[uClinux-dev] strace

Emilena Specht emilena.specht at gmail.com
Thu Aug 4 16:43:23 EDT 2005


Well, Lars, you were right, I was not running configure for m68k-linux.
Instead of following these steps:

 export CC="m68k-elf-gcc -m5307 -msep-data"
 export LDFLAGS="-Wl,-elf2flt"
 ../strace-4.4.94/configure --host m68k-linux

I´ve done these:

export CC="m68k-elf-gcc -m5307 -msep-data --version=2.95.3"
export LDFLAGS="-Wl,-elf2flt"
 ../strace-4.4.94/configure --host=m68k-linux

and then It worked. 
But configure shows some strange warnings, like:

 checking linux/ptrace.h usability... no
checking linux/ptrace.h presence... yes
configure: WARNING: linux/ptrace.h: present but cannot be compiled
configure: WARNING: linux/ptrace.h: check for missing prerequisite headers?
configure: WARNING: linux/ptrace.h: proceeding with the preprocessor's result
configure: WARNING:     ## ------------------------------------ ##
configure: WARNING:     ## Report this to bug-autoconf at gnu.org. ##
configure: WARNING:     ## ------------------------------------ ##
checking for linux/ptrace.h... yes
checking asm/reg.h usability... no
checking asm/reg.h presence... no

Maybe this is leading to errors when compiling:

make[1]: Entering directory `/moto/uClinux-dist/user/strace'
make[1]: *** No rule to make target `strace.', needed by `strace'.  Stop.

I´m working on it then....

Emilena


On 8/2/05, Friedrich, Lars <lars.friedrich at wago.com> wrote:
> > Thank you Lars, I used m68k-linux as the targed and configure
> > ran well.
> > Now I get some errors while compiling:
> >
> > file.c: In function `realprintstat':
> > file.c:737: error: structure has no member named `st_atim'
> > file.c:738: error: structure has no member named `st_mtim'
> > file.c:739: error: structure has no member named `st_ctim'
> [...]
> 
> The xstat structure in strace is newer than the one in
> uClibc/glibc/uC-libc. Previously, time for stat was stored
> in unsigned ints a la st_atime, st_mtime etc. This was then
> changed to struct timespec st_atim etc.
> 
> I suggest you change the source code for strace to either
> work with the old structure or remove the code completely
> that accesses the structure. realprintstat and printstat64
> doesn't sound very important for the overall strace
> functionality.
> The other option would be to upgrade the user space library,
> which might be much more painful.
> 
> Best regards,
> Lars Friedrich
> _______________________________________________
> uClinux-dev mailing list
> uClinux-dev at uclinux.org
> http://mailman.uclinux.org/mailman/listinfo/uclinux-dev
> This message was resent by uclinux-dev at uclinux.org
>



More information about the uClinux-dev mailing list