remove unneeded includes; ok miod@

fix double words in comments
feedback and ok jmc@ miod, ok millert@

remove tulip_mbuf_compress() unused since if_de.c 1.127
ok miod@

Constify struct cfattach.

Delete unnecessary #includes of <sys/domain.h> and/or <sys/protosw.h>

net/if_pppx.c pointed out by jsg@
ok gnezdo@ deraadt@ jsg@ mpi@ millert@

spelling
feedback and ok tb@ jmc@ ok ratchov@

Do not clear IFF_UP, even in the error path, clearing IFF_RUNNING
is enough.

This flag should only be set by the stack, drivers shouldn't mess
with it.

Discussed with dlg@ and mikeb@, ok mikeb@, stsp@

move counting if_opackets next to counting if_obytes in if_enqueue.

this means packets are consistently counted in one place, unlike the
many and various ways that drivers thought they should do it.

ok mpi@ deraadt@

G/C IFQ_SET_READY().

No trailers has been the default and only option for 20 years, yet some
drivers still set IFF_NOTRAILERS while others do not. Remove all usage of
the flag from the drivers which in ancient times used it (and the modern
drivers which blindly copied it from those drivers of yore).
suggested by guenther. ok mpi

dont need to repeatedly set if_start to the same function.

Network drivers should not include <net/route.h> or <net/netisr.h>

replace IFF_OACTIVE manipulation with mpsafe operations.

there are two things shared between the network stack and drivers
in the send path: the send queue and the IFF_OACTIVE flag. the send
queue is now protected by a mutex. this diff makes the oactive
functionality mpsafe too.

IFF_OACTIVE is part of if_flags. there are two problems with that.
firstly, if_flags is a short and we dont have any MI atomic operations
to manipulate a short. secondly, while we could make the IFF_OACTIVE
operates mpsafe, all changes to other flags would have to be made
safe at the same time, otherwise a read-modify-write cycle on their
updates could clobber the oactive change.

instead, this moves the oactive mark into struct ifqueue and provides
an API for changing it. there's ifq_set_oactive, ifq_clr_oactive,
and ifq_is_oactive. these are modelled on ifsq_set_oactive,
ifsq_clr_oactive, and ifsq_is_oactive in dragonflybsd.

this diff includes changes to all the drivers manipulating IFF_OACTIVE
to now use the ifsq_{set,clr_is}_oactive API too.

ok kettenis@ mpi@ jmatthew@ deraadt@

You only need <net/if_dl.h> if you're using LLADDR() or a sockaddr_dl.

The only network driver needing <net/if_types.h> is upl(4) for IFT_OTHER.

you cant touch m_pkthdr in between ifq_deq_begin and ifq_deq_commit.

TULIP_SETCTX() does though. this moves that call after ifq_deq_commit.

problem reported (with great detail) and fix tested by ricardo mestre
ok mpi@

shuffle struct ifqueue so in flight mbufs are protected by a mutex.

the code is refactored so the IFQ macros call newly implemented ifq
functions. the ifq code is split so each discipline (priq and hfsc
in our case) is an opaque set of operations that the common ifq
code can call. the common code does the locking, accounting (ifq_len
manipulation), and freeing of the mbuf if the disciplines enqueue
function rejects it. theyre kind of like bufqs in the block layer
with their fifo and nscan disciplines.

the new api also supports atomic switching of disciplines at runtime.
the hfsc setup in pf_ioctl.c has been tweaked to build a complete
hfsc_if structure which it attaches to the send queue in a single
operation, rather than attaching to the interface up front and
building up a list of queues.

the send queue is now mutexed, which raises the expectation that
packets can be enqueued or purged on one cpu while another cpu is
dequeueing them in a driver for transmission. a lot of drivers use
IFQ_POLL to peek at an mbuf and attempt to fit it on the ring before
committing to it with a later IFQ_DEQUEUE operation. if the mbuf
gets freed in between the POLL and DEQUEUE operations, fireworks
will ensue.

to avoid this, the ifq api introduces ifq_deq_begin, ifq_deq_rollback,
and ifq_deq_commit. ifq_deq_begin allows a driver to take the ifq
mutex and get a reference to the mbuf they wish to try and tx. if
there's space, they can ifq_deq_commit it to remove the mbuf and
release the mutex. if there's no space, ifq_deq_rollback simply
releases the mutex. this api was developed to make updating the
drivers using IFQ_POLL easy, instead of having to do significant
semantic changes to avoid POLL that we cannot test on all the
hardware.

the common code has been tested pretty hard, and all the driver
modifications are straightforward except for de(4). if that breaks
it can be dealt with later.

ok mpi@ jmatthew@

use IFQ_IS_EMPTY to see if if_snd is empty, not IF_IS_EMPTY

replace the ifqueues used for tulip_txq and _rxq with mbuf_lists.

arp_ifinit() is no longer needed.

There is no excuse for using dma_alloc(9) when a bus_dma_tag_t is available.

deraadt@ thinks this isn't worth it, but reyk@ and mlarkin@ tested it anyway

Ensure the setup block is DMA reachable.
tested by reyk, mlarkin, others

Increment if_ipackets in if_input().

Note that pseudo-drivers not using if_input() are not affected by this
conversion.

ok mikeb@, kettenis@, claudio@, dlg@

Convert to if_input().

ok dlg@

Remove annoying comment mentionning ether_output().

Remove some includes include-what-you-use claims don't
have any direct symbols used. Tested for indirect use by compiling
amd64/i386/sparc64 kernels.

ok tedu@ deraadt@

unifdef INET

remove uneeded route.h includes
ok miod@ mpi@

Fewer <netinet/in_systm.h>

Remove some altq tentacles.

ok pelikan@, henning@

another questionable "optimization": de used tulip_ifstart_one instead
of tulip_ifstart if the sendqueue was empty, but only if altq wasn't
compiled in (i. e., that's a _compile time_ decision and not based on
wether altq was actually used).
just use tulip_ifstart all the time, as before in our regular kernels.
kill tulip_ifstart_one completely.
makes sense to sthen

Instead of comparing the lower and higher addresses of all the multicast
entries to decide if the IFF_ALLMULTI flag should be set, check if there
is at least one real range between them.

This should not change the behavior of any driver but if you encounter
any problem, feel free to revert the offending chunk and ping me about
it.

ok naddy@, dlg@

Most network drivers include netinet/in_var.h, but apparently they
don't have to. Just remove these include lines.
Compiled on amd64 i386 sparc64; OK henning@ mikeb@

Prevent panic'ing on alpha after ifconfig'ing up an unplugged de interface,
by making tulip_txput() aware of whether or not the mbuf it is processing
is in if_snq or not, rather than abusing the TULIP_TXPROBE_ACTIVE flag.

