Commit Graph

138478 Commits

Author SHA1 Message Date
Sam Kolton
4a0cb73585 [AMDGPU] Disassembler: print label names in branch instructions
Summary: Add AMDGPUSymbolizer for finding names for labels from ELF symbol table.

Reviewers: vpykhtin, artem.tamazov, tstellarAMD

Subscribers: arsenm, kzhuravl, wdng, nhaehnle, yaxunl, tony-tye

Differential Revision: https://reviews.llvm.org/D24802

llvm-svn: 282394
2016-09-26 10:05:50 +00:00
James Molloy
462f650cad [ARM] Promote small global constants to constant pools
If a constant is unamed_addr and is only used within one function, we can save
on the code size and runtime cost of an indirection by changing the global's storage
to inside the constant pool. For example, instead of:

      ldr r0, .CPI0
      bl printf
      bx lr
    .CPI0: &format_string
    format_string: .asciz "hello, world!\n"

We can emit:

      adr r0, .CPI0
      bl printf
      bx lr
    .CPI0: .asciz "hello, world!\n"

This can cause significant code size savings when many small strings are used in one
function (4 bytes per string).

This recommit contains fixes for a nasty bug related to fast-isel fallback - because
fast-isel doesn't know about this optimization, if it runs and emits references to
a string that we inline (because fast-isel fell back to SDAG) we will end up
with an inlined string and also an out-of-line string, and we won't emit the
out-of-line string, causing backend failures.

It also contains fixes for emitting .text relocations which made the sanitizer
bots unhappy.

llvm-svn: 282387
2016-09-26 07:26:24 +00:00
Zvi Rackover
9ca8e1a54b [X86] Optimization for replacing LEA with MOV at frame index elimination time
Summary:
Replace a LEA instruction of the form 'lea (%esp), %ebx' --> 'mov %esp, %ebx'

MOV is preferable over LEA because usually there are more issue-slots available to execute MOVs than LEAs. Latest processors also support zero-latency MOVs.

Fixes pr29022.

Reviewers: hfinkel, delena, igorb, myatsina, mkuper

Differential Revision: https://reviews.llvm.org/D24705

llvm-svn: 282385
2016-09-26 06:42:07 +00:00
Chandler Carruth
8fcc7016c3 [PM] Refactor this unittest a bit to remove duplicated code. This was
suggested at one point during code review and I deferred it to
a follow-up commit.

llvm-svn: 282383
2016-09-26 06:29:21 +00:00
Ayman Musa
a9677386d1 [X86][avx512] Fix bug in masked compress store.
Differential Revision: https://reviews.llvm.org/D23984

llvm-svn: 282381
2016-09-26 06:22:08 +00:00
Chandler Carruth
59981672f6 [SCEV] Fix the order of members in the initializer list.
Noticed due to the warning on this line. Sanjoy is on
a less-than-awesome internet connection, so committing on his behalf.

llvm-svn: 282380
2016-09-26 04:49:58 +00:00
Chandler Carruth
7eee8eb33a [PM] Add a unittest covering the invalidation of a Module analysis from
a function pass nested inside of a CGSCC pass manager.

This is very similar to the previous unittest but makes sure the
invalidation logic works across all the layers here.

llvm-svn: 282378
2016-09-26 04:17:12 +00:00
Chandler Carruth
e62ef9f347 [PM] Add a unittest for invalidating module analyses with an SCC pass.
This reinstates r280447. Original commit log:
This wasn't really well explicitly tested with a nice unittest before.
It seems good to have reasonably broken out unittests for this kind of
functionality as I'm workin go other invalidation features to make sure
none of the existing ones regress.

This still has too much duplicated code, I plan to factor that out in
a subsequent commit to use common helpers for repeated parts of this.

llvm-svn: 282377
2016-09-26 04:01:55 +00:00
Sanjoy Das
ef4a695251 [SCEV] Assign LoopPropertiesCache in the move constructor
In a previous change I collapsed two different caches into one.  When
doing that I noticed that ScalarEvolution's move constructor was not
moving those caches.

