| Commit message (Collapse) | Author | Age | Files | Lines |
| ... | |
| |
|
|
|
|
|
|
|
|
| |
Reviewers: sdardis
Subscribers: petarj, mpf, dsanders, spetrovic, llvm-commits, sdardis
Differential Revision: http://reviews.llvm.org/D18899
llvm-svn: 269283
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
needsRelocateWithSymbol()
Summary:
This eliminates the default case for N64 that was left out of r269047.
The change to R_MIPS_SUB is needed in this patch to make this testable since
%lo(%neg(%gp_rel(foo))) and %hi(%neg(%gp_rel(foo))) remain the only ways to get
a compound relocation from the assembler.
Reviewers: sdardis, rafael
Subscribers: dsanders, llvm-commits, sdardis
Differential Revision: http://reviews.llvm.org/D20097
llvm-svn: 269280
|
| |
|
|
| |
llvm-svn: 269273
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
While promoting nodes in PPCTargetLowering::DAGCombineExtBoolTrunc, it is
possible for one of the nodes to be replaced by another. To make sure we do not
visit the deleted nodes, and to make sure we visit the replacement nodes, use a
list of HandleSDNodes to track the to-be-promoted nodes during the promotion
process.
The same fix has been applied to the analogous code in
PPCTargetLowering::DAGCombineTruncBoolExt.
Fixes PR26985.
llvm-svn: 269272
|
| |
|
|
|
|
|
|
|
| |
Shifts beyond the bitwidth are undef but SCCP resolved them to zero.
Instead, DTRT and resolve them to undef.
This reimplements the transform which caused PR27712.
llvm-svn: 269269
|
| |
|
|
| |
llvm-svn: 269268
|
| |
|
|
|
|
| |
Enabled by -force-precise-rotation-cost option
llvm-svn: 269267
|
| |
|
|
|
|
|
|
|
|
| |
Currently cost based loop rotation algo can only be turned on with
two conditions: the function has real profile data, and -precise-rotation-cost
flag is turned on. This is not convenient for developers to experiment
when profile is not available. Add a new option to force the new
rotation algorithm -force-precise-rotation-cost
llvm-svn: 269266
|
| |
|
|
|
|
|
|
|
|
|
|
|
| |
The promote alloca pass would attempt to promote an alloca with
a select, icmp, or phi user, even though the other operand was
from a non-promotable source, producing a select on two different
pointer types.
Only do this if we know that both operands derive from the same
alloca. In the future we should be able to relax this to an alloca
which will also be promoted.
llvm-svn: 269265
|
| |
|
|
|
|
| |
I figure if I'm adding Mach support I may as well use the new fancy Error model.
llvm-svn: 269264
|
| |
|
|
| |
llvm-svn: 269263
|
| |
|
|
| |
llvm-svn: 269262
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This new verifier rule lets us unambigously pick a calling convention
when creating a new declaration for
`@llvm.experimental.deoptimize.<ty>`. It is also congruent with our
lowering strategy -- since all calls to `@llvm.experimental.deoptimize`
are lowered to calls to `__llvm_deoptimize`, it is reasonable to enforce
a unique calling convention.
Some of the tests that were breaking this verifier rule have had to be
split up into different .ll files.
The inliner was violating this rule as well, and has been fixed to avoid
producing invalid IR.
llvm-svn: 269261
|
| |
|
|
|
|
|
| |
While not impossible, it's unlikely we'd be performing bitwise operations on FP
values.
llvm-svn: 269260
|
| |
|
|
|
|
|
|
|
|
|
|
| |
This is a large change, but it's pretty mechanical:
- Where we were returning a node before, call ReplaceNode instead.
- Where we would return null to fall back to another selector, rename
the method to try* and return a bool for success.
- Where we were calling SelectNodeTo, just return afterwards.
Part of llvm.org/pr26808.
llvm-svn: 269258
|
| |
|
|
|
|
|
|
|
| |
When we convert to the void Select interface, leaving unreferenced
nodes around won't be allowed anymore.
