update-sets:
- sort entries
- add NetBSD RCSID
- add "DO NOT EDIT" comment

(rin)

Satisfy -fno-common

(jmcneill)

Avoid duplicate "code_lines" symbols, which fails to link when building
tools with gcc 10.1.1 / ld 2.34-2.fc32.

(jmcneill)

This driver sleeps during iwm_media_change(), and thus requires an
adaptive mutex for the media lock.

(thorpej)

Remove the ubc_direct hack.

(ad)

fix to do backtrace properly for running LWPs and cpu_lwp_fork().
when dump of pcb_tf, only the switchframe part is now displayed instead of the whole trapframe.

(ryo)

uao_get(): handle PGO_OVERWRITE.

(ad)

Set frame = NULL in XENPV whether DIAGNOSTIC or not.

For Xen PV, the interrupt logic stores the relevant information in
the CPU's struct cpu_info, and CLKF_USERMODE and CLKF_PC don't use
the frame argument at all.  (For Xen PVH, we behave like normal x86
and use the frame as passed.)

DIAGNOSTIC shouldn't change the data flows that affect non-DIAGNOSTIC
logic, so let's just unconditionally set frame = NULL for XENPV.  If
this makes a difference, that's a bug -- and it's a bug that should
not be concealed by enabling DIAGNOSTIC.

ok bouyer

(riastradh)

remove NO_PCI_MSI_MSIX option, MSI is supported for XenPV Dom0

(jdolecek)

Sort

(skrll)

fix reversed test.

(christos)

Only build GENERIC kernel for armv7 releases.

(jmcneill)

DRY code in uvm_swap_io() for the write loop

(jdolecek)

Support get_framerate and provide a stub for set_framerate

(jmcneill)

Support set/get_framerate

(jmcneill)

Support VIDIOC_G_PARM/VIDIOC_S_PARM

(jmcneill)

instead of reading memory directly, db_read_bytes() is used to avoid faults in ddb.

(ryo)

Mark usage as dead

(joerg)

Don't return from dead functions.

(joerg)

- remove i2cbus etc. on drivers that don't use iic
- for private iic buses pull in only what we need and skip the bus attachment

(macallan)

prune unneeded includes

(macallan)

Fix merge error

(ad)

- Recalibrate the APIC timer using the TSC, once the TSC has in turn been
  recalibrated using the HPET.  This gets the clock interrupt firing more
  closely to HZ.

- Undo change with recent Xen merge and go back to starting the clocks in
  initclocks() on the boot CPU, and in cpu_hatch() on secondary CPUs.

- On reflection don't use HPET delay any more, it works very well but means
  going over the bus.  It's enough to use HPET to calibrate the TSC and
  APIC.

Tested on amd64 native, xen and xen PVH.

(ad)

Fix bugs in previous.

(ad)

Do the chmod last so that the kernels get consistent permissions for MKREPRO
Reported by wiz@.

(christos)

Let's not waste time decrypting garbage, shall we?

Skip to the end if the transfer failed.

(riastradh)

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.

(jakllsch)

xhci(4): dynamically allocate xhci_xfer->xx_trb buffer

Allocation size remains the same for now.

(jakllsch)

xhci(4): Check for overflows with KASSERT when putting TRBs on xfer TRB buffer.

Also more-correctly check for overruns when putting TRBs on a ring.

(jakllsch)

xhci(4): Be more concise in putting the setup request in TRB immediate data.

Or: memcpy(),htole64() is replaced by functionally-identical le64dec().

(jakllsch)

Ticket #919

(martin)

Pull up following revision(s) (requested by maxv in ticket #919):

sys/dev/nvmm/x86/nvmm_x86.c: revision 1.9
sys/dev/nvmm/x86/nvmm_x86_svm.c: revision 1.60
sys/dev/nvmm/x86/nvmm_x86_svm.c: revision 1.61
sys/dev/nvmm/x86/nvmm_x86_vmx.c: revision 1.56
sys/dev/nvmm/x86/nvmm_x86_vmx.c: revision 1.57
sys/dev/nvmm/x86/nvmm_x86_vmx.c: revision 1.58
sys/dev/nvmm/nvmm.c: revision 1.29

Improve the CPUID emulation of basic leaves:
- Hide DCA and PQM, they cannot be used in guests.
- On Intel, explicitly handle each basic leaf until 0x16.
- On AMD, explicitly handle each basic leaf until 0x0D.

