openbsd
18 months agoDo not rewrite MAN_LP and MAN_P to MAN_PP because doing that causes
schwarze [Fri, 28 Apr 2023 20:14:19 +0000 (20:14 +0000)]
Do not rewrite MAN_LP and MAN_P to MAN_PP because doing that causes
confusing warning messages complaining about macros that don't even
appear in the input file.
As a welcome side effect, this also shortens the code...

Fixing a minibug
reported by Alejandro Colomar <alx dot manpages at gmail dot com>.

18 months agoClarify -b usage by `vmctl start`.
dv [Fri, 28 Apr 2023 20:13:56 +0000 (20:13 +0000)]
Clarify -b usage by `vmctl start`.

18 months agoAdd rtentry refcnt type to dt(4).
mvs [Fri, 28 Apr 2023 20:03:13 +0000 (20:03 +0000)]
Add rtentry refcnt type to dt(4).

ok bluhm@

18 months agovmd(8)/vmctl(8): allow vm owners to override boot kernel.
dv [Fri, 28 Apr 2023 19:46:41 +0000 (19:46 +0000)]
vmd(8)/vmctl(8): allow vm owners to override boot kernel.

vmd allows non-root users to "own" a vm defined in vm.conf(5). While
the user can start/stop the vm, if they break their filesystem they
have no means of booting recovery media like a ramdisk kernel.

This change opens the provided boot kernel via vmctl and passes the
file descriptor through the control channel to vmd. The next boot
of the vm will use the provided file descriptor as boot kernel/bios.
Subsequent boots (e.g. a reboot) will return to using behavior
defined in vm.conf or the default bios image.

ok mlarkin@

18 months agoExecute each test as make target. Remove the shell wrapper. Mark
bluhm [Fri, 28 Apr 2023 19:41:07 +0000 (19:41 +0000)]
Execute each test as make target.  Remove the shell wrapper.  Mark
failing test so that claudio@ can fix them.

18 months agoRemove unneeded header includes in vmd.
dv [Fri, 28 Apr 2023 18:52:22 +0000 (18:52 +0000)]
Remove unneeded header includes in vmd.

No functional change. virtio block/networking emulation do not need
to know about vmm or any kernel types.

18 months agobump MAXDSIZ to 128G on amd64 and 64G on arm64
robert [Fri, 28 Apr 2023 18:33:22 +0000 (18:33 +0000)]
bump MAXDSIZ to 128G on amd64 and 64G on arm64
discussed with kettenis@, ok deraadt@

18 months agoFree all libcrypto global state memory before returning
job [Fri, 28 Apr 2023 18:32:40 +0000 (18:32 +0000)]
Free all libcrypto global state memory before returning

Found with the help of Otto's malloc memory leak detector!

18 months agoReturn a non-zero error exit code on any DER cache discrepancies
job [Fri, 28 Apr 2023 18:31:34 +0000 (18:31 +0000)]
Return a non-zero error exit code on any DER cache discrepancies

18 months agotimer(4/sparc64): remove driver
cheloha [Fri, 28 Apr 2023 18:27:55 +0000 (18:27 +0000)]
timer(4/sparc64): remove driver

The timer(4/sparc64) driver was effectively disabled during the
previous release.  Nobody has come forward asking for it to be adapted
to work with the new clockintr framework, so it's time to remove the
driver from the tree.

As of today, if you want to run OpenBSD on SPARC v9 hardware, that
hardware needs to sport either %tick and %tick_compare (%asr23), or
%stick (%asr24) and %stick_compare (%asr25).

All Sun/Oracle SPARC v9 hardware meets these conditions, from the
UltraSPARC I onward.

Most HAL/Fujitsu SPARC v9 hardware meets these conditions, from the
SPARC64 III onward.  The only HAL/Fujitsu hardware that might not have
%tick_compare are the HAL SPARC64 I and SPARC64 II, for which I can
find no documentation.  However, those processors are currently
unsupported by OpenBSD for other reasons, so their support status is
unchanged by the removal of this driver.

With help from miod@.

