Commit Graph

118910 Commits

Author SHA1 Message Date
Tom Stellard
9e7f0c8e77 R600/SI: Use ELF64 format instead of ELF32
Reviewers: arsenm, rafael

Subscribers: llvm-commits

Differential Revision: http://reviews.llvm.org/D10392

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@240331 91177308-0d34-0410-b5e6-96231b3b80d8
2015-06-22 21:03:54 +00:00
Tom Stellard
309f60c15a R600: Use EM_AMDGPU for the ELF Machine type
Reviewers: arsenm, rafael

Subscribers: llvm-commits

Differential Revision: http://reviews.llvm.org/D10390

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@240330 91177308-0d34-0410-b5e6-96231b3b80d8
2015-06-22 21:03:52 +00:00
Tom Stellard
b53953a84f Support/ELF: Add EM_AMDGPU
Summary: This will be used by the R600 backend.

Reviewers: chandlerc, rafael

Subscribers: llvm-commits

Differential Revision: http://reviews.llvm.org/D10389

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@240329 91177308-0d34-0410-b5e6-96231b3b80d8
2015-06-22 21:03:47 +00:00
Ahmed Bougacha
a3afb70a5d [X86] Teach load folding to accept scalar _Int users of MOVSS/MOVSD.
The _Int instructions are special, in that they operate on the full
VR128 instead of FR32.  The load folding then looks at MOVSS, at the
user, and bails out when it sees a size mismatch.

What we really know is that the rm_Int instructions don't load the
higher lanes, so folding is fine.

This happens for the straightforward intrinsic code, e.g.:

    _mm_add_ss(a, _mm_load_ss(p));

Fixes PR23349.

Differential Revision: http://reviews.llvm.org/D10554


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@240326 91177308-0d34-0410-b5e6-96231b3b80d8
2015-06-22 20:51:51 +00:00
Alex Lorenz
f654cdae59 MIR Serialization: Introduce a lexer for machine instructions.
This commit adds a function that tokenizes the string containing
the machine instruction. This commit also adds a struct called 
'MIToken' which is used to represent the lexer's tokens.

Reviewers: Sean Silva

Differential Revision: http://reviews.llvm.org/D10521


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@240323 91177308-0d34-0410-b5e6-96231b3b80d8
2015-06-22 20:37:46 +00:00
Peter Collingbourne
47e71da4b3 SafeStack: Create the unsafe stack pointer on demand.
This avoids creating an unnecessary undefined reference on targets such as
NVPTX that require such references to be declared in asm output.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@240321 91177308-0d34-0410-b5e6-96231b3b80d8
2015-06-22 20:26:54 +00:00
Pete Cooper
8b4b91acd8 Move MCSymbol Value in to the union of Offset and CommonSize.
This is a reapplication of r239440 which was reverted in r239441.
There are no changes to this patch from then, but this had instead exposed
a bug in .thumb_set which was fixed in r240318.  Having fixed that bug, it
is now safe to re-apply this code.

Original commit message below:

It wasn't possible to have a variable Symbol with offset or 'isCommon' so
this just enables better packing of the MCSymbol class.

Reviewed by Rafael Espindola.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@240320 91177308-0d34-0410-b5e6-96231b3b80d8
2015-06-22 19:57:33 +00:00
Pete Cooper
54d0bc35fe Change .thumb_set to have the same error checks as .set.
According to the documentation, .thumb_set is 'the equivalent of a .set directive'.

We didn't have equivalent behaviour in terms of all the errors we could throw, for
example, when a symbol is redefined.

This change refactors parseAssignment so that it can be used by .set and .thumb_set
and implements tests for .thumb_set for all the errors thrown by that method.

Reviewed by Rafael Espíndola.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@240318 91177308-0d34-0410-b5e6-96231b3b80d8
2015-06-22 19:35:57 +00:00
Krzysztof Parzyszek
18693ad79a Remove "const" from the MachineFunction reference in VLIWPacketizerList
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@240317 91177308-0d34-0410-b5e6-96231b3b80d8
2015-06-22 18:59:44 +00:00
Justin Bogner
0d1cc5aa82 docs: Update allowed values for LLVM_USE_SANITIZER
"Thread" and combinations of "Address" and "Undefined" have been
accepted for a while now.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@240316 91177308-0d34-0410-b5e6-96231b3b80d8
2015-06-22 18:55:46 +00:00
Sanjay Patel
73aa02eb09 [x86] set default reciprocal (division and square root) codegen to match GCC
D8982 ( checked in at http://reviews.llvm.org/rL239001 ) added command-line 
options to allow reciprocal estimate instructions to be used in place of
divisions and square roots.

