http://summit.ubuntu.com/lpc-2012/ < Thursday >

08:30 - 09:25 PDT
Not Attending Breakfast
Nautilus 1
Not Attending Breakfast
Nautilus 2
Not Attending Tracing Summit - Early Morning ( Tracing )
**THIS MEETING STARTS AT 8:30** This is the first session of the tracing summit Agenda: 1) Can mainstream tracing meet embedded needs?, Frank Rowand, Sony 2) What We Want in Our Toolkit: Thoughts From a Mission Critical Low Latency Environment, Vinod Kutty, CME Group 3) Troubleshooting complex problems with built-in diagnostic, Dominique Toupin, Ericsson 4) The Linux Perf Tools: Overview and Current Developments, Arnaldo Carvahlo de Melo, Red Hat

Participants:
attending mathieu-desnoyers (Mathieu Desnoyers)

Tracks:
  • Tracing
Nautilus 3
Not Attending Breakfast
Nautilus 4
Not Attending Breakfast
Nautilus 5
09:30 - 10:15 PDT
Not Attending Tracing Summit - Early Morning ( Tracing )
**THIS MEETING STARTS AT 8:30** This is the first session of the tracing summit Agenda: 1) Can mainstream tracing meet embedded needs?, Frank Rowand, Sony 2) What We Want in Our Toolkit: Thoughts From a Mission Critical Low Latency Environment, Vinod Kutty, CME Group 3) Troubleshooting complex problems with built-in diagnostic, Dominique Toupin, Ericsson 4) The Linux Perf Tools: Overview and Current Developments, Arnaldo Carvahlo de Melo, Red Hat

Participants:
attending mathieu-desnoyers (Mathieu Desnoyers)

Tracks:
  • Tracing
Nautilus 3
Not Attending USB port power off Kernel/Userspace API ( Refereed Presentations )
The Intel Lynx Point chipset includes a new power savings mechanism that allows software to turn off the USB port power (VBUS) via an ACPI call. This mechanism is designed to save power on unused internal USB ports, but it could be used to save power on external ports or even simulate a physical disconnect for misbehaving USB devices. This presentation will provide background on the mechanism, before opening the floor to discussion on what sort of API the kernel should provide userspace to power down ports, and how userspace could use this API to save power, while not confusing the user with "dead" USB ports. Target audience: kernel developers working with USB device drivers, userspace developers for applications that touch USB devices (e.g. libusb, ModemManager, ConnMan), and desktop developers for Gnome, KDE, XFCE, etc. Sarah Sharp (<email address hidden>) is the xHCI driver maintainer. She has been adding USB 3.0 support to the Linux kernel for the past five years. Sarah works in Intel's Open Source Technology Center, along with Tianyu Lan. Tianyu (<email address hidden>) is a Linux kernel ACPI developer, and is the author of the patchset for the port power off mechanism.

Participants:
attending ezannoni-7 (LPC Submitter)

Tracks:
  • Refereed Presentations
Nautilus 1
Not Attending Simplify volume setting at startup/shutdown ( Audio )
Currently, on a normal desktop session, volume is set four times on startup - initally by the kernel, then by alsactl, then by PulseAudio in the DM session, then by PulseAudio in the logged in session. When shutting down, both PulseAudio and alsactl saves volumes to restore them later. And then we also have suspend and hibernate to consider, and that cards can be plugged in at any time. First, isn't this quite complex for something as simple as setting volumes? Second, can we facilitate new features, such as 1) having a "set this volume as default, for all users, on startup" button in the volume control, or 2) "allow the DM user to introspect different users' volumes"? Topic Lead: David Henningsson (<email address hidden>) Working for Canonical with audio hardware enablement, fixing audio bugs and maintaining the audio stack, and also part of PulseAudio's current development team.

Participants:
attending broonie (Mark Brown)
attending tiwai (Takashi Iwai)

Tracks:
  • Audio
Nautilus 2
Not Attending System Storage Manager; A single tool to manage your storage ( File and Storage Systems )
In more sophisticated enterprise storage environments, management with Device Mapper (dm), Logical Volume Manager (LVM), or Multiple Devices (md) is becoming increasingly more difficult. With file systems added to the mix, the number of tools needed to configure and manage storage has grown so large that it is simply not user friendly. With so many options for a system administrator to consider, the opportunity for errors and problems is large. The btrfs administration tools have shown us that storage management can be simplified, and we are working to bring that ease of use to Linux storage in general. I would like to introduce the new easy to use command line interface to manage your storage using various technologies like lvm, btrfs, crypt and more. System Storage Manager is currently under development with lots of features already available and more to come. I will discuss those features and problems we are facing when getting the project ready. I will also describe the scope of this project as well as where we see it in the future and of course gather useful feedback from the audience. Topic Lead: Lukas Czerner (<email address hidden>) Lukas is one of the core ext4 developers employed by Red Hat, Inc located in Czech Republic. He has been involved in performance evaluation of Linux discard support and was examining of alternative approaches which led to establishing the interface for filesystem batched discard support aka FITRIM as well as implementation for Ext4/3 filesystems. He is actively working on improvements of ext4 file system, its user space utilities. He is currently is working on simplification of constructing and administrating heterogeneous storage using various technologies like dm, md file systems and brtfs which resulted in System Storage Manager tool.

