Rafael Espindola [Mon, 20 Apr 2015 12:44:06 +0000 (12:44 +0000)]
Look past locals in comdats.
We have to avoid converting a reference to a global into a reference to a local,
but it is fine to look past a local.
Patch by Vasileios Kalintiris.
I just moved the comment and added thet test.
llvm-svn: 235300
Daniel Sanders [Mon, 20 Apr 2015 12:44:01 +0000 (12:44 +0000)]
[mips] Support building compiler-rt with gcc.
Summary:
This will allow us to enable it on the buildbot (and then enable a
sanitizer buildbot).
The problem is that gcc for Mips does not accept -m32/-m64 like clang does.
We therefore need to use the nearest equivalent -mips32r2/-mips64r2. We must
also specify -mabi=64 in the -mips64r2 case since gcc's default ABI has
traditionally been N32 whereas clang's is N64.
Reviewers: sagar
Reviewed By: sagar
Subscribers: llvm-commits, samsonov, kcc, mohit.bhakkad, Anand.Takale, sagar
Differential Revision: http://reviews.llvm.org/D8892
llvm-svn: 235299
Jozef Kolek [Mon, 20 Apr 2015 12:42:08 +0000 (12:42 +0000)]
[mips][microMIPSr6] Implement initial mapping support
Differential Revision: http://reviews.llvm.org/D8387
llvm-svn: 235298
Viktor Kutuzov [Mon, 20 Apr 2015 12:25:11 +0000 (12:25 +0000)]
[Asan] Link lit tests shared libraries with -Wl,z,origin on FreeBSD
Differential Revision: http://reviews.llvm.org/D8834
llvm-svn: 235297
Jozef Kolek [Mon, 20 Apr 2015 12:23:06 +0000 (12:23 +0000)]
[mips][microMIPSr6] Implement initial subtarget support
Differential Revision: http://reviews.llvm.org/D8386
llvm-svn: 235296
Andrea Di Biagio [Mon, 20 Apr 2015 11:56:59 +0000 (11:56 +0000)]
[X86][FastIsel] Fix assertion failure when selecting int-to-double conversion (PR23273).
This fixes a regression introduced at revision 231243.
The target-independent selection algorithm in FastISel knows how to select
a SINT_TO_FP if the target is SSE but not AVX. That is because on X86, the
tablegen'd 'fastEmit' functions know how to select CVTSI2SSrr and CVTSI2SDrr.
Method X86FastISel::X86SelectSIToFP was therefore working under the
wrong assumption that the target was AVX. That assumption was incorrect since
we can have a target that is neither AVX nor SSE.
So, rather than asserting for the presence of AVX, we should have had an
early exit from 'X86SelectSIToFP' if the target was not AVX.
This patch fixes the issue replacing the invalid assertion with an early exit.
Thanks to Dimitry Andric for reporting this problem and for providing a small
reproducible testcase. Added test pr23273.ll.
llvm-svn: 235295
Dmitry Vyukov [Mon, 20 Apr 2015 10:35:21 +0000 (10:35 +0000)]
tsan: deflake a test
llvm-svn: 235294
Dmitry Vyukov [Mon, 20 Apr 2015 10:35:10 +0000 (10:35 +0000)]
tsan: fix handling of pthread_detach
Fixes https://llvm.org/bugs/show_bug.cgi?id=23235
If pthread_create is followed by pthread_detach,
the new thread may not acquire synchronize with
the parent thread.
llvm-svn: 235293
Tamas Berghammer [Mon, 20 Apr 2015 10:31:22 +0000 (10:31 +0000)]
Auto advance pc for signle stepping on arm when emulation failes
The arm instruction emulation handles only some of the opcode (including
all of them modifying the PC). For the rest of the instructions we can
advance the PC by the size of the instruction as they don't modify the
PC on any other way.
Differential revision: http://reviews.llvm.org/D9076
llvm-svn: 235292
Tamas Berghammer [Mon, 20 Apr 2015 09:52:47 +0000 (09:52 +0000)]
Stop read thread of Communication on EOF
Previously the read thread was only stopped if CloseOnEOF was set on the
communication channel. It caused it to spin in case of an EOF because
::select() always reported that we can read from the file descriptor.
This CL change this behavior with stopping the read thread on EOF but do
a disconnect only if CloseOnEOF is enabled.
