]> granicus.if.org Git - clang/commit
[OpenCL] Do not use vararg in emitted functions for enqueue_kernel
authorYaxun Liu <Yaxun.Liu@amd.com>
Sun, 3 Sep 2017 13:52:24 +0000 (13:52 +0000)
committerYaxun Liu <Yaxun.Liu@amd.com>
Sun, 3 Sep 2017 13:52:24 +0000 (13:52 +0000)
commitfc6ac77722c6bdc2508229af6f8aebfebe02f0fa
treef19af9a7f61a90000b35492d247dd58c6843f37c
parentc9dbb26f1933fc1a8ffaea4be96e3335dd5ba3b6
[OpenCL] Do not use vararg in emitted functions for enqueue_kernel

Not all targets support vararg (e.g. amdgpu). Instead of using vararg in the emitted functions for enqueue_kernel,
this patch creates a temporary array of size_t, stores the size arguments in the temporary array
and passes it to the emitted functions for enqueue_kernel.

Differential Revision: https://reviews.llvm.org/D36678

git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@312441 91177308-0d34-0410-b5e6-96231b3b80d8
lib/CodeGen/CGBuiltin.cpp
test/CodeGenOpenCL/cl20-device-side-enqueue.cl