ppb(4): Print out PCIe Gen5 link speed correctly

Use PCI-SIG official acronyms:

- RP stands for Root Port.
- RC stands for Root Complex.
- RCIEP stands for Root Complex Integrated End Point.

Merge thorpej-cfargs2.

branches: 1.72.4;
Forward the bridge's devhandle to the downstream bus.

branches: 1.71.2; 1.71.4;
Merge thorpej-cfargs branch:

Simplify and make extensible the config_search() / config_found() /
config_attach() interfaces: rather than having different variants for
which arguments you want pass along, just have a single call that
takes a variadic list of tag-value arguments.

Adjust all call sites:
- Simplify wherever possible; don't pass along arguments that aren't
actually needed.
- Don't be explicit about what interface attribute is attaching if
the device only has one. (More simplification.)
- Add a config_probe() function to be used in indirect configuiration
situations, making is visibly easier to see when indirect config is
in play, and allowing for future change in semantics. (As of now,
this is just a wrapper around config_match(), but that is an
implementation detail.)

Remove unnecessary or redundant interface attributes where they're not
needed.

There are currently 5 "cfargs" defined:
- CFARG_SUBMATCH (submatch function for direct config)
- CFARG_SEARCH (search function for indirect config)
- CFARG_IATTR (interface attribte)
- CFARG_LOCATORS (locators array)
- CFARG_DEVHANDLE (devhandle_t - wraps OFW, ACPI, etc. handles)

...and a sentinel value CFARG_EOL.

Add some extra sanity checking to ensure that interface attributes
aren't ambiguous.

Use CFARG_DEVHANDLE in MI FDT, OFW, and ACPI code, and macppc and shark
ports to associate those device handles with device_t instance. This
will trickle trough to more places over time (need back-end for pre-OFW
Sun OBP; any others?).

branches: 1.70.4;
Remove duplicate include of opt_ppb.h.
No binary changes.

branches: 1.69.2;
Identify 16GT/s.

- Almost all ppbreg.h's definitions are also in pcireg.h. Remove duplicated
definitions from ppbreg.h and move some definitions from ppbreg.h to
pcireg.h.
- Change fast back-to-back "capable" to "enable" in pci_subr.c.
- Print Primary Discard Timer, Secondary Discard Timer, Discard Timer Status
and Discard Timer SERR# Enable bit in pci_subr.c.
- PCI_BRIDGE_PREFETCHBASE32_REG and PCI_BRIDGE_PREFETCHLIMIT32_REG are
"upper" 32bit registers, rename to *UP32_REG to avoid confusion.
- Use macro.

If the secondary bus is configured and the bus mastering is not enabled,
enable it. Suggested by thorpej@.

Explicitly enable bus masterling in case BIOS, UEFI or firmware don't enable
it. Might fix PR kern/53811.

Merge the [pgoyette-compat] branch

- defflag PPB_USEINTR
- Print "interrupting at "

branches: 1.63.2; 1.63.8; 1.63.10;
Fix typos.

- Fix a bug that a device which has no PCIe capability incorrectly
accessess the PCI config area in ppbdetach().
- Don't add event counters if slot interrupt isn't used.

Make ppb(4) interrupt support stable:
- Disable all interrupts in the beginning of attach. Without this, interrupt
storm occurs while cold == 1 on some environment.
- Disable command complete interrput for a while to prevent hangup on some
enviroment. I'm sorry, I don't know what this bit is :-|
- Check all status bits and return 0 if an interrupt is not for me. It's
required for INTx. Tested on XEN3_DOM0 because it doesn't support MSI yet.
- Return 1 when a interrupt is processed.

branches: 1.60.2;
Disable ppb(4)'s interrupt for a while. It causes hangup on some environment.
Define PPB_USEINTR if you'd like to use interrupt.

Disable and disestablish interrupt in ppbdetach().

in ppbdetach(), try to detach the children before tearing down our own state,
to avoid trouble if detaching the children fails.

Enable PCIe's interrupt as much as possilbe in ppb(4) to detect and count
status change event. HotPlug function itself have not implemented yet.

- Interrupt and each event are counted by evcnt(9). Example:

ppb0 Interrupt 0 0 intr
ppb0 Attention Button Pressed 0 0 misc
ppb0 Power Fault Detected 0 0 misc
ppb0 MRL Sensor Changed 0 0 misc
ppb0 Presence Detect Changed 0 0 misc
ppb0 Command Completed 0 0 misc
ppb0 Data Link Layer State Changed 0 0 misc

- Print message if ppb_printevent is not zero. The default vaule is 0.
The output messages:

Attention Button Pressed
Power Fault Detected
MRL Sensor Changed
Presence Detect Changed
Command Completed
Data Link Layer State Changed

- Remove workaround code to disable interrupt (ppb.c rev. 1.35).

Tested with Dell Latitude 2120 without if_bge.c rev. 1.304's workaround.
dmesg when bge's device timeout occured:

ppb3: Presence Detect Changed
ppb3: Data Link Layer State Changed
ppb3: Presence Detect Changed

vmstat -e |grep ppb

ppb3 Interrupt 2 0 intr
ppb3 Presence Detect Changed 2 0 misc
ppb3 Data Link Layer State Changed 1 0 misc

Move struct ppb_softc into ppbvar.h.

branches: 1.55.2; 1.55.4;
Define PCIE_XCAP_{VER,TYPE}(x) and use them.

branches: 1.54.2;
Rename PCIE_XCAP_VER_* macros to avoid confusion.

- Modify message of PCIe capability version. This field (PCIE_XCAP_VER_MASK)
is not specification's version number but the capability structure's version
number. To avoid confusion, print "PCI Express capability version x".
- The max number of PCIe lane is not 16 but 32. Fix the bug using with macro.
- Use macro instead of magic number.
- Gb/s -> GT/s

branches: 1.52.10;
Delete "PCI_" from PCIX and PICE capability registers.

ppb: fix link speed print

Print out negotiated link width and speed for PCIe (merged from
matt-nb5-mips64).

branches: 1.49.6;
extend the pci_aprint_devinfo slightly to cover the cases commonly
used by drivers: a short name for the quiet/naive case and a string
to override the "pcidevs" based name by one provided by the driver,
ride on yesterday's kernel minor version bump

put printing of the pci_devinfo into its own function (not inlined
by purpose) - this is a stack hog, and with this change my uTCA amd64
system boots again
a lot of similar code can be eliminated from pci device drivers this way,
but before doing so (and making the new function part of the module API)
I'd like to consider a modification to make it work with drivers which
prefer to print names from other sources (like pciide)

branches: 1.47.2; 1.47.6;
Tell a pci(4) instance its subordinate PCI buses using a new member
in the pcibus_attach_args, pba_sub. pciN attaches to pba_bus itself.
If pba_bus < pba_sub, then [pba_bus + 1, pba_sub] are subordinate to
pba_bus.

On i386, make mainbus0 attach pci0 with pba_sub = 255 because all buses
1 and up must be subordinate to pci0.

XXX Deal with other architectures.

Redefine PCI_MSI_* and PCI_PCIE_* constants in terms of bits(3).

Use named constants and more conventional variable names in
pci_msi_establish() and pci_msi_disestablish(). Fix a couple of bugs:
pci_msi_establish() returned a pointer to the struct intrhand instead of
to the struct msi_hdl as it was intended to, and pci_msi_disestablish()
did not free(9) the msi_hdl.

add missing break

ppb_fix_pcix changes:
- rename to ppb_fix_pcie
- support version PCI-E 2.0
- print version and device/port type information
- use constants from pcireg.h instead of magic numbers

changes:

ppb2 at pci0 dev 21 function 0: vendor 0x15ad product 0x07a0 (rev. 0x01)
ppb2: unsupported PCI Express version

to:

ppb2 at pci0 dev 21 function 0: vendor 0x15ad product 0x07a0 (rev. 0x01)
ppb2: PCI Express 2.0 <Root Port of PCI-E Root Complex>

On powerpc, recognize PCI Express RC root bridges.

branches: 1.42.2;
A pointer typedef entails trading too much flexibility to declare const
and non-const types, and the kernel uses both const and non-const
PMF qualifiers and device suspensors, so change the pmf_qual_t and
device_suspensor_t typedefs from "pointers to const" to non-pointer,
non-const types.

branches: 1.41.2;
Expand PMF_FN_* macros.

During shutdown, detach devices in an orderly fashion.

Call the detach routine for every device in the device tree, starting
with the leaves and moving toward the root, expecting that each
(pseudo-)device driver will use the opportunity to gracefully commit
outstandings transactions to the underlying (pseudo-)device and to
relinquish control of the hardware to the system BIOS.

Detaching devices is not suitable for every shutdown: in an emergency,
or if the system state is inconsistent, we should resort to a fast,
simple shutdown that uses only the pmf(9) shutdown hooks and the
(deprecated) shutdownhooks. For now, if the flag RB_NOSYNC is set in
boothowto, opt for the fast, simple shutdown.

Add a device flag, DVF_DETACH_SHUTDOWN, that indicates by its presence
that it is safe to detach a device during shutdown. Introduce macros
CFATTACH_DECL3() and CFATTACH_DECL3_NEW() for creating autoconf
attachments with default device flags. Add DVF_DETACH_SHUTDOWN
to configuration attachments for atabus(4), atw(4) at cardbus(4),
cardbus(4), cardslot(4), com(4) at isa(4), elanpar(4), elanpex(4),
elansc(4), gpio(4), npx(4) at isa(4), nsphyter(4), pci(4), pcib(4),
pcmcia(4), ppb(4), sip(4), wd(4), and wdc(4) at isa(4).

Add a device-detachment "reason" flag, DETACH_SHUTDOWN, that tells the
autoconf code and a device driver that the reason for detachment is
system shutdown.

Add a sysctl, kern.detachall, that tells the system to try to detach
every device at shutdown, regardless of any device's DVF_DETACH_SHUTDOWN
flag. The default for kern.detachall is 0. SET IT TO 1, PLEASE, TO
HELP TEST AND DEBUG DEVICE DETACHMENT AT SHUTDOWN.

This is a work in progress. In future work, I aim to treat
pseudo-devices more thoroughly, and to gracefully tear down a stack of
(pseudo-)disk drivers and filesystems, including cgd(4), vnd(4), and
raid(4) instances at shutdown.

Also commit some changes that are not easily untangled from the rest:

(1) begin to simplify device_t locking: rename struct pmf_private to
device_lock, and incorporate device_lock into struct device.

(2) #include <sys/device.h> in sys/pmf.h in order to get some
definitions that it needs. Stop unnecessarily #including <sys/device.h>
in sys/arch/x86/include/pic.h to keep the amd64, xen, and i386 releases
building.

branches: 1.39.8; 1.39.10; 1.39.14; 1.39.18; 1.39.22;
device_t / softc split. Tested on amd64.
"looks good" spz

unuspported -> unsupported

branches: 1.37.2; 1.37.4;
Add methods for detaching self and for detaching children.

