if_stats(9): Add ifp argument to if_stat..._ref.

This will enable us to pass the ifp through to a dtrace probe inside.

No functional change intended in this change, but this is an API
change visible to modules so it shouldn't be pulled up.

PR kern/58377

fix spelling mistakes, mainly in comments and log messages.

Pass proper status values to rnd_add_uint32(9) as rnd(9) man page claims.

Use callout_setfunc() / callout_schedule().

Use ifmedia_fini().

Adopt <net/if_stats.h>.

branches: 1.156.2;
if_percpuq(9) automatically increments if_ipackets, so don't add number of
RX frames from device's statistics counter to if_ipackets to avoid double
count.

Fix direction of the loop.

Found by the lgtm bot.

branches: 1.154.2;
Don't automatically set ec_capenable's ETHERCAP_VLAN_HWTAGGING bit in
vlan_config() to make it user-controllable. Instead, set the bit in
xxx_attach().

Use ETHER_LOCK()/ETHER_UNLOCK() for all ethernet drivers to protect ec_multi*.

No functional change:
- Simplify MII structure initialization and reference.
- u_int*_t -> uint*_t.
- KNF

In drivers which use MII(4) and have hook SIOC[GS]IFMEDIA which just pass to
ifmedia_ioctl(), the hook is not required because ether_ioctl has it
(if_ethersubr.c rev. 1.160). These drivers don't return ENETRESET in
ifmedia_ioctl(), so no functional change.

Change MII PHY read/write API from:

int (*mii_readreg_t)(device_t, int, int);
void (*mii_writereg_t)(device_t, int, int, int);
to:

int (*mii_readreg_t)(device_t, int, int, uint16_t *);
int (*mii_writereg_t)(device_t, int, int, uint16_t);

Now we can test if a read/write operation failed or not by the return value.

In 802.3 spec says that the PHY shall not respond to read/write transaction
to the unimplemented register(22.2.4.3). Detecting timeout can be used to
check whether a register is implemented or not (if the register conforms to
the spec). ukphy(4) can be used this for MII_MMDACR and MII_MMDAADR.

Note that I noticed that the following code do infinite loop in the
read/wirte function. If it accesses unimplemented PHY register, it will hang.
It should be fixed:

arm/at91/at91emac.c
arm/ep93xx/epe.c
arm/omap/omapl1x_emac.c
mips/ralink/ralink_eth.c
arch/powerpc/booke/dev/pq3etsec.c(read)
dev/cadence/if_cemac.c <- hkenken
dev/ic/lan9118.c


Tested with the following device:

axe+ukphy
axe+rgephy
axen+rgephy (tested by Andrius V)
wm+atphy
wm+ukphy
wm+igphy
wm+ihphy
wm+makphy
sk+makphy
sk+brgphy
sk+gentbi
msk+makphy
sip+icsphy
sip+ukphy
re+rgephy
bge+brgphy
bnx+brgphy
gsip+gphyter
rtk+rlphy
fxp+inphy (tested by Andrius V)
tlp+acphy
ex+exphy
epic+qsphy
vge+ciphy (tested by Andrius V)
vr+ukphy (tested by Andrius V)
vte+ukphy (tested by Andrius V)

Not tested (MAC):
arm:at91emac
arm:cemac
arm:epe
arm:geminigmac
arm:enet
arm:cpsw
arm:emac(omac)
arm:emac(sunxi)
arm:npe
evbppc:temac
macppc:bm
macppc:gm
mips:aumac
mips:ae
mips:cnmac
mips:reth
mips:sbmac
playstation2:smap
powerpc:tsec
powerpc:emac(ibm4xx)
sgimips:mec
sparc:be
sf
ne(ax88190, dl10019)
awge
ep
gem
hme
smsh
mtd
sm
age
alc
ale
bce
cas
et
jme
lii
nfe
pcn
ste
stge
tl
xi
aue
mue
smsc
udav
url

Not tested (PHY):
amhphy
bmtphy
dmphy
etphy
glxtphy
ikphy
iophy
lxtphy
nsphyter
pnaphy
rdcphy
sqphy
tlphy
tqphy
urlphy

branches: 1.149.2;
Implement the BPF direction filter (BIOC[GS]DIRECTION). It provides backward
compatibility with BIOC[GS]SEESENT ioctl. The userland interface is the same
as FreeBSD.

This change also fixes a bug that the direction is misunderstand on some
environment by passing the direction to bpf_mtap*() instead of checking
m->m_pkthdr.rcvif.

branches: 1.148.2;
VLAN ID uses pkthdr instead of mtag now. Contributed by s-yamaguchi@IIJ.

I just commit by proxy. Reviewed by joerg@n.o and christos@n.o, thanks.
See http://mail-index.netbsd.org/tech-net/2017/09/26/msg006459.html

XXX need pullup to -8 branch

branches: 1.147.6;
Apply deferred if_start to more drivers

branches: 1.146.2;
Move bpf_mtap and if_ipackets++ on Rx of each driver to percpuq if_input

The benefits of the change are:
- We can reduce codes
- We can provide the same behavior between drivers
- Where/When if_ipackets is counted up
- Note that some drivers still update packet statistics in their own
way (periodical update)
- Moved bpf_mtap run in softint
- This makes it easy to MP-ify bpf

Proposed on tech-kern and tech-net

branches: 1.145.2;
Introduce m_set_rcvif and m_reset_rcvif

The API is used to set (or reset) a received interface of a mbuf.
They are counterpart of m_get_rcvif, which will come in another
commit, hide internal of rcvif operation, and reduce the diff of
the upcoming change.

No functional change.

Introduce softint-based if_input

This change intends to run the whole network stack in softint context
(or normal LWP), not hardware interrupt context. Note that the work is
still incomplete by this change; to that end, we also have to softint-ify
if_link_state_change (and bpf) which can still run in hardware interrupt.

This change softint-ifies at ifp->if_input that is called from
each device driver (and ieee80211_input) to ensure Layer 2 runs
in softint (e.g., ether_input and bridge_input). To this end,
we provide a framework (called percpuq) that utlizes softint(9)
and percpu ifqueues. With this patch, rxintr of most drivers just
queues received packets and schedules a softint, and the softint
dequeues packets and does rest packet processing.

To minimize changes to each driver, percpuq is allocated in struct
ifnet for now and that is initialized by default (in if_attach).
We probably have to move percpuq to softc of each driver, but it's
future work. At this point, only wm(4) has percpuq in its softc
as a reference implementation.

Additional information including performance numbers can be found
in the thread at tech-kern@ and tech-net@:
http://mail-index.netbsd.org/tech-kern/2016/01/14/msg019997.html

Acknowledgment: riastradh@ greatly helped this work.
Thank you very much!

Convert sys/dev to use <sys/rndsource.h>.

branches: 1.142.4;
Merge tls-earlyentropy branch into HEAD.

branches: 1.141.2;
Remove unused variable

branches: 1.140.2; 1.140.4;
Fix mii_statchg to take a 'struct ifnet *' instead of device_t. This fixes
problem with a common MDIO bus used for multiple interfaces.
Some drivers converted to CFATTACL_DECL_NEW.

Entropy-pool implementation move and cleanup.

1) Move core entropy-pool code and source/sink/sample management code
to sys/kern from sys/dev.

2) Remove use of NRND as test for presence of entropy-pool code throughout
source tree.

3) Remove use of RND_ENABLED in device drivers as microoptimization to
avoid expensive operations on disabled entropy sources; make the
rnd_add calls do this directly so all callers benefit.

4) Fix bug in recent rnd_add_data()/rnd_add_uint32() changes that might
have lead to slight entropy overestimation for some sources.

5) Add new source types for environmental sensors, power sensors, VM
system events, and skew between clocks, with a sample implementation
for each.

ok releng to go in before the branch due to the difficulty of later
pullup (widespread #ifdef removal and moved files). Tested with release
builds on amd64 and evbarm and live testing on amd64.

branches: 1.138.2; 1.138.6;
Add support for some fxp devices from FreeBSD and OpenBSD.
{Free,Open}BSD say that we have to do some work to make fxp
stable.

IFF_PROMISC implys IFF_ALLMULTI.
Should fix PR#43186.

branches: 1.136.2;
Include sys/proc.h for tsleep, wakeup.

Push the bpf_ops usage back into bpf.h. Push the common ifp->if_bpf
check into the inline functions as well the fourth argument for
bpf_attach.

Check whether the fxp(4) actually attached before calling fxp_stop().

In fxp_detach(), fxp_stop(), first. fxp_stop() stops the callout.
Destroy the callout in fxp_detach().

branches: 1.132.2;
Make fxp at cardbus detach during shutdown.

Stop calling (*cardbus_ctrl) to enable bus mastering, I/O and memory
spaces on the CardBus bridge. cbb(4) always enables that stuff,
anyway. In the process, avoid remembering what BAR we mapped by
writing CARDBUS_{IO,MEM}_ENABLE to sc_cben or sc_cbenable, and
record the BAR in use sc_bar, instead.

Replace more CARDBUS_ constants with PCI_ constants.

Compile-tested, only.

branches: 1.131.2;
Redefine bpf linkage through an always present op vector, i.e.
#if NBPFILTER is no longer required in the client. This change
doesn't yet add support for loading bpf as a module, since drivers
can register before bpf is attached. However, callers of bpf can
now be modularized.

Dynamically loadable bpf could probably be done fairly easily with
coordination from the stub driver and the real driver by registering
attachments in the stub before the real driver is loaded and doing
a handoff. ... and I'm not going to ponder the depths of unload
here.

Tested with i386/MONOLITHIC, modified MONOLITHIC without bpf and rump.

Simplify activation routines: don't call mii_activate(), it's a
no-op. Don't block interrupts, if_deactivate() does it.
Compile-tested, only.

