2021-03-25 03:09:34

by Yue Haibing

Subject: [PATCH -next] IB/srpt: Fix passing zero to 'PTR_ERR'

Fix smatch warning:

drivers/infiniband/ulp/srpt/ib_srpt.c:2341 srpt_cm_req_recv() warn: passing zero to 'PTR_ERR'

Use PTR_ERR_OR_ZERO instead of PTR_ERR

Fixes: 847462de3a0a ("IB/srpt: Fix srpt_cm_req_recv() error path (1/2)")
Signed-off-by: YueHaibing <[email protected]>
---
drivers/infiniband/ulp/srpt/ib_srpt.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/infiniband/ulp/srpt/ib_srpt.c b/drivers/infiniband/ulp/srpt/ib_srpt.c
index 6be60aa5ffe2..3ff24b5048ac 100644
--- a/drivers/infiniband/ulp/srpt/ib_srpt.c
+++ b/drivers/infiniband/ulp/srpt/ib_srpt.c
@@ -2338,7 +2338,7 @@ static int srpt_cm_req_recv(struct srpt_device *const sdev,

if (IS_ERR_OR_NULL(ch->sess)) {
WARN_ON_ONCE(ch->sess == NULL);
- ret = PTR_ERR(ch->sess);
+ ret = PTR_ERR_OR_ZERO(ch->sess);
ch->sess = NULL;
pr_info("Rejected login for initiator %s: ret = %d.\n",
ch->sess_name, ret);
--
2.22.0

2021-04-01 18:36:35

by Jason Gunthorpe

Subject: Re: [PATCH -next] IB/srpt: Fix passing zero to 'PTR_ERR'

On Wed, Mar 24, 2021 at 10:09:39PM +0800, YueHaibing wrote:
> Fix smatch warning:
>
> drivers/infiniband/ulp/srpt/ib_srpt.c:2341 srpt_cm_req_recv() warn: passing zero to 'PTR_ERR'
>
> Use PTR_ERR_OR_ZERO instead of PTR_ERR
>
> Fixes: 847462de3a0a ("IB/srpt: Fix srpt_cm_req_recv() error path (1/2)")
> Signed-off-by: YueHaibing <[email protected]>
> drivers/infiniband/ulp/srpt/ib_srpt.c | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/drivers/infiniband/ulp/srpt/ib_srpt.c b/drivers/infiniband/ulp/srpt/ib_srpt.c
> index 6be60aa5ffe2..3ff24b5048ac 100644
> +++ b/drivers/infiniband/ulp/srpt/ib_srpt.c
> @@ -2338,7 +2338,7 @@ static int srpt_cm_req_recv(struct srpt_device *const sdev,
>
> if (IS_ERR_OR_NULL(ch->sess)) {
> WARN_ON_ONCE(ch->sess == NULL);
> - ret = PTR_ERR(ch->sess);
> + ret = PTR_ERR_OR_ZERO(ch->sess);

This whole flow is a mess, if someone fixes properly then fine, but
I'm not convinced returning 0 here is correct either.

Jason

2021-04-01 19:16:45

by Bart Van Assche

Subject: Re: [PATCH -next] IB/srpt: Fix passing zero to 'PTR_ERR'

On 3/24/21 7:09 AM, YueHaibing wrote:
> Fix smatch warning:
>
> drivers/infiniband/ulp/srpt/ib_srpt.c:2341 srpt_cm_req_recv() warn: passing zero to 'PTR_ERR'
>
> Use PTR_ERR_OR_ZERO instead of PTR_ERR
>
> Fixes: 847462de3a0a ("IB/srpt: Fix srpt_cm_req_recv() error path (1/2)")
> Signed-off-by: YueHaibing <[email protected]>
> ---
> drivers/infiniband/ulp/srpt/ib_srpt.c | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/drivers/infiniband/ulp/srpt/ib_srpt.c b/drivers/infiniband/ulp/srpt/ib_srpt.c
> index 6be60aa5ffe2..3ff24b5048ac 100644
> --- a/drivers/infiniband/ulp/srpt/ib_srpt.c
> +++ b/drivers/infiniband/ulp/srpt/ib_srpt.c
> @@ -2338,7 +2338,7 @@ static int srpt_cm_req_recv(struct srpt_device *const sdev,
>
> if (IS_ERR_OR_NULL(ch->sess)) {
> WARN_ON_ONCE(ch->sess == NULL);
> - ret = PTR_ERR(ch->sess);
> + ret = PTR_ERR_OR_ZERO(ch->sess);
> ch->sess = NULL;
> pr_info("Rejected login for initiator %s: ret = %d.\n",
> ch->sess_name, ret);

(just noticed this patch)

target_setup_session() should either return a valid session pointer or
an ERR_PTR() but not NULL. Changing IS_ERR_OR_NULL() into IS_ERR() and
removing the WARN_ON_ONCE(ch->sess == NULL) may be a better solution.

Thanks,

Bart.