So when we VT switch back and attempt to flush the input devices,
we don't succeed because evdev won't return part of an event,
since we were only asking for 4 bytes, we'd only get -EINVAL back.
This could later cause events to be flushed that we shouldn't have
gotten.
This is a fix for CVE-2013-1940.
Signed-off-by: Dave Airlie <airlied@redhat.com>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Revert 70739e817b and mostly revert
c31eac647a.
Further investigation shows the encountered race condition is between
lightdm and plymouth-splash, as implemented in the Ubuntu distribution
within the limitations of upstart's job coordination logic, and can (and
should) be fixed within those limiations. Not in xserver itself.
This leaves some of the diagnostic improvements from the recent patch
series, in case others run into a similar situation.
Signed-off-by: Bryce Harrington <bryce@canonical.com>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
This path is technically executed through config/udev, but having two
messages in the form "config/udev: Adding drm device" makes it appear as if
the udev filters are wrong and it's trying to add the same device twice. In
fact, it's only one device, only added once, but a duplicate log message.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Dave Airlie <airlied@redhat.com>
We don't want to hotplug output devices while we are VT switched,
as we get races between multiple X servers on the device open, and
drm device master status. This just queues device opens until we return
from VT switch.
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Dave Airlie <airlied@redhat.com>
This replaces some previous uses of direct xf86Screens[0] accesses.
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Dave Airlie <airlied@gmail.com>
This is just a simple interface to avoid accessing x86Screens[0]
directly.
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Dave Airlie <airlied@gmail.com>
This removes a large number of redundant declaration warnings.
Signed-off-by: Jeremy White <jwhite@codeweavers.com>
Reviewed-by: Robert Morell <rmorell@nvidia.com>
Signed-off-by: Keith Packard <keithp@keithp.com>
Signed-off-by: Bryce Harrington <bryce@canonical.com>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Keith Packard <keithp@keithp.com>
If other processes have had drm open previously, xserver may attempt to
open the device too early and fail, with xserver error exit "Cannot
run in framebuffer mode" or Xorg.0.log messages about "setversion 1.4
failed".
In this situation, we're receiving back -EACCES from libdrm. To address
this we need to re-set ourselves as the drm master, and keep trying to
set the interface until it works (or until we give up).
See https://bugs.launchpad.net/ubuntu/+source/libdrm/+bug/982889
Signed-off-by: Bryce Harrington <bryce@canonical.com>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Keith Packard <keithp@keithp.com>
And if we've had to delay booting due to not being able to set the
interface, fess up.
Signed-off-by: Bryce Harrington <bryce@canonical.com>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Keith Packard <keithp@keithp.com>
Signed-off-by: Bryce Harrington <bryce@canonical.com>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Keith Packard <keithp@keithp.com>
Signed-off-by: Bryce Harrington <bryce@canonical.com>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Keith Packard <keithp@keithp.com>
Signed-off-by: Bryce Harrington <bryce@canonical.com>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Keith Packard <keithp@keithp.com>
Section "Monitor"
Identifier "a21inch"
Option "PreferredMode" "1600x1200"
Option "ZoomModes" "1600x1200 1280x1024 1280x1024 640x480"
EndSection
The option's effect is to search for and mark once each named mode in
the output modes list. So the specification order is free and the zoom
modes sequence follows the order of the output modes list. All marked
modes are available via the Ctrl+Alt+Keypad-{Plus,Minus} key
combination.
See also http://bugs.freedesktop.org/show_bug.cgi?id=17954.
This option has its use for combined monitor and television setups.
It allows for easy switching between 60 Hz and 50 Hz modes even when a
monitor refuses to display the input signal.
(Includes a few minor changes suggested by Aaron for v2)
Signed-off-by: Servaas Vandenberghe <vdb@picaros.org>
Reviewed-by: Aaron Plattner <aplattner@nvidia.com>
Signed-off-by: Keith Packard <keithp@keithp.com>
If a device is not primary, the PCI device match fails because the
xf86-video-modesetting driver looks specifically for a PCI class match of
0x30000 with a mask of 0xffffff. This fails to match, for example, a
non-primary Intel VGA device, because it is reported as having a class of
0x38000.
Fix that by ignoring the low 16 bits of the class in the pci_id_match table.
Signed-off-by: Aaron Plattner <aplattner@nvidia.com>
Reviewed on IRC by Adam Jackson <ajax@redhat.com>
Otherwise, displays driven by GPU screens remain on all the time.
Signed-off-by: Aaron Plattner <aplattner@nvidia.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Keith Packard <keithp@keithp.com>
So in the cold plug server shutdown case, we reap the resources
before we call CloseScreen handlers, so the config->randr_provider
is a dangling pointer when the xf86CrtcCloseScreen handler is called,
however in the hot screen unplug case, we can't rely on automatically
reaped resources, so we need to clean up the provider in the xf86CrtcCloseScreen
case.
This patch provides a cleanup callback from the randr provider removal
into the DDX so it can cleanup properly, this then gets called by the automatic
code for cold plug, or if hot unplug it gets called explicitly.
Fixes a number of random server crashes on shutdown
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=58174
Bugzilla: https://bugzilla.redhat.com/show_bug.cgi?id=891140
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
Signed-off-by: Dave Airlie <airlied@redhat.com>
The previous fix for the previous fix, didn't fully work,
If we don't set compat_output we end up doing derferences
of arrays with -1, leading to valgrind warnings.
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
Signed-off-by: Dave Airlie <airlied@redhat.com>
Due to another bug, the modesetting/udl driver would fail to init properly
on hotplug, when it did the code didn't clean up properly, and on removing
the device the server could crash.
Found in F18 testing.
Signed-off-by: Dave Airlie <airlied@redhat.com>
Reviewed-by: Aaron Plattner <aplattner@nvidia.com>
A fixed-mode output device like a panel will often only inform of its
preferred mode through its EDID. However, the driver will adjust user
specified modes for display through use of a panel-fitter allowing
greater flexibility in upscaling. This is often used by games to set a
low resolution for performance and use the panel fitter to fill the
screen.
v2: Use the presence of the 'scaling mode' connector property as an
indication that a panel fitter is attached to that pipe.
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=55564
xf86Cursor.c:19:18: warning: redundant redeclaration of 'inputInfo'
[-Wredundant-decls]
In file included from xf86Cursor.c:18:0:
../../../include/inputstr.h:614:57: note: previous declaration of
'inputInfo' was here
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Alan Coopersmith <alan.coopersmith@oracle.com>
Reviewed-by: Mark Kettenis <kettenis@openbsd.org>
Unused as of 5d309af2ed
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Alan Coopersmith <alan.coopersmith@oracle.com>
Reviewed-by: Mark Kettenis <kettenis@openbsd.org>
This is necessary when the input handler deletes itself from the
list. Bug found by Maarten Lankhorst, this patch uses the list macros
instead of open-coding the fix.
Signed-off-by: Keith Packard <keithp@keithp.com>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
man xorg.conf states that the 'Device' identifier is required in the
'Screen' section, yet current xserver defaults properly and boots up
fine without it.
Fixes: https://bugs.freedesktop.org/show_bug.cgi?id=20742
Signed-off-by: Bryce Harrington <bryce@canonical.com>
Signed-off-by: Keith Packard <keithp@keithp.com>
Instead of defaulting to -intel for Oaktrail, Medfield, and CDV chips,
default to -fbdev. For Poulsbo (only), attempt to use -psb if it's
installed, and fallback to fbdev otherwise. All other Intel chips
should use -intel.
This fixed an issue where -intel would load on these chips and cause a
boot failure. Newer -intel drivers avoid the boot hang, but it's still
the wrong driver to load, so why take chances.
The patch was originally created by Stefan Dirsch for OpenSUSE. We have
included it in our stable release (Ubuntu "quantal" 12.10) since
December.
ref: https://bugzilla.novell.com/show_bug.cgi?id=772279
ref: https://bugs.launchpad.net/ubuntu/+bug/1069031
Fixes: https://bugs.freedesktop.org/show_bug.cgi?id=60514
Signed-off-by: Bryce Harrington <bryce@canonical.com>
Signed-off-by: Keith Packard <keithp@keithp.com>
If we're about to abort, we're already in the signal handler and cannot call
down to the default device cleanup routines (which reset, free, alloc, and
do a bunch of other things).
Add a new DEVICE_ABORT mode to signal a driver's DeviceProc that it must
reset the hardware if needed but do nothing else. An actual HW reset is only
required for some drivers dealing with the HW directly.
This is largely backwards-compatible, hence the input ABI minor bump only.
Drivers we care about either return BadValue on a mode that's not
DEVICE_{INIT|ON|OFF|CLOSE} or print an error and return BadValue. Exception
here is vmmouse, which currently ignores it and would not reset anything.
This should be fixed if the reset is required.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Keith Packard <keithp@keithp.com>
Signed-off-by: Keith Packard <keithp@keithp.com>
If acpid sends a string in a format that we can't parse, bail out instead of
potentially dereferencing a NULL-pointer.
X.Org Bug 73227 <http://bugs.freedesktop.org/show_bug.cgi?id=73227>
Signed-off-by: Ted Felix <ted@tedfelix.com>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Call find_header first, returning on failure before calling malloc.
Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
Found by parfait 1.1 memory analyser:
Memory leak of pointer 'pAdapt' allocated with malloc((88 * num_adaptors))
at line 162 of hw/xfree86/common/xf86xvmc.c in function 'xf86XvMCScreenInit'.
'pAdapt' allocated at line 158 with malloc((88 * num_adaptors)).
Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
Also avoids leaving invalid pointers in structures if realloc had to
move them elsewhere to make them larger.
Found by parfait 1.1 code analyzer:
Memory leak of pointer 'newCallbacks' allocated with realloc(((char*)offman->FreeBoxesUpdateCallback), (8 * (offman->NumCallbacks + 1)))
at line 328 of hw/xfree86/common/xf86fbman.c in function 'localRegisterFreeBoxCallback'.
'newCallbacks' allocated at line 320 with realloc(((char*)offman->FreeBoxesUpdateCallback), (8 * (offman->NumCallbacks + 1))).
newCallbacks leaks when newCallbacks != NULL at line 327.
Memory leak of pointer 'newPrivates' allocated with realloc(((char*)offman->devPrivates), (8 * (offman->NumCallbacks + 1)))
at line 328 of hw/xfree86/common/xf86fbman.c in function 'localRegisterFreeBoxCallback'.
'newPrivates' allocated at line 324 with realloc(((char*)offman->devPrivates), (8 * (offman->NumCallbacks + 1))).
newPrivates leaks when newCallbacks == NULL at line 327.
Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
Reported by parfait 1.1 code analyzer:
Error: Null pointer dereference (CWE 476)
Read from null pointer 'p'
at line 746 of hw/xfree86/common/xf86Option.c in function 'xf86TokenToOptName'.
Function 'xf86TokenToOptinfo' may return constant 'NULL' at line 721, called at line 745.
Null pointer introduced at line 721 in function 'xf86TokenToOptinfo'.
Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
Our in-house parfait 1.1 code analysis tool complained that every exit
path from xf86ValidateModes() in hw/xfree86/common/xf86Mode.c leaks the
storeClockRanges allocation made at line 1501 with XNFalloc.
Investigating, it seems that this code to copy the clock range list to
the clockRanges list in the screen pointer is just plain insane, and
according to git, has been since we first imported it from XFree86.
We start at line 1495 by walking the linked list from scrp->clockRanges
until we find the end. But that was just a diversion, since we've found
the end and immediately forgotten it, and thus at 1499 we know that
storeClockRanges is NULL, but that's not a problem since we're going to
immediately overwrite that value as the first thing in the loop.
So we move on through this loop at 1499, which takes us through the
linked list from the clockRanges variable, and for every entry in
that list allocates a new structure and copies cp to it. If we've
not filled in the screen's clockRanges pointer yet, we set it to
the first storeClockRanges we copied from cp. Otherwise, as best
I can tell, we just drop it into memory and let it leak away, as
parfait warned.
And then we hit the loop action, which if we haven't hit the end of
the cp list, advances cp to the next item in the list, and then just
for the fun of it, also sets storeClockRanges to the ->next pointer it
has just copied from cp as well, even though it's going to overwrite
it as the very first instruction in the loop body.
v2: rewritten using nt_list_* macros from Xorg's list.h header
Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
The changes to miPointerSetPosition interface from int->double breaks
the SIS driver build, so time to bump this.
Signed-off-by: Dave Airlie <airlied@redhat.com>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
DGA only handles master devices but it does intercept slave device events as
well (since the event handlers are per event type, not per device).
The DGA code must thus call into UpdateDeviceState to reset the button/key
state on the slave device before it discards the remainder of the event.
Test case:
- Passive GrabModeSync on VCP
- Press button
- Enable DGA after ButtonPress
- AllowEvents(SyncPointer)
- Release button
The button release is handled by DGAProcessPointerEvent but the device state
is never updated, so the slave ends up with the button permanently down.
And since the master's button state is the union of the slave states, the
master has the button permanently down.
X.Org Bug 59100 <http://bugs.freedesktop.org/show_bug.cgi?id=59100>
Reported-by: Steven Elliott <selliott4@austin.rr.com>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Dave Airlie <airlied@redhat.com>
Reported-by: Steven Elliott <selliott4@austin.rr.com>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Dave Airlie <airlied@redhat.com>
Instead of requiring every mode set to complete successfully, start up
as long as at least one CRTC is working. This avoids failures when one
or more CRTCs can't start due to mode setting conflicts.
