summaryrefslogtreecommitdiff
path: root/src
AgeCommit message (Collapse)Author
2015-06-29Fix inconsistent default eg_tile_split in evergreen_accel.cMario Kleiner
eg_tile_split() maps the default: switch-case to tile split flag 6, like for a tile split size of 4096 Bytes. All other instances of tile split mappings or reverse mappings in both the ddx and in Mesa's radeon gallium drivers assign the default: case to tile split flag 4, consistent with a default of 1024 Bytes. Make this one outlier consistent with other instances of mappings in the code to avoid surprises. Signed-off-by: Mario Kleiner <mario.kleiner.de@gmail.com>
2015-06-25Disable CRTCs when disabling the outputsPiotr Redlewski
When turning the outputs off (DPMSModeOff), CRTCs stayed enabled. This led to higher gpu temperatures than with fglrx driver. v2: when entering DPMS also disable active CRTCs v3: use drmmode_set_mode_major() for enabling CRTCs when leaving DPMS Signed-off-by: Piotr Redlewski <predlewski@gmail.com> Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
2015-06-25Skip disabled CRTCs in radeon_scanout_(do_)updateMichel Dänzer
The vblank / page flip ioctls don't work as expected for a disabled CRTC. Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
2015-06-22Don't set TILE_SPLIT flags if surface.tile_split == 0.Mario Kleiner
On pre-Evergreen hw, libdrm's r6_surface_best() helper for the surface managers radeon_surface_best() routine is a no-op and therefore doesn't assign any tile_split settings to created surfaces, so it leaves surface.tile_split on its "undefined" value of 0. Mesa's DRI3/Present backend creates DRI3 Pixmaps via the DRIImage extension and the radeon gallium driver implementation of that extension uses the libdrm surface manager for backing bo creation and treats an undefined surface.tile_split==0, as returned by the surface manager for pre-evergreen, as a signal to not assign any tile_split flags to the DRI3 Pixmaps bo. The ddx also uses libdrm surface manager to create the x-screen pixmap, but so far treated the returned undefined surface.tile_split==0 by mapping it to eg_tile_split()'s default tile_split flags, which are different from Mesa's tiling flags for DRI3 pixmaps. Under DRI3/Present this causes a mismatch of src pixmap and destination root pixmaps tiling flags and thereby prevents page flipping for pixmap presents. Change the ddx code to treat surface.tile_split==0 the same way as the radeon gallium driver to avoid mismatched tiling flags and thereby allow DRI3/Present page-flip to work on pre-Evergreen hw. Tested on RV730 and Evergreen "Juniper". Signed-off-by: Mario Kleiner <mario.kleiner.de@gmail.com>
2015-06-11Don't attempt a DRI2/Present page flip while the other one is flippingMichel Dänzer
Fixes corrupted display and hangs when switching between DRI2 and DRI3 fullscreen apps, e.g. a compositor using DRI3 and a fullscreen app using DRI2 or vice versa. Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
2015-06-11Allow/Fix use of multiple ZaphodHead outputs per x-screen. (v2)Mario Kleiner
Defining multiple ZaphodHead outputs per x-screen in a multiple x-screen's per gpu configuration caused all outputs except one per x-screen to go dark, because there was a fixed mapping x-screen number -> crtc number, limiting the number of crtc's per x-screen to one. On a ZaphodHead's setup, be more clever and assign as many crtc's to a given x-screen as there are ZaphodHeads defined for that screen, assuming there are enough unused crtc's available. Tested on a triple display setup with different combos of one, two or three ZaphodHeads per one, two or three x-screens. This is a port of similar code from xf86-video-nouveau. v2: Implement suggestions by Michel Dänzer: Less verbose debug output, more clear warning message on crtc allocation failure. Move clearing of per gpu assigned_crtc mask to CloseScreen, indeed testing shows no need for the more complex new server generation check from v1. Signed-off-by: Mario Kleiner <mario.kleiner.de@gmail.com> (v1) Reviewed-by: Alex Deucher <alexander.deucher@amd.com> Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
2015-06-01DRI2: Fix handling of drmmode_crtc_get_ust_msc return codeMichel Dänzer
Fixes regression introduced in commit 76c2923ac5c7230a8b2f9f8329c308d28b44d9c0 ("DRI2: Split out helper for getting UST and MSC of a specific CRTC"). Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
2015-05-28glamor: Only wait for GPU writes in radeon_glamor_prepare_access_cpu_roMichel Dänzer
We don't need to wait for GPU reads to finish before reading with the CPU. Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
2015-05-21present: Remove get_drmmode_crtc helperMichel Dänzer
It was getting the drmmode_crtc_private_ptr in a roundabout way. Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
2015-05-20Replace SyncFence typedef with declaration of struct _SyncFenceMichel Dänzer
