BuildMI instructions were not including MachineMemOperand information.
This was discovered by 'SingleSource/Benchmarks/Stanford/Oscar' failing
due to a FrameIndex load incorrectly being hoisted by postra-machine-licm.
No other tests have been found to fail.
llvm-svn: 201562
ldrd r6, r7 [r2, #15]
simply gives an error and does not triggers an assertion.
As Jim points out, the diagnostic is really strange here,
but fixing that would be more complicated. The missing
comma results in the parser expecting a construct like r2[2],
which is the vector index thing the error message is talking
about. That's not what the user intended, though, and there's
nothing else in the instruction that looks at all like a vector.
Yet more fallout from not having a real parser here and trying
to do context-free generic matching for addressing modes.
rdar://15097243
llvm-svn: 201531
This is implemented by handling assignments to the '.' pseudo symbol
as ".org" directives.
Differential Revision: http://llvm-reviews.chandlerc.com/D2625
llvm-svn: 201530
Until this point only macro definition with named parameters were parsed but the
names were ignored. This adds support for using that information for named
parameter instantiation.
In order to support the full semantics of the keyword arguments, the arguments
are no longer lazily initialised since the keyword arguments can be specified
out of order and partially if they are defaulted. Prepopulate the arguments
with the default value for any defaulted parameters, and then parse the
specified arguments.
This simplies some of the handling of the arguments in the inner loop since
empty arguments simply increment the parameter index and move on.
Note that keyword and positional arguments cannot be mixed.
llvm-svn: 201499
NaCl's ARM ABI uses 16 byte stack alignment, so set that in
ARMSubtarget.cpp.
Using 16 byte alignment exposes an issue in code generation in which a
varargs function leaves a 4 byte gap between the values of r1-r3 saved
to the stack and the following arguments that were passed on the
stack. (Previously, this code only needed to support 4 byte and 8
byte alignment.)
With this issue, llc generated:
varargs_func:
sub sp, sp, #16
push {lr}
sub sp, sp, #12
add r0, sp, #16 // Should be 20
stm r0, {r1, r2, r3}
ldr r0, .LCPI0_0 // Address of va_list
add r1, sp, #16
str r1, [r0]
bl external_func
Fix the bug by checking for "Align > 4". Also simplify the code by
using OffsetToAlignment(), and update comments.
Differential Revision: http://llvm-reviews.chandlerc.com/D2677
llvm-svn: 201497
During LSR of one loop we can run into a situation where we have to expand the
start of a recurrence of a loop induction variable in this loop. This start
value is a value derived of the induction variable of a preceeding loop. SCEV
has cannonicalized this value to a different recurrence than the recurrence of
the preceeding loop's induction variable (the type and/or step direction) has
changed). When we come to instantiate this SCEV we created a second induction
variable in this preceeding loop. This patch tries to base such derived
induction variables of the preceeding loop's induction variable.
This helps twolf on arm and seems to help scimark2 on x86.
Reapply with a fix for the case of a value derived from a pointer.
radar://15970709
llvm-svn: 201496
alongside DIEBlock and replace uses accordingly. Use DW_FORM_exprloc
in DWARF4 and later code. Update testcases.
Adding a DIELoc instead of using extra forms inside DIEBlock so
that we can keep location expressions separate from other uses. No
direct use at the moment, however, it's not a lot of code and
using a separately named class keeps it somewhat more obvious
what's going on in various locations.
llvm-svn: 201481
The Linux kernel defines empty macros for compatibility with ARM UAL syntax.
The comma after the name is optional, and if present can be safely lexed. This
improves compatibility with the GNU assembler.
llvm-svn: 201474
This adds a partial implementation of the .arch_extension directive to the
integrated ARM assembler. There are a number of limitations to this
implementation arising from the target backend support rather than the
implementation itself. Namely, iWMMXT (v1 and v2), Maverick, and XScale support
is not present in the ARM backend. Currently, there is no check for A-class
only (needed for virt), and no ARMv6k detection (needed for os and sec). The
remainder of the extensions are fully supported.
llvm-svn: 201471
This broke in r185459 while TLS support was being generalized to handle
non-symbol TLS representations.
I thought about/tried having an enum rather than a bool to track the
TLS-ness of the address table entry, but namespaces and naming seemed
more hassle than it was worth for only one caller that needed to specify
this.
llvm-svn: 201469
During LSR of one loop we can run into a situation where we have to expand the
start of a recurrence of a loop induction variable in this loop. This start
value is a value derived of the induction variable of a preceeding loop. SCEV
has cannonicalized this value to a different recurrence than the recurrence of
the preceeding loop's induction variable (the type and/or step direction) has
changed). When we come to instantiate this SCEV we created a second induction
variable in this preceeding loop. This patch tries to base such derived
induction variables of the preceeding loop's induction variable.
This helps twolf on arm and seems to help scimark2 on x86.
radar://15970709
llvm-svn: 201465
Type units will share the statement list of their defining compile unit.
This is a tradeoff that reduces .o debug info size at the cost of some
linked debug info size (since the contents of those string tables won't
be deduplicated along with the type unit) which seems right for now.
llvm-svn: 201445
These types have an out of line virtual function each (emitHeader at
least) so they won't have weak vtables - no need for more than that.
llvm-svn: 201444
This probably also addresses the FIXME in the fission case regarding
multiple compile units, though I haven't tested that.
This code still confuses me (the literal zero offset makes little sense,
the limitations surrounding asm output I'm not sure about either - but
perhaps we should just always emit one line table? Or should we not rely
on .loc/.file even in assembly so we can produce the same output between
asm and object output?) but this maintains the existing functionality.
llvm-svn: 201441
This should be a small build time improvement in general and fixes
the build on OS X with -DBUILD_SHARED_LIBS=ON.
