This is the mail archive of the gcc-bugs@gcc.gnu.org mailing list for the GCC project.


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

Re: 3.0, sun, R_SPARC_32 failure




On Tue, 5 Jun 2001, Nathan Sidwell wrote:
> > What should such a check do (other than warn about binutils)?
> It depends on why these R_SPARC_UA32 things are being emitted.

Typically from .uaword directives, found in dwarf sections like .eh_frame.

> If they cannot be avoided, then the autoconf test should abort the
> configure, indicating that either a later binutils should be used,
> or the native as & ld used. Warning in this situation is unacceptable,
> as we know the resultant compiler will not be useable.

It can't be avoided, at least not without disabling EH entirely.

A simple test is:

echo ".byte 0x1\n.uaword foo" | $(AS) -o foo.o && \
  readelf -r foo.o | grep R_SPARC_32

I don't have the time or knowledge to properly autoconf-ize this test
at the moment :(

Jeff


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