XLibre/xf86-video-amdgpu
1
0
Fork 0
mirror of https://github.com/X11Libre/xf86-video-amdgpu.git synced 2026-03-24 01:24:31 +00:00
Code Issues Packages Projects Releases Wiki Activity
Files
533bd30ceaa373788b3d0bfd4d486f0f1c624d0c
xf86-video-amdgpu/man/amdgpu.man
Mario Kleiner 65c127366a Add option for non-vsynced flips for "secondary" outputs. 
This is a straightforward port of a patch with the same name
"modesetting: Add option for non-vsynced flips for "secondary"
outputs." from X-Server master / X-Server 21.1. See server MR 742.
The description below is therefore identical to that X-Server commit:

Whenever an unredirected fullscreen window uses pageflipping for a
DRI3/Present PresentPixmap() operation and the X-Screen has more than
one active output, multiple crtc's need to execute pageflips. Only
after the last flip has completed can the PresentPixmap operation
as a whole complete.

If a sync_flip is requested for the present, then the current
implementation will synchronize each pageflip to the vblank of
its associated crtc. This provides tear-free image presentation
across all outputs, but introduces a different artifact, if not
all outputs run at the same refresh rate with perfect synchrony:
The slowest output throttles the presentation rate, and present
completion is delayed to flip completion of the "latest" output
to complete. This means degraded performance, e.g., a dual-display
setup with a 144 Hz monitor and a 60 Hz monitor will always be
throttled to at most 60 fps. It also means non-constant present
rate if refresh cycles drift against each other, creating complex
"beat patterns", tremors, stutters and periodic slowdowns - quite
irritating!

Such a scenario will be especially annoying if one uses multiple
outputs in "mirror mode" aka "clone mode". One output will usually
be the "production output" with the highest quality and fastest
display attached, whereas a secondary mirror output just has a
cheaper display for monitoring attached. Users care about perfect
and perfectly timed tear-free presentation on the "production output",
but cares less about quality on the secondary "mirror output". They
are willing to trade quality on secondary outputs away in exchange
for better presentation timing on the "production output".

One example use case for such production + monitoring displays are
neuroscience / medical science applications where one high quality
display device is used to present visual animations to test subjects
or patients in a fMRI scanner room (production display), whereas
an operator monitors the same visual animations from a control room
on a lower quality display. Presentation timing needs to be perfect,
and animations high-speed and tear-free for the production display,
whereas quality and timing don't matter for the monitoring display.

This commit gives users the option to choose such a trade-off as
opt-in:

It adds a new boolean option "AsyncFlipSecondaries" to the device section
of xorg.conf. If this option is specified as true, then DRI3 pageflip
behaviour changes as follows:

1. The "reference crtc" for a windows PresentPixmap operation does a
   vblank synced flip, or a DRM_MODE_PAGE_FLIP_ASYNC non-synchronized
   flip, as requested by the caller, just as in the past. Typically
   flips will be requested to be vblank synchronized for tear-free
   presentation. The "reference crtc" is the one chosen by the caller
   to drive presentation timing (as specified by PresentPixmap()'s
   "target_msc", "divisor", "remainder" parameters and implemented by
   vblank events) and to deliver Present completion timestamps (msc
   and ust) extracted from its pageflip completion event.

2. All other crtc's, which also page-flip in a multi-display configuration,
   will try to flip with DRM_MODE_PAGE_FLIP_ASYNC, ie. immediately and
   not synchronized to vblank. This allows the PresentPixmap operation
   to complete with little delay compared to a single-display present,
   especially if the different crtc's run at different video refresh
   rates or their refresh cycles are not perfectly synchronized, but
   drift against each other. The downside is potential tearing artifacts
   on all outputs apart from the one of the "reference crtc".

Successfully tested on a AMD gpu with single-display and dual-display
setups, and with single-X-Screen as well as dual-X-Screen "ZaphodHeads"
configurations.

Signed-off-by: Mario Kleiner <mario.kleiner.de@gmail.com>
2022-01-25 03:59:15 +01:00

