There Might Be A New Linux DRM Driver

An interesting mailing list question was posed yesterday to DRM developers, which raises the question there might be a new Linux DRM kernel driver being hacked.

From Sascha Hauer, a developer at the German-based Pengutronix, wrote an email entitled libdrm and platform devices? It was quite a simple question, but in the email he mentioned he was working on a DRM driver, which is the interesting part. "I have a drm device on the platform bus, similar to the exynos driver. right now libdrm (at least the tests included in libdrm) refuses to open the device because i915, nouveau, radeon and vmwgfx is all they know about. Looking at the libdrm code it is not obvious how to fix this (except for adding "exynos", "mydevice", "myotherdevice" to the module table which seems awkward and not very futureproof)."

So what is "mydevice" and "myotherdevice" from his message? Pengutronix mostly works on ARM hardware and they describe themselves as providing "Linux solutions for science and industry." Last year this same German developer published a new open-source KMS driver that was for the Freescale i.MX51 and i.MX53 SoCs. These ARM SoCs aren't too popular, but are found in some Google Chrome OS netbooks. That driver from last year didn't have proper in-kernel memory management (GEM/TTM) nor was it accompanied by an open-source user-space driver (DDX and Mesa/Gallium3D). There was also other missing features to this driver from last year.

This yet-to-be-announced DRM driver from Pengutronix could be a re-hashed version of their Freescale DRM driver from last year (though it's a bit late to be worrying about libdrm with it), which has yet to make it into the mainline Linux kernel, or it might be something entirely new. It's been a while since hearing anything about the Freescale i.MX51/i.MX53 DRM driver and even when it was launched it wasn't too exciting, so hopefully this will be some new open-source driver initiative to hear about in the near future.

These patches published by Sascha Hauer against the generic xf86-video-modesetting driver would also seem to indicate that there isn't yet a DDX driver for the new KMS-enabled hardware. Additionally, these patches make the generic mode-setting DDX driver work with non-PCI devices (i.e. ARM hardware).