nicm [Wed, 13 Aug 2014 08:42:46 +0000 (08:42 +0000)]
Spelling - "alignment" not "alignement". ok ratchov
chrisz [Wed, 13 Aug 2014 08:08:55 +0000 (08:08 +0000)]
fix early loop termination in httpd path_info()
without this fix httpd always put at least the first
path component in SCRIPT_NAME even when it did not exist.
Now for completely non-existant paths everything goes into
PATH_INFO.
jsg [Wed, 13 Aug 2014 07:45:37 +0000 (07:45 +0000)]
match on Intel 9 series and 9 series LP PCH, and add 8 series KT to puc
ok millert@
jsg [Wed, 13 Aug 2014 07:36:57 +0000 (07:36 +0000)]
regen
jsg [Wed, 13 Aug 2014 07:36:21 +0000 (07:36 +0000)]
Add some more 9 series ids, correct some existing ones and add
9 series LP/Wildcat Point-LP entries.
ok millert@
deraadt [Wed, 13 Aug 2014 06:04:10 +0000 (06:04 +0000)]
munmap correct object in (extremely unlikely, and effectively terminal)
case of failing to map the 2nd object.
found by Paul Maurers
dlg [Wed, 13 Aug 2014 01:03:56 +0000 (01:03 +0000)]
use errc instead of juggling errno values in tftpd_listen.
dlg [Wed, 13 Aug 2014 01:00:16 +0000 (01:00 +0000)]
setsockopt sets errno on failure, so we can use err instead of errx when
handling the error.
schwarze [Tue, 12 Aug 2014 20:36:41 +0000 (20:36 +0000)]
Merge a patch that i successfully pushed to OpenSSL,
original OpenSSL commit message follows:
Fixed as shown; to be released post-1.0.2
commit
bebbb11d132cc149f7713d6693703f8bfae10072
Author: Ingo Schwarze <schwarze@usta.de>
Date: Sat Jan 18 11:46:25 2014 +0100
RT3239: Extra comma in NAME lines of two manpages
In two OpenSSL manual pages, in the NAME section, the last word of the
name list is followed by a stray trailing comma. While this may seem
minor, it is worth fixing because it may confuse some makewhatis(8)
implementations.
While here, also add the missing word "size" to the one line
description in SSL_CTX_set_max_cert_list(3).
Reviewed by: Dr Stephen Henson <shenson@drh-consultancy.co.uk>
miod [Tue, 12 Aug 2014 19:35:28 +0000 (19:35 +0000)]
Lower VM_PHYSSEG_MAX from the mips64 default to a generous 4, and change
allocation strategy to VM_PSTRAT_BIGFIRST.
miod [Tue, 12 Aug 2014 19:33:59 +0000 (19:33 +0000)]
addu -> PTR_ADDU when doing address arithmetic in the octeon-specific code path.
schwarze [Tue, 12 Aug 2014 19:27:57 +0000 (19:27 +0000)]
In mdoc(7) and man(7), if a width is given as a bare number without
specifying a unit, the implied unit is 'n' (on the terminal, one
character position; in PostScript, half of the current font size
in points), not 'u' (roff output device basic unit). No functional
change right now, but important for the upcoming scaling unit fixes.
schwarze [Tue, 12 Aug 2014 19:19:42 +0000 (19:19 +0000)]
The macro SCALE_HS_INIT() is always passed the result of strlen() or
an equivalent number as its argument, and strlen() measures the width
of a string in characters, not in basic units. No functional change
right now, but important for the upcoming scaling unit fixes.
mikeb [Tue, 12 Aug 2014 15:29:33 +0000 (15:29 +0000)]
Finally implement what's stated in the man page regarding parent
anchors for "once" rules: "In case this is the only rule in the
anchor, the anchor will be destroyed automatically after the rule
is matched." Employ an additional pointer pair to keep track of
the parent ruleset containing the anchor that we want to remove.
OK henning
bcook [Tue, 12 Aug 2014 15:02:52 +0000 (15:02 +0000)]
Replace intrinsic ROTATE macros with an inline.
Without the cast/mask, the compiler is allowed to optimize this directly
to the correct CPU intrinsic for rotate.
mikeb [Tue, 12 Aug 2014 14:42:06 +0000 (14:42 +0000)]
Make sure that pf_step_into_anchor always saves a pointer to the rule
that owns the anchor on the pf anchor stack. There's no reason why we
should check for depth here. As a side effect this makes sure that the
correct nested anchor gets it's counter bumped instead of the top most.
