See the following descriptions for details.
Proposed on tech-kern and tech-net
Overview
--------
We protect the routing table with a rwock and protect
rtcaches with another rwlock. Each rtentry is protected
from being freed or updated via reference counting and psref.
Global rwlocks
--------------
There are two rwlocks; one for the routing table (rt_lock) and
the other for rtcaches (rtcache_lock). rtcache_lock covers
all existing rtcaches; there may have room for optimizations
(future work).
The locking order is rtcache_lock first and rt_lock is next.
rtentry references
------------------
References to an rtentry is managed with reference counting
and psref. Either of the two mechanisms is used depending on
where a rtentry is obtained. Reference counting is used when
we obtain a rtentry from the routing table directly via
rtalloc1 and rtrequest{,1} while psref is used when we obtain
a rtentry from a rtcache via rtcache_* APIs. In both cases,
a caller can sleep/block with holding an obtained rtentry.
The reasons why we use two different mechanisms are (i) only
using reference counting hurts the performance due to atomic
instructions (rtcache case) (ii) ease of implementation;
applying psref to APIs such rtaloc1 and rtrequest{,1} requires
additional works (adding a local variable and an argument).
We will finally migrate to use only psref but we can do it
when we have a lockless routing table alternative.
Reference counting for rtentry
------------------------------
rt_refcnt now doesn't count permanent references such as for
rt_timers and rtcaches, instead it is used only for temporal
references when obtaining a rtentry via rtalloc1 and rtrequest{,1}.
We can do so because destroying a rtentry always involves
removing references of rt_timers and rtcaches to the rtentry
and we don't need to track such references. This also makes
it easy to wait for readers to release references on deleting
or updating a rtentry, i.e., we can simply wait until the
reference counter is 0 or 1. (If there are permanent references
the counter can be arbitrary.)
rt_ref increments a reference counter of a rtentry and rt_unref
decrements it. rt_ref is called inside APIs (rtalloc1 and
rtrequest{,1} so users don't need to care about it while
users must call rt_unref to an obtained rtentry after using it.
rtfree is removed and we use rt_unref and rt_free instead.
rt_unref now just decrements the counter of a given rtentry
and rt_free just tries to destroy a given rtentry.
See the next section for destructions of rtentries by rt_free.
Destructions of rtentries
-------------------------
We destroy a rtentry only when we call rtrequst{,1}(RTM_DELETE);
the original implementation can destroy in any rtfree where it's
the last reference. If we use reference counting or psref, it's
easy to understand if the place that a rtentry is destroyed is
fixed.
rt_free waits for references to a given rtentry to be released
before actually destroying the rtentry. rt_free uses a condition
variable (cv_wait) (and psref_target_destroy for psref) to wait.
Unfortunately rtrequst{,1}(RTM_DELETE) can be called in softint
that we cannot use cv_wait. In that case, we have to defer the
destruction to a workqueue.
rtentry#rt_cv, rtentry#rt_psref and global variables
(see rt_free_global) are added to conduct the procedure.
Updates of rtentries
--------------------
One difficulty to use refcnt/psref instead of rwlock for rtentry
is updates of rtentries. We need an additional mechanism to
prevent readers from seeing inconsistency of a rtentry being
updated.
We introduce RTF_UPDATING flag to rtentries that are updating.
While the flag is set to a rtentry, users cannot acquire the
rtentry. By doing so, we avoid users to see inconsistent
rtentries.
There are two options when a user tries to acquire a rtentry
with the RTF_UPDATING flag; if a user runs in softint context
the user fails to acquire a rtentry (NULL is returned).
Otherwise a user waits until the update completes by waiting
on cv.
The procedure of a updater is simpler to destruction of
a rtentry. Wait on cv (and psref) and after all readers left,
proceed with the update.
Global variables (see rt_update_global) are added to conduct
the procedure.
Currently we apply the mechanism to only RTM_CHANGE in
rtsock.c. We would have to apply other codes. See
"Known issues" section.
psref for rtentry
-----------------
When we obtain a rtentry from a rtcache via rtcache_* APIs,
psref is used to reference to the rtentry.
rtcache_ref acquires a reference to a rtentry with psref
and rtcache_unref releases the reference after using it.
rtcache_ref is called inside rtcache_* APIs and users don't
need to take care of it while users must call rtcache_unref
to release the reference.
struct psref and int bound that is needed for psref is
embedded into struct route. By doing so we don't need to
add local variables and additional argument to APIs.
However this adds another constraint to psref other than
reference counting one's; holding a reference of an rtentry
via a rtcache is allowed by just one caller at the same time.
So we must not acquire a rtentry via a rtcache twice and
avoid a recursive use of a rtcache. And also a rtcache must
be arranged to be used by a LWP/softint at the same time
somehow. For IP forwarding case, we have per-CPU rtcaches
used in softint so the constraint is guaranteed. For a h
rtcache of a PCB case, the constraint is guaranteed by the
solock of each PCB. Any other cases (pf, ipf, stf and ipsec)
are currently guaranteed by only the existence of the global
locks (softnet_lock and/or KERNEL_LOCK). If we've found the
cases that we cannot guarantee the constraint, we would need
to introduce other rtcache APIs that use simple reference
counting.
psref of rtcache is created with IPL_SOFTNET and so rtcache
shouldn't used at an IPL higher than IPL_SOFTNET.
