tobiasu [Fri, 5 Jun 2015 21:42:37 +0000 (21:42 +0000)]
Fix library search path so we link against the freshly built libcrypto.so
instead of a stale one.
ok miod@ mpi@
krw [Fri, 5 Jun 2015 21:41:43 +0000 (21:41 +0000)]
Try harder to avoid (very unlikely) NULL pointer de-ref by tweaking
code to use sotopf() like tcp_usrreq() does. Also following
tcp_usrreq(), put more stuff under splsoftnet. And as a result
in-line code in udp_detach() and nuke udp_detach().
Most ideas from and ok mikeb@
deraadt [Fri, 5 Jun 2015 19:36:28 +0000 (19:36 +0000)]
And ... more discussion occurs between miod and kettenis about
what register dance copyerr should do.....
deraadt [Fri, 5 Jun 2015 18:36:07 +0000 (18:36 +0000)]
And part 2 of the onfault repair. Do the actual clearing of pcb_onfault
in copyerr itself, like other architectures of this type do.
as a result of chatter between miod and kettenis
nicm [Fri, 5 Jun 2015 18:18:32 +0000 (18:18 +0000)]
Similarly, for sessions use a callback to free rather than checking
every loop.
miod [Fri, 5 Jun 2015 18:14:26 +0000 (18:14 +0000)]
Missing no-pie logic for bootblocks.
nicm [Fri, 5 Jun 2015 18:06:30 +0000 (18:06 +0000)]
Change deref to the more sensible unref, and add a couple I missed before.
nicm [Fri, 5 Jun 2015 18:01:12 +0000 (18:01 +0000)]
Instead of putting dead clients on a list and checking it every loop,
use event_once to queue a callback to deal with them. Also dead clients
with references would never actually be freed because the wrap-up
functions (the callback for stdin, or status_prompt_clear) would never
be called. So call them in server_client_lost.
deraadt [Fri, 5 Jun 2015 16:59:10 +0000 (16:59 +0000)]
Do not unconditionally clear pcb_onfault after a uvm_fault. That will
permit the active copyout/copyin to continue work on subsequent faulting
pages and not misinterpret & fault them as kernel bcopy against userland
addresses. Old bug -- fall of 1996. This should fix getentropy issues
on MP systems which have become more apparent recently, probably due to
some combo of increased ASLR with unlocked getentropy happening very soon
after vfork/fork...
ok miod
deraadt [Fri, 5 Jun 2015 16:45:24 +0000 (16:45 +0000)]
these days, curproc is never NULL, so skip those checks.
Discussed with kettenis
deraadt [Fri, 5 Jun 2015 16:35:24 +0000 (16:35 +0000)]
trunc_page() generally.... succeeds at clearing the page offset
bits the first time it is called, so don't do it again.
ok miod
kettenis [Fri, 5 Jun 2015 16:07:24 +0000 (16:07 +0000)]
These days p->p_addr will never be NULL.
ok deraadt@
millert [Fri, 5 Jun 2015 15:13:13 +0000 (15:13 +0000)]
For "ssh -L 12345:/tmp/sock" don't fail with "No forward host name."
(we have a path, not a host name). Based on a diff from Jared Yanovich.
OK djm@
nicm [Fri, 5 Jun 2015 15:10:13 +0000 (15:10 +0000)]
Do not use the key variable uninitialized (in a debug log statement),
reported by jungleboogie0 at gmail dot com.
vgross [Fri, 5 Jun 2015 13:35:08 +0000 (13:35 +0000)]
Fix coupling and decoupling operations.
With help and ok from mikeb@
mikeb [Fri, 5 Jun 2015 13:22:34 +0000 (13:22 +0000)]
Improve error handling and recovery during state insertion
Reshuffle the code around a bit and greatly improve error handling
fixing a few bugs along the way.
Problem reported by and fix was written with Alexandr Nedvedicky.
