nvme-fc: Zero out SGL before submitting AEN
Subsystems (2)
PATCH 1/1 nvme-fc: Zero out SGL before submitting AEN
nvme_fc_start_fc_op() will be updating the SGL entries of the sqe, so we need to zero those before resubmitting a received AEN. Otherwise we trigger the WARN_ON() in nvme_fc_start_fc_op().
Signed-off-by Hannes Reinecke <hare at suse.com>
--- drivers/nvme/host/fc.c | 3 +++ 1 file changed, 3 insertions(+)
drivers/nvme/host/fc.c +3 -0
@@ -2181,6 +2181,7 @@ enum { {struct nvme_fc_ctrl *ctrl =to_fc_ctrl (arg);struct nvme_fc_fcp_op *aen_op; +struct nvme_command *sqe;unsigned long flags;bool terminating = false;blk_status_t ret;
@@ -2197,6 +2198,8 @@ enum {return ; aen_op = &ctrl->aen_ops[aer_idx]; + sqe = &aen_op->cmd_iu.sqe; +memset (&sqe->common.dptr,0 ,sizeof (union nvme_data_ptr )); ret =nvme_fc_start_fcp_op (ctrl, aen_op->queue, aen_op,0 ,NVMEFC_FCP_NODATA );
Signature · 1 line
1.8.5.6
PATCH 1/1 nvme-fc: Zero out SGL before submitting AEN
Quoted 7 lines · Hannes Reinecke
nvme_fc_start_fc_op()will be updating the SGL entries of the sqe, so we need to zero those before resubmitting a received AEN. Otherwise we trigger theWARN_ON()innvme_fc_start_fc_op().Signed-off-by Hannes Reinecke <hare at suse.com>---
This patch is unnecessary. It was completely resolved by the following patch which has been committed:
nvme-fc: use transport-specific sgl format
http://lists.infradead.org/pipermail/linux-nvme/2017-September/012723.html
-- james