This is the mail archive of the
gdb@sources.redhat.com
mailing list for the GDB project.
HAVE_I387_REGS
- From: Kevin Nomura <nomura at netapp dot com>
- To: gdb at sources dot redhat dot com
- Date: Mon, 19 Aug 2002 17:47:16 -0700
- Subject: HAVE_I387_REGS
gdb 5.2.x introduced HAVE_I387_REGS in the configuration for
i386 targets. I am applying these changes to a target which
does not dump FP registers to the core file, so this may be
the first test of this #define not being set.
For this non-FP target, "i registers" complains about registers
like $SP not being convertible to floating point. From the
following defines, FP_REGNUM_P instead of giving a definitive
false when HAVE_I387_REGS is not defined, will just change the
range of numbers it says are floating point. In particular it
makes regnos 0-7 look like floating point; those are my GPRs.
gdb/config/i386/tm-i386.h:
> /* First FPU data register. */
> #ifdef HAVE_I387_REGS
> #define FP0_REGNUM 16
> #else
> #define FP0_REGNUM 0
> #endif
gdb/i386-tdep.h:
/* Return non-zero if N corresponds to a FPU data registers. */
#define FP_REGNUM_P(n) (FP0_REGNUM <= (n) && (n) < FPC_REGNUM)
/* "Generic" floating point control register. */
#define FPC_REGNUM (FP0_REGNUM + 8)