)]}' { "log": [ { "commit": "fe57faa2e0349418dda38e77ef1c0ac29db75f4d", "tree": "38ba7a406f8a86a1152bd6c9f2d0a6c677423211", "parents": [ "9e30c0e177adabaaf94a66c91130a19a7632fc7c" ], "author": { "name": "Mark Mendell", "email": "mark.p.mendell@intel.com", "time": "Fri Sep 18 09:26:15 2015 -0400" }, "committer": { "name": "Mark Mendell", "email": "mark.p.mendell@intel.com", "time": "Mon Sep 21 07:23:45 2015 -0400" }, "message": "[optimizing] Add basic PackedSwitch support\n\nAdd HPackedSwitch, and generate it from the builder. Code generators\nconvert this to a series of compare/branch tests. Better implementation\nin the code generators as a real jump table will follow as separate CLs.\n\nChange-Id: If14736fa4d62809b6ae95280148c55682e856911\nSigned-off-by: Mark Mendell \u003cmark.p.mendell@intel.com\u003e\n" }, { "commit": "b7d8e8cf7063fdec1cce6ebd33e33804976bd978", "tree": "8d60856999139cf5b0b0f145f69c35a84a60716c", "parents": [ "a201d5eeb0903408df925a1ed1686a55238a274c" ], "author": { "name": "Vladimir Marko", "email": "vmarko@google.com", "time": "Thu Sep 17 15:47:05 2015 +0100" }, "committer": { "name": "Vladimir Marko", "email": "vmarko@google.com", "time": "Thu Sep 17 16:06:59 2015 +0100" }, "message": "Optimizing: Do not use range-based loop when inserting elements.\n\nWhen we iterate over the elements of a container and we may\ninsert new elements into that container, it\u0027s wrong to use\nthe range-based loop.\n\nBug: 24133462\nChange-Id: Iee35fbcf88ed3bcd6155cbeba09bd256032a16be\n" }, { "commit": "71bf8090663d02869cafafdd530976f7f2a9db7f", "tree": "70f910498031be392dd1f851cf38fe68bdc3154c", "parents": [ "4919c34ead9a10d6cbdaba9967952c344b32e5b9" ], "author": { "name": "Vladimir Marko", "email": "vmarko@google.com", "time": "Tue Sep 15 15:33:14 2015 +0100" }, "committer": { "name": "Vladimir Marko", "email": "vmarko@google.com", "time": "Wed Sep 16 16:41:01 2015 +0100" }, "message": "Optimizing: Tag arena allocations in SsaBuilder.\n\nReplace GrowableArray with ArenaVector in SsaBuilder and\ntag allocations with a new arena allocation type.\n\nChange-Id: I27312c51d7be9d2ad02a974cce93b365c65c5fc4\n" }, { "commit": "fa6b93c4b69e6d7ddfa2a4ed0aff01b0608c5a3a", "tree": "3528c88e104dac8e58ae5370ab066b8b1dd0218f", "parents": [ "e295be4a95d7861f6ec179edf6565f58cad747cc" ], "author": { "name": "Vladimir Marko", "email": "vmarko@google.com", "time": "Tue Sep 15 10:15:55 2015 +0100" }, "committer": { "name": "Vladimir Marko", "email": "vmarko@google.com", "time": "Wed Sep 16 13:21:33 2015 +0100" }, "message": "Optimizing: Tag arena allocations in HGraph.\n\nReplace GrowableArray with ArenaVector in HGraph and related\nclasses HEnvironment, HLoopInformation, HInvoke and HPhi,\nand tag allocations with new arena allocation types.\n\nChange-Id: I3d79897af405b9a1a5b98bfc372e70fe0b3bc40d\n" }, { "commit": "b505997b2176bd29a108cb6c33d06d4ef29ba001", "tree": "05316a1837f77729f87b4fe25bfa7eb2b9e9549c", "parents": [ "5c1a3374adff0d5c6c8e2248fffd7549eb3a6bae", "77a48ae01bbc5b05ca009cf09e2fcb53e4c8ff23" ], "author": { "name": "David Brazdil", "email": "dbrazdil@google.com", "time": "Tue Sep 15 12:34:35 2015 +0000" }, "committer": { "name": "Gerrit Code Review", "email": "noreply-gerritcodereview@google.com", "time": "Tue Sep 15 12:34:35 2015 +0000" }, "message": "Merge \"Revert \"Revert \"ART: Register allocation and runtime support for try/catch\"\"\"" }, { "commit": "77a48ae01bbc5b05ca009cf09e2fcb53e4c8ff23", "tree": "780c7d6bdee784c2f8248979de348491cfb63b34", "parents": [ "659562aaf133c41b8d90ec9216c07646f0f14362" ], "author": { "name": "David Brazdil", "email": "dbrazdil@google.com", "time": "Tue Sep 15 12:34:04 2015 +0000" }, "committer": { "name": "David Brazdil", "email": "dbrazdil@google.com", "time": "Tue Sep 15 12:34:04 2015 +0000" }, "message": "Revert \"Revert \"ART: Register allocation and runtime support for try/catch\"\"\n\nThe original CL triggered b/24084144 which has been fixed\nby Ib72e12a018437c404e82f7ad414554c66a4c6f8c.\n\nThis reverts commit 659562aaf133c41b8d90ec9216c07646f0f14362.\n\nChange-Id: Id8980436172457d0fcb276349c4405f7c4110a55\n" }, { "commit": "baf89b8f2591df93686df545bc6ccc6eb3b128cc", "tree": "d66715e4cb93257b4903b997e1e09f37dd16d2cd", "parents": [ "cd9266e68246a42ae3eb65a0bd5e2c4ec62418a0" ], "author": { "name": "David Brazdil", "email": "dbrazdil@google.com", "time": "Tue Sep 15 11:36:54 2015 +0100" }, "committer": { "name": "David Brazdil", "email": "dbrazdil@google.com", "time": "Tue Sep 15 11:51:34 2015 +0100" }, "message": "ART: Fix bug in reference type propagation\n\nReference type propagation assumed that type is exact if the class is\nfinal. This does not hold for arrays which are always final and the\ncomponent type needs to be considered.\n\nBug: 24084144\nChange-Id: Ib72e12a018437c404e82f7ad414554c66a4c6f8c\n" }, { "commit": "659562aaf133c41b8d90ec9216c07646f0f14362", "tree": "be1beae390262bf2f5a17bfa44de93081a849d07", "parents": [ "b022fa1300e6d78639b3b910af0cf85c43df44bb" ], "author": { "name": "David Brazdil", "email": "dbrazdil@google.com", "time": "Mon Sep 14 21:26:33 2015 +0000" }, "committer": { "name": "David Brazdil", "email": "dbrazdil@google.com", "time": "Mon Sep 14 21:26:33 2015 +0000" }, "message": "Revert \"ART: Register allocation and runtime support for try/catch\"\n\nBreaks libcore test org.apache.harmony.security.tests.java.security.KeyStorePrivateKeyEntryTest#testGetCertificateChain. Need to investigate.\n\nThis reverts commit b022fa1300e6d78639b3b910af0cf85c43df44bb.\n\nChange-Id: Ib24d3a80064d963d273e557a93469c95f37b1f6f\n" }, { "commit": "b022fa1300e6d78639b3b910af0cf85c43df44bb", "tree": "780c7d6bdee784c2f8248979de348491cfb63b34", "parents": [ "e481c006e8b055a31d9c7cff27f4145e57e3c113" ], "author": { "name": "David Brazdil", "email": "dbrazdil@google.com", "time": "Thu Aug 20 17:47:48 2015 +0100" }, "committer": { "name": "David Brazdil", "email": "dbrazdil@google.com", "time": "Mon Sep 14 20:42:58 2015 +0100" }, "message": "ART: Register allocation and runtime support for try/catch\n\nThis patch completes a series of CLs that add support for try/catch\nin the Optimizing compiler. With it, Optimizing can compile all\nmethods containing try/catch, provided they don\u0027t contain catch loops.\nFuture work will focus on improving performance of the generated code.\n\nSsaLivenessAnalysis was updated to propagate liveness information of\ninstructions live at catch blocks, and to keep location information on\ninstructions which may be caught by catch phis.\n\nRegisterAllocator was extended to spill values used after catch, and\nto allocate spill slots for catch phis. Catch phis generated for the\nsame vreg share a spill slot as the raw value must be the same.\n\nLocation builders and slow paths were updated to reflect the fact that\nthrowing an exception may not lead to escaping the method.\n\nInstruction code generators are forbidden from using of implicit null\nchecks in try blocks as live registers need to be saved before handing\nover to the runtime.\n\nCodeGenerator emits a stack map for each catch block, storing locations\nof catch phis. CodeInfo and StackMapStream recognize this new type of\nstack map and store them separate from other stack maps to avoid dex_pc\nconflicts.\n\nAfter having found the target catch block to deliver an exception to,\nQuickExceptionHandler looks up the dex register maps at the throwing\ninstruction and the catch block and copies the values over to their\nrespective locations.\n\nThe runtime-support approach was selected because it allows for the\nbest performance in the normal control-flow path, since no propagation\nof catch phi values is necessary until the exception is thrown. In\naddition, it also greatly simplifies the register allocation phase.\n\nConstantHoisting was removed from LICMTest because it instantiated\n(now abstract) HConstant and was bogus anyway (constants are always in\nthe entry block).\n\nChange-Id: Ie31038ad8e3ee0c13a5bbbbaf5f0b3e532310e4e\n" }, { "commit": "9af47e27674e70e7f82315d39de5bc659c4e1b83", "tree": "e1cf3d95c94721cf9e411e544f37105b684d7666", "parents": [ "35717f8cfb6c1de5c89b2f9b990a050b2cbab66f", "6058455d486219994921b63a2d774dc9908415a2" ], "author": { "name": "Vladimir Marko", "email": "vmarko@google.com", "time": "Mon Sep 14 11:43:25 2015 +0000" }, "committer": { "name": "Gerrit Code Review", "email": "noreply-gerritcodereview@google.com", "time": "Mon Sep 14 11:43:25 2015 +0000" }, "message": "Merge \"Optimizing: Tag basic block allocations with their source.\"" }, { "commit": "3ecfd65143d95bd7c6cbe4f58c33af517d3761e0", "tree": "62b3194f0ac2814541170f970996b0a35511e78e", "parents": [ "426514cada274714244290d6447d9f62dc3394bf" ], "author": { "name": "Yevgeny Rouban", "email": "yevgeny.y.rouban@intel.com", "time": "Mon Sep 07 17:57:00 2015 +0600" }, "committer": { "name": "Nicolas Geoffray", "email": "ngeoffray@google.com", "time": "Wed Sep 09 14:12:13 2015 +0100" }, "message": "Add dex_pc to all HInstructions in builder.\n\nOptimizing compiler generates minimum debug line info that\nis built using the dex_pc information about suspend points.\nThis is not enough for performance and debugging needs.\n\nThis patch makes all HInstructions contain\ndex_pc and all allocations in the builder define this value.