RPCS3/llvm-mirror
1
0
Fork 0
mirror of https://github.com/RPCS3/llvm-mirror.git synced 2024-10-21 03:53:04 +02:00
Code Issues Projects Releases Wiki Activity
31,467 Commits 4 Branches 5 Tags 704 MiB
5bb6590218
Commit Graph

18270 Commits

Author SHA1 Message Date
Chris Lattner
0160bf7114 rearchitect the registration mechanism used by the command line option stuff. 
This dramatically reduce the amount of memory allocated by the commandline stuff
at static init time, changing it to build local data structures when ParseCommandLineOptions
is called.  In a dummy empty program that links some llvm libraries, this reduces
the number of malloc'd bytes from 4864 to 3360 on entry to main.  Most of that
memory is now allocated by non-commandline related stuff.

llvm-svn: 35701
 2007-04-06 21:06:55 +00:00
Chris Lattner
03c84be56b implement Transforms/InstCombine/malloc2.ll and PR1313 
llvm-svn: 35700
 2007-04-06 18:57:34 +00:00
Chris Lattner
b39a2df066 minor comment fix 
llvm-svn: 35696
 2007-04-06 17:47:14 +00:00
Reid Spencer
5d36a01dba For PR1209: 
Implement Type class's ContainedTys without using a std::vector.

llvm-svn: 35693
 2007-04-06 02:02:20 +00:00
Chris Lattner
d33f94f2d2 remove the dead removeArgument method, rename Options to OptionsMap. 
llvm-svn: 35690
 2007-04-05 21:58:17 +00:00
Chris Lattner
19b9e90f9c Fix Transforms/GlobalOpt/2007-04-05-Crash.ll 
llvm-svn: 35689
 2007-04-05 21:09:42 +00:00
Chris Lattner
997967979f Use a worklist-driven algorithm instead of a recursive one. 
llvm-svn: 35680
 2007-04-05 01:27:02 +00:00
Reid Spencer
aad0b4536b Change the bit_part_select (non)implementation from "return 0" to abort. 
llvm-svn: 35679
 2007-04-05 01:20:18 +00:00
Reid Spencer
6eb55df794 Implement the llvm.bit.part_select.iN.iN.iN overloaded intrinsic. 
llvm-svn: 35678
 2007-04-04 23:48:25 +00:00
Reid Spencer
a72beea861 Squelch a warning about mismatch between sign of constant and sign of return 
type.

llvm-svn: 35674
 2007-04-04 22:07:24 +00:00
Anton Korobeynikov
e16f421e0e Properly emit range comparisons for switch cases, where neighbour cases 
go to the same destination. Now we're producing really good code for
switch-lower-feature.ll testcase

llvm-svn: 35672
 2007-04-04 21:14:49 +00:00
Dale Johannesen
fb15913194 Prevent transformConstExprCastCall from generating conversions that assert 
elsewhere.

llvm-svn: 35668
 2007-04-04 19:16:42 +00:00
Chris Lattner
ed593d6789 stringmap memory managed with malloc now 
llvm-svn: 35666
 2007-04-04 17:24:28 +00:00
Jeff Cohen
01d4afe6da Fix 2007-04-04-BadFoldBitcastIntoMalloc.ll 
llvm-svn: 35665
 2007-04-04 16:58:57 +00:00
Evan Cheng
b7ec9433b3 Re-materialize all loads from fixed stack slots. 
llvm-svn: 35660
 2007-04-04 07:40:01 +00:00
Evan Cheng
48b94106d6 Trivially re-materializable instructions have spill weights that are half of what it would be otherwise. 
llvm-svn: 35658
 2007-04-04 07:04:55 +00:00
Reid Spencer
d8c3813f51 For PR1302: 
Use local variable names that match the function parameter name that it
is passed to so the code is more clear, to wit: is_bytecode -> is_native

llvm-svn: 35656
 2007-04-04 06:44:18 +00:00
Duncan Sands
de998e6599 Fix comment. 
llvm-svn: 35655
 2007-04-04 06:42:45 +00:00
Reid Spencer
87ede33a19 For PR1302: 
Implement file tests for both LinkInLibrary and LinkInFile to determine if
the file is native. Don't generate warnings if the file is native.

llvm-svn: 35653
 2007-04-04 06:33:17 +00:00
Reid Spencer
6df393cf38 For PR1302: 
Adjust useage of sys::Path::FileType for new enumerator names.

llvm-svn: 35651
 2007-04-04 06:31:04 +00:00
Reid Spencer
e825ec8a72 For PR1302: 
Implement recognition of COFF, ELF and Mach-O object/shared lib files.

llvm-svn: 35650
 2007-04-04 06:30:26 +00:00
Chris Lattner
8a4ca643dd use calloc instead of new/memset, it is more efficient 
llvm-svn: 35644
 2007-04-04 00:44:31 +00:00
Chris Lattner
dbf6a79a71 Extend StringMap to support being initialized as completely empty. When 
initialized this way, they do not do a malloc to allocate their buckets.

llvm-svn: 35642
 2007-04-04 00:29:37 +00:00
Evan Cheng
44488cb676 Implement inline asm modifier P. 
llvm-svn: 35640
 2007-04-04 00:13:29 +00:00
Evan Cheng
d5956e4409 Typo. 
llvm-svn: 35639
 2007-04-04 00:06:07 +00:00
Bill Wendling
a4aa65bc38 Adding more MMX instructions. 
llvm-svn: 35638
 2007-04-03 23:48:32 +00:00
Chris Lattner
78271f0596 make a new missing features section 
llvm-svn: 35637
 2007-04-03 23:41:34 +00:00
Evan Cheng
2cd9df6983 Remove unused constant pool entries. 
llvm-svn: 35635
 2007-04-03 23:39:48 +00:00
Bill Wendling
ff2c460267 Updated 
llvm-svn: 35634
 2007-04-03 23:37:20 +00:00
Chris Lattner
974e931689 Fix a bug I introduced with my patch yesterday which broke Qt (I converted 
some constant exprs to apints).

