♟ gchatelet

@arphaman do you mind having a look at the MIR parser / writer parts?
How is alignment supposed to be serialized for MIR? As power of two or a log of alignment?

Rebasing
Update documentation flags.

[LLVM][NFC] remove unused fields

Also removes JumpBufSize
Rebasing fix

[LLVM][Alignment] Introduce Alignment In MachineFrameInfo

Thx for working on this Teresa.

In D65859#1629358, @RKSimon wrote:

This isn't handling all the cases - its probably better to push at the top of the file (before namespace) and similarly pop at the end of the file

In D65800#1621111, @kparzysz wrote:

The Hexagon part looks ok.

Using Align for LocalFrameMaxAlign instead of MaybeAlign

Nice!
Please wait for other reviewers to have a look.

Also removes JumpBufSize

All the tests are passing but I'm not entirely sure of the validity of this patch.
Alignment, StackAlignment seems to be always defined alignment (aka. Align) but I'm not so sure of MaxAlignment and LocalFrameMaxAlign.
Hexagon/HexagonFrameLowering.cpp seem to rely on the fact that LocalFrameMaxAlign is undefined.
I'm not sure about MaxAlignment.

Converting all alignements in MachineFrameInfo.

[LLVM][Alignment] Introduce Alignment In Attributes

clang-format before submitting

[LLVM][Alignment] Introduce Alignment In GlobalObject

Added FIXME and used neither instead of not

[LLVM][Alignment] Introduce Alignment Type

[LLVM][Alignment] Introduce Alignment In CallingConv

[LLVM][Alignment] Introduce Alignment Type in DataLayout

In rG2bea69bf6503ffc9f3cde9a52b5dac1a25e94e1c#676020, @dsanders wrote:

The fixits indicate the right change to make (TargetRegisterInfo -> Register). I've applied them in r367697

By the way, do you have a bot? I didn't get any failure notices for ARC.

I will push this on Monday 5th of Aug. Let me know if you have any concerns.

The test appears to fail when the compilation path contains the debug_ substring.
This is not ideal.

Also changed getFunctionPtrAlign() return type to MaybeAlign.
Keeping getStackAlignment() return type as unsigned for now.

This patch broke ARC experimental target, activated by -DLLVM_EXPERIMENTAL_TARGETS_TO_BUILD="ARC"

/llvm-project/llvm/lib/Target/ARC/ARCOptAddrMode.cpp:142:10: error: no member named 'isVirtualRegister' in 'llvm::TargetRegisterInfo'; did you mean 'llvm::Register::isVirtualRegister'?
  assert(TargetRegisterInfo::isVirtualRegister(VReg) &&
         ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
         llvm::Register::isVirtualRegister
/usr/include/assert.h:93:27: note: expanded from macro 'assert'
     (static_cast <bool> (expr)                                         \
                          ^~~~
/llvm-project/llvm/include/llvm/CodeGen/Register.h:68:15: note: 'llvm::Register::isVirtualRegister' declared here
  static bool isVirtualRegister(unsigned Reg) {
              ^
/llvm/lib/Target/ARC/ARCOptAddrMode.cpp:209:27: error: no member named 'isStackSlot' in 'llvm::TargetRegisterInfo'
  if (TargetRegisterInfo::isStackSlot(B) ||
      ~~~~~~~~~~~~~~~~~~~~^
/llvm-project/llvm/lib/Target/ARC/ARCOptAddrMode.cpp:210:8: error: no member named 'isVirtualRegister' in 'llvm::TargetRegisterInfo'; did you mean 'llvm::Register::isVirtualRegister'?
      !TargetRegisterInfo::isVirtualRegister(B)) {
       ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
       llvm::Register::isVirtualRegister
/llvm-project/llvm/include/llvm/CodeGen/Register.h:68:15: note: 'llvm::Register::isVirtualRegister' declared here
  static bool isVirtualRegister(unsigned Reg) {

[LLVM][Alignment] Update documentation

Thx for the review, I'll proceed with the push.

[LLVM][Alignment] Fix AlignmentTest on platform where size_t != uint64_t

Make getFunctionPtrAlign() return MaybeAlign

@aprantl I've created https://reviews.llvm.org/D65558 to address your comments since this patch is already submitted.

In D64790#1608236, @RKSimon wrote:

@gchatelet This is breaking windows bots: http://lab.llvm.org:8011/builders/llvm-clang-x86_64-win-fast/builds/1933 - please can you take a look? I think we'll need to revert until you have a fix.

[LLVM] Fix Alignment death tests in Release Mode

In D64790#1608212, @foad wrote:

This also broke ninja check-llvm-unit on a Release build:

Fix mismatching vector type in AlignmentTest.cpp

This patch broke buildbot, I've submitted to fix MSVC https://reviews.llvm.org/rL367400

Add default constructor for MaybeAlign

[LLVM][NFC] Adding an Alignment type to LLVM

Thank you for the review @jfb and @jakehehrlich. I'll submit the patch now and start working on the migration.
Let me know if you have additional comments.

Remove final from Align and MaybeAlign

Assert when Log2 on undefined MaybeAlign.

Thx for the review @jfb. Anything else?

Make MaybeAlign an llvm::Optional.
Remove unnecessary documentation,

gchatelet (Guillaume Chatelet)
User

Projects

User Details

Recent Activity
View All

Today

Wed, Aug 21

Tue, Aug 20

Mon, Aug 19

Thu, Aug 15

Fri, Aug 9

Thu, Aug 8

Wed, Aug 7

Tue, Aug 6

Mon, Aug 5

Fri, Aug 2

Thu, Aug 1

Wed, Jul 31

Tue, Jul 30

Mon, Jul 29

gchatelet (Guillaume Chatelet)User