[jkhdigital]

I’m not going to argue against much of the content of this paper, but it should be pointed out that their argument in the middle section against the “confinement myth” seems pretty bogus. They say that you can isolate the capability read/write resource from the data read/write resource, but… this makes absolutely no sense. Bits are bits. If you assume some out-of-band isolation of capability distribution then you’ve changed the game, but even that isn’t enough for me to believe that isolation is possible.

[Animats]

Early thinking was in terms of capability handles. As with file descriptors, the handle is only meaningful when passed across a protection boundary to something which can check if the handle is valid.

Later, there were encrypted capabilities, which are signed data, like TLS certs. These get kind of bulky. And hardware support, in a few machines.

[jdougan]

alternate futures where the 33bit versions of the i960 became the processor family of choice.

[ryukafalz]

Consider two processes on a *nix system, and for the sake of argument let's say they're sufficiently isolated from each other as to have only one communications channel between them. If that communications channel is a unix domain socket, one process can send a file descriptor (effectively a capability) to the other over the socket. Each process has a file descriptor table in the kernel whose integer keys are only meaningful to that process in particular, and the kernel provides a mechanism to transmit file descriptors across a socket. The kernel mediates in this case.

If the communications channel is not a unix domain socket and is instead something like a TCP connection, you don't have this option available to you.

You aren't always just sending bits from one process to another!

[jkhdigital]

No, you’re using the same sleight of hand as the paper.

Boebert’s objection is about whether Alice can transmit unauthorized authority to Bob across a security boundary that’s supposed to prevent that flow. Your SCM_RIGHTS example is a case where the kernel is deliberately providing a sanctioned channel for authority transfer, with the kernel’s blessing, between two processes that the kernel does not consider to be on opposite sides of a mandatory access control boundary. Unix has no (*)-property. There is no “high” and “low” in the Bell-LaPadula sense on a standard Unix system. So of course the kernel mediates the transfer cleanly; it’s not enforcing any policy that would be violated by the transfer.

The moment you try to extend this to the actual case under dispute—Alice is “high,” Bob is “low,” and the security policy says high-to-low information flow is forbidden—then if the kernel refuses to deliver the fd across the boundary, the security property was enforced by the separate MAC layer, not the capability mechanism.

The conflation which is endemic in this whole debate is between “capabilities as a kernel-mediated authority mechanism” and “capabilities as a property that holds across all observable behavior of the system.” Unix file descriptors are the former. Boebert’s objection is about the latter.

[Veserv]

Your communication channel between Alice and Bob is, itself, a capability (or a collection of capabilitys) that grants Bob memory write, Alice memory read, but does not grant the ability to transmit a capability from Bob to Alice.

Absent a misunderstanding on your part, the only way I can coherently interpret your argument is that you are arguing that the presence of kernel data structures mediating the handles somehow makes it not a capability system. That there is some background element mediating the validity of your capability representation and thus that is just a MAC layer; unless you can write the byte representation of your handle into memory and somebody else can read it out and then have access to that resource it is not a capability.

One, that allows forging capabilitys unless they are cryptographically secure against collisions.

Two, the actual essence of capabilitys is not being bearer tokens, it is non-construction. Capabilitys are derived from existing capabilitys, not manifested into existence. They have provenance. It is the OS equivalent of not allowing programs to cast arbitrary integers to pointers and thus manifesting pointers into existence which breaks basically every high level memory safety guarantee. You do not allow programs to cast arbitrary data into handles to resources which is what ambient authority systems effectively require.

[jkhdigital]

I'm going to first apologize for engaging in rhetorical sleight of hand myself, since I indulged in a bit of the hand-wavy argumentation that happens so often in these nit-picky debates. I'm going to respond cleanly here mostly to sharpen my own argumentative saw.

The original PSOS paper makes a few claims that are in tension with one another, and then buries the lede about how that tension can be addressed. Here's a few passages, directly quoted from the paper:

> [...] there are several important pragmatic reasons why PSOS capabilities are useful as a naming and protection mechanism for supporting abstract objects.

> 1. The capability mechanism has a very simple implementation. This allows capabilities to be built into the system at the lowest level of abstraction, thus making capabilities available for the most primitive objects.

> 2. Capabilities are uniform in size, making them easy to manage.

