[llvm-commits] [llvm] r137145 - /llvm/trunk/docs/Atomics.html

[Chris Lattner]

Thanks!

On Aug 12, 2011, at 3:00 PM, Eli Friedman wrote:

> On Fri, Aug 12, 2011 at 11:47 AM, Chris Lattner <clattner at apple.com> wrote:
>> On Aug 11, 2011, at 4:46 PM, Eli Friedman wrote:
>>>> In the "Load and store" section, please give an example of the illegal transformation.
>>> 
>>> Done.
>> 
>> I meant a source code example. Something like:
>> 
>> For example, in this code:
>> 
>>  <pre>
>>   ...
>>  </pre>
>> 
>> It would be invalid for the optimizer to hoist out the store of ... because ...
> 
> Done.
> 
>>>> In the "Atomic Ordering" section, please make one subsection for each ordering, and then split out "optimizations allowed", "impact on code generation" and "things this corresponds to" in into their own paragraph or something.  The content looks great, but could use some structure to make it easier to digest.
>>> 
>>> Done.
>> 
>> Looks great.  Please add a subsection for non atomic accesses.  That gives you a chance to say that it is the default for C/C++ accesses, and briefly mention anything else interesting.
> 
> Done.
> 
>> It may be out of scope, but for monotonic and up, it would be useful to give a concrete synchronization idiom that the ordering is useful to represent.
> 
> Maybe... I'll think about it.
> 
> -Eli