]> git.lizzy.rs Git - plan9front.git/log
plan9front.git
5 years agolib: update pci database
mischief [Sun, 5 May 2019 21:36:11 +0000 (14:36 -0700)]
lib: update pci database

5 years agobootrc: fix $rootdir and $rootspec handling (thanks lucio)
cinap_lenrek [Sun, 5 May 2019 15:02:37 +0000 (17:02 +0200)]
bootrc: fix $rootdir and $rootspec handling (thanks lucio)

we did not interpret the $rootdir and $rootspec environment
variables right. $rootdir is what gets bound to / (usually /root)
and $rootspec is the mountspec of /root.

5 years agobcm: make sure fiq is on enabled on cpu0
cinap_lenrek [Sun, 5 May 2019 12:04:27 +0000 (14:04 +0200)]
bcm: make sure fiq is on enabled on cpu0

5 years agobcm64: provide idlehands() function
cinap_lenrek [Sun, 5 May 2019 11:59:03 +0000 (13:59 +0200)]
bcm64: provide idlehands() function

5 years agobcm64: get rid of usb workaround delay
cinap_lenrek [Sun, 5 May 2019 11:57:48 +0000 (13:57 +0200)]
bcm64: get rid of usb workaround delay

5 years agousbdwc: enable Slowbuilkin workarround, improve split transaction timing, handle...
cinap_lenrek [Sun, 5 May 2019 11:34:02 +0000 (13:34 +0200)]
usbdwc: enable Slowbuilkin workarround, improve split transaction timing, handle erroring sleep(), debugging

i'v been seeing the error condition described above in the
Slowbulkin comment. so i'm enabling the work arround which
seems to fix the lockup.

in the split transaction case where we want to start the
transaction at frame start, acquire the ctlr lock *before*
checking if we are in the right frame number. so the start
will happen atomically. checking the software ctlr->sofchan
instead of checking the interrupt mask register seems to
be quicker.

setting the haint mask bit for the chan under ctlr lock
in chanio() instead of chanwait() avoids needing to acquire
the ctlr lock twice.

mask wakechan bits with busychan bitmap in interrupt handlers
so we will not try to wake up released chans by accident.

sleep() and tsleep() might get interrupted so we have to
release the split qlock in the split transaction case and
in all cases, make sure to halt the channel before release.

add some common debug functions to dump channel and controller
registers.

5 years agolibc: fix NaN check precedence bug in modf() (thanks BurnZeZ)
cinap_lenrek [Sat, 4 May 2019 22:26:17 +0000 (00:26 +0200)]
libc: fix NaN check precedence bug in modf() (thanks BurnZeZ)

5 years agoacid: add arm64 support
cinap_lenrek [Sat, 4 May 2019 01:06:10 +0000 (03:06 +0200)]
acid: add arm64 support

5 years agorootstub: add /sys/lib/dist/ndb/dhcp/
cinap_lenrek [Fri, 3 May 2019 21:59:27 +0000 (23:59 +0200)]
rootstub: add /sys/lib/dist/ndb/dhcp/

5 years agopc kernel: remove countpagerefs() (thanks BurnZeZ)
cinap_lenrek [Fri, 3 May 2019 21:52:49 +0000 (23:52 +0200)]
pc kernel: remove countpagerefs() (thanks BurnZeZ)

forgot to commit this...

5 years agokernel: exec support for arm64 binaries
cinap_lenrek [Fri, 3 May 2019 21:15:42 +0000 (23:15 +0200)]
kernel: exec support for arm64 binaries

5 years agobcm64: add experimental work in progress arm64 kernel for raspberry pi 3
cinap_lenrek [Fri, 3 May 2019 21:14:57 +0000 (23:14 +0200)]
bcm64: add experimental work in progress arm64 kernel for raspberry pi 3

5 years agobcm: don't call nil on PADDR() in dmaaddr(), return busdram dummy address
cinap_lenrek [Fri, 3 May 2019 20:53:09 +0000 (22:53 +0200)]
bcm: don't call nil on PADDR() in dmaaddr(), return busdram dummy address

5 years agobcm: move fiq saved pc adjust into lexception.s so it can be shared with arm64
cinap_lenrek [Fri, 3 May 2019 20:02:07 +0000 (22:02 +0200)]
bcm: move fiq saved pc adjust into lexception.s so it can be shared with arm64

5 years agouartpl011: dont touch line control while uart is enabled
cinap_lenrek [Fri, 3 May 2019 19:34:20 +0000 (21:34 +0200)]
uartpl011: dont touch line control while uart is enabled

