dadb1571b7
Fix bug with APInt::getBitsNeeded with for base 10 numbers 0-9.
...
llvm-svn: 79593
2009-08-21 03:15:28 +00:00
1264bcb4de
Allow '+' to appear in APInt strings, and add more unit tests.
...
llvm-svn: 79592
2009-08-21 03:15:14 +00:00
51c7aa04ec
Remove Neon intrinsics for VZIP, VUZP, and VTRN. We will represent these as
...
vector shuffles. Temporarily remove the tests for these operations until the
new implementation is working.
llvm-svn: 79579
2009-08-21 00:01:42 +00:00
baa922cf16
Re-revert r79555. Apparently it's not just buildbot weirdness.
...
llvm-svn: 79578
2009-08-20 23:51:44 +00:00
da666c801c
Add support for including '+' in APFloat strings, more asserts,
...
and many new unit tests.
llvm-svn: 79574
2009-08-20 23:30:43 +00:00
9ca906e69d
Reapply r79555 for testing. Daniel's trying to work out some buildbot weirdnesss.
...
llvm-svn: 79572
2009-08-20 23:14:20 +00:00
e3836f99bf
--- Reverse-merging r79555 into '.':
...
U include/llvm/Target/TargetData.h
U lib/Target/TargetData.cpp
Temporarily revert 79555. It was causing hangs and test failures.
llvm-svn: 79568
2009-08-20 22:04:42 +00:00
0e749fe36b
Attempt to comment this code more.
...
llvm-svn: 79567
2009-08-20 22:02:24 +00:00
e7e865cf0b
Suppress build warning in -Asserts
...
llvm-svn: 79564
2009-08-20 20:01:34 +00:00
a874d82492
part of the previous commit for PIC16 ISR implementation.
...
llvm-svn: 79563
2009-08-20 19:34:18 +00:00
686e9f55dc
Implement support for ISRs.
...
Clone functions that are shared between the Main thread and Interrupt thread.
CallSites are changed in AsmPrinter currently. A better solution would have been to modify the legalizer (SoftenFloat) to allow targets to change the name of libcalls for float operations. But that currently breaks other targets.
Also, cloing of automatic variables is done AsmPrinter, a better approach would
be to use the ValueMap in CloneFunction itself.
llvm-svn: 79562
2009-08-20 19:28:24 +00:00
c1a3a471c4
Reduce contention on the Attributes lock by using atomic operations for reference counting rather than locking.
...
llvm-svn: 79560
2009-08-20 19:03:20 +00:00
d2354b8cc5
Make the StructType->StructLayout table private to TargetData, allowing us to avoid locking on it.
...
llvm-svn: 79555
2009-08-20 18:26:03 +00:00
46bb77f2cf
Fixed PCMPESTRM128 to have opcode 0x60 instead of 0x62, as specified by the
...
Intel documentation.
llvm-svn: 79554
2009-08-20 18:24:27 +00:00
05046085b6
Fix an x86 code size regression: prefer RIP-relative addressing
...
over absolute addressing even in non-PIC mode (unless the address
has an index or something else incompatible), because it has a
smaller encoding.
llvm-svn: 79553
2009-08-20 18:23:44 +00:00
e392e6236d
Add a comment explaining why llvm_unreachable_internal doesn't call
...
the ErrorHandler callback.
llvm-svn: 79541
2009-08-20 17:15:19 +00:00
bf5998edfb
Fix two APFloat bugs in converting hexadecimal constants.
...
llvm-svn: 79540
2009-08-20 17:12:33 +00:00
16f5415f5b
Rename hasNoUnsignedOverflow and hasNoSignedOverflow to hasNoUnsignedWrap
...
and hasNoSignedWrap, for consistency with the nuw and nsw properties.
llvm-svn: 79539
2009-08-20 17:11:38 +00:00
01de985ae6
Fix an obvious copy-n-paste bug.
...
llvm-svn: 79535
2009-08-20 17:01:04 +00:00
8a8ad7d8e8
Various comment and whitespace cleanups.
...
llvm-svn: 79533
2009-08-20 16:42:55 +00:00
f91a992c30
Update and fix some comments.
...
llvm-svn: 79532
2009-08-20 16:27:10 +00:00
8b4d51af58
Check for shared landing pads when assigning call site values. Invokes which
...
share a landing pad should also use the same call site value.
llvm-svn: 79501
2009-08-20 01:03:48 +00:00
aec3830ce7
Add an extra line to conform with preferred style.
