So it turns out that my Tegra-2 based device does not support depth texture.
I'm trying to come up with an alternative, but what are my options?
I've heard people encode/decode depth into RGB? That sound like a mess to me.
Is it not possible to render to a monochromatic luminance texture instead?
Attach a 16-bit single channel texture to a framebuffer, and render in that, putting the depth in the single channel?
Is such a thing supported, or does render-to-texture always have to be to RGBA or (if supported) depth?