Use device_t and accessors. Use aprint_*_dev().

branches: 1.36.6; 1.36.10;
Merge jmcneill-pm branch.

branches: 1.35.6; 1.35.8;
Merge 1.34.22.7 from jmcneill-pm:
Next attempt at trying to fix the irregular interrupt storms on my
Thinkpad: when we find a PCI Express device, check the list of
notification events and if any are sets, clear them. We can't handle
them ATM anyway.

branches: 1.34.8; 1.34.22; 1.34.24; 1.34.26;
__unused removal on arguments; approved by core.

- sprinkle __unused on function decls.
- fix a couple of unused bugs
- no more -Wno-unused for i386

branches: 1.32.20; 1.32.22;
merge ktrace-lwp.

branches: 1.31.2;
Use ANSI function decls and static.

de-__P

branches: 1.29.4; 1.29.6;
Phase out the use of a string as first "attach args" member to control
which bustype should be attached with a specific call to config_found()
(from a "mainbus" or a bus bridge).
Do it for isa/eisa/mca and pci/agp for now. These buses all attach to
an mi interface attribute "isabus", "eisabus" etc., and the autoconf
framework now allows to specify an interface attribute on config_found()
and config_search(), which limits the search of matching config data
to these which attach to that specific attribute.
So we basically have to call config_found_ia(..., "foobus", ...) where
such a bus is attached.
As a consequence, where a "mainbus" or alike also attaches other
devices (eg CPUs) which do not attach to a specific attribute yet,
we need at least pass an attribute name (different from "foobus") so
that the foo bus is not found at these places. This made some minor
changes necessary which are not obviously related to the mentioned buses.

pass string length (= boundary info) to pci_devinfo so that we do not run over
the end of memory region

Use aprint_*()

branches: 1.26.2;
Handle 64bit DMA addresses on PCI for platforms that can (currently only
enabled on amd64). Add a dmat64 field to various PCI attach structures,
and pass it down where needed. Implement a simple new function called
pci_dma64_available(pa) to test if 64bit DMA addresses may be used.
This returns 1 iff _PCI_HAVE_DMA64 is defined in <machine/pci_machdep.h>,
and there is more than 4G of memory.

Use aprint_normal() in cfprint routines.

Add trailing ; to CFATTACH_DECL.

Use CFATTACH_DECL().

Declare all cfattach structures const.

branches: 1.21.2;
* Add "pcitag_t *pba_bridgetag" to pci_attach_args. This is set to
NULL for root PCI busses. For busses behind a bridge, it points to
a persistent copy of the bridge's pcitag_t. This can be very useful
for machine-dependent PCI bus enumeration code.
* Implement a machine-dependent pci_enumerate_bus() for sparc64 which
uses OFW device nodes to enumerate the bus. When a PCI bus that is
behind a bridge is attached, pci_attach_hook() allocates a new PCI
chipset tag for the new bus and sets it's "curnode" to the OFW node
of the bridge. This is used as a starting point when enumerating
that bus. Root busses get the OFW node of the host bridge (psycho).
* Garbage-collect "ofpci" and "ofppb" from the sparc64 port.

add RCSID

branches: 1.19.6; 1.19.8;
Don't pass rd/mult capability though a PCI-PCI bridge. The bridge would
have to break it up into mutliple rd/line's anyhow, so why bother letting
the device issue it in the first place.

branches: 1.18.14; 1.18.16; 1.18.20;
Nuke __BROKEN_INDIRECT_CONFIG.

clean up slightly, correct a few comments

branches: 1.16.2;
Update for config changes.

branches: 1.15.2; 1.15.4;
Pass down bus_dma_tag_t's as appropriate (per Jason's bus_dma code).

Pass the I/O and memory enable flags through the bridge.

update these so they compile whether or not __BROKEN_INDIRECT_CONFIG
is defined.

New bus.h implementation/interface:
- No more distinction between i/o-mapped and memory-mapped
devices. It's all "bus space" now, and space tags
differentiate the space with finer grain than the
bus chipset tag.
- Add memory barrier methods.
- Implement space alloc/free methods.
- Implement region read/write methods (like memcpy to/from
bus space).
This interface provides a better abstraction for dealing with
machine-independent chipset drivers.

backout kprintf changes

printf -> kprintf, sprintf -> ksprintf

change cfprint_t type definition to take a const char *, rather than
a char *, because that's what was really intended, and because
if the print function modifies the string, various things could become
unhappy (so the string should _not_ be modified).

remove unused variables

modify these to provide a new, better-specified PCI interface
(soon to be documented on mailing lists; eventually in section 9 manual
pages), most importantly:
(1) support interrupt pin swizzling on non-i386 systems with
PCI-PCI bridges (per PPB spec; done, but meaningless, on i386).
(2) provide pci_{io,mem}_find(), to determine what I/O or memory
space is described by a given PCI configuration space
mapping register.
(3) provide pci_intr_map(), pci_intr_string(), and
pci_intr_{,dis}establish() to manipulate and print info about
PCI interrupts.
(4) make pci functions take as an argument a machine-dependent
cookie, to allow more flexibility in implementation.

spacing nit

New device attachment scheme:

- split softc size and match/attach out from cfdriver into
a new struct cfattach.

- new "attach" directive for files.*. May specify the name of
the cfattach structure, so that devices may be easily attached
to parents with different autoconfiguration semantics.

These devices don't actually need softc's that are supersets of 'struct
device.' No point in adding that complexity + space if they're not needed,
so axe them.

(1) provide #defines for cf_loc[] entries for devices that attach to
pcibus and pci.
(2) remove the #ifdef i386 from pci.c, and provide a machine-dependent
hook (pci_md_attach_hook()) to do any machine-dependent attachment
gunk, e.g. on the i386 printing out the configuration mode (if bus 0)
(3) don't pass max device number for a given bus in, use
PCI_MAX_DEVICE_NUMBER, which can be defined on a per-machine basis.
(defaults to 32. on i386, it's 32 if pci conf mode == 1, 16 if 2.)

fix aux argument to config_found() for secondary bus.

Preliminary support for PCI-PCI bridges. Recognize a PCI-PCI bridge
and attach the secondary pci bus as a 'pci' device. Note that this support
is incomplete and will not yet work for ports other than that i386. (The
i386 can rely on the PCI interrupt 'line' information to determine
interrupt mapping, which is not necessarily possible on other systems.)

Use PCI-SIG official acronyms:

- RP stands for Root Port.
- RC stands for Root Complex.
- RCIEP stands for Root Complex Integrated End Point.

Merge thorpej-cfargs2.

branches: 1.72.4;
Forward the bridge's devhandle to the downstream bus.

branches: 1.71.2; 1.71.4;
Merge thorpej-cfargs branch:

Simplify and make extensible the config_search() / config_found() /
config_attach() interfaces: rather than having different variants for
which arguments you want pass along, just have a single call that
takes a variadic list of tag-value arguments.

Adjust all call sites:
- Simplify wherever possible; don't pass along arguments that aren't
actually needed.
- Don't be explicit about what interface attribute is attaching if
the device only has one. (More simplification.)
- Add a config_probe() function to be used in indirect configuiration
situations, making is visibly easier to see when indirect config is
in play, and allowing for future change in semantics. (As of now,
this is just a wrapper around config_match(), but that is an
implementation detail.)

Remove unnecessary or redundant interface attributes where they're not
needed.

There are currently 5 "cfargs" defined:
- CFARG_SUBMATCH (submatch function for direct config)
- CFARG_SEARCH (search function for indirect config)
- CFARG_IATTR (interface attribte)
- CFARG_LOCATORS (locators array)
- CFARG_DEVHANDLE (devhandle_t - wraps OFW, ACPI, etc. handles)

...and a sentinel value CFARG_EOL.

Add some extra sanity checking to ensure that interface attributes
aren't ambiguous.

Use CFARG_DEVHANDLE in MI FDT, OFW, and ACPI code, and macppc and shark
ports to associate those device handles with device_t instance. This
will trickle trough to more places over time (need back-end for pre-OFW
Sun OBP; any others?).

branches: 1.70.4;
Remove duplicate include of opt_ppb.h.
No binary changes.

Identify 16GT/s.

- Almost all ppbreg.h's definitions are also in pcireg.h. Remove duplicated
definitions from ppbreg.h and move some definitions from ppbreg.h to
pcireg.h.
- Change fast back-to-back "capable" to "enable" in pci_subr.c.
- Print Primary Discard Timer, Secondary Discard Timer, Discard Timer Status
and Discard Timer SERR# Enable bit in pci_subr.c.
- PCI_BRIDGE_PREFETCHBASE32_REG and PCI_BRIDGE_PREFETCHLIMIT32_REG are
"upper" 32bit registers, rename to *UP32_REG to avoid confusion.
- Use macro.

If the secondary bus is configured and the bus mastering is not enabled,
enable it. Suggested by thorpej@.

Explicitly enable bus masterling in case BIOS, UEFI or firmware don't enable
it. Might fix PR kern/53811.

Merge the [pgoyette-compat] branch

- defflag PPB_USEINTR
- Print "interrupting at "

branches: 1.63.2; 1.63.8; 1.63.10;
Fix typos.

- Fix a bug that a device which has no PCIe capability incorrectly
accessess the PCI config area in ppbdetach().
- Don't add event counters if slot interrupt isn't used.

Make ppb(4) interrupt support stable:
- Disable all interrupts in the beginning of attach. Without this, interrupt
storm occurs while cold == 1 on some environment.
- Disable command complete interrput for a while to prevent hangup on some
enviroment. I'm sorry, I don't know what this bit is :-|
- Check all status bits and return 0 if an interrupt is not for me. It's
required for INTx. Tested on XEN3_DOM0 because it doesn't support MSI yet.
- Return 1 when a interrupt is processed.

branches: 1.60.2;
Disable ppb(4)'s interrupt for a while. It causes hangup on some environment.
Define PPB_USEINTR if you'd like to use interrupt.

Disable and disestablish interrupt in ppbdetach().

in ppbdetach(), try to detach the children before tearing down our own state,
to avoid trouble if detaching the children fails.

Enable PCIe's interrupt as much as possilbe in ppb(4) to detect and count
status change event. HotPlug function itself have not implemented yet.

