[PATCH] crypto: caam/qi2: remove redundant assignment to ret

[PATCH] crypto: caam/qi2: remove redundant assignment to ret

[Colin King]

From: Colin Ian King <colin.king@canonical.com>

The variable ret is being assigned a value that is never read, the
error exit path via label 'unmap' returns -ENOMEM anyhow, so assigning
ret with -ENOMEM is redundamt.

Addresses-Coverity: ("Unused value")
Signed-off-by: Colin Ian King <colin.king@canonical.com>
---
 drivers/crypto/caam/caamalg_qi2.c | 2 --
 1 file changed, 2 deletions(-)

diff --git a/drivers/crypto/caam/caamalg_qi2.c b/drivers/crypto/caam/caamalg_qi2.c
index 28669cbecf77..ef2c4e095db3 100644
--- a/drivers/crypto/caam/caamalg_qi2.c
+++ b/drivers/crypto/caam/caamalg_qi2.c
@@ -4044,7 +4044,6 @@ static int ahash_finup_no_ctx(struct ahash_request *req)
 					  DMA_TO_DEVICE);
 	if (dma_mapping_error(ctx->dev, edesc->qm_sg_dma)) {
 		dev_err(ctx->dev, "unable to map S/G table\n");
-		ret = -ENOMEM;
 		goto unmap;
 	}
 	edesc->qm_sg_bytes = qm_sg_bytes;
@@ -4055,7 +4054,6 @@ static int ahash_finup_no_ctx(struct ahash_request *req)
 	if (dma_mapping_error(ctx->dev, state->ctx_dma)) {
 		dev_err(ctx->dev, "unable to map ctx\n");
 		state->ctx_dma = 0;
-		ret = -ENOMEM;
 		goto unmap;
 	}
 
-- 
2.27.0.rc0

Re: [PATCH] crypto: caam/qi2: remove redundant assignment to ret

[Herbert Xu]

On Thu, Jun 11, 2020 at 04:39:34PM +0100, Colin King wrote:
> From: Colin Ian King <colin.king@canonical.com>
> 
> The variable ret is being assigned a value that is never read, the
> error exit path via label 'unmap' returns -ENOMEM anyhow, so assigning
> ret with -ENOMEM is redundamt.
> 
> Addresses-Coverity: ("Unused value")
> Signed-off-by: Colin Ian King <colin.king@canonical.com>
> ---
>  drivers/crypto/caam/caamalg_qi2.c | 2 --
>  1 file changed, 2 deletions(-)

Patch applied.  Thanks.
-- 
Email: Herbert Xu <herbert@gondor.apana.org.au>
Home Page: http://gondor.apana.org.au/~herbert/
PGP Key: http://gondor.apana.org.au/~herbert/pubkey.txt

Re: [PATCH] crypto: caam/qi2: remove redundant assignment to ret

[Horia Geantă]

On 6/18/2020 10:58 AM, Herbert Xu wrote:
> On Thu, Jun 11, 2020 at 04:39:34PM +0100, Colin King wrote:
>> From: Colin Ian King <colin.king@canonical.com>
>>
>> The variable ret is being assigned a value that is never read, the
>> error exit path via label 'unmap' returns -ENOMEM anyhow, so assigning
>> ret with -ENOMEM is redundamt.
>>
>> Addresses-Coverity: ("Unused value")
>> Signed-off-by: Colin Ian King <colin.king@canonical.com>
>> ---
>>  drivers/crypto/caam/caamalg_qi2.c | 2 --
>>  1 file changed, 2 deletions(-)
> 
> Patch applied.  Thanks.
> 
Unfortunately I missed this patch, and it doesn't look correct.

Do I need to send a revert?

Thanks,
Horia

Re: [PATCH] crypto: caam/qi2: remove redundant assignment to ret

[Herbert Xu]

On Thu, Jun 18, 2020 at 01:40:55PM +0300, Horia Geantă wrote:
> On 6/18/2020 10:58 AM, Herbert Xu wrote:
> > On Thu, Jun 11, 2020 at 04:39:34PM +0100, Colin King wrote:
> >> From: Colin Ian King <colin.king@canonical.com>
> >>
> >> The variable ret is being assigned a value that is never read, the
> >> error exit path via label 'unmap' returns -ENOMEM anyhow, so assigning
> >> ret with -ENOMEM is redundamt.
> >>
> >> Addresses-Coverity: ("Unused value")
> >> Signed-off-by: Colin Ian King <colin.king@canonical.com>
> >> ---
> >>  drivers/crypto/caam/caamalg_qi2.c | 2 --
> >>  1 file changed, 2 deletions(-)
> > 
> > Patch applied.  Thanks.
> > 
> Unfortunately I missed this patch, and it doesn't look correct.
> 
> Do I need to send a revert?