> 3. The inclusion of access rights in capabilities permits efficient fine-grained control of access to objects.

> 4. Capabilities can be written into storage (including secondary storage) and retrieved from storage in the same manner as other data, and therefore have many of the properties of other data.

Item 4 above is the one that should draw the most attention. I don't think anyone would contest the claim that PSOS has wonderful ergonomics for managing access to resources, but the moment you want to impose a system-wide access control policy then you must add another security mechanism, completely outside the capability abstraction, that adds some friction. This is fully acknowledged by the PSOS authors, although frankly they buried the lede since this is the only thing that the secure systems folks really cared about at the time. From the section on Store permissions:

> Because simplicity of the basic capability mechanism is extremely important to achieve the goals of PSOS, any means for restricting the propagation of capabilities should not add complexity to the capability mechanism. [...] A few access rights (only one is currently used by PSOS itself) are reserved as store permissions. This is the only burden placed on the capability mechanism.

> By properly choosing the segments that are capability store limited, some very useful restrictions on the propagation of capabilities can be achieved. The restriction used in PSOS is not allowing a process to pass certain capabilities to other processes or to place these capabilities in storage locations (e.g., a directory or interprocess communication channel) accessible to other processes. [...] The store permission mechanism has been selected as primitive in the system because it achieves the desired result with negligible additional complexity or cost.

This appears as claim 8 in the summary section of the paper near the end:

> Propagation of capabilities can be restricted by use of capability store permissions. The passage of a capability to other users can be prevented by not including process store permission in that capability's access rights.

Ok, so that's the PSOS paper and it's claims. Boebert's paper--really a note, since it is a mere 3 pages--states its argument in fairly direct terms:

> The attack is made possible by an inherent attribute of pure capability machines: the right to exercise access carries with it the right to propagate that access. Thus even if an omniscient oracle correctly creates capabilities, it cannot control their further propagation. If extra mechanisms are imposed to impose this control, the machine is no longer an unmodified capability machine.

The only issue here is, perhaps, semantic: Boebert (correctly) states that an unmodified capability machine cannot provide what is considered a very basic mandatory security policy, but the PSOS folks already acknowledged this by stating that the system needs a capability store permission manager for mandatory security policy enforcement. The phrasing that they used--"the store permission mechanism has been selected as primitive in the system"--is the bait-and-switch where they treat it like part of the capability model rather than making it clear that it is an entirely distinct mechanism that must be composed with pure capabilities to achieve the (genuinely difficult) security properties that systems designers were seeking.

I suspect the horse is already dead it's worth double-tapping to make sure, so let's continue. The Myths paper muddies the waters further by making this claim after supposedly debunking Boebert:

> Boebert’s result is valid in any capability system that cannot distinguish between data transfer and capability transfer. But partitioned and type-enforced capability systems do not have this problem, and password capability systems have been engineered to avoid this problem [1, 11].

> Moreover, it has been formally verified that any capability system enforcing independent controls on data transfer and capability transfer can enforce both confinement and the *-Property [22].

We'll focus on reference [22] since that is the stronger claim here. That paper is Shapiro & Weber (2000) "Verifying the EROS Confinement Mechanism": https://flint.cs.yale.edu/cs428/doc/eros-verify.pdf

This is the motivation for their paper, which is stated unambiguously:

> Boebert [1] and Karger [9] have argued that unmodified capability systems cannot enforce even basic mandatory access controls such as the *-property. Both have proposed solutions in the form of hybrid protection architectures. Karger has also argued that unmodified capability systems cannot enforce confinement [8]. Given that EROS is a pure capability system, and that its security design rests on its ability to enforce confinement, a rigorous verification of the EROS confinement mechanism is necessary.

For some reason, they decide to respond to these claims in the Related work section, just before their conclusion, although they do address them head-on:

> Boebert [1] and Karger [9] show that pure capability systems cannot enforce the *-property. While their conclusion is correct, capability systems do provide sufficient strength to construct mandatory policies at a higher level of abstraction with reasonable performance, as has been done in KeySafe [14].

