summaryrefslogtreecommitdiffstats
path: root/llvm
Commit message (Collapse)AuthorAgeFilesLines
...
* Skip liveness verification when MRI->tracksLiveness() is false.Jakob Stoklund Olesen2012-03-281-105/+112
| | | | | | | | | Extract the liveness verification into its own method. This makes it possible to run the machine code verifier after liveness information is no longer required to be valid. llvm-svn: 153596
* Reformat the LTOModule code to be more inline with LLVM's coding standards. AddBill Wendling2012-03-282-105/+170
| | | | | | a bunch of comments for the various functions. No intended functionality change. llvm-svn: 153595
* Revert r153516: "Invalidate liveness in Thumb2ITBlockPass."Jakob Stoklund Olesen2012-03-282-8/+0
| | | | | | | | | | | | | | Revert r153519: "ARMLoadStoreOptimizer invalidates register liveness." These patches caused miscompilations in povray by turning off branch folding's updating of live-in lists. It turns out the the late scheduler depends on the live-in lists, even if it doesn't need correct kill flags. <rdar://problem/11139228> llvm-svn: 153593
* Allow removeLiveIn to be called with a register that isn't live-in.Jakob Stoklund Olesen2012-03-281-2/+2
| | | | | | | | | This avoids the silly double search: if (isLiveIn(Reg)) removeLiveIn(Reg); llvm-svn: 153592
* Revert r153521 as it's causing large regressions on the nightly testers.Chad Rosier2012-03-283-56/+0
| | | | | | | | Original commit message for r153521 (aka r153423): Use the new range metadata in computeMaskedBits and add a new optimization to instruction simplify that lets us remove an and when loding a boolean value. llvm-svn: 153587
* Fixed commuteInstructions bug where if its called pre-regalloc the subreg ↵Pete Cooper2012-03-281-6/+15
| | | | | | indices weren't commuted llvm-svn: 153579
* GlobalOpt: If we have an inbounds GEP from a ConstantAggregateZero global ↵Benjamin Kramer2012-03-282-0/+17
| | | | | | that we just determined to be constant, replace all loads from it with a zero value. llvm-svn: 153576
* Add another note about a missed compare with nsw arithmetic instcombine.Benjamin Kramer2012-03-281-0/+7
| | | | llvm-svn: 153574
* Fixup VST1.32 with writeback instruction. Also re-factor non-writeback version.Richard Barton2012-03-282-21/+21
| | | | llvm-svn: 153573
* Switch to WeakVHs in the value mapper, and aggressively prune dead basicChandler Carruth2012-03-283-14/+24
| | | | | | | | | blocks in the function cloner. This removes the last case of trivially dead code that I've been seeing in the wild getting inlined, analyzed, re-inlined, optimized, only to be deleted. Nukes a FIXME from the cleanup tests. llvm-svn: 153572
* More debug output.Eric Christopher2012-03-281-1/+2
| | | | llvm-svn: 153571
* Fix the output of the DW_TAG_friend tag to include DW_AT_friendEric Christopher2012-03-283-8/+63
| | | | | | | | and not the rest of the member tag. Fixes PR11695 llvm-svn: 153570
* Some whitespace cleanup.Bill Wendling2012-03-281-19/+14
| | | | llvm-svn: 153567
* Use the correct filename for the error message.Bill Wendling2012-03-281-1/+1
| | | | llvm-svn: 153564
* Use Nakamura's suggestion of bypassing using 'filename' and just the ↵Bill Wendling2012-03-281-5/+2
| | | | | | pointers directly. llvm-svn: 153558
* Turn off post-RA scheduler by default.Akira Hatanaka2012-03-281-1/+1
| | | | llvm-svn: 153557
* Fix 80-column violation.Chad Rosier2012-03-281-2/+2
| | | | llvm-svn: 153556
* Fix test case.Akira Hatanaka2012-03-281-0/+2
| | | | llvm-svn: 153555
* Turn on post register allocation scheduler.Akira Hatanaka2012-03-284-0/+22
| | | | llvm-svn: 153554
* Sort relocation entries before they are written out to a file. MIPS ABIAkira Hatanaka2012-03-281-0/+103
| | | | | | | imposes a constraint that GOT16 referring to a local symbol or HI16 has to be followed immediately by a matching LO16 relocation. llvm-svn: 153553
