RegisterContextLLDB. I have core files of half a dozen tricky
unwind situations on x86/arm and they're all working pretty much
correctly at this point, but we'll need to keep an eye out for
unwinder regressions for a little while; it's tricky to get these
heuristics completely correct in all unwind situations.
<rdar://problem/18937193>
llvm-svn: 221866
MIPS ELF symbols might contain some additional MIPS-specific flags
in the st_other field besides visibility ones. These flags indicate
code properties like microMIPS / MIPS16 encoding, position independent
code etc. We need to transfer the flags from input objects to the
output linked file to write them into the symbol table, adjust symbols
addresses etc.
I add new attribute CodeModel to the DefinedAtom class to hold target
specific flag and to get over YAML/Native format conversion barrier.
Other architectures/targets can extend CodeModel enumeration by their
own flags.
MIPS specific part of this patch adds support for STO_MIPS_MICROMIPS
flag. This flag marks microMIPS symbols. Such symbol should:
a) Has STO_MIPS_MICROMIPS in the corresponding .symtab record.
b) Has adjusted (odd) address in the corresponding .symtab
and .dynsym records.
llvm-svn: 221864
The segment alignment for PT_LOAD segments is set to page size by default, but
if any of the sections require an alignment more than the page size, the segment
alignment property is set to the maximum alignment of the sections that are part
of the segment.
llvm-svn: 221862
between splitting a vector into 128-bit lanes and recombining them vs.
decomposing things into single-input shuffles and a final blend.
This handles a large number of cases in AVX1 where the cross-lane
shuffles would be much more expensive to represent even though we end up
with a fast blend at the root. Instead, we can do a better job of
shuffling in a single lane and then inserting it into the other lanes.
This fixes the remaining bits of Halide's regression captured in PR21281
for AVX1. However, the bug persists in AVX2 because I've made this
change reasonably conservative. The cases where it makes sense in AVX2
to split into 128-bit lanes are much more rare because we can often do
full permutations across all elements of the 256-bit vector. However,
the particular test case in PR21281 is an example of one of the rare
cases where it is *always* better to work in a single 128-bit lane. I'm
going to try to teach the logic to detect and form the good code even in
AVX2 next, but it will need to use a separate heuristic.
Finally, there is one pesky regression here where we previously would
craftily use vpermilps in AVX1 to shuffle both high and low halves at
the same time. We no longer pull that off, and not for any really good
reason. Ultimately, I think this is just another missing nuance to the
selection heuristic that I'll try to add in afterward, but this change
already seems strictly worth doing considering the magnitude of the
improvements in common matrix math shuffle patterns.
As always, please let me know if this causes a surprising regression for
you.
llvm-svn: 221861
Part of TestConcurrentEvents starts threads that are supposed to be
delayed by one second.
Test was adding "delay" threads to the "actions" thread list instead
of the "delay_actions" list, which caused them to be started without
delay.
llvm-svn: 221859
The user can use the max-page-size option and set the maximum page size. Dont
check for maximum allowed values for page size, as its what the kernel is
configured with.
Fix the test as well.
llvm-svn: 221858
Each entry in the delay-import address table had a wrong alignment
requirement if 32 bit. As a result it got wrong delay-import table.
Because llvm-readobj doesn't print out that field, we don't have a
test for that. I'll submit a test that would catch this bug after
improving llvm-readobj.
llvm-svn: 221853
re-combining shuffles because nothing was available in the wider vector
type.
The key observation (which I've put in the comments for future
maintainers) is that at this point, no further combining is really
possible. And so even though these shuffles trivially could be combined,
we need to actually do that as we produce them when producing them this
late in the lowering.
This fixes another (huge) part of the Halide vector shuffle regressions.
As it happens, this was already well covered by the tests, but I hadn't
noticed how bad some of these got. The specific patterns that turn
directly into unpckl/h patterns were occurring *many* times in common
vector processing code.
There are still more problems here sadly, but trying to incrementally
tease them apart and it looks like this is the core of the problem in
the splitting logic.
There is some chance of regression here, you can see it in the test
changes. Specifically, where we stop forming pshufb in some cases, it is
possible that pshufb was in fact faster. Intel "says" that pshufb is
slower than the instruction sequences replacing it.
llvm-svn: 221852
Prior to this patch the TypePromotionHelper was promoting only sign extensions.
Supporting zero extensions changes:
- How constants are extended.
- How sign extensions, zero extensions, and truncate are composed together.
- How the type of the extended operation is recorded. Now we need to know the
kind of the extension as well as its type.
Each change is fairly small, unlike the diff.
Most of the diff are comments/variable renaming to say "extension" instead of
"sign extension".
The performance improvements on the test suite are within the noise.
Related to <rdar://problem/18310086>.
llvm-svn: 221851
This folds the compare emission into the select emission when possible, so we
can directly use the flags and don't have to emit a separate compare.
Related to rdar://problem/18960150.
llvm-svn: 221847
The Autoconf build already does this, but it was never ported to
CMake. The host linker version affects the flags that Clang pass
to the linker, notably whether it passes -demangle or not.
http://reviews.llvm.org/D6239
llvm-svn: 221844
This reverts commit r221836.
The tests are asserting on some buildbots. This also reverts the
test part of r221837 as it relies on dwarfdump dumping the
accelerator tables.
llvm-svn: 221842
Windows normally limits the length of an absolute path name to 260
characters; directories can have lower limits. These limits increase
to about 32K if you use absolute paths with the special '\\?\'
prefix. Teach Support\Windows\Path.inc to use that prefix as needed.
