Commit Graph

67 Commits

Author SHA1 Message Date
Gleb Smirnoff
76039bc84f The r48589 promised to remove implicit inclusion of if_var.h soon. Prepare
to this event, adding if_var.h to files that do need it. Also, include
all includes that now are included due to implicit pollution via if_var.h

Sponsored by:	Netflix
Sponsored by:	Nginx, Inc.
2013-10-26 17:58:36 +00:00
Robert Watson
574644abac Use queue(9) instead of hand-crafted link lists for the global IPX
address list (ipx_ifaddr -> ipx_ifaddrhead), and generally adopt the
naming and usage conventions found in netinet.

MFC after:	6 weeks
2009-06-24 20:57:50 +00:00
Olivier Houchard
9d53036ff1 Include sys/lock.h before sys/rwlock.h. If anything used to bring it for us
before, it does not anymore.
2009-06-23 18:36:42 +00:00
Robert Watson
a0afe699ba Introduce basic locking of global IPX address list 'ipx_ifaddr' using
a new rwlock, ipx_ifaddr_rw, wrapped with macros.  This locking is
necessary but not sufficient, in isolation, to satisfy the stability
requirements of a fully parallel IPX input path during interface
reconfiguration.

MFC after:	3 weeks
2009-06-21 21:04:12 +00:00
Robert Watson
172aef7c6b Remove unuxed ipx_zerohost.
MFC after:	3 days
2009-06-21 16:11:40 +00:00
Robert Watson
941c2e1cd1 Remove historical support for capturing IPX packets in the output path
using raw IPX sockets.  While functional, this support is disabled
using a flag that can't be changed from userspace, and google reveals
no documentation or use of that flag anywhere.  This eliminates a
potential lock order reversal and code reentrance issue in which the
output path reentered the input path in IPX.

An alternative to removal would be to use the netisr, as a comment I
added in 2005 suggests.  While this change is fairly straight-forward,
the lack of any consumers or the easy possibility of consumers (kernel
modification and recompile required) suggests that this is simply an
unused feature.

Update README to remove this TODO, and a TODO regarding IPX/IP
encapsulation which was also removed a few years ago.

MFC after:	1 week
2009-06-21 10:10:44 +00:00
Robert Watson
d4b5cae49b Reimplement the netisr framework in order to support parallel netisr
threads:

- Support up to one netisr thread per CPU, each processings its own
  workstream, or set of per-protocol queues.  Threads may be bound
  to specific CPUs, or allowed to migrate, based on a global policy.

  In the future it would be desirable to support topology-centric
  policies, such as "one netisr per package".

- Allow each protocol to advertise an ordering policy, which can
  currently be one of:

  NETISR_POLICY_SOURCE: packets must maintain ordering with respect to
    an implicit or explicit source (such as an interface or socket).

  NETISR_POLICY_FLOW: make use of mbuf flow identifiers to place work,
    as well as allowing protocols to provide a flow generation function
    for mbufs without flow identifers (m2flow).  Falls back on
    NETISR_POLICY_SOURCE if now flow ID is available.

  NETISR_POLICY_CPU: allow protocols to inspect and assign a CPU for
    each packet handled by netisr (m2cpuid).

- Provide utility functions for querying the number of workstreams
  being used, as well as a mapping function from workstream to CPU ID,
  which protocols may use in work placement decisions.

- Add explicit interfaces to get and set per-protocol queue limits, and
  get and clear drop counters, which query data or apply changes across
  all workstreams.

- Add a more extensible netisr registration interface, in which
  protocols declare 'struct netisr_handler' structures for each
  registered NETISR_ type.  These include name, handler function,
  optional mbuf to flow ID function, optional mbuf to CPU ID function,
  queue limit, and ordering policy.  Padding is present to allow these
  to be expanded in the future.  If no queue limit is declared, then
  a default is used.

- Queue limits are now per-workstream, and raised from the previous
  IFQ_MAXLEN default of 50 to 256.

- All protocols are updated to use the new registration interface, and
  with the exception of netnatm, default queue limits.  Most protocols
  register as NETISR_POLICY_SOURCE, except IPv4 and IPv6, which use
  NETISR_POLICY_FLOW, and will therefore take advantage of driver-
  generated flow IDs if present.

- Formalize a non-packet based interface between interface polling and
  the netisr, rather than having polling pretend to be two protocols.
  Provide two explicit hooks in the netisr worker for start and end
  events for runs: netisr_poll() and netisr_pollmore(), as well as a
  function, netisr_sched_poll(), to allow the polling code to schedule
  netisr execution.  DEVICE_POLLING still embeds single-netisr
  assumptions in its implementation, so for now if it is compiled into
  the kernel, a single and un-bound netisr thread is enforced
  regardless of tunable configuration.

