[ECOS] Networkdebugging on i386

[ECOS] Networkdebugging on i386

[Christian Plessl]

Hi folks

Some days ago I've already made a posting to check whether somebody 
succeded in adapting RedBoot to i386.

I'm mainly interessted in debugging remotely via ethernet. RedBoot should 
give me that option, does it?

So, did anybody manage to build RedBoot for i386? I suppose it needs some 
work in the "obscure" virtual vector code (what ever this is, I could not 
figure out this yet...)

If it is not easy to get running RedBoot, is there an alternative for 
adding network debugging support for gdb?

Thanks in advance for any hint!

Chris


--
Christian Plessl <cplessl@ee.ethz.ch>

Re: [ECOS] Networkdebugging on i386

[Jesper Skov]

>>>>> "Christian" == Christian Plessl <cplessl@ee.ethz.ch> writes:

Christian> I'm mainly interessted in debugging remotely via
Christian> ethernet. RedBoot should give me that option, does it?

I believe there is contributed RedBoot support for x86. But I don't
know if it's available for download. Check through the archives.

Christian> So, did anybody manage to build RedBoot for i386? I suppose
Christian> it needs some work in the "obscure" virtual vector code
Christian> (what ever this is, I could not figure out this yet...)

Obscure? :)  Did you read this:
http://sources.redhat.com/ecos/docs-latest/porting/hal-calling-if.html

Granted, it needs updating as I've just made some changes today. And
it could do with a short primer on _why_ we need the virtual vectors,
not just what they are. That'll all come in due course.

Jesper

[ECOS] Re: Networkdebugging on i386

[Fabrice Gautier]

On Thu, 01 Feb 2001 15:46:04 +0100
Christian Plessl <cplessl@ee.ethz.ch> wrote:

> Hi folks
> 
> Some days ago I've already made a posting to check whether somebody 
> succeded in adapting RedBoot to i386.

Yep.

> I'm mainly interessted in debugging remotely via ethernet. RedBoot should 
> give me that option, does it?

Yep.

 
> So, did anybody manage to build RedBoot for i386? I suppose it needs some 
> work in the "obscure" virtual vector code (what ever this is, I could not 
> figure out this yet...)

Yep.

> 
> If it is not easy to get running RedBoot, is there an alternative for 
> adding network debugging support for gdb?

Well, I did get RedBoot running for x86 and I did a port of the intel
pro 100 ethernet driver for x86 too  so i did debug with ethernet.

I made some patches that you should find in the archive. But they're two
monthes old at least and haven't maintened them since them and i've seen
quite a few changes in the CVS since then so there maybe some works.

If you can't find the patches or the URL to those patches, i can send
the patches that I have left on my hard drive..


> Thanks in advance for any hint!

Y're welcome.

Fabrice


-- 
Fabrice Gautier <gautier@email.enstfr>

Re: [ECOS] Networkdebugging on i386

[Christian Plessl]

At 15:49 01.02.01 +0000, Jesper Skov wrote:
> >>>>> "Christian" == Christian Plessl <cplessl@ee.ethz.ch> writes:
>
>Christian> I'm mainly interessted in debugging remotely via
>Christian> ethernet. RedBoot should give me that option, does it?
>
>I believe there is contributed RedBoot support for x86. But I don't
>know if it's available for download. Check through the archives.

Yes, as far as I know yet, Fabrice Gautier's patches should provide RedBoot 
support for x86.

I saw an older posting of you, dating from January, 5th, where you (Jesper) 
announced to add virtual vector support to x86 target.

Jifl was working on including the Gautier patches into the official eCos 
cvs tree. What is the status of that work?

>Christian> So, did anybody manage to build RedBoot for i386? I suppose
>Christian> it needs some work in the "obscure" virtual vector code
>Christian> (what ever this is, I could not figure out this yet...)
>
>Obscure? :)  Did you read this:
> http://sources.redhat.com/ecos/docs-latest/porting/hal-calling-if.html
>
>Granted, it needs updating as I've just made some changes today. And
>it could do with a short primer on _why_ we need the virtual vectors,
>not just what they are. That'll all come in due course.

Thanks for this hint. I wasnt aware that the porting documents describe how 
this redboot stuff works. I will read through this and than I will 
reconsider if it is "obscure" or not :-)

Christian

--
Christian Plessl <cplessl@ee.ethz.ch>