Build date: 1736118003 - Sun Jan  5 23:00:03 UTC 2025
Build cvs date: 1736103123 - Sun Jan  5 18:52:03 UTC 2025
Build id: 2025-01-06.1
Build tags: amd64-regress ports sysupgrade

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

P lib/libcrypto/ec/ec_convert.c
P lib/libcrypto/ec/ec_lib.c
P lib/libcrypto/ec/ec_local.h
P lib/libcrypto/ec/ec_pmeth.c
P lib/libcrypto/ec/ecp_methods.c
P lib/libcrypto/man/RSA_get0_key.3
P lib/libcrypto/man/RSA_meth_new.3
P lib/libcrypto/rsa/rsa_local.h
P lib/libcrypto/rsa/rsa_sign.c
P regress/lib/libcrypto/rsa/rsa_method_test.c
P regress/usr.bin/mandoc/tbl/mod/width.in
P regress/usr.bin/mandoc/tbl/mod/width.out_ascii
P sys/kern/uipc_socket2.c
P sys/kern/uipc_syscalls.c
P sys/net/af_frame.c
P sys/netinet/in_pcb.h
P sys/netinet/in_proto.c
P sys/netinet/tcp_input.c
P sys/netinet/tcp_output.c
P sys/netinet/tcp_timer.c
P sys/netinet/tcp_usrreq.c
P sys/netinet/tcp_var.h
P sys/netinet6/in6_proto.c
P sys/sys/protosw.h
P usr.bin/mandoc/libmandoc.h
P usr.bin/mandoc/mandoc.1
P usr.bin/mandoc/mandoc.h
P usr.bin/mandoc/mandoc_msg.c
P usr.bin/mandoc/out.c
P usr.bin/mandoc/roff.c
P usr.bin/mandoc/tbl.c
P usr.bin/mandoc/tbl.h
P usr.bin/mandoc/tbl_layout.c

commit gSvy2z9EipQ48FHb
Author: schwarze
Date: 2025/01/05 18:52:03

  spelling; requested by jmc@

  usr.bin/mandoc/mandoc.1

commit QGAvPnVnJvTULlpt
Author: tb
Date: 2025/01/05 18:21:36

  Let's use RSA_3 rather than 3

  regress/lib/libcrypto/rsa/rsa_method_test.c

commit kuvd0BV0LWHVKe3K
Author: schwarze
Date: 2025/01/05 18:03:51

  In the tbl(7) layout parser, use the roff(7) numerical expression parser
  to support arithmetic operations in column width modifiers.

  Documentation authors ought to focus on semantic markup and avoid trying
  to manipulate presentational details.  Consequently, using tbl(7) is
  discouraged in the first place, using column width modifiers even more
  so, and complicating them by performing arithmetics yet more so.

  However, the madhouse called DocBook relies on this feature.
  For example, DocBook XSL version 1.79.2 generates man(7)+tbl(7) code
  that abuses arithmetics in column width modifiers to format things
  as simple and mundane as numbered lists, resulting in atrocious layout
  that is very hard to read.  Last year, tb@ reported git-reset(1)
  as an example of an affected manual page.

  Properly parsing the width in the tbl(7) parser allows removal of some
  incomplete ad-hoc parsing from the tbl(7) formatter, where it didn't
  really belong.

  regress/usr.bin/mandoc/tbl/mod/width.in
  regress/usr.bin/mandoc/tbl/mod/width.out_ascii
  usr.bin/mandoc/mandoc.1
  usr.bin/mandoc/mandoc.h
  usr.bin/mandoc/mandoc_msg.c
  usr.bin/mandoc/out.c
  usr.bin/mandoc/tbl.c
  usr.bin/mandoc/tbl.h
  usr.bin/mandoc/tbl_layout.c

commit TaVbCMrhcG1jIsWN
Author: schwarze
Date: 2025/01/05 16:56:48

  Add an additional argument to the functions implementing the roff(7)
  numerical parser to select the default scaling unit for numbers not
  followed by a scaling unit, rather than always using basic units ('u').

  Expose roff_evalnum() via the internal parser API such that, in addition
  to the roff(7) parser, other parsers become able to parse numerical
  expressions, too.

  This commit implies no functional change yet,
  but prepares for adding new functionality in future commits.

  usr.bin/mandoc/libmandoc.h
  usr.bin/mandoc/roff.c

commit Z9dzNQ26YbmSrL8A
Author: tb
Date: 2025/01/05 16:07:08

  Move BIGNUMs in EC_GROUP and EC_POINT to the heap

  The only way to get an EC_GROUP or an EC_POINT is by calling the relevant
  _new() function and to get rid of it, something must call _free(). Thus we
  can establish the invariant that every group has Weierstrass coefficients
  p, a, b as well as order and cofactor hanging off it. Similarly, Every
  point has allocated BIGNUMs for its Jacobian projective coordinates.

