openbsd
9 years agoRewrite of the target resolution internals to be simpler and more
nicm [Mon, 27 Apr 2015 16:25:57 +0000 (16:25 +0000)]
Rewrite of the target resolution internals to be simpler and more
consistent but with much less duplication, but keeping the same internal
API. Also adds more readable aliases for some of the special tokens used
in targets (eg "{start}" instead of "^"). Some behaviours may have
changed, for example prefix matches now happen before fnmatch.

9 years agoDo not call nd6_purge() before purging the IPv6 addresses of a detached
mpi [Mon, 27 Apr 2015 14:51:44 +0000 (14:51 +0000)]
Do not call nd6_purge() before purging the IPv6 addresses of a detached
interface.

Fix a use after free introduced in r1.98 of netinet6/in6.c and recently
exposed by a crazy pool/malloc damage finder being currently refined by
dlg@ and deraadt@.

ok mikeb@, henning@

9 years agoUse a systrace(4) sandbox with a short whitelist of allowed syscalls for
nicm [Mon, 27 Apr 2015 13:52:17 +0000 (13:52 +0000)]
Use a systrace(4) sandbox with a short whitelist of allowed syscalls for
the file(1) child process. Based on similar code in ssh sandbox-systrace.c.
Idea and help from deraadt@.

9 years agoAdd a _file user and use for privsep, ok deraadt
nicm [Mon, 27 Apr 2015 13:48:06 +0000 (13:48 +0000)]
Add a _file user and use for privsep, ok deraadt

9 years agoAdd simple privilege separation to file(1). Two processes, file
nicm [Mon, 27 Apr 2015 13:41:45 +0000 (13:41 +0000)]
Add simple privilege separation to file(1). Two processes, file
descriptors and a few other bits are opened in parent and passed to
child using imsg. Child currently drops to "nobody" but this will change.

9 years agoAdjust ld semantics to make static PIE the default. Forgotten by pascal@.
kettenis [Mon, 27 Apr 2015 13:25:10 +0000 (13:25 +0000)]
Adjust ld semantics to make static PIE the default.  Forgotten by pascal@.

Original commit message:

Change gcc and ld semantics to make static PIE the default when invoking
'cc -static'.  To explicitly request the legacy behaviour, use -nopie.

For the few port affected by this, bumps will follow shortly.

looks good to kettenis@, ok kurt@

9 years agoremove some extraneous text; ok nicm
jmc [Mon, 27 Apr 2015 11:12:49 +0000 (11:12 +0000)]
remove some extraneous text; ok nicm

9 years agosync with rev 1.99 of sbin/route/show.c
benno [Mon, 27 Apr 2015 09:54:04 +0000 (09:54 +0000)]
sync with rev 1.99 of sbin/route/show.c
requested by claudio@ and mpi@

9 years agoroute show does not need to filter unwanted af itself, the sysctl does
benno [Mon, 27 Apr 2015 09:51:58 +0000 (09:51 +0000)]
route show does not need to filter unwanted af itself, the sysctl does
that for us.
approach seems sound deraadt@
ok claudio@ mpi@ henning@ phessler@

9 years agoSince upd(4) currently supports a known but limited number of sensors,
mpi [Mon, 27 Apr 2015 09:14:45 +0000 (09:14 +0000)]
Since upd(4) currently supports a known but limited number of sensors,
parse the HID descriptor multiple times to find them.

This logic is necessary to later create a tree of sensors in order to
avoid lookups in the hot path for sensors that depend on the value of
others.

From David Higgs.

9 years agoUpdate list of card readers supported by rtsx(4). From brad.
stsp [Mon, 27 Apr 2015 09:07:49 +0000 (09:07 +0000)]
Update list of card readers supported by rtsx(4). From brad.

9 years agoIf the requested pane is already active, do not unzoom the window (or do
nicm [Mon, 27 Apr 2015 07:49:36 +0000 (07:49 +0000)]
If the requested pane is already active, do not unzoom the window (or do
anything else). Prevents mouse clicking when zoomed causing unzoom,
reported by Jose Antonio Delgado Alfonso (with a different fix).

