branches: 1.1.4; 1.1.6; 1.1.8; 1.1.12;
Descend into include/rump and install headers.

If the arch uses __BSWAP_RENAME, provide non-namespaced kernel
symbols which just call the libc symbols. Allows to get rid of
rump machine/bswap.h.

XXX:
1) this is unlikely to work on non-NetBSD
==>
2) should look into including libkern in librump

branches: 1.1.2; 1.1.4; 1.1.8; 1.1.14; 1.1.30; 1.1.34; 1.1.42;
Provide rumpmachine bswap.h, which makes bswapxx() call rumpuser
directly instead of relying on a symbol in rumpkern. I would like
to make it call the libc symbol directly, but I don't currently know
how to make it do that MI.

Makes hfs work (on i386), as it avoids endless recursion in bswap64().
Thanks to dillo for the image!

Use <sys/bus.h> not <machine/bus.h>.

branches: 1.2.2; 1.2.6;
Add missing public members of bus_dmamap_t per interface specification
(nothing uses them here for now, so untested).

Add a bus.h blanket header which forces definition of bus space /
bus dma interfaces as functions and therefore makes it possible to
use rump on archs which suffer from macro maladies, i.e. everything
except i386 and amd64.

(build-tested on sparc64 and vax, which are usually the two extremes
of weirdness in these cases)

Fix rump definition of cpu_number().

Add ci_onproc.

branches: 1.22.18;
Apparently not all ports define struct clockframe in cpu.h, so we cannot
provide our definition that way. Instead, generate the struct clockframe
passed to hardclock() in MD code.

emulgate clockframe

branches: 1.20.6;
Allow compile-time optimizations to curlwp. This can have a pretty
staggering impact on performance. When running sendto() in a loop,
the improvement is 200k more calls per second with an inlined __thread
curlwp as opposed to the default. In other words, it shaves off hundreds
of CPU cycles per call (~20%). Even just eliminating the x86_curlwp()
call in favor of an inline gives an improvement of 60k calls per second.

update comments

Kludges like to procreate. Add alpha cpu_frequency() kludge.

mark rump's CPU_INFO_ITERATOR as being __unused.

branches: 1.16.4;
Inform the rump kernel hypervisor of valid thread contexts so
that the implementation can allocate and release storage for them
in an optimal fashion, if necessary.

branches: 1.15.4; 1.15.14;
Update copyright statements.

no functional change.

branches: 1.14.2;
Add SMP support for all architectures.

tested on sparc64 by martin

add ci_curlwp since we use it now

branches: 1.12.4;
rump_get_curlwp() is dead

fix previous to make sense

adjust for new rump_cpu

powerpc md stuff for libkern

Assemble with _LOCORE.

My clever hack was not enough. Disable DELAY via ifndef _RUMPKERN in
subr_prf until I manage to get all archs in line with a proper MI
signature.

Sauce with some kludges:
* revert rev 1.68 of emul.c: delay() is hopelessly MD and needs
whipping (not the fun kind)
* #undef DELAY in cpu.h

This will hopefully fix build for all archs and give me time to
hop through our 46521 fine archs and edit the headers.

Apparently some ports rely on cpu.h including machine/pcb.h.

branches: 1.4.2;
Add some shameful kludges for the vax build. Eventually I'd like
to use the native kernel machine directory, so don't waste too much
effort fighting this now.

branches: 1.3.16; 1.3.20; 1.3.26;
Make it possible to run rumps multithreaded. This brings real
locking and makes it possible to run file systems which create
threads. It also makes rump file system behaviour better match
file system behaviour in the kernel.

branches: 1.2.2;
few stubs

branches: 1.1.2; 1.1.4; 1.1.6; 1.1.8; 1.1.12; 1.1.14;
Introduce RUMPs - Runnable Userspace Meta-Programs

/sys/rump contains programs which run unmodified kernel code in an
emulated userspace environment. The kernel environment is provided
by librump. Currently supported are a number of file systems,
which by using puffs integrate seamlessly into the system and
provide a similar user experience to if the code was running as
part of the kernel. Potential future rumpification targets include
for example parts of the networking stack and some device drivers.

This work was supported by Google Summer of Code 2007.

branches: 1.3.6;
Reinstance cpu_counter.h so that we don't have to work around the
MD implications everywhere else.

branches: 1.2.12; 1.2.18;
Remove rump cpu_counter.h in favour of the arch native one. I'm
sure there was a reason why I originally added this, but I can't
find what would break without it now. If something does pop up,
I'll fix it properly this time instead of adding silly headers.

branches: 1.1.2; 1.1.4; 1.1.8; 1.1.10; 1.1.24; 1.1.28; 1.1.34; 1.1.36;
It's non-trivial to say that "no, we don't have cpu counter on this
arch (rump)", so emulate it. But this is suboptimal. The crux of
the problem seems to be that types.h contains both information on
the machine architecture (which we want) as well as other defines
such as __HAVE_CPU_COUNTER (which we don't want).

fix typos in comments.

s/static inline/static __inline/g for consistency.

branches: 1.21.16;
Add a dummy splraiseipl()

branches: 1.20.6;
Make IPL_ values different.
This makes ni real difference for rump, but avoods
KASSERTs in common code when we try to assert
"ipl not higher as...".
Suggested by riastradh, OK: pooka.

branches: 1.19.2;
Provide splsoftserial.

GRRR RUMP

branches: 1.18.4; 1.18.10; 1.18.14; 1.18.16;
Update copyright statements.

no functional change.

branches: 1.17.2;
Make IPLs not equal IPL_NONE to avoid some KASSERTs.

branches: 1.16.4; 1.16.6;
Rip out the rwlock spl emulation code. It never did anything useful
except keep my feet warm by consuming an insane amount of cpu
cycles -- in rump our current "cpu" context is never interrupted
& we have MULTIPROCESSOR.

(itch i'm scratching: it made the networking stack 10-20% slower.
this is one of the places where fast code actually matters)

Define IPL_SOFTSERIAL and provide a makeiplcookie() implementation,
so that kern/subr_pool.c can build again.

OK'ed by martin@

Assemble with _LOCORE.

printf needs IPL_HIGH

Add a few symbols required by nfsd

branches: 1.11.2;
+IPL_SOFTNET

Add IPL_SOFTCLOCK, something required it (although I can't quite
recall what, this has been floating in my tree for a while)

branches: 1.9.6; 1.9.10; 1.9.12; 1.9.16;
Use vfs_subr.c from sys/kern. This brings differences in the vnode
life cycle between rump and a real kernel to a minimum.

Merge vmlocking2 to head.

branches: 1.7.2; 1.7.6;
Interrupt handling changes, in discussion since February:

- Reduce available SPL levels for hardware devices to none, vm, sched, high.
- Acquire kernel_lock only for interrupts at IPL_VM.
- Implement threaded soft interrupts.

Emulate spls by using pthread rwlocks: splfoo() takes a read lock
and when doing processing in an interrupt (effectively when calling
biodone()), we take the write lock.

Define ipl_cookie_t, ipl_t.

branches: 1.4.2; 1.4.4;
"Support" splhigh(), requested by reinoud. But the spl emulation
code needs to be actually written soon ...

branches: 1.3.4;
Define splclock() so that this will at least compile on architectures
without _HAVE_TIMECOUNTER.

branches: 1.2.2; 1.2.4; 1.2.6; 1.2.8;
interrupt headers necessary for vfs_bio

branches: 1.1.2;
Introduce RUMPs - Runnable Userspace Meta-Programs

/sys/rump contains programs which run unmodified kernel code in an
emulated userspace environment. The kernel environment is provided
by librump. Currently supported are a number of file systems,
which by using puffs integrate seamlessly into the system and
provide a similar user experience to if the code was running as
part of the kernel. Potential future rumpification targets include
for example parts of the networking stack and some device drivers.

