MFC r308529: intpm: clean up intsmb_bread and intsmb_pcall

MFC r306218,306290: amdsbwd, intpm: unify bits specific to AMD chipsets

MFC r305602: intpm: fix attachment to supported AMD FCHs

MFC r305606: intpm: make sure to register smbus driver before intpm driver

MFC r305604: intpm: better clean up resources after a failed attachment

MFC r305603: intpm: do not try attaching to unsupported controller revisions

MFC r304674: intpm: add support for SB800

MFC: r271864

Move rl(4) to dev/rl.

MFC r263957:
Increase the number of TX DMA segments from 32 to 35. It turned
out 32 is not enough to support a full sized TSO packet.
While I'm here fix a long standing bug introduced in r169632 in
bce(4) where it didn't include L2 header length of TSO packet in
the maximum DMA segment size calculation.

MFC: r261531

- Implement the RX EARLYOFF and RXDV GATED bits as done by RealTek's Linux
driver as iof version 8.037.00 for RTL8168{E-VL,EP,F,G,GU} and RTL8411B.
This makes reception of packets work with the RTL8168G (HW rev. 0x4c000000)
in my Shuttle DS47.
- Consistently use RL_MSI_MESSAGES.
In joint forces with: yongari

MFC: r261529

Try to make the style used here consistent.

MFC r257306:
Add preliminary support for RTL8168EP.
Approved by: re (delphij)

MFC r257305:
Add preliminary support for RTL8168G, RTL8168GU and RTL8411B.
RTL8168GU has two variants(GMII and MII) but it uses the same chip
revision id. Driver checks PCI device id of controller and
sets internal capability flag(i.e. jumbo frame and link speed down
in WOL).
Approved by: re (delphij)

MFC r256828:
Add preliminary support for RTL8106E PCIe FastEthernet.
Approved by: re (delphij)

r256827:
Correct MAC revision bits. Previously it always cleared bit 20 and
bit 21.
Approved by: re (delphij)

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

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

Update PCI drivers to no longer look at the MEMIO-enabled bit in the PCI
command register. The lazy BAR allocation code in FreeBSD sometimes
disables this bit when it detects a range conflict, and will re-enable
it on demand when a driver allocates the BAR. Thus, the bit is no longer
a reliable indication of capability, and should not be checked. This
results in the elimination of a lot of code from drivers, and also gives
the opportunity to simplify a lot of drivers to use a helper API to set
the busmaster enable bit.

This changes fixes some recent reports of disk controllers and their
associated drives/enclosures disappearing during boot.

Submitted by: jhb
Reviewed by: jfv, marius, achadd, achim
MFC after: 1 day

- Corrrect mispellings of word useful

Submitted by: Christoph Mallon <christoph.mallon@gmx.de> (via private mail)

Fix build.

Most other giant locked storage drivers in the tree don't use
splsoftclock to note the need for future locking, so remove it from
here.

Kill now-bogus splhigh() -- it is a nop

Mark 'ticks', 'time_second', and 'time_uptime' as volatile to prevent the
compiler from caching their values in tight loops.

Reviewed by: bde
MFC after: 1 week

Add D-Link DFE-520TX rev C1.

Tested by: Ruslan Makhmatkhanov < cvs-src <> yandex dot ru >
MFC After: 1 week

Mechanically substitute flags from historic mbuf allocator with
malloc(9) flags within sys.

Exceptions:

- sys/contrib not touched
- sys/mbuf.h edited manually

Remove duplicate const specifiers in many drivers (I hope I got all of
them, please let me know if not). Most of these are of the form:

static const struct bzzt_type {
[...list of members...]
} const bzzt_devs[] = {
[...list of initializers...]
};

The second const is unnecessary, as arrays cannot be modified anyway,
and if the elements are const, the whole thing is const automatically
(e.g. it is placed in .rodata).

I have verified this does not change the binary output of a full kernel
build (except for build timestamps embedded in the object files).

Reviewed by: yongari, marius
MFC after: 1 week

intpm: add ATI IXP400 pci id

PR: kern/136762
Submitted by: Aurelien Mere <freebsd@amc-os.com>
Tested by: Jens Link <jens.link@gmx.de>
MFC after: 5 days

intpm: return only SMB bus error codes from SMB methods

PR: kern/25733
MFC after: 5 days

intpm: reflect the fact that SB800 and later AMD chipsets are not supported

They do not have compatible configuration registers in PCI configuration
space. Instead their configuration resides in AMD "PM I/O" space
(accessed via a pair of I/O space registers).

MFC after: 5 days

Use correct Config registers for RTL8139 family. Unlike RTL8168 and
RTL810x family , RTL8139 has different register map for Config
registers.

While here, follow the lead of re(4) in WOL configuration.
- Disable WOL_UCAST and WOL_MCAST capabilities by default.
- Config5 register write does not need to unlock EEPROM access
on RTL8139 family but unlocking EEPROM access does not affect
its operation and make it consistent with re(4).

Reported by: Matt Renzelmann mjr <> cs dot wisc dot edu

To save more power, switch to 10/100Mbps link when controller is
put into suspend/shutdown. Old PCI controllers performed that
operation in firmware but for RTL8111C or newer controllers, it's
responsibility of driver. It's not clear whether the firmware of
RTL8111B still downgrades its speed to 10/100Mbps so leave it as it
was.

Make sure to stop TX MAC before freeing queued TX frames.
For RTL8111DP, check if the TX MAC is active by reading RL_GTXSTART
register. For RTL8402/8168E-VL/8168F/8411, wait until TX queue is
empty.

- There's no need to overwrite the default device method with the default
one. Interestingly, these are actually the default for quite some time
(bus_generic_driver_added(9) since r52045 and bus_generic_print_child(9)
since r52045) but even recently added device drivers do this unnecessarily.
Discussed with: jhb, marcel
- While at it, use DEVMETHOD_END.
Discussed with: jhb
- Also while at it, use __FBSDID.

Add preliminary support for RTL8168/8111F PCIe Gigabit ethernet.

H/W donated by: RealTek Semiconductor Corp.

Add preliminary support for second generation RTL8105E PCIe
FastEthernet.

H/W donated by: RealTek Semiconductor Corp.

Disable PCIe ASPM (Active State Power Management) for all
controllers.
More and more RealTek controllers started to implement EEE feature.
Vendor driver seems to load a kind of firmware for EEE with
additional PHY fixups. It is known that the EEE feature may need
ASPM support. Unfortunately there is no documentation for EEE of
the controller so enabling ASPM may cause more problems.

Add preliminary support for RTL8411 PCIe Gigabit ethernet with
integrated card reader.

H/W donated by: RealTek Semiconductor Corp.

Add preliminary support for RTL8402 PCIe FastEthernet with
integrated card reader.

H/W donated by: RealTek Semiconductor Corp.

- Import the common MII bitbang'ing code from NetBSD and convert drivers to
take advantage of it instead of duplicating it. This reduces the size of
the i386 GENERIC kernel by about 4k. The only potential in-tree user left
unconverted is xe(4), which generally should be changed to use miibus(4)
instead of implementing PHY handling on its own, as otherwise it makes not
much sense to add a dependency on miibus(4)/mii_bitbang(4) to xe(4) just
for the MII bitbang'ing code. The common MII bitbang'ing code also is
useful in the embedded space for using GPIO pins to implement MII access.
- Based on lessons learnt with dc(4) (see r185750), add bus barriers to the
MII bitbang read and write functions of the other drivers converted in
order to ensure the intended ordering. Given that register access via an
index register as well as register bank/window switching is subject to the
same problem, also add bus barriers to the respective functions of smc(4),
tl(4) and xl(4).
- Sprinkle some const.

Thanks to the following testers:
Andrew Bliznak (nge(4)), nwhitehorn@ (bm(4)), yongari@ (sis(4) and ste(4))
Thanks to Hans-Joerg Sirtl for supplying hardware to test stge(4).

Reviewed by: yongari (subset of drivers)
Obtained from: NetBSD (partially)

Close a race where SIOCGIFMEDIA ioctl get inconsistent link status.
Because driver is accessing a common MII structure in
mii_pollstat(), updating user supplied structure should be done
before dropping a driver lock.

Reported by: Karim (fodillemlinkarimi <> gmail dot com)

Add new device id of D-Link DGE-530T Rev. C controller. DGE-503T
Rev A1 and B1 is supported by sk(4) but the DGE-530T Rev. C
controller is re-branded RealTek 8169 controller.

PR: kern/159116
Approved by: re (kib)

Do a sweep of the tree replacing calls to pci_find_extcap() with calls to
pci_find_cap() instead.

Add initial support for RTL8401E PCIe Fast Ethernet.

PR: 154789

Add support for RTL8105E PCIe Fast Ethernet controller. It seems
the controller has a kind of embedded controller/memory and vendor
applies a large set of magic code via undocumented PHY registers in
device initialization stage. I guess it's a firmware image for the
embedded controller in RTL8105E since the code is too big compared
to other DSP fixups. However I have no idea what that magic code
does and what's purpose of the embedded controller. Fortunately
driver seems to still work without loading the firmware.

While I'm here change device description of RTL810xE controller.

H/W donated by: Realtek Semiconductor Corp.

Do not use interrupt taskqueue on controllers with MSI/MSI-X
capability. One of reason using interrupt taskqueue in re(4) was
to reduce number of TX/RX interrupts under load because re(4)
controllers have no good TX/RX interrupt moderation mechanism.
Basic TX interrupt moderation is done by hardware for most
controllers but RX interrupt moderation through undocumented
register showed poor RX performance so it was disabled in r215025.
Using taskqueue to handle RX interrupt greatly reduced number of
interrupts but re(4) consumed all available CPU cycles to run the
taskqueue under high TX/RX network load. This can happen even with
RTL810x fast ethernet controller and I believe this is not
acceptable for most systems.

To mitigate the issue, use one-shot timer register to moderate RX
interrupts. The timer register provides programmable one-shot timer
and can be used to suppress interrupt generation. The timer runs at
125MHZ on PCIe controllers so the minimum time allowed for the
timer is 8ns. Data sheet says the register is 32 bits but
experimentation shows only lower 13 bits are valid so maximum time
that can be programmed is 65.528us. This yields theoretical maximum
number of RX interrupts that could be generated per second is about
15260. Combined with TX completion interrupts re(4) shall generate
less than 20k interrupts. This number is still slightly high
compared to other intelligent ethernet controllers but system is
very responsive even under high network load.

Introduce sysctl variable dev.re.%d.int_rx_mod that controls amount
of time to delay RX interrupt processing in units of us. Value 0
completely disables RX interrupt moderation. To provide old
behavior for controllers that have MSI/MSI-X capability, introduce
a new tunable hw.re.intr_filter. If the tunable is set to non-zero
value, driver will use interrupt taskqueue. The default value of
the tunable is 0. This tunable has no effect on controllers that
has no MSI/MSI-X capability or if MSI/MSI-X is explicitly disabled
by administrator.

While I'm here cleanup interrupt setup/teardown since re(4) uses
single MSI/MSI-X message at this moment.

Remove TX taskqueue and directly invoke re_start in interrupt task.

Prefer MSI-X to MSI on controllers that support MSI-X. All
recent PCIe controllers(RTL8102E or later and RTL8168/8111C or
later) supports either 2 or 4 MSI-X messages. Unfortunately vendor
did not publicly release RSS related information yet. However
switching to MSI-X is one-step forward to support RSS.

Change model names of controller RTL_HWREV_8168_SPIN[123] to real ones.
s/RL_HWREV_8168_SPIN1/RL_HWREV_8168B_SPIN1/g
s/RL_HWREV_8168_SPIN2/RL_HWREV_8168B_SPIN2/g
s/RL_HWREV_8168_SPIN3/RL_HWREV_8168B_SPIN3/g
No functional changes.

Implement initial jumbo frame support for RTL8168/8111 C/D/E PCIe
GbE controllers. It seems these controllers no longer support
multi-fragmented RX buffers such that driver have to allocate
physically contiguous buffers.

o Retire RL_FLAG_NOJUMBO flag and introduce RL_FLAG_JUMBOV2 to
mark controllers that use new jumbo frame scheme.
o Configure PCIe max read request size to 4096 for standard frames
and reduce it to 512 for jumbo frames.
o TSO/checksum offloading is not supported for jumbo frames on
these controllers. Reflect it to ioctl handler and driver
initialization.
o Remove unused rl_stats_no_timeout in softc.
o Embed a pointer to structure rl_hwrev into softc to keep track
of controller MTU limitation and remove rl_hwrev in softc since
that information is available through a pointer to structure
rl_hwrev.

Special thanks to Realtek for donating sample hardwares which made
this possible.

H/W donated by: Realtek Semiconductor Corp.

Add initial support for RTL8168E/8111E-VL PCIe GbE.

H/W donated by: Realtek Semiconductor Corp.

Implement TSO on RealTek RTL8168/8111 C or later controllers.
RealTek changed TX descriptor format for later controllers so these
controllers require MSS configuration in different location of TX
descriptor. TSO is enabled by default for controllers that use new
descriptor format.
For old controllers, TSO is still disabled by default due to broken
frames under certain conditions but users can enable it.
Special thanks to Hayes Wang at RealTek.

MFC after: 2 weeks

Remove standard PCI configuration space register definitions.

Remove trailing white spaces.

Consistently use tab character instead of using space character.
No functional changes.

Follow the lead of vendor's interrupt moderation mechanism.
It seems RTL8169/RTL8168/RTL810xE has a kind of interrupt
moderation mechanism but it is not documented at all. The magic
value dramatically reduced number of interrupts without noticeable
performance drops so apply it to all RTL8169/RTL8169 controllers.
Vendor's FreeBSD driver also applies it to RTL810xE controllers but
their Linux driver explicitly cleared the register, so do not
enable interrupt moderation for RTL810xE controllers.

While I'm here sort 8169 specific registers.

Obtained from: RealTek FreeBSD driver

Add simple MAC statistics counter reading support. Unfortunately
useful counters like rl_missed_pkts is 16 bits quantity which is
too small to hold meaningful information happened in a second. This
means driver should frequently read these counters in order not to
lose accuracy and that approach is too inefficient in driver's
view. Moreover it seems there is no way to trigger an interrupt to
detect counter near-full or wraparound event as well as lacking
clearing the MAC counters. Another limitation of reading the
counters from RealTek controllers is lack of interrupt firing at
the end of DMA cycle of MAC counter read request such that driver
have to poll the end of the DMA which is a time consuming process
as well as inefficient. The more severe issue of the MAC counter
read request is it takes too long to complete the DMA. All these
limitation made maintaining MAC counters in driver impractical. For
now, just provide simple sysctl interface to trigger reading the
MAC counters. These counters could be used to track down driver
issues. Users can read MAC counters maintained in controller with
the following command.
#sysctl dev.re.0.stats=1

While I'm here add check for validity of dma map and allocated
memory before unloading/freeing them.

Tested by: rmacklem

Convert the PHY drivers to honor the mii_flags passed down and convert
the NIC drivers as well as the PHY drivers to take advantage of the
mii_attach() introduced in r213878 to get rid of certain hacks. For
the most part these were:
- Artificially limiting miibus_{read,write}reg methods to certain PHY
addresses; we now let mii_attach() only probe the PHY at the desired
address(es) instead.
- PHY drivers setting MIIF_* flags based on the NIC driver they hang
off from, partly even based on grabbing and using the softc of the
parent; we now pass these flags down from the NIC to the PHY drivers
via mii_attach(). This got us rid of all such hacks except those of
brgphy() in combination with bce(4) and bge(4), which is way beyond
what can be expressed with simple flags.

While at it, I took the opportunity to change the NIC drivers to pass
up the error returned by mii_attach() (previously by mii_phy_probe())
and unify the error message used in this case where and as appropriate
as mii_attach() actually can fail for a number of reasons, not just
because of no PHY(s) being present at the expected address(es).

Reviewed by: jhb, yongari

Rewrite interrupt handler to give fairness for both RX and TX.
Previously rl(4) continuously checked whether there are RX events
or TX completions in forever loop. This caused TX starvation under
high RX load as well as consuming too much CPU cycles in the
interrupt handler. If interrupt was shared with other devices which
may be always true due to USB devices in these days, rl(4) also
tried to process the interrupt. This means polling(4) was the only
way to mitigate the these issues.

To address these issues, rl(4) now disables interrupts when it
knows the interrupt is ours and limit the number of iteration of
the loop to 16. The interrupt would be enabled again before exiting
interrupt handler if the driver is still running. Because RX buffer
is 64KB in size, the number of iterations in the loop has nothing
to do with number of RX packets being processed. This change
ensures sending TX frames under high RX load.

RX handler drops a driver lock to pass received frames to upper
stack such that there is a window that user can down the interface.
So rl(4) now checks whether driver is still running before serving
RX or TX completion in the loop.

While I'm here, exit interrupt handler when driver initialized
controller.

With this change, now rl(4) can send frames under high RX load even
though the TX performance is still not good(rl(4) controllers can't
queue more than 4 frames at a time so low TX performance was one of
design issue of rl(4) controllers). It's much better than previous
TX starvation and you should not notice RX performance drop with
this change. Controller still shows poor performance under high
network load but for many cases it's now usable without resorting
to polling(4).

