]> git.lizzy.rs Git - plan9front.git/log
plan9front.git
9 years ago6l: no need to emit rex.w prefix for MOVBQZX and MOVWQZX
cinap_lenrek [Tue, 17 Feb 2015 21:25:55 +0000 (22:25 +0100)]
6l: no need to emit rex.w prefix for MOVBQZX and MOVWQZX

as with 32 bit operand size, the upper bits 63:32 are
automatically zeroed in 64bit mode. this gives a shoter
instruction encoding.

9 years agonusb/usbd: fix typo in debug prints, dont use %U format as its not installed
cinap_lenrek [Tue, 17 Feb 2015 21:17:30 +0000 (22:17 +0100)]
nusb/usbd: fix typo in debug prints, dont use %U format as its not installed

9 years agovc: eleminate strcpy in multab code
cinap_lenrek [Tue, 17 Feb 2015 21:16:02 +0000 (22:16 +0100)]
vc: eleminate strcpy in multab code

9 years ago[125678kqv][cl]: fix sprint() and strcpy() buffer overflows
cinap_lenrek [Tue, 17 Feb 2015 21:13:35 +0000 (22:13 +0100)]
[125678kqv][cl]: fix sprint() and strcpy() buffer overflows

9 years agotlssrv: produce meaningfull error message when unable to read certificate
cinap_lenrek [Tue, 17 Feb 2015 06:05:54 +0000 (07:05 +0100)]
tlssrv: produce meaningfull error message when unable to read certificate

9 years agotlssrv: fix usage, dont cat
cinap_lenrek [Tue, 17 Feb 2015 05:59:55 +0000 (06:59 +0100)]
tlssrv: fix usage, dont cat

9 years agotlssrv: fix this mess (thanks burnzez for reporting the issue)
cinap_lenrek [Tue, 17 Feb 2015 05:54:19 +0000 (06:54 +0100)]
tlssrv: fix this mess (thanks burnzez for reporting the issue)

tlsServer() closes the passed in fd, in our case fd=1 leaving it
with no std output which got occupied by pipe() filedescriptor
which it then closed after duping... a classic.

delete all this mess. theres no reason to fork() and copy traffic
on a pipe at all as tlsServer() gives us a perfectly valid filedescriptor.
just dup() and exec() and we'r done.

9 years agowc: flags dont need to be uvlong
cinap_lenrek [Sun, 15 Feb 2015 05:26:49 +0000 (06:26 +0100)]
wc: flags dont need to be uvlong

9 years agowc: bring back old wc but extended for 21 bit runes
cinap_lenrek [Sun, 15 Feb 2015 05:22:19 +0000 (06:22 +0100)]
wc: bring back old wc but extended for 21 bit runes

the imported wc from sources is arround 8 times slower
than our old one. it is common to run wc on large log files
to count lines. so i think the implementation complexity
is justified. (just like with grep)

9 years agowc: import new wc from sources
cinap_lenrek [Sat, 14 Feb 2015 15:20:40 +0000 (16:20 +0100)]
wc: import new wc from sources

9 years agousbehci: initial support for usb on zynq, remove uncached.h
cinap_lenrek [Sat, 14 Feb 2015 02:00:31 +0000 (03:00 +0100)]
usbehci: initial support for usb on zynq, remove uncached.h

the following hooks have been added to the ehci Ctlr
structore to handle cache coherency (on arm):

void* (*tdalloc)(ulong,int,ulong);
void* (*dmaalloc)(ulong);
void (*dmafree)(void*);
void (*dmaflush)(int,void*,ulong);

tdalloc() is used to allocate descriptors and the periodic
frame schedule array. on arm, this needs to return uncached
memory. tdalloc()ed memory is never freed.

dmaalloc()/dmafree() is used for io buffers. this can return
cached memory when when hardware maintains cache coherency (pc)
or dmaflush() is provided to flush/invalidate the cache (zynq),
otherwise needs to return uncached memory.

dmaflush() is used to flush/invalidate the cache. the first
argument tells us if we need to flush (non zero) or
invalidate (zero).

uncached.h is gone now. this change makes the handling explicit.