This patch changes the default settings for x86 targets to allow that recip
codegen (except for scalar division because that breaks too much code) when
using -ffast-math or its equivalent. 

This matches GCC behavior for this kind of codegen.

Differential Revision: http://reviews.llvm.org/D10396



git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@240310 91177308-0d34-0410-b5e6-96231b3b80d8
2015-06-22 18:29:44 +00:00
Sanjoy Das
060e38ab9f Fix MSVC build (again!).
Remove two `typename`s that I should have removed in r240307 but left in
by mistake.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@240309 91177308-0d34-0410-b5e6-96231b3b80d8
2015-06-22 18:24:50 +00:00
Rafael Espindola
84f9ca8e19 Do not pass -allow-shlib-undefined to the Solaris linker.
Patch by Xan López.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@240308 91177308-0d34-0410-b5e6-96231b3b80d8
2015-06-22 18:24:01 +00:00
Sanjoy Das
d5732752e8 Fix MSVC build.
I had some unnecessary `typename`s left in after addressing review.
This compiled successfully with clang++ but MSVC reported an error.  Fix
the build error by removing the redundant `typename`s.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@240307 91177308-0d34-0410-b5e6-96231b3b80d8
2015-06-22 18:20:10 +00:00
Rafael Espindola
21a540abc4 Remove the IsStreamed member variable.
Having different code paths for streamed and regular bitcode reading was a
source of bugs in the past and this defines them away.

It has a small but noticeable impact on performance. I timed running
"opt -disable-output -disable-verify" on a ltoed clang. It goes from

14.752845231 seconds time elapsed   ( +-  0.16% )

to

15.012463721 seconds time elapsed   ( +-  0.11% )

Extracted from a patch by Karl Schimpf.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@240305 91177308-0d34-0410-b5e6-96231b3b80d8
2015-06-22 18:06:15 +00:00
Sanjoy Das
e0ef46e5e7 [FaultMaps] Add a parser for the __llvm__faultmaps section.
Summary:
The parser is exercised by llvm-objdump using -print-fault-maps.  As is
probably obvious, the code itself was "heavily inspired" by
http://reviews.llvm.org/D10434.

Reviewers: reames, atrick, JosephTremoulet

Subscribers: llvm-commits

Differential Revision: http://reviews.llvm.org/D10491

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@240304 91177308-0d34-0410-b5e6-96231b3b80d8
2015-06-22 18:03:02 +00:00
Sanjoy Das
5a94e2a9e8 [NFC] Capitalization in documentation.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@240303 91177308-0d34-0410-b5e6-96231b3b80d8
2015-06-22 18:02:55 +00:00
Rafael Espindola
afd779f72d Bring r240130 back.
Now that pr23900 is fixed, we can bring it back with no changes.

Original message:

Make all temporary symbols unnamed.

What this does is make all symbols that would otherwise start with a .L
(or L on MachO) unnamed.

Some of these symbols still show up in the symbol table, but we can just
make them unnamed.

In order to make sure we produce identical results when going thought assembly,
all .L (not just the compiler produced ones), are now unnamed.

Running llc on llvm-as.opt.bc, the peak memory usage goes from 208.24MB to
205.57MB.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@240302 91177308-0d34-0410-b5e6-96231b3b80d8
2015-06-22 17:52:52 +00:00
Rafael Espindola
09bbd16112 Avoid a Symbol -> Name -> Symbol conversion.
Before this we were producing a TargetExternalSymbol from a MCSymbol.
That meant extracting the symbol name and fetching the symbol again
down the pipeline.

This patch adds a DAG.getMCSymbol that lets the MCSymbol pass unchanged on the
DAG.

