Date: Mon, 29 Jan 2001 14:00:43 GMT From: Brian Somers <brian@Awfulhak.org> To: Poul-Henning Kamp <phk@critter.freebsd.dk> Cc: Brian Somers <brian@Awfulhak.org>, freebsd-arch@FreeBSD.org Subject: Re: Cloned open support Message-ID: <200101291400.f0TE0hY61432@storm.FreeBSD.org.uk> In-Reply-To: Your message of "Mon, 29 Jan 2001 13:45:01 %2B0100" <20820.980772301@critter>
next in thread | previous in thread | raw e-mail | index | archive | help
> >The thing is that I believe the driver needs to be able to
> >decide what minor number will be opened. This idea is taken
> >from the Solaris implementation. For example, I have a
> >theoretical driver that does this:
>
> I'm not arguing that the driver should not do the decisionmaking,
> only that it should not be done in the d_open routine.
>
> By the time you get to the d_open, you have already committed,
> and possibly reused or even shared a vnode with somebody else.
I was thinking along the lines of just VN_RELE'ing it and making
up another at this point.
> If you do it at namei time, you can create the right vnode
> before you go (further) down the open path.
>
> That is how our cloning works:
>
> When devfs doesn't find what you are looking for, it will
> poll all drivers and ask if they can create the dev_t for it.
>
> If one of the drivers returns a dev_t, a devfs vnode is
> instantiated and everybody are happy.
>
> >o Makes minor 0 available in /dev as a control device
>
> No worries:
>
> make_dev(..., minor = 0, "/dev/foo");
>
> >o Makes minor 1 available as something that'll be cloned.
>
> Nope, decide on a name which is cloned; and recognize it
> in foo_clone(dev_t *dev, ...) Allocate the next free minor and
>
> if (strcmp(name, "fooclone"))
> return;
> ... /* find free minor */
> *dev = make_dev(..., minor = n, "dev/foo%03d", n);
> return;
>
> This should work as far as I know. Userland needs to find the
> right pathname, preferably without too much magic. fstat(2) +
> devname(3) is probably doing that just fine for that.
Hmm, I'm not sure I like the idea of a user being able to open
things they can't see beforehand (permissions seem to be one
problem)... My thoughts were along the lines of doing the
make_dev(... minor=1, "dev/fooclone" ...) in foo_attach() as
we do now, then tweaking the dev_t in foo_open().
The namei() bit is necessary to get to the point that you
know you're talking about a device and can call any foo_clone()
etc.
So, how about this then (or is this what you're already saying):
make_dev() is called during attach as normal.
devfs's vfs lookup routine checks if the cdevsw of the thing it
finds has D_CLONE, and if so asks the owning driver via a new
foo_isclone() cdevsw function if it'd like to change the dev_t
it's about to give back to the namei side of things.
Would that work ? I haven't looked at the devfs stuff, but I
wouldn't imagine there would be any problems doing this...
> --
> Poul-Henning Kamp | UNIX since Zilog Zeus 3.20
> phk@FreeBSD.ORG | TCP/IP since RFC 956
> FreeBSD committer | BSD since 4.3-tahoe
> Never attribute to malice what can adequately be explained by incompetence.
--
Brian <brian@Awfulhak.org> <brian@[uk.]FreeBSD.org>
<http://www.Awfulhak.org>; <brian@[uk.]OpenBSD.org>
Don't _EVER_ lose your sense of humour !
To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-arch" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?200101291400.f0TE0hY61432>