MFC r258779,r258780,r258787,r258822:

Fix undefined behavior: (1 << 31) is not defined as 1 is an int and this
shifts into the sign bit. Instead use (1U << 31) which gets the
expected result.

Similar to the (1 << 31) case it is not defined to do (2 << 30).

This fix is not ideal as it assumes a 32 bit int, but does fix the issue
for most cases.

A similar change was made in OpenBSD.

Copy head (r256279) to stable/10 as part of the 10.0-RELEASE cycle.

Approved by: re (implicit)
Sponsored by: The FreeBSD Foundation

Replace kernel virtual address space allocation with vmem. This provides
transparent layering and better fragmentation.

- Normalize functions that allocate memory to use kmem_*
- Those that allocate address space are named kva_*
- Those that operate on maps are named kmap_*
- Implement recursive allocation handling for kmem_arena in vmem.

Reviewed by: alc
Tested by: pho
Sponsored by: EMC / Isilon Storage Division

Remove duplicate code. Reduce diff between amd64 and i386.

Use volatile keywords properly.

Tidy up inline assembly. No functional change.

pciereg_cfg*: use assembly to access the mem-mapped cfg space

AMD BKDG for CPU families 10h and later requires that the memory
mapped config is always read into or written from al/ax/eax register.

Discussed with: kib, alc
Reviewed by: kib (earlier version)
MFC after: 25 days

Move {amd64,i386}/pci/pci_bus.c and {amd64,i386}/include/pci_cfgreg.h to
the x86 tree. The $PIR code is still only enabled on i386 and not amd64.
While here, make the qpi(4) driver on conditional on 'device pci'.

Use uintXX_t instead of u_intXX_t.

Add a helper routine to conditionally modify the start address of a
resource allocation from an x86 Host-PCI bridge driver so that it can be
reused by the ACPI Host-PCI bridge driver (and eventually the MPTable
Host-PCI bridge driver) instead of duplicating the same logic. Note that
this means that hw.acpi.host_mem_start is now replaced with the
hw.pci.host_mem_start tunable that was already used in the non-ACPI case.
This also removes hw.acpi.host_mem_start on ia64 where it was not
applicable (the implementation was very x86-specific).

While here, adjust the logic to apply the new start address on any
"wildcard" allocation even if that allocation comes from a subset of
the allowable address range.

Reviewed by: imp (1)

On multi-core, multi-threaded PPC systems, it is important that the threads
be brought up in the order they are enumerated in the device tree (in
particular, that thread 0 on each core be brought up first). The SLIST
through which we loop to start the CPUs has all of its entries added with
SLIST_INSERT_HEAD(), which means it is in reverse order of enumeration
and so AP startup would always fail in such situations (causing a machine
check or RTAS failure). Fix this by changing the SLIST into an STAILQ,
and inserting new CPUs at the end.

Reviewed by: jhb

Reimplement how PCI-PCI bridges manage their I/O windows. Previously the
driver would verify that requests for child devices were confined to any
existing I/O windows, but the driver relied on the firmware to initialize
the windows and would never grow the windows for new requests. Now the
driver actively manages the I/O windows.

This is implemented by allocating a bus resource for each I/O window from
the parent PCI bus and suballocating that resource to child devices. The
suballocations are managed by creating an rman for each I/O window. The
suballocated resources are mapped by passing the bus_activate_resource()
call up to the parent PCI bus. Windows are grown when needed by using
bus_adjust_resource() to adjust the resource allocated from the parent PCI
bus. If the adjust request succeeds, the window is adjusted and the
suballocation request for the child device is retried.

When growing a window, the rman_first_free_region() and
rman_last_free_region() routines are used to determine if the front or
end of the existing I/O window is free. From using that, the smallest
ranges that need to be added to either the front or back of the window
are computed. The driver will first try to grow the window in whichever
direction requires the smallest growth first followed by the other
direction if that fails.

Subtractive bridges will first attempt to satisfy requests for child
resources from I/O windows (including attempts to grow the windows). If
that fails, the request is passed up to the parent PCI bus directly
however.

The PCI-PCI bridge driver will try to use firmware-assigned ranges for
child BARs first and only allocate a "fresh" range if that specific range
cannot be accommodated in the I/O window. This allows systems where the
firmware assigns resources during boot but later wipes the I/O windows
(some ACPI BIOSen are known to do this) to "rediscover" the original I/O
window ranges.

The ACPI Host-PCI bridge driver has been adjusted to correctly honor
hw.acpi.host_mem_start and the I/O port equivalent when a PCI-PCI bridge
makes a wildcard request for an I/O window range.

The new PCI-PCI bridge driver is only enabled if the NEW_PCIB kernel option
is enabled. This is a transition aide to allow platforms that do not
yet support bus_activate_resource() and bus_adjust_resource() in their
Host-PCI bridge drivers (and possibly other drivers as needed) to use the
old driver for now. Once all platforms support the new driver, the
kernel option and old driver will be removed.

PR: kern/143874 kern/149306
Tested by: mav

Add implementations of BUS_ADJUST_RESOURCE() to the PCI bus driver,
generic PCI-PCI bridge driver, x86 nexus driver, and x86 Host to PCI bridge
drivers.

Add device IDs for two more ServerWorks Host-PCI bridges so that we can
read their starting PCI bus number for older systems that do not support
ACPI (or have a broken _BBN method).

PR: kern/148108
MFC after: 1 week

Change the MPTable and $PIR PCI-PCI bridge drivers to inherit from the
generic PCI-PCI bridge driver and only override specific methods. This
should fix suspend/resume of PCI-PCI bridges using these drivers.

number of cleanups in i386 and amd64 pci md code

o introduce PCIE_REGMAX and use it instead of ad-hoc constant
o where 'reg' parameter/variable is not already unsigned, cast it to
unsigned before comparison with maximum value to cut off negative
values
o use PCI_SLOTMAX in several places where 31 or 32 were explicitly used
o drop redundant check of 'bytes' in i386 pciereg_cfgread() - valid
values are already checked in the subsequent switch

Reviewed by: jhb
MFC after: 1 week

Match PCI Express root bridge _HID directly instead of
relying on _CID.

Reviewed by: jhb
Approved by: re (kib)

strict kobj signatures: fix legacy i386 pcib_write_config impl

Reviewed by: imp, current@
Approved by: jhb (mentor)

Add a read-only sysctl hw.pci.mcfg to mirror the tunable by the same name.

MFC after: 1 week

Fall back to using configuration type 1 accesses for PCI config requests if
the requested PCI bus falls outside of the bus range given in the ACPI
MCFG table. Several BIOSes seem to not include all of the PCI busses in
systems in their MCFG tables. It maybe that the BIOS is simply buggy and
does support all the busses, but it is more conservative to just fall back
to the old method unless it is certain that memory accesses will work.

Add a 'hw.pci.mcfg' tunable. It can be set to 0 to disable memory-mapped
PCI config access.

Some K8 chipsets don't expose all of the PCI devices on bus 0 via PCIe
memory-mapped config access. Add a workaround for these systems by
checking the first function of each slot on bus 0 using both the
memory-mapped config access and the older type 1 I/O port config access.
If we find a slot that is only visible via the type 1 I/O port config
access, we flag that slot. Future PCI config transactions to flagged
slots on bus 0 use type 1 I/O port config access rather than memory mapped
config access.

Extend the support for PCI-e memory mapped configuration space access:
- Rename pciereg_cfgopen() to pcie_cfgregopen() and expose it to the
rest of the kernel. It now also accepts parameters via function
arguments rather than global variables.
- Add a notion of minimum and maximum bus numbers and reject requests for
an out of range bus.
- Add more range checks on slot/func/reg/bytes parameters to the cfg reg
read/write routines. Don't panic on any invalid parameters, just fail
the request (writes do nothing, reads return -1). This matches the
behavior of the other cfg mechanisms.
- Port the memory mapped configuration space access to amd64. On amd64
we simply use the direct map (via pmap_mapdev()) for the memory mapped
window.
- During acpi_attach() just after loading the ACPI tables, check for a
MCFG table. If it exists, call pciereg_cfgopen() on each subtable
(memory mapped window). For now we only support windows for domain 0
that start with bus 0. This removes the need for more chipset-specific
quirks in the MD code.
- Remove the chipset-specific quirks for the Intel 5000P/V/Z chipsets
since these machines should all have MCFG tables via ACPI.
- Updated pci_cfgregopen() to DTRT if ACPI had invoked pcie_cfgregopen()
earlier.

