branches: 1.1.2; 1.1.10; 1.1.16;
Add the basic rump networking library, which provides e.g. sockets
and mbufs. This is required by all network-using rump applications.

netisr.c is gone.

Make pktq_rps_hash() pluggable for each interface type. Reviewed by gdt@n.o, thorpej@n.o, and riastradh@n.o, thanks.

Merge uipc_mbuf2.c into uipc_mbuf.c. Reorder the latter a little to gather
similar functions. No functional change.

branches: 1.22.10; 1.22.12;
Rumpify netipsec

Note that we should modularize netipsec and reduce reverse symbol references
(referencing symbols of netipsec from net, netinet and netinet6) though,
the task needs lots of code changes. Prior to doing so, rumpifying it and
having ATF tests should be useful.

Defer some pr_input to workqueue

pr_input is currently called in softint. Some pr_input such as ICMP, ICMPv6
and CARP can add/delete/update IP addresses and routing table entries. For
example, icmp6_redirect_input updates an a routing table entry and
nd6_ra_input may delete an IP address.

Basically such operations shouldn't be done in softint. That aside, we have
a reason to avoid the situation; psz/psref waits cannot be used in softint,
however they are required to work in such pr_input in the MP-safe world.

The change implements the workqueue pr_input framework called wqinput which
provides a means to defer pr_input of a protocol to workqueue easily.
Currently icmp_input, icmp6_input, carp_proto_input and carp6_proto_input
are deferred to workqueue by the framework.

Proposed and discussed on tech-kern and tech-net

Fix build (undefined reference to `rumpns_pfil_init' on usr.sbin/puffs/rump_nfs)

branches: 1.19.2; 1.19.4;
Add a COMMENT describing what each component roughly does.

"make describe" prints the comment.

Requested/inspired by Vincent Schwarzer on rumpkernel-users

remove librump/rumpnet/opt, consolidate in rump/include/opt

add printers.

Remove shlib_version files and just use Makefile SHLIB_MAJOR/MINOR,
with the default provided by Makefile.rump (they're all 0.0 anyway)

branches: 1.15.4;
- 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.

Eliminate weak symbols from rump kernel syscall handlers, part 7:

Build component constructors which establish syscalls at boottime.

branches: 1.13.6; 1.13.10;
Allow Makefile.rump to append to SRCS.

branches: 1.12.4; 1.12.14;
Hide the radix-trie implementation of the forwarding table so that we
will have an easier time replacing it with something different, even if
it is a second radix-trie implementation.

sys/net/route.c and sys/net/rtsock.c no longer operate directly on
radix_nodes or radix_node_heads.

Hopefully this will reduce the temptation to implement multipath or
source-based routing using grotty hacks to the grotty old radix-trie
code, too. :-)

Add a new AF/PF_ROUTE which is 64-bit clean which makes the routing socket
interface (and its associated sysctls) act identically for both 32 and 64 bit
programs. The old unclean one remains for backward compatibility.

branches: 1.10.4; 1.10.6; 1.10.8;
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.

We have real pollsuck() now, so remove this, ahem, less-than-perfect one.
(XXX: only thing in the kernel using pollsuck is netsmb)

Adjust rump sources for external/internal interfaces.
No functional change.

add a very simple version of pollsock()
XXX: it seems to exist purely for the pleasure of netsmb??

MAXUSERS comes from Makefile.rump

libkern is fully included in rumpkern, no need for cherry-picking
modules elsewhere.

branches: 1.4.2; 1.4.4;
include rtsock_50.c

branches: 1.3.4;
fix comment (no functional change)

Heave-ho radix.c from librumpnet_net to librumpnet.

branches: 1.1.2; 1.1.4;
Add the basic rump networking library, which provides e.g. sockets
and mbufs. This is required by all network-using rump applications.

We have real pollsuck() now, so remove this, ahem, less-than-perfect one.
(XXX: only thing in the kernel using pollsuck is netsmb)

branches: 1.1.2;
add a very simple version of pollsock()
XXX: it seems to exist purely for the pleasure of netsmb??

Move input processing of lagg(4) before ether_input
to get rid of dependence.

