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>
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>
DDX can now implement validation for swap_limit changes to prevent
configurations that are not support in driver.
Signed-off-by: Pauli Nieminen <ext-pauli.nieminen@nokia.com>
CC: Mario Kleiner <mario.kleiner@tuebingen.mpg.de>
Reviewed-by: Mario Kleiner <mario.kleiner@tuebingen.mpg.de>
Reviewed-by: Jesse Barnes <jbarnes@virtuousgeek.org>
This allows ddx to set swap_limit if there is more than one back
buffer for drawable. Setting swap_limit has to also check if change
affects a client that is blocked.
This can be used to implement N-buffering in driver with minimal
logic in allocation and selecting next back.
Signed-off-by: Pauli Nieminen <ext-pauli.nieminen@nokia.com>
Reviewed-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Reviewed-by: Francisco Jerez <currojerez@riseup.net>
Reviewed-by: Mario Kleiner <mario.kleiner@tuebingen.mpg.de>
Signed-off-by: Keith Packard <keithp@keithp.com>
ReuseBufferNotify hook is called whenever old buffer is reused in DRI2
code.
Driver can use this hook to rewrite the buffer name if hardware requires
shared buffers. Shared buffer might be some hardware limited resources like
framebuffer that is preallocated in boot.
Signed-off-by: Pauli Nieminen <ext-pauli.nieminen@nokia.com>
Reviewed-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Signed-off-by: Keith Packard <keithp@keithp.com>
ConfigNotify is set by DRI2ScreenInit, but not restored to
previous state on close. Fix it.
(I'm preparing a patch for xf86-video-nouveau which detects GPU lockup
after dri2 init and it needs to reinitialize dri2)
Signed-off-by: Marcin Slusarz <marcin.slusarz@gmail.com>
Reviewed-by: Daniel Stone <daniel@fooishbar.org>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Only send invalidate events for drawables if some client has requested
some buffers.
Signed-off-by: Ville Syrjälä <ville.syrjala@nokia.com>
Reviewed-by: Michel Dänzer <michel@daenzer.net>
Signed-off-by: Keith Packard <keithp@keithp.com>
update_dri2_drawable_buffers() doesn't modify out_count, so pass it
by value.
Signed-off-by: Ville Syrjälä <ville.syrjala@nokia.com>
Reviewed-by: Tiago Vignatti <tiago.vignatti@nokia.com>
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
Don't access invalid memory if calloc() fails to allocate the buffers
array.
Signed-off-by: Ville Syrjälä <ville.syrjala@nokia.com>
Reviewed-by: Tiago Vignatti <tiago.vignatti@nokia.com>
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
The calling for allocate_or_reuse_buffer may fail due to some reason, e.g. out of memory.
If the buffers[] were not initialized to be NULL, the following err_out may try to access an illegal memory, which will cause X crash afterward.
Reviewed-by: Kristian Høgsberg <krh@bitplanet.net>
Signed-off-by: Justin Dou <Justin.Dou@intel.com>
Signed-off-by: Keith Packard <keithp@keithp.com>
If the drawable size doesn't match the pixmap size page flipping should
not be allowed.
If the window is larger than the pixmap, page flipping might need to
reposition the CRTC somewhere in the middle of the pixmap. I didn't
spot any code that would handle that at least in the intel driver.
Also the root pixmap could then move to some negative screen
coordinates. Not sure if all bits of code could handle that. Perhaps
when composite is enabled screen_x/y would make it work, but without
composite there's no way that it would work AFAICS.
Signed-off-by: Ville Syrjälä <ville.syrjala@nokia.com>
Reviewed-by: Alex Deucher <alexdeucher@gmail.com>
DDX driver may implement schedule swap without GetMSC. In that case we
can't call GetMSC in DRI2SwapBuffers.