Signed-off-by: Keith Packard <keithp@keithp.com>
Reviewed-by: Dave Airlie <airlied@redhat.com>
So the kernel removes the device, and the driver processes the first
udev event, and gets no output back from the kernel, so it check
and don't fall over.
This fixes a couple of crashes seen when hotplugging USB devices.
Signed-off-by: Dave Airlie <airlied@redhat.com>
This commit regresses dri1 since it moves the drmSetServerInfo from being
called at module load time to extension init time. However DRIScreenInit
relies on this being called before it gets control.
This patches moves the call into DRIScreenInit and seems to work here.
Signed-off-by: Dave Airlie <airlied@redhat.com>
The formatter confused address operators preceded by casts with
bitwise-and expressions, placing spaces on either side of both.
That syntax isn't used by ordinary address operators, however,
so fix them for consistency.
Signed-off-by: Yaakov Selkowitz <yselkowitz@users.sourceforge.net>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
Recent versions of the X server no longer provide this function, which
has been obsolete for over 2 years now.
Signed-off-by: Thierry Reding <thierry.reding@avionic-design.de>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
Following commit 37d956e3ac
Author: Keith Packard <keithp@keithp.com>
Date: Mon Sep 10 11:14:20 2012 +1000
xf86: fix compat output selection for no output GPUs
headless servers can no longer startup as we no longer select a compat
output for the fake framebuffer.
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=56343
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Dave Airlie <airlied@redhat.com>
Signed-off-by: Keith Packard <keithp@keithp.com>
For non-PCI video devices, such as those found on many ARM embedded
systems, the X server currently requires the BusID option to specify the
full path to the DRM device's sysfs node in order to properly match it
against the probed platform devices.
In order to allow X to start up properly if either the BusID option was
omitted or no configuration is present at all, the first video device is
used by default.
Signed-off-by: Thierry Reding <thierry.reding@avionic-design.de>
Reviewed-by: Dave Airlie <airlied@redhat.com>
Signed-off-by: Keith Packard <keithp@keithp.com>
K_OFF is a slightly broken interface, since if some other process
(cough, systemd) sets the console state to K_UNICODE then it undoes
K_OFF, and now Alt-F2 will switch terminals instead of summoning the
Gnome "run command" dialog.
KDSKBMUTE separates the "don't enqueue events" logic from the keymap, so
doesn't have this problem. Try it first, then continue falling back to
older methods.
Bugzilla: https://bugzilla.redhat.com/show_bug.cgi?id=859485
Tested-by: Josh Boyer <jwboyer@redhat.com>
Signed-off-by: Adam Jackson <ajax@redhat.com>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
These functions are already declared in <X11/fonts/fontproto.h>.
Redeclaring them just for _X_EXPORT causes tons of warnings throughout
xserver, but they need to be declared somewhere to be picked up by
sdksyms.sh. Doing so in a private header limits the warnings to
sdksyms.c; fixing those as well would require changes to fontsproto.
Signed-off-by: Yaakov Selkowitz <yselkowitz@users.sourceforge.net>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
Recent Linux kernels reworked the linux/input.h header file, which is
now part of the "user-space API". The include guard therefore has an
additional additional _UAPI prefix.
Instead of adding another case to the #ifdef, drop any include guard
checks and instead always undefine the BUS_* definitions on Linux.
Signed-off-by: Thierry Reding <thierry.reding@avionic-design.de>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Fix compilation of Xorg DDX without XF86VIDMODE since 6e74fdda, by putting
xf86vmode.c back under the XF86VIDMODE automake conditional it was accidentally
taken out of.
Signed-off-by: Jon TURNEY <jon.turney@dronecode.org.uk>
Tested-by: Yaakov Selkowitz <yselkowitz@users.sourceforge.net>
Reviewed-by: Daniel Stone <daniel@fooishbar.org>
This call is required for external drivers (specifically NVIDIA) that do
not share the xfree86 infrastructure to update the desktop dimensions.
Without it, the driver would update the ScreenRecs but not update the total
dimensions the input code relies on for transformation.
This call is a thin wrapper around the already-existing internal call and
should be backported to all stable series servers, with the minor ABI bump.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
CC: Andy Ritger <aritger@nvidia.com>
Reviewed-by: Aaron Plattner <aplattner@nvidia.com>
I noticed that the build-in int10 driver always reports
"Unable to retrieve all of segment 0x0C0000."
even though the entire BIOS data is retrieved with success.
The associated code is in hw/xfree86/int10/generic.c, in the function
xf86ExtendedInitInt10():
if (pci_device_read_rom(pInt->dev, vbiosMem) < V_BIOS_SIZE) {
xf86DrvMsg(screen, X_WARNING,
"Unable to retrieve all of segment 0x0C0000.\n");
}
The function pci_device_read_rom() is from libpciaccess; its return
value is not a size but an error status code: 0 means success.
If pci_device_read_rom() returns 0 for success, the warning is generated.
The proposed patch corrects the evaluation of the return value of
pci_device_read_rom() and of the supplied BIOS size.
Debian bug#686153
Signed-off-by: Julien Cristau <jcristau@debian.org>
Reviewed-by: Dave Airlie <airlied@redhat.com>
Signed-off-by: Keith Packard <keithp@keithp.com>
Commit 09e4b78f missed a case.
Signed-off-by: Jason Gerecke <killertofu@gmail.com>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Remove any reference to mibstore.h and miInitializeBackingStore() from
the documentation.
Reviewed-by: Aaron Plattner <aplattner@nvidia.com>
Reviewed-by: Adam Jackson <ajax@redhat.com>
Signed-off-by: Keith Packard <keithp@keithp.com>
Remove more backing store leftovers.
Reviewed-by: Aaron Plattner <aplattner@nvidia.com>
Reviewed-by: Adam Jackson <ajax@redhat.com>
Signed-off-by: Keith Packard <keithp@keithp.com>
This is a really awkward interface, since we're calling it well before
the driver knows what device it's going to drive. Drivers with both KMS
and UMS support therefore don't know whether to say they need I/O port
access or not, and have to assume they do.
With this change we now call it only to query whether port access might
be needed; we don't use that to determine whether to call a driver's
probe function or not, instead we call them unconditionally. If the
driver doesn't check whether port access was enabled, they might crash
ungracefully. To accomodate this, we move xorgHWAccess to be explicitly
intentionally exported (sigh xf86Priv.h) so that drivers can check that
before they attempt port access.
v2: Move initial xf86EnableIO() nearer the logic that determines whether
to call it, suggested by Simon Farnsworth.
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Simon Farnsworth <simon.farnsworth@onelan.co.uk>
Signed-off-by: Adam Jackson <ajax@redhat.com>
We load the driver list, then enable I/O, then call driver probe based
on whether I/O enable succeeded. That's bad, because the loaded
security policy might forbid port access. We happen to treat that as
fatal for some reason, which means even drivers that don't need I/O
access (like kms and fbdev) don't get the chance to run. Facepalm.
How about we just make that non-fatal instead, that sounds like a much
better plan.
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Simon Farnsworth <simon.farnsworth@onelan.co.uk>
Signed-off-by: Adam Jackson <ajax@redhat.com>
Pull platform methods into their own sections for legibility, and
rewrite the ifdefs to be more concise.
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Simon Farnsworth <simon.farnsworth@onelan.co.uk>
Signed-off-by: Adam Jackson <ajax@redhat.com>
If we are not seat 0 the following apply:
don't probe any bus other than platform
don't probe any drivers other than platform
assume the first platform device we match on the bus is the primary GPU.
This just adds checks in the correct places to ensure this, and
with this X can now start on a secondary seat for an output device.
v2: fix Seat0 macros
Reviewed-by: Keith Packard <keithp@keithp.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
This solves a race if we are trying to dynamically power off
secondary GPUs. Its not the greatest fix ever but it probably
as good as we can do for now.
The GPU probing causes the devices to be powered up, then when
we scan the PCI bus we get the correct information from the kernel,
rather than a bunch of 0xff due to the device being powered off.
drop gratuitous '&'.
Reviewed-by: Keith Packard <keithp@keithp.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
After we share the pixmap, the backing storage may have changed,
and we need to invalidate and buffers pointing at it.
This fixes GL compositors and prime windows lacking contents initially.
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Dave Airlie <airlied@redhat.com>
Otherwise we can't do fast user switch properly for multiple GPUs.
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
This is set by pre_init not screen init, so if we free it here
and then recycle the server, we lose all the providers.
I think we need to wrap FreeScreen here to do this properly,
will investigate for 1.14 most likely, safer to just leak this
on server exit for now.
Reviewed-by: Keith Packard <keithp@keithp.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
Commit 9d457f9c55 added an array of
DevPrivateSetRec structures in the middle of the ScreenRec, which throws off
extension modules trying to call things like pScreen->DestroyPixmap.
Signed-off-by: Aaron Plattner <aplattner@nvidia.com>
Signed-off-by: Keith Packard <keithp@keithp.com>
The core RandR screen cleanup now involves cleaning up any GPU screen
associations, and those call down into DDX to clean up the driver. If
the pointers from the xf86 structures back to the core randr
structures are set to NULL at that point, bad things happen.
This patch "knows" that the core RandR close screen is underneath the
xf86 randr close screen function, and so makes sure it gets called
first.
Signed-off-by: Keith Packard <keithp@keithp.com>
Reviewed-by: Dave Airlie <airlied@redhat.com>
These are two minor changes, one to reset the pointer to NULL,
after freeing the pixmaps, one to make sure we use the right API for
the master pixmap, though I doubt it'll ever really matter.
Signed-off-by: Dave Airlie <airlied@redhat.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
When the drawable disappears we need to free the prime master/slave combos.
This fixes a leak after a prime app is run.
Signed-off-by: Dave Airlie <airlied@redhat.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
The indenter seems to have gotten confused by initializing arrays of
structs with the struct defined inline - for predefined structs it did
a better job, so match that.
Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>
The RandR CRTC structures are freed when their resource IDs are
destroyed during server shut down, which is before the screen is
closed. Calling back into RandR with stale pointers just segfaults the
server.
Signed-off-by: Keith Packard <keithp@keithp.com>
Reviewed-by: Dave Airlie <airlied@redhat.com>
Tested-by: Knut Petersen <knut_petersen@t-online.de>
Panning is at odds with CRTC cursor confinement. This disables CRTC cursor
confinement as long as panning is enabled.
Fixes regression introduced in 56c90e29f0.
Reviewed-by: Adam Jackson <ajax@redhat.com>
Signed-off-by: Rui Matos <tiagomatos@gmail.com>
Signed-off-by: Keith Packard <keithp@keithp.com>
Let's say - purely for the sake of argument, mind you - that you had a
server GPU with anemic memory bandwidth, and you walked up to it and
plugged in a monitor that was 1920x1080 because that's what happened to
be on the crash cart. Say the memory bandwidth is such that anything
larger than 1280x1024 gets filtered away. Now you're in trouble,
because the established timings section includes a 720x400 mode because
that's what DOS 80x25 is, and that happens to just about match the
physical aspect ratio.
Instead let's reuse the logic from the existing aspect-match path: pick
the larger mode of either the physical aspect ratio or 4:3.
Reviewed-by: Julien Cristau <jcristau@debian.org>
Signed-off-by: Adam Jackson <ajax@redhat.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Keith Packard <keithp@keithp.com>
We've had reports of two copies of the GLX bits, one in the server
and one in libglx.so causing problems, I didn't understand why the
X server needed a copy so drop it, however then we have to fix a missing
GlxExtensionInit that comes from sdksyms, so work around it by moving
that one declaration into a header that sdksyms doesn't scan.
Thanks to Jon Turney for debugging the actual problem.
(copyright header from extinit.h that seems most appropriate put on top).
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=52402
Tested-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Daniel Stone <daniel@fooishbar.org>
Tested-by: Jon TURNEY <jon.turney@dronecode.org.uk>
Signed-off-by: Dave Airlie <airlied@redhat.com>
Signed-off-by: Keith Packard <keithp@keithp.com>
Signed-off-by: Adam Jackson <ajax@redhat.com>
Reviewed-by: Dave Airlie <airlied@redhat.com>
Reviewed-by: Alan Coopersmith <alan.coopersmith@oracle.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Daniel Stone <daniel@fooishbar.org>
Acked-by: Aaron Plattner <aplattner@nvidia.com>
Signed-off-by: Keith Packard <keithp@keithp.com>
Signed-off-by: Adam Jackson <ajax@redhat.com>
Reviewed-by: Dave Airlie <airlied@redhat.com>
Reviewed-by: Alan Coopersmith <alan.coopersmith@oracle.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Daniel Stone <daniel@fooishbar.org>
Acked-by: Aaron Plattner <aplattner@nvidia.com>
Signed-off-by: Keith Packard <keithp@keithp.com>
This allows the driver to operate as an output slave.
It adds scan out pixmap, and the capability
checks to make sure they available.
Signed-off-by: Dave Airlie <airlied@redhat.com>
DRI2GetParam was going through review in parallel with main batch of
C99 initialization changes - sync up now that both have landed.
Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>
Reviewed-by: Keith Packard <keithp@keithp.com>
These flags were unexported by commit a1d41e311c,
which moved the declarations around and lost the _X_EXPORT attributes in the
process. Since drivers need these and it's late in the release cycle, just
re-export them for now.
Signed-off-by: Aaron Plattner <aplattner@nvidia.com>
Reviewed-by: Andy Ritger <aritger@nvidia.com>
Signed-off-by: Keith Packard <keithp@keithp.com>
This fixes an implicit declaration,
xf86AutoConfig.c:202:5: error: implicit declaration of function 'xf86PlatformMatchDriver' [-Werror=implicit-function-declaration]
xf86AutoConfig.c:202:5: warning: nested extern declaration of 'xf86PlatformMatchDriver' [-Wnested-externs]
Signed-off-by: Dave Airlie <airlied@redhat.com>
Reviewed-by: Keith Packard <keithp@keithp.com>
Same as DRI2CreateDrawable, except it can return the DRI2 specific XID of the
DRI2 drawable reference to the base drawable.
Signed-off-by: Michel Dänzer <michel.daenzer@amd.com>
Reviewed-by: Kristian Høgsberg <krh@bitplanet.net>
Signed-off-by: Keith Packard <keithp@keithp.com>
I hate this [redacted] script.
Signed-off-by: Daniel Stone <daniel@fooishbar.org>
Tested-by: Keith Packard <keithp@keithp.com>
Signed-off-by: Keith Packard <keithp@keithp.com>
Based on the original patch by Chris Wilson, which was a better fix than mine.
We stash a copy of the desiredMode on the crtc so that we can restore it
after a vt switch. This copy is a simple memcpy and so also stashes a
references to the pointers contained within the desiredMode. Those
pointers are freed the next time the outputs are probed and mode list
rebuilt, resulting in us chasing those dangling pointers on the next
mode switch.
==22787== Invalid read of size 1
==22787== at 0x40293C2: __GI_strlen (in
/usr/lib/valgrind/vgpreload_memcheck-amd64-linux.so)
==22787== by 0x668F875: strdup (strdup.c:42)
==22787== by 0x5DBA00: XNFstrdup (utils.c:1124)
==22787== by 0x4D72ED: xf86DuplicateMode (xf86Modes.c:209)
==22787== by 0x4CA848: xf86CrtcSetModeTransform (xf86Crtc.c:276)
==22787== by 0x4D05B4: xf86SetDesiredModes (xf86Crtc.c:2677)
==22787== by 0xA7479D0: sna_create_screen_resources
(sna_driver.c:220)
==22787== by 0x4CB914: xf86CrtcCreateScreenResources (xf86Crtc.c:725)
==22787== by 0x425498: main (main.c:216)
==22787== Address 0x72c60e0 is 0 bytes inside a block of size 9 free'd
==22787== at 0x4027AAE: free (in
/usr/lib/valgrind/vgpreload_memcheck-amd64-linux.so)
==22787== by 0x4A547E: xf86DeleteMode (xf86Mode.c:1984)
==22787== by 0x4CD84F: xf86ProbeOutputModes (xf86Crtc.c:1578)
==22787== by 0x4DC405: xf86RandR12GetInfo12 (xf86RandR12.c:1537)
==22787== by 0x518119: RRGetInfo (rrinfo.c:202)
==22787== by 0x51D997: rrGetScreenResources (rrscreen.c:335)
==22787== by 0x51E0D0: ProcRRGetScreenResources (rrscreen.c:475)
==22787== by 0x513852: ProcRRDispatch (randr.c:493)
==22787== by 0x4346DB: Dispatch (dispatch.c:439)
==22787== by 0x4256E4: main (main.c:287)
Signed-off-by: Maarten Lankhorst <maarten.lankhorst@canonical.com>
Reported-by: Zdenek Kabelac <zdenek.kabelac@gmail.com>
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=36108
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Keith Packard <keithp@keithp.com>
No drivers used this, so it got unexported, and now it's so unused it
got culled during the link. Take the poor function out behind the shed
and put it out of its misery.
Signed-off-by: Daniel Stone <daniel@fooishbar.org>
Reviewed-by: Adam Jackson <ajax@redhat.com>
Acked-by: Peter Hutterer <peter.hutterer@who-t.net>
Signed-off-by: Keith Packard <keithp@keithp.com>
Commit 0c6987df in June 2008 disabled XAA offscreen pixmaps per default,
as they were broken, leaving XAA only able to accelerate operations
directly on the screen pixmap and nowhere else, eliminating acceleration
for basically every modern toolkit, and any composited environment.
So, it hasn't worked for over four years. No-one's even come close to
fixing it.
RIP.
Signed-off-by: Daniel Stone <daniel@fooishbar.org>
Reviewed-by: Dave Airlie <airlied@redhat.com>
Reviewed-by: Alex Deucher <alexdeucher@gmail.com>
setupFunc was used as an early callback for half-modular extensions such
as Xv, XvMC and DGA to set up hooks between the core server and the
modular component. Now we've rid ourselves of that, we can also bin
setupFunc.
Signed-off-by: Daniel Stone <daniel@fooishbar.org>
Reviewed-by: Jamey Sharp <jamey@minilop.net>
Signed-off-by: Keith Packard <keithp@keithp.com>
Rather than having a non-Xorg and an Xorg-specific path which basically
just duplicated each other for no reason, we could ... just have one.
Signed-off-by: Daniel Stone <daniel@fooishbar.org>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
Signed-off-by: Keith Packard <keithp@keithp.com>
There was nothing XFree86-specific or loader-specific about this, aside
from using xf86MsgVerb instead of ErrorF.
Signed-off-by: Daniel Stone <daniel@fooishbar.org>
Reviewed-by: Jamey Sharp <jamey@minilop.net>
Signed-off-by: Keith Packard <keithp@keithp.com>
In preparation for gutting loadext.c, move the ExtensionModule struct to
the DIX, and unexport ExtensionModuleList (why, why, why, why was this
ever exported in the first place, tbqh).
Signed-off-by: Daniel Stone <daniel@fooishbar.org>
Reviewed-by: Jamey Sharp <jamey@minilop.net>
Signed-off-by: Keith Packard <keithp@keithp.com>
Extensions could previously declare initialisation dependencies on other
extensions, which would then get nicely sorted by the loader. We only
had one user for this, GLX, which had one pointless (Composite) and one
possibly useful dependency (DBE). As DBE is now a built-in, it will
always be sorted by GLX, so we no longer have any users for it.
Signed-off-by: Daniel Stone <daniel@fooishbar.org>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
Signed-off-by: Keith Packard <keithp@keithp.com>
GLX was the only user of extension init order dependencies, using them
to depend on Composite, which has always been built-in anyway, and DBE,
which is now built-in.
Signed-off-by: Daniel Stone <daniel@fooishbar.org>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
Signed-off-by: Keith Packard <keithp@keithp.com>
Make sure we add static extensions before anything in a module. This is
more or less a no-op at the moment, but will come in handy later when
extension dependency sorting is removed.
Signed-off-by: Daniel Stone <daniel@fooishbar.org>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
Signed-off-by: Keith Packard <keithp@keithp.com>
Instead of keeping a tiny amount of code in an external module, just man
up and build it into the core server.
v2: Fix test/Makefile.am to only link libdri2.la if DRI2 is set
Signed-off-by: Daniel Stone <daniel@fooishbar.org>
Reviewed-by: Jamey Sharp <jamey@minilop.net>
Signed-off-by: Keith Packard <keithp@keithp.com>
DRI2DestroyDrawable() was still being _X_EXPORTed, but hasn't existed
since 1da1f33f last year.
Signed-off-by: Daniel Stone <daniel@fooishbar.org>
Reviewed-by: Cyril Brulebois <kibi@debian.org>
Reviewed-by: Jamey Sharp <jamey@minilop.net>
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Signed-off-by: Keith Packard <keithp@keithp.com>
AM_CFLAGS will suffice, given we only have one target in this directory.
Signed-off-by: Daniel Stone <daniel@fooishbar.org>
Reviewed-by: Cyril Brulebois <kibi@debian.org>
Signed-off-by: Keith Packard <keithp@keithp.com>
Rather than building the tiny amount of code required for XFree86-DRI as
an external module, build it in if it's enabled at configure time.
v2: Fix test/Makefile.am to only link libdri.la if DRI is set
Signed-off-by: Daniel Stone <daniel@fooishbar.org>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
Signed-off-by: Keith Packard <keithp@keithp.com>
fixup for DRI1 move
Signed-off-by: Keith Packard <keithp@keithp.com>
extmod was originally a big pointless module. Now it's an empty,
pointless module. This commit makes it unexist.
Signed-off-by: Daniel Stone <daniel@fooishbar.org>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Signed-off-by: Keith Packard <keithp@keithp.com>
As with DGA, move VidMode from being part of extmod to a built-in part
of the server, if compiled as such. This is initialised from
xf86ExtensionInit rather than miinitext because it's wholly dependent on
the Xorg DDX.
Signed-off-by: Daniel Stone <daniel@fooishbar.org>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
The DGA event base used to have to be passed through a function pointer,
as the code was cleaved in two with half in a module, and half in the
core server. Now that's not the case, just access DGAEventBase
directly.
v2: Deal with Alan's event initialization cleanups
Signed-off-by: Tomas Carnecky <tom@dbservice.com>
Reviewed-by: Daniel Stone <daniel@fooishbar.org>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Signed-off-by: Keith Packard <keithp@keithp.com>
Rather than leave DGA languishing in extmod, move it to be a built-in
extension. As it's quite specific to the Xorg DDX, just move it
sideways to the rest of the DGA code in hw/xfree86/common, and
initialise it from xf86ExtensionInit, rather than miinitext.
Signed-off-by: Daniel Stone <daniel@fooishbar.org>
Acked-by: Keith Packard <keithp@keithp.com>
Signed-off-by: Keith Packard <keithp@keithp.com>
Instead of letting it languish in extmod just because we want to
configure bits of it from xf86, move XSELinux to the builtin part of
Xext, and do its configuration from xf86ExtensionInit.
Signed-off-by: Daniel Stone <daniel@fooishbar.org>
Reviewed-by: Jamey Sharp <jamey@minilop.net>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Signed-off-by: Keith Packard <keithp@keithp.com>
Xv used to call XvScreenInit and co. through function pointers, as
XvScreenInit may have been sitting on the other side of a module
boundary from xf86XvScreenInit. Why this was so is a mystery, but make
it not so any more.
Signed-off-by: Tomas Carnecky <tom@dbservice.com>
Reviewed-by: Daniel Stone <daniel@fooishbar.org>
Acked-by: Peter Hutterer <peter.hutterer@who-t.net>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Signed-off-by: Keith Packard <keithp@keithp.com>
We no longer have anything in the tree that checks for XorgLoader. This
was a fairly monumental hack: xvdi.h used to hide all its functions
behind #ifndef XorgLoader, solely to avoid sdksyms.sh picking up its
symbols, as it was previously a module rather than built-in.
This is no longer the case, so we can remove the define.
Signed-off-by: Daniel Stone <daniel@fooishbar.org>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
Signed-off-by: Keith Packard <keithp@keithp.com>
Always build these extensions into the core server, rather than letting
them languish in extmod.
Signed-off-by: Tomas Carnecky <tom@dbservice.com>
Reviewed-by: Daniel Stone <daniel@fooishbar.org>
Reviewed-by: Jamey Sharp <jamey@minilop.net>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Signed-off-by: Keith Packard <keithp@keithp.com>
Always build XRes support into the core server, rather than letting it
languish in extmod.
Signed-off-by: Tomas Carnecky <tom@dbservice.com>
Reviewed-by: Daniel Stone <daniel@fooishbar.org>
Reviewed-by: Jamey Sharp <jamey@minilop.net>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Signed-off-by: Keith Packard <keithp@keithp.com>
Always build DPMS support into the core server, rather than letting it
languish in extmod.
Signed-off-by: Tomas Carnecky <tom@dbservice.com>
Reviewed-by: Daniel Stone <daniel@fooishbar.org>
Reviewed-by: Jamey Sharp <jamey@minilop.net>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Signed-off-by: Keith Packard <keithp@keithp.com>
If we've built MIT-SCREEN-SAVER support, then just build it into the
main binary, rather than leaving it in extmod.
Signed-off-by: Tomas Carnecky <tom@dbservice.com>
Reviewed-by: Daniel Stone <daniel@fooishbar.org>
Reviewed-by: Jamey Sharp <jamey@minilop.net>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Signed-off-by: Keith Packard <keithp@keithp.com>
Rather than languishing in its own special module, move RECORD into the
core server.
Signed-off-by: Tomas Carnecky <tom@dbservice.com>
Reviewed-by: Daniel Stone <daniel@fooishbar.org>
Reviewed-by: Jamey Sharp <jamey@minilop.net>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Signed-off-by: Keith Packard <keithp@keithp.com>
If DBE support is compiled in the server, just man up and build it into
the server, rather than having it as an external module.
Signed-off-by: Tomas Carnecky <tom@dbservice.com>
Reviewed-by: Daniel Stone <daniel@fooishbar.org>
Reviewed-by: Jamey Sharp <jamey@minilop.net>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Signed-off-by: Keith Packard <keithp@keithp.com>
When the array gets down to size zero (which it does in later patches),
gcc complains that the index is out of bounds. Avoid this by using
ARRAY_SIZE on extensionModules instead.
