nicm [Fri, 4 Oct 2024 19:16:13 +0000 (19:16 +0000)]
Add -y flag to disable confirmation prompts in modes, GitHub issue 4152.
miod [Fri, 4 Oct 2024 19:00:27 +0000 (19:00 +0000)]
Remove unused SIGEMT.
bluhm [Fri, 4 Oct 2024 16:58:26 +0000 (16:58 +0000)]
Increase psp(4) timeouts.
On EPYC 9124 psp(4) seems to need more to time to respond. Otherwise
it will not attach. Increase timeouts for both polling during
autoconf and interrupt during normal operation.
from hshoexer@ OK mlarkin@
nicm [Fri, 4 Oct 2024 14:55:17 +0000 (14:55 +0000)]
Do not translate BSpace as Unicode, GitHub issue 4156.
kevlo [Fri, 4 Oct 2024 07:46:33 +0000 (07:46 +0000)]
Remove two MHI LOOPBACK channels, there's no need to match them.
Also move PCI ops registration ahead. That way the ops will be registered
if any subsequent operations within the case processing require the ops to
See Linux commit
fbb2a14afe00a5691b43cd19c946472e59e16fc6 and
515bcdf587f9911f2d5de51524cb7e048d295052
ok stsp@
nicm [Fri, 4 Oct 2024 07:03:08 +0000 (07:03 +0000)]
Rework of copy mode commands ("send-keys -X") to parse the arguments so
that flags may be detected propertly rather than just looking for
strings ("-O" and so on). Also add -C and -P flags to the copy commands:
-C prevents the commands from sending the text to the clipboard and -P
prevents them from adding the text as a paste buffer.
Note some of the default key bindings change to add "--" and any similar
custom key bindings using "send-keys -X" may need a similar change.
GitHub issue 4153.
jmatthew [Fri, 4 Oct 2024 06:37:22 +0000 (06:37 +0000)]
As with other multiqueue drivers, print the number of queues we set up
along with the interrupt and ethernet address details.
ok dlg@
yasuoka [Fri, 4 Oct 2024 05:22:10 +0000 (05:22 +0000)]
Expose {T,R}X desc trail and add link_enabled field to prepare VF
support. diff from naito.yuichiro at gmail.com
ok jan jmatthew
claudio [Thu, 3 Oct 2024 10:20:05 +0000 (10:20 +0000)]
In rucheck() sum up the tusage from all threads to get the real current
runtime of the process. Also this no longer needs the SCHED_LOCK().
OK mpi@
claudio [Thu, 3 Oct 2024 10:18:29 +0000 (10:18 +0000)]
Fix the clock_gettime() handler for pthread_getcpuclockid() to
use the proper way to read tu_runtime.
OK mpi@
sf [Thu, 3 Oct 2024 08:59:49 +0000 (08:59 +0000)]
vio: Increase rx mbuf size with lro
bluhm found that using bigger rx mbufs helps tcp splice performance if
lro is enabled. Use 4k in that case.
Also fix confusion in rx dmamap segment count. Even with lro/tso, we
only put unfragmented mbufs into the rx queue. Therefore we only need
max. 2 segments, one for the mbuf and one for the separate header for
legacy virtio devices.
OK bluhm@
tb [Thu, 3 Oct 2024 06:24:07 +0000 (06:24 +0000)]
Remove the unused field_mod_func from EC_GROUP
This was only used by the NIST method. For all other group methods it's
an uninitialized pointer (as EC_GROUP_new() still uses the malloc + set
all members to 0 idiom).
ok jsing
nicm [Thu, 3 Oct 2024 05:41:59 +0000 (05:41 +0000)]
Improve fix for shifted keys so it works for all the keys it should,
Stanislav Kljuhhin in GitHub issue 4146.
tb [Thu, 3 Oct 2024 05:07:49 +0000 (05:07 +0000)]
Make EC{,PK}PARAMETERS_it static
They aren't used outside of this file.
tb [Thu, 3 Oct 2024 04:39:09 +0000 (04:39 +0000)]
We've left release mode. Time to re-enable POOL_DEBUG
ok deraadt
tb [Thu, 3 Oct 2024 04:20:28 +0000 (04:20 +0000)]
Fix ASN1_INTEGER_to_BN() misuse
Same issue/leak as for BN_to_ASN1_INTEGER(). Stop reusing the elliptic
curve parameters a and b for order and cofacter. It's confusing.
