Generate hashes for bootable images in gzimg directory.
At one point this should probably be unified at least a bit with the hash
generation of regular sets, so that there is a single place that defines
the types of hashes generated for sets and gzimg.
iconv VIQR: Fix use-after-free during destruction.
PR lib/59019: various iconv issues
Prompted by investigating:
PR lib/60413: iconv_samples test crashes on sparc*
Fixed in FreeBSD back in 2022, and I confirmed -Wuse-after-free catches
this for us too, but we're not using it -- maybe we should:
commit 8f27c9d14a559f33aa7fc3245f841f7ce52fadd5
Author: John Baldwin <jhb at FreeBSD.org>
Date: Mon Oct 3 16:10:43 2022 -0700
libiconv VIQR: Fix a use after free.
Use TAILQ_FOREACH_SAFE to walk to list of children mnemonics to
free
[5 lines not shown]
Pull up following revision(s) (requested by mrg in ticket #372):
external/mit/xorg/lib/libEGL/Makefile: revision 1.10
external/mit/xorg/lib/driver.mk: revision 1.9
external/mit/xorg/lib/libgbm/Makefile: revision 1.11
external/mit/xorg/lib/libGL/Makefile: revision 1.36
ensure HAVE_NOATEXIT is properly set for all that need it.
Increase size of bpfjit interal stack, follow bpf_validate() closely.
Increase a size of M[] array allocated on the stack by bpfjit.
When bpf_validate() fails, bpfjit now fails too:
- Fail to validate/compile BVF_DIV+BPF_K and BPF_MOD+BPF_K with K=0.
- Fail unless the last instruction is BPF_RET+BPF_A or BPF_RET+BPF_K.
If a program contains additional BPF_RET instructions, it is accepted
by bpf_validate() and it compiles fine by bpfjit but both bpf_filter()
and bpfjit reject it (return 0) at runtime.
Adjust existing bpfjit test to follow the changes, add new tests.
defflag M68K_MMU_CUSTOM. Checks for this were added a while ago, but
the defflag was missed. It's not expected to be used directly by
configus, but is there as something other MMU options can add as a
dependency.
Pull up following revision(s) (requested by gson in ticket #371):
usr.sbin/sysinst/run.c: revision 1.17
usr.sbin/sysinst/defs.h: revision 1.98
usr.sbin/sysinst/util.c: revision 1.85
usr.sbin/sysinst/main.c: revision 1.39
Make sysinst intercept kernel console messages using TIOCCONS at all
times, not just when running an external command, and discard those
that occur when not running an external command so that they don't
mess up the display when running sysinst on the console. Fixes
PR install/54932.
ahcisata_pci: Add sanity check for BAR5 to prevent panic
Inte volume management can't be attach to ahcisata_pci.
It doesn't conform the AHCI specification. The AHCI specification says it's
32bit BAR. According to the PCI specification, if a BAR is the lower 32-bit
register of a 64-bit register, the next register represents the upper 32
bits. However, BARs only go up to 5, and there is no BAR 6.
Pull up following revision(s) (requested by msaitoh in ticket #2032):
sys/dev/usb/xhci.c: revision 1.193
xhci: Add missing UHF_PORT_POWER support to xhci_roothub_ctrl_locked()
Pull up following revision(s) (requested by msaitoh in ticket #1305):
sys/dev/usb/xhci.c: revision 1.193
xhci: Add missing UHF_PORT_POWER support to xhci_roothub_ctrl_locked()
Pull up following revision(s) (requested by msaitoh in ticket #367):
sys/dev/usb/xhci.c: revision 1.193
xhci: Add missing UHF_PORT_POWER support to xhci_roothub_ctrl_locked()
Pull up following revision(s) (requested by kre in ticket #366):
sbin/efi/devpath.c: revision 1.2
PR bin/60403 : sbin/efi - fix core dump from assert in devpath
assert(3) should never be used to validate data from external sources,
only to validate assumptions about how the code itself works.
Worse here, the assert() wasn't even validating that the externally sourced
data was correctly formed, which it was; it was just in a format that wasn't
common, and wasn't expected.
What the output should be (the textual representation of the variable
value) is unknown to me, I haven't been able to locate anything in the
EFI standard which says (or even gives an example of the case in question).
It might be there though somewhere I haven't found yet.
The unusual case is when a variable which contains a path, like:
Boot0000* NetBSD
[9 lines not shown]
Additionally pull up following revision(s) (requested by riastradh in ticket #308):
distrib/sets/lists/tests/mi: revision 1.1408
distrib/sets/lists/comp/mi: revision 1.2526
new openssl test file
new OpenSSL man page
