matthieu [Mon, 28 Nov 2022 07:38:40 +0000 (07:38 +0000)]
sync
matthieu [Mon, 28 Nov 2022 07:37:48 +0000 (07:37 +0000)]
Remove rex.x, obsolete remote execution protocol
that still used sgttyb struct. ok gnezdo@ miod@
tb [Mon, 28 Nov 2022 07:24:03 +0000 (07:24 +0000)]
Tweak x509_constraints_uri_host() regress to test for NULL deref in
fixed in x509_constraints.c r1.29.
tb [Mon, 28 Nov 2022 07:22:15 +0000 (07:22 +0000)]
Fix NULL dereference in x509_constraints_uri_host()
When called from v2i, hostpart in x509_constraints_uri_host() is NULL, so
add a NULL check before storing the strdup result in it.
From Anton Borowka
ok jsing miod
djm [Mon, 28 Nov 2022 01:38:22 +0000 (01:38 +0000)]
tighten pledge(2) after session establishment
feedback, ok & testing in snaps deraadt@
djm [Mon, 28 Nov 2022 01:37:36 +0000 (01:37 +0000)]
New EnableEscapeCommandline ssh_config(5) option
This option (default "no") controls whether the ~C escape is available.
Turning it off by default means we will soon be able to use a stricter
default pledge(2) in the client.
feedback deraadt@ dtucker@; tested in snaps for a while
kn [Sun, 27 Nov 2022 23:18:54 +0000 (23:18 +0000)]
Move UVM "swpgonly" from Daemon Counters to Swap Counters
Put it below "swpginuse" as systat(1) and uvm_init(9) already have it.
(uvm.c struct uvmline uvmline[] makes it easy to misplace counters without
double-checking against actual systat output.)
kn [Sun, 27 Nov 2022 22:55:31 +0000 (22:55 +0000)]
Remove last queue(3) *_END() usage from tree
queue(3) NOTES says they're deprecated and expand to NULL; indeed.
No object change.
OK kettenis mvs
kn [Sun, 27 Nov 2022 22:52:21 +0000 (22:52 +0000)]
Zap nonexistent "vnodepages" and "vtextpages" UVM Page Counters
Those members exists with an XXX comment in struct uvmexp/uvm_init(9),
but the kernel does not use them at all and systat(1) does not print them.
kettenis [Sun, 27 Nov 2022 22:04:59 +0000 (22:04 +0000)]
Implement support for the (optional) MSI controller of Synopsys Designware
PCIe host bridge. This MSI controller is quite retarded since it maps all
MSIs to a single hardware interrupt. So it doesn't really offer any
benefit over using classic INTx interrupts. Unfortunately we need to use
it on Amlogic SoCs since the PCIe device interrupt doesn't seem to work
correctly when configured as a level triggered interrupt and the workaround
of configuring it as an edge triggered interrupt causes problems when using
multiple disks connected to ahci(4) on the ODROID-HC4.
ok patrick@
job [Sun, 27 Nov 2022 20:50:09 +0000 (20:50 +0000)]
BIO_puts return values can be ambiguous, improve the check
OK tb@
schwarze [Sun, 27 Nov 2022 19:11:11 +0000 (19:11 +0000)]
In bio.h rev. 1.50 and rev. 1.51, tb@ provided BIO_set_retry_reason(3).
Merge the documentation from the OpenSSL 1.1.1 branch, which is still
under a free license, tweaked by me.
kn [Sun, 27 Nov 2022 15:31:36 +0000 (15:31 +0000)]
Remove useless casts
All *dp variables are of type 'struct dadq *'; no object change.
OK mvs
kn [Sun, 27 Nov 2022 15:27:17 +0000 (15:27 +0000)]
Zap qt3/4 remnants, MODQMAKE_RECURSIVE always defaults to yes
Like that since ports devel/qmake/qmake.port.mk r1.11
Zap qt3 remnants
kn [Sun, 27 Nov 2022 15:19:38 +0000 (15:19 +0000)]
Provide more accurate lock error message
When started manuall from single-user mode (/ still read-only), the current
error is misleading:
# slaacd -dv
slaacd: already running
The lock was specifically added to prevent multiple instances in the
installer, which discards the error message entirely anyway.
