6860ac7375
llvm-mc: Clean up some handling of symbol/section association to be more correct
...
(external was really undefined and there wasn't an explicit representation for
absolute symbols).
- This still needs some cleanup to how the absolute "pseudo" section is dealt
with, but I haven't figured out the nicest approach yet.
llvm-svn: 79733
2009-08-22 07:22:36 +00:00
63010bb65a
Reapply r79708 with the appropriate fix for the case that still requires locking.
...
llvm-svn: 79731
2009-08-22 06:32:36 +00:00
56d60eaa61
revert r79708 + r79711
...
llvm-svn: 79720
2009-08-22 04:07:34 +00:00
682d8c1881
Make x86 test actually test x86 code generation. Fix the
...
construct on ARM, which was breaking by coincidence, and add a similar
testcase for ARM.
llvm-svn: 79719
2009-08-22 03:13:10 +00:00
677c2287da
Actually remove unused static. Previous commit removed trailing
...
whitespace.
llvm-svn: 79711
2009-08-22 00:41:47 +00:00
dfda92b76e
Remove unused static.
...
llvm-svn: 79710
2009-08-22 00:40:45 +00:00
8e2456c254
Ease contention on this lock by noticing that all writes to the VTs array will
...
be of (dynamically) constant values, so races on it are immaterial. We just need
to ensure that at least one write has completed before return the pointer into it.
With this change, parllc exhibits essentially no overhead on 403.gcc.
llvm-svn: 79708
2009-08-22 00:29:12 +00:00
a7668ff636
Revert r79563
...
llvm-svn: 79691
2009-08-21 23:12:47 +00:00
773416e062
revert r79562 + r79563
...
llvm-svn: 79690
2009-08-21 23:12:15 +00:00
876955cbb0
Fix typo.
...
llvm-svn: 79688
2009-08-21 23:09:47 +00:00
f407dbbb1a
revert r79631
...
llvm-svn: 79686
2009-08-21 23:08:45 +00:00
6af22f02ef
revert 79631
...
llvm-svn: 79685
2009-08-21 23:08:09 +00:00
b1068e4fe3
llvm-mc/Mach-O: Support .o emission for .org and .align.
...
llvm-svn: 79684
2009-08-21 23:07:38 +00:00
13e8c7ef20
Implement APInt <-> APFloat conversion for IEEE 128-bit floats.
...
This fixes PR2555
llvm-svn: 79677
2009-08-21 22:10:30 +00:00
ceffeb6abd
Rename ARM "lane_cst" operands to "nohash_imm" since they are used for
...
several things other than Neon vector lane numbers. For inline assembly
operands with a "c" print code, check that they really are immediates.
llvm-svn: 79676
2009-08-21 21:58:55 +00:00
48ca915be9
Fix a bug where the DWARF emitter in the JIT was not initializing alignment
...
bytes. libgcc doesn't seem to mind, but if you pass this DWARF to GDB, it
doesn't like it.
Also make the JIT memory manager to initialize it's memory to garbage in debug
mode, so that it's easier to find bugs like these in the future.
llvm-svn: 79674
2009-08-21 21:03:57 +00:00
a70623102e
Match VTRN, VZIP, and VUZP shuffles. Restore the tests for these operations,
...
now using shuffles instead of intrinsics.
llvm-svn: 79673
2009-08-21 20:54:19 +00:00
f31a44ec01
Add fcopysign instructions
...
llvm-svn: 79664
2009-08-21 20:02:37 +00:00
2dd877d12a
Try again at privatizing the layout info map, with a rewritten patch.
...
This preserves the existing behavior much more closely than my previous attempt.
llvm-svn: 79663
2009-08-21 19:59:12 +00:00
47398930e4
Expand few nodes until someone will be crazy enough to implement them natively :)
...
llvm-svn: 79659
2009-08-21 18:52:42 +00:00
cb7444342b
Typo :(
...
llvm-svn: 79657
2009-08-21 18:41:02 +00:00
81be4b345a
Correct instruction names for subtract-with-borrow
...
llvm-svn: 79656
2009-08-21 18:37:28 +00:00
4ef7fb9530
llvm-mc/Mach-O: Support byte and fill value emission.
...
llvm-svn: 79652
2009-08-21 18:29:01 +00:00
c0b967dfb4
Don't assume that the operand of an inttoptr is an pointer-sized integer.
