To LUGNET HomepageTo LUGNET News HomepageTo LUGNET Guide Homepage
 Help on Searching
 
Post new message to lugnet.robotics.rcx.legosOpen lugnet.robotics.rcx.legos in your NNTP NewsreaderTo LUGNET News Traffic PageSign In (Members)
 Robotics / RCX / legOS / 1799
1798  |  1800
Subject: 
Re: Library use in LegOS
Newsgroups: 
lugnet.robotics.rcx.legos
Date: 
Tue, 27 Mar 2001 12:16:30 GMT
Viewed: 
1384 times
  
In lugnet.robotics.rcx.legos, Jakob Engblom writes:
Each program gets it's own copy

OK.

How are the calls to the kernel resolved? By the makelx program, or by some
magical stuff in the linking.

I can imagine compiling the kernel, getting a symbol table out, and then
when compiling user programs, you use a linker DEFINE or such to tell the
code that yes, there is such a function, at this absolute address.  However,
getting this to work automatically seems like a lot of work.

Not sure of the details, but the linker takes the .o file, and the symbol file
LegOS.lds, created during the kernel compile, and produces two files; .ds1 &
.ds2, which are then passed to makelx to produce the lx file. So it seems the
linker resolves kernel symbols, not makelx.

If I would use a different compiler, with a different set of libraries for
the program and the kernel, is there the potential for namn clashes?

Hmmmm. Not sure. Wish I knew more about compiler internals...

Well, if each program uses its own libraries, there is no risk.  The linked
binary from my compiler will call LegOS via some mechanism, and its own
libraries internally in the binary.  No problem.

How much of the LegOS-gcc standard libraires come from gcc and how much • have
been hacked specifically for legos?

It's all custom written - doesn't use the standard gcc libraries at all. In
fact when building the cross compiler, you have to tell it to forget about
the standard libraries.

Why is that?  For the IAR compiler we are going to use, the libraries are
completely operating-system neutral (strlen, memcpy, math libs, etc have
nothing to do with the operating system).

I think it's mainly for size - but I wasn't involved with the early versions of
LegOS, so maybe someone else knows more...

I guess libc could also be used for the H8 but I don't know if anyones tried
this.

Lots of questions, but I am trying what apparently nobody has tried before,
since I only have seen gcc discussed here in the legOS forum.



Message has 1 Reply:
  Re: Library use in LegOS
 
(...) of (...) For the IAR compiler, this should not a problem. The libraries are optimized for minimal size, so we can probably use the standard libraires. And give our students the benefit of a full C lib. /jakob (24 years ago, 28-Mar-01, to lugnet.robotics.rcx.legos)

Message is in Reply To:
  Re: Library use in LegOS
 
(...) OK. How are the calls to the kernel resolved? By the makelx program, or by some magical stuff in the linking. I can imagine compiling the kernel, getting a symbol table out, and then when compiling user programs, you use a linker DEFINE or (...) (24 years ago, 27-Mar-01, to lugnet.robotics.rcx.legos)

5 Messages in This Thread:

Entire Thread on One Page:
Nested:  All | Brief | Compact | Dots
Linear:  All | Brief | Compact

This Message and its Replies on One Page:
Nested:  All | Brief | Compact | Dots
Linear:  All | Brief | Compact
    

Custom Search

©2005 LUGNET. All rights reserved. - hosted by steinbruch.info GbR