To keep the previous change simple, I've moved that bugfix into this
separate change.

llvm-svn: 282376
2016-09-26 02:44:10 +00:00
Sanjoy Das
b9f8e32d12 [SCEV] Combine two predicates into one; NFC
Both `loopHasNoSideEffects` and `loopHasNoAbnormalExits` involve walking
the loop and maintaining similar sorts of caches.  This commit changes
SCEV to compute both the predicates via a single walk, and maintain a
single cache instead of two.

llvm-svn: 282375
2016-09-26 02:44:07 +00:00
Sanjoy Das
4dc0994173 [SCEV] Make it obvious BackedgeTakenInfo's constructor steals storage
Specifically, it moves SCEVUnionPredicates from its input into its own
storage.  Make this obvious at the type level.

llvm-svn: 282374
2016-09-26 01:10:27 +00:00
Sanjoy Das
6926238887 [SCEV] Further isolate incidental data structure; NFC
llvm-svn: 282373
2016-09-26 01:10:25 +00:00
Sanjoy Das
2025c6e5ef [SCEV] Simplify BackedgeTakenInfo::getMax; NFC
llvm-svn: 282372
2016-09-26 01:10:22 +00:00
Sanjoy Das
16ec3916da Appease MSVC
... by not default move constructors and operator= s. Defaulting these
works in clang, but not in MSVC.

llvm-svn: 282370
2016-09-26 00:22:18 +00:00
Sanjoy Das
db06a82fe8 Attempt to appease MSVC
... by explicitly deleting the copy constructor.

llvm-svn: 282369
2016-09-26 00:00:51 +00:00
Sanjoy Das
761e5aabb2 [SCEV] Reserve space in SmallVector; NFC
llvm-svn: 282368
2016-09-25 23:12:08 +00:00
Sanjoy Das
ea08482f83 [SCEV] Document a gotcha; NFC
We should re-consider the design decision that led to this gotcah, but
for now just document it.

llvm-svn: 282367
2016-09-25 23:12:06 +00:00
Sanjoy Das
23b6c4c2a8 [SCEV] Have ExitNotTakenInfo keep a pointer to its predicate; NFC
SCEVUnionPredicate is a "heavyweight" structure, so it is beneficial to
store the (optional) data out of line.

llvm-svn: 282366
2016-09-25 23:12:04 +00:00
Sanjoy Das
6bbb5f521c [SCEV] Simplify tracking ExitNotTakenInfo instances; NFC
This change simplifies a data structure optimization in the
`BackedgeTakenInfo` class for loops with exactly one computable exit.

I've sanity checked that this does not regress compile time performance,
using sqlite3's amalgamated build.

llvm-svn: 282365
2016-09-25 23:12:00 +00:00
Sanjoy Das
3b4c0c8086 [SCEV] Rename a couple of fields; NFC
llvm-svn: 282364
2016-09-25 23:11:57 +00:00
Sanjoy Das
e39d2ae7eb [SCEV] Remove incidental data structure; NFC
llvm-svn: 282363
2016-09-25 23:11:55 +00:00
Sanjoy Das
1133383a22 [SCEV] Clang format most of the SCEV header; NFC
The indentation for the declared classes was not as per LLVM coding
style.