Participants:
attending ricwheeler (Ric Wheeler)

Tracks:
  • File and Storage Systems
Nautilus 4
Not Attending UEFI Basics Tutorial ( Core OS )
Covers: - UEFI basics on what appears in UEFI 2.3.1 spec'd systems with UEFI operating systems - requirements of both the OS and firmware for UEFI 2.3.1 secure boot - PXE network boot services(IPv6 andIPV4). Abstract will be available soon Instructor: Harry Hsiung (Intel)

Participants:
attending srwarren (Stephen Warren)

Tracks:
  • Core OS
Nautilus 5
10:25 - 11:10 PDT
Not Attending UEFI Basics Tutorial ( Core OS )
Covers: - UEFI basics on what appears in UEFI 2.3.1 spec'd systems with UEFI operating systems - requirements of both the OS and firmware for UEFI 2.3.1 secure boot - PXE network boot services(IPv6 andIPV4). Abstract will be available soon Instructor: Harry Hsiung (Intel)

Participants:
attending srwarren (Stephen Warren)

Tracks:
  • Core OS
Nautilus 5
Not Attending Policies for transcending Linux memory efficiency with zcache ( Refereed Presentations )
Zcache is the first Linux memory management capability that uses compression to dramatically increase RAM efficiency for both the page cache and for swap pages. The mechanism has been in staging since 2.6.39, but has been complemented only by a very primitive policy. The key to promoting zcache to a full kernel citizen and, indeed, for turning zache into an enterprise-ready default feature is a bulletproof "first do no harm" policy that must improve performance on many workloads while avoiding performance degradation on others. We will present how zcache works, discuss core kernel memory management policies and how they might be extended to "harden" zcache, and solicit discussion on further improvements for zcache, as well as its Transcendent Memory brethren: RAMster, which builds on zcache and kernel sockets to load-balance RAM across a cluster, and Xen tmem and KVM tmem, which provide similar capability in a virtualized environment. Linux plumbing connections: Page cache pages in zcache are mapped pages obtained via hooks in the VFS/filesystems (c.f. cleancache); swap pages are obtained via hooks in the swap subsystem (c.f. frontswap). Increasing RAM efficiency ultimately reduces the total RAM needed in a system, thus reducing capital and power costs across the data center.

Tracks:
  • Refereed Presentations
Nautilus 1
Not Attending Multipath TCP && TCP Loss Probe && Client Congestion Manager ( Networking )
Networking Topics: 1. Linux Kernel Implementation of Multipath TCP 2. TCP Loss Probe (TLP): fast recovery for tail losses 3. Client-based Congestion Manager for TCP === Linux Kernel Implementation of Multipath TCP === MultiPath TCP (short MPTCP) is an extension to TCP that allows a single TCP-connection to be split among multiple interfaces, while presenting a standard TCP-socket API to the applications. Splitting a data-stream among different interfaces has multiple benefits. Data-center hosts may increase their bandwidth; smartphones with WiFi/3G may seamlessly handover traffic from 3G to WiFi,... MultiPath TCP works with unmodified applications over today's Internet with all its middleboxes and firewalls. A recent Google Techtalk about MultiPath TCP is available at [1] In this talk I will first present the basics of MultiPath TCP and how it works and show some of the performance results we obtained with our Linux Kernel implementation (freely available at [2]). Second, I will go into the details of our implementation in the Linux Kernel, and our plans to try submitting the MPTCP-patches to the upstream Linux Kernel. [1] http://www.youtube.com/watch?v=02nBaaIoFWU [2] http://mptcp.info.ucl.ac.be Topic Lead: Christoph Paasch <email address hidden> === TCP Loss Probe (TLP): fast recovery for tail losses === Fast recovery (FR) and retransmission timeouts (RTOs) are two mechanisms in TCP for detecting and recovering from packet losses. Fast recovery detects and repairs losses quicker than RTOs, however, it is only triggered when connections have a sufficiently large number of packets in transit. Short flows, such as the vast majority of Web transfers, are more likely to detect losses via RTOs which are expensive in terms of latency. While a single packet loss in a 1000 packet flow can be repaired within a round-trip time (RTT) by FR, the same loss in a one packet flow takes many RTTs to even detect. The problem is not just limited to short flows, but more generally losses near the end of transfers, aka tail losses, can only be recovered via RTOs. In this talk, I will describe TCP Loss Probe (TLP) - a mechanism that allows flows to detect and recover from tail losses much faster than an RTO, thereby speeding up short transfers. TLP also unifies the loss recovery regardless of the "position" of a loss, e.g., a packet loss in the middle of a packet train as well as at the tail end will now trigger the same fast recovery mechanisms. I will also describe experimental results with TLP and its impact on Web transfer latency on live traffic. Topic Lead: Nandita Dukkipati <email address hidden> Nandita is a software engineer at Google working on making Networking faster for Web traffic and Datacenter applications. She is an active participant at the IETF and in networking research. Prior to Google she obtained a PhD in Electrical Engineering, Stanford University. === Client-based Congestion Manager for TCP === Today, one of the most effective ways to improve the performance of chatty applications is to keep TCP connection open as long as possible to save the overhead of SYN exchange and slow start on later requests. However, due to Web domain sharing, NAT boxes often run out of ports or other resources and resort to dropping connections in ways that make later connections even slower to start. A better solution would be to enable TCP to start a new connection as quickly as restarting an idle connection. The approach is to have a congestion manager (CM) on the client that constantly learns about the network and adds some signaling information to requests from the client, indicating how the server can reply most quickly, for example by providing TCP metrics similar to today's destination cache. Such a CM could even indicate to the server what type of congestion control to use, such as the relentless congestion control algorithm such that opening more connections does not gain advantage on aggregate throughput. It also allows receiver-based congestion control which opens new possibilities to control congestion. The Linux TCP metrics have similar concept but there is a lot of room for improvement. Topic Lead: Yuchung Cheng <email address hidden> Yuchung Cheng is a software engineer at Google working on the Make-The-Web-Faster project. He works on the TCP protocol and the Linux TCP stack focusing on latency. He has contributed Fast Open, Proportional Rate Reduction, Early Retransmit implementation in Linux kernel and wrote a few papers and IETF drafts of them. He has also contributed to rate-limiting Youtube streaming and the cwnd-persist feature of the SPDY protocol.

