In D52050#2428735, @glaubitz wrote:

Hmm, I was pretty sure that autoconf can deal with x32 inside an x32 chroot.

In D52050#2428693, @glaubitz wrote:

Maybe sys::getDefaultTargetTriple() needs to be adjusted. I'll have a look.

In D52050#2428560, @glaubitz wrote:

Any idea how it could make clang default to -mx32 if this condition is met: getToolChain().getTriple().getEnvironment() == llvm::Triple::GNUX32?

In D91338#2421852, @pengfei wrote:

I'm OK with this patch. Can you please check if expensive check is happy with it?

Update to current LLVM.

In D16474#2421640, @RKSimon wrote:

@hvdijk Are you in a position to commandeer this phab revision and complete @hjl.tools's work?

In D16474#335360, @hjl.tools wrote:

LLVM ERROR: Cannot select: 0x29507a0: ch,glue = X86ISD::TLSADDR 0x28fc140, TargetGlobalTLSAddress:i32<i32* @x> 0 [TF=7]

@pengfei With the confirmation that we do need this, could you take another look? Does this look okay now?

In D91933#2421234, @RKSimon wrote:

If possible, please commit the pic.ll fix as a separate followup.

Changed auto to Register

In D52050#2415723, @glaubitz wrote:

Yeah, @hvdijk has made multiple other improvements which should finally allow the backend to be usable.

We still disagree on the search paths for libraries and headers though if I remember correctly.

In D91924#2418221, @RKSimon wrote:

We already have fast-isel-call tests - could we add gnux32 test coverage there?

Add a comment about skipping the optimization when constraining fails.

In D91339#2417918, @dstenb wrote:

What is the status of the expensive check failure? It has been present for five days now.

Fix FileCheck syntax at the same time, using CHECK-X32-DAG rather than CHECK-DAG-X32 that @pengfei noticed in D91339 after it had been committed already.

In D91339#2410165, @RKSimon wrote:

Thanks - please can you add -verify-machineinstrs to the pic.ll command lines as part of either of those patches?

As requested by @RKSimon, added -verify-machineinstrs to test/CodeGen/X86/pic.ll so that the problem is visible even when expensive checks are disabled.

I have now also opened https://reviews.llvm.org/D91933 hoping to fix the other pre-existing issue. With both https://reviews.llvm.org/D91924 and that, the test in here passes even when expensive checks are enabled.

In D91339#2409998, @RKSimon wrote:

@hvdijk On EXPENSIVE_CHECKS builds I'm seeing a build failure in pic.ll - please can you take a look?

Add MOV32rr to guarantee high bits will be zero.

Forgot to run clang-format-diff before uploading, updated now.

In D91913#2409688, @lebedev.ri wrote:

Then the comment needs to be fixed too i would think?

In D91339#2409588, @RKSimon wrote:

LGTM - @hvdijk do you have commit access yet?

In D91349#2408723, @ldionne wrote:

That looks correct to me with what I just read about x32 (I wasn't familiar with it). I can't imagine this is the only place that needs fixing for libc++ to work in x32?

In D91338#2395918, @hjl.tools wrote:

I never finished x32 work. You can use my branch as a reference when working on x32.

In D91338#2395610, @lebedev.ri wrote:

Was there ever a follow-up on https://lkml.org/lkml/2018/12/10/1145 ?

In D91338#2395497, @pengfei wrote:

Hi Harald, thanks for thoroughly answering my questions. I still have doubts intention of this patch. Is there any bug related to it?

Remove no longer used [[REG:.*]] from test

Remove no longer necessary FIXME comment

In D91339#2392968, @MaskRay wrote:

This is correct. Probably worth mentioning in the subject that this is for x32.

Thanks! I do not have commit access, can somebody commit it for me?

ping