OSDN Git Service
Sanjay Patel [Sat, 5 Dec 2015 19:54:59 +0000 (19:54 +0000)]
fix typo; NFC
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254860
91177308-0d34-0410-b5e6-
96231b3b80d8
Dan Gohman [Sat, 5 Dec 2015 19:43:19 +0000 (19:43 +0000)]
[WebAssembly] Update a stale comment. NFC.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254859
91177308-0d34-0410-b5e6-
96231b3b80d8
JF Bastien [Sat, 5 Dec 2015 19:36:33 +0000 (19:36 +0000)]
WebAssembly: improve readme, add placeholder for tests.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254857
91177308-0d34-0410-b5e6-
96231b3b80d8
Dan Gohman [Sat, 5 Dec 2015 19:27:18 +0000 (19:27 +0000)]
[WebAssembly] Move useAA() out of line to make it more convenient to experiment with.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254856
91177308-0d34-0410-b5e6-
96231b3b80d8
Dan Gohman [Sat, 5 Dec 2015 19:24:17 +0000 (19:24 +0000)]
[WebAssembly] Call TargetPassConfig base class functions in overriding functions.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254855
91177308-0d34-0410-b5e6-
96231b3b80d8
Dan Gohman [Sat, 5 Dec 2015 19:15:57 +0000 (19:15 +0000)]
[WebAssembly] Expand frem as a floating point library function.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254854
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Sat, 5 Dec 2015 17:34:07 +0000 (17:34 +0000)]
[Hexagon] Don't call getNumImplicitDefs and then iterate over the count. getNumImplicitDefs contains a loop so its better to just loop over the null terminated implicit def list. NFC
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254852
91177308-0d34-0410-b5e6-
96231b3b80d8
Keno Fischer [Sat, 5 Dec 2015 14:42:34 +0000 (14:42 +0000)]
[ASAN] Add doFinalization to reset state
Summary: If the same pass manager is used for multiple modules ASAN
complains about GlobalsMD being initialized twice. Fix this by
resetting GlobalsMD in a new doFinalization method to allow this
use case.
Reviewers: kcc
Subscribers: llvm-commits
Differential Revision: http://reviews.llvm.org/D14962
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254851
91177308-0d34-0410-b5e6-
96231b3b80d8
Simon Pilgrim [Sat, 5 Dec 2015 07:27:50 +0000 (07:27 +0000)]
[X86][ADX] Added memory folding patterns and stack folding tests
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254844
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Sat, 5 Dec 2015 07:13:35 +0000 (07:13 +0000)]
Replace uint16_t with the MCPhysReg typedef in many places. A lot of physical register arrays already use this typedef.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254843
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Sat, 5 Dec 2015 07:13:25 +0000 (07:13 +0000)]
Use std::fill instead of memset to initialize an array to avoid hardcoded count and a multiply. The outputed code is identical.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254842
91177308-0d34-0410-b5e6-
96231b3b80d8
Simon Pilgrim [Sat, 5 Dec 2015 07:07:42 +0000 (07:07 +0000)]
[X86][FMA4] Explicitly set the domain of FMA4 float/double scalar instructions
Both were defaulting to the float domain - now matches the packed instructions.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254841
91177308-0d34-0410-b5e6-
96231b3b80d8
Xinliang David Li [Sat, 5 Dec 2015 05:16:36 +0000 (05:16 +0000)]
[PGO] Add version to getPGOFuncName method
Different version of indexed format may use different
name uniquing schemes for static functions. Pass the
version info to the name interface so that different
schmes can be picked (for profile lookup).
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254838
91177308-0d34-0410-b5e6-
96231b3b80d8
Cong Hou [Sat, 5 Dec 2015 05:00:55 +0000 (05:00 +0000)]
Normalize successors' probabilities when building MBBs for jump table.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254837
91177308-0d34-0410-b5e6-
96231b3b80d8
David Blaikie [Sat, 5 Dec 2015 03:41:53 +0000 (03:41 +0000)]
[llvm-dwp] Add coverage for both the presence and absence of type units, and fix/remove the emission of a broken tu_index when no type units are present
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254833
91177308-0d34-0410-b5e6-
96231b3b80d8
David Blaikie [Sat, 5 Dec 2015 03:11:17 +0000 (03:11 +0000)]
[llvm-dwp] Fix the type_units.test since I renamed its inputs as well
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254830
91177308-0d34-0410-b5e6-
96231b3b80d8
David Blaikie [Sat, 5 Dec 2015 03:10:05 +0000 (03:10 +0000)]
[llvm-dwp] Rename the sufficiently-modified test to reflect it's non-simplicity
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254829
91177308-0d34-0410-b5e6-
96231b3b80d8
David Blaikie [Sat, 5 Dec 2015 03:06:30 +0000 (03:06 +0000)]
[llvm-dwp] clang-format this to catch anything I've missed along the way
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254828
91177308-0d34-0410-b5e6-
96231b3b80d8
David Blaikie [Sat, 5 Dec 2015 03:05:45 +0000 (03:05 +0000)]
[llvm-dwp] Support debug_tu_index
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254827
91177308-0d34-0410-b5e6-
96231b3b80d8
Dan Gohman [Sat, 5 Dec 2015 03:03:35 +0000 (03:03 +0000)]
[WebAssembly] Implement ReverseBranchCondition, and re-enable MachineBlockPlacement
This patch introduces a codegen-only instruction currently named br_unless,
which makes it convenient to implement ReverseBranchCondition and re-enable
the MachineBlockPlacement pass. Then in a late pass, it lowers br_unless
back into br_if.
