schwarze [Wed, 17 Aug 2016 21:23:01 +0000 (21:23 +0000)]
Add HISTORY section.
Information found on: http://www.tuhs.org/cgi-bin/utree.pl
Checked by Sevan Janiyan <venture37 at geeklan dot co dot uk>.
schwarze [Wed, 17 Aug 2016 20:46:06 +0000 (20:46 +0000)]
When the content of a manual page does not specify a section, the
empty string got added to the list of sections, breaking the database
format slightly and causing the page to not be considered part of
any section, not even if a section could be deduced from the directory
or from the file name.
Bug found due to the bogus pcredemo(3) "manual" in the pcre-8.38p0 package.
schwarze [Wed, 17 Aug 2016 18:58:31 +0000 (18:58 +0000)]
When reading back a mandoc.db(5) file in order to apply incremental
changes, do not prepend a stray NAME_FILE (0x10) byte to the first
names of pages.
Bug found while investigating another issue reported by sthen@.
deraadt [Wed, 17 Aug 2016 18:07:07 +0000 (18:07 +0000)]
Generate syslog warnings for %s fmt strings NULL to "(null)" conversions.
Over time we can repair software which performs this non-standard behaviour,
and fix bugs along the way. Let's first find out how bad the situation is
by deploying this in snapshots.
This type of logging is possible because OpenBSD syslog_r(3) -> sendsyslog(2)
is side-effect free enough to be used in the bowels of libc.
ok tedu
schwarze [Wed, 17 Aug 2016 18:05:40 +0000 (18:05 +0000)]
Make sure manuals in architecture-independent directories are treated
as architecture-independent even if they abuse the third (architecture)
argument of the .Dt macro for random stuff like "freetds reference manual".
While the .Dt syntax is not the same as the .TH syntax in man(7),
punishing offenders by treating them as architecture-dependent and
hence completely excluding them from searches is too severe.
Problem reported by sthen@.
mikeb [Wed, 17 Aug 2016 17:39:41 +0000 (17:39 +0000)]
Sync up monitor trigger groups handling to the upstream
Replace hand rolled atomic bit operations and use MI ones from DRM
and convert event matrixes to arrays of longs.
mikeb [Wed, 17 Aug 2016 17:18:38 +0000 (17:18 +0000)]
Replace hand rolled atomic bit operations and use MI ones from DRM
deraadt [Wed, 17 Aug 2016 16:13:26 +0000 (16:13 +0000)]
sync
krw [Wed, 17 Aug 2016 13:53:20 +0000 (13:53 +0000)]
Use more dhcp info when trying autoinstall/autoupgrade.
For the file name try 'filename' then 'option bootfile-name'.
For the server name try 'server-name' then 'option tftp-server-name' then
'next-server'.
Original idea from Patrick Lundin. Suggestions from rpe@.
Tweaks & ok halex@
bluhm [Wed, 17 Aug 2016 13:53:14 +0000 (13:53 +0000)]
Fix an mbuf leak and missing error propagation in uipc_usrreq(PRU_SEND)
in case sbappendcontrol() fails.
From Simon Mages; OK mikeb@
patrick [Wed, 17 Aug 2016 13:44:48 +0000 (13:44 +0000)]
Nothing from the original Wasabi code remains, so replace the copyright
with the standard ISC license.
ok kettenis@ jsg@
patrick [Wed, 17 Aug 2016 13:26:40 +0000 (13:26 +0000)]
Remove bogus suspend/resume code from armv7 com(4) attachment driver.
The code initially arrived when the zaurus driver was copied. As we
don't suspend on ARMv7, just get rid of it.
Prompted by jsg@ and kettenis@.
patrick [Wed, 17 Aug 2016 13:05:02 +0000 (13:05 +0000)]
Replace sxiuart(4) with the mostly compatible com(4).
The Synopsys DesignWare 8250 throws an interrupt if it detects a write
to LCR while it was busy. Read the USR register to clear the interrupt
so we don't get flooded.
