195322 2019-03-05 01:52:55 -0800 REGRESSION(r236862): early frame decoupling leaves JSC ArrayBuffer objects lingering 2019-03-18 14:28:25 -0700 1 1 1 Unclassified WebKit New Bugs WebKit Nightly Build Unspecified Unspecified RESOLVED FIXED InRadar P2 Normal --- 1 zan zan cdumez commit-queue ews-watchlist fpizlo ggaren keith_miller mark.lam rniwa saam webkit-bug-importer ysuzuki oldest_to_newest 1512572 0 zan 2019-03-05 01:52:55 -0800 With frame detachment changes made in r236862, it seems JSC resources can persist beyond frame destruction. https://trac.webkit.org/changeset/236862/webkit The observed case is the one of ArrayBuffer resources that are kept alive during the JetStream benchmark execution. After r236862, at the end of a JetStream run there can be about 25 ArrayBuffer objects still referenced in the JSC heap and thus deferred for allocation through the GCIncomingRefCountedSet mechanism. This totals to over 700MB of extra memory at the end of a JetStream run. Before r236862 these ArrayBuffer objects were properly cleaned up during the run with only one or two objects remaining alive at the end of the run due to being used in the last few test cases, totalling ~150MB in size. In r236862 specifically, the problem is a side effect of the early decoupling from the Frame object in DOMWindow::willDetachDocumentFromFrame(). This was further reorganized in later changes, but the problem remains present and can be observed on both Mac and Linux ports. 1512573 1 363624 zan 2019-03-05 01:57:20 -0800 Created attachment 363624 Heap logging patch This patch adds more detailed dumps of extra memory statistics during GC, and provides a rundown of to-be-swept ArrayBuffer objects contained in GCIncomingRefCountedSet<ArrayBuffer>. Here's the output on the last GC for a JetStream run: > GCIncomingRefCountedSet::sweep(): m_vector[24] > object 0x3f08a8780, deferred 1, refCount 0, 16 incoming references > object 0x3fc45a180, deferred 1, refCount 0, 16 incoming references > object 0x3f1c40140, deferred 1, refCount 0, 11 incoming references > object 0x3fb415b40, deferred 1, refCount 0, 12 incoming references > object 0x3fa48b900, deferred 1, refCount 0, 16 incoming references > object 0x3fb8cbdc0, deferred 1, refCount 0, 12 incoming references > object 0x3f2d489c0, deferred 1, refCount 0, 11 incoming references > object 0x3fbf72100, deferred 1, refCount 0, 11 incoming references > object 0x3f7278c00, deferred 1, refCount 0, 16 incoming references > object 0x4073a84c0, deferred 1, refCount 0, 16 incoming