)]}' { "log": [ { "commit": "9bf7817dd29d15ea49c88436db4067d87fc7e6c4", "tree": "83df3be7fc9aa5b879d22b9db6361baf7c83c17f", "parents": [ "ba198d508497ffeb146d61c5ec5fb66d84e864f3" ], "author": { "name": "Christopher Ferris", "email": "cferris@google.com", "time": "Wed May 20 15:37:30 2020 -0700" }, "committer": { "name": "Christopher Ferris", "email": "cferris@google.com", "time": "Wed May 20 18:18:54 2020 -0700" }, "message": "Fix deadlock/timeout in thread unwinding.\n\nWhen malloc debug is enabled, using libbacktrace to unwind can\nresult in a deadlock. This happens when an unwind of a thread\nis occuring which triggers a signal to be sent to that thread. If\nthat thread is interrupted while a malloc debug function is\nexecuting and owns a lock, that thread is then stuck in the signal\nhandler. Then the original unwinding thread attempts to do an\nallocation and gets stuck waiting for the same malloc debug lock.\n\nThis is not a complete deadlock since the unwinder has timeouts,\nbut it results in truncated unwinds that take at least five\nseconds to complete.\n\nOnly the backtrace signals needs to be blocked because it is the only\nknown signal that will result in a thread being paused in a signal\nhandler.\n\nAlso, added a named signal in the reserved signal list for the\nspecial bionic backtrace signal.\n\nBug: 150833265\n\nTest: New unit tests pass with fix, fail without fix.\nChange-Id: If3e41f092ebd40ce62a59ef51d636a91bc31ed80\n" }, { "commit": "3b21ada5f1bd98b337143ee5d5baba7373bfdbb1", "tree": "7e5bf3c1ae9008eccc61f7f6eefef9b0039af1b5", "parents": [ "73e9f242ed07631798433df4e35f5f73443cb9f9" ], "author": { "name": "Mitch Phillips", "email": "mitchp@google.com", "time": "Tue Jan 07 15:47:47 2020 -0800" }, "committer": { "name": "Evgenii Stepanov", "email": "eugenis@google.com", "time": "Thu Jan 23 10:37:24 2020 -0800" }, "message": "Add tagged pointers to bionic.\n\nThis patch introduces tagged pointers to bionic. We add a static tag to\nall pointers on arm64 compatible platforms (needs requisite\ntop-byte-ignore hardware feature and relevant kernel patches).\n\nWe dynamically detect TBI-compatible devices (a device with the TBI feature and\nkernel support) at process start time, and insert an implementation-dependent\ntag into the top byte of the pointer for all heap allocations. We then check\nthat the tag has not been truncated when deallocating the memory.\n\nIf an application incorrectly writes to the top byte of the pointer, we\nterminate the process at time of detection. This will allow MTE-incompatible\napplications to be caught early.\n\nBug: 135754954\nBug: 147147490\n\nTest: cd bionic \u0026\u0026 atest .\nChange-Id: Ie424325ba1e3c4443040ac265aeaa28d9e405d28\n" }, { "commit": "ce4c42e09fb9039df7c6a5b7306119279a694897", "tree": "13e38338ea7bc859b295b923be6a7acd73edbebc", "parents": [ "43d5f9d4dd83d15a859d9be1359c4a4a47381fea" ], "author": { "name": "Nicolas Geoffray", "email": "ngeoffray@google.com", "time": "Tue Jan 21 13:56:21 2020 +0000" }, "committer": { "name": "Nicolas Geoffray", "email": "ngeoffray@google.com", "time": "Tue Jan 21 13:57:26 2020 +0000" }, "message": "Revert \"Add tagged pointers to bionic.\"\n\nThis reverts commit 43d5f9d4dd83d15a859d9be1359c4a4a47381fea.\n\nBug: 135754954\nBug: 147147490\n\nExempt-From-Owner-Approval: clean revert\n\nReason for revert: Breaks ART gtest, see:\nhttps://ci.chromium.org/p/art/builders/ci/angler-armv8-non-gen-cc/561\n\nThe crash happens on mprotect of a page, the test crashes with ENOMEM.\n\nChange-Id: I52eea1abbfaf8d8e2226f92d30aa55aba3810528\n" }, { "commit": "43d5f9d4dd83d15a859d9be1359c4a4a47381fea", "tree": "95339c5cb5d055b8fd2b216b46e6264180c4fc36", "parents": [ "d2b8149ddad8b21ce535fb02ff752c85c7aa9d43" ], "author": { "name": "Mitch Phillips", "email": "mitchp@google.com", "time": "Tue Jan 07 15:47:47 2020 -0800" }, "committer": { "name": "Mitch Phillips", "email": "mitchp@google.com", "time": "Wed Jan 15 15:03:00 2020 -0800" }, "message": "Add tagged pointers to bionic.\n\nThis patch introduces tagged pointers to bionic. We add a static tag to\nall pointers on arm64 compatible platforms (needs requisite\ntop-byte-ignore hardware feature and relevant kernel patches).\n\nWe dynamically detect TBI-compatible devices (a device with the TBI feature and\nkernel support) at process start time, and insert an implementation-dependent\ntag into the top byte of the pointer for all heap allocations. We then check\nthat the tag has not been truncated when deallocating the memory.