Arnold Schwaighofer [Fri, 1 Nov 2013 22:18:19 +0000 (22:18 +0000)]
LoopVectorizer: Perform redundancy elimination on induction variables
When the loop vectorizer was part of the SCC inliner pass manager gvn would
run after the loop vectorizer followed by instcombine. This way redundancy
(multiple uses) were removed and instcombine could perform scalarization on the
induction variables. Having moved the loop vectorizer to later we no longer run
any form of redundancy elimination before we perform instcombine. This caused
vectorized induction variables to survive that did not before.
On a recent iMac this helps linpack back from 6000Mflops to 7000Mflops.
This should also help lpbench and paq8p.
I ran a Release (without Asserts) build over the test-suite and did not see any
negative impact on compile time.
radar://
15339680
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@193891
91177308-0d34-0410-b5e6-
96231b3b80d8
David Blaikie [Fri, 1 Nov 2013 20:59:53 +0000 (20:59 +0000)]
DebugInfo: Improve readability of test case added in r193878
The point is to ensure that the attribute in question
(DW_AT_data_member_location) is associated with the prior tag, so ensure
that we don't see another tag starting between the intended tag and the
desired attribute.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@193884
91177308-0d34-0410-b5e6-
96231b3b80d8
David Blaikie [Fri, 1 Nov 2013 18:25:55 +0000 (18:25 +0000)]
DebugInfo: add a test case for data member locations (coverage for r193835)
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@193878
91177308-0d34-0410-b5e6-
96231b3b80d8
David Blaikie [Fri, 1 Nov 2013 18:18:16 +0000 (18:18 +0000)]
Fix a test case broken by r193872
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@193876
91177308-0d34-0410-b5e6-
96231b3b80d8
Manman Ren [Fri, 1 Nov 2013 18:06:25 +0000 (18:06 +0000)]
Add comments.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@193874
91177308-0d34-0410-b5e6-
96231b3b80d8
David Blaikie [Fri, 1 Nov 2013 17:53:30 +0000 (17:53 +0000)]
DebugInfo: Make pubnames header printing similar to unit header printing
In a failed attempt to allow the gnu-public-names.ll test case to not
hardcode the size of the unit that the pubnames section referred to I've
at least managed to have unit headers and pubnames headers print out in
a similar style.
This failed to achieve the desired goal because the header in a unit
specifies the length of the unit without the length element of the
header whereas the length in the pubnames includes this element, so the
numbers are off by 4 bytes. I don't know of any arithmetic powers in
FileCheck so the test case can't simply say "CU_LENGTH + 4".
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@193872
91177308-0d34-0410-b5e6-
96231b3b80d8
Juergen Ributzka [Fri, 1 Nov 2013 17:53:27 +0000 (17:53 +0000)]
[Stackmap] Remove erroneous assert.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@193871
91177308-0d34-0410-b5e6-
96231b3b80d8
Matt Arsenault [Fri, 1 Nov 2013 17:39:26 +0000 (17:39 +0000)]
Use isa<> instead of dyn_cast<> with unused value
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@193869
91177308-0d34-0410-b5e6-
96231b3b80d8
Chad Rosier [Fri, 1 Nov 2013 17:13:44 +0000 (17:13 +0000)]
[AArch64] Simplify a few of the instruction patterns. No functional change intended.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@193867
91177308-0d34-0410-b5e6-
96231b3b80d8
Chad Rosier [Fri, 1 Nov 2013 17:13:42 +0000 (17:13 +0000)]
[AArch64] Fix assembly string formatting and other coding standard violations.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@193866
91177308-0d34-0410-b5e6-
96231b3b80d8
Rafael Espindola [Fri, 1 Nov 2013 17:09:14 +0000 (17:09 +0000)]
Remove linkonce_odr_auto_hide.
linkonce_odr_auto_hide was in incomplete attempt to implement a way
for the linker to hide symbols that are known to be available in every
TU and whose addresses are not relevant for a particular DSO.
It was redundant in that it all its uses are equivalent to
linkonce_odr+unnamed_addr. Unlike those, it has never been connected
to clang or llvm's optimizers, so it was effectively dead.
Given that nothing produces it, this patch just nukes it
(other than the llvm-c enum value).
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@193865
91177308-0d34-0410-b5e6-
96231b3b80d8
Aaron Ballman [Fri, 1 Nov 2013 15:12:23 +0000 (15:12 +0000)]
Commenting out this assert because it is causing the build bots to fail. This effectively reverts r193861, but needs to be fixed as part of r193769.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@193862
91177308-0d34-0410-b5e6-
96231b3b80d8
Aaron Ballman [Fri, 1 Nov 2013 14:53:14 +0000 (14:53 +0000)]
Fixing an order of evaluation error in an assert.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@193861
91177308-0d34-0410-b5e6-
96231b3b80d8
Benjamin Kramer [Fri, 1 Nov 2013 14:09:50 +0000 (14:09 +0000)]
LoopVectorize: Look for consecutive acces in GEPs with trailing zero indices
If we have a pointer to a single-element struct we can still build wide loads
and stores to it (if there is no padding).
