- 18 Oct, 2018 1 commit
-
-
Chunming Zhou authored
This patch is for VK_KHR_timeline_semaphore extension, semaphore is called syncobj in kernel side: This extension introduces a new type of syncobj that has an integer payload identifying a point in a timeline. Such timeline syncobjs support the following operations: * CPU query - A host operation that allows querying the payload of the timeline syncobj. * CPU wait - A host operation that allows a blocking wait for a timeline syncobj to reach a specified value. * Device wait - A device operation that allows waiting for a timeline syncobj to reach a specified value. * Device signal - A device operation that allows advancing the timeline syncobj to a specified value. v1: Since it's a timeline, that means the front time point(PT) always is signaled before the late PT. a. signal PT design: Signal PT fence N depends on PT[N-1] fence and signal opertion fence, when PT[N] fence is signaled, the timeline will increase to value of PT[N]. b. wait PT design: Wait PT fence is signaled by reaching timeline point value, when timeline is increasing, will compare wait PTs value with new timeline value, if PT value is lower than timeline value, then wait PT will be signaled, otherwise keep in list. syncobj wait operation can wait on any point of timeline, so need a RB tree to order them. And wait PT could ahead of signal PT, we need a sumission fence to perform that. v2: 1. remove unused DRM_SYNCOBJ_CREATE_TYPE_NORMAL. (Christian) 2. move unexposed denitions to .c file. (Daniel Vetter) 3. split up the change to drm_syncobj_find_fence() in a separate patch. (Christian) 4. split up the change to drm_syncobj_replace_fence() in a separate patch. 5. drop the submission_fence implementation and instead use wait_event() for that. (Christian) 6. WARN_ON(point != 0) for NORMAL type syncobj case. (Daniel Vetter) v3: 1. replace normal syncobj with timeline implemenation. (Vetter and Christian) a. normal syncobj signal op will create a signal PT to tail of signal pt list. b. normal syncobj wait op will create a wait pt with last signal point, and this wait PT is only signaled by related signal point PT. 2. many bug fix and clean up 3. stub fence moving is moved to other patch. v4： 1. fix RB tree loop with while(node=rb_first(...)). (Christian) 2. fix syncobj lifecycle. (Christian) 3. only enable_signaling when there is wait_pt. (Christian) 4. fix timeline path issues. 5. write a timeline test in libdrm v5: (Christian) 1. semaphore is called syncobj in kernel side. 2. don't need 'timeline' characters in some function name. 3. keep syncobj cb. v6: (Christian) 1. merge syncobj_timeline to syncobj structure. 2. simplify some check sentences. 3. some misc change. 4. fix CTS failed issue. v7: (Christian) 1. error handling when creating signal pt. 2. remove timeline naming in func. 3. export flags in find_fence. 4. allow reset timeline. v8: 1. use wait_event_interruptible without timeout 2. rename _TYPE_INDIVIDUAL to _TYPE_BINARY v9: 1. rename signal_pt->base to signal_pt->fence_array to avoid misleading 2. improve kerneldoc individual syncobj is tested by ./deqp-vk -n dEQP-VK*semaphore* timeline syncobj is tested by ./amdgpu_test -s 9 Signed-off-by:Chunming Zhou <david1.zhou@amd.com> Signed-off-by:
Christian König <christian.koenig@amd.com> Cc: Christian Konig <christian.koenig@amd.com> Cc: Dave Airlie <airlied@redhat.com> Cc: Daniel Rakos <Daniel.Rakos@amd.com> Cc: Daniel Vetter <daniel@ffwll.ch> Cc: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl> Cc: Jason Ekstrand <jason@jlekstrand.net> Reviewed-by:
Daniel Vetter <daniel.vetter@ffwll.ch> Link: https://patchwork.freedesktop.org/patch/257258/
-
- 17 Oct, 2018 3 commits
-
-
Deepak Rawat authored
With this patch split the kernel module specific code from actual selftest code. This is done to allow adding more selftests as separate file. Also added kernel module exit stub with this patch. Cc: Daniel Vetter <daniel.vetter@ffwll.ch> Cc: alexandru-cosmin.gheorghe@arm.com Signed-off-by:
Deepak Rawat <drawat@vmware.com> Reviewed-by:
Thomas Hellstrom <thellstrom@vmware.com> Signed-off-by:
-
Gustavo A. R. Silva authored