The issue is that not all users are including GenericDomTreeConstruction.h,
causing undefined references when ld64 managed to hide the
linkonce_odr symbols.
llvm-svn: 201440
Recommitting r201380 (reverted in r201389)
Recommitting r201351 and r201355 (reverted in r201351 and r201355)
We weren't emitting the an empty (header only) line table when the line
table was empty - this made the DWARF invalid (the compile unit would
point to the zero-size debug_lines section where there should've been an
empty line table but there was nothing at all). Fix that, and as a
consequence this works around/addresses PR18809.
Also, we emit a non-empty line table to workaround a darwin linker bug,
so XFAILing on darwin too.
Also, mark the test as 'REQUIRES: object-emission' because it does.
llvm-svn: 201429
Summary:
This adds support for emitting DWARF path discriminator values in
the object streamer. It also changes the DWARF dumper to show
discriminator values in the line table output.
Reviewers: echristo
CC: llvm-commits
Differential Revision: http://llvm-reviews.chandlerc.com/D2794
llvm-svn: 201427
1) Fix a specific bug when certain conversion functions are called in a program compiled as mips16 with hard float and
the program is linked as c++. There are two libraries that are reversed in the link order with gcc/g++ and clang/clang++ for
mips16 in this case and the proper stubs will then not be called. These stubs are normally handled in the Mips16HardFloat pass
but in this case we don't know at that time that we need to generate the stubs. This must all be handled later in code generation
and we have moved this functionality to MipsAsmPrinter. When linked as C (gcc or clang) the proper stubs are linked in from libc.
2) Set up the infrastructure to handle 90% of what is in the Mips16HardFloat pass in this new area of MipsAsmPrinter. This is a more
logical place to handle this and we have known for some time that we needed to move the code later and not implement it using
inline asm as we do now but it was not clear exactly where to do this and what mechanism should be used. Now it's clear to us
how to do this and this patch contains the infrastructure to move most of this to MipsAsmPrinter but the actual moving will be done
in a follow on patch. The same infrastructure is used to fix this current bug as described in #1. This change was requested by the list
during the original putback of the Mips16HardFloat pass but was not practical for us do at that time.
llvm-svn: 201426
As v1i1 is illegal, the type legalizer tries to scalarize such node. But if the type operands of SETCC is legal, the scalarization algorithm will cause an assertion failure.
llvm-svn: 201381
Recommitting r201351 and r201355 (reverted in r201351 and r201355)
We weren't emitting the an empty (header only) line table when the line
table was empty - this made the DWARF invalid (the compile unit would
point to the zero-size debug_lines section where there should've been an
empty line table but there was nothing at all). Fix that, and as a
consequence this works around/addresses PR18809.
llvm-svn: 201380
A simple register copy on X86 is just 3 bytes, whereas movabsq is a 10 byte
instruction. Marking movabsq as not beeing cheap will allow LICM to move it
out of the loop and it also prevents unnecessary rematerializations if the
value is needed in more than one register.
llvm-svn: 201377
Summary:
There should be a space before each of these two keywords to avoid
generating invalid assembly files.
NOTE: I could not find an obvious maintainers in CODE_OWNERS.TXT, but
this seems related to debug info.
Reviewers: echristo
CC: llvm-commits
Differential Revision: http://llvm-reviews.chandlerc.com/D2791
llvm-svn: 201359
This is an optimistic version of create_diretories: it tries to create the
directory first and looks at the parent only if that fails.
Running strace on "mkdir -p" shows that it is pessimistic, calling mkdir on
every element of the path. We could implement that if needed.
In any case, with both strategies there is no reason to call stat, just check
the return of mkdir.
llvm-svn: 201347
Summary:
AsmPrinter::EmitInlineAsm() will no longer use the EmitRawText() call for
targets with mature MC support. Such targets will always parse the inline
assembly (even when emitting assembly). Targets without mature MC support
continue to use EmitRawText() for assembly output.
The hasRawTextSupport() check in AsmPrinter::EmitInlineAsm() has been replaced
with MCAsmInfo::UseIntegratedAs which when true, causes the integrated assembler
to parse inline assembly (even when emitting assembly output). UseIntegratedAs
is set to true for targets that consider any failure to parse valid assembly
to be a bug. Target specific subclasses generally enable the integrated
assembler in their constructor. The default value can be overridden with
-no-integrated-as.
All tests that rely on inline assembly supporting invalid assembly (for example,
those that use mnemonics such as 'foo' or 'hello world') have been updated to
disable the integrated assembler.
Changes since review (and last commit attempt):
- Fixed test failures that were missed due to configuration of local build.
(fixes crash.ll and a couple others).
- Fixed tests that happened to pass because the local build was on X86
(should fix 2007-12-17-InvokeAsm.ll)
- mature-mc-support.ll's should no longer require all targets to be compiled.
(should fix ARM and PPC buildbots)
- Object output (-filetype=obj and similar) now forces the integrated assembler
to be enabled regardless of default setting or -no-integrated-as.
(should fix SystemZ buildbots)
Reviewers: rafael
Reviewed By: rafael
CC: llvm-commits
Differential Revision: http://llvm-reviews.chandlerc.com/D2686
llvm-svn: 201333
This fix checks the original LLVM IR node to identify opaque constants by
looking for the bitcast-constant pattern. Originally we looked at the generated
SDNode, but this might lead to incorrect results. The SDNode could have been
generated by an constant expression that was folded to a constant.
This fixes <rdar://problem/16050719>
llvm-svn: 201291
As defined in LangRef, aliases do not have sections. However, LLVM's
GlobalAlias class inherits from GlobalValue, which means we can read and
set its section. We should probably ban that as a separate change,
since it doesn't make much sense for an alias to have a section that
differs from its aliasee.