MFC after: 2 weeks

Rename rl_setmulti() to rl_rxfilter() as rl_rxfilter() will handle
IFF_ALLMULTI/IFF_PROMISC as well as multicast filter configuration.
Rewrite RX filter logic to reduce number of register accesses and
make it handle promiscuous/allmulti toggling without controller
reinitialization.

Previously rl(4) counted on controller reinitialization to reprogram
promiscuous configuration but r211767 resulted in avoiding
controller reinitialization whenever promiscuous mode is toggled.
To address this, keep track of driver's view of interface state and
handle IFF_ALLMULTI/IFF_PROMISC changes without reinitializing
controller. This should fix a regression introduced in r211267.

While I'm here remove unnecessary variable reassignment in ioctl
handler.

PR: kern/151079
MFC after: 1 week

Remove unnecessary controller reinitialization.

PR: kern/87506

It seems some newer RTL8139 controllers provides only memory space
register mapping. I'm not sure whether it comes from the fact that
controllers live behind certain PCI brdge(PLX PCI 6152 33BC) and
the bridge has some issues in handling I/O space register mapping.
Unfortunately it's not possible to narrow down to an exact
controller that shows this issue because RealTek used the same PCI
device/revision id again. In theory, it's possible to check parent
PCI bridge device and change rl(4) to use memory space register
mapping if the parent PCI bridge is PLX PCI 6152. But I didn't try
to do that and we wouldn't get much benefit with added complexity.

Blindly switching to use memory space register mapping for rl(4)
may make most old controllers not to work. At least, I don't want
to take potential risk from such change. So use I/O space register
mapping by default but give users chance to override it via a
tunable. The tunable to use memory space register mapping would be
given by adding the following line to /boot/loader.conf file.

dev.rl.%d.prefer_iomap="0"

This change makes P811B quad-port work with this tunable.

Tested by: Nikola Kalpazanov ( n.kalpazanov <> gmail dot com )
MFC after: 1 week

Implement WOL. WOL is supported on RTL8139B or newer controllers.

PR: kern/148013

Add new tunable 'net.link.ifqmaxlen' to set default send interface
queue length. The default value for this parameter is 50, which is
quite low for many of today's uses and the only way to modify this
parameter right now is to edit if_var.h file. Also add read-only
sysctl with the same name, so that it's possible to retrieve the
current value.

MFC after: 1 month

Consistently use capital letters.

Add preliminary support for 8168E/8111E PCIe controller.
While I'm here simplify device description string.

Tested by: Michael Beckmann < michael <> apfel dot de >
MFC after: 5 days

Add initial support for RTL8103E PCIe fastethernet.

PR: kern/142974

Add support for four more nfsmb controllers, shipping on at least the
ASUS Atom ION boards.

PR: kern/142571
Submitted by: oliver
Approved by: ed (mentor)
MFC after: 1 week

Remove extraneous semicolons, no functional changes.

Submitted by: Marc Balmer <marc@msys.ch>
MFC after: 1 week

intpm/sb700: force polling mode if configured interrupt is SMI

instead of failing to attach

intpm: add support for smbus controller found in AMD SB700

According to the specifications AMD/ATI SMBus controller is very
similar to SMBus controller found in PIIX4.
Some notable differences:
o different bit for enabling/signalling regular interrupt mode
o in practice seems to support only polling mode
Thus, intpm driver is modified to support polling-only mode
and to recognize SB700 PCI ID and differences.

Tested on: SB700 and PIIX4E platforms
Reviewed by: jhb
MFC after: 2 weeks
X-Perhaps-ToDo: rename the driver to reflect its function
and supported hardware

Add RTL8168DP/RTL8111DP device id. While I'm here append "8111D" to
the description of RTL8168D as RL_HWREV_8168D can be either
RTL8168D or RTL8111D.

PR: kern/137672
MFC after: 3 days

Adding hardware ID for RTL810x PCIe found on HP Pavilion DV2-1022AX.

Reviewed by: yongari
Approved by: re (kib, kensmith)

Use if_maddr_rlock()/if_maddr_runlock() rather than IF_ADDR_LOCK()/
IF_ADDR_UNLOCK() across network device drivers when accessing the
per-interface multicast address list, if_multiaddrs. This will
allow us to change the locking strategy without affecting our driver
programming interface or binary interface.

For two wireless drivers, remove unnecessary locking, since they
don't actually access the multicast address list.

Approved by: re (kib)
MFC after: 6 weeks

When user_frac in the polling subsystem is low it is going to busy the
CPU for too long period than necessary. Additively, interfaces are kept
polled (in the tick) even if no more packets are available.
In order to avoid such situations a new generic mechanism can be
implemented in proactive way, keeping track of the time spent on any
packet and fragmenting the time for any tick, stopping the processing
as soon as possible.

In order to implement such mechanism, the polling handler needs to
change, returning the number of packets processed.
While the intended logic is not part of this patch, the polling KPI is
broken by this commit, adding an int return value and the new flag
IFCAP_POLLING_NOCOUNT (which will signal that the return value is
meaningless for the installed handler and checking should be skipped).

Bump __FreeBSD_version in order to signal such situation.

Reviewed by: emaste
Sponsored by: Sandvine Incorporated

For RTL8139C+ controllers, have controller handle padding short
checksum offload frames. Software workaround used for broken
controllers(RTL8169, RTL8168, RTL8168B) seem to cause watchdog
timeouts on RTL8139C+.
Introduce a new flag RL_FLAG_AUTOPAD to mark automatic padding
feature of controller and set it for RTL8139C+ and controllers that
use new descriptor format. This fixes watchdog timeouts seen on
RTL8139C+.

Reported by: Dimitri Rodis < DimitriR <> integritasystems dot com >
Tested by: Dimitri Rodis < DimitriR <> integritasystems dot com >

intpm: minor enhancements

1. fix nointr check in intsmb_start, matters only if ENABLE_ALART is
defined (by default, it is not);
2. drop unnecessary inspection/reporting of power-management io registers
base address;
3. in verbose mode report errors from SMBus host controller and their
mapping to smbus(4) errors;

Approved by: jhb (mentor)

The callback takes a void *, not a caddr_t * (sic).
Except for the bb callback, which takes a caddr_t and not a caddr_t *.

Allocating 2 MSI messages do not seem to work on certain controllers
so use just 1 MSI message. This fixes regression introduced in
r188381.

Tested by: many

Destroy TX tag outside of loop scope.

Found with: Coverity Prevent(tm)
CID: 3886

- Add support for 8110SCe part. Some magic registers were taken from
Linux driver.
- Swap hardware revisions for 8110S and 8169S as Linux driver claims.

Reviewed by: yongari (early version)

Retire RL_FLAG_INVMAR bit to match its comment and reality.

Sometimes RTL8168B seems to take long time to access GMII registers
in device attach phase. Double GMII register access timeout value
to fix the issue.

Reported by: wkoszek
Tested by: wkoszek

Since we don't request reset for rlphy(4), the link state 'UP'
event from mii(4) may not be delivered if valid link was already
established. To address the issue, check current link state after
driving MII_TICK. This should fix a regression introduced in
r184245.

PR: kern/129647

It seems that RealTek PCIe controllers require an explicit Tx poll
command whenever Tx completion interrupt is raised. The Tx poll
bit is cleared when all packets waiting to be transferred have been
processed. This means the second Tx poll command can be silently
ignored as the Tx poll bit could be still active while processing
of previous Tx poll command is in progress.
To address the issue re(4) used to invoke the Tx poll command in Tx
completion handler whenever it detects there are pending packets in
TxQ. However that still does not seem to completely eliminate
watchdog timeouts seen on RealTek PCIe controllers. To fix the
issue kick Tx poll command only after Tx completion interrupt is
raised as this would indicate Tx is now idle state such that it can
accept new Tx poll command again. While here apply this workaround
for PCIe based controllers as other controllers does not seem to
have this limitation.

Tested by: Victor Balada Diaz < victor <> bsdes DOT net >

For RTL8168C SPIN2 controllers, make sure to take the controller
out of sleep mode prior to accessing to PHY. This should fix device
attach failure seen on these controllers. Also enable the sleep
mode when device is put into sleep state.

PR: kern/123123, kern/123053

By default assume a 8139 chip if the EEPROM contents prove inconclusive. The
same LOM hardware with goofed-up EEPROM programming also needed reading the
Ethernet address from the chips registers as the EEPROM did not have a
sensible address programmed.

Patch developed by: pyun@
Funky hardware on loan: www.id-it.nl

MFC after: 2 weeks

Make WOL work on RTL8168B. This controller seems to require
explicit command to enable Rx MAC prior to entering D3.

Tested by: Cyrus Rahman <crahman <> gmail DOT com>

Don't access undocumented register 0x82 on controllers that
have no such register. While here clear undocumented PHY
register 0x0B for RTL8110S.

Obtained from: RealTek FreeBSD driver

Newer RealTek controllers requires setting stop request bit to
terminate active Tx/Rx operation.

o Implemented miibus_statchg handler. It detects whether re(4)
established a valid link or not. In miibus_statchg handler add a
check for established link is valid one for the controller(e.g.
1000baseT is not a valid link for fastethernet controllers.)
o Added a flag RE_FLAG_FASTETHER to mark fastethernet controllers.
o Added additional check to know whether we've really encountered
watchdog timeouts or missed Tx completion interrupts. This change
may help to track down the cause of watchdog timeouts.
o In interrupt handler, removed a check for link state change
interrupt. Not all controllers have the bit and re(4) did not
rely on the event for a long time. In addition, re(4) didn't
request the interrupt in RL_IMR register.

Tested by: rpaulo

Update if_iqdrops instead of if_ierrors when m_devget(9) fails.

Add 8168D support.

Submitted by: Andrew < andrewwtulloch <> gmail DOT com >

Move mn over. One of the last stragglers in sys/pci. There's no
module built for this hardware, so no changes needed.

Make RL_TWISTER_ENABLE a tunable/sysctl. Eliminate it as an option.
Fix module build.

Submitted by: Kostik Belousov

Fix a few typos/spelling errors in my comments from the last commit,
plus a few others that had lingered in this driver...

Submitted by: "b." bf2006a att yahoo KIBO com

Add RL_TWISTER_ENABLE option. This enables the magic bits to do long
cable tuning. This has helped in some installations for hardware
deployed by a former employer. Made optional because the lists aren't
full of complaints about these cards... even when they were wildly
popular.

Reviewed by: attilio@, jhb@, trhodes@ (all an older version of the patch)

Implement miibus_statchg handler. It detects whether rl(4)
established a valid link or not.
In rl_start_locked, don't try to send packets unless we have valid
link. While I'm here add a check that verifies whether driver can
accept Tx requests by inspecting IFF_DRV_OACTIVE/IFF_DRV_RUNNING
flag.

After sending stop command to MAC, give hardware chance to drain
active DMA operation.

Make rl_init_locked() call rl_reset. This will put hardware into
sane state after resume/watchdog timeouts.

Don't rearm watchdog timer in rl_txeof(). The watchdog timer was
already set in rl_start_locked(). Touching the watchdog timer in
other places will hide the root cause of watchdog timeouts.

Various bus_dma(9) fixes.
- The hardware does not support DAC so limit DMA address space to
4GB.
- Removed BUS_DMA_ALLOC_NOW flag.
- Created separated Tx buffer and Rx buffer DMA tags. Previously
it used to single DMA tag and it was not possible to specify
different DMA restrictions.
- Apply 4 bytes alignment limitation of Tx buffer.
- Apply 8 bytes alignment limitation of Rx buffer.
- Tx side bus_dmamap_load_mbuf_sg(9) support.
- Preallocate Tx DMA maps as creating DMA maps take very long time
on architectures that require real DMA maps.
- Adjust guard buffer size to 1522 + 8 as it should include VLAN
and additional reserved bytes in Rx buffer.
- Plug memory leak in device detach. Previously wrong buffer
address was used to free allocated memory.
- Added rl_list_rx_init() to clear Rx buffer and cleared the
buffer.
- Don't destroy DMA maps in rl_txeof() as the DMA map should be
reused. There is no reason to destroy/recreate the DMA maps in
this driver.
- Removed rl_dma_map_rxbuf()/rl_dma_map_txbuf() callbacks.
- The hardware does not support descriptor based DMA on Tx side
and the Tx buffer address should be aligned on 4 bytes boundary
as well as manual padding for short frames. Because of this
hardware limitation rl(4) always used to invoke m_defrag(9) to
get a 4 bytes aligned single buffer. However m_defrag(9) takes
a lot of CPU cycles on slow machines and not all packets need
the help of m_defrag(9). Armed with the information, don't
invoke m_defrag(9) if the following conditions are true.
1. Buffer is not fragmented.
2. Buffer is aligned on 4 bytes boundary.
3. Manual padding is not necessary.
4. Or padding is necessary but upper stack passed a writable
buffer and the space needed for padding is satisfied.
This change combined with preallocated DMA maps greatly
increased Tx performance of driver on sparc64.
- Moved bus_dmamap_sync(9) in rl_start_locked() to rl_encap() and
corrected memory synchronization operation specifier of
bus_dmamap_sync(9).
- Removed bus_dmamap_unload(9) in rl_stop(). There is no need to
reload/unload Rx buffer as rl(4) always have to copy from the
buffer. It just needs proper bus_dmamap_sync(9) calls before
copying the received frame.

With this change rl(4) should work on systems with more than 4GB
memory.

PR: kern/128143

Move wb driver from sys/pci to sys/dev/wb.

Move pcn driver from sys/pci to sys/dev/pcn.

Move the ste driver from sys/pci to sys/dev/ste.

Move the tl driver form sys/pci to sys/dev/tl.

Move sis to sys/dev/sis for consistency.

Move the xl driver form sys/pci to sys/dev/xl for consistency.

Add locking to the various iicbus(4) bridge drivers:
- Just grab Giant in the ixp425_iic(4) driver since this driver uses
a shared address/data register window pair to access the actual
I2C registers. None of the other ixp425 drivers lock access to these
shared address/data registers yet and that would need to be done before
this could use any meaningful locking.
- Add locking to the interrupt handler and 'iicbus_reset' methods of the
at91_twi(4) driver.
- Add locking to the pcf(4) driver. Other pcf(4) fixes include:
- Don't needlessly zero the softc.
- Use bus_foo rather than bus_space_foo and remove bus space tag and
handle from softc.
- The lpbb(4) driver just grabs Giant for now. This will be refined later
when ppbus(4) is locked.
- As was done with smbus earlier, move the DRIVER_MODULE() lines to match
the bus driver (either iicbus or iicbb) to the bridge driver into the
bridge drivers.

Tested by: sam (arm/ixp425)

The number of bits reserved for MSS in RealTek controllers is
11bits. This limits the maximum interface MTU size in TSO case
as upper stack should not generate TCP segments with MSS greater
than the limit. Armed with this information, disable TSO if
interface MTU is greater than the limit.

Set all of the "optimum performance" PHY registers for the 15D parts as
well as the 15C since it seems to be required in practice. The Linux
natsemi.c driver mostly does this as well.

PR: kern/112179
Submitted by: Mark Willson mark - hydrus org uk
MFC after: 1 week

SIS_SETBIT() already does a read/modify/write operation, so there isn't
a reason to read the register twice.

MFC after: 1 week

Add driver support for RTL8102E and RTL8102EL which is the second
generation of RTL810x PCIe fast ethernet controller. Note, Tx/Rx
descriptor format is different from that of first generation of
RTL8101E series. Jumbo frame is not supported for RTL810x
family.

Tested by: NAGATA Shinya ( maya AT negeta DOT com )

Add a new RTL8169 variant, 8169SB(L)/8110SB(L).

Reported by: nork
Tested by: nork

Add basic support for RTL8168C, RTL8168CP, RTL8111C and RTL8111CP.
ATM Tx/Rx checksum offload is supported but TSO and jumbo frame is
not yet supported. Because these newer controllers use different
descriptor formats, a flag RL_FLAG_DESCV2 flag was introduced to
handle that case in Tx/Rx handler. Also newer controllers seems to
require to not touch 'enable Tx/Rx bit' in RL_CPLUS_CMD register
so don't blindly try to set that bits.

Note, it seems that there is still power-saving related issue where
driver fails to attach PHY. Rebooting seems to fix that issue but
number of required reboots varys.

Many thanks to users that helped during developement. I really
appreciate their patient and test/feedbacks.

Instead of allocating variables for each events/hardware flags, use
a dedicated flag that represents controller capabilities/events.
This will simplify many part of code that requires different
workaround for each controller revisions and will enhance
readability.
While I'm here move PHY wakeup code up before mii_phy_probe() which
seems to help to wake PHY in some cases.

Switch to memory space register mapping over IO space. If that
mapping fails fall back to traditional IO space access.

Add support for VT8237 ISA bridge.

PR: kern/120714
Event: Bugathon#5

Add the SMB functionality for the MCP65 chipset I happen to
have in my new motherboard.

Add another 8139D variant.