In the default configuration, the new netisr implementation maintains
the same basic assumptions as the previous implementation: a single,
un-bound worker thread processes all deferred work, and direct dispatch
is enabled by default wherever possible.

Performance measurement shows a marginal performance improvement over
the old implementation due to the use of batched dequeue.

An rmlock is used to synchronize use and registration/unregistration
using the framework; currently, synchronized use is disabled
(replicating current netisr policy) due to a measurable 3%-6% hit in
ping-pong micro-benchmarking.  It will be enabled once further rmlock
optimization has taken place.  However, in practice, netisrs are
rarely registered or unregistered at runtime.

A new man page for netisr will follow, but since one doesn't currently
exist, it hasn't been updated.

This change is not appropriate for MFC, although the polling shutdown
handler should be merged to 7-STABLE.

Bump __FreeBSD_version.

Reviewed by:	bz
2009-06-01 10:41:38 +00:00
Tom Rhodes
7521cd74f6 Begin the sysctl descriptions with a capital letter.
Make some slight wording tweaks.
2008-07-25 23:54:07 +00:00
Tom Rhodes
941b15a3e0 Document a few sysctls.
Reviewed by:	rwatson
2008-07-20 15:25:20 +00:00
Robert Watson
59dd72d040 Remove NETISR_MPSAFE, which allows specific netisr handlers to be directly
dispatched without Giant, and add NETISR_FORCEQUEUE, which allows specific
netisr handlers to always be dispatched via a queue (deferred).  Mark the
usb and if_ppp netisr handlers as NETISR_FORCEQUEUE, and explicitly
acquire Giant in those handlers.

Previously, any netisr handler not marked NETISR_MPSAFE would necessarily
run deferred and with Giant acquired.  This change removes Giant
scaffolding from the netisr infrastructure, but NETISR_FORCEQUEUE allows
non-MPSAFE handlers to continue to force deferred dispatch so as to avoid
lock order reversals between their acqusition of Giant and any calling
context.

It is likely we will be able to remove NETISR_FORCEQUEUE once
IFF_NEEDSGIANT is removed, as non-MPSAFE usb and if_ppp drivers will no
longer be supported.

Reviewed by:	bz
MFC after:	1 month
X-MFC note:	We can't remove NETISR_MPSAFE from stable/7 for KPI reasons,
		but the rest can go back.
2008-07-04 00:21:38 +00:00
Robert Watson
94ad7d68c7 Use ANSI C function declarations throughout netipx.
Remove 'register' use.
2007-05-11 10:38:34 +00:00
Robert Watson
28b2aee58f Factor out UCB and my copyrights from copyrights of Mike Mitchell;
the former use a three-clause BSD license (per UCB authorization
letter), whereas he uses a four-clause BSD license.

MFC after:	3 days
2007-01-08 22:14:00 +00:00
Robert Watson
1322b45c3d Canonicalize copyright order in one more file that contains my
copyright.

MFC after:	1 month
2006-03-27 01:12:58 +00:00
Robert Watson
6252bd331d Comment that raw output filter code for IPX should run in a netisr so as
to avoid recursing the socket code, as this input path can run in the
call stack of an output path.

MFC after:	1 month
2006-03-24 00:00:23 +00:00
Robert Watson
7d01b89631 Admit to ourselves that we don't actually implement pr_ctlinput() for
IPX or SPX, as the code in the implementing functions is essentially
a no-op.  Replace with a comment indicating we don't implement these
currently.
2006-03-23 19:50:00 +00:00
Robert Watson
bfcff7c78e Remove dead code associated with 'mcopy' in ipx_forward(): at no point
are the contents of the forwarded mbuf ever copied into mcopy, so there's
no need to have mcopy, conditionally look at mcopy, or conditionally free
it.

Noticed by:	Coverity Prevent analysis tool
MFC after:	3 days
2006-01-13 23:47:55 +00:00
Robert Watson
e1968df17f Compare (mbuf *) with NULL, not 0.
MFC after:	3 days
2005-04-10 18:05:02 +00:00
Robert Watson
66d165347d Mark the IPX netisr as MPSAFE so that inbound IPX traffic is processed
without Giant, and can be directly dispatched in the ithread when
net.isr.enable is turned on.

MFC after:	4 weeks
2005-01-09 07:34:55 +00:00
Robert Watson
e926c0ae48 Recent changes have locked down most of the highly dynamic data
structures in IPX/SPX -- primarily, sequence numbering, PCB lists,
and PCBs for IPX raw sockets, IPX datagram sockets, and IPX/SPX.
As such, remove remove NET_NEEDS_GIANT() for IPX, and remove the
assertion of Giant in the ipxintr() IPX input path.

Note that IPX/SPX is not fully MPSAFE, and that there are some
problems with IPX/SPX locking that will require some further work.
However, it is now safe enough to run in general without the Giant
lock.