5 years agoktrace: arm64 support
cinap_lenrek [Fri, 3 May 2019 19:16:23 +0000 (21:16 +0200)]
ktrace: arm64 support

5 years agoarm64: add mkfile
cinap_lenrek [Fri, 3 May 2019 19:12:31 +0000 (21:12 +0200)]
arm64: add mkfile

5 years agopython: arm64 support
cinap_lenrek [Fri, 3 May 2019 19:11:17 +0000 (21:11 +0200)]
python: arm64 support

5 years agogs: arm64 support
cinap_lenrek [Fri, 3 May 2019 19:10:45 +0000 (21:10 +0200)]
gs: arm64 support

5 years agogs: avoid stupid shifts by casting to uint64_t
cinap_lenrek [Fri, 3 May 2019 19:10:01 +0000 (21:10 +0200)]
gs: avoid stupid shifts by casting to uint64_t

5 years agoape: initial support for arm64
cinap_lenrek [Fri, 3 May 2019 19:06:46 +0000 (21:06 +0200)]
ape: initial support for arm64

5 years agolibthread: initial support for arm64
cinap_lenrek [Fri, 3 May 2019 19:03:12 +0000 (21:03 +0200)]
libthread: initial support for arm64

5 years agolibmp: add dummy mkfile for arm64
cinap_lenrek [Fri, 3 May 2019 19:02:13 +0000 (21:02 +0200)]
libmp: add dummy mkfile for arm64

5 years agolibsec: dummy mkfile for arm64
cinap_lenrek [Fri, 3 May 2019 19:01:29 +0000 (21:01 +0200)]
libsec: dummy mkfile for arm64

5 years agolibmach: initial arm64 support
cinap_lenrek [Fri, 3 May 2019 19:00:17 +0000 (21:00 +0200)]
libmach: initial arm64 support

5 years agolibc: initial arm64 support
cinap_lenrek [Fri, 3 May 2019 18:57:30 +0000 (20:57 +0200)]
libc: initial arm64 support

5 years ago7l: add missing AFMOVD with EXT/AUTO/LOREG operands to optab
cinap_lenrek [Wed, 1 May 2019 11:11:20 +0000 (13:11 +0200)]
7l: add missing AFMOVD with EXT/AUTO/LOREG operands to optab

5 years agokernel: get rid of checkpagerefs() debugging
cinap_lenrek [Wed, 1 May 2019 10:40:27 +0000 (12:40 +0200)]
kernel: get rid of checkpagerefs() debugging

was only implemented by the pc kernel. does not
account pages used by the mount cache.

5 years agokernel: export freepages() function so it can be used in mmurelease()
cinap_lenrek [Wed, 1 May 2019 08:07:39 +0000 (10:07 +0200)]
kernel: export freepages() function so it can be used in mmurelease()

5 years agokernel: include lock pointer in error report, stop spamming lockloop prints when...
cinap_lenrek [Wed, 1 May 2019 07:39:02 +0000 (09:39 +0200)]
kernel: include lock pointer in error report, stop spamming lockloop prints when we are panicing

5 years agokernel: insert memory barrier in the scheduler before setting up->mach = nil
cinap_lenrek [Wed, 1 May 2019 07:35:51 +0000 (09:35 +0200)]
kernel: insert memory barrier in the scheduler before setting up->mach = nil

we have to ensure that all stores saving the process state
have completed before setting up->mach = nil in the scheduler.
otherwise, another cpu could observe up->mach == nil while
the stores such as the processes p->sched label have not finnished.

5 years ago[5678vq]c: fix .safe node type for *FUNC() = *FUNC() sugen
cinap_lenrek [Wed, 1 May 2019 06:55:24 +0000 (08:55 +0200)]
[5678vq]c: fix .safe node type for *FUNC() = *FUNC() sugen

5 years agovmx: build vmx only for 386 or amd64
cinap_lenrek [Wed, 1 May 2019 06:19:29 +0000 (08:19 +0200)]
vmx: build vmx only for 386 or amd64

vmx uses non portable word unpacking macros, breaking
the build for arm64. vmx only works on a pc anyway.
this forces objtype to 386 on these machines, similar
to what the kernel mkfiles do.