\n\nIf an application incorrectly writes to the top byte of the pointer, we\nterminate the process at time of detection. This will allow MTE-incompatible\napplications to be caught early.\n\nBug: 135754954\nBug: 147147490\n\nTest: cd bionic \u0026\u0026 atest .\nChange-Id: I6e5b809fc81f55dd517f845eaf20f3c0ebd4d86e\n" }, { "commit": "6f517cd7a1142191fde2201b6c529758e7ff6895", "tree": "49105414620136e4a9bf2c613a3f8c29a5c00062", "parents": [ "ca1834de7a8180ee8db0a629fa950387c9da49b1" ], "author": { "name": "Christopher Ferris", "email": "cferris@google.com", "time": "Fri Nov 08 11:28:38 2019 -0800" }, "committer": { "name": "Christopher Ferris", "email": "cferris@google.com", "time": "Fri Nov 08 11:49:14 2019 -0800" }, "message": "Rename iterate to malloc_iterate internally.\n\nI have no idea why I used the iterate name internally which is\ncompletely unlike every other function name. Change this to match\neveryone else so that it\u0027s now malloc_iterate everywhere.\n\nThis is probably the last chance to change this before mainline\nmodules begin, so make everything consistent.\n\nTest: Compiles, unit tests passes.\nChange-Id: I56d293377fa0fe1a3dc3dd85d6432f877cc2003c\n" }, { "commit": "ff88fb0d3adbc67a4f94f5ef7e2b5bcc7a96c8f3", "tree": "1a78e188d9e51383579f697cb5532846b841cd69", "parents": [ "1994f28be2c0faf2b70b1ca8fff7d8d2fa68d922" ], "author": { "name": "Christopher Ferris", "email": "cferris@google.com", "time": "Mon Nov 04 18:40:00 2019 -0800" }, "committer": { "name": "Christopher Ferris", "email": "cferris@google.com", "time": "Wed Nov 06 10:42:42 2019 -0800" }, "message": "Fix allocations escaping malloc debug.\n\nWhen using a FILE object for some malloc debug functions, calling\nfprintf will trigger an allocation to be put in the object. The problem\nis that these allocations were not allocated by the malloc debug\nwrapper and they get freed during the fclose as if they are malloc\ndebug allocation. In most cases, the code will detect the bad pointer\nand leak the memory, but it might also cause a crash.\n\nThe fix is to avoid using fprintf so that no allocations are made\nin the object that survive and need to be freed in the fclose call.\n\nChange the MallocXmlElem.h to use a file decsriptor not a FILE object.\n\nAdd new unit and system tests to detect this case.\n\nBug: 143742907\n\nTest: Ran unit and system tests.\nTest: Ran bionic unit tests.\nChange-Id: I524392de822a29483aa5be8f14c680e70033eba2\n" }, { "commit": "705de3c6398d53b780d731008bafc3c2dafbbf85", "tree": "1118ec9cf1ce812c4d02d482a3b0f1a311036e32", "parents": [ "d277334059f5c0c08ea312cad84f003c0d9f3519" ], "author": { "name": "Christopher Ferris", "email": "cferris@google.com", "time": "Wed May 22 13:39:57 2019 -0700" }, "committer": { "name": "Christopher Ferris", "email": "cferris@google.com", "time": "Wed May 22 17:16:37 2019 -0700" }, "message": "Disable malloc debug when asan enabled.\n\nBug: 123312263\n\nTest: Verified with asan enabled, malloc debug does not initialize.\nTest: Ran tests on non-asan build and verify they pass.\nChange-Id: I3c37c170bf6c1de42740972f2113ae991351d931\n" }, { "commit": "d269fcc935b276502b9e47a575d76693fe1b8455", "tree": "cb9e09297fb247d3d8511565e25a930336b6bc72", "parents": [ "9433a0299e2493303776713072f0744d18c705dc" ], "author": { "name": "Christopher Ferris", "email": "cferris@google.com", "time": "Mon May 06 19:03:59 2019 -0700" }, "committer": { "name": "Christopher Ferris", "email": "cferris@google.com", "time": "Tue May 07 13:16:58 2019 -0700" }, "message": "Avoid using malloc debug code after exit.\n\nI wrote a new unit test that would fail on the old version of the\ncode.\n\nOn a walleye big cpu, this costs about 40ns-50ns (going from ~430ns to ~480ns).\nI think this is an acceptable performance degradation.\n\nBug: 131867816\n\nTest: New unit tests pass.\nChange-Id: I4c0f4373fb0694bf29c3824dbb1224a8a17e211e\n" }, { "commit": "8189e77bbb47dbaa1a09d565c62c2fed03e358c3", "tree": "2c8fc5d8fd1ec1ab881d35519670104ad86283e3", "parents": [ "b9a7c651f178323e27752406f89d741cfb143e61" ], "author": { "name": "Christopher Ferris", "email": "cferris@google.com", "time": "Tue Apr 09 16:37:23 2019 -0700" }, "committer": { "name": "Christopher Ferris", "email": "cferris@google.com", "time": "Tue Apr 16 11:22:06 2019 -0700" }, "message": "Remove gMallocLeakZygoteChild.\n\nRemove this global variable and change the setting of it to non-zero\nto a call to android_mallopt.