Fix DMA sync sizes.
Don't panic for inconsistent queue counter, just print an error to console.

s/ for for / for / in comment.

iwm(4): Disentangle attach.

Don't attach a half-baked interface and then detach it and then
reattach it after mountroot when we can read firmware; just defer
attaching the interface altogether until mountroot.

Likely fixes some panics I've seen every now and then at boot with
iwm(4).

if_attach and if_initialize cannot fail, don't test return value

These were originally made failable back in 2017 when if_initialize
allocated a softint in every interface for link state changes, so
that it could fail gracefully instead of panicking:

https://mail-index.NetBSD.org/source-changes/2017/10/23/msg089053.html

However, this spawned many seldom- or never-tested error branches,
which are risky to have around. And that softint in every interface
has since been replaced by a single global workqueue, because link
state changes require thread context but not low latency or high
throughput:

https://mail-index.NetBSD.org/source-changes/2020/02/06/msg113759.html

So there is no longer any reason for if_initialize to fail. (The
subroutine if_stats_init can't fail because percpu_alloc can't fail
either.)

There is a snag: the softint_establish in if_percpuq_create could
fail, potentially leading to bad consequences later on trying to use
the softint. This change doesn't introduce any new bugs because of
the snag -- if_percpuq_attach was already broken. However, the snag
can be better addressed without spawning error branches, either by
using a single softint or making softints less scarce.

(Separate commit will change the signatures of if_attach and
if_initialize to return void, scheduled to ride whatever is the next
convenient kernel bump.)

Patch and testing on amd64 and evbmips64-eb by maya@; commit message
soliloquy, and compile-testing on evbppc/i386/earmv7hf, by me.

branches: 1.85.6;
This driver sleeps during iwm_media_change(), and thus requires an
adaptive mutex for the media lock.

Adopt <net/if_stats.h>.

branches: 1.83.2; 1.83.10;
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.

iwm(4) workaround as ifp->if_hwdl can be set only once. Advised nonaka@n.o and ozaki-r@n.o, thanks.

To avoid for iwm(4) to set if_hwdl before loading firmware, use local address
as dummy. See sys/net/if.c::if_set_sadl().

Fix "ifconfig iwm0" failures reported by ryoon@n.o. Advised nonaka@n.o and ozaki-r@n.o, thanks.

Remove duplicate ;

Fix panic on boot with iwm(4). Advised by nonaka@n.o, thanks.

XXX pullup-8

branches: 1.78.2;
CID-1427736: Appease coverity using KASSERT

Match Intel Dual Band Wireless AC 3168. The code already supports it.

branches: 1.76.2;
- If if_initialize() failed in the attach function, free resources and return.
- Fix error path in the attach function correctly.

branches: 1.75.2;
kmem_intr_free kmem_intr_[z]alloced memory

the underlying pools are the same but api-wise those should match

fix typo.

reported by ryo@.

branches: 1.73.2;
remove checks for failure after memory allocation calls that cannot fail:

kmem_alloc() with KM_SLEEP
kmem_zalloc() with KM_SLEEP
percpu_alloc()
pserialize_create()
psref_class_create()

all of these paths include an assertion that the allocation has not failed,
so callers should not assert that again.

iwm(4): match Wireless 8265 device.

iwm(4): update firmware to maintained core release.

firmware from https://git.kernel.org/pub/scm/linux/kernel/git/iwlwifi/linux-firmware.git/tree/?id=4b9559fb0565f582a21a4070a81a3b6a7d77610b

wlan interfaces make interrupt routine running on softint context.

see http://mail-index.netbsd.org/tech-kern/2016/12/06/msg021281.html

tested device:
* ath at pci: AR5212, AR5424
* athn at pci: AR9287
* ipw at pci: 2100BG
* iwi at pci: 2915ABG
* iwm at pci: 3165, 7260, 8260
* iwn at pci: 4945, 6235
* ral at pci: RT2560
* rtwn at pci: RTL8192CE

PR/51797: initialize hardware completely when receiving a deauth frame from AP.

avoid "iwm0: fatal firmware error".

more block network interrupt.

increase command response timeout value.

include <sys/intr.h> for softint(9).

newstate should be blocked network interrupt, because it called iwm_start.

show NVM channel flags when AB_DEBUG is set.

more acknowledge interrupts from iwn(4).

only bus_dmamap_sync(9) once for ICT region.

branches: 1.61.2;
iwm(4): 3165 works fine with iwlwifi-7265D-17.ucode firmware.

firmware from http://git.kernel.org/cgit/linux/kernel/git/iwlwifi/linux-firmware.git/tree/?id=be8a6fcba2e48e00674d1f35e6def56e2268039c

check LAR support in NVM for 8260/4165.

Add apmg_wake_up_wa support from DragonFly BSD.

iwm(4): 8260 works now. 4165 may works too.

set NULL after free.

Apply OpenBSD if_iwm.c rev.1.148.

> While setting up the basic rate bitmask for iwm's firmware, if the AP does
> not specify basic rates for either the CCK or OFDM set, add just the most
> basic rate to that set (1 Mbit/s in case of CCK, 6 Mbit/s in case of OFDM).
> This behaviour matches what code comments seem to imply.

> The previous code would add all possible basic rates in such cases.
> So if all basic rates were CCK only, the code would add all possible OFDM basic
> rates on top. Then the firmware would send some frames at too high rates, e.g.
> RTS frames would be sent at 24Mbit/s which is a bit risky on noisy channels.

Do not use the "flags" field uninitialized

A backport if iwm(4) (to netbsd-7) does not build without this change.

LGTM nonaka@

Added enum iwm_tx_pm_timeouts from FreeBSD.

cosmetic.

use correct header size.

fix indent

iwm(4): make interrupt routine running on softint context.

see http://mail-index.netbsd.org/tech-kern/2016/12/06/msg021281.html

use pci_intr_establish_xname(9).

twiddle bpf related.

cosmetic

Export some 802.11 IE manipulate functions.

iwm(4): Sync with OpenBSD.

7265 works fine.
7260 and 3160 maybe work.
7265D, 3165 and 8260 not work yet.

if_iwm.c: r1.147.
if_iwmreg.h: r1.19 without r1.16.
if_iwmvar.h: r1.24.

firmware is taken from https://wireless.wiki.kernel.org/en/users/drivers/iwlwifi
- iwlwifi-7260-ucode-16.242414.0.tgz
- iwlwifi-3160-ucode-16.242414.0.tgz
- iwlwifi-7265-ucode-16.242414.0.tgz
- iwlwifi-8000-ucode-16.242414.0.tgz

Apply deferred if_start framework

if_schedule_deferred_start checks if the if_snd queue contains packets,
so drivers don't need to check it by themselves.

Fix inverted logic in iwm(4).

The PROT_REQUIRE flag in should be set for data frames above a certain
length, but we were setting it for !data frames above a certain length,
which makes no sense at all.

- taken from all other BSDs (OpenBSD, FreeBSD, DragonflyBSD...)

branches: 1.42.2;
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.

Use M_GETCTX

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!

Always use pci_intr_alloc(9)/pci_intr_release(9).

