Pull up following revision(s) (requested by riastradh in ticket #1280):
sys/dev/pci/virtio.c: revision 1.85
virtio(4): Add missing BUS_DMASYNC_PREREAD operations.
And one missing BUS_DMASYNC_POSTREAD operation.
With this change, loads from vq->vq_used->flags (which occur
immediately after a transfer is submitted to test whether we need to
kick the host device) is separated by a PREREAD/POSTREAD cycle from
loads from vq->vq_used->idx (which occur when we think a transfer may
have completed, e.g. upon receiving an interrupt, to test whether it
has, in fact, completed).
Additionally, with this change, consecutive loads from
*vq->vq_avail_event are separated by a PREREAD/POSTREAD cycle.
Should fix virtio(4) issues on m68k and other related architectures
like mips and armv<7:
PR kern/60144: virtio(4) cache coherence issue
[152 lines not shown]
Pull up following revision(s) (requested by riastradh in ticket #332):
sys/dev/pci/virtio.c: revision 1.85
virtio(4): Add missing BUS_DMASYNC_PREREAD operations.
And one missing BUS_DMASYNC_POSTREAD operation.
With this change, loads from vq->vq_used->flags (which occur
immediately after a transfer is submitted to test whether we need to
kick the host device) is separated by a PREREAD/POSTREAD cycle from
loads from vq->vq_used->idx (which occur when we think a transfer may
have completed, e.g. upon receiving an interrupt, to test whether it
has, in fact, completed).
Additionally, with this change, consecutive loads from
*vq->vq_avail_event are separated by a PREREAD/POSTREAD cycle.
Should fix virtio(4) issues on m68k and other related architectures
like mips and armv<7:
PR kern/60144: virtio(4) cache coherence issue
[152 lines not shown]
Pull up following revision(s) (requested by riastradh in ticket #1279):
sys/dev/pci/virtio_pci.c: revision 1.56
virtio(4): Allow virtio 0.9 BAR0 type to be memory rather than I/O.
This matches virtio>=1.0, and can't break working `hardware': any
existing virtio devices that worked must have reported I/O-type BAR0,
so they will continue to work; this will only enable previously
unusable virtio devices, reporting memory-type BAR0, to work.
Patch from Petri Koistinen.
PR kern/60247: virtio(4): legacy attach fails when BAR0 is MMIO
Pull up following revision(s) (requested by riastradh in ticket #331):
sys/dev/pci/virtio_pci.c: revision 1.56
virtio(4): Allow virtio 0.9 BAR0 type to be memory rather than I/O.
This matches virtio>=1.0, and can't break working `hardware': any
existing virtio devices that worked must have reported I/O-type BAR0,
so they will continue to work; this will only enable previously
unusable virtio devices, reporting memory-type BAR0, to work.
Patch from Petri Koistinen.
PR kern/60247: virtio(4): legacy attach fails when BAR0 is MMIO
Pull up following revision(s) (requested by riastradh in ticket #1277):
tests/net/if_wg/t_misc.sh: revision 1.17
tests/net/if_wg/t_misc.sh: revision 1.18
sys/net/if_wg.c: revision 1.136
tests/net/if_wg/t_basic.sh: revision 1.7
sys/net/if_wg.c: revision 1.137
tests/net/if_wg/t_basic.sh: revision 1.8
sys/net/if_wg.c: revision 1.138
tests/net/if_wg/t_basic.sh: revision 1.9
tests/net/if_wg/common.sh: revision 1.2
wg(4): Add test case for bad peer public keys.
wg(4) should not crash on an assertion if they are used -- it should
just gracefully accept them, with degraded security, since a peer
that maliciously provides an invalid public key is no worse than a
peer that voluntarily exposes all its plaintext anyway.
PR kern/60106: wg(4) should properly handle invalid or insecure
[45 lines not shown]
Pull up following revision(s) (requested by riastradh in ticket #329):
tests/net/if_wg/t_misc.sh: revision 1.17
tests/net/if_wg/t_misc.sh: revision 1.18
sys/net/if_wg.c: revision 1.136
tests/net/if_wg/t_basic.sh: revision 1.7
sys/net/if_wg.c: revision 1.137
tests/net/if_wg/t_basic.sh: revision 1.8
sys/net/if_wg.c: revision 1.138
tests/net/if_wg/t_basic.sh: revision 1.9
tests/net/if_wg/common.sh: revision 1.2
wg(4): Add test case for bad peer public keys.
wg(4) should not crash on an assertion if they are used -- it should
just gracefully accept them, with degraded security, since a peer
that maliciously provides an invalid public key is no worse than a
peer that voluntarily exposes all its plaintext anyway.