\n\nChange-Id: I1d14aefe075189b7b1b41b4384c3499474c19afc\nSigned-off-by: Yevgeny Rouban \u003cyevgeny.y.rouban@intel.com\u003e\nSigned-off-by: Serdjuk, Nikolay Y \u003cnikolay.y.serdjuk@intel.com\u003e\n" }, { "commit": "6058455d486219994921b63a2d774dc9908415a2", "tree": "3d205227f3ff54cd3a50bc5c0e7cb3ad6c175b86", "parents": [ "637ee0b9c10ab7732a7ee7b8335f3fff4ac1549c" ], "author": { "name": "Vladimir Marko", "email": "vmarko@google.com", "time": "Thu Sep 03 13:35:12 2015 +0000" }, "committer": { "name": "Vladimir Marko", "email": "vmarko@google.com", "time": "Tue Sep 08 18:09:49 2015 +0100" }, "message": "Optimizing: Tag basic block allocations with their source.\n\nReplace GrowableArray with ArenaVector in HBasicBlock and,\nto track the source of allocations, assign one new and two\nQuick\u0027s arena allocation types to these vectors. Rename\nkArenaAllocSuccessor to kArenaAllocSuccessors.\n\nBug: 23736311\nChange-Id: Ib52e51698890675bde61f007fe6039338cf1a025\n" }, { "commit": "145acc5361deb769eed998f057bc23abaef6e116", "tree": "3d66a0b44e1ac927156eec6e6488de5fd52b982b", "parents": [ "91e11c0c840193c6822e66846020b6647de243d5" ], "author": { "name": "Vladimir Marko", "email": "vmarko@google.com", "time": "Thu Sep 03 13:33:25 2015 +0000" }, "committer": { "name": "Vladimir Marko", "email": "vmarko@google.com", "time": "Thu Sep 03 13:33:25 2015 +0000" }, "message": "Revert \"Optimizing: Tag basic block allocations with their source.\"\n\nReverting so that we can have more discussion about the STL API.\n\nThis reverts commit 91e11c0c840193c6822e66846020b6647de243d5.\n\nChange-Id: I187fe52f2c16b6e7c5c9d49c42921eb6c7063dba\n" }, { "commit": "91e11c0c840193c6822e66846020b6647de243d5", "tree": "0c5398ef59c464c1848afd0113c74b6aeb75cf42", "parents": [ "f9f6441c665b5ff9004d3ed55014f46d416fb1bb" ], "author": { "name": "Vladimir Marko", "email": "vmarko@google.com", "time": "Wed Sep 02 17:03:22 2015 +0100" }, "committer": { "name": "Vladimir Marko", "email": "vmarko@google.com", "time": "Thu Sep 03 13:09:37 2015 +0100" }, "message": "Optimizing: Tag basic block allocations with their source.\n\nReplace GrowableArray with ArenaVector in HBasicBlock and,\nto track the source of allocations, assign one new and two\nQuick\u0027s arena allocation types to these vectors. Rename\nkArenaAllocSuccessor to kArenaAllocSuccessors.\n\nBug: 23736311\nChange-Id: I984aef6e615ae2380a532f5c6726af21015f43f5\n" }, { "commit": "bbd733e4ef277eff19bf9a6601032da081e9b68f", "tree": "47c5e9624378d8705e8e448ed951617f680d021f", "parents": [ "a35d4c9d66b372e7ae1014357c48ddfa6104085b" ], "author": { "name": "David Brazdil", "email": "dbrazdil@google.com", "time": "Tue Aug 18 17:48:17 2015 +0100" }, "committer": { "name": "David Brazdil", "email": "dbrazdil@google.com", "time": "Wed Aug 19 19:55:43 2015 +0100" }, "message": "ART: Enable basic optimizations for try/catch\n\nGenerating code for try/catch methods requires having run at least the\ninstruction simplifier to remove redundant suspend checks. This patch\nenables the first group of optimizations when try/catch is present.\n\nEnabled optimizations:\n1) IntrinsicsRecognizer\nDoes not modify the graph, only sets HInvoke::intrinsic_.\n\n2) ConstantFolding\nDoes not deal with throwing instructions.\n\n3) InstructionSimplifier\nMay remove a throwing instruction (e.g. LoadClass in VisitCheckCast),\nor may turn a throwing instruction into a non-throwing one (ArraySet).\nTheir corresponding catch phi inputs are not removed but correctness\nis preserved.\n\n4) ReferenceTypePropagation\nDoes not modify the graph, only sets type properties. Typing of\nLoadException from catch handler information was added.\n\n5) DeadCodeElimination\nRemoving individual instructions is fine (same as 3). Removal of dead\nblocks was disabled for try/catch.\n\nChange-Id: I2722c3229eb8aaf326391e07f522dbf5186774b8\n" }, { "commit": "ec16f79a4d0aeff319bf52139a0c82de3080d73c", "tree": "1306b2cbaa97bbacc2cbc100d6e157e14f9510cb", "parents": [ "32f264e67afa8654a5570d38b627515fb73fc333" ], "author": { "name": "David Brazdil", "email": "dbrazdil@google.com", "time": "Wed Aug 19 15:04:01 2015 +0100" }, "committer": { "name": "David Brazdil", "email": "dbrazdil@google.com", "time": "Wed Aug 19 17:43:13 2015 +0100" }, "message": "ART: Refactor try/catch block info, store exception type\n\nThis patch replaces HBasicBlock fields storing try/catch info with a\nsingle TryCatchInformation data structure, saving memory for the\nmajority of non-try/catch blocks. It also changes builder to store\nthe exception type for catch blocks.\n\nChange-Id: Ib3e43f7db247e6915d67c267fc62410420e230c9\n" }, { "commit": "29fc008c9689e9036a3f5e3bd186bbfb5de3cb82", "tree": "373a288e3bfcc44d6af87c78a3e114eed211eca3", "parents": [ "a539ec06503766fcad4be71480c194a225fb037c" ], "author": { "name": "David Brazdil", "email": "dbrazdil@google.com", "time": "Tue Aug 18 17:17:38 2015 +0100" }, "committer": { "name": "David Brazdil", "email": "dbrazdil@google.com", "time": "Tue Aug 18 17:17:38 2015 +0100" }, "message": "ART: Revert storing of exceptional predecessors\n\nAfter change of the approach for try/catch register allocation, it is\nno longer necessary to record instructions which might throw into a\ncatch block.\n\nChange-Id: I7ef12ed06c49a35280029810975fa2a50fe4a424\n" }, { "commit": "1339409345cc8b46d64570430ca3197408fda159", "tree": "4e2a3dfea6fab0f902beb70140d38b6edfecba3e", "parents": [ "902ec264e1040cbfd783c44809ed10b261bd78ee", "9867bc722f7c41e07a95397bc08b790cd21dc758" ], "author": { "name": "Roland Levillain", "email": "rpl@google.com", "time": "Wed Aug 05 12:52:12 2015 +0000" }, "committer": { "name": "Gerrit Code Review", "email": "noreply-gerritcodereview@google.com", "time": "Wed Aug 05 12:52:12 2015 +0000" }, "message": "Merge \"Have constant folding be more flexible.\"" }, { "commit": "902ec264e1040cbfd783c44809ed10b261bd78ee", "tree": "915f7d23caf192e440a585b9495f4d787e8d26bf", "parents": [ "d4fd591406f356dcb8feaa262c19983cacb9116d", "c90bc7c07f9bd24b5424cfb1e3f064fbae5334d6" ], "author": { "name": "Roland Levillain", "email": "rpl@google.com", "time": "Wed Aug 05 10:03:46 2015 +0000" }, "committer": { "name": "Gerrit Code Review", "email": "noreply-gerritcodereview@google.com", "time": "Wed Aug 05 10:03:46 2015 +0000" }, "message": "Merge \"Add constant folding for long unary operations in opt. compiler.\"" }, { "commit": "9867bc722f7c41e07a95397bc08b790cd21dc758", "tree": "6bad203433f24a5556fe98e6480326ed30956c75", "parents": [ "c90bc7c07f9bd24b5424cfb1e3f064fbae5334d6" ], "author": { "name": "Roland Levillain", "email": "rpl@google.com", "time": "Wed Aug 05 10:21:34 2015 +0100" }, "committer": { "name": "Roland Levillain", "email": "rpl@google.com", "time": "Wed Aug 05 11:01:53 2015 +0100" }, "message": "Have constant folding be more flexible.\n\n- Have Evaluate methods take as argument(s) and return value\n instances of HConstant (instead of built-in 32- or 64-bit\n integer values), to let the evaluated instruction choose\n the type of the statically evaluated node; for instance,\n art::HEqual::Evaluate shall return a HIntConstant\n node (as implementation of a Boolean constant) whatever\n the type of its inputs (a pair of HIntConstant or a pair\n of HLongConstant).\n- Split the evaluation job from the operation logic: the\n former is addressed by Evaluate methods, while the latter\n is done by a generic Compute method.\n- Adress valid BinOp(int, long) and BinOp(long, int) cases.\n- Add a constructor to art::HIntConstant to build an integer\n constant from a `bool` value.\n\nChange-Id: If84b6fe8406bb94ddb1aa8b02e36628dff526db3\n" }, { "commit": "c90bc7c07f9bd24b5424cfb1e3f064fbae5334d6", "tree": "a462fa0cc21489d4250febb6d098aacc7a265279", "parents": [ "c2abe2fe9036b57c581e3003d0b820d1c54dbd30" ], "author": { "name": "Roland Levillain", "email": "rpl@google.com", "time": "Thu Dec 11 12:14:33 2014 +0000" }, "committer": { "name": "Roland Levillain", "email": "rpl@google.com", "time": "Wed Aug 05 11:01:40 2015 +0100" }, "message": "Add constant folding for long unary operations in opt. compiler.\n\nAdd tests to exercise the constant folding of these\ninstructions.\n\nAlso, prevent Java methods from run-tests exercising the\ncode generation of these instruction from being inlined,\nso that they continue to check the generated code (and\nnot the code produced by the constant folding pass).\n\nChange-Id: I28efca7cdb5142ac2b6d158ba296fb9136d62481\n" }, { "commit": "b618adebbc19e50d7b1aa2f11b84341beb3c64dc", "tree": "02338a5656bb8a21b742f6506ce92a07d4a8162a", "parents": [ "7b926cdacc2b67241bc9cb5f2d4b04b13ca79d0e" ], "author": { "name": "David Brazdil", "email": "dbrazdil@google.com", "time": "Wed Jul 29 10:31:29 2015 +0100" }, "committer": { "name": "David Brazdil", "email": "dbrazdil@google.com", "time": "Thu Jul 30 17:02:41 2015 +0100" }, "message": "ART: Store and check exceptional predecessors\n\nFuture CL on register allocation for try/catch will require the\nknowledge of instructions which throw into a catch block. This patch\nstores that information with the basic block and verifies it in the\ngraph checker.\n\nMore checks on try catch also added to the graph checker and an order\nof exception handlers is enforced in TryBoundary successors.