Found the hard way by kurt@, tested on AlphaServer 1000A, I've been sleeping
on this diff for about 3 years.

use IF_LEN/IFQ_LEN to access and ifqueue's length field. ryan ok
with this nothing in the tree fiddles if ifqueue internals any more, of
course except if.c and if.h (and some altq)

cosnistently use IFQ_SET_MAXLEN, surfaced in a discussion with + ok bluhm

use nitems(); no binary change for drivers that are compiled on amd64.

ok claudio@

Stop doing shutdown hooks in network drivers where possible. We already
take all interfaces down, via their xxstop routines. Claudio and I have
verified that none of the shutdown hooks do much extra beyond what xxstop
was already doing; it is largely a pile of junk.
ok claudio, some early comments by sthen; also read by matthew, jsg

these files don't need to include proc.h anymore. ok oga for agp

remove trash left in debug changes; lindroos@nls.fi

- consistify cfdriver for the ethernet drivers (0 -> NULL)

ok dlg@

Tidy up allocation of transmit DMA maps and generalize it to also
tidy up allocation of receive DMA maps. Previously the driver was
using DMA maps off the free list without fully allocating them, in
order to save two or three lines releasing them on error paths.
This was causing it to reuse a map already in use when under load.
From NetBSD.

Revert a workaround against a NULL pointer dereference on alpha
when invoking bus_dmaamp_sync.

All submitted by Brad.

Eliminate the redundant bits of code for MTU and multicast handling
from the individual drivers now that ether_ioctl() handles this.

Shrinks the i386 kernels by..
RAMDISK - 2176 bytes
RAMDISKB - 1504 bytes
RAMDISKC - 736 bytes

Tested by naddy@/okan@/sthen@/brad@/todd@/jmc@ and lots of users.
Build tested on almost all archs by todd@/brad@

ok naddy@

First step towards cleaning up the Ethernet driver ioctl handling.
Move calling ether_ioctl() from the top of the ioctl function, which
at the moment does absolutely nothing, to the default switch case.
Thus allowing drivers to define their own ioctl handlers and then
falling back on ether_ioctl(). The only functional change this results
in at the moment is having all Ethernet drivers returning the proper
errno of ENOTTY instead of EINVAL/ENXIO when encountering unknown
ioctl's.

Shrinks the i386 kernels by..
RAMDISK - 1024 bytes
RAMDISKB - 1120 bytes
RAMDISKC - 832 bytes

Tested by martin@/jsing@/todd@/brad@
Build tested on almost all archs by todd@/brad@

ok jsing@

Add a workaround against a NULL pointer dereference on alpha when
invoking bus_dmaamp_sync, until the cause of the problem is found.
ok deraadt@

Remove unused tulip_21041_media_noprobe() function and tulip_21041np_boardsw
struct.

ok dlg@

Consistently spell FALLTHROUGH to appease lint.
ok kettenis@ cloder@ tom@ henning@

print the chipset revision too.

shorten dmesg entry from two lines to one.

tulip_pci_(probe/attach) -> tulip_(probe/attach)

remove splnet from attach routine.

de-static and formatting.

unknown ioctl is ENOTTY not EINVAL

fix non-ALTQ case.

restore ALTQ support which was lost with rev 1.87

allow bpf(4) to ignore packets based on their direction (inbound or
outbound), using a new BIOCSDIRFILT ioctl;
guidance, feedback and ok canacar@

branches: 1.87.2;
add bus_dma support.

From NetBSD

tested on i386/alpha.

ok martin@ krw@

oops, the include was just doing fine there

fix some comments and put if_devar.h include where it belongs

shuffle some stuff around

remove TULIP_PCI_ATTACH_ARGS and TULIP_SHUTDOWN_ARGS

remove more indirection through typedef's and macros

- delint: zero sized array initalisation is a gccism
- move timeout_set() from tulip_attach() to tulip_pci_attach()
- free -> freedescs in tulip_txput()
- remove duplicate flag setting in tulip_2114x_media_preset()
TULIP_MEDIA_10BASET_FD switch case
- some cleaning

reduces the delta from the NetBSD driver

add missing TULIP_STS_LINKFAIL flag in tulip_21041_media_probe().

From NetBSD rev 1.67 and verified to still be there in -current.

make de(4) use ether_input_mbuf()

From thorpej NetBSD

ok krw@ martin@

do not try to restore the spl level two times in a row.

ok martin@

- fix up the ioctl handler and make it use ether_ioctl() too
- remove more unused code
- remove some of the macros to make this more readable

ok krw@

another round of removing ifdef crud from this driver.

ok krw@ martin@

branches: 1.75.2;
do not set PCI_COMMAND_MASTER_ENABLE explicitly as it's already set in pcisubmatch(); kettenis@ testing; brad@ ok

remove __BROKEN_INDIRECT_CONFIG cruft

remove one more piece of the media detection from SRM code.

Remove media detection from the SRM on alpha. Initially a good idea, it
breaks some onboard de(4)'s. The conventional way of getting media settings
is just fine on alpha, so let's just use this. Shrinks de(4) a bit on alpha.

While there de-inline tulip_21140_mediainit().

ok brad@

remove netns crud.
some drivers actually had hooks for SIOCSIFADDR, most just useless includes
"looks good" deraadt miod brad

- identify Compex NICs (untested)
- wrap more functions in TULIP_PERF* macros

ok brad@

from NetBSD

in tulip_reset() remove a useless printf and check flags before resetting

- remomve unneeded tulip_reset() and always include intr.h
- make it possible to get media information from the SRM on alpha

ok brad@

ALTQ cleanout

ok brad@

make sure bus mastering is enabled

ok brad@

from NetBSD

remove lots of crud from this driver, makes it fit on the alpha media

pull over some minor fixes from NetBSD:

- use ETHER_ADDR_LEN where apropriate
- consistency fixes in mem-io vs io-io logic
- fixes for 21143 in full-duplex mode
- minor nits here and there

ok brad@

make sure interface is in RUNNING state before touching the multicast filters

From NetBSD

NetBSD PR 27678 for details

ok mcbride@

put more debug stuff inside #ifdef, to make the driver a wee bit smaller
for the alpha

don't need to set ifp->if_mtu or ifp->if_output in each driver,
{ether,atm,fddi}_ifattach already does this.

ok mcbride@ markus@ henning@

Replace tulip_crc32 with ether_crc32_le.

network drivers don't need to include bpfdesc.h
ok krw@ canacar@

Remove some double semicolons (hmm, do two semis equal a maxi?).
I've skipped the GNU stuff for now. From Patrick Latifi.

strcat/strcpy/sprintf cleanup. krw@, anil@ ok. art@ tested sparc64.

The #@$% message about the #$%@ cable down is going to drive me @#%$ insane.
I'm not using this built-in tulip because it sucks and I want it to SHUT UP!

withough -> without

Don't have dc and de return the same match value in any case.
Comment the reasoning for the various return values in dc_pci_match.
dc is preferred over de on all archs except for alpha where de is used.
The main reason dc is not preferred on alpha is that it doesn't fit
on the floppy, so RAMDISK and GENERIC would have different drivers for
the same device which is rather confusing.

Final __P removal plus some cosmetic fixups

branches: 1.52.2;
Replace inclusion of <vm/foo.h> with the correct <uvm/bar.h> when necessary.
(Look ma, I might have broken the tree)

Correct some pryntf() usage: get the correct number of arguments in the
correct order.

Don't include <vm/vm_kern.h> if you don't need foo_map.

Change pci_intr_map to take pci_attach_args as an argument.
All callers actually took all arguments to pci_intr_map from pci_attach_args
structs, so this simplifies code.
This also allows more complicated interrupt assignment schemes like the one
on sparc64.

This makes sparc64 pci interrupts work.

