CVS: cvs.openbsd.org: src
CVSROOT:/cvs Module name:src Changes by: js...@cvs.openbsd.org 2025/07/20 02:55:49 Modified files: lib/libcrypto/aes: aes.c Log message: Zero stack based IV and buffer when aes_ctr32_encrypt_generic() completes. ok tb@
CVS: cvs.openbsd.org: src
CVSROOT:/cvs Module name:src Changes by: t...@cvs.openbsd.org2025/07/20 01:49:47 Modified files: regress/usr.sbin/rpki-client: test-aspa.c test-gbr.c test-geofeed.c test-roa.c test-rsc.c test-spl.c test-tak.c Log message: rpki-client regress: adjust to new printing function signature
CVS: cvs.openbsd.org: src
CVSROOT:/cvs Module name:src Changes by: t...@cvs.openbsd.org2025/07/20 01:48:31 Modified files: usr.sbin/rpki-client: aspa.c extern.h filemode.c gbr.c geofeed.c parser.c print.c roa.c rsc.c spl.c tak.c Log message: rpki-client: remove copies of cert data from signed objects Stop copying AIA, AKI, SIA, SKI, notBefore and notAfter into the signed objects. Adjust the parser process to use the data from the cert instead of the copies. Annotate missing expiry handling for gbr and tak. Handling of manifests is a bit different and will be done in a separate step. This is a layer violation that was made early on and is needed only for file mode nowadays. For normal runs it led to pushing quite a bit of unnecessary data across the pipes. Now that file mode has direct access to the parsed EE cert corresponding to a signed object, this data is available without the need for copying, so all this complexity can go away without real cost for file mode. Only the printing functions need a slight adjustment. The parsing of the EE certificate also ensures the presence of all these fields, so no checks are lost. ok job
CVS: cvs.openbsd.org: src
CVSROOT:/cvs Module name:src Changes by: t...@cvs.openbsd.org2025/07/20 00:15:50 Modified files: usr.sbin/rpki-client: cert.c Log message: rpki-client: typo in comment: cretificate -> certificate
CVS: cvs.openbsd.org: src
CVSROOT:/cvs Module name:src Changes by: n...@cvs.openbsd.org2025/07/19 13:30:37 Modified files: usr.bin/tmux : cmd-parse.y Log message: yyerror should be void.
CVS: cvs.openbsd.org: src
CVSROOT:/cvs Module name:src Changes by: n...@cvs.openbsd.org2025/07/19 13:20:10 Modified files: usr.bin/tmux : tty-keys.c Log message: Only accept DA and DA2 when they end in 'c', stops them being confused with theme responses.. GitHub issue 4561.
CVS: cvs.openbsd.org: src
CVSROOT:/cvs Module name:src Changes by: m...@cvs.openbsd.org2025/07/19 10:40:40 Modified files: sys/net: if.c if.h if_var.h sys/netinet: in_proto.c ip_input.c ip_mroute.c Log message: Push netlock down to mrt_sysctl_vif(). Link the network interfaces to the temporary `if_tmplist' list protected with `if_tmplist_lock' rwlock(9). This moves copyuot() out of exclusive netlock and allows to relax it to shared netlock. ip_sysctl() became mp-safe, so mark it as PR_MPSYSCTL. ok bluhm
CVS: cvs.openbsd.org: src
CVSROOT:/cvs Module name:src Changes by: dera...@cvs.openbsd.org 2025/07/19 07:18:06 Modified files: lib/libc/sys : execve.2 Log message: improve description of wxneeded
CVS: cvs.openbsd.org: src
CVSROOT:/cvs Module name:src Changes by: bent...@cvs.openbsd.org 2025/07/19 06:23:30 Modified files: distrib/miniroot: install.sub Log message: Move font cache chown to upgrade conditional; update comment scope accordingly. ok deraadt@
CVS: cvs.openbsd.org: src
CVSROOT:/cvs Module name:src Changes by: schwa...@cvs.openbsd.org2025/07/19 03:58:44 src/regress/usr.bin/mandoc/roff/in Update of /cvs/src/regress/usr.bin/mandoc/roff/in In directory cvs.openbsd.org:/tmp/cvs-serv33860/in Log Message: Directory /cvs/src/regress/usr.bin/mandoc/roff/in added to the repository
CVS: cvs.openbsd.org: src
CVSROOT:/cvs Module name:src Changes by: schwa...@cvs.openbsd.org2025/07/19 04:02:37 Modified files: regress/usr.bin/mandoc/roff: Makefile Added files: regress/usr.bin/mandoc/roff/in: Makefile fraction.in fraction.out_ascii incr.in incr.out_ascii Log message: test the roff(7) .in request
CVS: cvs.openbsd.org: src
CVSROOT:/cvs Module name:src Changes by: t...@cvs.openbsd.org2025/07/19 03:27:41 Modified files: usr.sbin/rpki-client: parser.c Log message: proc_parser_tak: must initialize cert to NULL CID 582440
CVS: cvs.openbsd.org: src
CVSROOT:/cvs Module name:src Changes by: m...@cvs.openbsd.org2025/07/18 23:23:37 Modified files: sys/lib/libkern/arch/m88k: copy_subr.S Log message: Put table of constant function pointers into .rodata.
CVS: cvs.openbsd.org: src
CVSROOT:/cvs Module name:src Changes by: n...@cvs.openbsd.org2025/07/18 14:44:13 Modified files: usr.bin/tmux : grid.c screen-write.c Log message: Correctly remove overwritten wide characters at column 1.
CVS: cvs.openbsd.org: src
CVSROOT:/cvs Module name:src Changes by: k...@cvs.openbsd.org2025/07/18 12:11:51 Modified files: sbin/fdisk : disk.c gpt.c Log message: Make DEBUG compile again by removing duplicate declaration of 'i'. While here append /* DEBUG */ comments to various #endif.