Note that rtcache_free is used to invalidate a given rtcache.
We don't need another care by my change; just keep them as
they are.
Performance impact
------------------
When NET_MPSAFE is disabled the performance drop is 3% while
when it's enabled the drop is increased to 11%. The difference
comes from that currently we don't take any global locks and
don't use psref if NET_MPSAFE is disabled.
We can optimize the performance of the case of NET_MPSAFE
on by reducing lookups of rtcache that uses psref;
currently we do two lookups but we should be able to trim
one of two. This is a future work.
Known issues
------------
There are two known issues to be solved; one is that
a caller of rtrequest(RTM_ADD) may change rtentry (see rtinit).
We need to prevent new references during the update. Or
we may be able to remove the code (perhaps, need more
investigations).
The other is rtredirect that updates a rtentry. We need
to apply our update mechanism, however it's not easy because
rtredirect is called in softint and we cannot apply our
mechanism simply. One solution is to defer rtredirect to
a workqueue but it requires some code restructuring.
We no longer need to change rtentry below if_output.
The change makes it clear where rtentries are changed (or not)
and helps forthcoming locking (os psrefing) rtentries.
By this change, nexthop caches (IP-MAC address pair) are not stored
in the routing table anymore. Instead nexthop caches are stored in
each network interface; we already have lltable/llentry data structure
for this purpose. This change also obsoletes the concept of cloning/cloned
routes. Cloned routes no longer exist while cloning routes still exist
with renamed to connected routes.
Noticeable changes are:
- Nexthop caches aren't listed in route show/netstat -r
- sysctl(NET_RT_DUMP) doesn't return them
- If RTF_LLDATA is specified, it returns nexthop caches
- Several definitions of routing flags and messages are removed
- RTF_CLONING, RTF_XRESOLVE, RTF_LLINFO, RTF_CLONED and RTM_RESOLVE
- RTF_CONNECTED is added
- It has the same value of RTF_CLONING for backward compatibility
- route's -xresolve, -[no]cloned and -llinfo options are removed
- -[no]cloning remains because it seems there are users
- -[no]connected is introduced and recommended
to be used instead of -[no]cloning
- route show/netstat -r drops some flags
- 'L' and 'c' are not seen anymore
- 'C' now indicates a connected route
- Gateway value of a route of an interface address is now not
a L2 address but "link#N" like a connected (cloning) route
- Proxy ARP: "arp -s ... pub" doesn't create a route
You can know details of behavior changes by seeing diffs under tests/.
Proposed on tech-net and tech-kern:
http://mail-index.netbsd.org/tech-net/2016/03/11/msg005701.html
rt_refcnt of rtentry was used in bad manners, for example, direct rt_refcnt++
and rt_refcnt-- outside route.c, "rt->rt_refcnt++; rtfree(rt);" idiom, and
touching rt after rt->rt_refcnt--.
These abuses seem to be needed because rt_refcnt manages only references
between rtentry and doesn't take care of references during packet processing
(IOW references from local variables). In order to reduce the above abuses,
the latter cases should be counted by rt_refcnt as well as the former cases.
This change improves consistency of use of rt_refcnt:
- rtentry is always accessed with rt_refcnt incremented
- rtentry's rt_refcnt is decremented after use (rtfree is always used instead
of rt_refcnt--)
- functions returning rtentry increment its rt_refcnt (and caller rtfree it)
Note that rt_refcnt prevents rtentry from being freed but doesn't prevent
rtentry from being updated. Toward MP-safe, we need to provide another
protection for rtentry, e.g., locks. (Or introduce a better data structure
allowing concurrent readers during updates.)
and dom_sa_len members from struct domain. Pools of fixed-size
objects are too rigid for sockaddr_dls, whose size can vary over
a wide range.
Return sockaddr_dl to its "historical" size. Now that I'm using
malloc(9) instead of pool(9) to allocate sockaddr_dl, I can create
a sockaddr_dl of any size in the kernel, so expanding sockaddr_dl
is useless.
Avoid using sizeof(struct sockaddr_dl) in the kernel.
Introduce sockaddr_dl_alloc() for allocating & initializing an
arbitrary sockaddr_dl on the heap.
Add an argument, the sockaddr length, to sockaddr_alloc(),
sockaddr_copy(), and sockaddr_dl_setaddr().
Constify: LLADDR() -> CLLADDR().
Where the kernel overwrites LLADDR(), use sockaddr_dl_setaddr(),
instead. Used properly, sockaddr_dl_setaddr() will not overrun
the end of the sockaddr.
from the forwarding table's users:
Introduce rt_walktree() for walking the routing table and
applying a function to each rtentry. Replace most
rn_walktree() calls with it.
Use rt_getkey()/rt_setkey() to get/set a route's destination.