Differential revision: http://reviews.llvm.org/D9077
llvm-svn: 235291
Simon Atanasyan [Mon, 20 Apr 2015 07:36:27 +0000 (07:36 +0000)]
[Mips] Write DT_MIPS_OPTIONS tag to .dynamic section
If .MIPS.options section exists the DT_MIPS_OPTIONS tag should hold
an address of this section.
llvm-svn: 235290
Simon Atanasyan [Mon, 20 Apr 2015 07:34:52 +0000 (07:34 +0000)]
[Mips] Support R_MICROMIPS_PC18_S3/PC19_S2/PC21_S2/PC26_S2 relocations handling
llvm-svn: 235289
Simon Atanasyan [Mon, 20 Apr 2015 07:33:02 +0000 (07:33 +0000)]
[Mips] Support R_MICROMIPS_GOT_DISP / PAGE / OFST relocations handling
llvm-svn: 235288
Karthik Bhat [Mon, 20 Apr 2015 07:07:10 +0000 (07:07 +0000)]
Fix buildbot failure on darwin from r235284.
llvm-svn: 235287
Manuel Klimek [Mon, 20 Apr 2015 06:58:56 +0000 (06:58 +0000)]
Fix bug in Replacement's toString on Windows (missing flush).
Adapt function to LLVM coding style.
llvm-svn: 235286
Simon Atanasyan [Mon, 20 Apr 2015 05:34:48 +0000 (05:34 +0000)]
[Mips] Support DT_MIPS_OPTIONS dynamic section tag in the llvm-readobj
llvm-svn: 235285
Karthik Bhat [Mon, 20 Apr 2015 04:38:33 +0000 (04:38 +0000)]
[NFC] Refactor identification of reductions as common utility function.
This patch refactors reduction identification code out of LoopVectorizer and
exposes them as common utilities.
No functional change.
Review: http://reviews.llvm.org/D9046
llvm-svn: 235284
Hal Finkel [Mon, 20 Apr 2015 00:01:30 +0000 (00:01 +0000)]
[InlineAsm] Remove EarlyClobber on registers that are also inputs
When an inline asm call has an output register marked as early-clobber, but
that same register is also an input operand, what should we do? GCC accepts
this, and is documented to accept this for read/write operands saying,
"Furthermore, if the earlyclobber operand is also a read/write operand, then
that operand is written only after it's used." For write-only operands, the
situation seems less clear, but I have at least one existing codebase that
assumes this will work, in part because it has syscall macros like this:
({ \
register uint64_t r0 __asm__ ("r0") = (__NR_ ## name); \
register uint64_t r3 __asm__ ("r3") = ((uint64_t) (arg0)); \
register uint64_t r4 __asm__ ("r4") = ((uint64_t) (arg1)); \
register uint64_t r5 __asm__ ("r5") = ((uint64_t) (arg2)); \
__asm__ __volatile__ \
("sc" \
: "=&r"(r0),"=&r"(r3),"=&r"(r4),"=&r"(r5) \
: "0"(r0), "1"(r3), "2"(r4), "3"(r5) \
: "r6","r7","r8","r9","r10","r11","r12","cr0","memory"); \
r3; \
})
Furthermore, with register aliases and subregister relationships that only the
backend knows about, rejecting this in the frontend seems like a difficult
proposition (if we wanted to do so). However, keeping the early-clobber flag on
the INLINEASM MI does not work for us, because it will cause the register's
live interval to end to soon (so it will not appear defined to be used as an
input).
Fortunately, fixing this does not seem hard: When forming the INLINEASM MI,
check to see if any of the early-clobber outputs are also inputs, and if so,
remove the early-clobber flag.
llvm-svn: 235283
Davide Italiano [Sun, 19 Apr 2015 23:45:41 +0000 (23:45 +0000)]
Accept --dynamic-linker <file>.
Before we only accepted --dynamic-linker=<file> and -dynamic-linker <file>
but older versions of GNU ld (e.g. 2.17.50) accept this other form, so
try to be compatible.
PR: 23233
llvm-svn: 235282
Simon Pilgrim [Sun, 19 Apr 2015 22:16:49 +0000 (22:16 +0000)]
[X86][SSE] Fix for getScalarValueForVectorElement to detect scalar sources requiring truncation.
The fix ensures that scalar sources inserted into a vector are the correct bit size.
Integer scalar sources from BUILD_VECTOR and SCALAR_TO_VECTOR nodes may require truncation that this function doesn't currently support.
llvm-svn: 235281
Omair Javaid [Sun, 19 Apr 2015 21:36:06 +0000 (21:36 +0000)]
Fix LLDB ARM GCC4.7 broken build
llvm-svn: 235280
Simon Pilgrim [Sun, 19 Apr 2015 21:34:57 +0000 (21:34 +0000)]
[X86][SSE] Extended copysign tests to include llvm intrinsic implementation and constant folding.
llvm-svn: 235279
Simon Atanasyan [Sun, 19 Apr 2015 20:51:55 +0000 (20:51 +0000)]
[mips] Update MIPS relocations list
No functional changes.
llvm-svn: 235278
Hal Finkel [Sun, 19 Apr 2015 20:16:13 +0000 (20:16 +0000)]
Make ubsan respect log_path
As with the other sanitizers, it is desirable to allow ubsan's output to be
redirected to somewhere other than stderr (and into per-process log files).
llvm-svn: 235277
Argyrios Kyrtzidis [Sun, 19 Apr 2015 20:15:55 +0000 (20:15 +0000)]
[Sema] Check availability of ObjC super class and protocols of a container
in the context of the container itself.