No please check again.  The patch is correct.

Thanks,
-- 
Email: Herbert Xu <herbert@gondor.apana.org.au>
Home Page: http://gondor.apana.org.au/~herbert/
PGP Key: http://gondor.apana.org.au/~herbert/pubkey.txt

Re: [PATCH] crypto: caam/qi2: remove redundant assignment to ret

[Horia Geantă]

On 6/11/2020 6:39 PM, Colin King wrote:
> From: Colin Ian King <colin.king@canonical.com>
> 
> The variable ret is being assigned a value that is never read, the
> error exit path via label 'unmap' returns -ENOMEM anyhow, so assigning
> ret with -ENOMEM is redundamt.
> 
> Addresses-Coverity: ("Unused value")
> Signed-off-by: Colin Ian King <colin.king@canonical.com>
> ---
>  drivers/crypto/caam/caamalg_qi2.c | 2 --
>  1 file changed, 2 deletions(-)
> 
> diff --git a/drivers/crypto/caam/caamalg_qi2.c b/drivers/crypto/caam/caamalg_qi2.c
> index 28669cbecf77..ef2c4e095db3 100644
> --- a/drivers/crypto/caam/caamalg_qi2.c
> +++ b/drivers/crypto/caam/caamalg_qi2.c
> @@ -4044,7 +4044,6 @@ static int ahash_finup_no_ctx(struct ahash_request *req)
>  					  DMA_TO_DEVICE);
>  	if (dma_mapping_error(ctx->dev, edesc->qm_sg_dma)) {
>  		dev_err(ctx->dev, "unable to map S/G table\n");
> -		ret = -ENOMEM;
>  		goto unmap;
>  	}
>  	edesc->qm_sg_bytes = qm_sg_bytes;
> @@ -4055,7 +4054,6 @@ static int ahash_finup_no_ctx(struct ahash_request *req)
>  	if (dma_mapping_error(ctx->dev, state->ctx_dma)) {
>  		dev_err(ctx->dev, "unable to map ctx\n");
>  		state->ctx_dma = 0;
> -		ret = -ENOMEM;
>  		goto unmap;
>  	}
>  
The proper fix would be updating the ahash_finup_no_ctx() function
to return the specific error code:
	return ret;
instead of returning -ENOMEM for all error cases.

For example error code returned by dpaa2_caam_enqueue()
should be returned instead of -ENOMEM.

Horia

Re: [PATCH] crypto: caam/qi2: remove redundant assignment to ret

[Herbert Xu]

On Thu, Jun 18, 2020 at 01:54:55PM +0300, Horia Geantă wrote:
>
> The proper fix would be updating the ahash_finup_no_ctx() function
> to return the specific error code:
> 	return ret;
> instead of returning -ENOMEM for all error cases.
> 
> For example error code returned by dpaa2_caam_enqueue()
> should be returned instead of -ENOMEM.

You can do that as a follow-up.  The patch is correct as is.

Cheers,
-- 
Email: Herbert Xu <herbert@gondor.apana.org.au>
Home Page: http://gondor.apana.org.au/~herbert/
PGP Key: http://gondor.apana.org.au/~herbert/pubkey.txt

Re: [PATCH] crypto: caam/qi2: remove redundant assignment to ret

[Horia Geantă]

On 6/18/2020 2:00 PM, Herbert Xu wrote:
> On Thu, Jun 18, 2020 at 01:54:55PM +0300, Horia Geantă wrote:
>>
>> The proper fix would be updating the ahash_finup_no_ctx() function
>> to return the specific error code:
>> 	return ret;
>> instead of returning -ENOMEM for all error cases.
>>
>> For example error code returned by dpaa2_caam_enqueue()
>> should be returned instead of -ENOMEM.
> 
> You can do that as a follow-up.  The patch is correct as is.
> 
Just that the follow-up implies adding all the code back.

Anyway, not a big deal...

