<html> <head> <base href="https://bugzilla.gnome.org/" /> </head> <body> <div> <a class="bz_bug_link bz_status_NEW " title="NEW - memory leak" href="https://bugzilla.gnome.org/show_bug.cgi?id=761312#c2">Comment # 2</a> on <a class="bz_bug_link bz_status_NEW " title="NEW - memory leak" href="https://bugzilla.gnome.org/show_bug.cgi?id=761312">bug 761312</a> from <a href="page.cgi?id=describeuser.html&login=rstrode%40redhat.com" title="Ray Strode [halfline] <rstrode@redhat.com>"> Ray Strode [halfline]</a> <pre>Created <a href="attachment.cgi?id=320042&action=diff" name="attach_320042" title="wayland: Don't call set_busy twice on the same surface">attachment 320042</a> <a href="attachment.cgi?id=320042&action=edit" title="wayland: Don't call set_busy twice on the same surface">[details]</a> <a href='review?bug=761312&attachment=320042'>[review]</a> wayland: Don't call set_busy twice on the same surface If the compositor is using a shared memory buffer allocated by a client, then it's the client's responsibility to refrain from destroying the buffer until the compositor releases it. This is accomplished by taking a reference to the cairo surface assocatiated with the buffer after a frame, and dropping the reference when the compositor releases the buffer. In some cases though, the compositor doesn't release the buffer until a new buffer is set, so if we have staged drawing before the frame completes we can end up taking multiple references to the buffer and keeping it alive after it's released. This commit solves the problem by ensuring we only call _gdk_wayland_shm_surface_set_busy if isn't already busy.</pre> </div> <hr> You are receiving this mail because: <ul> <li>You are on the CC list for the bug.</li> </ul> </body> </html>