Chris Lattner 8a594489bf Fix a long standing deficiency in the X86 backend: we would ... sometimes emit "zero" and "all one" vectors multiple times, for example: _test2: pcmpeqd %mm0, %mm0 movq %mm0, _M1 pcmpeqd %mm0, %mm0 movq %mm0, _M2 ret instead of: _test2: pcmpeqd %mm0, %mm0 movq %mm0, _M1 movq %mm0, _M2 ret This patch fixes this by always arranging for zero/one vectors to be defined as v4i32 or v2i32 (SSE/MMX) instead of letting them be any random type. This ensures they get trivially CSE'd on the dag. This fix is also important for LegalizeDAGTypes, as it gets unhappy when the x86 backend wants BUILD_VECTOR(i64 0) to be legal even when 'i64' isn't legal. This patch makes the following changes: 1) X86TargetLowering::LowerBUILD_VECTOR now lowers 0/1 vectors into their canonical types. 2) The now-dead patterns are removed from the SSE/MMX .td files. 3) All the patterns in the .td file that referred to immAllOnesV or immAllZerosV in the wrong form now use *_bc to match them with a bitcast wrapped around them. 4) X86DAGToDAGISel::SelectScalarSSELoad is generalized to handle bitcast'd zero vectors, which simplifies the code actually. 5) getShuffleVectorZeroOrUndef is updated to generate a shuffle that is legal, instead of generating one that is illegal and expecting a later legalize pass to clean it up. 6) isZeroShuffle is generalized to handle bitcast of zeros. 7) several other minor tweaks. This patch is definite goodness, but has the potential to cause random code quality regressions. Please be on the lookout for these and let me know if they happen. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@44310 91177308-0d34-0410-b5e6-96231b3b80d8		2007-11-25 00:24:49 +00:00
..
Analysis	Ding dong, the DoesntAccessMemoryFns and	2007-11-23 19:30:27 +00:00
Archive	Delete the svn:executable property on these files, which aren't executable.	2007-07-23 19:26:08 +00:00
Assembler	Fix PR1645 by resolving forward alias references.	2007-09-10 23:23:53 +00:00
Bindings/Ocaml	C and Objective Caml bindings for PHINode::addIncoming etc.	2007-10-08 18:14:39 +00:00
Bitcode	This is the patch to provide clean intrinsic function overloading support in LLVM. It cleans up the intrinsic definitions and generally smooths the process for more complicated intrinsic writing. It will be used by the upcoming atomic intrinsics as well as vector and float intrinsics in the future.	2007-08-04 01:51:18 +00:00
BugPoint	Convert .cvsignore files	2007-06-29 16:35:07 +00:00
C++Frontend	New test.	2007-10-02 00:39:19 +00:00
CFrontend	Un XFAIL these tests, now that Bill has backported	2007-11-14 13:40:53 +00:00
CodeGen	Fix a long standing deficiency in the X86 backend: we would	2007-11-25 00:24:49 +00:00
DebugInfo	For PR1558:	2007-07-17 06:20:38 +00:00
ExecutionEngine	Convert .cvsignore files	2007-06-29 16:35:07 +00:00
Feature	Add support for vectors to int <-> float casts.	2007-11-17 03:58:34 +00:00
FrontendAda	For multipart tests, place the parts with no	2007-07-23 13:41:53 +00:00
FrontendObjC	Thanks to the XTARGET line, this test should still	2007-11-09 13:50:25 +00:00
Integer	Change all floating constants that are not exactly	2007-09-05 17:50:36 +00:00
lib	Implement the XTARGET feature for PR1778	2007-11-06 22:32:17 +00:00
Linker	Add a test to ensure that obvious link messages are actually produced on	2007-08-16 07:22:43 +00:00
Other	New test.	2007-09-10 18:12:52 +00:00
Scripts	Add a test script for counting lines. This can be used instead of the	2007-08-10 15:07:05 +00:00
TableGen	Convert tests using "| wc -l | grep ..." to use the count script.	2007-08-15 13:36:28 +00:00
Transforms	Fix PR1816, by correcting the broken definition of APInt::countTrailingZeros.	2007-11-23 22:42:31 +00:00
Verifier	Fix for PR1633: Verifier doesn't fully verify GC intrinsics	2007-09-17 20:30:04 +00:00
Makefile	Add explicit --enable-bindings option to configure.	2007-10-02 09:50:18 +00:00
Makefile.tests	Failure.sh is dead, remove it.	2007-04-21 20:46:55 +00:00
TestRunner.sh	Avoid an infinite loop when TestRunner.sh is run outside of the test dir.	2007-05-18 02:06:03 +00:00