Linux Plumbers Conf 2009

Video API Deathmatch: VDPAU vs. VAAPI

The perfcounters infrastructures is fast moving into being the unifying channel for hw and software performance analisys

This presentation gives an example how to integrate complex audio enhancement algorithms like acoustic echo canceler to Pulseaudio and how to implement really low audio latency applications as PA modules.

Aside from the internals of Linux Plumbing, some of the largest and most intricate examples of how Linux is plumbed is the variety of ways in which modern enterprise leverages Linux. One of the major issues affecting System Administrators and CIOs is how to appropriately manage legacy environments. Often systems reach the point where they have increasing TCO as vendors no longer supply security fixes and it becomes increasingly difficult to use current versions of our management tools. A number of the major linux distributions are now starting to twilight early enterprise distributions which still count for many thousands of Linux installs within businesses of every scale.

Video devices currently can only be open by one process at a time. This makes it impossible to do necessary things like software auto-focus or recording video while using ekiga.

A case study of the Portland Oregon Regional Transportation Archive Listing (PORTAL) database on PostgreSQL and Linux.

Describe the implementation, benefits and trade-offs of running without non-fatal System Management Interrupts in a Real-Time Enterprise level environment.

The Proportional IO controller allows to distribute disk time to tasks/cgroups in proportion to their assigned weights. It leverages existing cgroup infrastructure for task grouping and supports specification of weights hierarchically.

Discussion on a new real-time micro-benchmark suite and how it can help real-time overcome some challenges it faces, such as making apples-to-apples comparisons with other platforms, and how various programming languages compare in the real-time arena.

XACE can be used to make a nifty secure desktop on Xorg. But is XACE relevant with graphics interfaces moving into the kernel?

Discusses pro and cons of a unified mechanism to report platform errors.

I present a new algorithm for scalable concurrent hash tables, performance results for this algorithm (2-10x more scalable than Linux), and potential applications.

We take an in-depth look at two kinds of startup procedures: the initramfs system, and the process of loading and executing ELF binaries.

We aim to inform you about the archive formats you use every day. We will include an in-depth look at the tar, ar, cpio, gzip, bzip2, and deb formats, as well as the internals of the Git object store. Armed with this information, we will show you a practical application: removing the redundancy between files in version control and distributions of source and binaries.

Co-conspirators Jamey Sharp and Josh Triplett get up to a lot of miscellaneous hacking mischief together. Much of this hacking occurs while staring at the same screen, and tag-teaming the keyboard. Sometimes this happens with the two of them in different places. Learn how we manage this and why it's awesome.

The talk outlines various ways of establishing a networking communication between a network namespace (a container) and the outer world, compares their performance and features.

It is a solution to enable remote video acceleration on libVA for X-Window. It adds remote playback feature for libVA but does not require video-decoding capability on the client side.

A talk and demo on the current status of SELinux integration in Ubuntu.

Network devices have but a single name, and for systems with > 1 NIC, it's probably wrong. Let's fix that.

Explain the current state of kgdb, kdb. Discuss the future architectural plans for combining kgdb, kdb and kernel mode settings and proposed advanced single stepping APIs.

open discussion on locking issue on clustering file systems, especially associated with fs/dlm code

The network layer can accomplish some amazing feats with the rarely used multicast features in the Linux kernel. Not only is it possible to send a single message to multiple recipients but the network is also dynamically reconfiguring itself to provide for your communication needs using IGMP. Some degree of independence from the IP addresses of the machines in the network can be accomplished. The network state becomes important and one does not communicate with hosts but with communication channels that every host can tie into. The network develops an intelligence in itself. It can discover new services, fall back dynamically and trigger actions on multiple hosts in a single message. All of these feats occur with a much higher speed than TCP.

Running KVM in KVM, Xen in KVM, Hyper-V in KVM, ...

A description of a KVM port to PowerPC 970 cores, also known as PowerPC64

The Linux kernel Crypto API has come a long way since it was first added as a part of the IPsec stack.