MFC after: 2 weeks

- Add support for memory mapped PCI config space access on Intel 915GM
and 5000P/V/Z chipsets.
- If the base address of the config space BAR is above 4GB for some reason
and this isn't a PAE kernel, then warn about this (under bootverbose)
and don't use the BAR.

PR: kern/126525
Submitted by: Arthur Hartwig @ Nokia
MFC after: 2 weeks

Use switch statements instead of if-else for enabling PCI-express config
space support.

MFC after: 1 week

Integrate support for xen in to i386 common code.

MFC after: 1 month

More properly handle links who only have 1 valid IRQ in their bitmask. The
old code special cased them too early which caused a few differences for
these sort of links relative to other PCI links:

- They were always re-routed via the BIOS call instead of assuming that
they were already routed if the BIOS had programmed the IRQ into a
matching device during POST.
- If the BIOS did route that link to a different IRQ that was marked as
invalid, we trusted the $PIR table rather than the BIOS IRQ.

This change moves the special casing for "unique IRQ" links to only take
that into account when picking an IRQ for an unrouted link so that these
links will now not be routed if the BIOS appears to have routed it already
(some BIOSen have problems with that) and so that if the BIOS uses a
different IRQ than the $PIR, we trust the BIOS routing instead (this is
what we do for all other links as well).

Reported by: Bruce Walter walter of fortean com
MFC after: 1 week

MFamd64: 1.109 of pci_cfgreg.c which changes pci_cfgdisable() into a nop
for type #1 similar to what other OS's do.

MFC after: 3 days

Adjust the code to probe for the PCI config mechanism to use.
- On amd64, just assume type #1 is always used. PCI 2.0 mandated
deprecated type #2 and required type #1 for all future bridges which
was well before amd64 existed.
- For i386, ignore whatever value was in 0xcf8 before testing for type #1
and instead rely on the other tests to determine if type #1 works. Some
newer machines leave garbage in 0xcf8 during boot and as a result the
kernel doesn't find PCI at all (which greatly confuses ACPI which expects
PCI to exist when PCI busses are in the namespace).

MFC after: 3 days
Discussed with: scottl

Make the PCI code aware of PCI domains (aka PCI segments) so we can
support machines having multiple independently numbered PCI domains
and don't support reenumeration without ambiguity amongst the
devices as seen by the OS and represented by PCI location strings.
This includes introducing a function pci_find_dbsf(9) which works
like pci_find_bsf(9) but additionally takes a domain number argument
and limiting pci_find_bsf(9) to only search devices in domain 0 (the
only domain in single-domain systems). Bge(4) and ofw_pcibus(4) are
changed to use pci_find_dbsf(9) instead of pci_find_bsf(9) in order
to no longer report false positives when searching for siblings and
dupe devices in the same domain respectively.
Along with this change the sole host-PCI bridge driver converted to
actually make use of PCI domain support is uninorth(4), the others
continue to use domain 0 only for now and need to be converted as
appropriate later on.
Note that this means that the format of the location strings as used
by pciconf(8) has been changed and that consumers of <sys/pciio.h>
potentially need to be recompiled.

Suggested by: jhb
Reviewed by: grehan, jhb, marcel
Approved by: re (kensmith), jhb (PCI maintainer hat)

Revamp the MSI/MSI-X code a bit to achieve two main goals:
- Simplify the amount of work that has be done for each architecture by
pushing more of the truly MI code down into the PCI bus driver.
- Don't bind MSI-X indicies to IRQs so that we can allow a driver to map
multiple MSI-X messages into a single IRQ when handling a message
shortage.

The changes include:
- Add a new pcib_if method: PCIB_MAP_MSI() which is called by the PCI bus
to calculate the address and data values for a given MSI/MSI-X IRQ.
The x86 nexus drivers map this into a call to a new 'msi_map()' function
in msi.c that does the mapping.
- Retire the pcib_if method PCIB_REMAP_MSIX() and remove the 'index'
parameter from PCIB_ALLOC_MSIX(). MD code no longer has any knowledge
of the MSI-X index for a given MSI-X IRQ.
- The PCI bus driver now stores more MSI-X state in a child's ivars.
Specifically, it now stores an array of IRQs (called "message vectors" in
the code) that have associated address and data values, and a small
virtual version of the MSI-X table that specifies the message vector
that a given MSI-X table entry uses. Sparse mappings are permitted in
the virtual table.
- The PCI bus driver now configures the MSI and MSI-X address/data
registers directly via custom bus_setup_intr() and bus_teardown_intr()
methods. pci_setup_intr() invokes PCIB_MAP_MSI() to determine the
address and data values for a given message as needed. The MD code
no longer has to call back down into the PCI bus code to set these
values from the nexus' bus_setup_intr() handler.
- The PCI bus code provides a callout (pci_remap_msi_irq()) that the MD
code can call to force the PCI bus to re-invoke PCIB_MAP_MSI() to get
new values of the address and data fields for a given IRQ. The x86
MSI code uses this when an MSI IRQ is moved to a different CPU, requiring
a new value of the 'address' field.
- The x86 MSI psuedo-driver loses a lot of code, and in fact the separate
MSI/MSI-X pseudo-PICs are collapsed down into a single MSI PIC driver
since the only remaining diff between the two is a substring in a
bootverbose printf.
- The PCI bus driver will now restore MSI-X state (including programming
entries in the MSI-X table) on device resume.
- The interface for pci_remap_msix() has changed. Instead of accepting
indices for the allocated vectors, it accepts a mini-virtual table
(with a new length parameter). This table is an array of u_ints, where
each value specifies which allocated message vector to use for the
corresponding MSI-X message. A vector of 0 forces a message to not
have an associated IRQ. The device may choose to only use some of the
IRQs assigned, in which case the unused IRQs must be at the "end" and
will be released back to the system. This allows a driver to use the
same remap table for different shortage values. For example, if a driver
wants 4 messages, it can use the same remap table (which only uses the
first two messages) for the cases when it only gets 2 or 3 messages and
in the latter case the PCI bus will release the 3rd IRQ back to the
system.

MFC after: 1 month

Expand the MSI/MSI-X API to address some deficiencies in the MSI-X support.
- First off, device drivers really do need to know if they are allocating
MSI or MSI-X messages. MSI requires allocating powerof2() messages for
example where MSI-X does not. To address this, split out the MSI-X
support from pci_msi_count() and pci_alloc_msi() into new driver-visible
functions pci_msix_count() and pci_alloc_msix(). As a result,
pci_msi_count() now just returns a count of the max supported MSI
messages for the device, and pci_alloc_msi() only tries to allocate MSI
messages. To get a count of the max supported MSI-X messages, use
pci_msix_count(). To allocate MSI-X messages, use pci_alloc_msix().
pci_release_msi() still handles both MSI and MSI-X messages, however.
As a result of this change, drivers using the existing API will only
use MSI messages and will no longer try to use MSI-X messages.
- Because MSI-X allows for each message to have its own data and address
values (and thus does not require all of the messages to have their
MD vectors allocated as a group), some devices allow for "sparse" use
of MSI-X message slots. For example, if a device supports 8 messages
but the OS is only able to allocate 2 messages, the device may make the
best use of 2 IRQs if it enables the messages at slots 1 and 4 rather
than default of using the first N slots (or indicies) at 1 and 2. To
support this, add a new pci_remap_msix() function that a driver may call
after a successful pci_alloc_msix() (but before allocating any of the
SYS_RES_IRQ resources) to allow the allocated IRQ resources to be
assigned to different message indices. For example, from the earlier
example, after pci_alloc_msix() returned a value of 2, the driver would
call pci_remap_msix() passing in array of integers { 1, 4 } as the
new message indices to use. The rid's for the SYS_RES_IRQ resources
will always match the message indices. Thus, after the call to
pci_remap_msix() the driver would be able to access the first message
in slot 1 at SYS_RES_IRQ rid 1, and the second message at slot 4 at
SYS_RES_IRQ rid 4. Note that the message slots/indices are 1-based
rather than 0-based so that they will always correspond to the rid
values (SYS_RES_IRQ rid 0 is reserved for the legacy INTx interrupt).
To support this API, a new PCIB_REMAP_MSIX() method was added to the
pcib interface to change the message index for a single IRQ.

Tested by: scottl

Give Host-PCI bridge drivers their own pcib_alloc_msi() and
pcib_alloc_msix() methods instead of using the method from the generic
PCI-PCI bridge driver as the PCI-PCI methods will be gaining some PCI-PCI
specific logic soon.

Replace a few magic numbers.

