[PATCH] Fix thumbv5e stack coloring assertion failure

[Jim Grosbach]

Hi Logan,

I'm not 100% familiar with the lifetime intrinsics, but I agree with Bill that this change seems safe and totally reasonable. I have no objections.

-Jim

On Feb 19, 2013, at 2:25 PM, Bill Wendling <wendling at apple.com> wrote:

> Hi Logan,
> 
> This LGTM and fairly obvious. I'd like Jim or Owen to chime in though.
> 
> -bw
> 
> On Feb 19, 2013, at 8:03 AM, Logan Chien <tzuhsiang.chien at gmail.com> wrote:
> 
>> Ping?
>> 
>> 
>> On Fri, Feb 15, 2013 at 1:45 PM, Logan Chien <tzuhsiang.chien at gmail.com> wrote:
>> Hi,
>> 
>>  I found that some code with llvm.lifetime.* intrinsics will raise assertion failure in Thumb1 backend when stack coloring is enabled.  The root cause is that we are accessing the stack slot for frame pointer without checking the existence of the stack slot.  The test cases and the patch are attached.  Please have a look.  Thanks.
>> 
>> Sincerely,
>> Logan
>> 
>> _______________________________________________
>> llvm-commits mailing list
>> llvm-commits at cs.uiuc.edu
>> http://lists.cs.uiuc.edu/mailman/listinfo/llvm-commits
>