Signed-off-by: Daniel Stone <daniel@fooishbar.org>
Reviewed-by: Stéphane Marchesin <stephane.marchesin@gmail.com>
Signed-off-by: Keith Packard <keithp@keithp.com>
xf86ExtensionInit is called after configuration file parsing, so it can
perform the two parts of extension initialisation currently done by
extmod: enabling and disabling of extensions through an 'omit' option,
and SELinux configuration.
Signed-off-by: Daniel Stone <daniel@fooishbar.org>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Signed-off-by: Keith Packard <keithp@keithp.com>
EXTERN_MODULE was used to specify that we shouldn't worry about modules
lacking a ModuleData object. It was also completely unused. *shrug*
Signed-off-by: Daniel Stone <daniel@fooishbar.org>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Alan Coopersmith <alan.coopersmith@oracle.com>
Signed-off-by: Keith Packard <keithp@keithp.com>
Create extinit.h (and xf86Extensions.h, for Xorg-specific extensions) to
hold all our extension initialisation prototypes, rather than
duplicating them everywhere.
Signed-off-by: Daniel Stone <daniel@fooishbar.org>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
Signed-off-by: Keith Packard <keithp@keithp.com>
Rather than making poor old miinitext.c do it, including making DMX
have fake symbols just to keep it happy.
Signed-off-by: Daniel Stone <daniel@fooishbar.org>
Reviewed-by: Stéphane Marchesin <stephane.marchesin@gmail.com>
Signed-off-by: Keith Packard <keithp@keithp.com>
INITARGS was a hardcoded define to void. Since knowing the function
signature for your extensions is kinda useful, just replace it with a
hardcoded void, but leave the define there for API compatibility.
Signed-off-by: Tomas Carnecky <tom@dbservice.com>
Reviewed-by: Daniel Stone <daniel@fooishbar.org>
Reviewed-by: Cyril Brulebois <kibi@debian.org>
Reviewed-by: Jamey Sharp <jamey@minilop.net>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Signed-off-by: Keith Packard <keithp@keithp.com>
Huh, so I guess INITARGS used to be int argc, char *argv then. Either
way, it's now void, so fix that ...
Signed-off-by: Daniel Stone <daniel@fooishbar.org>
Reviewed-by: Cyril Brulebois <kibi@debian.org>
Reviewed-by: Jamey Sharp <jamey@minilop.net>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Signed-off-by: Keith Packard <keithp@keithp.com>
libxorgxkb.a contains a number of libraries which are used by XKB action
code to call back into the DDX, e.g. for VT switching, termination, grab
breaking, et al. Make sure libxkb.a comes first in the link order, so
it can mark XkbDDX* as used in order for the linker to not discard them.
Signed-off-by: Daniel Stone <daniel@fooishbar.org>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
Signed-off-by: Keith Packard <keithp@keithp.com>
Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>
Reviewed-by: Keith Packard <keithp@keithp.com>
Tested-by: Daniel Stone <daniel@fooishbar.org>
Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>
Reviewed-by: Keith Packard <keithp@keithp.com>
Tested-by: Daniel Stone <daniel@fooishbar.org>
Ensures padding bytes are zero-filled
Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>
Reviewed-by: Keith Packard <keithp@keithp.com>
Tested-by: Daniel Stone <daniel@fooishbar.org>
Seems silly waiting to check if the client failed to send us enough bytes
until after we've already tried using them.
Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>
Reviewed-by: Keith Packard <keithp@keithp.com>
Tested-by: Daniel Stone <daniel@fooishbar.org>
When passing variable pointers to functions or otherwise doing long
sequences to compute values for replies, create & use some new
temporary variables, to allow for simpler initialization of reply
structures in the following patches.
Move memsets & other initializations to group with the rest of the
filling in of the reply structure, now that they're not needed so
early in the code path.
Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>
Reviewed-by: Keith Packard <keithp@keithp.com>
Tested-by: Daniel Stone <daniel@fooishbar.org>
Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>
Reviewed-by: Keith Packard <keithp@keithp.com>
Tested-by: Daniel Stone <daniel@fooishbar.org>
Casting return to (void) was used to tell lint that you intended
to ignore the return value, so it didn't warn you about it.
Casting the third argument to (char *) was used as the most generic
pointer type in the days before compilers supported C89 (void *)
(except for a couple places it's used for byte-sized pointer math).
Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>
Reviewed-by: Keith Packard <keithp@keithp.com>
Tested-by: Daniel Stone <daniel@fooishbar.org>
Each DDX currently calls OsReleaseSIGIO in case it was suspended when
the server regen started. This causes a BUG to occur if SIGIO was
*not* blocked at that time. Instead of relying on each DDX, make the
OS layer reliably reset all signal state at server init time, ensuring
that signals are suitably unblocked and that the various signal state
counting variables are set back to zero.
Signed-off-by: Keith Packard <keithp@keithp.com>
Otherwise, OsReleaseSIGIO will complain, or perhaps something worse
will happen (if SIGIO actually needs to be blocked here).
Signed-off-by: Keith Packard <keithp@keithp.com>
Add a static mask of prime id and allocate them at screen time,
if the driver supports the prime interfaces and is a gpu screen.
This is instead of them changing due to user controlled randr commands,
as suggested by Keith.
Reviewed-by: Keith Packard <keithp@keithp.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
This adds the initial prime support for dri2 offload. The main thing is
when we get a connection from a prime client, we stored the information
and mark all drawables from that client as prime. We then create all
buffers for that drawable on the prime device dri2screen.
Then DRI2UpdatePrime is provided which drivers can call to get a shared
pixmap which they can use as the front buffer. The driver is then
responsible for doing the back->front copy to the shared buffer.
prime requires a compositing manager be run, but it handles the case where
a window get un-redirected by allocating a new pixmap and pointing the crtc
at it while the client is in that state.
Currently prime can't handle pageflipping, so always does straight copy swap,
v1.1: renumber on top of master.
v1.2: fix auth on top of master.
Reviewed-by: Keith Packard <keithp@keithp.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
This adds support for setting the offload sink to the xf86 ddx.
Reviewed-by: Keith Packard <keithp@keithp.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
This is so we can tell the scanout pixmap has changed between calls
to the crtc set function.
Reviewed-by: Keith Packard <keithp@keithp.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
This fixes a segfault where this code believes we are outside the screen
boundaries on a slave device, but we aren't.
Reviewed-by: Keith Packard <keithp@keithp.com>
Reviewed-by: Adam Jackson <ajax@redhat.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
Current USB devices have no hw rendered cursors, so we need the
master GPU to render the cursor, so whenever we plug in a
slave device, fallback to sw rendered cursors.
Reviewed-by: Adam Jackson <ajax@redhat.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
This adds support for the randr callback for setting the output source
for a device.
v2: drop root clip change on detach
Reviewed-by: Keith Packard <keithp@keithp.com>
Reviewed-by: Adam Jackson <ajax@redhat.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
This will detach any scanout pixmaps attached to slave crtcs.
Reviewed-by: Keith Packard <keithp@keithp.com>
Reviewed-by: Adam Jackson <ajax@redhat.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
Add the simple passthrough interface for drivers to use,
so they can set scanout pixmaps.
v2: detach scanout pixmap properly.
Reviewed-by: Keith Packard <keithp@keithp.com>
Reviewed-by: Adam Jackson <ajax@redhat.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
v2: Split out DGAAvailable into two interfaces, one for calls from protocol
decoding and one for internal usage, after discussion with ajax and keithp.
Reviewed-by: Keith Packard <keithp@keithp.com>
Reviewed-by: Adam Jackson <ajax@redhat.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
This is the default attachment, unbound gpu screens get
attached to the 0 protocol screen.
detach on hotunplug.
v2: detach after tearing down crtc/outputs.
Reviewed-by: Keith Packard <keithp@keithp.com>
Reviewed-by: Adam Jackson <ajax@redhat.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
This adds the framework for DDX provider support.
v2: as per keithp's suggestion remove the xf86 provider object
and just store it in the toplevel object.
v3: update for new protocol
v4: drop devPrivate, free name
Reviewed-by: Keith Packard <keithp@keithp.com>
Reviewed-by: Adam Jackson <ajax@redhat.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
This adds the initial provider object and provider property
support to the randr dix code.
v2: destroy provider in screen close
v2.1: fix whitespace
v3: update for latest rev of protocol + renumber after 1.4 tearout.
v4: fix logic issue, thanks Samsagax on irc
v5: keithp's review: fix current_role, fix copyrights, fix master
reporting crtc/outputs.
v6: port to new randr interface, drop all set role bits for now
v7: drop devPrivate in provider, not needed, add BadMatch returns
for NULL SetProviderOffloadSink and SetProviderOutputSource, drop
the old typedef.
Reviewed-by: Keith Packard <keithp@keithp.com>
Reviewed-by: Adam Jackson <ajax@redhat.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
This loop needs to count from 7 to 0, not only from 7 to 1.
The current code always skips the modes {1152, 864, 75, 0}, {1280, 1024, 85, 0},
{1400, 1050, 75, 0}, {1600, 1200, 70, 0} and {1920, 1200, 60, 0}.
Signed-off-by: Torsten Kaiser <x11@ariolc.dyndns.org>
Reviewed-by: Adam Jackson <ajax@redhat.com>
Signed-off-by: Keith Packard <keithp@keithp.com>
Using -O3 gcc notes that m could reach beyound the end of the EstIIIModes array,
if the last bits of the 11s byte where set.
Fix this, by extending the array to cover all possible bits from est.
https://bugs.freedesktop.org/show_bug.cgi?id=45623
Signed-off-by: Torsten Kaiser <x11@ariolc.dyndns.org>
Reviewed-by: Adam Jackson <ajax@redhat.com>
Signed-off-by: Keith Packard <keithp@keithp.com>
This option is to stop the X server adding non-primary devices as
gpu screens.
v2: fix per Keith's suggestion.
Reviewed-by: Keith Packard <keithp@keithp.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
This provides add/remove support for platform devices at xfree86 ddx level.
v2: cleanup properly if no driver found.
v3: load the modesetting driver before checking driver list.
Reviewed-by: Keith Packard <keithp@keithp.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
This just adds the structures and interfaces required for adding/deleteing
gpu screens at the DDX level. The platform probe can pass a new flag
to the driver, so they can call xf86AllocateScreen and pass back the new
gpu screen flag.
It also calls the gpu screens preinit and screeninit routines at
startup.
v2: fix delete screen use after free.
v3: split out pScrn into separate patch
Reviewed-by: Keith Packard <keithp@keithp.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
This is in preparation for gpu screens in here, just use
a pScrn pointer to point at the new screen.
suggested by Keith.
v1.1: fix spacing as suggested by Aaron.
Reviewed-by: Keith Packard <keithp@keithp.com>
Reviewed-by: Aaron Plattner <aplattner@nvidia.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
On Linux in order for future hotplug work, we are required to interface
to udev to detect device creation/removal. In order to try and get
some earlier testing on this, this patch adds the ability to use
udev for device enumeration on Linux.
At startup the list of drm/kms devices is probed and this info is
used to load drivers.
A new driver probing method is introduced that passes the udev
device info to the driver for probing.
The probing integrates with the pci probing code and will fallback
to the pci probe and old school probe functions in turn.
The flags parameter to the probe function will be used later
to provide hotplug and gpu screen flags for the driver to behave
in a different way.
This patch changes the driver ABI, all drivers should at least
be set with a NULL udev probe function after this commit.
v2: rename to platform bus, now with 100% less udev specific,
this version passes config_odev_attribs around which are an array
of id/string pairs, then the udev code can attach the set of attribs
it understands, the OS specific code can attach its attrib, and then
the core/drivers can lookup the required attribs.
also add MATCH_PCI_DEVICES macro.
This version is mainly to address concerns raised by ajax.
v3: Address comments from Peter.
fix whitespace that snuck in.
rework to use a linked list with some core functions that
xf86 wraps.
v4: add free list, fix struct whitespace.
ajax this address most of your issues?
v5: drop probe ifdef, fix logic issue
v6: some overhaul after more testing.
Implement primaryBus for platform devices.
document hotplug.h dev attribs - drop sysname attrib
fix build with udev kms disabled
make probing work like the PCI probe code,
match against bus id if one exists, or primary device.
RFC: add new bus id support "PLAT:syspath". we probably
want to match on this a bit different, or use a different
property maybe. I was mainly wanting this for use with
specifying usb devices in xorg.conf directly, but PLAT:path
could also work I suppose.
v6.1: add missing noop platform function
v7: fix two interactions with pci probing and slot claiming, prevents
pci and platform trying to load two drivers for same slot.
v8: test with zaphod mode on -ati driver, fixup resulting issue
clean up common probe code into another function, change busid
matching to allow dropping end of strings.
v9: fix platform probing logic so it actually works.
v9.1: fix pdev init to NULL properly.
v10: address most of Keith's concerns.
v4 was thanks to Reviewed-by: Adam Jackson <ajax@redhat.com>
v5 was Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Keith Packard <keithp@keithp.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
This fixes some really ugly code that got mangled by the indenting.
