From linuxhippy at gmail.com Tue Aug 2 04:25:08 2016 From: linuxhippy at gmail.com (Clemens Eisserer) Date: Tue, 2 Aug 2016 06:25:08 +0200 Subject: [Xcb] OpenGL applications spends 20% of CPU-cycles in discard_reply Message-ID: Hi, I am currently working on an OpenGL based application which streams video content asynchronously to the GPU. For simplicity I am using freeglut3 as "toolkit". When running the application I get quite a strange profile, a serious amount of time is spent in the function "discard_reply" contained in libxcb: samples % image name app name symbol name 3665 17.5401 libxcb.so.1.1.0 streamapp discard_reply Unfourtunately even with callgraph profiling I don't get any helpful information: ------------------------------------------------------------------------------- 3670 100.000 libxcb.so.1.1.0 testprog xcb_discard_reply 3665 17.5401 libxcb.so.1.1.0 testprog discard_reply 3665 99.8638 libxcb.so.1.1.0 testprog discard_reply [self] Ideas how to track this issue down, would be very welcome. Thank you in advance, Clemens From bugzilla-daemon at freedesktop.org Tue Aug 2 07:56:22 2016 From: bugzilla-daemon at freedesktop.org (bugzilla-daemon at freedesktop.org) Date: Tue, 02 Aug 2016 07:56:22 +0000 Subject: [Xcb] [Bug 93959] libxcb (git version) fails to build In-Reply-To: References: Message-ID: https://bugs.freedesktop.org/show_bug.cgi?id=93959 Jos van Wolput changed: What |Removed |Added ---------------------------------------------------------------------------- Resolution|--- |FIXED Status|REOPENED |RESOLVED -- You are receiving this mail because: You are the assignee for the bug. You are the QA Contact for the bug. -------------- next part -------------- An HTML attachment was scrubbed... URL: From linuxhippy at gmail.com Tue Aug 2 12:13:15 2016 From: linuxhippy at gmail.com (Clemens Eisserer) Date: Tue, 2 Aug 2016 14:13:15 +0200 Subject: [Xcb] OpenGL applications spends 20% of CPU-cycles in discard_reply In-Reply-To: References: Message-ID: Hi again, Sorry my fault - I didn't notice the window manager used for the tests was performing composition - which explains all the unnecessary context switches as well as the IPC overhead I observed. With another WM everything performs now as expected. Best regards, Clemens From lars.segerlund at gmail.com Wed Aug 31 08:06:17 2016 From: lars.segerlund at gmail.com (Lars Segerlund) Date: Wed, 31 Aug 2016 10:06:17 +0200 Subject: [Xcb] xcb-proto DBE extension Message-ID: Hi, I made an dbe extension to xcb-proto and libxcb, ( from some xml I got from Thien-Thi Nguyen a long time ago, I have a couple of questions. First there is a type BUFFER_ATTRIBUTES which can be window or none, so how do I specify this in the xml ? ( GetBackBufferAttributes ). Thien-Ti suggested a union of window and none ... but I haven't found out how to do this. So how to handle this ? Second what do I do to contribute ? checkout latest apply build test and post to the list ? ( please reply to my mail also I am not on the mailinglist ... ) / regards, Lars Segerlund. -------------- next part -------------- An HTML attachment was scrubbed... URL: From pharris at opentext.com Wed Aug 31 16:06:03 2016 From: pharris at opentext.com (Peter Harris) Date: Wed, 31 Aug 2016 12:06:03 -0400 Subject: [Xcb] xcb-proto DBE extension In-Reply-To: References: Message-ID: On 2016-08-31 4:06 AM, Lars Segerlund wrote: > I made an dbe extension to xcb-proto and libxcb, ( from some xml I got > from Thien-Thi Nguyen a long time ago, I have a couple of questions. > > First there is a type BUFFER_ATTRIBUTES which can be window or none, so > how do I specify this in the xml ? > ( GetBackBufferAttributes ). > Thien-Ti suggested a union of window and none ... but I haven't found > out how to do this. > So how to handle this ? You use an "altenum" (which means, may contain one of the enum, but is not required). The "xproto:Window" enum contains only "None". For example: xproto > Second what do I do to contribute ? > > checkout latest apply build test and post to the list ? Yes. > ( please reply to my mail also I am not on the mailinglist ... ) The list is low volume and high signal. Subscribing to the list will make the previous steps go more smoothly. Peter Harris -- Open Text Connectivity Solutions Group Peter Harris http://connectivity.opentext.com/ Research and Development Phone: +1 905 762 6001 pharris at opentext.com Toll Free: 1 877 359 4866