Tuesday, June 30, 2015

BSDCan 2015 Trip Report: Vsevolod Stakhov

During the BSDCan 2015 conference, I have attended the FreeBSD developers summit. I was particularly interested in the track called "Designing Universal Configuration Files for FreeBSD". As I'm the author of the library that was discussed,  I did a talk about library internals and discussed some open questions with Jonathan Anderson, David Chisnall and Allan Jude. We have planned some proposal changes, the interaction with libnv and casper and the following integration of UCL into FreeBSD base system. We have also discussed the desired features and Jonathan suggested a reasonable approach to implement missing ones.

Moreover, during the conference I have finished the feature of flexible dependencies in `pkg'. We have discussed this feature among all pkg developers that were also on BSDCan (namely, bapt@, matthew@ and bdrewery@). I've proposed my view of the future packages dependencies that would resolve the vast majority of the current issues with dependencies and upgrades. I'm going to write a detailed report about this feature to the pkg@ mailing list (I was just too busy with other tasks after the conference).

Another question we've discussed was the problem of digital signatures for packages and distributions. We have concluded that moving from RSA to ed25519 algorithm would simplify pkg architecture by avoiding linking to openssl (which is quite complicated for all openssl versions supported).

Further, after Ted Uagnst presentation I have a conversation with him and John-Mark Gourney (jmg@) about digital signatures formats, compatibility with OpenBSD signify tool and packages signing questions including the ways of how to verify signatures of untrusted sources with potentially malleable signing algorithms.

Afterwards, I've talked with Colin Percival (cpersiva@) asking for his comments about streamlined signatures scheme proposed by D.J. Bernstein. He agreed that this scheme might work securely providing a more convenient users' tool for digital signatures verification and creation.

Among other topics, I have discussed cryptography and security with John-Mark Gourney. We have also talked about '/dev/random' and fortuna upcoming patch. We discussed numerous topics about FreeBSD packages and pkg tool in particular with Baptiste Daroussin, Bryan Drewery and Matthew Seaman.

I have also extracted a lot of valuable information from BSDCan topics, namely from 'CloudABI' given by Ed Schouten and 'Protecting FreeBSD with Secure Virtual Architecture' given by John Criswell.

I'd like to thank the FreeBSD Foundation for giving me the possibility to attend the BSDCan 2015!

Wednesday, June 24, 2015

BSDCan 2015 Trip Report: Ahmed Kamal

I come from Cairo, Egypt, and this was my very first BSD conference! Needless to say, it was a blast! It was said many times during the conference, and I believe it to be true, so here it is again. “I’m here because so many people smarter than me are here!”. My flight landed on the 11th, time since I left home was roughly 18 hours, it was a couple of long flights. I seriously should have been tired but the excitement was keeping me awake. I took a walk around the city for a couple of hours to freshen up. Li-Wen Hsu joined me for the walk. We had an interesting discussion about jenkins, documentation, ways to contribute to BSD. We also had an interesting discussion about CJK languages, and their special needs. Most of that was new to me, although arabic script is complex too, and can get quite wild!! Afterwards I went back to the campus, I joined a documentation session. It was a kickoff to how the freebsd handbook and similar documentation were written. I learnt some useful tricks which I know I will probably put to use soon contributing some documentation. At roughly 21:00 my eyes turned red, and body was calling on me to go to sleep. I decided to sleep early and be fresh on day-1 of the conference!

Day one began with excitement .. The plenary was just inspiring! I was made aware of multiple new projects where BSD is helping the world. The idea of building “OS Coursework” for university level studies was quite interesting to me. The opening keynote also included legendary Steven Bourne (of /bin/sh fame!) explaining how it all started. He even showed annotated unix source code printed on paper. His quotes included “Took 2 years to get shell quoting straight, and I'm not sure it's straight now” and about “shell shock” he mentioned he didn’t write that code :) The next session for me was “Embedded FreeBSD Development and Package Building via QEMU”. This is an interesting topic about cross compiling packages for embedded devices using the power of Qemu. Qemu began as a JIT translation engine, later the PC hardware was added and now it can be used to build packages for say ARM platforms. The speaker mentioned the flow should hopefully soon work on Apple hardware.

The next big talk to me, was “A reimplementation of NetBSD using a MicroKernel” by Andrew Tanenbaum. It was quite exciting to watch prof Andrew describe his work on micro-kernels and how he’s rebuilding netbsd-3 through the microkernel approach. The basic idea is to separate IO devices, so that for example a random driver (like an audio driver), is no longer running with unlimited powers on your machine. Such a driver can no longer write to disk as it should(n’t). Also the idea is to Isolate communication, restrict kernel calls on a per kernel component basis. This protection can be enforced by the MMU. Various user-mode servers provide the needed services to applications. This includes VFS, Process manager, Memory manager, .. etc. The Reincarnation server is the parent of all drivers and servers. When a driver or server dies, RS collects it. Checks the kernel config table and restarts it. For improving IPC reliability in the new kernel, fixed length messages are used everywhere to avoid potential buffer overflows. In contrast to modern OSs, “drivers” are basically untrusted code! heavily isolated. Cannot touch kernel data structures. Andy mentioned how Infinite loops are detected, and how the driver is appropriately restarted if hung. Andy later described a fault injection experiment where his team injected 800,000 faults on binary drivers, the sequence was 100 faults injected, wait 1s and the drivers crashed 18,000 times, but never crashed the OS. Kernel reliability “proven” to me! In order to port Minix3 to ARM platform, various things needed to be done like (added uBoot support, rewrote context switching, removed x86 segmentation code, imported netbsd arm headers, ported build.sh for cross toolchain, wrote drivers for sdcard!). Things to do include adding crucial missing system calls, port more pkgs (java, browsers...), get it running on raspi, port rump! Minix-3 is a microkernel reimplementation of netbsd, which was primarily motivated by providing that microkernels can be practical OSs that have lots of applications on top. It has proven that drivers belong in user-mode. Future features Andy’s team is working on include live kernel patching which he described how it should work. Upgrade os and change data structures in a live way, without restarting the running processes! This was one hell of a talk, it was very dense and packed with information

