Commit Graph

103954 Commits

Author SHA1 Message Date
Eric Christopher
162180cba5 Reset the subtarget for DAGToDAG on every iteration of runOnMachineFunction.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@209374 91177308-0d34-0410-b5e6-96231b3b80d8
2014-05-22 01:07:21 +00:00
Eric Christopher
14fb9b6069 Sort includes.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@209373 91177308-0d34-0410-b5e6-96231b3b80d8
2014-05-22 01:07:18 +00:00
David Blaikie
b1a3b17e78 DebugInfo: Simplify dead variable collection slightly.
constructSubprogramDIE was already called for every subprogram in every
CU when the module was started - there's no need to call it again at
module finalization.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@209372 91177308-0d34-0410-b5e6-96231b3b80d8
2014-05-22 00:48:36 +00:00
Andrew Trick
facca6e3f3 Fix a bug in SCEV's backedge taken count computation from my prior fix in Jan.
This has to do with the trip count computation for loops with multiple
exits, which is quite subtle. Most passes just ask for a single trip
count number, so we must be conservative assuming any exit could be
taken.  Normally, we rely on the "exact" trip count, which was
correctly given as "unknown". However, SCEV also gives a "max"
back-edge taken count. The loops max BE taken count is conservatively
a maximum over the max of each exit's non-exiting iterations
count. Note that some exit tests can be skipped so the max loop
back-edge taken count can actually exceed the max non-exiting
iterations for some exits. However, when we know the loop *latch*
cannot be skipped, we can directly use its max taken count
disregarding other exits. I previously took the minimum here without
checking whether the other exit could be skipped. The correct, and
simpler thing to do here is just to directly use the loop latch's max
non-exiting iterations as the loops max back-edge count.

In the problematic test case, the first loop exit had a max of zero
non-exiting iterations, but could be skipped. The loop latch was known
not to be skipped but had max of one non-exiting iteration. We
incorrectly claimed the loop back-edge could be taken zero times, when
it is actually taken one time.

Fixes Loop %for.body.i: <multiple exits> Unpredictable backedge-taken count.
Loop %for.body.i: max backedge-taken count is 1.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@209358 91177308-0d34-0410-b5e6-96231b3b80d8
2014-05-22 00:37:03 +00:00
Eli Bendersky
107db21c2b Similar to bitcast, treat addrspacecast as a foldable operand.
Added a test sink-addrspacecast.ll to verify this change.

Patch by Jingyue Wu.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@209343 91177308-0d34-0410-b5e6-96231b3b80d8
2014-05-22 00:02:52 +00:00
Eric Christopher
a2f371b8eb Fix compilation issues.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@209342 91177308-0d34-0410-b5e6-96231b3b80d8
2014-05-21 23:51:57 +00:00
Eric Christopher
189fe78e2f Make early if conversion dependent upon the subtarget and add
a subtarget hook to enable. Unconditionally add to the pass pipeline
for targets that might want to use it. No functional change.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@209340 91177308-0d34-0410-b5e6-96231b3b80d8
2014-05-21 23:40:26 +00:00
Eric Christopher
595bdb7e8b Group the scheduling functions together.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@209339 91177308-0d34-0410-b5e6-96231b3b80d8
2014-05-21 23:40:18 +00:00
David Blaikie
8a0240ddf2 Revert "DebugInfo: Don't put fission type units in comdat sections."
This reverts commit r208930, r208933, and r208975.

