Build date: 1751493601 - Wed Jul  2 22:00:01 UTC 2025
Build cvs date: 1751491726 - Wed Jul  2 21:28:46 UTC 2025
Build id: 2025-07-03.1
Build tags: amd64-regress sysupgrade

Applied the following diff(s):
/home/anton/tmp/robsd/src-sys-em.diff
/home/anton/tmp/robsd/src-sys-uhidev-sispm.diff
/home/anton/tmp/robsd/src-sysupgrade.diff

P bin/ps/extern.h
P bin/ps/keyword.c
P bin/ps/nlist.c
P bin/ps/print.c
P bin/ps/ps.c
P lib/libc/sys/pledge.2
P lib/libc/sys/socket.2
P lib/libcrypto/evp/evp.h
P lib/libcrypto/evp/evp_cipher.c
P lib/libcrypto/evp/p_lib.c
P lib/libcrypto/man/EVP_PKEY_set1_RSA.3
P lib/libcrypto/pkcs7/pkcs7.h
P regress/usr.bin/mandoc/man/RS/paragraph.in
P regress/usr.bin/mandoc/man/RS/paragraph.out_ascii
P regress/usr.bin/mandoc/man/RS/paragraph.out_html
P sys/arch/amd64/amd64/fpu.c
P sys/arch/amd64/amd64/trap.c
P sys/dev/pci/drm/drm_drv.c
P sys/kern/uipc_socket.c
P sys/netinet/in_pcb.c
P sys/netinet/ip_icmp.c
P sys/netinet6/in6_src.c
P sys/sys/mutex.h
P usr.bin/mandoc/man_term.c
P usr.bin/mandoc/mdoc_man.c
P usr.bin/tmux/format.c
P usr.bin/tmux/options-table.c
P usr.bin/tmux/tmux.1
P usr.bin/tmux/tmux.h
M usr.sbin/bgpd/session.c
P usr.sbin/rpki-client/cert.c

commit CrW0Af2KEXxc4mN1
Author: bluhm
Date: 2025/07/02 21:28:46

  In AMD SEV-ES guest #VC trap asserts interrupts are disabled.

  Instead of disabling interrupts again, use kassert to verify that
  #VC trap stub in vector did not enable interrupts before calling
  vctrap().  Add a comment about the general workings of #VC handling.
  Prompted and provided by mlarkin@.

  from hshoexer@; OK mlarkin@

  sys/arch/amd64/amd64/trap.c

commit 6BfdTUa24bUFIjl0
Author: schwarze
Date: 2025/07/02 21:14:00

  more tests of .RS/.PP interaction, related to man_term.c rev. 1.199

  regress/usr.bin/mandoc/man/RS/paragraph.in
  regress/usr.bin/mandoc/man/RS/paragraph.out_ascii
  regress/usr.bin/mandoc/man/RS/paragraph.out_html

commit 7FzhldEDo3tJPvpj
Author: schwarze
Date: 2025/07/02 20:50:09

  Fix the logic for printing or skipping vertical spacing before
  a paragraph if the paragraph is at the beginning of an .RS block.
  In this case, rather than always printing the spacing as we did in
  the past, do the spacing as if the .RS block wouldn't be there.
  The new behaviour is compatible with groff-1.23.

  This matters for practical purposes because some LLVM manual pages
  are generated with the rst2man(1) program, which emits insane code
  in various respects.  In some versions, it relies on the finer quirks
  of how .RS interacts with paragraph macros.

  usr.bin/mandoc/man_term.c

commit Uqp3iX10YhQFMgtx
Author: schwarze
Date: 2025/07/02 19:57:41

  Putting .PP in front of .RS results in unreliable and surprising semantics.
  For example, with groff-1.23, the sequence
  text  .PD 0  .PP  .PD  .RS  .PP  text
  does not result in a blank line between the two text lines.
  Consequently, avoid .PP in front of .RS in general.

  usr.bin/mandoc/mdoc_man.c

commit Pr7PY325Xk9kX73B
Author: mvs
Date: 2025/07/02 16:44:40

  Move m_get(9) and m_split(9) calls out of mutex(9) protected section
  within the "while (((rcvstate & SS_RCVATMARK).." loop of somove(). The
  loop operates with local data so this is possible. The calls could sleep
  while somove() called from sosplice().

