visa [Tue, 23 Aug 2016 12:54:09 +0000 (12:54 +0000)]
The TLB refill and XTLB refill exceptions use distinct exception vectors
on Loongson 3A. Consequently, the kernel has to set up both vectors to
prevent panics with virtual address references.
On Loongson 2F, it is enough to set up the TLB refill vector because
the XTLB refill exception uses the same vector address.
ok miod@
dlg [Tue, 23 Aug 2016 12:37:44 +0000 (12:37 +0000)]
pool_setipl
dlg [Tue, 23 Aug 2016 12:37:11 +0000 (12:37 +0000)]
pool_setipl
mpi [Tue, 23 Aug 2016 11:03:10 +0000 (11:03 +0000)]
Use rtalloc(9) directly instead of in6_selectsr() in NS/NA output
routines.
This is another little step towards deprecating 'struct route{,_in6}'.
ok bluhm@
mpi [Tue, 23 Aug 2016 11:01:16 +0000 (11:01 +0000)]
Do not use a single global struct route_in6 to cache route lookups.
This is a little step towards deprecating 'struct route{,_in6}'.
ok jca@, claudio@
mpi [Tue, 23 Aug 2016 09:26:02 +0000 (09:26 +0000)]
Make the 'ifi' global local to dhclient.c and pass it as an argument to
functions needing it.
This is the first step to support multiple interfaces in one dhclient(8)
instance.
ok krw@
djm [Tue, 23 Aug 2016 08:17:42 +0000 (08:17 +0000)]
downgrade an error() to a debug2() to match similar cases
in addr_match_list()
djm [Tue, 23 Aug 2016 08:17:04 +0000 (08:17 +0000)]
add tests for addr_match_list()
kettenis [Tue, 23 Aug 2016 06:46:17 +0000 (06:46 +0000)]
Implement support for DT_INIT_ARRAY, DT_FINI_ARRAY and DT_PREINIT_ARRAY.
Don't skip DT_INIT and DT_FINI for the main executable. This matches what
Linux and Solaris do.
ok guenther@
djm [Tue, 23 Aug 2016 06:36:23 +0000 (06:36 +0000)]
remove Protocol directive from client/server configs that causes
spammy deprecation warnings
hardcode SSH_PROTOCOLS=2, since that's all we support on the server
now (the client still may support both, so it could get confused)
guenther [Tue, 23 Aug 2016 06:00:28 +0000 (06:00 +0000)]
Instead of doing strcmp(argv0), track the invocation mode (pax/tar/cpio)
in a separate variable
ok deraadt@
dlg [Tue, 23 Aug 2016 04:28:18 +0000 (04:28 +0000)]
pool_setipl for tmpfs.
jsg [Tue, 23 Aug 2016 04:04:23 +0000 (04:04 +0000)]
remove duplicated test for em_82574
guenther [Tue, 23 Aug 2016 03:31:44 +0000 (03:31 +0000)]
Only try to set the times on a directory once, at the end, to avoid
duplication of warning messages
ok krw@
guenther [Tue, 23 Aug 2016 03:28:01 +0000 (03:28 +0000)]
Convert %q to %ll in format strings
ok natano@ krw@
djm [Tue, 23 Aug 2016 03:24:10 +0000 (03:24 +0000)]
fix negated address matching where the address list consists of a
single negated match, e.g. "Match addr !192.20.0.1"
Report and patch from Jakub Jelen. bz#2397 ok dtucker@
djm [Tue, 23 Aug 2016 03:22:49 +0000 (03:22 +0000)]
fix matching for pattern lists that contain a single negated match,
e.g. "Host !example"
report and patch from Robin Becker. bz#1918 ok dtucker@
deraadt [Mon, 22 Aug 2016 22:48:14 +0000 (22:48 +0000)]
sync
kettenis [Mon, 22 Aug 2016 22:06:59 +0000 (22:06 +0000)]
Fix two small bugs in the new reset API code.
nicm [Mon, 22 Aug 2016 20:07:58 +0000 (20:07 +0000)]
Two minor fixes from dilyan palauzov.
kettenis [Mon, 22 Aug 2016 19:43:49 +0000 (19:43 +0000)]
Remove unused code and definitions.
schwarze [Mon, 22 Aug 2016 19:41:18 +0000 (19:41 +0000)]
This code served us well for many years,
but we are unlikely to reactivate it; espie@ agrees.