Doing so removes the need for MO_NOPREFIX and fixes the root cause of pr23900,
allowing r240130 to be committed again.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@240300 91177308-0d34-0410-b5e6-96231b3b80d8
2015-06-22 17:46:53 +00:00
Alex Lorenz
2f801faafb MIR Serialization: Serialize machine instruction names.
This commit implements initial machine instruction serialization. It
serializes machine instruction names. The instructions are represented
using a YAML sequence of string literals and are a part of machine
basic block YAML mapping.

This commit introduces a class called 'MIParser' which will be used to
parse the machine instructions and operands.

Reviewers: Duncan P. N. Exon Smith

Differential Revision: http://reviews.llvm.org/D10481


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@240295 91177308-0d34-0410-b5e6-96231b3b80d8
2015-06-22 17:02:30 +00:00
Pawel Bylica
39befc6ca8 Fix shl folding in DAG combiner.
Summary: The code responsible for shl folding in the DAGCombiner was assuming incorrectly that all constants are less than 64 bits. This patch simply changes the way values are compared.

Test Plan: A regression test included.

Reviewers: andreadb

Reviewed By: andreadb

Subscribers: andreadb, test, llvm-commits

Differential Revision: http://reviews.llvm.org/D10602

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@240291 91177308-0d34-0410-b5e6-96231b3b80d8
2015-06-22 15:58:11 +00:00
Rafael Espindola
0ed63077a2 Add a triple to the test to fix it on some hosts.
The slp vectorizer doesn't optimize this case in 32 bits.

Fixes PR23453.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@240289 91177308-0d34-0410-b5e6-96231b3b80d8
2015-06-22 15:44:20 +00:00
Joseph Tremoulet
c29284e6c3 [ORC] Add NullResolver
Summary:
This is an implementation of RuntimeDyld::SymbolResolver that simply
rejects all resolution requests; useful for clients that do not have any
cross-object symbol references.

Reviewers: lhames

Reviewed By: lhames

Subscribers: llvm-commits

Differential Revision: http://reviews.llvm.org/D10455

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@240288 91177308-0d34-0410-b5e6-96231b3b80d8
2015-06-22 15:27:58 +00:00
Rafael Espindola
fc3dae1cd0 Support Solaris unused sections' gc link syntax.
It is not clear if this would work or not with LLVM_NO_DEAD_STRIP
binaries, so be conservative for now.

Patch by Xan López.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@240287 91177308-0d34-0410-b5e6-96231b3b80d8
2015-06-22 15:06:17 +00:00
Toma Tabacu
8204fb27ac [mips] [IAS] Add support for LAReg with identical source and destination register operands.
Summary: In this case, we're supposed to load the immediate in AT and then ADDu it with the source register and put it in the destination register.

Reviewers: dsanders

Reviewed By: dsanders

Subscribers: llvm-commits

Differential Revision: http://reviews.llvm.org/D9367

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@240278 91177308-0d34-0410-b5e6-96231b3b80d8
2015-06-22 13:10:23 +00:00
Elena Demikhovsky
114489ab24 AVX-512: added VPSHUFB instruction - all SKX forms
Added intrinsics and encoding tests.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@240277 91177308-0d34-0410-b5e6-96231b3b80d8
2015-06-22 13:00:42 +00:00
Rafael Espindola
e16fa7fec8 Do not pass optimization flags to Solaris' linker.
It is not supported.

Patch by Xan López.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@240276 91177308-0d34-0410-b5e6-96231b3b80d8
2015-06-22 12:41:52 +00:00
Rafael Espindola
4c8ba2e74d Use right syntax to pass version script to Solaris' ld.
Patch by Xan Lopez!

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@240275 91177308-0d34-0410-b5e6-96231b3b80d8
2015-06-22 12:34:54 +00:00
Toma Tabacu
d174678830 [mips] [IAS] Add support for LASym with identical source and destination register operands.
Summary:
In this case, we're supposed to load the address of the symbol in AT and then ADDu it with the source register and
put it in the destination register.

