Linpus Lite (or maybe just aspire one mail) is only going to last a few more days on this aspire1, it's beginning to annoy me:
Update
sylpheed has now replaced aspire mail. OK, I still use aspire mail for calendar and addressbook stuff (for now), but sylpheed is a much better email client.

1. aspire one mail has no GnuPG support


2. aspire one mail has no non-HTML composition support. No, really, it doesn't even send plain text alongside the HTML. email sent from the aspire to myself comes in as an attachment in other email clients! (i.e. spam)

So, I would have to investigate a different email client but then there are niggles with Linpus itself:

1. the aspire default kernel build lacks the pl2303 kernel module so that I can't use the aspire to control my balloon3 board over usb serial


2. intermittent problem with network manager forgetting WPA passphrases

Testing continues - probably resulting in installing Lenny at the end of this week. Some of these things might be fixable using pages like taoofmac and Reg Hardware.
Unfortunately, the problem with usbserial means that I'll have to take two laptops if I want to demonstrate the balloon3 board in places.
I also don't seem to be able to locate any copyright files which makes me nervous. Maybe it's just a lack of knowledge of Fedora (at least not since FC5).

It’s not easy to choose a laptop, and part of the problem is that most people don’t seem to start from the use of the laptop. I believe that the following four categories cover the vast majority of the modern use of mobile computers.

1. PDA [1] - can be held in one hand and generally uses a touch-screen. They generally do not resemble a laptop in shape or design.


2. Subnotebook [2] AKA Netbook [3] - a very small laptop that is designed to be very portable. Typically weighing 1KG or less and having multiple design trade-offs to give light weight (such as a small screen and a slow CPU) while still being more powerful than a PDA. The EeePC [4] is a well known example.


3. Laptop [5] - now apparently defined to mean a medium size portable computer, light enough to be carried around but with a big enough screen and keyboard that many people will be happy to use them for 8 hours a day. The word is also used to mean all portable computers that could possibly fit on someone’s lap.


4. Desktop Replacement [6] - a big heavy laptop that is not carried much.

There is some disagreement about the exact number of categories and which category is most appropriate for each machine. There is a range of machines between the Subnotebook and Laptop categories. There is some amount of personal preference involved in determining which category a machine might fall in. For example I find a Thinkpad T series to fit into the “Laptop” category (and I expect that most people would agree with me). But when comparing the weight and height of an average 10yo child to an adult it seems that a 10yo would find an EeePC to be as much effort to carry as a T series Thinkpad is for an adult. It seems to me that the first thing that you need to do when choosing a laptop is to decide which of the above categories is most appropriate. While the boundaries between the categories are blurry and to some extent are limited by personal preference it’s an easy second step to determine which machines fit the category you have selected (in your opinion) once you have made a firm decision on the category. It’s also possible to choose a half-way point, for example if you wanted something on the border of the “Laptop” and NetBook categories then a Thinkpad X series might do the job. The next step of course is to determine which OSs and applications you want to run. There are some situations where the choice of OS and/or applications may force you to choose a category that has more powerful hardware (a CPU with more speed or features, more RAM, or more storage). For example a PDA generally won’t run a regular OS well (if at all) due to the limited options available for input devices and the very limited screen resolution. Even a NetBook has limitations as to what software runs well (for example many applications require a minimum resolution of 800×600 and don’t work well on an EeePC 701). Also Xen can not be used on the low-end CPUs used in some NetBooks which lack PAE. Once you have chosen a category you have to look for features which make sense for that category. A major criteria for a PDA is how fast you can turn it on, it should be possible to go from standby to full use in less than one second. Another major criteria is how long the battery lasts, it should compare to a mobile phone (several days on standby and 8 hours of active use). A criteria that is important to some people is the ability to use both portrait and landscape views for different actions (I use portrait for editing and landscape for reading). A NetBook is likely to be used in many places and needs to have a screen that will work well in adverse lighting conditions (a shiny reflective screen is a really bad thing), it also needs to be reasonably resilient as it is going to get bumped if it is transported a lot (a solid state disk is a good feature). It should also be as light as possible while having enough hardware to run a regular OS (an EeePC 701 with 512M of RAM and 4G of storage is about the minimum hardware for running a regular distribution of Linux). A desktop replacement needs to have all the features, lots of RAM, a fast CPU and video hardware, and a big screen - it also needs a good keyboard (test by typing your name several times). The “Laptop” category is much the same as the desktop replacement, but a bit smaller, a lot lighter, and better battery life. It seems very difficult to give any specific advice as to which laptop to buy when the person who wants the advice has not chosen a category (which is often the case).

• [1] http://en.wikipedia.org/wiki/Pda


• [2] http://en.wikipedia.org/wiki/Subnotebook


• [3] http://en.wikipedia.org/wiki/Netbook


• [4] http://en.wikipedia.org/wiki/Eeepc


• [5] http://en.wikipedia.org/wiki/Laptop


• [6] http://en.wikipedia.org/wiki/Desktop_replacement_computer

Share This

So, does anybody know of a command-line RSS reader? I want something that will save state and output recent entries in a nicely parsable manner. Maybe URL\tTitle\tDecscription\n or somesuch.

So here's why I'm asking.

I got to thinking that it might be nice to automatically post to Twitter when I've got a new blog post up, rather than manually have to say "just wrote a blog post."

Then I also got to thinking that when I see an interesting URL that I'm going to Twitter about, I'm also almost always adding it as a bookmark to Delicious. Why not make a fortwitter tag on Delicious, and automatically post my comments about them to Twitter, saving me having to do it twice?

So I've got Twidge that can be nicely used in a shell script to do this stuff. I'm hoping to avoid having to write a shell script-friendly RSS aggregator. But I'm just prone to do it if nobody else has already, though I'd really like to avoid reinventing the wheel.

