)]}' { "log": [ { "commit": "3056d0ca38acc3d829a13ec8f97d35a002ee354e", "tree": "395028dc5fbfaf00f52594fe2df34b57b88e65f4", "parents": [ "ec139de9e77663c13a907a5244267db8bed947b5" ], "author": { "name": "Mathieu Chartier", "email": "mathieuc@google.com", "time": "Fri Oct 19 10:49:56 2012 -0700" }, "committer": { "name": "Mathieu Chartier", "email": "mathieuc@google.com", "time": "Fri Oct 19 11:33:07 2012 -0700" }, "message": "Improve accuracy of heap trim times\n\nBefore, WaitForConcurrentGc was being called inside of Heap::Trim.\nThis caused the printed trim times to be larger than they should be.\n\nChange-Id: Icc76b5ed7fb99350536d48a5215e7c1fdb8b4567\n" }, { "commit": "6f1c94968ada57da433debf8e2d1b38a80ceb510", "tree": "3628c7d9f5273d2c83dd7c6f9e816816eed8b70a", "parents": [ "0051be6d118c511308ed7764ef41e0d594292c49" ], "author": { "name": "Mathieu Chartier", "email": "mathieuc@google.com", "time": "Mon Oct 15 12:08:41 2012 -0700" }, "committer": { "name": "Mathieu Chartier", "email": "mathieuc@google.com", "time": "Mon Oct 15 12:37:33 2012 -0700" }, "message": "Add root verification when we try to mark an invalid object.\n\nNow when we try to mark an object not contained by any spaces, We call verify\nroots. This prints the root\u0027s vreg and method when it finds an invalid root.\n\nFixed a error in the total paused time statistic.\n\nChange-Id: Id10e4097cce56bc54ee488de32183c18ba3f3780\n" }, { "commit": "0051be6d118c511308ed7764ef41e0d594292c49", "tree": "aadcb9811b6034bf85d5a1c09d9264514f27dfac", "parents": [ "17f6cb9d269b8289f751551c80fe8b0dcfad363c" ], "author": { "name": "Mathieu Chartier", "email": "mathieuc@google.com", "time": "Fri Oct 12 17:47:11 2012 -0700" }, "committer": { "name": "Mathieu Chartier", "email": "mathieuc@google.com", "time": "Mon Oct 15 09:52:41 2012 -0700" }, "message": "Add parsed options for heap min free, heap max free, target utilization.\n\nAdded options in runtime for heap min free, max free, and target utilization.\n\nChange-Id: Iaea988ffbf6cb7f07127640786168de7d341f1e1\n" }, { "commit": "1c23e1edb7361bbaec6e57fca86d8d3797960ad2", "tree": "af8aa791777271fc5105f19a70aa972768493856", "parents": [ "ed6d5ed3494648780b9e91422c90d6bc22c16b79" ], "author": { "name": "Mathieu Chartier", "email": "mathieuc@google.com", "time": "Fri Oct 12 14:14:11 2012 -0700" }, "committer": { "name": "Mathieu Chartier", "email": "mathieuc@google.com", "time": "Fri Oct 12 17:34:08 2012 -0700" }, "message": "More space refactoring.\n\nAdd common interface, AllocSpace.\n\nRenamed the old AllocSpace to DLMallocSpace.\n\nAdded an new option enforce_target_size_, which when enabled, doesn\u0027t let\nthe heap grow past ideal heap size calculated during last Gc.\n\nRemoved redundant AllocationSize calls.\n\nMoved large object space to its own file instead of being in space.h/cc.\n\nChange-Id: I15e60531114bf213800599737cbd66ef44b46b15\n" }, { "commit": "155dfe958412b71ab2adcc21611217c22277ec69", "tree": "91db1d5a368c8b9bc26a432064a272ce2a788051", "parents": [ "bac4334bb222e24fb3a080b69de0262a39e397c2" ], "author": { "name": "Mathieu Chartier", "email": "mathieuc@google.com", "time": "Tue Oct 09 14:24:49 2012 -0700" }, "committer": { "name": "Mathieu Chartier", "email": "mathieuc@google.com", "time": "Thu Oct 11 18:18:21 2012 -0700" }, "message": "More Gc statistics and logging\n\nAdded a few metrics to Gc performance measurements:\nMean GC throughput, total time spent in Gc, total time spent in allocation code, mean time per allocation.\n\nSlightly optimized object allocation.\n\nChange-Id: Ie68153a7100affafc4a01dab34b341077e754867\n" }, { "commit": "f0bbeabf0577627e9fb29f204a036e3cd51e8bae", "tree": "7b7789bcd8ed03deb25a7c68aa6a0ccf82201324", "parents": [ "7204db592a999318067b9c509e14aa8b30293e58" ], "author": { "name": "Ian Rogers", "email": "irogers@google.com", "time": "Wed Oct 10 18:26:27 2012 -0700" }, "committer": { "name": "Ian Rogers", "email": "irogers@google.com", "time": "Wed Oct 10 18:26:27 2012 -0700" }, "message": "Improve heap lock annotations.\n\nFix a deadlock in non-concurrent mark sweep caught by this.\nBroaden heap_bitmap_lock_ over bitmap swapping.\n\nChange-Id: I5e749f25d181217d530e2f573dc8aee2685108ad\n" }, { "commit": "7469ebf3888b8037421cb6834f37f946646265ec", "tree": "b1eb8cd8d1e3b6263c2de77616020de4b4b5019f", "parents": [ "d8195f19840911a73b1491dfc8e7c18139753731" ], "author": { "name": "Mathieu Chartier", "email": "mathieuc@google.com", "time": "Mon Sep 24 16:28:36 2012 -0700" }, "committer": { "name": "Mathieu Chartier", "email": "mathieuc@google.com", "time": "Mon Oct 08 18:02:42 2012 -0700" }, "message": "Refactor and remove copy mark bits.\n\nRefactor code GC realted code to be in a GC folder.\n\nRemove copy mark bits by using pointer changing instead.\n\nEnable concurrent sweeping of system weaks.\n\nFix non concurrent GC plan.\n\nChange-Id: I9c71478be27d21a75f8a4e6af6faabe896e5e263\n" }, { "commit": "d8195f19840911a73b1491dfc8e7c18139753731", "tree": "0df5a92659dc0472f5368695a2e33ac37c149666", "parents": [ "59e5384d894836ace0b8ea22ff12d26ffb3026d4" ], "author": { "name": "Mathieu Chartier", "email": "mathieuc@google.com", "time": "Fri Oct 05 12:21:28 2012 -0700" }, "committer": { "name": "Mathieu Chartier", "email": "mathieuc@google.com", "time": "Mon Oct 08 10:53:25 2012 -0700" }, "message": "Reduce mark stack / allocation stack / live stack address space usage.\n\nWe now have upper bounds on the stack sizes so that we don\u0027t run out of\nvirtual addresses with large heaps.\n\nRename mark stack to atomic stack, which now takes any data type.\n\nAdded behaviour to force GC when the allocation stack becomes too full.\n\nAdded a new special map for reserving the oat file address range.\n\nChange-Id: I5169dd98b5f5072ac67637798da50cb8fc68af2b\n" }, { "commit": "50b35e2fd1a68cd1240e4a9d9f363e11764957d1", "tree": "4f3c58b7b293380f19e4f33357cb352e3330772e", "parents": [ "8e9a1497f0d7da4d55f6e7ed8a7d96ba6db7222d" ], "author": { "name": "Ian Rogers", "email": "irogers@google.com", "time": "Thu Oct 04 10:09:15 2012 -0700" }, "committer": { "name": "Ian Rogers", "email": "irogers@google.com", "time": "Fri Oct 05 11:07:39 2012 -0700" }, "message": "Explicitly pass Thread::Current to MutexLock and Alloc.\n\nChange-Id: I8b75bc0617915465f102815b32306aa7760dcae4\n" }, { "commit": "2fde53367dbe721e5273c34b590e67112322cc9e", "tree": "c5e10aa7d9d41b29298a42a0b96807787e27c574", "parents": [ "2e447e0e8725503d975c50fbbd13df4c769ca85e" ], "author": { "name": "Mathieu Chartier", "email": "mathieuc@google.com", "time": "Fri Sep 14 14:51:54 2012 -0700" }, "committer": { "name": "Mathieu Chartier", "email": "mathieuc@google.com", "time": "Thu Oct 04 10:14:03 2012 -0700" }, "message": "Refactor spaces and add free list large object space\n\nAdded some more abstraction for spaces, we now have ContinuousSpaces and DiscontinousSpaces.\n\nAdded a free list version of large object space.\n\nPerformance should be better than the memory map version since we avoid creating more than\none memory map.\n\nAdded a cause for Gc which prints with the Gc message, dalvik has this as well.\n\nChange-Id: Ie4aa6b204fbde7193e8305eb246158fae0444cc1\n" }, { "commit": "1f5393447b9f45be7918042d9ee7b521376de866", "tree": "1793cb0ae00b38a63f8cfd3a25b6cb0ee4edca16", "parents": [ "870ea4772199f8d4cbd9a269f9093620ee50a59c" ], "author": { "name": "Ian Rogers", "email": "irogers@google.com", "time": "Wed Oct 03 21:09:42 2012 -0700" }, "committer": { "name": "Ian Rogers", "email": "irogers@google.com", "time": "Thu Oct 04 09:27:04 2012 -0700" }, "message": "Make PopSirt inlinable, pass self to SirtRef.\n\nChange-Id: Ieb91526b1cb1f8644ceb3c5b99649f658f43bbc1\n" }, { "commit": "81d425b0b232962441616f8b14f73620bffef5e5", "tree": "cd1e46656269acf20e78817d675e56d9b9133e6c", "parents": [ "bfaf917edbb1de8d158c3615e0da8ac3143d10c8" ], "author": { "name": "Ian Rogers", "email": "irogers@google.com", "time": "Thu Sep 27 16:03:43 2012 -0700" }, "committer": { "name": "Ian Rogers", "email": "irogers@google.com", "time": "Thu Sep 27 17:19:25 2012 -0700" }, "message": "Pass self to lock methods.\n\nThis avoids frequent recomputation of\nThread::Current/pthread_getspecific.\n\nAlso add a futex based reader/writer mutex that is disabled.\n\nChange-Id: I118fdb99ef1d1c4bfda6446ba3a0d8b6ab31eaee\n" }, { "commit": "e0f0cb3d855cb5e926452b5e1ec8457adc4e454e", "tree": "20d9b12d5c27761905c3e8aed007d53734b55556", "parents": [ "23435d0df17e19fe312a08ff151464d4b9f83e81" ], "author": { "name": "Mathieu Chartier", "email": "mathieuc@google.com", "time": "Tue Aug 28 11:26:00 2012 -0700" }, "committer": { "name": "Mathieu Chartier", "email": "mathieuc@google.com", "time": "Mon Sep 24 14:05:08 2012 -0700" }, "message": "Large object space\n\nThe large object space helps prevent fragmentation by putting large objects in mem maps insead of the alloc space.\n\nInstead of mark and live bitmaps it uses mark and live sets.\n\nChange-Id: Iada5db70b88a1572007d8af921fa353681a55dc7\n" }, { "commit": "23435d0df17e19fe312a08ff151464d4b9f83e81", "tree": "3d0e4ad4d6f0f0852fa1736eaa0077399622263e", "parents": [ "5a6220c9068f8401394664b56a725441dba17948" ], "author": { "name": "Ian Rogers", "email": "irogers@google.com", "time": "Mon Sep 24 11:23:12 2012 -0700" }, "committer": { "name": "Ian Rogers", "email": "irogers@google.com", "time": "Mon Sep 24 11:48:14 2012 -0700" }, "message": "Assignability checks for SetFieldObject.\n\nAdds check to SetFieldObject that is enabled with VERIFY_OBJECT_ENABLED.