visa [Thu, 17 Mar 2022 14:22:03 +0000 (14:22 +0000)]
Use the refcnt API in bpf.
OK sashan@ bluhm@
stsp [Thu, 17 Mar 2022 14:00:53 +0000 (14:00 +0000)]
Make tcpdump(8) show 802.11ac VHT capability and operation IEs in -v mode.
ok sthen@
nicm [Thu, 17 Mar 2022 13:39:13 +0000 (13:39 +0000)]
Check scroll-on-clear for ED also.
nicm [Thu, 17 Mar 2022 11:35:37 +0000 (11:35 +0000)]
Add an option (scroll-on-clear) to control if tmux scrolls into history
on clear, from Robert Lange in GitHub issue 3121.
mpi [Thu, 17 Mar 2022 10:15:13 +0000 (10:15 +0000)]
In swap_io() allocate the buffer before doing encryption.
If the allocation fails due to memory pressure no time is wasted doing
encryption. This also simplify the error path.
Tested by sthen@.
ok kn@, miod@, kettenis@, tb@
kettenis [Thu, 17 Mar 2022 00:28:29 +0000 (00:28 +0000)]
Fix previous commit; the FDT header is big-endian so we need to do the
appropriate byte swapping.
schwarze [Wed, 16 Mar 2022 23:23:24 +0000 (23:23 +0000)]
Just say that the databases are intended for use by apropos(1), whatis(1),
and man(1), without restricting that statement to "man -k".
Suggested by and patch OK'ed by jmc@.
While only apropos(1) and whatis(1) strictly require the database
and while our man(1) implementation can find many manual pages even
when no database is available or when the database is incomplete or
corrupt, it does use the database even without -k whenever possible.
Consequently, this change makes the manual page less confusing.
kettenis [Wed, 16 Mar 2022 22:32:50 +0000 (22:32 +0000)]
When we allocate space a buffer with some extra space for the FDT, adjust
the size of the FDT to reflect the size of that buffer. This prevents
an FDT overflow if the original FDT doesn't have enough space for the
additional properties that we add to it in our bootloader.
Fixes boot on the mcbin.
tested by bluhm@, ok patrick@
tobhe [Wed, 16 Mar 2022 18:17:20 +0000 (18:17 +0000)]
Make sure contents of vroute messages are aligned properly.
Fixes address autoconfiguration on octeon.
Found by and ok mbuhl@
krw [Wed, 16 Mar 2022 17:03:13 +0000 (17:03 +0000)]
The 'status' parameter to uuid_* functions is uint32_t, not int.
No intentional functional change.
nicm [Wed, 16 Mar 2022 17:00:17 +0000 (17:00 +0000)]
Add an option to set the character used for unused areas of the
terminal, GitHub issue 3110.
visa [Wed, 16 Mar 2022 16:17:46 +0000 (16:17 +0000)]
Remove an unneeded include.
visa [Wed, 16 Mar 2022 14:38:43 +0000 (14:38 +0000)]
Use the refcnt API in kqueue.
OK dlg@ bluhm@
visa [Wed, 16 Mar 2022 14:13:01 +0000 (14:13 +0000)]
Add refcnt_shared() and refcnt_read()
refcnt_shared() checks whether the object has multiple references.
When refcnt_shared() returns zero, the caller is the only reference
holder.
refcnt_read() returns a snapshot of the counter value.
refcnt_shared() suggested by dlg@.
OK dlg@ mvs@
bluhm [Wed, 16 Mar 2022 12:37:44 +0000 (12:37 +0000)]
Fix dependency to build bn_mod_sqrt.
tb [Wed, 16 Mar 2022 11:44:36 +0000 (11:44 +0000)]
Make gcc 4 happier about x509_addr.c
gcc 4 on sparc64 issues a few 'warning: value computed is not used'.
