Fixed GEP visitor in the InstCombine pass.
The current implementation of GEP visitor in InstCombine fails with assertion on Vector GEP with mix of scalar and vector types, like this:
getelementptr double, double* %a, <8 x i32> %i
(It fails to create a "sext" from <8 x i32> to <8 x i64>)
Teresa Johnson [Sun, 15 Nov 2015 02:00:09 +0000 (02:00 +0000)]
Use a different block id for block of metadata kind records
Summary:
There are currently two blocks with the METADATA_BLOCK id at module
scope. The first has the module-level metadata values (consisting of
some combination of METADATA_* record codes except for METADATA_KIND).
The second consists only of METADATA_KIND records. The latter is used
only in the METADATA_ATTACHMENT block within function blocks (for
metadata attached to instructions).
For ThinLTO we want to delay the parsing of module level metadata
until all functions have been imported from that module (there is some
bookkeeping used to suture it up when we read it during a post-pass).
However, we do need the METADATA_KIND records when parsing the function
body during importing, since those kinds are used as described above.
To simplify identification and parsing of just the block containing
the metadata kinds, use a different block id (METADATA_KIND_BLOCK_ID).
Support older bitcode without the new block id as well.
Akira Hatanaka [Sat, 14 Nov 2015 06:35:56 +0000 (06:35 +0000)]
Reduce the size of MCRelaxableFragment.
MCRelaxableFragment previously kept a copy of MCSubtargetInfo and
MCInst to enable re-encoding the MCInst later during relaxation. A copy
of MCSubtargetInfo (instead of a reference or pointer) was needed
because the feature bits could be modified by the parser.
This commit replaces the MCSubtargetInfo copy in MCRelaxableFragment
with a constant reference to MCSubtargetInfo. The copies of
MCSubtargetInfo are kept in MCContext, and the target parsers are now
responsible for asking MCContext to provide a copy whenever the feature
bits of MCSubtargetInfo have to be toggled.
With this patch, I saw a 4% reduction in peak memory usage when I
compiled verify-uselistorder.lto.bc using llc.
Don't recompute LCSSA after loop-unrolling when possible.
Summary:
Currently we always recompute LCSSA for outer loops after unrolling an
inner loop. That leads to compile time problem when we have big loop
nests, and we can solve it by avoiding unnecessary work. For instance,
if w eonly do partial unrolling, we don't break LCSSA, so we don't need
to rebuild it. Also, if all exits from the inner loop are inside the
enclosing loop, then complete unrolling won't break LCSSA either.
I replaced unconditional LCSSA recomputation with conditional recomputation +
unconditional assert and added several tests, which were failing when I
experimented with it.
Soon I plan to follow up with a similar patch for recalculation of dominators
tree.
Akira Hatanaka [Sat, 14 Nov 2015 05:20:05 +0000 (05:20 +0000)]
[MCTargetAsmParser] Move the member varialbes that reference
MCSubtargetInfo in the subclasses into MCTargetAsmParser and define a
member function getSTI.
This is done in preparation for making changes to shrink the size of
MCRelaxableFragment. (see http://reviews.llvm.org/D14346).
Quentin Colombet [Sat, 14 Nov 2015 01:55:17 +0000 (01:55 +0000)]
[ShrinkWrapping] Disable the optimization for functions with sanitize like
attribute.
Even if the target supports shrink-wrapping, the prologue and epilogue
must not move because a crash can happen anywhere and sanitizers need
to be able to unwind from the PC of the crash.
Justin Bogner [Fri, 13 Nov 2015 23:05:46 +0000 (23:05 +0000)]
AArch64: Default AArch64Subtarget::ReserveX18 to true on darwin
Darwin reserves x18, so it's never ABI compliant to generate code that
uses it. Set the default value based on the OS part of the triple
rather than forcing front-ends to set the +reserve-x18 target feature
in order to build correct code for Darwin.
This will make r243310 redundant, so I'll revert that shortly.
Reid Kleckner [Fri, 13 Nov 2015 21:27:00 +0000 (21:27 +0000)]
[WinEH] Fix ESP management with 32-bit __CxxFrameHandler3
The C++ EH personality automatically restores ESP from the C++ EH
registration node after a catchret. I mistakenly thought it was like
SEH, which does not restore ESP.
It makes sense for C++ EH to differ from SEH here because SEH does not
use funclets for catches, and does not allow catching inside of finally.
C++ EH may need to unwind through multiple catch funclets and eventually
catchret to some outer funclet. Therefore, the runtime has to keep track
of which ESP to use with catchret, rather than having the compiler
reload it manually.