Respect the convention for the hypervisor information: return the highest
hypervisor leaf in 0x40000000.EAX.

Improve the CPUID emulation on nvmm-intel: limit the highest basic and
hypervisor leaves.

Complete rev1.26: reset nvmm_impl to NULL in nvmm_fini().

(martin)

update Xen ldscript to add kASan .ctors section, so that Xen kernel with
enabled KASAN option at least compiles; no promises on it actually working
though

(jdolecek)

in wdctimeout(), do not schedule another timeout handler, as that only
creates race with eventual xfer resubmission - the c_intr()
method does all the necessary handling, including re-scheduling of the
timeout handler if appropriate

also make sure to clear ATACH_IRQ_WAIT before calling c_intr(), same
as real interrupt does; this is important so that eventual spurious timeout
would not interfere with xfer handling in the atabus thread

fixes another race in the atabus thread found by KASAN, reported by Paul Ripke

(jdolecek)

stop timeout handler while scheduling another part of partial I/O,
to avoid race between the timeout and I/O submission; the I/O
submission can sleep with xfer while waiting for the controller to
be ready once it gets to thread context, and timeout might cause
the xfer to be freed, leading to crashes due to use-after-free

this fixes another type of crashes with slow devices under QEMU reported
by Paul Ripke - thanks a lot with extensive debugging help

(jdolecek)

The Hacienda must be rebuilt.
https://twitter.com/McrHistory/status/1263386700249776130
https://twitter.com/Mr_Dave_Haslam/status/1263373996172640256

(sevan)

Mmh, should check cpuid_level first.

(maxv)

Complete rev1.26: reset nvmm_impl to NULL in nvmm_fini().

(maxv)

Improve the CPUID emulation on nvmm-intel: limit the highest basic and
hypervisor leaves.

(maxv)

Increase the number of ports to 8.

(maxv)

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

(riastradh)

Fix a problem that setcontext(2) sometimes fail on ARMv8.5-BTI cpu.

fixed to always get 0 for SPSR.BTYPE with getcontext(2).
a non-zero SPSR.BTYPE cannot be set with setcontext(2).

(ryo)

fix typo

(ryo)

In sleepq_insert(), in the SOBJ_SLEEPQ_SORTED case, if there are existing
waiters of lower priority, then the new LWP will be inserted in FIFO order
with respect to other LWPs of the same priority.  However, if all other
LWPs are of equal priority to the LWP being inserted, the new LWP would
be inserted in LIFO order.

Fix this to always insert in FIFO order with respect to equal priority LWPs.

OK ad@.

(thorpej)

sync with reality

(maxv)

future-proof-ness

(maxv)

zfs_netbsd_getpages:

- implement the PGO_LOCKED case
- handle npages > 1 for PGO_SYNCIO

(ad)

The boot CPU suffers a cache miss during TSC sync, before RDTSC.  Make the
secondary CPU take a miss as well to try and delay it an equal amount.

(ad)

use device_private() instead of casting self as it doesn't work here.
Bump rcs tag which was missed in r1.9

(sevan)

this is kmsan

(maxv)

Deal with KMSAN fussiness.  Pointed out by msaitoh@.

(ad)

uvm_loanuobjpages():

- there are no pages to unbusy in the error case
- always clear the caller's page array

(ad)

Make swap encryption MP-safe.

Not entirely sure the rest of the swap system is MP-safe, but let's
not make it worse!

XXX Why is swap_syscall_lock an rwlock?  We don't seem to take the
reader lock ever.

(riastradh)

Use more-semantically-correct types in xhci allocx/freex functions.

(jakllsch)

Fix EPERM vs EACCES on chtimes (thanks @hannken)

(christos)

fix accessx confusion (thanks hannken@)

(christos)

Switch sparc64 to binutils 2.34

(martin)

remove accmode_t typedef (not needed, breaks llvm) from maxv@

(christos)

Suppress GCC warnings and fix a UVMHIST_LOG() statement.

