]> git.lizzy.rs Git - plan9front.git/log
plan9front.git
9 years agomerge
cinap_lenrek [Fri, 12 Dec 2014 20:07:44 +0000 (21:07 +0100)]
merge

9 years agoacme/win: fix fswrite() not assuming iounit < EVENTSIZE
cinap_lenrek [Fri, 12 Dec 2014 20:06:16 +0000 (21:06 +0100)]
acme/win: fix fswrite() not assuming iounit < EVENTSIZE

9 years agofortunes: UX-wise (and generally speaking, for most common uses of a computer these...
stanley lieber [Thu, 11 Dec 2014 19:45:51 +0000 (14:45 -0500)]
fortunes: UX-wise (and generally speaking, for most common uses of a computer these days), Plan9 is, sadly, almost useless.

9 years agoqmail: call mail instead of smtp with the wrong path
stanley lieber [Thu, 11 Dec 2014 19:45:16 +0000 (14:45 -0500)]
qmail: call mail instead of smtp with the wrong path

9 years agocfs: change default mountpoint to /mnt/cfs and fix the manual (thanks mischief)
cinap_lenrek [Thu, 11 Dec 2014 19:38:04 +0000 (20:38 +0100)]
cfs: change default mountpoint to /mnt/cfs and fix the manual (thanks mischief)

9 years agoacme: fix buffer overrun in xfidutfread() and xfidruneread(), cleanup
cinap_lenrek [Thu, 11 Dec 2014 19:20:47 +0000 (20:20 +0100)]
acme: fix buffer overrun in xfidutfread() and xfidruneread(), cleanup

the utf8 buffers b1 where allocated from fbufalloc() which gives
us BUFSIZE bytes, but Xfid->count can be bigger than that. so just
emalloc() the requested number of bytes.

when converting from Runes to utf-8, we have to account for the
terminating '\0' byte snprint() places, so fix the maxrune number
calculation instead of using BUFSIZE+1 as buffer size.

9 years agokbmap: fix sprint() buffer overflow (thanks silasm)
cinap_lenrek [Thu, 11 Dec 2014 17:32:50 +0000 (18:32 +0100)]
kbmap: fix sprint() buffer overflow (thanks silasm)

A buffer can be overflowed in the init function of kbmap.c by using a filename of more than 112 characters.

sample output:
% cd /sys/lib/kbmap
% touch aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa
% kbmap
kbmap 1974: suicide: sys: trap: fault write addr=0xa6a96510 pc=0x000011df
offending code is most likely the call to sprint in the init function of /sys/src/cmd/kbmap.c,
which in this case writes /sys/lib/kbmap/$file to a 128-bit buffer.
I'm willing to submit a patch for this myself along with a few minor improvements/fixes to kbmap
if I can figure out the nuances of doing so.

--silasm

9 years agoplan9.ini(8): fix typo (thanks ftrvxmtrx)
cinap_lenrek [Thu, 11 Dec 2014 16:03:01 +0000 (17:03 +0100)]
plan9.ini(8): fix typo (thanks ftrvxmtrx)

9 years agoplan9.ini(8): etheryuk
cinap_lenrek [Wed, 10 Dec 2014 18:38:44 +0000 (19:38 +0100)]
plan9.ini(8): etheryuk

9 years agoplan9.ini(8): document bcm and virtio ethernet and hda and ac97 audio.
cinap_lenrek [Wed, 10 Dec 2014 18:29:46 +0000 (19:29 +0100)]
plan9.ini(8): document bcm and virtio ethernet and hda and ac97 audio.

9 years agoplan9.ini(8): remove factotumopts= (not implemented), add secstore=, clarify domain...
cinap_lenrek [Wed, 10 Dec 2014 17:34:08 +0000 (18:34 +0100)]
plan9.ini(8): remove factotumopts= (not implemented), add secstore=, clarify domain name use for fs= and auth=

9 years agobootrc: keep original address when ndb/dnsgetip fails and let dial complain
cinap_lenrek [Wed, 10 Dec 2014 17:02:41 +0000 (18:02 +0100)]
bootrc: keep original address when ndb/dnsgetip fails and let dial complain

9 years agobootrc: use "address" instead "ip" in wording netboot questions, as we domain names...
cinap_lenrek [Wed, 10 Dec 2014 16:37:53 +0000 (17:37 +0100)]
bootrc: use "address" instead "ip" in wording netboot questions, as we domain names are allowed now.

