Add powerpc64le Linux support #979

antonblanchard · 2016-01-14T20:43:20Z

This patch adds powerpc64le Linux support. While the patch also adds
the matching powerpc64 bits, there are endian issues that need to be
sorted out.

The PowerPC LLVM changes for the swift ABI (eg returning three element
non-homogeneous aggregates) are still in the works, but a simple LLVM
fix to allow those aggregates results in swift passing all but 8
test cases.

Note: It would be nice if swift had conditional compilation based on pointer size, a
number of the test cases only care if the target is 32 or 64 bit.

gribozavr · 2016-01-15T00:05:27Z

Is powerpc64 big-endian?

antonblanchard · 2016-01-15T00:08:41Z

Yes, powerpc64 is big endian. We hope to get big endian working, but for now are concentrating on the little endian port (powerpc64le).

gribozavr · 2016-01-15T00:13:15Z

include/swift/Runtime/Metadata.h

@@ -919,7 +919,7 @@ static const uintptr_t ObjCReservedBitsMask =
 static const unsigned ObjCReservedLowBits =
  SWIFT_ABI_X86_64_OBJC_NUM_RESERVED_LOW_BITS;

-#elif defined(__arm64__)
+#elif defined(__arm64__) || defined(__powerpc64__)


Is __powerpc64__ defined on BE, LE or both?

It is defined on both BE and LE targets:

$ uname -m ppc64le $ cpp -dM < /dev/null | grep __powerpc64__ #define __powerpc64__ 1

$ uname -m ppc64 $ cpp -dM < /dev/null | grep __powerpc64__ #define __powerpc64__ 1

The comment above this block says that lib/IRGen/SwiftTargetInfo.cpp needs adjustments (you need to add a configurePowerPC function there).

gribozavr · 2016-01-15T00:22:31Z

How far does this patch allows the port to go? Does the standard library build? Does hello world work? Do the tests run?

antonblanchard · 2016-01-15T01:00:55Z

The powerpc64le port builds and runs the test suite with a few failures:

********************
Testing Time: 275.66s
********************
Failing Tests (5):
    Swift :: 1_stdlib/VarArgs.swift
    Swift :: Driver/filelists.swift
    Swift :: IRGen/c_layout.sil
    Swift :: IRGen/type_layout_reference_storage.swift
    Swift :: Parse/BOM.swift

  Expected Passes    : 1698
  Expected Failures  : 67
  Unsupported Tests  : 651
  Unexpected Failures: 5
*** Failed while running tests for swift (check-swift-linux-powerpc64le)

And hello world works:

$ cat hello.swift 
print("Hello, world!")
$ swiftc hello.swift 
$ ./hello 
Hello, world!

gribozavr · 2016-01-15T01:03:45Z

@antonblanchard I'm really impressed!

gribozavr · 2016-01-15T01:04:13Z

Did you try the validation test suite build-script -T?

antonblanchard · 2016-01-15T01:31:28Z

Thanks, that's given me a few more failing tests to sort out:

********************
Testing Time: 1467.18s
********************
Failing Tests (8):
    Swift :: 1_stdlib/VarArgs.swift
    Swift :: Driver/filelists.swift
    Swift :: IRGen/c_layout.sil
    Swift :: IRGen/type_layout_reference_storage.swift
    Swift :: Parse/BOM.swift
    Swift :: stdlib/AtomicInt.swift
    Swift :: stdlib/CollectionType.swift.gyb
    Swift :: stdlib/FloatingPointConversionTraps.swift.gyb

  Expected Passes    : 6961
  Expected Failures  : 79
  Unsupported Tests  : 771
  Unexpected Failures: 8

gribozavr · 2016-01-15T01:33:34Z

Again, very impressed it is just 8 tests in total!

antonblanchard · 2016-01-15T01:41:06Z

Yes, I'm surprised by how clean the port is. Having the compiler written in c++ (vs the target language), and pulling in clang to get the ABI parameter passing rules helps. Other LLVM based languages have required a lot more work.

Big endian still requires work, but we can attack that once the little endian port is solid.

antonblanchard · 2016-01-15T01:43:04Z

I'm not sure of the preferred workflow, but I'm happy to rebase and squash these commits.

gribozavr · 2016-01-15T01:48:42Z

stdlib/public/core/FixedPoint.swift.gyb

    self = ${Self}(Builtin.int_bswap_${BuiltinName}(value._value) )
 #else
-    _UnsupportedArchitectureError()


Please keep the unsupported error so that we get a loud error for new architectures. Just add an explicit #elseif arch(powerpc64) for BE (here and below).

gribozavr · 2016-01-15T01:49:23Z

Yes, we prefer rebasing/squashing.

gribozavr · 2016-01-15T05:13:09Z

lib/Basic/LangOptions.cpp

+    break;
+  case llvm::Triple::ArchType::ppc64le:
+    addTargetConfigOption("arch", "powerpc64le");
+    break;


This part needs a test like test/BuildConfigurations/x64LinuxTarget.swift for each supported triple.

