Planet Debian

Subscribe to Planet Debian feed
Planet Debian - https://planet.debian.org/
Updated: 1 hour 32 min ago

Dirk Eddelbuettel: pkgKitten 0.2.0: Now with tinytest and new docs

27 September, 2020 - 21:13

A new release 0.2.0 of pkgKitten just hit on CRAN today, or about eleven months after the previous release.

This release brings support for tinytest by having pkgKitten::kitten() automagically call tinytest::puppy() if the latter package is installed (and the user did not opt out of calling it). So your newly created minimal package now also uses a wonderful yet tiny testing framework. We also added a new documentation site using the previously tweeted-about wrapper for Material for MkDocs I really dig. And last but not least we switched to BSPM-based Continued Integration (which I wrote about yesterday in R4 #30) and fixed one bug regarding the default NAMESPACE file.

Changes in version 0.2.0 (2020-09-27)
  • Continuous Integration uses the updated BSPM-based script on Travis and with GitHub Actions (Dirk in #11 plus earlier commits).

  • A new default NAMESPACE file is now installed (Dirk in #12).

  • A package documentation website was added (Dirk in #13).

  • Call tinytest::puppy if installed and not opted out (Dirk in #14).

More details about the package are at the pkgKitten webpage, the (new) pkgKitten docs site, and the pkgKitten GitHub repo.

Courtesy of my CRANberries site, there is also a diffstat report for this release.

If you like this or other open-source work I do, you can now sponsor me at GitHub. For the first year, GitHub will match your contributions.

This post by Dirk Eddelbuettel originated on his Thinking inside the box blog. Please report excessive re-aggregation in third-party for-profit settings.

Andrew Cater: Final post from media team for the day - most of the ordinary images and live images have been tested

27 September, 2020 - 04:51

 Winding down slightly - we've worked our way through most of the images and testing. Schweer tested all of the Debian Edu/Skolelinux images for which many thanks

Sledge, RattusRattus, Isy and I have been working pretty much solidly for 10 3/4 hours. There's still some images to build - mips, mipsel and s390x but these are all images that we don't have hardware to test on particularly.

Another good and useful day - bits and pieces done throughout. NOTE: There appear to have been some security updates since the main release this morning so, as ever, it's worth updating machines on a regular basis.

Waiting for the final images to finish building so that we can check the archive for completeness and then publish to the media mirrors. All the best until next time: thanks as ever to Sledge for his invaluable help. See you again in a couple of months in all likelihood. 

A much smaller release: some time in the next month we hope to be able to build and test an Alpha release for Bullseye. Bullseye is likely to be released somewhere round the middle of next year so we'll have additional Buster stable point releases in the meantime.

François Marier: Repairing a corrupt ext4 root partition

27 September, 2020 - 02:45

I ran into filesystem corruption (ext4) on the root partition of my backup server which caused it to go into read-only mode. Since it's the root partition, it's not possible to unmount it and repair it while it's running. Normally I would boot from an Ubuntu live CD / USB stick, but in this case the machine is using the mipsel architecture and so that's not an option.

Repair using a USB enclosure

I had to pull the shutdown the server and then pull the SSD drive out. I then moved it to an external USB enclosure and connected it to my laptop.

I started with an automatic filesystem repair:

fsck.ext4 -pf /dev/sde2

which failed for some reason and so I moved to an interactive repair:

fsck.ext4 -f /dev/sde2

Once all of the errors were fixed, I ran a full surface scan to update the list of bad blocks:

fsck.ext4 -c /dev/sde2

Finally, I forced another check to make sure that everything was fixed at the filesystem level:

fsck.ext4 -f /dev/sde2
Fix invalid alternate GPT

The other thing I noticed is this messge in my dmesg log:

scsi 8:0:0:0: Direct-Access     KINGSTON  SA400S37120     SBFK PQ: 0 ANSI: 6
sd 8:0:0:0: Attached scsi generic sg4 type 0
sd 8:0:0:0: [sde] 234441644 512-byte logical blocks: (120 GB/112 GiB)
sd 8:0:0:0: [sde] Write Protect is off
sd 8:0:0:0: [sde] Mode Sense: 31 00 00 00
sd 8:0:0:0: [sde] Write cache: enabled, read cache: enabled, doesn't support DPO or FUA
sd 8:0:0:0: [sde] Optimal transfer size 33553920 bytes
Alternate GPT is invalid, using primary GPT.
 sde: sde1 sde2

I therefore checked to see if the partition table looked fine and got the following:

$ fdisk -l /dev/sde
GPT PMBR size mismatch (234441643 != 234441647) will be corrected by write.
The backup GPT table is not on the end of the device. This problem will be corrected by write.
Disk /dev/sde: 111.8 GiB, 120034123776 bytes, 234441648 sectors
Disk model: KINGSTON SA400S3
Units: sectors of 1 * 512 = 512 bytes
Sector size (logical/physical): 512 bytes / 512 bytes
I/O size (minimum/optimal): 512 bytes / 512 bytes
Disklabel type: gpt
Disk identifier: 799CD830-526B-42CE-8EE7-8C94EF098D46

Device       Start       End   Sectors   Size Type
/dev/sde1     2048   8390655   8388608     4G Linux swap
/dev/sde2  8390656 234441614 226050959 107.8G Linux filesystem

It turns out that all I had to do, since only the backup / alternate GPT partition table was corrupt and the primary one was fine, was to re-write the partition table:

$ fdisk /dev/sde

Welcome to fdisk (util-linux 2.33.1).
Changes will remain in memory only, until you decide to write them.
Be careful before using the write command.

GPT PMBR size mismatch (234441643 != 234441647) will be corrected by write.
The backup GPT table is not on the end of the device. This problem will be corrected by write.

Command (m for help): w

The partition table has been altered.
Syncing disks.
Run SMART checks

Since I still didn't know what caused the filesystem corruption in the first place, I decided to do one last check: SMART errors.

I couldn't do this via the USB enclosure since the SMART commands aren't forwarded to the drive and so I popped the drive back into the backup server and booted it up.

First, I checked whether any SMART errors had been reported using smartmontools:

smartctl -a /dev/sda

That didn't show any errors and so I kicked off an extended test:

smartctl -t long /dev/sda

which ran for 30 minutes and then passed without any errors.

The mystery remains unsolved.

Dirk Eddelbuettel: #30: Easy, Reliable, Fast and Portable Linux and macOS Continuous Integration

26 September, 2020 - 23:46

Welcome to the 30th post in the rarified R recommendation resources series or R4 for short. The last post introduced BSPM. In the four weeks since, we have worked some more on BSPM to bring it to the point where it is ready for use with continuous integration. Building on this, it is now used inside the run.sh script that driven our CI use for many years (via the r-travis repo).

Which we actually use right now on three different platforms:

All three use the exact same script facilitating this, and run a ‘matrix’ over Linux and macOS. You read this right: one CI setup that is portable and which you can take to your CI provider of choice. No lock-in or tie-in. Use what works, change at will. Or run on all three if you like burning extra cycles.

This is already used by handful of my repos as well as by at least two repos of friends also deploying r-travis. How does it work? In a nutshell we are

  • downloading run.sh via curl and changing its mode;
  • running run.sh bootstrap which sets the operating system default:
    • on Linux we use Ubuntu,
      • add two PPAs repos for R itself and over 4600 r-cran-* binaries,
      • and enable BSPM to use these from install.packages()
    • on macOS we use the standard setup also used on Travis, GitHub Actions and elsewhere;
    • this provides us with fast, reliable, easy, and portable access to binaries on two OSs under dependency resolution;
  • running run.sh install_deps to install just the requireded Depends:, Imports: and LinkingTo:
  • running run.sh tests to build the tarball and test it via R CMD check --as-cran.

There are several customizations that are possible via environment variables

  • additional PPAs or drat repos can be added to offer even more package choice;
  • alternatively one could run run.sh install_all to also install Suggests:;
  • optionally one could run run.sh install_r pkgA pkgB ... to install packages explicitly listed;
  • optionally one could also run run.sh install_aptget r-cran-pkga r-cran-pkgb otherpackage to add more Ubuntu binaries.

We find this setup compelling. The scheme is simple: there really is just one shell script behind it which can also be downloaded and altered. The scheme is also portable as we can (as shown) rotate between CI provides. The scheme is also more flexible: in case of debugging needs one can simply run the script on a local Docker or VM instance. Lastly, the scheme moves away from single points of failure or breakage.

