]> git.lizzy.rs Git - plan9front.git/log
plan9front.git
7 years agomerge
cinap_lenrek [Fri, 21 Apr 2017 17:25:25 +0000 (19:25 +0200)]
merge

7 years agossh: actually handle flow control and channel id's
cinap_lenrek [Fri, 21 Apr 2017 17:23:56 +0000 (19:23 +0200)]
ssh: actually handle flow control and channel id's

7 years agofix yacc bug (found by dan cross)
aiju [Fri, 21 Apr 2017 17:16:44 +0000 (17:16 +0000)]
fix yacc bug (found by dan cross)

7 years agossh: increase initial window size for better throughput
cinap_lenrek [Thu, 20 Apr 2017 23:45:11 +0000 (01:45 +0200)]
ssh: increase initial window size for better throughput

7 years agorsa(8): fix description of rsa2ssh (now, ssh2 format only)
cinap_lenrek [Thu, 20 Apr 2017 22:23:46 +0000 (00:23 +0200)]
rsa(8): fix description of rsa2ssh (now, ssh2 format only)

7 years agossh: cleanup debug messages
cinap_lenrek [Thu, 20 Apr 2017 22:09:30 +0000 (00:09 +0200)]
ssh: cleanup debug messages

7 years agossh: do not try authentication methods that we know are not possible
cinap_lenrek [Thu, 20 Apr 2017 21:45:56 +0000 (23:45 +0200)]
ssh: do not try authentication methods that we know are not possible

7 years agossh: implement password and keyboard-interactive authentication methods
cinap_lenrek [Thu, 20 Apr 2017 20:28:21 +0000 (22:28 +0200)]
ssh: implement password and keyboard-interactive authentication methods

7 years agovt: cursoron, don't resize winow when replying history
cinap_lenrek [Wed, 19 Apr 2017 20:34:47 +0000 (22:34 +0200)]
vt: cursoron, don't resize winow when replying history

7 years agossh: fix locking, and key reexchange handling
cinap_lenrek [Wed, 19 Apr 2017 19:03:26 +0000 (21:03 +0200)]
ssh: fix locking, and key reexchange handling

when we initiate re-key exchange we data packets can still
come in. so instead we have everything that can come in all
the time in dispatch() function (including KEXINIT) and have
the receiver process just call that in a loop. exclude dispatch
and the sender proc from corrupting each others sendpkt() calls
with the QLock sl.

7 years ago/lib/rob, /lib/theo: I don't think so.
stanley lieber [Wed, 19 Apr 2017 17:38:30 +0000 (13:38 -0400)]
/lib/rob, /lib/theo: I don't think so.

7 years agoremove accidentally pushed change to .hgignore
aiju [Wed, 19 Apr 2017 10:16:00 +0000 (10:16 +0000)]
remove accidentally pushed change to .hgignore

7 years agoblit: always show top-left corner; add tony kaku's -m flag
aiju [Wed, 19 Apr 2017 10:13:36 +0000 (10:13 +0000)]
blit: always show top-left corner; add tony kaku's -m flag

7 years agochange cpurc to look for service folder in /cfg/$sysname and /cfg/default
aiju [Wed, 19 Apr 2017 09:00:44 +0000 (09:00 +0000)]
change cpurc to look for service folder in /cfg/$sysname and /cfg/default

7 years agodisable all services except 17019, 17020 by default
aiju [Wed, 19 Apr 2017 08:51:18 +0000 (08:51 +0000)]
disable all services except 17019, 17020 by default

7 years agovt: but not too fast :-)
cinap_lenrek [Wed, 19 Apr 2017 00:18:28 +0000 (02:18 +0200)]
vt: but not too fast :-)

if drawing can't keep up with the host, make sure we
eventually redraw the screen and check for user events.

