openbsd
18 months agoAvoid trailing whitespace in extension printing
tb [Mon, 8 May 2023 05:30:38 +0000 (05:30 +0000)]
Avoid trailing whitespace in extension printing

If an extension is non-critical, X509V3_extensions_print() would leave
trailing whitespace. This can be trivially avoided.

ok miod

18 months agoI preparation for TSO in software, cleanup the fragment code. Use
bluhm [Sun, 7 May 2023 16:23:23 +0000 (16:23 +0000)]
I preparation for TSO in software, cleanup the fragment code.  Use
if_output_ml() to send mbuf lists to interfaces.  This can be used
for TSO, fragments, ARP and ND6.  Rename variable fml to ml.  In
pf_route6() split the if else block.  Put the safety check (hlen +
firstlen < tlen) into ip_fragment().  It makes the code correct in
case the packet is too short to be fragmented.  This should not
happen, but other functions also have this logic.
No functional change.  OK sashan@

18 months agoRecommit -Wshadow now that the warning on BIG_ENDIAN is fixed
tb [Sun, 7 May 2023 14:39:55 +0000 (14:39 +0000)]
Recommit -Wshadow now that the warning on BIG_ENDIAN is fixed

18 months agoxts128 mode: avoid two -Wshadow warnings in the BIG_ENDIAN code path.
tb [Sun, 7 May 2023 14:38:04 +0000 (14:38 +0000)]
xts128 mode: avoid two -Wshadow warnings in the BIG_ENDIAN code path.

Found by, compile tested & ok bluhm.

18 months agocron: check for garbage after the '~' in a random range.
millert [Sun, 7 May 2023 13:43:13 +0000 (13:43 +0000)]
cron: check for garbage after the '~' in a random range.
A bug in the parsing of the optional number after the '~' in a
random range prevented proper syntax checking.  OK kn@

18 months agoGPT partitions with the 'Required' attribute are not deleted by -A.
krw [Sun, 7 May 2023 13:03:03 +0000 (13:03 +0000)]
GPT partitions with the 'Required' attribute are not deleted by -A.

18 months agoBackout -Wshadow, it breaks build on powerpc64.
bluhm [Sun, 7 May 2023 12:47:48 +0000 (12:47 +0000)]
Backout -Wshadow, it breaks build on powerpc64.

18 months agoRemove net lock from DIOCOSFP{FLUSH,ADD,GET} aka. OS fingerprinting
kn [Sun, 7 May 2023 12:45:21 +0000 (12:45 +0000)]
Remove net lock from DIOCOSFP{FLUSH,ADD,GET} aka. OS fingerprinting

pf_osfp.c contains all the locking for these three ioctls, everything is
protected by the pf lock;  assert/document it and inline acess to the global
list to eliminate useless function variables.

OK bluhm sashan

18 months agoRemove a misplaced empty line
tb [Sun, 7 May 2023 07:11:50 +0000 (07:11 +0000)]
Remove a misplaced empty line

18 months agoSupport random offsets when using ranges with a step value in cron.
millert [Sat, 6 May 2023 23:06:27 +0000 (23:06 +0000)]
Support random offsets when using ranges with a step value in cron.
This extends the random range syntax to support step values.  Instead
of choosing a random number between the high and low values, the
field is treated as a range with a random offset less than the step
value.  This can be used to avoid thundering herd problems where
multiple machines contact a server all at the same time via cron jobs.

The syntax is similar to the existing range/step syntax but uses a
random range.  For example, instead of "0-59/10" in the minutes
field, "0~59/10" can be used to run a command every 10 minutes where
the first command starts at a random offset in the range [0,9].
The high and low numbers are optional, "~/10" can be used instead.

Requested by job@, OK phessler@

18 months agoRegen cert.pem
tb [Sat, 6 May 2023 17:55:38 +0000 (17:55 +0000)]
Regen cert.pem

This drops a few certs per the CA's request and TrustCor because of drama.
Certainly, a new CA, is added as well as new certs for DigiCert, SECOM and
E-Tugra. Unizeto still haven't fixed one of their certs and we still don't
want the alternative Firmaprofesional with sha1WithRSAEncryption.

ok sthen

18 months agoReplace pointless %d formatting of 0 values with
krw [Sat, 6 May 2023 15:07:02 +0000 (15:07 +0000)]
Replace pointless %d formatting of 0 values with
literal '0's.

