From mboxrd@z Thu Jan  1 00:00:00 1970
From: Jesper Skov <jskov@cambridge.redhat.com>
To: "james chen" <james_ch1@sina.com>
Cc: <ecos-discuss@sources.redhat.com>
Subject: Re: [ECOS] Debug Error
Date: Tue, 20 Mar 2001 23:42:00 -0000
Message-id: <otzoefsh4s.fsf@thinktwice.zoftcorp.dk>
References: <001b01c0b1cb$9c7eb180$c408aa0a@inc.inventec>
X-SW-Source: 2001-03/msg00301.html

>>>>> "james" == james chen <james_ch1@sina.com> writes:

james>        When I debug eCos example between insight5.0 and
james> redboot, when goes into hal_if_init(hal_if.c). it will response
james> nothing(that is say redboot doesn't send packages back to
james> insight).

hal_if_init is a dead end debugging wise - it makes changes to the
channels used by the GDB stub for communication.

james>         When redboot already run in ROM, does I need to re-call
james> hal_if_init function in my hal_hardware_init function? 

Yes, you need to call it - but in this case it will be safe. It all
comes down to the configuration. What's probably happening is that
CYGSEM_HAL_VIRTUAL_VECTOR_INIT_WHOLE_TABLE is set in your RAM
application.

If you do indeed have redboot in ROM, you can safely disable that
option. And then you should be able to step through the hal_if_init
code.


I'll add some comments to the code to warn about this.

Jesper