noderemove - do not call puffs_setback here.  noderemove is called
from rename (for existing target) and calling setback is not
appropriate in that context.  Do that call instead directly in the
callers (remove, rmdir).

From Nicola Girardi, part of PR/54829.

(uwe)

puffs9p_node_rename - noderename takes puffs_node, not p9pnode, but
this goes undetected b/c pn_data we pass is a pointer to void.

From Nicola Girardi, part of PR/54829.

(uwe)

Avoid taking lwp_lock when calling process_sstep() for sibling LWPs

This makes the code consistend between single-threaded and multi-threaded
code.

(kamil)

Use the typedef to declare tag in AUTOVAR.  Do not allocate 9P tags in
advance in AUTOVAR, do it right before we need them.  Do no reuse tags.
If we are going to use sequential tags, we might as well try to make
them really sequential.

(uwe)

Do not allocate 9P tags in advance, do it right before we need them.
If we are going to use sequential tags, we might as well try to make
them really sequential.

(uwe)

puffs9p_node_lookup - use the right type for tfid.

(uwe)

Regen.

(ad)

Catch up after recent ACL changes

(ad)

uvm_page_array_fill(): return ENOENT in all cases when nothing's left.

(ad)

Avoid depending on namespace pollution in stdlib.h.

(joerg)

s/0/NULL/ for pointer (usb_async_proc)

(skrll)

Misc whitespace

(skrll)

proto_getstat - g/c debug printfs.

(uwe)

getdfwithoffset - do not issue dummy read to seek to offset zero on a
newly opened fid as it's already there.  This is redundant and also
seems to confuse inferno when issued against the root directory.  Now
you can ls the mount point of

    styxlisten -A 'tcp!*!styx' export /

While here fix the seek logic to actually repeatedly seek forward
instead of always doing it from zero.

(uwe)

Move sun2 and alpha to new binutils

(christos)

regen

(martin)

Add RTL8192EE Wireless LAN 802.11n PCI-E NIC

(martin)

Fix alpha build (relocation truncated to fit: GPREL16 against symbol ...)
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=47230

(christos)

Make vcache_tryvget() lockless.  Reviewed by hannken@.

(ad)

vmbus(4): Do not call hyperv_dma_alloc() in interrupt context.

The channel offer and rescind process is performed on another context.

(nonaka)

vmbus(4): Fixed incorrect use of vmbus_wait() in vmbus_channel_scan().

Found by kUBSan.

(nonaka)

it's "its"

(snj)

audio: Only restart recording mixer on resume if it's already been started

(nia)

put back NO_PCI_MSI_MSIX, need to fix MSI issues reported in PR port-xen/55285
first

(jdolecek)

Make the readahead-while-sending-output code conditional in
com_common_putc(), only erquest readahead before interrupts are enabled
and when called from comcnputc().
Fixes PR kern/55286.

(martin)

audio: remove comment that is no longer valid

(nia)

Add need-flags for kernfs.
Compile Xen kernfs support only if kernfs is compiled in the kernel.
Should fix MODULAR build.

(bouyer)

Always enable declarations needed by privcmd.c

(bouyer)

audio: Restore compat with Solaris for 8-bit LINEAR.

It's always signed, unless it's LINEAR8.

PR kern/55175

(nia)

Implement new ioctl, needed by Xen 4.13:
IOCTL_PRIVCMD_MMAPBATCH_V2
IOCTL_PRIVCMD_MMAP_RESOURCE
IOCTL_GNTDEV_MMAP_GRANT_REF
IOCTL_GNTDEV_ALLOC_GRANT_REF

(bouyer)

Ajust pmap_enter_ma() for upcoming new Xen privcmd ioctl:
pass flags to xpq_update_foreign()
Introduce a pmap MD flag: PMAP_MD_XEN_NOTR, which cause xpq_update_foreign()
to use the MMU_PT_UPDATE_NO_TRANSLATE flag.
make xpq_update_foreign() return the raw Xen error. This will cause
pmap_enter_ma() to return a negative error number in this case, but the
only user of this code path is privcmd.c and it can deal with it.

Add pmap_enter_gnt()m which maps a set of Xen grant entries at the
specified va in the specified pmap. Use the hooks implemented for EPT to
keep track of mapped grand entries in the pmap, and unmap them
when pmap_remove() is called. This requires pmap_remove() to be split
into a pmap_remove_locked(), to be called from pmap_remove_gnt().