Currently the script uses only BSPM as I had the hunch that it would a) work and b) be compelling. Adding support for RSPM would be equally easy, but I have no immediate need to do so. Adding BioConductor installation may be next. That is easy when BioConductor uses r-release; it may be little more challenging under r-devel to but it should work too. Stay tuned.

In the meantime, if the above sounds compelling, give run.sh from r-travis a go!

If you like this or other open-source work I do, you can now sponsor me at GitHub. For the first year, GitHub will match your contributions.

This post by Dirk Eddelbuettel originated on his Thinking inside the box blog. Please report excessive re-aggregation in third-party for-profit settings.

Andrew Cater: Chunking through the tests for various media images ...

26 September, 2020 - 22:47

We're working our way through some of the CD/DVD/Blu-Ray media images, doing test installs, noting failures and so on. It's repetitive work but vital if we're going to provide some assurance that folk can install from the images we make. 

There's always the few things that catch us out and there's always something to note for next time. Schweer has joined us and is busy chasing down debian-edu/Skolelinux installs from Germany. We're getting there, one way and another, and significantly ahead of where we were last time around when the gremlins got in and delayed us. All good :)


Andrew Cater: There are things that money can't buy - and sensible Debian colleagues are worth gold and diamonds :)

26 September, 2020 - 20:13

 Participating in the Debian media testing on debian-cd. One of my colleagues has just spent time to sort out an email issue having spent a couple of hours with me the other night. I now have good, working email for the first time in years - I can't value that highly enough.

Sledge, RattusRattus, Isy and myself are all engaged in testing various CD images. At the same time, they're debugging a new application to save us from wiki problems when we do this - and we're also able to use a video link which is really handy to chat backwards and forwards and means I can sit virtually in Cambridge :)

Lots of backchat and messages flying backwards and forwards - couldn't wish for a better way to spend an afternoon with friends.



Andrew Cater: There's a Debian point release for Debian stable happening this weekend - 10.6

26 September, 2020 - 17:58

 Nothing particularly new or unexpected: there's a point release happening at some point this weekend for Debian stable. Usual rules apply: if you've already got a system current and up to date, there's not much to do but the base files version will change at some point to reflect 10.6 when you next update. 

If you have media from 10.5, you may not _have_ to go and get media this weekend but it's always useful to get new media in due course.

This point release will contain security fixes, consequent changes etc. as usual - it is always good and useful to keep machines up to date.

Working with the CD team to eventually test, build and release CD / DVD images and media as and when files gradually become available. As ever, this may take 12-16 hours. As ever, I'll post some blog entries as we go.

Currently "sitting in Cambridge" with Sledge, RattusRattus and Isy who are all involved in the testing and we'll have a great day, as ever.


Russell Coker: Bandwidth for Video Conferencing

25 September, 2020 - 22:44

For the Linux Users of Victoria (LUV) I’ve run video conferences on Jitsi and BBB (see my previous post about BBB vs Jitsi [1]). One issue with video conferences is the bandwidth requirements.

The place I’m hosting my video conference server has a NBN link with allegedly 40Mb/s transmission speed and 100Mb/s reception speed. My tests show that it can transmit at about 37Mb/s and receive at speeds significantly higher than that but also quite a bit lower than 100Mb/s (around 60 or 70Mb/s). For a video conference server you have a small number of sources of video and audio and a larger number of targets as usually most people will have their microphones muted and video cameras turned off. This means that the transmission speed is the bottleneck. In every test the reception speed was well below half the transmission speed, so the tests confirmed my expectation that transmission was the only bottleneck, but the reception speed was higher than I had expected.

When we tested bandwidth use the maximum upload speed we saw was about 4MB/s (32Mb/s) with 8+ video cameras and maybe 20 people seeing some of the video (with a bit of lag). We used 3.5MB/s (28Mb/s) when we only had 6 cameras which seemed to be the maximum for good performance.

In another test run we had 4 people all sending video and the transmission speed was about 260KB/s.

I don’t know how BBB manages the small versions of video streams. It might reduce the bandwidth when the display window is smaller.

I don’t know the resolutions of the cameras. When you start sending video in BBB you are prompted for the “quality” with “medium” being default. I don’t know how different camera hardware and different choices about “quality” affect bandwidth.

These tests showed that for the cameras we had available a small group of people video chatting a 100/40 NBN link (the fastest Internet link in Australia that’s not really expensive) a small group of people can be all sending video or a medium size group of people can watch video streams from a small group.

For meetings of the typical size of LUV meetings we won’t have a bandwidth problem.

There is one common case that I haven’t yet tested, where there is a single video stream that many people are watching. If 4 people are all sending video with 260KB/s transmission bandwidth then 1 person could probably send video to 4 for 65KB/s. Doing some simple calculations on those numbers implies that we could have 1 person sending video to 240 people without running out of bandwidth. I really doubt that would work, but further testing is needed.

Related posts:

  1. Solving Rubik’s Cube and IO Bandwidth Solving Rubiks Cube by treating disk as RAM: Gene Cooperman...
  2. Digital Video Cameras I’ve just done some quick research on Digital Video Cameras...
  3. BBB vs Jitsi I previously wrote about how I installed the Jitsi video-conferencing...

Colin Watson: Porting Launchpad to Python 3: progress report

25 September, 2020 - 18:01

Launchpad still requires Python 2, which in 2020 is a bit of a problem. Unlike a lot of the rest of 2020, though, there’s good reason to be optimistic about progress.

I’ve been porting Python 2 code to Python 3 on and off for a long time, from back when I was on the Ubuntu Foundations team and maintaining things like the Ubiquity installer. When I moved to Launchpad in 2015 it was certainly on my mind that this was a large body of code still stuck on Python 2. One option would have been to just accept that and leave it as it is, maybe doing more backporting work over time as support for Python 2 fades away. I’ve long been of the opinion that this would doom Launchpad to being unmaintainable in the long run, and since I genuinely love working on Launchpad - I find it an incredibly rewarding project - this wasn’t something I was willing to accept. We’re already seeing some of our important dependencies dropping support for Python 2, which is perfectly reasonable on their terms but which is starting to become a genuine obstacle to delivering important features when we need new features from newer versions of those dependencies. It also looks as though it may be difficult for us to run on Ubuntu 20.04 LTS (we’re currently on 16.04, with an upgrade to 18.04 in progress) as long as we still require Python 2, since we have some system dependencies that 20.04 no longer provides. And then there are exciting new features like type hints and async/await that we’d like to be able to use.

However, until last year there were so many blockers that even considering a port was barely conceivable. What changed in 2019 was sorting out a trifecta of core dependencies. We ported our database layer, Storm. We upgraded to modern versions of our Zope Toolkit dependencies (after contributing various fixes upstream, including some substantial changes to Zope’s test runner that we’d carried as local patches for some years). And we ported our Bazaar code hosting infrastructure to Breezy. With all that in place, a port seemed more of a realistic possibility.

Still, even with this, it was never going to be a matter of just following some standard porting advice and calling it good. Launchpad has almost a million lines of Python code in its main git tree, and around 250 dependencies of which a number are quite Launchpad-specific. In a project that size, not only is following standard porting advice an extremely time-consuming task in its own right, but just about every strange corner case is going to show up somewhere. (Did you know that StringIO.StringIO(None) and io.StringIO(None) do different things even after you account for the native string vs. Unicode text difference? How about the behaviour of .union() on a subclass of frozenset?) Launchpad’s test suite is fortunately extremely thorough, but even just starting up the test suite involves importing most of the data model code, so before you can start taking advantage of it you have to make a large fraction of the codebase be at least syntactically-correct Python 3 code and use only modules that exist in Python 3 while still working in Python 2; in a project this size that turns out to be a large effort on its own, and can be quite risky in places.

Canonical’s product engineering teams work on a six-month cycle, but it just isn’t possible to cram this sort of thing into six months unless you do literally nothing else, and “please can we put all feature development on hold while we run to stand still” is a pretty tough sell to even the most understanding management. Fortunately, we’ve been able to grow the Launchpad team in the last year or so, and so it’s been possible to put “Python 3” on our roadmap in the understanding that we aren’t going to get all the way there in one cycle, while still being able to do other substantial feature development work as well.

