vax/dhu(4): Convert to ttylock/ttyunlock.

Add d_discard to all struct cdevsw instances I could find.

All have been set to "nodiscard"; some should get a real implementation.

branches: 1.56.14; 1.56.28;
Rename ttymalloc() to tty_alloc(), and ttyfree() to tty_free() for
consistency. Remove some unnecessary malloc.h inclusions as well.

branches: 1.55.18; 1.55.24;
fix some missing pieces of device/softc split

branches: 1.54.2;
PR kern/38763 "hanging in tty" bugfix breaks the build for vax (and
probably others)

Properly fix the "hanging in tty" bug that was worked around with cv_wakeup()
some time again.

branches: 1.52.2; 1.52.4; 1.52.6;
Rototill the vax code. Switch to devicet/PRIV_ALLOC. Cleanup vax autoconf
code. Move to prototype definitions. staticfy, constify, avoid casting.
Use device_* accessors.

branches: 1.51.10; 1.51.14;
Interrupt handling changes, in discussion since February:

- Reduce available SPL levels for hardware devices to none, vm, sched, high.
- Acquire kernel_lock only for interrupts at IPL_VM.
- Implement threaded soft interrupts.

- Factor out too many copies of the same bit of tty code.
- Fix another tty signalling/wakeup problem.

Merge tty changes from the vmlocking branch.

branches: 1.48.2;
machine/{bus,cpu,intr}.h -> sys/{bus,cpu,intr}.h

branches: 1.47.2; 1.47.14; 1.47.16; 1.47.20;
Kill caddr_t; there will be some MI fallout, but it will be fixed shortly.

branches: 1.46.4;
Adapt MD code to KAUTH_DEVICE_TTY_OPEN, batch #2 from Matt Fleming, thanks!

Also, add forgotten splx() calls in some places.

branches: 1.45.4; 1.45.6;
- Use the LWP cached credentials where sane.
- Minor cosmetic changes.

- include kauth.h for kauth_authorize_generic.
- wrap a long line.

integrate kauth.

Use device_private().

branches: 1.41.4; 1.41.6; 1.41.8; 1.41.10; 1.41.12;
make this compile again (from Johnny Billquist)

merge ktrace-lwp.

Change the driver open function's conditional for overriding exclusive tty
use from checking the proc's uid to suser(9), and account for the use of
privileges. Noted by David Holland in PR kern/31126.

Also change this to use the proc argument instead of curproc.

branches: 1.38.4;
White space nits.

de-__P

branches: 1.36.4; 1.36.6;
Remove UCB ad clause in Hugh Graham's license.
Ok'd by Hugh Graham and board.

Constify the speedtab arrays

Move UCB-licensed code from 4-clause to 3-clause licence.

Patches provided by Joel Baker in PR 22364, verified by myself.

branches: 1.33.2;
A bunch of fixes to make the DHU driver much more efficient,
from Hugh Graham (hugh@openbsd.org).

merge kqueue branch into -current

kqueue provides a stateful and efficient event notification framework
currently supported events include socket, file, directory, fifo,
pipe, tty and device changes, and monitoring of processes and signals

kqueue is supported by all writable filesystems in NetBSD tree
(with exception of Coda) and all device drivers supporting poll(2)

based on work done by Jonathan Lemon for FreeBSD
initial NetBSD port done by Luke Mewburn and Jason Thorpe

Add trailing ; to CFATTACH_DECL.

Use CFATTACH_DECL().

Declare all cfattach structures const.

Don't include <sys/map.h>.

Merge the gehenna-devsw branch into the trunk.

This merge changes the device switch tables from static array to
dynamically generated by config(8).

- All device switches is defined as a constant structure in device drivers.

- The new grammer ``device-major'' is introduced to ``files''.

device-major <prefix> char <num> [block <num>] [<rules>]

- All device major numbers must be listed up in port dependent majors.<arch>
by using this grammer.

- Added the new naming convention.
The name of the device switch must be <prefix>_[bc]devsw for auto-generation
of device switch tables.

- The backward compatibility of loading block/character device
switch by LKM framework is broken. This is necessary to convert
from block/character device major to device name in runtime and vice versa.

- The restriction to assign device major by LKM is completely removed.
We don't need to reserve LKM entries for dynamic loading of device switch.

- In compile time, device major numbers list is packed into the kernel and
the LKM framework will refer it to assign device major number dynamically.

branches: 1.26.4;
Convert ioctl code to use EPASSTHROUGH instead of -1 or ENOTTY for
indicating an unhandled "command". ERESTART is -1, which can lead to
confusion. ERESTART has been moved to -3 and EPASSTHROUGH has been
placed at -4. No ioctl code should now return -1 anywhere. The
ioctl() system call is now properly restartable.

add RCSIDs

branches: 1.24.2; 1.24.4;
Use cdev_decl() instead of local prototyping.

Add `l_poll' to `struct linesw' and provide an xxxpoll() entry point
in each tty driver to indirect through it.

This allows tty line-disciplines to handle poll(2) system calls.

Remove unnecessary test of tp->t_linesw against NULL; they are results
of confusion while correcting compilation error after t_line is
replaced with t_linesw.

branches: 1.21.2;
Adapt to the new line discipline scheme.

Adjust to new evcnt stuff.

More event counter stuff. Plus a little cruft cleanup while doing it.

branches: 1.18.2;
Remove register declarations.

Revamp the VAX interrupt handling code. Make it compatible with all the
other ports. Clean up some things.

branches: 1.16.2; 1.16.4;
Fix map register/DMA wait queues.
Still to do:
BDP handling. Currently missing.
Ubareset's won't work at all.

The DHU code changed to use bus_space.
XXX - still need to use bus_dma.

Copied from ../../arch/vax/uba/dhu.c,v

Allocate (almost) all interrupt vectors dynamically. Simplifies much
work when adding support for new machines and devices.

t_wopen fixes.

Remove __BROKEN_INDIRECT_CONFIG and change all drivers to use
config_search instead.

Update for changes to config.

Update due to changes to uba/mscp functions.

backout previous kprintf change

printf -> kprintf, sprintf -> ksprintf

tty stop functions really should return void, not int, and certainly not both.

Update from Ken Wellsch.

Added prototypes to everything. Made all files compile with -Wall.

Ubareset are now possible, devices that wants it now set it up
during autoconfig. ifubareset (if_reset) no longer used, actually
it's just a normal ubareset and is now handled like that.

Update from Ken Wellsch: The DHU driver can now handle DMA transfers.

Unibus/Q22 bus device driver for DHU-11 serial card.
Written by Ken Wellsch. (kcwellsc@math.uwaterloo.ca)

A bunch of fixes to make the DHU driver much more efficient,
from Hugh Graham (hugh@openbsd.org).

branches: 1.4.2;
The DHU code changed to use bus_space.
XXX - still need to use bus_dma.

Copied from ../../arch/vax/uba/dhureg.h,v

Update from Ken Wellsch: The DHU driver can now handle DMA transfers.

Unibus/Q22 bus device driver for DHU-11 serial card.
Written by Ken Wellsch. (kcwellsc@math.uwaterloo.ca)

Add d_discard to all struct cdevsw instances I could find.

All have been set to "nodiscard"; some should get a real implementation.

branches: 1.48.2;
Change (mostly mechanically) every cdevsw/bdevsw I can find to use
designated initializers.

