128549 2014-02-10 12:24:56 -0800 ENABLED(JIT) && !ENABLED(DFG_JIT) breaks the build 2014-06-30 02:26:03 -0700 1 1 1 Unclassified WebKit JavaScriptCore 528+ (Nightly build) Unspecified Unspecified RESOLVED FIXED P2 Normal --- 1 desrt webkit-unassigned benjamin berto cmarcelo commit-queue dbatyai.u-szeged fpizlo ggaren lantw44 mark.lam mhahnenberg msaboff oliver ossy zan oldest_to_newest 979000 0 desrt 2014-02-10 12:24:56 -0800 Building on FreeBSD results in JIT being enabled, but not DFG_JIT. This is a result of this check: #if !defined(ENABLE_DFG_JIT) && ENABLE(JIT) && !COMPILER(MSVC) /* Enable the DFG JIT on X86 and X86_64. Only tested on Mac and GNU/Linux. */ #if (CPU(X86) || CPU(X86_64)) && (OS(DARWIN) || OS(LINUX)) #define ENABLE_DFG_JIT 1 #endif As a result, some files in JavascriptScore fail to compile in a fairly simple (presumably as a result of nobody testing this combination). Source/JavaScriptCore/bytecode/CallLinkStatus.cpp:125:24: error: no member named 'hasExitSite' in 'JSC::CodeBlock' if (profiledBlock->hasExitSite(locker, DFG::FrequentExitSite(bytecodeIndex, BadCache)) ~~~~~~~~~~~~~ ^ Source/JavaScriptCore/bytecode/CallLinkStatus.cpp:125:49: error: no member named 'FrequentExitSite' in namespace 'JSC::DFG' if (profiledBlock->hasExitSite(locker, DFG::FrequentExitSite(bytecodeIndex, BadCache)) ~~~~~^ Source/JavaScriptCore/bytecode/CallLinkStatus.cpp:126:27: error: no member named 'hasExitSite' in 'JSC::CodeBlock' || profiledBlock->hasExitSite(locker, DFG::FrequentExitSite(bytecodeIndex, BadCacheWatchpoint)) ~~~~~~~~~~~~~ ^ Source/JavaScriptCore/bytecode/CallLinkStatus.cpp:126:52:Source/JavaScriptCore/bytecode/GetByIdStatus.cpp:40:27: error: no member named 'hasExitSite' in 'JSC::CodeBlock' return profiledBlock->hasExitSite(locker, DFG::FrequentExitSite(bytecodeIndex, BadCache, jitType)) ~~~~~~~~~~~~~ ^ error: no member named 'FrequentExitSite' in namespace 'JSC::DFG' || profiledBlock->hasExitSite(locker, DFG::FrequentExitSite(bytecodeIndex, BadCacheWatchpoint)) ~~~~~^ Source/JavaScriptCore/bytecode/CallLinkStatus.cpp:127:27: error: no member named 'hasExitSite' in 'JSC::CodeBlock' || profiledBlock->hasExitSite(locker, DFG::FrequentExitSite(bytecodeIndex, BadExecutable))) ~~~~~~~~~~~~~ ^ Source/JavaScriptCore/bytecode/GetByIdStatus.cpp:40:52: error: no member named 'FrequentExitSite' in namespace 'JSC::DFG' return profiledBlock->hasExitSite(locker, DFG::FrequentExitSite(bytecodeIndex, BadCache, jitType)) ~~~~~^ Source/JavaScriptCore/bytecode/GetByIdStatus.cpp:41:27: error: no member named 'hasExitSite' in 'JSC::CodeBlock' || profiledBlock->hasExitSite(locker, DFG::FrequentExitSite(bytecodeIndex, BadCacheWatchpoint, jitType)) ~~~~~~~~~~~~~ ^ Source/JavaScriptCore/bytecode/CallLinkStatus.cpp:127:52: error: no member named 'FrequentExitSite' in namespace 'JSC::DFG' || profiledBlock->hasExitSite(locker, DFG::FrequentExitSite(bytecodeIndex, BadExecutable))) ~~~~~^ Source/JavaScriptCore/bytecode/CallLinkStatus.cpp:142:24: error: no member named 'hasExitSite' in 'JSC::CodeBlock' if (profiledBlock->hasExitSite(locker, DFG::FrequentExitSite(bytecodeIndex, BadFunction))) ~~~~~~~~~~~~~ ^ Source/JavaScriptCore/bytecode/GetByIdStatus.cpp:41:52: error: no member named 'FrequentExitSite' in namespace 'JSC::DFG' || profiledBlock->hasExitSite(locker, DFG::FrequentExitSite(bytecodeIndex, BadCacheWatchpoint, jitType)) ~~~~~^ Source/JavaScriptCore/bytecode/GetByIdStatus.cpp:42:27: error: no member named 'hasExitSite' in 'JSC::CodeBlock' || profiledBlock->hasExitSite(locker, DFG::FrequentExitSite(bytecodeIndex, BadWeakConstantCache, jitType)) ~~~~~~~~~~~~~ ^ Source/JavaScriptCore/bytecode/CallLinkStatus.cpp:142:49: error: no member named 'FrequentExitSite' in namespace 'JSC::DFG' if (profiledBlock->hasExitSite(locker, DFG::FrequentExitSite(bytecodeIndex, BadFunction))) ~~~~~^ Source/JavaScriptCore/bytecode/GetByIdStatus.cpp:42:52: error: no member named 'FrequentExitSite' in namespace 'JSC::DFG' || profiledBlock->hasExitSite(locker, DFG::FrequentExitSite(bytecodeIndex, BadWeakConstantCache, jitType)) ~~~~~^ Source/JavaScriptCore/bytecode/GetByIdStatus.cpp:43:27: error: no member named 'hasExitSite' in 'JSC::CodeBlock' || profiledBlock->hasExitSite(locker, DFG::FrequentExitSite(bytecodeIndex, BadWeakConstantCacheWatchpoin... ~~~~~~~~~~~~~ ^ CXX Source/JavaScriptCore/bytecode/libjavascriptcoregtk_3_0_la-LazyOperandValueProfile.lo Source/JavaScriptCore/bytecode/GetByIdStatus.cpp:43:52: error: no member named 'FrequentExitSite' in namespace 'JSC::DFG' || profiledBlock->hasExitSite(locker, DFG::FrequentExitSite(bytecodeIndex, BadWeakConstantCacheWatchpoin... ~~~~~^ CXX Source/JavaScriptCore/bytecode/libjavascriptcoregtk_3_0_la-MethodOfGettingAValueProfile.lo 8 errors generated. The first of these errors, at CallLinkStatus.cpp:125:24, is caused by a use of DFG::FrequentExitSite guarded by only ENABLE(JIT) whereas its site of definition is guarded by ENABLE(DFG_JIT): #ifndef DFGExitProfile_h #define DFGExitProfile_h #if ENABLE(DFG_JIT) ... namespace JSC { namespace DFG { class FrequentExitSite { ... A earlier use of DFG::FrequentExitSite in the same file (line 94) is guarded by ENABLE(DFG_JIT) and there are no problems there. The fix to this bug is probably to change some cases of ENABLE(JIT) to ENABLE(DFG_JIT) but ultimately it seems like it might also be a good idea to enable DFG_JIT on FreeBSD. 979318 1 berto 2014-02-11 00:32:12 -0800 (In reply to comment #0) > The fix to this bug is probably to change some cases of ENABLE(JIT) > to ENABLE(DFG_JIT) Yes, probably, it seems that this was introduced when the jsCStack was merged, adding Michael Saboff to Cc. > but ultimately it seems like it might also be a good idea to enable > DFG_JIT on FreeBSD. I agree, can you test if it works or it doesn't build yet? :) Thanks! 