Participants:
attending therbert (Tom Herbert)

Tracks:
  • Networking
Nautilus 2
Not Attending Tracing Summit - Late Morning ( Tracing )
This is the second session of the tracing summit Agenda: 1) Tracing the Guest OS from Host via Shared Memory, Masami Hiramatsu, Hitachi 2) Shrinking core dump on the fly, Thomas Gleixner, Linutronix 3) Tracing Well With Others: Integration of GDB Tracepoints Into Trace Tools, Stan Shebs, Mentor Graphics 4) Ftrace and Multiple buffers, Steven Rostedt, Red Hat

Participants:
attending mathieu-desnoyers (Mathieu Desnoyers)

Tracks:
  • Tracing
Nautilus 3
Not Attending External Storage Array Management API ( File and Storage Systems )
The ability to manage external storage arrays and exploit the advanced features they provide in a programmatic way is an important capability for Linux. However, achieving storage array API nirvana isn't an easy thing to do with numerous obstacles in its path (e.g. licensing, terminology, features). In this session I would like to give a brief introduction to the libStorageMgmt project I am working on, discuss the difficulties it has encountered, and have an open discussion on how it can best work with other storage management components. Topic Lead: Tony Asleson Tony has a long and varied history of working in the storage industry and Linux. When he isn't sitting in front of his computer screen or spending time with his familiy, he can be found rock climbing and touring the back roads of MN on two wheels. He is currently a member of the Red Hat kernel storage team and lives in Rochester, MN.

Participants:
attending ricwheeler (Ric Wheeler)

Tracks:
  • File and Storage Systems
Nautilus 4
11:20 - 12:05 PDT
Not Attending UEFI Basics Tutorial ( Core OS )
Covers: - UEFI basics on what appears in UEFI 2.3.1 spec'd systems with UEFI operating systems - requirements of both the OS and firmware for UEFI 2.3.1 secure boot - PXE network boot services(IPv6 andIPV4). Abstract will be available soon Instructor: Harry Hsiung (Intel)

Participants:
attending srwarren (Stephen Warren)

Tracks:
  • Core OS
Nautilus 5
Not Attending Multipath TCP && TCP Loss Probe && Client Congestion Manager ( Networking )
Networking Topics: 1. Linux Kernel Implementation of Multipath TCP 2. TCP Loss Probe (TLP): fast recovery for tail losses 3. Client-based Congestion Manager for TCP === Linux Kernel Implementation of Multipath TCP === MultiPath TCP (short MPTCP) is an extension to TCP that allows a single TCP-connection to be split among multiple interfaces, while presenting a standard TCP-socket API to the applications. Splitting a data-stream among different interfaces has multiple benefits. Data-center hosts may increase their bandwidth; smartphones with WiFi/3G may seamlessly handover traffic from 3G to WiFi,... MultiPath TCP works with unmodified applications over today's Internet with all its middleboxes and firewalls. A recent Google Techtalk about MultiPath TCP is available at [1] In this talk I will first present the basics of MultiPath TCP and how it works and show some of the performance results we obtained with our Linux Kernel implementation (freely available at [2]). Second, I will go into the details of our implementation in the Linux Kernel, and our plans to try submitting the MPTCP-patches to the upstream Linux Kernel. [1] http://www.youtube.com/watch?v=02nBaaIoFWU [2] http://mptcp.info.ucl.ac.be Topic Lead: Christoph Paasch <email address hidden> === TCP Loss Probe (TLP): fast recovery for tail losses === Fast recovery (FR) and retransmission timeouts (RTOs) are two mechanisms in TCP for detecting and recovering from packet losses. Fast recovery detects and repairs losses quicker than RTOs, however, it is only triggered when connections have a sufficiently large number of packets in transit. Short flows, such as the vast majority of Web transfers, are more likely to detect losses via RTOs which are expensive in terms of latency. While a single packet loss in a 1000 packet flow can be repaired within a round-trip time (RTT) by FR, the same loss in a one packet flow takes many RTTs to even detect. The problem is not just limited to short flows, but more generally losses near the end of transfers, aka tail losses, can only be recovered via RTOs. In this talk, I will describe TCP Loss Probe (TLP) - a mechanism that allows flows to detect and recover from tail losses much faster than an RTO, thereby speeding up short transfers. TLP also unifies the loss recovery regardless of the "position" of a loss, e.g., a packet loss in the middle of a packet train as well as at the tail end will now trigger the same fast recovery mechanisms. I will also describe experimental results with TLP and its impact on Web transfer latency on live traffic. Topic Lead: Nandita Dukkipati <email address hidden> Nandita is a software engineer at Google working on making Networking faster for Web traffic and Datacenter applications. She is an active participant at the IETF and in networking research. Prior to Google she obtained a PhD in Electrical Engineering, Stanford University. === Client-based Congestion Manager for TCP === Today, one of the most effective ways to improve the performance of chatty applications is to keep TCP connection open as long as possible to save the overhead of SYN exchange and slow start on later requests. However, due to Web domain sharing, NAT boxes often run out of ports or other resources and resort to dropping connections in ways that make later connections even slower to start. A better solution would be to enable TCP to start a new connection as quickly as restarting an idle connection. The approach is to have a congestion manager (CM) on the client that constantly learns about the network and adds some signaling information to requests from the client, indicating how the server can reply most quickly, for example by providing TCP metrics similar to today's destination cache. Such a CM could even indicate to the server what type of congestion control to use, such as the relentless congestion control algorithm such that opening more connections does not gain advantage on aggregate throughput. It also allows receiver-based congestion control which opens new possibilities to control congestion. The Linux TCP metrics have similar concept but there is a lot of room for improvement. Topic Lead: Yuchung Cheng <email address hidden> Yuchung Cheng is a software engineer at Google working on the Make-The-Web-Faster project. He works on the TCP protocol and the Linux TCP stack focusing on latency. He has contributed Fast Open, Proportional Rate Reduction, Early Retransmit implementation in Linux kernel and wrote a few papers and IETF drafts of them. He has also contributed to rate-limiting Youtube streaming and the cwnd-persist feature of the SPDY protocol.

