This is the mail archive of the
ecos-devel@sources.redhat.com
mailing list for the eCos project.
AW: AW: contributing a failsafe update meachanism for FIS from within ecos applications
- From: "Neundorf, Alexander" <Alexander dot Neundorf at jenoptik dot com>
- To: "Andrew Lunn" <andrew at lunn dot ch>
- Cc: "Slawek" <sgp at telsatgp dot com dot pl>, <ecos-devel at sources dot redhat dot com>
- Date: Thu, 28 Oct 2004 10:56:08 +0200
- Subject: AW: AW: contributing a failsafe update meachanism for FIS from within ecos applications
> Von: Andrew Lunn [mailto:andrew@lunn.ch]
...
> > attached. It would be nice if the fisfs-implementation could get
> > the information about the two flash blocks reserved for the two fis
> > tables somehow from redboot during runtime. Is there a way to do
> > this ?
>
> You should take a look what the virtual vectors supports. There are a
> couple of examples of it being used in:
>
> packages/io/flash/current/src/flashiodev.c:flashiodev_init() and see
> packages/hal/common/current/src/hal_if.c:flash_fis_op().
>
> As i said before, i think this is the way the FISFS should work,
> redboot doing most of the work using VV as the interface between the
> application and redboot.
>
> Andrew
Ok, I'll have a look.
What would you say needs to be transferred to redboot ?
For implementing an ecos fisfs I need a bunch of functions.
open(), read(), close(), opendir(), readdir() are more or less trivial I'd say and transferring these functions to redboot wouldn't make much sense.
I guess what would make sense would be at least the mount() (i.e. finding the valid fis table and returning a pointer to it) call, and probably the eraseImage() and createImage() functions.
What do you think ?
Bye
Alex