27 May 2020

feedKernel Planet

Kees Cook: security things in Linux v5.5

Previously: v5.4.

I got a bit behind on this blog post series! Let's get caught up. Here are a bunch of security things I found interesting in the Linux kernel v5.5 release:

restrict perf_event_open() from LSM
Given the recurring flaws in the perf subsystem, there has been a strong desire to be able to entirely disable the interface. While the kernel.perf_event_paranoid sysctl knob has existed for a while, attempts to extend its control to "block all perf_event_open() calls" have failed in the past. Distribution kernels have carried the rejected sysctl patch for many years, but now Joel Fernandes has implemented a solution that was deemed acceptable: instead of extending the sysctl, add LSM hooks so that LSMs (e.g. SELinux, Apparmor, etc) can make these choices as part of their overall system policy.

generic fast full refcount_t
Will Deacon took the recent refcount_t hardening work for both x86 and arm64 and distilled the implementations into a single architecture-agnostic C version. The result was almost as fast as the x86 assembly version, but it covered more cases (e.g. increment-from-zero), and is now available by default for all architectures. (There is no longer any Kconfig associated with refcount_t; the use of the primitive provides full coverage.)

linker script cleanup for exception tables
When Rick Edgecombe presented his work on building Execute-Only memory under a hypervisor, he noted a region of memory that the kernel was attempting to read directly (instead of execute). He rearranged things for his x86-only patch series to work around the issue. Since I'd just been working in this area, I realized the root cause of this problem was the location of the exception table (which is strictly a lookup table and is never executed) and built a fix for the issue and applied it to all architectures, since it turns out the exception tables for almost all architectures are just a data table. Hopefully this will help clear the path for more Execute-Only memory work on all architectures. In the process of this, I also updated the section fill bytes on x86 to be a trap (0xCC, int3), instead of a NOP instruction so functions would need to be targeted more precisely by attacks.

KASLR for 32-bit PowerPC
Joining many other architectures, Jason Yan added kernel text base-address offset randomization (KASLR) to 32-bit PowerPC.

seccomp for RISC-V
After a bit of long road, David Abdurachmanov has added seccomp support to the RISC-V architecture. The series uncovered some more corner cases in the seccomp self tests code, which is always nice since then we get to make it more robust for the future!

seccomp USER_NOTIF continuation
When the seccomp SECCOMP_RET_USER_NOTIF interface was added, it seemed like it would only be used in very limited conditions, so the idea of needing to handle "normal" requests didn't seem very onerous. However, since then, it has become clear that the overhead of a monitor process needing to perform lots of "normal" open() calls on behalf of the monitored process started to look more and more slow and fragile. To deal with this, it became clear that there needed to be a way for the USER_NOTIF interface to indicate that seccomp should just continue as normal and allow the syscall without any special handling. Christian Brauner implemented SECCOMP_USER_NOTIF_FLAG_CONTINUE to get this done. It comes with a bit of a disclaimer due to the chance that monitors may use it in places where ToCToU is a risk, and for possible conflicts with SECCOMP_RET_TRACE. But overall, this is a net win for container monitoring tools.

Some EFI systems provide a Random Number Generator interface, which is useful for gaining some entropy in the kernel during very early boot. The arm64 boot stub has been using this for a while now, but Dominik Brodowski has now added support for x86 to do the same. This entropy is useful for kernel subsystems performing very earlier initialization whre random numbers are needed (like randomizing aspects of the SLUB memory allocator).

As has been enabled on many other architectures, Dmitry Korotin got MIPS building with CONFIG_FORTIFY_SOURCE, so compile-time (and some run-time) buffer overflows during calls to the memcpy() and strcpy() families of functions will be detected.

limit copy_{to,from}_user() size to INT_MAX
As done for VFS, vsnprintf(), and strscpy(), I went ahead and limited the size of copy_to_user() and copy_from_user() calls to INT_MAX in order to catch any weird overflows in size calculations.

That's it for v5.5! Let me know if there's anything else that I should call out here. Next up: Linux v5.6.

© 2020, Kees Cook. This work is licensed under a Creative Commons Attribution-ShareAlike 3.0 License.
Creative Commons License

27 May 2020 8:04pm GMT

Rusty Russell: 57 Varieties of Pyrite: Exchanges Are Now The Enemy of Bitcoin

TL;DR: exchanges are casinos and don't want to onboard anyone into bitcoin. Avoid.

There's a classic scam in the "crypto" space: advertize Bitcoin to get people in, then sell suckers something else entirely. Over the last few years, this bait-and-switch has become the core competency of "bitcoin" exchanges.

I recently visited the homepage of Australian exchange btcmarkets.net: what a mess. There was a list of dozens of identical-looking "cryptos", with bitcoin second after something called "XRP"; seems like it was sorted by volume?

