Dinesh Dwivedi [Mon, 2 Jun 2014 07:24:36 +0000 (07:24 +0000)]
Added inst combine transforms for single bit tests from Chris's note
if ((x & C) == 0) x |= C becomes x |= C
if ((x & C) != 0) x ^= C becomes x &= ~C
if ((x & C) == 0) x ^= C becomes x |= C
if ((x & C) != 0) x &= ~C becomes x &= ~C
if ((x & C) == 0) x &= ~C becomes nothing
Alp Toker [Mon, 2 Jun 2014 01:40:04 +0000 (01:40 +0000)]
GraphWriter: detect graph viewer programs at runtime
Replace the crufty build-time configure checks for program paths with
equivalent runtime logic.
This lets users install graphing tools as needed without having to reconfigure
and rebuild LLVM, while eliminating a long chain of inappropriate compile
dependencies that included GUI programs and the windowing system.
Additional features:
* Support the OS X 'open' command to view graphs generated by any of the
Graphviz utilities. This is an alternative to the Graphviz OS X UI which is
no longer available on Mountain Lion.
* Produce informative log output upon failure to indicate which programs can
be installed to view graphs.
Ping me if this doesn't work for your particular environment.
Since we cannot yet use variadic templates, add a specialisation for
6-parameters to format. This is motivated by a need for the additional
parameter for formatting information for an unwind decoder for Windows on ARM.
Introduce the support structures necessary to deal with the Windows ARM EH data.
These definitions are extremely aggressive about assertions to aid future use
for generation of the entries and subsequent decoding.
The names for the various fields are meant to reflect the names used by the
Visual Studio toolchain to aid communication.
Due to the complexity in reading a few of the values, there are a couple of
additional utility functions to decode the information.
In general, there are two ways to encode the unwinding information:
- packed, which places the data inline into the
_IMAGE_ARM_RUNTIME_FUNCTION_ENTRY structure.
- unpacked, which places the data into auxiliary structures placed into the
.xdata section.
The set of structures allow reading of data in either encoding, with the minor
caveat that epilogue scopes need to be decoded manually by constructing the
structure from the data returned by the RuntimeFunction structure.
These definitions are meant for read-only access at the current point as the
first use of them will be to decode the exception information.
David Blaikie [Sun, 1 Jun 2014 03:38:13 +0000 (03:38 +0000)]
DebugInfo: Assert that DbgVariables have associated DIEs
This was previously committed in r209680 and reverted in r209683 after
it caused sanitizer builds to crash.
The issue seems to be that the DebugLoc associated with dbg.value IR
intrinsics isn't necessarily accurate. Instead, we duplicate the
DIVariables and add an InlinedAt field to them to record their
location.
We were using this InlinedAt field to compute the LexicalScope for the
variable, but not using it in the abstract DbgVariable construction and
mapping. This resulted in a formal parameter to the current concrete
function, correctly having no InlinedAt information, but incorrectly
having a DebugLoc that described an inlined location within the
function... thus an abstract DbgVariable was created for the variable,
but its DIE was never constructed (since the LexicalScope had no such
variable). This DbgVariable was silently ignored (by testing for a
non-null DIE on the abstract DbgVariable).
So, fix this by using the right scoping information when constructing
abstract DbgVariables.
In the long run, I suspect we want to undo the work that added this
second kind of location tracking and fix the places where the DebugLoc
propagation on the dbg.value intrinsic fails. This will shrink debug
info (by not duplicating DIVariables), make it more efficient (by not
having to construct new DIVariable metadata nodes to try to map back to
a single variable), and benefit all instructions.
But perhaps there are insurmountable issues with DebugLoc quality that
I'm unaware of... I just don't know how we can't /just keep the DebugLoc
from the dbg.declare to the dbg.values and never get this wrong/.
Adam Nemet [Sat, 31 May 2014 16:23:20 +0000 (16:23 +0000)]
[SelectionDAG] Force cycle detection in AssignTopologicalOrder before aborting
DAG cycle detection is only enabled with ENABLE_EXPENSIVE_CHECKS. However we
can run it just before we would crash in order to provide more informative
diagnostics.