\nThe check verifies that the object being set to the field is assignable\nto that field.\n\nFix bug that iftable was typed to be Object[][] but being assigned a\nObject[].\n\nChange-Id: I3d3744347f2dd142ca90db321ed876eaebfe7f7f\n" }, { "commit": "c7b83a0d8ac73bdfff751619ae2a34948e3534b7", "tree": "6b01745b26a1f1abd30e58efe7aadd56964761de", "parents": [ "88c95be62362537b46393e0c7e69d48f22e14ef3" ], "author": { "name": "Mathieu Chartier", "email": "mathieuc@google.com", "time": "Tue Sep 11 18:07:39 2012 -0700" }, "committer": { "name": "Mathieu Chartier", "email": "mathieuc@google.com", "time": "Wed Sep 12 10:49:20 2012 -0700" }, "message": "Better verification: Detection of missing card marks and dead system weaks.\n\nAdded a few additional verification functions.\n\nVerification of missing card marks. Checks that all objects on clean cards do not reference objects in live stack.\n\nVerification of system weaks. Checks that all of the system weaks are live.\n\nVerify objects seems to also be fixed now. It is however, rediculously slow and barely usable.\n\nChange-Id: Ieb5765df59210faa8fdf937cadf6ee51532d8bcb\n" }, { "commit": "866fb2ad19180a5e5a50bf649b38ec9bfd010ac6", "tree": "157ffb8e68a2776cb06e1ba9c5b7798c5074d791", "parents": [ "660ed62e187d1d92d1193d4c596838982a2ca138" ], "author": { "name": "Mathieu Chartier", "email": "mathieuc@google.com", "time": "Mon Sep 10 10:47:49 2012 -0700" }, "committer": { "name": "Mathieu Chartier", "email": "mathieuc@google.com", "time": "Mon Sep 10 13:44:38 2012 -0700" }, "message": "Better WaitForConcurrentGcToComplete information\n\nWaitForConcurrentGcToComplete now returns which type of Gc we waited on. This enables us to have smarter logic so that we don\u0027t run as many redundant Gcs.\n\nFixes 074-gc-thrash occasionally failing due to sticky mark bits not clearing older weak references.\n\nChange-Id: I7adb84183816e6a2acc11afd7ae650686fe97a7d\n" }, { "commit": "0325e6296d2370c42e7be80d846bfc7f8b28423b", "tree": "3e7e2192b3756ed1ee70921910ded374189dd5ad", "parents": [ "581b2ea85d9b382f6b14efadc9d2d36b526ef7c5" ], "author": { "name": "Mathieu Chartier", "email": "mathieuc@google.com", "time": "Wed Sep 05 14:22:51 2012 -0700" }, "committer": { "name": "Mathieu Chartier", "email": "mathieuc@google.com", "time": "Thu Sep 06 15:57:22 2012 -0700" }, "message": "Cumulative timing logger\n\nAdded a new class: Cumulative timing logger. This logger enables us to get an average of how long each Gc phase takes.\n\nChange-Id: I87874d4c25b5935ef30ef593a9bd36a27e45ae14\n" }, { "commit": "b726dcb581bf72da46527378ccb6889020f0e6e9", "tree": "c0383ed788652926e93147e50d659ed226519548", "parents": [ "fd678beb171a4686a4f2d53ca4188a4ade8fa54e" ], "author": { "name": "Ian Rogers", "email": "irogers@google.com", "time": "Wed Sep 05 08:57:23 2012 -0700" }, "committer": { "name": "Ian Rogers", "email": "irogers@google.com", "time": "Wed Sep 05 09:06:19 2012 -0700" }, "message": "Rename GlobalSynchronization to Locks\n\nAlso address some review comments in common_throws relating to\nToStr\u003cInvokeType\u003e and exception detail messages.\n\nChange-Id: Ibf2c0f147689fa236d349bd7f01eed3c2522552b\n" }, { "commit": "fd678beb171a4686a4f2d53ca4188a4ade8fa54e", "tree": "ab932e92d683d8033241bb2248da98e76249d3eb", "parents": [ "5bfa60ffcc953340feb711ed05cf576ac821905e" ], "author": { "name": "Mathieu Chartier", "email": "mathieuc@google.com", "time": "Thu Aug 30 14:50:54 2012 -0700" }, "committer": { "name": "Mathieu Chartier", "email": "mathieuc@google.com", "time": "Tue Sep 04 14:58:14 2012 -0700" }, "message": "Additional heap verification for the Gc\n\nThere are now two verification booleans which can be enabled. When these get enabled, it verifies that each live object only references other live objects.\n\nChanged SetClass to use SetFieldPtr to avoid having an extra card mark. This is safe since all classes are held live by the class linker.\n\nChange-Id: I005bb59e5cc8153a79d3ccb3d7b5cabd29fb4051\n" }, { "commit": "037813dfa7f575ad31ab0275892726c9b9b2ea4e", "tree": "35de964b0e2d64b1c000c86b0670c8ebabaa5405", "parents": [ "1cd9c5ccb15a7d9dd6219851b5df3d8c9e3393be" ], "author": { "name": "Mathieu Chartier", "email": "mathieuc@google.com", "time": "Thu Aug 23 16:44:59 2012 -0700" }, "committer": { "name": "Mathieu Chartier", "email": "mathieuc@google.com", "time": "Thu Aug 23 18:43:14 2012 -0700" }, "message": "Record allocation after SetClass to have memory fence for SetClass card mark.\n\nRecord allocation is now called after SetClass to ensure that our SetClass card mark is behind a fence.\n\nAlso:\nRemoved remaining LOCKS_EXCLUDED mentionning statistics lock.\nCleaned up some of the code in heap.cc\n\nChange-Id: Id781c6a9780ad72e90330acea647432974b240c6\n" }, { "commit": "1cd9c5ccb15a7d9dd6219851b5df3d8c9e3393be", "tree": "c25a64a53ecf8a6577a18294fb4e89a5a7d0b9da", "parents": [ "b18e77abdb06a443744fbb6589e0932fa89f6073" ], "author": { "name": "Mathieu Chartier", "email": "mathieuc@google.com", "time": "Thu Aug 23 10:52:44 2012 -0700" }, "committer": { "name": "Mathieu Chartier", "email": "mathieuc@google.com", "time": "Thu Aug 23 16:04:33 2012 -0700" }, "message": "Fix JPC regression\n\nJPC had regressed in boot time since we were incorrectly doing too many GCs.\n\nHaving the ImageSpace allocations counted as used memory caused us to incorrectly schedule concurrent GCs.\n\nChange-Id: Ic466a1281ffc68141a5a19f3e7b8f7dea45fbdf8\n" }, { "commit": "637e34860a22a61f2a0aa47bf37f2316f0dd3947", "tree": "a095e4c3a07bbb42b2250e85693f811cf3279639", "parents": [ "c7c18df6effe46669c2c197e6b5d3926728687b7" ], "author": { "name": "Mathieu Chartier", "email": "mathieuc@google.com", "time": "Fri Aug 17 10:41:32 2012 -0700" }, "committer": { "name": "Mathieu Chartier", "email": "mathieuc@google.com", "time": "Mon Aug 20 10:22:39 2012 -0700" }, "message": "Remove statistics lock to improve performance.\n\nRemoves the statistics lock for performance improvements. We now use android atomic operations to avoid write race conditions.\n\nA few other fixes are bundled with this change list.\n\nDeltablue time now down to ~17s,\n\nChange-Id: Ib80ea66c5362903bf637a32eeb8140941457fb7f\n" }, { "commit": "357e9be24c17a6bc2ae9fb53f25c73503116101d", "tree": "a170276eabf319d107242f928a68463fe0b5ebfb", "parents": [ "66aee5cd571cf4739d2735769304202ea5051fb8" ], "author": { "name": "Mathieu Chartier", "email": "mathieuc@google.com", "time": "Wed Aug 01 11:00:14 2012 -0700" }, "committer": { "name": "Mathieu Chartier", "email": "mathieuc@google.com", "time": "Fri Aug 17 10:06:50 2012 -0700" }, "message": "Sticky mark bits \"generational\" GC\n\nSticky mark bits GC. Sticky mark bits implemented using allocation stack which enables us to use the previous GC live bitmap as the mark bitmap.\n\nRemoved heap_end_ since checking versus it caused more overhead than it saved.\n\nRemoved check for impossible allocations at the start of AllocObject since these allocations will just fall through and fail anyways.\nThese allocations do not happen often enough for it to be worth checking for.\n\nA bunch of constant optimization performance improvements.\n\nPre locking regression benchmark improvements:\nDeltablue: ~0.3 sec runtime.\nCaffeineMark: ~300 total score due to improved string score.\n\nChange-Id: I15016f1ae7fdf76fc3aadb5774b527bf802d9701\n" }, { "commit": "00f7d0eaa6bd93d33bf0c1429bf4ba0b3f28abac", "tree": "6a2172ece15c4699e6c2a67ce76f019db0a9a21d", "parents": [ "634ea28f78c4a138e6a1de54eae8696095422415" ], "author": { "name": "Ian Rogers", "email": "irogers@google.com", "time": "Thu Jul 19 15:28:27 2012 -0700" }, "committer": { "name": "Ian Rogers", "email": "irogers@google.com", "time": "Tue Aug 14 10:45:52 2012 -0700" }, "message": "Global lock levels.\n\nIntroduce the notion of the mutators/GC being a shared-exclusive (aka\nreader-writer) lock. Introduce globally ordered locks, analysable by\nannotalysis, statically at compile time. Add locking attributes to\nmethods.\n\nMore subtly, remove the heap_lock_ and split between various locks that\nare held for smaller periods (where work doesn\u0027t get blocked). Remove\nbuggy Dalvik style thread transitions. Make GC use CMS in all cases when\nconcurrent is enabled. Fix bug where suspend counts rather than debug\nsuspend counts were sent to JDWP. Move the PathClassLoader to\nWellKnownClasses. In debugger refactor calls to send request and\npossibly suspend. Break apart different VmWait thread states. Move\nidentity hash code to a shared method.\n\nChange-Id: Icdbfc3ce3fcccd14341860ac7305d8e97b51f5c6\n" }, { "commit": "dcf8d7283bd51714f3faa55b631ae4103dc98b51", "tree": "ce4f6f1b836387d9b448fa1583295a38d7d743b8", "parents": [ "259487a86ea29c840d586610f099e8177aab3f40" ], "author": { "name": "Mathieu Chartier", "email": "mathieuc@google.com", "time": "Thu Aug 02 14:55:54 2012 -0700" }, "committer": { "name": "Mathieu Chartier", "email": "mathieuc@google.com", "time": "Fri Aug 03 12:41:54 2012 -0700" }, "message": "Fix zygote live/mark bitmap size.\n\nFixed some errors with the sizes of mark/live bitmaps after zygote space creation.\n\nThis was causing us to occasionally have overlapping mark/live bitmaps.\n\nAdded a new verify objects mode called VERIFY_OBJECT_FAST which only checks objects and not their classes.\n\nRefactored/optimized some of the scanning code to use xor to clear bits instead of and+not.