...
llvm-svn: 79495
2009-08-19 23:44:01 +00:00
5c51639e3e
Modify an assert to avoid what looks like a GCC 4.2.4 signed-ness bug.
...
llvm-svn: 79494
2009-08-19 23:39:59 +00:00
7167f42769
Fix a few places to check if TargetData is available before using it.
...
llvm-svn: 79493
2009-08-19 23:38:22 +00:00
cb13b48969
Add a fast path for setName("") on an unnamed value.
...
llvm-svn: 79492
2009-08-19 23:37:23 +00:00
d6f7af652e
AttrListPtr operations need to be atomic.
...
llvm-svn: 79486
2009-08-19 22:58:34 +00:00
c59ba42b0f
Fix a bug in the over-index constant folding. When over-indexing an
...
array member of a struct, it's possible to land in an arbitrary position
inside that struct, such that attempting to find further getelementptr
indices will fail. In such cases, folding cannot be done.
llvm-svn: 79485
2009-08-19 22:46:59 +00:00
1d764f61ef
Handle 'a' modifier in X86 asms. PR 4742.
...
llvm-svn: 79484
2009-08-19 22:44:41 +00:00
d72b091840
Fixed error in CPPBackend from a contextification API change.
...
llvm-svn: 79483
2009-08-19 22:38:37 +00:00
15aafbbfa5
Add missing forward declaration.
...
llvm-svn: 79481
2009-08-19 22:19:44 +00:00
8230a93ae2
Add missing includes.
...
llvm-svn: 79480
2009-08-19 22:16:11 +00:00
2806a1e528
Add missing includes.
...
llvm-svn: 79479
2009-08-19 22:08:26 +00:00
ba540eb492
Add missing includes.
...
llvm-svn: 79478
2009-08-19 22:05:21 +00:00
40a0fbc1d0
Add missing includes.
...
llvm-svn: 79476
2009-08-19 22:02:07 +00:00
8704780390
Add missing includes.
...
llvm-svn: 79475
2009-08-19 21:59:18 +00:00
a5be1b174b
Add missing includes.
...
llvm-svn: 79474
2009-08-19 21:55:33 +00:00
829b3e8b14
Add missing includes.
...
llvm-svn: 79473
2009-08-19 21:52:55 +00:00
a83e868078
Get rid of a helgrind warning. If this is _actually_ a performance problem,
...
we can find a way to cache the answer that isn't racy.
llvm-svn: 79472
2009-08-19 21:48:34 +00:00
28c7bbf515
Add triple parsing support for TCE.
...
llvm-svn: 79461
2009-08-19 20:46:03 +00:00
8b0b11582d
Switch to SmallString::str from SmallString::c_str, and remove
...
SmallString::c_str.
llvm-svn: 79456
2009-08-19 20:07:03 +00:00
948f82b8e4
Change ValueSymbolTable to use raw_svector_ostream for string concatenation.
...
llvm-svn: 79449
2009-08-19 19:22:52 +00:00
e813cbaa21
Change raw_svector_ostream to reserve the input buffer if necessary, Ted was
...
right.
- This class turns out to be much more convenient to use if we do this; clients
can make sure the buffer is always big enough if they care (since our current
idiom tends to be to use a SmallString<256> for the input to this we should
generally be avoiding an unnecessary malloc).
Also, add a convenience raw_svector_ostream::str method which flushes the buffer
and returns a StringRef for the vector contents.
llvm-svn: 79446
2009-08-19 18:40:58 +00:00
de255fc8f6
Remove temporary testing code.
...
llvm-svn: 79443
2009-08-19 18:27:08 +00:00
915302c605
Make SROA and PredicateSimplifier cope if TargetData is not
...
available. This is very conservative for now.
llvm-svn: 79442
2009-08-19 18:22:18 +00:00
5d5bc6d000
Use hasDefinitiveInitializer() instead of testing the same thing
...
by hand, and fix a few places that were using hasInitializer() that
appear to depend on the initializer value.
llvm-svn: 79441
2009-08-19 18:20:44 +00:00
81ce842ecf
Canonicalize indices in a constantexpr GEP. If Indices exceed the
...
static extents of the static array type, it causes GlobalOpt and
other passes to be more conservative. This canonicalization also
allows the constant folder to add "inbounds" to GEPs.
llvm-svn: 79440
2009-08-19 18:18:36 +00:00
ac33a9061d
Add an x86 peep that narrows TEST instructions to forms that use
...
a smaller encoding. These kinds of patterns are very frequent in
sqlite3, for example.
llvm-svn: 79439
2009-08-19 18:16:17 +00:00
ea579e3d29
Switch Twine::str() to use toVector(), which is now efficient.