Participants:
attending therbert (Tom Herbert)

Tracks:
  • Networking
Nautilus 2
Not Attending Tracing Summit - Late Morning ( Tracing )
This is the second session of the tracing summit Agenda: 1) Tracing the Guest OS from Host via Shared Memory, Masami Hiramatsu, Hitachi 2) Shrinking core dump on the fly, Thomas Gleixner, Linutronix 3) Tracing Well With Others: Integration of GDB Tracepoints Into Trace Tools, Stan Shebs, Mentor Graphics 4) Ftrace and Multiple buffers, Steven Rostedt, Red Hat

Participants:
attending mathieu-desnoyers (Mathieu Desnoyers)

Tracks:
  • Tracing
Nautilus 3
Not Attending Letting go (A sensitive guide dealing with application caches under the pressures of low memory) ( Refereed Presentations )
Abstract: This talk will cover some of the recent work to allow applications to release memory to the kernel when the system is experiencing low memory conditions. The talk will focus on two basic approaches: applications pre-marking cache memory as freeable when not in use, allowing the kernel to reclaim it as needed, and APIs for kernel notification of lowmemory conditions that allow applications to then free up memory on their own. We'll talk about the upstream status of these different approaches, the pros and cons of the different methods, and talk about types of applications that could make use of these methods and how it affects both enterprise and embedded environments. Bio: John Stultz has worked with the IBM LTC for just over ten years. He has worked on x86 server enablement, Enterprise Realtime Linux, and now as a member of the Linaro.org effort. In the Linux community, he has worked mostly as a maintainer of the timekeeping subsystem, but has also worked on stability and scalability issues with the PREEMPT_RT patch, and most recently has been focused on upstreaming Android functionality into the mainline kernel.

Participants:
attending paulmck (Paul McKenney)

Tracks:
  • Refereed Presentations
Nautilus 1
Not Attending Anaconda, Snapper and Booting ( File and Storage Systems )
Topic Lead: Peter Jones Topic Lead: Matthias G. Eckermann Topic Lead: David Cantrell

Participants:
attending ricwheeler (Ric Wheeler)

Tracks:
  • File and Storage Systems
Nautilus 4
12:05 - 13:30 PDT
Lunch
13:30 - 14:15 PDT
Not Attending ACPI 5.0 in Linux ( Refereed Presentations )
Len Brown <email address hidden> ACPI (Advanced Configuration and Power Interface) abstracts platform hardware and firmware, helping to make operating systems portable. For over 15 years, this abstraction layer has allowed operating systems and hardware platforms to evolve independently. The 5th major revision of the ACPI spec was published December, 2011. This presentation will summarize the changes included in 5.0, and the state of Linux support for those changes. Bio: Len is a Principal Engineer at Intel’s Open Source Technology Center. He focuses on power management, and has maintained the Linux Kernel ACPI sub-system since 2003.

Participants:
attending ezannoni-7 (LPC Submitter)
attending lorenzo-pieralisi (Lorenzo Pieralisi)
attending srwarren (Stephen Warren)

Tracks:
  • Refereed Presentations
Nautilus 1
Not Attending The Core OS Wish List ( Core OS )
In the context of systemd we started collecting nice-to-have or that-should-just-work items, which we would wish the Linux kernel would provide us with. The emails to LKML have been called "A Plumber's Wish List for Linux". We give a quick update what problems have been solved, and what we still wish would work. Topic Lead: Lennart Poettering Topic Lead: Kay Sievers Lennart Poettering and Kay Sievers are the maintainers of systemd and udev and spend almost their entire work time on building infrastructure for the Linux Core OS.

Participants:
attending eblake (Eric Blake)
attending kaysievers (Kay Sievers)

Tracks:
  • Core OS
Nautilus 2
Not Attending Tracing Summit - Afternoon ( Tracing )
Third session of the tracing summit Agenda: 1) The Road ahead of Uprobes: Plans and features in pipeline, Srikar Dronamraju, IBM 2) Systemtap and new connections: dyninst, pcp, uprobes, David Smith and Josh Stone, Redhat 3) LTTng and Nexus Trace for Freescale QorIQ Devices, Ed Martinez, Freescale 4) Interoperability Between Tracing Tools with the Common Trace Format (CTF), Mathieu Desnoyers, EfficiOS

Participants:
attending mathieu-desnoyers (Mathieu Desnoyers)

Tracks:
  • Tracing
Nautilus 3
Not Attending Configuration and Management Open Discussion ( File and Storage Systems )

Participants:
attending ricwheeler (Ric Wheeler)

Tracks:
  • File and Storage Systems
Nautilus 4
Not Attending LLVM and Clang: Advancing Compiler Technology ( LLVM )
This talk introduces LLVM, giving a brief sense for its library based design. It then dives into Clang to describe the end-user benefits of LLVM compiler technology and the status of Clang Static Analyzer, LLDB, libc++ and the LLVM MC projects. Topic Leader: Chris Lattner Chris Lattner is best known as the primary author of the LLVM project and related projects, such as the clang compiler. He currently works at Apple Inc. as the Director of Low Level Tools and chief architect of the Compiler Group.

Tracks:
  • LLVM
Nautilus 5
14:25 - 15:10 PDT
Not Attending Tracing Summit - Afternoon ( Tracing )
Third session of the tracing summit Agenda: 1) The Road ahead of Uprobes: Plans and features in pipeline, Srikar Dronamraju, IBM 2) Systemtap and new connections: dyninst, pcp, uprobes, David Smith and Josh Stone, Redhat 3) LTTng and Nexus Trace for Freescale QorIQ Devices, Ed Martinez, Freescale 4) Interoperability Between Tracing Tools with the Common Trace Format (CTF), Mathieu Desnoyers, EfficiOS

Participants:
attending mathieu-desnoyers (Mathieu Desnoyers)

Tracks:
  • Tracing
Nautilus 3
Not Attending Moving Android and regular Linux closer together ( Refereed Presentations )
Abstract: Many people consider Linux and Android different operating systems - even though they're more alike than not. The communities are so split up that many Linux developers don't know anything about Android (other than the fact that it uses the Linux kernel) and many Android developers don't know anything about Linux (probably not even that Android is based on it). This is definitely a bad thing - because both sides have some good tools and features the other side could benefit from, if they just knew about them. Goal of the session is to bring the projects closer together, and try to start merging components where both sides can benefit. Should attend: People from "both sides" who aren't scared of learning about and working with the other side. Bio: Bernhard "Bero" Rosenkränzer is a developer of both "normal" Linux and Android, currently involved in the making of Linaro Android, Ark Linux and ROSA Linux.

