On 9/21/18 5:15 AM, Hannes Reinecke wrote:
The WARN_ON() is pointless as the rport is placed in SDEV_TRANSPORT_OFFLINE
at that time, so no new commands can be submitted via srp_queuecomment()
Signed-off-by: Hannes Reinecke <hare@xxxxxxxx>
---
drivers/infiniband/ulp/srp/ib_srp.c | 7 -------
1 file changed, 7 deletions(-)
diff --git a/drivers/infiniband/ulp/srp/ib_srp.c b/drivers/infiniband/ulp/srp/ib_srp.c
index 444d16520506..4f6e88136da3 100644
--- a/drivers/infiniband/ulp/srp/ib_srp.c
+++ b/drivers/infiniband/ulp/srp/ib_srp.c
@@ -1334,13 +1334,6 @@ static void srp_terminate_io(struct srp_rport *rport)
struct scsi_device *sdev;
int i, j;
- /*
- * Invoking srp_terminate_io() while srp_queuecommand() is running
- * is not safe. Hence the warning statement below.
- */
- shost_for_each_device(sdev, shost)
- WARN_ON_ONCE(sdev->request_queue->request_fn_active);
-
for (i = 0; i < target->ch_count; i++) {
ch = &target->ch[i];
Since this function is not in the hot path and since this function is
called indirectly, I'd like to keep the above code as documentation of
the requirements for callers of this function.
Please also Cc the linux-rdma mailing list for SRP initiator changes.
Thanks,
Bart.