Thanks to Anton for tracking down a small testcase that triggered this!

llvm-svn: 35633
 2007-04-03 23:29:39 +00:00
Chris Lattner
59994ede6b greatly reduce hte default size of stringmap. 
llvm-svn: 35632
 2007-04-03 22:15:38 +00:00
Evan Cheng
db15aa24f9 Fixed a bug that causes codegen of noop like add r0, r0, #0. 
llvm-svn: 35627
 2007-04-03 21:31:21 +00:00
Chris Lattner
bfe18d29f9 reinstate the previous two patches, with a bugfix :) 
ldecod now passes.

llvm-svn: 35626
 2007-04-03 17:43:25 +00:00
Nicolas Geoffray
681a87d9e8 Starting implementation of the ELF32 ABI specification of varargs handling. 
LowerVASTART emits the right code if the subtarget is ELF32, the other intrinsics
(VAARG, VACOPY and VAEND) are not yet implemented.

llvm-svn: 35625
 2007-04-03 13:59:52 +00:00
Nicolas Geoffray
5897c064a6 The PPC64 ELF ABI is "intended to use the same structure layout and calling convention rules 
as the 64-bit PowerOpen ABI" (Reference http://www.linux-foundation.org/spec/ELF/ppc64/).
Change all ELF tests to ELF32.

llvm-svn: 35624
 2007-04-03 12:35:28 +00:00
Nicolas Geoffray
45c3620d12 Addition to the previous commit for getCalleeSavedRegClasses: 
"The ELF ABI specifies F1-F8 registers as argument registers for double, not
F1-F10. This affects only ELF, not MachO."

llvm-svn: 35623
 2007-04-03 10:57:49 +00:00
Nicolas Geoffray
b7c0895529 The ELF ABI specifies F1-F8 registers as argument registers for double, not 
F1-F10. This affects only ELF, not MachO.

llvm-svn: 35622
 2007-04-03 10:27:07 +00:00
Evan Cheng
e30cfe8e9e Reverting back to 1.723. The last two commits broke JM (and possibily others) on ARM. 
llvm-svn: 35620
 2007-04-03 08:11:50 +00:00
Evan Cheng
16696f0ad8 Inverted logic. 
llvm-svn: 35619
 2007-04-03 06:44:25 +00:00
Evan Cheng
bad40843ec Bad bad bug. findRegisterUseOperand() returns -1 if a use if not found. 
llvm-svn: 35618
 2007-04-03 06:43:29 +00:00
Bill Wendling
a2eabd50c0 Changed to new MMX_ recipes. 
llvm-svn: 35617
 2007-04-03 06:18:31 +00:00
Bill Wendling
ca2124e5a9 Add FEMMS and ADDQ. Renamed MMX recipes to prepend the MMX_ to them. 
llvm-svn: 35616
 2007-04-03 06:00:37 +00:00
Chris Lattner
4302c8bcdb split some code out into a helper function 
llvm-svn: 35615
 2007-04-03 05:11:24 +00:00
Chris Lattner
23d58bde29 Split a whole ton of code out of visitICmpInst into visitICmpInstWithInstAndIntCst. 
llvm-svn: 35614
 2007-04-03 04:46:52 +00:00
Chris Lattner
9b66a372c8 Fix PR1253 and xor2.ll:test[01] 
llvm-svn: 35612
 2007-04-03 01:47:41 +00:00
Chris Lattner
4d702a890c Arm supports negative strides as well, add them. This lets us compile: 
CodeGen/ARM/arm-negative-stride.ll to:

LBB1_2: @bb
        str r1, [r3, -r0, lsl #2]
        add r0, r0, #1
        cmp r0, r2
        bne LBB1_2      @bb

llvm-svn: 35609
 2007-04-03 00:13:57 +00:00
Chris Lattner
3df2382275 allow -1 strides to reuse "1" strides. 
llvm-svn: 35607
 2007-04-02 22:51:58 +00:00
Scott Michel
d6b3d3d6ab 1. Insert custom lowering hooks for ISD::ROTR and ISD::ROTL. 
2. Help DAGCombiner recognize zero/sign/any-extended versions of ROTR and ROTL
patterns. This was motivated by the X86/rotate.ll testcase, which should now
generate code for other platforms (and soon-to-come platforms.) Rewrote code
slightly to make it easier to read.

llvm-svn: 35605
 2007-04-02 21:36:32 +00:00
Dale Johannesen
e1c3dc7c78 fix off by 1 error in displacement computation 
llvm-svn: 35602
 2007-04-02 20:31:06 +00:00
Chris Lattner
726d461daf fix the CodeGen/ARM/2007-03-13-InstrSched.ll regression: allow IV's with scales 
to be folded into non-store instructions.

llvm-svn: 35601
 2007-04-02 18:51:18 +00:00
Evan Cheng
d0366b0024 Ugh. Copy coalescer does not update register numbers. 
llvm-svn: 35600
 2007-04-02 18:49:18 +00:00
Chris Lattner
ff9caac83e add support for the 'w' inline asm register class. 
llvm-svn: 35598
 2007-04-02 17:24:08 +00:00
Zhou Sheng
98c161c290 1. Make use of APInt operation instead of using ConstantExpr::getXXX. 
2. Use cheaper APInt methods.

