schwarze [Sun, 10 Sep 2023 14:39:58 +0000 (14:39 +0000)]
Briefly mention SSLeay_add_all_algorithms(3) because
surprisingly large numbers of software packages still call it.
Mark the unused aliases OPENSSL_add_all_algorithms_conf(3),
OPENSSL_add_all_algorithms_noconf(3), SSLeay_add_all_ciphers(3),
and SSLeay_add_all_digests(3) as intentionally undicumented.
schwarze [Sun, 10 Sep 2023 13:58:46 +0000 (13:58 +0000)]
Document the deprecated functions EVP_set_pw_prompt(3) and
EVP_get_pw_prompt(3) because some software out there still uses them.
While here, also improve the description of EVP_read_pw_string(3).
Delete documentation for des_read_pw(3) and des_read_pw_string(3).
They couldn't be used in LibreSSL since at least 2016 because they
were never in Symbols.list, and in 2022, jsing@ also removed them
from <openssl/ui_compat.h>.
Delete the misleading AUTHORS section. Richard Levitte did not
write the original implementation of these functions, and the
compatibility wrapper around the UI_process(3) API that he did
write is not notable enough to be mentioned so prominently.
schwarze [Sun, 10 Sep 2023 11:20:52 +0000 (11:20 +0000)]
Mark EVP_ENCODE_LENGTH() and EVP_DECODE_LENGTH() as intentionally
undocumented because they do not describe properties of the Base64
encoding but add arbitrary constant lengths, hence being implementation
details of BIO_f_base64(3). Besides, they are practically unused
outside evp/bio_b64.c.
jsg [Sun, 10 Sep 2023 09:32:31 +0000 (09:32 +0000)]
load amd patch into a malloc'd region to make it page aligned
avoids a General-Protection Exception on patch loader wrmsr with
A10-5700, TN-A1
00610f01 15-10-01
the alignment requirement is not present on at least
Ryzen 5 2600X, PiR-B2
00800f82 17-08-02
problem reported and fix tested by espie@
jsg [Sun, 10 Sep 2023 06:25:09 +0000 (06:25 +0000)]
don't error on WSDISPLAYIO_SVIDEO/WSDISPLAYIO_GVIDEO in drm drivers
it is handled by wsdisplay(4)
ok dlg@
jsg [Sun, 10 Sep 2023 05:22:46 +0000 (05:22 +0000)]
fix Xr punctuation
jsg [Sun, 10 Sep 2023 04:05:26 +0000 (04:05 +0000)]
spelling
djm [Sun, 10 Sep 2023 03:51:55 +0000 (03:51 +0000)]
typo in comment
djm [Sun, 10 Sep 2023 03:25:53 +0000 (03:25 +0000)]
randomise keystroke obfuscation intervals and average interval rate.
ok dtucker@
cheloha [Sun, 10 Sep 2023 03:08:05 +0000 (03:08 +0000)]
clockintr: support an arbitrary callback function argument
Callers can now provide an argument pointer to clockintr_establish().
The pointer is kept in a new struct clockintr member, cl_arg. The
pointer is passed as the third parameter to clockintr.cl_func when it
is executed during clockintr_dispatch(). Like the callback function,
the callback argument is immutable after the clockintr is established.
At present, nothing uses this. All current clockintr_establish()
callers pass a NULL arg pointer. However, I am confident that dt(4)'s
profile provider will need this in the near future.
Requested by dlg@ back in March.
cheloha [Sun, 10 Sep 2023 01:41:16 +0000 (01:41 +0000)]
clockintr_dispatch: add intermediate pointer for clockintr_queue.cq_shadow
Adding an intermediate pointer lets me shortens "cq->cq_shadow" to
just "shadow". I think it makes the dispatch loop logic a little
easier to read.
While here, add a clarifying comment.
millert [Sat, 9 Sep 2023 18:59:43 +0000 (18:59 +0000)]
Update awk to Dec 15, 2022 version.
Force hex escapes in strings to be no more than two characters, as
they already are in regular expressions. This brings internal
consistency, as well as consistency with gawk.
tb [Sat, 9 Sep 2023 18:34:53 +0000 (18:34 +0000)]
sync
cheloha [Sat, 9 Sep 2023 18:19:03 +0000 (18:19 +0000)]
kernel: remove schedhz
Now that alpha no longer sets schedhz, schedhz is a dead variable.