979363 2 desrt 2014-02-11 03:54:14 -0800 It still doesn't build because of the 'too long argument list' to the linker problem. I can verify that adding OS(FREEBSD) to the list does get it past the build problem described here, though. 979372 3 desrt 2014-02-11 05:36:34 -0800 Okay. I managed to fix a couple of issues to get a complete build. I can verify that forcing DFG_JIT to be enabled on FreeBSD results in a working browser. It scores 100/100 on Acid3, at least. Is there any other 'more scientific' way that I can test that? 979378 4 223847 desrt 2014-02-11 06:00:25 -0800 Created attachment 223847 Enable DFG_JIT on FreeBSD 979803 5 berto 2014-02-12 01:18:36 -0800 (In reply to comment #4) > Enable DFG_JIT on FreeBSD It's probably fine to enable this on FreeBSD, but there's still the original problem of why it no longer builds without it. 979827 6 zan 2014-02-12 03:46:09 -0800 (In reply to comment #5) > (In reply to comment #4) > > Enable DFG_JIT on FreeBSD > > It's probably fine to enable this on FreeBSD, but there's still the > original problem of why it no longer builds without it. The reasons are probably the same as those outlined here: https://bugs.webkit.org/show_bug.cgi?id=126502#c4 979828 7 berto 2014-02-12 03:51:37 -0800 (In reply to comment #6) > > It's probably fine to enable this on FreeBSD, but there's still > > the original problem of why it no longer builds without it. > The reasons are probably the same as those outlined here: > https://bugs.webkit.org/show_bug.cgi?id=126502#c4 I was thinking the same, but then we should probably remove the #ifdefs altogether. 979854 8 desrt 2014-02-12 06:38:59 -0800 With DFG_JIT on FreeBSD, stress tests pass 8212/8212. The mozilla tests have 45 failures out of 1127, but apparently this is normal (and indeed, many of the failures have links to already-open bugs). 980224 9 224057 dbatyai.u-szeged 2014-02-13 03:51:56 -0800 Created attachment 224057 Build fix for enabled JIT with disabled DFG_JIT 981389 10 berto 2014-02-17 03:55:15 -0800 (In reply to comment #9) > Created an attachment (id=224057) [details] > Build fix for enabled JIT with disabled DFG_JIT Thanks for the patch, although I haven't had the chance to test it yet. Even if we fix the build we can also enable DFG_JIT on FreeBSD, I just filed bug 128898 for that. 981396 11 224057 ossy 2014-02-17 04:37:12 -0800 Comment on attachment 224057 Build fix for enabled JIT with disabled DFG_JIT CallLinkStatus.cpp, GetByIdStatus.cpp and PutByIdStatus.cpp were already fixed by https://trac.webkit.org/changeset/164164 981397 12 224057 ossy 2014-02-17 04:40:32 -0800 Comment on attachment 224057 Build fix for enabled JIT with disabled DFG_JIT View in context: https://bugs.webkit.org/attachment.cgi?id=224057&action=review r-, because it can't be applied now. > Source/JavaScriptCore/jit/Repatch.cpp:825 > +#if ENABLE(DFG_JIT) > MacroAssembler::Jump definitelyNotMarked = DFG::SpeculativeJIT::genericWriteBarrier(jit, owner, scratch1, scratch2); > - MacroAssembler::Call call = storeToWriteBarrierBuffer(jit, owner, scratch1, scratch2, allocator); > definitelyNotMarked.link(&jit); > +#endif // ENABLE(DFG_JIT) > + > + MacroAssembler::Call call = storeToWriteBarrierBuffer(jit, owner, scratch1, scratch2, allocator); Are you sure if this is the proper fix? I don't understand this code, but these lines looks to me as they can't be separated. 