v16i8 shuffles, and replace it with new facilities.
This uses precise patterns to match exact unpacks, and the new
generalized unpack lowering only when we detect a case where we will
have to shuffle both inputs anyways and they terminate in exactly
a blend.
This fixes all of the blend horrors that I uncovered by always lowering
blends through the vector shuffle lowering. It also removes *sooooo*
much of the crazy instruction sequences required for v16i8 lowering
previously. Much cleaner now.
The only "meh" aspect is that we sometimes use pshufb+pshufb+unpck when
it would be marginally nicer to use pshufb+pshufb+por. However, the
difference there is *tiny*. In many cases its a win because we re-use
the pshufb mask. In others, we get to avoid the pshufb entirely. I've
left a FIXME, but I'm dubious we can really do better than this. I'm
actually pretty happy with this lowering now.
For SSE2 this exposes some horrors that were really already there. Those
will have to fixed by changing a different path through the v16i8
lowering.
llvm-svn: 229846
on things not being marked as either custom or legal, but we now do
custom lowering of more VSELECT nodes. To cope with this, manually
replicate the legality tests here. These have to stay in sync with the
set of tests used in the custom lowering of VSELECT.
Ideally, we wouldn't do any of this combine-based-legalization when we
have an actual custom legalization step for VSELECT, but I'm not going
to be able to rewrite all of that today.
I don't have a test case for this currently, but it was found when
compiling a number of the test-suite benchmarks. I'll try to reduce
a test case and add it.
This should at least fix the test-suite fallout on build bots.
llvm-svn: 229844
lowering paths. I'm going to be leveraging this to simplify a lot of the
overly complex lowering of v8 and v16 shuffles in pre-SSSE3 modes.
Sadly, this isn't profitable on v4i32 and v2i64. There, the float and
double blending instructions for pre-SSE4.1 are actually pretty good,
and we can't beat them with bit math. And once SSE4.1 comes around we
have direct blending support and this ceases to be relevant.
Also, some of the test cases look odd because the domain fixer
canonicalizes these to floating point domain. That's OK, it'll use the
integer domain when it matters and some day I may be able to update
enough of LLVM to canonicalize the other way.
This restores almost all of the regressions from teaching x86's vselect
lowering to always use vector shuffle lowering for blends. The remaining
problems are because the v16 lowering path is still doing crazy things.
I'll be re-arranging that strategy in more detail in subsequent commits
to finish recovering the performance here.
llvm-svn: 229836
First, don't combine bit masking into vector shuffles (even ones the
target can handle) once operation legalization has taken place. Custom
legalization of vector shuffles may exist for these patterns (making the
predicate return true) but that custom legalization may in some cases
produce the exact bit math this matches. We only really want to handle
this prior to operation legalization.
However, the x86 backend, in a fit of awesome, relied on this. What it
would do is mark VSELECTs as expand, which would turn them into
arithmetic, which this would then match back into vector shuffles, which
we would then lower properly. Amazing.
Instead, the second change is to teach the x86 backend to directly form
vector shuffles from VSELECT nodes with constant conditions, and to mark
all of the vector types we support lowering blends as shuffles as custom
VSELECT lowering. We still mark the forms which actually support
variable blends as *legal* so that the custom lowering is bypassed, and
the legal lowering can even be used by the vector shuffle legalization
(yes, i know, this is confusing. but that's how the patterns are
written).
This makes the VSELECT lowering much more sensible, and in fact should
fix a bunch of bugs with it. However, as you'll see in the test cases,
right now what it does is point out the *hilarious* deficiency of the
new vector shuffle lowering when it comes to blends. Fortunately, my
very next patch fixes that. I can't submit it yet, because that patch,
somewhat obviously, forms the exact and/or pattern that the DAG combine
is matching here! Without this patch, teaching the vector shuffle
lowering to produce the right code infloops in the DAG combiner. With
this patch alone, we produce terrible code but at least lower through
the right paths. With both patches, all the regressions here should be
fixed, and a bunch of the improvements (like using 2 shufps with no
memory loads instead of 2 andps with memory loads and an orps) will
stay. Win!
There is one other change worth noting here. We had hilariously wrong
vectorization cost estimates for vselect because we fell through to the
code path that assumed all "expand" vector operations are scalarized.
However, the "expand" lowering of VSELECT is vector bit math, most
definitely not scalarized. So now we go back to the correct if horribly
naive cost of "1" for "not scalarized". If anyone wants to add actual
modeling of shuffle costs, that would be cool, but this seems an
improvement on its own. Note the removal of 16 and 32 "costs" for doing
a blend. Even in SSE2 we can blend in fewer than 16 instructions. ;] Of
course, we don't right now because of OMG bad code, but I'm going to fix
that. Next patch. I promise.
llvm-svn: 229835
If the thread receives a signal concurrently with PTRACE_ATTACH,
we can get notification about the signal before notification about stop.
In such case we need to forward the signal to the thread, otherwise
the signal will be missed (as we do PTRACE_DETACH with arg=0) and
any logic relying on signals will break. After forwarding we need to
continue to wait for stopping, because the thread is not stopped yet.
We do ignore delivery of SIGSTOP, because we want to make stop-the-world
as invisible as possible.
http://reviews.llvm.org/D7723
--This line, and those below, will be ignored--
M lib/sanitizer_common/sanitizer_stoptheworld_linux_libcdep.cc
M test/tsan/signal_segv_handler.cc
llvm-svn: 229832
Previously, subtarget features were a bitfield with the underlying type being uint64_t.