Incentives have driven exchanges to become casinos, and they're doing exactly what you'd expect unregulated casinos to do. This is no place you ever want to send anyone.

Incentives For Exchanges

Exchanges make money on trading, not on buying and holding. Despite the fact that bitcoin is the only real attempt to create an open source money, scams with no future are given false equivalence, because more assets means more trading. Worse than that, they are paid directly to list new scams (the crappier, the more money they can charge!) and have recently taken the logical step of introducing and promoting their own crapcoins directly.

It's like a gold dealer who also sells 57 varieties of pyrite, which give more margin than selling actual gold.

For a long time, I thought exchanges were merely incompetent. Most can't even give out fresh addresses for deposits, batch their outgoing transactions, pay competent fee rates, perform RBF or use segwit.

But I misunderstood: they don't want to sell bitcoin. They use bitcoin to get you in the door,but they want you to gamble. This matters: you'll find subtle and not-so-subtle blockers to simply buying bitcoin on an exchange. If you send a friend off to buy their first bitcoin, they're likely to come back with something else. That's no accident.

Looking Deeper, It Gets Worse.

Regrettably, looking harder at specific exchanges makes the picture even bleaker.

Consider Binance: this mainland China backed exchange pretending to be a Hong Kong exchange appeared out of nowhere with fake volume and demonstrated the gullibility of the entire industry by being treated as if it were a respected member. They lost at least 40,000 bitcoin in a known hack, and they also lost all the personal information people sent them to KYC. They aggressively market their own coin. But basically, they're just MtGox without Mark Karpales' PHP skills or moral scruples and much better marketing.

Coinbase is more interesting: an MBA-run "bitcoin" company which really dislikes bitcoin. They got where they are by spending big on regulations compliance in the US so they could operate in (almost?) every US state. (They don't do much to dispel the wide belief that this regulation protects their users, when in practice it seems only USD deposits have any guarantee). Their natural interest is in increasing regulation to maintain that moat, and their biggest problem is Bitcoin.

They have much more affinity for the centralized coins (Ethereum) where they can have influence and control. The anarchic nature of a genuine open source community (not to mention the developers' oft-stated aim to improve privacy over time) is not culturally compatible with a top-down company run by the Big Dog. It's a running joke that their CEO can't say the word "Bitcoin", but their recent "what will happen to cryptocurrencies in the 2020s" article is breathtaking in its boldness: innovation is mainly happening on altcoins, and they're going to overtake bitcoin any day now. Those scaling problems which the Bitcoin developers say they don't know how to solve? This non-technical CEO knows better.

So, don't send anyone to an exchange, especially not a "market leading" one. Find some service that actually wants to sell them bitcoin, like CashApp or Swan Bitcoin.

27 May 2020 12:49am GMT

20 May 2020

feedKernel Planet

Dave Airlie (blogspot): DirectX on Linux - what it is/isn't

This morning I saw two things that were Microsoft and Linux graphics related.


a) DirectX on Linux for compute workloads
b) Linux GUI apps on Windows

At first I thought these were related, but it appears at least presently these are quite orthogonal projects.

First up clarify for the people who jump to insane conclusions:

The DX on Linux is a WSL2 only thing. Microsoft are not any way bringing DX12 to Linux outside of the Windows environment. They are also in no way open sourcing any of the DX12 driver code. They are recompiling the DX12 userspace drivers (from GPU vendors) into Linux shared libraries, and running them on a kernel driver shim that transfers the kernel interface up to the closed source Windows kernel driver. This is in no way useful for having DX12 on Linux baremetal or anywhere other than in a WSL2 environment. It is not useful for Linux gaming.

Microsoft have submitted to the upstream kernel the shim driver to support this. This driver exposes their D3DKMT kernel interface from Windows over virtual channels into a Linux driver that provides an ioctl interface. The kernel drivers are still all running on the Windows side.

Now I read the Linux GUI apps bit and assumed that these things were the same, well it turns out the DX12 stuff doesn't address presentation at all. It's currently only for compute/ML workloads using CUDA/DirectML. There isn't a way to put the results of DX12 rendering from the Linux guest applications onto the screen at all. The other project is a wayland/RDP integration server, that connects Linux apps via wayland to RDP client on Windows display, integrating that with DX12 will be a tricky project, and then integrating that upstream with the Linux stack another step completely.

Now I'm sure this will be resolved, but it has certain implications on how the driver architecture works and how much of the rest of the Linux graphics ecosystem you have to interact with, and that means that the current driver might not be a great fit in the long run and upstreaming it prematurely might be a bad idea.

From my point of view the kernel shim driver doesn't really bring anything to Linux, it's just a tunnel for some binary data between a host windows kernel binary and a guest linux userspace binary. It doesn't enhance the Linux graphics ecosystem in any useful direction, and as such I'm questioning why we'd want this upstream at all.

20 May 2020 12:01am GMT