So, with all that preamble, what have we done this cycle? We’ve taken a two-pronged approach. From one end, we identified 147 classes that needed to be ported away from some compatibility code in our database layer that was substantially less friendly to Python 3: we’ve ported 38 of those, so there’s clearly a fair bit more to do, but we were able to distribute this work out among the team quite effectively. From the other end, it was clear that it would be very inefficient to do general porting work when any attempt to even run the test suite would run straight into the same crashes in the same order, so I set myself a target of getting the test suite to start up, and started hacking on an enormous git branch that I never expected to try to land directly: instead, I felt free to commit just about anything that looked reasonable and moved things forward even if it was very rough, and every so often went back to tidy things up and cherry-pick individual commits into a form that included some kind of explanation and passed existing tests so that I could propose them for review.

This strategy has been dramatically more successful than anything I’ve tried before at this scale. So far this cycle, considering only Launchpad’s main git tree, we’ve landed 137 Python-3-relevant merge proposals for a total of 39552 lines of git diff output, keeping our existing tests passing along the way and deploying incrementally to production. We have about 27000 more lines of patch at varying degrees of quality to tidy up and merge. Our main development branch is only perhaps 10 or 20 more patches away from the test suite being able to start up, at which point we’ll be able to get a buildbot running so that multiple developers can work on this much more easily and see the effect of their work. With the full unlanded patch stack, about 75% of the test suite passes on Python 3! This still leaves a long tail of several thousand tests to figure out and fix, but it’s a much more incrementally-tractable kind of problem than where we started.

Finally: the funniest (to me) bug I’ve encountered in this effort was the one I encountered in the test runner and fixed in zopefoundation/zope.testrunner#106: IDs of failing tests were written to a pipe, so if you have a test suite that’s large enough and broken enough then eventually that pipe would reach its capacity and your test runner would just give up and hang. Pretty annoying when it meant an overnight test run didn’t give useful results, but also eloquent commentary of sorts.

Reproducible Builds: ARDC sponsors the Reproducible Builds project

25 September, 2020 - 07:00

The Reproducible Builds project is pleased to announce a donation from Amateur Radio Digital Communications (ARDC) in support of its goals. ARDC’s contribution will propel the Reproducible Builds project’s efforts in ensuring the future health, security and sustainability of our increasingly digital society.

About Amateur Radio Digital Communications (ARDC)

Amateur Radio Digital Communications (ARDC) is a non-profit that was formed to further research and experimentation with digital communications using radio, with a goal of advancing the state of the art of amateur radio and to educate radio operators in these techniques.

It does this by managing the allocation of network resources, encouraging research and experimentation with networking protocols and equipment, publishing technical articles and number of other activities to promote the public good of amateur radio and other related fields. ARDC has recently begun to contribute funding to organisations, groups, individuals and projects towards these and related goals, and their grant to the Reproducible Builds project is part of this new initiative.

Amateur radio is an entirely volunteer activity performed by knowledgeable hobbyists who have proven their ability by passing the appropriate government examinations. No remuneration is permitted. “Ham radio,” as it is also known, has proven its value in advancements of the state of the communications arts, as well as in public service during disasters and in times of emergency.

For more information about ARDC, please see their website at ampr.org.

About the Reproducible Builds project

One of the original promises of open source software was that peer review would result in greater end-user security and stability of our digital ecosystem. However, although it is theoretically possible to inspect and build the original source code in order to avoid maliciously-inserted flaws, almost all software today is distributed in prepackaged form.

This disconnect allows third-parties to compromise systems by injecting code into seemingly secure software during the build process, as well as by manipulating copies distributed from ‘app stores’ and other package repositories.

In order to address this, ‘Reproducible builds’ are a set of software development practices, ideas and tools that create an independently-verifiable path from the original source code, all the way to what is actually running on our machines. Reproducible builds can reveal the injection of backdoors introduced by the hacking of developers’ own computers, build servers and package repositories, but can also expose where volunteers or companies have been coerced into making changes via blackmail, government order, and so on.

A world without reproducible builds is a world where our digital infrastructure cannot be trusted and where online communities are slower to grow, collaborate less and are increasingly fragile. Without reproducible builds, we leave space for greater encroachments on our liberties both by individuals as well as powerful, unaccountable actors such as governments, large corporations and autocratic regimes.

The Reproducible Builds project began as a project within the Debian community, but is now working with many crucial and well-known free software projects such as Coreboot, openSUSE, OpenWrt, Tails, GNU Guix, Arch Linux, Tor, and many others. It is now an entirely Linux distribution independent effort and serves as the central ‘clearing house’ for all issues related to securing build systems and software supply chains of all kinds.

For more about the Reproducible Builds project, please see their website at reproducible-builds.org.


If you are interested in ensuring the ongoing security of the software that underpins our civilisation, and wish to sponsor the Reproducible Builds project, please reach out to the project by emailing contact@reproducible-builds.org.

Vincent Fourmond: Tutorial: analyze Km data of CODHs

23 September, 2020 - 21:41
This is the first post of a series in which we will provide the readers with simple tutorial approaches to reproduce the data analysis of some of our published papers. All our data analysis is performed using QSoas. Today, we will show you how to analyze the experiments we used to characterize the behaviour of an enzyme, the Nickel-Iron CO dehydrogenase IV from Carboxytothermus hydrogenoformans. The experiments we analyze here are described in much more details in the original publication, Domnik et al, Angewandte Chemie, 2017. The only things you need to know for now are the following:
  • the data is the evolution of the current over time given by the absorbed enzyme;
  • at a given point, a certain amount of CO (50 µM) is injected, and its concentration decreases exponentially over time;
  • the enzyme has a Michaelis-Menten behaviour with respect to CO.
This means that we expect a response of the type: $$i(t) = \frac{i_m}{1 + \frac{K_m}{[\mathrm{CO}](t)}}$$ in which $$[\mathrm{CO}](t) = \begin{cases} 0, & \text{for } t < t_0 \\ C_0 \exp \frac{t_0 - t}{\tau}, & \text{for }t\geq t_0 %> \end{cases}$$

To begin this tutorial, first download the files from the github repository (direct links: data, parameter file and ruby script). Start QSoas, go to the directory where you saved the files, load the data file, and remove spikes in the data using the following commands:

QSoas> cd
QSoas> l Km-CODH-IV.dat
QSoas> R
First fitThen, to fit the above equation to the data, the simplest is to take advantage of the time-dependent parameters features of QSoas. Run simply:
QSoas> fit-arb im/(1+km/s) /with=s:1,exp
This simply launches the fit interface to fit the exact equations above. The im/(1+km/s) is simply the translation of the Michaelis-Menten equation above, and the /with=s:1,exp specifies that s is the result of the sum of 1 exponential like for the definition of above. Then, load the Km-CODH-IV.params parameter files (using the "Parameters.../Load from file" action at the bottom, or the Ctrl+L keyboard shortcut). Your window should now look like this: To fit the data, just hit the "Fit" button ! (or Ctrl+F). Including an offset The fit is not bad, but not perfect. In particular, it is easy to see why: the current predicted by the fit goes to 0 at large times, but the actual current is below 0. We need therefore to include an offset to take this into consideration. Close the fit window, and re-run a fit, but now with this command:
QSoas> fit-arb im/(1+km/s)+io /with=s:1,exp
Notice the +io bit that corresponds to the addition of an offset current. Load again the base parameters, run the fit again... Your fit window show now look like: See how the offset current is now much better taken into account. Let's talk a bit more about the parameters:
  • im is \(i_m\), the maximal current, around 120 nA (which matches the magnitude of the original current).
  • io is the offset current, around -3nA.
  • km is the \(K_m\), expressed in the same units as s_1, the first "injected" value of s (we used 50 because the injection is 50 µM CO). That means the value of \(K_m\) determined by this fit is around 9 nM ! (but see below).
  • s_t_1 is the time of the injection of CO (it was in the parameter files you loaded).
  • Finally, s_tau_1 is the time constant of departure of CO, noted \(\tau\) in the equations above.
Taking into account mass-transport limitations However, the fit is still unsatisfactory: the predicted curve fails to reproduce the curvature at the beginning and at the end of the decrease. This is due to issues linked to mass-transport limitations, which are discussed in details in Merrouch et al, Electrochimica Acta, 2017. In short, what you need to do is to close the fit window again, load the transport.rb Ruby file that contains the definition of the itrpt function, and re-launch the fit window using:
QSoas> ruby-run transport.rb
QSoas> fit-arb itrprt(s,km,nFAm,nFAmu)+io /with=s:1,exp
Load again the parameter file... but this time you'll have to play a bit more with the starting parameters for QSoas to find the right values when you fit. Here are some tips:
  • the curve predicted with the current parameters (use "Update" to update the display) should "look like" the data;
  • apart from io, no parameter should be negative;
  • there may be hints about the correct values in the papers...
A successful fit should look like this: Here you are ! I hope you enjoyed analyzing our data, and that it will help you analyze yours ! Feel free to comment and ask for clarifications.

