Commit Graph

264 Commits

Author SHA1 Message Date
matthieu
1c882161e8 In priv_open_device() allow opening tty[E-J]0. While there remove
ttyD[1-7] from the list, since those devices will never get used by X.
2012-08-14 15:57:57 +00:00
matthieu
1996326d50 Fix a logic introducred in rev 1.23. The parent pid is initialized
by the main X server too late in the privsep case (already in the
unpriviliged child). So keep the early init for this case.
2012-08-12 14:06:42 +00:00
matthieu
143961ca76 Un-blacklist radeon, blacklist radeonold. So X -configure will now
select the radeon driver.
2012-08-08 16:37:19 +00:00
matthieu
5f8132e311 Add privsep prototypes to osdep.h 2012-08-07 20:16:12 +00:00
matthieu
c7c0180b4c In non-privilege sepration mode, avoid accidentally sending
a SIGUSR1 signal to init(8).

It can happen that xdm dies before the X server that it started.
In that case X's is reparented by init...

This is handled correctly when privilege separation is not compiled
but got overlooked in the privilege separation case.
2012-08-07 20:15:23 +00:00
matthieu
4f2bf5df6d Rename 'socket' parameter to avoid shadowing the global declaration. 2012-08-07 20:13:18 +00:00
matthieu
20c1ecb828 Remove white space diff with X.Org. 2012-08-07 20:12:01 +00:00
matthieu
eb59960f12 regen autotools 2012-08-05 18:14:29 +00:00
matthieu
fe11647d1f Update to xserver 1.12.3. 2012-08-05 18:11:37 +00:00
matthieu
8ead23f263 Remove mac68k bits from Xenocara.
The imake config bits are not removed.
Those files are frozen, until they can be removed, once the last
port that still uses imake dies.
2012-06-20 19:56:34 +00:00
matthieu
baeff29842 screenFd -> consoleFd conversions that were missed in xserver 1.12 import. 2012-06-16 15:08:08 +00:00
matthieu
e60da74507 Update to xserver 1.12.2. tested by naddy@, krw@, mpi@. 2012-06-10 13:21:05 +00:00
matthieu
b4a75b3e96 Return an error much earlier if recvmsg fails. 2012-04-04 20:34:55 +00:00
matthieu
f0cf820ebf Disable doc formatting tools explicitely 2012-03-10 18:35:56 +00:00
matthieu
e4b13eedd8 Introduce a black list of drivers for Xorg -configure, and blacklist
all radeon driver variants, except radeonold.
This fixes the conflicting symbols warnings seen by Xorg -configure
and forces it to use radeonold in the generated xorg.conf file.

Xorg -configure is still broken in many ways, and should be
avoided.
2012-03-04 17:55:31 +00:00
matthieu
a682cc0ce7 Use the 'macintosh' xkb model on adb keyboards. From mpi@, reminded by
landry@. ok deraadt@
2012-02-11 18:30:29 +00:00
matthieu
2be03eee0e Disable build of Xserver on hp300 and mac68k. Building the Xserver
with gcc 2.95 is getting hopeless and there is not enough manpower at
the moment to switch these platforms to a more recent compiler.
ok miod@, Nick@
2012-02-07 19:46:11 +00:00
matthieu
9576ef223d Update to xserver 1.11.4. tested by krw@, shadchin@. 2012-01-31 07:52:35 +00:00
matthieu
d253c9004d Switch to the xf86-input-ws input driver for mouse by default.
ok shadchin@.
2012-01-08 17:15:52 +00:00
matthieu
05b4e5f308 Revert stuff from my local tree which leaked in xserver 1.11.3 update. 2011-12-29 10:47:37 +00:00
matthieu
4344ac3914 Bugfix Update to xserver 1.11.3 2011-12-18 16:08:59 +00:00
ajacoutot
289590f6a5 Merge patch from http://patchwork.freedesktop.org/patch/7707/,
Xi: allow passive keygrabs on the XIAll(Master)Devices fake deviceslogin.
Amongst other things, this fixes a regression that can be seen in media
keys handling under gnome-settings-daemon.

ok matthieu@
2011-11-15 00:09:30 +00:00
matthieu
fad3fc950a Bring back ChangeLog 2011-11-05 14:18:04 +00:00
matthieu
61a7d5427d Update to xserver 1.11.2 2011-11-05 13:32:40 +00:00
matthieu
a05754665a Fix CVE-2011-4028: File disclosure vulnerability.
use O_NOFOLLOW to open the existing lock file, so symbolic links
aren't followed, thus avoid revealing if it point to an existing file.

