- 03 Feb, 2017 3 commits
-
-
Chris Wilson authored
Check that if we request bottom-up allocation from drm_mm_insert_node() we receive the next available hole from the bottom. Signed-off-by:
Chris Wilson <chris@chris-wilson.co.uk> Reviewed-by:
Joonas Lahtinen <joonas.lahtinen@linux.intel.com> Signed-off-by:
Daniel Vetter <daniel.vetter@ffwll.ch> Link: http://patchwork.freedesktop.org/patch/msgid/20170202114434.3060-2-chris@chris-wilson.co.uk
-
Chris Wilson authored
The drm_mm range manager claimed to support top-down insertion, but it was neither searching for the top-most hole that could fit the allocation request nor fitting the request to the hole correctly. In order to search the range efficiently, we create a secondary index for the holes using either their size or their address. This index allows us to find the smallest hole or the hole at the bottom or top of the range efficiently, whilst keeping the hole stack to rapidly service evictions. v2: Search for holes both high and low. Rename flags to mode. v3: Discover rb_entry_safe() and use it! v4: Kerneldoc for enum drm_mm_insert_mode. Signed-off-by:
Alex Deucher <alexander.deucher@amd.com> Reviewed-by: Sinclair Yeh <syeh@vmware.com> # vmwgfx Reviewed-by: Lucas Stach <l.stach@pengutronix.de> #etnaviv Signed-off-by:
-
Daniel Vetter authored
Yet another backmerge to get at latest etnaviv code, which is need for Chris' drm_mm patch. Signed-off-by:Daniel Vetter <daniel.vetter@intel.com>
-
- 02 Feb, 2017 37 commits
-
-
https://git.pengutronix.de/git/lst/linuxDave Airlie authored
It includes code cleanups from Bhumika and Liviu, a significant shader performance fix and additions to the cmdstream validator from Wladimir and the addition of a cmdbuf suballocator by myself. The suballocator improves performance on all chips by reducing the CPU overhead of the kernel driver and side steps the GC3000 FE MMU flush erratum, now making the workarounds in IOVA allocation we had before unnecessary, which results in a nice cleanup of the code in that area. * 'drm-etnaviv-next' of https://git.pengutronix.de/git/lst/linux: drm/etnaviv: Remove duplicate header file include Revert "drm/etnaviv: trick drm_mm into giving out a low IOVA" drm/etnaviv: add cmdbuf suballocator drm/etnaviv: get cmdbuf physical address through the cmdbuf abstraction drm/etnaviv: wire up iova handling in new cmdbuf abstraction drm/etnaviv: move cmdbuf de-/allocation into own file drm/etnaviv: always flush MMU TLBs on map/unmap drm/etnaviv: constify etnaviv_iommu_ops structures drm/etnaviv: set up initial PULSE_EATER register drm/etnaviv: add new GC3000 sensitive states
-
Gabriel Krisman Bertazi authored
Commit be7f735cd5ea ("drm: Rely on mode_config data for fb_helper initialization") broke the build when CONFIG_DRM_FBDEV_EMULATION is disabled because it didn't update the prototype for drm_fb_helper_init in that case. Fixes: be7f735cd5ea ("drm: Rely on mode_config data for fb_helper initialization") Signed-off-by:Gabriel Krisman Bertazi <krisman@collabora.co.uk> Signed-off-by:
-
Gabriel Krisman Bertazi authored
Instead of receiving the num_crts as a parameter, we can read it directly from the mode_config structure. I audited the drivers that invoke this helper and I believe all of them initialize the mode_config struct accordingly, prior to calling the fb_helper. I used the following coccinelle hack to make this transformation, except for the function headers and comment updates. The first and second rules are split because I couldn't find a way to remove the unused temporary variables at the same time I removed the parameter. // <smpl> @r@ expression A,B,D,E; identifier C; @@ ( - drm_fb_helper_init(A,B,C,D) + drm_fb_helper_init(A,B,D) | - drm_fbdev_cma_init_with_funcs(A,B,C,D,E) + drm_fbdev_cma_init_with_funcs(A,B,D,E) | - drm_fbdev_cma_init(A,B,C,D) + drm_fbdev_cma_init(A,B,D) ) @@ expression A,B,C,D,E; @@ ( - drm_fb_helper_init(A,B,C,D) + drm_fb_helper_init(A,B,D) | - drm_fbdev_cma_init_with_funcs(A,B,C,D,E) + drm_fbdev_cma_init_with_funcs(A,B,D,E) | - drm_fbdev_cma_init(A,B,C,D) + drm_fbdev_cma_init(A,B,D) ) @@ identifier r.C; type T; expression V; @@ - T C; <... when != C - C = V; ...> // </smpl> Changes since v1: - Rebased on top of the tip of drm-misc-next. - Remove mention to sti since a proper fix got merged. Suggested-by:
