u invalidating %s cred %p\n",
+			task->tk_pid, cred->cr_auth->au_ops->au_name, cred);
+
 		clear_bit(RPCAUTH_CRED_UPTODATE, &cred->cr_flags);
+	}
+	else
+		dprintk("RPC: %5u invalidating is NULL\n", task->tk_pid);
 }
 
 int
-- 
1.7.10.4

Re: [PATCH] Fix for possible null pointer dereference in auth.c

[Trond Myklebust]

On Thu, May 15, 2014 at 5:56 PM, Rickard Strandqvist
<rickard_strandqvist-IW2WV5XWFqGZkjO+N0TKoMugMpMbD5Xr@public.gmane.org> wrote:
> There is otherwise a risk of a possible null pointer dereference.
>
> Was largely found by using a static code analysis program called cppcheck.
>
> Signed-off-by: Rickard Strandqvist <rickard_strandqvist@spectrumdigital.se>
> ---
>  net/sunrpc/auth.c |   10 +++++++---
>  1 fil ändrad, 7 tillägg(+), 3 borttagningar(-)
>
> diff --git a/net/sunrpc/auth.c b/net/sunrpc/auth.c
> index 5285ead..3a55698 100644
> --- a/net/sunrpc/auth.c
> +++ b/net/sunrpc/auth.c
> @@ -801,10 +801,14 @@ rpcauth_invalcred(struct rpc_task *task)
>  {
>         struct rpc_cred *cred = task->tk_rqstp->rq_cred;
>
> -       dprintk("RPC: %5u invalidating %s cred %p\n",
> -               task->tk_pid, cred->cr_auth->au_ops->au_name, cred);
> -       if (cred)
> +       if (cred) {
> +               dprintk("RPC: %5u invalidating %s cred %p\n",
> +                       task->tk_pid, cred->cr_auth->au_ops->au_name, cred);
> +
>                 clear_bit(RPCAUTH_CRED_UPTODATE, &cred->cr_flags);
> +       }
> +       else
> +               dprintk("RPC: %5u invalidating is NULL\n", task->tk_pid);
>  }
>
>

Can you please explain how we can get to rpcauth_invalcred() without a
valid cred pointer? I'm not seeing it...

Cheers
  Trond
--
To unsubscribe from this list: send the line "unsubscribe linux-nfs" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [PATCH] Fix for possible null pointer dereference in auth.c

[Rickard Strandqvist]

Hi

I do not know, I'm not an expert on this code.
But since there before was a if (cred) someone must have thought this
could happen.

Best regards
Rickard Strandqvist


2014-05-16 0:07 GMT+02:00 Trond Myklebust <trond.myklebust@primarydata.com>:
> On Thu, May 15, 2014 at 5:56 PM, Rickard Strandqvist
> <rickard_strandqvist@spectrumdigital.se> wrote:
>> There is otherwise a risk of a possible null pointer dereference.
>>
>> Was largely found by using a static code analysis program called cppcheck.
>>
>> Signed-off-by: Rickard Strandqvist <rickard_strandqvist@spectrumdigital.se>
>> ---
>>  net/sunrpc/auth.c |   10 +++++++---
>>  1 fil ändrad, 7 tillägg(+), 3 borttagningar(-)
>>
>> diff --git a/net/sunrpc/auth.c b/net/sunrpc/auth.c
>> index 5285ead..3a55698 100644
>> --- a/net/sunrpc/auth.c
>> +++ b/net/sunrpc/auth.c
>> @@ -801,10 +801,14 @@ rpcauth_invalcred(struct rpc_task *task)
>>  {
>>         struct rpc_cred *cred = task->tk_rqstp->rq_cred;
>>
>> -       dprintk("RPC: %5u invalidating %s cred %p\n",
>> -               task->tk_pid, cred->cr_auth->au_ops->au_name, cred);
>> -       if (cred)
>> +       if (cred) {
>> +               dprintk("RPC: %5u invalidating %s cred %p\n",
>> +                       task->tk_pid, cred->cr_auth->au_ops->au_name, cred);
>> +
>>                 clear_bit(RPCAUTH_CRED_UPTODATE, &cred->cr_flags);
>> +       }
>> +       else
>> +               dprintk("RPC: %5u invalidating is NULL\n", task->tk_pid);
>>  }
>>
>>
>
> Can you please explain how we can get to rpcauth_invalcred() without a
> valid cred pointer? I'm not seeing it...
>
> Cheers
>   Trond

Re: [PATCH] Fix for possible null pointer dereference in auth.c

[Sergei Shtylyov]

Hello.