CVS: cvs.openbsd.org: src
CVSROOT:/cvs Module name:src Changes by: m...@cvs.openbsd.org2025/07/18 11:34:29 Modified files: sys/dev/wscons : wsevent.c wseventvar.h wskbd.c wsmouse.c wsmux.c Log message: Wipe out WSEVENT_MPSAFE logic. All wsevent_filtops are mp-safe for a long time. We don't need it anymore. ok bluhm
CVS: cvs.openbsd.org: src
CVSROOT:/cvs Module name:src Changes by: schwa...@cvs.openbsd.org2025/07/18 11:19:45 Modified files: lib/libc/sys : intro.2 Log message: Add a BUGS section, including some points made by deraadt@ on tech@ expressed in my own words. OK deraadt@
CVS: cvs.openbsd.org: src
CVSROOT:/cvs Module name:src Changes by: schwa...@cvs.openbsd.org2025/07/18 11:10:05 Modified files: lib/libc/sys : intro.2 Log message: Add the missing STANDARDS section. OK deraadt@
CVS: cvs.openbsd.org: src
CVSROOT:/cvs Module name:src Changes by: schwa...@cvs.openbsd.org2025/07/18 10:50:30 Modified files: lib/libc/sys : intro.2 Log message: Document EREMOTE because this manual page explicitly claims to provide a "complete list" and because it already contains some other error numbers that are defined in , but not used in OpenBSD. Omission reported by Angelo dot Rossi dot homelab at gmail dot com on tech@. According to millert@, SunOS 4.x used it in the context of NFS. Mark both EPROCLIM and EREMOTE as "Not used in OpenBSD." OK deraadt@
CVS: cvs.openbsd.org: src
CVSROOT:/cvs Module name:src Changes by: schwa...@cvs.openbsd.org2025/07/18 10:32:39 Modified files: lib/libc/sys : intro.2 Log message: Document EWOULDBLOCK because POSIX requires it to be defined, because POSIX requires several functions to set it, and because several of our manual pages refer to it. POSIX explicitly allows EWOULDBLOCK to resolve to the same error number as EAGAIN, which it does in OpenBSD. OK deraadt@
CVS: cvs.openbsd.org: src
CVSROOT:/cvs Module name:src Changes by: t...@cvs.openbsd.org2025/07/18 10:21:00 Modified files: regress/usr.sbin/rpki-client: test-seqnum.c Log message: rpki-client: remove unused bn from test-seqnum.c
CVS: cvs.openbsd.org: src
CVSROOT:/cvs Module name:src Changes by: s...@cvs.openbsd.org2025/07/18 10:09:28 Modified files: sys/dev/pci: if_iwx.c Log message: load the correct iwx(4) firmware on QuZ devices which use RF JF1/JF2 problem reported by Prateek Kumar
CVS: cvs.openbsd.org: src
CVSROOT:/cvs Module name:src Changes by: schwa...@cvs.openbsd.org2025/07/18 09:46:58 Modified files: usr.bin/mandoc : term_ps.c Log message: Adjust viscol (the distance in basic units from the column offset) and minbl (the minimum whitespace in basic units before the next column) in ps_advance() and ps_endline() because that is what term.c now expects. Regression reported by Jan Stary on misc@. Also adjust ps_hspan() to the new definition of basic units in terminal output.
CVS: cvs.openbsd.org: src
CVSROOT:/cvs Module name:src Changes by: bl...@cvs.openbsd.org 2025/07/18 09:44:44 Modified files: sys/net: if.c Log message: Increase softnet kernel threads from 4 to 8. After unlocking the hot path of the IP/UDP/TCP network stack, increase the number of softnet threads. 8 CPU instead of 4 doing packet processing seems to be a good number. Note that the value is always capped by number of CPU in the system. OK mvs@
CVS: cvs.openbsd.org: src
CVSROOT:/cvs Module name:src Changes by: t...@cvs.openbsd.org2025/07/18 07:19:59 Modified files: usr.sbin/rpki-client: extern.h x509.c Log message: rpki-client: garbage collect unused x509_* API It has served us well. Improved versions of these are now in cert.c. ok job
CVS: cvs.openbsd.org: src
CVSROOT:/cvs Module name:src Changes by: t...@cvs.openbsd.org2025/07/18 06:22:07 Modified files: regress/usr.sbin/rpki-client: test-aspa.c test-gbr.c test-geofeed.c test-mft.c test-roa.c test-rsc.c test-spl.c test-tak.c Log message: rpki-client: adjust regress to X509 ** -> struct cert ** change
CVS: cvs.openbsd.org: src
CVSROOT:/cvs Module name:src Changes by: t...@cvs.openbsd.org2025/07/18 06:20:32 Modified files: usr.sbin/rpki-client: aspa.c cms.c extern.h filemode.c gbr.c geofeed.c mft.c parser.c roa.c rsc.c spl.c tak.c Log message: rpki-client: move from X509 ** to struct cert ** Make the CMS parsing function take a struct cert ** rather than an X509 **. Also pass the talid. The point is: cert_parse_ee_cert() now pulls all the information out of the X509 and ensures the presence of required fields in the extensions, so we don't need to do that again in the signed object handlers. Things become a bit simpler in CMS since we can directly use the cert->notafter and ditch the up-ref dance. In aspa_parse() and most other signed object parsers, we can do similar things: pass in struct cert, copy AIA/AKI/SIA/SKI from the cert (that's mostly for filemode and can be improved later on), copy validity, drop cert_parse_ee_cert() and assign the out_cert in the success path. gbr_parse() is nearly identical. geofeed_parse() handles a detached object, hence is slightly special, but really is mostly more of the same. For mft_parse() there's an extra twist with the CRLDP, the handling of which gets a bit simpler. Nothing special about roas, rscs have no SIA, and spl and tak are again more of the same. In parser.c the changes are straightforward with a shift from X509 to struct cert. filemode becomes a bit simpler since the signed object parsers now give us the EE cert, so we don't need to parse that by hand anymore (which, incidentally, moves a triple warning in -vf back to a single warning). Similarly to mft_parse, we can use the EE cert's CRLDP to parse the crl uri and we no longer need to deal with the X509 directly. The printers still do, but that's something to address later on. ok job
CVS: cvs.openbsd.org: src
CVSROOT:/cvs Module name:src Changes by: j...@cvs.openbsd.org2025/07/18 03:31:23 Modified files: share/man/man4 : qcdrm.4 qcdpc.4 Log message: order cross reference lines ok kettenis@
CVS: cvs.openbsd.org: src