9 years agowith more unit convertion when new sensors will appear.
mpi [Mon, 27 Apr 2015 07:41:41 +0000 (07:41 +0000)]
with more unit convertion when new sensors will appear.

From David Higgs.

9 years agoMake use of DEVNAME(), from David Higgs.
mpi [Mon, 27 Apr 2015 07:37:19 +0000 (07:37 +0000)]
Make use of DEVNAME(), from David Higgs.

9 years agoCorrectly write the 64bits of the HID 1, 4 and 5 registers.
mpi [Mon, 27 Apr 2015 07:20:57 +0000 (07:20 +0000)]
Correctly write the 64bits of the HID 1, 4 and 5 registers.

This makes the secondary cpu of my PowerMac as fast as the primary one,
and divide the build time by 3 with a GENERIC.MP kernel on MP G5s

Found thanks to MP kernel profiling.

ok dlg@, miod@

9 years agoNot all Linux libc's include linux/sysctl.h in sys/sysctl.h.
bcook [Mon, 27 Apr 2015 03:37:06 +0000 (03:37 +0000)]
Not all Linux libc's include linux/sysctl.h in sys/sysctl.h.

Include it if we have the sysctl syscall.

9 years agoSupport AIX versions without WPAR support.
bcook [Mon, 27 Apr 2015 03:34:43 +0000 (03:34 +0000)]
Support AIX versions without WPAR support.

From Michael Felt.

9 years agomore OPENSSL=no fixes; ok dtucker@
djm [Mon, 27 Apr 2015 01:53:37 +0000 (01:53 +0000)]
more OPENSSL=no fixes; ok dtucker@

9 years agofix compilation with OPENSSL=no; ok dtucker@
djm [Mon, 27 Apr 2015 01:52:30 +0000 (01:52 +0000)]
fix compilation with OPENSSL=no; ok dtucker@

9 years agoInclude stdio.h for FILE (used in sshkey.h) so it compiles with OPENSSL=no.
dtucker [Mon, 27 Apr 2015 00:37:53 +0000 (00:37 +0000)]
Include stdio.h for FILE (used in sshkey.h) so it compiles with OPENSSL=no.

9 years agoallow "sshd -f none" to skip reading the config file, much like
djm [Mon, 27 Apr 2015 00:21:21 +0000 (00:21 +0000)]
allow "sshd -f none" to skip reading the config file, much like
"ssh -F none" does. ok dtucker

9 years agoDon't support -s on FIFOs, it doesn't work well and the workarounds are
nicm [Sun, 26 Apr 2015 22:51:32 +0000 (22:51 +0000)]
Don't support -s on FIFOs, it doesn't work well and the workarounds are
a bit horrible.

9 years agoRemove panes from layout if spawning them fails, reported by Anthony J
nicm [Sun, 26 Apr 2015 20:25:20 +0000 (20:25 +0000)]
Remove panes from layout if spawning them fails, reported by Anthony J
Bentley.

9 years agomlarkin asks "bgpctl checks the length of the control socket path to
benno [Sun, 26 Apr 2015 20:12:03 +0000 (20:12 +0000)]
mlarkin asks "bgpctl checks the length of the control socket path to
make sure it fits. When browsing around last night I saw that bgpd
does not. Any reason it shouldn't? Please commit"

Add a check in parse.y to check this when reading the configuration.
ok phessler@ henning@

9 years agostat() the original link path not the resolved one which may be relative.
nicm [Sun, 26 Apr 2015 19:53:50 +0000 (19:53 +0000)]
stat() the original link path not the resolved one which may be relative.

