This is the mail archive of the
ecos-discuss@sources.redhat.com
mailing list for the eCos project.
Re: Still confused by virtual vectors
- To: Jonathan Larmour <jlarmour at redhat dot com>
- Subject: Re: [ECOS] Still confused by virtual vectors
- From: Grant Edwards <grante at visi dot com>
- Date: Wed, 3 Oct 2001 18:19:48 -0500
- Cc: ecos-discuss at sources dot redhat dot com
- References: <20011003163242.A12778@visi.com> <3BBB998B.431448C3@redhat.com>
On Thu, Oct 04, 2001 at 12:04:43AM +0100, Jonathan Larmour wrote:
> Grant Edwards wrote:
> >
> > I'm trying to update my HAL to work with the current CVS
> > snapshot and to use virtual vectors the "right" way. I've read
> > and re-read porting/hal-calling-if.html, but I just don't get
> > it.
> >
> > I've looked at a couple different HAL implimentations and they
> > don't all seem to support virtual vectors (or lack thereof) can
> > somebody point me to a good HAL example?
>
> Most do though. Since you're on ARM, choosing another ARM target is good.
> How about the ebsa285 since it's well established. Look at hal_diag.c but
> ignore the "compatibility with older stubs" bits. And then hal_diag.h and
> plf_stub.h. edb7xxx is another option.
>
> When you have it going, if you have any (preferably specific :-)) changes
> to the document we'd be happy to take them. That goes for anyone! We know
> it's still incomplete.
One thing that I'm still figuring out is what configuration
variables control various VV features and and whether or not my
HAL is supposed to pay attention to that variable. A few of
them are listed in the docs, but others aren't. For example:
CYGSEM_HAL_VIRTUAL_VECTOR_CLAIM_COMMS
determines whether the common HAL code will call the platform
HAL routine cyg_hal_plf_comms_init() to fill in the
communication channel functions.
[still figuring out some of the others...]
--
Grant Edwards
grante@visi.com