158619 2016-06-10 10:00:33 -0700 [JSC] Inline JSC::toInt32 to improve kraken 2016-06-13 09:51:05 -0700 1 1 1 Unclassified WebKit JavaScriptCore WebKit Nightly Build Unspecified Unspecified RESOLVED FIXED P2 Normal --- 1 ysuzuki ysuzuki commit-queue keith_miller mark.lam msaboff saam oldest_to_newest 1201257 0 ysuzuki 2016-06-10 10:00:33 -0700 Linux perf tool analysis on several kraken benchmarks shows that JSC::toInt32 is frequently called. But, the function JSC::toInt32 itself is small one. It's worth inlining. 1201261 1 ysuzuki 2016-06-10 10:10:24 -0700 Seems good. I'll check Octane, SunSpider, ASMBench and upload the patch. Benchmark report for Kraken on hanayamata. VMs tested: "baseline" at /home/yusukesuzuki/dev/WebKit/WebKitBuild/perf-master/Release/bin/jsc "patched" at /home/yusukesuzuki/dev/WebKit/WebKitBuild/perf/Release/bin/jsc Collected 100 samples per benchmark/VM, with 100 VM invocations per benchmark. Emitted a call to gc() between sample measurements. Used 1 benchmark iteration per VM invocation for warm-up. Used the jsc-specific preciseTime() function to get microsecond-level timing. Reporting benchmark execution times with 95% confidence intervals in milliseconds. baseline patched ai-astar 96.084+-0.548 96.015+-0.523 audio-beat-detection 48.621+-1.138 48.357+-1.020 audio-dft 123.803+-0.373 123.549+-0.109 audio-fft 37.399+-0.050 37.381+-0.039 audio-oscillator 53.682+-0.071 ? 53.718+-0.082 ? imaging-darkroom 87.673+-0.063 ? 87.724+-0.060 ? imaging-desaturate 56.400+-0.064 56.333+-0.041 imaging-gaussian-blur 79.772+-1.849 ? 79.936+-1.394 ? json-parse-financial 44.649+-0.208 ? 44.997+-0.302 ? json-stringify-tinderbox 27.110+-0.056 26.812+-0.428 might be 1.0111x faster stanford-crypto-aes 43.989+-0.157 ^ 43.001+-0.078 ^ definitely 1.0230x faster stanford-crypto-ccm 44.338+-0.688 ? 44.588+-0.686 ? stanford-crypto-pbkdf2 108.420+-1.044 ^ 106.231+-0.437 ^ definitely 1.0206x faster stanford-crypto-sha256-iterative 35.538+-0.077 ^ 35.314+-0.037 ^ definitely 1.0063x faster <arithmetic> 63.391+-0.168 63.140+-0.140 might be 1.0040x faster 1201491 2 281075 ysuzuki 2016-06-10 21:13:33 -0700 Created attachment 281075 Patch 1201492 3 ysuzuki 2016-06-10 21:16:47 -0700 Performance results. mandreel seems noise. To ensure that, we retake the mandreel 100 times, and the result is the following. Benchmark report for Octane on hanayamata. VMs tested: "baseline" at /home/yusukesuzuki/dev/WebKit/WebKitBuild/perf-master/Release/bin/jsc "patched" at /home/yusukesuzuki/dev/WebKit/WebKitBuild/perf/Release/bin/jsc Collected 100 samples per benchmark/VM, with 100 VM invocations per benchmark. Emitted a call to gc() between sample measurements. Used 1 benchmark iteration per VM invocation for warm-up. Used the jsc-specific preciseTime() function to get microsecond-level timing. Reporting benchmark execution times with 95% confidence intervals in milliseconds. baseline patched mandreel x2 49.08879+-0.28178 48.86048+-0.10656 <geometric> 49.08879+-0.28178 48.86048+-0.10656 might be 1.0047x faster Benchmark report for SunSpider, Octane, Kraken, and AsmBench on hanayamata. VMs tested: "baseline" at /home/yusukesuzuki/dev/WebKit/WebKitBuild/perf-master/Release/bin/jsc "patched" at /home/yusukesuzuki/dev/WebKit/WebKitBuild/perf/Release/bin/jsc Collected 30 samples per benchmark/VM, with 30 VM invocations per benchmark. Emitted a call to gc() between sample measurements. Used 1 benchmark iteration per VM invocation for warm-up. Used the jsc-specific preciseTime() function to get microsecond-level timing. Reporting benchmark execution times with 95% confidence intervals in milliseconds. baseline patched SunSpider: 3d-cube 6.1564+-0.2165 6.0395+-0.0198 might be 1.0194x faster 3d-morph 25.7981+-0.1100 ? 25.9357+-0.1825 ? 3d-raytrace 7.4267+-0.5293 6.8834+-0.2116 might be 1.0789x faster access-binary-trees 2.3713+-0.0980 ? 2.4013+-0.1151 ? might be 1.0127x slower access-fannkuch 6.9854+-0.3202 6.7167+-0.2763 might be 1.0400x faster access-nbody 2.8520+-0.0103 ? 2.9024+-0.0567 ? might be 1.0177x slower access-nsieve 3.2326+-0.0544 3.2281+-0.0483 bitops-3bit-bits-in-byte 1.2229+-0.0422 ? 1.2335+-0.0501 ? bitops-bits-in-byte 2.8118+-0.1010 2.7639+-0.0888 might be 1.0173x faster bitops-bitwise-and 2.0064+-0.0246 1.9841+-0.0082 might be 1.0112x faster bitops-nsieve-bits 3.5306+-0.2168 3.3985+-0.1914 might be 1.0389x faster controlflow-recursive 2.8202+-0.0975 ? 