summaryrefslogtreecommitdiffstats
path: root/llvm/lib
Commit message (Collapse)AuthorAgeFilesLines
* Do not check that the bodies of two defs of same linkonce global are the same.Anton Korobeynikov2010-10-081-1/+1
| | | | | | | Such a check does not make any sense in presense of inlining and other compiler-dependent stuff. This should fix bunch of warnings on mingw32. llvm-svn: 116113
* Implement a few more binary encoding bits. Still very early stage proof-of-Jim Grosbach2010-10-082-0/+49
| | | | | | | | | | concept level stuff at this point, but it is generally working for those instructions that know how to map the operands. This patch fills in the register operands for add/sub/or/etc instructions and adds the conditional execution predicate encoding. llvm-svn: 116112
* Classify value numbers into connected components in linear time.Jakob Stoklund Olesen2010-10-081-23/+15
| | | | llvm-svn: 116105
* Call InitSections in llc and clang so that the binaries produced by themRafael Espindola2010-10-081-0/+1
| | | | | | are easier to diff with those produced by llvm-mc. llvm-svn: 116095
* Filter out illegal formulae after updating offsets, not before, so thatDan Gohman2010-10-081-18/+18
| | | | | | | | | | formulae which become illegal as a result of the offset updating don't escape. This is for rdar://8529692. No testcase yet, because the given cases hit use-list ordering differences. llvm-svn: 116093
* Recommit 116056, now with the missing file...Cameron Esfahani2010-10-082-2/+7
| | | | llvm-svn: 116083
* Don't waste time unfolding simple loads. The unfolded copy won't be hoisted.Evan Cheng2010-10-081-0/+4
| | | | llvm-svn: 116081
* Fix operand latency computation in cases where the definition operand isEvan Cheng2010-10-081-0/+11
| | | | | | | | | | implicit. e.g. %D6<def>, %D7<def> = VLD1q16 %R2<kill>, 0, ..., %Q3<imp-def> %Q1<def> = VMULv8i16 %Q1<kill>, %Q3<kill>, ... The real definition indices are 0,1. llvm-svn: 116080
* CrashRecovery: Fix raise() override to actually send the right signal, *cough*.Daniel Dunbar2010-10-081-1/+1
| | | | llvm-svn: 116072
* Reapply 116059, this time without the fatfingered pasto at the top.Jim Grosbach2010-10-082-22/+22
| | | | | | ''const'ify getMachineOpValue() and associated helpers.' llvm-svn: 116067
* Reverting 116059. Bots are unhappy with it.Jim Grosbach2010-10-082-23/+23
| | | | llvm-svn: 116064
* reverting 116056: win64_params.ll may need to be conditionalized?Andrew Trick2010-10-081-5/+1
| | | | llvm-svn: 116063
* Line number 0 indicates there is no source line/file name info available for ↵Devang Patel2010-10-081-1/+7
| | | | | | this construct. llvm-svn: 116061
* 'const'ify getMachineOpValue() and associated helpers.Jim Grosbach2010-10-082-23/+23
| | | | llvm-svn: 116059
* Start separating out code metrics into code size metrics and code ↵Kenneth Uildriks2010-10-081-10/+53
| | | | | | performance metrics. Partial Specialization will apply the former to function specializations, and the latter to all callsites that can use a specialization, in order to decide whether to create a specialization llvm-svn: 116057
* Small patch to restore home register stack space allocation for the Win64 ↵Cameron Esfahani2010-10-081-1/+5
| | | | | | case. Add test case. This code eventually needs to be tighter, since it's always allocating it, even in leaf routines. llvm-svn: 116056
* Change register allocation order for ARM VFP and NEON registers to put theBob Wilson2010-10-081-6/+72