(bouyer)

audio: Fix logic for resuming when the device is in use.

audio_[r/p]mixer_start should never be called when the device is
marked busy.

Resolves a panic on resume when audio is playing, PR kern/55301

(nia)

clang assembler evaluates #'\r' as #0x72. Grrr

(ryo)

Replace vp->v_count with vrefcnt(vp) and
remove now unneeded "#define v_count ...".

(hannken)

fixed that BTI trap will be occured when AP jumps to mp_vstart on ARMV85_BTI+SMP evironment.

(ryo)

Copyright maintenance

(skrll)

disassemblable bti insns

(ryo)

Catch up with the usage of struct vmspace::vm_refcnt

Use the dedicated reference counting routines.

Change the type of struct vmspace::vm_refcnt and struct vm_map::ref_count
to volatile.

Remove the unnecessary vm->vm_map.misc_lock locking in process_domem().

Reviewed by <ad>

(kamil)

new postfix

(christos)

merge conflicts

(christos)

new acpica

(christos)

merge conflicts

(christos)

uao_get(): in the PGO_SYNCIO case use uvm_page_array and simplify control
flow a little bit.

(ad)

Make previous work as intended.  Bad programmer.

(ad)

Minor correction to previous.

(ad)

- Alter the convention for uvm_page_array slightly, so the basic search
  parameters can't change part way through a search: move the "uobj" and
  "flags" arguments over to uvm_page_array_init() and store those with the
  array.

- With that, detect when it's not possible to find any more pages in the
  tree with the given search parameters, and avoid repeated tree lookups if
  the caller loops over uvm_page_array_fill_and_peek().

(ad)

Remove support for the !DVF_PRIV_ALLOC case; device_t and driver private
storage separation are now mandatory.

(thorpej)

new ntp

(christos)

new ntp

(christos)

Add FALLTHROUGH

(christos)

Merge conflicts

(christos)

PR kern/55300: ubciomove triggers page not dirty assertion

If overwriting an existing page, mark it dirty since there may be no
managed mapping to track the modification.

(ad)

Improve RGMII TX / RX delay handling, from OpenBSD.

(jmcneill)

Add support for BCM54213PE RGMII clock delays, from OpenBSD

(jmcneill)

Add MIIF_RXID and MIIF_TXID flags to signal that RX or TX delays are required, from OpenBSD.

(jmcneill)

uvm_pageout_done(): do nothing when npages is zero.

(ad)

ubc_uiomove_direct(): if UBC_FAULTBUSY, the left-over portion of the final
page needs to be zeroed.

(ad)

unifdef everything-but-NetBSD.  Gets rid of a naked cfattach decl.

(thorpej)

disable downgrade of ATA mode from DMA, as generally not relevant
any more - while it has been instrumental to inadvertedly discover
driver bugs in PIO mode under QEMU recently, generally the switch
more hurts than helps, so now only warn when DMA errors happen

code kept under ATA_DOWNGRADE_MODE ifdef, disabled by default

(jdolecek)

This driver is incomplete and doens't even compile, but make it
use CFATTACH_DECL_NEW() anyway.

(thorpej)

make ata_downgrade_mode() static, it's not used anywhere else

(jdolecek)

Switch to CFATTACH_DECL_NEW.

This driver doesn't compile -- never has since it was imported from
OpenBSD -- but in case anyone wants to make it work, let's leave fewer
landmines to step on.

(riastradh)

Ticket #1549

(martin)

Pull up following revision(s) (requested by christos in ticket #1549):

sys/netinet/igmp.c: revision 1.70
sys/kern/kern_time.c: revision 1.204

igmp_sendpkt() expects ip_output() to set 'imo.imo_multicast_ttl' into
'ip->ip_ttl'; but ip_output() won't if the target is not a multicast
address, meaning that the uninitialized 'ip->ip_ttl' byte gets sent to
the network. This leaks one byte of kernel heap.

Fix this by filling 'ip->ip_ttl' with a TTL of one.
Found by KMSAN.

-

Fix uninitialized memory access. Found by KMSAN.