No functional change.

18 months agoMention RTL8153D.
kevlo [Sat, 6 May 2023 08:14:26 +0000 (08:14 +0000)]
Mention RTL8153D.

18 months agoAdd support for RTL8153D.
kevlo [Sat, 6 May 2023 08:07:10 +0000 (08:07 +0000)]
Add support for RTL8153D.

The RTL8153D chipset shares many similarities with the already supported
RTL8156B chip but additionally requires a few semi-unique configurations.

Tested by weerd@
ok jmatthew@

18 months agoUse -Wshadow with clang
tb [Fri, 5 May 2023 21:23:02 +0000 (21:23 +0000)]
Use -Wshadow with clang

ok jsing (a very long time ago)

18 months agoReinstate X9.31 padding mode support in rsautl
tb [Fri, 5 May 2023 18:01:27 +0000 (18:01 +0000)]
Reinstate X9.31 padding mode support in rsautl

18 months agoFix error handling in tls_check_common_name()
tb [Fri, 5 May 2023 14:05:33 +0000 (14:05 +0000)]
Fix error handling in tls_check_common_name()

A calloc failure should be a fatal error, so make it return -1.
Also switch the default rv to -1 and distinguish error cases with
acceptable situations with goto err/goto done.

ok jsing

18 months agocrontab: move spool temp file creation to spool_mkstemp()
millert [Fri, 5 May 2023 13:50:40 +0000 (13:50 +0000)]
crontab: move spool temp file creation to spool_mkstemp()
This fixes a bug introduced in rev 1.86 where if the second seteuid()
call failed, a temporary file would be left in the spool directory.

18 months agoSalt shares the blame of the continued existence of the X9.31 padding mode
tb [Fri, 5 May 2023 12:30:40 +0000 (12:30 +0000)]
Salt shares the blame of the continued existence of the X9.31 padding mode

18 months agoReinstate documentation of RSA_X931_PADDING
tb [Fri, 5 May 2023 12:22:21 +0000 (12:22 +0000)]
Reinstate documentation of RSA_X931_PADDING

18 months agoAdd back support for RSA_X931_PADDING
tb [Fri, 5 May 2023 12:21:44 +0000 (12:21 +0000)]
Add back support for RSA_X931_PADDING

This makes the custom stalt stack work again.

Tested by robert as part of a larger diff
ok jsing

18 months agoLink rsa_x931.c to build
tb [Fri, 5 May 2023 12:20:03 +0000 (12:20 +0000)]
Link rsa_x931.c to build

18 months agoBring back the X9.31 padding helpers
tb [Fri, 5 May 2023 12:19:37 +0000 (12:19 +0000)]
Bring back the X9.31 padding helpers

Nothing should be using this anymore, except that salt decided to use
it in its home-cooked protocol, which already had its share of issues.
Hopefully the efforts to switch salt to something more reasonable and
standardized like mTLS will succeed sooner rather than later.

tested as part of a larger patch by robert
ok jsing

18 months agoPass ASPA objects in reverse since this is what aspa_add_set() expects.
claudio [Fri, 5 May 2023 10:48:16 +0000 (10:48 +0000)]
Pass ASPA objects in reverse since this is what aspa_add_set() expects.
OK tb@

18 months agoConvert K&R function definition to ansi to make clang 15 happier
tb [Fri, 5 May 2023 10:26:50 +0000 (10:26 +0000)]
Convert K&R function definition to ansi to make clang 15 happier

18 months agoSync json.c with the one from rpki-client. This adds json_do_string().
claudio [Fri, 5 May 2023 07:42:40 +0000 (07:42 +0000)]
Sync json.c with the one from rpki-client. This adds json_do_string().
Convert json_do_printf() calls using static strings or "%s" fmt over
to json_do_string() in the json output.
OK tb@