Signed-off-by: Pauli Nieminen <ext-pauli.nieminen@nokia.com>
Reviewed-by: Mario Kleiner <mario.kleiner@tuebingen.mpg.de>
CC: Keith Packard <keithp@keithp.com>
CC: Kristian Høgsberg <krh@bitplanet.net>
Reviewed-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Signed-off-by: Keith Packard <keithp@keithp.com>
If a pixmap header is modified or the drawable serial changes, some
aspects of the drawable are likely to have changed so we should
re-allocate the corresponding DRI2 drawable in that case. This is one
way of catching when the root window pixmap changes through xrandr.
Fixes bug https://bugs.freedesktop.org/show_bug.cgi?id=28365.
Reviewed-by: Keith Packard <keithp@keithp.com>
Reviewed-by: Kristian Høgsberg <krh@bitplanet.net>
Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Signed-off-by: Keith Packard <keithp@keithp.com>
Detect if a drawable has been moved from an original crtc to a new crtc
with a lower current vblank count than the original crtc inbetween
glXSwapBuffers() calls. Reinitialize drawable's last_swap_target
before scheduling next swap if such a move has taken place.
last_swap_target defines the baseline for scheduling the next swap.
If a movement between crtc's is not taken into account, the swap may
schedule for a vblank count on the new crtc far in the future, resulting
in a apparent "hang" of the drawable for a long time.
Fixes Bugzilla bug #28383.
Signed-off-by: Mario Kleiner <mario.kleiner@tuebingen.mpg.de>
Reviewed-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Signed-off-by: Keith Packard <keithp@keithp.com>
This prevents DRI2GetScreen from being invoked with an uninitialized
private key which would cause an assert failure.
Signed-off-by: Keith Packard <keithp@keithp.com>
Tested-by: David Ronis <David.Ronis@McGill.CA>
This reverts commit fdb081b430
"dri2: Deal with input-only windows by using WindowDrawable()"
and replaces it as follows:
Reject the creation of a DRI2 drawable for UNDRAWABLE_WINDOW (input-only
windows) and DRAWABLE_BUFFER (whatever those are) drawables and only look up
privates for the supported drawable types.
The rest of the the code can continue pretending there's only output windows
and pixmaps, which are the only kinds of drawables relevant for DRI2.
Fixes server crash with GLX compositing managers such as compiz or kwin, due
to looking up a window private for a pixmap and getting a bogus pointer.
Signed-off-by: Michel Dänzer <daenzer@vmware.com>
Reviewed-by: Keith Packard <keithp@keithp.com>
Signed-off-by: Keith Packard <keithp@keithp.com>
DRI2WaitSBC() didn't block if requested targetSBC wasn't yet reached.
Instead it returned a xreply with uninitialized junk return values, then
blocked the connection until targetSBC was reached.
Therefore the client didn't block, but continued with bogus return
values from glXWaitForSbcOML.
This patch fixes the problem by implementing DRI2WaitSBC similar
to the clean and proven DRI2WaitMSC implementation.
Signed-off-by: Mario Kleiner <mario.kleiner@tuebingen.mpg.de>
Reviewed-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Signed-off-by: Keith Packard <keithp@keithp.com>
This patch was generated by the following Perl code:
perl -i -pe 's/([^_])return\s*\(\s*([^(]+?)\s*\)s*;(\s+(\n))?/$1return $2;$4/g;'
Signed-off-by: Mikhail Gusarov <dottedmag@dottedmag.net>
Reviewed-by: Jamey Sharp <jamey@minilop.net>
Reviewed-by: Daniel Stone <daniel@fooishbar.org>
Signed-off-by: Keith Packard <keithp@keithp.com>
Input only windows aren't DRAWABLE_WINDOW, but casting them to a
PixmapPtr is a bit harsh, and unlikely to get the appropriate privates
structure. use WindowDrawable instead which checks for both
input-output and input-only windows.
Signed-off-by: Keith Packard <keithp@keithp.com>
Tested-by: Eric Anholt <eric@anholt.net>
This patch only changes the API, not the implementation of the
devPrivates infrastructure. This will permit a new devPrivates
implementation to be layed into the server without requiring
simultaneous changes in every devPrivates user.