I have not built every extant kernel so I have probably broken at
least one build; however I've also found and fixed some wrong
cdevsw/bdevsw entries so even if so I think we come out ahead.

branches: 1.47.2;
split device_t/softc for all remaining drivers.
replace "struct device *" with "device_t".
use device_xname(), device_unit(), etc.

branches: 1.46.4; 1.46.14;
Rename ttymalloc() to tty_alloc(), and ttyfree() to tty_free() for
consistency. Remove some unnecessary malloc.h inclusions as well.

branches: 1.45.4; 1.45.6;
struct device * -> device_t, no functional changes intended.

struct cfdata * -> cfdata_t, no functional changes intended.

branches: 1.43.10;
use device_lookup_private() to get the softc -- while device/softc
are not split yet, it is better than a typecast

branches: 1.42.2;
PR kern/38763 "hanging in tty" bugfix breaks the build for vax (and
probably others)

branches: 1.41.2;
Remove clause 3 and 4 from TNF licenses

branches: 1.40.2; 1.40.4;
use aprint_*_dev and device_xname

branches: 1.39.14;
- Factor out too many copies of the same bit of tty code.
- Fix another tty signalling/wakeup problem.

branches: 1.38.2;
machine/{bus,cpu,intr}.h -> sys/{bus,cpu,intr}.h

branches: 1.37.2; 1.37.14; 1.37.16; 1.37.20;
Kill caddr_t; there will be some MI fallout, but it will be fixed shortly.

branches: 1.36.4;
Adapt MD code to KAUTH_DEVICE_TTY_OPEN, batch #2 from Matt Fleming, thanks!

Also, add forgotten splx() calls in some places.

branches: 1.35.4; 1.35.6;
- Use the LWP cached credentials where sane.
- Minor cosmetic changes.

- include kauth.h for kauth_authorize_generic.
- wrap a long line.

integrate kauth.

Use device_private().

branches: 1.31.4; 1.31.6; 1.31.8; 1.31.10; 1.31.12;
merge ktrace-lwp.

Change the driver open function's conditional for overriding exclusive tty
use from checking the proc's uid to suser(9), and account for the use of
privileges. Noted by David Holland in PR kern/31126.

branches: 1.29.4;
White space nits.

branches: 1.28.8; 1.28.10;
Move my copyright notice to the 3-clause UCB licence.

Move UCB-licensed code from 4-clause to 3-clause licence.

Patches provided by Joel Baker in PR 22364, verified by myself.

branches: 1.26.6;
merge kqueue branch into -current

kqueue provides a stateful and efficient event notification framework
currently supported events include socket, file, directory, fifo,
pipe, tty and device changes, and monitoring of processes and signals

kqueue is supported by all writable filesystems in NetBSD tree
(with exception of Coda) and all device drivers supporting poll(2)

based on work done by Jonathan Lemon for FreeBSD
initial NetBSD port done by Luke Mewburn and Jason Thorpe

Add trailing ; to CFATTACH_DECL.

Use CFATTACH_DECL().

Declare all cfattach structures const.

Don't include <sys/map.h>.

Merge the gehenna-devsw branch into the trunk.

This merge changes the device switch tables from static array to
dynamically generated by config(8).

- All device switches is defined as a constant structure in device drivers.

- The new grammer ``device-major'' is introduced to ``files''.

device-major <prefix> char <num> [block <num>] [<rules>]

- All device major numbers must be listed up in port dependent majors.<arch>
by using this grammer.

- Added the new naming convention.
The name of the device switch must be <prefix>_[bc]devsw for auto-generation
of device switch tables.

- The backward compatibility of loading block/character device
switch by LKM framework is broken. This is necessary to convert
from block/character device major to device name in runtime and vice versa.

- The restriction to assign device major by LKM is completely removed.
We don't need to reserve LKM entries for dynamic loading of device switch.

- In compile time, device major numbers list is packed into the kernel and
the LKM framework will refer it to assign device major number dynamically.

branches: 1.20.4;
Convert ioctl code to use EPASSTHROUGH instead of -1 or ENOTTY for
indicating an unhandled "command". ERESTART is -1, which can lead to
confusion. ERESTART has been moved to -3 and EPASSTHROUGH has been
placed at -4. No ioctl code should now return -1 anywhere. The
ioctl() system call is now properly restartable.

add RCSIDs

branches: 1.18.2; 1.18.4;
Use cdev_decl() instead of local prototyping.

Add `l_poll' to `struct linesw' and provide an xxxpoll() entry point
in each tty driver to indirect through it.

This allows tty line-disciplines to handle poll(2) system calls.

Remove unnecessary test of tp->t_linesw against NULL; they are results
of confusion while correcting compilation error after t_line is
replaced with t_linesw.

branches: 1.15.2;
Adapt to the new line discipline scheme.

Adjust to new evcnt stuff.

More event counter stuff. Plus a little cruft cleanup while doing it.

branches: 1.12.2;
Remove register declarations.

Revamp the VAX interrupt handling code. Make it compatible with all the
other ports. Clean up some things.

branches: 1.10.2; 1.10.4;
Fix map register/DMA wait queues.
Still to do:
BDP handling. Currently missing.
Ubareset's won't work at all.

Moved from ../dec/qbus/dl.c,v

Directory called qbus instead of uba, per request from Matt/Jason/...
(More describing name actually)

DL-11 driver bus'ified. UNTESTED.

Allocate (almost) all interrupt vectors dynamically. Simplifies much
work when adding support for new machines and devices.

t_wopen fixes.

Remove __BROKEN_INDIRECT_CONFIG and change all drivers to use
config_search instead.

Update for changes to config.

Copyright assigned to The NetBSD Foundation.

branches: 1.1.8;
Support for DL11 (-compatible) async ctlr.
Written by Ben Harris (bjh@mail.dotcom.fr).

merge ktrace-lwp.

White space nits.

branches: 1.4.28; 1.4.36; 1.4.38;
Release dlreg.h into the public domain. It's hardly a great creative
work, after all.

branches: 1.3.2; 1.3.4;
Moved from ../dec/qbus/dlreg.h,v

DL-11 driver bus'ified. UNTESTED.

Support for DL11 (-compatible) async ctlr.
Written by Ben Harris (bjh@mail.dotcom.fr).

Move the dz driver to dev/dec, since it can be shared with pmax.

add RCSIDs

branches: 1.28.2; 1.28.4;
Add `l_poll' to `struct linesw' and provide an xxxpoll() entry point
in each tty driver to indirect through it.

This allows tty line-disciplines to handle poll(2) system calls.

Remove unnecessary test of tp->t_linesw against NULL; they are results
of confusion while correcting compilation error after t_line is
replaced with t_linesw.

branches: 1.26.2;
Changes for dzkbd/dzms. This is not yet finished, but at least the
kernel is useable.

Adapt to the new line discipline scheme.

Adjust to new evcnt stuff.

Start couting interrupts.

branches: 1.22.2;
Change to use new ubareset() functionality.

Remove register declarations.

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.

Revamp the VAX interrupt handling code. Make it compatible with all the
other ports. Clean up some things.

branches: 1.18.2; 1.18.4;
Fix map register/DMA wait queues.
Still to do:
BDP handling. Currently missing.
Ubareset's won't work at all.

Moved from ../dec/qbus/dz.c,v

Directory called qbus instead of uba, per request from Matt/Jason/...
(More describing name actually)

DZ-11 routines bus'ified. Small fixes to uba routines.

Change vsbus to take a locator for all devices.
Figure out interrupt vector/mask on vsbus by forcing all devices to
generate interrupts in the match routine, like on unibus.

EGCS fixes.

Kickoff xmit interrupts, needed on chip DC7201.

defopt DDB.

Add catch function for keyboard.

Fixed bug with lost characters in DDB output.

Support for DDB on DZ console.

Split the DZ11 driver into two parts; can now use the same driver on
both unibus/qbus devices and vaxstations.

Remove __BROKEN_INDIRECT_CONFIG and change all drivers to use
config_search instead.

Update for changes to config.

backout previous kprintf change

printf -> kprintf, sprintf -> ksprintf

tty stop functions really should return void, not int, and certainly not both.

Drivers for DZ11/DZV11/DZQ11 serial cards.
Written (mostly) by Ken Wellsch.

Update copyright notice for Ludd (remove clause 3 & 4).

branches: 1.28.48; 1.28.68;
Switch dz driver to PRIVALLOC and device_* accessors.

