Errata i878 says that MPU should not be used to access RAM and DMM at
the same time. As it's not possible to prevent MPU accessing RAM, we
need to access DMM via a proxy.
This patch changes DMM driver to access DMM registers via sDMA. Instead
of doing a normal readl/writel call to read/write a register, we use
sDMA to copy 4 bytes from/to the DMM registers.
This patch provides only a partial workaround for i878, as not only DMM
register reads/writes are affected, but also accesses to the DMM mapped
buffers (framebuffers, usually).
Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ti.com>
Signed-off-by: Peter Ujfalusi <peter.ujfalusi@ti.com>
The interrupts should be enabled after the driver initialization to avoid
early interrupts while the driver is not yet ready to handle them.
On removal the interrupts must be disabled before other resources are
released, freed up.
Signed-off-by: Peter Ujfalusi <peter.ujfalusi@ti.com>
Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ti.com>
The driver probe would fail if the irq is not available.
Signed-off-by: Peter Ujfalusi <peter.ujfalusi@ti.com>
Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ti.com>
A DMM timeout "timed out waiting for done" has been observed on DRA7
devices. The timeout happens rarely, and only when the system is under
heavy load.
Debugging showed that the timeout can be made to happen much more
frequently by optimizing the DMM driver, so that there's almost no code
between writing the last DMM descriptors to RAM, and writing to DMM
register which starts the DMM transaction.
The current theory is that a wmb() does not properly ensure that the
data written to RAM is observable by all the components in the system.
This DMM timeout has caused interesting (and rare) bugs as the error
handling was not functioning properly (the error handling has been fixed
in previous commits):
* If a DMM timeout happened when a GEM buffer was being pinned for
display on the screen, a timeout error would be shown, but the driver
would continue programming DSS HW with broken buffer, leading to
SYNCLOST floods and possible crashes.
* If a DMM timeout happened when other user (say, video decoder) was
pinning a GEM buffer, a timeout would be shown but if the user
handled the error properly, no other issues followed.
* If a DMM timeout happened when a GEM buffer was being released, the
driver does not even notice the error, leading to crashes or hang
later.
This patch adds wmb() and readl() calls after the last bit is written to
RAM, which should ensure that the execution proceeds only after the data
is actually in RAM, and thus observable by DMM.
The read-back should not be needed. Further study is required to understand
if DMM is somehow special case and read-back is ok, or if DRA7's memory
barriers do not work correctly.
Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ti.com>
Signed-off-by: Peter Ujfalusi <peter.ujfalusi@ti.com>
tiler_reserve_2d allocates memory but does not check if it got the
memory. Add the check and return ENOMEM on failure.
Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ti.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20180329104038.29154-2-tomi.valkeinen@ti.com
Reviewed-by: Emil Velikov <emil.velikov@collabora.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Sean Paul <seanpaul@chromium.org>
The standard kernel integer types are [us]{8,16,32}. Use them instead of
the u?int{8,16,32}_t types.
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Sebastian Reichel <sebastian.reichel@collabora.co.uk>
Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ti.com>
Having the filename in the header serves little purpose and is
often wrong after renames as it is here in several places, just
drop it from all omapdrm files.
While we are here unify the copyright tags to the TI recommended style.
Signed-off-by: Andrew F. Davis <afd@ti.com>
Check the status of the DMM engine after it is reported that the
transaction was completed as in rare cases the engine might not reached a
working state.
The wait_status() will print information in case the DMM is not reached the
expected state and the dmm_txn_commit() will return with an error code to
make sure that we are not continuing with a broken setup.
Signed-off-by: Peter Ujfalusi <peter.ujfalusi@ti.com>
Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ti.com>
To help diagnose DMM errors, print out information if any of the error bits
are set in the interrupt status register.
Signed-off-by: Peter Ujfalusi <peter.ujfalusi@ti.com>
Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ti.com>
If the wait_status() fails either because of an error reported in the
STATUS register or because of a timeout waiting for the wait_mask, print
information which might help diagnose the reason.
Signed-off-by: Peter Ujfalusi <peter.ujfalusi@ti.com>
Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ti.com>
If we don't find a matching device node, we must free the memory allocated
in 'omap_dmm' a few lines above.
Fixes: 7cb0d6c17b ("drm/omap: fix TILER on OMAP5")
Signed-off-by: Christophe JAILLET <christophe.jaillet@wanadoo.fr>
Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ti.com>
Fix a few type issues that cause compile warnings on 64 bit ARM
compiler. The change should not affect 32bit platforms.
Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ti.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
By linking the sizeof to a variable type the code will be less prone to
bugs due to future type changes of variables.
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Reviewed-by: Tomi Valkeinen <tomi.valkeinen@ti.com>
The omapdrm driver relies on this header to be included
implicitly, but this does not always work, and I get
this error in randconfig builds:
gpu/drm/omapdrm/dss/hdmi_phy.c: In function 'hdmi_phy_dump':
gpu/drm/omapdrm/dss/hdmi_phy.c:34:2: error: implicit declaration of function 'seq_printf' [-Werror=implicit-function-declaration]
gpu/drm/omapdrm/dss/hdmi_wp.c: In function 'hdmi_wp_dump':
gpu/drm/omapdrm/dss/hdmi_wp.c:26:2: error: implicit declaration of function 'seq_printf' [-Werror=implicit-function-declaration]
gpu/drm/omapdrm/dss/hdmi_pll.c: In function 'hdmi_pll_dump':
gpu/drm/omapdrm/dss/hdmi_pll.c:30:2: error: implicit declaration of function 'seq_printf' [-Werror=implicit-function-declaration]
This adds the #include statements in all files that have
a seq_printf statement.
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ti.com>
Pull drm updates from Dave Airlie:
"This is the main drm pull request for 4.6 kernel.
Overall the coolest thing here for me is the nouveau maxwell signed
firmware support from NVidia, it's taken a long while to extract this
from them.
I also wish the ARM vendors just designed one set of display IP, ARM
display block proliferation is definitely increasing.
Core:
- drm_event cleanups
- Internal API cleanup making mode_fixup optional.
- Apple GMUX vga switcheroo support.
- DP AUX testing interface
Panel:
- Refactoring of DSI core for use over more transports.
New driver:
- ARM hdlcd driver
i915:
- FBC/PSR (framebuffer compression, panel self refresh) enabled by default.
- Ongoing atomic display support work
- Ongoing runtime PM work
- Pixel clock limit checks
- VBT DSI description support
- GEM fixes
- GuC firmware scheduler enhancements
amdkfd:
- Deferred probing fixes to avoid make file or link ordering.
amdgpu/radeon:
- ACP support for i2s audio support.
- Command Submission/GPU scheduler/GPUVM optimisations
- Initial GPU reset support for amdgpu
vmwgfx:
- Support for DX10 gen mipmaps
- Pageflipping and other fixes.
exynos:
- Exynos5420 SoC support for FIMD
- Exynos5422 SoC support for MIPI-DSI
nouveau:
- GM20x secure boot support - adds acceleration for Maxwell GPUs.
- GM200 support
- GM20B clock driver support
- Power sensors work
etnaviv:
- Correctness fixes for GPU cache flushing
- Better support for i.MX6 systems.
imx-drm:
- VBlank IRQ support
- Fence support
- OF endpoint support
msm:
- HDMI support for 8996 (snapdragon 820)
- Adreno 430 support
- Timestamp queries support
virtio-gpu:
- Fixes for Android support.
rockchip:
- Add support for Innosilicion HDMI
rcar-du:
- Support for 4 crtcs
- R8A7795 support
- RCar Gen 3 support
omapdrm:
- HDMI interlace output support
- dma-buf import support
- Refactoring to remove a lot of legacy code.
tilcdc:
- Rewrite of pageflipping code
- dma-buf support
- pinctrl support
vc4:
- HDMI modesetting bug fixes
- Significant 3D performance improvement.
fsl-dcu (FreeScale):
- Lots of fixes
tegra:
- Two small fixes
sti:
- Atomic support for planes
- Improved HDMI support"
* 'drm-next' of git://people.freedesktop.org/~airlied/linux: (1063 commits)
drm/amdgpu: release_pages requires linux/pagemap.h
drm/sti: restore mode_fixup callback
drm/amdgpu/gfx7: add MTYPE definition
drm/amdgpu: removing BO_VAs shouldn't be interruptible
drm/amd/powerplay: show uvd/vce power gate enablement for tonga.
drm/amd/powerplay: show uvd/vce power gate info for fiji
drm/amdgpu: use sched fence if possible
drm/amdgpu: move ib.fence to job.fence
drm/amdgpu: give a fence param to ib_free
drm/amdgpu: include the right version of gmc header files for iceland
drm/radeon: fix indentation.
drm/amd/powerplay: add uvd/vce dpm enabling flag to fix the performance issue for CZ
drm/amdgpu: switch back to 32bit hw fences v2
drm/amdgpu: remove amdgpu_fence_is_signaled
drm/amdgpu: drop the extra fence range check v2
drm/amdgpu: signal fences directly in amdgpu_fence_process
drm/amdgpu: cleanup amdgpu_fence_wait_empty v2
drm/amdgpu: keep all fences in an RCU protected array v2
drm/amdgpu: add number of hardware submissions to amdgpu_fence_driver_init_ring
drm/amdgpu: RCU protected amd_sched_fence_release
...
