29207 2009-09-11 13:33:13 -0700 Add checks for using WebCore JS context on secondary threads 2009-09-11 14:38:16 -0700 1 1 1 Unclassified WebKit JavaScriptCore 528+ (Nightly build) All All RESOLVED FIXED P2 Normal --- 1 ap ap oldest_to_newest 146729 0 ap 2009-09-11 13:33:13 -0700 WebCore JavaScript context can never be used on secondary threads, because any such use can cause GC to happen, and thus a lot of DOM code to run on this thread. We need some assertions to more easily catch such conditions with debug builds. 146730 1 39473 ap 2009-09-11 13:36:51 -0700 Created attachment 39473 proposed patch 146736 2 39473 ggaren 2009-09-11 14:01:50 -0700 Comment on attachment 39473 proposed patch r=me Is there an efficient way to make this assertion in release builds too, like we do for some WebKit APIs? 146748 3 ap 2009-09-11 14:38:16 -0700 Committed revision 48315. > Is there an efficient way to make this assertion in release builds too, like we > do for some WebKit APIs? It's hard for me to estimate - how do we measure JSC API overhead? 39473 2009-09-11 13:36:51 -0700 2009-09-11 14:01:50 -0700 proposed patch JSCAPIThreadAssertion.txt text/plain 3626 ap SW5kZXg6IEphdmFTY3JpcHRDb3JlL0NoYW5nZUxvZwo9PT09PT09PT09PT09PT09PT09PT09PT09 PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09Ci0tLSBKYXZhU2NyaXB0 Q29yZS9DaGFuZ2VMb2cJKHJldmlzaW9uIDQ4MzE0KQorKysgSmF2YVNjcmlwdENvcmUvQ2hhbmdl TG9nCSh3b3JraW5nIGNvcHkpCkBAIC0xLDMgKzEsMTcgQEAKKzIwMDktMDktMTEgIEFsZXhleSBQ cm9za3VyeWFrb3YgIDxhcEB3ZWJraXQub3JnPgorCisgICAgICAgIFJldmlld2VkIGJ5IE5PQk9E WSAoT09QUyEpLgorCisgICAgICAgIGh0dHBzOi8vYnVncy53ZWJraXQub3JnL3Nob3dfYnVnLmNn aT9pZD0yOTIwNworICAgICAgICBBZGQgY2hlY2tzIGZvciB1c2luZyBXZWJDb3JlIEpTIGNvbnRl eHQgb24gc2Vjb25kYXJ5IHRocmVhZHMKKworICAgICAgICAqIHJ1bnRpbWUvSlNHbG9iYWxEYXRh LmNwcDogKEpTQzo6SlNHbG9iYWxEYXRhOjpKU0dsb2JhbERhdGEpOgorICAgICAgICAqIHJ1bnRp bWUvSlNHbG9iYWxEYXRhLmg6CisgICAgICAgIEFkZGVkIGEgbmV3IG1haW5UaHJlYWRPbmx5IGZs YWcgdGhhdCBXZWJDb3JlIHdvdWxkIHNldC4KKworICAgICAgICAqIHJ1bnRpbWUvQ29sbGVjdG9y