This is the mail archive of the elix@sources.redhat.com mailing list for the Elix project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]

Re: 回覆 : subset of EL/IX for DSPs


Jwusheng Hu <jshu@cn.nctu.edu.tw> writes:

> Julian Rose <jhrose@dial.pipex.com> writes:
> 
> > Hello,
> >
> > Could anyone point me to info regarding a subset of EL/IX
> > or POSIX that targets DSPs? Or if anyone else is thinking
> > about this then their suggestions would be welcome. The
> > context of the question is the kernel I'm developing, please
> > see "http://www.jhrose.dial.pipex.com/dsp_K.htm". Thanks.
> >
> 
> > Nobody has put any real thought into the requirements of DSPs yet.
> > Embedded and real-time systems have proven to be difficult enough to
> > pin down. I suspect that DSPs would require a subset of the RT subset
> > of EL/IX, perhaps with special schedulers and variations of some
> > communications primitives. Whether we need to provide a specific DSP
> > subset, or whether this would just be a version of the real-time
> > subset I don't really know yet.
> 
> > Suggestions and comments about what would be good for DSPs are very
> > welcome. This would at least help to ensure that the next version of
> > EL/IX is not actively DSP-hostile.
> >
> > --
> > Nick Garnett, eCos Kernel Architect
> > Red Hat, Cambridge, UK
> 
> It appears that there are two types of requirements here and they are 
> related to a fundamental design question: will you design a system 
> completely based on DSP? Traditionally, DSP usually acts as a slave 
> processor preforming repeated calculations for time-critical algorithms. 
> Because of this, DSP chips have very little support of system interface like 
> ethernet, LCD etc.., compared with most microcontrollers.
> 
> However, when the MIPS of DSP exceeds the needs of algorithm complexity, it 
> is quite natural to use the extra MIPS for system interface and a RTOS would 
> be very helpful. However, most DSP algorithms must be coded in assembly to 
> fully utilize the computational capability, e.g., zero-overhead looping. 
> This creates a potential threat to RTOS that full control of the CPU 
> resource maynot be possible and users can easily crash the kernel by 
> mistake. Hence, unless DSP programmers understand the kernel well, she/he 
> may have difficulty using other people's RTOS.
> 
> On the other hand, the master-slave structure is still popular when DSP 
> takes on the computing role for software radio. For example, TI has a 
> DSP+ARM solution that is targeted for portable wireless devices. In this 
> case, the need would be how to have a single RTOS that handles both 
> processors together and provides a seamless communication interface between 
> them.

I agree with all of this. The main contribution that EL/IX can make to
these issues is to provide a uniform programming environment for both
the DSP processor and the control processor. This means that the
programmers do not have to master two sets of APIs and have the option
of moving code back and forth across the divide.

The issue of making a single RTOS that provides a seamless environment
across several processors is probably not something the EL/IX should
address beyond ensuring that the APIs allow it (but with my eCos hat
on, it is somthing that is of interest).

-- 
Nick Garnett, eCos Kernel Architect
Red Hat, Cambridge, UK


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]