Signed-off-by: Keith Packard <keithp@keithp.com>
Tested-by: Tiago Vignatti <tiago.vignatti@nokia.com>
This is a combination of a huge mechanical patch and a few small
fixups required to finish the job. They were reviewed separately, but
because the server does not build without both pieces, I've merged
them together at this time.
The mechanical changes were performed by running the included
'fix-region' script over the whole tree:
$ git ls-files | grep -v '^fix-' | xargs ./fix-region
And then, the white space errors in the resulting patch were fixed
using the provided fix-patch-whitespace script.
$ sh ./fix-patch-whitespace
Thanks to Jamey Sharp for the mighty fine sed-generating sed script.
The hand-done changes involve removing functions from dix/region.c
that duplicate inline functions in include/regionstr.h, along with
their declarations in regionstr.h, mi.h and mispans.h.
Reviewed-by: Jamey Sharp <jamey@minilop.net>
Signed-off-by: Keith Packard <keithp@keithp.com>
Since reallocating the backing pixmap can fail, we need to try and do
it before any other side effects of reconfiguring the window happen.
This changes the ConfigNotify hook to return status, and moves the
composite window reconfiguration wrappers to ConfigNotify. They all
basically did the same thing, so we can drop the MoveWindow,
ResizeWindow, ChangeBorderWidth wrappers, and allow ConfigNotify to do
all the work. If reallocation fails we fail before we send any
confiureNotify events, or enter the area we can't recover from.
The only place we now enforce 32k limits are in EXA/UXA/fb, so drivers
that don't use this should probably deal with it in their pixmap
allocate if they don't already.
This also breaks ABI, so we need an alternate fix for older servers,
working on the X server makes me realise why I'm a kernel hacker.
Signed-off-by: Dave Airlie <airlied@redhat.com>
Reviewed-by: Keith Packard <keithp@keithp.com>
Signed-off-by: Keith Packard <keithp@keithp.com>
Many references to the WindowTable array already had the corresponding
screen pointer handy, which meant they usually looked like
"WindowTable[pScreen->myNum]". Adding a field to ScreenRec instead of
keeping this information in a parallel array simplifies those
expressions, and eliminates a MAXSCREENS-sized array.
Since dix uses this data, a screen private entry isn't appropriate.
xf86-video-dummy currently uses WindowTable, so it needs to be updated
to reflect this change.
Signed-off-by: Jamey Sharp <jamey@minilop.net>
Reviewed-by: Tiago Vignatti <tiago.vignatti@nokia.com>
Tested-by: Tiago Vignatti <tiago.vignatti@nokia.com> (i686 GNU/Linux)
Some drivers use DRI protocol but implement their own kernel rendering
manager. For these drivers, libdrm becomes useless. --disable-libdrm
configure parameter can be used to disable libdrm support in dri2.
To provide ABI/API compatibility for libdrm based drivers, libdrm call
is wrapped in ifdef.
Signed-off-by: Tiago Vignatti <tiago.vignatti@nokia.com>
Signed-off-by: Pauli Nieminen <ext-pauli.nieminen@nokia.com>
Reviewed-by: Kristian Høgsberg <krh@bitplanet.net>
With this new hook, drmAuthMagic becomes useless and should be deprecated.
You might want to implement AuthMagic on driver side instead.
Signed-off-by: Tiago Vignatti <tiago.vignatti@nokia.com>
Signed-off-by: Pauli Nieminen <ext-pauli.nieminen@nokia.com>
Reviewed-by: Kristian Høgsberg <krh@bitplanet.net>
1461b31e40 tests to see if the allocation
of buffer[i] failed after i has been incremented, causing it to look
at the wrong pointer.
Signed-off-by: Kristian Høgsberg <krh@bitplanet.net>
Tested-by: Eric Anholt <eric@anholt.net>
Signed-off-by: Keith Packard <keithp@keithp.com>
If driver fails to allocate memory for dri2 buffer server would crash
in send_buffers_reply.
Solution is to handle the allocation failure and return BadAlloc to
client.