Inspired by the same change in NetBSD.

absolutely no need to include vm_parm.h after vm.h

ALTQ'ify network drivers.
- use the new queue macros.
- use IFQ_POLL() to peek at the next packet.
- use IFQ_IS_EMPTY() for empty check.
- drivers should always check if (m == NULL) after IFQ_DEQUEUE(),
since it could return NULL even when IFQ_IS_EMPTY() is FALSE
under rate-limiting.
- drivers are supposed to call if_start from tx complete interrupts
(in order to trigger the next dequeue under rate-limiting).

Make pci_mapreg_map take an extra argument where we can
put a size limitation of the PCI region to map. That makes the PERC 3/Di
raid controller usable, as it publishes too much PCI memory for us to map
in the kernel virtual memory. As we only access the first 256 byte it is
of no use to map 128MB of kvm.

Get rid of CLSIZE and all related stuff.
CLSIZE -> 1
CLBYTES -> PAGE_SIZE
OLOFSET -> PAGE_MASK
etc.
At the same time some archs needed some cleaning in vmparam.h so that
goes in at the same time.

new timeout, space control

Previous commit contained things that I want to keep local for now

Remove MTU cruft

branches: 1.41.4;


Added pci_mapreg* interface from NetBSD.
Added PCI 2.2 stuff from NetBSD
Modified if_de to use mapreg interface

compensate for the fact that the check for ownership of a unicast packet
has moved to ether_input()

correct attach printing for 4-port cards

remove pnic stuff

cogent-compatible adapters appearing in an adaptec ethernet block range

Partial merge from netBSD 102498. Seems to fix autonegotation difficulties
on Bay Networks FA310TX card.

Support for the Lite-On PNIC found on Netgear FA310TX rev D1 and
several other boards.

Dont swab the data twice, only when assiging to the structure.
This only affect big endian machines (powerpc).

Changes to make if_de driver work on big endian machines with a 21040 based
card. Tested on i386 and alpha. Could use some cleanup to swab constants
instead of variables, so that the compiler can do the work, not the processor.

com.c needs to know about mem mapped ports in ppc systems.

ncr.c and if_de.c needs some hooks for systems with ncr and dec
ether chips on the motherboard that runs on odd clocks and doesn't
support a srom. srom is emulated in machine dependent code.

big endian support for ncr and dec ether finished. the 21040
have some more coding requirements if it is to run BE since
it does not have the descriptor endian swap configuration bit.

do not panic

EM100FX

GCC 2.8 -Wall

now this is really pretty

new style pci attach printing

compile on alpha, needs testing

merge netbsd 1.56.2.1 for de500B (and other 21143) support

Few minor changes to make if_de work big endian on 21140

Support BSD/OS 2.1, add OpenBSD tags.

pretty

Force driver into store-forward mode on cards that support it. This increases
reliability (less underflows) at the expense of speed.

Comment out some superfluous const declarations that cause cast warnings
with gcc/alpha. We have bcopy(const void *...) so casting a const foo *bar
as in bcopy((caddr_t) bar,...) for bcopy seems to be fairly pointless in
the first place, but there are too many of these to hack out at the moment.

csapuntz@lcs.mit.edu: Port of NetBSD tulip driver. Works with new (21140-AC-based) SMC Etherpower cards.

These two was not ready for commit yet. Backed out.

Add Matrox Mystique

do SIOCGIFADDR in ether_ioctl

Adaption to new bus.h from NetBSD. Various driver has been updated to
NetBSD 961107 state. Other has not, partly because sometimes the bus.h
changes was coupled with other changes.

s/bus.h/bus.old.h/ to make the transit to the bus_space bus.h an easy road
possibly taken in small steps

__OpenBSD__ messup fixed

check for __OpenBSD__, prep for tossing __NetBSD__?

sync to 0607

sync 0521

if_name/if_unit -> if_xname/if_softc

sync with 0504 -- prototypes and bus.h

add general ether_ioctl call in net/if_ethersubr.c,
NS,IPX,X.25 special processing is now handled in there.
reflect this amazing addition in all the ether ifaces.
ppl, pls check the stuff.

partial sync with netbsd 960418, more to come

NetBSD 960317 merge

sc_dev -> tulip_dev

Pass device name to interrupt establish routines so it can be recorded in
the device interrupt chain structures (isa, pci)

Move interrupt chain structure definition to <machine/psl.h> so vmstat can
get at it (i386)

Remove hack to count interrupts the old way (i386)

from netbsd; merge latest Thomas version with our minimal changes

branches: 1.1.1;
Initial revision

fix double words in comments
feedback and ok jmc@ miod, ok millert@

remove tulip_mbuf_compress() unused since if_de.c 1.127
ok miod@

Constify struct cfattach.

Delete unnecessary #includes of <sys/domain.h> and/or <sys/protosw.h>

net/if_pppx.c pointed out by jsg@
ok gnezdo@ deraadt@ jsg@ mpi@ millert@

spelling
feedback and ok tb@ jmc@ ok ratchov@

Do not clear IFF_UP, even in the error path, clearing IFF_RUNNING
is enough.

This flag should only be set by the stack, drivers shouldn't mess
with it.

Discussed with dlg@ and mikeb@, ok mikeb@, stsp@

move counting if_opackets next to counting if_obytes in if_enqueue.

this means packets are consistently counted in one place, unlike the
many and various ways that drivers thought they should do it.

ok mpi@ deraadt@

G/C IFQ_SET_READY().

No trailers has been the default and only option for 20 years, yet some
drivers still set IFF_NOTRAILERS while others do not. Remove all usage of
the flag from the drivers which in ancient times used it (and the modern
drivers which blindly copied it from those drivers of yore).
suggested by guenther. ok mpi

dont need to repeatedly set if_start to the same function.

Network drivers should not include <net/route.h> or <net/netisr.h>

replace IFF_OACTIVE manipulation with mpsafe operations.

there are two things shared between the network stack and drivers
in the send path: the send queue and the IFF_OACTIVE flag. the send
queue is now protected by a mutex. this diff makes the oactive
functionality mpsafe too.

IFF_OACTIVE is part of if_flags. there are two problems with that.
firstly, if_flags is a short and we dont have any MI atomic operations
to manipulate a short. secondly, while we could make the IFF_OACTIVE
operates mpsafe, all changes to other flags would have to be made
safe at the same time, otherwise a read-modify-write cycle on their
updates could clobber the oactive change.

instead, this moves the oactive mark into struct ifqueue and provides
an API for changing it. there's ifq_set_oactive, ifq_clr_oactive,
and ifq_is_oactive. these are modelled on ifsq_set_oactive,
ifsq_clr_oactive, and ifsq_is_oactive in dragonflybsd.

this diff includes changes to all the drivers manipulating IFF_OACTIVE
to now use the ifsq_{set,clr_is}_oactive API too.

ok kettenis@ mpi@ jmatthew@ deraadt@

You only need <net/if_dl.h> if you're using LLADDR() or a sockaddr_dl.

The only network driver needing <net/if_types.h> is upl(4) for IFT_OTHER.

you cant touch m_pkthdr in between ifq_deq_begin and ifq_deq_commit.

