User: Password:
Subscribe / Log in / New account Weekly Edition for December 6, 2012

Google DocCamp: documentation unconference

By Nathan Willis
December 5, 2012

Google played host to its second annual "Summer of Code Documentation Camp" at its Mountain View offices over the first week of December, at which teams from three free software projects gathered to sort out and write sorely-needed documentation in an intensive workshop setting. This year's participating projects were the Evergreen integrated library system (ILS), the Etoys educational programming environment for children, and the open source font creator FontForge. During the last three days of the camp each team was tasked with producing its own self-contained manual through a "book sprint," but the first two days were composed of an unconference-style workshop about tackling the challenges of documentation in an open project.

Despite the name of the event, the camp is a distinct entity from Google's university-oriented Summer of Code mentorship program (although warm in December, Mountain View is demonstrably not in the Southern hemisphere...). But the two programs do share a common theme. The facilities and amenities were donated by Google, but the week's activities were run by Adam Hyde of FLOSSManuals and Allen "Gunner" Gunn of Aspiration Tech. FLOSSManuals, of course, is a project dedicated to enabling and encouraging open source communities to write and publish quality documentation; Aspiration Tech is a training and consulting group focused on working with nonprofits and software.


The unconference portion of the week consisted of a series of breakout sessions exploring documentation-related subjects. The unconference format means, first and foremost, that the agenda for the sessions is not fixed in advance, but is created by the participants during the event itself. The result is a program that tackles the specific needs of the participants — which, considering the time constraints of the week, was especially important.

Gunn first led the participants in a series of brainstorming exercises in order to generate a pool of topics of interest. The participants then read through all of the suggestions and attempted to group them into recurring subjects. From the resulting ad-hoc taxonomy, Gunn and Hyde selected the initial breakout session topics. But after the first round of breakouts was over, it was up to participants themselves to facilitate the sessions, exploring the topic at hand and deciding what (if any) additional sessions needed to follow later in the day.

At first glance, it might seem like the three projects represented at the unconference would have very little in common to discuss. Yes, every open source project struggles with writing documentation, just like every open source project struggles with recruiting new talent. But the projects themselves differed considerably in scope, purpose, and technical detail. Evergreen is an integrated, server-based application suite with a web interface, used and managed by a very specific type of professional institution. Etoys is a self-contained interpreted programming environment designed for classrooms and tailored toward young children. FontForge is a graphical desktop application used for an isolated task typically undertaken only by specialists.

But the brainstorming exercises quickly revealed that the projects grapple with a nearly identical set of documentation challenges. All three struggle with accessibility and translation tasks, all three have difficulty integrating updates to documentation with updates to the codebase, all three have trouble setting up and maintaining a workflow among documentation volunteers. Consequently, in each breakout session (there were three or four simultaneously during each session slot) Gunn ensured that there was at least one representative from each of the three teams, so the teams were able to exchange information and learn from each others' experiences.

Prep work

None of the session topics lent itself toward simple "how to" answers. Rather, the goal of the breakouts was to prime each group's thinking on the documentation tasks, in preparation for the writing portion of the week that followed. But there were still insights to be found in each discussion.

For example, in the translation session (in which I participated as a representative of the FontForge team), it came out that all of the projects have found success with open source tools for translating application strings and in-program help messages: Pootle, Launchpad, and other tools handle this task quite well, providing an overview that indicates the percentage of translation completed for each available language, offering translation suggestions based on other projects' work, and so on. But none of those features are available in the systems that projects use to maintain their documentation. These string translation tools are designed to handle short snippets of text only, and typically fail on paragraph-length content: long sections of text do not fit into the boxes in the interface, and the features that track completion expect the entire text to be translated at once (which is rarely possible or desirable with lengthy discussions). Conversely, the content management systems that support multi-lingual web sites do not offer the percentage-completion tracking or suggestions that make application translation more manageable.

The other topics examined during the breakout sessions included soliciting and incorporating audience feedback, iterating and versioning documentation, developing and maintaining a healthy community of documentation writing, finding tools to support remote documentation efforts, and many more. By Tuesday, however, the program shifted away from general topics and toward the specifics that each team needed to address for its own book sprint. This meant nailing down a specific target audience and drafting an outline of what subject matter would be covered, start to finish. That can be a daunting task, particularly in light of the knowledge that only three days will be available for content creation and editing, and with teams of only five or six people.

The writing itself began in earnest late on Tuesday, and although the deadline still looms large and ominous, it is remarkable how quickly the work starts to take on a concrete shape. But that is the secret that FLOSSManuals has set out to share with the free software community: writing documentation may seem hard, but then again writing software is hard, too. And just like writing software, once the community sets its mind to the task, it can accomplish considerable feats. That lesson does not end on Friday, either; already there has been talk among the projects about what subjects ought to be tackled in future book sprints. Even though Evergreen, Etoys, and FontForge will walk away from the week with a new book, hopefully the lesson will not be limited to just those three projects, either.

[The author would like to thank Google for support to attend the 2012 Summer of Code Documentation Camp]

Comments (3 posted)

Darktable 1.1

By Nathan Willis
December 5, 2012

Darktable is an open source raw photo converter with built-in image-library-management features, which puts it up against stiff competition. Previous releases boasted a wealth of functionality but were hampered by the application's peculiar user interface. The just-released 1.1 version, however, makes big strides forward in usability while still adding several interesting new features. The new features include front-end functionality, plus a command-line interface and GPU hardware acceleration through OpenCL.

We last looked at Darktable in November 2011, shortly after the release of version 0.9.3. At that time, the application offered a substantial collection of photo adjustment tools via plugins — including several tools not offered by competing raw converters like Rawstudio or RawTherapee. But the interface made them difficult to use: unlabeled controls, nonstandard widgets and status indicators, and a plugin selection palette composed entirely of cryptic, similar-looking logos.

Consequently, the biggest news for most users is that Darktable has evolved into a far more usable product, with interface updates touching most areas of functionality.

Photo editing

    1.1 UI]

All of the same plugins and their icons are still available; they have simply been organized into a scrollable list, with the name of the plugin next to its logo. The downside is that it might look a wee bit less like a TIE fighter's control panel — but it is far more usable. Darktable's editing interface allows you to stack multiple adjustments on top of each other by activating their plugins. This approach is different from the model employed by other raw editors (Rawstudio, for example) in which the adjustments available are presented as intrinsic qualities of the image. In Rawstudio's approach, for instance, an image has one tone curve, and you can change it or leave it alone. In Darktable, you can adjust the "base curve," but you can also apply (for example) high-pass or low-pass filters, each of which adjusts the curve in its own particular way. Order is important in Darktable's approach; if you desaturate an image then try to adjust its color balance, you will not have much to play with.

One of the benefits of Darktable's approach is that the developers can implement quirky and original features as plugins — effects that incorporate several types of adjustment at once. In older Darktable releases, though, some of these unique effects filters exhibited the most troublesome usability hangups. The new release fixes almost all of the issues: most of the curves, axes, and units are labeled, and in most cases it is clear what effect changing a widget will have on the image. Where controls remain unclear, there is usually a pop-up tooltip with a decent explanation. Many of the controls now feature a combination label-slider-spinbox akin to the "spin scales" (or as I call them, "spladers"...) now found in GIMP.

There are new plugins and adjustment features on display as well, including conditional blending — which is precisely what it sounds like. Conditional blending allows you to apply a blend mode (e.g., "multiply" or "soft light") only to regions of the image that fall within a particular color or brightness range. There is also a nice equalizer plugin that enhances local contrast, bringing out small image details without radically affecting the overall tone of the picture.

Speaking of user interfaces, the 1.1 release also introduces a command-line interface, darktable-cli. At the moment, it is only capable of resizing images, but the potential for using Darktable in scripting is intriguing.

Photo juggling

Those of us not blessed with obsessive-compulsive tendencies tend to let our file storage get messy. In the old days, photographers would have called this the "shoebox problem" in reference to stacks of boxes filled with negatives and prints. A fully digital workflow alleviates this to some degree, but hunting for a half-remembered image in the desktop environment's file manager remains a slow and frustrating ordeal. Although an entire category of application has sprung up to offer a hand (the "image manager" like Shotwell or Digikam), most raw photo editors are still forced to incorporate some file collection management and search functionality, simply to save the user from switching back and forth repeatedly.

Darktable 1.1 adds muscle to its file management skills. By far the flashiest new feature is similar-image-search, which scours the database of imported image files looking for photos that appear "similar" visually — as scored by histogram, color, and lightness . I am particularly partial to this feature because it was one of the main selling points of imgSeek, a now-defunct project that was the subject of my first-ever published review, and remains a rarely-seen feature. As was the case with imgSeek, the results of similar-image-searching are imprecise, but if you have imported your entire collection, it would surely assist you in finding the odd mis-labeled image buried in a strange directory.

Darktable also allows you to categorize images in "film rolls" (which despite the now-archaic terminology are merely named collections), in addition to applying keywords, tags, and other metadata. How keywords differ from tags is not explained, other than the fact that they reside on opposite sides of the screen. What is more distinctive is support for geotagging images, complete with a colorful map widget. The map feature is not fully integrated into the other image management tools — instead, it is one of the four top-level application tabs (the other three being image management, photo editing, and tethering). Finally, the application has a new "Group images" mode (which is toggled on or off with a G button in the upper toolbar) which, when activated, hides redundant images, such as the JPEG versions of existing raw photos.

Darktable 1.1 thus gives you multiple ways to find the image you are thinking of, based on automatic or user-assigned metadata, image properties, and the photo's point of origin. The application also gives you quick access to common editing tasks from within the image management tab. Direct export is the most obvious task, but Darktable can also take a select set of differently-exposed images and blend them into a high dynamic range (HDR) image, or it can immediately apply a user-defined "style" with one click of a button. These styles amount to templates; to define one you select an image to which you have made changes and save its current state as a new style.

Shoot first, accelerate later

The last of the new user-visible features in the new release is support for live previews in tethered shooting mode. Tethered shooting refers to capturing images from a camera connected to the computer over USB. There are several practical reasons for tethering, including the ability to see larger and higher-quality output than can be displayed on a camera's LCD screen. But tethered shooting can also be helpful when setting up complicated studio shots, such as time-lapses, macro-photography, or tricky-to-capture phenomena (imagine capturing the arrow-striking-an-apple shot, for instance). Live preview makes setting up and double-checking these carefully-managed shoots far simpler.

Darktable 1.1 has new features under the hood, too. The most prominent is support for GPU-based hardware acceleration, which arrives courtesy of OpenCL. OpenCL support benefits users by speeding up image transformations. When working with high-megapixel raw images, every speed increase is important, because users want to see even minute changes in settings reflected as soon as possible on the screen. The trend in camera-making is to add more megapixels with every release, of course, and these days the low end of the price spectrum offers more pixels than the high end did a few years ago — so time-saving is not the concern of professional photographers alone. Darktable, like GIMP, uses the GEGL library to perform image transformations. As we mentioned in May 2012, GEGL has been slowly but surely adding OpenCL support to its operations in recent years, most recently through the work of developer Victor Oliveira. OpenCL "kernels" are functions which can be executed in parallel on GPUs or CPUs, and they are architecture-independent (unlike, for example, NVIDIA's CUDA). Thus, systems with a supported GPU automatically get GPU acceleration, but all multi-core CPU systems automatically get multi-threading, too. At the moment, the proprietary graphics drivers from NVIDIA and ATI offer the best support for OpenCL, although the Nouveau driver project is making progress on its own.

