Duncan P. N. Exon Smith [Fri, 25 Apr 2014 18:15:50 +0000 (18:15 +0000)]
SCC: Un-inline long functions
These are long functions that really shouldn't be inlined. Otherwise,
no functionality change.
<rdar://problem/
14292693>
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207249
91177308-0d34-0410-b5e6-
96231b3b80d8
Duncan P. N. Exon Smith [Fri, 25 Apr 2014 18:10:23 +0000 (18:10 +0000)]
SCC: Remove redundant inline keywords, NFC
Functions declared in line in a class are inlined by default. There's
no reason for the `inline` keyword.
<rdar://problem/
14292693>
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207248
91177308-0d34-0410-b5e6-
96231b3b80d8
Reed Kotler [Fri, 25 Apr 2014 18:05:00 +0000 (18:05 +0000)]
Make sure that DSUB does not duplicate the pattern of DSUBU
Test Plan:
Run test suite to make sure there is no regression.
https://dmz-portal.mips.com/bb/builders/LLVM%20with%2064bit%20and%20delay%20slot%20optimizer%20and%20direct%20object%20emitter/builds/626
Reviewers: dsanders
Reviewed By: dsanders
Differential Revision: http://reviews.llvm.org/D3497
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207247
91177308-0d34-0410-b5e6-
96231b3b80d8
Saleem Abdulrasool [Fri, 25 Apr 2014 17:51:25 +0000 (17:51 +0000)]
ARM: remove @llvm.arm.sevl
This intrinsic is no longer needed with the new @llvm.arm.hint(i32) intrinsic
which provides a generic, extensible manner for adding hint instructions. This
functionality can now be represented as @llvm.arm.hint(i32 5).
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207246
91177308-0d34-0410-b5e6-
96231b3b80d8
Manman Ren [Fri, 25 Apr 2014 17:34:55 +0000 (17:34 +0000)]
[inline cold threshold] Command line argument for inline threshold will
override the default cold threshold.
When we use command line argument to set the inline threshold, the default
cold threshold will not be used. This is in line with how we use
OptSizeThreshold. When we want a higher threshold for all functions, we
do not have to set both inline threshold and cold threshold.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207245
91177308-0d34-0410-b5e6-
96231b3b80d8
David Blaikie [Fri, 25 Apr 2014 17:32:19 +0000 (17:32 +0000)]
Refactor some common logic in DwarfUnit::constructVariableDIE and pass non-null DIE by reference to DbgVariable::setDIE
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207244
91177308-0d34-0410-b5e6-
96231b3b80d8
Saleem Abdulrasool [Fri, 25 Apr 2014 17:24:24 +0000 (17:24 +0000)]
ARM: provide a new generic hint intrinsic
Introduce the llvm.arm.hint(i32) intrinsic that can be used to inject hints into
the instruction stream. This is particularly useful for generating IR from a
compiler where the user may inject an intrinsic (e.g. __yield). These are then
pattern substituted into the correct instruction which already existed.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207242
91177308-0d34-0410-b5e6-
96231b3b80d8
David Blaikie [Fri, 25 Apr 2014 17:07:55 +0000 (17:07 +0000)]
PR19554: Fix some memory leaks in DIEHashTest.cpp
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207240
91177308-0d34-0410-b5e6-
96231b3b80d8
Adrian Prantl [Fri, 25 Apr 2014 17:01:04 +0000 (17:01 +0000)]
Reapply r207135 without modifications.
Debug info: Let dbg.values inserted by LowerDbgDeclare inherit the location
of the dbg.value. This gets rid of tons of redundant variable DIEs in
subscopes.
rdar://problem/
14874886, rdar://problem/
16679936
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207236
91177308-0d34-0410-b5e6-
96231b3b80d8
Adrian Prantl [Fri, 25 Apr 2014 17:01:00 +0000 (17:01 +0000)]
This reapplies r207130 with an additional testcase+and a missing check for
AllocaInst that was missing in one location.
Debug info for optimized code: Support variables that are on the stack and
described by DBG_VALUEs during their lifetime.
Previously, when a variable was at a FrameIndex for any part of its
lifetime, this would shadow all other DBG_VALUEs and only a single
fbreg location would be emitted, which in fact is only valid for a small
range and not the entire lexical scope of the variable. The included
dbg-value-const-byref testcase demonstrates this.
This patch fixes this by
Local
- emitting dbg.value intrinsics for allocas that are passed by reference
- dropping all dbg.declares (they are now fully lowered to dbg.values)
SelectionDAG
- renamed constructors for SDDbgValue for better readability.
- fix UserValue::match() to handle indirect values correctly
- not inserting an MMI table entries for dbg.values that describe allocas.
- lowering dbg.values that describe allocas into *indirect* DBG_VALUEs.
