Oops + a favor

[Dominik Brodowski]

Hi Jean,

> 	When I boot 2.5.72, I get the Oops below. You may have already
> seen it reported from many source, sorry for the dup.
Thanks for this oops report -- haven't seen it before. Could you please 
check whether the attached patch helps?

> 	I've got a request for you. Could you find yourself a 16bit
> Pcmcia card and a 16bit ISA-Pcmcia adapter ? The 32bit is much more
> straightforward than the old 16bit stuff, and you really want it to do
> proper test coverage (which doesn't seem to be happening).

In fact, I test all my patches with a 16bit PCMCIA card. 32bit cardbus
testing is then done by Russell. And as I do not own any system with an ISA
bus [and I have no intention on spending money on this], I won't follow your
request.

	Dominik

diff -ruN linux-original/drivers/pcmcia/i82365.c linux/drivers/pcmcia/i82365.c
--- linux-original/drivers/pcmcia/i82365.c	2003-06-18 11:01:07.000000000 +0200
+++ linux/drivers/pcmcia/i82365.c	2003-06-18 11:38:39.000000000 +0200
@@ -1482,8 +1482,10 @@
 	add_timer(&poll_timer);
     }
 
-    class_device_create_file(&socket[i].socket.dev, &class_device_attr_info);
-    class_device_create_file(&socket[i].socket.dev, &class_device_attr_exca);
+    for (i = 0; i < sockets; i++) {
+	    class_device_create_file(&socket[i].socket.dev, &class_device_attr_info);
+	    class_device_create_file(&socket[i].socket.dev, &class_device_attr_exca);
+    }
     
     return 0;
     
@@ -1494,6 +1496,8 @@
     int i;
 
     for (i = 0; i < sockets; i++) {
+	    class_device_remove_file(&socket[i].socket.dev, &class_device_attr_info);
+	    class_device_remove_file(&socket[i].socket.dev, &class_device_attr_exca);
 	    pcmcia_unregister_socket(&socket[i].socket);
     }
     platform_device_unregister(&i82365_device);