Omnibus PCI commit:
o Save and restore bars for suspend/resume as well as for D3->D0
transitions.
o preallocate resources that the PCI devices use to avoid resource
conflicts
o lazy allocati

Omnibus PCI commit:
o Save and restore bars for suspend/resume as well as for D3->D0
transitions.
o preallocate resources that the PCI devices use to avoid resource
conflicts
o lazy allocation of resources not allocated by the BIOS.
o set unattached drivers to state D3. Set power state to D0
before probe/attach. Right now there's two special cases
for this (display and memory devices) that need work in other
areas of the tree.

Please report any bugs to me.

show more ...

It appears that the changes in the resources allocated is causing much
pain and suffering. Attempt to back it out by removing the 'if the
requested range is larger than the window, clip to the windo

It appears that the changes in the resources allocated is causing much
pain and suffering. Attempt to back it out by removing the 'if the
requested range is larger than the window, clip to the window' code.
This is a band-aide until the issues are better understood and the
issues with the lazy allocation patches are resolved.

show more ...

Add support for subtractive decoding bridges. These bridges pass all
signals to addresses to the child busses. Typically, ProgIf of 1
means a subtractive bridge. However, Intel has a whole lot of

Add support for subtractive decoding bridges. These bridges pass all
signals to addresses to the child busses. Typically, ProgIf of 1
means a subtractive bridge. However, Intel has a whole lot of ones
with a ProgIf of 80 that are also subtractive. We cope with these
bridges too. This eliminates hw.pci.allow_unsupported_io_range
because that had almost the same effect as these patches (almost means
'buggy'). Remove the bogus checks for ISA bus locations: these cycles
aren't special and are only passed by transparent bridges.

We allow any range to succeed. If the range is a superset of the
range that's decoded, trim the resource to that range. Otherwise,
pass the range unchanged. This will change the location that PC Card
and CardBus cards are attached. This might bogusly cause some
overlapping allocation that wasn't present before, but the overlapping
fixes need to be in the pci level.

There's also a few formatting changes here.

show more ...

Change all SYSCTLS which are readonly and have a related TUNABLE
from CTLFLAG_RD to CTLFLAG_RDTUN so that sysctl(8) can provide
more useful error messages.

Use __FBSDID().
Also some minor style cleanups.

Prefer new location of pci include files (which have only been in the
tree for two or more years now), except in a few places where there's
code to be compatible with older versions of FreeBSD.

Prefer the uintXX_t to the u_intXX_t names.

Calculate routed interrupts using the slot number from the device and
not that of the bridge.

Approved by: re (jhb)

Make the base pcib_route_interrupt method available to other pci-pci bridge
sub-classes.

This allows the powerpc kernel to build again.

Forgotten by: benno
Spotted by: grehan

Properly account for prefetchable memory when a request is being made.
We allow the request to go through if it matches either a prefetchable
or a non-prefetchable part of the bridge. We do not chec

Properly account for prefetchable memory when a request is being made.
We allow the request to go through if it matches either a prefetchable
or a non-prefetchable part of the bridge. We do not check to make
sure it is the right kind of memory because most drivers to not yet
properly set RF_PREFETCHABLE (only cardbus seems to do so, and I'm not
entirely sure it does it right). RF_PREFETCHABLE was invented for
cardbus, so hasn't been properly documented yet.

This is still overridable by hw.pci.allow_unsupported_io_ranges, but
the need for that is greatly reduced, especially for the nvida driver.

Approved by: re
Reviewed by: jhb and many testers
Submitted by: Matt Emmerton (although this has been reworked somewhat)

show more ...

Fix a couple of bugs in host_pcib_get_busno():
- If a PCI device is not present, then a 32-bit read_config() is going to
return 0xffffffff not 0xffff.
- For the 82454NX chipset, the MIOC that we re

Fix a couple of bugs in host_pcib_get_busno():
- If a PCI device is not present, then a 32-bit read_config() is going to
return 0xffffffff not 0xffff.
- For the 82454NX chipset, the MIOC that we read the bus numbers of the
various host-PCI bridges from is at function (slot) 0x10 not 0x0.

Approved by: re (rwatson)

show more ...

Add a function host_pcib_get_bnsno() that attempts to determine the bus
number of the child bus of a host to PCI bridge by reading from proprietary
configuration registers in the host to PCI bridge d

Add a function host_pcib_get_bnsno() that attempts to determine the bus
number of the child bus of a host to PCI bridge by reading from proprietary
configuration registers in the host to PCI bridge devices.

Approved by: re

show more ...

Unbreak kernel build.

Submitted by: Hiten Pandya <hiten@angelica.unixdaemons.com>

Convert kernel compile option PCI_ALLOW_UNSUPPORTED_IO_RANGE to
a loader tunable hw.pci.allow_unsupported_io_range.

Submitted by: Hiten Pandya <hiten@angelica.unixdaemons.com>
Approved by: re (mur

Convert kernel compile option PCI_ALLOW_UNSUPPORTED_IO_RANGE to
a loader tunable hw.pci.allow_unsupported_io_range.

Submitted by: Hiten Pandya <hiten@angelica.unixdaemons.com>
Approved by: re (murray)

show more ...

Export pcib_attach() as a "protected" for use in subclasses of the PCI-PCI
bridge driver.

Prefer the physical bus number of the PCI bus as the unit of the pciX
device created.

Make the printf messages when routing interrupts more consistent in the
various PCI bridge drivers.

Export a few symbols as globals to allow subclassing of this driver. In
OOP speak, you would mark these as 'protected' members. Specifically:
- Make the pcib_softc struct public so it can be used b

Export a few symbols as globals to allow subclassing of this driver. In
OOP speak, you would mark these as 'protected' members. Specifically:
- Make the pcib_softc struct public so it can be used by subclasses.
- Make pcib_{read,write}_ivar(), pcib_alloc_resource(), pcib_maxslots(),
and pcib_{read,write}_config() globals that can be used by subclasses.
- Make the pcib devclass a global variable.
- Move most of the pcib_attach() function into a global
pcib_attach_common() function that can be called by the attach routines
of subclasses.

Tested on: i386, alpha, sparc64, ia64

show more ...

Fix an edge case wrt membase, but more changes needed

Fix warnings introduced in the PCI_ALLOW_UNSUPPORTED_IO_RANGE case.

Allow PCI_ALLOW_UNSUPPORTED_IO_RANGE to leave broken setups broken enough
to work.

Put the stard/end adjustments back. They are needed. Also make start
== 0 a special case. I hope this fixes the real problem that phk and
others were seeing.

Don't claim to have routed an interrupt when the method actually returned an
error.

Make unsupported memory range message bootverbose only

Remove bogus range restrictions that attempted to restrict the range
of I/O in 1.5. It looks like I got it right only for some of the
cases. Instead, allow ISA addresses as a special case. Most PC

Remove bogus range restrictions that attempted to restrict the range
of I/O in 1.5. It looks like I got it right only for some of the
cases. Instead, allow ISA addresses as a special case. Most PCI
bridges decode this range. I need to investigate PCI bridges better
to know if this is always true or not, but for now assume that it is
since that seems to be the most common case.

# We need to allocate addresses better for the pccard stuff...

Submitted by: phk, mitsunaga-san

show more ...