CodeGenPrepare
- leaving dbg.values for an alloca were they are (see comment)
Other
- regenerated/updated instcombine.ll testcase and included source
rdar://problem/
16679879
http://reviews.llvm.org/D3374
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207235
91177308-0d34-0410-b5e6-
96231b3b80d8
Tilmann Scheller [Fri, 25 Apr 2014 13:43:18 +0000 (13:43 +0000)]
[ARM64] When compiling for ELF in PIC mode, local symbols shouldn't go through the GOT
There's no need for local symbols to go through the GOT, in fact it seems GNU ld is not even emitting GOT entries for local symbols and will error out when trying to resolve a GOT relocation for a local symbol.
This bug triggers when bootstrapping clang on AArch64 Linux with -fPIC and the ARM64 backend. The AArch64 backend is not affected.
With this commit it's now possible to bootstrap clang on AArch64 Linux with the ARM64 backend (-fPIC, -O3).
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207226
91177308-0d34-0410-b5e6-
96231b3b80d8
Jiangning Liu [Fri, 25 Apr 2014 12:07:03 +0000 (12:07 +0000)]
[ARM64] Handle fp128 for parameter passing on stack
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207222
91177308-0d34-0410-b5e6-
96231b3b80d8
Tim Northover [Fri, 25 Apr 2014 10:48:47 +0000 (10:48 +0000)]
ARM64: fix assertion in ISelDAGToDAG
Also an unused variable, so double bonus!
This should deal with PR19548.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207221
91177308-0d34-0410-b5e6-
96231b3b80d8
Bradley Smith [Fri, 25 Apr 2014 10:25:29 +0000 (10:25 +0000)]
[ARM64] Print preferred aliases for SFBM/UBFM in InstPrinter
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207219
91177308-0d34-0410-b5e6-
96231b3b80d8
Chandler Carruth [Fri, 25 Apr 2014 09:52:44 +0000 (09:52 +0000)]
[LCG] During the incremental update of an SCC, switch to using the
SCCMap to test for nodes that have been re-added to the root SCC rather
than a set vector. We already have done the SCCMap lookup, we juts need
to test it in two different ways. In turn, do most of the processing of
these nodes as they go into the root SCC rather than lazily. This
simplifies the final loop to just stitch the root SCC into its
children's parent sets. No functionlatiy changed.
However, this makes a few things painfully obvious, which was my intent.
=] There is tons of repeated code introduced here and elsewhere. I'm
splitting the refactoring of that code into helpers from this change so
its clear that this is the change which switches the datastructures used
around, and the other is a pure factoring & deduplication of code
change.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207217
91177308-0d34-0410-b5e6-
96231b3b80d8
Kevin Qin [Fri, 25 Apr 2014 09:44:20 +0000 (09:44 +0000)]
[ARM64] Add RUN lines for "–target arm64 –mattr=-fp-armv8" on AArch64 no-fp test.
This patch is a supplement of implementing predicate of FP, enabling aarch64 backend
no-fp tests on arm64 target for verification. During this, one bug is exposed and
fixed by this patch.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207215
91177308-0d34-0410-b5e6-
96231b3b80d8
Kevin Qin [Fri, 25 Apr 2014 09:25:42 +0000 (09:25 +0000)]
[ARM64] Support crc predicate on ARM64.
According to the specification, CRC is an optional extension of the
architecture.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207214
91177308-0d34-0410-b5e6-
96231b3b80d8
Chandler Carruth [Fri, 25 Apr 2014 09:08:10 +0000 (09:08 +0000)]
[LCG] During the incremental re-build of an SCC after removing an edge,
remove the nodes in the SCC from the SCC map entirely prior to the DFS
walk. This allows the SCC map to represent both the state of
not-yet-re-added-to-an-SCC and added-back-to-this-SCC independently. The
first is being missing from the SCC map, the second is mapping back to
'this'. In a subsequent commit, I'm going to use this property to
simplify the new node list for this SCC.
In theory, I think this also makes the contract for orphaning a node
from the graph slightly less confusing. Now it is also orphaned from the
SCC graph. Still, this isn't quite right either, and so I'm not adding
test cases here. I'll add test cases for the behavior of orphaning nodes
when the code *actually* supports it. The change here is mostly
incidental, my goal is simplifying the algorithm.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207213
91177308-0d34-0410-b5e6-
96231b3b80d8
Chandler Carruth [Fri, 25 Apr 2014 09:08:05 +0000 (09:08 +0000)]
[LCG] Rather than doing a linear time SmallSetVector removal of each
child from the worklist, wait until we actually need to pop another
element off of the worklist and skip over any that were already visited
by the DFS. This also enables swapping the nodes of the SCC into the
worklist. No functionality changed.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207212
91177308-0d34-0410-b5e6-
96231b3b80d8
Chandler Carruth [Fri, 25 Apr 2014 06:45:06 +0000 (06:45 +0000)]
[LCG] Remove a completely unnecessary loop. It wasn't even doing any
thing, just mucking up the code. I feel bad that I even wrote this loop.