Remove it.
For now, leave the schedclock() call in place in statclock(). It
still runs at its default rate of (stathz / 4).
Part of mpi@'s WIP scheduler patch. Suggested by mpi@.
Thread: https://marc.info/?l=openbsd-tech&m=
169419781317781&w=2
ok mpi@
cheloha [Sat, 9 Sep 2023 17:07:59 +0000 (17:07 +0000)]
clockintr_advance: tweak logic to eliminate early-return
With the switch to clockintr_schedule_locked(), clockintr_advance() is
now much shorter and the early-return from the non-mutex path doesn't
make the function any easier to read. Move the mutex path into the else
branch and always return 'count' at the end of the function.
cheloha [Sat, 9 Sep 2023 16:59:01 +0000 (16:59 +0000)]
clockintr_schedule: tweak logic to eliminate early return
With the introduction of clockintr_schedule_locked(), clockintr_schedule()
is short enough that the early-return from the non-mutex path doesn't make
the function easier to read. Move the mutex path into the else branch.
cheloha [Sat, 9 Sep 2023 16:34:39 +0000 (16:34 +0000)]
clockintr_advance: delete duplicate code, call clockintr_schedule_locked()
Replace a bunch of code in clockintr_advance() with just one call to
clockintr_schedule_locked() -- everyone's favorite hot new subroutine.
cheloha [Sat, 9 Sep 2023 16:20:48 +0000 (16:20 +0000)]
clockintr_schedule: refactor reusable code into clockintr_schedule_locked()
Move the code common to clockintr_advance() and clockintr_schedule()
from clockintr_schedule() into a new function, clockintr_schedule_locked().
In a subsequent commit we will replace that code in clockintr_advance()
with a call to clockintr_schedule_locked().
claudio [Sat, 9 Sep 2023 14:50:09 +0000 (14:50 +0000)]
Fix SCHED_LOCK() leak in single_thread_set()
In the (q->p_flag & P_WEXIT) branch is a continue that did not release
the SCHED_LOCK. Refactor the code a bit to simplify the places SCHED_LOCK
is grabbed and released.
Reported-by: syzbot+ea26d351acfad3bb3f15@syzkaller.appspotmail.com
OK kettenis@
schwarze [Sat, 9 Sep 2023 14:39:09 +0000 (14:39 +0000)]
new manual page EVP_PKEY_CTX_get_operation(3),
also documenting EVP_PKEY_CTX_get0_pkey(3)
schwarze [Sat, 9 Sep 2023 14:35:23 +0000 (14:35 +0000)]
document EVP_PKEY_CTX_get_data(3) and EVP_PKEY_CTX_set_data(3)
schwarze [Sat, 9 Sep 2023 14:31:38 +0000 (14:31 +0000)]
Document EVP_PKEY_CTX_set0_keygen_info(3).
While here, also add the missing RETURN VALUES entries
for EVP_PKEY_gen_cb(3), EVP_PKEY_CTX_get_cb(3),
and EVP_PKEY_CTX_get_keygen_info(3).
schwarze [Sat, 9 Sep 2023 14:26:35 +0000 (14:26 +0000)]
Document EVP_PKEY_CTX_get0_peerkey(3).
While here, also make the descriptions of the other functions more precise.
kevlo [Sat, 9 Sep 2023 14:24:06 +0000 (14:24 +0000)]
Regen
kevlo [Sat, 9 Sep 2023 14:23:37 +0000 (14:23 +0000)]
Correct TP-LINK bluetooth ID.