For the save/restore symmetry pf_step_out_of_anchor is made to always
restore previous value of the anchor rule. depth == 0 means what we a
at the top (main ruleset).
OK henning
mikeb [Tue, 12 Aug 2014 14:38:27 +0000 (14:38 +0000)]
Apart from some minor code reshuffling the big change is that we
start with a ruleset pointer assigned to pf_main_ruleset so that
pf_purge_rule doesn't get called with a NULL.
Prompted by the discussion with Alexandr Nedvedicky <alexandr !
nedvedicky at oracle ! com>.
OK henning
mpi [Tue, 12 Aug 2014 13:52:08 +0000 (13:52 +0000)]
route_cb is internal to rtsock.c
miod [Tue, 12 Aug 2014 04:30:21 +0000 (04:30 +0000)]
Better cache information gathering. Only affects information put in dmesg.
miod [Tue, 12 Aug 2014 04:29:05 +0000 (04:29 +0000)]
Revert 1.173 now that the real cause of the octeon regression has been fixed.
miod [Tue, 12 Aug 2014 04:28:07 +0000 (04:28 +0000)]
Pass 0 instead of uvm_map_hint() to uvm_map() in exec_md_map() to figure out
where to put the fpu assist page, for uvm_map_hint() may return an address
outside userland bounds due to aggressive randomization. Passing zero will
still get a random address, but correctly bounded.
dlg [Tue, 12 Aug 2014 01:31:43 +0000 (01:31 +0000)]
sigh. when returning ENOENT in the sysctl path, unlock on the way out.
dlg [Tue, 12 Aug 2014 01:25:21 +0000 (01:25 +0000)]
i accidentally removed the check for whether the requested pool in
the sysctl path exists. return ENOENT instead of trying a NULL
deref.
dlg [Tue, 12 Aug 2014 01:05:46 +0000 (01:05 +0000)]
bring back r1.135:
matthew@ noticed i wasnt populating npages in the kinfo_pool sent to
userland.
dlg [Tue, 12 Aug 2014 01:01:11 +0000 (01:01 +0000)]
bring back r1.134:
inline is the new __inline
dlg [Tue, 12 Aug 2014 00:59:27 +0000 (00:59 +0000)]
bring back r1.133. this is a bit different cos we're still using splvm to
protect pool_list rather than the rwlock that made i386 blow up:
use pool_count to report the number of pools to userland rather
than walking the list and counting the elements as we go.
use sysctl_rdint, sysctl_rdstring, and sysctl_rdstruct instead of
handcrafted copyouts.
nicm [Mon, 11 Aug 2014 22:39:57 +0000 (22:39 +0000)]
No need to repeat other-end more than once, from Juho Pohjala. Also add
it to the commands list while here.
nicm [Mon, 11 Aug 2014 22:18:16 +0000 (22:18 +0000)]
Fix two copy mode problems:
1. In vi mode the selection doesn't include the last character if you
moved the cursor up or left.
2. In emacs mode the selection includes the last character if you moved
the cursor to the left.
From Balazs Kezes.
nicm [Mon, 11 Aug 2014 22:14:30 +0000 (22:14 +0000)]
Add flags to selectp to enable and disable input to a pane, from Anish
Athalye.
millert [Mon, 11 Aug 2014 20:30:51 +0000 (20:30 +0000)]
regen
millert [Mon, 11 Aug 2014 20:30:22 +0000 (20:30 +0000)]
Add some Intel Z97 chipset devices; ok deraadt@
guenther [Mon, 11 Aug 2014 20:28:47 +0000 (20:28 +0000)]
Still need to separately set FD_CLOEXEC if the new fd was >= FDBASE.
Affects scripts that directly use 9 of the first 10 fds.
noted by miod@
ajacoutot [Mon, 11 Aug 2014 20:24:57 +0000 (20:24 +0000)]
Typo.
ok kili@
kili [Mon, 11 Aug 2014 20:16:04 +0000 (20:16 +0000)]
Let _rc_parse_conf stick to parsing /etc/rc.conf and /etc/rc.conf.local
if it's invoked without parameters. It's simpler than adding adding
/etc/rc.conf and /etc/rc.conf.local parameters whereever _rc_parse_conf
is used (e.g. /etc/rc and /etc/netstart).
