jsg [Sat, 9 Jul 2016 13:30:14 +0000 (13:30 +0000)]
ehci no longer attaches to imx
kettenis [Sat, 9 Jul 2016 12:39:28 +0000 (12:39 +0000)]
Perform PHY-specific initialization based on the PHY ID instead of the
board ID for the AR8031/AR8035.
ok jsg@
kettenis [Sat, 9 Jul 2016 12:32:50 +0000 (12:32 +0000)]
Dynamically attach i.MX6 ehci(4) using the FDT.
ok jsg@
kettenis [Sat, 9 Jul 2016 12:31:05 +0000 (12:31 +0000)]
Add interfaces to look up a device tree node by phandle.
ok patrick@, jsg@, visa@
stefan [Sat, 9 Jul 2016 09:06:22 +0000 (09:06 +0000)]
Prepare vionet to be handled asynchronously to the VCPU thread
This splits the handling of received data into a separate function
that can later be called in parallel to the VCPU thread instead of
handling received packets on VCPU exits only.
It also makes virtq accesses in the rx path safe to run in parallel
to the VCPU thread: the last index into the 'avail' ring the driver
has notified to the host is kept track of. It also makes sure that
the host only writes back to the 'avail' ring instead of modifying
the whole receive virtq.
While there, describe what virtio_vq_info and virtio_io_cfg are used
for, as suggested by mlarkin@
ok mlarkin@
jsg [Sat, 9 Jul 2016 07:19:52 +0000 (07:19 +0000)]
Mention the EFI and DTB requirement. List U-Boot 2016.07 as 2016.05
needs a patch to work with efiboot on non imx platforms that is
included in ports/packages but might not be present otherwise.
While U-Boot 2016.07 final is not released yet the
"efi_loader: Don't allocate from memory holes" patch is included
in rc1/rc2/rc3 and should make the final release.
jsg [Sat, 9 Jul 2016 04:25:44 +0000 (04:25 +0000)]
Use mac address and phy id from the fdt. Store settings for a second
port as well, though we still only handle a single port for now.
ok kettenis@
schwarze [Fri, 8 Jul 2016 22:27:58 +0000 (22:27 +0000)]
ISO C99 7.19.2.5 doesn't like mixing putchar(3) and putwchar(3) on
the same stream, and actually, it fails spectacularly on glibc.
Portability issue pointed out by Svyatoslav Mishyn <juef at openmailbox
dot org> after testing on Void Linux.
schwarze [Fri, 8 Jul 2016 20:41:13 +0000 (20:41 +0000)]
POSIX requires that a process calling tcsetpgrp(3) from the background
gets a SIGTTOU signal. In that case, do not stop.
Portability issue found while testing on commercial Solaris 9/10/11.
Thanks to opencsw.org for providing me with a testing environment.
millert [Fri, 8 Jul 2016 19:32:26 +0000 (19:32 +0000)]
When making a copy of svc_pollfd, use the correct size.
Also pass the correct fd count to svc_getreq_poll().
OK jca@
kettenis [Fri, 8 Jul 2016 18:20:48 +0000 (18:20 +0000)]
Fix check for "name" property. Restores synthesised "name" proprties in
eeprom -p output that were lost in revision 1.13.
millert [Fri, 8 Jul 2016 15:23:44 +0000 (15:23 +0000)]
Pass errret pointer to setupterm() to prevent setupterm()
from calling exit() when given an unknown terminal type.
From Anton Lindqvist, who also upstreamed the fix.
millert [Fri, 8 Jul 2016 14:26:04 +0000 (14:26 +0000)]
The XSLoader issue has been assigned CVE-2016-6185
kettenis [Fri, 8 Jul 2016 08:30:53 +0000 (08:30 +0000)]
regen
kettenis [Fri, 8 Jul 2016 08:30:06 +0000 (08:30 +0000)]
The Atheros PHYs with model ID 7 are the AR8031/AR8033/AR8035 family.