Part of llvm.org/pr26808.
llvm-svn: 269256
|
| |
|
|
|
|
|
|
| |
Do not precompute value counts for all sites. This
eliminates one more use of dynamic allocation
in profiler writer.
llvm-svn: 269253
|
| |
|
|
|
|
|
|
| |
defined."
This reverts commit r269105 as it caused PR27712.
llvm-svn: 269252
|
| |
|
|
|
|
|
|
|
| |
This should just be a compile-time change. Correct the check for whether
we have already analyzed the callee when making summary based decisions.
There is no need to reprocess one at the same threshold as when it was
last processed.
llvm-svn: 269251
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This is to fix the bug in https://llvm.org/bugs/show_bug.cgi?id=27612.
When spill is hoisted to a BB with landingpad successor, and if the VNI
of the spill reg lives into the landingpad successor, the spill should be
inserted before the call which may throw exception. InsertPointAnalysis
is used to compute the safe insert point.
http://reviews.llvm.org/D20027 is a preparing patch for this patch.
Differential Revision: http://reviews.llvm.org/D19884.
llvm-svn: 269249
|
| |
|
|
|
|
|
|
|
|
|
|
| |
InsertPointAnalysis.
Because both split and spill hoisting want to use LastSplitPoint computation
result, extract the LastSplitPoint computation from SplitAnalysis class which
also contains a bunch of other analysises only related to split.
Differential Revision: http://reviews.llvm.org/D20027.
llvm-svn: 269248
|
| |
|
|
|
|
|
|
|
| |
This was a typo in an earlier commit - there's no point in keeping the
old node around here.
Noticed by Meador Inge. Thanks!
llvm-svn: 269245
|
| |
|
|
|
|
|
|
| |
Adding the initial files for adding MachO support to yaml2obj. Passing a MachO file will result in an error.
I will be implementing obj2yaml and yaml2obj for MachO in parallel so that one can be used to test the other.
llvm-svn: 269244
|
| |
|
|
|
|
|
|
| |
Adding the initial files for adding MachO support to obj2yaml. Passing a MachO file will result in a new not_implemented error.
I will be implementing obj2yaml and yaml2obj for MachO in parallel so that one can be used to test the other.
llvm-svn: 269243
|
| |
|
|
| |
llvm-svn: 269241
|
| |
|
|
|
|
|
| |
Do not use getVRegDef() to print "the definition" of a vreg. If there
are multiple or none the function will fail.
llvm-svn: 269239
|
| |
|
|
| |
llvm-svn: 269238
|
| |
|
|
|
|
|
|
|
| |
It's very common to want to replace a node and then remove it since
it's dead, especially as we port backends from the SDNode *Select API
to the void Select one. This helper makes this sequence a bit less
verbose.
llvm-svn: 269236
|
| |
|
|
|
|
|
|
| |
It's awkward to force callers of SelectNodeTo to figure out whether
the node was morphed or CSE'd. Update uses here instead of requiring
callers to (sometimes) do it.
llvm-svn: 269235
|
| |
|
|
|
|
|
|
|
|
|
|
| |
This fixes a debug assert on Windows from the new iterator
implementation added in r269059. The Windows std::vector iterator
operator== checks in debug mode that the containers being iterated over
are the same, which they may not be.
Fixed by checking that we are iterating over the same container before
comparing the container iterators.
llvm-svn: 269232
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
For narrow stores (e.g., strb, srth) we know the upper bits of the register are
unused/not useful. In some cases we can use this information to eliminate
unnecessary instructions.
For example, without this patch we generate (from the 2nd test case):
ldr w8, [x0]
and w8, w8, #0xfff0
bfxil w8, w2, #16, #4
strh w8, [x1]
and after the patch the 'and' is removed:
ldr w8, [x0]
bfxil w8, w2, #16, #4
strh w8, [x1]
ret
During the lowering of the bitfield insert instruction the 'and' is eliminated
because we know the upper 16-bits that are masked off are unused and the lower
4-bits that are masked off are overwritten by the insert itself. Therefore, the
'and' is unnecessary.
