crossbow discuss - Jan 2009 - panic in mac_register()

Hi,

   Trying to get my driver working with the new interfaces caused a 
panic in mac_register(). The stack is a little confusing:

[2]> $c
kmdb_enter(10c8e08, 0, 182b670, 2a102739a90, 0, 1256800)
panicsys+0x130(2, 10c8c00, 2a102739ca0, 2a102738da0, 10c8c00, 18fc800)
vpanic+0xcc(10c2a38, 2a102738d68, 44, 0, 4, 2a102738da0)
panic+0x1c(10c2a38, 31, 2a102738fd0, 44, 0, 189ecd8)
die+0x74(10c2800, 2a102738fd0, 44, 0, 4, 2a102738da0)
trap+0x9e8(2a102738fd0, 1fff, 31, e2042004, 18284f0, 5)
ktl0+0x64(40, 3f00, fc0, 1000, 60012f68d80, 60012f68d80)
mac`mac_register+0x67c(0, 300000ec020, 0, 6001270bc08, 16, 1000)
mac`mac_fini_macaddr+8(6001270bc08, 600173aa020, 0, 60012f68e08, 
60012f68d80,
60012f68d80)
mac`mac_register+0x67c(0, 2a102739358, 30037569730, 6001270bc08, 16, 0)

but I believe the problem is in the fail path (starts at mac_provider.c, 
line 436 in my source). It''s calling mac_fini_macaddr() but can be 
jumped to before mac_init_macaddr() is called; hence it is trying to 
free memory that has not been allocated. Has anyone else tripped across 
this?

   Paul

On Jan 20, 2009, at 2:48 AM, Paul Durrant wrote:
> Hi,
>
>   Trying to get my driver working with the new interfaces caused a
> panic in mac_register(). The stack is a little confusing:
>
> [2]> $c
> kmdb_enter(10c8e08, 0, 182b670, 2a102739a90, 0, 1256800)
> panicsys+0x130(2, 10c8c00, 2a102739ca0, 2a102738da0, 10c8c00, 18fc800)
> vpanic+0xcc(10c2a38, 2a102738d68, 44, 0, 4, 2a102738da0)
> panic+0x1c(10c2a38, 31, 2a102738fd0, 44, 0, 189ecd8)
> die+0x74(10c2800, 2a102738fd0, 44, 0, 4, 2a102738da0)
> trap+0x9e8(2a102738fd0, 1fff, 31, e2042004, 18284f0, 5)
> ktl0+0x64(40, 3f00, fc0, 1000, 60012f68d80, 60012f68d80)
> mac`mac_register+0x67c(0, 300000ec020, 0, 6001270bc08, 16, 1000)
> mac`mac_fini_macaddr+8(6001270bc08, 600173aa020, 0, 60012f68e08,
> 60012f68d80,
> 60012f68d80)
> mac`mac_register+0x67c(0, 2a102739358, 30037569730, 6001270bc08, 16,  
> 0)
>
> but I believe the problem is in the fail path (starts at  
> mac_provider.c,
> line 436 in my source). It''s calling mac_fini_macaddr() but can be
> jumped to before mac_init_macaddr() is called; hence it is trying to
> free memory that has not been allocated. Has anyone else tripped  
> across
> this?
Yes, it''s 6793436 (panic in mac_fini_macaddr() on mac_register()  
failure).

Nicolas.
>
>
>   Paul
> _______________________________________________
> crossbow-discuss mailing list
> crossbow-discuss at opensolaris.org
> http://mail.opensolaris.org/mailman/listinfo/crossbow-discuss
-- 
Nicolas Droux - Solaris Kernel Networking - Sun Microsystems, Inc.
droux at sun.com - http://blogs.sun.com/droux