Pull a fix from hme.c rev 1.73 (to #if 0'ed out part):
> Fix a bug in calculation of checksum deduction:
> - To get 16 bit one's complement value from uint32_t variable,
> higher 16 bits should be ignored.
> - RFC 1624 describes methods to recalculate checksum field in headers,
> i.e. one's complement of one's complement sum that could be 0x0000,
> but we don't have to use the strategy to deduct one's complement sum
> itself which won't be zero but should be 0xffff.

Tweak comments and conditionals about EXT_RFA and IPCB.

u_intNN_t -> uintNN_t

Computed checksum value by the FXPF_82559_RXCSUM feature includes data
in IP headers, so we have to deduct not only IP option headers but all
IP headers. But in TCP/UDP layer we can assume the IP header is valid
and a sum of the IP header part should be 0xffff, so we don't have to
bother to deduct it from the computed checksum.

Add TCPv4/UDPv4 RX hardware checksum support for i82559 and later chips
which don't have EXT_RFA and IPCB support. From hme(4) driver and
FreeBSD's fxp(4). Tested on i82559.

XXX: Probably we should have a common function to parse RX packet headers
XXX: to handle a raw checksum value and share it among hme(4) and gem(4) etc.

Use FXPF_EXT_RFA flag instead of FXPF_EXT_TXCB to see IPCB capability
because EXT_RFA for RX cksum is always available with IPCB for TX cksum
but i82558 and i82559 have only EXT_TXCB without IPCB.

Tested on the following fxp cards:
fxp0 at pci0 dev 14 function 0: Intel i82557 Ethernet, rev 2
fxp0 at pci0 dev 14 function 0: i82559 Ethernet, rev 8
fxp0 at pci0 dev 14 function 0: i82550 Ethernet, rev 12

- remove FXPF_IPCB flag. it should always/only be used with the code
conditional on FXPF_EXT_TXCB, so, replace all uses with that
- for the pci frontend, reestablish some flags lost the the prior
changes and simplify one of the cases

this fixes PR 40677 and may fix PR 40431.

branches: 1.122.2;
The PCI revision numbers are unique to a PCI vendor/product
ID pair. Misuse of the revision numbers was causing some of the chip
features to be disabled on some integrated Intel chips. So, move the
determination of the features into the bus frontend, where the
vendor/product ID is known. (Note: sc_rev should be removed. The
microcode patch stuff is also busted and needs to be fixed.) Also,
poll the actual flow control status in inphy, rather than making
assumptions.

contributed anonymously.

Wrap long lines.

Don't pass uint8_t values to le16toh() in fxp_rx_hwcksum().
fxp(4)'s RX hwcksum results weren't used at all on big endian machines.

Checked by i82550 and vmstat -ev on macppc GENERIC kernel with
options INET_CSUM_COUNTERS,TCP_CSUM_COUNTERS,UDP_CSUM_COUNTERS.

Add a missed htole32() on the previous ip4csum-tx bug workaround.

Sort Tx/Rx macro in previous.

Add a workaround for hardware ip4csum-tx bug and enable it.
Confirmed on i82550 rev 12 and UDP fragment packets by ttcp(1).

Call BUS_DMASYNC_PREREAD more strictly on polling DMA descriptors.

branches: 1.115.2; 1.115.4;
Calling fxp_init within the interrupt handler results in
a hang (continuous assertion of FXP_SCB_STATACK_RNR).
Instead do it in the ioctl routine after receiving a
signal from the interrupt handler.

- device/softc split for fxp(4)

branches: 1.113.2; 1.113.4; 1.113.6;
Remove clause 3 and 4 from TNF licenses

branches: 1.112.2; 1.112.4;
use aprint_*_dev and device_xname

branches: 1.111.6;
Start patching up the kernel so that a network driver always has
the opportunity to handle an ioctl before generic ifioctl handling
occurs. This will ease extending the kernel and sharing of code
between drivers.

First steps: Make the signature of ifioctl_common() match struct
ifinet->if_ioctl. Convert SIOCSIFCAP and SIOCSIFMTU to the new
ifioctl() regime, throughout the kernel.

Make many ethernet drivers share the common code for MII media
handling, ether_mediastatus() and ether_mediachange(). Check for
a non-ENXIO error return from mii_mediachg(). (ENXIO indicates
that a PHY is suspended.)

This patch shrinks the source code size by 979 lines. There was
a 5100-byte savings on the NetBSD/i386 kernel configuration, ALL.

I have made a few miscellaneous changes, too:

gem(4): use LIST_EMPTY(), LIST_FOREACH().
mtd(4): handle media ioctls, for a change!
axe(4): do not track link status in sc->axe_link any longer
nfe(4), aue(4), axe(4), udav(4), url(4): do not reset all PHYs
on a change of media

Except for the change to mtd(4), no functional changes are intended.

XXX This patch affects more architectures than I can feasibly
XXX compile and run. I have compiled macppc, sparc64, i386. I
XXX have run the patches on i386 boxen with bnx(4) and sip(4).
XXX Compiling and running on evbmips (MERAKI, ADM5120) is in
XXX progress.

Fix typo in comments.

Fix from YAMAGUCHI Takahiro in PR kern/30023:
Pass correct EEPROM address in fxp_write_eeprom().

Remove the shutdown hook (pmf takes care about it already), as suggested by joerg@

branches: 1.106.2;
Register fxp @ pci to the pmf framework
Kill reference to old powerhook framework.

branches: 1.105.2;
Pull several fixes from OpenBSD's fxp.c:

Rev 1.41:
> use a nop with I bit set at the end of the tx chain. This avoids a race
> between status update and clearing the suspend bit on machines which can't
> write data smaller than 32bits at a time.
This should fix PR port-alpha/30560.

Rev 1.50:
> Fix nasty bug where driver would not correctly catch and handle an rnr
> condition when it was due to the the recieve buffers being exhausted with
> no packet transmits during that time. Symptom was that the fxp would
> simply stop interrupting for the next 15 seconds until the watchdog kicked
> in and reset the chip due to 15 seconds of inactivity, making the fxp very
> poorly behaved when hammered on hard.

Rev 1.61:
> Defer reinitialisation of the RU until after the interrupt handler has had
> a chance to process all pending packets, otherwise the chip may overwrite
> their mbuf clusters after we have freed them.
>
> Eliminates a race that can cause random pool corruption when reconfiguring
> an interface under heavy network load.

And one more change for RX sanity:
- put RU into suspend when the last RFA is processed.

These RNR/RU changes may fix pool corruption problems on fxp.

Tested on AlphaPC164 and i386 with i82559 and i82550.

branches: 1.104.2; 1.104.4;
machine/{bus,cpu,intr}.h -> sys/{bus,cpu,intr}.h

branches: 1.103.2; 1.103.6;
Constify: LLADDR -> CLLADDR. I'm aiming here to make it easier to
identify sockaddr_dl abuse that remains in the kernel, especially
the potential for overwriting memory past the end of a sockaddr_dl
with, e.g., memcpy(LLADDR(), ...).

branches: 1.102.2; 1.102.6;
Merge some of the less invasive changes from the vmlocking branch:

- kthread, callout, devsw API changes
- select()/poll() improvements
- miscellaneous MT safety improvements

branches: 1.101.2; 1.101.4; 1.101.6;
Kill caddr_t; there will be some MI fallout, but it will be fixed shortly.

branches: 1.100.2; 1.100.4; 1.100.8; 1.100.12;
__unused removal on arguments; approved by core.

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

Add "name" parameter to powerhook_establish, to aid debugging. No objections
on tech-kern@

branches: 1.97.14; 1.97.16;
Use device_is_active() rather than testing dv_flags for DVF_ACTIVE
directly.

branches: 1.96.2; 1.96.4; 1.96.6;
Remove leading __ from __(const|inline|signed|volatile) -- it is obsolete.

merge ktrace-lwp.

use VLAN_OUTPUT_TAG()

Make receiver lock-up workaround conditional on setting link1, as we have
too many false positives - should address PR/30505

Make microcode loading work on big endian machines.
Reported and patch supplied by yongari@freebsd a long time ago.

branches: 1.91.2;
- sprinkle const.
- avoid variable shadowing.

split IFCAP_CSUM_xxx to IFCAP_CSUM_xxx_Rx and IFCAP_CSUM_xxx_Tx.

branches: 1.89.10;
Use log() instead of printf().

When adding/deleting multicast addresses, only whack the address
filter if the interface is marked RUNNING.

Fixes kern/27678.

Fix a typo in a comment.

Patch from HITOSHI Osada:
* Add PAUSE-related event counters.
* Return flow control bits in fxp_mii_mediastatus().

Patch from HITOSHI Osada:
* Newer chips do not need the receiver lock-up workaround; detect when it
is required.

When we are recovering from a resource exhaustion on receive, make sure
that the RU is stopped before issuing an RU_START.
Fix suggested by Takahiro Kambe in PR kern/10622.

Back out revision 1.82. It seems to cause a number of device timeouts
on x86 boxes. There must be a better way to deal with this that works
well for everyone. In the meantime, back out to the version that works
for more people.

Be more careful about issuing CU_RESUME in fxp_start()--only doing
it if we think it's probably necessary. Then do it again in the
tx interrupt handler, if we again think it's necessary. This
reduces the number of commands we issue the chip. Prior to this
change, the i82550 (running without extended feature set, so like
a '559) would sometimes fail in fxp_scb_wait() prior to issuing
another CU_RESUME, resulting in
fxp0: WARNING: SCB timed out!
messages on the console, often followed by device timeouts.

branches: 1.81.2;
support h/w assisted vlan tag insertion.

support h/w assisted vlan tag removal.

