There was only one notable change in the src repo in July 2023. But first, a little background info:

A long time ago, I started on a project that makes anonymous remote code injection and PLT/GOT redireciton techniques over the ptrace boundary easy in one little consumable API. The end-goal of the tool is to support injection of shared objects in a completely anonymous manner, and to be able to hijack PLT/GOT entries to point to their counterpart in the newly-injected shared object. I started on this work well over a decade ago (it has roots back to ideas I had in 2003.) The project is aptly named libhijack[0].

One common technique is to rely on shared memory-backed file descriptors, writing the shared object to the shmfd, lseek(shmfd,0,seek_set), then fdlopen(shmfd). This causes the RTLD to load the shared object from anonymous memory. In fact, I've published a PoC[1] that shows this very technique.

I'm currently working on providing shmfd-based anonymous shared object injection support in libhijack, so it can be performed over the ptrace boundary with a simple API call. Imagine something like this on a webserver running nginx:


pid_t pid = get_nginx_pid();
const char *path_to_shared_object = "/lib/libpcap.so.8";
InjectSharedObject(pid, path_to_shared_object);


This would cause the target process to create a new shmfd, write the contents of libpcap.so.8 to it, seek to the beginning, then load it via fdlopen.

I have this mostly implemented, but I'm running into some ptrace oddities.

Where this ties into HardenedBSD's src is: while writing this code, I kept thinking to myself, how would I defend against this kind of technique? Results from fstat(2) aren't helpful as there's no way to detect that we're looking at an anonymous shared memory-backed file descriptor.

However, I noticed that calling fstatfs(2) on the shmfd causes undocumented behavior: fstatfs(2) will return EINVAL. The underlying code for shared memory file descriptors doesn't implement a handler for fstatfs(2), causing the syscall to return EINVAL.

I then added code to the RTLD to check the return value of a call to fstatfs(2) on the file descriptor passed in when RTLD hardening is enabled. If fstatfs(2) fails and sets errno to EINVAL, we prohibit loading the object.

This would force an attacker to fully implement what I call a "remote RTLD": an out-of-process RTLD that loads objects over a boundary (the boundary in this particular case being ptrace.) The attacker would have to force the application to call mmap (which libhijack supports), inject into those new mapping (which libhijack supports), but then perform all the RTLD logic and fixups over the boundary (which libhijack does not support.)

My hope is that one day, libhijack gains that last little bit. That last little bit is the most complex bit. That's why I'm going the shmfd route first: to prove the concept and to flush out a "rough draft" of what's in my head.

HardenedBSD's PaX NOEXEC-inspired strict W^X implementation is effective over the ptrace boundary, further frustrating the concept of a remote RTLD.

[0]: https://github.com/SoldierX/libhijack
[1]: https://git.hardenedbsd.org/shawn.webb/random-code/-/tree/main/memdlopen

May 2023 has been a very busy month behind-the-scenes here at HardenedBSD. I want to start off, though, by talking about infrastructure downtime as it's the most important take-away from this status report.

On 03 June 2023, in preparation for our move to Colorado, I will take down the following servers (hosted out of my home):

1. GitLab ( https://git.hardenedbsd.org/ )
2. The main installers mirror ( https://installers.hardenedbsd.org/pub )
3. The package build servers
4. The OS and update build servers
5. The Tor Onion Service nodes

The last pre-move builds will start tomorrow on 01 June 2023. They should complete by 02 June 2023. The builds will be accessible via our two mirrors. The link to the list of up-to-date HardenedBSD mirrors is: https://hardenedbsd.org/content/mirrors

We're grateful for those who provide mirrors. You ensure that HardenedBSD's installers (and hopefully one day: packages) are available regardless of our own infrastructure. The project, and I'm sure the entire community, thanks you.

At the time of writing this status report, we have a tentative temporary hosting solution with limited bandwidth capabilities. We're evaluating multiple hosting possibilities, with this tentative temporary hosting solution being only a backup in case those other options don't pan out. We do not have an ETA for bringing the infrastructure back online.

Now, let's get on with the fun stuff! The HardenedBSD Foundation is now fully registered in Colorado! We're moving the headquarters from Maryland to Colorado while still maintaining a presence my Maryland. Keeping a Maryland presence allows us to continue serving local communities. For those curious, our Colorado registration can be found at: https://www.coloradosos.gov/ccsa/ViewReports.do?ceId=325482

We will soon update the US federal side, the completion of which finishes the move to Colorado. The HardenedBSD Foundation Board of Directors has been a great help in supporting the move.