references > object 0x3f0da24c0, deferred 1, refCount 0, 11 incoming references > object 0x3f2696540, deferred 1, refCount 0, 12 incoming references > object 0x3fc663280, deferred 1, refCount 0, 16 incoming references > object 0x3f2d86c80, deferred 1, refCount 0, 12 incoming references > object 0x3f2bf1100, deferred 1, refCount 0, 11 incoming references > object 0x3f5dfca00, deferred 1, refCount 0, 11 incoming references > object 0x3f6c6a940, deferred 1, refCount 0, 16 incoming references > object 0x3f53a2340, deferred 1, refCount 0, 16 incoming references > object 0x415615980, deferred 1, refCount 0, 13 incoming references > object 0x3f8508ec0, deferred 1, refCount 0, 14 incoming references > object 0x3ed269140, deferred 1, refCount 0, 44 incoming references > object 0x3f6255700, deferred 1, refCount 0, 16 incoming references > object 0x40170e2c0, deferred 1, refCount 0, 13 incoming references > object 0x411f0f280, deferred 1, refCount 0, 13 incoming references > > bytesAllocatedThisCycle = 313249 > totalBytesVisited = 13801072, currentHeapSize = 13801072 > extraMemorySize() = 759166003, currentHeapSize = 772967075 > m_extraMemorySize = 4189747, m_deprecatedExtraMemorySize = 0, m_arrayBuffers.size() = 754976256 > Full: maxHeapSize = 1545934150 > Full: maxEdenSize = 772967075 > Full: sizeAfterLastFullCollect = 772967075 > Full: bytesAbandonedSinceLastFullCollect = 0 > sizeAfterLastCollect = 772967075 1514484 2 rniwa 2019-03-09 19:37:34 -0800 Hm... it's hard to believe that we'd only leak ArrayBuffer with that change. Perhaps we're leaking the entire document / frame? 1514492 3 cdumez 2019-03-09 20:23:01 -0800 We may need to call JSDOMWindowBase::fireFrameClearedWatchpointsForWindow(this); in DOMWindow::willDetachDocumentFromFrame(). 1514587 4 cdumez 2019-03-10 15:23:41 -0700 (In reply to Chris Dumez from comment #3) > We may need to call > JSDOMWindowBase::fireFrameClearedWatchpointsForWindow(this); in > DOMWindow::willDetachDocumentFromFrame(). It does not seem to help anything unfortunately. 1514592 5 cdumez 2019-03-10 15:35:16 -0700 (In reply to Chris Dumez from comment #4) > (In reply to Chris Dumez from comment #3) > > We may need to call > > JSDOMWindowBase::fireFrameClearedWatchpointsForWindow(this); in > > DOMWindow::willDetachDocumentFromFrame(). > > It does not seem to help anything unfortunately. That said, I do not see any difference either if I comment out the code in Document::frameWasDisconnectedFromOwner() which I believe would restore pre-r236862 behavior. 