\n\nIn addition, change the initialize function to use pass a bool* instead of\nint*.\n\nBug: 130028357\n\nTest: Ran malloc_debug/malloc_hooks/perfetto tests.\nChange-Id: I20d382bdeaaf38aac6b9dcabea5b3dfab3c945f6\nMerged-In: I20d382bdeaaf38aac6b9dcabea5b3dfab3c945f6\n(cherry picked from commit 5225b342f0810c027df3d09fbbcef4d324b19b93)\n" }, { "commit": "c328e4465d12d760f5fbcb2daddf5fa4138f50b1", "tree": "fd91c386332bbe1198cb0bedc11505b76d146057", "parents": [ "4699cf61746d37fc11c8eb8bdfc7eb296a476e4b" ], "author": { "name": "Christopher Ferris", "email": "cferris@google.com", "time": "Mon Apr 01 19:31:26 2019 -0700" }, "committer": { "name": "Christopher Ferris", "email": "cferris@google.com", "time": "Tue Apr 02 10:55:21 2019 -0700" }, "message": "Disable info messages by default for malloc debug.\n\nAdd a new option verbose for malloc debug that is not enabled by default.\nThis disables all of the info log messages. It turns out these log\nmessages can add a measurable amount of time and can change the boot up.\n\nBug: 129239269\n\nTest: Adjusted unit tests pass.\nTest: Verified no messages unless verbose option used.\nChange-Id: I805cb7c8ecb44de88119574e59d784877cacc383\n" }, { "commit": "6c619a0da3f96a26d91c1db48fd3e3be156aabe5", "tree": "aa0802d01d19ed0891e6b77a732cc72ed5617dcc", "parents": [ "4e167f35d602f9dcede3b74ff3e09e88f3edb398" ], "author": { "name": "Christopher Ferris", "email": "cferris@google.com", "time": "Fri Mar 01 17:59:51 2019 -0800" }, "committer": { "name": "Christopher Ferris", "email": "cferris@google.com", "time": "Thu Mar 07 08:39:55 2019 -0800" }, "message": "Refactor the malloc_info code.\n\nmalloc_info needs to be per native allocator, but the code treated it\nlike a global function that doesn\u0027t depend on the native memory allocator.\n\nUpdate malloc debug to dump the actual pointers that it has been tracking.\n\nTest: bionic-unit-tests pass.\nTest: malloc debug tests pass.\nTest: malloc hook tests pass.\nChange-Id: I3b0d4d748489dd84c16d16933479dc8b8d79013e\nMerged-In: I3b0d4d748489dd84c16d16933479dc8b8d79013e\n(cherry picked from commit a3656a98b10d2a4a6194a5d9705ad9c2cc5877b0)\n" }, { "commit": "a22f5d5175df5c42ec86d2c2db250edf1f64084c", "tree": "66d2980a1b04626c7e4fc1437e96194b958ea8f8", "parents": [ "0771b752f1b955e7ded6ccbbf825b0ae93439eb5" ], "author": { "name": "Christopher Ferris", "email": "cferris@google.com", "time": "Fri Mar 01 16:40:59 2019 -0800" }, "committer": { "name": "Christopher Ferris", "email": "cferris@google.com", "time": "Fri Mar 01 23:56:23 2019 -0800" }, "message": "Make aligned_alloc match the standard.\n\nJemalloc does not verify that the size parameter is a multiple of\nalignment. Fix this since it only went into P.\n\nFix the unit tests, and fix malloc debug/malloc hooks to handle this\nnew restrictive behavior.\n\nBug: 126944692\n\nTest: Ran bionic unit tests.\nTest: Ran bionic unit tests with malloc hooks enabled (no new tests fail).\nTest: Ran bionic unit tests with malloc debug enabled (no new tests fail).\nTest: Ran malloc debug unit tests.\nChange-Id: I4d50785928815679c781ca729f998454d76b9192\n" }, { "commit": "7f209a979c55a58e91946a2efd5f0b339ffe309c", "tree": "481e04cf25531a9ced30b032bb5ba93f06de674e", "parents": [ "822326db922ac5d0e4dea8cff1d774e8f04db94a" ], "author": { "name": "Iris Chang", "email": "iris.chang@mediatek.com", "time": "Wed Jan 16 11:17:15 2019 +0800" }, "committer": { "name": "Christopher Ferris", "email": "cferris@google.com", "time": "Tue Jan 22 15:54:36 2019 -0800" }, "message": "Bionic malloc debug: add a new option \"abort_on_error\"\n\nThis new option causes an abort after malloc debug detects an error.\nThis allows vendors to get process coredumps to analyze memory for\ncorruption.\n\nBug: 123009873\nTest: New test cases added for unit tests and config tests.\n\nChange-Id: I6b480af7f747d6a82f61e8bf3df204a5f7ba017f\n" }, { "commit": "32bc0fcf69dfccb3726fe572833a38b01179580e", "tree": "abca656f500087493e0997d5cf38cbcc9ec98eb8", "parents": [ "65f82092a17518080178ff7004cc6db362ebfbcd" ], "author": { "name": "Yi Kong", "email": "yikong@google.com", "time": "Thu Aug 02 17:31:13 2018 -0700" }, "committer": { "name": "Yi Kong", "email": "yikong@google.com", "time": "Thu Aug 02 18:09:44 2018 -0700" }, "message": "Modernize codebase by replacing NULL with nullptr\n\nFixes -Wzero-as-null-pointer-constant warning.