Do not require the display device to identify as VGA subclass.
If it's a display device and has AGP capability, we can trust it's a
proper AGP master. This fixes attachment for my Permedia 2 AGP board,
which identifies as subclass miscellaneous (0x80).
ddb(4)/crash(8): Print lock symbols in `show all locks'.
Reorder columns so the variable-width symbol comes last.
PR kern/60030: ddb/crash: show all locks without LOCKDEBUG
crash(8): Add support for `mach cpu N' on x86.
Requires core dump to have the registers saved on cpuN. This happens
with the ddb `sync' command, but not with the kern.dump_on_panic=1
ddb.onpanic=0 path yet -- TBD.
PR bin/58010: crash(8) doesn't support `mach cpu N' to examine
registers/stack of other CPUs
dwc2: Bring dwc2_wait_for_mode from upstream and use in dwc2_force_mode.
dwc2_wait_for_mode will wait upto 110milliseconds for the controller
to enter the expected mode.
This hopefully fixes
port-evbarm/60613: Recent dwc2 commit breaks usb based ethernet device...
Pull up following revision(s) (requested by rin in ticket #1327):
sys/dev/pci/pciconf.c: revision 1.57
sys/dev/pci/pciconf.c: revision 1.58
sys/dev/pci/pciconf.c: revision 1.59
pciconf: Fix alignments of device memory and I/O resources
Resources must be their-size-aligned. While here:
- Switch *align variable from int to more suitable integer types.
- Make some conditions clearer (NFC).
pciconf: Fix round-up logics for total sizes of bridge resources
The previous ones are valid only if alignments are 4K or 1M,
for I/O or memory, respectively.
pciconf: Sort resource windows in order of decreasing alignment
instead of size.
[8 lines not shown]
Pull up following revision(s) (requested by rin in ticket #418):
sys/dev/pci/pciconf.c: revision 1.57
sys/dev/pci/pciconf.c: revision 1.58
sys/dev/pci/pciconf.c: revision 1.59
pciconf: Fix alignments of device memory and I/O resources
Resources must be their-size-aligned. While here:
- Switch *align variable from int to more suitable integer types.
- Make some conditions clearer (NFC).
pciconf: Fix round-up logics for total sizes of bridge resources
The previous ones are valid only if alignments are 4K or 1M,
for I/O or memory, respectively.
pciconf: Sort resource windows in order of decreasing alignment
instead of size.
[8 lines not shown]
Pull up following revision(s) (requested by skrll in ticket #417):
sys/external/bsd/dwc2/dwc2.c: revision 1.84
PR/60021: USB-only boot: uhub0 attaches but uhub1 never appears...
Update the dwc2 intial reset sequence to more closely match the upstream
driver - it's changed a bit in the last 10 years.
This reset sequence gets the device out of the state it's left in by the
RaspberryPI firmware when booting from USB.
Pull up following revision(s) (requested by gutteridge in ticket #1326):
external/gpl2/grep/dist/src/grep.c: revision 1.6
external/gpl2/grep/dist/src/grep.c: revision 1.7
grep.c: fix ctype(3) issue leading to segfault
Addresses PR bin/60552 from Chavdar Ivanov, fix supplied by RVP.
grep.c: match upstream coding style for previous change (NFC)
Pull up following revision(s) (requested by gutteridge in ticket #416):
external/gpl2/grep/dist/src/grep.c: revision 1.6
external/gpl2/grep/dist/src/grep.c: revision 1.7
grep.c: fix ctype(3) issue leading to segfault
Addresses PR bin/60552 from Chavdar Ivanov, fix supplied by RVP.
grep.c: match upstream coding style for previous change (NFC)
Pull up following revision(s) (requested by riastradh in ticket #2045):
sys/kern/vfs_subr.c: revision 1.504
vflushbuf: Print `vflushbuf: dirty' warning only once per call.
And do it only under DEBUG, and rate-limit it globally.
This can happen when vflushbuf is competing with an onslaught of
concurrent I/O on a snapshot or block device (which is done without
holding the vnode lock, despite what the comment says). That
concurrent I/O might be happening indefinitely, so we are likely to
have to print a _lot_ of warnings, potentially overwhelming the
console, until the underlying problem is fixed by teaching vflushbuf
to wait only for the writes that began before have completed, not
also all writes that have begun since.
This stop-gap measure intended to be low-risk for pullup to release
branches.
[2 lines not shown]
Pull up following revision(s) (requested by riastradh in ticket #1325):
sys/kern/vfs_subr.c: revision 1.504
vflushbuf: Print `vflushbuf: dirty' warning only once per call.