Link: https://marc.info/?l=openbsd-tech&m=167898759928206&w=2
"after unlock" deraadt@, ok mlarkin@ miod@

18 months agoFix leaks reported by ASAN
tb [Fri, 28 Apr 2023 18:27:49 +0000 (18:27 +0000)]
Fix leaks reported by ASAN

debugged with job

18 months agoRevert amd64/i386 floppy change. Missing diff to vnconfig broke
krw [Fri, 28 Apr 2023 18:14:59 +0000 (18:14 +0000)]
Revert amd64/i386 floppy change. Missing diff to vnconfig broke
installboot'ing due to incorrect d_type (must be 'floppy' not
'vnd') in disklabel.

Noticed by deraadt@ and sthen@

18 months agoToo many stupid things whine about these being used uninitialized
tb [Fri, 28 Apr 2023 18:14:59 +0000 (18:14 +0000)]
Too many stupid things whine about these being used uninitialized
(which they aren't), so appease them.

18 months agoRemove preservation and use of cached DER/BER encodings in the d2i/i2d paths
job [Fri, 28 Apr 2023 17:59:53 +0000 (17:59 +0000)]
Remove preservation and use of cached DER/BER encodings in the d2i/i2d paths

A long time ago a workflow was envisioned for X509, X509_CRL, and X509_REQ
structures in which only fields modified after deserialization would need to
be re-encoded upon serialization.

Unfortunately, over the years, authors would sometimes forget to add code in
setter functions to trigger invalidation of previously cached DER encodings.

The presence of stale versions of structures can lead to very hard-to-debug
issues and cause immense sorrow.

Fully removing the concept of caching DER encodings ensures stale versions
of structures can never rear their ugly heads again.

OK tb@ jsing@

18 months agoSome wording tweaks to finish the polishing.
schwarze [Fri, 28 Apr 2023 17:31:58 +0000 (17:31 +0000)]
Some wording tweaks to finish the polishing.
While here, also correct the HISTORY section.
OK jmc@

18 months agoMark the obsolete PROXY_PARAM and SOCKS BIO_ctrl(3) command constants
schwarze [Fri, 28 Apr 2023 16:59:03 +0000 (16:59 +0000)]
Mark the obsolete PROXY_PARAM and SOCKS BIO_ctrl(3) command constants
as intentionally undocumented.  Do that here because no related
manual pages exist.

18 months agoEnable policy checking by default now that we are DAG implementation based.
beck [Fri, 28 Apr 2023 16:50:16 +0000 (16:50 +0000)]
Enable policy checking by default now that we are DAG implementation based.

This ensures that we will no longer silently ignore a certificate with
a critical policy extention by default.

ok tb@

18 months agoMark a number of BIO_ctrl(3) command constants as intentionally
schwarze [Fri, 28 Apr 2023 16:49:00 +0000 (16:49 +0000)]
Mark a number of BIO_ctrl(3) command constants as intentionally
undocumented because they are NOOPs or deprecated.

18 months agokill the .Xr to BN_nist_mod_521(3) which no longer exists
schwarze [Fri, 28 Apr 2023 16:39:19 +0000 (16:39 +0000)]
kill the .Xr to BN_nist_mod_521(3) which no longer exists

18 months agoUnifdef LIBRESSL_HAS_POLICY_DAG and remove it from the Makefile
tb [Fri, 28 Apr 2023 16:30:14 +0000 (16:30 +0000)]
Unifdef LIBRESSL_HAS_POLICY_DAG and remove it from the Makefile

with beck

18 months agoImport rsync regress provided by Martin Cracauer so that bluhm@ can work
claudio [Fri, 28 Apr 2023 16:28:28 +0000 (16:28 +0000)]
Import rsync regress provided by Martin Cracauer so that bluhm@ can work
improve it in tree.

18 months agoAdd BIO_C_SET_MD_CTX to the list of command constants.
schwarze [Fri, 28 Apr 2023 16:22:45 +0000 (16:22 +0000)]
Add BIO_C_SET_MD_CTX to the list of command constants.