PR kern/60106: wg(4) should properly handle invalid or insecure
[45 lines not shown]
Pull up following revision(s) (requested by jdc in ticket #328):
sys/arch/alpha/conf/INSTALL: revision 1.123
Add bge (3X-DEGXA-T*) to the INSTALL configuration too.
Requested by Janne Ramstedt.
Pull up following revision(s) (requested by mlelstv in ticket #326):
sys/dev/scsipi/sd.c: revision 1.346
For SPC-3 devices try READ_CAPACITY_16 first and fall back to older
READ_CAPACITY_10.
This fixes SCSI drivers like mpt(4) that announce SPC-3 conformity but
don't support READ_CAPACITY_16 in all configurations.
Add support for MPC5200B SoC, including:
- BestComm DMA controller
- Clock Distribution Module
- Fast Ethernet Controller, utilizing BestComm DMA
- AC97 controller, utilizing BestComm DMA
- ATA controller (PIO-only, hardware DMA is too broken)
- On-chip SRAM (needed for BestComm)
- PSC UART (for console)
- PCI along with bus mater DMA support
All of this is based on previous work by Robert Swindells and would not
be possible if it wasn't for his effort.
Pull up following revision(s) (requested by riastradh in ticket #325):
tests/usr.sbin/inetd/t_accept_max.sh: revision 1.3
tests/usr.sbin/inetd/t_accept_max.sh: revision 1.4
inetd(8): Fix sh(1) trap save/restore in accept-max test.
Fixes mysterious
Failed: 1
failures like this one:
https://releng.netbsd.org/b5reports/i386/2026/2026.06.22.22.27.17/test.html#usr.sbin_inetd_t_accept_max_max2_kv
PR bin/59645: inetd `rate-limiting' algorithm is stupid
tests/usr.sbin/inetd/t_accept_max: Handle SIGKRE.
Fix some cleanup issues and simplify according to most of kre's
suggestions. Mostly this fixes the kill-background-jobs-on-trap
[13 lines not shown]
Pull up following revision(s) (requested by wiz in ticket #318):
lib/libc/time/strptime.c: revision 1.68
Fix day of week calculcation in strptime().
The bug was that the calculation used the raw year value without
adding the 1900 offset.
While here, switch to the current Wikipedia version of Gauss' algorithm,
like FreeBSD did, and simplify a loop, like FreeBSD did.
Pull up following revision(s) (requested by tls in ticket #317):
sys/dev/acpi/acpi.c: revision 1.305
sys/dev/acpi/acpi.c: revision 1.306
Hyper-V passes bootstrap entropy to guests via the ACPI OEM0 table. Use it.
We verify that the OEM0 table has the correct Microsoft IDs, skip the header,
feed the remaining contents to the kernel entropy pool, and wipe them in-place
to prevent later extraction via /dev/acpi (e.g. with acpidump). This interface
doesn't seem to be documented but we handle it the same way Microsoft's own
Linux driver does.
acpi(4): Leave Hyper-V rndsource attached.