9 years agoFix typo in the buffer size value: 7680 in the manpage vs 7860 in
dcoppa [Sun, 26 Apr 2015 17:26:59 +0000 (17:26 +0000)]
Fix typo in the buffer size value: 7680 in the manpage vs 7860 in
the sndiod.c code.
Initially, I thought the error was in the manpage, but Alexander
told me it was the code, so fix the code instead.

ok ratchov@

9 years agoget_responsefile: instead of keeping the dhcp-supplied next-server where
henning [Sun, 26 Apr 2015 14:49:23 +0000 (14:49 +0000)]
get_responsefile: instead of keeping the dhcp-supplied next-server where
we fetch the response file from in a local _server var, put it in an
exported AI_SERVER one. last not least that allows install.site to see it.
ok krw rpe

9 years agoHave vio_start() check if the queue is empty.
sf [Sun, 26 Apr 2015 12:27:29 +0000 (12:27 +0000)]
Have vio_start() check if the queue is empty.

from brad@

9 years agoGet dwc2 working on octeon.
jmatthew [Sun, 26 Apr 2015 12:24:03 +0000 (12:24 +0000)]
Get dwc2 working on octeon.
- transplant the clock setup code from octhci
- add a bus space tag to deal with dwc2 using little endian addressing
- bump up the rx fifo size, necessary for umass/sd to work

tested on an edgerouter lite, which can almost boot by itself now
ok uebayasi@ (various parts), miod@ (bus space bits)

9 years agovio: Support checksum offloading for IPv4 TX
sf [Sun, 26 Apr 2015 12:19:24 +0000 (12:19 +0000)]
vio: Support checksum offloading for IPv4 TX

"Looks good to me" brad@

9 years agoregen
stsp [Sun, 26 Apr 2015 11:47:22 +0000 (11:47 +0000)]
regen

9 years agoAdd IDs for RTL8188ETV and RTL8188EU. From FreeBSD via Mikhail on tech@
stsp [Sun, 26 Apr 2015 11:47:13 +0000 (11:47 +0000)]
Add IDs for RTL8188ETV and RTL8188EU. From FreeBSD via Mikhail on tech@

9 years agofor every policy we write out, flush the output so we don't get an
phessler [Sun, 26 Apr 2015 11:32:54 +0000 (11:32 +0000)]
for every policy we write out, flush the output so we don't get an
irritating partially written line

9 years agoDisable PAE when switching to the hibernate resume pagetables. This involves
kettenis [Sun, 26 Apr 2015 11:09:32 +0000 (11:09 +0000)]
Disable PAE when switching to the hibernate resume pagetables.  This involves
a slightly conmplicated dance where we stash the PAE PDPTEs into the
hibernate resume pagetables and use those before turning off PAE.
Makes (un)hibernate work with the new PAE pmap.

ok mlarkin@

9 years agoAdd Digitus DN-7003GT to list of supported urtw(4) devices.
stsp [Sun, 26 Apr 2015 10:22:13 +0000 (10:22 +0000)]
Add Digitus DN-7003GT to list of supported urtw(4) devices.

9 years agoregen
stsp [Sun, 26 Apr 2015 10:18:30 +0000 (10:18 +0000)]
regen

9 years agoAdd USB device ID for RTL8812AU. Found in "TP-Link AC1200 T4U" device.
stsp [Sun, 26 Apr 2015 10:18:19 +0000 (10:18 +0000)]
Add USB device ID for RTL8812AU. Found in "TP-Link AC1200 T4U" device.

9 years agoEnable NX support in the resume path. Makes suspend/resume work with the
kettenis [Sun, 26 Apr 2015 09:49:42 +0000 (09:49 +0000)]
Enable NX support in the resume path.  Makes suspend/resume work with the
PAE pmap.

ok deraadt@, mlarkin@

9 years agoOnly enable PAE if the CPU we're running on has NX support. Without NX
kettenis [Sun, 26 Apr 2015 09:48:29 +0000 (09:48 +0000)]
Only enable PAE if the CPU we're running on has NX support.  Without NX
support we're only wasting memory on the larger PAE page tables without
any real benefit.  This allows some simplifications of the low-level
assembly code.

ok mlarkin@, deraadt@

9 years agobump i386 MAXDSIZ to 3GB. - "If you're running this, and presumably actually
sthen [Sun, 26 Apr 2015 09:22:33 +0000 (09:22 +0000)]
bump i386 MAXDSIZ to 3GB. - "If you're running this, and presumably actually
using that much memory, go for it" tedu@ "I don't see any immediate downsides"
kettenis@