> Karger has also shown that unmodified capability systems cannot enforce the confinement policy [8]. The apparent discrepancy results from differences in term definition. Karger’s confinement policy is a mandatory access control policy: "this piece of information must not be disclosed to that set of unauthorized parties." That is, it is a policy concerning the flow of information to subjects. Lampson’s confinement problem [10] imposes a weaker constraint: information can flow out of the subsystem only through authorized channels. That is, in the Lampson definition the channels define an encapsulation boundary to be enforced.

> We believe that the KeySafe architecture can enforce both the *-property and Karger’s confinement policy, but this does not directly contradict their claims. KeySafe is a reference monitor built on top of a more primitive capability mechanism; such a reference monitor constitutes a modified capability system in the sense of Karger’s discussion.

It's worth questioning whether the Myths authors were justified in citing this paper the way they did. But either way, I think it's pretty clear that once you pin down a precise definition of the terms used in the discussion, there is little disagreement among any of these authors. However, in casual arguments this precision is lost and you end up with a situation where two things are true at the same time but people choose to talk about only one at a time and think they're winning arguments:

1. An unmodified capability machine cannot enforce the *-property or mandatory access control confinement policies.

2. Modifying a capability machine to enforce such policies (and provide proof of enforcement) is straightforward because there is a single clearly-defined interface through which the systems may be composed.

My stance is that the PSOS folks screwed up their marketing. They really do have a superior product, so to speak, but they tried to downplay the fact that it did not provide a solution to the genuinely difficult problem of enforcing MAC policies (which was really about reference monitor discipline, not capabilities or ACLs). The right pitch for ocap design is "we offer a cleaner, more compositional, more auditable substrate for authority management--which is itself a substantial contribution and worth caring about--and on top of that substrate you can build the same MAC policies you'd build on any other substrate, but with better starting axioms and clearer proof structures." That's a contribution that doesn't need to be defended against Boebert because it doesn't claim (or appear to claim) what Boebert showed couldn't be claimed.

[adrian_b]

That argument assumes that the delegation of a capability to another process must happen through a path of interprocess communication that can be established only by the operating system, if the processes that want to communicate have the capabilites for this.

I have not studied to see how the existing capability-based operating systems solve this problem, because it seems that this is not a simple solution. If the capabilities are very fine-grained, to make certain that IPC really cannot happen, that might be cumbersome to use, while coarse-grained capabilities could be circumvented. To really prevent IPC without appropriate capabilities, a lot of the convenient features of a UNIX-like system must be forbidden, like the existence of files that can be read by any user, or directories like /tmp , where anyone can write a file.

[josephg]

> If the capabilities are very fine-grained, to make certain that IPC really cannot happen, that might be cumbersome to use, while coarse-grained capabilities could be circumvented.

In SeL4 it’s kinda like this: A capability is an opaque handle you can invoke to RPC into some other process or into the kernel. There’s no worry about how fine grained capabilities are because there’s no global table of permission bits or anything like that. Processes can invent capabilities whenever they want. Because caps just let other processes call your code, you can programmatically make them do anything.

Suppose I want to give a process read only access to a file I have RW access to. The OS doesn’t need a special “read only capability” type. It doesn’t need to. Instead, my process just makes capabilites for whatever I actually want on the fly. In this case, I just make a new capability. When it’s invoked I see the associated request, if the caller is making a read request, I proxy that request to the file handle I have. (Also another cap). And if it’s a write request, I can reject it. Voila!

This is how you can write the filesystem and drivers in userland. One process can be in charge of the block devices. That process creates some caps for reading and writing raw bytes to disk. It passes the “client side” of that cap to a filesystem process, which can produce its own file handle caps for interacting with directories and files, which can be passed to userland processes in turn. Its capabilities all the way down.

[zzo38computer]

This kind of proxy capabilities has other benefits as well, e.g. you can implement a disk quota, or transparent compression, or logging, or ask the user (if you have a capability which can do that), or provide access to a part of the file as though it is the entire file, etc.

Or, if a program requests access to a camera, you can provide a capability with a still picture, a video file, a filter (e.g. that resizes the picture or modifies the colour) from some source (including, but not limited to, a camera), etc; this can be helpful in case e.g. you do not have a camera on your computer, or for testing.

(Other people have similar ideas, sometimes independently than I do.)

There is also a way to transmit capabilities across a network; I had thought of how a protocol would be made to do such a thing.