Tracks:
  • Refereed Presentations
Nautilus 1
Not Attending Virtualization Memory Management ( Virtualization )
Virtualization Topics: 1) NUMA and Virtualization, the case of Xen 2) Automatic NUMA CPU scheduling and memory migration 3) One balloon for all - towards unified baloon driver === NUMA and Virtualization, the case of Xen === Having to deal with NUMA machines is becoming more and more common, and will likely continue to do so. Running typical virtualization workloads on such systems is particularly challenging, as Virtual Machines (VMs) are typically long lived processes with large memory footprints. This means one might incur really bad performance if the specific characteristics of the platform are not properly accounted for. Basically, it would be ideal to always run a VM on the CPUs of the node that host its memory, or at least as close as possible to that. Unfortunately, that is all but easy, and involves reconsidering the current approaches to scheduling and memory allocation. Extensive benchmarks have been performed, running memory intensive workloads inside Linux VMs hosted on NUMA hardware of different kinds and size. This has then driven the design and development of a suite of new VM placement, scheduling and memory allocation policies for the Xen hypervisor and its toolstack. The implementation of such changes has been benchmarked against the baseline performance and proved to be effective in yielding improvements, which will be illustrated during the talk. Although some of the work is hypervisor specific, it covers are issues that can be considered of interest for the whole Linux virtualization community. Whether and how to export NUMA topology related information to guests, just to give an example. We believe that the solutions we are working on, the ideas behing them and the performance evaluation we conducted are something the community would enjoy hearing and talking about. Topic Lead: Dario Faggioli Dario has interacted with the Linux kernel community in the domain of scheduling during his PhD on real-time systems. He now works for Citrix on the Xen Open Source project. He spent the last months on investigating and trying to improve the performance of virtualization workloads on NUMA systems. === Automatic NUMA CPU scheduling and memory migration === Topic Lead: Andrea Arcangeli === One balloon for all - towards unified baloon driver === During Google Summer of Code 2010 (Migration from memory ballooning to memory hotplug in Xen) it was discovered that in main line Linux Kernel exists 3 balloon driver implementations for 3 virtualization platforms (KVM, Xen, VMware). It quickly came out that they are almost identical but of course they have different controls and API/ABI. In view of e.g. memory hotplug driver which has generic base (not linked with specific hardware/software solution) this situation is not acceptable. The goal of this project is generic balloon driver which could be placed in MM subsystem and which could be linked with as little as possible platform specific code (placed e.g. in relevant arch directory). This solution could give unified ABI (which could ease administration) and unified API for developer (i.e. easier integration with e.g. tmem, memory hotplug, etc.). Additionally, balloon driver behavior would be almost identical on all platforms. Discussion should outline the goals and key solutions for such driver. Topic Lead: Daniel Kiper Daniel was a Google Summer of Code 2010 (memory hotplug/balloon driver) and Google Summer of Code 2011 (kexec/kdump) student. He is involved in *NIX administration/development since 1994. Currently his work and interests focuses on kexec/kdump implementation for Xen.

Participants:
attending amitshah (Amit Shah)
attending dkiper (Daniel Kiper)
attending lpc-virt-lead (LPC Virtualization Lead)
attending raistlin (Dario Faggioli)

Tracks:
  • Virtualization
Nautilus 2
Not Attending Local File Systems ( File and Storage Systems )
Topic Lead: Chris Mason Topic Lead: Ric Wheeler

Participants:
attending ricwheeler (Ric Wheeler)

Tracks:
  • File and Storage Systems
Nautilus 4
Not Attending LLVM Toolchain - Update and State of Building Linux with LLVM ( LLVM )
LLVM is a new toolchain that is becoming increasingly common in Linux environments and is already included in millions of Linux devices. Today, this is primarily as the JIT compiler for Renderscript in Android Ice-cream Sandwich, but its use is rapidly expanding into other areas of a Linux systems. This session will provide an update on the status of LLVM, some of the new uses that LLVM will be put to, and the state of building Linux with LLVM. Topic Lead: Mark Charlebois Mark Charlebois is Director of Open Source Software Strategy at QuIC. In his 13 years at Qualcomm he has lead diverse technical investigations working for various R&D divisions, and has worked on Unix-based systems since 1988 and embedded systems since 1990. Currently, Mark is a Linux evangelist, responsible for helping shape QuIC's Open Source SW strategy, and has been working on building Linux with LLVM. Mark has a bachelor's degree in Systems Design Engineering from the University of Waterloo in Canada and a masters degree in Engineering Science from Simon Fraser University in Canada. Topic Lead: Behan Webster Since graduating with a Bachelors degree in Computer Engineering from the University of Waterloo, Behan has spent the past two decades in diverse tech industries such as telecom, datacom, optical, and automotive. Throughout his career his work has most often involved kernel level programming, drivers, embedded software, board bring-ups, and build systems built on or for Linux (since 1996), and with UNIX before that. Currently Behan is the founder of Converse in Code and an embedded Linux engineer working on the LLVMLinux project as well as being a Trainer for the Linux Foundation. Behan is under the delusion he can fix most things with a “tiny little script”.

