| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
LWN.net Weekly Edition for June 19, 2008The Kernel Hacker's Bookshelf: Ultimate Physical Limits of Computation Moore's Law - we all know it (or at least think we do). To be annoyingly exact, Moore's Law is a prediction that the number of components per integrated circuit (for minimum cost per component) will double every 24 months (revised up from every 12 months in the original 1965 prediction). In slightly more useful form, Moore's Law is often used as a shorthand for the continuing exponential growth of computing technology in many areas - disk capacity, clock speed, random access memory. Every time we approach the limit of some key computer manufacturing technology, the same debate rages: Is this the end of Moore's Law? So far, the answer has always been no. But Moore's Law is inherently a statement about human ingenuity, market forces, and physics. Whenever exponential growth falters in one area - clock speed, or a particular mask technique - engineers find some new area or new technique to improve at an exponential pace. No individual technique experiences exponential growth for long, instead migration to new techniques occurs fast enough that the overall growth rate continues to be exponential. The discovery and improvement of manufacturing techniques is driven on one end by demand for computation and limited on the other end by physics. In between is a morass of politics, science, and plain old engineering. It's hard to understand the myriad forces driving demand and the many factors affect innovation including economies of scale, cultural attitudes towards new ideas, vast marketing campaigns, and the strange events that occur during the death throes of megacorporations. By comparison, understanding the limits of computation is easy, as long as you have a working knowledge of quantum physics, information theory, and the properties of black holes. The "Ultimate Laptop"In a paper published in Nature in 2000, Ultimate Physical Limits of Computation (free arXiv preprint [PDF] here), Dr. Seth Lloyd calculates (and explains) the limits of computing given our current knowledge of physics. Of course, we don't know everything about physics yet - far from it - but just as in other areas of engineering, we know enough to make some extremely interesting predictions about the future of computation. This paper wraps up existing work on the physical limits of computing and introduces several novel results, most notably the ultimate speed limit to computation. Most interesting in my mind is the calculation of a surprisingly specific upper bound on how many years a generalized Moore's Law can remain in effect (keep reading to find out exactly how long!). Dr. Lloyd begins by assuming that we have no idea what future computer manufacturing technology will look like. Many discussions of the future of Moore's Law center around physical limits on particular manufacturing techniques, such as the limit on feature size in optical masks imposed by the wavelength of light. Instead, he ignores manufacturing entirely and uses several key physical constants: the speed of light c, Planck's reduced constant h (normally written as h-bar, a symbol not available in standard HTML, so you'll have to just imagine the bar), the gravitational constant g, and Boltzmann's constant kB. These constants and our current limited understanding of general relativity and quantum physics are enough to derive many important limits on computing. Thus, these results don't depend on particular manufacturing techniques. The paper uses the device of the "Ultimate Laptop" to help make the calculations concrete. The ultimate laptop is one kilogram in mass and has a volume of one liter (coincidentally almost exactly the same specs as a 2008 Eee PC), and operates at the maximum physical limits of computing. Applying the limits to the ultimate laptop gives you a feel for the kind of computing power you can get in luggable format - disregarding battery life, of course. Energy limits speedSo, what are the limits? The paper begins with deriving the ultimate limit on the number of computations per second. This depends on the total energy, E, of the system, which can be calculated using Einstein's famous equation relating mass and energy, E = mc2. (Told you we'd need to know the speed of light.) Given the total energy of the system, we then need to know how quickly the system can change from one distinguishable state to another - i.e., flip bits. This turns out to be limited by the Heisenberg uncertainty principle. Lloyd has this to say about the Heisenberg uncertainty principle:
In particular, the correct interpretation of the time-energy
Heisenberg uncertainty principle ΔEΔt ≥ h
is not that it takes time Δt to measure energy to an accuracy
ΔE (a fallacy that was put to rest by Aharonov and Bohm) but
rather that that a quantum state with spread in energy ΔE takes
time at least Δt = πh/2ΔE to evolve to an
orthogonal (and hence distinguishable) state. More recently, Margolus
and Levitin extended this result to show that a quantum system with
average energy E takes time at least Δt = πh/2E
to evolve to an orthogonal state.
In other words, the Heisenberg uncertainty principle implies that a system will take a minimum amount of time to change in some observable way, and that the time is related to the total energy of the system. The result is that a system of energy E can perform 2E/πh logical operations per second (a logical operation is, for example, performing the AND operation on two bits of input - think of it as single bit operations, roughly). Since the ultimate laptop has a mass of 1 kilo, it has energy E = mc2 = 8.9874 x 1016 joules. The ultimate laptop can perform a maximum of 5.4258 x 1050 operations per second. How close are we to the 5 x 1050 operations per second today? Each of these operations is basically a single-bit operation, so we have to convert current measurements of performance to their single-bit operations per second equivalents. The most commonly available measure of operations per seconds is FLOPS (floating point operations per second) as measured by LINPACK (see the Wikipedia page on FLOPS). Estimating the exact number of actual physical single-bit operations involved in a single 32-bit floating point operation would require proprietary knowledge of the FPU implementation. The number of FLOPS as reported by LINPACK varies wildly depending on compiler optimization level as well. For this article, we'll make a wild estimate of 1000 single-bit operations per second (SBOPS) per FLOPS, and ask anyone with a better estimate to please post it in a comment. With our FLOPS to SBOPS conversion factor of 1000, the current LINPACK record holder, the Roadrunner supercomputer (near my home town, Albuquerque, New Mexico), reaches speeds of one petaflop, or 1000 x 1015 = 1 x 1018 SBOPS. But that's for an entire supercomputer - the ultimate laptop is only one kilo in mass and one liter in volume. Current laptop-friendly CPUs are around one gigaflop, or 1012 SBOPS, leaving us about 39 orders of magnitude to go before hitting the theoretical physical limit of computational speed. Finally, existing quantum computers have already attained the ultimate limit on computational speed - on a very small number of bits and in a research setting, but attained it nonetheless. Entropy limits memoryWhat we really want to know about the ultimate laptop is how many legally purchased DVDs we can store on it. The amount of data a system can store is a function of the number of distinguishable physical states it can take on - each distinct configuration of memory requires a distinct physical state. According to Lloyd, we have "known for more than a century that the number of accessible states of a physical system, W, is related to its thermodynamic entropy by the formula: S = kB ln W" (S is the thermodynamic entropy of the system). This means we can calculate the number of bits the ultimate laptop can store if we know what its total entropy is. Calculating the exact entropy of a system turns out to be hard. From the paper:
To calculate exactly the maximum entropy for a kilogram of matter in a
liter volume would require complete knowledge of the dynamics of
elementary particles, quantum gravity, etc. We do not possess such
knowledge. However, the maximum entropy can readily be estimated by a
method reminiscent of that used to calculate thermodynamic quantities
in the early universe. The idea is simple: model the volume occupied
by the computer as a collection of modes of elementary particles with
total average energy E.