Kernels ALL/amd64 and ALL/i386 and port sparc64 build again.

(hannken)

i.MX support merged into GENERIC kernel.

(hkenken)

radeonfb doesn't attach an iic*, so pull in iic, not i2cbus

(macallan)

Modified debug message.

(hkenken)

Use kernel API of PWM subsystems for i.MX PWM driver.

(hkenken)

Temporary back to lapic_initclocks() from lapci_reset() to avoid compile
error.

(msaitoh)

PR kern/32166: pgo_get protocol is ambiguous
Also problems with tmpfs+nfs noted by hannken@.

Don't pass PGO_ALLPAGES to pgo_get, and ignore PGO_DONTCARE in the
!PGO_LOCKED case.  In uao_get() have uvm_pagealloc() take care of page
zeroing and release busy pages on error.

(ad)

Remove duplicate def of tsc_get_timecount().

(ad)

Ignore x86_delay, for xen

(ad)

xen needs the TSC funcs too

(ad)

uvm_loanuobjpages():

- vmobjlock is shared between tmpfs vnodes and UAOs now
- split into two routines, to simplify
- fix error recovery

(ad)

Don't try to do readahead on tmpfs.

(ad)

Drop & re-acquire vmobjlock less often.

(ad)

If the the TSC timecounter is good then use the TSC for DELAY() too.

(ad)

Make cpu_counter(), cpu_counter32() and tsc_get_timecount() into a single
preemption-safe routine.

(ad)

lapic_delay() disable preemption since the state is very CPU dependent.

(ad)

Comment

(ad)

UVM_PAGE_TRKOWN: print the LID too

(ad)

Use fixed date in man page. For MKREPRO build.

(wiz)

put back extattr.3

(christos)

ohci(4): always use usb_syncmem() when reading from HCCA

(jakllsch)

ohci_device_isoc_done(): use ux_length rather than ux_bufsize

For consistency.  Nick may have missed this in previous because I'd
unexpectedly adjusted it.

(jakllsch)

Use usbd_destroy_xfer() in usbd_request_async() instead of usbd_free_xfer().

The only user of usbd_request_async() & co. does a usbd_create_xfer(), so use
the correct deconstructor.

(jakllsch)

Remove definition of unused struct xhci_endpoint.

(jakllsch)

Tickets #917 and #918

(martin)

Apply patch, requested by simonb in ticket #918:

the updated octeon_rnm(4) driver fails a bogus randomness
test in the netbsd-9 rnd(9) driver for RND_TYPE_RNG type devices so
uses the RND_TYPE_UNKNOWN type.  This approach is used by other drivers
on the netbsd-9 branch.

(martin)

Pull up following revision(s) (requested by simonb in ticket #918):

sys/arch/mips/cavium/dev/octeon_rnm.c: revision 1.3
sys/arch/mips/cavium/dev/octeon_rnm.c: revision 1.4
sys/arch/mips/cavium/dev/octeon_rnm.c: revision 1.5
sys/arch/mips/cavium/dev/octeon_rnm.c: revision 1.6 (+ patch)
sys/arch/mips/cavium/dev/octeon_rnmreg.h: revision 1.2
sys/arch/mips/cavium/dev/octeon_rnmreg.h: revision 1.3
sys/arch/mips/cavium/octeonvar.h: revision 1.7

Add a few more bits.
XXX convert to __BITS.
--
If bus_space_map fails, just don't attach the driver instead of panicing.
Check RNG built in self test, don't attach if that fails too.
--
Oceton RNG/RNM driver modernisation to fit new entropy world order by
riastradh@, with some tweaks to get working in RNG mode.
XXX TODO: work out how to get raw entropy mode working.
--
Rework octeon_rnm(4) random number generator driver.
- Do a little on-line self-test for fun.
- Draw raw samples from the ring oscillators.
- Draw substantially more samples:
  =3D> early RO samples seem to have considerably lower entropy
  =3D> consecutive RO samples are not independent
- Make sure to use rnd_add_data_sync in the callback.
  =3D> not technically needed in HEAD, but would be needed for pullup
--
Adjust entropy estimate for the Octeon.
We are hedging in serial and in parallel, and more conservative than
the Linux driver from Cavium seems to be, so although I don't know
exactly what the thermal jitter of the device is, this seems like a
reasonable compromise.