CVSROOT:/cvs Module name:src Changes by: j...@cvs.openbsd.org2025/07/18 03:27:53 Modified files: share/man/man4 : qcdpc.4 Log message: change some qcdrm(4) parts to qcdpc(4) ok kettenis@
CVS: cvs.openbsd.org: src
CVSROOT:/cvs Module name:src Changes by: m...@cvs.openbsd.org2025/07/18 02:39:14 Modified files: sys/netinet: ip_ipip.c tcp_subr.c udp_usrreq.c sys/netinet6 : icmp6.c in6_proto.c in6_src.c ip6_input.c Log message: Unlock IPV6CTL_DEFHLIM case of ip6_sysctl(). Note, traceroute(8) is the only user of IPV6CTL_DEFHLIM, but it doesn't included in the ramdisk. ok bluhm
CVS: cvs.openbsd.org: src
CVSROOT:/cvs Module name:src Changes by: t...@cvs.openbsd.org2025/07/17 21:16:28 Modified files: sbin/isakmpd : hash.c hash.h Log message: isakmpd: avoid undefined behavior in hash.c Calling a function through an incompatible function pointer cast is UB. Add wrapper functions for the hashing functions that pass the correct member of the union of hash contexts to the libc hashing functions, so we don't need to cast at all. This matches bin/mdt/md5.c r1.99. ok mvs yasuoka
CVS: cvs.openbsd.org: src
CVSROOT:/cvs Module name:src Changes by: j...@cvs.openbsd.org2025/07/17 18:15:45 Modified files: sys/dev/pci/drm: drm_framebuffer.c drm_gem.c drm_internal.h sys/dev/pci/drm/include/drm: drm_framebuffer.h Log message: drm/framebuffer: Acquire internal references on GEM handles >From Thomas Zimmermann 065bd940ee0a0033ea9a7dbb1a9110baef9415a4 in linux-6.12.y/6.12.39 f6bfc9afc7510cb5e6fbe0a17c507917b0120280 in mainline linux
CVS: cvs.openbsd.org: src
CVSROOT:/cvs Module name:src Changes by: j...@cvs.openbsd.org2025/07/17 18:12:49 Modified files: sys/dev/pci/drm: drm_gem.c sys/dev/pci/drm/include/drm: drm_file.h Log message: drm/gem: Fix race in drm_gem_handle_create_tail() >From Simona Vetter 2d2f07a9948756acdcd8aed14b4596482f51ac0f in linux-6.12.y/6.12.39 bd46cece51a36ef088f22ef0416ac13b0a46d5b0 in mainline linux
CVS: cvs.openbsd.org: src
CVSROOT:/cvs Module name:src Changes by: j...@cvs.openbsd.org2025/07/17 18:09:22 Modified files: sys/dev/pci/drm/ttm: ttm_bo_util.c Log message: drm/ttm: fix error handling in ttm_buffer_object_transfer >From Christian Koenig 38df1a5053bc8b2a1ba3aae562187304819a5ad0 in linux-6.12.y/6.12.39 97e000acf2e20a86a50a0ec8c2739f0846f37509 in mainline linux
CVS: cvs.openbsd.org: src
CVSROOT:/cvs Module name:src Changes by: j...@cvs.openbsd.org2025/07/17 18:07:17 Modified files: sys/dev/pci/drm/include/drm: spsc_queue.h Log message: drm/sched: Increment job count before swapping tail spsc queue >From Matthew Brost e2d6547dc8b9b332f9bc00875197287a6a4db65a in linux-6.12.y/6.12.39 8af39ec5cf2be522c8eb43a3d8005ed59e4daaee in mainline linux
CVS: cvs.openbsd.org: src
CVSROOT:/cvs Module name:src Changes by: j...@cvs.openbsd.org2025/07/17 18:04:27 Modified files: sys/dev/pci/drm: drm_gem.c drm_internal.h Log message: drm/gem: Acquire references on GEM handles for framebuffers >From Thomas Zimmermann 08480e285c6a82ce689008d643e4a51db0aaef8b in linux-6.12.y/6.12.39 5307dce878d4126e1b375587318955bd019c3741 in mainline linux
CVS: cvs.openbsd.org: src
CVSROOT:/cvs Module name:src Changes by: j...@cvs.openbsd.org2025/07/17 18:00:45 Modified files: sys/dev/pci/drm/amd/amdkfd: kfd_svm.c Log message: drm/amdkfd: Don't call mmput from MMU notifier callback >From Philip Yang e90ee15ce28c61f6d83a0511c3e02e2662478350 in linux-6.12.y/6.12.39 cf234231fcbc7d391e2135b9518613218cc5347f in mainline linux
CVS: cvs.openbsd.org: src
CVSROOT:/cvs Module name:src Changes by: j...@cvs.openbsd.org2025/07/17 17:58:24 Modified files: sys/dev/pci/drm/amd/amdgpu: amdgpu_device.c amdgpu_virt.c amdgpu_virt.h Log message: drm/amdgpu: Replace Mutex with Spinlock for RLCG register access to avoid Priority Inversion in SRIOV >From Srinivasan Shanmugam 07ed75bfa7ede8bfcfa303fd6efc85db1c8684c7 in linux-6.12.y/6.12.39 dc0297f3198bd60108ccbd167ee5d9fa4af31ed0 in mainline linux
CVS: cvs.openbsd.org: src
CVSROOT:/cvs Module name:src Changes by: j...@cvs.openbsd.org2025/07/17 17:53:31 Modified files: sys/dev/pci/drm/amd/amdgpu: amdgpu_discovery.c Log message: drm/amdgpu/ip_discovery: add missing ip_discovery fw >From Flora Cui 04513cf1581bfcd1f87fa50f9f96b601c8b536b9 in linux-6.12.y/6.12.39 2f6dd741cdcdadb9e125cc66d4fcfbe5ab92d36a in mainline linux
CVS: cvs.openbsd.org: src
CVSROOT:/cvs Module name:src Changes by: j...@cvs.openbsd.org2025/07/17 17:51:46 Modified files: sys/dev/pci/drm/amd/amdgpu: amdgpu_discovery.c Log message: drm/amdgpu/discovery: use specific ip_discovery.bin for legacy asics >From Flora Cui 39d6a607d531b05840cd095eaf0d93a0026406dc in linux-6.12.y/6.12.39 25f602fbbcc8271f6e72211b54808ba21e677762 in mainline linux
CVS: cvs.openbsd.org: src
CVSROOT:/cvs Module name:src Changes by: dera...@cvs.openbsd.org 2025/07/17 16:35:07 Modified files: distrib/sets/lists/man: mi Log message: sync
CVS: cvs.openbsd.org: src
CVSROOT:/cvs Module name:src Changes by: m...@cvs.openbsd.org2025/07/17 15:27:36 Modified files: gnu/usr.bin/gcc/gcc/config/m88k: m88k.c Log message: In expand_block_move(), the logic tries to emit inline load and stores (which can benefit from register renaming, especially in leaf functions) or a call to memcpy(), depending upon the alignment and size of the data. Double the limit for inline load and stores, unless optimizing for size: even with the limit double, this will still be faster than a memcpy() prologue + loop + epilogue. The code size increase remains quite small (a few KB for the kernel and libc), and worth the small speed gain.
CVS: cvs.openbsd.org: src
CVSROOT:/cvs Module name:src Changes by: m...@cvs.openbsd.org2025/07/17 15:23:44 Modified files: gnu/usr.bin/gcc/gcc/config/m88k: m88k.c m88k-protos.h m88k.md Log message: Remove dead code. This includes: - workarounds for bugs in non-gas assemblers - commented-out support for "fast" block move routines, which were subtly bogus back in the gcc 2.95 times when too much optimization (-O2) was used, and rotted enough to no longer work in gcc3 (unless you like internal compiler errors when that code triggers).