The following discussion is pretty heavy going; for example, it includes a note that baryon number may not be conserved in the case of black hole computing, something I'll have to take Lloyd's word on. But the end result is that the ultimate laptop, operating at maximum entropy, could store at least 2.13 x 1031 bits. Of course, maximum entropy means that all of the laptop's matter is converted to energy - basically, the equivalent of a thermonuclear explosion. As Lloyd notes, "Clearly, packaging issues alone make it unlikely that this limit can be obtained." Perhaps a follow-on paper can discuss the Ultimate Laptop Bag... How close are modern computers to this limit? A modern laptop in 2008 can store up to 250GB - about 2 x 1012 bits. We're about 19 orders of magnitude away from maximum storage capacity, or about 64 more doublings in capacity. Disk capacity as measured in bits per square inch has doubled about 30 times between 1956 and 2005, and at this historical rate, 64 more doublings will only take about 50 - 100 years. This isn't the overall limit on Moore's law as applied to computing, but it suggests the possibility of an end to Moore's law as applied to storage within some of our lifetimes. I guess we file system developers should think about second careers... Redundancy and error correctionExisting computers don't approach the physical limits of computing for many good reasons. As Lloyd wryly observes, "Most of the energy [of existing computers] is locked up in the mass of the particles of which the computer is constructed, leaving only an infinitesimal fraction for performing logic." Storage of a single bit in DRAM uses "billions and billions of degrees of freedom" - electrons, for example - instead of just one degree of freedom. Existing computers tend to conduct computation at temperatures at which matter remains in the form of atoms instead of plasma. Another fascinating practical limit on computation is the error rate of operations, which is bounded by the rate at which the computer can shed heat to the environment. As it turns out, logical operations don't inherently require the dissipation of energy, as von Neumann originally theorized. Reversible operations (such as NOT) which do not destroy information do not inherently require the dissipation of energy, only irreversible operations (such as AND). This makes some sense intuitively; the only way to destroy (erase) a bit is to turn that information into heat, otherwise the bit has just been moved somewhere else and the information it represents is still there. Reversible computation has been implemented and shown to have extremely low power dissipation. Of course, some energy will always be dissipated, whether or not the computation is reversible. However, the erasure of bits - in particular, errors - requires a minimum expenditure of energy. The rate at which the system can "reject errors to the environment" in the form of heat limits the rate of bit errors in the system; or, conversely, the rate of bit errors combined with the rate of heat transfer out of the system limits the rate of bit operations. Lloyd estimates the rate at which the system can reject error bits to the environment, relative to the surface area and assuming black-body radiation, as 7.195 x 1042 bits per meter2 per second. Computational limits of "smart dust"Right around the same time that I read the "Ultimate Limits" paper, I also read A Deepness in the Sky by Vernor Vinge, one of many science fiction books featuring some form of "smart dust." Smart dust is the concept of tiny computing elements scattered around the environment which operate as a sort of low-powered distributed computer. The smart dust in Vinge's book had enough storage for an entire systems manual, which initially struck me as a ludicrously large amount of storage for something the size of a grain of dust. So I sat down and calculated the limits of storage and computation for a computer one μm3 in size, under the constraint that its matter remain in the form of atoms (rather than plasma). Lloyd calculates that, under these conditions, the ultimate laptop (one kilogram in one liter) can store about 1025 bits and conduct 1040 single-bit operations per second. The ultimate laptop is one liter and there are 1015 μm3 in a liter. Dividing the total storage and operations per second by 1015 gives us 1010 bits and 1025 operations per second - about 1 gigabyte in data storage and so many FLOPS that the prefixes are meaningless. Basically, the computing potential of a piece of dust far exceeds the biggest supercomputer on the planet - sci-fi authors, go wild! Of course, none of these calculations take into account power delivery or I/O bandwidth, which may well turn out to be far more important limits on computation. Implications of the ultimate laptopCalculating the limits of the ultimate laptop has been a lot of fun, but what does it mean for computer science today? We know enough now to derive a theoretical upper bound for how long a generalized Moore's Law can remain in effect. Current laptops store 1012 bits and conduct 1012 single-bit operations per second. The ultimate laptop can store 1031 bits and conduct 1051 single-bit operations per second, a gap of a factor of 1019 and 1039 respectively. Lloyd estimates the rate of Moore's Law as 108 factor of improvement in areal bit density over the past 50 years. Assuming that both storage density and computational speed will improve by a factor of 108 per 50 years, the limit will be reached in about 125 years for storage and about 250 years for operations per second. One imagines the final 125 years being spent frantically developing better compression algorithms - or advanced theoretical physics research. Once Moore's Law comes to a halt, the only way to increase computing power will be to increase the mass and volume of the computer, which will also encounter fundamental limits. An unpublished paper entitled Universal Limits on Computation estimates that the entire computing capacity of the universe would be exhausted after only 600 years under Moore's Law. 250 years is a fascinating in-between length of time. It's too far away to be relevant to anyone alive today, but it's close enough that we can't entirely ignore it. Typical planning horizons for long-term human endeavors (like managing ecosystems) tend to max out around 300 years, so perhaps it's not unthinkable to begin planning for the end of Moore's Law. Me, I'm going to start work on the LZVH compression algorithm, tomorrow. One thing is clear: we live in the Golden Age of computing. Let's make the most of it. Valerie Henson is a Linux consultant specializing in file systems and owns a one kilo, one liter laptop.
Multi-system administration with Func Managing multiple computer systems can involve a lot of repetitive tasks: connecting to each, performing some update, status check, or configuration tweak, and then moving on to the next machine. These kinds of things can be scripted of course, but scripts of that nature typically need to be adjusted frequently as machines come and go or the tasks change. The Fedora Unified Network Controller (Func) is a tool that will help simplify system administration, but there is more to it than that—it is a framework for doing two-way secure communication, from the command line, scripts, or applications. Func is written in Python, providing an API for scripts written in that language, but it can also be used from the command line. Each client machine—minion in Func-speak—runs the funcd daemon which contacts the master server or overlord. From the overlord machine, commands can then be issued to individual minions or to subsets of them. Some of the power of Func can be seen in simple commands like:
func "*" call service restart httpd
which will restart the web server on all of the minions.
Similar kinds of tasks—but with more control—can be handled through the Python API. A somewhat contrived example from the Func website gives a sense of what can be done:
import func.overlord.client as fc
results = fc.Client("*").service.status("httpd")
for (host, returns) in results.iteritems():
if returns == 0:
fc.Client(host).reboot.reboot()
This example looks for minions that are running a web server and reboots
each that it finds.
Managing keys can be a hassle when using ssh as an administrative tool, so Func uses another tool, Certmaster, to assist with keys. Certmaster provides a set of utilities and a Python API for managing SSL certificates. Clients generate certificate signing requests (CSRs), which contain their public key, that are sent to the Certmaster on the overlord. Administrators can either sign them from the command line or enable auto-signing. The minion then retrieves the signed certificate so that the overlord and minion communicate over an encrypted channel after that. Func is not meant to replace ssh, instead it is intended to provide multi-system and scripting capabilities which are not the strengths of ssh. Like ssh, though, Func is meant to be easy to deploy—eventually ubiquitous, at least for Fedora—simple to use as well as easy to extend. It also has a pluggable architecture that allows Python modules to be integrated easily into Func, expanding the abilities of the minions. The documentation shows how to use the func-create-module command to generate template code which allows the administrator to ignore the Func requirements and concentrate on the task at hand. There is nothing particularly Fedora-specific about Func, that's just where it was born. There are some efforts underway to add it for other distributions. Most of the work would be in creating distribution-specific analogs for things like restarting services and querying hardware configurations. Red Hat has been releasing a steady stream of system administration tools over the last year or so. The Emerging Technology (ET) group has developed quite an ecosystem of tools to support installations with large numbers of servers that are frequently installed and upgraded. One might think they have a large infrastructure of such servers. One of those tools that is frequently discussed in conjunction with Func is Cobbler. It is meant to simplify the configuration of a server to handle network installation and booting for a large server farm. From the web page:
In short, Cobbler helps build and maintain network installation
infrastructure really easily. It's highly customizable to your particular
methods of operation through a wide variety of options, a powerful command
line, a Web interface, a pluggable extension mechanism, and (for
developers) its own Python API. Cobbler lets administrators forget how
software gets installed and delivered and lets them concentrate instead on
what they want to install where.
Cobbler and the other tools coming out of the ET group are not just targeted at physical machines, but also virtualized environments. By using Cobbler, the puppet configuration manager, and the oVirt virtual machine manager, thousands of systems of various kinds can be managed in a centralized fashion. As would be expected, all of the code is available as free software. These tools are quite interesting for system administrators, particularly those who use Fedora and have lots of systems to maintain. Even for small home networks, though, Func at least could come in handy. For overworked administrators—no matter the size of their domain—better tools are always welcome.
Deki helps Mozilla developers collaborate There was undoubtedly plenty of activity this week at the Mozilla Developer Center ahead of the release of Firefox 3. Thanks to a special tool created by the team at MindTouch and implemented into its latest product offering, Deki, Mozilla developers all across the globe were able view the site in their native tongue. The "polyglot" language feature is only one of several components that make up Deki, an open source collaboration tool for communities and the enterprise. The polyglot can distinguish between different languages across a single system so it's no longer necessary for IT professionals to allocate sections of a web site's infrastructure to overcome language barriers. Instead, multiple languages are consolidated into one system and a site's pages are then localized according to user settings. Deki functions similar to that of a traditional wiki, but with more features and practical applications. In fact, the company originally called the product "Deki Wiki" but realized it was too limiting and recently dropped "Wiki" from the name altogether. Developers can use Deki as a way to organize and aggregate project data, share documents and media, or even author and create collaborative applications from the ground up. Groups and organizations also use Deki as platform for managing a large knowledge base, coordinating team-based projects, or as a file repository. Deki is part application, part platform. It behaves much the same way as other content management frameworks like Drupal and Joomla!, but has the underpinnings of a wiki that give it collaborative features as well. Furthermore, everything under Deki's hood can be accessed via the API on which it was built, and can be extended in any programming language. At the heart of the platform is MindTouch Dream, which forms the application's architecture, and uses Deki as its interface. It's a .NET representational state transfer (REST) framework that runs on .NET 2.0 and Mono 1.2 — .NET runs on Microsoft Windows Servers 2003 and 2008, while Mono runs on Debian, Fedora, Ubuntu, openSUSE, and Apple OS X (see the web site for complete details). Data manipulation is done in XML using standard HTTP verbs, and data conversions to PHP, JSONP, etc. are done automatically behind the scenes. Licensed under the Gnu GPL and LGPL, together Deki and Dream can be completely customized and scaled to the needs of any size organization. Company co-founders Aaron Fulkerson and Steve Bjorg were approached last winter by Mozilla's Chief Evangelist Mike Shaver about implementing Deki in time for the upcoming re-launch of its Developer Center. "Mike had reviewed our API and architectural documentation and he was enthusiastic about MindTouch Deki," recalls Fulkerson. "Later on the phone, we discussed Mozilla's needs, pains, and how MindTouch Deki seemed to be the perfect solution. We also day-dreamed a little about what the Mozilla community might build on the MindTouch platform. By my recollection, we both were pretty excited about the opportunity." Given the Developer Center's wide geographical reach, barriers were to be expected as it struggled to cater to a group that collectively spoke dozens of different languages. In response, Bjorg and Fulkerson put together a design that allows for a multi-lingual Web site that scales as needed. As Mozilla's needs grow, additional languages can easily be added by translating a single file and submitting it for inclusion in the official Deki build. In fact, all current translations have come from the community, and more are on the way. Deki isn't just for large organizations. Development platform-as-a-service provider Bungee Connect uses it as a documentation repository at the moment, but according to the Director of Bungee Connect's Developer Community, Ted Haeger, the plan is to soon make it the community platform for its Developer Network. "Our developers are very interested in programmable Web technologies, and Deki will allow us to provide them the most feature-complete wiki API we have seen yet. We expect to see some interesting and exciting things built by combining Bungee Connect and MindTouch Deki," he says. The decision to choose Deki over other similar options "was driven overwhelmingly by the architecture of the product. Because Deki provides a complete RESTful API, it makes it an extremely attractive offering for us," notes Haeger. Indeed, he considers the API Deki's best feature. "MindTouch has done an outstanding job with it," Haeger says. "Additionally, they have written their PHP front-end to the Deki API, which means that the API is central to the product rather than an afterthought. However, we should note that Deki's default PHP user interface is extremely polished, too. That combined with other must-haves, such as a permissions system that is considerably more flexible than what other wikis provide, helped solidify our decision." Though there are varying levels of support options available, Haeger says Bungee Connect hasn't yet decided which to choose. They do plan, however, to lean on MindTouch for assistance as they migrate company documentation from MediaWiki to Deki. For organizations planning to take on the task themselves, Fulkerson points to the helpful guide on its site and the Mediawiki to Deki converter they have written: "As we always have done, we've released the source code to our public SVN repository. It's stable and has had generous test coverage, but this should be considered a beta release." As Deki continues to gain traction in the enterprise as an agile content management system, Fulkerson and Bjorg say they knew they were on to something when they caught wind of the first user-organized conference held in Belgium last fall. Notes Fulkerson, "This was a pretty clear indication people liked what we're doing."