Thanks,
Horia

[PATCH] crypto: caam/qi2 - fix return code in ahash_finup_no_ctx()

[Horia Geantă]

ahash_finup_no_ctx() returns -ENOMEM in most error cases,
and this is fine for almost all of them.

However, the return code provided by dpaa2_caam_enqueue()
(e.g. -EIO or -EBUSY) shouldn't be overridden by -ENOMEM.

Signed-off-by: Horia Geantă <horia.geanta@nxp.com>
---
 drivers/crypto/caam/caamalg_qi2.c | 10 ++++++----
 1 file changed, 6 insertions(+), 4 deletions(-)

diff --git a/drivers/crypto/caam/caamalg_qi2.c b/drivers/crypto/caam/caamalg_qi2.c
index 1e90412afea2..45e9ff851e2d 100644
--- a/drivers/crypto/caam/caamalg_qi2.c
+++ b/drivers/crypto/caam/caamalg_qi2.c
@@ -4004,7 +4004,7 @@ static int ahash_finup_no_ctx(struct ahash_request *req)
 	int digestsize = crypto_ahash_digestsize(ahash);
 	struct ahash_edesc *edesc;
 	struct dpaa2_sg_entry *sg_table;
-	int ret;
+	int ret = -ENOMEM;
 
 	src_nents = sg_nents_for_len(req->src, req->nbytes);
 	if (src_nents < 0) {
@@ -4017,7 +4017,7 @@ static int ahash_finup_no_ctx(struct ahash_request *req)
 					  DMA_TO_DEVICE);
 		if (!mapped_nents) {
 			dev_err(ctx->dev, "unable to DMA map source\n");
-			return -ENOMEM;
+			return ret;
 		}
 	} else {
 		mapped_nents = 0;
@@ -4027,7 +4027,7 @@ static int ahash_finup_no_ctx(struct ahash_request *req)
 	edesc = qi_cache_zalloc(GFP_DMA | flags);
 	if (!edesc) {
 		dma_unmap_sg(ctx->dev, req->src, src_nents, DMA_TO_DEVICE);
-		return -ENOMEM;
+		return ret;
 	}
 
 	edesc->src_nents = src_nents;
@@ -4044,6 +4044,7 @@ static int ahash_finup_no_ctx(struct ahash_request *req)
 					  DMA_TO_DEVICE);
 	if (dma_mapping_error(ctx->dev, edesc->qm_sg_dma)) {
 		dev_err(ctx->dev, "unable to map S/G table\n");
+		ret = -ENOMEM;
 		goto unmap;
 	}
 	edesc->qm_sg_bytes = qm_sg_bytes;
@@ -4054,6 +4055,7 @@ static int ahash_finup_no_ctx(struct ahash_request *req)
 	if (dma_mapping_error(ctx->dev, state->ctx_dma)) {
 		dev_err(ctx->dev, "unable to map ctx\n");
 		state->ctx_dma = 0;
+		ret = -ENOMEM;
 		goto unmap;
 	}
 
@@ -4080,7 +4082,7 @@ static int ahash_finup_no_ctx(struct ahash_request *req)
 unmap:
 	ahash_unmap_ctx(ctx->dev, edesc, req, DMA_FROM_DEVICE);
 	qi_cache_free(edesc);
-	return -ENOMEM;
+	return ret;
 }
 
 static int ahash_update_first(struct ahash_request *req)
-- 
2.17.1

Re: [PATCH] crypto: caam/qi2 - fix return code in ahash_finup_no_ctx()

[Herbert Xu]

On Fri, Jun 19, 2020 at 04:22:53PM +0300, Horia Geantă wrote:
> ahash_finup_no_ctx() returns -ENOMEM in most error cases,
> and this is fine for almost all of them.
> 
> However, the return code provided by dpaa2_caam_enqueue()
> (e.g. -EIO or -EBUSY) shouldn't be overridden by -ENOMEM.
> 
> Signed-off-by: Horia Geantă <horia.geanta@nxp.com>
> ---
>  drivers/crypto/caam/caamalg_qi2.c | 10 ++++++----
>  1 file changed, 6 insertions(+), 4 deletions(-)

Patch applied.  Thanks.
-- 
Email: Herbert Xu <herbert@gondor.apana.org.au>
Home Page: http://gondor.apana.org.au/~herbert/
PGP Key: http://gondor.apana.org.au/~herbert/pubkey.txt