MD support for PCI Message Signalled Interrupts on amd64 and i386:
- Add a new apic_alloc_vectors() method to the local APIC support code
to allocate N contiguous IDT vectors (aligned on a M >= N boundary).
This function is used to allocate IDT vectors for a group of MSI
messages.
- Add MSI and MSI-X PICs. The PIC code here provides methods to manage
edge-triggered MSI messages as x86 interrupt sources. In addition to
the PIC methods, msi.c also includes methods to allocate and release
MSI and MSI-X messages. For x86, we allow for up to 128 different
MSI IRQs starting at IRQ 256 (IRQs 0-15 are reserved for ISA IRQs,
16-254 for APIC PCI IRQs, and IRQ 255 is reserved).
- Add pcib_(alloc|release)_msi[x]() methods to the MD x86 PCI bridge
drivers to bubble the request up to the nexus driver.
- Add pcib_(alloc|release)_msi[x]() methods to the x86 nexus drivers that
ask the MSI PIC code to allocate resources and IDT vectors.

MFC after: 2 months

Don't dump the $PIR table under bootverbose. The pirtool program in
src/tools/tools works fine, and dumping this table can add a lot of noise.

MFC after: 1 week

- Make pcib_devclass private to sys/dev/pci/pci_pci.c and change all the
various pcib drivers to use their own private devclass_t variables for
their modules.
- Use the DEFINE_CLASS_0() macro to declare drivers for the various pcib
drivers while I'm here.

Move the hostb driver out of the i386 and amd64 PCI code (where it was
duplicated anyways) and into a single MI driver. Extend the driver a bit
to implement the bus and PCI kobj interfaces such that other drivers can
attach to it and transparently act as if their parent device is the PCI
bus (for the most part).

Add support for 7320 and 915 PCIe chipsets.

Submitted by: Gavin Atkinson <gavin.atkinson at ury dot york dot ac dot uk>
PR: kern/79139
Reviewed by: scottl

Provide a dummy NO_XBOX option that lives in opt_xbox.h for pc98.
This allows us to eliminate a three ifdef PC98 instances.

Fix pc98 build.

Add support for XBOX to the FreeBSD port. The xbox architecture is
nearly identical to wintel/ia32, with a couple of tweaks. Since it is
so similar to ia32, it is optionally added to a i386 kernel. This
port is preliminary, but seems to work well. Further improvements
will improve the interaction with syscons(4), port Linux nforce driver
and future versions of the xbox.

This supports the 64MB and 128MB boxes. You'll need the most recent
CVS version of Cromwell (the Linux BIOS for the XBOX) to boot.

Rink will be maintaining this port, and is interested in feedback.
He's setup a website http://xbox-bsd.nl to report the latest
developments.

Any silly mistakes are my fault.

Submitted by: Rink P.W. Springer rink at stack dot nl and
Ed Schouten ed at fxq dot nl

MFamd64: indent with tabs instead of spaces.

Undo the change to pci_cfgdisable() on i386 for now. It seems to fix
the amd64 case, but makes the i386 case fail even more often.

Modify the pci_cfgdisable() routine to bring it more in line with
other OSes (Solaris, Linux, VxWorks). It's not necessary to write a 0
to the config address register when using config mechanism 1 to turn
off config access. In fact, it can be downright troublesome, since it
seems to confuse the PCI-PCI bridge in the AMD8111 chipset and cause
it to sporadically botch reads from some devices. This is the cause
of the missing USP ports problem I was experiencing with my Sun Opteron
system.

Also correct the case for mechanism 2: it's only necessary to write
a 0 to the ENABLE port.

Expose legacy_pcib_alloc_resource, and use it in the mptable pci bus
implementation, like other routines in the legacy bus.

This should fix problems with resource allocation on MP systems without
ACPI enabled.

Commit a workaround to a problem with resource allocation. This helps
with some Dell servers that booted w/o a problem[*] on 5.4, but failed
with 6.0-BETA.

On the PCI bus, when we do lazy resource allocation, we narrow the
range requested as we pass through bridges to reflect how the bridges
are programmed and what addresses they pass. However, when we're
doing an allocation on a bus that's directly connected to a host
bridge, no such translation can take place. We already had a fallback
range for memory requests, but none for ioports. As such, provide a
fallback for I/O ports so we don't allocate location 0, which will
have undesired side effects when the resources are actually used.

This fixes a problem with booting a Dell server with usb in the
kernel. However, it is an unsatisfying solution. I don't like the
hard coded value, and I think we should start narrowing the resources
returned to not be in the so-called isa alias area (where the ranage &
0x0300 must be 0 iirc). Doing such filtering will have to wait for
another day.

This may be a good 6 candidate, maybe after its had a chance to be
refined.

Tested by: glebius@

Note that pc98 specific defines maybe would be better in a header file.

- Ignore BIOS IRQs (that is, IRQ settings left by the BIOS or a previous OS
in the PCI config registers) that are > 15 as $PIR can only route PCI
interrupts to ISA IRQs which are limited to the 0 to 15 range.
- Remove an extra word from a printf.

Reported by: othermark atkin901 at yahoo dot com
MFC after: 3 days

Trust the settings programmed by the BIOS over what the $PIR says.
Specifically, if the BIOS has programmed an IRQ for a device that doesn't
match the list of valid IRQs for the link, use it anyway as some BIOSes
don't correctly list the valid IRQs in the $PIR. Also, allow the user
to specify an IRQ that $PIR claims is invalid as an override, but emit a
warning in that case.

Add code to read the primary PCI bus number out of the Compaq/HP 6010
hotplug Host to PCI bridge. This is only needed for the non-ACPI case
as the BIOS includes a proper _BBN method in ACPI.

Make a bunch of malloc types static.

Found by: src/tools/tools/kernxref

/* -> /*- for copyright notices, minor format tweaks as necessary

Remove a stray critical_exit().

Submitted by: johan

Expand the scope of the critical section in the PCIe read and write methods
on the advice of Alan Cox.

Due to a significant addition of code, add my copyright to this file. Also
note that the PCIe work was made possible due to hardware donations from
the FreeBSD Foundation and Intel. Thanks!

Add support for the memory-mapped PCI Express configuration mechanism. This
actually is a property of the northbridge and applies to all PCI/PCI-X/PCIe
devices in the system, though only PCIe devices will respond to registers
higher than 256. This uses per-CPU pools of temporary mappings so that
the whole 256MB of configuration space doesn't have to be mapped all at
once. While the sf_buf API was considered for this, the fact that it
requires sleep locks and can return failure made it unsuitable for this use.

For now only the Intel Grantsdale and Lindenhurst (925 and 752x) chipsets are
supported. Since there doesn't appear to be a compatible way to determine
northbridge support, new chipsets will have to be explicitely added in the
future.

Add TUNABLE_LONG and TUNABLE_ULONG, and use the latter for the
hw.pci.host_mem_start tunable. Add comments to TUNABLE_INT and
TUNABLE_QUAD recommending against their use.

MFC after: 3 weeks

Whitespace cleanup

Make the lower range of the memory area 0x80000000 again. Also
introduce hw.{pci,acpi}.host_mem_start tunable to change this.

MFC: ASAP

Add missing 'static'

For legacy PCI bridges, limit memory allocation to the top 32MB of
RAM. Many older, legacy bridges only allow allocation from this
range. This only appies to devices who don't have their memory
assigned by the BIOS (since we allocate the ranges so assigned
exactly), so should have minimal impact.

Hoewver, for CardBus bridges (cbb), they rarely get the resources
allocated by the BIOS, and this patch helps them greatly. Typically
the 'bad Vcc' messages are caused by this problem.

Consistently use __inline instead of __inline__ as the former is an empty macro
in <sys/cdefs.h> for compilers without support for inline.

Trim a few things from the dmesg output and stick them under bootverbose to
cut down on the clutter including PCI interrupt routing, MTRR, pcibios,
etc.

Discussed with: USENIX Cabal

Remove atdevbase and replace it's remaining uses with direct references to
KERNBASE instead.

Allow the pir0 device add to fail since pir0 may already exist. This should
fix the panics in device_set_ivars() that people were seeing on boxes with
multiple Host-PCI bridges but not using ACPI.

