openbsd
6 years agoDo not collect top-bit-set characters in case they need to be replaced.
nicm [Fri, 12 Jan 2018 16:43:47 +0000 (16:43 +0000)]
Do not collect top-bit-set characters in case they need to be replaced.

6 years agoSimplify character replacement on non-UTF-8 terminals and make a common
nicm [Fri, 12 Jan 2018 16:41:00 +0000 (16:41 +0000)]
Simplify character replacement on non-UTF-8 terminals and make a common
function.

6 years agoSimplify UTF-8 states down into one state.
nicm [Fri, 12 Jan 2018 16:32:12 +0000 (16:32 +0000)]
Simplify UTF-8 states down into one state.

6 years agoUse correct size in allocarray() call.
kettenis [Fri, 12 Jan 2018 14:53:37 +0000 (14:53 +0000)]
Use correct size in allocarray() call.

ok patrick@

6 years agoRuntime services may (and do) use device mappings on some UEFI implementations.
kettenis [Fri, 12 Jan 2018 14:52:55 +0000 (14:52 +0000)]
Runtime services may (and do) use device mappings on some UEFI implementations.
Skip these mappings during the remap-pahse as they are likely to be in
a different 512G bloch as memory and SetVirtualAddressMap() shouldn't need
them.  But do assign a new virtual address and let efi(4) create a mapping.
Add a PMAP_DEVICE flag such that pmap_enter() can continue to be used
to create these mappings.

ok patrick@

6 years agoSome more tweaks to NOTES
jca [Fri, 12 Jan 2018 14:20:57 +0000 (14:20 +0000)]
Some more tweaks to NOTES

- it's safe to assume that there's no plan to write "an `agent' to
  execute unreadable/setuid/setgid shell scripts"
- the BUG-REPORTS file has been removed
- update the entry for FPATH, ksh93 also documents the described behavior
- kill entry about octal and hex notation in arithmetic expressions
  (supported)
- typos

Feedback from Klemens Nanni

6 years agoFix tyops
jca [Fri, 12 Jan 2018 14:04:49 +0000 (14:04 +0000)]
Fix tyops

6 years agoAdd tests for [[:foo:]] character classes in globs
jca [Fri, 12 Jan 2018 13:54:10 +0000 (13:54 +0000)]
Add tests for [[:foo:]] character classes in globs

6 years agoAdd basic tests for octal and hex notation in arithmetic expansions
jca [Fri, 12 Jan 2018 11:13:29 +0000 (11:13 +0000)]
Add basic tests for octal and hex notation in arithmetic expansions

POSIX requires only decimal, octal and hex, tests for the $((x#number))
notation could be useful too.

6 years agoConstrain alloc_pages() to the dma_constraint range. Avoids a
jsg [Fri, 12 Jan 2018 11:03:15 +0000 (11:03 +0000)]
Constrain alloc_pages() to the dma_constraint range.  Avoids a
"Non dma-reachable buffer" panic when trying to use it from ttm code on
a machine with 8GB of physmem and a radeon.

ok kettenis@ dlg@ visa@ deraadt@

6 years agoCorrectly iterate over all extended signatures.
patrick [Fri, 12 Jan 2018 10:33:07 +0000 (10:33 +0000)]
Correctly iterate over all extended signatures.

Found by markus@
With sf@

6 years agoImprove error message if creating socket parent directory fails, from
nicm [Fri, 12 Jan 2018 10:22:02 +0000 (10:22 +0000)]
Improve error message if creating socket parent directory fails, from
Thomas Adam for GitHub issue 1215.

6 years agoImprove logging for layout cells.
nicm [Fri, 12 Jan 2018 10:16:03 +0000 (10:16 +0000)]
Improve logging for layout cells.

6 years agoUnify <machine/mutex.h> a bit further.
mpi [Fri, 12 Jan 2018 09:19:32 +0000 (09:19 +0000)]
Unify <machine/mutex.h> a bit further.

`mtx_owner' becomes the first field of 'struct mutex' on i386/amd64/arm64.

ok visa@

6 years agoIBRS -> IBRS,IBPB in identifycpu lines
mlarkin [Fri, 12 Jan 2018 08:21:30 +0000 (08:21 +0000)]
IBRS -> IBRS,IBPB in identifycpu lines

6 years agorevert previous changes to enable dwiic on Dell precision as it seems
mlarkin [Fri, 12 Jan 2018 08:11:47 +0000 (08:11 +0000)]
revert previous changes to enable dwiic on Dell precision as it seems
to cause issues on Dell XPS 15 9560. Will investigate and recommit later
if a better fix can be found.