9 years agomerge
cinap_lenrek [Wed, 10 Dec 2014 02:24:21 +0000 (03:24 +0100)]
merge

9 years agobootfs: remove disk/kfs fileserver, nobody uses it
cinap_lenrek [Wed, 10 Dec 2014 02:22:59 +0000 (03:22 +0100)]
bootfs: remove disk/kfs fileserver, nobody uses it

9 years agobootrc: add ndb/dnsgetip resolver to bootfs so domain names can be used for fs=,...
cinap_lenrek [Wed, 10 Dec 2014 02:22:14 +0000 (03:22 +0100)]
bootrc: add ndb/dnsgetip resolver to bootfs so domain names can be used for fs=, auth= and secstore= (thanks mischief)

9 years agobootrc: preserve #ec/fs (thanks _sl)
cinap_lenrek [Wed, 10 Dec 2014 02:16:36 +0000 (03:16 +0100)]
bootrc: preserve #ec/fs (thanks _sl)

9 years agondb(8): document ndb/dnsgetip
mischief [Wed, 10 Dec 2014 01:33:09 +0000 (17:33 -0800)]
ndb(8): document ndb/dnsgetip

9 years agodnsgetip: filter negative answers, remove -f flag, use dnsgetip logfile
cinap_lenrek [Wed, 10 Dec 2014 01:30:25 +0000 (02:30 +0100)]
dnsgetip: filter negative answers, remove -f flag, use dnsgetip logfile

9 years agondb: add dnsgetip program to resolve A and AAAA records during bootstrapping
mischief [Wed, 10 Dec 2014 00:18:08 +0000 (16:18 -0800)]
ndb: add dnsgetip program to resolve A and AAAA records during bootstrapping

9 years agobootrc: handle multiple fs and auth addresses in netboot, always set auth= in /net/ndb
cinap_lenrek [Tue, 9 Dec 2014 23:30:07 +0000 (00:30 +0100)]
bootrc: handle multiple fs and auth addresses in netboot, always set auth= in /net/ndb

9 years agobootrc: do not pass authserver with -a flag to factotum, handle multiple secstore...
cinap_lenrek [Tue, 9 Dec 2014 21:26:30 +0000 (22:26 +0100)]
bootrc: do not pass authserver with -a flag to factotum, handle multiple secstore addresses

factotum is started in bootrc before the network is setup
(as we need it to negotiate wpa key for wifi). once, the network
is setup, the bootstrap authservers are passed in /net/ndb,
which factotum reads when doing bootstrap authdial. it does
this only when no authserver was specified earlier! but we
want net.rc to select the proper bootstrap authserver...

the $secstore variable takes precedence over $auth. as
there is no connection server yet, we have to select the set
of servers here and pass them to secstore with -s flag.
note that this will work if multiple addresses where
specified.

9 years agofactotum: accept multiple bootstrap auth servers in /net/ndb and -a arguments
cinap_lenrek [Tue, 9 Dec 2014 21:07:37 +0000 (22:07 +0100)]
factotum: accept multiple bootstrap auth servers in /net/ndb and -a arguments

we might have to deal with multiple bootstrap auth server
ip addresses (ipv4 and ipv6) in the future, so deal with them.

9 years agosecstore: allow multiple server addresses in $secstore and multiple -s arguments
cinap_lenrek [Tue, 9 Dec 2014 20:53:07 +0000 (21:53 +0100)]
secstore: allow multiple server addresses in $secstore and multiple -s arguments

on bootstrap, we might have to deal with multiple auth server ip
addresses (ipv4 and ipv6), so deal with it.