Add some missing <sys/module.h> includes which are masked by the
one on death-row in <sys/kernel.h>

- Create a pir0 psuedo device as a child of legacy0 if we attach a legacy
host-PCI bridge device and find a valid $PIR.
- Make pci_pir_parse() private to pci_pir.c and have pir0's attach routine
call it instead of having legacy_pcib_attach() call it.
- Implement suspend/resume support for the $PIR by giving pir0 a resume
method that calls the BIOS to reroute each link that was already routed
before the machine was suspended.
- Dump the state of the routed flag in the links display code.
- If a link's IRQ is set by a tunable, then force that link to be re-routed
the first time it is used.
- Move the 'Found $PIR' message under bootverbose as the pir0 description
line lists the number of entries already. The pir0 line also only shows
up if we are actually using the $PIR which is a bonus.
- Use BUS_CONFIG_INTR() to ensure that any IRQs used by a PCI link are
set to level/low trigger/polarity.

Make the legacy_pcib_attach() function static.

Don't call the BIOS to route a link that has already been routed by the
BIOS during POST as it apparently makes some machines unhappy.

Tested by: mux

Add back an include to fix the build for the CPU_ELAN case.

Switch to using the new $PIR interrupt routing code and remove the old
code. The pci_cfgreg.c file now just controls reading/writing PCI config
registers.

Rework the $PIR (aka PCIBIOS) PCI interrupt routing code and split it off
into its own file:
- All of the $PIR interrupt routing is now done in a link-centric fashion.
When a host-PCI bridge that uses the $PIR attaches, it calls pir_parse()
to parse the table. This scans for link devices and merges all the masks
for each link device from the table entries. It then looks at the intline
register of PCI devices connected to a link to figure out if the BIOS has
routed this link and if so to which IRQ.
- The IRQ for any given link can be overridden via a hint like so:
'hw.pci.link.0x62.irq=10' Any IRQ set in this matter is treated as if it
were set that way by the BIOS.
- We only call the BIOS to route each link device once.
- When a PCI device wants to route an interrupt, we look it up in the $PIR
to find the associated link. If the link is routed, we simply return the
IRQ it is using. If it is not routed, we have to pick one. This uses a
different algorithm from the old code. First off, when we try to pick
an interrupt from a mask of possible interrupts, we try to pick the one
that is least loaded as far as PCI devices. We maintain this weight based
on the number of devices attached to each link device. When choosing an
IRQ, we first attempt to route using any PCI only interrupts (the old
code did this as well). If that doesn't work, we try to use the list of
IRQs that the BIOS has used. This is a new step that the new code didn't
do and avoids using IRQ 3 or 4 for every virgin interrupt routing. If
none of the IRQs that the BIOS used worked, then we fall back to trying
anything.
- The fallback mask for !PC98 was fixed to include IRQ 3 and not allow IRQ
2.
- We don't use the $PIR to route interrupts on a PCI-PCI bridge unless it
has already been used to route on at least one Host-PCI bridge. This
helps to avoid mixing and matching x86 firmware PCI interrupt routing
methods (which is a Bad Thing(tm)).

Silence on: current@

Replace an outb() during the test for configuration mechanism #1 with a
DELAY(1) instead. After wading through old commit logs, I found that the
outb() was added not as part of the test but as an intentional delay. In
fact, according to Shanley's PCI book, the configuration 1 data and address
ports should only be accessed using aligned 32-bit accesses (i.e. inl()
and outl()). Thus, using outb() to just the last byte of the port violates
the PCI spec it would seem. On at least one box doing so broke the probe
for PCI, whereas changing it to a DELAY(1) fixed the probe.

Reported by: Sean Welch <welchsm@earthlink.net>
MFC after: 1 week

New APIC support code:

- The apic interrupt entry points have been rewritten so that each entry
point can serve 32 different vectors. When the entry is executed, it
uses one of the 32-bit ISR registers to determine which vector in its
assigned range was triggered. Thus, the apic code can support 159
different interrupt vectors with only 5 entry points.
- We now always to disable the local APIC to work around an errata in
certain PPros and then re-enable it again if we decide to use the APICs
to route interrupts.
- We no longer map IO APICs or local APICs using special page table
entries. Instead, we just use pmap_mapdev(). We also no longer
export the virtual address of the local APIC as a global symbol to
the rest of the system, but only in local_apic.c. To aid this, the
APIC ID of each CPU is exported as a per-CPU variable.
- Interrupt sources are provided for each intpin on each IO APIC.
Currently, each source is given a unique interrupt vector meaning that
PCI interrupts are not shared on most machines with an I/O APIC.
That mapping for interrupt sources to interrupt vectors is up to the
APIC enumerator driver however.
- We no longer probe to see if we need to use mixed mode to route IRQ 0,
instead we always use mixed mode to route IRQ 0 for now. This can be
disabled via the 'NO_MIXED_MODE' kernel option.
- The npx(4) driver now always probes to see if a built-in FPU is present
since this test can now be performed with the new APIC code. However,
an SMP kernel will panic if there is more than one CPU and a built-in
FPU is not found.
- PCI interrupts are now properly routed when using APICs to route
interrupts, so remove the hack to psuedo-route interrupts when the
intpin register was read.
- The apic.h header was moved to apicreg.h and a new apicvar.h header
that declares the APIs used by the new APIC code was added.

Lower the priority of the legacy host to pci bridge driver so that other
non-ACPI host-bridge drivers can preempt this driver.

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.

We represent PCI intpin's two different ways. One is the way that the
intpin register is expressed in hardware where 0 means none, 1 means INTA,
2 INTB, etc. The other way is commonly used in loops where 0 means INTA,
1 means INTB, etc. The matchpin argument to pci_cfgintr_search() is
supposed to be the first form, but we passsed in a loop index of the
second. This fix adds one to the loop index to convert to the first form.

Reported by: Pavlin Radoslavov <pavlin@icir.org>

- Rename PCIx_HEADERTYPE* to PCIx_HDRTYPE* so the constants aren't so long.
- Add a new PCIM_HDRTYPE constant for the field in PCIR_HDRTYPE that holds
the header type.
- Replace several magic numbers with appropriate constants for the header
type register and a couple of PCI_FUNCMAX.
- Merge to amd64 the fix to the i386 bridge code to skip devices with
unknown header types.

Requested by: imp (1, 2)

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.

PC98 uses different mask of IRQ.

The MI code was modified to filter the devices based on its header
type. We know about header types 0, 1 and 2. Ignore the rest in the
MD i386 code when we're looking for bridges. You cannot look at the
vendor tag. And if you don't you certainly can't look at function > 0
if the device isn't there.

The new soekris boards' GEODE cpu has issues with the old way. This
is reported to have fixed it.

MFC After: 2 days

Add hw.pci.irq_override_mask, which is a mask of interrupts that are
considered to be good to try when it otherwise has no clue about which
interrupts to try. This is a band-aide and we really should try to
balance the IRQs that we arbitrarily pick, but it should help some
people that would otherwise get bad IRQs.

- Rename nexus_pcib to legacy_pcib. I've been meaning to do this for a
while after the legacy device was added since this driver hangs from
legacy and not nexus.
- Make several methods non-static so they can be reused in a mptable
host -> pci bridge driver that will be added at a later date.
- Let legacy_pcib() use pcibios_pcib_route_interrupt() directly instead of
wrapping it in a private function. Originally, I thought I was going to
have the nexus_pcib() driver make a runtime APIC vs. 8259A check and call
the appropriate routing method (MPTable vs. PIR) that way, but it ended
up being cleaner to make nexus_pcib() just work with PIR and have a
separate host -> pci bridge driver for the mptable/apic case.

Use the secondary bus number instead of the number of the bus the PCI-PCI
bridge lives on (i.e., the parent bus) when probing the PIR table for a
bus. This could cause the PCIBIOS PCI-PCI bridge driver to bogusly attach
to bridges that weren't in the PIR but whose parent bus was in the PIR.

Use __FBSDID().

Remove unused variable.

Found by: FlexeLint

Initiate de-orbit burn for USE_PCI_BIOS_FOR_READ_WRITE. This has been
#if'ed out for a while. Complete the deed and tidy up some other bits.

We need to be able to call this stuff from outer edges of interrupt
handlers for devices that have the ISR bits in pci config space. Making
the bios code mpsafe was just too hairy. We had also stubbed it out some
time ago due to there simply being too much brokenness in too many systems.
This adds a leaf lock so that it is safe to use pci_read_config() and
pci_write_config() from interrupt handlers. We still will use pcibios
to do interrupt routing if there is no acpi.. [yes, I tested this]

Briefly glanced at by: imp

Outdent the string rather than use concatenation.

MFp4:
o Fix small style nit. This was supposed to be part of the last batch of
style fixes, but somehow didn't get merged.

Recognize the Serverworks CIOB30 host to pci bridge.

