@@ -937,27 +938,105 @@ bool CodeGenAction::loadLinkModules(CompilerInstance
&CI) {
for (const CodeGenOptions::BitcodeFileToLink &F :
CI.getCodeGenOpts().LinkBitcodeFiles) {
-auto BCBuf = CI.getFileManager().getBufferForFile(F.Filename);
-if (!BCBuf) {
+
+
Artem-B wrote:
> I don't think we're in a position to actually enable that at this time. We
> still don't have everything necessary to provide object linking, which this
> seems to rely on
OK. IR it is.
https://github.com/llvm/llvm-project/pull/71978
__
https://github.com/arsenm edited https://github.com/llvm/llvm-project/pull/71978
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
Artem-B wrote:
Would it be feasible to consider switching to the new offloading driver mode
and really link with the library instead? It may be a conveniently isolated use
case with little/no existing users that would disrupt.
https://github.com/llvm/llvm-project/pull/71978
__
https://github.com/yxsamliu created
https://github.com/llvm/llvm-project/pull/71978
Currently nvcc supports 128 bit int division in device code. This patch adds
support of 128 bit int division to HIP.
It builds lib functions for 128 bit division in compiler-rt for amdgcn target.
Then links co