Tracks:
  • LLVM
Nautilus 5
15:20 - 16:05 PDT
Not Attending Tracing Summit - Afternoon ( Tracing )
Third session of the tracing summit Agenda: 1) The Road ahead of Uprobes: Plans and features in pipeline, Srikar Dronamraju, IBM 2) Systemtap and new connections: dyninst, pcp, uprobes, David Smith and Josh Stone, Redhat 3) LTTng and Nexus Trace for Freescale QorIQ Devices, Ed Martinez, Freescale 4) Interoperability Between Tracing Tools with the Common Trace Format (CTF), Mathieu Desnoyers, EfficiOS

Participants:
attending mathieu-desnoyers (Mathieu Desnoyers)

Tracks:
  • Tracing
Nautilus 3
Not Attending Big and Little Endian inside/out ( Refereed Presentations )
Endian bugs are a common source of problems, mostly on big endian architectures such as PowerPC, sparc or s390. Getting endian'ness right isn't always trivial and some programming tricks will work on one endianness and not another. I will first show endian issues related to programs trying to access in memory data using the wrong size types. Then I will talk about the the problem of data "on the wire" (networking, storage, file formats) along with tools such as sparse that can help getting it right, and will then dive into IO busses and peripheral such as PCI and graphics. I will attempt to shed some light on the latter by describing what happens under the hood, the representation of data on typical processor bus and how it interconnects to a typical IO bus, where does swapping happens and when it needs to be handled in software. BIO: Ben has been part of the powerpc Linux kernel team for many years and is currently the maintainer of arch/powerpc. He works for IBM ADL/Ozlabs.

Participants:
attending paulmck (Paul McKenney)

Tracks:
  • Refereed Presentations
Nautilus 1
Not Attending Atomic upgrades, booting, and package systems ( Core OS )
Current major consumer operating systems like Microsoft Windows and the Playstation 3 explicitly warn the user "don't turn off your computer" for upgrades. But the state of the art in many Linux-based "distributions" is to simply ignore this; if you happen to lose power or the kernel crashes, your system is quite likely toast, and you need a recovery CD. This isn't acceptable. This presentation will discuss my research into the area, working prototype code, and further work necessary in the core plumbing (particularly bootup and configuration management) to get fully atomic upgrades. Topic Lead: Colin Walters Colin has contributed to many different Free and Open Source software projects, including GNU Emacs, Debian, rpm, and dbus, but primarily works on GNOME for Red Hat, Inc.

Participants:
attending kaysievers (Kay Sievers)

Tracks:
  • Core OS
Nautilus 2
Not Attending Hinting vs Heuristics: Plumbing I/O Cache Hints Through the Linux Storage Stack ( File and Storage Systems )
The Linux-storage and wider storage community are actively investigating ways to express and leverage the varying performance characterics of storage devices. A storage device may do a better job servicing the I/O stream if it can discern details deeper than just the currently requested block address range. The T10 committee is in the process of specifying a hinting scheme to classify the in-flight data in a SCSI request. Similarly, a filesystem can do a better job of allocation if it is given some explicit hints from the application about how a file will be used. EXT4 is investigating an O_HOT/O_COLD hint that applications could use to express a coarse quality of service for a given file. At the same time, bcache has arrived as a stacking block device driver that uses heuristics to guide the decision of whether an I/O request should be cached in a high performance device or passed on to the next tier in the storage hierarchy. This presentation investigates an approach to plumbing hints through the filesystem to be consumed by a modified bcache block device. The tradeoffs between hinting and heuristics, as well as a proposed mechanism for specifying cache policy in userspace, are explored. The target audience is kernel filesystem/block developers and application developers that want to express caching or other policies to a storage configuration. Topic Lead: Dan Williams (<email address hidden>) Dan is a Linux-storage developer at Intel. He contributed support for offloading raid5/6 calculations, developed bios-raid support for md/mdadm, and currently maintains the libsas based isci driver. He has presented at the Ottawa Linux Symposium, the Linux Storage Summit, and authored an article for LWN.net.

Participants:
attending ricwheeler (Ric Wheeler)

Tracks:
  • File and Storage Systems
Nautilus 4
Not Attending LLVM/Clang x86 Kernel Build ( LLVM )
LLVM as new toolchain is used in an increasing number of projects: RenderScript, Gallium3D, Minix, FreeBSD and others. Also in the Linux ecosystem usage of LLVM and clang is on the rise. This talk provides an overview about the current state and efforts to compile the Linux Kernel with LLVM on x86, issues remaining, and discussion on what challenges and how they can be solved. Topic Lead: Jan-Simon Moeller Jan-Simon Möller is an engineer and consultant familiar with Embedded Linux, Build Systems, the Kernel and he contributes to open source projects. For the Linux Foundation he works as trainer for Embedded Linux and device drivers classes.

Tracks:
  • LLVM
Nautilus 5
16:30 - 17:15 PDT
Not Attending The Tunnel Monster ( Refereed Presentations )
With the growth of virtual networking has come a new set of problems all caused by tunneling. Tunnels used to be limited to VPN's, getting around pesky firewalls (see RFC3093). Now everbody's doing them and inventing new standards with more confusing names like NVGRE, VXVLAN and STT. This talk will explore what these are, what problems they cause and how well Linux does (or does not) support htem. BIO: Stephen has been involved with Linux drivers, networking, routing and bridging. In addition to working for Vyatta, in developing their open source routing platform he has contributed the "netem" network emulator, several network drivers and is the maintainer several utilities.

Tracks:
  • Refereed Presentations
Nautilus 1
Not Attending Time Alignment && PulseAudio on Android ( Audio )
=== Time alignment in the Linux Audio Stack === The Linux Audio stack provides very little support for precise timing, despite the availability of hardware audio wall clocks and the adoption of new protocols such as IEEE1588 and Ethernet AVB, which align networked devices several orders of magnitude more precisely than NTP. In this presentation, we show how providing user-space applications with access to the audio wall clock can improve audio rendering/capture for local and networked devices. In the first case, the resolution of the wall clock can help PulseAudio track with more precision the drift between system time and audio time. Likewise for networked devices, the differences in audio wall clocks can help a server adjust asynchronous sample-rate conversions without large and frequent variations of the sample-rate ratio. We will present some ideas on modifications of the audio stack and data structures and gather feedback from the open-source community. Topic Lead: Pierre Bossart === PulseAudio on Android === As part of our efforts to make 'standard' Linux components available in the Android world, we are working on porting PulseAudio to Android. In this session, we talk about challenges in the initial porting effort, the approach we are taking to make PulseAudio an out-of-the-box replacement for the native system, and what advantages we hope to be able to provide with this work. Topic Lead: Arun Raghavan <email address hidden> Arun Raghavan is a long-time open source supporter and mainly hacks on the PulseAudio audio server at Collabora. He contributes to the GStremaer multimedia framework, and secretly is a developer on the Gentoo Linux distribution as well.