2.8671+-0.1323 ? might be 1.0166x slower crypto-aes 5.2740+-0.1032 ? 5.3000+-0.2027 ? crypto-md5 2.7913+-0.1429 2.7811+-0.1261 crypto-sha1 2.5246+-0.0264 ! 2.6164+-0.0640 ! definitely 1.0364x slower date-format-tofte 10.3640+-0.1394 ? 10.4484+-0.1300 ? date-format-xparb 6.0335+-0.1035 ? 6.0831+-0.1376 ? math-cordic 3.0874+-0.0865 ? 3.1688+-0.1189 ? might be 1.0264x slower math-partial-sums 10.5272+-0.0396 ? 10.5978+-0.1326 ? math-spectral-norm 2.1891+-0.0521 ? 2.2340+-0.0519 ? might be 1.0205x slower regexp-dna 7.3567+-0.0223 ? 7.3860+-0.0869 ? string-base64 3.9486+-0.0283 ? 4.0317+-0.0774 ? might be 1.0211x slower string-fasta 6.6015+-0.2154 6.4747+-0.1874 might be 1.0196x faster string-tagcloud 9.7379+-0.0710 9.6021+-0.1486 might be 1.0141x faster string-unpack-code 19.7744+-0.0478 ? 19.8322+-0.2298 ? string-validate-input 4.2980+-0.1012 ? 4.3578+-0.1089 ? might be 1.0139x slower <arithmetic> 6.2201+-0.0308 6.2026+-0.0274 might be 1.0028x faster baseline patched Octane: encrypt 0.17962+-0.00241 ? 0.18033+-0.00120 ? decrypt 2.98068+-0.01640 ? 3.00958+-0.04567 ? deltablue x2 0.15236+-0.00144 0.15025+-0.00117 might be 1.0141x faster earley 0.33712+-0.00074 ? 0.33722+-0.00079 ? boyer 5.32422+-0.00564 ? 5.32558+-0.00841 ? navier-stokes x2 4.78881+-0.00693 ? 4.79320+-0.01528 ? raytrace x2 0.93551+-0.00362 0.93233+-0.00309 richards x2 0.09723+-0.00092 ? 0.09801+-0.00167 ? splay x2 0.39334+-0.00123 ? 0.39364+-0.00119 ? regexp x2 18.71392+-0.06857 18.65126+-0.05944 pdfjs x2 42.39686+-0.35091 42.31642+-0.30435 mandreel x2 48.68664+-0.15582 ! 49.25694+-0.29339 ! definitely 1.0117x slower gbemu x2 39.31304+-1.31894 38.79220+-1.50068 might be 1.0134x faster closure 0.60327+-0.00395 0.60178+-0.00309 jquery 8.02553+-0.02883 ? 8.08426+-0.14237 ? box2d x2 14.63042+-0.09236 ? 14.69277+-0.19818 ? zlib x2 368.45610+-5.24313 364.03090+-2.28827 might be 1.0122x faster typescript x2 810.87181+-4.19509 810.07914+-4.30712 <geometric> 5.90654+-0.01651 5.90027+-0.01901 might be 1.0011x faster baseline patched Kraken: ai-astar 95.966+-1.109 ? 96.090+-0.925 ? audio-beat-detection 47.492+-1.701 46.657+-1.232 might be 1.0179x faster audio-dft 123.721+-0.402 ? 124.373+-0.666 ? audio-fft 37.322+-0.036 37.276+-0.015 audio-oscillator 53.789+-0.285 53.722+-0.157 imaging-darkroom 88.159+-0.917 87.730+-0.094 imaging-desaturate 56.332+-0.064 ? 56.345+-0.104 ? imaging-gaussian-blur 80.524+-2.687 79.091+-2.555 might be 1.0181x faster json-parse-financial 44.610+-0.441 ? 44.791+-0.325 ? json-stringify-tinderbox 27.087+-0.035 ^ 26.609+-0.144 ^ definitely 1.0180x faster stanford-crypto-aes 43.669+-0.210 ^ 42.862+-0.115 ^ definitely 1.0188x faster stanford-crypto-ccm 45.213+-1.424 44.490+-1.290 might be 1.0162x faster stanford-crypto-pbkdf2 107.665+-0.581 ^ 106.229+-0.807 ^ definitely 1.0135x faster stanford-crypto-sha256-iterative 35.470+-0.112 35.351+-0.078 <arithmetic> 63.359+-0.230 62.972+-0.281 might be 1.0061x faster baseline patched AsmBench: towers.c 280.7607+-10.5500 276.5852+-2.2406 might be 1.0151x faster n-body.c 930.6878+-6.5519 924.9891+-6.5493 float-mm.c 718.3752+-4.3448 ? 723.1201+-5.6769 ? container.cpp 3029.7668+-37.4858 ? 3048.7741+-25.0556 ? quicksort.c 439.4318+-1.6382 439.2427+-1.8137 gcc-loops.cpp 4163.1676+-18.3908 ? 4175.9266+-52.8059 ? bigfib.cpp 456.7496+-6.6840 ? 457.0188+-5.5533 ? hash-map 153.3622+-1.5466 ? 153.5558+-1.7462 ? dry.c 478.1047+-12.4209 477.8930+-9.9328 <geometric> 688.1226+-3.0102 688.0901+-2.4242 might be 1.0000x faster baseline patched Geomean of preferred means: <scaled-result> 35.5738+-0.0610 35.4851+-0.0812 might be 1.0025x faster 1201493 4 281075 ysuzuki 2016-06-10 21:33:39 -0700 Comment on attachment 281075 Patch View in context: https://bugs.webkit.org/attachment.cgi?id=281075&action=review > Source/JavaScriptCore/runtime/JSCJSValueInlines.h:1062 > +ALWAYS_INLINE int32_t toInt32(double number) It may be good to change this ALWAYS_INLINE to inline to rely on the compiler's decision. I'll check this. 1201496 5 281075 ysuzuki 2016-06-10 21:57:48 -0700 Comment on attachment 281075 Patch View in context: https://bugs.webkit.org/attachment.cgi?id=281075&action=review >> Source/JavaScriptCore/runtime/JSCJSValueInlines.h:1062 >> +ALWAYS_INLINE int32_t toInt32(double number) > > It may be good to change this ALWAYS_INLINE to inline