[PATCH] drivers/crypto/nx: saves chaining value from co-processor

[PATCH] drivers/crypto/nx: saves chaining value from co-processor

[Fionnuala Gunter]

The chaining value from co-processor was not being saved. This value is
needed because it is used as the IV, for example by cts(cbc(aes)).

Signed-off-by: Fionnuala Gunter <fin@linux.vnet.ibm.com>
Signed-off-by: Marcelo Cerri <mhcerri@linux.vnet.ibm.com>
---
 drivers/crypto/nx/nx-aes-cbc.c |    1 +
 1 file changed, 1 insertion(+)

diff --git a/drivers/crypto/nx/nx-aes-cbc.c b/drivers/crypto/nx/nx-aes-cbc.c
index 35d483f..a2f99a9 100644
--- a/drivers/crypto/nx/nx-aes-cbc.c
+++ b/drivers/crypto/nx/nx-aes-cbc.c
@@ -95,6 +95,7 @@ static int cbc_aes_nx_crypt(struct blkcipher_desc *desc,
        if (rc)
                goto out;

+       memcpy(desc->info, csbcpb->cpb.aes_cbc.cv, AES_BLOCK_SIZE);
        atomic_inc(&(nx_ctx->stats->aes_ops));
        atomic64_add(csbcpb->csb.processed_byte_count,
                     &(nx_ctx->stats->aes_bytes));
-- 
1.7.10.4

[PATCH] drivers/crypto/nx: saves chaining value from co-processor

[Fionnuala Gunter]

The chaining value from co-processor was not being saved. This value is
needed because it is used as the IV, for example by cts(cbc(aes)).

Signed-off-by: Fionnuala Gunter <fin@linux.vnet.ibm.com>
Signed-off-by: Marcelo Cerri <mhcerri@linux.vnet.ibm.com>
---
 drivers/crypto/nx/nx-aes-cbc.c |    1 +
 1 file changed, 1 insertion(+)

diff --git a/drivers/crypto/nx/nx-aes-cbc.c b/drivers/crypto/nx/nx-aes-cbc.c
index 35d483f..a2f99a9 100644
--- a/drivers/crypto/nx/nx-aes-cbc.c
+++ b/drivers/crypto/nx/nx-aes-cbc.c
@@ -95,6 +95,7 @@ static int cbc_aes_nx_crypt(struct blkcipher_desc *desc,
        if (rc)
                goto out;

+       memcpy(desc->info, csbcpb->cpb.aes_cbc.cv, AES_BLOCK_SIZE);
        atomic_inc(&(nx_ctx->stats->aes_ops));
        atomic64_add(csbcpb->csb.processed_byte_count,
                     &(nx_ctx->stats->aes_bytes));
-- 
1.7.10.4

Re: [PATCH] drivers/crypto/nx: saves chaining value from co-processor

[Michael Ellerman]

On Fri, Aug 02, 2013 at 05:16:56PM -0500, Fionnuala Gunter wrote:
> On Sat, 2013-08-03 at 07:59 +1000, Benjamin Herrenschmidt wrote:
> > On Fri, 2013-08-02 at 16:55 -0500, Fionnuala Gunter wrote:
> > > Sorry, I thought linuxppc-dev was the right tree since the patch is
> > > for
> > > a ppc driver and you've carried previous nx driver submissions
> > > upstream. 
> > > 
> > > I want to submit to the proper place, so what is your preference?
> > 
> > Well, that's the right place if you want me to merge it, which seems to
> > be what Marcelo is also expecting...
> 
> Please merge this patch, it fixes a hot bug.

You don't actually mention that anywhere in the changelog.

The description implies that it's probably a bug, but doesn't describe
the symptoms in any way - does it make it completely non-functional or
just break in certain corner cases.

And you make no mention of when this bug appeared, has it always been
there, or was it introduced by a particular commit? That's important
because it helps determine whether the patch should go into any of the
stable trees.

cheers

Re: [PATCH] drivers/crypto/nx: saves chaining value from co-processor

[Fionnuala Gunter]

On Sat, 2013-08-03 at 07:59 +1000, Benjamin Herrenschmidt wrote:
> On Fri, 2013-08-02 at 16:55 -0500, Fionnuala Gunter wrote:
> > Sorry, I thought linuxppc-dev was the right tree since the patch is
> > for
> > a ppc driver and you've carried previous nx driver submissions
> > upstream. 
> > 
> > I want to submit to the proper place, so what is your preference?
> 
> Well, that's the right place if you want me to merge it, which seems to
> be what Marcelo is also expecting...

Please merge this patch, it fixes a hot bug.

> However I do wonder why is that expectation there in the first place and
> why aren't those handled by the crypto maintainer...

In the future I can submit patches to the crypto maintainer.

