David Herrmann

Miracast is the name of a WiFi-Alliance certification program for the WiFi-Display standard. It basically defines a "wireless HDMI-cable" so you can connect monitors via WiFi. Some Android vendors implement it, Microsoft ships it with Windows 8.1 and with OpenWFD we now also have the first Open-Source implementation available. This talk shows what Miracast is, how it works, and how you can use it on your favourite linux distribution already.

The Wifi-Display standard (abbr. WFD) was created to define a common way to connect TVs and monitors to your PC or smartphone. It is not meant as media-file streaming protocol like DLNA, but rather as a "wireless HDMI-cable". It provides only a single video/audio stream from a source to sink, is optimized for low-latency and allows easy setup.

WFD uses Wifi-Direct (wifi peer-2-peer / Wifi-P2P) to create a direct connection between two devices. A mode-negotiation follows and once all parameters are clear, an mpeg stream is established. Several extended features like split video/audio-sinks, enhanced timing-protocols or more are available.

OpenWFD is the first Open-Source implementation of WFD. It is targeted at linux and provides some example source/sink daemons so you can already use it. However, proper integration into the linux software-stack is still ongoing and the final setup will probably differ highly from the current project state.

How software projects around kmscon try to provide multi-seat support, hardware-accelerated rendering, full internationalization support and more to the linux-console by moving CONFIG_VT into user-space.

CONFIG_VT is the kernel configuration option that enables virtual terminals in the kernel. Initially written by Linus himself, it has been around since 1991. However, it hasn't been updated much recently and thus lacks a big amount of functionality, namely multi-seat support, full Unicode font rendering, XKB-like keyboard handling, hardware-accelerated rendering, vt220-vt510 compatibility and much more.

This talk will be split into two parts. In the first half I will be talking about the terminal emulator in the kernel, its limitations and how kmscon provides a modern replacement that moves terminal-emulation out of the kernel into user-space where it belongs to.

The second half will be about the ancient POSIX VT API. I will explain why it was needed in the first place and discuss why current linux system still depend on it. We tried several replacements, including system-compositors, dbus-VTs and CUSE-based VTs and I will compare them to the classic POSIX VT API to show how the issues of VTs can actually be solved by dropping CONFIG_VT entirely.