About QSoasQSoas is a powerful open source data analysis program that focuses on flexibility and powerful fitting capacities. It is released under the GNU General Public License. It is described in Fourmond, Anal. Chem., 2016, 88 (10), pp 5050–5052. Current version is 2.2. You can download its source code or buy precompiled versions for MacOS and Windows there.

Vincent Fourmond: Define a function with inline Ruby code in QSoas

23 September, 2020 - 19:36
QSoas can read and execute Ruby code directly, while reading command files, or even at the command prompt. For that, just write plain Ruby code inside a ruby...ruby end block. Probably the most useful possibility is to define elaborated functions directly from within QSoas, or, preferable, from within a script; this is an alternative to defining a function in a completely separated Ruby-only file using ruby-run. For instance, you can define a function for plain Michaelis-Menten kinetics with a file containing:

ruby
def my_func(x, vm, km)
  return vm/(1 + km/x)
end
ruby end

This defines the function my_func with three parameters, , (vm) and (km), with the formula:

You can then test that the function has been correctly defined running for instance:

QSoas> eval my_func(1.0,1.0,1.0)
 => 0.5
QSoas> eval my_func(1e4,1.0,1.0)
 => 0.999900009999

This yields the correct answer: the first command evaluates the function with x = 1.0, vm = 1.0 and km = 1.0. For , the result is (here 0.5). For , the result is almost . You can use the newly defined my_func in any place you would use any ruby code, such as in the optional argument to generate-buffer, or for arbitrary fits:

QSoas> generate-buffer 0 10 my_func(x,3.0,0.6)
QSoas> fit-arb my_func(x,vm,km)

To redefine my_func, just run the ruby code again with a new definition, such as:
ruby
def my_func(x, vm, km)
  return vm/(1 + km/x**2)
end
ruby end
The previous version is just erased, and all new uses of my_func will refer to your new definition.


See for yourselfThe code for this example can be found there. Browse the qsoas-goodies github repository for more goodies !

About QSoasQSoas is a powerful open source data analysis program that focuses on flexibility and powerful fitting capacities. It is released under the GNU General Public License. It is described in Fourmond, Anal. Chem., 2016, 88 (10), pp 5050–5052. Current version is 2.1. You can download its source code or buy precompiled versions for MacOS and Windows there.

Vincent Fourmond: Release 2.2 of QSoas

23 September, 2020 - 19:36
The new release of QSoas is finally ready ! It brings in a lot of new features and improvements, notably greatly improved memory use for massive multifits, a fit for linear (in)activation processes (the one we used in Fourmond et al, Nature Chemistry 2014), a new way to transform "numbers" like peak position or stats into new datasets and even SVG output ! Following popular demand, it also finally brings back the peak area output in the find-peaks command (and the other, related commands) ! You can browse the full list of changes there.

The new release can be downloaded from the downloads page.
Freely available binary images for QSoas 1.0In addition to the new release, we are now releasing the binary images for MacOS and Windows for the release 1.0. They are also freely available for download from the downloads page.

About QSoasQSoas is a powerful open source data analysis program that focuses on flexibility and powerful fitting capacities. It is released under the GNU General Public License. It is described in Fourmond, Anal. Chem., 2016, 88 (10), pp 5050–5052. Current version is 2.2. You can download its source code or buy precompiled versions for MacOS and Windows there.

Russell Coker: Qemu (KVM) and 9P (Virtfs) Mounts

23 September, 2020 - 09:02

I’ve tried setting up the Qemu (in this case KVM as it uses the Qemu code in question) 9P/Virtfs filesystem for sharing files to a VM. Here is the Qemu documentation for it [1].

VIRTFS="-virtfs local,path=/vmstore/virtfs,security_model=mapped-xattr,id=zz,writeout=immediate,fmode=0600,dmode=0700,mount_tag=zz"
VIRTFS="-virtfs local,path=/vmstore/virtfs,security_model=passthrough,id=zz,writeout=immediate,mount_tag=zz"

Above are the 2 configuration snippets I tried on the server side. The first uses mapped xattrs (which means that all files will have the same UID/GID and on the host XATTRs will be used for storing the Unix permissions) and the second uses passthrough which requires KVM to run as root and gives the same permissions on the host as on the VM. The advantages of passthrough are better performance through writing less metadata and having the same permissions in host and VM. The advantages of mapped XATTRs are running KVM/Qemu as non-root and not having a SUID file in the VM imply a SUID file in the host.

Here is the link to Bonnie++ output comparing Ext3 on a KVM block device (stored on a regular file in a BTRFS RAID-1 filesystem on 2 SSDs on the host), a NFS share from the host from the same BTRFS filesystem, and virtfs shares of the same filesystem. The only tests that Ext3 doesn’t win are some of the latency tests, latency is based on the worst-case not the average. I expected Ext3 to win most tests, but didn’t expect it to lose any latency tests.

Here is a link to Bonnie++ output comparing just NFS and Virtfs. It’s obvious that Virtfs compares poorly, giving about half the performance on many tests. Surprisingly the only tests where Virtfs compared well to NFS were the file creation tests which I expected Virtfs with mapped XATTRs to do poorly due to the extra metadata.

Here is a link to Bonnie++ output comparing only Virtfs. The options are mapped XATTRs with default msize, mapped XATTRs with 512k msize (I don’t know if this made a difference, the results are within the range of random differences), and passthrough. There’s an obvious performance benefit in passthrough for the small file tests due to the less metadata overhead, but as creating small files isn’t a bottleneck on most systems a 20% to 30% improvement in that area probably doesn’t matter much. The result from the random seeks test in passthrough is unusual, I’ll have to do more testing on that.

SE Linux

On Virtfs the XATTR used for SE Linux labels is passed through to the host. So every label used in a VM has to be valid on the host and accessible to the context of the KVM/Qemu process. That’s not really an option so you have to use the context mount option. Having the mapped XATTR mode work for SE Linux labels is a necessary feature.

Conclusion

The msize mount option in the VM doesn’t appear to do anything and it doesn’t appear in /proc/mounts, I don’t know if it’s even supported in the kernel I’m using.

The passthrough and mapped XATTR modes give near enough performance that there doesn’t seem to be a benefit of one over the other.

NFS gives significant performance benefits over Virtfs while also using less CPU time in the VM. It has the issue of files named .nfs* hanging around if the VM crashes while programs were using deleted files. It’s also more well known, ask for help with an NFS problem and you are more likely to get advice than when asking for help with a virtfs problem.

Virtfs might be a better option for accessing databases than NFS due to it’s internal operation probably being a better map to Unix filesystem semantics, but running database servers on the host is probably a better choice anyway.

Virtfs generally doesn’t seem to be worth using. I had hoped for performance that was better than NFS but the only benefit I seemed to get was avoiding the .nfs* file issue.

The best options for storage for a KVM/Qemu VM seem to be Ext3 for files that are only used on one VM and for which the size won’t change suddenly or unexpectedly (particularly the root filesystem) and NFS for everything else.

Related posts:

  1. QEMU for ARM Processes I’m currently doing some embedded work on ARM systems. Having...
  2. Question about a “Secure Filesystem” I have just been asked for advice about “secure filesystem”...
  3. BTRFS and SE Linux I’ve had problems with systems running SE Linux on BTRFS...

Steve Kemp: Using a FORTH-like language for something useful

22 September, 2020 - 11:00

So my previous post was all about implementing a simple FORTH-like language. Of course the obvious question is then "What do you do with it"?

So I present one possible use - turtle-graphics:

\ Draw a square of the given length/width
: square
  dup dup dup dup
  4 0 do
    forward
    90 turn
  loop
;

\ pen down
1 pen

\ move to the given pixel
100 100 move

\ draw a square of width 50 pixels
50 square

\ save the result (png + gif)
save

Exciting times!