6 years agoCheck that mpls has been enabled on the input interface, lost in rev 1.66
jca [Fri, 12 Jan 2018 06:57:56 +0000 (06:57 +0000)]
Check that mpls has been enabled on the input interface, lost in rev 1.66

While here fix under MPLS_DEBUG.  ok dlg@

6 years agoAlign cases in conversion switch vertically. Normalize logic inside cases
tb [Fri, 12 Jan 2018 06:22:31 +0000 (06:22 +0000)]
Align cases in conversion switch vertically. Normalize logic inside cases
and drop an unnecessary test for intdata in '%c'.

reads ok to halex, ok millert

6 years agoAdjust references for sysctl(3) to sysctl(2)
deraadt [Fri, 12 Jan 2018 04:36:44 +0000 (04:36 +0000)]
Adjust references for sysctl(3) to sysctl(2)

6 years agosysctl(3) can now be renamed to sysctl(2)
deraadt [Fri, 12 Jan 2018 04:36:12 +0000 (04:36 +0000)]
sysctl(3) can now be renamed to sysctl(2)

6 years agomove gif_encap from gif_output to gif_start.
dlg [Fri, 12 Jan 2018 02:25:27 +0000 (02:25 +0000)]
move gif_encap from gif_output to gif_start.

the main win is the simplification of bpf in the ouput path.

ok visa@

6 years agodocument bpfattach and bpfdetach.
dlg [Fri, 12 Jan 2018 02:19:47 +0000 (02:19 +0000)]
document bpfattach and bpfdetach.

6 years agorestrict carp to configuring ethernet interfaces as carpdevs.
dlg [Fri, 12 Jan 2018 00:36:13 +0000 (00:36 +0000)]
restrict carp to configuring ethernet interfaces as carpdevs.

previously the driver only cared that a carp interface wasnt configured
as a carpdev. because the code only really works on ethernet, it makes
sense to restrict it.

ok visa@ mpi@

6 years agoImprove the handling of coordinate inputs and filters.
bru [Thu, 11 Jan 2018 23:50:49 +0000 (23:50 +0000)]
Improve the handling of coordinate inputs and filters.

Single-touch and multi-touch coordinate inputs are treated more uniformly,
and the hysteresis filters have a more consistent implementation.  If
possible, pointer control will be assigned to touches with coordinate
updates that pass the default hysteresis filter (the function has been
moved to wsmouse.c).  The "strong" variant of hysteresis has been
improved, the new version won't double the threshold when a movement
changes the orientation on an axis.

There is an additional change in wstpad_configure, which ensures that a
zero size disables an edge area even if the coordinate limits are
misconfigured.

6 years agocast the result of the ternary operators in __swapXX to the right type
dlg [Thu, 11 Jan 2018 23:13:37 +0000 (23:13 +0000)]
cast the result of the ternary operators in __swapXX to the right type

ok kettenis@
let's try it deraadt@

6 years agoComments
rpe [Thu, 11 Jan 2018 22:44:44 +0000 (22:44 +0000)]
Comments

6 years agoUpdate the Intel microcode once the root filesystem has been mounted.
patrick [Thu, 11 Jan 2018 22:31:09 +0000 (22:31 +0000)]
Update the Intel microcode once the root filesystem has been mounted.
This depends on the intel-firmware package that contains newer Intel
microcode which will be installed automatically by fw_update(1).

The update should happen much earlier since updating the microcode can
add or remove not only feature flags but also whole features.  For now
only update feature flags that are relevant to Spectre.

Initial diff from sf@
Tested by bluhm@
ok deraadt@

6 years agoInstall the intel firmware package which contains microcode updates
patrick [Thu, 11 Jan 2018 22:07:23 +0000 (22:07 +0000)]
Install the intel firmware package which contains microcode updates
if we match Intel CPUs in dmesg(8).

"do it do it" deraadt@

6 years agoCreate a second hash that maps all possible drivers to a default regex
patrick [Thu, 11 Jan 2018 22:04:39 +0000 (22:04 +0000)]
Create a second hash that maps all possible drivers to a default regex
to match in dmesg(8).  This is helpful if we want to install firmware
depending on dmesg(8) output that does not follow the typical driver
scheme.