Remove leading whitespaces and kind

Use pci_intr_alloc().

iwm(4) can read the firmware image from disk when interface up.
This change is useful for full-disk encryption environment.

Limit the number of dma segments used for transmitting packets to
IWM_NUM_OF_TBS - 2. We have IWM_NUM_OF_TBS slots, but use two of those
for sending commands to the firmware.

From OpenBSD rev. 1.41.

should be able to use 5GHz wireless network.

if_iwm use unified establish API.

Added control debugging output via sysctl.

Use MSI instead of legacy INTx, if available.

should fix PR/49663.

Register iwm(4) with pmf(9).

Patch from Pierre Pronchery.

branches: 1.29.2;
Not all iwm hardware supports the 5GHz band. Make setting 11a rates
and scanning conditional on the 5GHz support bit in the nvm.

From OpenBSD rev. 1.39. Arranged by me.

Fix radiotap and fixed rates. Radiotap was reporting the wrong rate for
data frames and fixed rates weren't really fixed and were converted into
the wrong hardware rate.

From OpenBSD rev.1.36.

Avoid kmem_free(NULL) in iwm_read_firmware()

This code path can be hit if the firmware failed to load, for instance
if the file is not present on the filesystem. In this case
firmware_open() fails, and fw->fw_rawdata never gets allocated in the
first place.

stop the calibration timeout when stopping the interface.

From OpenBSD rev.1.35.

Set/clear the IWM_MAC_FILTER_IN_BEACON flag under conditions that match what
Linux does.

From OpenBSD rev.1.34.

Fix 11a support.

From OpenBSD rev.1.33.

Make setting the MAC context bit for 11g protection when transmitting
OFDM frames conditional on the node via IEEE80211_F_USEPROT.

From OpenBSD rev.1.31.

Don't set CTS to self. Linux stopped setting this back in
dc271ee0d04d12d6bfabacbec803289a7072fbd9 as it is known
to cause problems.

From OpenBSD rev.1.30.

sc_fixed_ridx is initialised to 0 but tested as it it were initialised
to -1. The result of this is tx frames were always sent out at fixed
rate 0 instead of ni_txrate.

Match the iwn behaviour and test ic_fixed_rate for -1 instead.

From OpenBSD rev.1.29.

Correctly set lowest_present_ofdm in iwm_mvm_ack_rates().
From OpenBSD rev.1.28.

init error to 0 in iwm_send_cmd().
From OpenBSD rev.1.27.

match iwn and init txrate at lowest available bit-rate.
From OpenBSD rev.1.26.

workaround for interrupt coalescing bug not needed on 7265.
From OpenBSD if_iwm.c rev.1.25, if_iwmvar.h rev.1.7.

Correctly set lowest_present_ofdm in iwm_mvm_ack_rates().
From OpenBSD rev.1.23.

Display code number of unhandled packet.

Work around Coverity CIDs 127102{0,1}: Confused by the comparison
nrates > __arraycount(lq->rs_table), coverity erroneously binds nrates
to the range [0..16), whereas it is really bound to [0..15) by
ni->ni_rates.rs_nrates (IEEE80211_RATE_MAXSIZE). We add an explicit comparison
to unconfuse it that will always succeed.

Revert previous change.

CID 1271021: Overrunning array "in->in_ridx" of 15 bytes at byte offset 15 using index "i" (which evaluates to 15).

http://mail-index.netbsd.org/coverity-updates/2015/02/21/msg000115.html

Make the iwm(4) hardware receive filter pass multicast traffic from OpenBSD.
Fixes inet6 autoconf, for example.

Don't return success when hw is not ready.

fix compile failure with IWM_DEBUG on i386.

whitespace

Fix compile failure with clang.
Pointed out by Herbert J. Skuhra.

print error message when fail to load firmware.

iwm(4) works now, but 2.4GHz only.

> iwm0 at pci2 dev 0 function 0: Intel Dual Band Wireless AC 3160 (rev. 0x83)
> iwm0: interrupting at ioapic0 pin 18
> iwm0: hw rev: 0x160, fw ver 25.228 (API ver 9), address XX:XX:XX:XX:XX:XX
> iwm0: 11b rates: 1Mbps 2Mbps 5.5Mbps 11Mbps
> iwm0: 11g rates: 1Mbps 2Mbps 5.5Mbps 11Mbps 6Mbps 9Mbps 12Mbps 18Mbps 24Mbps 36Mbps 48Mbps 54Mbps

Make some functions and variables static.

Convert printf() to aprint_*().

Sync with OpenBSD if_iwm.c rev.1.18.

Add a driver for Intel Centrino 7260 and similar wireless cards.
Supported devices should more or less match ones supported
by the Linux iwlwifi mvm driver. Sponsored by genua mbh for OpenBSD.

This is probably the world's first Canadian cross device driver: it was
created for OpenBSD by writing and porting a NetBSD driver which was
developed in a rump kernel in Linux userspace.

Note: I don't have access to the hardware anymore, so this version is
not tested and not enabled by default. While I tried to be careful in
adding the NetBSD bits back, it's probable that there's a snafu or two.
Feel free to send private email in case you have the hardware and there
are issues.

s/ for for / for / in comment.

iwm(4): Disentangle attach.

Don't attach a half-baked interface and then detach it and then
reattach it after mountroot when we can read firmware; just defer
attaching the interface altogether until mountroot.

Likely fixes some panics I've seen every now and then at boot with
iwm(4).

if_attach and if_initialize cannot fail, don't test return value

These were originally made failable back in 2017 when if_initialize
allocated a softint in every interface for link state changes, so
that it could fail gracefully instead of panicking:

https://mail-index.NetBSD.org/source-changes/2017/10/23/msg089053.html

However, this spawned many seldom- or never-tested error branches,
which are risky to have around. And that softint in every interface
has since been replaced by a single global workqueue, because link
state changes require thread context but not low latency or high
throughput:

https://mail-index.NetBSD.org/source-changes/2020/02/06/msg113759.html

So there is no longer any reason for if_initialize to fail. (The
subroutine if_stats_init can't fail because percpu_alloc can't fail
either.)

There is a snag: the softint_establish in if_percpuq_create could
fail, potentially leading to bad consequences later on trying to use
the softint. This change doesn't introduce any new bugs because of
the snag -- if_percpuq_attach was already broken. However, the snag
can be better addressed without spawning error branches, either by
using a single softint or making softints less scarce.

(Separate commit will change the signatures of if_attach and
if_initialize to return void, scheduled to ride whatever is the next
convenient kernel bump.)

Patch and testing on amd64 and evbmips64-eb by maya@; commit message
soliloquy, and compile-testing on evbppc/i386/earmv7hf, by me.

branches: 1.85.6;
This driver sleeps during iwm_media_change(), and thus requires an
adaptive mutex for the media lock.

Adopt <net/if_stats.h>.

branches: 1.83.2; 1.83.10;
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.

iwm(4) workaround as ifp->if_hwdl can be set only once. Advised nonaka@n.o and ozaki-r@n.o, thanks.

To avoid for iwm(4) to set if_hwdl before loading firmware, use local address
as dummy. See sys/net/if.c::if_set_sadl().