Signed-off-by: Pauli Nieminen <ext-pauli.nieminen@nokia.com>
Reviewed-by: Michel Dänzer <michel@daenzer.net>
Reviewed-by: Kristian Høgsberg <krh@bitplanet.net>
Signed-off-by: Keith Packard <keithp@keithp.com>
This lets the DRI2 clients rely on the server to notify them when they
need to get new buffers. Without this, OpenGL clients poll the server
in glViewport() which can be a performance problems and also isn't
completely correct behaviour.
We bump the DRI2 protocol minor to indicate the availability of the
event, which the DRI2 clients can use to avoid polling. This speeds up
various piglit and oglc test cases as well as real applications.
Signed-off-by: Kristian Høgsberg <krh@bitplanet.net>
Reviewed-by: Adam Jackson <ajax@redhat.com>
Signed-off-by: Keith Packard <keithp@keithp.com>
The only remaining X-functions used in server are XNF*, the rest is converted to
plain alloc/calloc/realloc/free/strdup.
X* functions are still exported from server and x* macros are still defined in
header file, so both ABI and API are not affected by this change.
Signed-off-by: Mikhail Gusarov <dottedmag@dottedmag.net>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
If we don't have a fake front, nothing will happen. The fix was extracted
from a bigger patch from Francisco Jerez.
https://bugs.freedesktop.org/show_bug.cgi?id=27305
Signed-off-by: Kristian Høgsberg <krh@bitplanet.net>
Reviewed-by: Pauli Nieminen <suokkos@gmail.com>
Signed-off-by: Keith Packard <keithp@keithp.com>
Some pixmaps (window pixmaps and scratch pixmaps) don't have the
drawable->id set and thus DRI2 gets confused when using that field
for looking up the DRI2 drawable. Go back to using privates for getting
at the DRI2 drawable from a DrawablePtr. We need to keep the resource
tracking in place so we can remove the DRI2 drawable when the X resource
it was created for goes away. Additionally, we also now track the DRI2
drawable using a client XID so we can reclaim the DRI2 drawable even if
the client goes before the drawable and doesn't destroy the DRI2 drawable.
Tested-by: Owen W. Taylor <otaylor@fishsoup.net>
Signed-off-by: Kristian Høgsberg <krh@bitplanet.net>
The main motivation here is to have the resource system clean up the
DRI2 drawable automatically so glx doesn't have to. Right now, the
glx drawable resource must be destroyed before the X drawable, so that
calling DRI2DestroyDrawable doesn't crash. By making the DRI2
drawable a resource, GLX doesn't have to worry about that and the
resource destruction order becomes irrelevant.
https://bugs.freedesktop.org/show_bug.cgi?id=26394
Signed-off-by: Kristian Høgsberg <krh@bitplanet.net>
Signed-off-by: Keith Packard <keithp@keithp.com>
Export DDX swap control status from the DRI2 module and check for it in
GLX when initializing extensions.
Reviewed-by: Mario Kleiner <mario.kleiner@tuebingen.mpg.de>
Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
If a few swaps were queued leading to a throttle related block on the
client, and then the client submitted an MSC wait, one of the previous
swap wakeups could have caused the MSC wait to complete early. Add a
flag for this to prevent a swap wake from prematurely waking an MSC
waiter.
Reported-by: Mario Kleiner <mario.kleiner@tuebingen.mpg.de>
Reviewed-by: Mario Kleiner <mario.kleiner@tuebingen.mpg.de>
Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Avoid a potential swapsPending underflow by incrementing it before
ScheduleSwap, which may complete it immediately. And be sure to
decrement it again in case the schedule failed.
Reported-by: Mario Kleiner <mario.kleiner@tuebingen.mpg.de>
Reviewed-by: Mario Kleiner <mario.kleiner@tuebingen.mpg.de>
Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Update our supported DRI2 protocol version as each driver does
DRI2ScreenInit, since depending on available kernel features, each DDX
may support different callbacks and therefore protocol.
Reviewed-by: Mario Kleiner <mario.kleiner@tuebingen.mpg.de>
Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
A 0 swap interval means that swaps shouldn't be sync'd to vblank, so
just complete the swap immediately in that case.