Evgeniy Stepanov [Fri, 13 Nov 2015 21:21:42 +0000 (21:21 +0000)]
[safestack] Rewrite isAllocaSafe using SCEV.
Use ScalarEvolution to calculate memory access bounds.
Handle function calls based on readnone/nocapture attributes.
Handle memory intrinsics with constant size.
This change improves both recall and precision of IsAllocaSafe.
See the new tests (ex. BitCastWide) for the kind of code that was wrongly
classified as safe.
SCEV efficiency seems to be limited by the fact the SafeStack runs late
(in CodeGenPrepare), and many loops are unrolled or otherwise not in LCSSA.
Cong Hou [Fri, 13 Nov 2015 19:47:43 +0000 (19:47 +0000)]
[X86][SSE] Combine UNPCKL with vector_shuffle into UNPCKH to save one instruction for sext from v16i8 to v16i16 and v8i16 to v8i32.
This patch is enabling combining UNPCKL with vector_shuffle that moves the upper
half of a vector into the lower half, into a UNPCKH instruction. For example:
David Blaikie [Fri, 13 Nov 2015 19:18:49 +0000 (19:18 +0000)]
dwarfdump: Add support for dumping the table contents of DWP indexes
This is a recommit of 252842 which was reverted in 252859. The issue was
using %s format specifier for a StringRef - used Format's
left_justify(StringRef, int) instead.
It'd be nice to have __attribute__((format(..))) on llvm::format, but
apparently it's only implemented for c-style variadics, not C++ variadic
templates. Perhaps we could fix that & conditionalize the attribute on
such...
Reid Kleckner [Fri, 13 Nov 2015 17:00:36 +0000 (17:00 +0000)]
[Symbolizer] Don't use PE symbol tables to override PDB symbols
Summary:
PE files are stripped by default, and only contain the names of exported
symbols.
The actual reason that we bother to do this override by default is
actually due to a quirk of the way -gline-tables-only is implemented, so
I phrased the check as "if we are symbolizing from dwarf, do the symtab
override".
This fixes lots of Windows ASan tests that I broke in r250582.
Dan Liew [Fri, 13 Nov 2015 11:37:25 +0000 (11:37 +0000)]
[lit] Fix bug where ``lit.util.which()`` would return a directory
instead of executable if the argument was found inside a directory
contained in PATH.
An example where this could cause a problem is if there was a RUN line
that ran the ``test`` command and if the user had a directory in their
PATH that contained a directory called ``test/`` (that occured before
``/usr/bin/``). Lit would try to use the directory as the executable
which would fail with the rather cryptic message.
```
Could not create process due to [Errno 13] Permission denied
```
James Molloy [Fri, 13 Nov 2015 10:05:31 +0000 (10:05 +0000)]
[AArch64] Check the expansion of BITREVERSE in regression test
Something I missed from Hal's review, rightly pointed out by Ben Kramer - we should make sure the expansion is properly checked as it can be easy for bugs to creep in.
I've checked the scalar i8 expansion here and the vector i8 expansion in a previous commit.
James Molloy [Fri, 13 Nov 2015 10:02:36 +0000 (10:02 +0000)]
[SDAG] Fix expansion of BITREVERSE
Richard Trieu noted that UBSan detected an overflowing shift, and the obvious fix caused a crash.
What was happening was that the shiftee (1U) was indeed too small for the possible range of shifts it had to handle, but also we were using "VT.getSizeInBits()" to get the maximum type bitwidth, but we wanted "VT.getScalarSizeInBits()" to get the vector lane size instead of the entire vector size.
Use an APInt for the shift and VT.getScalarSizeInBits().
[llvm-profdata] Add check for text profile formats and improve error reporting (2nd try)
Summary:
This change addresses two possible instances of user error / confusion when
merging sampled profile data.
Previously any input that didn't match the raw or processed instrumented format
would automatically be interpreted as instrumented profile text format data.
No error would be reported during the merge.
Example:
If foo-sampled.profdata and bar-sampled.profdata are binary sampled profiles:
Old behavior:
$ llvm-profdata merge foo-sampled.profdata bar-sampled.profdata -output foobar-sampled.profdata
$ llvm-profdata show -sample foobar-sampled.profdata
error: foobar-sampled.profdata:1: Expected 'mangled_name:NUM:NUM', found lprofi
This change adds basic checks for valid input data when assuming text input.
It also makes error messages related to file format validity more specific about
the assumbed profile data type.
