The reason for the Change is because LLVM upstream did this commit last year:

commit a4776de44edaa3c84035632c60247b97eab08ff2
Author: Bill Wendling <isanbard@gmail.com>
Date:   Thu Oct 14 02:33:26 2010 +0000

    Add support for vmov.f64/.f32 encoding. There's a bit of a hack going on
    here. The f32 in FCONSTS is handled as a double instead of a float in the
    code. So the encoding of the immediate into the instruction isn't exactly in
    line with the documentation in that regard. But given that we know it's handled
    as a double, it doesn't cause any harm.

Wendling's Change caused USE_OLD_JIT to break, so, when merging a while ago, we only
took part of commit a4776de. Now when we switch to USE_MCJIT, we have to take the
entirety of commit a4776de.

However, the entirety of a4776de implies that we have to fix ARMCodeEmitter.cpp here,
to make sure that classical JIT and MC JIT work at the same time.
This should have been done upstream, but LLVM upstream obviously deprecates
classical JIT (USE_OLD_JIT) during Q3 last year. But we are now just in the
process of switching to USE_MCJIT. Anyway, LLVM has the right to deprecate things
that are not ToT. When we stay close to ToT, we can avoid such mess next time.

Change-Id: I20394c2c6f2c3943b2afd64ca418891e9361ebe7

0 files changed, 0 insertions, 0 deletions