Differential Revision: http://reviews.llvm.org/D14995
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254826
91177308-0d34-0410-b5e6-
96231b3b80d8
Kostya Serebryany [Sat, 5 Dec 2015 02:23:49 +0000 (02:23 +0000)]
[libFuzzer] one more trophie
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254825
91177308-0d34-0410-b5e6-
96231b3b80d8
Kostya Serebryany [Sat, 5 Dec 2015 02:09:22 +0000 (02:09 +0000)]
[libFuzzer] don't reload the corpus more than once every second
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254824
91177308-0d34-0410-b5e6-
96231b3b80d8
Lang Hames [Sat, 5 Dec 2015 01:44:20 +0000 (01:44 +0000)]
Whitespace.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254821
91177308-0d34-0410-b5e6-
96231b3b80d8
Keno Fischer [Sat, 5 Dec 2015 01:38:12 +0000 (01:38 +0000)]
[opt] Fix run-twice option for non-idempotent passes
Cloning the module was supposed to guard against the possibility
that the passes may be non-idempotent. However, for some reason
I decided to put that AFTER the passes had already run on the
module, defeating the point entirely. Fix that by moving up the
CloneModule as is done in llc.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254819
91177308-0d34-0410-b5e6-
96231b3b80d8
Keno Fischer [Sat, 5 Dec 2015 01:02:53 +0000 (01:02 +0000)]
[MC] Add a test for state reset in MCMachOStreamer
This was fixed in r254751, but untestable until r254774, which
added the necessary command line flag to llc. Add a test now
to make sure this doesn't regress again.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254814
91177308-0d34-0410-b5e6-
96231b3b80d8
Cong Hou [Sat, 5 Dec 2015 01:00:22 +0000 (01:00 +0000)]
Fix a typo in LoopVectorize.cpp. NFC.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254813
91177308-0d34-0410-b5e6-
96231b3b80d8
Dan Gohman [Sat, 5 Dec 2015 00:51:40 +0000 (00:51 +0000)]
[WebAssembly] Fix scheduling dependencies in register-stackified code
Add physical register defs to instructions used from stackified
instructions to prevent them from being scheduled into the middle of
a stack sequence. This is a conservative measure which may be loosened
in the future.
Differential Revision: http://reviews.llvm.org/D15252
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254811
91177308-0d34-0410-b5e6-
96231b3b80d8
Justin Bogner [Sat, 5 Dec 2015 00:39:14 +0000 (00:39 +0000)]
CodeGen: Let the BumpPtrAllocator free the elements of indexList
The indexList's nodes are all allocated on a BumpPtrAllocator, so it's
more efficient to let them be freed when it goes away, rather than
deleting them directly. This is a follow up to r254794.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254808
91177308-0d34-0410-b5e6-
96231b3b80d8
Derek Schuff [Sat, 5 Dec 2015 00:26:39 +0000 (00:26 +0000)]
[WebAssembly] Support constant offsets on loads and stores
This is just prototype for load/store for i32 types. I'll add them to
the rest of the types if we like this direction.
Differential Revision: http://reviews.llvm.org/D15197
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254807
91177308-0d34-0410-b5e6-
96231b3b80d8
Philip Reames [Sat, 5 Dec 2015 00:18:33 +0000 (00:18 +0000)]
[EarlyCSE] IsSimple vs IsVolatile naming clarification (NFC)
When the notion of target specific memory intrinsics was introduced to EarlyCSE, the commit confused the notions of volatile and simple memory access. Since I'm about to start working on this area, cleanup the naming so that patches aren't horribly confusing. Note that the actual implementation was always bailing if the load or store wasn't simple.
Reminder:
- "volatile" - C++ volatile, can't remove any memory operations, but in principal unordered
- "ordered" - imposes ordering constraints on other nearby memory operations
- "atomic" - can't be split or sheared. In LLVM terms, all "ordered" operations are also atomic so the predicate "isAtomic" is often used.
- "simple" - a load which is none of the above. These are normal loads and what most of the optimizer works with.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254805
91177308-0d34-0410-b5e6-
96231b3b80d8
Keno Fischer [Sat, 5 Dec 2015 00:06:37 +0000 (00:06 +0000)]
[opt] Fix sanitizer complaints about r254774
`Out` can be null if no output is requested, so move any access
to it inside the conditional. Thanks to Justin Bogner for finding
this.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254804
91177308-0d34-0410-b5e6-
96231b3b80d8
Philip Reames [Fri, 4 Dec 2015 23:48:19 +0000 (23:48 +0000)]
[PassManager] Ensure destructors of cached AnalysisUsage objects are run
In 254760, I introduced the usage of a BumpPtrAllocator for the AnalysisUsage instances held by the PassManger. This turns out to have been incorrect since a BumpPtrAllocator does not run the destructors of objects when deallocating memory. Since a few of our SmallVector's had grown beyond their small size, we end up with some leaked memory. We need to use a SpecificBumpPtrAllocator instead.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254803
91177308-0d34-0410-b5e6-
96231b3b80d8
Teresa Johnson [Fri, 4 Dec 2015 23:40:22 +0000 (23:40 +0000)]
[ThinLTO] Helper for performing renaming/promotion on a module
Creates a module and performs necessary renaming/promotion of locals
that may be exported to another module.