llvm-svn: 35594
 2007-04-02 13:45:30 +00:00
Zhou Sheng
47311533a7 Use uint32_t for bitwidth instead of unsigned. 
llvm-svn: 35593
 2007-04-02 08:20:41 +00:00
Chris Lattner
d7c96e25e4 Pass the type of the store access, not the type of the store, into the 
target hook.  This allows us to codegen a loop as:

LBB1_1: @cond_next
        mov r2, #0
        str r2, [r0, +r3, lsl #2]
        add r3, r3, #1
        cmn r3, #1
        bne LBB1_1      @cond_next

instead of:

LBB1_1: @cond_next
        mov r2, #0
        str r2, [r0], #+4
        add r3, r3, #1
        cmn r3, #1
        bne LBB1_1      @cond_next

This looks the same, but has one fewer induction variable (and therefore,
one fewer register) live in the loop.

llvm-svn: 35592
 2007-04-02 06:34:44 +00:00
Chris Lattner
dbce9ef4b8 Fix a bug which caused us to never be able to use signed comparisons for 
equality comparisons of a constant.  This allows us to codegen the 'sintzero'
loop in PR1288 as:

LBB1_1: ;cond_next
        li r4, 0
        addi r2, r2, 1
        stw r4, 0(r3)
        addi r3, r3, 4
        cmpwi cr0, r2, -1
        bne cr0, LBB1_1 ;cond_next

instead of:

LBB1_1: ;cond_next
        addi r2, r2, 1
        li r4, 0
        xoris r5, r2, 65535
        stw r4, 0(r3)
        addi r3, r3, 4
        cmplwi cr0, r5, 65535
        bne cr0, LBB1_1 ;cond_next

This implements CodeGen/PowerPC/compare-simm.ll, and also cuts 74
instructions out of kc++.

llvm-svn: 35590
 2007-04-02 05:59:42 +00:00
Chris Lattner
aaf29d3b5e Wrap long line 
llvm-svn: 35588
 2007-04-02 05:48:58 +00:00
Chris Lattner
7208dd1001 use more obvious function name. 
llvm-svn: 35587
 2007-04-02 05:42:22 +00:00
Chris Lattner
6668db1331 Treat xor of signbit like an add. 
llvm-svn: 35586
 2007-04-02 05:41:38 +00:00
Chris Lattner
35ece02bea simplify (x+c)^signbit as (x+c+signbit), pointed out by PR1288. This implements 
test/Transforms/InstCombine/xor.ll:test28

llvm-svn: 35584
 2007-04-02 05:36:22 +00:00
Chris Lattner
d59b6b3d90 reduce use of std::set 
llvm-svn: 35576
 2007-04-02 01:44:59 +00:00
Chris Lattner
2f94a80322 Various passes before isel split edges and do other CFG-restructuring changes. 
isel has its own particular features that it wants in the CFG, in order to
reduce the number of times a constant is computed, etc.  Make sure that we
clean up the CFG before doing any other things for isel.  Doing so can
dramatically reduce the number of split edges and reduce the number of
places that constants get computed.  For example, this shrinks
CodeGen/Generic/phi-immediate-factoring.ll from 44 to 37 instructions on X86,
and from 21 to 17 MBB's in the output.  This is primarily a code size win,
not a performance win.

This implements CodeGen/Generic/phi-immediate-factoring.ll and PR1296.

llvm-svn: 35575
 2007-04-02 01:35:34 +00:00
Lauro Ramos Venancio
7e21418680 - Divides the comparisons in two types: comparisons that only use N and Z 
flags (ARMISD::CMPNZ) and comparisons that use all flags (ARMISD::CMP).
- Defines the instructions: TST, TEQ (ARM) and TST (Thumb).

llvm-svn: 35573
 2007-04-02 01:30:03 +00:00
Reid Spencer
419407379c For PR1297: 
Make sure that the CTPOP result is casted to i32 as the bit counting
intrinsics all return i32 now (this affects CTLZ and CTTZ as well).

llvm-svn: 35567
 2007-04-02 01:01:49 +00:00
Chris Lattner
ac77daac29 print the type of an inserted IV in -debug mode. 
llvm-svn: 35563
 2007-04-01 22:21:39 +00:00
Chris Lattner
a66a97908c simplify this code, make it work for ap ints 
llvm-svn: 35561
 2007-04-01 20:57:36 +00:00
Chris Lattner
4fb170e47b fix breakage from last night, simplify code. 
llvm-svn: 35560
 2007-04-01 20:49:36 +00:00
Reid Spencer
34b5a64f15 The bit counting intrinsics return i32 not the operand type. This fixes 
last night's regression in SingleSource/UnitTests/2005-05-11-Popcount-ffs-fls

llvm-svn: 35556
 2007-04-01 18:42:20 +00:00
Zhou Sheng
8b2b6b9e36 Avoid unnecessary APInt construction. 
llvm-svn: 35555
 2007-04-01 17:13:37 +00:00
Evan Cheng
949a34fbad Add i16 address mode. 
llvm-svn: 35551
 2007-04-01 08:06:46 +00:00
Reid Spencer
88fb617695 For PR1297: 
Support overloaded intrinsics bswap, ctpop, cttz, ctlz.

llvm-svn: 35547
 2007-04-01 07:35:23 +00:00
Reid Spencer
4a28a16efb For PR1297: 
Adjust for changes in the bit counting intrinsics. They all return i32
now so we have to trunc/zext the DAG node accordingly.