Presenting Upstart 1.0

Overview of the current state of compositor interfaces and their interactions with OpenGL double-buffering.

Introduction of a new initramfs generation tool, which creates a general purpose udev event based initramfs.

It is compelling to integrate complex embedded systems with Linux USB userspace code. However, the complexity of userspace interface adds significant risk to the development cycle. Integrating a USB2.0 Isochronous mode ARM embedded system with Linux host software presents an end to end design challenge requiring not just a view of Linux from the 'inside' but also from the 'outside'. Some of the challenges Linux added to an already complex embedded system development will be presented, with some ideas to reduce the barries to entry and learning curve for embedded systems and host communication.

The Linux audio environment is a mess. Everybody who develops Linux audio software knows this; anybody who uses anything more than basic desktop playback applications (and maybe even just that) knows this too. How did things get to be this way? Can it be solved? What is required to solve it? This talk will be less of a presentation of novel technical material and will instead focus more on the differences (and similarities) between Linux, OS X and Windows and how this has affected the audio environment on each platform.

The moral of the story is, we still load too much stuff, and it takes too long to do it. Users (and developers) want time to go down, while things continue to get more complicated. How can we sort that out?

Routing performance is often used as a gauge of performance when evaluating a system. I will discuss some of the problems and solutions we found when trying to route line rate bi-directional traffic using off the shelf hardware.

The wireless extensions API is universally hated by wireless developers on both sides of the user/kernel divide. Fortunately the cfg80211 API is maturing as a modern alternative. What is needed is for drivers (both new and old) to adopt cfg80211 as their configuration mechanism so that the long wireless extensions nightmare can finally end.

This presentation addresses ways that developers can overcome development challenges, including using a Linux distribution that is tailored closely to their chosen hardware to help maintain the quality, control and consistency required to deliver commercial-ready devices to market faster.

USB 3.0 promises a faster, more power efficient common device bus. Is Linux ready for it?

The Oregon State Wireless Active Learning Device (OSWALD) is an open, fully featured Ultra-Mobile Personal Computer (UMPC) platform designed by and for students. Come learn about the design, the software hacks, and issues in getting Linux to run smoothly on a custom made hand-held platform.

A talk about Gallium and the Xorg state tracker and where they fit in and where they are going.

This talk will detail issues with Linux and large NUMA factor/COMA architectures.

SELinux policy is currently installed from a single or multiple package(s) as an application which breaks the linkage between policy and the software they are constraining. We will talk about a way to treat policy specially without adding downfalls such as a large increase in packages.

Over the last few years the graphics stack have been split up and refactored into shared libraries, kernel drivers and other components. The X server provides a lot of legacy functionality that isn't used by the modern, composited Linux desktop. Wayland is a new display server that builds on top of all those components to provide a minimal foundation for a composited destkop. X can run under Wayland with very little overhead for legacy applications.

This talk will discuss the requirements and design of a new SELinux policy infrastructure. It is hoped that this talk will lead to the SELinux community validation of the requirements, acceptance of the new architecture, and a plan to replace the old infrastructure.

Is it time to update or remove schedstats?

Shatter is a proposed upgrade to EXA that eliminates Virtual screen sizes and scanout/rendering limits.

Increasing networking performance by using threaded interrupts for the network devices.

The subject of this talk is a software based solution for scaling performance of the networking receive path. This is particularly useful with a single queue (legacy) NIC on a multi-core computer.

What happened in the last year in Linux audio?

State of threaded interrupt handlers in mainline. What is there and what needs to be dome.

Overview over the state of preempt-rt, the mainline merge status and key changes required.

Session to discuss and debate shared memory IPC in AMP systems. We will discuss the services desired, the implementation options, working with non-Linux OS instances, and whether or not it is even feasible to implement a common set of shared memory IPC tools

Session to discuss the work required to make the Flattened Device Tree method of describing hardware available to all architectures.