The sunxi machines also use a different frequency on the consoles. As
we have no clock framework, especially not for early consoles, match on
known Allwinner compatibles to set a known frequency.
ok kettenis@ mglocker@
bluhm [Wed, 17 Aug 2016 12:22:27 +0000 (12:22 +0000)]
syslogd(8) uses new libtls certificate and key file loading. So
the tests behave a little bit different now.
bluhm [Wed, 17 Aug 2016 12:18:29 +0000 (12:18 +0000)]
By using the new tls_config_load_file() feature, syslogd(8) certificate
and key loading can be simplified.
OK jsing@
millert [Wed, 17 Aug 2016 12:02:38 +0000 (12:02 +0000)]
When returning the file size info, only use z_total_in/z_total_out
for the read (inflate) case where there might be multiple streams.
For the write (deflate) case there can only be a single stream so
just use the per-stream total_in/total_out. Fixes "gzip -v" byte
counts. OK henning@.
kettenis [Wed, 17 Aug 2016 11:56:42 +0000 (11:56 +0000)]
Fix x86_atomic_{set|clear}bits_u64() by using the "er" constraint instead
of "ir" as the orq and andq instructions take a 32-bit immedate argument that
gets sign-extended.
ok mikeb@
stsp [Wed, 17 Aug 2016 11:52:29 +0000 (11:52 +0000)]
Document RT3900E (RT5390 / RT5392) support in ral(4).
Ported from FreeBSD by James Hastings.
stsp [Wed, 17 Aug 2016 11:50:52 +0000 (11:50 +0000)]
Add support for RT5390 and RT5392 chipsets to the ral(4) driver.
Tested by procter@ on RT2860 and by me on RT2790 and RT3090.
Ported from FreeBSD by James Hastings.
stsp [Wed, 17 Aug 2016 11:47:57 +0000 (11:47 +0000)]
Update ral(4) firmware for RT2860 family devices.
Required for upcoming RT5390 and RT5392 chipset support.
Tested by procter@ on RT2860 and by me on RT2790 and RT3090.
Ported from FreeBSD by James Hastings.
kevlo confirmed this version of the firmware is still ISC licensed.
dlg [Wed, 17 Aug 2016 11:09:01 +0000 (11:09 +0000)]
use a register variable to get at the curcpu pointer in %g7
this lets us define curcpu(), cpu_number(), etc as derefs of this
variable rather than having to write asm to copy it into another
register for the same derefs.
this provides shorter inlined code that wastes less registers.
ok kettenis@ who notes we use this these on other archs already
stsp [Wed, 17 Aug 2016 11:08:08 +0000 (11:08 +0000)]
wpi(4) scans all bands at once, as far as the net80211 stack is concerned.
So set the appropriate scan capability flags in the ic.
ok sthen@
dlg [Wed, 17 Aug 2016 10:49:09 +0000 (10:49 +0000)]
we have some spare interrupt levels, so move softnet above softclock.
ok kettenis@
stsp [Wed, 17 Aug 2016 10:01:34 +0000 (10:01 +0000)]
regen
stsp [Wed, 17 Aug 2016 10:01:20 +0000 (10:01 +0000)]
Add PCI ID for Ralink RT5392 and additional IDs for RT5390 devices.
Part of a larger patch by James Hastings.
stsp [Wed, 17 Aug 2016 09:43:27 +0000 (09:43 +0000)]
iwn(4) scans all bands at once, as far as the net80211 stack is concerned.
So set the appropriate scan capability flags in the ic.
ok sthen@
stsp [Wed, 17 Aug 2016 09:42:03 +0000 (09:42 +0000)]
If a wireless device or driver scans all bands at once give 5GHz APs
a slight priority in cases where good matches exist in either band.
ok sthen@
stsp [Wed, 17 Aug 2016 09:39:38 +0000 (09:39 +0000)]
If the stack demands protection by setting the USEPROT flag then set the
corresponding bit in the iwm(4) Tx command regardless of frame length.
ok sthen@
phessler [Wed, 17 Aug 2016 08:14:40 +0000 (08:14 +0000)]
pluralize a word
phessler [Wed, 17 Aug 2016 08:13:52 +0000 (08:13 +0000)]
The man page is a bit confusing for how we can control which address
families we support during capabilities negotiation. Re-organize the
man page to be more clear, and explicitly state what the default is.