...
llvm-svn: 79437
2009-08-19 18:09:47 +00:00
a7c2dfbca1
Update Cortex-A8 instruction itineraries for integer instructions.
...
llvm-svn: 79436
2009-08-19 18:00:44 +00:00
b090bf4626
Switch raw_svector_ostream to use the vector as the ostream buffer.
...
- This avoids unnecessary malloc/free overhead in the common case, and
unnecessary copying from the ostream buffer into the output vector.
llvm-svn: 79434
2009-08-19 17:54:29 +00:00
6d549d6450
Reapply my less-lock-contention-in-leak-detector patch, now with new files
...
actually added.
llvm-svn: 79429
2009-08-19 17:07:46 +00:00
32cd8550ce
Add support for Neon VEXT (vector extract) shuffles.
...
This is derived from a patch by Anton Korzh. I modified it to recognize
the VEXT shuffles during legalization and lower them to a target-specific
DAG node.
llvm-svn: 79428
2009-08-19 17:03:43 +00:00
465de3e2be
Speculatively revert r79375, which may be breaking bootstrap, although in a
...
rather obscure way (the other candidate is r79377).
llvm-svn: 79426
2009-08-19 16:25:25 +00:00
9b48cd4899
Use the schedule itinerary operand use/def cycle information to adjust dependence edge latency for post-RA scheduling.
...
llvm-svn: 79425
2009-08-19 16:08:58 +00:00
1e008c173a
PR4737: Fix a nasty bug in load narrowing with non-power-of-two types.
...
llvm-svn: 79415
2009-08-19 08:46:10 +00:00
cbfe9b195c
Fix up PHI nodes correctly in the presence of unreachable BBs, part two. Also
...
delete a newed pointer, and improve readability a little bit.
llvm-svn: 79411
2009-08-19 07:16:57 +00:00
78b23b0e3f
fix asmstreaming of 2/4 byte elements with pow-2 alignments.
...
llvm-svn: 79408
2009-08-19 06:35:36 +00:00
9ddc52d5b8
Be more careful when modifying PHI nodes. Patch by Andre Tavares.
...
llvm-svn: 79407
2009-08-19 06:24:33 +00:00
37b7234a75
switch asmprinter to emit alignments through OutStreamer.
...
llvm-svn: 79406
2009-08-19 06:12:02 +00:00
4b7dadb76e
eliminate AsmPrinter::SwitchToSection and just have clients
...
talk to the MCStreamer directly instead.
llvm-svn: 79405
2009-08-19 05:49:37 +00:00
21e66abd40
Broke line before break.
...
llvm-svn: 79404
2009-08-19 05:44:39 +00:00
acf0b258de
Make a SmallVector size more reasonable.
...
llvm-svn: 79403
2009-08-19 05:08:06 +00:00
11732add1e
Added an option to have the PBQP allocator attempt coalescing during allocation.
...
llvm-svn: 79397
2009-08-19 01:36:14 +00:00
077df2c4ce
Revert my last patch temporarily.
...
llvm-svn: 79388
2009-08-19 00:52:13 +00:00
3042539089
Privatize part of the leak detector mechanism, which turned out to be heavily contended
...
when trying to run opt in parallel. This lets parallel opt crunch 403.gcc in about a third
of the time.
llvm-svn: 79387
2009-08-19 00:37:02 +00:00
0cf168635e
raw_ostream: Simplify write(unsigned char) to match write(const char*, unsigned).
...
llvm-svn: 79386
2009-08-19 00:23:39 +00:00
835da36ee9
raw_ostream: Remove pointless redefinitions of tell().
...
- The base class implementation is correct.
llvm-svn: 79385
2009-08-19 00:14:25 +00:00
dea2358c68
Fix SimplifyLibcalls and ValueTracking to check mayBeOverridden
...
before performing optimizations based on constant string values.
llvm-svn: 79384
2009-08-19 00:11:12 +00:00
2fa67c9f70
Be tidy and use a break to exit from a switch block rather than
...
just falling through the end.
llvm-svn: 79383
2009-08-18 23:52:48 +00:00
317a6cd010
raw_ostream: Add the capability for subclasses to manually install an external
...
buffer.
llvm-svn: 79382
2009-08-18 23:42:36 +00:00
4906f73a9f
Legalize the shift amount operand of SRL_PARTS, SHL_PARTS, and
...