(martin)

Tickets #920 - #922, #924 - #926, #929 - #932

(martin)

Pull up following revision(s) (requested by rmind in ticket #932):

usr.sbin/npf/npfctl/npf_scan.l: revision 1.31

PR/55288: npfctl: change parameter syntax to be more permissive.

(martin)

Pull up following revision(s) (requested by rmind in ticket #931):

usr.sbin/npf/npfctl/npf_show.c: revision 1.30

PR/54670: Azuma OKAMOTO: Consistently use 'W' for TH_CWN, and bump buffer
size.

(martin)

Pull up following revision(s) (requested by rmind in ticket #930):

usr.sbin/npf/npfctl/npf_build.c: revision 1.54
sys/net/npf/npf_conn.h: revision 1.19
usr.sbin/npf/npfctl/npfctl.h: revision 1.52
usr.sbin/npf/npfctl/npf_show.c: revision 1.31
sys/net/npf/npf_conf.c: revision 1.16
sys/net/npf/npf_nat.c: revision 1.49
sys/net/npf/npf_inet.c: revision 1.56
sys/net/npf/npf_conndb.c: revision 1.8
sys/net/npf/npf_conn.c: revision 1.31

Backport selected NPF fixes from the upstream (to be pulled up):

- npf_conndb_lookup: protect the connection lookup with pserialize(9),
  instead of incorrectly assuming that the handler always runs at IPL_SOFNET.
  Should fix crashes reported on high load (PR/55182).

- npf_config_destroy: handle partially initialized config; fixes crashes
  with some invalid configurations.

- NAT policy creation / destruction: set the initial reference and do not
  wait for reference draining on destruction; destroy the policy on the
  last reference drop instead.  Fixes a lockup with the dynamic NAT rules.

- npf_nat_{export,import}: fix a regression since dynamic NAT rules.

- npfctl: fix a regression and restore the default group behaviour.

- Add npf_cache_tcp() and validate the TCP data offset (from maxv@).

(martin)

Pull up following revision(s) (requested by rmind in ticket #929):

sys/kern/subr_thmap.c: revision 1.6

thmap(9): merge changes from the upstream -- primarily, switch to the
C11-style memory fences and atomic primitives; in NetBSD, this translates
to using the atomic_loadstore(9) primitives.

To be pulled up (just in case).

(martin)

Pull up following revision(s) (requested by kamil in ticket #926):

sys/arch/sparc/sparc/process_machdep.c: revision 1.20

Reject setting unaligned Program Counter

Patch taken from OpenBSD.

Tested in qemu.
Fixes PR port-sparc/54734 by Andreas Gustafsson

(martin)

Apply patch, requested by kamil in ticket #925:
Adaption of:

tests/lib/libc/sys/t_ptrace_fork_wait.h 1.3,1.4,1.6

Ignore interception of SIGCHLD signals in the debugger

There is a race that SIGCHLD might be blocked during forking and dropped.

PR/55241 by Andreas Gustafsson

(martin)

Apply patch, requested by kamil in ticket #925:

Cherry-pick from:

tests/lib/libc/sys/t_ptrace_fork_wait.h 1.2

Fix race in fork_singalmasked

ELF RTLD after rtld.c r. 1.204 introduced locking that wraps the fork
syscall. This locking changes signal mask of the calling process during
the forking process.

Instead of comparing old and new signal mask, just after the forking
operation, check whether the expected signal is still masked and in
another test whether it is still ignored.

(martin)

Pull up following revision(s) (requested by rin in ticket #924):

external/gpl3/gdb/dist/gdb/alpha-tdep.c: revision 1.5

Revive entrypoint heuristics, accidentally removed in the previous.
Now, e.g., "b main" works again for "gdb /bin/echo" as expected.

See more details:
http://cvsweb.netbsd.org/bsdweb.cgi/src/external/gpl3/gdb/dist/gdb/alpha-tdep.c#rev1.2
XXX

Single-stepping into shared libraries still fails sometimes;
tracee does not stop at a function in shared library.
We need further analysis...

(martin)

Pull up following revision(s) (requested by mlelstv in ticket #922):

sys/arch/x86/x86/cpu_topology.c: revision 1.18

assert smt_bits value only after it is computed.