OK sthen@ benno@
otto [Wed, 17 Aug 2016 05:33:54 +0000 (05:33 +0000)]
wrterror() is fatal, delete dead code; ok tom@ natano@ tedu@
deraadt [Wed, 17 Aug 2016 05:07:13 +0000 (05:07 +0000)]
small bits of header cleanup; ok mlarkin
procter [Wed, 17 Aug 2016 03:24:11 +0000 (03:24 +0000)]
Reintroduce 5.3-style checksum modification to preserve end-to-end checksums
when fiddling with packets but without the mess that motivated Henning to
remove it. Affects only this one aspect of Henning's checksum work. Also tweak
the basic algorithm and supply a correctness argument.
OK dlg@ deraadt@ sthen@; no objection henning@
krw [Wed, 17 Aug 2016 01:17:54 +0000 (01:17 +0000)]
Move to iopool. No voluntary testers after several years of requests so now
everybody gets to test!
krw [Wed, 17 Aug 2016 01:16:11 +0000 (01:16 +0000)]
Move to iopool. No voluntary testers after several years of requests so now
everybody gets to test!
krw [Wed, 17 Aug 2016 01:02:31 +0000 (01:02 +0000)]
Make error handling (esp. DATA_UNDERRUN) clearer and avoid losing resid
value.
Specs from mikeb@, ok dlg@
krw [Wed, 17 Aug 2016 00:55:33 +0000 (00:55 +0000)]
KNF and clean up many comments.
krw [Wed, 17 Aug 2016 00:38:26 +0000 (00:38 +0000)]
Ansify function definitions.
Diff from Edgar Pettijohn.
dlg [Tue, 16 Aug 2016 23:49:35 +0000 (23:49 +0000)]
splraise() is an MI API now.
ok deraadt@ jmc@ mpi@
vgross [Tue, 16 Aug 2016 22:25:08 +0000 (22:25 +0000)]
Add regression tests for IP_SENDSRCADDR.
vgross [Tue, 16 Aug 2016 22:21:17 +0000 (22:21 +0000)]
Add IP_SENDSRCADDR cmsg for UDP sockets. As suggested by sthen@,
IP_SENDSRCADDR == IP_RECVDSTADDR.
OK sthen@ jca@ bluhm@
krw [Tue, 16 Aug 2016 21:57:51 +0000 (21:57 +0000)]
Track SSID in leases file and only consider leases from the current SSID when
starting up dhclient on wifi interfaces. In theory will preserve leases from
other SSID's and speed up obtaining a lease by not wasting time attempting
to re-acquire a lease from a different SSID.
Experimental feature from n2k16 needing some real world testing before g2k16.
natano [Tue, 16 Aug 2016 21:32:58 +0000 (21:32 +0000)]
There is no sense in doing caching in fusefs. In case of a non-local
filesystem the tree can change behind our back, resulting in stale cache
entries. "The only winning move is not to play."
ok tedu beck mpi
natano [Tue, 16 Aug 2016 20:04:46 +0000 (20:04 +0000)]
Piping to a shell command does not count as a save, so don't reset the
modified flag.
from Jerome Frgacic (jerome.frgacic (at) yahoo.fr)
ok tb
tb [Tue, 16 Aug 2016 19:00:59 +0000 (19:00 +0000)]
When the commit message is empty, don't default to c)ontinue
(committing without log message) but a)bort. Also, accept
the uppercase versions of the choices.
Matches the corresponding change of behavior of GNU cvs.
suggested by & ok jcs
tb [Tue, 16 Aug 2016 18:57:10 +0000 (18:57 +0000)]
When the commit message is empty, don't default to c)ontinue
(committing without log message), but to a)bort.
ok jcs, natano; millert and tj agree
schwarze [Tue, 16 Aug 2016 18:51:25 +0000 (18:51 +0000)]
Add, correct, and improve HISTORY sections.
From Sevan Janiyan <venture37 at geeklan dot co dot uk>.
Verified using http://minnie.tuhs.org/.
tedu [Tue, 16 Aug 2016 18:41:57 +0000 (18:41 +0000)]
stop including sys/param.h for nitems. define locally as needed.
ok natano reyk
schwarze [Tue, 16 Aug 2016 18:41:06 +0000 (18:41 +0000)]
Add HISTORY.
From Sevan Janiyan <venture37 at geeklan dot co dot uk>.