Reviewed-by: Keith Packard <keithp@keithp.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
LoaderSymbol calls dlsym with RTLD_DEFAULT pseudo handle making it search in
every loaded library. In addition glibc adds NODELETE flag to the library
containing the symbol.
It's used in doLoadModule to locate <modulename>ModuleData symbol, the
module's library gets the flag and is kept in memory even after it is
unloaded.
This patch adds LoaderSymbolFromModule function that looks for symbol only in
library specified by handle. That way the NODELETE flag isn't added.
This glibc behavior doesn't seem to be documented, but even if other
implementations differ, there is no reason to search ModuleData symbol outside
the module's library.
Signed-off-by: Michal Srb <msrb@suse.com>
Reviewed-by: Daniel Stone <daniel@fooishbar.org>
v2: Switch LoaderSymbolFromModule arguments order.
Correct description.
Reviewed-by: Adam Jackson <ajax@redhat.com>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
*dev is the condition of the while loop we're in, reset to NULL after
freeing
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Dave Airlie <airlied@redhat.com>
The hooks are left for this cycle, we can drop it next cycle once the
drivers that need it (e.g. wacom) have been updated.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Keith Packard <keithp@keithp.com>
We can ignore the "wasset" argument now since the DIX will keep proper
refcounting.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Keith Packard <keithp@keithp.com>
This merge includes a minor fixup for '%p' arguments; must cast to
uintptr_t instead of uint64_t as we use -Werror=pointer-to-int-cast
which complains when doing a cast (even explicitly) from a pointer
to an integer of different size.
The input ABI hasn't changed, but input drivers need something to hook on if
they want to log from within signal handlers and the input ABI is the
simplest way of doing so.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Already treated as const anyway by all drivers.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Chase Douglas <chase.douglas@canonical.com>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
This corresponds to XListInputDevice(3)'s "type" field (after being
converted to an Atom). Input drivers use the XI_KEYBOARD and similar
defines, even Wacom which falls out of the common defines uses constant
strings here. The use-case for having this non-const is small.
Input ABI break technically, since we never freed this information anyway it
is not a noticable change.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Chase Douglas <chase.douglas@canonical.com>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
The function can be called from a fatal signal handler.
Signed-off-by: Chase Douglas <chase.douglas@canonical.com>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
The function may be called from a fatal signal handler.
Signed-off-by: Chase Douglas <chase.douglas@canonical.com>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
The function may be called from a fatal signal handler.
Signed-off-by: Chase Douglas <chase.douglas@canonical.com>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
The Solaris linker recently added a -z parent flag for easier checking
of symbol definitions in plugins against the program that loads them.
If that's present, this enables it, along with -z defs to error on
undefined symbols to alert us if any modules call symbols that won't
be found at runtime.
This builds upon, and requires, the recent Cygwin work to build Xorg.
It moves a couple more modules to be after the Xorg binary in the build
order so that they can find the binary to check against, much as the
Cygwin changes did (these would be modules built on Solaris but not
Cygwin).
v2: This version only sets the flags for the xorg-server build itself,
and does not yet export them in xorg-server.pc to the drivers, since
most of the drivers are not ready to build with -no-undefined yet.
Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>
Reviewed-by: Yaakov Selkowitz <yselkowitz@users.sourceforge.net>
This code existed in 3 different forms, perhaps it should be
consolidated.
Reviewed-by: Keith Packard <keithp@keithp.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
Signed-off-by: Keith Packard <keithp@keithp.com>
The legacy logic was embarassingly wrong; AuthMagic should return errno,
so returning FALSE only when AuthMagic returns nonzero is exactly wrong.
v2: Match drmAuthMagic by returning -EINVAL rather than EINVAL
Fix trailing whitespace
https://bugs.freedesktop.org/show_bug.cgi?id=51400
Signed-off-by: Christopher James Halse Rogers <christopher.halse.rogers@canonical.com>
Tested-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
Tested-by: Knut Petersen <knut_petersen@t-online.de>
Signed-off-by: Keith Packard <keithp@keithp.com>
Drivers call xf86InstallSIGIOHandler() for their fd on DEVICE_ON. That
function does not actually enable the signal if it was blocked to begin
with. As a result, if one vt-switches away from the server (SIGIO is
blocked) and then triggers a server regeneration, the signal remains
blocked and input devices are dead.
Avoid this by always unblocking SIGIO when we start the server.
X.Org Bug 50957 <http://bugs.freedesktop.org/show_bug.cgi?id=50957>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Keith Packard <keithp@keithp.com>
This will be used for checking for proper logging when in signal
context.
Signed-off-by: Chase Douglas <chase.douglas@canonical.com>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
xwayland drivers need access to their screen private data to authenticate.
Now that drivers no longer have direct access to the global screen arrays,
this needs to be passed in as function context.
v2: Don't break ABI
v3: Paint the bikeshed blue; drop fd from AuthMagic2ProcPtr prototype
Signed-off-by: Christopher James Halse Rogers <christopher.halse.rogers@canonical.com>
Reviewed-by: Kristian Høgsberg <krh@bitplanet.net>
Signed-off-by: Keith Packard <keithp@keithp.com>
Rob Clark got an ARM tinderbox up and running, and this code is built there
but not here, this should fix it, though I hope that code never gets executed.
Signed-off-by: Dave Airlie <airlied@redhat.com>
Signed-off-by: Keith Packard <keithp@keithp.com>
Removed flags argument to match api rework in commit 1f0e8bd5eb
Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>
Reviewed-by: Dave Airlie <airlied@redhat.com>
Actual inline implementations are provided via external *.il files
generated from *.S files in hw/xfree86/os-support/solaris
Fixes missing prototype warnings that xorg-macros has recently elevated
to build-breaking errors.
Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>
Signed-off-by: Henry Zhao <henry.zhao@oracle.com>
Reviewed-by: Dave Airlie <airlied@redhat.com>
Spotted these in tinderbox:
xf86xv.c: In function 'xf86XVScreenInit':
xf86xv.c:282: warning: assignment from incompatible pointer type
xf86xv.c: In function 'xf86XVCloseScreen':
xf86xv.c:1275: warning: assignment from incompatible pointer type
xf86xv.c: In function 'xf86XVAdjustFrame':
xf86xv.c:1366: warning: assignment from incompatible pointer type
Missed removing one flags.
Signed-off-by: Dave Airlie <airlied@redhat.com>
Tested-by: Keith Packard <keithp@keithp.com>
Signed-off-by: Keith Packard <keithp@keithp.com>
This is a squash merge containing all the API changes, as
well as the video ABI bump.
Its been squashed to make bisection easier.
Full patch log below:
commit b202738bbf0c5a1c1172767119c2c71f1e7f8070
Author: Aaron Plattner <aplattner@nvidia.com>
Date: Mon May 14 15:16:11 2012 -0700
xfree86: Bump video ABI to 13.0
The ABI was broken by changes to convert from screen index numbers to ScreenPtr
/ ScrnInfoPtr in various structures and function signatures.
Signed-off-by: Aaron Plattner <aplattner@nvidia.com>
Reviewed-by: Dave Airlie <airlied@redhat.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
commit 3d5f7d9f8d408bcad3f83277d255f25d3b0edbf3
Author: Dave Airlie <airlied@redhat.com>
Date: Thu May 24 10:56:57 2012 +0100
xf86: xf86ClearEntityListForScreen should take a pScrn
When adding GPU screens this make life easier.
(also fix comment, as pointed out by Alan)
Reviewed-by: Alan Coopersmith <alan.coopersmith@oracle.com>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
Signed-off-by: Dave Airlie <airlied@redhat.com>
commit afee8b5ab4501597ecc1ade34124d7ca227ab055
Author: Dave Airlie <airlied@redhat.com>
Date: Thu May 24 07:07:32 2012 +0100
xf86i2c: add pscrn for drivers to use
This just adds a pScrn pointer into the struct for the drivers to use
instead of scrnIndex. Mostly scrnIndex is used for logging, but some
drivers use it to lookup xf86Screens, so let them stash a pScrn instead.
Removing the scrnIndex is a bit more involved and I'm not sure its worth
the effort. Doing i2c in the X server is legacy code as far as I'm concerned.
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
commit ea5092f1f679691d187f1eee9427e6057beec56e
Author: Dave Airlie <airlied@redhat.com>
Date: Wed May 23 19:25:20 2012 +0100
dix/gc: consolidate GC object creation in one place
The standard GC create and scratch GC create were 90% the same really,
and I have a need in the future for creating GC objects without the
other bits, so wanted to avoid a third copy.
Reviewed-by: Keith Packard <keithp@keithp.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
commit 3d91482ea9b4883e64e496f2768168e0ffa21ba1
Author: Dave Airlie <airlied@redhat.com>
Date: Wed May 23 10:24:06 2012 +0100
xf86: add a define to denote the new non-index interfaces are being used
This can be used by drivers to provide compatible APIs.
Reviewed-by: Keith Packard <keithp@keithp.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
commit 37c3ae3e6cd4f3dedc72f371096d6743f8f99df3
Author: Dave Airlie <airlied@redhat.com>
Date: Wed May 23 15:09:12 2012 +0100
dix: make Create/Free scratch pixmaps take a ScreenPtr
While technically an API/ABI change I doubt anyone uses it,
but it helps in splitting screens up.
Reviewed-by: Keith Packard <keithp@keithp.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
commit 75f2062a3fe94f04764ecc7d2ff2fbbeccb9da60
Author: Dave Airlie <airlied@redhat.com>
Date: Wed May 23 14:57:55 2012 +0100
xf86/xv: remove scrnIndexfrom xf86FindXvOptions.
Move this interface to taking an ScrnInfoPtr.
Reviewed-by: Keith Packard <keithp@keithp.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
commit f80c2374f40ea7b2ee0556e2e76cc07406f3d843
Author: Dave Airlie <airlied@redhat.com>
Date: Wed May 23 14:53:59 2012 +0100
xf86: make xf86DeleteScreen take a ScrnInfoPtr (v2)
stop passing indices into this function.
v2: drop flags argument.
Reviewed-by: Keith Packard <keithp@keithp.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
commit 58824e414f35682435f15bfe6c4b656bd90b9235
Author: Dave Airlie <airlied@redhat.com>
Date: Wed May 23 14:48:09 2012 +0100
xf86: fix xf86IsScreenPrimary interface to take a pScrn (API/ABI)
Reviewed-by: Keith Packard <keithp@keithp.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
commit 6b4fc1f9d391bcdf7ca288766e49bce60f4635cd
Author: Dave Airlie <airlied@gmail.com>
Date: Tue Apr 10 14:18:59 2012 +0100
xserver: convert block/wakeup handlers to passing ScreenPtr (ABI/API) (v2)
Instead of passing an index, pass the actual ScreenPtr. This allows
more moving towards not abusing xf86Screens + screenInfo.
v2: drop the blockData/wakeupData args as per ajax's suggestion.,
fix docs.
Reviewed-by: Keith Packard <keithp@keithp.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
commit 790d003de20fb47674420a24dadd92412d78620d
Author: Dave Airlie <airlied@gmail.com>
Date: Wed Apr 11 09:53:14 2012 +0100
xf86/common: remove some more pScrn->pScreen uses
remove some more conversions that appeared after api cleanups.
Reviewed-by: Alan Coopersmith <alan.coopersmith@oracle.com>
Acked-by: Aaron Plattner <aplattner@nvidia.com>
Reviewed-by: Adam Jackson <ajax@redhat.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
commit aac85e18d1dd093f2cad6bd29375e40bd7af0b8f
Author: Dave Airlie <airlied@gmail.com>
Date: Tue Apr 10 16:34:53 2012 +0100
ddc: change API to take ScrnInfoPtr (v2)
This removes all xf86Screens usage from ddc code,
it modifies the API for some functions to avoid taking indices.
v2: address Alan's comments about dropping DDC2Init parameter.
Reviewed-by: Alan Coopersmith <alan.coopersmith@oracle.com>
Acked-by: Aaron Plattner <aplattner@nvidia.com>
Reviewed-by: Adam Jackson <ajax@redhat.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
commit fe3f57b6eaf6860a33876a54f9439f69578f03a5
Author: Dave Airlie <airlied@gmail.com>
Date: Tue Apr 10 16:31:26 2012 +0100
vbe: don't use index for VBEInterpretPanelID (API)
Remove use of xf86screens from vbe module.
Reviewed-by: Alan Coopersmith <alan.coopersmith@oracle.com>
Acked-by: Aaron Plattner <aplattner@nvidia.com>
Reviewed-by: Adam Jackson <ajax@redhat.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
commit abf1965f4ed91529036d3fdb470d6a3ce6f29675
Author: Dave Airlie <airlied@gmail.com>
Date: Tue Apr 10 16:25:11 2012 +0100
int10/vbe: don't use xf86Screens. (ABI) (v3)
Pass the ScrnInfoPtr instead of the index in the int10 struct.
This saves us using it to dereference xf86Screens.
v2: address Alan's comment to fix struct alignment.
v3: squash in all the int10 fixes, test the vm86 code builds,
after comments by Keith.