Otherwise we will emit 'unavailable' errors when referencing an unavailable super class
even though the subclass is also marked 'unavailable'.
rdar://
20598702
llvm-svn: 235276
Ed Maste [Sun, 19 Apr 2015 18:27:00 +0000 (18:27 +0000)]
Correct machine type for 32-bit arm
Reported by Andrew Turner.
llvm-svn: 235275
Eric Fiselier [Sun, 19 Apr 2015 15:32:52 +0000 (15:32 +0000)]
Remove constexpr support for std::apply because it introduces regressions.
llvm-svn: 235274
Tobias Grosser [Sun, 19 Apr 2015 09:06:02 +0000 (09:06 +0000)]
Update isl to
a6523fb42c35
A minor update mostly documenting the isl coding style as well as adding
functions to inspect isl_schedule_constraints objects.
llvm-svn: 235273
David Majnemer [Sun, 19 Apr 2015 07:53:29 +0000 (07:53 +0000)]
[MS ABI] Correctly associate align attrs before the class-key
__declspec(align(...)) is unlike all other attributes in that it is not
applied to a variable if it appears before the class-key. If the
tag in question isn't part of a variable declaration, it is not ignored.
Instead, the alignment attribute is applied to the tag.
This fixes PR18024.
llvm-svn: 235272
Tobias Grosser [Sun, 19 Apr 2015 07:07:26 +0000 (07:07 +0000)]
Dependences: Allow to disable dependences computeout
llvm-svn: 235271
Bob Wilson [Sun, 19 Apr 2015 04:09:44 +0000 (04:09 +0000)]
Include __mulodi4 in the 32-bit ARM versions of libclang_rt.cc_kext.a.
This is needed when building the Darwin kernel with
-fsanitize=undefined-trap. rdar://problem/
12783702
llvm-svn: 235270
Eric Christopher [Sun, 19 Apr 2015 03:21:08 +0000 (03:21 +0000)]
Remove the FCFI option from TargetOptions as it is currently unused.
llvm-svn: 235269
Eric Christopher [Sun, 19 Apr 2015 03:21:04 +0000 (03:21 +0000)]
Remove CFIFuncName from TargetOptions as it is currently unused.
llvm-svn: 235268
Eric Christopher [Sun, 19 Apr 2015 03:20:59 +0000 (03:20 +0000)]
Remove the CFIEnforcing flag from TargetOptions as it is unused.
llvm-svn: 235267
Eric Christopher [Sun, 19 Apr 2015 03:20:55 +0000 (03:20 +0000)]
Remove the CFIType TargetOption as it is unused.
llvm-svn: 235266
Eric Christopher [Sun, 19 Apr 2015 03:20:51 +0000 (03:20 +0000)]
Remove the JITEmitDebugInfo TargetOptions as they're only set and
not used anywhere in llvm.
llvm-svn: 235265
Richard Smith [Sun, 19 Apr 2015 01:47:53 +0000 (01:47 +0000)]
Try to work around failure to convert this lambda to a function pointer in some versions of GCC.
llvm-svn: 235264
Richard Smith [Sun, 19 Apr 2015 01:34:23 +0000 (01:34 +0000)]
[modules] Refactor macro emission. No functionality change.
llvm-svn: 235263
Craig Topper [Sun, 19 Apr 2015 00:57:33 +0000 (00:57 +0000)]
Remove unnecessary include and probably a layering violation.
llvm-svn: 235262
Ahmed Bougacha [Sat, 18 Apr 2015 23:06:04 +0000 (23:06 +0000)]
[MemCpyOpt] Don't force i64 when promoting memset/memcpy sizes.
Harden r235258 to support any integer bitwidth. The quick glance at
the reference made me think only i32 and i64 were valid types, but
they're not special, so any overload is legal.