Verified using the CSRG archive CD.
tedu [Tue, 16 Aug 2016 18:21:54 +0000 (18:21 +0000)]
remove #define for /dev/mem minors. let the raw magic of the hardcoded
numbers serve as a warning not to touch them.
tedu [Tue, 16 Aug 2016 18:19:15 +0000 (18:19 +0000)]
move static variable up to first
tedu [Tue, 16 Aug 2016 18:17:36 +0000 (18:17 +0000)]
replace hand rolled tsleep physlock with rwlock. ok mlarkin
tedu [Tue, 16 Aug 2016 17:52:28 +0000 (17:52 +0000)]
doesn't need to link with crypto
tedu [Tue, 16 Aug 2016 17:51:38 +0000 (17:51 +0000)]
remove reference to 6th ed.
reyk [Tue, 16 Aug 2016 17:10:01 +0000 (17:10 +0000)]
Turn "TLS handshake failed -" log message into a debug message - it
happens way too often and does not provide much information.
OK jung@
tb [Tue, 16 Aug 2016 16:49:24 +0000 (16:49 +0000)]
Streamline and refactor the code a bit more:
There are two sentinels for infinite output: infinity == true and reps == 0.
Ensure that infinity is set to true whenever reps is set to zero, then we
can always use 'if (infinity)'.
This allows us to merge the loop for deterministic output into the first
if (!randomize) statement, which is a lot tidier.
Insert an occasional empty line.
"move fast and break stuff" tedu
krw [Tue, 16 Aug 2016 16:44:55 +0000 (16:44 +0000)]
Nuke a bunch of whitespace nits seen while '%q' hunting.
krw [Tue, 16 Aug 2016 16:41:46 +0000 (16:41 +0000)]
Nuke a bunch of whitespace nits seen while '%q' hunting.
krw [Tue, 16 Aug 2016 16:13:32 +0000 (16:13 +0000)]
Nuke some erroneous leading whitespace.
krw [Tue, 16 Aug 2016 16:09:24 +0000 (16:09 +0000)]
Bring types of variables used with struct stat into the modern world.
Replace a couple of u_long paramaters with int as they were only passed int
values and the function re-cast them to (int) anyway. Weird.
ok tedu@
tb [Tue, 16 Aug 2016 15:32:07 +0000 (15:32 +0000)]
Avoid a compiler warning about use of uninitialized l.beg emitted
when ksh is compiled with -Os (done in distrib/special): move the
bit using XPptrv(l) (which expands to l.beg) where it is actually
used and clearly initialized.
ok tedu, tweak + ok millert
tedu [Tue, 16 Aug 2016 15:24:01 +0000 (15:24 +0000)]
All that complexity stopped us from getting flying cars by today.
mikeb [Tue, 16 Aug 2016 14:43:31 +0000 (14:43 +0000)]
Update Hyper-V structure definitions to the 2016 version
Microsoft has performed a significant clean up and stylistic improvement
of Hyper-V structure definitions and updated FreeBSD port. We update to
stay in sync with the upstream and improve maintainability of this code
in the future.
schwarze [Tue, 16 Aug 2016 13:11:06 +0000 (13:11 +0000)]
Add HISTORY.
From Sevan Janiyan <venture37 at geeklan dot co dot uk>.
Verified using NetBSD CVS.
visa [Tue, 16 Aug 2016 13:03:58 +0000 (13:03 +0000)]
Remove RM7000/RM9000-specific performance counter code. It originates
from PMON2000 and has not been enabled on OpenBSD.
Suggested by and ok miod@ (after seeing a quad_t cleanup patch of mine)
tb [Tue, 16 Aug 2016 11:33:07 +0000 (11:33 +0000)]
Xr encrypt(1) here as well to be consistent with useradd(8).
ok jmc
kettenis [Tue, 16 Aug 2016 10:16:33 +0000 (10:16 +0000)]
Fix typo/inconsistensy where L1_S_DOMAIN was used instead of L1_C_DOMAIN.
These are functionally equivolent so it didn't matter and the resulting
code doesn't change.
mpi [Tue, 16 Aug 2016 09:01:14 +0000 (09:01 +0000)]
Mask the ``prio'' prior to comparing it to RTP_LOCAL.