The list of changes since the Darktable 0.9 series includes other features, too, but the most significant for the average user will no doubt be the improved user interface. There are still quirks, but the team has done an excellent job of fixing the biggest usability blockers — and doing so without sacrificing the design aesthetic that earlier releases established. Darktable's approach to image editing has always been different from the other open source raw converters. When its own user interface does not get in the way, it makes a much stronger case — and, more importantly, it lets the user experiment with the unique features, and stumble across interesting effects.

Comments (9 posted)

2012 Linux and free software timeline - Q2

By Michael Kerrisk
December 5, 2012

Here is LWN's fifteenth annual timeline of significant events in the Linux and free software world. We will be breaking the timeline up into quarters, and this is our report on April-June 2012. Timelines for the remaining quarters of the year will appear in the coming weeks.

This is version 0.8 of the 2012 timeline. There are almost certainly some errors or omissions; if you find any, please send them to

LWN subscribers have paid for the development of this timeline, along with previous timelines and the weekly editions. If you like what you see here, or elsewhere on the site, please consider subscribing to LWN.

If you'd like to look further back in time, our timeline index page has links to the previous timelines and some other retrospective articles going all the way back to 1998.


The 2012 Linux Storage, Filesystem, and Memory Management Summit is held in San Francisco, April 1-2 (LWN coverage: day 1 and day 2).

Debian joins the Open Source Initiative as an affiliate (announcement).

The udev maintainer announces that the udev and systemd projects will merge, noting that it will still be possible to run udev on a system that is not using systemd (announcement).

I think one of the things that makes Debian off-putting and unwelcoming is that we're a little *too* obsessed with criticizing everyone's ideas, and what some people see as "healthy discussion" other people see as "hurtful flamewars over bike shed colors."

-- Russ Allbery

Yukihiro "Matz" Matsumoto, creator of the Ruby language, wins the 2011 Free Software Foundation award for the advancement of free software (announcement).

Red Hat celebrates becoming the first open source company to turn over one billion dollars in a fiscal year with a US$100,000 donation to open source projects (LWN blurb).

The Kubuntu project acquires a new sponsor, as Blue Systems hires two former Kubuntu developers away from Canonical (LWN blurb and article).

The 2012 Linux Foundation Collaboration Summit take place in San Francisco, April 3-5 (LWN coverage: Trademarks for free software projects; The kernel panel; X and Wayland; The Linux System Definition; The future of GLIBC; LLVM and Linux).

We don't need a system to help us ignore bug reports; our existing process handles that with admirable efficiency.

-- Robert Haas

Maintenance of the Linux 2.4 kernel comes to an end, eight years after the release of Linux 2.6.0 (announcement).

PostGIS 2.0.0 is released (announcement).

The Samba team announce a fix for a remote code execution vulnerability (LWN blurb).

The 2012 Linux Audio Conference takes place in Palo Alto, California, April 12-15 (LWN coverage).

Stefano Zacchiroli is re-elected for a third term as leader of the Debian Project (announcement).

A couple of times I've said "It looks like you could use some help. Would you like me to co-maintain with you?" and have generally gotten a positive response. If it's put in terms of "Looks like you're busy, I can help" and not "You suck and should be fired so I can take over" people seem to be pretty open to it.

-- Scott Kitterman

MythTV 0.25 is released (LWN article).

FreeBSD 8.3 is released (announcement, release notes). [Calligra logo]

Calligra 2.4 is released (LWN blurb and article).

Nathan Willis joins LWN as an editor (LWN article).

gitolite v3.0 is released (announcement). [OpenSSH logo]

OpenSSH 6.0 released (announcement).


Geary 0.1 is released (LWN article on this GNOME-based email client).

The Defensive Patent License is released (LWN article).

OpenBSD 5.1 is released (announcement).

An Apple programmer, apparently by accident, left a debug flag in the most recent version of the Mac OS X operating system. In specific configurations, applying OS X Lion update 10.7.3 turns on a system-wide debug log file that contains the login passwords of every user who has logged in since the update was applied. The passwords are stored in clear text.

-- Emil Protalinski

The Tizen project announces the 1.0 ("Larkspur") release of its SDK and platform source code (LWN blurb and article). [Ubuntu logo]

Ubuntu 12.04 LTS "Precise Pangolin" released (announcement).

Yocto Project 1.2 is released (announcement). [Xfce logo]

Xfce 4.10 is released (LWN blurb).

The Libre Graphics Meeting 2012 is held in Austria, Vienna, May 2-5 (LWN coverage: Inkscape quietly evolves into a development platform; GIMP's new release, new über-core, and future; Unusual typography).

The inaugural Tizen conference takes place in San Francisco, 7-9 May (LWN coverage: Pitching HTML5 as a development framework).

Dell announces Project Sputnik, which is aimed at creating a commercial, Linux-based developer laptop (LWN blurb). [Apache OpenOffice logo]

Apache OpenOffice 3.4 is released (LWN blurb, pointer to an earlier timeline of the work on the project, and an earlier article looking at progress of the project).

The GNU nPth project makes a first release of its GNU portable threads library (announcement).

Open Build Service version 2.3 released (announcement).

GIMP 2.8 is released (release notes, LWN blurb and article previewing the release). [GIMP logo]

The Document Foundation announces a certification program "to foster the provision of professional services around LibreOffice" (announcement).

Red Hat Enterprise Linux turns 10 (press release).

Enough data has come in to satisfy me that with all the improvements in Linux over the last year, and with BQL, codel and fq_codel, that we've won a major battle in the war against bufferbloat

-- Dave Täht

ConnMan 1.0 is released (LWN blurb).

Kdenlive 0.9 is released (announcement).

PowerTOP v2.0 is released (LWN blurb).

PulseAudio 2.0 is released (announcement).

PGCon 2012 is held in Ottawa, Canada, May 17-18 (LWN coverage).

Mandriva SA announces it will return control of the distribution back "to the community". However, the Mageia community distribution that earlier forked from Mandriva declines to work with Mandriva's community effort (announcement, LWN article on the announcement and an earlier article on the status of Mandriva).

When I helped to develop the open standards that computers use to communicate with one another across the Net, I hoped for but could not predict how it would blossom and how much human ingenuity it would unleash. What secret sauce powered its success? The Net prospered precisely because governments — for the most part — allowed the Internet to grow organically, with civil society, academia, private sector and voluntary standards bodies collaborating on development, operation and governance.

-- Vint Cerf

The printerd project is announced (LWN article).

Linux 3.4 is released (announcement; KernelNewbies summary; LWN merge window summaries part 1, part 2, and part 3; LWN development statistics article).

Mageia 2 is released (announcement and LWN article).

LLVM 3.1 is released (announcement, release notes)

Nmap version 6 is released (announcement).

ownCloud 4 is released (LWN blurb).

Perl 5.16.0 is released (announcement and LWN article).

The jury in Oracle v. Google finds that Google did not infringe any of Oracle's patents (LWN blurb and earlier article on the case, Groklaw follow-up).

Simon Phipps becomes president of the Open Source Initiative (The H article).

The LibreOffice project embarks on a project to rebase and relicense the LibreOffice source code (LWN article).

I couldn't have told you the first thing about Java before this problem. I have done, and still do, a significant amount of programming in other languages. I've written blocks of code like rangeCheck a hundred times before. I could do it, you could do it. The idea that someone would copy that when they could do it themselves just as fast, it was an accident. There's no way you could say that was speeding them along to the marketplace. You're one of the best lawyers in America, how could you even make that kind of argument?

-- Judge Alsup (Oracle v. Google) has a clue

The Software Freedom Conservancy announces that it is expanding its license compliance efforts after signing up multiple Linux kernel and Samba developers whose copyrights can be used in license compliance efforts (article). [Fedora logo]

Fedora 17 is released (announcement).

GCC explorer is released (LWN blurb).

RPM 4.10 released (LWN blurb).

systemd 183 is released; this release merges the udev and systemd projects (announcement).

The Linux Foundation announces the existence of the FOSS Bar Code Tracker, a tool for tracking free and open source software components (announcement).

In the Oracle v. Google suit, Judge Alsup rules that the Java APIs are not copyrightable (LWN blurb).


Managing a volunteer open source project is a lot like herding kittens, except the kittens randomly appear and disappear because they have day jobs.

-- Matt Mackall

Obnam 1.0 is released (LWN blurb and article on this backup system).

LinuxCon Japan is held in Yokohama, June 6-8 (videos; LWN coverage: Making kernel developers less grumpy; OpenRelief launches; One zImage to rule them all; Advice for new kernel hackers; The business of contribution).

From the tone of the hearing, and the language of the House resolution, we are being asked to believe that "the position of the United States Government has been and is to advocate for the flow of information free from government control."

If only it were true. The reality is that Congress increasingly has its paws all over the Internet. Lawmakers and regulators are busier than ever trying to expand the horizons of cyber-control across the board: copyright mandates, cybersecurity rules, privacy regulations, speech controls, and much more.

-- Jerry Brito and Adam Thierer

Debian accepts a diversity statement (announcement).

Linus Torvalds co-wins the Millennium Technology Prize (BBC report).

The Apple versus Google-owned Motorola patent litigation takes a surprising turn as Judge Richard Posner dismisses the case, calling the patent system "dysfunctional" (GigaOm article). [Emacs logo]

Emacs 24.1 is released (announcement).

MPlayer 1.1 is released (LWN blurb).

X11R7.7 is released (announcement and LWN article). [X.Org logo]

SystemTap 1.8 is released (announcement).

Ulogd 2.0.0 is released (announcement).

The Electronic Frontier Foundation announces the Defend Innovation patent reform project (press release).

The Fedora and Ubuntu distributions outline their plans for dealing with UEFI secure boot (LWN article on the Fedora plan and the Ubuntu plan). [Red Hat logo]

Red Hat Enterprise Linux 6.3 is released (LWN blurb, release notes).

Grub 2.0.0 is released (announcement).

Documentation is the sort of thing that will never be great unless someone from outside contributes it (since the developers can never remember which parts are hard to understand).

-- Avery Pennarun

The GNU C library (glibc) version 2.16 is released (announcement).

Many Linux servers misbehave as a result of the leap second added at the end of the month (LWN article).

Comments (1 posted)

Page editor: Jonathan Corbet


Picking a MAC address for a FreedomBox

By Jonathan Corbet
December 5, 2012
The FreedomBox project is working toward the creation of an inexpensive, in-home device that can be used for secure and private communications. The initial plan is to create a version of the Debian distribution that can be installed on a device like the DreamPlug; the resulting configuration should "just work" for nontechnical users in potentially hostile situations. The project has many challenges to overcome, one of which — the choice of MAC address for the network interface — shows how tricky this problem space can be.

An interface's MAC address is a unique number identifying the interface to any other devices it may communicate directly with. Ethernet-style MAC addresses are six-byte quantities; half of those bytes identify the manufacturer while the other half are meant to be a unique serial number. The MAC address for the Ethernet interface on the system where this article being typed is:


This MAC address identifies the relevant system as having been manufactured by Dell. If Dell has done its job properly (and there is no evidence to the contrary), no other Ethernet interface on the planet should have that same MAC address.

FreedomBox developer Nick Daly recently started pondering the question of how a FreedomBox should set its MAC address. The hardware will come with an address provided by the manufacturer, of course, but that address can be changed by the kernel and there may well be good reasons for doing so. Many of those were outlined in this lengthy message from John Gilmore, which is well worth reading in its entirety; it forms the basis of this summary.