(martin)

Fix static build

(christos)

Pull up following revision(s) (requested by jakllsch in ticket #921):

sys/dev/usb/xhci.c: revision 1.129

xhci(4): only usbd_xfer_trycomplete() if certain to usb_transfer_complete()

Previously we'd cancel the timeout after the data stage of control xfers, and
then hope that the status stage would eventually interrupt.

(martin)

Pull up following revision(s) (requested by riastradh in ticket #920):

lib/libc/compat/gen/compat_ldexp_ieee754.c: revision 1.8

Teach libc's compat ldexp stub to raise fp exceptions.

This ldexp stub will shadow the ldexp weak alias for scalbn in libm,
which is unfortunate but hard to fix properly without chasing the
mythical libc bump beast.  With the change here, we should raise all
the same exceptions that libm's scalbn does -- overflow, underflow,
inexact-result, and (for signalling NaN only) invalid-operation.

This in turn should correct the missing overflow/underflow exceptions
of our portable C fma, and perhaps other routines.

XXX pullup

(martin)

Pointless to force use of 64 bit atomics on flags fields that fit in 32 bits.

(christos)

Fix stats counters to be 32 bits on ILP32

(christos)

Fix static linking.

(christos)

new tzcode2020a

(christos)

Bring in 2020a

(christos)

Be more careful with transfer descriptors and crossing page boundaries

Enable USBMALLOC_MULTISEG

(skrll)

No need to rename netbsd-GENERIC.img to kernel7.img

(jmcneill)

Switch to GPT w/ hybrid MBR for armv7 images

(jmcneill)

disassemblable pointer authentication insns

(ryo)

Add __KERNEL_RCSID

(jmcneill)

Use howmany() macro.

(nonaka)

Stop all processing related to rx before that related to tx for safety

(yamaguchi)

Use evcnt(9) to record error status in vioif(4)

(yamaguchi)

Introduce the lock for vioif_softc to avoid a race condition
in vioif_update_link_status()

The function is called in both vioif_init() and softint.

(yamaguchi)

Populate mbufs in the packet receiving process, not in a softint

(yamaguchi)

Always hold tx lock in deferred transmit to send all packets

There may be packets that enqueued before another transmission
releases the lock after finish of its transmission.
When using mutex_try_enter(), vioif_deferred_transmit() can not
sends them.

pointed out by knakahara@n.o

(yamaguchi)

Remove redundant checks.
There is the same check in vioif_send_common_locked()

(yamaguchi)

Replace macros with static functions for refactoring

(yamaguchi)

Fix typo in comments

(yamaguchi)

Fix the wrong segment size in vioif(4)

(yamaguchi)

Introduce packet handling in softint or kthread for vioif(4)

(yamaguchi)

Set handlers implemented in child device of virtio(4) to virtqueue
instead of the commonized function

(yamaguchi)

Register an interrupt handler for virtqueue to that for pci
to reduce function call

(yamaguchi)

Remove VIRTIO_F_PCI_INTR_SOFTINT support for multiqueue

The functionality is no longer used by obsolating VIOIF_SOFTINT_INTR.

(yamaguchi)

Obsolete VIOIF_SOFTINT_INTR

The kernel option is introduced to realize softint-based if_input.
Since the same scheme has been implemented in if_percpuq_enqueue(),
the option is no longer needed.

pointed out by ozaki-r@n.o.

(yamaguchi)

add ARMv8.1-8.5 definitions of TCR_EL1

(ryo)

cache information can be detected correctly on newer CPUs

- add VPIPT cache type
- adapt to 64-bit CCSIDR (ARMv8.3-CCIDX)
- CCSIDR:[WT,WB,PA,WA] are deprecated
- show number of cache lines when attaching cpu

(ryo)

account for already transferred data (partially done I/O) when
retrying an xfer, to avoid reading/writing data from/to wrong offset,
and eventually beyond the end of data buffer

fixes data corruption under QEMU observed by Paul Ripke for emulated
IDE drives

(jdolecek)

Add GNU_ARCH for earmv5hf{,eb}.

(rin)

Add -g to usage.

(wiz)

Use Nx, fix formatting nit.

(wiz)

Fix typo in macro. Sort option descriptions.

(wiz)

tmpfs_write(): use UBC_FAULTBUSY when extending files, if possible, to avoid
zeroing of newly allocated pages & fault processing.