PR: 124622
Submitted by: Evgeny Zhirnov <jirnov at gmail dot com>
Approved by: imp (mentor, implicit)
MFC after: 3 days

- Use bus_foo() rather than bus_space_foo() and remove bus space tag/handle
from softc.
- Mark interrupt handlers MPSAFE as these drivers have been locked for a
while.

Remove unused 'xl_unit' member from softc.

- Set sc->dev to the new-bus device_t so all the device_printf()s work.
- Add a missing newline to a printf.

MFC after: 1 week
Submitted by: Andriy Gapon avg <> icyb.net.ua

It seems that RealTek 8129/8139 chip reports invalid length of
received frame under certain conditions. wpaul said the length
0xfff0 is special meaning that indicates hardware is in the
process of copying a packet into host memory. But it seems
there are other cases that hardware is busy or stuck in bad
situation even if the received frame length is not 0xfff0.
To work-around this condition, add a check that verifys that
recevied frame length is in valid range. If received length is out
of range reinitialize hardware to recover from stuck condition.

Reported by: Mike Tancsa ( mike AT sentex DOT net )
Tested by: Mike Tancsa
Obtained from: OpenBSD
MFC after: 1 week

Padding more bytes than necessary one broke another variants of
PCIe RealTek chips. Only pad IP packets if the payload is less than
28 bytes.

Obtained from: NetBSD
PR: kern/122221

Replaced the misleading uses of a historical artefact M_TRYWAIT with M_WAIT.
Removed dead code that assumed that M_TRYWAIT can return NULL; it's not true
since the advent of MBUMA.

Reviewed by: arch

There are ongoing disputes as to whether we want to switch to directly using
UMA flags M_WAITOK/M_NOWAIT for mbuf(9) allocation.

- Take advantage of bus_dmamap_load_mbuf_sg(9).
- Take advantage of m_collapse(9).
- Sync with other NIC drivers and prepend a TX mbuf if the first attempt
to load it fails with an error other than EFBIG and stop trying instead
of freeing it and keeping on trying to enqueue more mbufs. Also ensure
the driver queue isn't empty before trying to enqueue mbufs in order to
reduce locking operations.
- In xl_ifmedia_upd() add a missing XL_UNLOCK(). [1]
- Const'ify the xl_devs array.
- Remove an outdated comment.

PR: 113406 [1]
MFC after: 1 month

For MSI capable hardwares, enable MSI enable bit in RL_CFG2
register. If MSI was disabled by hw.re.msi_disable tunable
expliclty clear the MSI enable bit.

vr(4) was repocopied to src/sys/dev/vr.

Don't map memory/IO resource in device probe and just use PCI
vendor/revision/sub device id of the hardware to probe it.
This is the same way as NetBSD does and it enhances readabilty
a lot.

Don't allow jumbo frame on 8139C+ controller.
While I'm here add a check for minimal MTU length.

Implement WOL.

Tested by: Fabian Keil ( freebsd-listen AT fabienkeli DOT de )

Give MEXTADD() another argument to make both void pointers to the
free function controlable, instead of passing the KVA of the buffer
storage as the first argument.

Fix all conventional users of the API to pass the KVA of the buffer
as the first argument, to make this a no-op commit.

Likely break the only non-convetional user of the API, after informing
the relevant committer.

Update the mbuf(9) manual page, which was already out of sync on
this point.

Bump __FreeBSD_version to 800016 as there is no way to tell how
many arguments a CPP macro needs any other way.

This paves the way for giving sendfile(9) a way to wait for the
passed storage to have been accessed before returning.

This does not affect the memory layout or size of mbufs.

Parental oversight by: sam and rwatson.

No MFC is anticipated.

Add copyrights.

PR: 119136

sf(4) was repocopied to src/sys/dev/sf.

Overhaul re(4).
o Increased number of Rx/Tx descriptors to 256 for 8169 GigEs
because it's hard to push the hardware to the limit with default
64 descriptors.
TSO requires large number of Tx descriptors to pass a full sized
TCP segment(65535 bytes IP packet) to hardware. Previously it
consumed 32 Tx descriptors, assuming MCLBYTES DMA segment size,
to send the TCP segment which means re(4) couldn't queue more
than two full sized IP packets.
For 8139C+ it still uses 64 Rx/Tx descriptors due to its hardware
limitations. With this changes there are (very) small waste of
memory for 8139C+ users but I don't think it would affect 8139C+
users for most cases.
o Various bus_dma(9) fixes.
- The hardware supports DAC so allow 64bit DMA operations.
- Removed BUS_DMA_ALLOC_NOW flag.
- Increased DMA segment size to 4096 from MCLBYTES because TSO
consumes too many descriptors with MCLBYTES DMA segment size.
- Tx/Rx side bus_dmamap_load_mbuf_sg(9) support. With these
changes the code is more readable than previous one and got a
(slightly) better performance as it doesn't need to pass/
decode arguments to/from callback function.
- Removed unnecessary callback function re_dmamap_desc() and
nuked rl_dmaload_arg structure which was used in the callback.
- Additional protection for DMA map load failure. In case of
failure reuse current map instead of returning a bogus DMA
map.
- Deferred DMA map unloading/sync operation for maximum
performance until we really need to load new DMA map. If we
happen to reuse current map(e.g. input error) there is no need
to sync/unload/load again.
- The number of allowable Tx DMA segments for a mbuf chains are
now 32 instead of magic nseg value. If the number of available
Tx descriptors are short enough to send highly fragmented mbuf
chains an optimized re_defrag() is called to collapse mbuf
chains which is supposed to be much faster than m_defrag(9).
re_defrag() was borrowed from ath(4).
- Separated Rx/Tx DMA tag from a common DMA tag such that Rx DMA
tag correctly uses DMA maps that were created with DMA alignment
restriction(8bytes alignments). Tx DMA tag does not have such
alignment limitation.
- Added additional sanity checks for DMA ring map load failure.
- Added additional spare Rx DMA map for graceful handling of Rx
DMA map load failure.
- Fixed misused bus_dmamap_sync(9) and added missing
bus_dmamap_sync(9) in re_encap()/re_txeof()/re_rxeof().
o Enabled TSO again as re(4) have reasonable number of Tx
descriptors.
o Don't touch DMA address of a Tx descriptor in re_txeof(). It's
not needed.
o Fix incorrect update of if_ierrors counter. For Rx buffer
shortage it should update if_qdrops as the buffer is reused.
o Added checks for unsupported H/W revisions and return ENXIO for
these hardwares. This is required to remove resource allocation
code in re_probe as other drivers do in device probe routine.
o Modified descriptor index manipulation macros as it's now possible
to have different number of descriptors for Rx/Tx.
o In re_start, to save a lock operation, use IFQ_DRV_IS_EMPTY before
trying to invoke IFQ_DRV_DEQUEUE. Also don't blindly call re_encap
since we already know the number of available Tx descriptors in
advance.
o Removed RL_TX_DESC_THLD which was used to reserve RL_TX_DESC_THLD
descriptors in Tx path. There is no such a limitation mentioned in
8139C+/8169/8110/8168/8101/8111 datasheet and it seems to work ok
without reserving RL_TX_DESC_THLD descriptors.
o Fix a comment for RL_GTXSTART. The register is 8bits register.
o Added comments for 8169/8139C+ hardware restrictions on descriptors.
o Removed forward declaration for "struct rl_softc", it's not needed.
o Added a new structure rl_txdesc for Tx descriptor managements and
a structure rl_rxdesc for Rx descriptor managements.
o Removed unused member variable rl_intlock in driver softc. There are
still several unused member variables which are supposed to be used
to access hardware statistics counters. But it seems that accessing
hardware counters were not implemented yet.

Add another RTL8168 revision 3 which is found on RTL8111-GR Gigabit
Ethernet Controller. Multicast filtering wasn't tested and needs more
expore. While I'm here change complex if statements with switch
statement which would improve readability.

Reported by: Abdullah Ibn Hamad Al-Marri < wearabnet AT yahoo DOT ca >
Tested by: Abdullah Ibn Hamad Al-Marri < wearabnet AT yahoo DOT ca >

Remove XRPU driver, after asking all the users.

Add the FNW3603TX Planex NIC.

PR: 76081
Approved by: imp (mentor)
Submitted by: umi at pocke dot org
MFC After: 3 days

Fix function prototype for device_shutdown method.

Move the agp(4) driver from sys/pci to sys/dev/agp. __FreeBSD_version was
bumped to 800004 to note the change though userland apps should not be
affected since they use <sys/agpio.h> rather than the headers in
sys/dev/agp.

Discussed with: anholt
Repocopy by: simon

Add SMB support for the MCP61 chipset.

PR: 108830
Submitted by: Edwin Mons <freebsd at edwinm dot ik dot nu>
Approbed by: imp (mentor)

Split agp_generic_detach() up into two routines: agp_free_cdev() destroys
/dev/agpgart and agp_free_res() frees resources like the BAR for the
aperture. Splitting this up lets chipset-specific detach routines
manipulate the aperture during their detach routines without panicing.

MFC after: 1 week
Reviewed by: anholt

Oops, convert a tsleep() to a msleep() that was missed when adding locking
to this driver.

Reported by: Michael Butler : imb of protected-networks net

Spelling fix for interupt -> interrupt

Not all VIA Rhine chips support 256 register space. So touching
VR_STICKHW register would result in unexpected results on these
hardwares. wpaul said the following for the issue.

The vr_attach() routine unconditionally does this for all supported
chips:

/*
* Windows may put the chip in suspend mode when it
* shuts down. Be sure to kick it in the head to wake it
* up again.
*/
VR_CLRBIT(sc, VR_STICKHW, (VR_STICKHW_DS0|VR_STICKHW_DS1));

The problem is, the VR_STICKHW register is not valid on all Rhine
devices. The VT86C100A chip, which is present on the D-Link DFE-530TX
boards, doesn't support power management, and its register space is
only 128 bytes wide. The VR_STICKHW register offset falls outside this
range. This may go unnoticed in most scenarios, but if you happen to have
another PCI device in your system which is assigned the register
space immediately after that of the Rhine, the vr(4) driver will
incorrectly stomp it. In my case, the BIOS on my test board decided
to put the register space for my PRO/100 ethernet board right next
to the Rhine, and the Rhine driver ended up clobbering the IMR register
of the PRO/100 device. (Long story short: the board kept locking up on
boot. Took me the better part of the morning suss out why.)

The strictly correct thing to do would be to check the PCI config space
to make sure the device supports the power management capability and only
write to the VR_STICKHW register if it does.

Instead of inspecting chip revision numbers for the availability of
VR_STICKHW register, check the existence of power management capability
of the hardware as wpaul suggested.

Reported by: wpaul
Suggested by: wpaul
OK'ed by: jhb

- Add the device ID for the VIA VT3324 (CX700) chipset.
- Set and Get aperture size correctly for VIA's AGP3 chipsets.

Approved by: re (kensmith)

Add the PCI id for the Intel 7221's integrated graphics controller. It is
similar to a 915G.

Approved by: re (kensmith)
Reviewed by: anholt
MFC after: 3 weeks

Remove the now-unused NET_{LOCK,UNLOCK,ASSERT}_GIANT() macros, which
previously conditionally acquired Giant based on debug.mpsafenet. As that
has now been removed, they are no longer required. Removing them
significantly simplifies error-handling in the socket layer, eliminated
quite a bit of unwinding of locking in error cases.

While here clean up the now unneeded opt_net.h, which previously was used
for the NET_WITH_GIANT kernel option. Clean up some related gotos for
consistency.

Reviewed by: bz, csjp
Tested by: kris
Approved by: re (kensmith)

Add the device ID for the VIA CX700 chipset.

Approved by: re (hrs)

Add MSI support.
Ever since switching to adaptive polling re(4) occasionally spews
watchdog timeouts on systems with MSI capability. This change is
minimal one for supporting MSI and re(4) also needs MSIX support
for RTL8111C in future. Because softc structure of re(4) is shared
with rl(4), rl(4) was touched to use the modified softc.

Reported by: cnst
Tested by: cnst
Approved by: re (kensmith)

Add support for G965/Q965/GM965/GME965/GME945 AGP.

This adds a function to agp.c to set the aperture resource ID if it's
not the usual AGP_APBASE. Previously, agp.c had been assuming
AGP_APBASE, which resulted in incorrect agp_info, and contortions by
agp_i810.c to work around it.

This also adds functions to agp.c for default AGP_GET_APERTURE() and
AGP_SET_APERTURE(), which return the aperture resource size and disallow
aperture size changes. Moving to these for our AGP drivers will likely
result in stability improvements. This should fix 855-class aperture
size detection.

Additionally, refuse to attach agp_i810 when some RAM is above 4GB and
the GART can't reference memory that high. This should be very rare.
The correct solution would be bus_dma conversion for agp, which is
beyond the scope of this change. Other AGP drivers could likely use
this change as well.

G33/Q35/Q33 AGP support is also included, but disconnected by default
due to lack of testing.

PR: kern/109724 (855 aperture issue)
Submitted by: FUJIMOTO Kou<fujimoto@j.dendai.ac.jp>
Approved by: re (hrs)

re(4) devices requires an external EEPROM. Depending on models it
would be 93C46(1Kbit) or 93C56(2Kbit). One of differences between them
is number of address lines required to access the EEPROM. For example,
93C56 EEPROM needs 8 address lines to read/write data. If 93C56
recevied premature end of required number of serial clock(CLK) to set
OP code/address of EEPROM, the result would be unexpected behavior.
Previously it tried to detect 93C46, which requires 6 address lines,
and then assumed it would be 93C56 if read data was not expected
value. However, this approach didn't work in some models/situations
as 93C56 requries 8 address lines to access its data. In order to fix
it, change EEPROM probing order such that 93C56 is detected reliably.

While I'm here change hard-coded address line numbers with defined
constant to enhance readability.

PR: 112710
Approved by: re (mux)

Prepare for future integration between CAM and newbus. xpt_bus_register
now takes a device_t to be the parent of the bus that is being created.
Most SIMs have been updated with a reasonable argument, but a few exceptions
just pass NULL for now. This argument isn't used yet and the newbus
integration likely won't be ready until after 7.0-RELEASE.

The mac address must be written a word length at a time, it was having no effect before.

MFC after: 1 week

Where I previously removed calls to kdb_enter(), now remove include of
kdb.h.

Pointed out by: bde

Implement assert() in ncr.c using KASSERT() rather than explicitly testing
the assertion and then calling kdb_enter().

tl(4) appears to support long frames.

Tested by: Peter Jeremy <peterjeremy at optushome dot com dot au>

Add VLAN capability.

Submitted by: Slawa Olhovchenkov <slw@zxy.spb.ru>

Remove the old software bit-banging MII interface, we started using
the Rhines shiftregisters in four years ago (1.60).

Remove further cobwebs: Two layers of pointless substructures.

Initialize the physical next pointer in the tx descriptors when we
initialize instead of in the start routine.

Don't rename fields with #define.
Collapse two semantically identical structs.
Add missing vr_ prefix.

Run if_vr(4) through FlexeLint and clean some of the cobwebs found.

Oops, fix intsmb(4) attach. Don't overwrite the 'value' holding the
interrupt mode with the SMB revision before checking 'value' for a valid
interrupt mode.

Reported by: Ulrich Spoerlein <uspoerlein of gmail fame>

Don't reinitialize the hardware if only PROMISC flag was changed.
Previously whenever PROMISC mode turned on/off link renegotiation
occurs and it could resulted in network unavailability for serveral
seconds.(Depending on switch STP settings it could last several tens
seconds.)

Reported by: Prokofiev S.P. < proks AT logos DOT uptel DOT net >
Tested by: Prokofiev S.P. < proks AT logos DOT uptel DOT net >

Add support for hw-assisted checksums on 6105M.

Sponsored by: Soekris Engineering

No need to throw tag+handle around on the stack.

Improve the if_vr driver ever so slightly.

The 6105M and 6102 does not have the DWORD alignment problem, so
don't m_defrag() every packet in the transmit path for those.

More stringent usage of tx-descriptor ring and its flags.

Tested on 6102 and 6105M, other chips may also be able to run
without the m_defrag() but I have neither hardware nor docs to
find out.

Sponsored by: Soekris Engineering

Remove Giant from CAM. Drivers (SIMs) now register a mutex that CAM will
use to synchornize and protect all data objects that are used for that
SIM. Drivers that are not yet MPSAFE register Giant and operate as
usual. RIght now, no drivers are MPSAFE, though a few will be changed
in the coming week as this work settles down.

The driver API has changed, so all CAM drivers will need to be recompiled.
The userland API has not changed, so tools like camcontrol do not need to
be recompiled.

Add new ST201 PCI id.

PR: kern/108150

Convert sis(4) to use its own watchdog procedure.