* Emit all directives except for ".cprestore" during asm printing rather than emitAkira Hatanaka2012-03-287-151/+188
| | | | | | | | | | | | | | | | them as machine instructions. Directives ".set noat" and ".set at" are now emitted only at the beginning and end of a function except in the case where they are emitted to enclose .cpload with an immediate operand that doesn't fit in 16-bit field or unaligned load/stores. Also, make the following changes: - Remove function isUnalignedLoadStore and use a switch-case statement to determine whether an instruction is an unaligned load or store. - Define helper function CreateMCInst which generates an instance of an MCInst from an opcode and a list of operands. llvm-svn: 153552
* Mark flag neverHasSideEffects of pattern-less instructions that do not haveAkira Hatanaka2012-03-281-0/+5
| | | | | | any side effects. llvm-svn: 153551
* MSVC doesn't like the mixing of declarations and statements in a .c file.Francois Pichet2012-03-271-1/+2
| | | | llvm-svn: 153549
* Add a note about a cute little fabs optimization.Benjamin Kramer2012-03-271-0/+5
| | | | llvm-svn: 153543
* Add two missed instcombines related to compares with nsw arithmetic.Benjamin Kramer2012-03-271-0/+12
| | | | llvm-svn: 153542
* Try to use the CWD if the path to the GCDA output is not available (e.g., theBill Wendling2012-03-271-0/+14
| | | | | | | | executable has been moved to another machine). If that's not available (read-only or something), then exit gracefully. <rdar://problem/11111686> llvm-svn: 153538
* Remove trailing white space.Akira Hatanaka2012-03-271-1/+1
| | | | llvm-svn: 153536
* Use a SmallVector and linear lookup instead of a DenseSet - SourceMap valuesLang Hames2012-03-271-11/+16
| | | | | | | will always be tiny sets, so DenseSet is overkill (SmallSet won't work as we need iteration support). llvm-svn: 153529
* Add member EmitNOAT and its setter and getter functions to class ↵Akira Hatanaka2012-03-271-1/+6
| | | | | | | | | MipsFunctionInfo. If EmitNOAT is true, directives ".set noat" and ".set at" are emitted at the beginning and end of a function. llvm-svn: 153528
* Add a test for the previous commit. Also, remove two tests that wereEric Christopher2012-03-273-117/+31
| | | | | | | testing a) the wrong behavior or b) something that I'm already testing in the new test. llvm-svn: 153525
* Use DW_AT_low_pc for a single entry point into a routine.Eric Christopher2012-03-271-3/+3
| | | | | | Fixes PR10105 llvm-svn: 153524
* Reapply r153423; the original commit was fine. The failing test, distray, had Chad Rosier2012-03-273-0/+56
| | | | | | | | | | undefined behavior, which Rafael was kind enough to fix. Original commit message for r153423: Use the new range metadata in computeMaskedBits and add a new optimization to instruction simplify that lets us remove an and when loding a boolean value. llvm-svn: 153521
* ARMLoadStoreOptimizer invalidates register liveness.Jakob Stoklund Olesen2012-03-271-0/+4
| | | | | | | | | | This pass tries to update kill flags, but there are still many bugs. Passes after the load/store optimizer don't need accurate liveness, so don't even try. <rdar://problem/11101911> llvm-svn: 153519
* Print SSA and liveness tracking flags in MF::print().Jakob Stoklund Olesen2012-03-271-1/+7
| | | | llvm-svn: 153518
* Branch folding may invalidate liveness.Jakob Stoklund Olesen2012-03-271-2/+9
| | | | | | | | Branch folding can use a register scavenger to update liveness information when required. Don't do that if liveness information is already invalid. llvm-svn: 153517
* Invalidate liveness in Thumb2ITBlockPass.Jakob Stoklund Olesen2012-03-271-0/+4
| | | | llvm-svn: 153516
* fix what looks like a real logic bug, found by PVS-Studio (part of PR12357)Chris Lattner2012-03-271-2/+2
| | | | llvm-svn: 153513
* Add an MRI::tracksLiveness() flag.Jakob Stoklund Olesen2012-03-273-1/+28
