nvmet_fc: correct broken add_port

prior patch to remove the add_port reference also removed the saving of the port pointer, which mean any io had no port pointer thus was rejected. Basically nothing works.

Restore the saving of the port pointer.

---
this restores operation but if the targetport is removed, it will no longer be functional if it comes back. I will be submitting a more involved patch for this shortly

 drivers/nvme/target/fc.c | 1 +
 1 file changed, 1 insertion(+)

index b652208cb37b..fc22415f84fa 100644
@@ -2522,6 +2522,7 @@ nvmet_fc_add_port(struct nvmet_port *port)
 	list_for_each_entry(tgtport, &nvmet_fc_target_list, tgt_list) {
 		if ((tgtport->fc_target_port.node_name == traddr.nn) &&
 		    (tgtport->fc_target_port.port_name == traddr.pn)) {
+			tgtport->port = port;
 			ret = 0;
 			break;
 		}

2.13.1

Looks good,
Reviewed-by: Johannes Thumshirn <jthumshirn at suse.de>

But a Fixes: and/or Cc: <stable at vger.kernel.org> might have been a good Idea as well.

Johannes Thumshirn                                          Storage
jthumshirn at suse.de                                +49 911 74053 689
SUSE LINUX GmbH, Maxfeldstr. 5, 90409 N?rnberg
GF: Felix Imend?rffer, Jane Smithard, Graham Norton
HRB 21284 (AG N?rnberg)
Key fingerprint = EC38 9CAB C2C4 F25D 8600 D0D0 0393 969D 2D76 0850

Fixes perhaps.? The patch that introduced the issue (a0b69cc8e5caff3cb6442dfa807f47b44e716aad) is only in nvme-4.16, so it doesn't quite apply.

-- james

Right, a Fixes: reference is useful even if the Cc: stable isn't relevant.

This fixes the issue I reported -- where I needed to revert commit a0b69cc8e5 ("nvmet-fc: cleanup nvmet add_port/remove_port") to allow nvme connect to work with fcloop.

Christoph, please feel free to add:

Thanks,

I've folded this into the patch introducing the problem.