Fix "ifconfig iwm0" failures reported by ryoon@n.o. Advised nonaka@n.o and ozaki-r@n.o, thanks.

Remove duplicate ;

Fix panic on boot with iwm(4). Advised by nonaka@n.o, thanks.

XXX pullup-8

branches: 1.78.2;
CID-1427736: Appease coverity using KASSERT

Match Intel Dual Band Wireless AC 3168. The code already supports it.

branches: 1.76.2;
- If if_initialize() failed in the attach function, free resources and return.
- Fix error path in the attach function correctly.

branches: 1.75.2;
kmem_intr_free kmem_intr_[z]alloced memory

the underlying pools are the same but api-wise those should match

fix typo.

reported by ryo@.

branches: 1.73.2;
remove checks for failure after memory allocation calls that cannot fail:

kmem_alloc() with KM_SLEEP
kmem_zalloc() with KM_SLEEP
percpu_alloc()
pserialize_create()
psref_class_create()

all of these paths include an assertion that the allocation has not failed,
so callers should not assert that again.

iwm(4): match Wireless 8265 device.

iwm(4): update firmware to maintained core release.

firmware from https://git.kernel.org/pub/scm/linux/kernel/git/iwlwifi/linux-firmware.git/tree/?id=4b9559fb0565f582a21a4070a81a3b6a7d77610b

wlan interfaces make interrupt routine running on softint context.

see http://mail-index.netbsd.org/tech-kern/2016/12/06/msg021281.html

tested device:
* ath at pci: AR5212, AR5424
* athn at pci: AR9287
* ipw at pci: 2100BG
* iwi at pci: 2915ABG
* iwm at pci: 3165, 7260, 8260
* iwn at pci: 4945, 6235
* ral at pci: RT2560
* rtwn at pci: RTL8192CE

PR/51797: initialize hardware completely when receiving a deauth frame from AP.

avoid "iwm0: fatal firmware error".

more block network interrupt.

increase command response timeout value.

include <sys/intr.h> for softint(9).

newstate should be blocked network interrupt, because it called iwm_start.

show NVM channel flags when AB_DEBUG is set.

more acknowledge interrupts from iwn(4).

only bus_dmamap_sync(9) once for ICT region.

branches: 1.61.2;
iwm(4): 3165 works fine with iwlwifi-7265D-17.ucode firmware.

firmware from http://git.kernel.org/cgit/linux/kernel/git/iwlwifi/linux-firmware.git/tree/?id=be8a6fcba2e48e00674d1f35e6def56e2268039c

check LAR support in NVM for 8260/4165.

Add apmg_wake_up_wa support from DragonFly BSD.

iwm(4): 8260 works now. 4165 may works too.

set NULL after free.

Apply OpenBSD if_iwm.c rev.1.148.

> While setting up the basic rate bitmask for iwm's firmware, if the AP does
> not specify basic rates for either the CCK or OFDM set, add just the most
> basic rate to that set (1 Mbit/s in case of CCK, 6 Mbit/s in case of OFDM).
> This behaviour matches what code comments seem to imply.

> The previous code would add all possible basic rates in such cases.
> So if all basic rates were CCK only, the code would add all possible OFDM basic
> rates on top. Then the firmware would send some frames at too high rates, e.g.
> RTS frames would be sent at 24Mbit/s which is a bit risky on noisy channels.

Do not use the "flags" field uninitialized

A backport if iwm(4) (to netbsd-7) does not build without this change.

LGTM nonaka@

Added enum iwm_tx_pm_timeouts from FreeBSD.

cosmetic.

use correct header size.

fix indent

iwm(4): make interrupt routine running on softint context.

see http://mail-index.netbsd.org/tech-kern/2016/12/06/msg021281.html

use pci_intr_establish_xname(9).

twiddle bpf related.

cosmetic

Export some 802.11 IE manipulate functions.

iwm(4): Sync with OpenBSD.

7265 works fine.
7260 and 3160 maybe work.
7265D, 3165 and 8260 not work yet.

if_iwm.c: r1.147.
if_iwmreg.h: r1.19 without r1.16.
if_iwmvar.h: r1.24.

firmware is taken from https://wireless.wiki.kernel.org/en/users/drivers/iwlwifi
- iwlwifi-7260-ucode-16.242414.0.tgz
- iwlwifi-3160-ucode-16.242414.0.tgz
- iwlwifi-7265-ucode-16.242414.0.tgz
- iwlwifi-8000-ucode-16.242414.0.tgz

Apply deferred if_start framework

if_schedule_deferred_start checks if the if_snd queue contains packets,
so drivers don't need to check it by themselves.

Fix inverted logic in iwm(4).

The PROT_REQUIRE flag in should be set for data frames above a certain
length, but we were setting it for !data frames above a certain length,
which makes no sense at all.

- taken from all other BSDs (OpenBSD, FreeBSD, DragonflyBSD...)

branches: 1.42.2;
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.

Use M_GETCTX

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!

Always use pci_intr_alloc(9)/pci_intr_release(9).

Remove leading whitespaces and kind

Use pci_intr_alloc().

iwm(4) can read the firmware image from disk when interface up.
This change is useful for full-disk encryption environment.

Limit the number of dma segments used for transmitting packets to
IWM_NUM_OF_TBS - 2. We have IWM_NUM_OF_TBS slots, but use two of those
for sending commands to the firmware.

From OpenBSD rev. 1.41.

should be able to use 5GHz wireless network.

if_iwm use unified establish API.

Added control debugging output via sysctl.

Use MSI instead of legacy INTx, if available.

should fix PR/49663.

Register iwm(4) with pmf(9).

Patch from Pierre Pronchery.

branches: 1.29.2;
Not all iwm hardware supports the 5GHz band. Make setting 11a rates
and scanning conditional on the 5GHz support bit in the nvm.

From OpenBSD rev. 1.39. Arranged by me.

Fix radiotap and fixed rates. Radiotap was reporting the wrong rate for
data frames and fixed rates weren't really fixed and were converted into
the wrong hardware rate.

From OpenBSD rev.1.36.

Avoid kmem_free(NULL) in iwm_read_firmware()

This code path can be hit if the firmware failed to load, for instance
if the file is not present on the filesystem. In this case
firmware_open() fails, and fw->fw_rawdata never gets allocated in the
first place.

stop the calibration timeout when stopping the interface.

From OpenBSD rev.1.35.

Set/clear the IWM_MAC_FILTER_IN_BEACON flag under conditions that match what
Linux does.

From OpenBSD rev.1.34.

Fix 11a support.

From OpenBSD rev.1.33.

Make setting the MAC context bit for 11g protection when transmitting
OFDM frames conditional on the node via IEEE80211_F_USEPROT.

From OpenBSD rev.1.31.

Don't set CTS to self. Linux stopped setting this back in
dc271ee0d04d12d6bfabacbec803289a7072fbd9 as it is known
to cause problems.

From OpenBSD rev.1.30.

sc_fixed_ridx is initialised to 0 but tested as it it were initialised
to -1. The result of this is tx frames were always sent out at fixed
rate 0 instead of ni_txrate.

Match the iwn behaviour and test ic_fixed_rate for -1 instead.

From OpenBSD rev.1.29.