There are two cases: sk_set_cmp_function() returns the old comparison
function of the stack which we don't care about. The one warning about
an sk_delete() is about a return value that we know already and which
we will free a few lines down.
ok inoguchi miod
florian [Wed, 16 Mar 2022 10:14:50 +0000 (10:14 +0000)]
Update to nsd 4.4.0
tested by sthen and me
OK sthen
bcook [Tue, 15 Mar 2022 21:15:08 +0000 (21:15 +0000)]
LibreSSL 3.5.2
jsing [Tue, 15 Mar 2022 18:47:22 +0000 (18:47 +0000)]
Initialise *out_name at the start of i2t_ASN1_OBJECT_name().
ok tb@
kettenis [Tue, 15 Mar 2022 18:46:15 +0000 (18:46 +0000)]
Implement additional error checking for aplsmc_read_key() such that we can
detect when we read a key that isn't supported by the firmware.
Only provide RTC functionality if the "CLKM" key is implemented.
Fixes reading the time on machines with old SMC firmware from macOS 11.x.
ok jsg@
tb [Tue, 15 Mar 2022 18:29:12 +0000 (18:29 +0000)]
Check BIO_reset() return value to make gcc happy.
krw [Tue, 15 Mar 2022 17:59:39 +0000 (17:59 +0000)]
Always use letoh32(gh.gh_part_num) instead of NGPTPARTITIONS when
scanning the GPT partition table. gh.gh_part_num reflects the actual
size of the table read from/written to the disk and scanning
possible entries between this value and the maximum allowed size
of a partition table is pointless.
No intentional functional change.
solene [Tue, 15 Mar 2022 17:13:50 +0000 (17:13 +0000)]
login class routing table should be honored when
doing a full login with su -l
ok millert@
patch from Matthew Martin < phy1729 at gmail dot com >
claudio [Tue, 15 Mar 2022 16:50:29 +0000 (16:50 +0000)]
Replace the eor member of struct prefix with a flag. Saves a byte that
will be reused soon.
OK denis@ tb@
tb [Tue, 15 Mar 2022 16:28:42 +0000 (16:28 +0000)]
Add a simple regress to verify that the infinite loop in BN_mod_sqrt()
is fixed.
tb [Tue, 15 Mar 2022 15:52:39 +0000 (15:52 +0000)]
Fix infinite loop in BN_mod_sqrt()
A bug in the implementation of the Tonelli-Shanks algorithm can lead to
an infinite loop. This loop can be hit in various ways, in particular on
decompressing an elliptic curve public key via EC_POINT_oct2point() - to
do this, one must solve y^2 = x^3 + ax + b for y, given x.
If a certificate uses explicit encoding for elliptic curve parameters,
this operation needs to be done during certificate verification, leading
to a DoS. In particular, everything dealing with untrusted certificates
is affected, notably TLS servers explicitly configured to request
client certificates (httpd, smtpd, various VPN implementations, ...).
Ordinary TLS servers do not consume untrusted certificates.
The problem is that we cannot assume that x^3 + ax + b is actually a
square on untrusted input and neither can we assume that the modulus
p is a prime. Ensuring that p is a prime is too expensive (it would
likely itself lead to a DoS). To avoid the infinite loop, fix the logic
to be more resilient and explicitly limit the number of iterations that
can be done. The bug is such that the infinite loop can also be hit for
primes = 3 (mod 4) but fortunately that case is optimized earlier.
It's also worth noting that there is a size bound on the field size
enforced via OPENSSL_ECC_MAX_FIELD_BITS (= 661), which help mitigate
further DoS vectors in presence of this fix.
Reported by Tavis Ormandy and David Benjamin, Google
Patch based on the fixes by David Benjamin and Tomas Mraz, OpenSSL
ok beck inoguchi
espie [Tue, 15 Mar 2022 14:53:53 +0000 (14:53 +0000)]
document noconfigurejunk, assuming it works as expected
claudio [Tue, 15 Mar 2022 14:39:34 +0000 (14:39 +0000)]
Refactor prefix_adjout_update(). Move the new prefix logic up and
then just fall through the update path. This is more in line with
prefix_update() and will make further work simpler.