OK henning
mpi [Fri, 5 Jun 2015 11:38:19 +0000 (11:38 +0000)]
Add bits missed in previous... I suck at cvs.
mpi [Fri, 5 Jun 2015 10:24:23 +0000 (10:24 +0000)]
Put spaces and commas where they belong.
mpi [Fri, 5 Jun 2015 10:15:54 +0000 (10:15 +0000)]
Finally protect VP lookups to guarantee that a pted won't be freed or
reused by a CPU while another CPU is manipulating it.
This races occurs because the virtual spill handlers are run without
taking the KERNEL_LOCK for obvious reasons. So use a per-pmap mutex
that CPUs must hold when modifying a pted in order to guarantee the
atomicity of operations *and* the coherence between pmap VPs tree and
what's in the HASH.
Thanks to dlg@ for assisting me debugging this. This change ends your
PowerPC pmap SMP show of the week. GENERIC.MP on macppc should now be
stable enough to build ports without corrupting its own memory.
ok kettenis@, deraadt@, dlg@
mpi [Fri, 5 Jun 2015 10:06:35 +0000 (10:06 +0000)]
Don't try to be clever when unrolling the loop in pmap_remove().
Needed for upcoming locking.
mpi [Fri, 5 Jun 2015 10:04:34 +0000 (10:04 +0000)]
Replace the per-entry locks by a global HASH lock.
Since this lock is recursive we can now guarantee the atomicity of
pte_inser{32,64}() when a pted has to be removed first. This fixes
one of the races.
Using a __mp_lock here also allowed dlg@ to provide me useful traces
to fix the next race. Thanks for your help!
ok kettenis@, deraadt@, dlg@
mpi [Fri, 5 Jun 2015 09:53:40 +0000 (09:53 +0000)]
Call pte_spill_v() from the real mode fault handler instead of rerolling
it. This will reduce the number of places to audit for locking.
Note that for profiling purposes pte_spill_v() is now marked a __noprof
since per-CPU profiling buffers are not guaranteed to be 1:1 mapped and
cannot be accessed from the real mode fault handler.
ok kettenis@, deraadt@, dlg@
mpi [Fri, 5 Jun 2015 09:48:01 +0000 (09:48 +0000)]
Rewrite PTE manipulation routines to better match the PEM.
Document every operation, make sure to call "sync" when appropriate so
that other CPUs see the bit changes and finally grab a lock where it was
missing to grantee atomicity.
ok kettenis@, deraadt@, dlg@
mpi [Fri, 5 Jun 2015 09:42:10 +0000 (09:42 +0000)]
Split pteclrbits() into pmap_{test,clear}_attrs().
This should not introduce any behavior change but makes the code easier
to read and later easier to protect. This also brings this pmap closer
to what others do.
Thanks to kettenis@ for spotting a bad typo!
ok kettenis@, deraadt@, dlg@
mpi [Fri, 5 Jun 2015 09:38:52 +0000 (09:38 +0000)]
More usages of pmap_ptedinhash().
If you wonder why pte_insert{32,64}() is not using pmap_hash_remove() if
it finds a conflicting PTE in the HASH, it's because in the current state
trying to grab the same lock a second time would lead to a deadlock.
This is much easier to reproduce on G5 (or G4 with BAT disabled).
ok kettenis@, deraadt@, dlg@
mpi [Fri, 5 Jun 2015 09:32:22 +0000 (09:32 +0000)]
Remove DEBUG stuff.
mpi [Fri, 5 Jun 2015 09:31:19 +0000 (09:31 +0000)]
Make use of ptesr() instead of rerolling it.
mpi [Fri, 5 Jun 2015 09:30:03 +0000 (09:30 +0000)]
Merge various copies of the same code into a new function to determine
if a PTE is present in the HASH.
Note that atomicity is currently not guaranteed between this check and
the following operations.
ok kettenis@, deraadt@, dlg@
mpi [Fri, 5 Jun 2015 09:25:21 +0000 (09:25 +0000)]
Introduce pmap_pted_ro() a simple wrapper for the 32/64 bits versions
that does not call pmap_vp_lookup().