981411 13 ossy 2014-02-17 05:29:53 -0800 I tried the proposed fix for Repatch.cpp, it builds but I got some test failures: stress/tricky-array-bounds-checks.js.always-trigger-copy-phase: Exception: Error: bad result for trick: ,,,false stress/tricky-array-bounds-checks.js.default: Exception: Error: bad result for trick: ,,,false stress/tricky-array-bounds-checks.js.no-llint: Exception: Error: bad result for trick: ,,,false stress/tricky-array-bounds-checks.js.no-cjit-validate-phases: Exception: Error: bad result for trick: ,,,false stress/tricky-array-bounds-checks.js.default: ERROR: Unexpected exit code: 3 stress/tricky-array-bounds-checks.js.always-trigger-copy-phase: ERROR: Unexpected exit code: 3 stress/tricky-array-bounds-checks.js.no-llint: ERROR: Unexpected exit code: 3 stress/tricky-array-bounds-checks.js.dfg-eager: Exception: Error: bad result for trick: ,,,false stress/tricky-array-bounds-checks.js.dfg-eager-no-cjit-validate: Exception: Error: bad result for trick: ,,,false stress/tricky-array-bounds-checks.js.no-cjit-validate-phases: ERROR: Unexpected exit code: 3 stress/tricky-array-bounds-checks.js.dfg-eager: ERROR: Unexpected exit code: 3 stress/tricky-array-bounds-checks.js.dfg-eager-no-cjit-validate: ERROR: Unexpected exit code: 3 8352/8352 (failed 6) ** The following JSC stress test failures have been introduced: stress/tricky-array-bounds-checks.js.default stress/tricky-array-bounds-checks.js.no-llint stress/tricky-array-bounds-checks.js.always-trigger-copy-phase stress/tricky-array-bounds-checks.js.no-cjit-validate-phases stress/tricky-array-bounds-checks.js.dfg-eager stress/tricky-array-bounds-checks.js.dfg-eager-no-cjit-validate Results for JSC stress tests: 6 failures found. I'm not sure if they were caused by this change or other bug. 981415 14 ossy 2014-02-17 05:40:26 -0800 I checked the failures, they aren't related to this fix, they fail on trunk with runtime dissabled DFG too. 981443 15 ossy 2014-02-17 07:15:29 -0800 (In reply to comment #14) > I checked the failures, they aren't related to this fix, > they fail on trunk with runtime dissabled DFG too. New bug report for this failire: https://bugs.webkit.org/show_bug.cgi?id=128910 984418 16 desrt 2014-02-25 12:00:51 -0800 (In reply to comment #14) > I checked the failures, they aren't related to this fix, > they fail on trunk with runtime dissabled DFG too. You still have a review- flag on this bug. 984724 17 ossy 2014-02-26 06:23:33 -0800 (In reply to comment #16) > (In reply to comment #14) > > I