llvm-svn: 282362
2016-09-25 23:11:51 +00:00
Craig Topper
f91537d654 [X86] Remove what appears to be leftover MMX code involving (v1i64 scalar_to_vector).
llvm-svn: 282361
2016-09-25 16:34:11 +00:00
Craig Topper
fe64841663 [X86] Remove patterns for scalar_to_vector from FR32/FR64 to 256-bit vectors. Lowering explicitly avoids creating this pattern.
llvm-svn: 282360
2016-09-25 16:34:09 +00:00
Craig Topper
7bc5d46631 [AVX-512] Replace get512BitSuperRegister with calls to TargetRegisterInfo::getMatchingSuperReg.
llvm-svn: 282359
2016-09-25 16:34:06 +00:00
Craig Topper
c35f693a30 [AVX-512] Fix some patterns predicates to properly enforce priority for various versions of CVTDQ2PD instruction.
llvm-svn: 282358
2016-09-25 16:34:02 +00:00
Craig Topper
8e5bd2f3f8 [AVX-512] Add rounding versions of instructions to hasUndefRegUpdate.
llvm-svn: 282357
2016-09-25 16:33:59 +00:00
Craig Topper
9977467fdd [AVX-512] Add the scalar unsigned integer to fp conversion instructions to hasUndefRegUpdate.
llvm-svn: 282356
2016-09-25 16:33:57 +00:00
Craig Topper
1f56afd8eb [AVX-512] Remove duplicate instructions for converting integer to scalar floating point. We can use patterns to point to the other instructions instead.
llvm-svn: 282355
2016-09-25 16:33:53 +00:00
Zachary Turner
bfc94bc7f8 Add a comment on StringRef::contains(char)
llvm-svn: 282350
2016-09-25 04:06:39 +00:00
Zachary Turner
e1a79703aa Fix signed / unsigned comparison.
llvm-svn: 282348
2016-09-25 03:57:34 +00:00
Zachary Turner
3c222df487 Add some predicated searching functions to StringRef.
This adds 4 new functions to StringRef, which can be used to
take or drop characters while a certain condition is met, or
until a certain condition is met.  They are:

take_while - Return characters until a condition is not met.
take_until - Return characters until a condition is met.
drop_while - Remove characters until a condition is not met.
drop_until - Remove characters until a condition is met.

Internally, all of these functions delegate to two additional
helper functions which can be used to search for the position
of a character meeting or not meeting a condition, which are:

find_if - Find the first character matching a predicate.
find_if_not - Find the first character not matching a predicate.

Differential Revision: https://reviews.llvm.org/D24842

llvm-svn: 282346
2016-09-25 03:27:29 +00:00
Craig Topper
0ccdd06243 [AVX-512] Don't use two opcodes for INTR_TYPE_SCALAR_MASK_RM. The handling was such that if the second opcode was present the first was ingored, so we can just have one opcode.
llvm-svn: 282344
2016-09-25 01:03:10 +00:00
Craig Topper
bab9b76656 [X86] Teach combineShuffle to avoid creating floating point operations with integer types and integer operations with floating point types. Seems isOperationLegal lies for mismatched types and operations.
Fixes PR30511.

llvm-svn: 282341
2016-09-24 21:42:49 +00:00
Craig Topper
306ed8539b [AVX-512] Split scalar version of X86ISD::SELECT into a separate opcode because isel is not robust with multiple type profiles for the same opcode.
llvm-svn: 282340
2016-09-24 21:42:47 +00:00
Craig Topper
eddbef5520 [AVX-512] Remove the patterns for selecting scalar VCOMI/VUCOMI instructions with SAE as there is no way to create the pattern.
llvm-svn: 282339
2016-09-24 21:42:43 +00:00
Duncan P. N. Exon Smith
315172f4ca ObjCARC: Don't look at users of ConstantData
Stop looking at users of UndefValue and ConstantPointerNull in the
objective C ARC optimizers.  The other users aren't actually
interesting, since they're not pointing at a particular object.  I
imagine these calls could be optimized through -instcombine... maybe
they already are?

These early returns will be required at some point in the future, with a
WIP patch that asserts when someone accesses a use-list on ConstantData.

llvm-svn: 282338
2016-09-24 21:01:20 +00:00
Duncan P. N. Exon Smith
2f789947a5 Analysis: Return early for UndefValue in computeKnownBits
There is no benefit in looking through assumptions on UndefValue to
guess known bits.  Return early to avoid walking their use-lists, and
assert that all instances of ConstantData are handled here for similar
reasons (UndefValue was the only integer/pointer holdout).

llvm-svn: 282337
2016-09-24 20:42:02 +00:00
Sanjay Patel
da8aa7cf2e [x86] don't try to create a vector integer inst for an SSE1 target (PR30512)
This bug was introduced with:
http://reviews.llvm.org/rL272511

We need to restrict the lowering to v4f32 comparisons because that's all SSE1 can handle.

