Move the function+line printing into GRE_DPRINTF().
Retire gre_closef(). Retire gre_join(). Constify gre_reconf(),
and don't pass it an LWP any longer.
Make this work in the new file descriptor regime. Add a kernel
thread per gre(4) instance whose purpose is to install the socket
into proc0's file descriptor table. Add gre_fp_send() and
gre_fp_recv() for passing file_t pointers to proc0.
Fix locking: don't solock() in the socket upcall, where it is
already held. Do solock() before calling soconnect().
Simplify reconfiguration.
Update a comment that mentions finding a less specific route, since
we don't do that any more.
proclist_mutex and proclist_lock into a single adaptive mutex (proc_lock).
Implications:
- Inspecting process state requires thread context, so signals can no longer
be sent from a hardware interrupt handler. Signal activity must be
deferred to a soft interrupt or kthread.
- As the proc state locking is simplified, it's now safe to take exit()
and wait() out from under kernel_lock.
- The system spends less time at IPL_SCHED, and there is less lock activity.
- Socket layer becomes MP safe.
- Unix protocols become MP safe.
- Allows protocol processing interrupts to safely block on locks.
- Fixes a number of race conditions.
With much feedback from matt@ and plunky@.
failure of wpa_supplicant(8) to re-key promptly, as reported in
http://mail-index.netbsd.org/tech-net/2008/04/18/msg000459.html
- Make bpf's read timeout work more correctly with select/poll.
- A fix for catchpacket() which delays calling bpf_wakeup() until
the state has been updated.
(rev 1.125): correct the check for fd_getsock() failure in
gre_socreate().
The second bug is more complicated to fix. Since rev 1.125,
gre_reconf() is using the file descriptor table of the current
process instead of the process 0's (the kernel's).
1. Please don't cast function pointers to (void *), use the full function
prototype cast; this is for archs where a function pointer is not a regular
pointer.
2. Compare pointers to NULL not 0.
Use fewer 'error = ...; break;' statements and more 'return
...;'
Make the SIOCSIFFLAGS case more clear by using a switch
statement instead of an if-else if-else chain.
Shorten a staircase, and remove two unnecessary curly
braces.
- Add a lot of missing selinit() and seldestroy() calls.
- Merge selwakeup() and selnotify() calls into a single selnotify().
- Add an additional 'events' argument to selnotify() call. It will
indicate which event (POLL_IN, POLL_OUT, etc) happen. If unknown,
zero may be used.
Note: please pass appropriate value of 'events' where possible.
Proposed on: <tech-kern>
the opportunity to handle an ioctl before generic ifioctl handling
occurs. This will ease extending the kernel and sharing of code
between drivers.
First steps: Make the signature of ifioctl_common() match struct
ifinet->if_ioctl. Convert SIOCSIFCAP and SIOCSIFMTU to the new
ifioctl() regime, throughout the kernel.
1 Extract subroutine if_dl_create() from if_alloc_sadl().
if_dl_create() allocates a link-layer ifaddr.
2 Extract subroutine ifioctl_common() from ifioctl(). ifioctl_common()
will be the basis for an ifnet "superclass" whose functions
drivers may inherit. Very simple drivers may set ifnet->if_ioctl
= ifioctl_common. More sophisticated drivers will set ifnet->if_ioctl
= driver_ioctl. driver_ioctl() will call ifioctl_common() to
re-use the common code.
reference, but mark the cache 'invalid'. Let the next user of the
route cache check to whether or not the cache is valid, and update
the rtentry reference if necessary. In this way, avoid hairy
splnet()/splx() protection of route caches, which I never did trust.
In rtcache_lookup2(), use the return values of rtcache_validate()
and _rtcache_init() instead of looking at _ro_rt. Also, check the
return code of rtcache_setdst() for an error.
ether_mediachange() to their own module that we compile only if
the kernel configuration demands support for both MII buses and
ethernet. Thanks to Tom Spindler for suggesting that these routines
move to dev/mii/.
Add if_set_sadl() that both sets the link-layer address length and
replaces the current link-layer address with a new one, and use it
throughout the tree.
to _ro_rt. Use rtcache_getrt() to access a route cache's struct
rtentry *.