checked the failures, they aren't related to this fix, > > they fail on trunk with runtime dissabled DFG too. > > You still have a review- flag on this bug. I r- -ed it, because it is obsolete now, see https://bugs.webkit.org/show_bug.cgi?id=128549#c11 for details. Maybe the fix in Source/JavaScriptCore/jit/Repatch.cpp is good. It worked for me without new regressions. But I'm not sure if it is the proper fix. Comment about it from a JSC expert would be welcome. 986813 18 225780 dbatyai.u-szeged 2014-03-04 08:16:08 -0800 Created attachment 225780 Proposed fix v2 987596 19 225780 mark.lam 2014-03-06 00:40:13 -0800 Comment on attachment 225780 Proposed fix v2 Thanks for looking into this. However, I think this issue has been solved in a more appropriate refactoring in https://bugs.webkit.org/show_bug.cgi?id=129760. Please re-test after that lands (or try the patch in that bug) to see if there are still other outstanding issue for the !ENABLE(DFG_JIT) build. 987674 20 dbatyai.u-szeged 2014-03-06 06:52:32 -0800 (In reply to comment #19) > (From update of attachment 225780 [details]) > Thanks for looking into this. However, I think this issue has been solved in a more appropriate refactoring in https://bugs.webkit.org/show_bug.cgi?id=129760. Please re-test after that lands (or try the patch in that bug) to see if there are still other outstanding issue for the !ENABLE(DFG_JIT) build. Thanks for the heads up. I tried the patch and managed to build with disabled DFG JIT without any problems. 1019306 21 ossy 2014-06-30 02:26:03 -0700 ENABLE(JIT) && !ENABLE(DFG_JIT) build works fine on ToT. 223847 2014-02-11 06:00:25 -0800 2014-02-11 06:00:25 -0800 Enable DFG_JIT on FreeBSD 0001-Enable-DFG_JIT-on-FreeBSD.patch text/plain 1192 desrt RnJvbSA3ZjY3NDkyNmExMWUzN2U1NjNkMjA0ZmE3ZDIyOTMzYTgxNGI2ZDVlIE1vbiBTZXAgMTcg MDA6MDA6MDAgMjAwMQpGcm9tOiBSeWFuIExvcnRpZSA8ZGVzcnRAZGVzcnQuY2E+CkRhdGU6IFR1 ZSwgMTEgRmViIDIwMTQgMDg6NTE6NTEgLTA1MDAKU3ViamVjdDogW1BBVENIIDEvNF0gRW5hYmxl IERGR19KSVQgb24gRnJlZUJTRAoKSSd2ZSB0cmllZCBhIGJ1aWxkIG9mIFdlYktpdCBvbiBGcmVl QlNEIGFuZCBpdCBzZWVtcyB0byBiZSB3b3JraW5nCm5pY2VseSB3aXRoIERGR19KSVQgZW5hYmxl ZC4KClRoaXMgYWxzbyB3b3JrcyBhcm91bmQgdGhlIGlzc3VlcyBjYXVzZWQgYnkgRU5BQkxFRChK SVQpIGJ1dCBub3QKRU5BQkxFRChERkdfSklUKSwgYnV0IHRoYXQgaXNzdWUgc2hvdWxkIHN0aWxs IGJlIGZpeGVkIHNlcGFyYXRlbHkuCi0tLQogU291cmNlL1dURi93dGYvUGxhdGZvcm0uaCB8IDQg KystLQogMSBmaWxlIGNoYW5nZWQsIDIgaW5zZXJ0aW9ucygrKSwgMiBkZWxldGlvbnMoLSkKCmRp ZmYgLS1naXQgYS9Tb3VyY2UvV1RGL3d0Zi9QbGF0Zm9ybS5oIGIvU291cmNlL1dURi93dGYvUGxh dGZvcm0uaAppbmRleCBhZWVjMjRmLi5lZDg0ZTdiIDEwMDY0NAotLS0gYS9Tb3VyY2UvV1RGL3d0 Zi9QbGF0Zm9ybS5oCisrKyBiL1NvdXJjZS9XVEYvd3RmL1BsYXRmb3JtLmgKQEAgLTcwOCw4ICs3 MDgsOCBAQAogI2VuZGlmCiAKICNpZiAhZGVmaW5lZChFTkFCTEVfREZHX0pJVCkgJiYgRU5BQkxF KEpJVCkgJiYgIUNPTVBJTEVSKE1TVkMpCi0vKiBFbmFibGUgdGhlIERGRyBKSVQgb24gWDg2IGFu ZCBYODZfNjQuICBPbmx5IHRlc3RlZCBvbiBNYWMgYW5kIEdOVS9MaW51eC4gKi8KLSNpZiAoQ1BV