\n\nChange-Id: I3034c610791ea51d96724bcca97f49ec6ecf2af3\n" }, { "commit": "2e76830f0b3f23825677436c0633714402715099", "tree": "71bd6bd7d327ff3ef8a46af2a7514c5434adb5fd", "parents": [ "a5ae3c3f468ffe3a317b498d7fde1f8e9325346a" ], "author": { "name": "Calin Juravle", "email": "calin@google.com", "time": "Tue Jul 28 14:41:11 2015 +0000" }, "committer": { "name": "Calin Juravle", "email": "calin@google.com", "time": "Wed Jul 29 17:08:28 2015 +0100" }, "message": "Revert \"Revert \"Revert \"Revert \"Use the object class as top in reference type propagation\"\"\"\"\n\nThis reverts commit b734808d0c93af98ec4e3539fdb0a8c0787263b0.\n\nChange-Id: Ifd925f166761bcb9be2268ff0fc9fa3a72f00c6f\n" }, { "commit": "b734808d0c93af98ec4e3539fdb0a8c0787263b0", "tree": "a9a0a67143d5af9c375dbb9db9d874b61586f7e1", "parents": [ "80caa1478cf3df4eac1214d8a63a4da6f4fe622b" ], "author": { "name": "Calin Juravle", "email": "calin@google.com", "time": "Tue Jul 28 11:52:02 2015 +0000" }, "committer": { "name": "Calin Juravle", "email": "calin@google.com", "time": "Tue Jul 28 11:52:02 2015 +0000" }, "message": "Revert \"Revert \"Revert \"Use the object class as top in reference type propagation\"\"\"\n\nThis reverts commit 80caa1478cf3df4eac1214d8a63a4da6f4fe622b.\n\nChange-Id: I63b51ca418b19b2bfb5ede3f8444f8fbeb8a339d\n" }, { "commit": "80caa1478cf3df4eac1214d8a63a4da6f4fe622b", "tree": "131652877e34099711e7d9fc4f1d840b99d1b5d3", "parents": [ "00e3b38be4b280d6d7a7e843cd336ffbd2ba4365" ], "author": { "name": "Calin Juravle", "email": "calin@google.com", "time": "Thu Jul 16 16:51:30 2015 +0100" }, "committer": { "name": "Calin Juravle", "email": "calin@google.com", "time": "Mon Jul 27 19:51:48 2015 +0100" }, "message": "Revert \"Revert \"Use the object class as top in reference type propagation\"\"\n\nThis reverts commit 7733bd644ac71f86d4b30a319624b23343882e53.\n\nChange-Id: I7d393a808c01c084c18d632a54e0554b4b455f2c\n" }, { "commit": "32d7806b34efda2af0f5a0eaaa384c781565d364", "tree": "365d25bc21b5af5ec6dfab828de2d28acbdb8b26", "parents": [ "d77aa00c583cc72f96bc4522d02d2070dcd72f1c", "7733bd644ac71f86d4b30a319624b23343882e53" ], "author": { "name": "Calin Juravle", "email": "calin@google.com", "time": "Wed Jul 22 17:19:20 2015 +0000" }, "committer": { "name": "Gerrit Code Review", "email": "noreply-gerritcodereview@google.com", "time": "Wed Jul 22 17:19:20 2015 +0000" }, "message": "Merge \"Revert \"Use the object class as top in reference type propagation\"\"" }, { "commit": "7733bd644ac71f86d4b30a319624b23343882e53", "tree": "7f897d8b0545e43da6b6059718685a8e396b3b70", "parents": [ "3fabec7a25d151b26ba7de13615bbead0dd615a6" ], "author": { "name": "Calin Juravle", "email": "calin@google.com", "time": "Wed Jul 22 17:14:50 2015 +0000" }, "committer": { "name": "Calin Juravle", "email": "calin@google.com", "time": "Wed Jul 22 17:14:50 2015 +0000" }, "message": "Revert \"Use the object class as top in reference type propagation\"\n\nThis reverts commit 3fabec7a25d151b26ba7de13615bbead0dd615a6.\n\nChange-Id: Id8614f6b6e3e0e4c9caeb9f771e4c145d9fec64f\n" }, { "commit": "e6e38ce021ef5e0d326d76172307c000e0e6fab3", "tree": "c7478d676ab6f0c229f154ab06fbce56f7ae507b", "parents": [ "a3073e24e00274d2957771518be5ecfa74591744", "3fabec7a25d151b26ba7de13615bbead0dd615a6" ], "author": { "name": "Calin Juravle", "email": "calin@google.com", "time": "Wed Jul 22 11:20:14 2015 +0000" }, "committer": { "name": "Gerrit Code Review", "email": "noreply-gerritcodereview@google.com", "time": "Wed Jul 22 11:20:14 2015 +0000" }, "message": "Merge \"Use the object class as top in reference type propagation\"" }, { "commit": "3fabec7a25d151b26ba7de13615bbead0dd615a6", "tree": "a73af9489665d5b033e961d1aafe7e0935c8af85", "parents": [ "b0d5fc0ac139da4aaa1440263416b9bde05630b0" ], "author": { "name": "Calin Juravle", "email": "calin@google.com", "time": "Thu Jul 16 16:51:30 2015 +0100" }, "committer": { "name": "Calin Juravle", "email": "calin@google.com", "time": "Tue Jul 21 14:33:42 2015 +0100" }, "message": "Use the object class as top in reference type propagation\n\nThis properly types all instructions, making it safe to query the type\nat any time.\n\nThis also moves a few functions from class.h to class-inl.h to please\ngcc linker when compiling for target.\n\nChange-Id: I6b7ce965c10834c994b95529ab65a548515b4406\n" }, { "commit": "ffee3d33f3ea39aa6031c3d2ff29c4806c8dcc51", "tree": "9b990552fbf7e0ba91954758b07e2a51d95da9cd", "parents": [ "045da88bd0230ddf48aaf037b3d6c16fbb97ba65" ], "author": { "name": "David Brazdil", "email": "dbrazdil@google.com", "time": "Mon Jul 06 11:48:53 2015 +0100" }, "committer": { "name": "David Brazdil", "email": "dbrazdil@google.com", "time": "Wed Jul 15 17:20:30 2015 +0200" }, "message": "ART: Build SSA form when try/catch is present\n\nThis patch implements support for try/catch in the SsaBuilder.\nValues of locals are propagated from throwing sites inside try\nblocks to their respective catch blocks and phis (\"catch phis\")\nare created when necessary.\n\nChange-Id: I0736565c2c4ff3f9f0924b6e3a785a50023f875a\n" }, { "commit": "761f281e6f78b1c48103cbd45385a5c410e2be07", "tree": "fe6d9c23cdd58c92708c5d6dd51348b6c3f3f1f9", "parents": [ "c7d5ae363bc43ebba0f21c6ee4a2164dfbb209a9", "beba9302bec33d72beb582970bf23d056f62641f" ], "author": { "name": "Calin Juravle", "email": "calin@google.com", "time": "Wed Jul 08 15:57:32 2015 +0000" }, "committer": { "name": "Gerrit Code Review", "email": "noreply-gerritcodereview@google.com", "time": "Wed Jul 08 15:57:33 2015 +0000" }, "message": "Merge \"Revert \"Use the object class as top in reference type propagation\"\"" }, { "commit": "beba9302bec33d72beb582970bf23d056f62641f", "tree": "6bc853b650f428b636f25594a6917d42cba1f86a", "parents": [ "20e6071362b84a9782b633a893c29ebde458205e" ], "author": { "name": "Calin Juravle", "email": "calin@google.com", "time": "Wed Jul 08 15:57:18 2015 +0000" }, "committer": { "name": "Calin Juravle", "email": "calin@google.com", "time": "Wed Jul 08 15:57:18 2015 +0000" }, "message": "Revert \"Use the object class as top in reference type propagation\"\n\nfailing on the build bot on some targets but not locally. needs more investigation.\n\nThis reverts commit 20e6071362b84a9782b633a893c29ebde458205e.\n\nChange-Id: I6965483f569fb862f9bdb66d459b747ded54de71\n" }, { "commit": "c7432f3b8cd7dd6fd198fe5cf7238650921cb306", "tree": "a5efcd21fb4ea77ca94124b3c6e0bd9244575e10", "parents": [ "c21dc873f5a6ce31d73299ee71805031ddb9b65a", "20e6071362b84a9782b633a893c29ebde458205e" ], "author": { "name": "Calin Juravle", "email": "calin@google.com", "time": "Wed Jul 08 15:00:00 2015 +0000" }, "committer": { "name": "Gerrit Code Review", "email": "noreply-gerritcodereview@google.com", "time": "Wed Jul 08 15:00:01 2015 +0000" }, "message": "Merge \"Use the object class as top in reference type propagation\"" }, { "commit": "20e6071362b84a9782b633a893c29ebde458205e", "tree": "8d1089cf494fc2b55d986c3084a80015925cc614", "parents": [ "2b252dc15995a708aead33aab6d307205d860a88" ], "author": { "name": "Calin Juravle", "email": "calin@google.com", "time": "Wed Jul 01 18:41:04 2015 +0100" }, "committer": { "name": "Calin Juravle", "email": "calin@google.com", "time": "Wed Jul 08 13:56:41 2015 +0100" }, "message": "Use the object class as top in reference type propagation\n\nThis properly types all instructions, making it safe to query the type\nat any time.\n\nChange-Id: I3ee2f0f79253cdf45b10ddab37ecb473345ca53a\n" }, { "commit": "c470193cfc522fc818eb2eaab896aef9caf0c75a", "tree": "9887d434f8d9e33c41b98ca406e7c060c68c9016", "parents": [ "c87c8939ea1bcfbddb954478d527cf1138f4f343" ], "author": { "name": "Mark Mendell", "email": "mark.p.mendell@intel.com", "time": "Fri Apr 10 13:18:51 2015 -0400" }, "committer": { "name": "Roland Levillain", "email": "rpl@google.com", "time": "Tue Jul 07 16:49:53 2015 +0100" }, "message": "Fuse long and FP compare \u0026 condition on x86/x86-64 in Optimizing.\n\nThis is a preliminary implementation of fusing long/float/double\ncompares with conditions to avoid materializing the result from the\ncompare and condition.\n\nThe information from a HCompare is transferred to the HCondition if it\nis legal. There must be only a single use of the HCompare, the HCompare\nand HCondition must be in the same block, the HCondition must not need\nmaterialization.\n\nAdded GetOppositeCondition() to HCondition to return the flipped\ncondition.\n\nBug: 21120453\nChange-Id: I1f1db206e6dc336270cd71070ed3232dedc754d6\nSigned-off-by: Mark Mendell \u003cmark.p.mendell@intel.com\u003e\n" }, { "commit": "56e1accf3966ae92e151567abf4561ef3f6466f4", "tree": "7dc23852fe89290f1e01626342dc2a2d5a7de492", "parents": [ "3b0667c1d68ba88c71b031757b757dca659afd69" ], "author": { "name": "David Brazdil", "email": "dbrazdil@google.com", "time": "Tue Jun 30 15:41:36 2015 +0100" }, "committer": { "name": "David Brazdil", "email": "dbrazdil@google.com", "time": "Wed Jul 01 14:51:21 2015 +0100" }, "message": "ART: Changes to try-catch in GraphBuilder\n\nThis patch adds an additional case into the insertion algorithm for\nHTryBoundary inside HGraphBuilder in order to better handle catch\nblocks covered by a TryItem.\n\nBuilding SSA form also required to stop combining HTryBoundaries for\nneighbouring TryItems because it was not clear which exception\nhandlers belong to which try block.