Identify as AR8035 since that one has the lowest revision number.
ok mlarkin@, millert@
mlarkin [Fri, 8 Jul 2016 06:35:12 +0000 (06:35 +0000)]
Return 0 on read from PIT control port. Intel explicitly says this is not
supported, and it looks like other emulators/hypervisors do a variety of
different things here. Most return 0, but at least one might return random
garbage. Returning 0 seems safest here, but leave a warning in place for
the logs in case a guest VM does this.
djm [Fri, 8 Jul 2016 03:44:42 +0000 (03:44 +0000)]
Improve crypto ordering for Encrypt-then-MAC (EtM) mode MAC algorithms.
Previously we were computing the MAC, decrypting the packet and then
checking the MAC. This gave rise to the possibility of creating a
side-channel oracle in the decryption step, though no such oracle has
been identified.
This adds a mac_check() function that computes and checks the MAC in
one pass, and uses it to advance MAC checking for EtM algorithms to
before payload decryption.
Reported by Jean Paul Degabriele, Kenny Paterson, Torben Hansen and
Martin Albrecht. feedback and ok markus@
rzalamena [Thu, 7 Jul 2016 19:56:27 +0000 (19:56 +0000)]
Updated the ldpd.conf man page examples
The man page already contains the definition of the new neighbor-addr and
neighbor-id, but the examples were outdated. Now we may have an LSR-ID that
is different from its address.
ok renato@
millert [Thu, 7 Jul 2016 19:16:15 +0000 (19:16 +0000)]
Apply perl5.git.perl.org/perl.git/commitdiff/
08e3451d7
This fixes a bug where XSLoader could try to load from a subdir
of the cwd when called via eval. OK afresh1@
deraadt [Thu, 7 Jul 2016 16:11:47 +0000 (16:11 +0000)]
sync
schwarze [Thu, 7 Jul 2016 14:32:24 +0000 (14:32 +0000)]
sync with mdocml.bsd.lv: mention httpd(8) and slowcgi(8)
jsing [Thu, 7 Jul 2016 14:09:44 +0000 (14:09 +0000)]
Revert previous since the libtls change has been reverted.
jsing [Thu, 7 Jul 2016 14:09:03 +0000 (14:09 +0000)]
Revert previous - it introduces problems with a common privsep use case.
bcook [Thu, 7 Jul 2016 13:25:37 +0000 (13:25 +0000)]
add ca cert error check and make the path configurable
from Kinichiro Inoguchi
bcook [Thu, 7 Jul 2016 11:53:12 +0000 (11:53 +0000)]
call BN_init on temporaries to avoid use-before-set warnings
ok beck@
semarie [Thu, 7 Jul 2016 09:26:25 +0000 (09:26 +0000)]
biff, mesg, vi: only consider ACCESSPERMS for setting tty mode.
it explicitly removes any S_ISUID|S_ISGID|S_ISTXT bits, instead of letting
pledge(2) silenciously remove them.
ok beck@ deraadt@
semarie [Thu, 7 Jul 2016 09:24:09 +0000 (09:24 +0000)]
tmux: only consider ACCESSPERMS for setting mode on socket_path.
it explicitly removes any S_ISUID|S_ISGID|S_ISTXT bits, instead of letting
pledge(2) silenciously remove them.
ok nicm@ beck@ deraadt@
mglocker [Thu, 7 Jul 2016 08:08:02 +0000 (08:08 +0000)]
Add man page for the maxrtc(4) I2C driver.
mlarkin [Thu, 7 Jul 2016 00:58:31 +0000 (00:58 +0000)]
sanity check vm create and run args earlier
bluhm [Wed, 6 Jul 2016 21:30:21 +0000 (21:30 +0000)]
Now libtls is always reading cert.pem during tls_config_new().