SRA_PARTS, as is done for SRL, SHL, and SRA.
llvm-svn: 79380
2009-08-18 23:36:17 +00:00
17a6fd2299
raw_ostream: Reduce FormattedStream's reliance on raw_ostream's implementation.
...
- Kill off begin(), end(), and iterator. It isn't clear what these
mean. Instead provide getBufferStart(), which can be used with
GetNumBytesInBuffer to the same effect.
- Update ComputeColumn to take arguments for the buffer to scan, this
simplifies the implementation of write_impl substantially.
- This should also fix possible problems with the scanning pointer pointing
outside of the current raw_ostream buffer.
llvm-svn: 79379
2009-08-18 23:36:04 +00:00
305be0e603
Changes DOUT references in the PBQP allocator to use DEBUG(errs() ...)
...
llvm-svn: 79378
2009-08-18 23:34:50 +00:00
9fe912de5f
Implement sse4.2 string/text processing instructions:
...
Add patterns and instruction encoding information.
Add custom lowering to deal with hardwired return register of
uncertain type (xmm0).
llvm-svn: 79377
2009-08-18 22:50:32 +00:00
bc96fe7246
Speed up raw_ostream::<<(unsigned long long) for 32-bit systems by doing most
...
div/mods in 32-bits.
llvm-svn: 79375
2009-08-18 22:24:00 +00:00
36d747745e
Simplify RegScavenger::FindUnusedReg.
...
- Drop the Candidates argument and fix all callers. Now that RegScavenger
tracks available registers accurately, there is no need to restict the
search.
- Make sure that no aliases of the found register are in use. This was a potential bug.
llvm-svn: 79369
2009-08-18 21:14:54 +00:00
2349fb4d45
Add support for mergeable sections back into the XCore backend.
...
llvm-svn: 79368
2009-08-18 21:14:31 +00:00
23f90c1d3a
Fix a bug in raw_ostream::write(char) introduced by the change to
...
allow underlying stream classes to decline buffering. After
calling SetBuffered(), re-check whether the stream is Unbuffered
in order to handle the case where the underlying stream has
declined buffering.
llvm-svn: 79362
2009-08-18 20:09:59 +00:00
d882f4b6ff
Revert r78924, disabling buffering defeats all the fast paths in raw_ostream.
...
llvm-svn: 79361
2009-08-18 20:07:36 +00:00
959323a260
Improve Triple to recognize the OS in i386-mingw32.
...
llvm-svn: 79359
2009-08-18 19:26:55 +00:00
60eb473ac9
Make various changes suggested by Chris.
...
llvm-svn: 79358
2009-08-18 19:22:55 +00:00
e8f2185a2f
Privatize the ValueHandle global map. Because this is used so heavily throughout the code base, locking all accesses to
...
it is not practical performance-wise.
llvm-svn: 79355
2009-08-18 18:28:58 +00:00
a9680df669
Fix an uninitialized value warning in APFloat.
...
llvm-svn: 79353
2009-08-18 18:20:37 +00:00
934d61648b
Put data with relocations in the same sections as data without relocations.
...
llvm-svn: 79351
2009-08-18 17:58:17 +00:00
03ded465d2
fix COFF targets (mingw/cygwin) to provide ehframe and LSDA sections
...
llvm-svn: 79346
2009-08-18 16:56:17 +00:00
bf2a9aedca
Generalize ScalarEvolution to be able to analyze GEPs when
...
TargetData is not present. It still uses TargetData when available.
This generalization also fixed some limitations in the TargetData
case; the attached testcase covers this.
llvm-svn: 79344
2009-08-18 16:46:41 +00:00
f446070b62
null streamer needs to maintain the current section as well.
...
llvm-svn: 79343
2009-08-18 16:46:29 +00:00
a41fa35992
Make tail merging handle blocks with repeated predecessors correctly, and
...
remove RemoveDuplicateSuccessor, as it is no longer necessary, and because
it breaks assumptions made in
MachineBasicBlock::isOnlyReachableByFallthrough.
