[6.0] glibc compile memcmp compile problem

Joel Miller cheeziologist at mail.isc.rit.edu
Tue Dec 14 16:56:59 PST 2004

Joel Mayes wrote:
> ../sysdeps/i386/i686/memcmp.S: Assembler messages:
> ../sysdeps/i386/i686/memcmp.S:390: Error: can't resolve `.rodata'
> {.rodata section} - `memcmp' {.text section}
> make[2]: *** [/mnt/lfs/tools/build/glibc-build/string/memcmp.o] Error 1
> make[2]: Leaving directory `/mnt/lfs/tools/build/glibc-2.3.4-20040701/string'
> make[1]: *** [string/subdir_lib] Error 2
> make[1]: Leaving directory `/mnt/lfs/tools/build/glibc-2.3.4-20040701'
> make: *** [all] Error 2

Ouch, assembly problems. Let me preface this by saying that assembly is 
very far from my strongsuit. From, what I can gather though, it appears 
this happens when the assembler tries to resolve two symbols that are in 
different address spaces. I came across one post that says this will 
happen with glibcs >20040501 when compiling with gcc <3.4 but that 
doesn't seem right becuase I know that the glibc in chapter five should 
be being built with the gcc built in chapter five just earlier. Are you 
sure you have /tools/bin at the beginning of your PATH and hashing is 
turned off on the shell? You want to be sure you are using the gcc thats 
installed in the /tools hierarchy.

Registered LFS User 6929
Registered Linux User 298182

More information about the lfs-support mailing list