The next talk was “Measure Twice, Code Once. Network Performance Analysis for FreeBSD” by George Neville-Neil. George discussed performance of the network under FreeBSD. This is critical for supporting modern 10G ethernet networks. He introduced the conductor python framework he wrote to perform the testing. He joked about how others always say “Yeah, we have something similar that we’re not yet ready to release”. George showed performance graphs from various systems such as pfsense, freebsd, openbsd pf and Linux’s iptables. Various interesting and yet unanswered questions were revealed such as “Why is pfsense much faster” than FreeBSD on which it is based! What kind of patches went into pfsense to get that effect. Also it was found that iptables offers much better scalability for multi-core performance. Obviously iptables has been tuned for SMP performance, but this necessitates the question of why it is much more scalable. And how the BSDs can improve their packet filters. George mentioned that this is only the beginning. Those questions need a lot of work to be answered.

My next session was “Molecular Evolution, Genomic Analysis and FreeBSD” by Joseph Mingrone. The session was in the sysadmin track, something I can easily relate to. Joseph runs a freebsd based cluster for molecular evolution analysis at his university. Joseph began by giving an introduction to genomics, something way outside my comfort zone. It’s always fun to me whenever I discover new worlds like that :) He moved on to explain the IT infrastructure powering his clusters and how they moved from Solaris to FreeBSD 7.x and now at FreeBSD 10.1. He explained how he had a hard time just installing FreeBSD where sometimes the machines would just “eat” the CDs (yes they’re still in there many years later :) He currently uses Poudriere, enjoys ZFS on the storage server and NFS mounts that on the compute nodes.

The next session was “FreeBSD on ARMv8” by Andrew Turner. It was great see’ing the kind of effort that goes into supporting ARMv8. Everything we sort of take for granted like pagetables, MMU, enabling virtual addressing, calling into C code ...etc was being made to work step by step. Later-on pmap support was added, then it was possible to use a 4MB in kernel filesystem. Afterwards on it was possible to expand this storage area. Now it was time to play with the dynamic linker. After that, it was time to add ThunderX support (bus dma, ITS, …) Future work includes (ACPI, hw-pmc, dtrace, gem5). This was an advanced but certainly fun and educational session.


Day-2 now begins with a very interesting session “CloudABI: Cloud computing meets fine-grained capabilities” by Ed Schouten. The overall goal is to improve Unix security and portability. For example, a web server should only read files and write to network, but in reality it can do a lot more! (run a bitcoin miner, install cronjobs that call binaries in /tmp..etc) .. Problem #2, running a 3rd party application, whether running them directly or thru jails/docker is not really safe, a VM might be acceptable though! Problem 3 is that UNIX programs are hard to reuse and test as a whole. The idea is to extract sockets and file systems from being embedded inside the program to being passed as an argument. This allows better flexibility and testability! Then he discussed capsicum, how an app can call the kernel and tell it to “lock me up”! I no longer need to access new resources. Capsicum is awesome, works well .. however it doesn’t scale! There is no guidance when something doesn’t work. With CloudABI a cloud provider can simply promise to run customers’ applications, pass those applications specific file descriptors and leave the app to do what it wants without touching any other part of the app! Cloudlibc is a C library built on top of the low level API Goal is 90% POSIX compliant! You get compiler errors when using unsupported functions.

After that is a session I was waiting for, “New OpenZFS features supporting remote replication” by Matt Ahrens. I have a soft spot for ZFS :) Matt started by giving an introduction to zfs send and receive and why it rocks! Only parts of the ZFS trees modified “after” the time we are sending from, need to be read! The more interesting part Matt discussed is what are the unique features to OpenZFS. This includes send stream size estimation and monitor its progress! Big receive improvements like receiving file with holes in them. Another cool thing added is bookmarks. This allows free’ing data on the sending side by deleting the relevant snapshot, and really only keep a bookmark to it. Now the really cool part is the new stuff. Resumable send and receive, hell yes! Another cool new feature, is a new checksum has been added per every record. As opposed to the old state of sending one checksum at the end of the stream. If a bad checksum is hit, ZFS receive aborts, and naturally you can resume it later! One last new feature, is receive performance improvements. In a benchmark, Matt said it improved performance by 6x which is great!

The next was “Multipath TCP for FreeBSD” by Nigel Williams. This one got me really excited! Part of that is that I knew very little previously about mptcp before. This got me too excited that I kept googling for more mptcp info all day and after the conference. Nigel explained mptcp design, and how it was done the way it is to work with the how the Internet is, as it is today. That is to work with  current middleboxes and NAT functions. MPTCP handshake piggybacks on top of the 3 way regular handshake. The two sides negotiate mptcp then the server informs client of other addresses it is reachable over and new sub connections are started. To me this is sort of the holy grail of WAN networking. A single TCP connection spanning multiple links for performance and redundancy, wow! I guess we’ll all just need to wait some more till the kernel implementations are more mature.


Next was “Packaging FreeBSD base system” by Baptiste Daroussin. He discussed packaging the base through pkgng and the challenges faced. I’m not really much of a packaging person, so it was good to hear many details that need to be take into account. Next up was “An Introduction to the Implementation of ZFS” by Kirk McKusick. If someone could make you appreciate the complexity of ZFS and operating system concepts it’s McKusick! He contrasted various filesystem features as they existed in UFS, and how ZFS is different. There was sort of a fire incident in the middle of the talk, we had to go to the street and then be back to continue. However overall it was an enjoyable session. The closing talk by “Dan Langille” was so much fun. The stats, the growth in the community, and even the goodbye auction were all something I won’t forget. Thanks to everyone who helped make BSDCan 2015 as fun, informative and enjoyable as it was to me.