Convert test/CodeGen/X86/omit-label.ll to FileCheck and add a testcase
for PR4732.
test/CodeGen/Thumb2/thumb2-ifcvt2.ll sees a diff with this commit due to
it being bugpoint-reduced to the point where it doesn't matter what the
condition for the branch is.
Add some more interesting code to
test/CodeGen/X86/2009-08-06-branchfolder-crash.ll, which is the testcase
that originally motivated the RemoveDuplicateSuccessor code, to help
verify that the original problem isn't being re-broken.
llvm-svn: 79338
2009-08-18 15:18:18 +00:00
82ac81b1cc
Fix a bug that caused globalopt to miscompile tramp3d: don't miss
...
unruly indices for arrays that are members of structs.
llvm-svn: 79337
2009-08-18 14:58:19 +00:00
e4d2e8465a
Text sections should have 'exec' flag set. This seems to unbreak libstdc++ on linux.
...
Patch by Dmitry Gorbachev!
llvm-svn: 79334
2009-08-18 14:06:12 +00:00
33d86f22f2
Fix Triple to recognize the 'bfin' arch.
...
llvm-svn: 79325
2009-08-18 07:06:26 +00:00
0adae25ec1
Make AsmStreamer maintain a notion of the current section, pushing it up from the
...
MCAsmStreamer. Based on this, eliminate the current section from AsmPrinter.
While I'm at it, clean up the last of the horrible "switch to null section" stuff
and add an assert. This change is in preparation for completely eliminating
asmprinter::switchtosection.
llvm-svn: 79324
2009-08-18 06:15:16 +00:00
5de2e2de93
add a horrible hack to the dwarf printer. It looks like mingw is not specifying
...
an EHFrame section, so we just emit ehframe data into a random section.
This is clearly bad.
llvm-svn: 79323
2009-08-18 06:13:03 +00:00
966d32dffb
remove some pointless null switchtosections. The IntelAsmPrinter doesn't really work anyway.
...
llvm-svn: 79321
2009-08-18 06:03:07 +00:00
dd406177de
Fix revsh pattern.
...
llvm-svn: 79318
2009-08-18 05:43:23 +00:00
40bbb85949
add support for some targetflags on GV operands. This allows us to
...
send instructions like:
NEW: movl "L___stack_chk_guard$non_lazy_ptr" - "L1$pb"(%esi), %eax
OLD: movl L___stack_chk_guard$non_lazy_ptr-"L1$pb"(%esi), %eax
through the streamer. Several fixmes.
llvm-svn: 79317
2009-08-18 05:33:27 +00:00
320d331311
Recognize xscale as an ARM arch.
...
- Patch by Yonggang Luo.
llvm-svn: 79315
2009-08-18 04:51:26 +00:00
781f94d7fe
Add Triple matching for pic16 arch and solaris OS.
...
- Patch by Yonggang Luo.
llvm-svn: 79314
2009-08-18 04:43:27 +00:00
63d3fa5f3b
fix another bozo bug
...
llvm-svn: 79313
2009-08-18 04:34:36 +00:00
249e1acf1b
fix accidentally inverted conditional and add comment.
...
llvm-svn: 79312
2009-08-18 04:33:15 +00:00
523d2f6e2c
turn this conditional into something humans might actually
...
be able to understand ;-)
llvm-svn: 79311
2009-08-18 04:30:35 +00:00
10f1471e2f
Make TargetData optional in MemCpyOptimizer.
...
llvm-svn: 79306
2009-08-18 01:17:52 +00:00
9f2b3db428
Make TargetData optional in SimplifyLibCalls.
...
llvm-svn: 79298
2009-08-18 00:48:13 +00:00
90e17e787f
The attached patches attempt to fix cross builds. For example, if you
...
try to use i686-darwin to build for arm-eabi, you'll quickly run into
several false assumptions that the target OS must be the same as the
host OS. These patches split $(OS) into $(HOST_OS) and $(TARGET_OS) to
help builds like "make check" and the test-suite able to cross
compile. Along the way a target of *-unknown-eabi is defined as
"Freestanding" so that TARGET_OS checks have something to work with.
Patch by Sandeep Patel!
llvm-svn: 79296
2009-08-18 00:40:33 +00:00
47a31a29c5
Fix function alignment at -Os on x86 to be 1, not 2. getFunctionAlignment
...
returns a log2 value.
llvm-svn: 79293
2009-08-18 00:20:06 +00:00
4a50e68b65
PowerPC inline asm was emitting two output operands
...
for a single "m" constraint; this is wrong because the
opcode of a load or store would have to change in parallel.