Fixes build failure in some environments: CC radeon_sync.lo In file included from radeon_sync.c:28: /home/tinderbox/xorg-build/include/xorg/misync.h:31: error: redefinition of typedef 'SyncFence' radeon.h:93: note: previous declaration of 'SyncFence' was here make[2]: *** [radeon_sync.lo] Error 1 Reported-Tested-and-Acked-by: David Airlie <airlied@redhat.com>
2015-05-15glamor: Deal with glamor_glyphs_init being removed from xserverMichel Dänzer
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
2015-05-13add new bonaire pci idAlex Deucher
Reviewed-by: Michel Dänzer <michel.daenzer@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2015-05-11present: Fix present notify timestamps and counts.Mario Kleiner
Invalid (msc,ust)=(0,0) were returned to the server because a wrong crtc_id was used to specify which crtc should deliver pageflip completion data. Fix it in accordance with the dri2 implementation. Signed-off-by: Mario Kleiner <mario.kleiner.de@gmail.com>
2015-05-11present: Move check for async flipsMario Kleiner
Check for and reject currently unsupported async_flip inside radeon_present_check_flip() instead of inside radeon_present_flip(). This way the server can detect early that async flips aren't supported by the ddx and can deal with this correctly by using its non-vsync'ed CopyRegion fallback. Signed-off-by: Mario Kleiner <mario.kleiner.de@gmail.com>
2015-04-28Skip disabled CRTCs in drmmode_set_desired_modes() even if set_hw == FALSEMichel Dänzer
Not skipping a disabled CRTC results in a crash. Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=90187 Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
2015-04-27Don't call radeon_get_pixmap_private() when using EXAMichel Dänzer
It only works when using glamor. Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=90169 Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
2015-04-23On screen resize, clear the new buffer before displaying itMichel Dänzer
Fixes garbage being intermittently visible during a screen resize. Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=27757#c7 Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
2015-04-23Make drmmode_copy_fb() work with glamor as wellMichel Dänzer
Needed for Xorg -background none. Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
2015-04-23Update scanout pixmap contents before setting a mode with itMichel Dänzer
This ensures the scanout pixmaps used for Option "TearFree" and Option "ShadowPrimary" have been initialized when their initial mode is set. Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
2015-04-23Defer initial modeset until the first BlockHandler invocationMichel Dänzer
This ensures that the screen pixmap contents have been initialized when the initial modes are set. Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=27757 Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
2015-04-23Defer initial drmmode_copy_fb call until root window creationMichel Dänzer
That's late enough for acceleration to be fully initialized, but still early enough to set pScreen->canDoBGNoneRoot. Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
2015-04-23Only copy fbcon BO contents if bgNoneRoot is TRUEMichel Dänzer
Otherwise, the X server will initialize the screen pixmap contents anyway. Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
2015-04-22DRI2: Clear old->devPrivate.ptr in fixup_glamorMichel Dänzer
It doesn't point to the memory of the newly allocated BO. Fixes crash running piglit with Option "ShadowPrimary" enabled. Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
2015-04-17Add Option "TearFree" v4Michel Dänzer
Avoids tearing by flipping between two scanout BOs per (non-rotated) CRTC v2: * Fix condition for TearFree log message (Richard Wilbur) * Log warning message about DRI page flipping being enabled because of TearFree (or ShadowPrimary) also when building without glamor support v3: * Only override fb_id/x/y if all scanout pixmaps have been successfully allocated v4: * Make log warning clearer if drmModePageFlip returns an error Reviewed-by: Alex Deucher <alexander.deucher@amd.com> (v1)
2015-04-03glamor: Remove the stride member of struct radeon_pixmapMichel Dänzer
Its value was always the same as that of the PixmapRec devKind member. Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
2015-04-03glamor: Add Option "ShadowPrimary" v2Michel Dänzer
When this option is enabled, most pixmaps (including the screen pixmap) are allocated in system RAM and mostly accessed by the CPU. Changed areas of the screen pixmap are copied to dedicated per-CRTC scanout pixmaps regularly, triggered by the vblank interrupt. v2: * Set region data pointer to NULL for keeping only the extents * Move pRegion and pBox local variable declarations closer to their uses in drmmode_set_mode_major()
2015-04-03glamor: Add wrappers for the X server rendering hooksMichel Dänzer
They can choose between using the GPU or CPU for the operation.
2015-04-03glamor: Add radeon_pixmap parameter to radeon_glamor_create_textured_pixmapMichel Dänzer
2015-04-03glamor: Remove unused function radeon_glamor_pixmap_is_offscreenMichel Dänzer
2015-04-03Add RADEON_CREATE_PIXMAP_SCANOUT flagMichel Dänzer
It means that the pixmap is used for scanout exclusively.