\n\nChange-Id: Iec87d9157f69e6a558e300950b51d8781679e3f7\n" }, { "commit": "cc236d74772dda5a4161d9bc5f497fd3d956eb87", "tree": "2cd44e980357edc5b1a477fb261adb590d34c7c7", "parents": [ "f1ea6b072b8b473eabf91d1d582fb5966ffea4ba" ], "author": { "name": "Mathieu Chartier", "email": "mathieuc@google.com", "time": "Fri Jul 20 10:29:05 2012 -0700" }, "committer": { "name": "Mathieu Chartier", "email": "mathieuc@google.com", "time": "Fri Jul 27 14:30:51 2012 -0700" }, "message": "Zygote space / partial collection support.\n\nThe zygote space is now created right before zygote fork. This space has new allocations into it disabled, this reduces memory usage since we have more shared pages.\n\nPartial collection works by marking all the zygote space -\u003e alloc space references by using a mod-union table and then recursively marking only over the alloc space.\n\nApproximate improvements;\n\nDeltablue time goes down ~0.5 seconds.\n\nCaffeinemark score goes up ~300.\n\nSystem memory usage goes down ~7MB.\n\nChange-Id: I198389371d23deacd9b4534f39727eb641786b34\n" }, { "commit": "654d3a217faf46310895a1825354d610c2f3d6c2", "tree": "58d5f7ffd6d561466c52f6f2753d0c4da54e8d9a", "parents": [ "1b55e2ad030939d4d79e5af2945f076c88d46150" ], "author": { "name": "Mathieu Chartier", "email": "mathieuc@google.com", "time": "Wed Jul 11 17:54:18 2012 -0700" }, "committer": { "name": "Mathieu Chartier", "email": "mathieuc@google.com", "time": "Thu Jul 12 14:24:31 2012 -0700" }, "message": "Enable concurrent sweeping\n\nNow we unlock the heap during the sweeping process. When sweeping, the heap is only locked when dealing with dlmalloc.\n\nChange-Id: I705ac499adbf0039a3e57d2c5d354b1087317032\n" }, { "commit": "b062fdd4cb097fbae69b4bcb479c34d83ecab8ca", "tree": "215ea8fb42828a0f753ac5bd424ca098ce748342", "parents": [ "ca314c6a1be1b4cc11f4d284da90af7dc8a4ce25" ], "author": { "name": "Mathieu Chartier", "email": "mathieuc@google.com", "time": "Tue Jul 03 09:51:48 2012 -0700" }, "committer": { "name": "Mathieu Chartier", "email": "mathieuc@google.com", "time": "Wed Jul 11 17:17:46 2012 -0700" }, "message": "Each space has its own bitmap(s)\n\nEach alloc space now has One mark+live bitmap. Each image space has only one live bitmap.\n\nChange-Id: I2e919d1bd7d9f4d35d0e95ed83a58df6f754df6e\n" }, { "commit": "c9965d7090fec7c0d0fb4da5c77078d4a12d6c32", "tree": "517fd8b809056b7372b5287bfad70810c1111bd7", "parents": [ "1258417d54c85a2cbb98267b397c238070f28338" ], "author": { "name": "Logan Chien", "email": "loganchien@google.com", "time": "Tue Jul 10 03:51:57 2012 -0700" }, "committer": { "name": "Shih-wei Liao", "email": "sliao@google.com", "time": "Wed Jul 11 10:27:01 2012 -0700" }, "message": "Remove unused member function from art::Heap.\n\nChange-Id: I3f5417745cd7a3453bbbbd5159a5da79b2d44a16\n" }, { "commit": "b43b7d42b0c497333564e76be953157066c2b995", "tree": "1b40050796f54ac8f1cd4db70cd1bad571492f18", "parents": [ "57b47f9b59b8850c5cf7ddba587cc0cb094ca878" ], "author": { "name": "Mathieu Chartier", "email": "mathieuc@google.com", "time": "Tue Jun 19 13:15:09 2012 -0700" }, "committer": { "name": "Mathieu Chartier", "email": "mathieuc@google.com", "time": "Mon Jun 25 16:39:55 2012 -0700" }, "message": "Enable mod union table\n\nEnable mod union table(bitmap). This change aims to reduce the second pause times by processing dirty image cards and cleaning them by storing their alloc space references in a bitmap. The bitmap is then scanned concurrently(if CMS) before recursive mark.\n\nReduces second pause time. Second pause time now typically \u003c 5 ms instead of 30-40ms.\n\nChange-Id: I831eeb8af7299d4e86cf2dd9c29d0b7e230cd387\n" }, { "commit": "fc8cface168b09536869976d6b6f4665cffb2369", "tree": "4cf591f8a3eeef0f0d77abac18deb3f56738d110", "parents": [ "c553b95527b78c1fcf9b719a87d8875f71280f71" ], "author": { "name": "Mathieu Chartier", "email": "mathieuc@google.com", "time": "Tue Jun 19 11:56:36 2012 -0700" }, "committer": { "name": "Mathieu Chartier", "email": "mathieuc@google.com", "time": "Tue Jun 19 12:50:47 2012 -0700" }, "message": "Add missing WaitForConcurrentGcToComplete calls.\n\nSome calls to WaitForConcurrentGcToComplete were missing. This was enabling recursive GCs to occasionally occur.\n\nChange-Id: I0dcce50ae1ebfb03681990ec3bffa7b227981383\n" }, { "commit": "f8349361a16a4e2796efe9f3586b994e8d4834e4", "tree": "0a3474cde4028e1915c59393f49c147dd76ad766", "parents": [ "44335e189951a863607049a33571932fb6a2a841" ], "author": { "name": "Elliott Hughes", "email": "enh@google.com", "time": "Mon Jun 18 15:00:06 2012 -0700" }, "committer": { "name": "Elliott Hughes", "email": "enh@google.com", "time": "Mon Jun 18 15:16:21 2012 -0700" }, "message": "Use static thread safety analysis when available, and fix the bugs GCC finds.\n\nIt\u0027s impossible to express the Heap locking and the ThreadList locking with\nGCC, but Clang is supposed to be able to do it. This patch does what\u0027s possible\nfor now.\n\nChange-Id: Ib64a890c9d27c6ce255d5003cb755c2ef1beba95\n" }, { "commit": "a6399035ee2af012e2ac898ee7a753a2f83b35ab", "tree": "47d90d21e242ec91345e61f3cf118cd1adabe78b", "parents": [ "69362a053e87efe0e3596618f723a90d97551947" ], "author": { "name": "Mathieu Chartier", "email": "mathieuc@google.com", "time": "Mon Jun 11 18:49:50 2012 -0700" }, "committer": { "name": "Mathieu Chartier", "email": "mathieuc@google.com", "time": "Thu Jun 14 11:26:06 2012 -0700" }, "message": "Fix and re-enable CMS\n\nFix for CardTable::Scan double length scan which caused occasional assert failures.\nFix for the case where CMS request JNI code caused a GC which freed newest allocated object.\nFix error with is_gc_running_ never being set and deadlock due to CV wait without thread state change.\nFix error where we were Unlocking the heap before doing ResumeAll in CollectGarbageInternal which resulted in a rare race condition.\nFix error where GC could be in progress when the runtime terminated.\n\nChange-Id: I8d243d4dca63170433a63429442788ecb09e55fe\n" }, { "commit": "7664f5cd118b355a5fe0c7536cb48ac991ed2b62", "tree": "0564c36316e2aa0a892ff9914258433f14641a07", "parents": [ "9eb5f03e0dd90ac25e01fcf81f832c50a1f09d2c" ], "author": { "name": "Mathieu Chartier", "email": "mathieuc@google.com", "time": "Fri Jun 08 18:15:32 2012 -0700" }, "committer": { "name": "Mathieu Chartier", "email": "mathieuc@google.com", "time": "Mon Jun 11 11:17:13 2012 -0700" }, "message": "Fix GC performance regression\n\nEnable CMS and fix performance regression due to recursive marking image spaces. Dependent on my java change list.\n\nChange-Id: I4765792aa8226e811ac158f04ab88217db755573\n" }, { "commit": "262e5ffa1d4b23f23af3dea762a71a0af4bfd4a9", "tree": "f8378de518cdf8a6067d74aaa5808624d8c7709f", "parents": [ "2792fade8cb6554f81008b7635cec6ed0e914b3c" ], "author": { "name": "Mathieu Chartier", "email": "mathieuc@google.com", "time": "Fri Jun 01 17:35:38 2012 -0700" }, "committer": { "name": "Mathieu Chartier", "email": "mathieuc@google.com", "time": "Tue Jun 05 17:15:28 2012 -0700" }, "message": "Enable card cleaning\n\nCards which do not map to image spaces are now cleared at the start of GC. Cards which map to image spaces are now processed when we go through all the dirty cards after recursive mark. Dirty cards are handled based on if they are either image space cards or cards invalidated by mutators during CMS.\n\nChange-Id: I6c2dc177ebd3f7dc7bbbbd19ae67221371851ebe\n" }, { "commit": "5301cd241b4d8afbfc1211e107c41f1b15c6bd48", "tree": "187ea24dc81ee1b6ea713d6d037fc22a71e22df7", "parents": [ "b45a1f5c0d70c3694fbf941a5419e2eaa5dcfb80" ], "author": { "name": "Mathieu Chartier", "email": "mathieuc@google.com", "time": "Thu May 31 12:11:36 2012 -0700" }, "committer": { "name": "Mathieu Chartier", "email": "mathieuc@google.com", "time": "Thu May 31 14:00:37 2012 -0700" }, "message": "Mark stack create once per heap.\n\nMark stack is now created during heap initialization and is then re-used for each GC. This helps to prevent fragmentation of the heap.\n\nChange-Id: I5dd1bdfb58452415b88bfeb0c05a41ecbca09696\n" }, { "commit": "a4f94740f9d66b21f4bcd3a225c434aa29cfa323", "tree": "88e3834418ac3e12d2c270fed3a370bda2643d0b", "parents": [ "214d307f3d5bbfe6254ae9e96d504d3e665baa2c" ], "author": { "name": "Elliott Hughes", "email": "enh@google.com", "time": "Tue May 29 16:28:38 2012 -0700" }, "committer": { "name": "Elliott Hughes", "email": "enh@google.com", "time": "Tue May 29 16:28:38 2012 -0700" }, "message": "Finish cleanup of class/field/method caching.\n\nChange-Id: I289ae724cbd98487429275837d23b7b2d7096156\n" }, { "commit": "c967f78cd29b6019f7cfca40a02e9b677112da70", "tree": "60458d56fc398c9fb7b67b6d834b85f626ac2b31", "parents": [ "fd8ea97f64fb6d32cbfe35729b816a22dfa18224" ], "author": { "name": "Elliott Hughes", "email": "enh@google.com", "time": "Mon Apr 16 10:23:15 2012 -0700" }, "committer": { "name": "Elliott Hughes", "email": "enh@google.com", "time": "Mon Apr 16 10:23:15 2012 -0700" }, "message": "Include a new heap summary line in the SIGQUIT output.\n\nLooks like this:\n\n Heap: 87% free, 4MB/32MB; 6327 objects\n\nWhile I\u0027m here, fix another long-standing TODO to make PrettySize have the\nusual google3 behavior. (I took the specific thresholds from Chromium.)\n\nAlso distinguish between the more general \"Dump\" member functions and the\nspecific SIGQUIT-related ones by consistently calling the latter DumpForSigQuit.