- Interrupt and each event are counted by evcnt(9). Example:

ppb0 Interrupt 0 0 intr
ppb0 Attention Button Pressed 0 0 misc
ppb0 Power Fault Detected 0 0 misc
ppb0 MRL Sensor Changed 0 0 misc
ppb0 Presence Detect Changed 0 0 misc
ppb0 Command Completed 0 0 misc
ppb0 Data Link Layer State Changed 0 0 misc

- Print message if ppb_printevent is not zero. The default vaule is 0.
The output messages:

Attention Button Pressed
Power Fault Detected
MRL Sensor Changed
Presence Detect Changed
Command Completed
Data Link Layer State Changed

- Remove workaround code to disable interrupt (ppb.c rev. 1.35).

Tested with Dell Latitude 2120 without if_bge.c rev. 1.304's workaround.
dmesg when bge's device timeout occured:

ppb3: Presence Detect Changed
ppb3: Data Link Layer State Changed
ppb3: Presence Detect Changed

vmstat -e |grep ppb

ppb3 Interrupt 2 0 intr
ppb3 Presence Detect Changed 2 0 misc
ppb3 Data Link Layer State Changed 1 0 misc

Move struct ppb_softc into ppbvar.h.

branches: 1.55.2; 1.55.4;
Define PCIE_XCAP_{VER,TYPE}(x) and use them.

branches: 1.54.2;
Rename PCIE_XCAP_VER_* macros to avoid confusion.

- Modify message of PCIe capability version. This field (PCIE_XCAP_VER_MASK)
is not specification's version number but the capability structure's version
number. To avoid confusion, print "PCI Express capability version x".
- The max number of PCIe lane is not 16 but 32. Fix the bug using with macro.
- Use macro instead of magic number.
- Gb/s -> GT/s

branches: 1.52.10;
Delete "PCI_" from PCIX and PICE capability registers.

ppb: fix link speed print

Print out negotiated link width and speed for PCIe (merged from
matt-nb5-mips64).

branches: 1.49.6;
extend the pci_aprint_devinfo slightly to cover the cases commonly
used by drivers: a short name for the quiet/naive case and a string
to override the "pcidevs" based name by one provided by the driver,
ride on yesterday's kernel minor version bump

put printing of the pci_devinfo into its own function (not inlined
by purpose) - this is a stack hog, and with this change my uTCA amd64
system boots again
a lot of similar code can be eliminated from pci device drivers this way,
but before doing so (and making the new function part of the module API)
I'd like to consider a modification to make it work with drivers which
prefer to print names from other sources (like pciide)

branches: 1.47.2; 1.47.6;
Tell a pci(4) instance its subordinate PCI buses using a new member
in the pcibus_attach_args, pba_sub. pciN attaches to pba_bus itself.
If pba_bus < pba_sub, then [pba_bus + 1, pba_sub] are subordinate to
pba_bus.

On i386, make mainbus0 attach pci0 with pba_sub = 255 because all buses
1 and up must be subordinate to pci0.

XXX Deal with other architectures.

Redefine PCI_MSI_* and PCI_PCIE_* constants in terms of bits(3).

Use named constants and more conventional variable names in
pci_msi_establish() and pci_msi_disestablish(). Fix a couple of bugs:
pci_msi_establish() returned a pointer to the struct intrhand instead of
to the struct msi_hdl as it was intended to, and pci_msi_disestablish()
did not free(9) the msi_hdl.

add missing break

ppb_fix_pcix changes:
- rename to ppb_fix_pcie
- support version PCI-E 2.0
- print version and device/port type information
- use constants from pcireg.h instead of magic numbers

changes:

ppb2 at pci0 dev 21 function 0: vendor 0x15ad product 0x07a0 (rev. 0x01)
ppb2: unsupported PCI Express version

to:

ppb2 at pci0 dev 21 function 0: vendor 0x15ad product 0x07a0 (rev. 0x01)
ppb2: PCI Express 2.0 <Root Port of PCI-E Root Complex>

On powerpc, recognize PCI Express RC root bridges.

branches: 1.42.2;
A pointer typedef entails trading too much flexibility to declare const
and non-const types, and the kernel uses both const and non-const
PMF qualifiers and device suspensors, so change the pmf_qual_t and
device_suspensor_t typedefs from "pointers to const" to non-pointer,
non-const types.

branches: 1.41.2;
Expand PMF_FN_* macros.

During shutdown, detach devices in an orderly fashion.

Call the detach routine for every device in the device tree, starting
with the leaves and moving toward the root, expecting that each
(pseudo-)device driver will use the opportunity to gracefully commit
outstandings transactions to the underlying (pseudo-)device and to
relinquish control of the hardware to the system BIOS.

Detaching devices is not suitable for every shutdown: in an emergency,
or if the system state is inconsistent, we should resort to a fast,
simple shutdown that uses only the pmf(9) shutdown hooks and the
(deprecated) shutdownhooks. For now, if the flag RB_NOSYNC is set in
boothowto, opt for the fast, simple shutdown.

Add a device flag, DVF_DETACH_SHUTDOWN, that indicates by its presence
that it is safe to detach a device during shutdown. Introduce macros
CFATTACH_DECL3() and CFATTACH_DECL3_NEW() for creating autoconf
attachments with default device flags. Add DVF_DETACH_SHUTDOWN
to configuration attachments for atabus(4), atw(4) at cardbus(4),
cardbus(4), cardslot(4), com(4) at isa(4), elanpar(4), elanpex(4),
elansc(4), gpio(4), npx(4) at isa(4), nsphyter(4), pci(4), pcib(4),
pcmcia(4), ppb(4), sip(4), wd(4), and wdc(4) at isa(4).

Add a device-detachment "reason" flag, DETACH_SHUTDOWN, that tells the
autoconf code and a device driver that the reason for detachment is
system shutdown.

Add a sysctl, kern.detachall, that tells the system to try to detach
every device at shutdown, regardless of any device's DVF_DETACH_SHUTDOWN
flag. The default for kern.detachall is 0. SET IT TO 1, PLEASE, TO
HELP TEST AND DEBUG DEVICE DETACHMENT AT SHUTDOWN.

This is a work in progress. In future work, I aim to treat
pseudo-devices more thoroughly, and to gracefully tear down a stack of
(pseudo-)disk drivers and filesystems, including cgd(4), vnd(4), and
raid(4) instances at shutdown.

Also commit some changes that are not easily untangled from the rest:

(1) begin to simplify device_t locking: rename struct pmf_private to
device_lock, and incorporate device_lock into struct device.

(2) #include <sys/device.h> in sys/pmf.h in order to get some
definitions that it needs. Stop unnecessarily #including <sys/device.h>
in sys/arch/x86/include/pic.h to keep the amd64, xen, and i386 releases
building.

branches: 1.39.8; 1.39.10; 1.39.14; 1.39.18; 1.39.22;
device_t / softc split. Tested on amd64.
"looks good" spz

unuspported -> unsupported

branches: 1.37.2; 1.37.4;
Add methods for detaching self and for detaching children.

Use device_t and accessors. Use aprint_*_dev().

branches: 1.36.6; 1.36.10;
Merge jmcneill-pm branch.

branches: 1.35.6; 1.35.8;
Merge 1.34.22.7 from jmcneill-pm:
Next attempt at trying to fix the irregular interrupt storms on my
Thinkpad: when we find a PCI Express device, check the list of
notification events and if any are sets, clear them. We can't handle
them ATM anyway.

branches: 1.34.8; 1.34.22; 1.34.24; 1.34.26;
__unused removal on arguments; approved by core.

- sprinkle __unused on function decls.
- fix a couple of unused bugs
- no more -Wno-unused for i386

branches: 1.32.20; 1.32.22;
merge ktrace-lwp.

branches: 1.31.2;
Use ANSI function decls and static.

de-__P

branches: 1.29.4; 1.29.6;
Phase out the use of a string as first "attach args" member to control
which bustype should be attached with a specific call to config_found()
(from a "mainbus" or a bus bridge).
Do it for isa/eisa/mca and pci/agp for now. These buses all attach to
an mi interface attribute "isabus", "eisabus" etc., and the autoconf
framework now allows to specify an interface attribute on config_found()
and config_search(), which limits the search of matching config data
to these which attach to that specific attribute.
So we basically have to call config_found_ia(..., "foobus", ...) where
such a bus is attached.
As a consequence, where a "mainbus" or alike also attaches other
devices (eg CPUs) which do not attach to a specific attribute yet,
we need at least pass an attribute name (different from "foobus") so
that the foo bus is not found at these places. This made some minor
changes necessary which are not obviously related to the mentioned buses.

pass string length (= boundary info) to pci_devinfo so that we do not run over
the end of memory region

Use aprint_*()

branches: 1.26.2;
Handle 64bit DMA addresses on PCI for platforms that can (currently only
enabled on amd64). Add a dmat64 field to various PCI attach structures,
and pass it down where needed. Implement a simple new function called
pci_dma64_available(pa) to test if 64bit DMA addresses may be used.
This returns 1 iff _PCI_HAVE_DMA64 is defined in <machine/pci_machdep.h>,
and there is more than 4G of memory.

Use aprint_normal() in cfprint routines.

Add trailing ; to CFATTACH_DECL.

Use CFATTACH_DECL().

Declare all cfattach structures const.

branches: 1.21.2;
* Add "pcitag_t *pba_bridgetag" to pci_attach_args. This is set to
NULL for root PCI busses. For busses behind a bridge, it points to
a persistent copy of the bridge's pcitag_t. This can be very useful
for machine-dependent PCI bus enumeration code.
* Implement a machine-dependent pci_enumerate_bus() for sparc64 which
uses OFW device nodes to enumerate the bus. When a PCI bus that is
behind a bridge is attached, pci_attach_hook() allocates a new PCI
chipset tag for the new bus and sets it's "curnode" to the OFW node
of the bridge. This is used as a starting point when enumerating
that bus. Root busses get the OFW node of the host bridge (psycho).
* Garbage-collect "ofpci" and "ofppb" from the sparc64 port.

add RCSID

branches: 1.19.6; 1.19.8;
Don't pass rd/mult capability though a PCI-PCI bridge. The bridge would
have to break it up into mutliple rd/line's anyhow, so why bother letting
the device issue it in the first place.

branches: 1.18.14; 1.18.16; 1.18.20;
Nuke __BROKEN_INDIRECT_CONFIG.

clean up slightly, correct a few comments

branches: 1.16.2;
Update for config changes.

branches: 1.15.2; 1.15.4;
Pass down bus_dma_tag_t's as appropriate (per Jason's bus_dma code).

Pass the I/O and memory enable flags through the bridge.

update these so they compile whether or not __BROKEN_INDIRECT_CONFIG
is defined.