Submitted by: Florian C. Smeets <flo kasimir com>

Catchup with filters

o break newbus api: add a new argument of type driver_filter_t to
bus_setup_intr()

o add an int return code to all fast handlers

o retire INTR_FAST/IH_FAST

For more info: http://docs.freebsd.org/cgi/getmsg.cgi?fetch=465712+0+current/freebsd-current

Reviewed by: many
Approved by: re@

Change the remainder of the drivers for DMA'ing devices enabled in the
sparc64 GENERIC and the sound device drivers known working on sparc64
to use bus_get_dma_tag() to obtain the parent DMA tag so we can get rid
of the sparc64_root_dma_tag kludge eventually. Except for ath(4), sk(4),
stge(4) and ti(4) these changes are runtime tested (unless I booted up
the wrong kernels again...).

o In re_newbuf() and re_encap() if re_dma_map_desc() aborts the mapping
operation as it ran out of free descriptors or if there are too many
segments in the first place, call bus_dmamap_unload() in order to
unload the already loaded segments.
For trying to map the defragmented mbuf (chain) in re_encap() this
introduces re_dma_map_desc() setting arg.rl_maxsegs to 0 as a new
failure mode. Previously we just ignored this case, corrupting our
view of the TX ring.
o In re_txeof():
- Don't clear IFF_DRV_OACTIVE unless there are at least 4 free TX
descriptors. Further down the road re_encap() will bail if there
aren't at least 4 free TX descriptors, causing re_start() to
abort and prepend the dequeued mbuf again so it makes no sense
to pretend we could process mbufs again when in fact we won't.
While at it replace this magic 4 with a macro RL_TX_DESC_THLD
throughout this driver.
- Don't cancel the watchdog timeout as soon as there's at least one
free TX descriptor but instead only if all descriptors have been
handled. It's perfectly normal, especially in the DEVICE_POLLING
case, that re_txeof() is called when only a part of the enqueued
TX descriptors have been handled, causing the watchdog to be
disarmed prematurely.
o In re_encap():
- If m_defrag() fails just drop the packet like other NIC drivers
do. This should only happen when there's a mbuf shortage, in which
case it was possible to end up with an IFQ full of packets which
couldn't be processed as they couldn't be defragmented as they
were taking up all the mbufs themselves. This includes adjusting
re_start() to not trying to prepend the mbuf (chain) if re_encap()
has freed it.
- Remove dupe initialization of members of struct rl_dmaload_arg to
values that didn't change since trying to process the fragmented
mbuf chain.
While at it remove an unused member from struct rl_dmaload_arg.
o In re_start() remove a abandoned, banal comment. The corresponding
code was moved to re_attach() some time ago.

With these changes re(4) now survives one day (until stopped) of
hammering out packets here.

Reviewed by: yongari
MFC after: 2 weeks

Remove duplicate variable initialization.

CID: 1706
Found by: Coverity Prevent (tm)

- Allow multiple (external) PHYs with Am79C97{2,6}, which actually
only support external PHYs (besides not connectable internal ones
which respond at the usual addresses, but which don't hurt if we
let them show up) and don't wedge when isolating PHYs. Actually,
this change special cases limiting PHYs to Am79C97{3,5,8}, for
which this driver doesn't implement swiching between the internal
and external PHYs, yet, and Am79C971, where isolating the external
PHY (at least in case it's a DP83840A) wedges the chip. Together
with sys/dev/mii/acphy.c rev. 1.21 this adds support for the
100baseFX port of AT-2700 series adaptors, which use two AC101,
one for the copper and one for the fibre port (there might be
variants which only use one PHY though).
- Fix a bug in the previous revision that prevented the address of
the used (external) PHY to be actually recorded.
- Don't bother to set if_mtu to ETHERMTU, ether_ifattach() does that.

MFC after: 1 week

Various updates to most of the smbus(4) drivers:
- Use printf() and device_printf() instead of log() in ichsmb(4).
- Create the mutex sooner during ichsmb(4) attach.
- Attach the interrupt handler later during ichsmb(4) attach to avoid
races.
- Don't try to set PCIM_CMD_PORTEN in ichsmb(4) attach as the PCI bus
driver does this already.
- Add locking to alpm(4), amdpm(4), amdsmb(4), intsmb(4), nfsmb(4), and
viapm(4).
- Axe ALPM_SMBIO_BASE_ADDR, it's not really safe to write arbitrary values
into BARs, and the PCI bus layer will allocate resources now if needed.
- Merge intpm(4) and intsmb(4) into just intsmb(4). Previously, intpm(4)
attached to the PCI device and created an intsmb(4) child. Now,
intsmb(4) just attaches to PCI directly.
- Change several intsmb functions to take a softc instead of a device_t
to make things simpler.

Restore agp aperture size after resume, in case it is modified after boot.

- Clean up Aperture Access Global Enable (APEN) bit access.
- Rename confusing AGP_INTEL_I845_MCHCFG to AGP_INTEL_I845_AGPM.
- Move E7205 and E7505 from i8x5 to i8x0 family. It probably worked
because the actual offset is the same.

In fact, all three families have the bit at the exact same place. Only
differences are name and width of the registers, i.e., NBXCFG (0x50, dword),
RDCR (0x51, byte), AGPM (0x51, byte), MCHCFG (0x50, word) depending on
the family of the chipsets.

Fix style(9).

Make agp_intel capable to work after resume from S3 state.

Don't assume IF_LLADDR returns aligned memory address.
Because accessing ID registers in rtl81x9 needs 32bit register access
and RL_IDR4/RL_IDR5 registers are reservered registers bzero() is
needed before copying ethernet address.
This fixes unaligned memory accesses panic in sparc64.

PR: kern/106801
MFC after: 3 days

- Revert the parts of the previous revision which reloaded the watchdog
timer in xl_txeof()/xl_txeof_90xB(); xl_poll_locked() unconditionally
invokes xl_txeof()/xl_txeof_90xB(), effectively circumventing that
the watchdog ever fires in the DEVICE_POLLING case as its timer is
constantly reloaded.
- Remove the banal and pedantically outdated comment regarding setting
xl_wdog_timer to 0 in xl_txeof().

Pointed out by: bde

- Use the xl_stats_update() callout instead of if_slowtimo() for
driving xl_watchdog() in order to avoid races accessing if_timer.
While at it relax the watchdog a bit by reloading it in xl_txeof()/
xl_txeof_90xB() if there are still packets enqueued.
- Use bus_get_dma_tag() so xl(4) works on platforms requiring it.
- Don't bother to set if_mtu to ETHERMTU, ether_ifattach() does that.

if_watchdog -> rl_watchdog

- Clear the PCN_MISC_ASEL bit so the media port can be actually set
via the PCN_CSR_MODE register. Along with sys/dev/mii/nsphy.c 1.26
this fixes the case of certain Am79c971-based HP cards and on-board
ones in IBM machines reporting link but not actually passing any
traffic. [1]
- Add support for the internal 10baseT PHY, which actually is used on
at least said HP cards (together with an external DP83840A in a
multiple PHYs configuration). With cards that don't make use of this
internal PHY it'll also show up in FreeBSD but not cause any harm.
This is still missing support for multiple PHYs configuration using
the internal 100baseTX and/or HomePNA PHYs together with external
PHYs or multiple external PHYs though.
- In pcn_ifmedia_upd() call pcn_reset() as otherwise the Am79C971 of
at least said HP cards can wedge when switching from the internal
10baseT PHY to the external PHY. This means that we need to also
initialize and possibly start the chip again in pcn_ifmedia_upd(),
which isn't that bad though as for setting the media port the chip
has to be powered down or stopped anyway and unlike documented
doesn't take effect until the next initialization.

PR: 27995, 25959, 72966 (likely) [1]
MFC after: 2 weeks

Add TSO support.

Tested by: wilko, Pieter de Goeje < pieter AT degoeje DOT nl >

Use #ifndef __NO_STRICT_ALIGNMENT rather than
#if !defined(__i386__) && !defined(__amd64__) for architectures
with alignment constraints.

- Don't bother to include IDs of PCnet chips which are not supported
by this driver and largely are not even PCI devices in pcn_chipid.
- Use device_printf(9)/if_printf(9) rather than implementing their
functionality with printf(9).
- Sprinkle some const.

2nd and final commit that moves us to CAM_NEW_TRAN_CODE
as the default.

Reviewed by multitudes.

Remove <sys/types.h>; including both <sys/param.h> and <sys/types.h>
violates style(9).

The first of 3 major steps to move the CAM layer forward to using
the CAM_NEW_TRAN_CODE that has been in the tree for some years now.

This first step consists solely of adding to or correcting
CAM_NEW_TRAN_CODE pieces in the kernel source tree such
that a both a GENERIC (at least on i386) and a LINT build
with CAM_NEW_TRAN_CODE as an option will compile correctly
and run (at least with some the h/w I have).

After a short settle time, the other pieces (making
CAM_NEW_TRAN_CODE the default and updating libcam
and camcontrol) will be brought in.

This will be an incompatible change in that the size of structures
related to XPT_PATH_INQ and XPT_{GET,SET}_TRAN_SETTINGS change
in both size and content. However, basic system operation and
basic system utilities work well enough with this change.

Reviewed by: freebsd-scsi and specific stakeholders

Wrap code optimized for architectures without alignment constraints
in #ifdef __NO_STRICT_ALIGNMENT rather than #if defined(__i386__) ||
defined(__amd64__). Currently this change is cosmetic only though.
While at it, fix a nearby style(9) bug and remove a no longer used
header.

The page queues lock is no longer required by vm_page_busy() or
vm_page_wakeup(). Reduce or eliminate its use accordingly.

Older incarnations of the device used non-standard BARs.

Reported by: Andriy Gapon
Confirmed by: many (including lm-sensors-2.10.1)
MFC after: 3 days

Fix the wraparound of memsize >=2GB.

Fix style(9) nits.

Fix 32-bit PTE in the GART table.

Noticed by: jmg

Add support for 945G/GM AGP chipsets.

The key problem was that the aperture size detection using the MSAC bit
doesn't work -- the bit appears to be set even when it shouldn't be. Linux
takes a different approach, testing for a bit of the GMADR (PCIR_BAR(2)) being
set. However, as I don't think that's a safe way to test aperture size, we
just allocate the resource and check its size. This also pointed out that
agp_generic_attach hadn't been allocating our aperture resource, which may
have caused problems in some cases.

Also corrected is a minor copy-and-pasteo in an error case.

PR: kern/103079
Submitted by: mnag
Tested on: i945GM, i915GM
MFC after: 2 weeks

Whitespace nits.

Consistently use if_printf() only in interface methods: if_start,
if_ioctl, if_watchdog, etc, or in functions that are used by
these methods only. In all other cases use device_printf().

This also fixes several panics, when if_printf() is called before
softc->ifp was initialized.

Submitted by: Alex Lyashkov <umka sevcity.net>

intpm(4) meet style(9). style(9) meet intpm(4).

Trim some unneeded includes.

Minor overhaul of SMBus support:
- Change smbus_callback() to pass a void * rather than caddr_t.
- Change smbus_bread() to pass a pointer to the count and have it be an
in/out parameter. The input is the size of the buffer (same as before),
but on return it will contain the actual amount of data read back from
the bus. Note that this value may be larger than the input value. It
is up to the caller to treat this as an error if desired.
- Change the SMB_BREAD ioctl to write out the updated struct smbcmd which
will contain the actual number of bytes read in the 'count' field. To
preserve the previous ABI, the old ioctl value is mapped to SMB_OLD_BREAD
which doesn't copy the updated smbcmd back out to userland. I doubt anyone
actually used the old BREAD anyway as it was rediculous to do a bulk-read
but not tell the using program how much data was actually read.
- Make the smbus driver and devclass public in the smbus module and
push all the DRIVER_MODULE()'s for attaching the smbus driver to
various foosmb drivers out into the foosmb modules. This makes all
the foosmb logic centralized and allows new foosmb modules to be
self-contained w/o having to hack smbus.c everytime a new smbus driver
is added.
- Add a new SMB_EINVAL error bit and use it in place of EINVAL to return
an error for bad arguments (such as invalid counts for bread and bwrite).
- Map SMB bus error bits to EIO in smbus_error().
- Make the smbus driver call bus_generic_probe() and require child drivers
such as smb(4) to create device_t's via identify routines. Previously,
smbus just created one anonymous device during attach, and if you had
multiple drivers that could attach it was just random chance as to which
driver got to probe for the sole device_t first.
- Add a mutex to the smbus(4) softc and use it in place of dummy splhigh()
to protect the 'owner' field and perform necessary synchronization for
smbus_request_bus() and smbus_release_bus().
- Change the bread() and bwrite() methods of alpm(4), amdpm(4), and
viapm(4) to only perform a single transaction and not try to use a
loop of multiple transactions for a large request. The framing and
commands to use for a large transaction depend on the upper-layer
protocol (such as SSIF for IPMI over SMBus) from what I can tell, and the
smb(4) driver never allowed bulk read/writes of more than 32-bytes
anyway. The other smb drivers only performed single transactions.
- Fix buffer overflows in the bread() methods of ichsmb(4), alpm(4),
amdpm(4), amdsmb(4), intpm(4), and nfsmb(4).
- Use SMB_xxx errors in viapm(4).
- Destroy ichsmb(4)'s mutex after bus_generic_detach() to avoid problems
from child devices making smb upcalls that would use the mutex during
their detach methods.

MFC after: 1 week
Reviewed by: jmg (mostly)

Add support for another ATI IGP 340M (RS200M) AGP bridge.

PR: kern/100958
Submitted by: Kazuo Dohzono <dohzono@axion-software.com>
MFC after: 1 week

Use aperture base address from north bridge. Some BIOS does not encode
misc. control registers correctly and it is inconsistent with north bridge.
In fact, there are too many broken BIOS implementations out there and we
cannot fix every possible combination but at least it is consistent with
what we advertise with ioctl(2).

Add more nForce4 SMBus 2.0 PCI IDs.

Don't reset Tx threshold value whenever xl_init_locked() is called.
Instead the threshould is initialized in device attach. Later the
threshold could be increased in Tx underrun error and the new
threshold should be used in xl_init_locked().

Make sure to check frames in Tx queue are empty before clearing
watchdog timer.

Explicitly set v3 mode only when it is requested. Don't bother otherwise.

Commit the results of the typo hunt by Darren Pilgrim.
This change affects documentation and comments only,
no real code involved.

PR: misc/101245
Submitted by: Darren Pilgrim <darren pilgrim bitfreak org>
Tested by: md5(1)
MFC after: 1 week

Another small update to the re(4) driver:

- Change the workaround for the autopad/checksum offload bug so that
instead of lying about the map size, we actually create a properly
padded mbuf and map it as usual. The other trick works, but is ugly.
This approach also gives us a chance to zero the pad space to avoid
possibly leaking data.

- With the PCIe devices, it looks issuing a TX command while there's
already a transmission in progress doesn't have any effect. In other
words, if you send two packets in rapid succession, the second one may
end up sitting in the TX DMA ring until another transmit command is
issued later in the future. Basically, if re_txeof() sees that there
are still descriptors outstanding, it needs to manually resume the
TX DMA channel by issuing another TX command to make sure all
transmissions are flushed out. (The PCI devices seem to keep the
TX channel moving until all descriptors have been consumed. I'm not
sure why the PCIe devices behave differently.)

(You can see this issue if you do the following test: plug an re(4)
interface into another host via crossover cable, and from the other
host do 'ping -c 2 <host with re(4) NIC>' to prime the ARP cache,
then do 'ping -c 1 -s 1473 <host with re(4) NIC>'. You're supposed
to see two packets sent in response, but you may only see one. If
you do 'ping -c 1 -s 1473 <host with re(4) NIC>' again, you'll
see two packets, but one will be the missing fragment from the last
ping, followed by one of the fragments from this ping.)

- Add the PCI ID for the US Robotics 997902 NIC, which is based on
the RTL8169S.

- Add a tsleep() of 1 second in re_detach() after the interrupt handler
is disconnected. This should allow any tasks queued up by the ISR
to drain. Now, I know you're supposed to use taskqueue_drain() for
this, but something about the way taskqueue_drain() works with
taskqueue_fast queues doesn't seem quite right, and I refuse to be
tricked into fixing it.

Fix the following bugs in re(4)

- Correct the PCI ID for the 8169SC/8110SC in the device list (I added
the macro for it to if_rlreg.h before, but forgot to use it.)

- Remove the extra interrupt spinlock I added previously. After giving it
some more thought, it's not really needed.

- Work around a hardware bug in some versions of the 8169. When sending
very small IP datagrams with checksum offload enabled, a conflict can
occur between the TX autopadding feature and the hardware checksumming
that can corrupt the outbound packet. This is the reason that checksum
offload sometimes breaks NFS: if you're using NFS over UDP, and you're
very unlucky, you might find yourself doing a fragmented NFS write where
the last fragment is smaller than the minimum ethernet frame size (60
bytes). (It's rare, but if you keep NFS running long enough it'll
happen.) If checksum offload is enabled, the chip will have to both
autopad the fragment and calculate its checksum header. This confuses
some revs of the 8169, causing the packet that appears on the wire
to be corrupted. (The IP addresses and the checksum field are mangled.)
This will cause the NFS write to fail. Unfortunately, when NFS retries,
it sends the same write request over and over again, and it keeps
failing, so NFS stays wedged.