Reported on bugs@ by Douglas Silva <dougsilva.0@outlook.com>.
ok jsg@
bentley [Sat, 9 Sep 2023 14:15:42 +0000 (14:15 +0000)]
an PT_OPENBSD_* -> a PT_OPENBSD_*
ok jmc@
espie [Sat, 9 Sep 2023 10:56:22 +0000 (10:56 +0000)]
gc WRKDIR_LINK_NAME
espie [Sat, 9 Sep 2023 10:06:59 +0000 (10:06 +0000)]
remove PATCH_DEBUG documentation, since it has been always on since 2016
tb [Sat, 9 Sep 2023 09:12:01 +0000 (09:12 +0000)]
sync
asou [Sat, 9 Sep 2023 06:52:40 +0000 (06:52 +0000)]
Print waring message when not allocated memory in putleakinfo().
ok otto.
cheloha [Sat, 9 Sep 2023 03:03:45 +0000 (03:03 +0000)]
clockintr_dispatch: copy cl_func to cq_shadow.cl_func before calling it
When we add support for disestablishing clockintrs in the near future,
the current call to cl->cl_func will be in a race with free(9) on
another CPU. Copying cl_func to cq_shadow.cl_func before leaving the
mutex and then calling the copy eliminates the race.
Correcting the race has a negligible performance impact and simplifies
the forthcoming clockintr_disestablish() patch.
Requested by dlg@ back in March.
bentley [Sat, 9 Sep 2023 00:10:20 +0000 (00:10 +0000)]
Mark up paths with Pa.
cheloha [Fri, 8 Sep 2023 22:23:30 +0000 (22:23 +0000)]
clockintr: rename "clockintr_schedule_locked" to "clockqueue_pend_insert"
The function clockintr_schedule_locked() is poorly named. "Schedule" is
a broad word that connotes too much and "clockintr" implies that we are
mainly operating on the clockintr itself, which is untrue.
clockqueue_pend_insert() is a better name: it is clearer that we are
inserting an element into cq_pend.
cheloha [Fri, 8 Sep 2023 22:14:57 +0000 (22:14 +0000)]
clockintr: rename "clockintr_cancel_locked" to "clockqueue_pend_delete"
The function clockintr_cancel_locked() is poorly named. "Cancel" is a
broad word that connotes too much. "clockintr" implies that we are
mainly operating on the clockintr itself, which is untrue: cq_pend,
the particular data structure, is important.
clockqueue_pend_delete() is a much better name: it specifies the data
structure we're operating on ("cq_pend") and the fundamental operation
at work ("delete").
lraab [Fri, 8 Sep 2023 21:57:09 +0000 (21:57 +0000)]
man5/bsd.port.mk: document USE_NOBTCFI
feedback from and ok bentley@ and tb@ thanks!
bluhm [Fri, 8 Sep 2023 21:16:55 +0000 (21:16 +0000)]
Print a message when test has finished successfully.
bluhm [Fri, 8 Sep 2023 21:15:02 +0000 (21:15 +0000)]
Test pf and stack with double atomic IPv6 fragments.
That means the IPv6 header chain contains two fragment header that
spawn the whole packet. Such packets are illegal and pf drops them.
Otherwise they could bypass pf rules as described in CVE-2023-4809.
OpenBSD is not affected as pf_walk_header6() drops them with "IPv6
multiple fragment" log message. This check exists since 2013 when
special support for atomic fragments was added to pf.
If pf is disabled, the IPv6 stack accepts such packets. I do not
consider this a security issue.
kn [Fri, 8 Sep 2023 20:47:22 +0000 (20:47 +0000)]
Clean up old console bootargs
7.3 is long gone, you must have new bootloaders and new kernels.
Zaps both condition and else block, unindent and merge lines where fit.
Feedback OK kettenis
Tests OK denis
mvs [Fri, 8 Sep 2023 20:00:27 +0000 (20:00 +0000)]
Remove the remnants of the leftover selinfo from vnode(9) layer. Just
mechanical 'selinfo' to 'klist' replacement in 'vnode' structure because
knote(9) API is already used.
<sys/selinfo.h> headers added where is was required.
ok bluhm
naddy [Fri, 8 Sep 2023 13:40:52 +0000 (13:40 +0000)]
revert previous
The change broke IPv6 neighbor discovery, and anton@ reports several
regression test failures.
ok bluhm@
schwarze [Fri, 8 Sep 2023 11:37:58 +0000 (11:37 +0000)]
Mark EVP_PKEY_save_parameters() as intentionally undocumented
because nothing uses it according to codesearch.debian.net
and it only affects X509_PUBKEY_set(3) for DSA and GOST2001 keys,
resulting in incomplete output without the public key parameters.
claudio [Fri, 8 Sep 2023 09:06:31 +0000 (09:06 +0000)]
Change how ps_threads and p_thr_link are locked away from using SCHED_LOCK.