9 years agodevsegment: fix segmentcreate function signature
mischief [Tue, 9 Dec 2014 07:16:22 +0000 (23:16 -0800)]
devsegment: fix segmentcreate function signature

9 years agomerge
cinap_lenrek [Tue, 9 Dec 2014 02:25:52 +0000 (03:25 +0100)]
merge

9 years agoethervirtio: fix queue notifications and interrupt flags, avoid useless notifications
cinap_lenrek [Tue, 9 Dec 2014 02:23:53 +0000 (03:23 +0100)]
ethervirtio: fix queue notifications and interrupt flags, avoid useless notifications

bug: Rnointerrupt was used on Vqueue.used.flags instead of
Vqueue.avail.flags.

introduce vqnotify() function that notifies the device
about available ring advancement.

avoid queue notifications there that can be slow by
checking Unonotify flag in Vqueue.used.flags.

keep track of the number of notifications in the queue.

9 years agohgwebfs: add fake add_handler method for HTTPRangeHandler()
cinap_lenrek [Tue, 9 Dec 2014 22:11:05 +0000 (23:11 +0100)]
hgwebfs: add fake add_handler method for HTTPRangeHandler()

provide a add_handler() so code trying to add
byterange.HTTPRangeHandler will succeed. webfs
does handle "206 Partial Content" already.

9 years agoip/ipconfig: treat /32 mask as /0
mischief [Tue, 9 Dec 2014 01:27:40 +0000 (17:27 -0800)]
ip/ipconfig: treat /32 mask as /0

some dhcp servers send a mask of 255.255.255.255 to indicate the gateway is directly routeable.

thanks to david du colombier for this patch.

9 years agoethervirtio: feature negotiation, allow setting mac address, dont read isr from ifsta...
cinap_lenrek [Mon, 8 Dec 2014 18:19:53 +0000 (19:19 +0100)]
ethervirtio: feature negotiation, allow setting mac address, dont read isr from ifstatus, avoid indirection

- properly negotiate Fctrlrx feature bit for promisc and multicast.
- allow setting mac address with ea= option from plan9.ini
- dont read the isr register from ifstats() as it has the side effect of reseting isr status
- embedd the Vqueue array in the Ctlr structure avoiding indirection
- add a interrupt counter Vqueue.nintr for statistical purposes
- only read network status register if the feature has been negotiated
- change name to "virtio" as "ethervirtio" is kind of redundant

9 years agondb/dns: ignore terminating authoritative flag for no-answer when more nameservers...
cinap_lenrek [Mon, 8 Dec 2014 00:26:07 +0000 (01:26 +0100)]
ndb/dns: ignore terminating authoritative flag for no-answer when more nameservers are provided

continue recursing when we get empty but non-negative answer
from a (claimed) authoritative nameserer that provides more
nameservers.

this fixes wordpress dns:

63766.3: sending to 192.0.80.93/ns1.wordpress.com bossypally.files.wordpress.com ip
63766: rcvd OK from 192.0.80.93 (authoritative)
Q:    bossypally.files.wordpress.com ip
Auth: files.wordpress.com              5 min           ns    mdns1.wordpress.com
      files.wordpress.com              5 min           ns    mdns2.wordpress.com
      files.wordpress.com              5 min           ns    mdns3.wordpress.com
      files.wordpress.com              5 min           ns    mdns4.wordpress.com
      files.wordpress.com              5 min           ns    mdns5.wordpress.com
Hint: mdns1.wordpress.com              4 hr            ip    192.0.75.7
      mdns2.wordpress.com              4 hr            ip    198.181.117.7
      mdns3.wordpress.com              4 hr            ip    198.181.116.7
      mdns4.wordpress.com              4 hr            ip    198.181.118.7
      mdns5.wordpress.com              4 hr            ip    192.0.74.7

63766.4: sending to 192.0.75.7/mdns1.wordpress.com bossypally.files.wordpress.com ip
63766: rcvd OK from 192.0.75.7 (authoritative)
Q:    bossypally.files.wordpress.com ip
Ans:  bossypally.files.wordpress.com   5 min           ip    192.0.72.2
      bossypally.files.wordpress.com   5 min           ip    192.0.72.3
----------------------------
answer bossypally.files.wordpress.com   5 min           ip    192.0.72.2
answer bossypally.files.wordpress.com   5 min           ip    192.0.72.3
----------------------------

note the authoritative flag in the first response from ns1.wordpress.com that
would otherwise terminate the search.

