Re: [patch rfc] Rationalize FIX_CALL_DUMMY

[Andrew Cagney <ac131313 at redhat dot com>]

> Hello,
>
> This one makes FIX_CALL_DUMMY optional.  It doesn't try to address the problem of HP/PA's version having a different function signature to the version in the architecture vector (the former returns a PC address).
>
> Long term, I think all the call dummy methods should be consolidated into a single push_dummy_return_point() method.  That would write the dummy call's return code directly to the stack.

I've checked this in.

Note that while the original change log mentioned changing 
fix_call_dummy to a function that returned an address, the actual change 
didn't include it - I abandoned the idea and fixed the ChangeLog.

Andrew


> 2003-03-29  Andrew Cagney  <cagney at redhat dot com>
>
> 	* gdbarch.sh (FIX_CALL_DUMMY): Change to function with predicate.
> 	* gdbarch.h, gdbarch.c: Regenerate.
> 	* inferior.h (FIX_CALL_DUMMY): Delete macro.
> 	* valops.c (hand_function_call): Only call FIX_CALL_DUMMY when
> 	available.
> 	* frame.h (generic_fix_call_dummy): Delete declaration.
> 	* dummy-frame.h: Update comment.
> 	* dummy-frame.c (generic_fix_call_dummy): Delete function.
> 	* xstormy16-tdep.c (xstormy16_gdbarch_init): Do not set
> 	fix_call_dummy.
> 	* sh-tdep.c (sh_gdbarch_init): Ditto.
> 	* s390-tdep.c (s390_gdbarch_init): Ditto.
> 	* mn10300-tdep.c (mn10300_gdbarch_init): Ditto.
> 	* mcore-tdep.c (mcore_gdbarch_init): Ditto.
> 	* m68hc11-tdep.c (m68hc11_gdbarch_init): Ditto.
> 	* ia64-tdep.c (ia64_gdbarch_init): Ditto.
> 	* i386-tdep.c (i386_gdbarch_init): Ditto.
> 	* h8300-tdep.c (h8300_gdbarch_init): Ditto.
> 	* frv-tdep.c (frv_gdbarch_init): Ditto.
> 	* d10v-tdep.c (d10v_gdbarch_init): Ditto.
> 	* cris-tdep.c (cris_gdbarch_init): Ditto.
> 	* avr-tdep.c (avr_gdbarch_init): Ditto.
> 	* arm-tdep.c (arm_gdbarch_init): Ditto.