This allows rt_if_remove() to remove RTF_BROACAST routes from down
interfaces.
Issue reported by Dimitris Papastamos on bugs@
ok dlg@, claudio@, phessler@
mpi [Tue, 16 Aug 2016 08:56:28 +0000 (08:56 +0000)]
In the IPv6 default route example, use a gateway that corresponds to a
RTF_HOST route and is always on the correct interface.
The gateway value doesn't really matter for routes on p2p interfaces
but this allow us to tighten the checks when adding RTF_GATEWAY routes.
ok stsp@, sthen@
reyk [Tue, 16 Aug 2016 08:36:29 +0000 (08:36 +0000)]
Rename server_handshake_tls() to server_tls_handshake() to align with
the other server_tls_* functions (and I like the prefix notation
better). No functional change.
reyk [Tue, 16 Aug 2016 08:23:18 +0000 (08:23 +0000)]
As suggested by the FastCGI spec, zero-pad the response buffers to be
aligned to 8 bytes. This matches what most other implementations are
doing. While here, make sure that the allocated response buffers are
zero'ed out.
OK florian@
kettenis [Tue, 16 Aug 2016 08:15:04 +0000 (08:15 +0000)]
Only probe the UART type if sc_uarttype is set to COM_UART_UNKNOWN.
This allows glue drivers to set the type based on other information (such as
Oopen Firmware or device tree properties) without going through to risky
code paths.
ok deraadt@, visa@
jmc [Tue, 16 Aug 2016 06:40:24 +0000 (06:40 +0000)]
shorten genrsa; ok jsing
jmc [Tue, 16 Aug 2016 06:39:46 +0000 (06:39 +0000)]
tweak previous;
tedu [Tue, 16 Aug 2016 04:55:33 +0000 (04:55 +0000)]
add a special edition doas that only lets root drop privs.
maybe the installer can use something like this, wink wink.
ok deraadt
tedu [Tue, 16 Aug 2016 04:48:04 +0000 (04:48 +0000)]
fread hasn't taken a char * in about 900 years
tedu [Tue, 16 Aug 2016 04:44:38 +0000 (04:44 +0000)]
remove unneeded casts
guenther [Tue, 16 Aug 2016 04:14:34 +0000 (04:14 +0000)]
Rewrite: ALL HAIL ELF
Mention the KERN_NOSUIDCOREDUMP sysctl(3)
ok deraadt@
stsp [Mon, 15 Aug 2016 22:16:46 +0000 (22:16 +0000)]
When running 'ifconfig scan' in hostap mode display the current Tx rate our
AP is using to send frames to an associated node. This used to always display
the node's highest supported Rx rate, which isn't all that interesting.
ok mpi@
stsp [Mon, 15 Aug 2016 22:14:19 +0000 (22:14 +0000)]
Expose more 802.11n information to userspace:
A flag which indicates whether HT has been negotiated with a node,
and the current Tx MCS value we use for a node.
This grows struct ieee80211_nodereq. Applications using it must be recompiled.
ok mpi@
tedu [Mon, 15 Aug 2016 22:01:59 +0000 (22:01 +0000)]
normalize some comments
kettenis [Mon, 15 Aug 2016 21:08:56 +0000 (21:08 +0000)]
Don't take shortcuts cleaning/invalidating the caches. The Cortex-A7 and most
other ARMv7 CPUs have a L2 cache that is larger than 32KB. And some of those
even have L1 caches larger than that. So bailing out after 32KB will leave
the cache partially dirty.
Fixes the SATA problems on Allwinner A20 based boards.
ok patrick@
patrick [Mon, 15 Aug 2016 21:04:32 +0000 (21:04 +0000)]
Now that com(4) uses a different iot for the initial console and the
main attachment, the serial port is now longer recognized as console.
