Re: [arm-gnu] Compiling Safety Related Systems?
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [arm-gnu] Compiling Safety Related Systems?
- To: arm-gnu@xxxxxxxxxxxxxxxx
- Subject: Re: [arm-gnu] Compiling Safety Related Systems?
- From: Charles Manning <manningc2@xxxxxxxxxxxxx>
- Date: Wed, 10 Feb 2010 12:45:07 +1300
On Tuesday 09 February 2010 18:10:12 42Bastian wrote:
> Hi Brian
>
> > I am working on a product that needs to be IEC-61508 SIL 2 compliant.
> > The documentation states that the compiler has to be either SIL2
> > certified or have been in use on many (3 or more) products without having
> > negative (or known) issues.
> >
> > So my question to everyone is: Has anyone produced a safety related
> > product using the Sourcery G++ Lite 2009q1-161compiler and a STM32F103RB
> > processor?
>
> We at SCIOPTA were in discussion with TUV about this while certifying our
> RTOS and it was clearly stated that GCC is not valid.
> Reasons (also against RealView btw.) were:
>
> - missing proven-in-use numbers
> There are a lot of users for "GCC" but not "known" number.
This seems a strange way to look at things.
Sure, nobody knows how many people are using gcc but if you establish a lower
acceptable bound then it should be easy to find that number of users.
For example, all of Android phones, iphone, Mac OSX, Linux are build with gcc
(though not the same versions) and we're only scratching the surface.
See http://www.linuxfordevices.com/ for a heap more.
I'd hazard a guess that gcc gets used in more products (and is thus tested
against more code bases) than any other compiler suite.
>
> - error/defect information (for a certain build)
Perhaps the only real concern.
>
> - some "vague" doubts
>
> So we went for IAR (or DIAB in case of PowerPC)
>
> But the commercial version of CS might work out.
That's hopefully the CS value-added.
-- Charles