Thread the determination of branch prediction hit rates back through the if-conversion heuristic APIs.  For now,
stick with a constant estimate of 90% (branch predictors are good!), but we might find that we want to provide
more nuanced estimates in the future.

llvm-svn: 115364

InstPrinter) subdir

llvm-svn: 115363

llvm-svn: 115362

llvm-svn: 115361

been MC-ized for assembly printing. MSP430 is mostly so, but still has the
asm printer and lowering code in the printer subdir for the moment.

llvm-svn: 115360

delay. Anton and PIC16 folks, if this is still good to keep, please go ahead
and add it back in with an updated comment about when would be a good time
to revisit.

llvm-svn: 115358

llvm-svn: 115356

a helper function (AdjustAPSInt) and use that
for adjusting the high bounds of case ranges
before APSInt comparisons. Fixes
http://llvm.org/bugs/show_bug.cgi?id=8135

Some minor refacorings while I am here.

llvm-svn: 115355

llvm-svn: 115354

llvm-svn: 115353

Need to specify SSE4 for machines which don't have SSE4. The code checked for is generated by SSE4. Otherwise, we get something else.

llvm-svn: 115352

llvm-svn: 115351

llvm-svn: 115350

llvm-svn: 115349

llvm-svn: 115348

Better diagnostic for superfluous scope specifier inside a class definition for member functions. + Fixit.

Example: 
class A {
   void A::foo(); //warning: extra qualification on member 'foo'
};

llvm-svn: 115347

and protocol expressions. Fixes <rdar://problem/7833565>.

llvm-svn: 115346

UnreachableCodeChecker does not need to inherit from CheckerVisitor, only Checker. It does not use any AST Stmt hooks, only VisitEndAnalysis.

llvm-svn: 115345

llvm-svn: 115344

its containing class was parsed, we would crash.

llvm-svn: 115343

llvm-svn: 115342

Make the spelling of the flags for old-style if-conversion heuristics consistent between ARM and Thumb2.

llvm-svn: 115341

we're missing the corresponding changes in the LLVM repository.

llvm-svn: 115340

llvm-svn: 115339

member function or an Objective-C method, which other member
functions/methods it overrides.

llvm-svn: 115338

Now that the profitable bits of EnableFullLoadPRE have been enabled by default, rip out the remainder.
Anyone interested in more general PRE would be better served by implementing it separately, to get real
anticipation calculation, etc.

llvm-svn: 115337

llvm-svn: 115336

Modify command options to use the new arguments mechanism.  Now all command option
arguments are specified in a standardized way, will have a standardized name, and
have functioning help.

The next step is to start writing useful help for all the argument types.

llvm-svn: 115335

llvm-svn: 115334

llvm-svn: 115333

Per Cortex-A9 pipeline diagram. AGU (core load / store issue) and NEON/FP issue are multiplexed. Model it correctly.

llvm-svn: 115332

llvm-svn: 115326

llvm-svn: 115325

Fix code gen crash reported in PR 8235. We still lose debug info for the unused argument here. This is a known limitation recorded debuginfo-tests/trunk/dbg-declare2.ll function 'f6' test case.

llvm-svn: 115323

  auto f(int) -> int

from Daniel Wallin!

(With a few minor bug fixes from me).

llvm-svn: 115322

Also, the break by location needs to be modified after changees made to the main.cpp source code.

llvm-svn: 115320

Add infrastructure for standardizing arguments for commands and
command options; makes it easier to ensure that the same type of
argument will have the same name everywhere, hooks up help for command
arguments, so that users can ask for help when they are confused about
what an argument should be; puts in the beginnings of the ability to
do tab-completion for certain types of arguments, allows automatic
syntax help generation for commands with arguments, and adds command
arguments into command options help correctly.

Currently only the breakpoint-id and breakpoint-id-range arguments, in
the breakpoint commands, have been hooked up to use the new mechanism.
The next steps will be to fix the command options arguments to use
this mechanism, and to fix the rest of the regular command arguments
to use this mechanism.  Most of the help text is currently missing or
dummy text; this will need to be filled in, and the existing argument
help text will need to be cleaned up a bit (it was thrown in quickly,
mostly for testing purposes).

Help command now works for all argument types, although the help may not
be very helpful yet.

Those commands that take "raw" command strings now indicate it in their
help text.

llvm-svn: 115318

llvm-svn: 115317

llvm-svn: 115316

llvm-svn: 115314