I found some weird opcodes in the s390 instruction set some time ago. I finally want to use it as I have access to such a machine. The result is rather slim:

.globl cu41
        .type   cu41, @function
cu41:
.L2:
        cu41    %r2,%r4
        ipm     %r1
        srl     %r1,28
        chi     %r1,3
        je      .L2
        ltr     %r1,%r1
        je      .L3
        lcr     %r1,%r1
        lgfr    %r3,%r1
.L3:
        lgr     %r2,%r3
        br      %r14

The cu41 opcode translates UTF32 to UTF8. Each "parameter" is a register pair [1], which describes address and length of the buffer. This opcode will return after a not specified number of translated characters. The rest is boiler plate code for condition code checking. The loop makes sure that anything is converted within the function unless another error occured.

[1]	An even pair. gcc is not yet able to allocate such pairs on its own, so it is hardcoded.

I found some weird opcodes in the s390 instruction set some time ago. I finally want to use it as I have access to such a machine. The result is rather slim:

.globl cu41
        .type   cu41, @function
cu41:
.L2:
        cu41    %r2,%r4
        ipm     %r1
        srl     %r1,28
        chi     %r1,3
        je      .L2
        ltr     %r1,%r1
        je      .L3
        lcr     %r1,%r1
        lgfr    %r3,%r1
.L3:
        lgr     %r2,%r3
        br      %r14

The cu41 opcode translates UTF32 to UTF8. Each "parameter" is a register pair [1], which describes address and length of the buffer. This opcode will return after a not specified number of translated characters. The rest is boiler plate code for condition code checking. The loop makes sure that anything is converted within the function unless another error occured.

[1]	An even pair. gcc is not yet able to allocate such pairs on its own, so it is hardcoded.

It seems to be a common idea by non-bloggers that the comment they enter on a blog is somehow special and should be taken seriously by the author of the blog (everyone is a legend in their own mind). In a recent discussion one anonymous commentator seemed offended that I didn’t take his comments seriously and didn’t understand why I would take little notice of an anonymous comment while taking note of a later comment on the same issue by the author of the project in question. In most forums (and I use the term in the broadest way) an anonymous comment is taken with a weight that is close to zero. That doesn’t mean that it will be ignored, it just means that the requirement for providing supporting evidence or of having a special insight and explaining it is much greater. One example of this is the comment weighting system used by Slashdot.org (AKA “/.”). The /. FAQ has a question “Why should I log in?” with the answer including “Posting in Discussions at Score:1 instead of Score:0 means twice as many people will see your comments” [1]. /. uses the term “Anonymous Coward” as the identification of users who are not logged in, this gives an idea of how they are regarded. Advogato uses a rating method for blog posts which shows you only posts from blogs that you directly rank well or which match the trust metric (based on rankings of people you rank) [2]. I believe that the automated systems developed by /. and other online forums emulate to a some extent the practices that occur off-line. For any discussion in a public place a comment from someone who does not introduce themself (or gives an introduction that gives no reason to expect quality) will be treated with much less weight than one from someone who is known. When someone makes a comment their background will be considered by people who hear it. If a comment is entirely a matter of opinion and can not be substantiated by facts and logical analysis then the acceptance of the comment is solely based on the background of the author (and little things like spelling errors can count against the author). Therefore if you want your blog comments to be considered by blog authors and readers you need to make sure that you are known. Using your full name is one way of not being as anonymous but most names are not unique on the Internet (I’ve previously described some ways of ensuring that you beat other people with the same name in Google rankings [3]). The person who owns the blog can use the email address that is associated with the comment to identify the author (if it’s a real email address and it’s known by Google). But for other readers the only option is the “Website” field. The most common practice is to use the “Website” field in the comment to store the URL of your blog (most blog comments are written by bloggers). But there is nothing stopping you from using any other URL, if you are not a blogger and want to write comments on blogs you could create a personal web page to use for the comments. If the web page you use for such purposes gives links to references as to your relevant experience then that would help. Someone who has skills in several areas could create a web page for each one and reference the appropriate page in their comment. One problem we face is that it is very easy to lie on the net. There is no technical obstacle to impersonation on the net, while I haven’t seen any evidence of people impersonating others in an attempt to add credibility to blog comments I expect it’s only a matter of time before that happens (I expect that people do it already but the evidence of them getting caught has not been published anywhere that I’ve read). People often claim university education to add weight to their comments (usually in email but sometimes in blog comments too). One problem with this is that anyone could falsely claim to have a university degree and no-one could disprove their claim without unreasonable effort, another is that a university degree actually doesn’t mean much (lots of people remain stupid after graduating). One way in which adding a URL to a comment adds weight is that for a small web site the author will check a reasonable portion of the sites that link to them, so if someone impersonates me and has a link to my web site in the comment then there’s a good chance that I will notice this. OpenID [4] has the potential to alleviate this by making it more difficult to forge an association with a web site. One thing that I am working on is enabling OpenID on all the web sites that are directly associated with me. I plan to use a hardware device to authenticate myself with the OpenID server (so I can securely enter blog comments from any location). I expect that it will become the standard practice that comments will not be accepted by most blogs if they are associated with a URL that is OpenID enabled unless the author of the comment authenticates themself via OpenID. Even when we get OpenID enabled everywhere there is still the issue of domain specific expertise. While I am well enough known for my work on SE Linux that most people will accept comments about it simply because I wrote them, the same can not be said for most topics that I write about. When writing about topics where I am not likely to be accepted as an expert I try and substantiate my main points with external web pages. Comments are likely to be regarded as spam if they have too many links so it seems best to only use one link per comment - which therefore has to be on an issue that is important to the conclusion and which might be doubted if evidence was not provided. The other thing that is needed is a reasonable chain of deduction. Simply stating your opinion means little, listing a series of logical steps that led you to the opinion and are based on provable facts will hold more weight. These issues are not only restricted to blog comments, I believe that they apply (to differing degrees) to all areas of online discussion.