\n\nTest: m\nBug: 68236239\nChange-Id: I5b4123bc6709641315120a191e36cc57541349b2\n" }, { "commit": "97b474710225bc4853ab314c3d9ca812dc375525", "tree": "dca19aa52db541c9ef7832d41d1630e81159b12a", "parents": [ "06f217d399b1bf6345e4c71b014cdaac3d1c183a" ], "author": { "name": "Christopher Ferris", "email": "cferris@google.com", "time": "Tue Jul 10 14:45:24 2018 -0700" }, "committer": { "name": "Christopher Ferris", "email": "cferris@google.com", "time": "Wed Jul 11 15:23:00 2018 -0700" }, "message": "Disable malloc debug intercepts when exiting.\n\nThere is a hang when enabling leak_track since the dumping of the\nleak data can wind up doing an allocation.\n\nAdd new system unit test to make sure this doesn\u0027t happen again.\n\nBug: 111146059\n\nTest: Test program that leaks does not hang forever.\nTest: Unit tests pass.\nChange-Id: Icf99be58ba5db98ee124a471b957a086045f5870\n" }, { "commit": "2e1a40a2033088e7f2e78424a8b163d3fcf84b1b", "tree": "51ec5d1703f6225d6dc58ad5b01f9c728e4f2883", "parents": [ "2583b0adb201233313bbe02f665f5c623f2c5a30" ], "author": { "name": "Christopher Ferris", "email": "cferris@google.com", "time": "Wed Jun 13 10:46:34 2018 -0700" }, "committer": { "name": "Christopher Ferris", "email": "cferris@google.com", "time": "Fri Jun 15 12:29:13 2018 -0700" }, "message": "Change heap dump format slightly.\n\nBump the version from v1.1 to v1.2 and add a build fingerprint line.\n\nUpdate the heap dump documentation to match the new format and reflect\nwhat made it in P and what made it in Q.\n\nUpdate the unit tests for this change.\n\nAdd -O0 to unit test code to make it easier to debug.\n\nAdd an external function that can be used by the framework code\nso that there is only one way to dump the heap.\n\nBug: 110095681\n\nTest: Ran unit tests.\nTest: Did a dump of a real process and verified fingerprint.\nTest: Did a dump of a process without malloc debug enabled.\nChange-Id: I769a476cbeaf4c85c5d75bd6d6385f0e3add948c\nMerged-In: I769a476cbeaf4c85c5d75bd6d6385f0e3add948c\n(cherry picked from commit c84a2a2601a4112ca6e43a33defb989c1da8c2f4)\n" }, { "commit": "93bdd6ae3ad8322b7be2be0201c8db7227631d14", "tree": "42ee959b96c2c2db42f0569554680dbe22ce39ef", "parents": [ "4c5c45346fcc6f066a89bfc455f287fe2f4e3e41" ], "author": { "name": "Christopher Ferris", "email": "cferris@google.com", "time": "Thu Apr 05 11:12:38 2018 -0700" }, "committer": { "name": "Christopher Ferris", "email": "cferris@google.com", "time": "Thu May 24 08:44:53 2018 -0700" }, "message": "Add support for using the new unwinder.\n\nThis adds a new option backtrace_full, when it is set, then it will use\nlibunwindstack.\n\nModify the dump to file data to dump the extra information from libunwindstack.\nAlong with the new dump file format, change the version to v1.1.\nUpdated document for new format of file data.\n\nAdd unit tests for the new functionality.\n\nBug: 74361929\n\nTest: Ran unit tests.\nChange-Id: I40fff795f5346bba7b9d7fde2e04f269ff4eb7f1\n" }, { "commit": "4da2503d70dc4bc1444454876e3794b69227d90d", "tree": "d904c750ebac96aee76df44baad1320c0db211a3", "parents": [ "a3f6f6c1b9135c90e410f5382b153db9a43a4db0" ], "author": { "name": "Christopher Ferris", "email": "cferris@google.com", "time": "Wed Mar 07 13:38:48 2018 -0800" }, "committer": { "name": "Christopher Ferris", "email": "cferris@google.com", "time": "Mon Apr 02 18:59:23 2018 -0700" }, "message": "Refactor malloc debug.\n\nChanges\n- Refactor the code so that only guards require creating a special header\n for every pointer allocated.\n- Store only a single copy of every backtrace. This saves memory so that\n turning on the backtrace option doesn\u0027t result in 10X memory usage.\n- Added new option track_allocs that only verifies pointers are valid for\n free/malloc_usable_size/realloc.\n- Remove suffix from test names.\n- Add the TRACK_ALLOCS options to all guard options.\n- Add new option verify_pointers that is a lightweight way to verify\n pointers that are passed to allocation routines.\n- Do auto-formatting of the code.\n- Updated documentation for all of these changes.\n\nBug: 74361929\n\nTest: Ran unit tests.\nTest: Ran libmemunreachable unit tests.\nTest: Ran an app with backtrace enabled.\n\nChange-Id: I3246c48ae4f9811f64622d90d0a9b4d9d818702c\n" }, { "commit": "cae21a9b53a10f0cba79bf6783c4a5af16228fed", "tree": "c60c6622d2b80f9ad4ae5a990b7fb3439f16444b", "parents": [ "be6fbae74fc3048647caadd58a9eaad81245e934" ], "author": { "name": "Christopher Ferris", "email": "cferris@google.com", "time": "Mon Feb 05 18:14:55 2018 -0800" }, "committer": { "name": "Christopher Ferris", "email": "cferris@google.com", "time": "Wed Feb 07 06:57:14 2018 -0800" }, "message": "Add aligned_alloc to libc.