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@193860
91177308-0d34-0410-b5e6-
96231b3b80d8
Bradley Smith [Fri, 1 Nov 2013 13:27:35 +0000 (13:27 +0000)]
[ARM] Add Virtualization subtarget feature and more build attributes in this area
Add a Virtualization ARM subtarget feature along with adding proper build
attribute emission for Tag_Virtualization_use (encodes Virtualization and
TrustZone) and Tag_MPextension_use.
Also rework test/CodeGen/ARM/2010-10-19-mc-elf-objheader.ll testcase to
something that is more maintainable. This changes the focus of this
testcase away from testing CPU defaults (which is tested elsewhere), onto
specifically testing that attributes are encoded correctly.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@193859
91177308-0d34-0410-b5e6-
96231b3b80d8
Bradley Smith [Fri, 1 Nov 2013 11:21:16 +0000 (11:21 +0000)]
[ARM] Fix Tag_ABI_HardFP_use build attribute
Fix Tag_ABI_HardFP_use build attribute to handle single precision FP,
replace deprecated Tag_ABI_HardFP_use value of 3 with 0 and also add
some tests for Tag_ABI_VFP_args.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@193856
91177308-0d34-0410-b5e6-
96231b3b80d8
Hal Finkel [Fri, 1 Nov 2013 10:58:22 +0000 (10:58 +0000)]
Consider (x == -1) unlikely in BranchProbabilityInfo
This adds another heuristic to BPI, similar to the existing heuristic that
considers (x == 0) unlikely to be true. As suggested in the PACT'98 paper by
Deitrich, Cheng, and Hwu, -1 is often used to indicate an invalid index, and
equality comparisons with -1 are also unlikely to succeed. Local
experimentation supports this hypothesis: This yields a 1-2% speedup in the
test-suite sqlite benchmark on the PPC A2 core, with no significant
regressions.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@193855
91177308-0d34-0410-b5e6-
96231b3b80d8
Arnold Schwaighofer [Fri, 1 Nov 2013 03:05:07 +0000 (03:05 +0000)]
LoopVectorizer: If dependency checks fail try runtime checks
When a dependence check fails we can still try to vectorize loops with runtime
array bounds checks.
This helps linpack to vectorize a loop in dgefa. And we are back to 2x of the
scalar performance on a corei7-avx.
radar://
15339680
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@193853
91177308-0d34-0410-b5e6-
96231b3b80d8
Arnold Schwaighofer [Fri, 1 Nov 2013 03:05:04 +0000 (03:05 +0000)]
LoopVectorizer: Clear all member data structures in RuntimeCheck.reset()
Clear all data structures when resetting the RuntimeCheck data structure.
No test case. This was exposed by an upcomming change.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@193852
91177308-0d34-0410-b5e6-
96231b3b80d8
Sylvestre Ledru [Fri, 1 Nov 2013 02:29:13 +0000 (02:29 +0000)]
OCaml bindings: fix typo
Patch by Peter Zotov
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@193851
91177308-0d34-0410-b5e6-
96231b3b80d8
Sylvestre Ledru [Fri, 1 Nov 2013 01:14:24 +0000 (01:14 +0000)]
OCaml bindings: remove unused DONT_BUILD_RELINKED from Makefiles
Patch by Peter Zotov
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@193847
91177308-0d34-0410-b5e6-
96231b3b80d8
Rafael Espindola [Fri, 1 Nov 2013 01:14:20 +0000 (01:14 +0000)]
Use \01 to disable the mangler. Should fix the 32 bit windows bots.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@193846
91177308-0d34-0410-b5e6-
96231b3b80d8
Sylvestre Ledru [Fri, 1 Nov 2013 01:10:58 +0000 (01:10 +0000)]
OCaml bindings: fix typo
Patch by Peter Zotov
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@193845
91177308-0d34-0410-b5e6-
96231b3b80d8
Sylvestre Ledru [Fri, 1 Nov 2013 00:30:02 +0000 (00:30 +0000)]
OCaml bindings: fix typo in documentation
Patch by Peter Zotov
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@193838
91177308-0d34-0410-b5e6-
96231b3b80d8
Sylvestre Ledru [Fri, 1 Nov 2013 00:26:01 +0000 (00:26 +0000)]
OCaml bindings: formatting
This commit only changes comments and documentation in OCaml bindings. The official name of the language is OCaml, and the usage is now consistent.
Patch by Peter Zotov
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@193836
91177308-0d34-0410-b5e6-
96231b3b80d8
David Blaikie [Fri, 1 Nov 2013 00:25:45 +0000 (00:25 +0000)]
DebugInfo: Emit member variable locations as data instead of expressions in blocks
Drive by space optimization. Also makes the DIEs more regular which
might speed up DWARF parsing.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@193835
91177308-0d34-0410-b5e6-
96231b3b80d8
Kevin Enderby [Fri, 1 Nov 2013 00:00:07 +0000 (00:00 +0000)]
Add to the disassembler C API output reference types for
Objective-C data structures.
This is allows tools such as darwin's otool(1) that uses the
LLVM disassembler take a pointer value being loaded by
an instruction and add a comment to what it is being referenced
to make following disassembly of Objective-C programs
more readable.
