Using Linux as a host for lightweight software cores specialized in delivering very short and bounded response times has been a popular way of supporting real-time applications in the embedded space over the years.
This dual kernel design introduces a small real-time infrastructure into the Linux kernel, which immediately handles time-critical, out-of-band activities independently from the ongoing general purpose kernel work. Application threads co-managed by this infrastructure still benefit from the common kernel services such as virtual memory management; they can leverage the rich GPOS feature set Linux provides such as networking, data storage or GUIs too.
There are significant upsides to keeping the real-time core separate from the GPOS infrastructure:
because the autonomous core does not depend on the main kernel logic for running time-critical operations, real-time activities are not serialized with GPOS operations internally. This removes potential delays induced by the latter, by construction. As a result, there is no need for keeping all GPOS operations fine-grained and highly preemptible at any time, which may otherwise induce noticeable overhead as task priority inheritance and IRQ threading have to apply system-wide, affecting all tasks including those which have zero real-time requirement. To sum up, the lesser the effort for the kernel to maintain real-time guarantees internally, the more CPU bandwidth is available to applications.
when debugging a real-time issue, the functional isolation of the real-time infrastructure from the rest of the kernel code restricts bug hunting to the scope of the small autonomous core, excluding most interactions with the very large GPOS kernel base.
with a dedicated infrastructure providing a specific, well-defined
set of real-time services, applications can unambiguously figure out
which API calls are available for supporting time-critical work,
excluding all the rest as being potentially non-deterministic with
respect to response time.
Said differently, would you assume that each and every routine from the glibc becomes real-time capable solely by virtue of running on a native preemption system? Of course you would not, therefore you would carefully select the set of services your real-time application may call from its time-critical work loop in any case. For this reason, providing a compact, dedicated API which exports a set of services specifically aimed at real-time usage is clearly an asset, not a limitation.
This documentation presents Dovetail, a kernel interface which introduces a high-priority execution stage into the main kernel logic. At any time, out-of-band activities running on this stage can preempt the common work. A task-specific software core - such as a real-time core - can connect to this interface for gaining bounded response time to external interrupts and ultra-low latency scheduling capabilities. This translates into the Dovetail implementation as follows:
the interrupt pipeline which creates a high-priority execution stage for an autonomous software core to run on.
support for the so-called alternate scheduling between the main kernel and the autonomous software core for sharing kthreads and user tasks.
Although both layers are likely to be needed for implementing some autonomous core, only the interrupt pipeline has to be enabled in the early stage of porting Dovetail. Support for alternate scheduling builds upon the latter, and may - and should - be postponed until the pipeline is fully functional on the target architecture or platform. The code base is specifically maintained in a way which allows such incremental process.
Dovetail only introduces the basic mechanisms for hosting an autonomous core into the kernel, enabling the common programming model for its applications in user-space. It does not implement the software core per se, which should be provided by a separate kernel component instead, such as the EVL core.
Linux-based dual kernel systems require some interface layer to couple the secondary core (typically a real-time capable one like the EVL core) to the logic of the kernel it is embedded in, so as to benefit from the rich Linux feature set while running dedicated applications with stringent real-time requirements. The archetypical implementation of such kind of interface is the I-pipe, which served both RTAI and Xenomai over the years. For several reasons explained in this document, maintaining the I-pipe proved to be difficult as changes to the mainline kernel regularly caused non-trivial code conflicts, sometimes nasty regressions to the I-pipe maintainers downstream. Although the concept of interrupt pipelining proved to be correct in delivering short response time with reasonably limited changes to the original kernel, the way this mechanism is integrated into the mainline code shows its age.
Dovetail is the successor to the I-pipe, with the following goals:
introduce a high-priority execution stage for time-critical operations into the kernel logic, enabling all device interrupts to behave like NMIs from the standpoint of the kernel.
provide a straightforward interface to autonomous cores for running Linux tasks on this high-priority execution stage when they have to, enabling ultra-fine grained preemption of all other kernel activities in such an event.
enable the common Linux programming model for applications which should be controlled by the autonomous core for delivering ultra-low latency services (private user address space, multi-threading, SMP capabilities, system calls etc).
make it possible to maintain Dovetail (and ultimately the EVL core which is based on it) with common kernel development knowledge, at a fraction of the engineering and maintenance cost native preemption requires. Typically, Dovetail always favors extending existing kernel subsystems with the ability to deal with the new execution stage, instead of taking sideways steps. For instance, the interrupt pipeline logic is directly integrated into the generic IRQ layer.
Working on Dovetail is customary Linux kernel development, following the common set of rules and guidelines which prevails with the mainline kernel.
The development tip of the Dovetail interface is maintained in the dovetail/master branch of the following GIT repository which tracks the mainline kernel:
This document is intended to people having common kernel development knowledge, who may be interested in building an autonomous software core on Dovetail, porting it to their architecture or platform of choice. Knowing about the basics of interrupt flow, IRQ chip and clock event device drivers in the kernel tree would be a requirement for porting this code.
However, this document is not suited for getting one’s feet wet with kernel development, as it assumes that the reader is already familiar with kernel fundamentals.
Last modified: Fri, 13 Mar 2020 12:44:32 CET