- Change register allocation hint to a pair of unsigned integers. The hint type is zero (which means prefer the register specified as second part of the pair) or entirely target dependent.
- Allow targets to specify alternative register allocation orders based on allocation hint.

Part 2.
- Use the register allocation hint system to implement more aggressive load / store multiple formation.
- Aggressively form LDRD / STRD. These are formed *before* register allocation. It has to be done this way to shorten live interval of base and offset registers. e.g.
v1025 = LDR v1024, 0
v1026 = LDR v1024, 0
=>
v1025,v1026 = LDRD v1024, 0

If this transformation isn't done before allocation, v1024 will overlap v1025 which means it more difficult to allocate a register pair.

- Even with the register allocation hint, it may not be possible to get the desired allocation. In that case, the post-allocation load / store multiple pass must fix the ldrd / strd instructions. They can either become ldm / stm instructions or back to a pair of ldr / str instructions.

This is work in progress, not yet enabled.

llvm-svn: 73381

llvm-svn: 73380

Patch by Benjamin Kramer!

llvm-svn: 73379

ranges more similar to the console output.  Consider:

#define FOO(X, Y) X/ Y

void foo(int *P, int *Q) {
  FOO(P, Q);
}

Before we emitted:

t.c:4:3:{4:3-4:6}{4:3-4:6}: error: invalid operands to binary expression ('int *' and 'int *')
   FOO(P, Q);
   ^~~~~~~~~
...

Note that while we underline the macro args that the range info just includes FOO 
without its macros.  This change teaches the printed ranges to include macro args
also so that we get:

t.c:4:3:{4:3-4:12}{4:3-4:12}: error: invalid operands to binary expression ('int *' and 'int *')
   FOO(P, Q);
   ^~~~~~~~~
...

This fixes rdar://6939599

llvm-svn: 73378

the file at the point of the pragma.  This allows clang to know that all
sourcelocations after the pragma are in a system header.

llvm-svn: 73376

llvm-svn: 73375

llvm-svn: 73374

llvm-svn: 73373

llvm-svn: 73372

like:
fatal error: could not find file '1.h' referenced by PCH file
instead of aborting with an assertion failure, PR4219

llvm-svn: 73371

llvm-svn: 73370

llvm-svn: 73369

absolute or not, based on a patch by Gregory Curfman!

llvm-svn: 73368

llvm-svn: 73367

llvm-svn: 73366

can occur in the middle of comment tokens.

llvm-svn: 73365

"GCC emits an __objc_class_name_{classname} symbol for every class, and a corresponding reference to this symbol for every compilation unit that references the class.  This causes linker errors when you try linking a program which references some classes but doesn't define them.  The attached patch implements this support in clang, so you can compile a class with clang, reference it in a file compiled with GCC, and have it all work correctly."

Patch by David Chisnall!

llvm-svn: 73364

llvm-svn: 73362

they contain multiplications of constants with add operations.
This helps simplify several kinds of things; in particular it
helps simplify expressions like ((-1 * (%a + %b)) + %a) to %b,
as expressions like this often come up in loop trip count
computations.

llvm-svn: 73361

used as keys in DenseMaps.

llvm-svn: 73360

the potentially expensive erase.

llvm-svn: 73359

even though the order doesn't matter at the top level of an expression,
it does matter when the constant is a subexpression of an n-ary
expression, because n-ary expressions are sorted lexicographically.

llvm-svn: 73358

SmallVector instead of std::vector.

llvm-svn: 73357

like "int x = + +3;".

llvm-svn: 73356

llvm-svn: 73355

llvm-svn: 73353

static intializers for structs.

llvm-svn: 73349

Move register allocation preference (or hint) from LiveInterval to MachineRegisterInfo. This allows more passes to set them.

llvm-svn: 73346

Patch by Jean-Daniel Dupas

llvm-svn: 73343

llvm-svn: 73342

llvm-svn: 73341

is that, for functions whose bodies are entirely guarded by an if-statement, it
can be profitable to pull the test out of the callee and into the caller.

This code has had some cursory testing, but still has a number of known issues
on the LLVM test suite.

llvm-svn: 73338

that were suppressed due to SFINAE. By checking whether any errors
occur at the end of template argument deduction, we avoid the
possibility of suppressing an error (due to SFINAE) and then
recovering so well that template argument deduction never detects that
there was a problem. Thanks to Eli for the push in this direction.

llvm-svn: 73336

Introduce new BinaryObject (blob) class, ELF Writer modified to use it. BinaryObject.h by Aaron Gray

llvm-svn: 73333

Implement support for C++ Substitution Failure Is Not An Error
(SFINAE), which says that errors that occur during template argument
deduction do *not* produce diagnostics and do not necessarily make a
program ill-formed. Instead, template argument deduction silently
fails. This is currently implemented for template argument deduction
during matching of class template partial specializations, although
the mechanism will also apply to template argument deduction for
function templates. The scheme is simple:

  - If we are in a template argument deduction context, any diagnostic
    that is considered a SFINAE error (or warning) will be
    suppressed. The error will be propagated up the call stack via the
    normal means.
  - By default, all warnings and errors are SFINAE errors. Add the
    NoSFINAE class to a diagnostic in the .td file to make it a hard
    error (e.g., for access-control violations).

Note that, to make this fully work, every place in Sema that emits an
error *and then immediately recovers* will need to check
Sema::isSFINAEContext() to determine whether it must immediately
return an error rather than recovering.

llvm-svn: 73332

llvm-svn: 73331

llvm-svn: 73323

foo has a constant address.

llvm-svn: 73321

llvm-svn: 73320

preprocessor and initialize it early in clang-cc.  This
ensures that __has_builtin works in all modes, not just
when ASTContext is around.

llvm-svn: 73319