Very sorry. The diff is huge because of the indent change, but I promise
all this is doing is realizing that the outer two loops were actually
the exact same loops, and we didn't need two of them.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207202
91177308-0d34-0410-b5e6-
96231b3b80d8
Chandler Carruth [Fri, 25 Apr 2014 06:38:58 +0000 (06:38 +0000)]
[LCG] Now that the loop structure of the core SCC finding routine is
factored into a more reasonable form, replace the tail call with
a simple outer-loop continuation. It's sad that C++ makes this so
awkward to write, but it seems more direct and clear than the tail call
at this point.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207201
91177308-0d34-0410-b5e6-
96231b3b80d8
Saleem Abdulrasool [Fri, 25 Apr 2014 06:29:36 +0000 (06:29 +0000)]
X86: convert object streamer selection to a switch
Change the object streamer selection to a switch from a series of if conditions.
Rather than defaulting to ELF, require that an ELF format is requested. The
Windows/!ELF is maintained as MachO would have been selected first and will
still provide a MachO format. Add an assertion that if COFF is requested that
the target platform is Windows as only WinCOFF object emission is currently
supported.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207200
91177308-0d34-0410-b5e6-
96231b3b80d8
Anders Waldenborg [Fri, 25 Apr 2014 06:25:15 +0000 (06:25 +0000)]
[python] Fix getting section contents.
The returnvalue was handled as c_char_p which ment that ctypes
handled it as a NUL-terminated string making it cut the contents
at first NUL (or even worse - overrunning the buffer if it doesn't
contain a NUL).
Differential Revision: http://reviews.llvm.org/D3474
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207199
91177308-0d34-0410-b5e6-
96231b3b80d8
David Blaikie [Fri, 25 Apr 2014 06:22:32 +0000 (06:22 +0000)]
Add missing cpp file header
Code review feedback from Paul Robinson on r207022
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207198
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Fri, 25 Apr 2014 05:30:21 +0000 (05:30 +0000)]
[C++] Use 'nullptr'. Target edition.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207197
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Fri, 25 Apr 2014 05:29:35 +0000 (05:29 +0000)]
[C++] Use 'nullptr'. Transforms edition.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207196
91177308-0d34-0410-b5e6-
96231b3b80d8
Duncan P. N. Exon Smith [Fri, 25 Apr 2014 04:38:43 +0000 (04:38 +0000)]
blockfreq: Only one mass distribution per node
Remove the concepts of "forward" and "general" mass distributions, which
was wrong. The split might have made sense in an early version of the
algorithm, but it's definitely wrong now.
<rdar://problem/
14292693>
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207195
91177308-0d34-0410-b5e6-
96231b3b80d8
Duncan P. N. Exon Smith [Fri, 25 Apr 2014 04:38:40 +0000 (04:38 +0000)]
blockfreq: Document assertion
<rdar://problem/
14292693>
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207194
91177308-0d34-0410-b5e6-
96231b3b80d8
Duncan P. N. Exon Smith [Fri, 25 Apr 2014 04:38:37 +0000 (04:38 +0000)]
blockfreq: Use better branch weights in multiexit test
The branch weights were even before. Make them different.
<rdar://problem/
14292693>
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207193
91177308-0d34-0410-b5e6-
96231b3b80d8
Duncan P. N. Exon Smith [Fri, 25 Apr 2014 04:38:35 +0000 (04:38 +0000)]
blockfreq: Clean up irreducible testcases
Strip irreducible testcases to pure control flow. The function calls
made the branch weights more believable but cluttered it up a lot.
There isn't going to be any constant analysis here, so just use dumb
branch logic to clarify the important parts.
<rdar://problem/
14292693>
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207192
91177308-0d34-0410-b5e6-
96231b3b80d8
Duncan P. N. Exon Smith [Fri, 25 Apr 2014 04:38:32 +0000 (04:38 +0000)]
blockfreq: Document high-level functions
<rdar://problem/
14292693>
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207191
91177308-0d34-0410-b5e6-
96231b3b80d8
Duncan P. N. Exon Smith [Fri, 25 Apr 2014 04:38:30 +0000 (04:38 +0000)]
blockfreq: Remove dead code
<rdar://problem/
14292693>
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207190
91177308-0d34-0410-b5e6-
96231b3b80d8
Duncan P. N. Exon Smith [Fri, 25 Apr 2014 04:38:27 +0000 (04:38 +0000)]
blockfreq: Scale LoopData::Scale on the way down
Rather than scaling loop headers and then scaling all the loop members
by the header frequency, scale `LoopData::Scale` itself, and scale the
loop members by it. It's much more obvious what's going on this way,
and doesn't cost any extra multiplies.
<rdar://problem/
14292693>
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207189
91177308-0d34-0410-b5e6-
96231b3b80d8
Duncan P. N. Exon Smith [Fri, 25 Apr 2014 04:38:25 +0000 (04:38 +0000)]
blockfreq: unwrapLoopPackage() => unwrapLoop()
<rdar://problem/
14292693>
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207188
91177308-0d34-0410-b5e6-
96231b3b80d8
Duncan P. N. Exon Smith [Fri, 25 Apr 2014 04:38:23 +0000 (04:38 +0000)]
blockfreq: Pass the Loop directly into unwrapLoopPackage()
<rdar://problem/
14292693>
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207187
91177308-0d34-0410-b5e6-
96231b3b80d8
Duncan P. N. Exon Smith [Fri, 25 Apr 2014 04:38:20 +0000 (04:38 +0000)]
blockfreq: Unwrap from Loops
When unwrapping loops, just visit the loops rather than all nodes.