New bus.h implementation/interface:
- No more distinction between i/o-mapped and memory-mapped
devices. It's all "bus space" now, and space tags
differentiate the space with finer grain than the
bus chipset tag.
- Add memory barrier methods.
- Implement space alloc/free methods.
- Implement region read/write methods (like memcpy to/from
bus space).
This interface provides a better abstraction for dealing with
machine-independent chipset drivers.

backout kprintf changes

printf -> kprintf, sprintf -> ksprintf

change cfprint_t type definition to take a const char *, rather than
a char *, because that's what was really intended, and because
if the print function modifies the string, various things could become
unhappy (so the string should _not_ be modified).

remove unused variables

modify these to provide a new, better-specified PCI interface
(soon to be documented on mailing lists; eventually in section 9 manual
pages), most importantly:
(1) support interrupt pin swizzling on non-i386 systems with
PCI-PCI bridges (per PPB spec; done, but meaningless, on i386).
(2) provide pci_{io,mem}_find(), to determine what I/O or memory
space is described by a given PCI configuration space
mapping register.
(3) provide pci_intr_map(), pci_intr_string(), and
pci_intr_{,dis}establish() to manipulate and print info about
PCI interrupts.
(4) make pci functions take as an argument a machine-dependent
cookie, to allow more flexibility in implementation.

spacing nit

New device attachment scheme:

- split softc size and match/attach out from cfdriver into
a new struct cfattach.

- new "attach" directive for files.*. May specify the name of
the cfattach structure, so that devices may be easily attached
to parents with different autoconfiguration semantics.

These devices don't actually need softc's that are supersets of 'struct
device.' No point in adding that complexity + space if they're not needed,
so axe them.

(1) provide #defines for cf_loc[] entries for devices that attach to
pcibus and pci.
(2) remove the #ifdef i386 from pci.c, and provide a machine-dependent
hook (pci_md_attach_hook()) to do any machine-dependent attachment
gunk, e.g. on the i386 printing out the configuration mode (if bus 0)
(3) don't pass max device number for a given bus in, use
PCI_MAX_DEVICE_NUMBER, which can be defined on a per-machine basis.
(defaults to 32. on i386, it's 32 if pci conf mode == 1, 16 if 2.)

fix aux argument to config_found() for secondary bus.

Preliminary support for PCI-PCI bridges. Recognize a PCI-PCI bridge
and attach the secondary pci bus as a 'pci' device. Note that this support
is incomplete and will not yet work for ports other than that i386. (The
i386 can rely on the PCI interrupt 'line' information to determine
interrupt mapping, which is not necessarily possible on other systems.)

Merge thorpej-cfargs2.

branches: 1.72.4;
Forward the bridge's devhandle to the downstream bus.

branches: 1.71.2; 1.71.4;
Merge thorpej-cfargs branch:

Simplify and make extensible the config_search() / config_found() /
config_attach() interfaces: rather than having different variants for
which arguments you want pass along, just have a single call that
takes a variadic list of tag-value arguments.

Adjust all call sites:
- Simplify wherever possible; don't pass along arguments that aren't
actually needed.
- Don't be explicit about what interface attribute is attaching if
the device only has one. (More simplification.)
- Add a config_probe() function to be used in indirect configuiration
situations, making is visibly easier to see when indirect config is
in play, and allowing for future change in semantics. (As of now,
this is just a wrapper around config_match(), but that is an
implementation detail.)

Remove unnecessary or redundant interface attributes where they're not
needed.

There are currently 5 "cfargs" defined:
- CFARG_SUBMATCH (submatch function for direct config)
- CFARG_SEARCH (search function for indirect config)
- CFARG_IATTR (interface attribte)
- CFARG_LOCATORS (locators array)
- CFARG_DEVHANDLE (devhandle_t - wraps OFW, ACPI, etc. handles)

...and a sentinel value CFARG_EOL.

Add some extra sanity checking to ensure that interface attributes
aren't ambiguous.

Use CFARG_DEVHANDLE in MI FDT, OFW, and ACPI code, and macppc and shark
ports to associate those device handles with device_t instance. This
will trickle trough to more places over time (need back-end for pre-OFW
Sun OBP; any others?).

branches: 1.70.4;
Remove duplicate include of opt_ppb.h.
No binary changes.

Identify 16GT/s.

- Almost all ppbreg.h's definitions are also in pcireg.h. Remove duplicated
definitions from ppbreg.h and move some definitions from ppbreg.h to
pcireg.h.
- Change fast back-to-back "capable" to "enable" in pci_subr.c.
- Print Primary Discard Timer, Secondary Discard Timer, Discard Timer Status
and Discard Timer SERR# Enable bit in pci_subr.c.
- PCI_BRIDGE_PREFETCHBASE32_REG and PCI_BRIDGE_PREFETCHLIMIT32_REG are
"upper" 32bit registers, rename to *UP32_REG to avoid confusion.
- Use macro.

If the secondary bus is configured and the bus mastering is not enabled,
enable it. Suggested by thorpej@.

Explicitly enable bus masterling in case BIOS, UEFI or firmware don't enable
it. Might fix PR kern/53811.

Merge the [pgoyette-compat] branch

- defflag PPB_USEINTR
- Print "interrupting at "

branches: 1.63.2; 1.63.8; 1.63.10;
Fix typos.

- Fix a bug that a device which has no PCIe capability incorrectly
accessess the PCI config area in ppbdetach().
- Don't add event counters if slot interrupt isn't used.

Make ppb(4) interrupt support stable:
- Disable all interrupts in the beginning of attach. Without this, interrupt
storm occurs while cold == 1 on some environment.
- Disable command complete interrput for a while to prevent hangup on some
enviroment. I'm sorry, I don't know what this bit is :-|
- Check all status bits and return 0 if an interrupt is not for me. It's
required for INTx. Tested on XEN3_DOM0 because it doesn't support MSI yet.
- Return 1 when a interrupt is processed.

branches: 1.60.2;
Disable ppb(4)'s interrupt for a while. It causes hangup on some environment.
Define PPB_USEINTR if you'd like to use interrupt.

Disable and disestablish interrupt in ppbdetach().

in ppbdetach(), try to detach the children before tearing down our own state,
to avoid trouble if detaching the children fails.

Enable PCIe's interrupt as much as possilbe in ppb(4) to detect and count
status change event. HotPlug function itself have not implemented yet.

- Interrupt and each event are counted by evcnt(9). Example:

ppb0 Interrupt 0 0 intr
ppb0 Attention Button Pressed 0 0 misc
ppb0 Power Fault Detected 0 0 misc
ppb0 MRL Sensor Changed 0 0 misc
ppb0 Presence Detect Changed 0 0 misc
ppb0 Command Completed 0 0 misc
ppb0 Data Link Layer State Changed 0 0 misc

- Print message if ppb_printevent is not zero. The default vaule is 0.
The output messages:

Attention Button Pressed
Power Fault Detected
MRL Sensor Changed
Presence Detect Changed
Command Completed
Data Link Layer State Changed

- Remove workaround code to disable interrupt (ppb.c rev. 1.35).

Tested with Dell Latitude 2120 without if_bge.c rev. 1.304's workaround.
dmesg when bge's device timeout occured:

ppb3: Presence Detect Changed
ppb3: Data Link Layer State Changed
ppb3: Presence Detect Changed

vmstat -e |grep ppb

ppb3 Interrupt 2 0 intr
ppb3 Presence Detect Changed 2 0 misc
ppb3 Data Link Layer State Changed 1 0 misc

Move struct ppb_softc into ppbvar.h.

branches: 1.55.2; 1.55.4;
Define PCIE_XCAP_{VER,TYPE}(x) and use them.

branches: 1.54.2;
Rename PCIE_XCAP_VER_* macros to avoid confusion.

- Modify message of PCIe capability version. This field (PCIE_XCAP_VER_MASK)
is not specification's version number but the capability structure's version
number. To avoid confusion, print "PCI Express capability version x".
- The max number of PCIe lane is not 16 but 32. Fix the bug using with macro.
- Use macro instead of magic number.
- Gb/s -> GT/s

branches: 1.52.10;
Delete "PCI_" from PCIX and PICE capability registers.

ppb: fix link speed print

Print out negotiated link width and speed for PCIe (merged from
matt-nb5-mips64).

branches: 1.49.6;
extend the pci_aprint_devinfo slightly to cover the cases commonly
used by drivers: a short name for the quiet/naive case and a string
to override the "pcidevs" based name by one provided by the driver,
ride on yesterday's kernel minor version bump

put printing of the pci_devinfo into its own function (not inlined
by purpose) - this is a stack hog, and with this change my uTCA amd64
system boots again
a lot of similar code can be eliminated from pci device drivers this way,
but before doing so (and making the new function part of the module API)
I'd like to consider a modification to make it work with drivers which
prefer to print names from other sources (like pciide)

branches: 1.47.2; 1.47.6;
Tell a pci(4) instance its subordinate PCI buses using a new member
in the pcibus_attach_args, pba_sub. pciN attaches to pba_bus itself.
If pba_bus < pba_sub, then [pba_bus + 1, pba_sub] are subordinate to
pba_bus.

On i386, make mainbus0 attach pci0 with pba_sub = 255 because all buses
1 and up must be subordinate to pci0.

XXX Deal with other architectures.

Redefine PCI_MSI_* and PCI_PCIE_* constants in terms of bits(3).

Use named constants and more conventional variable names in
pci_msi_establish() and pci_msi_disestablish(). Fix a couple of bugs:
pci_msi_establish() returned a pointer to the struct intrhand instead of
to the struct msi_hdl as it was intended to, and pci_msi_disestablish()
did not free(9) the msi_hdl.

add missing break

ppb_fix_pcix changes:
- rename to ppb_fix_pcie
- support version PCI-E 2.0
- print version and device/port type information
- use constants from pcireg.h instead of magic numbers

changes:

ppb2 at pci0 dev 21 function 0: vendor 0x15ad product 0x07a0 (rev. 0x01)
ppb2: unsupported PCI Express version

to:

ppb2 at pci0 dev 21 function 0: vendor 0x15ad product 0x07a0 (rev. 0x01)
ppb2: PCI Express 2.0 <Root Port of PCI-E Root Complex>

On powerpc, recognize PCI Express RC root bridges.

branches: 1.42.2;
A pointer typedef entails trading too much flexibility to declare const
and non-const types, and the kernel uses both const and non-const
PMF qualifiers and device suspensors, so change the pmf_qual_t and
device_suspensor_t typedefs from "pointers to const" to non-pointer,
non-const types.

branches: 1.41.2;
Expand PMF_FN_* macros.

During shutdown, detach devices in an orderly fashion.

Call the detach routine for every device in the device tree, starting
with the leaves and moving toward the root, expecting that each
(pseudo-)device driver will use the opportunity to gracefully commit
outstandings transactions to the underlying (pseudo-)device and to
relinquish control of the hardware to the system BIOS.

Detaching devices is not suitable for every shutdown: in an emergency,
or if the system state is inconsistent, we should resort to a fast,
simple shutdown that uses only the pmf(9) shutdown hooks and the
(deprecated) shutdownhooks. For now, if the flag RB_NOSYNC is set in
boothowto, opt for the fast, simple shutdown.