Fixes PR18757, where the section was being lost on the global in code
from Clang like:
extern "C" {
__attribute__((used, section("CUSTOM"))) static int in_custom_section;
}
Reviewers: rafael.espindola
Differential Revision: http://llvm-reviews.chandlerc.com/D2758
llvm-svn: 201286
logical operations on the i1's driving them. This is a bad idea for every
target I can think of (confirmed with micro tests on all of: x86-64, ARM,
AArch64, Mips, and PowerPC) because it forces the i1 to be materialized into
a general purpose register, whereas consuming it directly into a select generally
allows it to exist only transiently in a predicate or flags register.
Chandler ran a set of performance tests with this change, and reported no
measurable change on x86-64.
llvm-svn: 201275
'OK_NonUniformConstValue' to identify operands which are constants but
not constant splats.
The cost model now allows returning 'OK_NonUniformConstValue'
for non splat operands that are instances of ConstantVector or
ConstantDataVector.
With this change, targets are now able to compute different costs
for instructions with non-uniform constant operands.
For example, On X86 the cost of a vector shift may vary depending on whether
the second operand is a uniform or non-uniform constant.
This patch applies the following changes:
- The cost model computation now takes into account non-uniform constants;
- The cost of vector shift instructions has been improved in
X86TargetTransformInfo analysis pass;
- BBVectorize, SLPVectorizer and LoopVectorize now know how to distinguish
between non-uniform and uniform constant operands.
Added a new test to verify that the output of opt
'-cost-model -analyze' is valid in the following configurations: SSE2,
SSE4.1, AVX, AVX2.
llvm-svn: 201272
Instead of expanding a packed shift into a sequence of scalar shifts,
the backend now tries (when possible) to convert the vector shift into a
vector multiply.
Before this change, a shift of a MVT::v8i16 vector by a
build_vector of constants was always scalarized into a long sequence of "vector
extracts + scalar shifts + vector insert".
With this change, if there is SSE2 support, we emit a single vector multiply.
This change also affects SSE4.1, AVX, AVX2 shifts:
- A shift of a MVT::v4i32 vector by a build_vector of non uniform constants
is now lowered when possible into a single SSE4.1 vector multiply.
- Packed v16i16 shift left by constant build_vector are now expanded when
possible into a single AVX2 vpmullw.
This change also improves the lowering of AVX512f vector shifts.
Added test CodeGen/X86/vec_shift6.ll with some code examples that are affected
by this change.
llvm-svn: 201271
We are now no longer relying on the target-specific call lowering implementation
to lower a stackmap intrinsic call. Instead we perform the call lowering in a
target-independent way directly in the stackmap lowering code. This simplifies
the code and removes the need to fixup the code after the target-specific call
lowering.
llvm-svn: 201263
The ID type for the stackmap and patchpoint intrinsics are in both cases i64.
This fixes an zero extend in the SelectionDAGBuilder that still used i32. This
also updates the target independent instructions STACKMAP and PATCHPOINT to use
the correct type.
llvm-svn: 201262
required for all sections in a module. This can be useful when targets or
code-models place strict requirements on how sections must be laid out
in memory.
If RTDyldMemoryManger::needsToReserveAllocationSpace() is overridden to return
true then the JIT will call the following method on the memory manager, which
can be used to preallocate the necessary memory.
void RTDyldMemoryManager::reserveAllocationSpace(uintptr_t CodeSize,
uintptr_t DataSizeRO,
uintptr_t DataSizeRW)
Patch by Vaidas Gasiunas. Thanks very much Viadas!
llvm-svn: 201259
Summary:
AsmPrinter::EmitInlineAsm() will no longer use the EmitRawText() call for targets with mature MC support. Such targets will always parse the inline assembly (even when emitting assembly). Targets without mature MC support continue to use EmitRawText() for assembly output.
The hasRawTextSupport() check in AsmPrinter::EmitInlineAsm() has been replaced with MCAsmInfo::UseIntegratedAs which when true, causes the integrated assembler to parse inline assembly (even when emitting assembly output). UseIntegratedAs is set to true for targets that consider any failure to parse valid assembly to be a bug. Target specific subclasses generally enable the integrated assembler in their constructor. The default value can be overridden with -no-integrated-as.
All tests that rely on inline assembly supporting invalid assembly (for example, those that use mnemonics such as 'foo' or 'hello world') have been updated to disable the integrated assembler.
Reviewers: rafael
Reviewed By: rafael
CC: llvm-commits
Differential Revision: http://llvm-reviews.chandlerc.com/D2686
llvm-svn: 201237
I found that swapping the order of some header files helped fix a
build issue that we're seeing on mingw32. Without the swap, windows.h
was being included before _WIN32_WINNT was being defined and the
CreateHardLinkW function was #ifdef'd out.
It looks like the header is mainly used to get the SHGetFolderPathW
function, so I don't think that there'll be much fallout from the
switch.
Suggested by Alex Crichton. Thanks!
llvm-svn: 201230
There's still one piece missing here, which is adding the
DW_AT_stmt_list to the type unit that refer's to the compile unit's line
table. Working on that.
llvm-svn: 201198
This is preliminary work to fix type unit file strings so they appear in
their originating CU's line table - but it's also just good/simple
cleanup, so I'm committing it ahead of time.
llvm-svn: 201195
We used to be pretty vague about what debug entities were what, with
many conditionals to silently drop/skip/accept things. These don't seem
to be relevant anymore.
llvm-svn: 201194
* CPRCs may be allocated to co-processor registers or the stack – they may never be allocated to core registers
* When a CPRC is allocated to the stack, all other VFP registers should be marked as unavailable
The difference is only noticeable in rare cases where there are a large number of floating point arguments (e.g.