Russ Allbery: Review: Unconquerable Sun

22 September, 2020 - 09:58

Review: Unconquerable Sun, by Kate Elliott

Series: Sun Chronicles #1 Publisher: Tor Copyright: 2020 ISBN: 1-250-19725-2 Format: Kindle Pages: 526

Sun is the daughter and heir of the mercurial Queen-Marshal Eirene, ruler of the Republic of Chaonia. Chaonia, thanks to Eirene and her ancestors, has carved out a fiercely independent position between the Yele League and the Phene Empire. Sun's father, Prince João, is one of Eirene's three consorts, all chosen for political alliances to shore up that fragile position. João is Gatoi, a civilization of feared fighters and supposed barbarians from outside Chaonia who normally ally with the Phene, which complicates Sun's position as heir. Sun attempts to compensate for that by winning battles for the Republic, following in the martial footsteps of her mother.

The publisher's summary of this book is not great (I'm a huge fan of Princess Leia, but that is... not the analogy that comes to mind), so let me try to help. This is gender-swapped Alexander the Great in space. However, it is gender-swapped Alexander the Great in space with her Companions, which means the DNA of this novel is half space opera and half heist story (without, to be clear, an actual heist, although there are some heist-like maneuvers). It's also worth mentioning that Sun, like Alexander, is not heterosexual.

The other critical thing to know before reading, mostly because it will get you through the rather painful start, is that the most interesting character in this book is not Sun, the Alexander analogue. It's Persephone, who isn't introduced until chapter seven.

Significant disclaimer up front: I got a reasonably typical US grade school history of Alexander the Great, which means I was taught that he succeeded his father, conquered a whole swath of the middle of the Eurasian land mass at a very young age, and then died and left his empire to his four generals who promptly divided it four uninteresting empires that no one's ever heard of, and that's why Rome is more important than Greece. (I put in that last bit to troll one specific person.)

I am therefore not the person to judge the parallels between this story and known history, or to notice any damage done to Greek pride, or to pick up on elements that might cause someone with a better grasp of that history to break out in hives. I did enough research to know that one scene in this book is lifted directly out of Alexander's life, but I'm not sure how closely the other parallels track. Yele is probably southern Greece and Phene is probably Persia, but I'm not certain even of that, and some of the details don't line up. If I had to hazard a guess, I'd say that Elliott has probably mangled history sufficiently to make it clear that this isn't intended to be a retelling, but if the historical parallels are likely to bother you, you may want to do more research before reading.

What I can say is that the space opera setup, while a bit stock, has all the necessary elements to make me happy. Unconquerable Sun is firmly in the "lost Earth" tradition: The Argosy fleet fled the now-mythical Celestial Empire and founded a new starfaring civilization without any contact with their original home. Eventually, they invented (or discovered; the characters don't know) the beacons, which allow for instantaneous travel between specific systems without the long (but still faster-than-light) journeys of the knnu drive. More recently, the beacon network has partly collapsed, cutting off the characters' known world from the civilization that was responsible for the beacons and guarded their secrets. It's a fun space opera history with lots of lost knowledge to reference and possibly discover, and with plot-enabling military choke points at the surviving beacons that link multiple worlds.

This is all background to the story, which is the ongoing war between Chaonia and the Phene Empire mixed with cutthroat political maneuvering between the great houses of the Chaonian Republic. This is where the heist aspects come in. Each house sends one representative to join the household of the Queen-Marshal and (more to the point for this story) another to join her heir. Sun has encouraged the individual and divergent talents of her Companions and their cee-cees (an unfortunate term that I suspect is short for Companion's Companion) and forged them into a good working team. A team that's about to be disrupted by the maneuverings of a rival house and the introduction of a new team member whom no one wants.

A problem with writing tactical geniuses is that they often aren't good viewpoint characters. Sun's tight third-person chapters, which is a little less than half the book, advance the plot and provide analysis of the interpersonal dynamics of the characters, but aren't the strength of the story. That lies with the interwoven first-person sections that follow Persephone, an altogether more interesting character.

Persephone is the scion of the house that is Sun's chief rival, but she has no interest in being part of that house or its maneuverings. When the story opens, she's a cadet in a military academy for recruits from the commoners, having run away from home, hidden her identity, and won a position through the open entrance exams. She of course doesn't stay there; her past catches up with her and she gets assigned to Sun, to a great deal of mutual suspicion. She also is assigned an impeccably dressed and stunningly beautiful cee-cee, Tiana, who has her own secrets and who is by far the most interesting character in the book.

Somewhat unusually for the space opera tradition, this is a book that knows that common people exist and have interesting lives. It's primarily focused on the ruling houses, but that focus is not exclusive and the rulers do not have a monopoly on competence. Elliott also avoids narrowing the political field too far; the Gatoi are separate from the three rival powers, and there are other groups with traditions older than the Chaonian Republic and their own agendas. Sun and her Companions are following a couple of political threads, but there is clearly more going on in this world than that single plot.

This is exactly the kind of story I think of when I think space opera. It's not doing anything that original or groundbreaking, and it's not going to make any of my lists of great literature, but it's a fun romp with satisfyingly layered bits of lore, a large-scale setting with lots of plot potential, and (once we get through the confusing and somewhat tedious process of introducing rather too many characters in short succession) some great interpersonal dynamics. It's the kind of book in which the characters are in the middle of decisive military action in an interstellar war and are also near-teenagers competing for ratings in an ad hoc reality TV show, primarily as an excuse to create tactical distractions for Sun's latest scheme. The writing is okay but not great, and the first few chapters have some serious infodumping problems, but I thoroughly enjoyed the whole book and will pre-order the sequel.

One Amazon review complained that Unconquerable Sun is not a space opera like Hyperion or Use of Weapons. That is entirely true, but if that's your standard for space opera, the world may be a disappointing place. This is a solid entry in a subgenre I love, with some great characters, sarcasm, competence porn, plenty of pages to keep turning, a few twists, and the promise of more to come. Recommended.

Followed by the not-yet-published Furious Heaven.

Rating: 7 out of 10

Kees Cook: security things in Linux v5.7

22 September, 2020 - 06:32

Previously: v5.6

Linux v5.7 was released at the end of May. Here’s my summary of various security things that caught my attention:

arm64 kernel pointer authentication
While the ARMv8.3 CPU “Pointer Authentication” (PAC) feature landed for userspace already, Kristina Martsenko has now landed PAC support in kernel mode. The current implementation uses PACIASP which protects the saved stack pointer, similar to the existing CONFIG_STACKPROTECTOR feature, only faster. This also paves the way to sign and check pointers stored in the heap, as a way to defeat function pointer overwrites in those memory regions too. Since the behavior is different from the traditional stack protector, Amit Daniel Kachhap added an LKDTM test for PAC as well.

BPF LSM
The kernel’s Linux Security Module (LSM) API provide a way to write security modules that have traditionally implemented various Mandatory Access Control (MAC) systems like SELinux, AppArmor, etc. The LSM hooks are numerous and no one LSM uses them all, as some hooks are much more specialized (like those used by IMA, Yama, LoadPin, etc). There was not, however, any way to externally attach to these hooks (not even through a regular loadable kernel module) nor build fully dynamic security policy, until KP Singh landed the API for building LSM policy using BPF. With this, it is possible (for a privileged process) to write kernel LSM hooks in BPF, allowing for totally custom security policy (and reporting).

execve() deadlock refactoring
There have been a number of long-standing races in the kernel’s process launching code where ptrace could deadlock. Fixing these has been attempted several times over the last many years, but Eric W. Biederman and Ernd Edlinger decided to dive in, and successfully landed the a series of refactorings, splitting up the problematic locking and refactoring their uses to remove the deadlocks. While he was at it, Eric also extended the exec_id counter to 64 bits to avoid the possibility of the counter wrapping and allowing an attacker to send arbitrary signals to processes they normally shouldn’t be able to.

slub freelist obfuscation improvements
After Silvio Cesare observed some weaknesses in the implementation of CONFIG_SLAB_FREELIST_HARDENED‘s freelist pointer content obfuscation, I improved their bit diffusion, which makes attacks require significantly more memory content exposures to defeat the obfuscation. As part of the conversation, Vitaly Nikolenko pointed out that the freelist pointer’s location made it relatively easy to target too (for either disclosures or overwrites), so I moved it away from the edge of the slab, making it harder to reach through small-sized overflows (which usually target the freelist pointer). As it turns out, there were a few assumptions in the kernel about the location of the freelist pointer, which had to also get cleaned up.