ok jsing
tb [Thu, 3 Oct 2024 04:17:05 +0000 (04:17 +0000)]
Switch field_bits to be an int
ok jsing
tb [Thu, 3 Oct 2024 04:15:52 +0000 (04:15 +0000)]
Fix BN_to_ASN1_INTEGER() misuse
You can either let this API reuse an existing ASN1_INTEGER or you can let
it allocate a new one. If you try to do both at the same time, you'll leak.
ok jsing
tb [Thu, 3 Oct 2024 03:47:40 +0000 (03:47 +0000)]
Provide OPENSSL_INIT_NO_ATEXIT noop
The brilliant idea of installing a fragile non-idempotent cleanup atexit
handler as a library has bitten many people over time. This gets particularly
exciting when you can't control who dlopens the lib first (don't we all love
Python bindings) or if you are in a threaded context. Fake OpenSSL clones
chose not to do this but now get to carry a noop flag since people start
opting out of this madness (there's a good old tradition at work here).
ok beck joshua jsing millert miod
tb [Thu, 3 Oct 2024 03:31:47 +0000 (03:31 +0000)]
X509V3_EXT_get_nid.3: indicate what nid means
kettenis [Wed, 2 Oct 2024 21:21:32 +0000 (21:21 +0000)]
Improve the heuristic for detecting i2c devices. On the x1e some of the
USB controllers have an I2cSerialBusV2() (possibly defining a connection
to some sort of eUSB2 to USB2 redriver chip) but are not i2c devices
themselves. So check if a device has MMIO resources and assume it isn't
a proper i2c device if it has those. Makes the Type-A ports on my
vivobook work in acpi mode.
ok patrick@, mlarkin@
dv [Wed, 2 Oct 2024 18:18:27 +0000 (18:18 +0000)]
amd64: implement a pmap_shootept() for non-MP kernels.
Simplifies some of the recent INVEPT changes and mirrors the design
of other pmap tlb shootdown functions that provide simplified non-MP
implementations.
ok mlarkin@
millert [Wed, 2 Oct 2024 17:08:47 +0000 (17:08 +0000)]
Update to 2024bgtz from https://github.com/JodaOrg/global-tz
o Improve historical data for Mexico, Mongolia, and Portugal.
o System V names are now obsolescent.
o The main data form now uses %z.
dv [Wed, 2 Oct 2024 17:05:56 +0000 (17:05 +0000)]
Move some PCI MMIO defines from vmm(4) kernel headers to userland.
vmm(4) doesn't need this information anymore. vmd(8) is the only
consumer of this information.
ok mlarkin@
jsing [Wed, 2 Oct 2024 15:21:39 +0000 (15:21 +0000)]
Reorder functions.
Reorder functions so that things are somewhat more logical, moving internal
functions towards the top (and removing now unnecessary prototypes).
jsing [Wed, 2 Oct 2024 14:54:26 +0000 (14:54 +0000)]
Remove err_fns and associated machinery.
Like all good OpenSSL code, errors was built to be completely extensible.
Thankfully, the ERR_{get,set}_implementation() functions were removed in
r1.127 of err.c, which means that the extensibility can no longer be used.
Take the first of many steps to clean up this code - remove err_fns and
associated machinery, calling functions directly. Rename so that we have
an 'err_' prefix rather than 'int_' (or nothing).
ok joshua@ tb@
jsing [Wed, 2 Oct 2024 14:50:58 +0000 (14:50 +0000)]
Hook up the err regress.
jsing [Wed, 2 Oct 2024 14:41:46 +0000 (14:41 +0000)]
Add initial regress for the error stack and ERR_* APIs.
jsing [Wed, 2 Oct 2024 14:12:21 +0000 (14:12 +0000)]
Enable additional CRYPTO_get_ex_new_index() tests.