idx can be indirectly controlled by user-space, hence leading to a potential exploitation of the Spectre variant 1 vulnerability. This issue was detected with the help of Smatch: drivers/gpu/drm/drm_bufs.c:1420 drm_legacy_freebufs() warn: potential spectre issue 'dma->buflist' [r] (local cap) Fix this by sanitizing idx before using it to index dma->buflist Notice that given that speculation windows are large, the policy is to kill the speculation on the first load and not worry if it can be completed with a dependent load/store [1]. [1] https://marc.info/?l=linux-kernel&m=152449131114778&w=2Signed-off-by:
Gustavo A. R. Silva <gustavo@embeddedor.com> Signed-off-by:
-
Chunming Zhou authored
flags is new param for drm_syncob_find_fence, so need update doc, reproduce: make htmldocs reported: htmldocs: drivers/gpu/drm/drm_syncobj.c:230: warning: Function parameter or member 'flags' not described in 'drm_syncobj_find_fence' v2: rebased Signed-off-by:
-
- 16 Oct, 2018 3 commits
-
-
Enric Balletbo i Serra authored
'encoder' is dereferenced before it is null sanity checked, hence we potentially have a null pointer dereference bug. Instead, initialise drm_drv from encoder->dev->dev_private after we are sure 'encoder' is not null. Fixes: 5182c1a5 ("drm/rockchip: add an common abstracted PSR driver") Cc: stable@vger.kernel.org Signed-off-by:
Enric Balletbo i Serra <enric.balletbo@collabora.com> Signed-off-by:
Heiko Stuebner <heiko@sntech.de> Link: https://patchwork.freedesktop.org/patch/msgid/20181013105654.11827-1-enric.balletbo@collabora.com
-
Chunming Zhou authored
flags can be used by driver to decide whether need to block wait submission. Signed-off-by:
-
Mamta Shukla authored
Shift * to be adjacent to pointer name to follow Linux coding style. Issue found with checkpatch.pl ERROR: "foo * bar" should be "foo *bar" Signed-off-by:
Mamta Shukla <mamtashukla555@gmail.com> Signed-off-by:
-
- 15 Oct, 2018 4 commits
-
-
Eric Anholt authored
Fixes an oops reading this debugfs entry on BCM7278. Signed-off-by:
Eric Anholt <eric@anholt.net> Link: https://patchwork.freedesktop.org/patch/msgid/20180928232126.4332-4-eric@anholt.net Fixes: 57692c94 ("drm/v3d: Introduce a new DRM driver for Broadcom V3D V3.x+") Cc: <stable@vger.kernel.org> Reviewed-by:
Boris Brezillon <boris.brezillon@bootlin.com>
-
Eric Anholt authored
Since this is UAPI, it's good to document what exactly the guarantees we're providing are. Signed-off-by:
-
Eric Anholt authored
This adds just enough performance counter support to measure the clock. We don't have linux kernel drivers for the clock driving the HW, and this was useful for determining that the V3D HW is running on a slow clock, not that the driver was slow. Signed-off-by:
-
Eric Anholt authored
Once we push the job, the scheduler could run it and free it. So, if we want to reference their fences, we need to grab them before then. I haven't seen this happen in many days of conformance test runtime, but let's still close the race. Signed-off-by:
-
- 11 Oct, 2018 2 commits
-
-
José Roberto de Souza authored
drm_dp_cec_register_connector() is called when registering each DP connector in DRM, while sounds a good idea register CEC adapters as earlier as possible, it causes some driver initialization delay trying to do DPCD transactions in disconnected connectors. This change will cause no regressions as drm_dp_cec_set_edid() will still be called in further detection of connected connectors with a valid edid parameter. This change reduced the module load of i915 by average 0.5sec in a machine with just one DP port disconnected while reducing more than 3sec in a machine with 4 DP ports disconnected. Cc: Hans Verkuil <hans.verkuil@cisco.com> Signed-off-by:
José Roberto de Souza <jose.souza@intel.com> Acked-by:
Hans Verkuil <hans.verkuil@cisco.com> Signed-off-by:
-
YueHaibing authored
Use PTR_ERR_OR_ZERO rather than if(IS_ERR(...)) + PTR_ERR Signed-off-by:
YueHaibing <yuehaibing@huawei.com> Signed-off-by:
-
- 08 Oct, 2018 2 commits
-
-
Giulio Benetti authored
If tcon->panel pointer is NULL, trying to dereference from it (i.e. tcon->panel->connector) will cause a null pointer dereference. Add tcon->panel null pointer check before calling sun4i_tcon0_mode_set_dithering(). Signed-off-by:
Giulio Benetti <giulio.benetti@micronovasrl.com> Fixes: f11adcec ("drm/sun4i: tcon: Add dithering support for RGB565/RGB666 LCD panels") Reviewed-by:
Chen-Yu Tsai <wens@csie.org> Signed-off-by:
Maxime Ripard <maxime.ripard@bootlin.com> Link: https://patchwork.freedesktop.org/patch/msgid/20181005215951.99003-2-giulio.benetti@micronovasrl.com