RISCV page table dumping
Following v5.6’s generic page table dumping work, Zong Li landed the RISCV page dumping code. This means it’s much easier to examine the kernel’s page table layout when running a debug kernel (built with PTDUMP_DEBUGFS), visible in /sys/kernel/debug/kernel_page_tables.

array index bounds checking
This is a pretty large area of work that touches a lot of overlapping elements (and history) in the Linux kernel. The short version is: C is bad at noticing when it uses an array index beyond the bounds of the declared array, and we need to fix that. For example, don’t do this:

int foo[5];
...
foo[8] = bar;

The long version gets complicated by the evolution of “flexible array” structure members, so we’ll pause for a moment and skim the surface of this topic. While things like CONFIG_FORTIFY_SOURCE try to catch these kinds of cases in the memcpy() and strcpy() family of functions, it doesn’t catch it in open-coded array indexing, as seen in the code above. GCC has a warning (-Warray-bounds) for these cases, but it was disabled by Linus because of all the false positives seen due to “fake” flexible array members. Before flexible arrays were standardized, GNU C supported “zero sized” array members. And before that, C code would use a 1-element array. These were all designed so that some structure could be the “header” in front of some data blob that could be addressable through the last structure member:

/* 1-element array */
struct foo {
    ...
    char contents[1];
};

/* GNU C extension: 0-element array */
struct foo {
    ...
    char contents[0];
};

/* C standard: flexible array */
struct foo {
    ...
    char contents[];
};

instance = kmalloc(sizeof(struct foo) + content_size);

Converting all the zero- and one-element array members to flexible arrays is one of Gustavo A. R. Silva’s goals, and hundreds of these changes started landing. Once fixed, -Warray-bounds can be re-enabled. Much more detail can be found in the kernel’s deprecation docs.

However, that will only catch the “visible at compile time” cases. For runtime checking, the Undefined Behavior Sanitizer has an option for adding runtime array bounds checking for catching things like this where the compiler cannot perform a static analysis of the index values:

int foo[5];
...
for (i = 0; i < some_argument; i++) {
    ...
    foo[i] = bar;
    ...
}

It was, however, not separate (via kernel Kconfig) until Elena Petrova and I split it out into CONFIG_UBSAN_BOUNDS, which is fast enough for production kernel use. With this enabled, it's now possible to instrument the kernel to catch these conditions, which seem to come up with some regularity in Wi-Fi and Bluetooth drivers for some reason. Since UBSAN (and the other Sanitizers) only WARN() by default, system owners need to set panic_on_warn=1 too if they want to defend against attacks targeting these kinds of flaws. Because of this, and to avoid bloating the kernel image with all the warning messages, I introduced CONFIG_UBSAN_TRAP which effectively turns these conditions into a BUG() without needing additional sysctl settings.

Fixing "additive" snprintf() usage
A common idiom in C for building up strings is to use sprintf()'s return value to increment a pointer into a string, and build a string with more sprintf() calls:

/* safe if strlen(foo) + 1 < sizeof(string) */
wrote  = sprintf(string, "Foo: %s\n", foo);
/* overflows if strlen(foo) + strlen(bar) > sizeof(string) */
wrote += sprintf(string + wrote, "Bar: %s\n", bar);
/* writing way beyond the end of "string" now ... */
wrote += sprintf(string + wrote, "Baz: %s\n", baz);

The risk is that if these calls eventually walk off the end of the string buffer, it will start writing into other memory and create some bad situations. Switching these to snprintf() does not, however, make anything safer, since snprintf() returns how much it would have written:

/* safe, assuming available <= sizeof(string), and for this example
 * assume strlen(foo) < sizeof(string) */
wrote  = snprintf(string, available, "Foo: %s\n", foo);
/* if (strlen(bar) > available - wrote), this is still safe since the
 * write into "string" will be truncated, but now "wrote" has been
 * incremented by how much snprintf() *would* have written, so "wrote"
 * is now larger than "available". */
wrote += snprintf(string + wrote, available - wrote, "Bar: %s\n", bar);
/* string + wrote is beyond the end of string, and availabe - wrote wraps
 * around to a giant positive value, making the write effectively 
 * unbounded. */
wrote += snprintf(string + wrote, available - wrote, "Baz: %s\n", baz);

So while the first overflowing call would be safe, the next one would be targeting beyond the end of the array and the size calculation will have wrapped around to a giant limit. Replacing this idiom with scnprintf() solves the issue because it only reports what was actually written. To this end, Takashi Iwai has been landing a bunch scnprintf() fixes.

That's it for now! Let me know if there is anything else you think I should mention here. Next up: Linux v5.8.

© 2020, Kees Cook. This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 License.

Antoine Beaupré: PSA: Mailman used to harrass people

22 September, 2020 - 01:44

It seems that Mailman instances are being abused to harrass people with subscribe spam. If some random people complain to you that they "never wanted to subscribe to your mailing list", you may be a victim to that attack, even if you run the latest Mailman 2.

TL;DR: IKR! HOW DO I FIX THIS!?

Make sure you have SUBSCRIBE_FORM_SECRET set in your mailman configuration:

SECRET=$(tr -dc 'A-Za-z0-9' < /dev/urandom | head -c 30)'
echo "SUBSCRIBE_FORM_SECRET = '$SECRET'" >> /etc/mailman/mm.cfg

This will add a magic token to all forms in the Mailman web forms that will force the attacker to at least get a token before asking for registration. There are, of course, other ways of performing the attack then, but it's more expensive than a single request for the attacker and keeps most of the junk out.

Other solutions

I originally deployed a different fix, using referrer checks and an IP block list:

RewriteMap hosts-deny  txt:/etc/apache2/blocklist.txt
RewriteCond ${hosts-deny:%{REMOTE_ADDR}|NOT-FOUND} !=NOT-FOUND [OR]
RewriteCond ${hosts-deny:%{REMOTE_HOST}|NOT-FOUND} !=NOT-FOUND [OR]
RewriteCond %{HTTP_REFERER} !^https://lists.torproject.org/$ [NC]
RewriteRule ^/cgi-bin/mailman/subscribe/ - [F]
# see also https://www.w3.org/TR/referrer-policy/#referrer-policy-origin
Header always set Referrer-Policy "origin"

I kept those restrictions in place because it keeps the spammers from even hitting the Mailman CGI, which is useful to preserve our server resources. But if "they" escalate with smarter crawlers, the block list will still be useful.

You can use this query to extract the top 10 IP addresses used for subscription attempts:

awk '{ print $NF }' /var/log/mailman/subscribe | sort | uniq -c | sort -n | tail -10  | awk '{ print $2 " " $1 }'

Note that this might include email-based registration, but in our logs those are extremely rare: only two in three weeks, out of over 73,000 requests. I also use this to keep an eye on the logs:

tail -f  /var/log/mailman/subscribe /var/log/apache2/lists.torproject.org-access.log | grep -v 'GET /pipermail/'

The server-side mitigations might also be useful if you happen to run an extremely old version of Mailman, that is pre-2.1.18, but it's now over 6 years old and part of every supported Debian release out there (all the way back to Debian 8 jessie).

Why does that attack work?

Because Mailman 2 doesn't have CSRF tokens in its forms by default, anyone can send a POST request to /mailman/subscribe/LISTNAME to have Mailman send an email to the user. In the old "Internet is for nice people" universe, that wasn't a problem: all it does is ask the victim if they want to subscribe to LISTNAME. Innocuous, right?

But in the brave, new, post-Eternal-September, "Internet is for stupid" universe, some assholes think it's a good idea to make a form that collects hundreds of mailing list URLs and spam them through an iframe. To see what that looks like, you can look at the rendered source code behind samedyfreeday.co.uk (not linking to avoid promoting it). That site does what is basically a distributed cross-site scripting attack against Mailman servers.

Obviously, CSRF protection should be enabled by default in Mailman, but there you go. Hopefully this will help some folks...

(The latest Mailman 3 release doesn't suffer from such idiotic defaults and ships with proper CSRF protection out of the box.)

