When using DRI3+Present with PRIME render offload, sometimes there is
a mismatch between the stride of the to-be-presented Pixmap and the
frontbuffer. The current code would reject a pageflip present in this
case if atomic modesetting is not enabled, ie. always, as atomic
modesetting is disabled by default due to brokeness in the current
modesetting-ddx.
Fullscreen presents without page flipping however trigger the copy
path as fallback, which causes not only unreliable presentation timing
and degraded performance, but also massive tearing artifacts due to
rendering to the framebuffer without any hardware sync to vblank.
Tearing is extra awful on modesetting-ddx because glamor afaics seems
to use drawing of a textured triangle strip for the copy implementation,
not a dedicated blitter engine. The rasterization pattern creates extra
awful tearing artifacts.
We can do better: According to a tip from Michel Daenzer (thanks!),
at least atomic modesetting capable kms drivers should be able to
reliably change scanout stride during a pageflip, even if atomic
modesetting is not actually enabled for the modesetting client.
This commit adds detection logic to find out if the underlying kms
driver is atomic_modeset_capable, and if so, it no longer rejects
page flip presents on mismatched stride between new Pixmap and
frontbuffer.
We (ab)use a call to drmSetClientCap(ms->fd, DRM_CLIENT_CAP_ATOMIC, 0);
for this purpose. The call itself has no practical effect, as it
requests disabling atomic mode, although atomic mode is disabled by
default. However, the return value of drmSetClientCap() tells us if the
underlying kms driver is atomic modesetting capable: An atomic driver
will return 0 for success. A legacy non-atomic driver will return a
non-zero error code, either -EINVAL for early atomic Linux versions
4.0 - 4.19 (or for non-atomic Linux 3.x and earlier), or -EOPNOTSUPP
for Linux 4.20 and later.
Testing on a MacBookPro 2017 with Intel Kabylake display server gpu +
AMD Polaris11 as prime renderoffload gpu, X-Server master + Mesa 21.0.3
show improvement from unbearable tearing to perfect, despite a stride
mismatch between display gpu and Pixmap of 11776 Bytes vs. 11520
Bytes. That this is correct behaviour was also confirmed by comparing the
behaviour and .check_flip implementation of the patched modesetting-ddx
against the current intel-ddx SNA Present implementation.
Please consider merging this patch before the server-1.21 branch point.
This patch could also be cherry-picked into the server 1.20 branch to
fix the same limitation.
Signed-off-by: Mario Kleiner <mario.kleiner.de@gmail.com>
(cherry picked from commit 8f8ebf870b)
A lot of that code is the same as in xf86-amdgpu and xf86-nouveau drivers. By removing that functions from
ms namespace we can move that code to common implementation.
Signed-off-by: Łukasz Spintzyk <lukasz.spintzyk@synaptics.com>
(cherry picked from commit 5be3b80b8d)
The miPointerSpriteFunc swcursor code expects there to only be a single
framebuffer and when the cursor moves it will undo the damage of the
previous draw, potentially overwriting what ever is there in a new
framebuffer installed after a flip.
This leads to all kind of artifacts, so we need to disable pageflipping
when a swcursor is used.
The code for this has shamelessly been copied from the xf86-video-amdgpu
code.
Fixes: https://gitlab.freedesktop.org/xorg/xserver/issues/828
Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
(cherry picked from commit 0aaac8d783)
Signed-off-by: Łukasz Spintzyk <lukasz.spintzyk@displaylink.com>
The atomic driver has issues with modesetting when stealing
connectors from a different crtc, a black screen when doing rotation
on a different crtc, and in general is just a mapping of the legacy
helpers to atomic. This is already done in the kernel, so just
fallback to legacy by default until this is fixed.
Please backport to 1.20, as we don't want to enable it for everyone
there. It breaks for existing users.
The fixes to make the xserver more atomic have been pending on the
mailing list for ages.
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=110375
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=110030
References: https://gitlab.freedesktop.org/xorg/xserver/merge_requests/36/commits
Signed-off-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
(cherry picked from commit f0d78b47ac)
ms_present_get_crtc() returns an RRCrtcPtr, but derives it from a xf86CrtcPtr
found via ms_dri2_crtc_covering_drawable()=>ms_covering_crtc(). As a result, it
depends on all associated DIX ScreenRecs having an xf86CrtcConfigPtr DDX
private.