This work was supported by Google Summer of Code 2007.

Get rid of local machine/{mutex,rwlock}.h files by treating the
object storage as a single pointer (all archs have at least one
uintptr_t in the objects).

branches: 1.1.2; 1.1.6; 1.1.8; 1.1.10; 1.1.24; 1.1.28; 1.1.36;
Make it possible to run rumps multithreaded. This brings real
locking and makes it possible to run file systems which create
threads. It also makes rump file system behaviour better match
file system behaviour in the kernel.

provide pmap_wired_count macro.

Give the kernel/local pmaps actual storage.

That way friendly neighborhood macros won't go scribbling
in the wrong places.

branches: 1.6.18; 1.6.36;
Fix snafu in previous -- if your test it flawed, it doesn't matter
how many times you pass.

Reinstate the blanket pmap.h for archs which do not conform to the
kernel ABI (i.e. not i386 or amd64). Due to the "half function,
half macro, all noodles" nature of pmap.h, it's too entangling and
too brittle to keep up with an ifdeffy MI implementation.

branches: 1.4.4; 1.4.6;
Add a few MD ifdefs into the rump pmap stubs to enable us to get
rid of the local pmap.h and unify the pmap ABI with the kernel.
(XXX: compiles but doesn't work on vax)

Thanks once again to Greg Oster for helping with testbuilds.

Make pmap_kernel() a MI macro for struct pmap *kernel_pmap_ptr,
which is now the "API" provided by the pmap module. pmap_kernel()
remains as the syntactic sugar.

Bonus cosmetics round: move all the pmap_t pointer typedefs into
uvm_pmap.h.

Thanks to Greg Oster for providing cpu muscle for doing test builds.

branches: 1.2.2;
Add support bits necessary for rumpnet functionality.

branches: 1.1.2; 1.1.4; 1.1.8; 1.1.14; 1.1.30; 1.1.34; 1.1.40; 1.1.42;
Introduce RUMPs - Runnable Userspace Meta-Programs

/sys/rump contains programs which run unmodified kernel code in an
emulated userspace environment. The kernel environment is provided
by librump. Currently supported are a number of file systems,
which by using puffs integrate seamlessly into the system and
provide a similar user experience to if the code was running as
part of the kernel. Potential future rumpification targets include
for example parts of the networking stack and some device drivers.

This work was supported by Google Summer of Code 2007.

Get rid of local machine/{mutex,rwlock}.h files by treating the
object storage as a single pointer (all archs have at least one
uintptr_t in the objects).

branches: 1.2.4; 1.2.18; 1.2.22; 1.2.30;
Make it possible to run rumps multithreaded. This brings real
locking and makes it possible to run file systems which create
threads. It also makes rump file system behaviour better match
file system behaviour in the kernel.

branches: 1.1.4; 1.1.6; 1.1.8; 1.1.10;
add rw_tryenter() and rw_lock_held(), inspired by Adam Hamsik

Instead of punishing all archs for a fake vmparam.h, punish only
evbppc for its insolence and pick a ppc cpu family at random in
Makefile.rump.

branches: 1.3.2; 1.3.4;
Alpha requires 8k pages for mbuf size calculation to work (maybe
sparc64 too?). Make it so.

Bump virtual page shift from 10 to 12. Otherwise the mbuf generating
macros fail to operate properly due to M_EXT_MAXPAGES being too
large and hence struct _m_ext being over MSIZE already by itself.

branches: 1.1.2; 1.1.4; 1.1.6;
Give rump a faux vmparam.h to work around problematic ports (hooray
for evbppc). Obligatory inlined rant: it would be really nice if
we didn't have weird ports like that which sport totally incompatible
code for different kernels.

branches: 1.1.2; 1.1.18;
purge rump/net of component-specific opt directories

branches: 1.1.2; 1.1.18;
purge rump/net of component-specific opt directories

branches: 1.1.2; 1.1.18;
remove librump/rumpnet/opt, consolidate in rump/include/opt

branches: 1.1.2; 1.1.18;
opt files were nop't

branches: 1.1.2; 1.1.18;
purge rump/net of component-specific opt directories

branches: 1.1.2; 1.1.18;
purge rump/net of component-specific opt directories

branches: 1.1.2; 1.1.18;
Remove individual opt_foo.h files, continue consolidating to opt_rumpkernel.h

branches: 1.2.2; 1.2.18;
Define NETHER in rumpkernel

This fixes failures of ATF tests running on rump_server.

purge rump/net of component-specific opt directories

Retire EtherIP, we have L2TP instead.

branches: 1.1.2; 1.1.18; 1.1.20; 1.1.22;
purge rump/net of component-specific opt directories

branches: 1.1.2; 1.1.18;
purge rump/net of component-specific opt directories

branches: 1.1.2; 1.1.18;
purge rump/net of component-specific opt directories

branches: 1.2.16;
conform rump behavior to NetBSD kernel (in paticular net.inet.ip.gifttl sysctl)

branches: 1.1.2;
purge rump/net of component-specific opt directories

branches: 1.1.2; 1.1.18;
purge rump/net of component-specific opt directories

branches: 1.1.2; 1.1.18;
Remove individual opt_foo.h files, continue consolidating to opt_rumpkernel.h

branches: 1.1.2; 1.1.6; 1.1.14; 1.1.18;
l2tp(4) support rump.

branches: 1.1.2; 1.1.6;
Add a new link-aggregation pseudo interface named lagg(4)

- FreeBSD's lagg(4) based implementation
- MP-safe and MP-scalable

Added a kernel option to run LACP on a half duplex interface

Rump-ify ALTQ (librumpnet_altq.so)

branches: 1.6.8; 1.6.28;
merge the bouyer-socketcan branch to HEAD.

CAN stands for Controller Area Network, a broadcast network used
in automation and automotive fields. For example, the NMEA2000 standard
developped for marine devices uses a CAN network as the link layer.

This is an implementation of the linux socketcan API:
https://www.kernel.org/doc/Documentation/networking/can.txt
you can also see can(4).

This adds a new socket family (AF_CAN) and protocol (PF_CAN),
as well as the canconfig(8) utility, used to set timing parameter of
CAN hardware. Also inclued is a driver for the CAN controller
found in the allwinner A20 SoC (I tested it with an Olimex lime2 board,
connected with PIC18-based CAN devices).

There is also the canloop(4) pseudo-device, which allows to use
the socketcan API without CAN hardware.

At this time the CANFD part of the linux socketcan API is not implemented.
Error frames are not implemented either. But I could get the cansend and
canreceive utilities from the canutils package to build and run with minimal
changes. tcpudmp(8) can also be used to record frames, which can be
decoded with etherreal.

branches: 1.5.4;
Rump-ify if_pppoe

From s-yamaguchi@IIJ

use WSEMUL_NO_DUMB and WSEMUL_VT100

branches: 1.3.2;
Remove individual opt_foo.h files, continue consolidating to opt_rumpkernel.h

Remove a bunch of opt files.