Adapt ktrace count in syslogd test.
otto [Wed, 6 Jul 2016 20:32:02 +0000 (20:32 +0000)]
J/j is a three valued option, document and fix code to actuall support that
with a little help from jmc@ for the man page bits
ok jca@ and a reluctant tedu@
millert [Wed, 6 Jul 2016 19:29:13 +0000 (19:29 +0000)]
Allow space-deliminated fields in syslog.conf in addition to
traditional tabs-deliminated fields. This is consistent with what
FreeBSD, NetBSD and Linux do. Adapted from FreeBSD.
millert [Wed, 6 Jul 2016 19:26:35 +0000 (19:26 +0000)]
Return EINVAL for mknod/mknodat when dev is -1 (aka VNOVAL).
OK beck@ tedu@
jsing [Wed, 6 Jul 2016 16:47:18 +0000 (16:47 +0000)]
Check that the given ciphers string is syntactically valid and results in
at least one matching cipher suite.
ok doug@
jca [Wed, 6 Jul 2016 16:38:09 +0000 (16:38 +0000)]
route6d spring cleanup
Various tweaks and fixes:
- nuke util.h, not needed since pidfile(3) went away
- nuke the rrt_same member of struct riprt, "future use" since import
- mark rtdexit as __dead
- nuke progname handling
- fix pid handling: cache the pid *after* calling daemon(3)
- nuke setting rtm_pid. The kernel is responsible for setting this
in routing messages.
- nuke the useless myseq variable
ok florian@ benno@ millert@ deraadt@ renato@
jsing [Wed, 6 Jul 2016 16:31:18 +0000 (16:31 +0000)]
Remove manual file loading (now that libtls does this for us) and adjust
pledge to match. Also use tls_config_error() to provide friendlier error
messages.
jsing [Wed, 6 Jul 2016 16:16:36 +0000 (16:16 +0000)]
Always load CA, key and certificate files at the time the configuration
function is called. This simplifies code and results in a single memory
based code path being used to provide data to libssl. Errors that occur
when accessing the specified file are now detected and propagated
immediately. Since the file access now occurs when the configuration
function is called, we now play nicely with privsep/pledge.
ok beck@ bluhm@ doug@
tedu [Wed, 6 Jul 2016 15:53:01 +0000 (15:53 +0000)]
fix several places where calculating ticks could overflow.
it's not enough to assign to an unsigned type because if the arithmetic
overflows the compiler may decide to do anything. so change all the
long long casts to uint64_t so that we start with the right type.
reported by Tim Newsham of NCC.
ok deraadt
mpi [Wed, 6 Jul 2016 15:50:00 +0000 (15:50 +0000)]
Move Hop-by-Hop processing into its own function ip6_hbhchcheck().
This function will help splitting the IPv6 input path in two, in
order to run the first part without KERNEL_LOCK() held.
Tested by Hrvoje Popovski, ok bluhm@
visa [Wed, 6 Jul 2016 14:19:29 +0000 (14:19 +0000)]
sync
mlarkin [Wed, 6 Jul 2016 07:09:15 +0000 (07:09 +0000)]
clarify a comment about memory regions
guenther [Wed, 6 Jul 2016 04:35:12 +0000 (04:35 +0000)]
Use fstatat() to avoid path surgery.
bug catching and ok millert@
jsing [Wed, 6 Jul 2016 02:32:57 +0000 (02:32 +0000)]
Correctly handle an EOF that occurs prior to the TLS handshake completing.
Reported by Vasily Kolobkov, based on a diff from Marko Kreen.
ok beck@
schwarze [Tue, 5 Jul 2016 20:01:47 +0000 (20:01 +0000)]
Some new tests related to bin/cat.c rev. 1.25,
from Sevan Janiyan <venture37 at geeklan dot co dot uk>.
bru [Tue, 5 Jul 2016 19:33:14 +0000 (19:33 +0000)]
Improve the tracking functions in wsmouse.
ok mpi@
deraadt [Tue, 5 Jul 2016 18:45:39 +0000 (18:45 +0000)]
sync
millert [Tue, 5 Jul 2016 18:16:49 +0000 (18:16 +0000)]
Update to tzdata2016f from from ftp.iana.org.
tim [Tue, 5 Jul 2016 17:41:59 +0000 (17:41 +0000)]
Remove kern.random remnants; OK deraadt@
jca [Tue, 5 Jul 2016 16:41:40 +0000 (16:41 +0000)]
Move to svc_getreq_poll/svc_pollfd.