18 months agoTake the old policy code behind the barn
tb [Fri, 28 Apr 2023 16:21:57 +0000 (16:21 +0000)]
Take the old policy code behind the barn

It can go play in the fields with all the other exponential time policy
"code".

discussed with jsing
ok & commit message beck

18 months agoDocument BIO_set_md_ctx(3) and BIO_C_SET_MD_CTX.
schwarze [Fri, 28 Apr 2023 16:20:01 +0000 (16:20 +0000)]
Document BIO_set_md_ctx(3) and BIO_C_SET_MD_CTX.
Correct the return types of some macros.
Improve the RETURN VALUES section.

18 months agoThe policy test is no longer expected to fail
tb [Fri, 28 Apr 2023 16:18:17 +0000 (16:18 +0000)]
The policy test is no longer expected to fail

18 months agoEnable the new policy checking code in x509_policy.c
tb [Fri, 28 Apr 2023 16:14:46 +0000 (16:14 +0000)]
Enable the new policy checking code in x509_policy.c

ok beck jsing

18 months agoSilence gcc-4 warnings about sk_sort()
tb [Fri, 28 Apr 2023 15:57:38 +0000 (15:57 +0000)]
Silence gcc-4 warnings about sk_sort()

Tell it we deliberately ignore the return value, (we really don't
care what the old comparison function was).

18 months agoRemove misinformation, reason had nothing to do with efficiency
job [Fri, 28 Apr 2023 15:51:18 +0000 (15:51 +0000)]
Remove misinformation, reason had nothing to do with efficiency

"Failure to re-encode on modification is a bug not a feature."

OK jsing@

18 months agoremove superfluous/invalid KASSERT() in pfsync_q_del().
sashan [Fri, 28 Apr 2023 15:50:05 +0000 (15:50 +0000)]
remove superfluous/invalid KASSERT() in pfsync_q_del().

pointed and OK bluhm@

18 months agoRemove now no longer needed <assert.h>; sort headers
tb [Fri, 28 Apr 2023 15:39:29 +0000 (15:39 +0000)]
Remove now no longer needed <assert.h>; sort headers

ok jsing

18 months agoDeassert has_explicit_policy()
tb [Fri, 28 Apr 2023 15:37:28 +0000 (15:37 +0000)]
Deassert has_explicit_policy()

The only caller is X509_policy_check() which goes straight to error.

with beck
ok jsing

18 months agoDeassert delete_if() callbacks
tb [Fri, 28 Apr 2023 15:35:55 +0000 (15:35 +0000)]
Deassert delete_if() callbacks

Add sk_is_sorted() checks to the callers of sk_X509_POLICY_NODE_delete_if()
and add a comment that this is necessary.

with beck
ok jsing

18 months agoDeassert x509_policy_level_find()
tb [Fri, 28 Apr 2023 15:30:14 +0000 (15:30 +0000)]
Deassert x509_policy_level_find()

Move the check that level->nodes is sorted to the call site and make sure
that the logic is preserved and erroring does the right thing.

with beck
ok jsing

18 months agoDeassert X509_policy_check()
tb [Fri, 28 Apr 2023 15:27:15 +0000 (15:27 +0000)]
Deassert X509_policy_check()

Instead of asserting that i == num_certs - 2, simply make that an error
check.

with beck
ok jsing

18 months agoDeassert x509_policy_level_add_nodes()
tb [Fri, 28 Apr 2023 15:21:22 +0000 (15:21 +0000)]
Deassert x509_policy_level_add_nodes()

This assert is in debugging code that ensures that there are no duplicate
nodes on this level. This is an expensive and unnecessary check. Duplicates
already cause failures as ensured by regress.

with beck
ok jsing

18 months agoDeassert x509_policy_new()
tb [Fri, 28 Apr 2023 15:16:48 +0000 (15:16 +0000)]
Deassert x509_policy_new()

Turn the check into an error which will make all callers error.

with beck
ok jsing

18 months agoRearrange freeing of memory in the regress test
job [Fri, 28 Apr 2023 15:12:51 +0000 (15:12 +0000)]
Rearrange freeing of memory in the regress test

18 months agoReorder the text such that every function is discussed only once
schwarze [Fri, 28 Apr 2023 15:04:33 +0000 (15:04 +0000)]
Reorder the text such that every function is discussed only once
instead of discussing some of them at two different places.
Also follow a more logical order: initialization first, then reading
and writing, then retrieving the digest and reinitialization.
Leave context handling and chain duplication at the end because
both are rarely needed.