• [1] http://slashdot.org/faq/accounts.shtml


• [2] http://www.advogato.org/person/raph/diary.html?start=224


• [3] http://doc.coker.com.au/internet/personal-seo/


• [4] http://en.wikipedia.org/wiki/Openid

Share This

People frequently ask me what I am doing typing on my Palm right before logging in one of my remote systems. The answer is quite simple: generating my next password . People always seem puzzled by that answer Maybe I am just too old to have had only telnet sessions available to remote connections (that was the time before SSH took over)... or maybe I am just too paranoid regarding the access to my systems Anyway, I like One Time Passwords, and here is an article I can refer to when I get asked again ;-) One Time Passwords are just that: passwords that are good for one time use. They never repeat and once used you can throw it away securely. They were quite common when the authentication was done in clear text (so that any man-in-the-middle could steal your passwords), back in the r-tools age (rcp, rsh, rlogin, rexec, etc). After the SSH-related tools took over, providing easy tunneling and remote access, MITM attacks were of much concern, so OTPs became less relevant. What a shame, for they even have a couple of standards for those! But there s still room for OTPs Question: Is it easier to deploy a MITM attack or a keylogger? That s right We are always connecting to our remote systems in public terminals (well at least I am right now typing from the hospital computer), and a keylogger is so easily installed in one of those terminals, even remotely, that no one would ever get close to deploy a refined MITM attack just to steal someone s password. One would just install a keylogger! All the security SSH provides would just be gone by then. That s why I use SSH to connect to my remote machines, but use OTPs to authenticate myself. It s so easy to deploy it. In Debian you ll find packages opie-server and libpam-opie and those are the only things you ll need on the server side (besides SSH, obviously). OPIE means One-time Passwords In Everything , and combined with PAM, one can really use it everywhere. After installing it, you ll have to generate the password file for your account:

spectra@home:~$ opiepasswd -c
Adding spectra:
Only use this method from the console; NEVER from remote. If you are using
telnet, xterm, or a dial-in, type ^C now or exit with no password.
Then run opiepasswd without the -c parameter.
Using MD5 to compute responses.
Enter new secret pass phrase:
Again new secret pass phrase:
ID spectra OTP key is 499 ho6484
HAVE COOK LOLA AIRY NEIL ROAM

The pass phrase will be used to generate the passwords in a step before your login. Please, try not to forget it (specially if you re following this article and playing with some remote system at the same time). Now you can edit /etc/pam.d/ssh file (or its equivalent in non-Debian systems) to require that kind of authentication. Mine just looks like this:

# /etc/pam.d/ssh
auth       required     pam_env.so
auth       required     pam_env.so envfile=/etc/default/locale
auth       required     pam_opie.so

First two lines are unrelated and just load the environment variables. Last line is where the fun is. Please, note two things: (1) I removed references to pam_unix.so, which is what would ask for my real password, that I want to disable (no login is allowed with that password). And (2), I declared it as required, meaning that failing it will keep one out of the system. We are not ready yet! SSH will work by now, but will not present you the OTP challenge. Probably you could still login, but you d have to remember which is the current password (Trust me, you would not!). To get the challenge you ll need to enable it in /etc/ssh/sshd_config change the following line:

ChallengeResponseAuthentication yes

That s it. Now to login to your remote machine, that s what will usually happen:

spectra@hospital:~$ ssh home
otp-md5 498 ho6484 ext, Response:
spectra@home:~$

Voil ! It asks for password #498. By default, it starts with 500 passwords and goes down from that. Password #498 were asked, so the next will be #497. After that, #498 is not useable anymore, and #496 is not useable yet. You can generate a list of those passwords (let s say 10), print it and keep it in your pocket. This is the command you d use for that:

spectra@home:~$ opiekey -n 10 498 ho6484
Using the MD5 algorithm to compute response.
Reminder: Don't use opiekey from telnet or dial-in sessions.
Sorry, but you don't seem to be on the console or a secure terminal.
Warning: Continuing could disclose your secret pass phrase to an attacker!
Enter secret pass phrase:
489: CALM INTO WEEK APS LOON VIE
490: HASH GYM RAID GOSH HOYT DUAL
491: BELL GIN RIFT HELM GUY BUNK
492: HEBE OBOE SUP LEG LULU LANG
493: HOYT JOT ASK JOG GIBE BETH
494: NASH MOOT HIND YEAH  YAP CARL
495: MATE OF BARD LAVA LEAK AHOY
496: TAB BAG KEY GILT AVID VEAL
497: MOLE FORM NIB LEER ROSS HAVE
498: SING WERE OVEN SOD VEIN NIBS

That is not so secure, since you can loose the paper and be doomed Luckily there are lots of small softwares that does this generation for you. Some you can use from your Palm or from some J2ME-enabled phone (such as N95). Some your can use from another (trusted) computer. Some are even online, written in javascript! There are at least one other side-benefit of using OPIE as above: You can give away your user password (even root password OK, probably not a Good Thing^TM), that the system would still be secure, since it only allows SSH authentications via OPIE! If the session is started with a username whose opiepasswd was not activated (first step scroll back to the beginning of the article), SSH will greet you with a bogus challenge only opie-activated users will be allowed to login with the above configuration. Easy enough, isn t it? Now, next time you see me typing on my Palm before opening an SSH connection you ll know what I am doing ;-)