Stop using select to avoid the weird workarounds for fd_set size.
Also replace calloc with reallocarray. Prompted by a mail by Miod,
cluebat from guenther@.
ok millert@, prodding deraadt@
visa [Tue, 5 Jul 2016 13:41:46 +0000 (13:41 +0000)]
Remove debug code that slipped in.
visa [Tue, 5 Jul 2016 12:57:58 +0000 (12:57 +0000)]
Build eeprom(8) on octeon.
ok kettenis@ deraadt@ jasper@
visa [Tue, 5 Jul 2016 12:56:28 +0000 (12:56 +0000)]
regen
visa [Tue, 5 Jul 2016 12:55:32 +0000 (12:55 +0000)]
Add /dev/openprom.
ok kettenis@ deraadt@ jasper@
visa [Tue, 5 Jul 2016 12:53:40 +0000 (12:53 +0000)]
Add openprom(4) for octeon.
ok kettenis@ deraadt@ jasper@
tim [Tue, 5 Jul 2016 11:43:02 +0000 (11:43 +0000)]
Hook up the pcfrtc(4) manual; OK kettenis@
mpi [Tue, 5 Jul 2016 10:17:14 +0000 (10:17 +0000)]
Expand IN6_IFF_NOTREADY, ok bluhm@
sthen [Tue, 5 Jul 2016 09:45:02 +0000 (09:45 +0000)]
sync
mpi [Tue, 5 Jul 2016 09:17:10 +0000 (09:17 +0000)]
Do not use ``rt_addr'' in in{6,}_selectsrc() it doesn't work with magic
addresses set on p2p interfaces.
Found the hardway by naddy@
jsg [Tue, 5 Jul 2016 05:06:27 +0000 (05:06 +0000)]
Modify code added in rev 1.30 to use the correct variable instead of a
different uninitialised one.
ok martijn@
bcook [Tue, 5 Jul 2016 03:55:34 +0000 (03:55 +0000)]
remove extra assignment of s from 1.11, fix regression test
beck [Tue, 5 Jul 2016 03:24:38 +0000 (03:24 +0000)]
remove unneeded duplicate call - spotted by jsing@
bcook [Tue, 5 Jul 2016 02:54:35 +0000 (02:54 +0000)]
On systems where we do not have BN_ULLONG defined (most 64-bit systems),
BN_mod_word() can return incorrect results if the supplied modulus is
too big, so we need to fall back to BN_div_word.
Now that BN_mod_word may fail, handle errors properly update the man page.
Thanks to Brian Smith for pointing out these fixes from BoringSSL:
https://boringssl.googlesource.com/boringssl/+/
67cb49d045f04973ddba0f92fe8a8ad483c7da89
https://boringssl.googlesource.com/boringssl/+/
44bedc348d9491e63c7ed1438db100a4b8a830be
ok beck@
guenther [Tue, 5 Jul 2016 00:44:41 +0000 (00:44 +0000)]
Missed a reference to dl_prebind.h
problem noted by Andrew Ngo (andrew.ngo (at) gmail.com)
tedu [Tue, 5 Jul 2016 00:35:09 +0000 (00:35 +0000)]
remove some casts that aren't necessary.
beck [Tue, 5 Jul 2016 00:21:47 +0000 (00:21 +0000)]
Add several fixes from OpenSSL to make OCSP work with intermediate
certificates provided in the response. - makes our newly added
ocsp regress test pass too..
ok bcook@
beck [Tue, 5 Jul 2016 00:16:23 +0000 (00:16 +0000)]
make less awful.. test against cloudflare too
beck [Mon, 4 Jul 2016 23:43:30 +0000 (23:43 +0000)]
Add a nasty little ocsp regress test in the hope pedants will make it better.
mlarkin [Mon, 4 Jul 2016 23:03:52 +0000 (23:03 +0000)]
limit each viornd request to 64KB.