9 years agozynq: do fixed mapping for ocm memory on boot and make kaddr() and paddr() work with it
cinap_lenrek [Sat, 14 Feb 2015 01:44:19 +0000 (02:44 +0100)]
zynq: do fixed mapping for ocm memory on boot and make kaddr() and paddr() work with it

map the whole ocm memory on boot so we can translate physical to
virtual addresses and back for uncached memory using KADDR() and
PADDR().

replace ualloc() with ucalloc() returning virtual address. physical
address can be acquired with PADDR() now.

as ocm is now always mapped, use KADDR() instead of tmpmap() for
mp bootstrap.

9 years agoircrc: dont require nick and user to be the same for auth/userpasswd (thanks spew)
cinap_lenrek [Thu, 12 Feb 2015 17:40:20 +0000 (18:40 +0100)]
ircrc: dont require nick and user to be the same for auth/userpasswd (thanks spew)

9 years agoircrc(1): remove auth/userpasswd complaint from BUGS section
cinap_lenrek [Thu, 12 Feb 2015 17:24:41 +0000 (18:24 +0100)]
ircrc(1): remove auth/userpasswd complaint from BUGS section

9 years agobuild and document auth/userpasswd
cinap_lenrek [Thu, 12 Feb 2015 17:18:34 +0000 (18:18 +0100)]
build and document auth/userpasswd

9 years agolibc: check name in getvent()/putenv()
cinap_lenrek [Wed, 11 Feb 2015 22:54:28 +0000 (23:54 +0100)]
libc: check name in getvent()/putenv()

passing "", "." or ".." as name caused a crash in
getenv() as it would open the directory; then seek()
to determine the file size would fail and return -1.

now checking for these special names and set
error string when its bad.

doing a single read() will not work when /env has a
9p fileserver mounted onto it and the file size is bigger
than the i/o unit. so doing incremental reads until
we get eof.

9 years agoigfx: displayport support for ivy bridge
cinap_lenrek [Mon, 9 Feb 2015 18:17:04 +0000 (19:17 +0100)]
igfx: displayport support for ivy bridge

9 years agomothra: add side scrolling, controlled by left and right buttons on the keyboard...
stanley lieber [Sun, 8 Feb 2015 18:19:47 +0000 (13:19 -0500)]
mothra: add side scrolling, controlled by left and right buttons on the keyboard. if desired, x-scrollbar can be set visible via the visxbar variable in mothra.c. (thanks, jpm_)

9 years agovga/igfx: add 0x2a02 for GM965/GL960/X3100; comment vid/did with name of chipset
stanley lieber [Sat, 7 Feb 2015 06:12:46 +0000 (01:12 -0500)]
vga/igfx: add 0x2a02 for GM965/GL960/X3100; comment vid/did with name of chipset

9 years agovgadb: add vid=0x8086 did=0x2a02 for Intel GM965/GL960/X3100; comment recent igfx...
stanley lieber [Sat, 7 Feb 2015 06:10:35 +0000 (01:10 -0500)]
vgadb: add vid=0x8086 did=0x2a02 for Intel GM965/GL960/X3100; comment recent igfx ctlr definitions with the names of specific chipsets instead of the model numbers of laptops they may ship in

9 years agopc64: enable uartisa in default config
stanley lieber [Sat, 7 Feb 2015 06:01:11 +0000 (01:01 -0500)]
pc64: enable uartisa in default config

9 years agogames/c64: map : to \ key
aiju [Sat, 7 Feb 2015 17:55:03 +0000 (18:55 +0100)]
games/c64: map : to \ key

9 years agogames/c64: fix -2/-3
aiju [Sat, 7 Feb 2015 17:28:55 +0000 (18:28 +0100)]
games/c64: fix -2/-3

9 years agogames/mkfile: add c64
aiju [Sat, 7 Feb 2015 17:12:06 +0000 (18:12 +0100)]
games/mkfile: add c64

9 years agogames/c64: oops
aiju [Sat, 7 Feb 2015 17:11:19 +0000 (18:11 +0100)]
games/c64: oops

9 years agoadded games/c64
aiju [Sat, 7 Feb 2015 17:03:17 +0000 (18:03 +0100)]
added games/c64

9 years agozynq: fix print format warning in screeninit()
cinap_lenrek [Sat, 7 Feb 2015 02:24:07 +0000 (03:24 +0100)]
zynq: fix print format warning in screeninit()

9 years agokernel: make pagereclaim() a bit less stupid
cinap_lenrek [Sat, 7 Feb 2015 02:01:59 +0000 (03:01 +0100)]
kernel: make pagereclaim() a bit less stupid

put recently used pages at the head of ther image hash
chains, and reclaim pages from the tail first.