Now in addition to the "Overran sorted position" message we also get the Node
printed if a cycle was detected.
Tested by building several configs: Debug+Assert, Debug+Assert+Check (this is
ENABLE_EXPENSIVE_CHECKS), Release+Assert and Release. Also tried that the
AssignTopologicalOrder assert produces the expected results.
Benjamin Kramer [Sat, 31 May 2014 15:01:54 +0000 (15:01 +0000)]
[Reassociate] Similar to "X + -X" -> "0", added code to handle "X + ~X" -> "-1".
Handle "X + ~X" -> "-1" in the function Value *Reassociate::OptimizeAdd(Instruction *I, SmallVectorImpl<ValueEntry> &Ops);
This patch implements:
TODO: We could handle "X + ~X" -> "-1" if we wanted, since "-X = ~X+1".
Simon Atanasyan [Sat, 31 May 2014 04:51:07 +0000 (04:51 +0000)]
[yaml2obj] Add new command line option `-docnum`.
Input YAML file might contain multiple object file definitions.
New option `-docnum` allows to specify an ordinal number (starting from 1)
of definition used for an object file generation.
Andrea Di Biagio [Fri, 30 May 2014 23:17:53 +0000 (23:17 +0000)]
[X86] Add two combine rules to simplify dag nodes introduced during type legalization when promoting nodes with illegal vector type.
This patch teaches the backend how to simplify/canonicalize dag node
sequences normally introduced by the backend when promoting certain dag nodes
with illegal vector type.
This patch adds two new combine rules:
1) fold (shuffle (bitcast (BINOP A, B)), Undef, <Mask>) ->
(shuffle (BINOP (bitcast A), (bitcast B)), Undef, <Mask>)
Both rules are only triggered on the type-legalized DAG.
In particular, rule 1. is a target specific combine rule that attempts
to sink a bitconvert into the operands of a binary operation.
Rule 2. is a target independet rule that attempts to move a shuffle
immediately after a binary operation.
Convert a vselect into a concat_vector if possible
Summary:
If both vector args to vselect are concat_vectors and the condition is
constant and picks half a vector from each argument, convert the vselect
into a concat_vectors.
Added a test.
The ConvertSelectToConcatVector is assuming it doesn't get vselects with
arguments of, for example, <undef, undef, true, true>. Those get taken
care of in the checks above its call.
Summary:
Separate the check for blend shuffle_vector masks into isBlendMask.
This function will also be used to check if a vector shuffle is legal. No
change in functionality was intended, but we ended up improving codegen on
two tests, which were being (more) optimized only if the resulting shuffle
was legal.
Logan Chien [Fri, 30 May 2014 16:48:56 +0000 (16:48 +0000)]
Fix MIPS exception personality encoding.
For MIPS, we have to encode the personality routine with
an indirect pointer to absptr; otherwise, some link warning
warning will be raised, and the program might crash in some
early MIPS Android device.
Tim Northover [Fri, 30 May 2014 13:23:06 +0000 (13:23 +0000)]
ARM: use AAPCS-style prologues for embedded MachO.
Darwin prologues save their GPRs in two stages: a narrow push of r0-r7 & lr,
followed by a wide push of the remaining registers if there are any. AAPCS uses
a single push.w instruction.
It turns out that, on average, enough registers get pushed that code is smaller
in the AAPCS prologue, which is a nice property for M-class programmers. They
also have other options available for back-traces, so can hopefully deal with
the fact that FP & LR aren't adjacent in memory.
Tim Northover [Fri, 30 May 2014 10:09:59 +0000 (10:09 +0000)]
ARM & AArch64: make use of common cmpxchg idioms after expansion
The C and C++ semantics for compare_exchange require it to return a bool
indicating success. This gets mapped to LLVM IR which follows each cmpxchg with
an icmp of the value loaded against the desired value.
When lowered to ldxr/stxr loops, this extra comparison is redundant: its
results are implicit in the control-flow of the function.