It's a cute idea to have component-specific opt files, but also a
completely stupid one since there's no way of knowing how options
transcend component boundaries, and therefore if a set of options is
conflicting or not. So, just continue concentrating all of opt_foo.h in
opt_rumpkernel.h in accordance with the monolithic opt model.

Move constant cpp macros from Makefile.rump into opt_rumpkernel.h
(makes cc invocation lines a bit shorter)

branches: 1.1.2; 1.1.18;
remove librump/rumpnet/opt, consolidate in rump/include/opt

branches: 1.2.16;
Rump-ify if_pppoe

From s-yamaguchi@IIJ

branches: 1.1.2;
purge rump/net of component-specific opt directories

branches: 1.1.2; 1.1.18;
opt files were nop't

branches: 1.1.2; 1.1.18;
continue consolidating non-modular option files

branches: 1.1.2; 1.1.18;
continue consolidating non-modular option files

branches: 1.1.2; 1.1.18;
continue consolidating non-modular option files

branches: 1.1.2; 1.1.18;
purge rump/net of component-specific opt directories

branches: 1.1.2; 1.1.18;
Remove individual opt_foo.h files, continue consolidating to opt_rumpkernel.h

branches: 1.2.14;
Rumpify vlan(4)

branches: 1.1.2; 1.1.4;
purge rump/net of component-specific opt directories

branches: 1.1.2; 1.1.18;
opt files were nop't

[ozaki-r] Add wg files

branches: 1.1.2; 1.1.18;
continue consolidating non-modular option files

branches: 1.1.2; 1.1.18;
continue consolidating non-modular option files

branches: 1.1.2; 1.1.18;
continue consolidating non-modular option files

branches: 1.1.2; 1.1.18;
continue consolidating non-modular option files

Add capability to attach external memory to files on rumpfs. This
feature is useful e.g. for tight-memory systems where you don't need
block storage, but still need to provide some data via files.

Move "shotgun approach to rump syscalls" from rump.h to a separate header,
sort of as a hint that relying on those macros is not necessarily the
way you want to do things in 2016. Include things from rump.h for compat
for the time being, though.

install <rump/rumperrno2host.h>

branches: 1.7.4;
obsolete the installed <rump/scsitest.h> header and make it private
to tests/dev/scsipi

branches: 1.6.2;
install rumperr.h

branches: 1.5.2; 1.5.4;
don't install includes if MKRUMP = no

branches: 1.4.8;
whoopsie, need to commit this as part of scsitest

add rump_syscalls_compat.h

branches: 1.2.2; 1.2.4;
Adjust rump sources for external/internal interfaces.
No functional change.

branches: 1.1.4; 1.1.6; 1.1.10;
Descend into include/rump and install headers.

Unlike gnu sed our sed does not recognize inline labels (separated by
semi-colons) so split the lines, and make sure that command lists end
with a semi-colon.

rumpdefs.h, rumperr.h and rumperrno2host.h are generated from makerumpdefs.sh.
Don't edit those files directly. Use __inline instead of inline in the
script file. OK'd by christos.

branches: 1.29.16; 1.29.18;
snarf fcntl definitions

Generate rump_errno2host(): translates rump kernel errnos to host errnos.

Essentially, it's a partial I-know-what-I'm-doing syscall compat.

Functionality requested by Robert Millan.

Fix #define ERRNO EANOTHERRNO.

Was: #define RUMP_ERRNO EANOTHERRNO
Now: #define RUMP_ERRNO RUMP_EANOTHERRNO

pointed out by Sebastian Wicki on irc

Ignore multiline IOC macros by default.

branches: 1.25.6;
generate errno -> str routine (for portable programs)

need to grab KTRFACv[n] too

include ktrace constants

do enums, grab enum modctl

translate ino_t to a portability-friendlier type

Explicitly avoid MAXNAMELEN. I'm not claiming it's pretty, but most
likely it's more bang for the buck than adding cpp support to the
sed script.

grab struct dirent

branches: 1.18.2;
grab ioctl-related macros

Grab more network-related macros

grab mount-related macros too

Grab the sysvbfs args structure too, since the fs is useful for
simple testing. Incidentally, the structure is exactly the same as
for UFS...

fix sed command

Add the ability to define compat structs and include a few.

ok, rump_sockaddr_in was a pretty desperate kludge in the first place,
much better solved by syscall compat. but, now it's really time to let
it go since solaris thinks it's nice to define s_addr around any way
it likes.

Necessary _KERNEL wrap to previous.

make work for non-c++ (???)

While mucking in here, give the [local] client a chance to access all
rump kernel (i.e. NetBSD) errnos, not just EJUSTRETURN.

Finally remove the "temporary" __VTYPE_DEFINED stuff I added >5 years ago.

branches: 1.7.2;
add some network-related compat defs

branches: 1.6.8;
LK_* flags moved to sys/vnode.h

branches: 1.5.2; 1.5.4;
include AB_* macros from reboot.h

namespace RB_ from reboot.h

branches: 1.3.4; 1.3.6; 1.3.10;
adjust comments out of lock.h

include those appetizing lockmanger flags

Add script to statically generate rumpdefs.h instead of doing it buildtime.

PR/55781: Ruslan Nikolaev: rump_init() does differentiate when all CPUs are
initialized

branches: 1.72.2;
Remove old compat include of rump_syscallshotgun.h

It was separated in 2016 and is no longer needed.

Fix header compatibility with C++

This file is OS-agnostic and cannot include OS specific headers.

The fallback for NetBSD specific symbol __BEGIN_DECLS didn't work as it was
misspelled as _BEGIN_DECLS.

... but define RUMP_REGISTER_T if not defined.

Use the same exposure rules for register_t as <machine/types.h> does.
Eventually we could remove -D_KERNTYPES from all the rump libraries...

branches: 1.68.8;
Move "shotgun approach to rump syscalls" from rump.h to a separate header,
sort of as a hint that relying on those macros is not necessarily the
way you want to do things in 2016. Include things from rump.h for compat
for the time being, though.

PR/50631: Joerg Sonnenberger: Don't rename syscalls before you include
the relevant headers.

prop_dictionary_t hasn't been needed here in, um, 5 years, so remove
the type madness.

Don't include <rump/rumpvnode_if.h> from rump.h. It's not needed
unless you're doing something special, but requires register_t.
Adjust the few places which actually need rumpvnode_if.h.

fix typo: fnctl -> fcntl

from Robert Millan <rmh@gnu.org> via rumpkernel-users

branches: 1.63.4;
Add rump_boot_etfs_register(), which can be used to specify etfs
nodes that will be available immediately when mountroot is done
and file systems are available.

The intended use is for example for firmware images to be available when
config_mountroot() hooks run.

branches: 1.62.2;
Add Android support for rump kernel.

Reviewed by pooka@

add a compat prototype for rump_getversion()

rump_getversion() does not require a thread context, so make it a straight
call instead of being wrapped around rump_pub_getversion().

Move the "is arch capable of loading native kernel modules into
rump kernel" clauses from bsd.own.mk to Makefile.rump. Also,
add a rump_nativeabi_p() call to determine if rump kernel is
compiled with native ABI support.

Remove the "host" signal model. I'm not sure it ever made any sense.
I'm also quite sure nobody ever really wanted to use it.

Change the "raise" sigmodel to include the rump kernel PID in the call
to rumpuser_kill(). The purpose is to provide some hint for non-POSIX
platforms which don't sport signals and multiplex many clients in a
single address space.