7 doubles + additional float, double arguments). Although it's probably still better to avoid vmov as it can cause
stalls in some older ARM cores. The other, more subtle benefit, is to minimize difference between the various
calling conventions.
rdar://16039676
llvm-svn: 201193
Debug info: Emit values in subregisters that do not have a separate
DWARF register number by emitting a super-register + DW_OP_bit_piece.
This is necessary because on x86_64, there are no DWARF register numbers
for i386-style subregisters.
Fixes a bunch of FIXMEs.
rdar://problem/16015314
llvm-svn: 201190
This comes up in empty files or files containing #file directives that
never reference the actual source file name. Came up in a small test of
line tables I was playing with.
llvm-svn: 201187
DWARF register number by emitting a super-register + DW_OP_bit_piece.
This is necessary because on x86_64, there are no DWARF register numbers
for i386-style subregisters.
Fixes a bunch of FIXMEs.
rdar://problem/16015314
llvm-svn: 201180
sections. The call to data.getUnsigned(&Offset, AddressSize) only
increments Offset if the read succeeds, which will result in an infinite
loop.
llvm-svn: 201179
BUILD_VECTOR nodes, e.g.:
(concat_vectors (BUILD_VECTOR a1, a2, a3, a4), (BUILD_VECTOR b1, b2, b3, b4))
->
(BUILD_VECTOR a1, a2, a3, a4, b1, b2, b3, b4)
This fixes an issue with AVX, where a sequence was not recognized as a 256-bit
vbroadcast due to the concat_vectors.
llvm-svn: 201158
Fixes PR18753 and PR18782.
This is necessary for LICM to preserve LCSSA correctly and efficiently.
There is still some active discussion about whether we should be using
LCSSA, but we can't just immediately stop using it and we *need* LICM to
preserve it while we are using it. We can restore the old SSAUpdater
driven code if and when there is a serious effort to remove the reliance
on LCSSA from all of the loop passes.
However, this also serves as a great example of why LCSSA is very nice
to have. This change significantly simplifies the process of sinking
instructions for LICM, and makes it quite a bit less expensive.
It wouldn't even be as complex as it is except that I had to start the
process of removing the big recursive LCSSA formation hammer in order to
switch even this much of the re-forming code to asserting that LCSSA was
preserved. I'll fully remove that next just to tidy things up until the
LCSSA debate settles one way or the other.
llvm-svn: 201148
Xcore target ABI requires const data that is externally visible
to be handled differently if it has C-language linkage rather than
C++ language linkage.
Clang now emits ".cp.rodata" section information.
All other externally visible constant data will be placed in the DP section.
llvm-svn: 201144
uintptr_t. An unsigned could overflow for large sections.
No test case - anything big enough to overflow an unsigned is going to take an
appreciable time to zero when the test passes.
The choice of uintptr_t was made to match the RTDyldMemoryManager APIs, but
these should probably be hardcoded to uint64_ts: It is legitimate to JIT for
64-bit targets from a 32-bit host/compiler.
llvm-svn: 201127
The addressing mode matcher checks at some point the profitability of folding an
instruction into the addressing mode. When the instruction to be folded has
several uses, it checks that the instruction can be folded in each use.
To do so, it creates a new matcher for each use and check if the instruction is
in the list of the matched instructions of this new matcher.
The new matchers may promote some instructions and this has to be undone to keep
the state of the original matcher consistent.
A test case will follow.
<rdar://problem/16020230>
llvm-svn: 201121
This function adds an extra path argument to lto_module_create_from_memory.
The path argument will be passed to makeBuffer to make sure the MemoryBuffer
has a name and the created module has a module identifier.
This is mainly for emitting warning messages from the linker. When we emit
warning message on a module, we can use the module identifier.
rdar://15985737
llvm-svn: 201114
A const ObjectFile needs to be able to provide its name. For an IRObjectFile,
that means being able to call the mangler. Since each IRObjectFile can have
a different mangling, it is natural for them to contain a Mangler which is
therefore also const.
llvm-svn: 201113
The crux of the issue is that LCSSA doesn't preserve stateful alias
analyses. Before r200067, LICM didn't cause LCSSA to run in the LTO pass
manager, where LICM runs essentially without any of the other loop
passes. As a consequence the globalmodref-aa pass run before that loop
pass manager was able to survive the loop pass manager and be used by
DSE to eliminate stores in the function called from the loop body in
Adobe-C++/loop_unroll (and similar patterns in other benchmarks).
When LICM was taught to preserve LCSSA it had to require it as well.
This caused it to be run in the loop pass manager and because it did not
preserve AA, the stateful AA was lost. Most of LLVM's AA isn't stateful
and so this didn't manifest in most cases. Also, in most cases LCSSA was
already running, and so there was no interesting change.
The real kicker is that LCSSA by its definition (injecting PHI nodes
only) trivially preserves AA! All we need to do is mark it, and then
everything goes back to working as intended. It probably was blocking
some other weird cases of stateful AA but the only one I have is
a 1000-line IR test case from loop_unroll, so I don't really have a good
test case here.
Hopefully this fixes the regressions on performance that have been seen
since that revision.
llvm-svn: 201104
DS instructions that access local memory can only uses addresses that
are less than or equal to the value of M0. When M0 is uninitialized,
then we experience undefined behavior.
