deal with shortages of the VM maps where the backing pages are mapped
(usually kmem_map). Try to deal with this:
* Group all information about the backend allocator for a pool in a
separate structure. The pool references this structure, rather than
the individual fields.
* Change the pool_init() API accordingly, and adjust all callers.
* Link all pools using the same backend allocator on a list.
* The backend allocator is responsible for waiting for physical memory
to become available, but will still fail if it cannot callocate KVA
space for the pages. If this happens, carefully drain all pools using
the same backend allocator, so that some KVA space can be freed.
* Change pool_reclaim() to indicate if it actually succeeded in freeing
some pages, and use that information to make draining easier and more
efficient.
* Get rid of PR_URGENT. There was only one use of it, and it could be
dealt with by the caller.
From art@openbsd.org.
Don't copy ttl from the inner packet to the encapsulating packet. Make
the outer ttl sysctl'able. This should close PR 14269 from Jasper Wallace
(change partly from there) and it makes traceroute work over gre tunnels.
that other parts of the kernel won't lose gratuitously. There are
places where it's assumed that it won't grow that large.
(2) Avoid accidental reuse of occupied slots in the ifindex2ifnet[]
table.
timeout connection that made it to phase NETWORK yet. (For drivers using
the internal timeout mechanism; isdnd, that does the timeout handling for
ISDN drivers, still needs to be fixed.)
Thanks to Wolfgang Solfrank for finding this.
a configurable maximum (default: 5).
Some ISPs shut down accounts (at least temporarily) after to many bad
retries. This hit me recently due to a stupid pilot error and the fast
retry rate.
containing the userland visible thinks (i.e. ioctl definitions).
Remove all (both) old ioctls, as they had a brain dead API and made keeping
binary compatibility more or less impossible.
Replace by several new ioctls. While there, remove any arbitrary limits
(resulting from the old, broken ioctls) and allow any length of names
and passwords.
values).
Implement a secondary connection-reestablishement mode, which is only
entered after (1) we have successfully transfered payload data over this
connection and (2) if initial retries did not reestablish a session.
In this mode we retry (infrequently) forever, until adminstrator stops
us (by "ifconfig ppppoe0 down"). XXX - need to display this mode in
pppoectl.
It is now possible to pull the DSL modems plug for say 15 minutes, plug
it back in again and just wait. The connection will be reestablished within
three minutes.
it again when going from INITIAL to STARTING. This has been done for
passive or auto-conecting interfaces always, but not for permanent
ones.
This fixes session reestablishement for PPPoE interfaces without LINK1 set,
and probably also closes PR kern/11161.
Thanks to Jared D. McNeill and Ross Harvey for sugesting debug methology.
Collect both local and remote address and set them to the interface in
one step (the peer adress was not set at all before).
This causes the peer address now to show up on the interface and all
messages to the routing socket to be send with correct data. The latter
has been the last missing piece to complete PPPoE support.
from sppp_attach.
When destroying the interface, call sppp_detach for proper cleanup.
This avoids a crash from the slow timeout handler for no longer existing
interfaces (spotted by Rémi Zara).
does not use software interrupts; remove these bridge netisr
hooks left over from a previous incarnation of the bridge code.
Noted by Andrew Brown <atatat@atatdot.net>.
The type of ALIGN() is vary on architecture and casting pointer to u_int
is incorrect for MI code.
Since the code is to make sure aligned access to IP header and requires
bcopy if the test fails. So the performance implication is not necessary
and we can use ALIGNED_POINTER() here.
pointed out by nathanw.
XXX: FH chanset should be calculated by FH hop pattern, but BayStack 650 AP
always specify chanset to fixed value 1. The previous code try to this
hack into awi driver, but it is insufficient because the chanset value
in awi driver may change while scan and it may be different from the
value in receiving beacon/probe-response. So we save encoded FH chanset
into channel in 802.11 common bss information for now.
if_ieee80211subr.c, which can be shared between any IEEE 802.11
drivers.
However, most of current working IEEE 802.11b wireless LAN cards
have rich firmware and we cannot have a control to management frames
for such cards.
IBSS creation is now supported for the awi driver.
complete for some reason, we defer it for a bit and then try again. This
gets ping down to 0% packet loss.
Of course, ping _should_ have been at 0% packet loss anyway, and that's the
next thing to deal with.
Add capabilities bits that indicate an interface can only perform
in-bound TCPv4 or UDPv4 checksums. There is at least one Gig-E chip
for which this is true (Level One LXT-1001), and this is also the
case for the Intel i82559 10/100 Ethernet chips.
worked out by observing RISC iX's behaviour, so it may be technically
wrong. The only implementations of IP-over-Econet for which I've got
sources don't support broadcasts.
Tested using broadcast ping from RISC iX to NetBSD, and using rwhod.
Use m_copydata() to preserve the Econet header, so we don't depend on
notionally-unused areas of an mbuf remaining untouched.
Check that ARP-over-Econet requests are exactly eight bytes long.
Use m_pullup() before trusting mtod().
Between them, these make reception of unicast ARP responses work properly.
protocols, and lacking any timeouts, but it basically works, doing four-way
handshakes in both directions and incoming Machine Peek operations.
Oh, and Econet is Acorn's ancient, proprietary 500kbit/s networking
technology.
fvdl