Page editor: Jonathan Corbet Security The Application Security Desk Reference The Open Web Application Security Project (OWASP) has undertaken an ambitious project to create a reference manual—in the same vein as the Physician's Desk Reference—covering application security. The book, along with a companion wiki are meant to be the starting point for researchers, developers, and code reviewers when performing a number of security-related tasks. The book is currently in an alpha state, with OWASP looking for more reviewers and authors to get the book into a finished state by August. The Application Security Desk Reference (ASDR) will be a 900+ page book, extensively tagged—cross-referenced in the wiki—to provide a multi-dimensional view of security threats, attacks, vulnerabilities, and impacts. The book introduces a set of principles that will help guide developers in avoiding these problems along with controls (aka countermeasures) to evade or eliminate them. The authors provide a description of why they took this approach:
Application security information cannot be organized into a one-dimensional
taxonomy that is useful for all
purposes, although many have tried. For example, organizing application
security by vulnerability helps tool
vendors, but makes it very difficult for architects to select
controls. We've adopted the folksonomy tagging
approach to solving this problem. We simply tag our articles with a number
of different categories. You can use
these categories to help get different views into the complex,
interconnected set of topics that is application
security.
The PDF 0.9 version is available, and it is already quite useful, though there is still a fair amount of work to do. An important goal is to provide a foundation:
The ASDR is helpful as basic reference material when performing such
activities as threat modeling, security
architecture review, security testing, code review, and metrics. We intend
to encourage understanding and
consistency when discussing these basic foundational elements of
application security. Security only works if
people can make informed decisions about risk. The ASDR provides that basic
information to help ensure all
stakeholders are involved.
Technical books have a unfortunate tendency to rapidly go stale because the industry moves so quickly. Maintaining the wiki will help alleviate this problem by allowing for a dynamic reference that can be periodically produced in dead tree form as well. Much of this kind of information can be found in books and on the web, but collecting it up into one place is very valuable. Three sections of the current draft stand out as being closest to completion: Principles, Attacks, and Vulnerabilities. Principles contains 17 basic things to keep in mind as part of gaining a "security consciousness". It defines terms in clear language and provides reasons why the principle should be followed. An example:
Security through obscurity is a weak security control, and nearly always
fails when it is the only control. This is not
to say that keeping secrets is a bad idea, it simply means that the
security of key systems should not be reliant
upon keeping details hidden.
More than 50 attacks are listed, along with examples and concise descriptions. In addition, there are several hundred vulnerabilities listed, each with examples as well as information on which platforms or languages are affected. It clearly sets out to be a clearinghouse of application security information and looks like it is succeeding in that. For anyone with an interest in security, it is well worth a look. For those who are skilled in security techniques, assisting with the review and content creation might be in order.
Security news SSL Certificates Vulnerable to OpenSSL Flaw on Debian (Netcraft) Netcraft has discovered a "significant number" of bad SSL certificates due to the recent Debian OpenSSL flaw. Some Extended Validation (EV) certificates are among those they found that were generated with the vulnerable code. "The vulnerable certificates afford opportunities to create deceptive sites which use apparently valid SSL certificates, giving the user the impression that the site belongs to the certified organisation. In the case of EV certificates, browsers will also turn the address bar green, even though the certificate may be cloned."
New vulnerabilities cbrpager: execution of arbitrary code
freetype: multiple vulnerabilities
openoffice.org: arbitrary code execution
roundcubemail: cross-site scripting
typo3: several vulnerabilities
xorg-server: multiple vulnerabilities
Page editor: Jake Edge
Kernel development Release status Kernel release status The current 2.6 development kernel is 2.6.26-rc6, released by Linus on June 12. "I'd like to say that the diffs are shrinking and things are calming down, but I'd be lying. Another week, another -rc, and I another 350 commits." See the long-format changelog for all the details.As of this writing, some 140 commits have gone into the mainline git repository since the 2.6.26-rc6 release. They include a number of fixes and a new driver for FM3130 realtime clocks. The current -mm tree is 2.6.26-rc5-mm3. Says Andrew: "The aim here is to get all the stupid bugs out of the way so that some serious MM testing can be performed." Among other things, this release contains the latest version of the pageout scalability patches (see below). The current stable 2.6 kernel is 2.6.25.7, released on June 16. It contains a rather long list of important fixes.
Kernel development news Why some drivers are not merged early Arjan van de Ven's kernel oops report always makes for interesting reading; it is a quick summary of what is making the most kernels crash over the past week. It thus points to where some of the most urgent bugs are to be found. Sometimes, though, this report can raise larger issues as well. Consider the June 16 report, which notes that quite a few kernel crashes were the result of a not-quite-ready wireless update shipped by Fedora. Ingo Molnar was quick to jump on this report with a process-related complaint:
i suspect Fedora has done this to enable more hardware, and/or to
fix mainline wireless bugs? I wish we would do such new driver
merging in mainline instead, so that we had a single point of
testing and single point of effort.
Same for Nouveau: Fedora carries it and i dont understand why such a major piece of work is not done in mainline and not _helped by_ mainline. He then took the discussion further with this observation:
That's my main point: when we mess up and dont merge OSS driver
code that was out there in time - and we messed up big time with
wireless - we should admit the screwup and swallow the bitter pill.
This comment drew some unhappy responses from the networking developers, who feel that they have been unfairly targeted for criticism. Wireless drivers have been merged at the first real opportunity, they say, and trying to put them in earlier would have only made things worse. In fact, your editor will submit that mistakes were made with wireless drivers, but those mistakes have little to do with delaying their inclusion into the mainline. What went wrong with wireless is this:
The result of all this (and the occasional legal hassle as well) is that wireless networking on Linux lagged for years, and is only now reaching something close to a stable state. So it is not surprising that there has been a lot of code churn in this area, or that things occasionally break. But it is hard to see how trying to merge wireless drivers sooner would have helped the situation significantly. The non-merging of the Nouveau driver - the reverse-engineered driver for NVIDIA adapters - also has a simple explanation: the developers have not yet asked for this merge to happen. Nouveau is not considered to be at a point where it works yet, and, importantly, there are still user-space API issues which must be worked out. Breaking user-space code is severely frowned upon, so merging of code is nearly impossible if its user-space interfaces are still in flux. James Bottomley put forward another reason why a driver may stay out of the mainline even though the author would like to see it merged:
For the record, my own view is that when a new driver does appear
we have a limited time to get the author to make any necessary
changes, so I try to get it reviewed and most of the major issues
elucidated as soon as possible. However, since the only leverage I
have is inclusion, I tend to hold it out of tree until the problems
are sorted out.
In other words, their control over access to the mainline tree is the one club subsystem maintainers have at hand when they feel the need to push a developer to make changes to a driver. It may well be that simply merging drivers regardless of technical objections (something which a number of developers are pushing for) will reduce the incentive for developers to get their code into top shape - and it's not always clear that others will step in and do the work for them. On the other hand, the idea that in-tree code tends to be less buggy than out-of-tree code is relatively uncontroversial. So, for many drivers at least, a "merge first and fix it up later" policy may well lead to the best results in the shortest period of time. One thing that is clear is that this discussion will not be going away anytime soon; chances are good that this year's kernel summit (happening in September) will end up revisiting the issue.