On some 'fast' i386 motherboards, the timing between consecutive reads
and writes of the EEPROM are too fast so invalid data is returned. So we
increase the DELAY()s. The right thing would be to check a ready bit
on the E^2 if such a thing exists. (checked by potr)

- disable ipv4 checksum offloading as it seems to have a problem.
- add comments to describe it.

Fix some typos. From Tom Cosgrove via jmc@openbsd.

In fxp_intr(), bail early without touching any registers if sc_enabled is
zero, otherwise we could cause PCI master aborts if the card is in D3 power
state and we happen to share a PCI interrupt line with some other device.

branches: 1.75.2;
ip/tcp/udp checksum offloading support for i82550.
largely from FreeBSD.

check and MCLAIM correct mbuf in fxp_copy_small path.

Add MBUFTRACE kernel option.
Do a little mbuf rework while here. Change all uses of MGET*(*, M_WAIT, *)
to m_get*(M_WAIT, *). These are not performance critical and making them
call m_get saves considerable space. Add m_clget analogue of MCLGET and
make corresponding change for M_WAIT uses.
Modify netinet, gem, fxp, tulip, nfs to support MBUFTRACE.
Begin to change netstat to use sysctl.

Use bus_dmamap_load_mbuf() in fxp_add_rfabuf().

Use aprint_*().

synchronous, not syncronous.

Cosmetic changes.

Fix signed/unsigned comparison warnings.

Change the "dontcare bits" argument to ifmedia_init() to IFM_IMASK
so that it is possible select PHY instances other than the first
one (instance zero), if there is more than one PHY attached.

Add newlines to the WARNING messages printed for the dynamic standby bug.

branches: 1.65.2;
The FIX_RESUME_BUG hack does not work, so remove it.
Also, attempt to clarify the message slightly when updating the EEPROM.

branches: 1.64.2;
Add support for the "CPUSaver" receive interrupt mitigating microcode
on the following PRO/100 chips:

* i82558 step A4
* i82558 step B0
* i82559 step A0
* i82559S step A
* i82550
* i82550 step C

The interrupt delay is configurable on all microcodable chips. The
maximum "bundle" size (packet count) is configurable on all but the
i82558.

The microcode is enabled by setting IFF_LINK0 on the interface.

Derived from code in FreeBSD.

Address Intel 82801BA/82801BAM Specification Update, Errata #30:

The ICH on-board Ethernet and some i82559 chips have a bug which
will cause a PCI protocol violation if the chip receives a CU_RESUME
command as it is entering the IDLE state by deasserting #CLKRUN.
(This is the so-called "resume bug" that we previously had an incomplete
work-around for on ICH chipsets.)

The work-around is to disable Dynamic Standby Mode, such that the
chip will never deasert #CLKRUN. Dynamic Standby Mode is disabled
by clearing a bit in the EEPROM and updating the EEPROM (and EEPROM
checksum).

Unfortunately, the chip will only consult the EEPROM setting after
a PCI bus reset, so a system reboot is required once the EEPROM
has been updated (the EEPROM update only needs to happen once,
and the driver usses a warning instructing the user to reboot the
system once the work-around has been applied).

Issue pointed out by David Brownlee, and code more-or-less lifted
from FreeBSD.

Move the code which shifts bits into the EEPROM into its own function.

add/cleanup RCSID

Spell 'occurred' with two 'r's.

branches: 1.59.2;
Advertize pause capability (802.3x flow control) to peer.

Use BUS_DMA_READ and BUS_DMA_WRITE in some obvious places.

branches: 1.57.2;
bzero -> memset

bcopy -> strcpy

Split the Tx and Rx interrupt routines into separate functions,
and add some (optional) event counters.

Don't need INET or NS includes here.

Define some extra configuration block bits used on the i82550 (a.k.a.
PRO 100/S), including the extended RFA format. Don't hard-code the
RFA size to allow us to use the extended format in the future.

Some changes from the new FreeBSD `fxp' driver:
- Add some additional config block bits for the i82558/i82559.
- Change the config block template to only fill in the must-be-one
reserved bits, leave fxp_init() to fill in all the important things.
- On the i82558/i82559, we can use "Receive Long Frames" rather than
"Save Bad Frames" to support the VLAN MTU.
- Use 802.3x flow control on the i82558/i82559. This is all handled
transparently by the hardware. When in promiscuous mode, allow
wire-watchers to see the flow control frames.
- Use the Extended TxCB format on the i82558/i82559. This places two
Transmit Buffer Descriptors directly in the TxCB, which should cover
the vast majority of packets transmitted. This saves PCI transactions
that would otherwise be required to fetch the TBD list.

With the tansmit queue length changes from yesterday, this gets us up to
11.5MB/s TCP transmit speed, out of an absolute maximum of 12MB/s possible
on a 100Mb/s link.

A little slight cleanup.

Re-arrange the transmit control data somewhat so that we can
use the extended TxCB format (that change will happen later).

Sigh, typo in last.

Fix non-fatal typo -- CB_NOP -> CU_NOP in the ICH work-around (they
have the same value).

The 82801BA built-in Ethernet has a bug which requires us to
issue a NOP before a CU_RESUME when in 10baseT mode. Handle
this.

ANSI'ify.

delint: remove volatile from fxp_mdi_read() decl

branches: 1.44.2;
Fix a problem with the ALTQ changes that can cause bogus memory
refernces. Problem reported by Luke Mewburn.

ALTQ'ify.

Add new powerhook argument values, PWR_SOFTSUSPEND, PWR_SOFTSTANDBY and
PWR_SOFTRESUME. Apm calls powerhook with the values in normal interrupt
priority level while others are protected with splhigh().

Move bpfattach()/bpfdetach() calls into ether_ifattach()/ether_ifdetach().

Use ether_ioctl().

Support ETHERCAP_VLAN_MTU by doing extra error checking upon packet
reception and saving other "bad frames" (i.e. ones that are larger
than the standard Ethernet frame length) if we have VLANs configured
on the interface.

Move the check for "promisc + unicast + not for us" into ether_input(),
and change Ethernet drivers to always pass all received frames to
ether_input() (with a few exceptions, which are documented in the
code).

6 -> ETHER_ADDR_LEN

Since the RFA is allocated in a normal mbuf, as opposed to DMA-safe
memory that is explicitly mapped in a DMA-coherent manner, we must
make sure to PREREAD sync the RFA after noticing a clear "complete"
bit. Without this, the clear bit will linger in the cache, and the
CPU will not notice when the chip updates the bit via DMA later.

From Izumi Tsutsui on port-arm32@netbsd.org.

remove include of <vm/vm.h>. <vm/vm.h> -> <uvm/uvm_extern.h>

branches: 1.34.2;
For all network drivers that call ether_ifattach(), and also
have _detach() functions:
Ensure that softc keeps state about whether the attach succeeded,
and make the detach function return immediately if the attach did
not complete.

branches: 1.33.2;
Increase delay on EEPROM access.
CATS sometimes cannot read MAC address correctly.

Call bus_dmamap_sync() properly before/after reading fcd_stats
in fxp_control_data.

Make two previous changes more correct, as shown by Izumi Tsutsui.

The EEPROM is little-endian.

Fix DMA status reading on big-endian systems. From Sanjay Lal on port-macppc.

And a newline after error string.

Calibrate the timeouts from rev 1.26 by using DELAY(1) so their
values are less machine-dependant. Fixes port-i386/10141, where
spurrious timeouts were being seen.
Also note the line numbers of the timeouts so it can be determined
which is being seen (via __LINE__).

Catch some potentially infinite loops in while() loops if the chip happens
to "go insane" and never complete some operations (dmasync, etc.).

Delete redundant decls of fxp_enable(), fxp_disable() - they're in
i82557var.h.

New callout mechanism with two major improvements over the old
timeout()/untimeout() API:
- Clients supply callout handle storage, thus eliminating problems of
resource allocation.
- Insertion and removal of callouts is constant time, important as
this facility is used quite a lot in the kernel.

The old timeout()/untimeout() API has been removed from the kernel.

Get rid of the powerhook when we detach.

No longer necessary to futz with ifp->if_baudrate here.

(fxp_init): don't clear the FXPF_MII flag as this results in a panic
when detaching

Add activate routine and check device active flag also in one second
tick handler.

Cosmetic changes.

add detach code

Don't dry to diving MIIF_NOISOLATE in the PHY drivers. Instead, pass
flags down from the parent to child vi mii_attach().

Bring some order to the chaos which was the MII code function naming
"conventions".

Take a stab at making this work on big-endian systems.

change comment to mention that this driver also handles the '558 and '559

try to do a better job of figuring out the EEPROM size; the old code
didn't work on (some?) 557-based cards

Call mii_down() as appropriate.

Adapt to mii_phy_probe() change.

Add enable/disable support and EEPROM size selection, prerequisites
for getting the cardbus fxp attachment to work.

Tested on the built-in 82559 in a VAIO Z505S.

From Johan Danielsson / PR8631; changed slightly to KNF.

Fix suspend/resume-related problems observed on fxp0 on Sony Z505S:

Symptoms: system would crash with "data modified on free list" pool
panic from the mbuf cluster pool shortly after a resume. The cluster
in question contained a valid 82557 receive descriptor and an IP
packet. Happened sporadically in normal use. Easiest way for me to
reproduce it was to run tcpdump and a flood ping and do a
suspend/resume cycle or two.

Changes:
- in interrupt handler, if the interface isn't in IFF_RUNNING state,
just ack interrupts and return; don't try to receive packets, queue
new descriptors, etc., etc.,
- add power control hook to take interface down on suspend,
and restart it (if it was up) on resume.
- tweaks to fxp_stop and fxp_shutdown to avoid recursive panics due
to the (now fixed) bug.

branches: 1.8.2; 1.8.4; 1.8.6;
Since we have to go through fxp_init() to properly handle IFF_ALLMULTI
anyway, take advantage of this and greatly simplify the programming
of the multicast filter. This solves the last reported "device timeout"
problem with this driver.