18 months agoLimit the socket buffer size to 64k for all sessions. Long time ago
claudio [Fri, 5 May 2023 07:28:08 +0000 (07:28 +0000)]
Limit the socket buffer size to 64k for all sessions. Long time ago
setting the size to 64k was increasing the size but now auto-scaling
could grow the buffer size much more. The origianl idea was that sessions
without protection had a small window size to make window attacks
harder. This problem was fixed long time ago and is no longer relevant.
Limiting the buffer size to a reasonable size ensures that not too many
updates end up queued in the TCP stack.
OK benno@ (some time ago)

18 months agoThe mbuf_queue API allows read access to integer variables which
bluhm [Fri, 5 May 2023 01:19:51 +0000 (01:19 +0000)]
The mbuf_queue API allows read access to integer variables which
another CPU may change simultaneously.  To prevent miss optimisation
by the compiler, they need the READ_ONCE() macro.  Otherwise there
could be two read operations with inconsistent values.  Writing to
integer in mq_set_maxlen() needs mutex protection.  Otherwise the
value could change within critical sections.  Again the compiler
could optimize to multiple read operations within the critical
section.  With inconsistent values, the behavior is undefined.
OK dlg@

18 months agosymbols.awk: Remove cfb dance
tb [Thu, 4 May 2023 20:15:27 +0000 (20:15 +0000)]
symbols.awk: Remove cfb dance

With e_old.c gone, we no longer need this.

18 months agoStraightforward conversion form K&R to ANSI function definitions
tb [Thu, 4 May 2023 16:08:29 +0000 (16:08 +0000)]
Straightforward conversion form K&R to ANSI function definitions

18 months agofix stupid thinko, as noticed while converting to v5.36 (later)
espie [Thu, 4 May 2023 14:02:44 +0000 (14:02 +0000)]
fix stupid thinko, as noticed while converting to v5.36 (later)

18 months agoRewrite ECParameters_dup()
tb [Thu, 4 May 2023 13:51:59 +0000 (13:51 +0000)]
Rewrite ECParameters_dup()

This should leak slightly less than the direct expansion of ASN1_dup_of().
Use freezero() since the DER could contain a private key.

ok jsing

18 months agosigh. typo
tb [Thu, 4 May 2023 13:50:14 +0000 (13:50 +0000)]
sigh. typo

18 months agoLet ecdsatest exercise ECParameters_dup() a bit
tb [Thu, 4 May 2023 13:49:29 +0000 (13:49 +0000)]
Let ecdsatest exercise ECParameters_dup() a bit

This currently leaks, which will fixed in a follow-on commit.

18 months agoRemove x9_62_test_internal()
tb [Thu, 4 May 2023 13:41:20 +0000 (13:41 +0000)]
Remove x9_62_test_internal()

This test depends on RAND_set_rand_method() allowing stupid things like
making ECDSA signatures deterministic. This was gutted a long time ago
and the function should have followed its wrappers into the attic.

18 months agoBe more economical with returning bodys in bounce messages
chrisz [Thu, 4 May 2023 12:43:44 +0000 (12:43 +0000)]
Be more economical with returning bodys in bounce messages
according to rfc3461 4.3

OK millert@

18 months agoregen
mvs [Thu, 4 May 2023 09:41:15 +0000 (09:41 +0000)]
regen

18 months agoPush kernel lock deep down to sys_sysctl(). At least network subset of
mvs [Thu, 4 May 2023 09:40:36 +0000 (09:40 +0000)]
Push kernel lock deep down to sys_sysctl(). At least network subset of
sysctl(8) MIBs relies on netlock or another locks and doesn't require
kernel lock, so unlock it. The protocols layer *_sysctl()s are left
under kernel lock and will be sequentially unlocked later.

ok bluhm@

18 months agoIntroduce a neighbor discovery mutex like ARP uses it. For now it
bluhm [Thu, 4 May 2023 06:56:56 +0000 (06:56 +0000)]
Introduce a neighbor discovery mutex like ARP uses it.  For now it
only protects nd6_list.  It does not unlock ND6 from kernel lock
yet.
OK kn@

18 months agoUse size_t instead of int in EC_POINT_point2oct()
tb [Thu, 4 May 2023 06:45:51 +0000 (06:45 +0000)]
Use size_t instead of int in EC_POINT_point2oct()

An int would be perfectly sufficient for this, but then again there would
be fewer traps.

ok jsing

18 months agoFix line wrapping
tb [Thu, 4 May 2023 05:59:38 +0000 (05:59 +0000)]
Fix line wrapping

18 months agoFix function name in doc comment
tb [Thu, 4 May 2023 05:57:18 +0000 (05:57 +0000)]
Fix function name in doc comment

18 months agoregen
kevlo [Thu, 4 May 2023 05:38:50 +0000 (05:38 +0000)]
regen

18 months agoAdd another Lenovo NVMe device id found in my ThinkPad X1 Extreme Gen 1.
kevlo [Thu, 4 May 2023 05:38:17 +0000 (05:38 +0000)]
Add another Lenovo NVMe device id found in my ThinkPad X1 Extreme Gen 1.

ok miod@

18 months agoadd virtio(4); ok (in principal) miod
jmc [Wed, 3 May 2023 22:03:17 +0000 (22:03 +0000)]
add virtio(4); ok (in principal) miod

18 months agoadd onewire(4); ok miod
jmc [Wed, 3 May 2023 21:56:13 +0000 (21:56 +0000)]
add onewire(4); ok miod

18 months agoavoid use after free
jsg [Wed, 3 May 2023 15:25:25 +0000 (15:25 +0000)]
avoid use after free
ok miod@ millert@

18 months agoImprove setenv markup
kn [Wed, 3 May 2023 14:29:57 +0000 (14:29 +0000)]
Improve setenv markup

- braces are not commands, use 'Brq'
- markup variable and value each, leave = (like mdoc(7) 'Ns' example)

18 months agoSome checks in nd6_resolve() do not require kernel lock. The analog
bluhm [Wed, 3 May 2023 11:43:31 +0000 (11:43 +0000)]
Some checks in nd6_resolve() do not require kernel lock.  The analog
code for ARP has been unlocked a while ago.
OK kn@

18 months agoRemove net lock from DIOCGETRULESET and DIOCGETRULESETS
kn [Wed, 3 May 2023 10:32:47 +0000 (10:32 +0000)]
Remove net lock from DIOCGETRULESET and DIOCGETRULESETS

Both walk the list of rulesets aka. anchors, to yield a total count and
specific anchor name, respectively.  Same access, different copy out.

pf_anchor_global are contained within pf_ioctl.c and pf_ruleset.c and
fully protected by the pf lock, as is pf_main_ruleset and its pf.c usage.

Rely on and assert for pf lock alone.  'pfctl -sr' on 60k unique rules gets
noticably faster, around 2.1s instead of 3.5s.

OK sashan

18 months agoFix a use-after-free in filemode
tb [Wed, 3 May 2023 10:22:30 +0000 (10:22 +0000)]
Fix a use-after-free in filemode

In case the TAL of a self-signed is unavailable, cert would be freed but
we'd still hold a reference to its expired time in expires, so invalidate
that pointer as well.

Found by, initial fix and ok job

18 months agoConvert json_do_printf() with "%s" a fmt string to json_do_string().
claudio [Wed, 3 May 2023 09:54:25 +0000 (09:54 +0000)]
Convert json_do_printf() with "%s" a fmt string to json_do_string().
OK tb@

18 months agoRevert utf-8 fix for X509_NAME_get_index_by_NID to avoid libtls
beck [Wed, 3 May 2023 08:10:23 +0000 (08:10 +0000)]
Revert utf-8 fix for X509_NAME_get_index_by_NID to avoid libtls
regress for the moment.  this will come back after we rethink
the failure versus not there case.

ok tb@ jsing@

18 months agoIntroduce json_do_string() a function that JSON escapes a string.
claudio [Wed, 3 May 2023 07:56:05 +0000 (07:56 +0000)]
Introduce json_do_string() a function that JSON escapes a string.

Implement json_do_printf() using json_do_string() and vasprintf().
json_do_string() only escapes the basic control chars (\b, \f, \n, \r and \t)
other control chars are considered an error. Also the forward slash is
not escaped since the JSON data is not embedded into HTML or XML.
With feedback from tb@ & millert@
OK tb@

18 months agoOn read failure just abort the rrdp request.
claudio [Wed, 3 May 2023 07:51:08 +0000 (07:51 +0000)]
On read failure just abort the rrdp request.

Before a read error would not be removed from the poll fds and trigger
constantly. RRDP_STATE_PARSE_ERROR should only be used for errors from
xml parser since then the remaining data from the socket still needs to
be consumed.
OK tb@

18 months agoBring back length check tb ok'ed and I managed to remove while
beck [Wed, 3 May 2023 07:13:18 +0000 (07:13 +0000)]
Bring back length check tb ok'ed and I managed to remove while
changing tests.

ok tb@

18 months agoFix a few KNF/whitespace issues
tb [Wed, 3 May 2023 06:30:11 +0000 (06:30 +0000)]
Fix a few KNF/whitespace issues

18 months agoSometimes the touchpad doesn't attach. Try to solve this issue be
kettenis [Tue, 2 May 2023 19:39:10 +0000 (19:39 +0000)]
Sometimes the touchpad doesn't attach.  Try to solve this issue be
reloading the firmware.

ok tobhe@

18 months agoAbsolutely astounding that custom code was written for the softraid case
deraadt [Tue, 2 May 2023 15:55:58 +0000 (15:55 +0000)]
Absolutely astounding that custom code was written for the softraid case
to require "no", instead of "n" or "no" which works at every other damn prompt
in the installer.  This seems to be an artifact of relentlessly pushing people
towards softraid by default, and I think that is a bit nasty and pushy.

18 months agoChange X509_NAME_get_index_by[NID|OBJ] to be safer.
beck [Tue, 2 May 2023 14:13:05 +0000 (14:13 +0000)]
Change X509_NAME_get_index_by[NID|OBJ] to be safer.

Currently these functions return raw ASN1_STRING bytes as
a C string and ignore the encoding in a "hold my beer I am
a toolkit not a functioning API surely it's just for testing
and you'd never send nasty bytes" kind of way.

Sadly some callers seem to use them to fetch things liks
subject name components for comparisons, and often just
use the result as a C string.

Instead, encode the resulting bytes as UTF-8 so it is
something like "text",

Add a failure case if the length provided is inadequate
or if the resulting text would contain an nul byte.

based on boringssl.

nits by dlg@
ok tb@

18 months agoMark the BIO_F_* function codes as intentionally undocumented
schwarze [Tue, 2 May 2023 13:15:05 +0000 (13:15 +0000)]
Mark the BIO_F_* function codes as intentionally undocumented
and for now, skip the the BIO_R_* reason codes.
It looks like all public symbols in the BIO library
are now documented or marked as intentionally undocumented.

18 months agofree the correct pointer. Missed in rev 1.16.
jsg [Tue, 2 May 2023 13:02:51 +0000 (13:02 +0000)]
free the correct pointer.  Missed in rev 1.16.
ok dv@

18 months agoRename P into generator
tb [Tue, 2 May 2023 13:01:57 +0000 (13:01 +0000)]
Rename P into generator

ok jsing

18 months agoRemove bogus newline in printf.
kettenis [Tue, 2 May 2023 12:32:22 +0000 (12:32 +0000)]
Remove bogus newline in printf.

ok jmatthew@

18 months agoSimplify EC_GROUP_new_by_curve_name()
tb [Tue, 2 May 2023 10:44:20 +0000 (10:44 +0000)]
Simplify EC_GROUP_new_by_curve_name()

Pull the setting of the name a.k.a. nid into ec_group_new_from_data().
This way, we can return early on finding the nid in the curve_list[].
This also avoids a silly bug where a bogus ERR_R_UNKNOWN_BUG is pushed
onto the error stack when ec_group_new_from_data() failed.

While there rework the exit path of ec_group_new_from_data() a bit.
Instead of an ok variable we can use an additional pointer to keep
track of the return value and free the EC_GROUP unconditionally.

ok jsing

18 months agoStyle tweaks for SMIME_write_PKCS7()
tb [Tue, 2 May 2023 09:56:12 +0000 (09:56 +0000)]
Style tweaks for SMIME_write_PKCS7()

Initialize the mdalgs stack at the top and test and assign for ctype_nid.
Use an empty line to separate variable declarations from the actual code
and zap an extra empty line.

ok jsing

18 months agoSwitch K&R function definition to ANSI to make clang 15 happier
tb [Tue, 2 May 2023 09:51:22 +0000 (09:51 +0000)]
Switch K&R function definition to ANSI to make clang 15 happier

18 months agoUnwrap a line
tb [Tue, 2 May 2023 09:38:33 +0000 (09:38 +0000)]
Unwrap a line

18 months agoSimplify slightly and use i2d_PKCS7_bio_stream()
tb [Tue, 2 May 2023 09:30:37 +0000 (09:30 +0000)]
Simplify slightly and use i2d_PKCS7_bio_stream()

This is a wrapper of i2d_ASN1_bio_stream() that doesn't require us to
pass in PKCS7_it.

18 months agoadd Brussels South Charleroi airport
stsp [Tue, 2 May 2023 08:48:06 +0000 (08:48 +0000)]
add Brussels South Charleroi airport

18 months agosync with userland
tb [Tue, 2 May 2023 08:05:18 +0000 (08:05 +0000)]
sync with userland

18 months agoSync with upstream
tb [Tue, 2 May 2023 08:03:54 +0000 (08:03 +0000)]
Sync with upstream

Update some links in the README, remove a duplicate word in a zlib.h doc
comment. The only code change is guarded by #if defined(_WIN32).

18 months agoCall nd6_ns_output() without kernel lock from nd6_resolve().
bluhm [Tue, 2 May 2023 06:06:13 +0000 (06:06 +0000)]
Call nd6_ns_output() without kernel lock from nd6_resolve().
OK kn@

18 months agostray whitespace
tb [Mon, 1 May 2023 21:15:26 +0000 (21:15 +0000)]
stray whitespace

18 months agoBump to 8.0
claudio [Mon, 1 May 2023 19:44:42 +0000 (19:44 +0000)]
Bump to 8.0

18 months agoAdd a missing pair of braces.
tb [Mon, 1 May 2023 17:53:01 +0000 (17:53 +0000)]
Add a missing pair of braces.

18 months agoUse uppercase for the CURVE_LIST_LENGTH macro
tb [Mon, 1 May 2023 17:49:33 +0000 (17:49 +0000)]
Use uppercase for the CURVE_LIST_LENGTH macro

18 months agoConsistently use lowercase hex digits for curve parameters
tb [Mon, 1 May 2023 17:31:15 +0000 (17:31 +0000)]
Consistently use lowercase hex digits for curve parameters

18 months agoNow that we have C99 initializers, garbage collect some comments
tb [Mon, 1 May 2023 17:29:36 +0000 (17:29 +0000)]
Now that we have C99 initializers, garbage collect some comments

18 months agoRework the curve list to use actual structs instead of a custom
tb [Mon, 1 May 2023 17:28:03 +0000 (17:28 +0000)]
Rework the curve list to use actual structs instead of a custom
serialized format.

ok jsing

18 months agoDrop the now unnecessary and unused field_type from the curve data
tb [Mon, 1 May 2023 13:49:26 +0000 (13:49 +0000)]
Drop the now unnecessary and unused field_type from the curve data

ok jsing

18 months agoConvert EC_CURVE_DATA to C99 initializers
tb [Mon, 1 May 2023 13:14:00 +0000 (13:14 +0000)]
Convert EC_CURVE_DATA to C99 initializers

Also clean up the definition of EC_CURVE_DATA a bit.

ok jsing

18 months agoSimplify ec_group_new_from_data() further
tb [Mon, 1 May 2023 12:39:38 +0000 (12:39 +0000)]
Simplify ec_group_new_from_data() further

We have a BN_CTX available, so we may as well use it. This simplifies
the cleanup path at the cost of a bit more code in the setup. Also use
an extra BIGNUM for the cofactor. Reusing x for this is just silly. If
you were really going to avoid extra allocations, this entire function
could easily have been written with three BIGNUMs.

ok jsing

18 months agoMake warnings more precise
job [Mon, 1 May 2023 11:02:23 +0000 (11:02 +0000)]
Make warnings more precise

18 months agoImport regenerated moduli.
dtucker [Mon, 1 May 2023 08:57:29 +0000 (08:57 +0000)]
Import regenerated moduli.

18 months agoThe built-in 10G Ethernet on Apple arm64 hardware does not have a MAC
kettenis [Mon, 1 May 2023 08:25:55 +0000 (08:25 +0000)]
The built-in 10G Ethernet on Apple arm64 hardware does not have a MAC
address programmed into the hardware.  Get it from the device tree instead.

ok dlg@, jmatthew@

18 months agoDrop some dead code
tb [Mon, 1 May 2023 08:16:17 +0000 (08:16 +0000)]
Drop some dead code

No member of the curve_list[] table has a method set. Thus, curve.meth
is always NULL and we never take the EC_GROUP_new(meth) code path.

ok jsing

18 months agoRemove pointless/wrong .meth = 0 entries from curves_list[]
tb [Mon, 1 May 2023 07:58:34 +0000 (07:58 +0000)]
Remove pointless/wrong .meth = 0 entries from curves_list[]

18 months agoMechanically convert curve_list[] to C99 initializers
tb [Mon, 1 May 2023 07:56:05 +0000 (07:56 +0000)]
Mechanically convert curve_list[] to C99 initializers

ok jsing

18 months agoClean up handling of nist_curves[]
tb [Mon, 1 May 2023 07:54:08 +0000 (07:54 +0000)]
Clean up handling of nist_curves[]

There's no point in introducing a typedef only for two sizeof() calls.
We might as well use an anonymous struct for this list. Make it const
while there, drop some braces and compare strcmp() return value to 0.

ok jsing

18 months agoRemove ASN1_item_ndef_i2d(3) documentation
tb [Mon, 1 May 2023 07:37:45 +0000 (07:37 +0000)]
Remove ASN1_item_ndef_i2d(3) documentation

This was the last public API explicitly named ndef/NDEF for indefinite
length encoding, so remove that explanation as well.

18 months agosync
tb [Mon, 1 May 2023 07:29:12 +0000 (07:29 +0000)]
sync

18 months agoFirst pass of removing low-level ASN.1 streaming docs
tb [Mon, 1 May 2023 07:28:11 +0000 (07:28 +0000)]
First pass of removing low-level ASN.1 streaming docs

18 months agoregen
dlg [Mon, 1 May 2023 07:24:20 +0000 (07:24 +0000)]
regen

18 months agoIntel Braswell SDIO
dlg [Mon, 1 May 2023 07:24:04 +0000 (07:24 +0000)]
Intel Braswell SDIO

18 months agospelling
jsg [Mon, 1 May 2023 07:04:38 +0000 (07:04 +0000)]
spelling

18 months agoadd acpi(4) listing; ok miod
jmc [Mon, 1 May 2023 06:04:46 +0000 (06:04 +0000)]
add acpi(4) listing; ok miod

18 months agodrm/fb-helper: set x/yres_virtual in drm_fb_helper_check_var
jsg [Mon, 1 May 2023 01:24:02 +0000 (01:24 +0000)]
drm/fb-helper: set x/yres_virtual in drm_fb_helper_check_var

From Daniel Vetter
d27acf15c8fac00a251e2a24da09fcc1bb3337dd in linux-6.1.y/6.1.27
1935f0deb6116dd785ea64d8035eab0ff441255b in mainline linux

18 months agoavoid use after free
jsg [Sun, 30 Apr 2023 23:49:14 +0000 (23:49 +0000)]
avoid use after free
ok jmatthew@

18 months agoavoid use after free
jsg [Sun, 30 Apr 2023 23:46:52 +0000 (23:46 +0000)]
avoid use after free
ok florian@