While this is technically an ABI change, see above. Also, since
rumpuser_kill() was not documented, I'll just claim I'm fixing a bug in
the implementation (and will be sure to document the interface correctly
very soon).

Give reader-friendly aliases to rfork flags.

branches: 1.56.4;
Get rid of the rump__init() macro. it was never useful and just makes
debugging less obvious.

export register_t to rump kernel users on solaris

branches: 1.54.2;
Revert rev 1.51. Turns out the defines are still useful when you really
want a local client.

branches: 1.53.4;
Update copyright statements.

no functional change.

Duh, the nfsd hacks in tests still used RUMP_SYS_NETWORKING. It
appears that using nxr to search for users wasn't a very good idea.
Put networking back and make the test of the defines give out
#errors.

me be fixink this

g/c old-style syscall selection method

branches: 1.50.2;
There is a use case where preserving the parent's fd table is
relevant, so to accommodate that change rump_lwproc_newproc() to
rump_lwproc_rfork(). The new interface has the rfork() fd table
semantics. The equivalent of rump_lwproc_newproc() is
rump_lwproc_rfork(RUMP_RFCFDG).

Require server to be explicitly initialized with rump_init_server(url).
Also, add rump_daemonize_begin() / rump_daemonize_end() to help
with the "can't daemon() after pthread_create()" problem. Applications
could accomplish the same, but since it's such a common operation,
provide a little help.

RUMPCN_FORCEFREE is no longer needed.

Add a new signal model RUMP_SIGMODEL_RECORD which records all
signals which are posted to a process.

Start rework of system call proxying over socket ("sysproxy").
This incarnation is written in the user namespace as opposed to
the previous one which was done in kernel namespace. Also, rump
does all the handshaking now instead of excepting an application
to come up with the user namespace socket.

There's still a lot to do, including making code "a bit" more
robust, actually running different clients in a different process
inside the kernel and splitting the client side library from librump.
I'm committing this now so that I don't lose it, plus it generally
works as long as you don't use it in unexcepted ways: i've tested
ifconfig(8), route(8), envstat(8) and sysctl(8).

add a few more system calls

_RUMPKERNEL -> _KERNEL

forward-declare struct uio (required by rumpvnode_if.h)

Add rump_printevcnts() which dumps values from all event counters.
(NOTE! it is run unscheduled for various experimentation etcetc. reasons)

Add directory support to etfs: host directories and their contents
will be mapped to rumpfs based on the given key. The directory
can be mapped either for a single level or recursively down the
entire subtree.

Remove trailing comma in enum, per PR kern/43200.

(trailing commas in enums are allowed in C99, but not C89, and possibly
not in C++, so let's be nice to older and deviant compilers)

Move all signal-related from emul.c to signals.c. Additionally,
define a few alternate signal models for the rump kernel, including
ones where signals are ignored or sent to host processes.

Use "struct kauth_cred *" instead of kauth_cred_t in all exported
interfaces. Allows to remove hairbrained _t typedef dance.

branches: 1.37.2;
Expose rump_schedule() and rump_unschedule(). These can be used
for wrapping application space calls to the rump kernel without
having to do heavyweight interface specification with ifspec.

Add RUMP_SYS_FILEOPS for syscalls which operate on files (or at least
close to that).

branches: 1.35.2;
Fixed ukfs build on non-NetBSD system.

Add RUMP_SYS_OPEN define to replace open calls with their RUMP version.

Add rump_boot_set/gethowto(), which can be called before rump_init()
to control .... *tadaa* boothowto.

"rumppriv" goes back to "rump" per internal interface naming change.

Adjust rump sources for external/internal interfaces.
No functional change.

don't define curlwp, it's just (unused) sugar

add recvmsg and sendmsg to networking rump syscalls

Provide an interface for reboot.

g/c some prehistoric interfaces which have been superceded by others.

Allow to set size and host file offset for etfs files and rumpblk.

Give a p2k node an internal state. This allows us to do proper
reference counting and not release nodes based just on puffs'
impression of if they are free.

This also allows us to reclaim vnodes already in inactive if the
file system so desires. Some file systems, most notably ffs, change
file state already in inactive. This could lead to a deadlock in
the middle of inactive and reclaim if some other puffs operation
was processed in between (as exposed by haad's open(at) test
program).

Also, properly thread the componentname from lookup to the actual
vnode operation. This required the changes the rump componentname
routines. Yes, the rename case is truly mindbogglingly disgusting.
Puke for yourself.

* get rid of the old fakeblk registration interface. Instead,
introduce a new and improved "etfs" interface, which can be used
to register host files accessible from rump fs namespace. This
new interface is not restriced to block devices, and neither does
it require the same pathname in host namespace and rump namespace.
Therefore, the same host file can be represented both as a char
and block device in rump namespace.

* adjust rumpblk to make the above possible

* improve rumpfs: nodes are now created properly and not implicitly
tied to the vnode lifecycle

add ioctl to list of compile-time translated calls

forward declare struct stat to avoid implicit header dependencies

Add compat routines for vattr translation over time_t change.

Add handrolled stat30 compat syscalls for calling post-time_t rump
kernels from pre-time_t userlands.

Add rump_getversion(), which returns the version of NetBSD for the
rump kernel.

Add rump_set_curlwp() for context switch to an existing lwp/proc.

Nuke RUMPCN_HASNTBUF. The inspiration behind it must've been ... deep.

branches: 1.16.2;
Rename rump_cred_destroy() to rump_cred_put() -- they've been
refcounted since forever now.

Add rump_vfs_getmp(), which returns struct mount for a given path.
(yes, it does not take a reference to the mountpoint)

Since rump_module_load() doesn't actually load the module, change
the name to rump_module_init(). Also, adjust the signature to take
a direct pointer to modinfo and allow passing of props. Finally,
provide rump_module_fini().

Add proof-of-concept code for enabling system calls to rump virtual
kernels running in other processes on the same machine or on an
entirely different host. I wrote this a while ago and am now
committing it mainly to avoid losing it. It works, but could do
with a little tuning here and there.

What this will hopefully eventually buy us is the ability to use
standard userland tools to configure rump kernels, e.g. ifconfig(8)
and route(8) could be used to configure the networking stack provided
by a rump kernel. Also some distributed OS implications may apply.

fun fact: a system call which just does copyin/copyout takes >1000x
longer when made over the LAN as compared to when made on the same
machine.

rump_virtif_create prototype

shotgun approach to rump syscalls

PR kern/26878 FFSv2 + softdep = livelock (no free ram)
PR kern/16942 panic with softdep and quotas
PR kern/19565 panic: softdep_write_inodeblock: indirect pointer #1 mismatch
PR kern/26274 softdep panic: allocdirect_merge: ...
PR kern/26374 Long delay before non-root users can write to softdep partitions
PR kern/28621 1.6.x "vp != NULL" panic in ffs_softdep.c:4653 while unmounting a softdep (+quota) filesystem
PR kern/29513 FFS+Softdep panic with unfsck-able file-corruption
PR kern/31544 The ffs softdep code appears to fail to write dirty bits to disk
PR kern/31981 stopping scsi disk can cause panic (softdep)
PR kern/32116 kernel panic in softdep (assertion failure)
PR kern/32532 softdep_trackbufs deadlock
PR kern/37191 softdep: locking against myself
PR kern/40474 Kernel panic after remounting raid root with softdep

Retire softdep, pass 2. As discussed and later formally announced on the
mailing lists.

missed one prototype

typedef rump_sysproxy_t

branches: 1.7.2;
Rip out the rwlock spl emulation code. It never did anything useful
except keep my feet warm by consuming an insane amount of cpu
cycles -- in rump our current "cpu" context is never interrupted
& we have MULTIPROCESSOR.