<rdar://problem/
14292693>
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207186
91177308-0d34-0410-b5e6-
96231b3b80d8
Duncan P. N. Exon Smith [Fri, 25 Apr 2014 04:38:17 +0000 (04:38 +0000)]
blockfreq: Separate unwrapLoops() from finalizeMetrics()
<rdar://problem/
14292693>
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207185
91177308-0d34-0410-b5e6-
96231b3b80d8
Duncan P. N. Exon Smith [Fri, 25 Apr 2014 04:38:15 +0000 (04:38 +0000)]
blockfreq: LoopData::MemberList => NodeList
<rdar://problem/
14292693>
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207184
91177308-0d34-0410-b5e6-
96231b3b80d8
Duncan P. N. Exon Smith [Fri, 25 Apr 2014 04:38:12 +0000 (04:38 +0000)]
blockfreq: Expose getPackagedNode()
Make `getPackagedNode()` a member function of
`BlockFrequencyInfoImplBase` so that it's available for templated code.
<rdar://problem/
14292693>
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207183
91177308-0d34-0410-b5e6-
96231b3b80d8
Duncan P. N. Exon Smith [Fri, 25 Apr 2014 04:38:09 +0000 (04:38 +0000)]
blockfreq: Store the header with the members
<rdar://problem/
14292693>
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207182
91177308-0d34-0410-b5e6-
96231b3b80d8
Duncan P. N. Exon Smith [Fri, 25 Apr 2014 04:38:06 +0000 (04:38 +0000)]
blockfreq: Encapsulate LoopData::Header
<rdar://problem/
14292693>
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207181
91177308-0d34-0410-b5e6-
96231b3b80d8
Duncan P. N. Exon Smith [Fri, 25 Apr 2014 04:38:03 +0000 (04:38 +0000)]
blockfreq: Embed Loop hierarchy in LoopData
Continue refactoring to make `LoopData` first-class. Here I'm making
the `LoopData` hierarchy explicit, instead of bouncing back and forth
with `WorkingData`. This simplifies the logic and better matches the
`LoopInfo` design. (Eventually, `LoopInfo` should be restructured so
that it supports this pass, and `LoopData` can be removed.)
<rdar://problem/
14292693>
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207180
91177308-0d34-0410-b5e6-
96231b3b80d8
Duncan P. N. Exon Smith [Fri, 25 Apr 2014 04:38:01 +0000 (04:38 +0000)]
blockfreq: Use LoopData directly
Instead of passing around loop headers, pass around `LoopData` directly.
<rdar://problem/
14292693>
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207179
91177308-0d34-0410-b5e6-
96231b3b80d8
Duncan P. N. Exon Smith [Fri, 25 Apr 2014 04:37:58 +0000 (04:37 +0000)]
blockfreq: Stop using range-based for to traverse Loops
A follow-up commit will need the actual iterators.
<rdar://problem/
14292693>
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207178
91177308-0d34-0410-b5e6-
96231b3b80d8
Duncan P. N. Exon Smith [Fri, 25 Apr 2014 04:30:06 +0000 (04:30 +0000)]
blockfreq: Use a std::list for Loops
As pointed out by David Blaikie in code review, a `std::list<T>` is
simpler than a `std::vector<std::unique_ptr<T>>`. Another option is a
`std::deque<T>` (which allocates in chunks), but I'd like to leave open
the option of inserting in the middle of the sequence for handling
irreducible control flow on the fly.
<rdar://problem/
14292693>
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207177
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Fri, 25 Apr 2014 04:24:47 +0000 (04:24 +0000)]
[C++] Use 'nullptr'. Tools edition.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207176
91177308-0d34-0410-b5e6-
96231b3b80d8
Karthik Bhat [Fri, 25 Apr 2014 03:33:48 +0000 (03:33 +0000)]
Allow vectorization of bit intrinsics in BB Vectorizer.
This patch adds support for vectorization of bit intrinsics such as bswap,ctpop,ctlz,cttz.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207174
91177308-0d34-0410-b5e6-
96231b3b80d8
Justin Bogner [Fri, 25 Apr 2014 02:45:33 +0000 (02:45 +0000)]
ProfileData: Treat missing function counts as malformed
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207172
91177308-0d34-0410-b5e6-
96231b3b80d8
Reid Kleckner [Fri, 25 Apr 2014 01:44:20 +0000 (01:44 +0000)]
Change llvm-config --ldflags to report ${CMAKE_CXX_LINK_FLAGS}
Should fix PR19526.
