jan [Fri, 22 Mar 2024 21:49:52 +0000 (21:49 +0000)]
Increase /usr/src partition to 2GB - 5GB
ok miod@, deraadt@, otto@ and bluhm@
bluhm [Fri, 22 Mar 2024 21:48:38 +0000 (21:48 +0000)]
Remove padding from union inpaddru.
Alignment of IPv4 address with lower part of IPv6 address looks
like a leftover from times when IPv6 compatible addresses should
contain IPv4 addreses. Better use a simple union for both IPv4 and
IPv6 addresses like everywhere else. Use this type also for common
zero address.
OK mvs@
jcs [Fri, 22 Mar 2024 19:22:23 +0000 (19:22 +0000)]
add "-V none" to prevent making any backups
from FreeBSD
ok bluhm deraadt
bluhm [Fri, 22 Mar 2024 19:14:28 +0000 (19:14 +0000)]
Fix chroot(2) call in control process.
Use /var/empty as chroot directory. Call chroot(2) before setresuid(2).
Do the error check correctly. Call chdir(2) after chroot(2).
from spiros thanasoulas; with florian@ tb@; OK millert@
bluhm [Fri, 22 Mar 2024 18:05:01 +0000 (18:05 +0000)]
Make local port which is bound during connect(2) unique per laddr.
in_pcbconnect() did not pass down the address it got from in_pcbselsrc()
to in_pcbpickport(). As a consequence local port numbers selected
during connect(2) were globally unique although they belong to
different addresses. This strict uniqueness is not necessary and
wastes usable ports for outgoing connections.
To solve this, pass ina from in_pcbconnect() to in_pcbbind_locked().
This does not interfere how wildcard sockets are matched with
specific sockets during bind(2). It only allows non-wildcard sockets
to share a local port during connect(2).
OK mvs@ deraadt@
mvs [Fri, 22 Mar 2024 17:34:11 +0000 (17:34 +0000)]
Use sorflush() instead of direct unp_scan(..., unp_discard) to discard
dead unix(4) sockets.
The difference in direct unp_scan() and sorflush() is the mbuf(9) chain.
For the first case it is still linked to the `so_rcv', for the second it
is not. This is required to make `sb_mtx' mutex(9) the only `so_rcv'
sockbuf protection and remove socket re-locking from the most of
uipc_*send() paths. The unlinked mbuf(9) chain doesn't require any
protection, so this allows to perform sleeping unp_discard() lockless.
Also, the mbuf(9) chain of the discarded socket still contains addresses
of file descriptors and it is much safer to unlink it before FRELE()
them. This is the reason to commit this diff standalone.
ok bluhm
claudio [Fri, 22 Mar 2024 15:41:34 +0000 (15:41 +0000)]
Rework the cease shutdown reason to work in both directions by looking
at the ibuf payload passed to log_notification().
Because of this move ibuf_get_string() and the log_notification() call
in parse_notification().
OK tb@
jsg [Fri, 22 Mar 2024 12:29:33 +0000 (12:29 +0000)]
correct history of /dev/mem and /dev/kmem
The history section (added in CSRG) claimed both first appeared in v6.
Looking at the manuals in the TUHS archive, /dev/mem was in v1 and
/dev/kmem was introduced in v5.
ok jan@
claudio [Fri, 22 Mar 2024 07:19:28 +0000 (07:19 +0000)]
Rework parse_notification() to use the ibuf API for everything.
While there fix the RFC5492 handling of ERR_OPEN_CAPA (the current code
has the logic inversed). ERR_OPEN_CAPA is there to signal that a needed
capability is missing in our OPEN message. Just add the handling of
ERR_OPEN_CAPA to log_notification().
Also rework the handling of the shutdown reason and move the printing
into log_notification().
OK tb@
claudio [Fri, 22 Mar 2024 07:15:04 +0000 (07:15 +0000)]
Limit NFS connections to originate from a reserved port.
