Michael Van Reenen <mvr@qnx.com> wrote:
: Patrick Mueller <patrick_mueller@oti.com> wrote:
: : OK, the “Odd mode list” thread answer was interesting, as we can now see
: : that the banshee driver uses the generic stuff, but we can’t get the mode
: : set. It ends up trapping. I’ll see if we can get a straight C sample, but
: : in the meantime, one thing I noticed is that for the generic modes, the
: : <disp_crtc_settings_t *settings> parm is used. We are setting a couple of
: : fields in there, and the doc mentions a
: : modefuncs.get_current_crtc_settings() that isn’t in the mode.h include file.
: : Bummer, as I was thinking we could initialize the disp_crtc_settings_t
: : structure with that function, then tweak what we wanted to.
: Unfortunately the docs are not quite up to date with the source code.
: (source code is current - docs are older) They should be updated in
: a future release.
: The get_current_crtc_settings() no longer exists.
: : So, we need to know which fields in disp_crtc_settings need to be set.
: If you fill in fields xres, yres, and refresh, you can call disp_crtc_calc()
: which will fill in the remaining members of the disp_crtc_settings_t
: structure. See the graphics DDK docs section Libraries for info on
: disp_crtc_calc.
It has been pointed out to me that you should also fill in the h_granularity
and v_granularity fields of the disp_crtc_settings_t structure.
Another thing of note relating to the “Odd mode list” thread is that when
before each call to modef.get_modeinfo the disp_mode_info_t structure should
be memset to 0.
: : –
: : ----------------------------------------------------
: : Patrick_Mueller@oti.com
: : ----------------------------------------------------