While here, also tweak the wording of the shuffled text
and add some precision in a few places.

18 months agomake the policy test compile on sparc64
tb [Fri, 28 Apr 2023 14:45:51 +0000 (14:45 +0000)]
make the policy test compile on sparc64

18 months agoInbound portion of RFC9131. Routers can create new neighbor cache entries
phessler [Fri, 28 Apr 2023 14:09:06 +0000 (14:09 +0000)]
Inbound portion of RFC9131.  Routers can create new neighbor cache entries
when receiving a valid Neighbor Advertisement.

OK florian@ kn@

18 months agoThis change speeds up DIOCGETRULE ioctl(2) which pfctl(8) uses to
sashan [Fri, 28 Apr 2023 14:08:38 +0000 (14:08 +0000)]
This change speeds up DIOCGETRULE ioctl(2) which pfctl(8) uses to
retrieve rules from kernel. The current implementation requires
like O((n^2)/2) operation to read the complete rule set, because
each DIOCGETRULE operation must iterate over previous n
rules to find (n + 1)-th rule to read.

To address the issue diff introduces a pf_trans structure to keep
pointer to next rule to read, thus  reading process does not need
to iterate from beginning of rule set to reach the next rule.
All transactions opened by process get closed either when process
is done (reads all rules) or when /dev/pf device is closed.

the diff also comes with lots of improvements from dlg@ and kn@

OK dlg@, kn@

18 months agoRelax the "pass all" rule so all forms of neighbor advertisements are allowed
phessler [Fri, 28 Apr 2023 14:08:34 +0000 (14:08 +0000)]
Relax the "pass all" rule so all forms of neighbor advertisements are allowed
in either direction.

This more closely matches the IPv4 ARP behaviour.

From sashan@
discussed with kn@ deraadt@

18 months agoAdd X509_REQ_add_extensions and to X509_REQ_add1_attr to DER cache test
job [Fri, 28 Apr 2023 13:48:38 +0000 (13:48 +0000)]
Add X509_REQ_add_extensions and to X509_REQ_add1_attr to DER cache test

These new tests won't bubble up a non-zero error exit code because
other libcrypto bits still need to land first.

18 months agoSame change as in bgpd:
claudio [Fri, 28 Apr 2023 13:24:25 +0000 (13:24 +0000)]
Same change as in bgpd:
Add explicit default labels in switch() statements with error handling.
Right now these are not reachable. Should also clear some gcc warnings.
OK tb@

18 months agoAdd explicit default labels in switch() statements with error handling.
claudio [Fri, 28 Apr 2023 13:23:52 +0000 (13:23 +0000)]
Add explicit default labels in switch() statements with error handling.
Right now these are not reachable. Should also clear some gcc warnings.
OK tb@

18 months agoAdd a membar_consumer() for the taskq_create() in sosplice(). Membar
bluhm [Fri, 28 Apr 2023 12:53:42 +0000 (12:53 +0000)]
Add a membar_consumer() for the taskq_create() in sosplice().  Membar
producer and consumer must come in pair and the latter was missing.
Also move the code a bit to make clear which check is needed for
what.
OK mvs@

18 months agoMove FSDISKTYPE uses from disklabel(8) invocations to vnconfig(8)
krw [Fri, 28 Apr 2023 12:26:43 +0000 (12:26 +0000)]
Move FSDISKTYPE uses from disklabel(8) invocations to vnconfig(8)
invocations, making the geometry information written to the
disklabel a bit more logically related to the disktab information
from whence it came. Also makes FSDISKTYPE usage consistent.