7 years agovt: make it fast
cinap_lenrek [Tue, 18 Apr 2017 23:46:20 +0000 (01:46 +0200)]
vt: make it fast

7 years agovt: send interrupt on exit, open logfile OCEXEC, run host after environment got expor...
cinap_lenrek [Tue, 18 Apr 2017 01:13:15 +0000 (03:13 +0200)]
vt: send interrupt on exit, open logfile OCEXEC, run host after environment got exported, send rest of arguments to host

7 years agossh: close the channel when vt hangs up
cinap_lenrek [Tue, 18 Apr 2017 01:06:39 +0000 (03:06 +0200)]
ssh: close the channel when vt hangs up

7 years agovt: allow for a bit of buffering
cinap_lenrek [Mon, 17 Apr 2017 22:21:51 +0000 (00:21 +0200)]
vt: allow for a bit of buffering

7 years agovt: convert from event to threads
cinap_lenrek [Mon, 17 Apr 2017 22:18:03 +0000 (00:18 +0200)]
vt: convert from event to threads

the event library doesnt do any flow control. if the host
keeps sending data while the user is mousing arround,
extract() will collect the data until memory fills up.

so instead we abandon the event library and convert
the program to use channels.

7 years agovt: increase history buffer to 64K runes
cinap_lenrek [Mon, 17 Apr 2017 03:32:01 +0000 (05:32 +0200)]
vt: increase history buffer to 64K runes

7 years agossh: weekend project
cinap_lenrek [Mon, 17 Apr 2017 03:22:16 +0000 (05:22 +0200)]
ssh: weekend project

manpage should follow...

7 years agofactotum: append public rsa encyption exponent after the modulus
cinap_lenrek [Mon, 17 Apr 2017 03:18:24 +0000 (05:18 +0200)]
factotum: append public rsa encyption exponent after the modulus

this makes implementing ssh-rsa authentication easier, as we
then can convert the public key directly to ssh format and check
if the server will accept that public key.

tlshand just needs the modulus to see if the public key matches
the one it has from the certificate.

7 years agorsa2ssh: drop support for version 1 key format
cinap_lenrek [Mon, 17 Apr 2017 03:13:46 +0000 (05:13 +0200)]
rsa2ssh: drop support for version 1 key format

7 years agovt: implement snarf support
cinap_lenrek [Mon, 17 Apr 2017 03:12:07 +0000 (05:12 +0200)]
vt: implement snarf support

7 years agopc64: state mp.h dependency for archacpi.$O
cinap_lenrek [Tue, 11 Apr 2017 00:26:39 +0000 (02:26 +0200)]
pc64: state mp.h dependency for archacpi.$O

7 years agoupas/fs: remove imap lastread debounding
cinap_lenrek [Sun, 9 Apr 2017 02:36:47 +0000 (04:36 +0200)]
upas/fs: remove imap lastread debounding

its unclear why this was here, waketime logic should handle
this already.

7 years agoupas/fs: fix sync condition
cinap_lenrek [Sun, 9 Apr 2017 02:32:38 +0000 (04:32 +0200)]
upas/fs: fix sync condition

we used to check for mb->d->name != nil before stating, to avoid
stating pop3/imap mailboxes who don't really have a local file,
but this breaks when the md->d is reconstructed (faked) from
the index! resulting in the mailbox stop being refreshed.

the solution is to not have mb->d == nil for imap/pop mailoxes.

7 years agoupas/fs: make use of Maxmsg consistent
cinap_lenrek [Sat, 8 Apr 2017 21:01:19 +0000 (23:01 +0200)]
upas/fs: make use of Maxmsg consistent

7 years agoupas/fs: simplify imap4read()
cinap_lenrek [Sat, 8 Apr 2017 21:00:48 +0000 (23:00 +0200)]
upas/fs: simplify imap4read()

7 years agoupas/fs: fix dir comparsion and skip directories in mdir, avoid stat
cinap_lenrek [Sat, 8 Apr 2017 20:59:15 +0000 (22:59 +0200)]
upas/fs: fix dir comparsion and skip directories in mdir, avoid stat