-
Giulio Benetti authored
Since tcon->panel is a pointer returned by of_drm_find_panel() need to check if it is not NULL, hence a valid pointer. IS_ERR() instead checks return error values, not NULL pointers. Substitute "if (!IS_ERR(tcon->panel))" with "if (tcon->panel)". Signed-off-by:
-
- 05 Oct, 2018 20 commits
-
-
Daniel Vetter authored
It's for legacy drivers only (atomic ones should use drm_atomic_helper_check_plane_state() instead), and there's no users left except the one in the primary plane helpers. Signed-off-by:
Daniel Vetter <daniel.vetter@intel.com> Reviewed-by:
Ville Syrjälä <ville.syrjala@linux.intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20181004202446.22905-18-daniel.vetter@ffwll.ch
-
Thomas Hellstrom authored
Use the correct helper and also return early on helper success rather than on helper failure. Also explicitly return 0 in the case of no fb. v2: Check for !fb after updating state->visible (Ville). Reviewed-by:
Dan Carpenter <dan.carpenter@oracle.com> Reported-by:
Daniel Vetter <daniel@ffwll.ch> Cc: Ville Syrjälä <ville.syrjala@linux.intel.com> Signed-off-by:
-
Daniel Vetter authored
Well except the destroy helper, which isn't really a primary helper but generally useful, if mislabelled. v2: Keep some of the nice comments about the limitations of the primarmy plane helpers, and put them into the kerneldoc for drm_crtc_init() (Sam). Reviewed-by:
-
Daniel Vetter authored
Motivated by review comments from Ville&Sean. Cc: Ville Syrjälä <ville.syrjala@linux.intel.com> Cc: Sean Paul <sean@poorly.run> Reviewed-by:
-
Daniel Vetter authored
Didn't get updated in a rework of the original patch. Fixes: 059b5eb5 ("drm: move native byte order quirk to new drm_driver_legacy_fb_format function") Cc: Gerd Hoffmann <kraxel@redhat.com> Reviewed-by:
Gerd Hoffmann <kraxel@redhat.com> Signed-off-by:
-
Daniel Vetter authored
With armada the last bigger driver that realistically needed these to convert from legacy kms to atomic is converted. These helpers have been broken more often than not the past 2 years, and as this little patch series shows, tricked a bunch of people into using the wrong helpers for their functions. Aside: I think a lot more drivers should be using the device-level drm_atomic_helper_shutdown/suspend/resume helpers and related functions. In almost all the cases they get things exactly right. Reviewed-by:
-
Daniel Vetter authored
drm_plane_helper_disable is a non-atomic drivers only function, and will blow up (since no one passes the locking context it needs). Atomic drivers which want to quiescent their hw on unload should use drm_atomic_helper_shutdown() instead. Reviewed-by:
Shawn Guo <shawnguo@kernel.org> Link: https://patchwork.freedesktop.org/patch/msgid/20181004202446.22905-15-daniel.vetter@ffwll.ch
-
Daniel Vetter authored
drm_plane_helper_disable is a non-atomic drivers only function, and will blow up (since no one passes the locking context it needs). Atomic drivers which want to quiescent their hw on unload should use drm_atomic_helper_shutdown() instead. v2: Rebase. Reviewed-by:
-
Daniel Vetter authored
drm_plane_helper_disable is a non-atomic drivers only function, and will blow up (since no one passes the locking context it needs). Atomic drivers which want to quiescent their hw on unload should use drm_atomic_helper_shutdown() instead. The sti cleanup code seems supremely confused: - In the load error path it calls drm_mode_config_cleanup before it stops various kms services like poll worker or fbdev emulation. That's going to oops. - The actual unload code doesn't even bother with the cleanup and just leaks. Try to fix this while at it. Reviewed-by:
-
Daniel Vetter authored
drm_plane_helper_disable is a non-atomic drivers only function, and will blow up (since no one passes the locking context it needs). Atomic drivers which want to quiescent their hw on unload should use drm_atomic_helper_shutdown() instead. Reviewed-by:
-
Daniel Vetter authored
drm_plane_helper_disable is a non-atomic drivers only function, and will blow up (since no one passes the locking context it needs). Atomic drivers which want to quiescent their hw on unload should use drm_atomic_helper_shutdown() instead. Reviewed-by:
-
Daniel Vetter authored
These do absolutely nothing for atomic drivers. Reviewed-by:
-
Daniel Vetter authored
These do absolutely nothing for atomic drivers. Reviewed-by:
-
Daniel Vetter authored
Motivated by vmwgfx digging around in core uapi bits it shouldn't dig around in. Reviewed-by:
-
Daniel Vetter authored
We already have a separate overview doc for this, makes sense to untangle it from the overall atomic helpers. v2: Rebase v3: Rebase more. Acked-by:
-
Daniel Vetter authored
It's the default. The exported version was kinda a transition state, before we made this the default. To stop new atomic drivers from using it (instead of just relying on the default) let's unexport it. v2: rename the default implementation to a more fitting name and add a comment (Laurent) Signed-off-by:
Laurent Pinchart <laurent.pinchart@ideasonboard.com> Link: https://patchwork.freedesktop.org/patch/msgid/20181004202446.22905-3-daniel.vetter@ffwll.ch
-
Daniel Vetter authored
For atomic driver this is the default, no need to reimplement it. We still need to keep the copypasta for not-atomic drivers though, since no one polished the legacy crtc helpers as much as the atomic ones. v2: amdgpu uses ->best_encoder internally, give it a local copy. It might be a good idea to merge the connector and encoder into one amdgpu_dm_sink structure, that might match DC internals better. At least for non-DPMST outputs. Kudos to Ville for spotting this. v3: Rebase onto a487411a ("drm/amd/display: Use DRM helper for best_encoder"). Cc: Ville Syrjälä <ville.syrjala@linux.intel.com> Signed-off-by:
Harry Wentland <harry.wentland@amd.com> Link: https://patchwork.freedesktop.org/patch/msgid/20181004202446.22905-2-daniel.vetter@ffwll.ch
-
Chris Wilson authored
Since commit 09ea0dfb ("dma-buf: make map_atomic and map function pointers optional"), we no longer need to provide stub no-op functions as the core now provides them directly. References: 09ea0dfb ("dma-buf: make map_atomic and map function pointers optional") Signed-off-by:
Chris Wilson <chris@chris-wilson.co.uk> Cc: Daniel Vetter <daniel@ffwll.ch> Cc: Gerd Hoffmann <kraxel@redhat.com> Cc: Alex Deucher <alexander.deucher@amd.com> Cc: "Christian König" <christian.koenig@amd.com> Reviewed-by:
-
Daniel Vetter authored
Somehow I forgot a few when typing all the shiny new kerneldoc. Drop them. Reviewed-by:
-
Gerd Hoffmann authored
drm_connector_init doesn't attach the edid property for some connector types, drm_connector_attach_edid_property() can be used to enable the edid property in these cases. Signed-off-by:
-
- 03 Oct, 2018 3 commits
-
-
Manasi Navare authored
This patch explains the DRM_MODE_PROP_IMMUTABLE flag a bit better by telling which function to call if kernel wants to update drm object's immutable properties. Suggested-by:
Manasi Navare <manasi.d.navare@intel.com> Reviewed-by:
-
Daniel Vetter authored
Typed up all the items Rodrigo capture at the vkms workshop at xdc2018: https://etherpad.net/p/vkms_todo v2: Review from Rodrigo. - Add eBPF for atomic check, I forgot it. - Improve spelling. Signed-off-by:
Rodrigo Siqueira <rodrigosiqueiramelo@gmail.com> Link: https://patchwork.freedesktop.org/patch/msgid/20181003132103.29309-1-daniel.vetter@ffwll.ch
-
Ville Syrjälä authored
Let's make the infoframe pack functions usable with a const infoframe structure. This allows us to precompute the infoframe earlier, and still pack it later when we're no longer allowed to modify the structure. So now we end up with a _check()+_pack_only() or _pack() functions depending on whether you want to precompute the infoframes or not. The names aren't great but I was lazy and didn't want to change all the drivers. v2: Deal with exynos churn Actually export the new funcs v3: Fix various documentation fails (Hans) Cc: Thierry Reding <thierry.reding@gmail.com> Cc: Hans Verkuil <hans.verkuil@cisco.com> Cc: linux-media@vger.kernel.org Signed-off-by:
-
- 01 Oct, 2018 2 commits
-
-
Satendra Singh Thakur authored
Use the drm_display_mode_to_videomode function to calculate front/ back porches and sync length. Cc: Madhur Verma <madhur.verma@samsung.com> Cc: Hemanshu Srivastava <hemanshu.s@samsung.com> Signed-off-by:
Satendra Singh Thakur <satendra.t@samsung.com> Signed-off-by:
Stefan Agner <stefan@agner.ch> Acked-by:
-
Stefan Agner authored
The result of drm_crtc_index() is unused. Remove the call and the unused variable. Signed-off-by:
-