This patch also changes the behavior to emit S_WQM_B64 on pixel shaders
no matter what kind of DS instruction is used.
llvm-svn: 201097
This doesn't change any functionality, since we only have two shader
types (compute and pixel) that use local memory. We're just changing
the logic to match the documentation.
llvm-svn: 201096
Similarly to the vshrn instructions, these are simple zext/sext + trunc
operations. Using normal LLVM IR should allow for better code, and more sharing
with the AArch64 backend.
llvm-svn: 201093
For A- and R-class processors, r12 is not normally callee-saved, but is for
interrupt handlers. See AAPCS, 5.3.1.1, "Use of IP by the linker".
llvm-svn: 201089
vshrn is just the combination of a right shift and a truncate (and the limits
on the immediate value actually mean the signedness of the shift doesn't
matter). Using that representation allows us to get rid of an ARM-specific
intrinsic, share more code with AArch64 and hopefully get better code out of
the mid-end optimisers.
llvm-svn: 201085
Original commits messages:
Add MRMXr/MRMXm form to X86 for use by instructions which treat the 'reg' field of modrm byte as a don't care value. Will allow for simplification of disassembler code.
Simplify a bunch of code by removing the need for the x86 disassembler table builder to know about extended opcodes. The modrm forms are sufficient to convey the information.
llvm-svn: 201065
r201059 appears to cause a crash in a bootstrapped build of clang. Craig
isn't available to look at it right now, so I'm reverting it while he
investigates.
llvm-svn: 201064
Some of the more complex directive and macro handling for GAS compatibility
requires lookahead. Add a single token lookahead in the MCAsmLexer.
llvm-svn: 201058
These methods normally call each other and it is really annoying if the
arguments are in different order. The more common rule was that the arguments
specific to call are first (GV, Encoding, Suffix) and the auxiliary objects
(Mang, TM) come after. This patch changes the exceptions.
llvm-svn: 201044
Before conditional store vectorization/unrolling we had only one
vectorized/unrolled basic block. After adding support for conditional store
vectorization this will not only be one block but multiple basic blocks. The
last block would have the back-edge. I updated the code to use a vector of basic
blocks instead of a single basic block and fixed the users to use the last entry
in this vector. But, I forgot to add the basic blocks to this vector!
Fixes PR18724.
llvm-svn: 201028
The bitcast instruction during constant materialization was not placed correcly
in the presence of phi nodes. This commit fixes the insertion point to be in the
idom instead.
This fixes PR18768
llvm-svn: 201009
This fix first traverses the whole use list of the constant expression and
keeps track of the instructions that need to be updated. Then perform the
fixup afterwards.
llvm-svn: 201008
An intermediate solution until the problems with analyzer plugins linking with
llvm/Support and causing assertions due to duplicate GeneralCategory are solved.
llvm-svn: 200981
According to the AAPCS, when a CPRC is allocated to the stack, all other
VFP registers should be marked as unavailable.
I have also modified the rules for allocating non-CPRCs to the stack, to make
it more explicit that all GPRs must be made unavailable. I cannot think of a
case where the old version would produce incorrect answers, so there is no test
for this.
llvm-svn: 200970
Generalize the AArch64 .td nodes for AssertZext and AssertSext. Use
them to match the relevant pextr store instructions.
The test widen_load-2.ll requires a slight change because with the
stores gone, the remaining instructions are scheduled in a different
order.
Add test cases for SSE4 and AVX variants.
Resolves rdar://13414672.
Patch by Adam Nemet <anemet@apple.com>.
llvm-svn: 200957
mode.
Basically the idea is to transform code like this:
%idx = add nsw i32 %a, 1
%sextidx = sext i32 %idx to i64
%gep = gep i8* %myArray, i64 %sextidx
load i8* %gep
Into:
%sexta = sext i32 %a to i64
%idx = add nsw i64 %sexta, 1
%gep = gep i8* %myArray, i64 %idx
load i8* %gep
That way the computation can be folded into the addressing mode.
This transformation is done as part of the addressing mode matcher.
If the matching fails (not profitable, addressing mode not legal, etc.), the
matcher will revert the related promotions.
<rdar://problem/15519855>
llvm-svn: 200947
This solves a problem where a def machine operand has no uses but has
not been marked dead. In this case, the initial RP analysis was being
extra precise and determining from LiveIntervals the the register was
actually dead. This caused us to omit the register from the RP
tracker's block live out. That's all good, but the per-instruction
summary still accounted for it as a valid def. This could cause an
assertion in the tracker later when we underflow pressure.
This is from a bug report on an out-of-tree target. It is not
reproducible on well-behaved targets. I'm just making an obvious fix
without unit test.
llvm-svn: 200941
There was a problem with the old pattern, so we were copying some
larger immediates into registers when we could have been encoding
them in the instruction.
llvm-svn: 200932
In a previous commit (r199818) we added a const_cast to an existing
subtarget info instead of creating a new one so that we could reuse
it when creating the TargetAsmParser for parsing inline assembly.
This cast was necessary because we needed to reuse the existing STI
to avoid generating incorrect code when the inline asm contained
mode-switching directives (e.g. .code 16).
The root cause of the failure was that there was an implicit sharing
of the STI between the parser and the MCCodeEmitter. To fix a
different but related issue, we now explicitly pass the STI to the
MCCodeEmitter (see commits r200345-r200351).
The const_cast is no longer necessary and we can now create a fresh
STI for the inline asm parser to use.
Differential Revision: http://llvm-reviews.chandlerc.com/D2709
llvm-svn: 200929
The most important part of this is probably adding any cost at all for
operations like zext <8 x i8> to <8 x i32>. Before they were being
recorded as extremely costly (24, I believe) which made LLVM fall back
on a 4-wide vectorisation of a loop.
It also rebalances the values for sext, zext and trunc. Lacking any
other sane metric that might work across CPU microarchitectures I went
for instructions. This seems to be in reasonable accord with the rest
of the table (sitofp, ...) though no doubt at least one value is
sub-optimal for some bizarre reason.