\n\nBug: 72969374\n\nTest: Bionic unit tests pass.\nTest: Malloc debug unit tests pass.\nChange-Id: I235985bbc638855d94249c97c98f14ab2924bda0\n(cherry picked from commit d69ee59594088c0d92ba9273188ef53ea5e6cd6a)\n" }, { "commit": "a613d0df5c682617380728a7d151483de05a4f92", "tree": "5a73110db954e5fc92116ad0930ed83495607729", "parents": [ "3e8584d4a7d617289e8a954711ba895415d781c4" ], "author": { "name": "Dan Albert", "email": "danalbert@google.com", "time": "Thu Oct 05 16:39:33 2017 -0700" }, "committer": { "name": "Dan Albert", "email": "danalbert@google.com", "time": "Thu Oct 05 23:41:47 2017 -0700" }, "message": "Add a legacy inline for mmap64.\n\nWhile this was never an inline, this function alone has caused most of\nthe bug reports related to _FILE_OFFSET_BITS\u003d64. Providing an inline\nfor it should allow a lot more code to build with _FILE_OFFSET_BITS\u003d64\nwhen targeting pre-L.\n\nTest: make checkbuild\nTest: built trivial cc_binary for LP32 against API 14 with\n _FILE_OFFSET_BITS\u003d64 set\nBug: lots\nChange-Id: I8479d34af4da358c11423bee43d45b59e9d4143e\n" }, { "commit": "602b88ccddfafdefec023e7599d49c811ad3214f", "tree": "11730f3f0cbc308860c4e45d6f0310c7f2083041", "parents": [ "e0e2798f11ca082fe0a796f167f9006c51dc3304" ], "author": { "name": "Christopher Ferris", "email": "cferris@google.com", "time": "Fri Aug 04 13:04:04 2017 -0700" }, "committer": { "name": "Christopher Ferris", "email": "cferris@google.com", "time": "Tue Sep 05 15:57:00 2017 -0700" }, "message": "Provide method to dump backtrace heap data.\n\nFor non-zygote spawned processes, we might want to dump the backtrace\ndata. Provide a method to send a signal to a process and then dump the\ndata to a file.\n\nAdds a method to dump the backtrace data on exit.\n\nUpdate documentation and explain format of heap dump data.\n\nTest: Ran unit tests, enabled new options and used them.\nChange-Id: Ie2fa706694160731afe02c1382b037d06df1d069\n" }, { "commit": "a1c0d2fd4ce96e123c4ae6506c9d637d747e1fe2", "tree": "b1f89edbd67d6b09b9cbdfb1bf03b65b2dac9ad4", "parents": [ "ac58d33d343eba6a3370f541467f09802c48b24b" ], "author": { "name": "Christopher Ferris", "email": "cferris@google.com", "time": "Mon May 15 15:50:19 2017 -0700" }, "committer": { "name": "Christopher Ferris", "email": "cferris@google.com", "time": "Mon May 15 16:59:16 2017 -0700" }, "message": "Add support for modifying decay timer.\n\nAdd the mallopt function, and only a single option so far.\n\nBug: 36401135\n\nTest: Built and booted bullhead.\nTest: Ran jemalloc unit tests.\nTest: Ran bionic unit tests.\nTest: Ran a test that allocated and free\u0027d a large piece of memory,\nTest: and verified that after changing the parameter, the PSS\nTest: sticks around (decay timer set to 1), the PSS is purged (decay\nTest: timer set to 0).\nChange-Id: I6927929b0c539c1023d34772d9e26bb6a8a45877\n" }, { "commit": "2b2b25b87827102671cdd4b25c01aa22a9971a63", "tree": "d318825c7548379a307beee5b78580e51ce97a52", "parents": [ "e06c69d07324f3da67aa063202e698dedd831365" ], "author": { "name": "Christopher Ferris", "email": "cferris@google.com", "time": "Wed Apr 05 19:13:03 2017 -0700" }, "committer": { "name": "Christopher Ferris", "email": "cferris@google.com", "time": "Thu Apr 06 15:30:42 2017 -0700" }, "message": "Refactor Config from a struct to a class.\n\nThis should make it easier to add new options, and to add options that\nare complex. For example, I want to modify the behavior of\nrecord_allocs_file so that it also enables record_allocs to a default\nstate.\n\nTest: All unit tests pass.\nTest: Enable the backtrace option and restart.\nChange-Id: Idf5cdeed06ade3bc2c8ae39d228734bf65209b4f\n" }, { "commit": "ac81fe8657119c265edcc2f05b3e5b7c5b17ae9f", "tree": "99a6e825926829bb7c8d955cbc96ae1777691ef9", "parents": [ "8945d5e43a25737f3f888a90742cc918b1124a05" ], "author": { "name": "Tamas Berghammer", "email": "tberghammer@google.com", "time": "Fri Aug 26 15:54:59 2016 +0100" }, "committer": { "name": "Christopher Ferris", "email": "cferris@google.com", "time": "Thu Nov 17 13:13:27 2016 -0800" }, "message": "Enable malloc debug using environment variables\n\nPreviously malloc debug can be enabled only using global settings\naccessible to the root user only. This CL adds a new option to enable\nit using environment variables making it possible to use it with pure\nnative (shell) applications on production builds (from shell user) and\nprepares it for using it from logwrapper on production devices.