to rely on the compiler's decision. I'll check this. Performance results recommend us to use ALWAYS_INLINE for this toInt32 function. So use ALWAYS_INLINE. Benchmark report for Kraken on hanayamata. VMs tested: "baseline" at /home/yusukesuzuki/dev/WebKit/WebKitBuild/perf-master/Release/bin/jsc "patched" at /home/yusukesuzuki/dev/WebKit/WebKitBuild/perf/Release/bin/jsc "always" at /home/yusukesuzuki/dev/WebKit/WebKitBuild/perf2/Release/bin/jsc Collected 100 samples per benchmark/VM, with 100 VM invocations per benchmark. Emitted a call to gc() between sample measurements. Used 1 benchmark iteration per VM invocation for warm-up. Used the jsc-specific preciseTime() function to get microsecond-level timing. Reporting benchmark execution times with 95% confidence intervals in milliseconds. baseline patched always always v. baseline stanford-crypto-pbkdf2 107.863+-0.376 ^ 106.828+-0.409 106.087+-0.450 ^ definitely 1.0167x faster <arithmetic> 107.863+-0.376 ^ 106.828+-0.409 106.087+-0.450 ^ definitely 1.0167x faster 1201501 6 281075 mark.lam 2016-06-10 22:06:46 -0700 Comment on attachment 281075 Patch View in context: https://bugs.webkit.org/attachment.cgi?id=281075&action=review r=me with fix. >>> Source/JavaScriptCore/runtime/JSCJSValueInlines.h:1062 >>> +ALWAYS_INLINE int32_t toInt32(double number) >> >> It may be good to change this ALWAYS_INLINE to inline to rely on the compiler's decision. I'll check this. > > Performance results recommend us to use ALWAYS_INLINE for this toInt32 function. So use ALWAYS_INLINE. > > Benchmark report for Kraken on hanayamata. > > VMs tested: > "baseline" at /home/yusukesuzuki/dev/WebKit/WebKitBuild/perf-master/Release/bin/jsc > "patched" at /home/yusukesuzuki/dev/WebKit/WebKitBuild/perf/Release/bin/jsc > "always" at /home/yusukesuzuki/dev/WebKit/WebKitBuild/perf2/Release/bin/jsc > > Collected 100 samples per benchmark/VM, with 100 VM invocations per benchmark. Emitted a call to gc() between sample measurements. > Used 1 benchmark iteration per VM invocation for warm-up. Used the jsc-specific preciseTime() function to get microsecond-level > timing. Reporting benchmark execution times with 95% confidence intervals in milliseconds. > > baseline patched always always v. baseline > > stanford-crypto-pbkdf2 107.863+-0.376 ^ 106.828+-0.409 106.087+-0.450 ^ definitely 1.0167x faster > > <arithmetic> 107.863+-0.376 ^ 106.828+-0.409 106.087+-0.450 ^ definitely 1.0167x faster I know that this function came from JSCJSValue.cpp, but I think a better place to put it is in runtime/MathCommon.h since it doesn't have anything to do with JSValues. 1201502 7 281075 ysuzuki 2016-06-10 22:10:02 -0700 Comment on attachment 281075 Patch View in context: https://bugs.webkit.org/attachment.cgi?id=281075&action=review >>>> Source/JavaScriptCore/runtime/JSCJSValueInlines.h:1062 >>>> +ALWAYS_INLINE int32_t toInt32(double number) >>> >>> It may be good to change this ALWAYS_INLINE to inline to rely on the compiler's decision. I'll check this. >> >> Performance results recommend us to use ALWAYS_INLINE for this toInt32 function. So use ALWAYS_INLINE. >> >> Benchmark report for Kraken on hanayamata. >> >> VMs tested: >> "baseline" at /home/yusukesuzuki/dev/WebKit/WebKitBuild/perf-master/Release/bin/jsc >> "patched" at /home/yusukesuzuki/dev/WebKit/WebKitBuild/perf/Release/bin/jsc >> "always" at /home/yusukesuzuki/dev/WebKit/WebKitBuild/perf2/Release/bin/jsc >> >> Collected 100 samples per benchmark/VM, with 100 VM invocations per benchmark. Emitted a call to gc() between sample measurements. >> Used 1 benchmark iteration per VM invocation for warm-up. Used the jsc-specific preciseTime() function to get microsecond-level >> timing. Reporting benchmark execution times with 95% confidence intervals in milliseconds. >> >> baseline patched always always v. baseline >> >> stanford-crypto-pbkdf2 107.863+-0.376 ^ 106.828+-0.409 106.087+-0.450 ^ definitely 1.0167x faster >> >> <arithmetic> 107.863+-0.376 ^ 106.828+-0.409 106.087+-0.450 ^ definitely 1.0167x faster > > I know that this function came from JSCJSValue.cpp, but I think a better place to put it is in runtime/MathCommon.h since it doesn't have anything to do with JSValues. Right! I'll move this to MathCommon.h (And that is also the better place to put operationToInt32 since operationMathPow resides there). 