When Oscar added this code in the intial CMake build system port, he had
a TODO saying that ${CMAKE_SHARED_LINKER_FLAGS} was probably wrong. I
agree. I'm using ${CMAKE_CXX_LINK_FLAGS} to point LLVM at my custom
installation of gcc 4.recent, so that seems more correct. With this
change, I can build creduce against an installed clang, and it picks up
the write flags from --ldflags.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207171
91177308-0d34-0410-b5e6-
96231b3b80d8
David Blaikie [Fri, 25 Apr 2014 00:48:01 +0000 (00:48 +0000)]
Fix quadratic performance during debug compression due to sections x symbols iteration.
When fixing the symbols in each compressed section we were iterating
over all symbols for each compressed section. In extreme cases this
could snowball severely (5min uncompressed -> 35min compressed) due to
iterating over all symbols for each compressed section (large numbers of
compressed sections can be generated by DWARF type units).
To address this, build a map of the symbols in each section ahead of
time, and access that map if a section is being compressed. This brings
compile time for the aforementioned example down to ~6 minutes.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207167
91177308-0d34-0410-b5e6-
96231b3b80d8
Adrian Prantl [Fri, 25 Apr 2014 00:42:50 +0000 (00:42 +0000)]
Revert "This reapplies r207130 with an additional testcase+and a missing check for"
Typo in testcase.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207166
91177308-0d34-0410-b5e6-
96231b3b80d8
Adrian Prantl [Fri, 25 Apr 2014 00:38:40 +0000 (00:38 +0000)]
This reapplies r207130 with an additional testcase+and a missing check for
AllocaInst that was missing in one location.
Debug info for optimized code: Support variables that are on the stack and
described by DBG_VALUEs during their lifetime.
Previously, when a variable was at a FrameIndex for any part of its
lifetime, this would shadow all other DBG_VALUEs and only a single
fbreg location would be emitted, which in fact is only valid for a small
range and not the entire lexical scope of the variable. The included
dbg-value-const-byref testcase demonstrates this.
This patch fixes this by
Local
- emitting dbg.value intrinsics for allocas that are passed by reference
- dropping all dbg.declares (they are now fully lowered to dbg.values)
SelectionDAG
- renamed constructors for SDDbgValue for better readability.
- fix UserValue::match() to handle indirect values correctly
- not inserting an MMI table entries for dbg.values that describe allocas.
- lowering dbg.values that describe allocas into *indirect* DBG_VALUEs.
CodeGenPrepare
- leaving dbg.values for an alloca were they are (see comment)
Other
- regenerated/updated instcombine.ll testcase and included source
rdar://problem/
16679879
http://reviews.llvm.org/D3374
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207165
91177308-0d34-0410-b5e6-
96231b3b80d8
Adrian Prantl [Fri, 25 Apr 2014 00:04:49 +0000 (00:04 +0000)]
Revert "Debug info for optimized code: Support variables that are on the stack and"
This reverts commit 207130 for buildbot breakage.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207162
91177308-0d34-0410-b5e6-
96231b3b80d8
Adrian Prantl [Thu, 24 Apr 2014 23:53:29 +0000 (23:53 +0000)]
Revert "Debug info: Let dbg.values inserted by LowerDbgDeclare inherit the location"
This reverts commit 207130 for buildbot breakage.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207159
91177308-0d34-0410-b5e6-
96231b3b80d8
Richard Smith [Thu, 24 Apr 2014 23:29:25 +0000 (23:29 +0000)]
Add missing include, found by modules build.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207158
91177308-0d34-0410-b5e6-
96231b3b80d8
Richard Smith [Thu, 24 Apr 2014 23:14:32 +0000 (23:14 +0000)]
Function defined in a header should be inline. Found by modules build.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207157
91177308-0d34-0410-b5e6-
96231b3b80d8
Alexey Samsonov [Thu, 24 Apr 2014 23:08:56 +0000 (23:08 +0000)]
[DWARF parser] Make a few methods non-public
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207156
91177308-0d34-0410-b5e6-
96231b3b80d8
Alexey Samsonov [Thu, 24 Apr 2014 22:51:03 +0000 (22:51 +0000)]
[DWARF parser] DWARFUnit ctor doesn't need both parsed and raw .debug_abbrev section. Remove the former.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207153
91177308-0d34-0410-b5e6-
96231b3b80d8
Alexey Samsonov [Thu, 24 Apr 2014 22:41:09 +0000 (22:41 +0000)]
[DWARF parser] Simplify and re-format a method
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207151
91177308-0d34-0410-b5e6-
96231b3b80d8
Chandler Carruth [Thu, 24 Apr 2014 21:19:30 +0000 (21:19 +0000)]
[LCG] Switch a weird do/while loop that actually couldn't fail its
condition into an obviously infinite loop with an assert about the
degenerate condition. No functionality changed.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207147
91177308-0d34-0410-b5e6-
96231b3b80d8
Chandler Carruth [Thu, 24 Apr 2014 21:10:35 +0000 (21:10 +0000)]
[ADT] Generalize pointee_iterator to smart pointers by using decltype.
Based on review feedback from Dave on the original patch.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207146
91177308-0d34-0410-b5e6-
96231b3b80d8
Benjamin Kramer [Thu, 24 Apr 2014 20:51:37 +0000 (20:51 +0000)]
X86: Don't transform shifts into ands when the sign bit is tested.