The per process thread list can be traversed (read) by holding either
the KERNEL_LOCK or the per process ps_mtx (instead of SCHED_LOCK).
Abusing the SCHED_LOCK for this makes it impossible to split up the
scheduler lock into something more fine grained.
Tested by phessler@, ok mpi@
nicm [Fri, 8 Sep 2023 07:05:06 +0000 (07:05 +0000)]
On second thoughts, do check DA2 for DECFRA and DECSLRM since that will
catch terminals that say they are VT520 even if we can't use DA1
(because of VTE).
nicm [Fri, 8 Sep 2023 06:52:31 +0000 (06:52 +0000)]
Use DECSLRM and DECFRA only at level 4 rather than checking the terminal
id.
djm [Fri, 8 Sep 2023 06:34:24 +0000 (06:34 +0000)]
fix sizeof(*ptr) instead sizeof(ptr) in realloc (pointer here is
char**, so harmless); spotted in CID 416964
djm [Fri, 8 Sep 2023 06:10:57 +0000 (06:10 +0000)]
regress test recursive remote-remote directories copies where the
directory contains a symlink to another directory.
also remove errant `set -x` that snuck in at some point
djm [Fri, 8 Sep 2023 06:10:02 +0000 (06:10 +0000)]
fix recursive remote-remote copies of directories that contain symlinks
to other directories (similar to bz3611)
yasuoka [Fri, 8 Sep 2023 05:56:22 +0000 (05:56 +0000)]
Add request or response decoration feature which is used through the
radiusd module interface. This makes additional modules can modify
RADIUS request or response messages. Also add new "radius_standard"
module which uses this new feature, provides some generic features
like "strip-atmark-realm" which removes the realm part from the
User-Name attribute. from IIJ.
djm [Fri, 8 Sep 2023 05:56:13 +0000 (05:56 +0000)]
the sftp code was one of my first contributions to OpenSSH and it
shows - the function names are terrible.
Rename do_blah() to sftp_blah() to make them less so.
Completely mechanical except for sftp_stat() and sftp_lstat() which
change from returning a pointer to a static variable (error-prone) to
taking a pointer to a caller-provided receiver.
djm [Fri, 8 Sep 2023 05:50:57 +0000 (05:50 +0000)]
regress test for recursive copies of directories containing symlinks
to other directories. bz3611, ok dtucker@
djm [Fri, 8 Sep 2023 05:50:12 +0000 (05:50 +0000)]
fix scp in SFTP mode recursive upload and download of directories
that contain symlinks to other directories. In scp mode, the links
would be followed, but in SFTP mode they were not. bz3611, ok dtucker@
espie [Thu, 7 Sep 2023 23:32:56 +0000 (23:32 +0000)]
blast from the past... be gone!
tb [Thu, 7 Sep 2023 21:31:00 +0000 (21:31 +0000)]
sync
schwarze [Thu, 7 Sep 2023 19:59:58 +0000 (19:59 +0000)]
Below RETURN VALUES,
* mention that EVP_MD_CTX_md(3) also returns NULL
if no message digest is configured yet; and
* omplete the list of functions returning const EVP_MD *,
also making the wording more precise.
Delete EVP_MAX_MD_SIZE from the NAME, SYNOPSIS, and HISTORY sections
because we do not usually document preprocessor macro constants in
this way. There is nothing special about this constant justifying
an exception.
schwarze [Thu, 7 Sep 2023 19:28:37 +0000 (19:28 +0000)]
add the missing entries for EVP_MD_CTX_pkey_ctx(3) and EVP_MD_CTX_md_data(3)
to the RETURN VALUES section
espie [Thu, 7 Sep 2023 17:26:14 +0000 (17:26 +0000)]
always say "the sqlports package"
espie [Thu, 7 Sep 2023 17:24:39 +0000 (17:24 +0000)]
be more direct about SUPDISTFILES
espie [Thu, 7 Sep 2023 17:21:10 +0000 (17:21 +0000)]
We prefer SITES.sufx (lowercase)
espie [Thu, 7 Sep 2023 17:19:19 +0000 (17:19 +0000)]
MASTER_SITES -> SITES
espie [Thu, 7 Sep 2023 17:17:43 +0000 (17:17 +0000)]
mention that SITE_BACKUP won't work transparently until we completely
get rid of MASTER_SITES.
