Make ifq and ifiq interface MP safe.

Rename ifq_set_maxlen() to ifq_init_maxlen(). This function neither
uses WRITE_ONCE() nor a mutex and is called before the ifq mutex
is initialized. The new na

Make ifq and ifiq interface MP safe.

Rename ifq_set_maxlen() to ifq_init_maxlen(). This function neither
uses WRITE_ONCE() nor a mutex and is called before the ifq mutex
is initialized. The new name expresses that it should be used only
during interface attach when there is no concurrency.

Protect ifq_len(), ifq_empty(), ifiq_len(), and ifiq_empty() with
READ_ONCE(). They can be used without lock as they only read a
single integer.

OK dlg@

show more ...

Protect interface media list with a mutex. This is just a start
to make make media structures MP safe.
OK mvs@

Rename the macro MCLGETI to MCLGETL and removes the dead parameter ifp.

OK dlg@, bluhm@
No Opinion mpi@
Not against it claudio@

Change users of IFQ_SET_MAXLEN() and IFQ_IS_EMPTY() to use the "new" API.

ok dlg@ tobhe@

use ifiq_input and use it's return value to apply backpressure to rxrs.

this is a step toward deprecating softclock based livelock detection.

Sporadically the network over gem(4) interface hang on sparc64 and
macppc. Receiving packets stopped, ifconfig down/up made it work
again.
In the tick timeout handler refill the receive ring if it i

Sporadically the network over gem(4) interface hang on sparc64 and
macppc. Receiving packets stopped, ifconfig down/up made it work
again.
In the tick timeout handler refill the receive ring if it is empty.
The logic is taken from hme(4). Also protect the register access
and ifp counters with splnet().
In gem_rx_watchdog() is a workaround for a hardware bug. It resets
the hardware when there is no progress. If the fifo pointers
advanced a bit, it got stuck anyway. So restart the receive watchdog
timeout in that case.
OK mpi@

show more ...

make the gem tx path MPSAFE.

this mostly follows the pattern in ifq.h

ok jmatthew@

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.

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@

show more ...

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

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

show more ...

Rework gem_start to check that there's enough space in the ring before
dequeueing a packet, then drop the packet if it can't be sent, rather than
using two phase dequeue then commit.

Adjust sc_tx_cn

Rework gem_start to check that there's enough space in the ring before
dequeueing a packet, then drop the packet if it can't be sent, rather than
using two phase dequeue then commit.

Adjust sc_tx_cnt using atomic operations, take the kernel lock before
calling gem_start or gem_init from the interrupt handler, and use an
interrupt barrier when taking the interface down.

With all this done, we can mark the interrupt handler for pci (but not sbus)
attached gem(4) as mpsafe.

mpi@ wrote this initially, I fixed bugs to keep sparc64 happy and kept it up
to date with ifq changes.

ok mpi@ dlg@

show more ...

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 i

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@

show more ...

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

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

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@

show more ...

arp_ifinit() is no longer needed.

Make room for media types of the future. Extend the ifmedia word to 64 bits.
This changes numbers of the SIOCSIFMEDIA and SIOCGIFMEDIA ioctls and
grows struct ifmediareq.

Old ifconfig and dhclient b

Make room for media types of the future. Extend the ifmedia word to 64 bits.
This changes numbers of the SIOCSIFMEDIA and SIOCGIFMEDIA ioctls and
grows struct ifmediareq.

Old ifconfig and dhclient binaries can still assign addresses, however
the 'media' subcommand stops working. Recompiling ifconfig and dhclient
with new headers before a reboot should not be necessary unless in very
special circumstances where non-default media settings must be used to
get link and console access is not available.

There may be some MD fallout but that will be cleared up later.

ok deraadt miod
with help and suggestions from several sharks attending l2k15

show more ...

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@

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@

tweak the new if_input function so it takes an mbuf_list instead
of a single mbuf. this forces us to batch work between the hardware
rx handlers and the stack.

this includes a converstion of bge fro

tweak the new if_input function so it takes an mbuf_list instead
of a single mbuf. this forces us to batch work between the hardware
rx handlers and the stack.

this includes a converstion of bge from ether_input to if_input.

ok claudio@ pelikan@ mpi@

show more ...

Convert to if_input().

ok pelikan@, reyk@, blambert@, henning@

unifdef INET

Use <sys/endian.h> instead of <machine/endian.h>

ok dlg@ mpi@ bcook@ millert@ miod@

rxr ioctl handling.

Fewer <netinet/in_systm.h>

cut things that relied on mclgeti for rx ring accounting/restriction over
to using if_rxr.

cut the reporting systat did over to the rxr ioctl.

tested as much as i can on alpha, amd64, and sparc64.

cut things that relied on mclgeti for rx ring accounting/restriction over
to using if_rxr.

cut the reporting systat did over to the rxr ioctl.

tested as much as i can on alpha, amd64, and sparc64.
mpi@ has run it on macppc.
ok mpi@

show more ...