CVS: cvs.openbsd.org: src
CVSROOT:/cvs Module name:src Changes by: m...@cvs.openbsd.org2025/07/17 15:16:15 Modified files: gnu/usr.bin/gcc/gcc/config/m88k: m88k.h Log message: Remove commented out macros which have no longer matter (and already didn't in the gcc 2.95 days). NFC
CVS: cvs.openbsd.org: src
CVSROOT:/cvs Module name:src Changes by: m...@cvs.openbsd.org2025/07/17 15:13:46 Modified files: gnu/usr.bin/gcc/gcc/config/m88k: m88k.c Log message: For the sake of proper superscalar operation, consecutive accesses to volatile variables need to be separated by weak barriers to enforce that load and stores to these variables happen in order. That wasn't required until gcc grew support for the 88110 processor (the 88100 wasn't superscalar enough for that cinematic). But then one (of the many...) errata of the 88110 processor caused the weak barrier to have to be replaced with a strong one, which also impacts the 88100, which doesn't need such a barrier in the first place! Since then, better knowledge of the 88110 errata tell us that it's ok to use the original weak (88100-friendly) barrier as long as it isn't the last instruction of a page. So force it to be aligned to an 8 byte boundary. At worst we waste a NOP, and even that way this lets the 88100 systems run faster. Well, less slowed down.
CVS: cvs.openbsd.org: src
CVSROOT:/cvs Module name:src Changes by: t...@cvs.openbsd.org2025/07/17 12:21:37 Modified files: usr.sbin/rpki-client: rpki-client.8 Log message: I just made myself a target of axe murderers
CVS: cvs.openbsd.org: src
CVSROOT:/cvs Module name:src Changes by: m...@cvs.openbsd.org2025/07/17 11:30:47 Modified files: sys/kern : uipc_mbuf.c Log message: While modifying `mq_maxlen' through sysctl_mq(), use "oldval != newval" style to avoid `mq_mtx' mutex(9) acquisition if `mq_maxlen' was not changed. `mq_maxlen' has type of unsigned int, but protected with `mq_mtx' mutex(9), so use READ_ONCE() lockless load to ensure that `mq_maxlen' actually loaded to `oldval'. ok bluhm
CVS: cvs.openbsd.org: src
CVSROOT:/cvs Module name:src Changes by: m...@cvs.openbsd.org2025/07/17 11:31:45 Modified files: sys/netinet: if_ether.c ip_input.c Log message: Unlock IPCTL_ARPDOWN case of ip_sysctl(). ok bluhm
CVS: cvs.openbsd.org: src
CVSROOT:/cvs Module name:src Changes by: kette...@cvs.openbsd.org2025/07/17 09:53:36 Modified files: share/man/man4 : Makefile Added files: share/man/man4 : qcdpc.4 qcdrm.4 Log message: qcdpc(4) and qcdrm(4)
CVS: cvs.openbsd.org: src
CVSROOT:/cvs Module name:src Changes by: kette...@cvs.openbsd.org2025/07/17 09:52:10 Modified files: sys/dev/fdt: files.fdt sys/arch/arm64/conf: GENERIC Added files: sys/dev/fdt: qcdpc.c qcdrm.c Log message: Add initial support for the MSM Mobile Display Subsystem. This initial support brings us support for the AUX channel of the DisplayPort controllers. This in turn fives us support for backlight control on eDP panels. ok tobhe@
CVS: cvs.openbsd.org: src
CVSROOT:/cvs Module name:src Changes by: lan...@cvs.openbsd.org 2025/07/17 08:44:42 Modified files: etc/etc.alpha : login.conf etc/etc.amd64 : login.conf etc/etc.arm64 : login.conf etc/etc.armv7 : login.conf etc/etc.hppa : login.conf etc/etc.i386 : login.conf etc/etc.landisk: login.conf etc/etc.loongson: login.conf etc/etc.luna88k: login.conf etc/etc.macppc : login.conf etc/etc.octeon : login.conf etc/etc.powerpc64: login.conf etc/etc.riscv64: login.conf etc/etc.sparc64: login.conf Log message: login.conf: bump pbuild's datasize-cur to 12G on arm64, amd64, powerpc64 & riscv64 time -l shows that 10G are now required to build gkrust crate in thunderbird 140, as reported by tb@, ajacoutot@ and naddy@. going to 12G gives us a bit of headspace.. while here, assorted alignments wrt MAXDSIZ, with hints from sthen@: - use G instead of M when appropriate - bump sparc64 from 6G to 8G - bump armv7 from 1843M to 2G - bump macppc from 1.5G to 2G - clamp to 256M on luna88k and 512M on landisk, also for default/daemon/staff classes. ok sthen@ phessler@
CVS: cvs.openbsd.org: src
CVSROOT:/cvs Module name:src Changes by: clau...@cvs.openbsd.org 2025/07/17 06:43:43 Modified files: sys/net: if_wg.c Log message: KNF switch statements. OK tb@
CVS: cvs.openbsd.org: src
CVSROOT:/cvs Module name:src Changes by: n...@cvs.openbsd.org2025/07/17 05:39:23 Modified files: usr.bin/tmux : colour.c Log message: Allow uppercase letters in gray/grey color names, GitHub issue 4560 from Pavel Roskin.
CVS: cvs.openbsd.org: src
CVSROOT:/cvs Module name:src Changes by: bent...@cvs.openbsd.org 2025/07/17 04:47:36 Modified files: distrib/miniroot: install.sub Log message: Ignore stderr if (harmlessly) trying to chown a nonexistent font cache. Requested by deraadt@
CVS: cvs.openbsd.org: src
CVSROOT:/cvs Module name:src Changes by: t...@cvs.openbsd.org2025/07/17 04:34:05 Modified files: distrib/sets/lists/comp: mi Log message: sync
CVS: cvs.openbsd.org: src
CVSROOT:/cvs Module name:src Changes by: schwa...@cvs.openbsd.org2025/07/17 04:31:50 Modified files: lib/libcrypto/man: Makefile Added files: lib/libcrypto/man: PEM_X509_INFO_read_bio.3 Removed files: lib/libcrypto/man: PEM_X509_INFO_read.3 Log message: Rename the file PEM_X509_INFO_read.3 to PEM_X509_INFO_read_bio.3 because PEM_X509_INFO_read(3) no longer exists. Requested by tb@.
