The xf86-video-modesetting v0.5.0 driver was released on Thursday with a few important changes.
This updated X.Org video driver, which is used in cases where a kernel mode-setting (KMS) driver is available on the system but not any hardware-specific X.Org (DDX) driver, is a relatively minor update but some changes that will affect users. The xf86-video-modesetting driver is like a VESA driver for the KMS-enabled world. Among the DRM/KMS drivers lacking a DDX driver -- where xf86-video-modesetting is most useful -- include the DisplayLink KMS USB setup
, a virtual QEMU KMS driver based on Cirrus
, Matrox MGA
, and ASpeed
Changes by David Airlie of Red Hat to the xf86-video-modesetting 0.5.0 release add in the platform bus support and output slave support, which is part of the ongoing DRI2 PRIME
work. In the context of this generic mode-setting driver, the output slave support could allow this updated mode-setting driver to be used in conjunction with the DisplayLink
KMS driver where the USB graphics adapter is displaying a window that's been rendered by another GPU that's playing in the PRIME-enabled world.
Basically, hot-plug the USB GPU
with the different DisplayLink adapters being USB-based (and working fairly well based upon my testing so far of a Plugable.com
devices) while letting another GPU (Intel / Radeon / Nouveau are the other driver players right now with this new X.Org technology) take care of the actual rendering prior to being offloaded to the USB-based display.
The output slave concept in an X.Org world was talked about last year in X.Org Server GPU Hot-Plugging Moves Along
but only now is coming to mainline fruition. But if this output slave support doesn't interest you, there isn't much to see out of the xf86-video-modesetting 0.5.0 update. The release announcement is on xorg-announce