For TCP connections do the check when adding the socket via nfssvc(2).
For UDP do the check early after soreceive().
On top of this limit the sockets added via nfssvc(2) to IPv4 TCP and UDP
sockets.
OK millert@ deraadt@
tb [Fri, 22 Mar 2024 06:24:54 +0000 (06:24 +0000)]
Simplify X509_STORE_free()
This had an inlined version of sk_pop_free(). We can just call it the
right way.
ok jsing
ratchov [Fri, 22 Mar 2024 06:20:48 +0000 (06:20 +0000)]
aucat: Allow any device sample encoding.
If the device doesn't support the aucat internal encoding,
then setup a conversion layer instead of failing. This allows
aucat to be used for audio equipment testing/debugging
without involving the full sndiod processing chain.
ratchov [Fri, 22 Mar 2024 06:06:06 +0000 (06:06 +0000)]
aucat: Add a bytes-per-sample argument to allcobuf()
No behavior change.
ratchov [Fri, 22 Mar 2024 05:54:25 +0000 (05:54 +0000)]
pledge: Allow the AUDIO_GETDEV ioctl in "audio"
ok deraadt, kn, phessler
job [Fri, 22 Mar 2024 03:38:12 +0000 (03:38 +0000)]
Replace protocol literal strings and strlen() calls with defined constants
OK tb@ claudio@
patrick [Thu, 21 Mar 2024 23:12:33 +0000 (23:12 +0000)]
Never bindings use the phy-handle property instead of the phy property.
ok kettenis@
tobhe [Thu, 21 Mar 2024 22:08:49 +0000 (22:08 +0000)]
Avoid redundant allocation in ikev2_prfplus()
from markus@
bluhm [Thu, 21 Mar 2024 16:46:04 +0000 (16:46 +0000)]
Fix white space. OK tobhe@
nicm [Thu, 21 Mar 2024 12:10:57 +0000 (12:10 +0000)]
Write padding character into the right position.
nicm [Thu, 21 Mar 2024 11:53:11 +0000 (11:53 +0000)]
Reduce escape-time default to 10 milliseconds, 500 is far too long for
modern terminals and networks. Case made by Kurtis Rader in GitHub issue
3844.
nicm [Thu, 21 Mar 2024 11:51:32 +0000 (11:51 +0000)]
Add -M to always turn mouse on in a menu, GitHub issue 3779.
nicm [Thu, 21 Mar 2024 11:47:55 +0000 (11:47 +0000)]
Look for feature code 21 for DECSLRM and 28 for DECFRA in the device
attributes and also accept level 1 (there is no hardware with this but
some emulators may use it). Pointed out by James Holderness.
nicm [Thu, 21 Mar 2024 11:32:49 +0000 (11:32 +0000)]
Use -p for default paste-buffer command in buffer mode, it will only do
anything if the application asked for it. From Gregory Anders.
nicm [Thu, 21 Mar 2024 11:30:42 +0000 (11:30 +0000)]
Do not notify window-layout-changed if the window is about to be
destroyed (since it may have been freed by the time the notify happens),
from Romain Francoise in GitHub issue 3860.
nicm [Thu, 21 Mar 2024 11:27:18 +0000 (11:27 +0000)]
Revert detach-client part of last, did not intend this to go in.
nicm [Thu, 21 Mar 2024 11:26:28 +0000 (11:26 +0000)]
Do not consider a selection present if it is empty, from Michael Grant
(GitHub issue 3869). Also a typo fix from GitHub issue 3877.
claudio [Thu, 21 Mar 2024 10:58:15 +0000 (10:58 +0000)]
Do not exit the master process if a client TCP connection fails to be
added.