(ad)

- ubc_uiomove():  Always use direct access in the UBC_FAULTBUSY case, since
  it works basically the same way as !direct minus temporary mappings, and
  there are no concurrency issues.

- ubc_alloc_direct(): In the PGO_OVERWRITE case blocks are allocated
  beforehand.  Avoid waking or activating pages unless needed.

(ad)

Add libuv

(christos)

Adjust for bind-9.16.3

(christos)

new bind and libuv

(christos)

new bind+libuv

(christos)

Add uvm_pagewanted_p(): return true if someone is waiting on the page and
assert caller has correct lock to observe that.

(ad)

Merge conflicts

(christos)

Add libuv

(christos)

Add libuv

(christos)

Descend into libuv

(christos)

Add build glue

(christos)

Add build glue

(christos)

Enable GPT support w/ hybrid MBR for Raspberry Pi compatibility.

(jmcneill)

Add support for hybrid MBR/GPT images.

(jmcneill)

Add -g flag to preserve GPT headers when updating MBR.

(jmcneill)

Add -H flag to ignore existing MBR (Hybrid MBR/GPT mode).

(jmcneill)

uvideo(4): clamp (micro)frames-per-xfer to at most 80

Previously, on a 30fps YUV422 640x480 webcam, we were putting over 250
USB (micro)frames per video frame in the host controller queue.

xhci(4) is currently limited to 256-1 TRBs per xHC Transfer Ring, and as
such, trying to place 3 xfers each of 250+ microframes in the queue fails.

As there is no UVC requirement that whole video frames be in one logical
chunk of isoc transactions, and there doesn't seem to be compelling reason
to keep the xfer completion rate slower than 1 in 10ms, we can limit each
of the 3 uvideo xfers to 80 (micro)frames of bus time, and solve the
Transfer Ring constraint for upcoming xhci(4) Isochronous pipe support.
This works out to using only 240 TRBs on the 255-usable-TRB Transfer Ring.

(jakllsch)

Fix races in sdmmc tasks and teach ld@sdmmc to abort xfers on detach.

- Teach sdmmc_add_task to queue it only if not already queued.
- Remove now-redundant logic to avoid repeated queueing elsewhere.
- Teach sdmmc_del_task to wait until task has completed.
- Call sdmmc_del_task in various needful places.
- Replace abuse of pcq by a lock and a tailq.
  (pcq is multi-producer, _single_-consumer, but there are potentially
  multiple consumers here and really only one producer.)
- Teach ld_sdmmc to abort xfers on detach.
  (Mechanism is kinda kludgey but it'll do for now; any effort one is
  tempted to spend overhauling this should be spent overhauling sdmmc
  to support proper asynchronous commands.)
- Make sure ld_sdmmc_discard either returns failure or eventually calls
  ldenddiscard.

XXX Currently ld_sdmmc_detach aborts xfers _before_ ldbegindetach has
has committed to detaching or not.  This is currently necessary to
avoid a deadlock because ldbegindetach waits for xfers to drain --
which strikes me as wrong; ldbegindetach shouldn't wait for anything,
and should only make the decision to commit to detaching or not so
the caller can decide whether to abort xfers before we actually wait
for them in ldenddetach.

XXX pullup -- although this changes some kernel symbols (sdmmc_add_task
and sdmmc_del_task), it shouldn't affect any existing modules; the only
module that uses sdmmc is ld_sdmmc.kmod, which is `.if 0' in the build
so there shouldn't be any of them floating around.

(riastradh)

PR/55288: npfctl: change parameter syntax to be more permissive.

(rmind)

Set resize_gpt=NO by default

(jmcneill)

Add GPT support to mkimage.

(jmcneill)

Pass TOOL_GPT to mkimage

(jmcneill)

Support resizing wedges

(jmcneill)

Exit gracefully when auto-expanding a partition and it is already the
correct size. Add a -q flag to "resize" and "resizedisk" commands to skip
printing warnings in the already resize paths.

(jmcneill)

Use getfsspecname and getrawdiskname so we can resize by NAME= label.

(jmcneill)

dkwedge_add: Allow for expanding the size of an existing wedge without
having to delete it first, provided that no other parameters have changed.

(jmcneill)

fix KASAN PoolUseAfterFree for async write - can't read bp after VOP_STRATEGY()

problem found and fix provided by Paul Ripke

(jdolecek)

Change proc_lock -> &proc_lock.

