This is the mail archive of the
dwarf2@corp.sgi.com
mailing list for the dwarf2 project.
Re: Request for clarification - call frame instruction DW_CFA_remember_state
- To: BRENDER at gemevn dot zko dot dec dot com, DWARF2 at corp dot sgi dot com, davea at quasar dot engr dot sgi dot com, brender at gemevn dot zko dot dec dot com (Ron 603-884-2088)
- Subject: Re: Request for clarification - call frame instruction DW_CFA_remember_state
- From: David B Anderson <davea at quasar dot engr dot sgi dot com>
- Date: Wed, 28 Feb 2001 10:00:41 -0800 (PST)
- Reply-To: David B Anderson <davea at quasar dot engr dot sgi dot com>
Ron writes:
>Besides, if CFA is (ncessarily) "merely" one of the logical registers,
>why are there several DW_CFA instructions that set/modify the
>CFA specifically and a distinct set of instructions that apply to the
>registers? If the CFA is one of "the registers", then all of the former
>are redundant, yes?
Well, 'merely' is not a word I would use as CFA is a bit
special since other stuff refers to it implicitly. But it is
given a 'dwarf register number' as that is the way it is given
a 'rule' number.
I see no way for the CFA 'rule' number to be other than a
register number, though it is not (that is, not necessarily,
depending on the implementation) a real register. Logically.
There is just one 'namespace' here of rules/registers and if
they overlap...disaster.
I grant it's not as clear as I once thought it was.
Specifically the rules/registers/columns description has at
least one more interpretation than I could see till earlier
today... But I think only one really 'works'. Clarifying
this would be good.
>Me thinks we agree that we want CFA to be saved/restored and only the
>means to assure clarity is at issue. I am open to suggestions...
Agreed.
davea@sgi.com