Re: [RFC] killing boilerplate checks in ->link/->mkdir/->rename

[Al Viro <viro@ZenIV.linux.org.uk>]

On Fri, Feb 03, 2012 at 01:16:12AM +0000, Al Viro wrote:

> After looking a bit more: nlink_t is a f*cking mess.  Almost any code
> using that type kernel-side is broken.  Crap galore:
> 	* sometimes it's 32 bits, sometimes 16, sometimes 64.  Essentially
> at random.
> 	* almost all have it unsigned, except for sparc32, where it's
> signed short [inherited from v7 via SunOS?  BTW, in v6 it used to be even
> funnier - char, which is where ridiculous LINK_MAX == 127 comes from]
> 
> IOW, nlink_t is an attractive nuisance - it's nearly impossible to use in
> a portable way and we are lucky that almost nobody tries to.

Incidentally, why the hell do we have
	typedef __kernel_nlink_t        nlink_t;
anyway?  It's *not* exposed to userland and it's different from the
userland nlink_t (which is unsigned int on 32bit and unsigned long on 64bit).
Why not use __kernel_nlink_t (or explicitly-sized __uNN) in
arch/*/include/asm/stat.h and declare nlink_t kernel-side as __u32?

Why do we have daddr_t, while we are at it?  There is exactly one user -
fs/freevxfs and there we definitely want a fixed-sized type.