Rototill the vax code. Switch to devicet/PRIV_ALLOC. Cleanup vax autoconf
code. Move to prototype definitions. staticfy, constify, avoid casting.
Use device_* accessors.

branches: 1.26.12; 1.26.16;
machine/{bus,cpu,intr}.h -> sys/{bus,cpu,intr}.h

branches: 1.25.18; 1.25.32; 1.25.34; 1.25.38;
integrate kauth.

Use device_private().

branches: 1.23.4; 1.23.6; 1.23.8; 1.23.10; 1.23.12;
merge ktrace-lwp.

branches: 1.22.4;
White space nits.

de-__P

branches: 1.20.8; 1.20.10;
Add write barriers (a no-op on vax).

branches: 1.19.6;
Add trailing ; to CFATTACH_DECL.

Use CFATTACH_DECL().

Declare all cfattach structures const.

Update for dzattach() change.

Don't include <sys/map.h>.

Move the dz driver to dev/dec, since it can be shared with pmax.

add RCSIDs

branches: 1.12.4; 1.12.6;
Adjust to new evcnt stuff.

More event counter stuff. Plus a little cruft cleanup while doing it.

branches: 1.10.2;
Change to use new ubareset() functionality.

Remove register declarations.

Revamp the VAX interrupt handling code. Make it compatible with all the
other ports. Clean up some things.

branches: 1.7.2; 1.7.4;
Fix map register/DMA wait queues.
Still to do:
BDP handling. Currently missing.
Ubareset's won't work at all.

Moved from ../dec/qbus/dz_uba.c,v

Directory called qbus instead of uba, per request from Matt/Jason/...
(More describing name actually)

DZ-11 routines bus'ified. Small fixes to uba routines.

Avoid warning if !QBA.

Allocate (almost) all interrupt vectors dynamically. Simplifies much
work when adding support for new machines and devices.

Split the DZ11 driver into two parts; can now use the same driver on
both unibus/qbus devices and vaxstations.

Move the dz driver to dev/dec, since it can be shared with pmax.

branches: 1.4.2; 1.4.16; 1.4.18;
Moved from ../dec/qbus/dzreg.h,v

DZ-11 routines bus'ified. Small fixes to uba routines.

Split the DZ11 driver into two parts; can now use the same driver on
both unibus/qbus devices and vaxstations.

Drivers for DZ11/DZV11/DZQ11 serial cards.
Written (mostly) by Ken Wellsch.

Move the dz driver to dev/dec, since it can be shared with pmax.

branches: 1.10.2; 1.10.4;
Changes for dzkbd/dzms. This is not yet finished, but at least the
kernel is useable.

Adjust to new evcnt stuff.

Start couting interrupts.

branches: 1.7.2;
Change to use new ubareset() functionality.

Revamp the VAX interrupt handling code. Make it compatible with all the
other ports. Clean up some things.

branches: 1.5.2; 1.5.4;
Moved from ../dec/qbus/dzvar.h,v

DZ-11 routines bus'ified. Small fixes to uba routines.

Change vsbus to take a locator for all devices.
Figure out interrupt vector/mask on vsbus by forcing all devices to
generate interrupts in the match routine, like on unibus.

Add catch function for keyboard.

Split the DZ11 driver into two parts; can now use the same driver on
both unibus/qbus devices and vaxstations.

add missing ifuba attribute for qe(4), from Josh Moyer on port-vax.

qbus/qt(4): Add missing dependency to `ifuba`

branches: 1.17.130;
Rototill the vax code. Switch to devicet/PRIV_ALLOC. Cleanup vax autoconf
code. Move to prototype definitions. staticfy, constify, avoid casting.
Use device_* accessors.

branches: 1.16.46; 1.16.66; 1.16.70;
merge ktrace-lwp.

branches: 1.15.16;
The qt driver must be probed first, so that the qe driver is skipped
if it is a qt card.

Add qt.

branches: 1.13.2;
Driver for RX01/02 floppies, written by Jochen Kunz.

Include DMF32/DMZ32 files. The actual device drivers will be checked in when
the licensing of the 4.4BSD encumbered files are solved.

branches: 1.11.8;
Move the dz driver to dev/dec, since it can be shared with pmax.

branches: 1.10.2;
be more specific about which interlan controller the qbus/if_il supports

Add il, dmc, ts and if_uba.

branches: 1.8.2;
Changes for dzkbd/dzms. This is not yet finished, but at least the
kernel is useable.

Add the famous "rl" disk.

branches: 1.6.2; 1.6.4;
add qe and qd.

Fix map register/DMA wait queues.
Still to do:
BDP handling. Currently missing.
Ubareset's won't work at all.

Moved from ../dec/qbus/files.uba,v

Directory called qbus instead of uba, per request from Matt/Jason/...
(More describing name actually)

DZ-11 routines bus'ified. Small fixes to uba routines.

First step towards MI Unibus/Q22 bus code.

fix typos in word "otherwise".

branches: 1.35.10;
Adopt <net/if_stats.h>.

branches: 1.34.2; 1.34.10;
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.

It's not required to include net/bpfdesc.h. Remove it.

branches: 1.32.2; 1.32.8;
Update copyright notice for Ludd (remove clause 3 & 4).

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: 1.30.2;
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!

branches: 1.29.18; 1.29.36;
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.

branches: 1.28.2; 1.28.4;
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.

KNF

Remove extra whitespace added by a stupid tool.
XXX: more in src/sys/arch

bcopy -> memcpy

branches: 1.24.2;
replace bitmask_snprintf(9) with snprintb(3)

branches: 1.23.4; 1.23.12;
Rototill the vax code. Switch to devicet/PRIV_ALLOC. Cleanup vax autoconf
code. Move to prototype definitions. staticfy, constify, avoid casting.
Use device_* accessors.

branches: 1.22.12; 1.22.16;
machine/{bus,cpu,intr}.h -> sys/{bus,cpu,intr}.h

branches: 1.21.2; 1.21.14; 1.21.16; 1.21.20;
Kill caddr_t; there will be some MI fallout, but it will be fixed shortly.

branches: 1.20.14;
Use device_private().

Use device_parent().

branches: 1.18.4; 1.18.6; 1.18.8; 1.18.10; 1.18.12;
merge ktrace-lwp.

branches: 1.17.2;
Fixes due to cast-qual addition.

White space nits.

branches: 1.15.8; 1.15.10;
Move UCB-licensed code from 4-clause to 3-clause licence.

Patches provided by Joel Baker in PR 22364, verified by myself.

branches: 1.14.6;
Add trailing ; to CFATTACH_DECL.

Use CFATTACH_DECL().

Declare all cfattach structures const.

add RCSIDs

branches: 1.10.2;
Use if_uba + ether_ioctl routines. Removed ~150 lines of duplicated code.

Use ubmemalloc() to get mapped uba memory.
Only copy mbufs if there are more than two in a mbuf chain (on transmit).
This squeezed another 5% out of the DEUNA. (now closing up to 100 K/s :-)

branches: 1.8.2;
ALTQ'ify.

branches: 1.7.2;
Move bpfattach()/bpfdetach() calls into ether_ifattach()/ether_ifdetach().

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).

Fix a transmit bug and get a little more out from the receive logic.

Adjust to new evcnt stuff.

More event counter stuff. Plus a little cruft cleanup while doing it.

Some updates to the DEUNA driver:
- Change the transmit logic to send from a pre-allocated buffer instead
of doing DMA out of the mbufs directly. Timing shows that it takes
4 times longer to map in the mbufs via the sgmap routines instead of
just copying the data! This must be fixed sometine in the future.
- A bunch of other minor fixes.

branches: 1.1.2;
Rewritten (most of) the DEUNA/DELUA driver. No more data copy, packet
header fiddling and forget driver support for trailers.

merge ktrace-lwp.