This commit makes two changes: it replaces that icmp with appropriate PHI
nodes, and then makes sure earlyCSE is called after expansion to actually make
use of the opportunities revealed.
I've also added -{arm,aarch64}-enable-atomic-tidy options, so that
existing fragile tests aren't perturbed too much by the change. Many
of them either rely on undef/unreachable too pervasively to be
restored to something well-defined (particularly while making sure
they test the same obscure assert from many years ago), or depend on a
particular CFG shape, which is disrupted by SimplifyCFG.
Karthik Bhat [Fri, 30 May 2014 04:31:24 +0000 (04:31 +0000)]
Allow vectorization of intrinsics such as powi,cttz and ctlz in Loop and SLP Vectorizer.
This patch adds support to vectorize intrinsics such as powi, cttz and ctlz in Vectorizer. These intrinsics are different from other
intrinsics as second argument to these function must be same in order to vectorize them and it should be represented as a scalar.
Review: http://reviews.llvm.org/D3851#inline-32769 and http://reviews.llvm.org/D3937#inline-32857
Chandler Carruth [Fri, 30 May 2014 03:50:06 +0000 (03:50 +0000)]
Add a trivial ArrayRef helper overload to the SelectionDAG interface.
I'm using this pretty frequently in a patch I'm working on and it seems
generally useful.
Nick Lewycky [Fri, 30 May 2014 02:31:27 +0000 (02:31 +0000)]
When analyzing params/args for readnone/readonly, don't forget to consider that a pointer argument may be passed through a callsite to the return, and that we may need to analyze it. Fixes a bug reported on llvm-dev: http://lists.cs.uiuc.edu/pipermail/llvmdev/2014-May/073098.html
Adam Nemet [Thu, 29 May 2014 23:35:36 +0000 (23:35 +0000)]
[X86] Remove AVX1 vbroadcast intrinsics
The corresponding CFE patch replaces these intrinsics with vector initializers
in avxintrin.h. This patch removes the LLVM intrinsics from the backend.
We now stop lowering at X86ISD::VBROADCAST custom node rather than lowering
that further to the intrinsics.
The patch only changes VBROADCASTS* and leaves VBROADCAST[FI]128 to continue
to use intrinsics. As explained in the CFE patch, the reason is that we
currently don't generate as good code for them without the intrinsics.
CodeGen/X86/avx-vbroadcast.ll already provides coverage for this change. It
checks that for a series of insertelements we generate the appropriate
vbroadcast instruction.
Also verified that there was no assembly change in the test-suite before and
after this patch.
Adam Nemet [Thu, 29 May 2014 23:35:33 +0000 (23:35 +0000)]
[X86] Auto-upgrade AVX1 vbroadcast intrinsics
They are replaced with the same IR that is generated for the
vector-initializers in avxintrin.h.
The test verifies that we get back the original instruction. I haven't seen
this approach to be used in other auto-upgrade tests (i.e. llc + FileCheck)
but I think it's the most direct way to test this case. I believe this should
work because llc upgrades calls during parsing. (Other tests mostly check
that assembling and disassembling yields the upgraded IR.)
Chandler Carruth [Thu, 29 May 2014 23:21:12 +0000 (23:21 +0000)]
And fix my fix to sink down through the type at the right time. My
original fix would actually trigger the *exact* same crasher as the
original bug for a different reason. Awesomesauce.
Working on test cases now, but wanted to get bots healthier.
Chandler Carruth [Thu, 29 May 2014 23:05:52 +0000 (23:05 +0000)]
Fix one bug in the latest incarnation of r209843 -- combining GEPs
across PHI nodes. The code was computing the Idxs from the 'GEP'
variable's indices when what it wanted was Op1's indices. This caused an
ASan heap-overflow for me that pin pointed the issue when Op1 had more
indices than GEP did. =] I'll let Louis add a specific test case for
this if he wants.
Added tests for shufflevector lowering to blend instrs.
These tests ensure that a change I will propose in clang works as
expected.