While here, replace a
for foo in "$@"; do something; done
by
for foo; do something; done
ok aja@
miod [Mon, 11 Aug 2014 19:16:56 +0000 (19:16 +0000)]
Do not pmap_update(pmap_kernel()) within the `create mappings' loop, but only
once after it is over.
miod [Mon, 11 Aug 2014 19:00:50 +0000 (19:00 +0000)]
Do not extern octeon_boot_info, <machine/octeonvar.h> declares it for you.
miod [Mon, 11 Aug 2014 18:56:49 +0000 (18:56 +0000)]
Do not use CKSEG0 to quickly map physical addresses, but XKPHYS, for we are
not limited to 512MB physmem.
miod [Mon, 11 Aug 2014 18:52:54 +0000 (18:52 +0000)]
Get rid of the `quirks' member, which is never initialized (but by zero at
softc creation) anyway. Sorry, we are not trying to support SEIL-specific
hardware at this point.
tobias [Mon, 11 Aug 2014 18:41:13 +0000 (18:41 +0000)]
When parsing 32 bit values, verify that we received 4 bytes.
ok krw@
tobias [Mon, 11 Aug 2014 18:39:41 +0000 (18:39 +0000)]
validate len field for proper length, not just "not zero."
ok krw@
miod [Mon, 11 Aug 2014 18:33:36 +0000 (18:33 +0000)]
-current dammit
miod [Mon, 11 Aug 2014 18:29:56 +0000 (18:29 +0000)]
Remove evcnt and %b format strings relying upon SEIL extensions. This gives
OCTEON_ETH_DEBUG kernels a chance to build. No functional change for regular
kernels.
miod [Mon, 11 Aug 2014 18:08:17 +0000 (18:08 +0000)]
De-static so that my ddb backtrace get closer to reality.
miod [Mon, 11 Aug 2014 16:54:20 +0000 (16:54 +0000)]
Add option USER_PCICONF for pcidump(8) to be able to do its magic.
deraadt [Mon, 11 Aug 2014 15:26:33 +0000 (15:26 +0000)]
make a few variables more local
mpi [Mon, 11 Aug 2014 13:51:07 +0000 (13:51 +0000)]
Use the RTF_LOCAL flag instead of abusing the RTF_LLINFO flag when
adding local route entries.
This hack made sense when we didn't have the RTF_LOCAL flag, but since
some months it is set on every local route.
dlg [Mon, 11 Aug 2014 13:31:42 +0000 (13:31 +0000)]
bring back r1.132:
provide a pool_count global so we can figure out how many pools there are
active without having to walk the global pool_list.
bcook [Mon, 11 Aug 2014 13:29:43 +0000 (13:29 +0000)]
Guard RSA / RC4-5 ASM when NO_ASM is not defined
Most assembly blocks remain inactive if OPENSSL_NO_ASM is not defined,
only enabling inline assembly, but the RSA / RC4-5 blocks (used only in
amd64 systems) turn on implicitly. Guard these two as well.
This simplifies enabling just inline ASM in portable, no effective
change in OpenBSD.
ajacoutot [Mon, 11 Aug 2014 13:25:23 +0000 (13:25 +0000)]
Make it possible to pass arguments to _rc_parse_conf().
ok robert@
jmc [Mon, 11 Aug 2014 12:57:14 +0000 (12:57 +0000)]
document "tc";
From: Edward
ok millert
mpi [Mon, 11 Aug 2014 12:45:45 +0000 (12:45 +0000)]
Fewer <netinet/in_systm.h>
dlg [Mon, 11 Aug 2014 12:37:36 +0000 (12:37 +0000)]
bring back r1.131:
take the pools mutex when copying stats out of it in the sysctl
path so we are guaranteed a consistent snapshot.
mpi [Mon, 11 Aug 2014 11:59:05 +0000 (11:59 +0000)]
Put back the checks about RTF_LOCAL routes now that userland tools are
aware of them.