(A simple way to provoke the failure is to connect the failing system
to a network with a known good machine and do "ping -s 1473 <badhost>"
from the good system. The ping will fail.)

Someone had previously worked around this using the heavy-handed
approahch of just disabling checksum offload. The correct fix is to
manually pad short frames where the TCP/IP stack has requested
checksum offloading. This allows us to have checksum offload turned
on by default but still let NFS work right.

- Not a bug, but change the ID strings for devices with hardware rev
0x30000000 and 0x38000000 to both be 8168B/8111B. According to RealTek,
they're both the same device, but 0x30000000 is an earlier silicon spin.

Properly detect the RTL8168(B?) again. RealTek sent me a bunch of sample
cards: the chips are all marked "RTL8111B", but they put stickers on the
back that say "RTL8168B/8111B". The manual says there's only one HWREV code
for both the 8111B and 8168B devices, which is 0x30000000, but the cards
they sent me actually report HWREV of 0x38000000. Deciding to trust the
hardware in front of me rather than a possibly incorrect manual (it wouldn't
be the first time the HWREVs were incorrectly documented), I changed the
8168 revision code. It turns out this was a mistake though: 0x30000000
really is a valid for the 8168.

There are two possible reasons for there to be two different HWREVs:

1) 0x30000000 is used only for the 8168B and 0x38000000 is only for
the 8111B.
2) There were 8111/8168 rev A devices which both used code 0x30000000,
and the 8111B/8168B both use 0x38000000.

The product list on the RealTek website doesn't mention the existence of
any 8168/8111 rev A chips being in production though, and I've never seen
one, so until I get clarification from RealTek, I'm going to assume that
0x30000000 is just for the 8168B and 0x38000000 is for the 8111B only.

So, the HWREV code for the 8168 has been put back to 0x30000000,
a new 8111 HWREV code has been added, and there are now separate
entries for recognizing both devices in the device list. This will
allow all devices to work, though if it turns out I'm wrong I may
need to change the ID strings

Fix breakage of CHIP_I855 in the last revision.

Submitted by: Ted Faber <faber@ISI.EDU>

Add support for the RealTek 8169SC/8110SC and RTL8101E devices. The
latter is a PCIe 10/100 chip.

Finally fix the EEPROM reading code so that we can access the EEPROMs on all
devices. In order to access the EEPROM, we must select 'EEPROM programming'
mode, and then set the EEPROM chip select bit. Previously, we were setting
both bits simultaneously, which doesn't work: they must be set in the
right sequence.

Always obtain the station address from the EEPROM, now that EEPROM
reading works correctly.

Make the TX interrupt moderation code based on the internal timer
optional and turned off by default.

Make the re_diag() routine conditional and off by default. When it is
on, only use it for the original 8169, which was the only device that
that really needed it.

Modify interrupt handling to use a fast interrupt handler and fast
taskqeueue.

Correct the rgephy driver so that it only applies the DSP fixup for
PHY revs 0 and 1. Later chips are fixed and don't need the fixup.

Make the rgephy driver advertise both 1000_FD and 1000_HD bits in
autoneg mode. A couple of the devices don't autoneg correctly unless
configured this way.

Replace the three copies of the list of pci ids with a single centralized list.
Add the i945 PCI IDs commented out -- I think it should just work, but it hasn't
been tested yet.

Move SiS 760 to where it belongs.

PR: 98094
Submitted by: Mike M < mmcgus at yahoo dot com >

Rename device name in the last commit. According to PR, the ID is
more likely to belong to chips of 8168 family.

PR: kern/96734
Submitted by: Sven Petai <hadara bsd.ee>

Add support for allocating one larger than page-sized contiguous block of memory
with a physical address. This is used for hardware ARGB cursor support on newer
chipsets.

Since DELAY() was moved, most <machine/clock.h> #includes have been
unnecessary.

Replace references to lnc(4) with references to le(4) (so far the notes
still apply to le(4)) now that lnc(4) is removed and le(4) is going to
replace it.

Remove various bits of conditional Alpha code and fixup a few comments.

The sk(4) driver has moved to /sys/dev/sk

Add support for RTL8111B chip, that can be found on some mainboards,
for example ASUS P5PL2.

Tested by: Vadim Frolov <vadim uch.net>

Remove de driver from old location.

Add support for the Intel E7205 chipset.

PR: kern/91315
Submitted by: Joerg Pulz <Joerg.Pulz@frm2.tum.de>

Do not touch ifp->if_baudrate in miibus aware drivers.

Don't add an agp child in vgapci's attach routine if the PCIY_AGP
capability is present as not all devices supported by the agp_i810 driver
(such as i915) have the AGP capability. Instead, add an identify routine
to the agp_i810 driver that uses the PCI ID to determine if it should
create an agp child device.

Check ifp before dereferencing it in xl_detach(). xl_detach() can be called
from xl_attach(), when ifp is not defined yet.

Found with: Coverity Prevent(tm)

Fix a memory leak I introduced with the hostb/vgapci stuff.

Reported by: Coverity (via dfr's clue-bat)

The number of ticks per usec for YUKON_EC is 125.

fix interrupt moderation timer frequencies for Yukon

Obtained from: OpenBSD

remove trailing spaces

Remove unused code.

Found with: Coverity Prevent(tm)

Use the copy of the card's MAC address saved in tulip_enaddr() in the softc
if we need a valid MAC address (for probing the media for example) before
ether_ifattach() has been called since IF_LLADDR() is NULL then.

Tested by: tisco

Drivers for AMD-8111 and NVIDIA nForce2/3/4 SMBus 2.0 controllers.

Use ETHER_ADDR_LEN rather than hardcoding 6.

- Use PCIR_BAR() macro for the BAR for the aperture.
- Axe macros used for walking PCI capabilities list. We now ask the PCI
bus to find caps for us rather than doing it in the drm and agp drivers.

- Bump FreeBSD version for the hostb(4) and vgapci(4) drivers as well as
the addition of pci_find_extcap().
- Change the drm drivers to attach to vgapci. This is #ifdef'd so the
code can be shared across branches.
- Use pci_find_extcap() to look for AGP and PCIE capabilities in drm.
- GC all the drmsub stuff for i810/i830/i915. The agp and drm devices are
now both children of vgapci.

Change the various AGP drivers that attach to the Host-PCI bridge device to
attach to the hostb driver instead. This means that agp can now be loaded
at runtime (in theory at least). Also, the drivers no longer have to
explicity call device_verbose() to cancel out any earlier calls to
device_quiet() by the hostb(4) driver (this shows a limitation in new-bus,
drivers really shouldn't be doing device_quiet() until they know they are
going to drive that device, i.e. in attach).

Change the agp_find_device() to return the first agp device that has been
attached to a driver rather than always returning agp0.

Don't map the AGP aperture into contiguous KVA. The various graphics
drivers already map sections into KVA as needed anyway. Note that this
will probably break the nvidia driver, but I will coordinate to get that
fixed.

MFC after: 2 weeks

Destroy the /dev device before destroying the mutex or releasing resources
rather than afterwards.

MFC after: 1 week

Use pci_find_extcap() to search for AGP capabilities (PCIY_AGP).

Backout pseudo nForce2/3/4 support. These devices (as well as
AMD-8111 SMBus 2.0 controller) are all SMBus 2.0 controllers,
and need another implementation of SMBus access methods, while
this driver supports AMD-756 SMBus 1.0 controller and clones,
including AMD-8111 SMBus 1.0 controller.

Tested by: Vladimir Timofeev (0x006410de),
mezz (0x008410de),
ru (0x00d410de)

All of us got the same(!) nonsense when running ``mbmon -S'',
repeated every four rows.

Fix PCI ID of the AMD-8111 System Management controller so it matches
SMBus 1.0 and not SMBus 2.0.

AMD-8111 hub (datasheet is publically available) implements both SMBus
2.0 (a separate PCI device) and SMBus 1.0 (a subfunction of the System
Management Controller device with the base I/O address is accessible
through the CSR 0x58). This driver only supports AMD-756 SMBus 1.0
compatible devices.

With the patched sysutils/xmbmon port (to also fix PCI ID and to enable
smb(4) support), I now get:

pciconf:
none0@pci0:7:2: class=0x0c0500 card=0x746a1022 chip=0x746a1022 rev=0x02 hdr=0x00
vendor = 'Advanced Micro Devices (AMD)'
device = 'AMD-8111 SMBus 2.0 Controller'
class = serial bus
subclass = SMBus
amdpm0@pci0:7:3: class=0x068000 card=0x746b1022 chip=0x746b1022 rev=0x05 hdr=0x00
vendor = 'Advanced Micro Devices (AMD)'
device = 'AMD-8111 ACPI System Management Controller'
class = bridge

dmesg:
amdpm0: <AMD 756/766/768/8111 Power Management Controller> port 0x10e0-0x10ff at device 7.3 on pci0
smbus0: <System Management Bus> on amdpm0

# mbmon -A -d
Summary of Detection:
* SMB monitor(s)[ioctl:AMD8111]:
** Winbond Chip W83627HF/THF/THF-A found at slave address: 0x50.
** Analog Dev. Chip ADM1027 found at slave address: 0x5C.
* ISA monitor(s):
** Winbond Chip W83627HF/THF/THF-A found.

I think the confusion comes from the fact that nobody really tried
SMBus with xmbmon :-), since sysutils/xmbmon port doesn't come with
SMBus support enabled, neither in FreeBSD 4, nor in later versions,
so mbmon(1) was just showing the values from the Winbond sensors
accessible through the ISA I/O method (mbmon -I), for me anyway.

On my test machine, the amdpm(4) didn't even attach due to I/O port
allocation failure (who knows what the hell it read from CSR 0x58
of the SMBus 2.0 device :-), which isn't in the CSR space).

I've also checked that lm_sensors.org uses correct PCI ID for SMBus
1.0 of AMD-8111:

i2c-amd756.c: {PCI_VENDOR_ID_AMD, 0x746B, PCI_ANY_ID, PCI_ANY_ID, 0, 0, AMD8111 },

This driver is analogous to our amdpm.c which supports SMBus 1.0
AMD-756 and compatible devices, including SMBus 1.0 on AMD-8111.

i2c-amd8111.c: { 0x1022, 0x746a, PCI_ANY_ID, PCI_ANY_ID, 0, 0, 0 },

This driver is analogous to nForce-2/3/4, i2c-nforce2.c, which
supports SMBus 2.0, and which our amdpm.c does NOT support
(SMBus 2.0 uses a different, ACPI-unified, API to talk to SMBus).
At least I know for sure it doesn't work with my nForce3. :-)

(The xmbmon port will be fixed to correct the PCI ID too and to
enable the smb(4) support.)

Add support for the nForce2/3/4 SMBus controllers which all contain two
SMBus busses. Because of limitations in smbus_if.m, the second smbus is
attached to an amdpm1 device that is a child of amdpm0.

Submitted by: Artemiev Igor ai (at) bmc dot brk dot ru

The if_ti Tigon I/II driver has moved to /sys/dev/ti

Fix -Wundef from compiling the amd64 LINT.

Add support for i915 GMCH AGP. This diff is a combination of work by myself
and some fixes from Motomichi Matsuzaki. Testing involved many people, but the
final, successful testing was from rwatson who endured several rounds of "it
crashes at XYZ stage" "oh, please correct this typo and try again." The Linux
driver, and to a small extent the limited specs, were both used as a reference
for how to program the chipset.

PR: kern/80396
Submitted by: Martin Mersberger

Fix the type of "eaddr" to guarantee the required alignment.

Suggested by: marcel

Byte copy IF_LLADDR() on stack to align it to be safe for typecasts.

Tested by: jhb

Add support for the i855GM, tested by an r300 user.

Various fixes to make de(4) not panic after ru@'s IF_LLADDR() changes:
- Don't call tulip_addr_filter() to reset the RX address filter in
tulip_reset() since that gets called before ether_ifattach(). Just
call it in tulip_init_locked().
- Use be16dec() and le16dec() to parse MAC addresses when programming
the RX filter.
- Let ether_ioctl() handle SIOCSIFMTU since we were doing the exact same
thing with the added bonus that we leaked the driver lock if the MTU
was > ETHERMTU in the homerolled version. This part will be MFC'd.

Clue from: wpaul (1)
Stolen from: marcel (2 via patch for dc(4))
MFC after: 1 week

0xb1881106 seems to be an AGP bridge and some BIOSes incorrectly handle
the bridge. Therefore, we give the same treatment as we did for nForce3-250
and ULi chipsets. VIA AGPv3 code was copied from agp_via.c.

- Store pointer to the link-level address right in "struct ifnet"
rather than in ifindex_table[]; all (except one) accesses are
through ifp anyway. IF_LLADDR() works faster, and all (except
one) ifaddr_byindex() users were converted to use ifp->if_addr.

- Stop storing a (pointer to) Ethernet address in "struct arpcom",
and drop the IFP2ENADDR() macro; all users have been converted
to use IF_LLADDR() instead.

Catch up with IFP2ENADDR() type change (array -> pointer).

- Use callout_*() to manage the callout and make it MPSAFE.
- Fix locking in detach(), we only need to lock across vr_stop().

Tested by: Mike Tancsa mike at sentex dot net
MFC after: 1 week

Add the device ID for the VIA VT8235 south bridge.

PR: kern/62438
Submitted by: FUJIMOTO Kou fujimoto at j dot dendai dot ac dot jp
Tested by: Oliver Fromme olli at secnetix dot de
MFC after: 1 week

Some of the VIA pm and propm devices are actually the same device as the
PCI-ISA bridge. Thus, when viapm0 or viapropm0 attaches, isab0 dosen't
attach so there is no isa0 bus hung off of that bridge. In the non-ACPI
case, legacy0 will add an isa0 anyway as a fail-safe, but ACPI assumes that
any ISA bus will be enumerated via a bridge. To fix this, call
isab_attach() to attach an isa0 ISA child bus device if the pm or propm
device we are probing is a PCI-ISA bridge. Both drivers now have to
implement the bus_if interface via the generic methods for resource
allocation, etc. to work. Also, we now add 2 new ISA bus drivers that
attach to viapm and viapropm devices.

PR: kern/87363
Reported by: Oliver Fromme olli at secnetix dot de
Tested by: glebius
MFC after: 1 week

Remove forgotten, no longer needed WB_UNLOCK() from the end wb_ioctl().
With this change, the driver tests good (at least on i386):

wb0: <Winbond W89C840F 10/100BaseTX> port 0xb800-0xb87f mem 0xe6800000-0xe680007f irq 12 at device 10.0 on pci0
miibus1: <MII bus> on wb0
amphy0: <Am79C873 10/100 media interface> on miibus1
amphy0: 10baseT, 10baseT-FDX, 100baseTX, 100baseTX-FDX, auto
wb0: Ethernet address: 00:00:e8:18:2a:02
wb0: link state changed to DOWN
wb0: link state changed to UP

Fixup locking and mark MPSAFE.
- Add locked variants of init() and start().
- Use callout_*() to manage callout.
- Test IFF_DRV_RUNNING rather than IFF_UP in wb_intr() to see if we are
still active when an interrupt comes in.

I couldn't find any of these cards anywhere to test on myself, and google
turns up references to FreeBSD and OpenBSD manpages for this driver when
trying to locate a card that way. I'm not sure anyone actually uses these
cards with FreeBSD.

Tested by: NO ONE (despite repeated requests)

- Use if_printf() and device_printf() and remove vr_unit from the softc.
I had to initialize the ifnet a bit earlier in attach so that the
if_printf()'s in vr_reset() didn't explode with a page fault.
- Use M_ZERO with contigmalloc() rather than an explicit bzero.

Replace FreeBSD 3.x syntax (controller miibus0) with 4.x syntax
(device miibus) in time for 7.0 :-)

Remove dc after repo copy.

Only allow the sk(4) driver to attach to revision 2 of the LinkSys EG1032
cards and teach the re(4) driver to attach to revision 3 cards.

Submitted by: Fredrik Lindberg fli+freebsd-current at shapeshifter dot se
MFC after: 2 weeks
Reviewed by: imp, mdodd

In detach method, move if_free() after bus_teardown_intr().

Mark sk(4) as capable of handling extended VLAN frames. NICs
based on XMAC II chip should be ready for this in their initial
mode of operation, and Yukon-based NICs are configured so by
the driver.

PR: kern/79998
MFC after: 1 month

- Don't pollute opt_global.h with DEVICE_POLLING and introduce
opt_device_polling.h
- Include opt_device_polling.h into appropriate files.
- Embrace with HAVE_KERNEL_OPTION_HEADERS the include in the files that
can be compiled as loadable modules.

Reviewed by: bde

Big polling(4) cleanup.

o Axe poll in trap.

o Axe IFF_POLLING flag from if_flags.

o Rework revision 1.21 (Giant removal), in such a way that
poll_mtx is not dropped during call to polling handler.
This fixes problem with idle polling.

o Make registration and deregistration from polling in a
functional way, insted of next tick/interrupt.

o Obsolete kern.polling.enable. Polling is turned on/off
with ifconfig.