Participants:
attending broonie (Mark Brown)
attending tiwai (Takashi Iwai)

Tracks:
  • Audio
Nautilus 2
Not Attending Tracing Summit - Late Afternoon ( Tracing )
Last meeting of the tracing summit Agenda: 1) Making linsched useful, Dhaval Giani, University of Toronto 2) LTTngTop: Human Readable Trace Viewer, Julien Desfossez, EfficiOS 3) Extensible trace analysis using the Tracing and Monitoring Framework, Alexandre Montplaisir, Ericsson 4) Open Discussion, Where do we go from here?

Participants:
attending mathieu-desnoyers (Mathieu Desnoyers)
attending paulmck (Paul McKenney)

Tracks:
  • Tracing
Nautilus 3
Not Attending NFS Advanced Projects ( File and Storage Systems )
Topic Lead: Jeff Layton Jeff is a long-time Linux enthusiast. After working as a Unix System Administrator for almost a decade, he joined the Red Hat kernel engineering team in 2007, focusing mainly on NFS and CIFS. He is also a member of the worldwide Samba team by virtue of his work on the Linux kernel CIFS filesystem. Topic Lead: Bruce Fields Bruce has worked on the Linux NFS code since 2002, first at the University of Michigan and then since 2010 at Red Hat. He maintains the kernel's NFS server, contributes to the IETF's NFSv4 working group, and generally enjoys solving problems wherever they turn up. Topic Lead: Chuck Lever

Participants:
attending ricwheeler (Ric Wheeler)

Tracks:
  • File and Storage Systems
Nautilus 4
Not Attending PM Constraints: OMAP ( Constraint Framework )
Power Management Constraints - OMAP 1) OMAP SoC Thermal Containment 2) Power Management in Linux with a coprocessor 3) New Model for System and Devices Latency === OMAP SoC Thermal Containment === The thermal challenge is to design an end-product with high performance while keeping the junction temperature of the IC components used on this product within their limitations and which does not present a thermal discomfort for the user. OMAP4/OMAP5 System on Chips, operating at highest Operating Performance Points (OPP), is a powerful mobile applications processor. However operating at higher voltage and higher frequency in a sustained manner may cause thermal limits to be exceeded, both for silicon and user comfort. We propose extensions on existing frameworks to model per device power constraints, for containment of thermal limitations across major heat sources of a end-product device, e.g. LCD, CPU, charging, etc. The framework shall facilitate the power and thermal management performed by governor and policies, depending on device context and use case knowledge. Topic Lead: Eduardo Valentin Eduardo currently acting as System Software Engineer at Texas Instruments, working on OMAP Linux kernel. He has been involved with embedded Linux for some years already, contributing on products of companies like Texas Instruments, Nokia, Motorola and Samsung. The main areas of interest are (not limited to): power management, real time, performance, scheduling, system and software otimization, and recently, thermal management. === Power Management in Linux with a coprocessor === Shutting down the main processor of an SoC in the idle and standby state results in significant power savings. However, doing so requires the responsibility of reviving the system to be passed onto an independent entity in the system. Texas Instruments has taken the lead in introducing a novel approach for system power management in its AM335x processor family involving a Cortex-M3 to assist the main processor. In the future there will other devices from Texas Instruments and possibly other silicon vendors which adopt this technique. Integrating a co-processor which is not running Linux with the PM framework comes with a new set of challenges. The co-processor needs to interact with the host processor for idle as well as standby power management. How do we communicate with a co-processor without significant overhead in the idle thread? In case the co-processor stops responding what should be recovery mechanism in the PM framework? What should be the mechanism for exporting the core details like the configured wakeup sources to the co-processor? This session will focus on the above mentioned challenges and other issues surrounding the usage of a co-processor for power management in Linux. Topic Lead: Vaibhav Bedia Vaibhav Bedia, Software Systems Engineer, Texas Instruments, works on Linux kernel development for Sitara ARM microprocessors. === New Model for System and Devices Latency === Due to the nature of the new SoC architectures the Power Management needs a new model for the various system latencies. The session discusses: - Concepts of system, devices, wake-up and resume latencies, - Recent changes in the devices framework for the latency, why and how to make it generic, - Links with the other PM QoS frameworks: thermal, cpuidle, - Recent changes in the ARM/OMAP platform code for the system latency, - Problems encountered while modelling and measuring the various latencies, - A proposed model and how to implement it, - Planned changes in the device framework, the platform code and the APIs. This session is oriented towards Linux power management developers. The goal is to agree on a framework implementation and the interfaces within the kernel and with the user space. Topic Lead: Jean Pihet <email address hidden> Jean is working with embedded Linux since many years now, for companies like Texas Instruments, MontaVista, Motorola and Philips. Recently NewOldBits.com has been founded to provide high quality consulting services. The area of work is mainly OMAP Power Management, tracing and profiling tools (perf, ftrace, oprofile...) for recent ARM cores.

Participants:
attending apm (Antti P Miettinen)
attending mark97229 (Mark Gross)
attending srwarren (Stephen Warren)

Tracks:
  • Constraint Framework
Nautilus 5
17:25 - 18:10 PDT
Not Attending Tracing Summit - Late Afternoon ( Tracing )
Last meeting of the tracing summit Agenda: 1) Making linsched useful, Dhaval Giani, University of Toronto 2) LTTngTop: Human Readable Trace Viewer, Julien Desfossez, EfficiOS 3) Extensible trace analysis using the Tracing and Monitoring Framework, Alexandre Montplaisir, Ericsson 4) Open Discussion, Where do we go from here?