TODO: Other parts of Support could also learn to use this prefix.
llvm-svn: 221841
This avoids an issue where AtEndOfStream mistakenly returns true at the /start/ of
a stream.
(In the rare case that the size is known and actually 0, the slow path will still
handle it correctly.)
llvm-svn: 221840
If x is known to have the range [a, b), in a loop predicated by (icmp
ne x, a) its range can be sharpened to [a + 1, b). Get
ScalarEvolution and hence IndVars to exploit this fact.
This change triggers an optimization to widen-loop-comp.ll, so it had
to be edited to get it to pass.
This change was originally landed in r219834 but had a bug and broke
ASan. It was reverted in r219878, and is now being re-landed after
fixing the original bug.
phabricator: http://reviews.llvm.org/D5639
reviewed by: atrick
llvm-svn: 221839
The DIE offset in the accel tables is an offset relative to the start
of the debug_info section, but we were encoding the offset to the
start of the containing CU.
llvm-svn: 221837
Currently FormValues are only used for attributes of DIEs and thus
uers always have a CU lying around when calling into the FormValue
API.
Accelerator tables encode their information using the same Forms
as the attributes, thus it is natural to use DWARFFormValue to
extract/dump them. There is no CU in that case though. Allow the
API to be called with a null CU arguemnt by making the RelocMap
lookup conditional on the CU pointer validity. And document this
new behvior in the header. (Test coverage for this use of the API
comes in the DwarfAccelTable support patch)
llvm-svn: 221835
expansion into a parameter pack; we know that we're still filling in that
parameter's arguments. Previously, if we hit this case for an alias template,
we'd try to substitute using non-canonical template arguments.
llvm-svn: 221832
The GOT slots were being laid out in a random order by the GOTPass which
caused randomness in the output file.
Note: With this change lld now bootstraps on darwin. That is:
1) link lld using system linker to make lld.1
2) link lld using lld.1 to make lld.2
3) link lld using lld.2 to make lld.3
Now lld.2 and lld.3 are identical.
llvm-svn: 221831
r221820 fixed a problem (PR21548) where an iPTR was used in TLI legality checks,
which isn't valid and resulted in a failed assertion.
The solution was to lower pointer types into the correct target's VT, by
using TL::getValueType instead of EVT::getEVT.
This commit changes 3 other uses of EVT::getEVT, but without any tests:
- One of these non-lowered EVTs is passed to allowsMisalignedMemoryAccesses,
which goes into target's TL implementation and doesn't cause any problem (yet.)
- Two others are passed to TLI.isOperationLegalOrCustom:
- one only looks at extensions, so doesn't concern pointers.
- one only looks at binary operators, so also isn't a problem.
The latter might some day be exposed to pointers and cause the same assert as
the original PR, because there's a comment hinting at also supporting cast ops.
For consistency, update all of them and be done with it.
llvm-svn: 221827
The code in {add,sub}v.i3 routines does not trap when it should, because
it performs the actual add/subtract operation in signed arithmetic,
rather than unsigned.
Patch by Francois-Xavie Coudert!
llvm-svn: 221826
On darwin in final linked images, the __TEXT segment covers that start of the
file. That means in memory a process can see the mach_header (and load commands)
for every loaded image in a process. There are APIs that take and return the
mach_header addresses as a way to specify a particular loaded image.
For completeness, any code can get the address of the mach_header of the image
it is in by using &__dso_handle. In addition there are mach-o type specific
symbols like __mh_execute_header.
The linker needs to supply a definition for any of these symbols if used. But
the address the symbol it resolves to is not in any section. Instead it is the
address of the start of the __TEXT segment.
I needed to make a small change to SimpleFileNode to not override
resetNextIndex() because the Driver creates a SimpleFileNode to hold the
internal/implicit files that the context/writer can create. For some reason
SimpleFileNode overrode resetNextIndex() to do nothing instead of reseting
the index (which mach-o needs if the internal file is an archive).
llvm-svn: 221822
Summary:
Consider the following nifty 1 liner: (0 ? csqrtl(2.0f) : sqrtl(2.0f)). One can easily obtain such code from e.g. tgmath. Right now it produces an assertion because we fail to do the promotion real => _Complex real.
The case was properly handled previously (old handleOtherComplexFloatConversion routine), but was forgotten in the current version. This seems to be about fallout from r219557
Reviewers: chandlerc, rsmith
Reviewed By: rsmith
Subscribers: cfe-commits
Differential Revision: http://reviews.llvm.org/D6217
llvm-svn: 221821
This is a follow-on to r221706 and r221731 and discussed in more detail in PR21385.
This patch also loosens the testcase checking for btver2. We know that the "1.0" will be loaded, but
we can't tell exactly when, so replace the CHECK-NEXT specifiers with plain CHECKs. The CHECK-NEXT
sequence relied on a quirk of post-RA-scheduling that may change independently of anything in these tests.
llvm-svn: 221819
The check for unnamed members was intended to skip unnamed bitfields,
but it ended up skipping unnamed structs. This lead to an assertion in
IRGen.
llvm-svn: 221818
One of them (__memcpy_chk) was already there, the others were checked
by comparing function names.
Note that the fortified libfuncs are now part of TLI, but are always
available, because they aren't generated, only optimized into the
non-checking versions.
Differential Revision: http://reviews.llvm.org/D6179
llvm-svn: 221817
Jason Molenda