It seems not all fission consumers are ready to handle this behavior.
Reverting until tools are brought up to spec.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@209338 91177308-0d34-0410-b5e6-96231b3b80d8
2014-05-21 23:27:41 +00:00
Saleem Abdulrasool
cba7ac7bda MC: correct IMAGE_REL_ARM_MOV32T relocation emission
This corrects the emission of IMAGE_REL_ARM_MOV32T relocations.  Previously, we
were avoiding the high portion of the relocation too early.  If there was a
section-relative relocation with an offset greater than 16-bits (65535), you
would end up truncating the high order bits of the offset.  Allow the current
relocation representation to flow through out the MC layer to the object writer.
Use the new ability to restrict recorded relocations to avoid emitting the
relocation into the final object.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@209337 91177308-0d34-0410-b5e6-96231b3b80d8
2014-05-21 23:17:56 +00:00
Saleem Abdulrasool
159ccc8f51 MC: introduce ability to restrict recorded relocations
Add support to allow a target specific COFF object writer to restrict the
recorded resolutions in the emitted object files.  This is motivated by the need
in Windows on ARM, where an intermediate relocation needs to be prevented from
being emitted in the object file.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@209336 91177308-0d34-0410-b5e6-96231b3b80d8
2014-05-21 23:17:50 +00:00
David Blaikie
111bad385a DebugInfo: Use the SPMap to find the parent CU of inlined functions as they may not be in the current CU
Committed in r209178 then reverted in r209251 due to LTO breakage,
here's a proper fix for the case of the missing subprogram DIE. The DIEs
were there, just in other compile units. Using the SPMap we can find the
right compile unit to search for and produce cross-unit references to
describe this kind of inlining.

One existing test case needed to be updated because it had a function
that wasn't in the CU's subprogram list, so it didn't appear in the
SPMap.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@209335 91177308-0d34-0410-b5e6-96231b3b80d8
2014-05-21 23:14:12 +00:00
Matt Arsenault
afd1747bbc R600: Add comment describing problems with LowerConstantInitializer
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@209333 91177308-0d34-0410-b5e6-96231b3b80d8
2014-05-21 22:59:17 +00:00
Matt Arsenault
bd124c85ef R600: Partially fix constant initializers for structs and vectors.
This should extend the current workaround to work with structs
that only contain legal, scalar types.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@209331 91177308-0d34-0410-b5e6-96231b3b80d8
2014-05-21 22:42:42 +00:00
Matt Arsenault
713096fb68 R600: Add failing testcases for constant initializers.
Constant initializers involving illegal types hit an assertion.

Patch by: Jan Vesely <jan.vesely@rutgers.edu>

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@209330 91177308-0d34-0410-b5e6-96231b3b80d8
2014-05-21 22:42:38 +00:00
Eric Christopher
5fe59a2e10 Remove getTargetLowering from TargetPassConfig as the target lowering
can change depending upon subtarget/subtarget features for a function.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@209329 91177308-0d34-0410-b5e6-96231b3b80d8
2014-05-21 22:42:07 +00:00
Eric Christopher
9bfc3f592d Remove unused member variable from hexagon pass.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@209328 91177308-0d34-0410-b5e6-96231b3b80d8
2014-05-21 22:42:02 +00:00
David Blaikie
30436451e2 DebugInfo: Ensure concrete out of line variables from inlined functions reference their abstract origins.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@209327 91177308-0d34-0410-b5e6-96231b3b80d8
2014-05-21 22:41:17 +00:00
Quentin Colombet
fd0096a42c [X86] Fix a bug in the lowering of BLENDI introduced in r209043.
ISD::VSELECT mask uses 1 to identify the first argument and 0 to identify the
second argument.
On the other hand, BLENDI uses 0 to identify the first argument and 1 to
identify the second argument.
Fix the generation of the blend mask to account for this difference.

The bug did not show up with r209043, because we were not checking for the
actual arguments of the blend instruction!
This commit also fixes the test cases.

Note: The same mask works for the BLENDr variant because the arguments are
swapped during instruction selection (see the BLENDXXrr patterns).