llvm-svn: 35546
 2007-04-01 07:34:11 +00:00
Reid Spencer
77c46f6cf5 For PR1297: 
Change getOperationName to return std::string instead of const char*

llvm-svn: 35545
 2007-04-01 07:32:19 +00:00
Reid Spencer
d371e2e492 For PR1297: 
Implement "actual" argument types for the Intrinsic member functions. This
involves changing the getName, getType, and getDeclaration methods to have
optional parameters for the actual types. These are necessary in order for
the type/name to be constructed properly for overloaded intrinsics. Only
the caller knows the actual argument types desired.

llvm-svn: 35541
 2007-04-01 07:25:33 +00:00
Reid Spencer
1db694fe4b For PR1297: 
1. Clear up confusion between "GotBits" and "ExpectBits". GotBits is the
   type actually provided. ExpectedBits is the type expected for the
   intrinsics. Before this patch, it was reversed!
2. Implement checks for overloaded intrinsics. This involves computing the
   suffix expected and making sure the suffix matches the function name. It
   also includes some intrinsic-specific checks such as ensuring that the
   bswap parameter and result are the same width and an even number of bytes.

llvm-svn: 35540
 2007-04-01 07:22:57 +00:00
Chris Lattner
aa4b06bdae Fix InstCombine/2007-03-31-InfiniteLoop.ll 
llvm-svn: 35536
 2007-04-01 05:36:37 +00:00
Nick Lewycky
03f9c0ffc9 Implement union of wrapped sets. 
llvm-svn: 35534
 2007-04-01 03:47:44 +00:00
Andrew Lenharth
ff1ff37be9 Readme 
llvm-svn: 35533
 2007-03-31 15:05:44 +00:00
Anton Korobeynikov
db09256c9f Consistency with native compilers 
llvm-svn: 35532
 2007-03-31 13:11:52 +00:00
Bill Wendling
8f16f4bf9c Fix comment. 
llvm-svn: 35531
 2007-03-31 09:36:12 +00:00
Chris Lattner
cea5e599ed add a note 
llvm-svn: 35530
 2007-03-31 07:06:25 +00:00
Chris Lattner
f01b0a800b move a bunch of code out of the sdisel pass into its own opt pass "codegenprepare". 
llvm-svn: 35529
 2007-03-31 04:18:03 +00:00
Chris Lattner
579b1c63d5 Split the sdisel code munging stuff out into its own opt-pass, CodeGenPrepare. 
llvm-svn: 35528
 2007-03-31 04:06:36 +00:00
Chris Lattner
8747705792 switch TL::getValueType to use MVT::getValueType. 
llvm-svn: 35527
 2007-03-31 04:05:24 +00:00
Chris Lattner
546138e3dd add a method to turn a type into a VT. 
llvm-svn: 35526
 2007-03-31 04:03:02 +00:00
Zhou Sheng
967239685a Delete dead code. 
llvm-svn: 35525
 2007-03-31 02:50:26 +00:00
Zhou Sheng
ebf4c32f0b Use APInt operators to calculate the carry bits, remove this loop. 
llvm-svn: 35524
 2007-03-31 02:38:39 +00:00
Bill Wendling
0cba5d3c15 Match GCC's MMX calling convention. 
llvm-svn: 35523
 2007-03-31 01:03:53 +00:00
Chris Lattner
330464a91c Add a -print-lsr-output option to LLC, to print the output of the LSR pass. 
llvm-svn: 35522
 2007-03-31 00:24:43 +00:00
Chris Lattner
c0405a348d implement the new addressing mode description hook. 
llvm-svn: 35521
 2007-03-30 23:15:24 +00:00
Chris Lattner
143464e05f add one addressing mode description hook to rule them all. 
llvm-svn: 35520
 2007-03-30 23:14:50 +00:00
Dale Johannesen
d4ab7d28e9 Fix incorrect combination of different loads. Reenable zext-over-truncate 
combination.

llvm-svn: 35517
 2007-03-30 21:38:07 +00:00
Evan Cheng
d814ad9d50 Don't add the same MI to register reuse "last def/use" twice if it reads the 
register more than once.

llvm-svn: 35513
 2007-03-30 20:21:35 +00:00
Evan Cheng
fc6519b457 Bug fix for PR1279. When isDead is propagate by copy coalescing, we keep length 
of dead def live interval at 1 to avoid multiple def's targeting the same
register. The previous patch missed a case where the source operand is live-in.
In that case, remove the whole interval.

llvm-svn: 35512
 2007-03-30 20:18:35 +00:00
Zhou Sheng
67bcf1cfb6 Make sure the use of ConstantInt::getZExtValue() for shift amount safe. 
llvm-svn: 35510
 2007-03-30 17:20:39 +00:00
Reid Spencer
e1fed6f90d For PR1293: 
* Rename the FunctionType* parameter from M to FT on all the functions.
* Implement a fix for PR1293 by just asserting that library functions that
  must return pointers should have pointer typed results. This just makes
  sure that we don't attempt to use an uninitialized integer or something
  later on.

llvm-svn: 35508
 2007-03-30 16:41:50 +00:00
Zhou Sheng
d667b5db0c 1. Make sure the use of ConstantInt::getZExtValue() for getting shift 
amount is safe.
2. Use new method on ConstantInt instead of (? :) operator.
3. Use new method uge() on ConstantInt to simplify codes.

llvm-svn: 35505
 2007-03-30 09:29:48 +00:00
Zhou Sheng
382723d8b4 Use APInt operation instead of ConstantExpr::getXX. 
llvm-svn: 35503
 2007-03-30 05:45:18 +00:00
Reid Spencer
bc87d8dbef For PR1289: Regenerate 
llvm-svn: 35500
 2007-03-30 01:37:39 +00:00
Reid Spencer
8a51bc0379 For PR1289: 
Check at the end of the parse that there are no unresolved types and no
undefined values. Issue errors if there are. This gets rid of the need
for implementation or checkpoint by ensuring you can't finish a parse with
undefined things.