(martin)

Pull up following revision(s) (requested by hannken in ticket #917):

sys/kern/vnode_if.src: revision 1.80
sys/kern/vnode_if.c: revision 1.112

VOP_STRATEGY() may still deadlock with devices.

Change FSTRANS from LAZY to NO.

Regen.

(martin)

postfix-3.5.2 out.

(wiz)

Add RISC-V's birthday
https://riscv.org/2020/05/happy-10th-birthday-risc-v/

(sevan)

Fix outdated or misleading comments for MEMORY_DISK_ROOT_SIZE.
No functional changes.

(rin)

Fix comment; MEMORY_DISK_SIZE --> MEMORY_DISK_ROOT_SIZE

(rin)

Oops, fix comment in the previous for accuracy.

(rin)

Bump ramdisk size to 2MB.

(rin)

Fix dtblist; our echo(1) does not interpret \t.

(rin)

Round amiga_clk_interval to the nearest integer, in order to emulate
100 Hz clock better by that running at PAL frequency (709379 Hz).

No functional changes for systems running at NTSC freq (715909 Hz).

(rin)

in wdctimeout() ignore timeout if not actually waiting for IRQ, to avoid
processing xfer which is not quite setup

in wdcintr() actually write the error message on invocation when
not waiting for IRQ even without ATADEBUG option, as usually that
is a driver bug; might revisit this if this ends up too noisy for
PCI-IDE which seems to ignore WDCTL_IDS and hence trigger irq even
for polled commands

(jdolecek)

only start the timeout machinery once the I/O is completely setup
and successful, particularly after PIO write is finished

fixes crashes in case the setup is so slow that timeout is triggered
e.g. while still waiting in wdc_wait_for_unbusy() or shortly after, without
drive actually having chance to complete the I/O, as seen in some
configuration under QEMU by Paul Ripke

(jdolecek)

Try to fix fallout from addition of dtb.tgz; MKDTB was added to both
_MKVARS.yes and _MKVARS.no. Remove it from the former.

(rin)

Add tadpmu to Makefile and set list.

(jdc)

Note rockchip RNG.

(riastradh)

Typo fixes from FreeBSD via Ed Maste

(christos)

fix build on netbsd-9 hosts.

(mrg)

Separate devicetree .dts -> .dtb building from kernel builds. They are now
part of a separate set, "dtb.tgz", and only built when MKDTB=yes. This
defaults to yes for earmv[67]* and aarch64, and no everywhere else.

(jmcneill)

Add Margaret Hamilton, Florian Schneider, Ron Hardy, Little Richard

(sevan)

BMP, DGEMM, DIB, GEMM, WMF

(sevan)

Heads up on Bell patents
P21 on https://minnie.tuhs.org/Archive/Documentation/AUUGN/AUUGN-V01.1.pdf

(sevan)

remove debugging, it is just clutter.

(christos)

Fix EPERM vs EACCES return.

(christos)

Tickets #908 - #916

(martin)

Pull up following revision(s) (requested by maxv in ticket #916):

sys/kern/kern_time.c: revision 1.204

Fix uninitialized memory access. Found by KMSAN.

(martin)

Pull up following revision(s) (requested by maxv in ticket #915):

sys/netinet/igmp.c: revision 1.70

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.

(martin)

Pull up following revision(s) (requested by nia in ticket #914):

sys/kern/subr_cprng.c: revision 1.37 (via patch, adapted)

Make kern.arandom truncate the output instead of failing with ETOOBIG
when the requested data exceeds 256 bytes in size. The actual size of
the returned data is output to oldlenp.

This matches FreeBSD's behaviour and seems to be more in line with
what software in the wild expects.

