[PATCH] Use ARM-style representation for C++ method pointers under PNaCl

[Mark Seaborn]

On 19 June 2013 13:17, John McCall <rjmccall at apple.com> wrote:

> On Jun 19, 2013, at 1:05 PM, Mark Seaborn <mseaborn at chromium.org> wrote:
>
> On 19 June 2013 13:01, Mark Seaborn <mseaborn at chromium.org> wrote:
>
>> Use ARM-style representation for C++ method pointers under PNaCl
>>
>> Before this change, Clang uses the x86 representation for C++ method
>> pointers when generating code for PNaCl.  However, the resulting code
>> will assume that function pointers are 0 mod 2.  This assumption is
>> not safe for PNaCl, where function pointers could have any value
>> (especially in future sandboxing models).
>>
>> So, switch to using the ARM representation for PNaCl code, which makes
>> no assumptions about the alignment of function pointers.
>>
>> See: https://code.google.com/p/nativeclient/issues/detail?id=3450
>>
>
> Oops, I meant to send this to cfe-commits rather than llvm-commits.
>
>
> I do not think you should just unconditionally opt in to random ARM
> behavior.  In particular, ARM uses 32-bit guard variables because that's
> the size of a pointer on ARM;  PNaCl needs to be able to efficiently
> support 64-bit platforms as well.
>

The code does always use 64-bit guard variables on 64-bit systems.  It does
this:

    // Guard variables are 64 bits in the generic ABI and size width on ARM
    // (i.e. 32-bit on AArch32, 64-bit on AArch64).
    guardTy = (IsARM ? CGF.SizeTy : CGF.Int64Ty);

Having said that, PNaCl is 32-bit-only:  PNaCl programs assume a 32-bit
address space.  We don't support 64-bit pointers in PNaCl.  In Clang,
targeting PNaCl is identified by "le32" being in the triple, and I assume
there's no way to get 64-bit pointers with "le32". :-)

Cheers,
Mark
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/cfe-commits/attachments/20130619/55bc2007/attachment.html>