2020-11-06 09:05:22

by 王擎

[permalink] [raw]
Subject: [PATCH] crypto: Fix return value check in aead_crypt()

Fix passing zero to 'PTR_ERR' warning

Signed-off-by: Wang Qing <[email protected]>
---
drivers/crypto/caam/caamalg_qi.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/crypto/caam/caamalg_qi.c b/drivers/crypto/caam/caamalg_qi.c
index 66f60d7..add60e8
--- a/drivers/crypto/caam/caamalg_qi.c
+++ b/drivers/crypto/caam/caamalg_qi.c
@@ -1166,7 +1166,7 @@ static inline int aead_crypt(struct aead_request *req, bool encrypt)
/* allocate extended descriptor */
edesc = aead_edesc_alloc(req, encrypt);
if (IS_ERR_OR_NULL(edesc))
- return PTR_ERR(edesc);
+ return PTR_ERR_OR_ZERO(edesc);

/* Create and submit job descriptor */
ret = caam_qi_enqueue(ctx->qidev, &edesc->drv_req);
--
2.7.4


2020-11-12 09:22:36

by Horia Geanta

[permalink] [raw]
Subject: crypto: caam/qi - simplify error path for context allocation

On 11/6/2020 11:01 AM, Wang Qing wrote:
> Fix passing zero to 'PTR_ERR' warning
>
Thanks.

> Signed-off-by: Wang Qing <[email protected]>
> ---
> drivers/crypto/caam/caamalg_qi.c | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/drivers/crypto/caam/caamalg_qi.c b/drivers/crypto/caam/caamalg_qi.c
> index 66f60d7..add60e8
> --- a/drivers/crypto/caam/caamalg_qi.c
> +++ b/drivers/crypto/caam/caamalg_qi.c
> @@ -1166,7 +1166,7 @@ static inline int aead_crypt(struct aead_request *req, bool encrypt)
> /* allocate extended descriptor */
> edesc = aead_edesc_alloc(req, encrypt);
> if (IS_ERR_OR_NULL(edesc))
> - return PTR_ERR(edesc);
> + return PTR_ERR_OR_ZERO(edesc);
>
Digging a bit into the logic, it turns out aead_edesc_alloc() can't return
a NULL pointer.

-- >8 --

Subject: [PATCH] crypto: caam/qi - simplify error path for context allocation

Wang Qing reports that IS_ERR_OR_NULL() should be matched with
PTR_ERR_OR_ZERO(), not PTR_ERR().

As it turns out, the error path always returns an error code,
i.e. NULL is never returned.
Update the code accordingly - s/IS_ERR_OR_NULL/IS_ERR.

Reported-by: Wang Qing <[email protected]>
Signed-off-by: Horia Geantă <[email protected]>
---
drivers/crypto/caam/caamalg_qi.c | 8 ++++----
1 file changed, 4 insertions(+), 4 deletions(-)

diff --git a/drivers/crypto/caam/caamalg_qi.c b/drivers/crypto/caam/caamalg_qi.c
index a24ae966df4a..189a7438b29c 100644
--- a/drivers/crypto/caam/caamalg_qi.c
+++ b/drivers/crypto/caam/caamalg_qi.c
@@ -852,7 +852,7 @@ static struct caam_drv_ctx *get_drv_ctx(struct caam_ctx *ctx,

cpu = smp_processor_id();
drv_ctx = caam_drv_ctx_init(ctx->qidev, &cpu, desc);
- if (!IS_ERR_OR_NULL(drv_ctx))
+ if (!IS_ERR(drv_ctx))
drv_ctx->op_type = type;

ctx->drv_ctx[type] = drv_ctx;
@@ -955,7 +955,7 @@ static struct aead_edesc *aead_edesc_alloc(struct aead_request *req,
struct caam_drv_ctx *drv_ctx;

drv_ctx = get_drv_ctx(ctx, encrypt ? ENCRYPT : DECRYPT);
- if (IS_ERR_OR_NULL(drv_ctx))
+ if (IS_ERR(drv_ctx))
return (struct aead_edesc *)drv_ctx;

/* allocate space for base edesc and hw desc commands, link tables */
@@ -1165,7 +1165,7 @@ static inline int aead_crypt(struct aead_request *req, bool encrypt)

/* allocate extended descriptor */
edesc = aead_edesc_alloc(req, encrypt);
- if (IS_ERR_OR_NULL(edesc))
+ if (IS_ERR(edesc))
return PTR_ERR(edesc);

/* Create and submit job descriptor */
@@ -1259,7 +1259,7 @@ static struct skcipher_edesc *skcipher_edesc_alloc(struct skcipher_request *req,
struct caam_drv_ctx *drv_ctx;

drv_ctx = get_drv_ctx(ctx, encrypt ? ENCRYPT : DECRYPT);
- if (IS_ERR_OR_NULL(drv_ctx))
+ if (IS_ERR(drv_ctx))
return (struct skcipher_edesc *)drv_ctx;

src_nents = sg_nents_for_len(req->src, req->cryptlen);
--
2.17.1

2020-11-13 03:40:07

by Herbert Xu

[permalink] [raw]
Subject: Re: crypto: caam/qi - simplify error path for context allocation

Horia Geantă <[email protected]> wrote:
>
> -- >8 --
>
> Subject: [PATCH] crypto: caam/qi - simplify error path for context allocation

You can't do this. Patchwork takes any replies with the same
Subject line as a comment.

You need to resend this patch with a different subject.

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

2020-11-13 03:42:58

by Herbert Xu

[permalink] [raw]
Subject: Re: crypto: caam/qi - simplify error path for context allocation

On Fri, Nov 13, 2020 at 02:39:24PM +1100, Herbert Xu wrote:
> Horia Geantă <[email protected]> wrote:
> >
> > -- >8 --
> >
> > Subject: [PATCH] crypto: caam/qi - simplify error path for context allocation
>
> You can't do this. Patchwork takes any replies with the same
> Subject line as a comment.
>
> You need to resend this patch with a different subject.

Nevermind, you did change the email subject too.

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

2020-11-20 06:57:48

by Herbert Xu

[permalink] [raw]
Subject: Re: crypto: caam/qi - simplify error path for context allocation

On Thu, Nov 12, 2020 at 11:21:46AM +0200, Horia Geantă wrote:
>
> Wang Qing reports that IS_ERR_OR_NULL() should be matched with
> PTR_ERR_OR_ZERO(), not PTR_ERR().
>
> As it turns out, the error path always returns an error code,
> i.e. NULL is never returned.
> Update the code accordingly - s/IS_ERR_OR_NULL/IS_ERR.
>
> Reported-by: Wang Qing <[email protected]>
> Signed-off-by: Horia Geantă <[email protected]>
> ---
> drivers/crypto/caam/caamalg_qi.c | 8 ++++----
> 1 file changed, 4 insertions(+), 4 deletions(-)

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