This should fix:
https://llvm.org/bugs/show_bug.cgi?id=28044

llvm-svn: 282336
2016-09-24 20:24:06 +00:00
Duncan P. N. Exon Smith
1463b59de9 Scalar: Ignore ConstantData in processAssumption
Assumptions on UndefValue and ConstantPointerNull aren't relevant to
other users.  Ignore them entirely to avoid wasting cycles walking
through their (possibly extremely extensive (cross-module)) use-lists.

It wasn't clear how to add a specific test for this, and it'll be
covered anyway by an eventual patch that asserts when trying to access
the use-list of an instance of ConstantData.

llvm-svn: 282334
2016-09-24 20:00:38 +00:00
Duncan P. N. Exon Smith
b94ad7ec1b Analysis: Return early in isKnownNonNullAt for ConstantData
Check and return early for ConstantPointerNull and UndefValue
specifically in isKnownNonNullAt, and assert that ConstantData never
make it to isKnownNonNullFromDominatingCondition.

This confirms that isKnownNonNullFromDominatingCondition never walks
through the use-list of an instance of ConstantData.  Given that such
use-lists cross module boundaries, it never really made sense to do so,
and was potentially very expensive.

llvm-svn: 282333
2016-09-24 19:39:47 +00:00
Dylan McKay
87118e6a0f [AVR] Update signature of AVRTargetObjectFile::SelectSectionForGlobal
It was changed recently, and was breaking compilation of the backend.

llvm-svn: 282329
2016-09-24 11:38:08 +00:00
Quentin Colombet
3d5fa3f4d2 [RegisterBankInfo] Constify the member of the XXXMapping maps.
This makes it obvious that items in those maps behave like statically
created objects.

llvm-svn: 282327
2016-09-24 04:54:03 +00:00
Quentin Colombet
c04a3667f5 [RegisterBankInfo] Document the design choice for the BreakDown.
This adds a comment explaining why we will duplicate PartialMapping to
represent the breakdown for complex mappings (mappings with more than
one partial mapping), instead of using an array of pointer.

NFC

llvm-svn: 282326
2016-09-24 04:53:58 +00:00
Quentin Colombet
3292cd26b1 [RegisterBankInfo] Add statistics for dynamic value mappings.
Like partial mappings, as we move toward TableGen'ed information, the
number should reach zero eventually.

llvm-svn: 282325
2016-09-24 04:53:55 +00:00
Quentin Colombet
3708fb9829 [RegisterBankInfo] Uniquely generate ValueMapping.
This is a step toward statically allocate ValueMapping. Like the
previous few commits, the goal is to move toward a TableGen'ed like
structure with no dynamic allocation at all.

llvm-svn: 282324
2016-09-24 04:53:52 +00:00
Quentin Colombet
ffe7332f86 [RegisterBankInfo] Keep valid pointers for PartialMappings.
Previously we were using the address of the unique instance of a partial
mapping in the related map to access this instance. However, when the
map grows, the whole set of instances may be moved elsewhere and the
previous addresses are not valid anymore.

Instead, keep the address of the unique heap allocated instance of a
partial mapping.

Note: I did not see any actual bugs for that problem as the number of
partial mappings dynamically allocated is small (<= 4).

llvm-svn: 282323
2016-09-24 04:53:48 +00:00
Kostya Serebryany
810d04f0e6 [libFuzzer] add a standalone build script
llvm-svn: 282321
2016-09-24 04:00:00 +00:00
Duncan P. N. Exon Smith
f0fc8836d4 GlobalStatus: Don't walk use-lists of ConstantData
Return early from llvm::isSafeToDestroyConstant() whenever the value
`isa<ConstantData>()`.  These constants are shared across the
LLVMContext.  We never really want to delete them here, and walking
their use-lists can be very expensive.

(This is motivated by an eventual goal of removing use-lists entirely
from ConstantData.)

llvm-svn: 282320
2016-09-24 02:30:11 +00:00
Kostya Serebryany
496fab274a [libFuzzer] simplify HandleTrace again, start re-running interesting units and collecting their features.
llvm-svn: 282316
2016-09-23 23:51:58 +00:00