5 years agoether8169: fix thinkpad A485 ethernet (thanks mischief)
cinap_lenrek [Sat, 27 Apr 2019 19:55:16 +0000 (21:55 +0200)]
ether8169: fix thinkpad A485 ethernet (thanks mischief)

attached is a patch to fix receive in the 8169 chip on my thinkpad
A485. i'm not sure why, but the same thing was done in 3d56a0fc4645
for Macv45.

nick

5 years ago7c: don't emit SXTW for non-register source operand
cinap_lenrek [Thu, 25 Apr 2019 17:44:16 +0000 (19:44 +0200)]
7c: don't emit SXTW for non-register source operand

5 years agodistproto: add /lib/news/ (thanks lyndon)
cinap_lenrek [Tue, 23 Apr 2019 13:15:26 +0000 (15:15 +0200)]
distproto: add /lib/news/ (thanks lyndon)

5 years ago7l: there is no BIC* $bimm variant
cinap_lenrek [Mon, 22 Apr 2019 01:05:51 +0000 (03:05 +0200)]
7l: there is no BIC* $bimm variant

5 years agousbxhci: make stuck usb transactions interruptable.
cinap_lenrek [Fri, 19 Apr 2019 21:39:47 +0000 (23:39 +0200)]
usbxhci: make stuck usb transactions interruptable.

some control transactions can confuse the xhci controller so
much that it even fails to respond to command abort or STOPEP
control command. with no way for us to abort the transaction
but a full controller reset.

we give the controller 5 seconds to abort our initial
transaction and if that fails we wake the recover process
to reset the controller.

thanks mischief for testing.

5 years agodistproto: add /lib/ndb/dhcp/ directory (thanks lyndon)
cinap_lenrek [Fri, 19 Apr 2019 21:31:24 +0000 (23:31 +0200)]
distproto: add /lib/ndb/dhcp/ directory (thanks lyndon)

5 years agonusb/usbd: stop sending port enable commands
mischief [Thu, 18 Apr 2019 09:48:35 +0000 (02:48 -0700)]
nusb/usbd: stop sending port enable commands

from what i can tell, sending port enable is a spec violation.

this fixes a hang during hub enumeration in the ASMedia
xhci controller when i plug in my IBM UltraNav SK-8845.

also, send unsuspend when port is suspended instead of enable.

from the USB 2 specification:

11.24.2.7.1.2 PORT_ENABLE
...
This bit may be set only as a result of a SetPortFeature(PORT_ENABLE).
...
The hub response to a SetPortFeature(PORT_ENABLE) request is not specified.

5 years ago7l: deal with huge (negative or > 24bit) register offsets, fix LACON, avoid DWORD...
cinap_lenrek [Wed, 17 Apr 2019 21:43:59 +0000 (23:43 +0200)]
7l: deal with huge (negative or > 24bit) register offsets, fix LACON, avoid DWORD in constant pool when we can sign extend

5 years ago7c: fix long to vlong/pointer conversion, avoid negative immediate offsets
cinap_lenrek [Wed, 17 Apr 2019 21:38:00 +0000 (23:38 +0200)]
7c: fix long to vlong/pointer conversion, avoid negative immediate offsets

we have to explicitely convert to vlong by sign or
zero extending as not every operation leaves a proper
zero/sign extended result in the register. for example
NEGW will zero extend, breaking negative int offsets
on pointers.

we explicitely insert SXTW or MOVWU instructions which
the peephole optimizer takes out again when it is safe
todo so.

when promoting constant offsets to immediate offsets,
make sure the offset will be in range. otherwise the
linker will produce not so optimal pointer arithmetic
instructions to calculate the offset.

5 years agodevsd: fix sddelpart() AGAIN
cinap_lenrek [Tue, 16 Apr 2019 14:25:33 +0000 (16:25 +0200)]
devsd: fix sddelpart() AGAIN

the previous "fix" missed to advance pp so after
we hit a invalid partition all following partitions
would be ignored.

5 years agoawk: fix nextfile crash (thanks leetspete)
cinap_lenrek [Tue, 16 Apr 2019 14:13:30 +0000 (16:13 +0200)]
awk: fix nextfile crash (thanks leetspete)

to reproduce the bug:

term% awk 'BEGIN{nextfile; nextfile;}'
sys: trap: fault read addr=0x10 pc=0x00019a3a
awk 6584: suicide: sys: trap: fault read addr=0x10 pc=0x00019a3a

5 years agodevsd: fix panic when using "delpart" to remove a partition that was already removed
BurnZeZ [Mon, 15 Apr 2019 19:48:03 +0000 (19:48 +0000)]
devsd: fix panic when using "delpart" to remove a partition that was already removed