Correctly set lowest_present_ofdm in iwm_mvm_ack_rates().
From OpenBSD rev.1.28.

init error to 0 in iwm_send_cmd().
From OpenBSD rev.1.27.

match iwn and init txrate at lowest available bit-rate.
From OpenBSD rev.1.26.

workaround for interrupt coalescing bug not needed on 7265.
From OpenBSD if_iwm.c rev.1.25, if_iwmvar.h rev.1.7.

Correctly set lowest_present_ofdm in iwm_mvm_ack_rates().
From OpenBSD rev.1.23.

Display code number of unhandled packet.

Work around Coverity CIDs 127102{0,1}: Confused by the comparison
nrates > __arraycount(lq->rs_table), coverity erroneously binds nrates
to the range [0..16), whereas it is really bound to [0..15) by
ni->ni_rates.rs_nrates (IEEE80211_RATE_MAXSIZE). We add an explicit comparison
to unconfuse it that will always succeed.

Revert previous change.

CID 1271021: Overrunning array "in->in_ridx" of 15 bytes at byte offset 15 using index "i" (which evaluates to 15).

http://mail-index.netbsd.org/coverity-updates/2015/02/21/msg000115.html

Make the iwm(4) hardware receive filter pass multicast traffic from OpenBSD.
Fixes inet6 autoconf, for example.

Don't return success when hw is not ready.

fix compile failure with IWM_DEBUG on i386.

whitespace

Fix compile failure with clang.
Pointed out by Herbert J. Skuhra.

print error message when fail to load firmware.

iwm(4) works now, but 2.4GHz only.

> iwm0 at pci2 dev 0 function 0: Intel Dual Band Wireless AC 3160 (rev. 0x83)
> iwm0: interrupting at ioapic0 pin 18
> iwm0: hw rev: 0x160, fw ver 25.228 (API ver 9), address XX:XX:XX:XX:XX:XX
> iwm0: 11b rates: 1Mbps 2Mbps 5.5Mbps 11Mbps
> iwm0: 11g rates: 1Mbps 2Mbps 5.5Mbps 11Mbps 6Mbps 9Mbps 12Mbps 18Mbps 24Mbps 36Mbps 48Mbps 54Mbps

Make some functions and variables static.

Convert printf() to aprint_*().

Sync with OpenBSD if_iwm.c rev.1.18.

Add a driver for Intel Centrino 7260 and similar wireless cards.
Supported devices should more or less match ones supported
by the Linux iwlwifi mvm driver. Sponsored by genua mbh for OpenBSD.

This is probably the world's first Canadian cross device driver: it was
created for OpenBSD by writing and porting a NetBSD driver which was
developed in a rump kernel in Linux userspace.

Note: I don't have access to the hardware anymore, so this version is
not tested and not enabled by default. While I tried to be careful in
adding the NetBSD bits back, it's probable that there's a snafu or two.
Feel free to send private email in case you have the hardware and there
are issues.

iwm(4): Disentangle attach.

Don't attach a half-baked interface and then detach it and then
reattach it after mountroot when we can read firmware; just defer
attaching the interface altogether until mountroot.

Likely fixes some panics I've seen every now and then at boot with
iwm(4).

if_attach and if_initialize cannot fail, don't test return value

These were originally made failable back in 2017 when if_initialize
allocated a softint in every interface for link state changes, so
that it could fail gracefully instead of panicking:

https://mail-index.NetBSD.org/source-changes/2017/10/23/msg089053.html

However, this spawned many seldom- or never-tested error branches,
which are risky to have around. And that softint in every interface
has since been replaced by a single global workqueue, because link
state changes require thread context but not low latency or high
throughput:

https://mail-index.NetBSD.org/source-changes/2020/02/06/msg113759.html

So there is no longer any reason for if_initialize to fail. (The
subroutine if_stats_init can't fail because percpu_alloc can't fail
either.)

There is a snag: the softint_establish in if_percpuq_create could
fail, potentially leading to bad consequences later on trying to use
the softint. This change doesn't introduce any new bugs because of
the snag -- if_percpuq_attach was already broken. However, the snag
can be better addressed without spawning error branches, either by
using a single softint or making softints less scarce.

(Separate commit will change the signatures of if_attach and
if_initialize to return void, scheduled to ride whatever is the next
convenient kernel bump.)

Patch and testing on amd64 and evbmips64-eb by maya@; commit message
soliloquy, and compile-testing on evbppc/i386/earmv7hf, by me.

branches: 1.85.6;
This driver sleeps during iwm_media_change(), and thus requires an
adaptive mutex for the media lock.

Adopt <net/if_stats.h>.

branches: 1.83.2; 1.83.10;
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.

iwm(4) workaround as ifp->if_hwdl can be set only once. Advised nonaka@n.o and ozaki-r@n.o, thanks.

To avoid for iwm(4) to set if_hwdl before loading firmware, use local address
as dummy. See sys/net/if.c::if_set_sadl().

Fix "ifconfig iwm0" failures reported by ryoon@n.o. Advised nonaka@n.o and ozaki-r@n.o, thanks.

Remove duplicate ;

Fix panic on boot with iwm(4). Advised by nonaka@n.o, thanks.

XXX pullup-8

branches: 1.78.2;
CID-1427736: Appease coverity using KASSERT

Match Intel Dual Band Wireless AC 3168. The code already supports it.

branches: 1.76.2;
- If if_initialize() failed in the attach function, free resources and return.
- Fix error path in the attach function correctly.

branches: 1.75.2;
kmem_intr_free kmem_intr_[z]alloced memory

the underlying pools are the same but api-wise those should match

fix typo.

reported by ryo@.

branches: 1.73.2;
remove checks for failure after memory allocation calls that cannot fail:

kmem_alloc() with KM_SLEEP
kmem_zalloc() with KM_SLEEP
percpu_alloc()
pserialize_create()
psref_class_create()

all of these paths include an assertion that the allocation has not failed,
so callers should not assert that again.

iwm(4): match Wireless 8265 device.

iwm(4): update firmware to maintained core release.

firmware from https://git.kernel.org/pub/scm/linux/kernel/git/iwlwifi/linux-firmware.git/tree/?id=4b9559fb0565f582a21a4070a81a3b6a7d77610b

wlan interfaces make interrupt routine running on softint context.

see http://mail-index.netbsd.org/tech-kern/2016/12/06/msg021281.html

tested device:
* ath at pci: AR5212, AR5424
* athn at pci: AR9287
* ipw at pci: 2100BG
* iwi at pci: 2915ABG
* iwm at pci: 3165, 7260, 8260
* iwn at pci: 4945, 6235
* ral at pci: RT2560
* rtwn at pci: RTL8192CE

PR/51797: initialize hardware completely when receiving a deauth frame from AP.

avoid "iwm0: fatal firmware error".

more block network interrupt.

increase command response timeout value.

include <sys/intr.h> for softint(9).

newstate should be blocked network interrupt, because it called iwm_start.

show NVM channel flags when AB_DEBUG is set.

more acknowledge interrupts from iwn(4).

only bus_dmamap_sync(9) once for ICT region.

branches: 1.61.2;
iwm(4): 3165 works fine with iwlwifi-7265D-17.ucode firmware.

firmware from http://git.kernel.org/cgit/linux/kernel/git/iwlwifi/linux-firmware.git/tree/?id=be8a6fcba2e48e00674d1f35e6def56e2268039c

check LAR support in NVM for 8260/4165.