KFg4NikgfHwgQ1BVKFg4Nl82NCkpICYmIChPUyhEQVJXSU4pIHx8IE9TKExJTlVYKSkKKy8qIEVu YWJsZSB0aGUgREZHIEpJVCBvbiBYODYgYW5kIFg4Nl82NC4gIE9ubHkgdGVzdGVkIG9uIE1hYywg R05VL0xpbnV4IGFuZCBGcmVlQlNELiAqLworI2lmIChDUFUoWDg2KSB8fCBDUFUoWDg2XzY0KSkg JiYgKE9TKERBUldJTikgfHwgT1MoTElOVVgpIHx8IE9TKEZSRUVCU0QpKQogI2RlZmluZSBFTkFC TEVfREZHX0pJVCAxCiAjZW5kaWYKIC8qIEVuYWJsZSB0aGUgREZHIEpJVCBvbiBBUk12Ny4gIE9u bHkgdGVzdGVkIG9uIGlPUyBhbmQgUXQvR1RLKyBMaW51eC4gKi8KLS0gCjEuOC41LjIKCg== 224057 2014-02-13 03:51:56 -0800 2014-02-26 01:19:05 -0800 Build fix for enabled JIT with disabled DFG_JIT proposedpatch.patch text/plain 4209 dbatyai.u-szeged ZGlmZiAtLWdpdCBhL1NvdXJjZS9KYXZhU2NyaXB0Q29yZS9DaGFuZ2VMb2cgYi9Tb3VyY2UvSmF2 YVNjcmlwdENvcmUvQ2hhbmdlTG9nCmluZGV4IDNlYTNkN2UuLmM5Yjk0ZmIgMTAwNjQ0Ci0tLSBh L1NvdXJjZS9KYXZhU2NyaXB0Q29yZS9DaGFuZ2VMb2cKKysrIGIvU291cmNlL0phdmFTY3JpcHRD b3JlL0NoYW5nZUxvZwpAQCAtMSwzICsxLDE3IEBACisyMDE0LTAyLTEzICBEw6FuaWVsIELDoXR5 YWkgIDxkYmF0eWFpLnUtc3plZ2VkQHBhcnRuZXIuc2Ftc3VuZy5jb20+CisKKyAgICAgICAgRU5B QkxFRChKSVQpICYmICFFTkFCTEVEKERGR19KSVQpIGJyZWFrcyB0aGUgYnVpbGQKKyAgICAgICAg aHR0cHM6Ly9idWdzLndlYmtpdC5vcmcvc2hvd19idWcuY2dpP2lkPTEyODU0OQorCisgICAgICAg IFJldmlld2VkIGJ5IE5PQk9EWSAoT09QUyEpLgorCisgICAgICAgICogYnl0ZWNvZGUvQ2FsbExp bmtTdGF0dXMuY3BwOgorICAgICAgICAoSlNDOjpDYWxsTGlua1N0YXR1czo6Y29tcHV0ZUZvcik6 CisgICAgICAgICogYnl0ZWNvZGUvR2V0QnlJZFN0YXR1cy5jcHA6CisgICAgICAgICogYnl0ZWNv ZGUvUHV0QnlJZFN0YXR1cy5jcHA6CisgICAgICAgICogaml0L1JlcGF0Y2guY3BwOgorICAgICAg ICAoSlNDOjp3cml0ZUJhcnJpZXIpOgorCiAyMDE0LTAyLTA5ICBGaWxpcCBQaXpsbyAgPGZwaXps b0BhcHBsZS5jb20+CiAKICAgICAgICAgR0MgYmxvY2tzIG9uIEZUTCBhbmQgdGhlbiBiYWRuZXNz CmRpZmYgLS1naXQgYS9Tb3VyY2UvSmF2YVNjcmlwdENvcmUvYnl0ZWNvZGUvQ2FsbExpbmtTdGF0 dXMuY3BwIGIvU291cmNlL0phdmFTY3JpcHRDb3JlL2J5dGVjb2RlL0NhbGxMaW5rU3RhdHVzLmNw cAppbmRleCBkNDZhNjhmLi42YjMwMDZiIDEwMDY0NAotLS0gYS9Tb3VyY2UvSmF2YVNjcmlwdENv cmUvYnl0ZWNvZGUvQ2FsbExpbmtTdGF0dXMuY3BwCisrKyBiL1NvdXJjZS9KYXZhU2NyaXB0Q29y ZS9ieXRlY29kZS9DYWxsTGlua1N0YXR1cy5jcHAKQEAgLTEyMiwxMSArMTIyLDEzIEBAIENhbGxM aW5rU3RhdHVzIENhbGxMaW5rU3RhdHVzOjpjb21wdXRlRm9yKENvZGVCbG9jayogcHJvZmlsZWRC bG9jaywgdW5zaWduZWQgYnl0CiAgICAgVU5VU0VEX1BBUkFNKHByb2ZpbGVkQmxvY2spOwogICAg IFVOVVNFRF9QQVJBTShieXRlY29kZUluZGV4KTsKICNpZiBFTkFCTEUoSklUKQorI2lmIEVOQUJM RShERkdfSklUKQogICAgIGlmIChwcm9maWxlZEJsb2NrLT5oYXNFeGl0U2l0ZShsb2NrZXIsIERG Rzo6RnJlcXVlbnRFeGl0U2l0ZShieXRlY29kZUluZGV4LCBCYWRDYWNoZSkpCiAgICAgICAgIHx8 IHByb2ZpbGVkQmxvY2stPmhhc0V4aXRTaXRlKGxvY2tlciwgREZHOjpGcmVxdWVudEV4aXRTaXRl KGJ5dGVjb2RlSW5kZXgsIEJhZENhY2hlV2F0Y2hwb2ludCkpCiAgICAgICAgIHx8IHByb2ZpbGVk QmxvY2stPmhhc0V4aXRTaXRlKGxvY2tlciwgREZHOjpGcmVxdWVudEV4aXRTaXRlKGJ5dGVjb2Rl SW5kZXgsIEJhZEV4ZWN1dGFibGUpKSkKICAgICAgICAgcmV0dXJuIHRha2VzU2xvd1BhdGgoKTsK LSAgICAKKyNlbmRpZiAvLyBFTkFCTEUoREZHX0pJVCkKKwogICAgIGlmICghcHJvZmlsZWRCbG9j ay0+aGFzQmFzZWxpbmVKSVRQcm9maWxpbmcoKSkKICAgICAgICAgcmV0dXJuIGNvbXB1dGVGcm9t TExJbnQobG9ja2VyLCBwcm9maWxlZEJsb2NrLCBieXRlY29kZUluZGV4KTsKICAgICAKQEAgLTEz