Jonathan McDowell: Mainline Linux on the MikroTik RB3011

22 September, 2020 - 00:53

I upgraded my home internet connection to fibre (FTTP) last October. I’m still on an 80M/20M service, so it’s no faster than my old VDSL FTTC connection was, and as a result for a long time I continued to use my HomeHub 5A running OpenWRT. However the FTTP ONT meant I was using up an additional ethernet port on the router, and I was already short, so I ended up with a GigE switch in use as well. Also my wifi is handled by a UniFi, which takes its power via Power-over-Ethernet. That mean I had a router, a switch and a PoE injector all in close proximity. I wanted to reduce the number of devices, and ideally upgrade to something that could scale once I decide to upgrade my FTTP service speed.

Looking around I found the MikroTik RB3011UiAS-RM, which is a rack mountable device with 10 GigE ports (plus an SFP slot) and a dual core Qualcomm IPQ8064 ARM powering it. There’s 1G RAM and 128MB NAND flash, as well as a USB3 port. It also has PoE support. On paper it seemed like an ideal device. I wasn’t particularly interested in running RouterOS on it (the provided software), but that’s based on Linux and there was some work going on within OpenWRT to add support, so it seemed like a worthwhile platform to experiment with (what, you expected this to be about me buying an off the shelf device and using it with only the supplied software?). As an added bonus a friend said he had one he wasn’t using, and was happy to sell it to me for a bargain price.

I did try out RouterOS to start with, but I didn’t find it particularly compelling. I’m comfortable configuring firewalling and routing at a Linux command line, and I run some additional services on the router like my MQTT broker, and mqtt-arp, my wifi device presence monitor. I could move things around such that they ran on the house server, but I consider them core services and as a result am happier with them on the router.

The first step was to get something booting on the router. Luckily it has an RJ45 serial console port on the back, and a reasonably featured bootloader that can manage to boot via tftp over the network. It wants an ELF binary rather than a plain kernel, but Sergey Sergeev had done the hard work of getting u-boot working for the IPQ8064, which mean I could just build normal u-boot images to try out.

Linux upstream already had basic support for a lot of the pieces I was interested in. There’s a slight fudge around AUTO_ZRELADDR because the network coprocessors want a chunk of memory at the start of RAM, but there’s ongoing discussions about how to handle this cleanly that I’m hopeful will eventually mean I can drop that hack. Serial, ethernet, the QCA8337 switches (2 sets of 5 ports, tied to different GigE devices on the processor) and the internal NOR all had drivers, so it was a matter of crafting an appropriate DTB to get them working. That left niggles.

First, the second switch is hooked up via SGMII. It turned out the IPQ806x stmmac driver didn’t initialise the clocks in this mode correctly, and neither did the qca8k switch driver. So I need to fix up both of those (Sergey had handled the stmmac driver, so I just had to clean up and submit his patch). Next it turned out the driver for talking to the Qualcomm firmware (SCM) had been updated in a way that broke the old method needed on the IPQ8064. Some git archaeology figured that one out and provided a solution. Ansuel Smith helpfully provided the DWC3 PHY driver for the USB port. That got me to the point I could put a Debian armhf image onto a USB stick and mount that as root, which made debugging much easier.

At this point I started to play with configuring up the device to actually act as a router. I make use of a number of VLANs on my home network, so I wanted to make sure I could support those. Turned out the stmmac driver wasn’t happy reconfiguring its MTU because the IPQ8064 driver doesn’t configured the FIFO sizes. I found what seem to be the correct values and plumbed them in. Then the qca8k driver only supported port bridging. I wanted the ability to have a trunk port to connect to the upstairs switch, while also having ports that only had a single VLAN for local devices. And I wanted the switch to handle this rather than requiring the CPU to bridge the traffic. Thankfully it’s easy to find a copy of the QCA8337 datasheet and the kernel Distributed Switch Architecture is pretty flexible, so I was able to implement the necessary support.

I stuck with Debian on the USB stick for actually putting the device into production. It makes it easier to fix things up if necessary, and the USB stick allows for a full Debian install which would be tricky on the 128M of internal NAND. That means I can use things like nftables for my firewalling, and use the standard Debian packages for things like collectd and mosquitto. Plus for debug I can fire up things like tcpdump or tshark. Which ended up being useful because when I put the device into production I started having weird IPv6 issues that turned out to be a lack of proper Ethernet multicast filter support in the IPQ806x ethernet device. The driver would try and setup the multicast filter for the IPv6 NDP related packets, but it wouldn’t actually work. The fix was to fall back to just receiving all multicast packets - this is what the vendor driver does.

Most of this work will be present once the 5.9 kernel is released - the basics are already in 5.8. Currently not queued up that I can think of are the following:

  • stmmac IPQ806x FIFO sizes. I sent out an RFC patch for these, but didn’t get any replies. I probably just need to submit this.
  • NAND. This is missing support for the QCOM ADM DMA engine. I’ve sent out the patch I found to enable this, and have had some feedback, so I’m hopeful it will get in at some point.
  • LCD. AFAICT LCD is an ST7735 device, which has kernel support, but I haven’t spent serious effort getting the SPI configuration to work.
  • Touchscreen. Again, this seems to be a zt2046q or similar, which has a kernel driver, but the basic attempts I’ve tried don’t get any response.
  • Proper SFP functionality. The IPQ806x has a PCS module, but the stmmac driver doesn’t have an easy way to plumb this in. I have ideas about how to get it working properly (and it can be hacked up with a fixed link config) but it’s not been a high priority.
  • Device tree additions. Some of the later bits I’ve enabled aren’t yet in the mainline RB3011 DTB. I’ll submit a patch for that at some point.

Overall I consider the device a success, and it’s been entertaining getting it working properly. I’m running a mostly mainline kernel, it’s handling my house traffic without breaking a sweat, and the fact it’s running Debian makes it nice and easy to throw more things on it as I desire. However it turned out the RB3011 isn’t as perfect device as I’d hoped. The PoE support is passive, and the UniFi wants 802.1af. So I was going to end up with 2 devices. As it happened I picked up a cheap D-Link DGS-1210-10P switch, which provides the PoE support as well as some additional switch ports. Plus it runs Linux, so more on that later…

Russ Allbery: Review: Lower Ed

21 September, 2020 - 10:59

Review: Lower Ed, by Tressie McMillan Cottom

Publisher: The New Press Copyright: 2017 Printing: 2018 ISBN: 1-62097-472-X Format: Kindle Pages: 217

Lower Ed (subtitled The Troubling Rise of For-Profit Colleges in the New Economy) is the first book by sociologist Tressie McMillan Cottom. (I previously reviewed her second book, the excellent essay collection Thick.) It is a deep look at the sociology of for-profit higher education in the United States based on interviews with students and executives, analysis of Wall Street filings, tests of the admissions process, and her own personal experiences working for two of the schools. One of the questions that McMillan Cottom tries to answer is why students choose to enroll in these institutions, particularly the newer type of institution funded by federal student loans and notorious for being more expensive and less valuable than non-profit colleges and universities.

I was hesitant to read this book because I find for-profit schools depressing. I grew up with the ubiquitous commercials, watched the backlash develop, and have a strongly negative impression of the industry, partly influenced by having worked in traditional non-profit higher education for two decades. The prevailing opinion in my social group is that they're a con job. I was half-expecting a reinforcement of that opinion by example, and I don't like reading infuriating stories about people being defrauded.

I need not have worried. This is not that sort of book (nor, in retrospect, do I think McMillan Cottom would approach a topic from that angle). Sociology is broader than reporting. Lower Ed positions for-profit colleges within a larger social structure of education, credentialing, and changes in workplace expectations; takes a deep look at why they are attractive to their students; and humanizes and complicates the motives and incentives of everyone involved, including administrators and employees of for-profit colleges as well as the students. McMillan Cottom does of course talk about the profit motive and the deceptions surrounding that, but the context is less that of fraud that people are unable to see through and more a balancing of the drawbacks of a set of poor choices embedded in institutional failures.

One of my metrics for a good non-fiction book is whether it introduces me to a new idea that changes how I analyze the world. Lower Ed does that twice.

The first idea is the view of higher education through the lens of risk shifting. It used to be common for employers to hire people without prior job-specific training and do the training in-house, possibly through an apprenticeship structure. More notably, once one was employed by a particular company, the company routinely arranged or provided ongoing training. This went hand-in-hand with a workplace culture of long tenure, internal promotion, attempts to avoid layoffs, and some degree of mutual loyalty. Companies expected to invest significantly in an employee over their career and thus also had an incentive to retain that employee rather than train someone for a competitor.