MFp4:
o It turns out that we always need to try to route the interrupts for
the case where the $PIR tells us there can be only one. Some machines
require this, while others fail when we try to do this (bogusly, imho).
Since we have no apriori way of knowing which is which, we always try to
do the routing and hope for the best if things fail.
o Add some additional comments that state the obvious, but amplify it in
non-obvious ways (judging from the questions I've gotten).

This should un-break older laptops that still have to use PCIBIOS to route
interrupts.

Tested by: sam

Use 0xffffffff instead of -1 for id to compare against.
Use exact width types, since this is a MD file and won't be used elsewhere.
Fix a couple of resulting printf breakages

Bug found by: phk using Flexlint

Revert last commit, there actually was a -1 waaaaay down in pcireg_cfgread().

"id" is never going to be -1 when it is unsigned.

Spotted by: FlexeLint

Use the global pcib devclass instead of our own static copy.

o go ahead and route the interupt, even if it is supposedly unique.
there are some strange machines that seem to need this.
o delete bogus comment.
o don't use the the bios for read/writing config space. They interact badly
with SMP and being called from ISR. This brings -current in line with
-stable.

# make the latter #ifdef on USE_PCI_BIOS_FOR_READ_WRITE in case we
# need to go back in a hurry.

Add 2 Ids for new ServerWorks host to PCI bridge chipset.
These are still unknown name but these are working as well
as the other ServerWorks chipset.
Description strings should be corrected when the chipsets
are known.

MFC after: 1 week

Don't call function in return() for a void function.

Now that we only probe host-PCI bridges once, we no longer have to check to
see if we have been probed before by checking for a pciX bus device.

Put verbose printf's in the PCI BIOS interrupt routing code under
if (bootverbose).

Change the nexus_pcib driver (eventually to be renamed to legacy_pcib) to
hang off of the legacy driver instead of the nexus.

Axe unused include.

Make sure a $PIR table header has a valid length before accepting the table
as valid.

Submitted by: Michal Mertl <mime@traveller.cz>

#include "opt_bla.h" goes first says Bruce.

Fix style(9) bugs.

Brucified by: bde

Add a subclass of the PCI-PCI bridge driver that uses the PCIBIOS to
route interrupts if the child bus is described in the PCIBIOS interrupt
routing table. For child busses that are in the routing table, they do
not necessarily use a 'swizzle' on their pins on the parent bus to route
interrupts for child devices. If the child bus is an embedded device then
the pins on the child devices can be (and usually are) directly connected
either to a PIC or to a Interrupt Router. This fixes PCIBIOS interrupt
routing across PCI-PCI bridges for embedded devices.

Add a function pci_probe_route_table() that returns true if our PCI BIOS
supports interrupt routing and if the specified PCI bus is present in the
routing table.

Dump the $PIR table if booting verbose.

- Add a pci_cfgintr_valid() function to see if a given IRQ is a valid
IRQ for an entry in a PCIBIOS interrupt routing ($PIR) table.
- Change pci_cfgintr() to except the current IRQ of a device as a fourth
argument and to use that IRQ for the device if it is valid.
- If an intpin entry in a $PIR entry has a link of 0, it means that that
intpin isn't connected to anything that can trigger an interrupt. Thus,
test the link against 0 to find invalid entries in the table instead of
implicitly relying on the irqs field to be zero. In the machines I have
looked at, intpin entries with a link of 0 often have the bits for all
possible interrupts for PCI devices set.

Add support for printing out the contents of a PCI BIOS $PIR interrupt
routing table on the console. Eventually it will be printed during
verbose boots.

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

Test PCIbios.ventry against 0 to see if we found a PCIbios entry point,
not the 'entry' member. The entry point is formed from both a base and
a relative entry point. 'entry' is that relative offset. It is perfectly
valid to have an entry point with a relative offset of 0. PCIbios.ventry
is the virtual address of the entry point that takes both 'base' and
'entry' into account, thus it is the proper variable to test to see if we
have an entry point or not.

Change the support for AMDs ElanSC520 CPU from being a device driver to
be
options CPU_ELAN
(NB: Soekris.com users!)

It is cleaner this way. We still recognize the cpu on the host-pci bridge.

Move a prototype to the least wrong place.

Suggested by: bde

style(9)ize the whole file

Approved in concept a long time ago by: msmith

Add support for probing secondary buses on the ServerWorks Grand Champion
chipset used for P4-Xeon machines

PR: kern/38894
Tested-by: "Marc G. Fournier" <scrappy@hub.org>
Submitted-by: Mark Tinguely (partially)

Add initialization code for the AMD Elan sc520 which maps the MMCR
into KVM and sets the i8254 frequency to the correct value.

Add an entry for the AMD Elan SC520 hostbridge. I do not belive we can
identify this gadget on the CPUID result alone, so I intend to activate
the necessary magic (i8254 frequency for instance) for it based on the
precense of the on-chip host to PCI bridge.

Use a common function to map the bogus intlines.
Don't require pin be non-zero before we map bogus intlines, always do it.
This fixes a number of problems on HP Omnibook computers.

Tested/Reviewed by: Brooks Davis

Restore the irq=0 => irq=255 hack to pci_cfgintr_search(). Just having
it in pci_cfgregread() wasn't sufficent on at least the HP Omnibook 500.

Reviewed by: imp

o Work around bugs in the powerof2 macro: It thinks that 0 is a power of
2, but that's not the case. This fixes the case where there were slots
in the PIR table that had no bits set, but we assumed they did and used
strange results as a result.
o Map invalid INTLINE registers to 255 in pci_cfgreg.c. This should allow
us to remove the bogus checks in MI code for non-255 values.

I put these changes out for review a while ago, but no one responded
to them, so into current they go.

This should help us work better on machines that don't route
interrupts in the traditional way.

MFC After: 4286 millifortnights

Fix a PNPID in a comment

Submitted by: David Xu

Major rework of the iicbus/smbus framework:

- VIA chipset SMBus controllers added
- alpm driver updated
- Support for dynamic modules added
- bktr FreeBSD smbus updated but not tested
- cleanup

Don't call the bios if the interrupt appaers to be already routed. Some
older PCI BIOSes hate this and this leads to panics when it is done. Also,
assume that a uniquely routed interrupt is already routed. This also
seems to help some older laptops with feable BIOSes cope.

The Libretto L series has no $PIR table, but does have a _PIR table.
This typo keeps us from properly routing an interrupt for CardBus
bridges on this machine. So, now we look for $PIR and then _PIR to
cope. With these changes, the Libretto L1 now works properly.
Evidentally, the idea comes from patch that the Japanese version of
RedHat (or against a Japanese version of Red Hat), but my Japanese
isn't good enough to to know for sure.

Reported by: Hiroyuki Aizu-san <eyes@navi.org>

# This may be an MFC candidate, but I'm not yet sure.

Add identification string for AMD-761 host to PCI bridge.

PR: kern/32255

MFS: I was confused. This code wasn't in -current after all.

Merge in the irq 0 detection. Add comment about why.

If we have irq 0, ignore it like we do irq 255. Some BIOS writers aren't
careful like they should be.

It turns out that while Toshiba laptops don't want to route interrupts
multiple times, others do. The last strategy, which was to assume
that already routed interrupts were good and just return them doesn't
work for some laptops. So, instead, we have a new strategy: we notice
that we have an interrupt that's already routed. We go ahead and try
to route it, none the less. We will assume that it is correctly
routed, even if the route fails. We still assume that other failures
in the bios32 call are because the interrupt is NOT routed.

Note: some laptops do not support the bios32 interface to PCI BIOS and
we need to call it via the INT 2A interface. That is another windmill
to till at later.

Also correct a minor typo and minor whitespace nits.

Strong MFC candidate.

MFS: IRQ ordering, PRVERB and more whining in pcibios_get_version on failure.
Check return value from bios32.

[[ Yes, I was bad and committed this to stable first. I should have done
the commit in the other order. ]]

The general conesnsus on irc was that pci bios for config registers
and such was just a bad idea and one that users should be forced to
enable if they want it. This patch introduces a hw.pci.enable_pcibios
tunable for those people. This does not impact the pcibios interrupt
routing at all.

Approved by: peter, msmith

Detect a certain type of PCIBIOS brain damage. For some reason,
some bios vendors took it apon themselves to "censor" the
host->pci bridges from PCIBIOS callers, even when the caller
explicitly asks for them. This includes certain Compaq machines
(eg: DL360) and some laptops.

If we detect this, shut down pcibios and revert to using IO
port bashing.

Under -current, apcica does a better job anyway.