OCwxMCArMTQwLDEyIEBAIENhbGxMaW5rU3RhdHVzIENhbGxMaW5rU3RhdHVzOjpjb21wdXRlRm9y KENvZGVCbG9jayogcHJvZmlsZWRCbG9jaywgdW5zaWduZWQgYnl0CiAgICAgQ2FsbExpbmtTdGF0 dXMgcmVzdWx0ID0gY29tcHV0ZUZvcihsb2NrZXIsIGNhbGxMaW5rSW5mbyk7CiAgICAgaWYgKCFy ZXN1bHQpCiAgICAgICAgIHJldHVybiBjb21wdXRlRnJvbUxMSW50KGxvY2tlciwgcHJvZmlsZWRC bG9jaywgYnl0ZWNvZGVJbmRleCk7Ci0gICAgCisKKyNpZiBFTkFCTEUoREZHX0pJVCkKICAgICBp ZiAocHJvZmlsZWRCbG9jay0+aGFzRXhpdFNpdGUobG9ja2VyLCBERkc6OkZyZXF1ZW50RXhpdFNp dGUoYnl0ZWNvZGVJbmRleCwgQmFkRnVuY3Rpb24pKSkKICAgICAgICAgcmVzdWx0Lm1ha2VDbG9z dXJlQ2FsbCgpOwotICAgIAorI2VuZGlmIC8vIEVOQUJMRShERkdfSklUKQorCiAgICAgcmV0dXJu IHJlc3VsdDsKICNlbHNlCiAgICAgcmV0dXJuIENhbGxMaW5rU3RhdHVzKCk7CmRpZmYgLS1naXQg YS9Tb3VyY2UvSmF2YVNjcmlwdENvcmUvYnl0ZWNvZGUvR2V0QnlJZFN0YXR1cy5jcHAgYi9Tb3Vy Y2UvSmF2YVNjcmlwdENvcmUvYnl0ZWNvZGUvR2V0QnlJZFN0YXR1cy5jcHAKaW5kZXggYTNjNGZi Zi4uZjc5NmYzYjkgMTAwNjQ0Ci0tLSBhL1NvdXJjZS9KYXZhU2NyaXB0Q29yZS9ieXRlY29kZS9H ZXRCeUlkU3RhdHVzLmNwcAorKysgYi9Tb3VyY2UvSmF2YVNjcmlwdENvcmUvYnl0ZWNvZGUvR2V0 QnlJZFN0YXR1cy5jcHAKQEAgLTM0LDcgKzM0LDcgQEAKIAogbmFtZXNwYWNlIEpTQyB7CiAKLSNp ZiBFTkFCTEUoSklUKQorI2lmIEVOQUJMRShERkdfSklUKQogYm9vbCBHZXRCeUlkU3RhdHVzOjpo YXNFeGl0U2l0ZShjb25zdCBDb25jdXJyZW50SklUTG9ja2VyJiBsb2NrZXIsIENvZGVCbG9jayog cHJvZmlsZWRCbG9jaywgdW5zaWduZWQgYnl0ZWNvZGVJbmRleCwgRXhpdGluZ0pJVFR5cGUgaml0 VHlwZSkKIHsKICAgICByZXR1cm4gcHJvZmlsZWRCbG9jay0+aGFzRXhpdFNpdGUobG9ja2VyLCBE Rkc6OkZyZXF1ZW50RXhpdFNpdGUoYnl0ZWNvZGVJbmRleCwgQmFkQ2FjaGUsIGppdFR5cGUpKQpk aWZmIC0tZ2l0IGEvU291cmNlL0phdmFTY3JpcHRDb3JlL2J5dGVjb2RlL1B1dEJ5SWRTdGF0dXMu Y3BwIGIvU291cmNlL0phdmFTY3JpcHRDb3JlL2J5dGVjb2RlL1B1dEJ5SWRTdGF0dXMuY3BwCmlu ZGV4IDIzMjM4NGYuLjk2YjBmODIgMTAwNjQ0Ci0tLSBhL1NvdXJjZS9KYXZhU2NyaXB0Q29yZS9i eXRlY29kZS9QdXRCeUlkU3RhdHVzLmNwcAorKysgYi9Tb3VyY2UvSmF2YVNjcmlwdENvcmUvYnl0 ZWNvZGUvUHV0QnlJZFN0YXR1cy5jcHAKQEAgLTM1LDcgKzM1LDcgQEAKIAogbmFtZXNwYWNlIEpT QyB7CiAKLSNpZiBFTkFCTEUoSklUKQorI2lmIEVOQUJMRShERkdfSklUKQogYm9vbCBQdXRCeUlk U3RhdHVzOjpoYXNFeGl0U2l0ZShjb25zdCBDb25jdXJyZW50SklUTG9ja2VyJiBsb2NrZXIsIENv ZGVCbG9jayogcHJvZmlsZWRCbG9jaywgdW5zaWduZWQgYnl0ZWNvZGVJbmRleCwgRXhpdGluZ0pJ VFR5cGUgZXhpdFR5cGUpCiB7CiAgICAgcmV0dXJuIHByb2ZpbGVkQmxvY2stPmhhc0V4aXRTaXRl KGxvY2tlciwgREZHOjpGcmVxdWVudEV4aXRTaXRlKGJ5dGVjb2RlSW5kZXgsIEJhZENhY2hlLCBl eGl0VHlwZSkpCmRpZmYgLS1naXQgYS9Tb3VyY2UvSmF2YVNjcmlwdENvcmUvaml0L1JlcGF0Y2gu Y3BwIGIvU291cmNlL0phdmFTY3JpcHRDb3JlL2ppdC9SZXBhdGNoLmNwcAppbmRleCBkYjdjYjVi Li41N2ZlM2ExIDEwMDY0NAotLS0gYS9Tb3VyY2UvSmF2YVNjcmlwdENvcmUvaml0L1JlcGF0Y2gu Y3BwCisrKyBiL1NvdXJjZS9KYXZhU2NyaXB0Q29yZS9qaXQvUmVwYXRjaC5jcHAKQEAgLTgxNyw5 ICs4MTcsMTIgQEAgc3RhdGljIE1hY3JvQXNzZW1ibGVyOjpDYWxsIHdyaXRlQmFycmllcihDQ2Fs bEhlbHBlcnMmIGppdCwgR1BSUmVnIG93bmVyLCBHUFJSZWcKICAgICBBU1NFUlQob3duZXIgIT0g c2NyYXRjaDEpOwogICAgIEFTU0VSVChvd25lciAhPSBzY3JhdGNoMik7CiAKKyNpZiBFTkFCTEUo REZHX0pJVCkKICAgICBNYWNyb0Fzc2VtYmxlcjo6SnVtcCBkZWZpbml0ZWx5Tm90TWFya2VkID0g REZHOjpTcGVjdWxhdGl2ZUpJVDo6Z2VuZXJpY1dyaXRlQmFycmllcihqaXQsIG93bmVyLCBzY3Jh