\n\nChange-Id: Ic68bd6ef98fee784609fa593cb08dca1f00a15e0\n" }, { "commit": "fc6a86ab2b70781e72b807c1798b83829ca7f931", "tree": "90201491e811cf7be0e0469d7a06a828f4384cad", "parents": [ "d3eaade87ac079accca30473ef0a3b38ab600828" ], "author": { "name": "David Brazdil", "email": "dbrazdil@google.com", "time": "Fri Jun 26 10:33:45 2015 +0000" }, "committer": { "name": "David Brazdil", "email": "dbrazdil@google.com", "time": "Fri Jun 26 13:49:50 2015 +0100" }, "message": "Revert \"Revert \"ART: Implement try/catch blocks in Builder\"\"\n\nThis patch enables the GraphBuilder to generate blocks and edges which\nrepresent the exceptional control flow when try/catch blocks are\npresent in the code. Actual compilation is still delegated to Quick\nand Baseline ignores the additional code.\n\nTo represent the relationship between try and catch blocks, Builder\nsplits the edges which enter/exit a try block and links the newly\ncreated blocks to the corresponding exception handlers. This layout\nwill later enable the SsaBuilder to correctly infer the dominators of\nthe catch blocks and to produce the appropriate reverse post ordering.\nIt will not, however, allow for building the complete SSA form of the\ncatch blocks and consequently optimizing such blocks.\n\nTo this end, a new TryBoundary control-flow instruction is introduced.\nCodegen treats it the same as a Goto but it allows for additional\nsuccessors (the handlers).\n\nThis reverts commit 3e18738bd338e9f8363b26bc895f38c0ec682824.\n\nChange-Id: I4f5ea961848a0b83d8db3673763861633e9bfcfb\n" }, { "commit": "3e18738bd338e9f8363b26bc895f38c0ec682824", "tree": "708013ef06cfb524f040b2b5c494f7f3cb84ac2c", "parents": [ "0b5c7d1994b76090afcc825e737f2b8c546da2f8" ], "author": { "name": "David Brazdil", "email": "dbrazdil@google.com", "time": "Fri Jun 26 09:59:52 2015 +0000" }, "committer": { "name": "David Brazdil", "email": "dbrazdil@google.com", "time": "Fri Jun 26 09:59:52 2015 +0000" }, "message": "Revert \"ART: Implement try/catch blocks in Builder\"\n\nCauses OutOfMemory issues, need to investigate.\n\nThis reverts commit 0b5c7d1994b76090afcc825e737f2b8c546da2f8.\n\nChange-Id: I263e6cc4df5f9a56ad2ce44e18932ca51d7e349f\n" }, { "commit": "0b5c7d1994b76090afcc825e737f2b8c546da2f8", "tree": "057eddf8830b1991f02af3c3ce1b63dee90dd2ad", "parents": [ "1dd3136d9f6b1c7d551897a2d96c8314e40f7324" ], "author": { "name": "David Brazdil", "email": "dbrazdil@google.com", "time": "Thu Jun 11 11:17:49 2015 +0100" }, "committer": { "name": "David Brazdil", "email": "dbrazdil@google.com", "time": "Thu Jun 25 16:58:08 2015 +0100" }, "message": "ART: Implement try/catch blocks in Builder\n\nThis patch enables the GraphBuilder to generate blocks and edges which\nrepresent the exceptional control flow when try/catch blocks are\npresent in the code. Actual compilation is still delegated to Quick\nand Baseline ignores the additional code.\n\nTo represent the relationship between try and catch blocks, Builder\nsplits the edges which enter/exit a try block and links the newly\ncreated blocks to the corresponding exception handlers. This layout\nwill later enable the SsaBuilder to correctly infer the dominators of\nthe catch blocks and to produce the appropriate reverse post ordering.\nIt will not, however, allow for building the complete SSA form of the\ncatch blocks and consequently optimizing such blocks.\n\nTo this end, a new TryBoundary control-flow instruction is introduced.\nCodegen treats it the same as a Goto but it allows for additional\nsuccessors (the handlers).\n\nChange-Id: I415b985596d5bebb7b1bb358a46e08b7b04bb53a\n" }, { "commit": "b809daaffdd42dc01457ae06ae83f2d7ebdb5e65", "tree": "16c3d455297fbf79a1d084392aada209ac2e8704", "parents": [ "1c7f2f10230baf7a549af5bfd16fd3f0fb73ba21", "fe659462e7d58bb2585b1bd029f9e08fd9dd32ae" ], "author": { "name": "David Brazdil", "email": "dbrazdil@google.com", "time": "Wed Jun 24 13:56:48 2015 +0000" }, "committer": { "name": "Gerrit Code Review", "email": "noreply-gerritcodereview@google.com", "time": "Wed Jun 24 13:56:49 2015 +0000" }, "message": "Merge \"ART: Stop creating a fallthrough block for Goto\"" }, { "commit": "fe659462e7d58bb2585b1bd029f9e08fd9dd32ae", "tree": "987b9dd7f8278766d2b6a4a65397f2852f1d7ef4", "parents": [ "582195dfcee32586b71b0ed00a973cfc7d7b8b57" ], "author": { "name": "David Brazdil", "email": "dbrazdil@google.com", "time": "Wed Jun 24 14:23:56 2015 +0100" }, "committer": { "name": "David Brazdil", "email": "dbrazdil@google.com", "time": "Wed Jun 24 14:25:48 2015 +0100" }, "message": "ART: Stop creating a fallthrough block for Goto\n\nOptimizing\u0027s Builder used to create a basic block after a Goto under\nthe assumption that control flow can fall through.\n\nBug: 19084197\nChange-Id: Id85f31df98a4177466750d3cd0bc8bb74782ca2d\n" }, { "commit": "1f82ecc6a0c9f88d03d6d1a6d95eeb8707bd06c1", "tree": "50d71ef83f27827f0acfe6708acdb46e159f7506", "parents": [ "a12f729e7165780e73857e6248af0df00ad6ce2b" ], "author": { "name": "Nicolas Geoffray", "email": "ngeoffray@google.com", "time": "Wed Jun 24 12:20:24 2015 +0100" }, "committer": { "name": "Nicolas Geoffray", "email": "ngeoffray@google.com", "time": "Wed Jun 24 12:37:52 2015 +0100" }, "message": "Recompute dominator tree after DCE.\n\nbug:22031382\nChange-Id: Ifebe169897b76872015e3ce0ed7d0a9662f80cef\n" }, { "commit": "8b20f88b0a8d1b374dd5eaae289d19734c77b8f8", "tree": "7bb855bee3c93bd2d4edec359480b4606606559d", "parents": [ "d8d7b90ca57e8e3762951bb419b6f2645bb414c5" ], "author": { "name": "Nicolas Geoffray", "email": "ngeoffray@google.com", "time": "Fri Jun 19 16:17:05 2015 +0100" }, "committer": { "name": "Nicolas Geoffray", "email": "ngeoffray@google.com", "time": "Fri Jun 19 16:46:00 2015 +0100" }, "message": "Be careful with predecessor/successor index.\n\nWhen we simplify the CFG, we must preserve things that were already\nsimplified. For example, the index in the predecessor list or\nsuccessor list of a block must be preserved for ensuring the\nfirst block is a loop pre header.\n\nbug:21867463\n\nChange-Id: Ic3fcb3eb2c3fb109d8a57ee2a6b6d4d65fdb9410\n" }, { "commit": "01b6ec8c982400fe415533296ab0b563971c239d", "tree": "eab23a382221898f016db154fc047fecaef32724", "parents": [ "92fe57fe6e3512616f71b2b2183344c56d6753e6", "f78848f2ced8466b5fb2d7148d608288ee88757b" ], "author": { "name": "Nicolas Geoffray", "email": "ngeoffray@google.com", "time": "Thu Jun 18 15:51:54 2015 +0000" }, "committer": { "name": "Gerrit Code Review", "email": "noreply-gerritcodereview@google.com", "time": "Thu Jun 18 15:51:54 2015 +0000" }, "message": "Merge \"Don\u0027t special case HCurrentMethod in DCE.\"" }, { "commit": "18e6873c469b48aaed22148451523479eece98e3", "tree": "6df7efad2f7777bb0a4dbcbbe719750393eac531", "parents": [ "edb83c606e034d76bed1331f34cdc435df47bb95" ], "author": { "name": "Nicolas Geoffray", "email": "ngeoffray@google.com", "time": "Wed Jun 17 23:09:05 2015 +0100" }, "committer": { "name": "Nicolas Geoffray", "email": "ngeoffray@google.com", "time": "Wed Jun 17 23:11:39 2015 +0100" }, "message": "Fix a bug in optimizing when the null constant has been DCE.\n\nIf it has been DCE, we should create a new one, instead of\nusing the old one.\n\nAlso move the first DCE to a place where it could actually\nbe useful.\n\nbug:21870788\n\nChange-Id: I28fc52ae481ef92cba45fc1b5abcf07c995f524c\n" }, { "commit": "f78848f2ced8466b5fb2d7148d608288ee88757b", "tree": "72bfc782c51995b8c9d8ec41fafdae32f5341090", "parents": [ "edb83c606e034d76bed1331f34cdc435df47bb95" ], "author": { "name": "Nicolas Geoffray", "email": "ngeoffray@google.com", "time": "Wed Jun 17 11:57:56 2015 +0100" }, "committer": { "name": "Nicolas Geoffray", "email": "ngeoffray@google.com", "time": "Wed Jun 17 22:15:01 2015 +0100" }, "message": "Don\u0027t special case HCurrentMethod in DCE.\n\nInstead, re-create the HCurrentMethod if it is needed\nafter it has been removed.\n\nChange-Id: Id3bf15ae87b00a1d7eb35bf36d58fe96f788fba4\n" }, { "commit": "3584bce5b1f45e5741d3a6ca24884a36320ecb6b", "tree": "d9ccce473db996594e84135cb6d50d260063aabf", "parents": [ "cbc1e2ed4a7d3d43d59c20713888058097c8188f" ], "author": { "name": "Mingyao Yang", "email": "mingyao@google.com", "time": "Tue May 19 16:01:59 2015 -0700" }, "committer": { "name": "Mingyao Yang", "email": "mingyao@google.com", "time": "Wed Jun 10 16:51:57 2015 -0700" }, "message": "Fix premature deoptimization if the loop body isn\u0027t entered.\n\nAdd a test between initial_ and end_ to see if the loop body is entered.\nIf the loop body isn\u0027t entered at all, we jump to the loop header. Loop header is\nstill executed and is going to test the condition again and loop body won\u0027t be\nentered. This makes sure no deoptimization is triggered if the loop body isn\u0027t\neven entered.\n\nBug: 21034044\nChange-Id: I2b6de1f22fbc4568ca419f76382ebd87806d9694\n" }, { "commit": "e401d146407d61eeb99f8d6176b2ac13c4df1e33", "tree": "17927f9bfe7d2041b5942c89832d55f9dedb24c5", "parents": [ "2006b7b9b8e32722bd0d640c62549d8a0ac624b6" ], "author": { "name": "Mathieu Chartier", "email": "mathieuc@google.com", "time": "Wed Apr 22 13:56:20 2015 -0700" }, "committer": { "name": "Mathieu Chartier", "email": "mathieuc@google.com", "time": "Fri May 29 18:45:49 2015 -0700" }, "message": "Move mirror::ArtMethod to native\n\nOptimizing + quick tests are passing, devices boot.\n\nTODO: Test and fix bugs in mips64.\n\nSaves 16 bytes per most ArtMethod, 7.5MB reduction in system PSS.\nSome of the savings are from removal of virtual methods and direct\nmethods object arrays.