9 years agokernel: reduce Page structure size by changing Page.cachectl[]
cinap_lenrek [Sat, 7 Feb 2015 01:52:23 +0000 (02:52 +0100)]
kernel: reduce Page structure size by changing Page.cachectl[]

there are no kernels currently that do page coloring,
so the only use of cachectl[] is flushing the icache
(on arm and ppc).

on pc64, cachectl consumes 32 bytes in each page resulting
in over 200 megabytes of overhead for 32gb of ram with 4K
pages.

this change removes cachectl[] and adds txtflush ulong
that is set to ~0 by pio() to instruct putmmu() to flush
the icache.

9 years agofplot: add modulus % operator
cinap_lenrek [Thu, 5 Feb 2015 18:55:46 +0000 (19:55 +0100)]
fplot: add modulus % operator

9 years agolibsec: add TLS_DHE_RSA_WITH_3DES_EDE_CBC_SHA client cipher suit support
cinap_lenrek [Thu, 5 Feb 2015 03:29:28 +0000 (04:29 +0100)]
libsec: add TLS_DHE_RSA_WITH_3DES_EDE_CBC_SHA client cipher suit support

9 years agoaux/vga: remove vbs/vbe from mode, use shs/ehs when sync is ment, prefer detailed...
cinap_lenrek [Thu, 5 Feb 2015 22:08:46 +0000 (23:08 +0100)]
aux/vga: remove vbs/vbe from mode, use shs/ehs when sync is ment, prefer detailed timing in edid

vbs/vbe members in Mode was only used in the vesadb
and cannot be changed from vgadb.

use shs/ehs in drivers when refering to the horizontal
sync pulse. clarify the matter in a comment.

link detailed timing modes at the head of the edid
modelist. these are the modes we'r interested in,
not the ones from vesadb.

9 years agofortunes: The NTP codebase is larger than the SSH codebase.
stanley lieber [Tue, 3 Feb 2015 02:30:07 +0000 (21:30 -0500)]
fortunes: The NTP codebase is larger than the SSH codebase.

9 years agoaux/vga: fix mistale :-)
cinap_lenrek [Mon, 2 Feb 2015 02:05:56 +0000 (03:05 +0100)]
aux/vga: fix mistale :-)

9 years agoaux/vga: use optional edid information to determine mode when vgadb fails
cinap_lenrek [Mon, 2 Feb 2015 01:58:59 +0000 (02:58 +0100)]
aux/vga: use optional edid information to determine mode when vgadb fails

igfx and vesa can determine monitor timing information from ddc
and store the edid info for connected monitors in vga->edid[].
when monitor type cannot be found in vgadb, we consult the edid
information and make a mode based on the edid info.

this avoids having to maintain a vgadb entry for each monitor.

monitor can be set to "[width]x[height]@[freq]Hz" for a specific
edid setting. when not found, a mode is searched based on the
size.

so the following should work:

aux/vga -m 1366x768@60Hz -l 1366x768x32
aux/vga -m auto -l 1366x768x32

9 years agoaux/vga: cleanup vesa code
cinap_lenrek [Sun, 1 Feb 2015 18:28:24 +0000 (19:28 +0100)]
aux/vga: cleanup vesa code

dbvesamode() modified the passed in size string in the process
of option parsing. this is a no-go because the string might be
constant in the read only section. provide cracksize() function
for the parsing and make a static copy.

do the vendor specific monitor detection in vbesnarf() instead
of vbecheck(). vbecheck()'s purpose is to check if vesa bios
service is avialable, not snarf graphics card state.

nvidiascale() was a no-op because it missed the vbecall() at
the end of the function. this means it was never tested so i
add the missing vbecall(), but disable nvidiascale for now
until someone tests this.

keep fancy stuff out of the Vbe structure. it is just there for
making bios calls, not keep state about the graphics card.