[adrian_b]

That works perfectly fine for an embedded computer, which is where systems like SeL4 are used.

On the other hand, I cannot see how this approach can be scaled to something like a personal computer.

For some programs that I run, e.g. for an Internet browser, I may want to not authorize them to access anything on SSDs/HDDs, except for a handful of configuration files and for any files they may create in some cache directories.

For other programs that I run, I may want to let them access most or all files in certain file systems. Any file system that I use contains typically many millions of files.

Therefore it is obvious that using one capability per file is not acceptable. Moreover, such programs may need to access immediately many thousands of files that have been just created by some other program that has run before them, for instance a linker running after a compiler.

Assuming that a pure capability-based system is used, not some hybrid between ACLs and capabilities, there must be some more general kinds of capabilities, that would grant access simultaneously to a great number of resources, based on some rules, e.g. all files in some directory can be read, all files with a certain extension or some other kind of name pattern from some directory may be written or deleted or renamed, and so on.

[josephg]

> On the other hand, I cannot see how this approach can be scaled to something like a personal computer.

Personally I think the biggest challenge is UX. The systems engineering is good, and it works just fine.

> For other programs that I run, I may want to let them access most or all files in certain file systems. Any file system that I use contains typically many millions of files. Therefore it is obvious that using one capability per file is not acceptable.

Yeah, of course! Just make a capability representing the containing directory or filesystem. Then the program is free to open and browse files within that directory, but nothing outside of it.

I agree with others in this thread. Think of the capability like a bearer token. You wouldn't make a token per file. Just make one for the directory.

[nullpoint420]

Then make a userspace server to do that. If you want to see how this works in practice, macOS and iOS are great “pragmatic” implementations of this pattern. They use a Mach/BSD hybrid

[convolvatron]

you're absolutely right. this is just a terminology confusion I think. we can talk about capabilities as 'a replacement for ACLs', in which case, yes we need to think about policy rules and not a gigantic list of possible atoms.

from a mechanism point of view a 'capability' is really more a bearer token, the result of a policy decision, a credential that we can give to the OS to show that we have been given access without going through the rules-based machine for every operation.

[newpavlov]

>Its capabilities all the way down.

IIUC one problem with such layering of capability processing is that each passed layer results in a context switch (i.e. switch of memory mappings, thrashing of caches, etc.) and its on top of the cost of passing through the kernel. In other words, you may need to pay cost of N syscalls for one multi-layered capability-based operation.

[josephg]

True, but capability calls in SeL4 are supposedly faster than linux syscalls. Because caps are such an important primitive, they're extremely heavily optimised.

As an example, when you invoke a capability, your process hands the callee your scheduler time-slice. So its not like linux where your process yields to the scheduler. The same CPU core will handle the entire call -> process -> return computation pipeline between multiple processes.

I'm not sure how fast it ends up in practice compared to a similar system built on top of linux. I suspect a lot of the difference would come down to implementation choices. And if its still not fast enough, you can always just set up a ring buffer or something between processes to share data directly.

[daymanstep]

Doesn't that mean that your process is then responsible for ensuring that an app with a read-only capability cannot do a write ?

You're moving the burden of enforcement from the kernel to the user level ?

[josephg]

Yes, microkernels like SeL4 do almost all real work out of the kernel, and in userland processes. It’s much more secure that way.

[jkhdigital]

Covert channels are a thing. Shared access to resources always opens the possibility of covert information passing through e.g. modulation of resource usage. This isn’t even out-of-band, it’s just a hard fact that a shared resource always creates a potential covert channel (source: Lampson 1972, A Note on the Confinement Problem).

[zzo38computer]

How I had idea of a computer and operating system design, measuring time also requires a capability. Creating files (including temporary files) also requires capabilities. Shared memory is read-only by everyone; to be able to write to memory you must have exclusive access. All of these capabilities are not necessarily what the program using them intended them to be; they may also be proxies, or capabilities of the wrong type (the kernel does not know anything about the types of capabilities except those it created itself), etc. A proxy may limit communication from one program to a service. Using these as well as other things (including, but not limited to, the CPU design), there are things that can be done to mitigate these problems (including things necessary to mitigate other kind of timing attacks based on other capabilities, e.g. slowing down network access for purpose of testing its working on slow networks).