On Thu, 23 Feb 2017, 11:44am -0000, Hannes Reinecke wrote:
> The driver didn't follow the atomic_t vs refcount_t change, and
> anyway one should be using kref_read() instead of accessing the
> counter inside an kref.
>
> Fixes: 61d8658b4a435e ("scsi: qedf: Add QLogic FastLinQ offload FCoE driver
> framework.)
> Cc: Chad Dupuis <chad.dup...@cavium.com>
> Cc: Nilesh Javali <nilesh.jav...@cavium.com>
> Signed-off-by: Hannes Reinecke <h...@suse.com>
> ---
> drivers/scsi/qedf/qedf_io.c | 8 ++++----
> 1 file changed, 4 insertions(+), 4 deletions(-)
>
> diff --git a/drivers/scsi/qedf/qedf_io.c b/drivers/scsi/qedf/qedf_io.c
> index 486c045..ee0dcf9 100644
> --- a/drivers/scsi/qedf/qedf_io.c
> +++ b/drivers/scsi/qedf/qedf_io.c
> @@ -998,7 +998,7 @@ static void qedf_trace_io(struct qedf_rport *fcport,
> struct qedf_ioreq *io_req,
> io_log->sg_count = scsi_sg_count(sc_cmd);
> io_log->result = sc_cmd->result;
> io_log->jiffies = jiffies;
> - io_log->refcount = atomic_read(&io_req->refcount.refcount);
> + io_log->refcount = kref_read(&io_req->refcount);
>
> if (direction == QEDF_IO_TRACE_REQ) {
> /* For requests we only care abot the submission CPU */
> @@ -1340,7 +1340,7 @@ void qedf_scsi_completion(struct qedf_ctx *qedf, struct
> fcoe_cqe *cqe,
> /* Good I/O completion */
> sc_cmd->result = DID_OK << 16;
> } else {
> - refcount = atomic_read(&io_req->refcount.refcount);
> + refcount = kref_read(&io_req->refcount);
> QEDF_INFO(&(qedf->dbg_ctx), QEDF_LOG_IO,
> "%d:0:%d:%d xid=0x%0x op=0x%02x "
> "lba=%02x%02x%02x%02x cdb_status=%d "
> @@ -1425,7 +1425,7 @@ void qedf_scsi_done(struct qedf_ctx *qedf, struct
> qedf_ioreq *io_req,
> qedf_unmap_sg_list(qedf, io_req);
>
> sc_cmd->result = result << 16;
> - refcount = atomic_read(&io_req->refcount.refcount);
> + refcount = kref_read(&io_req->refcount);
> QEDF_INFO(&(qedf->dbg_ctx), QEDF_LOG_IO, "%d:0:%d:%d: Completing "
> "sc_cmd=%p result=0x%08x op=0x%02x lba=0x%02x%02x%02x%02x, "
> "allowed=%d retries=%d refcount=%d.\n",
> @@ -1556,7 +1556,7 @@ static void qedf_flush_els_req(struct qedf_ctx *qedf,
> {
> QEDF_INFO(&(qedf->dbg_ctx), QEDF_LOG_IO,
> "Flushing ELS request xid=0x%x refcount=%d.\n", els_req->xid,
> - atomic_read(&els_req->refcount.refcount));
> + kref_read(&els_req->refcount));
>
> /*
> * Need to distinguish this from a timeout when calling the
>
Thanks for the patch Hannes though it looks like the atomic_read's in
qedf_els.c need to be converted as well. Will post a patch shortly.