Ahmed Kamal

BSDCan 2015 Trip Report: Steven Douglas

With the Foundation's help, I was able to meet and network with new people. Where I live, there are only a handful of people that even know what BSD is, let alone can talk at a high level about it. That was one of my favorite things, being around like minded people. I made many new contacts, most notably Peter Toth. Peter is working on iocage, which is a modern jail management utility with some very impressive features. In speaking with Peter, I met Sean Chittenden from Groupon. Sean is looking to perhaps implement iocage in conjunction with a new orchestration software that he is testing. Other than the social aspect, I also got some much needed help with my GSOC project.

At the conference, there were opportunities to learn every hour of every day. An expert in a field is never more than arm's reach away. It is very helpful to have questions that can be answered by the people who wrote the code. All of the talks I went to were absolutely fantastic, and I can't wait to watch the ones I missed when they are posted. My favorite talks were Steven Bourne's talk about his past, Matt Ahren's talk about code flow between our community and OpenZFS, and Multipath TCP by Nigel Williams. The speaker's all did fantastic jobs, and I hope that I am able to speak in the future.

This was my first BSDCan, and first BSD conference. I enjoyed every minute of it, and it happened so quickly. I hope to be back to BSDCan next year, and hope to make it VBSDCon and EuroBSDCon. Thank you to the Foundation for the financial assistance to help get me there!

Steven Douglas

Thursday, May 28, 2015

Another Data Center Site Visit - NYI

No Systems Administrators Were Harmed While Writing This Blog Entry

Mmm...  Freshly-unboxed servers.  There really is nothing better to wake up to in the morning.

Well, okay, coffee.  But new servers - definitely second.

In late April, the FreeBSD Foundation generously purchased more machines to keep the FreeBSD.org infrastructure operating smoothly.  While the new servers are not yet in production (a task the Cluster Administrators will undertake while at BSDCan in June), we have planned far in advance what we intend to do with the new hardware.

In mid-May, I spent several days at our East-Coast US colocation facility, racking, cabling, installing, and configuring the new servers.

As They Say in Real-Estate: Colocation, Colocation, Colocation

The new hardware is located at New York Internet in Bridgewater, New Jersey, who generously provides colocation services to the FreeBSD Project.  They have an amazing staff, and whether we are on-site or working with them through their ticket system, are always friendly, knowledgeable, and of course, helpful.

New Hardware Specs

In total, 14 new SuperMicro-based machines were purchased from iXsystems, for this site. They are all 1U servers, each with four 1TB drives, redundant power supplies, and gigabit ethernet.

Ten of the machines are single-socket Intel(R) Xeon(R) E3-1230 v3 CPUs, running at 3.30GHz (4-core with hyper-threading, providing 8 logical CPU threads), with 32 GB RAM.

Four of the machines are dual-socket Intel(R) Xeon(R) E5-2630 v3 CPUs, running at 2.40GHz (8-core with hyper-threading, providing 16 logical CPU threads per socket, 32 threads total), with 64 GB RAM.

For the Inner Geek in All of Us

No blog post about new hardware would be complete without pictures, right?

This is the front view of the ten single-socket machines:


And this is the back view of the same machines:


(Yes, I really do label each end of the network and serial console cables.)

Then, this is the front view of the four dual-socket machines (sorry for the blurry picture, though):


And the back view of the dual-socket machines:


Last but not Least...

On behalf of the FreeBSD Foundation and the FreeBSD Cluster Administration Team, I would like to thank New York Internet for generously providing the colocation space for our east-coast site, the NYI network operations team for all of their assistance during my visit, and especially all of the community investors that have generously donated to the FreeBSD Foundation.  All of your support is greatly appreciated.

Thank you for helping make FreeBSD better!

Monday, April 27, 2015

From the Trenches, Tips & Tricks Edition: Hacking "/ on ZFS" and GELI Encrypted Drives, the Old-School Way

Glen Barber is back to kick off our latest From The Trenches series: The Tips and Tricks Edition. 

All my personal machines run FreeBSD.

In fact, all my personal machines run FreeBSD-CURRENT. I do this primarily to keep track of changes that get committed to the head branch, so I can personally test changes (for the things I use, at least) before they get merged to the stable branches.

As one of the Release Engineers, I find it essential that, whenever possible, I find issues so they can be corrected before they are part of a release.

My primary work machine is a laptop, currently a Lenovo Thinkpad T540p. I picked this laptop, and all the other laptops before it, because it met my minimum requirements for a primary workstation: it is capable of supporting a large amount of RAM (16GB for my Thinkpad, 8GB for all previous laptops), an Intel Core i7 CPU, and I could replace the DVD drive with a second hard drive.

In addition to these hardware requirements, I also have a few personal requirements of any workstation - the drives must be encrypted, and the underlying filesystem must be ZFS.

For me, it is not so much about the data I have *on* the laptop that I need to protect, but the kinds of things within the FreeBSD Project I am permitted access. Without encrypted drives, a lost or stolen laptop would absolutely be my worst possible nightmare, because I only have my login passphrase protecting my data (GPG key, SSH keys, and so on).

Recent FreeBSD releases allow "/ on ZFS" installation with the option to enable GELI-based encryption. This predates my original installation, however, since each laptop I have purchased for the past several years used the hard drives from the previous laptop. According to zpool history, the installation was at least two and a half years ago, but I know it is much longer than that, because of zfs recv being one of the first things zpool history reports.