espie [Thu, 7 Sep 2023 17:15:29 +0000 (17:15 +0000)]
beyond libtool
schwarze [Thu, 7 Sep 2023 16:32:41 +0000 (16:32 +0000)]
document EVP_MD_nid(3) and EVP_MD_name(3)
schwarze [Thu, 7 Sep 2023 15:57:04 +0000 (15:57 +0000)]
improve the descriptions of most functions
espie [Thu, 7 Sep 2023 15:28:48 +0000 (15:28 +0000)]
default value for PORTHOME (todo: add it as diagnostic)
espie [Thu, 7 Sep 2023 15:23:00 +0000 (15:23 +0000)]
don't share patches lightly
espie [Thu, 7 Sep 2023 15:12:38 +0000 (15:12 +0000)]
even though it's annotated, users don't see .Cm, so note we're talking
about the target since 'package' is used all over the place.
schwarze [Thu, 7 Sep 2023 14:22:11 +0000 (14:22 +0000)]
Split two new manual pages EVP_MD_nid(3) and EVP_MD_CTX_ctrl(3)
out of the large EVP_DigestInit(3). No text change.
tobhe [Thu, 7 Sep 2023 11:17:32 +0000 (11:17 +0000)]
Treat consecutive paragraph indicators as different paragraphs
Consecutive empty lines count toward the same state, so there're
2x states (to get in and out). ^L and .PP are counted as text,
hitting those in the text should be treated as getting out of a
paragraph and then getting in.
From Walter Alejandro Iglesias and Zhihao Yuan in nvi2
ok bluhm@
nicm [Thu, 7 Sep 2023 10:21:46 +0000 (10:21 +0000)]
Use DECSLRM and DECFRA on terminals pretending to be VT520 or VT525 as
well as VT420.
sashan [Thu, 7 Sep 2023 09:59:43 +0000 (09:59 +0000)]
pf(4) ignores 'keep state' and 'nat-to' actions for unsolicited
icmp error responses. Fix tightens rule matching logic so icmp
error responses no longer match 'keep state' rule. In typical
scenarios icmp errors (if solicited) should match existing state.
The change is going to bite firewalls which deal with asymmetric
routes. In those cases the 'keep state' action should be relaxed
to sloppy or new 'no state' rule to explicitly match icmp
errors should be added.
The issue has been reported by Peter J. Philip (pjp _at_ delphinusdns.org).
Discussed with bluhm@ and florian@
OK bluhm@
nicm [Thu, 7 Sep 2023 07:19:21 +0000 (07:19 +0000)]
Accept 65 for VT525 as well.
jsg [Thu, 7 Sep 2023 06:32:24 +0000 (06:32 +0000)]
convert to ansi function decl
daniel [Thu, 7 Sep 2023 02:12:07 +0000 (02:12 +0000)]
regen
daniel [Thu, 7 Sep 2023 02:11:26 +0000 (02:11 +0000)]
add Realtek 8852AE
ok kevlo@
jsg [Thu, 7 Sep 2023 01:45:53 +0000 (01:45 +0000)]
mention ConnectX-6 Dx and ConnectX-6 Lx
jsg [Thu, 7 Sep 2023 01:43:46 +0000 (01:43 +0000)]
match on Mellanox ConnectX-6 Lx
from and tested by Olivier Croquin
ok dlg@
jsg [Thu, 7 Sep 2023 01:41:42 +0000 (01:41 +0000)]
regen
jsg [Thu, 7 Sep 2023 01:41:09 +0000 (01:41 +0000)]
add Mellanox ConnectX-6 Lx
from Olivier Croquin
jca [Thu, 7 Sep 2023 01:08:36 +0000 (01:08 +0000)]
Fix missing function declarations to let mkhybrid build with clang-16
ok millert@
jsg [Wed, 6 Sep 2023 23:47:51 +0000 (23:47 +0000)]
regen
jsg [Wed, 6 Sep 2023 23:47:14 +0000 (23:47 +0000)]
add a Navi 32 device id
747e rev c8 is Radeon RX 7800 XT
747e rev ff is Radeon RX 7700 XT
found in AMD Software: Adrenalin Edition 23.9.1
djm [Wed, 6 Sep 2023 23:36:09 +0000 (23:36 +0000)]
regression test for override of subsystem in match blocks
djm [Wed, 6 Sep 2023 23:35:35 +0000 (23:35 +0000)]
allow override of Sybsystem directives in sshd Match blocks
djm [Wed, 6 Sep 2023 23:26:37 +0000 (23:26 +0000)]
allocate the subsystems array as necessary and remove the fixed limit of
subsystems. Saves a few kb of memory in the server and makes it more like
the other options.