Detailed kern_poll.c changes:
- Remove polling handler flags, introduced in 1.21. The are not
needed now.
- Forget and do not check if_flags, if_capenable and if_drv_flags.
- Call all registered polling handlers unconditionally.
- Do not drop poll_mtx, when entering polling handlers.
- In ether_poll() NET_LOCK_GIANT prior to locking poll_mtx.
- In netisr_poll() axe the block, where polling code asks drivers
to unregister.
- In netisr_poll() and ether_poll() do polling always, if any
handlers are present.
- In ether_poll_[de]register() remove a lot of error hiding code. Assert
that arguments are correct, instead.
- In ether_poll_[de]register() use standard return values in case of
error or success.
- Introduce poll_switch() that is a sysctl handler for kern.polling.enable.
poll_switch() goes through interface list and enabled/disables polling.
A message that kern.polling.enable is deprecated is printed.

Detailed driver changes:
- On attach driver announces IFCAP_POLLING in if_capabilities, but
not in if_capenable.
- On detach driver calls ether_poll_deregister() if polling is enabled.
- In polling handler driver obtains its lock and checks IFF_DRV_RUNNING
flag. If there is no, then unlocks and returns.
- In ioctl handler driver checks for IFCAP_POLLING flag requested to
be set or cleared. Driver first calls ether_poll_[de]register(), then
obtains driver lock and [dis/en]ables interrupts.
- In interrupt handler driver checks IFCAP_POLLING flag in if_capenable.
If present, then returns.This is important to protect from spurious
interrupts.

Reviewed by: ru, sam, jhb

- Use if_printf() and device_printf() in re(4) and remove rl_unit from
the softc.
- Use callout_init_mtx() and rather than timeout/untimeout in both rl(4)
and re(4).
- Fix locking for ifmedia by locking the driver in the ifmedia handlers
rather than in the miibus functions. (re(4) didn't lock the mii stuff
at all!)
- Fix some locking in re_ioctl().

Note: the two drivers share the same softc declared in if_rlreg.h, so they
had to be change simultaneously.

MFC after: 1 week
Tested by: several on rl(4), none on re(4)

Use if_printf() and device_printf().

Typo.

- Add a work-around for nForce3-250. Aperture base address encoded in misc.
control register and AGP bridge seems to be inconsistent with some BIOS.
Instead of relying on BIOS settings, we just take the initial aperture size
and encode them for both miscellaneous control register and AGP bridge.
Some idea was borrowed from agp_nvidia.c.

- Add preliminary ULi M1689 chipset support. The idea was taken from Linux
because hardware and documentation are unavailable. Not tested.

- Add more VIA chipset PCI IDs taken from Linux driver.

Approved by: anholt (mentor)
Tested by: Adam Gregoire <ebola at psychoholics dot org>
Ganael Laplanche <ganael.laplanche at martymac dot com>
K Wieland <kwieland at wustl dot edu>

Remove bridge(4) from the tree. if_bridge(4) is a full functional
replacement and has additional features which make it superior.

Discussed on: -arch
Reviewed by: thompsa
X-MFC-after: never (RELENG_6 as transition period)

Small fixes to sis(4):
- Remove sis_unit and use device_printf() and if_printf() instead.
- Use callout_init_mtx() for the callout.
- Remove spls.
- Fix locking for ifmedia to happen in the ifmedia handlers rather than in
sis_ioctl().
- Log an error message if we fail to allocate any resources. Perform
cleanup if we fail to allocate any resources so that we don't leave
a mutex hanging around.

Tested by: Jason Tsai jason dot tsai at newcyberian dot com (1-4)
MFC after: 3 days

Use the new bus_space/resource convenience functions.

Move code macros from if_sis*reg*.h to if_sis.*c*

Add a new AGP driver for ATI IGP chipsets. The driver is based on reading of
the Linux driver, since specs are unavailable. Many thanks to Adam Kirchhoff
for multiple useful testing cycles, and Ralf Wostrack for the final fix to get
it working.

PR: i386/75251
Submitted by: anholt

Fix agp_nvidia.c to behave more like the linux driver, fixing DRI on Radeon
9200 according to one responder. The primary issue was not setting some bits
to say that the entries were active, but also fix one place where some memory
wasn't being used as volatile as it should. While here, change some use of ffs
to a relatively short case statement, to make it more obvious what's going on.

PR: kern/71638, kern/72372, kern/71547?
Submitted by: Andrew J. Caines <A.J.Caines@halplant.com>,
Robin Schoonover <end@endif.cjb.net>,
Jason Henson <jason@ec.rr.com>

Fix "struct ifnet" leaks when attach() fails in the middle, e.g.
when mii_phy_probe() or bus_setup_intr() fails. For drivers that
call their detach() in this case, call if_free() there to cover
this case too.

When bus_alloc_resource_any() fails, dc_detach() is called and it
attempts to deallocate busdma tags and resources that haven't been
allocated yet, causing a panic every time a dc interface fails to
attach. Fix by checking that we really have something to dealloc
before calling bus_dma*() functions.

Approved by: jhb
MFC after: 1 week

- Fixup locking and mark MPSAFE.
- Use callout_init_mtx() and static callouts rather than timeout().
- m_getcl() in one place to simplify the code.

Tested by: Gavin Atkinson gavin dot atkinson at ury dot york dot ac dot uk
MFC after: 1 week

Fixed "Memory modified after free" panic in rl_detach() due
to rl_stop() accessing already freed "struct ifnet".

Fixed LOR between rl mutex and some ACPI mutex in rl_detach().

Use if_printf() and device_printf() for printf's and remove the unit number
from the softc.

MFC after: 3 days

Various locking fixes and mark MPSAFE:
- Add locked variants of start(), init(), ifmedia_upd(), and poll() and stop
recursing on the driver lock.
- Add locking to ifmedia_upd() and ifmedia_sts().
- Use callout_*() instead of timeout/untimeout.
- Fix locking in ioctl().

Tested by: Bob Bishop rb at gid dot co dot uk
MFC after: 3 days

Better comment

Fix another instance of old info re: miibus

Fix obsolete comment.

MFC After: 2 days

- Use m_defrag() instead of homerolling our own variant
tulip_mbuf_compress(). If we fail to allocate a new mbuf to copy the
data into, put the mbuf back in the driver's send queue so that we can
retry it later rather than throwing the packet away.
- Use m_devget() instead of doing it inline ourselves in the
TULIP_COPY_RXDATA case. If we fail to allocate an mbuf to copy the data
into, don't forget about the original mbuf cluster. The old code would
lose the pointer and leak the cluster in that case. Now it doesn't lose
it but always sticks the original rx buffer back into the receive ring
after trying to copy the data out and send it up the stack. Also, if we
fail to allocate a new mbuf to copy the data into, log an input error.
Also, don't combine the priming case with the received-a-packet case to
make the code flow a bit clearer and easier to follow.

- Remove non-bus-dma code.
- Remove form feed characters.
- Fixup style of function declarations.
- Assume that an mbuf cluster is big enough to hold an ethernet frame.
(This should really be using m_defrag(), but this diff is just simple
changes for now.)

Major rototill of this driver to add FreeBSD bus-dma support:
- Allocate arrays of metadata for the descriptors in the rx and tx rings
and change the ring pointers to walk the metadata array rather than the
actual descriptor rings. Each metadata object contains a pointer to its
descriptor, a pointer to any associated mbuf, and a pointer to the
associated bus_dmamap_t in the bus_dma case. The mbuf pointers replace
the tulip_txq and tulip_rxq local ifqueue's in the softc.
- Add lots of KTR trace entries using a local KTR_TULIP level which
defaults to 0, but can be changed to KTR_DEV at the top of the file
when debugging.
- Rename tulip_init(), tulip_start(), tulip_ifinit(), and tulip_ifstart()
to tulip_init_locked(), tulip_start_locked(), tulip_init(), and
tulip_start(), respectively, to match the convention in other drivers.
- Add a TULIP_SP_MAC() macro to encode two bytes of the MAC address into
the setup buffer and use that in place of lots of BYTE_ORDER #ifdef's.
Also, remove an incorrect XXX comment I added earlier, the driver was
correct (at least it does the same thing dc(4) does). TULIP_SP_MAC
was shamelessly copied from DC_SP_MAC() in dc(4).
- Remove the #ifdef'd NetBSD bus-dma code and replace it with FreeBSD
bus-dma code that not only compiles but even works at runtime.
- Use callout_init_mtx() instead of just callout_init().
- Correct the various wrapper macros for bus_dmamap_sync() for the rx
and tx buffers to only ask for the sync ops that they actually need.
- Tidy the #ifdef TULIP_COPY_RXDATA code by expanding an #ifdef a bit
so it becomes easier to read at the expense of a couple of duplicated
lines of code. Also, use m_getcl() to get an mbuf cluster rather than
MGETHDR() followed by MCLGET().
- Maintain the ring free (ri_free) count for the rx ring metadata since
we no longer have tulip_rxq.ifq_len around to indicate how many mbuf's
are currently in the rx ring.
- Add code to teardown bus_dma resources when attach fails and generally
fixup attach to do a better job of cleaning up when it fails. This
gets us a good bit closer to possibly having a detach method someday
and making this driver an unloadable module.
- Add some functions that can be called from ddb to dump the state of
a descriptor ring and to dump the state of an individual descriptor.
- Various comment grammer and spelling fixes.

I have bus-dma turned on by default, but I've left the non-bus-dma code
around so that it can be turned off to aid in debugging should any problems
turn up later on. I'll be removing the non-bus-dma code in a subsequent
commit.

Various fixups to locking:
- Remove a lot of superfluous locking during attach. There is no need
to lock access to the driver until some other thread has a way of getting
to it. For ethernet drivers the other ways include registering an
interrupt handler via bus_setup_intr(), calling ether_ifattach() to hook
into the network stack, and kicking off a callout-driven timer via
callout_reset().
- Use callout_* rather than timeout/untimeout.
- Break out of xl_rxeof() if IFF_DRV_RUNNING is clear after ifp->if_input
returns to handle the case where the interface was stopped while we were
passing a packet up the stack. Don't call xl_rxeof() in xl_rxeof_task()
unless IFF_DRV_RUNNING is set. With these fixes in place, any
outstanding task will gracefully terminate as soon as it gets a chance to
run after the interface has been stopped via xl_stop(). As a result,
taskqueue_drain() is no longer required in xl_stop(). The task is still
drained in detach() however to make sure that detach() can safely destroy
the driver mutex at the end of the function.
- Lock the driver lock in the ifmedia callouts and don't lock across
ifmedia_ioctl() in xl_ioctl().

Note: glebius came up with most of (3) as well independently. I took a
rather roundabout way of arriving at the same conclusion.

MFC after: 3 days

Fixup locking and mark MPSAFE:
- Add locked versions of start and init. The SRM_MEDIA code in dc_init()
stayed in dc_init() instead of moving to dc_init_locked() to make the
locking saner.
- Use callout_init_mtx().
- Fixup locking in detach and ioctl.
- Lock the driver in the ifmedia callouts.
- Don't recurse on the driver lock.
- De-spl.

MFC after: 3 days

Fixup locking for sf(4) and mark MPSAFE:
- Add locked variants of start, init, and ifmedia_upd.
- Use callout_* instead of timeout/untimeout.
- Don't recurse on the driver lock.
- Fixup locking in ioctl.
- Lock the driver lock in the ifmedia handlers rather than across
ifmedia_ioctl().

Tested by: brueffer
MFC after: 3 days

- Use htole32() instead of using bswap32() conditional on #if BYTE_ORDER.
- Don't set IFF_ALLMULTI in our ifnet's if_flags if we end up allowing
all multicast due to limits in the MAC receive filters in hardware.

Requested by: rwatson (2)

Remove the unused TULIP_CSR_{READ,WRITE}BYTE macros.

Use callout_init_mtx() to simplify the stats callout.

MFC after: 3 days

Use device_printf() and if_printf() and remove ste_unit from the softc.

* Solve "No PHY found" problem for more Yukon Lite variants.
* Catch a bus attach error.
* Improve locking.

MFC after: 6 days

Use device_printf() and if_printf() and remove dc_unit from softc.

Use if_printf() and device_printf() and axe sf_unit from the softc as a
result.

Call tulip_start() rather than tulip_ifstart() from the interrupt handler
to avoid recursing on the driver lock. Not sure why my test box didn't
catch this earlier.

MFC after: 3 days

Propagate rename of IFF_OACTIVE and IFF_RUNNING to IFF_DRV_OACTIVE and
IFF_DRV_RUNNING, as well as the move from ifnet.if_flags to
ifnet.if_drv_flags. Device drivers are now responsible for
synchronizing access to these flags, as they are in if_drv_flags. This
helps prevent races between the network stack and device driver in
maintaining the interface flags field.

Many __FreeBSD__ and __FreeBSD_version checks maintained and continued;
some less so.

Reviewed by: pjd, bz
MFC after: 7 days

Fix up the locking in pcn(4) and mark it MPSAFE.
- Add locked versions of the init() and start() methods.
- Use callout_*() rather than timeout().
- Make the driver lock non-recursive.
- Push down locking in detach() and ioctl().
- Fix the tick routine to bail if the interface has been stopped and use
callout_drain() in detach() after the call to stop().
- Lock the driver lock in the ifmedia handlers.

Tested by: Ketrien I. Saihr-Kesenchedra ketrien at error404.nls.net
MFC after: 1 week

Recognize D-Link DGE-528(T) Gigabit as an re(4) device.

Submitted by: Andrus Nomm <andrus@members.ee>
PR: 76780
MFC After: 1 week

Modify device drivers supporting multicast addresses to lock if_addr_mtx
over iteration of their multicast address lists when synchronizing the
hardware address filter with the network stack-maintained list.

Problem reported by: Ed Maste (emaste at phaedrus dot sandvine dot ca>
MFC after: 1 week

- Use callout_*() rather than timeout() to periodically poll the media.
- Add locking to protect the softc and mark this driver as MP safe. There
are still some edge cases with multiport cards that need more locking
work.

MFC after: 1 week
Tested on: alpha

Fix a typo in the name of the dmamap for a bus_dmamap_sync().

MFC after: 3 days

The 575A doesn't have funcregs in memio. So don't claim that it does.
This gets my 575A card probing.

Card provided by: James Flemer
MFC After: 3 days

- Use the PCIR_BAR() macro rather than hardcoding rids.
- Use pci_enable_busmaster() rather than fiddling with the PCI command
register directly.

Don't set if_start to tulip_ifstart all over the place. It is already
set in tulip_attach() and its value is never changed, so all the extra sets
are redundant. I'm guessing that at some point in time de(4) had an
alternate start routine, but that hasn't been true in recent history.

Remove conditional code that has largely rotted that is also not on by
default:
- TULIP_NEED_FASTTIMEOUT - tulip_fasttimeout() wasn't called anywhere
- BIG_PACKET - only worked on i386 anyway
- TULIP_USE_SOFTINTR - doesn't compile and was never updated to handle
new netisr registration
- non-FreeBSD code

Use get_cyclecount() rather than hardcoding rdtsc and rpcc in asm for i386
and alpha, respectively.

Fix a typo and some whitespace nits.

NET_LOCK_GIANT() when entering network code.

Pointy hat to: glebius
Reported by: rodrigc

ALTQify ste(4).

Requested and tested by: <nike_d at cytexbg dot com>
Approved by: re (scottl)

Fix ifnet fallout in if_ti.

Reviewed by: brooks
Approved by: re

Make the initialization in the AGPv3 case match that of Linux. Fixes hangs on
X startup with DRI enabled, with a v3-capable card.

Tested by: Tom McLaughlin <tmclaugh@sdf.lonestar.org>
Approved by: re (scottl)

Move if_alloc() up so it's before mii_phy_probe().

Stop embedding struct ifnet at the top of driver softcs. Instead the
struct ifnet or the layer 2 common structure it was embedded in have
been replaced with a struct ifnet pointer to be filled by a call to the
new function, if_alloc(). The layer 2 common structure is also allocated
via if_alloc() based on the interface type. It is hung off the new
struct ifnet member, if_l2com.

This change removes the size of these structures from the kernel ABI and
will allow us to better manage them as interfaces come and go.

Other changes of note:
- Struct arpcom is no longer referenced in normal interface code.
Instead the Ethernet address is accessed via the IFP2ENADDR() macro.
To enforce this ac_enaddr has been renamed to _ac_enaddr.
- The second argument to ether_ifattach is now always the mac address
from driver private storage rather than sometimes being ac_enaddr.

Reviewed by: sobomax, sam

Remove bus_{mem,p}io.h and related code for a micro-optimization on i386
and amd64. The optimization is a trivial on recent machines.

Reviewed by: -arch (imp, marcel, dfr)

Calling xl_rxeof() at the end of xl_start_locked() leads to recursion
in case of IP fast forwarding. Enqueue a taskqueue(9) task instead of
calling xl_rxeof() directly.

Reported & tested by: Slava Alpatov
Reviewed by: wpaul
MFC after: 1 week

Deal with failed malloc calls[1].
While there also check for failed device_add_child calls.

