On Mon, 2006-08-28 at 11:00 +0100, David Woodhouse wrote:

On Mon, 2006-08-28 at 10:53 +0200, Andi Kleen wrote:

quoted

quoted

/usr/include/linux is _not_ a place to dump "reference code" in lieu of
documentation on using kernel interfaces.

At least for the system call interface it was always. It is not
my fault you're trying to suddenly redefine it to be something else.

I'm trying to 'suddenly redefine' kernel headers as something that
_isn't_ just a library of random crap for people to abuse in userspace
as they see fit, then whine when something breaks even though it was
never really guaranteed to work when abused in that way anyway.

So far, you're just reminding me why that needed to be done.

quoted

quoted

Besides, the _syscallX implementations in the kernel were generally
unsuitable for use [as a reference implementation]

I disagree. I used them and they worked great for me.

the x86_64 macros use the VDSO page? great! I didn't know that.
But that would make them unusable for kernel use I suspect...


-- 
if you want to mail me at work (you don't), use arjan (at) linux.intel.com