Some people think that you can recognise a good restaurant by the presence of obscure dishes on the menu or having high prices. The reality is that there are two ways of quickly identifying a good restaurant, one is the Michelin Guide [1] (or a comparable guide - if such a thing exists), the other is how quiet the restaurant is. By a quiet restaurant I certainly don’t mean a restaurant with no customers (which may become very noisy once customers arrive). I mean a restaurant which when full will still be reasonably quiet. Making a restaurant quiet is not in itself a sufficient criteria to be a good restaurant - but it’s something that is usually done after the other criteria (such as hiring good staff and preparing a good menu) are met. The first thing to do to make a room quiet is to have good carpet. Floor boards are easy to clean and the ratio of investment to lifetime is very good (particularly for hard wood), but they reflect sound and the movement of chairs and feet makes noise. A thick carpet with a good underlay is necessary to absorb sound. Booths are also good for containing sound if the walls extend above head height. Decorations on the walls such as curtains and thick wallpaper also absorb sound. A quiet environment allows people to talk at a normal volume which improves the dining experience. It seems to me that the same benefits apply to server rooms and offices, with the benefit being more efficient work. I found it exciting when I first had my desk in a server room (surrounded by tens of millions of pounds worth of computer gear). But as I got older I found it less interesting to work in that type of environment just as I found it less interesting to have dinner in a noisy bar - and for the same reasons. For a server room there is no escaping the fact that it will be noisy. But if the noise can be minimised then it will allow better communication between the people who are there and less distraction which should result in higher quality of work - which matters if you want good uptime! One thing I have observed is that physically larger servers tend to make less noise per volume and per compute power. For example a 2RU server with four CPUs seems to always make less noise than two 1RU servers that each have two CPUs. I believe that this is because a fan with a larger diameter can operate at a lower rotational speed which results in less bearing noise and the larger fans also give less turbulence. While it’s obvious that using fewer servers via virtualisation has the potential to avoid noise (both directly through fans and disks and indirectly through the cooling system for the server room [2]). A less obvious way of reducing noise is to swap two 1RU servers for one 2RU server - although my experience is that for machines in a similar price band, a 2RU server often has comparable compute power (in terms of RAM and disk capacity) to three or four 1RU servers. To reduce noise both directly and indirectly it is a requirement to increase disk IO capacity (in terms of the number of random IOs per second) without increasing the number of spindles (disks). I just read an interesting Sun blog covering some concepts related to using Solid State Disks (SSDs) on ZFS for best performance [3]. It seems that using such techniques is one way of significantly increasing the IO capacity per server (and thus allowing more virtual servers on one physical machine) - it’s a pity that we currently don’t have access to ZFS or a similar filesystem for Linux servers (ZFS has license issues and the GPL alternatives are all in a beta state AFAIK). Another possibility that seems to have some potential is the use of NetApp Filers [4] for the main storage of virtual machines. A NetApp Filer gives a better ratio of IO requests per second to the number of spindles used than most storage array products due to the way they use NVRAM caching and their advanced filesystem features (which also incidentally gives some good options for backups and for detecting and correcting errors). So a set of 2RU servers that have the maximum amount of RAM installed and which use a NetApp Filer (or two if you want redundancy) for the storage with the greatest performance requirements should give the greatest density of virtual machines. Blade servers also have potential to reduce noise in the server room. The most significant way that they do this is by reducing the number of power supplies, instead of having one PSU per server (or two if you want redundancy) you might have three or five PSUs for a blade enclosure that has 8 or more blades. HP blade enclosures support shutting down some PSUs when the blades are idling and don’t need much power (I don’t know whether blade enclosures from other vendors do this - I expect that some do). A bigger problem however is the noise in offices where people work. It seems that the major responsible for this is the cheap cubicles that are used in most offices (and almost all computer companies). More expensive cubicles that are at almost head-height (for someone who is standing) and which have a cloth surface absorb sound better significantly improve the office environment, and separate offices are better still. One thing I would like to see is more use of shared desktop computers, it’s not difficult to set up a desktop machine with multiple video cards, so with appropriate software support (which is really difficult) you could have one desktop machine for two, or even four users which would save electricity and reduce noise. Better quality carpet on the floors would also be a good thing. While office carpet wears out fast adding some underlay would not increase the long-term cost (it can remain as the top layer gets replaced). Better windows in offices are necessary to provide a quiet working environment. The use of double-glazed windows with reflective plastic film significantly decreases the amount of heating and cooling that is required in the office. This would permit a lower speed of air flow for heating and cooling which means less noise. Also an office in a central city area will have a noise problem outside the building, again double (or even triple) glazed windows help a lot. Some people seem to believe that an operations room should have no obstacles (one ops room where I once worked had all desks facing a set of large screens that displayed network statistics and the desks were like school desks with no dividers), I think that even for an ops room there should be some effort made to reduce the ambient noise. If the room is generally reasonably quiet then it should be easy to shout the news of an outage so that everyone can hear it. Let’s assume for the sake of discussion that a quieter working environment can increase productivity by 5% (I think this is a conservative assumption). For an office full of skilled people who are doing computer work the average salary may be about $70,000, and it’s widely regarded that to factor in the management costs etc you should double the salary - so the average cost of an employee would be about $140,000. If there are 50 people in the office then the work of those employees has a cost of $7,000,000 per annum. A 5% increase in that would be worth $350,000 per annum - you could buy a lot of windows for that!