So, I needed to do things the old-fashioned way, and manually create the GELI-backed providers and perform the "/ on ZFS" installation myself.

While bsdinstall(8) may now cover the majority of use cases for such installations, there may be cases where someone specifically needs to do something a certain way that the installer does not provide.

Because I only had one hard drive in the system when the system was initially installed (a long time ago), I will only refer to one hard drive when describing the steps I used to perform the installation, for now.

I installed the system using the 9.0-RELEASE or 9.1-RELEASE memory stick installer (memstick.img), I cannot remember which, but that detail is not as important, since I did not use the installer anyway.

When I booted from the memory stick, the two drives recognized on the system were the internal hard drive, /dev/ada0, and the external USB flash drive for the installation, /dev/da0. The first menu screen has three options available: "Install", "Shell", "Live CD".

I selected "Live CD", and logged in as root (no password is necessary for the "Live CD" functionality). The hard drive did not have an operating system. Because I purchased the hard drive, in addition to the laptop, with the intention of replacing the laptop's drive, I did not need to remove any partitions from an existing installation. If I did need to remove partitions, I would have done so with:

# gpart destroy -F ada0
Here is where some technical details become important:
  • While you can install "/ on ZFS" on a drive partitioned with MBR (Master Boot Record), using GPT is far easier. In fact, I have forgotten much about how MBR partitioning is actually done.
  • When doing full disk encryption, you must keep /boot contents separate, otherwise loader(8) and the kernel will not be available when the BIOS hands over control to the operating system. As such, /boot should be given its own partition on the disk left unencrypted, and the rest of the system on its own encrypted partition.
I created four partitions on the drive. The first partition is for the boot blocks (not to be confused with the /boot contents), the second partition is for /boot, the third is for the encrypted system, and the fourth is for swap.
# gpart create -s gpt ada0
# gpart add -t freebsd-boot -s 512k -i 1 -l gptboot ada0
# gpart add -t freebsd-zfs -s 10G -i 2 -l bootfs ada0
# gpart add -t freebsd-swap -s 10G -i 3 -l swapfs ada0
# gpart add -t freebsd-zfs -s 180G -i 4 -l rootfs ada0
I decided to put the swap partition between the /boot partition and the rest of the system, in case I needed to increase or decrease the size of the /boot partition, it would be far easier (and safer) to do.

Then, I loaded the necessary kernel modules for ZFS and GELI:
# kldload /boot/kernel/opensolaris.ko
# kldload /boot/kernel/zfs.ko
# kldload /boot/kernel/geom_eli.ko
Now that GELI functionality is available, I created the backend provider for the ZFS dataset:
# geli init -b -a HMAC/SHA256 -e AES-CBC -l 256 \
    -s 4096 /dev/ada0p4
Then I attached the GELI provider, and wrote data from /dev/random to the new device /dev/ada0p4.eli:
# geli attach ada0p4
# dd if=/dev/random of=/dev/ada0p4.eli bs=4096

This took a while on the system this hard drive was originally installed, so I probably got coffee at this point. :-)

When the dd(1) command finished, I continued the installation.

I created temporary directories to use to import the pools after they were created:
# mkdir /tmp/zroot
# mkdir /tmp/zboot
Keep in mind, I am installing from a memory stick image, which by default, is read-only. The /tmp directory is writable, however, because it is a md(4)-backed memory disk filesystem.
# zpool create -O checksum=fletcher4 -O atime=off \
    -m /tmp/zboot zboot /dev/ada0p2
# zpool create -O checksum=fletcher4 -O atime=off \
    -m /tmp/zroot zroot /dev/ada0p4.eli
Then I made a few ZFS datasets for various paths:
# for i in var var/log var/tmp var/db usr usr/home \
    usr/compat usr/ports \
    usr/local tmp; do \
    zfs create zroot/${i} \
    done
I also made a separate ZFS dataset for the "bootfs" contents, and set the mountpoint to the /boot directory in the temporary working directory:
# zfs create zboot/boot
# zfs set mountpoint=/tmp/zroot/boot zboot/boot
On the memory stick installation media, the distribution sets are located in /usr/freebsd-dist. I extracted their contents into the newly-created filesystem:
# cd /tmp/zroot
# for i in base kernel lib32; do \
    tar -xf /usr/freebsd-dist/${i}.txz -C . \
    done
Then I wrote the bootcode to the first partition of the drive:
# gpart bootcode -b /tmp/zroot/boot/pmbr \
    -p /tmp/zroot/boot/gptzfsboot -i 1 ada0
Because the "bootfs" (/boot) and "rootfs" (everything else) are both ZFS, I needed to use the gptzfsboot bootcode for the "freebsd-boot" partition.

Now the system is installed, but I needed to make a few modifications before I was ready to reboot. In particular, set a root password, edit /etc/fstab to enable swap, edit /etc/rc.conf to enable the zfs rc(8) startup script, and edit /boot/loader.conf to load the geom_eli.ko, opensolaris.ko, and zfs.ko kernel modules at boot.
# chroot /tmp/zroot
# passwd root
[enter password]
# echo '/dev/gpt/swapfs none swap sw 0 0' \
    >> /etc/fstab
# echo 'zfs_enable="YES"' >> /etc/rc.conf
# echo 'geom_eli_load="YES"' >> /boot/loader.conf
# echo 'zfs_load="YES"' >> /boot/loader.conf
# exit
Before rebooting, I needed to make a few adjustments to where /boot from the zboot/boot dataset would be mounted at boot.
# zfs umount zboot/boot
# zfs set mountpoint=/realboot zboot/boot
This now makes the /boot directory mount as /realboot, so I then needed to point /boot in the zroot dataset to the correct place. This was easily solved with a symbolic link:
# cd /tmp/zroot
# ln -s boot /realboot
Now when the system boots, the filesystem will look something like this:
/bin
/sbin
/boot -> /realboot
/realboot
[...]
Finally, I needed to unmount the zroot dataset, and fix its mountpoints. I only needed to change the zroot mountpoint itself, since all children datasets adjusted their paths automatically.
# zfs umount -a
# zfs set mountpoint=/ zroot
At this point, the installation was complete. I rebooted the laptop, entered the GELI passphrase for /dev/ada0p4.eli when prompted, and was greeted by the "login: " prompt we have all grown to love.