It was previously possible to call CRYPTO_get_ex_new_index() with either
a negative index or a positive index that equaled or exceeded
CRYPTO_EX_INDEX__COUNT. The reimplementation of exdata treats these as
error cases.
claudio [Wed, 2 Oct 2024 12:31:33 +0000 (12:31 +0000)]
On i386 long double is 80bit expanded to 96bits or 12 bytes. This is the
size that the ctftools ctfconvert uses and I think we should do as well.
Fixes regress/usr.bin/ctfdump on i386.
OK miod@
nicm [Wed, 2 Oct 2024 11:51:15 +0000 (11:51 +0000)]
Add MSYSTEM to default update-environment.
nicm [Wed, 2 Oct 2024 11:48:16 +0000 (11:48 +0000)]
Add a define for the socket permissions check so it can be overridden
more easily (for Cgywin).
mpi [Wed, 2 Oct 2024 10:36:33 +0000 (10:36 +0000)]
Modify uvmpd_scan_inactive() to access `uvmexp.pdfreed' only once.
ok kettenis@
mpi [Wed, 2 Oct 2024 10:24:11 +0000 (10:24 +0000)]
Improve responsiveness in OOM situations & make free target checks coherent.
Remove a change introduced in NetBSD to pageout 4 times as many pages as
required to meet the low water mark of free pages. With todays' Gbs of
RAMs, it makes the pagedaemon hog the CPU for too long when the amount of
free pages is close to none.
ok sthen@, kettenis@
mpi [Wed, 2 Oct 2024 10:17:28 +0000 (10:17 +0000)]
Return number of freed pages in uvm_pmr_cache_drain().
ok kettenis@
mpi [Wed, 2 Oct 2024 10:12:52 +0000 (10:12 +0000)]
Switch PV pool allocator to use km_alloc(9) instead of uvm_km_kmemalloc(9).
Reduce differences between pmaps & allow us to remove another deprecated
allocator.
ok phessler@, miod@
claudio [Wed, 2 Oct 2024 09:45:29 +0000 (09:45 +0000)]
No need to check as->established before calling pfkey_remove().
Remove the extra checks in the caller and simplify some code because of that.
OK tb@
tb [Wed, 2 Oct 2024 09:14:08 +0000 (09:14 +0000)]
Adding pkg_add from stdin hasn't worked in ages. It was something
we inherited from the initial implemention on FreeBSD which has
made no sense in years.
prompted by a diff by Johannes Thyssen Tishman
from espie
nicm [Wed, 2 Oct 2024 08:06:45 +0000 (08:06 +0000)]
Report shifted keys like S-A as A not as S-A in mode 1 extended keys,
from Stanislav Kljuhhin.
tb [Tue, 1 Oct 2024 18:48:29 +0000 (18:48 +0000)]
Extend Log->system to support the same features as BaseState->system
In order to support privsep in tags, we need to be able to pass some
code values in child/parent.
from espie, tested by sthen, ok giovanni
claudio [Tue, 1 Oct 2024 18:33:16 +0000 (18:33 +0000)]
Use auth_conf.method instead of auth_state.method. The latter is always 0.
claudio [Tue, 1 Oct 2024 18:31:10 +0000 (18:31 +0000)]
Explicitly set peer.auth_conf.method, the auth_state value is never
initialized in the SE so bgpctl did always see 0, auth_conf.method
is on the other hand properly shared.
OK tb@
claudio [Tue, 1 Oct 2024 18:29:34 +0000 (18:29 +0000)]
In merge_peers also copy over the auth_conf from the new peer to the old
since that one is kept.
OK tb@
claudio [Tue, 1 Oct 2024 18:28:17 +0000 (18:28 +0000)]
Fix typo in comment
OK tb@
claudio [Tue, 1 Oct 2024 11:50:15 +0000 (11:50 +0000)]
Adjust to the pfkey changes in bgpd.
OK tb@
claudio [Tue, 1 Oct 2024 11:49:24 +0000 (11:49 +0000)]
Rework the pfkey and tcp md5 API to not depend on struct peer.
Instead use struct auth_config and struct auth_state in the pfkey calls
and those tcp_md5 calls where it matters.
This is preparation work to allow RTR to use TCP MD5 as well.
