Re: [RFC PATCH 1/2] crypto: caam - properly set IV after {en,de}crypt

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



On 6/2/2017 3:25 PM, David Gstir wrote:
> Certain cipher modes like CTS expect the IV (req->info) of
> ablkcipher_request (or equivalently req->iv of skcipher_request) to
> contain the last ciphertext block when the {en,de}crypt operation is done.
> This is currently not the case for the CAAM driver which in turn breaks
> e.g. cts(cbc(aes)) when the CAAM driver is enabled.
> 
> This patch fixes the CAAM driver to properly set the IV after the
> {en,de}crypt operation of ablkcipher finishes.
> 
> Signed-off-by: David Gstir <david@xxxxxxxxxxxxx>
> ---
>  drivers/crypto/caam/caamalg.c | 26 ++++++++++++++++++++++++--
>  1 file changed, 24 insertions(+), 2 deletions(-)
> 
> diff --git a/drivers/crypto/caam/caamalg.c b/drivers/crypto/caam/caamalg.c
> index 398807d1b77e..d13c1aee4427 100644
> --- a/drivers/crypto/caam/caamalg.c
> +++ b/drivers/crypto/caam/caamalg.c
> @@ -882,10 +882,11 @@ static void ablkcipher_encrypt_done(struct device *jrdev, u32 *desc, u32 err,
>  {
>  	struct ablkcipher_request *req = context;
>  	struct ablkcipher_edesc *edesc;
> -#ifdef DEBUG
>  	struct crypto_ablkcipher *ablkcipher = crypto_ablkcipher_reqtfm(req);
>  	int ivsize = crypto_ablkcipher_ivsize(ablkcipher);
> +	int nents;
>  
> +#ifdef DEBUG
>  	dev_err(jrdev, "%s %d: err 0x%x\n", __func__, __LINE__, err);
>  #endif
>  
> @@ -904,6 +905,19 @@ static void ablkcipher_encrypt_done(struct device *jrdev, u32 *desc, u32 err,
>  #endif
>  
>  	ablkcipher_unmap(jrdev, edesc, req);
> +
> +	if (req->src == req->dst)
> +		nents = edesc->src_nents;
> +	else
> +		nents = edesc->dst_nents;
> +
> +	/*
> +	 * The crypto API expects us to set the IV (req->info) to the last
> +	 * ciphertext block. This is used e.g. by the CTS mode.
> +	 */

IIUC, IV update is required only in case of CBC.
Since this callback is used also for CTR, we should avoid the copy:
if ((ctx->cdata.algtype & OP_ALG_AAI_MASK) == OP_ALG_AAI_CBC) ...

> +	sg_pcopy_to_buffer(req->dst, nents, req->info, ivsize,
> +			   req->nbytes - ivsize);

scatterwalk_map_and_copy() should be used instead.

> +
>  	kfree(edesc);
>  
>  	ablkcipher_request_complete(req, err);
> @@ -914,10 +928,10 @@ static void ablkcipher_decrypt_done(struct device *jrdev, u32 *desc, u32 err,
>  {
>  	struct ablkcipher_request *req = context;
>  	struct ablkcipher_edesc *edesc;
> -#ifdef DEBUG
>  	struct crypto_ablkcipher *ablkcipher = crypto_ablkcipher_reqtfm(req);
>  	int ivsize = crypto_ablkcipher_ivsize(ablkcipher);
>  
> +#ifdef DEBUG
>  	dev_err(jrdev, "%s %d: err 0x%x\n", __func__, __LINE__, err);
>  #endif
>  
> @@ -935,6 +949,14 @@ static void ablkcipher_decrypt_done(struct device *jrdev, u32 *desc, u32 err,
>  #endif
>  
>  	ablkcipher_unmap(jrdev, edesc, req);
> +
> +	/*
> +	 * The crypto API expects us to set the IV (req->info) to the last
> +	 * ciphertext block.
> +	 */
> +	sg_pcopy_to_buffer(req->src, edesc->src_nents, req->info, ivsize,
> +			   req->nbytes - ivsize);
> +
>  	kfree(edesc);
>  
>  	ablkcipher_request_complete(req, err);
>




[Index of Archives]     [Kernel]     [Gnu Classpath]     [Gnu Crypto]     [DM Crypt]     [Netfilter]     [Bugtraq]

  Powered by Linux