djm [Wed, 6 Sep 2023 23:23:53 +0000 (23:23 +0000)]
preserve quoting of Subsystem commands and arguments. This may change
behaviour of exotic configurations, but the most common subsystem
configuration (sftp-server) is unlikely to be affected.
djm [Wed, 6 Sep 2023 23:21:36 +0000 (23:21 +0000)]
downgrade duplicate Subsystem directives from being a fatal error
to being a debug message to match behaviour with just about all other
directives.
djm [Wed, 6 Sep 2023 23:18:15 +0000 (23:18 +0000)]
handle cr+lf (instead of just cr) in sshsig signature files
espie [Wed, 6 Sep 2023 22:28:50 +0000 (22:28 +0000)]
point GH_* to DIST_TUPLE for complicated stuff
remove description of MTREE_FILE that was removed a few years ago
espie [Wed, 6 Sep 2023 21:14:43 +0000 (21:14 +0000)]
ROACH_URL and ROACH_SITES, to help portroach figure out the new things.
dv [Wed, 6 Sep 2023 19:27:54 +0000 (19:27 +0000)]
Correct error message: vionet -> vioblk. No functional change.
dv [Wed, 6 Sep 2023 19:26:39 +0000 (19:26 +0000)]
vmd(8): clean up struct ioinfo.
In prep for fixing some vioblk device issues, simplify the ioinfo
struct by dropping members that aren't needed.
ok mlarkin@
schwarze [Wed, 6 Sep 2023 16:26:49 +0000 (16:26 +0000)]
Mark EVP_CIPHER_set_asn1_iv() and EVP_CIPHER_get_asn1_iv() as intentionally
undocumented because they are unused outside libcrypto according
to codesearch.debian.net and should probably not be public: they seem
hardly useful even for implementing custom EVP_CIPHER algorithms.
tb@ came to similar conclusions regarding these two functions.
schwarze [Wed, 6 Sep 2023 16:03:29 +0000 (16:03 +0000)]
minimal linting, no output change: .Ed with an argument;
stray .Pp before .It; new sentence, new line; trailing whitespace
job [Wed, 6 Sep 2023 15:53:07 +0000 (15:53 +0000)]
Initialize afi & safi to zero
OK tb@
tb [Wed, 6 Sep 2023 15:32:54 +0000 (15:32 +0000)]
Avoid use-of-uninitialized in i2r_IPAddrBlocks()
Reported by Viktor Szakats in
https://github.com/libressl/portable/issues/910
ok job
espie [Wed, 6 Sep 2023 15:02:51 +0000 (15:02 +0000)]
reflect new reality
jca [Wed, 6 Sep 2023 15:02:32 +0000 (15:02 +0000)]
Use casts to force two function pointers assignements and fix build with clang-16
"looks fine to me" robert@, ok kettenis@
caspar [Wed, 6 Sep 2023 13:33:46 +0000 (13:33 +0000)]
bsd.port.mk.5: fix markup of NO_ARCH variable
OK kn@ schwarze@
robert [Wed, 6 Sep 2023 12:45:19 +0000 (12:45 +0000)]
explicitly cast ULLONG_MAX to double when comparing to a double
ok krw@
schwarze [Wed, 6 Sep 2023 12:26:59 +0000 (12:26 +0000)]
Correct the RETURN VALUES of OBJ_add_object(3).
The text was misleading before and after the improvement
in obj_dat.c rev. 1.61. The way i'm fixing the documentation
here takes that improvement into account.
Also add a CAVEATS section about adding incomplete objects.