• [1] http://www.michelinguide.com/


• [2] http://etbe.coker.com.au/2008/09/10/efficiency-of-cooling-servers/


• [3] http://blogs.sun.com/studler/entry/why_you_should_avoid_placing


• [4] http://en.wikipedia.org/wiki/Netapp

Share This

I have previously written about the execmod permission check in SE Linux [1] and in a post about SE Linux on the desktop I linked to some bug reports about it [2] (which probably won’t be fixed in Debian). One thing I didn’t mention is the proof of the implication of this. When running a program in the unconfined_t domain on a SE Linux system (the domain for login sessions on a default configuration), if you set the boolean allow_execmod then the following four tests from paxtest will be listed as vulnerable: Executable bss (mprotect)
Executable data (mprotect)
Executable shared library bss (mprotect)
Executable shared library data (mprotect) This means that if you have a single shared object which uses text relocations and therefore needs the execmod permission then the range of possible vectors for attack against bugs in the application has just increased by four. This doesn’t necessarily require that the library in question is actually being used either! If a program is linked against many shared objects that it might use, then even if it is not going to ever use the library in question it will still need execmod access to start and thus give extra possibilities to the attacker. For reference when comparing a Debian system that doesn’t run SE Linux (or has SE Linux in permissive mode) to a SE Linux system with execmod enabled the following tests fail (are reported as vulnerable): Executable anonymous mapping (mprotect)
Executable heap (mprotect)
Executable stack (mprotect)
Writable text segments If you set the booleans allow_execstack and allow_execheap then you lose those protections. But if you use the default settings of all three booleans then a program running in the unconfined_t domain will be protected against 8 different memory based attacks. Based on discussions with other programmers I get the impression that fixing all the execmod issues on i386 is not going to be possible. The desire for a 15% performance boost (the expected result of using an extra register) is greater than the desire for secure systems among the people who matter most (the developers). Of course we could solve some of these issues by using statically linked programs and have statically linked versions of the libraries in question which can use the extra register without any issues. This does of course mean that updates to the libraries (including security updates) will require rebuilding the statically linked applications in question - if a rebuild was missed then this could be reduce the security of the system. To totally resolve that issue we need to have i386 machines (the cause of the problem due to their lack of registers) go away. Fortunately in the mainstream server, desktop, and laptop markets that is already pretty much done. I’m still running a bunch of P3 servers (and I know many other people who have similar servers), but they are not used for tasks that involve running programs that are partially written in assembly code (mplayer etc). One problem is that there are still new machines being released with the i386 ISA as the only instruction set. For example the AMD Geode CPU [2] is used by the One Laptop Per Child (OLPC) project [3] and the new Intel Atom line of CPUs [4] apparently only supports the AMD64 ISA on the “desktop” models and the versions used in ultra-mobile PCs are i386 only. I think that these issues are particularly difficult in regard to the OLPC. It’s usually not difficult to run “yum upgrade” or “apt-get dist-upgrade” on an EeePC or similar ultra-mobile PC. But getting an OLPC machine upgraded in some of the remote places where they may be deployed might be more difficult. Past experience has shown that viruses and trojans can get to machines that are supposed to be on isolated networks, so it seems that malware can get access to machines that can not communicate with servers that contain security patches… One mitigating factor however is that the OLPC OS is based on Fedora, and Fedora seems to be taking the strongest efforts to improve security of any mainstream distribution, a choice between 15% performance and security seems to be a no-brainer for the Fedora developers.

• [1] http://etbe.coker.com.au/2007/02/10/execmod/


• [2] http://en.wikipedia.org/wiki/AMD_Geode


• [3] http://en.wikipedia.org/wiki/One_Laptop_per_Child


• [4] http://en.wikipedia.org/wiki/Intel_Atom

Share This

My primary e-mail domain deuxchevaux.org was offline for one and a half week due to my (former) domain provider Korypet went bankrupt and probably hadn’t paid the registrar’s invoice for deuxchevaux.org although I had paid Korypet’s invoice. (The domain wasn’t expired. Its expiration date was somewhen in 2009.) Korypet itself just told its customers (and told it only on request) that they a) will shut down their business because it doesn’t pay off and that b) their service won’t be as good as usual since their head is currently in hospital. I should have noticed that there is something fishy and some details were missing at lastest when they told customers (also only on request) that they have to shut down some services earlier than announced because one of their providers has terminated their agreement with Korypet at short notice. I also should have expected that they couldn’t keep their promise to continue domains and DNS until May 2009 as they couldn’t keep other promised grace periods. I’m quite sad about how Korypet went down and how bad blood they caused — not only for me — since they offered not only good services for money but also had really good (and personal) support. But I guess that this quality at very low prices was also one of the reasons why they couldn’t keep up their business for longer. I had one of their UML based, low-end virtual “VD” servers for about six years, and used it as secondary DNS server and IRC client host. I also often thought about getting a second one at some other geographical location (they offered virtual servers in three German cities). That was the reason why I started to move my domains to them a while ago. They even could fully answer me domain registration questions for which the eDNS and the Hetzner staff only had partial and therefore confusing answers. Since it was the easiest way, I tried to transfer the domains I had registered with Korypet to the B2C division of their registrar Key Systems, to DomainDiscount24. The web interface isn’t as bad as the name “DomainDiscount24” suggests, but they seem to have communication problems if – as in this case – more than one of their business divisions are involved. It took at least two phone calls until they did the all the necessary things to get the domains transferred from Korypet’s business customer account to my newly created end customer account although they have a special form for former Korypet customers. Domains I registered after Korypet stopped accepting new domains (well, they just didn’t answer on my domain registration requests anymore) I have registered via eDNS which was a recommendation form someone on the DaLUG mailing list. My conclusion after these hassles: No more domain registration resellers, only directly interacting with registrars. Never register all domains at the same company. Have more than two DNS servers at different hosters. Don’t have domain registrations and DNS servers at the same company. So I now register my domains either at DomainDiscount24 or at eDNS. And .ch domains of course directly at SWITCH. (If only all domain registrations would be so simple and uncomplicated as with SWITCH!) And my DNS servers are currently hosted at x encon in Hannover (Yes, it’s a Xen DomU :-) and Hetzner in Erlangen. And a third one in Switzerland is already planned.

I apply (2007-07-26)
I am advocated (2007-07-26)
Mail from FD(myon) (2007-08-12)
I reply (2007-08-15)
FD(myon) says OK (2007-08-17)
AM assigned (2007-12-05)
First mails from my AM(white) (2007-12-06)
ID check done (2007-12-07)
P&P1 check done (2007-12-21)
P&P2 check done (2008-01-17)
T&S1 check done (2008-05-04)
T&S2 check done (2008-05-13)
AM report (2008-05-13)
FD(wouter) check done (2008-07-23)
DAM(myon) approval (2008-09-03)
Account created by DSA(weasel) (2008-09-08)
Debian Beer/Pastis Party in Marseille (2008-09-15) Thanks to all Debian people in particular opal, lmamane, white, madcoder, myon, wouter and weasel.