One obvious problem is that a static MAC address is a unique number identifying a particular system. Most interfaces never operate with anything but the vendor-supplied address; if a hostile party learns that address, they can quickly identify the system it belongs to. So, while a FreedomBox device might move around, a suitably informed observer will always know which device it is. That allows the correlation of activities over time and the monitoring of specific devices.

Current technologies make things worse. Quoting John:

Apple iPhones record the MAC addresses that are nearby, report these to Apple, and Apple uses them to return a physical position fix. This is used to more rapidly cause the GPS algorithm to converge on a position, and also used when GPS isn't working. The phones often report their GPS position and any nearby MAC addresses back to Apple servers... It's easy for hackers to query that database of MAC addresses and locations, by pretending to be an iPhone seeking its location.

In other words, a hostile entity might not have to drive around a city in an attempt to detect a device with a specific MAC address; instead, it is just a matter of asking Apple, which has a widespread surveillance network in place and can simply say where that device is to be found. Similar information is maintained by other parties — Google, for example.

John also pointed out that it is often trivial to determine which IP address is assigned to a device; it is often just a matter of sending a DNS query to the MAC address of interest. That can enable the identification of the location from which specific network activity has been generated.

Finally, there is the matter of that manufacturer identification number found in every MAC address. If FreedomBox becomes a widely used and effective system, certain authorities might develop a strong interest in knowing where DreamPlug systems are to be found. The identifying information found in the MAC address makes that identification a relatively simple task. Turning on a DreamPlug could be a way of painting a target on a specific location — not the sort of dream the owner may have been looking for.

The obvious conclusion is that FreedomBox systems should not normally run with the default MAC address provided by the vendor. They should, instead, generate a new address, and that address should be changed frequently. Fortunately, much of this is easy to do; any even remotely contemporary hardware will allow the host system to provide a new MAC address, and the data link layer (and above) protocols are pretty good about responding to MAC address changes. So there is no obvious technical barrier preventing frequent changing of a system's MAC address.

But there is still the question of what that address should be. Nick had suggested defaulting to 00:00:00:00:00:00 by default, a choice that would clearly prevent the identification of specific FreedomBoxes. But there are problems with that choice, starting with the fact that confusion would result as soon as two FreedomBoxes appeared on the same network. So something a little smarter is needed.

One obvious possibility is to simply generate a six-byte random number and use that. Care would have to be taken to avoid MAC address collisions on any given net, but that is not a particularly hard problem to solve. There are also the usual issues with having enough entropy available to generate a proper random number at boot time; without an adequate level of care, that random address might be far less random than people expect. Once again, that is a problem that should be amenable to a proper solution.

But, as John pointed out, there is another problem: real-world MAC addresses follow a specific pattern; a random address, being unlikely to fit that pattern, would probably stand out like a neon sign to anybody who is looking for it. To be convincing, a system-chosen MAC address cannot be completely random. It should have a recognized manufacturer number, preferably a manufacturer that actually makes contemporary wireless network interfaces. The serial number also needs to fit into a range that was actually shipped by that manufacturer. In other words, a random MAC address will only blend in if it makes the device look like some other random piece of real-world hardware.

These problems are all tractable, but the solution requires a great deal of due care if it is not to expose its users to unwanted consequences. Indeed, the whole system must be designed and implemented with that level of care; that is part of why the FreedomBox has not come to fruition as quickly as many would have liked. Privacy is a surprisingly difficult problem, with many pitfalls for those who try for a quick solution.

Comments (36 posted)

Brief items

Security quotes of the week

My computer was arrested before me
-- Syrian protester Dr. Taymour Karim

… the FBI has access to … the emails of virtually everybody in the country.
-- NSA whistle-blower William Binney, interviewed on

Jellyfish are interesting to trojan writers. Deep at their heart they are colony creatures, with stealth capabilities. What could be more pertinent to those of us who feed on the world's information like opportunistic predators?

Right now my feeling is that the world has been lucky because most of the malicious software on the internet has been, at worst, a rapscallion, or a scofflaw, or perhaps a ne'er-do-well. And there are modern networks who fare pretty well against that kind of adversary. But longer term, there's going to be malware that resembles science fiction <>...or maybe jellyfish? :>

-- Dave Aitel

These companies own us, so they can sell us off -- again, like serfs -- to rival lords... or turn us in to the authorities.
-- Bruce Schneier

Comments (1 posted)

New vulnerabilities

apache2: denial of service

Package(s):apache2 CVE #(s):CVE-2012-4557
Created:November 30, 2012 Updated:December 5, 2012

From the Debian advisory:

A flaw was found when mod_proxy_ajp connects to a backend server that takes too long to respond. Given a specific configuration, a remote attacker could send certain requests, putting a backend server into an error state until the retry timeout expired. This could lead to a temporary denial of service.

Ubuntu USN-1765-1 apache2 2013-03-18
CentOS CESA-2013:0512 httpd 2013-03-09
Scientific Linux SL-http-20130228 httpd 2013-02-28
Oracle ELSA-2013-0512 httpd 2013-02-25
Red Hat RHSA-2013:0512-02 httpd 2013-02-21
openSUSE openSUSE-SU-2013:0248-1 apache2 2013-02-05
openSUSE openSUSE-SU-2013:0243-1 apache2 2013-02-05
Debian DSA-2579-1 apache2 2012-11-30

Comments (none posted)

claws-mail: user credential leak

Package(s):claws-mail CVE #(s):CVE-2012-5527
Created:December 3, 2012 Updated:January 18, 2013
Description: From the Red Hat bugzilla:

A security flaw was found in the way vCalendar plug-in of Claws Mail displayed user credential information in the system tray display when using https scheme. A local attacker could use this flaw to obtain user credentials (username and password) used for connection to remote point.

Mageia MGASA-2013-0014 claws-mail-plugins 2013-01-18
Fedora FEDORA-2012-18559 claws-mail-plugins 2012-12-03
Fedora FEDORA-2012-18558 claws-mail-plugins 2012-12-03
Fedora FEDORA-2012-18559 claws-mail 2012-12-03
Fedora FEDORA-2012-18558 claws-mail 2012-12-03

Comments (none posted)

firefox: multiple vulnerabilities

Package(s):Mozilla Firefox CVE #(s):CVE-2012-5837 CVE-2012-4206
Created:November 29, 2012 Updated:December 5, 2012

From the Mozilla advisory:

MFSA 2012-102 / CVE-2012-5837: Security researcher Masato Kinugawa reported that when script is entered into the Developer Toolbar, it runs in a chrome privileged context. This allows for arbitrary code execution or cross-site scripting (XSS) if a user can be convinced to paste malicious code into the Developer Toolbar.

MFSA 2012-98 / CVE-2012-4206: Security researcher Robert Kugler reported that when a specifically named DLL file on a Windows computer is placed in the default downloads directory with the Firefox installer, the Firefox installer will load this DLL when it is launched. In circumstances where the installer is run by an administrator privileged account, this allows for the downloaded DLL file to be run with administrator privileges. This can lead to arbitrary code execution from a privileged account.

openSUSE openSUSE-SU-2014:1100-1 Firefox 2014-09-09
openSUSE openSUSE-SU-2013:0175-1 mozilla 2013-01-23
Gentoo 201301-01 firefox 2013-01-07
SUSE SUSE-SU-2012:1592-1 Mozilla Firefox 2012-11-29
Ubuntu USN-1638-3 firefox 2012-12-03

Comments (none posted)

kernel: information leak

Package(s):kernel CVE #(s):CVE-2012-4530
Created:December 3, 2012 Updated:January 15, 2013
Description: From the Red Hat bugzilla:

A memory disclosure flaw has been found in the way binfmt_script load_script() function handled excessive recursions. An unprivileged local user could use this flaw to leak kernel memory.

Oracle ELSA-2013-1645 kernel 2013-11-26
SUSE SUSE-SU-2013:0674-1 Linux kernel 2013-04-13
Red Hat RHSA-2013:0566-01 kernel-rt 2013-03-06
openSUSE openSUSE-SU-2013:0396-1 kernel 2013-03-05
Oracle ELSA-2013-2507 kernel 2013-02-28
Oracle ELSA-2013-2504 kernel 2013-02-07
Oracle ELSA-2013-2503 kernel 2013-02-07
Scientific Linux SL-kern-20130206 kernel 2013-02-06
Oracle ELSA-2013-0223 kernel 2013-02-06
CentOS CESA-2013:0223 kernel 2013-02-06
Red Hat RHSA-2013:0223-01 kernel 2013-02-05
Ubuntu USN-1704-2 Quantal kernel 2013-02-01
Ubuntu USN-1698-2 OMAP4 kernel 2013-02-01
Ubuntu USN-1696-2 kernel 2013-02-01
Ubuntu USN-1700-2 OMAP4 kernel 2013-02-01
Ubuntu USN-1699-2 kernel 2013-02-01
Mageia MGASA-2013-0016 kernel-rt 2013-01-24
Ubuntu USN-1704-1 kernel 2013-01-22
Ubuntu USN-1699-1 linux 2013-01-17
Ubuntu USN-1696-1 linux 2013-01-17
Ubuntu USN-1691-1 linux-ti-omap4 2013-01-15
Ubuntu USN-1689-1 linux 2013-01-15
Ubuntu USN-1688-1 linux-lts-backport-oneiric 2013-01-15
Ubuntu USN-1684-1 linux-ec2 2013-01-10
Ubuntu USN-1683-1 linux 2013-01-10
Ubuntu USN-1700-1 linux-ti-omap4 2013-01-17
Ubuntu USN-1698-1 linux-ti-omap4 2013-01-17
Fedora FEDORA-2012-20240 kernel 2012-12-18
Fedora FEDORA-2012-19337 kernel 2012-12-01

Comments (none posted)

kernel: privilege escalation

Package(s):kernel CVE #(s):CVE-2012-5513
Created:December 5, 2012 Updated:December 24, 2012
Description: From the Red Hat advisory:

A flaw in the way the Xen hypervisor implementation range checked guest provided addresses in the XENMEM_exchange hypercall could allow a malicious, para-virtualized guest administrator to crash the hypervisor or, potentially, escalate their privileges, allowing them to execute arbitrary code at the hypervisor level.

