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

[Andi Kleen <ak@suse.de>]

> Your main objection is that *incorrect* programs that assume they can
> execute malloc() code without PROT_EXEC protection. For legacy binaries
> keeping this behavior makes sense, no objection from me.
> 
> For newly compiled programs this is just wrong and incorrect.

That's not true as the grub/mono/... experience shows. 


> You mention grub (which has RWE and the patch below thus makes that work)
> and mono. mono has patches for this on their mailinglist and bugzilla since
> 2003 according to google, I'm surprised the novell mono guys haven't fixed
> this bug in their code.

There are probably more.

> FWIW all jvm's don't suffer from this. They are either legacy binaries or
> mprotect properly (only i386 traditionally had this behavior, all others
> already required PROT_EXEC anyway so any half portable app already did this,
> as well as any app portable to BSD since they enforce this on x86 as well)
> 
> So I rather see the patch below merged instead; it fixes the worst problems
> (RWE not marking the heap executable) while keeping this useful feature
> enabled.

It still keeps x86-64 32bit emulation as a proving point for experimental
i386 features which is really not its purpose.

-Andi