[PATCH] Disable EHABI when -fno-exceptions option is chosen.

[Keith Walker]

Renato,

See comments below ...

> -----Original Message-----
> From: Renato Golin [mailto:renato.golin at linaro.org]
> Sent: 31 January 2014 12:38
> To: samsonov at google.com; rnk at google.com; Keith Walker;
> t.p.northover at gmail.com
> Cc: cfe-commits at cs.uiuc.edu; joey.gouly at gmail.com
> Subject: Re: [PATCH] Disable EHABI when -fno-exceptions option is
> chosen.
> 
>   Simpler check (thanks Joey) and test.
> 
> Hi samsonov, rnk, keith.walker.arm,
> 
> http://llvm-reviews.chandlerc.com/D2666
> 
> CHANGE SINCE LAST DIFF
>   http://llvm-reviews.chandlerc.com/D2666?vs=6792&id=6794#toc
> 
> Files:
>   lib/Driver/Tools.cpp
>   test/Driver/arm-no-exception.c
> 
> Index: lib/Driver/Tools.cpp
> ===================================================================
> --- lib/Driver/Tools.cpp
> +++ lib/Driver/Tools.cpp
> @@ -830,6 +830,10 @@
>        CmdArgs.push_back("-backend-option");
>        CmdArgs.push_back("-arm-reserve-r9");
>      }
> +
> +    // Exception handling
> +    if (Args.hasArg(options::OPT_fno_exceptions))

I think there should be a call to CmdArgs.push_back("-backend-option") here

> +      CmdArgs.push_back("-arm-disable-ehabi");
>  }
> 
>  // Get CPU and ABI names. They are not independent
> Index: test/Driver/arm-no-exception.c
> ===================================================================
> --- /dev/null
> +++ test/Driver/arm-no-exception.c
> @@ -0,0 +1,4 @@
> +// RUN: %clang -target arm-none-gnueeabi -fno-exceptions -### %s 2> %t
> +// RUN: FileCheck --check-prefix=CHECK-NOEH < %t %s
> +
> +// CHECK-NOEH: "-arm-disable-ehabi"

So also update expected test result.

Otherwise looks good to me.

Keith