This is the mail archive of the libc-alpha@sources.redhat.com mailing list for the glibc 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]

Build failures in powerpc32/powerpc64 related to _exit


I pulled the cvs head this morning and saw the following on powerpc32
(and something similar for powerpc64):

/home/sjmunroe/work/build32-23/libc.so.6: undefined reference to
`__syscall_exit'
/home/sjmunroe/work/build32-23/libc.so.6: undefined reference to
`__GI__exit'
/home/sjmunroe/work/build32-23/libc.so.6: undefined reference to `_exit'

There are a number of issues but the two most interesting are:

1a) A libc_hidden_def(_exit) statement has been added to
    ./sysdeps/unix/sysv/linux/_exit.c

1b) A HIDDEN_JUMPTARGET(_exit) has been added to:
    ./sysdeps/unix/sysv/linux/powerpc/powerpc32/clone.S
    (and ./sysdeps/unix/sysv/linux/powerpc/powerpc64/clone.S) which
    generates the "__GI" prefix for _exit.

    but ./include/unistd.h contains the following comment:

// XXX Normally we should be able to hide _exit.  But in the thread library
// XXX we might need to overload this function so that all threads can be
// XXX killed.
// libc_hidden_proto (_exit)

Does any one know the final resolution of this? (will _exit be hidden or
not?)

The other issue is that powerpc32 does not define its own INLINE_SYSCALL
macro and the default INLINE_SYSCALL macro generates a  reference to
__syscall_exit which is unresolved.

Is any one else working on this? Powerpc64 has its own INLINE_SYSCALL
macro and I could take a stab at it ...



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