Should unbreak MultiSource/Benchmarks/mediabench/g721/g721encode/encode.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207145
91177308-0d34-0410-b5e6-
96231b3b80d8
Reid Kleckner [Thu, 24 Apr 2014 20:19:22 +0000 (20:19 +0000)]
Remove dead inline function that doesn't compile
MSVC doesn't diagnose this, interestingly.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207144
91177308-0d34-0410-b5e6-
96231b3b80d8
Reid Kleckner [Thu, 24 Apr 2014 20:14:34 +0000 (20:14 +0000)]
Add 'musttail' marker to call instructions
This is similar to the 'tail' marker, except that it guarantees that
tail call optimization will occur. It also comes with convervative IR
verification rules that ensure that tail call optimization is possible.
Reviewers: nicholas
Differential Revision: http://llvm-reviews.chandlerc.com/D3240
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207143
91177308-0d34-0410-b5e6-
96231b3b80d8
Reid Kleckner [Thu, 24 Apr 2014 20:14:08 +0000 (20:14 +0000)]
Fix rdtsc.ll test to match r8 on win64
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207142
91177308-0d34-0410-b5e6-
96231b3b80d8
Richard Smith [Thu, 24 Apr 2014 18:49:15 +0000 (18:49 +0000)]
Remove C++11ism (specializing a template in a surrounding namespace) to appease the buildbots.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207136
91177308-0d34-0410-b5e6-
96231b3b80d8
Adrian Prantl [Thu, 24 Apr 2014 18:44:15 +0000 (18:44 +0000)]
Debug info: Let dbg.values inserted by LowerDbgDeclare inherit the location
of the dbg.value. This gets rid of tons of redundant variable DIEs in
subscopes.
rdar://problem/
14874886, rdar://problem/
16679936
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207135
91177308-0d34-0410-b5e6-
96231b3b80d8
Richard Smith [Thu, 24 Apr 2014 18:27:29 +0000 (18:27 +0000)]
[modules] "Specialize" a function by actually specializing a function template
rather than by adding an overload and hoping that it's declared before the code
that calls it. (In a modules build, it isn't.)
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207133
91177308-0d34-0410-b5e6-
96231b3b80d8
Adrian Prantl [Thu, 24 Apr 2014 17:41:45 +0000 (17:41 +0000)]
Debug info for optimized code: Support variables that are on the stack and
described by DBG_VALUEs during their lifetime.
Previously, when a variable was at a FrameIndex for any part of its
lifetime, this would shadow all other DBG_VALUEs and only a single
fbreg location would be emitted, which in fact is only valid for a small
range and not the entire lexical scope of the variable. The included
dbg-value-const-byref testcase demonstrates this.
This patch fixes this by
Local
- emitting dbg.value intrinsics for allocas that are passed by reference
- dropping all dbg.declares (they are now fully lowered to dbg.values)
SelectionDAG
- renamed constructors for SDDbgValue for better readability.
- fix UserValue::match() to handle indirect values correctly
- not inserting an MMI table entries for dbg.values that describe allocas.
- lowering dbg.values that describe allocas into *indirect* DBG_VALUEs.
CodeGenPrepare
- leaving dbg.values for an alloca were they are (see comment)
Other
- regenerated/updated instcombine-intrinsics testcase and included source
rdar://problem/
16679879
http://reviews.llvm.org/D3374
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207130
91177308-0d34-0410-b5e6-
96231b3b80d8
Andrea Di Biagio [Thu, 24 Apr 2014 17:18:27 +0000 (17:18 +0000)]
[X86] Add support for Read Time Stamp Counter x86 builtin intrinsics.
This patch:
- Adds two new X86 builtin intrinsics ('int_x86_rdtsc' and
'int_x86_rdtscp') as GCCBuiltin intrinsics;
- Teaches the backend how to lower the two new builtins;
- Introduces a common function to lower READCYCLECOUNTER dag nodes
and the two new rdtsc/rdtscp intrinsics;
- Improves (and extends) the existing x86 test 'rdtsc.ll'; now test 'rdtsc.ll'
correctly verifies that both READCYCLECOUNTER and the two new intrinsics
work fine for both 64bit and 32bit Subtargets.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207127
91177308-0d34-0410-b5e6-
96231b3b80d8
Matt Arsenault [Thu, 24 Apr 2014 17:08:26 +0000 (17:08 +0000)]
R600/SI: Use address space in allowsUnalignedMemoryAccesses
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207126
91177308-0d34-0410-b5e6-
96231b3b80d8
David Blaikie [Thu, 24 Apr 2014 16:59:40 +0000 (16:59 +0000)]
Spread some const around for non-mutating uses of MCSymbolData.
