jchlanda wrote:
> In principle I am against this, it adds a relatively brittle hook, and
> bypasses the pre-existing mechanisms (use CL2 or enable the generic-as
> extension) for obtaining this behaviour, in a way that does not ensure that
> the pre-existing mechanisms are available (e.g. it a
https://github.com/jchlanda closed
https://github.com/llvm/llvm-project/pull/117588
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
https://github.com/AlexVlx requested changes to this pull request.
In principle I am against this, it adds a relatively brittle hook, and bypasses
the pre-existing mechanisms (use CL2 or enable the generic-as extension) for
obtaining this behaviour, in a way that does not ensure that the pre-ex
AlexVlx wrote:
Thank you for this, unfortunately I don't quite see the reason for adding
another twiddly bit here. We are long term migrating away from the
default-is-private hack, and to get the benefits you are looking for you can
simply compile for CL2, or enable the generic as extension, f
llvmbot wrote:
@llvm/pr-subscribers-backend-amdgpu
Author: Jakub Chlanda (jchlanda)
Changes
`opencl-def-is-generic-addrspace` sets the default address space from private
to generic. This feature allows for building bitcode libraries written in
OpenCL that can then be linked against modul
https://github.com/jchlanda created
https://github.com/llvm/llvm-project/pull/117588
`opencl-def-is-generic-addrspace` sets the default address space from private
to generic. This feature allows for building bitcode libraries written in
OpenCL that can then be linked against modules compiled f