\n\nBug: 19264997\nChange-Id: I622469a0cfa0e7082a2119f3d6a9491eb61e3f3d\n" }, { "commit": "d23eeef3492b53102eb8093524cf37e2b4c296db", "tree": "57d3e9ab2853d5b8092568bb3d29bc850c113315", "parents": [ "a15c78d3cc28f514a482ffd792a767e97fe53c95" ], "author": { "name": "Nicolas Geoffray", "email": "ngeoffray@google.com", "time": "Mon May 18 22:31:29 2015 +0100" }, "committer": { "name": "Nicolas Geoffray", "email": "ngeoffray@google.com", "time": "Fri May 29 14:15:04 2015 +0100" }, "message": "Support for inlining methods that call/throw.\n\nMostly fixes here and there to make it working.\n\nChange-Id: I1b535e895105d78b65634636d675b818551f783e\n" }, { "commit": "76b1e1799a713a19218de26b171b0aef48a59e98", "tree": "897d0d22d246367eb09d8b825b43c384074083f4", "parents": [ "382f5c24eb663ca8fa39a94a038349138a00272a" ], "author": { "name": "Nicolas Geoffray", "email": "ngeoffray@google.com", "time": "Wed May 27 17:18:33 2015 +0100" }, "committer": { "name": "Nicolas Geoffray", "email": "ngeoffray@google.com", "time": "Thu May 28 10:44:36 2015 +0100" }, "message": "Add a HCurrentMethod node.\n\nThis enables register allocation for the current method, so\nthat users of it don\u0027t always load it from the stack.\n\nCurrently only used by HLoadClass. Will make follow-up\nCLs for the other users.\n\nChange-Id: If73324d85643102faba47fabbbd2755eb258c59c\n" }, { "commit": "dc31313c4577bb4c00d90774159c84e738ed348d", "tree": "5f850028ebc8843cdcaf56c5fc122bd660937603", "parents": [ "03b5a398a96ea29c39ddfe3d810245b868d2871f", "80afd02024d20e60b197d3adfbb43cc303cf29e0" ], "author": { "name": "Vladimir Marko", "email": "vmarko@google.com", "time": "Tue May 26 17:42:31 2015 +0000" }, "committer": { "name": "Gerrit Code Review", "email": "noreply-gerritcodereview@google.com", "time": "Tue May 26 17:42:31 2015 +0000" }, "message": "Merge \"ART: Clean up arm64 kNumberOfXRegisters usage.\"" }, { "commit": "80afd02024d20e60b197d3adfbb43cc303cf29e0", "tree": "ef054c7b4f2a739f7cf063e0bc4c501c2c7e41b5", "parents": [ "559b178e34c5d92e7932f92e5d8a981ac334606f" ], "author": { "name": "Vladimir Marko", "email": "vmarko@google.com", "time": "Tue May 19 18:08:00 2015 +0100" }, "committer": { "name": "Vladimir Marko", "email": "vmarko@google.com", "time": "Tue May 26 15:59:02 2015 +0100" }, "message": "ART: Clean up arm64 kNumberOfXRegisters usage.\n\nAvoid undefined behavior for arm64 stemming from 1u \u003c\u003c 32 in\nloops with upper bound kNumberOfXRegisters.\n\nCreate iterators for enumerating bits in an integer either\nfrom high to low or from low to high and use them for\n\u003carch\u003eContext::FillCalleeSaves() on all architectures.\n\nRefactor runtime/utils.{h,cc} by moving all bit-fiddling\nfunctions to runtime/base/bit_utils.{h,cc} (together with\nthe new bit iterators) and all time-related functions to\nruntime/base/time_utils.{h,cc}. Improve test coverage and\nfix some corner cases for the bit-fiddling functions.\n\nBug: 13925192\nChange-Id: I704884dab15b41ecf7a1c47d397ab1c3fc7ee0f7\n" }, { "commit": "c7af85dad0dc392cfc0b373b0c1cb4b4197c89f4", "tree": "b49bf148c462e9b5f1911037b2ae830ed3beed6e", "parents": [ "8a8d8252f52bec14a8fe2b6f88e1f184c63cf351" ], "author": { "name": "David Brazdil", "email": "dbrazdil@google.com", "time": "Tue May 26 12:05:55 2015 +0100" }, "committer": { "name": "David Brazdil", "email": "dbrazdil@google.com", "time": "Tue May 26 12:05:55 2015 +0100" }, "message": "ART: Update graph\u0027s exit block field if removed\n\nRunning DCE on an infinite loop will delete the exit block but the\ncorresponding field is currently not cleared in the parent graph.\nThis does not cause any problems at the moment as that information is\nonly used in codegens to DCHECK that a block is not the exit block.\nHowever, it will be necessary to update the inliner once we start to\ninline methods with loops.\n\nWith this patch, DCE will update the HGraph::exit_block_ field. DCHECK\nwas also added to HGraph::InlineInto to make sure that the inlined\ngraph does have an exit block.\n\nChange-Id: Ia8ddca375bbc6830cd919af6059a52cc9b73a023\n" }, { "commit": "e82549b14c7def0a45461183964f7e6a34cbb70c", "tree": "9293e5bf58657883923fe08ff1964e92e81e8851", "parents": [ "c3912c8a2db109a15603554fd456f56cd0a69ad0" ], "author": { "name": "Mark Mendell", "email": "mark.p.mendell@intel.com", "time": "Wed May 06 10:55:34 2015 -0400" }, "committer": { "name": "Mark Mendell", "email": "mark.p.mendell@intel.com", "time": "Wed May 13 16:20:48 2015 -0400" }, "message": "[optimizing] Fold HTypeConversion of constants\n\nWhile looking into optimizing long shifts on x86, I found that the\ncompiler wasn\u0027t folding HTypeConversion of constants. Add simple\nconversions of constants, taking care of float/double values\nwith NaNs and small/large values, ensuring Java conversion semantics.\n\nAdd checker cases to see that constant folding of HTypeConversion is\ndone.\n\nEnsure 422-type-conversion type conversion routiness don\u0027t get\ninlined to avoid compile time folding.\n\nChange-Id: I5a4eb376b64bc4e41bf908af5875bed312efb228\nSigned-off-by: Mark Mendell \u003cmark.p.mendell@intel.com\u003e\n" }, { "commit": "a4b8c21dae70ae34aee13628632c39a675c06022", "tree": "3a72a4fcb64d61b7913f6ee3599a46b3248437b6", "parents": [ "faf935eb3a8565de98d9ec1f79c8c43c5c6eb2a7" ], "author": { "name": "David Brazdil", "email": "dbrazdil@google.com", "time": "Thu May 07 09:59:30 2015 +0100" }, "committer": { "name": "David Brazdil", "email": "dbrazdil@google.com", "time": "Tue May 12 11:37:53 2015 +0100" }, "message": "ART: Rediscover loops after deleting blocks in DCE\n\nThe way DCE currently updates loop information does not cover all\ncases. This patch removes the logic, resets loop information of live\nblocks to pre-SSA state and reanalyzes the affected loops.\n\nChange-Id: I0b996a70235b95a8db0de9a23a03f71db57a21b8\n" }, { "commit": "0a23d74dc2751440822960eab218be4cb8843647", "tree": "39d69de5d812826c4065d0acd38a58cd983f21f0", "parents": [ "cdeb0b5fede4c06488f43a212591e661d946bc78" ], "author": { "name": "Nicolas Geoffray", "email": "ngeoffray@google.com", "time": "Thu May 07 11:57:35 2015 +0100" }, "committer": { "name": "Nicolas Geoffray", "email": "ngeoffray@google.com", "time": "Mon May 11 14:17:22 2015 +0100" }, "message": "Add a parent environment to HEnvironment.\n\nThis code has no functionality change. It adds a placeholder\nfor chaining inlined frames.\n\nChange-Id: I5ec57335af76ee406052345b947aad98a6a4423a\n" }, { "commit": "8c0c91a845568624815df026cfdac8c42ecccdf6", "tree": "48f4a1f6158234c3257ed75405122b1fb0941f96", "parents": [ "46a4b26855ae9fcd14c6e4435f37f09ceb134f61" ], "author": { "name": "Nicolas Geoffray", "email": "ngeoffray@google.com", "time": "Thu May 07 11:46:05 2015 +0100" }, "committer": { "name": "Nicolas Geoffray", "email": "ngeoffray@google.com", "time": "Thu May 07 11:46:05 2015 +0100" }, "message": "Use a growable array instead of an environment during SSA.\n\nUsing an environment was convenient because it contains\na growable array. But there\u0027s no need for the environment\nabstraction when being used as a temporary holder for values\nof locals.\n\nChange-Id: Idf2883fe4b8f97a31ee70b3627c1bdd23ebfff0e\n" }, { "commit": "db216f4d49ea1561a74261c29f1264952232728a", "tree": "8b7914435ad1ba519a3d88b5cca7f0f6e842cd4f", "parents": [ "bc3b93eadd155342b6124d2d5ef3806ecec5dfd6" ], "author": { "name": "Nicolas Geoffray", "email": "ngeoffray@google.com", "time": "Tue May 05 17:02:20 2015 +0100" }, "committer": { "name": "Nicolas Geoffray", "email": "ngeoffray@google.com", "time": "Thu May 07 09:54:09 2015 +0100" }, "message": "Relax the only one back-edge restriction.\n\nThe rule is in the way for better register allocation, as\nit creates an artificial join point between multiple paths.\n\nChange-Id: Ia4392890f95bcea56d143138f28ddce6c572ad58\n" }, { "commit": "5d7b7f81ed5455893f984752c00571ef27cc97c5", "tree": "a497582ae4ab5a69ba7035c1f95af0ee0f0b025b", "parents": [ "01ce498499eed47e87fceb8736d26fe49b2a4346" ], "author": { "name": "Nicolas Geoffray", "email": "ngeoffray@google.com", "time": "Tue Apr 28 00:52:43 2015 +0100" }, "committer": { "name": "Nicolas Geoffray", "email": "ngeoffray@google.com", "time": "Fri May 01 10:07:52 2015 +0100" }, "message": "Update the remaining input index of phis after deleting an input.\n\nbug:20715803\nbug:20690906\n\nChange-Id: Iaf08f0c30d629e766be2b04815dc3e38b6e7ff35\n" }, { "commit": "69a2804c3bb48cf4fd00a66080f613a4fd96c422", "tree": "aab3f079d972bae71bbfa27fdca139738f41dbf0", "parents": [ "f073a36d1ed866e786f7d7784c709c86b00bc58e" ], "author": { "name": "David Brazdil", "email": "dbrazdil@google.com", "time": "Wed Apr 29 17:16:07 2015 +0100" }, "committer": { "name": "David Brazdil", "email": "dbrazdil@google.com", "time": "Wed Apr 29 17:36:52 2015 +0100" }, "message": "ART: Fix loop information after dead code elimination\n\nCompilation failed when only some blocks of a loop were removed during\ndead code elimination.\n\nBug: 20680703\nChange-Id: If31025169ca493f0d7f7f2788576e98d05f03394\n" }, { "commit": "fadf977d85bfbf3f9d5d4730afb660cbed34f989", "tree": "d64b1cd421d92a52930c44fc96cc822b19ff1430", "parents": [ "d474b1d0fd69fdeb481adda8c0814512956f2d51", "2967ec6c3dad1c1dc15fc827188bd5ecfa75493b" ], "author": { "name": "Roland Levillain", "email": "rpl@google.com", "time": "Wed Apr 29 10:23:07 2015 +0000" }, "committer": { "name": "Gerrit Code Review", "email": "noreply-gerritcodereview@google.com", "time": "Wed Apr 29 10:23:07 2015 +0000" }, "message": "Merge \"Add InsertInstructionAfter in HBasicBlock.