I discovered this const-hole while attempting to coalesnce the Symbol
and SymbolMap data structures. There's some pending issues with that,
but I figured this change was easy to flush early.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207124
91177308-0d34-0410-b5e6-
96231b3b80d8
Matheus Almeida [Thu, 24 Apr 2014 16:31:10 +0000 (16:31 +0000)]
[mips] Remove non-ascii character.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207123
91177308-0d34-0410-b5e6-
96231b3b80d8
Tim Northover [Thu, 24 Apr 2014 15:04:26 +0000 (15:04 +0000)]
AArch64/ARM64: add ARM64 runs to more MC tests.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207120
91177308-0d34-0410-b5e6-
96231b3b80d8
Tim Northover [Thu, 24 Apr 2014 15:04:20 +0000 (15:04 +0000)]
AArch64/ARM64: run AArch64 NEON MC tests through ARM64 too.
This skips a couple of compare ones due to the different syntaxt for
floating-point 0.0. AArch64 does it more canonically, and we'll need to fiddle
ARM64 to make it work.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207119
91177308-0d34-0410-b5e6-
96231b3b80d8
David Blaikie [Thu, 24 Apr 2014 14:33:36 +0000 (14:33 +0000)]
Fix memory leak of MCSymbolData in MCAsmStreamer.
Leak identified by LSan and reported by Kostya Serebryany.
Let's get a bit experimental here... in theory our minimum compiler
versions support unordered_map.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207118
91177308-0d34-0410-b5e6-
96231b3b80d8
Tim Northover [Thu, 24 Apr 2014 14:06:20 +0000 (14:06 +0000)]
AArch64: print NEON lists with a space.
This matches ARM64 behaviour, which I think is clearer. It also puts all the
churn from that difference into one easily ignored commit.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207116
91177308-0d34-0410-b5e6-
96231b3b80d8
Evgeniy Stepanov [Thu, 24 Apr 2014 13:29:34 +0000 (13:29 +0000)]
[asan] Use MCInstrInfo in inline asm instrumentation.
Patch by Yuri Gorshenin.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207115
91177308-0d34-0410-b5e6-
96231b3b80d8
Tim Northover [Thu, 24 Apr 2014 12:56:41 +0000 (12:56 +0000)]
AArch64/ARM64: enable remaining MC elf tests.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207112
91177308-0d34-0410-b5e6-
96231b3b80d8
Tim Northover [Thu, 24 Apr 2014 12:56:38 +0000 (12:56 +0000)]
AArch64/ARM64: allow negative addends, at least on ELF.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207111
91177308-0d34-0410-b5e6-
96231b3b80d8
Tim Northover [Thu, 24 Apr 2014 12:56:34 +0000 (12:56 +0000)]
ARM64: support relocated "TBZ/TBNZ" instructions.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207110
91177308-0d34-0410-b5e6-
96231b3b80d8
Tim Northover [Thu, 24 Apr 2014 12:56:30 +0000 (12:56 +0000)]
AArch64/ARM64: support relocated ADR instruction
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207109
91177308-0d34-0410-b5e6-
96231b3b80d8
Tim Northover [Thu, 24 Apr 2014 12:56:27 +0000 (12:56 +0000)]
AArch64/ARM64: add support for :abs_gN_s: MOVZ modifiers
We only need assembly support, so it's fairly easy.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207108
91177308-0d34-0410-b5e6-
96231b3b80d8
Tim Northover [Thu, 24 Apr 2014 12:22:12 +0000 (12:22 +0000)]
ARM64: shut up warning about variable only used in assert.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207106
91177308-0d34-0410-b5e6-
96231b3b80d8
Tim Northover [Thu, 24 Apr 2014 12:12:10 +0000 (12:12 +0000)]
AArch64/ARM64: disentangle the "B.CC" and "LDR lit" operands
These can have different relocations in ELF. In particular both:
b.eq global
ldr x0, global
are valid, giving different relocations. The only possible way to distinguish
them is via a different fixup, so the operands had to be separated throughout
the backend.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207105
91177308-0d34-0410-b5e6-
96231b3b80d8
Tim Northover [Thu, 24 Apr 2014 12:12:01 +0000 (12:12 +0000)]
AArch64/ARM64: enable some MC tests on ARM64
This will also (as with CodeGen) disable testing when the ARM64 backend is not
present.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207104
91177308-0d34-0410-b5e6-
96231b3b80d8
Tim Northover [Thu, 24 Apr 2014 12:11:56 +0000 (12:11 +0000)]
AArch64/ARM64: port bitfield test to ARM64.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207103
91177308-0d34-0410-b5e6-
96231b3b80d8
Tim Northover [Thu, 24 Apr 2014 12:11:53 +0000 (12:11 +0000)]
AArch64/ARM64: implement BFI optimisation
ARM64 was not producing pure BFI instructions for bitfield insertion
operations, unlike AArch64. The approach had to be a little different (in
ISelDAGToDAG rather than ISelLowering), and the outcomes aren't identical but
hopefully this gives it similar power.
This should address PR19424.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207102
91177308-0d34-0410-b5e6-
96231b3b80d8
Tim Northover [Thu, 24 Apr 2014 12:11:46 +0000 (12:11 +0000)]
AArch64/ARM64: port more tests
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207101
91177308-0d34-0410-b5e6-
96231b3b80d8
Chandler Carruth [Thu, 24 Apr 2014 11:05:20 +0000 (11:05 +0000)]
[LCG] Incorporate the core trick of improvements on the naive Tarjan's
algorithm here: http://dl.acm.org/citation.cfm?id=177301.