The new code is in /usr/src/usr.bin/mandoc/mandocdb.c.
kettenis [Mon, 22 Aug 2016 19:38:42 +0000 (19:38 +0000)]
Use new clock API.
kettenis [Mon, 22 Aug 2016 19:31:27 +0000 (19:31 +0000)]
Use new clock API to manipulate clocks.
kettenis [Mon, 22 Aug 2016 19:29:32 +0000 (19:29 +0000)]
Add support for the gmac clock.
kettenis [Mon, 22 Aug 2016 19:28:27 +0000 (19:28 +0000)]
Add an API to set the clock frequency.
kettenis [Mon, 22 Aug 2016 18:31:07 +0000 (18:31 +0000)]
Move all the platform-specific code into a new function sxiehci_attach_phy()
making use of pinctrl, clock and reset APIs where appropriate.
kettenis [Mon, 22 Aug 2016 18:18:35 +0000 (18:18 +0000)]
Add support for the usb clock. Also implements reset signal support.
kettenis [Mon, 22 Aug 2016 18:16:58 +0000 (18:16 +0000)]
Add a reset signal API alongside the clock API.
mpi [Mon, 22 Aug 2016 17:43:29 +0000 (17:43 +0000)]
Use a reachable next hop for test 13, which make it fail as it should.
jsing [Mon, 22 Aug 2016 17:12:35 +0000 (17:12 +0000)]
Various clean up and reorganisation of the connection info handling code.
In particular, rename tls_free_conninfo() to tls_conninfo_free() and make
it a real free function. Rename tls_get_conninfo() to
tls_conninfo_populate() and have it allocate the struct tls_conninfo (after
freeing any existing one).
ok beck@
jsing [Mon, 22 Aug 2016 17:08:10 +0000 (17:08 +0000)]
Stick with the usual 'if NULL return NULL' idiom.
ok beck@
mpi [Mon, 22 Aug 2016 16:55:14 +0000 (16:55 +0000)]
Sync counters now that ifa_ifwithroute() no longer uses ifa_ifwithnet().
mpi [Mon, 22 Aug 2016 16:53:59 +0000 (16:53 +0000)]
Use rtalloc(9) instead of ifa_ifwithnet() to find an interface
when adding a route to gateway to ensure a most specific match.
This makes "# route add" coherent to "# route get" even with
p2p interfaces. Fix a problem reported by Mart Tõnso.
This also fix rttest20 after the introduction of RTF_CACHED.
ok vgross@
millert [Mon, 22 Aug 2016 16:27:00 +0000 (16:27 +0000)]
Replace "union sockunion" with "union sockaddr_union" which is also
used in the kernel. This makes it possible to remove the casts to
"struct sockaddr *" when calling networking syscalls.
OK jsing@ jca@
schwarze [Mon, 22 Aug 2016 16:12:52 +0000 (16:12 +0000)]
When trying to edit an existing database with makewhatis(8) -d or -u
but reading the database fails, report the full path to the database
on standard error, and mention that the database is automatically
recreated from scratch.
Suggested by espie@.
schwarze [Mon, 22 Aug 2016 16:05:56 +0000 (16:05 +0000)]
When running into a mandoc.db(5) file still using the obsolete
format based on SQLite 3, say so in words that mortals can
understand rather than babbling about hex magic.
Suggested by espie@.
mpi [Mon, 22 Aug 2016 16:03:03 +0000 (16:03 +0000)]
Sync refcount with recent change.
mpi [Mon, 22 Aug 2016 16:01:52 +0000 (16:01 +0000)]
Make the ``rt_gwroute'' pointer of RTF_GATEWAY entries immutable.
This means that no protection is needed to guarantee that the next hop
route wont be modified by CPU1 while CPU0 is dereferencing it in a L2
resolution functions.
While here also fix an ``ifa'' leak resulting in RTF_GATEWAY being always
invalid.
dlg@ likes it, inputs and ok bluhm@
mpi [Mon, 22 Aug 2016 15:37:23 +0000 (15:37 +0000)]
Do not dereference ``rt->rt_ifa'' after calling rtfree(9).