Retain the useful EAGAIN/"already running" message, but otherwise print the
real error reason:
# slaacd -dv
slaacd: /dev/slaacd.lock: Read-only file system
Feedback OK deraadt millert
anton [Sun, 27 Nov 2022 15:13:15 +0000 (15:13 +0000)]
reformat REGRESS_TARGETS
anton [Sun, 27 Nov 2022 15:12:57 +0000 (15:12 +0000)]
Applying mimmutable(2) to bss, data and text made the sys/kern/noexec tests fail
since they try to mutate the permissions of the aforementioned sections which is
no longer possible. Instead, "mimic" the sections by allocating and operating on
memory with the same initial permissions.
With help from deraadt@
kn [Sun, 27 Nov 2022 15:08:16 +0000 (15:08 +0000)]
x11/qt3 and qt4 are no more, say x11/qt* like the module's error message
schwarze [Sun, 27 Nov 2022 14:35:27 +0000 (14:35 +0000)]
delete duplicate OpenBSD CVS keyword line
tb [Sun, 27 Nov 2022 14:31:22 +0000 (14:31 +0000)]
Plug leak of tmp in case allocation of pq->abuf fails
ok florian
tb [Sun, 27 Nov 2022 14:29:06 +0000 (14:29 +0000)]
KNF nit: place brace correctly
ok florian
otto [Sun, 27 Nov 2022 13:19:00 +0000 (13:19 +0000)]
Once we are synced, we can validate the certificate in the standard way.
ok tb@
tb [Sat, 26 Nov 2022 23:05:22 +0000 (23:05 +0000)]
Missing return value check for BIO_new()
kettenis [Sat, 26 Nov 2022 21:35:22 +0000 (21:35 +0000)]
Adjust for changes in the device tree bindings. Since it is relatively
easy to support both the "official" and "preliminary" bindings, we support
both. Support for the "preliminary" bindings will be removed at some point
in the future.
ok tobhe@
mvs [Sat, 26 Nov 2022 17:52:35 +0000 (17:52 +0000)]
Turn sowriteable(), sballoc() and sbfree() macro to inline functions.
soreadable() is already presented as inline function, but corresponding
sowriteable() is still macro. Also it's no reason to keep sballoc() and
sbfree() as macro.
The first argument of sballoc() and sbfree() is not used, but keep it for
a while.
ok kn@ bluhm@
mvs [Sat, 26 Nov 2022 17:51:18 +0000 (17:51 +0000)]
Merge uipc_bind() with unp_bind(). Unlike other unp_*() functions,
unp_bind() has the only uipc_bind() caller. In the uipc_usrreq() times,
it made sense to have dedicated unp_bind() for prevent tne code mess
within giant switch(), but now it doesn't.
ok bluhm@
mvs [Sat, 26 Nov 2022 17:50:26 +0000 (17:50 +0000)]
Next step of netlock pressure decreasing in pppx(4).
The kernel lock is still taken when we access pppx(4) layer through
device node. Since pipex(4) layer doesn't rely on netlock anymore, and we
don't acquire it when we access pipex(4) from pppx(4) layer, kernel lock
is enough to protect pppx(4) data. Such data doesn't accessed from packet
processing path, so there is no reason to block it by netlock acquiring.
Assume kernel lock as protection for `pxd_pxis' lists and `pppx_ifs' tree.
The search in `pppx_ifs' tree has no context switch. There is no context
switch between the `pxi' free unit search and tree insertion.
Use reference counters to make `pxi' dereference safe, instead of holding
netlock. Now pppx_if_find() returns `pxi' with reference counter bumped,
and newly introduced pppx_if_rele() used for release this `pxi'.
Introduce pppx_if_find_locked() which returns `pxi' but doesn't bump
reference counter. pppx_if_find_locked() and pppx_if_find() both called
with kernel lock held, but keep existing notation where _locked()
function returned data with non bumped counter.
Mark dying `pxi' by setting `pxi_ready' to null, so concurrent thread
can't receive it by pppx_if_find().
The netlock is left around modification of associated ifnet's
`if_description'. This is unwanted because `if_description' never accessed
within packet processing path, but this require ifnet locking
modification, so keep this to the following diffs.
ok bluhm@
tb [Sat, 26 Nov 2022 17:23:17 +0000 (17:23 +0000)]
Make header guards of internal headers consistent
Not all of them, only those that didn't leak into a public header...
