Bug #217989 forces cloop when USE(64KB_PAGE_BLOCK) is enabled. Xan recommends we use llint's assembler backend instead, since it is faster and bugs should be rare.
Created attachment 413291 [details] Patch
(Mike: this reverses my previous recommendation that we should use cloop on aarch64.)
Committed r269586: <https://trac.webkit.org/changeset/269586> All reviewed patches have been landed. Closing bug and clearing flags on attachment 413291 [details].
This doesn't build. Turns out aarch64 llint depends on JIT due to USE(JUMP_ISLANDS). It seems this configuration is maybe not as well-tested as we had hoped? Let's follow-up in bug #219288.