Since several targets (X86 and ARM, in particular) have hit or were very close to hitting this bound, switching the features to use a bitset.
No functional change.
Differential Revision: http://reviews.llvm.org/D7065
llvm-svn: 229831
Scops that only read seem generally uninteresting and scops that only write are
most likely initializations where there is also little to optimize. To not
waste compile time we bail early.
Differential Revision: http://reviews.llvm.org/D7735
llvm-svn: 229820
For projects depending on LLVM, I find it very useful to combine a
release-no-asserts build of LLVM with a debug+asserts build of the dependent
project. The motivation is that when developing a dependent project, you are
debugging that project itself, not LLVM. In my usecase, a significant part of
the runtime is spent in LLVM optimization passes, so I would like to build LLVM
without assertions to get the best performance from this combination.
Currently, `lib/Support/Debug.cpp` changes the set of symbols it provides
depending on NDEBUG, while `include/llvm/Support/Debug.h` requires extra
symbols when NDEBUG is not defined. Thus, it is not possible to enable
assertions in an external project that uses facilities of `Debug.h`.
This patch changes `Debug.cpp` and `Valgrind.cpp` to always define the symbols
that other code may depend on when #including LLVM headers without NDEBUG.
http://reviews.llvm.org/D7662
llvm-svn: 229819
Patch improves lookup into dependendt bases of dependent class and adds lookup
into non-dependent bases.
Differential Revision: http://reviews.llvm.org/D7173
llvm-svn: 229817
Previously we wrongly emitted a base relocation entry for an absolute symbol.
That made the loader to rewrite some instruction operands with wrong values
only when a DLL is not loaded at the default address. That caused a
misterious crash of some executable.
Absolute symbols will of course never change value wherever the binary is
loaded to memory. We shouldn't emit base relocations for absolute symbols.
llvm-svn: 229816
I didn't realize how easily the hostname could change - for example just
changing wireless networks seems to prompt it in some cases.
Users can always set their own local module cache path to avoid this.
This reverts commits r228592, 228594, 228601 and 228613.
rdar://19287368
llvm-svn: 229815
When this test was written, no llvm tool could print out contents
of base relocation section. Now llvm-readobj is able to dump it in
a text format. Use that tool to make this test readable.
llvm-svn: 229814
When tools like llvm-cov show regions, it's much easier to understand
what's happening if the condition of an if shows a counter as well as
the body.
llvm-svn: 229813
Summary:
The definition for _mm256_insert_epi64 was taking an int, which would get
truncated before being inserted in the vector.
Original patch by Joshua Magee!
Reviewers: bruno, craig.topper
Subscribers: cfe-commits
Differential Revision: http://reviews.llvm.org/D7179
llvm-svn: 229811
Summary: No declaration for the type `tuple` is given in c++03 or c++98 modes. Mark all tests that use the actual `tuple` type as UNSUPPORTED.
Reviewers: jroelofs, mclow.lists, danalbert
Reviewed By: danalbert
Subscribers: cfe-commits
Differential Revision: http://reviews.llvm.org/D5956
llvm-svn: 229808
This patch removes the huge blob of code that is dealing with
rtti/exceptions/sanitizers and replaces it with:
A ToolChain function which, for a given set of Args, figures out if rtti
should be:
- enabled
- disabled implicitly
- disabled explicitly
A change in the way SanitizerArgs figures out what sanitizers to enable
(or if it should error out, or warn);
And a check for exceptions/rtti interaction inside addExceptionArgs.
The RTTIMode algorithm is:
- If -mkernel, -fapple-kext, or -fno-rtti are passed, rtti was disabled explicitly;
- If -frtti was passed or we're not targetting the PS4, rtti is enabled;
- If -fexceptions or -fcxx-exceptions was passed and we're targetting
the PS4, rtti was enabled implicitly;
- If we're targetting the PS4, rtti is disabled implicitly;
- Otherwise, rtti is enabled;
Since the only flag needed to pass to -cc1 is -fno-rtti if we want to
disable it, there's no problem in saying rtti is enabled if we're
compiling C code, so we don't look at the input file type.
addExceptionArgs now looks at the RTTIMode and warns that rtti is being
enabled implicitly if targetting the PS4 and exceptions are on. It also
errors out if, targetting the PS4, -fno-rtti was passed, and exceptions
were turned on.
SanitizerArgs now errors out if rtti was disabled explicitly and the vptr
sanitizer was enabled implicitly, but just turns off vptr if rtti is
disabled but -fsanitize=undefined was passed.
Also fixed tests, removed duplicate name from addExceptionArgs comment,
and added one or two surrounding lines when running clang-format.
This changes test/Driver/fsanitize.c to make it not expect a warning when
passed -fsanitize=undefined -fno-rtti, but expect vptr to not be on.
Removed all users and definition of SanitizerArgs::sanitizesVptr().
Reviewers: samsonov
Subscribers: llvm-commits, samsonov, rsmith
Differential Revision: http://reviews.llvm.org/D7525
llvm-svn: 229801
A null MCTargetStreamer allows IRObjectFile to ignore target-specific
directives. Previously we were crashing.
Differential Revision: http://reviews.llvm.org/D7711
llvm-svn: 229797
Jean-Daniel Dupas