LGTM thanks

LGTM thanks

LGTM thanks

LGTM thanks

minor formatting fixes, rebased

added extra checks for not host, rebased

In D152414#4493174, @arsenm wrote:

In D152414#4492691, @JonChesterfield wrote:

LG here. Can we get rid of EnableFunctionCalls?

If we can get rid of hipcc which is still using it

hipcc only needs a way to force EarlyInlineAll but that pass is predicated by !EnableFunctionCalls. If we don't think it's warranted then we can deprecate EnableFunctionCalls by first removing it in hipcc.

In rGac3e720dc1af9555c5591017bd2b6c7f6641a34d#1225535, @yaxunl wrote:

In rGac3e720dc1af9555c5591017bd2b6c7f6641a34d#1225506, @arsenm wrote:

Can you either

1. Add test coverage
2. Delete this entirely

Is it really worth holding a static_assert for the size of the FP types? The implementation for nan should be moved out of the header anyway (and go through builtin_nan/builtin_nanf)

+ @scchan I think we might still need these changes. We could add a lit test to clang/test/Headers about this.

Yea, I think we can make nan/nanf to go through the builtins and then get rid of this implementation.

addressed comments from yaxunl, rebased

LGTM

In D152414#4406186, @arsenm wrote:

In D152414#4406127, @scchan wrote:

We should probably drop the amdgpu-function-calls too but I think
hipcc is still using it.

The amdgpu-function-calls switch is needed to disable function calls as a prereq to enable alwaysinline, which I don't understand why.

I'm failing to parse this sentence

rebased

We should probably drop the amdgpu-function-calls too but I think
hipcc is still using it.

Removed redundant comment.

added a matching --no-whole-archive

fix some formatting issue

remove ws

Updated patch to address review feedback

LGTM

In D142506#4078402, @arsenm wrote:

The resource directory should be strictly controlled. why would lib64 ever be used here?

In D140315#4076614, @mgorny wrote:

I'm sorry for reporting it this late (clang's been broken for over two weeks, so we couldn't periodically test it) but this change is causing test regressions on Gentoo/amd64:

Thanks for reporting the issue. I'm working on a patch to fix the test.

In D140315#4011440, @yaxunl wrote:

files under clang/test/Driver/Inputs/rocm_resource_dir/lib/amdgcn/bitcode-no-abi-ver seem not used.

Remove unused bitcode files

LGTM

Just briefly glanced through the code and it seems that it doesn't look at whether a kernel has dynamic LDS or not because LDS spilling has to be disabled in that case, can you confirm?

In D127052#3560815, @JonChesterfield wrote:

Hoping this is uncontroversial - got more patches to follow up with if we can land this,

• fixed the hip-link-bundle-archive.hip test issue on Windows
• fixed a couple of formatting issues

update the failing lit tests to recognize the new switches

update the clang-offload-bundler-asserts-on test

fix lit test failures on Windows, rebased

In D117494#3254762, @yaxunl wrote:

I am OK for now. We may need keep an eye for the perf regression and be prepared to figure a way to alleviate that.

Check for zero-sized and move the check into the existing loop.

LGTM

Minor clean up in the hip-header.hip test

Adding test

LGTM

LGTM

LGTM

In D31210#706955, @yaxunl wrote:

In D31210#706890, @kzhuravl wrote:

In D31210#706880, @rampitec wrote:

I'm concerned about the default address space to be 64 bit. It would move alloca into generic address space effectively making private address to be 64 bit.
This may have very undesirable performance implications, like address arithmetic can become expensive 64 bit and only be truncated at load or store.
I realize you will use addrspacecast on an alloca's value, though I'm not sure that is sufficient to mitigate performance hit.
I believe such change shall not be made without a good performance comparison with the feature enabled, provided the very likely performance issues.

Did not we want to use this: http://lists.llvm.org/pipermail/llvm-dev/2017-March/111199.html and use non-0 for our allocas?

Our final goal is to let alloca return private pointer. The Clang changes are mostly common whether alloca returns generic pointer or private pointer. Actually to be able to test the above patch we need to get the changes in Clang done first.

LGTM