Un-swap irq/link byte values so that printf works.

Free the memory we get from devclass_get_devices and device_get_children.

Submitted by: wpaul

Fix a warning due to missing prototype.

Remove declaration of airq variable from outer block. There were two
declarations of a variable of the same name. The one in the outer block
was unused and probably just slipped in at one point or another. This
silences a compiler warning.

Next phase in the PCI subsystem cleanup.

- Move PCI core code to dev/pci.
- Split bridge code out into separate modules.
- Remove the descriptive strings from the bridge drivers. If you
want to know what a device is, use pciconf. Add support for
broadly identifying devices based on class/subclass, and for
parsing a preloaded device identification database so that if
you want to waste the memory, you can identify *anything* we know
about.
- Remove machine-dependant code from the core PCI code. APIC interrupt
mapping is performed by shadowing the intline register in machine-
dependant code.
- Bring interrupt routing support to the Alpha
(although many platforms don't yet support routing or mapping
interrupts entirely correctly). This resulted in spamming
<sys/bus.h> into more places than it really should have gone.
- Put sys/dev on the kernel/modules include path. This avoids
having to change *all* the pci*.h includes.

Hack to work around a probe which will lock up at least some i450GX-based
systems.

From the PR:

When 'probe.slot' is PCI_SLOTMAX (== 31) and 'probe.func' is 7,
call to 'pci_cfgread()' here and machine suddenly hangs up.
I don't know why... (or 450GX chipset's bug?)

PR: i386/20379
Submitted by: Masayuki FUKUI <fukui@sonic.nm.fujitsu.co.jp>

Improve the PCI interrupt routing code. Now the process is as follows:

- Look for a hardwired interrupt in the routing table for this
bus/device/pin (we already did this).
- Look for another device with the same link byte which has a hardwired
interrupt.
- Look for a PCI device matching an entry with the same link byte
which has already been assigned an interrupt, and use that.
- Look for a routable interrupt listed in the "PCI only" interrupts
field and use that.
- Pick the first interrupt that's marked as routable and use that.

Return -10000 in pci_hostb_probe to allow agp driver (disabled otherwise)

Add i815 Host to Hub

Call the BIOS to route the selected interrupt. Correctly calculate the
interrupt from the PCI routing table (ffs returns 1 for the rightmost
bit, not 0).

Remove debug writes introduced in prior commit

Add the ability to use the $PIR table in the BIOS to route interrupts
on demand.

Submitted by: msmith

Change the text for the ServerWorks north bridge chips. RCC is now
officially listed as ServerWorks by www.pcisig.com.

When testing for PCI bus overlap with another enumerator, make sure we
check for the right bus number. This is still not quite right, but
fixes things for multi-bus machines again.

Submitted by: tegge

Only attach "legacy" PCI busses if none have been attached via any other
method.

Move the i386 PCI attachment code out of i386/isa back into i386/pci.

Split out the configuration space access primitives, as these are needed
elsewhere as well.

Get out the roto-rooter and clean up the abuse of nexus ivars by the
i386/isa/pcibus.c. This gets -current running again on multiple host->pci
machines after the most recent nexus commits. I had discussed this with
Mike Smith, but ended up doing it slightly differently to what we
discussed as it turned out cleaner this way. Mike was suggesting creating
a new resource (SYS_RES_PCIBUS) or something and using *_[gs]et_resource(),
but IMHO that wasn't ideal as SYS_RES_* is meant to be a global platform
property, not a quirk of a given implementation. This does use the ivar
methods but does so properly. It also now prints the physical pci bus that
a host->pci bridge (pcib) corresponds to.

Catch a few more bogosities in certain chipsets before they mess us up.
Some have dual host->PCI bridges for the same logical pci bus (!), eg:
some of the RCC chipsets. This is a 32/64 bit 33/66MHz and dual pci
voltage motherboard so persumably there are electical or signalling
differences but they are otherwise the same logical bus.
The new PCI probe code however was getting somewhat upset about it and
ended up creating two pci bridges to the same logical bus, which caused
devices on that logical bus to appear and be probed twice.

The ACPI data on this box correctly identifies this stuff, so bring on
ACPI! :-)

Take a shot at fixing multiple pci busses on i386.
pcib_set_bus() cannot be used on the new child because it is
meant to be used on the *pci* device (it looks at the parent internally)
not the pcib being added. Bite the bullet and use ivars for the bus
number to avoid any doubts about whether the softc is consistant between
probe and attach. This should not break the Alpha code.

* Completely rewrite the alpha busspace to hide the implementation from
the drivers.
* Remove legacy inx/outx support from chipset and replace with macros
which call busspace.
* Rework pci config accesses to route through the pcib device instead of
calling a MD function directly.

With these changes it is possible to cleanly support machines which have
more than one independantly numbered PCI busses. As a bonus, the new
busspace implementation should be measurably faster than the old one.

Add PnP probe methods to some common AT hardware drivers. In each case,
the PnP probe is merely a stub as we make assumptions about some of this
hardware before we have probed it.

Since these devices (with the exception of the speaker) are 'standard',
suppress output in the !bootverbose case to clean up the probe messages
somewhat.

Add OPTi 82C700 chipset.

Submitted by: sanpei@sanpei.org
PR: kern/18155 (part of)

Add 440MX chipset.

Submitted by: YOSHIMURA Hideaki <hideakiy@cs-tokyo01.chuosystem.co.jp>
References: [bsd-nomads:13764]

Don't assume that the PCI BIOS is going to clear the unused bits in %ecx
when it returns.

Some more i386-only BIOS-friendliness:

- Add support for using the PCI BIOS functions for configuration space
accesses, and make this the default.

- Make PNPBIOS the default (obsoletes the PNPBIOS config option).

- Add two new boot-time tunables to disable each of the above.

Add a workaround to allow us to detect the second pci bus on an HP
Netserver LS/2.

Approved by: jkh

Fix an uninitialised variable which affected probing on some machines.

Approved by: jkh
Reviewed by: gallatin

Allow allows peer pci buses which are directly connected to the RCC host pci
chipset to be probed & attached on newer Dell PowerEdge servers, such as
the 2400 and 4400.

Reviewed by: dfr, msmith, jlemon
Tested by: hnokubi@yyy.or.jp (in a previous incantation)
Approved by: jkh

Add PCI Id's for i810 chipsets.

PR: 16517
Submitted by: SAKIYAMA Nobuo <sakichan@lares.dti.ne.jp>
Approved by: jhk

Clean up the cfgmech/pci_mechanism debris. The reason for the existance
of this is no longer an issue as we have a replacement driver for the
one that needed it.

Reviewed by: dfr

Don't use a bogus bus number for Ross host-pci bridges.

PR: kern/15278
Submitted by: Ahmed Benani <ahmed_benani@urbanet.ch>

Remove the 'ivars' arguement to device_add_child() and
device_add_child_ordered(). 'ivars' may now be set using the
device_set_ivars() function.

This makes it easier for us to change how arbitrary data structures are
associated with a device_t. Eventually we won't be modifying device_t
to add additional pointers for ivars, softc data etc.

Despite my best efforts I've probably forgotten something so let me know
if this breaks anything. I've been running with this change for months
and its been quite involved actually isolating all the changes from
the rest of the local changes in my tree.

Reviewed by: peter, dfr

If we have found pci devices via pci_cfgopen(), but don't find a
host->pci bridge specifically, then add a pcib0 device on the motherboard
for the pci bus to hang off.

Requested by: Anders Andersson <anders@sanyusan.se>
Obtained from: dfr

Add text for the AMD-751 host-to-PCI and PCI-to-PCI (AGP) bridges.

$Id$ -> $FreeBSD$

Make the identify routine add itself with priority 100 to make sure it
goes after the npx/apm devices and any other motherboard devices that
may get added down the track.

Hopefully fix the previous commit, it caused *all* bridges to be detected
as PCI->HOST bridges on my (440BX) box.

My change is to remove the test at the beginning entirely, letting the
switch on the device ID happen first. If the device ID is unknown, then
(in the default case) check for the generic PCIS_BRIDGE_HOST tag. This
should allow wierd cases (eg: wpaul's IMS VL bridge) to work by using the
id override. This strategy is more in line with the other PCI match
methods we use elsewhere,

I only have a limited testbed, but having my USB etc devices detected as
PCI->HOST bridges doesn't look good.

