Re: [PROPOSAL/PATCH] Remove PT_GNU_STACK support before 2.6.11

[Arjan van de Ven <arjan@infradead.org>]

On Sun, 2005-02-06 at 09:31 -0800, Linus Torvalds wrote:
> 
> On Sun, 6 Feb 2005, Arjan van de Ven wrote:
> > 
> > Note that these techniques all exist today. The only issue is that the
> > current code doesn't do the RWE->READIMPLIESEXEC binding, which my patch
> > fixed. 
> 
> My main objection to your patch is the naming. If 'executable_stack' 
> affects the heap, then why is it called "executable_STACK"?

fair point.

> 
> Wouldn't it be much nicer to
> 
>  - get rid of "EXSTACK_DEFAULT" as a special case, and instead just have 
>    the architecture _initialize_ the damn variable to what it wants? In 
>    other words, make it a nice understandable binary value (or maybe a 
>    bitmask, if you want to have separate flags for stack/heap/mmap), 
>    rather than a ternary value where one of the values means something 
>    arch-dependent.
> 
>  - just rename the dang thing to "read_implies_exec" and be done with it?
> 
> Hmm? Wouldn't that make a lot more sense?

it would. It'll be a bit bigger than I'd be comfortable with this late
in the 2.6.11 cycle though. 

> 
> And if you want to split things up, there's at least three flags there:  
> "stack" vs "file mapping" vs "anonymous mapping". For example, it might

lets add "brk" as 4th I guess.

Ok so what to do for 2.6.11... the setarch workaround is there; that
works. My patch patches the worst issues and is quite minimal. What you
propose will be more invasive and more suitable for 2.6.11-bk1... 
I can do such a patch no problem (although the next two days I won't
have time).