(itch i'm scratching: it made the networking stack 10-20% slower.
this is one of the places where fast code actually matters)

branches: 1.6.2; 1.6.4;
Put syspuffs_glueinit in the rump kernel namespace.

_rump_init() - > rump__init()

rump_vfs_load() is really rump_module_load(), so call it that and
move it from rumpvfs to rumpkern.

Split vfs out of rumpkern into rumpvfs. Non-fs rumps no longer
include the file system code. File system rumps explicitly need
to include rumpvfs from now on.

Remove rump_vp locking interfaces, use RUMP_VOP instead.

Move rump public headers to include/rump

regen

branches: 1.52.6;
Regen.

Back out recent change to replace pool_cache with then general allocator.
Will return to this when I have time again.

Regen.

Regen.

Regen.

Regen.

branches: 1.46.6;
Regen.

Regen.

Regen.

Regen.

Regen.

Regen.

Regen.

Regen.

Regen.

branches: 1.37.2;
Regen.

Regen.

Regen.

namecache changes:

- Delete the per-entry lock, and borrow the associated vnode's v_interlock
instead. We need to acquire it during lookup anyway. We can revisit this
in the future but for now it's a stepping stone, and works within the
quite limited context of what we have (BSD namecache/lookup design).

- Implement an idea that Mateusz Guzik (mjg@FreeBSD.org) gave me. In
cache_reclaim(), we don't need to lock out all of the CPUs to garbage
collect entries. All we need to do is observe their locks unheld at least
once: then we know they are not in the critical section, and no longer
have visibility of the entries about to be garbage collected.

- The above makes it safe for sysctl to take only namecache_lock to get stats,
and we can remove all the crap dealing with per-CPU locks.

- For lockstat, make namecache_lock a static now we have __cacheline_aligned.

- Avoid false sharing - don't write back to nc_hittime unless it has changed.
Put a a comment in place explaining this. Pretty sure this was there in
2008/2009 but someone removed it (understandably, the code looks weird).

- Use a mutex to protect the garbage collection queue instead of atomics, and
adjust the low water mark up so that cache_reclaim() isn't doing so much
work at once.

Regen

branches: 1.32.2;
Regen.

branches: 1.31.6; 1.31.14;
regen

regen

regen

regen

branches: 1.27.2; 1.27.4;
Cull unused INRENAME and INRELOOKUP from callers.

Regen.

Update stats-keeping in sys/kern/vfs_cache.c to remove (most)
races while allowing consistent lockless sampling of the per-cpu
statistics without atomic operations. Update comment describing
the locking protocol to include this.

branches: 1.24.4;
Regen.

Regen.

branches: 1.22.10;
Regen from updated namei.src.

regen

regen

regen

regen

regen

branches: 1.16.8;
Regen.

branches: 1.15.4;
Regen for ISSYMLINK removal.

branches: 1.14.2;
regen: fix PARAMASK

Final regen for tonight.

Regen yet again.

Regen again.

Regen.

Regen for both preceding commits of namei.src together. I suppose I should
technically have regen'd separately for each but it didn't seem worthwhile.

Regen.

branches: 1.7.4;
regen: INRENAME

regen again (forgot both patches touched namei.h)

regen

regen

branches: 1.3.2;
Regen.

branches: 1.2.4; 1.2.6; 1.2.8;
reregen to get those most important rcsids right

regen

regen

regen

regen

branches: 1.130.2;
regen

regen

regen

Regen

regen for "struct sigaltstack" -> "stack_t"

Regen for eventfd(2) and timerfd(2).

regen

branches: 1.123.2;
regen

Regen with new makesyscalls.

(the large diff of rump_syscalls.h has been checked and was the point
of the makesyscalls change)

Regenerate syscalls

Respect syscall alias names in rump.

No binary change in this version.

The previous commit was interrupted in the middle by CVS network outage.

Regenerate native NetBSD syscalls

The only change is rump repecting syscall alias names.

No binary change.

Add ACL support for FFS. From FreeBSD.

Regen for futex call relocation.

Regen for native futex calls.

Merge the bouyer-xenpvh branch, bringing in Xen PV drivers support under HVM
guests in GENERIC.
Xen support can be disabled at runtime with
boot -c
disable hypervisor

Regen for removal of _lwp_gettid(2).

branches: 1.114.2;
Regen for _lwp_gettid(2).

Regen

branches: 1.112.2;
regen

Regen

regen.

Remove the kernel PMC code. Sent yesterday on tech-kern@.

This change:

* Removes "options PERFCTRS", the associated includes, and the associated
ifdefs. In doing so, it removes several XXXSMPs in the MI code, which is
good.

* Removes the PMC code of ARM XSCALE.

* Removes all the pmc.h files. They were all empty, except for ARM XSCALE.

* Reorders the x86 PMC code not to rely on the legacy pmc.h file. The
definitions are put in sysarch.h.

* Removes the kern/sys_pmc.c file, and along with it, the sys_pmc_control
and sys_pmc_get_info syscalls. They are marked as OBSOL in kern,
netbsd32 and rump.

* Removes the pmc_evid_t and pmc_ctr_t types.

* Removes all the associated man pages. The sets are marked as obsolete.

branches: 1.108.2; 1.108.4;
Revert vadvise(2) removal

This system call was used in legacy Lisp code, that was inherited to modern
age and still compiled against supported compat layers (e.g. in clisp,
oaklisp, Franz Lisp).

It used to instruct the kernel about paging policy (G/C aware, flush etc).

Newly compiled code (assuming that it will detect vadvise()) will use the
libc stub for vadvise(). The headers for this interface are gone.

vadvise(2) could be marked as COMPAT_80, but as long as we support ultrix,
sunos or aout68k ABI, don't bother with this.

Requested by <mrg>

Drop SYS_vadvise

The (o)vadvise syscall is dummy since the beginning of NetBSD.

It is an obsolete remnant from the old UNIX.

Sponsored by <The NetBSD Foundation>

regen
XXX: pullup-8

regen

branches: 1.104.6;
regen

Forgot these two generated files...

branches: 1.102.2;
regen

regen syscall files

regen

regen

regen

regen

regen syscall files

Regen

Regen

Finish the regen - some of the files from sys/kern got committed
accidentally with the fix to makesyscalls.sh - sorry about that.

Regen for 7.99.23

Regen

regen

regen

regen

Regenerate

Regen for changes related to separation of compat_sysv syscalls into a
separate module.

Regen again, this time without the extra word in the "generated from" line.

Regen from syscalls.master

regen after syscall update

regen

PR/49684: Pierre Pronchery: readlinkat(2) return type is wrong.

regen

branches: 1.79.2; 1.79.4;
Regen for fdiscard and posix_fallocate.

regen

regen

Regenerate

Eliminate weak symbols from rump kernel syscall handlers, part 5:

regen syscalls to eliminate weak aliases and link-time initialization

regen

branches: 1.73.2;
regen: time/timer related syscalls for rump kernels

reregengen

regen

regen

regen

regen

Regen for mknodat(2) device argument type change.

regen

regen

regen

regen

branches: 1.62.4;
regen

regen

branches: 1.60.2;
reregen

regen

reregen

regen

Implement sem_timedwait.

branches: 1.55.2;
Regen for posix_spawn

Regen syscalls with proper id info.