5 years agopc: audiohda: recognize AMD Family 17h hd audio controller
mischief [Sun, 14 Apr 2019 06:11:19 +0000 (23:11 -0700)]
pc: audiohda: recognize AMD Family 17h hd audio controller

5 years agopc: devarch: recognize AMD Ryzen model
mischief [Sun, 14 Apr 2019 06:08:26 +0000 (23:08 -0700)]
pc: devarch: recognize AMD Ryzen model

5 years agopc: pci: add AMD FCH LPC bridge to southbridges
mischief [Sun, 14 Apr 2019 06:07:57 +0000 (23:07 -0700)]
pc: pci: add AMD FCH LPC bridge to southbridges

5 years agonusb/usbd: fix dump %U formatter
mischief [Sun, 14 Apr 2019 05:37:52 +0000 (22:37 -0700)]
nusb/usbd: fix dump %U formatter

5 years agoip(3): document TCP ctl close command
BurnZeZ [Sun, 14 Apr 2019 04:50:55 +0000 (04:50 +0000)]
ip(3): document TCP ctl close command

5 years agodevip: reject bad numeric ports (such as 9fs -> 9)
cinap_lenrek [Sun, 14 Apr 2019 01:22:05 +0000 (03:22 +0200)]
devip: reject bad numeric ports (such as 9fs -> 9)

5 years agobcm: move CONFADDR parsing into bootargs.c, simplify initcode start() args handling
cinap_lenrek [Thu, 11 Apr 2019 17:10:47 +0000 (19:10 +0200)]
bcm: move CONFADDR parsing into bootargs.c, simplify initcode start() args handling

5 years agobcm: don't use PADDR() to convert bus address to physical in vcore
cinap_lenrek [Thu, 11 Apr 2019 15:12:20 +0000 (17:12 +0200)]
bcm: don't use PADDR() to convert bus address to physical in vcore

5 years agokernel: get rid of KSTKSIZE alias for KSTACK
cinap_lenrek [Thu, 11 Apr 2019 12:05:23 +0000 (14:05 +0200)]
kernel: get rid of KSTKSIZE alias for KSTACK

5 years agomerge
cinap_lenrek [Thu, 11 Apr 2019 11:54:03 +0000 (13:54 +0200)]
merge

5 years agokernel: get rid of PTR2UINT() and UINT2PTR() macros
cinap_lenrek [Thu, 11 Apr 2019 11:51:38 +0000 (13:51 +0200)]
kernel: get rid of PTR2UINT() and UINT2PTR() macros

5 years agobcm: dont assume PHYSDRAM 0 in dmaaddr(), fix dmaioaddr()
cinap_lenrek [Thu, 11 Apr 2019 11:49:41 +0000 (13:49 +0200)]
bcm: dont assume PHYSDRAM 0 in dmaaddr(), fix dmaioaddr()

5 years agobcm: add pl011 uart driver
cinap_lenrek [Thu, 11 Apr 2019 11:21:06 +0000 (13:21 +0200)]
bcm: add pl011 uart driver

the raspi has two uarts, the pl011 and the mini. only one
can be used at a time due to pin muxing. the bcm kernel
uses the mini by default.

5 years agobcm: move okay() from uartmini.c to devarch.c
cinap_lenrek [Wed, 10 Apr 2019 13:06:56 +0000 (15:06 +0200)]
bcm: move okay() from uartmini.c to devarch.c

5 years agobcm: add irq.c to mkfile
cinap_lenrek [Wed, 10 Apr 2019 13:06:24 +0000 (15:06 +0200)]
bcm: add irq.c to mkfile

5 years agobcm: move interrupt handling out of trap.c into irq.c
cinap_lenrek [Wed, 10 Apr 2019 13:04:59 +0000 (15:04 +0200)]
bcm: move interrupt handling out of trap.c into irq.c

5 years agofix bad function declarations in manpages
BurnZeZ [Tue, 9 Apr 2019 18:34:17 +0000 (18:34 +0000)]
fix bad function declarations in manpages

5 years agothread(2): threadsetname takes a fmt string
BurnZeZ [Mon, 8 Apr 2019 22:18:15 +0000 (22:18 +0000)]
thread(2): threadsetname takes a fmt string

5 years agoaudiohda: add support for Gemini-Lake audio in Intel NUC nuc7pjyh (thanks sam-d)
cinap_lenrek [Mon, 8 Apr 2019 13:50:13 +0000 (15:50 +0200)]
audiohda: add support for Gemini-Lake audio in Intel NUC nuc7pjyh (thanks sam-d)

