Commit Graph

31155 Commits

Author SHA1 Message Date
Chris Lattner 50490d54f2 use more obvious function name.
llvm-svn: 35587
2007-04-02 05:42:22 +00:00
Chris Lattner a3e0bb4ebb Treat xor of signbit like an add.
llvm-svn: 35586
2007-04-02 05:41:38 +00:00
Chris Lattner d14447833a add a helper function.
llvm-svn: 35585
2007-04-02 05:41:00 +00:00
Chris Lattner b24acc7bee 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 2d81c6d706 creative way to add one.
llvm-svn: 35583
2007-04-02 05:35:08 +00:00
Chris Lattner 0a844d8df1 my patch fixed the fixme.
llvm-svn: 35582
2007-04-02 05:08:54 +00:00
Reid Spencer e51961b5ba Fix illegal assembly syntax.
llvm-svn: 35581
2007-04-02 03:24:47 +00:00
Reid Spencer f361c4f4f8 bwsap -> bswap
llvm-svn: 35580
2007-04-02 02:25:19 +00:00
Reid Spencer 69e56c3461 Regenerate
llvm-svn: 35579
2007-04-02 02:08:35 +00:00
Reid Spencer 940fccb088 Upgrade the bit count intrinsics to have an i32 result.
llvm-svn: 35578
2007-04-02 02:08:05 +00:00
Reid Spencer a3bc850712 Add a test case to make sure that constant folding of the bit counting
intrinsics works.

llvm-svn: 35577
2007-04-02 01:45:31 +00:00
Chris Lattner b7b75145f1 reduce use of std::set
llvm-svn: 35576
2007-04-02 01:44:59 +00:00
Chris Lattner c3748562bd 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
Chris Lattner 4cb30067c2 new testcase for PR1296
llvm-svn: 35574
2007-04-02 01:32:59 +00:00
Lauro Ramos Venancio 6be85337b0 - 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
Chris Lattner 7c773e6850 add range version of insert()
llvm-svn: 35572
2007-04-02 01:21:19 +00:00
Reid Spencer 1bd88bdab3 Regenerate
llvm-svn: 35571
2007-04-02 01:14:00 +00:00
Reid Spencer 86207ee54d Don't upgrade functions that look like a bswap intrinsic but aren't.
llvm-svn: 35570
2007-04-02 01:13:36 +00:00
Reid Spencer 84c90a9431 Let llvm-upgrade upgrade the bswap intrinsic.
llvm-svn: 35569
2007-04-02 01:09:19 +00:00
Reid Spencer 0b544372d4 Let llvm-upgrade upgrade bswap intrinsic.
llvm-svn: 35568
2007-04-02 01:08:02 +00:00
Reid Spencer fad9bd6b92 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
Reid Spencer a5f996bd27 Revert the name changes for llvm.bswap to allow (and test) llvm-upgrade of
this intrinsic.

llvm-svn: 35566
2007-04-02 00:51:15 +00:00
Reid Spencer ef592214e2 Handle upgrade of llvm.bswap.iXX to llvm.bswap.iXX.iXX per new naming
rules for overloaded intrinsic functions.

llvm-svn: 35565
2007-04-02 00:50:28 +00:00
Reid Spencer 403a1c4517 Fix some bswap typos that Chris noticed.
llvm-svn: 35564
2007-04-02 00:19:52 +00:00
Chris Lattner 8fe3cbe6bd print the type of an inserted IV in -debug mode.
llvm-svn: 35563
2007-04-01 22:21:39 +00:00
Anton Korobeynikov ecc769e4f2 Add testcase for correct sret handling.
llvm-svn: 35562
2007-04-01 22:03:04 +00:00
Chris Lattner c3eeb42809 simplify this code, make it work for ap ints
llvm-svn: 35561
2007-04-01 20:57:36 +00:00
Chris Lattner 59a6fa7af6 fix breakage from last night, simplify code.
llvm-svn: 35560
2007-04-01 20:49:36 +00:00
Chris Lattner 489ad3313a gcc 3.4.4 is known-bad on x86-64
llvm-svn: 35559
2007-04-01 20:14:46 +00:00
Reid Spencer 8bc7d957cf Preview documentation for additional intrinsic functions.
llvm-svn: 35558
2007-04-01 19:00:37 +00:00
Jeff Cohen 3e96df52d8 Unbreak VC++ build.
llvm-svn: 35557
2007-04-01 18:58:22 +00:00
Reid Spencer add6123405 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 150f3bbab2 Avoid unnecessary APInt construction.
llvm-svn: 35555
2007-04-01 17:13:37 +00:00
Duncan Sands 2e45201769 Testcase for the problem worked around in
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20070319/046204.html

llvm-svn: 35554
2007-04-01 15:29:02 +00:00
Zhou Sheng 63b6bc1c88 Remove unused methods.
llvm-svn: 35553
2007-04-01 12:45:33 +00:00
Reid Spencer 8a5799f558 For PR1297:
Preview documentation for the llvm.bit.concat intrinsic (yet to be
implemented).

