• Change is_md5sum to is_md5

• Correct the options we pass to csum

• Fix typo

(hopefully someday we can set this based on something internal to the CMake configuration rather than OBJECT_MODE, but that's all we have for now)

LGTM, thanks!

LGTM with minor comment and naming fixes

LGTM, thanks!

Noting this is to fix assertions in llvm/test/LTO/X86/remangle_intrinsics.ll on AIX.

Thanks, LGTM

LGTM, with minor nit

• Address comments

• Fix formatting

• Handle specialized cases for pwr 4/5/6

• Simplify
• Comment

• Remove stray program.h

• Use systemcfg interfaces instead. Move the getsysttype call to the test.

LGTM

LGTM

The traceback decode format here is quite different from what binutils objdump would generate on AIX:

LGTM, thanks!

• Suppress linter warnings for header stub

• Prefer cmakedefine to cmakedefine01 because the later doesn't work with defined() macro checks

• Add back in SEH and SJLJ eh guards

Just a note on status: This patch is potentially being replaced by https://reviews.llvm.org/D89492

• Cut out the middle man of 'HAVE_EHTABLE_SUPPORT' and just say what we mean.

LGTM, thanks!

Thanks! I've verified this fixes the test on AIX.

In D89696#2344797, @jhuber6 wrote:

In D89696#2344753, @daltenty wrote:

In D89696#2344508, @jhuber6 wrote:

@daltenty Do you think this will fix the problem on AIX?

It's not just AIX that will have this problem I suspect. If you configure any cross-compiling build with a 64-bit host, targeting a 32-bit arch, your going to run into the problem. I think checking against a list of known 64-bit arches is good enough for the purposes here, but the check still needs to be against the target, not the host.

Also, the frontend diagnostic looks like it is off for the same reason:

else if (getArchPtrSize(T) != getArchPtrSize(TT))
  Diags.Report(diag::err_drv_incompatible_omp_arch)

T seems to be the target triple, not the host, but the diagnostic reads: pointer size is incompatible with host.

TT is the target triple, otherwise the error message would be backwards, x86 is the OpenMP target architecture and PPC is the host architecture. I'm not sure what you mean by checking the target rather than the host. The target is set explicitly by -fopenmp-targets=<triple>, since it's being set in the test file we try to make sure that the host matches it. If this was just a frontend test we could just specify the host Triple and be done with it. Maybe we could add some CMake options for which offloading libraries were built?

In D89696#2344508, @jhuber6 wrote:

@daltenty Do you think this will fix the problem on AIX?

LGTM, with slight adjustment. Though I will note that this requires a clean build because CMake will complain about the change in source directory in cache:

In D89492#2333051, @phosek wrote:

This is an alternative to D70744. I'm not a fan of introducing compiler-rt/builtins as a new top-level CMake, so I was trying to come up with an alternative and this is what I came up with, let me know what you think.