Re: [SeaBIOS] [PATCH v6 1/3] Boot up with a splash picture configurable

[Wayne Xia]

于 2011-7-8 8:05, Kevin O'Connor 写道:
> On Thu, Jul 07, 2011 at 05:38:09PM +0800, Wayne Xia wrote:
> > added an option to let seabios show BMP file as a logo.
> >
> > Signed-off-by: Wayne Xia
>
> Thanks.  As before, I'm okay with patch 1 and 2 and if there are no
> other comments I'll commit them.
>
> I don't see the need for patch 3 - if someone wants to show a
> bootsplash they can pass in "menu=on" or QEmu can imply "menu=on" when
> a bootsplash is requested.

thanks, agree that -menu would be enough.
> [...]
> > +static void raw_data_format_adjust_24bpp(u8 *src, u8 *dest, int width,
> >                                    int height, int bytes_per_line_dest)
> >
> > +{
> > int cy = height;
> > u8 *pd, *ps, *pd_valid, *ps_lastline;
> > int bytes_per_line_src = 3 * width;
> > while (cy>  0) {
> >    ps = bytes_per_line_src*cy + src - 1;
> >    /*  verical line sequence reverse */
> >    pd = bytes_per_line_dest * (height - cy + 1) + dest - 1;
> >    ps_lastline = ps - bytes_per_line_src;
> >    pd_valid = pd - (bytes_per_line_dest - bytes_per_line_src);
> >    while (pd>  pd_valid) {
> >        *(pd--) = 0;
> >    }
> >    while (ps>  ps_lastline) {
> >        *(pd--) = *(ps--);
> >    }
> >    cy--;
> > }
> >
> > +}
>
> Can't this be written as:
>
> static void raw_data_format_adjust_24bpp(u8 *src, u8 *dest, int width,
>                                           int height, int bytes_per_line_dest)
> {
>      int bytes_per_line_src = 3 * width;
>      int i;
>      for (i=0; i<height; i++)
>          memcpy(&dest[(height - i - 1) * bytes_per_line_dest]
>                 ,&src[i * bytes_per_line_src]
>                 , bytes_per_line_src);
> }

changed, now code become more simple.
> [...]
> > @@ -241,7 +293,10 @@ done:
> >       free(picture);
> >       free(vesa_info);
> >       free(mode_info);
> > free(jpeg);
> > if (jpeg != NULL)
> >    free(jpeg);
> > if (bmp != NULL)
> >    free(bmp);
>
> Minor nit - free() is defined to check for NULL, so it is safe to just
> unconditionally call free().

changed.
> -Kevin