9 years agosync
deraadt [Sun, 26 Apr 2015 07:25:10 +0000 (07:25 +0000)]
sync

9 years agoTwo extra messages required after sending a baud rate update, as observed
jmatthew [Sun, 26 Apr 2015 06:38:04 +0000 (06:38 +0000)]
Two extra messages required after sending a baud rate update, as observed
in the linux mct_u232 driver, which apparently got them by sniffing usb
traffic from the vendor's windows 98 driver.  Makes this device work at
115200:

umct0 at uhub0 port 2 "Belkin Components F5U109 Serial" rev 1.10/1.02 addr 2

ok dlg@

9 years agoRequire a PT_LOAD segment's p_filesz to be no larger than its p_memsz.
guenther [Sun, 26 Apr 2015 05:30:42 +0000 (05:30 +0000)]
Require a PT_LOAD segment's p_filesz to be no larger than its p_memsz.

test cases provided by Alejandro Herna'ndez (nitrousenador (at) gmail.com)
ok deraadt@ jsg@

9 years agoFix glitches in previous commit: strip the \n and only complain on failure
guenther [Sun, 26 Apr 2015 01:23:19 +0000 (01:23 +0000)]
Fix glitches in previous commit: strip the \n and only complain on failure

ok deraadt@

9 years agosome people are capitalizing keywords, so ignore case when we test the tokens
phessler [Sat, 25 Apr 2015 21:44:26 +0000 (21:44 +0000)]
some people are capitalizing keywords, so ignore case when we test the tokens

OK henning@ benno@

9 years agoCut the aliases part of /etc/rpc lines before feeding them to strtonum(), for
miod [Sat, 25 Apr 2015 21:38:22 +0000 (21:38 +0000)]
Cut the aliases part of /etc/rpc lines before feeding them to strtonum(), for
it will fail otherwise; ok deraadt@ millert@

9 years agoWe now following the ABI and always clear cld on function entry, so remove
guenther [Sat, 25 Apr 2015 21:31:24 +0000 (21:31 +0000)]
We now following the ABI and always clear cld on function entry, so remove
the extra CLD instructions from when that wasn't true

testing miod@ krw@

9 years agocpu_busy_cycle_fcn callback has never been used
guenther [Sat, 25 Apr 2015 21:21:02 +0000 (21:21 +0000)]
cpu_busy_cycle_fcn callback has never been used

9 years agoOnly print MIME warnings when warnings are enabled.
nicm [Sat, 25 Apr 2015 21:17:09 +0000 (21:17 +0000)]
Only print MIME warnings when warnings are enabled.

9 years agoRepair boot device detection when booting off the second SCSI controller on
miod [Sat, 25 Apr 2015 21:15:08 +0000 (21:15 +0000)]
Repair boot device detection when booting off the second SCSI controller on
AV530.

9 years agoDisable kq-tun test, as it's been broken for a while and not because of kq
guenther [Sat, 25 Apr 2015 20:47:49 +0000 (20:47 +0000)]
Disable kq-tun test, as it's been broken for a while and not because of kq

9 years agoApply pre-built unicore patch, remove excess files - perl-5.20.2
afresh1 [Sat, 25 Apr 2015 19:30:14 +0000 (19:30 +0000)]
Apply pre-built unicore patch, remove excess files - perl-5.20.2

9 years agoApply local patches, remove excess files - perl-5.20.2
afresh1 [Sat, 25 Apr 2015 19:26:15 +0000 (19:26 +0000)]
Apply local patches, remove excess files - perl-5.20.2

9 years agoFix merge issues, remove excess files - match perl-5.20.2 dist
afresh1 [Sat, 25 Apr 2015 19:14:36 +0000 (19:14 +0000)]
Fix merge issues, remove excess files - match perl-5.20.2 dist

9 years agoImport perl-5.20.2
afresh1 [Sat, 25 Apr 2015 19:10:05 +0000 (19:10 +0000)]
Import perl-5.20.2

9 years agoGet rid of window_choose_list type.
nicm [Sat, 25 Apr 2015 18:56:05 +0000 (18:56 +0000)]
Get rid of window_choose_list type.