\n\nRemove the old environment variable and property since they are not\nnecessary.\n\nTest: Enable malloc debug using environment variable and verify\nTest: that it only affects the commands launched from the shell.\nTest: Enable malloc debug using the property variable and verify\nTest: that it affects all commands.\nTest: Run all unit tests in 32 bit and 64 bit.\nChange-Id: Iecb75a3471552f619f196ad550c5f41fcd9ce8e5\n" }, { "commit": "7bd01783a830f72c1245c262a7fd9a199e90aed1", "tree": "5bd657e6c452d5f9d81a6c9b156ee748c8a8f070", "parents": [ "a0e205e3b4f7993affb864260e924ce9f20573d0" ], "author": { "name": "Christopher Ferris", "email": "cferris@google.com", "time": "Wed Apr 20 12:30:58 2016 -0700" }, "committer": { "name": "Christopher Ferris", "email": "cferris@google.com", "time": "Fri Jul 08 14:05:17 2016 -0700" }, "message": "Add the record alloc option.\n\nThis option adds the ability to record all of the allocation requests\nand dump them to a file when a signal is sent to the process.\n\nIncluded in this change, redo the option processing to add a new\nstring option.\n\nBug: 27747898\n\nChange-Id: Ida043362e38b5eb1d459c99db9c2581015dab366\n" }, { "commit": "55a89a48f41da349735627c2ae8985895955b1ca", "tree": "c946bd9b7e1353787f278e5face49cae5d8f238d", "parents": [ "03a16d3db7c5fff733d28250212821ec2ff5eccd" ], "author": { "name": "Christopher Ferris", "email": "cferris@google.com", "time": "Thu Apr 07 17:14:53 2016 -0700" }, "committer": { "name": "Christopher Ferris", "email": "cferris@google.com", "time": "Tue Apr 12 15:36:53 2016 -0700" }, "message": "Small refactor.\n\n- Move all ScopedDisableDebugCalls into the debug_XXX calls. This avoids\nany issues that might arise where every part of the code needs to properly\nguard anything that might allocate. Instead everything is already guarded.\n- Add a pointer to debug_data in all of the XXData classes. This avoids\ncalling individual functions passing in the debug_data pointer.\n- Flip the NO_HEADER_OPTIONS to an explicit HEADER_OPTIONS list since fewer\noptions actually require a header.\n- Move the extern of g_debug to the DebugData.h header.\n\nChange-Id: Ia213a391b4a44d9ce122a709d09fe4f1b5426f36\n" }, { "commit": "d0919623a2ef56107590eca9a9522a250fb8bd4a", "tree": "b0e66ab45ca64f3fb90e03205ebd814f5c55c4f0", "parents": [ "8d0af0bf8004c65f13b985643004a915d7e382eb" ], "author": { "name": "Christopher Ferris", "email": "cferris@google.com", "time": "Tue Mar 15 22:39:39 2016 -0700" }, "committer": { "name": "Christopher Ferris", "email": "cferris@google.com", "time": "Wed Mar 16 17:38:08 2016 -0700" }, "message": "Fix race in malloc debug option free_track.\n\nThe free track mechanism could fail if, at the same time a free occurs,\nanother thread is trying to free and verify the same allocation. This\ndoesn\u0027t work if the freed allocation is added to the list and we still\ndo work on it. The fix is to only add to the free list when we are done\nwith the allocation.\n\nAlso fix a problem where the usable size is computed incorrectly because\ntwo of the arguments where reversed.\n\nIn addition, add a check that the allocation being verified has the correct\ntag before trying to check the body of the allocation.\n\nAdd a test to catch the original failure.\n\nAdd a test for the tag being different.\n\nBug: 27601650\nChange-Id: Ie9200677d066255b8e668a48422f23f909f4ddee\n" }, { "commit": "7877df66c27c8cd27dcb809e5cbfdc38b798e5cb", "tree": "d77a82183c705cb4a0e2ebf290de217d86bc8af6", "parents": [ "9567c7b82cba21d3046bab5c9b1fe4e73f2629bf" ], "author": { "name": "Colin Cross", "email": "ccross@android.com", "time": "Thu Mar 10 13:01:27 2016 -0800" }, "committer": { "name": "Colin Cross", "email": "ccross@android.com", "time": "Thu Mar 10 14:53:02 2016 -0800" }, "message": "malloc_debug: fix multiplication overflow in debug_calloc\n\nThe over flow check for nmemb * bytes in debug_calloc is incorrect,\nuse the builtin overflow functions to check for multiplication and\naddition overflow.\n\nChange-Id: I3f1c13102621bc5380be1f69caa88dba2118f3cb\n(cherry picked from commit 239838608dbe9917acddfe5a51d92350a4c8e135)\n" }, { "commit": "9567c7b82cba21d3046bab5c9b1fe4e73f2629bf", "tree": "7804baf7adca1beafa840fdb020e1ffc50cd48be", "parents": [ "baa7c6f307569ec3ed17688495e29000cd8bc617" ], "author": { "name": "Colin Cross", "email": "ccross@android.com", "time": "Wed Mar 09 17:56:14 2016 -0800" }, "committer": { "name": "Colin Cross", "email": "ccross@android.com", "time": "Thu Mar 10 14:53:02 2016 -0800" }, "message": "malloc_debug: round 0 byte allocations up to 1 byte\n\n0 byte allocations can cause problems if they are immediately followed\nby another allocation with no header, as both allocations will have the\nsame address. Treat 0 byte allocations as 1 byte allocations so that\ndebug_iterate will return separate addresses for them.