Hi,

On Fri, 2025-07-11 at 11:01 +0400, Marc-André Lureau wrote:
> Hi
> 
> On Thu, Jul 10, 2025 at 4:24 PM Peter Maydell
> <[email protected]> wrote:
> > 
> > On Sat, 24 May 2025 at 18:37, <[email protected]> wrote:
> > > 
> > > From: Weifeng Liu <[email protected]>
> > > 
> > > When gl=on, scale_x and scale_y were set to 1 on startup that
> > > didn't
> > > reflect the real situation of the scan-out in free scale mode,
> > > resulting
> > > in incorrect cursor coordinates to be sent when moving the mouse
> > > pointer. Simply updating the scales before rendering the image
> > > fixes
> > > this issue.
> > > 
> > > Signed-off-by: Weifeng Liu <[email protected]>
> > > Message-ID: <[email protected]>
> > > Acked-by: Gerd Hoffmann <[email protected]>
> > > Acked-by: Marc-André Lureau <[email protected]>
> > 
> > Hi; Coverity complains about this change CID 1610328):
> > 
> > > @@ -50,8 +52,14 @@ void gd_gl_area_draw(VirtualConsole *vc)
> > > 
> > >      gtk_gl_area_make_current(GTK_GL_AREA(vc->gfx.drawing_area));
> > >      gs = gdk_window_get_scale_factor(gtk_widget_get_window(vc-
> > > >gfx.drawing_area));
> > > -    pw = gtk_widget_get_allocated_width(vc->gfx.drawing_area) *
> > > gs;
> > > -    ph = gtk_widget_get_allocated_height(vc->gfx.drawing_area) *
> > > gs;
> > > +    fbw = surface_width(vc->gfx.ds);
> > > +    fbh = surface_height(vc->gfx.ds);
> > 
> > Here we now unconditionally dereference vc->gfx.ds at the start of
> > gd_gl_area_draw().
> > 
> > But towards the end of this function we have a NULL check:
> > 
> >         if (!vc->gfx.ds) {
> >             return;
> >         }
> > 
> > Either vc->gfx.ds can be NULL, in which case we need some
> > kind of guard on these surface_width() and surface_height()
> > calls; or else it can't, and the NULL check later is dead code.
> > Which is correct ?
> 
> Given that it's simply called from a GTK callback, it can be NULL. I
> think we should simply return in this case, or perhaps use the older
> code path as a fallback. Weifeng, wdyt?

It makes sense to skip drawing when ds is NULL, although that shouldn’t
happen since dpy_gfx_replace_surface installs a dummy DisplaySurface
for any NULL input. I’ll draft a patch shortly to add this check (and
possibly other refinements).

Best regards,
Weifeng

Reply via email to