Found by: Coventry Analysis tool[1].
Submitted by: sam[1]
Approved by: pjd (mentor)
MFC after: 1 week

Fix newer Xircom CBE2-100 cards that were reporting
dc0: MII without any PHY!
We have to enable the connection to the MII first. Doing so fixes the
problem cards without breaking the older, working cards.

Bad card provided by: deischen

Invert conditional and use continue to reduce nesting.

Add nForce3-250.

nVidia AGP chipsets beyond nForce2 are AMD64-specific.
So move the AGP support to there.

Submitted by: Jung-uk Kim <jkim@niksun.com>

If resource allocation fails, we could wind up freeing the cdev without it
being allocated. Add a simple check for this.

Submitted by: yongari

Bring rev 1.31 to the AMD64 platform.
This adds support for the SiS intergrated NIC on some Athlon64 motherboards.
The MAC address is stored in the APC CMOS RAM and this fixes the
sis driver ending up with a 00:00:00:00:00:00 MAC address.

Submitted by: Stasys Smailys <ssmailys@komvista.lt>

deal with malloc failure when setting up the multicast filter

Noticed by: Coverity Prevent analysis tool

handle malloc failure and sk_vpd_prodname potentially being null for
other reasons

Noticed by: Coverity Prevent analysis tool
Reviewed by: bz, jmg

xl(4) meets polling(4). Hardware for this work kindly provided by
Eric Masson.

MFC after: 3 weeks

Start the process of modernizing the Tigon driver by using busdma for the
descriptor and configuration data. Thanks to Martin Nilsson for providing
hardware.

Do not try to free non allocated memory in error case.
Do our best to plug some memory leaks (VPD data, jumbo memory buffer,...).
Log if we cannot free because memory still in use[1].
Change locking to avoid ''acquiring duplicate lock of same
type: "network driver"'' and potential deadlock. Also seems to fix LOR #063.

[1] This change does not solve problems if buffers are still in use when
unloading if_sk.ko. There is ongoing work which will address jumbogram
allocations in a more general way.

PR: kern/75677 (with changes, no mii fixes in here)
Tested by: net, Antoine Brodin (slightly different version)
Approved by: rwatson (mentor)
MFC after: 5 days

* Lower interrupt moderation timer 200->100.
Obtained from: NetBSD if_sk.c rev. 1.11

* Make interrupt moderation configurable via sysctl/tuneable.
PR: kern/41220
Approved by: rwatson (mentor)

* Improve chip identification.
Obtained from: NetBSD if_sk.c rev. 1.11

* Take PHY out of reset for Yukon Lite Rev. A3.
Submitted by: postings on net@ in thread "skc0: no PHY found", 2005-02-22

Tested by: net
Approved by: rwatson (mentor)
MFC after: 5 days

* When adding/deleting multicast addresses, only whack the address filter
if the interface is marked RUNNING.
Obtained from: NetBSD if_sk.c rev. 1.12

* Don't initialize the card (and start an autonegotiation) every time the IP
address changes. Makes 'dhclient sk0' invocations way faster and more
consistant. i.e. one DHCPREQUEST elicits the DHCPACK.
Obtained from: OpenBSD if_sk.c rev. 1.56

* Additional locking changes in sk_ioctl.

PR: kern/61296 should see improvements by the last two.
Approved by: rwatson (mentor)
MFC after: 5 days

Bugger, wiped out a needed comma in the previous commit.

The DC driver asks for an alignment of PAGE_SIZE for data buffers, but also
asks that each buffer be (2048 * 256) bytes long. I suspect that alignment
isn't a real requirement since busdma only recently started honoring it. The
size is also bogus. Fix both of these and stop busdma from trying to
exhaust the system memory pool with bounce pages.

Submitted by: Kevin Oberman
MFC After: 7 days

Add device id for the Ali M1671 host to AGP bridge.

Add back 'n' that went missing in last commit

Fix style(9) issues with __P removal.

Noticed by: bde

Return BUS_PROBE_DEFAULT instead of 0.

Correct the SiS 755 PCI ID. Confirmed against Linux code.

PR: kern/76411
Submitted by: Jonathan Fosburgh, jonathan at fosburgh dot org
Obtained from: Jung-uk Kim, jkim at niksun.com

No use for this AMD64 special-case "return NULL;" in probe now that we don't
do fake "generic" support.

Adjust TX threshold on every TX underrun. Some supported cards (8139-based)
can retransmit on TX underrun and set TOK in addition to TUND. Also add a
check to prevent overflow of the addressable threshold.

This fixes some reports of rl(4) slowness, believed to be related to ALTQ
before.

PR: kern/61448
Submitted by: Tim Draegen-Gilman <timNOeudaemonSPAMnet> (with changes)
MFC after: 1 week

Cleanup debugging code and put it under bootverbose
(includes minor style polishing).

Approved by: rwatson (mentor)

Bring in support for SUGOI LAN GIGA NIC made by System TALKS, Inc from
a RealTek 8169SB.

PR: 74262
Submitted by: Yoshikazu GOTO-san

# Submitter notes that he's unsure of the revision string for 8169SB

Reimplement the fix in rev. 1.126.

OK'ed by: phk

correct direction for bus_dma sync of rx buffer

Submitted by: Tai-hwa Liang

Revert local experiment which leaked into commit.

Fix compilation of DEVICE_POLLING code.

Fix typos from previous commit.

/* -> /*- for license, minor formatting changes

We no longer recurse the mutex.

If we get an interrupt and the interface is down, return before we
grab the lock. This should help a tiny bit on machines where unused
if_sis interfaces share IRQ.

Don't tweak DSP on the ..16 chips.

Fix for an issue with excessive collisions in half duplex mode.

Nail the short cable problem the exact way National says it should be.

Simplify and fix bugs in rx/tx ring cleanup.

Rewrite the rx/tx ring initialization to use pointers instead of arrays.

Eliminate a bunch of unnecessary prototypes.

Move the module related stuff to the bottom of the file. This will
allow us to save prototypes.

Make sure to clear any pending interrupts when we stop the interface.

Close a theoretical race: By the time the watchdog comes around
the interface may have been stopped, so we should not restart it.

Align if else if properly

Style: new-speak functions, remove pointless "return".

Make sis_initl() take a typed argument.
Expect caller to lock before calling sis_stop()
Various style stuff.

Add locked/unlocked variants of sis_init()

Make a locked and unlocked variant of sis_start()

Don't declare variables "register", the compiler ought to know what to do.

Instead of keeping track of the index into the receive ring use the already
implemented "sis_nextdesc" pointer to keep a pointer instead.

Forget about the sis_list_data and sis_ring_data structures and embedd
their fields directly in the softc structure.

This is a no-op which shortens most of the affected source lines
by N * 10 characters.

Prototype busdma callback using the typedef.

Add PCI and device ID's to if_xl to support:

3C920B-EMB-WNM Integrated Fast Ethernet Controller

Submitter reports that the card appears to autonegotiate properly, and
operate well with high levels of NFS traffic.

PR: 75253
Submitted by: "Oleg V. Nauman" <oleg at reis dot zp dot ua>
MFC after: 2 weeks

[1] Remove the generic bridge support from those drivers that had it. The
generic bridge support was biting us more than it helped, whenever a new chipset
came out from a vendor and misprogramming it caused strange hangs or corruption.
[2] Add a large number of PCI IDs based on what the linux drivers support.
Note that the new PCI IDs haven't been tested, they're just *likely* to work.
In particular the VIA AGP 8x chipsets are concerning, due to lack of testing,
possible issues (kern/69953), and not having a nice "does this bridge say it
would do 8x" function. However, this shouldn't make the situation worse, since
these chips would have probed in the past anyway.

Enable sk(4) for ATLQ.

Thoroughly tested by: Ender <ender NO tog SPAM net>
MFC after: 4 weeks

Enable jumbo frames on Yukon variants of sk(4).

In contrast to OpenBSD we enable jumbo frame support
depending on MTU setting (like done for xmac).

Approved by: pjd (mentor)
Obtained from: OpenBSD if_sk.c r1.52 (YU_SMR_MFL_JUMBO flag)
Tested by: Heinz Knocke <knockefreebsd at o2 dot pl>
MFC after: 5 days

Update the Tigon 1 and 2 driver to use the sf_buf API for implementing
zero-copy receive of jumbo frames. This eliminates the need for the
jumbo frame allocator implemented in kern/uipc_jumbo.c and sys/jumbo.h.
Remove it.

Note: Zero-copy receive of jumbo frames did not work without these changes;
I believe there was insufficient locking on the jumbo vm object.

Tested by: ken@
Discussed with: gallatin@

- Simplify pcn_probe() by moving vendor/device matching code to pcn_match().
- Avoid LOR in pcn_probe() by removing useless mutex stuff.

- Move chip ID code into separate function.
- Initialize sc->pcn_type during ATTACH as softc contents may not surivive
from PROBE.
- Print out chip-id to assist with ongoing pcn(4) debugging efforts.

Additional register definitions.

Obtained from: NetBSD

Correct URL of the programming manual.

Add missing /* DEVICE_POLLING */

only clear the IFF_OACTIVE flag when we have a chance of being able to
queue a packet to the hardware... instead of when the hardware queue is
empty..

don't initalize cur_tx now that it doesn't need to be...

Pointed out by: bde

move the lock after the NULL check so we don't have a hard(er) to diagnose
panic...

Pointed out by: Bjoern A. Zeeb

fix the missing lock in sk_jfree (verified w/ an assert)
also fix up handling and proding of the tx, _OACTIVE is now handled
better...

Submitted by: Peter Edwards (sk_jfree)
Obtained from: OpenBSD and/or NetBSD (tx prod)

Commit more debugging output. This is a little bit of using a large hammer,
but sk(4) is so prevalent on AMD64 motherboards we need to reduce the number
of round trips in the mailing lists trying to get sufficient information to
make sure we've got a handle on all the problems and are working towards
making sk(4) solid.

Submitted by: bz

Set ramsize depending on a value from eprom instead of using
hardcoded 128k for Yukon devices. 88E8001 only has 64k of on-chip RAM[1].

[1] http://www.marvell.com/products/pcconn/yukon/Yukon_88E8001_10_073103_final.pdf

Tested by: amd64, current
Approved by: rwatson (mentor)
MFC after: 1 week

Add altq support.

Patch by mlaier.

Approved by: mlaier
MFC after: 2 weeks

Add device polling support

Original patch by me, improvements by ru

Happy birthday to: BSDforen.de!
Approved by: ru
MFC after: 2 weeks

Hide link up/down/media printfs behind bootverbose

Another missing ! in front of IFQ_DRV_IS_EMPTY, this time in an even more
sensitive, but less excercised location (the watchdog). While here use the
*_start_locked function directly to avoid drop, grab, drop lock.

I have to be very careful with future ALTQ patches!

Found & reviewed by: rwatson
MFC after: 3 days

* Correct an off-by-one reading vpd ro data.
* Announce some more fields from ro area for better debugging of broken
sk(4)s on various boards.

Submitted by: Bjoern A. Zeeb <bzeeb-lists@lists.zabbadoz.net>

Forcibly disable interrupts, if we find ourselves servicing one when
the device is suspended or shutting down. This will need to be rethought
slightly if we implement suspend/resume support within vr(4).
This appears to fix the vr_shutdown() panic on SMP machines.

My theory here is there's a race somewhere during vr_detach() with
vr_intr() in the SMP case which was sometimes being triggered,
although quite why this was happening is unclear (vr_stop() also
explicitly disables interrupts by writing to the IMR register).

MFC-to-RELENG_5* candidate.

PR: kern/62889
Tested by: seb at struchtrup dot com
MFC after: 10 days

Workaround for a recursive acquisition of the driver mutex during device
detach; triggered by ether_ifdetach() -> if_delmulti() -> vr_ioctl().
MFC candidate.

PR: kern/62889
MFC after: 3 days

Use VM_ALLOC_NOBUSY to eliminate an unneeded vm_page_wakeup() call and the
synchronization that one entails.

Avoid repeated acquisition and release of the vm object lock inside of
two loops in agp_generic_bind_memory(). As an intended side-effect, all
of the calls to vm_page_wakeup() are now performed with the containing
vm object lock held.

Really really fix typo this time.

Fix typo sc -> dev.

Detach the Rhine completely on shutdown, rather than merely stopping it
as the original logic did. This fixes a race with vr_intr() which was
masked on UP systems and manifested on SMP systems.

PR: kern/62889
MFC after: 1 day

fix (for me) the problems where if_de gets really slow after time
(usually taking 20 seconds to transmit a packet).. no longer fall back
to only transmitting one packet (instead of the entire queue) after we
have processed the entire send queue... I have no idea why we didn't
start seeing this problem ~6 years ago when this code was introduced...

Fix sis, bfe and ndis in the same way dc was fixed:
Do not tell the hardware to send when there were no packets enqueued.

Found and reviewed by: green
MFC after: 1 days

Add PCI ID for VIA K8T800Pro chipset. Tested with agptest and X with DRI
enabled, but not 3D.

Read the MAC address in the EEPROM in the correct byte order. This
is a no-op on little endian architectures, but fixes getting the MAC
address for some dc(4) cards on big endian architectures.

This is a RELENG_5 candidate.

Tested by: gallatin (powerpc), marius (sparc64)
First version of the patch written by: gallatin

Add ALTQ support for dc(4), based upon a mostly-working patch from mlaier.

Conditionalize IFF_NEEDSGIANT, like everything else here, on IS_MPSAFE.
The driver doesn't look any less safe without Giant than with, and works
with IS_MPSAFE set to 1 here, so others should probably test it as such.

fix jumbo frames as much as they can be fixed for re. We now cap the MTU
to 7422 since it appears that the 8169S can't transmit anything larger..
The 8169S can receive full jumbo frames, but we don't have an mru to let
the upper layers know this...

add fixup so that this driver should work on alignment constrained platforms
(!i386 && !amd64)

MFC after: 5 days

style: Move the { back to the else line to match the }.

Discussed with: glebius
X-MFC after: 5.3-Release

Prefer C99's __func__ over GCC's __FUNCTION__.

fix misspelling of TX...

Submitted by: Johan Karlsson
MFC after: 3 days

comment requirement that rx/tx descriptor counts must be equal due to
shared code...
define rx descriptor count in terms of tx
align defines

MFC after: 3 days

Fix oversight reported by Norikatsu Shigemura for the "sym" driver
(which was derived from the "ncr" driver) and add a MODULE_DEPEND
on "cam".

MT5 candidate, IMHO.
MFC after: 1 week

Do not call xl_init_locked() unconditionally when we are bringed UP. Call
it only if we weren't UP before. In some cases xl_init causes long media
re-negotiation, and ppp(8) fails to open PPPoE connection because it sets
IFF_UP every time before opening PPPoE connection.

PR: kern/69133
Patch by: mdodd
Approved by: wpaul, julian (mentor)
MFC after: 1 week

Fix a couple of typos in comments

MFC after: 3 days

Tag a last set of PCI network interfaces as IFF_NEEDSGIANT until they
are either locked down or demonstrated MPSAFE.

Add support Corega CG-LAPCIGT Gigabit Ethernet(8169S)

PR: [FreeBSD-users-jp 80667]
Submitted by: FUJIMOTO Kou <fujimoto@j.dendai.ac.jp>
MFC after: 1 week

if_dc includes locking, but that locking is disabled by a #ifdef
by default. As such, mark if_dc as IFF_NEEDSGIANT until such
time as appropriate locking review and testing can take place,
and the locking can be enabled by default.

RELENG_5 candidate.

Add support for Intel E7205 AGP.

PR: kern/69858
Submitted by: Jacobo Arvelo <unix4all at gulic dot org>

Fix aperture size detection on some ALi chipsets by only using the lowest 4 bits
to check aperture size, avoiding hangs. Maintain the rest of the bits when
setting/unsetting ATTBASE. This essentially matches Linux's AGP driver as well.

PR: kern/70037
Submitted by: Mark Tinguely <tinguely at casselton dot net>
Obtained from: NetBSD

Apply some stylistic changes based off of kern/70037 (content changes to
follow).

fix LOR's in sk. Original patch from dwhite. This moves the memory
allocation earlier on in sk_attach so we don't have to lock until a bit
later.

PR: 69752

put function's name at begining of column...

add pci id for Belkin F5D5005 Gigabit ethernet card.

Unconditionally support the AMD64 GART HW.

AMD64 on-CPU GART support.
This also applies to AMD64 HW running 'i386' OS.

Submitted by: Jung-uk Kim <jkim@niksun.com>
Integration by: obrien

style.9.

- Make OF_getetheraddr() honour the "local-mac-address?" system config
variable. If set to "true" OF_getetheraddr() will now return the unique
MAC address stored in the "local-mac-address" property of the device's
OFW node if present and the host address/system default MAC address if
the node doesn't doesn't have such a property. If set to "false" the
host address will be returned for all devices like before this change.
This brings the behaviour of device drivers for NICs with OFW support/
FCode, i.e. dc(4) for on-board DM9102A on Sun machines, gem(4) and hme(4),
regarding "local-mac-address?" in line with NetBSD and Solaris.
The man pages of the respective drivers will be updated separately to
reflect this change.
- Remove OF_getetheraddr2() which was used as a stopgap in dc(4). Its
functionality is now part of OF_getetheraddr().