TULIP_SETCTX() does though. this moves that call after ifq_deq_commit.

problem reported (with great detail) and fix tested by ricardo mestre
ok mpi@

shuffle struct ifqueue so in flight mbufs are protected by a mutex.

the code is refactored so the IFQ macros call newly implemented ifq
functions. the ifq code is split so each discipline (priq and hfsc
in our case) is an opaque set of operations that the common ifq
code can call. the common code does the locking, accounting (ifq_len
manipulation), and freeing of the mbuf if the disciplines enqueue
function rejects it. theyre kind of like bufqs in the block layer
with their fifo and nscan disciplines.

the new api also supports atomic switching of disciplines at runtime.
the hfsc setup in pf_ioctl.c has been tweaked to build a complete
hfsc_if structure which it attaches to the send queue in a single
operation, rather than attaching to the interface up front and
building up a list of queues.

the send queue is now mutexed, which raises the expectation that
packets can be enqueued or purged on one cpu while another cpu is
dequeueing them in a driver for transmission. a lot of drivers use
IFQ_POLL to peek at an mbuf and attempt to fit it on the ring before
committing to it with a later IFQ_DEQUEUE operation. if the mbuf
gets freed in between the POLL and DEQUEUE operations, fireworks
will ensue.

to avoid this, the ifq api introduces ifq_deq_begin, ifq_deq_rollback,
and ifq_deq_commit. ifq_deq_begin allows a driver to take the ifq
mutex and get a reference to the mbuf they wish to try and tx. if
there's space, they can ifq_deq_commit it to remove the mbuf and
release the mutex. if there's no space, ifq_deq_rollback simply
releases the mutex. this api was developed to make updating the
drivers using IFQ_POLL easy, instead of having to do significant
semantic changes to avoid POLL that we cannot test on all the
hardware.

the common code has been tested pretty hard, and all the driver
modifications are straightforward except for de(4). if that breaks
it can be dealt with later.

ok mpi@ jmatthew@

use IFQ_IS_EMPTY to see if if_snd is empty, not IF_IS_EMPTY

replace the ifqueues used for tulip_txq and _rxq with mbuf_lists.

arp_ifinit() is no longer needed.

There is no excuse for using dma_alloc(9) when a bus_dma_tag_t is available.

deraadt@ thinks this isn't worth it, but reyk@ and mlarkin@ tested it anyway

Ensure the setup block is DMA reachable.
tested by reyk, mlarkin, others

Increment if_ipackets in if_input().

Note that pseudo-drivers not using if_input() are not affected by this
conversion.

ok mikeb@, kettenis@, claudio@, dlg@

Convert to if_input().

ok dlg@

Remove annoying comment mentionning ether_output().

Remove some includes include-what-you-use claims don't
have any direct symbols used. Tested for indirect use by compiling
amd64/i386/sparc64 kernels.

ok tedu@ deraadt@

unifdef INET

remove uneeded route.h includes
ok miod@ mpi@

Fewer <netinet/in_systm.h>

Remove some altq tentacles.

ok pelikan@, henning@

another questionable "optimization": de used tulip_ifstart_one instead
of tulip_ifstart if the sendqueue was empty, but only if altq wasn't
compiled in (i. e., that's a _compile time_ decision and not based on
wether altq was actually used).
just use tulip_ifstart all the time, as before in our regular kernels.
kill tulip_ifstart_one completely.
makes sense to sthen

Instead of comparing the lower and higher addresses of all the multicast
entries to decide if the IFF_ALLMULTI flag should be set, check if there
is at least one real range between them.

This should not change the behavior of any driver but if you encounter
any problem, feel free to revert the offending chunk and ping me about
it.

ok naddy@, dlg@

Most network drivers include netinet/in_var.h, but apparently they
don't have to. Just remove these include lines.
Compiled on amd64 i386 sparc64; OK henning@ mikeb@

Prevent panic'ing on alpha after ifconfig'ing up an unplugged de interface,
by making tulip_txput() aware of whether or not the mbuf it is processing
is in if_snq or not, rather than abusing the TULIP_TXPROBE_ACTIVE flag.

Found the hard way by kurt@, tested on AlphaServer 1000A, I've been sleeping
on this diff for about 3 years.

use IF_LEN/IFQ_LEN to access and ifqueue's length field. ryan ok
with this nothing in the tree fiddles if ifqueue internals any more, of
course except if.c and if.h (and some altq)

cosnistently use IFQ_SET_MAXLEN, surfaced in a discussion with + ok bluhm

use nitems(); no binary change for drivers that are compiled on amd64.

ok claudio@

Stop doing shutdown hooks in network drivers where possible. We already
take all interfaces down, via their xxstop routines. Claudio and I have
verified that none of the shutdown hooks do much extra beyond what xxstop
was already doing; it is largely a pile of junk.
ok claudio, some early comments by sthen; also read by matthew, jsg

these files don't need to include proc.h anymore. ok oga for agp

remove trash left in debug changes; lindroos@nls.fi

- consistify cfdriver for the ethernet drivers (0 -> NULL)

ok dlg@

Tidy up allocation of transmit DMA maps and generalize it to also
tidy up allocation of receive DMA maps. Previously the driver was
using DMA maps off the free list without fully allocating them, in
order to save two or three lines releasing them on error paths.
This was causing it to reuse a map already in use when under load.
From NetBSD.

Revert a workaround against a NULL pointer dereference on alpha
when invoking bus_dmaamp_sync.

All submitted by Brad.

Eliminate the redundant bits of code for MTU and multicast handling
from the individual drivers now that ether_ioctl() handles this.

Shrinks the i386 kernels by..
RAMDISK - 2176 bytes
RAMDISKB - 1504 bytes
RAMDISKC - 736 bytes

Tested by naddy@/okan@/sthen@/brad@/todd@/jmc@ and lots of users.
Build tested on almost all archs by todd@/brad@

ok naddy@

First step towards cleaning up the Ethernet driver ioctl handling.
Move calling ether_ioctl() from the top of the ioctl function, which
at the moment does absolutely nothing, to the default switch case.
Thus allowing drivers to define their own ioctl handlers and then
falling back on ether_ioctl(). The only functional change this results
in at the moment is having all Ethernet drivers returning the proper
errno of ENOTTY instead of EINVAL/ENXIO when encountering unknown
ioctl's.

Shrinks the i386 kernels by..
RAMDISK - 1024 bytes
RAMDISKB - 1120 bytes
RAMDISKC - 832 bytes

Tested by martin@/jsing@/todd@/brad@
Build tested on almost all archs by todd@/brad@

ok jsing@

Add a workaround against a NULL pointer dereference on alpha when
invoking bus_dmaamp_sync, until the cause of the problem is found.
ok deraadt@

Remove unused tulip_21041_media_noprobe() function and tulip_21041np_boardsw
struct.

ok dlg@

Consistently spell FALLTHROUGH to appease lint.
ok kettenis@ cloder@ tom@ henning@

print the chipset revision too.

shorten dmesg entry from two lines to one.

tulip_pci_(probe/attach) -> tulip_(probe/attach)

remove splnet from attach routine.

de-static and formatting.

unknown ioctl is ENOTTY not EINVAL

fix non-ALTQ case.

restore ALTQ support which was lost with rev 1.87

allow bpf(4) to ignore packets based on their direction (inbound or
outbound), using a new BIOCSDIRFILT ioctl;
guidance, feedback and ok canacar@

branches: 1.87.2;
add bus_dma support.