Almost completely rewrite the receive logic, making it as close as possible
to the EPIC/100 driver's (adjusting for the fact that Intel Ethernet chips
are from Pluto):
* Don't allocate receive buffers until the interface is actually brought
up, and release all of them if the interface is taken down.
* Add a knob (defaults to off) which will copy an incoming packet to
a single header mbuf if it is small enough to fit in one, rather than
burning an entire cluster on it. Note that this change will be mostly
moot if/when sbcompress() is changed to handle compressing clusters.

Simplify some of the receive list logic:
* Rather than using a homegrown queue and additional software RX descriptors,
use an ifqueue to queue receive buffers, and M_{GET,SET}CTX() to hook DMA
maps and receive buffers together.

Clean up a bit:
* Macroize a bunch of things to make the code a bit easier to follow.

Only tick the MII if we are using it.

Be a little less selective in the transmit interrupt path.

Use the Interrupt bit in the command block to generate interrupts, rather
than interrupting after every Command Unit Not Active event.

Don't bother with interrupt-driven multicast setup.

Completely rewrite the transmit logic, making it look more like the
EPIC/100 driver's. Also, fix the "all multicast" logic. Also do some
general cleanup.

branches: 1.1.2;
Bus-independent back-end driver for Intel i82557 fast Ethernet chips.

fix spelling mistakes, mainly in comments and log messages.

Pass proper status values to rnd_add_uint32(9) as rnd(9) man page claims.

Use callout_setfunc() / callout_schedule().

Use ifmedia_fini().

Adopt <net/if_stats.h>.

branches: 1.156.2;
if_percpuq(9) automatically increments if_ipackets, so don't add number of
RX frames from device's statistics counter to if_ipackets to avoid double
count.

Fix direction of the loop.

Found by the lgtm bot.

branches: 1.154.2;
Don't automatically set ec_capenable's ETHERCAP_VLAN_HWTAGGING bit in
vlan_config() to make it user-controllable. Instead, set the bit in
xxx_attach().

Use ETHER_LOCK()/ETHER_UNLOCK() for all ethernet drivers to protect ec_multi*.

No functional change:
- Simplify MII structure initialization and reference.
- u_int*_t -> uint*_t.
- KNF

In drivers which use MII(4) and have hook SIOC[GS]IFMEDIA which just pass to
ifmedia_ioctl(), the hook is not required because ether_ioctl has it
(if_ethersubr.c rev. 1.160). These drivers don't return ENETRESET in
ifmedia_ioctl(), so no functional change.

Change MII PHY read/write API from:

int (*mii_readreg_t)(device_t, int, int);
void (*mii_writereg_t)(device_t, int, int, int);
to:

int (*mii_readreg_t)(device_t, int, int, uint16_t *);
int (*mii_writereg_t)(device_t, int, int, uint16_t);

Now we can test if a read/write operation failed or not by the return value.

In 802.3 spec says that the PHY shall not respond to read/write transaction
to the unimplemented register(22.2.4.3). Detecting timeout can be used to
check whether a register is implemented or not (if the register conforms to
the spec). ukphy(4) can be used this for MII_MMDACR and MII_MMDAADR.

Note that I noticed that the following code do infinite loop in the
read/wirte function. If it accesses unimplemented PHY register, it will hang.
It should be fixed:

arm/at91/at91emac.c
arm/ep93xx/epe.c
arm/omap/omapl1x_emac.c
mips/ralink/ralink_eth.c
arch/powerpc/booke/dev/pq3etsec.c(read)
dev/cadence/if_cemac.c <- hkenken
dev/ic/lan9118.c


Tested with the following device:

axe+ukphy
axe+rgephy
axen+rgephy (tested by Andrius V)
wm+atphy
wm+ukphy
wm+igphy
wm+ihphy
wm+makphy
sk+makphy
sk+brgphy
sk+gentbi
msk+makphy
sip+icsphy
sip+ukphy
re+rgephy
bge+brgphy
bnx+brgphy
gsip+gphyter
rtk+rlphy
fxp+inphy (tested by Andrius V)
tlp+acphy
ex+exphy
epic+qsphy
vge+ciphy (tested by Andrius V)
vr+ukphy (tested by Andrius V)
vte+ukphy (tested by Andrius V)

Not tested (MAC):
arm:at91emac
arm:cemac
arm:epe
arm:geminigmac
arm:enet
arm:cpsw
arm:emac(omac)
arm:emac(sunxi)
arm:npe
evbppc:temac
macppc:bm
macppc:gm
mips:aumac
mips:ae
mips:cnmac
mips:reth
mips:sbmac
playstation2:smap
powerpc:tsec
powerpc:emac(ibm4xx)
sgimips:mec
sparc:be
sf
ne(ax88190, dl10019)
awge
ep
gem
hme
smsh
mtd
sm
age
alc
ale
bce
cas
et
jme
lii
nfe
pcn
ste
stge
tl
xi
aue
mue
smsc
udav
url

Not tested (PHY):
amhphy
bmtphy
dmphy
etphy
glxtphy
ikphy
iophy
lxtphy
nsphyter
pnaphy
rdcphy
sqphy
tlphy
tqphy
urlphy

branches: 1.149.2;
Implement the BPF direction filter (BIOC[GS]DIRECTION). It provides backward
compatibility with BIOC[GS]SEESENT ioctl. The userland interface is the same
as FreeBSD.

This change also fixes a bug that the direction is misunderstand on some
environment by passing the direction to bpf_mtap*() instead of checking
m->m_pkthdr.rcvif.

branches: 1.148.2;
VLAN ID uses pkthdr instead of mtag now. Contributed by s-yamaguchi@IIJ.

I just commit by proxy. Reviewed by joerg@n.o and christos@n.o, thanks.
See http://mail-index.netbsd.org/tech-net/2017/09/26/msg006459.html

XXX need pullup to -8 branch

branches: 1.147.6;
Apply deferred if_start to more drivers

branches: 1.146.2;
Move bpf_mtap and if_ipackets++ on Rx of each driver to percpuq if_input

The benefits of the change are:
- We can reduce codes
- We can provide the same behavior between drivers
- Where/When if_ipackets is counted up
- Note that some drivers still update packet statistics in their own
way (periodical update)
- Moved bpf_mtap run in softint
- This makes it easy to MP-ify bpf

Proposed on tech-kern and tech-net

branches: 1.145.2;
Introduce m_set_rcvif and m_reset_rcvif

The API is used to set (or reset) a received interface of a mbuf.
They are counterpart of m_get_rcvif, which will come in another
commit, hide internal of rcvif operation, and reduce the diff of
the upcoming change.

No functional change.

Introduce softint-based if_input

This change intends to run the whole network stack in softint context
(or normal LWP), not hardware interrupt context. Note that the work is
still incomplete by this change; to that end, we also have to softint-ify
if_link_state_change (and bpf) which can still run in hardware interrupt.

This change softint-ifies at ifp->if_input that is called from
each device driver (and ieee80211_input) to ensure Layer 2 runs
in softint (e.g., ether_input and bridge_input). To this end,
we provide a framework (called percpuq) that utlizes softint(9)
and percpu ifqueues. With this patch, rxintr of most drivers just
queues received packets and schedules a softint, and the softint
dequeues packets and does rest packet processing.

To minimize changes to each driver, percpuq is allocated in struct
ifnet for now and that is initialized by default (in if_attach).
We probably have to move percpuq to softc of each driver, but it's
future work. At this point, only wm(4) has percpuq in its softc
as a reference implementation.

Additional information including performance numbers can be found
in the thread at tech-kern@ and tech-net@:
http://mail-index.netbsd.org/tech-kern/2016/01/14/msg019997.html

Acknowledgment: riastradh@ greatly helped this work.
Thank you very much!

Convert sys/dev to use <sys/rndsource.h>.

branches: 1.142.4;
Merge tls-earlyentropy branch into HEAD.

branches: 1.141.2;
Remove unused variable

branches: 1.140.2; 1.140.4;
Fix mii_statchg to take a 'struct ifnet *' instead of device_t. This fixes
problem with a common MDIO bus used for multiple interfaces.
Some drivers converted to CFATTACL_DECL_NEW.

Entropy-pool implementation move and cleanup.

1) Move core entropy-pool code and source/sink/sample management code
to sys/kern from sys/dev.

2) Remove use of NRND as test for presence of entropy-pool code throughout
source tree.

3) Remove use of RND_ENABLED in device drivers as microoptimization to
avoid expensive operations on disabled entropy sources; make the
rnd_add calls do this directly so all callers benefit.

4) Fix bug in recent rnd_add_data()/rnd_add_uint32() changes that might
have lead to slight entropy overestimation for some sources.

5) Add new source types for environmental sensors, power sensors, VM
system events, and skew between clocks, with a sample implementation
for each.

ok releng to go in before the branch due to the difficulty of later
pullup (widespread #ifdef removal and moved files). Tested with release
builds on amd64 and evbarm and live testing on amd64.

branches: 1.138.2; 1.138.6;
Add support for some fxp devices from FreeBSD and OpenBSD.
{Free,Open}BSD say that we have to do some work to make fxp
stable.

IFF_PROMISC implys IFF_ALLMULTI.
Should fix PR#43186.

branches: 1.136.2;
Include sys/proc.h for tsleep, wakeup.

Push the bpf_ops usage back into bpf.h. Push the common ifp->if_bpf
check into the inline functions as well the fourth argument for
bpf_attach.

Check whether the fxp(4) actually attached before calling fxp_stop().