For example disassembling the Mac OS X TextEdit app one
will see comments like the following:
movq 0x20684(%rip), %rsi ## Objc selector ref: standardUserDefaults
movq 0x21985(%rip), %rdi ## Objc class ref: _OBJC_CLASS_$_NSUserDefaults
movq 0x1d156(%rip), %r14 ## Objc message: +[NSUserDefaults standardUserDefaults]
leaq 0x23615(%rip), %rdx ## Objc cfstring ref: @"SelectLinePanel"
callq 0x10001386c ## Objc message: -[[%rdi super] initWithWindowNibName:]
These diffs also include putting quotes around C strings
in literal pools and uses "symbol address" in the comment
when adding a symbol name to the comment to tell these
types of references apart:
leaq 0x4f(%rip), %rax ## literal pool for: "Hello world"
movq 0x1c3ea(%rip), %rax ## literal pool symbol address: ___stack_chk_guard
Of course the easy changes are in the LLVM disassembler and
the hard work is up to the implementer of the SymbolLookUp()
call back.
rdar://
10602439
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@193833
91177308-0d34-0410-b5e6-
96231b3b80d8
Dan Gohman [Thu, 31 Oct 2013 22:58:11 +0000 (22:58 +0000)]
Fix unused variable warnings.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@193823
91177308-0d34-0410-b5e6-
96231b3b80d8
Andrew Trick [Thu, 31 Oct 2013 22:46:51 +0000 (22:46 +0000)]
These test cases for experimental features are a bit too darwin-specific still. Use a triple.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@193820
91177308-0d34-0410-b5e6-
96231b3b80d8
Andrew Trick [Thu, 31 Oct 2013 22:42:20 +0000 (22:42 +0000)]
Unused variable
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@193819
91177308-0d34-0410-b5e6-
96231b3b80d8
Chad Rosier [Thu, 31 Oct 2013 22:36:59 +0000 (22:36 +0000)]
[AArch64] Add support for NEON scalar fixed-point convert to floating-point instructions.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@193816
91177308-0d34-0410-b5e6-
96231b3b80d8
Rafael Espindola [Thu, 31 Oct 2013 22:13:41 +0000 (22:13 +0000)]
Modify CMakeLists.txt to work around pr17763 and bring some bots back.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@193813
91177308-0d34-0410-b5e6-
96231b3b80d8
Andrew Trick [Thu, 31 Oct 2013 22:12:01 +0000 (22:12 +0000)]
Add new calling convention for WebKit Java Script.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@193812
91177308-0d34-0410-b5e6-
96231b3b80d8
Andrew Trick [Thu, 31 Oct 2013 22:11:56 +0000 (22:11 +0000)]
Add support for stack map generation in the X86 backend.
Originally implemented by Lang Hames.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@193811
91177308-0d34-0410-b5e6-
96231b3b80d8
Rafael Espindola [Thu, 31 Oct 2013 22:07:46 +0000 (22:07 +0000)]
Relax check line to match what llvm-nm prints for COFF.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@193810
91177308-0d34-0410-b5e6-
96231b3b80d8
Manman Ren [Thu, 31 Oct 2013 21:56:03 +0000 (21:56 +0000)]
Do not convert "call asm" to "invoke asm" in Inliner.
Given that backend does not handle "invoke asm" correctly ("invoke asm" will be
handled by SelectionDAGBuilder::visitInlineAsm, which does not have the right
setup for LPadToCallSiteMap) and we already made the assumption that inline asm
does not throw in InstCombiner::visitCallSite, we are going to make the same
assumption in Inliner to make sure we don't convert "call asm" to "invoke asm".
If it becomes necessary to add support for "invoke asm" later on, we will need
to modify the backend as well as remove the assumptions that inline asm does
not throw.
Fix rdar://
15317907
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@193808
91177308-0d34-0410-b5e6-
96231b3b80d8
Rafael Espindola [Thu, 31 Oct 2013 21:27:02 +0000 (21:27 +0000)]
XFAIL on ppc64 too.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@193804
91177308-0d34-0410-b5e6-
96231b3b80d8
Rafael Espindola [Thu, 31 Oct 2013 21:22:43 +0000 (21:22 +0000)]
XFAIL this for now.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@193802
91177308-0d34-0410-b5e6-
96231b3b80d8
Rafael Espindola [Thu, 31 Oct 2013 20:51:58 +0000 (20:51 +0000)]
Use LTO_SYMBOL_SCOPE_DEFAULT_CAN_BE_HIDDEN instead of the "dso list".
There are two ways one could implement hiding of linkonce_odr symbols in LTO:
* LLVM tells the linker which symbols can be hidden if not used from native
files.
* The linker tells LLVM which symbols are not used from other object files,
but will be put in the dso symbol table if present.
GOLD's API is the second option. It was implemented almost 1:1 in llvm by
passing the list down to internalize.
LLVM already had partial support for the first option. It is also very similar
to how ld64 handles hiding these symbols when *not* doing LTO.
This patch then
* removes the APIs for the DSO list.
* marks LTO_SYMBOL_SCOPE_DEFAULT_CAN_BE_HIDDEN all linkonce_odr unnamed_addr
global values and other linkonce_odr whose address is not used.