This patch makes it always compute addresses into a register,
which is correct but not as efficient as possible. 7144566.
llvm-svn: 79292
2009-08-18 00:18:39 +00:00
c6faffde6c
Add prefix only if it is needed.
...
llvm-svn: 79289
2009-08-17 23:17:17 +00:00
f933c409e2
cleanups per review. Mostly cosmetic, plus use SmallVector in place of std::vector.
...
llvm-svn: 79287
2009-08-17 21:40:03 +00:00
993e527b30
Oops. find all llvm.dbg.global_variables.
...
llvm-svn: 79274
2009-08-17 20:36:20 +00:00
43bbb9de66
Remove a bit more cruft from the sjlj moving to a backend pass.
...
llvm-svn: 79272
2009-08-17 20:25:04 +00:00
436bd9f770
Update CMakeLists.
...
llvm-svn: 79264
2009-08-17 18:47:11 +00:00
7cb92a1e3d
Update comments to new-style syntax.
...
llvm-svn: 79263
2009-08-17 18:45:31 +00:00
f9bdf988a2
Fix build warning.
...
llvm-svn: 79262
2009-08-17 18:41:42 +00:00
9801560ace
Privatize the last bits of static type state.
...
llvm-svn: 79258
2009-08-17 17:59:35 +00:00
286b3af6ae
Move the TypeMap lock to a member on LLVMContextImpl.
...
llvm-svn: 79256
2009-08-17 17:34:27 +00:00
d91e6b097d
Add locking around the attributes list.
...
llvm-svn: 79255
2009-08-17 17:10:58 +00:00
820739dccc
Add locking around signal handler registration.
...
llvm-svn: 79254
2009-08-17 17:07:22 +00:00
bd3c033713
Clear the uniquing table when initializing TLOF to avoid a crash when the TLOF is reinitialized with a different MCContext.
...
llvm-svn: 79253
2009-08-17 17:05:44 +00:00
4c737147e1
Don't crash on critical edge. Patch by Andre Tavares.
...
llvm-svn: 79252
2009-08-17 17:00:57 +00:00
486be66dbd
Move the sjlj exception handling conversions to a back-end pass where they
...
more properly belong. This allows removing the front-end conditionalized
SJLJ code, and cleans up the generated IR considerably. All of the
infrastructure code (calling _Unwind_SjLj_Register/Unregister, etc) is
added by the SjLjEHPrepare pass.
llvm-svn: 79250
2009-08-17 16:41:22 +00:00
94a2c1acae
Update getSectionForConstant() to to allow mergable sections to be nulled out
...
if not supported by the ELF subtarget.
llvm-svn: 79249
2009-08-17 16:37:11 +00:00
d813cbfe72
Extend the instruction itinerary model to include the ability to indicate the def and use cycle for each operand. This additional information is optional, so existing itineraries do not need to be changed.
...
llvm-svn: 79247
2009-08-17 16:02:57 +00:00
ee97b8b11c
the MinPad argument to PadToColumn only really makes sense to be 1,
...
just remove the argument and replace it with 1.
llvm-svn: 79246
2009-08-17 15:48:08 +00:00
74cb168a9e
Fix more missing newlines.
...
llvm-svn: 79244
2009-08-17 15:29:31 +00:00
d04f9bf01a
Fix printing of instructions with null operands.
...
llvm-svn: 79243
2009-08-17 15:28:08 +00:00
d0cc3f31c0
Always print at least one space before adding a comment.
...
llvm-svn: 79242
2009-08-17 15:27:30 +00:00
8dd69f88ea
Fix debug output to include a newline after printing a Value, now
...
that Value's operator<< doesn't include one.
llvm-svn: 79240
2009-08-17 15:25:05 +00:00
c4ce58d8fe
Don't access the first element of a potentially empty
...
vector (&Formals[0]). With this change llvm-gcc builds
with expensive checking enabled for C, C++ and Fortran.
While there, change a std::vector into a SmallVector.
This is partly gratuitous, but mostly because not all
STL vector implementations define the data method (and
it should be faster).
llvm-svn: 79237
2009-08-17 14:33:27 +00:00
aa1526419c
change AsmPrinter to switch sections using AsmStreamer instead of
...
doing it directly. This requires const'izing a bunch of stuff that
took sections, but this seems like the right semantic thing to do:
emitting a label to a section shouldn't mutate the MCSection object
itself, for example.
llvm-svn: 79227
2009-08-17 05:49:08 +00:00
aa464002f0
Don't crash trying to promote VLAs.