Finally, separate AVX and AVX2 values are provided where appropriate.
The CodeGen is quite different in many cases.
rdar://problem/15981990
llvm-svn: 200928
The aim in this patch is to reduce work that VirtRegRewriter needs to do when
telling MachineRegisterInfo which physregs are in use. Up until now
VirtRegRewriter::rewrite has been doing rewriting and populating def info and
then proceeding to set whether a physreg is used based this info for every
physreg that the target provides. This can be expensive when a target has an
unusually high number of supported physregs, and is a noticeable chunk of
compile time for small programs on such targets.
So to reduce compile time, this patch simply adds the use of a SparseSet to the
rewrite function that is used to flag each physreg that is encountered in a
MachineFunction. Afterward, rather than iterating over the set of all physregs
for a given target to set the physregs used in MachineRegisterInfo, the new way
is to iterate over the set of physregs that were actually encountered and set
in the SparseSet. This improves compile time because the existing rewrite
function was iterating over all MachineOperands already, and because the
iterations afterward to setPhysRegUsed is reduced by use of the SparseSet data.
llvm-svn: 200919
I believe VZEXT_MOVL means "zero all vector elements except the first" (and
should have identical input & output types) whereas VZEXT means "zero extend
each element of a vector (discarding higher elements if necessary)".
For example:
(v4i32 (vzext (v16i8 ...)))
should zero extend the low 4 bytes of the incoming vector to 32-bits,
discarding higher bytes.
However, somewhere in the past, these two concepts had become confused, even
leading to a nonsensical VSEXT_MOVL.
This re-merges the nodes where appropriate (all VSEXT_MOVL -> VSEXT, VZEXT_MOVL
-> VZEXT when it's an actual extension).
rdar://problem/15981990
llvm-svn: 200918
programs on targets with large register files. The root of the compile time
overhead was in the use of llvm::SmallVector to hold PhysRegEntries, which
resulted in slow-down from calling llvm::SmallVector::assign(N, 0). In contrast
std::vector uses the faster __platform_bzero to zero out primitive buffers when
assign is called, while SmallVector uses an iterator.
The fix for this was simply to replace the SmallVector with a dynamically
allocated buffer and to initialize or reinitialize the buffer based on the
total registers that the target architecture requires. The changes support
cases where a pass manager may be reused for different targets, and note that
the PhysRegEntries is allocated using calloc mainly for good for, and also to
quite tools like Valgrind (see comments for more info on this).
There is an rdar to track the fact that SmallVector doesn't have platform
specific speedup optimizations inside of it for things like this, and I'll
create a bugzilla entry at some point soon as well.
TL;DR: This fix replaces the expensive llvm::SmallVector<unsigned
char>::assign(N, 0) with a call to calloc for N bytes which is much faster
because SmallVector's assign uses iterators.
llvm-svn: 200917
large register files. The omission of Queries.clear() is perfectly safe because
LiveIntervalUnion::Query doesn't contain any data that needs freeing and
because LiveRegMatrix::runOnFunction happens to reset the OwningArrayPtr
holding Queries every time it is run, so there's no need to zero out the
queries either. Not having to do this for very large numbers of physregs
is a noticeable constant cost reduction in compilation of small programs.
llvm-svn: 200913
build but spectacularly changed behavior of the C++98 build. =]
This shows my one problem with not having unittests -- basic API
expectations aren't well exercised by the integration tests because they
*happen* to not come up, even though they might later. I'll probably add
a basic unittest to complement the integration testing later, but
I wanted to revive the bots.
llvm-svn: 200905
The primary motivation for this pass is to separate the call graph
analysis used by the new pass manager's CGSCC pass management from the
existing call graph analysis pass. That analysis pass is (somewhat
unfortunately) over-constrained by the existing CallGraphSCCPassManager
requirements. Those requirements make it *really* hard to cleanly layer
the needed functionality for the new pass manager on top of the existing
analysis.
However, there are also a bunch of things that the pass manager would
specifically benefit from doing differently from the existing call graph
analysis, and this new implementation tries to address several of them:
- Be lazy about scanning function definitions. The existing pass eagerly
scans the entire module to build the initial graph. This new pass is
significantly more lazy, and I plan to push this even further to
maximize locality during CGSCC walks.
- Don't use a single synthetic node to partition functions with an
indirect call from functions whose address is taken. This node creates
a huge choke-point which would preclude good parallelization across
the fanout of the SCC graph when we got to the point of looking at
such changes to LLVM.
- Use a memory dense and lightweight representation of the call graph
rather than value handles and tracking call instructions. This will
require explicit update calls instead of some updates working
transparently, but should end up being significantly more efficient.
The explicit update calls ended up being needed in many cases for the
existing call graph so we don't really lose anything.
- Doesn't explicitly model SCCs and thus doesn't provide an "identity"
for an SCC which is stable across updates. This is essential for the
new pass manager to work correctly.
- Only form the graph necessary for traversing all of the functions in
an SCC friendly order. This is a much simpler graph structure and
should be more memory dense. It does limit the ways in which it is
appropriate to use this analysis. I wish I had a better name than
"call graph". I've commented extensively this aspect.
This is still very much a WIP, in fact it is really just the initial
bits. But it is about the fourth version of the initial bits that I've
implemented with each of the others running into really frustrating
problms. This looks like it will actually work and I'd like to split the
actual complexity across commits for the sake of my reviewers. =] The
rest of the implementation along with lots of wiring will follow
somewhat more rapidly now that there is a good path forward.
Naturally, this doesn't impact any of the existing optimizer. This code
is specific to the new pass manager.
