Hello Mike Marciniszyn,
The patch 45842abbb292: "staging/rdma/hfi1: move txreq header code"
from Feb 14, 2016, leads to the following static checker warning:
drivers/infiniband/hw/hfi1/verbs_txreq.h:89 get_txreq()
warn: 'tx' can also be NULL
drivers/infiniband/hw/hfi1/verbs_txreq.h
75 static inline struct verbs_txreq *get_txreq(struct hfi1_ibdev *dev,
76 struct rvt_qp *qp)
77 __must_hold(&qp->slock)
78 {
79 struct verbs_txreq *tx;
80 struct hfi1_qp_priv *priv = qp->priv;
81
82 tx = kmem_cache_alloc(dev->verbs_txreq_cache, GFP_ATOMIC);
83 if (unlikely(!tx)) {
84 /* call slow path to get the lock */
85 tx = __get_txreq(dev, qp);
^^^^^^^^^^^^^^^^^^^^
The problem is that __get_txreq() returns a mix of error pointers and
NULL. I skimmed the function but it's not totally clear to me how it's
supposed to work.
86 if (IS_ERR(tx))
87 return tx;
88 }
89 tx->qp = qp;
90 tx->mr = NULL;
91 tx->sde = priv->s_sde;
92 tx->psc = priv->s_sendcontext;
93 /* so that we can test if the sdma decriptors are there */
94 tx->txreq.num_desc = 0;
95 /* Set the header type */
96 tx->phdr.hdr.hdr_type = priv->hdr_type;
97 return tx;
98 }
regards,
dan carpenter
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