Reviewed-by: Mario Kleiner <mario.kleiner@tuebingen.mpg.de>
Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Handle drawable destruction and lifetime correctly.
Check whether the drawable priv is valid in DRI2SwapInterval(),
DRI2WaitSBC() and DRI2WaitMSC(); it may have gone away, so be sure to
check it before using it.
If more than 1 outstanding swap is queued, we may complete several after
an app has exited. If we free it after the first one completes and the
refcount reaches 0, we'll crash the server on subsequent completions.
So delay freeing until all swaps complete and remove the error message
as this is a normal occurence. To do this properly, we must also avoid
destroying drawables in DRI2DestroyDrawable() if a swap or wait event is
pending.
And finally, make sure we free drawables in DRI2WaitMSCComplete() if
necessary (i.e. if the refcount has reached 0 and this MSC was the last
pending event on the object).
Reported-by: Mario Kleiner <mario.kleiner@tuebingen.mpg.de>
Reviewed-by: Mario Kleiner <mario.kleiner@tuebingen.mpg.de>
Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Returns expected SBC after completion of swap to caller, as required by
OML_sync_control spec, instead of the last_swap_target value.
Passes target_msc, divisor, remainder, correctly for
glXSwapBuffersMscOML() call, while retaining old behaviour for simple
glXSwapBuffers() call.
An OML swap can have a 0 target_msc, which just means it needs to
satisfy the divisor/remainder equation. Pass this down to the driver as
needed so we can support it.
Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Signed-off-by: Mario Kleiner <mario.kleiner@tuebingen.mpg.de>
Added implementation for case target_sbc == 0. In that case, the
function shall schedule a wait until all pending swaps for the drawable
have completed.
Fix for non-blocking case. Old implementation returned random,
uninitialized values for (ust,msc,sbc) if it returned immediately
without scheduling a wait due to sbc >= target_sbc.
Now if function doesn't schedule a wait, but returns immediately,
it returns the (ust,msc,sbc) of the most recently completed swap,
i.e., the UST and MSC corresponding to the time when the returned
current SBC was reached.
Signed-off-by: Mario Kleiner <mario.kleiner@tuebingen.mpg.de>
DRI2SwapComplete(): Increment pPriv->swap_count++; before calling
into callback for INTEL_swap_events extension, so the swap event
contains the current SBC after swap completion instead of the
previous one.
DRI2WakeupClient: Check for pPriv->target_sbc <= pPriv->swap_count,
had wrong comparison pPriv->target_sbc >= pPriv->swap_count for
unblocking of clients of DRI2WaitSBC().
Signed-off-by: Mario Kleiner <mario.kleiner@tuebingen.mpg.de>
We need to track invalid targets as well as 0 targets, so just make it
signed so our comparisons work like they should.
Reviewed-by: Mario Kleiner <mario.kleiner@tuebingen.mpg.de>
Reported-by: Kristian Høgsberg <krh@bitplanet.net>
Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
We need to initialize the swap target, which is passed to the driver to
schedule events. Rather than using -1 to indicate that the field is
uninitialized, just make sure we initialize it at drawable creation
time.
Reviewed-by: Mario Kleiner <mario.kleiner@tuebingen.mpg.de>
Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
It can be quite an expensive operation, so we're better off not doing
it unless it's totally required.
Signed-off-by: Francisco Jerez <currojerez@riseup.net>
Signed-off-by: Kristian Høgsberg <krh@bitplanet.net>
Reviewed-by: Kristian Høgsberg <krh@bitplanet.net>
Simon reported an issue with kwin that turned out to be a general problem. If
a drawable goes away before its swap completes, we'll try to free it up.
However, we free it improperly, which causes a server crash in
DRI2DestroyDrawable. Fix that up by splitting the free code out and calling
it from DRI2SwapComplete.
Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Reviewed-by: Michel Dänzer <michel@daenzer.net>
Signed-off-by: Keith Packard <keithp@keithp.com>