OK millert@
jsing [Thu, 21 Mar 2024 08:17:23 +0000 (08:17 +0000)]
Remove now unused certificates (which are also soon to expire).
jsg [Thu, 21 Mar 2024 05:14:35 +0000 (05:14 +0000)]
add uao_read_to_iosys_map() alternative to shmem_read_to_iosys_map()
jsg [Thu, 21 Mar 2024 05:06:42 +0000 (05:06 +0000)]
unstub intel_uc_fw_copy_rsa()
jmatthew [Thu, 21 Mar 2024 04:10:37 +0000 (04:10 +0000)]
mention that SR-IOV virtual functions work too
afresh1 [Thu, 21 Mar 2024 01:02:29 +0000 (01:02 +0000)]
Clear spinner after each cycle
Otherwise we can get left with a piece of the spinner if all firmware
gets updates and don't print a "kept" value.
While here, replace \010 with the ksh supported \b, as suggested by cheloha@
Noticed by deraadt@
bluhm [Wed, 20 Mar 2024 22:52:44 +0000 (22:52 +0000)]
Fix build with bitmap_copy() on i386 by adding const qualifier.
OK jsg@
tb [Wed, 20 Mar 2024 22:11:07 +0000 (22:11 +0000)]
Rename the remaining pmod to mod
tb [Wed, 20 Mar 2024 22:08:22 +0000 (22:08 +0000)]
pmod -> mod in CONF_IMODULE
Hungarian notation occasionally helps, but only if it is done consistently.
Steve and consistency... cf e.g the last few commits. Accordingly,
Hensonian Hungarian notation is a complete disaster. Start cleaning this
mess up.
tb [Wed, 20 Mar 2024 21:53:57 +0000 (21:53 +0000)]
tmod -> mod
requested by jsing
tb [Wed, 20 Mar 2024 21:51:23 +0000 (21:51 +0000)]
md -> imod for CONF_IMODULEs
A CONF_IMODULE is neither an EVP_MD nor a CONF_MODULE, so call it imod
instead of md or mod.
tb [Wed, 20 Mar 2024 21:49:00 +0000 (21:49 +0000)]
md -> mod for CONF_MODULEs
A CONF_MODULE is no EVP_MD, so call it mod instead of md.
tb [Wed, 20 Mar 2024 21:41:09 +0000 (21:41 +0000)]
Change return type of module_add()
There is one caller of this function which returns module_add() != NULL.
Make the function return an int instead.
suggested by and ok jsing
tb [Wed, 20 Mar 2024 21:31:31 +0000 (21:31 +0000)]
Implement imodule_free() and call it from module_finish()
ok jsing
tb [Wed, 20 Mar 2024 21:21:03 +0000 (21:21 +0000)]
Make module_free() NULL safe
ok jsing
op [Wed, 20 Mar 2024 17:52:43 +0000 (17:52 +0000)]
fix ORCPT handling
due to a swapped strlcpy() arguments we don't save the ORCPT argument
after validation. There's no buffer overflow since dsn_orcpt is zeroed.
Spotted by Tassilo Philipp, thanks!
ok millert gilles
jsing [Wed, 20 Mar 2024 10:38:05 +0000 (10:38 +0000)]
Use the new certificates/chains in regress.
The new certificates are more representative of the real world. The old
certificates use weak algorithms and expire in the very near future. Most
of our regress has already been switched over, this changes the remainder.
Thanks to Bernhard M. Wiedemann for reminding us of the upcoming expiry.
ok tb@
claudio [Wed, 20 Mar 2024 09:35:46 +0000 (09:35 +0000)]
Cleanup AID handling.
- Loops over all valid AID should start with AID_MIN and go up to AID_MAX - 1
e.g. for (i = AID_MIN; i < AID_MAX; i++)
If for some reason AID_UNSPEC must be handled make that explicit in the
for loop.
- aid2afi() now returns an error for AID_UNSPEC since there is no valid
AFI SAFI combo for AID_UNSPEC.
- Add additional checks for AID_MIN where currently only AID_MAX was checked.
This affects imsg for route refresh and graceful restart.
- Simplify add-path capability handling. Only the negotiated add_path capa
sets the flag for AID_UNSPEC to help code to quickly check if any add-path
is active.
