Re: [avr-gcc-list] mcall-prologues completely broken for >128k

[Anatoly Sokolov]

----- Original Message ----- 
From: "Sean D'Epagnier" <address@hidden>
To: "Anatoly Sokolov" <address@hidden>
Cc: "Weddington, Eric" <address@hidden>; 
<address@hidden>
Sent: Wednesday, May 13, 2009 6:35 PM
Subject: Re: [avr-gcc-list] mcall-prologues completely broken for >128k


> On 5/13/09, Anatoly Sokolov <address@hidden> wrote:
> > Hi.
> >
> >
> > > On 5/12/09, Weddington, Eric <address@hidden> wrote:
> > > > I don't understand what you're doing here.
> > > >
> > > > You've completely removed EIJMP from gcc.
> > >
> > > That is correct.  It is better for gcc not to use this instruction.
> >
> > This prevent using indirect function call in bootloader code for >128KB
> > FLASH devices.
>
> Well as it stands now, gcc emits eijmp and eicall, but never sets
> EIND, so if EIND is non-zero, then the call will most likely fail.

avr-libc/crt1/gcrt1.S :

.section .init0,"ax",@progbits
        .weak   __init
;       .func   __init
__init:
.......
#ifdef __AVR_3_BYTE_PC__
        ldi     r16, hh8(pm(__vectors))
        out     _SFR_IO_ADDR(EIND), r16
#endif  /* __AVR_3_BYTE_PC__ */

This code set EIND to third byte of '__vectors' address.

1. This code is not clear, it is better use start address of *(.trampolines) 
section instead of '__vectors' .
2. More logical to move this code from crt1/gcrt1.S to 
gcc\config\avr\libgcc.S. Btcause the GCC uses EIND internally, then it must 
initialize its.

Anatoly.