deraadt [Fri, 5 Jan 2018 05:53:56 +0000 (05:53 +0000)]
backout, because it breaks builds in dev/microcode.
Always build the parts of the tree that use a file.
dlg [Fri, 5 Jan 2018 05:08:53 +0000 (05:08 +0000)]
cast the result of the ternary ops in __swapXX to the relevant types.
it appears clang and guenther@ have a different way of interpreting
the c standards around type promotion and ternary operators compared
to gcc and me.
"no you do it" guenther@
patrick [Thu, 4 Jan 2018 23:34:06 +0000 (23:34 +0000)]
Receiving an AUTH event means that we successfully authenticated, thus
we have to move to the "trying to" ASSOC state. When association has
finished we will receive an ASSOC event, so that we can move into the
RUN state. After that point we will receive ethernet packets and the
WPA2 handshake can occur. The SET_SSID event will now only be used to
bring us back into SCAN when it notes an error, as it often completes
after we have already done the WPA2 handshake and can not be used to
switch between the different states.
bluhm [Thu, 4 Jan 2018 22:58:25 +0000 (22:58 +0000)]
Add more tests that check the timeout behavior of relayd.
jmc [Thu, 4 Jan 2018 20:38:12 +0000 (20:38 +0000)]
from paul de weerd: provide a more helpful Xr to ocspcheck, and note
that the path to "file" is not relative to the chroot;
jmc [Thu, 4 Jan 2018 20:09:56 +0000 (20:09 +0000)]
add HISTORY; requested by jiri navratil
krw [Thu, 4 Jan 2018 19:52:30 +0000 (19:52 +0000)]
Compact some whitespace, unsplit now shorter lines.
kettenis [Thu, 4 Jan 2018 19:46:31 +0000 (19:46 +0000)]
Add ARM EABI aliases and remove functions that are also provided by our libc.
This allows linking code compiled by clang with the gcc compiler driver
and makes sure we always use the softfloat implementation in libc. The
libc softfloat implementation is preferred over the one in libgcc as it
implements rounding modes and floating point exceptions.
ok patrick@
millert [Thu, 4 Jan 2018 19:06:16 +0000 (19:06 +0000)]
Back out sign compare changes that appear to cause problems on hppa.
Requested by deraadt@
krw [Thu, 4 Jan 2018 18:26:04 +0000 (18:26 +0000)]
Comments and grammar use '_decl' so use _decl instead of _declaration
in the names of the respective parsing functions.
deraadt [Thu, 4 Jan 2018 17:44:20 +0000 (17:44 +0000)]
glaring KNF violation
ccardenas [Thu, 4 Jan 2018 15:19:56 +0000 (15:19 +0000)]
Address TOCTOU issue with checking to ensure disks are regular files.
Reported by jca@.
Ok mlarkin@ and deraadt@
kettenis [Thu, 4 Jan 2018 14:30:08 +0000 (14:30 +0000)]
Implement support for calling EFI runtime services and use it to implement
a time-of-day clock device based on the GetTime() and SetTime() services.
The virtual memory mappings for the runtime services calls are implemented
through a separate pmap that is only activated when we make a runtime services
call.
ok tom@, visa@
tested by naddy@
mpi [Thu, 4 Jan 2018 14:21:00 +0000 (14:21 +0000)]
space -> tab
No object change.
mpi [Thu, 4 Jan 2018 11:03:48 +0000 (11:03 +0000)]
Unify <machine/mutex.h> a bit further.
Remove `mtx_lock' from i386, add volatile before `mtx_owner' where it
was missing.
Inputs from kettenis@, ok visa@
tb [Thu, 4 Jan 2018 11:02:57 +0000 (11:02 +0000)]
Back out tx mitigation again because it breaks suspend and resume at
least on x230 and x240. Problem noted by claudio on icb.
ok dlg
mpi [Thu, 4 Jan 2018 10:51:11 +0000 (10:51 +0000)]
Do a FREF/FRELE dance after calling fd_getfile().