-Fin

Re: [PATCH] drivers/crypto/nx: saves chaining value from co-processor

[Benjamin Herrenschmidt]

On Fri, 2013-08-02 at 16:55 -0500, Fionnuala Gunter wrote:
> Sorry, I thought linuxppc-dev was the right tree since the patch is
> for
> a ppc driver and you've carried previous nx driver submissions
> upstream. 
> 
> I want to submit to the proper place, so what is your preference?

Well, that's the right place if you want me to merge it, which seems to
be what Marcelo is also expecting...

However I do wonder why is that expectation there in the first place and
why aren't those handled by the crypto maintainer...

Cheers,
Ben.

Re: [PATCH] drivers/crypto/nx: saves chaining value from co-processor

[Fionnuala Gunter]

On Sat, 2013-08-03 at 07:09 +1000, Benjamin Herrenschmidt wrote:
> On Fri, 2013-08-02 at 15:58 -0500, Fionnuala Gunter wrote:
> > The chaining value from co-processor was not being saved. This value is
> > needed because it is used as the IV, for example by cts i.e.
> > cts(cbc(aes)).
> > 
> > Signed-off-by: Fionnuala Gunter <fin@linux.vnet.ibm.com>
> > Signed-off-by: Marcelo Cerri <mhcerri@linux.vnet.ibm.com>
> 
> So now you guys are sending the patch to linuxppc-dev and not the crypto
> list ..
> 
> I need to understand things better here. Any reason why those drivers
> aren't handled by the crypto maintainer ?
> 
> Ben.

Sorry, I thought linuxppc-dev was the right tree since the patch is for
a ppc driver and you've carried previous nx driver submissions
upstream. 

I want to submit to the proper place, so what is your preference?

-Fin

Re: [PATCH] drivers/crypto/nx: saves chaining value from co-processor

[Benjamin Herrenschmidt]

On Fri, 2013-08-02 at 15:58 -0500, Fionnuala Gunter wrote:
> The chaining value from co-processor was not being saved. This value is
> needed because it is used as the IV, for example by cts i.e.
> cts(cbc(aes)).
> 
> Signed-off-by: Fionnuala Gunter <fin@linux.vnet.ibm.com>
> Signed-off-by: Marcelo Cerri <mhcerri@linux.vnet.ibm.com>

So now you guys are sending the patch to linuxppc-dev and not the crypto
list ..

I need to understand things better here. Any reason why those drivers
aren't handled by the crypto maintainer ?

Ben.

> ---
>  drivers/crypto/nx/nx-aes-cbc.c |    1 +
>  1 file changed, 1 insertion(+)
> 
> diff --git a/drivers/crypto/nx/nx-aes-cbc.c
> b/drivers/crypto/nx/nx-aes-cbc.c
> index 35d483f..a2f99a9 100644
> --- a/drivers/crypto/nx/nx-aes-cbc.c
> +++ b/drivers/crypto/nx/nx-aes-cbc.c
> @@ -95,6 +95,7 @@ static int cbc_aes_nx_crypt(struct blkcipher_desc
> *desc,
>         if (rc)
>                 goto out;
> 
> +       memcpy(desc->info, csbcpb->cpb.aes_cbc.cv, AES_BLOCK_SIZE);
>         atomic_inc(&(nx_ctx->stats->aes_ops));
>         atomic64_add(csbcpb->csb.processed_byte_count,
>                      &(nx_ctx->stats->aes_bytes));

[PATCH] drivers/crypto/nx: saves chaining value from co-processor

[Fionnuala Gunter]

The chaining value from co-processor was not being saved. This value is
needed because it is used as the IV, for example by cts i.e.
cts(cbc(aes)).

Signed-off-by: Fionnuala Gunter <fin@linux.vnet.ibm.com>
Signed-off-by: Marcelo Cerri <mhcerri@linux.vnet.ibm.com>
---
 drivers/crypto/nx/nx-aes-cbc.c |    1 +
 1 file changed, 1 insertion(+)

diff --git a/drivers/crypto/nx/nx-aes-cbc.c
b/drivers/crypto/nx/nx-aes-cbc.c
index 35d483f..a2f99a9 100644
--- a/drivers/crypto/nx/nx-aes-cbc.c
+++ b/drivers/crypto/nx/nx-aes-cbc.c
@@ -95,6 +95,7 @@ static int cbc_aes_nx_crypt(struct blkcipher_desc
*desc,
        if (rc)
                goto out;

+       memcpy(desc->info, csbcpb->cpb.aes_cbc.cv, AES_BLOCK_SIZE);
        atomic_inc(&(nx_ctx->stats->aes_ops));
        atomic64_add(csbcpb->csb.processed_byte_count,
                     &(nx_ctx->stats->aes_bytes));
-- 
1.7.10.4