5 years ago7l: fix vlong constants in literal pool for non-MOV instructions
cinap_lenrek [Mon, 8 Apr 2019 12:15:39 +0000 (14:15 +0200)]
7l: fix vlong constants in literal pool for non-MOV instructions

5 years ago7l: fix bitcon instruction selection and encoding
cinap_lenrek [Mon, 8 Apr 2019 12:12:01 +0000 (14:12 +0200)]
7l: fix bitcon instruction selection and encoding

the possible bitmasks generated depend on the data width
of the instruction, so we introduce C_BITCON32 and C_BITCON64
operand types to keep them apart.

the encoding of the bitcon operation was wrong.

5 years ago7l: add arm64 linker (initial sync)
cinap_lenrek [Mon, 8 Apr 2019 12:05:27 +0000 (14:05 +0200)]
7l: add arm64 linker (initial sync)

5 years ago7a: fix indexreg operations
cinap_lenrek [Mon, 8 Apr 2019 11:53:41 +0000 (13:53 +0200)]
7a: fix indexreg operations

5 years ago7a: fix post increment operands
cinap_lenrek [Mon, 8 Apr 2019 11:50:11 +0000 (13:50 +0200)]
7a: fix post increment operands

5 years ago7c: fix registerization of vlong constants
cinap_lenrek [Mon, 8 Apr 2019 11:47:45 +0000 (13:47 +0200)]
7c: fix registerization of vlong constants

5 years agoadd arm64 c compiler and assembler (thanks charles forsyth)
cinap_lenrek [Mon, 8 Apr 2019 11:45:49 +0000 (13:45 +0200)]
add arm64 c compiler and assembler (thanks charles forsyth)

this is the the initial sync of charles forsyths plan9 c
compiler suite from http://bitbucket.org/plan9-from-bell-labs/9-cc
at changeset version 54:65fb8bb56c59

5 years agorootstub: create arm64 directories
cinap_lenrek [Mon, 8 Apr 2019 11:07:18 +0000 (13:07 +0200)]
rootstub: create arm64 directories

5 years agouptime: rewrite in awk
BurnZeZ [Sun, 7 Apr 2019 17:09:52 +0000 (17:09 +0000)]
uptime: rewrite in awk

5 years agolibauthsrv: authdial smprint() out of memory autism
cinap_lenrek [Sat, 6 Apr 2019 18:23:00 +0000 (20:23 +0200)]
libauthsrv: authdial smprint() out of memory autism

5 years agosgi: remove obsolete TSTKTOP constant
cinap_lenrek [Thu, 4 Apr 2019 13:28:04 +0000 (15:28 +0200)]
sgi: remove obsolete TSTKTOP constant

5 years agoxen: remove obsolete TSTKTOP constant
cinap_lenrek [Thu, 4 Apr 2019 13:25:10 +0000 (15:25 +0200)]
xen: remove obsolete TSTKTOP constant

5 years agobcm: remove obsolete TSTKTOP constant
cinap_lenrek [Thu, 4 Apr 2019 13:23:20 +0000 (15:23 +0200)]
bcm: remove obsolete TSTKTOP constant

5 years agopc64: get rid of TSTKTOP, USTKTOP is the end of user address space
cinap_lenrek [Thu, 4 Apr 2019 13:21:25 +0000 (15:21 +0200)]
pc64: get rid of TSTKTOP, USTKTOP is the end of user address space

the temporary stack segment used to be at a fixed address above or
below the user stack. these days, the temp stack is mapped dynamically
by sysexec so TSTKTOP is obsolete.

5 years agosshnet: get rid of Announced state, simplify
cinap_lenrek [Wed, 3 Apr 2019 21:49:22 +0000 (23:49 +0200)]
sshnet: get rid of Announced state, simplify

5 years agosshnet: implement listen (port forwarding)
cinap_lenrek [Wed, 3 Apr 2019 20:15:47 +0000 (22:15 +0200)]
sshnet: implement listen (port forwarding)

5 years agosshnet: don't leak error string
cinap_lenrek [Wed, 3 Apr 2019 12:29:49 +0000 (14:29 +0200)]
sshnet: don't leak error string

5 years agosshnet: don't leak "Dialing" connections when "connect" ctl write gets interrupted
cinap_lenrek [Wed, 3 Apr 2019 12:16:28 +0000 (14:16 +0200)]
sshnet: don't leak "Dialing" connections when "connect" ctl write gets interrupted

5 years agosshnet: pass on open failure error message, simplify
cinap_lenrek [Wed, 3 Apr 2019 11:45:54 +0000 (13:45 +0200)]
sshnet: pass on open failure error message, simplify

return the error message from MSG_CHANNEL_OPEN_FAILURE
in the "connect" control write.

use a extra state "Finished" to distinguish server from client
initiated teardown. that way we do not need to track if we
send the MSG_CHANNEL_CLOSE message in closeclient(). this way
we also cannot be fooled by misbehaving server.

simplify hangupclient() by removing state transitions and doing
them in the caller explicitely. that way we can use hangupclient()
instead of dialedclient().