Fix nexus_pcib_is_host_bridge() so that it detects my 486's PCI bus
correctly. It has the following code:

if (class != PCIC_BRIDGE || subclass != PCIS_BRIDGE_HOST)
return NULL;

My 486 has an Integrated Micro Solutions PCI bridge which identifies
itself as subclass PCIS_BRIDGE_OTHER, not PCIS_BRIDGE_HOST. Consequently,
it gets ignored. In my opinion, the correct test should be:

if ((class != PCIC_BRIDGE) && (subclass != PCIS_BRIDGE_HOST))
return NULL;

That way the test still succeeds because the chip's class is PCIC_BRIDGE.
Clearly it's not reasonable to expect all host to PCI bridges to always
have a subclass of PCIS_BRIDGE_HOST since I've got one that doesn't.
This way the sanity test should remain relatively sane while still allowing
some oddball yet correct hardware to work. If somebody has a better way
to do it, go ahead and tweak the test, but be aware that
class == PCIC_BRIDGE and subclass == PCIS_BRIDGE_OTHER is a valid case.

While I was here, I also added an explicit ID string for the IMS chipset.
I also dealt with a minor style nit: it's bad karma not to have a default
case for your switch statements, but the one in this routine doesn't have
one. The default string of "Host to PCI bridge" is now assigned in a
default case of the switch statement instead of initializing "s" with the
string before the switch and then not having any default case.

Don't probe if pci_cfgopen() fails to find pci hardware, like we used to
to. This might have caused interesting things on non-PCI hardware if
PCI was compiled in.

Add support for multiple PCI busses directly connected to the nexus.
This is only partially complete, but allows 450NX-based systems with
more than one PCI bus to be used again.

Submitted by: dfr

Move pcibus (host -> pci bus) probe/attach routines from nexus
to pcibus.c. pci_cfgopen() becomes static and there are no more
bus #ifdef's in nexus.c.

Make the class code checks in function pci_cfgcheck less strict.
It failed to recognize the PCI bus in a system that had only an
old chip-set (class code 000000) and a Cyclom multiport serial
card on PCI bus 0, but no VGA card or disk or network controller.

PR: i386/5300
Submitted by: Nickolay N. Dudorov <nnd@itfs.nsk.su>

Removed unused #includes.

Yet another fix for configuration mechanism 1 register accesses:
Adjust the data port address by adding the two low order bits of
the register number. The address port takes only a word address
(i.e. ignores the two low order bits written to it).

Fix previous fix: The enable bit is bit 31 (0x8000000) and not bit 15.

Set enable bit when writing the configuration address in configuration
mode 1. Omission of this bit makes all config register accesses fail in
on recent chip sets ...

(The problem was reported and debug output provided by: Steve Passe)

Completely replace the PCI bus driver code to make it better reflect
reality. There will be a new call interface, but for now the file
pci_compat.c (which is to be deleted, after all drivers are converted)
provides an emulation of the old PCI bus driver functions. The only
change that might be visible to drivers is, that the type pcici_t
(which had been meant to be just a handle, whose exact definition
should not be relied on), has been converted into a pcicfgregs* .

The Tekram AMD SCSI driver bogusly relied on the definition of pcici_t
and has been converted to just call the PCI drivers functions to access
configuration space register, instead of inventing its own ...

This code is by no means complete, but assumed to be fully operational,
and brings the official code base more in line with my development code.

A new generic device descriptor data type has to be agreed on. The PCI
code will then use that data type to provide new functionality:

1) userconfig support
2) "wired" PCI devices
3) conflicts checking against ISA/EISA
4) maps will depend on the command register enable bits
5) PCI to Anything bridges can be defined as devices,
and are probed like any "standard" PCI device.

The following features are currently missing, but will be added back,
soon:

1) unknown device probe message
2) suppression of "mirrored" devices caused by ancient, broken chip-sets

This code relies on generic shared interrupt support just commited to
kern_intr.c (plus the modifications of isa.c and isa_device.h).

Mask out revision register in consistency test of class register.

Fix spelling of align and interrupt in comments.

Fix consistency test to not fail on pre PCI 2.0 motherboards

improve pcibus_check: Only assume PCI if at least one PCI to anything bridge
on bus 0.
This fixes problems with EISA-only systems mistakenly being assumed to support PCI.

Back out part 1 of the MCFH that changed $Id$ to $FreeBSD$. We are not
ready for it yet.

Sync with <pci/pcibus.h>. pcibus.c unfortunately still compiled (with
only 3 or 4 warnings) when pb_maxirq went away.

Make the long-awaited change from $Id$ to $FreeBSD$

This will make a number of things easier in the future, as well as (finally!)
avoiding the Id-smashing problem which has plagued developers for so long.

Boy, I'm glad we're not using sup anymore. This update would have been
insane otherwise.

Make the code more consistant by using the INTR*MASK macros througout the
code.

Reviewed by: bde

[
Bruce suggest removing the macros completely, but I'm not up to that
task quite yet.
]

More merge and update.

(1) deleted #if 0

pc98/pc98/mse.c

(2) hold per-unit I/O ports in ed_softc

pc98/pc98/if_ed.c
pc98/pc98/if_ed98.h

(3) merge more files by segregating changes into headers.

new file (moved from pc98/pc98):

i386/isa/aic_98.h

deleted:

well, it's already in the commit message so I won't repeat the
long list here ;)

Submitted by: The FreeBSD(98) Development Team

Removed unused #includes of <i386/isa/icu.h> and <i386/isa/icu.h>. icu.h
is only used by the icu support modules and by a few drivers that know
too much about the icu (most only use it to convert `n' to `IRQn'). isa.h
is only used by ioconf.c and by a few drivers that know too much about
isa addresses (a few have to, because config is deficient).

Change CONF1_ENABLE_MSK to 0x7ff00000 in another attempt to decide
whether a system could possibly support PCI configuration mechanism 1
(or whether it rather is an EISA only system ...).

Make pcibus_check() ignore Device/Vendor IDs of all 0.

Removed now-unused #includes of <machine/cpu.h>. They were for bootverbose
being declared in the wrong place.

Count PCI irqs in up to 4 ISAish counters named `pci irqnn' instead of
in the clk0 counter.

Reviewed by: s

recording cvs-1.6 file death

Completed function declarations and/or added prototypes and/or added
#includes to get prototypes.

pci now uses a different interrupt handler type for interrupts that it
dispatches and the isa interrupt handler type for the interrupts that
it handles.

Staticize and cleanup.

Make CONF1_ENABLE_MSK1 even less restriktive: Ignore slot ID ...

At least the ASUS Triton motherboards don't disable the PCI bus configuration
accesses after the BIOS bus scan. The previous revision made the assumption,
that every PCI motherboard did ...

Change the test on the initial value of the CONF1_ADDR_PORT register in a way
that makes the probe succeed on triton based motherboards, without breaking
the EISA motherboard that has some non-PCI register at the same address.

Go back to separate tests for configuration mechanism 1 and mechanism 2.
Require the state of the configuration enable bits to be OFF assuming
that the BIOS left them that way, as it should anyway to avoid bad things
to happen.

The tests themselves are copied from the previous release, with the
exception of CONF1_ENABLE_MSK1 having the LSB set. This bit should be
read back as '0', since only DWORD addresses are legal.

Fix bad typo: CONF1_ENABLE_RES1 was written CONF1_ENABLE_CHK1 ...

New approach to the PCI bus configuration mechanism probe problem:
- try to make sure there is any kind of PCI device
- if there is anything at port 0x0cf8, then check for mech. 1 or 2

Revert most changes of previous commit.
Changes relative to 1.12:
- Put extra instruction between outl()/inl() sequence to prevent the
old value being read back because of the bus capacitance.
- Additional check for existence of register at CONF2_ENABLE_PORT.

Another try to determine the PCI bus configuration mode (and whether
there is a PCI bus at all) ...

- Do not expect the chip sets to follow even very clearly expressed
requirements of the PCI 2.0 spec.
- Do not read back the value just written to an I/O port without making
sure that some other data have crossed the bus in between ...

Improved verification of configuration space accesses working:
Scan for devices instead of assuming that device 0 is present on bus 0
of every PCI motherboard.

Make the PCI host bridge probe code more robust when dealing with chip sets
that use configuration mode 1, but still violate the PCI 2.0 specs ...
(Required for the Compaq Proliant, for example.)

The PCI config mechanism 1 test failed for the Intel Aries.
Make it less strict ...

Submitted by: NIIMI Satoshi <sa2c@and.or.jp>

PCI configuration mechanism now determined by a method, that doesn't
fail on new hardware (Compaq Prolinea and Compaq Prosignea), and that
doesn't erroneously identify old mech. 2 chip sets as using mech. 1.
(See section 3.6.4.1.1 of the PCI bus specs rev. 2.0)

Correct pcibus_setup() to return as soon as one test succeeds.