9 years agoFix some char* -> char *.
nicm [Sat, 25 Apr 2015 18:49:01 +0000 (18:49 +0000)]
Fix some char* -> char *.

9 years agoUse a char **,u_int pair for cfg_causes.
nicm [Sat, 25 Apr 2015 18:47:01 +0000 (18:47 +0000)]
Use a char **,u_int pair for cfg_causes.

9 years agoFail if a \ appears at EOL rather than continuing off the end of the
nicm [Sat, 25 Apr 2015 18:44:28 +0000 (18:44 +0000)]
Fail if a \ appears at EOL rather than continuing off the end of the
buffer, from Sebastien Marie.

9 years agoMake message log a TAILQ.
nicm [Sat, 25 Apr 2015 18:33:59 +0000 (18:33 +0000)]
Make message log a TAILQ.

9 years agoMove the functions to convert ids from strings into session.c and window.c.
nicm [Sat, 25 Apr 2015 18:09:28 +0000 (18:09 +0000)]
Move the functions to convert ids from strings into session.c and window.c.

9 years agodouble word fix; from david vasek
jmc [Sat, 25 Apr 2015 16:37:17 +0000 (16:37 +0000)]
double word fix; from david vasek

9 years agoSwitch to using the POSIX endian macros, i.e: be64toh vs. betoh64.
brynet [Sat, 25 Apr 2015 16:35:47 +0000 (16:35 +0000)]
Switch to using the POSIX endian macros, i.e: be64toh vs. betoh64.

ok nicm@

9 years agoDon't ignore the reference count in X509_STORE_free.
doug [Sat, 25 Apr 2015 16:02:55 +0000 (16:02 +0000)]
Don't ignore the reference count in X509_STORE_free.

Based on this upstream commit: bff9ce4db38b297c72a6d84617d71ae2934450f7
which didn't make it into a release until 1.0.2.

Thanks to william at 25thandclement dot com for reporting this!

ok deraadt@ jsing@ beck@

9 years agoExplicitly cancel mouse "button" mode, this happens implicitly with some
nicm [Sat, 25 Apr 2015 15:57:48 +0000 (15:57 +0000)]
Explicitly cancel mouse "button" mode, this happens implicitly with some
one of the other things we send with xterm, but not with urxvt. Reported
by sthen@.

9 years agoadd check for overflow while doubling (very unlikely in practice, but still
espie [Sat, 25 Apr 2015 15:33:47 +0000 (15:33 +0000)]
add check for overflow while doubling (very unlikely in practice, but still
better style code). Problem noticed by deraadt@ in m4.

okay doug@ deraadt@

9 years agoCheck for invalid leading zeros in CBS_get_asn1_uint64.
doug [Sat, 25 Apr 2015 15:28:47 +0000 (15:28 +0000)]
Check for invalid leading zeros in CBS_get_asn1_uint64.

ASN.1 integers cannot have all zeros or all ones for the first 9 bits.
This rule ensures the numbers are encoded with the smallest number of
content octets (see ITU-T Rec X.690 section 8.3.2).

Based on BoringSSL commit 5933723b7b592e9914f703d630b596e140c93e16

ok deraadt@ jsing@

9 years agoallow us to write rules that match directly on the peer AS
phessler [Sat, 25 Apr 2015 15:28:18 +0000 (15:28 +0000)]
allow us to write rules that match directly on the peer AS

...
allow from AS 1 prefix 192.0.2.0/24
...

Also adjust the IRR ruleset output to include the declared peer AS,
instead of hoping they listed their neighbor IP address!

OK benno@
older version OK: claudio@ henning@

9 years agoPrepend files or directories containing ":" with "./" in directory
florian [Sat, 25 Apr 2015 14:40:35 +0000 (14:40 +0000)]
Prepend files or directories containing ":" with "./" in directory
indexes as per RFC 3986:
A path segment that contains a colon character (e.g., "this:that")
cannot be used as the first segment of a relative-path reference, as
it would be mistaken for a scheme name.  Such a segment must be
preceded by a dot-segment (e.g., "./this:that") to make a relative-
path reference.

While here add a "/" to the end of directory names, this saves us one
redirect round trip.