171 lines
5.1 KiB
Groff
Raw Blame History

.ds q \N'34'
.TH AMDGPU __drivermansuffix__ __vendorversion__
.SH NAME
amdgpu \- AMD RADEON GPU video driver
.SH SYNOPSIS
.nf
.B "Section \*qDevice\*q"
.BI " Identifier \*q" devname \*q
.B " Driver \*qamdgpu\*q"
\ \ ...
.B EndSection
.fi
.SH DESCRIPTION
.B amdgpu
is an __xservername__ driver for AMD RADEON-based video cards with the
following features:
.PP
.PD 0
.TP 2
\(bu
Support for 8-, 15-, 16-, 24- and 30-bit pixel depths;
.TP
\(bu
RandR support up to version 1.4;
.TP
\(bu
3D acceleration;
.PD
.SH SUPPORTED HARDWARE
The
.B amdgpu
driver supports SI and newer families' video cards.
.PD
.SH CONFIGURATION DETAILS
Please refer to __xconfigfile__(__filemansuffix__) for general configuration
details. This section only covers configuration details specific to this
driver.
.PP
The following driver
.B Options
are supported:
.TP
.BI "Option \*qSWcursor\*q \*q" boolean \*q
Selects software cursor. The default is
.B off.
.TP
.BI "Option \*qAccel\*q \*q" boolean \*q
Enables or disables all hardware acceleration.
.br
The default is
.B on.
.TP
.BI "Option \*qZaphodHeads\*q \*q" string \*q
Specify the RandR output(s) to use with zaphod mode for a particular driver
instance. If you use this option you must use this option for all instances
of the driver.
.br
For example:
.B
Option \*qZaphodHeads\*q \*qLVDS,VGA-0\*q
will assign xrandr outputs LVDS and VGA-0 to this instance of the driver.
.TP
.BI "Option \*qDRI\*q \*q" integer \*q
Define the maximum level of DRI to enable. Valid values are 2 for DRI2 or 3 for DRI3.
The default is
.B 3 for DRI3
if the Xorg version is >= 1.18.3, otherwise
.B 2 for DRI2.
.TP
.BI "Option \*qEnablePageFlip\*q \*q" boolean \*q
Enable DRI2 page flipping. The default is
.B on.
.TP
.BI "Option \*qTearFree\*q \*q" boolean \*q
Set the default value of the per-output 'TearFree' property, which controls
tearing prevention using the hardware page flipping mechanism. TearFree is
on for any CRTC associated with one or more outputs with TearFree on. Two
separate scanout buffers need to be allocated for each CRTC with TearFree
on. If this option is set, the default value of the property is 'on' or 'off'
accordingly. If this option isn't set, the default value of the property is
.B auto,
which means that TearFree is on for rotated outputs, outputs with RandR
transforms applied and for RandR 1.4 secondary outputs, otherwise off.
.TP
.BI "Option \*qVariableRefresh\*q \*q" boolean \*q
Enables support for enabling variable refresh on the Screen's CRTCs
when an suitable application is flipping via the Present extension.
.br
The default is
.B off.
.TP
.BI "Option \*qAsyncFlipSecondaries\*q \*q" boolean \*q
Use async flips for secondary video outputs on multi-display setups. If a screen
has multiple displays attached and DRI3 page flipping is used, then only one of
the displays will have its page flip synchronized to vblank for tear-free
presentation. This is the display that is used for presentation timing and
timestamping, usually the one covering the biggest pixel area of the screen.
All other displays ("Secondaries") will not synchronize their flips. This may
cause some tearing on these displays, but it prevents a permanent or periodic
slowdown or irritating judder of animations if not all video outputs are running
synchronized with each other and with the same refresh rate. There is no perfect
solution apart from perfectly synchronized outputs, but this option may give
preferrable results if the displays in a multi-display setup mirror or clone
each other. The default is
.B off.
.TP
.BI "Option \*qAccelMethod\*q \*q" string \*q
Setting this option to
.B none
disables use of the glamor acceleration architecture. In that case, all 2D
rendering is done by the CPU, but 3D and video hardware acceleration can still
work. This is mainly useful for OpenGL driver bring-up.
.br
The default is to use glamor.
.PP
The following driver
.B Option
is supported for
.B glamor
:
.TP
.BI "Option \*qShadowPrimary\*q \*q" boolean \*q
This option enables a so-called "shadow primary" buffer for fast CPU access to
pixel data, and separate scanout buffers for each display controller (CRTC).
This may improve performance for some 2D workloads, potentially at the expense
of other (e.g. 3D, video) workloads.
Note in particular that enabling this option currently disables page flipping.
The default is
.B off.
.SH SEE ALSO
__xservername__(__appmansuffix__), __xconfigfile__(__filemansuffix__), Xserver(__appmansuffix__), X(__miscmansuffix__)
.IP " 1." 4
Wiki page:
.RS 4
https://www.x.org/wiki/radeon
.RE
.IP " 2." 4
Overview about amdgpu development code:
.RS 4
https://cgit.freedesktop.org/xorg/driver/xf86-video-amdgpu/
.RE
.IP " 3." 4
Mailing list:
.RS 4
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
.RE
.IP " 4." 4
IRC channel:
.RS 4
#radeon on irc.freenode.net
.RE
.IP " 5." 4
Query the bugtracker for amdgpu bugs:
.RS 4
https://bugs.freedesktop.org/query.cgi?product=xorg&component=Driver/AMDgpu
.RE
.IP " 6." 4
Submit bugs & patches:
.RS 4
https://bugs.freedesktop.org/enter_bug.cgi?product=xorg&component=Driver/AMDgpu
.RE
.SH AUTHORS
.nf
Authors include:
Michel D\(:anzer \fImichel@daenzer.net\fP
Alex Deucher \fIalexdeucher@gmail.com\fP
Reference in New Issue View Git Blame Copy Permalink