A bunch of thanks are deserved for the various folks that have helped
with the design of this, especially Nick Lewycky who actually sat with
me to go through the fundamentals of the final version here.
llvm-svn: 200903
During DAGCombine visitShiftByConstant assumes that certain binary operations
with only constant operands can always be folded successfully. This is no longer
true when the constant is opaque. This commit fixes visitShiftByConstant by not
performing the optimization for opaque constants. Otherwise we would end up in
an infinite DAGCombine loop.
llvm-svn: 200900
225 is the default value of inline-threshold. This change will make sure
we have the same inlining behavior as prior to r200886.
As Chandler points out, even though we don't have code in our testing
suite that uses cold attribute, there are larger applications that do
use cold attribute.
r200886 + this commit intend to keep the same behavior as prior to r200886.
We can later on tune the inlinecold-threshold.
The main purpose of r200886 is to help performance of instrumentation based
PGO before we actually hook up inliner with analysis passes such as BPI and BFI.
For instrumentation based PGO, we try to increase inlining of hot functions and
reduce inlining of cold functions by setting inlinecold-threshold.
Another option suggested by Chandler is to use a boolean flag that controls
if we should use OptSizeThreshold for cold functions. The default value
of the boolean flag should not change the current behavior. But it gives us
less freedom in controlling inlining of cold functions.
llvm-svn: 200898
This is a nop. doesSectionRequireSymbols is only used from
isSymbolLinkerVisible. isSymbolLinkerVisible only use from ELF was in
if (!Asm.isSymbolLinkerVisible(Symbol) && !Symbol.isUndefined())
return false;
if (Symbol.isTemporary())
return false;
If the symbol is a temporary this code returns false and it is irrelevant if
we take the first if or not. If the symbol is not a temporary,
Asm.isSymbolLinkerVisible returns true without ever calling
doesSectionRequireSymbols.
This was an horrible leftover from when support for ELF was first added.
llvm-svn: 200894
Ideally only those transform passes that run at -O0 remain enabled,
in reality we get as close as we reasonably can.
Passes are responsible for disabling themselves, it's not the job of
the pass manager to do it for them.
llvm-svn: 200892
Added command line option inlinecold-threshold to set threshold for inlining
functions with cold attribute. Listen to the cold attribute when it would
decrease the inline threshold.
llvm-svn: 200886
find a register.
The idea is to choose a color for the variable that cannot be allocated and
recolor its interferences around. Unlike the current register allocation scheme,
it is allowed to change the color of an already assigned (but maybe not
splittable or spillable) live interval while propagating this change to its
neighbors.
In other word, there are two things that may help finding an available color:
- Already assigned variables (RS_Done) can be recolored to different color.
- The recoloring allows to catch solutions that needs to touch more that just
the neighbors of the current allocated variable.
E.g.,
vA can use {R1, R2 }
vB can use { R2, R3}
vC can use {R1 }
Where vA, vB, and vC cannot be split anymore (they are reloads for instance) and
they all interfere.
vA is assigned R1
vB is assigned R2
vC tries to evict vA but vA is already done.
=> Regular register allocation heuristic fails.
Last chance recoloring kicks in:
vC does as if vA was evicted => vC uses R1.
vC is marked as fixed.
vA needs to find a color.
None are available.
vA cannot evict vC: vC is a fixed virtual register now.
vA does as if vB was evicted => vA uses R2.
vB needs to find a color.
R3 is available.
Recoloring => vC = R1, vA = R2, vB = R3.
<rdar://problem/15947839>
llvm-svn: 200883
I think this was just over-eagerness on my part. The analysis results
need to often be non-const because they need to (in some cases at least)
be updated by the transformation pass in order to remain correct. It
also makes lazy analyses (a common case) needlessly annoying to write in
order to make their entire state mutable.
llvm-svn: 200881
This patch adds NaCl target for Mips. It also forbids indexed loads and
stores if the target is NaCl.
Patch by Sasa Stankovic.
Differential Revision: http://llvm-reviews.chandlerc.com/D2690
llvm-svn: 200855
Summary:
The check performed in the comparator is invalid, as some STL
implementations enforce strict weak ordering by calling the comparator with the
same value. This check was also in a wrong place: the assertion would only fire
when -help was used. The new check is performed each time the category is
registered (we are not going to have thousands of them, so it's fine to do it in
O(N^2)).
Reviewers: jordan_rose
Reviewed By: jordan_rose
CC: cfe-commits, alexmc
Differential Revision: http://llvm-reviews.chandlerc.com/D2699
llvm-svn: 200853
In Thumb1 mode, bl instruction might be selected for branches between
basic blocks in the function if the offset is greater than 2KB.
However, this might cause SEGV because the destination symbol
is not marked as thumb function and the execution mode will be reset
to ARM mode.
Since we are sure that these symbols are in the same data fragment, we
can simply resolve these local symbols, and don't emit any relocation
information for this bl instruction.
llvm-svn: 200842
It is not clear how much we should try to expose in getFlags. For example,
should there be a SF_Object and a SF_Text?
But for information that is already being exposed, we may as well use it in
llvm-nm.
llvm-svn: 200820
ISSUE:
On Ubuntu 12.04 LTS, arc4random is provided by libbsd.so, which is a
transitive dependency of libedit. If a system had libedit on it that
was implemented in terms of libbsd.so, then the arc4random test,
previously implemented as a linker test, would succeed with -ledit.
However, on Ubuntu this would also require a #include <bsd/stdlib.h>.
This caused a build breakage on configure-based Ubuntu 12.04 with
libedit installed.
FIX:
This fix changes configure to test for arc4random by searching for it
in the standard header files. On Ubuntu 12.04, this test now properly
fails to find arc4random as it is not defined in the default header
locations. It also tweaks the #define names to match the output of the
header check command, which is slightly different than the linker
function check #defines.
