AsmPrinter: Document why DIEValueList uses a linked-list, NFC
There are two main reasons why a linked-list makes sense for
`DIEValueList`.
1. We want `DIE` to be on a `BumpPtrAllocator` to improve teardown
efficiency. Making `DIEValueList` array-based would make that much
more complicated.
2. The singly-linked list is fairly memory efficient. The histogram
[1] shows that most DIEs have relatively few values, so we often pay
less than the 2/3-pointer static overhead of a vector. Furthermore,
we don't know ahead of time exactly how many values a `DIE` needs,
so a vector-like scheme will on average over-allocate by ~50%. As
it happens, that's the same memory overhead as the linked list node.
Allow callers of `Value::print()` and `Metadata::print()` to pass in a
`ModuleSlotTracker`. This allows them to pay only once for calculating
module-level slots (such as Metadata).
This is related to PR23865, where there was a huge cost for
`MachineFunction::print()`. Although I don't have a *particular* user
in mind for this new code, I have hit big slowdowns before when running
`opt -debug`, and I think this will be useful. Going forward, if
someone hits a big slowdown with `print()` statements, they can create a
`ModuleSlotTracker` and send it through. Similarly, adding support to
`Value::dump()` and `Metadata::dump()` should be trivial.
I added unit tests to be sure the `print()` functions actually behave
the same way with and without the slot tracker.
LowerBitSets: Ignore bitset entries that do not directly refer to a global.
It is possible for a global to be substituted with another global of a
different type or a different kind (i.e. an alias) at IR link time. One
example of this scenario is when a Microsoft ABI vtable is substituted with
an alias referring to a larger vtable containing an RTTI reference.
This will cause the global to be RAUW'd with a possibly bitcasted reference
to the other global. This will of course also affect any references to the
global in bitset metadata.
The right way to handle such metadata is simply to ignore it. This is sound
because the linked module should contain another copy of the bitset entries as
applied to the new global.
Lang Hames [Fri, 26 Jun 2015 23:56:53 +0000 (23:56 +0000)]
[StackMaps] Add a lightweight parser for stackmap version 1 sections.
The parser provides a convenient interface for reading llvm stackmap v1 sections
in object files.
This patch also includes a new option for llvm-readobj, '-stackmap', which uses
the parser to pretty-print stackmap sections for debugging/testing purposes.
CodeGen: Create a proper ModuleSlotTracker for MachineInstr
Another follow-up related to r240848: try a little harder to share slot
tracking calculations within a single `MachineInstr` dump. This is
unrelated to `MachineFunction::print()`, since that should be passing
through the function's `ModuleSlotTracker` by now, but could affect the
speed of dumping from a debugger if there is more than one IR-level
operand.
Alex Lorenz [Fri, 26 Jun 2015 22:56:48 +0000 (22:56 +0000)]
MIR Serialization: Serialize global address machine operands.
This commit serializes the global address machine operands.
This commit doesn't serialize the operand's offset and target
flags, it serializes only the global value reference.
Philip Reames [Fri, 26 Jun 2015 22:47:37 +0000 (22:47 +0000)]
[RewriteStatepointsForGC] Generalized vector phi/select handling for base pointers
This change extends the detection of base pointers for vector constructs to handle arbitrary phi and select nodes. The existing non-vector code already handles those, so this is basically just extending the vector special case to be less special cased. It still isn't generalized vector handling since we can't handle arbitrary vector instructions (e.g. shufflevectors), but it's a lot closer.
The general structure of the change is as follows:
* Extend the base defining value relation over a subset of vector instructions and vector typed phi & select instructions.
* Move scalarization from before base pointer rewriting to after base pointer rewriting. The extension of the BDV relation is sufficient to find vector base phis for vector inputs.
* Preserve the existing special case logic for when the base of a vector element is locally obvious. This general idea could be extended to the scalar case as well.
