Commit Graph

1463 Commits

Author SHA1 Message Date
Adam Jackson a8f9936f55 Prefer driver-provided modes when matching name strings to modelines.
Well, kinda.  Strictly we prefer M_T_BUILTIN strongest since those are modes
where the driver has said it absolutely can't do anything else (VBE).  Then
we look for user-defined modes, ie, modelines from the config file.  Then
we consider modes reported by the monitor via EDID.  Finally if nothing has
matched yet we consider the default mode pool.

Within each of the above-mentioned classes, modes with the M_T_PREFERRED bit
take priority over other modes in the same class.

This logic ensures that the timings sent to the monitor exactly match the
timings it reported as supported, which occasionally don't match the numbers
you might get for that mode from CVT or GTF.
2006-09-14 19:24:41 -04:00
Adam Jackson 81ef1b6d60 Mark EDID modes as driver modes. Infer virtual size from driver modes.
This allows the server to guess an appropriate initial virtual size and
resolution.  The heuristic is to select the largest driver-reported mode
that matches the monitor's physical aspect ratio.  We revalidate this
estimate after mode validation, since we may have filtered away all
modes that would fill that size.

Also, the EDID preferred timing is now marked as M_T_PREFERRED as well.
2006-09-14 19:18:58 -04:00
Adam Jackson 43d9edd31e Attempt to add the 'mouse' driver in more situations.
Always add a mouse driver instance configured to send core events, unless
a core pointer already exists using either the mouse or void drivers.  This
handles the laptop case where the config file only specifies, say,
synaptics, which causes the touchpad to work but not the pointing stick.
We don't double-instantiate the mouse driver to avoid the mouse moving twice
as fast, and we skip this logic when the user asked for a void core pointer
since that probably means they want to run with no pointer at all.
2006-09-14 19:09:02 -04:00
Adam Jackson 739224d05e Load the default module set when no Module section is given in the config.
Also, synchronize that list with the list for the pseudoconfig file used
when starting with no config file.  These really need to be better unified.
2006-09-14 19:03:32 -04:00
Adam Jackson beac2bf1e4 Expand the default sync ranges to be large enough for 800x600@60. 2006-09-14 19:01:13 -04:00
Adam Jackson 71a15a7d57 Publish the raw EDID block as a property on the root window.
This was removed in the patch for bug #5386, but is still useful.
2006-09-14 19:00:10 -04:00
Adam Jackson 7939c8dfb7 Bump the default pixel depth to 24, and default bpp to 32. 2006-09-14 18:57:57 -04:00
Adam Jackson 72af975f9c Fix up EDID blocks where the max pixclock exceeds the preferred mode clock.
Base EDID only lets you specify the maximum dotclock in tens of MHz, which
is too fuzzy for some monitors.  1600x1200@60 is just over 160MHz, but if
the monitor really can't handle any mode at 170MHz, then 160 is more
correct.  Fix up the EDID block before the driver can see it in this case,
so we don't spuriously reject modes.
2006-09-14 18:56:34 -04:00
Adam Jackson 334f7db9f6 Allow hsync and vsync ranges to be overridden independently again. 2006-09-14 18:46:10 -04:00
Adam Jackson ced46e1777 Record all standard timings from EDID as modes, instead of just the first five. 2006-09-14 18:41:59 -04:00
Adam Jackson d89fee68d0 Record the maximum dot clock of the monitor, and filter by it. 2006-09-14 18:41:15 -04:00
Adam Jackson fa8ef71668 Don't translate monitor gamma to X gamma.
The X gamma is used to set the output ramp of the card.  Setting a 2.2 output
gamma going into a 2.2 monitor gives an effective gamma of 4.84, which is
very much not what you want.
2006-09-14 18:33:00 -04:00
Luc Verhaegen 511c60bc73 Bug #5386: Synthesize modelines from EDID data. 2006-09-14 18:30:36 -04:00
Bill Nottingham 0a62840e2c Bug 7641: fix comment written to Xorg.conf (s/VertSync/VertRefresh/)
X.Org Bugzilla #7641 <https://bugs.freedesktop.org/show_bug.cgi?id=7641>
Patch #6349 <https://bugs.freedesktop.org/attachment.cgi?id=6349>
2006-09-13 15:40:23 -07:00
Drew Parsons 182e5e0f4b Xprint: revert installation of /etc/X11/Xsession.d/cde_xsessiond_xprint.sh
pending resolution of #8232.
2006-09-12 14:30:46 +10:00
Donnie Berkholz 2b357e9a2f If we're installing libxf86config, install headers needed to build against it. 2006-09-10 22:17:20 -07:00
Jesse Barnes b3a3020fd0 the new PCI mapping routines are broken on sparc64 (in fact they look
broken for any 32 bit X server running on a 64 bit kernel) so #ifdef
them out for now.  the PCI rework tree will make all this crap go away,