(hannken)

hvn(4): Prevent from occur panic when acquiring the media status.

hvn(4) sleeps during hvn_media_status(), and thus requires an adaptive mutex
for the media lock.  Taken from recent iwm(4) change.

(nonaka)

Remove trailing whitespace, add missing word.

(wiz)

Misc whitespace

(skrll)

Gather the conditions to return from the VCPU loops in nvmm_return_needed(),
and use it in nvmm_do_vcpu_run() as well. This fixes two undesired behaviors:

- When a VM initializes, the many nested page faults that need processing
  could cause the calling thread to occupy the CPU too much if we're unlucky
  and are only getting repeated nested page faults thousands of times in a
  row.

- When the emulator calls nvmm_vcpu_run() and immediately sends a signal to
  stop the VCPU, it's better to check signals earlier and leave right away,
  rather than doing a round of VCPU run that could increase the time spent
  by the emulator waiting for the return.

(maxv)

Be more careful with OHCI_PAGE limitations in ohci_reset_std_chain and
ohci_device_isoc_enter

Enable USBMALLOC_MULTISEG

(skrll)

Note GENERIC and GENREIC64

(skrll)

Add mipsNN_cp0_rdhwr_cpunum() which returns the current CPU number
read from the CPUNum hardware register on MIPS{32,64}R2.

(simonb)

Add missing MACHINE_ARCH=earmv5hf{,eb} for evbarm.

(rin)

new binutils/gcc started emitting R_PPC_UADDR32 reloc entries, so for now
treat them like R_PPC_ADDR32
XXX
these relocs have been observed in the xf86-video-radeon driver, but only in
the AtomBios support code, which is unused on PowerPC. This allows the driver
to load but it's not clear if the affected parts would actually function.

(macallan)

Move proc_lock into the data segment.  It was dynamically allocated because
at the time we had mutex_obj_alloc() but not __cacheline_aligned.

(ad)

Oops.  If a SCHED_RR thread is preempted and has exceeded its timeslice it
needs to go to the back of the run queue so round-robin actually happens,
otherwise it should go to the front.

(ad)

PR kern/55237: Panic: vrelel: bad ref count (9.99.54)

Adjust v_writecount with v_interlock held.

(ad)

NetBSD 9.99.64 - struct lwp changed

(ad)

Oops.  If a SCHED_RR thread is preempted and has exceeded its timeslice it
needs to go to the back of the run queue so round-robin actually happens,
otherwise it should go to the front.

(ad)

sched_bestcpu(): There is a fallback CPU, in case the user manages to set
the system up so no CPU is permitted to run a given LWP.  Fix a bug where
that fallback CPU would sometimes get picked even if there was another CPU
that could legitimately run the LWP.

(ad)

- Replace pid_table_lock with a lockless lookup covered by pserialize, with
  the "writer" side being pid_table expansion.  The basic idea is that when
  doing an LWP lookup there is usually already a lock held (p->p_lock), or a
  spin mutex that needs to be taken (l->l_mutex), and either can be used to
  get the found LWP stable and confidently determine that all is correct.

- For user processes LSLARVAL implies the same thing as LSIDL ("not visible
  by ID"), and lookup by ID in proc0 doesn't really happen.  In-tree the new
  state should be understood by top(1), the tty subsystem and so on, and
  would attract the attention of 3rd party kernel grovellers in time, so
  remove it and just rely on LSIDL.

(ad)

uvm_pglistfree(): just use uvm_pagefree().

(ad)

Backport selected NPF fixes from the upstream (to be pulled up):

- npf_conndb_lookup: protect the connection lookup with pserialize(9),
  instead of incorrectly assuming that the handler always runs at IPL_SOFNET.
  Should fix crashes reported on high load (PR/55182).

- npf_config_destroy: handle partially initialized config; fixes crashes
  with some invalid configurations.

- NAT policy creation / destruction: set the initial reference and do not
  wait for reference draining on destruction; destroy the policy on the
  last reference drop instead.  Fixes a lockup with the dynamic NAT rules.

- npf_nat_{export,import}: fix a regression since dynamic NAT rules.

- npfctl: fix a regression and restore the default group behaviour.

- Add npf_cache_tcp() and validate the TCP data offset (from maxv@).

(rmind)