Add apmg_wake_up_wa support from DragonFly BSD.

iwm(4): 8260 works now. 4165 may works too.

set NULL after free.

Apply OpenBSD if_iwm.c rev.1.148.

> While setting up the basic rate bitmask for iwm's firmware, if the AP does
> not specify basic rates for either the CCK or OFDM set, add just the most
> basic rate to that set (1 Mbit/s in case of CCK, 6 Mbit/s in case of OFDM).
> This behaviour matches what code comments seem to imply.

> The previous code would add all possible basic rates in such cases.
> So if all basic rates were CCK only, the code would add all possible OFDM basic
> rates on top. Then the firmware would send some frames at too high rates, e.g.
> RTS frames would be sent at 24Mbit/s which is a bit risky on noisy channels.

Do not use the "flags" field uninitialized

A backport if iwm(4) (to netbsd-7) does not build without this change.

LGTM nonaka@

Added enum iwm_tx_pm_timeouts from FreeBSD.

cosmetic.

use correct header size.

fix indent

iwm(4): make interrupt routine running on softint context.

see http://mail-index.netbsd.org/tech-kern/2016/12/06/msg021281.html

use pci_intr_establish_xname(9).

twiddle bpf related.

cosmetic

Export some 802.11 IE manipulate functions.

iwm(4): Sync with OpenBSD.

7265 works fine.
7260 and 3160 maybe work.
7265D, 3165 and 8260 not work yet.

if_iwm.c: r1.147.
if_iwmreg.h: r1.19 without r1.16.
if_iwmvar.h: r1.24.

firmware is taken from https://wireless.wiki.kernel.org/en/users/drivers/iwlwifi
- iwlwifi-7260-ucode-16.242414.0.tgz
- iwlwifi-3160-ucode-16.242414.0.tgz
- iwlwifi-7265-ucode-16.242414.0.tgz
- iwlwifi-8000-ucode-16.242414.0.tgz

Apply deferred if_start framework

if_schedule_deferred_start checks if the if_snd queue contains packets,
so drivers don't need to check it by themselves.

Fix inverted logic in iwm(4).

The PROT_REQUIRE flag in should be set for data frames above a certain
length, but we were setting it for !data frames above a certain length,
which makes no sense at all.

- taken from all other BSDs (OpenBSD, FreeBSD, DragonflyBSD...)

branches: 1.42.2;
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.

Use M_GETCTX

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!

Always use pci_intr_alloc(9)/pci_intr_release(9).

Remove leading whitespaces and kind

Use pci_intr_alloc().

iwm(4) can read the firmware image from disk when interface up.
This change is useful for full-disk encryption environment.

Limit the number of dma segments used for transmitting packets to
IWM_NUM_OF_TBS - 2. We have IWM_NUM_OF_TBS slots, but use two of those
for sending commands to the firmware.

From OpenBSD rev. 1.41.

should be able to use 5GHz wireless network.

if_iwm use unified establish API.

Added control debugging output via sysctl.

Use MSI instead of legacy INTx, if available.

should fix PR/49663.

Register iwm(4) with pmf(9).

Patch from Pierre Pronchery.

branches: 1.29.2;
Not all iwm hardware supports the 5GHz band. Make setting 11a rates
and scanning conditional on the 5GHz support bit in the nvm.

From OpenBSD rev. 1.39. Arranged by me.

Fix radiotap and fixed rates. Radiotap was reporting the wrong rate for
data frames and fixed rates weren't really fixed and were converted into
the wrong hardware rate.

From OpenBSD rev.1.36.

Avoid kmem_free(NULL) in iwm_read_firmware()

This code path can be hit if the firmware failed to load, for instance
if the file is not present on the filesystem. In this case
firmware_open() fails, and fw->fw_rawdata never gets allocated in the
first place.

stop the calibration timeout when stopping the interface.

From OpenBSD rev.1.35.

Set/clear the IWM_MAC_FILTER_IN_BEACON flag under conditions that match what
Linux does.

From OpenBSD rev.1.34.

Fix 11a support.

From OpenBSD rev.1.33.

Make setting the MAC context bit for 11g protection when transmitting
OFDM frames conditional on the node via IEEE80211_F_USEPROT.

From OpenBSD rev.1.31.

Don't set CTS to self. Linux stopped setting this back in
dc271ee0d04d12d6bfabacbec803289a7072fbd9 as it is known
to cause problems.

From OpenBSD rev.1.30.

sc_fixed_ridx is initialised to 0 but tested as it it were initialised
to -1. The result of this is tx frames were always sent out at fixed
rate 0 instead of ni_txrate.

Match the iwn behaviour and test ic_fixed_rate for -1 instead.

From OpenBSD rev.1.29.

Correctly set lowest_present_ofdm in iwm_mvm_ack_rates().
From OpenBSD rev.1.28.

init error to 0 in iwm_send_cmd().
From OpenBSD rev.1.27.

match iwn and init txrate at lowest available bit-rate.
From OpenBSD rev.1.26.

workaround for interrupt coalescing bug not needed on 7265.
From OpenBSD if_iwm.c rev.1.25, if_iwmvar.h rev.1.7.

Correctly set lowest_present_ofdm in iwm_mvm_ack_rates().
From OpenBSD rev.1.23.

Display code number of unhandled packet.

Work around Coverity CIDs 127102{0,1}: Confused by the comparison
nrates > __arraycount(lq->rs_table), coverity erroneously binds nrates
to the range [0..16), whereas it is really bound to [0..15) by
ni->ni_rates.rs_nrates (IEEE80211_RATE_MAXSIZE). We add an explicit comparison
to unconfuse it that will always succeed.

Revert previous change.

CID 1271021: Overrunning array "in->in_ridx" of 15 bytes at byte offset 15 using index "i" (which evaluates to 15).

http://mail-index.netbsd.org/coverity-updates/2015/02/21/msg000115.html

Make the iwm(4) hardware receive filter pass multicast traffic from OpenBSD.
Fixes inet6 autoconf, for example.

Don't return success when hw is not ready.

fix compile failure with IWM_DEBUG on i386.

whitespace

Fix compile failure with clang.
Pointed out by Herbert J. Skuhra.

print error message when fail to load firmware.

iwm(4) works now, but 2.4GHz only.

> iwm0 at pci2 dev 0 function 0: Intel Dual Band Wireless AC 3160 (rev. 0x83)
> iwm0: interrupting at ioapic0 pin 18
> iwm0: hw rev: 0x160, fw ver 25.228 (API ver 9), address XX:XX:XX:XX:XX:XX
> iwm0: 11b rates: 1Mbps 2Mbps 5.5Mbps 11Mbps
> iwm0: 11g rates: 1Mbps 2Mbps 5.5Mbps 11Mbps 6Mbps 9Mbps 12Mbps 18Mbps 24Mbps 36Mbps 48Mbps 54Mbps

Make some functions and variables static.