so I think we can tolerate the extra #ifdef for the next release.
2006-09-10 11:13:18 -07:00
Drew Parsons 65256109bb * Define XPSERVERLIST with `/etc/init.d/xprint get_xpserverlist`
instead of `/bin/sh /etc/init.d/xprint get_xpserverlist`
	- allows the initscript to set its own different shell under #!
	- allows disabling of XPSERVERLIST by making the script non-executable
* Allow files to be installed by using dist_*_DATA instead of EXTRA_DIST.
  Also, use dist_*_SCRIPTS to install scripts.
* Fix minor typos in man pages.
2006-09-10 17:40:37 +10:00
Michel Dänzer 8356be492c Make sure _XSERVER64 is defined when it should be and gets tested. 2006-09-06 15:20:55 +02:00
Matthew Allum 69d5becce4 Fix previous commit breaking other kdrives pulling in fbdev.a 2006-08-31 17:30:24 +01:00
Matthew Allum fd609956f2 Add framebuffer device command line switch for Xfbdev 2006-08-31 17:18:57 +01:00
Matthew Allum 5ddbf4bcd4 Re-add support for tslib (1.0 release) and Xcalibrate extension. 2006-08-29 22:07:15 +01:00
Daniel Stone 393dc0a338 kdrive: remove @KDRIVE_LIBS@ from Xfoo_DEPENDENCIES 2006-08-29 22:53:54 +03:00
Adam Jackson ebbdc1342a Remove calls to LoaderCheckUnresolved(), since it's now a stub. 2006-08-28 18:17:32 -04:00
Adam Jackson b879356ce9 More #ifdef USE_DEPRECATED_KEYBOARD_DRIVER. 2006-08-24 15:50:15 -04:00
Matthias Hopf 4ed311cf1c Merge branch 'master' of git://anongit.freedesktop.org/git/xorg/xserver 2006-08-24 20:17:10 +02:00
Alan Hourihane ce4a0a4dda Apply patch in bug #7919, blit improvements in
multiwindow mode for Xming/CygwinX
2006-08-24 13:56:22 +01:00
David Nusinow 733c4beb16 Add xorg.conf IgnoreABI option which does the same thing as -ignoreABI 2006-08-23 22:39:42 +00:00
Adam Jackson bdec9680fa Make sure Composite is never enabled for Xnest. 2006-08-23 14:43:23 -04:00
Adam Jackson 9f2a108051 Make 'Xvfb -render' also disable Composite, lest we segfault on startup. 2006-08-23 14:38:34 -04:00
Alan Hourihane 70ddd0f39d Fix bug #5735, Serious flaw in CygwinX clipboard
integration prevents paste from X to Windows apps
(Brett Stahlman & Colin Harrison)
2006-08-18 14:43:10 +01:00
Alan Hourihane 708b225689 Merge branch 'master' of git+ssh://git.freedesktop.org/git/xorg/xserver 2006-08-18 09:13:52 +01:00
Alan Hourihane 1880defe4e Fix bug #7281, clipboard viewer should not
call SetClipboard viewer when bogus in Xming/CygwinX
(Colin Harrison)
2006-08-18 09:09:53 +01:00
Alan Hourihane a1a8e4f7f5 Fix bug #7280, round title corner background
should be transparent not black in Xming/CygwinX
(Colin Harrison)
2006-08-18 09:08:12 +01:00
Adam Jackson ee5e2cbd2b Un-cut-and-paste the mode rejection message. 2006-08-17 17:29:32 -04:00
Adam Jackson 43e42eef1f Fix default mouse device on Linux, again.
It would be really nice if we only did this in one place instead of 40.
2006-08-17 16:22:07 -04:00
Matthias Hopf 73e58adda9 Fixed segfault w/ broken Xinerama configs. 2006-08-16 18:18:51 +02:00
Bastian Blank f7919c2879 xfree86: don't do legacy IO on ARM or S/390 (Debian #362641)
Don't attempt to poke legacy IO ranges on ARM or S/390.
2006-08-12 20:43:25 +03:00
Samuel Thibault e641000b98 xfree86: add Hurd support (#5613)
Add support for GNU/Hurd.
2006-08-12 19:41:59 +03:00
Adam Jackson c4951e0a6b Fix a mode sanity check to not break reduced-blanking setups (LCDs). 2006-08-10 20:49:06 -04:00
Adam Jackson e1921f014b Rename some mode tokens to better reflect their use.
Per #5386, M_T_EDID -> M_T_DRIVER, since it's really for any driver-detected
mode.  Also add M_T_PREFERRED bit, to select a 'best' mode out of a set.
2006-08-10 20:43:15 -04:00
Adam Jackson db82e12fac Remove TargetRefresh option from the autoconfig logic.
The default target of 75Hz is almost always wrong for LCDs.
2006-08-09 14:55:17 -04:00
Daniel Stone 12dbd8a02f remove optional R3 backwards compatibility
Remove the permitOldBugs flag, which enabled backwards compatbility with
broken R2/R3 era clients.
2006-08-07 23:43:40 +03:00
David Nusinow f54b71b772 Document enable/disable flag for AIGLX in xorg.conf manpage. 2006-08-06 18:11:00 +00:00
Matthew Allum 39169fd373 Back out 'mystery' spurious host window hints. 2006-08-01 13:39:22 +01:00
Ian Romanick 3112a6c4f2 Noting uses libdummy.a, so don't build it. Only libdummy-nonserver.a
is actually used.
2006-07-31 10:25:31 -07:00
Daniel Stone 24051ef974 remove filename that's too long for tar 2006-07-30 12:15:33 +03:00
Daniel Stone ecb7d43a76 add sym.h to sources 2006-07-30 11:52:41 +03:00
Daniel Stone e87e68634d remove .cvsignores from EXTRA_DIST 2006-07-30 11:11:59 +03:00
Daniel Stone a68dc013a3 remove README (which doesn't exist) from EXTRA_DIST 2006-07-30 11:08:47 +03:00