Date: Tue, 23 Aug 2016 09:20:18 -0600 From: Warner Losh <imp@bsdimp.com> To: Ravi Pokala <rpokala@mac.com> Cc: "freebsd-hackers@freebsd.org" <freebsd-hackers@freebsd.org>, karu.pruun@gmail.com Subject: Re: problem attaching driver at LPC bus Message-ID: <CANCZdfpzvvsA8JTLiKZAjyWjifFMzJN3KZ76HWW=%2BYnH326w3w@mail.gmail.com> In-Reply-To: <20A27669-0B16-4199-853F-46D84E876AE9@panasas.com> References: <20A27669-0B16-4199-853F-46D84E876AE9@panasas.com>
next in thread | previous in thread | raw e-mail | index | archive | help
On Tue, Aug 23, 2016 at 9:13 AM, Ravi Pokala <rpokala@mac.com> wrote:
> Hi Peeter,
>
> I wrote a driver for a pair of LPC-attached GPIO controllers last year, s=
o this is (somewhat) in my head. In my case, ACPI didn't actually have the =
correct information for either controller.
>
> For the controller that was embedded in the chipset, I got the address in=
fo from the parent ISA bridge. In my case, that involved reading and maskin=
g various registers, as defined by the chipset specs. I also got the resour=
ce size from the specs. For the standalone controller attached to the LPC b=
us, I had to do something similar - inb() the address registers, mask accor=
dingly, and set up the resource. During their respective probe()s:
>
> rc =3D bus_set_resource(dev, SYS_RES_IOPORT, sc->gp_nextrid, XXX_=
base,
> XXX_RESOURCE_BYTES);
> if (rc !=3D 0) {
> device_printf(dev, " probe failed to setup resource: %d\n=
", rc);
> return ENXIO;
> }
> /* Because we needed multiple controllers to be connected to the =
same
> * devnode, increment to ensure they get their own RID
> */
> sc->gp_XXX_ctlr.gpio_ctlr_rid =3D sc->gp_nextrid;
> sc->gp_nextrid++;
>
>
> Then, during attach():
>
> rid =3D sc->gp_XXX_ctlr.gpio_ctlr_rid;
> /* Allocate resource setup in probe method */
> res =3D bus_alloc_resource_any(dev, SYS_RES_IOPORT, &rid, RF_ACTI=
VE);
> if (res =3D=3D NULL) {
> /* Failed to allocate I/O space resource */
> device_printf(dev, "Failed to allocate resource for RID %=
d\n", rid);
> return ENOMEM;
> }
> sc->gp_XXX_ctlr.gpio_ctlr_res =3D res;
> device_printf(dev, "Attached: %s @ 0x%08lx\n", cfg->gc_desc, rman=
_get_start(res));
>
> One thing to note is that I was careful about keeping track of the RIDs. =
Several of the existing drivers in the tree seem to just use 0 indiscrimina=
tely, and it works because they only use one resource.
For ISA drivers, RID is just a number, best thought of as an index.
So incrementing here like you've done is the right call.
Warner
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?CANCZdfpzvvsA8JTLiKZAjyWjifFMzJN3KZ76HWW=%2BYnH326w3w>