Sebastien Roy wrote:
> softmac_destroy() is being passed a dip for a device that doesn't exist 
> in the softmac_hash:

This bug appears to have been fixed by Cathy's putback to the Clearview 
gate this morning:

Comment:
change DLADM_INVALID_LINKID to DATALINK_INVALID_LINKID
split class definition to class and media
remove -v option
not check "lo0" in dladm_valid_linkname()
fix confusing comments
change mac_open() to take macname as the input
system panicked because it fails to find softmacs in the predetach process

I should have looked at her putback comments more closely before trying 
to root-cause this bug using the fixed source-code as a reference.  No 
wonder it was an elusive bug to find. :-)  Alexander was running bits 
from June 26th when it panicked (i.e., it panicked after the bfu had 
completed but before Dan could reboot the system to the new bits.)

-Seb

Reply via email to