Flip the disklabel(8) invocations to the "echo '/ *'"
idiom to make it obvious that the desire is to create a single
'a' partition containing all free space.

No intentional functional change. MBRs, disklabels and newfs
outputs appear identical.

reads good to kn@

18 months agoEnable kernel-address sanitizer for clang openbsd target
gnezdo [Fri, 28 Apr 2023 12:03:49 +0000 (12:03 +0000)]
Enable kernel-address sanitizer for clang openbsd target

OK deraadt@

18 months agoImplement --size-only and --ignore-times
claudio [Fri, 28 Apr 2023 10:24:38 +0000 (10:24 +0000)]
Implement --size-only and --ignore-times

Flags are passed to the remote system but --size-only is only set
if local system is sender since this is the behaviour of rsync.
Initial diff from Martin Cracauer but mostly reimplemented and extended
by myself.
OK kn@

18 months agoRemove net lock from DIOCGETQUEUES
kn [Fri, 28 Apr 2023 10:19:35 +0000 (10:19 +0000)]
Remove net lock from DIOCGETQUEUES

Both ticket and number of queues stem from the pf_queues_active list which
is effectively static to pf_ioctl.c and fully protected by the pf lock.

OK sashan

18 months agoFix memory constraints in the inline-assembly stub that calls into secure
patrick [Fri, 28 Apr 2023 10:19:07 +0000 (10:19 +0000)]
Fix memory constraints in the inline-assembly stub that calls into secure
mode.  Without this change the compiler doesn't realize that the memory
behind the array that contains the return values might have changed and
optimizes the access away.  With this change it properly access the array
to retrieve the returned values.

ok drahn@

18 months agoRemove error handling around mallocarray(9). I cannot fail when
bluhm [Fri, 28 Apr 2023 10:18:57 +0000 (10:18 +0000)]
Remove error handling around mallocarray(9).  I cannot fail when
called with M_WAITOK.
OK kevlo@

18 months agofix lfindent (newline-and-indent) comment and description in the man page
op [Fri, 28 Apr 2023 10:02:03 +0000 (10:02 +0000)]
fix lfindent (newline-and-indent) comment and description in the man page

for a while it has used only spaces when no-tab-mode is enabled and respected
the current buffer tab width.

18 months agoCleanup pass over x509_check_policy.c
tb [Fri, 28 Apr 2023 09:56:09 +0000 (09:56 +0000)]
Cleanup pass over x509_check_policy.c

This hoists variable declarations to the top and compiles with -Wshadow.

ok beck

18 months agomark up all commands in the man page
op [Fri, 28 Apr 2023 09:50:50 +0000 (09:50 +0000)]
mark up all commands in the man page

18 months agoHook up the the x509 policy regression tests to x509 regress.
beck [Fri, 28 Apr 2023 09:11:35 +0000 (09:11 +0000)]
Hook up the the x509 policy regression tests to x509 regress.

These were adapted from BoringSSL's regress tests for x509
policy. They are currently marked as expected to fail as
we have not enabled LIBRESSL_HAS_POLICY_DAG by default yet, and
the old tree based policy code from OpenSSL is special.

These tests pass when we build with LIBRESSL_HAS_POLICY_DAG.

18 months agoFix copyright, convert boringssl comments to C style
beck [Fri, 28 Apr 2023 09:02:04 +0000 (09:02 +0000)]
Fix copyright, convert boringssl comments to C style

18 months agoKNF
beck [Fri, 28 Apr 2023 08:53:20 +0000 (08:53 +0000)]
KNF
ok knfmt

18 months agoremove unused code.
beck [Fri, 28 Apr 2023 08:50:08 +0000 (08:50 +0000)]
remove unused code.

18 months agoremove debugging printf
beck [Fri, 28 Apr 2023 08:45:50 +0000 (08:45 +0000)]
remove debugging printf

18 months agoRetire -E's "expert" mode. Introduced 23 years ago to avoid
krw [Fri, 28 Apr 2023 08:45:24 +0000 (08:45 +0000)]
Retire -E's "expert" mode. Introduced 23 years ago to avoid
confusing users with FFS attributes that only experts should
fiddle with. Actual use has withered away with functionality
rendered moot or moved elsewhere.