kettenis [Mon, 4 Jul 2016 21:51:40 +0000 (21:51 +0000)]
Add nep(4).
guenther [Mon, 4 Jul 2016 21:15:06 +0000 (21:15 +0000)]
Remove prebind support: binding to symbol table indices is too fragile
for our development process.
ok kettenis@ deraadt@
kettenis [Mon, 4 Jul 2016 20:56:50 +0000 (20:56 +0000)]
Nuke prebind support; it's unworkable and we're never going to finish it.
ok guenther@, deraadt@
natano [Mon, 4 Jul 2016 18:34:03 +0000 (18:34 +0000)]
Add missing vput() in error path to prevent a vnode getting stuck with a
stale reference and lock, while it shouldn't hold either.
"makes sense to me" beck@
guenther [Mon, 4 Jul 2016 18:01:44 +0000 (18:01 +0000)]
DEBUGLIBS has been broken since the gcc4 switch, so delete it. CFLAGS
contains -g by default anyway
problem noted by Edgar Pettijohn (edgar (at) pettijohn-web.com)
ok millert@ kettenis@ deraadt@
guenther [Mon, 4 Jul 2016 17:56:54 +0000 (17:56 +0000)]
gcc's -fvisibility=hidden isn't the behavior we wanted when cleaning up
symbol exports, so delete ${VISIBILITY_HIDDEN} as unused
ok kettenis@ deraadt@
tedu [Mon, 4 Jul 2016 16:12:52 +0000 (16:12 +0000)]
switch calculuated thrsleep timeout to unsigned to prevent overflow
into negative values, which later causes a panic.
reported by Tim Newsham at NCC.
ok guenther
mpi [Mon, 4 Jul 2016 09:30:18 +0000 (09:30 +0000)]
On Quad-G5 make hpb(4) attach first when iterating PCI buses.
This allows openpic(4) to properly map interrupt for the devices
instead of possibly dereferencing garbage.
Found the hardway by and ok jmatthew@
mpi [Mon, 4 Jul 2016 08:11:48 +0000 (08:11 +0000)]
Use the _SAFE_ version of SRPL_FOREACH() in rtable_walk_helper() to
prevent an off-by-one when removing entries from the mpath list.
Fix a regression introduced by the refactoring needed to serialize
rtable_walk() with create/delete.
ok jca@
guenther [Mon, 4 Jul 2016 04:33:35 +0000 (04:33 +0000)]
The GOT has been initally mapped RW for *years*; ld.so doesn't need to
mprotect it to RW when filling in the references from the PLT
in snaps for a week, ok deraadt@
guenther [Mon, 4 Jul 2016 04:05:29 +0000 (04:05 +0000)]
Drop support for the undocumented second argument (same as -N option)
ok deraadt@
guenther [Mon, 4 Jul 2016 03:24:48 +0000 (03:24 +0000)]
Use fstatat() instead of crafting a filename to use with stat()
ok millert@
millert [Mon, 4 Jul 2016 01:39:12 +0000 (01:39 +0000)]
The -I flag is documented but not implemented. This fixes that and
also honors the -I flag from ci/co when prompting like GNU RCS.
OK jca@
deraadt [Mon, 4 Jul 2016 00:40:17 +0000 (00:40 +0000)]
sync
mglocker [Sun, 3 Jul 2016 20:05:44 +0000 (20:05 +0000)]
Move videovar.h in to video.c since it isn't used anywhere else.
Suggested by mpi@ diff from Patrick Keshishian.
ok mpi
gilles [Sun, 3 Jul 2016 14:30:33 +0000 (14:30 +0000)]
add -r option to enqueuer as compat interface for mailx
diff by Richard <richard@aaazen.com>
semarie [Sun, 3 Jul 2016 04:36:08 +0000 (04:36 +0000)]
introduces new promise "chown" to allow changing owner/group with *chown(2) family
it splits PLEDGE_FATTR in two ("fattr" stills grant the 2 flags, so no functional changes):
- PLEDGE_CHOWN : to be able to call *chown(2) syscalls
- PLEDGE_FATTR : the rest
it introduces "chown" which grant:
- PLEDGE_CHOWN : be able to call *chown(2)
- PLEDGE_CHOWNUID : be able to modifying owner/group
ok deraadt@ tedu@
deraadt [Sun, 3 Jul 2016 03:19:02 +0000 (03:19 +0000)]
sync
afresh1 [Sun, 3 Jul 2016 01:07:47 +0000 (01:07 +0000)]
Update to perl 5.20.3
OK bluhm@
tedu [Sat, 2 Jul 2016 17:09:09 +0000 (17:09 +0000)]
check cache tree for collisions when inserting replies.
