fss(4): Clarify FSSIOCSET operation for persistent snapshots.
Prompted by the persistent confusion on my part that led to:
PR bin/60383: fssconfig(8): distinguish between taking and loading a
snapshot
pmap_enter(): When overwriting an existing mapping for one to a different
PA, don't pass the vm_page * computed earlier to the pmap_remove_mapping()
call; that vm_page * is for the new PA, not the existing one we're removing
the mapping for.
rump: New rule for RUMPCOMP_USER_SRCS .go objects for MKDEBUGLIB.
Similar to the .o/.po/.pico rules already here.
PR lib/60418: -current rump/wg(4) build failure
Pull up the following revisions(s) (requested by msaitoh in ticket #2034):
usr.sbin/syslogd/tls.c: revision 1.22
Fix memory leak in match_certfile().
Found and fixed by Masakazu OOHASHI at IIJ.
Pull up the following revisions(s) (requested by msaitoh in ticket #2033):
usr.sbin/syslogd/syslogd.c: revision 1.149
syslogd: Fix a bug that tls_ca mistakenly matches tls_cadir.
Found and fixed by Masakazu OOHASHI at IIJ.
usmsc(4): Check the return value of smsc_readreg() in smsc_uno_read_reg()
Fixes PR kern/58634:
- Check the return value of smsc_readreg() in smsc_uno_read_reg() and return
the error to the caller.
- Return 0 as the register value when error occurred just for the consistency
in smsc_uno_read_reg(). Note that it's not required because
usbnet.c::uno_read_reg() doesn't access the value if error occurred.
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.