CVS: cvs.openbsd.org: src
CVSROOT:/cvs Module name:src Changes by: s...@cvs.openbsd.org2025/07/17 03:25:11 Modified files: share/man/man4 : ice.4 Log message: document ice(4) E823-L support and list currently supported devices
CVS: cvs.openbsd.org: src
CVSROOT:/cvs Module name:src Changes by: s...@cvs.openbsd.org2025/07/17 03:19:21 Modified files: sys/dev/pci: if_ice.c Log message: make ice(4) match on E823-L devices tested by me on E823-L with a 10G RJ45 transceiver
CVS: cvs.openbsd.org: src
CVSROOT:/cvs Module name:src Changes by: s...@cvs.openbsd.org2025/07/17 03:17:15 Modified files: sys/dev/pci: if_ice.c Log message: set MAC type to "generic" for E823-L ice(4) devices
Re: CVS: cvs.openbsd.org: src
Congratulations on pushing this through and thank you, not just for digging out the remaining non-compliance but also banging your head on the ports that are affected. I bow my head to you for your persistence. Philip Guenther On Wed, Jul 16, 2025 at 8:33 AM YASUOKA Masahiko wrote: > > CVSROOT:/cvs > Module name:src > Changes by: yasu...@cvs.openbsd.org 2025/07/16 09:33:05 > > Modified files: > lib/libc : Symbols.list shlib_version > lib/libc/hidden: stdio.h wchar.h > lib/libc/stdio : Makefile.inc fclose.3 fclose.c findfp.c > lib/libcrypto : shlib_version > lib/libcurses : shlib_version > lib/libedit: shlib_version > lib/libexpat : shlib_version > lib/libfido2 : shlib_version > lib/libfuse: shlib_version > lib/libkvm : shlib_version > lib/libpcap: shlib_version > lib/librthread : shlib_version > lib/libskey: shlib_version > lib/libsndio : shlib_version > lib/libutil: shlib_version > include: stdio.h wchar.h > gnu/lib/libcxx : shlib_version > gnu/lib/libcxxabi: shlib_version > gnu/lib/libexecinfo: shlib_version > gnu/lib/libiberty: shlib_version > gnu/lib/libreadline: shlib_version > gnu/lib/libstdc++: shlib_version > gnu/lib/libstdc++-v3: shlib_version > gnu/usr.bin/perl: shlib_version > Added files: > lib/libc/hidden: _stdio.h > lib/libc/stdio : fdclose.c > > Log message: > Make the FILE object opaque. Move the __sFILE definition into > libc/hidden/_stdio.h. All programs that refer to the internal > structure of the FILE object can't be compiled from now on. > std{in,out,err} doesn't refer __sF[] now, but the hidden __sF along > with __srget and __swbuf symbols are kept temporarily to make our > transition easier. But those symbols will be deleted soon. The shared > library versions are bumped for libc and all the other libraries that > refer to std{in,out,err}. > > diff from guenther, tweak by me, tested by many > ok sthen tb >
CVS: cvs.openbsd.org: src
CVSROOT:/cvs Module name:src Changes by: t...@cvs.openbsd.org2025/07/16 23:30:41 Modified files: bin/md5: md5.c Log message: md5(1): avoid incompatible function pointer casts This is a bit of a stupid change with lots of boilerplate. The problem is undefined behavior due to the second sentence of C99 6.3.2.3, 8: A pointer to a function of one type may be converted to a pointer to a function of another type and back again; the result shall compare equal to the original pointer. If a converted pointer is used to call a function whose type is not compatible with the pointed-to type, the behavior is undefined. The notion of compatibility for the functions here boils down to compatible return type and compatible argument list, which in particular means the same number of arguments. See 6.7.5.3, 15 for the gory details. There's also a version understandable to a human being lacking the brain damage necessary for successfully deciphering the C99 standard's language: https://en.cppreference.com/w/c/language/type.html So calling 'void CKSUM_Final(CKSUM_CTX *);' through the final member of struct hash_function is clearly UB since the numbers of parameters don't match. If that was all, the fix would be trivial. But the same goes for void MD5Final(MD5_CTX *, char *); because MD5_CTX * and void * aren't compatible in this sense: by definition, two pointer types are compatible if they point to compatible types. Or, if you prefer: void * has looser alignment requirements than MD5_CTX *, so they aren't compatible. The below avoids this by adding wrapper functions that have the correct types that then propagate their parameters by using the appropriate union member. with/ok millert
CVS: cvs.openbsd.org: src
CVSROOT:/cvs Module name:src Changes by: dera...@cvs.openbsd.org 2025/07/16 22:58:00 Modified files: sys/kern : kern_pledge.c Log message: When pledge was designed, I noticed that the libc DNS code would poll for newer "/etc/resolv.conf" using stat(2), so I permitted that operation in the pledge_namei() callback for the "dns" promise. Recent review found that libc DNS should actually be doing open(2) then fstat(2), and obviously the callback permits open(2). So libc was changed a few weeks ago to not use stat(2). We have now cranked libc for other reasons, and since we expect people to keep their kernel and libc in the same era (certainly less than a month), so I'm deleting the stat(2) approval in pledge_namei(). Something remains that I am uncomfortable with. Some go libraries also do stat(2)+open(2), rather than open(2)+fstat(2), and many years ago tedu wanted to do pledge in go programs. I'm uncomfortable because pledge is a contract with libc, not with some arbitrary external library that isn't actually trying to jive with the kernel's increasingly tightened security model. I leave that code until a grumpy day.
