Eli Friedman 251919f579 [ARM] Adjust AND immediates to make them cheaper to select. ... LLVM normally prefers to minimize the number of bits set in an AND immediate, but that doesn't always match the available ARM instructions. In Thumb1 mode, prefer uxtb or uxth where possible; otherwise, prefer a two-instruction sequence movs+ands or movs+bics. Some potential improvements outlined in ARMTargetLowering::targetShrinkDemandedConstant, but seems to work pretty well already. The ARMISelDAGToDAG fix ensures we don't generate an invalid UBFX instruction due to a larger-than-expected mask. (It's orthogonal, in some sense, but as far as I can tell it's either impossible or nearly impossible to reproduce the bug without this change.) According to my testing, this seems to consistently improve codesize by a small amount by forming bic more often for ISD::AND with an immediate. Differential Revision: https://reviews.llvm.org/D50030 llvm-svn: 339472		2018-08-10 21:21:53 +00:00
..
2007-01-31-RegInfoAssert.ll
2007-02-02-JoinIntervalsCrash.ll
2007-05-05-InvalidPushPop.ll
2009-06-18-ThumbCommuteMul.ll
2009-07-20-TwoAddrBug.ll
2009-07-27-PEIAssert.ll
2009-08-12-ConstIslandAssert.ll
2009-08-12-RegInfoAssert.ll
2009-08-20-ISelBug.ll
2009-12-17-pre-regalloc-taildup.ll
2010-06-18-SibCallCrash.ll
2010-07-01-FuncAlign.ll
2010-07-15-debugOrdering.ll
2011-05-11-DAGLegalizer.ll	Remove alignment argument from memcpy/memmove/memset in favour of alignment attributes (Step 1)	2018-01-19 17:13:12 +00:00
2011-06-16-NoGPRs.ll
2011-EpilogueBug.ll
2012-04-26-M0ISelBug.ll
2014-06-10-thumb1-ldst-opt-bug.ll
addr-modes.ll
and_neg.ll
asmprinter-bug.ll
barrier.ll
bic_imm.ll	[ARM] Adjust AND immediates to make them cheaper to select.	2018-08-10 21:21:53 +00:00
branchless-cmp.ll	[ARM] Materialise some boolean values to avoid a branch	2018-02-16 09:23:59 +00:00
callee_save.ll
cmp-add-fold.ll
cmp-fold.ll
constants.ll	[ARM] Materialise some boolean values to avoid a branch	2018-02-16 09:23:59 +00:00
copy_thumb.ll
cortex-m0-unaligned-access.ll
DbgValueOtherTargets.test
dyn-stackalloc.ll	Remove alignment argument from memcpy/memmove/memset in favour of alignment attributes (Step 1)	2018-01-19 17:13:12 +00:00
fastcc.ll
fpconv.ll
fpow.ll
frame_thumb.ll
frame-access.ll	[ARM] Fix access to stack arguments when re-aligning SP in Armv6m	2018-03-02 15:47:14 +00:00
i8-phi-ext.ll	[CodeGen] Emit more precise AssertZext/AssertSext nodes.	2018-07-11 23:26:35 +00:00
iabs.ll
inlineasm-imm-thumb.ll
inlineasm-thumb.ll
ispositive.ll
large-stack.ll
ldm-merge-call.ll
ldm-merge-struct.ll
ldm-stm-base-materialization-thumb2.ll	Remove alignment argument from memcpy/memmove/memset in favour of alignment attributes (Step 1)	2018-01-19 17:13:12 +00:00
ldm-stm-base-materialization.ll	Remove alignment argument from memcpy/memmove/memset in favour of alignment attributes (Step 1)	2018-01-19 17:13:12 +00:00
ldm-stm-postinc.ll
ldr_ext.ll
ldr_frame.ll
lit.local.cfg
litpoolremat.ll
long_shift.ll
long-setcc.ll	[ARM] Materialise some boolean values to avoid a branch	2018-02-16 09:23:59 +00:00
long.ll	[ARM] Return true in enableMultipleCopyHints().	2018-02-16 09:51:01 +00:00
machine-cse-physreg.mir	Followup on Proposal to move MIR physical register namespace to '$' sigil.	2018-01-31 22:04:26 +00:00
mature-mc-support.ll
mul.ll
mvn.ll	[ARM] Fix issue with large xor constants.	2018-02-22 09:38:57 +00:00
optionaldef-scheduling.ll
pop.ll
pr35836_2.ll	[ARM] Allow the scheduler to clone a node with glue to avoid a copy CPSR ↔ GPR.	2018-01-31 09:23:43 +00:00
pr35836.ll	Re-enable "[MachineCopyPropagation] Extend pass to do COPY source forwarding"	2018-02-27 16:59:10 +00:00
PR17309.ll
PR35481.ll
PR36658.mir	[ARM] Fix "Constant pool entry out of range!" in Thumb1 mode	2018-03-23 17:53:27 +00:00
push.ll
remove-unneeded-push-pop.ll
rev.ll
segmented-stacks-dynamic.ll
segmented-stacks.ll
select.ll	Revert "[ARM] Lower lower saturate to 0 and lower saturate to -1 using bit-operations"	2018-01-31 22:55:19 +00:00
shift-and.ll	[ARM] Adjust AND immediates to make them cheaper to select.	2018-08-10 21:21:53 +00:00
sjljehprepare-lower-vector.ll
stack_guard_remat.ll
stack-access.ll
stack-coloring-without-frame-ptr.ll	Remove alignment argument from memcpy/memmove/memset in favour of alignment attributes (Step 1)	2018-01-19 17:13:12 +00:00
stack-frame.ll
stm-deprecated.ll
stm-merge.ll
stm-scavenging.ll	[LivePhysRegs] Fix handling of return instructions.	2018-02-06 23:00:17 +00:00
tbb-reuse.mir	Followup on Proposal to move MIR physical register namespace to '$' sigil.	2018-01-31 22:04:26 +00:00
thumb1-cmp.ll	[ARM] Testcase for Thumb1 cmp with constants.	2018-06-19 00:12:13 +00:00
thumb-imm.ll
thumb-ldm.ll
thumb-shrink-wrapping.ll	[ARM] Allow CMPZ transforms even if the input has multiple uses.	2018-06-08 21:16:56 +00:00
trap.ll
triple.ll
tst_teq.ll
unord.ll
vargs.ll