if two identical requests are sent out, the first will create a cache
entry. the second will not go into the cache tree, but will linger around,
causing a crash when we free it and try to remove from the tree. instead,
give up if insert fails.
diagnosis and initial patch from Duncan.
mglocker [Sat, 2 Jul 2016 16:28:50 +0000 (16:28 +0000)]
Get 'AOAShasta' soundchip working found on PowerMac9,1.
ok deraadt
eric [Sat, 2 Jul 2016 09:32:30 +0000 (09:32 +0000)]
remove misleading comment. it's not true anymore.
ok gilles@
eric [Sat, 2 Jul 2016 08:47:30 +0000 (08:47 +0000)]
datain counter is part of the transaction state
ok gilles@
eric [Sat, 2 Jul 2016 07:55:59 +0000 (07:55 +0000)]
set the msgid on the transaction
ok gilles@ jung@
jmc [Sat, 2 Jul 2016 05:58:00 +0000 (05:58 +0000)]
do not uppercase "hop limit";
renato [Fri, 1 Jul 2016 23:36:38 +0000 (23:36 +0000)]
More renaming and whitespace cleanup.
No binary change after "strip -s".
renato [Fri, 1 Jul 2016 23:33:46 +0000 (23:33 +0000)]
Be more compliant with RFC 4447.
When sending a label withdraw during the pseudowire Control Word
negotiation, append a "Wrong C-bit" status TLV after the FEC TLV (in
conformance to RFC 4447 section 6.2). Apparently this has no use other
than aiding in troubleshooting.
Also, extend the recv_labelmessage() function to accept Status TLVs and
ignore them instead of shutting down the session.
renato [Fri, 1 Jul 2016 23:29:55 +0000 (23:29 +0000)]
Several minor tweaks.
renato [Fri, 1 Jul 2016 23:22:42 +0000 (23:22 +0000)]
Decrease the initialization FSM timeout.
The previous value of 180 was just too long. If a neighbor get stuck in
the initialization FSM for more than 15 seconds, then there's certainly
something wrong and the session should be dropped.
A potential case of a neighbor getting stuck in the initialization
FSM is when both the local and the remote LSRs disable the LDPv4 GTSM
negotiation and there's a mismatch in their GTSM configuration (one is
enabled for GTSM while the other is not).
In this case, a smaller timeout allows for a quicker recovery of the
session when the configuration is fixed on either side.
renato [Fri, 1 Jul 2016 23:18:24 +0000 (23:18 +0000)]
Rename hello flags to keep consistent with the rest of the code.
Flag constants should start with F_.
renato [Fri, 1 Jul 2016 23:14:31 +0000 (23:14 +0000)]
Add GTSM support (RFC 6720).
This also finishes the missing bits from our RFC 7552 implementation
because GTSM is mandatory for LDPv6.
To avoid any kind of interoperability problems, I included a few
knobs to enable/disable GTSM on a per-address-family and per-neighbor
basis. Cisco's LDPv6 implementation, for instance, doesn't support GTSM.
"reads good" claudio@
schwarze [Fri, 1 Jul 2016 22:40:44 +0000 (22:40 +0000)]
For -be, indent the $ on blank lines.
Patch from Giles Lean (NetBSD PR bin/4841), tweaked by kleink at
NetBSD (rev. 1.17 1998-01-27), version for OpenBSD sent in by Sevan
Janiyan <venture37 at geeklan dot co dot uk>.
OK deraadt@