From mboxrd@z Thu Jan  1 00:00:00 1970
From: hjl@lucon.org (H.J. Lu)
To: law@cygnus.com
Cc: hjl@lucon.org, gafton@redhat.com, rth@cygnus.com, egcs@cygnus.com
Subject: Re: egcs 1.0.2 is broken on x86 and a patch for it.
Date: Mon, 20 Apr 1998 13:06:00 -0000
Message-id: <m0yRHEa-00058wC@ocean.lucon.org>
References: <17624.893055503@hurl.cygnus.com>
X-SW-Source: 1998-04/msg00812.html

> 
>   > Basically convert_regs () calls change_stack () with INSN to emit
>   > some insn after INSN. But some insn has been added after INSN so
>   > that change_stack () will emit some insn at the wrong place. As the
>   > result, the floating point is broken on x86.
> This is an extremely confusing explanation, primarily because you
> use "insn"/INSN to refer to three different things.
> 
> Basically is sounds like you have
> 
> 	insn 1
>         insn 2
> 
> 
> It sounds like we thought we wanted to insert after insn 1, but
> because of other reg-stack actions we really wanted to insert
> after insn 2.

It is close. Originally, we have

	insn 1

and we should insert something after insn 1. But in reg-stack.c,
something is added to fit the register stack and we get

	insn 1
	insn 2

before we are just about to insert something after insn 1.
Now we want to insert something after the original insn 1. Where
should we insert? After insn 1 or insn 2? insn 2 is added to
adjust the register stack and it is really the extention of
insn 1. If we add anything after insn 1 now, we just screw
up the purpose of adding insn 2 which is fit the register stack.
So things should be inserted after the original insn 1 should
be inserted after insn 2 now.

> 
>   > This patch seems to fix it. Could someone please take a look? Given
>   > the bugs we have seen in egcs 1.0.2, I suggest egcs 1.0.3 be made.
> I must confess I don't undersatnd reg-stack all that well.  Assuming
> your analysis and explanation are correct, then I think your change
> is OK, though possibly incomplete.
> 
> In particular I worry that we need to pass "new" instead of "insn"
> to the call to emit_pop_insn near the end of convert_regs.  I think
> the call to goto_block_pat in convert_regs is OK.
> 
> I also worry that there may be cases were we need to insert after
> insn 1 instead of after insn 2.  But I don't know reg-stack well
> enough to know if that can actually happen.
> 
> You might ask Bernd Schmidt or Stan Cox to look at the change and
> comment.
> 

Are they reading this list?

-- 
H.J. Lu (hjl@gnu.org)