Carreful readers would have notice the removal of the bits on the virtual
address with a page mask, this change allows me to find the 13 years old
bug fixed in r1.145.
ok kettenis@, deraadt@, dlg@
mpi [Fri, 5 Jun 2015 09:18:50 +0000 (09:18 +0000)]
Do only one VP lookup when removing a page.
This simplify pmap_remove() & friends by re-using an already fetched PTE
descriptor.
There's currently a race on MP system where one CPU can reuse a pted
while another one is still trying to insert it in the HASH. This commit
starts reducing the number of pmap_vp_lookup() calls to help fix this
race.
ok kettenis@, deraadt@, dlg@
mpi [Fri, 5 Jun 2015 09:09:58 +0000 (09:09 +0000)]
Remove the MANAGED flag when removing a PV entry.
Even if this change is not strickly needed, because the memory will be
returned to the pool it helped me track the use-after-free.
nicm [Fri, 5 Jun 2015 09:09:08 +0000 (09:09 +0000)]
Fix a warning.
mpi [Fri, 5 Jun 2015 09:05:35 +0000 (09:05 +0000)]
Remove unneeded splvm() calls and the pool_setipl(9) hack of r1.140.
By instructing spl(9) calls on MP machines I figured out that their high
cost was hiding a race condition involving PTE reuse in our pmap. Thanks
to deraadt@ for finding a way to trigger such panic by adding a couple of
splvm().
This should make the races easier to trigger but will be addressed
shortly.
This commit starts your PowerPC pmap SMP show of the week.
ok kettenis@, deraadt@, dlg@
nicm [Fri, 5 Jun 2015 08:14:16 +0000 (08:14 +0000)]
Make it so that if a window or session target is prefixed with an =,
only an exact name or index match is accepted, no special character,
prefix match, or fnmatch.
jsg [Fri, 5 Jun 2015 05:39:54 +0000 (05:39 +0000)]
Use p15 for the coprocessor name not 15. gas accepts both forms, clang
won't accept just a number which a comment in the gas code mentions is
for backward compatibility.
stsp [Fri, 5 Jun 2015 05:24:19 +0000 (05:24 +0000)]
regen
stsp [Fri, 5 Jun 2015 05:24:08 +0000 (05:24 +0000)]
Add another PCI ID for Intel Dual Band Wireless AC 3165. From brad.
djm [Fri, 5 Jun 2015 03:44:14 +0000 (03:44 +0000)]
typo: accidental repetition; bz#2386
deraadt [Thu, 4 Jun 2015 23:30:14 +0000 (23:30 +0000)]
sync
nicm [Thu, 4 Jun 2015 23:27:51 +0000 (23:27 +0000)]
Move the nested check from client to server and compare the client tty
name to all the pane pty names instead of comparing socket paths. This
means that "new -d" will work without unsetting $TMUX.
stsp [Thu, 4 Jun 2015 23:08:21 +0000 (23:08 +0000)]
Mention rtwn(4) in pci(4). Reminded by brad.
nicm [Thu, 4 Jun 2015 22:56:33 +0000 (22:56 +0000)]
systrace(4) STRIOCATTACH can only fail for file(1) if it is already
systraced. If so, silently ignore the error and do not attempt to apply
our own systrace policy. From Patrick Keshishian.
stsp [Thu, 4 Jun 2015 22:29:31 +0000 (22:29 +0000)]
man page for rtwn(4)
stsp [Thu, 4 Jun 2015 22:14:51 +0000 (22:14 +0000)]
enable rtwn(4) on i386
stsp [Thu, 4 Jun 2015 21:29:06 +0000 (21:29 +0000)]
Enable rtwn(4) on amd64.
stsp [Thu, 4 Jun 2015 21:13:12 +0000 (21:13 +0000)]
attach rtwn at pci
stsp [Thu, 4 Jun 2015 21:08:40 +0000 (21:08 +0000)]
New rtwn(4) driver for RTL8188CE wifi cards.
