Hans de Goede

BoF for people interested in / working on MIPI camera support / libcamera.

MIPI cameras are cameras where there is a sensor directly connected to the laptop/tablet/phone's SoC, as typically found in smartphones but recently also showing up in x86 laptops. For these raw sensors a whole bunch of the processing and sensor control needs to be done by software (optionally with assistance from an ISP in the SoC).

This makes these cameras a lot more complex then standard USB cameras which generate ready to use video frames.

Related talks @FOSDEM: https://fosdem.org/2023/schedule/event/sharpphotos/ https://fosdem.org/2023/schedule/event/linuxcameraapps/ https://fosdem.org/2023/schedule/event/allwinnercamera/

Presentation on the recent "flickerfree" boot work, where the goal is to go from the firmware splashscreen/logo to the GUI in one fluid experience without any modesets (causing the screen to blank) or transitions to text-mode. Topics discussed during the presentation are:

• The changes made to each of the involved components
• The upstream status of these changes
• Steps to take to enable this

A typical UEFI boot of a Linux system looks like this: * firmware shows a boot logo * shim loads, clears the screen * grub loads, shows a text menu * kernel efifb loads, clears the screen, may show a bunch of text messages * kernel drm/kms driver loads * bootsplash (plymouth) loads * GUI loads

Where the kms driver, plymouth and GUI load all may do a modeset resulting in the display loosing signal for a while. The constant screen clearing, various text-messages and the modesets all result in a quite restless boot experience.

My latest project at Red Hat is fixing all this so that we go straight from the logo shown into the GUI, while still being able to show error messages at all stages if necessary.

During this presentation I describe: * The changes made to each of the involved components * The upstream status of these changes * Steps to take to enable this

Benefits to the eco-system:

1) A nicer boot experience for all Linux systems makes Linux more user-friendly and will hopefully lead to wider boot experience

2) Customized embedded systems can already offer a similar experience by disabling the kernel's fbcon driver. There are 2 downsides to the current solutions: 2.1) They are per project customizations, there is no standard way to do this, 2.2) Without fbcon, no text output is possible at all and if something causes the GUI to not load there is no way to show a message to the end user (who will not have serial console access).

My work offers a standardized way to achieve the desired boot experience while keeping fbcon enabled so that error messages can be shown at any stage if necessary.

Modern laptops can use a lot less energy then laptops from a decade ago. But in order to actually get this low energy usage the operating system needs to make efficient use of the hardware. Linux supports a lot of hardware power-saving features, but many of them are disabled by default because they cause problems on certain devices or in certain, often corner-case, circumstances. This talk will describe and look into recent efforts to enable more power-saving features by default in such a way that this will not cause regressions. The end goal of these efforts is to shave of at least 2 watt of the typical idle power-consumption of 6-9 watt for recent (Hasswell or newer) laptops. The goal is to achieve this saving on an OOTB Fedora Workstation install without the user needing to do any manual tweaks.

This talk will look at the work recently done to enable compute support on GPUs supported by Nouveau. The focus of the talk will be on generating a representation of the OpenCL code that can be consumed by Nouveau, which is the biggest missing piece to get OpenCL running on Nouveau.

GPGPU support is something which is becoming more and more important nowadays. This means that besides trying to catch-up with the latest OpenGL revisions, opensource gpu drivers also need to add support for Compute programs through OpenCL and/or CUDA.

This talk will look at the various bits and pieces involved in adding Compute support to Nouveau. The main focus of the talk will be the translation from OpenCL to Nouveau GPU machine-code. This translation typically includes multiple passes translating from one intermediate representation (IR) to another. Which new passes can we introduce to get Compute support? What are the advantages / disadvantages of chosing a specific translation / compile pass to get Compute support?

After this global overview this talk will look in detail on two different approaches that implement different backends. The first work reuses the fact that Nouveau can consume TGSI, by implementing a LLVM IR to TGSI pass, meaning that it is driver independent. On the other hand, the second work teaches Nouveau how to consume SPIR-V, which is the intermediate representation being used by Vulkan for all shaders and compute programs. This makes it possible for Nouveau to consume directly a SPIR-V binary (if the shader / compute program has been pre-compiled), rather than going back to LLVM IR before being converted to TGSI and fed into Nouveau.

Other people: Pierre Moreau

This presentation will discuss the plans to move Xorg to use libinput too through an input driver called xf86-input-libinput, as well as the status of this move. xf86-input-libinput is scheduled to be the default Xorg input driver for Fedora 22.

Currently xorg uses a 1 driver per input device model, this makes it impossible to do things like middle button scrolling on the trackpoint on laptops where the trackpoint buttons are softbuttons on the touchpad. Besides this the xf86-input-synaptics driver was never really designed for multi-touch touchpads and this causes various issues.

For Wayland we've been working on a new improved input stack, which is to be shared by all compositors and lives inside libinput. This talk is about switching the xf86-input-* part of the Xorg input stack over to libinput by using a wrapper called xf86-input-libinput.

This talk will focus on the process and status of making this switch, can we make the switch without loosing any features, what updates are needed to the control-panel-applets of various desktop-environments to work with this new stack ?

What is the current status of Allwinner support in upstream u-boot and the kernel, which SoCs are supported, and which features (sound, video, etc.) are supported ?

The linux-sunxi community has been slowly but steadily working on getting Allwinner SoCs like the A10 supported in upstream u-boot and the kernel.

This talk will present the current status of Allwinner support upstream. Which SoCs are supported and which ones are not (yet) supported ? Which blocks if the supported SoCs are supported, and which are not ? Why are some SoCs / blocks not supported, and what are the plans to get them supported ? This are some of the questions this talk tries to answer.

Xorg (the X-server) is a big and complex beast. Currently it runs as root as it needs root privileges for various reasons. But with the latest systemd-logind all necessary infrastructure is in place to allow the server to run as a normal user and use systemd-logind to do input and graphics device management.

This talk looks at the work being done to leverage this new infrastructure to run Xorg without root rights.

This talk gives an overview on the state of the qemu usb subsystem. What happened last year? What are the plans for the future? Where do we stand in terms of USB 3.0 support? The intended audience are people who are interested in usb in virtual machines (vusb) both developers and users. The audience isexpected to be familiar with generic virtualization concepts, but no deep technical knowledge is required.

USB redirection is using USB devices attached to a real machine inside a virtual machine (guest). The focus of the talk will be the special case where the USB device being redirected is not attached to the physical machine hosting the virtual machine, but to some other machine. The talk will present and discuss the ongoing work to create an FOSS solution for this, getting this merged into upstream qemu, as well is integrating this into Spice and the Spice-client. The intended audience are people who are interested in usb in virtual machines (vusb) both developers and users. The audience is expected to be familiar with generic virtualization concepts, but no deep technical knowledge is required.

The purposes of this talk are: * To inform people about the new USB redirection support in qemu and Spice * Get feedback for future versions * Hopefully get people to use Spice and maybe even become contributors I gave the same talk last year, but quite a few things have changed: * USB-2.0 support * Integrated into Spice

In one sentence: "The Spice project aims to provide a complete open source solution for interaction with virtualized desktop services". This talk will give an overview of: * What Spice is * How Spice works (globally) * What makes it better then using ie vnc to access a remote virtual machine * Future features

The purposes of this talk are: * To inform people about the Spice protocol and the Spice project * Get feedback for future versions * Hopefully get people to use Spice and maybe even become contributors