This implementation is similar with that of bridge(4).

vlan: Register vlan_ifdetach to ether_ifdetach hook

bridge: Register bridge_ifdetach to ether_ifdetach hook

vlan: Register the callback to update link-state of vlan I/F
to link-state change hook

The callback is registered in every vlan I/F even if the parent
interface is the same. Therefore it is not needed to search the
vlan I/F by the parent interface unlike the previous callback.

lagg: Register lagg_linkstate_changed to link-state change hook

bridge: Register bridge_calc_link_state to link-state change hook

Rump-ify ALTQ (librumpnet_altq.so)

Add a new link-aggregation pseudo interface named lagg(4)

- FreeBSD's lagg(4) based implementation
- MP-safe and MP-scalable

branches: 1.41.4; 1.41.6;
PR/55777: Ruslan Nikolaev: Move the unp_sysctl_create to uipc_usrreq.c to
facilitate splitting rump modules and does not require a dummy function.

branches: 1.40.2;
rump: Add weak alias for bridge_calc_link_state

rump: Try to fix build

ipsec: rename ipsec_ip_input to ipsec_ip_input_checkpolicy

Because it just checks if a packet passes security policies.

PR kern/53562

Handle TX offload in software when a packet is sent via
bridge_output(). We can send it as is in the following
exceptional cases:

For unicast:

(1) When the destination interface is the same as source.

(2) When the destination supports all TX offload options
specified in a packet.

For multicast/broadcast:

(3) When all the members of the bridge support the specified
TX offload options.

For (3), add sc_csum_flags_tx flag to bridge softc, which is
logical AND b/w capabilities of TX offload options in member
interface (ifp->if_csum_flags_tx). The flag is updated when a
member is (i) added to or (ii) removed from a bridge, or (iii)
if_csum_flags_tx flag of a member interface is manipulated via
ifconfig(8).

Turn on M_CSUM_TSOv[46] bit in ifp->if_csum_flags_tx flag when
TSO[46] is enabled for that interface.

OK msaitoh thorpej

branches: 1.36.2;
Merge ipsec4_input and ipsec6_input into ipsec_ip_input. Make the argument
a bool for clarity. Optimize the function: if M_CANFASTFWD is not there
(because already removed by the firewall) leave now.

Makes it easier to see that M_CANFASTFWD is not removed on IPv6.

Rename ipsec4_forward -> ipsec_mtu, and switch to void.

add an empty stub

Stop using a macro, rename the function to ipsec_init_pcbpolicy directly.

remove now unused crap.

branches: 1.31.2;
Dedup: merge ipsec4_set_policy and ipsec6_set_policy. The content of the
original ipsec_set_policy function is inlined into the new one.

Oops, forgot this file; I just merged two IPsec functions, so adapt
the rump stubs accordingly.

Dedup: merge ipsec4_in_reject and ipsec6_in_reject into ipsec_in_reject.
While here fix misleading comment.

ok ozaki-r@

Dedup: merge ipsec4_hdrsiz and ipsec6_hdrsiz into ipsec_hdrsiz.

ok ozaki-r@

Make IPsec SPD MP-safe

We use localcount(9), not psref(9), to make the sptree and secpolicy (SP)
entries MP-safe because SPs need to be referenced over opencrypto
processing that executes a callback in a different context.

SPs on sockets aren't managed by the sptree and can be destroyed in softint.
localcount_drain cannot be used in softint so we delay the destruction of
such SPs to a thread context. To do so, a list to manage such SPs is added
(key_socksplist) and key_timehandler_spd deletes dead SPs in the list.

For more details please read the locking notes in key.c.

Proposed on tech-kern@ and tech-net@

branches: 1.26.4;
Rumpify netipsec

Note that we should modularize netipsec and reduce reverse symbol references
(referencing symbols of netipsec from net, netinet and netinet6) though,
the task needs lots of code changes. Prior to doing so, rumpifying it and
having ATF tests should be useful.

Use if_acquire and if_release instead of using psref API directly

- Provide if_release for consistency to if_acquire
- Use if_acquire and if_release for ifp iterations
- Make ifnet_psref_class static

branches: 1.24.2;
Rumpify vlan(4)

branches: 1.23.2;
Protect ifnet list with psz and psref

The change ensures that ifnet objects in the ifnet list aren't freed during
list iterations by using pserialize(9) and psref(9).