Summary:
Added tests for the generation of blend+immediate instructions from a
shufflevector.
These tests were proposed along with a patch that was dropped. I'm
committing the tests anyway to protect against possible regressions in
codegen.
Louis Gerbarg [Thu, 29 May 2014 20:29:47 +0000 (20:29 +0000)]
Add support for combining GEPs across PHI nodes
Currently LLVM will generally merge GEPs. This allows backends to use more
complex addressing modes. In some cases this is not happening because there
is PHI inbetween the two GEPs:
GEP1--\
|-->PHI1-->GEP3
GEP2--/
This patch checks to see if GEP1 and GEP2 are similiar enough that they can be
cloned (GEP12) in GEP3's BB, allowing GEP->GEP merging (GEP123):
Sebastian Pop [Thu, 29 May 2014 19:44:09 +0000 (19:44 +0000)]
implement missing SCEVDivision case
without this case we would end on an infinite recursion: the remainder is zero,
so Numerator - Remainder is equal to Numerator and so we would recursively ask
for the division of Numerator by Denominator.
Sebastian Pop [Thu, 29 May 2014 19:44:05 +0000 (19:44 +0000)]
fail to find dimensions when ElementSize is nullptr
when ScalarEvolution::getElementSize returns nullptr it is safe to early return
in ScalarEvolution::findArrayDimensions such that we avoid later problems when
we try to divide the terms by ElementSize.
Adrian Prantl [Thu, 29 May 2014 16:56:48 +0000 (16:56 +0000)]
Debug Info: Remove unused code. The MInsn of an _abstract_ variable is
never used again and updating the abstract variable for each inlined
instance of it was questionable in the first place.
Rafael Espindola [Thu, 29 May 2014 15:41:38 +0000 (15:41 +0000)]
[PPC] Use alias symbols in address computation.
This seems to match what gcc does for ppc and what every other llvm
backend does.
This is a fixed version of r209638. The difference is to avoid any change
in behavior for functions. The logic for using constant pools for function
addresseses is spread over a few places and we have to keep them in sync.
Simon Atanasyan [Thu, 29 May 2014 11:05:31 +0000 (11:05 +0000)]
[elf2yaml][ELF] Move Info field to the RelocationSection structure. This
field represents ELF section header sh_info field and does not have any
sense for regular sections. Its interpretation depends on section type.
Dinesh Dwivedi [Thu, 29 May 2014 06:47:23 +0000 (06:47 +0000)]
LCSSA should be performed on the outermost affected loop while unrolling loop.
During loop-unroll, loop exits from the current loop may end up in in different
outer loop. This requires to re-form LCSSA recursively for one level down from
the outer most loop where loop exits are landed during unroll. This fixes PR18861.
Alexey Samsonov [Thu, 29 May 2014 01:44:13 +0000 (01:44 +0000)]
[ASan] Hoist blacklisting globals from init-order checking to Clang.
Clang knows about the sanitizer blacklist and it makes no sense to
add global to the list of llvm.asan.dynamically_initialized_globals if it
will be blacklisted in the instrumentation pass anyway. Instead, we should
do as much blacklisting as possible (if not all) in the frontend.
[x86] Fold extract_vector_elt of a load into the Load's address computation.
An address only use of an extract element of a load can be simplified to a
load. Without this the result of the extract element is spilled to the
stack so that an address is available.
Alexey Samsonov [Thu, 29 May 2014 00:51:15 +0000 (00:51 +0000)]
[ASan] Use llvm.global_ctors to insert init-order checking calls into ASan runtime.
Don't assume that dynamically initialized globals are all initialized from
_GLOBAL__<module_name>I_ function. Instead, scan the llvm.global_ctors and
insert poison/unpoison calls to each function there.
Reid Kleckner [Wed, 28 May 2014 22:49:12 +0000 (22:49 +0000)]
Add a simple helper function to create a 64-bit integer.
Add a function to combine two 32-bit integers into a 64-bit integer.
There are no calls to this function yet, although a subsequent change
will add some in LLDB.