Thanks to David Majnemer for noticing!
llvm-svn: 235261
Simon Pilgrim [Sat, 18 Apr 2015 21:24:16 +0000 (21:24 +0000)]
[X86][AVX2] Force execution domain on broadcast folding tests.
llvm-svn: 235260
Simon Pilgrim [Sat, 18 Apr 2015 18:50:55 +0000 (18:50 +0000)]
[X86][SSE] Force execution domain on float/double unpack shuffle tests.
llvm-svn: 235259
Ahmed Bougacha [Sat, 18 Apr 2015 17:57:41 +0000 (17:57 +0000)]
[MemCpyOpt] Promote both memset/memcpy sizes if differently typed.
Followup to r235232, which caused PR23278.
We can't assume the memset and memcpy sizes have the same type, as
nothing in the language reference prevents that.
Instead, zext both to i64 if they disagree.
While there, robustify tests by using i8 %c rather than i8 0 for the
memset character.
llvm-svn: 235258
Benjamin Kramer [Sat, 18 Apr 2015 16:52:08 +0000 (16:52 +0000)]
[InstCombine] Create zero constants on demand.
No functional change intended.
llvm-svn: 235257
Tobias Grosser [Sat, 18 Apr 2015 15:57:56 +0000 (15:57 +0000)]
Drop unused extern declaration
llvm-svn: 235256
Oleksiy Vyalov [Sat, 18 Apr 2015 14:08:16 +0000 (14:08 +0000)]
Remove usages of host architecture within ELF object file for unknown parts of ArchSpec.
http://reviews.llvm.org/D9078
llvm-svn: 235255
Tobias Grosser [Sat, 18 Apr 2015 11:01:25 +0000 (11:01 +0000)]
Clear InsnToMemAcc map in ScopDetection after each runOnFunction
Otherwise, instructions in different functions that share the same pointer (due
to earlier modifications), might get assigned incorrect memory access
information (belonging to instructions in previous functions), which can result
in arbitrary memory corruption and assertion failures.
This fixes llvm.org/PR23160 and possibly also llvm.org/PR23167.
Note: InsnToMemAcc is a global variable that should never have existed in the
first place. We will clean this up in a subsequent patch.
Reported-by: Jeremy Huddleston Sequoia <jeremyhu@apple.com>
Debugged-by: Johannes Doerfert <doerfert@cs.uni-saarland.de>
llvm-svn: 235254
Benjamin Kramer [Sat, 18 Apr 2015 10:00:10 +0000 (10:00 +0000)]
Make helper function static. NFC.
llvm-svn: 235253
Tobias Grosser [Sat, 18 Apr 2015 09:40:01 +0000 (09:40 +0000)]
Fix typo
llvm-svn: 235252
David Majnemer [Sat, 18 Apr 2015 04:55:51 +0000 (04:55 +0000)]
[Sema] Don't crash if array bound calculation overflowed constexpr array
We didn't correctly expect a QualifiedTypeLoc when faced with fixing a
variable array type into a constant array type.
Differential Revision: http://reviews.llvm.org/D8958
llvm-svn: 235251
David Majnemer [Sat, 18 Apr 2015 04:41:30 +0000 (04:41 +0000)]
[InstCombine] (mul nsw 1, INT_MIN) != (shl nsw 1, 31)
Multiplying INT_MIN by 1 doesn't trigger nsw. However, shifting 1 into
the sign bit *does* trigger nsw.
llvm-svn: 235250
Ahmed Bougacha [Sat, 18 Apr 2015 01:21:58 +0000 (01:21 +0000)]
[GlobalMerge] Look at uses to create smaller global sets.
Instead of merging everything together, look at the users of
GlobalVariables, and try to group them by function, to create
sets of globals used "together".
Using that information, a less-aggressive alternative is to keep merging
everything together *except* globals that are only ever used alone, that
is, those for which it's clearly non-profitable to merge with others.
In my testing, grouping by Function is too aggressive, but grouping by
BasicBlock is too conservative. Anything in-between isn't trivially
available, so stick with Function grouping for now.
cl::opts are added for testing; both enabled by default.
A few of the testcases aren't testing the merging proper, but just
various edge cases when merging does occur. Update them to use the
previous grouping behavior. Also, one of the tests is unrelated to
GlobalMerge; change it accordingly.
While there, switch to r234666' flags rather than the brutal -O3.
Differential Revision: http://reviews.llvm.org/D8070
llvm-svn: 235249
Duncan P. N. Exon Smith [Sat, 18 Apr 2015 00:35:36 +0000 (00:35 +0000)]
DebugInfo: Delete DIDescriptor (but not its subclasses)
Delete `DIDescriptor` and update the remaining users. I'll follow-up by
deleting subclasses in manageable groups (top-down).
llvm-svn: 235248
Davide Italiano [Sat, 18 Apr 2015 00:32:38 +0000 (00:32 +0000)]
[GNU/ELF] Add support for -x/--discard-all.