Yes.
tobhe [Sat, 26 Nov 2022 17:23:15 +0000 (17:23 +0000)]
Add arm64 lid_action sysctl for Apple Silicon laptops.
ok kettenis@
job [Sat, 26 Nov 2022 17:15:44 +0000 (17:15 +0000)]
Remove RFC 9092 example file because it contains inherit elements
job [Sat, 26 Nov 2022 17:14:40 +0000 (17:14 +0000)]
Make error messages about 'inherit' elements in End-Entity certs consistent
OK tb@
tobhe [Sat, 26 Nov 2022 17:12:11 +0000 (17:12 +0000)]
Bump to 7.2
job [Sat, 26 Nov 2022 17:06:43 +0000 (17:06 +0000)]
Disallow 'inherit' elements in geofeed authenticators
RFC 9092 is underspecified in this regard, but other signed
objects relating to Internet number resources (ROA, BGPsec,
ASPA, RSC) all disallow inherit.
See https://mailarchive.ietf.org/arch/msg/opsawg/JXjxCA14BkW4DWyVoUMwqDvB17I/
OK tb@
job [Sat, 26 Nov 2022 16:42:04 +0000 (16:42 +0000)]
Fix warning message
(Geofeed authenticators don't have a SIA)
tb [Sat, 26 Nov 2022 16:11:36 +0000 (16:11 +0000)]
bn_lcl.h wanted special treatment.
tb [Sat, 26 Nov 2022 16:08:50 +0000 (16:08 +0000)]
Make internal header file names consistent
Libcrypto currently has a mess of *_lcl.h, *_locl.h, and *_local.h names
used for internal headers. Move all these headers we inherited from
OpenSSL to *_local.h, reserving the name *_internal.h for our own code.
Similarly, move dtls_locl.h and ssl_locl.h to dtls_local and ssl_local.h.
constant_time_locl.h is moved to constant_time.h since it's special.
Adjust all .c files in libcrypto, libssl and regress.
The diff is mechanical with the exception of tls13_quic.c, where
#include <ssl_locl.h> was fixed manually.
discussed with jsing,
no objection bcook
tb [Sat, 26 Nov 2022 15:45:47 +0000 (15:45 +0000)]
Two small tweaks to the geofeed code
Only allocate b64 when it is needed. This way we can avoid allocating extra
memory for the signed data itself. Also, only check for the end signature
marker when it is actually expected. It's not forbidden - if stupid - to
have a comment '# End Signature:' in the signed data.
ok job
jsing [Sat, 26 Nov 2022 13:56:33 +0000 (13:56 +0000)]
Remove BIGNUM consistency macros.
Compiling with BN_DEBUG (and if you want to take it further, BN_DEBUG_RAND)
supposedly adds consistency checks to the BN code. These are rarely if ever
used and introduce a bunch of clutter in the code. Furthermore, there are
hacks in place to undo things that the debugging code does.
Remove all of this mess and instead rely on always enabled checks, more
readable code and proper regress coverage to ensure correct behaviour.
"Good riddance." tb@
tb [Sat, 26 Nov 2022 12:36:19 +0000 (12:36 +0000)]
Split eContent extration into a small helper
job didn't like jumping over a bunch of code, so handle this via a small
helper. It's not as if cms_parse_validate_internal() could not do with a
bit of splitting in general.
ok job
tb [Sat, 26 Nov 2022 12:34:31 +0000 (12:34 +0000)]
Add test-geofeed.c (forgotten cvs add in previous).
job [Sat, 26 Nov 2022 12:09:34 +0000 (12:09 +0000)]
Add regress for geofeed
job [Sat, 26 Nov 2022 12:02:36 +0000 (12:02 +0000)]
Add support for authenticating geofeed data CSV files in filemode
RFC 9092 describes a scheme in which an authenticator is appended to a
geofeed (RFC 8805) file. It is a digest of the main body of the file
signed by the private key of the relevant RPKI certificate for a covering
address range. The authenticator is a detached CMS signature.
with and OK tb@
tb [Sat, 26 Nov 2022 11:18:49 +0000 (11:18 +0000)]
cms_lcl.h should not be part of SRCS
kettenis [Sat, 26 Nov 2022 09:05:32 +0000 (09:05 +0000)]
Enable aplpwm(4) and pwmleds(4).
jmc [Sat, 26 Nov 2022 07:26:43 +0000 (07:26 +0000)]
- in SYNOPSIS, redo the formatting for "address" and "dest address" to avoid
an ugly line split on narrower terminals
- in usage(), match the output
anton [Sat, 26 Nov 2022 06:30:08 +0000 (06:30 +0000)]
Remove unused battery fields.
anton [Sat, 26 Nov 2022 06:29:50 +0000 (06:29 +0000)]
Shove more battery feature logic down to hidpp20_battery_get_level_status().