1201508 8 281075 ysuzuki 2016-06-10 22:29:10 -0700 Comment on attachment 281075 Patch View in context: https://bugs.webkit.org/attachment.cgi?id=281075&action=review >>>>> Source/JavaScriptCore/runtime/JSCJSValueInlines.h:1062 >>>>> +ALWAYS_INLINE int32_t toInt32(double number) >>>> >>>> It may be good to change this ALWAYS_INLINE to inline to rely on the compiler's decision. I'll check this. >>> >>> Performance results recommend us to use ALWAYS_INLINE for this toInt32 function. So use ALWAYS_INLINE. >>> >>> Benchmark report for Kraken on hanayamata. >>> >>> VMs tested: >>> "baseline" at /home/yusukesuzuki/dev/WebKit/WebKitBuild/perf-master/Release/bin/jsc >>> "patched" at /home/yusukesuzuki/dev/WebKit/WebKitBuild/perf/Release/bin/jsc >>> "always" at /home/yusukesuzuki/dev/WebKit/WebKitBuild/perf2/Release/bin/jsc >>> >>> Collected 100 samples per benchmark/VM, with 100 VM invocations per benchmark. Emitted a call to gc() between sample measurements. >>> Used 1 benchmark iteration per VM invocation for warm-up. Used the jsc-specific preciseTime() function to get microsecond-level >>> timing. Reporting benchmark execution times with 95% confidence intervals in milliseconds. >>> >>> baseline patched always always v. baseline >>> >>> stanford-crypto-pbkdf2 107.863+-0.376 ^ 106.828+-0.409 106.087+-0.450 ^ definitely 1.0167x faster >>> >>> <arithmetic> 107.863+-0.376 ^ 106.828+-0.409 106.087+-0.450 ^ definitely 1.0167x faster >> >> I know that this function came from JSCJSValue.cpp, but I think a better place to put it is in runtime/MathCommon.h since it doesn't have anything to do with JSValues. > > Right! I'll move this to MathCommon.h (And that is also the better place to put operationToInt32 since operationMathPow resides there). Looking into MathCommon, it includes several headers to gain some define flags. But I think these flags should be moved to WTF/Compiler.h. I'll create a new patch later. 1201509 9 ysuzuki 2016-06-10 22:30:20 -0700 Committed r201964: <http://trac.webkit.org/changeset/201964> 1201516 10 ysuzuki 2016-06-10 23:00:20 -0700 Committed r201966: <http://trac.webkit.org/changeset/201966> 1201611 11 281075 darin 2016-06-11 12:35:02 -0700 Comment on attachment 281075 Patch View in context: https://bugs.webkit.org/attachment.cgi?id=281075&action=review > Source/JavaScriptCore/runtime/JSCJSValueInlines.h:1057 > +// of the resulting bit-pattern (as such this metod is also called to implement Typo here "metod". > Source/JavaScriptCore/runtime/JSCJSValueInlines.h:1071 > + // Note this case handles 0, -0, and all infinte, NaN, & denormal value. Typo here "infinte" 1201865 12 281075 ysuzuki 2016-06-13 09:47:28 -0700 Comment on attachment 281075 Patch View in context: https://bugs.webkit.org/attachment.cgi?id=281075&action=review Thanks! I'll land the follow up patch that fixes pointed typos. >> Source/JavaScriptCore/runtime/JSCJSValueInlines.h:1057 >> +// of the resulting bit-pattern (as such this metod is also called to implement > > Typo here "metod". Thanks! Fixed in the follow up patch. >> Source/JavaScriptCore/runtime/JSCJSValueInlines.h:1071 >> + // Note this case handles 0, -0, and all infinte, NaN, & denormal value. > > Typo here "infinte" Fixed. 1201866 13 ysuzuki 2016-06-13 09:51:05 -0700 Committed r201993: <http://trac.webkit.org/changeset/201993> 281075 2016-06-10 21:13:33 -0700 2016-06-10 22:06:46 -0700 Patch bug-158619-20160611131503.patch text/plain 13604 ysuzuki U3VidmVyc2lvbiBSZXZpc2lvbjogMjAxOTE0CmRpZmYgLS1naXQgYS9Tb3VyY2UvSmF2YVNjcmlw dENvcmUvQ2hhbmdlTG9nIGIvU291cmNlL0phdmFTY3JpcHRDb3JlL0NoYW5nZUxvZwppbmRleCBm MzMzNjhiZjljMjcyNjE4MWQxOGIwNWIyNmY0ZmIyYmUyOGRmZmJiLi5lMmVhMTg5ODBjYmIzYmJk YmI0YTg1ODFiMjgzNjIzMTRkYWYyMTk4IDEwMDY0NAotLS0gYS9Tb3VyY2UvSmF2YVNjcmlwdENv cmUvQ2hhbmdlTG9nCisrKyBiL1NvdXJjZS9KYXZhU2NyaXB0Q29yZS9DaGFuZ2VMb2cKQEAgLTEs MyArMSw1MiBAQAorMjAxNi0wNi0xMCAgWXVzdWtlIFN1enVraSAgPHV0YXRhbmUudGVhQGdtYWls LmNvbT4KKworICAgICAgICBbSlNDXSBJbmxpbmUgSlNDOjp0b0ludDMyIHRvIGltcHJvdmUga3Jh a2VuCisgICAgICAgIGh0dHBzOi8vYnVncy53ZWJraXQub3JnL3Nob3dfYnVnLmNnaT9pZD0xNTg2 MTkKKworICAgICAgICBSZXZpZXdlZCBieSBOT0JPRFkgKE9PUFMhKS4KKworICAgICAgICBTZXZl cmFsIGtyYWtlbiBiZW5jaG1hcmtzIHNob3cgdGhhdCBKU0M6OnRvSW50MzIgaXMgZnJlcXVlbnRs eSBjYWxsZWQuCisgICAgICAgIEZvciBleGFtcGxlLCBzdGFuZm9yZC1jcnlwdG8tcGJrZGYyIHJl cG9ydHMgdGhhdCB0aGUgaG90dGVzdCBydW50aW1lIGZ1bmN0aW9uIGlzIEpTQzo6dG9JbnQzMi4K