vitalybuka added a comment. In D80046#2042874 <https://reviews.llvm.org/D80046#2042874>, @vitalybuka wrote:
> In D80046#2040309 <https://reviews.llvm.org/D80046#2040309>, @efriedma wrote: > > > If I'm understanding this correctly, the stack-safe metadata on allocas is > > both produced and consumed at LTO-time. So at the point where the > > stack-safe metadata would be produced, we can compute whether any later > > passes will query it. > > > > Given that, why do you need a module flag? > > > I assumed a pass knows nothing about consumers. Can you point to some > examples how this can be done? Oh, I assume the idea to take that from AArch64TargetMachine. Still AArch64TargetMachine is going to use this info only if sanitizers where enabled. Also right now SSI is used for memtag. but asan, hwasan may follow. I assume LTO PM knows nothing about if clang was using sanitizers, so we have always to insert the pass. Then we'd like to return from pass if nothing to do there. The pass may check if there are functions with corresponding sanitizer attributes. However with module flag we can make it more efficient and avoid scanning all functions. BTW. function attribute was used in the first snapshot, but @eugenis proposed the flag and I agree with him that now it's nicer. Repository: rG LLVM Github Monorepo CHANGES SINCE LAST ACTION https://reviews.llvm.org/D80046/new/ https://reviews.llvm.org/D80046 _______________________________________________ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
