Re: [ECOS] eCos on 80186

Re: [ECOS] eCos on 80186

[Bruno Cabral Goncalves]

Thanks Bart,
but I read im someplace that is possible do this changing the makefile to compile on other compiler like Borland C++. And I would like known if someone did this port.

Bruno


De: "Bart Veer" bartv@ecoscentric.com
Para: brunogon@terra.com.br
Cópia: ecos-discuss@sources.redhat.com
Data: Tue, 29 Mar 2005 10:20:03 +0100 (BST)
Assunto: Re: [ECOS] eCos on 80186

> >>>>> "Bruno" == Bruno Cabral Goncalves <brunogon@terra.com.br> writes:
>
>     Bruno> I read on archieves that is possible run eCos on 16 bits
>     Bruno> processor, but I would like known if someone have did this
>     Bruno> port, and if possible what I need do to do this port to a
>     Bruno> aMD 80186 processor
>
> The term "16 bit processor" is poorly defined.
>
> In the case of an 80186, an eCos port is unlikely. eCos depends on gcc
> and the compiler only supports 386 onwards. Both gcc and eCos assume a
> simple linear address space, not the segment/offset complications in
> early x86 processors.
>
> Bart
>
> --
> Bart Veer                       eCos Configuration Architect
> http://www.ecoscentric.com/     The eCos and RedBoot experts
>
>
>


--
Before posting, please read the FAQ: http://ecos.sourceware.org/fom/ecos
and search the list archive: http://ecos.sourceware.org/ml/ecos-discuss

Re: [ECOS] eCos on 80186

[Bart Veer]

>>>>> "Bruno" == Bruno Cabral Goncalves <brunogon@terra.com.br> writes:

    Bruno> Thanks Bart,
    Bruno> but I read im someplace that is possible do this changing
    Bruno> the makefile to compile on other compiler like Borland C++.
    Bruno> And I would like known if someone did this port.

I suspect you are referring to the ecos-discuss thread
http://ecos.sourceware.org/ml/ecos-discuss/2004-06/msg00220.html
The thread included a suggestion to use Borland, but I am not aware of
anybody writing any code along those lines. eCos depends heavily on a
number of gcc extensions which makes it hard to use any other
compiler.

Bart

-- 
Bart Veer                       eCos Configuration Architect
http://www.ecoscentric.com/     The eCos and RedBoot experts


-- 
Before posting, please read the FAQ: http://ecos.sourceware.org/fom/ecos
and search the list archive: http://ecos.sourceware.org/ml/ecos-discuss

[ECOS] Re: eCos on 80186

[Grant Edwards]

In gmane.os.ecos.general, you wrote:

>    Bruno> but I read im someplace that is possible do this changing
>    Bruno> the makefile to compile on other compiler like Borland C++.
>    Bruno> And I would like known if someone did this port.
>
> I suspect you are referring to the ecos-discuss thread
> http://ecos.sourceware.org/ml/ecos-discuss/2004-06/msg00220.html
> The thread included a suggestion to use Borland, but I am not
> aware of anybody writing any code along those lines. eCos
> depends heavily on a number of gcc extensions which makes it
> hard to use any other compiler.

As a purely academic question, what is the "smallest"
processor/platform to which eCos has been ported?  There are
some fairly "wimpy" processors with gcc support (e.g. 6811,
AVR), but they tend to have rather limited address spaces by
eCos standards.

-- 
Grant Edwards                   grante             Yow!  ... I want to perform
                                  at               cranial activities with
                               visi.com            Tuesday Weld!!

-- 
Before posting, please read the FAQ: http://ecos.sourceware.org/fom/ecos
and search the list archive: http://ecos.sourceware.org/ml/ecos-discuss

Re: [ECOS] Re: eCos on 80186

[Bart Veer]

>>>>> "Grant" == Grant Edwards <grante@visi.com> writes:

    Grant> As a purely academic question, what is the "smallest"
    Grant> processor/platform to which eCos has been ported? There are
    Grant> some fairly "wimpy" processors with gcc support (e.g. 6811,
    Grant> AVR), but they tend to have rather limited address spaces
    Grant> by eCos standards.

Obviously this depends on the definition of "smallest".

I have RedBoot running on an old vanilla 68000, but haven't finished
the eCos side yet. In terms of transistor count I suspect that is the
smallest - obviously the lack of on-chip peripherals helps.

Anoncvs has support for h8300 and calmrisc16 (the latter RedBoot-only
IIRC). From a computer science/architecture perspective I suspect one
of those two would qualify as "smallest". I have not been involved
with those ports so don't know much about them.

For practical hardware design purposes I suspect the LPC2xxx ARM chips
currently qualify as smallest.
Bart

-- 
Bart Veer                       eCos Configuration Architect
http://www.ecoscentric.com/     The eCos and RedBoot experts


-- 
Before posting, please read the FAQ: http://ecos.sourceware.org/fom/ecos
and search the list archive: http://ecos.sourceware.org/ml/ecos-discuss

[ECOS] Re: eCos on 80186

[Grant Edwards]

In gmane.os.ecos.general, you wrote:
>>>>>> "Grant" == Grant Edwards <grante@visi.com> writes:
>
>    Grant> As a purely academic question, what is the "smallest"
>    Grant> processor/platform to which eCos has been ported? There are
>    Grant> some fairly "wimpy" processors with gcc support (e.g. 6811,
>    Grant> AVR), but they tend to have rather limited address spaces
>    Grant> by eCos standards.
>
> Obviously this depends on the definition of "smallest".
>
> I have RedBoot running on an old vanilla 68000, but haven't finished
> the eCos side yet. In terms of transistor count I suspect that is the
> smallest - obviously the lack of on-chip peripherals helps.
>
> Anoncvs has support for h8300 and calmrisc16 (the latter
> RedBoot-only IIRC). From a computer science/architecture
> perspective I suspect one of those two would qualify as
> "smallest".

The H8/300H is a 32-bit CISC processor a 24-bit linear address
space. I don't really consider it "small" from an embedded OS
POV.

CalmRISC seems to be more what I hand in mind. It has a 20 bit
program address space and a 16 bit data address space.

> For practical hardware design purposes I suspect the LPC2xxx
> ARM chips currently qualify as smallest.

I was thinking more along the lines of address space and ALU
width (e.g. something like the 6811 with an 8/16 bit ALU and 16
bit linear address space).  I can see how you might be able to
shoehorn a stripped down eCos[1] into a harvard architecture
with two 16-bit address spaces, but a single 16-bit address
space doesn't sound practical.

[1] A minimal multi-tasking kernel with a few services like
    mutexes and semaphores.  A BSD network stack is obviously
    out of the question.

-- 
Grant Edwards                   grante             Yow!  I feel like I'm
                                  at               in a Toilet Bowl with a
                               visi.com            thumbtack in my forehead!!

-- 
Before posting, please read the FAQ: http://ecos.sourceware.org/fom/ecos
and search the list archive: http://ecos.sourceware.org/ml/ecos-discuss