- ignore directories in dirskip()
- use sortkey of 0 for invalid items, otherwise it could confuse qsort()
- use file size from dirreadall() instead of doing stat
- various cleanups

7 years agoupas/send: use openfolder() to deliver mail to avoid code duplication
cinap_lenrek [Sat, 8 Apr 2017 20:52:06 +0000 (22:52 +0200)]
upas/send: use openfolder() to deliver mail to avoid code duplication

7 years agoupas/common: deliver mail to mdir as .tmp file and rename after it has been fully...
cinap_lenrek [Sat, 8 Apr 2017 20:50:17 +0000 (22:50 +0200)]
upas/common: deliver mail to mdir as .tmp file and rename after it has been fully written

theres a race condition when mail delivery to mdir is slow,
then upas/fs sees partial mail file and caches the truncated
file size.

to avoid this, delivery will create the new mail file with
the .tmp extension (which is ignored by upas/fs) and after
everything has been written, rename it to the final name.

7 years agokernel: avoid waserror() botch in devwalk (from drawterm, thanks aiju)
cinap_lenrek [Tue, 4 Apr 2017 22:34:07 +0000 (00:34 +0200)]
kernel: avoid waserror() botch in devwalk (from drawterm, thanks aiju)

calculate alloc flag before waserror(), as compilers like
gcc will not notice the value changing later because
setjump() restores the old value due to callee-saves.

change is applies here to make it easier to merge with
drawterm.

thanks to aiju for debugging this; used to cause drawterm
memory leak until compiled with gcc -O0.

7 years agokernel: fix memory leak in checkpagerefs() debug function (thanks aiju)
cinap_lenrek [Tue, 4 Apr 2017 18:13:31 +0000 (20:13 +0200)]
kernel: fix memory leak in checkpagerefs() debug function (thanks aiju)

7 years agotlshand: simplify tlsReadN()
cinap_lenrek [Tue, 4 Apr 2017 17:01:57 +0000 (19:01 +0200)]
tlshand: simplify tlsReadN()

7 years agotlshand: fix mpint to bytes conversion, reorganize send/recv buffer, check for overfl...
cinap_lenrek [Mon, 3 Apr 2017 23:59:17 +0000 (01:59 +0200)]
tlshand: fix mpint to bytes conversion, reorganize send/recv buffer, check for overflow in msgSend()

when converting mpint to bytes, always pad it to the size of
the modulus (RSA,DHE,ECDHE). mptobytes() now takes a byte len
parameter which the caller usually calculates from the group
modulus using mpsignif(). this bug sometimes caused "bad record mac"
after the handshake.

use a shared buffer, given that msgSend()/msgRecv() don't overlap
we can use the first half for sending, and the top half for
receiving, shifting down as neccesary. the space beween sendp and
recvp is free.

explicitely check for overflow in msgSend().

7 years agomerge
cinap_lenrek [Mon, 3 Apr 2017 00:50:36 +0000 (02:50 +0200)]
merge

7 years agotlshand: fix ECDHE and DHE for SSLv3
cinap_lenrek [Mon, 3 Apr 2017 00:48:47 +0000 (02:48 +0200)]
tlshand: fix ECDHE and DHE for SSLv3

7 years agofix typo in 9p(2) manpage
aiju [Mon, 3 Apr 2017 00:11:40 +0000 (00:11 +0000)]
fix typo in 9p(2) manpage

7 years agodevvga: remove unused copy of checkport() function
cinap_lenrek [Sun, 2 Apr 2017 18:17:44 +0000 (20:17 +0200)]
devvga: remove unused copy of checkport() function

7 years agodevvga: fix mistake
cinap_lenrek [Sun, 2 Apr 2017 18:10:44 +0000 (20:10 +0200)]
devvga: fix mistake

7 years agodevvga: include hwgc in vgactl file
cinap_lenrek [Sun, 2 Apr 2017 18:08:15 +0000 (20:08 +0200)]
devvga: include hwgc in vgactl file