Since the if_de driver doesn't contain locking, mark it as
IFF_NEEDSGIANT so that ifp->if_start won't be called without Giant
when running debug.mpsafenet=1.

Revert rev 1.93 and replace it by grabbing the vr lock before calling
mii_pollstat(). The previous was causing the vr lock to recurse.

PR: kern/70189

Minimal fix to prevent crashes when an AGP v2 card is used with the new v3 VIA
chipsets, based on Linux's via-agp.c. On boot, the system selects which AGP
version to use based on the inserted card. If v2 was chosen, the chipset
needs to be programmed with the v2 registers still. Also included in kern/69953
are changes to make the programming of the v3 registers match linux, but that
will be left out until the need to do so is confirmed (want specs or a tester).

PR: kern/69953
Submitted by: Oleg Sharoiko <os@rsu.ru>
Tested by: Oleg Sharoiko <os@rsu.ru>, Geoff Speicher <geoff@speicher.org>
(full version from PR)

Fix rl(4)'s lock behavior upon deinitialization. I would get a panic
when kldunloading due to its private locking being acquired recursively.

Acquire vr lock before entering vr_setcfg() in vr_miibus_statchg(),
since vr_setcfg() expects it.

Reported by: Mike Bristow <mike@urgle.com>

Do not attempt to clean up data that has not been initialized yet.
This fixes two kernel panics on boot when the xl driver fails to
allocate bus/port/memory resources.

Reviewed by: silence on -net

Unbreak DEVICE_POLLING build / LINT. Sorry!

Submitted by: roam

Second part of ALTQ driver modifications, covering:
an(4), ath(4), hme(4), ndis(4), vr(4) and wi(4)

Please help testing: http://people.freebsd.org/~mlaier/ALTQ_driver/

Tested by: Vaidas Damosevicius (an, ath, wi)
Roman Divacky (vr)
Submitted by: yongari (hme)

Update for the KDB framework:
o Call kdb_enter() instead of Debugger().

Further improve locking in xl(4):

- Avoid an additional lock acquire/release when leaving xl_intr(), by
changing xl_start*() to xl_start*_locked(), and calling the appropriate
routine by chip revision (as the DMA descriptors are different).

- Simplify the appropriate routines now that they are called with the
lock held.

This should save a significant amount of CPU cycles spent on servicing
each interrupt for both UP and SMP whilst remaining MPSAFE.

Tested by: rwatson

Apply the long-overdue hatchet of style(9) death to this file.

Further locking improvements for vr(4):
- Add *_locked() entry points as needed to avoid unnecessary lock thrashing.
- Use these entry points wisely.
- Only acquire the lock once when servicing an interrupt.
- Check 'suspended' on interrupt to avoid racing detach.
- Correct a mis-spelled comment.
- Don't take the lock in vr_reset() to avoid lock thrashing in attach.
- Comment this.

Reviewed by: -net (silence)

Add a 'suspended' flag to softc so that we can avoid races on detach.

Actually turn on driver locking in xl(4).

Further rl(4) locking improvements:
- Avoid unnecessary re-acquisition elsewhere by adding *_locked()
entry points as needed.
- Correct locking for the DEVICE_POLLING case.
- Hold the driver lock for the entire duration of interrupt servicing,
to avoid unneeded, expensive re-acquisition; use *_locked() entry
points as needed.

Reviewed by: -net (silence)

Whitespace nit

Eliminate redundant return keywords.

Whitespace pass.

style(9):
- Space before bracketized non-void function returns.
- Space before condition for conditional blocks.

Eliminate redundant return keywords.

Whitespace nits

style(9) pass on prototypes.

Locking cleanup for rl(4).
- Eliminate the use of a recursive mutex.
- Mark the driver INTR_MPSAFE.

This work is incomplete and will be refined in a future commit.
- Most notably, _locked() variants of entry points need to be introduced.
- The mii upcall/downcall may still be racy.
- Add a stubbed-out guard against racing rl_detach() for the time being.

Tested on: UP, debug.mpsafenet && !debug.mpsafenet
Reviewed by: silence on -net

style(9) and whitespace cleanup.
Use C99 types. Use ANSI function definitions. Sort prototypes.
Split long lines correctly. Punctuate/wordsmith comments.
Use device_printf()/if_printf() where possible.

Reviewed by: -net (silence)

Reintroduce and clean up locking in xl(4).

- Eliminate the use of a recursive mutex.
- Mark the driver as INTR_MPSAFE.
- Split the default media choice code out into xl_choose_media() to
avoid making poor assumptions about the state of the lock during attach.
- The miibus upcall/downcall paths may still be racy.
Change to commented-out locking assertions there for now.
- Tested with nfsclient, routed, ssh, ntp, dhclient and quagga bgpd.
- This needs SMP test coverage. I do not have such resources.

Tested on: UP, !debug.mpsafenet && debug.mpsafenet
Hardware: 3C905B-TX (0x905510b7)

Use if_printf() and device_printf() where appropriate, i.e.:
- Use device_printf() during device probe/attach.
- Move if_xname initialization to before xl_reset() is called.
- Use if_printf() at all other times after struct ifnet has been
initialized.

ANSIfy function definitions.
Remove unnecessary return keywords.
Other minor stylistic changes.

Fix whitespace, indentation, long line wrapping and comments.

Fix whitespace and comments.

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

SMPng locking cleanup for vr(4).
- Remove recursive locking situations. Remove the MTX_RECURSE bit.
- Take the lock for any routine which is not called from within if_vr.c
itself; this includes entry points called by newbus, ifnet, callout,
ifmedia, and polling subsystems.
- Remove spl references from the code added to miibus callbacks in rev 1.60.
- Add the INTR_MPSAFE bit.
- Tidy up some assignments; locks are not needed for taking the address
of something at a known offset, for example.
- Tested on the machine this was committed from.

Tested on: UP only, !debug.mpsafenet && debug.mpsafenet
Reviewed by: rwatson

- Another whitespace pass; make locking calls more obvious.
- Use C99 types for vr_miibus_readreg().

style(9) compliance.
Put some braces around the busy-wait loop in vr_rxeoc() to make the
no-op semicolon more obvious.
No functional changes.
Running on the machine I am committing from without problems.

Reviewed by: jmallett

Bring in the first chunk of altq driver modifications. This covers the
following drivers: bfe(4), em(4), fxp(4), lnc(4), tun(4), de(4) rl(4),
sis(4) and xl(4)

More patches are pending on: http://peoples.freebsd.org/~mlaier/ Please take
a look and tell me if "your" driver is missing, so I can fix this.

Tested-by: many
No-objection: -current, -net

Add support for the VIA Apollo KT400/400A/600 AGP host bridges which use
the VIA v3 register offsets.

PR: 68545
Submitted by: Ariff Abdullah <skywizard@mybsd.org.my>

Remove saved_* from dc_softc. They are now no longer needed.

Submitted by: Marius Strobl

Remove burn bridges code that saved/restored the pci config registers
that are now handled in the pci bus layer. They are no longer
necessary.

Be BURN_BRIDGES compliant.

Do the dreaded s/dev_t/struct cdev */
Bump __FreeBSD_version accordingly.

Use NG_FREE_MSG() instead of FREE().

Replace handrolled CRC calculation with ether_crc32_[lb]e().

Add missing <sys/module.h> includes

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

Switch to using C99 sparse initialisers for the type methods array.
Should make no binary difference.

Submitted by: Gleb Smirnoff <glebius@cell.sick.ru>
Reviewed by: Harti Brandt <harti@freebsd.org>
MFC after: 1 week

Wrap the code to save/restore PCI config registers on suspend/resume in
#ifndef BURN_BRIDGES.

Noticed by: phk

Only initialize the if_sis callout as MPSAFE if debug.mpsafenet is set
true. Otherwise, assertion checks for Giant in the network stack will
fail causing a panic.

Reported by: simon

A handler for ioctl(SIOCSIFCAP) should not alter a bit in
if_capenable unless the interface driver is actually able
to toggle the respective capability on and off.

Reviewed by: ru

Use __FBSDID.

In agp_generic_bind_memory(), grab the needed pages before acquiring
the agp mutex. We do this because vm_page_grab() called with the
VM_ALLOC_RETRY flag can sleep.

Pointed out by: alc

Get rid of a lockmgr consumer by making agp(4) use a standard mutex,
since it's always acquiring the lock exclusively. This was tested
with X on an SMP box, with and without WITNESS.

Plug three lock leaks.

Add explicit list of SiS AGP chipsets based on Linux kernel's list.

Prompted by: i386/59503

Replace the lame big endian crc with wpaul's standard big endian crc
algorithm, supplied by wpaul himself. The lame one has an origin
that's been called into question, so rather than argue about that (one
could make an excellent fair use argument), replace it with better
code since that's what FreeBSD is about.

Submitted by: wpaul[1], Klaus Klein

[1] Bill called this a silly bikeshed. Maybe his is not incorrect.

Different VIA host bridges use different offsets to their AGP config
registers, so add a register offset array to the softc. We key off the
device ID to determine which set of register offsets. Currently the 8385
host bridge used on amd64 is the only bridge to use the AGP3_VIA_*
register offsets and all other bridges use the AGP_VIA_* offsets. It is
currently unclear if the AGP3_VIA_* offsets are for VIA bridges that
implement AGP 3.0 bridges or just for amd64 bridges.

Submitted by: Kenneth Culver culverk at sweetdreamsracing dot biz

Remove old cy driver files. They have been repo-copied to sys/dev/cy and
sys/dev/ic and adjusted to work there.

Push down the responsibility for zeroing a physical page from the
caller to vm_page_grab(). Although this gives VM_ALLOC_ZERO a
different meaning for vm_page_grab() than for vm_page_alloc(), I feel
such change is necessary to accomplish other goals. Specifically, I
want to make the PG_ZERO flag immutable between the time it is
allocated by vm_page_alloc() and freed by vm_page_free() or
vm_page_free_zero() to avoid locking overheads. Once we gave up on
the ability to automatically recognize a zeroed page upon entry to
vm_page_free(), the ability to mutate the PG_ZERO flag became useless.
Instead, I would like to say that "Once a page becomes valid, its
PG_ZERO flag must be ignored."

Add support for the AMD 8111.

Use BSD spelling, no SysV.

Boomerang 10/100BT (found in 2c905-TX) chips apparently suffer the
same problems as their Hurricane 575* bretheren in that one could set
the memory mapped port, but that has no effect. Add a quirk for this.

# I'll have to see if I can dig up documentation on these parts to see
# if there's someway software can know this other than a table...

Implemented per-interface polling(4) control.

Implemented per-interface polling(4) control.

Implemented per-interface polling(4) control.

Implemented per-interface polling(4) control.

First driver with user-configurable polling(4).

Actually fix the TX performance with polling(4) enabled
by increasing the TX list size from 64 to 128, which is
adequate for HZ=1000.

Submitted by: Vsevolod Lobko

- Improved the TX performance with polling(4) by only checking the
status registers for error conditions and updating statistics
when there are cycles left (inspired by the nge(4) driver).

- Removed the TX list counter and the producer/consumer gap; it's
enough to just ensure we don't reuse the last (free) descriptor,
as the chip may not have read its next pointer yet. If we reuse
it, the TX may stall under a heavy TX load with polling enabled.

- Dropped code to recharge the watchdog timer, it's pointless; the
watchdog routine will re-init the chip and both RX and TX lists.

- Rewritten TX to use only two pointers to track producer/consumer.
- Added polling(4) support!
- Bugfix: don't forget to set IFF_OACTIVE when TX list is full.
- Minor: tidy up vr_encap().

Converted the isa probe and attach to new-bus so that this driver works
without the (defunct) isa compatibility shims. The new-bus-specific
parts are very similar to the ones for the pci probe and attach.

This was held up too long waiting for a repo copy to src/sys/dev/cy,
so I decided to fix the files in their old place. This gives easier
to read and merge diffs anyway.

The "count" line in src/sys/conf/files won't be changed until after
the repo copy, so old kernel configs that specify a count need not be
(and must not be) changed until then. The count is just ignored in
the driver. One unfinished detail is dynamic allocation of arrays
with <count> and (<count> * 32) entries, and iteration over the arrays.
This is now kludged with a fixed count of 10 (up to 10 cards with up
to 32 ports each).

Prodded by: imp
Submitted by: mostly by imp
Approved by: imp

Add register definitions for the status and command registers for AGP.

PR: 64846
Submitted by: Samy Al Bahra

Add the ability to disable agp devices at the loader prompt. Usage is
hint.agp.0.disabled="1"

Submitted by: jhb

Recharge the watchdog timer if there's still some TX work left.

Before MFC'ing the previous commit, I noticed I'd left out a case.
Add in missing case for i845G in the attach routine. I'll MFC this
with the rest of the change after the 4.10 codefreeze lifts.

Reviewed By: Doug Rabson

Fixed a few bugs in the rl(4) driver:

Under polling(4), we counted non-existent output packets and wasted
CPU cycles, corrected. (PR kern/64975.)

The fix in revision 1.71 to correct resetting of the watchdog timer
was wrong.

In rl(4), the TX list does not have a gap between the consumer and
producer, so the "empty TX list" test was wrong, corrected.

Also, resetting the timer to five each time we know there is still
some TX work to do was a bad idea -- under polling(4), if the chip
goes out to lunch, this results in the watchdog routine to _never_
be called. Instead, let the timer downgrade to zero and fire the
watchdog, then reset it to five when it is zero AND there is some
TX work left. (Most other network drivers need this fix too.)

MFC after: 3 days

Performance tuning.

Moved the RX ring resyncing code to ste_rxeoc(), and only run it
if we were asked to POLL_AND_CHECK_STATUS, under DEVICE_POLLING.
(This significantly reduces the CPU load.)

Improved the RX ring resyncing code by re-checking if the head
is still empty before doing resyncing. This mostly affects the
DEVICE_POLLING mode, where we run this code periodically. We
could start checking with an empty head (well, an empty ring
even), and after doing a few iterations, the chip might write
a few entries, including the head, and we would bogusly consider
this case as requiring resyncing. On a test box, this reduced
the number of resyncs done by a factor of 10.

In ste_txeof(sc), only reset the watchdog timer to zero when
the TX list is completely empty.

Converted ste_tx_prev_idx to a pointer -- faster.

Removed some bitrot.

Moved the statistical counter under hw.ste.rxsyncs.

Suggested by: njl

Under a heavy RX load, at least with D-Link DFE-550TX adapters,
the driver's RX ring head may fall behind the chip, causing the
stuck traffic, disordered packets, etc. Work around this by
adopting the technique of resyncing RX head used in dc(4) and
xl(4) drivers, but do it in a slightly different place to reduce
the number of resyncs needed.

Also, set the NIC's RX polling period to a more meaningful value,
to stop overloading the PCI bus (this also reduces the number of
resyncs by a factor of 3 or more in a long run; the actual number
is very dependent on a nature of the traffic).

Maintain the statistics counter as the hw.ste_rxsyncs sysctl.

In cooperation with: Vsevolod Lobko
OK'ed by: ambrisko
MFC after: 5 days

Added polling(4) support for ste(4).

MFC after: 5 days

Support the D-Link DGE-530T. Mine appears to have a blank eeprom, so assume
they all do and handle that without alarming the user. Also pull in a bit
of defensive code from OpenBSD that triggers when a card is recognised but
not properly classified as either Genesis or Yukon. Not that I could ever
have needed this. :-)

Obtained from: OpenBSD/NetBSD (partially)
MFC after: 2 weeks

Switch ste_encap() over to using m_defrag().

No functional change, the previous ste_encap() was correct WRT
long mbuf chains; this just reduces code duplication.

MFC after: 3 days
Prodded by: ambrisko

Properly reprogram the hardware when IFF_ALLMULTI flag is set.

solid reports that it is buggy *and* that it slows down transmit
speed.

Buggy report: Matt Dillon & others
Slowness report: I can't find the e-mail

MFC After: 1 minute

Convert callers to the new bus_alloc_resource_any(9) API.

Submitted by: Mark Santcroos <marks@ripe.net>
Reviewed by: imp, dfr, bde

Announce ethernet MAC addresss in ether_ifattach().

Recognise the 82845G AGP bridge, and poke it appropriately at
attach/detach time.

Assigning the default behaviour to this particular device is
incorrect, corrupting the video BIOS aperture, and breaking
VESA support in the kernel and XFree86.

Reviewed By: dfr
MFC after: 1 week
PR: kern/62906

Stop setting ifp->if_output to ether_output() since ether_ifattach()
does it for us already.

Device megapatch 4/6:

Introduce d_version field in struct cdevsw, this must always be
initialized to D_VERSION.

Flip sense of D_NOGIANT flag to D_NEEDGIANT, this involves removing
four D_NOGIANT flags and adding 145 D_NEEDGIANT flags.

Device megapatch 1/6:

Free approx 86 major numbers with a mostly automatically generated patch.

A number of strategic drivers have been left behind by caution, and a few
because they still (ab)use their major number.