9 years agoethervirtio: fix missing */
cinap_lenrek [Sun, 7 Dec 2014 17:20:47 +0000 (18:20 +0100)]
ethervirtio: fix missing */

9 years agoethervirtio: implement promisc and multicast mode, cleanup
cinap_lenrek [Sun, 7 Dec 2014 16:58:51 +0000 (17:58 +0100)]
ethervirtio: implement promisc and multicast mode, cleanup

add vctlcmd() function to setup and comlete control commands.
handle Vctlq and implement promiscuous and multicast mode commands.
remove Vqueue.block[] and Vqueue.header. these are not properties
of the queue (Vctlq as no block array).

the block[] array only needs to be half the queue size as we use
two descriptors per packet.

fix broken shutdown() and remove useless ctl() function.

9 years agoethervirtio: avoid dumping packets in txproc when ring gets full
cinap_lenrek [Sun, 7 Dec 2014 14:34:53 +0000 (15:34 +0100)]
ethervirtio: avoid dumping packets in txproc when ring gets full

9 years ago9boot(8): -from
cinap_lenrek [Sun, 7 Dec 2014 13:34:15 +0000 (14:34 +0100)]
9boot(8): -from

9 years agostrcat(2): fix strchr() and strrchr() function prototypes (thanks BurnZeZ)
cinap_lenrek [Sun, 7 Dec 2014 13:26:41 +0000 (14:26 +0100)]
strcat(2): fix strchr() and strrchr() function prototypes (thanks BurnZeZ)

9 years agopc, pc64: add and enable ethervirtio driver
mischief [Sun, 7 Dec 2014 06:13:44 +0000 (22:13 -0800)]
pc, pc64: add and enable ethervirtio driver

many thanks to cinap_lenrek, erik quanstrom and david du colombier for feedback and testing.

9 years agoresize: fix typo
mischief [Thu, 27 Nov 2014 03:32:16 +0000 (19:32 -0800)]
resize: fix typo

9 years agoastro: fix typo
mischief [Thu, 27 Nov 2014 02:56:00 +0000 (18:56 -0800)]
astro: fix typo

9 years agogames/md: fix interlace at scale 2
cinap_lenrek [Sat, 22 Nov 2014 19:35:28 +0000 (20:35 +0100)]
games/md: fix interlace at scale 2

9 years agogames/md: interlacing, sort of
aiju [Sat, 22 Nov 2014 18:40:36 +0000 (19:40 +0100)]
games/md: interlacing, sort of

9 years agoaudio(1): add mixfs to NAME section
cinap_lenrek [Sat, 22 Nov 2014 17:17:06 +0000 (18:17 +0100)]
audio(1): add mixfs to NAME section

9 years agopc, pc64: more conservative pcirouting
cinap_lenrek [Sat, 22 Nov 2014 16:23:46 +0000 (17:23 +0100)]
pc, pc64: more conservative pcirouting

when we hit a conflict where the pci INTL register gives us
a different irq than we get from southbridge irq router, dont
just change the router setting to the bios assigned irq (that
was previously known as the BIOS workarround), but assume the
southbridge setting to be valid and change the pci INTL register
on the device to it.

only when the router link doesnt seem to be configured or
disabled, then program the router to the irq that bios asisgned
in the INTL register.

the reason is that changing the router setting changes the
irq routing for *all* devices on the same link and changing
it breaks previously checked and valid interrupt routings.
(so happend with virtualbox where the last device on the bus
is some powermanagement device that has wrong INTL setting
and changing the routing breaks the ethernet interrupts)

this change shouldnt affect modern machines which use ioapic
and mp tables or acpi for pci interrupt routing.

9 years agopc, pc64: preserve last KB of conventional memory (might contain bios tables)
cinap_lenrek [Thu, 20 Nov 2014 18:05:43 +0000 (19:05 +0100)]
pc, pc64: preserve last KB of conventional memory (might contain bios tables)

we add new function convmemsize() that returns the size of
*usable* conventional memory that does some sanity checking
and reserves the last KB below the top of memory pointer.

this avoids lowraminit() overriding potential bios tables
and sigsearch() going off the rails looking for tables
at above 640K.

9 years agogames/snes: do message display in screenproc
cinap_lenrek [Thu, 20 Nov 2014 00:26:31 +0000 (01:26 +0100)]
games/snes: do message display in screenproc

do the message drawing from the screenproc so there are no
race conditions with libdraw.