OK tb@
claudio [Wed, 20 Mar 2024 09:26:42 +0000 (09:26 +0000)]
Remove some else statements in if () else if () chains that don't need
to be written this way. Makes the code more readable.
OK tb@
ratchov [Wed, 20 Mar 2024 08:42:11 +0000 (08:42 +0000)]
aucat: Add generic channel mapping in place of -j and -c options.
The argument to the -m option specifies the source and destination
channel ranges to be mapped. Compatibility is maintained: if -m
is not used, the -c and -j options still work.
Help and suggestions from Jan Stary <hans@stare.cz>, thanks.
jsg [Wed, 20 Mar 2024 06:11:41 +0000 (06:11 +0000)]
remove i915_gem_object_map() removed in linux 5.10
jsg [Wed, 20 Mar 2024 06:02:19 +0000 (06:02 +0000)]
switch ggtt from agp_map to io_mapping
reduces the diff to linux, no functional change
jsg [Wed, 20 Mar 2024 05:44:19 +0000 (05:44 +0000)]
add io_mapping_map_atomic_wc()
job [Wed, 20 Mar 2024 04:39:10 +0000 (04:39 +0000)]
Run most of regress explicitly in filemode to avoid hitting location checks
with tb@
job [Wed, 20 Mar 2024 04:36:30 +0000 (04:36 +0000)]
Check whether filename and SIA match
Verify whether the filename as presented by the publication point (which
is unsigned information) matches the filename in the SIA attribute
(which is signed information). Based on RFC 6487 section 4.8.8.
with and OK tb@
jsg [Wed, 20 Mar 2024 04:28:39 +0000 (04:28 +0000)]
unstub i915_ttm_io_mem_pfn()
jsg [Wed, 20 Mar 2024 03:02:24 +0000 (03:02 +0000)]
handle ttm_sg_tt_init() not being called before ttm_tt_fini()
Check that dmamap and segs are non-NULL as not all drivers call
ttm_sg_tt_init().
jsg [Wed, 20 Mar 2024 03:00:21 +0000 (03:00 +0000)]
unstub i915_ttm_memcpy_init() i915_ttm_move_memcpy()
jsg [Wed, 20 Mar 2024 02:58:40 +0000 (02:58 +0000)]
init bus_space tags in ttm struct
jsg [Wed, 20 Mar 2024 02:56:49 +0000 (02:56 +0000)]
enable io_mapping paths for lmem
jsg [Wed, 20 Mar 2024 02:54:40 +0000 (02:54 +0000)]
unstub xehp_load_dss_mask()
jsg [Wed, 20 Mar 2024 02:51:06 +0000 (02:51 +0000)]
add bitmap_intersects() bitmap_from_arr32()
jsg [Wed, 20 Mar 2024 02:48:20 +0000 (02:48 +0000)]
unstub i915_gem_object_map_pfn()
jsg [Wed, 20 Mar 2024 02:47:24 +0000 (02:47 +0000)]
add replacements for io_mapping_init_wc() paths
jsg [Wed, 20 Mar 2024 02:44:57 +0000 (02:44 +0000)]
start implementing io_mapping
jsg [Wed, 20 Mar 2024 02:42:17 +0000 (02:42 +0000)]
add vmap_pfn()
jsg [Wed, 20 Mar 2024 02:38:35 +0000 (02:38 +0000)]
in vga_get_uninterruptible() return early if behind a bridge
remove KASSERT() that triggered on dg2
tb [Tue, 19 Mar 2024 19:27:33 +0000 (19:27 +0000)]
Move the OPENSSL_init() stub from o_init.c to crypto_init.c
tb [Tue, 19 Mar 2024 17:48:57 +0000 (17:48 +0000)]
Remove OPENSSL/CRYPTO_realloc documentation
tb [Tue, 19 Mar 2024 17:34:05 +0000 (17:34 +0000)]
Remove X509_ALGOR_set_md() documentation
As far as LibreSSL is concerned, this terrible API is pushing up the
daisies.