From NetBSD

tested on i386/alpha.

ok martin@ krw@

oops, the include was just doing fine there

fix some comments and put if_devar.h include where it belongs

shuffle some stuff around

remove TULIP_PCI_ATTACH_ARGS and TULIP_SHUTDOWN_ARGS

remove more indirection through typedef's and macros

- delint: zero sized array initalisation is a gccism
- move timeout_set() from tulip_attach() to tulip_pci_attach()
- free -> freedescs in tulip_txput()
- remove duplicate flag setting in tulip_2114x_media_preset()
TULIP_MEDIA_10BASET_FD switch case
- some cleaning

reduces the delta from the NetBSD driver

add missing TULIP_STS_LINKFAIL flag in tulip_21041_media_probe().

From NetBSD rev 1.67 and verified to still be there in -current.

make de(4) use ether_input_mbuf()

From thorpej NetBSD

ok krw@ martin@

do not try to restore the spl level two times in a row.

ok martin@

- fix up the ioctl handler and make it use ether_ioctl() too
- remove more unused code
- remove some of the macros to make this more readable

ok krw@

another round of removing ifdef crud from this driver.

ok krw@ martin@

branches: 1.75.2;
do not set PCI_COMMAND_MASTER_ENABLE explicitly as it's already set in pcisubmatch(); kettenis@ testing; brad@ ok

remove __BROKEN_INDIRECT_CONFIG cruft

remove one more piece of the media detection from SRM code.

Remove media detection from the SRM on alpha. Initially a good idea, it
breaks some onboard de(4)'s. The conventional way of getting media settings
is just fine on alpha, so let's just use this. Shrinks de(4) a bit on alpha.

While there de-inline tulip_21140_mediainit().

ok brad@

remove netns crud.
some drivers actually had hooks for SIOCSIFADDR, most just useless includes
"looks good" deraadt miod brad

- identify Compex NICs (untested)
- wrap more functions in TULIP_PERF* macros

ok brad@

from NetBSD

in tulip_reset() remove a useless printf and check flags before resetting

- remomve unneeded tulip_reset() and always include intr.h
- make it possible to get media information from the SRM on alpha

ok brad@

ALTQ cleanout

ok brad@

make sure bus mastering is enabled

ok brad@

from NetBSD

remove lots of crud from this driver, makes it fit on the alpha media

pull over some minor fixes from NetBSD:

- use ETHER_ADDR_LEN where apropriate
- consistency fixes in mem-io vs io-io logic
- fixes for 21143 in full-duplex mode
- minor nits here and there

ok brad@

make sure interface is in RUNNING state before touching the multicast filters

From NetBSD

NetBSD PR 27678 for details

ok mcbride@

put more debug stuff inside #ifdef, to make the driver a wee bit smaller
for the alpha

don't need to set ifp->if_mtu or ifp->if_output in each driver,
{ether,atm,fddi}_ifattach already does this.

ok mcbride@ markus@ henning@

Replace tulip_crc32 with ether_crc32_le.

network drivers don't need to include bpfdesc.h
ok krw@ canacar@

Remove some double semicolons (hmm, do two semis equal a maxi?).
I've skipped the GNU stuff for now. From Patrick Latifi.

strcat/strcpy/sprintf cleanup. krw@, anil@ ok. art@ tested sparc64.

The #@$% message about the #$%@ cable down is going to drive me @#%$ insane.
I'm not using this built-in tulip because it sucks and I want it to SHUT UP!

withough -> without

Don't have dc and de return the same match value in any case.
Comment the reasoning for the various return values in dc_pci_match.
dc is preferred over de on all archs except for alpha where de is used.
The main reason dc is not preferred on alpha is that it doesn't fit
on the floppy, so RAMDISK and GENERIC would have different drivers for
the same device which is rather confusing.

Final __P removal plus some cosmetic fixups

branches: 1.52.2;
Replace inclusion of <vm/foo.h> with the correct <uvm/bar.h> when necessary.
(Look ma, I might have broken the tree)

Correct some pryntf() usage: get the correct number of arguments in the
correct order.

Don't include <vm/vm_kern.h> if you don't need foo_map.

Change pci_intr_map to take pci_attach_args as an argument.
All callers actually took all arguments to pci_intr_map from pci_attach_args
structs, so this simplifies code.
This also allows more complicated interrupt assignment schemes like the one
on sparc64.

This makes sparc64 pci interrupts work.

Inspired by the same change in NetBSD.

absolutely no need to include vm_parm.h after vm.h

ALTQ'ify network drivers.
- use the new queue macros.
- use IFQ_POLL() to peek at the next packet.
- use IFQ_IS_EMPTY() for empty check.
- drivers should always check if (m == NULL) after IFQ_DEQUEUE(),
since it could return NULL even when IFQ_IS_EMPTY() is FALSE
under rate-limiting.
- drivers are supposed to call if_start from tx complete interrupts
(in order to trigger the next dequeue under rate-limiting).

Make pci_mapreg_map take an extra argument where we can
put a size limitation of the PCI region to map. That makes the PERC 3/Di
raid controller usable, as it publishes too much PCI memory for us to map
in the kernel virtual memory. As we only access the first 256 byte it is
of no use to map 128MB of kvm.

Get rid of CLSIZE and all related stuff.
CLSIZE -> 1
CLBYTES -> PAGE_SIZE
OLOFSET -> PAGE_MASK
etc.
At the same time some archs needed some cleaning in vmparam.h so that
goes in at the same time.

new timeout, space control

Previous commit contained things that I want to keep local for now

Remove MTU cruft

branches: 1.41.4;


Added pci_mapreg* interface from NetBSD.
Added PCI 2.2 stuff from NetBSD
Modified if_de to use mapreg interface

compensate for the fact that the check for ownership of a unicast packet
has moved to ether_input()

correct attach printing for 4-port cards

remove pnic stuff

cogent-compatible adapters appearing in an adaptec ethernet block range

Partial merge from netBSD 102498. Seems to fix autonegotation difficulties
on Bay Networks FA310TX card.

Support for the Lite-On PNIC found on Netgear FA310TX rev D1 and
several other boards.

Dont swab the data twice, only when assiging to the structure.
This only affect big endian machines (powerpc).

Changes to make if_de driver work on big endian machines with a 21040 based
card. Tested on i386 and alpha. Could use some cleanup to swab constants
instead of variables, so that the compiler can do the work, not the processor.

com.c needs to know about mem mapped ports in ppc systems.

ncr.c and if_de.c needs some hooks for systems with ncr and dec
ether chips on the motherboard that runs on odd clocks and doesn't
support a srom. srom is emulated in machine dependent code.

big endian support for ncr and dec ether finished. the 21040
have some more coding requirements if it is to run BE since
it does not have the descriptor endian swap configuration bit.

do not panic

EM100FX

GCC 2.8 -Wall

now this is really pretty

new style pci attach printing

compile on alpha, needs testing

merge netbsd 1.56.2.1 for de500B (and other 21143) support

Few minor changes to make if_de work big endian on 21140

Support BSD/OS 2.1, add OpenBSD tags.

pretty

Force driver into store-forward mode on cards that support it. This increases
reliability (less underflows) at the expense of speed.