1516615 6 zan 2019-03-14 13:39:30 -0700 The problem seems to be a lack of call to InspectorInstrumentation::frameWindowDiscarded(). Before r236862 this was called from Frame::willDetachPage(), through the FrameDestructionObserver interface. http://trac.webkit.org/browser/webkit/trunk/Source/WebCore/page/DOMWindow.cpp?rev=236862#L484 Starting in r236862 this was not being called anymore because DOMWindow was already decoupled from Frame when Frame::willDetachPage() was invoked and subsequently the m_destructionObservers set was empty. The situation in trunk is similar, only difference is that after further refactoring the InspectorInstrumentation::frameWindowDiscarded() is now placed in Document::willDetachPage(). 1516619 7 cdumez 2019-03-14 13:40:52 -0700 (In reply to Zan Dobersek from comment #6) > The problem seems to be a lack of call to > InspectorInstrumentation::frameWindowDiscarded(). Before r236862 this was > called from Frame::willDetachPage(), through the FrameDestructionObserver > interface. > http://trac.webkit.org/browser/webkit/trunk/Source/WebCore/page/DOMWindow. > cpp?rev=236862#L484 > > Starting in r236862 this was not being called anymore because DOMWindow was > already decoupled from Frame when Frame::willDetachPage() was invoked and > subsequently the m_destructionObservers set was empty. > > The situation in trunk is similar, only difference is that after further > refactoring the InspectorInstrumentation::frameWindowDiscarded() is now > placed in Document::willDetachPage(). Then we probably need a call to InspectorInstrumentation::frameWindowDiscarded() in DOMWindow::willDetachDocumentFromFrame(). 1516967 8 364796 zan 2019-03-15 07:50:06 -0700 Created attachment 364796 EWS run 1516968 9 ews-watchlist 2019-03-15 07:54:06 -0700 Attachment 364796 did not pass style-queue: ERROR: Source/WebCore/ChangeLog:8: You should remove the 'No new tests' and either add and list tests, or explain why no new tests were possible. [changelog/nonewtests] [5] Total errors found: 1 in 2 files If any of these errors are false positives, please file a bug against check-webkit-style. 1517887 10 365054 cdumez 2019-03-18 13:41:54 -0700 Created attachment 365054 Patch 1517925 11 365054 commit-queue 2019-03-18 14:27:14 -0700 Comment on attachment 365054 Patch Clearing flags on attachment: 365054 Committed r243104: <https://trac.webkit.org/changeset/243104> 1517926 12 commit-queue 2019-03-18 14:27:16 -0700 All reviewed patches have been landed. Closing bug. 1517927 13 webkit-bug-importer 2019-03-18 14:28:25 -0700 <rdar://problem/48993555> 363624 2019-03-05 01:57:20 -0800 2019-03-15 07:50:00 -0700 Heap logging patch heaplogging.patch text/plain 1946 zan ZGlmZiAtLWdpdCBhL1NvdXJjZS9KYXZhU2NyaXB0Q29yZS9oZWFwL0dDSW5jb21pbmdSZWZDb3Vu dGVkU2V0SW5saW5lcy5oIGIvU291cmNlL0phdmFTY3JpcHRDb3JlL2hlYXAvR0NJbmNvbWluZ1Jl ZkNvdW50ZWRTZXRJbmxpbmVzLmgKaW5kZXggYjEwOTcyZGEwYTQuLmExOWEzYzM1ZTBlIDEwMDY0 NAotLS0gYS9Tb3VyY2UvSmF2YVNjcmlwdENvcmUvaGVhcC9HQ0luY29taW5nUmVmQ291bnRlZFNl dElubGluZXMuaAorKysgYi9Tb3VyY2UvSmF2YVNjcmlwdENvcmUvaGVhcC9HQ0luY29taW5nUmVm Q291bnRlZFNldElubGluZXMuaApAQCAtNjEsNiArNjEsMTAgQEAgYm9vbCBHQ0luY29taW5nUmVm Q291bnRlZFNldDxUPjo6YWRkUmVmZXJlbmNlKEpTQ2VsbCogY2VsbCwgVCogb2JqZWN0KQogdGVt cGxhdGU8dHlwZW5hbWUgVD4KIHZvaWQgR0NJbmNvbWluZ1JlZkNvdW50ZWRTZXQ8VD46OnN3ZWVw KCkKIHsKKyAgICBmcHJpbnRmKHN0ZGVyciwgIkdDSW5jb21pbmdSZWZDb3VudGVkU2V0Ojpzd2Vl cCgpOiBtX3ZlY3RvclslenVdXG4iLCBtX3ZlY3Rvci5zaXplKCkpOworICAgIGZvciAoYXV0byog b2JqZWN0IDogbV92ZWN0b3IpCisgICAgICAgIGZwcmludGYoc3RkZXJyLCAiICBvYmplY3QgJXAs IGRlZmVycmVkICVkLCByZWZDb3VudCAlZCwgJXp1IGluY29taW5nIHJlZmVyZW5jZXNcbiIsCisg ICAgICAgICAgICBvYmplY3QsIG9iamVjdC0+aXNEZWZlcnJlZCgpLCBvYmplY3QtPnJlZkNvdW50 KCksIG9iamVjdC0+bnVtYmVyT2ZJbmNvbWluZ1JlZmVyZW5jZXMoKSk7CiAgICAgZm9yIChzaXpl X3QgaSA9IDA7IGkgPCBtX3ZlY3Rvci5zaXplKCk7ICsraSkgewogICAgICAgICBUKiBvYmplY3Qg PSBtX3ZlY3RvcltpXTsKICAgICAgICAgc2l6ZV90IHNpemUgPSBvYmplY3QtPmdjU2l6ZUVzdGlt YXRlSW5CeXRlcygpOwpkaWZmIC0tZ2l0IGEvU291cmNlL0phdmFTY3JpcHRDb3JlL2hlYXAvSGVh cC5jcHAgYi9Tb3VyY2UvSmF2YVNjcmlwdENvcmUvaGVhcC9IZWFwLmNwcAppbmRleCAwNmVhM2I2 MmU2NC4uNGMwYmFkOThlZmEgMTAwNjQ0Ci0tLSBhL1NvdXJjZS9KYXZhU2NyaXB0Q29yZS9oZWFw L0hlYXAuY3BwCisrKyBiL1NvdXJjZS9KYXZhU2NyaXB0Q29yZS9oZWFwL0hlYXAuY3BwCkBAIC0y MTc5LDcgKzIxNzksNyBAQCB2b2lkIEhlYXA6Om5vdGlmeUluY3JlbWVudGFsU3dlZXBlcigpCiAK IHZvaWQgSGVhcDo6dXBkYXRlQWxsb2NhdGlvbkxpbWl0cygpCiB7Ci0gICAgc3RhdGljIGNvbnN0 IGJvb2wgdmVyYm9zZSA9IGZhbHNlOworICAgIHN0YXRpYyBjb25zdCBib29sIHZlcmJvc2UgPSB0 cnVlOwogICAgIAogICAgIGlmICh2ZXJib3NlKSB7CiAgICAgICAgIGRhdGFMb2coIlxuIik7CkBA IC0yMjEzLDYgKzIyMTMsOSBAQCB2b2lkIEhlYXA6OnVwZGF0ZUFsbG9jYXRpb25MaW1pdHMoKQog CiAgICAgaWYgKHZlcmJvc2UpCiAgICAgICAgIGRhdGFMb2coImV4dHJhTWVtb3J5U2l6ZSgpID0g IiwgZXh0cmFNZW1vcnlTaXplKCksICIsIGN1cnJlbnRIZWFwU2l6ZSA9ICIsIGN1cnJlbnRIZWFw U2l6ZSwgIlxuIik7CisgICAgaWYgKHZlcmJvc2UpCisgICAgICAgIGRhdGFMb2coIiAgbV9leHRy YU1lbW9yeVNpemUgPSAiLCBtX2V4dHJhTWVtb3J5U2l6ZSwgIiwgbV9kZXByZWNhdGVkRXh0cmFN ZW1vcnlTaXplID0gIiwKKyAgICAgICAgICAgIG1fZGVwcmVjYXRlZEV4dHJhTWVtb3J5U2l6ZSwg IiwgbV9hcnJheUJ1ZmZlcnMuc2l6ZSgpID0gIiwgbV9hcnJheUJ1ZmZlcnMuc2l6ZSgpLCAiXG4i