Nate Begeman 058d4eeccf Adapt the x86 build_vector dagcombine to the current state of the legalizer. ... build vectors with i64 elements will only appear on 32b x86 before legalize. Since vector widening occurs during legalize, and produces i64 build_vector elements, the dag combiner is never run on these before legalize splits them into 32b elements. Teach the build_vector dag combine in x86 back end to recognize consecutive loads producing the low part of the vector. Convert the two uses of TLI's consecutive load recognizer to pass LoadSDNodes since that was required implicitly. Add a testcase for the transform. Old: subl $28, %esp movl 32(%esp), %eax movl 4(%eax), %ecx movl %ecx, 4(%esp) movl (%eax), %eax movl %eax, (%esp) movaps (%esp), %xmm0 pmovzxwd %xmm0, %xmm0 movl 36(%esp), %eax movaps %xmm0, (%eax) addl $28, %esp ret New: movl 4(%esp), %eax pmovzxwd (%eax), %xmm0 movl 8(%esp), %eax movaps %xmm0, (%eax) ret llvm-svn: 72957		2009-06-05 21:37:30 +00:00
..
CallingConvLower.cpp	Minor top-level comment fix.	2009-03-31 16:51:18 +00:00
CMakeLists.txt	Update for CMakeLists; untested, so tell me if there are issues.	2009-05-24 09:13:13 +00:00
DAGCombiner.cpp	Adapt the x86 build_vector dagcombine to the current state of the legalizer.	2009-06-05 21:37:30 +00:00
FastISel.cpp	Split the Add, Sub, and Mul instruction opcodes into separate	2009-06-04 22:49:04 +00:00
LegalizeDAG.cpp	Re-commit r72514 and r72516 with a fixed version of BR_CC lowering.	2009-05-28 20:40:34 +00:00
LegalizeFloatTypes.cpp	Rename CustomLowerResults to CustomLowerNode, since	2009-05-31 04:15:38 +00:00
LegalizeIntegerTypes.cpp	Allow libcalls for i16 sdiv/udiv/rem operations.	2009-06-05 14:41:10 +00:00
LegalizeTypes.cpp	Rename CustomLowerResults to CustomLowerNode, since	2009-05-31 04:15:38 +00:00
LegalizeTypes.h	Revert 72707 and 72709, for the moment.	2009-06-02 03:12:52 +00:00
LegalizeTypesGeneric.cpp	Refactor the code to grab the low and high parts of a value	2009-04-25 17:55:53 +00:00
LegalizeVectorOps.cpp	Remove special cases for more opcodes.	2009-05-27 07:58:35 +00:00
LegalizeVectorTypes.cpp	Implement review feedback for vector shuffle work.	2009-04-29 05:20:52 +00:00
Makefile	Removed trailing whitespace from Makefiles.	2009-01-09 16:44:42 +00:00
ScheduleDAGFast.cpp	Instead of passing in an unsigned value for the optimization level, use an enum,	2009-04-29 23:29:43 +00:00
ScheduleDAGList.cpp	Instead of passing in an unsigned value for the optimization level, use an enum,	2009-04-29 23:29:43 +00:00
ScheduleDAGRRList.cpp	Instead of passing in an unsigned value for the optimization level, use an enum,	2009-04-29 23:29:43 +00:00
ScheduleDAGSDNodes.cpp	Revert 72707 and 72709, for the moment.	2009-06-02 03:12:52 +00:00
ScheduleDAGSDNodes.h	Rename COPY_TO_SUBCLASS to COPY_TO_REGCLASS, and generalize	2009-04-13 21:06:25 +00:00
ScheduleDAGSDNodesEmit.cpp	Silence Release-Asserts warnings.	2009-05-18 16:43:04 +00:00
SelectionDAG.cpp	Don't do the X * 0.0 -> 0.0 transformation in instcombine, because	2009-06-04 17:12:12 +00:00
SelectionDAGBuild.cpp	Split the Add, Sub, and Mul instruction opcodes into separate	2009-06-04 22:49:04 +00:00
SelectionDAGBuild.h	Split the Add, Sub, and Mul instruction opcodes into separate	2009-06-04 22:49:04 +00:00
SelectionDAGISel.cpp	Add a new step to legalization to legalize vector math operations. This	2009-05-23 12:35:30 +00:00
SelectionDAGPrinter.cpp	Fix some significant problems with constant pools that resulted in unnecessary paddings between constant pool entries, larger than necessary alignments (e.g. 8 byte alignment for .literal4 sections), and potentially other issues.	2009-03-13 07:51:59 +00:00
TargetLowering.cpp	Adapt the x86 build_vector dagcombine to the current state of the legalizer.	2009-06-05 21:37:30 +00:00