Lonnie VanZandt, Sr. Systems Engineer of TimeSys Corporation, takes issue with a comment about “disadvantages of real-time Linux” that appeared in a report from the second annual Real-Time Linux Workshop, held recently in Orlando, Florida. VanZandt writes . . .

In an otherwise informative report . . . Kenneth J. Hendrickson remarks . . .

Disadvantages of Real-Time Linux: The real-time portions of a system run in kernel space in real-time Linux. Therefore, no memory protection can be offered. A bug in the real-time code can take down the entire system. (Note: this disadvantage also exists for all other real-time OSes as well, so this is not really a disadvantage to real-time Linux!) The real-time Linux community is hard at work to eliminate this disadvantage. A goal for the near-term future is to be able to run real-time tasks in userland with all of the memory protections provided by unix, on all real-time Linux variants. Emanuele Bianchi has already done this on RTAI! Real-time Linux (as far as I know) is the first real-time OS to offer this capability.

TimeSys Corporation is compelled to address the errors in the above remark. The statement that RTLinux real-time tasks run as kernel-mode code and that this circumvents intertask memory protection is indeed fact.

However, the claim that this correctly [characterizes] serious [drawbacks] in all real-time operating systems is false. Several non-Linux [RTOSes] support user-mode real-time tasks that use per-task memory management protection; two well-known operating systems are LynxOS and VxWorks AE — and the respective vendors of many more can substantiate claims for their products.

Within the Linux space, which is of more interest to Kenneth's audience, the TimeSys Linux/RT operating system was specifically designed to instantiate real-time tasks as user-level Linux processes benefiting from full memory management protection. TimeSys Linux/RT has been available since May, 2000.

Attempts by RTLinux and RTAI developers to address the difficulties in performing intertask communication and in providing minimal per-task protections should not be viewed as innovations but as evidence of fundamental flaws in the original architecture of a multi-kernel operating system.

Lonnie VanZandt
Sr. Systems Engineer
TimeSys Corporation

 
This article was originally published on LinuxDevices.com and has been donated to the open source community by QuinStreet Inc. Please visit LinuxToday.com for up-to-date news and articles about Linux and open source.