Original commit message was:
Reserve the highest route priority for kernel-managed routes and prevent
userland from playing with the local and broadcast flags.
ok claudio@
mpi [Mon, 11 Aug 2014 11:50:41 +0000 (11:50 +0000)]
Convert ARP llinfo allocation to pool(9).
ok henning@, dlg@
jsing [Mon, 11 Aug 2014 10:46:19 +0000 (10:46 +0000)]
Check the return value of sk_SSL_CIPHER_new_null(), since it allocates
memory and can return NULL.
ok miod@
gilles [Mon, 11 Aug 2014 09:50:51 +0000 (09:50 +0000)]
missing memset in table_api.c (currently only used by pluggable tables)
diff from Martijn van Duren
mpi [Mon, 11 Aug 2014 09:47:56 +0000 (09:47 +0000)]
Do not try to delete RTF_LOCAL entries. Such routes are automatically
created /deleted by the kernel when an IP address is added/removed
on/from an interface and should not be manipulated by userland tools.
ok henning@, jca@
ajacoutot [Mon, 11 Aug 2014 09:03:27 +0000 (09:03 +0000)]
Steal stripcom() from /etc/rc to parse sysmerge.ignore.
jmc [Mon, 11 Aug 2014 08:21:55 +0000 (08:21 +0000)]
add a caveat about databases;
From: Matthew Weigel
ok gilles
ajacoutot [Mon, 11 Aug 2014 06:30:39 +0000 (06:30 +0000)]
Move a sentence to the corresponding option description.
from jmc@
miod [Mon, 11 Aug 2014 04:46:42 +0000 (04:46 +0000)]
Unchecked memory allocation and potential leak upon error in
ssl3_get_cert_verify().
ok guenther@ jsing@
miod [Mon, 11 Aug 2014 04:45:19 +0000 (04:45 +0000)]
Remove now-unused SSL2_STATE as well as ssl2-specific state machine values.
ok guenther@ jsing@
schwarze [Mon, 11 Aug 2014 04:18:18 +0000 (04:18 +0000)]
start catching up with recent code changes, but no new files yet
jsing [Mon, 11 Aug 2014 01:10:42 +0000 (01:10 +0000)]
Currently, ssl3_put_char_by_bytes(NULL, NULL) is just a long handed way
of writing "2". Add a define for the SSL3_CIPHER_VALUE_SIZE (rather than
using a less-readable hardcoded constant everywhere) and replace the
ssl3_put_char_by_bytes(NULL, NULL) calls with it.
ok bcook@ miod@
jsing [Mon, 11 Aug 2014 01:06:22 +0000 (01:06 +0000)]
Provide a ssl3_get_cipher_by_id() function that allows ciphers to be looked
up by their ID. For one, this avoids an ugly mess in ssl_sess.c, where the
cipher value is manually written into a buffer, just so the cipher can be
located using ssl3_get_cipher_by_char().
ok bcook@ miod@
bluhm [Sun, 10 Aug 2014 23:44:20 +0000 (23:44 +0000)]
Write the core file of a non-suid process into the current directory
even if sysctl kern.nosuidcoredump has been set to 2 or 3. This
allows a regular user to debug his programs again.
OK guenther@ deraadt@
bluhm [Sun, 10 Aug 2014 21:55:17 +0000 (21:55 +0000)]
There was a race in relayd that caused connections to hang. It
happend with non-persistent PUT connections that had a very short
body. If the whole body was read from the client before the
connection to the server was set up, the event callback was not
called. Do the regular checks after relay_connect() succeeded.
OK reyk@
zhuk [Sun, 10 Aug 2014 21:09:26 +0000 (21:09 +0000)]
Improve variables listing in x11/kde4 module description.
rpe [Sun, 10 Aug 2014 20:54:17 +0000 (20:54 +0000)]
[] -> [[]]
OK krw@
"well ok" halex@
jmc [Sun, 10 Aug 2014 17:15:18 +0000 (17:15 +0000)]
- add -p to SYNOPSIS
- the argument to -s is not optional
ok ajacoutot
tedu [Sun, 10 Aug 2014 16:36:13 +0000 (16:36 +0000)]
guenther wants a copyright
rpe [Sun, 10 Aug 2014 16:32:28 +0000 (16:32 +0000)]
Convert if foo; then bar; fi blocks to foo && bar but leave out enable_network()
because it shares code&style with /etc/netstart. No functional change.
with feedback and OK krw@ halex@
jsing [Sun, 10 Aug 2014 15:06:15 +0000 (15:06 +0000)]
Tweak cipher list comments and add missing cipher value comments.
rpe [Sun, 10 Aug 2014 14:59:22 +0000 (14:59 +0000)]
{ foo ; bar ; } -> { foo; bar; }
OK krw@
jsing [Sun, 10 Aug 2014 14:57:04 +0000 (14:57 +0000)]
Remove disabled (weakened export and non-ephemeral DH) cipher suites from
the cipher list. This reduces code size, saves data segment space and
prevents them from being turned back on at runtime by flipping a bit in
memory.
ok guenther@
jsing [Sun, 10 Aug 2014 14:42:55 +0000 (14:42 +0000)]
Since we no longer need to support SSLv2-style cipher lists, start
unravelling the maze of function pointers and callbacks by directly
calling ssl3_{get,put}_cipher_by_char() and removing the
ssl_{get,put}_cipher_by_char macros.