\n\nChange-Id: I76e783adc18dd089bac9b348f53dc9860a0fe4b9\n" }, { "commit": "250455229aa0cc07bbd18174efe510bd52631a99", "tree": "1499cfc216d6092057a4c4bf996c2ce1a2c360ca", "parents": [ "eb5701498830a406e957d11481296eb11734e850" ], "author": { "name": "jeffhao", "email": "jeffhao@google.com", "time": "Tue Mar 13 19:34:37 2012 -0700" }, "committer": { "name": "jeffhao", "email": "jeffhao@google.com", "time": "Fri Mar 23 16:49:39 2012 -0700" }, "message": "Added option to allow verification of objects on the stack.\n\nCurrently it\u0027s defined out by VERIFY_OBJECT_ENABLED. Also made various\nfixes to allow heap object verification to work again.\n\n(cherry picked from commit d4983495782be8d54c9c1f0266dc52395b947724)\n\nChange-Id: Icd625e1e0dc240c0ad1d10e2c58c7d5d63fec030\n" }, { "commit": "1bac54ffa933fbe9b92b62437577f2f4583eff1a", "tree": "a27134343ed3cd45a286c3a68b9a4469e70f13e7", "parents": [ "96970cdab3e28511322ce37fa578f6eff8931d72" ], "author": { "name": "Elliott Hughes", "email": "enh@google.com", "time": "Fri Mar 16 12:48:31 2012 -0700" }, "committer": { "name": "Elliott Hughes", "email": "enh@google.com", "time": "Fri Mar 16 12:48:31 2012 -0700" }, "message": "Fix all our unused parameter warnings so we let GCC report them.\n\nThere were a couple of genuine bugs here (fixed), plus there\u0027s a missing\nfeature in trace.cc that I\u0027ve just added a TODO for.\n\nAlso note that I haven\u0027t touched the compilers; this warning is still\nexplicitly disabled for that code. I\u0027ll do that when there\u0027s less going\non in those directories.\n\nChange-Id: Ic3570bf82411a07c7530bfaf1995ac995b9fc00f\n" }, { "commit": "fddf6f6b4af42c32dd2e92234bab0741118d092e", "tree": "3c6db33df447f729abfa0671e40936a89384dbd6", "parents": [ "cbf0b61f5415e96a76284deff87e7cb9aba022b9" ], "author": { "name": "Brian Carlstrom", "email": "bdc@google.com", "time": "Thu Mar 15 16:56:45 2012 -0700" }, "committer": { "name": "Brian Carlstrom", "email": "bdc@google.com", "time": "Thu Mar 15 16:57:03 2012 -0700" }, "message": "Add Heap::GetImageSpace()\n\nChange-Id: Ib52898043948a3db2539f7b689d5d97af370e9ca\n" }, { "commit": "b3bd5f07884f5a1f2b84224363b1372d7c28d447", "tree": "6e2997ab64b4a4f32d7ef539a4649adc736ea553", "parents": [ "ddbd01ac1660d57416879d5a576482f1048dde64" ], "author": { "name": "Elliott Hughes", "email": "enh@google.com", "time": "Thu Mar 08 21:05:27 2012 -0800" }, "committer": { "name": "Elliott Hughes", "email": "enh@google.com", "time": "Fri Mar 09 15:37:46 2012 -0800" }, "message": "Refactor the compilers out of libart.\n\nThis builds three separate compilers and dynamically links with the right one\nat runtime.\n\nChange-Id: I59d22b9884f41de733c09f97e29ee290236d5f4b\n" }, { "commit": "223f20fa939c386c695977263780dea2195093db", "tree": "642d1abe5dfbf2239bf239e8e91bebd333384378", "parents": [ "a56fcd60596ae8694da21fccde5c56832e437c56" ], "author": { "name": "Brian Carlstrom", "email": "bdc@google.com", "time": "Sat Feb 04 23:06:55 2012 -0800" }, "committer": { "name": "Brian Carlstrom", "email": "bdc@google.com", "time": "Sat Feb 04 23:24:56 2012 -0800" }, "message": "Remove old code to support multiple images\n\nChange-Id: I29bc2f216361ac4ed0cc0fefb7e3c46ee64e0ae8\n" }, { "commit": "8cf5bc0ab8742df107f0f4fc8ec0e991b6568f11", "tree": "2c21156aa0314dfa4f5b0cea9489e3aa16ebe743", "parents": [ "c27273f0508fb56ba35bead311151fbe45724d35" ], "author": { "name": "Elliott Hughes", "email": "enh@google.com", "time": "Thu Feb 02 16:32:16 2012 -0800" }, "committer": { "name": "Elliott Hughes", "email": "enh@google.com", "time": "Thu Feb 02 17:07:51 2012 -0800" }, "message": "Simple heap trimming.\n\nChange-Id: I8a94b637e64c6b5586db3c41b6e9230e7c8250c8\n" }, { "commit": "3bb17a644e2945c3913cfbde245d2f520d62a3ff", "tree": "5b9abb315069c2103d7bbaabc5a07d2a13bddb87", "parents": [ "715112b4551b6ead2643b5e4e9bd796331417115" ], "author": { "name": "Ian Rogers", "email": "irogers@google.com", "time": "Fri Jan 27 23:56:44 2012 -0800" }, "committer": { "name": "Ian Rogers", "email": "irogers@google.com", "time": "Sun Jan 29 12:50:16 2012 -0800" }, "message": "Space trim and other unit tests. General space clean up.\n\nThe space unit tests now include checks on space invariants, in\nparticular relating to footprint and size.\nOut-of-date comments have been removed.\nThis patch adds PrettySize and PrettyDuration methods to make these\nstrings more human readable.\n\nChange-Id: I6bc05b2db0d0115b97d666b832fce57bcdd2e091\n" }, { "commit": "30fab40ee5a07af6b8c3b6b0e9438071695a57f4", "tree": "86514e3535c8b23cf52249a202076b59ac532021", "parents": [ "f3778f66879fa038a7c9ebe46d5bc4167ddef7d8" ], "author": { "name": "Ian Rogers", "email": "irogers@google.com", "time": "Mon Jan 23 15:43:46 2012 -0800" }, "committer": { "name": "Ian Rogers", "email": "irogers@google.com", "time": "Thu Jan 26 17:32:10 2012 -0800" }, "message": "Upgrade to latest dlmalloc. Refactor Heap and related APIs to use STL like naming.\n\nWe fail assertions in the existing heap code, as does Dalvik. This refactoring\nis to clean the heap and space APIs and to reduce duplication of data\nand thereby solve a failing assertion in the card table.\n\nThis change also wires up clearing of soft references including before\nout-of-memory errors are reported.\n\nIn doing this change it was made clear that mspaces are buggy (and\nviolating invariants with the garbage collector). This\nchange upgrades to an un-Android molested version of dlmalloc-2.8.5 and\nimplements a version of the mspace morecore routine under ART control.\n\nrun-test 061-out-of-memory is updated for current heap sizes.\n\nChange-Id: I377e83ab2a8c78afb9b1881f03356929e2c9dc64\n" }, { "commit": "cd74c4b3a6893c876c6e03fd99a1264249653d80", "tree": "fd59dad11cb63ab6ff40be13e1572e1c2becab67", "parents": [ "844f9a072454eb9dca1022299b6bf99ef4294008" ], "author": { "name": "Brian Carlstrom", "email": "bdc@google.com", "time": "Mon Jan 23 13:21:00 2012 -0800" }, "committer": { "name": "Brian Carlstrom", "email": "bdc@google.com", "time": "Wed Jan 25 11:07:15 2012 -0800" }, "message": "Fix thread hang\n\n- Primary problem was ScopedThreadListLock was releasing heap lock in constructor instead of destructor\n- Secondary problem was ScopedThreadListLock should not be used with Mutex::Wait\n- Added Thread.getStackTrace case to ThreadStress that reproduces YouTube problem\n- Added Mutex::GetDepth and related methods that were useful in diagnoising this issue\n\nChange-Id: I1bdc7245e9b411378b98f4dcf498ad66eb96366d\n" }, { "commit": "4dd9b4d95eec9db5338fb9bf132f9bb8facf6cf4", "tree": "f9dc6f6e6a55acea3a08f1f568ae6900c1dd7dd1", "parents": [ "20a409c99a649a01cb7c65bd1089988445d8cdfd" ], "author": { "name": "Elliott Hughes", "email": "enh@google.com", "time": "Mon Dec 12 18:29:24 2011 -0800" }, "committer": { "name": "Elliott Hughes", "email": "enh@google.com", "time": "Tue Dec 13 08:23:08 2011 -0800" }, "message": "Clean up verbose logging.\n\nChange-Id: I6b01980e7145c4f8d25a5b50f620903f61a1b427\n" }, { "commit": "5d4bdc29737a693027daaf6ed3f0792368eb4bae", "tree": "6f1e9e59c2899488ab018d5cd8279641041f7634", "parents": [ "5e863ddd72a70d33525f7403a695f7bc1c218938" ], "author": { "name": "Ian Rogers", "email": "irogers@google.com", "time": "Wed Nov 02 22:15:43 2011 -0700" }, "committer": { "name": "Ian Rogers", "email": "irogers@google.com", "time": "Wed Nov 02 22:56:53 2011 -0700" }, "message": "Various runtime/JNI related fixes.\n\nNewObject returns NULL if exception is thrown in constructor.\nAllocate the peer then call constructor to ensure peer_ field is\ninitialized.\nChange thread state for AddFinalizer and pass current thread through for\nease.\n\nChange-Id: Ib578b6d44b08aef10fde5d8bc27cc6a2acbf6fae\n" }, { "commit": "5d76c435082332ef79a22962386fa92a0870e378", "tree": "db8fdd7bfba3617494157b483e9df979f76dfdcc", "parents": [ "4b6fe5a568ca2bc5e8fa110bf7af692cab220a15" ], "author": { "name": "Ian Rogers", "email": "irogers@google.com", "time": "Mon Oct 31 21:42:49 2011 -0700" }, "committer": { "name": "Ian Rogers", "email": "irogers@google.com", "time": "Wed Nov 02 01:25:57 2011 -0700" }, "message": "Mark non-image spaces and use write barrier for image spaces.\n\nDon\u0027t mark string and class roots that are in the image, alloc space\nreferences will be caught by the write barrier.\n\nChange-Id: Idcf9e4ede3b83556d4f8a01276273726dc6eea46\n" }, { "commit": "352a4244b62ad4a24eec6228dc4dd9983dfbf78e", "tree": "401e74441c510807691c83020eecfd2d4ba5c725", "parents": [ "23d5bc99a97e22323adc4b1a4237e5c45bdf7096" ], "author": { "name": "Elliott Hughes", "email": "enh@google.com", "time": "Mon Oct 31 15:15:21 2011 -0700" }, "committer": { "name": "Elliott Hughes", "email": "enh@google.com", "time": "Mon Oct 31 16:37:43 2011 -0700" }, "message": "Make it easier to run an optimized build, turn the GC timings back on, and remove some unnecessary code.\n\nThe Class* field is a regular instance field now, so no need to scan it twice.\nWe do our null checking in ScanObject, so we don\u0027t need to duplicate the checks\nin each child.\n\nAlso count the number of class, array, and other objects we scan.\n\nChange-Id: I9034708d5e0e1edffaccf3cad4785c5fccf3b38e\n" }, { "commit": "6a5bd495ff2f614f1495f652c86f3902d3bde537", "tree": "421434975640d4ddc592ac3caba6198277f83369", "parents": [ "e1fab56ffa50364fd08d4a935a2a3339f7faa509" ], "author": { "name": "Elliott Hughes", "email": "enh@google.com", "time": "Fri Oct 28 14:33:57 2011 -0700" }, "committer": { "name": "Elliott Hughes", "email": "enh@google.com", "time": "Fri Oct 28 15:30:08 2011 -0700" }, "message": "Implement the DDMS heap walking (for native and managed heaps).\n\nThis gets you the DDMS histograms of what\u0027s on your heaps.