LmNwcDogKEpTQzo6SGVhcDo6cmVnaXN0ZXJUaHJlYWQpOiBKU0MgQVBJIG1ldGhvZHMgYWx3YXlz IGNhbGwgdGhpcywKKyAgICAgICAgc28gdGhpcyBpcyBhIGdvb2QgcGxhY2UgdG8gY2hlY2sgdGhh dCB0aGUgQVBJIGlzbid0IHVzZWQgZm9ybSBhIHdyb25nIHRocmVhZC4KKwogMjAwOS0wOS0xMSAg Sm9jZWx5biBUdXJjb3R0ZSAgPGpvY2VseW4udHVyY290dGVAbm9raWEuY29tPgogCiAgICAgICAg IFJldmlld2VkIGJ5IFNpbW9uIEhhdXNtYW5uLgpJbmRleDogSmF2YVNjcmlwdENvcmUvcnVudGlt ZS9Db2xsZWN0b3IuY3BwCj09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09 PT09PT09PT09PT09PT09PT09PT09PT09PT0KLS0tIEphdmFTY3JpcHRDb3JlL3J1bnRpbWUvQ29s bGVjdG9yLmNwcAkocmV2aXNpb24gNDgyNTUpCisrKyBKYXZhU2NyaXB0Q29yZS9ydW50aW1lL0Nv bGxlY3Rvci5jcHAJKHdvcmtpbmcgY29weSkKQEAgLTY1MSw2ICs2NTEsOCBAQCB2b2lkIEhlYXA6 Om1ha2VVc2FibGVGcm9tTXVsdGlwbGVUaHJlYWRzCiAKIHZvaWQgSGVhcDo6cmVnaXN0ZXJUaHJl YWQoKQogeworICAgIEFTU0VSVCghbV9nbG9iYWxEYXRhLT5tYWluVGhyZWFkT25seSB8fCBpc01h aW5UaHJlYWQoKSk7CisKICAgICBpZiAoIW1fY3VycmVudFRocmVhZFJlZ2lzdHJhciB8fCBwdGhy ZWFkX2dldHNwZWNpZmljKG1fY3VycmVudFRocmVhZFJlZ2lzdHJhcikpCiAgICAgICAgIHJldHVy bjsKIApJbmRleDogSmF2YVNjcmlwdENvcmUvcnVudGltZS9KU0dsb2JhbERhdGEuY3BwCj09PT09 PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09 PT09PT0KLS0tIEphdmFTY3JpcHRDb3JlL3J1bnRpbWUvSlNHbG9iYWxEYXRhLmNwcAkocmV2aXNp b24gNDgyNTUpCisrKyBKYXZhU2NyaXB0Q29yZS9ydW50aW1lL0pTR2xvYmFsRGF0YS5jcHAJKHdv cmtpbmcgY29weSkKQEAgLTE0Nyw2ICsxNDcsOSBAQCBKU0dsb2JhbERhdGE6OkpTR2xvYmFsRGF0 YShib29sIGlzU2hhcmVkCiAgICAgLCBmdW5jdGlvbkNvZGVCbG9ja0JlaW5nUmVwYXJzZWQoMCkK ICAgICAsIGZpcnN0U3RyaW5naWZpZXJUb01hcmsoMCkKICAgICAsIG1hcmtTdGFjayh2cHRyU2V0 LmpzQXJyYXlWUHRyKQorI2lmbmRlZiBOREVCVUcKKyAgICAsIG1haW5UaHJlYWRPbmx5KGZhbHNl KQorI2VuZGlmCiB7CiAjaWYgUExBVEZPUk0oTUFDKQogICAgIHN0YXJ0UHJvZmlsZXJTZXJ2ZXJJ Zk5lZWRlZCgpOwpJbmRleDogSmF2YVNjcmlwdENvcmUvcnVudGltZS9KU0dsb2JhbERhdGEuaAo9 PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09 PT09PT09PT09Ci0tLSBKYXZhU2NyaXB0Q29yZS9ydW50aW1lL0pTR2xvYmFsRGF0YS5oCShyZXZp c2lvbiA0ODI1NSkKKysrIEphdmFTY3JpcHRDb3JlL3J1bnRpbWUvSlNHbG9iYWxEYXRhLmgJKHdv