cc1 SEGV (2.94.17 19990411 - current cvs)

Matti Aarnio matti.aarnio@sonera.fi
Fri Apr 30 23:15:00 GMT 1999


Richard Henderson <rth@cygnus.com> replied:
> On Sun, Apr 11, 1999 at 11:26:17PM +0300, Matti Aarnio wrote:
> > Program received signal SIGSEGV, Segmentation fault.
> > find_replacement (loc=0x8) at ../../gcc/reload.c:5853
> > 5853      if (GET_CODE (*loc) == PLUS || GET_CODE (*loc) == MINUS
> > (gdb) where
> > #0  find_replacement (loc=0x8) at ../../gcc/reload.c:5853
> > #1  0x1201b00e4 in get_unaligned_address (ref=0x8, extra_offset=0) at ...
> 
> Out of range stack slots suck.  Dynamically allocate
>   u_char answer[32*1024];
> and you'll get much better code.  Not that you really care for
> something like dig...
> 
> Anyway, this should fix the segv. 
> r~

	Yep, it did fix that.  Now even that pathological animal
	known as 'dig' did compile successfully.

/Matti Aarnio <matti.aarnio@sonera.fi>



More information about the Gcc-bugs mailing list