Ultimate Debian Database (UDD) is a GSoC project (now finished) which aimed at importing all different data sources that we have in Debian in a single SQL database, to make it easy to combine that data. Currently, we import information about source and binary packages, all bugs (both archived and unarchived), lintian, carnivore, popcon, history of uploads, history of migrations to testing, and orphaned packages. The goal is really to have that data, without really thinking of a specific use cases: there will be lots of use cases. Buildstat is a project by Gon ri Le Bouder, that provides a framework for running QA tests (rebuilds and lintian currently, but buildstat is built in a very extensible way) on packages, using both packages in the archive, and packages in the VCS repositories of teams. This is pretty cool: it allows teams to get an overview of the status of their packages, not using the archive as reference, but using their VCS. Buildstat schedules and runs the tests, store the data in an SQL database, and allows to browse the data using a web interface. Buildstat also import some data from other sources (only the BTS currently, using the LDAP dump) to display it on the web interface. Since both projects are using an SQL DB, people have been asking why we don’t simply merge them. The big advantage would be that the data is synchronized: buildstat would display more up-to-date info about the data sources it doesn’t generate locally (like bugs), and UDD would get fresh buildstat data. We have been talking a lot with Gon ri, considering the different possibilities. But I don’t think it’s a good idea. I think that both projects should try to do one thing, but do it very well, instead of trying to fix the world. UDD focuses on importing data that exists elsewhere. Sometimes it means doing some complex processing. But data should be be generated by UDD. Merging the projects would mean having a very big piece of software that does everything (or tries to do everything). Both databases were designed differently, with different goals. UDD tries to stay close to the data it imports. There might be some incoherences in the data sources, but that’s fine: one of the goal of UDD is to make it easy to find them (and fix them), so we need them in the DB. In buildstat, since the goal of importing data is to display it on the web interface, with a strict use case, you can freely “simplify” data if it helps. Another big difference is that UDD is designed to be easy to use (ie write and run queries) by a human user: UDD uses multi-column primary keys, while buildstat uses surrogate keys (integer “id” keys), that ORM tools usually require. There are also more technical concerns: currently UDD makes a compromise, for each data source, on what it imports: it tries to import the data that is useful, not all the data available. Merging buildstat and UDD would mean increasing the DB size significantly, by adding all the “private” data that buildstat needs. Another problem is the stable API problem: if buildstat and UDD are merged, it means that buildstat cannot change its DB schema without making sure that it wouldn’t break what UDD users are doing. So, what should we do, from my POV, instead of merging? - Continue to talk, and get Gon ri into the UDD “team”. He gathered a lot of experience working on buildstat, and he probably would be able to help a lot. - Data that is not generated by buildstat (bugs data) should be imported from UDD. Doing SQL->SQL will probably make things easier there. - A summary of buildstat’s infos should be imported into UDD. There’s also the issue of providing the data through a web interface, to the DDs, which buildstat tries to address partially. The Debian Developer Packages Overview (DDPO)’s main limitations are: - lack of knowledge about VCS (what buildstat solves) - lack of knowledge about complex organizations (you can’t get any list of packages, or list of packages maintained by teams not using a consistant Maintainer/Uploaders scheme, or list of packages from tasks, etc) - poor handling of large amount of packages. In teams with lots of packages, it’s useful to have restricted views, such as “packages outdated compared to upstream”, “packages which have bugs/RC bugs”, “packages which are newer in the VCS than in the archive”. The perl team’s work on PET clearly shows the kind of things that are needed. At this point, I think that DDPO would benefit from a full rewrite (using the existing code as a source of inspiration, and making sure that there are no regressions, of course). Using UDD as the data source, it should be easy to get something done (even if UDD still lacks some of the data DDPO has currently). But it still requires web developer skills, which I don’t have. If you are interested, contact me!

One thing I had wondered was why home air-conditioning systems are more efficient than air-conditioning systems for server rooms. I received some advice on this matter from the manager of a small server room (which houses about 30 racks of very powerful and power hungry servers). The first issue is terminology, the efficiency of a “chiller” is regarded as the number of Watts of heat energy removed divided by the number of Watts of electricity consumed by the chiller. For example when using a 200% efficient air cooling plant, a 100W light bulb is rated as being a 150W heat source. 100W to Heat it, 50W from the cooling plant to cool it. For domestic cooling I believe that 300% is fairly common for modern “split systems” (it’s the specifications for the air-conditioning on my house and the other air-conditioners on display had similar ratings). For high-density server rooms with free air cooling I have been told that a typical efficiency range is between 80% and 110%! So it’s possible to use MORE electricity on cooling than on running the servers! One difficulty in cooling a server room is that the air often can’t flow freely (unlike a big open space such as the lounge room of your house). Another is the range of temperatures and the density of heat production in some parts (a 2RU server can dissipate 1000W of heat in a small space). These factors can be minimised by extracting hot air at the top and/or rear of racks and forcing cold air in the bottom and/or the front and by being very careful when planning where to place equipment. HP offers some services related to designing a server room to increase cooling efficiency, one of the services is using computational fluid dynamics to simulate the air-flow in the server-room [1]! CFD is difficult and expensive (the complete package from HP for a small server room costs more than some new cars), I believe that the fact that it is necessary for correct operation of some server rooms is an indication of the difficulty of the problem. The most effective ways of cooling servers involve tight coupling of chillers and servers. This often means using chilled water or another liquid to extract the heat. Chilled water refrigeration systems largely remove the problem of being unable to extract the heat from the right places, but instead you have some inefficiency in pumping the water and the servers are fixed in place. I have not seen or heard of chilled water being used for 2RU servers (I’m not saying that it doesn’t get used or that it wouldn’t make sense - merely that I haven’t seen it). When installing smaller servers (2RU and below) there is often a desire to move them and attaching a chilled-water cooling system would make such a move more difficult and expensive. When a server weighs a ton or more then you aren’t going to move it in a hurry (big servers have to be mostly disassembled before the shell can be moved, and the shell might require the efforts of four men to move it). Another issue related to water cooling is the weight. Managing a moderate amount of water involves a lot of heavy pipes (a leak would be really bad) and the water itself can weigh a lot. A server room that is based around 20Kg servers might have some issues with the extra weight of water cooling (particularly the older rooms), but a server room designed for a single rack that weighs a ton can probably cope. I have been told that the cooling systems for low density server rooms are typically as efficient as those used for houses, and may even be more efficient. I expect that when designing an air-conditioner the engineering trade-offs when designing for home use favor low purchase price. But someone who approves the purchase of an industrial cooling system will be more concerned about the overall cost of operations and will be prepared to spend some extra money up-front and recover it over the course of a few years. The fact that server rooms run 24*7 also gives more opportunity to recover the money spent on the purchase (my home A-C system runs for about 3 months a year for considerably less than 24 hours a day). So it seems that the way to cool servers efficiently is to have low density server rooms (to the largest extent possible). One step towards this goal would be to have servers nearer the end users. For example having workgroup servers near the workgroup (instead of in the server room). Of course physical security of those servers would be more challenging - but if all the users have regular desktop PCs that can be easily 0wned then having the server for them in the same room probably doesn’t make things any worse. Modern tower servers are more powerful than rack mounted servers that were available a few years ago while also being very quiet. A typical rack-mounted server is not something you would want near your desk, but one of the quiet tower servers works quite well.