\n\nChange-Id: I7133d044030b10a787277faf3a77e20c565e69c5\n" }, { "commit": "c11607024d2e7a52ca690b8339306cab040e6653", "tree": "96ff9bd6a47dd71d18221a9037a6c1102ed64724", "parents": [ "371814fd131d5b7b54a5839d292240f9c4ed4160" ], "author": { "name": "jeffhao", "email": "jeffhao@google.com", "time": "Thu Oct 27 15:48:45 2011 -0700" }, "committer": { "name": "jeffhao", "email": "jeffhao@google.com", "time": "Fri Oct 28 11:57:50 2011 -0700" }, "message": "Added support for -XX:HeapGrowthLimit switch\n\nRuntime can now take in a growth limit for the heap, which can be\nsmaller than the maximum size of the heap, and will prevent the heap\nfrom growing beyond that size. The growth limit can also be cleared to\nincrease the size of the heap to the specified maximum size. This gives\napps some control of the size of the heap, but should be removed when we\nhave better ways to resize the heap.\n\nChange-Id: I338655bccd20bfd32e2318365f0f3283dbaaab1e\n" }, { "commit": "7162ad937f5f6bec32bf78d4675ff65cd6d1a233", "tree": "78bd960316746977400a9fc1772fbedeb821faae", "parents": [ "f7c3b6625d710a8700325eea447f65e9f963b7f2" ], "author": { "name": "Elliott Hughes", "email": "enh@google.com", "time": "Thu Oct 27 14:08:42 2011 -0700" }, "committer": { "name": "Elliott Hughes", "email": "enh@google.com", "time": "Thu Oct 27 14:08:42 2011 -0700" }, "message": "Implement DDMS heap info (\"HPIF\") chunks.\n\nThis lets you see how many bytes/objects are in your managed heap.\n\nChange-Id: Ie925207e9c48989a24968633e60b99314d220865\n" }, { "commit": "6b4ef025af12b158d117fc80fc79acf620f411a0", "tree": "223c580477cbb7bde240b599da4bbee793b8dcaf", "parents": [ "21b9f1355f0311b5c67bfdfa9dd63cacb1fba502" ], "author": { "name": "Brian Carlstrom", "email": "bdc@google.com", "time": "Sun Oct 23 14:59:04 2011 -0700" }, "committer": { "name": "Brian Carlstrom", "email": "bdc@google.com", "time": "Tue Oct 25 00:03:03 2011 -0700" }, "message": "Make ResolveField not rely on Field::GetType resolution\n\nChange-Id: I10f4a874809ac9db2cd54e200cf10eb7c8979fce\n" }, { "commit": "5e71b52a6f2c16fd02b389d220e60808cbc8f74a", "tree": "f13a4f98b80c6e48f11be02aff3e84de5e62c63e", "parents": [ "fb8a127fd2ca8eb6cbbc9d62b708c1fc2b3003a4" ], "author": { "name": "Elliott Hughes", "email": "enh@google.com", "time": "Thu Oct 20 13:12:32 2011 -0700" }, "committer": { "name": "Elliott Hughes", "email": "enh@google.com", "time": "Thu Oct 20 13:12:32 2011 -0700" }, "message": "Rename object_bitmap to heap_bitmap (since that\u0027s what the class is called).\n\nChange-Id: Idce6e9062545eb13b701e6b7e371c262977814d1\n" }, { "commit": "24a3c2e9924e8765c4a9b4d383cb8f3b922f9c9f", "tree": "1f61ecb377be4abcba4838651bc9d98b548a3f81", "parents": [ "3535c22d05bb4600255db90b490e625eda874b53" ], "author": { "name": "Brian Carlstrom", "email": "bdc@google.com", "time": "Mon Oct 17 18:07:52 2011 -0700" }, "committer": { "name": "Brian Carlstrom", "email": "bdc@google.com", "time": "Mon Oct 17 18:17:36 2011 -0700" }, "message": "ClassLinker changes\n\n- Add ClassLinker::LockOwner for SignalCatcher\n- ClassLinker::ResolveMethod now sets CodeAndDirectMethod when code is available\n\nChange-Id: I404b80d1994808e018d00e0b30e8ba4f9d08fecc\n" }, { "commit": "accd83d1523545ac69bafd38e72a7d5cff8e2fac", "tree": "1bad139a9b741483e34adc23b249e9f7fd1b6913", "parents": [ "6f7c958469d9b5a44ca2472058961a13d437b7bf" ], "author": { "name": "Elliott Hughes", "email": "enh@google.com", "time": "Mon Oct 17 14:25:58 2011 -0700" }, "committer": { "name": "Elliott Hughes", "email": "enh@google.com", "time": "Mon Oct 17 14:25:58 2011 -0700" }, "message": "Fix a heap lock/thread list lock deadlock.\n\nWe had an uncaught OOME whose uncaught exception handler -- running\nwith the thread lock held -- was trying to cause a GC while some\nother thread had the heap lock and was waiting for the thread list\nlock.\n\nChange-Id: I22177129562268837127d9edcc63ef5e93054bdf\n" }, { "commit": "307f75d6bcf7c32db7e1b43124dead628cc7ce96", "tree": "81fdf8690b0387e6535c85d11e24c2cc7b45cef3", "parents": [ "caab8c4ef372db5c119bfac1911fa27b174a935c" ], "author": { "name": "Elliott Hughes", "email": "enh@google.com", "time": "Wed Oct 12 18:04:40 2011 -0700" }, "committer": { "name": "Elliott Hughes", "email": "enh@google.com", "time": "Wed Oct 12 18:27:18 2011 -0700" }, "message": "Add a C++ equivalent of Android\u0027s TimingLogger, and use it for the heap.\n\nThis also changes us over to one big stats update at the end of sweeping,\nlike Dalvik, rather than recording every single free individually.\n\nChange-Id: Ib8e2a83d41d36e35e154183a4e173e915af79ae9\n" }, { "commit": "c33a32bccc4c66ed82ce3a580b16636399385cb4", "tree": "ef101d11157784c0724998d39e988a275ed4f69c", "parents": [ "a465c2034094eeb827ff9204951703dc5f269d60" ], "author": { "name": "Elliott Hughes", "email": "enh@google.com", "time": "Tue Oct 11 18:18:07 2011 -0700" }, "committer": { "name": "Elliott Hughes", "email": "enh@google.com", "time": "Tue Oct 11 18:18:07 2011 -0700" }, "message": "Sweep the monitor list.\n\nChange-Id: I343261206f8bbabd245b404dd95d532255e5d870\n" }, { "commit": "7f1caabe2936b5b6e34d5740607f5f5d63356fea", "tree": "c6676fd887d88565b2dac39d266ff8f3f81df5b2", "parents": [ "efdbac54eef6209772dadf7d24c7b3d37dcc8713" ], "author": { "name": "Shih-wei Liao", "email": "sliao@google.com", "time": "Thu Oct 06 12:11:04 2011 -0700" }, "committer": { "name": "Shih-wei Liao", "email": "sliao@google.com", "time": "Thu Oct 06 16:47:17 2011 -0700" }, "message": "Add an interface to Space, so Heap no longer needs to be friend of Space.\n\nAlso, fix and add comments with respect to utilization.\n\nChange-Id: Ia9d27cc16d2b85ac97a48fee80567829f41224f4\n" }, { "commit": "8c2f6414e1cff9a024a14369352f7ce112528b9e", "tree": "f3d9a60e81d416f9e88971091ae163f228b36de5", "parents": [ "fbfdce6952bee54a189af4b3984300301a5d62c1" ], "author": { "name": "Shih-wei Liao", "email": "sliao@google.com", "time": "Mon Oct 03 22:58:14 2011 -0700" }, "committer": { "name": "Shih-wei Liao", "email": "sliao@google.com", "time": "Wed Oct 05 18:42:36 2011 -0700" }, "message": "Grow for improving target utilization. Applicable to the target side only.\n\nChange-Id: If57ff026b66953abdfb56263589bb1b2521552d6\n" }, { "commit": "adb460d96702573de6e732ad58ca0380405cd928", "tree": "f70ad27d5fc007b70765a76644519234466dee61", "parents": [ "68e76526e98432625464022cb26f66b9ef6f5af4" ], "author": { "name": "Elliott Hughes", "email": "enh@google.com", "time": "Wed Oct 05 17:02:34 2011 -0700" }, "committer": { "name": "Elliott Hughes", "email": "enh@google.com", "time": "Wed Oct 05 17:02:34 2011 -0700" }, "message": "Implement finalization.\n\nAlso make System.gc no longer a no-op. This replaces some of the\nMemberOffsets exposed by Heap with intention-revealing functions\nthat we\u0027ll want to share between all the actual GC implementations,\nbut there\u0027s still more to be done.\n\nChange-Id: I57ba26c0416fbbfe20142b7b6e27108684add7f9\n" }, { "commit": "58ae9416e197ae68ed12ed43d87407d4dfb15093", "tree": "9fa93ce844f931155bf4a1d060f091f3285fb205", "parents": [ "4d0207c3ed28bbbb8c583a3c3a37f00d1cd4dedc" ], "author": { "name": "Brian Carlstrom", "email": "bdc@google.com", "time": "Tue Oct 04 00:56:06 2011 -0700" }, "committer": { "name": "Brian Carlstrom", "email": "bdc@google.com", "time": "Tue Oct 04 10:39:42 2011 -0700" }, "message": "Simplify command line arguments\n\nChange-Id: I9d6902b0c447fd8fbe5600fd36139791b2ceefb7\n" }, { "commit": "395520eaa47eca25b92e86188accf3095d60af49", "tree": "23cff36a71ea43eeb6a4bb7a7544c45aa3534c85", "parents": [ "14357e842b611279d467b64d450c569af33a6936" ], "author": { "name": "Brian Carlstrom", "email": "bdc@google.com", "time": "Sun Sep 25 19:35:00 2011 -0700" }, "committer": { "name": "Brian Carlstrom", "email": "bdc@google.com", "time": "Mon Sep 26 10:56:37 2011 -0700" }, "message": "Assorted fixes for running frameworks apps\n\n- Adding Heap::target_utililization_\n- Convert ClassNotFoundError to ClassNotFoundException in Class.forName\n- Finish String::GetHashCode\n- Changed Runtime.gc to unimplemented\n- Disable preload simulation in oat_process\n- Empty native placeholder for org.apache.harmony.dalvik.ddmc.DdmServer\n- Bonus: Fix problem with reflective invocation of constructors and private methods\n\nChange-Id: If11b273e2f82e8e0fa7161a756098514e1156a19\n" }, { "commit": "27ec961a1da540ba7f16c07a682585ab167317ad", "tree": "0aa6cb8345e394fe0ebe608bc26aa2acbdac1393", "parents": [ "03fa263ba90e9f6b11231576c6b9cc434e67141d" ], "author": { "name": "Brian Carlstrom", "email": "bdc@google.com", "time": "Mon Sep 19 20:20:38 2011 -0700" }, "committer": { "name": "Brian Carlstrom", "email": "bdc@google.com", "time": "Wed Sep 21 10:36:24 2011 -0700" }, "message": "Adding oat_process\n\n- Added oat_process, a version of app_process use to launch frameworks apps\n- Added liboat_runtime, a version of libandroid_runtime that uses art instead of dvm\n This is just a special makefile, frameworks/base/core/jni code is used for source\n- Added support for build a boot.oat with the full BOOTCLASSPATH\n The older smaller boat.oat is now core.oat\n- Added mem_map code for making sure a requested memory region is available\n Moved mem_map code to cc file to make easier to debug with smaller rebuild\n- Moved oat base address to 0x6000000 as a work around to host addres conflict\n- Added -Xms and -Xmx options to dex2oat to allow build specific memory options\n- Fixed miranda method initialization problem found compiling full bootclasspath\n- Made compiler.cc tolerant of verification errors found compiling full bootclasspath\n- Bumped arena block alloc warning to avoid noise when compiling full bootclasspath\n- Marked implicit GC unimplemented to fail fast\n- Added --output argument to oatdump\n- Made many object asserts tolerate access in IsErroneous state\n now that verifier is failing validation of some classes during compilation\n- Made runtime tolerate unknown access as short term solution for oat_process\n- Workaround SSA issue to restore full bootclasspath compilation\n- Added test-art-target-processs to excercise oat_process with \"am\"\n \"am\" found bug where class_linker was using Method::GetClass and not ::GetDeclaringClass\n\nChange-Id: I1a645a142b163e06bab9e72eb094ae1f1dbfbd97\n" }, { "commit": "9d5ccec86d60c9ddd811836b9a2bc28d0b3d11fe", "tree": "3a65213e8238770bb62ec54c8d82d230be405621", "parents": [ "9086572fa809d1a19d886b467e4da3ce42016982" ], "author": { "name": "Elliott Hughes", "email": "enh@google.com", "time": "Mon Sep 19 13:19:50 2011 -0700" }, "committer": { "name": "Elliott Hughes", "email": "enh@google.com", "time": "Mon Sep 19 15:06:39 2011 -0700" }, "message": "Implement as much of VMDebug as we can reasonably do right now.