And now, in the src tree:

1. Shawn fixed NULL pointer derefs in the following portions of the kernel:
   • mrsas(4)
   • powerpc code
   • NVIDIA Tegra CPU Frequency code <- integer overflow, too
   • iw_cxgbe
   • drm2 <- integer overflow, too
   • hdspe-pcm
   • linsysfs(5)
   • beri's virtio support <- integer overflow, too
   • firewire(4)
   • qnnxe(4)
   • axgbe pci code <- integer overflows, too
   • mmc(4)
2. On 14-CURRENT (and soon in hbsdfw), SHM hardening places restricitons on what can be done with the shared memory subsystem (see `shm_open(2)`.) This feature is launching with one technique (with more planned):
   • Use of `shm_open(2)/__sys_shm_open2` system calls is prohibited when:
     1. The `hardening.harden_shm` sysctl tunable is enabled;
     2. The process has not opted out of the feature;
     3. The process has entered capability mode (aka, Capsicum mode)
3. FreeBSD introduced an old vulnerability, the infamous (argc == 0) CVE, but this time in the linuxulator. Shawn ensured that HardenedBSD remains invulnerable.
4. Shawn fixed the 14-CURRENT builds by fixing a FreeBSD commit for building ELF Toolchain's `strings(1)` utility. By default, we use llvm's.

And in ports:

1. MrUnix disabled the JIT and re-enabled MPROTECT for www/ungoogled-chromium

In last month's status report, I mentioned that I decided to punish myself by running HardenedBSD 14-CURRENT/amd64 with Cross-DSO CFI enabled for base userland on my primary laptop, even during this move to Colorado. Shawn is happy to report that two months in, and he's still rocking (and keeping up-to-date with) Cross-DSO CFI! :-)

As a reminder, we mirror our src and ports repos on GitHub. They will remain available (and updated as best as I can). Links below.

I missed February's status report, so March's will contain both months. Let's start off with a bit of personal news, though. My wife, our two dogs, and I are officially moving to Colorado! Our move-out date is June 10th.

Additionally, my wife and I have set up a GoFundMe. Please note that donations to the GoFundMe are NOT eligible for tax deductions, but donations directly to The HardenedBSD Foundation are.

I plan to purchase a number of pieces of equipment to help with the move to make sure that the equipment arrives intact. HardenedBSD could still definitely use your continued support, for which I and the rest of the community are grateful.

This move obviously will impact the project. I've spent most of March working on getting the project and the HardenedBSD Foundation ready administratively for the move. There's only three things left to do, and the Foundation will have officially moved headquarters from Maryland to Colorado. We will be leaving a skeleton organziation in Maryland in case we want to do Maryland-specific things. For example, there could be some future Maryland-specific fundraiser or service opportunity.

I plan to keep the infrastructure online as long as I can. I suspect the infrastructure will be taken offline on around 09 June 2023, when we start loading the truck. However, there's a chance that the infrastructure will be taken down up to a week in advance. Since we don't know where we're going to land, I cannot provide an ETA as to when the infrastructure would be back online.

We plan to rent only for up to one year at most. We want to explore Colorado a bit to find out where we want to plant our roots. So there will be another infrastructure-disrupting move on the inside of a year from June. Hopefully that move will be to the home in which we retire in two-ish decades. I'm going to make having full self-hosting capabilities a priority in our "forever" home. So, now, let's get on to the src and ports updates since the end of January 2023.

In src:

1. tarfs is marked as insecure.
2. tarfs has been opted into -ftrivial-var-auto-init=zero.
3. geli has been opted into -ftrivial-var-auto-init=zero.
4. FreeBSD merged llvm 15 into base, so some fixups were needed on our side.
5. pf has been opted into -ftrivial-var-auto-init=zero.
6. pfsync has been opted into -ftrivial-var-auto-init=zero.
7. pflog has been opted into -ftrivial-var-auto-init=zero.
8. Default net.inet6.icmp6.nodeinfo to 0. This prevents host information disclosures via ICMP6.
9. Added some safty nets in the new netlink code
10. Dislabed netlink multiple times in multiple ways, chasing FreeBSD's continued attempt at enabling netlink by default.
11. cfi-icall has been disabled for various NFS-related applications.
12. The TTY pushback vulnerability has been mitigated by hardening the TIOCSTI ioctl. Attempts to use the TIOCSTI ioctl node will fail with EPERM. A new per-jail sysctl tunable (`hardening.harden_tty`) has been added, defaulted to 1.
13. The default packet TTL is randomized at boot-time, with a minimum of 32 hops. This feature is controlled by the new `HBSD_RESIST_FINGERPRINTING` kernel option, which is enabled by default. This feature can help with resisting fingerprinting attacks and preventing information leaks.
14. 0x1eef fixed a few typos in libhbsdcontrol.3.