CVS: cvs.openbsd.org: src
CVSROOT:/cvs Module name:src Changes by: d...@cvs.openbsd.org2025/07/16 21:00:45 Modified files: sys/netinet6 : icmp6.c Log message: implement route sourceaddr handling in icmp6 reflection. this brings it into line with ipv4 and ipv6 pcb source address selection and ipv4 icmp_reflect source address selection. ok bluhm@
CVS: cvs.openbsd.org: src
CVSROOT:/cvs Module name:src Changes by: dera...@cvs.openbsd.org 2025/07/16 20:21:44 Modified files: lib/librthread : Makefile rthread.c Log message: Oops! Two files missing from pthread gmon diff: This thread-aware profiling support diff comes from Yuichiro NAITO. At runtime, profiling data is stored per-thread. Upon termination, the per-thread profiling data pools are merged into a into a single record, which is then written out (using the new kernel-assisted system). I think the original gmon merging parts may have come from or been inspired by someone at NetBSD. This has been delayed because there were concerns about adding a TLS object which would require other library cranks, so this rides all the major cranks today. ok kettenis tb claudio sthen etc etc etc
CVS: cvs.openbsd.org: src
CVSROOT:/cvs Module name:src Changes by: kenj...@cvs.openbsd.org 2025/07/16 18:55:48 Modified files: lib/libcrypto/curve25519: curve25519.c Log message: avoid undefined behavior when shifting into sign bit Shifting a signed int64_t into the sign bit is undefined behavior in C. /dev/portable/crypto/curve25519/curve25519.c:3900:18: runtime error: left shift of negative value -222076011 SUMMARY: UndefinedBehaviorSanitizer: undefined-behavior /dev/portable To avoid this, import int64_lshift21() from BoringSSL, a helper function that casts the input to uint64_t before shifting and back to int64_t afterward. This ensures defined behavior when shifting left by 21 bits, avoiding undefined behavior in expressions like `carry << 21`. This change addresses potential runtime issues detected by sanitizers when shifting signed values with high bits set. ok tb beck
CVS: cvs.openbsd.org: src
CVSROOT:/cvs Module name:src Changes by: t...@cvs.openbsd.org2025/07/16 12:33:05 Modified files: distrib/sets/lists/comp: mi Log message: sync
CVS: cvs.openbsd.org: src
CVSROOT:/cvs Module name:src Changes by: t...@cvs.openbsd.org2025/07/16 12:12:54 Modified files: lib/libcrypto/bio: bio.h Log message: Remove BIO_s_log() prototype, pointed out by schwarze
CVS: cvs.openbsd.org: src
CVSROOT:/cvs Module name:src Changes by: t...@cvs.openbsd.org2025/07/16 12:07:12 Removed files: lib/libcrypto/bio: bss_log.c Log message: Remove bss_log.c: no longer linked to build
CVS: cvs.openbsd.org: src
CVSROOT:/cvs Module name:src Changes by: schwa...@cvs.openbsd.org2025/07/16 11:59:10 Modified files: lib/libcrypto/man: Makefile PEM_ASN1_read.3 PEM_X509_INFO_read.3 PEM_bytes_read_bio.3 PEM_read.3 PEM_read_bio_PrivateKey.3 X509_INFO_new.3 X509_new.3 Removed files: lib/libcrypto/man: X509_PKEY_new.3 Log message: Undocument PEM_X509_INFO_read(3), X509_PKEY_new(3) and X509_PKEY_free(3) because tb@ removed them from Symbols.list rev. 1.220 today.
CVS: cvs.openbsd.org: src
CVSROOT:/cvs Module name:src Changes by: yasu...@cvs.openbsd.org 2025/07/16 10:43:22 Modified files: lib/libc/gmon : gmon.c Log message: Add the copyright lines for the previous commit. ok deraadt
CVS: cvs.openbsd.org: src
CVSROOT:/cvs Module name:src Changes by: schwa...@cvs.openbsd.org2025/07/16 10:36:04 Modified files: regress/usr.bin/mandoc/roff/ti: Makefile Added files: regress/usr.bin/mandoc/roff/ti: fraction.in fraction.out_ascii Log message: test temporary indentation by non-integer multiples of EN
CVS: cvs.openbsd.org: src
CVSROOT:/cvs Module name:src Changes by: dera...@cvs.openbsd.org 2025/07/16 10:22:58 Modified files: lib/libc : Symbols.list lib/libc/include: thread_private.h lib/libc/thread: rthread.c lib/libc/gmon : gmon.c mcount.c share/mk : bsd.lib.mk sys/lib/libkern: mcount.c sys/sys: gmon.h Log message: This thread-aware profiling support diff comes from Yuichiro NAITO. At runtime, profiling data is stored per-thread. Upon termination, the per-thread profiling data pools are merged into a into a single record, which is then written out (using the new kernel-assisted system). I think the original gmon merging parts may have come from or been inspired by someone at NetBSD. This has been delayed because there were concerns about adding a TLS object which would require other library cranks, so this rides all the major cranks today. ok kettenis tb claudio sthen etc etc etc
CVS: cvs.openbsd.org: src
CVSROOT:/cvs Module name:src Changes by: t...@cvs.openbsd.org2025/07/16 10:02:36 Modified files: distrib/sets/lists/base: md.alpha md.amd64 md.arm64 md.armv7 md.hppa md.i386 md.landisk md.loongson md.luna88k md.octeon md.powerpc64 md.riscv64 md.sparc64 mi Log message: sync the opaque stdio earthquake
CVS: cvs.openbsd.org: src
CVSROOT:/cvs Module name:src Changes by: t...@cvs.openbsd.org2025/07/16 10:01:40 Modified files: lib/libssl : shlib_version lib/libtls : shlib_version Log message: bump major for libssl and libtls to match libcrypto
CVS: cvs.openbsd.org: src
CVSROOT:/cvs Module name:src Changes by: t...@cvs.openbsd.org2025/07/16 09:59:26 Modified files: lib/libcrypto : Makefile Symbols.list lib/libcrypto/hidden/openssl: bio.h pem.h x509.h lib/libcrypto/pem: pem.h pem_info.c lib/libcrypto/x509: x509.h Log message: Ride the libcrypto bump with some simple cleanup: Remove BIO_s_log(): already unhooked in portable, completely unused. Remove X509_PKEY_new/free from public API. Remove PEM_X509_INFO_read() PEM_X509_INFO_write_bio(): all unused garbage. The simplify X509_PKEY_new/free was ok kenjiro.