SUSE SUSE-SU-2014:0470-1 Xen 2014-04-01
SUSE SUSE-SU-2014:0446-1 Xen 2014-03-25
Gentoo 201309-24 xen 2013-09-27
openSUSE openSUSE-SU-2013:0637-1 xen 2013-04-08
openSUSE openSUSE-SU-2013:0636-1 xen 2013-04-08
openSUSE openSUSE-SU-2013:0133-1 xen 2013-01-23
openSUSE openSUSE-SU-2012:1687-1 xen 2012-12-23
openSUSE openSUSE-SU-2012:1685-1 xen 2012-12-23
Debian DSA-2582-1 xen 2012-12-07
Oracle ELSA-2012-1540 kernel 2012-12-05
Scientific Linux SL-kern-20121206 kernel 2012-12-06
SUSE SUSE-SU-2012:1615-1 Xen 2012-12-06
SUSE SUSE-SU-2012:1606-1 Xen 2012-12-04
CentOS CESA-2012:1540 kernel 2012-12-05
Red Hat RHSA-2012:1540-01 kernel 2012-12-04

Comments (none posted)

keystone: multiple vulnerabilities

Package(s):keystone CVE #(s):CVE-2012-5571 CVE-2012-5563
Created:November 29, 2012 Updated:December 11, 2012

From the Ubuntu advisory:

Vijaya Erukala discovered that Keystone did not properly invalidate EC2-style credentials such that if credentials were removed from a tenant, an authenticated and authorized user using those credentials may still be allowed access beyond the account owner's expectations. (CVE-2012-5571)

It was discovered that Keystone did not properly implement token expiration. A remote attacker could use this to continue to access an account that is disabled or has a changed password. This issue was previously fixed as CVE-2012-3426 but was reintroduced in Ubuntu 12.10. (CVE-2012-5563)

Fedora FEDORA-2012-19341 openstack-keystone 2012-12-11
Red Hat RHSA-2012:1557-01 openstack-keystone 2012-12-10
Red Hat RHSA-2012:1556-01 openstack-keystone 2012-12-10
Ubuntu USN-1641-1 keystone 2012-11-28

Comments (none posted)

libxml2: code execution

Package(s):libxml2 CVE #(s):CVE-2012-5134
Created:November 30, 2012 Updated:March 1, 2013

From the Red hat advisory:

A heap-based buffer underflow flaw was found in the way libxml2 decoded certain entities. A remote attacker could provide a specially-crafted XML file that, when opened in an application linked against libxml2, would cause the application to crash or, potentially, execute arbitrary code with the privileges of the user running the application. (CVE-2012-5134)

Gentoo 201311-06 libxml2 2013-11-10
SUSE SUSE-SU-2013:1627-1 libxml2 2013-11-04
SUSE SUSE-SU-2013:1625-1 libxml2 2013-11-04
Oracle ELSA-2013-0581 libxml2 2013-03-01
Mandriva MDVSA-2013:056 libxml2 2013-04-08
Scientific Linux SL-ming-20130201 mingw32-libxml2 2013-02-01
Oracle ELSA-2013-0217 mingw32-libxml2 2013-02-01
CentOS CESA-2013:0217 mingw32-libxml2 2013-02-01
Red Hat RHSA-2013:0217-01 mingw32-libxml2 2013-01-31
openSUSE openSUSE-SU-2013:0178-1 libxml2 2013-01-23
openSUSE openSUSE-SU-2012:1647-1 libxml2 2012-12-17
SUSE SUSE-SU-2012:1636-1 libxml2 2012-12-12
Slackware SSA:2012-341-03 libxml2 2012-12-06
Debian DSA-2580-1 libxml2 2012-12-02
Scientific Linux SL-libx-20121129 libxml2 2012-11-29
Oracle ELSA-2012-1512 libxml2 2012-11-29
CentOS CESA-2012:1512 libxml2 2012-11-29
CentOS CESA-2012:1512 libxml2 2012-11-29
Red Hat RHSA-2012:1512-01 libxml2 2012-11-29
Oracle ELSA-2012-1512 libxml2 2012-11-29
Ubuntu USN-1656-1 libxml2 2012-12-05
Mandriva MDVSA-2012:176 libxml2 2012-12-02
Mageia MGASA-2012-0350 libxml2 2012-11-30

Comments (none posted)

lynx: multiple vulnerabilities

Package(s):lynx-cur CVE #(s):CVE-2010-2810 CVE-2012-5821
Created:November 30, 2012 Updated:April 10, 2013

From the Ubuntu advisory:

Dan Rosenberg discovered a heap-based buffer overflow in Lynx. If a user were tricked into opening a specially crafted page, a remote attacker could cause a denial of service via application crash, or possibly execute arbitrary code as the user invoking the program. This issue only affected Ubuntu 10.04 LTS. (CVE-2010-2810)

It was discovered that Lynx did not properly verify that an HTTPS certificate was signed by a trusted certificate authority. This could allow an attacker to perform a "man in the middle" (MITM) attack which would make the user believe their connection is secure, but is actually being monitored. This update changes the behavior of Lynx such that self-signed certificates no longer validate. Users requiring the previous behavior can use the 'FORCE_SSL_PROMPT' option in lynx.cfg. (CVE-2012-5821)

Mandriva MDVSA-2013:101 lynx 2013-04-10
Ubuntu USN-1642-1 lynx-cur 2012-11-29
Mageia MGASA-2012-0351 lynx 2012-11-30

Comments (none posted)

mod_security: multipart/invalid part ruleset bypass

Package(s):mod_security CVE #(s):CVE-2012-4528
Created:December 3, 2012 Updated:April 5, 2013
Description: From the Red Hat bugzilla:

ModSecurity <= 2.6.8 is vulnerable to multipart/invalid part ruleset bypass, this was fixed in 2.7.0 (released on2012-10-16)

openSUSE openSUSE-SU-2013:1342-1 apache2-mod_security2 2013-08-14
openSUSE openSUSE-SU-2013:1336-1 apache2-mod_security2 2013-08-14
openSUSE openSUSE-SU-2013:1331-1 apache2-mod_security2 2013-08-14
Mandriva MDVSA-2013:029 apache-mod_security 2013-04-05
Mandriva MDVSA-2013:016 apache-mod_security 2013-04-04
Mageia MGASA-2012-0371 apache-mod_security 2012-12-31
Mandriva MDVSA-2012:183 apache-mod_security 2012-12-23
Mandriva MDVSA-2012:182 apache-mod_security 2012-12-23
Fedora FEDORA-2012-18315 mod_security_crs 2012-12-01
Fedora FEDORA-2012-18315 mod_security 2012-12-01

Comments (none posted)

mysql: code execution

Package(s):mysql-5.1 CVE #(s):CVE-2012-5611
Created:December 4, 2012 Updated:September 9, 2013
Description: From the CVE entry:

Stack-based buffer overflow in MySQL 5.5.19, 5.1.53, and possibly other versions, and MariaDB 5.5.2.x before 5.5.28a, 5.3.x before 5.3.11, 5.2.x before 5.2.13 and 5.1.x before 5.1.66, allows remote authenticated users to execute arbitrary code via a long argument to the GRANT FILE command.

Gentoo 201308-06 mysql 2013-08-29
openSUSE openSUSE-SU-2013:1412-1 mariadb 2013-09-08
Gentoo GLSA 201308-06:02 mysql 2013-08-30
Mandriva MDVSA-2013:102 mariadb 2013-04-10
SUSE SUSE-SU-2013:0262-1 MySQL 2013-02-09
Mageia MGASA-2013-0046 mariadb 2013-02-09
Mandriva MDVSA-2013:008 mysql 2013-02-06
Mandriva MDVSA-2013:007 mysql 2013-02-05
Scientific Linux SL-mysq-20130123 mysql 2013-01-23
Oracle ELSA-2013-0180 mysql 2013-01-22
CentOS CESA-2013:0180 mysql 2013-01-22
Ubuntu USN-1703-1 mysql-5.1, mysql-5.5, mysql-dfsg-5.1 2013-01-22
openSUSE openSUSE-SU-2013:0013-1 mysql-community-server 2013-01-07
openSUSE openSUSE-SU-2013:0014-1 mariadb 2013-01-07
openSUSE openSUSE-SU-2013:0011-1 mariadb 2013-01-07
openSUSE openSUSE-SU-2013:0156-1 mariadb 2013-01-23
openSUSE openSUSE-SU-2013:0135-1 mysql-community-server 2013-01-23
Red Hat RHSA-2013:0180-01 mysql 2013-01-22
Fedora FEDORA-2012-19823 mysql 2012-12-21
Fedora FEDORA-2012-19833 mysql 2012-12-15
Ubuntu USN-1658-1 mysql-5.1, mysql-5.5, mysql-dfsg-5.1 2012-12-10
Oracle ELSA-2012-1551 mysql 2012-12-07
Mandriva MDVSA-2012:178 mysql 2012-12-09
Scientific Linux SL-mysq-20121207 mysql 2012-12-07
CentOS CESA-2012:1551 mysql 2012-12-07
Red Hat RHSA-2012:1551-01 mysql 2012-12-07
Mageia MGASA-2012-0349 mysql 2012-12-07
Debian DSA-2581-1 mysql-5.1 2012-12-04

Comments (none posted)

perl: code execution

Package(s):perl CVE #(s):CVE-2012-5195
Created:November 30, 2012 Updated:January 28, 2013

From the Ubuntu advisory:

It was discovered that Perl's 'x' string repeat operator is vulnerable to a heap-based buffer overflow. An attacker could use this to execute arbitrary code. (CVE-2012-5195)

Gentoo 201401-11 perl 2014-01-19
Mandriva MDVSA-2013:113 perl 2013-04-10
Scientific Linux SL-perl-20130327 perl 2013-03-27
Oracle ELSA-2013-0685 perl 2013-03-27
Oracle ELSA-2013-0685 perl 2013-03-26
CentOS CESA-2013:0685 perl 2013-03-26
CentOS CESA-2013:0685 perl 2013-03-26
Red Hat RHSA-2013:0685-01 perl 2013-03-26
Mandriva MDVSA-2013:005 perl 2013-01-28
Debian DSA-2586-1 perl 2012-12-11
Mageia MGASA-2012-0352 perl 2012-12-07
Ubuntu USN-1643-1 perl 2012-11-29

Comments (none posted)

Page editor: Michael Kerrisk

Kernel development

Brief items

Kernel release status

The current development kernel is 3.7-rc8, reluctantly released by Linus on December 3. "I really didn't want it to come to this, but I was uncomfortable doing the 3.7 release yesterday due to last-minute issues, and decided to sleep on it. And today, I ended up even *less* comfortable about it due to the resurrection of a kswapd issue, so I decided that I'm going to do another -rc after all." As he points out, that implies that the 3.8 merge window will run close to the holidays.

Stable updates: 3.6.9, 3.4.21 and 3.0.54 were released on December 3. Meanwhile, 3.2.35 is in the review process; its release can be expected at any time.

Comments (none posted)

Quotes of the week

I'm all in favour of "whence", which is indeed the name of that lseek argument - since mediaeval times I believe.

It's good to have words like that in the kernel source: while you're in the mood, please see if you can find good homes for "whither" and "thrice" and "widdershins".

Hugh Dickins

Took vacation last week, spent most of it doing userspace coding. It was joyous.
Rusty Russell

If yes, yet again this illustrates why the use of atomic types leads people down the path of believing that their code somehow becomes magically safe through the use of this smoke-screen. IMHO, every use of atomic_t must be questioned and carefully analysed before it gets into the kernel - many are buggy through assumptions that atomic_t buys you something magic.
Russell King

Comments (14 posted)


By Jonathan Corbet
December 5, 2012
Last week's edition included an article on the addition of the FALLOC_FL_NO_HIDE_STALE flag to the fallocate() system call. Some developers, objecting to the patch and the way it got into the kernel, had called for it to be reverted before the 3.7 release went final. At the time, Linus had not made any remarks in the discussion or indicated whether he would accept the revert.

That situation changed after Linus was prompted by Martin Steigerwald. His response was clear enough:

If you want something reverted, you show me the *technical* reason for it. Not the "ooh, I'm so annoyed by how this was done" reason for it.

And if your little feelings got hurt, get your mommy to tuck you in, don't email me about it. Because I'm not exactly known for my deep emotional understanding and supportive personality, am I?

There were some technical reasons offered in the discussion, along with the more general process-oriented complaints. But it seems clear that Linus has not found that discussion convincing. So, in the absence of a surprise from somewhere, it seems that the new fallocate() flag will remain for the 3.7 release, at which point it will become part of the kernel's user-space ABI.

Comments (18 posted)

Kernel development news

Stepping closer to practical containers: "syslog" namespaces

By Michael Kerrisk
December 5, 2012

The abstract goal of containers is, in effect, to provide a group of processes with the illusion that that they are the only processes on the system. When fully implemented, this feature has the potential to realize many practical benefits, such as light-weight virtualization and checkpoint/restore.