OK tb@
tobhe [Tue, 15 Mar 2022 13:07:33 +0000 (13:07 +0000)]
Fix build after recent IKEv2 message fragment retransmit change.
Found by anton@
sthen [Tue, 15 Mar 2022 13:00:30 +0000 (13:00 +0000)]
sync
jan [Tue, 15 Mar 2022 11:22:10 +0000 (11:22 +0000)]
Enable IP header checksum offloading in ix(4).
ok jmatthew@
claudio [Tue, 15 Mar 2022 11:16:25 +0000 (11:16 +0000)]
Add a test for expanding macros in large-community strings.
claudio [Tue, 15 Mar 2022 11:13:48 +0000 (11:13 +0000)]
Change how $macros are expanded in the config.
Expand $macros not only at the start of a yacc token but also inside STRING
elements. STRING elements are used e.g. for community specifications and
it makes sense to allow $FOO:$BAR to correctly expand. There is no expansion
of macros in quoted strings (both single and double quotes).
Factor out the macro expand logic and with this introduce its own lookup
buffer for the macro name. For expansion to work inside STRING the char
after the makro name must be a character not allowed in macro names (not
alpha-numerical or '_').
Add extra checks to set variables. Mainly restrict length of the name and
also make sure it does not include not allowed characters.
OK tb@
stsp [Tue, 15 Mar 2022 09:23:01 +0000 (09:23 +0000)]
Fix Tx rate used by rtwn(4) and urtwn(4) for RTS frames.
Using ni_txrate for RTS is a bad choice since it could go up to 54 Mbit/s.
The AP needs to receive our RTS frame reliably. Usually, 1 Mbit/s is used
for RTS, but this hurts throughput and does not really make sense on today's
wifi networks.
Use 1 Mbit/s in 11b mode, and otherwise use 24 Mbit/s, as damien@
already hard-coded in urtwn long ago.
ok kevlo@
stsp [Tue, 15 Mar 2022 09:22:56 +0000 (09:22 +0000)]
Fix Tx rate used by rtwn(4) and urtwn(4) for RTS frames.
Using ni_txrate for RTS is a bad choice since it could go up to 54 Mbit/s.
The AP needs to receive our RTS frame reliably. Usually, 1 Mbit/s is used
for RTS, but this hurts throughput and does not really make sense on today's
wifi networks.
Use 1 Mbit/s in 11b mode, and otherwise use 24 Mbit/s, as damien@
already hard-coded in urtwn long ago.
ok kevlo@
jsg [Tue, 15 Mar 2022 08:15:23 +0000 (08:15 +0000)]
remove if_vxlan.h unused after if_vxlan.c rev 1.84
ok kevlo@ claudio@
espie [Tue, 15 Mar 2022 08:12:53 +0000 (08:12 +0000)]
if we pass exact pkg args, just check that their files are there instead
of the full pkgtree, useful for debug
ratchov [Tue, 15 Mar 2022 05:53:37 +0000 (05:53 +0000)]
Fix crash caused by confusion between requested device and last used one
Found by and ok semarie@
djm [Tue, 15 Mar 2022 05:27:37 +0000 (05:27 +0000)]
improve DEBUG_CHANNEL_POLL debugging message
jmatthew [Tue, 15 Mar 2022 02:07:21 +0000 (02:07 +0000)]
Enable checksum offloads. The nic does all the work here, so we don't
need to calculate header offsets for it.
ok dlg@
dlg [Mon, 14 Mar 2022 23:41:42 +0000 (23:41 +0000)]
unload the dmamap in bnxt_dmamem_free.
this is technically not necessary, but it makes it feel symmetrical
with bnxt_dmamem_alloc which loads it.
ok jmatthew@
tb [Mon, 14 Mar 2022 22:38:43 +0000 (22:38 +0000)]
Unbreak the tree, revert commitid aZ8fm4iaUnTCc0ul
This reverts the commit protecting the list and hashes in the PCB tables
with a mutex since the build of sysctl(8) breaks, as found by kettenis.
ok sthen
solene [Mon, 14 Mar 2022 21:52:08 +0000 (21:52 +0000)]
add -k flag to gzip and gunzip
When using this flag, the input file is kept after
compression or decompression, this makes our gzip
more compatible with the other gzip
changes reworked by jca@, thanks
ok jca@ millert@
tb [Mon, 14 Mar 2022 21:30:48 +0000 (21:30 +0000)]
Add a few regress test cases for name constraints.