...
llvm-svn: 79226
2009-08-17 05:37:31 +00:00
a61e93d4b5
give MCAsmStreamer a TargetAsmInfo.
...
llvm-svn: 79222
2009-08-17 04:23:44 +00:00
962c5bd803
formatting cleanups, no functionality change.
...
llvm-svn: 79221
2009-08-17 04:17:34 +00:00
dd99135721
Expose creating constant ints and floats from strings in llvm-c.
...
llvm-svn: 79213
2009-08-16 23:36:46 +00:00
fc2280d874
Add helper functions to ConstantInt and ConstantFP to accept strings.
...
llvm-svn: 79212
2009-08-16 23:36:33 +00:00
19f63b2e4d
Modify APFloat to take a StringRef instead of a c string.
...
This also adds unit tests to APFloat that mainly tests the
string handling of APFloat, but not much else of it's api.
llvm-svn: 79210
2009-08-16 23:36:19 +00:00
ff3af725ea
Add a getOffsetOf, for building a target-independent expression for
...
offsetof, similar to getSizeOf for sizeof.
llvm-svn: 79208
2009-08-16 21:26:11 +00:00
f0b984021c
Avoid emitting XMM save code in soft-float or no-implicit-float mode
...
or some other situation where no xmm registers need to be saved.
llvm-svn: 79207
2009-08-16 21:24:25 +00:00
4bb9357921
Delete an unused field.
...
llvm-svn: 79206
2009-08-16 21:19:53 +00:00
84b4d2241f
Replace RegScavenger::DistanceMap with a simpler local algorithm.
...
llvm-svn: 79195
2009-08-16 17:41:39 +00:00
ac09a9a1f5
Clean up the public interface of RegScavenger.
...
Remove unused methods and make others private.
llvm-svn: 79194
2009-08-16 17:41:20 +00:00
7f91fee62b
Be more clever about regclasses in ScheduleDAGSDNodes::EmitCopyFromReg.
...
If two uses of a CopyFromReg want different regclasses, first try a common
sub-class, then fall back on the copy emitted in AddRegisterOperand. There is
no need for an assert here. The cross-class joiner usually cleans up nicely.
llvm-svn: 79193
2009-08-16 17:40:59 +00:00
2c64130c43
Fix use after free in Thumb2SizeReduction (PR4707). A MachineInstr was used after erasure.
...
llvm-svn: 79189
2009-08-16 11:56:42 +00:00
195fd9d970
Styalistic and format changes. No functionality change.
...
llvm-svn: 79187
2009-08-16 11:00:26 +00:00
8a4c35dd18
add support for external symbols + X86::MOVPC32r.
...
llvm-svn: 79175
2009-08-16 04:28:14 +00:00
d56fca4708
Fix for PR3016: detect the tricky case, where there are
...
unfoldable references to a PHI node in the block being folded, and disable
the transformation in that case. The correct transformation of such PHI
nodes depends on whether BB dominates Succ, and dominance is expensive
to compute here. (Alternatively, it's possible to check whether any
uses are live, but that's also essentially a dominance calculation.
Another alternative is to use reg2mem, but it probably isn't a good idea to
use that in simplifycfg.)
Also, remove some incorrect code from CanPropagatePredecessorsForPHIs
which is made unnecessary with this patch: it didn't consider the case
where a PHI node in BB has multiple uses.
llvm-svn: 79174
2009-08-16 04:23:49 +00:00
c659caa848
implement support for lowering references to global addresses. For example, we now
...
can asmprint:
NEW: movl "L___stack_chk_guard$non_lazy_ptr", %eax
OLD: movl L___stack_chk_guard$non_lazy_ptr, %eax
where 'new' is coming out of the MCInst version of the printer.
llvm-svn: 79170
2009-08-16 03:12:25 +00:00
10f605ce2b
more formatting improvements, no functionality change.
...
llvm-svn: 79167
2009-08-16 02:45:18 +00:00
043bb021da
code formatting improvements, no functionality change.
...
llvm-svn: 79165
2009-08-16 02:36:40 +00:00
9813beadcd
Add an llvm-c function that lets you insert an instruction with a name.
...
llvm-svn: 79163
2009-08-16 02:20:57 +00:00
Erick Tryzelaar