Peter Zijlstra: From DOS to kernel hacking In a linux-kernel thread about fixing the Kernel Janitors project, Peter Zijlstra spoke up, with a bit of his perspective on attracting better kernel contributors. As he is a relatively recent addition to the kernel community, his path from Linux user to kernel hacker may serve as a template of sorts for others who are starting out now. We asked Peter to answer a few questions by email to help fill in some more of the details. LWN: How did you get started with Linux? What attracted you?
Peter: Around the time Win95 came around, IIRC [if I remember
correctly]. I used to do demo
coding on
DOS, which involved rebooting your machine every time you messed up, and
whereas DOS reboots quite quickly, doing the same on Win95 was anything
but quick.
A friend of mine introduced me to Unix/Linux at the time, and I started learning all about programming in a real environment. Basically all programming up to that point was in a freestanding environment where you had to poke the hardware to get anything done. So initially it was the charm of a proper multitasking OS (with memory protection) that got me to use it – not having to reboot your machine every time, and the luxury of being able to run a debugger. LWN: How quickly did you start poking around in the kernel? What did you first start to look at and why?
Peter: The kernel ... well that took a seriously long while. The
above introduction to Linux was around 95/96 IIRC. My first real kernel
patches were about 10 years later.
In those 10 years I learnt a lot about programming. I learnt about Unix system programming, I learnt about C++, multi-threading, database engines, and a whole range of interesting things. Somewhere along I got a real internet connection and started lurking on mailing lists, including LKML – I must have been reading that on and off for about 5 years by the time I really sat down and wrote some patches. During that time I might have sent in some trivial build fixes, and I remember finding a priority leak in one of the realtime patches. But I wasn't actively coding on the kernel – I just liked running real exotic stuff, you know Gentoo and building just about everything from CVS. So what got me started on the kernel ... I can't quite remember how it happened, but I ran into some of Rik's [van Riel] Advanced Page Replacement stuff. I had worked on that problem space earlier while doing database engines, and had recently run into it again at work. So I started reading those papers and some of the proposed kernel patches, and I started to itch. I dropped basically everything I was working on in my spare time (hacking WindowMaker, writing a C++ ASN.1-DER serialization class, writing a new LDAP server and I'm sure some other projects that are rotting away on a harddrive somewhere :-) and started hacking. Why ... I'm not sure – it sure got me back to where I started out – crashing machines (and boot times haven't improved over those past 10 years at all). I think because of the challenge – I knew I could write whatever it was I was coding and this page replacement stuff was a whole new challenge, and TBH [to be honest] the kernel code didn't look too hard at the time (phew how ignorant I was..) LWN: How well were your contributions received by kernel hackers? Did you make any missteps along the way?
Peter: Some better than others. I think its natural for every kernel
hacker to
grow a huge pile of discarded patches. Not everything will make it. But
don't get discouraged by that, you did get to learn something from doing
them.
Mis-steps, feh, still do ;-) Unlike most people seem to think, kernel hackers are human too. LWN: What suggestions do you have for folks that are looking at getting involved in kernel hacking today?
Peter: Just do it – seriously it's that easy. Oh and don't be
afraid
of
criticism, you'll get it anyway – in spades. Criticism is not personal,
it's about your patch, there are two things you can do:
OK it can get personal, but that is only if you repeatedly fail the above two points. LWN: There has been a lot of talk about the Kernel Janitors project recently, do you think that is a good way to get started with kernel development? What do you think should be done differently in that (or other) project(s) to attract more and better contributors?
Peter: I'm not sure. The Kernel Janitors thing doesn't really seem to
work out.
I think that might be due to two things:
So I guess what I'm saying is that you need to really want to do it. There is no other way to become a kernel hacker than by simply doing it. LWN: Do you work on Linux for your job, as a hobby, or both?
Peter: Both; initially it was spare time besides $JOB. But after
keeping this
up for about a year my wife nudged me to look for a kernel job, since I
obviously enjoyed hacking the kernel more than $JOB, and she'd get some
of that spare time back ;-)
So I applied for a kernel position at a few of the larger vendors, and Red Hat won the race. Already having had a year's worth of exposure to kernel code and LKML, certainly helped in getting this amazing opportunity. Have I already mentioned I absolutely love working on the kernel? So now I get to poke at the kernel all day, every day... LWN: What are your current kernel projects? What kinds of things do you see yourself doing in the kernel in the future?
Peter: Current active projects are group scheduling and some -rt
work. I should
pick up the swap over network code again, and there are some other loose
ends.
The future ... well we'll see what happens, loads of interesting stuff to do. We would like to thank Peter for taking the time to answer our questions.
The state of the pageout scalability patches The virtual memory scalability improvement patch set overseen by Rik van Riel has been under construction for well over a year; LWN last looked at it in November, 2007. Since then, a number of new features have been added and the patch set, as a whole, has gotten closer to the point where it can be considered for mainline inclusion. So another look would appear to be in order.One of the core changes in this patch set remains the same: it still separates the least-recently-used (LRU) lists for pages backed up by files and those backed up by swap. When memory gets tight, it is generally preferable to evict page cache pages (those backed up by files) rather than anonymous memory. File-backed pages are less likely to need to be written back to disk and they are more likely to be well laid-out on disk, making it quicker to read them back in if necessary. Current Linux kernels keep both types of pages on the same LRU list, though, forcing the pageout code to scan over (potentially large numbers of) pages which it is not interested in evicting. Rik's patch improves this situation by splitting the LRU list in two, allowing the pageout code to only look at pages which might actually be candidates for eviction. There comes a point, though, where anonymous pages need to be reclaimed as well. The kernel will make an effort to pick the best pages to evict by going for those which have not been recently referenced. Doing that, however, requires going through the entire list of anonymous pages, clearing the "referenced" bit on each. A large system can have many millions of anonymous pages; iterating over the entire set can take a long time. And, as it turns out, it's not really necessary. The VM scalability patch set now changes that behavior by simply keeping a certain percentage of the system's anonymous pages on the inactive list - the first place the system looks for pages to evict. Those pages will drift toward the front of the list over time, but will be returned to the active list if they are used. Essentially, this patch is applying a form of the "referenced" test to a portion of anonymous memory - whether or not anonymous pages are being evicted at the time - rather than trying to check the referenced state of all anonymous pages when the kernel decides it needs to reclaim some of them. Another set of patches addresses a different situation: pages which cannot be evicted at all. These pages might have been locked into memory with a system call like mlock(), be part of a locked SYSV shared memory region, or be part of a RAM disk, for example. They can be either page cache or anonymous pages. Either way, there is little point in having the reclaim code scan them, since it will not be possible to evict them. But, of course, the current reclaim code does have to scan over these pages. This unneeded scanning, as it turns out, can be a problem. The extensive unevictable LRU document included with the patch claims:
For example, a non-numal x86_64 platform with 128GB of main memory
will have over 32 million 4k pages in a single zone. When a large
fraction of these pages are not evictable for any reason [see
below], vmscan will spend a lot of time scanning the LRU lists
looking for the small fraction of pages that are evictable. This
can result in a situation where all cpus are spending 100% of their
time in vmscan for hours or days on end, with the system completely
unresponsive.
Most of us are not currently working with systems of this size; one must spend a fair amount of money to gain the benefits of this sort of pathological behavior. Still, it seems like something which is worth fixing. The solution, of course, is yet another list. When a page is determined to be unevictable, that page will go onto the special, per-zone unevictable list, after which the pageout code will simply not see it anymore. As a result of the variety of ways in which a page can become unevictable, the kernel will not always know at mapping time whether a specific page can go onto the unevictable list or not. So the pageout code must keep an eye out for those pages as it scans for reclaim candidates and shunt them over to the unevictable list as they are found. In relatively short order, the locked-down pages will accumulate in this list, freeing the pageout code to concentrate on pages it can actually do something about. Many of the concerns which have been raised about this patch set over the last year have been addressed. A few remain, though. Some of the new features require new page flags; these flags are in extremely short supply, so there is always pressure to find ways of implementing things which do not allocate more of them. There are a few too many configuration options and associated #ifdef blocks. And so on. Addressing these may take a while, but convincing everybody that these (rather fundamental) memory management changes are beneficial under all circumstances may take rather longer. So, while this patch set is making progress, a 2.6.27 merge is probably not in the cards.