Split out of D15024.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254802
91177308-0d34-0410-b5e6-
96231b3b80d8
Hans Wennborg [Fri, 4 Dec 2015 23:32:19 +0000 (23:32 +0000)]
Add FeatureLAHFSAHF to amdfam10 as well.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254801
91177308-0d34-0410-b5e6-
96231b3b80d8
Dan Gohman [Fri, 4 Dec 2015 23:22:35 +0000 (23:22 +0000)]
[WebAssembly] Initial varargs support.
Full varargs support will depend on prologue/epilogue support, but this patch
gets us started with most of the basic infrastructure.
Differential Revision: http://reviews.llvm.org/D15231
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254799
91177308-0d34-0410-b5e6-
96231b3b80d8
Philip Reames [Fri, 4 Dec 2015 23:06:33 +0000 (23:06 +0000)]
Address a memory leak in 254760
The issue appears to have been that the copy constructor of the SmallVector was being invoked and this was somehow leading to leaked memory. This patch avoids the symptom, but likely doesn't address the underlying problem. I'm still investigating the root cause, but wanted to avoid the memory leak in the mean time. Even with the underlying fix, avoiding the redundant allocation is worthwhile.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254795
91177308-0d34-0410-b5e6-
96231b3b80d8
Justin Bogner [Fri, 4 Dec 2015 23:00:54 +0000 (23:00 +0000)]
CodeGen: Move the SlotIndexes BumpPtrAllocator before the list it allocates
When a `SlotIndexes` is destroyed, `ileAllocator` will currently be
destructed before `IndexList`, but all of `IndexList`'s storage has
been allocated by `ileAllocator`. This means we'll call destructors on
garbage data, which is very bad. This can be avoided by putting the
BumpPtrAllocator earlier in the class than anything it allocates.
Unfortunately, I don't know how to test this. It depends very much on
memory layout, and the only evidence I have that this is actually
happening in practice are backtraces that might be explained by this.
By inspection though, the code is obviously dangerous/wrong, and this
is the right thing to do.
I'll follow up later with a patch that calls clearAndLeakNodesUnsafely
on the list, since there isn't much point in destructing them when
they're allocated in a BPA anyway, but I figured it makes sense to
commit the correctness fix separately from that optimization.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254794
91177308-0d34-0410-b5e6-
96231b3b80d8
Hans Wennborg [Fri, 4 Dec 2015 23:00:33 +0000 (23:00 +0000)]
X86: Don't emit SAHF/LAHF for 64-bit targets unless explicitly supported
These instructions are not supported by all CPUs in 64-bit mode. Emitting them
causes Chromium to crash on start-up for users with such chips.
(GCC puts these instructions behind -msahf on 64-bit for the same reason.)
This patch adds FeatureLAHFSAHF, enables it by default for 32-bit targets
and modern CPUs, and changes X86InstrInfo::copyPhysReg back to the lowering
from before r244503 when the instructions are not available.
Differential Revision: http://reviews.llvm.org/D15240
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254793
91177308-0d34-0410-b5e6-
96231b3b80d8
Derek Schuff [Fri, 4 Dec 2015 22:47:58 +0000 (22:47 +0000)]
Add TransformUtils to list of required libraries for llc
This dependency was added in r254774
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254786
91177308-0d34-0410-b5e6-
96231b3b80d8
Kostya Serebryany [Fri, 4 Dec 2015 22:29:39 +0000 (22:29 +0000)]
[libFuzzer] compute base64 in-process instead of using an external lib. Since libFuzzer should not depend on anything, just re-implement base64 encoder. PR25746
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254784
91177308-0d34-0410-b5e6-
96231b3b80d8
Rafael Espindola [Fri, 4 Dec 2015 22:26:21 +0000 (22:26 +0000)]
MSVC complains about this being ambiguous.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254782
91177308-0d34-0410-b5e6-
96231b3b80d8
Lang Hames [Fri, 4 Dec 2015 22:09:19 +0000 (22:09 +0000)]
[Orc] Move some code up into the JITCompileCallbackManager base class. NFC.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254778
91177308-0d34-0410-b5e6-
96231b3b80d8
Rafael Espindola [Fri, 4 Dec 2015 22:08:53 +0000 (22:08 +0000)]
Always pass a diagnostic handler to the linker.
Before this patch the diagnostic handler was optional. If it was not
passed, the one in the LLVMContext was used.
That is probably not a pattern we want to follow. If each area has an
optional callback, there is a sea of callbacks and it is hard to follow
which one is called.