Comment out some superfluous const declarations that cause cast warnings
with gcc/alpha. We have bcopy(const void *...) so casting a const foo *bar
as in bcopy((caddr_t) bar,...) for bcopy seems to be fairly pointless in
the first place, but there are too many of these to hack out at the moment.

csapuntz@lcs.mit.edu: Port of NetBSD tulip driver. Works with new (21140-AC-based) SMC Etherpower cards.

These two was not ready for commit yet. Backed out.

Add Matrox Mystique

do SIOCGIFADDR in ether_ioctl

Adaption to new bus.h from NetBSD. Various driver has been updated to
NetBSD 961107 state. Other has not, partly because sometimes the bus.h
changes was coupled with other changes.

s/bus.h/bus.old.h/ to make the transit to the bus_space bus.h an easy road
possibly taken in small steps

__OpenBSD__ messup fixed

check for __OpenBSD__, prep for tossing __NetBSD__?

sync to 0607

sync 0521

if_name/if_unit -> if_xname/if_softc

sync with 0504 -- prototypes and bus.h

add general ether_ioctl call in net/if_ethersubr.c,
NS,IPX,X.25 special processing is now handled in there.
reflect this amazing addition in all the ether ifaces.
ppl, pls check the stuff.

partial sync with netbsd 960418, more to come

NetBSD 960317 merge

sc_dev -> tulip_dev

Pass device name to interrupt establish routines so it can be recorded in
the device interrupt chain structures (isa, pci)

Move interrupt chain structure definition to <machine/psl.h> so vmstat can
get at it (i386)

Remove hack to count interrupts the old way (i386)

from netbsd; merge latest Thomas version with our minimal changes

branches: 1.1.1;
Initial revision

remove tulip_mbuf_compress() unused since if_de.c 1.127
ok miod@

Constify struct cfattach.

Delete unnecessary #includes of <sys/domain.h> and/or <sys/protosw.h>

net/if_pppx.c pointed out by jsg@
ok gnezdo@ deraadt@ jsg@ mpi@ millert@

spelling
feedback and ok tb@ jmc@ ok ratchov@

Do not clear IFF_UP, even in the error path, clearing IFF_RUNNING
is enough.

This flag should only be set by the stack, drivers shouldn't mess
with it.

Discussed with dlg@ and mikeb@, ok mikeb@, stsp@

move counting if_opackets next to counting if_obytes in if_enqueue.

this means packets are consistently counted in one place, unlike the
many and various ways that drivers thought they should do it.

ok mpi@ deraadt@

G/C IFQ_SET_READY().

No trailers has been the default and only option for 20 years, yet some
drivers still set IFF_NOTRAILERS while others do not. Remove all usage of
the flag from the drivers which in ancient times used it (and the modern
drivers which blindly copied it from those drivers of yore).
suggested by guenther. ok mpi

dont need to repeatedly set if_start to the same function.

Network drivers should not include <net/route.h> or <net/netisr.h>

replace IFF_OACTIVE manipulation with mpsafe operations.

there are two things shared between the network stack and drivers
in the send path: the send queue and the IFF_OACTIVE flag. the send
queue is now protected by a mutex. this diff makes the oactive
functionality mpsafe too.

IFF_OACTIVE is part of if_flags. there are two problems with that.
firstly, if_flags is a short and we dont have any MI atomic operations
to manipulate a short. secondly, while we could make the IFF_OACTIVE
operates mpsafe, all changes to other flags would have to be made
safe at the same time, otherwise a read-modify-write cycle on their
updates could clobber the oactive change.

instead, this moves the oactive mark into struct ifqueue and provides
an API for changing it. there's ifq_set_oactive, ifq_clr_oactive,
and ifq_is_oactive. these are modelled on ifsq_set_oactive,
ifsq_clr_oactive, and ifsq_is_oactive in dragonflybsd.

this diff includes changes to all the drivers manipulating IFF_OACTIVE
to now use the ifsq_{set,clr_is}_oactive API too.

ok kettenis@ mpi@ jmatthew@ deraadt@

You only need <net/if_dl.h> if you're using LLADDR() or a sockaddr_dl.

The only network driver needing <net/if_types.h> is upl(4) for IFT_OTHER.

you cant touch m_pkthdr in between ifq_deq_begin and ifq_deq_commit.

TULIP_SETCTX() does though. this moves that call after ifq_deq_commit.

problem reported (with great detail) and fix tested by ricardo mestre
ok mpi@

shuffle struct ifqueue so in flight mbufs are protected by a mutex.

the code is refactored so the IFQ macros call newly implemented ifq
functions. the ifq code is split so each discipline (priq and hfsc
in our case) is an opaque set of operations that the common ifq
code can call. the common code does the locking, accounting (ifq_len
manipulation), and freeing of the mbuf if the disciplines enqueue
function rejects it. theyre kind of like bufqs in the block layer
with their fifo and nscan disciplines.

the new api also supports atomic switching of disciplines at runtime.
the hfsc setup in pf_ioctl.c has been tweaked to build a complete
hfsc_if structure which it attaches to the send queue in a single
operation, rather than attaching to the interface up front and
building up a list of queues.

the send queue is now mutexed, which raises the expectation that
packets can be enqueued or purged on one cpu while another cpu is
dequeueing them in a driver for transmission. a lot of drivers use
IFQ_POLL to peek at an mbuf and attempt to fit it on the ring before
committing to it with a later IFQ_DEQUEUE operation. if the mbuf
gets freed in between the POLL and DEQUEUE operations, fireworks
will ensue.

to avoid this, the ifq api introduces ifq_deq_begin, ifq_deq_rollback,
and ifq_deq_commit. ifq_deq_begin allows a driver to take the ifq
mutex and get a reference to the mbuf they wish to try and tx. if
there's space, they can ifq_deq_commit it to remove the mbuf and
release the mutex. if there's no space, ifq_deq_rollback simply
releases the mutex. this api was developed to make updating the
drivers using IFQ_POLL easy, instead of having to do significant
semantic changes to avoid POLL that we cannot test on all the
hardware.

the common code has been tested pretty hard, and all the driver
modifications are straightforward except for de(4). if that breaks
it can be dealt with later.

ok mpi@ jmatthew@

use IFQ_IS_EMPTY to see if if_snd is empty, not IF_IS_EMPTY

replace the ifqueues used for tulip_txq and _rxq with mbuf_lists.

arp_ifinit() is no longer needed.

There is no excuse for using dma_alloc(9) when a bus_dma_tag_t is available.

deraadt@ thinks this isn't worth it, but reyk@ and mlarkin@ tested it anyway

Ensure the setup block is DMA reachable.
tested by reyk, mlarkin, others

Increment if_ipackets in if_input().

Note that pseudo-drivers not using if_input() are not affected by this
conversion.

ok mikeb@, kettenis@, claudio@, dlg@

Convert to if_input().

ok dlg@

Remove annoying comment mentionning ether_output().