llvm-svn: 35499
 2007-03-30 01:37:13 +00:00
Bill Wendling
503f3e41f6 Add MMX calling conventions. 
llvm-svn: 35489
 2007-03-30 00:35:22 +00:00
Evan Cheng
526a67c09b New entry. 
llvm-svn: 35480
 2007-03-29 21:40:13 +00:00
Evan Cheng
3a6b488265 Can't re-materialize mov r, imm in thumb since mov would clobber the condition code. 
llvm-svn: 35479
 2007-03-29 21:38:31 +00:00
Reid Spencer
e35e9d07b0 Use the stat information in the Path object, if it is already obtained. This 
avoids a call to ::fstat by MappedFile when the file size information was
already obtained by the Path object.

llvm-svn: 35477
 2007-03-29 19:11:22 +00:00
Reid Spencer
4269ae1274 For PR789: 
Make the sys::Path::getFileStatus function more efficient by having it
return a pointer to the FileStatus structure rather than copy it. Adjust
uses of the function accordingly. Also, fix some memory issues in sys::Path.

llvm-svn: 35476
 2007-03-29 19:05:44 +00:00
Chris Lattner
3a52e7b06d the bytecode reader supports dematerializeFunction 
llvm-svn: 35475
 2007-03-29 18:58:08 +00:00
Reid Spencer
6125321928 Regenerate. 
llvm-svn: 35473
 2007-03-29 18:50:01 +00:00
Reid Spencer
b4730512d7 Give users a grace period on the implementation keyword. This *will* get 
removed in a few days.

llvm-svn: 35472
 2007-03-29 18:49:28 +00:00
Jeff Cohen
67dbe2064f MS C does have inlining after all, just uses _inline instead of inline. 
llvm-svn: 35467
 2007-03-29 17:42:21 +00:00
Jeff Cohen
6e6a8a9128 Determine absolute paths the correct way :) 
llvm-svn: 35464
 2007-03-29 17:27:38 +00:00
Reid Spencer
d0eb672f1e Add a uniqueID field to the FileStatus structure for Paths. This will map 
to the inode number on Unix and something far less unique on Windows. The
windows case needs to be improved.

llvm-svn: 35461
 2007-03-29 17:00:31 +00:00
Reid Spencer
50b83584c4 For PR789: 
Updates for change in interface of getFileStatus method of sys::Path class.

llvm-svn: 35458
 2007-03-29 16:48:16 +00:00
Reid Spencer
fc67ef7cfe For PR789: 
* Add a method: bool isAbsolute() const, which determines if the path name
  is absolute or not.
* Implement caching of file status information in the Path object. Allow it
  to be updated forcefully or lazily re-fetched from the cached value.

llvm-svn: 35456
 2007-03-29 16:43:20 +00:00
Zhou Sheng
ac2b833c7a 1. Make more use of APInt::getHighBitsSet/getLowBitsSet. 
2. Let APInt variable do the binary operation stuff instead of using
   ConstantExpr::getXXX.

llvm-svn: 35450
 2007-03-29 08:15:12 +00:00
Evan Cheng
2d09850760 Disable load width reduction xform of variant (zext (truncate load x)) for 
big endian targets until llvm-gcc build issue has been resolved.

llvm-svn: 35449
 2007-03-29 07:56:46 +00:00
Evan Cheng
bde0409264 Add support for hidden visibility to darwin/arm. 
llvm-svn: 35448
 2007-03-29 07:49:34 +00:00
Zhou Sheng
684c2133f2 Clean up some codes in InstCombiner::SimplifyDemandedBits(). 
llvm-svn: 35446
 2007-03-29 04:45:55 +00:00
Evan Cheng
a72a3b7a74 New entries. 
llvm-svn: 35445
 2007-03-29 02:48:56 +00:00
Zhou Sheng
067efd7806 Clean up codes in InstCombiner::SimplifyDemandedBits(): 
1. Line out nested call of APInt::zext/trunc.
2. Make more use of APInt::getHighBitsSet/getLowBitsSet.
3. Use APInt[] operator instead of expression like "APIntVal & SignBit".

llvm-svn: 35444
 2007-03-29 02:26:30 +00:00
Zhou Sheng
74db84591c 1. Make the APInt variable do the binary operation stuff if possible 
instead of using ConstantExpr::getXX.
2. Use constant reference to APInt if possible instead of expensive
   APInt copy.

llvm-svn: 35443
 2007-03-29 01:57:21 +00:00
Jeff Cohen
9193778f39 Fix C Backend to generate code that works with Microsoft C for the benefit of 
front ends that do not depend on the GCC runtime (someday...).

llvm-svn: 35441
 2007-03-28 23:08:37 +00:00
Bill Wendling
0549ae5cdb Made this into a bug report: PR1286 
llvm-svn: 35439
 2007-03-28 19:07:34 +00:00
Anton Korobeynikov
1a8740c88b Oops :) 
llvm-svn: 35438
 2007-03-28 18:38:33 +00:00
Anton Korobeynikov
d59c4e54c7 Don't allow MatchAddress recurse too much. This trims exponential 
behaviour in some cases.