  ok bluhm

  sys/kern/uipc_socket.c

commit bEZHCEU2HJ4jrReA
Author: schwarze
Date: 2025/07/02 16:07:02

  sprinkle some missing .Dv macros in .It heads, no text change

  lib/libc/sys/socket.2

commit kFRjqYsIKYBe2Gnt
Author: schwarze
Date: 2025/07/02 15:56:32

  fix one instance of ".Va struct stat" to use the correct macro, .Vt

  lib/libc/sys/pledge.2

commit zCIsZDiikn7oMc9q
Author: schwarze
Date: 2025/07/02 15:42:13

  tweak previous: use .Xr for manual page references
  and .Dv for object-like preprocessor macros

  lib/libc/sys/pledge.2

commit 89ajE7KzlmwZxXLW
Author: kettenis
Date: 2025/07/02 14:51:31

  Add an splassert(IPL_NONE) in fpu_kernel_enter() to catch unintended use
  of this function in interrupt context.

  ok claudio@, deraadt@

  sys/arch/amd64/amd64/fpu.c

commit sheBppzKGXPsCxIM
Author: claudio
Date: 2025/07/02 14:50:05

  Rewrite m_getuio() to better align data in the mbufs

  Especially for datagram traffic data should be put at the end of a buffer
  that also holds at least max_hdr bytes extra space. Doing so should prevent
  the stack from calling m_prepend() later on.

  Also handle cluster allocations differently. If there is no cluster available
  just wait for memory and stop building stupidly long mbuf chains.
  In the very old times the number of clusters was very limited and so this
  was kind of needed but this has not been the case for a long time.
  Also most drivers will need to m_defrag() such a chain anyway which requires
  the same mbuf cluster to succeed.

  Joint work with dlg@
  OK bluhm@

  sys/kern/uipc_socket.c

commit VH6lxrXSMxhoBgmX
Author: claudio
Date: 2025/07/02 14:36:56

  Make the mtx_owner pointer volatile in struct db_mutex else
  db_mtx_enter() may not work properly.
  OK kettenis@ miod@

  sys/sys/mutex.h

commit EYZ6fog7H4v3cMyR
Author: tb
Date: 2025/07/02 14:30:00

  rpki-client: move AKI handling to cert_parse_extensions()

  With this, the extension parsing switch is complete.

  ok job

  usr.sbin/rpki-client/cert.c

commit fn4bbllgdtPyBvcJ
Author: tb
Date: 2025/07/02 14:28:33

  rpki-client: move SKI handling to cert_parse_extensions()

  ok job

  usr.sbin/rpki-client/cert.c

commit hKrKk9u8PeXc9HMs
Author: deraadt
Date: 2025/07/02 13:25:05

  use a local definition of MINIMUM

  bin/ps/print.c

commit lvxkLh9gP574qF2w
Author: deraadt
Date: 2025/07/02 13:24:48

  stop using an extern variable to collect errors; ok tedu millert

  bin/ps/extern.h
  bin/ps/keyword.c
  bin/ps/nlist.c
  bin/ps/ps.c

commit w4EvaKXGKzrXshbX
Author: deraadt
Date: 2025/07/02 13:08:21

  Describe the "dns" promise in more detail, as being distinct from "inet".
  This is a pretty important factor that makes it impossible for other
  operating systems to do pledge..

  lib/libc/sys/pledge.2

commit lffKqHoSzn36lAzi
Author: jsg
Date: 2025/07/02 12:19:26

  run drm_managed cleanup from drm detach

  avoids a 'pool busy: still out' panic seen when radeondrm firmware is
  missing on non-efi installs

  reported by landry@

  sys/dev/pci/drm/drm_drv.c

commit cIXKdc5MtcbxTtS6
Author: tb
Date: 2025/07/02 11:23:25

  rpki-client: add an X509 reference to cert early on

  In order to deal with SKI and AKI, it's cleaner to have the libcrypto
  cert available in struct cert when parsing them so that the extension
  handlers can all have the same signature. Hoisting the assignment up
  in cert_parse_ee_cert() for this is very simple.

  In cert_parse_pre() we own the X509 from the start, so we take an extra
  reference which we must release before exit. In the error path there's
  an X509_free() and cert_free() releases the extra reference.