* makes the gold plugin responsible for handling the API mismatch.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@193800
91177308-0d34-0410-b5e6-
96231b3b80d8
Chad Rosier [Thu, 31 Oct 2013 20:11:32 +0000 (20:11 +0000)]
[AArch64] Add diagnostic tests for NEON scalar shift immediate instructions (see: r193790).
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@193798
91177308-0d34-0410-b5e6-
96231b3b80d8
Rui Ueyama [Thu, 31 Oct 2013 19:59:55 +0000 (19:59 +0000)]
Use StringRef::startswith_lower. No functionality change.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@193796
91177308-0d34-0410-b5e6-
96231b3b80d8
Nuno Lopes [Thu, 31 Oct 2013 19:53:53 +0000 (19:53 +0000)]
[ConstantRange] improve my previous patch per Nick suggestion
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@193795
91177308-0d34-0410-b5e6-
96231b3b80d8
Chad Rosier [Thu, 31 Oct 2013 19:28:44 +0000 (19:28 +0000)]
[AArch64] Add support for NEON scalar shift immediate instructions.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@193790
91177308-0d34-0410-b5e6-
96231b3b80d8
Roman Divacky [Thu, 31 Oct 2013 19:22:33 +0000 (19:22 +0000)]
SparcV9 doesnt have rem instruction either.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@193789
91177308-0d34-0410-b5e6-
96231b3b80d8
Reid Kleckner [Thu, 31 Oct 2013 19:12:36 +0000 (19:12 +0000)]
Use a larger invalid attribute bitcode number
That way the test won't start faililng when someone adds a new attribute
and wants to use the next logical enum (38) for bitcode. The new
bitcode file tries to use the number 48 as an attribute instead.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@193787
91177308-0d34-0410-b5e6-
96231b3b80d8
Alexey Samsonov [Thu, 31 Oct 2013 18:54:20 +0000 (18:54 +0000)]
DWARFDebugArangeSet: remove dead code
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@193785
91177308-0d34-0410-b5e6-
96231b3b80d8
Matt Arsenault [Thu, 31 Oct 2013 18:18:09 +0000 (18:18 +0000)]
Add FileCheck tests for @LINE
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@193782
91177308-0d34-0410-b5e6-
96231b3b80d8
Petar Jovanovic [Thu, 31 Oct 2013 18:10:25 +0000 (18:10 +0000)]
[mips] XFAIL several MCJIT remote tests
Two of the tests are new test cases (cross-module-a.ll, multi-module-a.ll)
not yet supported on MIPS, while XFAIL for the other two tests was
accidentally removed in r193570 and this change reverts those lines.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@193781
91177308-0d34-0410-b5e6-
96231b3b80d8
Alexey Samsonov [Thu, 31 Oct 2013 18:05:02 +0000 (18:05 +0000)]
DWARFUnit: kill dead code and make a couple of functions private. No functionality change.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@193780
91177308-0d34-0410-b5e6-
96231b3b80d8
Manman Ren [Thu, 31 Oct 2013 17:54:35 +0000 (17:54 +0000)]
Debug Info: remove duplication of DIEs when a DIE can be shared across CUs.
We add a map in DwarfDebug to map MDNodes that are shareable across CUs to the
corresponding DIEs: MDTypeNodeToDieMap. These DIEs can be shared across CUs,
that is why we keep the maps in DwarfDebug instead of CompileUnit.
We make the assumption that if a DIE is not added to an owner yet, we assume
it belongs to the current CU. Since DIEs for the type system are added to
their owners immediately after creation, and other DIEs belong to the current
CU, the assumption should be true.
A testing case is added to show that we only create a single DIE for a type
MDNode and we use ref_addr to refer to the type DIE.
We also add a testing case to show ref_addr relocations for non-darwin
platforms.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@193779
91177308-0d34-0410-b5e6-
96231b3b80d8
Roman Divacky [Thu, 31 Oct 2013 17:50:45 +0000 (17:50 +0000)]
Merge and filecheckize.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@193778
91177308-0d34-0410-b5e6-
96231b3b80d8
Manman Ren [Thu, 31 Oct 2013 17:25:22 +0000 (17:25 +0000)]
Cleanup: update comments.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@193773
91177308-0d34-0410-b5e6-
96231b3b80d8
Alexey Samsonov [Thu, 31 Oct 2013 17:20:14 +0000 (17:20 +0000)]
DWARFAbbreviationDeclaration: remove dead code, refactor parsing code and make it more robust. No functionality change.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@193770
91177308-0d34-0410-b5e6-
96231b3b80d8
Andrew Trick [Thu, 31 Oct 2013 17:18:24 +0000 (17:18 +0000)]
Lower stackmap intrinsics directly to their target opcode in the DAG builder.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@193769
91177308-0d34-0410-b5e6-
96231b3b80d8
Andrew Trick [Thu, 31 Oct 2013 17:18:17 +0000 (17:18 +0000)]
Add Verifier test case for variable argument intrinsics.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@193768
91177308-0d34-0410-b5e6-
96231b3b80d8
Andrew Trick [Thu, 31 Oct 2013 17:18:14 +0000 (17:18 +0000)]
Add experimental stackmap intrinsics to definition file and documenation.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@193767
91177308-0d34-0410-b5e6-
96231b3b80d8
Andrew Trick [Thu, 31 Oct 2013 17:18:11 +0000 (17:18 +0000)]
Enable variable arguments support for intrinsics.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@193766
91177308-0d34-0410-b5e6-
96231b3b80d8
Andrew Trick [Thu, 31 Oct 2013 17:18:07 +0000 (17:18 +0000)]
whitespace
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@193765
91177308-0d34-0410-b5e6-
96231b3b80d8
Rafael Espindola [Thu, 31 Oct 2013 15:58:33 +0000 (15:58 +0000)]
Remove another unused flag.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@193756
91177308-0d34-0410-b5e6-
96231b3b80d8
Rafael Espindola [Thu, 31 Oct 2013 15:49:39 +0000 (15:49 +0000)]
Remove unused flag.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@193752
91177308-0d34-0410-b5e6-
96231b3b80d8
Rafael Espindola [Thu, 31 Oct 2013 14:35:00 +0000 (14:35 +0000)]
Rules adjustments in order to build on DragonFly BSD.