Note that the change adds a pslist(9) for ifnet but doesn't remove the
original ifnet list (ifnet_list) to avoid breaking kvm(3) users. We
shouldn't use the original list in the kernel anymore.

Rump-ify if_pppoe

From s-yamaguchi@IIJ

Get rid of weak_alias for bridge_input

bridge_input is now not directly called from ether_input,
so we can remove weak_alias for it.

Revert "Pull if_drain routine out of m_reclaim"

The commit broke dlopen()'d rumpnet on platforms where ld.so does not
override weak aliases (e.g. musl, Solaris, potentially OS X, ...).

Requested by pooka@.

branches: 1.19.2;
Pull if_drain routine out of m_reclaim

It's if-specific and should be in if.c.

No functional change.

Fix RUMP build.

don't invade the __toolchain namespace

Move a few stubs to where they really belong; allows us to create
them as strong instead of weak symbols.

kill weak aliases which are no longer necessary

branches: 1.14.2; 1.14.6;
Remove the route_info weak alias. It's unnecessary since 2011
and confused Linux ld.so, thus preventing dlopen() of librumpnet.
Reported in private email by Justin Cormack.

Also, g/c a few other stubs which are no longer necessary.

branches: 1.13.2;
Remove stubs which have been provided elsewhere for almost 2 years now.

branches: 1.12.4;
Add a new AF/PF_ROUTE which is 64-bit clean which makes the routing socket
interface (and its associated sysctls) act identically for both 32 and 64 bit
programs. The old unclean one remains for backward compatibility.

branches: 1.11.2; 1.11.4;
weak stubs for agr(4)

add weak stubs for bridge (not the game)

Don't alias the same symbol twice.

branches: 1.8.4; 1.8.8;
Add a few stubs to allow to compile - i'll have to sort these out later.

branches: 1.7.2;
Provide ifunit stub in the same place as others.

__weak_alias some more networking symbols for the benefit of the
nfs boot code in case not using the full networking stack.

__KERNEL_RCSID

branches: 1.4.2; 1.4.4;
...... except that the compat converter is actually supposed to convert
the ioctl, not return random garbage. It returning random garbage
causes an interface with an interesting netmask to get set and
hence a very interesting effect when you try to add a route for
your system.

All in all, a very healthy debugging session (if you're a vampire
or a wraith).

... except that compat_ifioctl() must know how to route the
commands further.

Add a few compat stubs. They shouldn't be required at all, but I
don't currently have the time to wage a holy war against how the
kernel decides it wants them.

Add the basic rump networking library, which provides e.g. sockets
and mbufs. This is required by all network-using rump applications.

Garbage-collect the remaining vestiges of netisr.

Put the kernelside rump kernel headers into <rump-sys> instead of
sprinkling them around the faction directories. Avoids having
to add a CPPFLAGS (or several) to pretty much every component
Makefile.

Leave compat headers around in the old locations.

The commit changes some autogenerated files, but I'll fix the
generators shortly and regen.

branches: 1.9.4;
- 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.

branches: 1.8.2;
Schedule only those netisr's that have registered handlers.

Register netisr's from component constructors instead of via a hardcoded
global list.

Add librumpnet_netmpls that provides MPLS features into rump kernels
ok'ed pooka@

branches: 1.5.8; 1.5.18; 1.5.22; 1.5.30;
Refetch netisr function pointers with dlsym(). This gives the
desired values in case the components containing the netisr handlers
were not linked in but dlopen()'d before calling rump_init().

(could simplify a little in case static linking is declared dead)

branches: 1.4.4;
Support IPv6 in rump. I'd have liked to introduce a netinet6
component, but due to ifdef happiness permeating the sources, it's
a compile decision for now, so netinet pulls in both inet and inet6.