"sounds reasonble" - Riastradh

(martin)

Pull up following revision(s) (requested by riastradh in ticket #913):

sys/arch/arm/dts/rk3399-crypto.dtsi: revision 1.1
sys/arch/arm/rockchip/rk_v1crypto.c: revision 1.1
sys/arch/arm/rockchip/rk_v1crypto.c: revision 1.2 (plus patch)
sys/arch/arm/rockchip/rk_v1crypto.h: revision 1.1
sys/arch/arm/dts/rk3399-pinebook-pro.dts: revision 1.3
sys/arch/arm/rockchip/rk3399_cru.c: revision 1.20
sys/arch/evbarm/conf/GENERIC64: revision 1.158
sys/arch/arm/dts/rk3399-rockpro64.dts: revision 1.11
sys/arch/arm/rockchip/files.rockchip: revision 1.24

Rockchip crypto engine RNG driver.

As found on the rk3288 and rk3399.  This driver only supports the
TRNG, not the rest of the crypto engine, although it uses the AES unit
to do a self-test at attach time to verify that the engine works.
There seem to be two versions of the Rockchip crypto engine, v1 and
v2; this one is for v1.  Can't name a driver `rkcryptov1' so we'll
clumsily call it `rkv1crypto' instead to leave room for `rkv2crypto'
later on.

The crypto binding derived from the Rockchip BSP Linux kernel, in the
location it appears on the rk3399, is in rk3399-crypto.dtsi, since
there doesn't seem to be a better place to put it at the moment among
this twisty maze of inclusions, all different.

Use rnd_add_data_sync from the callback.

(Doesn't make a difference in HEAD but this is the stated API
contract and it matters if we want to pull this up.)

Prime the pool on attach.

(martin)

Pull up following revision(s) (requested by riastradh in ticket #912):

sys/arch/arm/dts/sun50i-a64.dtsi: revision 1.14
sys/arch/evbarm/conf/GENERIC64: revision 1.116
sys/arch/evbarm/conf/GENERIC64: revision 1.131
sys/arch/arm/sunxi/sun50i_a64_ccu.c: revision 1.21
sys/arch/arm/sunxi/sun50i_a64_ccu.c: revision 1.22
sys/arch/arm/sunxi/sun8i_crypto.c: revision 1.10
sys/arch/arm/sunxi/sun8i_crypto.c: revision 1.11
sys/arch/arm/sunxi/sun8i_crypto.c: revision 1.12
sys/arch/arm/sunxi/sun8i_crypto.c: revision 1.1
sys/arch/arm/sunxi/sun8i_crypto.c: revision 1.2
sys/arch/arm/sunxi/sun8i_crypto.c: revision 1.3
sys/arch/arm/sunxi/sun8i_crypto.h: revision 1.1
sys/arch/arm/sunxi/sun8i_crypto.c: revision 1.4
sys/arch/arm/sunxi/sun8i_crypto.c: revision 1.5
sys/arch/arm/sunxi/sun8i_crypto.c: revision 1.6
sys/arch/arm/sunxi/sun8i_crypto.c: revision 1.7
sys/arch/arm/sunxi/sun8i_crypto.c: revision 1.8
sys/arch/arm/sunxi/sun8i_crypto.c: revision 1.9
sys/arch/arm/sunxi/files.sunxi: revision 1.67

Add crypto engine clock

Add SUNXI_CCU_NM_ROUND_DOWN to CE clock, fix pll parents to use 2X outputs

Add crypto engine block.

Draft driver for Allwinner Crypto Engine.
Found on, e.g., the Pinebook.

Only used for TRNG at the moment, but hooking it up to opencrypto(9)
shouldn't be too hard if anyone still cares about that these days.

The distribution of the alleged TRNG is very nonuniform distributed
seems to alternate between toward runs with exceptionally high
fractions of 0 bits and runs with exceptionally high fractions of 1
bits -- initially all my samples were mostly 0's, and then all my
samples were mostly 1's, and now I'm seeing more oscillation between
these runs.

So I've wired it up as RND_TYPE_UNKNOWN, not RND_TYPE_RNG (it will
immediately flunk our rngtest and be disabled), and I estimated it to
provide at most one bit of entropy per byte of data -- which may
still be optimistic.  I also added a sysctl node hw.sun8icryptoN.rng
to read out 1024-byte samples for analysis, and I left the driver
commented out in GENERIC64 for now.
(If anyone has contacts at Allwinner who can tell us about how the
alleged TRNG is supposed to work, please let me know!)