9 years agogames/md, games/snes: fix memory leak on resize
cinap_lenrek [Wed, 19 Nov 2014 20:01:36 +0000 (21:01 +0100)]
games/md, games/snes: fix memory leak on resize

9 years agohjfs(8): theres no chown in plan9, link to chgrp(1) manpage.
cinap_lenrek [Wed, 19 Nov 2014 00:31:20 +0000 (01:31 +0100)]
hjfs(8): theres no chown in plan9, link to chgrp(1) manpage.

9 years agogames/snes: fix extern pic declaration to match reality
cinap_lenrek [Tue, 18 Nov 2014 16:03:45 +0000 (17:03 +0100)]
games/snes: fix extern pic declaration to match reality

9 years agogames/md: apply same screen flush changes from games/snes
cinap_lenrek [Tue, 18 Nov 2014 16:00:44 +0000 (17:00 +0100)]
games/md: apply same screen flush changes from games/snes

9 years agogames/snes: flush screen in parallel to audio (fixes buffer underruns on x200s)
cinap_lenrek [Tue, 18 Nov 2014 15:07:34 +0000 (16:07 +0100)]
games/snes: flush screen in parallel to audio (fixes buffer underruns on x200s)

the x200s is too slow on a single core to keep up without
audio buffer underruns, so the idea is to flush screen
in parallel to witing audio samples in a separate process.

with the proc, we also can keep updating the screen on resize
when paused.

9 years agogames/snes: save more audio state
aiju [Mon, 17 Nov 2014 16:00:41 +0000 (17:00 +0100)]
games/snes: save more audio state

9 years agoaudiohda: add pci ids for Intel 8 Series Lynx Point (thanks quux)
cinap_lenrek [Mon, 17 Nov 2014 09:12:29 +0000 (10:12 +0100)]
audiohda: add pci ids for Intel 8 Series Lynx Point (thanks quux)

Based on OpenBSD driver:
- /sys/dev/pci/azalia.c rev 1.209
- /sys/dev/pci/pcidevs rev 1.1689
- only tested on amd64; machine is an Acer V5-573G
    exact model: V5-573G-74518G1Takk

9 years agowpa: handle essid: with whitespaces from ether ifstats file
cinap_lenrek [Sun, 16 Nov 2014 23:53:02 +0000 (00:53 +0100)]
wpa: handle essid: with whitespaces from ether ifstats file

9 years agowpa: fix quoting for essid with spaces
cinap_lenrek [Sun, 16 Nov 2014 23:12:55 +0000 (00:12 +0100)]
wpa: fix quoting for essid with spaces

9 years agoabaco: cleanup, handle image/x-icon, don't use backspace as a hotkey, and remove...
BurnZeZ [Sun, 16 Nov 2014 14:58:21 +0000 (09:58 -0500)]
abaco: cleanup, handle image/x-icon, don't use backspace as a hotkey, and remove urlconvience()/related functions

9 years agonintendo(1): NTSC only
aiju [Sun, 16 Nov 2014 14:17:34 +0000 (15:17 +0100)]
nintendo(1): NTSC only

9 years agohpost: break up tags with newlines to avoid sed line buffer limit
cinap_lenrek [Sun, 16 Nov 2014 00:59:29 +0000 (01:59 +0100)]
hpost: break up tags with newlines to avoid sed line buffer limit

9 years agosed: fix error handling for "Output line too long" and other stuff
cinap_lenrek [Sun, 16 Nov 2014 00:48:18 +0000 (01:48 +0100)]
sed: fix error handling for "Output line too long" and other stuff

sed just continued writing past genbuf when it should stop
with "Output line too long".

quit when we get unspecified options.

stupid casts from long to char* for no reason.

some 0 vs nil cleanup.

9 years agopc: get rid of fixed 8MB memory map (now dynamically between 4 to 16 MB depending...
cinap_lenrek [Sat, 15 Nov 2014 14:21:24 +0000 (15:21 +0100)]
pc: get rid of fixed 8MB memory map (now dynamically between 4 to 16 MB depending on kernel size)

we now do mapping of KZERO to ROUND(end, 4*MB) where
end needs not to be above 16MB. this allows for bigger
kernels.