dGNoMSwgc2NyYXRjaDIpOwotICAgIE1hY3JvQXNzZW1ibGVyOjpDYWxsIGNhbGwgPSBzdG9yZVRv V3JpdGVCYXJyaWVyQnVmZmVyKGppdCwgb3duZXIsIHNjcmF0Y2gxLCBzY3JhdGNoMiwgYWxsb2Nh dG9yKTsKICAgICBkZWZpbml0ZWx5Tm90TWFya2VkLmxpbmsoJmppdCk7CisjZW5kaWYgLy8gRU5B QkxFKERGR19KSVQpCisKKyAgICBNYWNyb0Fzc2VtYmxlcjo6Q2FsbCBjYWxsID0gc3RvcmVUb1dy aXRlQmFycmllckJ1ZmZlcihqaXQsIG93bmVyLCBzY3JhdGNoMSwgc2NyYXRjaDIsIGFsbG9jYXRv cik7CiAgICAgcmV0dXJuIGNhbGw7CiB9CiAjZW5kaWYgLy8gRU5BQkxFKEdHQykK 225780 2014-03-04 08:16:08 -0800 2014-03-06 06:52:44 -0800 Proposed fix v2 proposedpatch.patch text/plain 1759 dbatyai.u-szeged ZGlmZiAtLWdpdCBhL1NvdXJjZS9KYXZhU2NyaXB0Q29yZS9DaGFuZ2VMb2cgYi9Tb3VyY2UvSmF2 YVNjcmlwdENvcmUvQ2hhbmdlTG9nCmluZGV4IGM3NGVlNzkuLjZjMDI2YTQgMTAwNjQ0Ci0tLSBh L1NvdXJjZS9KYXZhU2NyaXB0Q29yZS9DaGFuZ2VMb2cKKysrIGIvU291cmNlL0phdmFTY3JpcHRD b3JlL0NoYW5nZUxvZwpAQCAtMSwzICsxLDE0IEBACisyMDE0LTAzLTA0ICBEw6FuaWVsIELDoXR5 YWkgIDxkYmF0eWFpLnUtc3plZ2VkQHBhcnRuZXIuc2Ftc3VuZy5jb20+CisKKyAgICAgICAgRU5B QkxFRChKSVQpICYmICFFTkFCTEVEKERGR19KSVQpIGJyZWFrcyB0aGUgYnVpbGQKKyAgICAgICAg aHR0cHM6Ly9idWdzLndlYmtpdC5vcmcvc2hvd19idWcuY2dpP2lkPTEyODU0OQorCisgICAgICAg IFJldmlld2VkIGJ5IE5PQk9EWSAoT09QUyEpLgorCisgICAgICAgICogaml0L1JlcGF0Y2guY3Bw OgorICAgICAgICAoSlNDOjpnZW5lcmljV3JpdGVCYXJyaWVyKToKKyAgICAgICAgKEpTQzo6d3Jp dGVCYXJyaWVyKToKKwogMjAxNC0wMy0wMyAgSm9zZXBoIFBlY29yYXJvICA8cGVjb3Jhcm9AYXBw bGUuY29tPgogCiAgICAgICAgIFdlYiBJbnNwZWN0b3I6IEF2b2lkIHRvbyBlYXJseSBkZXJlZiBj YXVzZWQgYnkgUmVtb3RlSW5zcGVjdG9yWFBDQ29ubmVjdGlvbjo6Y2xvc2UKZGlmZiAtLWdpdCBh L1NvdXJjZS9KYXZhU2NyaXB0Q29yZS9qaXQvUmVwYXRjaC5jcHAgYi9Tb3VyY2UvSmF2YVNjcmlw dENvcmUvaml0L1JlcGF0Y2guY3BwCmluZGV4IDY1MjM0NmMuLjA5ZTRlMjAgMTAwNjQ0Ci0tLSBh L1NvdXJjZS9KYXZhU2NyaXB0Q29yZS9qaXQvUmVwYXRjaC5jcHAKKysrIGIvU291cmNlL0phdmFT Y3JpcHRDb3JlL2ppdC9SZXBhdGNoLmNwcApAQCAtODE5LDEyICs4MTksMTcgQEAgc3RhdGljIE1h Y3JvQXNzZW1ibGVyOjpDYWxsIHN0b3JlVG9Xcml0ZUJhcnJpZXJCdWZmZXIoQ0NhbGxIZWxwZXJz JiBqaXQsIEdQUlJlZwogICAgIHJldHVybiBjYWxsOwogfQogCitzdGF0aWMgTWFjcm9Bc3NlbWJs ZXI6Okp1bXAgZ2VuZXJpY1dyaXRlQmFycmllcihDQ2FsbEhlbHBlcnMmIGppdCwgR1BSUmVnIG93 bmVyKQoreworICAgIHJldHVybiBqaXQuYnJhbmNoVGVzdDgoTWFjcm9Bc3NlbWJsZXI6Olplcm8s IE1hY3JvQXNzZW1ibGVyOjpBZGRyZXNzKG93bmVyLCBKU0NlbGw6OmdjRGF0YU9mZnNldCgpKSk7 Cit9CisKIHN0YXRpYyBNYWNyb0Fzc2VtYmxlcjo6Q2FsbCB3cml0ZUJhcnJpZXIoQ0NhbGxIZWxw ZXJzJiBqaXQsIEdQUlJlZyBvd25lciwgR1BSUmVnIHNjcmF0Y2gxLCBHUFJSZWcgc2NyYXRjaDIs IFNjcmF0Y2hSZWdpc3RlckFsbG9jYXRvciYgYWxsb2NhdG9yKQogewogICAgIEFTU0VSVChvd25l ciAhPSBzY3JhdGNoMSk7CiAgICAgQVNTRVJUKG93bmVyICE9IHNjcmF0Y2gyKTsKIAotICAgIE1h Y3JvQXNzZW1ibGVyOjpKdW1wIGRlZmluaXRlbHlOb3RNYXJrZWQgPSBERkc6OlNwZWN1bGF0aXZl SklUOjpnZW5lcmljV3JpdGVCYXJyaWVyKGppdCwgb3duZXIpOworICAgIE1hY3JvQXNzZW1ibGVy OjpKdW1wIGRlZmluaXRlbHlOb3RNYXJrZWQgPSBnZW5lcmljV3JpdGVCYXJyaWVyKGppdCwgb3du ZXIpOwogICAgIE1hY3JvQXNzZW1ibGVyOjpDYWxsIGNhbGwgPSBzdG9yZVRvV3JpdGVCYXJyaWVy QnVmZmVyKGppdCwgb3duZXIsIHNjcmF0Y2gxLCBzY3JhdGNoMiwgYWxsb2NhdG9yKTsKICAgICBk ZWZpbml0ZWx5Tm90TWFya2VkLmxpbmsoJmppdCk7CiAgICAgcmV0dXJuIGNhbGw7Cg==