7 years agoaux/wpa: go to background when not prompting, handle open networks
cinap_lenrek [Sun, 2 Apr 2017 17:05:56 +0000 (19:05 +0200)]
aux/wpa: go to background when not prompting, handle open networks

7 years agopc(1): fix memory leak and add clog() function
aiju [Sun, 2 Apr 2017 14:47:23 +0000 (14:47 +0000)]
pc(1): fix memory leak and add clog() function

7 years agonusb/lib: make usbcmd() return value symmetic; returning size of data phase (if any...
cinap_lenrek [Sat, 1 Apr 2017 20:19:58 +0000 (22:19 +0200)]
nusb/lib: make usbcmd() return value symmetic; returning size of data phase (if any) (thanks aiju)

usbcmd() with Rh2d used to return the command size (8+ndata) wile returning
only ndata for Rd2h. this changes it to always return ndata for Rh2d. it
mostly doesnt matter as Rh2d callers only check r < 0 for error, but this
makes the interface symmetic.

7 years agohjfs: avoid 8c "non-interruptable temporary" warning
cinap_lenrek [Sat, 1 Apr 2017 13:21:28 +0000 (15:21 +0200)]
hjfs: avoid 8c "non-interruptable temporary" warning

7 years agochgrp: remove unused function declaration (thanks archeus)
aiju [Sat, 1 Apr 2017 09:01:26 +0000 (09:01 +0000)]
chgrp: remove unused function declaration (thanks archeus)

7 years agogames/blit: remove strange debugging line
aiju [Sat, 1 Apr 2017 08:57:03 +0000 (08:57 +0000)]
games/blit: remove strange debugging line

7 years agondb/dns: double Maxretries for long cname redirection chains
cinap_lenrek [Sat, 1 Apr 2017 01:35:30 +0000 (03:35 +0200)]
ndb/dns: double Maxretries for long cname redirection chains

7 years agonusb/serial: pl2303: better error handling
aiju [Fri, 31 Mar 2017 20:20:25 +0000 (20:20 +0000)]
nusb/serial: pl2303: better error handling

7 years agomerge
cinap_lenrek [Thu, 30 Mar 2017 21:39:03 +0000 (23:39 +0200)]
merge

7 years agosdnvme: enable in pcf, pccpuf, pc64 kernel configuration
cinap_lenrek [Thu, 30 Mar 2017 21:37:28 +0000 (23:37 +0200)]
sdnvme: enable in pcf, pccpuf, pc64 kernel configuration

7 years agosdnvme: don't write completion queue doorbell register when nothing has been processed
cinap_lenrek [Thu, 30 Mar 2017 21:33:46 +0000 (23:33 +0200)]
sdnvme: don't write completion queue doorbell register when nothing has been processed

turns out on real hardware, the front falls off if we write
the completion queue doorbell registers without consuming
an entry. so only write the register when we have processed
something.

7 years agohjfs: disable hjfs check until more functionality is complete
spew [Wed, 29 Mar 2017 18:54:50 +0000 (13:54 -0500)]
hjfs: disable hjfs check until more functionality is complete

7 years agohjfs: improve error messaging around blocks that are not found
spew [Wed, 29 Mar 2017 04:04:28 +0000 (23:04 -0500)]
hjfs: improve error messaging around blocks that are not found

7 years agohjfs: Fix bugs in ref count scan check. Enable as a console command (caveat: command...
spew [Wed, 29 Mar 2017 01:39:14 +0000 (20:39 -0500)]
hjfs: Fix bugs in ref count scan check. Enable as a console command (caveat: command arguments will change as I implement more functionality)

7 years agomove blit roms to /sys/lib/blit
cinap_lenrek [Tue, 28 Mar 2017 23:33:16 +0000 (01:33 +0200)]
move blit roms to /sys/lib/blit

7 years agoblit: fix mkfile
cinap_lenrek [Tue, 28 Mar 2017 23:20:13 +0000 (01:20 +0200)]
blit: fix mkfile

7 years agokernel: fix twakeup()/timerdel() race condition
cinap_lenrek [Tue, 28 Mar 2017 22:30:53 +0000 (00:30 +0200)]
kernel: fix twakeup()/timerdel() race condition

timerdel() did not make sure that the timer function
is not active (on another cpu). just acquiering the
Timer lock in the timer function only blocks the caller
of timerdel()/timeradd() but not the other way arround
(on a multiprocessor).

this changes the timer code to track activity of
the timer function, having timerdel() wait until
the timer has finished executing.