tb [Tue, 19 Mar 2024 05:04:13 +0000 (05:04 +0000)]
Rename parent to issuer in struct auth
Parent is confusing and issuer is the appropriate terminology. This is
a mechanical diff. The only remaining uses of 'parent' in this code
base now mean 'parent process'.
discussed with beck and job
ok job
dlg [Tue, 19 Mar 2024 03:49:11 +0000 (03:49 +0000)]
count if_enqueue/ifq_enqueue errors as oqdrops.
this helps narrow down where some "output failures" on sec interfaces
occur.
based on discussion with jason tubnor
kettenis [Mon, 18 Mar 2024 21:57:22 +0000 (21:57 +0000)]
Implement Spectre-V4 mitigations. The only real effect of this change is
that we now make a firmware call to enable the mitigations if the
firmware tells us mitigations are implemented and needed. But according
to the specification these mitigations should be enabled by default.
The open source TF-A implementation only implements mitigations for older
Cortex-A76 cores. Newer Cortex-A76 revisions are not vulnerable and as
far as I can tell we only support SoCs with the newer cores.
ok patrick@
patrick [Mon, 18 Mar 2024 21:37:44 +0000 (21:37 +0000)]
Pass PHY OF node to the MII layer for use by PHY drivers.
kettenis [Mon, 18 Mar 2024 21:20:46 +0000 (21:20 +0000)]
Reduce dmesg spam and only print about resource conflicts for resources
that are actually enabled.
ok dlg@, deraadt@
kettenis [Mon, 18 Mar 2024 18:35:21 +0000 (18:35 +0000)]
Add support for the new layout of the CCSIDR_EL1 register that was
introduced in Armv8.3 when the CCIDX feature is advertised. This
makes us properly detect the cache size on newer CPU cores like
Neoverse N2, at least when emulated by QEMU.
ok jsg@
claudio [Mon, 18 Mar 2024 14:54:52 +0000 (14:54 +0000)]
Typecast char argument to isxdigit() to unsigned char since isxdigit()
only works that way correctly.
OK deraadt@
claudio [Mon, 18 Mar 2024 10:49:24 +0000 (10:49 +0000)]
Double PEER_DESCR_LEN to 64 characters since 32 is a bit on the short side.
OK sthen@, deraadt@, dlg@, tb@
claudio [Mon, 18 Mar 2024 10:16:50 +0000 (10:16 +0000)]
Use same markup for origin-set arguments as for roa-set. The maxlen argument
is optional.
OK tb@
op [Mon, 18 Mar 2024 08:50:54 +0000 (08:50 +0000)]
fix markup of _PATH_DEFPATH (Li -> Dv); ok jmc
op [Mon, 18 Mar 2024 08:48:50 +0000 (08:48 +0000)]
improve the MDA documentation
- add a pointer to the section when documenting the `mda' keyword
- rename the section to MDA COMMANDS
- document also what happens when the MDA doesn't exit with status 0
- add the missing environment variables
- sort the variables
- minor other tweaks to the text
with several improvements from jmc, ok jmc
dlg [Mon, 18 Mar 2024 06:14:50 +0000 (06:14 +0000)]
expose per port information via kstats.
the most interesting information exposed here is the number of times
a port changes state according to the lacp state machine. if a port
is flapping, it's hard to see if you only look at the current state.
getting a count of changes over time makes problems a lot more
visible and therefore fixable.
this also exposes counters around how the lacp protocol packets.
all of these can be useful when trying to line up behaviors with
another system (eg, a switch).
ok jmatthew@
dlg [Mon, 18 Mar 2024 06:05:23 +0000 (06:05 +0000)]
use high bits from the mbuf flowid to pick a port to transmit on.
a port here is a physical interface used by an aggr.
this leaves the low bits for a physical interface to use to pick a
tx ring. without this, aggr used low bits for port selection, which
takes bits away from the ring selection, which can lead to uneven
distribution of packets over tx rings.
ive been running this in production for well over a year now.
mvs [Sun, 17 Mar 2024 19:47:08 +0000 (19:47 +0000)]
Do UNP_CONNECTING and UNP_BINDING flags check in uipc_listen() and
return EINVAL if set. This prevents concurrent solisten() thread to make
this socket listening while socket is unlocked.