This could result in a use after free if the route entry was holding
the last reference of the address descriptor.
ok jca@, bluhm@, claudio@
jsing [Mon, 22 Aug 2016 15:02:18 +0000 (15:02 +0000)]
Enable SNI support in httpd(8).
ok reyk@
jsing [Mon, 22 Aug 2016 14:58:26 +0000 (14:58 +0000)]
Bump TLS_API due to the addition of server side SNI functions.
jsing [Mon, 22 Aug 2016 14:57:49 +0000 (14:57 +0000)]
Bump libtls minor due to the addition of symbols.
jsing [Mon, 22 Aug 2016 14:55:59 +0000 (14:55 +0000)]
Provide an API that enables server side SNI support - add the ability to
provide additional keypairs (via tls_config_add_keypair_{file,mem}()) and
allow the server to determine what servername the client requested (via
tls_conn_servername()).
ok beck@
jsing [Mon, 22 Aug 2016 14:51:37 +0000 (14:51 +0000)]
Create contexts for server side SNI - these include the additional SSL_CTX
that is required for certificate switching with libssl and the certificate
itself so that we can match against the subject and SANs. Hook up the
servername callback and switch to the appropriate SSL_CTX if we find a
matching certificate.
ok beck@
jmc [Mon, 22 Aug 2016 13:39:52 +0000 (13:39 +0000)]
shorten the pkcs8 text;
kettenis [Mon, 22 Aug 2016 11:24:45 +0000 (11:24 +0000)]
Use the new clock API.
kettenis [Mon, 22 Aug 2016 11:23:54 +0000 (11:23 +0000)]
Implement interfaces to disable clocks and add interfaces that enable or
disable all clocks for a device. The latter interfaces are useful for
devices that have multiple clocks that don't have specific names/purposes
such as sxiahci(4).
mpi [Mon, 22 Aug 2016 10:40:36 +0000 (10:40 +0000)]
Call csignal() and selwakeup() from a KERNEL_LOCK'd task.
This will allow us make bpf_tap() KERNEL_LOCK() free.
Discussed with dlg@ and input from guenther@
mpi [Mon, 22 Aug 2016 10:33:22 +0000 (10:33 +0000)]
Sizes for free(9) from David Hill.
claudio [Mon, 22 Aug 2016 10:23:42 +0000 (10:23 +0000)]
Refactor the uio to mbuf code out of sosend and start to make use of
MCLGETI and large mbuf clusters. This should speed up local connections
a fair bit. OK dlg@ and bluhm@ (after reverting the M_WAIT change on the
cluster allocation)
kettenis [Mon, 22 Aug 2016 09:37:27 +0000 (09:37 +0000)]
Use the new clock API to get the clock frequency from the device tree and the
actual hardware that controls the clock. This gets rid of all the hardcoded
clock frequencies for specific hardware.
Tested by jsg@ on the BBB, mglocker@ on the C.H.I.P. and myself on the
Banana Pi and Raspberry Pi 3.
ok jsg@
kettenis [Mon, 22 Aug 2016 06:48:38 +0000 (06:48 +0000)]
Add code to enable the pll6 clock.
tedu [Mon, 22 Aug 2016 06:45:12 +0000 (06:45 +0000)]
fix typos
deraadt [Mon, 22 Aug 2016 05:21:10 +0000 (05:21 +0000)]
sync
deraadt [Mon, 22 Aug 2016 04:33:07 +0000 (04:33 +0000)]
Sorry Andrew and Luke, I'm pretty sure we deleted your IRIX and VMS code.
jsg [Mon, 22 Aug 2016 01:41:59 +0000 (01:41 +0000)]
Before pmap7.c rev 1.35 and pmap.h rev 1.44 DMA'able memory with the
BUS_DMA_COHERENT flag was mapped as device memory which does not use the
store buffer. It is now mapped as normal inner and outer non-cacheable
which does.
While we drain the cpu store buffer for this case, on cortex a9 systems we
also need to explicitly drain the PL310 L2's store buffer. With PL310
revisions r3p2 and later this is done automatically after being present in
the store buffer for 256 cycles. On i.MX6 PL310 is rev r3p1 which does
not have this behaviour. This issue is i.MX6 errata ERR055199 and PL310
errata 769419.
This change restores io performance with a usb flash drive attached to
my cubox. Raw reads go from 3 MB/s to 19 MB/s for example.
