Fix several RAIDframe bugs relating to spare disks (most minor, one or two not)
1) Incomplete component swap - this can leave the replacement's auto_configured,
partitionSize, etc. set to the old component's values or in some cases
to zero (component missing at boot etc).
2) A write error on the spare during reconstruction should fail the spare.
Otherwise, multi-spare arrays will never proceed from a bad spare to a
good one.
3) rf_remove_hot_spare and rf_delete_component fail to close the vnode
in the autoconfigured case. rf_remove_hot_spare also appears to have
an off-by-one error.
4) When we bring a spare into operation and update its label, we fail to
call rf_component_label_set_partitionsize(). It could get stuck at 0
or whatever was in there before, since at present spares don't get labels
written in most cases until they are used.
[2 lines not shown]
mozilla-certdata: Mark two nixed certs as obsolete.
Missed a spot, must not have had a clean destdir somehow.
PR security/60908: mozilla-certdata: update for compromised CA
Pull up following revision(s) (requested by riastradh in ticket #535):
tests/lib/libc/gen/t_arc4random.c: revision 1.6
t_arc4random: Fix test to guarantee global state initialization.
Previously, if pthread_atfork failed, arc4random_global_buf would
fail to initialize the global state as intended. (There are separate
tests for intentional failure of global state initialization.) It is
unlikely for pthread_atfork to fail, but with RLIMIT_AS cranked down
to zero, it may. To ensure that the global state is successfully
initialized for the tests that are trying to exercise it, call
arc4random from another thread first before forcing failure of the
local state initialization.
PR lib/59117: arc4random has some failure modes it shouldn't
Pull up following revision(s) (requested by jnemeth in ticket #1367):
etc/mtree/NetBSD.dist.base: revision 1.265
distrib/sets/lists/base/mi: revision 1.1391
distrib/sets/lists/base/mi: revision 1.1392
etc/namedb/bind.keys: revision 1.3
Make namedb subdirectories for non-chroot case.
Part of fix for PR/52465 and PR/60864
Update bind.keys from src/external/mpl/dist/bind.keys.
managed-keys is ancient history and just generated warnings doing
nothing useful. The proper keyword to use is trust-anchors and
this file is now current. The latest root key is only partially
specified as it wasn't published at the time this file was updated,
but enough information is provided for named to download the full
key using RFC 5011.
[3 lines not shown]
Pull up following revision(s) (requested by jnemeth in ticket #534):
etc/mtree/NetBSD.dist.base: revision 1.265
distrib/sets/lists/base/mi: revision 1.1391
distrib/sets/lists/base/mi: revision 1.1392
etc/namedb/bind.keys: revision 1.3
Make namedb subdirectories for non-chroot case.
Part of fix for PR/52465 and PR/60864
Update bind.keys from src/external/mpl/dist/bind.keys.
managed-keys is ancient history and just generated warnings doing
nothing useful. The proper keyword to use is trust-anchors and
this file is now current. The latest root key is only partially
specified as it wasn't published at the time this file was updated,
but enough information is provided for named to download the full
key using RFC 5011.
[3 lines not shown]
Pull up following revision(s) (requested by tls in ticket #1366):
tools/gmp/Makefile: revision 1.7
Fix compiler core dumps on MacOS on Apple Silicon (PR toolchain/60904),
by disabling the use of the broken arm64 asm that comes with the present
version of GMP (we can re-enable it when we upgrade).
Pull up following revision(s) (requested by tls in ticket #533):
tools/gmp/Makefile: revision 1.7
Fix compiler core dumps on MacOS on Apple Silicon (PR toolchain/60904),
by disabling the use of the broken arm64 asm that comes with the present
version of GMP (we can re-enable it when we upgrade).
Pull up following revision(s) (requested by andvar in ticket #1365):
sys/dev/pci/if_vr.c: revision 1.140
vr(4): set or clear VR_RXCFG_RX_PROMISC bit in vr_setmulti().
Previously the promiscuous bit was only handled in vr_init().
When an interface is added to or removed from bridge(4), its
IFF_PROMISC flag is changed through the interface flags (SIOCSIFFLAGS).
On a running interface it does not reinitialize device, but calls
vr_setmulti() directly, which did not update the promiscuous bit in RX filter.
PR port-i386/60884
Pull up following revision(s) (requested by andvar in ticket #532):
sys/dev/pci/if_vr.c: revision 1.140
vr(4): set or clear VR_RXCFG_RX_PROMISC bit in vr_setmulti().
Previously the promiscuous bit was only handled in vr_init().
When an interface is added to or removed from bridge(4), its
IFF_PROMISC flag is changed through the interface flags (SIOCSIFFLAGS).
On a running interface it does not reinitialize device, but calls
vr_setmulti() directly, which did not update the promiscuous bit in RX filter.
PR port-i386/60884
t_arc4random: Fix test to guarantee global state initialization.
Previously, if pthread_atfork failed, arc4random_global_buf would
fail to initialize the global state as intended. (There are separate
tests for intentional failure of global state initialization.) It is
unlikely for pthread_atfork to fail, but with RLIMIT_AS cranked down
to zero, it may. To ensure that the global state is successfully
initialized for the tests that are trying to exercise it, call
arc4random from another thread first before forcing failure of the
local state initialization.
PR lib/59117: arc4random has some failure modes it shouldn't
fix the m68ksf build.
avoid using bare ${MACHINE_ARCH} when referring to subdirs that
mknative-gcc creates, and avoid overwriting ${GCC_MACHINE_ARCH}
if it is already set.
fix a test in libgomp, and explain why it exists.