\n\nBug: 27578580\nChange-Id: Ia8dc3481fa7062391e9b3ae58a36e8d47e7ee557\n(cherry picked from commit 15af478080cfbfa800fb8172fdf70a84075925e3)\n" }, { "commit": "baa7c6f307569ec3ed17688495e29000cd8bc617", "tree": "ae68b98d1ec02f896f0d7af99012710c1f3dee74", "parents": [ "ffd19ee64d89d5a9ba3757e8ecb1017e0ec5f135" ], "author": { "name": "Colin Cross", "email": "ccross@android.com", "time": "Wed Mar 09 16:33:44 2016 -0800" }, "committer": { "name": "Colin Cross", "email": "ccross@android.com", "time": "Thu Mar 10 14:53:02 2016 -0800" }, "message": "malloc_debug: iterate: use usable_size\n\nmalloc_iterate with malloc debug enabled was returning allocation sizes\nusing the requested size instead of usable size. If anything wrote\npointers above the requested size but below the usable size, those\npointers would be invisible to libmemunreachable and referenced\nallocations could be reported as a leak.\n\nBug: 27107100\nChange-Id: I6b19fd631f68ce93b3aee408cc3d296ec457bd01\n(cherry picked from commit b23c6606d2326f8835fc7fc35b23ff7ee5a48f48)\n" }, { "commit": "2c75991359df165ca7cc7a6213fb227c0b5ed87c", "tree": "4dc02df00481962e80acb454ee857fd5b2f8b2d4", "parents": [ "2d4721c0c57fe2f7c1e1b40df4763a561b3cf856" ], "author": { "name": "Colin Cross", "email": "ccross@android.com", "time": "Fri Feb 05 16:17:39 2016 -0800" }, "committer": { "name": "Colin Cross", "email": "ccross@android.com", "time": "Thu Feb 18 16:09:17 2016 -0800" }, "message": "Add backtrace_string and export to libmemunreachable\n\nAdd backtrace_string to convert a malloc_debug backtrace to a string.\nAlso move the backtrace functions to libc_malloc_debug_backtrace so that\nlibmemunreachable can reuse them.\n\nChange-Id: I5ad67001c0b4d184903c762863a8588181d4873b\n" }, { "commit": "2d4721c0c57fe2f7c1e1b40df4763a561b3cf856", "tree": "256b5519e32df5b56b6bc6a343f5ff50d16ddea0", "parents": [ "869691c6f7691dd5bf343617aa185c32e0d3d160" ], "author": { "name": "Colin Cross", "email": "ccross@android.com", "time": "Tue Feb 02 11:57:54 2016 -0800" }, "committer": { "name": "Colin Cross", "email": "ccross@android.com", "time": "Thu Feb 18 16:09:16 2016 -0800" }, "message": "Export malloc_backtrace\n\nChange-Id: Ic1adb4dfd86b9ca698443a36263a3df2c91edda3\n" }, { "commit": "869691c6f7691dd5bf343617aa185c32e0d3d160", "tree": "aa0cb5619aa90a899a31ca0ba250cb180c5d1302", "parents": [ "aca201358b910e5a4b925c639902387fbdfd9016" ], "author": { "name": "Colin Cross", "email": "ccross@android.com", "time": "Fri Jan 29 12:48:18 2016 -0800" }, "committer": { "name": "Colin Cross", "email": "ccross@android.com", "time": "Thu Feb 18 15:05:38 2016 -0800" }, "message": "Export malloc_iterate and friends\n\nExport malloc_iterate, malloc_enable, and malloc_disable to be used by\nlibmemunreachable.\n\nChange-Id: I08a50349af82a95d096b6b4cbac37ababe4b9b06\n" }, { "commit": "598cb89790ba93edbb392c567ecda32b39d64ff1", "tree": "28f8dc579b1462a232620d180ab5f32a88f687f4", "parents": [ "2de48bc809bba48a7f209e209b3822bad010166f", "7a28a3cf1f8df36e30724e8b4021cddde0596118" ], "author": { "name": "Colin Cross", "email": "ccross@android.com", "time": "Wed Feb 17 20:38:02 2016 +0000" }, "committer": { "name": "Gerrit Code Review", "email": "noreply-gerritcodereview@google.com", "time": "Wed Feb 17 20:38:02 2016 +0000" }, "message": "Merge \"malloc_debug: reset TrackData mutex after fork\"" }, { "commit": "72df6708c829a4c6494936fdfbda6dc7e68e647b", "tree": "17ede0d54b6f50c3598a3fcec56f50754067690d", "parents": [ "72bca4b4105e24058f3f2eca024382bedb122a30" ], "author": { "name": "Christopher Ferris", "email": "cferris@google.com", "time": "Thu Feb 11 15:51:31 2016 -0800" }, "committer": { "name": "Christopher Ferris", "email": "cferris@google.com", "time": "Fri Feb 12 17:30:30 2016 -0800" }, "message": "Fix the default alignment of the allocations.\n\nIn order to enforce this constraint:\n\nThe pointer returned if the allocation succeeds shall be suitably\naligned so that it may be assigned to a pointer to any type of object\nand then used to access such an object in the space allocated.