Introduce struct ifnet->if_dl that always points at the interface
identifier/link-layer address. Make code that treated the first
ifaddr on struct ifnet->if_addrlist as the interface address use
if_dl, instead.
Remove stale debugging code from net/route.c. Move the rtflush()
code into rtcache_clear() and delete rtflush(). Delete rtalloc(),
because nothing uses it any more.
Make ND6_HINT an inline, lowercase subroutine, nd6_hint.
I've done my best to convert IP Filter, the ISO stack, and the
AppleTalk stack to rtcache_getrt(). They compile, but I have not
tested them. I have given the changes to PF, GRE, IPv4 and IPv6
stacks a lot of exercise.
- Reduce available SPL levels for hardware devices to none, vm, sched, high.
- Acquire kernel_lock only for interrupts at IPL_VM.
- Implement threaded soft interrupts.
instead of adding/subtracting our own IPv4 header.
There are many benefits: gre(4) needn't grok the outer encapsulation
header any longer, so this simplifies the gre(4) code. The IP
stack needn't grok GRE, so it is simplified, too. gre(4) will
benefit from optimizations in the socket code. Eventually, gre(4)
will gain an IPv6 encapsulation with very few new lines of code.
There is a small performance loss. A 133 MHz, 486-class AMD Elan
sinks/sources a TCP stream over GRE with about 93% the throughput
of the old code. TCP throughput on a 266 MHz, 586-class AMD Geode
is about 96% the throughput of the old code. A 175-MHz ADM5120
(MIPS) only sinks a TCP stream over GRE at about 90% of the old
code; I am still investigating that.
I produced stripped-down versions of sosend() and soreceive() for
gre(4) to use. They are guaranteed not to block, so they can be
called from a software interrupt and from a socket upcall,
respectively.
A kernel thread is no longer necessary for socket transmit/receive,
but I didn't get around to removing it, yet.
Thanks to Matt Thomas for suggesting the use of stripped-down socket
code and software interrupts, and to Andrew Doran for advice and
answers concerning software interrupts, threads, and performance.
ifreq * arguments to ether_addmulti() and ether_delmulti() to const
struct sockaddr *, since ether_{add,del}multi() only ever read the
sockaddr ifreq member, ifr_addr. Update uses in carp(4) and in
vlan(4).
Somehow having it above interfered with ctags(1) producing a tag
for etherip_softc.
Remove the sole member of the union etherip_softc.sc_scr; call it
sc_ro. Delete the union. Delete the #define for sc_ro. The union
was a holdover from days before the route caches were unified.
Copy the entire sockaddr to the buffer to be written to user space,
according to its length, not just the part that fits in struct
sockaddr.
This fixes the 'bad MAC address' problem in dhclient.
sockaddrs bigger than struct sockaddr. Tightly bind decrementing
available space and using it, avoiding incorrect accounting in an
error case. Document invariants. Document calling convention for
SIOCGIFCONF. Simplify by removing code to handle sockaddrs that don't
fit in struct ifreq; with sockaddr_storage this can no longer occur.
Add several KASSERTs.
This commit resolves the problem with racoon failing to list
interfaces.
Proposed on tech-net@ with no objections.
acquisition and final release out into gre_thread(). This will
fix a locking bug that LOCKDEBUG exposed: holding a spinlock over
an sosend() call is a no-no.
Cosmetic: join some lines, remove some unnecessary curly braces.
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.
identify sockaddr_dl abuse that remains in the kernel, especially
the potential for overwriting memory past the end of a sockaddr_dl
with, e.g., memcpy(LLADDR(), ...).
Use sockaddr_dl_setaddr() in a few places.
* Create the kernel thread in gre_clone_create() instead of trying
to create it in gre_ioctl(). (Thanks ad@ for suggesting it, and
pointing out that I can't kthread_create while I hold a spin
lock.) Run the thread always, but put it to sleep while the
gre(4) is not in UDP mode.
* Use sockaddr_in_init().
* Move some thread state off of the stack and into the softc.
* Extract subroutines gre_do_recv(), gre_do_send(), and gre_reconf()
from gre_thread1(), making the code more readable.