Prompted by similar changes in boringssl.
ok guenther.
mpi [Sun, 10 Aug 2014 13:32:14 +0000 (13:32 +0000)]
Ask for the descriptor size corresponding to the number of ports present
in the hub, not from some random value from the stack.
mpi [Sun, 10 Aug 2014 12:58:49 +0000 (12:58 +0000)]
Do not allocate space for the whole structure when we just want to store
a pointer, found by clang.
mpi [Sun, 10 Aug 2014 12:48:43 +0000 (12:48 +0000)]
sizeof() a pointer of the right struct to appease clang.
sf [Sun, 10 Aug 2014 12:28:09 +0000 (12:28 +0000)]
Note that qemu 1.7.2 has the VLAN bug fix, too
jasper [Sun, 10 Aug 2014 11:26:07 +0000 (11:26 +0000)]
zap trailing whitespace
jmc [Sun, 10 Aug 2014 11:23:02 +0000 (11:23 +0000)]
some Ev fixes;
mpi [Sun, 10 Aug 2014 11:21:49 +0000 (11:21 +0000)]
Merge xhci_device_setup() into xhci_pipe_init() there's no reason to
have a separate function anymore, it is just a wrapper around the "set
address" command.
mpi [Sun, 10 Aug 2014 11:18:57 +0000 (11:18 +0000)]
Since USB xfer pools are accessed in interrupt context, initialize them
with the correct ipl to prevent your CPU from locking against itself.
mpi [Sun, 10 Aug 2014 11:00:35 +0000 (11:00 +0000)]
Set and check for XFER_BUSY in the common methods instead of doing it
in every HC driver.
ratchov [Sun, 10 Aug 2014 10:25:35 +0000 (10:25 +0000)]
typo in debug message
ratchov [Sun, 10 Aug 2014 10:25:04 +0000 (10:25 +0000)]
initialize mmcpos, found by guenther
espie [Sun, 10 Aug 2014 10:10:21 +0000 (10:10 +0000)]
Document SUBST_*
espie [Sun, 10 Aug 2014 10:04:39 +0000 (10:04 +0000)]
comment what's going on
espie [Sun, 10 Aug 2014 10:03:46 +0000 (10:03 +0000)]
Ouch... fuck cvs
espie [Sun, 10 Aug 2014 10:01:03 +0000 (10:01 +0000)]
if pkg_create is run as non-root, restory correct group/owner to root/bin.
also, remove write permissions without explicit modes. Allows fake installs
to keep directories/files writable while producing correct package.
jsg [Sun, 10 Aug 2014 09:23:06 +0000 (09:23 +0000)]
fix kqueue read/write filters for msdosfs and fusefs
ok tedu@
matthieu [Sun, 10 Aug 2014 09:15:39 +0000 (09:15 +0000)]
Typo. ok deraadt@ jmc@
jsg [Sun, 10 Aug 2014 09:11:07 +0000 (09:11 +0000)]
debug level test for clock_gettime() calls should match
ok ratchov@
guenther [Sun, 10 Aug 2014 07:31:58 +0000 (07:31 +0000)]
AF_IMPLINK and AF_BLUETOOTH are gone, but add printing of SOCK_SEQPACKET
guenther [Sun, 10 Aug 2014 07:29:45 +0000 (07:29 +0000)]
Don't need to know how to format bluetooth socket info here
guenther [Sun, 10 Aug 2014 07:28:32 +0000 (07:28 +0000)]
No more bluetooth in netintr(), so delete NETISR_BT. Zap the declaration
for btintr(), as well as the also gone atintr() and clnlintr()
guenther [Sun, 10 Aug 2014 07:23:59 +0000 (07:23 +0000)]
AF_BLUETOOTH is gone. AF_LOCAL isn't involved pipe (and portals are gone).
guenther [Sun, 10 Aug 2014 07:19:08 +0000 (07:19 +0000)]
AF_IMPLINK and AF_BLUETOOTH are gone
guenther [Sun, 10 Aug 2014 05:09:31 +0000 (05:09 +0000)]
Only need <stdint.h> and not all of <inttypes.h> here