\"" }, { "commit": "a0ee862288b702468f8c2b6d0ad0f1c61be0b483", "tree": "9917112ec01e4b734900cde975e465cf3bb2b670", "parents": [ "0fbfe6f92a2481daf914043262b5854e65d8c3cc", "f213e05cef6d38166cfe0cce8f3b0a53225a1b39" ], "author": { "name": "Nicolas Geoffray", "email": "ngeoffray@google.com", "time": "Mon Apr 27 14:12:24 2015 +0000" }, "committer": { "name": "Gerrit Code Review", "email": "noreply-gerritcodereview@google.com", "time": "Mon Apr 27 14:12:25 2015 +0000" }, "message": "Merge \"Add support for caching float and double constants.\"" }, { "commit": "f213e05cef6d38166cfe0cce8f3b0a53225a1b39", "tree": "2fd9573056062ae085f84f1b63b8ec3298927da1", "parents": [ "76bf84a196576f902a76a1165516a49dac15856f" ], "author": { "name": "Nicolas Geoffray", "email": "ngeoffray@google.com", "time": "Mon Apr 27 08:53:46 2015 +0000" }, "committer": { "name": "Nicolas Geoffray", "email": "ngeoffray@google.com", "time": "Mon Apr 27 15:05:20 2015 +0100" }, "message": "Add support for caching float and double constants.\n\nChange-Id: Ib5205bad1006bc5e3c9cc86bc82a6b4b1ce9bef9\n" }, { "commit": "c7508e93fa3df3a3890f6b62550cbd5e35bdd8df", "tree": "85795f0df38a741ab14c34e0403ba4bc3c317371", "parents": [ "76bf84a196576f902a76a1165516a49dac15856f" ], "author": { "name": "David Brazdil", "email": "dbrazdil@google.com", "time": "Mon Apr 27 13:28:57 2015 +0100" }, "committer": { "name": "David Brazdil", "email": "dbrazdil@google.com", "time": "Mon Apr 27 13:40:06 2015 +0100" }, "message": "ART: Fix removing a Phi with RemoveInstruction\n\nBoolean simplifier might attempt to remove a Phi from the Instruction\nlist.\n\nChange-Id: I698cc616549bd88dac96395cb2e5d09b5433d157\n" }, { "commit": "2967ec6c3dad1c1dc15fc827188bd5ecfa75493b", "tree": "f923c51b3a6aaa2a05e6fcf6fe0c68cbab718cef", "parents": [ "9f3565a632d12c9cadd7d966da308fd26dbc899c" ], "author": { "name": "Guillaume \"Vermeille\" Sanchez", "email": "guillaumesa@google.com", "time": "Fri Apr 24 16:36:52 2015 +0100" }, "committer": { "name": "Guillaume \"Vermeille\" Sanchez", "email": "guillaumesa@google.com", "time": "Mon Apr 27 13:04:35 2015 +0100" }, "message": "Add InsertInstructionAfter in HBasicBlock.\n\nChange-Id: I56e4e6edb39d1aab747877b7e517e94f0393f296\n" }, { "commit": "206d6fd6cae5ba8c4d5f0e230111fe77b9d5c0a5", "tree": "0684e7dd7ca2768991a839a5ddbe78f20bf1e4bf", "parents": [ "76f1413492c228bfa710e1eaa4c60370eaffbb8a" ], "author": { "name": "Mingyao Yang", "email": "mingyao@google.com", "time": "Mon Apr 13 16:46:28 2015 -0700" }, "committer": { "name": "Mingyao Yang", "email": "mingyao@google.com", "time": "Sun Apr 26 10:47:01 2015 -0700" }, "message": "Deoptimization-based BCE for unknown loop bounds.\n\nFor loop like:\n for (int i \u003d start; i \u003c end; i++) {\n array[i] \u003d 1;\n }\nWe add the following to the loop pre-header:\n if (start \u003c 0) deoptimize();\n if (end \u003e array.length) deoptimize();\nThen we can eliminate bounds-check of array[i] inside the loop.\n\nWe also take care of indexing with induction variable plus some offsets,\nlike array[i - 1]/array[i + 1] inside the loop, and adjust the condition\nfor deoptimization accordingly.\n\nChange-Id: I9e24c6b5e134ff95eff5b5605ff8f95d6546616f\n" }, { "commit": "eb5459ca861b58ee8a9907789f11400dcdddb87b", "tree": "36c1f328c48f0ec111ee31702cc4a82ddb2ad784", "parents": [ "ae803f6efbe8378b5423c51ee3c5564cae0e6e59", "4c0eb42259d790fddcd9978b66328dbb3ab65615" ], "author": { "name": "Roland Levillain", "email": "rpl@google.com", "time": "Fri Apr 24 18:17:40 2015 +0000" }, "committer": { "name": "Gerrit Code Review", "email": "noreply-gerritcodereview@google.com", "time": "Fri Apr 24 18:17:41 2015 +0000" }, "message": "Merge \"Ensure inlined static calls perform clinit checks in Optimizing.\"" }, { "commit": "4c0eb42259d790fddcd9978b66328dbb3ab65615", "tree": "9d1ac505dfd4d0225f479d860b72a58747c8f6ce", "parents": [ "223f2f5b2a20ca8246da1523494900a2424d5956" ], "author": { "name": "Roland Levillain", "email": "rpl@google.com", "time": "Fri Apr 24 16:43:49 2015 +0100" }, "committer": { "name": "Roland Levillain", "email": "rpl@google.com", "time": "Fri Apr 24 16:43:49 2015 +0100" }, "message": "Ensure inlined static calls perform clinit checks in Optimizing.\n\nCalls to static methods have implicit class initialization\n(clinit) checks of the method\u0027s declaring class in\nOptimizing. However, when such a static call is inlined,\nthe implicit clinit check vanishes, possibly leading to an\nincorrect behavior.\n\nTo ensure that inlining static methods does not change the\nbehavior of a program, add explicit class initialization\nchecks (art::HClinitCheck) as well as load class\ninstructions (art::HLoadClass) as last input of static\ncalls (art::HInvokeStaticOrDirect) in Optimizing\u0027 control\nflow graphs, when the declaring class is reachable and not\nknown to be already initialized. Then when considering the\ninlining of a static method call, proceed only if the method\nhas no implicit clinit check requirement.\n\nThe added explicit clinit checks are already removed by the\nart::PrepareForRegisterAllocation visitor. This CL also\nextends this visitor to turn explicit clinit checks from\nstatic invokes into implicit ones after the inlining step,\nby removing the added art::HLoadClass nodes mentioned\nhereinbefore.\n\nChange-Id: I9ba452b8bd09ae1fdd9a3797ef556e3e7e19c651\n" }, { "commit": "2d7352ba5311b8f57427b91b7a891e61497373c1", "tree": "3f3426f4f30663ee252ebc1f02ecd0eb114bad85", "parents": [ "c5cb691ca6a746a193bfbe3525aafa7cbb281d40" ], "author": { "name": "David Brazdil", "email": "dbrazdil@google.com", "time": "Mon Apr 20 14:52:42 2015 +0100" }, "committer": { "name": "David Brazdil", "email": "dbrazdil@google.com", "time": "Fri Apr 24 16:19:31 2015 +0100" }, "message": "ART: Dead block removal\n\nAdds a new pass which finds all unreachable blocks, typically due to\nsimplifying an if-condition to a constant, and removes them from the\ngraph. The patch also slightly generalizes the graph-transforming\noperations.\n\nChange-Id: Iff7c97f1d10b52886f3cd7401689ebe1bfdbf456\n" }, { "commit": "2cebb24bfc3247d3e9be138a3350106737455918", "tree": "d04d27d21b3c7733d784e303f01f873bb99e7770", "parents": [ "1f02f1a7b3073b8fef07770a67fbf94afad317f0" ], "author": { "name": "Mathieu Chartier", "email": "mathieuc@google.com", "time": "Tue Apr 21 16:50:40 2015 -0700" }, "committer": { "name": "Mathieu Chartier", "email": "mathieuc@google.com", "time": "Wed Apr 22 12:44:27 2015 -0700" }, "message": "Replace NULL with nullptr\n\nAlso fixed some lines that were too long, and a few other minor\ndetails.\n\nChange-Id: I6efba5fb6e03eb5d0a300fddb2a75bf8e2f175cb\n" }, { "commit": "c3d743fa2a26effcb35627d8a1338029c86e582a", "tree": "cf883e9b7f707ac2e6c2a6627e462b20e634c17c", "parents": [ "8d1ba74fe48703fc2b5a5920dd5eea0c65db5b15" ], "author": { "name": "David Brazdil", "email": "dbrazdil@google.com", "time": "Wed Apr 22 13:40:50 2015 +0100" }, "committer": { "name": "David Brazdil", "email": "dbrazdil@google.com", "time": "Wed Apr 22 14:43:19 2015 +0100" }, "message": "ART: Update last_instruction when adding Phis\n\nHBasicBlock::InsertPhiAfter would not update the last_instruction\npointer when adding at the end of the list. This could cause problems\nwhen iterating over phis backwards. Fortunately, we don\u0027t do that\nanywhere in the existing code.\n\nChange-Id: I4487265bf2cf3d3819623fafd7ce7c359bac190e\n" }, { "commit": "7d275379bf490a87805852129e3fe2e8afe961e7", "tree": "d0bbe856872c92773931c62e6881b0305a916236", "parents": [ "223f2f5b2a20ca8246da1523494900a2424d5956" ], "author": { "name": "David Brazdil", "email": "dbrazdil@google.com", "time": "Tue Apr 21 16:36:35 2015 +0100" }, "committer": { "name": "David Brazdil", "email": "dbrazdil@google.com", "time": "Tue Apr 21 17:17:11 2015 +0100" }, "message": "ART: Update loop info of all nested loops when inlining\n\nWhen inlining into a nested loop, the inliner would only add the new\nblocks into the innermost loop info object. This patch fixes that and\nmodifies SsaChecker to verify the property.\n\nChange-Id: I21d343a6f7d972f5b7420701f816c65ab3f20566\n" }, { "commit": "05144f4322eed049f4878015bf1f0381d419b785", "tree": "1608d15ead8f9f897b954560779b30232956b25d", "parents": [ "3e6f1c41220c84f31479a88aa37882307244110b" ], "author": { "name": "David Brazdil", "email": "dbrazdil@google.com", "time": "Thu Apr 16 15:18:00 2015 +0100" }, "committer": { "name": "David Brazdil", "email": "dbrazdil@google.com", "time": "Thu Apr 16 15:20:34 2015 +0100" }, "message": "ART: Hot fix for an inliner issue\n\nChange 147130 caused a problem with duplicit instruction ids when\ninlining methods with constants. This is a hot fix to unblock build.