9 years agomothra: people do not like it. reverting
cinap_lenrek [Sat, 31 Jan 2015 19:18:45 +0000 (20:18 +0100)]
mothra: people do not like it. reverting

9 years agomothra: wrap long text and images to fit
cinap_lenrek [Sat, 31 Jan 2015 19:01:24 +0000 (20:01 +0100)]
mothra: wrap long text and images to fit

9 years agokernel: increase size of palloc.mem[] user page bank array
cinap_lenrek [Fri, 30 Jan 2015 13:50:28 +0000 (14:50 +0100)]
kernel: increase size of palloc.mem[] user page bank array

we'r hitting the limit of user page banks on some asrock mainboard,
so doubling the size of the array twice to make running out unlikely.

9 years agowifi: fix recvbeacon()
cinap_lenrek [Fri, 30 Jan 2015 12:41:23 +0000 (13:41 +0100)]
wifi: fix recvbeacon()

we used to read beyond the boundaries of the becon because of
the end pointer was offset by the beacon header. this is
also what caused the double entries.

9 years agowifi: ignore truncated tlv in beacon/probes
cinap_lenrek [Fri, 30 Jan 2015 01:00:07 +0000 (02:00 +0100)]
wifi: ignore truncated tlv in beacon/probes

9 years agofactotum(4): fix typo
cinap_lenrek [Thu, 29 Jan 2015 13:11:43 +0000 (14:11 +0100)]
factotum(4): fix typo

9 years agofactotum(4): document mschapv2 and wpapsk protocols
cinap_lenrek [Thu, 29 Jan 2015 12:57:27 +0000 (13:57 +0100)]
factotum(4): document mschapv2 and wpapsk protocols

9 years agotftpfs(8): fix manpage reference dhcp(8) -> dhcpd(8)
cinap_lenrek [Thu, 29 Jan 2015 12:21:12 +0000 (13:21 +0100)]
tftpfs(8): fix manpage reference dhcp(8) -> dhcpd(8)

9 years agoircrc(1): tlsclient(8) -> tlssrv(8)
cinap_lenrek [Thu, 29 Jan 2015 12:17:34 +0000 (13:17 +0100)]
ircrc(1): tlsclient(8) -> tlssrv(8)

9 years ago6in4(8): document ayiya
cinap_lenrek [Thu, 29 Jan 2015 12:16:08 +0000 (13:16 +0100)]
6in4(8): document ayiya

9 years agoircrc(8), ndb(8): fix manpage reference
cinap_lenrek [Thu, 29 Jan 2015 11:36:14 +0000 (12:36 +0100)]
ircrc(8), ndb(8): fix manpage reference

9 years agodevmnt: handle rpc buffer exhaustion on mntflushalloc()
cinap_lenrek [Tue, 27 Jan 2015 21:14:26 +0000 (22:14 +0100)]
devmnt: handle rpc buffer exhaustion on mntflushalloc()

this bug happens when the kernel runs out of mount rpc
buffers when allocating a flush rpc. in this case, mntflushalloc()
will errorjump out of mountio() leaving the currently in
flight rpc in the mount. the caller of mountrpc()/mountio()
frees the rpc thats still queued in the mount leaving
to interesting results.

for the fix, we add a waserror() arround mntflushalloc() and
handle the error case like a mount rpc failure which will
properly dequeue the rpc's in flight.

9 years agowpa: fix password avp padding, recognize wpa1 authentication oui
cinap_lenrek [Tue, 27 Jan 2015 13:09:24 +0000 (14:09 +0100)]
wpa: fix password avp padding, recognize wpa1 authentication oui

we have to padd the data portion in the password avp,
not add padding after the avp.

9 years agowpa: do not forward alert tls records, close connection
cinap_lenrek [Tue, 27 Jan 2015 09:56:56 +0000 (10:56 +0100)]
wpa: do not forward alert tls records, close connection

9 years agopc, efi loaders: erase characters when backspace is pressed
mischief [Mon, 26 Jan 2015 01:21:07 +0000 (17:21 -0800)]
pc, efi loaders: erase characters when backspace is pressed

9 years agowpa: cleanup, pad eap-ttls/pap password, zero passwords and key material after use
cinap_lenrek [Mon, 26 Jan 2015 00:26:42 +0000 (01:26 +0100)]
wpa: cleanup, pad eap-ttls/pap password, zero passwords and key material after use

9 years agowpa: experimental wpa2 enterprise support
cinap_lenrek [Sun, 25 Jan 2015 06:58:20 +0000 (07:58 +0100)]
wpa: experimental wpa2 enterprise support

this adds support for eap-peap/mschapv2 and eap-ttls/pap.

code has only been tested with freeradius and a cheap
access point, not tested with actual eduroam network.