Change the syscall API for quotas over to the new non-proplib one.

- struct vfs_quotactl_args -> struct quotactl_args
- add sys/stdint.h to sys/quotactl.h for clean userland build
- install sys/quotactl.h in /usr/include
- update set lists for same
- add new marshalling code in libquota
- add new unmarshalling code in vfs_syscalls.c
- discard proplib interpreter code in vfs_quotactl.c
- add dispatching code for the 14 quotactl ops in vfs_quotactl.c
- mark the proplib quotactl syscall obsolete
- add a new syscall number for the new quotactl syscall
- change the name of the syscall to __quotactl()
- remove the decl of the old quotactl from quota/quotaprop.h
- add a decl of the new quotactl to sys/quotactl.h
- update the libc build
- update ktruss
- remove proplib marshalling code from libquota
- update copy of syscall table in gdb ppc sources
- hack rumphijack to accomodate new quotactl name (as I recall,
pooka wanted such a name change to simplify something, but I
don't really see what/how)

This change appears to require a kernel version bump for rumpish
reasons.

branches: 1.52.2; 1.52.6;
regen

regen after *setxattr constification

regen

regen: include rumpclient syscall headers from source tree instead of host

merge the bouyer-quota2 branch. This adds a new on-disk format
to store disk quota usage and limits, integrated with ffs
metadata. Usage is checked by fsck_ffs (no more quotacheck)
and is covered by the WAPBL journal. Enabled with kernel
option QUOTA2 (added where QUOTA was enabled in kernel config files),
turned on with tunefs(8) on a per-filesystem
basis. mount_mfs(8) can also turn quotas on.

See http://mail-index.netbsd.org/tech-kern/2011/02/19/msg010025.html
for details.

omstart

regenagain: make returning off_t work (without breaking other return
types on some archs)

unregen

regen: cast rval to return type instead of just using rval[0]

regen: NOERR syscalls

regen: preadv/pwritev

regen: always explicitly set errno (fixes some apps)

commit regen for int -> pid_t fix

branches: 1.39.2;
regen: more compat syscalls

regen: rework rump syscall stubs (see makesyscalls.sh log for details)

branches: 1.37.2;
regen

regen for comments

regen: SYCALL_NOSYS in rump_sysent[]

regen: optimize local syscall path a bit

+posix_fadvise

Refactor the sysproxy code so that rumpuser contains only the server side.

regen: getcwd

regen: umask

regen: rump syscalls for kern_prot interfaces

Empty commit to show makesyscalls.sh rev 1.99 didn't change anything (yet).

Regen syscalls to get compat header included.

regen: ksem syscalls

Regen for pathconf/fpathconf rumpification.

Regen for multiple inclusion protection.

regen: _RUMPKERNEL -> _KERNEL

regen: rump_enosys -> sys_nomodule for modular system calls

regen: get/setrlimit

branches: 1.20.2;
regen (for lfs syscalls #ifdef removal).

branches: 1.19.2;
regen for rump_sys_pipe()

regen: retval -> retval[2]

regen: rump_sys_modctl()

regen: fix rump varargs syscalls prototypes

regen: fh syscalls

regen: removal of pad arguments from the public interfaces

branches: 1.13.2;
regen

regen: new syscalls

regen: rump_sys_mount()

regen: kevent and kqueue

regen: make syscalls through a table

branches: 1.8.2;
regen for new syscalls

regen: int *error is dead.

branches: 1.6.2; 1.6.4;
regen: internal prototype for rump_syscalls.c

Regen. We now have compat __RENAME(). E.g. what was previously
rump_sys___stat50() is now rump_sys_stat() from the code.

Regen to pull proper "created from" revision strings.

merge christos-time_t

regen

branches: 1.1.2;
regen

Explicitly produce an error if the mythical kernel folk try to include
this header.

Give includers a way to say "yea I'll just deal with the type
compat on my own, thank you very many"

type compat header list doesn't need to be autogenerated

branches: 1.10.2;
Helps ELF platforms if I spell __ELF__ correctly ...

For non-NetBSD, assume leading underscore with !__ELF__ (fixes PE-COFF)

branches: 1.8.2;
add simple compat for other OSs to be able to use rump_syscalls.h

branches: 1.7.4; 1.7.8;
Missed pselect50 in previous (file was open in editor ... d0h).
Add mknod50 and fhstat50 too.

use newstyle rump syscall compat

branches: 1.5.2;
Add NetBSD 5 compat wrapping for *utimes()

do 5.0 compat for fstat

branches: 1.3.2; 1.3.4;
Add select to list on compat syscalls. Makes some code work for
me with a nb5 userland without having to compile with -g ;)

Add some forward declarations used by the interfaces.

add rump_syscalls_compat.h

branches: 1.1.2; 1.1.18;
Move "shotgun approach to rump syscalls" from rump.h to a separate header,
sort of as a hint that relying on those macros is not necessarily the
way you want to do things in 2016. Include things from rump.h for compat
for the time being, though.

regen

branches: 1.37.8; 1.37.10;
merge the bouyer-socketcan branch to HEAD.

CAN stands for Controller Area Network, a broadcast network used
in automation and automotive fields. For example, the NMEA2000 standard
developped for marine devices uses a CAN network as the link layer.

This is an implementation of the linux socketcan API:
https://www.kernel.org/doc/Documentation/networking/can.txt
you can also see can(4).

This adds a new socket family (AF_CAN) and protocol (PF_CAN),
as well as the canconfig(8) utility, used to set timing parameter of
CAN hardware. Also inclued is a driver for the CAN controller
found in the allwinner A20 SoC (I tested it with an Olimex lime2 board,
connected with PIC18-based CAN devices).

There is also the canloop(4) pseudo-device, which allows to use
the socketcan API without CAN hardware.

At this time the CANFD part of the linux socketcan API is not implemented.
Error frames are not implemented either. But I could get the cansend and
canreceive utilities from the canutils package to build and run with minimal
changes. tcpudmp(8) can also be used to record frames, which can be
decoded with etherreal.

branches: 1.36.4;
regen

periodic regen (because having new rcs ids in comments is most useful)

regen

regen

regen

regen

branches: 1.30.6;
regen

regen

regen

regen

regen

regen

revert pending fix: MAXNAMLEN causes trouble

regen

branches: 1.22.2;
regen

regen

regen

regen

Translate error codes between the hypervisor and rump kernel.

regen

regen

regen

regen

regen

regen

branches: 1.11.2;
regen

branches: 1.10.4;
regen for O_DIRECTORY

branches: 1.9.2; 1.9.4;
Regen. LK_INTERLOCK is gone.

Regen. LK_TYPEMASK and LK_RELEASE are gone.

Regen. LK_EXCLOTHER has gone.