| | | | | | | | | | | | | | | | callee-saved registers at the end of the lists. Also prefer to avoid using the low registers that are in register subclasses required by certain instructions, so that those registers will more likely be available when needed. This change makes a huge improvement in spilling in some cases. Thanks to Jakob for helping me realize the problem. Most of this patch is fixing the testsuite. There are quite a few places where we're checking for specific registers. I changed those to wildcards in places where that doesn't weaken the tests. The spill-q.ll and thumb2-spill-q.ll tests stopped spilling with this change, so I added a bunch of live values to force spills on those tests. llvm-svn: 116055
* fix a subtle bug I introduced in my refactoring, where we stopped preferringChris Lattner2010-10-081-24/+36
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | the i8 versions of instructions in some cases. In test6, we started generating: cmpq $0, -8(%rsp) ## encoding: [0x48,0x81,0x7c,0x24,0xf8,0x00,0x00,0x00,0x00] ## <MCInst #478 CMP64mi32 ## <MCOperand Reg:114> ## <MCOperand Imm:1> ## <MCOperand Reg:0> ## <MCOperand Imm:-8> ## <MCOperand Reg:0> ## <MCOperand Imm:0>> instead of: cmpq $0, -8(%rsp) ## encoding: [0x48,0x83,0x7c,0x24,0xf8,0x00] ## <MCInst #479 CMP64mi8 ## <MCOperand Reg:114> ## <MCOperand Imm:1> ## <MCOperand Reg:0> ## <MCOperand Imm:-8> ## <MCOperand Reg:0> ## <MCOperand Imm:0>> Fix this and add some comments. llvm-svn: 116053
* reapply: Use the new TB_NOT_REVERSABLE flag instead of specialChris Lattner2010-10-083-38/+77
| | | | | | | | | reapply: reimplement the second half of the or/add optimization. We should now with no changes. Turns out that one missing "Defs = [EFLAGS]" can upset things a bit. llvm-svn: 116040
* reapply the patch reverted in r116033:Chris Lattner2010-10-084-65/+126
| | | | | | | | "Reimplement (part of) the or -> add optimization. Matching 'or' into 'add'" With a critical fix: the add pseudos clobber EFLAGS. llvm-svn: 116039
* Update CMake.Daniel Dunbar2010-10-081-1/+0
| | | | llvm-svn: 116034
* Revert "Reimplement (part of) the or -> add optimization. Matching 'or' intoDaniel Dunbar2010-10-084-126/+65
| | | | | | 'add'", which seems to have broken just about everything. llvm-svn: 116033
* Revert "Use the new TB_NOT_REVERSABLE flag instead of special ", which dependsDaniel Dunbar2010-10-081-26/+20
| | | | | | on r116007, which I am about to revert. llvm-svn: 116032
* Revert "reimplement the second half of the or/add optimization. We should now",Daniel Dunbar2010-10-083-51/+18
| | | | | | which depends on r116007, which I am about to revert. llvm-svn: 116031
* Move to thumb2 loads, fixes a problem with incoming registersEric Christopher2010-10-081-16/+13
| | | | | | | | as thumb1. Fixes lencod. llvm-svn: 116027
* reimplement the second half of the or/add optimization. We should nowChris Lattner2010-10-083-18/+51
| | | | | | | | | | only end up emitting LEA instead of OR. If we aren't able to promote something into an LEA, we should never be emitting it as an ADD. Add some testcases that we emit "or" in cases where we used to produce an "add". llvm-svn: 116026
* Enable binary encoding of some simple instructions.Jim Grosbach2010-10-081-0/+8
| | | | llvm-svn: 116022
* Make <target>CodeEmitter::getBinaryCodeForInstr() a const method.Jim Grosbach2010-10-084-8/+8
| | | | llvm-svn: 116018
* Use the new TB_NOT_REVERSABLE flag instead of special Chris Lattner2010-10-081-20/+26
| | | | | | casing FsMOVAPDrr/FsMOVAPSrr. llvm-svn: 116016
* simplify some map operations.Chris Lattner2010-10-072-16/+14
| | | | llvm-svn: 116014
* MC-COFF: Handle relaxation in COFF better. Fixes PR8321.Michael J. Spencer2010-10-071-1/+29