Reported-by: syzbot+4acfcd73d15382a3e7cf@syzkaller.appspotmail.com
ok mpi
kettenis [Sun, 17 Mar 2024 13:05:40 +0000 (13:05 +0000)]
The feature is called SSBS instead of SBSS.
tb [Sun, 17 Mar 2024 07:10:00 +0000 (07:10 +0000)]
Annotate RSA-PSS SHA parameter encoding as wrong
A historic blunderfest in the ASN.1 module for RSA-PSS led to very
confusing text in various RFCs. davidben and my current reading of
this is that parameters for SHA-* should be encoded as an ASN.1 NULL
rather than omitted. The use of X509_ALGOR_set_evp_md() leads to them
being omitted, and is therefore counter to the specification (but
allowed. We should fix this. For now, leave a reminder.
See https://boringssl-review.googlesource.com/c/boringssl/+/67088
for a lot more details.
ok davidben
guenther [Sun, 17 Mar 2024 05:49:41 +0000 (05:49 +0000)]
Use VERW to mitigate the RFDS (Register File Data Sampling) vulnerability
present in Intel Atom CPUs, reordering some ASM in return-to-userspace and
start/resume-vmx-guest to reduce the number of kernel values still live in
registers when VERW is used. This mitigation requires updated firmware which
has affected CPUs report RFDS_CLEAR in dmesg.
Firmware packaging by jsg@ and sthen@
Logic for interpreting intel's flags by jsg@ after lots of discussion
between him, deraadt@, and I
ok deraadt@
tb [Sun, 17 Mar 2024 01:44:59 +0000 (01:44 +0000)]
Remove unused enum rsc_resourceblock_tag
This was used in rsc.c prior to the switch to ASN.1 templates.
ok job
patrick [Sun, 17 Mar 2024 00:06:43 +0000 (00:06 +0000)]
Some PHYs need board-specific initializations, e.g. to correctly configure
LED settings, which might be stored in the marvell,reg-init property. With
these applied, the LEDs on the SolidRun ClearFog CN9130 Base work correctly.
ok kettenis@
tb [Sat, 16 Mar 2024 21:42:20 +0000 (21:42 +0000)]
Remove ugly parens and thereby fix KNF
kettenis [Sat, 16 Mar 2024 20:46:28 +0000 (20:46 +0000)]
Set the HCR_API and HCR_APK bits in the HCR_EL2 when CPUs boot in EL2.
Otherwise using PAC instructions in EL1 will trigger a trap into EL2
that we don't handle.
ok jsg@, deraadt@
tb [Sat, 16 Mar 2024 20:42:33 +0000 (20:42 +0000)]
Fix signed integer overflow in bnrand()
If more bits than INT_MAX - 7 are requested, the calculation of number
of bytes required to store the bignum triggers undefined behavior due to
signed integer overflow. This will typically result in bytes becoming
negative which will then make malloc() fail. If the ulimit should be
high enough to make malloc() succeed, there is a bad out of bounds write
in case bottom is set (an odd number was requested).
On jsing's request this does not deal with another bug which we could
catch with a similar check due to BN_bn2bin() failing later on as the
number of words in a BIGNUM is some fraction of INT_MAX.
ok jsing
cheloha [Sat, 16 Mar 2024 17:42:37 +0000 (17:42 +0000)]
btrace(8): cache ELF symbols in sorted array
Currently, every kelf_snprintsym() call performs a linear search
through the .symtab for a symbol matching the given PC. The search is
expensive and seems to be a major source of dropped profiling events.