9 years agofactotum: implement proto=mschapv2 client role
cinap_lenrek [Sun, 25 Jan 2015 06:49:50 +0000 (07:49 +0100)]
factotum: implement proto=mschapv2 client role

this is used for wpa2 enterprise peap/mschapv2. server role
is not implemented as that would require changing the
wire format on the auth server.

the naming is unfortunate as we already have proto=mschap2 which
really refers to ntlmv2.

9 years agoauth(2): fix prototype for auth_getinfo() in manual page
cinap_lenrek [Sun, 25 Jan 2015 06:40:52 +0000 (07:40 +0100)]
auth(2): fix prototype for auth_getinfo() in manual page

9 years agopc: dont handle pending floating point exception in fpsave thru FPOFF
cinap_lenrek [Fri, 23 Jan 2015 13:01:56 +0000 (14:01 +0100)]
pc: dont handle pending floating point exception in fpsave thru FPOFF

the FPOFF macro that follows the FXSAVE/FSAVE instructions in l.s
used to execute WAIT instruction when the TS flag was not set. this
is wrong and causes pending exceptions to be raised from fpsave which
is called from provsave() which holds up->rlock making it deadlock
when matherror() tries to postnote() to itself.

so making FPOFF non-waiting (just set TS flag).

we handle pending exception when restoring the context.

9 years agolibdraw: fix broken eenter()
cinap_lenrek [Sun, 18 Jan 2015 09:34:05 +0000 (10:34 +0100)]
libdraw: fix broken eenter()

eenter would go into a endless loop in the redraw avoidance case
because the label was misplaced.

9 years agovgadb: use X60t without hardware cursor (broken currently)
cinap_lenrek [Sat, 17 Jan 2015 04:15:59 +0000 (05:15 +0100)]
vgadb: use X60t without hardware cursor (broken currently)

9 years agoigfx: remove hardware cursor support for X60t, does not work
cinap_lenrek [Sat, 17 Jan 2015 04:14:23 +0000 (05:14 +0100)]
igfx: remove hardware cursor support for X60t, does not work

9 years agoigfx: fix wrong cursor position register address
cinap_lenrek [Sat, 17 Jan 2015 03:00:36 +0000 (04:00 +0100)]
igfx: fix wrong cursor position register address

9 years agoiwl: add "reset" ctl message to software tigger device reset
cinap_lenrek [Sat, 17 Jan 2015 00:52:24 +0000 (01:52 +0100)]
iwl: add "reset" ctl message to software tigger device reset

9 years agovgadb: add lcd2190uxp, x60t
stanley lieber [Fri, 16 Jan 2015 19:19:45 +0000 (14:19 -0500)]
vgadb: add lcd2190uxp, x60t

9 years agoigfx: determine amount of preallocated stolen graphics memory for hw cursor
cinap_lenrek [Fri, 16 Jan 2015 12:38:52 +0000 (13:38 +0100)]
igfx: determine amount of preallocated stolen graphics memory for hw cursor

allocate the hardware cursor plane at the end of
preallocated stolen graphics memory.

9 years agodevvga: use uintptr for VGAscr.paddr and VGAscr.storage
cinap_lenrek [Fri, 16 Jan 2015 10:36:11 +0000 (11:36 +0100)]
devvga: use uintptr for VGAscr.paddr and VGAscr.storage

some vga drivers use virtual addresses in VGAscr.storage,
so has to be uintptr sized for amd64.

9 years agoigfx: support for X60t with 1400x1050 panel
cinap_lenrek [Fri, 16 Jan 2015 06:55:38 +0000 (07:55 +0100)]
igfx: support for X60t with 1400x1050 panel

program secret plane size and position registers described as "reserved"
in g45_vol_3_register_0_0.pdf that was found by inspecting vesa bios
port traces.

also, we have to set 18:19 (Cursor/Dispaly/Overlay Planes Off) in
PIPExCONF while programming the planes on this card. this is what
vesa bios does on modeset.