\n\nNo hprof and no method tracing, but everything else.\n\nChange-Id: Ifccd1f08e31f34b947c30f1211db788aae674d81\n" }, { "commit": "85d1545e985ac689db4bad7849880e843707c862", "tree": "dbda9214657085a3cbb77b323766da585e133a36", "parents": [ "dcb3dd8bfd6e2d5935442fe8671a2f6326a1e4e5" ], "author": { "name": "Elliott Hughes", "email": "enh@google.com", "time": "Fri Sep 16 17:33:01 2011 -0700" }, "committer": { "name": "Elliott Hughes", "email": "enh@google.com", "time": "Fri Sep 16 17:38:56 2011 -0700" }, "message": "Add packing to class Thread.\n\nOtherwise, code compiled on the host can use different offsets than\nwe need for the same structure on the target.\n\nAlso add commented-out code to start up the various GC daemon threads.\nMore Class native methods need to be implemented before that will work.\n\nChange-Id: I618b647b92378eec1b25cee469c8cfccf42f21fd\n" }, { "commit": "78128a63b2615744760b7f8ab83df9764a5d4a95", "tree": "544b24f88e1f8d3df2ab28ec2973476ad1b85a97", "parents": [ "d03935aea5192b32e3ca60a94db90b78c6702046" ], "author": { "name": "Brian Carlstrom", "email": "bdc@google.com", "time": "Thu Sep 15 17:21:19 2011 -0700" }, "committer": { "name": "Brian Carlstrom", "email": "bdc@google.com", "time": "Fri Sep 16 10:07:10 2011 -0700" }, "message": "oatdump\n\nFROM MAKE:\nmm dump-boot-oat\n\nON HOST:\noatdump --dex-file\u003d$ANDROID_PRODUCT_OUT/system/framework/core.jar --image\u003d$ANDROID_PRODUCT_OUT/system/framework/boot.oat --strip-prefix\u003d$ANDROID_PRODUCT_OUT\noatdump --boot-dex-file\u003d$ANDROID_PRODUCT_OUT/system/framework/core.jar --boot\u003d$ANDROID_PRODUCT_OUT/system/framework/boot.oat --dex-file\u003d$ANDROID_PRODUCT_OUT/system/framework/art-test-dex-HelloWorld.jar --image\u003d$ANDROID_PRODUCT_OUT//system/framework/art-test-dex-HelloWorld.oat --strip-prefix\u003d$ANDROID_PRODUCT_OUT\n\nON TARGET:\nadb shell oatdump --dex-file\u003d/system/framework/core.jar --image\u003d/system/framework/boot.oat\nadb shell oatdumpd --boot-dex-file\u003d/system/framework/core.jar --boot\u003d/system/framework/boot.oat --dex-file\u003d/system/framework/art-test-dex-HelloWorld.jar --image\u003d/system/framework/art-test-dex-HelloWorld.oat\n\nChange-Id: Iad2ae40a9cf2dc28799ff5dc5222d50f1bc6b39f\n" }, { "commit": "7ede61eceed2f8da14a4fff05c57c748c160d57f", "tree": "f3fc2ac910aaf0ad0295d914e48da820b5bbffd6", "parents": [ "8d768a954b101a9532f980253ac46be2c53aba11" ], "author": { "name": "Elliott Hughes", "email": "enh@google.com", "time": "Wed Sep 14 18:18:06 2011 -0700" }, "committer": { "name": "Elliott Hughes", "email": "enh@google.com", "time": "Wed Sep 14 18:18:06 2011 -0700" }, "message": "Implement VMRuntime.\n\n(Some of this forwards to unimplemented stuff in Heap.)\n\nChange-Id: I01e51d5673e61ddfdbfa2098310122502c3afcf1\n" }, { "commit": "5ee7a8b56dee896245e0f4ae5215f8b7376c1787", "tree": "da49e925c8a0833e6afc31e1e6558290ae9bc6ce", "parents": [ "3a4f8df8f75cbb9495b84a96c60e12fcce65ccf9" ], "author": { "name": "Elliott Hughes", "email": "enh@google.com", "time": "Tue Sep 13 16:40:07 2011 -0700" }, "committer": { "name": "Elliott Hughes", "email": "enh@google.com", "time": "Wed Sep 14 11:36:28 2011 -0700" }, "message": "Implement sun.misc.Unsafe and fix the jfieldID/jmethodID leak.\n\nChange-Id: I2997e19d9ae5e332a6731e7e10ebeff32bff085a\n" }, { "commit": "3a4f8df8f75cbb9495b84a96c60e12fcce65ccf9", "tree": "2e8b9d3a2ddd6e987162c48f0b0fb50577d3e7fd", "parents": [ "f0cde549bed96e16401a347a4511b59130c61e84" ], "author": { "name": "Elliott Hughes", "email": "enh@google.com", "time": "Tue Sep 13 15:22:36 2011 -0700" }, "committer": { "name": "Elliott Hughes", "email": "enh@google.com", "time": "Tue Sep 13 15:22:36 2011 -0700" }, "message": "A few more trivial write barrier calls.\n\nChange-Id: I4b0f3c4347eed9a2e0a494ae1382065c026d5c40\n" }, { "commit": "5ea047b386c5dac78eda62305d14dedf7b5611a8", "tree": "67f105b78df9d734a463b1194af11bf37ec6fe48", "parents": [ "1d3f114124b140629a7d22fa5cfa20ab8fc96934" ], "author": { "name": "Elliott Hughes", "email": "enh@google.com", "time": "Tue Sep 13 14:38:18 2011 -0700" }, "committer": { "name": "Elliott Hughes", "email": "enh@google.com", "time": "Tue Sep 13 14:38:18 2011 -0700" }, "message": "Support 64-bit volatiles and reference/pointer volatiles.\n\nChange-Id: I62c0751747767c02d6c57afd0fce3db6c8b02510\n" }, { "commit": "be759c63c6bb58b76ac71cad2c5a736bd31f374d", "tree": "60c5ab8fdb9d47bb0db5bda996aaac21caa4b1f3", "parents": [ "92b3b5623ec8b65f3e099c076e247bb8273692f8" ], "author": { "name": "Elliott Hughes", "email": "enh@google.com", "time": "Thu Sep 08 19:38:21 2011 -0700" }, "committer": { "name": "Elliott Hughes", "email": "enh@google.com", "time": "Thu Sep 08 19:50:39 2011 -0700" }, "message": "Add a heap HWM to the Thread*, for compiled code.\n\nAlso fix a bug in thread detach, and implement the thread exit callback.\n\nDestroy our pthread_mutex_t instances, and check for success. (This will\ncatch us deleting locked Mutex instances.)\n\nChange-Id: I26cf8117b825234f6c790e0cf70b2c025a743f84\n" }, { "commit": "92b3b5623ec8b65f3e099c076e247bb8273692f8", "tree": "94f85f0a8d33884b41227a1a560a755deccfffa2", "parents": [ "15041f433341f111665943ecfb9ca88c77d5c19a" ], "author": { "name": "Elliott Hughes", "email": "enh@google.com", "time": "Thu Sep 08 16:32:26 2011 -0700" }, "committer": { "name": "Elliott Hughes", "email": "enh@google.com", "time": "Thu Sep 08 18:11:55 2011 -0700" }, "message": "Add locking to the heap.\n\nNow tsan is happy with us too, at least on the host.\n\nChange-Id: Ib4657f56be2014de832dff8886b63843a40ea788\n" }, { "commit": "693267ad474039981e9be20a592ac2e4e3bd742e", "tree": "87a5bcfc0fd528e14f34f3e4212718100b58748e", "parents": [ "5fe594f576225dd7d333835e39c448a71ea9b433" ], "author": { "name": "Brian Carlstrom", "email": "bdc@google.com", "time": "Tue Sep 06 09:25:34 2011 -0700" }, "committer": { "name": "Brian Carlstrom", "email": "bdc@google.com", "time": "Thu Sep 08 15:05:14 2011 -0700" }, "message": "Fix GarbageCollector to work with VERIFY_OBJECT_ENABLED\n\nThere were two problems with the GC:\n1.) roots were being missed for most of the C++ fields\n2.) Heap::RecordFree(Object) was not being called, only Space::Free\n\nTo solve #1, added all C++ shadow fields to libcore.\nThis involed updating dalvik to cope with the new sizes and offsets.\nThis had the positive side effect of allowing a lot of special cases\nin the object scanning and image writing.\n\nTo solve #2, added a call to the now public Heap::RecordFree from MarkSweep\n\nGiven the now better working GC:\n- Reenabled HeapTest.GarbageCollectClassLinkerInit which is now passing.\n- ImageWriter now GC\u0027s before writing an image to avoid garbage.\n\nChange-Id: Ie7d1cc89e0bcf314cb37f0cabcb8593bf6e4d4be\n" }, { "commit": "69b15fb098162f19a4c20e6dccdcead04d9c77f0", "tree": "84fdfc26f560fb618b652a0aff98347f875a2653", "parents": [ "109bd6a38d0cd7c4b7797a9f2db8324c797d1368" ], "author": { "name": "Brian Carlstrom", "email": "bdc@google.com", "time": "Sat Sep 03 12:25:21 2011 -0700" }, "committer": { "name": "Brian Carlstrom", "email": "bdc@google.com", "time": "Tue Sep 06 15:28:47 2011 -0700" }, "message": "Working dex2oat and oatexec\n\nadb shell dex2oatd --dex-file\u003d/system/framework/core.jar --image\u003d/system/framework/boot.oat --base\u003d0x50000000 \"\u0027--method\u003dLjava/lang/System;logI(Ljava/lang/String;)V\u0027\" \"\u0027--method\u003dLjava/lang/System;log(CLjava/lang/String;Ljava/lang/Throwable;)V\u0027\"\nadb shell dex2oatd --boot-dex-file\u003d/system/framework/core.jar --boot\u003d/system/framework/boot.oat --dex-file\u003d/system/framework/art-test-dex-HelloWorld.jar --image\u003d/system/framework/art-test-dex-HelloWorld.oat\nadb shell oatexecd -Xbootclasspath:/system/framework/core.jar -Xbootimage:/system/framework/boot.oat -classpath /system/framework/art-test-dex-HelloWorld.jar -Ximage:/system/framework/art-test-dex-HelloWorld.oat HelloWorld\n\n09-05 17:58:18.912 2385 2385 I System : Hello, world!\n\nChange-Id: I53e534068584f0c3a837313e4d517a0e4a7154fc\n" }, { "commit": "9baa4aefc370f48774b6104680193d9a7e4fb631", "tree": "26ed5735022c29f77d6b025c357674ecf50b27c8", "parents": [ "e9ea8d1f002dd86976f0e1493ab8aa23a66c0494" ], "author": { "name": "Brian Carlstrom", "email": "bdc@google.com", "time": "Thu Sep 01 21:14:14 2011 -0700" }, "committer": { "name": "Brian Carlstrom", "email": "bdc@google.com", "time": "Sat Sep 03 13:25:31 2011 -0700" }, "message": "Remove use of MemMap for code allocation.