7 years agosdnvme: NVMe controller driver (work in progress)
cinap_lenrek [Tue, 28 Mar 2017 22:21:35 +0000 (00:21 +0200)]
sdnvme: NVMe controller driver (work in progress)

basic NVMe controller driver, reads and writes work.
"namespaces" show up as logical units.
uses pin/msi interrupts (no msi-x support yet).
one submission queue per cpu, shared completion queue.
no recovery from fatal controller errors.
only tested in qemu (no hardware available).

commiting this so it can be found by someone who has
hardware.

7 years agogames/blit: mkfile: install into /bin/games
aiju [Tue, 28 Mar 2017 19:48:50 +0000 (19:48 +0000)]
games/blit: mkfile: install into /bin/games

7 years agoadd games/blit
aiju [Tue, 28 Mar 2017 19:43:09 +0000 (19:43 +0000)]
add games/blit

7 years agohjfs: fix broken dprint
spew [Tue, 28 Mar 2017 03:05:53 +0000 (22:05 -0500)]
hjfs: fix broken dprint

7 years agohjfs: simplify dprinting
spew [Tue, 28 Mar 2017 02:38:12 +0000 (21:38 -0500)]
hjfs: simplify dprinting

7 years agohjfs: add simple scan check of directory entry blocks
spew [Tue, 28 Mar 2017 02:27:45 +0000 (21:27 -0500)]
hjfs: add simple scan check of directory entry blocks

7 years agohjfs: check: check a block if its ref count is _not_ zero. Also check all the ref...
spew [Mon, 27 Mar 2017 22:57:10 +0000 (17:57 -0500)]
hjfs: check: check a block if its ref count is _not_ zero. Also check all the ref counts of blocks of a directory and clean up messages

7 years agovarious fortune corpuses: Nil means nothing.
stanley lieber [Mon, 27 Mar 2017 17:33:09 +0000 (13:33 -0400)]
various fortune corpuses: Nil means nothing.

7 years agohjfs: merge start of hjfs check implementation
spew [Mon, 27 Mar 2017 14:57:08 +0000 (09:57 -0500)]
hjfs: merge start of hjfs check implementation

7 years agohjfs: start implementation of checking a directory
spew [Mon, 27 Mar 2017 14:55:56 +0000 (09:55 -0500)]
hjfs: start implementation of checking a directory

7 years agodevsd: handle case where theres no ifc->enable() function
cinap_lenrek [Sun, 26 Mar 2017 14:53:19 +0000 (16:53 +0200)]
devsd: handle case where theres no ifc->enable() function

7 years agomerge
cinap_lenrek [Sun, 26 Mar 2017 14:46:58 +0000 (16:46 +0200)]
merge

7 years agodevsd: check return value of ifc->enable(), don't leak unit name/user strings
cinap_lenrek [Sun, 26 Mar 2017 14:45:34 +0000 (16:45 +0200)]
devsd: check return value of ifc->enable(), don't leak unit name/user strings

7 years agosdvirtio: return 1 for success in vioenable()/viodisable()
cinap_lenrek [Sun, 26 Mar 2017 14:43:32 +0000 (16:43 +0200)]
sdvirtio: return 1 for success in vioenable()/viodisable()

7 years agofile(1): recognise Xilinx bitstreams
aiju [Sat, 25 Mar 2017 22:29:29 +0000 (22:29 +0000)]
file(1): recognise Xilinx bitstreams