Patches and updates Kernel trees
Core kernel code
Development tools
Device drivers
Documentation
Filesystems and block I/O
Memory management
Networking
Architecture-specific
Security-related
Benchmarks and bugs
Miscellaneous
Page editor: Jonathan Corbet
Distributions News and Editorials Looking ahead to Mandriva 2009 Mandriva developer Adam Williamson recently announced the plans for Mandriva Linux 2009. The schedule and other details are available at 2009 development wiki.There will be two alpha releases, two beta releases and two release candidates before the final release in October 2008. The first alpha will be available very soon as the scheduled date is June 25, 2008. As usual Mandriva 2009 will be available in the Free, One (live CD) and PowerPack editions. So what's in store? Users of Cooker, Mandriva's development branch, will have already noticed the churn as gcc is upgraded to 4.3. There's also the switch to newer technologies such as libata and PolicyKit. The final kernel is not yet fixed but will likely be 2.6.26; with server, desktop and desktop586 flavors. The technical specifications available in SVN, where they are changed to reflect progress. I looked at the PDF snapshot for more information. KDE 4.1 and GNOME 2.24 will both be available, along with updated packages such as OpenOffice.org 3 and Firefox 3. There's a new design for the installer, and live distribution upgrade mode for MandrivaUpdate. The package management tools will be smarter about the removal of packages that are no longer required. The Windows migration tools have also gotten smarter, making it easier than ever for new users to get started with Linux. That's just the beginning. There is much more coming up in Mandriva Linux 2009.
Distribution News Fedora Fedora Board & FESCo Recap 2008-JUN-09 The Fedora board and the Fedora Engineering Steering Committee (FESCo) had a joint meeting on June 9, 2008. Click below for the notes.
Fedora Board Recap 2008-JUN-10 A recap of the June 10 meeting of the Fedora board is available (click below). Topics discussed include Codeina, secondary architectures, and a lengthy discussion about the path forward.
No more updates for Fedora 7 Fedora 7 has officially reached end of life and will no longer be updated. This is in keeping with the Fedora lifecycle where releases of a given Fedora N are supported until Fedora N+2 is released plus one month. Fedora 8 will go the same route one month after Fedora 10 is released, which is currently scheduled for late October.
Gentoo Linux Council meeting summary for 12 June 2008 A summary of the June 12 Gentoo council meeting is available. Click below for the summary.
SUSE Linux and openSUSE Accessibility Wiki Launched Accessibility pages have been created on opensuse.org's Wiki. "The intent is to highlight efforts made within the openSUSE community to improve Accessibility (a11y) standards within our distribution, as well as become more aware of who within our community works on a11y issues."
Ubuntu family Changes to Gobuntu The Gobuntu development team has announced that after the 8.04 release of Gobuntu, the project will aim to merge many of the Gobuntu changes into mainline Ubuntu. Ubuntu will feature a "Free Software Only" installer option which only installs software considered free by the Free Software Foundation's definition of software freedom. "This installer option now obviates the need for a separate derivative project, and in the interest of reducing the workload of Ubuntu core developers, the Gobuntu project will instead focus on merging as many changes as possible into mainline Ubuntu."
Intrepid Alpha 1 delayed The first alpha of Ubuntu's Intrepid Ibex release has been delayed. "We are making every effort to get an installable alpha image together in the coming days. Watch this space for further information..."
Distribution Newsletters Ubuntu Weekly Newsletter #95 The Ubuntu Weekly Newsletter for June 14, 2008 covers Intrepid Alpha 1 delayed, more info about Global Bug Jam, future Brainstorm plans, Server Team Intrepid blueprints, new Ubuntu Members, future of Gobuntu, Kubuntu Tutorial Days, Mark Shuttleworth's response to accusations of proprietary codecs in Ubuntu, open source in UK schools, and much more.
OpenSUSE Weekly News/26 This issue of the openSUSE Weekly News looks at openSUSE Launches Merged Forums, Announcing the openSUSE Marketing Team, People of openSUSE: Cornelius Schumacher, Sneak Peeks at openSUSE 11.0, Tips and Tricks: Jigish Gohil: Useful openSUSE One-Click installs from command line, and more.
OpenSUSE Weekly News/27 This issue of the openSUSE Weekly News covers Upcoming openSUSE 11.0, People of openSUSE: Rupert Horstkötter, Sneak Peeks at openSUSE 11.0: KDE with Stephan Binner, Tips and Tricks: Jigish Gohil: Getting NVIDIA and ATI drivers on openSUSE 11.0, Planet SUSE: Ben Kevan: Why upgrade to openSUSE 11 from openSUSE 10.x, and more.
Fedora Weekly News Issue 131 The Fedora Weekly News for June 15, 2008 is out. Topics include Board Elections, website developers wanted, Fedora 7 end of life, and much more.
DistroWatch Weekly, Issue 257 The DistroWatch Weekly for June 16, 2008 is out. "This is openSUSE's week as one of the oldest and most popular Linux distributions prepares for its highly ambitious release. Will the project's switch to Qt 4.x toolkit be a success? And how will the integration of the shaky KDE 4.0.x code into the distribution be received? These are some of the questions many readers are asking before the Thursday release of openSUSE 11.0. In the news section, Mandriva releases Flash 2008.1, a portable distribution on an 8 GB USB key, Debian clarifies the beta status of "Lenny", Linux Mint publishes an important security advisory for one of its utilities, and Sabayon Linux announces the imminent arrival of the final beta for its upcoming version 3.5. Also in this issue, a reader-contributed review of PC/OS 8.04, an Ubuntu based distribution with a BeOS-like user interface, links to two excellent interviews with Mark Shuttleworth, a hands-on guide on turning FreeBSD into a desktop system, and a report on how Microsoft intends to prevent Linux from becoming the operating system of choice on low-cost laptops."
Interviews Mark Shuttleworth Interview with Linux-Magazine Italia Vincenzo Ciaglia from Linux-Magazine Italia talks with Mark Shuttleworth about the release of 8.04 LTS and more. "What do you think about your competitors? Fedora/Red Hat, openSUSE and Mandriva are doing good work as well as Ubuntu. What GNU/Linux distribution do you prefer if you couldn't use Ubuntu? Yes, all of the distributions make contributions to the art and industry of free software. I'm very glad that lots of companies continue to invest in Linux, it makes it a much healthier and more vibrant ecosystem than it would be if just one company dominated it. So I'm very happy with the competition. If Ubuntu didn't exist, I would use Debian."
Distribution reviews GNewSense, the Present and the Future (IT Management) Bruce Byfield takes a look at GNewSense. "At first GNewSense seems virtually identical to Hardy Heron, the latest Ubuntu version. Both use GNOME 2.22, and include such packages as OpenOffice.org 2.4 and GIMP 2.4.5. The branded wallpaper is different, but the muddy brown of Ubuntu's default theme is even visible on the title bar of GNewSense windows. Dig deeper, though, and you'll start to see differences. If you are on a laptop, your wireless card has a strong chance of not working. Your video card probably does only 2-D acceleration, and the desktop has no Restricted Drivers Manager to help you add proprietary 3-D ones. Instead of Firefox, you have Epiphany, another Mozilla-based browser, not because Firefox is proprietary, but because the Mozilla Corporation is strict about enforcing the trademark on its products. At the kernel level, you'll find some 119 firmware files gone, too."
Page editor: Rebecca Sobol
Development The Wine project releases version 1 Wine (Wine Is Not an Emulator) is one of the long-standing Windows interoperability projects that runs under Linux and other Unix-based systems:
Wine is an Open Source implementation of the Windows API on top of X, OpenGL, and Unix.
Think of Wine as a compatibility layer for running Windows programs. Wine does not require Microsoft Windows, as it is a completely free alternative implementation of the Windows API consisting of 100% non-Microsoft code, however Wine can optionally use native Windows DLLs if they are available. Wine provides both a development toolkit for porting Windows source code to Unix as well as a program loader, allowing many unmodified Windows programs to run on x86-based Unixes, including Linux, FreeBSD, Mac OS X, and Solaris. Wine is free software, released under the GNU LGPL.
Although not game-specific, the ability to run Windows games has always been one of the major driving forces behind Wine. The Wine AppDB page lists the numerous Windows applications that have been made to work under Wine. Photoshop CS2 stands out as one of the few most-popular Wine-compatible Windows applications that is not a game. The Wine Features document lists Wine's capabilities, it is capable of running DOS through Windows XP applications, Windows Vista compatibility is not yet mentioned. The About Wine document explores the project's history, contributors, myths and more. The history document details the magnitude of the project: "Wine has grown to over 1.4 million lines of C code over the past decade. Nearly 700 people have contributed in some fashion. As always, you can expect Wine to be released sometime this year; or maybe early next year." ![[WineHQ]](/images/ns/WineHQlogo.png)
Version 1.0 of Wine was announced (see the LWN reader comments) on June 17, 2008:
The Wine team is proud to announce that Wine 1.0 is now available.
This is the first stable release of Wine after 15 years of development
and beta testing. Many thanks to everybody who helped us along that
long road!
There have been a series of Wine 1.0 release candidates over the last month involving a ton of bug fixes, janitorial code work, translation improvements and more. The details are available in the series of release notes for RC1, RC2, RC3, RC4, RC5 and finally version 1.0. Binary packages and source code for Wine 1.0 are available for download. While fairly unusual for most open-source projects, a commercial distribution of Wine known as CrossOver is available from Code Weavers. CrossOver Linux 7.0, which is synchronized with Wine 1.0, was announced this week.