However, from a purely financial perspective, this is a risk and an inefficiency, similar to the risk of carrying a large inventory of parts and components. Companies have responded to investor-driven focus on profits and efficiency by reducing overhead and shifting risk. This leads to the lean supply chain, where no one pays for parts to sit around in warehouses and companies aren't caught with large stockpiles of now-useless components, but which is more sensitive to any disruption (such as from a global pandemic). And, for employment, it leads to a desire to hire pre-trained workers, retain only enough workers to do the current amount of work, and replace them with new workers who already have appropriate training rather than retrain them.

The effect of the corporate decision to only hire pre-trained employees is to shift the risk and expense of training from the company to the prospective employee. The individual has to seek out training at their own expense in the hope (not guarantee) that at the conclusion of that training they will get or retain a job. People therefore turn to higher education to both provide that training and to help them decide what type of training will eventually be valuable. This has a long history with certain professional fields (doctors and lawyers, for example), but the requirements for completing training in those fields are relatively clear (a professional license to practice) and the compensation reflects the risk. What's new is the shift of training risk to the individual in more mundane jobs, without any corresponding increase in compensation.

This, McMillan Cottom explains, is the background for the growth in demand for higher education in general and the the type of education offered by for-profit colleges in particular. Workers who in previous eras would be trained by their employers are now responsible for their own training. That training is no longer judged by the standards of a specific workplace, but is instead evaluated by a hiring process that expects constant job-shifting. This leads to increased demand by both workers and employers for credentials: some simple-to-check certificate of completion of training that says that this person has the skills to immediately start doing some job. It also leads to a demand for more flexible class hours, since the student is now often someone older with a job and a family to balance. Their ongoing training used to be considered a cost of business and happen during their work hours; now it is something they have to fit around the contours of their life because their employer has shifted that risk to them.

The risk-shifting frame makes sense of the "investment" language so common in for-profit education. In this job economy, education as investment is not a weird metaphor for the classic benefits of a liberal arts education: broadened perspective, deeper grounding in philosophy and ethics, or heightened aesthetic appreciation. It's an investment in the literal financial sense; it is money that you spend now in order to get a financial benefit (a job) in the future. People have to invest in their own training because employers are no longer doing so, but still require the outcome of that investment. And, worse, it's primarily a station-keeping investment. Rather than an optional expenditure that could reap greater benefits later, it's a mandatory expenditure to prevent, at best, stagnation in a job paying poverty wages, and at worst the disaster of unemployment.

This explains renewed demand for higher education, but why for-profit colleges? We know they cost more and have a worse reputation (and therefore their credentials have less value) than traditional non-profit colleges. Flexible hours and class scheduling explains some of this but not all of it. That leads to the second perspective-shifting idea I got from Lower Ed: for-profit colleges are very good at what they focus time and resources on, and they focus on enrolling students.

It is hard to enroll in a university! More precisely, enrolling in a university requires bureaucracy navigation skills, and those skills are class-coded. The people who need them the most are the least likely to have them.

Universities do not reach out to you, nor do they guide you through the process. You have to go to them and discover how to apply, something that is often made harder by the confusing state of many university web sites. The language and process is opaque unless other people in your family have experience with universities and can explain it. There might be someone you can reach on the phone to ask questions, but they're highly unlikely to proactively guide you through the remaining steps. It's your responsibility to understand deadlines, timing, and sequence of operations, and if you miss any of the steps (due to, for example, the overscheduled life of someone in need of better education for better job prospects), the penalty in time and sometimes money can be substantial. And admission is just the start; navigating financial aid, which most students will need, is an order of magnitude more daunting. Community colleges are somewhat easier (and certainly cheaper) than universities, but still have similar obstacles (and often even worse web sites).

It's easy for people like me, who have long professional expertise with bureaucracies, family experience with higher education, and a support network of people to nag me about deadlines, to underestimate this. But the application experience at a for-profit college is entirely different in ways far more profound than I had realized. McMillan Cottom documents this in detail from her own experience working for two different for-profit colleges and from an experiment where she indicated interest in multiple for-profit colleges and then stopped responding before signing admission paperwork. A for-profit college is fully invested in helping a student both apply and get financial aid, devotes someone to helping them through that process, does not expect them to understand how to navigate bureaucracies or decipher forms on their own, does not punish unexpected delays or missed appointments, and goes to considerable lengths to try to keep anyone from falling out of the process before they are enrolled. They do not expect their students to already have the skills that one learns from working in white-collar jobs or from being surrounded by people who do. They provide the kind of support that an educational institution should provide to people who, by definition, don't understand something and need to learn.

Reading about this was infuriating. Obviously, this effort to help people enroll is largely for predatory reasons. For-profit schools make their money off federal loans and they don't get that money unless they can get someone to enroll and fill out financial paperwork (and to some extent keep them enrolled), so admissions is their cash cow and they act accordingly. But that's not why I found it infuriating; that's just predictable capitalism. What I think is inexcusable is that nothing they do is that difficult. We could being doing the same thing for prospective community college students but have made the societal choice not to. We believe that education is valuable, we constantly advocate that people get more job training and higher education, and yet we demand prospective students navigate an unnecessarily baroque and confusing application process with very little help, and then stereotype and blame them for failing to do so.

This admission support is not a question of resources. For-profit colleges are funded almost entirely by federally-guaranteed student loans. We are paying them to help people apply. It is, in McMillan Cottom's term, a negative social insurance program. Rather than buffering people against the negative effects of risk-shifting of employers by helping them into the least-expensive and most-effective training programs (non-profit community colleges and universities), we are spending tax dollars to enrich the shareholders of for-profit colleges while underfunding the alternatives. We are choosing to create a gap that routes government support to the institution that provides worse training at higher cost but is very good at helping people apply. It's as if the unemployment system required one to use payday lenders to get one's unemployment check.

There is more in this book I want to talk about, but this review is already long enough. Suffice it to say that McMillan Cottom's analysis does not stop with market forces and the admission process, and the parts of her analysis that touch on my own personal experience as someone with a somewhat unusual college path ring very true. Speaking as a former community college student, the discussion of class credit transfer policies and the way that institutional prestige gatekeeping and the desire to push back against low-quality instruction becomes a trap that keeps students in the for-profit system deserves another review this length. So do the implications of risk-shifting and credentialism on the morality of "cheating" on schoolwork.

As one would expect from the author of the essay "Thick" about bringing context to sociology, Lower Ed is personal and grounded. McMillan Cottom doesn't shy away from including her own experiences and being explicit about her sources and research. This is backed up by one of the best methodological notes sections I've seen in a book. One of the things I love about McMillan Cottom's writing is that it's solidly academic, not in the sense of being opaque or full of jargon (the text can be a bit dense, but I rarely found it hard to follow), but in the sense of being clear about the sources of knowledge and her methods of extrapolation and analysis. She brings her receipts in a refreshingly concrete way.

I do have a few caveats. First, I had trouble following a structure and line of reasoning through the whole book. Each individual point is meticulously argued and supported, but they are not always organized into a clear progression or framework. That made Lower Ed feel at times like a collection of high-quality but somewhat unrelated observations about credentials, higher education, for-profit colleges, their student populations, their business models, and their relationships with non-profit schools.

Second, there are some related topics that McMillan Cottom touches on but doesn't expand sufficiently for me to be certain I understood them. One of the big ones is credentialism. This is apparently a hot topic in sociology and is obviously important to this book, but it's referenced somewhat glancingly and was not satisfyingly defined (at least for me). There are a few similar places where I almost but didn't quite follow a line of reasoning because the book structure didn't lay enough foundation.

Caveats aside, though, this was meaty, thought-provoking, and eye-opening, and I'm very glad that I read it. This is a topic that I care more about than most people, but if you have watched for-profit colleges with distaste but without deep understanding, I highly recommend Lower Ed.

Rating: 8 out of 10

Pages

Creative Commons License ลิขสิทธิ์ของบทความเป็นของเจ้าของบทความแต่ละชิ้น
ผลงานนี้ ใช้สัญญาอนุญาตของครีเอทีฟคอมมอนส์แบบ แสดงที่มา-อนุญาตแบบเดียวกัน 3.0 ที่ยังไม่ได้ปรับแก้