In fxp_detach(), fxp_stop(), first. fxp_stop() stops the callout.
Destroy the callout in fxp_detach().

branches: 1.132.2;
Make fxp at cardbus detach during shutdown.

Stop calling (*cardbus_ctrl) to enable bus mastering, I/O and memory
spaces on the CardBus bridge. cbb(4) always enables that stuff,
anyway. In the process, avoid remembering what BAR we mapped by
writing CARDBUS_{IO,MEM}_ENABLE to sc_cben or sc_cbenable, and
record the BAR in use sc_bar, instead.

Replace more CARDBUS_ constants with PCI_ constants.

Compile-tested, only.

branches: 1.131.2;
Redefine bpf linkage through an always present op vector, i.e.
#if NBPFILTER is no longer required in the client. This change
doesn't yet add support for loading bpf as a module, since drivers
can register before bpf is attached. However, callers of bpf can
now be modularized.

Dynamically loadable bpf could probably be done fairly easily with
coordination from the stub driver and the real driver by registering
attachments in the stub before the real driver is loaded and doing
a handoff. ... and I'm not going to ponder the depths of unload
here.

Tested with i386/MONOLITHIC, modified MONOLITHIC without bpf and rump.

Simplify activation routines: don't call mii_activate(), it's a
no-op. Don't block interrupts, if_deactivate() does it.
Compile-tested, only.

Pull a fix from hme.c rev 1.73 (to #if 0'ed out part):
> Fix a bug in calculation of checksum deduction:
> - To get 16 bit one's complement value from uint32_t variable,
> higher 16 bits should be ignored.
> - RFC 1624 describes methods to recalculate checksum field in headers,
> i.e. one's complement of one's complement sum that could be 0x0000,
> but we don't have to use the strategy to deduct one's complement sum
> itself which won't be zero but should be 0xffff.

Tweak comments and conditionals about EXT_RFA and IPCB.

u_intNN_t -> uintNN_t

Computed checksum value by the FXPF_82559_RXCSUM feature includes data
in IP headers, so we have to deduct not only IP option headers but all
IP headers. But in TCP/UDP layer we can assume the IP header is valid
and a sum of the IP header part should be 0xffff, so we don't have to
bother to deduct it from the computed checksum.

Add TCPv4/UDPv4 RX hardware checksum support for i82559 and later chips
which don't have EXT_RFA and IPCB support. From hme(4) driver and
FreeBSD's fxp(4). Tested on i82559.

XXX: Probably we should have a common function to parse RX packet headers
XXX: to handle a raw checksum value and share it among hme(4) and gem(4) etc.

Use FXPF_EXT_RFA flag instead of FXPF_EXT_TXCB to see IPCB capability
because EXT_RFA for RX cksum is always available with IPCB for TX cksum
but i82558 and i82559 have only EXT_TXCB without IPCB.

Tested on the following fxp cards:
fxp0 at pci0 dev 14 function 0: Intel i82557 Ethernet, rev 2
fxp0 at pci0 dev 14 function 0: i82559 Ethernet, rev 8
fxp0 at pci0 dev 14 function 0: i82550 Ethernet, rev 12

- remove FXPF_IPCB flag. it should always/only be used with the code
conditional on FXPF_EXT_TXCB, so, replace all uses with that
- for the pci frontend, reestablish some flags lost the the prior
changes and simplify one of the cases

this fixes PR 40677 and may fix PR 40431.

branches: 1.122.2;
The PCI revision numbers are unique to a PCI vendor/product
ID pair. Misuse of the revision numbers was causing some of the chip
features to be disabled on some integrated Intel chips. So, move the
determination of the features into the bus frontend, where the
vendor/product ID is known. (Note: sc_rev should be removed. The
microcode patch stuff is also busted and needs to be fixed.) Also,
poll the actual flow control status in inphy, rather than making
assumptions.

contributed anonymously.

Wrap long lines.

Don't pass uint8_t values to le16toh() in fxp_rx_hwcksum().
fxp(4)'s RX hwcksum results weren't used at all on big endian machines.

Checked by i82550 and vmstat -ev on macppc GENERIC kernel with
options INET_CSUM_COUNTERS,TCP_CSUM_COUNTERS,UDP_CSUM_COUNTERS.

Add a missed htole32() on the previous ip4csum-tx bug workaround.

Sort Tx/Rx macro in previous.

Add a workaround for hardware ip4csum-tx bug and enable it.
Confirmed on i82550 rev 12 and UDP fragment packets by ttcp(1).

Call BUS_DMASYNC_PREREAD more strictly on polling DMA descriptors.

branches: 1.115.2; 1.115.4;
Calling fxp_init within the interrupt handler results in
a hang (continuous assertion of FXP_SCB_STATACK_RNR).
Instead do it in the ioctl routine after receiving a
signal from the interrupt handler.

- device/softc split for fxp(4)

branches: 1.113.2; 1.113.4; 1.113.6;
Remove clause 3 and 4 from TNF licenses

branches: 1.112.2; 1.112.4;
use aprint_*_dev and device_xname

branches: 1.111.6;
Start patching up the kernel so that a network driver always has
the opportunity to handle an ioctl before generic ifioctl handling
occurs. This will ease extending the kernel and sharing of code
between drivers.

First steps: Make the signature of ifioctl_common() match struct
ifinet->if_ioctl. Convert SIOCSIFCAP and SIOCSIFMTU to the new
ifioctl() regime, throughout the kernel.

Make many ethernet drivers share the common code for MII media
handling, ether_mediastatus() and ether_mediachange(). Check for
a non-ENXIO error return from mii_mediachg(). (ENXIO indicates
that a PHY is suspended.)

This patch shrinks the source code size by 979 lines. There was
a 5100-byte savings on the NetBSD/i386 kernel configuration, ALL.

I have made a few miscellaneous changes, too:

gem(4): use LIST_EMPTY(), LIST_FOREACH().
mtd(4): handle media ioctls, for a change!
axe(4): do not track link status in sc->axe_link any longer
nfe(4), aue(4), axe(4), udav(4), url(4): do not reset all PHYs
on a change of media

Except for the change to mtd(4), no functional changes are intended.

XXX This patch affects more architectures than I can feasibly
XXX compile and run. I have compiled macppc, sparc64, i386. I
XXX have run the patches on i386 boxen with bnx(4) and sip(4).
XXX Compiling and running on evbmips (MERAKI, ADM5120) is in
XXX progress.

Fix typo in comments.

Fix from YAMAGUCHI Takahiro in PR kern/30023:
Pass correct EEPROM address in fxp_write_eeprom().

Remove the shutdown hook (pmf takes care about it already), as suggested by joerg@

branches: 1.106.2;
Register fxp @ pci to the pmf framework
Kill reference to old powerhook framework.

branches: 1.105.2;
Pull several fixes from OpenBSD's fxp.c:

Rev 1.41:
> use a nop with I bit set at the end of the tx chain. This avoids a race
> between status update and clearing the suspend bit on machines which can't
> write data smaller than 32bits at a time.
This should fix PR port-alpha/30560.

Rev 1.50:
> Fix nasty bug where driver would not correctly catch and handle an rnr
> condition when it was due to the the recieve buffers being exhausted with
> no packet transmits during that time. Symptom was that the fxp would
> simply stop interrupting for the next 15 seconds until the watchdog kicked
> in and reset the chip due to 15 seconds of inactivity, making the fxp very
> poorly behaved when hammered on hard.

Rev 1.61:
> Defer reinitialisation of the RU until after the interrupt handler has had
> a chance to process all pending packets, otherwise the chip may overwrite
> their mbuf clusters after we have freed them.
>
> Eliminates a race that can cause random pool corruption when reconfiguring
> an interface under heavy network load.

And one more change for RX sanity:
- put RU into suspend when the last RFA is processed.

These RNR/RU changes may fix pool corruption problems on fxp.

Tested on AlphaPC164 and i386 with i82559 and i82550.

branches: 1.104.2; 1.104.4;
machine/{bus,cpu,intr}.h -> sys/{bus,cpu,intr}.h

branches: 1.103.2; 1.103.6;
Constify: LLADDR -> CLLADDR. I'm aiming here to make it easier to
identify sockaddr_dl abuse that remains in the kernel, especially
the potential for overwriting memory past the end of a sockaddr_dl
with, e.g., memcpy(LLADDR(), ...).

branches: 1.102.2; 1.102.6;
Merge some of the less invasive changes from the vmlocking branch:

- kthread, callout, devsw API changes
- select()/poll() improvements
- miscellaneous MT safety improvements

branches: 1.101.2; 1.101.4; 1.101.6;
Kill caddr_t; there will be some MI fallout, but it will be fixed shortly.

branches: 1.100.2; 1.100.4; 1.100.8; 1.100.12;
__unused removal on arguments; approved by core.

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

Add "name" parameter to powerhook_establish, to aid debugging. No objections
on tech-kern@

branches: 1.97.14; 1.97.16;
Use device_is_active() rather than testing dv_flags for DVF_ACTIVE
directly.

branches: 1.96.2; 1.96.4; 1.96.6;
Remove leading __ from __(const|inline|signed|volatile) -- it is obsolete.

merge ktrace-lwp.

use VLAN_OUTPUT_TAG()

Make receiver lock-up workaround conditional on setting link1, as we have
too many false positives - should address PR/30505

Make microcode loading work on big endian machines.
Reported and patch supplied by yongari@freebsd a long time ago.

branches: 1.91.2;
- sprinkle const.
- avoid variable shadowing.

split IFCAP_CSUM_xxx to IFCAP_CSUM_xxx_Rx and IFCAP_CSUM_xxx_Tx.

branches: 1.89.10;
Use log() instead of printf().

When adding/deleting multicast addresses, only whack the address
filter if the interface is marked RUNNING.