Add a device flag, DVF_DETACH_SHUTDOWN, that indicates by its presence
that it is safe to detach a device during shutdown. Introduce macros
CFATTACH_DECL3() and CFATTACH_DECL3_NEW() for creating autoconf
attachments with default device flags. Add DVF_DETACH_SHUTDOWN
to configuration attachments for atabus(4), atw(4) at cardbus(4),
cardbus(4), cardslot(4), com(4) at isa(4), elanpar(4), elanpex(4),
elansc(4), gpio(4), npx(4) at isa(4), nsphyter(4), pci(4), pcib(4),
pcmcia(4), ppb(4), sip(4), wd(4), and wdc(4) at isa(4).

Add a device-detachment "reason" flag, DETACH_SHUTDOWN, that tells the
autoconf code and a device driver that the reason for detachment is
system shutdown.

Add a sysctl, kern.detachall, that tells the system to try to detach
every device at shutdown, regardless of any device's DVF_DETACH_SHUTDOWN
flag. The default for kern.detachall is 0. SET IT TO 1, PLEASE, TO
HELP TEST AND DEBUG DEVICE DETACHMENT AT SHUTDOWN.

This is a work in progress. In future work, I aim to treat
pseudo-devices more thoroughly, and to gracefully tear down a stack of
(pseudo-)disk drivers and filesystems, including cgd(4), vnd(4), and
raid(4) instances at shutdown.

Also commit some changes that are not easily untangled from the rest:

(1) begin to simplify device_t locking: rename struct pmf_private to
device_lock, and incorporate device_lock into struct device.

(2) #include <sys/device.h> in sys/pmf.h in order to get some
definitions that it needs. Stop unnecessarily #including <sys/device.h>
in sys/arch/x86/include/pic.h to keep the amd64, xen, and i386 releases
building.

branches: 1.39.8; 1.39.10; 1.39.14; 1.39.18; 1.39.22;
device_t / softc split. Tested on amd64.
"looks good" spz

unuspported -> unsupported

branches: 1.37.2; 1.37.4;
Add methods for detaching self and for detaching children.

Use device_t and accessors. Use aprint_*_dev().

branches: 1.36.6; 1.36.10;
Merge jmcneill-pm branch.

branches: 1.35.6; 1.35.8;
Merge 1.34.22.7 from jmcneill-pm:
Next attempt at trying to fix the irregular interrupt storms on my
Thinkpad: when we find a PCI Express device, check the list of
notification events and if any are sets, clear them. We can't handle
them ATM anyway.

branches: 1.34.8; 1.34.22; 1.34.24; 1.34.26;
__unused removal on arguments; approved by core.

- sprinkle __unused on function decls.
- fix a couple of unused bugs
- no more -Wno-unused for i386

branches: 1.32.20; 1.32.22;
merge ktrace-lwp.

branches: 1.31.2;
Use ANSI function decls and static.

de-__P

branches: 1.29.4; 1.29.6;
Phase out the use of a string as first "attach args" member to control
which bustype should be attached with a specific call to config_found()
(from a "mainbus" or a bus bridge).
Do it for isa/eisa/mca and pci/agp for now. These buses all attach to
an mi interface attribute "isabus", "eisabus" etc., and the autoconf
framework now allows to specify an interface attribute on config_found()
and config_search(), which limits the search of matching config data
to these which attach to that specific attribute.
So we basically have to call config_found_ia(..., "foobus", ...) where
such a bus is attached.
As a consequence, where a "mainbus" or alike also attaches other
devices (eg CPUs) which do not attach to a specific attribute yet,
we need at least pass an attribute name (different from "foobus") so
that the foo bus is not found at these places. This made some minor
changes necessary which are not obviously related to the mentioned buses.

pass string length (= boundary info) to pci_devinfo so that we do not run over
the end of memory region

Use aprint_*()

branches: 1.26.2;
Handle 64bit DMA addresses on PCI for platforms that can (currently only
enabled on amd64). Add a dmat64 field to various PCI attach structures,
and pass it down where needed. Implement a simple new function called
pci_dma64_available(pa) to test if 64bit DMA addresses may be used.
This returns 1 iff _PCI_HAVE_DMA64 is defined in <machine/pci_machdep.h>,
and there is more than 4G of memory.

Use aprint_normal() in cfprint routines.

Add trailing ; to CFATTACH_DECL.

Use CFATTACH_DECL().

Declare all cfattach structures const.

branches: 1.21.2;
* Add "pcitag_t *pba_bridgetag" to pci_attach_args. This is set to
NULL for root PCI busses. For busses behind a bridge, it points to
a persistent copy of the bridge's pcitag_t. This can be very useful
for machine-dependent PCI bus enumeration code.
* Implement a machine-dependent pci_enumerate_bus() for sparc64 which
uses OFW device nodes to enumerate the bus. When a PCI bus that is
behind a bridge is attached, pci_attach_hook() allocates a new PCI
chipset tag for the new bus and sets it's "curnode" to the OFW node
of the bridge. This is used as a starting point when enumerating
that bus. Root busses get the OFW node of the host bridge (psycho).
* Garbage-collect "ofpci" and "ofppb" from the sparc64 port.

add RCSID

branches: 1.19.6; 1.19.8;
Don't pass rd/mult capability though a PCI-PCI bridge. The bridge would
have to break it up into mutliple rd/line's anyhow, so why bother letting
the device issue it in the first place.

branches: 1.18.14; 1.18.16; 1.18.20;
Nuke __BROKEN_INDIRECT_CONFIG.

clean up slightly, correct a few comments

branches: 1.16.2;
Update for config changes.

branches: 1.15.2; 1.15.4;
Pass down bus_dma_tag_t's as appropriate (per Jason's bus_dma code).

Pass the I/O and memory enable flags through the bridge.

update these so they compile whether or not __BROKEN_INDIRECT_CONFIG
is defined.

New bus.h implementation/interface:
- No more distinction between i/o-mapped and memory-mapped
devices. It's all "bus space" now, and space tags
differentiate the space with finer grain than the
bus chipset tag.
- Add memory barrier methods.
- Implement space alloc/free methods.
- Implement region read/write methods (like memcpy to/from
bus space).
This interface provides a better abstraction for dealing with
machine-independent chipset drivers.

backout kprintf changes

printf -> kprintf, sprintf -> ksprintf

change cfprint_t type definition to take a const char *, rather than
a char *, because that's what was really intended, and because
if the print function modifies the string, various things could become
unhappy (so the string should _not_ be modified).

remove unused variables

modify these to provide a new, better-specified PCI interface
(soon to be documented on mailing lists; eventually in section 9 manual
pages), most importantly:
(1) support interrupt pin swizzling on non-i386 systems with
PCI-PCI bridges (per PPB spec; done, but meaningless, on i386).
(2) provide pci_{io,mem}_find(), to determine what I/O or memory
space is described by a given PCI configuration space
mapping register.
(3) provide pci_intr_map(), pci_intr_string(), and
pci_intr_{,dis}establish() to manipulate and print info about
PCI interrupts.
(4) make pci functions take as an argument a machine-dependent
cookie, to allow more flexibility in implementation.

spacing nit

New device attachment scheme:

- split softc size and match/attach out from cfdriver into
a new struct cfattach.

- new "attach" directive for files.*. May specify the name of
the cfattach structure, so that devices may be easily attached
to parents with different autoconfiguration semantics.

These devices don't actually need softc's that are supersets of 'struct
device.' No point in adding that complexity + space if they're not needed,
so axe them.

(1) provide #defines for cf_loc[] entries for devices that attach to
pcibus and pci.
(2) remove the #ifdef i386 from pci.c, and provide a machine-dependent
hook (pci_md_attach_hook()) to do any machine-dependent attachment
gunk, e.g. on the i386 printing out the configuration mode (if bus 0)
(3) don't pass max device number for a given bus in, use
PCI_MAX_DEVICE_NUMBER, which can be defined on a per-machine basis.
(defaults to 32. on i386, it's 32 if pci conf mode == 1, 16 if 2.)

fix aux argument to config_found() for secondary bus.

Preliminary support for PCI-PCI bridges. Recognize a PCI-PCI bridge
and attach the secondary pci bus as a 'pci' device. Note that this support
is incomplete and will not yet work for ports other than that i386. (The
i386 can rely on the PCI interrupt 'line' information to determine
interrupt mapping, which is not necessarily possible on other systems.)

Forward the bridge's devhandle to the downstream bus.

Merge thorpej-cfargs branch:

Simplify and make extensible the config_search() / config_found() /
config_attach() interfaces: rather than having different variants for
which arguments you want pass along, just have a single call that
takes a variadic list of tag-value arguments.

Adjust all call sites:
- Simplify wherever possible; don't pass along arguments that aren't
actually needed.
- Don't be explicit about what interface attribute is attaching if
the device only has one. (More simplification.)
- Add a config_probe() function to be used in indirect configuiration
situations, making is visibly easier to see when indirect config is
in play, and allowing for future change in semantics. (As of now,
this is just a wrapper around config_match(), but that is an
implementation detail.)

Remove unnecessary or redundant interface attributes where they're not
needed.

There are currently 5 "cfargs" defined:
- CFARG_SUBMATCH (submatch function for direct config)
- CFARG_SEARCH (search function for indirect config)
- CFARG_IATTR (interface attribte)
- CFARG_LOCATORS (locators array)
- CFARG_DEVHANDLE (devhandle_t - wraps OFW, ACPI, etc. handles)

...and a sentinel value CFARG_EOL.

Add some extra sanity checking to ensure that interface attributes
aren't ambiguous.

Use CFARG_DEVHANDLE in MI FDT, OFW, and ACPI code, and macppc and shark
ports to associate those device handles with device_t instance. This
will trickle trough to more places over time (need back-end for pre-OFW
Sun OBP; any others?).

branches: 1.70.4;
Remove duplicate include of opt_ppb.h.
No binary changes.

Identify 16GT/s.

- Almost all ppbreg.h's definitions are also in pcireg.h. Remove duplicated
definitions from ppbreg.h and move some definitions from ppbreg.h to
pcireg.h.
- Change fast back-to-back "capable" to "enable" in pci_subr.c.
- Print Primary Discard Timer, Secondary Discard Timer, Discard Timer Status
and Discard Timer SERR# Enable bit in pci_subr.c.
- PCI_BRIDGE_PREFETCHBASE32_REG and PCI_BRIDGE_PREFETCHLIMIT32_REG are
"upper" 32bit registers, rename to *UP32_REG to avoid confusion.
- Use macro.

If the secondary bus is configured and the bus mastering is not enabled,
enable it. Suggested by thorpej@.