From espie@
"do it do it" deraadt@

6 years ago- change [] tests to [[]]
rpe [Thu, 11 Jan 2018 21:09:26 +0000 (21:09 +0000)]
- change [] tests to [[]]
- change -a, -o to &&, || inside [[]]
- remove unecessary quoting inside [[]]
- remove X"" constructs inside [[]]
- remove \ (line continuation) in case of &&, || and pipes
- replace backticks with $()

discussed with and OK aja@
OK tb

6 years agoChange the shebang line from /bin/sh to /bin/ksh in all base rc.d
rpe [Thu, 11 Jan 2018 19:52:12 +0000 (19:52 +0000)]
Change the shebang line from /bin/sh to /bin/ksh in all base rc.d
daemon scripts.

discussed with and OK aja@
OK tb

6 years agoExtract the control channel number from the chanspec information and
patrick [Thu, 11 Jan 2018 19:33:34 +0000 (19:33 +0000)]
Extract the control channel number from the chanspec information and
apply the channel to the node, like iwm(4) does, when supplying the
scan response to the net80211 stack.  Our stack uses this information
for node selection so it's elementary for it to be correct.

Tested by jcs@

6 years agofix parsing of long TXT records, this prevents 'smtpctl spf walk' from
gilles [Thu, 11 Jan 2018 19:20:41 +0000 (19:20 +0000)]
fix parsing of long TXT records, this prevents 'smtpctl spf walk' from
producing bogus entries with some SPF records.

ok ajacoutot@

6 years agoPostpone secondary CPUs until after mounthook activities. This is
patrick [Thu, 11 Jan 2018 18:58:17 +0000 (18:58 +0000)]
Postpone secondary CPUs until after mounthook activities.  This is
useful for loading CPU microcode from the disk before the CPUs are
let go.

Tested by visa@ on sgi, loongson and octeon
"don't see immediate issues" kettenis@
ok deraadt@

6 years agoThe PCI bwfm(4) chips have no TX rings in the traditional sense, as on
patrick [Thu, 11 Jan 2018 16:09:19 +0000 (16:09 +0000)]
The PCI bwfm(4) chips have no TX rings in the traditional sense, as on
the actual rings we only share messages.  Sending a TX packet means
putting a message on the ring which contains a pktid (which for us maps
to an mbuf) and the physical address of the mbuf.  On jcs@'s macbook he
seems to run out of TX pktids pretty quickly during a speedtest.  This
would mean that there are 2048 TX packets in flight that we either want
to send out or that have not been "acked" by the firmware yet.  Either
way, recover from that situation when we hit that arbitrary limit by
restarting the queue after we free'd a packet from the TX pktid list.

Tested by jcs@

6 years agoAdvance data pointer after exporting SA type to userland. Otherwise
bluhm [Thu, 11 Jan 2018 16:02:31 +0000 (16:02 +0000)]
Advance data pointer after exporting SA type to userland.  Otherwise
ipsecctl(8) may display bundled SA incorrectly.
from markus@

6 years agoAdd size for free.
visa [Thu, 11 Jan 2018 15:49:34 +0000 (15:49 +0000)]
Add size for free.

OK mpi@

6 years agoRemove arbitary length limits for -s, -b and -w. Based on a diff from kshe.
tb [Thu, 11 Jan 2018 14:53:42 +0000 (14:53 +0000)]
Remove arbitary length limits for -s, -b and -w. Based on a diff from kshe.
schwarze suggested using asprintf, which led to a substantial cleanup.

ok schwarze, help & ok millert

6 years agosync
kevlo [Thu, 11 Jan 2018 09:27:20 +0000 (09:27 +0000)]
sync

6 years agoAdd support for Huawei E3372.
kevlo [Thu, 11 Jan 2018 09:26:36 +0000 (09:26 +0000)]
Add support for Huawei E3372.

ok phessler@

6 years agoAdd tests for incoming route-to and outgoing reply-to rules. Disable
bluhm [Thu, 11 Jan 2018 03:23:16 +0000 (03:23 +0000)]
Add tests for incoming route-to and outgoing reply-to rules.  Disable
the cases that currently fail.

6 years agocarp_ourether gets passed the parent interface, not the carp interface.
dlg [Thu, 11 Jan 2018 00:14:15 +0000 (00:14 +0000)]
carp_ourether gets passed the parent interface, not the carp interface.