| | | | | | | | | | | | | | | | | | | | Late optimization passes like branch folding and tail duplication can transform the machine code in a way that makes it expensive to keep the register liveness information up to date. There is a fuzzy line between register allocation and late scheduling where the liveness information degrades. The MRI::tracksLiveness() flag makes the line clear: While true, liveness information is accurate, and can be used for register scavenging. Once the flag is false, liveness information is not accurate, and can only be used as a hint. Late passes generally don't need the liveness information, but they will sometimes use the register scavenger to help update it. The scavenger enforces strict correctness, and we have to spend a lot of code to update register liveness that may never be used. llvm-svn: 153511
* llvm/docs/*.html: Fix markups.NAKAMURA Takumi2012-03-2713-22/+52
| | | | llvm-svn: 153508
* Make a seemingly tiny change to the inliner and fix the generated codeChandler Carruth2012-03-271-1/+1
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | size bloat. Unfortunately, I expect this to disable the majority of the benefit from r152737. I'm hopeful at least that it will fix PR12345. To explain this requires... quite a bit of backstory I'm afraid. TL;DR: The change in r152737 actually did The Wrong Thing for linkonce-odr functions. This change makes it do the right thing. The benefits we saw were simple luck, not any actual strategy. Benchmark numbers after a mini-blog-post so that I've written down my thoughts on why all of this works and doesn't work... To understand what's going on here, you have to understand how the "bottom-up" inliner actually works. There are two fundamental modes to the inliner: 1) Standard fixed-cost bottom-up inlining. This is the mode we usually think about. It walks from the bottom of the CFG up to the top, looking at callsites, taking information about the callsite and the called function and computing th expected cost of inlining into that callsite. If the cost is under a fixed threshold, it inlines. It's a touch more complicated than that due to all the bonuses, weights, etc. Inlining the last callsite to an internal function gets higher weighth, etc. But essentially, this is the mode of operation. 2) Deferred bottom-up inlining (a term I just made up). This is the interesting mode for this patch an r152737. Initially, this works just like mode #1, but once we have the cost of inlining into the callsite, we don't just compare it with a fixed threshold. First, we check something else. Let's give some names to the entities at this point, or we'll end up hopelessly confused. We're considering inlining a function 'A' into its callsite within a function 'B'. We want to check whether 'B' has any callers, and whether it might be inlined into those callers. If so, we also check whether inlining 'A' into 'B' would block any of the opportunities for inlining 'B' into its callers. We take the sum of the costs of inlining 'B' into its callers where that inlining would be blocked by inlining 'A' into 'B', and if that cost is less than the cost of inlining 'A' into 'B', then we skip inlining 'A' into 'B'. Now, in order for #2 to make sense, we have to have some confidence that we will actually have the opportunity to inline 'B' into its callers when cheaper, *and* that we'll be able to revisit the decision and inline 'A' into 'B' if that ever becomes the correct tradeoff. This often isn't true for external functions -- we can see very few of their callers, and we won't be able to re-consider inlining 'A' into 'B' if 'B' is external when we finally see more callers of 'B'. There are two cases where we believe this to be true for C/C++ code: functions local to a translation unit, and functions with an inline definition in every translation unit which uses them. These are represented as internal linkage and linkonce-odr (resp.) in LLVM. I enabled this logic for linkonce-odr in r152737. Unfortunately, when