CodeGen: Push the ModuleSlotTracker through MachineOperands
Push `ModuleSlotTracker` through `MachineOperand`s, dropping the time
for `llc -print-machineinstrs` on the testcase in PR23865 from ~13
seconds to ~9 seconds. Now `SlotTracker::processFunctionMetadata()`
accounts for only 8% of the runtime, which seems reasonable.
CodeGen: Use a single SlotTracker in MachineFunction::print()
Expose enough of the IR-level `SlotTracker` so that
`MachineFunction::print()` can use a single one for printing
`BasicBlock`s. Next step would be to lift this through a few more APIs
so that we can make other print methods faster.
Fixes PR23865, changing the runtime of `llc -print-machineinstrs` from
many minutes (killed after 3 minutes, but it wasn't very close) to
13 seconds for a 502185 line dump.
Philip Reames [Fri, 26 Jun 2015 21:39:44 +0000 (21:39 +0000)]
[Verifier] Verify invokes of intrinsics
We support invoking a subset of llvm's intrinsics, but the verifier didn't account for this. We had previously added a special case to verify invokes of statepoints. By generalizing the code in terms of CallSite, we can verify invokes of other intrinsics as well. Interestingly, this found one test case which was invalid.
Note: I'm deliberately leaving the naming change from CI to CS to a follow up change. That will happen shortly, I just wanted to reduce the diff to make it clear what was happening with this one.
Philip Reames [Fri, 26 Jun 2015 20:51:17 +0000 (20:51 +0000)]
Teach InlineCost to account for a null check which can be folded away
If we have a caller that knows a particular argument can never be null, we can exploit this fact while simplifying values in the inline cost analysis. This has the effect of reducing the cost for inlining when a null check is present in the callee, but the value is known non null in the caller. In particular, any dependent control flow can be discounted from the cost estimate.
Note that we use the parameter attributes at the call site to memoize the analysis within the caller's code. The setting of this attribute is done in InstCombine, the inline cost analysis just consumes it. This is intentional and important because we want the inline cost analysis results to be easily cachable themselves. We're not currently doing so, but initial results on LTO indicate this will quickly become important.
Marek Olsak [Fri, 26 Jun 2015 20:29:10 +0000 (20:29 +0000)]
AMDGPU: really don't commute REV opcodes if the target variant doesn't exist
If pseudoToMCOpcode failed, we would return the original opcode, so operands
would be swapped, but the instruction would remain the same.
It resulted in LSHLREV a, b ---> LSHLREV b, a.
This fixes Glamor text rendering and
piglit/arb_sample_shading-builtin-gl-sample-mask on VI.
Nemanja Ivanovic [Fri, 26 Jun 2015 19:26:53 +0000 (19:26 +0000)]
Add missing builtins to the PPC back end for ABI compliance (vol. 1)
This patch corresponds to review:
http://reviews.llvm.org/D10638
This is the back end portion of patch
http://reviews.llvm.org/D10637
It just adds the code gen and intrinsic functions necessary to support that patch to the back end.
Pete Cooper [Fri, 26 Jun 2015 18:17:36 +0000 (18:17 +0000)]
Add op_values() to iterate over the SDValue operands of an SDNode.
SDNode already had ops() which would iterate over the operands and return
SDUse*. This version instead gets the SDValue's out of the SDUse's so that
we can use foreach in more places.
Adam Nemet [Fri, 26 Jun 2015 17:25:43 +0000 (17:25 +0000)]
[LAA] Try to prove non-wrapping of pointers if SCEV cannot
Summary:
Scalar evolution does not propagate the non-wrapping flags to values
that are derived from a non-wrapping induction variable because
the non-wrapping property could be flow-sensitive.
This change is a first attempt to establish the non-wrapping property in
some simple cases. The main idea is to look through the operations
defining the pointer. As long as we arrive to a non-wrapping AddRec via
a small chain of non-wrapping instruction, the pointer should not wrap
either.
I believe that this essentially is what Andy described in
http://article.gmane.org/gmane.comp.compilers.llvm.cvs/220731 as the way
forward.