\n\nForce all allocations on 32 bit systems to have 8 byte alignment,\nand all allocations on 64 bit systems to have 16 byte alignment.\n\nAdd a test to verify that the allocator returns the correct alignments.\n\nBug: 26739265\nChange-Id: I9af53279617408676b94e4ec6481b3ed7ffafc6a\n" }, { "commit": "7a28a3cf1f8df36e30724e8b4021cddde0596118", "tree": "2add78d5dcf97d4cb4f360fba7b512656be70313", "parents": [ "9750a77b31f2fc6d548a02b3a9750c2794648fea" ], "author": { "name": "Colin Cross", "email": "ccross@android.com", "time": "Sun Feb 07 22:51:15 2016 -0800" }, "committer": { "name": "Colin Cross", "email": "ccross@android.com", "time": "Fri Feb 12 16:41:14 2016 -0800" }, "message": "malloc_debug: reset TrackData mutex after fork\n\nAdd a pthread_atfork handler to malloc_debug to lock the TrackData mutex\nduring fork and reset it in the child. Ensures that the TrackData is\nconsistent when forking from a multi-threaded process, and that the\nmutex is in a defined state in the child.\n\nChange-Id: I0dfed30045a28551217ceac227d33b2e18894932\n" }, { "commit": "7993b80f894db20af4d1d154221c42fea6171a3d", "tree": "fc99d36000a8f0853224d75296d96509a730ee05", "parents": [ "50647711ebaf2360aca05cc94a2fdf431c1a153e" ], "author": { "name": "Christopher Ferris", "email": "cferris@google.com", "time": "Thu Jan 28 18:35:05 2016 -0800" }, "committer": { "name": "Christopher Ferris", "email": "cferris@google.com", "time": "Fri Jan 29 15:19:22 2016 -0800" }, "message": "Add better free tracking.\n\nIncluded in this change:\n\n- Change the tag when a pointer is freed so it\u0027s easy to detect if\n an already freed pointer is being used.\n- Move the free backtrace out of the header. This backtrace is only\n used under only some circumstances, so no need to allocate space\n in all headers for it.\n- Add new option free_track_backtrace_num_frames to specify how many\n frames to record when the free occurs. This removes the dependency\n on the backtrace option to get backtraces.\n\nBug: 26739265\nChange-Id: I76f5209507dcf46af67ada162a7cb2bf282116f2\n" }, { "commit": "63860cb8fd1adf3f679b9b4ad876323a8d65cd9d", "tree": "25aeae686d92efcb5e08a54e8cba1e8c31efe33b", "parents": [ "ad9c3f34f762ed45cce5dbb93218124ed31f6873" ], "author": { "name": "Christopher Ferris", "email": "cferris@google.com", "time": "Mon Nov 16 17:30:32 2015 -0800" }, "committer": { "name": "Christopher Ferris", "email": "cferris@google.com", "time": "Mon Jan 25 10:54:21 2016 -0800" }, "message": "Malloc debug rewrite.\n\nThe major components of the rewrite:\n\n- Completely remove the qemu shared library code. Nobody was using it\n and it appears to have broken at some point.\n- Adds the ability to enable/disable different options independently.\n- Adds a new option that can enable the backtrace on alloc/free when\n a process gets a specific signal.\n- Adds a new way to enable malloc debug. If a special property is\n set, and the process has an environment variable set, then debug\n malloc will be enabled. This allows something that might be\n a derivative of app_process to be started with an environment variable\n being enabled.\n- get_malloc_leak_info() used to return one element for each pointer that\n had the exact same backtrace. The new version returns information for\n every one of the pointers with same backtrace. It turns out ddms already\n automatically coalesces these, so the old method simply hid the fact\n that there where multiple pointers with the same amount of backtrace.\n- Moved all of the malloc debug specific code into the library.\n Nothing related to the malloc debug data structures remains in libc.\n- Removed the calls to the debug malloc cleanup routine. Instead, I\n added an atexit call with the debug malloc cleanup routine. This gets\n around most problems related to the timing of doing the cleanup.\n\nThe new properties and environment variables:\n\nlibc.debug.malloc.options\n Set by option name (such as \"backtrace\"). Setting this to a bad value\n will cause a usage statement to be printed to the log.\n\nlibc.debug.malloc.program\n Same as before. If this is set, then only the program named will\n be launched with malloc debug enabled. This is not a complete match,\n but if any part of the property is in the program name, malloc debug is\n enabled.\n\nlibc.debug.malloc.env_enabled\n If set, then malloc debug is only enabled if the running process has the\n environment variable LIBC_DEBUG_MALLOC_ENABLE set.\n\nBug: 19145921\n\nChange-Id: I7b0e58cc85cc6d4118173fe1f8627a391b64c0d7\n" } ] }