This is the mail archive of the ecos-patches@sourceware.org mailing list for the eCos project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: Initialisation for the AT91... PIT if there is no kernel...


On Fri, Mar 24, 2006 at 10:55:14PM +0100, oliver munz @ s p e a g wrote:
> If i use only "Clean" and "Library F7" or only "Library F7" from the 
> configtool, then changes in my CDL-seting's don't work the right way.
> Thats because the "Clean" command don't delete the makefiles. And in the 
> "\andrew1_build\hal\arm\at91\at91sam7s\current\makefile"  is written:

I don't use the configtool, but i would of expected that when you
change a configuration option it does the equivelent of

ecosconfig tree

so it remakes all the makefile. Please could you check if this is
true. If not, it sounds like a configtool bug.

Anyway, this won't solve the problem. The build system is known to be
broken in another way. Doing a

make clean 

does not delete 

install/lib/libtarget.a

So when you build the new configuration you end up with both
timer_pit.o and timer_tc.o in the library. Which is then used by the
linker is undefined, probably the first one in the library.

The only guaranteed way to ensure your build really is what you have
configured is to

rm -fr [^e]* error
ecosconfig tree
make

Just doing a 

make clean ; make

works most of the time, but not always.

Your #ifdef won't help either becasue the object file was already been
compiled and put into the library.

        Andrew


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