Friday, March 13, 2015

15th Anniversary and Spring Fundraising Kickoff

I'm so excited to announce our spring fundraising campaign. I know it's not officially spring yet, but it sure feels like it here at Foundation headquarters in Boulder, Colorado. We're kicking off our fundraising campaign in conjunction with some other exciting events. There's so much to celebrate. First, we are proud to be a Platinum sponsor of AsiaBSDCon. This is the tenth AsiaBSDCon, with over 140 attendees planned, and 31 talks, providing a venue for all things BSD in Asia. People from around the world attend this conference to learn about the BSD operating systems, share their knowledge and experience, and work together to develop, hack, fix, improve, and document the various BSD operating systems.

The most exciting news we have is that we are celebrating our 15th anniversary supporting the FreeBSD Project and community worldwide! We have grown from our president and founder, Justin Gibbs, creating a non-profit to support FreeBSD, to an eight member board with 7 staff members. In case you missed it earlier, check out Justin's interview about the history of the Foundation on BSDNow. 

As the first employee, 9 years ago, I've witnessed incredible growth in our ability to support the Project and community. The year we were founded we raised a whopping $7,000. My first year with the Foundation, in 2006, we raised a little over $100,000. And, last year we raised $2,436,194, spending $877,412 on the project.

When we first started out, we focused on funding project development, conference sponsorships, and travel grants. Fifteen years later, we have increased support in those areas and have now grown to providing legal support for the Project; purchasing and helping manage hardware for FreeBSD infrastructure; providing release engineering support for consistent and timely releases; creating marketing literature and presentations that not only inform people of what FreeBSD is, but also provides detailed information on what's in new releases; attending more conferences to promote FreeBSD; and publishing a professional online FreeBSD magazine, The FreeBSD Journal.

To celebrate our anniversary, we are kicking off a fundraising campaign to help broaden the reach of our mission, by adding 500 new community investors in the next four weeks. What's a new community investor? An individual or organization that makes their first 2015 donation during this spring campaign. 

Why donate to the Foundation? Your donations will help us continue and increase our support in the following areas:
  • Funding improvement and development projects, including: Native ISCSI kernel Stack, Updated video console (Newcons), UEFI system boot support, Capsicum component framework, IPv6 support in FreeBSD, Auditdistd improvements for FreeBSD cluster, and adding modern AES modes to OpenCrypto (to support IP/SEC).
  • Helping to provide consistent and on-time releases.
  • Educating the public and promoting FreeBSD with tools like our high-quality FreeBSD 10X Brochure and company visits to help
  • facilitate collaboration efforts with the Project.
  • Sponsoring BSD conferences and summits in Europe, Japan, Canada, and the US.
  • Protecting FreeBSD IP and providing legal support to the Project.
  • Purchasing hardware to build and improve FreeBSD project infrastructure.
For the last 15 years, you as a community have allowed us to make an major impact on the FreeBSD Project and Community. Please help us continue and increase our support by making a donation today.


Deb Goodkin, Executive Director

FreeBSD From the Trenches: Using autofs(5) to Mount Removable Media

This next FreeBSD From the Trenches story come to us from Edward Tomasz NapieraÅ‚a who shares his work on the new FreeBSD automounter.

My big project for 2014 was the new FreeBSD automounter.  Like any proper FreeBSD Foundation sponsored project, it included the usual kind of documentation - man pages and the Handbook chapter.  But there is no document that shows how it works inside, from the advanced system administrator or a power user point of view.

So, here it is.  The article demonstrates how modular the automounter is, and how easy it is to adopt to any mount-related situation you might have, using recently added removable media support as an example.  (And it shows some related mechanisms as a bonus.)

autofs(5) Basics

The purpose of autofs(5) is to mount filesystems on access, in a way that's transparent to the application. In other words, filesystems get mounted when they are first accessed, and then unmounted after some time passes. The application trying to access the filesystem doesn't even notice this event, apart from a slight delay on first access.  It's a mechanism similar to ones available in other systems, in particular to OS X.  It's a completely independent implementation, it's just that OS X is the other operating system I use.

Automounting requires cooperation of four things: the kernel filesystem, autofs.ko, which is responsible, among other things, for "pausing" the application until the filesystem is actually there; the automountd(8) daemon, which is the component that retrieves configuration information from maps (this includes fetching it from remote sources, such as LDAP) and actually mounts the filesystems; the automount(8) utility for various administrative purposes; and then the autounmountd(8) daemon to, well, unmount the filesystems mounted by automountd(8) after a timeout.

Setting it up is fairly simple: you obviously need to have autofs(5) enabled in /etc/rc.conf:
autofs_enable="YES"
And you need to have the autofs(5) daemons running - just like other deamons in FreeBSD those will get started at system bootup if autofs_enable was set; otherwise you need to start them by hand:
# /etc/rc.d/automount start
# /etc/rc.d/automountd start
# /etc/rc.d/autounmountd start
The kernel driver will get loaded automatically, you can see it in kldstat(8) output.

autofs(5) and Removable Media

Note that at the time of this writing, this is only available in FreeBSD 11-CURRENT. This will change soon.