This should be enought to prevent `fp' to disapear while sleeping in
malloc(9).
ok helg@
mpi [Thu, 4 Jan 2018 10:48:02 +0000 (10:48 +0000)]
Include timeout & tasks in 'struct ifnet' instead of always allocating
them as M_TEMP.
ok visa@
mpi [Thu, 4 Jan 2018 10:45:30 +0000 (10:45 +0000)]
Unifdef snd/rcv.
ok visa@, claudio@
krw [Thu, 4 Jan 2018 03:02:05 +0000 (03:02 +0000)]
parse_lease_time() is an unnecessary abstraction. Just use
parse_number('L') since it is parsing unsigned 32bit integers.
krw [Thu, 4 Jan 2018 02:27:55 +0000 (02:27 +0000)]
We only parse decimal numbers, so parse_decimal() -> parse_number() to
match grammar comments and improve euphony.
dlg [Thu, 4 Jan 2018 00:33:54 +0000 (00:33 +0000)]
remove support for decapsulating LLC/SNAP frames.
the code was broken and noone noticed. this argues that we don't
need it.
ok mpi@
deraadt [Wed, 3 Jan 2018 23:39:02 +0000 (23:39 +0000)]
sync
dlg [Wed, 3 Jan 2018 23:20:10 +0000 (23:20 +0000)]
rework __swapXX to avoid the use of __statement.
the primary motivation of this was to allow the use of things like
htons() and htole16() as case labels. previously gcc would vomit
with "expression is not an integer constant expression" if you tried
that.
ok guenther@
dlg [Wed, 3 Jan 2018 23:11:06 +0000 (23:11 +0000)]
assign 1 instead of the result of htole16(1) to a uint8_t.
on big endian archs the 1 is shifted to the high byte, which then
gets lost when it's assigned to the uint8_t. at worst we lose the
value, at best the compiler has a teary and fixes it.
this is the fix for a compiler teary.
ok claudio@
patrick [Wed, 3 Jan 2018 21:01:16 +0000 (21:01 +0000)]
Since the PCI attachment code already uses mbufs for RX packets, we can
push the mbuf allocation down into the USB attachment code and now pass
an mbuf to the bwfm(4) receive function.
kettenis [Wed, 3 Jan 2018 20:41:31 +0000 (20:41 +0000)]
Instead of adding each and every clock to the list of clocks that are enabled
by default, simply assume they are. But do check for this when we attach.
kettenis [Wed, 3 Jan 2018 20:10:40 +0000 (20:10 +0000)]
Add ASMedia ASM1061 SATA to the list of devices as it is an AHCI controller
but doesn't advertise itself as such.
ok dlg@, jmatthew@
denis [Wed, 3 Jan 2018 19:39:36 +0000 (19:39 +0000)]
Add support for IPv6 over MPLS pseudowire aka mpw(4)
OK claudio@ jca@
denis [Wed, 3 Jan 2018 19:39:07 +0000 (19:39 +0000)]
Add support for IPv6 over MPLS pseudowire aka mpw(4)
OK claudio@ jca@
schwarze [Wed, 3 Jan 2018 19:12:20 +0000 (19:12 +0000)]
Stop "dd if=/dev/zero of=/dev/null bs=1 count=2 skip=0x7fffffffffffffff"
from erroring out with "dd: skip: Undefined error: 0", mostly for
consistency and to avoid the unidiomatic, wrong looking code.
Patch from Bulat Musin <mbulatka at yandex dot ru>.
OK guenther@ millert@
sunil [Wed, 3 Jan 2018 11:12:21 +0000 (11:12 +0000)]
Use crypt_checkpass(3) instead of crypt(3).
Based on a diff from Edgar Pettijohn.
Ok gilles@ eric@
rpe [Wed, 3 Jan 2018 10:22:38 +0000 (10:22 +0000)]
Always present the same question at the end of installs or upgrades.
Offer to exit to shell, halt or reboot the system, where 'reboot' is
the default answer. Change default answer to 'halt' for installs if
MDHALT is set to 'y'.
suggested by landry@
OK halex@
patrick [Wed, 3 Jan 2018 08:43:10 +0000 (08:43 +0000)]
Add size for free(9) in the bwfm(4) PCI attachment code.
From Michael W. Bombardieri
bentley [Wed, 3 Jan 2018 08:43:09 +0000 (08:43 +0000)]
Include https:// in links, so HTML output doesn't assume they're relative.
ok jmc@ schwarze@
jmc [Wed, 3 Jan 2018 08:17:18 +0000 (08:17 +0000)]
pretty up the "start" command;
jmc [Wed, 3 Jan 2018 08:13:16 +0000 (08:13 +0000)]
fix SEE ALSO;
dlg [Wed, 3 Jan 2018 06:27:42 +0000 (06:27 +0000)]
remove __swap16_multi because endian.h provides swap16_multi without it
ok guenther@
ccardenas [Wed, 3 Jan 2018 05:39:56 +0000 (05:39 +0000)]
Add initial CD-ROM support to VMD via vioscsi.