Patch by Robin Hahling.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@193750
91177308-0d34-0410-b5e6-
96231b3b80d8
Rafael Espindola [Thu, 31 Oct 2013 14:07:59 +0000 (14:07 +0000)]
Remove the --shrink-wrap option.
It had no tests, was unused and was "experimental at best".
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@193749
91177308-0d34-0410-b5e6-
96231b3b80d8
Cameron McInally [Thu, 31 Oct 2013 13:56:31 +0000 (13:56 +0000)]
Add AVX512 unmasked integer broadcast intrinsics and support.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@193748
91177308-0d34-0410-b5e6-
96231b3b80d8
Elena Demikhovsky [Thu, 31 Oct 2013 13:15:32 +0000 (13:15 +0000)]
AVX-512: Implemented CMOV for 512-bit vectors
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@193747
91177308-0d34-0410-b5e6-
96231b3b80d8
Richard Sandiford [Thu, 31 Oct 2013 12:14:17 +0000 (12:14 +0000)]
[SystemZ] Automatically detect zEC12 and z196 hosts
As on other hosts, the CPU identification instruction is priveleged,
so we need to look through /proc/cpuinfo. I copied the PowerPC way of
handling "generic".
Several tests were implicitly assuming z10 and so failed on z196.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@193742
91177308-0d34-0410-b5e6-
96231b3b80d8
Amara Emerson [Thu, 31 Oct 2013 09:32:11 +0000 (09:32 +0000)]
[AArch64] Make the use of FP instructions optional, but enabled by default.
This adds a new subtarget feature called FPARMv8 (implied by NEON), and
predicates the support of the FP instructions and registers on this feature.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@193739
91177308-0d34-0410-b5e6-
96231b3b80d8
NAKAMURA Takumi [Thu, 31 Oct 2013 06:21:00 +0000 (06:21 +0000)]
llvm/test/Bitcode/invalid.ll: Tweak expresion to mach "llvm-dis.EXE:"
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@193738
91177308-0d34-0410-b5e6-
96231b3b80d8
Rafael Espindola [Thu, 31 Oct 2013 04:20:23 +0000 (04:20 +0000)]
Fix a use after free on invalid input.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@193737
91177308-0d34-0410-b5e6-
96231b3b80d8
Rafael Espindola [Thu, 31 Oct 2013 04:07:41 +0000 (04:07 +0000)]
Fix most memory leaks in tablegen.
Found by the valgrind bot.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@193736
91177308-0d34-0410-b5e6-
96231b3b80d8
Rafael Espindola [Thu, 31 Oct 2013 03:03:55 +0000 (03:03 +0000)]
Merge CallGraph and BasicCallGraph.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@193734
91177308-0d34-0410-b5e6-
96231b3b80d8
Yuchen Wu [Thu, 31 Oct 2013 02:01:24 +0000 (02:01 +0000)]
Updated llvm-cov's OVERVIEW description
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@193732
91177308-0d34-0410-b5e6-
96231b3b80d8
Jim Grosbach [Thu, 31 Oct 2013 00:20:48 +0000 (00:20 +0000)]
Legalize: Improve legalization of long vector extends.
When an extend more than doubles the size of the elements (e.g., a zext
from v16i8 to v16i32), the normal legalization method of splitting the
vectors will run into problems as by the time the destination vector is
legal, the source vector is illegal. The end result is the operation
often becoming scalarized, with the typical horrible performance. For
example, on x86_64, the simple input of:
define void @bar(<16 x i8> %a, <16 x i32>* %p) nounwind {
%tmp = zext <16 x i8> %a to <16 x i32>
store <16 x i32> %tmp, <16 x i32>*%p
ret void
}
Generates:
.section __TEXT,__text,regular,pure_instructions
.section __TEXT,__const
.align 5
LCPI0_0:
.long 255 ## 0xff
.long 255 ## 0xff
.long 255 ## 0xff
.long 255 ## 0xff
.long 255 ## 0xff
.long 255 ## 0xff
.long 255 ## 0xff
.long 255 ## 0xff
.section __TEXT,__text,regular,pure_instructions
.globl _bar
.align 4, 0x90
_bar:
vpunpckhbw %xmm0, %xmm0, %xmm1
vpunpckhwd %xmm0, %xmm1, %xmm2
vpmovzxwd %xmm1, %xmm1
vinsertf128 $1, %xmm2, %ymm1, %ymm1
vmovaps LCPI0_0(%rip), %ymm2
vandps %ymm2, %ymm1, %ymm1
vpmovzxbw %xmm0, %xmm3
vpunpckhwd %xmm0, %xmm3, %xmm3
vpmovzxbd %xmm0, %xmm0
vinsertf128 $1, %xmm3, %ymm0, %ymm0
vandps %ymm2, %ymm0, %ymm0
vmovaps %ymm0, (%rdi)
vmovaps %ymm1, 32(%rdi)
vzeroupper
ret
So instead we can check if there are legal types that enable us to split
more cleverly when the input vector is already legal such that we don't
turn it into an illegal type. If the extend is such that it's more than
doubling the size of the input we check if
- the number of vector elements is even,
- the source type is legal,
- the type of a split source is illegal,
- the type of an extended (by doubling element size) source is legal, and
- the type of that extended source when split is legal.
