Build date: 1755295203 - Fri Aug 15 22:00:03 UTC 2025
Build cvs date: 1755292189 - Fri Aug 15 21:09:49 UTC 2025
Build id: 2025-08-16.1
Build tags: amd64-regress ports 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/ps.1
P distrib/sets/lists/comp/mi
P etc/rc.d/rc.subr
P gnu/usr.sbin/mkhybrid/mkhybrid/Makefile
P regress/lib/libcrypto/mlkem/Makefile
P regress/lib/libcrypto/mlkem/mlkem_iteration_tests.c
P regress/lib/libcrypto/mlkem/mlkem_tests.c
P regress/lib/libcrypto/mlkem/mlkem_tests_util.c
P regress/lib/libcrypto/mlkem/mlkem_tests_util.h
P regress/lib/libcrypto/mlkem/mlkem_unittest.c
P share/man/man8/rc.d.8
P sys/arch/amd64/amd64/machdep.c
P sys/arch/amd64/amd64/pmap.c
P sys/arch/arm64/conf/GENERIC
P sys/arch/arm64/conf/RAMDISK
P sys/arch/i386/i386/pmap.c
P sys/arch/i386/i386/pmapae.c
U sys/dev/fdt/bcmstbpinctrl.c
P sys/dev/fdt/files.fdt
P sys/dev/fdt/sdhc_fdt.c
P sys/dev/pci/if_ice.c
P sys/kern/exec_subr.c
P sys/kern/kern_exec.c
P sys/kern/kern_resource.c
P sys/kern/subr_prof.c
P sys/net/art.c
P sys/sys/proc.h
P sys/uvm/uvm_mmap.c
P sys/uvm/uvm_vnode.c
M usr.sbin/bgpd/session.c
P usr.sbin/vmd/x86_vm.c

commit hcHBEjqJ1Nk1vGnR
Author: dv
Date: 2025/08/15 21:09:49

  vmd(8): load the full SeaBIOS image in lower bios memory.

  My change to make space for edk2 firmware images trimmed the amount
  of the image we load into the lower bios area. Instead of trimming
  to 128KB, trim to the bios area size. In the case of SeaBIOS this
  loads the full image.

  This fixes booting certain Linux guests that use boot loaders that
  (I believe) try calling into parts of the bios in this lower memory.
  (We don't shadow upper and lower bios regions.) It puts them typically
  into a spiral of hammering the PIC.

  usr.sbin/vmd/x86_vm.c

commit OUz6xEPkK2TmV04k
Author: stsp
Date: 2025/08/15 16:35:58

  print ice(4) health event debug messages in debug builds only

  sys/dev/pci/if_ice.c

commit mM8T15a7DyCcDcB4
Author: deraadt
Date: 2025/08/15 14:49:43

  delete comment about mkisofs transition; noticed by Kyle Ackerman

  gnu/usr.sbin/mkhybrid/mkhybrid/Makefile

commit PeCHZryn2qzg3yEW
Author: tb
Date: 2025/08/15 14:47:54

  Fix includes in mlkem unit and iteration tests

  regress/lib/libcrypto/mlkem/mlkem_iteration_tests.c
  regress/lib/libcrypto/mlkem/mlkem_tests_util.c
  regress/lib/libcrypto/mlkem/mlkem_tests_util.h
  regress/lib/libcrypto/mlkem/mlkem_unittest.c

commit 4NFLc7jXpq3yhYI2
Author: tb
Date: 2025/08/15 14:46:37

  mlkem_tests no longer depends on mlkem_tests_util

  regress/lib/libcrypto/mlkem/Makefile

commit 7fAw3xVSjOKwfzDL
Author: tb
Date: 2025/08/15 14:45:33

  mlkem_tests: use public mlkem.h, no longer needs mlkem_tests_util.h

  regress/lib/libcrypto/mlkem/mlkem_tests.c

commit 3dYuDcV2NK1kniKE
Author: kettenis
Date: 2025/08/15 13:40:43

  Fix pmap_write_protect() such that it supports kernel addresses by
  replacing some broken code with code from pmap_do_remove().

  ok miod@, mpi@

  sys/arch/amd64/amd64/pmap.c
  sys/arch/i386/i386/pmap.c
  sys/arch/i386/i386/pmapae.c

commit xdUI5a7XTQ94XWWI
Author: kettenis
Date: 2025/08/15 13:35:49

  Add bcmstbpinctrl(4), a driver for the pin muxing controller found on
  the rpi5.  The driver intents to support both the "C0" and the "D0"
  silicon variants, but so far we've only been to attach bwfm(4)
  sucessfully on "C0" silicon.

  This driver is incomplete and currently only supports the pins that are
  used to connect the WiFi chip.

  ok mglocker@

  sys/arch/arm64/conf/GENERIC
  sys/arch/arm64/conf/RAMDISK
  sys/dev/fdt/bcmstbpinctrl.c
  sys/dev/fdt/files.fdt

commit y5RnQ2lScmy2xKqH
Author: kettenis
Date: 2025/08/15 13:31:58

  Implement support for "vmmc-supply".  Needed to power on the WiFi chip on
  the rpi5.

  ok mglocker@

  sys/dev/fdt/sdhc_fdt.c

commit 5qNcgzV6Wx0rZlXf
Author: ajacoutot
Date: 2025/08/15 11:35:18

  Make "quiet" mode *not* quiet on error.

  ok landry@

  etc/rc.d/rc.subr
  share/man/man8/rc.d.8

commit eCSyo1kdfRnzAm5n
Author: mpi
Date: 2025/08/15 09:53:53

  No need for the KERNEL_LOCK() around uvm_map_protect().

  ok miod@, kettenis@

  sys/kern/kern_resource.c

commit Dr1HmTPt6wpNtWsV
Author: jsg
Date: 2025/08/15 09:15:55

  remove prototypes for removed functions

  sys/net/art.c

commit KZAhBA5Mp95rNIJQ
Author: mpi
Date: 2025/08/15 08:21:41

  Prevent lock recursion in error path.

  This is a fallout from pushing uvm_pageqlock dance inside uvm_pagefree().

  Reported-by: syzbot+69f651609f3f00ee7e78@syzkaller.appspotmail.com

  sys/uvm/uvm_vnode.c

commit WPcJ100z0j9C67ch
Author: mpi
Date: 2025/08/15 07:50:53

  Third argument of uvm_map_protect() is `end' address not a `length'.

  From David Higgs.

  sys/kern/exec_subr.c

commit 3hCkq4hSUIt5I40H
Author: tb
Date: 2025/08/15 07:49:12

  mlkem_tests: mop up after feral openssl devs were here

  regress/lib/libcrypto/mlkem/mlkem_tests.c

commit VYIC60C0iy4mb6uL
Author: tb
Date: 2025/08/15 07:25:19

  mlkem_unittest: undo unnecessary variable renaming

  regress/lib/libcrypto/mlkem/mlkem_unittest.c

commit d5Wn3scVI1qh8n5h
Author: anton
Date: 2025/08/15 05:29:02

  sync; ok tb@

  distrib/sets/lists/comp/mi

commit rPGyVr0iYzH4kSxH
Author: guenther
Date: 2025/08/15 04:21:00

  Move the proc.ps_flags PS_* bits that are only changed by sys_execve()
  and not tested (by name) in userland to a new ps_iflags member.  This
  frees up some ps_flags bits for other uses.

  ok mpi@ claudio@

  bin/ps/ps.1
  sys/arch/amd64/amd64/machdep.c
  sys/kern/kern_exec.c
  sys/kern/subr_prof.c
  sys/sys/proc.h
  sys/uvm/uvm_mmap.c