7 years agogames/galaxy: fix exit race condition by pausing the galaxy before threadexitsall
spew [Sat, 25 Mar 2017 18:51:33 +0000 (13:51 -0500)]
games/galaxy: fix exit race condition by pausing the galaxy before threadexitsall

7 years agohjfs: Add comment to change the OFF size to 8 when given the chance
spew [Sat, 25 Mar 2017 18:49:13 +0000 (13:49 -0500)]
hjfs: Add comment to change the OFF size to 8 when given the chance

7 years agogames/galaxy: parallelize gravitational force calculations
spew [Sat, 25 Mar 2017 18:05:47 +0000 (13:05 -0500)]
games/galaxy: parallelize gravitational force calculations

Once the Barnes-Hut tree is constructed, the gravitational
force calculations can be done in parallel by dividing the
bodies up between a number of procs.

7 years agopc kernel: handle PCMP and RSD being in low (kaddr) or reserved (vmap) memory
cinap_lenrek [Sat, 25 Mar 2017 03:08:14 +0000 (04:08 +0100)]
pc kernel: handle PCMP and RSD being in low (kaddr) or reserved (vmap) memory

on thinkpad x1v4, the PCMP structure resides in upper reserved memory
pa=0xd7f49000 - while system memory ends at 0x0ffff000; so we have to
vmap() it instead of KADDR().

the RSD structure for ACPI might reside in low memory, so we sould
KADDR() in that case.

7 years agoupas/smtp: fix cram-md5 auth, simplify doauth(), check varargs for dBprint()
cinap_lenrek [Thu, 23 Mar 2017 00:31:35 +0000 (01:31 +0100)]
upas/smtp: fix cram-md5 auth, simplify doauth(), check varargs for dBprint()

- smtpcram() was replying with the challenge instead of the response... m(
- simplify doauth, use error string for auth_getuserpasswd() error logging
- enable #pragma varargck for dBprint()

7 years agoupas/smtp: generate 128-bit random message id (was 32 bit), use dBprint(), cleanup...
cinap_lenrek [Wed, 22 Mar 2017 23:36:57 +0000 (00:36 +0100)]
upas/smtp: generate 128-bit random message id (was 32 bit), use dBprint(), cleanup unused variables

7 years agoplumb(2): Fix typo in description of Plumbdelattr (thanks sam-d)
spew [Wed, 22 Mar 2017 23:13:51 +0000 (18:13 -0500)]
plumb(2): Fix typo in description of Plumbdelattr (thanks sam-d)

7 years agowebfs(4): timeout is in milliseconds not seconds (thanks sam-d)
spew [Wed, 22 Mar 2017 23:11:23 +0000 (18:11 -0500)]
webfs(4): timeout is in milliseconds not seconds (thanks sam-d)

7 years agoupas/smtp: Revert smtp dial string behavior to match that of old upas (thanks sam-d)
spew [Wed, 22 Mar 2017 22:10:35 +0000 (17:10 -0500)]
upas/smtp: Revert smtp dial string behavior to match that of old upas (thanks sam-d)
Also do some cleanup around buffers and memory management

7 years ago[012568kqv]a: correctly lex full range of integers in the assemblers (thanks Ori_B)
spew [Wed, 22 Mar 2017 05:04:24 +0000 (00:04 -0500)]
[012568kqv]a: correctly lex full range of integers in the assemblers (thanks Ori_B)
The Plan 9 assemblers use strtoll to parse the integer literals
in their input. It turns out that this is almost correct, but
VLONG_MIN is clamped. This patch changes to use strtoull
in order to allow the full range of integers.