OK tb@
nicm [Tue, 1 Oct 2024 10:10:29 +0000 (10:10 +0000)]
Add a way to make the preview larger in tree mode, GitHub issue 4124.
claudio [Tue, 1 Oct 2024 09:22:25 +0000 (09:22 +0000)]
Make it more obvious which ruadd argument is the source and which one is
the destination by adding const to the source pointer.
OK tb@
claudio [Tue, 1 Oct 2024 08:28:34 +0000 (08:28 +0000)]
Adjust ptrace interface to properly suport single threaded continue.
Introduce P_TRACESINGLE flag to instruct the trapped thread to not
wakeup the other threads (via single_thread_clear). This must be done
like this since ptrace must wake just the single thread to ensure it
runs first and gets the ps_xsig value from ptrace.
Modern gdb depends on this for multi-threaded processes, when a breakpoint
is hit gdb fixes up the trapping instruction and then single steps over
it with only that thread. After that single step gdb continues with all
threads. If all threads are run like now it is possible that one of the
other threads hits a breakpoint before the single step is done which results
in an assertion in gdb (because that is not expected).
OK mpi@
nicm [Tue, 1 Oct 2024 08:01:19 +0000 (08:01 +0000)]
Use global cursor style and colour options for modes instead of default,
GitHub issue 4117.
jsg [Tue, 1 Oct 2024 06:46:19 +0000 (06:46 +0000)]
remove unused variable
jsg [Tue, 1 Oct 2024 06:41:22 +0000 (06:41 +0000)]
drm: Expand max DRM device number to full MINORBITS
From Michal Winiarski
e615cd84dcf834e83b333bfb690fc2032b3fdb85 in linux-6.6.y/6.6.53
071d583e01c88272f6ff216d4f867f8f35e94d7d in mainline linux
jsg [Tue, 1 Oct 2024 06:36:45 +0000 (06:36 +0000)]
accel: Use XArray instead of IDR for minors
From Michal Winiarski
f6b589e361538285fdad8cf62143e3cf3b2c8b2a in linux-6.6.y/6.6.53
45c4d994b82b08f0ce5eb50f8da29379c92a391e in mainline linux
jsg [Tue, 1 Oct 2024 06:30:40 +0000 (06:30 +0000)]
drm: Use XArray instead of IDR for minors
From Michal Winiarski
d2e3d344e20e605378556610811549249298d0aa in linux-6.6.y/6.6.53
5fbca8b48b3050ae7fb611a8b09af60012ed6de1 in mainline linux
jsg [Tue, 1 Oct 2024 06:19:39 +0000 (06:19 +0000)]
add DEFINE_XARRAY_ALLOC for 6.6.53 drm
jsg [Tue, 1 Oct 2024 06:18:20 +0000 (06:18 +0000)]
add MINORBITS for 6.6.53 drm
jsg [Tue, 1 Oct 2024 06:17:46 +0000 (06:17 +0000)]
add ACCEL_MAX_MINORS for 6.6.53 drm
nicm [Tue, 1 Oct 2024 06:15:47 +0000 (06:15 +0000)]
Change pasting to bypass the output key processing entirely and write
what was originally received. Fixes problems with pasted text being
interpreted as extended keys reported by Mark Kelly.
deraadt [Mon, 30 Sep 2024 14:31:56 +0000 (14:31 +0000)]
change release date
claudio [Mon, 30 Sep 2024 12:54:12 +0000 (12:54 +0000)]
Improve some currently impossible error path in log_ext_subtype().
Mainly handle unknown ext-communities better and handle the special
case of type == -1.
OK tb@
claudio [Mon, 30 Sep 2024 12:32:26 +0000 (12:32 +0000)]
Use ps_ppid instead of ps_pptr->ps_pid in all places.
OK mpi@
jsg [Mon, 30 Sep 2024 12:21:17 +0000 (12:21 +0000)]
handle non xa_limit_32b ranges in xarray
jsg [Mon, 30 Sep 2024 12:09:04 +0000 (12:09 +0000)]
store xarray flags for XA_FLAGS_ALLOC1 test in alloc
claudio [Mon, 30 Sep 2024 11:49:44 +0000 (11:49 +0000)]
Remove code after exit1() and NOTREACHED comment. Nothing will ever get there.
