On Mon, Sep 08, 2014 at 09:32:34PM +0200, Joerg Sonnenberger wrote:
 > Hi all,
 > please review the attached patch to remove the two remaining MI uses of
 > random(9). Neither should be in a performance critical code path and
 > both may actually benefit from the better distribution. The vfs_bio.c
 > change is a bit more involved as I am trying to replace a 32bit division
 > with a long multiplication, which should be cheaper pretty much
 > anywhere.

When doing that sort of thing, please expand the explanatory comments,
not shrink them...

-- 
David A. Holland
[email protected]

Reply via email to