On 16-05-2014 1:56, Rickard Strandqvist wrote:

> There is otherwise a risk of a possible null pointer dereference.

> Was largely found by using a static code analysis program called cppcheck.

> Signed-off-by: Rickard Strandqvist <rickard_strandqvist@spectrumdigital.se>
> ---
>   net/sunrpc/auth.c |   10 +++++++---
>   1 fil ändrad, 7 tillägg(+), 3 borttagningar(-)

> diff --git a/net/sunrpc/auth.c b/net/sunrpc/auth.c
> index 5285ead..3a55698 100644
> --- a/net/sunrpc/auth.c
> +++ b/net/sunrpc/auth.c
> @@ -801,10 +801,14 @@ rpcauth_invalcred(struct rpc_task *task)
>   {
>   	struct rpc_cred *cred = task->tk_rqstp->rq_cred;
>
> -	dprintk("RPC: %5u invalidating %s cred %p\n",
> -		task->tk_pid, cred->cr_auth->au_ops->au_name, cred);
> -	if (cred)
> +	if (cred) {
> +		dprintk("RPC: %5u invalidating %s cred %p\n",
> +			task->tk_pid, cred->cr_auth->au_ops->au_name, cred);
> +
>   		clear_bit(RPCAUTH_CRED_UPTODATE, &cred->cr_flags);
> +	}
> +	else

    } and *else* should be on the same line, and there should be {} in the 
*else* arm since there's {} in the *if* arm already, according to 
Documentation/CodingStyle.

> +		dprintk("RPC: %5u invalidating is NULL\n", task->tk_pid);

    That's not a proper English, I'm afraid.

>   }

WBR, Sergei

Re: [PATCH] Fix for possible null pointer dereference in auth.c

[Rickard Strandqvist]

[-- Attachment #1: Type: text/plain, Size: 1734 bytes --]

Hi

I have made a new patch according Sergei specification.

But there remains Tronds question though.

Best regards
Rickard Strandqvist


2014-05-16 14:19 GMT+02:00 Sergei Shtylyov <sergei.shtylyov@cogentembedded.com>:
> Hello.
>
>
> On 16-05-2014 1:56, Rickard Strandqvist wrote:
>
>> There is otherwise a risk of a possible null pointer dereference.
>
>
>> Was largely found by using a static code analysis program called cppcheck.
>
>
>> Signed-off-by: Rickard Strandqvist
>> <rickard_strandqvist@spectrumdigital.se>
>> ---
>>   net/sunrpc/auth.c |   10 +++++++---
>>   1 fil ändrad, 7 tillägg(+), 3 borttagningar(-)
>
>
>> diff --git a/net/sunrpc/auth.c b/net/sunrpc/auth.c
>> index 5285ead..3a55698 100644
>> --- a/net/sunrpc/auth.c
>> +++ b/net/sunrpc/auth.c
>> @@ -801,10 +801,14 @@ rpcauth_invalcred(struct rpc_task *task)
>>   {
>>         struct rpc_cred *cred = task->tk_rqstp->rq_cred;
>>
>> -       dprintk("RPC: %5u invalidating %s cred %p\n",
>> -               task->tk_pid, cred->cr_auth->au_ops->au_name, cred);
>> -       if (cred)
>> +       if (cred) {
>> +               dprintk("RPC: %5u invalidating %s cred %p\n",
>> +                       task->tk_pid, cred->cr_auth->au_ops->au_name,
>> cred);
>> +
>>                 clear_bit(RPCAUTH_CRED_UPTODATE, &cred->cr_flags);
>> +       }
>> +       else
>
>
>    } and *else* should be on the same line, and there should be {} in the
> *else* arm since there's {} in the *if* arm already, according to
> Documentation/CodingStyle.
>
>
>> +               dprintk("RPC: %5u invalidating is NULL\n", task->tk_pid);
>
>
>    That's not a proper English, I'm afraid.
>
>>   }
>
>
> WBR, Sergei
>
>

[-- Attachment #2: 0001-Fix-for-possible-null-pointer-dereferenc.patch --]
[-- Type: text/x-patch, Size: 1202 bytes --]

From 210994905f7f7b13f6c621f110699f28fd134980 Mon Sep 17 00:00:00 2001
From: Rickard Strandqvist <rickard_strandqvist@spectrumdigital.se>
Date: Sun, 11 May 2014 18:51:42 +0200
Subject: [PATCH] net: sunrpc: auth.c: Fix for possible null pointer dereference

There is otherwise a risk of a possible null pointer dereference.