Move UCB-licensed code from 4-clause to 3-clause licence.

Patches provided by Joel Baker in PR 22364, verified by myself.

branches: 1.2.4; 1.2.28;
Some updates to the DEUNA driver:
- Change the transmit logic to send from a pre-allocated buffer instead
of doing DMA out of the mbufs directly. Timing shows that it takes
4 times longer to map in the mbufs via the sgmap routines instead of
just copying the data! This must be fixed sometine in the future.
- A bunch of other minor fixes.

branches: 1.1.2;
Rewritten (most of) the DEUNA/DELUA driver. No more data copy, packet
header fiddling and forget driver support for trailers.

sys: Drop redundant NULL check before m_freem(9)

m_freem(9) safely has accepted NULL argument at least since 4.2BSD:
https://www.tuhs.org/cgi-bin/utree.pl?file=4.2BSD/usr/src/sys/sys/uipc_mbuf.c

Compile-tested on amd64/ALL.

Suggested by knakahara@

branches: 1.29.8; 1.29.10;
fix various typos, mainly in comments.

Adopt <net/if_stats.h>

Adopt <net/if_stats.h>

branches: 1.26.16; 1.26.24;
Apply pserialize to some iterations of IP address lists

branches: 1.25.2;
Switch the address list of intefaces to pslist(9)

As usual, we leave the old list to avoid breaking kvm(3) users.

IFQ_ENQUEUE refactor (3/3) : eliminate pktattr argument from IFQ_ENQUEUE caller

branches: 1.23.4;
- Implement pktqueue interface for lockless IP input queue.
- Replace ipintrq and ip6intrq with the pktqueue mechanism.
- Eliminate kernel-lock from ipintr() and ip6intr().
- Some preparation work to push softnet_lock out of ipintr().

Discussed on tech-net.

Save a NETISR_* value in a variable and call schednetisr() after enqueue
a packet for readability and future modification.

branches: 1.21.2; 1.21.10;
split device_t/softc for all remaining drivers.
replace "struct device *" with "device_t".
use device_xname(), device_unit(), etc.

branches: 1.20.12; 1.20.22;
struct device * -> device_t, no functional changes intended.

struct cfdata * -> cfdata_t, no functional changes intended.

branches: 1.18.2;
replace bitmask_snprintf(9) with snprintb(3)

*** Summary ***

When a link-layer address changes (e.g., ifconfig ex0 link
02:de:ad:be:ef:02 active), send a gratuitous ARP and/or a Neighbor
Advertisement to update the network-/link-layer address bindings
on our LAN peers.

Refuse a change of ethernet address to the address 00:00:00:00:00:00
or to any multicast/broadcast address. (Thanks matt@.)

Reorder ifnet ioctl operations so that driver ioctls may inherit
the functions of their "class"---ether_ioctl(), fddi_ioctl(), et
cetera---and the class ioctls may inherit from the generic ioctl,
ifioctl_common(), but both driver- and class-ioctls may override
the generic behavior. Make network drivers share more code.

Distinguish a "factory" link-layer address from others for the
purposes of both protecting that address from deletion and computing
EUI64.

Return consistent, appropriate error codes from network drivers.

Improve readability. KNF.

*** Details ***

In if_attach(), always initialize the interface ioctl routine,
ifnet->if_ioctl, if the driver has not already initialized it.
Delete if_ioctl == NULL tests everywhere else, because it cannot
happen.

In the ioctl routines of network interfaces, inherit common ioctl
behaviors by calling either ifioctl_common() or whichever ioctl
routine is appropriate for the class of interface---e.g., ether_ioctl()
for ethernets.

Stop (ab)using SIOCSIFADDR and start to use SIOCINITIFADDR. In
the user->kernel interface, SIOCSIFADDR's argument was an ifreq,
but on the protocol->ifnet interface, SIOCSIFADDR's argument was
an ifaddr. That was confusing, and it would work against me as I
make it possible for a network interface to overload most ioctls.
On the protocol->ifnet interface, replace SIOCSIFADDR with
SIOCINITIFADDR. In ifioctl(), return EPERM if userland tries to
invoke SIOCINITIFADDR.

In ifioctl(), give the interface the first shot at handling most
interface ioctls, and give the protocol the second shot, instead
of the other way around. Finally, let compatibility code (COMPAT_OSOCK)
take a shot.

Pull device initialization out of switch statements under
SIOCINITIFADDR. For example, pull ..._init() out of any switch
statement that looks like this:

switch (...->sa_family) {
case ...:
..._init();
...
break;
...
default:
..._init();
...
break;
}

Rewrite many if-else clauses that handle all permutations of IFF_UP
and IFF_RUNNING to use a switch statement,

switch (x & (IFF_UP|IFF_RUNNING)) {
case 0:
...
break;
case IFF_RUNNING:
...
break;
case IFF_UP:
...
break;
case IFF_UP|IFF_RUNNING:
...
break;
}

unifdef lots of code containing #ifdef FreeBSD, #ifdef NetBSD, and
#ifdef SIOCSIFMTU, especially in fwip(4) and in ndis(4).

In ipw(4), remove an if_set_sadl() call that is out of place.

In nfe(4), reuse the jumbo MTU logic in ether_ioctl().

Let ethernets register a callback for setting h/w state such as
promiscuous mode and the multicast filter in accord with a change
in the if_flags: ether_set_ifflags_cb() registers a callback that
returns ENETRESET if the caller should reset the ethernet by calling
if_init(), 0 on success, != 0 on failure. Pull common code from
ex(4), gem(4), nfe(4), sip(4), tlp(4), vge(4) into ether_ioctl(),
and register if_flags callbacks for those drivers.

Return ENOTTY instead of EINVAL for inappropriate ioctls. In
zyd(4), use ENXIO instead of ENOTTY to indicate that the device is
not any longer attached.

Add to if_set_sadl() a boolean 'factory' argument that indicates
whether a link-layer address was assigned by the factory or some
other source. In a comment, recommend using the factory address
for generating an EUI64, and update in6_get_hw_ifid() to prefer a
factory address to any other link-layer address.

Add a routing message, RTM_LLINFO_UPD, that tells protocols to
update the binding of network-layer addresses to link-layer addresses.
Implement this message in IPv4 and IPv6 by sending a gratuitous
ARP or a neighbor advertisement, respectively. Generate RTM_LLINFO_UPD
messages on a change of an interface's link-layer address.

In ether_ioctl(), do not let SIOCALIFADDR set a link-layer address
that is broadcast/multicast or equal to 00:00:00:00:00:00.

Make ether_ioctl() call ifioctl_common() to handle ioctls that it
does not understand.

In gif(4), initialize if_softc and use it, instead of assuming that
the gif_softc and ifp overlap.

Let ifioctl_common() handle SIOCGIFADDR.

Sprinkle rtcache_invariants(), which checks on DIAGNOSTIC kernels
that certain invariants on a struct route are satisfied.

In agr(4), rewrite agr_ioctl_filter() to be a bit more explicit
about the ioctls that we do not allow on an agr(4) member interface.

bzero -> memset. Delete unnecessary casts to void *. Use
sockaddr_in_init() and sockaddr_in6_init(). Compare pointers with
NULL instead of "testing truth". Replace some instances of (type
*)0 with NULL. Change some K&R prototypes to ANSI C, and join
lines.

branches: 1.16.4; 1.16.10; 1.16.12;
use aprint_*_dev and device_xname

branches: 1.15.16;
machine/{bus,cpu,intr}.h -> sys/{bus,cpu,intr}.h

branches: 1.14.2; 1.14.14; 1.14.16; 1.14.20;
Kill caddr_t; there will be some MI fallout, but it will be fixed shortly.

branches: 1.13.14;
Use device_private().

Use device_cfdata().