\n\nIn preparation of moving to an executable code space.\n\nChange-Id: I6cb6fec93c66005f8020d718fe3c1c5aa1c85356\n" }, { "commit": "3e465b1a72d89f7c30f18efb90dc688d33d46c16", "tree": "2bac1709132fd261d54abdb3601f64065ea210b7", "parents": [ "0db460b881a3fdaa63d581a1dc39e417d682faaa" ], "author": { "name": "Elliott Hughes", "email": "enh@google.com", "time": "Fri Sep 02 18:26:12 2011 -0700" }, "committer": { "name": "Elliott Hughes", "email": "enh@google.com", "time": "Fri Sep 02 18:26:12 2011 -0700" }, "message": "Get the test speed back to where it was.\n\n(This patch actually has an active ingredient, unlike my last screwup.)\n\nChange-Id: Ic9f48a1c7c62ea024c51dc57e245a7d4df068db0\n" }, { "commit": "0cfe1fb7060576d047f7f894fc0d8b87de84fcab", "tree": "4255eae2034dcb0febe28411c50cba76189f39ae", "parents": [ "a09576416788b916095739e43a16917e7948f3a4" ], "author": { "name": "Ian Rogers", "email": "irogers@google.com", "time": "Fri Aug 26 03:29:44 2011 -0700" }, "committer": { "name": "Ian Rogers", "email": "irogers@google.com", "time": "Fri Sep 02 14:48:01 2011 -0700" }, "message": "Use accessor methods for Object fields.\n\nEnsure that Object fields are modified via accessors so that it\u0027s easy\nto insert barriers (make all fields within Objects private). Add validity\nchecks to Field and Method accessors to ensure they are accessed when a\nClass is in a suitable state. Add validity checks to all Object\naccessors to check heap isn\u0027t corrupted. Remove hacked in strings from Field\nand Method; make type fields used the dex cache that is lazily initialized.\nClean up various other TODOs and lint issues.\n\nChange-Id: Iac0afc515c01f5419874d9cdcdb9a7b45443e3fb\n" }, { "commit": "410c0c876f326e14c176a39ba21fc4dd3f7db8ab", "tree": "e4b77ac5ed99b7206c51f6519689fdbc623e843d", "parents": [ "1b4c85959b3d9a4a33bc2160c46c1bbde67350c7" ], "author": { "name": "Elliott Hughes", "email": "enh@google.com", "time": "Thu Sep 01 17:58:25 2011 -0700" }, "committer": { "name": "Elliott Hughes", "email": "enh@google.com", "time": "Thu Sep 01 22:16:18 2011 -0700" }, "message": "Mark more roots.\n\nThis is most of the stuff. (Everything that currently exists, though there\u0027s\nmore to come.)\n\nChange-Id: I235a21b006820a027c494374a5b52ffefed89c32\n" }, { "commit": "cf4c6c41b0084dc4567ff709fb8ce9ebd72b26ac", "tree": "ca3d0314b5c4111d7ab6fd97b4e68d5b2679e237", "parents": [ "2f0ce9d60a4a9371c63a32a3764320fb02341acb" ], "author": { "name": "Elliott Hughes", "email": "enh@google.com", "time": "Thu Sep 01 15:16:42 2011 -0700" }, "committer": { "name": "Elliott Hughes", "email": "enh@google.com", "time": "Thu Sep 01 15:51:43 2011 -0700" }, "message": "Fully implement string interning.\n\nAlso, more const.\n\nChange-Id: I09cae88d677e8e6e42d0fa9b5d1093c79d225e66\n" }, { "commit": "bf86d0438e9ef9c145ebcf16a2e74c4efaa2686a", "tree": "e8641925dd68c6e03b41ceb3de58108436f988a0", "parents": [ "ad7c2a3b4daa2abd707375444f4b0db7d69a3838" ], "author": { "name": "Elliott Hughes", "email": "enh@google.com", "time": "Wed Aug 31 17:53:14 2011 -0700" }, "committer": { "name": "Elliott Hughes", "email": "enh@google.com", "time": "Thu Sep 01 08:51:12 2011 -0700" }, "message": "Start adding implementations for runtime-provided native methods.\n\nThe library can\u0027t do everything...\n\nChange-Id: Ib808c00570c7214aeb2ca058b1a66cacbeb372f1\n" }, { "commit": "1f87008b165d26541d832ff805250afdc89c253d", "tree": "a84122b3757b8f856eb0656a951ed6621b6d01a8", "parents": [ "c5bfa8f49d8548d7c685a99b411311ef56bedffa" ], "author": { "name": "Brian Carlstrom", "email": "bdc@google.com", "time": "Tue Aug 23 16:02:11 2011 -0700" }, "committer": { "name": "Brian Carlstrom", "email": "bdc@google.com", "time": "Wed Aug 31 09:46:57 2011 -0700" }, "message": "Add HeapTest and make GC work enough to pass it\n\nChange-Id: If06eaef2921b64b3226bfd347acaec60ec993e67\n" }, { "commit": "a2501990dd0f68baf38ce19251949d7bb3ecfe5a", "tree": "3f72ceee0297d213dac66a9ce2c72592906c1803", "parents": [ "82870720a9a2709fcde4a03d96be383cee4b25d3" ], "author": { "name": "Elliott Hughes", "email": "enh@google.com", "time": "Fri Aug 26 19:39:54 2011 -0700" }, "committer": { "name": "Elliott Hughes", "email": "enh@google.com", "time": "Mon Aug 29 22:06:55 2011 -0700" }, "message": "Add -Xcheck:jni.\n\nI think this is as complete as possible right now. The remaining\ntwo #if 0 sections require:\n\n1. a way to get the Method* of the current native method.\n2. a way to get the Class* of the type of a given Field*.\n\nChange-Id: I331586022095fb36ccc10c9ac1890a59a9224d01\n" }, { "commit": "408f79aeb676251ba35667a64e86c20638d7cb0b", "tree": "419c9dca6bc1fd5c13a997d4fad62f92960e5e45", "parents": [ "72025e5f8f2743b4af33453df5b3bddada2078db" ], "author": { "name": "Ian Rogers", "email": "irogers@google.com", "time": "Tue Aug 23 18:22:33 2011 -0700" }, "committer": { "name": "Ian Rogers", "email": "irogers@google.com", "time": "Tue Aug 23 18:56:54 2011 -0700" }, "message": "Add validate object, clean up stack indirect reference table.\n\nChange-Id: Ifb329ae7d3bede3e95d48e6761cee1412a33d867\n" }, { "commit": "a663ea5de4c9ab6b1510fdebd6d8eca77ba699ae", "tree": "37078335c64cb760a0379a33eb4ceaf6b08356ea", "parents": [ "68f4fa09cfd72b4292cbe7a213f03da7902ba775" ], "author": { "name": "Brian Carlstrom", "email": "bdc@google.com", "time": "Fri Aug 19 23:33:41 2011 -0700" }, "committer": { "name": "Brian Carlstrom", "email": "bdc@google.com", "time": "Sun Aug 21 12:30:48 2011 -0700" }, "message": "Initialize ClassLinker from image\n\nChange-Id: Ibaf47b4181f7c6603a8b37e2eba8fa6509c927ed\n" }, { "commit": "9cff8e13d41825c4f3f0127af061e94b06114fc8", "tree": "3425941132a8b1406126015cf9c45fd661b574a2", "parents": [ "67bf885d62b1473c833bece1c9e0bb624e6ba391" ], "author": { "name": "Brian Carlstrom", "email": "bdc@google.com", "time": "Thu Aug 18 16:47:29 2011 -0700" }, "committer": { "name": "Brian Carlstrom", "email": "bdc@google.com", "time": "Fri Aug 19 22:01:22 2011 -0700" }, "message": "Convert Class::descriptor_ from StringPiece to String (as part of image loading)\n\nChange-Id: Iab5ffa353661a7c06ee79af1f40d399a53777174\n" }, { "commit": "4a289ed61242964b921434de7d375f46480472a1", "tree": "fa1a25f1eec3ef79fd9765e9c89ce21afbd8144c", "parents": [ "44753c3908fcd5ccc066bf0dceddeabfeacf9f24" ], "author": { "name": "Brian Carlstrom", "email": "bdc@google.com", "time": "Tue Aug 16 17:17:49 2011 -0700" }, "committer": { "name": "Brian Carlstrom", "email": "bdc@google.com", "time": "Thu Aug 18 12:25:10 2011 -0700" }, "message": "Work on heap and space initialization to support image loading\n\nChange-Id: Icab25efa4dee17e4b6c6e97e38f63f5ab8a8a005\n" }, { "commit": "8a436595d36c1e4935984fcac249d7d877e00383", "tree": "66e6ecb9afb1c9950285c52fe98b93206436fccd", "parents": [ "a5b897eae4b6f9f9608faa9eada7ddf42bf1bfd2" ], "author": { "name": "Brian Carlstrom", "email": "bdc@google.com", "time": "Mon Aug 15 21:27:23 2011 -0700" }, "committer": { "name": "Brian Carlstrom", "email": "bdc@google.com", "time": "Tue Aug 16 13:50:46 2011 -0700" }, "message": "Work on option parsing as prelude to image loading\n\nChange-Id: I13edbd9b341e603817941beaca676535a7e590c7\n" }, { "commit": "db4d54081f09abcbe97ffdf615874f2809a9e777", "tree": "86d4b6cdbfda6538ebf3fbb2f362209eb6867ce5", "parents": [ "350dad1bb785281d6740633f219a2832feda6ff3" ], "author": { "name": "Brian Carlstrom", "email": "bdc@google.com", "time": "Tue Aug 09 12:18:28 2011 -0700" }, "committer": { "name": "Brian Carlstrom", "email": "bdc@google.com", "time": "Mon Aug 15 10:03:22 2011 -0700" }, "message": "First pass of ImageWriter\n\nChange-Id: I4f189587a2e3cc1c265200b8fa64321b299947eb\n" }, { "commit": "ad107abee00fb523b9fda575a40395e9d6d9510d", "tree": "9db5ff4af4613224ed136bedce10b88538933bff", "parents": [ "650a07d237f219ee6f23cfe939118e6f9c48f4d3" ], "author": { "name": "Carl Shapiro", "email": "cshapiro@google.com", "time": "Sun Aug 14 15:54:24 2011 -0700" }, "committer": { "name": "Carl Shapiro", "email": "cshapiro@google.com", "time": "Sun Aug 14 15:54:24 2011 -0700" }, "message": "Remove author attributions.\n\nChange-Id: I58d2983ce41f6b3f3a4e1bbd1d075e794b8257e3\n" }, { "commit": "744ad05e473be49033d38f30aab63891ca4ee5c3", "tree": "6c14472de037544af316af24b24a37851bbced7d", "parents": [ "f88c95285b2d105a504d18736fe477f74c8c5b45" ], "author": { "name": "Carl Shapiro", "email": "cshapiro@google.com", "time": "Sat Aug 06 15:53:36 2011 -0700" }, "committer": { "name": "Carl Shapiro", "email": "cshapiro@google.com", "time": "Sat Aug 06 15:53:36 2011 -0700" }, "message": "Remove stray trailing semicolons.\n\nChange-Id: I151191cc3ffdd9e8d5ea85496e435ee6f3716f56\n" }, { "commit": "7e93b50433cde2a44d99212e8040299bde498546", "tree": "5365abc77a3ada7f50bfc06b5772e03b86ee8f3c", "parents": [ "8860c0e8491a27aa03b55f2417478fb3629611d6" ], "author": { "name": "Brian Carlstrom", "email": "bdc@google.com", "time": "Thu Aug 04 14:16:22 2011 -0700" }, "committer": { "name": "Brian Carlstrom", "email": "bdc@google.com", "time": "Thu Aug 04 22:59:02 2011 -0700" }, "message": "String intern table and support for unordered_map\n\nChange-Id: I22d86d060780552675c5d7f14a98ffde480eac82\n" }, { "commit": "a40f9bc48afe3a9d38be2fa298fece13ed82ba28", "tree": "ee1597be70dc8f064f9c7e86e1eb11670c894129", "parents": [ "3aaddbb438daaaa37419d281d95fbe21e64982a6" ], "author": { "name": "Brian Carlstrom", "email": "bdc@google.com", "time": "Tue Jul 26 21:26:07 2011 -0700" }, "committer": { "name": "Brian Carlstrom", "email": "bdc@google.com", "time": "Tue Jul 26 21:27:40 2011 -0700" }, "message": "Add ObjectArray::CopyOf\n\nAs part of doing this, moved Heap::Alloc* helpers to the corresponding\nclasses in object.h to break heap.h\u003c-\u003eobject.h cycle due to use of\nHeap from ObjectArray::CopyOf (which is templatized and needs to be\ndefined in .h file).