5 years agosshnet: fix eof and close handling, use proper packet size, cleanup
cinap_lenrek [Wed, 3 Apr 2019 08:49:47 +0000 (10:49 +0200)]
sshnet: fix eof and close handling, use proper packet size, cleanup

5 years agovt: dont make fs procs hang arround when closing rio window
cinap_lenrek [Tue, 2 Apr 2019 18:44:35 +0000 (20:44 +0200)]
vt: dont make fs procs hang arround when closing rio window

5 years agosshnet(4): fix user@host syntax, add BUGS section, add ip(3) reference
cinap_lenrek [Tue, 2 Apr 2019 17:42:09 +0000 (19:42 +0200)]
sshnet(4): fix user@host syntax, add BUGS section, add ip(3) reference

5 years agossh: print usage for unknown flags, cleanup
cinap_lenrek [Tue, 2 Apr 2019 17:22:19 +0000 (19:22 +0200)]
ssh: print usage for unknown flags, cleanup

5 years agossh(1): add sshnet(4) reference
cinap_lenrek [Tue, 2 Apr 2019 16:46:00 +0000 (18:46 +0200)]
ssh(1): add sshnet(4) reference

5 years agosshnet: reduce memory consumption by lowering stack sizes
cinap_lenrek [Tue, 2 Apr 2019 16:29:22 +0000 (18:29 +0200)]
sshnet: reduce memory consumption by lowering stack sizes

5 years agosshnet: fix memory leak
cinap_lenrek [Tue, 2 Apr 2019 16:06:25 +0000 (18:06 +0200)]
sshnet: fix memory leak

5 years agosshnet: allow 0 port
cinap_lenrek [Tue, 2 Apr 2019 15:51:15 +0000 (17:51 +0200)]
sshnet: allow 0 port

5 years agosshnet: fix write count for ctl message writes
cinap_lenrek [Tue, 2 Apr 2019 15:32:34 +0000 (17:32 +0200)]
sshnet: fix write count for ctl message writes

5 years agosshnet: actually make sure ssh established connection before exiting main proc
cinap_lenrek [Tue, 2 Apr 2019 15:28:56 +0000 (17:28 +0200)]
sshnet: actually make sure ssh established connection before exiting main proc

this fixes password prompts and handles errors properly.

5 years agosshnet(4): add documentation
cinap_lenrek [Tue, 2 Apr 2019 14:40:14 +0000 (16:40 +0200)]
sshnet(4): add documentation

5 years agosshnet: bring back sshnet using ssh(1) mux mode
cinap_lenrek [Tue, 2 Apr 2019 14:23:01 +0000 (16:23 +0200)]
sshnet: bring back sshnet using ssh(1) mux mode

this is a port of the original ssh1 sshnet to our
ssh2 client using mux mode.

5 years agossh: add experimental mux mode
cinap_lenrek [Tue, 2 Apr 2019 09:18:50 +0000 (11:18 +0200)]
ssh: add experimental mux mode

in mux mode, ssh relays raw MSG_CHANNEL_*
messages on standard input and output while
still handling authentication and key exchange
internally.

the intend is to use the mux mode to implement
something like the old sshnet ontop of ssh.

5 years agomerge
cinap_lenrek [Tue, 2 Apr 2019 07:05:01 +0000 (09:05 +0200)]
merge

5 years agossh: implement -W option for making remote tcp connections
cinap_lenrek [Tue, 2 Apr 2019 07:03:35 +0000 (09:03 +0200)]
ssh: implement -W option for making remote tcp connections

5 years agomerge
Alex Musolino [Mon, 1 Apr 2019 12:44:14 +0000 (23:14 +1030)]
merge

5 years agofix DST begin/end timestamps for South Australia
Alex Musolino [Mon, 1 Apr 2019 12:25:49 +0000 (22:55 +1030)]
fix DST begin/end timestamps for South Australia