What can the Userspace RCU LGPL library do to help your application scale well on large multiprocessor systems ?

Discussing present and future git bisect features and how they can help application and kernel users and developers.

This talk will discuss problems faced by designers of multicore systems with multiple kernels, with an emphasis on solutions across developer communities.

SELinux is often disabled immediately or at the first sign of trouble. How can we make SELinux something users actually want to leave on?

This presentation aims at sharing lessons learned with the Linux audio stack (gstreamer, PulseAudio, ALSA) on recent consumer and mobile Intel devices.

A walk through the pieces of a modern Linux desktop, which manage the hardware that is coming and going and failing.

A discussion on how to manage KVM guests using CIM.

Requirements and challenges in implementation of Checkpoint/ Restart in Linux mainline.

Discussing limitations of current userspace storage APIs for use in QEMU/KVM.

Data de-duplication is a effective way to reduce large storage needs by eliminating redundant data, a hot demanded feature for virtualization OS image sharing and efficient data storage backups. It's really valuable to add data de-duplication support to Linux filesystem, however the feature is quite challenging too. How to get it right? What's the performance impact? Block level or file level? On the fly data de-duplication in filesystem or background userspace de-duplication?

The networking world in Linux has undergone some significant changes in the past two years. With the expansion of multiqueue networking, coupled with the growing abundance of multi-core computers with 10 Gigabit ethernet, the concept of efficiently converging different network flows becomes a real possibility.

This talk will focus on some of the challenges in migrating data from old, potentially failing hardware to new hardware: dealing quickly with IO errors, how to optimize the list of files to move and suggestions about how to handle failures during migration.

Linux 2.6.30 includes the Integrity Measurement Architecture (IMA) system, which measures (hashes) files before they are accessed, and which can use a TPM for hardware signed attestation for centralized management of client integrity.

The Out of Memory (OOM) killer has consistently been a hotly debated topic in Linux. Why does it kill? What is being done to improve it? What do users really need it to do?

With stabilisation for a 2.0 release almost complete, the FFADO project (www.ffado.org) is looking to solve the timing difficulties it has faced using a new kernel-based helper module.

Last years libv4l2 presentation ended with a slide future with bullet points of features to be added, so where we stand today? And what will happen in the coming year?

The initial impetus for cairo-drm came from a desire to experiment with GEM on my i915, and a need to provide an acceleration architecture for the wayland system compositor. (wayland is a minimalistic compositor and input multiplexer that requires client-side rendering.) It proved very easy to incorporate a drm/i915 backend into cairo; first by adapting the existing EXA driver, then to extend it to provide full acceleration for all patterns and to accelerate some of cairo's higher level operations. (The i915 hardware does not seem amenable for offloading tessellation, which has limited just how much we can accelerate in the backend. On more capable hardware, such as the fully programmable i965, we should be able to achieve much more.)

Our sample application is a commercial database server. It provides database services over the network using TCP connections. The security goal we're using Smack to address is the isolation of the database files from the users on the server. We'll cover two different ways to provide access to the database, allowing either remote or local users access to the services of the database while protecting the database itself.

Implementing power management is the job of the kernel - however, it's userspace's job to let it know what functionality is needed. What interfaces do we need to expose, and how should userspace be making use of them?

We have introduced the concept of the general purpose sandbox using SELinux.

e-Paper Development, Progress and Problems

This talk addresses relationship between embedded developers and the Linux community

This talk is about decoupling package management systems from build environments, making the decisions "what to build" and "how to build" orthogonal.

Per backing device dirty data writeback replaces pdflush driven writeback in an attempt to speed up this operation.

oFono telephony stack

The objective of this talk is to define common guidelines for security testing on Linux, sharing our concern about software security and the importance of software security testing as part of the development process.

Kernel Development and Testing with Autotest

This talk gives an overview of the new core V4L2 framework that was created in the past year, and the future developments planned for this framework.

Fine grained locking to improve scalability of the Linux VFS