Found the hard way & "functionality wise, OK" ajacoutot@
RFC pointer & OK benno@

9 years agohandle an IRR record of "export ... action X" the same way we handle
phessler [Sat, 25 Apr 2015 13:23:01 +0000 (13:23 +0000)]
handle an IRR record of "export ... action X" the same way we handle
"import ... action X".

OK benno@ henning@

9 years agoAdd quirk for Cirrus Logic CS4208 which is needed for MacBookAir6,1.
jturner [Sat, 25 Apr 2015 11:37:24 +0000 (11:37 +0000)]
Add quirk for Cirrus Logic CS4208 which is needed for MacBookAir6,1.
ok deraadt@ jung@ ratchov@

9 years agoConvert clients list into a TAILQ.
nicm [Fri, 24 Apr 2015 23:17:11 +0000 (23:17 +0000)]
Convert clients list into a TAILQ.

9 years agoSet working directory for run-shell and if-shell.
nicm [Fri, 24 Apr 2015 22:19:36 +0000 (22:19 +0000)]
Set working directory for run-shell and if-shell.

9 years agostruct timespec/clock_gettime(3) conversion for vi(1).
brynet [Fri, 24 Apr 2015 21:48:31 +0000 (21:48 +0000)]
struct timespec/clock_gettime(3) conversion for vi(1).

ok guenther@

9 years agoAllow choice options (multiple states) to be toggled between states 0
nicm [Fri, 24 Apr 2015 21:38:18 +0000 (21:38 +0000)]
Allow choice options (multiple states) to be toggled between states 0
and 1.

9 years agoWrite (uid_t)-1 as -1 instead of UID_MAX in the v7 passwd file.
millert [Fri, 24 Apr 2015 21:14:48 +0000 (21:14 +0000)]
Write (uid_t)-1 as -1 instead of UID_MAX in the v7 passwd file.
OK deraadt@

9 years agoUse strtonum() instead of strtoul() when parsing uid/gid so we get
millert [Fri, 24 Apr 2015 21:13:56 +0000 (21:13 +0000)]
Use strtonum() instead of strtoul() when parsing uid/gid so we get
consistent handling of negative ids on 32bit/64bit systems.
The only negative uid/gid allowed is -1 which is special-cased
so it can be preserved when writing the new master.passwd file
instead of being written as an unsigned number.  OK deraadt@

9 years agono more tip(1); ok nicm
jmc [Fri, 24 Apr 2015 21:03:58 +0000 (21:03 +0000)]
no more tip(1); ok nicm

9 years agoSet up signal handler earlier so that we don't get zombies, reported by
nicm [Fri, 24 Apr 2015 20:58:44 +0000 (20:58 +0000)]
Set up signal handler earlier so that we don't get zombies, reported by
sobrado@.

9 years agoDocument -h, and bring some consistency to the option descriptions. ok jmc
nicm [Fri, 24 Apr 2015 20:57:51 +0000 (20:57 +0000)]
Document -h, and bring some consistency to the option descriptions. ok jmc

9 years agotweaks; ok nicm
jmc [Fri, 24 Apr 2015 20:32:34 +0000 (20:32 +0000)]
tweaks; ok nicm

9 years agoEnable PAE mode for those CPUs that support it. This allows us to use the
mlarkin [Fri, 24 Apr 2015 19:53:43 +0000 (19:53 +0000)]
Enable PAE mode for those CPUs that support it. This allows us to use the
NX bit for userland and kernel W^X. Unlike the previous c.2008 PAE
experiment, this does not provide > 4GB phys ram on i386 - PAE is solely
being used for NX capability this time. If you need > 4GB phys, use amd64.

Userland W^X was committed yesterday by kettenis@, and we will shortly
start reworking the kernel like we did for amd64 a few months back to get
kernel W^X.

