panic in amap_wipeout (amd64)?

["Steven M. Bellovin" <smb%cs.columbia.edu@localhost>]

I need to gather a few more details before I send-pr, but is anyone
else seeing a reproducible panic in amap_wipeout?  I first noticed this
on a kernel from Friday, Jan 25 (amd64) on a Thinkpad T61.  If I boot
single-user and do

        sysctl -w machdep.sleep_state=3

the machine will suspend and resume properly.  If, however, I do a
second

        sysctl -w machdep.sleep_state=3

it will panic on resume.  Normal behavior, which may involve a few
hours between suspend/resumes, doesn't trigger this.  (Well, I do see
crashes, but not that frequently.)  I haven't yet been able to get a
traceback from a dump.


                --Steve Bellovin, http://www.cs.columbia.edu/~smb