Subject: Re: error building mipsel-linux gmon.c - msg#00032

Previous Message by Date: click to view message preview

Re: mips build

Hello, Im trying to complile uclibc 0.9.19 for mips. I am using mips-linux- binutils 2.13.1 and egcs-mips-linux-1.1.2. My complitation doesnt go thru and i get the follwing errors. Can anyone help me !? Thanks in advance, Avinash ----------------------------------------------------------------------- --- mips-linux-gcc -Wall -Wstrict-prototypes -Wno-trigraphs -fno-strict-aliasing -Os -fno-builtin -nostdinc -D_LIBC -I../../include -I. -g -I/usr/lib/gcc-lib/mips-linux/egcs-2.91.66/include -DNDEBUG -fPIC -DL_llabs stdlib.c -c -o llabs.o stdlib.c:52: warning: integer constant out of range stdlib.c:52: warning: integer constant is so large that it is unsigned stdlib.c:101: warning: integer constant out of range stdlib.c:101: warning: integer constant is so large that it is unsigned stdlib.c:180: warning: integer constant out of range stdlib.c:180: warning: integer constant is so large that it is unsigned stdlib.c:182: warning: integer constant out of range stdlib.c:182: warning: integer constant is so large that it is unsigned stdlib.c:182: warning: integer constant out of range stdlib.c:183: `llabs' undeclared here (not in a function) stdlib.c:183: warning: type defaults to `int' in declaration of `imaxabs' make[2]: *** [llabs.o] Error 1 make[2]: Leaving directory `/home/avinash/ulibc/uClibc-0.9.19/libc/stdlib' make[1]: *** [_dir_stdlib] Error 2 make[1]: Leaving directory `/home/avinash/ulibc/uClibc-0.9.19/libc' make: *** [_dir_libc] Error 2

Next Message by Date: click to view message preview

mips, floats, and stdio

Hey guys. I built my toolchain using the toolchain script and the latest 0.9.19 release (not the snapshot.) I notice that whenever I use printf with a float, it segfaults. The processor is a MIPS R4k core, and there is Kernel FPU emulation. Interestingly enough, cout with floats seems to work. I was hoping someone could give me a hint or two. I just spent a week getting the built-in CompactFlash driver working on this part and my 900k statically linked (to glibc) iwconfig makes my initrd not fit in the flash heh. Steve

Previous Message by Thread: click to view message preview

Re: error building mipsel-linux gmon.c

On Mon Apr 14, 2003 at 04:09:28PM +0100, Liam Girdwood wrote: > On Mon, 2003-04-14 at 15:30, Joel Coltoff wrote: > > On 9 Apr 2003, Liam Girdwood wrote: > > > > > Hi, > > > > > > I've hit another problem building for mipsel-linux. > > > > > > gmon.c: At top level: > > > gmon.c:517: parse error before "__mcount" > > > gmon.c:518: warning: return type defaults to `int' > > > > > > I'm using gcc 3.2, and it's having difficulty with __attribute_used__ > > > in machine-gmon.h. I know that gcc 3.1 needs this definition in order to > > > emit the proper code, but I'm not so sure about gcc 3.2 > > > > > > I've removed __attribute_used__ from machine-gmon.h and it now builds, > > > but I haven't had a change to properly test it yet. > > > > Have you gotten any feedback on this. We are in the middle of a long > > needed update (from 0.9.12) and will bump into this. I'd like to know > > if it's safe to make the same change here. I'm probably not in a position > > to determine if it's a safe change. We'll be using 3.2.2. A co-worker > > downloaded the binaries of a native compiler and hasn't had any problems > > using it. > > > > I've not had any feedback on this yet, but it only applies to the CVS > version when profiling support is enabled. The last release 0.9.19 > compiles ok. I see no reason why removing the __attribute_used__ should cause you any problems. Perhaps we need to add in something like #if __GNUC__ <= 3 && __GNUC_MINOR__ <= 1 before adding the __attribute_used__? > As for testing the profiling, I've not yet had a chance. I have been hoping for feedback on whether profiling actually works on mips, so when you get a chance... -Erik -- Erik B. Andersen http://codepoet-consulting.com/ --This message was written using 73% post-consumer electrons--

Next Message by Thread: click to view message preview

Re: error building mipsel-linux gmon.c

On Tue Apr 15, 2003 at 01:55:08PM +0100, Liam Girdwood wrote: > I agree, this would be the best solution. Here is the patch > > diff -u -r1.1 machine-gmon.h > --- libc/sysdeps/linux/mips/bits/machine-gmon.h 3 Mar 2003 20:58:25 > -0000 1.1 > +++ libc/sysdeps/linux/mips/bits/machine-gmon.h 15 Apr 2003 12:58:40 > -0000 > @@ -17,8 +17,13 @@ > Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA > 02111-1307 USA. */ > > +#if __GNUC__ <= 3 && __GNUC_MINOR__ <= 1 > #define _MCOUNT_DECL(frompc,selfpc) \ > static void __attribute_used__ __mcount (u_long frompc, u_long selfpc) > +#else > +#define _MCOUNT_DECL(frompc,selfpc) \ > +static void __mcount (u_long frompc, u_long selfpc) > +#endif > > /* Call __mcount with our the return PC for our caller, > and the return PC our caller will return to. */ > > > > However, it may only apply to a cross gcc3.2. Has anyone tried building > the current CVS using a native gcc3.2 ? I am upgrading my mips box at the moment. When that finishes, (another 88 MB or stuff still to download, so around 20 minutes) I will then have gcc 3.2 installed and running natively. I'll give it a try then both with and without the patch, -Erik -- Erik B. Andersen http://codepoet-consulting.com/ --This message was written using 73% post-consumer electrons--