summaryrefslogtreecommitdiff
AgeCommit message (Collapse)Author
2023-09-03Allow UDP for built-in inetd(8) services on 127.0.0.1.Alexander Bluhm
This restriction was added in year 2000 due to IPv6 compatible and mapped addresses. Nowadays our kernel does not support these IPv6 features and blocks localhost addresses on non-loopback interfaces. Make IPv4 127.0.0.1/8 and IPv6 ::1 behave identically and provide local services if configured. OK mvs@ deraadt@
2023-09-03sync perl syscall emulator headerAlexander Bluhm
2023-09-03Avoid a useless increment and decrement of the tcp syn cache refcountAlexander Bluhm
by unexpanding the SYN_CACHE_TIMER_ARM() macro in the timer callback. OK mvs@
2023-09-03Document rdomain-related behaviour of dhcpd(8).Stefan Sperling
ok jmc
2023-09-03if UNLOCK_CMD is a user settings, so is LOCK_CMD !Marc Espie
2023-09-03gc ancient stuffMarc Espie
point people to pkg_create wrt FULLPKGNAME
2023-09-03Shuffle the order in which Manifest entries are processedJob Snijders
Previously work items were enqueued in the order the CA intended them to appear on a Manifest. However, there is no obvious benefit to letting third parties decide the order in which things are processed. Instead, randomize: ordering has no meaning anyway, and the number of concurrent repository synchronization operations is limited & timeboxed. As they say, a fox is not taken twice in the same snare OK tb@
2023-09-03Let builtin process names be stored in maps as string.Martin Pieuchot
ok dv@
2023-09-03Use EVBUFFER_DATA instead of reaching into struct evbuffer. ok tbNicholas Marriott
2023-09-03Use EVBUFFER_DATA instead of reaching into struct evbuffer. ok tbNicholas Marriott
2023-09-03vmm(4): Suppress AMD HwPstate visibility to guestsMike Larkin
On newer Ryzen/EPYC, we need to hide the HwPstate CPUID 80000007:EDX field for HwPstate, or guests will try to access the MSRs associated with those, and that will fail with #GP. ok deraadt
2023-09-03Replace perl's use of syscall(2) with a dispatcher to libcAndrew Fresh
This removes the ability to do direct syscalls from perl, instead calling the appropriate libc functions. Currently we generate the dispatcher via a perl script duing build. requested by deraadt@ nits from espie@ education from miod@ infrastructure fixes from sthen@ many improvements and ok gkoehler@ Please commit soon. OK bluhm@
2023-09-03Zero initialize Perl AV elements after unshift and resize.Alexander Bluhm
https://github.com/Perl/perl5/issues/21235 https://github.com/Perl/perl5/pull/21265 https://github.com/Perl/perl5/commit/9e298ab597b2c6fc0741749d7e29119e2d6cd628 Bug reported by Anton Borowka. Fix from upstream Perl. OK afresh1@
2023-09-03Adapt tlb flush calls following arm64/pmap.cJeremie Courreges-Anglas
1. in pmap_enter() no need to call tlb_flush_page() if we don't actually insert a pted 2. all callers of pmap_pte_remove() already call tlb_flush_page() This seems to result in some performance improvement (18mn -> 17mn15) while building libc on a Hifive Unmatched). Also zap whitespace and useless comments to further reduce the diff with arm64/pmap.c ok kettenis@
2023-09-03Inline PTED_* functions and actually use PTED_WIRED()Jeremie Courreges-Anglas
As noted by drahn@ the compiler did inline said functions, but it also provided them as unused symbols. ok miod@ mlarkin@ kettenis@
2023-09-03pmap_page_protect() should not unmap pages after making them readonly.Jeremie Courreges-Anglas
This brings riscv64/pmap.c in line with arm64/pmap.c, original fix by drahn@ ok miod@ kettenis@ mlarkin@
2023-09-02Request terminal colours again on SIGWINCH but at most once every 30Nicholas Marriott
seconds, GitHub issue 3582.
2023-09-02btrace(8): allow empty statements in probes.Dave Voutila
Allows for probes like `BEGIN {}`, in parity with bpftrace. Also fixes an incorrect syntax error parsing argN builtins in subsequent probes after an empty BEGIN block. ok mpi@
2023-09-02Make sure cert_type is not 0 to prevent leak of certid->id_buf.Tobias Heider
Found by David Linder ok patrick@
2023-09-02Remove unneccessary id == NULL check after dereferencing it. id can neverTobias Heider
be NULL here. Found by tb@
2023-09-02document ALL_DISTFILES and ALL_PATCHFILESMarc Espie
2023-09-02btrace(8): fix probe builtin in BEGIN/END.Dave Voutila
Resolves segfaults when using probe within BEGIN or END. ok mpi@
2023-09-02Fix comment about skipped interfacesKlemens Nanni
After r1.44 "Start on DOWN interfaces" this sentence makes no sense and just repeats the obvious conditions, so zap it. With/OK stsp
2023-09-02Setulc only does RGB colour so add Setulc1 to do non-RGB colours, GitHubNicholas Marriott
issue 3627.