One issue, one single issue: the loopback interface still needs to
be created for IPv6 to work. I have patches to take care of it
automatically if the appropriate component (net) is present, but
they require a bit more testing before commit.

branches: 1.3.2;
Ansify function definitions w/o arguments. Generated with sed.

branches: 1.2.2; 1.2.4;
__KERNEL_RCSID

branches: 1.1.2; 1.1.4;
Add the basic rump networking library, which provides e.g. sockets
and mbufs. This is required by all network-using rump applications.

Don't start callouts for domains before attaching domains on rump kernels

On rump kernels, the callouts for domains, pffasttimo and pfslowtimo, started
before domains were attached. Normally the callouts were dispatched after
domain attaches (initializations) finished, however, under load the callouts
could be executed prior to the attaches, resulting in that the callouts accessed
unallocated or uninitialized resources.

branches: 1.20.6;
call pfil_init() here.

branches: 1.19.2; 1.19.4;
Sweep unncessary radix.h inclusions

Put the kernelside rump kernel headers into <rump-sys> instead of
sprinkling them around the faction directories. Avoids having
to add a CPPFLAGS (or several) to pretty much every component
Makefile.

Leave compat headers around in the old locations.

The commit changes some autogenerated files, but I'll fix the
generators shortly and regen.

branches: 1.17.6;
Register netisr's from component constructors instead of via a hardcoded
global list.

branches: 1.16.2;
Remove routines which don't make sense (maybe they're some intermediate
phase of the component framework?)

Set networking soft interrupt vectors earlier, just in case interface
component initialization causes network traffic to be received.

Don't use __weak_alias where the component framework is sufficient.

branches: 1.13.8; 1.13.18;
Add one more component level to networking: IFCFG.
It is executed after IF and the purposes to guarantee the right
order in cross-component interface address configuration.
(e.g. lo0 is attached by net but 127.0.0.1 is configured by netinet)

branches: 1.12.2;
Introduce RUMP_COMPONENT. It behaves mostly like a simplified
module which is linked into the kernel and cannot be unloaded.
The main purpose is to get the proper constructors run and create
any /dev nodes necessary for said component. Once more of the
kernel (e.g. networking stack and device drivers) are converted to
MODULE and devfs pops up from somewhere, rump components can be
retired.

branches: 1.11.2;
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 suckets before domains since some domains install timers
which take softnet_lock and might run before the lock is actually
initialized. Also, soinit() itself already calls soinit2(), so no
need to call it twice.

create interfaces only after ifnef is initialized

Use a bunch of weak symbols to determine which network components
are present. This works in userspace as opposed relying in link
sets, which fail miserably. Later, when the networking stack
becomes modularized, we can move to a dynamic scheme like with file
systems.

Also, this change allows us to do proper autoconfig, namely attach
the loopback interface iff it is present.

branches: 1.7.2;
Ansify function definitions w/o arguments. Generated with sed.

branches: 1.6.2; 1.6.4;
Provide ifunit stub in the same place as others.

__KERNEL_RCSID

"Manually" call rn_init() to make sure it's done. The real kernel
hooks it up with domain constructors in a fascinating way.

branches: 1.3.2; 1.3.4;
Before I sort all the COMPAT_RIDICULOUSLYPREHISTORIC ioctl stuff
out, add a weak alias for ifunit() so that utils can be linked with
sockin.

call rump_netisr_init()

Add the basic rump networking library, which provides e.g. sockets
and mbufs. This is required by all network-using rump applications.

Put the kernelside rump kernel headers into <rump-sys> instead of
sprinkling them around the faction directories. Avoids having
to add a CPPFLAGS (or several) to pretty much every component
Makefile.

Leave compat headers around in the old locations.

The commit changes some autogenerated files, but I'll fix the
generators shortly and regen.

branches: 1.10.2;
Nuke the DOMAINADD() macro and just call domain_attach(), now that things
work correctly that way.

Register netisr's from component constructors instead of via a hardcoded
global list.

g/c prototype of imaginary routine

g/c dummyif, not useful these days

branches: 1.6.2;
Don't use __weak_alias where the component framework is sufficient.

branches: 1.5.10; 1.5.20;
Introduce RUMP_COMPONENT. It behaves mostly like a simplified
module which is linked into the kernel and cannot be unloaded.
The main purpose is to get the proper constructors run and create
any /dev nodes necessary for said component. Once more of the
kernel (e.g. networking stack and device drivers) are converted to
MODULE and devfs pops up from somewhere, rump components can be
retired.

branches: 1.4.2;
Adjust rump sources for external/internal interfaces.
No functional change.