PR: 23232
llvm-svn: 235247
Duncan P. N. Exon Smith [Sat, 18 Apr 2015 00:26:49 +0000 (00:26 +0000)]
Remove dead code, NFC
llvm-svn: 235246
Duncan P. N. Exon Smith [Sat, 18 Apr 2015 00:07:30 +0000 (00:07 +0000)]
DebugInfo: Prepare for deletion of DIDescriptor
Prepare for upcoming LLVM change to delete `DIDescriptor`.
llvm-svn: 235245
Duncan P. N. Exon Smith [Sat, 18 Apr 2015 00:01:35 +0000 (00:01 +0000)]
docs: Update Kaleidoscope for recent DI changes
This has been bit-rotting, so fix it up. I'll have to edit this again
once the MD* classes have been renamed to DI* -- I'll try to remember to
do that with the commit that renames them.
llvm-svn: 235244
Ahmed Bougacha [Fri, 17 Apr 2015 23:43:33 +0000 (23:43 +0000)]
[AArch64] Don't force MVT::Untyped when selecting LD1LANEpost.
The result is either an Untyped reg sequence, on ldN with N > 1, or
just the type of the input vector, on ld1. Don't force Untyped.
Instead, just use the type of the reg sequence.
This mirrors the behavior of createTuple, which feeds the LD1*_POST.
The narrow code path wasn't actually covered by tests, because V64
insert_vector_elt are widened to V128 before the LD1LANEpost combine
has the chance to run, usually.
The only case where it does run on V64 vectors is if the vector ops
legalizer ran. So, tickle the code with a ctpop.
Fixes PR23265.
llvm-svn: 235243
Ahmed Bougacha [Fri, 17 Apr 2015 23:38:46 +0000 (23:38 +0000)]
Fix another typo in r235224 testcase. NFC.
Third time's the charm!
llvm-svn: 235242
Andrew Kaylor [Fri, 17 Apr 2015 23:20:24 +0000 (23:20 +0000)]
Fix build wanrings and line endings
llvm-svn: 235241
Duncan P. N. Exon Smith [Fri, 17 Apr 2015 23:20:10 +0000 (23:20 +0000)]
DebugInfo: Remove DIDescriptor from the DebugInfo API
Stop using `DIDescriptor` and its subclasses in the `DebugInfoFinder`
API, as well as the rest of the API hanging around in `DebugInfo.h`.
llvm-svn: 235240
Andrew Kaylor [Fri, 17 Apr 2015 23:05:43 +0000 (23:05 +0000)]
[WinEH] Fixes for a few cppeh failures.
Differential Review: http://reviews.llvm.org/D9065
llvm-svn: 235239
Adam Nemet [Fri, 17 Apr 2015 22:43:10 +0000 (22:43 +0000)]
[LoopAccesses] Improve debug output
llvm-svn: 235238
Zachary Turner [Fri, 17 Apr 2015 22:40:36 +0000 (22:40 +0000)]
[PDB] Support executables and source/line info.
Previously DebugInfoPDB could only load data for a PDB given a
path to the PDB. It could not open an EXE and find the matching
PDB and verify it matched, etc. This patch adds support for that
so that we can simply load debug information for a PDB directly.
Additionally, this patch extends DebugInfoPDB to support getting
source and line information for symbols.
llvm-svn: 235237
David Blaikie [Fri, 17 Apr 2015 22:32:20 +0000 (22:32 +0000)]
[opaque pointer type] Access the pointee of the result type from the GEP rather than pulling it out of the pointer result type
The implementation of this GEP::getResultElementType will be refactored
to either rely on a member variable, or recompute the value from the
indicies (any preferences?).
llvm-svn: 235236
David Blaikie [Fri, 17 Apr 2015 22:32:17 +0000 (22:32 +0000)]
[opaque pointer type] Query the GEP for its source element type directly rather than finding it through the pointer type of the first operand in the Verifier
llvm-svn: 235235
David Blaikie [Fri, 17 Apr 2015 22:32:16 +0000 (22:32 +0000)]
[opaque pointer type] Avoid creating (and then unwrapping) a pointer type to compute the result type of a GEP
llvm-svn: 235234
David Blaikie [Fri, 17 Apr 2015 22:32:13 +0000 (22:32 +0000)]
[opaque pointer type] Use the parsed explicit pointee type when error-checking geps during LL parsing
llvm-svn: 235233
Ahmed Bougacha [Fri, 17 Apr 2015 22:20:57 +0000 (22:20 +0000)]
[MemCpyOpt] Optimize double-storing by memset+memcpy.