<rdar://problem/16975435>


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@209324 91177308-0d34-0410-b5e6-96231b3b80d8
2014-05-21 22:00:39 +00:00
Eric Christopher
d1b5bdaebd Move MCOptions that aren't shared between programs into their specific
program and have them initialize the MCOptions struct explicitly.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@209321 91177308-0d34-0410-b5e6-96231b3b80d8
2014-05-21 21:05:09 +00:00
Eric Christopher
9a37986328 Make a couple of command lines static and remove an unnecessary
initialization.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@209320 91177308-0d34-0410-b5e6-96231b3b80d8
2014-05-21 21:05:05 +00:00
David Blaikie
c10a1edf3f DebugInfo: Simplify subprogram declaration creation/references and accidentally refix PR11300.
Also simplifies the linkage name handling a little too.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@209311 91177308-0d34-0410-b5e6-96231b3b80d8
2014-05-21 18:04:33 +00:00
Matt Arsenault
49323774b4 Use cast<> instead of unchecked dyn_cast
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@209310 91177308-0d34-0410-b5e6-96231b3b80d8
2014-05-21 18:03:59 +00:00
Saleem Abdulrasool
5dd4b9f43b MC: loosen an overzealous assertion
Permit active macro expansions when terminating the assembler if there were
errors during the expansion.  This would only trigger on invalid input when
built with assertions.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@209309 91177308-0d34-0410-b5e6-96231b3b80d8
2014-05-21 17:53:18 +00:00
Greg Fitzgerald
03b003b1cb Use llvm-lit if LLVM source tree is unavailable.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@209308 91177308-0d34-0410-b5e6-96231b3b80d8
2014-05-21 16:44:03 +00:00
Dave Estes
c12c3d0753 Test comment commit.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@209306 91177308-0d34-0410-b5e6-96231b3b80d8
2014-05-21 16:19:51 +00:00
Daniel Sanders
820861536c [mips][mips64r6] Add bc[12](eq|ne)z
Summary: Depends on D3691

Reviewers: jkolek, zoran.jovanovic, vmedic

Reviewed By: vmedic

Differential Revision: http://reviews.llvm.org/D3760

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@209292 91177308-0d34-0410-b5e6-96231b3b80d8
2014-05-21 12:56:39 +00:00
Daniel Sanders
3046dcbce4 [asm matcher] Fix incorrect assertion when there are exactly 32 SubtargetFeatures
Summary:
The minimal type needs to hold a value of '1ULL << 31' but
getMinimalTypeForRange() is called with a value of '1ULL << 32'.

This patch will also reduce the size of the matcher table when there are 8
or 16 SubtargetFeatures.

Also added a dump of the SubtargetFeatures to the -debug output and corrected getMinimalTypeInRange() to consider 0xffffffffull to be a 32-bit value.

The testcase is that no existing code is broken and that LLVM still successfully
compiles after adding MIPS64r6 CodeGen support.

Reviewers: rafael

Reviewed By: rafael

Subscribers: llvm-commits

Differential Revision: http://reviews.llvm.org/D3787

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@209288 91177308-0d34-0410-b5e6-96231b3b80d8
2014-05-21 10:11:24 +00:00
Evgeniy Stepanov
c0e669384d [asan] Fix x86-32 asm instrumentation to preserve flags.
Patch by Yuri Gorshenin.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@209280 91177308-0d34-0410-b5e6-96231b3b80d8
2014-05-21 08:14:24 +00:00
Saleem Abdulrasool
8235ed1808 MC: mark COFF .drectve section as REMOVE
The .drectve section should be marked as IMAGE_SCN_LNK_REMOVE.  This matches what
the MSVC toolchain does and accurately reflects that this section should not be
emitted into the final binary.  This section is merely information for the
linker, comprising of additional linker directives.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@209273 91177308-0d34-0410-b5e6-96231b3b80d8
2014-05-21 05:15:01 +00:00
Richard Smith
651f680b79 [modules] Add module maps for LLVM. These are not quite ready for prime-time
yet, but only a few more Clang patches need to land. (I have 'ninja check'
passing locally.)


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@209269 91177308-0d34-0410-b5e6-96231b3b80d8
2014-05-21 02:46:14 +00:00
Saleem Abdulrasool
91bbc253bd ARM: correct bundle generation for MOV32T relocations
Although the previous code would construct a bundle and add the correct elements
to it, it would not finalise the bundle.  This resulted in the InternalRead
markers not being added to the MachineOperands nor, more importantly, the
externally visible defs to the bundle itself.  So, although the bundle was not
exposing the def, the generated code would be correct because there was no
optimisations being performed.  When optimisations were enabled, the post
register allocator would kick in, and the hazard recognizer would reorder
operations around the load which would define the value being operated upon.

Rather than manually constructing the bundle, simply construct and finalise the
bundle via the finaliseBundle call after both MIs have been emitted.  This
improves the code generation with optimisations where IMAGE_REL_ARM_MOV32T
relocations are emitted.