In preparation for supporting the unified battery feature.
anton [Sat, 26 Nov 2022 06:29:24 +0000 (06:29 +0000)]
Pave the way for checking presence of more features in
uhidpp_device_features().
anton [Sat, 26 Nov 2022 06:29:07 +0000 (06:29 +0000)]
Stop checking the link status upon receiving connect notifications as
the Bolt receiver uses another bit for this which I haven't been able to
identify.
anton [Sat, 26 Nov 2022 06:28:50 +0000 (06:28 +0000)]
Group function and response defines.
anton [Sat, 26 Nov 2022 06:28:34 +0000 (06:28 +0000)]
Pass a uhidpp_device to hidpp20_battery_get_capability() and
hidpp20_battery_get_level_status().
anton [Sat, 26 Nov 2022 06:28:08 +0000 (06:28 +0000)]
Take note of the needed feature indices already in
uhipp_device_features().
anton [Sat, 26 Nov 2022 06:27:48 +0000 (06:27 +0000)]
Reduce indentation, no functional change.
anton [Sat, 26 Nov 2022 06:26:51 +0000 (06:26 +0000)]
Stop printing the device serial number during attach, it's not that
useful after all.
anton [Sat, 26 Nov 2022 06:26:14 +0000 (06:26 +0000)]
Rename type argument to buf in hidpp10_get_type(), no functional change.
anton [Sat, 26 Nov 2022 06:20:18 +0000 (06:20 +0000)]
While emulating the bell, wsdisplay could end up sleeping when reaching
down to wsmux. This does not work since we're in interrupt context, as
pointed out by witness.
Instead, defer the work to a task which in turn will execute in process
context.
Problem reported by beck@
ok visa@
deraadt [Fri, 25 Nov 2022 23:09:20 +0000 (23:09 +0000)]
ifconfig -M <mac> finds the address on an interface and prints it.
cloned (virtual) interfaces are skipped, and if the MAC is on more
than 1 interface, no answer either. The mac must be in same format
as the ifconfig lladdr output (complete lowercase with :)
idea from florian, ok afresh1
tobhe [Fri, 25 Nov 2022 20:33:11 +0000 (20:33 +0000)]
Disable screen backlight on Apple silicon laptops when lid is closed.
ok kettenis@
bluhm [Fri, 25 Nov 2022 20:27:53 +0000 (20:27 +0000)]
revert pf.c r1.1152 again: move pf_purge out from under the kernel lock
Using systqmp for pf_purge creates a deadlock between pf_purge()
and ixgbe_stop() and possibly other drivers. On systqmp pf(4) needs
netlock which the interface ioctl(2) is holding. ix(4) waits in
sched_barrier() which is also scheduled on the systqmp task queue.
Removing the netlock from pf_purge() as a quick fix caused other
problems.
backout suggested by deraadt@
kettenis [Fri, 25 Nov 2022 18:03:53 +0000 (18:03 +0000)]
Revert previous commit. It was not properly tested and produces splassert
warnings. Rushing to pile more stuff on top of it isn't the answer. This
needs a rethink.
ok deraadt@
schwarze [Fri, 25 Nov 2022 17:44:01 +0000 (17:44 +0000)]
In bio.h rev. 1.46/1.47 (Oct/Nov 2021), tb@ provided BIO_get_init(3).
Document it.
bluhm [Fri, 25 Nov 2022 16:10:07 +0000 (16:10 +0000)]
Do not crash when a tcp query is larger than the length field
indicated.
Found by kn with amap.
Input bluhm.
OK deraadt, tb, otto, kn
from florian@
kn [Fri, 25 Nov 2022 15:03:24 +0000 (15:03 +0000)]
Revert hunk accidentially committed in r1.248 "Clarify/typofix comments"
visa [Fri, 25 Nov 2022 14:56:56 +0000 (14:56 +0000)]
Add ld.so linker script for mips64
Since the introduction of automatic immutable from the kernel, the munmap()
of ld.so boot.text region is now (silently) failing because the region is
contained within the text LOAD, which is immutable. So create a new btext
LOAD with flags PF_X|PF_R|PF_OPENBSD_MUTABLE, and place all boot.text objects
in there. This LOAD must also be page-aligned so it doesn't skip unmapping
some of the object region, previously it was hilariously unaligned.