Storing all STT_FUNC .symtab entries and their names in a sorted array
cuts search time from O(n) to O(lg n). In practice, the faster
lookups seem to dramatically reduce the profiling drop rate.
With tweaks from mpi@.
Thread: https://marc.info/?l=openbsd-tech&m=
170830125132105&w=2
ok mpi@
jsg [Sat, 16 Mar 2024 09:15:04 +0000 (09:15 +0000)]
recognise Cortex-A520AE (Hayes AE), Cortex-A720AE (Hunter AE)
jmc [Sat, 16 Mar 2024 06:29:36 +0000 (06:29 +0000)]
add -S to usage();
job [Sat, 16 Mar 2024 05:18:01 +0000 (05:18 +0000)]
Move RPKI Trust Anchor constraints from etc set to base
The cadence of updates being applied to the RPKI Trust Anchor constraints
seems sufficiently low, while the required understanding of context to make
educated decisions quite high, so centralized coordination of updates through
tech@openbsd.org is more appropriate.
requested by & OK deraadt@, OK tb@
millert [Sat, 16 Mar 2024 02:00:31 +0000 (02:00 +0000)]
whois: trim output after ">>> Last update of WHOIS database:"
Currently, whois(1) displays the full output it receives from the
server. With this change, any text after a line starting with ">>>
Last update of WHOIS database:" is dropped. This trims a lot of
useless text that would otherwise cause the data you care about to
scroll off the screen. From FreeBSD. OK deraadt@
op [Fri, 15 Mar 2024 21:56:22 +0000 (21:56 +0000)]
add some initial documentation regarding MDAs
this adds some initial commentary for how MDAs should behave and
in what environment they are executed.
diff from Philipp (philipp+openbsd [at] bureaucracy [dot] de) with
some tweaks from Richard Toohey and me.
ok gilles@
op [Fri, 15 Mar 2024 21:52:20 +0000 (21:52 +0000)]
set ORIGINAL_RECIPIENT in the environment of mda scripts
mostly for compatibility with postfix since some mdas (like
public-inbox) make use of it.
diff from Philipp (philipp+openbsd [at] bureaucracy [dot] de)
ok gilles@
tb [Fri, 15 Mar 2024 21:32:54 +0000 (21:32 +0000)]
zlib: sync with src
tb [Fri, 15 Mar 2024 21:32:21 +0000 (21:32 +0000)]
zlib: sync with upstream
More Windows #ifdef shuffling. Only one change relevant for OpenBSD:
Make deflateBound() more conservative and handle Z_STREAM_END.
stsp [Fri, 15 Mar 2024 17:45:36 +0000 (17:45 +0000)]
Ignore ADDBA requests if we are not ready to receive data frames.
This prevents potential firmware errors in Intel wifi drivers when APs
send an ADDBA request before the driver's state machine has settled
into RUN state.
The driver's addba task would race the driver's newstate task, and the
hardware would see an incorrect sequence of commands.
Ignoring an early ADDBA request is harmless. The AP will retry later.
Reported by zxystd from the OpenIntelWireless project, thanks!
ok phessler@
kn [Fri, 15 Mar 2024 17:31:21 +0000 (17:31 +0000)]
Backout "Move code into new stop_watchdog()"
An upgrade stalled on me, either my testing was flawed or my diff is...
Having stop_watchdog() is fine, but calling it in a different place has
is apparently too subtle for me to get right.
kn [Fri, 15 Mar 2024 16:29:32 +0000 (16:29 +0000)]
Move code into new stop_watchdog()
We have {reset,start}_watchdog() which are only used in unattended upgrade
code, but stopping the background timer is done inline for all upgrades,
incl. interactive ones.
Relocate it out of the very end of do_upgrade() right after its only caller
and limit it to unattended upgrades to match where/how the timer is started.
OK afresh1