Alex Lorenz [Fri, 26 Jun 2015 16:46:11 +0000 (16:46 +0000)]
MIR Serialization: Serialize machine basic block operands.
This commit serializes machine basic block operands. The
machine basic block operands use the following syntax:
%bb.<id>[.<name>]
This commit also modifies the YAML representation for the
machine basic blocks - a new, required field 'id' is added
to the MBB YAML mapping.
The id is used to resolve the MBB references to the
actual MBBs. And while the name of the MBB can be
included in a MBB reference, this name isn't used to
resolve MBB references - as it's possible that multiple
MBBs will reference the same BB and thus they will have the
same name. If the name is specified, the parser will verify
that it is equal to the name of the MBB with the specified id.
Aaron Ballman [Fri, 26 Jun 2015 14:51:22 +0000 (14:51 +0000)]
Silencing spurious MSVC C4189 warnings regarding local variables that are initialized but not used; NFC. This bug has been reported to Microsoft (https://connect.microsoft.com/VisualStudio/feedback/details/1475983).
Toma Tabacu [Fri, 26 Jun 2015 13:20:17 +0000 (13:20 +0000)]
[mips] [IAS] Add partial support for the ULW pseudo-instruction.
Summary:
This only adds support for ULW of an immediate address with/without a source register.
It does not include support for ULW of the address of a symbol.
Rafael Espindola [Fri, 26 Jun 2015 12:18:49 +0000 (12:18 +0000)]
Simplify getSymbolType.
This is still a really odd function. Most calls are in object format specific
contexts and should probably be replaced with a more direct query, but at least
now this is not too obnoxious to use.
Hao Liu [Fri, 26 Jun 2015 02:32:07 +0000 (02:32 +0000)]
[AArch64] Lower interleaved memory accesses to ldN/stN intrinsics. This patch also adds a function to calculate the cost of interleaved memory accesses.
E.g. Lower an interleaved load:
%wide.vec = load <8 x i32>, <8 x i32>* %ptr
%v0 = shuffle %wide.vec, undef, <0, 2, 4, 6>
%v1 = shuffle %wide.vec, undef, <1, 3, 5, 7>
into:
%ld2 = { <4 x i32>, <4 x i32> } call llvm.aarch64.neon.ld2(%ptr)
%vec0 = extractelement { <4 x i32>, <4 x i32> } %ld2, i32 0
%vec1 = extractelement { <4 x i32>, <4 x i32> } %ld2, i32 1
E.g. Lower an interleaved store:
%i.vec = shuffle <8 x i32> %v0, <8 x i32> %v1, <0, 4, 8, 1, 5, 9, 2, 6, 10, 3, 7, 11>
store <12 x i32> %i.vec, <12 x i32>* %ptr
into:
%sub.v0 = shuffle <8 x i32> %v0, <8 x i32> v1, <0, 1, 2, 3>
%sub.v1 = shuffle <8 x i32> %v0, <8 x i32> v1, <4, 5, 6, 7>
%sub.v2 = shuffle <8 x i32> %v0, <8 x i32> v1, <8, 9, 10, 11>
call void llvm.aarch64.neon.st3(%sub.v0, %sub.v1, %sub.v2, %ptr)
Hao Liu [Fri, 26 Jun 2015 02:10:27 +0000 (02:10 +0000)]
[InterleavedAccess] Add a pass InterleavedAccess to identify interleaved memory accesses and transform into target specific intrinsics.
E.g. An interleaved load (Factor = 2):
%wide.vec = load <8 x i32>, <8 x i32>* %ptr
%v0 = shuffle <8 x i32> %wide.vec, <8 x i32> undef, <0, 2, 4, 6>
%v1 = shuffle <8 x i32> %wide.vec, <8 x i32> undef, <1, 3, 5, 7>
It can be transformed into a ld2 intrinsic in AArch64 backend or a vld2 intrinsic in ARM backend.