Reviewed-by: Alan Coopersmith <alan.coopersmith@oracle.com>
Acked-by: Aaron Plattner <aplattner@nvidia.com>
Reviewed-by: Adam Jackson <ajax@redhat.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
commit 23cca612b4fb5efc33683c7624b803b457387e3d
Author: Dave Airlie <airlied@gmail.com>
Date: Tue Apr 10 15:30:18 2012 +0100
xserver: drop index argument to ScreenInit (ABI/API) (v2)
This drops the index argument, its the same as pScreen->myNum,
and its the last major index abuse I can find.
v2: address Alan's review - update docs, fix xwin/xnest/darwin
Reviewed-by: Alan Coopersmith <alan.coopersmith@oracle.com>
Acked-by: Aaron Plattner <aplattner@nvidia.com>
Reviewed-by: Adam Jackson <ajax@redhat.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
commit 40d360e2d7e832407f3ed64e3a02c27ecc89a960
Author: Dave Airlie <airlied@gmail.com>
Date: Tue Apr 10 15:23:01 2012 +0100
xf86: migrate PointerMoved from index to ScrnInfoPtr (ABI/API)
This migrates PointerMoved from an index to ScrnInfoPtr.
Reviewed-by: Alan Coopersmith <alan.coopersmith@oracle.com>
Acked-by: Aaron Plattner <aplattner@nvidia.com>
Reviewed-by: Adam Jackson <ajax@redhat.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
commit aa60a2f38679d0eeb979a9c2648c9bc771409bf9
Author: Dave Airlie <airlied@gmail.com>
Date: Tue Apr 10 15:20:46 2012 +0100
xf86: migrate PMEvent to a ScrnInfoPtr (ABI/API)
This migrates the PMEvent from index to ScrnInfoPtr.
Reviewed-by: Alan Coopersmith <alan.coopersmith@oracle.com>
Acked-by: Aaron Plattner <aplattner@nvidia.com>
Reviewed-by: Adam Jackson <ajax@redhat.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
commit d3f28ef44371ed4a039ffc5dd7eb6408d1269ba2
Author: Dave Airlie <airlied@gmail.com>
Date: Tue Apr 10 15:18:30 2012 +0100
xf86: migrate SetDGAMode from index to ScrnInfoPtr (ABI/API)
This migrates the SetDGAMode callback from an index to ScrnInfoPtr.
Reviewed-by: Alan Coopersmith <alan.coopersmith@oracle.com>
Acked-by: Aaron Plattner <aplattner@nvidia.com>
Reviewed-by: Adam Jackson <ajax@redhat.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
commit baf5e4818a74f2b68c3dfdcc56f54322351039a0
Author: Dave Airlie <airlied@gmail.com>
Date: Tue Apr 10 15:14:11 2012 +0100
xf86: migrate ChangeGamma from index to ScrnInfoPtr (ABI/API) (v2)
This migrates the ChangeGamma interface to avoid passing a index.
v2: fix xf86RandR12.c + xf86cmap.c call
Reviewed-by: Alan Coopersmith <alan.coopersmith@oracle.com>
Acked-by: Aaron Plattner <aplattner@nvidia.com>
Reviewed-by: Adam Jackson <ajax@redhat.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
commit 51e5f90ada929d6b23176090badbb42fdb3fa550
Author: Dave Airlie <airlied@gmail.com>
Date: Tue Apr 10 15:11:09 2012 +0100
xf86/exa: migrate index to screen types for EnableDisableFBAccess (ABI/API)
The EXA interface migrates to ScreenPtr,
and the xf86 interface migrated to ScrnInfoPtr.
Reviewed-by: Alan Coopersmith <alan.coopersmith@oracle.com>
Acked-by: Aaron Plattner <aplattner@nvidia.com>
Reviewed-by: Adam Jackson <ajax@redhat.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
commit 94f1f21d17e86f96d4a54292a399160950087675
Author: Dave Airlie <airlied@gmail.com>
Date: Tue Apr 10 15:02:11 2012 +0100
xf86: migrate ValidMode callback to ScrnInfoPtr (ABI/API)
This migrates the ValidMode to passing a ScrnInfoPtr instead
of an index.
Reviewed-by: Alan Coopersmith <alan.coopersmith@oracle.com>
Acked-by: Aaron Plattner <aplattner@nvidia.com>
Reviewed-by: Adam Jackson <ajax@redhat.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
commit 3f8f18198fed4f39ec805b508a3482e91eea26b2
Author: Dave Airlie <airlied@gmail.com>
Date: Tue Apr 10 14:59:46 2012 +0100
xf86: migrate SwitchMode to taking ScrnInfoPtr (ABI/API) (v2)
This migrate the SwitchMode interface to take a ScrnInfoPtr
instead of an index.
v2: drop flags.
Reviewed-by: Alan Coopersmith <alan.coopersmith@oracle.com>
Acked-by: Aaron Plattner <aplattner@nvidia.com>
Reviewed-by: Adam Jackson <ajax@redhat.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
commit d06a038a5c49328ab3a8d969d24f9fcd22c63202
Author: Dave Airlie <airlied@gmail.com>
Date: Tue Apr 10 14:50:37 2012 +0100
xf86: move AdjustFrame to passing ScrnInfoPtr (ABI/API) (v2)
This converts AdjustFrame code paths to passing a ScrnInfoPtr
instead of an integer index.
v2: drop flags args.
Reviewed-by: Alan Coopersmith <alan.coopersmith@oracle.com>
Acked-by: Aaron Plattner <aplattner@nvidia.com>
Reviewed-by: Adam Jackson <ajax@redhat.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
commit 53d2f8608ffd4090d08e7d5cf2e92fb954959b90
Author: Dave Airlie <airlied@gmail.com>
Date: Tue Apr 10 14:41:27 2012 +0100
xf86: modify FreeScreen callback to take pScrn instead of index. (ABI/API) (v2)
Another index->pScrn conversion.
v2: drop flags arg.
Reviewed-by: Alan Coopersmith <alan.coopersmith@oracle.com>
Acked-by: Aaron Plattner <aplattner@nvidia.com>
Reviewed-by: Adam Jackson <ajax@redhat.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
commit 60db37c0b247052e0f5c54b1921fe58a3609c2e3
Author: Dave Airlie <airlied@gmail.com>
Date: Tue Apr 10 14:35:41 2012 +0100
xf86: change EnterVT/LeaveVT to take a ScrnInfoPtr (ABI/API break) (v2)
This modifies the EnterVT/LeaveVT interfaces to take a ScrnInfoPtr
instead of an index into xf86Screens. This allows dropping more
public dereferences of the xf86Screens and screenInfo.
v2: drop flags args as suggested by Keith, fix docs.
Reviewed-by: Alan Coopersmith <alan.coopersmith@oracle.com>
Acked-by: Aaron Plattner <aplattner@nvidia.com>
Reviewed-by: Adam Jackson <ajax@redhat.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
commit 06729dbbc804a20242e6499f446acb5d94023c3c
Author: Dave Airlie <airlied@gmail.com>
Date: Tue Apr 10 14:04:59 2012 +0100
xserver: remove index from CloseScreen (API/ABI breakage)
This drops the index from the CloseScreen callback,
its always been useless really, since the pScreen contains it.
Reviewed-by: Alan Coopersmith <alan.coopersmith@oracle.com>
Acked-by: Aaron Plattner <aplattner@nvidia.com>
Reviewed-by: Adam Jackson <ajax@redhat.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
Bump dri2proto dependency to 2.7.
Bump DRI2INFOREC_VERSION to 7.
This new protocol request effectively allows clients to perform feature
detection on the DDX. The request was added in DRI2 protocol 1.4.
If I had DRI2GetParam in June 2011, when I was implementing support in the
Intel DDX and Mesa for new hardware that required a new DRI2 attachment
format, then I could have avoided a week of pain caused by the necessity
to write a horrid feature detection hack [1] in Mesa. In the future, when
the work begins to add MSAA support to the Intel DDX, having a clean way
to do feature detection will allow us to avoid revisiting and expanding
that hack.
[1] mesa, commit aea2236a, function intel_verify_dri2_has_hiz
v2: If driver doesn't define ds->GetParam, dont' crash. Fall back to
default behavior, per keithp.
Reviewed-by: Keith Packard <keithp@keithp.com>
Reviewed-by: Eric Anholt <eric@anholt.net> (v1)
Reviewed-by: Ian Romanick <idr@freedesktop.org> (v1)
Signed-off-by: Chad Versace <chad.versace@linux.intel.com>
xf86-input-evdev (since "smooth scrolling" support was added) can send mouse
motion and wheel events in one batch, so we need to handle it properly.
Otherwise mouse wheel events which come with motion events are lost
and separate mouse wheel events are handled through non-DGA path.
Signed-off-by: Marcin Slusarz <marcin.slusarz@gmail.com>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
The swapped implementation of DRI2Connect is always responding with empty
device and driver values. However the response was only prepared and never
sent (also had undefined .type member), causing e.g. glxinfo get stuck waiting
for response when started remotely from machine with different endianity.
Signed-off-by: Michal Srb <msrb@suse.com>
Reviewed-by: Jeremy Huddleston <jeremyhu@apple.com>
Tested-by: Michel Dänzer <michel@daenzer.net>
Signed-off-by: Keith Packard <keithp@keithp.com>
This macro did lookups via privates but we can just use the ScreenToScrn
conversion instead.
This patch drops all in-server uses, we should drop the macro later,
once drivers have been converted to not use it.
Reviewed-by: Adam Jackson <ajax@redhat.com>
Reviewed-by: Keith Packard <keithp@keithp.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
Just simple conversion to use the lookup function.
Reviewed-by: Adam Jackson <ajax@redhat.com>
Reviewed-by: Keith Packard <keithp@keithp.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
This uses a standard conversion function to do the conversion.
Reviewed-by: Adam Jackson <ajax@redhat.com>
Reviewed-by: Keith Packard <keithp@keithp.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
Just use new macros to access scrn->screen.
Reviewed-by: Adam Jackson <ajax@redhat.com>
Reviewed-by: Keith Packard <keithp@keithp.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
migrate to new helper API.
This just wraps all the obvious uses of xf86Screens[pScreen->myNum],
and should be fairly simple to review.
v2: remove commented out lines.
Reviewed-by: Aaron Plattner <aplattner@nvidia.com>
Reviewed-by: Keith Packard <keithp@keithp.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
These are just simple functions that we should start migrating drivers
to using.
The end goal is to remove xf86Screens and screenInfo from the ABI.
This includes a define XF86_HAS_SCRN_CONV that drivers can ifdef to provide
their own copies. I'll probably post a generic compat.h file for drivers later.
v2: add asserts.
Reviewed-by: Aaron Plattner <aplattner@nvidia.com>
Reviewed-by: Keith Packard <keithp@keithp.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
Currently if we claim a slot for a PCI driver, we never let it go properly,
this prevents the fallback probe from reusing the slot, even though it
isn't claimed for that pci slot.
So if you set the modesetting driver to point at a specific kms device,
that isn't a PCI device (i.e. USB dongle), then the modesetting driver
loads, the pci probe tries to bind the config slot to the primary PCI
device, however we then check the kms device bus id to discover it
isn't valid. However we don't remove the claim on the slot. Next the
old probe function is called and there is no slots to claim.
This patch fixes that and converts the pciSlotClaimed boolean into
a counter, and changes the unclaim api to take a device pointer
to remove from the entity.
Reviewed-by: Adam Jackson <ajax@redhat.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
In order to use udev for gpu enumeration, we need to init udev earlier
than input initialisations. This splits the config init stuff so that udev
pre init sets up before output initialisation.
this is just a prepatory patch, doesn't change anything major.
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
Signed-off-by: Dave Airlie <airlied@redhat.com>
Add the modesetting driver to the fallback list on Linux, after vesa
before fbdev.
Acked-by: Peter Hutterer <peter.hutterer@who-t.net>
Signed-off-by: Dave Airlie <airlied@redhat.com>
Code was deleted in commit affec10635
Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>
Reviewed-by: Dave Airlie <airlied@redhat.com>
Reviewed-by: Adam Jackson <ajax@redhat.com>
The code to implement was deleted when BaseModules[] was emptied by
the replacement of the "pcidata" module with libpciaccess calls
in commit 46f55f5dea.
Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>
Reviewed-by: Dave Airlie <airlied@redhat.com>
Reviewed-by: Adam Jackson <ajax@redhat.com>
Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>
Reviewed-by: Mark Kettenis <kettenis@openbsd.org>
Reviewed-by: Dave Airlie <airlied@redhat.com>
The functions are already declared in xf86xv.h
Signed-off-by: Michal Suchanek <hramrach@gmail.com>
Reviewed-by: Dave Airlie <airlied@redhat.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
If we get asked to pci open a device with a kms path override,
make sure they match, otherwise this driver can steal the primary
device binding for a usb adaptor.
The driver should fallback to the old probe entry point in this case.
Signed-off-by: Dave Airlie <airlied@redhat.com>
In order to generate a 256-entry ramp in [0,65535] which covers the full
range, one must mupliply eight-bit values not by 256 but rather by 257.
Many years back – well before the RANDR extension was written, and
before xorg@fdo – a similar bug fix was made to the DIX for converting
client-supplied eight-bit color values into sixteen-bit values.
Noticed by: Elle Stone and Graeme Gill.
Signed-off-by: James Cloos <cloos@jhcloos.com>
the cirrus driver presents certain challenges, and this is a
workaround, until we can possibly agree some sane interface
for exposing this information.