\n\nChange-Id: Ieddadcd94135930a1f29ad64ad57349a384da07f\n" }, { "commit": "da93333d568f3c5bd8eeb58341d10a332e1d42bf", "tree": "ce6aba4bb0214f733707c58634fa98e73a03563b", "parents": [ "669d8a1edbb2a78e08731a9cd6d8e815b0ec49db", "4a3faecbe4157225a3fe83a9ef7f4992dfc9c19d" ], "author": { "name": "David Brazdil", "email": "dbrazdil@google.com", "time": "Thu Apr 16 09:49:55 2015 +0000" }, "committer": { "name": "Gerrit Code Review", "email": "noreply-gerritcodereview@google.com", "time": "Thu Apr 16 09:49:56 2015 +0000" }, "message": "Merge \"ART: Don\u0027t duplicate null/int/long constants when inlining\"" }, { "commit": "4a3faecbe4157225a3fe83a9ef7f4992dfc9c19d", "tree": "90cd60a8f7e39f6d160c3e24598fc898cdac3d7e", "parents": [ "8e8bb8aab6f19ccb5b5869a632d9bc882891e17e" ], "author": { "name": "David Brazdil", "email": "dbrazdil@google.com", "time": "Thu Apr 16 10:38:44 2015 +0100" }, "committer": { "name": "David Brazdil", "email": "dbrazdil@google.com", "time": "Thu Apr 16 10:40:53 2015 +0100" }, "message": "ART: Don\u0027t duplicate null/int/long constants when inlining\n\nChange-Id: I7e6a3393fcbbcf76b4ba2000915ba6bbbfb7c70e\n" }, { "commit": "f776b92a0d52bb522043812dacb9c21ac11858e2", "tree": "619ae49853b201fc4ea9d0ac4b113c6226e3c339", "parents": [ "acf9b7b7616a9b104e6f2146051d8e14d9cb9030" ], "author": { "name": "Nicolas Geoffray", "email": "ngeoffray@google.com", "time": "Wed Apr 15 18:22:45 2015 +0100" }, "committer": { "name": "Nicolas Geoffray", "email": "ngeoffray@google.com", "time": "Wed Apr 15 21:51:12 2015 +0100" }, "message": "Remove dead blocks for the blocks_ array.\n\nThis prevents crashing because of structurally incorrect\nblocks. Also we now don\u0027t need to remove its instructions.\n\nTest case courtesy of Serguei I Katkov.\n\nChange-Id: Ia3ef9580549fc3546e8cd5f346079b1f0ceb2a61\n" }, { "commit": "d43b3ac88cd46b8815890188c9c2b9a3f1564648", "tree": "6c599c3f40d57e92786bd7f41c0541d9eaa2643b", "parents": [ "a109632b240f3c9355ca95500f6f48e4478e3c51" ], "author": { "name": "Mingyao Yang", "email": "mingyao@google.com", "time": "Wed Apr 01 14:03:04 2015 -0700" }, "committer": { "name": "Mingyao Yang", "email": "mingyao@google.com", "time": "Wed Apr 01 14:05:13 2015 -0700" }, "message": "Revert \"Revert \"Deoptimization-based bce.\"\"\n\nThis reverts commit 0ba627337274ccfb8c9cb9bf23fffb1e1b9d1430.\n\nChange-Id: I1ca10d15bbb49897a0cf541ab160431ec180a006\n" }, { "commit": "8d5b8b295930aaa43255c4f0b74ece3ee8b43a47", "tree": "c26fc49bbc74615e7f0b9657aaf3757a8282d7a9", "parents": [ "c8924c6ea9e83ba3832dd5551df38ab06f4aaca9" ], "author": { "name": "David Brazdil", "email": "dbrazdil@google.com", "time": "Tue Mar 24 10:51:52 2015 +0000" }, "committer": { "name": "David Brazdil", "email": "dbrazdil@google.com", "time": "Thu Mar 26 14:10:03 2015 +0000" }, "message": "ART: Force constants into the entry block\n\nOptimizations such as GVN and BCE make the assumption that all\nconstants are located in the entry block of the CFG, but not all\npasses adhere to this rule.\n\nThis patch makes constructors of constants private and only accessible\nto friend classes - HGraph for int/long constants and SsaBuilder for\nfloat/double - which ensure that they are placed correctly and not\nduplicated.\n\nNote that the ArenaAllocatorAdapter was modified to not increment\nthe ArenaAllocator\u0027s internal reference counter in order to allow\nfor use of ArenaSafeMap inside an arena-allocated objects. Because\ntheir destructor is not called, the counter does not get decremented.\n\nChange-Id: I36a4fa29ae34fb905cdefd482ccbf386cff14166\n" }, { "commit": "b2bd1c5f9171f35fa5b71ada42d1a9e11189428d", "tree": "db9165b3daa18d1d430b690b78c2d125bade3021", "parents": [ "11e99b19f48576f1bb6d0993635b34b6e09c9832" ], "author": { "name": "David Brazdil", "email": "dbrazdil@google.com", "time": "Wed Mar 25 11:17:37 2015 +0000" }, "committer": { "name": "David Brazdil", "email": "dbrazdil@google.com", "time": "Wed Mar 25 12:30:26 2015 +0000" }, "message": "ART: Formatting and comments in BooleanSimplifier\n\nChange-Id: I9a5aa3f2aa8b0a29d7b0f1e5e247397cf8e9e379\n" }, { "commit": "46e2a3915aa68c77426b71e95b9f3658250646b7", "tree": "2b0a4470b05291894db73c631fe94f0fdff8c46b", "parents": [ "bce0855ca1dbb1fa226c5b6a81760272ce0b64ef" ], "author": { "name": "David Brazdil", "email": "dbrazdil@google.com", "time": "Mon Mar 16 17:31:52 2015 +0000" }, "committer": { "name": "David Brazdil", "email": "dbrazdil@google.com", "time": "Tue Mar 24 17:28:37 2015 +0000" }, "message": "ART: Boolean simplifier\n\nThe optimization recognizes the negation pattern generated by \u0027javac\u0027\nand replaces it with a single condition. To this end, boolean values\nare now consistently assumed to be represented by an integer.\n\nThis is a first optimization which deletes blocks from the HGraph and\ndoes so by replacing the corresponding entries with null. Hence,\nexisting code can continue indexing the list of blocks with the block\nID, but must check for null when iterating over the list.\n\nChange-Id: I7779da69cfa925c6521938ad0bcc11bc52335583\n" }, { "commit": "0ba627337274ccfb8c9cb9bf23fffb1e1b9d1430", "tree": "0e1d0813c1d8d1c7239a900c1653296975713df0", "parents": [ "e295e6ec5beaea31be5d7d3c996cd8cfa2053129" ], "author": { "name": "Andreas Gampe", "email": "agampe@google.com", "time": "Tue Mar 24 02:39:46 2015 +0000" }, "committer": { "name": "Andreas Gampe", "email": "agampe@google.com", "time": "Tue Mar 24 02:39:46 2015 +0000" }, "message": "Revert \"Deoptimization-based bce.\"\n\nThis breaks compiling the core image:\n\n Error after BCE: art::SSAChecker: Instruction 219 in block 1 does not dominate use 221 in block 1.\n\nThis reverts commit e295e6ec5beaea31be5d7d3c996cd8cfa2053129.\n\nChange-Id: Ieeb48797d451836ed506ccb940872f1443942e4e\n" }, { "commit": "e295e6ec5beaea31be5d7d3c996cd8cfa2053129", "tree": "4d8a657d23d511743ce35bee596544d7f652efdb", "parents": [ "d24ba2c44c76a2b2dd13aafe8f7981c15be31a98" ], "author": { "name": "Mingyao Yang", "email": "mingyao@google.com", "time": "Sat Mar 07 06:37:59 2015 -0800" }, "committer": { "name": "Mingyao Yang", "email": "mingyao@google.com", "time": "Mon Mar 23 16:39:37 2015 -0700" }, "message": "Deoptimization-based bce.\n\nA mechanism is introduced that a runtime method can be called\nfrom code compiled with optimizing compiler to deoptimize into\ninterpreter. This can be used to establish invariants in the managed code\nIf the invariant does not hold at runtime, we will deoptimize and continue\nexecution in the interpreter. This allows to optimize the managed code as\nif the invariant was proven during compile time. However, the exception\nwill be thrown according to the semantics demanded by the spec.\n\nThe invariant and optimization included in this patch are based on the\nlength of an array. Given a set of array accesses with constant indices\n{c1, ..., cn}, we can optimize away all bounds checks iff all 0 \u003c\u003d min(ci) and\nmax(ci) \u003c array-length. The first can be proven statically. The second can be\nestablished with a deoptimization-based invariant. This replaces n bounds\nchecks with one invariant check (plus slow-path code).\n\nChange-Id: I8c6e34b56c85d25b91074832d13dba1db0a81569\n" }, { "commit": "97d30aca68af2aa4f114a01d7a7d73e4bf009fe0", "tree": "e441187dfab14ffab41aff6576df6aaa14ce8b1c", "parents": [ "3848c4fbc189287bca1298d45d07e21ec90c7c82", "b2fd7bca70b580921eebf7c45769c39d2dfd8a5a" ], "author": { "name": "Roland Levillain", "email": "rpl@google.com", "time": "Thu Mar 12 17:00:14 2015 +0000" }, "committer": { "name": "Gerrit Code Review", "email": "noreply-gerritcodereview@google.com", "time": "Thu Mar 12 17:00:15 2015 +0000" }, "message": "Merge \"Opt compiler: Basic simplification for arithmetic operations.\"" }, { "commit": "4f1a384762bf49fe8f3ecae8dd2bcb0e19d044a9", "tree": "3cb359f31e1f3080fb7cfffd5a6fd417f61753d7", "parents": [ "d304629202eec2ab053a47fc7bcf10223b5ccea4" ], "author": { "name": "Nicolas Geoffray", "email": "ngeoffray@google.com", "time": "Thu Mar 12 10:34:11 2015 +0000" }, "committer": { "name": "Nicolas Geoffray", "email": "ngeoffray@google.com", "time": "Thu Mar 12 10:57:47 2015 +0000" }, "message": "Give an expected type to phis created for multiple returns.\n\nWhen inlining, we used to take the type of the inlined method\nfor the phi in case of multiple returns. I recently changed the\nlogic of phi types to only be of int/float/double/ref, so we\nneed to call ToPhiType when creating the phi.\n\nChange-Id: I960067ca8a8814509c2a7c52c08387d892ebf4a3\n" }, { "commit": "b2fd7bca70b580921eebf7c45769c39d2dfd8a5a", "tree": "c5dae29519df73f889ba14495eb79d545cd7d6e5", "parents": [ "356286f989941ac495417195e4129aaceaf36a83" ], "author": { "name": "Alexandre Rames", "email": "alexandre.rames@arm.com", "time": "Wed Mar 11 16:48:16 2015 +0000" }, "committer": { "name": "Alexandre Rames", "email": "alexandre.rames@arm.com", "time": "Wed Mar 11 16:48:16 2015 +0000" }, "message": "Opt compiler: Basic simplification for arithmetic operations.\n\nThe optimisations in this patch do not look further than the\ninputs of each operation.\n\nChange-Id: Iddd0ab6b360b9e7bb042db22086d51a31be85530\n" }, { "commit": "817bce7658918b7a70c17b70aa5e6a46b1ae8b3d", "tree": "4dc41c5a319d948ff09666cec28f530282191f72", "parents": [ "0b6daeb0f0014474b542cbba1f713eb0dbefb7f9" ], "author": { "name": "Nicolas Geoffray", "email": "ngeoffray@google.com", "time": "Tue Feb 24 13:35:38 2015 +0000" }, "committer": { "name": "Nicolas Geoffray", "email": "ngeoffray@google.com", "time": "Tue Feb 24 14:44:46 2015 +0000" }, "message": "Fix inlining in the presence of multiple returns.\n\nOne return could actually return a phi, so doing a phi check for\nknowing if the result phi was already created was bogus.