Doing this also found cases where the callback is a nice addition, like
testing that no errors or warnings are reported.
The other option is to always use the diagnostic handler in the
LLVMContext. That has a few problems
* To implement the C API we would have to set the diag handler and then
set it back to the original value.
* Code that creates the context might be far away from code that wants
the diagnostics.
I do have a patch that implements the second option and will send that as
an RFC.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254777
91177308-0d34-0410-b5e6-
96231b3b80d8
Weiming Zhao [Fri, 4 Dec 2015 22:00:47 +0000 (22:00 +0000)]
[SimplifyLibCalls] Optimization for pow(x, n) where n is some constant
Summary:
In order to avoid calling pow function we generate repeated fmul when n is a
positive or negative whole number.
For each exponent we pre-compute Addition Chains in order to minimize the no.
of fmuls.
Refer: http://wwwhomes.uni-bielefeld.de/achim/addition_chain.html
We pre-compute addition chains for exponents upto 32 (which results in a max of
7 fmuls).
For eg:
4 = 2+2
5 = 2+3
6 = 3+3 and so on
Hence,
pow(x, 4.0) ==> y = fmul x, x
x = fmul y, y
ret x
For negative exponents, we simply compute the reciprocal of the final result.
Note: This transformation is only enabled under fast-math.
Patch by Mandeep Singh Grang <mgrang@codeaurora.org>
Reviewers: weimingz, majnemer, escha, davide, scanon, joerg
Subscribers: probinson, escha, llvm-commits
Differential Revision: http://reviews.llvm.org/D13994
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254776
91177308-0d34-0410-b5e6-
96231b3b80d8
Pete Cooper [Fri, 4 Dec 2015 21:59:04 +0000 (21:59 +0000)]
Fix incorrect quote. NFC
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254775
91177308-0d34-0410-b5e6-
96231b3b80d8
Keno Fischer [Fri, 4 Dec 2015 21:56:46 +0000 (21:56 +0000)]
[llc/opt] Add an option to run all passes twice
Summary: Lately, I have submitted a number of patches to fix bugs that
only occurred when using the same pass manager to compile multiple
modules (generally these bugs are failure to reset some persistent
state). Unfortunately I don't think there is currently a way to test
that from the command line. This adds a very simple flag to both llc
and opt, under which the tools will simply re-run their respective
pass pipelines using the same pass manager on (a clone of the same
module). Additionally, we verify that both outputs are bitwise the
same.
Reviewers: yaron.keren
Subscribers: loladiro, yaron.keren, kcc, llvm-commits
Differential Revision: http://reviews.llvm.org/D14965
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254774
91177308-0d34-0410-b5e6-
96231b3b80d8
Chad Rosier [Fri, 4 Dec 2015 21:38:44 +0000 (21:38 +0000)]
[AArch64] Expand vector SDIVREM/UDIVREM operations.
http://reviews.llvm.org/D15214
Patch by Ana Pazos <apazos@codeaurora.org>!
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254773
91177308-0d34-0410-b5e6-
96231b3b80d8
David Blaikie [Fri, 4 Dec 2015 21:38:39 +0000 (21:38 +0000)]
[llvm-dwp] Remove some out of date comments
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254772
91177308-0d34-0410-b5e6-
96231b3b80d8
David Blaikie [Fri, 4 Dec 2015 21:30:23 +0000 (21:30 +0000)]
[llvm-dwp] Implement the required on-disk probed hash table
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254770
91177308-0d34-0410-b5e6-
96231b3b80d8
Reid Kleckner [Fri, 4 Dec 2015 21:29:53 +0000 (21:29 +0000)]
Fix llvm-readobj build on Windows, match noreturn attribute on reportError in headers
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254769
91177308-0d34-0410-b5e6-
96231b3b80d8
David Blaikie [Fri, 4 Dec 2015 21:16:42 +0000 (21:16 +0000)]
[llvm-dwp] Include the debug_line.dwo section
This probably shouldn't be generated in the .dwo file for CUs, only for
TUs, but it's in the sample .dwos (generated by clang) so dwp should
reflect that.
Arguably the DWP tool could be smart enough to know that the CUs
shouldn't need a debug_line.dwo section and skip that even when it's
legitimately generated for TUs, but that's a bit more off-book.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254767
91177308-0d34-0410-b5e6-
96231b3b80d8
Sanjoy Das [Fri, 4 Dec 2015 20:34:37 +0000 (20:34 +0000)]
[OperandBundles] Allow operand-specific attributes in operand bundles
Currently `OperandBundleUse::operandsHaveAttr` computes its result
without being given a specific operand. This is problematic because it
forces us to say that, e.g., even non-pointer operands in `"deopt"`
operand bundles are `readonly`, which doesn't make sense.
This commit changes `operandsHaveAttr` to work in the context of a
specific operand, so that we can give the operand attributes that make
sense for the operands's `llvm::Type`.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254764
91177308-0d34-0410-b5e6-
96231b3b80d8
Philip Reames [Fri, 4 Dec 2015 20:05:04 +0000 (20:05 +0000)]
[LegacyPassManager] Reduce memory usage for AnalysisUsage
The LegacyPassManager was storing an instance of AnalysisUsage for each instance of each pass. In practice, most instances of a single pass class share the same dependencies. We can't rely on this because passes can (and some do) have dynamic dependencies based on instance options.
