<html xmlns:v="urn:schemas-microsoft-com:vml" xmlns:o="urn:schemas-microsoft-com:office:office" xmlns:w="urn:schemas-microsoft-com:office:word" xmlns:m="http://schemas.microsoft.com/office/2004/12/omml" xmlns="http://www.w3.org/TR/REC-html40"> <head> <meta http-equiv="Content-Type" content="text/html; charset=us-ascii"> <meta name="Generator" content="Microsoft Word 14 (filtered medium)"> <style></style> </head> <body lang="ZH-CN" link="blue" vlink="purple" style="text-justify-trim:punctuation"> <div class="WordSection1"> More questions:<o:p></o:p> <![if !supportLists]>1. <![endif]>I searched inside mesa code, there is wl_display_create_queue(), but no wl_event_queue_destroy(), is it a bug?<o:p></o:p> <![if !supportLists]>2. <![endif]>I searched inside Weston code, there is no wl_buffer_send_release(), is there other way to send the event of WL_BUFFER_RELEASE?<o:p></o:p> <![if !supportLists]>3. <![endif]>What’s the design philosophy of wl_buffer_send_release?<o:p></o:p> <![if !supportLists]>a) <![endif]>I had thought the event is sent when server doesn’t use wl_buffer any longer<o:p></o:p> <![if !supportLists]> i. <![endif]>when a new wl_buffer is committed (release previous committed buffer)<o:p></o:p> <![if !supportLists]> ii. <![endif]>When the corresponding wl_surface is destroying (release the committed buffer)<o:p></o:p> <![if !supportLists]> iii. <![endif]>When a second wl_buffer is attached before commit (release the previous attached one)<o:p></o:p> <![if !supportLists]>b) <![endif]>However, the above assumption doesn’t work well with mesa. With nested server in webkit.<o:p></o:p> Seems mesa requires wayland server send wl_buffer_send_release upon attach.<o:p></o:p> <o:p> </o:p> <o:p> </o:p> <o:p> </o:p> <div> <div style="border:none;border-top:solid #B5C4DF 1.0pt;padding:3.0pt 0cm 0cm 0cm"> From: wayland-devel-bounces+halley.zhao=intel.com@lists.freedesktop.org [mailto:wayland-devel-bounces+halley.zhao=intel.com@lists.freedesktop.org] On Behalf Of Zhao, Halley Sent: Monday, September 02, 2013 10:12 AM To: wayland-devel@lists.freedesktop.org Subject: how to deal with the last video frame buffer (wl_buffer)?<o:p></o:p> </div> </div> <o:p> </o:p> When a wl_buffer (of video frame) is committed to wayland server, this buffer keeps being used by server until a second buffer is committed.<o:p></o:p> It means a wl_buffer isn’t ready to destroy in frame callback, but wl_buffer_send_release(). And the buffer (in media pipeline) should also be retained for some time after it is committed to server side.<o:p></o:p> <o:p> </o:p> This expose additional questions for the last video buffer.<o:p></o:p> <![if !supportLists]>1. <![endif]>Should we commit empty wl_buffer upon stopping of video playback?<o:p></o:p> <![if !supportLists]>a) <![endif]>Does Weston support it ? (use empty buffer to replace previous committed buffer)<o:p></o:p> <![if !supportLists]>b) <![endif]>If yes, is there any example to use it?<o:p></o:p> <![if !supportLists]>2. <![endif]>When to destroy the event queue if it is used?<o:p></o:p> Usually media creates an event queue, wl_buffer is also set to this queue. It comes a question that when to destroy this event queue, since wl_buffer_send_release() (on this event queue) happens after media stop. <o:p></o:p> <![if !supportLists]>a) <![endif]>Should we have an event like wl_surface_send_release() to give the chance to destroy this event queue?<o:p></o:p> <![if !supportLists]>b) <![endif]>If not, then we have to destroy this event queue upon wl_buffer_send_release of the last buffer. <o:p></o:p> <![if !supportLists]> i. <![endif]>Is it ok to destroy the event queue in the callback function of the event in this queue?<o:p></o:p> <![if !supportLists]> ii. <![endif]>It creates complexity to sycn on the last buffer release event in media.<o:p></o:p> <o:p> </o:p> <o:p> </o:p> </div> </body> </html>