This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: Updated PATCH for expr.c problems (MIPS, 386), 1.1.2 and current
- To: Hans-Peter Nilsson <hp at bitrange dot com>
- Subject: Re: Updated PATCH for expr.c problems (MIPS, 386), 1.1.2 and current
- From: Jeffrey A Law <law at upchuck dot cygnus dot com>
- Date: Sun, 11 Apr 1999 22:31:48 -0600
- cc: plai at Lynx dot COM, egcs-patches at egcs dot cygnus dot com
- Reply-To: law at cygnus dot com
In message <Pine.BSF.4.02A.9903301943390.13846-100000@dair.pair.com>you write
:
> Here's an update to the patch in
> <URL:http://egcs.cygnus.com/ml/egcs-patches/1999-02/msg00796.html>.
> (It's close, but does not solve the problem you report in
> <URL:http://egcs.cygnus.com/ml/egcs-bugs/1999-03/msg00973.html>.)
>
> The force_operand is necessary since the address is often (as in
> <URL:http://egcs.cygnus.com/ml/egcs-bugs/1999-02/msg00786.html>
> and <URL:http://www.cygnus.com/ml/egcs-bugs/1998-Oct/0928.html>)
> a (plus ...), which is not correct input for force_reg. The ptr_mode
> is because we're dealing with an address; there's no need to look inside
> the source rtx for the mode. That's also fatal when dealing with
> addresses that are const_int:s...
You are _still_ missing the point.
If you fix the mips backend to not accept addresses that are invalid, then you
will not get the bogus argument to force_reg.
I will not install this patch as-is. We need to fix the MIPS backend first.
After that we can re-evaluate whether or not we need the expr.c patch.
jeff