• [1] http://en.wikipedia.org/wiki/Computational_fluid_dynamics

Share This

As some might know, I've been working at the university for a few weeks now. I'm very happy how it all worked out, because of the people there and the subjects to work on. It's just great to think through some mind-bending index structure to improve reverse-k-nearest-neighbor queries. And the professor (who is very well known for his work on R*-Trees, the X-Tree, and the DBSCAN and OPTICS algorithms) really manages to aggregate excellent people.(rkNN is the problem of finding those points for which a Nearest Neighbor Search [Wikipedia] would return the given point. There are many use cases for example in location based services.)A lot of my work involves a framework which has been published recently on the SSDBM 2008 conference: ELKI - Environment for DeveLoping KDD-Applications Supported by Index-StructuresThe goal of ELKI is to become what WEKA is for machine learning. And it's well on it's way for that.ELKI aims at the data mining researcher, it's not designed to squeeze out the last bit of performance. Instead it allows you to compare different algorithms, try different index structures (including various spatial and metrical trees) and offers all kinds of functionality you can reuse. If you are planning to do some real world applications, it likely is still useful for prototyping.In the last weeks I've already put quite some effort in ELKI, so expect the next version to be released with significant changes and new functionality. In fact some working functionality was just left out of the release because we hadn't cleaned up the code yet.Currently, getting started with ELKI can be a bit tricky. It's not hard to use, but you just need to find out where to start. The next release will therefore include a "dummy algorithm" that mostly serves as being a template for implementing custom algorithms (it can also be used for benchmarking an index structure, though).

I have released second beta of Adept a few hours ago. For the unaware, Adept is an APT front-end for KDE: it lets you install, remove, upgrade software and such your Debian and Kubuntu boxes (and maybe on some other Debian derivatives, too). Changes since Alpha 6 Now, alpha 6 has been the last alpha release, therefore there are no more new features in beta 2. Most of the issues should have been resolved. I have also added a menubar, resolving a potential usability problem (and making the about box accessible, too!). There will be another beta that should fix the few remaining known issues. Please report any issues you have encountered to me (either mail, irc or bugs.kde.org). Where to get I have prepared binary packages. These should be now part of Debian Sid (unstable) and also of Kubuntu Intrepid. You should be able to get the new version from your distribution:

apt-get install adept

There is no Hardy backport, since my time is tight. If anyone is willing to do a backport (you can use the stuff that is in my PPA; however, some things need updating), please drop me a note: I will gladly publish installation instructions. Heroes of Beta 2 There has been a (silent) beta 1 a while back, which introduced a couple of regressions against alpha 6 (but fixed bugs, too!).

• Jonathan Thomas (JontheEchidna) for catching the updater regression
• Harald Sitter (apachelogger):
  • for providing new .desktop files
  • for catching the non-root mode regression

Jobs for Beta 2 Sorting through Launchpad for bugs that have been fixed is still needed, as far as I know. JontheEchidna has been working on this (and maybe some others have? Drop me a note if you worked on this and claim your fair share of fame… Sorry that I don’t keep track of that, but you know the story… too much stuff to do in too little time, all the time.) Notifier is not included. Kubuntu has an independent sort of notifier now, I believe. Maybe I’ll find time to reconcile it in 3.1 or so.

For the last few days I’ve been in the city Badajoz in region Extremadura of Spain at the FAI developer meeting. Check out the Linex Euronews Video at youtube to get a short impression of Linux in the region. Lessions learnt:

• debcommit is great, it generates a commit message based on new text in debian/changelog (thanks for the pointer, Holger)
• rc-alert lists installed packages which have release-critical bugs, nice (again part of devscripts, I should take a whole day to take a closer look at all the nifty stuff in devscripts; thanks again, Holger)
• use caff (Debian package signing-party) for keysigning (thanks tokkee for the reminder, I always forget the name of it)
• collectd (statistics collection and monitoring daemon) seems to rock (thanks again, tokkee)
• chdk is a firmware add-on for Canon’s Digic II & Digic III cameras and really rocks (thanks for the pointer Thomas, allbest-a700-100b-50.zip works fine for my Canon Powershot A700)
• ganglia (cluster system monitoring daemon) looks nice as well (thanks Thomas)
• workrave (Repetitive Strain Injury prevention tool) seems to be perfect for me (thanks again, Thomas)

<insider>

</insider>

Two reasons:

1. 'foo FTBFS on m68k' is not RC. 'foo FTBFS on m68k because it segfaults during the build since it does bar and baz, which just happens to result in a segfault on m68k at startup but might randomly result in a segfault on every architecture, really' in most cases is.
2. It, well, is just fun to debug someone else's code. On m68k, it's easy to do that, too, if you know a bit of assembly language.

Today's (and last week's) bug: #497262. Xemacs21 segfaults at startup on m68k (and FTBFSes because of that), which is RC for m68k. Reason: confusion between a pointer and its value. In other words: the sanity check might succeed, or it might not, depending on your luck. In fact, the first time the code flows through this buggy path, it does not bomb. Once we do get through this bit, it's also very likely that regular error handling will catch it—the only reason why xemacs segfaults here is because we're still in initialization, and the error object in the lisp stuff hasn't been initialized yet, so we get a nice null pointer dereference... That alone isn't enough to make it RC everywhere, but it's good enough to make it important and fixable for the maintainer. Like I said, fun. Sorry for those of you who want to release Lenny ASAP, but I think I'll be making your life a bit harder :-> Next week's bug: a similar issue in emacs, or the build failure of ghc6 on m68k, depending on time and complexity.