6 years agoget rid of struct carp_if by moving the srpl into struct ifnet if_carp.
dlg [Wed, 10 Jan 2018 23:50:39 +0000 (23:50 +0000)]
get rid of struct carp_if by moving the srpl into struct ifnet if_carp.

currently carp uses a struct carp_if to hold an srp list head, which
is accessed by both if_carp in struct ifnet, and via the if input
handlers list.

this gets rid of some indirection by making if_carp itself the list
head, rather than a pointer to the list head via a struct carp_if.
it also makes accessing the list consistent by only using if_carp
to get to it.

ok mpi@

6 years agoImplement FUCKWIT for arm64; unmap the kernel almost entirely while userland
kettenis [Wed, 10 Jan 2018 23:27:18 +0000 (23:27 +0000)]
Implement FUCKWIT for arm64; unmap the kernel almost entirely while userland
is running.  This provides protection against meltown on cores that are
vilnerable (just Cortex-A75 so far) but also seems to be an essential to
protect against spectre-like attacks against the kernel.

This implementation only exposes a single treampoline page that does not
contain any kernel virtual addresses and also hides the real virtual address
of the exception vectors, which helps on cores vulnerable to "variant 3a"
(Cortex-A57, Cortex-A72).  The implementation is inspired by the work done
by Will Deacon for Linux, but there are no knobs to turn it off.  The
overhead is fairly limited: around 3-4% slowdown on Cortex-A57.

ok patrick@, deraadt@

6 years agoMark sosplice task mp safe, do not grab kernel lock for tcp output.
bluhm [Wed, 10 Jan 2018 18:14:34 +0000 (18:14 +0000)]
Mark sosplice task mp safe, do not grab kernel lock for tcp output.
OK mpi@

6 years agoAttach bwfm to the Broadcom 4350 found in the 2017 MacBook.
jcs [Wed, 10 Jan 2018 17:54:26 +0000 (17:54 +0000)]
Attach bwfm to the Broadcom 4350 found in the 2017 MacBook.

Easily handles >150Mbps transfers through a 5Ghz AP.

ok patrick

(Committed via bwfm0, of course)

6 years agoregen
jcs [Wed, 10 Jan 2018 15:52:36 +0000 (15:52 +0000)]
regen

6 years agoadd broadcom bcm4350
jcs [Wed, 10 Jan 2018 15:52:23 +0000 (15:52 +0000)]
add broadcom bcm4350

6 years agoAttach nvme to additional Apple NVMe device, limit targets to 1
jcs [Wed, 10 Jan 2018 15:45:46 +0000 (15:45 +0000)]
Attach nvme to additional Apple NVMe device, limit targets to 1

ok dlg

6 years agoregen
jcs [Wed, 10 Jan 2018 15:44:56 +0000 (15:44 +0000)]
regen

6 years agoadd another Apple NVMe
jcs [Wed, 10 Jan 2018 15:44:41 +0000 (15:44 +0000)]
add another Apple NVMe

6 years agoDon't require "disk" or "kernel", also allow just "cdrom" instead, a VM can
sthen [Wed, 10 Jan 2018 14:59:59 +0000 (14:59 +0000)]
Don't require "disk" or "kernel", also allow just "cdrom" instead, a VM can
still be useful with only cdrom storage. ok ccardenas@

6 years agoadd bwfm; ok patrick
jmc [Wed, 10 Jan 2018 14:30:07 +0000 (14:30 +0000)]
add bwfm; ok patrick

6 years agoUse pf_send_icmp() consistently in pf_route(). It sets the routing
bluhm [Wed, 10 Jan 2018 13:57:17 +0000 (13:57 +0000)]
Use pf_send_icmp() consistently in pf_route().  It sets the routing
domain and other mbuf flags.  In pf_route6() the bad packet counter
and dup-to check were missing.
OK visa@

6 years agosimplify the input interface type check in carp_proto_input_if.
dlg [Wed, 10 Jan 2018 10:25:52 +0000 (10:25 +0000)]
simplify the input interface type check in carp_proto_input_if.

carp6_proto_input_if only handles packets "received" on real carp
interfaces, which the ethernet stack goes to a lot of trouble to
provide. since carp assumes ethernet, carp_proto_input_if can assume
the packets will come in right too.

ok mpi@

6 years agoSimplify bounds checking of client numbers of channels and rate. From
ratchov [Wed, 10 Jan 2018 09:05:48 +0000 (09:05 +0000)]
Simplify bounds checking of client numbers of channels and rate. From
Michael W. Bombardieri, thanks.