Reduce some duplicated bus_dma clutter.

Factor out some of the self-test logic used for debugging.

Add missing bus_dmamap_sync(POSTWRITE) while here.

Make sure ERESTART doesn't come flying out to userland.
I picked ERESTART to mean `all channels are occupied' because that's
what opencrypto(9) uses to decide whether to queue a request, but
it's not appropriate for sysctl(2) to return that.

Avoid a race between interruption and reacquisition of lock.
Otherwise, we would have leaked the memory in this case.

Tidy up comments.

Oops -- forgot to kmem_free.

Fix typo -- acknowledge interrupts _and_ errors.

Reduce entropy estimate for sun8icrypto TRNG.

NIST's SP800-90B entropy estimation tools put it at no more than .08
bits of entropy per byte of data(!), so estimate 100 bits of data per
bit of entropy.  This is probably not conservative enough -- the NIST
tools were written without knowledge of how this alleged TRNG works!
Knowledge of the physics of how the TRNG is supposed to work could
probably enable a better job at predicting the outputs.
While here, bump the size of data we can sample directly with sysctl
to 4096 bytes.

Enable sun8icrypto in GENERIC64.

But set its entropy estimate to zero until we have a better idea of
how it works.  Can't really hurt this way unless sun8icrypto is
maliciously dependent on all other inputs to the entropy pool, which
seems unlikely.

Fix (presently harmless) psato.

Don't overwrite cy_root_node; use cy_trng_node as intended.

Fix previous brainfart.

Don't use the uninitialized trng node as the root node -- derp.
Instead, use the root node as the root node, and initialize the trng
node here.

(martin)

Pull up following revision(s) (requested by isaki in ticket #911):

sys/dev/audio/audiodef.h: revision 1.13
sys/dev/audio/audiodef.h: revision 1.14
sys/arch/sparc/include/param.h: revision 1.74
sys/arch/evbarm/conf/GENERIC64: revision 1.148
sys/arch/sh3/include/param.h: revision 1.25
sys/arch/evbarm/conf/GENERIC: revision 1.75
sys/arch/m68k/include/param.h: revision 1.23
sys/arch/amd64/conf/GENERIC: revision 1.564
sys/arch/hppa/include/param.h: revision 1.27
sys/dev/audio/audio.c: revision 1.68
sys/arch/vax/include/param.h: revision 1.63
sys/dev/audio/audio.c: revision 1.69

Reduce default AUDIO_BLK_MS from 40msec to 10msec on all platform except m68k
(m68k uses 40msec default as before).  And remove the option from GENERIC.

- It's not good idea to set such parameter in individual GENERICs.
- 4msec is (probably no problem for most modern real hardware but)
  too aggressive to be default.
- 10msec is too severe for antique machines but it's hard to draw a line.

Set AUDIO_BLK_MS 40 msec on other old(slow) architectures not only m68k.

Thanks tsutsui@ for comment about architecture choice.
And move it from audiodef.h to audio.c as suggested by joerg@.

Move machine dependent AUDIO_BLK_MS default value to <machine/param.h>.
If the port has __AUDIO_BLK_MS in <machine/param.h>, it will be used.

Otherwise the default value (currently 10 msec) defined in audio.c will
be used.  This mechanism is for very old ports which cannot satisfactorily
handle 10 msec block.  Currently hppa, m68k, sh3, sparc(!64) and vax are.

For port maintainers, if general models in your port cannot satisfactorily
handle 10 msec block, please consider to define your suitable longer period
(40 msec would be a good first choice).

But please don't be eager to make the default value shorter.
<machine/param.h> was discussed in source-changes-d.  It's better than
ifdef storm, or adding 60+ new header files in every arch/*/include/
directories for this.  Thanks mrg@, ad@, and everyone.
http://mail-index.netbsd.org/source-changes-d/2020/05/01/msg012572.html

(martin)

Pull up following revision(s) (requested by isaki in ticket #910):

sys/dev/audio/audio.c: revision 1.49
sys/dev/audio/audio.c: revision 1.61
sys/dev/audio/audio.c: revision 1.66
sys/dev/audio/audio.c: revision 1.67

Move #if DIAGNOSTIC..#endif to correct place.
It should ignore stray interrupts regardless of DIAGNOSTIC.