Convert printf() to aprint_*().

Sync with OpenBSD if_iwm.c rev.1.18.

Add a driver for Intel Centrino 7260 and similar wireless cards.
Supported devices should more or less match ones supported
by the Linux iwlwifi mvm driver. Sponsored by genua mbh for OpenBSD.

This is probably the world's first Canadian cross device driver: it was
created for OpenBSD by writing and porting a NetBSD driver which was
developed in a rump kernel in Linux userspace.

Note: I don't have access to the hardware anymore, so this version is
not tested and not enabled by default. While I tried to be careful in
adding the NetBSD bits back, it's probable that there's a snafu or two.
Feel free to send private email in case you have the hardware and there
are issues.

if_attach and if_initialize cannot fail, don't test return value

These were originally made failable back in 2017 when if_initialize
allocated a softint in every interface for link state changes, so
that it could fail gracefully instead of panicking:

https://mail-index.NetBSD.org/source-changes/2017/10/23/msg089053.html

However, this spawned many seldom- or never-tested error branches,
which are risky to have around. And that softint in every interface
has since been replaced by a single global workqueue, because link
state changes require thread context but not low latency or high
throughput:

https://mail-index.NetBSD.org/source-changes/2020/02/06/msg113759.html

So there is no longer any reason for if_initialize to fail. (The
subroutine if_stats_init can't fail because percpu_alloc can't fail
either.)

There is a snag: the softint_establish in if_percpuq_create could
fail, potentially leading to bad consequences later on trying to use
the softint. This change doesn't introduce any new bugs because of
the snag -- if_percpuq_attach was already broken. However, the snag
can be better addressed without spawning error branches, either by
using a single softint or making softints less scarce.

(Separate commit will change the signatures of if_attach and
if_initialize to return void, scheduled to ride whatever is the next
convenient kernel bump.)

Patch and testing on amd64 and evbmips64-eb by maya@; commit message
soliloquy, and compile-testing on evbppc/i386/earmv7hf, by me.

This driver sleeps during iwm_media_change(), and thus requires an
adaptive mutex for the media lock.

Adopt <net/if_stats.h>.

branches: 1.83.2; 1.83.10;
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.

iwm(4) workaround as ifp->if_hwdl can be set only once. Advised nonaka@n.o and ozaki-r@n.o, thanks.

To avoid for iwm(4) to set if_hwdl before loading firmware, use local address
as dummy. See sys/net/if.c::if_set_sadl().

Fix "ifconfig iwm0" failures reported by ryoon@n.o. Advised nonaka@n.o and ozaki-r@n.o, thanks.

Remove duplicate ;

Fix panic on boot with iwm(4). Advised by nonaka@n.o, thanks.

XXX pullup-8

branches: 1.78.2;
CID-1427736: Appease coverity using KASSERT

Match Intel Dual Band Wireless AC 3168. The code already supports it.

branches: 1.76.2;
- If if_initialize() failed in the attach function, free resources and return.
- Fix error path in the attach function correctly.

branches: 1.75.2;
kmem_intr_free kmem_intr_[z]alloced memory

the underlying pools are the same but api-wise those should match

fix typo.

reported by ryo@.

branches: 1.73.2;
remove checks for failure after memory allocation calls that cannot fail:

kmem_alloc() with KM_SLEEP
kmem_zalloc() with KM_SLEEP
percpu_alloc()
pserialize_create()
psref_class_create()

all of these paths include an assertion that the allocation has not failed,
so callers should not assert that again.

iwm(4): match Wireless 8265 device.

iwm(4): update firmware to maintained core release.

firmware from https://git.kernel.org/pub/scm/linux/kernel/git/iwlwifi/linux-firmware.git/tree/?id=4b9559fb0565f582a21a4070a81a3b6a7d77610b

wlan interfaces make interrupt routine running on softint context.

see http://mail-index.netbsd.org/tech-kern/2016/12/06/msg021281.html

tested device:
* ath at pci: AR5212, AR5424
* athn at pci: AR9287
* ipw at pci: 2100BG
* iwi at pci: 2915ABG
* iwm at pci: 3165, 7260, 8260
* iwn at pci: 4945, 6235
* ral at pci: RT2560
* rtwn at pci: RTL8192CE

PR/51797: initialize hardware completely when receiving a deauth frame from AP.

avoid "iwm0: fatal firmware error".

more block network interrupt.

increase command response timeout value.

include <sys/intr.h> for softint(9).

newstate should be blocked network interrupt, because it called iwm_start.

show NVM channel flags when AB_DEBUG is set.

more acknowledge interrupts from iwn(4).

only bus_dmamap_sync(9) once for ICT region.

branches: 1.61.2;
iwm(4): 3165 works fine with iwlwifi-7265D-17.ucode firmware.

firmware from http://git.kernel.org/cgit/linux/kernel/git/iwlwifi/linux-firmware.git/tree/?id=be8a6fcba2e48e00674d1f35e6def56e2268039c

check LAR support in NVM for 8260/4165.

Add apmg_wake_up_wa support from DragonFly BSD.

iwm(4): 8260 works now. 4165 may works too.

set NULL after free.

Apply OpenBSD if_iwm.c rev.1.148.

> While setting up the basic rate bitmask for iwm's firmware, if the AP does
> not specify basic rates for either the CCK or OFDM set, add just the most
> basic rate to that set (1 Mbit/s in case of CCK, 6 Mbit/s in case of OFDM).
> This behaviour matches what code comments seem to imply.

> The previous code would add all possible basic rates in such cases.
> So if all basic rates were CCK only, the code would add all possible OFDM basic
> rates on top. Then the firmware would send some frames at too high rates, e.g.
> RTS frames would be sent at 24Mbit/s which is a bit risky on noisy channels.

Do not use the "flags" field uninitialized

A backport if iwm(4) (to netbsd-7) does not build without this change.

LGTM nonaka@

Added enum iwm_tx_pm_timeouts from FreeBSD.

cosmetic.

use correct header size.

fix indent

iwm(4): make interrupt routine running on softint context.

see http://mail-index.netbsd.org/tech-kern/2016/12/06/msg021281.html

use pci_intr_establish_xname(9).

twiddle bpf related.

cosmetic

Export some 802.11 IE manipulate functions.

iwm(4): Sync with OpenBSD.

7265 works fine.
7260 and 3160 maybe work.
7265D, 3165 and 8260 not work yet.

if_iwm.c: r1.147.
if_iwmreg.h: r1.19 without r1.16.
if_iwmvar.h: r1.24.

firmware is taken from https://wireless.wiki.kernel.org/en/users/drivers/iwlwifi
- iwlwifi-7260-ucode-16.242414.0.tgz
- iwlwifi-3160-ucode-16.242414.0.tgz
- iwlwifi-7265-ucode-16.242414.0.tgz
- iwlwifi-8000-ucode-16.242414.0.tgz

Apply deferred if_start framework

if_schedule_deferred_start checks if the if_snd queue contains packets,
so drivers don't need to check it by themselves.

Fix inverted logic in iwm(4).

The PROT_REQUIRE flag in should be set for data frames above a certain
length, but we were setting it for !data frames above a certain length,
which makes no sense at all.