Fixes kern/27678.

Fix a typo in a comment.

Patch from HITOSHI Osada:
* Add PAUSE-related event counters.
* Return flow control bits in fxp_mii_mediastatus().

Patch from HITOSHI Osada:
* Newer chips do not need the receiver lock-up workaround; detect when it
is required.

When we are recovering from a resource exhaustion on receive, make sure
that the RU is stopped before issuing an RU_START.
Fix suggested by Takahiro Kambe in PR kern/10622.

Back out revision 1.82. It seems to cause a number of device timeouts
on x86 boxes. There must be a better way to deal with this that works
well for everyone. In the meantime, back out to the version that works
for more people.

Be more careful about issuing CU_RESUME in fxp_start()--only doing
it if we think it's probably necessary. Then do it again in the
tx interrupt handler, if we again think it's necessary. This
reduces the number of commands we issue the chip. Prior to this
change, the i82550 (running without extended feature set, so like
a '559) would sometimes fail in fxp_scb_wait() prior to issuing
another CU_RESUME, resulting in
fxp0: WARNING: SCB timed out!
messages on the console, often followed by device timeouts.

branches: 1.81.2;
support h/w assisted vlan tag insertion.

support h/w assisted vlan tag removal.

On some 'fast' i386 motherboards, the timing between consecutive reads
and writes of the EEPROM are too fast so invalid data is returned. So we
increase the DELAY()s. The right thing would be to check a ready bit
on the E^2 if such a thing exists. (checked by potr)

- disable ipv4 checksum offloading as it seems to have a problem.
- add comments to describe it.

Fix some typos. From Tom Cosgrove via jmc@openbsd.

In fxp_intr(), bail early without touching any registers if sc_enabled is
zero, otherwise we could cause PCI master aborts if the card is in D3 power
state and we happen to share a PCI interrupt line with some other device.

branches: 1.75.2;
ip/tcp/udp checksum offloading support for i82550.
largely from FreeBSD.

check and MCLAIM correct mbuf in fxp_copy_small path.

Add MBUFTRACE kernel option.
Do a little mbuf rework while here. Change all uses of MGET*(*, M_WAIT, *)
to m_get*(M_WAIT, *). These are not performance critical and making them
call m_get saves considerable space. Add m_clget analogue of MCLGET and
make corresponding change for M_WAIT uses.
Modify netinet, gem, fxp, tulip, nfs to support MBUFTRACE.
Begin to change netstat to use sysctl.

Use bus_dmamap_load_mbuf() in fxp_add_rfabuf().

Use aprint_*().

synchronous, not syncronous.

Cosmetic changes.

Fix signed/unsigned comparison warnings.

Change the "dontcare bits" argument to ifmedia_init() to IFM_IMASK
so that it is possible select PHY instances other than the first
one (instance zero), if there is more than one PHY attached.

Add newlines to the WARNING messages printed for the dynamic standby bug.

branches: 1.65.2;
The FIX_RESUME_BUG hack does not work, so remove it.
Also, attempt to clarify the message slightly when updating the EEPROM.

branches: 1.64.2;
Add support for the "CPUSaver" receive interrupt mitigating microcode
on the following PRO/100 chips:

* i82558 step A4
* i82558 step B0
* i82559 step A0
* i82559S step A
* i82550
* i82550 step C

The interrupt delay is configurable on all microcodable chips. The
maximum "bundle" size (packet count) is configurable on all but the
i82558.

The microcode is enabled by setting IFF_LINK0 on the interface.

Derived from code in FreeBSD.

Address Intel 82801BA/82801BAM Specification Update, Errata #30:

The ICH on-board Ethernet and some i82559 chips have a bug which
will cause a PCI protocol violation if the chip receives a CU_RESUME
command as it is entering the IDLE state by deasserting #CLKRUN.
(This is the so-called "resume bug" that we previously had an incomplete
work-around for on ICH chipsets.)

The work-around is to disable Dynamic Standby Mode, such that the
chip will never deasert #CLKRUN. Dynamic Standby Mode is disabled
by clearing a bit in the EEPROM and updating the EEPROM (and EEPROM
checksum).

Unfortunately, the chip will only consult the EEPROM setting after
a PCI bus reset, so a system reboot is required once the EEPROM
has been updated (the EEPROM update only needs to happen once,
and the driver usses a warning instructing the user to reboot the
system once the work-around has been applied).

Issue pointed out by David Brownlee, and code more-or-less lifted
from FreeBSD.

Move the code which shifts bits into the EEPROM into its own function.

add/cleanup RCSID

Spell 'occurred' with two 'r's.

branches: 1.59.2;
Advertize pause capability (802.3x flow control) to peer.

Use BUS_DMA_READ and BUS_DMA_WRITE in some obvious places.

branches: 1.57.2;
bzero -> memset

bcopy -> strcpy

Split the Tx and Rx interrupt routines into separate functions,
and add some (optional) event counters.

Don't need INET or NS includes here.

Define some extra configuration block bits used on the i82550 (a.k.a.
PRO 100/S), including the extended RFA format. Don't hard-code the
RFA size to allow us to use the extended format in the future.

Some changes from the new FreeBSD `fxp' driver:
- Add some additional config block bits for the i82558/i82559.
- Change the config block template to only fill in the must-be-one
reserved bits, leave fxp_init() to fill in all the important things.
- On the i82558/i82559, we can use "Receive Long Frames" rather than
"Save Bad Frames" to support the VLAN MTU.
- Use 802.3x flow control on the i82558/i82559. This is all handled
transparently by the hardware. When in promiscuous mode, allow
wire-watchers to see the flow control frames.
- Use the Extended TxCB format on the i82558/i82559. This places two
Transmit Buffer Descriptors directly in the TxCB, which should cover
the vast majority of packets transmitted. This saves PCI transactions
that would otherwise be required to fetch the TBD list.

With the tansmit queue length changes from yesterday, this gets us up to
11.5MB/s TCP transmit speed, out of an absolute maximum of 12MB/s possible
on a 100Mb/s link.

A little slight cleanup.

Re-arrange the transmit control data somewhat so that we can
use the extended TxCB format (that change will happen later).

Sigh, typo in last.

Fix non-fatal typo -- CB_NOP -> CU_NOP in the ICH work-around (they
have the same value).

The 82801BA built-in Ethernet has a bug which requires us to
issue a NOP before a CU_RESUME when in 10baseT mode. Handle
this.

ANSI'ify.

delint: remove volatile from fxp_mdi_read() decl

branches: 1.44.2;
Fix a problem with the ALTQ changes that can cause bogus memory
refernces. Problem reported by Luke Mewburn.

ALTQ'ify.

Add new powerhook argument values, PWR_SOFTSUSPEND, PWR_SOFTSTANDBY and
PWR_SOFTRESUME. Apm calls powerhook with the values in normal interrupt
priority level while others are protected with splhigh().

Move bpfattach()/bpfdetach() calls into ether_ifattach()/ether_ifdetach().

Use ether_ioctl().

Support ETHERCAP_VLAN_MTU by doing extra error checking upon packet
reception and saving other "bad frames" (i.e. ones that are larger
than the standard Ethernet frame length) if we have VLANs configured
on the interface.

Move the check for "promisc + unicast + not for us" into ether_input(),
and change Ethernet drivers to always pass all received frames to
ether_input() (with a few exceptions, which are documented in the
code).

6 -> ETHER_ADDR_LEN

Since the RFA is allocated in a normal mbuf, as opposed to DMA-safe
memory that is explicitly mapped in a DMA-coherent manner, we must
make sure to PREREAD sync the RFA after noticing a clear "complete"
bit. Without this, the clear bit will linger in the cache, and the
CPU will not notice when the chip updates the bit via DMA later.

From Izumi Tsutsui on port-arm32@netbsd.org.

remove include of <vm/vm.h>. <vm/vm.h> -> <uvm/uvm_extern.h>

branches: 1.34.2;
For all network drivers that call ether_ifattach(), and also
have _detach() functions:
Ensure that softc keeps state about whether the attach succeeded,
and make the detach function return immediately if the attach did
not complete.

branches: 1.33.2;
Increase delay on EEPROM access.
CATS sometimes cannot read MAC address correctly.

Call bus_dmamap_sync() properly before/after reading fcd_stats
in fxp_control_data.

Make two previous changes more correct, as shown by Izumi Tsutsui.

The EEPROM is little-endian.

Fix DMA status reading on big-endian systems. From Sanjay Lal on port-macppc.

And a newline after error string.

Calibrate the timeouts from rev 1.26 by using DELAY(1) so their
values are less machine-dependant. Fixes port-i386/10141, where
spurrious timeouts were being seen.
Also note the line numbers of the timeouts so it can be determined
which is being seen (via __LINE__).

Catch some potentially infinite loops in while() loops if the chip happens
to "go insane" and never complete some operations (dmasync, etc.).

Delete redundant decls of fxp_enable(), fxp_disable() - they're in
i82557var.h.

New callout mechanism with two major improvements over the old
timeout()/untimeout() API:
- Clients supply callout handle storage, thus eliminating problems of
resource allocation.
- Insertion and removal of callouts is constant time, important as
this facility is used quite a lot in the kernel.

The old timeout()/untimeout() API has been removed from the kernel.

Get rid of the powerhook when we detach.

No longer necessary to futz with ifp->if_baudrate here.

(fxp_init): don't clear the FXPF_MII flag as this results in a panic
when detaching

Add activate routine and check device active flag also in one second
tick handler.

Cosmetic changes.

add detach code

Don't dry to diving MIIF_NOISOLATE in the PHY drivers. Instead, pass
flags down from the parent to child vi mii_attach().

Bring some order to the chaos which was the MII code function naming
"conventions".