Some DIX ScreenRecs don't have an xf86CrtcConfigPtr DDX private, but do have an
rrScrPrivPtr DDX private. Given that we can derive all of the information we
need from RandR, we can support these screens by avoiding the use of xf86Crtc.
This change implements an RandR-based path for ms_present_get_crtc(), allowing
drawables to successfully fall back to syncing to the primary output, even if
the slave doesn't have an xf86CrtcConfigPtr DDX private.
Without this change, if a slave doesn't have an xf86CrtcConfigPtr DDX private,
drawables will fall back to 1 FPS if they overlap an output on that slave.
Signed-off-by: Alex Goins <agoins@nvidia.com>
(cherry picked from commit 562c7888be)
This lets an application open a suitable DRM device and pass the file
descriptor to the mode setting driver through an X server command line
option, '-masterfd'.
There's a companion application, xlease, which creates a DRM master by
leasing an output from another X server. That is available at
git clone git://people.freedesktop.org/~keithp/xlease
v2:
Always print usage, but note that it can't be used if
setuid/gid
Suggested-by: Lyude Paul <lyude@redhat.com>
Signed-off-by: Keith Packard <keithp@keithp.com>
Reviewed-by: Lyude Paul <lyude@redhat.com>
(cherry picked from commit 38ff29ec8e)
The old 32-Bit wraparound handling didn't actually work, due to some
integer casting bug, and the mapping was ill equipped to deal with input
from the new true 64-bit GetCrtcSequence/QueueCrtcSequence api's
introduced in Linux 4.15.
For 32-Bit truncated input from pageflip events and old vblank events
and old drmWaitVblank ioctl, implement new wraparound handling, which
also allows to deal with wraparound in the other direction, e.g., if a
32-Bit truncated sequence value is passed in, whose true 64-Bit
in-kernel hw value is within 2^30 counts of the previous processed
value, but whose 32-bit truncated sequence value happens to lie just
above or below a 2^32 boundary, iow. one of the two values 'sequence'
vs. 'msc_prev' lies above a 2^32 border, the other one below it.
The method is directly translated from Mesa's proven implementation of
the INTEL_swap_events extension, where a true underlying 64-Bit wide
swapbuffers count (SBC) needs to get reconstructed from a 32-Bit LSB
truncated SBC transported over the X11 protocol wire. Same conditions
apply, ie. successive true 64-Bit SBC values are close to each other,
but don't always get received in strictly monotonically increasing
order. See Mesa commit cc5ddd584d17abd422ae4d8e83805969485740d9 ("glx:
Handle out-of-sequence swap completion events correctly. (v2)") for
explanation.
Additionally add a separate path for true 64-bit msc input originating
from Linux 4.15+ drmCrtcGetSequence/QueueSequence ioctl's and
corresponding 64-bit vblank events. True 64-bit msc's don't need
remapping and must be passed through.
As a reliability bonus, they are also used here to update the tracking
values msc_prev and ms_high with perfect 64-Bit ground truth as baseline
for mapping msc from pageflip completion events, because pageflip events
are always 32-bit wide, even when the new kernel api's are used. Because
each pageflip(-event) is always preceeded close in time (and vblank
count) by a drmCrtcQueueSequence queued event or drmCrtcGetSequence
query as part of DRI2 or DRI3+Present swap scheduling, we can be certain
that each pageflip event will get its truncated 32-bit msc remapped
reliably to the true 64-bit msc of flip completion whenever the sequence
api is available, ie. on Linux 4.15 or later.
Note: In principle at least the 32-bit mapping path could also be
backported to earlier server branches, as this seems to be broken for at
least server 1.16 to 1.19.
Signed-off-by: Mario Kleiner <mario.kleiner.de@gmail.com>
Reviewed-by: Adam Jackson <ajax@redhat.com>
Cc: Keith Packard <keithp@keithp.com>
Cc: Michel Dänzer <michel.daenzer@amd.com>
The function is ported from intel-ddx uxa backend around 2013, where its
stated purpose was to apply a vblank_offset to msc values to correct for
problems with those kernel provided msc values. Some (somewhat magic and
puzzling to myself) heuristic tried to guess if provided values were
unreasonable and tried to adapt the corrective vblank_offset to account
for that.
Except: It wasn't applied to kernel provided msc values, but the values
delivered by clients via DRI2 or Present, so valid client targetmsc
values, e.g., requesting a vblank event > 1000 vblanks in the future,
triggered the offset correction in arbitrarily wrong ways, leading to
wrong msc values being returned and thereby vblank events queued to the
kernel for the wrong time. This causes glXSwapBuffersMscOML and
glXWaitForMscOML to swap / return immediately whenever a swap/wait in >
1000 vblanks is requested.