Participants:
attending mathieu-desnoyers (Mathieu Desnoyers)
attending paulmck (Paul McKenney)

Tracks:
  • Tracing
Nautilus 3
Not Attending NFS Advanced Projects ( File and Storage Systems )
Topic Lead: Jeff Layton Jeff is a long-time Linux enthusiast. After working as a Unix System Administrator for almost a decade, he joined the Red Hat kernel engineering team in 2007, focusing mainly on NFS and CIFS. He is also a member of the worldwide Samba team by virtue of his work on the Linux kernel CIFS filesystem. Topic Lead: Bruce Fields Bruce has worked on the Linux NFS code since 2002, first at the University of Michigan and then since 2010 at Red Hat. He maintains the kernel's NFS server, contributes to the IETF's NFSv4 working group, and generally enjoys solving problems wherever they turn up. Topic Lead: Chuck Lever

Participants:
attending ricwheeler (Ric Wheeler)

Tracks:
  • File and Storage Systems
Nautilus 4
Not Attending Reducing Network Latency in Linux ( Refereed Presentations )
This presentation will cover some development-in-progress of a new in-kernel interface to allow applications to achieve lower network latency and jitter. These proposed patches create a new driver interface to allow an application to drive a poll all the way down to the device driver. A major benefit of this design is that applications do not have to change in order to speed up, and the Linux networking stack is not bypassed in any way. The design, implementation and results from an early prototype will be shown, and current efforts to refine, refactor, and upstream the design will be discussed. Affected areas include the core networking stack, and network drivers. Owners: Jesse Brandeburg BIO: Jesse Brandeburg is a senior Linux developer in the Intel LAN Access Division, producing the Intel Ethernet product lines. Jesse has been with Intel since 1994, and has worked on the Linux e100, e1000, e1000e, igb, ixgb, ixgbe drivers since 2002. Jesse splits his time between solving customer issues, performance tuning Intel's drivers, and working on bleeding edge development for the Linux networking stack.

Participants:
attending eblake (Eric Blake)
attending paulmck (Paul McKenney)

Tracks:
  • Refereed Presentations
Nautilus 1
Not Attending ALSA channel-mapping API ( Audio )
The functionality to query and/or set the PCM channel-mapping is a long-standing missing feature in ALSA. The session will cover the requirement by the actual hardware and discuss the pros and cons of proposed implementations. * REQUIRED AUDIENCE ALSA devs, PulseAudio devs, gstreamer devs Topic Lead: Takashi Iwai <email address hidden>: Working as a member of hardware-enablement team in SUSE Labs at SUSE Linux Products GmbH in Nuremberg, Germany, while playing a role as a gatekeepr of the Linux sound subsystem tree over years.

Participants:
attending broonie (Mark Brown)
attending diwic (David Henningsson)
attending srwarren (Stephen Warren)
attending tiwai (Takashi Iwai)

Tracks:
  • Audio
Nautilus 2
Not Attending ARM Virtualization ( Virtualization )
Virtualization Topics: 1. Xen on ARM Cortex A15 2. Porting KVM to the ARM Architecture === Xen on ARM Cortex A15 === During the last few months of 2011 the Xen Community started an effort to port Xen to ARMv7 with virtualization extensions, using the Cortex A15 processor as reference platform. The new Xen port is exploiting this set of hardware capabilities to run guest VMs in the most efficient way possible while keeping the ARM specific changes to the hypervisor and the Linux kernel to a minimum. Developing the new port we took the chance to remove legacy concepts like PV or HVM guests and only support a single kind of guests that is comparable to "PV on HVM" in the Xen X86 world. This talk will explain the reason behind this and other design choices that we made during the early development process and it will go through the main technical challenges that we had to solve in order to accomplish our goal. Notable examples are the way Linux guests issue hypercalls and receive event channels notifications from Xen. Is there anything that we could have done better? Is the architecture that we lied down in the Linux kernel generic enough to be re-used by other hypervisors? Topic Lead: Stefano Stabellini Stefano is a Senior Software Engineer at Citrix, working on the Open Source Xen Platform team. He has been working on Xen since 2007, focusing on several different projects, spanning from Qemu to the Linux kernel. He currently maintains libxenlight, Xen support in Qemu and PV on HVM in the Linux kernel. Before joining Citrix he was a researcher at the Institute for Human and Machine Cognition, working on mobile ad hoc networks. === Porting KVM to the ARM Architecture === With the introduction of the Virtualization Extensions to the ARM architecture (as implemented in the Cortex A7 and A15 processors), it is possible to implement a hardware assisted hypervisor. The KVM port to the ARM architecture, started by Christoffer Dall (University of Columbia) is an example of such a hypervisor. Our proposal is to describe the current state of the project, explain how the various virtualization extensions (hypervisor mode, second stage translation, virtual interrupt controller, timers) are used, how the KVM implementation on ARM differs from other architectures, and what our plans are for upstreaming the code. Topic Lead: Marc Zyngier <email address hidden> Marc has been toying with the Linux kernel since 1993, and has been involved over time with the RAID subsystem (MD), all kind of ancient architectures (by maintaining the EISA bus), messed with consumer electronics, and now focuses on the ARM architecture.

Participants:
attending amitshah (Amit Shah)
attending eblake (Eric Blake)
attending lpc-virt-lead (LPC Virtualization Lead)
attending marc-zyngier (Marc Zyngier)
attending srwarren (Stephen Warren)
attending stefano-stabellini (Stabe)

Tracks:
  • Virtualization
Nautilus 5
18:30 - 21:00 PDT [PLENARY]
Not Attending Evening Event @ Del Coronado Hotel
Offsite
< Thursday >

PLEASE NOTE The Linux Plumbers Conference 2012 schedule is still in a draft format and is subject to changes at any time.