[PATCH v2] drivers/net/tulip/de4x5.c

[PATCH v2] drivers/net/tulip/de4x5.c

[Dan Rosenberg]

Fixed formatting (tabs and line breaks).

This was previously reported as a security issue due to leakage of
uninitialized stack memory.  Jeff Mahoney pointed out that this is
incorrect since the copied data is from a union (rather than a struct).
Therefore, this patch is only under consideration for the sake of
correctness, and is not security relevant. 

Signed-off-by: Dan Rosenberg <dan.j.rosenberg@gmail.com>
Acked-by: Grant Grundler <grundler@parisc-linux.org>

--- linux-2.6.35.4.orig/drivers/net/tulip/de4x5.c	2010-08-26 19:47:12.000000000 -0400
+++ linux-2.6.35.4/drivers/net/tulip/de4x5.c	2010-09-14 21:26:52.499474207 -0400
@@ -5474,7 +5474,7 @@ de4x5_ioctl(struct net_device *dev, stru
 	tmp.lval[6] = inl(DE4X5_STRR); j+=4;
 	tmp.lval[7] = inl(DE4X5_SIGR); j+=4;
 	ioc->len = j;
-	if (copy_to_user(ioc->data, tmp.addr, ioc->len)) return -EFAULT;
+	if (copy_to_user(ioc->data, tmp.lval, ioc->len)) return -EFAULT;
 	break;
 
 #define DE4X5_DUMP              0x0f /* Dump the DE4X5 Status */

Re: [PATCH v2] drivers/net/tulip/de4x5.c

[David Miller]

From: Dan Rosenberg <drosenberg@vsecurity.com>
Date: Wed, 15 Sep 2010 17:43:18 -0400

> Fixed formatting (tabs and line breaks).
> 
> This was previously reported as a security issue due to leakage of
> uninitialized stack memory.  Jeff Mahoney pointed out that this is
> incorrect since the copied data is from a union (rather than a struct).
> Therefore, this patch is only under consideration for the sake of
> correctness, and is not security relevant. 
> 
> Signed-off-by: Dan Rosenberg <dan.j.rosenberg@gmail.com>
> Acked-by: Grant Grundler <grundler@parisc-linux.org>

You need to give something more descriptive than just the name of the
file you're editing in the Subject: line of this commit message.

Otherwise people scanning "shortlog" and similar output won't have
even a general idea of what kind of change your commit is making.

Please formally resubmit this with a proper commit header line.

[PATCH v3] drivers/net/tulip/de4x5.c: fix union member name in DE4X5_GET_REG ioctl

[Dan Rosenberg]

Now with descriptive title, as requested.

This was previously reported as a security issue due to leakage of
uninitialized stack memory.  Jeff Mahoney pointed out that this is
incorrect since the copied data is from a union (rather than a struct).
Therefore, this patch is only under consideration for the sake of
correctness, and is not security relevant. 

Signed-off-by: Dan Rosenberg <dan.j.rosenberg@gmail.com>
Acked-by: Grant Grundler <grundler@parisc-linux.org>

--- linux-2.6.35.4.orig/drivers/net/tulip/de4x5.c       2010-08-26 19:47:12.000000000 -0400
+++ linux-2.6.35.4/drivers/net/tulip/de4x5.c    2010-09-14 21:26:52.499474207 -0400
@@ -5474,7 +5474,7 @@ de4x5_ioctl(struct net_device *dev, stru
        tmp.lval[6] = inl(DE4X5_STRR); j+=4;
        tmp.lval[7] = inl(DE4X5_SIGR); j+=4;
        ioc->len = j;
-       if (copy_to_user(ioc->data, tmp.addr, ioc->len)) return -EFAULT;
+       if (copy_to_user(ioc->data, tmp.lval, ioc->len)) return -EFAULT;
        break;
 
 #define DE4X5_DUMP              0x0f /* Dump the DE4X5 Status */

Re: [PATCH v3] drivers/net/tulip/de4x5.c: fix union member name in DE4X5_GET_REG ioctl

[David Miller]

From: Dan Rosenberg <drosenberg@vsecurity.com>
Date: Fri, 17 Sep 2010 01:05:44 -0400

> Now with descriptive title, as requested.
> 
> This was previously reported as a security issue due to leakage of
> uninitialized stack memory.  Jeff Mahoney pointed out that this is
> incorrect since the copied data is from a union (rather than a struct).
> Therefore, this patch is only under consideration for the sake of
> correctness, and is not security relevant. 
> 
> Signed-off-by: Dan Rosenberg <dan.j.rosenberg@gmail.com>
> Acked-by: Grant Grundler <grundler@parisc-linux.org>

Tab et al. damaged by your email client.