Was largely found by using a static code analysis program called cppcheck.

Signed-off-by: Rickard Strandqvist <rickard_strandqvist@spectrumdigital.se>
---
 net/sunrpc/auth.c |   10 +++++++---
 1 file changed, 7 insertions(+), 3 deletions(-)

diff --git a/net/sunrpc/auth.c b/net/sunrpc/auth.c
index 5285ead..a9e1866 100644
--- a/net/sunrpc/auth.c
+++ b/net/sunrpc/auth.c
@@ -801,10 +801,14 @@ rpcauth_invalcred(struct rpc_task *task)
 {
 	struct rpc_cred *cred = task->tk_rqstp->rq_cred;
 
-	dprintk("RPC: %5u invalidating %s cred %p\n",
-		task->tk_pid, cred->cr_auth->au_ops->au_name, cred);
-	if (cred)
+	if (cred) {
+		dprintk("RPC: %5u invalidating %s cred %p\n",
+			task->tk_pid, cred->cr_auth->au_ops->au_name, cred);
+
 		clear_bit(RPCAUTH_CRED_UPTODATE, &cred->cr_flags);
+	} else {
+		dprintk("RPC: %5u cred pointer is 0", task->tk_pid);
+	}
 }
 
 int
-- 
1.7.10.4

Re: [PATCH] Fix for possible null pointer dereference in auth.c

[Trond Myklebust]

On Sat, May 17, 2014 at 12:36 PM, Rickard Strandqvist
<rickard_strandqvist-IW2WV5XWFqGZkjO+N0TKoMugMpMbD5Xr@public.gmane.org> wrote:
> Hi
>
> I have made a new patch according Sergei specification.
>

I'll be happy to apply this patch once someone explains to me why we
can't just remove that check for cred != NULL, however I'm not
applying a patch purely in order to silence a warning from cppcheck.

Cheers,
   Trond

> But there remains Tronds question though.
>
> Best regards
> Rickard Strandqvist
>
>
> 2014-05-16 14:19 GMT+02:00 Sergei Shtylyov <sergei.shtylyov@cogentembedded.com>:
>> Hello.
>>
>>
>> On 16-05-2014 1:56, Rickard Strandqvist wrote:
>>
>>> There is otherwise a risk of a possible null pointer dereference.
>>
>>
>>> Was largely found by using a static code analysis program called cppcheck.
>>
>>
>>> Signed-off-by: Rickard Strandqvist
>>> <rickard_strandqvist-IW2WV5XWFqGZkjO+N0TKoMugMpMbD5Xr@public.gmane.org>
>>> ---
>>>   net/sunrpc/auth.c |   10 +++++++---
>>>   1 fil ändrad, 7 tillägg(+), 3 borttagningar(-)
>>
>>
>>> diff --git a/net/sunrpc/auth.c b/net/sunrpc/auth.c
>>> index 5285ead..3a55698 100644
>>> --- a/net/sunrpc/auth.c
>>> +++ b/net/sunrpc/auth.c
>>> @@ -801,10 +801,14 @@ rpcauth_invalcred(struct rpc_task *task)
>>>   {
>>>         struct rpc_cred *cred = task->tk_rqstp->rq_cred;
>>>
>>> -       dprintk("RPC: %5u invalidating %s cred %p\n",
>>> -               task->tk_pid, cred->cr_auth->au_ops->au_name, cred);
>>> -       if (cred)
>>> +       if (cred) {
>>> +               dprintk("RPC: %5u invalidating %s cred %p\n",
>>> +                       task->tk_pid, cred->cr_auth->au_ops->au_name,
>>> cred);
>>> +
>>>                 clear_bit(RPCAUTH_CRED_UPTODATE, &cred->cr_flags);
>>> +       }
>>> +       else
>>
>>
>>    } and *else* should be on the same line, and there should be {} in the
>> *else* arm since there's {} in the *if* arm already, according to
>> Documentation/CodingStyle.
>>
>>
>>> +               dprintk("RPC: %5u invalidating is NULL\n", task->tk_pid);
>>
>>
>>    That's not a proper English, I'm afraid.
>>
>>>   }
>>
>>
>> WBR, Sergei
>>
>>



-- 
Trond Myklebust

Linux NFS client maintainer, PrimaryData

trond.myklebust-7I+n7zu2hftEKMMhf/gKZA@public.gmane.org
--
To unsubscribe from this list: send the line "unsubscribe linux-nfs" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html