llvm-svn: 35437
 2007-03-28 18:36:33 +00:00
Chris Lattner
a710f7442c add a note 
llvm-svn: 35435
 2007-03-28 18:17:19 +00:00
Chris Lattner
26a37bfd6a Compile CodeGen/X86/lea-3.ll:test2 to: 
_test3:
        leaq (,%rdi,4), %rax
        orq %rdi, %rax
        ret

instead of:
_test2:
        movq %rdi, %rax
        shlq $2, %rax
        orq %rdi, %rax
        ret

llvm-svn: 35434
 2007-03-28 18:12:31 +00:00
Zhou Sheng
a098617383 Avoid unnecessary APInt construction. 
llvm-svn: 35431
 2007-03-28 17:38:21 +00:00
Chris Lattner
1f18012438 Dan pointed out that this is done, remove it! 
llvm-svn: 35430
 2007-03-28 17:26:52 +00:00
Scott Michel
ae7090641f First test check-in. 
llvm-svn: 35429
 2007-03-28 17:04:43 +00:00
Zhou Sheng
8fecc8c3a0 1. Make more use of getLowBitsSet/getHighBitsSet. 
2. Use APInt[] instead of "X & SignBit".
3. Clean up some codes.
4. Make the expression like "ShiftAmt = ShiftAmtC->getZExtValue()" safe.

llvm-svn: 35424
 2007-03-28 15:02:20 +00:00
Devang Patel
7cde72cbc1 It is not possible to determine dominance between two PHI nodes 
based on their ordering. This is applicable to ETForest::dominates() also.

llvm-svn: 35423
 2007-03-28 14:57:43 +00:00
Zhou Sheng
cf7d99a731 1. Make more use of getLowBitsSet/getHighBitsSet. 
2. Make the APInt value do the zext/trunc stuff instead of using
   ConstantExpr::getZExt().

llvm-svn: 35422
 2007-03-28 09:19:01 +00:00
Evan Cheng
29fcdd1b2f Notes on re-materialization. 
llvm-svn: 35420
 2007-03-28 08:30:04 +00:00
Evan Cheng
13037fbfb4 Move rematerialization out of beta. 
llvm-svn: 35419
 2007-03-28 08:26:40 +00:00
Zhou Sheng
2dc7cf0114 Use UnknownBIts[BitWidth-1] instead of UnknownBIts & SignBits. 
llvm-svn: 35418
 2007-03-28 05:15:57 +00:00
Zhou Sheng
e43a68551e Remove unused APInt variable. 
llvm-svn: 35414
 2007-03-28 03:02:21 +00:00
Reid Spencer
c50c59b4f6 Regenerate. 
llvm-svn: 35411
 2007-03-28 02:37:40 +00:00
Reid Spencer
f88fd1096c Replace the "implementation" keyword with "checkpoint". This more accurately 
describes its function as check against unresolved types and definitions.
This is handy for forcing error messages in assembly test cases or otherwise
ensuring that everything is resolved at that point in the assembly.

llvm-svn: 35410
 2007-03-28 02:36:42 +00:00
Zhou Sheng
7245616070 Clean up codes in ComputeMaskedBits(): 
1. Line out nested use of zext/trunc.
2. Make more use of getHighBitsSet/getLowBitsSet.
3. Use APInt[] != 0 instead of "(APInt & SignBit) != 0".

llvm-svn: 35408
 2007-03-28 02:19:03 +00:00
Evan Cheng
13cc34e91b Scale 1 is always ok. 
llvm-svn: 35407
 2007-03-28 01:55:52 +00:00
Evan Cheng
a55449c051 Remove isLegalAddressImmediate. 
llvm-svn: 35406
 2007-03-28 01:53:55 +00:00
Reid Spencer
bd89a6ea1c No need to generate the implementation keyword any more. Its frivolous. 
llvm-svn: 35405
 2007-03-28 01:53:20 +00:00
Evan Cheng
6056fd729d GEP index sinking fixes: 
1) Take address scale into consideration. e.g. i32* -> scale 4.
2) Examine all the users of GEP.
3) Generalize to inter-block GEP's (no longer uses loopinfo).
4) Don't do xform if GEP has other variable index(es).

llvm-svn: 35403
 2007-03-28 01:49:39 +00:00
Reid Spencer
3b614f9dad For PR1280: 
When converting an add/xor/and triplet into a trunc/sext, only do so if the
intermediate integer type is a bitwidth that the targets can handle.

llvm-svn: 35400
 2007-03-28 01:36:16 +00:00
Evan Cheng
a182608846 Fix for PR1279. Dead def has a live interval of length 1. Copy coalescing should 
not violate that.

llvm-svn: 35396
 2007-03-28 01:30:37 +00:00
Bill Wendling
e8eccb1684 Remove cruft I put in there... 
llvm-svn: 35394
 2007-03-28 01:02:54 +00:00
Chris Lattner
9f59529c61 Fix a problem building llvm-gcc on amd64-unknown-freebsd6.2, due to the 
system assembler not groking legal instructions like "leal (,%esi,8), %ecx".

llvm-svn: 35393
 2007-03-28 00:58:40 +00:00
Bill Wendling
1087888176 Unbreak mmx arithmetic. It was barfing trying to do v8i8 arithmetic. 
llvm-svn: 35392
 2007-03-28 00:57:11 +00:00
Bill Wendling
6b555c80c0 Add the "unpack low packed data" instructions. This should be the last of 
the MMX instructions that are needed...

llvm-svn: 35389
 2007-03-27 21:20:36 +00:00
Devang Patel
67eede36f7 Use SmallPtrSet. 
llvm-svn: 35388
 2007-03-27 20:50:46 +00:00
Bill Wendling
d43819da2f Fix so that pandn is emitted instead of an xor/and combo. Add integer 
comparison operators.