Based on code written by patrick@ some time ago.
ok kettenis@ patrick@
kettenis [Sun, 21 Aug 2016 23:02:32 +0000 (23:02 +0000)]
Use generic clock API to enable the module clock.
kettenis [Sun, 21 Aug 2016 23:01:57 +0000 (23:01 +0000)]
Add support for all the ahb, apb0 and apb1 clock gating devices on sun4i,
sun5i and sun7i.
kettenis [Sun, 21 Aug 2016 21:39:59 +0000 (21:39 +0000)]
Implement some fundamental clocks using the new clock framework.
ok patrick@
kettenis [Sun, 21 Aug 2016 21:38:05 +0000 (21:38 +0000)]
Add a minimal clock "framework". Build it on armv7.
ok patrick@
tedu [Sun, 21 Aug 2016 21:23:48 +0000 (21:23 +0000)]
introduce a union of sockaddr types and eliminate a lot of casts.
schwarze [Sun, 21 Aug 2016 21:10:52 +0000 (21:10 +0000)]
bugfix: when fgetwc(3) fails, fgetwln(3) must fail as well;
OK jca@ martijn@ millert@
halex [Sun, 21 Aug 2016 19:22:15 +0000 (19:22 +0000)]
Make lease_value() unescape quoted strings. To be fully compliant, we
should unvis() it too, but I think this is enough, at least for now.
ok krw@
rpe [Sun, 21 Aug 2016 18:47:01 +0000 (18:47 +0000)]
tweak previous
OK halex, krw
kettenis [Sun, 21 Aug 2016 14:41:51 +0000 (14:41 +0000)]
Fix a pasto in a comment.
schwarze [Sun, 21 Aug 2016 13:31:21 +0000 (13:31 +0000)]
fix obvious typo in the .Dt section number
natano [Sun, 21 Aug 2016 09:23:33 +0000 (09:23 +0000)]
There are three callers of update_vattr(). Two of them don't use the
updated struct vattr afterwards, so the call can be removed. Remove both
calls and the function itself, inlining the last remaining call.
ok millert
jsg [Sun, 21 Aug 2016 07:08:46 +0000 (07:08 +0000)]
When detecting the console on attach test fa_node and stdout_node not fa_reg
and pl011consaddr as the pl011consaddr addresss may have been translated.
Set cn_dev at this point as well as in com(4) and imxuart(4) to handle cases
where stdout is not serial0.
jsg [Sun, 21 Aug 2016 06:47:47 +0000 (06:47 +0000)]
Remove the remains of a workaround to get GIC ranges for exynos.
Exynos differed from everything else by not having GIC at fixed offsets
from PERIPHBASE. Now that ampintc/GIC attaches using FDT we get the ranges
out of the FDT instead of using offsets from PERIPHBASE.
jsg [Sun, 21 Aug 2016 06:36:23 +0000 (06:36 +0000)]
Dynamically attach exuart using the FDT.
beck [Sun, 21 Aug 2016 03:26:04 +0000 (03:26 +0000)]
Change sleep from crazy backoff to one second of sleep between retries
ok deraadt@
deraadt [Sat, 20 Aug 2016 23:05:50 +0000 (23:05 +0000)]
sync
kettenis [Sat, 20 Aug 2016 21:08:16 +0000 (21:08 +0000)]
Fix indentation.
kettenis [Sat, 20 Aug 2016 21:07:07 +0000 (21:07 +0000)]
Argh, commit from the wrong tree. Revert previous commit.
kettenis [Sat, 20 Aug 2016 21:04:18 +0000 (21:04 +0000)]
Fix indentation.
eric [Sat, 20 Aug 2016 20:47:08 +0000 (20:47 +0000)]
Properly initialize the message parser. fix a regression where the
message headers would not be altered as expected.
ok gilles@
jmc [Sat, 20 Aug 2016 20:23:29 +0000 (20:23 +0000)]
add "armv7" to the Dt line;
millert [Sat, 20 Aug 2016 20:22:28 +0000 (20:22 +0000)]
Sync connect_wait() example with its real usage in ftp(1).
millert [Sat, 20 Aug 2016 20:18:42 +0000 (20:18 +0000)]
Use connect(2) + a connect_wait() function instead of connect_sync(),
similar to the example in connect(2). OK tedu@
kettenis [Sat, 20 Aug 2016 19:53:31 +0000 (19:53 +0000)]
sximmc(4)
kettenis [Sat, 20 Aug 2016 19:44:02 +0000 (19:44 +0000)]
Enable sximmc(4).
kettenis [Sat, 20 Aug 2016 19:42:55 +0000 (19:42 +0000)]
Increase MSGBUFSIZE to 8 pages.
ok deraadt@
kettenis [Sat, 20 Aug 2016 19:41:14 +0000 (19:41 +0000)]
Various improvements to make this work better.