The changes to the other tests are the result of the bundle generation
preventing the scheduler from hoisting the moves across the loads.  The net
effect of the generated code is equivalent, but, is much more identical to what
is actually being lowered.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@209267 91177308-0d34-0410-b5e6-96231b3b80d8
2014-05-21 01:25:24 +00:00
Eric Christopher
e648d75c60 This command line option is only used in one place. Move it there and
rename it to something more descriptive.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@209263 91177308-0d34-0410-b5e6-96231b3b80d8
2014-05-21 00:20:01 +00:00
Eric Christopher
110260377a Add a comment here.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@209262 91177308-0d34-0410-b5e6-96231b3b80d8
2014-05-21 00:02:50 +00:00
Eric Christopher
3d5f46bb8a Move this test to the backend from the frontend.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@209259 91177308-0d34-0410-b5e6-96231b3b80d8
2014-05-20 23:59:54 +00:00
Eric Christopher
4f6d26dbe8 Move the verbose asm option to be part of the options struct and
set appropriately.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@209258 91177308-0d34-0410-b5e6-96231b3b80d8
2014-05-20 23:59:50 +00:00
Kevin Enderby
b38059ab25 Update MachOObjectFile::getSymbolAddress so it returns UnknownAddressOrSize
for undefined symbols, so it matches what COFFObjectFile::getSymbolAddress
does.  This allows llvm-nm to print spaces instead of 0’s for the value
of undefined symbols in Mach-O files.

To make this change other uses of MachOObjectFile::getSymbolAddress
are updated to handle when the Value is returned as UnknownAddressOrSize.
Which is needed to keep two of the ExecutionEngine tests working for example.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@209253 91177308-0d34-0410-b5e6-96231b3b80d8
2014-05-20 23:04:47 +00:00
Alexey Samsonov
acf621545d Fix test added in r209242: llc shouldn't create files in source tree
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@209252 91177308-0d34-0410-b5e6-96231b3b80d8
2014-05-20 22:40:31 +00:00
David Blaikie
ed3abe4994 Revert "DebugInfo: Assume all subprogram DIEs have been created before any abstract subprograms are constructed."
This reverts commit r209178.

This seems to be asserting in an LTO build on some internal Apple
buildbots. No upstream reproduction (and I don't have an LLVM-aware gold
built right now to reproduce it personally) but it's a small patch & the
failure's semi-plausible so I'm going to revert first while I try to
reproduce this.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@209251 91177308-0d34-0410-b5e6-96231b3b80d8
2014-05-20 22:33:09 +00:00
Adam Nemet
adf1668bec [ARM64] PR19792: Fix cycle in DAG after performPostLD1Combine
Povray and dealII currently assert with "Overran sorted position" in
AssignTopologicalOrder.  The problem is that performPostLD1Combine can
introduce cycles.

Consider:

(insert_vector_elt (INSERT_SUBREG undef,
                                  (load (add %vreg0, Constant<8>), undef),  <= A
                                  TargetConstant<2>),
                   (load %vreg0, undef),                                    <= B
                   Constant<1>)

This is turned into a LD1LANEpost node.  However the address in A is not a
valid user of the post-incremented address of B in LD1LANEpost.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@209242 91177308-0d34-0410-b5e6-96231b3b80d8
2014-05-20 21:47:07 +00:00
David Blaikie
e94103adcd Unbreak the sanitizer buildbots after r209226 due to SROA issue described in http://reviews.llvm.org/D3714
Undecided whether this should include a test case - SROA produces bad
dbg.value metadata describing a value for a reference that is actually
the value of the thing the reference refers to. For now, loosening the
assert lets this not assert, but it's still bogus/wrong output...

If someone wants to tell me to add a test, I'm willing/able, just
undecided. Hopefully we'll get SROA fixed soon & we can tighten up this
assertion again.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@209240 91177308-0d34-0410-b5e6-96231b3b80d8
2014-05-20 21:40:13 +00:00
Eric Christopher
6a9366c0c6 Move the function and data section flags into the options struct and
make the functions to set them non-static.
Move and rename the llvm specific backend options to avoid conflicting
with the clang option.