New behavior:
$ llvm-profdata merge foo-sampled.profdata bar-sampled.profdata -o foobar-sampled.profdata
error: foo.profdata: Unrecognized instrumentation profile encoding format
Perhaps you forgot to use the -sample option?
Akira Hatanaka [Fri, 13 Nov 2015 01:23:11 +0000 (01:23 +0000)]
Provide a way to specify inliner's attribute compatibility and merging.
This reapplies r252949. I've changed the type of FuncName to be
std::string instead of StringRef in emitFnAttrCompatCheck.
Original commit message for r252949:
Provide a way to specify inliner's attribute compatibility and merging
rules using table-gen. NFC.
This commit adds new classes CompatRule and MergeRule to Attributes.td,
which are used to generate code to check attribute compatibility and
merge attributes of the caller and callee.
Dan Gohman [Fri, 13 Nov 2015 00:50:04 +0000 (00:50 +0000)]
[WebAssembly] Un-mangle the conversion instruction names.
This arranges the types in the LLVM instruction names in the same order that
they appear in the WebAssembly opcode names, and eliminates
double-underscores.
Joseph Tremoulet [Fri, 13 Nov 2015 00:39:23 +0000 (00:39 +0000)]
[WinEH] Find root frame correctly in CLR funclets
Summary:
The value that the CoreCLR personality passes to a funclet for the
establisher frame may be the root function's frame or may be the parent
funclet's (mostly empty) frame in the case of nested funclets. Each
funclet stores a pointer to the root frame in its own (mostly empty)
frame, as does the root function itself. All frames allocate this slot at
the same offset, measured from the post-prolog stack pointer, so that the
same sequence can accept any ancestor as an establisher frame parameter
value, and so that a single offset can be reported to the GC, which also
looks at this slot.
This change allocate the slot when processing function entry, and records
its frame index on the WinEHFuncInfo object, then inserts the code to
set/copy it during prolog emission.
X86-FMA3: Implemented commute transformations FMA*_Int instructions.
It made it possible to apply the memory folding optimization for the 2nd
operand of FMA*_Int instructions.
Akira Hatanaka [Thu, 12 Nov 2015 20:59:43 +0000 (20:59 +0000)]
Provide a way to specify inliner's attribute compatibility and merging
rules using table-gen. NFC.
This commit adds new classes CompatRule and MergeRule to Attributes.td,
which are used to generate code to check attribute compatibility and
merge attributes of the caller and callee.
r243347 was intended to support a change to LSR (r243348). That change
to LSR has since had to be reverted (r243939) because it was buggy, and
now the code added in r243347 is untested and unexercised. Given that,
I think it is appropriate to revert r243347 for now, with the intent of
adding it back in later if I get around to checking in a fixed version
of r243348.
Teresa Johnson [Thu, 12 Nov 2015 19:31:46 +0000 (19:31 +0000)]
[ThinLTO] Update test to be more tolerant of ordering changes
Update the ThinLTO function importing test to use DAG forms of checks so
that it is more tolerant of changes to relative ordering between
imported decls/defs. This reduces the number of changes required by the
comdat importing patch I am sending for review shortly.
Quentin Colombet [Thu, 12 Nov 2015 18:13:42 +0000 (18:13 +0000)]
[ShrinkWrap] Make sure we do not mess up with EH funclet lowering.
ShrinkWrapping does not understand exception handling constraints for now, so
make sure we do not mess with them by aborting on functions that use EH
funclets.
[llvm-profdata] Add check for text profile formats and improve error reporting
Summary:
This change addresses two possible instances of user error / confusion when
merging sampled profile data.
Previously any input that didn't match the raw or processed instrumented format
would automatically be interpreted as instrumented profile text format data.
No error would be reported during the merge.
Example:
If foo-sampled.profdata and bar-sampled.profdata are binary sampled profiles:
Old behavior:
$ llvm-profdata merge foo-sampled.profdata bar-sampled.profdata -output foobar-sampled.profdata
$ llvm-profdata show -sample foobar-sampled.profdata
error: foobar-sampled.profdata:1: Expected 'mangled_name:NUM:NUM', found lprofi
This change adds basic checks for valid input data when assuming text input.
It also makes error messages related to file format validity more specific about
the assumbed profile data type.
New behavior:
$ llvm-profdata merge foo-sampled.profdata bar-sampled.profdata -o foobar-sampled.profdata
error: foo.profdata: Unrecognized instrumentation profile encoding format
Perhaps you forgot to use the -sample option?
projects / llvm / log