Restrict the block size to 512 bytes for now. While Linux and NetBSD seem
to allow larger block sizes (4096 and 8192 bytes) I'm getting errors wth an
older 2G Sandisk SD card that advertises support for 1024 byte blocks.
Implement switching between the 400 KHz "discovery" clock and the 25 MHz
"operational" clock. Enable the right clock based on the SD/MMC "unit"
number. Don't advertise highspeed capabilities for now as I'm not sure
how the adjust the clock timing in that case.
Fix the way we calculate the number of blocks and the size of the total data
transfer to matc what we do in sdhc(4).
kettenis [Sat, 20 Aug 2016 19:34:44 +0000 (19:34 +0000)]
Add some code to set the SD/MMC clocks.
kettenis [Sat, 20 Aug 2016 19:31:01 +0000 (19:31 +0000)]
Don't set MSGBUFSIZE here such that the setting in <machine/param.h> takes
effect. This will let us have different settings on armv7 and zaurus and
also unconfuses this developer.
ok tom@, deraadt@
stefan [Sat, 20 Aug 2016 19:22:05 +0000 (19:22 +0000)]
Format string fixes in debug code: need %llx to print 64 bit values
ok kettenis@ deraadt@ mlarkin@
jca [Sat, 20 Aug 2016 19:08:57 +0000 (19:08 +0000)]
Declare all _asr_* debug functions as hidden.
Reported by & similar diff by guenther@ some time ago, ok eric@
schwarze [Sat, 20 Aug 2016 18:44:07 +0000 (18:44 +0000)]
already in v2 according to tuhs.org/Archive/PDP-11/Distributions/research/1972_stuff/unix_2nd_edition_manual.pdf
patch from Sevan Janiyan <venture37 at geeklan dot co dot uk>
schwarze [Sat, 20 Aug 2016 17:58:09 +0000 (17:58 +0000)]
When a mismatching end macro occurs while at least two nested blocks
are open, all except the innermost open block got a bogus MDOC_ENDED
marker, in some situations triggering segfaults down the road
which tb@ found with afl(1).
Fix the logic error by figuring out up front whether an end macro
has a matching body, and if it hasn't, don't mark any blocks as broken.
schwarze [Sat, 20 Aug 2016 15:58:16 +0000 (15:58 +0000)]
When scanning upwards for a column list to put a .Ta macro in,
ignore body end markers of lists breaking other blocks.
Fixing a logical error that caused a NULL deref found by tb@ with afl(1).
kettenis [Sat, 20 Aug 2016 15:44:04 +0000 (15:44 +0000)]
Replace bcmmuart(4) with com(4).
ok patrick@
jca [Sat, 20 Aug 2016 15:10:18 +0000 (15:10 +0000)]
Tweak timer debug output.
-RA timer on em0 is set to 16:0
+RA timer on em0 is set to 16.0s
jca [Sat, 20 Aug 2016 15:05:52 +0000 (15:05 +0000)]
Move counters from u_quad_t to uint64_t.
Printing is already done with %llu/(unsigned long long) casts.
tb [Sat, 20 Aug 2016 15:04:21 +0000 (15:04 +0000)]
fsdb(8) sucks in and is pledged by fsck(8). Since it uses editline(3),
add a special case for the missing "rpath" and "tty" promises.
Issue found and initial analysis by Jan Stary, thanks!
ok deraadt
schwarze [Sat, 20 Aug 2016 14:43:39 +0000 (14:43 +0000)]
If a column list starts with implicit rows (that is, rows without .It)
and roff-level nodes (e.g. tbl or eqn) follow, don't run into an
assertion. Instead, wrap the roff-level nodes in their own row.
Issue found by tb@ with afl(1).
patrick [Sat, 20 Aug 2016 14:35:28 +0000 (14:35 +0000)]
Remove obsolete sxiuart(4) header.
Noticed by mglocker@.
patrick [Sat, 20 Aug 2016 13:24:38 +0000 (13:24 +0000)]
sxiuart(4) has been replaced with com(4).
ok kettenis@
jmc [Sat, 20 Aug 2016 12:54:49 +0000 (12:54 +0000)]
shorten pkcs7 text;
kettenis [Sat, 20 Aug 2016 12:36:59 +0000 (12:36 +0000)]
Correctly enter a mapping as writable if no "page modified" emulation
is needed.
ok visa@, patrick@