| Commit message (Collapse) | Author | Age | Files | Lines |
| |
|
|
|
|
|
|
| |
Change MachineInstr API to prefer MachineInstr& over MachineInstr*
whenever the parameter is expected to be non-null. Slowly inching
toward being able to fix PR26753.
llvm-svn: 262149
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
In the case where op = add, y = base_ptr, and x = offset, this
transform:
(op y, (op x, c1)) -> (op (op x, y), c1)
breaks the canonical form of add by putting the base pointer in the
second operand and the offset in the first.
This fix is important for the R600 target, because for some address
spaces the base pointer and the offset are stored in separate register
classes. The old pattern caused the ISel code for matching addressing
modes to put the base pointer and offset in the wrong register classes,
which required no-trivial code transformations to fix.
llvm-svn: 262148
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Summary:
iOS on ARM64 doesn't unique RTTI.
Ref: clang's iOS64CXXABI::shouldRTTIBeUnique()
Due to this, pointer-equality will not necessarily work in this
architecture, across dylib boundaries.
dynamic_cast<>() will (as expected) still work, since Apple ships with
one prepared for this, but we can't rely on the type names being
pointer-equal.
I've limited the expensive strcmp check to the specific architecture
which needs it.
Example which triggers this bug:
lib.h:
struct X {
virtual ~X() {}
};
X *libCall();
lib.mm:
X *libCall() {
return new X;
}
prog.mm:
int main() {
X *px = libCall();
delete px;
}
Expected output: Nothing
Actual output:
<unknown>: runtime error: member call on address 0x00017001ef50 which does not point to an object of type 'X'
0x00017001ef50: note: object is of type 'X'
00 00 00 00 60 00 0f 00 01 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
^~~~~~~~~~~~~~~~~~~~~~~
vptr for ‘X’
Reviewers: kubabrecka, samsonov, eugenis, rsmith
Subscribers: aemerson, llvm-commits, rengolin
Differential Revision: http://reviews.llvm.org/D11502
llvm-svn: 262147
|
| |
|
|
|
|
| |
Also update HashEndOfMBB to take MachineBasicBlock&.
llvm-svn: 262146
|
| |
|
|
|
|
|
|
| |
Take parameters as MachineInstr& instead of MachineInstr* in
AntiDepBreaker API, since these are required to be non-null. No
functionality change intended. Looking toward PR26753.
llvm-svn: 262145
|
| |
|
|
|
|
|
| |
This already assumes a valid MI, since it dereferences the MI in an
assertion before checking for null. At an explicit assert.
llvm-svn: 262144
|
| |
|
|
| |
llvm-svn: 262143
|
| |
|
|
|
|
|
|
|
| |
In all but one case, change the DFAPacketizer API to take MachineInstr&
instead of MachineInstr*. In DFAPacketizer::endPacket(), take
MachineBasicBlock::iterator. Besides cleaning up the API, this is in
search of PR26753.
llvm-svn: 262142
|
| |
|
|
|
|
|
|
| |
Update APIs in MachineInstrBundle.h to take and return MachineInstr&
instead of MachineInstr* when the instruction cannot be null. Besides
being a nice cleanup, this is tacking toward a fix for PR26753.
llvm-svn: 262141
|
| |
|
|
|
|
| |
It was broken by the work for PR26753.
llvm-svn: 262140
|
| |
|
|
|
|
| |
This reverts commit r262103, as it broke all ARM and AArch64 bots.
llvm-svn: 262139
|
| |
|
|
| |
llvm-svn: 262138
|
| |
|
|
|
|
| |
NFCI.
llvm-svn: 262137
|
| |
|
|
| |
llvm-svn: 262136
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
in Sparc.
Previous check-in message was:
The patch adds missing registers and instructions to complete all the registers supported by the Sparc v8 manual.
These are all co-processor registers, with the exception of the floating-point deferred-trap queue register.
Although these will not be lowered automatically by any instructions, it allows the use of co-processor
instructions implemented by inline-assembly.
Code Reviewed at http://reviews.llvm.org/D17133, with the exception of a very small change in brace placement in SparcInstrInfo.td,
which was formerly causing a problem in the disassembly of the %fq register.
llvm-svn: 262135
|
| |
|
|
| |
llvm-svn: 262134
|
| |
|
|
|
|
|
|
|
|
|
|
|
| |
registers supported by the Sparc v8 manual.
These are all co-processor registers, with the exception of the floating-point deferred-trap queue register.
Although these will not be lowered automatically by any instructions, it allows the use of co-processor
instructions implemented by inline-assembly.
Code Reviewed at http://reviews.llvm.org/D17133, with the exception of a very small change in brace placement in SparcInstrInfo.td,
which was formerly causing a problem in the disassembly of the %fq register.
llvm-svn: 262133
|
| |
|
|
| |
llvm-svn: 262132
|
| |
|
|
| |
llvm-svn: 262131
|
| |
|
|
|
|
| |
with 'class' keywords in the template arguments and other silliness.
llvm-svn: 262130
|
| |
|
|
|
|
|
|
|
|
| |
manager as some compilers print the typedef name and others print the
"canonical" name of the underlying class template.
This isn't really an important artifact of the test anyways so it seems
fine to just loosen the test assertions here.
llvm-svn: 262129
|
| |
|
|
|
|
| |
PassManager and AnalysisManager template specializations as well.
llvm-svn: 262128
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
manager proxies and use those rather than repeating their definition
four times.
There are real differences between the two directions: outer AMs are
const and don't need to have invalidation tracked. But every proxy in
a particular direction is identical except for the analysis manager type
and the IR unit they proxy into. This makes them prime candidates for
nice templates.
I've started introducing explicit template instantiation declarations
and definitions as well because we really shouldn't be emitting all this
everywhere. I'm going to go back and add the same for the other
templates like this in a follow-up patch.