System Applications Database Software cx_Oracle 4.4 released Version 4.4 of cx_Oracle has been announced. "cx_Oracle is a Python extension module that allows access to Oracle and conforms to the Python database API 2.0 specifications with a few exceptions."
Firebird 2.1.1 Release Candidate available A new release candidate for version 2.1.1 of the Firebird DBMS has been announced. "The Firebird Team is pleased to make a release candidate available for field testing the first V.2.1.x patch release on Windows, Linux and MacOSX Intel platforms. Please test it hard and report any problems to the firebird-devel list."
PostgreSQL 8.3.3 and 8.2.9 update releases Update releases 8.3.3 and 8.2.9 of the PostgreSQL DBMS are available. "Updates for all maintained versions of PostgreSQL are available today: 8.3.3, 8.2.9, 8.1.13, 8.0.17 and 7.4.21. These releases fix more than two dozen minor issues reported and patched over the last few months. All PostgreSQL users should plan to update at their earliest convenience. Users of UTF-8 databases on Windows and people in affected time zones, in particular, should upgrade as soon as possible."
python/pg_proboscis 1.0 released Version 1.0 of python/pg_proboscis has been announced. "pg_proboscis is a Python programmer's client for PostgreSQL(driver/interface)."
PostgreSQL Weekly News The June 15, 2008 edition of the PostgreSQL Weekly News is online with the latest PostgreSQL DBMS articles and resources.
Filesystem Utilities Allmydata.org "Tahoe" v1.1 announced Version 1.1 of Allmydata.org, a secure decentralized filesystem, has been announced. "This is the successor to Allmydata.org "Tahoe" Least Authority Filesystem v1.0, which was released March 25, 2008 [1]. This release fixes several serious issues in Tahoe v1.0, and improves the user interfaces."
Interoperability Samba 3.2.0rc2 is available Version 3.2.0rc2 of Samba has been announced. "This is the second release candidate of Samba 3.2.0. This is *not* intended for production environments and is designed for testing purposes only."
Virtualization Software Flashlight-VNC: 1.0.0 released (SourceForge) Version 1.0.0 of Flashlight-VNC has been announced. "Flashlight-VNC is a VNC Viewer / Player written in Flash. It can connect to a VNC server and play recorded VNC sessions in FBS format. It supports Tight Encoding. License in LGPL. Cross-platform, web-based, it can easily be integrated in a web page or in a flash interface."
Web Site Development Apache HTTP Server 2.2.9 released Version 2.2.9 of the Apache web server has been announced. "This version of Apache is principally a bug and security fix release."
nginx 0.7.2 released Version 0.7.2 of the nginx web server has been announced, it includes new features and bug fixes. See the CHANGES file for details.
webminstats: 0.13.0 released (SourceForge) Stable version 0.13.0 of webminstats has been announced. "Webminstats is a Webmin's module which display a graphical log of historic information. It's modular in design, as to be able to log everything from CPU usage to email box size. this is a major release : many changes on interface and core code".
Desktop Applications Audio Applications PulseAudio meets BlueTooth There is an effort to add support for Bluetooth devices to the PulseAudio networked sound server project, as documented by notes from the recent BlueZ meeting in Helsinki. "There should be two new PA modules, module-bluetooth-discover and module-bluetooth-device. The former will use D-Bus to connect to the BlueZ system services and whenever a new BT audio devices appears load one m-bt-device instance for it. (as a side node: in contrast to linux kernel modules, PA modules can be loaded more than once at the same time). The latter, m-bt-device, then connects to the BlueZ audio service via one BlueZ specific well known unix socket, configures a connection to the BT device, gets a BT socket fd passed in via the unix socket and then hands this over to its RT thread." (Thanks to Ernst Persson).
QjackCtl 0.3.3 (unstable-qt4) released! Version 0.3.3 of QjackCtl, a control application for the Jack Audio Connection Kit (JACK), has been announced. "Release highlights are mainly about final JACK-MIDI support for the "evil" Patchbay, new Messages file logging and the most intriguing application window instance uniqueness which will make X11 desktop life easier for everyone (ie. no more duplicates as JACK server gets auto-started as candy bonus:)"
XMMS2: DrLecter escaped A new version of the XMMS2 music player has been announced. "Finally a new XMMS2 release has arrived. This time it contains even finer meat, such as a whole bunch of new plugins and support for ruby 1.9. The XMMS2 Team would like to extend a big THANK YOU to all who have helped out with this release. Hope you will have as good time eating it as we had cooking it!"
Data Visualization Asymptote: 1.43 released (SourceForge) Version 1.43 of Asymptote has been announced, it adds some new capabilities. "Asymptote is a powerful descriptive vector graphics language for technical drawing, inspired by MetaPost but with an improved C++-like syntax. Asymptote provides for figures the same high-quality level of typesetting that LaTeX does for scientific text.
Desktop Environments GNOME 2.23.4 released Development version 2.23.4 of the GNOME Desktop has been announced. "You all know what you have to do now. Go download it. Go compile it. Go test it. And go hack on it, document it, translate it, fix it."
GNOME Software Announcements The following new GNOME software has been announced this week:
KDE Software Announcements The following new KDE software has been announced this week:
Xorg Software Announcements Xorg Software Announcements The following new Xorg software has been announced this week:
Financial Applications SQL-Ledger 2.8.15 released Version 2.8.15 of SQL-Ledger, a web-based accounting system, has been announced: "changed code to get rid of the "Bizarre copy of ARRAY in aassign" error cropping up on newer perl versions".
Games Imperium: Sticks version 0.1.5 released (SourceForge) Version 0.1.5 of Imperium: Sticks has been announced. "Imperium: Sticks is a cross-platform, real-time strategy game spanning the dawn of human history to the end of civilization. A new version of Imperium: Sticks is out. Version 0.1.5 now allows for 8-directional sprites, art upgrades to the female worker unit, and bugfixes."
pymunk 0.8 released Version 0.8 of pymunk, a Python wrapper for the Chipmunk 2d rigid body physics library, has been announced. "It puts a pythonic layer above chipmunk to make it easy to use for python programmers. The main goal with pymunk is to make 2d physics easy to include in your game/project." This is the initial public release.
GUI Packages JUCE: Version 1.46 released (SourceForge) Version 1.46 of JUCE has been announced. "JUCE is a C++ toolkit for building cross-platform applications on PC/Mac/Linux. Particularly good for complex, customised GUIs and audio/midi processing, it also includes a vast range of classes to help with all your day-to-day programming tasks."
Multimedia aTunes: 1.9.0 released (SourceForge) Version 1.9.0 of aTunes has been announced, it features some new capabilities and bug fixes. "aTunes is a powerful, full-featured, cross-platform player and manager, with audio cd rip frontend. Currently supported formats are mp3, ogg, wav, wma, flac, mp4, ape, mpc, mac, radio streaming and podcasts."
Office Suites KOffice 2.0 Alpha 8 (KDE.News) KDE.News takes a look at KOffice 2.0 Alpha 8. "The KDE Project today announced the eighth alpha release of KOffice 2, a technology preview of the upcoming version 2.0. Work continues in the same vein as before, with a strong focus on finishing and polishing our new features that will set KOffice. This is a work in progress, showing the changes that have been made over the last month by the KOffice developers. Most features that will be part of the final release are present now, and bug reports are welcome for the more stable components."
Video Applications Burn 360: New Release (SourceForge) Version 0.7 of Burn 360 has been announced. "This is a front-end to ffmpeg and standard VCD/DVD ripping creation programs written in perl-gtk2. It is designed to enable (as separate processes) ;VCD/DVD ripping;Any Media-transcoding that ffmpeg supports ;VCD/DVD creation".
Web Browsers Field Guide to Mozilla Firefox 3 Details New and Improved Features (MozillaZine) A new guide to Firefox 3 has been announced on MozillaZine. The Field Guide to Firefox 3 provides a in-depth look, with many screenshots, of the new features in Firefox 3. "Firefox 3 is going to be launched very soon. In anticipation of this long-awaited event, the folks in the Mozilla community have been writing extensively about the new and improved features you'll see in the browser. The new features cover the full range from huge and game-changing to ones so subtle you may not notice them until you realize that using Firefox is just somehow easier and better. The range of improved features is similar — whole back-end systems have been rebuilt from scratch, while other features have been tweaked slightly or redesigned in small ways."
Languages and Tools C Converting GCC to C++ Ian Lance Taylor has announced the creation of a branch to explore the idea of converting the GCC code base from C to C++. He has also posted slides from a talk [PDF] on why he thinks this is a reasonable thing to do. In short, he thinks the change can help to simplify the code and make interfaces more robust while, with luck, avoiding the C++ language's worst problems.
Caml Caml Weekly News The June 17, 2008 edition of the Caml Weekly News is out with new articles about the Caml language.
Java CACAO 0.99 released Version 0.99 of CACAO, a Java Virtual Machine, has been announced. "The major feature enhancement of this release is the OpenJDK support. CACAO's libjvm.so can now be used as drop-in replacement for Sun's HotSpot libjvm.so in OpenJDK. There is also support in IcedTea available to use CACAO as JVM (--with-cacao option). CACAO uses GNU Classpath as default Java runtime library and supports upstream releases or CVS snapshots."