KworICAgICAgICBUaGUgZGF0YSBpcyBiZWxvdyAodGFrZW4gYnkgTGludXggcGVyZiB0b29scyku CisgICAgICAgIDUuNTAlICBqc2MgICAgICBsaWJKYXZhU2NyaXB0Q29yZS5zby4xLjAuMCAgWy5d IF9aTjNKU0M3dG9JbnQzMkVkCisgICAgICAgIDMuOTYlICBqc2MgICAgICBsaWJKYXZhU2NyaXB0 Q29yZS5zby4xLjAuMCAgWy5dIF9aTjNKU0MyMGFycmF5UHJvdG9GdW5jQ29uY2F0RVBOU185RXhl Y1N0YXRlRQorICAgICAgICAyLjQ4JSAganNjICAgICAgbGliSmF2YVNjcmlwdENvcmUuc28uMS4w LjAgIFsuXSBfWk4zSlNDMTlhcnJheVByb3RvRnVuY1NsaWNlRVBOU185RXhlY1N0YXRlRQorICAg ICAgICAxLjY5JSAganNjICAgICAgbGliSmF2YVNjcmlwdENvcmUuc28uMS4wLjAgIFsuXSBfWk5L M0pTQzlTdHJ1Y3R1cmUyN2hvbGVzTXVzdEZvcndhcmRUb1Byb3RvdHlwZUVSTlNfMlZNRQorCisg ICAgICAgIFRoaXMgaXMgYmVjYXVzZSBvZiBDb21tb25TbG93UGF0aHMnIGJpdCBvcGVyYXRpb25z J3MgSlNWYWx1ZTo6dG9JbnQzMi4KKyAgICAgICAgRHVlIHRvIHRoZSBzbG93IHBhdGgsIGluIGB2 YWx1ZSB8IDBgLCBgdmFsdWVgIG1heSBiZSBhIGRvdWJsZSBudW1iZXIgdmFsdWUuIEluIHRoYXQg Y2FzZSwgSlNDOjp0b0ludDMyIGlzIGNhbGxlZC4KKworICAgICAgICBXaGlsZSBKU0M6OnRvSW4z MiBpcyBob3QsIHRoZSBmdW5jdGlvbiBpdHNlbGYgaXMgdmVyeSBzbWFsbC4gSXQncyB3b3J0aCBp bmxpbmluZy4KKworICAgICAgICBUaGlzIGNoYW5nZSBvZmZlcnMgdGhlIGZvbGxvd2luZyBrcmFr ZW4gaW1wcm92ZW1lbnRzLgorCisgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAg ICAgICAgICAgICAgICAgICAgICBiYXNlbGluZSAgICAgICAgICAgICAgICAgIHBhdGNoZWQKKyAg ICAgICAgS3Jha2VuOgorICAgICAgICAgICBhdWRpby1iZWF0LWRldGVjdGlvbiAgICAgICAgICAg ICAgICAgICAgICAgNDcuNDkyKy0xLjcwMSAgICAgICAgICAgICA0Ni42NTcrLTEuMjMyICAgICAg ICAgICBtaWdodCBiZSAxLjAxNzl4IGZhc3RlcgorICAgICAgICAgICBzdGFuZm9yZC1jcnlwdG8t YWVzICAgICAgICAgICAgICAgICAgICAgICAgNDMuNjY5Ky0wLjIxMCAgICAgIF4gICAgICA0Mi44 NjIrLTAuMTE1ICAgICAgICAgXiBkZWZpbml0ZWx5IDEuMDE4OHggZmFzdGVyCisgICAgICAgICAg IHN0YW5mb3JkLWNyeXB0by1jY20gICAgICAgICAgICAgICAgICAgICAgICA0NS4yMTMrLTEuNDI0 ICAgICAgICAgICAgIDQ0LjQ5MCstMS4yOTAgICAgICAgICAgIG1pZ2h0IGJlIDEuMDE2MnggZmFz dGVyCisgICAgICAgICAgIHN0YW5mb3JkLWNyeXB0by1wYmtkZjIgICAgICAgICAgICAgICAgICAg IDEwNy42NjUrLTAuNTgxICAgICAgXiAgICAgMTA2LjIyOSstMC44MDcgICAgICAgICBeIGRlZmlu aXRlbHkgMS4wMTM1eCBmYXN0ZXIKKworICAgICAgICBUaGlzIHBhdGNoIG9ubHkgZm9jdXNlZCBv biB0aGUgY2FsbCB0byB0b0ludDMyIGZyb20gdGhlIHJ1bnRpbWUgZnVuY3Rpb25zLgorICAgICAg ICBTbyBKU0M6OnRvSW50MzIgY2FsbHMgZnJvbSB0aGUgYmFzZWxpbmUgLyBERkcgcmVtYWluLgor ICAgICAgICBXZSBlbnN1cmUgdGhhdCBKSVQgY29kZSB1c2VzIG9wZXJhdGlvblRvSW50MzIgaW5z dGVhZCBvZiBKU0M6OnRvSW50MzIgc2luY2UgSlNDOjp0b0ludDMyIGlzIG5vdyBtYXJrZWQgYXMg QUxXQVlTX0lOTElORS4KKyAgICAgICAgTGludXggcGVyZiBwcm9maWxlciBhbHNvIGZpbmRzIHRo YXQgdGhpcyBgb3BlcmF0aW9uVG9JbnQzMmAgaXMgZnJlcXVlbnRseSBjYWxsZWQgaW4gdGhlIGFi b3ZlIGJlbmNobWFya3MuCisgICAgICAgIEl0IG1heSBiZSBnb29kIHRvIGludHJvZHVjZSBhc20g ZW1pdCBmb3IgdGhhdCBpbnN0ZWFkIG9mIGNhbGxpbmcgSlNDOjp0b0ludDMyIG9wZXJhdGlvbiBp biB0aGUgc2VwYXJhdGVkIHBhdGNoLgorCisgICAgICAgICogZGZnL0RGR1NwZWN1bGF0aXZlSklU LmNwcDoKKyAgICAgICAgKEpTQzo6REZHOjpTcGVjdWxhdGl2ZUpJVDo6Y29tcGlsZVZhbHVlVG9J bnQzMik6CisgICAgICAgIChKU0M6OkRGRzo6U3BlY3VsYXRpdmVKSVQ6OmNvbXBpbGVQdXRCeVZh bEZvckludFR5cGVkQXJyYXkpOgorICAgICAgICAqIGZ0bC9GVExMb3dlckRGR1RvQjMuY3BwOgor ICAgICAgICAoSlNDOjpGVEw6OkRGRzo6TG93ZXJERkdUb0IzOjpkb3VibGVUb0ludDMyKToKKyAg ICAgICAgKEpTQzo6RlRMOjpERkc6Okxvd2VyREZHVG9CMzo6c2Vuc2libGVEb3VibGVUb0ludDMy KToKKyAgICAgICAgKiBqaXQvSklUT3BlcmF0aW9ucy5jcHA6CisgICAgICAgICogaml0L0pJVE9w ZXJhdGlvbnMuaDoKKyAgICAgICAgKiBydW50aW1lL0pTQ0pTVmFsdWUuY3BwOgorICAgICAgICAo SlNDOjp0b0ludDMyKTogRGVsZXRlZC4KKyAgICAgICAgKiBydW50aW1lL0pTQ0pTVmFsdWVJbmxp bmVzLmg6CisgICAgICAgIChKU0M6OnRvSW50MzIpOgorCiAyMDE2LTA2LTA5ICBBbGV4IENocmlz dGVuc2VuICA8YWNocmlzdGVuc2VuQHdlYmtpdC5vcmc+CiAKICAgICAgICAgQ2xlYW4gdXAgSmF2 YVNjcmlwdENvcmUudmN4cHJvaiBkaXJlY3RvcnkgYWZ0ZXIgc3dpdGNoaW5nIHRvIENNYWtlLgpk aWZmIC0tZ2l0IGEvU291cmNlL0phdmFTY3JpcHRDb3JlL2RmZy9ERkdTcGVjdWxhdGl2ZUpJVC5j cHAgYi9Tb3VyY2UvSmF2YVNjcmlwdENvcmUvZGZnL0RGR1NwZWN1bGF0aXZlSklULmNwcAppbmRl eCAwNTgxYWNkMjM4MDE5MmI0ZGE2ZDg2MDU2OGFkNzFjZjVlNDZkMmIzLi4wZmZiMmI1NTViMDkw ZGU2NjA1OTAxNmYxNTk1ZjY4ODQ0YjNhNDIzIDEwMDY0NAotLS0gYS9Tb3VyY2UvSmF2YVNjcmlw dENvcmUvZGZnL0RGR1NwZWN1bGF0aXZlSklULmNwcAorKysgYi9Tb3VyY2UvSmF2YVNjcmlwdENv cmUvZGZnL0RGR1NwZWN1bGF0aXZlSklULmNwcApAQCAtMjExNCw3ICsyMTE0LDcgQEAgdm9pZCBT cGVjdWxhdGl2ZUpJVDo6Y29tcGlsZVZhbHVlVG9JbnQzMihOb2RlKiBub2RlKQogICAgICAgICBH UFJSZWcgZ3ByID0gcmVzdWx0LmdwcigpOwogICAgICAgICBKSVRDb21waWxlcjo6SnVtcCBub3RU cnVuY2F0ZWRUb0ludGVnZXIgPSBtX2ppdC5icmFuY2hUcnVuY2F0ZURvdWJsZVRvSW50MzIoZnBy LCBncHIsIEpJVENvbXBpbGVyOjpCcmFuY2hJZlRydW5jYXRlRmFpbGVkKTsKICAgICAgICAgCi0g ICAgICAgIGFkZFNsb3dQYXRoR2VuZXJhdG9yKHNsb3dQYXRoQ2FsbChub3RUcnVuY2F0ZWRUb0lu