From Alex Wilson
tb [Mon, 14 Mar 2022 21:29:46 +0000 (21:29 +0000)]
Allow constraints of the form @domain.com
Some things issue and expect that we support a non-standard extension of
accepting any email address from a host by prefixing an email name
constraint with @. This used to be the case with the old code as well.
Pointed out and based on a diff by Alex Wilson.
ok jsing
tb [Mon, 14 Mar 2022 21:15:49 +0000 (21:15 +0000)]
Rework ownership handling in x509_constraints_validate()
Instead of having the caller allocate and pass in a new
x509_constraints_name struct, handle allocation inside
x509_constraints_validate(). Also make the error optional.
All this is done to simplify the call sites and to make it
more obvious that there are no leaks.
ok jsing
miod [Mon, 14 Mar 2022 20:07:57 +0000 (20:07 +0000)]
There is no objectionable-C compiler in the comp set anymore.
kettenis [Mon, 14 Mar 2022 19:09:32 +0000 (19:09 +0000)]
The current FDT code we use in the bootloader is buggy and will write into
memory beyond the actual FDT data structure when adding information to the
device tree. This is especially problematic on ACPI systems where we add
lots of information to the device tree based on ACPI tables. Fix the FDT
code to never write beyond the end of the data structure and panic if we
run out of free space. Raise the amount of free space frm 4K to 16K for
the proto-FDT we use on ACPI systems. Bump the version number of the
arm64 bootloader.
ok visa@, patrick@
bluhm [Mon, 14 Mar 2022 17:23:00 +0000 (17:23 +0000)]
pf_socket_lookup() calls in_pcbhashlookup() in the PCB layer. To
run pf in parallel, make parts of the stack MP safe. Protect the
list and hashes in the PCB tables with a mutex.
Note that the protocol notify functions may call pf via tcp_output().
As the pf lock is a sleeping rw_lock, we must not hold a mutex. To
solve this for now, collect these PCBs in inp_notify list and protect
it with exclusive netlock.
OK sashan@
krw [Mon, 14 Mar 2022 17:11:44 +0000 (17:11 +0000)]
Abstract duplicated code scanning gpt_types[] into a helper
function find_gpt_type().
Use find_gpt_type() to simplify the functions obtaining
information from gpt_types[]. Add not yet used
PRT_uuid_to_protection() to allow simplification of GPT partition
protection code..
No intentional functional change.
jsing [Mon, 14 Mar 2022 16:49:35 +0000 (16:49 +0000)]
Factor out unexpected handshake message handling code in the legacy stack.
The TLS record layer has to be able to handle unexpected handshake messages
that result when it has been asked to read application data. The way that
this is currently done in the legacy stack is a layering violation - the
record layer knows about DTLS/TLS handshake messages, parsing them and then
deciding what action to take. This is further complicated by the need to
handle handshake message fragments.
For now, factor this code out with minimal changes - since it is a layering
violation we have to retain separate code for DTLS and TLS.
ok beck@ inoguchi@ tb@
jsing [Mon, 14 Mar 2022 16:35:45 +0000 (16:35 +0000)]
Factor out ASN1_STRING clearing code.
This fixes a bug in ASN1_STRING_set0() where it does not respect the
ASN1_STRING_FLAG_NDEF flag and potentially frees memory that we do not own.
ok inguchi@ tb@
jsing [Mon, 14 Mar 2022 16:23:29 +0000 (16:23 +0000)]
First pass clean up of ASN1_STRING code.