llvm-svn: 35385
 2007-03-27 20:22:40 +00:00
Evan Cheng
a23cc41b11 Unbreaks non-debug builds. 
llvm-svn: 35383
 2007-03-27 16:44:48 +00:00
Lauro Ramos Venancio
99fca527d3 "The C standards do say that "char" may either be a "signed char" or "unsigned 
char" and it is up to the compilers implementation or the platform which is
followed."
http://www.arm.linux.org.uk/docs/faqs/signedchar.php

llvm-svn: 35382
 2007-03-27 16:33:08 +00:00
Lauro Ramos Venancio
8b2e32de5a bugfix: sometimes the spiller puts a load between the "mov lr, pc" and "bx" of a CALL_NOLINK. 
llvm-svn: 35381
 2007-03-27 16:19:21 +00:00
Anton Korobeynikov
64622a0ddf Remove dead code 
llvm-svn: 35380
 2007-03-27 12:05:48 +00:00
Anton Korobeynikov
b58a93156f Split big monster into small helpers. No functionality change. 
llvm-svn: 35379
 2007-03-27 11:29:11 +00:00
Evan Cheng
7218d782fe SDISel does not preserve all, it changes CFG and other info. 
llvm-svn: 35376
 2007-03-27 00:53:36 +00:00
Evan Cheng
c81bdb4100 Don't call getOperandConstraint() if operand index is greater than 
TID->numOperands.

llvm-svn: 35375
 2007-03-27 00:48:28 +00:00
Devang Patel
8330ee3120 Spell check. 
llvm-svn: 35374
 2007-03-27 00:16:08 +00:00
Reid Spencer
e4b331a4f2 Implement some minor review feedback. 
llvm-svn: 35373
 2007-03-26 23:58:26 +00:00
Reid Spencer
0cdfe48fc8 For PR1271: 
Fix another incorrectly converted shift mask.

llvm-svn: 35371
 2007-03-26 23:45:51 +00:00
Devang Patel
c2708c8ccb Reduce malloc/free traffic. 
llvm-svn: 35370
 2007-03-26 23:19:29 +00:00
Devang Patel
58a8e74de6 It is not possible to determie dominance between two PHI nodes 
based on their ordering.

llvm-svn: 35369
 2007-03-26 23:18:28 +00:00
Devang Patel
3062ae001d Use std::vector<DFCalculateWorkObject> instead of 
std::vector<DFCalculateWorkObject *> to reduce malloc/free traffic.

llvm-svn: 35368
 2007-03-26 23:17:19 +00:00
Evan Cheng
968135bd3f findRegisterUseOperand() changed. 
llvm-svn: 35366
 2007-03-26 22:41:48 +00:00
Evan Cheng
0ff19780f9 Fix for PR1266. Don't mark a two address operand IsKill. 
llvm-svn: 35365
 2007-03-26 22:40:42 +00:00
Evan Cheng
201771637e Change findRegisterUseOperand() to return operand index instead. 
llvm-svn: 35363
 2007-03-26 22:37:45 +00:00
Dale Johannesen
3ad8ab7b61 Fix reversed logic in getRegsUsed. Rename RegStates to RegsAvailable to 
hopefully forestall similar errors.

llvm-svn: 35362
 2007-03-26 22:23:54 +00:00
Chris Lattner
16700ef424 eliminate use of std::set 
llvm-svn: 35361
 2007-03-26 20:40:50 +00:00
Reid Spencer
4af99334ea Add a comment to explain a folding transform. 
llvm-svn: 35360
 2007-03-26 20:09:02 +00:00
Reid Spencer
73d778edb7 Get better debug output by having modified instructions print both the 
original and new instruction. A slight performance hit with ostringstream
but it is only for debug.
Also, clean up an uninitialized variable warning noticed in a release build.

llvm-svn: 35358
 2007-03-26 17:44:01 +00:00
Reid Spencer
3fea170ecd Get the number of bits to set in a mask correct for a shl/lshr transform. 
llvm-svn: 35357
 2007-03-26 17:18:58 +00:00
Bill Wendling
8065cc3173 Promote to v1i64 type... 
llvm-svn: 35353
 2007-03-26 08:03:33 +00:00
Bill Wendling
3c4e130b77 Updated. 
llvm-svn: 35352
 2007-03-26 07:55:58 +00:00
Bill Wendling
a42484728c Add support for the v1i64 type. This makes better code for this: 
#include <mmintrin.h>

extern __m64 C;

void baz(__v2si *A, __v2si *B)
{
  *A = C;
  _mm_empty();
}

We get this:

_baz:
        call "L1$pb"
"L1$pb":
        popl %eax
        movl L_C$non_lazy_ptr-"L1$pb"(%eax), %eax
        movq (%eax), %mm0
        movl 4(%esp), %eax
        movq %mm0, (%eax)
        emms
        ret

GCC gives us this:

_baz:
        pushl   %ebx
        call    L3
"L00000000001$pb":
L3:
        popl    %ebx
        subl    $8, %esp
        movl    L_C$non_lazy_ptr-"L00000000001$pb"(%ebx), %eax
        movl    (%eax), %edx
        movl    4(%eax), %ecx
        movl    16(%esp), %eax
        movl    %edx, (%eax)
        movl    %ecx, 4(%eax)
        emms
        addl    $8, %esp
        popl    %ebx
        ret

llvm-svn: 35351
 2007-03-26 07:53:08 +00:00
Evan Cheng
5b1c21d27b SIGN_EXTEND_INREG requires one extra operand, a ValueType node. 
llvm-svn: 35350
 2007-03-26 07:12:51 +00:00
Reid Spencer
a6a4dbd24e For PR1271: 
Fix SingleSource/Regression/C/2003-05-21-UnionBitFields.c by changing a
getHighBitsSet call to getLowBitsSet call that was incorrectly converted
from the original lshr constant expression.