MFC after:	4 weeks
2005-01-09 05:34:37 +00:00
Robert Watson
2082ca5d57 Use the IPX PCB list mutex and IPX PCB mutexes to lock down the SPX
portion of IPX/SPX:

- Protect IPX PCB lists with the IPX PCB list mutex, in particular
  when calling PCB and PCB list manipulation routines in ipx_pcb.c.
- Protect both IPX PCB state and SPX PCB state using the IPX PCB
  mutex.
- Generally annotate locking, as well as adding liberal use of lock
  assertions to document locking requirements.
- Where possible, use unlocked reads when reading integer or smaller
  sized socket options on SPX sockets.
- De-spl throughout.

Notes:

- spx_input() expects both the list mutex and PCB mutex to be held
  on entry, but will release both on return.  Because sonewconn() is
  called from spx_input(), it may actually drop one PCB lock and
  acquire another during generation of a new connection, meaning the
  caller is not in a position to unlock the PCB mutex.

MFC after:	3 weeks
2005-01-09 05:31:16 +00:00
Robert Watson
971365a711 Protect ipx_pexseq with the IPX PCB list mutex.
When processing socket options against IPX PCBs, generally protect
PCB fields using the IPX PCB mutex.  Where possible, use unlocked
reads on integer values to avoid locking overhead.

MFC after:	3 weeks
2005-01-09 05:15:59 +00:00
Robert Watson
31f1a840d9 Hold the IPX PCB mutex around calls to ipx_input() in the IPX input
path.

MFC after:	3 weeks
2005-01-09 05:08:47 +00:00
Robert Watson
992e1a5842 Hold the global IPX PCB list mutex in the IPX input path when walking
the IPX PCB list.

MFC after:	3 weeks
2005-01-09 05:06:19 +00:00
Robert Watson
c2b563b532 Introduce a global mutex, ipxpcb_list_mtx, to protect the global
IPX PCB lists.  Add macros to initialize, destroy, lock, unlock,
and assert the mutex.  Initialize the mutex when IPX is started.

Add per-IPX PCB mutexes, ipxp_mtx in struct ipxpcb, to protect
per-PCB IPX/SPX state.  Add macros to initialize, destroy, lock,
unlock, and assert the mutex.  Initialize the mutex when a new
PCB is allocated; destroy it when the PCB is free'd.

MFC after:	2 weeks
2005-01-09 05:00:41 +00:00
Robert Watson
f7bca2686a Eliminate jump to 'bad' label in order to clean up the ipx_input()
return/unwind path for locking work.

MFC after:	2 weeks
2005-01-09 04:39:16 +00:00
Warner Losh
c398230b64 /* -> /*- for license, minor formatting changes 2005-01-07 01:45:51 +00:00
Robert Watson
928944eeb5 Trim trailing whitespace. 2005-01-02 15:13:59 +00:00
Robert Watson
c2b8a29d33 Prefer rtalloc_ign() API to rtalloc() API. 2005-01-02 01:39:38 +00:00
Robert Watson
86c788d323 Move the definition of ipxpcb_lport_cache from ipx_input.c to ipx_pcb.c,
the only source file where it is actually used.
2005-01-01 22:04:03 +00:00
Robert Watson
96979ee67d Marginally reformat copyright statements to remove an excess ','. 2004-12-31 17:05:37 +00:00
Robert Watson
08e044cb89 Use a global variable, ipxpcb_lport_cache, to cache the most recently
used IPX port number, rather than using the global ipxpcb list head.
2004-12-30 17:54:53 +00:00
Robert Watson
80a4dabe7d Convert netipx to use queue(9) doubly-linked lists instead of home-brew
linked lists for ipxpcb's.
2004-12-30 17:49:40 +00:00
Robert Watson
05b4b08b61 Constify ipx_zeronet, ipx_zerohost, ipx_broadnet, ipx_broadhost.
Remove 'allones' since the values of the broadcast network and
host variables are set statically.
2004-12-30 16:56:07 +00:00
Sam Leffler
7902224c6b o add a flags parameter to netisr_register that is used to specify
whether or not the isr needs to hold Giant when running; Giant-less
  operation is also controlled by the setting of debug_mpsafenet
o mark all netisr's except NETISR_IP as needing Giant
o add a GIANT_REQUIRED assertion to the top of netisr's that need Giant
o pickup Giant (when debug_mpsafenet is 1) inside ip_input before
  calling up with a packet
o change netisr handling so swi_net runs w/o Giant; instead we grab
  Giant before invoking handlers based on whether the handler needs Giant
o change netisr handling so that netisr's that are marked MPSAFE may
  have multiple instances active at a time
o add netisr statistics for packets dropped because the isr is inactive

Supported by:	FreeBSD Foundation
2003-11-08 22:28:40 +00:00
Ruslan Ermilov
78f94aa951 Fix a bunch of off-by-one errors in the range checking code. 2003-09-11 21:40:21 +00:00
David E. O'Brien
ab0de15baf Use __FBSDID(). 2003-06-11 05:37:42 +00:00
Jonathan Lemon
1cafed3941 Update netisr handling; Each SWI now registers its queue, and all queue
drain routines are done by swi_net, which allows for better queue control
at some future point.  Packets may also be directly dispatched to a netisr
instead of queued, this may be of interest at some installations, but
currently defaults to off.

