Re: Microkernels & Capabilities (was Re: Byte ordering)

On Wed, 16 Oct 2024 11:34:31 -0400, Paul A. Clayton wrote:

While the argument that only microkernels can provide modularity with
respect to software development seems highly flawed, modularity with
respect to privilege seems more challenging (impossible?) for a
monolithic kernel and modularity with respect to fault isolation seems
to require substantially more discipline/ constraint than typical for a
monolithic design.

The CHERI project has been trying to revive the capability concept. They
chose BSD over Linux (neither being microkernel-based) for their research
purely on the basis that BSD was a slower-moving target. (Maybe they
didn’t know about LTS Linux kernels.)

Seems like the idea of using a microkernel never occurred to them. Not
seriously, anyway.

Software like FUSE (Filesystem in Userspace) hints that some microkernel
aspects are desirable even in a monolithic kernel system.

It is useful for less performance-intensive code. For example, NTFS
support has primarily been primarily provided over the past couple of
decades via FUSE; the equivalent kernel-based filesystem module has been
lagging somewhat in features, even though that is still seen as the better
approach for high performance.

Obviously this is not true of Linux filesystems in general, it just
happens to be the case with NTFS.

I know relatively little about OSes, but the arguments I have read on
both sides seem to have been very biased.

Can’t argue with reality, though.