Differential Revision: http://reviews.llvm.org/D20175
llvm-svn: 269226
|
| |
|
|
| |
llvm-svn: 269225
|
| |
|
|
|
|
| |
Differential Revision: http://reviews.llvm.org/D20082
llvm-svn: 269222
|
| |
|
|
|
|
| |
Fixed incorrect operands indices used to access src registers
llvm-svn: 269221
|
| |
|
|
|
|
| |
This is similar to how getName is handled.
llvm-svn: 269218
|
| |
|
|
| |
llvm-svn: 269217
|
| |
|
|
|
|
|
| |
Differential Revision: http://reviews.llvm.org/D20138
Reviewed By: rnk
llvm-svn: 269216
|
| |
|
|
|
|
|
|
| |
Don't bother returning a result we don't use here. I've also renamed
this from selectGather to tryGather to better indicate that it may not
do anything.
llvm-svn: 269215
|
| |
|
|
|
|
|
|
|
| |
SCEVExpander::replaceCongruentIVs assumes the backedge value of an
SCEV-analysable PHI to always be an instruction, when this is not
necessarily true. For now address this by bailing out of the
optimization if the backedge value of the PHI is a non-Instruction.
llvm-svn: 269213
|
| |
|
|
|
|
|
|
|
|
|
|
| |
`SCEVExpander::replaceCongruentIVs` bypasses `hoistIVInc` if both the
original and the isomorphic increments are PHI nodes. Doing this can
break SSA if the isomorphic increment is not dominated by the original
increment. Get rid of the bypass, and let `hoistIVInc` do the right
thing.
Fixes PR27232 (compile time crash/hang).
llvm-svn: 269212
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
... for AddRec's in loops for which SCEV is unable to compute a max
tripcount. This is not a problem for "normal" loops[0] that don't have
guards or assumes, but helps in cases where we have guards or assumes in
the loop that can be used to constrain incoming values over the backedge.
This partially fixes PR27691 (we still don't handle the NUW case).
[0]: for "normal" loops, in the cases where we'd be able to prove
no-wrap via isKnownPredicate, we'd also be able to compute a max
tripcount.
llvm-svn: 269211
|
| |
|
|
| |
llvm-svn: 269209
|
| |
|
|
|
|
|
|
|
|
| |
Reenable R600 testing
reviewer: arsenm
Differential Revision: http://reviews.llvm.org/D20031
llvm-svn: 269207
|
| |
|
|
| |
llvm-svn: 269206
|
| |
|
|
|
|
|
| |
DbgInfoIntrinsic::StripCast() is dead since r79977
The only function that creates Comdat objects seems to be in Module, and always creates them using the default constructor.
llvm-svn: 269204
|
| |
|
|
| |
llvm-svn: 269203
|
| |
|
|
| |
llvm-svn: 269199
|
| |
|
|
|
|
|
|
|
|
|
|
| |
Equivalent GEP indices with different types are treated as different
indices altogether, leading to an incorrect AA result. Fix the issue
by comparing indices based on their values.
Thanks to Mikael Holmén for reporting the issue!
Differential Revision: http://reviews.llvm.org/D19935
llvm-svn: 269197
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
r268900
microMIPS has a special case that is not correctly implemented in LLVM. If we
have a symbol 'foo' which is equivalent to '.text+0x10'. The value of an
R_MICROMIPS_LO16 relocation using 'foo' is 'foo+0x11' and not 'foo+0x10'. The
in-place addend should therefore be 0x11.
Work around this by partially reverting the effect of r268900 by keeping the
symbol when the STO_MIPS_MICROMIPS flag is set. This fixes
SingleSource/Regression/C/PR640 for microMIPS.
llvm-svn: 269196
|
| |
|
|
| |
llvm-svn: 269193
|