Use consistent variable names (astr/src) rather than 'a', 'bs', 'str', 'v'
or 'x', add some whitespace and remove some unneeded parentheses.
ok inoguchi@ tb@
stsp [Mon, 14 Mar 2022 15:17:17 +0000 (15:17 +0000)]
update iwx(4) man page CAVEATS section in light of 11ac support
stsp [Mon, 14 Mar 2022 15:08:50 +0000 (15:08 +0000)]
Add initial support for 802.11ac (VHT) to the iwx(4) driver.
This makes it possible to use 80MHz channels and VHT-specific MCS.
Other 11ac features remain disabled for now.
Tested:
ax200: Matthias Schmidt, phessler, dv, kevlo, Joel Carnat, hrvoje, jmc, stsp
ax201: mlarkin, stsp
iwm (regression testing): stsp
stsp [Mon, 14 Mar 2022 15:07:24 +0000 (15:07 +0000)]
Add initial 802.11ac (VHT) support to net80211.
Add VHT capability and operation IE definitions to ieee80211.h.
Introduce channel flags to identify 80MHz and 160MHz capable channels.
Parse VHT IEs in beacons, announce the driver's VHT capabilities in
probe requests and assoc requests, and hop into 11ac mode after
association to the AP if possible.
Enable VHT by default if the driver announces support for it.
ok claudio@
stsp [Mon, 14 Mar 2022 15:06:04 +0000 (15:06 +0000)]
sync ieee80211 channel flags over to radiotap
krw [Mon, 14 Mar 2022 14:31:23 +0000 (14:31 +0000)]
And the prototype too.
krw [Mon, 14 Mar 2022 14:28:58 +0000 (14:28 +0000)]
PRT_uuid_to_typename() can return const char *.
krw [Mon, 14 Mar 2022 14:22:32 +0000 (14:22 +0000)]
Nuke extraneous blank line.
stsp [Mon, 14 Mar 2022 14:15:33 +0000 (14:15 +0000)]
Fix Tx performance on urtwn(4) RTL8192EU devices.
The "driver rate" bit which tells firmware to use the driver-provided
Tx rate is in a different spot of Tx descriptor double word 3.
Fix the corresponding macro definition.
The device was sending all frames at CCK1 because of this bug, regardless
of the Tx rate the driver had chosen.
tcpbench before:
Conn: 1 Mbps: 0.808 Peak Mbps: 1.108 Avg Mbps: 0.808
tcpbench with fix:
Conn: 1 Mbps: 21.078 Peak Mbps: 21.078 Avg Mbps: 21.078
ok jmatthew@
kettenis [Mon, 14 Mar 2022 13:08:32 +0000 (13:08 +0000)]
Make sure the apldart(4) implementation of bus_dmamap_destroy(9) does the
equivalent of bus_dmamap_unload(9) if active mappings exist. This fixes
the kerenal panics seen with bringing bwfm(4) down and up again.
ok jsg@, patrick@
tobhe [Mon, 14 Mar 2022 12:58:55 +0000 (12:58 +0000)]
Improve retransmission of message fragments. RFC 7383 states that loss of
a single fragment results in a retransmit of all fragments belonging
to the same message. Instead of treating each fragment as message with
seperate retransmit timer, keep only a single timer for all fragments of
a message and retransmit all fragments in order on timeout.
Improves reliability in case of packet loss when fragmentation is enabled.
Found by and diff from Daniel Herzinger
ok patrick@
sthen [Mon, 14 Mar 2022 10:58:35 +0000 (10:58 +0000)]
minor lint fixes
sthen [Mon, 14 Mar 2022 10:56:55 +0000 (10:56 +0000)]
document changes to MODPERL_REGEN_PPPORT
bluhm [Mon, 14 Mar 2022 00:51:27 +0000 (00:51 +0000)]
Increase timeout to make test pass on sparc64.
cheloha [Sun, 13 Mar 2022 23:27:54 +0000 (23:27 +0000)]
ssh: xstrdup(): use memcpy(3)
Copying the given string into the buffer with strlcpy(3) confers no
benefit in this context because we have already determined the
string's length with strlen(3) in order to allocate that buffer.