\n\nChange-Id: I1870af6e9dc6552820034ead5e20d13ae4da67da\n" }, { "commit": "4a96b60e45fba4a9d4a2e9c8fc849660eacef684", "tree": "4c48587a07fdafd7438d8b2b34bee3d5a8dfbea2", "parents": [ "2ed144c2b49ae1da6c464d7a1be0062870530802" ], "author": { "name": "Brian Carlstrom", "email": "bdc@google.com", "time": "Tue Jul 26 16:40:23 2011 -0700" }, "committer": { "name": "Brian Carlstrom", "email": "bdc@google.com", "time": "Tue Jul 26 18:50:03 2011 -0700" }, "message": "ClassLinker can now FindClass all libcore classes\n\nSummary:\n- added ClassLinkerTest/LibCore test of finding all libcore classes\n- fixed bug in LinkInterfaceMethods appending mirant methods to vtable_\n- fixed bug in LinkVirtualMethods allocating subclass vtable_\n- fixed mmap bug in MarkStack::Init\n- bumped default (and max) heap sizes to handle ClassLinkerTest/LibCore\n- made ObjectArray a templated class\n- heap allocate Class::vtable_\n- \"mm test-art\" and \"mm test-art-target\" added\n- new is_host_ for use by tests\n\nDetails:\n\n Added support for \"mm test-art\" and \"mm test-art-target\" in addition to test-art-host\n\tAndroid.mk\n\n Made ObjectArray a template class for clearer declarations and to\n remove need the need for down_cast.\n\n\tsrc/object.h\n\tsrc/object_test.cc\n\tsrc/class_linker.cc\n\tsrc/class_linker.h\n\tsrc/dex_cache.cc\n\tsrc/dex_cache.h\n\tsrc/dex_cache_test.cc\n\tsrc/mark_sweep.cc\n\n Made Class::vtable_ a heap allocated ObjectArray\u003cMethod\u003e\n\n\tsrc/class_linker.cc\n\n Fix bug in ClassLinker::LinkInterfaceMethods where we were not\n extending the vtable_ before appending miranda methods.\n\n\tsrc/class_linker.cc\n\n Changed all uses of DexFile* in ClassLinker to be const\n\n\tsrc/class_linker.cc\n\tsrc/class_linker.h\n\n Fix bug in ClassLinker::LinkVirtualMethods where we were using\n NumVirtualMethods vs vtable-\u003eGetLength when determining new max\n vtable_ length.\n\n\tsrc/class_linker.cc\n\n New ClassLinkerTest/LibCore that enumerates the libcore dex file\n and tries to FindClass every descriptor found.\n\n\tsrc/class_linker_test.cc\n\n Added if_host_ check for host only tests. In the future will use\n for picking proper location of files, etc. on host vs target.\n\n\tsrc/common_test.h\n\n Fixed test to use ClassLinker::AllocDexCache\n\n\tsrc/dex_cache_test.cc\n\n Fixed fooIds comments to foo_ids_\n\n\tsrc/dex_file.h\n\n Bumped default heap size (and max as well) to make ClassLinkerTest/LibCore run\n\n\tsrc/heap.h\n\n Fixed bug where we were not using MAP_ANONYMOUS for MarkStack\n allocation found running ClassLinkerTest/LibCore.\n\n\tsrc/mark_stack.cc\n\nChange-Id: I204e2ec7205210e2b60f5b81d126ab6e1da5a71c\n" }, { "commit": "58551df3c2646ed507feec9e9eb3768085a76059", "tree": "892bc05ab5b2959c64e34cfe7cb541da0967f013", "parents": [ "f615a61aef972cfc1dc23931ac2ed0da14c3fedb" ], "author": { "name": "Carl Shapiro", "email": "cshapiro@google.com", "time": "Sun Jul 24 03:09:51 2011 -0700" }, "committer": { "name": "Carl Shapiro", "email": "cshapiro@google.com", "time": "Sun Jul 24 03:09:51 2011 -0700" }, "message": "Add a collection implementation.\n\nChange-Id: I47330dc14dd2ecee052b6a5165c6d48bd6c89dc4\n" }, { "commit": "578bbdc684db8ed68e9fedbc678669d27fa68b6e", "tree": "d752941d81ad6d7e248c349ce21371a8ae369693", "parents": [ "45a76cb99104a222d6a9bd768a084893dcb7cf30" ], "author": { "name": "Brian Carlstrom", "email": "bdc@google.com", "time": "Thu Jul 21 14:07:47 2011 -0700" }, "committer": { "name": "Brian Carlstrom", "email": "bdc@google.com", "time": "Fri Jul 22 15:23:33 2011 -0700" }, "message": "Move DexFiles to be allocated on the heap\n\nRemoved pointer from DexFile to RawDexfile to allow heap allocation of\nDexFile. DexFile is now an ObjectArray of ObjectArrays. Removing the\npointer from DexFile to RawDexfile meant adding arguments to pass the\nRawDexfile along to ClassLinker::LoadClass, LoadInterfaces, LoadField,\nLoadMethod, LinkClass, LinkInterfaces, ResolveClass to avoid the need\nto look a RawDexfile up from a DexFile. ClassLinker::CreateArrayClass\nnow takes the raw_dex_file to assist in looking up the component class\nfrom the proper source. Moved Method::HasSame* methods to ClassLinker\nsince for access to the RawDexfile (and tests of HasSame from\nObjectTest to ClassLinkerTest.\n\n\tsrc/dex_file.cc\n\tsrc/dex_file.h\n\tsrc/class_linker.cc\n\tsrc/class_linker.h\n\tsrc/object.h\n\nRunTime::Create/RunTime::Init/ClassLinker::Create/ClassLinker::Init\nnow take the boot class path. A boot class path with\nLjava/lang/Object; must be provided to initalize [Ljava/lang/Object;\nduring startup in order to heap allocate DexFiles.\n\n\tsrc/class_linker.cc\n\tsrc/class_linker.h\n\tsrc/runtime.cc\n\tsrc/runtime.h\n\nRestored FindClass to load from a specific dex file. This is for use\nby class loaders, otherwise it is assumed the caller wants to use the\nboot classpath. We now distinguish and track the boot classpath as\nseparate from other dex files known to the class linker. Renamed\nAppendToClassPath to AppendToBootClassPath and FindInClassPath to\nFindInBootClassPath to clarify.\n\n\tsrc/class_linker.cc\n\tsrc/class_linker.h\n\nCleaned up AllocCharArray to generic AllocArray and added\nAllocObjectArray. Added a functional ObjectArray implementation and a\nObjectTest to cover it.\n\n\tsrc/heap.h\n\tsrc/object.h\n\tsrc/object_test.cc\n\nSwitched more Object* downcasting to down_cast\n\n\tsrc/class_linker.cc\n\nSwitched order of arguments for DexFile::SetResolved to follow more\nconventional collection ordering ( \"(index, value)\" vs \"(value, index)\" )\n\n\tsrc/dex_file.h\n\tsrc/class_linker.cc\n\tsrc/class_linker.h\n\nAdded class_linker_ and java_lang_raw_dex_file_ to RuntimeTest as\nconvenience to test subclasses. ClassLinkerTest now can use these to\nsimplify its ::Assert* methods. JniCompilerTest now uses it for\nsetting up its boot class path. Removed now unneeded\nOpenDexFileBase64.\n\n\tsrc/common_test.h\n\tsrc/class_linker_test.cc\n\tsrc/jni_compiler_test.cc\n\nAdd external/gtest/include to non-test include path so FRIEND_TEST can be used.\nAdd src to include path to remove the need of using in in art/src files.\n\n\tbuild/Android.libart.host.mk\n\tbuild/Android.libart.mk\n\tsrc/assembler.cc\n\tsrc/assembler.h\n\tsrc/assembler_arm.cc\n\tsrc/assembler_arm.h\n\tsrc/assembler_x86.cc\n\tsrc/assembler_x86.h\n\tsrc/assembler_x86_test.cc\n\tsrc/base64.cc\n\tsrc/base64.h\n\tsrc/calling_convention.cc\n\tsrc/calling_convention.h\n\tsrc/calling_convention_arm.cc\n\tsrc/calling_convention_x86.cc\n\tsrc/casts.h\n\tsrc/class_linker.h\n\tsrc/class_linker_test.cc\n\tsrc/common_test.h\n\tsrc/constants.h\n\tsrc/constants_arm.h\n\tsrc/constants_x86.h\n\tsrc/dex_file.cc\n\tsrc/dex_file.h\n\tsrc/dex_file_test.cc\n\tsrc/dex_instruction.cc\n\tsrc/dex_instruction.h\n\tsrc/dex_instruction_visitor.h\n\tsrc/dex_instruction_visitor_test.cc\n\tsrc/dex_verifier.cc\n\tsrc/dex_verifier.h\n\tsrc/heap.cc\n\tsrc/heap.h\n\tsrc/jni_compiler.cc\n\tsrc/jni_compiler_test.cc\n\tsrc/jni_internal.cc\n\tsrc/jni_internal.h\n\tsrc/leb128.h\n\tsrc/managed_register.h\n\tsrc/managed_register_arm.cc\n\tsrc/managed_register_arm.h\n\tsrc/managed_register_arm_test.cc\n\tsrc/managed_register_x86.cc\n\tsrc/managed_register_x86.h\n\tsrc/managed_register_x86_test.cc\n\tsrc/mark_stack.cc\n\tsrc/mark_stack.h\n\tsrc/mark_sweep.cc\n\tsrc/mark_sweep.h\n\tsrc/memory_region.cc\n\tsrc/memory_region.h\n\tsrc/monitor.h\n\tsrc/object.cc\n\tsrc/object.h\n\tsrc/object_bitmap.cc\n\tsrc/object_bitmap.h\n\tsrc/object_test.cc\n\tsrc/offsets.cc\n\tsrc/offsets.h\n\tsrc/raw_dex_file.cc\n\tsrc/raw_dex_file.h\n\tsrc/raw_dex_file_test.cc\n\tsrc/runtime.cc\n\tsrc/runtime.h\n\tsrc/scoped_ptr.h\n\tsrc/space.cc\n\tsrc/space.h\n\tsrc/space_test.cc\n\tsrc/stringpiece.cc\n\tsrc/thread.cc\n\tsrc/thread.h\n\tsrc/thread_arm.cc\n\tsrc/thread_x86.cc\n\tsrc/utils.h\n\nChange-Id: Ib633cea878c36921e9037b0464cb903aec318c3e\n" }, { "commit": "69759eaa6fd4386f1e6d8748052ad221087b3476", "tree": "9f37e49625a6e528405ed28ae8f5ac515ffedd67", "parents": [ "dbd05642d4157b65d837da30c1d8df0ef577a378" ], "author": { "name": "Carl Shapiro", "email": "cshapiro@google.com", "time": "Thu Jul 21 18:13:35 2011 -0700" }, "committer": { "name": "Carl Shapiro", "email": "cshapiro@google.com", "time": "Thu Jul 21 18:13:35 2011 -0700" }, "message": "Add allocation and garbage collection infrastructure.\n\nChange-Id: I4b04cdfdf18afb75a7b0df87b509e8156b4a932b\n" }, { "commit": "a080803c2f0af91cc0de18919a762f4fcd5cce29", "tree": "9089f2704cf5eee7020142c87c35d960835a5c7e", "parents": [ "a7f4f4888502aaa5c137cb17c9b482053285aebf" ], "author": { "name": "Brian Carlstrom", "email": "bdc@google.com", "time": "Mon Jul 18 00:39:23 2011 -0700" }, "committer": { "name": "Brian Carlstrom", "email": "bdc@google.com", "time": "Mon Jul 18 12:11:08 2011 -0700" }, "message": "Setting klass_ fields of heap allocated objects\n\nChange-Id: Ib1ab43254430497cce0660156394eebddd2cc445\n" } ], "next": "a7f4f4888502aaa5c137cb17c9b482053285aebf" }