Re: Ugh in 2.1.60
Alan Cox (alan@lxorguk.ukuu.org.uk)
Sun, 9 Nov 1997 00:23:31 +0000 (GMT)
> It should be also considered to move most copy_*_user()/get/put_user() calls
> out-of-line to save code bloat: there is evidence(1) that most of the 2.1
> binary image bloat (my 2.1 kernel is over 100k bigger than the equivalent 2.0
> one) comes from these functions. Actually the kernel would not only become
Evidence (2)
16Mbyte pentium - using a crude non inline of copy*user for the same kernel
gives me no real measurable difference on lmbench, and far better performance
due to 50K or so of saved memory