And do it only under DEBUG, and rate-limit it globally.
This can happen when vflushbuf is competing with an onslaught of
concurrent I/O on a snapshot or block device (which is done without
holding the vnode lock, despite what the comment says). That
concurrent I/O might be happening indefinitely, so we are likely to
have to print a _lot_ of warnings, potentially overwhelming the
console, until the underlying problem is fixed by teaching vflushbuf
to wait only for the writes that began before have completed, not
also all writes that have begun since.
This stop-gap measure intended to be low-risk for pullup to release
branches.
[2 lines not shown]
Pull up following revision(s) (requested by riastradh in ticket #415):
sys/kern/vfs_subr.c: revision 1.504
vflushbuf: Print `vflushbuf: dirty' warning only once per call.
And do it only under DEBUG, and rate-limit it globally.
This can happen when vflushbuf is competing with an onslaught of
concurrent I/O on a snapshot or block device (which is done without
holding the vnode lock, despite what the comment says). That
concurrent I/O might be happening indefinitely, so we are likely to
have to print a _lot_ of warnings, potentially overwhelming the
console, until the underlying problem is fixed by teaching vflushbuf
to wait only for the writes that began before have completed, not
also all writes that have begun since.
This stop-gap measure intended to be low-risk for pullup to release
branches.
[2 lines not shown]
Pull up following revision(s) (requested by msaitoh in ticket #1324):
usr.sbin/syslogd/syslogd.c: revision 1.156
usr.sbin/syslogd/syslogd.c: revision 1.157
syslogd: Fix buffering of partial klog lines
When a /dev/klog read ends without a newline, the remaining data
is buffered for the next read. The code mistakenly copied the data
into linebuf instead of klog_linebuf, breaking reconstruction of
split log messages.
Found and fixed by nonaka@ and Kenichi Suzuki at IIJ.
syslogd: Preserve full partial klog line when buffering
Buffer incomplete /dev/klog lines from the start of the original
line instead of the parse position. Otherwise, split messages
[3 lines not shown]
Pull up following revision(s) (requested by msaitoh in ticket #414):
usr.sbin/syslogd/syslogd.c: revision 1.156
usr.sbin/syslogd/syslogd.c: revision 1.157
syslogd: Fix buffering of partial klog lines
When a /dev/klog read ends without a newline, the remaining data
is buffered for the next read. The code mistakenly copied the data
into linebuf instead of klog_linebuf, breaking reconstruction of
split log messages.
Found and fixed by nonaka@ and Kenichi Suzuki at IIJ.
syslogd: Preserve full partial klog line when buffering
Buffer incomplete /dev/klog lines from the start of the original
line instead of the parse position. Otherwise, split messages
[3 lines not shown]
Pull up following revision(s) (requested by he in ticket #1323):
sys/dev/pci/mpii.c: revision 1.32
sys/dev/pci/mpii.c: Reduce MPII_REQUEST_CREDIT to improve stability.
This should fix PR#58684, at least it does in my case on a
host with a "Symbios Logic SAS3008 (rev. 0x02)" controller.
Admittedly, this is a band-aid fix, but it's the best we could come
up with to solve the PR, and it does not appear to have significant
performance implications, at least not in a single-drive configuration.
So, better to be stable / functional than not.
Pull up following revision(s) (requested by he in ticket #412):
sys/dev/pci/mpii.c: revision 1.32
sys/dev/pci/mpii.c: Reduce MPII_REQUEST_CREDIT to improve stability.
This should fix PR#58684, at least it does in my case on a
host with a "Symbios Logic SAS3008 (rev. 0x02)" controller.
Admittedly, this is a band-aid fix, but it's the best we could come
up with to solve the PR, and it does not appear to have significant
performance implications, at least not in a single-drive configuration.
So, better to be stable / functional than not.
Pull up following revision(s) (requested by riastradh in ticket #411):
external/cddl/osnet/lib/libdtrace/Makefile: revision 1.35
libdtrace/Makefile: Add to, don't overwrite, CLEANFILES.
PR lib/60530: NetBSD 11.0 cleandir leaves
external/cddl/osnet/lib/libdtrace/obj/dtrace.expsym around
uvm_pdaemon.c: Undo mistake in previous uvm_lwp_is_pagedaemon change.
This was not a test for whether curlwp is the page daemon; this is a
test for whether the page daemon exists at all, but it was changed
nevertheless to use the new uvm_lwp_is_pagedaemon() in rev. 1.139.