Explicitly enable bus masterling in case BIOS, UEFI or firmware don't enable
it. Might fix PR kern/53811.

Merge the [pgoyette-compat] branch

- defflag PPB_USEINTR
- Print "interrupting at "

branches: 1.63.2; 1.63.8; 1.63.10;
Fix typos.

- Fix a bug that a device which has no PCIe capability incorrectly
accessess the PCI config area in ppbdetach().
- Don't add event counters if slot interrupt isn't used.

Make ppb(4) interrupt support stable:
- Disable all interrupts in the beginning of attach. Without this, interrupt
storm occurs while cold == 1 on some environment.
- Disable command complete interrput for a while to prevent hangup on some
enviroment. I'm sorry, I don't know what this bit is :-|
- Check all status bits and return 0 if an interrupt is not for me. It's
required for INTx. Tested on XEN3_DOM0 because it doesn't support MSI yet.
- Return 1 when a interrupt is processed.

branches: 1.60.2;
Disable ppb(4)'s interrupt for a while. It causes hangup on some environment.
Define PPB_USEINTR if you'd like to use interrupt.

Disable and disestablish interrupt in ppbdetach().

in ppbdetach(), try to detach the children before tearing down our own state,
to avoid trouble if detaching the children fails.

Enable PCIe's interrupt as much as possilbe in ppb(4) to detect and count
status change event. HotPlug function itself have not implemented yet.

- Interrupt and each event are counted by evcnt(9). Example:

ppb0 Interrupt 0 0 intr
ppb0 Attention Button Pressed 0 0 misc
ppb0 Power Fault Detected 0 0 misc
ppb0 MRL Sensor Changed 0 0 misc
ppb0 Presence Detect Changed 0 0 misc
ppb0 Command Completed 0 0 misc
ppb0 Data Link Layer State Changed 0 0 misc

- Print message if ppb_printevent is not zero. The default vaule is 0.
The output messages:

Attention Button Pressed
Power Fault Detected
MRL Sensor Changed
Presence Detect Changed
Command Completed
Data Link Layer State Changed

- Remove workaround code to disable interrupt (ppb.c rev. 1.35).

Tested with Dell Latitude 2120 without if_bge.c rev. 1.304's workaround.
dmesg when bge's device timeout occured:

ppb3: Presence Detect Changed
ppb3: Data Link Layer State Changed
ppb3: Presence Detect Changed

vmstat -e |grep ppb

ppb3 Interrupt 2 0 intr
ppb3 Presence Detect Changed 2 0 misc
ppb3 Data Link Layer State Changed 1 0 misc

Move struct ppb_softc into ppbvar.h.

branches: 1.55.2; 1.55.4;
Define PCIE_XCAP_{VER,TYPE}(x) and use them.

branches: 1.54.2;
Rename PCIE_XCAP_VER_* macros to avoid confusion.

- Modify message of PCIe capability version. This field (PCIE_XCAP_VER_MASK)
is not specification's version number but the capability structure's version
number. To avoid confusion, print "PCI Express capability version x".
- The max number of PCIe lane is not 16 but 32. Fix the bug using with macro.
- Use macro instead of magic number.
- Gb/s -> GT/s

branches: 1.52.10;
Delete "PCI_" from PCIX and PICE capability registers.

ppb: fix link speed print

Print out negotiated link width and speed for PCIe (merged from
matt-nb5-mips64).

branches: 1.49.6;
extend the pci_aprint_devinfo slightly to cover the cases commonly
used by drivers: a short name for the quiet/naive case and a string
to override the "pcidevs" based name by one provided by the driver,
ride on yesterday's kernel minor version bump

put printing of the pci_devinfo into its own function (not inlined
by purpose) - this is a stack hog, and with this change my uTCA amd64
system boots again
a lot of similar code can be eliminated from pci device drivers this way,
but before doing so (and making the new function part of the module API)
I'd like to consider a modification to make it work with drivers which
prefer to print names from other sources (like pciide)

branches: 1.47.2; 1.47.6;
Tell a pci(4) instance its subordinate PCI buses using a new member
in the pcibus_attach_args, pba_sub. pciN attaches to pba_bus itself.
If pba_bus < pba_sub, then [pba_bus + 1, pba_sub] are subordinate to
pba_bus.

On i386, make mainbus0 attach pci0 with pba_sub = 255 because all buses
1 and up must be subordinate to pci0.

XXX Deal with other architectures.

Redefine PCI_MSI_* and PCI_PCIE_* constants in terms of bits(3).

Use named constants and more conventional variable names in
pci_msi_establish() and pci_msi_disestablish(). Fix a couple of bugs:
pci_msi_establish() returned a pointer to the struct intrhand instead of
to the struct msi_hdl as it was intended to, and pci_msi_disestablish()
did not free(9) the msi_hdl.

add missing break

ppb_fix_pcix changes:
- rename to ppb_fix_pcie
- support version PCI-E 2.0
- print version and device/port type information
- use constants from pcireg.h instead of magic numbers

changes:

ppb2 at pci0 dev 21 function 0: vendor 0x15ad product 0x07a0 (rev. 0x01)
ppb2: unsupported PCI Express version

to:

ppb2 at pci0 dev 21 function 0: vendor 0x15ad product 0x07a0 (rev. 0x01)
ppb2: PCI Express 2.0 <Root Port of PCI-E Root Complex>

On powerpc, recognize PCI Express RC root bridges.

branches: 1.42.2;
A pointer typedef entails trading too much flexibility to declare const
and non-const types, and the kernel uses both const and non-const
PMF qualifiers and device suspensors, so change the pmf_qual_t and
device_suspensor_t typedefs from "pointers to const" to non-pointer,
non-const types.

branches: 1.41.2;
Expand PMF_FN_* macros.

During shutdown, detach devices in an orderly fashion.

Call the detach routine for every device in the device tree, starting
with the leaves and moving toward the root, expecting that each
(pseudo-)device driver will use the opportunity to gracefully commit
outstandings transactions to the underlying (pseudo-)device and to
relinquish control of the hardware to the system BIOS.

Detaching devices is not suitable for every shutdown: in an emergency,
or if the system state is inconsistent, we should resort to a fast,
simple shutdown that uses only the pmf(9) shutdown hooks and the
(deprecated) shutdownhooks. For now, if the flag RB_NOSYNC is set in
boothowto, opt for the fast, simple shutdown.

Add a device flag, DVF_DETACH_SHUTDOWN, that indicates by its presence
that it is safe to detach a device during shutdown. Introduce macros
CFATTACH_DECL3() and CFATTACH_DECL3_NEW() for creating autoconf
attachments with default device flags. Add DVF_DETACH_SHUTDOWN
to configuration attachments for atabus(4), atw(4) at cardbus(4),
cardbus(4), cardslot(4), com(4) at isa(4), elanpar(4), elanpex(4),
elansc(4), gpio(4), npx(4) at isa(4), nsphyter(4), pci(4), pcib(4),
pcmcia(4), ppb(4), sip(4), wd(4), and wdc(4) at isa(4).

Add a device-detachment "reason" flag, DETACH_SHUTDOWN, that tells the
autoconf code and a device driver that the reason for detachment is
system shutdown.

Add a sysctl, kern.detachall, that tells the system to try to detach
every device at shutdown, regardless of any device's DVF_DETACH_SHUTDOWN
flag. The default for kern.detachall is 0. SET IT TO 1, PLEASE, TO
HELP TEST AND DEBUG DEVICE DETACHMENT AT SHUTDOWN.

This is a work in progress. In future work, I aim to treat
pseudo-devices more thoroughly, and to gracefully tear down a stack of
(pseudo-)disk drivers and filesystems, including cgd(4), vnd(4), and
raid(4) instances at shutdown.

Also commit some changes that are not easily untangled from the rest:

(1) begin to simplify device_t locking: rename struct pmf_private to
device_lock, and incorporate device_lock into struct device.

(2) #include <sys/device.h> in sys/pmf.h in order to get some
definitions that it needs. Stop unnecessarily #including <sys/device.h>
in sys/arch/x86/include/pic.h to keep the amd64, xen, and i386 releases
building.

branches: 1.39.8; 1.39.10; 1.39.14; 1.39.18; 1.39.22;
device_t / softc split. Tested on amd64.
"looks good" spz

unuspported -> unsupported

branches: 1.37.2; 1.37.4;
Add methods for detaching self and for detaching children.

Use device_t and accessors. Use aprint_*_dev().

branches: 1.36.6; 1.36.10;
Merge jmcneill-pm branch.

branches: 1.35.6; 1.35.8;
Merge 1.34.22.7 from jmcneill-pm:
Next attempt at trying to fix the irregular interrupt storms on my
Thinkpad: when we find a PCI Express device, check the list of
notification events and if any are sets, clear them. We can't handle
them ATM anyway.

branches: 1.34.8; 1.34.22; 1.34.24; 1.34.26;
__unused removal on arguments; approved by core.

- sprinkle __unused on function decls.
- fix a couple of unused bugs
- no more -Wno-unused for i386

branches: 1.32.20; 1.32.22;
merge ktrace-lwp.

branches: 1.31.2;
Use ANSI function decls and static.

de-__P

branches: 1.29.4; 1.29.6;
Phase out the use of a string as first "attach args" member to control
which bustype should be attached with a specific call to config_found()
(from a "mainbus" or a bus bridge).
Do it for isa/eisa/mca and pci/agp for now. These buses all attach to
an mi interface attribute "isabus", "eisabus" etc., and the autoconf
framework now allows to specify an interface attribute on config_found()
and config_search(), which limits the search of matching config data
to these which attach to that specific attribute.
So we basically have to call config_found_ia(..., "foobus", ...) where
such a bus is attached.
As a consequence, where a "mainbus" or alike also attaches other
devices (eg CPUs) which do not attach to a specific attribute yet,
we need at least pass an attribute name (different from "foobus") so
that the foo bus is not found at these places. This made some minor
changes necessary which are not obviously related to the mentioned buses.

pass string length (= boundary info) to pci_devinfo so that we do not run over
the end of memory region

Use aprint_*()

branches: 1.26.2;
Handle 64bit DMA addresses on PCI for platforms that can (currently only
enabled on amd64). Add a dmat64 field to various PCI attach structures,
and pass it down where needed. Implement a simple new function called
pci_dma64_available(pa) to test if 64bit DMA addresses may be used.
This returns 1 iff _PCI_HAVE_DMA64 is defined in <machine/pci_machdep.h>,
and there is more than 4G of memory.

Use aprint_normal() in cfprint routines.

Add trailing ; to CFATTACH_DECL.

Use CFATTACH_DECL().