6 years agoRemove assignement of sc->round to itself that makes no sense. From
ratchov [Wed, 10 Jan 2018 09:03:26 +0000 (09:03 +0000)]
Remove assignement of sc->round to itself that makes no sense. From
Michael W. Bombardieri, thanks.

6 years agoMake azalia_free_dmamem() void, as its return value is not used. From
ratchov [Wed, 10 Jan 2018 09:00:40 +0000 (09:00 +0000)]
Make azalia_free_dmamem() void, as its return value is not used. From
Michael W. Bombardieri, thanks.

6 years agoAdd firmware names for the two revisions of the Broadcom 4350 as seen
patrick [Wed, 10 Jan 2018 02:15:22 +0000 (02:15 +0000)]
Add firmware names for the two revisions of the Broadcom 4350 as seen
on a MacBook 12-inch (2017).

Tested by and with jcs@

6 years agoDon't reset the internal memory core on chips other than the Broadcom
patrick [Wed, 10 Jan 2018 02:07:11 +0000 (02:07 +0000)]
Don't reset the internal memory core on chips other than the Broadcom
43602, as it's only necessary on that specific chip.

Found the hard way by jcs@ on a MacBook 12-inch (2017)

6 years agoUpdate manpage to reflect changes to bwfm(4) which allow Host AP mode
patrick [Wed, 10 Jan 2018 02:02:55 +0000 (02:02 +0000)]
Update manpage to reflect changes to bwfm(4) which allow Host AP mode
and the PCI attachment driver.

6 years agoMove line for readability.
patrick [Wed, 10 Jan 2018 01:43:01 +0000 (01:43 +0000)]
Move line for readability.

6 years agoshuffle how the protocol family input is done in ether_input.
dlg [Wed, 10 Jan 2018 00:14:38 +0000 (00:14 +0000)]
shuffle how the protocol family input is done in ether_input.

the main change is to defer chopping the ethernet header off until the
ether_type is looked at. the main advantage of this is pppoe doesn't have
to reattach the ether header anymore, which simplifies the code a lot.

ok mpi@ visa@ bluhm@

6 years agofix some poop that snuck into the last commit.
dlg [Wed, 10 Jan 2018 00:05:06 +0000 (00:05 +0000)]
fix some poop that snuck into the last commit.

noticed by bijanebrahimi at riseup dot net and confirmed by anton@

6 years agoExtract compile_mode command status with WEXITSTATUS.
cheloha [Tue, 9 Jan 2018 17:59:29 +0000 (17:59 +0000)]
Extract compile_mode command status with WEXITSTATUS.

Also tell the user if the parent shell spawned by popen(3)
was killed by a signal.

Discussed with jca@.

ok millert@ jca@ deraadt@

6 years agoStop grabing the KERNEL_LOCK() for running protocol input routines.
mpi [Tue, 9 Jan 2018 17:50:57 +0000 (17:50 +0000)]
Stop grabing the KERNEL_LOCK() for running protocol input routines.

The NET_LOCK() is already held in this thread and is now enough.

People interested in ARP/bridge(4)/switch(4)/pipex(4)/pppoe(4)
performances can now push the KERNEL_LOCK() without depending on
other subsystems/drivers.

Tested by Hrvoje Popovski.

ok bluhm@, visa@

6 years agoCreating a cloned interface could return ENOMEM due to temporary
bluhm [Tue, 9 Jan 2018 15:24:24 +0000 (15:24 +0000)]
Creating a cloned interface could return ENOMEM due to temporary
memory shortage.  As it is invoked from a system call, it should
not fail and wait instead.
OK visa@ mpi@

6 years agoChange `so_state' and `so_error' to unsigned int such that they can
mpi [Tue, 9 Jan 2018 15:14:23 +0000 (15:14 +0000)]
Change `so_state' and `so_error' to unsigned int such that they can
be atomically read from any context.

ok bluhm@, visa@

6 years agolock->ctx != NULL => lock->ctx
jsg [Tue, 9 Jan 2018 14:23:03 +0000 (14:23 +0000)]
lock->ctx != NULL => lock->ctx
requested by kettenis@

6 years agoWork around a problem with ww_mutexes in the drm modeset lock.
jsg [Tue, 9 Jan 2018 13:48:36 +0000 (13:48 +0000)]
Work around a problem with ww_mutexes in the drm modeset lock.
Screen blanks often result in a NULL dereference in __ww_mutex_lock with
lock->acquired being non zero and lock->ctx NULL.  mpi@ also reported
it occuring when switching from Xorg to a virtual terminal.

ok mpi@

6 years agoDo not truncate 64bit integers when pretty-printing types.
mpi [Tue, 9 Jan 2018 10:19:25 +0000 (10:19 +0000)]
Do not truncate 64bit integers when pretty-printing types.