Remove some includes include-what-you-use claims don't
have any direct symbols used. Tested for indirect use by compiling
amd64/i386/sparc64 kernels.

ok tedu@ deraadt@

unifdef INET

remove uneeded route.h includes
ok miod@ mpi@

Fewer <netinet/in_systm.h>

Remove some altq tentacles.

ok pelikan@, henning@

another questionable "optimization": de used tulip_ifstart_one instead
of tulip_ifstart if the sendqueue was empty, but only if altq wasn't
compiled in (i. e., that's a _compile time_ decision and not based on
wether altq was actually used).
just use tulip_ifstart all the time, as before in our regular kernels.
kill tulip_ifstart_one completely.
makes sense to sthen

Instead of comparing the lower and higher addresses of all the multicast
entries to decide if the IFF_ALLMULTI flag should be set, check if there
is at least one real range between them.

This should not change the behavior of any driver but if you encounter
any problem, feel free to revert the offending chunk and ping me about
it.

ok naddy@, dlg@

Most network drivers include netinet/in_var.h, but apparently they
don't have to. Just remove these include lines.
Compiled on amd64 i386 sparc64; OK henning@ mikeb@

Prevent panic'ing on alpha after ifconfig'ing up an unplugged de interface,
by making tulip_txput() aware of whether or not the mbuf it is processing
is in if_snq or not, rather than abusing the TULIP_TXPROBE_ACTIVE flag.

Found the hard way by kurt@, tested on AlphaServer 1000A, I've been sleeping
on this diff for about 3 years.

use IF_LEN/IFQ_LEN to access and ifqueue's length field. ryan ok
with this nothing in the tree fiddles if ifqueue internals any more, of
course except if.c and if.h (and some altq)

cosnistently use IFQ_SET_MAXLEN, surfaced in a discussion with + ok bluhm

use nitems(); no binary change for drivers that are compiled on amd64.

ok claudio@

Stop doing shutdown hooks in network drivers where possible. We already
take all interfaces down, via their xxstop routines. Claudio and I have
verified that none of the shutdown hooks do much extra beyond what xxstop
was already doing; it is largely a pile of junk.
ok claudio, some early comments by sthen; also read by matthew, jsg

these files don't need to include proc.h anymore. ok oga for agp

remove trash left in debug changes; lindroos@nls.fi

- consistify cfdriver for the ethernet drivers (0 -> NULL)

ok dlg@

Tidy up allocation of transmit DMA maps and generalize it to also
tidy up allocation of receive DMA maps. Previously the driver was
using DMA maps off the free list without fully allocating them, in
order to save two or three lines releasing them on error paths.
This was causing it to reuse a map already in use when under load.
From NetBSD.

Revert a workaround against a NULL pointer dereference on alpha
when invoking bus_dmaamp_sync.

All submitted by Brad.

Eliminate the redundant bits of code for MTU and multicast handling
from the individual drivers now that ether_ioctl() handles this.

Shrinks the i386 kernels by..
RAMDISK - 2176 bytes
RAMDISKB - 1504 bytes
RAMDISKC - 736 bytes

Tested by naddy@/okan@/sthen@/brad@/todd@/jmc@ and lots of users.
Build tested on almost all archs by todd@/brad@

ok naddy@

First step towards cleaning up the Ethernet driver ioctl handling.
Move calling ether_ioctl() from the top of the ioctl function, which
at the moment does absolutely nothing, to the default switch case.
Thus allowing drivers to define their own ioctl handlers and then
falling back on ether_ioctl(). The only functional change this results
in at the moment is having all Ethernet drivers returning the proper
errno of ENOTTY instead of EINVAL/ENXIO when encountering unknown
ioctl's.

Shrinks the i386 kernels by..
RAMDISK - 1024 bytes
RAMDISKB - 1120 bytes
RAMDISKC - 832 bytes

Tested by martin@/jsing@/todd@/brad@
Build tested on almost all archs by todd@/brad@

ok jsing@

Add a workaround against a NULL pointer dereference on alpha when
invoking bus_dmaamp_sync, until the cause of the problem is found.
ok deraadt@

Remove unused tulip_21041_media_noprobe() function and tulip_21041np_boardsw
struct.

ok dlg@

Consistently spell FALLTHROUGH to appease lint.
ok kettenis@ cloder@ tom@ henning@

print the chipset revision too.

shorten dmesg entry from two lines to one.

tulip_pci_(probe/attach) -> tulip_(probe/attach)

remove splnet from attach routine.

de-static and formatting.

unknown ioctl is ENOTTY not EINVAL

fix non-ALTQ case.

restore ALTQ support which was lost with rev 1.87

allow bpf(4) to ignore packets based on their direction (inbound or
outbound), using a new BIOCSDIRFILT ioctl;
guidance, feedback and ok canacar@

branches: 1.87.2;
add bus_dma support.

From NetBSD

tested on i386/alpha.

ok martin@ krw@

oops, the include was just doing fine there

fix some comments and put if_devar.h include where it belongs

shuffle some stuff around

remove TULIP_PCI_ATTACH_ARGS and TULIP_SHUTDOWN_ARGS

remove more indirection through typedef's and macros

- delint: zero sized array initalisation is a gccism
- move timeout_set() from tulip_attach() to tulip_pci_attach()
- free -> freedescs in tulip_txput()
- remove duplicate flag setting in tulip_2114x_media_preset()
TULIP_MEDIA_10BASET_FD switch case
- some cleaning

reduces the delta from the NetBSD driver

add missing TULIP_STS_LINKFAIL flag in tulip_21041_media_probe().

From NetBSD rev 1.67 and verified to still be there in -current.

make de(4) use ether_input_mbuf()

From thorpej NetBSD

ok krw@ martin@

do not try to restore the spl level two times in a row.

ok martin@

- fix up the ioctl handler and make it use ether_ioctl() too
- remove more unused code
- remove some of the macros to make this more readable

ok krw@

another round of removing ifdef crud from this driver.

ok krw@ martin@

branches: 1.75.2;
do not set PCI_COMMAND_MASTER_ENABLE explicitly as it's already set in pcisubmatch(); kettenis@ testing; brad@ ok

remove __BROKEN_INDIRECT_CONFIG cruft

remove one more piece of the media detection from SRM code.

Remove media detection from the SRM on alpha. Initially a good idea, it
breaks some onboard de(4)'s. The conventional way of getting media settings
is just fine on alpha, so let's just use this. Shrinks de(4) a bit on alpha.

While there de-inline tulip_21140_mediainit().

ok brad@

remove netns crud.
some drivers actually had hooks for SIOCSIFADDR, most just useless includes
"looks good" deraadt miod brad

- identify Compex NICs (untested)
- wrap more functions in TULIP_PERF* macros

ok brad@

from NetBSD

in tulip_reset() remove a useless printf and check flags before resetting

- remomve unneeded tulip_reset() and always include intr.h
- make it possible to get media information from the SRM on alpha

ok brad@

ALTQ cleanout

ok brad@

make sure bus mastering is enabled

ok brad@

from NetBSD

remove lots of crud from this driver, makes it fit on the alpha media

pull over some minor fixes from NetBSD:

- use ETHER_ADDR_LEN where apropriate
- consistency fixes in mem-io vs io-io logic
- fixes for 21143 in full-duplex mode
- minor nits here and there

ok brad@

make sure interface is in RUNNING state before touching the multicast filters

From NetBSD

NetBSD PR 27678 for details

ok mcbride@

put more debug stuff inside #ifdef, to make the driver a wee bit smaller
for the alpha

don't need to set ifp->if_mtu or ifp->if_output in each driver,
{ether,atm,fddi}_ifattach already does this.

ok mcbride@ markus@ henning@

Replace tulip_crc32 with ether_crc32_le.

network drivers don't need to include bpfdesc.h
ok krw@ canacar@

Remove some double semicolons (hmm, do two semis equal a maxi?).
I've skipped the GNU stuff for now. From Patrick Latifi.

strcat/strcpy/sprintf cleanup. krw@, anil@ ok. art@ tested sparc64.