If the conditions are met, instead of just splitting both the
destination and the source types, we create an extend that only goes up
one "step" (doubling the element width), and the continue legalizing the
rest of the operation normally. The result is that this operates as a
new, more effecient, termination condition for the loop of "split the
operation until the destination type is legal."
With this change, the above example now compiles to:
_bar:
vpxor %xmm1, %xmm1, %xmm1
vpunpcklbw %xmm1, %xmm0, %xmm2
vpunpckhwd %xmm1, %xmm2, %xmm3
vpunpcklwd %xmm1, %xmm2, %xmm2
vinsertf128 $1, %xmm3, %ymm2, %ymm2
vpunpckhbw %xmm1, %xmm0, %xmm0
vpunpckhwd %xmm1, %xmm0, %xmm3
vpunpcklwd %xmm1, %xmm0, %xmm0
vinsertf128 $1, %xmm3, %ymm0, %ymm0
vmovaps %ymm0, 32(%rdi)
vmovaps %ymm2, (%rdi)
vzeroupper
ret
This generalizes a custom lowering that was added a while back to the
ARM backend. That lowering is no longer necessary, and is removed. The
testcases for it, however, provide excellent ARM tests for this change
and so remain.
rdar://
14735100
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@193727
91177308-0d34-0410-b5e6-
96231b3b80d8
Matt Arsenault [Wed, 30 Oct 2013 23:43:29 +0000 (23:43 +0000)]
Fix a few typos
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@193723
91177308-0d34-0410-b5e6-
96231b3b80d8
Matt Arsenault [Wed, 30 Oct 2013 23:30:05 +0000 (23:30 +0000)]
Fix CodeGen for unaligned loads with address spaces
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@193721
91177308-0d34-0410-b5e6-
96231b3b80d8
Matt Arsenault [Wed, 30 Oct 2013 22:54:58 +0000 (22:54 +0000)]
Teach scalarrepl about address spaces
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@193720
91177308-0d34-0410-b5e6-
96231b3b80d8
Rafael Espindola [Wed, 30 Oct 2013 22:37:51 +0000 (22:37 +0000)]
Add calls to doInitialization() and doFinalization() in verifyFunction()
The function verifyFunction() in lib/IR/Verifier.cpp misses some
calls. It creates a temporary FunctionPassManager that will run a
single Verifier pass. Unfortunately, FunctionPassManager is no
PassManager and does not call doInitialization() and doFinalization()
by itself. Verifier does important tasks in doInitialization() such as
collecting type information used to check DebugInfo metadata and
doFinalization() does some additional checks. Therefore these checks
were missed and debug info couldn't be verified at all, it just
crashed if the function had some.
verifyFunction() is currently not used in llvm unless -debug option is
enabled, and in unittests/IR/VerifierTest.cpp
VerifierTest had to be changed to create the function in a module from
which the type debug info can be collected.
Patch by Michael Kruse.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@193719
91177308-0d34-0410-b5e6-
96231b3b80d8
Rafael Espindola [Wed, 30 Oct 2013 22:08:11 +0000 (22:08 +0000)]
Produce .weak_def_can_be_hidden for some linkonce_odr values
With this patch llvm produces a weak_def_can_be_hidden for linkonce_odr
if they are also unnamed_addr or don't have their address taken.
There is not a lot of documentation about .weak_def_can_be_hidden, but
from the old discussion about linkonce_odr_auto_hide and the name of
the directive this looks correct: these symbols can be hidden.