6 years agoUse ip{,6}_send() instead of ip{,6}_output() to prevent a recursion.
mpi [Tue, 9 Jan 2018 10:08:01 +0000 (10:08 +0000)]
Use ip{,6}_send() instead of ip{,6}_output() to prevent a recursion.

if_start() is a boundary between the network stack and drivers.  The
states it modifies must be protected by the driver, so we should not
require the NET_LOCK() there.

ok bluhm@, visa@

6 years agoStop printing <not displayed> for wireless keys we know the kernel no
mpi [Tue, 9 Jan 2018 10:02:02 +0000 (10:02 +0000)]
Stop printing <not displayed> for wireless keys we know the kernel no
longer export them to userland.

ok stsp@, deraadt@, jca@

6 years agoDo not silently truncate the firmware channel number.
mpi [Tue, 9 Jan 2018 10:00:12 +0000 (10:00 +0000)]
Do not silently truncate the firmware channel number.

Fix build with gcc, ok stsp@

6 years agomake mpls_input take a struct ifnet *ifp argument.
dlg [Tue, 9 Jan 2018 06:24:14 +0000 (06:24 +0000)]
make mpls_input take a struct ifnet *ifp argument.

this makes it like all our other protocol family input functions.

mpls_input always looks up the interface the mbuf was received on,
but it's always called by code that already has a reference to that
interface anyway. the result of this is a few less if_get/if_put
calls.

ok mpi@ bluhm@ visa@ claudio@

6 years agosync
deraadt [Tue, 9 Jan 2018 04:05:49 +0000 (04:05 +0000)]
sync

6 years agoRename test file =.err to eq.err for simpler make file scripts.
bluhm [Tue, 9 Jan 2018 00:43:10 +0000 (00:43 +0000)]
Rename test file =.err to eq.err for simpler make file scripts.

6 years agoFix indentation.
patrick [Mon, 8 Jan 2018 23:30:36 +0000 (23:30 +0000)]
Fix indentation.

6 years agoInitial support for HOSTAP mode. With this bwfm(4) can spawn an access
patrick [Mon, 8 Jan 2018 23:30:11 +0000 (23:30 +0000)]
Initial support for HOSTAP mode.  With this bwfm(4) can spawn an access
point including WPA2 support.  We now have a different event mask per
mode, so that events that are only useful for STA mode don't interfere
with HOSTAP mode.  Power savings is disabled when we act as AP. The
connection events generate 802.11 frames for handling auth/assoc and
deauth/deassoc so that our stack takes note of the connecting nodes.

6 years agoConvert IF_CLONE_INITIALIZER() into C99 initializer.
bluhm [Mon, 8 Jan 2018 23:05:21 +0000 (23:05 +0000)]
Convert IF_CLONE_INITIALIZER() into C99 initializer.
OK mpi@

6 years agoDo not pass a userland pointer to ktrabstimespec().
mpi [Mon, 8 Jan 2018 22:33:13 +0000 (22:33 +0000)]
Do not pass a userland pointer to ktrabstimespec().

Prevents an infinite pagefault/pmap_enter() loop when ktracing apps
doing a lot of futex(2) calls like firefox & chrome.

6 years agoupdate comments that were left behind in the recent unifdef JOB commit
benno [Mon, 8 Jan 2018 22:22:28 +0000 (22:22 +0000)]
update comments that were left behind in the recent unifdef JOB commit
from Klemens Nanni, thanks.

6 years agoFix typo in comment; ok mpi@
anton [Mon, 8 Jan 2018 18:21:22 +0000 (18:21 +0000)]
Fix typo in comment; ok mpi@

6 years agoIn AP mode multicast packets share the flowrings with broadcast
patrick [Mon, 8 Jan 2018 17:57:48 +0000 (17:57 +0000)]
In AP mode multicast packets share the flowrings with broadcast
packets.