Eric Anholt <eric@anholt.net> Signed-off-by:
-
Chris Wilson authored
Some state is coupled into the device lifetime outside of the load/unload timeframe and requires teardown during final unreference from drm_dev_release(). For example, dmabufs hold both a device and module reference and may live longer than expected (i.e. the current pattern of the driver tearing down its state and then releasing a reference to the drm device) and yet touch driver private state when destroyed. v2: Export drm_dev_fini() and move the responsibility for finalizing the drm_device and freeing it to the release callback. (If no callback is provided, the core will call drm_dev_fini() and kfree(dev) as before.) v3: Remember to add drm_dev_fini() to drm_drv.h v4: Tidy language for kerneldoc v5: Cross reference from drm_dev_init() to note that driver->release() allows for arbitrary embedding. v6: Refer to driver data rather than driver state, as state is now becoming associated with the struct drm_atomic_state and friends. Signed-off-by:
Laurent Pinchart <laurent.pinchart@ideasonboard.com> [danvet: Use the proper reference for struct members, which is &drm_driver.release.] Signed-off-by:
-
Neil Armstrong authored
The platform driver name is currently "meson" which can lead to some confusion, this patch renames it to "meson-drm" and removes the owner attribute since is always added by __platform_driver_register called by the module_platform_driver() macro. Signed-off-by:
Neil Armstrong <narmstrong@baylibre.com> Acked-by:
Daniel Vetter <daniel@ffwll.ch> Link: http://patchwork.freedesktop.org/patch/msgid/1486028864-19622-3-git-send-email-narmstrong@baylibre.com
-
Neil Armstrong authored
The module is currently named "meson.ko" which can lead to some confusion, this patches renames it "meson-drm.ko" Signed-off-by:
-
Andrzej Hajda authored
Bug in DECON(CRTC) driver prevented interlace modes from proper work. Since DECON is fixed interlace modes can be enabled in MHL. Signed-off-by:
Andrzej Hajda <a.hajda@samsung.com> Signed-off-by:
Archit Taneja <architt@codeaurora.org> Link: http://patchwork.freedesktop.org/patch/msgid/1485935272-17337-26-git-send-email-a.hajda@samsung.com
-
Andrzej Hajda authored
Since all sub-protocols of MHL3 are already supported MHL3 mode can be enabled. With this patch it is possible to use packed pixel modes and clocks up to 300MHz - 1920x1080@60Hz and 4K modes. Signed-off-by:
-
Andrzej Hajda authored
In case of MHL3 HSIC should be initialized. Signed-off-by:
-
Andrzej Hajda authored
The patch adds code to report back feature complete IRQ, and code to read and drop burst writes from peer. Signed-off-by:
-
Andrzej Hajda authored
Device should report to the peer which features are really supported. Signed-off-by:
-
Andrzej Hajda authored
MHL3 protocol requires registry adjustments depending on chosen video mode. Necessary information is gathered in mode_fixup callback. In case of HDMI video modes driver should also send special AVI and MHL3 infoframes. The patch introduces generic helpers for handling MHL3 infoframes, in case of appearance of other users of MHL3 infoframes these function can be moved to common library. Signed-off-by:
-
Andrzej Hajda authored
MHL3 protocol uses vendor specific infoframes to transmit additional information to the sink. This patch adds definitions of structures and constants used to create such frames. Signed-off-by:
-
Andrzej Hajda authored
The patch allows to avoid rare cases when discovery fails. Signed-off-by:
-
Andrzej Hajda authored
MHL3 requires that after reading EDID from the sink source should ask peer for features. To make both protocols happy the patch splits the code accordingly. Signed-off-by:
-
Andrzej Hajda authored
Without delay CBUS sometimes was not reset properly. Signed-off-by:
-
Andrzej Hajda authored