  Again this will become a bit simpler in a few more steps.

  ok job

  usr.sbin/rpki-client/cert.c

commit Qb835v6zzmuPsxW1
Author: tb
Date: 2025/07/02 11:13:34

  rpki-client: add CRLDP parsing to cert_parse_extensions()

  Add a version of x509_get_crl() to cert_parse_extensions(). One change
  is that we disallow TA certs and the other one is that the deserialization
  matches all the other handlers. Also populate cert->crl directly.

  As with the other nearly duplicated code, the old x509_get_crl()
  will be removed later on.

  ok job

  usr.sbin/rpki-client/cert.c

commit 7X2v9O5VRCRfhY7w
Author: tb
Date: 2025/07/02 11:11:39

  rpki-client: annotate basic constraints and (extended) key usage

  These are handled by cert_parse_pre(), add comments to make it more
  obvious that we're close to being done with this switch (finally).

  ok job (as part of a larger diff)

  usr.sbin/rpki-client/cert.c

commit rIKwmlY1gRlwSkRm
Author: tb
Date: 2025/07/02 10:24:17

  pkcs7.h: remove whitespace before opening paren.

  fixes in particular ./check_complete.pl pkcs7

  lib/libcrypto/pkcs7/pkcs7.h

commit JXxJGFJodyagp5hi
Author: nicm
Date: 2025/07/02 08:13:09

  Add sorting to W, P, L operators as well, and add some new session
  format variables. From Michael Grant in GitHub issue 4516.

  usr.bin/tmux/format.c
  usr.bin/tmux/options-table.c
  usr.bin/tmux/tmux.1
  usr.bin/tmux/tmux.h

commit B0BjViuz5keEJgzK
Author: tb
Date: 2025/07/02 06:40:28

  Const correct the documentation of EVP_PKEY_get{0,1}_{DH,DSA,EC_KEY,RSA}()

  lib/libcrypto/man/EVP_PKEY_set1_RSA.3

commit YhmtMjAbNK6lqype
Author: tb
Date: 2025/07/02 06:36:52

  Const correct EVP_PKEY_get{0,1}_{DH,DSA,EC_KEY,RSA}()

  These are safe to call concurrently and they don't modify the memory
  region pointed to by the pkey - they only bump the refcount of the
  key hanging off of it. The returned "legacy" key has to be handled with
  care in threaded constexts, so it is handed back as non-const. This also
  matches what EVP_PKEY_get0() always had.

  This way our signature is identical to BoringSSL's and doesn't cause
  compiler warnings in code that overuses const because one of the many
  API incoherencies added by OpenSSL 3 was to turn get0 into a function
  that takes and returns const while leaving get1 as it was.

  dlg agrees
  ok kenjiro

  lib/libcrypto/evp/evp.h
  lib/libcrypto/evp/p_lib.c

commit wi8Smuv3mKJVjsDc
Author: tb
Date: 2025/07/02 06:27:44

  Fix documented EVP_PKEY{,_base}_id() signature

  These have been taking a const pkey ever since they were added in
  OpenSSL 1.0.0.

  lib/libcrypto/man/EVP_PKEY_set1_RSA.3

commit pJ89ANYMzEM8eXhg
Author: tb
Date: 2025/07/02 06:19:46

  EVP_CipherInit_ex(): normalize EVP_CIPHER_CTX_ctrl() error check

  While EVP_CIPHER_CTX_ctrl() can return a negative value this can't
  actually happen currently as all ciphers with EVP_CIPH_CTRL_INIT set
  normalize the EVP_CTRL_INIT return value to boolean in their ctrl()
  methods. Still, this check looks weird in grep, so align it.

  ok beck kenjiro

  lib/libcrypto/evp/evp_cipher.c

commit 9IZKK57PTOzpMUzQ
Author: dlg
Date: 2025/07/02 05:44:46

  have route sourceaddr use RTF_GATEWAY to decide when to kick in.

  previously it used !RTF_HOST and !RTF_LLINFO.

  the intention with route sourceaddr was to use it except when a
  peer was on link. however, it is possible to have host routes (ie,
  RTF_HOST) via a gateway, which ended up not using the route sourceaddr
  when it should not have. by definition any route with RTF_GATEWAY
  set is not directly connected, so using it seems to better suit
  what route sourceaddr is doing.

  discussed with and ok claudio@ denis@
  tested by denis@

  sys/netinet/in_pcb.c
  sys/netinet/ip_icmp.c
  sys/netinet6/in6_src.c