Declare all cfattach structures const.

branches: 1.21.2;
* Add "pcitag_t *pba_bridgetag" to pci_attach_args. This is set to
NULL for root PCI busses. For busses behind a bridge, it points to
a persistent copy of the bridge's pcitag_t. This can be very useful
for machine-dependent PCI bus enumeration code.
* Implement a machine-dependent pci_enumerate_bus() for sparc64 which
uses OFW device nodes to enumerate the bus. When a PCI bus that is
behind a bridge is attached, pci_attach_hook() allocates a new PCI
chipset tag for the new bus and sets it's "curnode" to the OFW node
of the bridge. This is used as a starting point when enumerating
that bus. Root busses get the OFW node of the host bridge (psycho).
* Garbage-collect "ofpci" and "ofppb" from the sparc64 port.

add RCSID

branches: 1.19.6; 1.19.8;
Don't pass rd/mult capability though a PCI-PCI bridge. The bridge would
have to break it up into mutliple rd/line's anyhow, so why bother letting
the device issue it in the first place.

branches: 1.18.14; 1.18.16; 1.18.20;
Nuke __BROKEN_INDIRECT_CONFIG.

clean up slightly, correct a few comments

branches: 1.16.2;
Update for config changes.

branches: 1.15.2; 1.15.4;
Pass down bus_dma_tag_t's as appropriate (per Jason's bus_dma code).

Pass the I/O and memory enable flags through the bridge.

update these so they compile whether or not __BROKEN_INDIRECT_CONFIG
is defined.

New bus.h implementation/interface:
- No more distinction between i/o-mapped and memory-mapped
devices. It's all "bus space" now, and space tags
differentiate the space with finer grain than the
bus chipset tag.
- Add memory barrier methods.
- Implement space alloc/free methods.
- Implement region read/write methods (like memcpy to/from
bus space).
This interface provides a better abstraction for dealing with
machine-independent chipset drivers.

backout kprintf changes

printf -> kprintf, sprintf -> ksprintf

change cfprint_t type definition to take a const char *, rather than
a char *, because that's what was really intended, and because
if the print function modifies the string, various things could become
unhappy (so the string should _not_ be modified).

remove unused variables

modify these to provide a new, better-specified PCI interface
(soon to be documented on mailing lists; eventually in section 9 manual
pages), most importantly:
(1) support interrupt pin swizzling on non-i386 systems with
PCI-PCI bridges (per PPB spec; done, but meaningless, on i386).
(2) provide pci_{io,mem}_find(), to determine what I/O or memory
space is described by a given PCI configuration space
mapping register.
(3) provide pci_intr_map(), pci_intr_string(), and
pci_intr_{,dis}establish() to manipulate and print info about
PCI interrupts.
(4) make pci functions take as an argument a machine-dependent
cookie, to allow more flexibility in implementation.

spacing nit

New device attachment scheme:

- split softc size and match/attach out from cfdriver into
a new struct cfattach.

- new "attach" directive for files.*. May specify the name of
the cfattach structure, so that devices may be easily attached
to parents with different autoconfiguration semantics.

These devices don't actually need softc's that are supersets of 'struct
device.' No point in adding that complexity + space if they're not needed,
so axe them.

(1) provide #defines for cf_loc[] entries for devices that attach to
pcibus and pci.
(2) remove the #ifdef i386 from pci.c, and provide a machine-dependent
hook (pci_md_attach_hook()) to do any machine-dependent attachment
gunk, e.g. on the i386 printing out the configuration mode (if bus 0)
(3) don't pass max device number for a given bus in, use
PCI_MAX_DEVICE_NUMBER, which can be defined on a per-machine basis.
(defaults to 32. on i386, it's 32 if pci conf mode == 1, 16 if 2.)

fix aux argument to config_found() for secondary bus.

Preliminary support for PCI-PCI bridges. Recognize a PCI-PCI bridge
and attach the secondary pci bus as a 'pci' device. Note that this support
is incomplete and will not yet work for ports other than that i386. (The
i386 can rely on the PCI interrupt 'line' information to determine
interrupt mapping, which is not necessarily possible on other systems.)

Merge thorpej-cfargs branch:

Simplify and make extensible the config_search() / config_found() /
config_attach() interfaces: rather than having different variants for
which arguments you want pass along, just have a single call that
takes a variadic list of tag-value arguments.

Adjust all call sites:
- Simplify wherever possible; don't pass along arguments that aren't
actually needed.
- Don't be explicit about what interface attribute is attaching if
the device only has one. (More simplification.)
- Add a config_probe() function to be used in indirect configuiration
situations, making is visibly easier to see when indirect config is
in play, and allowing for future change in semantics. (As of now,
this is just a wrapper around config_match(), but that is an
implementation detail.)

Remove unnecessary or redundant interface attributes where they're not
needed.

There are currently 5 "cfargs" defined:
- CFARG_SUBMATCH (submatch function for direct config)
- CFARG_SEARCH (search function for indirect config)
- CFARG_IATTR (interface attribte)
- CFARG_LOCATORS (locators array)
- CFARG_DEVHANDLE (devhandle_t - wraps OFW, ACPI, etc. handles)

...and a sentinel value CFARG_EOL.

Add some extra sanity checking to ensure that interface attributes
aren't ambiguous.

Use CFARG_DEVHANDLE in MI FDT, OFW, and ACPI code, and macppc and shark
ports to associate those device handles with device_t instance. This
will trickle trough to more places over time (need back-end for pre-OFW
Sun OBP; any others?).

branches: 1.70.4;
Remove duplicate include of opt_ppb.h.
No binary changes.

Identify 16GT/s.

- Almost all ppbreg.h's definitions are also in pcireg.h. Remove duplicated
definitions from ppbreg.h and move some definitions from ppbreg.h to
pcireg.h.
- Change fast back-to-back "capable" to "enable" in pci_subr.c.
- Print Primary Discard Timer, Secondary Discard Timer, Discard Timer Status
and Discard Timer SERR# Enable bit in pci_subr.c.
- PCI_BRIDGE_PREFETCHBASE32_REG and PCI_BRIDGE_PREFETCHLIMIT32_REG are
"upper" 32bit registers, rename to *UP32_REG to avoid confusion.
- Use macro.

If the secondary bus is configured and the bus mastering is not enabled,
enable it. Suggested by thorpej@.

Explicitly enable bus masterling in case BIOS, UEFI or firmware don't enable
it. Might fix PR kern/53811.

Merge the [pgoyette-compat] branch

- defflag PPB_USEINTR
- Print "interrupting at "

branches: 1.63.2; 1.63.8; 1.63.10;
Fix typos.

- Fix a bug that a device which has no PCIe capability incorrectly
accessess the PCI config area in ppbdetach().
- Don't add event counters if slot interrupt isn't used.

Make ppb(4) interrupt support stable:
- Disable all interrupts in the beginning of attach. Without this, interrupt
storm occurs while cold == 1 on some environment.
- Disable command complete interrput for a while to prevent hangup on some
enviroment. I'm sorry, I don't know what this bit is :-|
- Check all status bits and return 0 if an interrupt is not for me. It's
required for INTx. Tested on XEN3_DOM0 because it doesn't support MSI yet.
- Return 1 when a interrupt is processed.

branches: 1.60.2;
Disable ppb(4)'s interrupt for a while. It causes hangup on some environment.
Define PPB_USEINTR if you'd like to use interrupt.

Disable and disestablish interrupt in ppbdetach().

in ppbdetach(), try to detach the children before tearing down our own state,
to avoid trouble if detaching the children fails.

Enable PCIe's interrupt as much as possilbe in ppb(4) to detect and count
status change event. HotPlug function itself have not implemented yet.

- Interrupt and each event are counted by evcnt(9). Example:

ppb0 Interrupt 0 0 intr
ppb0 Attention Button Pressed 0 0 misc
ppb0 Power Fault Detected 0 0 misc
ppb0 MRL Sensor Changed 0 0 misc
ppb0 Presence Detect Changed 0 0 misc
ppb0 Command Completed 0 0 misc
ppb0 Data Link Layer State Changed 0 0 misc

- Print message if ppb_printevent is not zero. The default vaule is 0.
The output messages:

Attention Button Pressed
Power Fault Detected
MRL Sensor Changed
Presence Detect Changed
Command Completed
Data Link Layer State Changed

- Remove workaround code to disable interrupt (ppb.c rev. 1.35).

Tested with Dell Latitude 2120 without if_bge.c rev. 1.304's workaround.
dmesg when bge's device timeout occured:

ppb3: Presence Detect Changed
ppb3: Data Link Layer State Changed
ppb3: Presence Detect Changed

vmstat -e |grep ppb

ppb3 Interrupt 2 0 intr
ppb3 Presence Detect Changed 2 0 misc
ppb3 Data Link Layer State Changed 1 0 misc

Move struct ppb_softc into ppbvar.h.

branches: 1.55.2; 1.55.4;
Define PCIE_XCAP_{VER,TYPE}(x) and use them.

branches: 1.54.2;
Rename PCIE_XCAP_VER_* macros to avoid confusion.

- Modify message of PCIe capability version. This field (PCIE_XCAP_VER_MASK)
is not specification's version number but the capability structure's version
number. To avoid confusion, print "PCI Express capability version x".
- The max number of PCIe lane is not 16 but 32. Fix the bug using with macro.
- Use macro instead of magic number.
- Gb/s -> GT/s

branches: 1.52.10;
Delete "PCI_" from PCIX and PICE capability registers.

ppb: fix link speed print

Print out negotiated link width and speed for PCIe (merged from
matt-nb5-mips64).

branches: 1.49.6;
extend the pci_aprint_devinfo slightly to cover the cases commonly
used by drivers: a short name for the quiet/naive case and a string
to override the "pcidevs" based name by one provided by the driver,
ride on yesterday's kernel minor version bump

put printing of the pci_devinfo into its own function (not inlined
by purpose) - this is a stack hog, and with this change my uTCA amd64
system boots again
a lot of similar code can be eliminated from pci device drivers this way,
but before doing so (and making the new function part of the module API)
I'd like to consider a modification to make it work with drivers which
prefer to print names from other sources (like pciide)

branches: 1.47.2; 1.47.6;
Tell a pci(4) instance its subordinate PCI buses using a new member
in the pcibus_attach_args, pba_sub. pciN attaches to pba_bus itself.
If pba_bus < pba_sub, then [pba_bus + 1, pba_sub] are subordinate to
pba_bus.

On i386, make mainbus0 attach pci0 with pba_sub = 255 because all buses
1 and up must be subordinate to pci0.

XXX Deal with other architectures.

Redefine PCI_MSI_* and PCI_PCIE_* constants in terms of bits(3).

