[LLVMdev] Proposal: Loads/stores with deterministic trap/unwind behavior

Philip Reames listmail at philipreames.com
Thu Apr 10 09:40:07 PDT 2014

I'll respond to the rest of this thread separately, but want to respond 
to one bit in particular.

On 04/06/2014 08:52 PM, Peter Collingbourne wrote:
> I did initially consider implementing the checked load/store as an 
> intrinsic. But there are relatively boring reasons why this won't work 
> at present. For example, there is no current way to represent a struct 
> load using an intrinsic, as there is no mangling for struct types. 
> Also, named structs would need a mangling that is resistant to 
> renames. Rather than solve these problems, I decided to avoid 
> intrinsics entirely.
I have out of tree code which adds support for intrinsics with struct 
type.  I'm happy to share this.

More over, if you identify a weakness in the intrinsic mechanism, we 
should fix it.  :)

(Says the guy with a couple of such fixes out of tree... really should 
get around to upstreaming that...)


More information about the llvm-dev mailing list