| | | | llvm-svn: 116013
* Delete the FormulaSorter class and inline its one method into itsDan Gohman2010-10-071-28/+10
| | | | | | | one user. This code will be restructured soon and FormulaSorter is getting in the way. llvm-svn: 116012
* Fix a spello.Dan Gohman2010-10-071-1/+1
| | | | llvm-svn: 116011
* Charge a formula for explicit multiplies on scaled registers too,Dan Gohman2010-10-071-3/+3
| | | | | | not just base registers. llvm-svn: 116010
* Use size_t for consistency.Dan Gohman2010-10-071-1/+1
| | | | llvm-svn: 116009
* When merging one use into another, transfer the offsets fromDan Gohman2010-10-071-0/+8
| | | | | | the old use to the new one. llvm-svn: 116008
* Reimplement (part of) the or -> add optimization. Matching 'or' into 'add'Chris Lattner2010-10-074-65/+126
| | | | | | | | | | | | | | | | | | | is general goodness because it allows ORs to be converted to LEA to avoid inserting copies. However, this is bad because it makes the generated .s file less obvious and gives valgrind heartburn (tons of false positives in bitfield code). While the general fix should be in valgrind, we can at least try to avoid emitting ADD instructions that *don't* get promoted to LEA. This is more work because it requires introducing pseudo instructions to represents "add that knows the bits are disjoint", but hey, people really love valgrind. This fixes this testcase: https://bugs.kde.org/show_bug.cgi?id=242137#c20 the add r/i cases are coming next. llvm-svn: 116007
* After splitting, the remaining LiveInterval may be fragmented into multipleJakob Stoklund Olesen2010-10-072-11/+134
| | | | | | | | | | | | | connected components. These components should be allocated different virtual registers because there is no reason for them to be allocated together. Add the ConnectedVNInfoEqClasses class to calculate the connected components, and move values to new LiveIntervals. Use it from SplitKit::rewrite by creating new virtual registers for the components. llvm-svn: 116006
* Fix LSR to keep the RegUseTracker up to date when combining users.Dan Gohman2010-10-071-8/+19
| | | | | | | | This doesn't usually matter, because the other heuristics usually succeed regardless, but it's good to keep the register use bookkeeping consistent. llvm-svn: 116005
* Remove LoopIndexSplit pass. It is neither maintained nor used by anyone.Devang Patel2010-10-072-1275/+0
| | | | llvm-svn: 116004
* Unbreak cmake build.John Thompson2010-10-071-0/+1
| | | | llvm-svn: 116003
* Code refactoring.Evan Cheng2010-10-072-104/+160
| | | | llvm-svn: 116002
* Reduce casting in various tables by defining the tableChris Lattner2010-10-072-27/+26
| | | | | | with the right types. llvm-svn: 116001
* simplify code: don't build up vector only to assert it is empty.Chris Lattner2010-10-071-8/+4
| | | | llvm-svn: 115997
* Now with fewer extraneous semicolons!Owen Anderson2010-10-07131-160/+161
| | | | llvm-svn: 115996
* Trivial MC code emitter shell. No instruction forms actually handled yet.Jim Grosbach2010-10-071-3/+19
| | | | llvm-svn: 115993
* Correctly check if a path is a directory. Fix by Brian Korver.Evan Cheng2010-10-071-1/+1
| | | | llvm-svn: 115991
* Provie a clearner interface so that FE can decide whether a function has ↵Devang Patel2010-10-071-6/+1
| | | | | | prototype or not. llvm-svn: 115988
* Include the auto-generated bits for machine encoding.Jim Grosbach2010-10-071-0/+20
| | | | llvm-svn: 115987
* Remember to promote load/store types for stack to register size.Eric Christopher2010-10-071-0/+6
| | | | llvm-svn: 115984
OpenPOWER on IntegriCloud