Do you know this experience: A program, in my case subversion, has a bug and starts to eat memory. You can not interact with your system any more, only watch the memory and swap run full (if you have a display for that). Then it takes a while, while the kernel kills the (hopefully right) program. Things start to move again, until they are fully recovered from the swap and you can continue your work. Or the kernel does somehow not kill the right program, and you are screwed.During regular work, though, your swap is hardy ever needed. Only after a while, a few megabytes of never-used RAM is swapped out, to make space for using the RAM as a file cache.I d like to see the kernel not give out more memory to processes than there is physical memory, because that s plenty for normal work, and if there is more requested, then that s most likely wrong. But I still want the kernel to use the rest of the memory for caching files, and also move some unused RAM pages to the swap file.Unfortunately, there does not seem to be a settings that achieves this directly. But if you happen to have the swap about the same size as your RAM, then these settings, when written to /etc/sysctl.d/vm.conf, will do the job:

vm.overcommit_memory = 2
vm.overcommit_ratio = 0

The first one is to make sure the kernel does not hand out more memory than you tell it to, and the second is to make sure that it only hands out (swap size + 0 * RAM size) to processes.Beware that things go wrong if you happen to have no swap any more for some reason, beause then the kernel will hand out zero memory! Therefore, you need to make sure that these settings are applied after swap was enabled. On a Debian machine, rename /etc/rcS.d/S30procps to /etc/rcS.d/S37procps. This would not be possible if you could also specify the ration of swap memory to be used. Then I could set that to zero and the RAM ratio to 100.If anyone knows better ways to achieve this, I m interested to hear them.Update: For my qemu based armel package builder, this is not enough it seems. I m now running it with overcommit_ratio = 50.

It's the X Development Summit in Edinburgh this week, so I've been hanging out with the graphics crowd. There hasn't been a huge amount of work done in the field of power management in graphics so far - Intel have framebuffer compression and there's the lvds reclocking patch I wrote (I've cleaned this up somewhat since then, but it probably wants some refactoring to avoid increasing CPU usage based on its use of damage for update notifications). That still leaves us with some fun things to look at, though.

The most obvious issue is the gpu clock. Intel's chipset implements clock gating, where unused sections of chip automatically unclock themselves. This is pleasingly transparent to the OS, and we get power savings without any complexity we have to care about. However, there's no way to control the core clock of the GPU - it's part of the northbridge and fiddling with the clocking of that would be likely to upset things. Nvidia and Radeon hardware is more interesting in this respect, since we can control the gpu clock independently of anything else. The problem is trying to do so in a reasonable way.

In an ideal universe, we can change these clock frequencies quickly and without any visual artifacts. That way it's possible to leave it in the lowest state and clock it up as load develops. There's a couple of problems with this - non ideal hardware, and the software in the first place. Jerome's been testing a little on Radeon and discovered that changing the memory clock through Atom results in visual corruption. It's conceivable that this is due to some memory transaction cycles getting corrupted as the clock gets changed. If we could ensure that the reclock happens during the vertical blank interval, that's something that could potentially be avoided (of course, then we have the entertainment of working out when the vertical blank interval actually is when you have a dual head output...). The other problem is that 3D software tends to consume as many resources as are available. Games will produce as many frames per second as possible. Demand-based clocking will simply ramp the gpu to full speed in that situation, which isn't necessarily what you want in the battery case (as the number of frames per second goes up, so does the cpu usage - even more power draw) but is probably pretty reasonable in the powered case.

Handwavy testing suggests that this can save a pretty significant amount of power, so it's something that I'm planning on working on. Further optimisations include things like making sure that we're not running any PLLs that aren't being used at the time (oscillators chew power), not powering up output ports when you're not outputting to them and enabling any hardware-level features that we're currently ignoring. And, ideally, doing all of this without causing the machine to hang on a regular basis.

I am surely not the only one to complain about the ridiculous liquids restrictions for airplane travel. Since these new regulation are in place, I ve challenged them, found holes, and compiled tips for those trying to blow up planes. I wouldn t write another story if it weren t for an idea I ve had at Zurich airport on my last trip to London: let s swamp the airports with trash so that they ll be forced to deal with Brussels and IATA to return to normal. After checking in for my flight, I stopped by the supermarket to buy two containers of yoghurt that would make someone happy. The containers each said 150g (that s weight, not volume) on them, and I put them into a clear, resealable one-litre bag, placed them into a tray to be x-ray-scanned separately, only to have them confiscated. After discovering (not much to my surprise) that the security staff didn t know the difference between weight and volume, nor understood the concept of density, I got a chance to speak to the head security officer (surrounded by five police whose attention I d gotten), and learnt that Zurich airport has one ton of trash to discard every day, Frankfurt supposedly has to deal with four. As I was walking onto my plane, I tried to think of non-recyclable containers that we could fill with liquids to bring along to increase that amount. My theory was that once the trash problem became too massive, the airports would have to deal with the authorities to resolve this liquid restriction, because it seems quite clear that normal people have no way to influence choices made that affect our safety . Unfortunately, I see two problems: First, we d be dealing with trash and hence face all the environmental concerns. The airports do not recycle the millions of PET bottles they confiscate every day, so we shouldn t make that worse. Unfortunately, I cannot think of another liquid container that wouldn t come with similar concerns. Second, the airports might have the burden, but they won t carry the cost of all the trash. In fact, thanks to the security theatre related to liquids, we already pay higher airport taxes and charges. Surely it can t be in our interest to push that further up the scale. So in the end, swamping airports with trash doesn t seem like a viable way forward, unfortunately. I wish I knew what to do. I wish that the decision makers at IATA would finally admit that they overreacted and revert to normal, with sensible security measures, which focus on fending off the real threats, not fake ones. Unfortunately, nobody likes to admit that they were wrong, especially not when the decision is heavily backed up by the lobby of vending machine companies and restaurant owners, who benefit greatly from these ridiculous liquid safety measures. NP: Pulp: We Love Life