Improve an error message about round_blocksize.

Improve diagnostic messages.

Improve the blocksize notation.
The blocksize is expressed in bytes, and the millisecond notation
is supplementary information to make it easier to understand.

(martin)

Pull up following revision(s) (requested by isaki in ticket #909):

sys/dev/audio/audio.c: revision 1.65

Fix to start recording immediately when open() with READ mode is
called (unless pause).
- Opening /dev/audio always initializes pause with false.  Therefore
  it always starts recording.
- Opening /dev/sound inherites pause from the sticky parameter.
  Therefore whether /dev/sound starts recording or not depends on the
  sticky pause parameter.

This fixes two problems:
- Opening /dev/audio didn't start recording after merging isaki-audio2.
- Opening /dev/sound didn't start recording regardless of the sticky
  pause, probably since long time ago (at least netbsd-7).

(martin)

Pull up following revision(s) (requested by tsutsui in ticket #908):

sys/dev/ic/stivar.h: revision 1.10
sys/dev/ic/sti.c: revision 1.21
sys/arch/hp300/dev/sti_sgc.c: revision 1.3

Add bitmap access ops support for EVRX framebuffer on HP9000/425e.

8bpp Xorg wsfb server and mlterm-wscons (formerly mlterm-fb) work.
No particular comment on port-hp300@ and port-hppa@:
https://mail-index.netbsd.org/port-hp300/2020/05/02/msg000170.html

Special thanks to Miod Vallat, for his advice about HP-UX implementation
and binutils patches to disassemble old HP-UX a.out-hp300hpux binaries
(and also contributing his 425e back in 2014).

(martin)

Whitespace fixes and a serial comma.

(wiz)

Adjust entropy estimate for the Octeon.

We are hedging in serial and in parallel, and more conservative than
the Linux driver from Cavium seems to be, so although I don't know
exactly what the thermal jitter of the device is, this seems like a
reasonable compromise.

(riastradh)

Remove misleading comment.

I've successfully transferred 16-byte command. I've checked scsipi source and
see no reason for this limitation.

(rkujawa)

Add a basic manual page for tadpmu.

(jdc)

Regen.

(hannken)

VOP_STRATEGY() may still deadlock with devices.

Change FSTRANS from LAZY to NO.

(hannken)

Assert ufs_strategy() always gets used while current thread
holds a fstrans lock.

(hannken)

vrele_flush(): yield() every 100ms like we do it in vflush().

(hannken)

Fix typo in comment.

(msaitoh)

No ACL support for install media (SMALLPROG)

(christos)

Mask all the MMC counter interrupts if the MMC module is present.

(chs)

Use rnd_add_data_sync from the callback.

(Doesn't make a difference in HEAD but this is the stated API
contract and it matters if we want to pull this up.)

(riastradh)

Rockchip crypto engine RNG driver.

As found on the rk3288 and rk3399.  This driver only supports the
TRNG, not the rest of the crypto engine, although it uses the AES unit
to do a self-test at attach time to verify that the engine works.

There seem to be two versions of the Rockchip crypto engine, v1 and
v2; this one is for v1.  Can't name a driver `rkcryptov1' so we'll
clumsily call it `rkv1crypto' instead to leave room for `rkv2crypto'
later on.

The crypto binding derived from the Rockchip BSP Linux kernel, in the
location it appears on the rk3399, is in rk3399-crypto.dtsi, since
there doesn't seem to be a better place to put it at the moment among
this twisty maze of inclusions, all different.

(riastradh)

Minor correction to previous.

(ad)

PR kern/55268: tmpfs is slow

tmpfs_getpages(): handle the PGO_LOCKED case and implement lazy update of
atime/mtime.

(ad)

Start trying to reduce cache misses on vm_page during fault processing.

- Make PGO_LOCKED getpages imply PGO_NOBUSY and remove the latter.  Mark
  pages busy only when there's actually I/O to do.