Thanks, will add these tests

gribozavr · 2016-01-15T05:14:01Z

@antonblanchard Thank you! Looks good, but I have two more comments.

lattner · 2016-01-15T05:39:13Z

Very cool, thank you for working on adding this!

This patch adds powerpc64le Linux support. While the patch also adds the matching powerpc64 bits, there are endian issues that need to be sorted out. The PowerPC LLVM changes for the swift ABI (eg returning three element non-homogeneous aggregates) are still in the works, but a simple LLVM fix to allow those aggregates results in swift passing all but 8 test cases.

gribozavr · 2016-01-16T01:37:28Z

@antonblanchard The ObjC values should be dynamically dead as far as I understand. If you find they are also statically unused -- feel free to remove them!

gribozavr · 2016-01-16T02:14:43Z

I ran tests on Ubuntu 14.04 x86_64 and OS X, and see no regressions. Merging. Thank you!

Add powerpc64le Linux support

antonblanchard · 2016-01-16T07:21:10Z

Thanks for your help @gribozavr!

gribozavr reviewed Jan 15, 2016
View reviewed changes

antonblanchard force-pushed the powerpc64_merge branch from 1e3c7fb to 8ce2507 Compare January 15, 2016 02:07

gribozavr reviewed Jan 15, 2016
View reviewed changes

antonblanchard force-pushed the powerpc64_merge branch from 8ce2507 to b1827d8 Compare January 15, 2016 06:51

gribozavr added a commit that referenced this pull request Jan 16, 2016

Merge pull request #979 from antonblanchard/powerpc64_merge

5a1f8b2

Add powerpc64le Linux support

gribozavr merged commit 5a1f8b2 into apple:master Jan 16, 2016

Add powerpc64le Linux support #979

Add powerpc64le Linux support #979

antonblanchard commented Jan 14, 2016

gribozavr commented Jan 15, 2016

antonblanchard commented Jan 15, 2016

gribozavr Jan 15, 2016

antonblanchard Jan 15, 2016

gribozavr Jan 15, 2016

gribozavr commented Jan 15, 2016

antonblanchard commented Jan 15, 2016

gribozavr commented Jan 15, 2016

gribozavr commented Jan 15, 2016

antonblanchard commented Jan 15, 2016

gribozavr commented Jan 15, 2016

antonblanchard commented Jan 15, 2016

antonblanchard commented Jan 15, 2016

gribozavr Jan 15, 2016

gribozavr commented Jan 15, 2016

gribozavr Jan 15, 2016

antonblanchard Jan 15, 2016

gribozavr commented Jan 15, 2016

lattner commented Jan 15, 2016

gribozavr commented Jan 16, 2016

gribozavr commented Jan 16, 2016

antonblanchard commented Jan 16, 2016

Add powerpc64le Linux support #979

Add powerpc64le Linux support #979

Conversation

antonblanchard commented Jan 14, 2016

gribozavr commented Jan 15, 2016

antonblanchard commented Jan 15, 2016

gribozavr Jan 15, 2016

Choose a reason for hiding this comment

antonblanchard Jan 15, 2016

Choose a reason for hiding this comment

gribozavr Jan 15, 2016

Choose a reason for hiding this comment

gribozavr commented Jan 15, 2016

antonblanchard commented Jan 15, 2016

gribozavr commented Jan 15, 2016

gribozavr commented Jan 15, 2016

antonblanchard commented Jan 15, 2016

gribozavr commented Jan 15, 2016

antonblanchard commented Jan 15, 2016

antonblanchard commented Jan 15, 2016

gribozavr Jan 15, 2016

Choose a reason for hiding this comment

gribozavr commented Jan 15, 2016

gribozavr Jan 15, 2016

Choose a reason for hiding this comment

antonblanchard Jan 15, 2016

Choose a reason for hiding this comment

gribozavr commented Jan 15, 2016

lattner commented Jan 15, 2016

gribozavr commented Jan 16, 2016

gribozavr commented Jan 16, 2016

antonblanchard commented Jan 16, 2016