CVS: cvs.openbsd.org: src
CVSROOT:/cvs Module name:src Changes by: schwa...@cvs.openbsd.org2025/07/16 09:51:27 Modified files: regress/usr.bin/mandoc/roff/po: Makefile Added files: regress/usr.bin/mandoc/roff/po: fraction.in fraction.out_ascii Log message: test page offsets by non-integer multiples of EN
CVS: cvs.openbsd.org: src
CVSROOT:/cvs Module name:src Changes by: dera...@cvs.openbsd.org 2025/07/16 09:43:26 Modified files: lib/libc : Symbols.list sys/sys: gmon.h lib/libc/gmon : gmon.c Log message: The kernel-assisted gmon support requires initialization at program start, so calling monstartup() later on is impossible. It was converted to call abort(), and has found no obvious or common observable run-time users. Remove the symbol from libc, alongside the recent major crank. ok sthen tb claudio
CVS: cvs.openbsd.org: src
CVSROOT:/cvs Module name:src Changes by: yasu...@cvs.openbsd.org 2025/07/16 09:33:05 Modified files: lib/libc : Symbols.list shlib_version lib/libc/hidden: stdio.h wchar.h lib/libc/stdio : Makefile.inc fclose.3 fclose.c findfp.c lib/libcrypto : shlib_version lib/libcurses : shlib_version lib/libedit: shlib_version lib/libexpat : shlib_version lib/libfido2 : shlib_version lib/libfuse: shlib_version lib/libkvm : shlib_version lib/libpcap: shlib_version lib/librthread : shlib_version lib/libskey: shlib_version lib/libsndio : shlib_version lib/libutil: shlib_version include: stdio.h wchar.h gnu/lib/libcxx : shlib_version gnu/lib/libcxxabi: shlib_version gnu/lib/libexecinfo: shlib_version gnu/lib/libiberty: shlib_version gnu/lib/libreadline: shlib_version gnu/lib/libstdc++: shlib_version gnu/lib/libstdc++-v3: shlib_version gnu/usr.bin/perl: shlib_version Added files: lib/libc/hidden: _stdio.h lib/libc/stdio : fdclose.c Log message: Make the FILE object opaque. Move the __sFILE definition into libc/hidden/_stdio.h. All programs that refer to the internal structure of the FILE object can't be compiled from now on. std{in,out,err} doesn't refer __sF[] now, but the hidden __sF along with __srget and __swbuf symbols are kept temporarily to make our transition easier. But those symbols will be deleted soon. The shared library versions are bumped for libc and all the other libraries that refer to std{in,out,err}. diff from guenther, tweak by me, tested by many ok sthen tb
CVS: cvs.openbsd.org: src
CVSROOT:/cvs Module name:src Changes by: t...@cvs.openbsd.org2025/07/16 09:21:45 Modified files: regress/usr.sbin/rpki-client: Makefile regress/usr.sbin/rpki-client/openssl: Makefile Log message: rpki-client regress: use openssl 3.5 if available and fall back to 3.3. The packages will be updated from 3.3 to 3.5 automatically. No action needed for those running full regress.
CVS: cvs.openbsd.org: src
CVSROOT:/cvs Module name:src Changes by: schwa...@cvs.openbsd.org2025/07/16 08:42:21 Modified files: regress/usr.bin/mandoc/tbl/mod: misalign.out_ascii Log message: As a side effect of tbl_term.c rev. 1.67, positioning of the vertical line on the right side of the table became better in this unusual edge case.
CVS: cvs.openbsd.org: src
CVSROOT:/cvs Module name:src Changes by: schwa...@cvs.openbsd.org2025/07/16 08:23:55 Modified files: usr.bin/mandoc : man_term.c mdoc_term.c out.c out.h roff_term.c tbl.h tbl_html.c tbl_layout.c tbl_term.c term.c term.h term_ascii.c term_ps.c term_tab.c Log message: In 2009, kristaps@ decided that using EN as the basic unit for horizontal widths on terminal output devices was the simplest option. It made sense because one EN is the smallest possible horizontal cursor movement on a terminal. In 2014, kristaps@ implemented the 'u' unit (BU, "basic unit") for scaled width input as 1/24 EN, but all internal variables continued to store widths in units of EN. This causes compatibility problems with groff, which internally represents widths in units of BU rather than EN. Various high-level macros internally calculate sums of user-controlled widths with a sub-EN precision. In mandoc, the coarse internal representation causes rounding errors, in some cases resulting in wrong indentations. Fix this by changing many internal variables in the terminal formatter to store widths in units of BU instead of EN, improving compatibility. An example of a manual page where mandoc output becomes more compatible with groff output is clang(1) - a page generated by the atrocious rst2man(1) tool, which insists, among other insanity, on indenting by non-integer multiples of EN. This patch includes the following related changes: * The termp.advance and endline functions now update termp.viscol. * Rename termp.width to termp.getwidth to help grep(1). * Delete term_hen() which becomes the same as term_hspan(). * Delete the unused struct termp_tbl. * Delete the unused tbl_sulen functions. In the tbl_term formatter, pass an additional coloff argument to all the data and fill functions, to deal with mismatches between the desired and the actual cursor position: the desired position is now maintained with BU precision, but the actual position is of course still an integer multiple of EN. Also simplify tbl_literal() by leaving TERMP_CENTER and TERMP_RIGHT adjustment to term_flushln(), which already contains code for that very purpose. We are really lucky to have the regress/ test suite. Without it, tricky and pervasive low-level refactoring of this kind would be impossible because it would unavoidably cause lots of regressions.
CVS: cvs.openbsd.org: src
CVSROOT:/cvs Module name:src Changes by: j...@cvs.openbsd.org2025/07/16 07:48:38 Modified files: sys/net: rtable.c Log message: avoid uninitialised return value in rtable_walk() with empty tables found with smatch, ok claudio@
Re: CVS: cvs.openbsd.org: src
On Wed, Jul 16, 2025 at 05:50:21AM -0600, Stefan Sperling wrote: > CVSROOT: /cvs > Module name: src > Changes by: s...@cvs.openbsd.org2025/07/16 05:50:21 > > Modified files: > sys/dev/pci: pcidevs > > Log message: > add E823-L PCI device IDs > This was ok jan@ phessler@
CVS: cvs.openbsd.org: src
CVSROOT:/cvs Module name:src Changes by: s...@cvs.openbsd.org2025/07/16 05:50:45 Modified files: sys/dev/pci: pcidevs.h pcidevs_data.h Log message: regen
CVS: cvs.openbsd.org: src
CVSROOT:/cvs Module name:src Changes by: s...@cvs.openbsd.org2025/07/16 05:50:21 Modified files: sys/dev/pci: pcidevs Log message: add E823-L PCI device IDs
CVS: cvs.openbsd.org: src
