[llvm] r247450 - Use function attribute "stackrealign" to decide whether stack

[Hal Finkel via llvm-commits]

----- Original Message -----
> From: "Reid Kleckner via llvm-commits" <llvm-commits at lists.llvm.org>
> To: "Akira Hatanaka" <ahatanak at gmail.com>
> Cc: "llvm-commits" <llvm-commits at lists.llvm.org>
> Sent: Friday, September 11, 2015 5:08:38 PM
> Subject: Re: [llvm] r247450 - Use function attribute "stackrealign" to decide whether stack
> 
> 
> 
> I recall some users from nvidia or Intel did something like arranging
> to call a function using AVX with 32-byte stack alignment, and using
> -mno-stackrealign to suppress the funky prologue. They would
> probably be better served with an ABI stack alignment override, so
> maybe that use case is not worth considering.
> 

It seems like, for this use case, we really want an attribute to specify what the enhanced caller-provided stack alignment actually is, so that the backend can then make the appropriate choice about generating a stack realignment based on the needs of its local variables and spill slots.

 -Hal

> 
> 
> On Fri, Sep 11, 2015 at 2:31 PM, Akira Hatanaka < ahatanak at gmail.com
> > wrote:
> 
> 
> 
> I was thinking "no-stack-realign" should be a cl::opt debug option,
> not a function attribute.
> 
> 
> 
> Is there a use case that requires the end-user to disallow stack
> realignment even when it's necessary to do so to be functionally
> correct (for example, there is a local variable with a large
> alignment)?
> 
> 
> 
> 
> 
> On Fri, Sep 11, 2015 at 12:57 PM, Reid Kleckner via llvm-commits <
> llvm-commits at lists.llvm.org > wrote:
> 
> 
> 
> 
> 
> 
> 
> On Fri, Sep 11, 2015 at 12:40 PM, Joerg Sonnenberger via llvm-commits
> < llvm-commits at lists.llvm.org > wrote:
> 
> 
> On Fri, Sep 11, 2015 at 06:54:39PM -0000, Akira Hatanaka via
> llvm-commits wrote:
> > URL: http://llvm.org/viewvc/llvm-project?rev=247450&view=rev
> > Log:
> > Use function attribute "stackrealign" to decide whether stack
> > realignment should be forced.
> 
> Can you please actually call it force_stackrealign or
> always_realign_stack to avoid confusion?
> 
> 
> 
> The X86 backend looks for a "no-stack-realign" attribute. Perhaps we
> should have a string attribute "stackrealign" with three values,
> "always", "auto", and "never". The default behavior is "auto".
> _______________________________________________
> llvm-commits mailing list
> llvm-commits at lists.llvm.org
> http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-commits
> 
> 
> 
> 
> _______________________________________________
> llvm-commits mailing list
> llvm-commits at lists.llvm.org
> http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-commits
> 

-- 
Hal Finkel
Assistant Computational Scientist
Leadership Computing Facility
Argonne National Laboratory