* Adds 'cdrom' keyword to vm.conf(5) and '-r' to vmctl(8)
* Support various sized ISOs (Limitation of 4G ISOs on Linux guests)
* Known working guests: OpenBSD (primary), Alpine Linux (primary),
CentOS 6 (secondary), Ubuntu 17.10 (secondary).
NOTE: Secondary indicates some issue(s) preventing full/reliable
functionality outside the scope of the vioscsi work.
* If the attached disks are non-bootable (i.e. empty), SeaBIOS (vmd's
default BIOS) will boot from CD-ROM.
ok mlarkin@, jca@
kettenis [Wed, 3 Jan 2018 04:15:51 +0000 (04:15 +0000)]
Add remap_bynode() since I use it in the rkpcie(4) implementation.
kettenis [Tue, 2 Jan 2018 22:56:01 +0000 (22:56 +0000)]
rkpcie(4)
kettenis [Tue, 2 Jan 2018 22:47:45 +0000 (22:47 +0000)]
Enable rkpcie(4).
kettenis [Tue, 2 Jan 2018 22:33:17 +0000 (22:33 +0000)]
Initial stab at a driver for the PCIe interface on the Rockhip RK3399. For
now it cheats when setting up an interrupt handler. This cheat only works
because it currently effectively only supports a single device. But the
cheat works well enough to support the Firefly SATA adapter board.
kettenis [Tue, 2 Jan 2018 21:24:02 +0000 (21:24 +0000)]
regen
kettenis [Tue, 2 Jan 2018 21:23:13 +0000 (21:23 +0000)]
Add Rockchip vendor and RK3399 Root Complex device.
gilles [Tue, 2 Jan 2018 19:11:06 +0000 (19:11 +0000)]
we haven't updated the version in a while despite many commits which is
confusing for people running the portable version
anton [Tue, 2 Jan 2018 17:39:34 +0000 (17:39 +0000)]
Add missing mouse type. Reported by bit shifter on misc@; ok bru@
mpi [Tue, 2 Jan 2018 12:57:30 +0000 (12:57 +0000)]
Move the cloning bits in their own function.
This will be needed to select the proper cloning route in a multipath
scenario.
While here remove a NET_ASSERT_LOCKED(), the routing table doesn't need
the lock.
ok dlg@
mpi [Tue, 2 Jan 2018 12:54:07 +0000 (12:54 +0000)]
Do not memset() the whole structure in sorflush() to keep `sb_flagsintr'
untouched.
ok bluhm@, visa@
mpi [Tue, 2 Jan 2018 12:52:17 +0000 (12:52 +0000)]
Move the NET_LOCK() inside the switch and start documenting which field
is protected by which lock.
ok bluhm@, visa@
jmc [Tue, 2 Jan 2018 08:05:03 +0000 (08:05 +0000)]
some grammar fixes; from dholland@netbsd, -r1.68
jmc [Tue, 2 Jan 2018 08:03:40 +0000 (08:03 +0000)]
update currency exchange rates;
dlg [Tue, 2 Jan 2018 07:08:10 +0000 (07:08 +0000)]
reintroduce tx mitigation
to quote the previous commit:
> this replaces ifq_start with code that waits until at least 4 packets
> have been queued on the ifq before calling the drivers start routine.
> if less than 4 packets get queued, the start routine is called from
> a task in a softnet tq.
>
> 4 packets was chosen this time based on testing sephe did in dfly
> which showed no real improvement when bundling more packets. hrvoje
> popovski tested this on several nics and found an improvement of
> 10 to 20 percent when forwarding across the board.
>
> because some of the ifq's work could be sitting on a softnet tq,
> ifq_barrier now calls taskq_barrier to guarantee any work that was
> pending there has finished.
>
> ok mpi@ visa@
this was backed out because of a race in the net80211 stack that
anton@ hit. mpi@ committed a workaround for it in revision 1.30 of
src/sys/net80211/ieee80211_pae_output.c.
im putting this in again so we can see what breaks next.
guenther [Tue, 2 Jan 2018 06:56:41 +0000 (06:56 +0000)]
When invoked with no file arguments, display usage and fail instead of
silently doing nothing.
suggested by Klemens Nanni (kl3 (at) posteo.org), from FreeBSD r92772
ok millert@
guenther [Tue, 2 Jan 2018 06:40:55 +0000 (06:40 +0000)]
Don't #include fcntl.h when _KERNEL is defined.