SPECIAL NOTE ON ITEM 13: This has the potential for network disruption. I probably need to adjust the minimum hops as 32 might be too limiting. Please reach out to me directly if you are unable to get to resources that you used to get to. The sysctl node being set is `net.inet.ip.ttl`, and that can be adjusted anytime post-boot.

In ports:

1. DTRACE is disabled for www/node18.
2. CFI is disabled for security/sudo.
3. java/openjdk11 was fixed.
4. MrUNIX disabled PaX MPROTECT for bin/kdeconnect-app
5. MrUNIX disabled PaX MPROTECT for multimedia/mpv
6. MrUNIX disabled PaX MPROTECT for www/librewolf
7. MrUNIX disabled PaX MPROTECT and PaX PAGEEXEC for www/otter-browser
8. MrUNIX disabled PaX MPROTECT for www/tor-browser
9. MrUNIX enabled CFI and SafeStack for x11-wm/i3-gaps

HardenedBSD Firewall (hbsdfw)

hbsdfw is a fork of OPNsense based on HardenedBSD 13-STABLE.

A new build of hbsdfw is out. We still don't provide official in-place updatesupport, so your update process should be as follows:

1. Backup your existing config
2. Reinstall with the new image
3. Restore the config

The default username and password is:

Username: root
Password: hbsdfw

HardenedBSD Infrastructure

We now regularly build 13-STABLE and 14-CURRENT, on the first and fourteenth day of each month. The 13-STABLE and 14-CURRENT package build jobs are still kicked off manually, but usually follow on the same day or the day after the installation media build.

We're looking for more mirrors for the installation media. We're grateful for those who currently run public mirrors.

On Sunday, 27 Nov 2022, we will be temporarily taking down the HardenedBSD development and build infrastructure. We are working to find a new home for it and will be working those details out over the next few months.

We provide a read-only mirror of src and ports at GitHub:

src: https://github.com/HardenedBSD/hardenedBSD
ports: https://github.com/HardenedBSD/ports

Here are the short-term physical infrastructure changes:

1. GitLab will (hopefully) be brought back up within one week.
2. The Tor Onion Service endpoints will be brought back up at around the same time.
3. The nightly build server will be taken offline for an extended period of time until we find a new home for it.
4. The binary update server will be taken offline and brought back up at around the same time as the nightly build server.
5. The package building servers will be taken offline for an even more extended period of time. We encourage our community to learn how to use Poudriere to build their own packages.

Here are the short-term infrastructure policy changes:

1. Once the nightly build server is back online, we will no longer produce nightly builds. We will likely build twice per month on a bi-weekly schedule.
2. We will do the same for the binary update server. We will do our best to pair the hbsd-update builds with the bi-weekly builds to keep our users more in-sync with what is installed versus the latest update artifact.

If you know of a hosting provider within a two to four hour drive of Baltimore, Maryland, USA that would host our infrastructure for free with at least one public IPv4 address, please reach out. As a reminder, all donations--whether that's funding, equipment, services, or otherwise--are eligible for tax deductions in the US.

In full, our infrastructure takes around 24U, but we can get around with 16U. The physical infrastructure changes outlined above would be around 5U worth of equipment.

We appreciate the continued help and support from the community. HardenedBSD's continued success is enabled by the many generous ways in which the community contributes to the project.

Please reach out to me (shawn.webb@hardenedbsd.org and/or foundation@hardenedbsd.org) if you have any questions, comments, or concerns.

I apologize for the delay in getting the September 2022 status report out. But alas, it has arrived!

My time was spent mostly on infrastructure. We're slowly aging out some incredibly old servers in our infrastructure, occasionally not by choice. The Dell R410 server that ran our auto-sync cron jobs decided to die. So I rebuilt the auto-sync jail on another Dell R410 server of the same age, but the PERC controller decided to die. So now, the auto-sync jail is hosted on another server--but this time on a performant, stable system.