OK deraadt@
tb [Fri, 25 Nov 2022 09:32:10 +0000 (09:32 +0000)]
Units generally help...
dlg [Fri, 25 Nov 2022 03:45:39 +0000 (03:45 +0000)]
get rid of NET_LOCK in the pf purge work
pf purge was moved to systqmp (to get it away from KERNEL_LOCK)
which is also used as the backend for things like intr_barrier and
sched_barrier. it is common for network cards to call intr_barrier
while holding NET_LOCK, and if pf is trying to get the NET_LOCK in
the purge tasks that are now running in systqmp, it's a deadlock.
bluhm@ hit this exact issue.
sashan@ has been working to get rid of the need for NET_LOCK in pf,
so now we can remove the NET_LOCKs here rather than create a pf
specific taskq to run these tasks in.
ok sashan@ bluhm@
dlg [Fri, 25 Nov 2022 03:20:09 +0000 (03:20 +0000)]
nvme can do 64bit dma.
ok jmatthew@
schwarze [Thu, 24 Nov 2022 19:06:38 +0000 (19:06 +0000)]
Major overhaul.
Remove many statements that are no longer true after tb@, in July,
massively improved the algorithms used by these functions
and also did some cleanup of the interface. Instead, explain
many aspects that were missing. Also use more descriptive argument
names, drop some redundancy, and improve ordering in various respects.
Feedback and enthusiastic OK from tb@.
schwarze [Thu, 24 Nov 2022 18:11:26 +0000 (18:11 +0000)]
Mark BN_options() and BN_prime_checks as obsolete;
it appears that all BN public symbols are now documented,
except those intentionally undocumented.
deraadt [Thu, 24 Nov 2022 17:44:35 +0000 (17:44 +0000)]
sycn
tb [Thu, 24 Nov 2022 16:35:32 +0000 (16:35 +0000)]
Merge the second y_bit check into the first one where it belongs
suggested by jsing
tb [Thu, 24 Nov 2022 16:34:13 +0000 (16:34 +0000)]
Simplify y_bit handling in compressed coordinates
If y_bit is set for a zero y, something is wrong and we can error directly.
No need to run the non-trivial BN_kronecker() to check if BN_mod_sqrt()
lied or not, only to set a more specific error code.
ok jsing
tb [Thu, 24 Nov 2022 16:29:09 +0000 (16:29 +0000)]
Clean up EC_METHOD and EC_GROUP definitions
Remove obvious comments, wrap long lines and general KNF cleanup. Format
and rephrase the more important comments.
Discussed with jsing
kettenis [Thu, 24 Nov 2022 14:43:16 +0000 (14:43 +0000)]
Expose the complete set of ID registers as defined in the current version
of ARMv8/ARMv9. Make sure we only expose the features that we know about
and support in our kernel. This matches what Linux does. For now, mostly
restrict ourselves to features defined in ARMv8.5 which means that we only
actually implement support for ID_AA64ISAR0_EL1, ID_AA64ISAR1_EL1,
ID_AA64PFR0_EL1 and ID_AA64PFR1_EL1. For the other registers we simply
always return 0.
ok deraadt@
kettenis [Thu, 24 Nov 2022 14:36:07 +0000 (14:36 +0000)]
Document a few more ID register bits. This should add most of the stuff
in ARMv8.5 as far as the ISAR and PFR registers are concerned.
ok deraadt@
jmatthew [Thu, 24 Nov 2022 04:04:39 +0000 (04:04 +0000)]
Don't report unknown apic structure types if they're in the OEM reserved
range (0x80 to 0xff). We're not going to understand OEM specific
structures, and some machines such as the Dell R6515 have significant
numbers of them.
ok kettenis@
jsing [Thu, 24 Nov 2022 01:30:01 +0000 (01:30 +0000)]
Change bn_expand()/bn_wexpand() to indicate failure/success via 0/1.
Currently bn_expand()/bn_wexpand() return a BIGNUM *, however none of the
callers use this (and many already treat it as a true/false value).
Change these functions to return 0 on failure and 1 on success, revising
callers that test against NULL in the process.
ok tb@
jsing [Thu, 24 Nov 2022 01:24:37 +0000 (01:24 +0000)]
Call bn_expand() rather than handrolling an equivalent.