OpenSwing: 1.6.4 released (SourceForge) Version 1.6.4 of OpenSwing has been announced, some new features have been added. "OpenSwing is a component library that provides a rich set of advanced graphics components and a framework for developing java applications based on Swing front-end. It can be applied both to rich client applications and Rich Internet Applications."
JSP MicroNova YUZU: 20080616 released (SourceForge) Version 20080616 of MicroNova YUZU has been announced. "MicroNova YUZU is a BSD-licensed JSP tag library designed to augment JSTL (JSP Standard Tag Library) using EL (Expression Language). YUZU is compatible with both JSP 1.2 and JSP 2.0 specifications (tomcat 4.x/5.x). JSTL/tagfiles along with YUZU transforms JSP into a powerful framework-independent XML-style "scripting language" for web applications and DSL (domain specific languages)."
Lisp SBCL 1.0.17 has been released Version 1.0.17 of Steel Bank Common Lisp has been announced. "This version adds a runtime argument to adjust the default control stack size, optimizes several list operations, and fixes many bugs."
Perl This Week on perl5-porters (use Perl) The June 1-6, 2008 edition of This Week on perl5-porters is out with the latest Perl 5 news.
Python AVC 0.6.0 released Version 0.6.0 of AVC has been announced. "AVC is a multiplatform, fully automatic, live connection among graphical interface widgets and application variables for the python language. AVC supports in a uniform way the most popular widget toolkits: GTK+, Qt3, Qt4, Tk, wxWidgets"
eGenix pyOpenSSL Distribution 0.7.0-0.9.8h-1 released Version 0.7.0-0.9.8h-1 of eGenix pyOpenSSL, a repackaged distribution of the pyOpenSSL Python interface for OpenSSL, has been announced. "This is the first release of the eGenix.com pyOpenSSL Distribution. It includes pyOpenSSL 0.7.0 and the OpenSSL 0.9.8h libraries on all supported platforms."
Python-URL! - weekly Python news and links The June 16, 2008 edition of the Python-URL! is online with a new collection of Python article links.
Tcl/Tk Tcl-URL! - weekly Tcl news and links The June 13, 2008 edition of the Tcl-URL! is online with new Tcl/Tk articles and resources.
Miscellaneous LLVM 2.3 released Version 2.3 of LLVM (Low Level Virtual Machine) has been announced. "LLVM 2.3 includes many bug fixes, vastly improved support for the X86-64 ABI, support for SSE 4.1 on X86 chips, support for functions that return multiple results in memory, a new 'llvmc' tool, support for atomic operations, improved gfortran support, and many new and improved optimizer and code generator passes. Overall, LLVM 2.3 generates significantly better code in less time than LLVM 2.2, which was released less than 4 months ago - an amazing rate of progress."
Page editor: Forrest Cook
Linux in the news Recommended Reading Red Hat Makes History With Patent Settlement (Groklaw) Groklaw discusses Red Hat's patent settlement with an eye toward compliance with the GPLv3 license. "You've probably been wondering why I've been quiet, when there is news about a patent settlement between Red Hat and Firestar and DataTern in the JBoss litigation. It's because I wanted to be positive I was correct that this is the first known settlement involving patents that is harmonious with GPLv3. It is. It's also harmonious with GPLv2, of course, but this is history in the making, friends. They settled a lawsuit brought against them in a way that licenses patents without violating the GPL."
Ari Jaaksi of Nokia Wants to Educate the Linux Community (Technocrat) Bruce Perens discusses comments by Nokia's Ari Jaaksi on DRM and intellectual property rights issues under Linux. "The key is knowing how to draw bright lines between different parts of the system. That's a legal term, and in this case it means a line between the Free Software and the rest of the system, that is "bright" in that the two pieces are very well separated, and there is no dispute that one could be a derivative work of the other, or infringes on the other in any way. All of the Free Software goes on one side of that line, and all of the lock-down stuff on the other side."
Linux Adoption Why Linux Failed at Walmart -- And What to Do About It (Datamation) Matt Hartley opines on problems selling Linux machines at big-box stores. "Stop selling junk to those who don’t want it in the first place. From Xandros to Linspire, even more recently gOS: In each instance, I have watched in amazement as otherwise perfectly good Linux distributions were bundled with extremely low-end hardware and sold to people who honestly were not in the market for what was being offered."
Legal Copyright, Technology, and Access to the Law: An Opinionated Primer Attorney James Grimmelmann has put together a detailed history of issues surrounding access to the law as a way of looking at the state of Oregon's attempts to assert copyright privileges on its legal code. He concludes with a set of principles for access to legal code which has a lot of similarities with our principles regarding source code. "The right to distribute: Once primary legal information is online, anyone should be able to pick it up and republish it as they see fit. That could be coordinated with other materials in larger collections, it could be pushed out to mobile devices, it could be reformatted or sliced and diced in some clever new way."
Interviews Interview with Marten Mickos (ODBMS Industry Watch) Roberto Zicari interviews Sun's database group head Marten Mickos. "Q1. It appears as if the positioning of MySQL has been refocused more predominately on the Web applications / SaaS / ASP market in the last year or so. Would you agree with this, and if so, what does that mean regarding the potential of MySQL to penetrate further into the enterprise? Marten Mickos: Great question. We believe that enterprises will move to web-based architectures, and with that wave, MySQL is penetrating the web market. Goldman Sachs stated in 2006 that "the shift to more web-based applications in the enterprise is unstoppable". The percentage is still relatively low (10-20% I think) but it is growing."
Resources Get ready for Firefox 3.0 (developerWorks) developerWorks has a web developer's guide to the many new features in Firefox 3.0. "The reason for all the excitement around Firefox 3.0 is that it promises an impressive array of improvements for the Web user and developer alike. It's an important milestone because Web developers have often used Firefox as a favorite development platform, even though they know they have to make eventual concessions toward cross-browser compatibility. Firefox's popularity among developers stems from the vibrant community, the rigorous standards support, and the constant innovation in the platform. The platform's innovation gives developers a head start on Web trends, and its standards support means that even on the cutting edge of these trends, Firefox offers transparency in its features that accelerates adoption and improves compatibility. Firefox 3 has even more to offer."
Getting down to basics: Running Linux on a 32-/64-bit RISC architecture - Part 1 (Embedded.com) Embedded.com has the first in a series of six articles about running Linux on MIPS-based systems. This installment provides definitions of Linux terms as a basis for the rest of the series. "To get to grips with any artifact you need to attach some good working meaning to the terms used by its experts, and you are particularly likely to be confused by terms you already know, but with not quite the same meaning. The UNIX/Linux heritage is long enough that there are lots of magic words: thread, file, user mode and system calls: interrupt context, Interrupt service routine (ISR), scheduler, memory map/address space, thread group, high memory, libraries and applications. "
Visualizing time-dependent data with distortion portals (developerWorks) developerWorks creates an SDL-enabled application that allows you to create distortion portals in sequential image frames to explore the relationship of data sets through time. "There are many ways to visualize data as snapshots, or sequential images showing trending and time progressions. Few options exist for exploring the relationships between data sets through time with an interactive interface. This article demonstrates code and techniques to create what I call "animated distortion portals" in the data to provide time-dependent visualizations of various parts of the image. Additionally, certain aspects of the code are presented that allow for effective visualization on slower-computing platforms without sacrificing usefulness. The code presented here will allow new insights into application-flow models and usage patterns by exploring various data sets and how they move through time."
Companies, Developers Contributing To The X Server (Phoronix) Phoronix has dug through the X.org source repository to see where the changes are coming from. "The biggest software company contributing to the X server has been Red Hat, with the second largest contributor being Tungsten Graphics and they had more than three times fewer commits than Red Hat. Novell / SuSE was in third. Combined, these software vendors had made up nearly 33% of the total commits over the past nine years."
Reviews Hands on: Flock 2 steps up the social browser game (ars technica) ars technica reviews the Flock 2 beta release. "While Flock is based on the Firefox code base, the most noticeable addition is a toolbar of socially-themed icons to the left of the bookmark bar. This toolbar is perhaps the most important launching pad for Flock's strengths as a social web browser, as it can open various sidebars and media bars that act as portals to services like Flickr, Twitter, del.icio.us, Blogger, YouTube Facebook, TypePad, Gmail, and more."
Movie download device runs embedded Linux framework (LinuxDevices) LinuxDevices takes a look at Roku's Netflix Player. "Roku's Netflix Player (pictured at left) was announced last month as a way to reduce tech support calls for Netflix. The device is designed as an embedded Linux alternative to the Netflix Instant application, which runs on Windows PCs. Subscribers can visit the Netflix website using a laptop or PC and add movies to their Instant queue. The selected items can then be browsed using a TV or projector, using the Player's remote control. The remote can also be used to control playback, rate movies, or stop a movie for later resumption."