In src and ports land, I spent most of my time just resolving the occasional merge conflict.

In src:

1. Shawn ensured that the HardenedBSD copyright is always applied
2. Loic did some house cleaning with a few files in src
3. Loic removed leftover cruft from our LibreSSL-in-base experiment
4. MrUnix0 changed the HardenedBSD pkg repo configuration in 14-CURRENT to use HTTPS rather than HTTP. We're still exploring whether this change can be safely MFC'd to 13-STABLE, but we're being very conservative here.
5. FreeBSD updated `less(1)` to v608, which introduced a number of CFI violations. Shawn fixed two that were readily apparent.
6. Loic set `-fstack-protector-strong` for the kernel.
7. Loic fixed a few compiler warnings/errors when using a modified kernel config.

In ports:

1. Shawn enabled PulseAudio support for net/freerdp. Having audio over RDP seems pretty useful.
2. Loic enabled the sort plugin for editors/pluma.
3. Loic added games/scratch
4. Loic fixed the uname output in sysutils/mate-system-monitor
5. Loic disabled Java support by default for editors/libreoffice
6. Loic Fixed textproc/docbook2mdoc
7. Shawn fixed the llvm compiler toolchain component tests, fixing CFI applicability detection
8. Loic forced lld for graphics/cimg, science/cdo, and math/octave
9. Loic disabled PaX MPROTECT for emulators/qemu70
10. Loic fixed java/openjdk11
11. MrUnix disabled PaX MPROTECT and PaX PAGEEXEC for games/assaultcube
12. MrUnix disabled PaX MPROTECT and PaX PAGEEXEC for x11/lumina-core
13. MrUnix disabled PaX MPROTECT and PaX PAGEEXEC for games/xonotic

I did a new build of hbsdfw in late September, but I didn't get around to deploying it at home as a good first test. I'm following some of the work the OPNsense folks are doing and it seems best to hold off on a new build until some things settle down in their core repo. I plan to kick off a new build once I'm confident the dust has settled.

Upcoming plans:

Many of those in the HardenedBSD community know that I've worked (incredibly slowly) off-and-on throughout the years on Cross-DSO CFI support in HardenedBSD. In October, I plan to resume that work starting mid-October. Here's where we stand on Cross-DSO CFI today:

I can compile (nearly) the entire dynamic world with Cross-DSO CFI. However, there is an interesting recursion issue at early application startup with some applications. The Cross-DSO CFI runtime intercepts calls to dlopen and dlclose. In certain cases, libc itself may call dlopen and/or dlclose. Some applications, even some in base (like `id(1)`) call libc functions that call into dlopen/dlclose. This presents problems with llvm's Cross-DSO CFI runtime.

libc is an incredibly attractive target given its large surface area. It's incredibly complex. At this time, I feel applying Cross-DSO CFI to libc itself may be too large of an undertaking, preventing tangible progress. Thus, my initial goal will be to apply CFI to as many shared libraries in base as I can, but likely not libc at this time. As Rome was not built in a single day, neither
will a Cross-DSO CFI HardenedBSD be. It is my hope that we will indeed apply CFI in the future to libc (in whole or in part), but that day is not today.

Building ports/packages will be another huge aspect of this. Back in 2018, the last time I made tangible progress on Cross-DSO CFI, the memory footprint ballooned when building packages due to CFI'd libraries in base. Eventually, the experimental package build failed due to memory pressure.

My main objective: end 2023 (yes: 2023) with Cross-DSO CFI enabled in HardenedBSD by default. Whether libc is a part of that is unknown, but we can hope.

To fit that main objective, I plan to take a back seat to most other development aspects of the project, with the exception of hbsdfw. I will definitely be involved in all other aspects of the project (the infrastructure, the Foundation, etc.) The only thing that is changing: I am formally delegating the implementation of new security and hardening techniques to the wider HardenedBSD community so that I can focus on Cross-DSO CFI.

I appreciate all the help the community has given the project to date. I'm especially grateful for the continued contributions, the advocacy, the support. This little project would not exist in its current state without the recurring love and support you, the community, provide. As I focus my attention on a more difficult and involved goal (that of Cross-DSO CFI), I'm hopeful for a renewed sense of excitement and support from the community.