r189542 - use the last passed -munaligned-access / -mno-unaligned-access

Renato Golin renato.golin at linaro.org
Thu Aug 29 02:05:30 PDT 2013


Hal,

Absolutely, I think the reason why it's there is that we didn't know
better. Please, check with Jeroen for a revised patch.

Cheers,
Renato
On 29 Aug 2013 04:50, "Hal Finkel" <hfinkel at anl.gov> wrote:

> ----- Original Message -----
> > Author: rengolin
> > Date: Wed Aug 28 18:56:07 2013
> > New Revision: 189542
> >
> > URL: http://llvm.org/viewvc/llvm-project?rev=189542&view=rev
> > Log:
> > use the last passed -munaligned-access / -mno-unaligned-access
> >
> > Passing inconsistent munaligned-access / mno-unaligned-access
> > flags, intentionally resulted in a warning and the flag
> > no-unaligned-access being used.
> >
> > Gcc does, at least in practice, use the last flag in such a
> > case. This patch updates clang behaviour accordingly; use the
> > last flag or base alignment behaviour on the target (which
> > llvm will do if no flag is explicitly passed)
> >
> > Patch by Jeroen Hofstee.
> >
> > Modified:
> >     cfe/trunk/lib/Driver/Tools.cpp
> >     cfe/trunk/test/Driver/arm-alignment.c
> >
> > Modified: cfe/trunk/lib/Driver/Tools.cpp
> > URL:
> >
> http://llvm.org/viewvc/llvm-project/cfe/trunk/lib/Driver/Tools.cpp?rev=189542&r1=189541&r2=189542&view=diff
> >
> ==============================================================================
> > --- cfe/trunk/lib/Driver/Tools.cpp (original)
> > +++ cfe/trunk/lib/Driver/Tools.cpp Wed Aug 28 18:56:07 2013
> > @@ -2944,12 +2944,15 @@ void Clang::ConstructJob(Compilation &C,
> >    }
> >    // -mkernel implies -mstrict-align; don't add the redundant
> >    option.
> >    if (!KernelOrKext) {
> > -    if (Args.hasArg(options::OPT_mno_unaligned_access)) {
> > -      CmdArgs.push_back("-backend-option");
> > -      CmdArgs.push_back("-arm-strict-align");
> > -    } else if (Args.hasArg(options::OPT_munaligned_access)) {
> > -      CmdArgs.push_back("-backend-option");
> > -      CmdArgs.push_back("-arm-no-strict-align");
> > +    if (Arg *A = Args.getLastArg(options::OPT_mno_unaligned_access,
> > +                                 options::OPT_munaligned_access)) {
> > +      if (A->getOption().matches(options::OPT_mno_unaligned_access))
> > {
> > +        CmdArgs.push_back("-backend-option");
> > +        CmdArgs.push_back("-arm-strict-align");
> > +      } else {
> > +        CmdArgs.push_back("-backend-option");
> > +        CmdArgs.push_back("-arm-no-strict-align");
>
> From what Chandler tells me, using -backend-option plays horribly when the
> compiler is used via the tooling interface, and should really be avoided.
> This seems like something that belongs in TargetOptions; can we move it
> there?
>
>  -Hal
>
> > +      }
> >      }
> >    }
> >
> >
> > Modified: cfe/trunk/test/Driver/arm-alignment.c
> > URL:
> >
> http://llvm.org/viewvc/llvm-project/cfe/trunk/test/Driver/arm-alignment.c?rev=189542&r1=189541&r2=189542&view=diff
> >
> ==============================================================================
> > --- cfe/trunk/test/Driver/arm-alignment.c (original)
> > +++ cfe/trunk/test/Driver/arm-alignment.c Wed Aug 28 18:56:07 2013
> > @@ -1,9 +1,25 @@
> >  // RUN: %clang -target arm-none-gnueeabi -munaligned-access -### %s
> >  2> %t
> >  // RUN: FileCheck --check-prefix=CHECK-UNALIGNED < %t %s
> >
> > +// RUN: %clang -target arm-none-gnueeabi -mstrict-align
> > -munaligned-access -### %s 2> %t
> > +// RUN: FileCheck --check-prefix=CHECK-UNALIGNED < %t %s
> > +
> > +// RUN: %clang -target arm-none-gnueeabi -mno-unaligned-access
> > -munaligned-access -### %s 2> %t
> > +// RUN: FileCheck --check-prefix=CHECK-UNALIGNED < %t %s
> > +
> >  // CHECK-UNALIGNED: "-backend-option" "-arm-no-strict-align"
> >
> > +
> >  // RUN: %clang -target arm-none-gnueeabi -mno-unaligned-access -###
> >  %s 2> %t
> >  // RUN: FileCheck --check-prefix=CHECK-ALIGNED < %t %s
> >
> > +// RUN: %clang -target arm-none-gnueeabi -mstrict-align -### %s 2>
> > %t
> > +// RUN: FileCheck --check-prefix=CHECK-ALIGNED < %t %s
> > +
> > +// RUN: %clang -target arm-none-gnueabi -munaligned-access
> > -mno-unaligned-access -### %s 2> %t
> > +// RUN: FileCheck --check-prefix=CHECK-ALIGNED < %t %s
> > +
> > +// RUN: %clang -target arm-none-gnueabi -munaligned-access
> > -mstrict-align -### %s 2> %t
> > +// RUN: FileCheck --check-prefix=CHECK-ALIGNED < %t %s
> > +
> >  // CHECK-ALIGNED: "-backend-option" "-arm-strict-align"
> >
> >
> > _______________________________________________
> > cfe-commits mailing list
> > cfe-commits at cs.uiuc.edu
> > http://lists.cs.uiuc.edu/mailman/listinfo/cfe-commits
> >
>
> --
> Hal Finkel
> Assistant Computational Scientist
> Leadership Computing Facility
> Argonne National Laboratory
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/cfe-commits/attachments/20130829/58039855/attachment.html>


More information about the cfe-commits mailing list