In order to give the processes in a container the illusion that there are no other processes on the system, various global system resources must be wrapped in abstractions that make it appear that each container has its own instance of the resources. This has been achieved by the addition of "namespaces" for a number of global resources. Each namespace provides an isolated view of a particular global resource to the set of processes that are members of that namespace.

Step by step, more and more global resources have been wrapped in namespaces, and before we look at another step in this path it's worth reviewing the progress to date.

Namespaces so far

The first step in the journey was mount namespaces, which can be used to provide a group of processes with a private view of the mount points that make up the filesystem hierarchy. Mount namespaces first appeared in the mainline kernel in 2002, with the release of Linux 2.4.19. The clone() flag used to create mount namespaces was given the rather generic name CLONE_NEWNS for "new namespace", implying that no one was then really considering the possibility that there might be other kinds of namespaces; at that time, of course, containers were no more than a gleam in the eyes of some developers.

However, as the concept of containers took hold, a number of other namespaces have followed. Network namespaces were added to provide a group of processes with a private view of the network (network devices, IP addresses, IP routing tables, port number space, and so on). PID namespaces isolated the global "PID number space" resource, so that processes in separate PID namespaces can have the same PIDs—in particular, each namespace can have its own 'init' (PID 1), the "ancestor of all processes". PID namespaces also allow techniques such as freezing the processes in a container and then restoring them on another system while maintaining the same PIDs.

Several other global resources have likewise been wrapped in namespaces, so that there are also IPC namespaces (initially implemented to isolate System V IPC identifiers and later to isolate instances of the virtual filesystems used in the implementation of POSIX message queues) and UTS namespaces (which wrap the nodename and domainname identifiers returned by uname(2)). Work on one of the more complex namespaces, user namespaces, was started in about Linux 2.6.23 and seems to be edging towards completion. When complete, user namespaces will allow per-namespace mappings of user and group IDs, so that, for example, it will be possible for a process to be root inside a container without having root privileges in the system as a whole.

Of course, a Linux system has a large number of global resources, each of which could conceivably be wrapped in a namespace. At the more extreme end, for example, even a resource such as the system time could be wrapped, so that different containers could maintain different concepts of the time. (A time namespace was once proposed, but the implementation was not merged.) The trick is to determine the minimum set of resources that need to be wrapped for the practical implementation of containers. (Of course, this "minimum set" may well grow over time, as people develop new uses for containers.) A related question is how those wrappings should be grouped so as to avoid an explosion of namespaces that would increase application complexity. So, for example, System V IPC and POSIX message queues could conceivably have been wrapped in different namespaces, but the kernel developers concluded that it makes practical sense to group them in a single "IPC" namespace.

The global kernel log problem

What is necessary for the practical implementation of containers sometimes only becomes clear when one starts doing, well, practical things. Thus, it was that in early 2010 Jean-Marc Pigeon reported that he had written a small utility to build containers using the clone() system call that worked fine, except that "HOST and all containers share the SAME /proc/kmsg, meaning kernel syslog information are scrambled (useless)".

What Jean-Marc was discovering is that the kernel log is one of the global resources that is not wrapped in a namespace. He went on to note another ill-effect: "I have in iptables, reject packet logging on the HOST, [but as soon as] rsyslog is started on one container, I can't see my reject packet log any more." In other words, starting a syslog daemon on the host or any container sucks up all of the kernel log messages produced on the host or in any container. The point here about iptables is particularly relevant: the inability to isolate kernel log messages from iptables is a significant practical problem when trying to employ the network namespaces facility that the kernel already provides.

In response to Jean-Marc's question about how the problem could be fixed, Serge Hallyn replied:

Well, the results of do_syslog() should be containerized. Kernel messages (oopses for instance) should always go to the initial container. Shouldn't be hard to do, but the question is what do we tie it to? User namespace? Network namespace? … I'm tempted to say userns makes the most sense - if you start a new userns you likely always want private syslog, whereas with netns and pidns you may not.

do_syslog() is the kernel function that encapsulates the main logic of the syslog(2) system call. That system call retrieves messages from the kernel log ring buffer (and performs a range of control operations on the log buffer) that is populated by messages created using the kernel's printk() function. Thus, though discussions on this topic have tended to use the term "syslog namespace", that is something of a misnomer: what is really meant is wrapping the kernel log resource in a namespace.

To avoid possible confusion, it is probably worth noting that the syslog(2) system call is a quite different thing from the syslog(3) library function, which writes messages to the UNIX domain datagram socket (/dev/log) from which the user-space syslog daemon (rsyslogd or similar) retrieves messages. (Because of this collision of names, the GNU C library exposes the syslog(2) system call under a quite different name: klogctl().) A picture helps clarify things:

[Interrelationship of
logging primitives]

First attempts at a solution

In the event, "containerizing" do_syslog() turned out to be more difficult than Serge thought. His first shot at addressing the problem (a "gross hack" to "provide each user namespace with its own syslog ring buffer") quickly uncovered a further difficulty: the kernel's printk() is sometimes called in contexts where there is no way to determine in which of the per-namespace ring buffers a message should be logged. For example, if the kernel is executing a network interrupt (to process an incoming network packet) and wants to log a message, that message should not be sent to the per-namespace kernel log of the interrupted process. Rather, the message should be sent to the kernel log associated with the network namespace for the network device; however, the kernel data structures provide no way to obtain a reference to that kernel log.

Jean-Marc himself also made an attempt at implementing a solution. However, Serge pointed out that Jean-Marc's patch suffered some of the same problems as his own earlier attempt. Serge went on to describe what he thought would be the correct solution, which would require the creation of a separate syslog namespace. His proposed solution can be paraphrased as follows:

  1. The core of vprintk_emit() (which contains most of implementation of the printk() function) should be moved into a new nsvprintk_emit() function that takes an argument that specifies a syslog namespace.

  2. vprintk_emit() would then become a wrapper around nsvprintk_emit() that specifies the "initial" syslog namespace (i.e., the syslog namespace of the host system).

  3. A namespace-aware version of printk(), called (say) nsprintk(), should be created. That function would take a syslog namespace argument and pass it to nsvprintk_emit().

  4. The kernel log ring buffer should be "containerized" as per Serge's initial patch. Thus each syslog namespace would have its own ring buffer, and syslog(2) would operate on the per-namespace ring buffer of the calling process.

  5. At call sites in the kernel code where it is not appropriate to use the syslog namespace of the currently executing process, calls to printk() should be replaced with calls to nsprintk() that pass a suitable syslog namespace argument.

Although Jean-Marc made a few more efforts to rework his patch in the following weeks, the effort ultimately petered out without much further comment or consensus on a solution. It seems that Serge and other kernel developers realized that the problem was more complex than first thought and they had neither the time to implement a solution themselves nor to help Jean-Marc toward implementing a solution.

The main difficulty lies in the last of the points above, and its solution was not really elaborated in Serge's mail. The kernel data structures and code need to be modified to add suitable hooks to handle the "no current process context problem"—the cases where printk() is called from a context in which the currently executing process can't be used to identify a suitable syslog namespace to which a message should be logged.

Restarting work on a solution

Work in this area then seems to have gone quiet for more than two years, until a few days ago when Serge proposed a new proof-of-concept patch set, pretty much along the lines he described two years earlier. His description of the patch noted that:

The syslog ns is tied to a user namespace. You must create a new user namespace before you can create a new sylog ns. The syslog ns is created through a new command (11) to the __NR_syslog system call.

Once a task enters a new syslog ns, it's "dmesg", "dmesg -c" and /dev/kmsg actions affect only itself, so that user-created syslog messages no longer are confusingly combined in the host's syslog.

In other words, Serge's patch provides isolation for the kernel log by implementing a new dedicated namespace for that purpose (rather than providing the isolation by attaching the implementation to one of the existing namespaces). Each syslog namespace instance would be tied to a particular user namespace.

Normally, new namespaces of each type are created by suitable flags to the clone() system call. Thus, for example, there are clone flags such as CLONE_NEWUTS and CLONE_NEWUSER. However, a while ago, the kernel developers realized that the flag space for clone() was exhausted. (Providing additional flag space was one of the motivations behind the proposal to add an eclone() system call, a proposal that was ultimately unsuccessful.) For this reason, Serge proposed instead to use a new command to the syslog() system call to create syslog namespace instances.

Serge went on to note:

"printk" itself always goes to the initial syslog_ns, and consoles belong only to the initial syslog_ns. However printks relating to a specific network namespace, for instance, can now be targeted to the syslog ns for the user ns which owns the network ns, aiding in debugging in a container.

Serge's patch would solve the "no current process context problem" as follows. As noted above, this case is handled by an nsprintk()-style function that takes an argument (of type struct syslog_ns *) that identifies the syslog namespace to which the log message should be sent. The value for that argument can be obtained via the struct net structure for the network namespace instance: in the current user namespace implementation (git tree), when a network namespace is created using clone(), a pointer to the corresponding user namespace instance of the caller is stored in the net structure. Serge's patch in turn provides a linkage from that user namespace structure to the corresponding syslog namespace.

Eric Biederman, the maintainer of the user namespace git tree, agreed with Serge's overall approach, but queried one particular point:

I am not a fan of how this ties into the user namespace. I would prefer closer or looser ties. The recursive reference count loop where a userns refers to a syslogns and that syslogns refers to the same userns is unpleasant.

In Serge's implementation, the syslog and user namespaces are maintained as separate structures, but, as the recursive pointers between the two namespace structures and the need to create a new user namespace before creating a syslog namespace indicate, instances of each namespace are not truly independent. In Eric's view then, the syslog and user namespace structures should either be more fully decoupled, or they should be much more tightly coupled.

Eric went on later to note that:

There is an argument to be made that syslog messages are the kind of security identifiers like uid, gids, and keys that should be part of a user namespace. I'm not fully convinced but there are some DOS attacks that would naturally prevent.

The discussion ultimately led Serge to conclude that the syslog resource should instead be grouped as part of the user namespace rather than as a separate namespace:

I can't really think of a good case for not putting the syslogns straight into the userns (i.e. not having a separate syslogns), so I'd say let's go that route.

Serge's patch seems to have inspired another group to try implementing syslog namespaces. A couple of days after Serge's patch, Rui Xiang posted some patches that he and his colleague Libo Chen had developed to implement similar functionality. Rui began by noting a couple of the obvious differences in their patch set:

In Serge's patch [...] syslog_namespace was tied to a user namespace. We add syslog_ns tied to nsproxy instead, and implement ns_printk in ip_table context.

We add syslog_namespace as a part of nsproxy, and a new flag CLONE_SYSLOG to unshare syslog area.

Using nsproxy is the conventional way of dealing with the namespaces associated with a process: it is a structure that contains pointers to structures describing each of the namespaces that a process is associated with. This contrasts with Serge's original approach, which hung the syslog namespace off the user namespace.

Rui's team also took advantage of a detail that Serge perhaps overlooked: there happens to be one spare bit in the flag space for clone() because the CLONE_STOPPED flag was removed several kernel releases ago. Therefore, Rui's team repurposed that bit. Normally, it would not be safe to recycle flag bits in this way, but the CLONE_STOPPED flag has a special history. It was initially proposed for use specifically in the NPTL threading implementation, but the final implementation abandoned the flag in favor of a different approach. As such, CLONE_STOPPED is likely never to have had serious user-space users.

Unsurprisingly, the overall approaches of the two patch sets have many similarities, but there are differences in details such as how a syslog namespace is associated with a struct net in order to solve the "no current process context problem".