CVSROOT:/cvs Module name:src Changes by: j...@cvs.openbsd.org2025/07/16 01:15:42 Modified files: sys/arch/amd64/amd64: amd64_mem.c sys/arch/amd64/include: specialreg.h sys/arch/armv7/exynos: ec_commands.h sys/arch/armv7/omap: omehci.c sys/arch/hppa/dev: ssio.c viper.h sys/arch/hppa/hppa: locore.S sys/arch/hppa/include: iomod.h pdc.h trap.h sys/arch/i386/i386: i686_mem.c sys/arch/i386/include: pmap.h sys/arch/loongson/dev: sisfb.c smfb.c sys/arch/luna88k/cbus: nec86hw.c sys/arch/luna88k/stand/boot: sc.c scsivar.h sys/arch/m88k/m88k: m88100_fp.c sys/arch/octeon/dev: cn30xxpko.c sys/arch/powerpc64/powerpc64: trap.c sys/arch/sh/include: mmu_sh4.h sys/arch/sparc64/dev: viommu.c sys/arch/sparc64/include: ctlreg.h instr.h sys/arch/sparc64/sparc64: autoconf.c db_disasm.c Log message: spelling
CVS: cvs.openbsd.org: src
CVSROOT:/cvs Module name:src Changes by: j...@cvs.openbsd.org2025/07/15 21:01:53 Modified files: sys/arch/armv7/include: conf.h Log message: remove outdated comment about BEAGLEBOARD config
CVS: cvs.openbsd.org: src
CVSROOT:/cvs Module name:src Changes by: m...@cvs.openbsd.org2025/07/15 16:12:49 Modified files: sys/netinet: if_ether.c ip_input.c Log message: Unlock IPCTL_ARPTIMEOUT case of ip_sysctl(). ok bluhm
CVS: cvs.openbsd.org: src
CVSROOT:/cvs Module name:src Changes by: b...@cvs.openbsd.org2025/07/15 14:27:59 Modified files: sys/dev/hid: hidmt.c Log message: hidmt: separate contact inputs correctly When processing input reports, hidmt_input identifies groups of contact- level items by checking for the repetition of an item type - the first type it has encountered in the report. This fails for reports starting with report-level items, which only occur once. The 'firstu' value must not be set to a report-level type. Thanks to Thomas Dettbarn for analyzing and reporting the problem. ok stsp@
CVS: cvs.openbsd.org: src
CVSROOT:/cvs Module name:src Changes by: m...@cvs.openbsd.org2025/07/15 12:28:57 Modified files: sys/netinet: ip_input.c ip_output.c Log message: Unlock IPCTL_MFORWARDING case of ip_sysctl(). ok bluhm
CVS: cvs.openbsd.org: src
CVSROOT:/cvs Module name:src Changes by: j...@cvs.openbsd.org2025/07/15 07:40:02 Modified files: sys/dev/ata: atavar.h satareg.h sys/dev/ic : psp.c sys/dev/isa: pasreg.h sys/dev/mii: ciphyreg.h sys/dev/onewire: owsbm.c sys/dev/pci: amdpm.c if_bge.c if_bnxtreg.h if_ice.c if_iwxreg.h if_tl.c jmb.c neo.c sys/dev/pckbc : pms.c sys/dev/vmm: vmm.h Log message: spelling
CVS: cvs.openbsd.org: src
CVSROOT:/cvs Module name:src Changes by: bl...@cvs.openbsd.org 2025/07/15 06:52:52 Modified files: sys/kern : uipc_socket.c Log message: Use shared netlock and socket lock for closing sockets. Functions soclose() and sofree() do not need exclusive netlock anymore. Allow to run them in parallel. OK mvs@
CVS: cvs.openbsd.org: src
CVSROOT:/cvs Module name:src Changes by: clau...@cvs.openbsd.org 2025/07/15 06:28:05 Modified files: sys/arch/hppa/hppa: lock_machdep.c sys/arch/powerpc/powerpc: lock_machdep.c Log message: Use the same MP_LOCKDEBUG pattern as it is used in all other places like the MI mutex implementation. OK jca@
CVS: cvs.openbsd.org: src
CVSROOT:/cvs Module name:src Changes by: d...@cvs.openbsd.org2025/07/15 04:14:46 Modified files: sys/net: rtable.c Log message: make rtable_mpath_insert put routes after the same prio rather than before. this is another semantic i accidentally screwed up when moving art from srp to smr. pointed out by anton@ ok claudio@
CVS: cvs.openbsd.org: src
CVSROOT:/cvs Module name:src Changes by: d...@cvs.openbsd.org2025/07/15 03:55:49 Modified files: sys/sys: sysctl.h sys/net: rtable.c rtable.h rtsock.c Log message: add rtable_read(), a "reader" variant of rtable_walk. rtable_read doesn't give up the rtable lock when calling the rtentry handler, and consequently doesnt need to mess around with rtentry refs. rtable_walk is often used to update or delete routes. when i moved art from srp to smr i also made rtable_walk properly account for the rtentry references when it gives up the rtable lock. this in turn meant that the routing table dump sysctl reported extra rtentry refs than it used to. dumping the routing table doesn't change the shape of the routing table, so we don't have to give up the rtable lock for it's callback. the rt ref damage was pointed out by bluhm (which i didnt absorb somehow) and later anton. discussed with, tested by, and ok anton@ claudio@ bluhm@
CVS: cvs.openbsd.org: src
CVSROOT:/cvs Module name:src Changes by: t...@cvs.openbsd.org2025/07/15 03:26:19 Modified files: regress/usr.sbin/rpki-client: test-cert.c Log message: rpki-client: adjust test-cert.c to cert_parse_pre() removal
CVS: cvs.openbsd.org: src
CVSROOT:/cvs Module name:src Changes by: m...@cvs.openbsd.org2025/07/15 02:27:47 Modified files: sys/netinet: ip_input.c Log message: Unlock IPCTL_IPPORT_MAXQUEUE case of ip_sysctl() We could set `ip_maxqueue' to be less than (ip_frags + 1) just after the netlock is released. So we don't need to serializes `ip_frags' and `ip_maxqueue' with the `ipq_mutex', the cached value is enough for ip_fragcheck() run. ok bluhm
CVS: cvs.openbsd.org: src
CVSROOT:/cvs Module name:src Changes by: t...@cvs.openbsd.org2025/07/15 01:23:39 Modified files: usr.sbin/rpki-client: cert.c extern.h filemode.c parser.c Log message: rpki-client: merge cert_parse() and cert_parse_pre() The split between cert_parse_pre() and cert_parse() is no longer meaningful, really. cert_parse_pre() has checks depending on the purpose which are now all part of cert_parse_extensions() except for the check that a cert coming through here is not supposed to be an EE cert, which I kept. cert_parse() itself only has some checks on extensions which are done as part of cert_parse_extensions() already, so it can go away. For TA certs we could not call cert_parse() previously because of AIA, CRLDP and SKI vs AKI, which we now handle depending on purpose in cert_parse_extensions(), so this limiattion is now gone. We don't call cert_parse() or cert_parse_internal() directly from ta_parse() since that complicates filemode and doesn't provide a big benefit. ok job
CVS: cvs.openbsd.org: src
CVSROOT:/cvs Module name:src Changes by: an...@cvs.openbsd.org 2025/07/14 23:09:26 Modified files: regress/sys/net/rtable: util.c Log message: Perform the same rt_plen2mask() refcnt_read() const corrections in regress.
CVS: cvs.openbsd.org: src
CVSROOT:/cvs Module name:src Changes by: j...@cvs.openbsd.org2025/07/14 19:09:32 Modified files: sys/arch/amd64/include: smbiosvar.h sys/arch/arm64/include: smbiosvar.h sys/arch/i386/include: smbiosvar.h sys/arch/powerpc64/include: smbiosvar.h Log message: Wnclosure -> Enclosure