9 years agopc64: get rid of fixed 8MB INIMAP and dynamically map KZERO-end in l.s
cinap_lenrek [Sat, 15 Nov 2014 10:43:05 +0000 (11:43 +0100)]
pc64: get rid of fixed 8MB INIMAP and dynamically map KZERO-end in l.s

traditionally, the pc kernel mapped the first 8MB of physical
address space. when the kernel size grows beyond that memory mapping,
it will crash on boot and theres no checking in the build process
making sure it fits.

with the pc64 kernel, it is not hard to always map the whole
kernel memory image from KZERO to end[], so that the kernel will
always fit into the initial mapping.

9 years agohpost: fix multipart/form variable name mangeling when field value contains newlines
cinap_lenrek [Sat, 15 Nov 2014 22:51:01 +0000 (23:51 +0100)]
hpost: fix multipart/form variable name mangeling when field value contains newlines

com='MH3U
pic unrelated
' hpost -u ... -m post com:$com

resulted in wrong field name like:

--HJBOUNDARY
 Content-Disposition: form-data; name="com pic unrelated"

 MH3U
 pic unrelated

9 years agofortunes: (no plan9 to hand)
stanley lieber [Fri, 14 Nov 2014 17:11:42 +0000 (12:11 -0500)]
fortunes: (no plan9 to hand)

9 years agoupas/fs: handle ramfs and hjfs "file locked" error
cinap_lenrek [Fri, 14 Nov 2014 09:09:29 +0000 (10:09 +0100)]
upas/fs: handle ramfs and hjfs "file locked" error

9 years agoupas/fs: handle cwfs "file is locked" error message
cinap_lenrek [Fri, 14 Nov 2014 22:36:18 +0000 (23:36 +0100)]
upas/fs: handle cwfs "file is locked" error message

9 years agoip: implement "hangup" ctl for udp protocol
cinap_lenrek [Thu, 13 Nov 2014 15:47:19 +0000 (16:47 +0100)]
ip: implement "hangup" ctl for udp protocol

9 years agokernel: make use of nil vs 0 consistent in qio.c (sorry)
cinap_lenrek [Thu, 13 Nov 2014 15:46:41 +0000 (16:46 +0100)]
kernel: make use of nil vs 0 consistent in qio.c (sorry)

9 years agorootstub: create /sys/src/pkg and /sys/lib/pkg directories
cinap_lenrek [Thu, 13 Nov 2014 09:23:53 +0000 (10:23 +0100)]
rootstub: create /sys/src/pkg and /sys/lib/pkg directories

9 years agohttpfile(4): document webfs dependency
cinap_lenrek [Wed, 12 Nov 2014 11:33:14 +0000 (12:33 +0100)]
httpfile(4): document webfs dependency

9 years agohttpfile: use webfs, fix 9p flushes
cinap_lenrek [Wed, 12 Nov 2014 11:03:51 +0000 (12:03 +0100)]
httpfile: use webfs, fix 9p flushes

we can improve performance alot by using webfs which
does http keep alives for us, so connection setup
overhead is eleminated.

fix 9p flushes and double frees.

9 years agonintendo(1): more detail about gba
stanley lieber [Tue, 11 Nov 2014 19:57:45 +0000 (14:57 -0500)]
nintendo(1): more detail about gba

9 years agonintendo(1): document games/gba
stanley lieber [Mon, 10 Nov 2014 06:05:18 +0000 (01:05 -0500)]
nintendo(1): document games/gba

9 years agopc64: fix comment description of pc64 kernel configuration
cinap_lenrek [Mon, 10 Nov 2014 11:22:50 +0000 (12:22 +0100)]
pc64: fix comment description of pc64 kernel configuration

9 years agoether8169: add support for RTL8411B (thanks quux)
cinap_lenrek [Mon, 10 Nov 2014 11:13:07 +0000 (12:13 +0100)]
ether8169: add support for RTL8411B (thanks quux)

from quux 9fans post:

- based on the following changes in FreeBSD/OpenBSD:
http://svnweb.freebsd.org/base?view=revision&revision=257305
http://cvsweb.openbsd.org/cgi-bin/cvsweb/src/sys/dev/ic/re.c.diff?r1=1.144&r2=1.145&f=h
http://cvsweb.openbsd.org/cgi-bin/cvsweb/src/sys/dev/ic/rtl81x9reg.h.diff?r1=1.76&r2=1.77&f=h
http://cvsweb.openbsd.org/cgi-bin/cvsweb/src/sys/dev/ic/rtl81x9reg.h.diff?r1=1.83&r2=1.84&f=h
- sysinfo _with_ patch, on amd64 (please disregard non-working audio and iwl stuff):
http://sysinfo.9front.org/src/86/body
- tested only on amd64; machine is an Acer V5-573G (exact model: V5-573G-74518G1Takk)

9 years agopc, pc64: command flag sometimes doesnt reset, just check for IBF
cinap_lenrek [Sun, 9 Nov 2014 23:53:22 +0000 (00:53 +0100)]
pc, pc64: command flag sometimes doesnt reset, just check for IBF

9 years agoio: add -E flag to access embedded controller space
cinap_lenrek [Sun, 9 Nov 2014 23:35:07 +0000 (00:35 +0100)]
io: add -E flag to access embedded controller space

9 years agomerge
cinap_lenrek [Sun, 9 Nov 2014 23:06:01 +0000 (00:06 +0100)]
merge

9 years agopc, pc64: provide access to embedded controller with #P/ec file
cinap_lenrek [Sun, 9 Nov 2014 23:04:37 +0000 (00:04 +0100)]
pc, pc64: provide access to embedded controller with #P/ec file

9 years agofilter(1): fix typo i introduced
mischief [Sun, 9 Nov 2014 21:41:06 +0000 (13:41 -0800)]
filter(1): fix typo i introduced

this is a good example of why blindly copy-pasting patches is a bad idea.

9 years agokernel: remove implicit Proc* argument from procctl()
cinap_lenrek [Sun, 9 Nov 2014 07:19:28 +0000 (08:19 +0100)]
kernel: remove implicit Proc* argument from procctl()

procctl() is always called with up and it would not
work correctly if passed a different process, so
remove the Proc* argument and use up directly.

9 years agodont flush screen when hiding software cursor
cinap_lenrek [Sat, 8 Nov 2014 10:48:38 +0000 (11:48 +0100)]
dont flush screen when hiding software cursor

we can avoid some flickering when removing the software cursor
from the shadow framebuffer by avoiding the flushscreenimage()
call.

once the cursor is redrawn, we flush the combined rect of its
old and new position in one go.

9 years agosmtpd: cleanup hello() domain check code
cinap_lenrek [Sat, 8 Nov 2014 08:00:04 +0000 (09:00 +0100)]
smtpd: cleanup hello() domain check code

theres no point in doing domain checks on literal ip addresses,
so handle that case early.

9 years agomerge
cinap_lenrek [Sat, 8 Nov 2014 04:23:59 +0000 (05:23 +0100)]
merge

9 years agosmtpd: accept literal ipv6 addresses in HELO/EHLO
cinap_lenrek [Sat, 8 Nov 2014 04:22:25 +0000 (05:22 +0100)]
smtpd: accept literal ipv6 addresses in HELO/EHLO

9 years agoall: typo fixes from ray@raylai.com
mischief [Fri, 7 Nov 2014 18:39:50 +0000 (10:39 -0800)]
all: typo fixes from ray@raylai.com

9 years agoDon't quit ircrc on blank line (thanks Ray)
khm [Fri, 7 Nov 2014 17:51:16 +0000 (12:51 -0500)]
Don't quit ircrc on blank line (thanks Ray)

9 years agofix dangerous werrstr() usages
cinap_lenrek [Fri, 7 Nov 2014 11:51:59 +0000 (12:51 +0100)]
fix dangerous werrstr() usages

werrstr() takes a format string as its first argument.
a common error is to pass user controlled string buffers
into werrstr() that might contain format string escapes
causing werrstr() to take bogus arguments from the stack
and crash.

so instead of doing:
werrstr(buf);

we want todo:
werrstr("%s", buf);

or if we have a local ERRMAX sized buffer that we can override:
errstr(buf, sizeof buf);

9 years agolibc: import cleaned up syslog() function from sources
cinap_lenrek [Fri, 7 Nov 2014 07:42:19 +0000 (08:42 +0100)]
libc: import cleaned up syslog() function from sources

this fixes a potential format string problem where the
error string is passed to werrstr() as fmt. also, the
directory comparsion is simplified in this version using
a helper function.