Thread: https://marc.info/?l=openbsd-tech&m=
164687525802691&w=2
ok dtucker@ millert@
bluhm [Sun, 13 Mar 2022 22:16:59 +0000 (22:16 +0000)]
Document membar_enter_after_atomic and membar_exit_before_atomic.
OK kettenis@
bluhm [Sun, 13 Mar 2022 21:38:32 +0000 (21:38 +0000)]
Hrvoje has hit a crash with IPsec acquire while testing the parallel
IP forwarding diff. Add mutex and refcount to make memory management
of struct ipsec_acquire MP safe.
testing Hrvoje Popovski; input sashan@; OK mvs@
kettenis [Sun, 13 Mar 2022 21:17:52 +0000 (21:17 +0000)]
Revert previous commit. It breaks the serial console on my rpi4.
tb [Sun, 13 Mar 2022 17:23:02 +0000 (17:23 +0000)]
Relax the check of x509_constraints_dirname()
The dirname constraint must be a prefix in DER format, so relax the
check from requiring equal-length strings to allow shorter names also.
From Alex Wilson
ok jsing
tb [Sun, 13 Mar 2022 17:08:04 +0000 (17:08 +0000)]
Add x509_constraints_validate() to x509_internal.h
From Alex Wilson
ok jsing
tb [Sun, 13 Mar 2022 16:48:49 +0000 (16:48 +0000)]
Check name constraints using the proper API
The previous versions were too strict and disallowed leading dots.
From Alex Wilson
ok jsing
tb [Sun, 13 Mar 2022 16:30:31 +0000 (16:30 +0000)]
style tweak
tb [Sun, 13 Mar 2022 16:25:58 +0000 (16:25 +0000)]
Add missing error check after strdup()
From Alex Wilson
ok jsing
florian [Sun, 13 Mar 2022 15:14:01 +0000 (15:14 +0000)]
parse_packet() is used by unbound to parse response packets, not
queries. There is no need to do all this work just to get access to
the query id and flags.
OK bket, sthen
jsing [Sun, 13 Mar 2022 14:58:14 +0000 (14:58 +0000)]
Remove free_cont from asn1_d2i_ex_primitive()/asn1_ex_c2i().
The constructed ASN.1 handling in asn1_d2i_ex_primitive() and asn1_ex_c2i()
currently has code to potentially avoid a malloc/memcpy - this is a less
common code path and it introduces a bunch of complexity for minimal gain.
In particular, we're manually adding a trailing NUL when ASN1_STRING_set()
would already do that for us, plus we currently manually free() the data on
an ASN1_STRING, rather than using freezero().
ok inoguchi@ tb@
espie [Sun, 13 Mar 2022 14:39:56 +0000 (14:39 +0000)]
better without warnings
kettenis [Sun, 13 Mar 2022 14:33:46 +0000 (14:33 +0000)]
Now that fdisk(8) retains the EFI System Partition on Apple systems, skip
formatting the partition such that its contents actually survive.
ok krw@
espie [Sun, 13 Mar 2022 13:42:04 +0000 (13:42 +0000)]
show the realname in verbose mode for tieing.
if we're "fixing" a partial-* package, we want to see exactly what filename
we're tieing up to, concretely, not the "theoretic" filename
mpi [Sun, 13 Mar 2022 13:34:54 +0000 (13:34 +0000)]
Constify struct cfattach.
ok miod@
espie [Sun, 13 Mar 2022 13:28:09 +0000 (13:28 +0000)]
make those checks less chatty... once the file is not there, we don't need
to say anything more
krw [Sun, 13 Mar 2022 12:57:29 +0000 (12:57 +0000)]
Missing 'or'.