KTsKICAgICAKICAgICBpZiAobV9jb2xsZWN0aW9uU2NvcGUgJiYgbV9jb2xsZWN0aW9uU2NvcGUu dmFsdWUoKSA9PSBDb2xsZWN0aW9uU2NvcGU6OkZ1bGwpIHsKICAgICAgICAgLy8gVG8gYXZvaWQg cGF0aG9sb2dpY2FsIEdDIGNodXJuIGluIHZlcnkgc21hbGwgYW5kIHZlcnkgbGFyZ2UgaGVhcHMs IHdlIHNldAo= 364796 2019-03-15 07:50:06 -0700 2019-03-18 13:41:52 -0700 EWS run bug-195322-20190315155004.patch text/plain 1260 zan U3VidmVyc2lvbiBSZXZpc2lvbjogMjQyOTQxCmRpZmYgLS1naXQgYS9Tb3VyY2UvV2ViQ29yZS9D aGFuZ2VMb2cgYi9Tb3VyY2UvV2ViQ29yZS9DaGFuZ2VMb2cKaW5kZXggMmQyMTVlNjk2ODcyMWJj ZTk1MzFhOTI1NDBjMDMyODFkMzZhYmM2MC4uYjNmMmJiZGM1MmZkYWMwNTEyNTZmNWZjYzJmOWQ0 ZDI3YjMxYmMwZiAxMDA2NDQKLS0tIGEvU291cmNlL1dlYkNvcmUvQ2hhbmdlTG9nCisrKyBiL1Nv dXJjZS9XZWJDb3JlL0NoYW5nZUxvZwpAQCAtMSwzICsxLDE1IEBACisyMDE5LTAzLTE1ICBaYW4g RG9iZXJzZWsgIDx6ZG9iZXJzZWtAaWdhbGlhLmNvbT4KKworICAgICAgICBSRUdSRVNTSU9OKHIy MzY4NjIpOiBlYXJseSBmcmFtZSBkZWNvdXBsaW5nIGxlYXZlcyBKU0MgQXJyYXlCdWZmZXIgb2Jq ZWN0cyBsaW5nZXJpbmcKKyAgICAgICAgaHR0cHM6Ly9idWdzLndlYmtpdC5vcmcvc2hvd19idWcu Y2dpP2lkPTE5NTMyMgorCisgICAgICAgIFJldmlld2VkIGJ5IE5PQk9EWSAoT09QUyEpLgorCisg ICAgICAgIE5vIG5ldyB0ZXN0cyAoT09QUyEpLgorCisgICAgICAgICogcGFnZS9ET01XaW5kb3cu Y3BwOgorICAgICAgICAoV2ViQ29yZTo6RE9NV2luZG93Ojp3aWxsRGV0YWNoRG9jdW1lbnRGcm9t RnJhbWUpOgorCiAyMDE5LTAzLTE0ICBEZXZpbiBSb3Vzc28gIDxkcm91c3NvQGFwcGxlLmNvbT4K IAogICAgICAgICBXZWIgSW5zcGVjdG9yOiBBdWRpdDogcHJvdmlkZSBhIHdheSB0byBnZXQgdGhl IGNvbnRlbnRzIG9mIHJlc291cmNlcwpkaWZmIC0tZ2l0IGEvU291cmNlL1dlYkNvcmUvcGFnZS9E T01XaW5kb3cuY3BwIGIvU291cmNlL1dlYkNvcmUvcGFnZS9ET01XaW5kb3cuY3BwCmluZGV4IDhj MzFkYmQwNWNhMzliODRhY2EwYzIwYmY3ZDZjNjZjNzVmMmUzOWMuLjNlNDY1ZDYwNTNkYzc1Y2Ri ZjRiZDZmOTZmYTlhMzRjZWZiMGMxNzIgMTAwNjQ0Ci0tLSBhL1NvdXJjZS9XZWJDb3JlL3BhZ2Uv RE9NV2luZG93LmNwcAorKysgYi9Tb3VyY2UvV2ViQ29yZS9wYWdlL0RPTVdpbmRvdy5jcHAKQEAg LTQ4NSw2ICs0ODUsOCBAQCB2b2lkIERPTVdpbmRvdzo6d2lsbERldGFjaERvY3VtZW50RnJvbUZy YW1lKCkKIAogICAgIGlmIChtX3BlcmZvcm1hbmNlKQogICAgICAgICBtX3BlcmZvcm1hbmNlLT5j bGVhclJlc291cmNlVGltaW5ncygpOworCisgICAgSW5zcGVjdG9ySW5zdHJ1bWVudGF0aW9uOjpm cmFtZVdpbmRvd0Rpc2NhcmRlZCgqZnJhbWUoKSwgdGhpcyk7CiB9CiAKICNpZiBFTkFCTEUoR0FN RVBBRCkK 365054 2019-03-18 13:41:54 -0700 2019-03-18 14:27:14 -0700 Patch bug-195322-20190318134153.patch text/plain 2147 cdumez U3VidmVyc2lvbiBSZXZpc2lvbjogMjQzMDcxCmRpZmYgLS1naXQgYS9Tb3VyY2UvV2ViQ29yZS9D aGFuZ2VMb2cgYi9Tb3VyY2UvV2ViQ29yZS9DaGFuZ2VMb2cKaW5kZXggYzdhZTg5MGQwODBkOWE3 NjFhMjgyOWFlMGZhZDQzMzIxMDgwN2JhOC4uNGU3ODRiZTFiMjY4NjdjN2YyZjIzNTFmNWFjMWZi MDk0M2IxYTJjZiAxMDA2NDQKLS0tIGEvU291cmNlL1dlYkNvcmUvQ2hhbmdlTG9nCisrKyBiL1Nv dXJjZS9XZWJDb3JlL0NoYW5nZUxvZwpAQCAtMSwzICsxLDI2IEBACisyMDE5LTAzLTE4ICBDaHJp