7 years agoremove accidentally commited binary /sys/src/cmd/upas/fs/chkidx file (thanks mischief)
cinap_lenrek [Tue, 21 Mar 2017 19:31:15 +0000 (20:31 +0100)]
remove accidentally commited binary /sys/src/cmd/upas/fs/chkidx file (thanks mischief)

7 years agohjfs: Clear all refs to zero when reaming.
spew [Tue, 21 Mar 2017 19:07:18 +0000 (14:07 -0500)]
hjfs: Clear all refs to zero when reaming.
And a couple clarity/formatting changes

7 years agoip/torrent: exit immidiately when file is complete after verification and not being...
cinap_lenrek [Mon, 20 Mar 2017 18:15:40 +0000 (19:15 +0100)]
ip/torrent: exit immidiately when file is complete after verification and not being in seed mode

7 years agoip/torrent: avoid peerid collision using truerand() instead of time(0)
cinap_lenrek [Mon, 20 Mar 2017 18:11:03 +0000 (19:11 +0100)]
ip/torrent: avoid peerid collision using truerand() instead of time(0)

7 years agofortune: avoid buffer overflow for lines >= 2K, make sure index has at least one...
cinap_lenrek [Sun, 19 Mar 2017 21:04:26 +0000 (22:04 +0100)]
fortune: avoid buffer overflow for lines >= 2K, make sure index has at least one entry, use nrand()/ntruerand() for uniform distribution

7 years ago5l,6l,8l,kl,ql,vl: allow duplicate GLOBAL symbols (from Ori Bernstein)
cinap_lenrek [Sun, 19 Mar 2017 02:05:24 +0000 (03:05 +0100)]
5l,6l,8l,kl,ql,vl: allow duplicate GLOBAL symbols (from Ori Bernstein)

The plan 9 assemblers support the DUPOK flag on text symbols. They parse and
ignore it on GLOBL symbols. This patch makes it work in the linkers.

The reason I ran into this is because my programming language (Myrddin) uses
data symbols to generate type information, and it's useful to avoid
duplicating all of the type info in every file that gets generated.

7 years agoinst: make command window not overlap stats window in glendas riostart
cinap_lenrek [Sat, 18 Mar 2017 18:35:06 +0000 (19:35 +0100)]
inst: make command window not overlap stats window in glendas riostart

7 years agocdproto: make mdir for glenda
cinap_lenrek [Sat, 18 Mar 2017 18:07:29 +0000 (19:07 +0100)]
cdproto: make mdir for glenda

7 years agodiskparts(8), prep(8): add edisk in NAME section
cinap_lenrek [Sat, 18 Mar 2017 17:08:30 +0000 (18:08 +0100)]
diskparts(8), prep(8): add edisk in NAME section

7 years agomus(1): fix manpage references
cinap_lenrek [Sat, 18 Mar 2017 16:56:34 +0000 (17:56 +0100)]
mus(1): fix manpage references

7 years agovga(3): complete list of vga controllers and software cursors
cinap_lenrek [Sat, 18 Mar 2017 16:13:09 +0000 (17:13 +0100)]
vga(3): complete list of vga controllers and software cursors

7 years agodevmouse: refactor screen blanking logic
cinap_lenrek [Sat, 18 Mar 2017 15:58:27 +0000 (16:58 +0100)]
devmouse: refactor screen blanking logic

devmouse controls the screen blanking timeout, so move the
code there avoiding cross calls between modules. the only
function that needs to be provided is blankscreen(), which
gets called with drawlock locked.

the blank timeout is set thru /dev/mousectl now, so kernels
without devvga can set it.

blanking now only happens while /dev/mouse is read. so this
avoids accidentally blanking the screen on cpu servers that
do not have a mouse to unblank it.

7 years agogames/doom: don't spawn midi process when opening /dev/audio failed (thanks qwx)
cinap_lenrek [Fri, 17 Mar 2017 00:02:01 +0000 (01:02 +0100)]
games/doom: don't spawn midi process when opening /dev/audio failed (thanks qwx)

7 years agotapefs: handle more cpio formats (thanks qrstuv)
cinap_lenrek [Thu, 16 Mar 2017 23:45:48 +0000 (00:45 +0100)]
tapefs: handle more cpio formats (thanks qrstuv)