Pointed out by jmc@
mpi [Sun, 13 Mar 2022 12:33:01 +0000 (12:33 +0000)]
Constify struct cfattach.
ok miod@
jmatthew [Sun, 13 Mar 2022 10:13:54 +0000 (10:13 +0000)]
Enable vlan promisc and header stripping, and use vlan rx/tx offloads.
The vlan tx information takes up an extra slot on the tx ring, so this
affects the point at which we stop putting packets on the ring too.
ok dlg@
mpi [Sun, 13 Mar 2022 08:04:38 +0000 (08:04 +0000)]
Constify struct cfattach.
ok miod@
mpi [Sun, 13 Mar 2022 08:04:13 +0000 (08:04 +0000)]
Constify struct cfattach.
ok miod@
jmatthew [Sat, 12 Mar 2022 23:54:53 +0000 (23:54 +0000)]
Transmit segmented packets using multiple tx ring slots rather than
defragging them first.
ok dlg@
mpi [Sat, 12 Mar 2022 14:40:41 +0000 (14:40 +0000)]
Constify struct cfattach.
ok patrick@
florian [Sat, 12 Mar 2022 14:35:29 +0000 (14:35 +0000)]
Probe for DNS64 presence with an absolute name so that asr doesn't add
search domains and retry.
jsing [Sat, 12 Mar 2022 12:53:03 +0000 (12:53 +0000)]
Factor out change cipher spec handing code in the legacy stack.
Factor out the code that handles the processing of a change cipher spec
message that has been read in the legacy stack, deduplicating code in the
DTLS stack.
ok inoguchi@ tb@
mpi [Sat, 12 Mar 2022 12:34:22 +0000 (12:34 +0000)]
Uncompress some one line comments to reduce the difference with NetBSD.
No functionnal change.
sthen [Sat, 12 Mar 2022 12:20:03 +0000 (12:20 +0000)]
document MODPERL_REGEN_PPPORT
kettenis [Sat, 12 Mar 2022 11:28:55 +0000 (11:28 +0000)]
Unfortunately some last-minute changes were made to the AIC2 device tree
bindings before it was accepted in Linux. But handling both the
preliminary and (hopefully) final bindings can be done in a fairly clean
way.
ok jsg@
mpi [Sat, 12 Mar 2022 08:11:07 +0000 (08:11 +0000)]
Revert holding a read lock on the map while copying out data during sysctl(2).
This introduced a lock ordering issue reported by naddy@, anton@ and syzkaller.
Reported-by: syzbot+739bb901045d9b193bde@syzkaller.appspotmail.com
jmatthew [Sat, 12 Mar 2022 06:54:05 +0000 (06:54 +0000)]
Constify struct cfattach, not struct cfdriver
krw [Fri, 11 Mar 2022 22:29:55 +0000 (22:29 +0000)]
Simple hack to preserve the EFI System partition when
auto-allocating the space on a disk whose GPT contains
any of the magic Apple partitions.
One step towards saving the firmware files on Apple
M1 machines when installing OpenBSD.
Tested & ok kettenis@
kettenis [Fri, 11 Mar 2022 19:24:19 +0000 (19:24 +0000)]
Hold a read lock on the map while copying out data during a sysctl(2) call
to prevent another thread from unmapping the memory and triggering
an assertion or even corrupting random physical memory pages.
This fix is similar to the change in uvm_glue.c rev. 1.74. However in this
case we need to be careful since some sysctl(2) calls look at the map of
the current process. In those cases we must not attempt to lock the map
again.
ok mpi@
Should fix:
Reported-by: syzbot+be89fe83d6c004fcb412@syzkaller.appspotmail.com
bluhm [Fri, 11 Mar 2022 19:02:15 +0000 (19:02 +0000)]
Declare membar_datadep_consumer() inline to fix build on sparc64.
mpi [Fri, 11 Mar 2022 18:00:45 +0000 (18:00 +0000)]
Constify struct cfattach.