Use device_parent().

branches: 1.10.4; 1.10.6; 1.10.8; 1.10.10; 1.10.12;
merge ktrace-lwp.

branches: 1.9.4;
White space nits.

branches: 1.8.2;
Add IFNET_FOREACH and IFADDR_FOREACH macros and start using them.

branches: 1.7.8;
Move UCB-licensed code from 4-clause to 3-clause licence.

Patches provided by Joel Baker in PR 22364, verified by myself.

branches: 1.6.6;
Add trailing ; to CFATTACH_DECL.

Use CFATTACH_DECL().

Declare all cfattach structures const.

bring in latest ALTQ from kjc. ALTQify some of the drivers.

add RCSIDs

branches: 1.1.2; 1.1.4;
Driver for the DMC-11/DMR-11 DDCMP interface, (untested) from 4.4BSD.

merge ktrace-lwp.

White space nits.

branches: 1.2.8; 1.2.10;
Move UCB-licensed code from 4-clause to 3-clause licence.

Patches provided by Joel Baker in PR 22364, verified by myself.

branches: 1.1.2; 1.1.24;
Driver for the DMC-11/DMR-11 DDCMP interface, (untested) from 4.4BSD.

fix typos in word "otherwise".

branches: 1.36.12;
if_percpuq(9) automatically increments if_ipackets, so don't increment it in
the driver itself to prevent double count.

branches: 1.35.2;
Use ETHER_LOCK()/ETHER_UNLOCK() for all ethernet drivers to protect ec_multi*.

Make if_il.c compilable again.

Revert previous.

Make it compilable. Not tested.

Whitespace fix (mainly tabify).

-No functional change:
- KNF
- u_int*_t -> uint*_t.

branches: 1.29.10; 1.29.18;
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!

add missing splx() found by brainy

branches: 1.27.4;
Add missing braces. From Henning Petersen in PR 48846.

branches: 1.26.10;
split device_t/softc for all remaining drivers.
replace "struct device *" with "device_t".
use device_xname(), device_unit(), etc.

branches: 1.25.12; 1.25.22;
Fix incomplete line (deleto?), found by cppcheck and reported by
Henning Petersen in PR 42719.

struct device * -> device_t, no functional changes intended.

struct cfdata * -> cfdata_t, no functional changes intended.

Remove extra whitespace added by a stupid tool.
XXX: more in src/sys/arch

bcopy -> memcpy

branches: 1.20.2;
replace bitmask_snprintf(9) with snprintb(3)

branches: 1.19.4; 1.19.12;
use aprint_*_dev and device_xname

branches: 1.18.16;
machine/{bus,cpu,intr}.h -> sys/{bus,cpu,intr}.h

branches: 1.17.2; 1.17.6;
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: 1.16.2; 1.16.10; 1.16.14;
Kill caddr_t; there will be some MI fallout, but it will be fixed shortly.

branches: 1.15.8;
remove more vestiges of CCITT, LLC, HDLC, NS, and NSIP.

Use device_private().

Use device_parent().

branches: 1.12.4; 1.12.6; 1.12.8; 1.12.10; 1.12.12;
merge ktrace-lwp.

branches: 1.11.4;
White space nits.

branches: 1.10.8; 1.10.10;
Move UCB-licensed code from 4-clause to 3-clause licence.

Patches provided by Joel Baker in PR 22364, verified by myself.

branches: 1.9.2;
Ragge said this chip should autopad, so just remove the check.

Add trailing ; to CFATTACH_DECL.

Use CFATTACH_DECL().

Declare all cfattach structures const.

bcmp -> memcmp.

branches: 1.4.6; 1.4.8;
bring in latest ALTQ from kjc. ALTQify some of the drivers.

#endif ... -> #endif /* ... */

add RCSIDs

branches: 1.1.2; 1.1.4;
Driver for the Interlan NI1010 Ethernet Controller, from 4.4BSD.

merge ktrace-lwp.

Move UCB-licensed code from 4-clause to 3-clause licence.

Patches provided by Joel Baker in PR 22364, verified by myself.

branches: 1.1.2; 1.1.24;
Driver for the Interlan NI1010 Ethernet Controller, from 4.4BSD.

merge ktrace-lwp.

White space nits.

branches: 1.3.8; 1.3.10;
Move UCB-licensed code from 4-clause to 3-clause licence.

Patches provided by Joel Baker in PR 22364, verified by myself.

branches: 1.2.2;
DMA, not dma nor Dma.

branches: 1.1.2;
Driver for the Interlan NI1010 Ethernet Controller, from 4.4BSD.

sys: Drop redundant NULL check before m_freem(9)

m_freem(9) safely has accepted NULL argument at least since 4.2BSD:
https://www.tuhs.org/cgi-bin/utree.pl?file=4.2BSD/usr/src/sys/sys/uipc_mbuf.c

Compile-tested on amd64/ALL.

Suggested by knakahara@

branches: 1.82.2;
vax/qe(4): supply an ipf->if_init() so that if_init() doesn't crash.

convert the existing qeinit() to one compatible with if_init.

should fix PR#58068.

branches: 1.81.2; 1.81.28;
Use ETHER_LOCK()/ETHER_UNLOCK() for all ethernet drivers to protect ec_multi*.

-No functional change:
- KNF
- u_int*_t -> uint*_t.

branches: 1.79.2;
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.

It's not required to include net/bpfdesc.h. Remove it.

branches: 1.77.2; 1.77.8;
Update copyright notice for Ludd (remove clause 3 & 4).

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: 1.75.2;
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!

branches: 1.73.2;
Avoid leak in error branch, noted by maxv@, compile-tested for vax.

branches: 1.72.4;
Fix dmamap cleanup in an error path

branches: 1.71.8; 1.71.14; 1.71.18; 1.71.20; 1.71.22; 1.71.28;
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.

branches: 1.70.2; 1.70.4;
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.

bzero -> memset

branches: 1.68.4;
*** Summary ***

When a link-layer address changes (e.g., ifconfig ex0 link
02:de:ad:be:ef:02 active), send a gratuitous ARP and/or a Neighbor
Advertisement to update the network-/link-layer address bindings
on our LAN peers.

Refuse a change of ethernet address to the address 00:00:00:00:00:00
or to any multicast/broadcast address. (Thanks matt@.)

Reorder ifnet ioctl operations so that driver ioctls may inherit
the functions of their "class"---ether_ioctl(), fddi_ioctl(), et
cetera---and the class ioctls may inherit from the generic ioctl,
ifioctl_common(), but both driver- and class-ioctls may override
the generic behavior. Make network drivers share more code.

Distinguish a "factory" link-layer address from others for the
purposes of both protecting that address from deletion and computing
EUI64.

Return consistent, appropriate error codes from network drivers.

Improve readability. KNF.

*** Details ***

In if_attach(), always initialize the interface ioctl routine,
ifnet->if_ioctl, if the driver has not already initialized it.
Delete if_ioctl == NULL tests everywhere else, because it cannot
happen.

In the ioctl routines of network interfaces, inherit common ioctl
behaviors by calling either ifioctl_common() or whichever ioctl
routine is appropriate for the class of interface---e.g., ether_ioctl()
for ethernets.

Stop (ab)using SIOCSIFADDR and start to use SIOCINITIFADDR. In
the user->kernel interface, SIOCSIFADDR's argument was an ifreq,
but on the protocol->ifnet interface, SIOCSIFADDR's argument was
an ifaddr. That was confusing, and it would work against me as I
make it possible for a network interface to overload most ioctls.
On the protocol->ifnet interface, replace SIOCSIFADDR with
SIOCINITIFADDR. In ifioctl(), return EPERM if userland tries to
invoke SIOCINITIFADDR.

In ifioctl(), give the interface the first shot at handling most
interface ioctls, and give the protocol the second shot, instead
of the other way around. Finally, let compatibility code (COMPAT_OSOCK)
take a shot.