Reviewers: dsanders

Reviewed By: dsanders

Subscribers: llvm-commits

Differential Revision: http://reviews.llvm.org/D9366

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@240273 91177308-0d34-0410-b5e6-96231b3b80d8
2015-06-22 12:08:39 +00:00
Elena Demikhovsky
4f1ddd396b AVX-512: All forms of VCOPMRESS VEXPAND instructions,
encoding tests.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@240272 91177308-0d34-0410-b5e6-96231b3b80d8
2015-06-22 11:16:30 +00:00
Alexander Kornienko
948713ad4f Fixed/added namespace ending comments using clang-tidy. NFC
A few more files that were fixed while preparing r240270.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@240271 91177308-0d34-0410-b5e6-96231b3b80d8
2015-06-22 09:57:54 +00:00
Alexander Kornienko
a8c8b145fd Removing empty directories.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@240269 91177308-0d34-0410-b5e6-96231b3b80d8
2015-06-22 09:34:18 +00:00
Elena Demikhovsky
42ceb12123 Reverted AVX-512 vector shuffle
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@240258 91177308-0d34-0410-b5e6-96231b3b80d8
2015-06-22 09:01:15 +00:00
Michael Kuperstein
12219f8c85 [X86] Allow more call sequences to use push instructions for argument passing
This allows more call sequences to use pushes instead of movs when optimizing for size.
In particular, calling conventions that pass some parameters in registers (e.g. thiscall) are now supported.

Differential Revision: http://reviews.llvm.org/D10500

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@240257 91177308-0d34-0410-b5e6-96231b3b80d8
2015-06-22 08:31:22 +00:00
Elena Demikhovsky
c768510422 AVX-512: Added intrinsics for VPERMT2W/D/Q/PS/PD and
VPERMI2W/D/Q/PS/PD instructions.
Added tests.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@240256 91177308-0d34-0410-b5e6-96231b3b80d8
2015-06-22 06:45:48 +00:00
Chandler Carruth
1e3557de0d [PM/AA] Hoist the AliasResult enum out of the AliasAnalysis class.
This will allow classes to implement the AA interface without deriving
from the class or referencing an internal enum of some other class as
their return types.

Also, to a pretty fundamental extent, concepts such as 'NoAlias',
'MayAlias', and 'MustAlias' are first class concepts in LLVM and we
aren't saving anything by scoping them heavily.

My mild preference would have been to use a scoped enum, but that
feature is essentially completely broken AFAICT. I'm extremely
disappointed. For example, we cannot through any reasonable[1] means
construct an enum class (or analog) which has scoped names but converts
to a boolean in order to test for the possibility of aliasing.

[1]: Richard Smith came up with a "solution", but it requires class
templates, and lots of boilerplate setting up the enumeration multiple
times. Something like Boost.PP could potentially bundle this up, but
even that would be quite painful and it doesn't seem realistically worth
it. The enum class solution would probably work without the need for
a bool conversion.

Differential Revision: http://reviews.llvm.org/D10495

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@240255 91177308-0d34-0410-b5e6-96231b3b80d8
2015-06-22 02:16:51 +00:00
Chandler Carruth
ab6ddadac7 [PM/AA] Rework the names and comments in AliasSetTracker to more
accurately describe what is being tracked.

While these two enums do track mod/ref information and aliasing
information, they don't represent the exact same things as either the
mod/ref enums or the alias result enum in AA. They're definitions are
dominated by the structure of their lattice and the bit's various
semantics. This patch just calls them what they are and tries to spell
out usefully distinct names for these things.

This will clear the path for using a raw unscoped enum to represent some
of these concepts across LLVM's analysis library.

No functionality changed here.

Differential Revision: http://reviews.llvm.org/D10494

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@240254 91177308-0d34-0410-b5e6-96231b3b80d8
2015-06-22 02:12:52 +00:00
Rafael Espindola
2faef9ccb0 Add the testcase from pr23900.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@240253 91177308-0d34-0410-b5e6-96231b3b80d8
2015-06-22 01:29:24 +00:00
Simon Pilgrim
da5f3d8f76 [X86] Code tidyup - Use SDValue bool operator. NFC.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@240249 91177308-0d34-0410-b5e6-96231b3b80d8
2015-06-21 21:34:32 +00:00
Duncan P. N. Exon Smith
f1c527b5c1 AsmPrinter: Don't emit empty .debug_loc entries
If we don't know how to represent a .debug_loc entry, skip the entry
entirely rather than emitting an empty one.  Similarly, if a .debug_loc
list has no entries, don't create the list.