E.g. An interleaved store (Factor = 3):
%i.vec = shuffle <8 x i32> %v0, <8 x i32> %v1, <0, 4, 8, 1, 5, 9, 2, 6, 10, 3, 7, 11>
store <12 x i32> %i.vec, <12 x i32>* %ptr
It can be transformed into a st3 intrinsic in AArch64 backend or a vst3 intrinsic in ARM backend.
Alexey Samsonov [Fri, 26 Jun 2015 00:00:47 +0000 (00:00 +0000)]
[ASan] Use llvm::getDISubprogram() to get function entry debug location.
It can be more robust than copying debug info from first non-alloca
instruction in the entry basic block. We use the same strategy in
coverage instrumentation.
AsmPrinter: Use an intrusively linked list for DIE::Children
Replace the `std::vector<>` for `DIE::Children` with an intrusively
linked list. This is a strict memory improvement: it requires no
auxiliary storage, and reduces `sizeof(DIE)` by one pointer. It also
factors out the DIE-related malloc traffic.
This drops llc memory usage from 735 MB down to 718 MB, or ~2.3%.
(I'm looking at `llc` memory usage on `verify-uselistorder.lto.opt.bc`;
see r236629 for details.)
Change `DIE::Values` to a singly linked list, where each node is
allocated on a `BumpPtrAllocator`. In order to support `push_back()`,
the list is circular, and points at the tail element instead of the
head. I abstracted the core list logic out to `IntrusiveBackList` so
that it can be reused for `DIE::Children`, which also cares about
`push_back()`.
This drops llc memory usage from 799 MB down to 735 MB, about 8%.
(I'm looking at `llc` memory usage on `verify-uselistorder.lto.opt.bc`;
see r236629 for details.)
Anna Zaks [Thu, 25 Jun 2015 23:35:48 +0000 (23:35 +0000)]
[asan] Do not instrument special purpose LLVM sections.
Do not instrument globals that are placed in sections containing "__llvm"
in their name.
This fixes a bug in ASan / PGO interoperability. ASan interferes with LLVM's
PGO, which places its globals into a special section, which is memcpy-ed by
the linker as a whole. When those goals are instrumented, ASan's memcpy wrapper
reports an issue.
Adrian Prantl [Thu, 25 Jun 2015 23:19:19 +0000 (23:19 +0000)]
Debug Info: Add basic test coverage for the DWARF encoding of bitfields.
While looking at a couple of bugs in the debug info output for bitfields
I noticed that there wasn't a single regression test to test my changes
against, so here's a start.
Frederic Riss [Thu, 25 Jun 2015 21:57:33 +0000 (21:57 +0000)]
IAS: Use the root macro instanciation for location
r224810 fixed the handling of macro debug locations in AsmParser. This patch
fixes the logic to actually do what was intended: it uses the first macro of
the macro stack instead of the last one. The updated testcase shows that the
current scheme doesn't work when macro instanciations are nested and multiple
files are used.
Split out code to patch up the `DW_AT_stmt_list` for the cloned DIE, and
reorganize it so that it doesn't depend on `DIE::values_begin()` and
`DIE::values_end()` (which I'm trying to kill off).
David Blaikie and I talked about adding a range-algorithm version of
`std::find_if()`, but the assertion *still* required getting at the end
iterator. IMO, a separate helper function with an early return is
easier to reason about here.
A follow-up commit that removes `DIE::setValue()` and mutates the
`DIEValue` directly is coming shortly.
Jingyue Wu [Thu, 25 Jun 2015 20:14:47 +0000 (20:14 +0000)]
[InstCombine] call SimplifyICmpInst with correct context
Summary:
Fixes PR23809. Without passing the context to SimplifyICmpInst, we would
use the assume to prove that the condition feeding the assume is
trivially true (see isValidAssumeForContext in ValueTracking.cpp),
causing the removal of the assume which may be useful for later
optimizations.
projects / llvm / log