labath accepted this revision. labath added a comment. In D121348#3373396 <https://reviews.llvm.org/D121348#3373396>, @jingham wrote:
> In D121348#3373170 <https://reviews.llvm.org/D121348#3373170>, @labath wrote: > >> I don't think this is a peculiarity of the darwin ABI. I'm pretty sure i've >> ran into this on the "sysv" ABI as well. >> >> If we're going to disallow this, then I think we should do the same for >> ABISysV_arm64 as well -- it looks like an identical change to ABISysV_arm64 >> should suffice. > > The ABI docs say "you aren't required to restore the value of x8 on function > return"... Not all that surprising everybody took the speed over > debuggability option. Yeah, I'm actually pretty fine with this. I dread the day when they start routinely (I think it already happens in a limited fashion) start using ABI's made up on the spot for translation-unit-local functions. In D121348#3376345 <https://reviews.llvm.org/D121348#3376345>, @jingham wrote: > At Pavel's request I extended this to the SysV_arm64 ABI. Since I don't want > to have to guess which arch's are which ABI's, I exposed the ABI plugin's > name and then used that in the test. Heh.. one of my reasons for requesting this was to avoid complex logic in the test -- I thought you'd just say that the functionality is not available on all arm64 architectures. Though I suppose having the plugin name around could be useful in general... Repository: rG LLVM Github Monorepo CHANGES SINCE LAST ACTION https://reviews.llvm.org/D121348/new/ https://reviews.llvm.org/D121348 _______________________________________________ lldb-commits mailing list lldb-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/lldb-commits