We can exploit the likely commonality by uniqueing the usage information after querying the pass, but before storing it into the pass manager. This greatly reduces memory consumption by the AnalysisUsage objects. For a long pass pipeline, I measured a decrease in memory consumption for this storage of about 50%. I have not measured on the default O3 pipeline, but I suspect it will see some benefit as well since many passes are repeated (e.g. InstCombine).
Differential Revision: http://reviews.llvm.org/D14677
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254760
91177308-0d34-0410-b5e6-
96231b3b80d8
Matthias Braun [Fri, 4 Dec 2015 19:54:24 +0000 (19:54 +0000)]
ScheduleDAGInstrs: Move LiveIntervals field to ScheduleDAGMI
Now that ScheduleDAGInstrs doesn't need it anymore we can move the field
down the class hierarcy to ScheduleDAGMI.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254759
91177308-0d34-0410-b5e6-
96231b3b80d8
Davide Italiano [Fri, 4 Dec 2015 19:29:49 +0000 (19:29 +0000)]
[llvm-readobj] reportError() never returns. Mark with the correct attribute.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254752
91177308-0d34-0410-b5e6-
96231b3b80d8
Davide Italiano [Fri, 4 Dec 2015 19:27:58 +0000 (19:27 +0000)]
[llvm-readobj/ELF] Simplify Verdef handling.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254751
91177308-0d34-0410-b5e6-
96231b3b80d8
Mike Aizatsky [Fri, 4 Dec 2015 19:11:54 +0000 (19:11 +0000)]
fixing Makefile
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254749
91177308-0d34-0410-b5e6-
96231b3b80d8
Mike Aizatsky [Fri, 4 Dec 2015 18:50:18 +0000 (18:50 +0000)]
adding MC dependencies in hopes to pacify the hexagon build.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254745
91177308-0d34-0410-b5e6-
96231b3b80d8
Mike Aizatsky [Fri, 4 Dec 2015 18:35:37 +0000 (18:35 +0000)]
sancov -not-covered-functions.
Summary: The command prints out list of functions that were not entered.
To do this, addresses are first converted to function locations. Set
operations are used for function locations.
Differential Revision: http://reviews.llvm.org/D14889
review
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254742
91177308-0d34-0410-b5e6-
96231b3b80d8
Dan Gohman [Fri, 4 Dec 2015 18:27:03 +0000 (18:27 +0000)]
[WebAssembly] Add several more calling conventions to the supported list.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254741
91177308-0d34-0410-b5e6-
96231b3b80d8
Sanjay Patel [Fri, 4 Dec 2015 17:54:31 +0000 (17:54 +0000)]
don't repeat function names in comments; NFC
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254740
91177308-0d34-0410-b5e6-
96231b3b80d8
Sanjay Patel [Fri, 4 Dec 2015 17:51:55 +0000 (17:51 +0000)]
fix formatting; NFC
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254739
91177308-0d34-0410-b5e6-
96231b3b80d8
Manman Ren [Fri, 4 Dec 2015 17:40:13 +0000 (17:40 +0000)]
[CXX TLS calling convention] Add CXX TLS calling convention.
This commit adds a new target-independent calling convention for C++ TLS
access functions. It aims to minimize overhead in the caller by perserving as
many registers as possible.
The target-specific implementation for X86-64 is defined as following:
Arguments are passed as for the default C calling convention
The same applies for the return value(s)
The callee preserves all GPRs - except RAX and RDI
The access function makes C-style TLS function calls in the entry and exit
block, C-style TLS functions save a lot more registers than normal calls.
The added calling convention ties into the existing implementation of the
C-style TLS functions, so we can't simply use existing calling conventions
such as preserve_mostcc.
rdar://
9001553
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254737
91177308-0d34-0410-b5e6-
96231b3b80d8
David Blaikie [Fri, 4 Dec 2015 17:20:04 +0000 (17:20 +0000)]
[llvm-dwp] Retrieve the DWOID from the CU for the cu_index entry
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254731
91177308-0d34-0410-b5e6-
96231b3b80d8
Dan Gohman [Fri, 4 Dec 2015 17:19:44 +0000 (17:19 +0000)]
[WebAssembly] Give names to the callseq begin and end instructions.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254730
91177308-0d34-0410-b5e6-
96231b3b80d8
Dan Gohman [Fri, 4 Dec 2015 17:18:32 +0000 (17:18 +0000)]
[WebAssembly] clang-format CallingConvSupported. NFC.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254729
91177308-0d34-0410-b5e6-
96231b3b80d8
Dan Gohman [Fri, 4 Dec 2015 17:16:07 +0000 (17:16 +0000)]
[WebAssembly] Factor out the list of supported calling conventions.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254728
91177308-0d34-0410-b5e6-
96231b3b80d8
Dan Gohman [Fri, 4 Dec 2015 17:12:52 +0000 (17:12 +0000)]
[WebAssembly] Check for more unsupported ABI flags.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254727
91177308-0d34-0410-b5e6-
96231b3b80d8
Dan Gohman [Fri, 4 Dec 2015 17:09:42 +0000 (17:09 +0000)]
[WebAssembly] Use SelectionDAG::getUNDEF. NFC.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254726
91177308-0d34-0410-b5e6-
96231b3b80d8
Krzysztof Parzyszek [Fri, 4 Dec 2015 16:18:15 +0000 (16:18 +0000)]
[Hexagon] Simplify LowerCONCAT_VECTORS, handle different types better
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254724
91177308-0d34-0410-b5e6-
96231b3b80d8
Rafael Espindola [Fri, 4 Dec 2015 16:14:31 +0000 (16:14 +0000)]
Modernize the C++ APIs for creating LTO modules.