cyBEdW1leiAgPGNkdW1lekBhcHBsZS5jb20+CisKKyAgICAgICAgUkVHUkVTU0lPTihyMjM2ODYy KTogZWFybHkgZnJhbWUgZGVjb3VwbGluZyBsZWF2ZXMgSlNDIEFycmF5QnVmZmVyIG9iamVjdHMg bGluZ2VyaW5nCisgICAgICAgIGh0dHBzOi8vYnVncy53ZWJraXQub3JnL3Nob3dfYnVnLmNnaT9p ZD0xOTUzMjIKKworICAgICAgICBSZXZpZXdlZCBieSBOT0JPRFkgKE9PUFMhKS4KKworICAgICAg ICBTaW5jZSByMjM2ODYyLCBET01XaW5kb3cgb2JqZWN0cyBnZXQgZGlzY29ubmVjdGVkIGZyb20g dGhlaXIgRnJhbWUgb2JqZWN0IGFzIHNvb24gYXMKKyAgICAgICAgdGhlaXIgaWZyYW1lIGVsZW1l bnQgZ2V0cyByZW1vdmVkIGZyb20gdGhlIGRvY3VtZW50LiBQcmV2aW91c2x5LCBET01XaW5kb3cg d2FzIGEKKyAgICAgICAgRnJhbWVEZXN0cnVjdGlvbk9ic2VydmVyIGFuZCB3b3VsZCBzdGF5IGNv bm5lY3RlZCB0byBpdHMgZnJhbWUgdW50aWwgdGhlIGZyYW1lIGRpZWQuCisKKyAgICAgICAgVGhp cyBtZWFucyB0aGF0IHNvbWUgb2YgdGhlIHdvcmsgdGhhdCB3ZSB3ZXJlIGRvaW5nIGluIERPTVdp bmRvdzo6ZnJhbWVEZXN0cm95ZWQoKSBhbmQKKyAgICAgICAgRG9jdW1lbnQ6OndpbGxEZXRhY2hQ YWdlKCkgbm8gbG9uZ2VyIGhhcHBlbnMgZm9yIHN1YmZyYW1lIHdpbmRvd3MgYmVjYXVzZSB0aGV5 IGdldAorICAgICAgICBkaXNjb25uZWN0ZWQgZnJvbSB0aGVpciBmcmFtZSBiZWNhdXNlIHRoZXkg Z2V0IGEgY2hhbmNlIHRvIGdldCBzdWNoIG5vdGlmaWNhdGlvbnMuCisgICAgICAgIFRvIGFkZHJl c3MgdGhpcyBpc3N1ZSwgd2Ugbm93IGFsc28gZG8gdGhpcyB3b3JrIGluIERPTVdpbmRvdzo6d2ls bERldGFjaERvY3VtZW50RnJvbUZyYW1lKCkKKyAgICAgICAgd2hpY2ggZ2V0cyBjYWxsZWQgd2hl biB0aGUgaWZyYW1lIGdldHMgcmVtb3ZlZCBmcm9tIHRoZSBkb2N1bWVudCBhbmQgdGhlIGRvY3Vt ZW50IC8gd2luZG93CisgICAgICAgIGdldCBkaXNjb25uZWN0ZWQgZnJvbSB0aGUgRnJhbWUgZWxl bWVudC4KKworICAgICAgICBObyBuZXcgdGVzdHMsIHZlcmlmaWVkIGxvY2FsbHkgdGhhdCB0aGUg bGVhayBpcyBnb25lIG9uIEpldFN0cmVhbS4KKworICAgICAgICAqIHBhZ2UvRE9NV2luZG93LmNw cDoKKyAgICAgICAgKFdlYkNvcmU6OkRPTVdpbmRvdzo6d2lsbERldGFjaERvY3VtZW50RnJvbUZy YW1lKToKKwogMjAxOS0wMy0xNSAgQW50b2luZSBRdWludCAgPGdyYW91dHNAYXBwbGUuY29tPgog CiAgICAgICAgIE1ha2UgbW91c2UgZXZlbnQgc2ltdWxhdGlvbiBhIHF1aXJrCmRpZmYgLS1naXQg YS9Tb3VyY2UvV2ViQ29yZS9wYWdlL0RPTVdpbmRvdy5jcHAgYi9Tb3VyY2UvV2ViQ29yZS9wYWdl L0RPTVdpbmRvdy5jcHAKaW5kZXggNGQ3YjNhZmU3ZWM5MTY0ZWJkMDMxYThiZmMwMDQ1NWNlNTU2 OWEzOC4uZWYzMTliY2NiYzg1NWIwNzllYWNjY2EzY2NiYmI5NTNhNTc0NDJkZCAxMDA2NDQKLS0t IGEvU291cmNlL1dlYkNvcmUvcGFnZS9ET01XaW5kb3cuY3BwCisrKyBiL1NvdXJjZS9XZWJDb3Jl L3BhZ2UvRE9NV2luZG93LmNwcApAQCAtNDg1LDYgKzQ4NSw5IEBAIHZvaWQgRE9NV2luZG93Ojp3 aWxsRGV0YWNoRG9jdW1lbnRGcm9tRnJhbWUoKQogCiAgICAgaWYgKG1fcGVyZm9ybWFuY2UpCiAg ICAgICAgIG1fcGVyZm9ybWFuY2UtPmNsZWFyUmVzb3VyY2VUaW1pbmdzKCk7CisKKyAgICBKU0RP TVdpbmRvd0Jhc2U6OmZpcmVGcmFtZUNsZWFyZWRXYXRjaHBvaW50c0ZvcldpbmRvdyh0aGlzKTsK KyAgICBJbnNwZWN0b3JJbnN0cnVtZW50YXRpb246OmZyYW1lV2luZG93RGlzY2FyZGVkKCpmcmFt ZSgpLCB0aGlzKTsKIH0KIAogI2lmIEVOQUJMRShHQU1FUEFEKQo=