- taken from all other BSDs (OpenBSD, FreeBSD, DragonflyBSD...)

branches: 1.42.2;
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.

Use M_GETCTX

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!

Always use pci_intr_alloc(9)/pci_intr_release(9).

Remove leading whitespaces and kind

Use pci_intr_alloc().

iwm(4) can read the firmware image from disk when interface up.
This change is useful for full-disk encryption environment.

Limit the number of dma segments used for transmitting packets to
IWM_NUM_OF_TBS - 2. We have IWM_NUM_OF_TBS slots, but use two of those
for sending commands to the firmware.

From OpenBSD rev. 1.41.

should be able to use 5GHz wireless network.

if_iwm use unified establish API.

Added control debugging output via sysctl.

Use MSI instead of legacy INTx, if available.

should fix PR/49663.

Register iwm(4) with pmf(9).

Patch from Pierre Pronchery.

branches: 1.29.2;
Not all iwm hardware supports the 5GHz band. Make setting 11a rates
and scanning conditional on the 5GHz support bit in the nvm.

From OpenBSD rev. 1.39. Arranged by me.

Fix radiotap and fixed rates. Radiotap was reporting the wrong rate for
data frames and fixed rates weren't really fixed and were converted into
the wrong hardware rate.

From OpenBSD rev.1.36.

Avoid kmem_free(NULL) in iwm_read_firmware()

This code path can be hit if the firmware failed to load, for instance
if the file is not present on the filesystem. In this case
firmware_open() fails, and fw->fw_rawdata never gets allocated in the
first place.

stop the calibration timeout when stopping the interface.

From OpenBSD rev.1.35.

Set/clear the IWM_MAC_FILTER_IN_BEACON flag under conditions that match what
Linux does.

From OpenBSD rev.1.34.

Fix 11a support.

From OpenBSD rev.1.33.

Make setting the MAC context bit for 11g protection when transmitting
OFDM frames conditional on the node via IEEE80211_F_USEPROT.

From OpenBSD rev.1.31.

Don't set CTS to self. Linux stopped setting this back in
dc271ee0d04d12d6bfabacbec803289a7072fbd9 as it is known
to cause problems.

From OpenBSD rev.1.30.

sc_fixed_ridx is initialised to 0 but tested as it it were initialised
to -1. The result of this is tx frames were always sent out at fixed
rate 0 instead of ni_txrate.

Match the iwn behaviour and test ic_fixed_rate for -1 instead.

From OpenBSD rev.1.29.

Correctly set lowest_present_ofdm in iwm_mvm_ack_rates().
From OpenBSD rev.1.28.

init error to 0 in iwm_send_cmd().
From OpenBSD rev.1.27.

match iwn and init txrate at lowest available bit-rate.
From OpenBSD rev.1.26.

workaround for interrupt coalescing bug not needed on 7265.
From OpenBSD if_iwm.c rev.1.25, if_iwmvar.h rev.1.7.

Correctly set lowest_present_ofdm in iwm_mvm_ack_rates().
From OpenBSD rev.1.23.

Display code number of unhandled packet.

Work around Coverity CIDs 127102{0,1}: Confused by the comparison
nrates > __arraycount(lq->rs_table), coverity erroneously binds nrates
to the range [0..16), whereas it is really bound to [0..15) by
ni->ni_rates.rs_nrates (IEEE80211_RATE_MAXSIZE). We add an explicit comparison
to unconfuse it that will always succeed.

Revert previous change.

CID 1271021: Overrunning array "in->in_ridx" of 15 bytes at byte offset 15 using index "i" (which evaluates to 15).

http://mail-index.netbsd.org/coverity-updates/2015/02/21/msg000115.html

Make the iwm(4) hardware receive filter pass multicast traffic from OpenBSD.
Fixes inet6 autoconf, for example.

Don't return success when hw is not ready.

fix compile failure with IWM_DEBUG on i386.

whitespace

Fix compile failure with clang.
Pointed out by Herbert J. Skuhra.

print error message when fail to load firmware.

iwm(4) works now, but 2.4GHz only.

> iwm0 at pci2 dev 0 function 0: Intel Dual Band Wireless AC 3160 (rev. 0x83)
> iwm0: interrupting at ioapic0 pin 18
> iwm0: hw rev: 0x160, fw ver 25.228 (API ver 9), address XX:XX:XX:XX:XX:XX
> iwm0: 11b rates: 1Mbps 2Mbps 5.5Mbps 11Mbps
> iwm0: 11g rates: 1Mbps 2Mbps 5.5Mbps 11Mbps 6Mbps 9Mbps 12Mbps 18Mbps 24Mbps 36Mbps 48Mbps 54Mbps

Make some functions and variables static.

Convert printf() to aprint_*().

Sync with OpenBSD if_iwm.c rev.1.18.

Add a driver for Intel Centrino 7260 and similar wireless cards.
Supported devices should more or less match ones supported
by the Linux iwlwifi mvm driver. Sponsored by genua mbh for OpenBSD.

This is probably the world's first Canadian cross device driver: it was
created for OpenBSD by writing and porting a NetBSD driver which was
developed in a rump kernel in Linux userspace.

Note: I don't have access to the hardware anymore, so this version is
not tested and not enabled by default. While I tried to be careful in
adding the NetBSD bits back, it's probable that there's a snafu or two.
Feel free to send private email in case you have the hardware and there
are issues.

This driver sleeps during iwm_media_change(), and thus requires an
adaptive mutex for the media lock.

Adopt <net/if_stats.h>.

branches: 1.83.2; 1.83.10;
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.

iwm(4) workaround as ifp->if_hwdl can be set only once. Advised nonaka@n.o and ozaki-r@n.o, thanks.

To avoid for iwm(4) to set if_hwdl before loading firmware, use local address
as dummy. See sys/net/if.c::if_set_sadl().

Fix "ifconfig iwm0" failures reported by ryoon@n.o. Advised nonaka@n.o and ozaki-r@n.o, thanks.

Remove duplicate ;

Fix panic on boot with iwm(4). Advised by nonaka@n.o, thanks.

XXX pullup-8

branches: 1.78.2;
CID-1427736: Appease coverity using KASSERT

Match Intel Dual Band Wireless AC 3168. The code already supports it.

branches: 1.76.2;
- If if_initialize() failed in the attach function, free resources and return.
- Fix error path in the attach function correctly.

branches: 1.75.2;
kmem_intr_free kmem_intr_[z]alloced memory

the underlying pools are the same but api-wise those should match

fix typo.

reported by ryo@.

branches: 1.73.2;
remove checks for failure after memory allocation calls that cannot fail:

kmem_alloc() with KM_SLEEP
kmem_zalloc() with KM_SLEEP
percpu_alloc()
pserialize_create()
psref_class_create()

all of these paths include an assertion that the allocation has not failed,
so callers should not assert that again.

iwm(4): match Wireless 8265 device.

iwm(4): update firmware to maintained core release.

firmware from https://git.kernel.org/pub/scm/linux/kernel/git/iwlwifi/linux-firmware.git/tree/?id=4b9559fb0565f582a21a4070a81a3b6a7d77610b

wlan interfaces make interrupt routine running on softint context.

