public inbox for gas2@sourceware.org
 help / color / mirror / Atom feed
* possible bug in bfd/sunos.c
@ 1996-05-28  8:56 Joel Sherrill
  1996-05-28  9:07 ` Ian Lance Taylor
  0 siblings, 1 reply; 3+ messages in thread
From: Joel Sherrill @ 1996-05-28  8:56 UTC (permalink / raw)
  To: gas2

Version binutils-2.6.

I am configured for sparc-rtems which is essentially the same as 
sparc-aout with different CPP_PREDEFINES.  I use the following assembly 
file (cut down to the point of uselessness :) ):

 
        .text
        sethi   %hi(stack_space + 16 * 1024), %g1
        call      _main
 
        .comm   stack_space, 16 * 1024

I then run this script on it:

sparc-rtems-as -o test.o test.s
sparc-rtems-ld test.o

and get the following message:

bfd assertion fail ../../src/bfd/sunos.c:1839

If I delete any line in the file other than the .text, the problem goes 
away.  Is this something I am doing which is stupid?  Is this a real 
problem?  I looked at the source for this problem but (in my usual luck) 
requires real expertise to even read.

I noticed that the assert in question has changed in the latest binutils 
snapshot.  Is this the bug?  If so is there a simple patch for this 
specific problem?

Any assistance is appreciated.

--joel

^ permalink raw reply	[flat|nested] 3+ messages in thread

* Re: possible bug in bfd/sunos.c
  1996-05-28  8:56 possible bug in bfd/sunos.c Joel Sherrill
@ 1996-05-28  9:07 ` Ian Lance Taylor
  1996-05-28 12:14   ` Joel Sherrill
  0 siblings, 1 reply; 3+ messages in thread
From: Ian Lance Taylor @ 1996-05-28  9:07 UTC (permalink / raw)
  To: joel; +Cc: gas2

   Date: Tue, 28 May 1996 10:50:25 -0500 (CDT)
   From: Joel Sherrill <joel@merlin.gcs.redstone.army.mil>

   I am configured for sparc-rtems which is essentially the same as 
   sparc-aout with different CPP_PREDEFINES.  I use the following assembly 
   file (cut down to the point of uselessness :) ):


	   .text
	   sethi   %hi(stack_space + 16 * 1024), %g1
	   call      _main

	   .comm   stack_space, 16 * 1024

   I then run this script on it:

   sparc-rtems-as -o test.o test.s
   sparc-rtems-ld test.o

   and get the following message:

   bfd assertion fail ../../src/bfd/sunos.c:1839

If you configure for a SunOS target, the linker will automatically
generate a shared library if there are any undefined symbols in the
link.  This odd behaviour is for compatibility with the SunOS linker.

In the current snapshots, the sparc-aout target does not have this
behaviour, since shared libraries are not useful for embedded systems,
but in 2.6 the sparc-aout target is the same as SunOS.

In your example, the linker is created a shared library.

The assertion failure you are getting is, as you guessed, a bug in the
assertion condition.  You can just replace the BFD_ASSERT with one
from a current snapshot to avoid it.  It is likely, though, that you
do not want to create a shared library at all.

Ian


^ permalink raw reply	[flat|nested] 3+ messages in thread

* Re: possible bug in bfd/sunos.c
  1996-05-28  9:07 ` Ian Lance Taylor
@ 1996-05-28 12:14   ` Joel Sherrill
  0 siblings, 0 replies; 3+ messages in thread
From: Joel Sherrill @ 1996-05-28 12:14 UTC (permalink / raw)
  To: Ian Lance Taylor; +Cc: gas2

Thanks for the tip.  Adding the third condition to the assert in question 
fixed the problem.  Now I get the undefined message I did not even know 
to expect. :)

--joel

^ permalink raw reply	[flat|nested] 3+ messages in thread

end of thread, other threads:[~1996-05-28 12:14 UTC | newest]

Thread overview: 3+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
1996-05-28  8:56 possible bug in bfd/sunos.c Joel Sherrill
1996-05-28  9:07 ` Ian Lance Taylor
1996-05-28 12:14   ` Joel Sherrill

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).