[clang] [compiler-rt] [llvm] [HIP] support 128 bit int division (PR #71978)

2023-12-07 Thread Yaxun Liu via cfe-commits
@@ -3630,10 +3631,17 @@ SDValue SITargetLowering::LowerCall(CallLoweringInfo &CLI, std::vector Ops; Ops.push_back(Chain); + bool AddTargetGlobalAddr = true; + // Try to find the callee in the current module. + if (isa(Callee)) { +Callee = DAG.getSymbolFunctionGloba

[clang] [compiler-rt] [llvm] [HIP] support 128 bit int division (PR #71978)

2023-12-05 Thread Yaxun Liu via cfe-commits
@@ -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) { + +

[clang] [compiler-rt] [llvm] [HIP] support 128 bit int division (PR #71978)

2023-12-04 Thread Yaxun Liu via cfe-commits
@@ -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) { + +

[clang] [compiler-rt] [llvm] [HIP] support 128 bit int division (PR #71978)

2023-11-12 Thread Joseph Huber via cfe-commits
jhuber6 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. I've thought a reasonable amount about a `compiler-rt` for GP

[clang] [compiler-rt] [llvm] [HIP] support 128 bit int division (PR #71978)

2023-11-12 Thread Matt Arsenault via cfe-commits
@@ -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) { + +