OK mpi@
claudio [Mon, 30 Sep 2024 09:42:24 +0000 (09:42 +0000)]
In bgpd_rtr_connect() do the setsockopt dance for IP_TOS and TCP_NODELAY
before calling connect(). This way it happens for sure and on top the TOS
is already set on the initial SYN.
OK tb@
nicm [Mon, 30 Sep 2024 08:10:20 +0000 (08:10 +0000)]
On some Windows terminals, if TIOCWINSZ does not return xpixel and
ypixel (they are zero), if this is the case then try the query escape
sequences. From Dmitry Galchinsky in GitHub issue 4099.
mpi [Mon, 30 Sep 2024 08:09:39 +0000 (08:09 +0000)]
Return the number of freed pages and handle SHRINK_STOP in drmbackoff().
ok jsg@
nicm [Mon, 30 Sep 2024 07:54:51 +0000 (07:54 +0000)]
Only use default-shell for popups, return to /bin/sh for run-shell,
if-shell and #() - these have been documented as using /bin/sh for a
long time and scripts rely on it. Pointed out by Gregory Pakosz.
jsg [Mon, 30 Sep 2024 03:55:46 +0000 (03:55 +0000)]
init xarray pool in drm_linux_init(), needed for static initialisation
jsg [Mon, 30 Sep 2024 01:41:49 +0000 (01:41 +0000)]
remove WSDISPLAYIO_OGINFO compat ioctl, planned for after 7.3
ok miod@
nicm [Sun, 29 Sep 2024 20:05:42 +0000 (20:05 +0000)]
Fix grey colour, from Magnus Gross.
jmc [Sun, 29 Sep 2024 18:00:38 +0000 (18:00 +0000)]
add some commas, to aid readability;
from thorsten blum
kn [Sun, 29 Sep 2024 14:36:13 +0000 (14:36 +0000)]
sync synopsis and usage, sort commands, fix their spacing
OK input lucas
jsg [Sun, 29 Sep 2024 12:22:57 +0000 (12:22 +0000)]
remove unused cruft; ok kettenis@
jsg [Sun, 29 Sep 2024 11:22:39 +0000 (11:22 +0000)]
fix HWCAP_ASIMDHP test; ok kettenis@
jsg [Sun, 29 Sep 2024 09:25:37 +0000 (09:25 +0000)]
correct a prototype
sthen [Sat, 28 Sep 2024 17:09:52 +0000 (17:09 +0000)]
catch up with the u-boot-aarch64 package split in install docs. changing
the dtb isn't normally needed (and can cause problems) so don't suggest that.
add some info for rk3588. partly from jsg.
tb [Fri, 27 Sep 2024 13:13:14 +0000 (13:13 +0000)]
rsync: make blkhash_free() callable with NULL.
It is in principle possible that blkhash_free() (which obviously frees
a struct blktab, sigh) is called with NULL. This would lead to a crash.
ok claudio
tb [Fri, 27 Sep 2024 13:10:39 +0000 (13:10 +0000)]
rsync: avoid reallocarray() with a nmemb of 0.
bset->blksz can be 0 here and how reallocarray() behaves is implementation
defined.
tb [Fri, 27 Sep 2024 13:06:21 +0000 (13:06 +0000)]
rsync: fix reallocarray() usage in blkhash_set()
The well-named ERR() macro doesn't error out. Therefore an incorrect use
of reallocarray() is actually a leak that is easily overlooked. Do it the
right way by assigning to a temporary variable and preserve behavior by
freeing and NULL-ing.
ok claudio
tb [Fri, 27 Sep 2024 12:55:03 +0000 (12:55 +0000)]
rpki-client: free deltas in rrdp_session_parse() as in rrdp_repo_free()
ok claudio
tb [Fri, 27 Sep 2024 12:52:58 +0000 (12:52 +0000)]
rpki-client: check for getline error when reading the skiplist file
ok claudio
djm [Fri, 27 Sep 2024 01:05:54 +0000 (01:05 +0000)]
test some more Match syntax, including criteria=arg and negations
yasuoka [Fri, 27 Sep 2024 00:38:49 +0000 (00:38 +0000)]
Previous pipex.c,v 1.155 was broken if the client was not behind a NAT.
ok mvs
djm [Thu, 26 Sep 2024 23:55:08 +0000 (23:55 +0000)]
fix previous change to ssh_config Match, which broken on negated
Matches; spotted by phessler@ ok deraadt@
dv [Thu, 26 Sep 2024 21:55:42 +0000 (21:55 +0000)]
Fix amd64/GENERIC not linking due to missing pmap_shootept symbol.