To fix this, store the OFW node of the initial console and check it
in the attachment driver.
ok kettenis@
kettenis [Mon, 15 Aug 2016 21:03:27 +0000 (21:03 +0000)]
First stab at porting the awinmmc(4) driver from NetBSD. The driver will
be known as sximmc(4) and is somewhat functional. Still needs proper clock
support and some further cleanup.
sthen [Mon, 15 Aug 2016 19:31:57 +0000 (19:31 +0000)]
sync
kettenis [Mon, 15 Aug 2016 18:31:28 +0000 (18:31 +0000)]
For some reason the RTL8211E PHY on the Banana Pi responds to both address 0
and 1. As a result rgephy(4) attaches twice. Prevent this from happening
by passing the PHY address, which is part of the device tree, down from the
bus-specific glue to the generic driver code.
mglocker [Mon, 15 Aug 2016 17:49:43 +0000 (17:49 +0000)]
Fix typo -> s/drirect/direct/.
From Daniel Bolgheroni
krw [Mon, 15 Aug 2016 16:57:53 +0000 (16:57 +0000)]
%*qd -> %*lld + (long long) for off_t.
Avoid some casts by changing printsize() to take an int (the only
type actually passed via that parameter) instead of size_t.
Tweaks & ok guenther@
jsing [Mon, 15 Aug 2016 16:12:34 +0000 (16:12 +0000)]
Move server_match() from parse.y to server.c; use env instead of conf,
which is actually the same thing (cluebat from reyk@).
jsing [Mon, 15 Aug 2016 15:44:58 +0000 (15:44 +0000)]
The tls_conninfo serial is also unused.
jsing [Mon, 15 Aug 2016 15:41:50 +0000 (15:41 +0000)]
Group conninfo fields by connection and peer cert based information,
sort and remove unused fingerprint.
jsing [Mon, 15 Aug 2016 14:47:41 +0000 (14:47 +0000)]
Fix some style(9) issues.
patrick [Mon, 15 Aug 2016 14:17:34 +0000 (14:17 +0000)]
Allow com(4) to make use of the simplebus address translation.
This typically works for every simplebus client, but com(4) did not
make use of the supplied bus tag. Instead it references the global
a4x bus tag. This is needed as the ARM com(4) controllers use 4-byte
spaced registers, while com(4) operates on 1-byte. To be able to make
use of the address translation, copy the a4x bus tag and replace the
cookie and map function with the one supplied by simplebus.
ok kettenis@
jsing [Mon, 15 Aug 2016 14:14:55 +0000 (14:14 +0000)]
Use lowercase 'tls' in debug and log messages for consistency.
Requested by reyk@
jsing [Mon, 15 Aug 2016 14:04:23 +0000 (14:04 +0000)]
Explicitly pass in an SSL_CTX * to the functions that operate on one,
instead of assuming that they should use the one associated with the TLS
context. This allows these functions to be used with the additional
SSL contexts that are needed to support server-side SNI.
Also rename tls_configure_keypair() to tls_configure_ssl_keypair(), so that
these functions have a common prefix.
ok reyk@
jsing [Mon, 15 Aug 2016 13:48:24 +0000 (13:48 +0000)]
Make httpd stricter with respect to TLS configuration - in particular, do
not allow TLS and non-TLS to be configured on the same port, do not allow
TLS options to be specified without a TLS listener and ensure that the TLS
options are the same when a server is specified on the same address/port.
Currently, these configurations are permitted but do not work as intended.
Also factor out and reuse the server matching code, which was previously
duplicated.
ok reyk@
patrick [Mon, 15 Aug 2016 13:42:49 +0000 (13:42 +0000)]
Move the com(4) fdt attachment driver to a more generic location and
rename it to fit the typical scheme of such an attachment driver.
ok kettenis@ jsg@
stsp [Mon, 15 Aug 2016 12:59:53 +0000 (12:59 +0000)]
Remove IWM_CMD_SYNC, which expands to zero while all the other
IWM_CMD_* values are bit flags. The same was done 2 years ago
in Linux git commit
a10229271946731959b2269370d0492d88cfab23.
naddy [Mon, 15 Aug 2016 12:32:04 +0000 (12:32 +0000)]
Catch up with the SSH1 code removal and delete all mention of
protocol 1 particularities, key files and formats, command line
options, and configuration keywords from the server documentation
and examples. ok jmc@
naddy [Mon, 15 Aug 2016 12:27:56 +0000 (12:27 +0000)]
Remove more SSH1 server code:
* Drop sshd's -k option.
* Retire configuration keywords that only apply to protocol 1, as well as
the "protocol" keyword.
* Remove some related vestiges of protocol 1 support.
ok markus@