Note that xserver on OpenBSD isn't affected by CVE-2011-4029.
2011-10-18 14:58:36 +00:00
mpi
9fa376d08a Use the synaptics driver with ALPS touchpads
with and ok matthieu@
2011-09-21 17:42:03 +00:00
shadchin
4e607c59bd Cleaning of current solution for 'Kill X on Ctrl+Alt+Backspace on default'
ok matthieu@
2011-09-17 13:39:30 +00:00
matthieu
5e01252a10 unifdef support for xkeyboard-config.
I don't know why cvs didn't pick up this chunk. Noticed by okan@
2011-08-31 12:31:51 +00:00
matthieu
ec7d9704ca Fix default xkboptions settings here too. 2011-08-21 09:18:40 +00:00
matthieu
b9f30b39b5 Remove warnings emitted when a device can't be opened. This
is just noise and now happens while the X autoconfiguration
code probes all /dev/wsmouse<n> devices.
If the error matters, the driver will emit a proper error
nevertheless.
2011-08-20 17:30:37 +00:00
kettenis
b1e129123c Fix transformation of events supplying only a valuator for the y-coordinate.
ok matthieu@
2011-07-04 20:40:19 +00:00
matthieu
394dd9fc87 A wscons keyboard type of zaurus should map to an XKB model of zaurus 2011-07-04 20:27:57 +00:00
matthieu
a4d630d049 regen 2011-06-29 19:57:45 +00:00
matthieu
b6047701ef Introduce new autoconfiguration code for X input drivers, based on X
hotplug framework.This also takes care of the configuration of the
keyboard layout from the wscons layout that was hacked inside the
keyboard driver.

Tested by many, ok oga@.
2011-06-29 19:55:01 +00:00
shadchin
89a8ad0171 If you do not have XkbOptions in xorg.conf, it is automatically set to
"terminate: ctrl_alt_bksp" (only if XENOCARA_USE_XKEYBOARD_CONFIG=Yes)

tested ajacoutot@, ok matthieu@
2011-05-30 10:52:57 +00:00
matthieu
9c5fcc7f6c Enable X-ACE and XC-SECURITY extensions, now that they build on
OpenBSD. Requested and tested by tedu@.
2011-04-27 20:46:31 +00:00
matthieu
4a238ea6a4 Update to xserver 1.9.5. Tested by jasper@, ajacoutot@ and krw@ 2011-04-02 16:08:38 +00:00
matthieu
eb61d3c991 Xi: add XI_Focus{In,Out} to swapped events. 2011-04-01 21:28:00 +00:00
matthieu
e03328b4d1 Use the array size for checking the validity of request numbers
instead of hard-coded constants. Suggested by Julien Cristau.
2011-04-01 21:27:04 +00:00
matthieu
90f7a9841a Xext: fix test on extension number for the swapped case. 2011-04-01 21:24:37 +00:00
matthieu
a282fbce7f WriteReplyToClient() swaps rep contents if client and server endianess
differ. Cache rep.length so that it can be used after returning from
WriteReplyToClient().
Fixes xinput --list when client and server have different endianess.
gtk3-demo now fails a bit further in this case.
2011-03-28 21:10:15 +00:00
matthieu
6e1b78e41d Give XI2 requests a chance if server and client endianess differ.
There are more bugs in this code though.
2011-03-27 19:51:09 +00:00
matthieu
2cc47a1202 add a new build switch 'XENOCARA_USE_XKEYBOARD_CONFIG' to enable
the use of xkeyboard-config xkb definitions instead of the old xkbdata.
This will make testing easier. No changes in default builds.
ok shadchin@
2011-03-08 07:59:23 +00:00
matthieu
e169dc846e Revert stuff that should not have been committed in version 1.33.
Noticed by krw@
2011-02-20 13:42:17 +00:00
matthieu
f98a26f8f7 Disable Xorg on alpha. Xrender is broken since 2 releases and I still
don't know how to fix it. ok miod@, todd@
2011-02-19 19:51:53 +00:00
matthieu
da0820c3bb Merge another EXA fix from X.Org 1.9.4RC1:
EXA: Pad size of system memory copy for 1x1 pixmaps (bug #32803).
2011-02-02 20:47:08 +00:00
stsp
fba58a8a71 Apply fix from upstream:
commit e06fa804009798ea95efa8babaabb0228dfdfe65
  EXA: Fix crash with fill using 1x1 tile of depth < 8 (bug #24703).
Fixes http://bugs.freedesktop.org/show_bug.cgi?id=24703
Reported by Barry Grumbine on ports@
ok matthieu@
2011-01-31 21:10:06 +00:00
matthieu
3870417379 restart recvmsg() if returning with errno==EINTR.
Fixes a crash on server reset on some machines.
Code inspired by ssh, with feedback from guenther@ and millert@
ok guenther@ miod@
2011-01-28 19:37:55 +00:00
matthieu
805ddd83d6 Effectively close the AGP device fd in xf86GARTCloseScreen() if it's
different from the consoleFd. Fixes crashes at server reset.
"looks good" miod@, oga@. Tested by shadchin@.
2011-01-28 19:34:22 +00:00
kettenis
31dc726d5a Call xf86Reconfigure() after readjusting the physical dimensions of the screen
based on EDID information.  Fixes issues with the mouse getting stuck on dual
card setups.  Still trying to push this one upstream.

ok matthieu@
2011-01-10 20:45:49 +00:00