This is a PCI card from the same chip family as supported by urtwn(4) on USB.
Development started in 2013 using urtwn(4) as a starting point but was dormant
for much of the time since. I finally unslacked after uwe@ provided help with
lifting this driver on its feet. As usual we got helpful hints from Theo.
Requires firmware which will be available in ports soon.
There are rate adaptation issues that still need to be fixed, cause unknown.
In my testing the hardware rarely transmits more than 1Mbit/s.
Committing over MAC/BB RTL8188CE, RF 6052 1T1R.
jmc [Thu, 4 Jun 2015 20:34:22 +0000 (20:34 +0000)]
tweak SYNOPSIS and usage();
eric [Thu, 4 Jun 2015 19:26:39 +0000 (19:26 +0000)]
force reseeding if pid has changed.
ok deraadt@
eric [Thu, 4 Jun 2015 19:23:17 +0000 (19:23 +0000)]
make sure to check for resolv.conf update the first time the resolver
is used after pid has changed.
ok deraadt@
deraadt [Thu, 4 Jun 2015 18:49:28 +0000 (18:49 +0000)]
sync
deraadt [Thu, 4 Jun 2015 18:46:43 +0000 (18:46 +0000)]
+ REALTEK RT8191SE
dms [Thu, 4 Jun 2015 18:33:41 +0000 (18:33 +0000)]
Add support for em(4) on Teak 3020, a Tolopai (EP80579)
based devices. This introduces Realtek PHY into em driver
code and is only a temporary solution to the problem.
OK deraadt@
kettenis [Thu, 4 Jun 2015 18:01:44 +0000 (18:01 +0000)]
The (no quite so) new kernel perfpolicy code calls cpu_setperf() from a
timeout. Unfortunately the smu(4) CPU voltage slewing code sleeps, which
causes a kernel panic. Prevent this by delegating the CPU frequency switching
and voltage slewing to a task.
ok mpi@
dms [Thu, 4 Jun 2015 17:42:59 +0000 (17:42 +0000)]
regen
dms [Thu, 4 Jun 2015 17:42:12 +0000 (17:42 +0000)]
Add more of Tolopai (EP80579) Gigabit controllers id.
mikeb [Thu, 4 Jun 2015 17:10:33 +0000 (17:10 +0000)]
Check if interface was stopped before calling rx/tx interrupt routines.
Report & tests by mxb@alumni.chalmers.se, thanks!
OK deraadt, chris
nicm [Thu, 4 Jun 2015 14:29:33 +0000 (14:29 +0000)]
Make unsetting a global option restore it to the default. Diff lying
around for a while, I have forgotten who suggested it :-/.
sobrado [Thu, 4 Jun 2015 14:23:00 +0000 (14:23 +0000)]
spacing, makes example fit on display.
no objection from gilles@
nicm [Thu, 4 Jun 2015 11:43:51 +0000 (11:43 +0000)]
Add support for a single "marked pane". There is one marked pane in the
server at a time; it may be toggled or cleared with select-pane -m and
-M (the border is highlighted). A new target '~' or '{marked}' specifies
the marked pane to commands and it is the default target for the
swap-pane and join-pane -s flag (this makes them much simpler to use -
mark the source pane and then change to the target pane to run swapp or
joinp).
mpi [Thu, 4 Jun 2015 09:51:15 +0000 (09:51 +0000)]
Remove unused sa_dl.
nicm [Thu, 4 Jun 2015 09:42:29 +0000 (09:42 +0000)]
teminal -> terminal, from Corey Farwell.
jsg [Thu, 4 Jun 2015 06:11:21 +0000 (06:11 +0000)]
Switch intel_gmbus_is_port_valid and intel_gmbus_is_forced_bit from
extern inline back to static inline so the kernel will build on
compilers that default to c99 inline semantics.
jsg [Thu, 4 Jun 2015 06:07:23 +0000 (06:07 +0000)]
use mallocarray instead of malloc in drm_calloc
jsg [Thu, 4 Jun 2015 06:03:11 +0000 (06:03 +0000)]
remove the now unused alawtolin16 table
ok ratchov@
bcallah [Wed, 3 Jun 2015 23:40:01 +0000 (23:40 +0000)]
Remove unused defines and functions.