Reviewed by: hsu, silby, jayanth, sam
Sponsored by: DARPA, NAI Labs
2003-03-04 23:19:55 +00:00
Warner Losh
a163d034fa Back out M_* changes, per decision of the TRB.
Approved by: trb
2003-02-19 05:47:46 +00:00
Alfred Perlstein
44956c9863 Remove M_TRYWAIT/M_WAITOK/M_WAIT. Callers should use 0.
Merge M_NOWAIT/M_DONTWAIT into a single flag M_NOWAIT.
2003-01-21 08:56:16 +00:00
John Baldwin
6008862bc2 Change callers of mtx_init() to pass in an appropriate lock type name. In
most cases NULL is passed, but in some cases such as network driver locks
(which use the MTX_NETWORK_LOCK macro) and UMA zone locks, a name is used.

Tested on:	i386, alpha, sparc64
2002-04-04 21:03:38 +00:00
Mike Smith
bedbd47e6a Initialise the intrq_present fields at runtime, not link time. This allows
us to load protocols at runtime, and avoids the use of common variables.

Also fix the ip6_intrq assignment so that it works at all.
2002-01-08 10:34:03 +00:00
Poul-Henning Kamp
37d4006626 Another round of the <sys/queue.h> FOREACH transmogriffer.
Created with:   sed(1)
Reviewed by:    md5(1)
2001-02-04 16:08:18 +00:00
Poul-Henning Kamp
fc2ffbe604 Mechanical change to use <sys/queue.h> macro API instead of
fondling implementation details.

Created with: sed(1)
Reviewed by: md5(1)
2001-02-04 13:13:25 +00:00
Jonathan Lemon
df5e198723 Lock down the network interface queues. The queue mutex must be obtained
before adding/removing packets from the queue.  Also, the if_obytes and
if_omcasts fields should only be manipulated under protection of the mutex.

IF_ENQUEUE, IF_PREPEND, and IF_DEQUEUE perform all necessary locking on
the queue.  An IF_LOCK macro is provided, as well as the old (mutex-less)
versions of the macros in the form _IF_ENQUEUE, _IF_QFULL, for code which
needs them, but their use is discouraged.

Two new macros are introduced: IF_DRAIN() to drain a queue, and IF_HANDOFF,
which takes care of locking/enqueue, and also statistics updating/start
if necessary.
2000-11-25 07:35:38 +00:00
David E. O'Brien
b0e56cde37 * Use sys/sys/random.h rather than a i386 specific one.
* There was nothing that should be machine dependant about
  i386/isa/random_machdep.c, so it is now sys/kern/kern_random.c.
2000-04-24 17:30:08 +00:00
Peter Wemm
242c5536ea Clean up some loose ends in the network code, including the X.25 and ISO
#ifdefs.  Clean out unused netisr's and leftover netisr linker set gunk.
Tested on x86 and alpha, including world.

Approved by:	jkh
2000-02-13 03:32:07 +00:00
Brian Somers
367d34f853 Move the *intrq variables into net/intrq.c and unconditionally
include this in all kernels.  Declare some const *intrq_present
variables that can be checked by a module prior to using *intrq
to queue data.

Make the if_tun module capable of processing atm, ip, ip6, ipx,
natm and netatalk packets when TUNSIFHEAD is ioctl()d on.

Review not required by: freebsd-hackers
2000-01-24 20:39:02 +00:00
John Hay
01d3efc077 Get rid of the old XNS checksum code and implement it the IPX way.
PR:		13374
Submitted by:	Boris Popov <bp@butya.kz>
1999-08-28 18:21:55 +00:00
Peter Wemm
c3aac50f28 $Id$ -> $FreeBSD$ 1999-08-28 01:08:13 +00:00
Poul-Henning Kamp
1de24c3a4e When system act as IPX router (sysctl -w net.ipx.ipx.ipxforwarding=1,
running IPXRouted -s) between IPX configured interfaces, it generate
syslog messages "ipx_ctlinput: cmd 15." even if kernel compiled with
IPXPRINTFS=0 and IPX_ERRPRINTFS=0 options.

PR:		6875
Reviewed by:	phk
Submitted by:	Vladimir A. Jakovenko <vovik@ntu-kpi.kiev.ua>
1998-06-10 12:52:11 +00:00