97646 2012-09-26 02:03:51 -0700 [EFL][WK2] WebProcess keeps on waiting for UIProcess 2012-10-11 21:56:54 -0700 1 1 1 Unclassified WebKit WebKit EFL 528+ (Nightly build) Other Linux RESOLVED FIXED P2 Normal --- 0 heejin.r.chung webkit-unassigned gyuyoung.kim lucas.de.marchi rakuco webkit.review.bot yael oldest_to_newest 728327 0 heejin.r.chung 2012-09-26 02:03:51 -0700 WebKit2 Efl with Coordinated Graphics is not working because WebProcess never gets RenderNextFrame message from UIProcess. 728355 1 165759 heejin.r.chung 2012-09-26 02:38:01 -0700 Created attachment 165759 Patch 730884 2 165759 yael 2012-09-28 12:27:02 -0700 Comment on attachment 165759 Patch View in context: https://bugs.webkit.org/attachment.cgi?id=165759&action=review > Source/WebKit2/UIProcess/API/efl/ewk_view.cpp:533 > + priv->viewportHandler->setRendererActive(true); Is this really needed every time we enter AC or only the first time? As an aside, PageClientImpl::enterAcceleratedCompositingMode() is not implemented in any other port, so we should probably move the initialization code out of it instead of adding to it. 739346 3 yael 2012-10-10 15:37:16 -0700 (In reply to comment #2) > (From update of attachment 165759 [details]) > View in context: https://bugs.webkit.org/attachment.cgi?id=165759&action=review > > > Source/WebKit2/UIProcess/API/efl/ewk_view.cpp:533 > > + priv->viewportHandler->setRendererActive(true); > > Is this really needed every time we enter AC or only the first time? > As an aside, PageClientImpl::enterAcceleratedCompositingMode() is not implemented in any other port, so we should probably move the initialization code out of it instead of adding to it. After http://trac.webkit.org/changeset/130389 ewk_view_accelerated_compositing_mode_enter is called only once. This patch looks good to me. 739570 4 gyuyoung.kim 2012-10-10 20:45:41 -0700 Don't you need to set setRendererActive(false) in ewk_view_accelerated_compositing_mode_exit() ? 740195 5 yael 2012-10-11 13:14:33 -0700 (In reply to comment #4) > Don't you need to set setRendererActive(false) in ewk_view_accelerated_compositing_mode_exit() ? What is the benefit of doing that? We need to call setActive(true) to trigger the first renderNextFrame, so that we can start rendering. 740220 6 165759 kenneth 2012-10-11 13:33:27 -0700 Comment on attachment 165759 Patch OK, r=me then. Anyway we should rename the class to follow current trunk style and inherit from the ViewportPageController, like Qt. But that is not directly related to this, but maybe something for you Yael? 740229 7 165759 webkit.review.bot 2012-10-11 13:41:31 -0700 Comment on attachment 165759 Patch Clearing flags on attachment: 165759 Committed r131091: <http://trac.webkit.org/changeset/131091> 740230 8 webkit.review.bot 2012-10-11 13:41:35 -0700 