Testing this with the ld64 in Xcode 5 linking clang reduces the number of
exported symbols from 21053 to 19049.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@193718
91177308-0d34-0410-b5e6-
96231b3b80d8
David Blaikie [Wed, 30 Oct 2013 20:42:41 +0000 (20:42 +0000)]
DebugInfo: Push header handling down into CompileUnit
This is a preliminary step to handling type units by abstracting over
all (type or compile) units.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@193714
91177308-0d34-0410-b5e6-
96231b3b80d8
Simon Atanasyan [Wed, 30 Oct 2013 20:41:45 +0000 (20:41 +0000)]
[Mips] Add more SHF_MIPS_xxx ELF section flags.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@193713
91177308-0d34-0410-b5e6-
96231b3b80d8
Will Dietz [Wed, 30 Oct 2013 20:27:17 +0000 (20:27 +0000)]
Add DebugInfo testcase for high_pc encoded as constant, fixed in r193555.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@193711
91177308-0d34-0410-b5e6-
96231b3b80d8
Matt Arsenault [Wed, 30 Oct 2013 19:05:41 +0000 (19:05 +0000)]
Fix GVN creating bitcast between address spaces
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@193710
91177308-0d34-0410-b5e6-
96231b3b80d8
Tom Roeder [Wed, 30 Oct 2013 18:47:25 +0000 (18:47 +0000)]
This commit adds some (but not all) of the x86-64 relocations that are not
currently supported in the ELF object writer, along with a simple test case.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@193709
91177308-0d34-0410-b5e6-
96231b3b80d8
Rui Ueyama [Wed, 30 Oct 2013 18:32:26 +0000 (18:32 +0000)]
Add {start,end}with_lower methods to StringRef.
startswith_lower is ocassionally useful and I think worth adding.
endwith_lower is added for completeness.
Differential Revision: http://llvm-reviews.chandlerc.com/D2041
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@193706
91177308-0d34-0410-b5e6-
96231b3b80d8
Artyom Skrobov [Wed, 30 Oct 2013 18:10:09 +0000 (18:10 +0000)]
[ARM] NEON instructions were erroneously decoded from certain invalid encodings
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@193705
91177308-0d34-0410-b5e6-
96231b3b80d8
Tom Stellard [Wed, 30 Oct 2013 17:22:05 +0000 (17:22 +0000)]
R600: Custom lower f32 = uint_to_fp i64
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@193701
91177308-0d34-0410-b5e6-
96231b3b80d8
David Blaikie [Wed, 30 Oct 2013 17:14:24 +0000 (17:14 +0000)]
DwarfDebug: Change Abbreviations member from pointer to reference
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@193699
91177308-0d34-0410-b5e6-
96231b3b80d8
Benjamin Kramer [Wed, 30 Oct 2013 17:09:32 +0000 (17:09 +0000)]
fix RST reference in Writing an LLVM Pass
Currently, instead of showing up as link, it is rendered as
...of FunctionPass <writing-an-llvm-pass-FunctionPass>. The...
PR17733. Patch by Tay Ray Chuan!
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@193698
91177308-0d34-0410-b5e6-
96231b3b80d8
Hans Wennborg [Wed, 30 Oct 2013 16:10:10 +0000 (16:10 +0000)]
Add #include of raw_ostream.h to MipsSEISelLowering.cpp
Fixing this Windows build error:
..\lib\Target\Mips\MipsSEISelLowering.cpp(997) : error C2027: use of undefined type 'llvm::raw_ostream'
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@193696
91177308-0d34-0410-b5e6-
96231b3b80d8
Daniel Sanders [Wed, 30 Oct 2013 15:45:42 +0000 (15:45 +0000)]
[mips][msa] Correct definition of bins[lr] and CHECK-DAG-ize related tests
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@193695
91177308-0d34-0410-b5e6-
96231b3b80d8
Nuno Lopes [Wed, 30 Oct 2013 15:36:50 +0000 (15:36 +0000)]
make ConstantRange::signExtend() optimal
the case [x, INT_MIN) was not handled optimally
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@193694
91177308-0d34-0410-b5e6-
96231b3b80d8
Daniel Sanders [Wed, 30 Oct 2013 15:20:38 +0000 (15:20 +0000)]
[mips][msa] Added support for matching bmnz, bmnzi, bmz, and bmzi from normal IR (i.e. not intrinsics)
Also corrected the definition of the intrinsics for these instructions (the
result register is also the first operand), and added intrinsics for bsel and
bseli to clang (they already existed in the backend).
These four operations are mostly equivalent to bsel, and bseli (the difference
is which operand is tied to the result). As a result some of the tests changed
as described below.
bitwise.ll:
- bsel.v test adapted so that the mask is unknown at compile-time. This stops
it emitting bmnzi.b instead of the intended bsel.v.
- The bseli.b test now tests the right thing. Namely the case when one of the
values is an uimm8, rather than when the condition is a uimm8 (which is
covered by bmnzi.b)
compare.ll:
- bsel.v tests now (correctly) emits bmnz.v instead of bsel.v because this
is the same operation (see MSA.txt).
i8.ll
- CHECK-DAG-ized test.
- bmzi.b test now (correctly) emits equivalent bmnzi.b with swapped operands
because this is the same operation (see MSA.txt).
- bseli.b still emits bseli.b though because the immediate makes it
distinguishable from bmnzi.b.
vec.ll:
- CHECK-DAG-ized test.
- bmz.v tests now (correctly) emits bmnz.v with swapped operands (see
MSA.txt).
- bsel.v tests now (correctly) emits bmnz.v with swapped operands (see
MSA.txt).
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@193693
91177308-0d34-0410-b5e6-
96231b3b80d8
Chad Rosier [Wed, 30 Oct 2013 15:19:37 +0000 (15:19 +0000)]
[AArch64] Add support for NEON scalar floating-point compare instructions.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@193691
91177308-0d34-0410-b5e6-
96231b3b80d8
Cameron McInally [Wed, 30 Oct 2013 15:19:10 +0000 (15:19 +0000)]
Refactor the AVX512 intrinsics. Cluster the intrinsics into the appropriate vector extension class within the .td file.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@193690
91177308-0d34-0410-b5e6-
96231b3b80d8
Howard Hinnant [Wed, 30 Oct 2013 15:10:54 +0000 (15:10 +0000)]
Rehash but don't grow when full of tombstones.