"Looks fine." deraadt@
mikeb [Wed, 3 Jun 2015 22:01:07 +0000 (22:01 +0000)]
Adjust yielding condition slightly in the if_input_process thread.
ok mpi, kettenis
reyk [Wed, 3 Jun 2015 20:43:21 +0000 (20:43 +0000)]
Turn off -Werror in relayctl.
reyk [Wed, 3 Jun 2015 19:57:53 +0000 (19:57 +0000)]
Remove -Werror from the Makefile: as it happened with asprintf, it
breaks the tree if any new compiler warnings get enabled, especially
with -Wall. We do not want to use it in the tree, except for some
very special places.
As discussed with deraadt@
reyk [Wed, 3 Jun 2015 19:51:16 +0000 (19:51 +0000)]
It is better to check the asprintf return value for -1 instead of the
input buffer for NULL.
Found by adding __attribute((__warn_unused_result__)) to asprintf.
OK deraadt@
millert [Wed, 3 Jun 2015 18:08:54 +0000 (18:08 +0000)]
Treat a missing mail spool the same as a zero-length mail spool
unless the -f option was specified. OK deraadt@
millert [Wed, 3 Jun 2015 16:26:04 +0000 (16:26 +0000)]
snprintf(3) is available on all modern systems and asprintf(3) is
available on more systems these days. OK deraadt@
mpi [Wed, 3 Jun 2015 15:43:55 +0000 (15:43 +0000)]
Unify .Nd blocks, prodded by and ok jmc@
yasuoka [Wed, 3 Jun 2015 11:57:37 +0000 (11:57 +0000)]
Fix pf_map_addr() not to cause dividing by 0. This fixes problem when
using table or dynamic interface addresses for source-hash. Also
avoid calling arc4random_uniform() with upper_bound == 0.
ok mikeb
kettenis [Wed, 3 Jun 2015 11:45:52 +0000 (11:45 +0000)]
Actually commit the bits that make the compiler emit visibility information
for (undefined) references with non-default visibility on mips64.
mpi [Wed, 3 Jun 2015 08:41:43 +0000 (08:41 +0000)]
Support for U4, missed in previous.
ok kettenis@, miod@, dlg@
mpi [Wed, 3 Jun 2015 08:30:27 +0000 (08:30 +0000)]
Check for the correct bit in the interrupts property of the device-tree
to determin if interrupts should be established as edge/level triggered.
Fix audio interrupts on U4 systems. On such systems the "interrupts"
fields in the device-tree has more bits set and cannot be compared to
0.
mpi [Wed, 3 Jun 2015 08:14:23 +0000 (08:14 +0000)]
Add hpb(4), reminded by Brad.
mpi [Wed, 3 Jun 2015 08:10:53 +0000 (08:10 +0000)]
Explicitly request the sockaddr_dl when doing a RTM_GET rather than
assuming that it will be in the gateway sa.
Fixes a regression introduced with the support of multiple connected
routes, found the hardway by deraadt@.
ok claudio@
millert [Wed, 3 Jun 2015 02:35:50 +0000 (02:35 +0000)]
Move warning on open failure to open_mbox() where we have the file name.
millert [Wed, 3 Jun 2015 02:24:36 +0000 (02:24 +0000)]
Do not assume that asprintf() clears the pointer on failure, which
is non-portable. Also add missing asprintf() return value checks.