see http://mail-index.netbsd.org/tech-kern/2016/12/06/msg021281.html

tested device:
* ath at pci: AR5212, AR5424
* athn at pci: AR9287
* ipw at pci: 2100BG
* iwi at pci: 2915ABG
* iwm at pci: 3165, 7260, 8260
* iwn at pci: 4945, 6235
* ral at pci: RT2560
* rtwn at pci: RTL8192CE

PR/51797: initialize hardware completely when receiving a deauth frame from AP.

avoid "iwm0: fatal firmware error".

more block network interrupt.

increase command response timeout value.

include <sys/intr.h> for softint(9).

newstate should be blocked network interrupt, because it called iwm_start.

show NVM channel flags when AB_DEBUG is set.

more acknowledge interrupts from iwn(4).

only bus_dmamap_sync(9) once for ICT region.

branches: 1.61.2;
iwm(4): 3165 works fine with iwlwifi-7265D-17.ucode firmware.

firmware from http://git.kernel.org/cgit/linux/kernel/git/iwlwifi/linux-firmware.git/tree/?id=be8a6fcba2e48e00674d1f35e6def56e2268039c

check LAR support in NVM for 8260/4165.

Add apmg_wake_up_wa support from DragonFly BSD.

iwm(4): 8260 works now. 4165 may works too.

set NULL after free.

Apply OpenBSD if_iwm.c rev.1.148.

> While setting up the basic rate bitmask for iwm's firmware, if the AP does
> not specify basic rates for either the CCK or OFDM set, add just the most
> basic rate to that set (1 Mbit/s in case of CCK, 6 Mbit/s in case of OFDM).
> This behaviour matches what code comments seem to imply.

> The previous code would add all possible basic rates in such cases.
> So if all basic rates were CCK only, the code would add all possible OFDM basic
> rates on top. Then the firmware would send some frames at too high rates, e.g.
> RTS frames would be sent at 24Mbit/s which is a bit risky on noisy channels.

Do not use the "flags" field uninitialized

A backport if iwm(4) (to netbsd-7) does not build without this change.

LGTM nonaka@

Added enum iwm_tx_pm_timeouts from FreeBSD.

cosmetic.

use correct header size.

fix indent

iwm(4): make interrupt routine running on softint context.

see http://mail-index.netbsd.org/tech-kern/2016/12/06/msg021281.html

use pci_intr_establish_xname(9).

twiddle bpf related.

cosmetic

Export some 802.11 IE manipulate functions.

iwm(4): Sync with OpenBSD.

7265 works fine.
7260 and 3160 maybe work.
7265D, 3165 and 8260 not work yet.

if_iwm.c: r1.147.
if_iwmreg.h: r1.19 without r1.16.
if_iwmvar.h: r1.24.

firmware is taken from https://wireless.wiki.kernel.org/en/users/drivers/iwlwifi
- iwlwifi-7260-ucode-16.242414.0.tgz
- iwlwifi-3160-ucode-16.242414.0.tgz
- iwlwifi-7265-ucode-16.242414.0.tgz
- iwlwifi-8000-ucode-16.242414.0.tgz

Apply deferred if_start framework

if_schedule_deferred_start checks if the if_snd queue contains packets,
so drivers don't need to check it by themselves.

Fix inverted logic in iwm(4).

The PROT_REQUIRE flag in should be set for data frames above a certain
length, but we were setting it for !data frames above a certain length,
which makes no sense at all.

- taken from all other BSDs (OpenBSD, FreeBSD, DragonflyBSD...)

branches: 1.42.2;
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.

Use M_GETCTX

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!

Always use pci_intr_alloc(9)/pci_intr_release(9).

Remove leading whitespaces and kind

Use pci_intr_alloc().

iwm(4) can read the firmware image from disk when interface up.
This change is useful for full-disk encryption environment.

Limit the number of dma segments used for transmitting packets to
IWM_NUM_OF_TBS - 2. We have IWM_NUM_OF_TBS slots, but use two of those
for sending commands to the firmware.

From OpenBSD rev. 1.41.

should be able to use 5GHz wireless network.

if_iwm use unified establish API.

Added control debugging output via sysctl.

Use MSI instead of legacy INTx, if available.

should fix PR/49663.

Register iwm(4) with pmf(9).

Patch from Pierre Pronchery.

branches: 1.29.2;
Not all iwm hardware supports the 5GHz band. Make setting 11a rates
and scanning conditional on the 5GHz support bit in the nvm.

From OpenBSD rev. 1.39. Arranged by me.

Fix radiotap and fixed rates. Radiotap was reporting the wrong rate for
data frames and fixed rates weren't really fixed and were converted into
the wrong hardware rate.

From OpenBSD rev.1.36.

Avoid kmem_free(NULL) in iwm_read_firmware()

This code path can be hit if the firmware failed to load, for instance
if the file is not present on the filesystem. In this case
firmware_open() fails, and fw->fw_rawdata never gets allocated in the
first place.

stop the calibration timeout when stopping the interface.

From OpenBSD rev.1.35.

Set/clear the IWM_MAC_FILTER_IN_BEACON flag under conditions that match what
Linux does.

From OpenBSD rev.1.34.

Fix 11a support.

From OpenBSD rev.1.33.

Make setting the MAC context bit for 11g protection when transmitting
OFDM frames conditional on the node via IEEE80211_F_USEPROT.

From OpenBSD rev.1.31.

Don't set CTS to self. Linux stopped setting this back in
dc271ee0d04d12d6bfabacbec803289a7072fbd9 as it is known
to cause problems.

From OpenBSD rev.1.30.

sc_fixed_ridx is initialised to 0 but tested as it it were initialised
to -1. The result of this is tx frames were always sent out at fixed
rate 0 instead of ni_txrate.

Match the iwn behaviour and test ic_fixed_rate for -1 instead.

From OpenBSD rev.1.29.

Correctly set lowest_present_ofdm in iwm_mvm_ack_rates().
From OpenBSD rev.1.28.

init error to 0 in iwm_send_cmd().
From OpenBSD rev.1.27.

match iwn and init txrate at lowest available bit-rate.
From OpenBSD rev.1.26.

workaround for interrupt coalescing bug not needed on 7265.
From OpenBSD if_iwm.c rev.1.25, if_iwmvar.h rev.1.7.

Correctly set lowest_present_ofdm in iwm_mvm_ack_rates().
From OpenBSD rev.1.23.

Display code number of unhandled packet.

Work around Coverity CIDs 127102{0,1}: Confused by the comparison
nrates > __arraycount(lq->rs_table), coverity erroneously binds nrates
to the range [0..16), whereas it is really bound to [0..15) by
ni->ni_rates.rs_nrates (IEEE80211_RATE_MAXSIZE). We add an explicit comparison
to unconfuse it that will always succeed.

Revert previous change.

CID 1271021: Overrunning array "in->in_ridx" of 15 bytes at byte offset 15 using index "i" (which evaluates to 15).

http://mail-index.netbsd.org/coverity-updates/2015/02/21/msg000115.html

Make the iwm(4) hardware receive filter pass multicast traffic from OpenBSD.
Fixes inet6 autoconf, for example.