The current code manually calculates words from bits and then calls
bn_wexpand() - call bn_expand() with bits instead.
ok tb@
mvs [Thu, 24 Nov 2022 00:04:32 +0000 (00:04 +0000)]
Remove netlock assertion within PF_LOCK(). The netlock should be taken
first, but only if both locks are taken.
ok dlg@ sashan@
kettenis [Wed, 23 Nov 2022 23:47:05 +0000 (23:47 +0000)]
pwmleds(4)
kettenis [Wed, 23 Nov 2022 23:43:08 +0000 (23:43 +0000)]
Add pwmleds(4), a driver for PWM controlled LEDs.
For now this only implements keyboard backlight support.
ok kn@
kn [Wed, 23 Nov 2022 23:06:16 +0000 (23:06 +0000)]
Fix sparc64 build/run
constraints.c:269: warning: ISO C90 forbids mixed declarations and code
from tb
tb [Wed, 23 Nov 2022 22:52:25 +0000 (22:52 +0000)]
Add void casts since gcc 4.2.1 on sparc64 doesn't like the missing return
checks for BIO_reset().
kn [Wed, 23 Nov 2022 19:35:42 +0000 (19:35 +0000)]
Remove unused NDPRF_* defines; dead since 2017 sys/netinet6/nd6.c r1.210
Remove sending of router solicitations and processing of router
advertisements from the kernel. It's handled by slaacd(8) these days.
Outside of _KERNEL, but nothing in base uses it, either.
codesearch.debian.net seems to agree.
OK mvs claudio bluhm
kn [Wed, 23 Nov 2022 19:35:18 +0000 (19:35 +0000)]
Remove unused struct in6_ndifreq; dead since 2013 sys/netinet6/in6_var.h r1.37
Remove unused code manipulating a default interface and its index
This is a leftover from the on-link assumption behavior removal,
which has been deprecated by RFC4861 anyway.
Outside of _KERNEL, but nothing in base uses it, either.
codesearch.debian.net seems to agree.
OK mvs claudio bluhm
kn [Wed, 23 Nov 2022 19:34:59 +0000 (19:34 +0000)]
Remove unused struct prf_ra; dead since 2017 sys/netinet/icmp6.h r1.45
Get rid of ICMPV6CTL_ND6_DRLIST and ICMPV6CTL_ND6_PRLIST sysctls
With this we can also get rid of in6_prefix and in6_defrouter. They
are meaningless, the kernel no longer tracks this information.
Outside of _KERNEL, but nothing in base uses it, either.
codesearch.debian.net seems to agree.
OK mvs claudio bluhm
schwarze [Wed, 23 Nov 2022 17:20:23 +0000 (17:20 +0000)]
Several improvements required for <openssl/bn.h>:
* List internal constants and types that are intentionally undocumented.
* List unused constants and types that are intentionally undocumented.
* Cope with intentionally undocumented identifiers being declared
more than once (in this case, because of #if and #else).
* Require exact matches for man -k searches (in this case,
such that BN_BITS does not match BN_BITS2).
* Handle the weird BN_ULONG, which is #define'd instead of using typedef.
kn [Wed, 23 Nov 2022 16:59:10 +0000 (16:59 +0000)]
Inline useless ND_IFINFO() macro
A single cast-free struct pointer dereference needs no indirection.
ND_IFINFO() is under _KERNEL.
OK mvs
kn [Wed, 23 Nov 2022 16:57:37 +0000 (16:57 +0000)]
Let nd6_if{at,de}tach() be void and take an ifp argument
Do it like the rest of at/detach routines which modify a struct ifnet
pointer without returning anything.
OK mvs
tb [Wed, 23 Nov 2022 15:52:43 +0000 (15:52 +0000)]
Make a stupid compiler on a stupid OS happy.
from bcook
mvs [Wed, 23 Nov 2022 15:12:27 +0000 (15:12 +0000)]
Make `so' dereference safe within pflow_output_process().
sosend() has sleep points, so the kernel lock serialisation within
pflow_output_process() doesn't work as expected. The pflow(4) interface
associated socket `so' could be overwritten by concurrent pflowioctl()
thread.
