sh3, newlib, and sprintf

David Williams davidwilliams@ozemail.com.au
Sun Dec 19 14:17:00 GMT 1999


I had a problem with newlib and sprintf doing floating point on a 68000. 
The problem was that dynamic memory was not set up properly and only 
floating point format required dynamic memory, all other types were ok. As 
soon as I setup dynamic memory all was ok.

Dave.

----------
From: 	Bryan Ogawa[SMTP:bogawa@vovida.com]
Sent: 	Friday, December 17, 1999 3:16 PM
To: 	Steve Goodrich
Cc: 	crossgcc@sourceware.cygnus.com
Subject: 	Re: sh3, newlib, and sprintf

On Mon, 8 Nov 1999, Steve Goodrich wrote:

> I have an Hitachi HS7709 eval board that I am trying to get up and
> running on. I am having problems with the sprintf function. sprintf
> seems to work unless I use a floating point parameter, i.e.
> "sprintf(ca, "%f", a);" where ca is a 32 byte character array and a is
> a double.

Steve, I know you may have already resolved this, as this message is quite
old.  I can't offer you clear advice, but just a possible rememberance.

When we were trying to build a toolchain for sh3 boards, I seem to
remember that there were places where the toolchain needs to know things
about the floating point architecture of the system, and that those things
weren't exactly there for the SH3 processor.

I seem to recall that it was in newlib that this was the case.  We didn't
have to resolve it because we could bootstrap someone else's IO libs, but
you may need to define stuff there for floating point.

--
Bryan K. Ogawa  <bogawa@vovida.com>  http://www.vovida.com/


------
Want more information?  See the CrossGCC FAQ, 
http://www.objsw.com/CrossGCC/
Want to unsubscribe? Send a note to 
crossgcc-unsubscribe@sourceware.cygnus.com



------
Want more information?  See the CrossGCC FAQ, http://www.objsw.com/CrossGCC/
Want to unsubscribe? Send a note to crossgcc-unsubscribe@sourceware.cygnus.com



More information about the crossgcc mailing list