Take a stab at making this work on big-endian systems.

change comment to mention that this driver also handles the '558 and '559

try to do a better job of figuring out the EEPROM size; the old code
didn't work on (some?) 557-based cards

Call mii_down() as appropriate.

Adapt to mii_phy_probe() change.

Add enable/disable support and EEPROM size selection, prerequisites
for getting the cardbus fxp attachment to work.

Tested on the built-in 82559 in a VAIO Z505S.

From Johan Danielsson / PR8631; changed slightly to KNF.

Fix suspend/resume-related problems observed on fxp0 on Sony Z505S:

Symptoms: system would crash with "data modified on free list" pool
panic from the mbuf cluster pool shortly after a resume. The cluster
in question contained a valid 82557 receive descriptor and an IP
packet. Happened sporadically in normal use. Easiest way for me to
reproduce it was to run tcpdump and a flood ping and do a
suspend/resume cycle or two.

Changes:
- in interrupt handler, if the interface isn't in IFF_RUNNING state,
just ack interrupts and return; don't try to receive packets, queue
new descriptors, etc., etc.,
- add power control hook to take interface down on suspend,
and restart it (if it was up) on resume.
- tweaks to fxp_stop and fxp_shutdown to avoid recursive panics due
to the (now fixed) bug.

branches: 1.8.2; 1.8.4; 1.8.6;
Since we have to go through fxp_init() to properly handle IFF_ALLMULTI
anyway, take advantage of this and greatly simplify the programming
of the multicast filter. This solves the last reported "device timeout"
problem with this driver.

Almost completely rewrite the receive logic, making it as close as possible
to the EPIC/100 driver's (adjusting for the fact that Intel Ethernet chips
are from Pluto):
* Don't allocate receive buffers until the interface is actually brought
up, and release all of them if the interface is taken down.
* Add a knob (defaults to off) which will copy an incoming packet to
a single header mbuf if it is small enough to fit in one, rather than
burning an entire cluster on it. Note that this change will be mostly
moot if/when sbcompress() is changed to handle compressing clusters.

Simplify some of the receive list logic:
* Rather than using a homegrown queue and additional software RX descriptors,
use an ifqueue to queue receive buffers, and M_{GET,SET}CTX() to hook DMA
maps and receive buffers together.

Clean up a bit:
* Macroize a bunch of things to make the code a bit easier to follow.

Only tick the MII if we are using it.

Be a little less selective in the transmit interrupt path.

Use the Interrupt bit in the command block to generate interrupts, rather
than interrupting after every Command Unit Not Active event.

Don't bother with interrupt-driven multicast setup.

Completely rewrite the transmit logic, making it look more like the
EPIC/100 driver's. Also, fix the "all multicast" logic. Also do some
general cleanup.

branches: 1.1.2;
Bus-independent back-end driver for Intel i82557 fast Ethernet chips.

Pass proper status values to rnd_add_uint32(9) as rnd(9) man page claims.

Use callout_setfunc() / callout_schedule().

Use ifmedia_fini().

Adopt <net/if_stats.h>.

branches: 1.156.2;
if_percpuq(9) automatically increments if_ipackets, so don't add number of
RX frames from device's statistics counter to if_ipackets to avoid double
count.

Fix direction of the loop.

Found by the lgtm bot.

branches: 1.154.2;
Don't automatically set ec_capenable's ETHERCAP_VLAN_HWTAGGING bit in
vlan_config() to make it user-controllable. Instead, set the bit in
xxx_attach().

Use ETHER_LOCK()/ETHER_UNLOCK() for all ethernet drivers to protect ec_multi*.

No functional change:
- Simplify MII structure initialization and reference.
- u_int*_t -> uint*_t.
- KNF

In drivers which use MII(4) and have hook SIOC[GS]IFMEDIA which just pass to
ifmedia_ioctl(), the hook is not required because ether_ioctl has it
(if_ethersubr.c rev. 1.160). These drivers don't return ENETRESET in
ifmedia_ioctl(), so no functional change.

Change MII PHY read/write API from:

int (*mii_readreg_t)(device_t, int, int);
void (*mii_writereg_t)(device_t, int, int, int);
to:

int (*mii_readreg_t)(device_t, int, int, uint16_t *);
int (*mii_writereg_t)(device_t, int, int, uint16_t);

Now we can test if a read/write operation failed or not by the return value.

In 802.3 spec says that the PHY shall not respond to read/write transaction
to the unimplemented register(22.2.4.3). Detecting timeout can be used to
check whether a register is implemented or not (if the register conforms to
the spec). ukphy(4) can be used this for MII_MMDACR and MII_MMDAADR.

Note that I noticed that the following code do infinite loop in the
read/wirte function. If it accesses unimplemented PHY register, it will hang.
It should be fixed:

arm/at91/at91emac.c
arm/ep93xx/epe.c
arm/omap/omapl1x_emac.c
mips/ralink/ralink_eth.c
arch/powerpc/booke/dev/pq3etsec.c(read)
dev/cadence/if_cemac.c <- hkenken
dev/ic/lan9118.c


Tested with the following device:

axe+ukphy
axe+rgephy
axen+rgephy (tested by Andrius V)
wm+atphy
wm+ukphy
wm+igphy
wm+ihphy
wm+makphy
sk+makphy
sk+brgphy
sk+gentbi
msk+makphy
sip+icsphy
sip+ukphy
re+rgephy
bge+brgphy
bnx+brgphy
gsip+gphyter
rtk+rlphy
fxp+inphy (tested by Andrius V)
tlp+acphy
ex+exphy
epic+qsphy
vge+ciphy (tested by Andrius V)
vr+ukphy (tested by Andrius V)
vte+ukphy (tested by Andrius V)

Not tested (MAC):
arm:at91emac
arm:cemac
arm:epe
arm:geminigmac
arm:enet
arm:cpsw
arm:emac(omac)
arm:emac(sunxi)
arm:npe
evbppc:temac
macppc:bm
macppc:gm
mips:aumac
mips:ae
mips:cnmac
mips:reth
mips:sbmac
playstation2:smap
powerpc:tsec
powerpc:emac(ibm4xx)
sgimips:mec
sparc:be
sf
ne(ax88190, dl10019)
awge
ep
gem
hme
smsh
mtd
sm
age
alc
ale
bce
cas
et
jme
lii
nfe
pcn
ste
stge
tl
xi
aue
mue
smsc
udav
url

Not tested (PHY):
amhphy
bmtphy
dmphy
etphy
glxtphy
ikphy
iophy
lxtphy
nsphyter
pnaphy
rdcphy
sqphy
tlphy
tqphy
urlphy

branches: 1.149.2;
Implement the BPF direction filter (BIOC[GS]DIRECTION). It provides backward
compatibility with BIOC[GS]SEESENT ioctl. The userland interface is the same
as FreeBSD.

This change also fixes a bug that the direction is misunderstand on some
environment by passing the direction to bpf_mtap*() instead of checking
m->m_pkthdr.rcvif.

branches: 1.148.2;
VLAN ID uses pkthdr instead of mtag now. Contributed by s-yamaguchi@IIJ.

I just commit by proxy. Reviewed by joerg@n.o and christos@n.o, thanks.
See http://mail-index.netbsd.org/tech-net/2017/09/26/msg006459.html

XXX need pullup to -8 branch

branches: 1.147.6;
Apply deferred if_start to more drivers

branches: 1.146.2;
Move bpf_mtap and if_ipackets++ on Rx of each driver to percpuq if_input

The benefits of the change are:
- We can reduce codes
- We can provide the same behavior between drivers
- Where/When if_ipackets is counted up
- Note that some drivers still update packet statistics in their own
way (periodical update)
- Moved bpf_mtap run in softint
- This makes it easy to MP-ify bpf

Proposed on tech-kern and tech-net

branches: 1.145.2;
Introduce m_set_rcvif and m_reset_rcvif

The API is used to set (or reset) a received interface of a mbuf.
They are counterpart of m_get_rcvif, which will come in another
commit, hide internal of rcvif operation, and reduce the diff of
the upcoming change.

No functional change.

Introduce softint-based if_input

This change intends to run the whole network stack in softint context
(or normal LWP), not hardware interrupt context. Note that the work is
still incomplete by this change; to that end, we also have to softint-ify
if_link_state_change (and bpf) which can still run in hardware interrupt.

This change softint-ifies at ifp->if_input that is called from
each device driver (and ieee80211_input) to ensure Layer 2 runs
in softint (e.g., ether_input and bridge_input). To this end,
we provide a framework (called percpuq) that utlizes softint(9)
and percpu ifqueues. With this patch, rxintr of most drivers just
queues received packets and schedules a softint, and the softint
dequeues packets and does rest packet processing.

To minimize changes to each driver, percpuq is allocated in struct
ifnet for now and that is initialized by default (in if_attach).
We probably have to move percpuq to softc of each driver, but it's
future work. At this point, only wm(4) has percpuq in its softc
as a reference implementation.

Additional information including performance numbers can be found
in the thread at tech-kern@ and tech-net@:
http://mail-index.netbsd.org/tech-kern/2016/01/14/msg019997.html

Acknowledgment: riastradh@ greatly helped this work.
Thank you very much!

Convert sys/dev to use <sys/rndsource.h>.

branches: 1.142.4;
Merge tls-earlyentropy branch into HEAD.

branches: 1.141.2;
Remove unused variable

branches: 1.140.2; 1.140.4;
Fix mii_statchg to take a 'struct ifnet *' instead of device_t. This fixes
problem with a common MDIO bus used for multiple interfaces.
Some drivers converted to CFATTACL_DECL_NEW.

Entropy-pool implementation move and cleanup.

1) Move core entropy-pool code and source/sink/sample management code
to sys/kern from sys/dev.