The original code was also written to only deal with 32 bit mscs, but
server 1.20 modesetting ddx can now use new Linux 4.15+ kernel vblank
api to process true 64 bit msc's, which may confuse the heuristic even
more due to 32 bit integer truncation/wrapping.
This code caused various problems in the intel-ddx in the past since
year 2013, and was removed there in 2015 by Chris Wilson in commit
42ebe2ef9646be5c4586868cf332b4cd79bb4618:
" uxa: Remove the filtering of bogus Present MSC values
If the intention was to filter the return values from the kernel, the
filtering would have been applied to the kernel values and not to the
incoming values from Present. This filtering introduces crazy integer
promotion and truncation bugs all because Present feeds garbage into its
vblank requests.
"
Indeed, i found a Mesa bug yesterday which can cause Mesa's
PresentPixmap request to spuriously feed garbage targetMSC's into the
driver under some conditions. However, while other video drivers seem to
cope relatively well with that, modesetting ddx causes KDE-5's
plasmashell to lock up badly quite frequently, and my suspicion is that
the code removed in this commit is one major source of the extra
fragility.
Also my own tests fail for any swap scheduled more than 1000 vblanks
into the future, which is not uncommon for some scientific applications.
Iow. modesetting's swap scheduling seems to be more robust without this
function afaics.
Signed-off-by: Mario Kleiner <mario.kleiner.de@gmail.com>
Reviewed-by: Adam Jackson <ajax@redhat.com>
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Keith Packard <keithp@keithp.com>
Tested-by: Mike Lothian <mike@fireburn.co.uk>
Use the DRM_CAP_ADDFB2_MODIFIERS query to make sure the kms
driver supports modifiers in the addfb2 ioctl, and fall back
to addfb ioctl without modifiers if modifiers are unsupported.
E.g., as of Linux 4.17, nouveau-kms so far does not suppport
modifiers and gets angry if drmModeAddFB2WithModifiers() is
called (-> failure to set a video mode -> blank screen), but
Mesa's nvc0+ gallium driver causes gbm_bo_get_modifier() to
return a valid modifier by translating the default tiling of
bo's created via gbm_bo_create() into a modifier other than
DRM_FORMAT_MOD_INVALID (see Mesa's nvc0_miptree_get_modifier()).
Testing for != DRM_FORMAT_MOD_INVALID is apparently not
sufficient for safe use of drmModeAddFB2WithModifiers.
Bonus: Handle potential failure of populate_format_modifiers().
The required DRM_CAP is defined since libdrm v2.4.65, and we
require v2.4.89+ for the server, so we can use it unconditionally.
Tested on intel-kms, radeon-kms, nouveau-kms. Fixes failure on
NVidia Pascal.
Fixes: 2f807c2324 ("modesetting: Add support for multi-plane pixmaps when page-flipping")
Signed-off-by: Mario Kleiner <mario.kleiner.de@gmail.com>
Cc: Daniel Stone <daniels@collabora.com>
Cc: Louis-Francis Ratté-Boulianne <lfrb@collabora.com>
Reviewed-by: Louis-Francis Ratté-Boulianne <lfrb@collabora.com>
CRTCs and outputs needs to be enabled/disabled when the current
DPMS mode is changed. We also try to do it in an atomic commit
when possible.
Signed-off-by: Louis-Francis Ratté-Boulianne <lfrb@collabora.com>
Tested-by: Daniel Stone <daniels@collabora.com>
Reviewed-by: Adam Jackson <ajax@redhat.com>
In order to flip between compressed and uncompressed buffers -
something drmModePageFlip explicitly bans us from doing - we need
to port use the atomic modesetting API. It's only 'fake' atomic
though given we still commit for each CRTC separately and
CRTC and connector properties are not set with the atomic API.
The helper functions to retrieve DRM properties have been borrowed
from Weston.
Signed-off-by: Louis-Francis Ratté-Boulianne <lfrb@collabora.com>
Reviewed-by: Daniel Stone <daniels@collabora.com>
Acked-by: Keith Packard <keithp@keithp.com>
Reviewed-by: Adam Jackson <ajax@redhat.com>
This really just wants to be the list of disable booleans and
initialization functions, and nothing else. Stop including the protocol
headers from extinit.h, remove a stray mention of xgl, and move an
XInput declaration to a better place.
v2: A bunch of drivers assume they'll get the DPMS tokens implicitly,
so add it to globals.h.