This is a continuation of r253367.
These functions return is owned by the caller, so they return
std::unique_ptr now.
The call can fail, so the return is wrapped in ErrorOr.
They have a context where to report diagnostics, so they don't need to
take a string out parameter.
With this there are no call to getGlobalContext in lib/LTO.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254721
91177308-0d34-0410-b5e6-
96231b3b80d8
Tim Northover [Fri, 4 Dec 2015 16:10:48 +0000 (16:10 +0000)]
ARM/AArch64: update reference documentation.
There's a more comprehensive ACLE and a real v8 ARM ARM now.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254720
91177308-0d34-0410-b5e6-
96231b3b80d8
Colin LeMahieu [Fri, 4 Dec 2015 15:48:45 +0000 (15:48 +0000)]
[Hexagon] Using multiply instead of shift on signed number which can be UB
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254719
91177308-0d34-0410-b5e6-
96231b3b80d8
Jonas Paulsson [Fri, 4 Dec 2015 12:48:51 +0000 (12:48 +0000)]
[SystemZ] Bugfix: Don't add CC twice to new three-address instruction.
Since BuildMI() automatically adds the implicit operands for a new instruction,
adding the old instructions CC operand resulted in that there were two CC imp-def
operands, where only one was marked as dead. This caused buildSchedGraph() to
miss dependencies on the CC reg.
Review by Ulrich Weigand
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254714
91177308-0d34-0410-b5e6-
96231b3b80d8
Alexey Bataev [Fri, 4 Dec 2015 10:53:15 +0000 (10:53 +0000)]
LEA code size optimization pass (Part 1): Remove redundant address recalculations, by Andrey Turetsky
Add new x86 pass which replaces address calculations in load or store instructions with def register of existing LEA (must be in the same basic block), if the LEA calculates address that differs only by a displacement. Works only with -Os or -Oz.
Differential Revision: http://reviews.llvm.org/D13294
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254712
91177308-0d34-0410-b5e6-
96231b3b80d8
Oliver Stannard [Fri, 4 Dec 2015 09:45:18 +0000 (09:45 +0000)]
[AArch64] Clean up statistical profiling test
This check has nothing to do with the statistical profiling extension, so
shouldn't be in this test.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254709
91177308-0d34-0410-b5e6-
96231b3b80d8
Yury Gribov [Fri, 4 Dec 2015 09:19:14 +0000 (09:19 +0000)]
[asan] Fix dynamic allocas unpoisoning on PowerPC64.
For PowerPC64 we cannot just pass SP extracted from @llvm.stackrestore to
_asan_allocas_unpoison due to specific ABI requirements
(http://refspecs.linuxfoundation.org/ELF/ppc64/PPC-elf64abi.html#DYNAM-STACK).
This patch adds the value returned by @llvm.get.dynamic.area.offset to
extracted from @llvm.stackrestore stack pointer, so dynamic allocas unpoisoning
stuff would work correctly on PowerPC64.
Patch by Max Ostapenko.
Differential Revision: http://reviews.llvm.org/D15108
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254707
91177308-0d34-0410-b5e6-
96231b3b80d8
Rafael Espindola [Fri, 4 Dec 2015 04:15:05 +0000 (04:15 +0000)]
Revert "[BranchFolding] Merge MMOs during tail merge"
This reverts commit r254694.
It broke bootstrap.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254700
91177308-0d34-0410-b5e6-
96231b3b80d8
Rafael Espindola [Fri, 4 Dec 2015 02:42:28 +0000 (02:42 +0000)]
Move a call to getGlobalContext out of lib/LTO.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254696
91177308-0d34-0410-b5e6-
96231b3b80d8
Lang Hames [Fri, 4 Dec 2015 02:32:32 +0000 (02:32 +0000)]
[Orc] Fix Kaleidoscope example for change in r254693.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254695
91177308-0d34-0410-b5e6-
96231b3b80d8
Junmo Park [Fri, 4 Dec 2015 02:29:25 +0000 (02:29 +0000)]
[BranchFolding] Merge MMOs during tail merge
Summary:
If we remove the MMOs from Load/Store instructions,
they are treated as volatile. This makes other optimization passes unhappy.
eg. Load/Store Optimization
So, it looks better to merge, not remove.