Pull device initialization out of switch statements under
SIOCINITIFADDR. For example, pull ..._init() out of any switch
statement that looks like this:

switch (...->sa_family) {
case ...:
..._init();
...
break;
...
default:
..._init();
...
break;
}

Rewrite many if-else clauses that handle all permutations of IFF_UP
and IFF_RUNNING to use a switch statement,

switch (x & (IFF_UP|IFF_RUNNING)) {
case 0:
...
break;
case IFF_RUNNING:
...
break;
case IFF_UP:
...
break;
case IFF_UP|IFF_RUNNING:
...
break;
}

unifdef lots of code containing #ifdef FreeBSD, #ifdef NetBSD, and
#ifdef SIOCSIFMTU, especially in fwip(4) and in ndis(4).

In ipw(4), remove an if_set_sadl() call that is out of place.

In nfe(4), reuse the jumbo MTU logic in ether_ioctl().

Let ethernets register a callback for setting h/w state such as
promiscuous mode and the multicast filter in accord with a change
in the if_flags: ether_set_ifflags_cb() registers a callback that
returns ENETRESET if the caller should reset the ethernet by calling
if_init(), 0 on success, != 0 on failure. Pull common code from
ex(4), gem(4), nfe(4), sip(4), tlp(4), vge(4) into ether_ioctl(),
and register if_flags callbacks for those drivers.

Return ENOTTY instead of EINVAL for inappropriate ioctls. In
zyd(4), use ENXIO instead of ENOTTY to indicate that the device is
not any longer attached.

Add to if_set_sadl() a boolean 'factory' argument that indicates
whether a link-layer address was assigned by the factory or some
other source. In a comment, recommend using the factory address
for generating an EUI64, and update in6_get_hw_ifid() to prefer a
factory address to any other link-layer address.

Add a routing message, RTM_LLINFO_UPD, that tells protocols to
update the binding of network-layer addresses to link-layer addresses.
Implement this message in IPv4 and IPv6 by sending a gratuitous
ARP or a neighbor advertisement, respectively. Generate RTM_LLINFO_UPD
messages on a change of an interface's link-layer address.

In ether_ioctl(), do not let SIOCALIFADDR set a link-layer address
that is broadcast/multicast or equal to 00:00:00:00:00:00.

Make ether_ioctl() call ifioctl_common() to handle ioctls that it
does not understand.

In gif(4), initialize if_softc and use it, instead of assuming that
the gif_softc and ifp overlap.

Let ifioctl_common() handle SIOCGIFADDR.

Sprinkle rtcache_invariants(), which checks on DIAGNOSTIC kernels
that certain invariants on a struct route are satisfied.

In agr(4), rewrite agr_ioctl_filter() to be a bit more explicit
about the ioctls that we do not allow on an agr(4) member interface.

bzero -> memset. Delete unnecessary casts to void *. Use
sockaddr_in_init() and sockaddr_in6_init(). Compare pointers with
NULL instead of "testing truth". Replace some instances of (type
*)0 with NULL. Change some K&R prototypes to ANSI C, and join
lines.

branches: 1.67.4; 1.67.10; 1.67.12;
Rototill the vax code. Switch to devicet/PRIV_ALLOC. Cleanup vax autoconf
code. Move to prototype definitions. staticfy, constify, avoid casting.
Use device_* accessors.

branches: 1.66.2; 1.66.6;
Apply changes for struct ifnet->if_sadl constify:
http://mail-index.netbsd.org/source-changes/2007/12/20/0033.html

branches: 1.65.4; 1.65.8;
machine/{bus,cpu,intr}.h -> sys/{bus,cpu,intr}.h

branches: 1.64.4;
Remove a now-unused local variable.

Change a bazillion occurrences of code resembling this,

error = (cmd == SIOCADDMULTI) ?
ether_addmulti(ifr, &sc->sc_ec) :
ether_delmulti(ifr, &sc->sc_ec);

if (error == ENETRESET) {

to this,

if ((error = ether_ioctl(ifp, cmd, data)) == ENETRESET) {

which does the same thing.

(A bazillion is a very large number. This seems to make the i386
ALL kernel smaller by 3kB to 4kB.)

Use ifreq_getaddr() twice in es(4).

Whitespace nits.

branches: 1.62.2; 1.62.10; 1.62.14; 1.62.16;
Kill caddr_t; there will be some MI fallout, but it will be fixed shortly.

branches: 1.61.14;
Use device_private().

branches: 1.60.4; 1.60.6; 1.60.8; 1.60.10; 1.60.12;
merge ktrace-lwp.

branches: 1.59.4;
White space nits.

branches: 1.58.4; 1.58.6;
When adding/deleting multicast addresses, only whack the address
filter if the interface is marked RUNNING.

Fixes kern/27678.

branches: 1.57.2; 1.57.6;
Redo previous in a better way:
allocate a static buffer, which is added at the end of short packets
to pad the buffer to ETHER_MIN_LEN - ETHER_CRC_LEN.
While I'm there fix 2 bugs:
in qeinit(), unload the right dmamap if bus_dmamap_load fails
in qestart, don't dmamap_load the mbuf if its len is 0.
Tested on simh-vax.

Zero out the extra space stolen from mbuf to pad packet to ETHER_MIN_LEN
XXX should it be ETHER_MIN_LEN - ETHER_CRC_LEN ?
XXX2 we assume there are enouth space in the mbuf for the padding bytes.
But other places in the code assumes this already.

Add trailing ; to CFATTACH_DECL.

Use CFATTACH_DECL().

Declare all cfattach structures const.

remove trailing \n in panic(). approved perry.

Detection bug: If a packet is received while in the match() routine the stack
would be overwritten, so allocate memory for a full packet via malloc()
instead. Use correct buffer maxlen (must be 2 complemet). Bug reported
and helped to find by Roar Thron�s.

bcmp -> memcmp.

branches: 1.49.8; 1.49.10;
add RCSIDs

branches: 1.48.2;
Use uballoc()/ubmemalloc() for data structure allocation.

splimp -> splnet

branches: 1.46.2;
Do not send looped-back setup packets into the IP stack.
Handle xmit interrupts due to invalid transmit list (``cannot happen'')
Shoot the one that wrote the software for those cards.

ALTQ'ify.

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

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).

Adjust to new evcnt stuff.

More event counter stuff. Plus a little cruft cleanup while doing it.

branches: 1.40.2;
Remove register declarations.

Revamp the VAX interrupt handling code. Make it compatible with all the
other ports. Clean up some things.

branches: 1.38.2;
Add a timeout-routine that checks if transmit logic dies, and reset it
in that case. This is not uncommon when the interface is heavy loaded.

branches: 1.37.2;
Totally rewritten DEQNA device driver, that uses the bus.h interface
instead. This almost doubled the receive speed of the interface :-)

Copied from ../../arch/vax/if/if_qe.c,v

Make this file compile again.

Rework layer 2 protocol input routines. Instead of calling e.g. ether_input()
directly, call the function pointer (*if_input)(ifp, m). The input routine
expects the packet header to be at the head of the packet, and will adjust
as necessary. Privatize the layer 2 input and output routines, allowing
*_ifattach() to set them up as appropriate.

branches: 1.33.2;
the btoc/ctob/btop macros must use CLBYTES, not NBPG. How this should
work in the long run is an open issue; some parts must be reworked
in a MI way.

EGCS fixes.

defopt NS, NSIP.

defopt ISO TPIP.

defopt LLC

defopt CCITT.

defopt INET, NETATALK.

Fix some compile-time warnings. Pointed out by David Brownlee.

Add support for "root on qe".

Remove __BROKEN_INDIRECT_CONFIG and change all drivers to use
config_search instead.

Update for changes to config.

BPF and multicast support added, patches provided by Jonathan Stone
and Ken Wellsch. Fixes PR#2796.

Fix so that these files compile again after recent ARP changes.