A common idiom in some code is to do the following:
memset(dst, 0, dst_size);
memcpy(dst, src, src_size);
Some of the memset is redundant; instead, we can do:
memcpy(dst, src, src_size);
memset(dst + src_size, 0,
dst_size <= src_size ? 0 : dst_size - src_size);
Original patch by: Joel Jones
Differential Revision: http://reviews.llvm.org/D498
llvm-svn: 235232
Pete Cooper [Fri, 17 Apr 2015 22:16:24 +0000 (22:16 +0000)]
Change AArch64 i128 returns to use [2 x i64] when possible.
Something like { void*, void * } would be passed to a function as a [2 x i64], but returned as an i128. This patch unifies the 2 behaviours so that we also return it as a [2 x i64].
This is better for the quality of the IR, and the size of the final LLVM binary as we tend to want to insert/extract values from these types and do so with the insert/extract instructions is less IR than shifting, truncating, and or'ing values.
Reviewed by Tim Northover.
llvm-svn: 235231
Sean Silva [Fri, 17 Apr 2015 21:58:55 +0000 (21:58 +0000)]
[LangRef] Remove redundant and inconsistent condition.
Just above, 'op2' is stated to be unsigned, so 'negative' doesn't make
sense (and is handled by "larger than" anyway). The descriptions for
lshr and ashr don't say 'negative or' either.
llvm-svn: 235230
Duncan P. N. Exon Smith [Fri, 17 Apr 2015 21:34:47 +0000 (21:34 +0000)]
AsmPrinter: Create a unified .debug_loc stream
This commit removes `DebugLocList` and replaces it with
`DebugLocStream`.
- `DebugLocEntry` no longer contains its byte/comment streams.
- The `DebugLocEntry` list for a variable/inlined-at pair is allocated
on the stack, and released right after `DebugLocEntry::finalize()`
(possible because of the refactoring in r231023). Now, only one
list is in memory at a time now.
- There's a single unified stream for the `.debug_loc` section that
persists, stored in the new `DebugLocStream` data structure.
The last point is important: this collapses the nested `SmallVector<>`s
from `DebugLocList` into unified streams. We previously had something
like the following:
vec<tuple<Label, CU,
vec<tuple<BeginSym, EndSym,
vec<Value>,
vec<char>,
vec<string>>>>>
A `SmallVector` can avoid allocations, but is statically fairly large
for a vector: three pointers plus the size of the small storage, which
is the number of elements in small mode times the element size).
Nesting these is expensive, since an inner vector's size contributes to
the element size of an outer one. (Nesting any vector is expensive...)
In the old data structure, the outer vector's *element* size was 632B,
excluding allocation costs for when the middle and inner vectors
exceeded their small sizes. 312B of this was for the "three" pointers
in the vector-tree beneath it. If you assume 1M functions with an
average of 10 variable/inlined-at pairs each (in an LTO scenario),
that's almost 6GB (besides inner allocations), with almost 3GB for the
"three" pointers.
This came up in a heap profile a little while ago of a `clang -flto -g`
bootstrap, with `DwarfDebug::collectVariableInfo()` using something like
10-15% of the total memory.
With this commit, we have:
tuple<vec<tuple<Label, CU, Offset>>,
vec<tuple<BeginSym, EndSym, Offset, Offset>>,
vec<char>,
vec<string>>
The offsets are used to create `ArrayRef` slices of adjacent
`SmallVector`s. This reduces the number of vectors to four (unrelated
to the number of variable/inlined-at pairs), and caps the number of
allocations at the same number.
Besides saving memory and limiting allocations, this is NFC.
I don't know my way around this code very well yet, but I wonder if we
could go further: why stream to a side-table, instead of directly to the
output stream?
llvm-svn: 235229
Pete Cooper [Fri, 17 Apr 2015 21:31:25 +0000 (21:31 +0000)]
AArch64: Add test for returning [2 x i64] in registers. NFC.
llvm-svn: 235228
Rafael Espindola [Fri, 17 Apr 2015 21:15:17 +0000 (21:15 +0000)]
Compute A-B when A or B is weak.
Similar to r235222, but for the weak symbol case.
In an "ideal" assembler/object format an expression would always refer to the
final value and A-B would only be computed from a section in the same
comdat as A and B with A and B strong.
Unfortunately that is not the case with debug info on ELF, so we need an
heuristic. Since we need an heuristic, we may as well use the same one as
gas:
* call weak_sym : produces a relocation, even if in the same section.
* A - weak_sym and weak_sym -A: don't produce a relocation if we can
compute it.