Reviewers: gberry, mcrosier
Subscribers: gberry, llvm-commits
Differential Revision: http://reviews.llvm.org/D14797
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254694
91177308-0d34-0410-b5e6-
96231b3b80d8
Lang Hames [Fri, 4 Dec 2015 02:15:39 +0000 (02:15 +0000)]
[Orc] Rename JITCompileCallbackManagerBase to JITCompileCallbackManager.
This class is turning into a useful interface, rather than an implementation
detail, so I'm dropping the 'Base' suffix.
No functional change.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254693
91177308-0d34-0410-b5e6-
96231b3b80d8
Justin Bogner [Fri, 4 Dec 2015 02:14:34 +0000 (02:14 +0000)]
IR: Use format_hex instead of handrolling the conversion. NFC
Cleans up some very old code in AsmWriter's WriteConstantInternal.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254688
91177308-0d34-0410-b5e6-
96231b3b80d8
Nathan Slingerland [Fri, 4 Dec 2015 02:13:58 +0000 (02:13 +0000)]
Revert "[llvm-profdata] Add support for weighted merge of profile data"
This reverts commit
b7250858d96b8ce567681214273ac0e62713c661.
Reverting in order to investigate Windows test failure.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254687
91177308-0d34-0410-b5e6-
96231b3b80d8
Junmo Park [Fri, 4 Dec 2015 02:06:59 +0000 (02:06 +0000)]
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254686
91177308-0d34-0410-b5e6-
96231b3b80d8
NAKAMURA Takumi [Fri, 4 Dec 2015 02:00:12 +0000 (02:00 +0000)]
Move llvm/test/CodeGen/Generic/function-alias.ll to X86. It is incompatible to PECOFF.
FIXME: It may be ELF-generic.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254685
91177308-0d34-0410-b5e6-
96231b3b80d8
Quentin Colombet [Fri, 4 Dec 2015 01:53:14 +0000 (01:53 +0000)]
[ARM] When a bitcast is about to be turned into a VMOVDRR, try to combine it
with its source instead of forcing the values on GPRs.
This improves the lowering of vector code when such bitcasts happen in the
middle of vector computations.
rdar://problem/
23691584
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254684
91177308-0d34-0410-b5e6-
96231b3b80d8
Matthias Braun [Fri, 4 Dec 2015 01:51:19 +0000 (01:51 +0000)]
ScheduleDAGInstrs: Rework schedule graph builder.
Re-comitting with a change that avoids undefined uses getting put into
the VRegUses list.
The new algorithm remembers the uses encountered while walking backwards
until a matching def is found. Contrary to the previous version this:
- Works without LiveIntervals being available
- Allows to increase the precision to subregisters/lanemasks
(not used for now)
The changes in the AMDGPU tests are necessary because the R600 scheduler
is not stable with respect to the order of nodes in the ready queues.
Differential Revision: http://reviews.llvm.org/D9068
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254683
91177308-0d34-0410-b5e6-
96231b3b80d8
Matthias Braun [Fri, 4 Dec 2015 01:31:59 +0000 (01:31 +0000)]
raw_ostream: << operator for callables with raw_ostream argument
This is a revised version of r254655 which uses a Printable wrapper
class to avoid ambiguous overload problems.
Differential Revision: http://reviews.llvm.org/D14348
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254681
91177308-0d34-0410-b5e6-
96231b3b80d8
JF Bastien [Fri, 4 Dec 2015 01:18:17 +0000 (01:18 +0000)]
X86InstrInfo::copyPhysReg: workaround reg liveness
Summary:
computeRegisterLiveness and analyzePhysReg are currently getting
confused about liveness in some cases, breaking copyPhysReg's
calculation of whether AX is dead in some cases. Work around this issue
temporarily by assuming that AX is always live.
See detail in: https://llvm.org/bugs/show_bug.cgi?id=25033#c7
And associated bugs PR24535 PR25033 PR24991 PR24992 PR25201.
This workaround makes the code correct but slightly inefficient, but it
seems to confuse the machine instr verifier which now things EAX was
undefined in some cases where it's being conservatively saved /
restored.
Reviewers: majnemer, sanjoy
Subscribers: llvm-commits
Differential Revision: http://reviews.llvm.org/D15198
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254680
91177308-0d34-0410-b5e6-
96231b3b80d8
Justin Bogner [Fri, 4 Dec 2015 01:14:24 +0000 (01:14 +0000)]
IR: Update a comment and a bool that've been out of date since 2012
It became impossible to get here with a half in r157393, over 3 years
ago.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254679
91177308-0d34-0410-b5e6-
96231b3b80d8
Xinliang David Li [Fri, 4 Dec 2015 01:02:10 +0000 (01:02 +0000)]
[PGO] Unify VP data format between raw and indexed profile (Reader)
With the latest refactoring and code sharing patches landed,
it is possible to unify the value profile implementation between
raw and indexed profile. This is the patch in raw profile reader
that uses the common interface.