Although kernel flame wars between competing implementations are what often make the biggest headlines in the online press, the subsequent exchange between Serge, Rui, and Libo demonstrated that life on developer mailing lists is usually more cordial. Serge asked:

I understand that user namespaces aren't 100% usable yet, but looking long term, is there a reason to have the syslog namespace separate from user namespace?

In response, Rui noted:

Actually we don't have strong preference. We'll think more about it. Hope we can make consensus with Eric.

That in turn led Serge to ask Rui and Libo if his patch set might suffice for their needs, with the gracious note that:

I'm not at all wedded to my patchset. I'm happy to go with something else entirely. My set was just a proof of concept.

There is one other notable difference in functionality between the two patch sets. In Serge's patch set, system consoles belonged (by intention) only to the initial syslog namespace, meaning that kernel log messages from other syslog namespace instances can't be displayed on consoles. By contrast, Rui and Libo's patches include consoles in the syslog namespace, so that kernel messages from syslog namespaces other than the initial namespace can be displayed on consoles. Rui and Libo would like this functionality in order to be able to obtain kernel log messages from containers when monitoring embedded devices that provide access to the console over a serial port.

The summary of the discussion is that there are useful pieces in both patches. Serge plans to revise his patch to merge the syslog namespace functionality into user namespaces, add the console functionality desired by Rui and Libo, and add some in-kernel uses of the namespace-aware printk() interface as a proof-of-concept for the implementation (as was done in the patches by Rui and Libo).

Concluding remarks

The history of the work to provide syslog namespaces (or as it might better be termed, namespace isolation for the kernel log) presents a microcosm of work on namespaces in general. As has often been the case, the implementation of namespaces often turns out to be surprisingly complex. Much of that complexity hinges on detailed questions of functionality (for example, the behavior of consoles in this case) and the question of whether resources should be grouped inside a new namespace or within an existing namespace. In the case of syslog namespaces, it looks like a number of decisions have been made; there will probably be a few more rounds of patches, but there seems to be general consensus on the direction forward. Thus, there is a reasonable chance that proper namespace isolation of kernel logging will appear in the kernel sometime around Linux 3.9 or soon afterward.

Comments (9 posted)

Optimizing stable pages

By Jonathan Corbet
December 5, 2012
The term "stable pages" refers to the concept that the system should not modify the data in a page of memory while that page is being written out to its backing store. Much of the time, writing new data to in-flight pages is not actively harmful; it just results in the writing of the newer data sooner than might be expected. But sometimes, modification of in-flight pages can create trouble; examples include hardware where data integrity features are in use, higher-level RAID implementations, or filesystem-implemented compression schemes. In those cases, unexpected data modification can cause checksum failures or, possibly, data corruption.

To avoid these problems, the stable pages feature was merged for the 3.0 development cycle. This relatively simple patch set simply ensures that any thread trying to modify an under-writeback page blocks until the pending write operation is complete. This patch set, by Darrick Wong, appeared to solve the problem; by blocking inopportune data modifications, potential problems were avoided and everybody would be happy.

Except that not everybody was happy. In early 2012, some users started reporting performance problems associated with stable pages. In retrospect, such reports are not entirely surprising; any change that causes processes to block and wait for asynchronous events is unlikely to make things go faster. In any case, the reported problems were more severe than anybody expected, with multi-second stalls being observed at times. As a result, some users (Google, for example) have added patches to their kernels to disable the feature. The performance costs are too high, and, in the absence of a use case like those described above, there is no real advantage to using stable pages in the first place.

So now Darrick is back with a new patch set aimed at improving this situation. The core idea is simple enough: a new flag (BDI_CAP_STABLE_WRITES) is added to the backing_dev_info structure used to describe a storage device. If that flag is set, the memory management code will enforce stable pages as is done in current kernels. Without the flag, though, attempts to write a page will not be forced to wait for any current writeback activity. So the flag gives the ability to choose between a slow (but maybe safer) mode or a higher-performance mode.

Much of the discussion around this patch set has focused on just how that flag gets set. One possibility is that the driver for the low-level storage device will turn on stable pages; that can happen, for example, when hardware data integrity features are in use. Filesystem code could also enable stable pages if, for example, it is compressing data transparently as that data is written to disk. Thus far, things work fine: if either the storage device or the filesystem implementation requests stable pages, they will be enforced; otherwise things will run in the faster mode.

The real question is whether the system administrator should be able to change this setting. Initial versions of the patch gave complete control over stable pages to the user by way of a sysfs attribute, but a number of developers complained about that option. Neil Brown pointed out that, if the flag could change at any time, he could never rely on it within the MD RAID code; stable pages that could disappear without warning at any time might as well not exist at all. So there was little disagreement that users should never be able to turn off the stable-pages flag. That left the question of whether they should be able to enable the feature, even if neither the hardware nor the filesystem needs it, presumably because it would make them feel safer somehow. Darrick had left that capability in, saying:

I dislike the idea that if a program is dirtying pages that are being written out, then I don't really know whether the disk will write the before or after version. If the power goes out before the inevitable second write, how do you know which version you get? Sure would be nice if I could force on stable writes if I'm feeling paranoid.

Once again, the prevailing opinion seemed to be that there is no actual value provided to the user in that case, so there is no point in making the flag user-settable in either direction. As a result, subsequent updates from Darrick took that feature out.

Finally, there was some disagreement over how to handle the ext3 filesystem, which is capable of modifying journal pages during writeback even when stable pages are enabled. Darrick's patch changed the filesystem's behavior in a significant way: if the underlying device indicates that stable pages are needed and the filesystem is to be mounted in the data=ordered mode, the filesystem will complain and mount it read-only. The idea was that, now that the kernel could determine that a specific configuration was unsafe, it should refuse to operate in that mode.

At this point, Neil returned to point out that, with this behavior, he would not be able to set the "stable pages required" flag in the MD RAID code. Any system running an ext3 filesystem over an MD volume would break, and he doesn't want to deal with the subsequent bug reports. Neil has requested a variant on the flag whereby the storage level could request stable pages on an optional basis. If stable pages are available, the RAID code can depend on that behavior to avoid copying the data internally. But that code can still work without stable pages (by copying the data, thus stabilizing it) as long as it knows that stable pages are unavailable.

Thus far, no patches adding that feature have appeared; Darrick did, however, post a patch set aimed at simply fixing the ext3 problem. It works by changing the stable page mechanism to not depend on the PG_writeback page flag; instead, it uses a new flag called PG_stable. That allows the journaling layer to mark its pages as being stable without making them look like writeback pages, solving the problem. Comments from developers have pointed out some issues with the patches, not the least of which is that page flags are in extremely short supply. Using a flag to work around a problem with a single, old filesystem may not survive the review process.

The end result is that, while the form of the solution to the stable page performance issue is reasonably clear, there are still a few details to be dealt with. There appears to be enough interest in fixing this problem to get something worked out. Needless to say, that will not happen for the 3.8 development cycle, but having something in place for 3.9 looks like a reasonable goal.

Comments (26 posted)

The PowerClamp driver

By Jonathan Corbet
December 5, 2012
The kernel's power management subsystem has become increasingly effective over recent years, to the point that our CPU power management is said to be second to none. But, while the kernel endeavors to minimize the power consumed by a given workload, it lacks mechanisms to put an overall limit on the amount of power consumed. The recently-announced PowerClamp driver by Jacob Pan and Arjan van de Ven is intended to change that situation on Intel processors.

Most users will never want to use PowerClamp. As a general rule, when one has purchased hardware with a given computational capability, one wants that full capability to be available when needed. But there are situations where it makes sense to run a system below its full speed. Data centers have power-consumption and cooling constraints that can argue against running all systems flat-out all the time. Even the owner of an individual laptop or handheld system may wish to ensure that its operating temperature does not exceed a given value; an overly hot laptop can be uncomfortable to work with, even if it is still working within its specified temperature range. So there can be value in telling the system to run slower at times.

The PowerClamp driver allows the system administrator to set a desired idle percentage by way of a sysfs attribute. That percentage is capped at 50% in the current implementation. Once a percentage has been set, the kernel monitors the actual idle time for each processor in the system. Should a processor's idle time fall below the desired idle percentage, a special kernel thread (called kidle_inject/N, where N is the number of the CPU to which the thread is assigned) is created to take corrective action.

That thread operates as a high-priority realtime process, so it is able to respond quickly when needed. Its job is relatively simple: look at the amount of idle time on its assigned CPU and calculate the difference from the desired idle time. Then, periodically, the thread will run, disable the clock tick, and force the CPU into a sleep state for the required amount of time. The sleeping is done for a given number of jiffies, so the sleep states tend to be relatively long — a necessary condition for an effective reduction in power usage.

Naturally, the PowerClamp thread will continue to monitor actual idle time as it operates, adjusting the amount of forced sleep time as needed. It also monitors the amount of desired sleep time that is lost to interrupts. Interrupts remain enabled during the forced sleep, so they can bring the processor back to an operational state before the PowerClamp driver would have otherwise done so. Over time, the amount of sleep time lost in this manner is tracked; the driver will then attempt to compensate by increasing the amount of forced sleep time to try to pull the CPU back to the original idle time target.

By itself, PowerClamp can come close to achieving the desired level of idle time on a system with a changing workload. Often, though, the real goal is not idle time as such; instead, the purpose is to keep the system within a given level of power consumption or a set of thermal limits. Doing that will require the implementation of additional logic in user space. By monitoring the parameter of interest, a user-space process can implement a control loop that adjusts the desired level of idle time as needed. The PowerClamp driver can respond relatively quickly to those changes, giving the control process an effective tool for the management of the amount of power used by the system.

The driver has been through a couple of revisions with little in the way of substantive comments. This patch poses a relatively small risk to the system, since it does not do anything if the feature is not in use. It could thus conceivably be ready for merging as soon as the 3.8 development cycle. Some more information can be found in the documentation file included with the patch.

Comments (17 posted)

Patches and updates

Kernel trees


Core kernel code

Development tools

Device drivers

Filesystems and block I/O


Virtualization and containers


  • Lucas De Marchi: kmod 12 . (December 5, 2012)

Page editor: Jonathan Corbet


CyanogenMod 10 on the Nexus 7

By Jonathan Corbet
December 5, 2012
Last week's review of Ubuntu core 12.10 on the Nexus 7 tablet showcased one of the alternative operating systems that can be installed onto this particular device. But Ubuntu is certainly not the only choice out there. While he was busy installing software onto the tablet, your editor decided to give the latest CyanogenMod build a try. As is usual with CyanogenMod, the results were good, but also a bit discouraging with regard to how the Android development community works.

The page suggests that there is no stable build for the Nexus 7, but there is a 10.0.0 release listed on the stable releases page. Beyond that, "nightly" development builds are available under the "grouper" code name. Naturally, the nightly build was chosen; when would a self-respecting editor pick a stable build over something leading-edge and potentially dangerous? As it happens, the installation of the December 4 nightly build went without a hitch. Or, at least, it did once your editor remembered to wipe the device prior to trying to boot the new system; otherwise it simply hung at the boot splash screen. As usual, one also needs to install the Google applications separately.

The CM10 nightly release works flawlessly, as far as your editor can tell. It has some interesting differences from the stock Android install, many of which are reminiscent of a handset-oriented system. For example, the number of applications is far below what stock Android has; CyanogenMod lacks Google+, Chrome (it has the standard Android browser), Maps and more, but it does include the camera application by default. The missing applications can, naturally, be installed easily from the "Play" store afterward.

