[PATCH] D14755: Handle ARMv6-J as an alias, instead of fake architecture
Jonathan Roelofs via llvm-commits
llvm-commits at lists.llvm.org
Thu Nov 19 11:31:37 PST 2015
jroelofs added a comment.
In http://reviews.llvm.org/D14755#293154, @jroelofs wrote:
> GCC seems to think `arm1136j-s` and `arm1136jz-s` are real things. FWIW, it treats `arm1136j-s` as v6J + no VFP2, `arm1136jf-s` as v6J + VFP2, `arm1136jz-s` as v6zk + no VFP2, and `arm1136jz-s` as v6zk + VFP2. That being said, it's possible GCC was wrong for adding them in the first place. I have no idea how many people rely on this behavior in gcc, or in clang.
>
> Perhaps it's best to "break" this with an error message that says: "x is not supported, did you mean y?"... that way users whose build systems are dependent on this mcpu at least get an explanation on what they have to fix, rather than them complaining that something just doesn't work anymore.
erm... ignore this ^, I didn't read the thread carefully enough.
http://reviews.llvm.org/D14755
More information about the llvm-commits
mailing list