llvm-svn: 35552
2007-04-01 08:27:01 +00:00
Evan Cheng 17d48a8bc2 Add i16 address mode.
llvm-svn: 35551
2007-04-01 08:06:46 +00:00
Reid Spencer 4eefaab3f6 For PR1297:
Document changes to bswap and bit counting intrinsics. bswap's name now
requires two types in the suffix per overloaded intrinsic naming rules.
The ctpop, cttz, and ctlz intrinsics were changed to always return i32.

llvm-svn: 35550
2007-04-01 08:04:23 +00:00
Reid Spencer 9ade28d260 Mention how to create overloaded intrinsics.
llvm-svn: 35549
2007-04-01 07:44:52 +00:00
Reid Spencer c3d87685ad For PR1297:
Update these test cases to use proper signatures for bswap which is now
and overloaded intrinsic. Its name must be of the form llvm.bswap.i32.i32
since both the parameter and the result or of type "iAny". Also, the
bit counting intrinsics changed to always return i32.

llvm-svn: 35548
2007-04-01 07:36:28 +00:00
Reid Spencer 6bba6c8143 For PR1297:
Support overloaded intrinsics bswap, ctpop, cttz, ctlz.

llvm-svn: 35547
2007-04-01 07:35:23 +00:00
Reid Spencer 3a0843e734 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 a090ffb2ab 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 eba021453b For PR1297:
Implement bswap, ctpop, ctlz, and cttz utilizing the new overloaded
intrinsic feature of tblgen. By defining llvm_int_ty to have an ValueType
of iAny, any intrinsic using that type becomes an overloaded intrinsic.
This signals that a) any integer type could be used for that argument and
b) the intrinsic needs a suffix to match each such argument for the
actual intrinsic.

llvm-svn: 35544
2007-04-01 07:30:57 +00:00
Reid Spencer c7c4880ab1 For PR1297:
Intrinsic functions might need to be generated (if they are overloaded) so
returning a const char* is no longer possible. Change getOperationName to
return a std::string instead.

llvm-svn: 35543
2007-04-01 07:28:37 +00:00
Reid Spencer be23fb0370 For PR1297:
Change the interface to the getName, getType, getDeclaration methods to
include the "actual" types of the result/arguments. This is necessary to
support overloaded intrinsics.

llvm-svn: 35542
2007-04-01 07:26:35 +00:00
Reid Spencer 2a2117c82d 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 7c57b88b27 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
Reid Spencer e67d0c226d For PR1297:
Implement code generation for overloaded intrinsic functions. The basic
difference is that "actual" argument types must be provided when
constructing intrinsic names and types. Also, for recognition, only the
prefix is examined. If it matches, the suffix is assumed to match. The
suffix is checked by the Verifier, however.

llvm-svn: 35539
2007-04-01 07:20:02 +00:00
Reid Spencer 51455e00dc For PR1297:
Add an "iAny" ValueType as a place holder for an integer type of any
bit width. This is used for intrinsics overloading on integer type.

llvm-svn: 35538
2007-04-01 07:17:45 +00:00
Chris Lattner f517e9578d not not!
llvm-svn: 35537
2007-04-01 06:19:41 +00:00
Chris Lattner 0427799531 Fix InstCombine/2007-03-31-InfiniteLoop.ll
llvm-svn: 35536
2007-04-01 05:36:37 +00:00
Chris Lattner 9729bdd8e3 New testcase
llvm-svn: 35535
2007-04-01 05:34:53 +00:00
Nick Lewycky f22938af1b Implement union of wrapped sets.
llvm-svn: 35534
2007-04-01 03:47:44 +00:00
Andrew Lenharth 427ec7fa51 Readme
llvm-svn: 35533
2007-03-31 15:05:44 +00:00
Anton Korobeynikov a8cc1ebae1 Consistency with native compilers
llvm-svn: 35532
2007-03-31 13:11:52 +00:00
Bill Wendling b72fcddd23 Fix comment.
llvm-svn: 35531
2007-03-31 09:36:12 +00:00
Chris Lattner 075b4db621 add a note
llvm-svn: 35530
2007-03-31 07:06:25 +00:00
Chris Lattner f6a6d3c8b0 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 f2836d17b6 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 f2d71d49e2 switch TL::getValueType to use MVT::getValueType.
llvm-svn: 35527
2007-03-31 04:05:24 +00:00
Chris Lattner 516f38b35f add a method to turn a type into a VT.
llvm-svn: 35526
2007-03-31 04:03:02 +00:00
Zhou Sheng 82c42284f4 Delete dead code.
llvm-svn: 35525
2007-03-31 02:50:26 +00:00
Zhou Sheng 4f16402e0d Use APInt operators to calculate the carry bits, remove this loop.
llvm-svn: 35524
2007-03-31 02:38:39 +00:00
Bill Wendling afddb2c6f8 Match GCC's MMX calling convention.
llvm-svn: 35523
2007-03-31 01:03:53 +00:00
Chris Lattner bafc837c83 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 1eb94d973a implement the new addressing mode description hook.
llvm-svn: 35521
2007-03-30 23:15:24 +00:00
Chris Lattner ac3f81508c add one addressing mode description hook to rule them all.
llvm-svn: 35520
2007-03-30 23:14:50 +00:00
Evan Cheng 67a49736af Missing '&&'
llvm-svn: 35519
2007-03-30 22:03:49 +00:00
Jeff Cohen 2cdce430e2 This is working again.
llvm-svn: 35518
2007-03-30 22:02:18 +00:00
Dale Johannesen 4bbd2eefba Fix incorrect combination of different loads. Reenable zext-over-truncate
combination.

