[patch] rfa: fix problem with reassociating pointer additions

[Thiemo Seufer]

Eric Christopher wrote:
[snip]
> > If so, I don't see how that applies; the effective address produced
> > will be properly sign extended.
> > 
> > I would need to see a test case...
> 
> I gave one in the thread...

Can this one be caused by either

 - an old Linux bug, where the first invalid address for a 32bit task
   running on a 64bit kernel was erraneously defined as 0x80000000
   instead of 0x7fff8000? (The trick of reserving the last 32k is
   commonly used in MIPS OSes to avoid such problems).

 - an old assembler bug, which garbled sign extensions in such cases?


Thiemo