llvm-svn: 35348
 2007-03-26 05:25:00 +00:00
Dale Johannesen
43fc14d6e4 Look through bitcast when finding IVs. (Chris' patch really.) 
llvm-svn: 35347
 2007-03-26 03:01:27 +00:00
Reid Spencer
4d3d7d0f2a For PR1271: 
Remove a use of getLowBitsSet that caused the mask used for replacement of
shl/lshr pairs with an AND instruction to be computed incorrectly. Its not
clear exactly why this is the case. This solves the disappearing shifts
problem, but it doesn't fix Regression/C/2003-05-21-UnionBitFields. It
seems there is more going on.

llvm-svn: 35342
 2007-03-25 21:11:44 +00:00
Chris Lattner
26cfd91db8 implement Transforms/InstCombine/cast2.ll:test3 and PR1263 
llvm-svn: 35341
 2007-03-25 20:43:09 +00:00
Reid Spencer
1677543a3a Some cleanup from review: 
* Don't assume shift amounts are <= 64 bits
* Avoid creating an extra APInt in SubOne and AddOne by using -- and ++
* Add another use of getLowBitsSet
* Convert a series of if statements to a switch

llvm-svn: 35339
 2007-03-25 19:55:33 +00:00
Anton Korobeynikov
6f78c59650 First step of switch lowering refactoring: perform worklist-driven 
strategy, emit JT's where possible.

llvm-svn: 35338
 2007-03-25 15:07:15 +00:00
Anton Korobeynikov
a5e854461f Fix authorship 
llvm-svn: 35337
 2007-03-25 13:44:26 +00:00
Chris Lattner
f4ff1dc689 fold constantexprs more aggressively, fixing PR1265 
llvm-svn: 35336
 2007-03-25 05:47:04 +00:00
Reid Spencer
8449e5fa0e Refactor several ConstantExpr::getXXX calls with ConstantInt arguments 
using the facilities of APInt. While this duplicates a tiny fraction of
the constant folding code, it also makes the code easier to read and
avoids large ConstantExpr overhead for simple, known computations.

llvm-svn: 35335
 2007-03-25 05:33:51 +00:00
Chris Lattner
ee90a087a6 add a note 
llvm-svn: 35334
 2007-03-25 05:10:46 +00:00
Zhou Sheng
3312a86598 1. Avoid unnecessary APInt construction if possible. 
2. Use isStrictlyPositive() instead of isPositive() in two places where
   they need APInt value > 0 not only >=0.

llvm-svn: 35333
 2007-03-25 05:01:29 +00:00
Chris Lattner
6f17a615cb Implement support for vector operands to inline asm, implementing 
CodeGen/X86/2007-03-24-InlineAsmVectorOp.ll

llvm-svn: 35332
 2007-03-25 05:00:54 +00:00
Chris Lattner
6d8bbccee0 add a note 
llvm-svn: 35330
 2007-03-25 04:46:28 +00:00
Chris Lattner
31c6763051 Fix CodeGen/PowerPC/2007-03-24-cntlzd.ll 
llvm-svn: 35329
 2007-03-25 04:44:03 +00:00
Chris Lattner
4a7feb72f5 implement initial support for the silly X constraint. Testcase here: CodeGen/X86/2007-03-24-InlineAsmXConstraint.ll 
llvm-svn: 35327
 2007-03-25 04:35:41 +00:00
Reid Spencer
e601991a5e Make more uses of getHighBitsSet and get rid of some pointless & of an 
APInt with its type mask.

llvm-svn: 35325
 2007-03-25 04:26:16 +00:00
Chris Lattner
91adfbe0ff Implement CodeGen/X86/2007-03-24-InlineAsmMultiRegConstraint.ll 
llvm-svn: 35324
 2007-03-25 02:18:14 +00:00
Chris Lattner
b19069959d switch TargetLowering::getConstraintType to take the entire constraint, 
not just the first letter.  No functionality change.

llvm-svn: 35322
 2007-03-25 02:14:49 +00:00
Reid Spencer
676eeb9c94 More APIntification: 
* Convert the last use of a uint64_t that should have been an APInt.
* Change ComputeMaskedBits to have a const reference argument for the Mask
  so that recursions don't cause unneeded temporaries. This causes temps
  to be needed in other places (where the mask has to change) but this
  change optimizes for the recursion which is more frequent.
* Remove two instances of &ing a Mask with getAllOnesValue. Its not
  needed any more because APInt is accurate in its bit computations.
* Start using the getLowBitsSet and getHighBits set methods on APInt
  instead of shifting. This makes it more clear in the code what is
  going on.

llvm-svn: 35321
 2007-03-25 02:03:12 +00:00
Chris Lattner
18c3c6a01d Allow the b/h/w/k constraints to be applied to values that have multiple alternatives, and end up not being registers. 
llvm-svn: 35320
 2007-03-25 02:01:03 +00:00
Chris Lattner
104e73382c enforce the proper range for the i386 N constraint. 
llvm-svn: 35319
 2007-03-25 01:57:35 +00:00
Chris Lattner
92e2ee7b2d Fix test/CodeGen/X86/2007-03-24-InlineAsmPModifier.ll 
llvm-svn: 35318
 2007-03-25 01:44:57 +00:00
Chris Lattner
3f71a6b498 fix a regression on vector or instructions. 
llvm-svn: 35314
 2007-03-24 23:56:43 +00:00
Chris Lattner
57d0c197e6 don't rely on ADL 
llvm-svn: 35299
 2007-03-24 17:37:03 +00:00