In message <tkrznb5ia4e.fsf@memory-leak.sm.sony.co.jp>, you write: 

-> christos@zoulas.com (Christos Zoulas) writes:
-> 
-> > My guess is to prevent it from blocking/sleeping because we are
-> > careless and we are calling ktrace functions from points where
-> > sleeping is not allowed?
-> 
-> If so, we need to avoid the write operation also, to prevent a process
-> waiting for I/O completion.  And if the write operation is done by
-> other thread or is defered until other safe ktrxxx() call, the fp
-> itself isn't necessary to be marked non-blocking, is it?

I seem to vaguely recall some similar-sounding changes so that ktrace
could write out the trace file to NFS, but my memory is pretty hazy on 
this, so I could be out in left field :-)

--rafal

----
Rafal Boni                                                     rafal@pobox.com
  We are all worms.  But I do believe I am a glowworm.  -- Winston Churchill