2023-09-02Use a hardware based number of KDF rounds by default for passphrasesKlemens Nanni
When creating new crypto volumes with a passphrase or updating one, pick a number of rounds that aims to take around 1s instead of just 16 (on X230 and T14 machines, 16 rounds unlock pretty much instantly). New default [-r auto] never decreases rounds, only explicit '-r N' can. 16 is the absolute minimum. Motivation is to provide a saner and more modern default, especially for fresh installations utilizing new disk encryption question. Prodding for new default from and OK jsing on early "-r auto" installer diff idea to to pick MAX(auto, old-rounds) from Lucas[AT sexy DOT is] "seems acceptable to me" deraadt Feedback kettenis sthen OK op
2023-09-02Avoid trying to remove keys if the station is not active in iwm firmware.Stefan Sperling
Same fix as applied to iwx(4) in March. Since iwm(4) does not use a task for key installation and removal the same race might not manifest here. However, trying to delete a key while the station is not in firmware is wrong in any case. And this code might get moved into a task later.
2023-09-02Make iwm(4) not crash when aircrack-ng attempts to inject frames viaStefan Sperling
bpf in monitor mode. Crash reported by brad and sthen. I reproduced the issue to verify my fix but I did not check whether aircrack-ng actually works with this.
2023-09-02Set visited flag on last windows when linking session.Nicholas Marriott
2023-09-02Zap anon pages mappings in uvm_anon_release() instead of in the fault handler.Martin Pieuchot
This makes all code paths deactivating or freeing anons consistent. No objection from the usual suspects.
2023-09-02whitespaceTheo Buehler
2023-09-02Align EVP_PKEY_get1_RSA() with EVP_PKEY_get0_RSA()Theo Buehler
2023-09-01Implement drive strength tweaking for the YT8531. Needed on the v1.3bMark Kettenis
variant of the visionfive2 board. ok jsing@
2023-09-01Call rtm_send() with netlock held to protect dereference of sockaddrVitaliy Makkoveev
structure data returned by rtable_getsource(). Netlock can't be pushed within rtm_send() because we have paths where caller already holds it. tested by jca ok bluhm jca
2023-09-01vmd(8): ignore masks on asserts, use synchronous deasserts.Dave Voutila
The i8259 was considering the state of the mask register when a device requested raising the bit in the interrupt request register. This caused a race condition where if the virtio device asserted the irq while it was masked in the i8259 by the vm, we'd miss the interrupt request. The device and the pic would become out of sync and users reported virtio block device stalls as the vioblk(4) driver would starve, waiting for an interrupt that will never arrive. The mask is now considered only at ack time, when finding possible interrupts to inject. This bug was never a problem previously as virtio devices were emulated synchronously. Deasserts related to the vcpu reading the virtio isr register are also made now in response to the read request instead of issued asynchronously. This removes a subsequent race condition. Testing from mbuhl@, stsp@, and Florian Riehm. ok mlarkin@
2023-09-01Only compare the actual size of the UTF-8 character, not all of it.Nicholas Marriott
2023-09-01Many improvements, almost amounting to a partial rewrite:Ingo Schwarze
* more precision what the CIPHER_CTX functions do * more precision what an NID is * avoid talking about RC2, use AES-256 for an example instead * clarify that block sizes are measured in bytes * mention additional restrictions regarding valid block sizes * add the missing description of the *_flags(3) functions * mention the public mask constant EVP_CIPH_MODE * add three missing modes that can occur as return values * add the missing entries for *_flags(3) and *_mode(3) below RETURN VALUES * tweak various wordings for precision and conciseness
2023-09-01Fix EVP_PKEY_get0_RSA() for RSA-PSSTheo Buehler
It currently returns NULL. This is OpenSSL 4088b926 + De Morgan. ok jsing
2023-09-01Add missing -T to getopt string.Nicholas Marriott
2023-09-01Update with pin definitions from Linux 6.4.Mark Kettenis
ok patrick@
2023-09-01Clear combine flag when a non-UTF-8 set of characters is encountered.Nicholas Marriott
2023-09-01Rewrite combined character handling to be more consistent and to supportNicholas Marriott
newer Unicode combined characters (which we have to "know" are combined since they are not width zero). GitHub issue 3600.
2023-09-01Expand name before looking for window with -S, GitHub issue 3670.Nicholas Marriott
2023-09-01Add detach-on-destroy previous and next, mostly from Alexis Hildebrandt.Nicholas Marriott
2023-09-01fix an obvious typo in the OBJ_NAME_add(3) prototypeIngo Schwarze
2023-08-31Add spaces for failure statusAndrew Fresh
2023-08-31sched_cpu_init: remove unnecessary NULL-checks for clockintr pointersScott Soule Cheloha
sched_cpu_init() is only run once per cpu_info struct, so we don't need these NULL-checks. The NULL-checks are a vestige of clockintr_cpu_init(), which runs more than once per CPU and uses the checks to avoid leaking clockintr handles. Thread: https://marc.info/?l=openbsd-tech&m=169349579804340&w=2 ok claudio@
2023-08-31Exit fw_update(8) lock_db subprocess when parent exitsAndrew Fresh
If fw_update exits unexpectedly the package database would never unlock. select solution from millert@
2023-08-31Improve feedback from fw_update(8)Andrew Fresh
Show status as we go with spinner rather than printing only at the end. Suggestions from deraadt@ Most of this has been in snapshots for a while
2023-08-31syncTheo Buehler
2023-08-31Split three new manual pages EVP_CIPHER_nid(3), EVP_CIPHER_CTX_ctrl(3),Ingo Schwarze
and EVP_CIPHER_CTX_set_flags(3) out of the excessively large and unwieldy EVP_EncryptInit(3). This causes a number of inaccuracies and gaps to stand out, but i'm not mixing text changes or content additions into this split. Using very useful feedback from tb@ regarding what belongs together and how important the various functions are. I refrained from bothering him with the complete patch, but he likes the general direction.
generated by cgit v1.2.3 (git 2.25.1) at 2025-01-10 08:33:21 +0000