The idea of isolating the roots has even more relevance when using the
stack not just to implement the DFS but also to implement the recursive
step. Because we use it for the recursive step, to isolate the roots we
need to maintain two stacks: one for our recursive DFS walk, and another
of the nodes that have been walked. The nice thing is that the latter
will be half the size. It also fixes a complete hack where we scanned
backwards over the stack to find the next potential-root to continue
processing. Now that is always the top of the DFS stack.
While this is a really nice improvement already (IMO) it further opens
the door for two important simplifications:
1) De-duplicating some of the code across the two different walks. I've
actually made the duplication a bit worse in some senses with this
patch because the two are starting to converge.
2) Dramatically simplifying the loop structures of both walks.
I wanted to do those separately as they'll be essentially *just* CFG
restructuring. This patch on the other hand actually uses different
datastructures to implement the algorithm itself.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207098
91177308-0d34-0410-b5e6-
96231b3b80d8
Chandler Carruth [Thu, 24 Apr 2014 09:59:59 +0000 (09:59 +0000)]
[LCG] Rotate logic applied to the top of the DFSStack to instead be
applied prior to pushing a node onto the DFSStack. This is the first
step toward avoiding the stack entirely for leaf nodes. It also
simplifies things a bit and I think is pointing the way toward factoring
some more of the shared logic out of the two implementations.
It is also making it more obvious how to restructure the loops
themselves to be a bit easier to read (although no different in terms of
functionality).
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207095
91177308-0d34-0410-b5e6-
96231b3b80d8
Chandler Carruth [Thu, 24 Apr 2014 09:59:56 +0000 (09:59 +0000)]
[LCG] Re-order expectations to provide more useful output when debugging
an issue. This way you see that the number of nodes was wrong before
a crash due to accessing too many nodes.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207094
91177308-0d34-0410-b5e6-
96231b3b80d8
Evgeniy Stepanov [Thu, 24 Apr 2014 09:56:15 +0000 (09:56 +0000)]
[asan] Fix instrumentation of x86 intel syntax inline assembly.
Patch by Yuri Gorshenin.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207092
91177308-0d34-0410-b5e6-
96231b3b80d8
Chandler Carruth [Thu, 24 Apr 2014 09:22:31 +0000 (09:22 +0000)]
[LCG] Switch the parent SCC tracking from a SmallSetVector to
a SmallPtrSet. Currently, there is no need for stable iteration in this
dimension, and I now thing there won't need to be going forward.
If this is ever re-introduced in any form, it needs to not be
a SetVector based solution because removal cannot be linear. There will
be many SCCs with large numbers of parents. When encountering these, the
incremental SCC update for intra-SCC edge removal was quadratic due to
linear removal (kind of).
I'm really hoping we can avoid having an ordering property here at all
though...
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207091
91177308-0d34-0410-b5e6-
96231b3b80d8
Chandler Carruth [Thu, 24 Apr 2014 08:55:36 +0000 (08:55 +0000)]
[LCG] We don't actually need a set in each SCC to track the nodes. We
can use the node -> SCC mapping in the top-level graph to test this on
the rare occasions we need it.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207090
91177308-0d34-0410-b5e6-
96231b3b80d8
Zinovy Nis [Thu, 24 Apr 2014 08:42:58 +0000 (08:42 +0000)]
[CLNUP] Test commit. Remove newline.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207089
91177308-0d34-0410-b5e6-
96231b3b80d8
Benjamin Kramer [Thu, 24 Apr 2014 08:15:31 +0000 (08:15 +0000)]
X86: Emit test instead of constant shift + compare if the shift result is unused.
This allows us to compile
return (mask & 0x8 ? a : b);
into
testb $8, %dil
cmovnel %edx, %esi
instead of
andl $8, %edi
shrl $3, %edi
cmovnel %edx, %esi
which we formed previously because dag combiner canonicalizes setcc of and into shift.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207088
91177308-0d34-0410-b5e6-
96231b3b80d8
Chandler Carruth [Thu, 24 Apr 2014 07:48:18 +0000 (07:48 +0000)]
[LCG] Switch the SCC's parent iterators to be value iterators rather
than pointer iterators.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207086
91177308-0d34-0410-b5e6-
96231b3b80d8
Karthik Bhat [Thu, 24 Apr 2014 07:29:55 +0000 (07:29 +0000)]
Allow vectorization of few missed llvm intrinsic calls in BBVectorizor by handling them in isVectorizableIntrinsic function.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207085
91177308-0d34-0410-b5e6-
96231b3b80d8
Chandler Carruth [Thu, 24 Apr 2014 06:59:50 +0000 (06:59 +0000)]
[ADT] Attempt to appease another MSVC oddity by moving the injected
class name usage into a context we can put typename on it.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207084
91177308-0d34-0410-b5e6-
96231b3b80d8