> Yes, I do plan to fix this properly fairly soon.
   
   Then may I suggest that, unless "fairly soon" is in the next few days (say
   this weekend), you put back Christos's "hack", and then revert it when the
   proper fix goes in?
   
   While I agree whole-heartedly with your expressed desire to keep the main
   code clean, by reverting and not adding a compat-specific fix, we lose
   useful functionality.


i dunno...


i've worked hard to remove & keep compat-specific code out of
MI code (as has jd) and hacks like this are exactly the sort
of thing i really dislike.  i know where jd is coming from and
i think i support his position more than christos'.

in this case, the hack is always broken somehow:

	- if it's only there with COMPAT_LINUX it means the LKM
	won't work

	- if it's always present it means that _everyone_ gets
	his behaviour not just linux binaries.



.mrg.