This fixes pr23272 and changes the fix of pr22815 to match what gas does.
llvm-svn: 235227
Ahmed Bougacha [Fri, 17 Apr 2015 21:11:58 +0000 (21:11 +0000)]
Fix typo in r235224 testcase. NFC.
llvm-svn: 235226
Duncan P. N. Exon Smith [Fri, 17 Apr 2015 21:06:49 +0000 (21:06 +0000)]
Remove dead code, NFC
llvm-svn: 235225
Ahmed Bougacha [Fri, 17 Apr 2015 21:02:30 +0000 (21:02 +0000)]
[AArch64] Avoid vector->load dependency cycles when creating LD1*post.
They would break the SelectionDAG.
Note that the opposite load->vector dependency is already obvious in:
(LD1*post vec, ..)
llvm-svn: 235224
David Majnemer [Fri, 17 Apr 2015 20:12:09 +0000 (20:12 +0000)]
[WinEH] Reusing HandlerType entries leads to small CatchHigh values
CatchHigh may be smaller than TryHigh if we reuse an outlined catch
handler for two different invokes with different EH states. We have no
evidence which shows that CatchHigh must be greater than TryHigh or
TryLow. We can revisit this if we turn out to be wrong.
llvm-svn: 235223
Rafael Espindola [Fri, 17 Apr 2015 20:05:17 +0000 (20:05 +0000)]
Compute A-B if both A and B are in the same comdat section.
Part of pr23272.
A small annoyance with the assembly syntax we implement is that given an
expression there is no way to know if what is desired is the value of that
expression for the symbols in this file or for the final values of those
symbols in a link.
The first case is useful for use in sections that get discarded or ignored
if the section they are describing is discarded.
For axample, consider A-B where A and B are in the same comdat section.
We can compute the value of the difference in the section that is present in
the current .o and if that section survives to the final DSO the value will
still will be correct.
But the section is in a comdat. Another section from another object file
might be used istead. We know that that section will define A and B, but
we have no idea what the value of A-B might be.
In practice we have to assume that the intention is to compute the value
in the current section since otherwise the is no way to create something like
the debug aranges section.
llvm-svn: 235222
David Blaikie [Fri, 17 Apr 2015 19:56:21 +0000 (19:56 +0000)]
[opaque pointer types] Use the pointee type loaded from bitcode when constructing a LoadInst
Now (with a few carefully placed suppressions relating to general type
serialization, etc) we can round trip a simple load through bitcode and
textual IR without calling getElementType on a PointerType.
llvm-svn: 235221
Pirama Arumuga Nainar [Fri, 17 Apr 2015 19:51:44 +0000 (19:51 +0000)]
Fix build errors introduced by r235215
Summary:
- Handle TypePromoteFloat in switch statements
- Move an expression into an assert to avoid unused variable in
non-assert builds.
Reviewers: srhines, ab
Subscribers: llvm-commits
Differential Revision: http://reviews.llvm.org/D9086
llvm-svn: 235220
Aaron Ballman [Fri, 17 Apr 2015 19:35:44 +0000 (19:35 +0000)]
Suppressing the C4324 warnings generated by MSVC. This is the only declarative instance that would generate the warning, but it accounted for about 525+ warnings due to template instantiations. This is a marginal-value warning which we may decide to disable more broadly, but since this header is in Support and may be used out of tree, it's a low burden for us to be warning-free in this case.
llvm-svn: 235219
Jason Molenda [Fri, 17 Apr 2015 19:15:02 +0000 (19:15 +0000)]
Whitespace-only tweaks to Colin's r235109 patch to match the lldb
coding style a little more closely.
llvm-svn: 235218
Rui Ueyama [Fri, 17 Apr 2015 18:58:34 +0000 (18:58 +0000)]
Delete dead code (again).
llvm-svn: 235217
Rui Ueyama [Fri, 17 Apr 2015 18:54:29 +0000 (18:54 +0000)]
Delete dead code.
llvm-svn: 235216
Pirama Arumuga Nainar [Fri, 17 Apr 2015 18:36:25 +0000 (18:36 +0000)]
Add support to promote f16 to f32
Summary:
This patch adds legalization support to operate on FP16 as a load/store type
and do operations on it as floats.
Tests for ARM are added to test/CodeGen/ARM/fp16-promote.ll
Reviewers: srhines, t.p.northover
Differential Revision: http://reviews.llvm.org/D8755
llvm-svn: 235215
Chaoren Lin [Fri, 17 Apr 2015 17:33:43 +0000 (17:33 +0000)]
TestMiBreak - func isn't always ??.
Summary: Caused a failure because it showed up as `__printf`.