This is getting tiring for me, I guess it's getting tiring for
you too? :-)

[PATCH v4] drivers/net/tulip/de4x5.c: fix union member name in DE4X5_GET_REG ioctl

[Dan Rosenberg]

Tiring doesn't begin to describe it.  Formatting undamaged.

This was previously reported as a security issue due to leakage of
uninitialized stack memory.  Jeff Mahoney pointed out that this is
incorrect since the copied data is from a union (rather than a struct).
Therefore, this patch is only under consideration for the sake of
correctness, and is not security relevant. 

Signed-off-by: Dan Rosenberg <dan.j.rosenberg@gmail.com>
Acked-by: Grant Grundler <grundler@parisc-linux.org>

--- linux-2.6.35.4.orig/drivers/net/tulip/de4x5.c	2010-08-26 19:47:12.000000000 -0400
+++ linux-2.6.35.4/drivers/net/tulip/de4x5.c	2010-09-14 21:26:52.499474207 -0400
@@ -5474,7 +5474,7 @@ de4x5_ioctl(struct net_device *dev, stru
 	tmp.lval[6] = inl(DE4X5_STRR); j+=4;
 	tmp.lval[7] = inl(DE4X5_SIGR); j+=4;
 	ioc->len = j;
-	if (copy_to_user(ioc->data, tmp.addr, ioc->len)) return -EFAULT;
+	if (copy_to_user(ioc->data, tmp.lval, ioc->len)) return -EFAULT;
 	break;
 
 #define DE4X5_DUMP              0x0f /* Dump the DE4X5 Status */

Re: [PATCH v4] drivers/net/tulip/de4x5.c: fix union member name in DE4X5_GET_REG ioctl

[David Miller]

From: Dan Rosenberg <drosenberg@vsecurity.com>
Date: Fri, 17 Sep 2010 01:30:36 -0400

> Tiring doesn't begin to describe it.  Formatting undamaged.

:-)  Thanks.

> This was previously reported as a security issue due to leakage of
> uninitialized stack memory.  Jeff Mahoney pointed out that this is
> incorrect since the copied data is from a union (rather than a struct).
> Therefore, this patch is only under consideration for the sake of
> correctness, and is not security relevant. 
> 
> Signed-off-by: Dan Rosenberg <dan.j.rosenberg@gmail.com>
> Acked-by: Grant Grundler <grundler@parisc-linux.org>

Applied.

Re: [Security] [PATCH v4] drivers/net/tulip/de4x5.c: fix union member name in DE4X5_GET_REG ioctl

[Willy Tarreau]

On Thu, Sep 16, 2010 at 10:34:41PM -0700, David Miller wrote:
> From: Dan Rosenberg <drosenberg@vsecurity.com>
> Date: Fri, 17 Sep 2010 01:30:36 -0400
> 
> > Tiring doesn't begin to describe it.  Formatting undamaged.
> 
> :-)  Thanks.
> 
> > This was previously reported as a security issue due to leakage of
> > uninitialized stack memory.  Jeff Mahoney pointed out that this is
> > incorrect since the copied data is from a union (rather than a struct).
> > Therefore, this patch is only under consideration for the sake of
> > correctness, and is not security relevant. 
> > 
> > Signed-off-by: Dan Rosenberg <dan.j.rosenberg@gmail.com>
> > Acked-by: Grant Grundler <grundler@parisc-linux.org>
> 
> Applied.

David, just for the record, as was already reported on the list, this
fix is finally more a cleanup than a security fix because "tmp" is a
union and not a struct, so tmp.addr == tmp.lval.