'-e' remains for the truly obscure corner cases.

Simply excise the code for now to see if hidden users/uses are
exposed.  Further simplifications are possible if no such
users/uses surface.

ok with sthen@ millert@ kn@ otto@

18 months agoThis test should not have V_EXPLICIT_POLICY set. with this
beck [Fri, 28 Apr 2023 08:43:18 +0000 (08:43 +0000)]
This test should not have V_EXPLICIT_POLICY set.  with this
corrected we pass

18 months agoAdd the rest of the boringssl policy unit tests.
beck [Fri, 28 Apr 2023 08:15:11 +0000 (08:15 +0000)]
Add the rest of the boringssl policy unit tests.

We currently still fail two of these, looks like one more bug in
extracting the depth for require policy from the certificate..

18 months agoescape the "D1" revisions, as mandoc thinks this is a macro call;
jmc [Fri, 28 Apr 2023 06:21:42 +0000 (06:21 +0000)]
escape the "D1" revisions, as mandoc thinks this is a macro call;

18 months agoAdd options to change the confirm key and default behaviour of
nicm [Fri, 28 Apr 2023 06:12:27 +0000 (06:12 +0000)]
Add options to change the confirm key and default behaviour of
confirm-before. From Elias Assaf in GitHub issue 3548; prompted by an
earlier change from Yutaro Yoshii in GitHub issue 3496.

18 months agoDo not fatal if tparm fails, instead just log it (not working sequences
nicm [Fri, 28 Apr 2023 05:59:35 +0000 (05:59 +0000)]
Do not fatal if tparm fails, instead just log it (not working sequences
are better than exiting).

18 months agoAdd a driver for the Qualcomm rng device found on the Thinkpad X13s, based on
phessler [Fri, 28 Apr 2023 05:13:37 +0000 (05:13 +0000)]
Add a driver for the Qualcomm rng device found on the Thinkpad X13s, based on
kettenis's amlrng driver.

suggestions and OK patrick@

18 months agosync
kevlo [Fri, 28 Apr 2023 01:25:51 +0000 (01:25 +0000)]
sync

18 months agoregen
kevlo [Fri, 28 Apr 2023 01:24:51 +0000 (01:24 +0000)]
regen

18 months agoAdd support for RTL8188FTV chip to urtwn(4).
kevlo [Fri, 28 Apr 2023 01:24:14 +0000 (01:24 +0000)]
Add support for RTL8188FTV chip to urtwn(4).

Tested with Comfast CF-WU710N v4.

"go ahead" deraadt@
OK stsp@

18 months agoUse __size_t which is available in syslog.h
gnezdo [Thu, 27 Apr 2023 23:16:18 +0000 (23:16 +0000)]
Use __size_t which is available in syslog.h

Directly including sys/syslog.h would fail due to size_t
being unknown.

OK millert, miod

18 months agovmd(8): introduce multi-process model for virtio devices.
dv [Thu, 27 Apr 2023 22:47:27 +0000 (22:47 +0000)]
vmd(8): introduce multi-process model for virtio devices.

Isolate virtio network and block device emulation in dedicated
processes, forked and exec'd from the vm process. This allows for
tightening pledge promises to just "stdio".

Communication between the vcpu's and these devices now occurs via
imsg channels, which adds the benefit of not always blocking the
vcpu thread while emulating the device.

With this commit, it's possible that vmd is the first open source
hypervisor that *defaults* to a multi-process device emulation
model without requiring any additional configuration from the
operator.