All reviewed patches have been landed. Closing bug. 740277 9 yael 2012-10-11 14:49:22 -0700 (In reply to comment #6) > (From update of attachment 165759 [details]) > OK, r=me then. Anyway we should rename the class to follow current trunk style and inherit from the ViewportPageController, like Qt. But that is not directly related to this, but maybe something for you Yael? Filed https://bugs.webkit.org/show_bug.cgi?id=99101 for that. 740616 10 gyuyoung.kim 2012-10-11 21:56:54 -0700 (In reply to comment #5) > (In reply to comment #4) > > Don't you need to set setRendererActive(false) in ewk_view_accelerated_compositing_mode_exit() ? > > What is the benefit of doing that? > We need to call setActive(true) to trigger the first renderNextFrame, so that we can start rendering. First, I'm not expert for graphics area. To my experience, however, we have been set true/false in a pair of functions generally if there is needs to set specific setting. In this case, ewk_view_accelerated_compositing_mode_exit() might don't need to set setActive(false) because ewk_view_accelerated_compositing_mode_exit() will be called only once when existing program, right ? If so, we might not call setActive(false) from a functionality perspective. But, IMO, it is not bad to call setActive(false) in _exit() from a pair of functions perspective. If the _exit() will be called many times, isn't it better or safer to call setActive(false) in _exit() ? Anyway, patch was already landed. I'm just wondering about this. 165759 2012-09-26 02:38:01 -0700 2012-10-11 13:41:31 -0700 Patch bug-97646-20120926183815.patch text/plain 2939 heejin.r.chung SW5kZXg6IFNvdXJjZS9XZWJLaXQyL0NoYW5nZUxvZwo9PT09PT09PT09PT09PT09PT09PT09PT09 PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09Ci0tLSBTb3VyY2UvV2Vi S2l0Mi9DaGFuZ2VMb2cJKHJldmlzaW9uIDEyOTYxNCkKKysrIFNvdXJjZS9XZWJLaXQyL0NoYW5n ZUxvZwkod29ya2luZyBjb3B5KQpAQCAtMSwzICsxLDIxIEBACisyMDEyLTA5LTI2ICBSZWdpbmEg Q2h1bmcgIDxoZWVqaW4uci5jaHVuZ0BzYW1zdW5nLmNvbT4KKworICAgICAgICBbRUZMXVtXSzJd IFdlYlByb2Nlc3Mga2VlcHMgb24gd2FpdGluZyBmb3IgVUlQcm9jZXNzCisgICAgICAgIGh0dHBz Oi8vYnVncy53ZWJraXQub3JnL3Nob3dfYnVnLmNnaT9pZD05NzY0NgorCisgICAgICAgIFJldmll d2VkIGJ5IE5PQk9EWSAoT09QUyEpLgorCisgICAgICAgIFNldCBMYXllclRyZWVSZW5kZXJlciB0 byBhY3RpdmUgc28gVUlQcm9jZXNzIGNhbiBzZW5kIFJlbmRlck5leHRGcmFtZSBtZXNzYWdlIHRv IFdlYlByb2Nlc3MKKyAgICAgICAgYWZ0ZXIgYWNjZWxlcmF0ZWQgY29tcG9zaXRpbmcoQUMpIG1v ZGUgaXMgZW50ZXJlZCwgc28gV2ViUHJvY2VzcyBjYW4gc3RhcnQgZmx1c2hpbmcgY2hhbmdlcyB0 byBVSVByb2Nlc3MuCisKKyAgICAgICAgKiBVSVByb2Nlc3MvQVBJL2VmbC9FZmxWaWV3cG9ydEhh bmRsZXIuY3BwOiBBZGRlZCBBUEkgZm9yIHNldHRpbmcgTGF5ZXJUcmVlUmVuZGVyZXIncyBhY3Rp dmUgc3RhdGUuCisgICAgICAgIChXZWJLaXQ6OkVmbFZpZXdwb3J0SGFuZGxlcjo6c2V0QWN0aXZl KToKKyAgICAgICAgKFdlYktpdCk6CisgICAgICAgICogVUlQcm9jZXNzL0FQSS9lZmwvRWZsVmll