llvm-svn: 35517
2007-03-30 21:38:07 +00:00
Chris Lattner d1440c429f add a testcase for x86
llvm-svn: 35516
2007-03-30 21:22:46 +00:00
Evan Cheng 1d78d3deeb New test case.
llvm-svn: 35515
2007-03-30 20:41:18 +00:00
Evan Cheng 9a2a7b174a 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 48dd5e7d44 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
Evan Cheng 9ca63d605c Test case for PR1279 (part 2).
llvm-svn: 35511
2007-03-30 20:15:22 +00:00
Zhou Sheng fd28a33031 Make sure the use of ConstantInt::getZExtValue() for shift amount safe.
llvm-svn: 35510
2007-03-30 17:20:39 +00:00
Zhou Sheng 02d3ac6895 Make sure this method just return value equal or less than Limit.
llvm-svn: 35509
2007-03-30 16:50:28 +00:00
Reid Spencer 8c3d3dce2b 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
Duncan Sands 3bf01520eb Remove reference to checkpoint.
llvm-svn: 35507
2007-03-30 16:09:58 +00:00
Duncan Sands 16f122e6ad Various hopefully correct easy fixes.
llvm-svn: 35506
2007-03-30 12:22:09 +00:00
Zhou Sheng b25806fa5f 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
Duncan Sands 3dd14331ac Correct typo.
llvm-svn: 35504
2007-03-30 06:39:42 +00:00
Zhou Sheng 5e60a4a6b0 Use APInt operation instead of ConstantExpr::getXX.
llvm-svn: 35503
2007-03-30 05:45:18 +00:00
Zhou Sheng 9737575772 Rename and correct the method "greaterOrEqual".
llvm-svn: 35502
2007-03-30 05:10:59 +00:00
Zhou Sheng 12b3549f5d Add two utility methods into ConstantInt.
llvm-svn: 35501
2007-03-30 03:22:55 +00:00
Reid Spencer ef01c47512 For PR1289: Regenerate
llvm-svn: 35500
2007-03-30 01:37:39 +00:00
Reid Spencer cf2ccbff1a 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
Reid Spencer 3c401f0613 For PR1289:
Remove an unneeded instance of "implementation" that soon won't be supported.

llvm-svn: 35498
2007-03-30 01:35:32 +00:00
Bill Wendling 2fbcbc16b4 Add MMX calling conventions.
llvm-svn: 35489
2007-03-30 00:35:22 +00:00
Evan Cheng 045414aa8e New entry.
llvm-svn: 35480
2007-03-29 21:40:13 +00:00
Evan Cheng cc44b1e743 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 d232f311cb Remove some non-sensical logic that prevented llvm-nm from working on any
file other than one named "-".

llvm-svn: 35478
2007-03-29 19:49:07 +00:00
Reid Spencer c7f485fd71 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 200c6f9c3d 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 7c82675afa the bytecode reader supports dematerializeFunction
llvm-svn: 35475
2007-03-29 18:58:08 +00:00
Reid Spencer b9d7f2a3ba checkpoint -> implementation. This really needs to just drop the need for
a keyword by having the assembler check for incomplete types and values
upon end-of-file.

llvm-svn: 35474
2007-03-29 18:56:58 +00:00
Reid Spencer 0833b81248 Regenerate.
llvm-svn: 35473
2007-03-29 18:50:01 +00:00
Reid Spencer 01a599a202 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
Chris Lattner c3f4d87c90 add a hook to demat functions.
llvm-svn: 35471
2007-03-29 18:30:26 +00:00