cmtpbmcgY29weSkKQEAgLTE1MCw2ICsxNTAsMTEgQEAgbmFtZXNwYWNlIEpTQyB7CiAgICAgICAg IFN0cmluZ2lmaWVyKiBmaXJzdFN0cmluZ2lmaWVyVG9NYXJrOwogCiAgICAgICAgIE1hcmtTdGFj ayBtYXJrU3RhY2s7CisKKyNpZm5kZWYgTkRFQlVHCisgICAgICAgIGJvb2wgbWFpblRocmVhZE9u bHk7CisjZW5kaWYKKwogICAgIHByaXZhdGU6CiAgICAgICAgIEpTR2xvYmFsRGF0YShib29sIGlz U2hhcmVkLCBjb25zdCBWUHRyU2V0Jik7CiAgICAgICAgIHN0YXRpYyBKU0dsb2JhbERhdGEqJiBz aGFyZWRJbnN0YW5jZUludGVybmFsKCk7CkluZGV4OiBXZWJDb3JlL0NoYW5nZUxvZwo9PT09PT09 PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09 PT09Ci0tLSBXZWJDb3JlL0NoYW5nZUxvZwkocmV2aXNpb24gNDgzMTQpCisrKyBXZWJDb3JlL0No YW5nZUxvZwkod29ya2luZyBjb3B5KQpAQCAtMSwzICsxLDEzIEBACisyMDA5LTA5LTExICBBbGV4 ZXkgUHJvc2t1cnlha292ICA8YXBAd2Via2l0Lm9yZz4KKworICAgICAgICBSZXZpZXdlZCBieSBO T0JPRFkgKE9PUFMhKS4KKworICAgICAgICBodHRwczovL2J1Z3Mud2Via2l0Lm9yZy9zaG93X2J1 Zy5jZ2k/aWQ9MjkyMDcKKyAgICAgICAgQWRkIGNoZWNrcyBmb3IgdXNpbmcgV2ViQ29yZSBKUyBj b250ZXh0IG9uIHNlY29uZGFyeSB0aHJlYWRzCisKKyAgICAgICAgKiBiaW5kaW5ncy9qcy9KU0RP TVdpbmRvd0Jhc2UuY3BwOiAoV2ViQ29yZTo6SlNET01XaW5kb3dCYXNlOjpjb21tb25KU0dsb2Jh bERhdGEpOgorICAgICAgICBTZXQgdGhlIG5ldyBkZWJ1Zy1vbmx5IG1haW5UaHJlYWRPbmx5IGZs YWcuCisKIDIwMDktMDktMTEgIEJyaWFuIFdlaW5zdGVpbiAgPGJ3ZWluc3RlaW5AYXBwbGUuY29t PgogCiAgICAgICAgIFJldmlld2VkIGJ5IFRpbW90aHkgSGF0Y2hlci4KSW5kZXg6IFdlYkNvcmUv YmluZGluZ3MvanMvSlNET01XaW5kb3dCYXNlLmNwcAo9PT09PT09PT09PT09PT09PT09PT09PT09 PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09Ci0tLSBXZWJDb3JlL2Jp bmRpbmdzL2pzL0pTRE9NV2luZG93QmFzZS5jcHAJKHJldmlzaW9uIDQ4MjU1KQorKysgV2ViQ29y ZS9iaW5kaW5ncy9qcy9KU0RPTVdpbmRvd0Jhc2UuY3BwCSh3b3JraW5nIGNvcHkpCkBAIC0xNjcs NiArMTY3LDkgQEAgSlNHbG9iYWxEYXRhKiBKU0RPTVdpbmRvd0Jhc2U6OmNvbW1vbkpTRwogICAg IGlmICghZ2xvYmFsRGF0YSkgewogICAgICAgICBnbG9iYWxEYXRhID0gSlNHbG9iYWxEYXRhOjpj cmVhdGVMZWFrZWQoKS5yZWxlYXNlUmVmKCk7CiAgICAgICAgIGdsb2JhbERhdGEtPnRpbWVvdXRD aGVja2VyLnNldFRpbWVvdXRJbnRlcnZhbCgxMDAwMCk7IC8vIDEwIHNlY29uZHMKKyNpZm5kZWYg TkRFQlVHCisgICAgICAgIGdsb2JhbERhdGEtPm1haW5UaHJlYWRPbmx5ID0gdHJ1ZTsKKyNlbmRp ZgogICAgIH0KIAogICAgIHJldHVybiBnbG9iYWxEYXRhOwo=