================
@@ -6440,7 +6440,8 @@ const ToolChain &Driver::getToolChain(const ArgList &Args,
TC = std::make_unique<toolchains::NVPTXToolChain>(*this, Target, Args);
break;
case llvm::Triple::AMDHSA:
- TC = std::make_unique<toolchains::ROCMToolChain>(*this, Target, Args);
+ TC = std::make_unique<toolchains::ROCMToolChain>(*this, Target, Args,
+ Target.isOSWindows());
----------------
jhuber6 wrote:
I need to wrap my head around this more. We create multiple toolchains for
offloading, the host toolchain and then the additional ones for offloading. All
of those toolchains inherit from `Generic_GCC` in this case which initializes
the detector. For the host toolchain this will just be the normal host triple,
but for the GPU it should be the GPU target. Since we have the triple in the
installation detector, we should just be able to change behavior if the OS is
windows? We have the triple argument for ROCm we just don't use it, what's
stopping you from just saving that triple inside the class and using it when we
do our detection?
https://github.com/llvm/llvm-project/pull/113628
_______________________________________________
cfe-commits mailing list
[email protected]
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