The main configuration file for autofs(5) is /etc/auto_master; you need to uncomment this line:
/media -media -nosuid
This basically says that there is a /media directory, and automount will mount the "-media" map there, and everything that gets mounted there will have the "nosuid" mount option, for security reasons.

If you already had autofs(5) running before uncommenting the line, you must refresh its configuration by running automount(8) as root; run it as "automount -v" for a detailed explanation of what it does.  It looks like this:
# automount -v
automount: parsing auto_master file at "/etc/auto_master"
automount: done parsing "/etc/auto_master"
automount: unmounting stale autofs mounts
automount: skipping /, filesystem type is not autofs
automount: skipping /dev, filesystem type is not autofs
automount: leaving autofs mounted on /net
automount: mounting new autofs mounts
automount: autofs already mounted on /net
automount: nothing mounted on /media; mounting
automount: mounting map -media on /media, prefix "/media", options "nosuid"
If you run mount(8), you will see so called "trigger nodes" of type autofs(5):
# mount
/dev/ada0p2 on / (ufs, local, noatime, journaled soft-updates)
devfs on /dev (devfs, local, multilabel)
map -hosts on /net (autofs)
map -media on /media (autofs)

Basic usage

With all that done, plug a drive into USB, and here is what happens in a real-world case:
[trasz@brick:~]% ll /media
total 9
drwxr-xr-x  3 root wheel  512 Feb 24 12:54 .
drwxr-xr-x 30 root wheel 1024 Feb 24 12:28 ..
drwxr-xr-x  1 root wheel 4096 Jan  1 1980 ADATA UFD
drwxr-xr-x  3 root wheel  512 Feb 24 12:54 md0
[trasz@brick:~]% cd /media/ADATA\ UFD
[trasz@brick:/media/ADATA UFD]% ll
total 10117
drwxr-xr-x 1 root wheel    4096 Jan  1  1980 .
drwxr-xr-x 3 root wheel     512 Feb 24 12:54 ..
drwxr-xr-x 1 root wheel    4096 Nov 24 00:03 .Spotlight-V100
drwxr-xr-x 1 root wheel    4096 Nov 24 00:03 .Trashes
-rwxr-xr-x 1 root wheel    4096 Nov 24 00:03 ._.Trashes
drwxr-xr-x 1 root wheel    4096 Jan 13 11:24 .fseventsd
drwxr-xr-x 1 root wheel    4096 Nov 22 22:44 Bonus
-rwxr-xr-x 1 root wheel 3309568 Nov 24 14:50 DSC05996.JPG
-rwxr-xr-x 1 root wheel 4063232 Nov 24 14:50 DSC05997.JPG
-rwxr-xr-x 1 root wheel 2953199 Nov 25 21:40 DSC05998.JPG
drwxr-xr-x 1 root wheel    4096 Nov 22 18:24 Meshuggah
drwxr-xr-x 1 root wheel    4096 Nov 22 21:06 System Volume Information
[trasz@brick:/media/ADATA UFD]% mount
/dev/ada0p2 on / (ufs, local, noatime, journaled soft-updates)
devfs on /dev (devfs, local, multilabel)
map -hosts on /net (autofs)
map -media on /media (autofs)
/dev/da0s1 on /media/ADATA UFD (msdosfs, local, nosuid, automounted)
[trasz@brick:/media/ADATA UFD]% cd /
[trasz@brick:/media/ADATA UFD]% sudo automount -u
[trasz@brick:/media/ADATA UFD]% mount
/dev/ada0p2 on / (ufs, local, noatime, journaled soft-updates)
devfs on /dev (devfs, local, multilabel)
map -hosts on /net (autofs)
map -media on /media (autofs)
Two things to notice here: first, the "ADATA UFD" is a factory default
filesystem label on the flash drive.  If there was no filesystem label,
autofs(5) would use device name instead - in this case, that would
be "da0s1".  Second - if you don't want to wait for autounmountd(8)
to unmount the automounted volume, you can use "automount -u".  Or
"automount -fu", if you want to force unmount.

Not So Basic Usage

Take a close look at the directory listing for /media in previous example. Did you notice the "md0" there?  It looks like a device node for memory disk (md(4)), but is a directory.  That's a leftover from my earlier experimentation, and shows an interesting feature of autofs(5)-based automounter: it's not limited to removable media, it can mount everything that's available for mounting.  In this case it's a memory disk (kind of ramdisk, see "man mdconfig").  It can also be an iSCSI lun.  And, of course, a removable media.  How does that work?

GEOM

In FreeBSD, GEOM is a name of what could otherwise be called a block device layer.  It's a piece of code that manages all the "disk-like devices", both physical and virtual: SATA/SAS/FC/NVME/USB drives, memory disks, iSCSI LUNs, partitions, encrypted GELI volumes etc.

GEOM has another meaning: an instance of GEOM class.  The "class" here means the "kind" of device, and the instance is an actual device of that kind. It's easiest to explain it with an example:
# geom disk list
Geom name: cd0
Providers:
1. Name: cd0
   Mediasize: 0 (0B)
   Sectorsize: 2048
   Mode: r0w0e0
   descr: MATSHITA DVD/CDRW UJDA775
   ident: (null)
   fwsectors: 0
   fwheads: 0

Geom name: ada0
Providers:
1. Name: ada0
   Mediasize: 250059350016 (233G)
   Sectorsize: 512
   Stripesize: 4096
   Stripeoffset: 0
   Mode: r2w2e3
   descr: Samsung SSD 850 EVO 250GB
   lunid: 5002538da000f602
   ident: S21PNSAFC02149R
   fwsectors: 63
   fwheads: 16

Geom name: da0
Providers:
1. Name: da0
   Mediasize: 7654604800 (7.1G)
   Sectorsize: 512
   Mode: r0w0e0
   descr: ADATA USB Flash Drive
   lunname: USB MEMORY BAR
   lunid: 2020030102060804
   ident: 14A0711312300023
   fwsectors: 63
   fwheads: 255