New ARP system, supports IPv4 over any hardware link.

Some of the stuff (e.g., rarpd, bootpd, dhcpd etc., libsa) still will
only support Ethernet. Tcpdump itself should be ok, but libpcap needs
lot of work.

For the detailed change history, look at the commit log entries for
the is-newarp branch.

Fix so it works with root over NFS.

branches: 1.18.6;
backout previous kprintf change

printf -> kprintf, sprintf -> ksprintf

Change uba* functions to take pointers, rather than uba unit numbers.

Fixed all (proto)type errors. Fixes PR 2377.

Changed struct ifnet to have a pointer to the softc of the underlying
device and a printable "external name" (name + unit number), thus eliminating
if_name and if_unit. Updated interface to (*if_watchdog)() and (*if_start)()
to take a struct ifnet *, rather than a unit number.

Ubareset are now possible, devices that wants it now set it up
during autoconfig. ifubareset (if_reset) no longer used, actually
it's just a normal ubareset and is now handled like that.

Convert all devices according to the changes to config.

Add prototypes.

Fix #includes.

Fix #includes.

The IST_* and IPL_* constants are not bus-specific; don't treat them as such.
Change splimp -> splnet in Ethernet, ARCnet, and FDDI drivers.

DELUA/DENUA & DELQA/DEQNA Ethernet adapters converted to new config.

merge in changes from 1.1 release branch

Dynamic allocate uba_softc and interrupt vectors.
Allow for more that one uba.

branches: 1.4.2;
Changes due to common network code mods.

DELAY() removed, now common.

Sync with the rest of the world. Say goodbye to Mr. Trailer.

Support for DEQNA/DELQA added.

Fix various typos, mainly in comments:
s/invaid/invalid/
s/instad/instead/
s/wich/with/
s/tranform/transform/
s/tranmist/transmit/
s/tranceiver/transceiver/
s/Tranparent/Transparent/
s/tranlated/translated/
s/tranfer/transfer/
s/tranmissions/transmissions/
s/condtions/conditions/
s/Recient/Recent/

s/discriptor/descriptor/ in comments.

merge ktrace-lwp.

White space nits.

branches: 1.7.8; 1.7.10;
Move UCB-licensed code from 4-clause to 3-clause licence.

Patches provided by Joel Baker in PR 22364, verified by myself.

branches: 1.6.22;
`existent', not `existant'

branches: 1.5.2; 1.5.16;
Totally rewritten DEQNA device driver, that uses the bus.h interface
instead. This almost doubled the receive speed of the interface :-)

Copied from ../../arch/vax/if/if_qereg.h,v

Ungarble the copyright notice.

branches: 1.2.16;
Multicast definitions, forgotten in previous checkins.

Support for DEQNA/DELQA added.

Use ETHER_CRC_LEN where appropriate.

Fix misc style, especially weird indent, per KNF.

branches: 1.27.10;
Remove unnecessary inclusion of <net/netisr.h>.

Set up multicast (input) filter on qt (DELQA-Turbo).

Adopt <net/if_stats.h>.

branches: 1.24.2;
if_percpuq(9) automatically increments if_ipackets, so don't increment it in
the driver itself to prevent double count.

branches: 1.23.2; 1.23.8;
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.

It's not required to include net/bpfdesc.h. Remove it.

branches: 1.21.8; 1.21.14;
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

acknowleg -> acknowledg, proceedure -> procedure.
only comments were changed.

from miod

branches: 1.19.2;
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!

branches: 1.18.18; 1.18.36;
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.

branches: 1.17.2; 1.17.4;
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.

Initialize sc->sc_dev before we use it.

replace bitmask_snprintf(9) with snprintb(3)

branches: 1.14.4; 1.14.12;
Rototill the vax code. Switch to devicet/PRIV_ALLOC. Cleanup vax autoconf
code. Move to prototype definitions. staticfy, constify, avoid casting.
Use device_* accessors.

branches: 1.13.2; 1.13.6;
Add missing sys/device.h.

branches: 1.12.2; 1.12.8;
machine/{bus,cpu,intr}.h -> sys/{bus,cpu,intr}.h

branches: 1.11.2; 1.11.14; 1.11.16; 1.11.20;
Kill caddr_t; there will be some MI fallout, but it will be fixed shortly.

branches: 1.10.8;
remove more vestiges of CCITT, LLC, HDLC, NS, and NSIP.

Use device_private().

Use device_parent().

branches: 1.7.4; 1.7.6; 1.7.8; 1.7.10; 1.7.12;
merge ktrace-lwp.

branches: 1.6.4;
White space nits.

branches: 1.5.4; 1.5.6;
- qtinit(): Don't skip initialization if we're already marked RUNNING.
ether_ioctl() will call (*if_init)() when flags change. Instead, do
what other drivers do and stop the interface in this case before
re-initializing.
- qtioctl(): Now that qtinit() does the right thing, remove the check
for IFF_PROMISC change.

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

Fixes kern/27678.

branches: 1.4.4; 1.4.6;
Add stop routine. Add support for turning the device into promiscous mode.
Force interrupt in probe so that the interrupt vector can be detected.

bus-ify. Make bpf work, but no promiscuous mode yet.

Make the qt driver work under NetBSD. Still to do:
- bus-ify.
- bpf support.
- multicast support.

Driver for the DELQA-PLUS card in Turbo mode, from 2.11BSD, written by
Steven M. Schultz.

Set up multicast (input) filter on qt (DELQA-Turbo).

merge ktrace-lwp.

White space nits.

branches: 1.3.4; 1.3.10; 1.3.12;
bus-ify. Make bpf work, but no promiscuous mode yet.

Make the qt driver work under NetBSD. Still to do:
- bus-ify.
- bpf support.
- multicast support.

Driver for the DELQA-PLUS card in Turbo mode, from 2.11BSD, written by
Steven M. Schultz.

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.

branches: 1.31.18; 1.31.36;
Don't pull in the whole uvm(9) API to access only PAGE_SIZE and
some other constants. These are provided by sys/param.h now.

branches: 1.30.26;
Rototill the vax code. Switch to devicet/PRIV_ALLOC. Cleanup vax autoconf
code. Move to prototype definitions. staticfy, constify, avoid casting.
Use device_* accessors.

branches: 1.29.12; 1.29.16;
machine/{bus,cpu,intr}.h -> sys/{bus,cpu,intr}.h

branches: 1.28.2; 1.28.14; 1.28.16; 1.28.20;
Cast to char* before doing pointer arithmetic.

Kill caddr_t; there will be some MI fallout, but it will be fixed shortly.

branches: 1.26.26;
merge ktrace-lwp.

branches: 1.25.4;
White space nits.

branches: 1.24.8; 1.24.10;
Move UCB-licensed code from 4-clause to 3-clause licence.

Patches provided by Joel Baker in PR 22364, verified by myself.

branches: 1.23.2;
Use PAGE_SIZE rather than NBPG.

add RCSIDs

branches: 1.21.2;
Trying to get these files out of Attic.

Trying to get files out of Attic.

branches: 1.19.2;
Back from the dead, but now changed to use the bus_dma interface.
Still to do:
- ubaresets
- BDP handling

Discard these old leftovers from the IMP and trailer days!

Remove register declarations.

branches: 1.16.2; 1.16.4;
Copied from ../../arch/vax/if/if_uba.c,v

Giant change: NBPG now set to 4k and CLSIZE == 1 for vax. This change
made a whole bunch of annoying bugs disappear; mostly depending on
bad use of NBPG in non-MD code. The VAX port was the only port that
used this historical "feature".

The CL* macros should probably go away totally, there is no reason
at all to keep them.

branches: 1.14.2;
the btoc/ctob/btop macros must use CLBYTES, not NBPG. How this should
work in the long run is an open issue; some parts must be reworked
in a MI way.

EGCS fixes.

Change uba* functions to take pointers, rather than uba unit numbers.