[Navbar configuration] As reported here in July, recent CyanogenMod builds seem to have fewer shiny features above stock Android than they did in the past. There is a whole set of configuration options, especially with regard to how the interface works. CyanogenMod also adds profiles, a more configurable lock screen (though stock Android is catching up and taking its own direction in this area), a more useful user space for those who get to the command-line level, and a set of scary "performance" knobs. That is about it; many users might not ever notice or make use of the additional features that CyanogenMod offers. Given that, many users might well wonder why they should bother installing CyanogenMod; for many of them, the best answer might be that they shouldn't.

That is doubly true for Nexus 7 users at this particular point in time; CM10, while not yet released in stable form, is already obsolete: it is based on the Android 4.1.2 release. Anybody running a stock Nexus 7 is likely to have already been updated to 4.2, which offers a number of new features. The CyanogenMod developers are busily trying to catch up with this release and the list of devices supported by the experimental, 4.2-based CM10.1 release is growing, but the Nexus 7 does not yet appear there. So running CyanogenMod on this device means accepting a net loss in new features: no fancy lock screen, no swipe typing, no screen magnifier, etc.

Unfortunately, that state of affairs looks to be a permanent part of the experience of running CyanogenMod (or any other Android derivative). As has been pointed out many times, Android is (mostly) open source, but it is not run as an open-source project. Instead, the world outside of Google gets an occasional code dump after an official Android release is made. Thanks to the heroic efforts of the Google folks working on the Android Open Source Project, those code dumps are both timely and useful for the community. They are a great gift, and we should never forget the value of that gift.

It is worth keeping in mind why things are done that way as well. Clearly, it is easier to run a large software project without having to involve all those pesky community people; there is a whole level of bikeshedding behavior that the Android developers simply do not have to deal with. Keeping the code under wraps also allows Google to control when it first appears on devices — and which devices those will be. The Nexus-branded handsets and tablets have a lot of nice features, including their relative openness. Not the least of those features is that they tend to be the first showcase for new versions of the Android system. If the public Android repositories were always current, a new Android release could be old news by the time it appeared on an officially blessed device.

So Google's reasoning is understandable, but it is still hard not to wish for a different situation. An always-current public repository would allow the CyanogenMod developers to keep up with the tree as it evolved, rather than having to figure out a new code dump a couple of times each year. Perhaps they could even manage to upstream more of their interesting work, helping Android to evolve more quickly, and in more interesting directions. It would help Android to be a real open-source project.

That, however, does not appear to be in the cards. So CyanogenMod and others will tend to lag a bit behind what official Android can do, at least if Google continues to develop and release the system at the current fast pace. The result is that, for those who have devices running current, relatively pure Android software, CyanogenMod may not have a lot to offer. On the other hand, CyanogenMod retains its value as a laboratory where new features can be tested. It is unparalleled in its support for older devices that are no longer supported by their manufacturers — and, in this industry, "older" can have a value of less than one year. Users who have devices that are infected with manufacturer- or carrier-supplied "enhancements" will continue to appreciate the work that the CyanogenMod developers do. So there is a an important role for CyanogenMod, even if that role is changing over time.

Comments (8 posted)

Brief items

Distribution quotes of the week

So next time you're not happy about something: just prefix your criticism with "I think". You may be surprised what difference it makes to the conversation.

Oh, two other magic words: "for me". Compare "This workflow is completely broken" vs "This workflow is completely broken for me". Amazing what difference those two words make...

-- Peter Hutterer

Documenting a distribution is a lot like law enforcement. You might get results from routine patrols, but it is largely a complaint-driven venture.
-- pete

Paraphrasing Star Trek's Bones, that would be "Gentoo, Jim, but not as we know it."
-- Duncan

Comments (none posted)

NetBSD 5.2 Released

The NetBSD Project has announced the release of NetBSD 5.2. "NetBSD 5.2 is the second feature update of the NetBSD 5.0 release branch. It represents a selected subset of fixes deemed critical for security or stability reasons, as well as new features and enhancements."

Comments (none posted)

Red Hat releases RHEL 6.4 Beta

Red Hat has released a beta version of Red Hat Enterprise Linux (RHEL) 6.4. "The beta release includes a broad set of updates to the platform's existing feature set, while also providing rich new functionality in the areas of identity management, file system, virtualization, storage and productivity tools."

Comments (none posted)

Distribution News

Mageia Linux

Mageia 1 End Of Life

Mageia 1 has reached its end-of-life. There will be no further updates. Users of Mageia 1 are encouraged to upgrade to Mageia 2.

Full Story (comments: none)

Red Hat Enterprise Linux

Red Hat Enterprise Linux Extended Update Support 6.0 - End Of Life

Red Hat Enterprise Linux Extended Update Support Add-On 6.0 is no longer supported.

Full Story (comments: none)

Newsletters and articles of interest

Distribution newsletters

Comments (none posted)

New versions of aptosid and siduction continue Sidux legacy (The H)

The H looks at the latest releases of two distributions based on Debian unstable (sid), aptosid and siduction. "Shortly after that release, the developers of siduction, a distribution forked from aptosid to be more community-focused, announced its second release candidate of siduction 2012.2 "Riders on the Storm". The final release of the distribution will add the Razor-qt desktop environment for the first time to its range of desktops. Both distributions aim to provide a usable desktop Linux solution based on Debian Sid with a stable upgrade path."

Comments (none posted)

Opinion: Best Linux Distro for Me Has a Refreshing Burst of Cinnamon (

Brian Warner likes Linux Mint with Cinnamon. "The developers behind Linux Mint have historically done a great job in striking a comfortable balance between beauty and usability, but it's also worth calling out the team's ability to execute. While its roots go back to a number of Mint-specific GNOME Shell extensions, Cinnamon itself is quite stable despite being not quite one year old."

Comments (1 posted)

Page editor: Rebecca Sobol


The synthesizers of Sean Bolton

December 5, 2012

This article was contributed by Dave Phillips

The community of Linux audio developers is a relatively small group that includes a core of programmers committed to the evolution of the Linux audio ecosystem. Some developers work on the unglamorous but necessary system components — sound card drivers, kernel interfaces, function libraries, etc. — while others work on the shinier applications level. Developer/musician Sean M. Bolton wears hats in both domains, with a special attachment to the DSSI plugin interface. His contributions include code for the interface specification, a DSSI plugin host, and three software synthesizers.

This article profiles Sean's synthesizers, Xsynth, WhySynth, and Hexter. As we might expect, they are DSSI-compatible plugins, so any host that supports DSSI should run them without complaint. You'll need the DSSI library and utilities first, but most mainstream distributions include the software in their package repositories. Check your system's package manager for an installable bundle. If you need to build the DSSI system yourself, have no fear, the software is easy to build, with no extraordinary dependencies. The synthesizers are equally straightforward to build and install.

For those who want to quickly play with these synthesizers, the complete DSSI system includes a handy command-line utility for running DSSI plugins as standalone applications. My tests were made with this invocation from an xterm command prompt:

	jack-dssi-host /usr/lib/dssi/

The synthesizers were correctly listed in the audio and ALSA MIDI tabs in QJackCtl. I connected my Akai LP25 keyboard and various sequencers to play each synthesizer, and I'm happy to report that I experienced no problems. Sean's synthesizers are stable applications, ALSA and JACK compatible, with great sound and copious presets. Let's look at each one in some detail.

Xsynth 0.9.4

Xsynth is a 2-oscillator subtractive synthesizer based on an original design by Steve Brookes. The synthesizer's [Xsynth patch editor] architecture is revealed by its patch editor (right) — the output from the oscillators is mixed, filtered, amplified, and modulated before reaching the audio output stage, a typical arrangement for a subtractive synthesizer. Each oscillator can select one of seven cyclic waveforms, the pulse-width duty cycle is user-definable, and oscillator sync is available. The filter is relatively simple, with controls only for cutoff frequency, resonance, and rolloff mode. The amplifier envelope is likewise uncomplicated, with a typical ADSR (attack/decay/sustain/release) envelope to control the evolution of the sound's amplitude. The LFO includes a set of six waveforms, a single control for the frequency of the selected waveform, and two controls for pitch modulation and filter modulation sensitivity. No on-board effects are available, but I prefer mine external anyway. JACK-Rack or the awesome Rakarrak make good processing companions for Xsynth.

Xsynth's documentation consists of a README file in the source package and a default collection of 50+ presets. The README contains much valuable information regarding the synthesizer's architecture and controls, and the presets include excellent examples of commonly encountered analog synthesizer sounds such as string pads, resonant filter sweeps, and fat basses. From the README and the presets you can learn all you need to know to master the synthesis method. You can also learn that Xsynth responds to various MIDI messages, including note-on/off, velocity, aftertouch, program change, mod wheel, volume control, and others. See the file for the complete details.

Xsynth is a standard item in the full DSSI package, so if you've installed DSSI you've already installed Xsynth and you're ready to roll with it. The DSSI software is available in the repositories for most mainstream Linux distributions, but if yours doesn't have it you can pick up the source code on the DSSI site.

There's not much more to say about Xsynth. It's an uncomplicated realization of a classic analog synthesis method, presented in an uncluttered interface, easy to learn and use. I like the sound of many of its presets, especially when they're routed through an external effects processor, and it's great fun to program for my own sounds. But Xsynth is no mere toy — even simple subtractive synthesis is capable of making wonderful rich sounds, and you may well lose track of time while you explore its capabilities.

WhySynth 120903

WhySynth is what Xsynth dreams of becoming when it grows up. Both synthesizers are designed with the same large-scale architecture, that of the analog subtractive synthesizer, but WhySynth's implementation of the synthesis method differs profoundly at the detail level. The number of oscillators has been doubled over what Xsynth offers, the waveform selection has increased to more than 168 waves, and each oscillator can be configured to a unique modality (e.g., wavecycle, frequency modulation (FM), noise). The number of filters has increased to two, we now have three LFOs, and the number of envelope generators (EGs) has grown from two to five. WhySynth adds an effects stage to the classic design, though it is wisely restricted to only two reverbs and a delay line. The parameter set for each stage has likewise expanded for considerably finer control over your sounds.

[WhySynth patch editor] WhySynth's UI follows the design set by Xsynth. The program opens to a single panel with tabs for the preset patch list and the synthesizer's global configuration. The patch editor is considerably more complex than Xsynth's, so it gets its own window, available from the Edit menu. Like its sibling, WhySynth's editor clarifies the organization of the synthesis method. WhySynth is a deep synthesizer, meaning that it allows very fine control of the sound-shaping process, and it is capable of timbres and effects not available with a simpler architecture.

WhySynth can import Xsynth patches and banks, so work done in that synthesizer can be brought into WhySynth for more detailed design. WhSynth also includes a facility for interpreting patches in sysex format for Kawai's K4 synthesizer. Actually, Sean indicates that WhySynth mis-interprets them, but it's obvious that those patches provided a great resource during the creation of WhySynth's 280+ presets, of which more than 130 have been derived from patches for the K4.

The documentation is slim but informative. The WhySynth web site and the source package README contain the same material describing the synthesizer's design in detail. Given the complexity of the program, I suggest reading the documentation thoroughly if you plan on making your own sounds. Of course the default patches are instructive, and WhySynth provides a number of "development" patches to be used as starter material. You'll also want to read the docs to find out what MIDI controllers have been mapped to WhySynth's synthesis parameters for dynamic control of your sound's evolution.

