Use 64-bit inode numbers internally in the kernel

These patches make the kernel use 64-bit inode numbers internally, even on a
32-bit system.  They are required because some filesystems have intrinsic
64-bit inode numbers: NFS and XFS for example.  The 64-bit inode numbers are
then propagated to userspace automatically where the arch supports it.

Problems have been seen with userspace (eg: ld.so) using the 64-bit inode
number returned by stat64() or getdents64() to differentiate files, and failing
because the 64-bit inode number space was compressed to 32-bits, and so
overlaps occur.


There are four patches:

 (1) Supply __udivdi3() and __umoddi3() so that the compiler can do 64-bit
     division and modulus.  __udivmoddi4() is also supplied as the other two
     are wrappers around that.

 (2) Make __kernel_ino_t consistently typedef'd to "unsigned long long" on all
     archs.

     Where ino_t has been used in userspace facing structures (such as struct
     stat), it has been replaced with whatever ino_t used to be defined as on
     that arch.

     struct inode::i_ino and struct kstat::ino are changed to ino_t type.

 (3) Fix print format warnings in general code due to the change in type of
     i_ino.  These should now all carry an "ll" modifier.

 (4) As for (3), but fix the CacheFiles code.

David