Convert all devices according to the changes to config.

Add prototypes.

Fix #includes.

Fix #includes.

Dynamic allocate uba_softc and interrupt vectors.
Allow for more that one uba.

GC old comment about trailers.

Sync with the rest of the world. Say goodbye to Mr. Trailer.

Bug fixes to get network function OK.

new RCS ID format.

fixed minor problem with mtpr()

Initial VAX port merging.

Kill caddr_t; there will be some MI fallout, but it will be fixed shortly.

branches: 1.14.26;
merge ktrace-lwp.

branches: 1.13.4;
White space nits.

branches: 1.12.8; 1.12.10;
Move UCB-licensed code from 4-clause to 3-clause licence.

Patches provided by Joel Baker in PR 22364, verified by myself.

branches: 1.11.2; 1.11.22;
Trying to get these files out of Attic.

Trying to get files out of Attic.

branches: 1.9.2;
Back from the dead, but now changed to use the bus_dma interface.
Still to do:
- ubaresets
- BDP handling

Discard these old leftovers from the IMP and trailer days!

branches: 1.7.2; 1.7.4;
Copied from ../../arch/vax/if/if_uba.h,v

Change uba* functions to take pointers, rather than uba unit numbers.

Added prototypes to everything. Made all files compile with -Wall.

Add prototypes.

KERNEL -> _KERNEL

new RCS ID format.

Initial VAX port merging.

Fix few typos in comments.

fix various typos in comments.

fix various typos in comments and log messages.

fix various typos in comments.

Change the kqueue filterops::f_isfd field to filterops::f_flags, and
define a flag FILTEROP_ISFD that has the meaning of the prior f_isfd.
Field and flag name aligned with OpenBSD.

This does not constitute a functional or ABI change, as the field location
and size, and the value placed in that field, are the same as the previous
code, but we're bumping __NetBSD_Version__ so 3rd-party module source code
can adapt, as needed.

NetBSD 9.99.89

Use sel{record,remove}_knote().

branches: 1.57.18;
Use C99 initializer for filterops

Mostly done with spatch with touchups for indentation

@@
expression a;
identifier b,c,d;
identifier p;
@@
const struct filterops p =
- { a, b, c, d
+ {
+ .f_isfd = a,
+ .f_attach = b,
+ .f_detach = c,
+ .f_event = d,
};

src is too big these days to tolerate superfluous apostrophes. It's
"its", people!

Add d_discard to all struct cdevsw instances I could find.

All have been set to "nodiscard"; some should get a real implementation.

branches: 1.54.2;
Change (mostly mechanically) every cdevsw/bdevsw I can find to use
designated initializers.

I have not built every extant kernel so I have probably broken at
least one build; however I've also found and fixed some wrong
cdevsw/bdevsw entries so even if so I think we come out ahead.

branches: 1.53.2;
split device_t/softc for all remaining drivers.
replace "struct device *" with "device_t".
use device_xname(), device_unit(), etc.

branches: 1.52.4; 1.52.14;
Rename ttymalloc() to tty_alloc(), and ttyfree() to tty_free() for
consistency. Remove some unnecessary malloc.h inclusions as well.

branches: 1.51.2;
Don't pull in the whole uvm(9) API to access only PAGE_SIZE and
some other constants. These are provided by sys/param.h now.

branches: 1.50.4;
Don't pass a buffer to physio(), let it be allocated dynamically.

This leaves only scsipi and atapi doing the same.

Remove extra whitespace added by a stupid tool.
XXX: more in src/sys/arch

bcopy -> memcpy

Ansify function definitions w/o arguments. Generated with sed.

ANSIfy another 1261 function definitions.
The only ones left in sys are beyond by sed script!
(or in sys/dist or sys/external)
Mostly they have function pointer parameters.

Change about 4500 of the K&R function definitions to ANSI ones.
There are still about 1600 left, but they have ',' or /* ... */
in the actual variable definitions - which my awk script doesn't handle.
There are also many that need () -> (void).
(The script does handle misordered arguments.)

branches: 1.44.4; 1.44.10;
use device_lookup_private to get softc

branches: 1.43.2; 1.43.4; 1.43.6; 1.43.8;
Rototill the vax code. Switch to devicet/PRIV_ALLOC. Cleanup vax autoconf
code. Move to prototype definitions. staticfy, constify, avoid casting.
Use device_* accessors.

Welcome to 4.99.55:

- Add a lot of missing selinit() and seldestroy() calls.

- Merge selwakeup() and selnotify() calls into a single selnotify().

- Add an additional 'events' argument to selnotify() call. It will
indicate which event (POLL_IN, POLL_OUT, etc) happen. If unknown,
zero may be used.

Note: please pass appropriate value of 'events' where possible.
Proposed on: <tech-kern>

branches: 1.41.8; 1.41.12;
Do not "return 1" from kqfilter for errors. That value is passed
directly to the userland caller and results in a mysterious EPERM.
Instead, return EINVAL or something else sensible depending on the
case.

branches: 1.40.2;
- Factor out too many copies of the same bit of tty code.
- Fix another tty signalling/wakeup problem.

branches: 1.39.2;
machine/{bus,cpu,intr}.h -> sys/{bus,cpu,intr}.h

branches: 1.38.4; 1.38.6; 1.38.10; 1.38.12;
It's not a good idea for device drivers to modify b_flags, as they don't
need to understand the locking around that field. Instead of setting
B_ERROR, set b_error instead. b_error is 'owned' by whoever completes
the I/O request.

branches: 1.37.2; 1.37.10;
Kill caddr_t; there will be some MI fallout, but it will be fixed shortly.

branches: 1.36.14;
Use device_unit().

Use device_parent().

branches: 1.34.4; 1.34.6; 1.34.8; 1.34.10; 1.34.12;
merge ktrace-lwp.

branches: 1.33.4;
White space nits.

de-__P

branches: 1.31.8; 1.31.10;
Move UCB-licensed code from 4-clause to 3-clause licence.

Patches provided by Joel Baker in PR 22364, verified by myself.

branches: 1.30.6;
si_ -> sel_

merge kqueue branch into -current

kqueue provides a stateful and efficient event notification framework
currently supported events include socket, file, directory, fifo,
pipe, tty and device changes, and monitoring of processes and signals

kqueue is supported by all writable filesystems in NetBSD tree
(with exception of Coda) and all device drivers supporting poll(2)

based on work done by Jonathan Lemon for FreeBSD
initial NetBSD port done by Luke Mewburn and Jason Thorpe

Add trailing ; to CFATTACH_DECL.

Use CFATTACH_DECL().

Declare all cfattach structures const.

Merge the gehenna-devsw branch into the trunk.

This merge changes the device switch tables from static array to
dynamically generated by config(8).

- All device switches is defined as a constant structure in device drivers.

- The new grammer ``device-major'' is introduced to ``files''.

device-major <prefix> char <num> [block <num>] [<rules>]

- All device major numbers must be listed up in port dependent majors.<arch>
by using this grammer.

- Added the new naming convention.
The name of the device switch must be <prefix>_[bc]devsw for auto-generation
of device switch tables.

- The backward compatibility of loading block/character device
switch by LKM framework is broken. This is necessary to convert
from block/character device major to device name in runtime and vice versa.

- The restriction to assign device major by LKM is completely removed.
We don't need to reserve LKM entries for dynamic loading of device switch.

- In compile time, device major numbers list is packed into the kernel and
the LKM framework will refer it to assign device major number dynamically.

branches: 1.24.4;
Convert ioctl code to use EPASSTHROUGH instead of -1 or ENOTTY for
indicating an unhandled "command". ERESTART is -1, which can lead to
confusion. ERESTART has been moved to -3 and EPASSTHROUGH has been
placed at -4. No ioctl code should now return -1 anywhere. The
ioctl() system call is now properly restartable.

add RCSIDs