I did that, I also introduced a subtle bug. There was an implicit assumption that the last caller of the function within the TU was the last caller of the function in the program. We want to bonus the last caller of the function in the program by a huge amount for inlining because inlining that callsite has very little cost. Unfortunately, the last caller in the TU of a linkonce-odr function is *not* the last caller in the program, and so we don't want to apply this bonus. If we do, we can apply it to one callsite *per-TU*. Because of the way deferred inlining works, when it sees this bonus applied to one callsite in the TU for 'B', it decides that inlining 'B' is of the *utmost* importance just so we can get that final bonus. It then proceeds to essentially force deferred inlining regardless of the actual cost tradeoff. The result? PR12345: code bloat, code bloat, code bloat. Another result is getting *damn* lucky on a few benchmarks, and the over-inlining exposing critically important optimizations. I would very much like a list of benchmarks that regress after this change goes in, with bitcode before and after. This will help me greatly understand what opportunities the current cost analysis is missing. Initial benchmark numbers look very good. WebKit files that exhibited the worst of PR12345 went from growing to shrinking compared to Clang with r152737 reverted. - Bootstrapped Clang is 3% smaller with this change. - Bootstrapped Clang -O0 over a single-source-file of lib/Lex is 4% faster with this change. Please let me know about any other performance impact you see. Thanks to Nico for reporting and urging me to actually fix, Richard Smith, Duncan Sands, Manuel Klimek, and Benjamin Kramer for talking through the issues today. llvm-svn: 153506
* Prune some includesCraig Topper2012-03-2730-33/+6
| | | | llvm-svn: 153502
* Remove unnecessary llvm:: qualificationsCraig Topper2012-03-2718-261/+261
| | | | llvm-svn: 153500
* Pass the llvm IR pointer value and offset to the constructor ofAkira Hatanaka2012-03-271-9/+13
| | | | | | | | | | | | MachinePointerInfo when getStore is called to create a node that stores an argument passed in register to the stack. Without this change, the post RA scheduler will fail to discover the dependencies between the stores instructions and the instructions that load from a structure passed by value. The link to the related discussion is here: http://lists.cs.uiuc.edu/pipermail/llvmdev/2012-March/048055.html llvm-svn: 153499
* Fix bug in LowerConstantPool. Akira Hatanaka2012-03-271-1/+1
| | | | llvm-svn: 153498
* Add T9 to the list of live-in registers of the entry basic block. Akira Hatanaka2012-03-271-0/+2
| | | | llvm-svn: 153497
* Retrieve and add the offset of a symbol in applyFixup rather than retrieve andAkira Hatanaka2012-03-272-67/+67
| | | | | | | set it in MipsMCCodeEmitter::getMachineOpValue. Assert in getMachineOpValue if MachineOperand MO is of an unexpected type. llvm-svn: 153494
* Define function MipsGetSymAndOffset which returns a fixup's symbol and theAkira Hatanaka2012-03-271-0/+30
| | | | | | offset applied to it. llvm-svn: 153493
* Post-ra LICM should take care not to hoist an instruction that would clobber aEvan Cheng2012-03-272-4/+87
| | | | | | | | register that's read by the preheader terminator. rdar://11095580 llvm-svn: 153492
* Rewrite computation of Value in adjustFixupValue so that the upper 48-bits areAkira Hatanaka2012-03-271-1/+1
| | | | | | cleared. No functionality change. llvm-svn: 153491
* During MachineCopyPropagation a register may be the source operand of multipleLang Hames2012-03-271-17/+26
| | | | | | | | | | copies being considered for removal. Make sure to track all of the copies, rather than just the most recent encountered, by holding a DenseSet instead of an unsigned in SrcMap. No test case - couldn't reduce something with a sane size. llvm-svn: 153487
OpenPOWER on IntegriCloud