My recent commit to add an INVEPT ipi missed an ifdef related to
MP machines.
ok krw@
dv [Thu, 26 Sep 2024 13:18:25 +0000 (13:18 +0000)]
Add an ipi for executing INVEPT to flush EPT on remote cpus.
Similar to how the fast ipi for tlb flush is implemented, this adds
one for calling INVEPT to invalidate EPT caches on the cpu. This
is the first step to allowing guest memory to not be wired by UVM
and decreases the behavioral differences between Intel and AMD's
nested paging in vmm(4) and pmap(9).
This change does not hook EPT ptes into the PV list, so the ipi is
only used during address space teardown and pte removal. (With the
removal of the "mprotect" ioctl, vmm(4) no longer modifies EPT ptes
other than inserting them and removing them.)
ok mlarkin@
jsg [Thu, 26 Sep 2024 10:12:02 +0000 (10:12 +0000)]
remove unneeded psl.h include
claudio [Thu, 26 Sep 2024 08:12:30 +0000 (08:12 +0000)]
Remove debug leftover. There is no need to run tcpdump anymore.
jmc [Thu, 26 Sep 2024 06:10:34 +0000 (06:10 +0000)]
document that there are nvme sensors available and nudge the reader to sysctl;
ok bmercer
jsg [Thu, 26 Sep 2024 01:45:13 +0000 (01:45 +0000)]
remove unneeded includes; ok dv@ mlarkin@
jsg [Wed, 25 Sep 2024 23:01:39 +0000 (23:01 +0000)]
remove some unused defines; ok djm@
bru [Wed, 25 Sep 2024 19:56:33 +0000 (19:56 +0000)]
Correct an indexing error that may leave stale data in the mousecfg buffer.
ok mpi@
bluhm [Wed, 25 Sep 2024 18:24:13 +0000 (18:24 +0000)]
Do not crash in witness code if interrupt fires early.
It can happen that an interrupt handler is called immediately on a
secondary CPU when it is spun up during boot. At that moment no
process is schduled yet, so curproc is NULL. To prevent a crash
when p->p_sleeplocks is dereferenced, skip witness check in this
case.
OK visa@
florian [Wed, 25 Sep 2024 16:26:37 +0000 (16:26 +0000)]
Improve description of -f.
ok sthen, florian
Committing on behalf of jmc as requested.
claudio [Wed, 25 Sep 2024 15:38:39 +0000 (15:38 +0000)]
Extend maxattr test to also include IPv6 and with that test for the
up_generate_mp_reach() failure.
jmc [Wed, 25 Sep 2024 15:08:42 +0000 (15:08 +0000)]
correct macro for -R entry;
claudio [Wed, 25 Sep 2024 14:46:51 +0000 (14:46 +0000)]
When generating UPDATE handle the message size limit better.
First of all warn that a prefix was dropped. In the generate an update
code handle possible overflows of attributes and NLRI and withdraw the
affected prefix. This way the peer will not have stale data.
OK tb@
claudio [Wed, 25 Sep 2024 14:42:39 +0000 (14:42 +0000)]
Add a maxattr regress test that causes the rdomain1 bgpd instance to trip
over 4096 bytes for a BGP UPDATE message. Check on the 3rd system that the
affected prefixes are properly withdrawn.
sthen [Wed, 25 Sep 2024 13:55:23 +0000 (13:55 +0000)]
sysupgrade: add -R #.# to try to use a specific release version rather
than the immediate +0.1. print an https://ftp.openbsd.org/... URL where
the new signify pubkey can be found if not present.
no guarantees: we only test +0.1, but jumping further does work quite
often (and if tight on disk, can work better than multiple steps) -
this avoids editing the script if you're going to do it anyway.
"Only upgrades from one version to the next are tested. Skipping
versions may work. Downgrading is unlikely to work."
discussed with deraadt chris florian, ok deraadt