The #@$% message about the #$%@ cable down is going to drive me @#%$ insane.
I'm not using this built-in tulip because it sucks and I want it to SHUT UP!

withough -> without

Don't have dc and de return the same match value in any case.
Comment the reasoning for the various return values in dc_pci_match.
dc is preferred over de on all archs except for alpha where de is used.
The main reason dc is not preferred on alpha is that it doesn't fit
on the floppy, so RAMDISK and GENERIC would have different drivers for
the same device which is rather confusing.

Final __P removal plus some cosmetic fixups

branches: 1.52.2;
Replace inclusion of <vm/foo.h> with the correct <uvm/bar.h> when necessary.
(Look ma, I might have broken the tree)

Correct some pryntf() usage: get the correct number of arguments in the
correct order.

Don't include <vm/vm_kern.h> if you don't need foo_map.

Change pci_intr_map to take pci_attach_args as an argument.
All callers actually took all arguments to pci_intr_map from pci_attach_args
structs, so this simplifies code.
This also allows more complicated interrupt assignment schemes like the one
on sparc64.

This makes sparc64 pci interrupts work.

Inspired by the same change in NetBSD.

absolutely no need to include vm_parm.h after vm.h

ALTQ'ify network drivers.
- use the new queue macros.
- use IFQ_POLL() to peek at the next packet.
- use IFQ_IS_EMPTY() for empty check.
- drivers should always check if (m == NULL) after IFQ_DEQUEUE(),
since it could return NULL even when IFQ_IS_EMPTY() is FALSE
under rate-limiting.
- drivers are supposed to call if_start from tx complete interrupts
(in order to trigger the next dequeue under rate-limiting).

Make pci_mapreg_map take an extra argument where we can
put a size limitation of the PCI region to map. That makes the PERC 3/Di
raid controller usable, as it publishes too much PCI memory for us to map
in the kernel virtual memory. As we only access the first 256 byte it is
of no use to map 128MB of kvm.

Get rid of CLSIZE and all related stuff.
CLSIZE -> 1
CLBYTES -> PAGE_SIZE
OLOFSET -> PAGE_MASK
etc.
At the same time some archs needed some cleaning in vmparam.h so that
goes in at the same time.

new timeout, space control

Previous commit contained things that I want to keep local for now

Remove MTU cruft

branches: 1.41.4;


Added pci_mapreg* interface from NetBSD.
Added PCI 2.2 stuff from NetBSD
Modified if_de to use mapreg interface

compensate for the fact that the check for ownership of a unicast packet
has moved to ether_input()

correct attach printing for 4-port cards

remove pnic stuff

cogent-compatible adapters appearing in an adaptec ethernet block range

Partial merge from netBSD 102498. Seems to fix autonegotation difficulties
on Bay Networks FA310TX card.

Support for the Lite-On PNIC found on Netgear FA310TX rev D1 and
several other boards.

Dont swab the data twice, only when assiging to the structure.
This only affect big endian machines (powerpc).

Changes to make if_de driver work on big endian machines with a 21040 based
card. Tested on i386 and alpha. Could use some cleanup to swab constants
instead of variables, so that the compiler can do the work, not the processor.

com.c needs to know about mem mapped ports in ppc systems.

ncr.c and if_de.c needs some hooks for systems with ncr and dec
ether chips on the motherboard that runs on odd clocks and doesn't
support a srom. srom is emulated in machine dependent code.

big endian support for ncr and dec ether finished. the 21040
have some more coding requirements if it is to run BE since
it does not have the descriptor endian swap configuration bit.

do not panic

EM100FX

GCC 2.8 -Wall

now this is really pretty

new style pci attach printing

compile on alpha, needs testing

merge netbsd 1.56.2.1 for de500B (and other 21143) support

Few minor changes to make if_de work big endian on 21140

Support BSD/OS 2.1, add OpenBSD tags.

pretty

Force driver into store-forward mode on cards that support it. This increases
reliability (less underflows) at the expense of speed.

Comment out some superfluous const declarations that cause cast warnings
with gcc/alpha. We have bcopy(const void *...) so casting a const foo *bar
as in bcopy((caddr_t) bar,...) for bcopy seems to be fairly pointless in
the first place, but there are too many of these to hack out at the moment.

csapuntz@lcs.mit.edu: Port of NetBSD tulip driver. Works with new (21140-AC-based) SMC Etherpower cards.

These two was not ready for commit yet. Backed out.

Add Matrox Mystique

do SIOCGIFADDR in ether_ioctl

Adaption to new bus.h from NetBSD. Various driver has been updated to
NetBSD 961107 state. Other has not, partly because sometimes the bus.h
changes was coupled with other changes.

s/bus.h/bus.old.h/ to make the transit to the bus_space bus.h an easy road
possibly taken in small steps

__OpenBSD__ messup fixed

check for __OpenBSD__, prep for tossing __NetBSD__?

sync to 0607

sync 0521

if_name/if_unit -> if_xname/if_softc

sync with 0504 -- prototypes and bus.h

add general ether_ioctl call in net/if_ethersubr.c,
NS,IPX,X.25 special processing is now handled in there.
reflect this amazing addition in all the ether ifaces.
ppl, pls check the stuff.

partial sync with netbsd 960418, more to come

NetBSD 960317 merge

sc_dev -> tulip_dev

Pass device name to interrupt establish routines so it can be recorded in
the device interrupt chain structures (isa, pci)

Move interrupt chain structure definition to <machine/psl.h> so vmstat can
get at it (i386)

Remove hack to count interrupts the old way (i386)

from netbsd; merge latest Thomas version with our minimal changes

branches: 1.1.1;
Initial revision

Constify struct cfattach.

Delete unnecessary #includes of <sys/domain.h> and/or <sys/protosw.h>

net/if_pppx.c pointed out by jsg@
ok gnezdo@ deraadt@ jsg@ mpi@ millert@

spelling
feedback and ok tb@ jmc@ ok ratchov@

Do not clear IFF_UP, even in the error path, clearing IFF_RUNNING
is enough.

This flag should only be set by the stack, drivers shouldn't mess
with it.

Discussed with dlg@ and mikeb@, ok mikeb@, stsp@

move counting if_opackets next to counting if_obytes in if_enqueue.

this means packets are consistently counted in one place, unlike the
many and various ways that drivers thought they should do it.

ok mpi@ deraadt@

G/C IFQ_SET_READY().

No trailers has been the default and only option for 20 years, yet some
drivers still set IFF_NOTRAILERS while others do not. Remove all usage of
the flag from the drivers which in ancient times used it (and the modern
drivers which blindly copied it from those drivers of yore).
suggested by guenther. ok mpi