On 07/31/15 12:10, Jakub Jelinek wrote:

> This will hopefully be just GOMP_4.1 instead in the end, but it can
> change when gomp-4_1-branch is merged to trunk, we don't guarantee
> ABI stability at this point.

Sure.

> I'd prefer version to go after devicep argument rather than before.
Fixed.

> And really don't have ver_func vs. unver_func, just a single
> callback that will take the version argument too (again, if possible
> after target_id).

Fixed (& elsewhere).  The patch should be checked for intelmic if possible 
(Ilya?).  The  changes there are very mechanical so I'm not expecting a problem.

We don't need to  make the initial value of GOMP_VERSION non-zero, because the 
absence of the GOMP_OFFLOAD_version func will distinguish out of date plugins at 
this point.


>> +
>> +  if (DLSYM_OPT (version, version))
>
> I'd prefer requiring version always (i.e. DLSYM (version);
> plus the v != GOMP_VERSION checking).

Fixed.

ok?  I'll commit a version of this to gomp4 branch shortly.

nathan