thakis added a comment. Heads-up: This very visibly broke web page rendering in Chromium (but only when using LTO) (https://crbug.com/1383873) . Not clear yet if this exposed UB on our end or not.
One thing that we did notice is that adding `-fno-global-isel` to cflags and ldflags only had an effect when manually blowing away the thinlto cache. Maybe whatever decides when to invalidate the thinlto cache needs to learn to invalidate it when the `-global-isel` flag changes? ================ Comment at: clang/lib/Driver/ToolChains/Clang.cpp:7218-7219 + CmdArgs.push_back("-global-isel=1"); + CmdArgs.push_back("-mllvm"); + CmdArgs.push_back("-global-isel-abort=0"); + } ---------------- aemerson wrote: > arsenm wrote: > > Why abort=0? I can understand abort=1 or 2 > Abort=1 would mean the compiler crashes on fallback, and `2` would mean we > emit diagnostics, neither of which we want for an on-by-default configuration. Any reason why we pass different flags to ld's LTO than to cc1? Repository: rG LLVM Github Monorepo CHANGES SINCE LAST ACTION https://reviews.llvm.org/D137269/new/ https://reviews.llvm.org/D137269 _______________________________________________ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits