kettenis [Sun, 4 Aug 2024 12:01:18 +0000 (12:01 +0000)]
The devicetree spec says in section 2.4:
If an interrupt-generating device does not have an interrupt-parent
property, its interrupt parent is assumed to be its devicetree
parent.
Implement this aspect of the spec in the same way as Linux does by not
just looking at the immediate parent but keep on walking up the tree
until we hit a node with an "interrupt-controller" property or a node
that has an "interrupt-parent" property.
ok patrick@
kettenis [Sun, 4 Aug 2024 11:05:18 +0000 (11:05 +0000)]
Add intelpmc(4), a driver for the power management controller found on
various Intel SoCs. The driver takes care of calling the AML methods
needed to enter low power idle states during suspend-to-idle (S0i).
The driver also implements some debug code that prints the residency of
various power states in dmesg. Based on some earlier code by jcs@
ok jcs@
nicm [Sun, 4 Aug 2024 09:42:23 +0000 (09:42 +0000)]
Make a little effort to treate CRLF as LF in config files. GitHub issue
3720.
nicm [Sun, 4 Aug 2024 09:35:30 +0000 (09:35 +0000)]
The Linux console has some bugs with bright colours. It seems likely
that it is emulating them by setting a bright (or bold) flag; however,
when the colour is changed from a bright colour (say SGR 96) to a
non-bright (say SGR 36), the flag is not reset, so the new colour
remains as bright. SGR 39 (default colour) also does not reset, so you
end up with the bright default colour. Work around this by sending SGR 0
when switching away from a bright colour, and disable AX for TERM=linux.
Also make the check for AX simpler and do not check for the op
capability is not actually used. GitHub issue 3976.
nicm [Sun, 4 Aug 2024 09:01:18 +0000 (09:01 +0000)]
-l should be before -r, pointed out by jmc a while ago.
nicm [Sun, 4 Aug 2024 08:53:43 +0000 (08:53 +0000)]
Adjust the logic when deleting last buffer to better preserve the
selection: if selecting the element below the deleted one fails (because
as the last one), select the one above it instead. From Daniel Mueller,
GitHub issue 4043.
nicm [Sun, 4 Aug 2024 08:39:38 +0000 (08:39 +0000)]
Use paste_is_empty to check if there are no buffers since paste_get_top
will return NULL if there are no automatic buffers.
tb [Sun, 4 Aug 2024 08:15:36 +0000 (08:15 +0000)]
x509_vfy.c: drop some unnecessary parentheses
jmc [Sun, 4 Aug 2024 05:18:28 +0000 (05:18 +0000)]
- sort SEE ALSO
- radiusd_file(8) is section 8, not 5
yasuoka [Sun, 4 Aug 2024 03:56:57 +0000 (03:56 +0000)]
Use "module" for the modules. Add a link to radiusd_file(8) from
radiusd_eap2mschap(8).
guenther [Sat, 3 Aug 2024 23:06:56 +0000 (23:06 +0000)]
A dozen interfaces in <endian.h> were standardized in POSIX-2024 as expected
millert [Sat, 3 Aug 2024 22:23:32 +0000 (22:23 +0000)]
Set system ID field in the PVD to OpenBSD
This matches what we use in mkybrid. From Crystal Kolipe.
millert [Sat, 3 Aug 2024 22:00:31 +0000 (22:00 +0000)]
The maximum count is SKEY_MAX_SEQ + 1 not SKEY_MAX_SEQ - 1.
Sequence numbers are in the range [0, SKEY_MAX_SEQ] inclusive.
This means the maximum value for the -n option (count) should be
SKEY_MAX_SEQ + 1. From Denis Bodor.
millert [Sat, 3 Aug 2024 21:12:16 +0000 (21:12 +0000)]
Update awk to the July 28, 2024 version.
* Fixed readcsvrec resize segfault when reading csv records longer than 8k.
* Rewrite if-else chain in quoted as a switch.
guenther [Sat, 3 Aug 2024 20:13:23 +0000 (20:13 +0000)]
The improbable occurred: strlcpy(3) and strlcat(3) are in POSIX-2024.
memmem(3) was also added. Update #include visibility and manpages
and add restrict qualifiers.
"never thought I'd see this day" millert@
guenther [Sat, 3 Aug 2024 20:09:24 +0000 (20:09 +0000)]
mkostemp(3) and reallocarray(3) are in POSIX-2024:
adjust #include visibility and update the reallocarray(3) manpage
ok millert@
tb [Sat, 3 Aug 2024 13:06:37 +0000 (13:06 +0000)]
The EC_KEY method should use the EC_KEY index, not RSA
Unbreaks ssh's t-agent-pkcs11-cert regress reported by anton.
ok jsing
tb [Sat, 3 Aug 2024 07:45:26 +0000 (07:45 +0000)]
Bump CRYPTO_EX_DATA_MAX_INDEX to 32
rust-openssl tests do something weird and need lots of ex data (one index
for each registered callback, for example). This makes the regress pass
again.
noticed by anton
ok jsing
tb [Sat, 3 Aug 2024 07:43:33 +0000 (07:43 +0000)]
Use proper size for allocating indexes
It's a double pointer, so we should allocate a pointer size, not the entire
struct. This saves roughly 500B per class.
CID 507397
ok jsing
tb [Sat, 3 Aug 2024 04:53:01 +0000 (04:53 +0000)]
Document SSL_CTX_set_cert_store
From Kenjiro Nakayama
tb [Sat, 3 Aug 2024 04:50:27 +0000 (04:50 +0000)]
Prepare to provide SSL_CTX_set1_cert_store()
SSL_CTX_set_cert_store() should have been called SSL_CTX_set0_cert_store()
since it takes ownership of the store argument. Apparently a few people ran
into the issue of not bumping the refcount themselves, leading to use after
frees about 10 years ago. This is a quite rarely used API and there are no
misuses in the ports tree, but since someone did the work of writing a diff,
we can still add it.
Needless to say that SSL_CTX_get_cert_store() obviously has the exact same
issue and nobody seems to have thought of adding a get0 or get1 version to
match...
Fixes https://github.com/libressl/openbsd/issues/71
From Kenjiro Nakayama
guenther [Fri, 2 Aug 2024 22:24:51 +0000 (22:24 +0000)]
Code patch out the retpoline in the signal trampoline just
like the others.
Nudged, test, and ok deraadt@
guenther [Fri, 2 Aug 2024 22:14:54 +0000 (22:14 +0000)]
The {get,set}res[ug]id(2) family are in POSIX-2024's XSI option,
so adjust the #include visibility and update the manpage.
ok millert@
jan [Fri, 2 Aug 2024 21:08:47 +0000 (21:08 +0000)]
nc(1): remove useless function pointer dereferencing
ok tb@
tb [Fri, 2 Aug 2024 16:02:35 +0000 (16:02 +0000)]
Fix previous
Arguably the want_protocol entries in various of these tests are incorrect
but I'll leave that for another day.
tb [Fri, 2 Aug 2024 15:02:22 +0000 (15:02 +0000)]
Adjust tls regress for protocol parsing fixes
This mostly reverts what was done by beck in Tallinn and adjust tlstest
to add new test cases and now failing connection tests.
tb [Fri, 2 Aug 2024 15:00:01 +0000 (15:00 +0000)]
libtls: fix legacy protocol parsing
Redefining TLS_PROTOCOL_TLSv1_0 and TLS_PROTOCOL_TLSv1_1 to be the same
as TLS_PROTOCOL_TLSv1_2 had undesired side effects, as witnessed in the
accompanying regress tests. The protocol string all:tlsv1.0 would disable
TLSv1.2 (so only enable TLSv1.3) and tlsv1.2:!tlsv1.1 would disable all
protocols.
It makes more sense to ignore any setting of TLSv1.0 and TLSv1.1, so if
you request 'tlsv1.1' you get no protocol, but 'all:!tlsv1.1' will enable
the two supported protocols TLSv1.3 and TLSv1.2.
Restore the defines to their original values and adjust the parsing code
to set/unset them.
Issue reported by Kenjiro Nakayama
Fixes https://github.com/libressl/openbsd/issues/151
with/ok jsing
mvs [Fri, 2 Aug 2024 14:35:56 +0000 (14:35 +0000)]
regen
mvs [Fri, 2 Aug 2024 14:34:45 +0000 (14:34 +0000)]
Push kernel lock down to sysctl(2).
Unlock few obvious immutable or read-only variables from "kern.*" and
"hw.*" paths. Keep the rest variables locked as before, include pages
wiring. Use new sysctl_vs{,un}lock() functions introduced for thar
purpose.
In kern.* path:
- KERN_OSTYPE, KERN_OSRELEASE, KERN_OSVERSION, KERN_VERSION -
immutable;
- KERN_NUMVNODES - read-only access to integer;
- KERN_MBSTAT - read-only access to per-CPU counters;
In hw.* path:
- HW_MACHINE, HW_MODEL, HW_NCPUONLINE, HW_PHYSMEM, HW_VENDOR,
HW_PRODUCT, HW_VERSION, HW_SERIALNO, HW_UUID, HW_PHYSMEM64 -
immutable;
- HW_USERMEM and HW_USERMEM64 - `physmem' is immutable, uvmexp.wired
is mutable but integer; read-only access to localy stored difference
between `physmem' and uvmexp.wired;
- `hw_vars' - read-only access to integers; some of them like
HW_BYTEORDER and HW_PAGESIZE are immutable;
ok bluhm kettenis
tb [Fri, 2 Aug 2024 14:02:33 +0000 (14:02 +0000)]
free class->indexes in CRYPTO_cleanup_all_ex_data()
ok jsing
yasuoka [Fri, 2 Aug 2024 12:43:55 +0000 (12:43 +0000)]
Backout previous. It was incompleted and too early.
jsing [Fri, 2 Aug 2024 10:48:54 +0000 (10:48 +0000)]
Rewrite CRYPTO_EX_DATA.
CRYPTO_EX_DATA exists as a way to allow an application to attach data to
various components in libcrypto and libssl. The general idea is that there
are various "classes" (e.g. RSA) and an application can get an "index"
(which can have new/dup/free functions provided). The application can then
use the index to store a pointer to some form of data within that class,
for later retrieval.
However, even by OpenSSL standards, this is an insane API. The current
implementation allows for data to be set without calling new, indexes
can be used without allocation, new can be called without actually getting
an index and dup can be called either after new or without new (see regress
and RSA_get_ex_new_index(3)/CRYPTO_set_ex_data(3) for more details). On
top of this, the previous "overhaul" of the code was written to be
infinitely extensible.
For now, the rewrite intends to maintain the existing behaviour - once we
bed this down we can attempt to ratchet the API requirements and require
some sort of sensible sequence. The only intentional change is that there
is now a hard limit on the number of indexes that can be allocated
(previously there was none, relying only on ENOMEM).
ok tb@
kettenis [Fri, 2 Aug 2024 09:28:35 +0000 (09:28 +0000)]
Switch off unreferenced power resources in DVACT_POWERDOWN instead of
DVACT_SUSPEND. Fixes an issue with (un)hibernate where we don't attach
xhci(4) controllers, which would make is turn the xhci(4) off inadvertedly
because its power resource wasn't referenced.
ok mlarkin@
tb [Fri, 2 Aug 2024 05:00:35 +0000 (05:00 +0000)]
sync
tb [Fri, 2 Aug 2024 04:59:55 +0000 (04:59 +0000)]
bump minor after yesterday's symbol addition
requested by kettenis
tb [Fri, 2 Aug 2024 04:59:23 +0000 (04:59 +0000)]
Put exported symbols of libz under our own control
This adds a Symbols.list containing the exported symbols like in other
libraries in base. If upstream adds new API this won't necessarily need
a bump on our side.
help/ok kettenis
guenther [Fri, 2 Aug 2024 01:53:21 +0000 (01:53 +0000)]
getentropy(2) is in POSIX-2024, so adjust the #include visibility,
change the "too much" error to EINVAL, add GETENTROPY_MAX to
<limits.h> (via sys/syslimits.h), and update the manpage.
ok deraadt@
mglocker [Thu, 1 Aug 2024 21:55:48 +0000 (21:55 +0000)]
Add qcgpio at fdt.
ok kettenis@
bluhm [Thu, 1 Aug 2024 17:19:01 +0000 (17:19 +0000)]
Run socket splice idle timeout without kernel lock.
OK mvs@
ratchov [Thu, 1 Aug 2024 16:25:34 +0000 (16:25 +0000)]
Bump libsndio pkg-config version to 1.10.0
ok tb@
jmc [Thu, 1 Aug 2024 14:44:34 +0000 (14:44 +0000)]
bump posix spec from 2008 -> 2024;
no documented changes.
ratchov [Thu, 1 Aug 2024 14:36:27 +0000 (14:36 +0000)]
sndiod: Fix insufficent check of input from clients
Fixes possible sndiod(8) crashes caused by a global table overread
triggered by the client.
Found and analysed by Henry Ford <henryfordkjv at gmail.com>, thanks!
ok blumh, help from millert, mlarkin
jmc [Thu, 1 Aug 2024 14:30:17 +0000 (14:30 +0000)]
bump posix spec 2008 -> 2024;
no change noted from 2008 spec, but i've added -h to the list
of extensions. that was an omission from our 2008 notes,
rather than a change in the spec.
jmc [Thu, 1 Aug 2024 14:08:38 +0000 (14:08 +0000)]
bump posix spec 2008 -> 2024;
jmc [Thu, 1 Aug 2024 14:08:07 +0000 (14:08 +0000)]
bump posix spec 2008 -> 2024;
mglocker [Thu, 1 Aug 2024 11:53:03 +0000 (11:53 +0000)]
Add device tree mapping for Samsung Galaxy Book4 Edge.
ok kettenis@
sf [Thu, 1 Aug 2024 11:13:19 +0000 (11:13 +0000)]
virtio: Fix dmamap_sync calls
Add some missing bus_dmamap_sync calls, noticed with SEV and based on
an earlier diff by hshoexer@.
Some of the required syncing is done in virtio_check_vq(). Make sure
to use that function instead of calling call the virtqueue done
function directly from device specific drivers.
For viogpu this means that we cannot poll with virtio_dequeue() but
must use virtio_check_vq() instead. To make this more clear, rename
viogpu_vq_wait() into viogpu_vq_done(). While there, set the DRIVER_OK
flag even earlier. It must be set before using any virtqueue.
ok kettenis@
tb [Thu, 1 Aug 2024 06:04:22 +0000 (06:04 +0000)]
Remove a comment that was committed by accident
tb [Thu, 1 Aug 2024 04:03:10 +0000 (04:03 +0000)]
sync with src
tb [Thu, 1 Aug 2024 04:02:26 +0000 (04:02 +0000)]
sync with upstream
This adds a new public API, deflateUsed(), so is technically a minor bump.
Nothing will be using this anytime soon, so no shared library bump.
discussed with deraadt during c2k24
asou [Thu, 1 Aug 2024 01:43:08 +0000 (01:43 +0000)]
Possible to read the core file of a process.
yasuoka [Thu, 1 Aug 2024 00:58:14 +0000 (00:58 +0000)]
Use arc4random_uniform(n) instead of "% n".
yasuoka [Thu, 1 Aug 2024 00:05:16 +0000 (00:05 +0000)]
Change the routing entry of the pool addresses from blackhole to
reject. This is better because the routing cache can be purged by the
reject.
jmc [Wed, 31 Jul 2024 17:09:23 +0000 (17:09 +0000)]
bump posix spec 2008 -> 2024;
dlg [Wed, 31 Jul 2024 12:00:18 +0000 (12:00 +0000)]
add a random amount of time (up to 4 seconds) to the grace login time.
ok deraadt@ djm@
mglocker [Wed, 31 Jul 2024 10:07:33 +0000 (10:07 +0000)]
Add ufshci(4) at fdt support.
This lets me boot the Samsung Galaxy Book4 Edge in DT mode with ufshci(4),
based on a WIP DTB file.
ok kettenis@
jmc [Wed, 31 Jul 2024 05:36:13 +0000 (05:36 +0000)]
quote the book title (%T) reference, to prevent formatters (in this case
groff) stumbling over literal words that match macro names;
mglocker [Tue, 30 Jul 2024 19:47:05 +0000 (19:47 +0000)]
Make the Samsung Galaxy Book4 Edge (x1e80100) boot in ACPI mode;
Until now the kernel did panic during early boot because we couldn't
respond to the battery requests sent to the 0xa1 "bus". This (dummy)
driver registers a 0xa1 region space very early, and responds with a
static zero value to satisfy the battery requests on 0xa1, which let us
boot successfully to multi-user.
In the future this machine should be operated in FDT mode, and an improved
AML parsing interface might make this workaround obsolete also for RAMDISK.
Help and OK kettenis@, deraadt@
jmc [Tue, 30 Jul 2024 19:16:49 +0000 (19:16 +0000)]
bump posix ref 2008 -> 2024;
jmc [Tue, 30 Jul 2024 13:58:02 +0000 (13:58 +0000)]
fold the "timespec" info into the options list (since "time_arg" is already
there). move the information of "time_arg" matching touch -t to the "time_arg"
description, instead of "timespec".
ok millert
jmc [Tue, 30 Jul 2024 13:55:11 +0000 (13:55 +0000)]
bump posix spec from 2008 to 2024; ok millert
yasuoka [Tue, 30 Jul 2024 13:41:15 +0000 (13:41 +0000)]
Exports the statistics when PIPEXDSESSION. Found by ymatsui at iij.
ok mvs
claudio [Tue, 30 Jul 2024 13:28:27 +0000 (13:28 +0000)]
sched_yield() is not strong enough to overflow the recv buffer on some
systems. Use a proper sleep using usleep(100) instead.
kettenis [Tue, 30 Jul 2024 09:07:00 +0000 (09:07 +0000)]
On arm64, check whether the CPU has the BT feature to determine the expected
outcome of the test.
kettenis [Tue, 30 Jul 2024 08:59:33 +0000 (08:59 +0000)]
Populate most of the remaining hwcap and hwcap2 flags based on the detected
CPU features.
ok naddy@
guenther [Tue, 30 Jul 2024 05:57:31 +0000 (05:57 +0000)]
Recognize the features defines for POSIX 1003.1-2024, giving new values
for __POSIX_VISIBLE and __XPG_VISIBLE. ok millert@
jmc [Mon, 29 Jul 2024 20:25:55 +0000 (20:25 +0000)]
bump posix spec from 2008 -> 2024:
- "now" is a permissable timespec
- split the at and batch notes in STANDARDS
- note batch behaviour differences
- note default queue name differences
ok millert
kettenis [Mon, 29 Jul 2024 18:43:11 +0000 (18:43 +0000)]
Set the target ACPI to S5 when we're powering down the machine. Prevents
us from trying to put devices into the D3 power state which some hardware
doesn't appreciate.
ok ratchov@, deraadt@
claudio [Mon, 29 Jul 2024 12:42:53 +0000 (12:42 +0000)]
Move the signal related kqueue filters to kern_event.c.
Since proc and signal filters share the same klist it makes sense
to keep them together.
OK mvs@
bluhm [Mon, 29 Jul 2024 12:41:30 +0000 (12:41 +0000)]
Use shared net lock instead of exclusive when frag6 calls icmp6_error().
OK mvs@ a while ago as part of a larger diff
patrick [Mon, 29 Jul 2024 12:32:02 +0000 (12:32 +0000)]
add SBZ
mvs [Mon, 29 Jul 2024 10:35:22 +0000 (10:35 +0000)]
Remove `sb_sel' from sobuf_print() output, no sense to print "...".
ok bluhm
claudio [Mon, 29 Jul 2024 09:50:30 +0000 (09:50 +0000)]
Sync with sys/proc.h after P_CONTINUED -> PS_CONTINUED change.
OK mpi@
claudio [Mon, 29 Jul 2024 09:49:49 +0000 (09:49 +0000)]
Replace per thread P_CONTINUED with per process PS_CONTINUED flag
dowait6() can only look at per process state so switch this over.
Right now SIGCONT handling in ptsignal is recursive and not quite
right but this is a step in the right direction. It fixes dowait6()
handling for multithreaded processes where the main thread exited.
OK mpi@
tb [Mon, 29 Jul 2024 06:05:31 +0000 (06:05 +0000)]
typo: regresss -> regress
jsg [Mon, 29 Jul 2024 01:26:15 +0000 (01:26 +0000)]
drm/amdgpu: Fix signedness bug in sdma_v4_0_process_trap_irq()
From Dan Carpenter
4edb0a84e6b32e75dc9bd6dd085b2c2ff19ec287 in linux-6.6.y/6.6.43
6769a23697f17f9bf9365ca8ed62fe37e361a05a in mainline linux
jsg [Mon, 29 Jul 2024 01:18:32 +0000 (01:18 +0000)]
regen
jsg [Mon, 29 Jul 2024 01:17:53 +0000 (01:17 +0000)]
add "Strix Point" APU device id, and another Navi 31
150e rev c1 is Radeon 880M/890M
150e rev c4 is Radeon 880M/890M
150e rev c6 is Radeon 890M
744a rev 00 is Radeon Pro W7900 Dual Slot
found in AMD Software: Adrenalin Edition 24.7.1
kn [Sun, 28 Jul 2024 21:44:42 +0000 (21:44 +0000)]
The dash must not come first in the getopt(3) string
Broken/sorted in previous -u addition, fix/move it to the end.
By tb
millert [Sun, 28 Jul 2024 19:13:26 +0000 (19:13 +0000)]
pwd_mkdb: limit db entries to _PW_BUF_LEN to match libc
Otherwise, it is possible to create a passwd(5) entry that is too
large for getpwent(3), which ignores database entries larger than
_PW_BUF_LEN. This adds a check in db_store() so that we do not
store an entry larger than getpwent(3) can read. Callers of
pwd_mkdb(8), typically via pw_mkdb(3), already check for failure.
In most cases, the checks in chpass(1) will prevent a user from
creating an entry that is too large by changing their gecos field.
However, it is only when storing the db record that we know the
true size.
OK deraadt@
deraadt [Sun, 28 Jul 2024 15:31:22 +0000 (15:31 +0000)]
block SIGHUP in the same places where SIGCHLD is blocked to protect
the process lists, because the SIGHUP handler looks at them (and
it is very difficult to rewrite the that handler a different way)
ok millert
kn [Sun, 28 Jul 2024 10:08:44 +0000 (10:08 +0000)]
Support "-u name" to remove variable from environment
OK aisha millert
Feedback jmc
miod [Sat, 27 Jul 2024 17:31:49 +0000 (17:31 +0000)]
Allow for another TEMPerHUM version. From Sebastiaan Indesteege.
kettenis [Sat, 27 Jul 2024 13:31:26 +0000 (13:31 +0000)]
Use ".arch_extension fp" for the signal trampoline since it saves and
restores the FPU registers (running in userland). And add an
".arch_extension nofp" afterwards such that the remaining kernel code
can't touch the FPU.
Needed to build with clang18
ok jsg@
deraadt [Sat, 27 Jul 2024 03:26:12 +0000 (03:26 +0000)]
sync
deraadt [Sat, 27 Jul 2024 03:26:04 +0000 (03:26 +0000)]
shorten strings by removing pointless information
ok jsg
guenther [Sat, 27 Jul 2024 02:10:26 +0000 (02:10 +0000)]
Sort and regularize all the ktrstruct() macros
ok deraadt@
guenther [Fri, 26 Jul 2024 19:16:31 +0000 (19:16 +0000)]
Trace struct itimerval
ok deraadt@ claudio@
bluhm [Fri, 26 Jul 2024 15:59:04 +0000 (15:59 +0000)]
On AMD vmm(4) set SVM_INTERCEPT_INVLPGA in intercept1.
According to the AMD64 Architecture Programmer's Manual volume 2
the intercept SVM_INTERCEPT_INVLPGA needs to be set in vmcb.intercept1
(vector 3, offest 00Ch) instead of intercept2 (vector 4, offset
010h). SVM_INTERCEPT_INVLPGA is bit 26, so before vcpu_reset_regs_svm()
was actually setting an intercept for CR10, which does not exist.
from hshoexer@; OK mlarkin@
yasuoka [Fri, 26 Jul 2024 15:51:09 +0000 (15:51 +0000)]
Mark ipsecflowinfo immutable.
ok mvs
yasuoka [Fri, 26 Jul 2024 15:45:31 +0000 (15:45 +0000)]
In pipex_l2tp_input(), check if ipsecflowinfo is not changed instead
of updating it blindly.
ok mvs
naddy [Fri, 26 Jul 2024 15:24:49 +0000 (15:24 +0000)]
document the reduced logingrace penalty
bluhm [Fri, 26 Jul 2024 14:38:20 +0000 (14:38 +0000)]
Run UDP input on multiple CPU in parallel.
The socket layer of UDP has been made fully MP safe. UDP output
is MP safe for a while. mvs@ has fixed the missing pieces in socket
splicing recently. This means that complete UDP stack can be
processed by multiple threads now. Activate multi processing for
udp_input() when called with IPv4 or IPv6 packets.
Usually IP processing runs on multiple softnet threads with shared
net lock. From there local packets are queued and processed by one
thread with exclusive net lock. If the PR_MPINPUT flag is set,
protocol input is called directly from IP input on multiple threads,
with shared net lock and no additional queueing.
tested by Hrvoje Popovski; OK mvs@
tb [Fri, 26 Jul 2024 13:34:56 +0000 (13:34 +0000)]
Mark X509at_get_attr{,_count}() and X509at_delete_attr() as unused
ok jsing
tb [Fri, 26 Jul 2024 13:33:39 +0000 (13:33 +0000)]
Disable X509at_get_attr{,_count}() and X509at_delete_attr()
These are (not so) thin wrappers around the stack API and only make
things unreadable by adding an unneccesary layer of indirection and
repeating checks already present in the stack API. X509at_delete_attr()
is a masterpiece.
ok jsing
tb [Fri, 26 Jul 2024 13:30:40 +0000 (13:30 +0000)]
Inline last user of X509at_get_attr()
ok jsing
tb [Fri, 26 Jul 2024 13:24:39 +0000 (13:24 +0000)]
Inline trivial X509at_* calls in x509_req
ok jsing
tb [Fri, 26 Jul 2024 13:23:52 +0000 (13:23 +0000)]
Inline a few trivial X509at_* calls in cms
ok jsing
sf [Fri, 26 Jul 2024 07:55:23 +0000 (07:55 +0000)]
virtio: add/fix feature bits
There was a off-by-one in unused vioblk feature defines. Fix this.
Add missing feature bits from virtio 1.2 so that they are displayed
with VIRTIO_DEBUG.
ok jan@
sf [Fri, 26 Jul 2024 06:29:01 +0000 (06:29 +0000)]
vio: Don't request csum offload if not negotiated
The standard says "A driver MUST NOT enable an offload for which the
appropriate feature has not been negotiated."
ok jan@