This has been in snaps for a few days and tested by kettenis and myself
as well.

ok deraadt@, kettenis@

9 years agoMake sure we keep the whole recursive mapping of the PDP instead of just the
kettenis [Fri, 24 Apr 2015 19:41:58 +0000 (19:41 +0000)]
Make sure we keep the whole recursive mapping of the PDP instead of just the
mapping for the first page when tearing things down.  Seems to fix the last
bug mlarkin@ has been chasing for a while.

ok mlarkin@

9 years agosync
deraadt [Fri, 24 Apr 2015 18:49:54 +0000 (18:49 +0000)]
sync

9 years agoTrim down remote(5) to reflect what is supported by cu(1).
nicm [Fri, 24 Apr 2015 18:06:54 +0000 (18:06 +0000)]
Trim down remote(5) to reflect what is supported by cu(1).

ok sthen jmc

9 years agoremove the old tip SCCS note
deraadt [Fri, 24 Apr 2015 18:06:31 +0000 (18:06 +0000)]
remove the old tip SCCS note

9 years agoRemove tip(1), superseded by cu(1).
nicm [Fri, 24 Apr 2015 18:05:21 +0000 (18:05 +0000)]
Remove tip(1), superseded by cu(1).

ok deraadt sthen bmercer

9 years agoDo not attempt to use ~/.magic if running as root (or issetugid()).
nicm [Fri, 24 Apr 2015 17:34:57 +0000 (17:34 +0000)]
Do not attempt to use ~/.magic if running as root (or issetugid()).

9 years agomove an ugly space
deraadt [Fri, 24 Apr 2015 17:19:06 +0000 (17:19 +0000)]
move an ugly space

9 years agoWe need to loop over the files once to work out the width of the labels
nicm [Fri, 24 Apr 2015 17:10:50 +0000 (17:10 +0000)]
We need to loop over the files once to work out the width of the labels
before testing them.

9 years agoTrying to drop privileges means we can't handle ARG_MAX arguments, so
nicm [Fri, 24 Apr 2015 16:47:32 +0000 (16:47 +0000)]
Trying to drop privileges means we can't handle ARG_MAX arguments, so
remove it for now.

9 years agoAdd a couple of missing spaces (style nits).
nicm [Fri, 24 Apr 2015 16:45:32 +0000 (16:45 +0000)]
Add a couple of missing spaces (style nits).

9 years agoUse symbolic constants for st_mode flags, no binary change.
nicm [Fri, 24 Apr 2015 16:30:06 +0000 (16:30 +0000)]
Use symbolic constants for st_mode flags, no binary change.

9 years agoIf ~/.magic exists but can't be used, fail rather than silently falling
nicm [Fri, 24 Apr 2015 16:28:00 +0000 (16:28 +0000)]
If ~/.magic exists but can't be used, fail rather than silently falling
back to /etc/magic.

9 years agoFix a couple of errors in the magic source files.
nicm [Fri, 24 Apr 2015 16:24:42 +0000 (16:24 +0000)]
Fix a couple of errors in the magic source files.

9 years agoNew implementation of the file(1) utility. This is a simplified,
nicm [Fri, 24 Apr 2015 16:24:11 +0000 (16:24 +0000)]
New implementation of the file(1) utility. This is a simplified,
modernised version with a nearly complete magic(5) parser but omits some
of the complex builtin tests (notably ELF) and has a reduced set of
options.

ok deraadt

9 years agoRevert back to using GCC builtins. This code triggers an off by one in
mpi [Fri, 24 Apr 2015 15:26:22 +0000 (15:26 +0000)]
Revert back to using GCC builtins.  This code triggers an off by one in
device_unref() as found by deraadt@.

9 years agoanother thinkpad event to ignore; from mp39590@gmail
deraadt [Fri, 24 Apr 2015 14:44:17 +0000 (14:44 +0000)]
another thinkpad event to ignore; from mp39590@gmail

9 years agovio: If enqueue fails, drop packet.
sf [Fri, 24 Apr 2015 12:53:35 +0000 (12:53 +0000)]
vio: If enqueue fails, drop packet.

From brad@

9 years agoEnable the NX bit and use it in the PAE pmap code. PAE is still disabled
kettenis [Fri, 24 Apr 2015 12:52:38 +0000 (12:52 +0000)]
Enable the NX bit and use it in the PAE pmap code.  PAE is still disabled
while we're chasing at least one remaining bug.

ok mlarkin@, deraadt@