dGVnZXIsIHRoaXMsIHRvSW50MzIsIGdwciwgZnByLCBOZWVkVG9TcGlsbCwgRXhjZXB0aW9uQ2hl Y2tSZXF1aXJlbWVudDo6Q2hlY2tOb3ROZWVkZWQpKTsKKyAgICAgICAgYWRkU2xvd1BhdGhHZW5l cmF0b3Ioc2xvd1BhdGhDYWxsKG5vdFRydW5jYXRlZFRvSW50ZWdlciwgdGhpcywgb3BlcmF0aW9u VG9JbnQzMiwgZ3ByLCBmcHIsIE5lZWRUb1NwaWxsLCBFeGNlcHRpb25DaGVja1JlcXVpcmVtZW50 OjpDaGVja05vdE5lZWRlZCkpOwogICAgICAgICAKICAgICAgICAgaW50MzJSZXN1bHQoZ3ByLCBu b2RlKTsKICAgICAgICAgcmV0dXJuOwpAQCAtMjE2NSw3ICsyMTY1LDcgQEAgdm9pZCBTcGVjdWxh dGl2ZUpJVDo6Y29tcGlsZVZhbHVlVG9JbnQzMihOb2RlKiBub2RlKQogICAgICAgICAgICAgdW5i b3hEb3VibGUoZ3ByLCByZXN1bHRHcHIsIGZwcik7CiAKICAgICAgICAgICAgIHNpbGVudFNwaWxs QWxsUmVnaXN0ZXJzKHJlc3VsdEdwcik7Ci0gICAgICAgICAgICBjYWxsT3BlcmF0aW9uKHRvSW50 MzIsIHJlc3VsdEdwciwgZnByKTsKKyAgICAgICAgICAgIGNhbGxPcGVyYXRpb24ob3BlcmF0aW9u VG9JbnQzMiwgcmVzdWx0R3ByLCBmcHIpOwogICAgICAgICAgICAgc2lsZW50RmlsbEFsbFJlZ2lz dGVycyhyZXN1bHRHcHIpOwogCiAgICAgICAgICAgICBjb252ZXJ0ZWQuYXBwZW5kKG1faml0Lmp1 bXAoKSk7CkBAIC0yMjI0LDcgKzIyMjQsNyBAQCB2b2lkIFNwZWN1bGF0aXZlSklUOjpjb21waWxl VmFsdWVUb0ludDMyKE5vZGUqIG5vZGUpCiAgICAgICAgICAgICAgICAgdW5ib3hEb3VibGUodGFn R1BSLCBwYXlsb2FkR1BSLCBmcHIsIHNjcmF0Y2guZnByKCkpOwogCiAgICAgICAgICAgICAgICAg c2lsZW50U3BpbGxBbGxSZWdpc3RlcnMocmVzdWx0R3ByKTsKLSAgICAgICAgICAgICAgICBjYWxs T3BlcmF0aW9uKHRvSW50MzIsIHJlc3VsdEdwciwgZnByKTsKKyAgICAgICAgICAgICAgICBjYWxs T3BlcmF0aW9uKG9wZXJhdGlvblRvSW50MzIsIHJlc3VsdEdwciwgZnByKTsKICAgICAgICAgICAg ICAgICBzaWxlbnRGaWxsQWxsUmVnaXN0ZXJzKHJlc3VsdEdwcik7CiAKICAgICAgICAgICAgICAg ICBjb252ZXJ0ZWQuYXBwZW5kKG1faml0Lmp1bXAoKSk7CkBAIC0yNzY0LDcgKzI3NjQsNyBAQCB2 b2lkIFNwZWN1bGF0aXZlSklUOjpjb21waWxlUHV0QnlWYWxGb3JJbnRUeXBlZEFycmF5KEdQUlJl ZyBiYXNlLCBHUFJSZWcgcHJvcGVydAogICAgICAgICAgICAgICAgIE1hY3JvQXNzZW1ibGVyOjpK dW1wIGZhaWxlZCA9IG1faml0LmJyYW5jaFRydW5jYXRlRG91YmxlVG9JbnQzMigKICAgICAgICAg ICAgICAgICAgICAgZnByLCBncHIsIE1hY3JvQXNzZW1ibGVyOjpCcmFuY2hJZlRydW5jYXRlRmFp bGVkKTsKICAgICAgICAgICAgICAgICAKLSAgICAgICAgICAgICAgICBhZGRTbG93UGF0aEdlbmVy YXRvcihzbG93UGF0aENhbGwoZmFpbGVkLCB0aGlzLCB0b0ludDMyLCBncHIsIGZwciwgTmVlZFRv U3BpbGwsIEV4Y2VwdGlvbkNoZWNrUmVxdWlyZW1lbnQ6OkNoZWNrTm90TmVlZGVkKSk7CisgICAg ICAgICAgICAgICAgYWRkU2xvd1BhdGhHZW5lcmF0b3Ioc2xvd1BhdGhDYWxsKGZhaWxlZCwgdGhp cywgb3BlcmF0aW9uVG9JbnQzMiwgZ3ByLCBmcHIsIE5lZWRUb1NwaWxsLCBFeGNlcHRpb25DaGVj a1JlcXVpcmVtZW50OjpDaGVja05vdE5lZWRlZCkpOwogICAgICAgICAgICAgICAgIAogICAgICAg ICAgICAgICAgIGZpeGVkLmxpbmsoJm1faml0KTsKICAgICAgICAgICAgICAgICB2YWx1ZS5hZG9w dChyZXN1bHQpOwpkaWZmIC0tZ2l0IGEvU291cmNlL0phdmFTY3JpcHRDb3JlL2Z0bC9GVExMb3dl ckRGR1RvQjMuY3BwIGIvU291cmNlL0phdmFTY3JpcHRDb3JlL2Z0bC9GVExMb3dlckRGR1RvQjMu Y3BwCmluZGV4IGVhY2YyNGE2MGMzMjk2NDUxODFiODRlNDZkYzM1MDllOGZlYmY5Y2UuLmEzNmY0 ZTBjNTg5YzcyMWZhOGVjYzY2NDM0OTg3MmUzZDc2YWZmY2MgMTAwNjQ0Ci0tLSBhL1NvdXJjZS9K YXZhU2NyaXB0Q29yZS9mdGwvRlRMTG93ZXJERkdUb0IzLmNwcAorKysgYi9Tb3VyY2UvSmF2YVNj cmlwdENvcmUvZnRsL0ZUTExvd2VyREZHVG9CMy5jcHAKQEAgLTkwOTEsNyArOTA5MSw3IEBAIGNs YXNzIExvd2VyREZHVG9CMyB7CiAgICAgICAgIG1fb3V0Lmp1bXAoY29udGludWF0aW9uKTsKICAg ICAgICAgCiAgICAgICAgIG1fb3V0LmFwcGVuZFRvKHNsb3dQYXRoLCBjb250aW51YXRpb24pOwot ICAgICAgICByZXN1bHRzLmFwcGVuZChtX291dC5hbmNob3IobV9vdXQuY2FsbChtX291dC5pbnQz MiwgbV9vdXQub3BlcmF0aW9uKHRvSW50MzIpLCBkb3VibGVWYWx1ZSkpKTsKKyAgICAgICAgcmVz dWx0cy5hcHBlbmQobV9vdXQuYW5jaG9yKG1fb3V0LmNhbGwobV9vdXQuaW50MzIsIG1fb3V0Lm9w ZXJhdGlvbihvcGVyYXRpb25Ub0ludDMyKSwgZG91YmxlVmFsdWUpKSk7CiAgICAgICAgIG1fb3V0 Lmp1bXAoY29udGludWF0aW9uKTsKICAgICAgICAgCiAgICAgICAgIG1fb3V0LmFwcGVuZFRvKGNv bnRpbnVhdGlvbiwgbGFzdE5leHQpOwpAQCAtOTEyMCw3ICs5MTIwLDcgQEAgY2xhc3MgTG93ZXJE RkdUb0IzIHsKICAgICAgICAgCiAgICAgICAgIExCYXNpY0Jsb2NrIGxhc3ROZXh0ID0gbV9vdXQu YXBwZW5kVG8oc2xvd1BhdGgsIGNvbnRpbnVhdGlvbik7CiAgICAgICAgIFZhbHVlRnJvbUJsb2Nr IHNsb3dSZXN1bHQgPSBtX291dC5hbmNob3IoCi0gICAgICAgICAgICBtX291dC5jYWxsKG1fb3V0 LmludDMyLCBtX291dC5vcGVyYXRpb24odG9JbnQzMiksIGRvdWJsZVZhbHVlKSk7CisgICAgICAg ICAgICBtX291dC5jYWxsKG1fb3V0LmludDMyLCBtX291dC5vcGVyYXRpb24ob3BlcmF0aW9uVG9J bnQzMiksIGRvdWJsZVZhbHVlKSk7CiAgICAgICAgIG1fb3V0Lmp1bXAoY29udGludWF0aW9uKTsK ICAgICAgICAgCiAgICAgICAgIG1fb3V0LmFwcGVuZFRvKGNvbnRpbnVhdGlvbiwgbGFzdE5leHQp OwpkaWZmIC0tZ2l0IGEvU291cmNlL0phdmFTY3JpcHRDb3JlL2ppdC9KSVRPcGVyYXRpb25zLmNw cCBiL1NvdXJjZS9KYXZhU2NyaXB0Q29yZS9qaXQvSklUT3BlcmF0aW9ucy5jcHAKaW5kZXggMTc0 NWQ2ZTIzN2NjMDA5MmQzNjZkM2YxN2IwZTg1YzY0MTZmZWJhMS4uOGNkZDViMGQ0ZDY1NzI5MDJh MWJmMmNkZmI3YjE0MTRkMjEzMTJjMSAxMDA2NDQKLS0tIGEvU291cmNlL0phdmFTY3JpcHRDb3Jl