2015-04-03Split out struct drmmode_scanout for rotation shadow buffer informationMichel Dänzer
Will be used for other kinds of dedicated scanout buffers as well.
2015-04-03Rename scanout_pixmap_x field to prime_pixmap_xMichel Dänzer
To avoid confusion with upcoming changes.
2015-04-03present: Don't flip between BOs with different tiling parametersMichel Dänzer
The kernel driver doesn't handle that correctly yet. Fixes or at least avoids issues with OpenGL fullscreen apps with DRI3 enabled and using PRIME or with (2D) tiling disabled. Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=89720 Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
2015-04-03Add radeon_get_pixmap_tiling_flags helperMichel Dänzer
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
2015-04-02Only enable SYNC extension fences and the Present extension along with DRI3Michel Dänzer
This avoids some trouble with the Gallium nine state tracker, which uses the Present extension even when DRI3 is disabled. Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=89772 Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
2015-03-30DRI2: Use radeon_get_pixmap_handleMichel Dänzer
Now we can share pixmaps with no struct radeon_bo via DRI2. Fixes VDPAU video playback freezing when using an OpenGL compositor with DRI3 enabled. Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=89755 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> Tested-by: Nick Sarnie <commendsarnex@gmail.com>
2015-03-30Move get_pixmap_handle helper to radeon_bo_helper.cMichel Dänzer
No functional change. Reviewed-by: Alex Deucher <alexander.deucher@amd.com> Tested-by: Nick Sarnie <commendsarnex@gmail.com>
2015-03-30Move radeon_drm_handler/abort_proc fields to drmmode_flipdata_recMichel Dänzer
Their values are the same for all DRM flip ioctl calls within a single radeon_do_pageflip() call. Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
2015-03-30Simplify radeon_do_pageflip() error handling slightly moreMichel Dänzer
We don't need the local variable old_fb_id. Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
2015-03-30Increase robustness against DRM page flip ioctl failures v3Michel Dänzer
Centralize cleanup, only clean up things that have been allocated for the failed ioctl call. Fixes double-free after a flip ioctl failure. Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=89681 v2: Only call drmModeRmFB for flipdata->old_fb_id on receipt of last DRM page flip event. Fixes Black screen on making glxgears fullscreen with DRI3 enabled. v3: Avoid double-free of flipdata in the unlikely case that calloc fails for flipcarrier, but only for the second or later CRTC. Reviewed-by: Alex Deucher <alexander.deucher@amd.com> (v2)
2015-03-30glamor: Handle GLAMOR_* flags removed from xserverMichel Dänzer
The behaviour is the same as when the removed flags were passed in. Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
2015-03-25glamor: Avoid generating GEM flink names for BOs shared via DRI3 v2Michel Dänzer
We can't create our own struct radeon_bo representation in this case because destroying that would make the GEM handle inaccessible to glamor as well. So just get the handle directly via dma-buf. v2: Close dma-buf file descriptor, pointed out by Axel Davy. Reviewed-by: Axel Davy <axel.davy@ens.fr> Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
2015-03-25Present: Add radeon_present_get_pixmap_handle helperMichel Dänzer
Reviewed-by: Axel Davy <axel.davy@ens.fr> Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
2015-03-25Make radeon_do_pageflip take a BO handle directlyMichel Dänzer
Reviewed-by: Axel Davy <axel.davy@ens.fr> Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
2015-03-24Adapt radeon_sync.c for misyncshm.h vs misync.h as wellMichel Dänzer
This should have been part of the previous commit. Without this, SYNC extension fences (and by extension DRI3) were disabled because it checked for HAVE_MISYNC_H, but configure now only defines HAVE_MISYNCSHM_H.
2015-03-19EXA: Return NULL from radeon_get_pixmap_bo if there is no driver privateMichel Dänzer
This mirrors what the glamor part of the function does. Fixes a crash running glxgears_pixmap with DRI3 enabled, reported by "marvin24" on IRC.
2015-03-18dri3: Use screen->CreatePixmap instead of fbCreatePixmap directlyMichel Dänzer
Fixes crash with EXA reported by "marvin24" on IRC.
2015-03-18glamor: Add glamor_fd_from_pixmap define for standalone glamor treeMichel Dänzer
2015-03-18Always include misync.h before other misync headersMichel Dänzer
Older versions of xserver didn't include misync.h from other misync headers as needed.
2015-03-18DRI3: Use open hook instead of open_clientMichel Dänzer
We don't need the client pointer, and this allows it to work with older versions of xserver as well.
generated by cgit v1.2.3 (git 2.25.1) at 2025-01-21 09:10:06 +0000