Differential Revision: http://reviews.llvm.org/D15056
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254677
91177308-0d34-0410-b5e6-
96231b3b80d8
Evgeniy Stepanov [Fri, 4 Dec 2015 00:57:25 +0000 (00:57 +0000)]
Fix function-alias.ll test on non-X86 targets.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254676
91177308-0d34-0410-b5e6-
96231b3b80d8
Rafael Espindola [Fri, 4 Dec 2015 00:45:57 +0000 (00:45 +0000)]
Simplify the error handling in llvm-lto a bit.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254675
91177308-0d34-0410-b5e6-
96231b3b80d8
Evgeniy Stepanov [Fri, 4 Dec 2015 00:45:43 +0000 (00:45 +0000)]
Emit function alias to data as a function symbol.
CFI emits jump slots for indirect functions as a byte array
constant, and declares function-typed aliases to these constants.
This change fixes AsmPrinter to emit these aliases as function
symbols and not data symbols.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254674
91177308-0d34-0410-b5e6-
96231b3b80d8
Cong Hou [Fri, 4 Dec 2015 00:36:58 +0000 (00:36 +0000)]
Don't punish vectorized arithmetic instruction whose type will be split to multiple registers
Currently in LLVM's cost model, a vectorized arithmetic instruction will have
high cost if its type is split into multiple registers. However, this
punishment is too heavy and unnecessary. The overhead of the split should not
be on arithmetic instructions but instructions that implement the split. Note
that during vectorization we have calculated the register pressure, and we
only choose proper interleaving factor (and also vectorization factor) so
that we don't use more registers than the maximum number.
Here is a very simple example: if a vadd has the cost 1, and if we double VF
so that we need two registers to perform it, then its cost will become 4 with
the current implementation, which will prevent us to use larger VF.
Differential revision: http://reviews.llvm.org/D15159
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254671
91177308-0d34-0410-b5e6-
96231b3b80d8
Nathan Slingerland [Fri, 4 Dec 2015 00:00:20 +0000 (00:00 +0000)]
[llvm-profdata] Add support for weighted merge of profile data
This change adds support for an optional weight when merging profile data with the llvm-profdata tool.
Weights are specified by adding an option ':<weight>' suffix to the input file names.
Adding support for arbitrary weighting of input profile data allows for relative importance to be placed on the
input data from multiple training runs.
Both sampled and instrumented profiles are supported.
Reviewers: dnovillo, bogner, davidxl
Subscribers: llvm-commits
Differential Revision: http://reviews.llvm.org/D14547
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254669
91177308-0d34-0410-b5e6-
96231b3b80d8
Kevin B. Smith [Fri, 4 Dec 2015 00:00:10 +0000 (00:00 +0000)]
[CodeGen] Minor correction to comment on PhysRegInfo.
Differential revision: http://reviews.llvm.org/D15216
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254668
91177308-0d34-0410-b5e6-
96231b3b80d8
Rafael Espindola [Thu, 3 Dec 2015 23:56:42 +0000 (23:56 +0000)]
Simplify since this function never fails.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254667
91177308-0d34-0410-b5e6-
96231b3b80d8
JF Bastien [Thu, 3 Dec 2015 23:43:56 +0000 (23:43 +0000)]
CodeGen peephole: fold redundant phys reg copies
Code generation often exposes redundant physical register copies through
virtual registers such as:
%vreg = COPY %PHYSREG
...
%PHYSREG = COPY %vreg
There are cases where no intervening clobber of %PHYSREG occurs, and the
later copy could therefore be removed. In some cases this further allows
us to remove the initial copy.
This patch contains a motivating example which comes from the x86 build
of Chrome, specifically cc::ResourceProvider::UnlockForRead uses
libstdc++'s implementation of hash_map. That example has two tests live
at the same time, and after machine sinking LLVM has confused itself
enough and things spilling EFLAGS is a great idea even though it's
never restored and the comparison results are both live.
Before this patch we have:
DEC32m %RIP, 1, %noreg, <ga:@L>, %noreg, %EFLAGS<imp-def>
%vreg1<def> = COPY %EFLAGS; GR64:%vreg1
%EFLAGS<def> = COPY %vreg1; GR64:%vreg1
JNE_1 <BB#1>, %EFLAGS<imp-use>
Both copies are useless. This patch tries to eliminate the later copy in
a generic manner.
dec is especially confusing to LLVM when compared with sub.
I wrote this patch to treat all physical registers generically, but only
remove redundant copies of non-allocatable physical registers because
the allocatable ones caused issues (e.g. when calling conventions weren't
properly modeled) and should be handled later by the register allocator
anyways.
The following tests used to failed when the patch also replaced allocatable
registers:
CodeGen/X86/StackColoring.ll
CodeGen/X86/avx512-calling-conv.ll
CodeGen/X86/copy-propagation.ll
CodeGen/X86/inline-asm-fpstack.ll
CodeGen/X86/musttail-varargs.ll
CodeGen/X86/pop-stack-cleanup.ll
CodeGen/X86/preserve_mostcc64.ll
CodeGen/X86/tailcallstack64.ll
CodeGen/X86/this-return-64.ll
This happens because COPY has other special meaning for e.g. dependency
breakage and x87 FP stack.
Note that all other backends' tests pass.
Reviewers: qcolombet
Subscribers: llvm-commits
Differential Revision: http://reviews.llvm.org/D15157
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254665
91177308-0d34-0410-b5e6-
96231b3b80d8