Re: RFA: ia64 tdep patch

["J. Johnston" <jjohnstn@redhat.com>]

Kevin Buettner wrote:
> On Oct 23, 12:22pm, J. Johnston wrote:
> 
> 
>>I have included a patch which contains responses to your comments thus far.
>>
>>The ChangeLog is:
>>
>>2003-10-23  Jeff Johnston  <jjohnstn@redhat.com>
>>	* ia64-tdep.c: (ia64_frame_cache): Add new prev_cfm field.
>>	(pseudo_regs): Add comment regarding register stack registers.
>>	(ia64_alloc_frame_cache):  Initialize new prev_cfm field to 0.
>>	(floatformat_valid): New static routine.
>>	(floatformat_ia64_ext): Add name field and set up is_valid routine
>>	to floatformat_valid().
>>	(examine_prologue):  For the previous cfm, use
>>	frame_unwind_register()
>>	if the cfm is not stored in a register-stack register.  Save the
>>	previous cfm value in the prev_cfm field.  Add debug output.
>>	(ia64_frame_this_id): Use frame_id_build_special() to also register
>>	the bsp.  Add debug output.
>>	(ia64_sigtramp_frame_this_id): Ditto.
>>	(ia64_frame_prev_register):  Look at cache saved_regs for a few more
>>	registers and also add some checks for framelessness before accepting
>>	current register values for fields such as return address.  For cfm,
>>	use the cached prev_cfm field if available.  Add comment to explain
>>	PSR logic.  Add debug output.
>>	(ia64_sigtramp_frame_init_saved_regs): Bump up base by 16 to get
>>	sp needed for calling lower level
>>	ia64_linux_sigcontext_register_address().  Also save the
>>	bsp and sp address as part of initialization.
>>	(ia64_sigtramp_frame_cache): Hard-code stack size as it can't be
>>	calculated.  Cache the bsp and cfm values.
>>	(ia64_sigtramp_frame_prev_register): Add logic to this routine out
>>	instead of using ia64_frame_prev_register() which doesn't expect most
>>	registers to be saved.  The saved values for bsp and sp
>>	can be taken from the cache.  Add debug output.
>>	(ia64_push_dummy_call): Use frame_id_build_special() to also register
>>	the bsp.
>>
>>Ok?
> 
> 
> Okay.
> 

Thanks.  Patch checked into mainline.

-- Jeff J.