Signed-off-by: Dave Airlie <airlied@redhat.com>
due to interaction between option handling and set depth, we need
to what fbdev does to get the device path early.
Signed-off-by: Dave Airlie <airlied@redhat.com>
If a device was enabled before the VT switch, re-enabled it. Otherwise leave
it as is, there was probably a reason why it was disabled.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Chase Douglas <chase.douglas@canonical.com>
Reviewed-by: Daniel Stone <daniel@fooishbar.org>
This is the result of re-running the 'x-indent.sh' script over
xf86vmode.c to clean up the disaster caused by broken syntax in the
file.
Signed-off-by: Keith Packard <keithp@keithp.com>
Inside the unfinished XF86VIDMODE_EVENTS #ifdef block the
function definition for xf86VidModeNotifyEvent had an extra ');'
before the prototype argument declarations. This was harmless for the
compiler as the code never gets used, but completely messed up the
file re-indentation. This patch removes the spurious characters in
preparation for re-indenting the file.
Signed-off-by: Keith Packard <keithp@keithp.com>
This reverts commit 55f552adb6.
This appears to cause a crash at init time instead of close.
Reported-by: Knut Petersen <Knut_Petersen@t-online.de>
Acked-by: Jeremy Huddleston <jeremyhu@apple.com>
Signed-off-by: Keith Packard <keithp@keithp.com>
As a PE platform, all symbols in both EXEs and DLLs must be resolved
at link time. As Xorg modules depend on symbols in the Xorg
executable, we must build Xorg before its modules, creating an implib
from the former which is used to link the latter. This implib must
then be installed in order to build the drivers.
Currently only two drivers are supported on Cygwin: xf86-video-dummy
(to replace Xvfb/Xfake) and xf86-video-nested (to replace Xnest/Xephyr).
Signed-off-by: Yaakov Selkowitz <yselkowitz@users.sourceforge.net>
Reviewed-by: Jon TURNEY <jon.turney@dronecode.org.uk>
Reviewed-by: Jeremy Huddleston <jeremyhu@apple.com>
Cygwin doesn't have ELF rpath capabilities, so these libraries need
to be loaded before the drivers (namely dummy and nested) which
depend on their symbols.
Signed-off-by: Yaakov Selkowitz <yselkowitz@users.sourceforge.net>
Reviewed-by: Jon TURNEY <jon.turney@dronecode.org.uk>
Reviewed-by: Jeremy Huddleston <jeremyhu@apple.com>
This is necessary when building Xorg and XWin simultaneously, otherwise
undefined symbol errors result in sdksyms.c.
Signed-off-by: Yaakov Selkowitz <yselkowitz@users.sourceforge.net>
Reviewed-by: Jon TURNEY <jon.turney@dronecode.org.uk>
Reviewed-by: Jeremy Huddleston <jeremyhu@apple.com>
Cygwin libraries use the .dll extension and "cyg" prefix in place of "lib".
Signed-off-by: Yaakov Selkowitz <yselkowitz@users.sourceforge.net>
Reviewed-by: Jon TURNEY <jon.turney@dronecode.org.uk>
Reviewed-by: Jeremy Huddleston <jeremyhu@apple.com>
This will be necessary to port Xorg to Cygwin, but other platforms may
find this useful as well.
Signed-off-by: Yaakov Selkowitz <yselkowitz@users.sourceforge.net>
Reviewed-by: Jon TURNEY <jon.turney@dronecode.org.uk>
Reviewed-by: Jeremy Huddleston <jeremyhu@apple.com>
Excerpt from http://lists.x.org/archives/xorg-devel/2011-March/020481.html:
The Xorg & xorg.conf substitutions are leftover from the transitional
period where some distros were building our sources with the XFree86
and XF86config names until they had time to adjust the rest of their
packages/installer/config code to the new names.
This will fix inconsistencies and prevent the creation of new unneeded
sed patterns.
Reviewed-by: Alan Coopersmith <alan.coopersmith@oracle.com>
Signed-off-by: Gaetan Nadon <memsize@videotron.ca>
Signed-off-by: Keith Packard <keithp@keithp.com>
Excerpt from http://lists.x.org/archives/xorg-devel/2011-March/020481.html:
The Xorg & xorg.conf substitutions are leftover from the transitional
period where some distros were building our sources with the XFree86
and XF86config names until they had time to adjust the rest of their
packages/installer/config code to the new names.
This will fix inconsistencies and prevent the creation of new unneeded
sed patterns.
Reviewed-by: Alan Coopersmith <alan.coopersmith@oracle.com>
Signed-off-by: Gaetan Nadon <memsize@videotron.ca>
Signed-off-by: Keith Packard <keithp@keithp.com>
This is strictly the application of the script 'x-indent-all.sh'
from util/modular. Compared to the patch that Daniel posted in
January, I've added a few indent flags:
-bap
-psl
-T PrivatePtr
-T pmWait
-T _XFUNCPROTOBEGIN
-T _XFUNCPROTOEND
-T _X_EXPORT
The typedefs were needed to make the output of sdksyms.sh match the
previous output, otherwise, the code is formatted badly enough that
sdksyms.sh generates incorrect output.
The generated code was compared with the previous version and found to
be essentially identical -- "assert" line numbers and BUILD_TIME were
the only differences found.
The comparison was done with this script:
dir1=$1
dir2=$2
for dir in $dir1 $dir2; do
(cd $dir && find . -name '*.o' | while read file; do
dir=`dirname $file`
base=`basename $file .o`
dump=$dir/$base.dump
objdump -d $file > $dump
done)
done
find $dir1 -name '*.dump' | while read dump; do
otherdump=`echo $dump | sed "s;$dir1;$dir2;"`
diff -u $dump $otherdump
done
Signed-off-by: Keith Packard <keithp@keithp.com>
Acked-by: Daniel Stone <daniel@fooishbar.org>
Acked-by: Alan Coopersmith <alan.coopersmith@oracle.com>
indent sometimes adds a blank line between the type and the name in a
function declaration that includes _X_EXPORT, so handle that before
the files are re-indented.
Signed-off-by: Keith Packard <keithp@keithp.com>
Currently the driver only probes a device when it has a
busID. The busID is optional so don't depend on it.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
Signed-off-by: Dave Airlie <airlied@redhat.com>
To make the driver work on nin PCI devices we shouldn't bail
out in this case.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
Signed-off-by: Dave Airlie <airlied@redhat.com>
When no devicename is found in the option then the driver probes
by PciInfo no matter if it's valid or not. Instead of doing this
use PciInfo only when it's valid and fall back to the devicename
otherwise. With devicename probing use open_hw() to fall back
on the KMSDEVICE environment variable or to the default device.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
Signed-off-by: Dave Airlie <airlied@redhat.com>
probe_hw opens the hardware in the order we want it:
first try devname, if this is NULL then try the KMSDEVICE
environment variable and as a last fallback use "/dev/dri/card0".
Instead of implementing the same code again when really opening
the device move the code to a open_hw() function and let probe_hw
use it.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
Signed-off-by: Dave Airlie <airlied@redhat.com>
in Probe() the indention shows what's meant but there are no
brackets. Add them.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
Signed-off-by: Dave Airlie <airlied@redhat.com>
The current code only adds -Wall and only for gcc.
Automake reserves the use of CPPFLAGS for the user to override
on the command line.
This also breaks the option --enable-strict-compilation
The variable CWARNFLAGS contains the complete set of warnings
and is platform sensitive.
Signed-off-by: Gaetan Nadon <memsize@videotron.ca>
Signed-off-by: Dave Airlie <airlied@redhat.com>
This reflects the copyright license text in the source code
Signed-off-by: Gaetan Nadon <memsize@videotron.ca>
Signed-off-by: Dave Airlie <airlied@redhat.com>
If the kernel rejects a cursor, cause a fallback, this isn't 100% as
we can lose the initial cursor, but it works fine once wm starts.
Signed-off-by: Dave Airlie <airlied@redhat.com>
Rename functions/macros from list_* to xorg_list_*
Rename struct from struct list to struct xorg_list.
Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>
Reviewed-by: Keith Packard <keithp@keithp.com>
In-sed-I-trust: Peter Hutterer <peter.hutterer@who-t.net>
We don't need anything from that header (which defines /proc & kernel
structures for process information), and it causes some namespace conflicts.
Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>
Reviewed-by: Keith Packard <keithp@keithp.com>
342f3eac84 introduced a bug, 'base' is
incremented before use. The old code corrected this when unmapping, so
the new code should too.
Signed-off-by: Adam Jackson <ajax@redhat.com>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
Signed-off-by: Keith Packard <keithp@keithp.com>
libdix.a is already provided by XSERVER_LIBS. Including it in libxorgxkb
results can result in duplicate symbols landing in the Xorg binary on some
configurations (buggy glibtool on darwin).
Signed-off-by: Jeremy Huddleston <jeremyhu@apple.com>
Always install XAA SDK headers so drivers still build even with
--disable-xaa
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Jeremy Huddleston <jeremyhu@apple.com>
Signed-off-by: Keith Packard <keithp@keithp.com>
Some driver modules try to unload submodules that are now built-in.
Signed-off-by: Matthieu Herrb <matthieu.herrb@laas.fr>
Reviewed-by: Julien Cristau <jcristau@debian.org>
Reviewed-by: Adam Jackson <ajax@redhat.com>
Signed-off-by: Keith Packard <keithp@keithp.com>
We don't want to unconditionally use I/O routines here, since if the
driver is using mmap'd VGA ports then the I/O handle won't be set up.
Tested-by: Jeff Chua <jeff.chua.linux@gmail.com>
Signed-off-by: Adam Jackson <ajax@redhat.com>
Reviewed-by: Mark Kettenis <kettenis@openbsd.org>
Signed-off-by: Keith Packard <keithp@keithp.com>
Linux kernels since 2.6.38 (March 2011) have an VT KB mode K_OFF in
which special keys (like Ctrl+C) are not interpreted and input is not
buffered. Use of this mode over K_RAW removes the need for a
xf86ConsoleHandler to drain the VT input buffer, removing the grief it
causes when it goes wrong or is (de)initialized out-of-order. (This
also saves a few needless context switches per key event.)
If K_OFF is not defined or not understood by the kernel, K_RAW and the
previous method is used as a fall-back.
Reviewed-by: Adam Jackson <ajax@redhat.com>
Signed-off-by: Arthur Taylor <art@ified.ca>
Signed-off-by: Keith Packard <keithp@keithp.com>
We don't do anything with EDID v2 blocks besides publish them on the
root window. Worse, the check deleted by this patch would attempt to
take a checksum of arbitrary memory if the rawData array isn't 256+
bytes long (and, for the monitors mentioned, it probably is only 128).
Reviewed-by: Mark Kettenis <kettenis@openbsd.org>
Signed-off-by: Adam Jackson <ajax@redhat.com>
Check for identifier first and bail if it's missing (also remove the current
identifier check after we've already bailed due to missing identifiers)
If a driver is missing, warn but also say that we may have added this device
already. I see too many bugreports with incorrectly shortened log files.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Cyril Brulebois <kibi@debian.org>
Side effect of aa0bfb0f13:
| CCLD Xorg
| sdksyms.o:(.data.rel+0x27d8): undefined reference to `outl'
| collect2: ld returned 1 exit status
Since the linux/ia64 domain I/O support code got removed in that
commit, there's no reason to keep on declaring those functions
(inb, inl, inw, outb, outl, outw).
Bugzilla: https://bugs.freedesktop.org/43985
Reviewed-by: Jeremy Huddleston <jeremyhu@apple.com>
Signed-off-by: Cyril Brulebois <kibi@debian.org>
Debian's QA tool “lintian” reported a bad whatis entry for the
xorg.conf(.d) manpages.
It comes with the following pointers:
For manual pages that document multiple programs, functions, files, or
other things, the part before "\-" should list each separated by a
comma and a space. […]
Refer to the lexgrog(1) manual page, the groff_man(7) manual page, and
the groff_mdoc(7) manual page for details.
Indeed, the current situation is:
$ whatis xorg.conf; whatis xorg.conf.d
xorg.conf (5) - (unknown subject)
xorg.conf.d (5) - (unknown subject)
With this patch:
xorg.conf (5) - configuration files for Xorg X server
xorg.conf.d (5) - configuration files for Xorg X server
Reviewed-by: Alan Coopersmith <alan.coopersmith@oracle.com>
Signed-off-by: Cyril Brulebois <kibi@debian.org>
While a redirected window is flipped, its pixmap may still be used as
and EGL image and should also get invalidated. When sending invalidate
events for a window, also send the events for its pixmap.
Signed-off-by: Ville Syrjälä <syrjala@sci.fi>
Reviewed-by: Keith Packard <keithp@keithp.com>
Signed-off-by: Keith Packard <keithp@keithp.com>
Without this, when a compositing manager unredirects a fullscreen window which
uses DRI2 and page flipping, the DRI2 buffer information for the compositing
manager's output window (typically the Composite Overlay Window or root window)
may become stale, resulting in all kinds of hilarity.
Fixes https://bugs.freedesktop.org/show_bug.cgi?id=35452 .
[Original patch by Michel Dänzer <michel@daenzer.net>]
[Tree walk optimized version by Keith Packard <keithp@keithp.com>]
Signed-off-by: Ville Syrjälä <syrjala@sci.fi>
Signed-off-by: Keith Packard <keithp@keithp.com>
Otherwise we might keep stale cached information, e.g. after the driver
performed page flipping.