- When doing COW on a uvm_object, don't mess with neighbouring pages.  In
  all likelyhood they're already entered.

- Don't mess with neighbouring VAs that have existing mappings as replacing
  those mappings with same can be quite costly.

- Don't enqueue pages for neighbour faults unless not enqueued already, and
  don't activate centre pages unless uvmpdpol says its useful.

Also:

- Make PGO_LOCKED getpages on UAOs work more like vnodes: do gang lookup in
  the radix tree, and don't allocate new pages.

- Fix many assertion failures around faults/loans with tmpfs.

(ad)

fix sun2 build

(christos)

Reorganise the locking and allocation of fstrans_lwp_info slightly, to
reduce contention.  "please go ahead" hannken@.

(ad)

reverse the order for sun2

(christos)

Fix sun2 build. Still there are some duplicate symbols in the rump libraries
that make this awkward.

(christos)

put librumpvfs_nofifofs after librumpvfs for the benefit of sun2

(christos)

Don't set PG_AOBJ on a page unless UVM_OBJ_IS_AOBJ(), otherwise it can
catch pages from e.g. uvm_loanzero_object.

(ad)

pool_cache_prime, not pool_cache_lowat, as suggested by nick.

(riastradh)

Missed a spot.

(riastradh)

Fix lists for RUMP-less builds

(martin)

- If the hardware provided NUMA info, then use it to decide how to set up
  the allocator's buckets, instead of doing round robin distribution.  There
  are open questions here but this is better than doing nothing.

- Kernel reserve pages are for the kernel not realtime threads.

(ad)

Mark amappl with PR_LARGECACHE.

(ad)

Flag EOI on TLB shootdown IPIs after the shootdown has been processed.

(ad)

softint_disestablish(): use a high priority xcall to determine that the
handler is no longer running.

(ad)

Install sparc/intr.h

(ad)

usr.bin/make: demonstrate actual behavior of .INCLUDEDFROMFILE

(rillig)

comments

(ad)

comments

(ad)

Fixed a problem that caused a page fault when attaching vmbus(4).

Dynamically allocate a page of memory with uvm_km_alloc(kernel_map, ...)
for Hyper-V hypercall.  However, this method can no longer be used to
make an executable page.
So we prevent it by using statically allocated memory for text segment.

(nonaka)

With the recent change of the EC address space handler, we no longer get
an ACPI_INTEGER in host byte order but a byte sized buffer with little
endian data.

Extract only the low 8 bits from buffer to get the fan speed again.

(mlelstv)

usr.bin/make: fix test for dollar and backslash at eol

The previous version of this test relied on the way how the shell
interprets a lonely backslash at the end of the line.  The NetBSD and
FreeBSD shells print the backslash, while Bash doesn't.

While here, make the escaping a bit simpler and align the test
descriptions with the actual test data.

(rillig)

Copyright maintenance

(skrll)

Use BUS_ADDR_{LO,HI}32.  NFCI.

(skrll)

__HAVE_CPU_DATA_FIRST is still true for sparc64 (which shares this file)

(martin)

Remove obsolete paragraph.

(riastradh)

Rework hifn(4) RNG logic.

Previously we would read 34 bits of entropy into the pool and stop,
never to try again.  This change finishes the conversion to an
on-demand source, although we still use a callout to (a) try again in
another second if the 7811 FIFO was emptied, and (b) wait at least
0.4 sec after we start the RNG to read anything from it as the
alleged documentation allegedly recommends.

(riastradh)

Tweak locking and use a pool cache for commands and dmamaps.

This is enough to get the crypto decelerator working in LOCKDEBUG;
previously it would crash the moment you looked at it.

(riastradh)

Merge updates from OpenBSD's hifn(4) crypto decelerator driver.

Highlights:

- Apply a ten-year-old fix to the driver's misuse of previously
  exposed ciphertext as IV for the next ciphertext with CBC.

  (Someone^TM should make sure there are no other drivers still doing
  this...)

- Various zeroing of buffers.

- Various minor cosmetic changes.

(riastradh)

Reduce diffs from OpenBSD hifn(4) header files.

(riastradh)

Remove bogus comment.

(riastradh)