Correct the LLVM target of aarch64-unknown-uefi

[tgross35]

As noticed in #132975 (comment), the LLVM target string for aarch64 UEFI is not correct. Fix it here.

[rustbot]

r? @chenyukang

rustbot has assigned @chenyukang.
They will have a look at your PR within the next two weeks and either review your PR or reassign to another reviewer.

Use r? to explicitly pick a reviewer

[rustbot]

These commits modify compiler targets.
(See the Target Tier Policy.)

[arichardson]

I just looked at the LLVM source history for the UEFI triple subcomponent - it was added in https://reviews.llvm.org/D131594 (llvm/llvm-project@30198bd). I just looked where the UEFI triple is used and it appears it's only being used for x86.

It looks like it does not work for aarch64:

echo "int x = 1;" | ~/upstream-llvm-project/cmake-build-debug/bin/clang -x c -c - -o /tmp/foo.o -target aarch64-unknown-uefi && file /tmp/foo.o
error: backend data layout 'e-m:w-p270:32:32-p271:32:32-p272:64:64-p:64:64-i32:32-i64:64-i128:128-n32:64-S128-Fn32' does not match expected target description 'e-m:e-p270:32:32-p271:32:32-p272:64:64-i8:8:32-i16:16:32-i64:64-i128:128-n32:64-S128-Fn32'

But it does work for x86_64, so we could update that target info:

❯ echo "int x = 1;" | ~/upstream-llvm-project/cmake-build-debug/bin/clang -x c -c - -o /tmp/foo.o -target x86_64-unknown-uefi && file /tmp/foo.o
/tmp/foo.o: Intel amd64 COFF object file, not stripped, 4 sections, symbol offset=0xb8, 12 symbols, 1st section name ".text"

[arichardson]

CC: @Prabhuk was the UEFI support in LLVM supposed to only support x86 or is AArch64 support planned eventually?

[tgross35]

I just found #132570 which does the same thing, plus a couple others. I'll close this one, @arichardson mind taking a look at that PR?

[Prabhuk]

CC: @Prabhuk was the UEFI support in LLVM supposed to only support x86 or is AArch64 support planned eventually?

Yes. Currently only x86_64 is being worked on. The backend patch for it still needs to land which I am hoping will happen by end of the month. AArch64 is the next target I want to work on followed by 32 bit targets. I think the Driver also needs some fixes to provide an error if unsupported UEFI triple is passed.