Regen (LK_CANRECURSE is gone).

branches: 1.5.2; 1.5.4;
regen: add AB_* macros

regen

branches: 1.3.4; 1.3.6; 1.3.10;
regen

regen

regen

regen

branches: 1.7.2;
s/static inline/static __inline/g for consistency.

branches: 1.6.16;
regen

periodic regen (because having new rcs ids in comments is most useful)

regen

regen

regen

branches: 1.1.4; 1.1.6; 1.1.10; 1.1.12;
regen

regen

branches: 1.4.2;
s/static inline/static __inline/g for consistency.

branches: 1.3.16; 1.3.18;
regen

periodic regen (because having new rcs ids in comments is most useful)

branches: 1.1.2;
regen

branches: 1.1.2; 1.1.18;
Add capability to attach external memory to files on rumpfs. This
feature is useful e.g. for tight-memory systems where you don't need
block storage, but still need to provide some data via files.

regen rump kernel interfaces for header change.
(they were already manually edited for a prior commit, so not much change)

branches: 1.17.4;
regen

regen for etfs move from rumpvfs to rumpkern

branches: 1.15.2;
rump_getversion() does not require a thread context, so make it a straight
call instead of being wrapped around rump_pub_getversion().

branches: 1.14.6;
regen

branches: 1.13.8; 1.13.18;
regen: rump_lwproc_rfork

regen: rump_reboot is gone

regen: -rump_pub_syscall

regen:
- rump_sysproxy
+ rump_syscall

regen: new lwp/proc interfaces

regen from ages ago (to get rcsid consistent)

regen: kauth_cred_t -> struct kauth_cred *

branches: 1.6.2; 1.6.4;
regen for rump_module interface change

branches: 1.5.2;
regen: rump_kernelfsym_load()

regen: lwp interface changes

regen: scheduling points in rump_pub wrappers

regen: put all public interfaces created by ifspec into a rump_pub namespace

generate rump local interfaces

regen rump kernel interfaces for header change.
(they were already manually edited for a prior commit, so not much change)

branches: 1.5.8;
regen

branches: 1.4.4; 1.4.14; 1.4.24; 1.4.28;
regen

regen: scheduling points in rump_pub wrappers

regen: put all public interfaces created by ifspec into a rump_pub namespace

generate rump local interfaces

PR kern/57626: instead of an (arbitrary) FAKE_PAGE_SHIFT (and always
using 4k pages), query the hypervisor for the real page size of the host
kernel and use that for architectures that do not have compile time
constant page sizes.

Teach rump to process __link_set_evcnts entries. (Second part of
fix for PR kern/55088)

branches: 1.115.4;
Distinguish spin mutex and adaptive mutex on rump kernels for LOCKDEBUG

Formerly rump kernels treated the two types of mutexes as both adaptive for
LOCKDEBUG for some reasons.

Now we can detect violations of mutex restrictions on rump kernels such as
taking an adaptive mutex with holding a spin mutex as well as normal kernels.

branches: 1.114.10;
limit visibility of sysproxy calls to where they are needed

branches: 1.113.4;
Revert last change. This is not the right place for rumpuser_sync_icache().

Implement rumpuser_sync_icache hypercall.

weak alias removal implication: rumpuser_dl_globalsym() can now sleep
with the fishes.

branches: 1.110.2;
remove unnecesary verbosity: s/RUMP_CURLWP_MODEL/RUMP_CURLWP/

Allow compile-time optimizations to curlwp. This can have a pretty
staggering impact on performance. When running sendto() in a loop,
the improvement is 200k more calls per second with an inlined __thread
curlwp as opposed to the default. In other words, it shaves off hundreds
of CPU cycles per call (~20%). Even just eliminating the x86_curlwp()
call in favor of an inline gives an improvement of 60k calls per second.

branches: 1.108.2;
Looks like I'm done revamping, so bump interface version to 17,
where it will hopefully stay for all eternity (unless some snafus
are discovered before NetBSD 7).

Add a generalized rumpuser_syncfd() call which allows the caller
to request a sync or a barrier for fd.

Pass off_t over the hypercall interface as int64_t to avoid potential
problems in setups where off_t doesn't want to be 64bit.

I'm already excitedly waiting for 128bit off_t and _LARGE_LARGE_FILES

Pass enums over the hypercall interface as ints to avoid some
pathological scenarios from arising in setups where enums might be
of a different size in the rump kernel and hypervisor.

Add RUMPUSER_LWP_CLEAR instead of overloading RUMPUSER_LWP_SET.
This simplifies some alternative hypervisor implementations.

Ok, maybe using int64 for nanoseconds in the (sec,nsec) tuple was
a bit too future-proof. I think long is enough there (let's just
hope nobody redefines "nano"). Also, make seconds signed just in
case someone wants their clock to be in 1901.

Push rwlock upgrade and downgrade into the hypervisor where there's
at least a chance to implement them with minimal fuss.

Inform the rump kernel hypervisor of valid thread contexts so
that the implementation can allocate and release storage for them
in an optimal fashion, if necessary.

Flip parameter order in the clock hypercalls to make them consistent
with everything else wrt in/out parameters.

In kthread_create(), pass the priority and cpu index down to the
rump kernel hypervisor (though the current implementation doesn't take
advantage of them).

Make hypercall calling conventions consistent: iff a hypercall can fail,
it returns an int containing the error value.

Reduce hypercalls related to reading to essentially an amalgamation
of readv and preadv. ditto for writing. Hypercalls are so seldomly
used that it doesn't justify 3x the calls for syntactic sugar.

remove accidentally committed bit

Replace the various "get info from hypervisor" interfaces with one
unified rumpuser_getparam(), and make it return a plist. The
contents can come e.g. from the env or a config file. Make
identifiers starting with an underscore denote system identifiers
which must be implemented by hypervisor. (yea, j/k about the plist bit)

The fsync hypercall is now used only with close, so get rid of it.
rumpuser_bio() should do the necessary syncing.

Make the rump kernel upcalls which were previous available only to
the sysproxy module available for the entire hypervisor.

regroup calls a bit and sprinkle comments
no functional change for a change

RUMPUSER_OPEN_DIRECT is no longer necessary

changes to bio hypercalls, part 3/n:

retire the filemmap/memsync hypercalls, they're no longer used

Rework how the bio hypercalls work, part 1/n:

Reduce the set of hypercalls to one: "do block i/o". This not only
eliminates a lot of pseudo-duplicate code, it also gives the
hypervisor a lot more freedom on how to optimize the i/o.

Improve the time-related hypercalls so that's it's possible to
sleep until an absolute time on the host's monotonic clock (should
something like that be supported).

make the file monitoring hypercalls private to shmif

The ioctl hypercall is used only by the usb host controller driver,
so make it exclusive to that component. I'm happy to see it gone.

make the random hypercall more generic

roll mutex init hypercalls into one (one of them already took a flag anyway)

move struct lwp forward decl before its first use. otherwise it's rather
pointless. some whitespace landscaping too.

wrap unportable register_t with a long. should do it in the front
of the house too, but at least this is a start.

* treat kern_malloc.c as an unreal allocator (it's so lightweight)
* get rid of the rumpuser_realloc() hypercall
* pass size to rumpuser_free()

Let the includer figure out where to get stdint or equivalent from.

rumpuser_poll() is used only by sockin, so make the hypercall
private to that component.

* integrate rumpuser_getversion() with the hypervisor initialization
* rename some identifiers to better match reality

remove prototypes for interfaces no longer provided

Try to make sure that the appropriate calls to mutex_enter() takes
a spin mutex (i.e. does not relinquish cpu context while trying to
take the mutex).

Bump the hypercall interface version number. I'll be doing a bunch
of other cleanups to simplify the interface for the benefit of
alternative hypervisor implementations. I'll be riding this bump
and doing a second one only after I'm finished with all of the
changes.