Delete PCI PCI bridge simulator code ...

Submitted by: Wolfgang Stanglmeier <wolf@kintaro.cologne.de>

Remove spurious declaration of printf().

Submitted by: Michael Reifenberger <root@rz-wb.fh-sw.de>

New ISA specific PCI code.
Supports shared PCI interrupts.

Submitted by: Wolfgang Stanglmeier <wolf@kintaro.cologne.de>

Replace all remaining instances of `i386/include' by `machine' and fix
nearby #include inconsistencies.

Keep PCI_CONF_MODE in a safe place for later reference, if #defined.

Reviewed by: se
Submitted by: seb@erix.ericsson.se (Sebastian Strollo)

Initialisation of interrupt masks changed.

Reviewed by: se
Submitted by: wolf (Wolfgang Stanglmeier)

Reviewed by: se
Submitted by: wolf (Wolfgang Stanglmeier)
New ISA dependend file for PCI bus support.
Replaces sys/i386/pci/pcibios.c.

Change the string returned in the aic7870 motherboard probe case. Change
#define to a more appropriate name.

Add $Id. Recognize motherboard aic7870 based controllers.

Add in aic7770.c (EISA/VL Adaptors) and aic7870.c (PCI adaptor) dependancies
for the ahc driver.

Submitted by: Mikael Hybsch <micke@dynas.se>

Add support for NCR 53c815 PCI SCSI chip.

Restore my changes in rev 1.11 that Garrett killed in his commit.

Move ARP interface initialization into if_ether.c:arp_ifinit().

Add support for -v option passed to boot loader (bootverbose).
Reviewed by:
Submitted by:
Obtained from:

Unbogify the size being passed to bzero when clearing struct softc.

The physical memory allocated for input DMA must be contiguous. The driver
worked in the past only because of good fortune. Anyway, use the contig alloc
routine I wrote awhile ago (vm_page_alloc_contig) for the sound code to do
this allocation. Also, specify read+write on the permissions to pmap_enter().
Specifying just read can have unexpected consquences.

Fix bug I introduced that broke BPF support. Caused by a byte order problem
in an if () expression. Problem fixed by removing the test for ETHERTYPE_*
before passing packet to higher layers.

Really deactivated the code that puts scsi stats into dk0.
This should have been disabled for some time, but I had screwed up ...
This made spurious values appear for fd0 in systat, when there was
NCR SCSI activity.

Moved conversion of ether_type to host byte order out of ethernet drivers
and into ether_input(). It was silly to have bpf want this one way and
ether_input want it another way. Ripped out trailer support from the few
remaining drivers that still had it.

Bunch of fixes from Matt Thomas:

1) make #includes correct
2) fix bugs in address check macros
3) fixed bugs in, and enabled, recopy if heavily fragmented code
4) moved call to bpf tap to be before enqueing packet (probably gratuitous)
5) fixed bug that caused "abnormal interrupt" at boot time/first use
6) added support for reading Zynx address ROM
7) fixed bug that caused broadcasts to not work shortly after booting (only
manifested if not using multicast - e.g. not in FreeBSD 2.0)
8) fixed spelling errors in comments

Submitted by: Matt Thomas

Added support for SIOCSIFMTU ioctl. Fixed bug that caused panic at boot time
related to interrupts being enabled before the device attach. The interrupt
should be mapped *after* the device attach.

Added missing call to bpf on the transmit side. Nuked revision log.

Fixed bug that somehow made it into here: the ifp must be stored in the
mbuf header for received packets. Minor performance optimizations. Removed
#ifdef MULTICAST's as this isn't optional in our kernel.

Submitted by:
Added hooks for "lsdev" ...
PCI devices should need no individual code for lsdev.

Submitted by: Paul F. Werkowski
Add a quirk line for the SONY SDT-5000 like it is done for the WangDAT tapes.

Make this documentation more sane in the context of 2.0.

Change some compile time defaults, which may be overridden from the
kernel config file by options lines.
Now the default settings are FAST SCSI, max. 4 TAGS, WIDE transfers,
if supported by the hardware ...

Modified fifth parameter (imask) to register_intr() according to
new definition of that function.

Added decoding of chipset registers for Pentium 90/100 CPUs.

#ifdef'ed the verbose probe messages. Only until I find a good way of
making it run-time selectable, then it will be back (at will).

Fixed typo and message format change to avoid line wrapping.

Submitted by: "Randall W. Dean" <rwd@osf.org>
Fixed typo in initialisation of DMODE:
PCI burst length now really 16 transfers as advertised ...

Submitted by: Wolfgang Stanglmeier <wolf@dentaro.GUN.de>
removed PCI mapping call from vga_attach() in pcisupport.c.

Submitted by: Wolfgang Stanglmeier <wolf@dentaro.GUN.de>
Debugging option changed, getirr() removed.

Submitted by: Matt Thomas <thomas@lkg.dec.com>
Preliminary FAST Ethernet support added (DEC21140).

Temporary disable scanning for VGA cards, was causing problems.

Submitted by: Bruce Evans <bde@zeta.org.au>
Function getirr() could clobber interrupt controller state.

Small corrections:
pci config line now without irq
boot message changed

Submitted by: Wolfgang Stanglmeier <wolf@dentaro.GUN.de>
Bug fixed, that caused system hang on first interrupt on some motherboards.

New version of PCI bus configuration code, now supports dynamic interrupt
configuration (using BIOS supplied values).
NCR SCSI and DEC Ethernet driver patched to use this feature.
*** Remove PCI IRQ specifications from your kernel config file ! ***

Submitted by: Wolfgang Stanglmeier <wolf@dentaro.GUN.de>
Name change: pci_intel.c -> pcisupport.c

Submitted by: Wolfgang Stanglmeier <wolf@dentaro.GUN.de>
Files merged and name change: pci_device.h + pcibios.h -> pcireg.h

Submitted by: Wolfgang Stanglmeier <wolf@dentaro.GUN.de>
Name change: ncrstat -> ncrcontrol

Submitted by: Wolfgang Stanglmeier <wolf@dentaro.GUN.de>
Name change: ncr_reg.h -> ncrreg.h

Add Matt Thomas's DC21040 PCI Ethernet driver. (This is turning out
to be quite a popular chip, so expect to see a number of products
based on it.)

Correct `de' device ID.
Add pd_npresent field for future loadable PCI drivers.

Correct DEC -> Digital Equipment Corporation.

Submitted by: Wolfgang Stanglmeier <wolf@dentaro.GUN.de>
New version with improved support for WIDE SCSI using the NCR 53c825.
Test for buggy secondary cache implementations.
PCI Int to IRQ mapping now specified per slot.

Add missing closing comment. I have taken the non-knf format
of this code up with the authors and they are looking at converting it,
but until then leave the format alone.

Removed inclusion of pio.h and cpufunc.h (cpufunc.h is included from
systm.h). Merged functionality of pio.h into cpufunc.h. Cleaned up some
related code.

Submitted by: Wolfgang Stanglmeier <wolf@dentaro.GUN.de> + <se>
Improved bus probing, symbolic names for registers..
Chip set parameters get dumped for intel PCI chip sets
(82424+82434 only, for now).

Submitted by: Wolfgang Stanglmeier <wolf@dentaro.GUN.de>
New version with support for the NCR 53c810 and 53c825.
Support for WIDE SCSI devices.

Added a bit of missing functionality to make this work correctly on a
wider variety of systems. Include the deivers from pci_intel.c in
pci_config.c (I hope this is what was intended; my system works ok).
Use pmap_mapdev(). Automatically map any large linear frame buffers
or whatnot in VGA-style devices which ordinarily would not have their
own drivers, and don't call not_supported() for them. (This shuts up
complaints about my Matrox card.) Include the beginnings of what could
eventually become dynamically-loadable PCI devices. Allow for the
possibility of PCI devices simply providing a PCI veneer over an existing
ISA device, and shut up about them, too.
Make autoconfiguration text conform more to the style of other supported
buses.

This commit was generated by cvs2svn to compensate for changes in r2528,
which included commits to RCS files with non-trunk default branches.

Submitted by: Wolfgang Stanglmeier <wolf@dentaro.GUN.de>
Merged in changes required for NetBSD support (by mycroft@gnu.ai.mit.edu)
and support for multiple NCR chips.

This commit was generated by cvs2svn to compensate for changes in r2432,
which included commits to RCS files with non-trunk default branches.