Stopping output causes full re-detection of the sink and slows down the process. Signed-off-by:
-
Andrzej Hajda authored
Write burst should be enabled for MHL_INT_RC_FEAT_REQ and disabled for other commands. The patch moves functions up and adds delay setting for MHL3 burst mode. Signed-off-by:
-
Andrzej Hajda authored
MHL1 and MHL3 have different initialization paths. To make both protocols happy sink detection is put into continuation after link mode enablement. Signed-off-by:
-
Andrzej Hajda authored
MHL2 receiver require disabling transmitter on initialization. Signed-off-by:
-
Andrzej Hajda authored
Messages queue can be cleaned up by message callbacks, to avoid repeated removal of current message it should be removed from the queue before calling these callbacks. Signed-off-by:
-
Andrzej Hajda authored
In case of MHL3 CBUS is bring-up already in sii8620_got_ecbus_speed. Signed-off-by:
-
Andrzej Hajda authored
Peer capabilities should be read differently depending on protocol version. Signed-off-by:
-
Andrzej Hajda authored
MHL3 protocol requires device to respond to feature request from peer. This minimal answer fulfills the requirement and allows to continue negotiation. Signed-off-by:
-
Andrzej Hajda authored
Burst transmissions are used in MHL3 mode negotiation. Signed-off-by:
-
Andrzej Hajda authored
Burst transmissions are required in MHL3 modes. Signed-off-by:
-
Andrzej Hajda authored
The Single-ended eCBUS Mode (eCBUS-S) mode provides 60 Mb/s full-duplex bidirectional traffic for three channels: - CBUS data (CBUS1 channel), - High-bandwidth MHL data (eMSC channel), - tunneling data (T-CBUS channel). It is required to fully support MHL3 dongles. Signed-off-by:
-
Andrzej Hajda authored
Due to asynchronous nature of MHL flow of execution is dispersed. Logical continuation of some actions happens after response of peer, i.e in interrupt handler. To simplify coding continuation mechanism has been added - it is now possible to provide continuation callback, which will be called after peer responds to given action. Signed-off-by:
-
Andrzej Hajda authored
This functionality is necessary to implement MHL3 modes. Signed-off-by:
-
Andrzej Hajda authored
MHL3 modes differs significantly from MHL1 mode, this helper will be used frequently to clearly distinguish them. Signed-off-by:
-
Andrzej Hajda authored
It is not necessary to set REG_COC_CTL0, REG_MHL_COC_CTL1 registers. Signed-off-by:
-
Liviu Dudau authored
etnaviv_gem.h header gets included twice. Remove duplicate. Signed-off-by:
Liviu Dudau <liviu@dudau.co.uk> Reviewed-by:
Christian Gmeiner <christian.gmeiner@gmail.com> Signed-off-by:
Lucas Stach <l.stach@pengutronix.de>
-
Lucas Stach authored
Now that commandstreams are handled through the cmdbuf suballocator the workaround to make the IOVA games work is not needed anymore. Signed-off-by:
-
Lucas Stach authored
There are 3 big benefits to suballocating a single big DMA buffer for command submission: 1. Avoid hammering CMA. The old way of allocating and freeing a DMA buffer for each submission was hitting some of the real slow pathes in CMA, as this allocator was not designed for a concurrent small buffers load. 2. Less TLB flushes on IOMMUv2. If a new command buffer is mapped into the GPU address space the MMU TLBs need to be flushed. By having one big buffer statically mapped to the GPU, a lot of those flushes can be avoided. 3. No funky workarounds for GC3000. The FE TLB flush on GC3000 isn't reliable. To work around that we tried to lay out the cmdbufs in the GPU address space in a way to avoid this issue. This hasn't always worked if the address space is crowded. A single statically mapped buffer avoids the erratum completely. Signed-off-by:
-
Lucas Stach authored
Don't allow IOMMUv2 to peek directly into the cmdbuf, but get the needed PA through a dedicated function. Signed-off-by:
-
Lucas Stach authored
Don't call the IOMMU directly, but go through the new cmdbuf abstraction. Signed-off-by:
-
Lucas Stach authored
This will get more complex with the following changes, so move it into its own place. Signed-off-by:
-
