[Spice-devel] [PATCH spice-gtk 6/7] egl: don't terminate display

Victor Toso lists at victortoso.com
Mon May 23 10:19:31 UTC 2016


Hi,

On Mon, May 23, 2016 at 05:53:35AM -0400, Marc-André Lureau wrote:
> Hi
>
> ----- Original Message -----
> > Hi,
> >
> > On Mon, May 23, 2016 at 11:31:52AM +0200, Marc-André Lureau wrote:
> > > Hi
> > >
> > > On Mon, May 23, 2016 at 10:21 AM, Victor Toso <lists at victortoso.com> wrote:
> > > > Could you also include a fix that calls eglTerminate ?
> > >
> > > Since this is global resources, spice-gtk (a library) can't do that by
> > > itself without risking to break somebody else code.
> >
> > Sorry, I don't really follow. The spice-gtk lib does eglInitialize on
> > spice_egl_init but we should not eglTerminate at some point?
>
> Yes, " Initializing an already initialized EGL display connection has
> no effect besides returning the version numbers. "

Right.

> this is not the case with eglterminate, we can't call it as egl may be
> used by somebody else code.

So far I understand that we might be calling eglTerminate when we still
will use it the EGLDisplay which I thought we could eglTerminate when
session ends, for instance.

If spice-gtk does eglInitialize (not the application) I would like to
understaund in this case why can't eglTerminate. Is it because the egl
API does not inform that spice_egl_init was the one that initialized the
EGLDisplay ?

Either way, this really needs a comment in the unrealize function
instead of just removing the function call. We will see leaks from this.

>
> >
> > >
> > > --
> > > Marc-André Lureau
> > _______________________________________________
> > Spice-devel mailing list
> > Spice-devel at lists.freedesktop.org
> > https://lists.freedesktop.org/mailman/listinfo/spice-devel
> > 


More information about the Spice-devel mailing list