L2ppdC9KSVRPcGVyYXRpb25zLmNwcAorKysgYi9Tb3VyY2UvSmF2YVNjcmlwdENvcmUvaml0L0pJ VE9wZXJhdGlvbnMuY3BwCkBAIC0yMzQyLDYgKzIzNDIsMTEgQEAgaW50MzJfdCBKSVRfT1BFUkFU SU9OIG9wZXJhdGlvbkNoZWNrSWZFeGNlcHRpb25Jc1VuY2F0Y2hhYmxlQW5kTm90aWZ5UHJvZmls ZXIoRXgKICAgICAgICAgcmV0dXJuIDA7CiB9CiAKK2ludDMyX3QgSklUX09QRVJBVElPTiBvcGVy YXRpb25Ub0ludDMyKGRvdWJsZSB2YWx1ZSkKK3sKKyAgICByZXR1cm4gSlNDOjp0b0ludDMyKHZh bHVlKTsKK30KKwogfSAvLyBleHRlcm4gIkMiCiAKIC8vIE5vdGU6IGdldEhvc3RDYWxsUmV0dXJu VmFsdWVXaXRoRXhlY1N0YXRlKCkgbmVlZHMgdG8gYmUgcGxhY2VkIGJlZm9yZSB0aGUKZGlmZiAt LWdpdCBhL1NvdXJjZS9KYXZhU2NyaXB0Q29yZS9qaXQvSklUT3BlcmF0aW9ucy5oIGIvU291cmNl L0phdmFTY3JpcHRDb3JlL2ppdC9KSVRPcGVyYXRpb25zLmgKaW5kZXggYTk3MjZlYmQ2YjY4MTlk ZWFkNmRjYzRlM2E1MDBiZDk5MWVkMDAyNC4uZTVhNDBkMDcyNGRjYTUwNDVhZjkwMjY1NjY0ODY0 OWQzZDMxMmYwZiAxMDA2NDQKLS0tIGEvU291cmNlL0phdmFTY3JpcHRDb3JlL2ppdC9KSVRPcGVy YXRpb25zLmgKKysrIGIvU291cmNlL0phdmFTY3JpcHRDb3JlL2ppdC9KSVRPcGVyYXRpb25zLmgK QEAgLTQxOSw2ICs0MTksOCBAQCBFbmNvZGVkSlNWYWx1ZSBKSVRfT1BFUkFUSU9OIG9wZXJhdGlv blZhbHVlU3ViUHJvZmlsZWQoRXhlY1N0YXRlKiwgRW5jb2RlZEpTVmFsdQogdm9pZCBKSVRfT1BF UkFUSU9OIG9wZXJhdGlvblByb2Nlc3NUeXBlUHJvZmlsZXJMb2coRXhlY1N0YXRlKikgV1RGX0lO VEVSTkFMOwogdm9pZCBKSVRfT1BFUkFUSU9OIG9wZXJhdGlvblByb2Nlc3NTaGFkb3dDaGlja2Vu TG9nKEV4ZWNTdGF0ZSopIFdURl9JTlRFUk5BTDsKIAoraW50MzJfdCBKSVRfT1BFUkFUSU9OIG9w ZXJhdGlvblRvSW50MzIoZG91YmxlKTsKKwogfSAvLyBleHRlcm4gIkMiCiAKIH0gLy8gbmFtZXNw YWNlIEpTQwpkaWZmIC0tZ2l0IGEvU291cmNlL0phdmFTY3JpcHRDb3JlL3J1bnRpbWUvSlNDSlNW YWx1ZS5jcHAgYi9Tb3VyY2UvSmF2YVNjcmlwdENvcmUvcnVudGltZS9KU0NKU1ZhbHVlLmNwcApp bmRleCBjYTRmNTU4YTBkMzhiN2Y4MTM1YjhlMWM3ZTY5ZTUzYTJmZDExOGE4Li5kZGYzYjVmZjU3 NmQwNzA5MzA1MDgxZGE5MzUxMjJlZTA1NTA5MWNkIDEwMDY0NAotLS0gYS9Tb3VyY2UvSmF2YVNj cmlwdENvcmUvcnVudGltZS9KU0NKU1ZhbHVlLmNwcAorKysgYi9Tb3VyY2UvSmF2YVNjcmlwdENv cmUvcnVudGltZS9KU0NKU1ZhbHVlLmNwcApAQCAtMzE0LDUxICszMTQsNiBAQCB2b2lkIEpTVmFs dWU6OmR1bXBGb3JCYWNrdHJhY2UoUHJpbnRTdHJlYW0mIG91dCkgY29uc3QKICAgICAgICAgb3V0 LnByaW50KCJJTlZBTElEIik7CiB9CiAKLS8vIFRoaXMgaW4gdGhlIFRvSW50MzIgb3BlcmF0aW9u IGlzIGRlZmluZWQgaW4gc2VjdGlvbiA5LjUgb2YgdGhlIEVDTUEtMjYyIHNwZWMuCi0vLyBOb3Rl IHRoYXQgdGhpcyBvcGVyYXRpb24gaXMgaWRlbnRpY2FsIHRvIFRvVUludDMyIG90aGVyIHRoYW4g dG8gaW50ZXJwcmV0YXRpb24KLS8vIG9mIHRoZSByZXN1bHRpbmcgYml0LXBhdHRlcm4gKGFzIHN1 Y2ggdGhpcyBtZXRvZCBpcyBhbHNvIGNhbGxlZCB0byBpbXBsZW1lbnQKLS8vIFRvVUludDMyKS4K LS8vCi0vLyBUaGUgb3BlcmF0aW9uIGNhbiBiZSBkZXNjaWJlZCBhcyByb3VuZCB0b3dhcmRzIHpl cm8sIHRoZW4gc2VsZWN0IHRoZSAzMiBsZWFzdAotLy8gYml0cyBvZiB0aGUgcmVzdWx0aW5nIHZh bHVlIGluIDJzLWNvbXBsZW1lbnQgcmVwcmVzZW50YXRpb24uCi1pbnQzMl90IHRvSW50MzIoZG91 YmxlIG51bWJlcikKLXsKLSAgICBpbnQ2NF90IGJpdHMgPSBXVEY6OmJpdHdpc2VfY2FzdDxpbnQ2 NF90PihudW1iZXIpOwotICAgIGludDMyX3QgZXhwID0gKHN0YXRpY19jYXN0PGludDMyX3Q+KGJp dHMgPj4gNTIpICYgMHg3ZmYpIC0gMHgzZmY7Ci0KLSAgICAvLyBJZiBleHBvbmVudCA8IDAgdGhl cmUgd2lsbCBiZSBubyBiaXRzIHRvIHRoZSBsZWZ0IG9mIHRoZSBkZWNpbWFsIHBvaW50Ci0gICAg Ly8gYWZ0ZXIgcm91bmRpbmc7IGlmIHRoZSBleHBvbmVudCBpcyA+IDgzIHRoZW4gbm8gYml0cyBv ZiBwcmVjaXNpb24gY2FuIGJlCi0gICAgLy8gbGVmdCBpbiB0aGUgbG93IDMyLWJpdCByYW5nZSBv ZiB0aGUgcmVzdWx0IChJRUVFLTc1NCBkb3VibGVzIGhhdmUgNTIgYml0cwotICAgIC8vIG9mIGZy YWN0aW9uYWwgcHJlY2lzaW9uKS4KLSAgICAvLyBOb3RlIHRoaXMgY2FzZSBoYW5kbGVzIDAsIC0w LCBhbmQgYWxsIGluZmludGUsIE5hTiwgJiBkZW5vcm1hbCB2YWx1ZS4gCi0gICAgaWYgKGV4cCA8 IDAgfHwgZXhwID4gODMpCi0gICAgICAgIHJldHVybiAwOwotCi0gICAgLy8gU2VsZWN0IHRoZSBh cHByb3ByaWF0ZSAzMi1iaXRzIGZyb20gdGhlIGZsb2F0aW5nIHBvaW50IG1hbnRpc3NhLiAgSWYg dGhlCi0gICAgLy8gZXhwb25lbnQgaXMgNTIgdGhlbiB0aGUgYml0cyB3ZSBuZWVkIHRvIHNlbGVj dCBhcmUgYWxyZWFkeSBhbGlnbmVkIHRvIHRoZQotICAgIC8vIGxvd2VzdCBiaXRzIG9mIHRoZSA2 NC1iaXQgaW50ZWdlciByZXByZXNlbnRhdGlvbiBvZiB0Z2hlIG51bWJlciwgbm8gbmVlZAotICAg IC8vIHRvIHNoaWZ0LiAgSWYgdGhlIGV4cG9uZW50IGlzIGdyZWF0ZXIgdGhhbiA1MiB3ZSBuZWVk IHRvIHNoaWZ0IHRoZSB2YWx1ZQotICAgIC8vIGxlZnQgYnkgKGV4cCAtIDUyKSwgaWYgdGhlIHZh bHVlIGlzIGxlc3MgdGhhbiA1MiB3ZSBuZWVkIHRvIHNoaWZ0IHJpZ2h0Ci0gICAgLy8gYWNjb3Jk aW5nbHkuCi0gICAgaW50MzJfdCByZXN1bHQgPSAoZXhwID4gNTIpCi0gICAgICAgID8gc3RhdGlj X2Nhc3Q8aW50MzJfdD4oYml0cyA8PCAoZXhwIC0gNTIpKQotICAgICAgICA6IHN0YXRpY19jYXN0 PGludDMyX3Q+KGJpdHMgPj4gKDUyIC0gZXhwKSk7Ci0KLSAgICAvLyBJRUVFLTc1NCBkb3VibGUg cHJlY2lzaW9uIHZhbHVlcyBhcmUgc3RvcmVkIG9taXR0aW5nIGFuIGltcGxpY2l0IDEgYmVmb3Jl Ci0gICAgLy8gdGhlIGRlY2ltYWwgcG9pbnQ7IHdlIG5lZWQgdG8gcmVpbnNlcnQgdGhpcyBub3cu ICBXZSBtYXkgYWxzbyB0aGUgc2hpZnRlZAotICAgIC8vIGludmFsaWQgYml0cyBpbnRvIHRoZSBy ZXN1bHQgdGhhdCBhcmUgbm90IGEgcGFydCBvZiB0aGUgbWFudGlzc2EgKHRoZSBzaWduCi0gICAg