  Unfortunately, a group has the generator as an optional component in
  addition to seed and montgomery context/one (where optionality makes
  more sense).

  This is a mostly mechanical diff and only drops a few silly comments and
  a couple of unnecessary NULL checks since in our part of the wrold the
  word invariant has a meaning.

  This should also appease Coverity who likes to throw fits at calling
  BN_free() for BIGNUM on the stack (yes, this is actually a thing).

  ok jsing

  lib/libcrypto/ec/ec_convert.c
  lib/libcrypto/ec/ec_lib.c
  lib/libcrypto/ec/ec_local.h
  lib/libcrypto/ec/ec_pmeth.c
  lib/libcrypto/ec/ecp_methods.c

commit TjOkjogYPH5Vxd0y
Author: tb
Date: 2025/01/05 15:42:48

  Stop setting RSA_FLAG_SIGN_VER

  With rsa_sign.c r1.37 this is no longer needed.

  regress/lib/libcrypto/rsa/rsa_method_test.c

commit EmW1syk8Oq5An44z
Author: tb
Date: 2025/01/05 15:40:42

  Remove most of the RSA_FLAG_SIGN_VER documentation

  ok jsing kn

  lib/libcrypto/man/RSA_get0_key.3
  lib/libcrypto/man/RSA_meth_new.3

commit FqDlgOXR1uOh3NEz
Author: tb
Date: 2025/01/05 15:39:12

  Stop requiring the RSA_FLAG_SIGN_VER

  You can set custom sign and verify handlers on an RSA method (wihch is
  used to create RSA private and public key handles). However, even if you
  set them explicitly with RSA_meth_set_{sign,verify}(3), these handlers
  aren't used for the sake of "backward compatibility" (with what?). In order
  to use them, you need to opt your objects into using the custom methods
  you set by setting the RSA_FLAG_SIGN_VER flag.

  OpenSSL 1.1 dropped this requirement and therefore nobody sets this flag
  anyore. Like most of the mechanically added accessors, almost nothing
  uses them, but, as found by kn, the yubco-piv-tool does. This resulted
  in a public key being passed to rsa_private_encrypt(), which of course
  doesn't end well.

  So follow OpenSSL 1.1 and drop this muppetry. This makes kn's problem
  with yubico-piv-tool go away.

  ok jsing kn

  lib/libcrypto/rsa/rsa_local.h
  lib/libcrypto/rsa/rsa_sign.c

commit 32lFGvX1P5GLoq9n
Author: bluhm
Date: 2025/01/05 12:36:48

  Retire PR_MPSOCKET flag.

  TCP socket layer is MP safe for more than a week now.  That means
  all protocols with pr_usrreqs have the PR_MPSOCKET flag.  Remove
  PR_MPSOCKET and use the logic that was used when set.

  OK mvs@

  sys/kern/uipc_socket2.c
  sys/net/af_frame.c
  sys/netinet/in_proto.c
  sys/netinet6/in6_proto.c
  sys/sys/protosw.h

commit xi7EhXsgqInnDiGc
Author: bluhm
Date: 2025/01/05 12:23:38

  TCP integer sysctl variables are all atomic.  Remove net lock.

  OK mvs@

  sys/netinet/tcp_usrreq.c

commit 6PPYN8YyhjczvPCp
Author: bluhm
Date: 2025/01/05 12:18:48

  Read tcp_do_ecn only once during packet processing and mark atomic.

  OK mvs@

  sys/netinet/tcp_input.c
  sys/netinet/tcp_output.c
  sys/netinet/tcp_timer.c
  sys/netinet/tcp_var.h

commit zfVRdiL9YV5ybBOV
Author: bluhm
Date: 2025/01/05 12:10:39

  Reorder struct inpcb fields and shorten struct inpcb_iterator.

  OK mvs@

  sys/netinet/in_pcb.h

commit Od0ri2dKBkhn80oY
Author: mvs
Date: 2025/01/05 11:33:45

  Unlock accept(4) for tcp sockets.

  Makes sense because accept(2) could be fast path. tcp_accept() is the
  only in_setpeeraddr() which copies `inp_fport' and `inp_faddr' to passed
  mbuf(9). Shared netlock with `so_lock' taken is pretty enough, but both
  sockets should be locked simultaneously.

  So yet another solock() variation. The second arg of doaccept_so*lock()
  controls shared netlock acquisition and release. tcp(4) sockets are
  PR_MPSOCKET sockets, so soleep_nsec() will be happy. We have some raw
  inet6 sockets which are not PR_MPSOCKET, but they never follow this
  path.

  Note, we modify `so_qlen' of listening socket but filt_soread() takes
  only shared netlock. This should be enough because we cache `so_qlen'
  to local variable.

  ok bluhm

  sys/kern/uipc_syscalls.c