inspired by FreeBSD r24131
ok millert@ sthen@
guenther [Tue, 2 Jan 2018 06:38:45 +0000 (06:38 +0000)]
Stop assuming <sys/file.h> will pull in fcntl.h when _KERNEL is defined.
ok millert@ sthen@
guenther [Tue, 2 Jan 2018 06:07:21 +0000 (06:07 +0000)]
Fix an off-by-one in the free(9) "passed size was too small" check:
if the size passed is exactly half the size of the bucket that the
allocation was actually from, then it was incorrect.
problem noted by florian@
ok florian@ visa@
krw [Tue, 2 Jan 2018 00:13:27 +0000 (00:13 +0000)]
More grammar comments improvements.
patrick [Mon, 1 Jan 2018 22:41:56 +0000 (22:41 +0000)]
For whatever reason the firmware needs more RX buffers available as
we typically use, which unfortunately creates a bigger memory foot-
print. With this the receive path can be made to work.
patrick [Mon, 1 Jan 2018 22:41:03 +0000 (22:41 +0000)]
Put the code that prints the firmware's debug console into a function
so we can read and print the messages printed by the firmware when we
are debugging the driver.
millert [Mon, 1 Jan 2018 19:45:56 +0000 (19:45 +0000)]
Add WARNINGS=yes to ksh and fix the resulting sign compare warnings.
OK tb@
denis [Mon, 1 Jan 2018 19:41:01 +0000 (19:41 +0000)]
Fix wrong ETHERTYPE_IPV6
OK jca@ kettenis@ job@
kettenis [Mon, 1 Jan 2018 19:18:30 +0000 (19:18 +0000)]
Make lld respect ALIGN directives on output sections even with -r. Fixes
the generation of gap.o.
ok patrick@
bluhm [Mon, 1 Jan 2018 16:16:23 +0000 (16:16 +0000)]
The missing NULL check for the return value of malloc(9) with
M_NOWAIT is wrong. Use M_WAITOK in ipmi_match() during attach.
OK kettenis@
krw [Mon, 1 Jan 2018 15:07:58 +0000 (15:07 +0000)]
Tweak syntax comments into consistency and completeness.
claudio [Mon, 1 Jan 2018 15:01:09 +0000 (15:01 +0000)]
RSA_private_{en,de}crypt() can fail and will return -1 in that case.
Check for this in the ca process and return a valid answer to the
relay process. This fixes rsae_send_imsg poll timeouts blocking relay
processes as seen by Mischa Peters and myself.
OK benno@
espie [Mon, 1 Jan 2018 14:02:37 +0000 (14:02 +0000)]
placeholder for new keyword that doesn't do anything yet, so that when it
gets used, tools already know about it a bit.
(meant to simplify wantlib updates down the line)
otto [Mon, 1 Jan 2018 12:41:48 +0000 (12:41 +0000)]
Only init chunk_info once, plus some moving of code to group related functions.
nicm [Mon, 1 Jan 2018 11:19:08 +0000 (11:19 +0000)]
Prefer PWD for current directory if present in client, from Wei Zhao in
GitHub issue 1183.
nicm [Mon, 1 Jan 2018 11:03:54 +0000 (11:03 +0000)]
Add C-g at command prompt for emacs people, GitHub issue 1213.
florian [Mon, 1 Jan 2018 09:01:21 +0000 (09:01 +0000)]
free(9) sizes for sys_execve.
Convert the hand rolled loop to strlcpy which gives us the size for
free(9).
OK visa
florian [Mon, 1 Jan 2018 08:55:43 +0000 (08:55 +0000)]
We are either allocating 2 or three array members. Unroll while loop
to be able to call free(9) with sizes.
off-by-one pointed out by guenther
OK visa
jsg [Mon, 1 Jan 2018 08:23:19 +0000 (08:23 +0000)]
copyright++;
nicm [Sun, 31 Dec 2017 20:00:44 +0000 (20:00 +0000)]
Initialize the size of new panes created by the even-* layout correctly;
reported by Andreas Kahari and Anton Lindqvist.
kettenis [Sun, 31 Dec 2017 15:55:02 +0000 (15:55 +0000)]
sxitemp(4)
kettenis [Sun, 31 Dec 2017 15:41:25 +0000 (15:41 +0000)]
Add Allwinner R40/V40 support.
kettenis [Sun, 31 Dec 2017 15:35:09 +0000 (15:35 +0000)]
Implement R40/V40 thermal sensor clock.
kettenis [Sun, 31 Dec 2017 13:54:48 +0000 (13:54 +0000)]
Enable sxitemp(4).