I tested the following scenarios:
(1) Ubuntu 12.04 without the libedit package [did not find arc4random,
as expected]
(2) Ubuntu 12.04 with libedit package [properly did not find
arc4random, as expected]
(3) Ubuntu 12.04 with most recent libedit, custom built, and not
dependent on libbsd.so [properly did not find arc4random, as
expected].
(4) FreeBSD 10.0B1 [properly found arc4random, as expected]
llvm-svn: 200819
Commuting the 231 and 132 variants would swap addends and
multiplicands/multipliers, which isn't valid.
I'm still trying to reduce a decent test case for this.
llvm-svn: 200792
No functional change. Updated loops from:
for (I = scc_begin(), E = scc_end(); I != E; ++I)
to:
for (I = scc_begin(); !I.isAtEnd(); ++I)
for teh win.
llvm-svn: 200789
Patch implements %hi(sym1 - sym2) and %lo(sym1 - sym2) expressions for MIPS
by creating target expression class MipsMCExpr.
Patch by Sasa Stankovic.
Differential Revision: http://llvm-reviews.chandlerc.com/D2592
llvm-svn: 200783
This patch fixes the ldr-pseudo implementation to work when used in
inline assembly. The fix is to move arm assembler constant pools
from the ARMAsmParser class to the ARMTargetStreamer class.
Previously we kept the assembler generated constant pools in the
ARMAsmParser object. This does not work for inline assembly because
a new parser object is created for each blob of inline assembly.
This patch moves the constant pools to the ARMTargetStreamer class
so that the constant pool will remain alive for the entire code
generation process.
An ARMTargetStreamer class is now required for the arm backend.
There was no existing implementation for MachO, only Asm and ELF.
Instead of creating an empty MachO subclass, we decided to make the
ARMTargetStreamer a non-abstract class and provide default
(llvm_unreachable) implementations for the non constant-pool related
methods.
Differential Revision: http://llvm-reviews.chandlerc.com/D2638
llvm-svn: 200777
This fixes a crashes in the OpenCV test suite and also the scrypt
kernel in bfgminer.
I was unable to come up with a reduced test case for this.
https://bugs.freedesktop.org/show_bug.cgi?id=72785
llvm-svn: 200776
There is no lit test for this, because it would be too big and
complicated, but it does fix a crash in the Arithm/Absdiff.* OpenCV test.
llvm-svn: 200775
The OpenCL specs say: "The vector versions of the math functions operate
component-wise. The description is per-component."
Patch by: Jan Vesely
Reviewed-by: Tom Stellard <thomas.stellard@amd.com>
Signed-off-by: Jan Vesely <jan.vesely@rutgers.edu>
llvm-svn: 200773
There was an extremely confusing proliferation of LLVM intrinsics to implement
the vacge & vacgt instructions. This combines them all into two polymorphic
intrinsics, shared across both backends.
llvm-svn: 200768
Until now, when a path in a gcno file included a directory, we would
emit our .gcov file in that directory, whereas gcov always emits the
file in the current directory. In doing so, this implements gcov's
strange name-mangling -p flag, which is needed to avoid clobbering
files when two with the same name exist in different directories.
The path mangling is a bit ugly and only handles unix-like paths, but
it's simple, and it doesn't make any guesses as to how it should
behave outside of what gcov documents. If we decide this should be
cross platform later, we can consider the compatibility implications
then.
llvm-svn: 200754
V_ADD_F32 with source modifier does not produce -0.0 for this. Just
manipulate the sign bit directly instead.
Also add a pattern for (fneg (fabs ...)).
Fixes a bunch of bit encoding piglit tests with radeonsi.
Reviewed-by: Tom Stellard <thomas.stellard@amd.com>
llvm-svn: 200743
Add the missing transformation strchr(p, 0) -> p + strlen(p) to SimplifyLibCalls
and remove the ToDo comment.
Reviewer: Duncan P.N. Exan Smith
llvm-svn: 200736
A bunch of test cases needed to be cleaned up for this, many my fault -
when implementid imported modules I updated test cases by simply
duplicating the prior metadata field - which wasn't always the empty
metadata entry.
llvm-svn: 200731
For some anachronistic reason we were producing {i32 0} for zero-length
debug info arrays.
(this change is paired with a Clang change and may cause temporary
buildbot noise)
Let's not.
llvm-svn: 200721
It disturbs the layout of the parameters in memory and registers,
leading to problems in the backend.
The plan for optimizing internal inalloca functions going forward is to
essentially SROA the argument memory and demote any captured arguments
(things that aren't trivially written by a load or store) to an indirect
pointer to a static alloca.
llvm-svn: 200717
Some of the SHA instructions take a scalar i32 as one argument (largely because
they work on 160-bit hash fragments). This wasn't reflected in the IR
previously, with ARM and AArch64 choosing different types (<4 x i32> and <1 x
i32> respectively) which was ugly.
This makes all the affected intrinsics take a uniform "i32", allowing them to
become non-polymorphic at the same time.
llvm-svn: 200706
'SmallPtrSetImplBase'. This more closely matches the organization of
SmallVector and should allow introducing a SmallPtrSetImpl which serves
the same purpose as SmallVectorImpl: isolating the element type from the
particular small size chosen. This in turn allows a lot of
simplification of APIs by not coding them against a specific small size
which is rarely needed.
llvm-svn: 200687
LowerExpectIntrinsic previously only understood the idiom of an expect
intrinsic followed by a comparison with zero. For llvm.expect.i1, the
comparison would be stripped by the early-cse pass.
Patch by Daniel Micay.
llvm-svn: 200664
Robert Lytton