Regards,
Willy

Re: [Security] [PATCH v4] drivers/net/tulip/de4x5.c: fix union member name in DE4X5_GET_REG ioctl

[Willy Tarreau]

On Fri, Sep 17, 2010 at 08:36:02AM +0200, Willy Tarreau wrote:
> On Thu, Sep 16, 2010 at 10:34:41PM -0700, David Miller wrote:
> > From: Dan Rosenberg <drosenberg@vsecurity.com>
> > Date: Fri, 17 Sep 2010 01:30:36 -0400
> > 
> > > Tiring doesn't begin to describe it.  Formatting undamaged.
> > 
> > :-)  Thanks.
> > 
> > > This was previously reported as a security issue due to leakage of
> > > uninitialized stack memory.  Jeff Mahoney pointed out that this is
> > > incorrect since the copied data is from a union (rather than a struct).
> > > Therefore, this patch is only under consideration for the sake of
> > > correctness, and is not security relevant. 
> > > 
> > > Signed-off-by: Dan Rosenberg <dan.j.rosenberg@gmail.com>
> > > Acked-by: Grant Grundler <grundler@parisc-linux.org>
> > 
> > Applied.
> 
> David, just for the record, as was already reported on the list, this
> fix is finally more a cleanup than a security fix because "tmp" is a
> union and not a struct, so tmp.addr == tmp.lval.

sorry, that was already explained in the commit message, next time
I'll read it fully before replying :-(

Willy

Re: [Security] [PATCH v4] drivers/net/tulip/de4x5.c: fix union member name in DE4X5_GET_REG ioctl

[David Miller]

From: Willy Tarreau <w@1wt.eu>
Date: Fri, 17 Sep 2010 08:36:02 +0200

> On Thu, Sep 16, 2010 at 10:34:41PM -0700, David Miller wrote:
>> From: Dan Rosenberg <drosenberg@vsecurity.com>
>> Date: Fri, 17 Sep 2010 01:30:36 -0400
>> 
>> > Tiring doesn't begin to describe it.  Formatting undamaged.
>> 
>> :-)  Thanks.
>> 
>> > This was previously reported as a security issue due to leakage of
>> > uninitialized stack memory.  Jeff Mahoney pointed out that this is
>> > incorrect since the copied data is from a union (rather than a struct).
>> > Therefore, this patch is only under consideration for the sake of
>> > correctness, and is not security relevant. 
>> > 
>> > Signed-off-by: Dan Rosenberg <dan.j.rosenberg@gmail.com>
>> > Acked-by: Grant Grundler <grundler@parisc-linux.org>
>> 
>> Applied.
> 
> David, just for the record, as was already reported on the list, this
> fix is finally more a cleanup than a security fix because "tmp" is a
> union and not a struct, so tmp.addr == tmp.lval.

I can read, thanks :-)

That's why I applied it to net-next-2.6 and not net-2.6

Re: [PATCH v3] drivers/net/tulip/de4x5.c: fix union member name in DE4X5_GET_REG ioctl

[Eric Dumazet]

Le vendredi 17 septembre 2010 à 01:05 -0400, Dan Rosenberg a écrit :
> Now with descriptive title, as requested.
> 
> This was previously reported as a security issue due to leakage of
> uninitialized stack memory.  Jeff Mahoney pointed out that this is
> incorrect since the copied data is from a union (rather than a struct).
> Therefore, this patch is only under consideration for the sake of
> correctness, and is not security relevant. 
> 
> Signed-off-by: Dan Rosenberg <dan.j.rosenberg@gmail.com>
> Acked-by: Grant Grundler <grundler@parisc-linux.org>
> 
> --- linux-2.6.35.4.orig/drivers/net/tulip/de4x5.c       2010-08-26 19:47:12.000000000 -0400
> +++ linux-2.6.35.4/drivers/net/tulip/de4x5.c    2010-09-14 21:26:52.499474207 -0400
> @@ -5474,7 +5474,7 @@ de4x5_ioctl(struct net_device *dev, stru
>         tmp.lval[6] = inl(DE4X5_STRR); j+=4;
>         tmp.lval[7] = inl(DE4X5_SIGR); j+=4;
>         ioc->len = j;
> -       if (copy_to_user(ioc->data, tmp.addr, ioc->len)) return -EFAULT;
> +       if (copy_to_user(ioc->data, tmp.lval, ioc->len)) return -EFAULT;
>         break;
>  
>  #define DE4X5_DUMP              0x0f /* Dump the DE4X5 Status */
> 


While you are at it

	if (copy_to_user(ioc->data, tmp.lval, ioc->len))
		return -EFAULT;

Thanks