kettenis [Sun, 31 Dec 2017 13:54:09 +0000 (13:54 +0000)]
Add sxitemp(4), a driver for the temperature sensors on the Allwinner H5 SoC.
kettenis [Sun, 31 Dec 2017 13:52:18 +0000 (13:52 +0000)]
Implement H3/H5 thermal sensor clocks.
kettenis [Sun, 31 Dec 2017 11:14:31 +0000 (11:14 +0000)]
Enable sypwr(4).
kettenis [Sun, 31 Dec 2017 11:13:59 +0000 (11:13 +0000)]
Add sypwr(4), a driver for the Silergy SY8106A regulator found on many boards
using the Allwinner H3/H5 SoC.
kettenis [Sun, 31 Dec 2017 10:54:39 +0000 (10:54 +0000)]
Implement H3/H5 CPU clock.
anton [Sun, 31 Dec 2017 09:40:41 +0000 (09:40 +0000)]
Remove an unused variable and instruct lex not to produce the input() function
since it's never used; ok guenther@
kettenis [Sun, 31 Dec 2017 08:42:04 +0000 (08:42 +0000)]
Tighten the permissions used in the early stage page tables somewhat.
Add an interface to establish additional VA=PA 1G block mappings for use by
upcoming EFI runtime services support.
ok guenther@
guenther [Sat, 30 Dec 2017 23:08:29 +0000 (23:08 +0000)]
Don't pull in <sys/file.h> just to get fcntl.h
ok deraadt@ krw@
guenther [Sat, 30 Dec 2017 20:46:59 +0000 (20:46 +0000)]
Delete unnecessary <sys/file.h> includes
ok millert@ krw@
kettenis [Sat, 30 Dec 2017 19:04:00 +0000 (19:04 +0000)]
Somewhow I botched rev 1.3, which has a = where a |= is needed. Fix this.
kettenis [Sat, 30 Dec 2017 16:32:52 +0000 (16:32 +0000)]
Add support for the USB PHY of the Allwinner R40/V40.
kettenis [Sat, 30 Dec 2017 16:30:37 +0000 (16:30 +0000)]
Implement Allwinner A40/V40 support.
kettenis [Sat, 30 Dec 2017 13:34:56 +0000 (13:34 +0000)]
The generic clock framework expects frequencies in Hz whereas the imx clock
code was providing them in kHz. Get rid of this discrepancy.
Hook up the CPU clock to the generic clock framework and remove the
cpuspeed function as it is no longer needed.
ok patrick@
kettenis [Sat, 30 Dec 2017 12:45:36 +0000 (12:45 +0000)]
Add support for the PRCM clocks on Allwinner A23/A33.
Tested by kevlo@
kettenis [Sat, 30 Dec 2017 10:20:34 +0000 (10:20 +0000)]
Add copyin32 implementation.
ok guenther@
guenther [Sat, 30 Dec 2017 10:10:34 +0000 (10:10 +0000)]
get_date(), from getdate.y, was last used in the sparc support, so zap it
ok kettenis@
kettenis [Sat, 30 Dec 2017 08:39:49 +0000 (08:39 +0000)]
Implement CPU_BUSY_CYCLE and intr_disable/restore.
ok patrick@, tom@
otto [Sat, 30 Dec 2017 07:35:02 +0000 (07:35 +0000)]
The `F' option no longer disables delayed freeing; from kshe
Also, fix a verb-noun mixup; ok jmc@
jmc [Sat, 30 Dec 2017 07:32:09 +0000 (07:32 +0000)]
backslash also escapes newlines when quoted;
from kshe
tb [Sat, 30 Dec 2017 07:21:10 +0000 (07:21 +0000)]
Simplify switch by removing some unneeded cases.
from kshe
tb [Sat, 30 Dec 2017 07:19:05 +0000 (07:19 +0000)]
Avoid one-byte overflow in error path. If the format string ends in an
invalid specifier like `%l', p will already point to the trailing NUL
upon entering the switch, wherein the instruction
*++p = '\0';
will write another NUL after it, but there is no guarantee that the
buffer extends beyond that first NUL; thus, in the rare case where it
does not, this assignment will write one byte past its end.
from kshe
bluhm [Fri, 29 Dec 2017 23:55:22 +0000 (23:55 +0000)]
Make sure that pf_mbuf_link_state_key() does not overwrite an
existing statekey in the mbuf header. Reset the statekey in
m_dup_pkthdr().
suggested by and OK sahan@