9 years agowebcookies: only sync the jar when dirty on clunk
cinap_lenrek [Fri, 16 Jan 2015 03:46:52 +0000 (04:46 +0100)]
webcookies: only sync the jar when dirty on clunk

avoid sync the jar file when fids get clunked.
the only reason to sync the jar on clunk is when it has
been marked dirty (cookies added or deleted) and we
want to flush the changes to disk.

9 years agowebcookies: avoid rereading and rewriting the cookie jar file all the time, dont...
cinap_lenrek [Fri, 16 Jan 2015 03:04:55 +0000 (04:04 +0100)]
webcookies: avoid rereading and rewriting the cookie jar file all the time, dont return deleted cookies on search

- rewrite when jar->dirty != 0 (caller modified the in memory jar)
- reread when the jar->qid != stat(jar->file)->qid (on disk file changed)
- ignore deleted cookies in cookiesearch()

9 years agoigfx: cursor/plane pipe assign is G45 only, add magic dsp a toggle from enable sequence
cinap_lenrek [Tue, 13 Jan 2015 17:04:33 +0000 (18:04 +0100)]
igfx: cursor/plane pipe assign is G45 only, add magic dsp a toggle from enable sequence

9 years agokbfds: delete autogenerated latin1.h file when mklatin fails
cinap_lenrek [Mon, 12 Jan 2015 23:19:18 +0000 (00:19 +0100)]
kbfds: delete autogenerated latin1.h file when mklatin fails

9 years agoplan9.ini(8) man page: update links to iwn and ral firmware blobs
stanley lieber [Mon, 12 Jan 2015 19:23:32 +0000 (14:23 -0500)]
plan9.ini(8) man page: update links to iwn and ral firmware blobs

9 years agovgadb: add x301
mischief [Mon, 12 Jan 2015 17:52:45 +0000 (09:52 -0800)]
vgadb: add x301

9 years agoigfx: set the display/overlay/cursor plane override disbale bits for G45 (like bios)
cinap_lenrek [Mon, 12 Jan 2015 15:52:49 +0000 (16:52 +0100)]
igfx: set the display/overlay/cursor plane override disbale bits for G45 (like bios)

9 years agoigfx: fix typo, initialize more lvds bits for G45, T60 testing
cinap_lenrek [Mon, 12 Jan 2015 15:40:02 +0000 (16:40 +0100)]
igfx: fix typo, initialize more lvds bits for G45, T60 testing

9 years agovgadb: alias X200s for X200
cinap_lenrek [Mon, 12 Jan 2015 03:14:01 +0000 (04:14 +0100)]
vgadb: alias X200s for X200

9 years agoigfx: reverting previous change, pci id is for the second graphics controller pci...
cinap_lenrek [Mon, 12 Jan 2015 03:07:16 +0000 (04:07 +0100)]
igfx: reverting previous change, pci id is for the second graphics controller pci device

9 years agoigfx: support for X200 (thanks bigato)
cinap_lenrek [Mon, 12 Jan 2015 02:09:58 +0000 (03:09 +0100)]
igfx: support for X200 (thanks bigato)

9 years agoigfx: properly turn vga monitor off
cinap_lenrek [Mon, 12 Jan 2015 01:10:00 +0000 (02:10 +0100)]
igfx: properly turn vga monitor off

9 years agoigfx: aaand vga support for x200s :-)
cinap_lenrek [Mon, 12 Jan 2015 01:04:13 +0000 (02:04 +0100)]
igfx: aaand vga support for x200s :-)

9 years agoigfx: vga support on x230, fix fdi link setup, LG Flatron L1730P vgadb entry
cinap_lenrek [Mon, 12 Jan 2015 00:51:51 +0000 (01:51 +0100)]
igfx: vga support on x230, fix fdi link setup, LG Flatron L1730P vgadb entry

9 years agofix missing games/gba target in mkfile (thanks qu7uux)
cinap_lenrek [Sun, 11 Jan 2015 23:02:15 +0000 (00:02 +0100)]
fix missing games/gba target in mkfile (thanks qu7uux)

9 years agoigfx: only use mmio... left from debugging
cinap_lenrek [Sun, 11 Jan 2015 03:03:14 +0000 (04:03 +0100)]
igfx: only use mmio... left from debugging

9 years agoigfx: cleanup
cinap_lenrek [Sun, 11 Jan 2015 02:44:56 +0000 (03:44 +0100)]
igfx: cleanup

