From mboxrd@z Thu Jan  1 00:00:00 1970
From: Jesper Skov <jskov@redhat.com>
To: Jonathan Larmour <jlarmour@redhat.com>
Cc: ness <root@t2-design.com>, ecos-discuss@sourceware.cygnus.com
Subject: Re: [ECOS] Problem about redboot
Date: Wed, 25 Oct 2000 23:52:00 -0000
Message-id: <otn1fs14dy.fsf@thinktwice.zoftcorp.dk>
References: <Pine.LNX.4.21.0010260855130.762-100000@localhost.localdomain> <39F79129.AE97222C@redhat.com>
X-SW-Source: 2000-10/msg00308.html

>>>>> "Jonathan" == Jonathan Larmour <jlarmour@redhat.com> writes:
Jonathan> The vrc4373 doesn't implement the "standard" virtual vector
Jonathan> support. What you are seeing is an oversight that the
Jonathan> hal_ctrl_isr* functions in hal/common were not sufficiently
Jonathan> conditionally compiled, which I will fix.  [ Jesper, I will
Jonathan> make these functions conditional on
Jonathan> CYGSEM_HAL_VIRTUAL_VECTOR_SUPPORT ]

But, but, but, that shouldn't be necessary... :)

        cdl_component CYGSEM_HAL_VIRTUAL_VECTOR_SUPPORT {
            display     "Enable use of virtual vector calling interface"
            active_if   CYGINT_HAL_VIRTUAL_VECTOR_SUPPORT
            calculated  1
            description "
                Virtual vector support allows the HAL to let the ROM
                monitor handle certain operations. The virtual vector table
                defines a calling interface between applications running in
                RAM and the ROM monitor."
            compile     hal_if.c hal_misc.c
        }

The file is only ever compiled when there is virtual vector
support. The problem is the fact that the vrc4373 is not converted
properly/at all and the virtual vector support was still attempted
used.

Jesper