Page editor: Forrest Cook
Announcements Non-Commercial announcements Announcing the Django Software Foundation The Django Software Foundation has been launched. "Django, started nearly five years ago by programmers affiliated with The World Company, now joins a lineup of pervasive computer languages and systems — including Mozilla, Apache and Linux — to be overseen by a nonprofit organization. We're still breaking this baby in, so we're a little light on details for now. You can read a bit about our goals now, and as you can imagine we'll be talking a lot about this in the days and weeks to come."
EFF: Judges Urged to Curtail Random Searches of Travelers' Laptops The Electronic Frontier Foundation has issued a press release concerning random searches of travelers' laptops at the US border. "The news media has reported extensively on these searches as well as the surprise and anger felt by American travelers when they are singled out for inspection. In a typical search, U.S. border officials will turn on the computer and then open and review files. If agents see something of interest, they may confiscate the computer, copy its contents, and sometimes provide a copy to the Department of Justice -- even when the traveler is not suspected of criminal activity. In some cases, travelers have never gotten their computers back from the government."
Commercial announcements CrossOver Linux 7.0 announced Celebrating Wine 1.0 CodeWeavers has announced version 7.0 of CrossOver Linux and CrossOver Mac. "I am pleased to announce that we have shipped CrossOver 7 for both Macintosh and Linux. New in Version 7 is support for Microsoft Office 2007, dramatically improved support for Outlook 2003 and Internet Explorer 6, and a broad range of improvements that should bring improvements to all Windows applications. For our Linux customers, it also brings expanded support for most Adobe programs, with Photoshop CS and CS2 working particularly well."
IronKey announces Linux support IronKey Inc. has announced that it will support Linux on its flash drives. "IronKey brings unprecedented mobile data security to individual, enterprise and government users. Its combination of security applications and services along with hardware-based encryption deliver unmatched security for its USB drives, and protection of the data stored on them. Extending these capabilities onto Linux platforms provides a comprehensive solution for USB data security to network administrators running multiple operating systems."
Novell enters Pre-Load Agreement with Micro-Star International Novell, Inc. has announced a partnership with Micro-Star International. "In order to meet increasing global customer demand for Linux on client systems, Novell today announced that Micro-Star International (MSI), a leading manufacturer of computer hardware products and solutions based in Taiwan, is pre-installing SUSE Linux Enterprise Desktop 10 on MSI's low-cost mini-notebooks. The initial pre-loaded model, the Wind Notebook, is one of the first in the rapidly growing segment of low-cost mini-notebooks, termed "netbooks" by Intel Corporation."
Opera 9.5 is now available Version 9.5 of the Opera browser has been announced. "Opera Software today released to the world the final public release of its flagship Web browser, Opera 9.5. Opera's cross-device expertise, support for open Web standards and commitment to speed and performance culminate to create the most powerful Opera browser yet. Making its desktop debut in Opera 9.5, Opera Link blurs the boundaries between computers and mobile phones by enabling a seamless Web experience from device to device."
Red Hat Puts Patent Issue to Rest Red Hat has announced the settlement of patent litigation involving Firestar Software, Inc. and DataTern, Inc. "Like most settlements, this one ends the specific lawsuit between the parties, but it does much more than that. It assures that upstream developers are protected against patent suits by DataTern and Amphion with respect to projects incorporated into Red Hat products. In addition, our distributors, customers, and anyone else who uses Red Hat products are protected with respect to Red Hat products. This broad coverage is a significant benefit to the open source community." (Thanks to Andreas Thieneman)
Terra Soft Solutions to resell Allinea tools Terra Soft Solutions and Allinea have announced a partnership. "Allinea Software, a leading provider of innovative products for large-scale scalar and parallel high performance computing applications, today announced a new reseller agreement whereby Terra Soft Solutions offers Allinea's Distributed Debugging Tool (DDT) and Optimization and Profiling Tool (OPT), both recently made available for Cell Broadband Engine (Cell BE) platforms."
Third Brigade acquires OSSEC Third Brigade has announced the acquisition of OSSEC. "Third Brigade, a security software company specializing in host intrusion detection and prevention systems today announced that it has acquired OSSEC, a leading open source, host-based intrusion detection system, and that Daniel Cid, creator and primary developer for OSSEC, has joined the Third Brigade team as Principal Researcher, OSSEC Development."
New Books Learning Flex 3--New from O'Reilly O'Reilly has published the book Learning Flex 3 by Alaric Cole.
Ubuntu for Non-Geeks, 3rd Edition--New from No Starch Press No Starch Press has published the book Ubuntu for Non-Geeks, 3rd Edition by Rickford Grant.
Resources PostgreSQL Software Catalogue launched The PostgreSQL DBMS project has announced a new software catalogue. "Linked from the Download section of the PostgreSQL website, the Software Catalogue aims to list all the drivers, tools, applications and other software available to work with and on PostgreSQL in a single place to avoid all that tedious googling previously required!"
Meeting Minutes Perl 6 Design Meeting Minutes (use Perl) The minutes from the June 4, 2008 Perl 6 Design Meeting have been published. "The Perl 6 design team met on 04 June 2008 by phone. Larry, Allison, Patrick, Jerry, Jesse, and chromatic attended."
Calls for Presentations FOMS2009 Call for Participation A Call for Participation has gone out for the S2009 Foundations of Open Media Software Developer Workshop. The event takes place in Hobart (Tasmania), Australia on January 15-16, 2009, submissions are due by August 15.
Linux Plumbers Conference call for proposals and registration The Linux Plumbers Conference will be held in Portland, Oregon, on September 17 to 19 - immediately after the kernel summit. The event organizers have released the call for proposals for those who would like to speak at the event. The Plumbers Conference is an interesting attempt to bring a low-level development conference back to the U.S.; it should certainly be worth attending.For those who are interested in going to the conference but not in speaking, registration for attendees has also opened.
OSDC 2008 CFP closes June 30 A final call for papers has gone out for Australia's Open Source Developers' Conference. The submission deadline is June 30. "Papers on Open Source languages, technologies, projects, tools and well as topics of interest to Open Source developers are being solicited for submission to OSDC 2008. So if you are an Open Source maintainer, developer or user, the organising committee would encourage you to submit a talk proposal on the open source tools, solutions, languages or technologies you are working with. The conference will be held at SMC Conference venue in the Sydney CBD, New South Wales, from the 2nd to the 5th of December, 2008."
Upcoming Events EuroPython 2008: Timetable Published The conference timetable for EuroPython 2008 is available. "This year? schedule includes a significant Jython presence, with talks, tutorials and a panel covering the technology, applications and the road ahead for the project. Once again, Web and Internet technologies are well represented, with distributed computing being an emerging topic this time around. Not to be outdone, representatives of the PyPy project will be on hand to describe the current state of their work."
Vilnius/Post EuroPython PyPy Sprint The Vilnius/Post EuroPython PyPy Sprint will take place on July 10-12. "The PyPy team is sprinting at EuroPython again and we invite you to participate in our 3 day long sprint at the conference hotel - Reval Hotel Lietuva. If you plan to attend the sprint we recommend you to listen to the PyPy technical talks during the conference since it will give you a good overview of the status of development."
EuroSciPy - Early Bird Deadline Extended The Early Bird Deadline for EuroSciPy has been extended. "I would like to announce that the early bird deadline for the first-annual EuroSciPy conference is extended to June 20, 2008. So you have a few more days to take advantage of the reduced rate."
2nd ImageJ User and Developer Conference (LinuxMedNews) LinuxMedNews has announced the 2008 ImageJ User and Developer Conference, the event takes place in Luxembourg on November 6-7, 2008. "After the successful first ImageJ User and Developer Conference in 2006, the Public Research Center Henri Tudor is currently planning the second edition of this event in Luxembourg. ImageJ is a public domain Java image processing that runs, either as an online applet or as a downloadable application, on any computer with a Java 1.4 or later virtual machine."
LinuxWorldExpo PostgreSQL Day and More The LinuxWorldExpo PostgreSQL Day has been announced. "The San Francisco PostgreSQL User Group, together with PostgreSQL.us, will be having several events around LinuxWorldExpo in San Francisco On August 5th, in space generously donated by IDG, we will have a full day of PostgreSQL technical sessions. This will be followed by some kind of social event with drinks & food. We will also have a booth, of course and are seeking booth volunteers."
Plans for PyCon 2009: Dates and Venue The initial plans are being made for PyCon 2009, the event will be held in Chicago, IL. "Mark your calendars: * Tutorial days: Wednesday March 25 & Thursday March 26, 2009. (Two days of tutorials next year instead of one!) * Conference days: Friday March 27 through Sunday March 29, 2009. * Development sprints: Monday March 30 through Thursday April 2, 2009."
Events: June 26, 2008 to August 25, 2008 The following event listing is taken from the LWN.net Calendar.
If your event does not appear here, please tell us about it. Web sites OpenOffice.org forum adds Japanese language support The OpenOffice.org forum has added support for the Japanese language. "This web-based Community Forum brings experts and users together to provide support to users of OpenOffice.org, the free alternative to Microsoft Office. The Forum is free to users and is free of advertisements. Since its launch six months ago, the Forum has greatly exceeded expectations, attracting a registered membership of nearly 13,000."
Page editor: Forrest Cook
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||