This problem was found and fixed by José Fonseca in March 2011 for
SmallPtrSet, committed r128566. But as far as I can tell, all other
llvm hash tables retain the same problem: the bucket count can grow
without bound while size() remains near constant by repeated
insert/erase cycles that tend to fill the container with tombstones.
Here is a demo that has been reduced to a trivial case:
int
main()
{
llvm::DenseSet<unsigned> d;
for (unsigned i = 0; i < 0xFFFFFFF; ++i)
{
d.insert(i);
d.erase(i);
}
}
While the container size() never grows above 1, the bucket count grows
like this:
nb = 64
nb = 128
nb = 256
nb = 512
nb = 1024
nb = 2048
nb = 4096
nb = 8192
nb = 16384
nb = 32768
nb = 65536
nb = 131072
nb = 262144
nb = 524288
nb =
1048576
nb =
2097152
nb =
4194304
nb =
8388608
nb =
16777216
nb =
33554432
nb =
67108864
nb =
134217728
nb =
268435456
The above program currently consumes a few GB ram. This patch brings
the memory consumption down by several orders of magnitude, and keeps
the bucket count at 64 for the above test.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@193689
91177308-0d34-0410-b5e6-
96231b3b80d8
Daniel Sanders [Wed, 30 Oct 2013 14:45:14 +0000 (14:45 +0000)]
[mips][msa] Added support for matching bins[lr]i.[bhwd] from normal IR (i.e. not intrinsics)
This required correcting the definition of the bins[lr]i intrinsics because
the result is also the first operand.
It also required removing the (arbitrary) check for 32-bit immediates in
MipsSEDAGToDAGISel::selectVSplat().
Currently using binsli.d with 2 bits set in the mask doesn't select binsli.d
because the constant is legalized into a ConstantPool. Similar things can
happen with binsri.d with more than 10 bits set in the mask. The resulting
code when this happens is correct but not optimal.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@193687
91177308-0d34-0410-b5e6-
96231b3b80d8
Daniel Sanders [Wed, 30 Oct 2013 13:51:01 +0000 (13:51 +0000)]
[mips][msa] Combine binsri-like DAG of AND and OR into equivalent VSELECT
(or (and $a, $mask), (and $b, $inverse_mask)) => (vselect $mask, $a, $b).
where $mask is a constant splat. This allows bitwise operations to make use
of bsel.
It's also a stepping stone towards matching bins[lr], and bins[lr]i from
normal IR.
Two sets of similar tests have been added in this commit. The bsel_* functions
test the case where binsri cannot be used. The binsr_*_i functions will
start to use the binsri instruction in the next commit.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@193682
91177308-0d34-0410-b5e6-
96231b3b80d8
Daniel Sanders [Wed, 30 Oct 2013 13:31:27 +0000 (13:31 +0000)]
[mips] MipsSETargetLowering now reports DAGCombiner changes when using -debug-only=mips-isel
No test since -debug output is intended for developers and not end-users.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@193681
91177308-0d34-0410-b5e6-
96231b3b80d8
Daniel Sanders [Wed, 30 Oct 2013 13:07:44 +0000 (13:07 +0000)]
[mips][msa] Added support for matching splat.[bhw] from normal IR (i.e. not intrinsics)
splat.d is implemented but this subtest is currently disabled. This is because
it is difficult to match the appropriate IR on MIPS32. There is a patch under
review that should help with this so I hope to enable the subtest soon.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@193680
91177308-0d34-0410-b5e6-
96231b3b80d8
Juergen Ributzka [Wed, 30 Oct 2013 06:36:19 +0000 (06:36 +0000)]
Revert "SelectionDAG: Teach the legalizer to split SETCC if VSELECT needs splitting too."
Now Hexagon and SystemZ are not happy with it :-(
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@193677
91177308-0d34-0410-b5e6-
96231b3b80d8
Juergen Ributzka [Wed, 30 Oct 2013 05:48:18 +0000 (05:48 +0000)]
SelectionDAG: Teach the legalizer to split SETCC if VSELECT needs splitting too.
The Type Legalizer recognizes that VSELECT needs to be split, because the type
is to wide for the given target. The same does not always apply to SETCC,
because less space is required to encode the result of a comparison. As a result
VSELECT is split and SETCC is unrolled into scalar comparisons.
This commit fixes the issue by checking for VSELECT-SETCC patterns in the DAG
Combiner. If a matching pattern is found, then the result mask of SETCC is
promoted to the expected vector mask type for the given target. This mask has
usually the same size as the VSELECT return type (except for Intel KNL). Now the
type legalizer will split both VSELECT and SETCC.
This allows the following X86 DAG Combine code to sucessfully detect the MIN/MAX
pattern. This fixes PR16695, PR17002, and <rdar://problem/
14594431>.
Reviewed by Nadav
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@193676
91177308-0d34-0410-b5e6-
96231b3b80d8