Testing help from phessler@ and Mischa Peters.

ok mlarkin@

18 months agoadd missing entries; ok miod
jmc [Thu, 27 Apr 2023 20:35:16 +0000 (20:35 +0000)]
add missing entries; ok miod

18 months agoadd missing entries; ok miod
jmc [Thu, 27 Apr 2023 19:26:06 +0000 (19:26 +0000)]
add missing entries; ok miod

18 months agoadd missing entries;
jmc [Thu, 27 Apr 2023 19:10:53 +0000 (19:10 +0000)]
add missing entries;
remove Tn macro usage;

feedback/ok miod

18 months agoMark the fan speed sensor as invalid when it reports -1 RPM. This happens for
miod [Thu, 27 Apr 2023 19:06:57 +0000 (19:06 +0000)]
Mark the fan speed sensor as invalid when it reports -1 RPM. This happens for
a short while after suspend.
ok deraadt@ kn@

18 months agoBetter pcic(4) description.
miod [Thu, 27 Apr 2023 19:01:01 +0000 (19:01 +0000)]
Better pcic(4) description.

18 months agoadd support for I/O statistics so that tape speeds can be observed with
robert [Thu, 27 Apr 2023 18:21:44 +0000 (18:21 +0000)]
add support for I/O statistics so that tape speeds can be observed with
iostat(8)

18 months agosort options;
jmc [Thu, 27 Apr 2023 17:18:40 +0000 (17:18 +0000)]
sort options;

18 months agoarm64 install.md: fix softraid crypto installation on Mac
caspar [Thu, 27 Apr 2023 17:04:17 +0000 (17:04 +0000)]
arm64 install.md: fix softraid crypto installation on Mac

Make sure we don't newfs the EFI Sys partition on systems that have an
"apfsisc" partition in the case we're installing with softraid crypto.

Debugged with help from and came up with a fix with kn@

"go ahead" kettenis@
"no objections" krw@
OK kn@

18 months agoRFC 9096 changes the default timers for prefix preferred and valid lifetimes,
phessler [Thu, 27 Apr 2023 16:56:52 +0000 (16:56 +0000)]
RFC 9096 changes the default timers for prefix preferred and valid lifetimes,
so update rad(8) to the new ones.

OK florian@

18 months agotiny wording tweak from Ted Bullock to make misunderstandings less likely;
schwarze [Thu, 27 Apr 2023 16:48:53 +0000 (16:48 +0000)]
tiny wording tweak from Ted Bullock to make misunderstandings less likely;
OK jmc@

18 months agoImplement -V as an alias to --version.
claudio [Thu, 27 Apr 2023 16:28:18 +0000 (16:28 +0000)]
Implement -V as an alias to --version.
From Martin Cracauer
OK kn@

18 months agoConvert size_t's used in conjuction with sk_X509_num back to int.
beck [Thu, 27 Apr 2023 16:12:08 +0000 (16:12 +0000)]
Convert size_t's used in conjuction with sk_X509_num back to int.

The lets the regress in x509/policy pass instead of infinite looping.

The changes are necessry  because our sk_num() returns an int with
0 for empty and -1 for NULL, wheras BoringSSL's returns a size_t with
0 for both an empty stack and a NULL stack.

pair work with tb@
ok tb@ jsing@

18 months agovarious markup tweaks, no content change; OK jmc@
schwarze [Thu, 27 Apr 2023 16:10:11 +0000 (16:10 +0000)]
various markup tweaks, no content change; OK jmc@

18 months agovarious minor content corrections and improvements;
schwarze [Thu, 27 Apr 2023 15:44:36 +0000 (15:44 +0000)]
various minor content corrections and improvements;
feedback and OK jmc@ and Ted Bullock

18 months agoRemove efi32 and efi64. These are leftovers from a project that didn't go
kettenis [Thu, 27 Apr 2023 15:06:35 +0000 (15:06 +0000)]
Remove efi32 and efi64.  These are leftovers from a project that didn't go
very far and the presence of these directories just confuse people and
make them do more work than necessary.

ok patrick@, kn@, mlarkin@

18 months agoAlso list the command constants not associated with any macros,
schwarze [Thu, 27 Apr 2023 14:44:33 +0000 (14:44 +0000)]
Also list the command constants not associated with any macros,
and point to their documentation.

18 months agoRemove kernel lock from rtfree(9).
mvs [Thu, 27 Apr 2023 14:41:09 +0000 (14:41 +0000)]
Remove kernel lock from rtfree(9).