9 years agolibc: improve dial error handling
cinap_lenrek [Fri, 7 Nov 2014 07:10:19 +0000 (08:10 +0100)]
libc: improve dial error handling

when dial is called with a generic dialstring, it will try
/net and /net.alt in sequence. error out if the /net dial
gets interrupted and do not continue dialing /net.alt.

reduce stack usage by using the swaping nature of errstr()
instead of keeping two error string buffers on the stack.

9 years agodevproc: reset p->pdbg under p->debug qlock in procstopwait()
cinap_lenrek [Fri, 7 Nov 2014 04:21:42 +0000 (05:21 +0100)]
devproc: reset p->pdbg under p->debug qlock in procstopwait()

theres a race where procstopwait() is interrupted by a note,
setting p->pdbg to nil *before* acquiering the lock and
and pexit() and procctl() accessing it assuming it doesnt
change under them while they are holding the lock.

9 years agomouse(2): fix atomouse (thanks qrstuv)
cinap_lenrek [Thu, 6 Nov 2014 03:59:07 +0000 (04:59 +0100)]
mouse(2): fix atomouse (thanks qrstuv)

9 years agolibdraw: fix atomouse
cinap_lenrek [Thu, 6 Nov 2014 03:42:20 +0000 (04:42 +0100)]
libdraw: fix atomouse

9 years agolibc: allow dial to be interrupted
mischief [Thu, 6 Nov 2014 01:24:55 +0000 (17:24 -0800)]
libc: allow dial to be interrupted

previously, if dial was interrupted by an alarm or other note while connecting to a host that resolved to multiple ips, dial would ignore the interruption and try the next host. now dial properly returns with error when it is interrupted.

9 years agoinst: copy efi bootloader in bootsetup
cinap_lenrek [Wed, 5 Nov 2014 20:38:28 +0000 (21:38 +0100)]
inst: copy efi bootloader in bootsetup

9 years agofix some manpage cross references
cinap_lenrek [Wed, 5 Nov 2014 20:00:23 +0000 (21:00 +0100)]
fix some manpage cross references

9 years agotermrc: setup mouse when we have a framebuffer
cinap_lenrek [Wed, 5 Nov 2014 19:42:47 +0000 (20:42 +0100)]
termrc: setup mouse when we have a framebuffer

previously, we setup mouse only when vgasize= was specifid in
plan9.ini. with efi systems, the framebuffer is already setup
for us and theres no requirement for going thru aux/vga setup,
but we still want to setup the mouse.

so do the mouseport= check once theres a framebuffer by testing
the existence of '#i/winname' (which fails when thers no
framebuffer).

9 years agoacpi: fix pcibusno() when PCI0._ADR = 0
cinap_lenrek [Tue, 4 Nov 2014 05:42:39 +0000 (06:42 +0100)]
acpi: fix pcibusno() when PCI0._ADR = 0

libaml comresses zero integer as nil, so remove the
nil check. this makes interrupts work in vmware with
efi.

9 years agopc: enable page size extension early in apbootstrap
cinap_lenrek [Tue, 4 Nov 2014 04:52:42 +0000 (05:52 +0100)]
pc: enable page size extension early in apbootstrap

vmware in efi mode brings application processors up
with CR4 = 0 (pse disabled) which makes us page fault
when accessing the ap's pdb which might be in a 4MB
mapping when the boot processor used pse to setup
page tables.

so we unconditionally enable pse in apbootstrap
(and disable pae in case of surprises).

9 years ago9boot(8): remove redundant "the"
cinap_lenrek [Sun, 2 Nov 2014 20:25:37 +0000 (21:25 +0100)]
9boot(8): remove redundant "the"

9 years ago9boot(8): improve efi documentation
cinap_lenrek [Sun, 2 Nov 2014 20:22:03 +0000 (21:22 +0100)]
9boot(8): improve efi documentation

9 years agobullshit: don't pollute the environment
stanley lieber [Sun, 2 Nov 2014 18:16:33 +0000 (13:16 -0500)]
bullshit: don't pollute the environment