Keep a pointer to the sockaddr key in the rtentry, so that
rtentry users do not have to grovel in the radix_node for
the key.
Add a RTM_GET method to rtrequest. Use that instead of
radix_node lookups in, e.g., carp(4).
Add sys/net/link_proto.c, which supplies sockaddr routines for
link-layer socket addresses (sockaddr_dl).
Cosmetic:
Constify. KNF. Stop open-coding LIST_FOREACH, TAILQ_FOREACH,
et cetera. Use NULL instead of 0 for null pointers. Use
__arraycount(). Reduce gratuitous parenthesization.
Stop using variadic arguments for rip6_output(), it is
unnecessary.
Remove the unnecessary rtentry member rt_genmask and the
code to maintain it, since nothing actually used it.
Make rt_maskedcopy() easier to read by using meaningful variable
names.
Extract a subroutine intern_netmask() for looking up a netmask in
the masks table.
Start converting backslash-ridden IPv6 macros in
sys/netinet6/in6_var.h into inline subroutines that one
can read without special eyeglasses.
One functional change: when the kernel serves an RTM_GET, RTM_LOCK,
or RTM_CHANGE request, it applies the netmask (if supplied) to a
destination before searching for it in the forwarding table.
I have changed sys/netinet/ip_carp.c, carp_setroute(), to remove
the unlawful radix_node knowledge.
Apart from the changes to carp(4), netiso, ATM, and strip(4), I
have run the changes on three nodes in my wireless routing testbed,
which involves IPv4 + IPv6 dynamic routing acrobatics, and it's
working beautifully so far.
parentheses in return statements.
Cosmetic: don't open-code TAILQ_FOREACH().
Cosmetic: change types of variables to avoid oodles of casts: in
in6_src.c, avoid casts by changing several route_in6 pointers
to struct route pointers. Remove unnecessary casts to caddr_t
elsewhere.
Pave the way for eliminating address family-specific route caches:
soon, struct route will not embed a sockaddr, but it will hold
a reference to an external sockaddr, instead. We will set the
destination sockaddr using rtcache_setdst(). (I created a stub
for it, but it isn't used anywhere, yet.) rtcache_free() will
free the sockaddr. I have extracted from rtcache_free() a helper
subroutine, rtcache_clear(). rtcache_clear() will "forget" a
cached route, but it will not forget the destination by releasing
the sockaddr. I use rtcache_clear() instead of rtcache_free()
in rtcache_update(), because rtcache_update() is not supposed
to forget the destination.
Constify:
1 Introduce const accessor for route->ro_dst, rtcache_getdst().
2 Constify the 'dst' argument to ifnet->if_output(). This
led me to constify a lot of code called by output routines.
3 Constify the sockaddr argument to protosw->pr_ctlinput. This
led me to constify a lot of code called by ctlinput routines.
4 Introduce const macros for converting from a generic sockaddr
to family-specific sockaddrs, e.g., sockaddr_in: satocsin6,
satocsin, et cetera.
have sys/net/route.c:rtrequest1(), which takes rt_addrinfo * as the argument.
pass rt_addrinfo all the way down to rtrequest, and ifa->ifa_rtrequest.
3rd arg of ifa->ifa_rtrequest is now rt_addrinfo * instead of sockaddr *
(almost noone is using it anyways).
benefit: the follwoing command now works. previously we need two route(8)
invocations, "add" then "change".
# route add -inet6 default ::1 -ifp gif0
remove unsafe typecast in rtrequest(), from rtentry * to sockaddr *. it was
introduced by 4.3BSD-reno and never corrected.
XXX is eon_rtrequest() change correct regarding to 3rd arg?
eon_rtrequest() and rtrequest() were incorrect since 4.3BSD-reno,
so i do not have correct answer in the source code.
someone with more clue about netiso-over-ip, please help.
(Sorry for a big commit, I can't separate this into several pieces...)
Pls check sys/netinet6/TODO and sys/netinet6/IMPLEMENTATION for details.
- sys/kern: do not assume single mbuf, accept chained mbuf on passing
data from userland to kernel (or other way round).
- "midway" ATM card: ATM PVC pseudo device support, like those done in ALTQ
package (ftp://ftp.csl.sony.co.jp/pub/kjc/).
- sys/netinet/tcp*: IPv4/v6 dual stack tcp support.
- sys/netinet/{ip6,icmp6}.h, sys/net/pfkeyv2.h: IETF document assumes those
file to be there so we patch it up.
- sys/netinet: IPsec additions are here and there.
- sys/netinet6/*: most of IPv6 code sits here.
- sys/netkey: IPsec key management code
- dev/pci/pcidevs: regen
In my understanding no code here is subject to export control so it
should be safe.
[1] if user tries to enter in a bogus PVC don't leave it in the routing
table ... remove it
[2] change ioctl arg to include rxso for lower layer
[3] add hooks (inside "NATM" ifdef) for native mode atm sockets so that
they don't clash with IP PVCs. [i am still debugging the native
mode atm socket protosw code]
ozaki-r