CVS: The file 'linux/drivers/char/bluetooth/bluetooth.c' has been modified.

[Gordon M. <gm...@us...>]

The following file was modified in linux/drivers/char/bluetooth:

Name            Old version     New version     Comment
----            -----------     -----------     -------
bluetooth.c     1.161           1.162=20=20=20=20=20=20=20=20=20=20=20

The accompanying log:

--Someone correct me if I'm wrong, but I don't think we can get a useful SD=
P connection without this change. The only time we'll ever see BT_INACTIVE =
on a session after starting a BTCONNECT is if the line is /dev/ttyBTC. But =
if we establish an SDP connection on /dev/ttyBTC we'll never be able to use=
 it for BT_SDP_REQUEST ioctls because it's session state will be stuck in B=
T_LOWERCONNECTED. The BT_ACTIVE state is impossible on /dev/ttyBTC. So we'l=
l need to use one of the data devices, and as soon as one of these opens th=
e session state goes to BT_UPPERCONNECTED.


The diff of the modified file(s):

--- bluetooth.c	2001/04/12 12:06:53	1.161
+++ bluetooth.c	2001/04/17 05:33:19	1.162
@@ -2412,7 +2412,8 @@
 		D_WARN("bt_register_sdp: Pending SDP data for this new connection @ line=
 %d\n", line);
 	}
=20
-	if (SESSIONSTATE(line) =3D=3D BT_INACTIVE) {
+	if ((SESSIONSTATE(line) =3D=3D BT_INACTIVE) ||
+	    (SESSIONSTATE(line) =3D=3D BT_UPPERCONNECTED)){
 		/* now register ! */
 		DSYS("bt_register_sdp : line %d\n", line);
 		SESSIONSTATE(line) =3D BT_ACTIVE;
@@ -2503,7 +2504,7 @@
 		/* Data line ! */
 		if (!bt_stack_initiated) {
 			D_WARN("bt_register_tty : Bluetooth stack not initiated\n");
-			return -1;
+			return -EPERM;
 		}
 	}
=20