Ly8gYW5kIGV4cG9uZW50IGJpdHMgZnJvbSB0aGUgZmxvYXRpbmdwb2ludCByZXByZXNlbnRhdGlv bik7IG1hc2sgdGhlc2Ugb3V0LgotICAgIGlmIChleHAgPCAzMikgewotICAgICAgICBpbnQzMl90 IG1pc3NpbmdPbmUgPSAxIDw8IGV4cDsKLSAgICAgICAgcmVzdWx0ICY9IG1pc3NpbmdPbmUgLSAx OwotICAgICAgICByZXN1bHQgKz0gbWlzc2luZ09uZTsKLSAgICB9Ci0KLSAgICAvLyBJZiB0aGUg aW5wdXQgdmFsdWUgd2FzIG5lZ2F0aXZlICh3ZSBjb3VsZCB0ZXN0IGVpdGhlciAnbnVtYmVyJyBv ciAnYml0cycsCi0gICAgLy8gYnV0IHRlc3RpbmcgJ2JpdHMnIGlzIGxpa2VseSBmYXN0ZXIpIGlu dmVydCB0aGUgcmVzdWx0IGFwcHJvcHJpYXRlbHkuCi0gICAgcmV0dXJuIGJpdHMgPCAwID8gLXJl c3VsdCA6IHJlc3VsdDsKLX0KLQogYm9vbCBKU1ZhbHVlOjppc1ZhbGlkQ2FsbGVlKCkKIHsKICAg ICByZXR1cm4gYXNPYmplY3QoYXNDZWxsKCkpLT5nbG9iYWxPYmplY3QoKTsKZGlmZiAtLWdpdCBh L1NvdXJjZS9KYXZhU2NyaXB0Q29yZS9ydW50aW1lL0pTQ0pTVmFsdWVJbmxpbmVzLmggYi9Tb3Vy Y2UvSmF2YVNjcmlwdENvcmUvcnVudGltZS9KU0NKU1ZhbHVlSW5saW5lcy5oCmluZGV4IGNmY2Iz NWZhMTA4YTNhODI5ZjgyYTgxYWI0ODcxMjFjZjY3MTk0NGEuLjZiYjI4NGUxMDM1NDM1ZDIxYWUw OTZiMTgzY2RlYmQxZTc5MTlkNmYgMTAwNjQ0Ci0tLSBhL1NvdXJjZS9KYXZhU2NyaXB0Q29yZS9y dW50aW1lL0pTQ0pTVmFsdWVJbmxpbmVzLmgKKysrIGIvU291cmNlL0phdmFTY3JpcHRDb3JlL3J1 bnRpbWUvSlNDSlNWYWx1ZUlubGluZXMuaApAQCAtMTA1Miw2ICsxMDUyLDUxIEBAIEFMV0FZU19J TkxJTkUgYm9vbCBpc1RoaXNWYWx1ZUFsdGVyZWQoY29uc3QgUHV0UHJvcGVydHlTbG90JiBzbG90 LCBKU09iamVjdCogYmFzCiAgICAgcmV0dXJuIHRydWU7CiB9CiAKKy8vIFRoaXMgaW4gdGhlIFRv SW50MzIgb3BlcmF0aW9uIGlzIGRlZmluZWQgaW4gc2VjdGlvbiA5LjUgb2YgdGhlIEVDTUEtMjYy IHNwZWMuCisvLyBOb3RlIHRoYXQgdGhpcyBvcGVyYXRpb24gaXMgaWRlbnRpY2FsIHRvIFRvVUlu dDMyIG90aGVyIHRoYW4gdG8gaW50ZXJwcmV0YXRpb24KKy8vIG9mIHRoZSByZXN1bHRpbmcgYml0 LXBhdHRlcm4gKGFzIHN1Y2ggdGhpcyBtZXRvZCBpcyBhbHNvIGNhbGxlZCB0byBpbXBsZW1lbnQK Ky8vIFRvVUludDMyKS4KKy8vCisvLyBUaGUgb3BlcmF0aW9uIGNhbiBiZSBkZXNjaWJlZCBhcyBy b3VuZCB0b3dhcmRzIHplcm8sIHRoZW4gc2VsZWN0IHRoZSAzMiBsZWFzdAorLy8gYml0cyBvZiB0 aGUgcmVzdWx0aW5nIHZhbHVlIGluIDJzLWNvbXBsZW1lbnQgcmVwcmVzZW50YXRpb24uCitBTFdB WVNfSU5MSU5FIGludDMyX3QgdG9JbnQzMihkb3VibGUgbnVtYmVyKQoreworICAgIGludDY0X3Qg Yml0cyA9IFdURjo6Yml0d2lzZV9jYXN0PGludDY0X3Q+KG51bWJlcik7CisgICAgaW50MzJfdCBl eHAgPSAoc3RhdGljX2Nhc3Q8aW50MzJfdD4oYml0cyA+PiA1MikgJiAweDdmZikgLSAweDNmZjsK KworICAgIC8vIElmIGV4cG9uZW50IDwgMCB0aGVyZSB3aWxsIGJlIG5vIGJpdHMgdG8gdGhlIGxl ZnQgb2YgdGhlIGRlY2ltYWwgcG9pbnQKKyAgICAvLyBhZnRlciByb3VuZGluZzsgaWYgdGhlIGV4 cG9uZW50IGlzID4gODMgdGhlbiBubyBiaXRzIG9mIHByZWNpc2lvbiBjYW4gYmUKKyAgICAvLyBs ZWZ0IGluIHRoZSBsb3cgMzItYml0IHJhbmdlIG9mIHRoZSByZXN1bHQgKElFRUUtNzU0IGRvdWJs ZXMgaGF2ZSA1MiBiaXRzCisgICAgLy8gb2YgZnJhY3Rpb25hbCBwcmVjaXNpb24pLgorICAgIC8v IE5vdGUgdGhpcyBjYXNlIGhhbmRsZXMgMCwgLTAsIGFuZCBhbGwgaW5maW50ZSwgTmFOLCAmIGRl bm9ybWFsIHZhbHVlLgorICAgIGlmIChleHAgPCAwIHx8IGV4cCA+IDgzKQorICAgICAgICByZXR1 cm4gMDsKKworICAgIC8vIFNlbGVjdCB0aGUgYXBwcm9wcmlhdGUgMzItYml0cyBmcm9tIHRoZSBm bG9hdGluZyBwb2ludCBtYW50aXNzYS4gSWYgdGhlCisgICAgLy8gZXhwb25lbnQgaXMgNTIgdGhl biB0aGUgYml0cyB3ZSBuZWVkIHRvIHNlbGVjdCBhcmUgYWxyZWFkeSBhbGlnbmVkIHRvIHRoZQor ICAgIC8vIGxvd2VzdCBiaXRzIG9mIHRoZSA2NC1iaXQgaW50ZWdlciByZXByZXNlbnRhdGlvbiBv ZiB0Z2hlIG51bWJlciwgbm8gbmVlZAorICAgIC8vIHRvIHNoaWZ0LiBJZiB0aGUgZXhwb25lbnQg aXMgZ3JlYXRlciB0aGFuIDUyIHdlIG5lZWQgdG8gc2hpZnQgdGhlIHZhbHVlCisgICAgLy8gbGVm dCBieSAoZXhwIC0gNTIpLCBpZiB0aGUgdmFsdWUgaXMgbGVzcyB0aGFuIDUyIHdlIG5lZWQgdG8g c2hpZnQgcmlnaHQKKyAgICAvLyBhY2NvcmRpbmdseS4KKyAgICBpbnQzMl90IHJlc3VsdCA9IChl eHAgPiA1MikKKyAgICAgICAgPyBzdGF0aWNfY2FzdDxpbnQzMl90PihiaXRzIDw8IChleHAgLSA1 MikpCisgICAgICAgIDogc3RhdGljX2Nhc3Q8aW50MzJfdD4oYml0cyA+PiAoNTIgLSBleHApKTsK KworICAgIC8vIElFRUUtNzU0IGRvdWJsZSBwcmVjaXNpb24gdmFsdWVzIGFyZSBzdG9yZWQgb21p dHRpbmcgYW4gaW1wbGljaXQgMSBiZWZvcmUKKyAgICAvLyB0aGUgZGVjaW1hbCBwb2ludDsgd2Ug bmVlZCB0byByZWluc2VydCB0aGlzIG5vdy4gV2UgbWF5IGFsc28gdGhlIHNoaWZ0ZWQKKyAgICAv LyBpbnZhbGlkIGJpdHMgaW50byB0aGUgcmVzdWx0IHRoYXQgYXJlIG5vdCBhIHBhcnQgb2YgdGhl IG1hbnRpc3NhICh0aGUgc2lnbgorICAgIC8vIGFuZCBleHBvbmVudCBiaXRzIGZyb20gdGhlIGZs b2F0aW5ncG9pbnQgcmVwcmVzZW50YXRpb24pOyBtYXNrIHRoZXNlIG91dC4KKyAgICBpZiAoZXhw IDwgMzIpIHsKKyAgICAgICAgaW50MzJfdCBtaXNzaW5nT25lID0gMSA8PCBleHA7CisgICAgICAg IHJlc3VsdCAmPSBtaXNzaW5nT25lIC0gMTsKKyAgICAgICAgcmVzdWx0ICs9IG1pc3NpbmdPbmU7 CisgICAgfQorCisgICAgLy8gSWYgdGhlIGlucHV0IHZhbHVlIHdhcyBuZWdhdGl2ZSAod2UgY291 bGQgdGVzdCBlaXRoZXIgJ251bWJlcicgb3IgJ2JpdHMnLAorICAgIC8vIGJ1dCB0ZXN0aW5nICdi aXRzJyBpcyBsaWtlbHkgZmFzdGVyKSBpbnZlcnQgdGhlIHJlc3VsdCBhcHByb3ByaWF0ZWx5Lgor ICAgIHJldHVybiBiaXRzIDwgMCA/IC1yZXN1bHQgOiByZXN1bHQ7Cit9CisKIH0gLy8gbmFtZXNw YWNlIEpTQwogCiAjZW5kaWYgLy8gSlNWYWx1ZUlubGluZXNfaAo=