We still want to create the variables, just in an optimized-out form
that doesn't have a DW_AT_location.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@240244 91177308-0d34-0410-b5e6-96231b3b80d8
2015-06-21 16:54:56 +00:00
Duncan P. N. Exon Smith
01a8dc8ca6 AsmPrinter: Rewrite initialization of DbgVariable, NFC
There are three types of `DbgVariable`:
  - alloca variables, created based on the MMI table,
  - register variables, created based on DBG_VALUE instructions, and
  - optimized-out variables.

This commit reconfigures `DbgVariable` to make it easier to tell which
kind we have, and make initialization a little clearer.

For MMI/alloca variables, `FrameIndex.size()` must always equal
`Expr.size()`, and there shouldn't be an `MInsn`.  For register
variables (with a `MInsn`), `FrameIndex` must be empty, and `Expr`
should have 0 or 1 element depending on whether it has a complex
expression (registers with multiple locations use `DebugLocListIndex`).
Optimized-out variables shouldn't have any of these fields.

Moreover, this separates DBG_VALUE initialization until after the
variable is created, simplifying logic in a future commit that changes
`collectVariableInfo()` to stop creating empty .debug_loc entries/lists.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@240243 91177308-0d34-0410-b5e6-96231b3b80d8
2015-06-21 16:50:43 +00:00
Simon Pilgrim
87f860c737 [X86][SSE] Added missing stack folding test for CVTSD2SS instruction.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@240241 91177308-0d34-0410-b5e6-96231b3b80d8
2015-06-21 16:07:47 +00:00
NAKAMURA Takumi
b87f6217b3 OptTable.h: Prune a couple of \param(s), since Arg has been ArrayRef-ized. [-Wdocumentation]
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@240240 91177308-0d34-0410-b5e6-96231b3b80d8
2015-06-21 13:44:46 +00:00
David Blaikie
caca524562 Devirtualize ArgList's dtor now that -Wvirtual-dtor and C++11 allow a better way to describe this situation
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@240238 91177308-0d34-0410-b5e6-96231b3b80d8
2015-06-21 06:51:35 +00:00
David Blaikie
fa6a5de95b ArrayRef-ify libDriverMain
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@240234 91177308-0d34-0410-b5e6-96231b3b80d8
2015-06-21 06:31:56 +00:00
David Blaikie
006bce7764 ArrayRef-ify ParseArgs
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@240233 91177308-0d34-0410-b5e6-96231b3b80d8
2015-06-21 06:31:53 +00:00
Rui Ueyama
87ae2e8ab0 Object: Add load configuration structure for 64 bit PE/COFF.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@240230 91177308-0d34-0410-b5e6-96231b3b80d8
2015-06-21 01:12:32 +00:00
Hans Wennborg
ee16995e3d Switch lowering: add heuristic for filling leaf nodes in the weight-balanced binary search tree
Sparse switches with profile info are lowered as weight-balanced BSTs. For
example, if the node weights are {1,1,1,1,1,1000}, the right-most node would
end up in a tree by itself, bringing it closer to the top.

However, a leaf in this BST can contain up to 3 cases, and having a single
case in a leaf node as in the example means the tree might become
unnecessarily high.

This patch adds a heauristic to the pivot selection algorithm that moves more
cases into leaf nodes unless that would lower their rank. It still doesn't
yield the optimal tree in every case, but I believe it's conservatibely correct.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@240224 91177308-0d34-0410-b5e6-96231b3b80d8
2015-06-20 17:14:07 +00:00
Simon Pilgrim
1eabc9fb9d [X86][SSE] Fix PerformSExtCombine bug that accessed the wrong return value of an aggregate type.
Fix to rL237885 to ensure that it accesses the correct return value of an aggregate type.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@240223 91177308-0d34-0410-b5e6-96231b3b80d8
2015-06-20 16:19:24 +00:00
Benjamin Kramer
793ca75174 [PPC] Factor vector removal into a function and remove O(n^2) behavior.
No functionality change intended.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@240222 91177308-0d34-0410-b5e6-96231b3b80d8
2015-06-20 15:59:41 +00:00