Signed-off-by: Adam Jackson <ajax@redhat.com>
Reviewed-by: Eric Anholt <eric@anholt.net>
This provides an API wrapper around the kernel interface for queueing
a vblank event, simplifying all of the callers.
v2: Fix missing '|' in computing vbl.request.type
v3: Remove spurious bit of next patch (thanks, Michel Dänzer)
Signed-off-by: Keith Packard <keithp@keithp.com>
Reviewed-by: Adam Jackson <ajax@redhat.com>
The glamor_egl module that the GLAMOR paths are using is only built if
GLAMOR_HAS_GBM is true, and there's no plan for implementing the
module without GBM. Simplify modesetting's code as a result.
Reviewed-by: Adam Jackson <ajax@redhat.com>
Signed-off-by: Eric Anholt <eric@anholt.net>
Move ms_flush_drm_events out of GLAMOR ifdef.
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=97586
Signed-off-by: Qiang Yu <Qiang.Yu@amd.com>
Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
Signed-off-by: Keith Packard <keithp@keithp.com>
Remove unused arguments from ms_covering_crtc, make it static as it is
only used in vblank.c.
While at it also change its first argument from a ScrnInfoPtr to a
ScreenPtr, this makes the next patch in this patch-set cleaner.
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Reviewed-by: Keith Packard <keithp@keithp.com>
The common page flip handle framework can be shared with DRI2
page flip.
Signed-off-by: Qiang Yu <Qiang.Yu@amd.com>
Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
Server GPUs often have a VNC feature attached to allow remote console.
The controller implementing this feature is usually not very powerful,
and we can easily swamp it with work. This is made somewhat worse by
damage over-reporting the size of the dirty region, and a whole lot
worse by applications (or shells) that update the screen with identical
pixel content as was already there.
Fix this by double-buffering the shadow fb, using memcmp to identify
dirty tiles on each update pass. Since both shadows are in host memory
the memcmp is cheap, and worth it given the win in network bandwidth.
The tile size is somewhat arbitrarily chosen to be one cacheline wide at
32bpp on Intel Core.
By default we enable this behaviour for (a subset of) known server GPUs;
the heuristic could use work.
Signed-off-by: Adam Jackson <ajax@redhat.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
This adds support using glamor for background None.
loosely based off the amdgpu code. relies on the glamor_finish code.
Acked-by: Eric Anholt <eric@anholt.net>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
This isn't used anywhere, so no point storing it until we need it.
Reviewed-by: Mark Kettenis <kettenis@openbsd.org>
Signed-off-by: Dave Airlie <airlied@redhat.com>
This adds zaphod and ZaphodHeads support
to the the in-server modesetting driver.
this is based on a request from Mario,
and on the current radeon driver, along
with some patches from Mario to bring things
up to the state of the art in Zaphod.
v2: fixup vblank fd registring.
v3: squash Mario's fixes.
modesetting: Allow/Fix use of multiple ZaphodHead outputs per x-screen.
modesetting: Take shift in crtc positions for ZaphodHeads configs into account.
modesetting: Add ZaphodHeads description to man page.
small cleanups (airlied).
Reviewed-and-tested-by: Mario Kleiner <mario.kleiner.de@gmail.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
Based on code by Keith Packard, Eric Anholt, and Jason Ekstrand.
v2:
- Fix double free and flip_count underrun (caught by Mario Kleiner).
- Don't leak flip_vblank_event on the error_out path (Mario).
- Use the updated ms_flush_drm_events API (Mario, Ken).
v3: Hack around DPMS shenanigans. If all monitors are DPMS off, then
there is no active framebuffer; attempting to pageflip will hit the
error_undo paths, causing us to drmModeRmFB with no framebuffer,
which confuses the kernel into doing full modesets and generally
breaks things. To avoid this, make ms_present_check_flip check that
some CRTCs are enabled and DPMS on. This is an ugly hack that would
get better with atomic modesetting, or some core Present work.
v4:
- Don't do pageflipping if CRTCs are rotated (caught by Jason Ekstrand).
- Make pageflipping optional (Option "PageFlip" in xorg.conf.d), but
enabled by default.
v5: Initialize num_crtcs_on to 0 (caught by Michel Dänzer).