d3BvcnRIYW5kbGVyLmg6CisgICAgICAgIChFZmxWaWV3cG9ydEhhbmRsZXIpOgorICAgICAgICAq IFVJUHJvY2Vzcy9BUEkvZWZsL2V3a192aWV3LmNwcDogU2V0cyBMYXllclRyZWVSZW5kZXJlcidz IGFjdGl2ZSBzdGF0dXMgdG8gdHJ1ZSBhZnRlciBlbnRlcmluZyBBQyBtb2RlLgorICAgICAgICAo ZXdrX3ZpZXdfYWNjZWxlcmF0ZWRfY29tcG9zaXRpbmdfbW9kZV9lbnRlcik6CisKIDIwMTItMDkt MjUgIFJhcGhhZWwgS3VibyBkYSBDb3N0YSAgPHJhcGhhZWwua3Viby5kYS5jb3N0YUBpbnRlbC5j b20+CiAKICAgICAgICAgW0RSVF1bV1RSXSBTdXBwb3J0IG92ZXJyaWRpbmcgdGhlICdXZWJLaXRE aXNwbGF5SW1hZ2VzS2V5JyBwcmVmZXJlbmNlCkluZGV4OiBTb3VyY2UvV2ViS2l0Mi9VSVByb2Nl c3MvQVBJL2VmbC9FZmxWaWV3cG9ydEhhbmRsZXIuY3BwCj09PT09PT09PT09PT09PT09PT09PT09 PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT0KLS0tIFNvdXJjZS9X ZWJLaXQyL1VJUHJvY2Vzcy9BUEkvZWZsL0VmbFZpZXdwb3J0SGFuZGxlci5jcHAJKHJldmlzaW9u IDEyOTYwMSkKKysrIFNvdXJjZS9XZWJLaXQyL1VJUHJvY2Vzcy9BUEkvZWZsL0VmbFZpZXdwb3J0 SGFuZGxlci5jcHAJKHdvcmtpbmcgY29weSkKQEAgLTUzLDYgKzUzLDExIEBAIERyYXdpbmdBcmVh UHJveHkqIEVmbFZpZXdwb3J0SGFuZGxlcjo6ZHIKICAgICByZXR1cm4gZXdrX3ZpZXdfcGFnZV9n ZXQobV92aWV3V2lkZ2V0KS0+ZHJhd2luZ0FyZWEoKTsKIH0KIAordm9pZCBFZmxWaWV3cG9ydEhh bmRsZXI6OnNldFJlbmRlcmVyQWN0aXZlKGJvb2wgYWN0aXZlKQoreworICAgIGRyYXdpbmdBcmVh KCktPmxheWVyVHJlZUNvb3JkaW5hdG9yUHJveHkoKS0+bGF5ZXJUcmVlUmVuZGVyZXIoKS0+c2V0 QWN0aXZlKGFjdGl2ZSk7Cit9CisKIHZvaWQgRWZsVmlld3BvcnRIYW5kbGVyOjpkaXNwbGF5KGNv bnN0IEludFJlY3QmIHJlY3QpCiB7CiAgICAgV2ViQ29yZTo6VHJhbnNmb3JtYXRpb25NYXRyaXgg bWF0cml4OwpJbmRleDogU291cmNlL1dlYktpdDIvVUlQcm9jZXNzL0FQSS9lZmwvRWZsVmlld3Bv cnRIYW5kbGVyLmgKPT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09 PT09PT09PT09PT09PT09PT09PT09PQotLS0gU291cmNlL1dlYktpdDIvVUlQcm9jZXNzL0FQSS9l ZmwvRWZsVmlld3BvcnRIYW5kbGVyLmgJKHJldmlzaW9uIDEyOTYwMSkKKysrIFNvdXJjZS9XZWJL aXQyL1VJUHJvY2Vzcy9BUEkvZWZsL0VmbFZpZXdwb3J0SGFuZGxlci5oCSh3b3JraW5nIGNvcHkp CkBAIC00OCw3ICs0OCw3IEBAIHB1YmxpYzoKICAgICB2b2lkIHVwZGF0ZVZpZXdwb3J0U2l6ZShj b25zdCBXZWJDb3JlOjpJbnRTaXplJiB2aWV3cG9ydFNpemUpOwogICAgIHZvaWQgc2V0VmlzaWJs ZUNvbnRlbnRzUmVjdChjb25zdCBXZWJDb3JlOjpJbnRQb2ludCYsIGZsb2F0LCBjb25zdCBXZWJD b3JlOjpGbG9hdFBvaW50Jik7CiAgICAgdm9pZCBkaWRDaGFuZ2VDb250ZW50c1NpemUoY29uc3Qg V2ViQ29yZTo6SW50U2l6ZSYgc2l6ZSk7Ci0KKyAgICB2b2lkIHNldFJlbmRlcmVyQWN0aXZlKGJv b2wpOwogcHJpdmF0ZToKICAgICBleHBsaWNpdCBFZmxWaWV3cG9ydEhhbmRsZXIoRXZhc19PYmpl Y3QqKTsKIApJbmRleDogU291cmNlL1dlYktpdDIvVUlQcm9jZXNzL0FQSS9lZmwvZXdrX3ZpZXcu Y3BwCj09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09 PT09PT09PT09PT09PT0KLS0tIFNvdXJjZS9XZWJLaXQyL1VJUHJvY2Vzcy9BUEkvZWZsL2V3a192 aWV3LmNwcAkocmV2aXNpb24gMTI5NjAxKQorKysgU291cmNlL1dlYktpdDIvVUlQcm9jZXNzL0FQ SS9lZmwvZXdrX3ZpZXcuY3BwCSh3b3JraW5nIGNvcHkpCkBAIC01MzAsNiArNTMwLDcgQEAgYm9v bCBld2tfdmlld19hY2NlbGVyYXRlZF9jb21wb3NpdGluZ19tbwogICAgICAgICByZXR1cm4gZmFs c2U7CiAgICAgfQogCisgICAgcHJpdi0+dmlld3BvcnRIYW5kbGVyLT5zZXRSZW5kZXJlckFjdGl2 ZSh0cnVlKTsKICAgICByZXR1cm4gdHJ1ZTsKIH0KIAo=