Route timers and route labels protected by corresponding mutexes. `ifa'
uses references counting for protection. rt_mpls_clear() could be called
lockless because this is the last reference of `rt'.

ok bluhm@ kn@

18 months agoRetire -E's "expert" mode. Introduced 23 years ago to avoid
krw [Thu, 27 Apr 2023 14:19:28 +0000 (14:19 +0000)]
Retire -E's "expert" mode. Introduced 23 years ago to avoid
confusing users with FFS attributes that only experts should
fiddle with. Actual use has withered away with functionality
rendered moot or moved elsewhere.

'-e' remains for the truly obscure corner cases.

Simply excise the code for now to see if hidden users/uses are
exposed.  Further simplifications are possible if no such
users/uses surface.

ok with sthen@ millert@ kn@ otto@

18 months agoTemporarily workaround double calls into vioscsi_req_done()
krw [Thu, 27 Apr 2023 13:52:58 +0000 (13:52 +0000)]
Temporarily workaround double calls into vioscsi_req_done()
causing NULL de-reference.

Reported, initial patch and tests by Antun Matanovic. Thanks!

ok miod@

18 months agocorrect test cases to add expected errors.
beck [Thu, 27 Apr 2023 13:26:57 +0000 (13:26 +0000)]
correct test cases to add expected errors.

18 months agorevert cache lookup for full pathnames
robert [Thu, 27 Apr 2023 12:27:56 +0000 (12:27 +0000)]
revert cache lookup for full pathnames

18 months agoStart of an x509 policy regress test. test cases from BoringSSL.
beck [Thu, 27 Apr 2023 12:23:31 +0000 (12:23 +0000)]
Start of an x509 policy regress test. test cases from BoringSSL.

Still a work in progress adapting tests from boringssl x509_test.cc
but dropping in here for tb to be able to look at and run as well
since the new stuff still has bugs.

18 months agoRemove net lock from DIOCGETTIMEOUT
kn [Thu, 27 Apr 2023 12:10:30 +0000 (12:10 +0000)]
Remove net lock from DIOCGETTIMEOUT

'pfctl -s timeouts' values are only used inside of pf, entirely protected
by the pf lock through the ioctl interface;  the net lock is useless.

Previous attempts to remove net lock usage showed that the pf lock cannot
yet entirely replace it, so start with small pieces like this one.

Contrary to IPv4/6 read-only ioctls, some pf ioctls without FWRITE flag do
modify internal pf state, which is not entirely obvious when approached
from the ioctl layer.

OK sashan dlg

18 months agoAdd `rttimer_mtx' to the locking description.
mvs [Thu, 27 Apr 2023 11:11:04 +0000 (11:11 +0000)]
Add `rttimer_mtx' to the locking description.

No functional changes.

18 months agotlsexttest: check additional logic in tlsext randomization
tb [Thu, 27 Apr 2023 10:53:58 +0000 (10:53 +0000)]
tlsexttest: check additional logic in tlsext randomization

This verifies that we put PSK always last and that the Apache 2 special
does what it is supposed to do. There is also some weak validation of
the Fisher-Yates shuffle that will likely catch errors introduced in
tlsext_randomize_build_order()

18 months agozap APM_CANCEL, dead since import; OK tb
kn [Thu, 27 Apr 2023 10:51:27 +0000 (10:51 +0000)]
zap APM_CANCEL, dead since import;  OK tb

18 months agossl_tlsext.c: Add an accessor for the tls extension type.
tb [Thu, 27 Apr 2023 10:50:37 +0000 (10:50 +0000)]
ssl_tlsext.c: Add an accessor for the tls extension type.

Needed for the tlsexttest.c

ok jsing

18 months agoSomehow I managed not to bump LIBRESSL_VERSION_NUMBER
tb [Thu, 27 Apr 2023 10:43:47 +0000 (10:43 +0000)]
Somehow I managed not to bump LIBRESSL_VERSION_NUMBER

reported by aja