[airlied: took over]
v6: merge async flip support from Mario Kleiner
free sequence after failed vblank queue
handle unflip while DPMS'ed off (Michel)
move flip tracking into its own structure, and
fix up reference counting issues, and add comments.
Signed-off-by: Dave Airlie <airlied@redhat.com>
Acked-by: Eric Anholt <eric@anholt.net>
Signed-off-by: Kenneth Graunke <kenneth@whitecape.org>
This is a specialization of ms_drm_abort that matches based on the drm
event queue's sequence number.
Based on code by Keith Packard.
Reviewed-by: Eric Anholt <eric@anholt.net>
Reviewed-by: Dave Airlie <airlied@redhat.com>
Signed-off-by: Kenneth Graunke <kenneth@whitecape.org>
I want to use this in present.c.
Reviewed-by: Dave Airlie <airlied@redhat.com>
Reviewed-by: Eric Anholt <eric@anholt.net>
Signed-off-by: Kenneth Graunke <kenneth@whitecape.org>
modesetting hooked up vblank support for DRI2, but was missing support
for vblanks in Present.
This is mostly copy and pasted from Keith's code in the intel driver.
v2: Use ms_crtc_msc_to_kernel_msc in ms_present_queue_vblank to hook
up the vblank_offset workaround for bogus MSC values (which the
DRI2 code already did).
Also simplify the ms_present_get_crtc function. vblank.c already
implements the functionality; we just need to convert types.
v3: Fix ms_flush_drm_events return code. I'd copied code where 0 meant
success into a function that returned a boolean, so the return code
was always backwards.
Also add DebugPresent calls in ms_present_vblank_{handler,abort}.
Signed-off-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Keith Packard <keithp@keithp.com>
Tested-by: Jason Ekstrand <jason.ekstrand@intel.com>
Signed-off-by: Keith Packard <keithp@keithp.com>
We basically want it throughout the driver.
Signed-off-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Keith Packard <keithp@keithp.com>
Tested-by: Jason Ekstrand <jason.ekstrand@intel.com>
Signed-off-by: Keith Packard <keithp@keithp.com>
Move the boolean glamor from struct modesetting into struct drmmode for
later re-use in drmmode_display.
Signed-off-by: Daniel Martin <consume.noise@gmail.com>
Reviewed-by: Eric Anholt <eric@anholt.net>
Reviewed-by: Keith Packard <keithp@keithp.com>
Signed-off-by: Keith Packard <keithp@keithp.com>
This is derived from the intel driver DRI2 code, with swapchain and
pageflipping dropped, functions renamed, and vblank event management
shared code moved to a vblank.c for reuse by Present.
This allows AIGLX to load, which means that you get appropriate
visuals exposed in GL, along with many extensions under
direct-rendering that require presence in GLX (which aren't supported
in glxdriswrast.c).
v2: Drop unused header includes in pageflip.c, wrap in #ifdef GLAMOR.
Drop triple-buffering, which was totally broken in practice (I'll
try to fix this later). Fix up some style nits. Document the
general flow of pageflipping and why, rename the DRI2 frame event
type enums to reflect what they're for, and handle them in a
single switch statement so you can understand the state machine
more easily.
v3: Drop pageflipping entirely -- it's unstable on my Intel laptop
(not that the normal 2D driver is stable with pageflipping for
me), and I won't get it fixed before the merge window. It now
passes all of the OML_sync_control tests from Jamey and Theo
(except for occasional warns in timing -fullscreen -divisor 2).
v4: Fix doxygen at the top of vblank.c
Signed-off-by: Eric Anholt <eric@anholt.net>
Reviewed-by: Adam Jackson <ajax@redhat.com>
By default modesetting now tries to enable X acceleration using
glamor, but falls back to normal shadowfb if GL fails to initialize.
Signed-off-by: Eric Anholt <eric@anholt.net>
Reviewed-by: Keith Packard <keithp@keithp.com>
As I was editing code, the top-level .dir-locals.el was making my new
stuff conflict with the existing style. Make it consistently use the
xorg style, instead.
Signed-off-by: Eric Anholt <eric@anholt.net>
Reviewed-by: Keith Packard <keithp@keithp.com>
Mario Kleiner
Łukasz Spintzyk
Hans de Goede
Maarten Lankhorst
Alex Goins
Keith Packard
Louis-Francis Ratté-Boulianne
Adam Jackson
Daniel Martin
Eric Anholt
Qiang Yu
Dave Airlie
Dave Airlie
Kenneth Graunke
Alex Deucher
Alan Hourihane