# geom part list
Geom name: ada0
modified: false
state: OK
fwheads: 16
fwsectors: 63
last: 488397127
first: 34
entries: 128
scheme: GPT
Providers:
1. Name: ada0p1
   Mediasize: 65536 (64K)
   Sectorsize: 512
   Stripesize: 4096
   Stripeoffset: 1024
   Mode: r0w0e0
   rawuuid: 42dc1b8b-c49b-11e3-8066-001c257ac65f
   rawtype: 83bd6b9d-7f41-11dc-be0b-001560b84f0f
   label: (null)
   length: 65536
   offset: 17408
   type: freebsd-boot
   index: 1
   end: 161
   start: 34
2. Name: ada0p2
   Mediasize: 236223201280 (220G)
   Sectorsize: 512
   Stripesize: 4096
   Stripeoffset: 1024
   Mode: r1w1e1
   rawuuid: 42dc921f-c49b-11e3-8066-001c257ac65f
   rawtype: 516e7cb6-6ecf-11d6-8ff8-00022d09712b
   label: (null)
   length: 236223201280
   offset: 82944
   type: freebsd-ufs
   index: 2
   end: 461373601
   start: 162
3. Name: ada0p3
   Mediasize: 13836045312 (13G)
   Sectorsize: 512
   Stripesize: 4096
   Stripeoffset: 1024
   Mode: r1w1e0
   rawuuid: 21a8eef9-a0d4-11e4-ab80-001c257ac65f
   rawtype: 516e7cb5-6ecf-11d6-8ff8-00022d09712b
   label: (null)
   length: 13836045312
   offset: 236223284224
   type: freebsd-swap
   index: 3
   end: 488397127
   start: 461373602
Consumers:
1. Name: ada0
   Mediasize: 250059350016 (233G)
   Sectorsize: 512
   Stripesize: 4096
   Stripeoffset: 0
   Mode: r2w2e3

Geom name: da0
modified: false
state: OK
fwheads: 255
fwsectors: 63
last: 14950399
first: 1
entries: 4
scheme: MBR
Providers:
1. Name: da0s1
   Mediasize: 7654576128 (7.1G)
   Sectorsize: 512
   Stripesize: 0
   Stripeoffset: 28672
   Mode: r0w0e0
   rawtype: 12
   length: 7654576128
   offset: 28672
   type: !12
   index: 1
   end: 14950399
   start: 56
Consumers:
1. Name: da0
   Mediasize: 7654604800 (7.1G)
   Sectorsize: 512
   Mode: r0w0e0
See?  I've used the geom(8) command to get the information about two GEOM classes: "disk", and "part".  The first one returned information about three instances of the disk class: the DVD drive, the SSD, and the flash.  The second one returned information on the partitions known to the system. Everything that is potentially mountable - a physical disk, a partition, encrypted ELI volume, multipath device, RAID3 volume, memory disks, even volume labels - it all has its GEOM class and can be queried in a similar way.  To see all the GEOM instances in the running system, use:
# sysctl kern.geom.conftxt
Now, notice the "Mode" lines.  Like the one for ada0: "r2w2e3".  Those are three usage counters for ada0 GEOM: read, write, and exclusive.  They are non-zero, because ada0 is used: there are three partitions on it; three instances of PART GEOM class hold it opened.  The partitions, just like any other GEOM nodes, have their own counters.  Take a look at the first one, ada0p1: the mode there is "r0w0e0".  This means it's not open by anything.  It's, in other words, available for mounting.  If you check the MD geom class:
# geom md list 
Geom name: md0 
Providers: 
1. Name: md0
   Mediasize: 1073741824 (1.0G) 
   Sectorsize: 512
   Mode: r0w0e0
   type: swap
   access: read-write
   compression: off
   length: 1073741824
   fwsectors: 0
   fwheads: 0
   unit: 0
You will see the same thing: it's not opened.  That's the first thing the autofs(5) "-media" map checks for: zero access counts; if the counts are not zero, it means the node is used by something: it's either mounted (like ada0p2, mounted on /), or there is something "on top of it" - like ada0.

But why there is no /media/ada0p1?  Because it's not mountable; there is no filesystem there.  It's a boot loader partition.  How does autofs(5) figure it out?

fstyp(8)

Before we can do anything with a filesystem, we need to determine what kind of filesystem it is - and whether it actually is a supported filesystem in the first place.  That means we need a piece of code that can take a look at it and determine if it has a format it recognizes.

It is possible to use file(1) for this, eg:
# file -s /dev/md0
Vermaden's sysutils/automount port uses this approach.  There are a few problems with doing it this way, though.  First, the output, for a typical FAT filesystem, looks like this:
/dev/md0: DOS/MBR boot sector, code offset 0x3c+2, \
 OEM-ID "BSD4.4  ", sectors/cluster 32, root entries \
 512, sectors/FAT 256, sectors/track 63, heads 255, \
 sectors 2097144 (volumes > 32 MB) , serial number \
 0x668a120e, unlabeled, FAT (16 bit)
It's not particularly easy to parse.  It's even harder to extract the volume label.

Second, file(1) can recognize all kinds of file types, from JPEG to 6502 assembly.  This means that if there are some strange data on the removable media, instead of a filesystem we expect, the file(1) will output something our script wasn't tested against, making the first problem even harder.

Third, file(1) had its share of security bugs, eg CVE-2014-1943, CVE-2014-9620, or CVE-2014-3710.

For this reason I've decided the proper fix would be to just write a new utility. The strange name - "fstyp" - comes from the utility of the same name, installed by default on Solaris, IRIX, OS X, and perhaps most other UNIX systems.

