cvs commit: src/sys/dev/pci vga_pci.c

[John Baldwin]

On Tuesday 20 December 2005 05:47 pm, Nate Lawson wrote:
> John Baldwin wrote:
> > jhb         2005-12-20 22:41:33 UTC
> >
> >   FreeBSD src repository
> >
> >   Added files:
> >     sys/dev/pci          vga_pci.c
> >   Log:
> >   Add a vgapci(4) stub device driver for VGA PCI devices.  This device
> > serves as a bus so that other drivers such as drm(4), acpi_video(4), and
> > agp(4) can attach to it thus allowing multiple drivers for the same
> > device.  It also removes the need for the drmsub hack for the
> > i8[13]0/i915 drm and agp drivers.
> >
> >   Revision  Changes    Path
> >   1.3       +262 -0    src/sys/dev/pci/vga_pci.c (new)
>
> All the methods seem to just be a straight-through mapping or a debug
> print + mapping.  For the ones you don't plan to do anything special in,
> can you just make the methods map directly to the generic versions?

The ones that do map, yes.  I was mostly paranoid to see if functions whose 
stubs weren't really right were getting called or not.  There is room for 
cleaning.

> > +static int
> > +vga_pci_suspend(device_t dev)
> > +{
> > +
> > +	return (bus_generic_suspend(dev));
> > +}
> > +
> > +static int
> > +vga_pci_resume(device_t dev)
> > +{
> > +
> > +	return (bus_generic_resume(dev));
> > +}

These are generics and could just use bus_generic_foo().

> > +
> > +/* Bus interface. */
> > +
> > +static int
> > +vga_pci_read_ivar(device_t dev, device_t child, int which, uintptr_t
> > *result) +{
> > +
> > +	return (BUS_READ_IVAR(device_get_parent(dev), dev, which, result));
> > +}
> > +
> > +static int
> > +vga_pci_write_ivar(device_t dev, device_t child, int which, uintptr_t
> > value) +{
> > +
> > +	return (EINVAL);
> > +}

These are not generics.  The generic would pass child as the second arg rather 
than dev.

> > +static struct resource *
> > +vga_pci_alloc_resource(device_t dev, device_t child, int type, int *rid,
> > +    u_long start, u_long end, u_long count, u_int flags)
> > +{
> > +
> > +	return (bus_alloc_resource(dev, type, rid, start, end, count, flags));
> > +}
> > +
> > +static int
> > +vga_pci_release_resource(device_t dev, device_t child, int type, int
> > rid, +    struct resource *r)
> > +{
> > +
> > +	return (bus_release_resource(dev, type, rid, r));
> > +}

Same with these two.

> > +
> > +/* PCI interface. */
> > +
> > +static uint32_t
> > +vga_pci_read_config(device_t dev, device_t child, int reg, int width)
> > +{
> > +
> > +	return (pci_read_config(dev, reg, width));
> > +}
> > +
> > +static void
> > +vga_pci_write_config(device_t dev, device_t child, int reg,
> > +    uint32_t val, int width)
> > +{
> > +
> > +	pci_write_config(dev, reg, val, width);
> > +}

And all of the PCI ones including these and the others you quoted.  The 
different is subtle, but important.

-- 
John Baldwin <jhb at FreeBSD.org>  <><  http://www.FreeBSD.org/~jhb/
"Power Users Use the Power to Serve"  =  http://www.FreeBSD.org