2) Remove use of NRND as test for presence of entropy-pool code throughout
source tree.

3) Remove use of RND_ENABLED in device drivers as microoptimization to
avoid expensive operations on disabled entropy sources; make the
rnd_add calls do this directly so all callers benefit.

4) Fix bug in recent rnd_add_data()/rnd_add_uint32() changes that might
have lead to slight entropy overestimation for some sources.

5) Add new source types for environmental sensors, power sensors, VM
system events, and skew between clocks, with a sample implementation
for each.

ok releng to go in before the branch due to the difficulty of later
pullup (widespread #ifdef removal and moved files). Tested with release
builds on amd64 and evbarm and live testing on amd64.

branches: 1.138.2; 1.138.6;
Add support for some fxp devices from FreeBSD and OpenBSD.
{Free,Open}BSD say that we have to do some work to make fxp
stable.

IFF_PROMISC implys IFF_ALLMULTI.
Should fix PR#43186.

branches: 1.136.2;
Include sys/proc.h for tsleep, wakeup.

Push the bpf_ops usage back into bpf.h. Push the common ifp->if_bpf
check into the inline functions as well the fourth argument for
bpf_attach.

Check whether the fxp(4) actually attached before calling fxp_stop().

In fxp_detach(), fxp_stop(), first. fxp_stop() stops the callout.
Destroy the callout in fxp_detach().

branches: 1.132.2;
Make fxp at cardbus detach during shutdown.

Stop calling (*cardbus_ctrl) to enable bus mastering, I/O and memory
spaces on the CardBus bridge. cbb(4) always enables that stuff,
anyway. In the process, avoid remembering what BAR we mapped by
writing CARDBUS_{IO,MEM}_ENABLE to sc_cben or sc_cbenable, and
record the BAR in use sc_bar, instead.

Replace more CARDBUS_ constants with PCI_ constants.

Compile-tested, only.

branches: 1.131.2;
Redefine bpf linkage through an always present op vector, i.e.
#if NBPFILTER is no longer required in the client. This change
doesn't yet add support for loading bpf as a module, since drivers
can register before bpf is attached. However, callers of bpf can
now be modularized.

Dynamically loadable bpf could probably be done fairly easily with
coordination from the stub driver and the real driver by registering
attachments in the stub before the real driver is loaded and doing
a handoff. ... and I'm not going to ponder the depths of unload
here.

Tested with i386/MONOLITHIC, modified MONOLITHIC without bpf and rump.

Simplify activation routines: don't call mii_activate(), it's a
no-op. Don't block interrupts, if_deactivate() does it.
Compile-tested, only.

Pull a fix from hme.c rev 1.73 (to #if 0'ed out part):
> Fix a bug in calculation of checksum deduction:
> - To get 16 bit one's complement value from uint32_t variable,
> higher 16 bits should be ignored.
> - RFC 1624 describes methods to recalculate checksum field in headers,
> i.e. one's complement of one's complement sum that could be 0x0000,
> but we don't have to use the strategy to deduct one's complement sum
> itself which won't be zero but should be 0xffff.

Tweak comments and conditionals about EXT_RFA and IPCB.

u_intNN_t -> uintNN_t

Computed checksum value by the FXPF_82559_RXCSUM feature includes data
in IP headers, so we have to deduct not only IP option headers but all
IP headers. But in TCP/UDP layer we can assume the IP header is valid
and a sum of the IP header part should be 0xffff, so we don't have to
bother to deduct it from the computed checksum.

Add TCPv4/UDPv4 RX hardware checksum support for i82559 and later chips
which don't have EXT_RFA and IPCB support. From hme(4) driver and
FreeBSD's fxp(4). Tested on i82559.

XXX: Probably we should have a common function to parse RX packet headers
XXX: to handle a raw checksum value and share it among hme(4) and gem(4) etc.

Use FXPF_EXT_RFA flag instead of FXPF_EXT_TXCB to see IPCB capability
because EXT_RFA for RX cksum is always available with IPCB for TX cksum
but i82558 and i82559 have only EXT_TXCB without IPCB.

Tested on the following fxp cards:
fxp0 at pci0 dev 14 function 0: Intel i82557 Ethernet, rev 2
fxp0 at pci0 dev 14 function 0: i82559 Ethernet, rev 8
fxp0 at pci0 dev 14 function 0: i82550 Ethernet, rev 12

- remove FXPF_IPCB flag. it should always/only be used with the code
conditional on FXPF_EXT_TXCB, so, replace all uses with that
- for the pci frontend, reestablish some flags lost the the prior
changes and simplify one of the cases

this fixes PR 40677 and may fix PR 40431.

branches: 1.122.2;
The PCI revision numbers are unique to a PCI vendor/product
ID pair. Misuse of the revision numbers was causing some of the chip
features to be disabled on some integrated Intel chips. So, move the
determination of the features into the bus frontend, where the
vendor/product ID is known. (Note: sc_rev should be removed. The
microcode patch stuff is also busted and needs to be fixed.) Also,
poll the actual flow control status in inphy, rather than making
assumptions.

contributed anonymously.

Wrap long lines.

Don't pass uint8_t values to le16toh() in fxp_rx_hwcksum().
fxp(4)'s RX hwcksum results weren't used at all on big endian machines.

Checked by i82550 and vmstat -ev on macppc GENERIC kernel with
options INET_CSUM_COUNTERS,TCP_CSUM_COUNTERS,UDP_CSUM_COUNTERS.

Add a missed htole32() on the previous ip4csum-tx bug workaround.

Sort Tx/Rx macro in previous.

Add a workaround for hardware ip4csum-tx bug and enable it.
Confirmed on i82550 rev 12 and UDP fragment packets by ttcp(1).

Call BUS_DMASYNC_PREREAD more strictly on polling DMA descriptors.

branches: 1.115.2; 1.115.4;
Calling fxp_init within the interrupt handler results in
a hang (continuous assertion of FXP_SCB_STATACK_RNR).
Instead do it in the ioctl routine after receiving a
signal from the interrupt handler.

- device/softc split for fxp(4)

branches: 1.113.2; 1.113.4; 1.113.6;
Remove clause 3 and 4 from TNF licenses

branches: 1.112.2; 1.112.4;
use aprint_*_dev and device_xname

branches: 1.111.6;
Start patching up the kernel so that a network driver always has
the opportunity to handle an ioctl before generic ifioctl handling
occurs. This will ease extending the kernel and sharing of code
between drivers.

First steps: Make the signature of ifioctl_common() match struct
ifinet->if_ioctl. Convert SIOCSIFCAP and SIOCSIFMTU to the new
ifioctl() regime, throughout the kernel.

Make many ethernet drivers share the common code for MII media
handling, ether_mediastatus() and ether_mediachange(). Check for
a non-ENXIO error return from mii_mediachg(). (ENXIO indicates
that a PHY is suspended.)

This patch shrinks the source code size by 979 lines. There was
a 5100-byte savings on the NetBSD/i386 kernel configuration, ALL.

I have made a few miscellaneous changes, too:

gem(4): use LIST_EMPTY(), LIST_FOREACH().
mtd(4): handle media ioctls, for a change!
axe(4): do not track link status in sc->axe_link any longer
nfe(4), aue(4), axe(4), udav(4), url(4): do not reset all PHYs
on a change of media

Except for the change to mtd(4), no functional changes are intended.

XXX This patch affects more architectures than I can feasibly
XXX compile and run. I have compiled macppc, sparc64, i386. I
XXX have run the patches on i386 boxen with bnx(4) and sip(4).
XXX Compiling and running on evbmips (MERAKI, ADM5120) is in
XXX progress.

Fix typo in comments.

Fix from YAMAGUCHI Takahiro in PR kern/30023:
Pass correct EEPROM address in fxp_write_eeprom().

Remove the shutdown hook (pmf takes care about it already), as suggested by joerg@

branches: 1.106.2;
Register fxp @ pci to the pmf framework
Kill reference to old powerhook framework.

branches: 1.105.2;
Pull several fixes from OpenBSD's fxp.c:

Rev 1.41:
> use a nop with I bit set at the end of the tx chain. This avoids a race
> between status update and clearing the suspend bit on machines which can't
> write data smaller than 32bits at a time.
This should fix PR port-alpha/30560.

Rev 1.50:
> Fix nasty bug where driver would not correctly catch and handle an rnr
> condition when it was due to the the recieve buffers being exhausted with
> no packet transmits during that time. Symptom was that the fxp would
> simply stop interrupting for the next 15 seconds until the watchdog kicked
> in and reset the chip due to 15 seconds of inactivity, making the fxp very
> poorly behaved when hammered on hard.

Rev 1.61:
> Defer reinitialisation of the RU until after the interrupt handler has had
> a chance to process all pending packets, otherwise the chip may overwrite
> their mbuf clusters after we have freed them.
>
> Eliminates a race that can cause random pool corruption when reconfiguring
> an interface under heavy network load.

And one more change for RX sanity:
- put RU into suspend when the last RFA is processed.

These RNR/RU changes may fix pool corruption problems on fxp.

Tested on AlphaPC164 and i386 with i82559 and i82550.

branches: 1.104.2; 1.104.4;
machine/{bus,cpu,intr}.h -> sys/{bus,cpu,intr}.h

branches: 1.103.2; 1.103.6;
Constify: LLADDR -> CLLADDR. I'm aiming here to make it easier to
identify sockaddr_dl abuse that remains in the kernel, especially
the potential for overwriting memory past the end of a sockaddr_dl
with, e.g., memcpy(LLADDR(), ...).

branches: 1.102.2; 1.102.6;
Merge some of the less invasive changes from the vmlocking branch:

- kthread, callout, devsw API changes
- select()/poll() improvements
- miscellaneous MT safety improvements