Native surface creation

[Brian Paul]

Jon Smirl wrote:
> On Sun, 13 Mar 2005 15:50:03 -0700, Brian Paul
> <brian.paul at tungstengraphics.com> wrote:
> 
>>In any case, the point of all this isn't really stereo, but
>>consideration of other mode attributes that we might have to be
>>concerned with.  Other examples might be interlacing or syncing to
>>external signals and the fact that the application may need to control
>>those things to some degree.
> 
> 
> My plan was to add all of these as drop downs in a list box. The mode
> names are just strings that select from a set of fbdev fb_mode
> structures.

You've mentioned this drop-down list box twice now but I don't know 
what you're talking about.  Is this a particular application you're 
writing?


> One upside is that the list could contain things like:
> HDTV 1080p
> HDTV 1080i
> HDTV 720p
> 
> On the other hand let's say you have a movie that plays at 640x480 and
> you want to go full screen. I do need to make a way for you to figure
> out the string for the 640x480 mode.
> 
> I could make a sysfs attribute named match. Write an fb_mode structure
> to it and then read the attribute back and you get the mode string for
> the closest matching mode.
> 
> struct fb_videomode {
>         const char *name;       /* optional */
>         u32 refresh;            /* optional */
>         u32 xres;
>         u32 yres;
>         u32 pixclock;
>         u32 left_margin;
>         u32 right_margin;
>         u32 upper_margin;
>         u32 lower_margin;
>         u32 hsync_len;
>         u32 vsync_len;
>         u32 sync;
>         u32 vmode;
>         u32 flag;
> };
> 
> #define FB_SYNC_HOR_HIGH_ACT    1       /* horizontal sync high active  */
> #define FB_SYNC_VERT_HIGH_ACT   2       /* vertical sync high active    */
> #define FB_SYNC_EXT             4       /* external sync                */
> #define FB_SYNC_COMP_HIGH_ACT   8       /* composite sync high active   */
> #define FB_SYNC_BROADCAST       16      /* broadcast video timings      */
>                                         /* vtotal = 144d/288n/576i => PAL  */
>                                         /* vtotal = 121d/242n/484i => NTSC */
> #define FB_SYNC_ON_GREEN        32      /* sync on green */
> 
> #define FB_VMODE_NONINTERLACED  0       /* non interlaced */
> #define FB_VMODE_INTERLACED     1       /* interlaced   */
> #define FB_VMODE_DOUBLE         2       /* double scan */
> 
> flags
> #define FB_MODE_IS_DETAILED     1
> #define FB_MODE_IS_STANDARD     2
> #define FB_MODE_IS_VESA         4
> #define FB_MODE_IS_CALCULATED   8
> #define FB_MODE_IS_FIRST        16
> #define FB_MODE_IS_FROM_VAR     32
> 
> I need to fix this to use enums. We could remove the unneeded fields
> from this structure and add  it to egl.h.

This would not be appropriate for egl.h.  Traditionally, the various 
OpenGL interfaces don't expose structures or machine-specific details.

Here's more of what I hand in mind for an mode API in EGL:

typedef unsigned int EGLMode;

EGLBoolean eglChooseMode(EGLDisplay dpy, const EGLint *attrib_list, 
EGLMode *modes, EGLint modes_size, EGLint *num_modes);

EGLBoolean eglGetModes(EGLDisplay dpy, EGLint screen_number, EGLMode 
*modes, EGLint modes_size, EGLint *num_modes);

EGLBoolean eglGetModeAttribMESA(EGLDisplay dpy, EGLMode mode, EGLint 
attribute, EGLint *value);

They mirror the EGL Config functions.  Among the attribute tokens 
would be: EGL_WIDTH, EGL_HEIGHT, EGL_DEPTH, EGL_REFRESH_RATE, etc.


It's clear that you're working at a lower level of implementation 
detail than I am in trying to specify an EGL API extension.

-Brian