[PATCH] D67926: Fix endianness handling in AVR MC

[serge via Phabricator via llvm-commits]

serge-sans-paille added a comment.

> How have you been testing this?

Rebuilding llvm package with AVR support on alla supported arch using Fedora buildsystem

> How many regression test failures are there on s390x? The first version of the patch should not have passed all the regression tests on a big-endian target, I think. (In particular, it should not have worked for 32-bit instructions, like llvm/test/MC/AVR/inst-jmp.s).

Yeah, the first version of the patch was not tested correctly (I was explicilty removing AVR from the targets for s390x). When tested correctly, most MC/AVR tests where failing. This is no longer the case (tests activated and pass).

> Probably want to wait for Dylan to comment, but looks fine. (Changing the code to use support::endian::write isn't necessary, but it's more readable.)

OK, let's wait for @dylanmckay input for a few days!


CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D67926/new/

https://reviews.llvm.org/D67926