This way, it will be reported through rndctl(8) rather than only by
grepping through dmesg, and rndctl(8) can control it, e.g. if it
turns out something is wrong with that but not with RDSEED on the
same virtual CPU, you can use `rndctl -Ed "Hyper-V OEM0"' to discount
[9 lines not shown]
Pull up the following, requested by christos in ticket #315:
external/mpl/bind/dist/bin/tests/system/class/ns1/chaos.db.in up to 1.1.1.1
external/mpl/bind/dist/bin/tests/system/class/ns1/named.conf.j2 up to 1.1.1.1
external/mpl/bind/dist/bin/tests/system/class/ns2/example.db.in up to 1.1.1.1
external/mpl/bind/dist/bin/tests/system/class/ns2/localhost.db.in up to 1.1.1.1
external/mpl/bind/dist/bin/tests/system/class/ns2/named.conf.j2 up to 1.1.1.1
external/mpl/bind/dist/bin/tests/system/class/setup.sh up to 1.1.1.1
external/mpl/bind/dist/bin/tests/system/class/tests_class_chaos.py up to 1.1.1.1
external/mpl/bind/dist/bin/tests/system/class/tests_class_update.py up to 1.1.1.1
external/mpl/bind/dist/bin/tests/system/class/ns3/named.conf.j2 up to 1.1.1.1
external/mpl/bind/dist/bin/tests/system/cap_glues/ns1/named.conf.j2 up to 1.1.1.1
external/mpl/bind/dist/bin/tests/system/cap_glues/ns1/root.db up to 1.1.1.1
external/mpl/bind/dist/bin/tests/system/cap_glues/ns2/named.conf.j2 up to 1.1.1.1
external/mpl/bind/dist/bin/tests/system/cap_glues/ns2/tld.db up to 1.1.1.1
external/mpl/bind/dist/bin/tests/system/cap_glues/tests_cap_glues.py up to 1.1.1.1
external/mpl/bind/dist/bin/tests/system/cap_glues/ns3/named.conf.j2 up to 1.1.1.1
external/mpl/bind/dist/bin/tests/system/cap_glues/ns3/root.hint up to 1.1.1.1
external/mpl/bind/dist/bin/tests/system/catz/ns1/catalog-bad6.example.db up to 1.1.1.1
[396 lines not shown]
Pull up following revision(s) (requested by christos in ticket #1274):
sys/kern/kern_exec.c: revision 1.533
Prevent a DoS attack by crafted binaries that trigger KASSERTS. Change the
KASSERTS into errors instead. Reported by Andrew Griffiths.
Pull up following revision(s) (requested by christos in ticket #314):
sys/kern/kern_exec.c: revision 1.533
Prevent a DoS attack by crafted binaries that trigger KASSERTS. Change the
KASSERTS into errors instead. Reported by Andrew Griffiths.
Pull up following revision(s) (requested by christos in ticket #1273):
sys/nfs/nfs_serv.c: revision 1.185
sys/nfs/nfs_srvsocket.c: revision 1.7
sys/nfs/nfs_syscalls.c: revision 1.165
sys/nfs/nfs.h: revision 1.83
Prevent a DoS attack by stuffing the server with too many fragments that
can lead to resource exhaustion and memory corruption. Reported by Andrew Griffiths
Validate the rpc readdir and readdirplus cnt, siz to prevent integer overflows
that will corrupt memory. Reported by Andrew Griffiths
Fix memory leak. ns_frag holds the unprocessed fragments that needs to be freed.
Reported by Andrew Griffiths
Pull up following revision(s) (requested by christos in ticket #313):
sys/nfs/nfs_serv.c: revision 1.185
sys/nfs/nfs_srvsocket.c: revision 1.7
sys/nfs/nfs_syscalls.c: revision 1.165
sys/nfs/nfs.h: revision 1.83
Prevent a DoS attack by stuffing the server with too many fragments that
can lead to resource exhaustion and memory corruption. Reported by Andrew Griffiths
Validate the rpc readdir and readdirplus cnt, siz to prevent integer overflows
that will corrupt memory. Reported by Andrew Griffiths
Fix memory leak. ns_frag holds the unprocessed fragments that needs to be freed.
Reported by Andrew Griffiths
uvm_km_pgremove_intrsafe: drain TLB shootdown before freeing pages
Call pmap_update(pmap_kernel()) between pmap_kremove and the page
free so the shootdown completes first.
Reviewed by skrll@
mips: NMI handler must also clear STATUS.ERL
STATUS.ERL may be set in addition to STATUS.EXL and STATUS.NMI
when entering the NMI handler.
Clearing only EXL and NMI leaves the CPU in error level.
Reviewed by skrll@
mips: avoid redundant STATUS write in _splraise
The previous test only short-circuited when the requested IPL was
strictly less than the current IPL. When newipl == curipl,
short-circuit as well.
Reviewed by skrll@
Pull up following revision(s) (requested by martin in ticket #344):
usr.sbin/sysinst/Makefile.inc: revision 1.55
usr.sbin/sysinst/configmenu.c: revision 1.26
usr.sbin/sysinst/defs.h: revision 1.97
usr.sbin/sysinst/main.c: revision 1.38
usr.sbin/sysinst/menus.mi: revision 1.32-1.33
usr.sbin/sysinst/net.c: revision 1.48
Do not remove support for https neither at compile time (SMALLPROG)
nor at runtime (no SSL trust anchors available).
Instead dynamically disable the https option when it won't work, but
fully support it later (after unpacking the base set) when downloading
stuff for binary pkgs or pkgsrc.
--
#ifdef remove_menu_options() like all its callers to avoid stupid
compiler warnings.