Reviewers: ki.stfu
Subscribers: lldb-commits
Differential Revision: http://reviews.llvm.org/D9060
llvm-svn: 235214
Vasileios Kalintiris [Fri, 17 Apr 2015 17:29:58 +0000 (17:29 +0000)]
[mips][FastISel] Implement FastMaterializeAlloca in Mips fast-isel.
Summary: Implement the method FastMaterializeAlloca in Mips fast-isel
Based on a patch by Reed Kotler.
Test Plan:
Passes test-suite at O0/O2 for mips32 r1/r2
fastalloca.ll
Reviewers: dsanders, rkotler
Subscribers: rfuhler, llvm-commits
Differential Revision: http://reviews.llvm.org/D6742
llvm-svn: 235213
David Majnemer [Fri, 17 Apr 2015 17:20:30 +0000 (17:20 +0000)]
[WinEH] Allow CatchHigh to be equal to TryHigh
Catch blocks which are empty may be in the same state as their try
blocks. It is not meaningful to give the catch block its own state
number in this case because it can't do anything exceptional.
llvm-svn: 235212
Manman Ren [Fri, 17 Apr 2015 17:10:09 +0000 (17:10 +0000)]
[LTO API] add lto_codegen_set_should_internalize.
When debugging LTO issues with ld64, we use -save-temps to save the merged
optimized bitcode file, then invoke ld64 again on the single bitcode file.
The saved bitcode file is already internalized, so we can call
lto_codegen_set_should_internalize and skip running internalization again.
rdar://
20227235
llvm-svn: 235211
Sanjay Patel [Fri, 17 Apr 2015 17:02:37 +0000 (17:02 +0000)]
[X86, AVX] add an exedepfix entry for vmovq == vmovlps == vmovlpd
This is the AVX extension of r235014:
http://llvm.org/viewvc/llvm-project?view=revision&revision=235014
Review:
http://reviews.llvm.org/D8691
llvm-svn: 235210
Pete Cooper [Fri, 17 Apr 2015 17:00:16 +0000 (17:00 +0000)]
Add a missing forward def of CheckerManager. NFC.
This file doesn't include CheckerManager or forward declare it, so is sensitive to include order.
llvm-svn: 235209
Ilia K [Fri, 17 Apr 2015 16:55:36 +0000 (16:55 +0000)]
Fix the MiVarTestCase.test_lldbmi_var_create_register test to expect 32bit value on 32bit systems
llvm-svn: 235204
Duncan P. N. Exon Smith [Fri, 17 Apr 2015 16:36:10 +0000 (16:36 +0000)]
AsmPrinter: Store MDExpression directly instead of MDNode, NFC
Clean up `DebugLocEntry::Value::Expression`'s type while I'm messing
around in here anyway.
llvm-svn: 235203
Duncan P. N. Exon Smith [Fri, 17 Apr 2015 16:33:37 +0000 (16:33 +0000)]
AsmPrinter: Stop storing MDLocalVariable in DebugLocEntry
Stop storing the `MDLocalVariable` in the `DebugLocEntry::Value`s. We
generate the list of `DebugLocEntry`s separately for each
variable/inlined-at pair, so the variable never actually changes here.
This is effectively NFC (aside from saving some memory and CPU time).
llvm-svn: 235202
Duncan P. N. Exon Smith [Fri, 17 Apr 2015 16:28:58 +0000 (16:28 +0000)]
AsmPrinter: Calculate type upfront for location lists, NFC
We can calculate the variable type up front before calling
`DebugLocEntry::finalize()`. In fact, since we only care about the type
if it's an `MDBasicType`, don't even bother resolving it using the type
identifier map.
llvm-svn: 235201
David Blaikie [Fri, 17 Apr 2015 16:28:26 +0000 (16:28 +0000)]
[opaque pointer type] Serialize the type of an llvm::Function as a function type rather than a function pointer type
llvm-svn: 235200
Aidan Dodds [Fri, 17 Apr 2015 16:12:58 +0000 (16:12 +0000)]
test commit.
llvm-svn: 235199
Kit Barton [Fri, 17 Apr 2015 16:11:05 +0000 (16:11 +0000)]
Add support for v1i128 type.
The v1i128 type is needed for the quadword add/substract instructions introduced
in POWER8. Futhermore, the PowerPC ABI specifies that parameters of type v1i128
are to be passed in a single vector register, while parameters of type i128 are
passed in pairs of GPRs. Thus, it is necessary to be able to differentiate
between v1i128 and i128 in LLVM.
http://reviews.llvm.org/D8564
llvm-svn: 235198
Ilia K [Fri, 17 Apr 2015 15:45:32 +0000 (15:45 +0000)]
Fix the MiVarTestCase.test_lldbmi_var_create_register test on 32bit systems
llvm-svn: 235197