Use named constants and more conventional variable names in
pci_msi_establish() and pci_msi_disestablish(). Fix a couple of bugs:
pci_msi_establish() returned a pointer to the struct intrhand instead of
to the struct msi_hdl as it was intended to, and pci_msi_disestablish()
did not free(9) the msi_hdl.

add missing break

ppb_fix_pcix changes:
- rename to ppb_fix_pcie
- support version PCI-E 2.0
- print version and device/port type information
- use constants from pcireg.h instead of magic numbers

changes:

ppb2 at pci0 dev 21 function 0: vendor 0x15ad product 0x07a0 (rev. 0x01)
ppb2: unsupported PCI Express version

to:

ppb2 at pci0 dev 21 function 0: vendor 0x15ad product 0x07a0 (rev. 0x01)
ppb2: PCI Express 2.0 <Root Port of PCI-E Root Complex>

On powerpc, recognize PCI Express RC root bridges.

branches: 1.42.2;
A pointer typedef entails trading too much flexibility to declare const
and non-const types, and the kernel uses both const and non-const
PMF qualifiers and device suspensors, so change the pmf_qual_t and
device_suspensor_t typedefs from "pointers to const" to non-pointer,
non-const types.

branches: 1.41.2;
Expand PMF_FN_* macros.

During shutdown, detach devices in an orderly fashion.

Call the detach routine for every device in the device tree, starting
with the leaves and moving toward the root, expecting that each
(pseudo-)device driver will use the opportunity to gracefully commit
outstandings transactions to the underlying (pseudo-)device and to
relinquish control of the hardware to the system BIOS.

Detaching devices is not suitable for every shutdown: in an emergency,
or if the system state is inconsistent, we should resort to a fast,
simple shutdown that uses only the pmf(9) shutdown hooks and the
(deprecated) shutdownhooks. For now, if the flag RB_NOSYNC is set in
boothowto, opt for the fast, simple shutdown.

Add a device flag, DVF_DETACH_SHUTDOWN, that indicates by its presence
that it is safe to detach a device during shutdown. Introduce macros
CFATTACH_DECL3() and CFATTACH_DECL3_NEW() for creating autoconf
attachments with default device flags. Add DVF_DETACH_SHUTDOWN
to configuration attachments for atabus(4), atw(4) at cardbus(4),
cardbus(4), cardslot(4), com(4) at isa(4), elanpar(4), elanpex(4),
elansc(4), gpio(4), npx(4) at isa(4), nsphyter(4), pci(4), pcib(4),
pcmcia(4), ppb(4), sip(4), wd(4), and wdc(4) at isa(4).

Add a device-detachment "reason" flag, DETACH_SHUTDOWN, that tells the
autoconf code and a device driver that the reason for detachment is
system shutdown.

Add a sysctl, kern.detachall, that tells the system to try to detach
every device at shutdown, regardless of any device's DVF_DETACH_SHUTDOWN
flag. The default for kern.detachall is 0. SET IT TO 1, PLEASE, TO
HELP TEST AND DEBUG DEVICE DETACHMENT AT SHUTDOWN.

This is a work in progress. In future work, I aim to treat
pseudo-devices more thoroughly, and to gracefully tear down a stack of
(pseudo-)disk drivers and filesystems, including cgd(4), vnd(4), and
raid(4) instances at shutdown.

Also commit some changes that are not easily untangled from the rest:

(1) begin to simplify device_t locking: rename struct pmf_private to
device_lock, and incorporate device_lock into struct device.

(2) #include <sys/device.h> in sys/pmf.h in order to get some
definitions that it needs. Stop unnecessarily #including <sys/device.h>
in sys/arch/x86/include/pic.h to keep the amd64, xen, and i386 releases
building.

branches: 1.39.8; 1.39.10; 1.39.14; 1.39.18; 1.39.22;
device_t / softc split. Tested on amd64.
"looks good" spz

unuspported -> unsupported

branches: 1.37.2; 1.37.4;
Add methods for detaching self and for detaching children.

Use device_t and accessors. Use aprint_*_dev().

branches: 1.36.6; 1.36.10;
Merge jmcneill-pm branch.

branches: 1.35.6; 1.35.8;
Merge 1.34.22.7 from jmcneill-pm:
Next attempt at trying to fix the irregular interrupt storms on my
Thinkpad: when we find a PCI Express device, check the list of
notification events and if any are sets, clear them. We can't handle
them ATM anyway.

branches: 1.34.8; 1.34.22; 1.34.24; 1.34.26;
__unused removal on arguments; approved by core.

- sprinkle __unused on function decls.
- fix a couple of unused bugs
- no more -Wno-unused for i386

branches: 1.32.20; 1.32.22;
merge ktrace-lwp.

branches: 1.31.2;
Use ANSI function decls and static.

de-__P

branches: 1.29.4; 1.29.6;
Phase out the use of a string as first "attach args" member to control
which bustype should be attached with a specific call to config_found()
(from a "mainbus" or a bus bridge).
Do it for isa/eisa/mca and pci/agp for now. These buses all attach to
an mi interface attribute "isabus", "eisabus" etc., and the autoconf
framework now allows to specify an interface attribute on config_found()
and config_search(), which limits the search of matching config data
to these which attach to that specific attribute.
So we basically have to call config_found_ia(..., "foobus", ...) where
such a bus is attached.
As a consequence, where a "mainbus" or alike also attaches other
devices (eg CPUs) which do not attach to a specific attribute yet,
we need at least pass an attribute name (different from "foobus") so
that the foo bus is not found at these places. This made some minor
changes necessary which are not obviously related to the mentioned buses.

pass string length (= boundary info) to pci_devinfo so that we do not run over
the end of memory region

Use aprint_*()

branches: 1.26.2;
Handle 64bit DMA addresses on PCI for platforms that can (currently only
enabled on amd64). Add a dmat64 field to various PCI attach structures,
and pass it down where needed. Implement a simple new function called
pci_dma64_available(pa) to test if 64bit DMA addresses may be used.
This returns 1 iff _PCI_HAVE_DMA64 is defined in <machine/pci_machdep.h>,
and there is more than 4G of memory.

Use aprint_normal() in cfprint routines.

Add trailing ; to CFATTACH_DECL.

Use CFATTACH_DECL().

Declare all cfattach structures const.

branches: 1.21.2;
* Add "pcitag_t *pba_bridgetag" to pci_attach_args. This is set to
NULL for root PCI busses. For busses behind a bridge, it points to
a persistent copy of the bridge's pcitag_t. This can be very useful
for machine-dependent PCI bus enumeration code.
* Implement a machine-dependent pci_enumerate_bus() for sparc64 which
uses OFW device nodes to enumerate the bus. When a PCI bus that is
behind a bridge is attached, pci_attach_hook() allocates a new PCI
chipset tag for the new bus and sets it's "curnode" to the OFW node
of the bridge. This is used as a starting point when enumerating
that bus. Root busses get the OFW node of the host bridge (psycho).
* Garbage-collect "ofpci" and "ofppb" from the sparc64 port.

add RCSID

branches: 1.19.6; 1.19.8;
Don't pass rd/mult capability though a PCI-PCI bridge. The bridge would
have to break it up into mutliple rd/line's anyhow, so why bother letting
the device issue it in the first place.

branches: 1.18.14; 1.18.16; 1.18.20;
Nuke __BROKEN_INDIRECT_CONFIG.

clean up slightly, correct a few comments

branches: 1.16.2;
Update for config changes.

branches: 1.15.2; 1.15.4;
Pass down bus_dma_tag_t's as appropriate (per Jason's bus_dma code).

Pass the I/O and memory enable flags through the bridge.

update these so they compile whether or not __BROKEN_INDIRECT_CONFIG
is defined.

New bus.h implementation/interface:
- No more distinction between i/o-mapped and memory-mapped
devices. It's all "bus space" now, and space tags
differentiate the space with finer grain than the
bus chipset tag.
- Add memory barrier methods.
- Implement space alloc/free methods.
- Implement region read/write methods (like memcpy to/from
bus space).
This interface provides a better abstraction for dealing with
machine-independent chipset drivers.

backout kprintf changes

printf -> kprintf, sprintf -> ksprintf

change cfprint_t type definition to take a const char *, rather than
a char *, because that's what was really intended, and because
if the print function modifies the string, various things could become
unhappy (so the string should _not_ be modified).

remove unused variables

modify these to provide a new, better-specified PCI interface
(soon to be documented on mailing lists; eventually in section 9 manual
pages), most importantly:
(1) support interrupt pin swizzling on non-i386 systems with
PCI-PCI bridges (per PPB spec; done, but meaningless, on i386).
(2) provide pci_{io,mem}_find(), to determine what I/O or memory
space is described by a given PCI configuration space
mapping register.
(3) provide pci_intr_map(), pci_intr_string(), and
pci_intr_{,dis}establish() to manipulate and print info about
PCI interrupts.
(4) make pci functions take as an argument a machine-dependent
cookie, to allow more flexibility in implementation.

spacing nit

New device attachment scheme:

- split softc size and match/attach out from cfdriver into
a new struct cfattach.

- new "attach" directive for files.*. May specify the name of
the cfattach structure, so that devices may be easily attached
to parents with different autoconfiguration semantics.

These devices don't actually need softc's that are supersets of 'struct
device.' No point in adding that complexity + space if they're not needed,
so axe them.

(1) provide #defines for cf_loc[] entries for devices that attach to
pcibus and pci.
(2) remove the #ifdef i386 from pci.c, and provide a machine-dependent
hook (pci_md_attach_hook()) to do any machine-dependent attachment
gunk, e.g. on the i386 printing out the configuration mode (if bus 0)
(3) don't pass max device number for a given bus in, use
PCI_MAX_DEVICE_NUMBER, which can be defined on a per-machine basis.
(defaults to 32. on i386, it's 32 if pci conf mode == 1, 16 if 2.)

fix aux argument to config_found() for secondary bus.

Preliminary support for PCI-PCI bridges. Recognize a PCI-PCI bridge
and attach the secondary pci bus as a 'pci' device. Note that this support
is incomplete and will not yet work for ports other than that i386. (The
i386 can rely on the PCI interrupt 'line' information to determine
interrupt mapping, which is not necessarily possible on other systems.)

Remove duplicate include of opt_ppb.h.
No binary changes.

Identify 16GT/s.

- Almost all ppbreg.h's definitions are also in pcireg.h. Remove duplicated
definitions from ppbreg.h and move some definitions from ppbreg.h to
pcireg.h.
- Change fast back-to-back "capable" to "enable" in pci_subr.c.
- Print Primary Discard Timer, Secondary Discard Timer, Discard Timer Status
and Discard Timer SERR# Enable bit in pci_subr.c.
- PCI_BRIDGE_PREFETCHBASE32_REG and PCI_BRIDGE_PREFETCHLIMIT32_REG are
"upper" 32bit registers, rename to *UP32_REG to avoid confusion.
- Use macro.

If the secondary bus is configured and the bus mastering is not enabled,
enable it. Suggested by thorpej@.