Ensure that rump kernel component constructors from the main object
get processed. This applies to ones which were linked statically.
Unfortunately, that's where it got a bit tricky, since the dlsym()
interface searches the handle and all its dependencies. For the main
object the list of dependencies includes all the dynamic rump kernel
components that were included when the binary is linked. So, a long
story short, make only one pass through the objects to harvest all the
component entries, weed out the dupes, and initialize components from
an in-memory dupe-free list when so requested.

limit use of rumpuser.h to where it is meant to be used from

Do not assume that O_FOO flags have matching values in the rump kernel
and hypervisor.

branches: 1.72.2;
Add __printflike attribution to use vprintf and friends with an argument
as format string.

branches: 1.71.2;
Remove arc4random() and arc4randbytes() from the kernel API. Replace
arc4random() hacks in rump with stubs that call the host arc4random() to
get numbers that are hopefully actually random (arc4random() keyed with
stack junk is not). This should fix some of the currently failing anita
tests -- we should no longer generate duplicate "random" MAC addresses in
the test environment.

branches: 1.70.2;
Use __dead

Mark rumpuser_exit and rumpuser_thread_exit as dead.

Update copyright statements.

no functional change.

Nuke all threads belonging to a process calling exec before allowing
the exec handshake to return.

In addition to being The Right Thing To Do, fixes some nasty
conditions for CLOEXEC fd's (or at least does so in theory, I
couldn't create any problems although I tried).

Add an "exec" callback for the proxy code. The client can now
notify the rump kernel of an exec having taken place.

Pass the value of getprogname() from the client to the server and
record it in p_comm. This is nice for things like sockstat, since
they now display the client command name:

pain-rustique:43:~> rump.sockstat
USER COMMAND PID FD PROTO LOCAL ADDRESS FOREIGN ADDRESS
root xulrunner- 16 0 tcp 192.168.2.114.65507 204.152.190.12.80
root xulrunner- 16 1 tcp 192.168.2.114.65501 204.152.190.12.80
root xulrunner- 16 2 tcp 192.168.2.114.65500 204.152.190.12.80
root xulrunner- 16 3 tcp 192.168.2.114.65499 204.152.190.12.80
root xulrunner- 16 5 tcp 192.168.2.114.65498 204.152.190.12.80
root xulrunner- 16 6 tcp 192.168.2.114.65497 204.152.190.12.80
root socket 62 0 tcp6 *.http *.*
root socket 62 1 tcp *.http *.*
root socket 63 0 tcp6 *.81 *.*
root socket 63 1 tcp *.81 *.*

In case sys_reboot() was called by a remote client, put the response
in the socket before we shut down. This way the response to the
syscall travels to the caller and they know things worked correctly
instead of having to just assume.

branches: 1.63.2; 1.63.4;
Support SIGMODEL_RAISE for non-local clients.

add spop_procexit

There is a use case where preserving the parent's fd table is
relevant, so to accommodate that change rump_lwproc_newproc() to
rump_lwproc_rfork(). The new interface has the rfork() fd table
semantics. The equivalent of rump_lwproc_newproc() is
rump_lwproc_rfork(RUMP_RFCFDG).

Allow rump kernel to call dlsym(RTLD_DEFAULT).

Print a banner a connecting client. The banner contains rump sp
protocol version, os name, os revision and machine.

add rumpuser_sp_fini, which surpringly is the opposite of rumpuser_sp_init

implement mutex_owner()

rumpuser interface changed

Track lwp as the rumpuser mutex owner instead of pthread_t (this
is done in rumpuser for simplicity, since on the kernel side things
we assume we have only one pointer of space). As a side-effect,
we can no longer know if the current thread is holding on to a
mutex locked without curlwp context (basically all mutexes inited
outside of mutex_init()). The only thing that called rumpuser_mutex_held()
for a non-kmutex was the giant lock. So, instead implement recursive
locking for the giant lock in the rump kernel and get rid of the
now-unused recursive pthread mutex in the hypercall interface.

Require server to be explicitly initialized with rump_init_server(url).
Also, add rump_daemonize_begin() / rump_daemonize_end() to help
with the "can't daemon() after pthread_create()" problem. Applications
could accomplish the same, but since it's such a common operation,
provide a little help.

*facepalm*, adjust remote copyinstr to work in cases where the end
of the max copyin extends to an unmapped page.

Noticed, as usual, by tests.

Support physio for remote processes.
==> add support for remote vmspace vmapbuf/vunmapbuf
==> add proper support for copyin/out_vmspace
==> add support for remote vmspace uvm_io
==> add support for non-curproc rumpuser_sp_copyin/out
==> store remote context in vm_map->pmap instead of
pthread_specificdata

In short, makes read/write of most (all?) block devices work from
a remote rump client via rump syscalls.

Pass routines necessary for multithreaded operation down to rumpuser_sp.

bump rumpuser version

Refactor the sysproxy code so that rumpuser contains only the server side.

Missed a few symbols in previous ...

Also, reorganize rumpuser header inclusion to make sure problem is
caught already by the compiler.

Make librumpuser linkage once again free of librump.

problem pointed out by <he>

Start rework of system call proxying over socket ("sysproxy").
This incarnation is written in the user namespace as opposed to
the previous one which was done in kernel namespace. Also, rump
does all the handshaking now instead of excepting an application
to come up with the user namespace socket.

There's still a lot to do, including making code "a bit" more
robust, actually running different clients in a different process
inside the kernel and splitting the client side library from librump.
I'm committing this now so that I don't lose it, plus it generally
works as long as you don't use it in unexcepted ways: i've tested
ifconfig(8), route(8), envstat(8) and sysctl(8).

Version rumpuser ABI with a lightweight mechanism.

Add the ability to specify a preferred address the "map anon memory"
hypercall.

Don't pass "canfail" down to rumpuser_malloc -- there's quite little
we can do with that info way down there. Instead, pass alignment.
Implement rumpuser_malloc() with posix_memalign().

Support KTHREAD_JOINABLE/kthread_join(). Also fixes earlier bug
where all pthreads were created non-detached.

Make it possible to use the scheduler lock as the rumpuser condvar
interlock. This is applicable in cases where the actual interlock
is the CPU the currently running thread is scheduled on. Borrowing
the scheduler lock as the mutex mandated by pthread_cond_wait()
does away with need to have an additional mutex. This both optimizes
runtime execution and simplifies code, as the extra lock typically
lead to quite some trickeries to avoid the dungeon collapsing due
to zaps from the wand of deadlock.

Add rumpuser interface to fetch number of host cpus.

Add rumpuser_kill, which sends a signal to a host process.

branches: 1.38.2;
During bootstrap, process all modinfos in a DSO in one go. Get
rid of dependency tricks, since they are no longer necessary.

Introduce RUMP_COMPONENT. It behaves mostly like a simplified
module which is linked into the kernel and cannot be unloaded.
The main purpose is to get the proper constructors run and create
any /dev nodes necessary for said component. Once more of the
kernel (e.g. networking stack and device drivers) are converted to
MODULE and devfs pops up from somewhere, rump components can be
retired.

Rename rumpuser_dl_module_bootstrap() to rumpuser_dl_bootstrap(),
since it hasn't been involved only with modules for quite a while now.

branches: 1.35.2;
Fixed ukfs build on non-NetBSD system.

Build the kernel symbol table in rumpuser bootstrap and feed it to
the rump kernel. After this change it is possible to use the
in-kernel linker and rump_sys_modctl() to load kernel modules at
runtime.

Previously loading modules at runtime was possible only through
using the host system ld.so. Note that it is still preferred to
use shared libs when possible, since they are PIC and n virtual
kernels will only require one copy of r/o segments. However, when
there is no access to source code, a binary kernel module is the
only thing available ...