\n\nBug: 19454010\n\nChange-Id: Iee703a2d1071ae263092354465eda368e5d6770d\n" }, { "commit": "1abb4191a2e56d8dbf518efcaeefb266c1acdf2b", "tree": "e9cd0006df96b167c5da9f8f872713ff05b72803", "parents": [ "735969139b162f9d45a3c0e47dc24a8aec63c736" ], "author": { "name": "David Brazdil", "email": "dbrazdil@google.com", "time": "Tue Feb 17 18:33:36 2015 +0000" }, "committer": { "name": "David Brazdil", "email": "dbrazdil@google.com", "time": "Mon Feb 23 15:12:24 2015 +0000" }, "message": "Optimizing: Speed up HInstruction use removal\n\nSimilarly to a previous commit on HEnvironment use removal, this patch\nadds links from instructions to their respective inputs\u0027 use lists for\ncontant-time removal at the cost of doubling the size of input lists\n(from one pointer per entry to two). Manual testing shows that this\nsignificantly reduces the time required to transform HGraph to SSA\nform for some huge methods.\n\nChange-Id: I8dc3e4b0c48a50ac1481eb55c31093b99f4dc29f\n" }, { "commit": "acf735c13998ad2a175f5a17e7bfce220073279d", "tree": "94969f2387f0e6dad8c7e5712aa8187c9de2be56", "parents": [ "39109a06015c91188232e59fa9e60e0915d24cd7" ], "author": { "name": "Calin Juravle", "email": "calin@google.com", "time": "Thu Feb 12 15:25:22 2015 +0000" }, "committer": { "name": "Calin Juravle", "email": "calin@google.com", "time": "Thu Feb 19 17:07:52 2015 +0000" }, "message": "Reference type propagation\n\n- propagate reference types between instructions\n- remove checked casts when possible\n- add StackHandleScopeCollection to manage an arbitrary number of stack\nhandles (see comments)\n\nChange-Id: I31200067c5e7375a5ea8e2f873c4374ebdb5ee60\n" }, { "commit": "d6138ef1ea13d07ae555542f8898b30d89e9ac9a", "tree": "a8ffd5fd966512fd280bc1b3214f4e57a9e1805f", "parents": [ "92095533ac28879ddd8b44b559d700527ca12b8a" ], "author": { "name": "Nicolas Geoffray", "email": "ngeoffray@google.com", "time": "Wed Feb 18 14:48:53 2015 +0000" }, "committer": { "name": "Nicolas Geoffray", "email": "ngeoffray@google.com", "time": "Thu Feb 19 14:01:18 2015 +0000" }, "message": "Ensure the graph is correctly typed.\n\nWe used to be forgiving because of HIntConstant(0) also being\nused for null. We now create a special HNullConstant for such uses.\n\nAlso, we need to run the dead phi elimination twice during ssa\nbuilding to ensure the correctness.\n\nChange-Id: If479efa3680d3358800aebb1cca692fa2d94f6e5\n" }, { "commit": "be31ff94d66a0037c445eb57dc82f2a51bb46d9e", "tree": "2eb41a8786b2f0d1d42bc9bbecfe831b3b9bc17c", "parents": [ "a6919983bfc7e2c1eb69a4408f940146a2b78d59" ], "author": { "name": "Nicolas Geoffray", "email": "ngeoffray@google.com", "time": "Wed Feb 04 14:52:20 2015 +0000" }, "committer": { "name": "Nicolas Geoffray", "email": "ngeoffray@google.com", "time": "Wed Feb 04 15:08:37 2015 +0000" }, "message": "Fix a bug in the inliner.\n\nCode did not work in the presence of multiple returns.\nSpotted by Mark P. Mendell.\n\nChange-Id: I237050a0d79c0cfaa479e9b886f7450879e84713\n" }, { "commit": "276d9daaedfbff716339f94d55e6eff98b7434c6", "tree": "fdef0e2d95396aedae484dc5163689f282a8dd12", "parents": [ "bd17c3b46b035861ced434d6ed4d4fe183736e65" ], "author": { "name": "Nicolas Geoffray", "email": "ngeoffray@google.com", "time": "Mon Feb 02 18:24:11 2015 +0000" }, "committer": { "name": "Nicolas Geoffray", "email": "ngeoffray@google.com", "time": "Tue Feb 03 18:11:46 2015 +0000" }, "message": "Inline methods with multiple blocks.\n\nChange-Id: I3431af60e97fae230e0b6e98bcf0acc0ee9abf8c\n" }, { "commit": "82091dad38f3e5bfaf3b6984c9ab73069fb68310", "tree": "f66bd397f64a13ee102e45e0b6267c5d55e77505", "parents": [ "28acb6feb50951645c37c077bd3897ea760ca322" ], "author": { "name": "Nicolas Geoffray", "email": "ngeoffray@google.com", "time": "Mon Jan 26 10:02:45 2015 +0000" }, "committer": { "name": "Nicolas Geoffray", "email": "ngeoffray@google.com", "time": "Fri Jan 30 09:34:25 2015 +0000" }, "message": "Implement LICM in optimizing compiler.\n\nChange-Id: I9c8afb0a58ef45e568576015473cbfd5f011c242\n" }, { "commit": "ed59619b370ef23ffbb25d1d01f615e60a9262b6", "tree": "6c93bb6ceff95f7aaf232825e050eecc05c7282d", "parents": [ "f90eec005997f98c1a9f874fbbf68414e5f9c766" ], "author": { "name": "David Brazdil", "email": "dbrazdil@google.com", "time": "Fri Jan 23 10:39:45 2015 +0000" }, "committer": { "name": "David Brazdil", "email": "dbrazdil@google.com", "time": "Mon Jan 26 16:13:57 2015 +0000" }, "message": "Optimizing: Speed up HEnvironment use removal\n\nRemoval of use records from HEnvironment vregs involved iterating over\npotentially large linked lists which made compilation of huge methods\nvery slow. This patch turns use lists into doubly-linked lists, stores\npointers to the relevant nodes inside HEnvironment and subsequently\nturns the removals into constant-time operations.\n\nChange-Id: I0e1d4d782fd624e7b8075af75d4adf0a0634a1ee\n" }, { "commit": "9dec5a74014e96d41f84373990b4d4ac83206a88", "tree": "3aacf08f959a3d2558b9dc5985b6a3069506f4c8", "parents": [ "b6b114c02b8bacd3b5d64e646fdaefa03c069c61", "0ada95d8de4b04b5f201b4b7e9c3c2fd2cc321ae" ], "author": { "name": "Roland Levillain", "email": "rpl@google.com", "time": "Wed Jan 21 17:36:33 2015 +0000" }, "committer": { "name": "Gerrit Code Review", "email": "noreply-gerritcodereview@google.com", "time": "Wed Jan 21 17:36:33 2015 +0000" }, "message": "Merge \"ART: Replace NULL to nullptr in the optimizing compiler\"" }, { "commit": "0ada95d8de4b04b5f201b4b7e9c3c2fd2cc321ae", "tree": "97231800ac31c9ae1ce6351ca94d1d35dfc8caf8", "parents": [ "03c45ccf104fade857cb01a2b760b08fad5e304a" ], "author": { "name": "Jean Christophe Beyler", "email": "jean.christophe.beyler@intel.com", "time": "Thu Dec 04 11:20:20 2014 -0800" }, "committer": { "name": "Roland Levillain", "email": "rpl@google.com", "time": "Wed Jan 21 17:33:12 2015 +0000" }, "message": "ART: Replace NULL to nullptr in the optimizing compiler\n\nReplace macro NULL to the nullptr variation for C++.\n\nChange-Id: Ib6e48dd4bb3c254343383011b67372622578ca76\nSigned-off-by: Jean Christophe Beyler \u003cjean.christophe.beyler@intel.com\u003e\n" }, { "commit": "73d8fe409fbf2cb9665779690660ccc852d60431", "tree": "820f35d44a41fb0ff76f7d185f300862ce7a6f86", "parents": [ "36e8a91d975b4096b40c2b4bb6c6cd663cc5166d", "53d9da8507a1b68f036ce8669ad3f2ae9fc3d225" ], "author": { "name": "Roland Levillain", "email": "rpl@google.com", "time": "Wed Jan 21 17:26:11 2015 +0000" }, "committer": { "name": "Gerrit Code Review", "email": "noreply-gerritcodereview@google.com", "time": "Wed Jan 21 17:26:12 2015 +0000" }, "message": "Merge \"ART: Create a RemoveBlock method\"" }, { "commit": "6c2dff8ff8e1440fa4d9e1b2ba2a44d036882801", "tree": "da2d48b3d84733ac6b29194cb2f624693a643d48", "parents": [ "22c9285142169691eb2a9e2d4a49751fc7e57c2a" ], "author": { "name": "Nicolas Geoffray", "email": "ngeoffray@google.com", "time": "Wed Jan 21 14:56:54 2015 +0000" }, "committer": { "name": "Nicolas Geoffray", "email": "ngeoffray@google.com", "time": "Wed Jan 21 15:19:06 2015 +0000" }, "message": "Revert \"Revert \"Fully support pairs in the register allocator.\"\"\n\nThis reverts commit c399fdc442db82dfda66e6c25518872ab0f1d24f.\n\nChange-Id: I19f8215c4b98f2f0827e04bf7806c3ca439794e5\n" }, { "commit": "77520bca97ec44e3758510cebd0f20e3bb4584ea", "tree": "2e3be6fdc182e5cf5ae390019457af5e9c1ed242", "parents": [ "4d2c611bf17ff309abfa152e56c0b98a21ec8787" ], "author": { "name": "Calin Juravle", "email": "calin@google.com", "time": "Mon Jan 12 18:45:46 2015 +0000" }, "committer": { "name": "Calin Juravle", "email": "calin@google.com", "time": "Wed Jan 21 14:26:35 2015 +0000" }, "message": "Record implicit null checks at the actual invoke time.\n\nImplicitNullChecks are recorded only for instructions directly (see NB\nbelow) preceeded by NullChecks in the graph. This way we avoid recording\nredundant safepoints and minimize the code size increase.\n\nNB: ParallalelMoves might be inserted by the register allocator between\nthe NullChecks and their uses. These modify the environment and the\ncorrect action would be to reverse their modification. This will be\naddressed in a follow-up CL.\n\nChange-Id: Ie50006e5a4bd22932dcf11348f5a655d253cd898\n" }, { "commit": "c399fdc442db82dfda66e6c25518872ab0f1d24f", "tree": "6f0841ad5e8e80b09e34e084ae8eac336bce73a2", "parents": [ "41aedbb684ccef76ff8373f39aba606ce4cb3194" ], "author": { "name": "Nicolas Geoffray", "email": "ngeoffray@google.com", "time": "Wed Jan 21 12:42:57 2015 +0000" }, "committer": { "name": "Nicolas Geoffray", "email": "ngeoffray@google.com", "time": "Wed Jan 21 12:42:57 2015 +0000" }, "message": "Revert \"Fully support pairs in the register allocator.\"\n\nLibcore tests fail.\n\nThis reverts commit 41aedbb684ccef76ff8373f39aba606ce4cb3194.\n\nChange-Id: I2572f120d4bbaeb7a4d4cbfd47ab00c9ea39ac6c\n" }, { "commit": "41aedbb684ccef76ff8373f39aba606ce4cb3194", "tree": "94929237a0fe9b24dda7409d9433f07e82af4461", "parents": [ "97c89e4c081dcf4bacbde70b6609e366c9da417e" ], "author": { "name": "Nicolas Geoffray", "email": "ngeoffray@google.com", "time": "Wed Jan 14 10:49:16 2015 +0000" }, "committer": { "name": "Nicolas Geoffray", "email": "ngeoffray@google.com", "time": "Wed Jan 21 11:27:57 2015 +0000" }, "message": "Fully support pairs in the register allocator.\n\nEnabled on ARM for longs and doubles.\n\nChange-Id: Id8792d08bd7ca9fb049c5db8a40ae694bafc2d8b\n" } ], "next": "9ee66183d8e046ea661f642ba884626f16b46e06" }