...
llvm-svn: 79651
2009-08-21 18:27:26 +00:00
dff54eff8e
Fix typo. Should check both values of RangeUse for 0. Patch by Marius Wachtler.
...
llvm-svn: 79649
2009-08-21 18:16:06 +00:00
a39f96c6ed
Handle 'r' inline asm constraint
...
llvm-svn: 79648
2009-08-21 18:15:41 +00:00
08b437d600
Fix a problem noticed by gcc-4.4:
...
warning: comparison is always true due to limited range of data type.
llvm-svn: 79642
2009-08-21 17:16:10 +00:00
23e62c532c
Change getelementptr folding to use APInt instead of uint64_t for
...
offset computations. This fixes a truncation bug on targets that
don't have 64-bit pointers.
llvm-svn: 79639
2009-08-21 16:52:54 +00:00
2f0e8f55f5
Fix -Asserts warning.
...
llvm-svn: 79636
2009-08-21 16:17:36 +00:00
7950510b29
Fix a typo
...
llvm-svn: 79634
2009-08-21 15:41:56 +00:00
9ae3bcb79e
Add a pass to do call graph analyis to overlay the autos and frame sections of
...
leaf functions. This pass will be extended to color other nodes of the call tree
as well in future.
llvm-svn: 79631
2009-08-21 15:22:33 +00:00
eaaca496c5
More cpp backend fixes. Now for FP stuff.
...
llvm-svn: 79626
2009-08-21 12:50:54 +00:00
232b19c3d5
Fix some typos and use type-based isel for VZIP/VUZP/VTRN
...
llvm-svn: 79625
2009-08-21 12:41:42 +00:00
9a232f46a8
Add lowering of ARM 4-element shuffles to multiple instructios via perfectshuffle-generated table.
...
llvm-svn: 79624
2009-08-21 12:41:24 +00:00
ce3ff1be8a
Add nodes & dummy matchers for some v{zip,uzp,trn} instructions
...
llvm-svn: 79622
2009-08-21 12:40:50 +00:00
e3046618de
Expand EXTRACT_SUBVECTOR
...
llvm-svn: 79621
2009-08-21 12:40:35 +00:00
38f284f2ae
Provide vext.{16,32}
...
llvm-svn: 79620
2009-08-21 12:40:21 +00:00
c32e99e3ed
Use masks not nodes for vector shuffle predicates. Provide set of 'legal' masks, so legalizer won't infinite cycle
...
llvm-svn: 79619
2009-08-21 12:40:07 +00:00
c20529945a
Update CMakeLists.
...
llvm-svn: 79617
2009-08-21 11:28:56 +00:00
9e0c4f50ec
bug 4530: Make debug information static to it is preservered during bitcode linking.
...
llvm-svn: 79616
2009-08-21 11:10:31 +00:00
3016db39dd
llvm-mc: Start MCAssembler and MCMachOStreamer.
...
- Together these form the (Mach-O) back end of the assembler.
- MCAssembler is the actual assembler backend, which is designed to have a
reasonable API. This will eventually grow to support multiple object file
implementations, but for now its Mach-O/i386 only.
- MCMachOStreamer adapts the MCStreamer "actions" API to the MCAssembler API,
e.g. converting the various directives into fragments, managing state like
the current section, and so on.
- llvm-mc will use the new backend via '-filetype=obj', which may eventually
be, but is not yet, since I hear that people like assemblers which actually
assemble.
- The only thing that works at the moment is changing sections. For the time
being I have a Python Mach-O dumping tool in test/scripts so this stuff can
be easily tested, eventually I expect to replace this with a real LLVM tool.
- More doxyments to come.
I assume that since this stuff doesn't touch any of the things which are part of
2.6 that it is ok to put this in not so long before the freeze, but if someone
objects let me know, I can pull it.
llvm-svn: 79612
2009-08-21 09:11:24 +00:00
64965aafe7
Remove #include <iostream>.
...
llvm-svn: 79603
2009-08-21 06:52:44 +00:00
6096409238
Clean up the APInt function getDigit.
...
llvm-svn: 79602
2009-08-21 06:48:37 +00:00
43a1decc8c
Update error messages for '+'. Fix grammar and make the two
...
negative checks resemble each other.
llvm-svn: 79595
2009-08-21 04:10:31 +00:00
820256bc99
Fix trailing whitespace and 80-col violation.
...
llvm-svn: 79594
2009-08-21 04:06:45 +00:00
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
Daniel Dunbar