The fstyp(8) addresses the file(1) issues: the output is easily parsable (just a filesystem name, one word), it only recognizes filesystems supported by FreeBSD, and uses Capsicum sandboxing to make sure that even if there is a vulnerability, its impact is limited to incorrectly reporting the filesystem type.  It's a good topic for another article, but in short - in FreeBSD, every process can enter what's called a "capability mode". It's one way - a process can enter it, but there is no way to exit it.  Child processes inherit the mode. In capability mode, kernel will deny all attempts to open new files, create sockets, attach the shared memory segments etc, but the process is pretty much free to do anything it likes with the file descriptors it already had opened before entering the capability mode - and it can receive other file descriptors over a UNIX socket.  So, the fstyp(8) utility opens the device file, then calls cap_enter(2), which switches it into capability mode, and then continues execution, reading from the device to determine what's there.  Should it be compromised, it won't be able to execute /bin/sh, it won't be able to open a socket to transmit the data to some external host, etc.

The "-media" Map

Those are the components underneath the autofs(5), but how does it fit together? Let's start with the actual map.  In FreeBSD, special maps (the ones with names starting with "-") are just executables in /etc/autofs/:
# ls -al /etc/autofs
total 36
drwxr-xr-x   2 root  wheel  512 Feb 14 21:18 .
drwxr-xr-x  25 root  wheel 3072 Feb 24 11:22 ..
-rwxr-xr-x   1 root  wheel 1010 Oct 17 11:26 include_ldap
-rwxr-xr-x   1 trasz wheel   43 Aug 17  2014 include_nis
-rwxr-xr-x   1 root  wheel  367 Oct 17 11:26 special_hosts
-rwxr-xr-x   1 root  wheel 2294 Dec  6 10:15 special_media
-rwxr-xr-x   1 root  wheel  355 Feb 14 21:17 special_noauto
-rwxr-xr-x   1 root  wheel   97 Oct 17 11:26 special_null
-rwxr-xr-x   1 root  wheel  357 Aug 22  2014 special_smb
See the special_media?  That's the one.  It's a shell script.  The reason it's in /etc is that the system administrator can modify it if required, or add new special maps.

Now, let's try to run it by hand, as root:
# /etc/autofs/special_media
ADATA UFD
md0 

# /etc/autofs/special_media md0
-fstype=msdosfs,nosuid  :/dev/md0
That's exactly how automountd(8) uses it, after the kernel component notifies it that it needs the /media directory taken care of.  It's described in more detail in the auto_master(5) manual page.  The shell script is pretty well commented, and I don't think there is any point in explaining it here.

Bottom line:
the core autofs itself doesn't know anything about removable devices; the special map "-media" does: it queries GEOM for the list of all disk-like nodes that are not in use, and then uses fstyp(8) to determine whether they contain a useful filesystem.  UNIX.  Modularity.  Plain text.  ;-)

Cache

Now, let's create a second memory disk, 1GB in size (the "1g" below) to see if it all works as intended:
# mdconfig -s1g
md1
# newfs_msdos /dev/md1
newfs_msdos: cannot get number of sectors per track: \
 Operation not supported
newfs_msdos: cannot get number of heads: \
 Operation not supported
newfs_msdos: trim 8 sectors to adjust to a multiple of 63
/dev/md1: 2096576 sectors in 65518 FAT16 clusters \
 (16384 bytes/cluster)
BytesPerSec=512 SecPerClust=32 ResSectors=1 FATs=2 \
 RootDirEnts=512 Media=0xf0 FATsecs=256 SecPerTrack=63 \
 Heads=255 HiddenSecs=0 HugeSectors=2097144

# ll /media
total 5
drwxr-xr-x   3 root wheel  512 Feb 24 12:25 .
drwxr-xr-x  30 root wheel 1024 Feb 23 09:04 ..
drwxr-xr-x   1 root wheel 4096 Jan  1 1980 ADATA UFD
drwxr-xr-x   3 root wheel  512 Feb 24 12:25 md0
Whoops.  Where is /media/md1?

There is one more mechanism for the whole thing to work correctly: the autofs(5) cache needs to be dealt with.

The first paragraph mentioned that it's automountd(8) that does all the map parsing - including running the /etc/autofs/special_media - and actual mounting. Doing it every time someone accesses the /media directory - or any directory, for that matter - would kill performance.  For this reason, after the kernel component asks the automound(8) to do its magic, it doesn't do that again until some time later.  In most cases it doesn't matter - the list of NFS exports for a given host doesn't change too often - but in case of removable media it's not acceptable.  The cache needs to be flushed, using "automount -c".  After that, the subsequent lookup in /media will trigger automountd(8), which will query the devices list and refresh the directory contents.

This obviously needs to happen automatically.  And if you actually went and opened /etc/auto_master in a text editor, you would have noticed this:
# When using the -media special map, make sure to edit devd.conf(5)
# to move the call to "automount -c" out of the comments section.
The devd(8) is a daemon responsible for listening for notifications from the kernel and running whatever is configured in its config, /etc/devd.conf. There are all kinds of things there, from running utilities to upload firmware for various USB devices, to launch moused(8) when a mouse gets connected, to switching power profiles, to... discarding autofs(5) caches.  It looks like this:
notify 100 {
   match "system" "GEOM";
   match "subsystem" "DEV";
   action "/usr/sbin/automount -c";
}
If you do "man devd.conf", you will see the description of those events. Note that, just like the "-media" map works the same way for flash drives and encrypted volumes over multipath over iSCSI, this mechanism does not care about any specific hardware either.

Caveats

Two, really.  First: you need to run 11-CURRENT.  Second: the nodes in /media never disappear.  I expect to merge this support to 10-STABLE after the second issue is addressed.