I've left the analysis manager as an opaque type rather than using two
IR units and requiring it to be an AnalysisManager template
specialization. I think its important that users retain the ability to
provide their own custom analysis management layer and provided it has
the appropriate API everything should Just Work.
llvm-svn: 262127
|
| |
|
|
| |
llvm-svn: 262126
|
| |
|
|
|
|
| |
This is OK for +0 since compares to +/-0 give the same result.
llvm-svn: 262125
|
| |
|
|
|
|
| |
amdgcn should not be defining __R600__
llvm-svn: 262124
|
| |
|
|
| |
llvm-svn: 262123
|
| |
|
|
| |
llvm-svn: 262122
|
| |
|
|
|
|
|
| |
This will be more useful for marking builtins acceptable for which
subtargets.
llvm-svn: 262121
|
| |
|
|
| |
llvm-svn: 262120
|
| |
|
|
|
|
|
|
|
|
| |
This matches the behavior of the HSAIL clock instruction.
s_realmemtime is used if the subtarget supports it, and falls
back to s_memtime if not.
Also introduces new intrinsics for each of s_memtime / s_memrealtime.
llvm-svn: 262119
|
| |
|
|
|
|
|
|
| |
Avoid another implicit conversion from MachineInstrBundleIterator to
MachineInstr*, this time in MachineInstrBuilder.h (this is in pursuit of
PR26753).
llvm-svn: 262118
|
| |
|
|
|
|
|
|
|
| |
This debug output distracts from the -debug-only=polly-scops output. As it is
rather verbose and only really needed for debugging the domain construction
I drop this output. The domain construction is meanwhile stable enough to
not require regular debugging.
llvm-svn: 262117
|
| |
|
|
|
|
|
| |
Remove a couple of implicit conversions from MachineInstrBundleIterator
to MachineInstr*.
llvm-svn: 262116
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
| |
Take MachineInstr by reference instead of by pointer in SlotIndexes and
the SlotIndex wrappers in LiveIntervals. The MachineInstrs here are
never null, so this cleans up the API a bit. It also incidentally
removes a few implicit conversions from MachineInstrBundleIterator to
MachineInstr* (see PR26753).
At a couple of call sites it was convenient to convert to a range-based
for loop over MachineBasicBlock::instr_begin/instr_end, so I added
MachineBasicBlock::instrs.
llvm-svn: 262115
|
| |
|
|
|
|
|
| |
Some of this functionality is useful beyond the generation of a normal OpenMP
loop.
llvm-svn: 262114
|
| |
|
|
|
|
|
|
|
|
|
|
| |
In case the underlying basepointer of a ScopArrayInfo object is moved to another
module while the scop is still processed is it necessary to free dependent
ScopArrayInfo objects as they might otherwise be looked accidentally when a
new llvm basepointer value is reassigned the very same memory location as the
llvm value that has been moved earlier.
This function is not yet used in Polly itself, but is useful for external users.
llvm-svn: 262113
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Summary:
The PS4 linker seems to handle this fine.
Hi David, it seems that indeed most ELF linkers support
__{start,stop}_SECNAME, as our proprietary linker does as well.
This follows the pattern of r250679 w.r.t. the testing.
Maggie, Phillip, Paul: I've tested this with the PS4 SDK 3.5 toolchain
prerelease and it seems to work fine.
Reviewers: davidxl
Subscribers: probinson, phillip.power, MaggieYi
Differential Revision: http://reviews.llvm.org/D17672
llvm-svn: 262112
|
| |
|
|
| |
llvm-svn: 262111
|
| |
|
|
|
|
| |
-fsanitize-coverage=trace-pc mode; update libFuzzer doc for previous commit
llvm-svn: 262110
|
| |
|
|
| |
llvm-svn: 262109
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
merged into a loop that was subsequently unrolled (or otherwise nuked).
In this case it can't merge in the ASTs for any remaining nested loops,
it needs to re-add their instructions dircetly.
The fix is very isolated, but I've pulled the code for merging blocks
into the AST into a single place in the process. The only behavior
change is in the case which would have crashed before.
This fixes a crash reported by Mikael Holmen on the list after r261316
restored much of the loop pass pipelining and allowed us to actually do
this kind of nested transformation sequenc. I've taken that test case
and further reduced it into the somewhat twisty maze of loops in the
included test case. This does in fact trigger the bug even in this
reduced form.
llvm-svn: 262108
|
| |
|
|
| |
llvm-svn: 262106
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
with a prior UsingDecl -- those should not even really be found by the lookup
here, except that we use the same lookup results for two different checks, and
the other check needs them.
This happens to work in *almost all* cases, because either the lookup results
list the UsingDecl first (and the NonTag result gets replaced by something
else) or because the problematic declaration is a function (which causes us to
use different logic to detect conflicts). This can also be triggered from a
state only reachable through modules (where the name lookup results can contain
multiple UsingDecls in the same scope).
llvm-svn: 262105
|
| |
|
|
|
|
| |
Differential Revision: http://reviews.llvm.org/D17670
llvm-svn: 262104
|
| |
|
|
|
|
|
|
|
|
|
| |
Summary:
Without tree pruning clang has 2,667,552 points.
Wiht only dominators pruning: 1,515,586.
With both dominators & predominators pruning: 1,340,534.
Differential Revision: http://reviews.llvm.org/D17671
llvm-svn: 262103
|
| |
|
|
| |
llvm-svn: 262102
|
| |
|
|
| |
llvm-svn: 262101
|
| |
|
|
|
|
| |
Differential Revision: http://reviews.llvm.org/D17250
llvm-svn: 262100
|
| |
|
|
| |
llvm-svn: 262096
|