Re: sizeof (siginfo_t) problem

[Jamie Lokier]

Jakub Jelinek wrote:
> When siginfo_t was added, the intent obviously was that its size
> be 128 bytes on all arches.
> 
> The kernel unfortunately does this right on sparc64 and alpha from 64-bit
> arches only; ia64, s390x, ppc64 etc. got it wrong.

That's not the only siginfo_t problem:

	- Take a look at the placement of the _uid32 field on m68k.
	  It varies from sub-structure to sub-structure - yet it is
	  always written to the same offset by the kernel.  Borken!

Other cleanups:

	- SI_* codes on S390 are identical to the generic ones.
	  They can be removed from the S390 file.

	- Comment in MIPS siginfo about IRIX compatibility is no
	  longer true; the layout was changed when uid32 added.

	- __ARCH_SI_UID_T can be removed.  Only Sparc sets it,
	  and to the same type as uid_t (unsigned int).

-- Jamie