This is part of the fix for
https://bugs.freedesktop.org/show_bug.cgi?id=35452 .
Signed-off-by: Michel Dänzer <daenzer@vmware.com>
Reviewed-by: Ville Syrjälä <ville.syrjala@nokia.com>
Reviewed-by: Mario Kleiner <mario.kleiner@tuebingen.mpg.de>
Signed-off-by: Keith Packard <keithp@keithp.com>
If the client is not behaving correctly and swaps buffers before
getting them, Valgrind will complain about uninitialized memory being
used in DRI2InvalidateDrawable.
Signed-off-by: Rami Ylimäki <rami.ylimaki@vincit.fi>
Reviewed-by: Ville Syrjälä <syrjala@sci.fi>
Signed-off-by: Keith Packard <keithp@keithp.com>
New additions to the API:
- InitTouchClassDeviceStruct
- xf86PostTouchEvent
Changes to the ABI:
- DeviceIntRec now contains a TouchClassPtr
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Chase Douglas <chase.douglas@canonical.com>
xf86PostTouchEvent is the driver API to submit touch events to the server.
This API doesn't do anything yet though but now we can at least bump the
API.
For valuators, drivers should use the existing xf86InitValuatorAxisStruct
function.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Chase Douglas <chase.douglas@canonical.com>
This allows us to run the server as a normal user whilst still
being able to use the -modulepath, -logfile and -config switches
We define a xf86PrivsElevated which will do the checks and cache
the result in case it is called more than once.
Also renamed the paths #defines to match their new meaning.
Original discussion which led to this patch can be found here:
http://lists.freedesktop.org/archives/xorg-devel/2011-September/025853.html
Signed-off-by: Antoine Martin <antoine@nagafix.co.uk>
Tested-by: Michal Suchanek <hramrach at centrum.cz>
Reviewed-by: Jamey Sharp <jamey at minilop.net>
Reviewed-by: Adam Jackson <ajax@redhat.com>
Previously it always passed a format string with exactly one argument,
using NULL when the format string needed none. Now pass the right number
of arguments to clear gcc warnings of 'too many arguments for format'.
Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>
Reviewed-by: Jamey Sharp <jamey@minilop.net>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
Both functions call exit() at the end and have no other return path.
Also correct comment/heading to reflect commit 6450f6ca7e moving
DoShowOptions into xf86Configure.c.
Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>
Reviewed-by: Jamey Sharp <jamey@minilop.net>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
Clears gcc warning in every file that includes xf86Modes.h:
xf86Modes.h:102:1: warning: redundant redeclaration of 'xf86ValidateModesFlags'
xf86Modes.h:72:1: note: previous declaration of 'xf86ValidateModesFlags' was here
Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>
Reviewed-by: Jamey Sharp <jamey@minilop.net>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
Since all we do with the option list is walk down the list printing
the names, there's no need to cast away its constness.
Clears gcc warning:
xf86Configure.c: In function 'DoShowOptions':
xf86Configure.c:781:4: warning: cast discards qualifiers from pointer target type
Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>
Reviewed-by: Jamey Sharp <jamey@minilop.net>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
Strings are all pointers to literal constants, just used as input
to printf calls when debugging is enabled.
Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>
Reviewed-by: Jamey Sharp <jamey@minilop.net>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
If we didn't go into the if (!autoconfig) { } block, the filename,
dirname, and sysdirname pointers were never initialized, but we
freed them outside the block, leading to potential memory corruption.
Move the frees inside the block where they're initialized to avoid this.
To avoid similar problems, move the declarations of the variables that
are only used in this block inside the block.
Regression introduced by commit 3d635fe84d
Found by gcc warning:
xf86Config.c: In function 'xf86HandleConfigFile':
xf86Config.c:2303:11: warning: 'filename' may be used uninitialized in this function
xf86Config.c:2303:22: warning: 'dirname' may be used uninitialized in this function
xf86Config.c:2303:32: warning: 'sysdirname' may be used uninitialized in this function
Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>
Reviewed-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
It will never return NULL, so don't try to handle a NULL condition,
since that just confuses programmers and static analyzers.
It uses calloc, so all the allocated memory is cleared, so there's
no point looping over the memory to manually initialize it NULL.
And just because it's annoying, it doesn't need to be the only
place in this file to do if (NULL==...) instead of if (... == NULL).
Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>
Reviewed-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Avoids the dummy-event dance if we have an event type and need to get the
matching XI2 type.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Chase Douglas <chase.douglas@canonical.com>
Treat a scandir error from a missing (or unusable) directory return as
if it simply returned no files at all, which is what we want.
cc: Paulo Zanoni <przanoni@gmail.com>
Signed-off-by: Keith Packard <keithp@keithp.com>
If the pGCPriv->flags == 2, then we try to assign the freed pGCPriv->XAAOps
avoid this by clearing the flags in to be destroyed pGCPriv.
Reported by coverity.
Signed-off-by: Dave Airlie <airlied@redhat.com>
Reviewed-by: Jeremy Huddleston <jeremyhu@apple.com>
So on RHEL5 anaconda sets an xorg.conf with a fixed 800x600 mode in it,
we run radeonfb and fbdev since ati won't work in userspace due to domain
issues in the older codebase.
On certain pseries blades the built-in KVM can't accept an 800x600-43 mode,
it requires the 800x600-60 mode, so we have to have the kernel radeonfb
driver reject the 800x600-43 mode when it sees it. However then fbdev
doesn't try any of the other 800x600 modes in the modelist, and we end up
getting a default 640x480 mode we don't want.
This patch changes the mode validation loop to continue on with the other modes
that match to find one that works.
v2: move code around to avoid extra loop, after comment from Jamey.
v3: move loop setup back into loop as per Jeremy's review.
Signed-off-by: Dave Airlie <airlied@redhat.com>
Reviewed-by: Jamey Sharp <jamey@minilop.net>
Reviewed-by: Jeremy Huddleston <jeremyhu@apple.com>
Fixes Sun cc warning that was recently elevated to error by the
stricter default CFLAGS changes to xorg-macros:
"loadmod.c", line 914: improper pointer/integer combination: op "<"
Should have been changed when commit ab7f057ce9 changed the
LoaderOpen return type from int to void *.
Changes log message when file is found but dlopen() fails from:
(EE) LoadModule: Module dbe does not have a dbeModuleData data object.
(EE) Failed to load module "dbe" (invalid module, 0)
to:
(EE) Failed to load module "dbe" (loader failed, 7)
Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>
Reviewed-by: Adam Jackson <ajax@redhat.com>
Ensure ffs, strndup, strlcat, etc. aren't defined by our headers
if they're already defined in the system headers.
This does export the HAVE_FFS, HAVE_STRNDUP, etc. definitions to drivers,
but if you built the Xserver with a libc that had those, and then build
the drivers with a less capable libc, you're going to have problems anyway,
and this should solve some reported problems with conflicts between our
strndup definition and gcc magic for it.
Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>
Reviewed-by: Adam Jackson <ajax@redhat.com>
I wonder if there are any other patterns we haven't seen yet?
Reviewed-by: Jamey Sharp <jamey@minilop.net>
Signed-off-by: Matt Turner <mattst88@gmail.com>
Signed-off-by: Keith Packard <keithp@keithp.com>
After we tokenize val.str, we discard it.
This is just one example:
6 bytes in 1 blocks are definitely lost in loss record 24 of 652
at 0x4C2779D: malloc (in vgpreload_memcheck-amd64-linux.so)
by 0x4D744D: xf86getToken (scan.c:400)
by 0x4D75F1: xf86getSubToken (scan.c:462)
by 0x4DB060: xf86parseInputClassSection (InputClass.c:145)
by 0x4D664C: xf86readConfigFile (read.c:184)
by 0x490556: xf86HandleConfigFile (xf86Config.c:2360)
by 0x49AA77: InitOutput (xf86Init.c:365)
by 0x425A7A: main (main.c:204)
Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Dan Nicholson <dbn.lists@gmail.com>
After we convert the value to a boolean, we discard the string.
This is just one example:
3 bytes in 1 blocks are definitely lost in loss record 5 of 657
at 0x4C2779D: malloc (vgpreload_memcheck-amd64-linux.so)
by 0x4D744D: xf86getToken (scan.c:400)
by 0x4D75F1: xf86getSubToken (scan.c:462)
by 0x4DB3E0: xf86parseInputClassSection (InputClass.c:189)
by 0x4D664C: xf86readConfigFile (read.c:184)
by 0x490556: xf86HandleConfigFile (xf86Config.c:2360)
by 0x49AA77: InitOutput (xf86Init.c:365)
by 0x425A7A: main (main.c:204)
Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Dan Nicholson <dbn.lists@gmail.com>
v2: move the free()s to the function that calls scandir
80 bytes in 1 blocks are definitely lost in loss record 411 of 631
at 0x4C2779D: malloc (vgpreload_memcheck-amd64-linux.so)
by 0x4C27927: realloc (vgpreload_memcheck-amd64-linux.so)
by 0x696A80D: scandir (scandir.c:108)
by 0x4D8828: OpenConfigDir (scan.c:854)
by 0x4D8A43: xf86openConfigDirFiles (scan.c:952)
by 0x49031F: xf86HandleConfigFile (xf86Config.c:2327)
by 0x49A9E3: InitOutput (xf86Init.c:365)
by 0x425A7A: main (main.c:204)
Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
We call xf86penConfigDirFiles twice, so we overwrite the configDirPath
variable, losing the pointer. If we move the pointer management to the
upper layer (the function callers), they will be able to call these
functions as many times as they want, but they'll have to free those
returned values.
v2: don't leak inside XWin
4,097 bytes in 1 blocks are definitely lost in loss record 625 of 632
at 0x4C2779D: malloc (in vgpreload_memcheck-amd64-linux.so)
by 0x4D7899: DoSubstitution (scan.c:615)
by 0x4D87B0: OpenConfigDir (scan.c:845)
by 0x4D8A2D: xf86openConfigDirFiles (scan.c:955)
by 0x49031F: xf86HandleConfigFile (xf86Config.c:2327)
by 0x49A9BF: InitOutput (xf86Init.c:365)
by 0x425A7A: main (main.c:204)
Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Reviewed-by: Dan Nicholson <dbn.lists@gmail.com>
Not needed since 6cf844ab69 split out the allocation/manipulation
into the helper function, leaving FindModule just copying the pointer
around, and causing gcc warnings and an unreachable call to free.
Also no longer need to store the combined strlen results in dirlen.
Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>
Reviewed-by: Jeremy Huddleston <jeremyhu@apple.com>
Exposed by recent addition of -Wredundant-decls to default CWARNFLAGS
Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>
Reviewed-by: Jeremy Huddleston <jeremyhu@apple.com>
Matches what linux_agp already does and prevents gcc from throwing up:
sun_agp.c: In function 'xf86DeallocateGARTMemory':
sun_agp.c:236:40: error: cast to pointer from integer of different size
Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>
Reviewed-by: Jeremy Huddleston <jeremyhu@apple.com>
This batch is the straightforward set - others are more complex and
need more analysis to determine right size to pass.
Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>
Reviewed-by: Jeremy Huddleston <jeremyhu@apple.com>
Found no calls from current driver modules
Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>
Reviewed-by: Jeremy Huddleston <jeremyhu@apple.com>
Gets rid of duplicate static copy of optionTypeToString by putting
both callers of that helper function in the same source file.
Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>
Reviewed-by: Daniel Stone <daniel@fooishbar.org>
Reviewed-by: Jeremy Huddleston <jeremyhu@apple.com>
When we want to print a string, it's okay to just print it.
We don't need to first allocate a buffer 2 bytes bigger than the
string, copy the entire string unmodified to the buffer, print the
buffer, and then leak the buffer (though we AbortDDX 8 lines later,
and then just in case we survived that, call exit as well, so the
leak is short lived, just oh so pointless).
Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>
Reviewed-by: Daniel Stone <daniel@fooishbar.org>
Reviewed-by: Jeremy Huddleston <jeremyhu@apple.com>
As long as we're carrying around a compatibility copy in os/strl*.c,
might as well use them.
Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>
Reviewed-by: Jeremy Huddleston <jeremyhu@apple.com>
We don't ship either one, so don't waste time and make confusing log
entries trying to load them.
Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>
Reviewed-by: Daniel Stone <daniel@fooishbar.org>
At least one revision of the AAO reports a 190x110mm maximum size but a
451x113mm mode.
X.Org Bug 41141 <https://bugs.freedesktop.org/show_bug.cgi?id=41141>
Signed-off-by: Ross Burton <ross@linux.intel.com>
Reviewed-by: Daniel Stone <daniel@fooishbar.org>
Reviewed-by: Jeremy Huddleston <jeremyhu@apple.com>
The Resource database is reset upon regeneration and so the dri2 module
needs to re-register its RESTYPE for the drawable or else it will
clobber the next unsuspecting user of the database. Fortunately, DRI2 is
loaded late in the initialisation sequence and was last up until
xf86-video-intel started using the Resource database to track
outstanding swaps...
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Jeremy Huddleston <jeremyhu@apple.com>
Tested-by: Paulo Zanoni <paulo.r.zanoni@intel.com>