VxWorks

What is VxWorks? VxWorks, from Wind River Systems, is a networked real-time operating system designed to be used in a distributed environment. It runs on a wide variety of hardware. It requires a host workstation for program development; supported host platforms include Sun3, Sun4…. Is the run-time component of the Tornado 3 embedded development platform.

What is VxWorks? The VxWorks RTOS comprises the core capabilities of the wind microkernel (not monolithic) along with advanced networking support, powerful file system and I/O management, and C++ and other standard run-time support.

What is VxWorks? The development environment is based on cross-development or remote-development method. You will need a UNIX machine of some sort (e.g. SUN's) to run the compilers and debuggers. The compiled application code can be downloaded to the target and runs as part of the VxWorks image. During the development phase or thereafter, individual object code (.o files) can be downloaded dynamically to running target system. Finished applications can be ROM'ed or whatever.

What is VxWorks? VxWorks most definitely is not a "realtime UNIX", or a varient of UNIX as often misunderstood by some people. The confusion perhaps is due to the fact that UNIX hosts are used most widely to develop applications for VxWorks (and VxWorks itself).

Brief History of VxWorks WRS was started by Jerry Fiddler and David Wilner in Jerry's garage as a contract/consultant shop for realtime, embedded systems and other fun things. They wrote a bunch of neat programs for their work and eventually adding some things unheard of in embedded/realtime market in those days such TCP/IP networking. And continued to pioneer in this area by adding NFS, etc.

Brief History of VxWorks VxWorks was the name given the collection of software which ran on top of various realtime kernels including VRTX and pSOS as well an earlier slower version of WIND kernel. VxWorks now no longer runs on other kernels; it now runs exclusively on its own WIND kernel since the 5.0 release, for which the WIND was rewritten by John Fogelin.

Wind microkernel Efficient task management multitasking, unlimited number of tasks Preemptive and round-robin scheduling Fast, deterministic context switching 256 priority levels

Wind microkernel Fast, flexible intertask communications Binary, counting, and mutual exclusion semaphores with priority inheritance Message queues – local and locale-transparent distributed messaging POSIX pipes, counting semaphores, message queues, signals and scheduling Control sockets Shared memory Standard COM VxFusion

Wind semaphores binary mutual exclusion Counting The fastest, most general purpose semaphore. Optimized for synchronization and can also be used for mutual exclusion. mutual exclusion A special binary semaphore optimized for problems inherent in mutual exclusion: priority inheritance, deletion safety and recursion. Counting Like the binary semaphore, but keeps track of the number of times the semaphore is given. Optimized for guarding multiple instances of a resource.

Wind microkernel High scalability Incremental linking and loading of components Fast, efficient interrupt and exception handling Optimized floating-point support Dynamic memory management System clock and timing facilities

Networking support BSD 4.4 TCP/IP networking IP, IGMP, CIDR, TCP, UDP, ARP PPP, SLIP/CSLIP Standard Berkeley sockets Berkeley packet filter RIPv1/v2 and routing sockets TFTP, rlogin, telnet, rsh NFS dient and server, ONC RPC SNTP SNMPv1/v2c/v3 Network Protocol Toolkit BGP-4 Ipsec/IKE NAT L2TP MPLS ATM Frame relay ISDN SS7

Fast, flexible I/O and local file system POSIX asynchronous I/O and directory handling SCSI support MS-DOS-compatible file system TrueFFS flash file system (optional) ISO 9660 CD-ROM file system PCMCIA support

Other Features Protection domain management Resource tracking and management System overrun protection Powerful loader facilities

Protection domain management

Overlapped memory model

Multasking ready – pending – delayed – suspended – the task is waiting only for CPU time pending – the task is blocked because some resource is unavailable delayed – the task is inactive for a pre-determined time interval suspended – the task is prevented from becoming active

Multasking The highest-priority ready task is executing. delayed pending ready suspended taskInit()

Priority Preemption

Round-Robin Scheduling

differences between traditional UNIX and VxWorks VxWorks runs in one mode. No protected vs. user mode switching is done. Running in supervisor mode on most processors, and not using traps for system calls, VxWorks can achieve minimal overhead on a given piece of hardware than UNIX. Programming on VxWorks can be more tricky than UNIX for the same reason. UNIX provides resource reclamation; by default, VxWorks does not. VxWorks does not have full "process"; it only has tasks, or "threads".

differences between traditional UNIX and VxWorks VxWorks: minimal interrupt latency (e.g. spl's are quasi-implemented as semaphores). Traditional UNIX: high interrupt latency (e.g. spl's are implemented as interrupt lock and unlock calls). VxWorks: priority interrupt-driven preemption, optional round-robin time-slicing. Traditional UNIX: prioritized round-robin preemptive time-slicing. Since VxWorks is just a glorified "program" it can be changed and customized pretty easily. Task scheduling can be customized as desired. VxWorks networking code, however, is very UNIX compatible. It is essentially "ported" version of BSD UNIX TCP/IP code -- tahoe release .

Example: Porting VxWorks to Gmicro/200 Portable kernel modifications Establishing the VxWorks shell environment Adding Gmicro-specific debug functions Adding board-specific functions Adding networking support Building BootROM Optimizing and tuning testing

Impl. Architecture

development Tornado in Mobile Internet Wind River rolls out safety-critical software