9 years agomerge
cinap_lenrek [Sun, 11 Jan 2015 02:37:07 +0000 (03:37 +0100)]
merge

9 years agoigfx: get edid information from lvds
cinap_lenrek [Sun, 11 Jan 2015 02:35:30 +0000 (03:35 +0100)]
igfx: get edid information from lvds

9 years agoip/ipconfig: set ndb database file with -f argument
mischief [Sat, 10 Jan 2015 19:27:45 +0000 (11:27 -0800)]
ip/ipconfig: set ndb database file with -f argument

9 years agoxargs(1) man page: rewritten -> implemented
stanley lieber [Sat, 10 Jan 2015 18:48:18 +0000 (13:48 -0500)]
xargs(1) man page: rewritten -> implemented

9 years agoxargs(1) man page: add HISTORY (thanks, jpm_)
stanley lieber [Sat, 10 Jan 2015 18:46:47 +0000 (13:46 -0500)]
xargs(1) man page: add HISTORY (thanks, jpm_)

9 years agoigfx: use mmio to access registers instead of pio, fix wrong igfxmmio segment size
cinap_lenrek [Sat, 10 Jan 2015 02:07:29 +0000 (03:07 +0100)]
igfx: use mmio to access registers instead of pio, fix wrong igfxmmio segment size

initially, pio was used to access registers so i didnt need
a kernel driver for initial testing.

pio does not work under efi, so use mmio to access registers.

9 years agomerge
cinap_lenrek [Sat, 10 Jan 2015 01:16:50 +0000 (02:16 +0100)]
merge

9 years agoigfx: perserve frame start delay "magic" bits 27:28 in pipe conf register
cinap_lenrek [Sat, 10 Jan 2015 01:15:30 +0000 (02:15 +0100)]
igfx: perserve frame start delay "magic" bits 27:28 in pipe conf register

it turns out, the "magic" bits 27:28 are the frame start
delay setting that need to be preserved.

9 years agolibdraw: don't redraw input box in enter/eenter when mouse is moved outside the rectangle
mischief [Fri, 9 Jan 2015 23:41:00 +0000 (15:41 -0800)]
libdraw: don't redraw input box in enter/eenter when mouse is moved outside the rectangle

prevents some flickering when devdraw is used over a high latency connection.

Fixes issue 236

9 years agoigfx: vgadb entries for x230 and x200s
cinap_lenrek [Fri, 9 Jan 2015 23:19:16 +0000 (00:19 +0100)]
igfx: vgadb entries for x230 and x200s

9 years agoigfx: implement hardware cursor
cinap_lenrek [Fri, 9 Jan 2015 21:23:25 +0000 (22:23 +0100)]
igfx: implement hardware cursor

this can even be used with the vesa driver, just
enable the cursor after mode switch like:

echo hwgc igfxhwgc >/dev/vgactl

9 years agoigfx: fix some comments
cinap_lenrek [Fri, 9 Jan 2015 02:18:44 +0000 (03:18 +0100)]
igfx: fix some comments

9 years agoigfx: aaaand the kernel part :(
cinap_lenrek [Fri, 9 Jan 2015 02:03:23 +0000 (03:03 +0100)]
igfx: aaaand the kernel part :(

9 years agoigfx: just kidding, heres the code :)
cinap_lenrek [Fri, 9 Jan 2015 01:58:14 +0000 (02:58 +0100)]
igfx: just kidding, heres the code :)

9 years agoigfx: work in progress intel graphics driver
cinap_lenrek [Fri, 9 Jan 2015 01:55:12 +0000 (02:55 +0100)]
igfx: work in progress intel graphics driver

this driver can modeset lvds on x200s and x230.
everything else is completely untested.
no hardware cursor implemented.
no vgadb entries are provided.

9 years agomerge
cinap_lenrek [Fri, 9 Jan 2015 01:45:14 +0000 (02:45 +0100)]
merge

9 years agovncs: add -A for no auth to usage
mischief [Wed, 7 Jan 2015 04:13:16 +0000 (20:13 -0800)]
vncs: add -A for no auth to usage

9 years agovncs: turn off auth with -A, do not require -c for -x
mischief [Tue, 6 Jan 2015 04:21:40 +0000 (20:21 -0800)]
vncs: turn off auth with -A, do not require -c for -x