6 years agoThe bwfm(4) TX ring expects the ethernet header as part of the TX info
patrick [Mon, 8 Jan 2018 17:53:07 +0000 (17:53 +0000)]
The bwfm(4) TX ring expects the ethernet header as part of the TX info
struct.  The data length is the length of the frame without the header.
In the previous version m_adj(9) is used, but since that was changed we
need to decrease the length ourselves.

6 years agoGuard the debug printf function behind BWFM_DEBUG as well. Also only
patrick [Mon, 8 Jan 2018 17:29:21 +0000 (17:29 +0000)]
Guard the debug printf function behind BWFM_DEBUG as well.  Also only
print the firmware's dmesg(8) if we're running with a higher debug
mode.

Prompted by Michael W. Bombardieri

6 years agoMove SC_* defines into <machine/setjmp.h>, and rename them,
visa [Mon, 8 Jan 2018 16:44:32 +0000 (16:44 +0000)]
Move SC_* defines into <machine/setjmp.h>, and rename them,
so that <machine/signal.h> is not needed in setjmp.S.

Suggested by kettenis@ long ago, OK deraadt@

6 years agoPass correct size to uvm_vnp_setsize() for large files.
millert [Mon, 8 Jan 2018 16:16:16 +0000 (16:16 +0000)]
Pass correct size to uvm_vnp_setsize() for large files.

6 years agoAdd kqueue support for ext2fs based on ffs. OK deraadt@
millert [Mon, 8 Jan 2018 16:15:34 +0000 (16:15 +0000)]
Add kqueue support for ext2fs based on ffs.  OK deraadt@

6 years agogroup shared source files (e.g. SRCS_KEX) and allow compilation w/o OPENSSL
markus [Mon, 8 Jan 2018 15:37:21 +0000 (15:37 +0000)]
group shared source files (e.g. SRCS_KEX) and allow compilation w/o OPENSSL
ok djm@

6 years agomove subprocess() so scp/sftp do not need uidswap.o; ok djm@
markus [Mon, 8 Jan 2018 15:21:49 +0000 (15:21 +0000)]
move subprocess() so scp/sftp do not need uidswap.o; ok djm@

6 years agoswitch ssh-pkcs11-helper to new API; ok djm@
markus [Mon, 8 Jan 2018 15:18:46 +0000 (15:18 +0000)]
switch ssh-pkcs11-helper to new API; ok djm@

6 years agoonly ssh-keygen needs uuencode.o; only scp/sftp use progressmeter.o
markus [Mon, 8 Jan 2018 15:15:17 +0000 (15:15 +0000)]
only ssh-keygen needs uuencode.o; only scp/sftp use progressmeter.o

6 years agouuencode.h is not used
markus [Mon, 8 Jan 2018 15:14:44 +0000 (15:14 +0000)]
uuencode.h is not used

6 years agosplit client/server kex; only ssh-keygen needs uuencode.o;
markus [Mon, 8 Jan 2018 15:15:36 +0000 (15:15 +0000)]
split client/server kex; only ssh-keygen needs uuencode.o;
only scp/sftp use progressmeter.o; ok djm@

6 years agoOverride membar_producer() for octeon.
visa [Mon, 8 Jan 2018 13:44:43 +0000 (13:44 +0000)]
Override membar_producer() for octeon.

OK dlg@

6 years agoDrop two outdated entries
jca [Mon, 8 Jan 2018 13:39:06 +0000 (13:39 +0000)]
Drop two outdated entries

Namely POSIX character classes and tilde expansion within parameter
substitution.  While here, kill a useless line and fix a typo.
From Klemens Nanni

6 years agoRemove unused "args" param from kb_add(), it is always NULL.
millert [Mon, 8 Jan 2018 13:01:31 +0000 (13:01 +0000)]
Remove unused "args" param from kb_add(), it is always NULL.
OK anton@ jca@

6 years agooptimization and some cleanup; mostly from kshe (except the unmap() part)
otto [Mon, 8 Jan 2018 12:20:23 +0000 (12:20 +0000)]
optimization and some cleanup; mostly from kshe (except the unmap() part)

6 years agoKill entries that are obviously outdated/out of scope
jca [Mon, 8 Jan 2018 12:08:17 +0000 (12:08 +0000)]
Kill entries that are obviously outdated/out of scope

Pointed out by Klemens Nanni who had ~ the same diff