Introduce pflow(4) interface's `sc_lock' rwlock(9) to make `so'
dereference safe. Since the solock() of udp(4) sockets uses netlock as
backend, the `sc_lock' should be taken first. This expands a little
netlock relocking within pflowioctl().
pflow_sendout_mbuf() called by pflow_output_process(), now called
without kernel lock held, so the mp safe counters_pkt(9) used instead
of manual `if_opackets' increment.
Since if_detach() does partial ifnet destruction, now it can't be called
before we finish pflow_output_process() task, otherwise we introduce use
after free for interface counters. In other hand, we need to deny
pflowioctl() to reschedule pflow_output_process() task. The `sc_dyind'
flag introduced for that.
Tested by Hrvoje Popovski.
ok bluhm@
kn [Wed, 23 Nov 2022 14:50:59 +0000 (14:50 +0000)]
Remove unused struct ifnet's *if_afdata[] and struct domain's dom_if{at,de}tach()
Both made obsolete through struct ifnet's previous *if_nd addition.
IPv6 Neighbour Discovery handles per-interface data directly, nothing
else uses this generic domain API anymore.
Outside of _KERNEL, but nothing in base uses them, either.
OK bluhm mvs claudio
kn [Wed, 23 Nov 2022 14:48:27 +0000 (14:48 +0000)]
Add *if_nd to struct ifnet, call nd6_if{at,de}tach() directly
*if_afdata[] and struct domain's dom_if{at,de}tach() are only used with
IPv6 Neighbour Discovery in6_dom{at,de}tach(), which allocate/init and
free single struct nd_ifinfo.
Set up a new ND-specific *if_nd member directly to avoid yet another
layer of indirection and thus make the generic domain API obsolete.
The per-interface data is only accessed in nd6.c and nd6_nbr.c through
the ND_IFINFO() macro; it is allocated and freed exactly once during
interface at/detach, so document it as [I]mmutable.
OK bluhm mvs claudio
kn [Wed, 23 Nov 2022 14:22:17 +0000 (14:22 +0000)]
iked(8) is proposing as well
Since sbin/iked/vroute.c r1.13 (sep 2021):
Add client side support for DNS configuration. Use RTM_PROPOSAL_STATIC
route messages to propose the name server to resolvd(8).
mbuhl [Wed, 23 Nov 2022 11:00:27 +0000 (11:00 +0000)]
cache ps_auxinfo inside the kernel, to avoid codedump() reading the
copy on userland stack which points at an illicit region.
ok kettenis, deraadt
tb [Wed, 23 Nov 2022 08:58:34 +0000 (08:58 +0000)]
bn_unit: appease coverity
Apparently, the '0' in memset(a, '0', size - 1); could be a typo for '\0'.
Randomize the decimal digit to make the intent clear.
CID 377009
tb [Wed, 23 Nov 2022 08:51:05 +0000 (08:51 +0000)]
asn1_string_to_utf8 test: appease coverity
Check for ASN_STRING_to_UTF8() failure before checking it matches our
expectations. This should convey clearly that test->want_len is never
negative.
CID 377011
Diagnosed by jsing
kn [Wed, 23 Nov 2022 08:05:49 +0000 (08:05 +0000)]
ND_IFINFO() cannot be NULL, use inline read-only idiom for clarity
ND_IFINFO() always points at a valid struct nd_ifinfo; ND6_LLINFO_DELAY
checks for NULL, while other cases in nd6_llinfo_timer() dereference it
unconditionally.
Inline all three per-case read-only usages rather than having one hoisted
*ndi pointer which could be used to write.
nd6_nbr.c already uses this `ND_IFINFO(ifp)->retrans' idiom which makes it
immediately clear that data is only read.
OK bluhm
tb [Wed, 23 Nov 2022 08:01:05 +0000 (08:01 +0000)]
Neuter getrlimit dance, it's not portable enough. Stupid Windows.
kn [Wed, 23 Nov 2022 07:57:39 +0000 (07:57 +0000)]
Recommit previous "Remove useless struct in6_ifextra"
This was the right diff after all, I just confused myself between trees.
OK bluhm
---
Remove useless struct in6_ifextra
in6_var.h r1.75 removed all other struct members.
Now It only contains a single struct nd_ifinfo pointer, so address family
specific data might as well be just that.
ND_IFINFO() is the only way nd6_nbr.c and nd6.c access this data, there is
no other usage of if_afdata[].
One allocation and unhelpful indirection less per interface.
All under _KERNEL.
OK claudio