Use a bunch of weak symbols to determine which network components
are present. This works in userspace as opposed relying in link
sets, which fail miserably. Later, when the networking stack
becomes modularized, we can move to a dynamic scheme like with file
systems.

Also, this change allows us to do proper autoconfig, namely attach
the loopback interface iff it is present.

Add a dummyif, which doesn't actually traffic any cargo, but since
it has no backend it can always be attached and is therefore
convenient for testing ifconfig.

branches: 1.1.2; 1.1.10; 1.1.12; 1.1.16;
Add the basic rump networking library, which provides e.g. sockets
and mbufs. This is required by all network-using rump applications.

generate privhdrs to new location

branches: 1.3.8;
Having a system-supplied function interface to create virt(4) is not
fully useful when the system doesn't supply anything to configure the
interface with. So remove rump_virtif_create(), that functionality
is now taken care of with ifcloner.

branches: 1.2.4; 1.2.14; 1.2.24; 1.2.28;
add shmif_create

Create rump public interfaces from description tables. This allows
us to control and wrap all entry points from "userspace" into rump.
This in turn is necessary for the upcoming rump cpu scheduler.

For each interface "foo" a public wrapper called "rump_foo" is
created. It calls the internal implementation "rumppriv_foo". In
case foo is to be called from inside of rump kernel space, the
private interface "rumppriv_foo" is used -- the userspace wrapper
prototypes are not even exported into the rump kernel namespace.
Needless to say, the rump kernel internal interfaces are not exported
for users.

Now, three classes of interfaces fight for control of rump:
+ the noble local control interfaces (which this commit addresses)
+ the insidious rump system calls (which are generated from syscalls.master)
+ and the evil vnode interfaces (which are generated from vnode_if.src)

Put the kernelside rump kernel headers into <rump-sys> instead of
sprinkling them around the faction directories. Avoids having
to add a CPPFLAGS (or several) to pretty much every component
Makefile.

Leave compat headers around in the old locations.

The commit changes some autogenerated files, but I'll fix the
generators shortly and regen.

branches: 1.5.8;
regen

branches: 1.4.4; 1.4.14; 1.4.24; 1.4.28;
regen

regen: scheduling points in rump_pub wrappers

regen: put all public interfaces created by ifspec into a rump_pub namespace

generate rump local interfaces

regen rump kernel interfaces for header change.
(they were already manually edited for a prior commit, so not much change)

Put the kernelside rump kernel headers into <rump-sys> instead of
sprinkling them around the faction directories. Avoids having
to add a CPPFLAGS (or several) to pretty much every component
Makefile.

Leave compat headers around in the old locations.

The commit changes some autogenerated files, but I'll fix the
generators shortly and regen.

branches: 1.5.8;
regen

branches: 1.4.4; 1.4.14; 1.4.24; 1.4.28;
regen

regen: scheduling points in rump_pub wrappers

regen: put all public interfaces created by ifspec into a rump_pub namespace

generate rump local interfaces

regen

regen

rump: Regen again to fix missing __kevent100 symbol.

makerumpsyscalls.sh requires rebuilt rump libraries installed in
/usr/lib or directory specified by its first argument.

Now, new failures observed for ATF after __kevent100 addition are
fixed at least for aarch64.

Regen

regen for lpathconf.

regen

branches: 1.4.18;
Put the kernelside rump kernel headers into <rump-sys> instead of
sprinkling them around the faction directories. Avoids having
to add a CPPFLAGS (or several) to pretty much every component
Makefile.

Leave compat headers around in the old locations.

The commit changes some autogenerated files, but I'll fix the
generators shortly and regen.

regen

branches: 1.2.2; 1.2.4; 1.2.6; 1.2.10; 1.2.12;
regen

Eliminate weak symbols from rump kernel syscall handlers, part 6:

Generate component constructors which establish non-modular syscalls.