Hexter 1.0.1

According to its "About" panel, Hexter is a "Yamaha DX7 modeling software synthesizer for the DSSI Soft Synth Interface". The DX7 was Yamaha's most famous synthesizer built with FM (frequency modulation) audio technology licensed from Stanford University, where Dr. John Chowning invented the method. This review is not the place for an explanation of FM, please see the Wikipedia on FM synthesis for a good summary and some excellent external links. It suffices here to note that FM differs substantially from the synthesis methods seen and heard in Xsynth and WhySynth.

Hexter opens to a display similar to its siblings, a single panel with tabs for patch selection, global configuration, and performance settings. The patch selector is self-explanatory. The "Configuration" tab sets the synthesizer master tuning, output volume, and polyphony. It also toggles access to ALSA "sysex" editing, about which I'll have more to say later. The additional "Performance" tab includes settings for pitch bend, modulation wheel, foot controller, and breath controller. The bend and mod wheels were parts of the original DX7 hardware, the controllers were external devices. All these devices were designed to add greater expressive capabilities during performance, for example, by modulating a sustained tone by controlling its vibrato with the mod wheel.

The DX7 spawned an industry of third-party extensions, add-ons, and designer patch libraries. The hardware-based DX7 additions are history, but a vast number of DX7 patches are still available as systems-exclusive (sysex) data files. Hexter can read a variety of DX-related sysex files, but only so far as they contain the basic elements of the original DX7. Alas, there's no support for the extended features of other members of Yamaha's DX family. However, Hexter also supports the file format for the Sideman D/TX patch editor/librarian by Voyetra. That's great news [Hexter's patch editor] for me, as I used that software for many years to program my beloved TX802, a rack-mount synthesizer in the mkII family of Yamaha's DX/TX synthesizers. Alas, the 802 is gone, but my patch collection has received a new lease on life thanks to Hexter.

Catching up with Xsynth and WhySynth, Hexter now has a graphic editor for basic DX7 patch parameters. Unlike its siblings, Hexter's editor includes graphic envelope displays for relevant parameters. Alas, the envelope breakpoints can't be adjusted directly, but the visual feedback is immediate when values are set in the scroll boxes.

As I mentioned earlier, the Configuration tab includes a toggle for enabling patch edits via system-exclusive messages. A separate ALSA MIDI port is opened for receiving sysex messages from an appropriate device or program, such as a hardware DX/TX synthesizer or a software editor/librarian. The integrated editor will be sufficient for most users, but I found a very helpful purpose for the sysex connection. The Voyetra Sideman editor includes a well-designed patch randomizer, and Hexter conveniently provides a compatible target synthesizer for the editor's patches. The Sideman software runs smoothly under DOSemu, I use a MIDI connections utility such as aconnect or QJackCtl to route its output to Hexter's sysex port, [Hexter and Sideman] and voilà: I can program Hexter in realtime with an editor running in an emulated MS-DOS environment.

Hexter includes no on-board effects processors. No problem, it's a JACK-savvy application, so just route its output to JACK-Rack or Rakarrak. The original DX never sounded so good.

Regarding external control: Hexter's synthesis parameters are controllable with a set of pre-assigned MIDI continuous controllers, including a lengthy list of parameters addressed by the NRPN (non-registered parameter numbers). The default assignments can be found in the source package's README. Hexter also supports OpenSound Control (OSC) but I haven't yet looked into its possibilities.

I've owned four FM synthesizers from Yamaha. My first synthesizer was a TX81Z, followed by an FB01, both of which were 4-operator FM boxes. I stepped up to the 6-operator machines with the TX802 and the massively powerful TG77. I've also tested FM7 and FM8, excellent FM synthesis programs from Native Instruments. Alas, my hardware boxes are gone now, and the proprietary software isn't made for Linux. However, I'm not complaining at all — I have Hexter. I won't compare it to the hardware or the Native Instruments synthesizers, I'll just say that I'm happy to hear my favorite FM sounds again.

Sounding Out

By now I hear you say, "But how do they sound?". Well, you can check out the following links for some demonstrations of the Sean Bolton synthesizer collection:

Subtractive and FM synthesizers are staple items in the complete modern computer-controlled studio. The two synthesis methods make a good pair; subtractive synthesizers are often used for dense, rich, layered sounds and strings, while FM excels in metallic and bell-like sounds. Sean Bolton has given us three capable synthesizers, now it's up to us to show off their capabilities. Try them all — they're all free and open-source Linux software licensed under the GPL — and do let us know if you make some joyful noises with them.

Comments (3 posted)

Brief items

Quotes of the week

I find it fascinating that DVCS aficionados haven't noticed that GitHub takes the D out of DVCS very effectively, thereby making git actually useful for most normal people.
Branko Čibej (virtual hat-tip to Markus Schaber)

I'm pretty sure that HURD stands for "Hurd Users Relish Deviance" so I would expect Hurd folks to actually appreciate these test failures.
Matt Mackall

Comments (11 posted)

GNU Guile 2.0.7 released

Version 2.0.7 of the GNU Guile language is out. It adds an implementation of "curly infix expressions," per-port reader options, a number of extension loading improvements, and something known as "nested futures": "Futures may now be nested: a future can itself spawn and then `touch' other futures. In addition, any thread that touches a future that has not completed now processes other futures while waiting for the touched future to completed."

Full Story (comments: 104)

Mozilla joins Internet Society

Mozilla's Josh Aas announced on his blog that the browser maker was joining the Internet Society (ISOC) as a "Silver" member, in order to support its Internet Engineering Task Force (IETF) work on core protocols beyond HTTP. "Today we’re heavily involved in IETF working groups relating to key Internet technologies such as TLS, HTTP and HTTP/2, RTCWeb, WebSockets, and others." ISOC, of course, is the parent organization of IETF and an alphabet soup of other Internet standards bodies.

Comments (none posted)

Ekiga 4.0 released

Version 4.0 of the Ekiga telephony application is out. It features a new user interface, some new codecs, auto-answer functionality, a number of improvements in SIP support, and more.

Comments (7 posted)

The first "shim" UEFI secure bootloader released

Matthew Garrett has announced the availability of the first "usable" version of the "shim" UEFI secure bootloader. "If you want, you're then free to impose any level of additional signing restrictions - it's entirely possible to use this signing as the basis of a complete chain of trust, including kernel lockdowns and signed module loading. However, since the end-user has explicitly indicated that they trust your code, you're under no obligation to do so. You should make it clear to your users what level of trust they'll be able to place in their system after installing your key, if only to allow them to make an informed decision about whether they want to or not."

Comments (none posted)

Nmap 6.25 released

Version 6.25 of the Nmap network scanner is out; it contains a lot of new stuff. "Nmap 6.25 contains hundreds of improvements, including 85 new NSE scripts, nearly 1,000 new OS and service detection fingerprints, performance enhancements such as the new kqueue and poll I/O engines, better IPv6 traceroute support, Windows 8 improvements, and much more."

Full Story (comments: 2)

Buildroot 2012.11 available

Version 2012.11 of the buildroot embedded Linux system-creation tool has been released. This stable release adds initial support for Aarch64, a large number of new packages, and updates to several core components, including: "Binutils 2.23.1, GCC 4.7.2. We're now defaulting to GCC 4.6. Newer Codebench and Linaro external toolchains. Libtirpc support for modern Glibc variants"

Full Story (comments: none)

Newsletters and articles

Development newsletters from the past week

Comments (none posted)

Taylor: Avoiding Jitter in Composited Frame Display

Owen Taylor has written a detailed update about his ongoing work on compositor frame timing (to which we provided an introduction back in August). This time, the issue at hand is coping with video playback or other sources that need a fixed frame rate not equal to the display's refresh rate. "I’m pretty happy with how this algorithm works out in testing, and it may be as good as we can get for X. The main downside I know of is that it only individually solves the two problems – handling clients that need all the rendering resources of the system and handling clients that want minimum jitter for displayed frames, it doesn’t solve the combination."

Comments (31 posted)

GIMP Magazine issue 2 available

The second issue of Gimp Magazine has been released as a 100-page, 65MB PDF file. covered topics include a graphic novel tutorial, oil painting, using graphic tablets, a number of artist interviews, and more.

Comments (none posted)

Page editor: Nathan Willis


Brief items

MariaDB Foundation announced

MariaDB developers Michael 'Monty' Widenius, David Axmark, and Allan Larsson have announced the MariaDB Foundation. "In its mission statement, the MariaDB Foundation exists to improve database technology, including standards implementation, interoperability with other databases, and building bridges to other types of database such as transactional and NoSQL. To deliver this the Foundation provides technical work in reviewing, merging, testing, and releasing the MariaDB product suite. The Foundation also provides infrastructure for the MariaDB project and the user and developer communities." (Thanks to Dan Shearer)

Comments (1 posted)

2012 LinuxCon Europe videos

Some videos from the 2012 LinuxCon in Barcelona have been posted. (Thanks to Scott Dowdle)

Comments (1 posted)

Articles of interest

Free Software Supporter -- Issue 56, November 2012

Free Software Foundation's monthly newsletter is out, with a look at Cyber Monday, Amazon books, MediaGoblin, software patents, VLC licensed, and much more.

Full Story (comments: none)

FSFE Newsletter - December 2012

The Free Software Foundation Europe's monthly newsletter covers Free Software in the UK, secure boot, Free Software in Germany, and several other topics.

Full Story (comments: none)

Calls for Presentations

European Lisp Symposium 2013: Call for Papers

The European Lisp Symposium will take place June 1-4, 2013 in Madrid, Spain. The call for papers deadline is March 1. "The purpose of the European Lisp Symposium is to provide a forum for the discussion and dissemination of all aspects of design, implementation and application of any of the Lisp and Lisp-inspired dialects, including Common Lisp, Scheme, Emacs Lisp, AutoLisp, ISLISP, Dylan, Clojure, ACL2, ECMAScript, Racket, SKILL, Hop and so on. We encourage everyone interested in Lisp to participate."

Full Story (comments: none)

Upcoming Events

Bdale Garbee to keynote LCA has announced that Bdale Garbee will be a keynote speaker at the 2013 conference next January. "Bdale Garbee is best known for his pioneering work with Debian, and for open source community-building efforts with the Linux Foundation, Freedombox, and Software in the Public Interest (SPI). He is a regular presence at, wowing many recent conference-goers with his rocketry exploits and other hobby activities turned into open source projects."

Full Story (comments: none)

Events: December 6, 2012 to February 4, 2013

The following event listing is taken from the Calendar.

December 5
December 7
Qt Developers Days 2012 North America Santa Clara, CA, USA
December 5
December 7
Open Source Developers Conference Sydney 2012 Sydney, Australia
December 7
December 9
CISSE 12 Everywhere, Internet
December 9
December 14
26th Large Installation System Administration Conference San Diego, CA, USA
December 27
December 30
29th Chaos Communication Congress Hamburg, Germany
December 27
December 29
SciPy India 2012 IIT Bombay, India
December 28
December 30
Exceptionally Hard & Soft Meeting 2012 Berlin, Germany
January 18
January 19
Columbus Python Workshop Columbus, OH, USA
January 18
January 20
FUDCon:Lawrence 2013 Lawrence, Kansas, USA
January 20 Berlin Open Source Meetup Berlin, Germany
January 28
February 2 2013 Canberra, Australia
February 2
February 3
Free and Open Source software Developers' European Meeting Brussels, Belgium

If your event does not appear here, please tell us about it.

Page editor: Rebecca Sobol

Copyright © 2012, Eklektix, Inc.
Comments and public postings are copyrighted by their creators.
Linux is a registered trademark of Linus Torvalds