Paired with a backend commit to update.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@209238 91177308-0d34-0410-b5e6-96231b3b80d8
2014-05-20 21:25:34 +00:00
Kevin Enderby
95aa960b71 Revert r209235 as it broke two tests:
Failing Tests (2):
	    LLVM :: ExecutionEngine/MCJIT/stubs-sm-pic.ll
	    LLVM :: ExecutionEngine/MCJIT/stubs.ll


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@209236 91177308-0d34-0410-b5e6-96231b3b80d8
2014-05-20 21:10:15 +00:00
Kevin Enderby
e3f74564ff Update MachOObjectFile::getSymbolAddress so it returns UnknownAddressOrSize
for undefined symbols.  Allowing llvm-nm to print spaces instead of 0’s for
the value of undefined symbols in Mach-O files.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@209235 91177308-0d34-0410-b5e6-96231b3b80d8
2014-05-20 20:32:18 +00:00
Quentin Colombet
50d4008b47 [LSR] Canonicalize reg1 + ... + regN into reg1 + ... + 1*regN.
This commit introduces a canonical representation for the formulae.
Basically, as soon as a formula has more that one base register, the scaled
register field is used for one of them. The register put into the scaled
register is preferably a loop variant.
The commit refactors how the formulae are built in order to produce such
representation.
This yields a more accurate, but still perfectible, cost model.

<rdar://problem/16731508>


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@209230 91177308-0d34-0410-b5e6-96231b3b80d8
2014-05-20 19:25:04 +00:00
Duncan P. N. Exon Smith
8598e290c3 GlobalValue: Automatically reset visibility when setting local linkage
r208264 started asserting in `setLinkage()` and `setVisibility()` that
visibility and linkage are compatible.  There are a few places in clang
where visibility is set first, and then linkage later, so the assert
fires.  In `setLinkage()`, it's clear what the visibility *should* be,
so rather than updating all the call sites just automatically fix the
visibility.

The testcase for this is for *clang*, so it'll follow separately in cfe.

PR19760

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@209227 91177308-0d34-0410-b5e6-96231b3b80d8
2014-05-20 19:00:58 +00:00
David Blaikie
1e3493b04a Fix test breakage introduced in r209223.
Oops, broke the broken enum constants again.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@209226 91177308-0d34-0410-b5e6-96231b3b80d8
2014-05-20 18:36:35 +00:00
Alexey Samsonov
447a7ce76c Rewrite calculateDbgValueHistory to make it (hopefully) more transparent.
This change preserves the original algorithm of generating history
for user variables, but makes it more clear.

High-level description of algorithm:
Scan all the machine basic blocks and machine instructions in the order
they are emitted to the object file. Do the following:
1) If we see a DBG_VALUE instruction, add it to the history of the
corresponding user variable. Keep track of all user variables, whose
locations are described by a register.
2) If we see a regular instruction, look at all the registers it clobbers,
and terminate the location range for all variables described by these registers.
3) At the end of the basic block, terminate location ranges for all
user variables described by some register.

Although this change shouldn't be user-visible (the contents of .debug_loc section
should be the same), it changes some internal assumptions about the set
of instructions used to track the variable locations. Watching the bots.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@209225 91177308-0d34-0410-b5e6-96231b3b80d8
2014-05-20 18:34:54 +00:00
David Blaikie
e2ab467b90 PR19767: DebugInfo emission of pointer constants.
In refactoring DwarfUnit::isUnsignedDIType I restricted it to only work
on values with signedness (unsigned or signed), asserting on anything
else (which did uncover some bugs). But it turns out that we do need to
emit constants of signless data, such as pointer constants - only null
pointer constants are known to need this so far, but it's conceivable
that there might be non-null pointer constants at some point (hardcoded
address offsets for device drivers?).

This patch just uses 'unsigned' for signless data such as pointer
constants. Arguably we could use signless representations
(DW_FORM_dataN) instead, allowing a trinary result from isUnsignedDIType
(signed, unsigned, signless), but this seems reasonable for now.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@209223 91177308-0d34-0410-b5e6-96231b3b80d8
2014-05-20 18:21:51 +00:00
Renato Golin
5d7afdb2ec Avoids DCE on write_register
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@209222 91177308-0d34-0410-b5e6-96231b3b80d8
2014-05-20 17:40:03 +00:00