Windows defines NULL to 0, which when used as an argument to a variadic
function, is not a null pointer constant. As a result, Clang's
-Wsentinel fires on this code. Using '0' would be wrong on most 64-bit
platforms, but both MSVC and Clang make it work on Windows. Sidestep the
issue with nullptr.
llvm-svn: 221940
in-lane shuffles that aren't always handled well by the current vector
shuffle lowering.
No functionality change yet, that will follow in a subsequent commit.
llvm-svn: 221938
of new warning for deprecated method call for receiver
of type 'id'. This addresses rdar://18960378 where
unintended warnings being issued.
llvm-svn: 221933
Summary:
As a follow-up to D6167, this patch fixes the test to account for
another difference between PowerPC and x86 systems.
On x86 if you branch to an invalid address, you get a SIGSEGV with PC
set to the invalid address (and si_addr in the siginfo struct also set
to the same address).
On PowerPC, you get a SIGSEGV with PC pointing at the branch
instruction, but si_addr set to the invalid address.
You can see this difference if you run the test case under gdb.
Reviewers: kcc, glider, samsonov
Reviewed By: samsonov
Subscribers: llvm-commits
Differential Revision: http://reviews.llvm.org/D6226
llvm-svn: 221929
Summary: These checksums are ignored by llgs but some implementations require them to be specified properly.
Test Plan: Re-run llgs tests with the checksums and make sure we don't break anything.
Reviewers: tfiala, clayborg
Subscribers: lldb-commits
Differential Revision: http://reviews.llvm.org/D6254
llvm-svn: 221927
The generic FastISel code would bail, because it can't emit a sign-extend for
AArch64. This copies the code over and uses AArch64 specific emit functions.
This is not ideal and 'computeAddress' should handles this, so it can fold the
address computation into the memory operation.
I plan to clean up 'computeAddress' anyways, so I will add that in a future
commit.
Related to rdar://problem/18962471.
llvm-svn: 221923
These were directly using the old base instruction
class, and specifying the wrong register classes
for operands. The operands can be the other special
inputs besides SGPRs. The op name was also being
directly used for the asm string, so this was printed
without any operands.
llvm-svn: 221921
If a function is just an unreachable, this would hit a
"this is not a MachO target" assertion because of setting
HasSubsectionViaSymbols.
llvm-svn: 221920
e.g. v_mad_f32 a, b, c -> v_mad_f32 b, a, c
This simplifies matching v_madmk_f32.
This looks somewhat surprising, but it appears to be
OK to do this. We can commute src0 and src1 in all
of these instructions, and that's all that appears
to matter.
llvm-svn: 221910
The issues were:
- If you called this function with any arch other than the default target architecture, creating the target would fail because the Target::GetDefaultArchitecture() would not match the single architecture in the file specified. This caused running the test suite remotely with lldb-platform to fail many many tests due to the bad target.
- It would specify the currently selected platform which might not work for the specified platform
All other SBDebugger::CreateTarget calls do not assume an architecture or platform and if they aren't specified, they don't auto select the wrong one for you.
With this fix, SBTarget SBDebugger::CreateTarget (const char *filename) now behaves like the other SBDebugger::CreateTarget() variants.
llvm-svn: 221908
Normally entries can only move to a lower address, but when that wasn't viable,
the user's block was considered anyway. Unfortunately, it went via
createNewWater which wasn't designed to handle the case where there's already
an island after the block.
Unfortunately, the test we have is slow and fragile, and I couldn't reduce it
to anything sane even with the @llvm.arm.space intrinsic. The test change here
is recreating the previous one after the change.
rdar://problem/18545506
llvm-svn: 221905
We were using a naive heuristic to determine whether a basic block already had
an unconditional branch at the end. This mostly corresponded to reality
(assuming branches got optimised) because there's not much point in a branch to
the next block, but could go wrong.
llvm-svn: 221904
Creating tests for the ConstantIslands pass is very difficult, since it depends
on precise layout details. Having the ability to precisely inject a number of
bytes into the stream helps greatly.
llvm-svn: 221903
namespace blocks act as if KeepEmptyLinesAtTheStartOfBlocks is always true,
and aren't collapsed to a single line even if they would fit. Do the same
for extern "C" blocks.
Before,
extern "C" {
void ExternCFunction();
}
was collapsed into `extern "C" { void ExternCFunction(); }`. Now it stays like
it was.
Fixes http://crbug.com/432640 and part of PR21419.
llvm-svn: 221897
Reid Kleckner