OK deraadt@ guenther@ doug@
deraadt [Wed, 3 Jun 2015 01:49:35 +0000 (01:49 +0000)]
better description of internal copyn() function
jsg [Wed, 3 Jun 2015 01:30:29 +0000 (01:30 +0000)]
missing __EOT spotted by rpe@
dlg [Wed, 3 Jun 2015 00:50:09 +0000 (00:50 +0000)]
there's been a long standing issue in ppp on a tty/serial line where it allocates mbufs at IPL_SOFTTTY, which is above the IPL_NET the mbuf layer protects itself at.
recent improvements to diagnostics in pools and mbufs now panic
instead of letting these things silently corrupt.
this reworks the ppp handling in the tty layer so it has its own
private pool to allocate packet memory out of. these packets get
built and then queued for softnet to process. softnet dequeues the
packet and attaches it to mbufs as external storage before handing
it on to the rest of the stack.
this was reported on bugs@ and tested by both Walter Daugherity and
Martin van den Nieuwelaar
ok deraadt@ mpi@
millert [Tue, 2 Jun 2015 23:05:05 +0000 (23:05 +0000)]
Don't leak value returned by cgetstr(). OK deraadt@
rpe [Tue, 2 Jun 2015 19:54:06 +0000 (19:54 +0000)]
Remove the $FSTABFLAG variable and use -F directly. The installer
uses disklabel UIDs unconditionally for a while already.
OK krw@
"looks good" deraadt@
rpe [Tue, 2 Jun 2015 19:39:18 +0000 (19:39 +0000)]
Replace
[[ -n $(foo | grep 'bar') ]]
with
foo | grep -q 'bar'
OK halex@ krw@
deraadt [Tue, 2 Jun 2015 17:59:19 +0000 (17:59 +0000)]
sync
jung [Tue, 2 Jun 2015 17:29:28 +0000 (17:29 +0000)]
fix variable name typo
ok ajacoutot@
sobrado [Tue, 2 Jun 2015 16:02:44 +0000 (16:02 +0000)]
write Mobile IP{,v6} in a consistent way, as given in RFCs 5944 and 6275.
ok jmc@
millert [Tue, 2 Jun 2015 15:44:17 +0000 (15:44 +0000)]
Modernize from(1) a bit. Use getline() to support longer lines and
document the environment variables used. OK deraadt@
mpi [Tue, 2 Jun 2015 15:32:53 +0000 (15:32 +0000)]
Mention hpb(4), prodded by jmc@
mpi [Tue, 2 Jun 2015 15:26:19 +0000 (15:26 +0000)]
Add missing define for hpb(4).
mpi [Tue, 2 Jun 2015 13:54:50 +0000 (13:54 +0000)]
Man page for hpb(4).
mpi [Tue, 2 Jun 2015 13:53:43 +0000 (13:53 +0000)]
Add a new HT-PCI bridge driver and the necessary glue to openpic(4)
to get interrupts working on U4 machines.
With this OpenBSD can run on PowerMac11,2 (Quad G5).
With inputs from and ok kettenis@, miod@, dlg@
mpi [Tue, 2 Jun 2015 13:23:55 +0000 (13:23 +0000)]
RIP ether_input_mbuf().
mpi [Tue, 2 Jun 2015 13:21:21 +0000 (13:21 +0000)]
Use if_input() instead of ether_input_mbuf().
ok jasper@, bluhm@
mpi [Tue, 2 Jun 2015 09:38:24 +0000 (09:38 +0000)]
Take carp(4) out of ether_input().
Tested by <mxb AT alumni DOT chalmers DOT se>, thanks!
ok bluhm@, dlg@
djm [Tue, 2 Jun 2015 09:10:40 +0000 (09:10 +0000)]
mention CheckHostIP adding addresses to known_hosts;
bz#1993; ok dtucker@
deraadt [Tue, 2 Jun 2015 06:10:14 +0000 (06:10 +0000)]
missing " spotted by Kyle Milz