pgt(4): do not tsleep in the scan ioctl.

The only action taken during the scan ioctl was to sleep. This does not make
sense as the scanning process happens in the state machine behind the scenes.

Initialization indeed performs newstate(SCAN), but that is done no matter
if the user waits for the full results or not. Instead, this change unifies
the (new) user experience when performing ifconfig scan and shows the
nodes list produced by the latest background scan.

This was not tested on real hardware.
Please message me directly in case this broke your card.

OK stsp@, deraadt@

1 files changed, 1 insertions, 8 deletions

diff --git a/sys/dev/ic/pgt.c b/sys/dev/ic/pgt.c
index 45a85503221..8ff77a99ccb 100644
--- a/sys/dev/ic/pgt.c
+++ b/sys/dev/ic/pgt.c
@@ -1,4 +1,4 @@
-/*	$OpenBSD: pgt.c,v 1.90 2017/10/26 15:00:28 mpi Exp $  */
+/*	$OpenBSD: pgt.c,v 1.91 2018/04/28 11:17:08 pirofti Exp $  */
 
 /*
  * Copyright (c) 2006 Claudio Jeker <claudio@openbsd.org>
@@ -2250,13 +2250,6 @@ pgt_ioctl(struct ifnet *ifp, u_long cmd, caddr_t req)
 		/*
 		 * This chip scans always as soon as it gets initialized.
 		 */
-
-		/*
-		 * Give us a bit time to scan in case we were not
-		 * initialized before and let the userland process wait.
-		 */
-		tsleep(&sc->sc_flags, 0, "pgtsca", hz * SCAN_TIMEOUT);
-
 		break;
 	case SIOCG80211ALLNODES: {
 		struct ieee80211_nodereq *nr = NULL;