Rename dma_*_writecombine() to dma_*_wc(), so that the naming
is coherent across the various write-combining APIs. Keep the
old names for compatibility for a while, these can be removed
at a later time. A guard is left to enable backporting of the
rename, and later remove of the old mapping defines seemlessly.
Build tested successfully with allmodconfig.
The following Coccinelle SmPL patch was used for this simple
transformation:
@ rename_dma_alloc_writecombine @
expression dev, size, dma_addr, gfp;
@@
-dma_alloc_writecombine(dev, size, dma_addr, gfp)
+dma_alloc_wc(dev, size, dma_addr, gfp)
@ rename_dma_free_writecombine @
expression dev, size, cpu_addr, dma_addr;
@@
-dma_free_writecombine(dev, size, cpu_addr, dma_addr)
+dma_free_wc(dev, size, cpu_addr, dma_addr)
@ rename_dma_mmap_writecombine @
expression dev, vma, cpu_addr, dma_addr, size;
@@
-dma_mmap_writecombine(dev, vma, cpu_addr, dma_addr, size)
+dma_mmap_wc(dev, vma, cpu_addr, dma_addr, size)
We also keep the old names as compatibility helpers, and
guard against their definition to make backporting easier.
Generated-by: Coccinelle SmPL
Suggested-by: Ingo Molnar <mingo@kernel.org>
Signed-off-by: Luis R. Rodriguez <mcgrof@suse.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: airlied@linux.ie
Cc: akpm@linux-foundation.org
Cc: benh@kernel.crashing.org
Cc: bhelgaas@google.com
Cc: bp@suse.de
Cc: dan.j.williams@intel.com
Cc: daniel.vetter@ffwll.ch
Cc: dhowells@redhat.com
Cc: julia.lawall@lip6.fr
Cc: konrad.wilk@oracle.com
Cc: linux-fbdev@vger.kernel.org
Cc: linux-pci@vger.kernel.org
Cc: luto@amacapital.net
Cc: mst@redhat.com
Cc: tomi.valkeinen@ti.com
Cc: toshi.kani@hp.com
Cc: vinod.koul@intel.com
Cc: xen-devel@lists.xensource.com
Link: http://lkml.kernel.org/r/1453516462-4844-1-git-send-email-mcgrof@do-not-panic.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
This patch adds wrapper functions for readl() and writel(), dmm_read()
and dmm_write(), so that we can implement workaround for DRA7 errata
i878.
Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ti.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
The current driver uses non-blocking DMM fill when releasing memory.
This gives us a small performance increase as we don't have to wait for
the fill operation to finish.
However, the driver does not have any error handling for non-blocking
fill. In case of an error, the fill operation may silently fail, leading
to leaking DMM engines, which may eventually lead to deadlock if we run
out of DMM engines.
This patch makes the DMM driver always use blocking fills, so that we
can catch the errors. A more complex option would be to allow
non-blocking fills, and implement proper error handling, but that is
left for the future.
This patch is a HACK, as the proper fix is to either decide to always
use sync fills and remove all the async related code, or fix the async
code.
Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ti.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Modified Tiler placement to utilize bitmaps for bookkeeping and
all placement algorithms. This resulted in a substantial savings
in time for all Tiler reservation and free operations. Typical
savings are in the range of 28% decrease in time taken with larger
buffers showing a 80%+ decrease.
Signed-off-by: Andy Gross <andy.gross@ti.com>
Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ti.com>
omap_dmm_tiler.c can't be compiled as a module and it is built
unconditionally as part of omapdrm. Since it can't be used as a module,
there is no need for it to have an unused MODULE_ALIAS().
Signed-off-by: Luis de Bethencourt <luisbg@osg.samsung.com>
Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ti.com>
The DMM driver uses a timeout of 1 ms to wait for DMM transaction to
finish. While DMM should always finish the operation within that time,
the timeout is rather strict. Small misbehavior of the system (e.g. an
irq taking too long) could trigger the timeout.
As the DMM is a critical piece of code for display memory management,
let's increase the timeout to 100 ms so that we are less likely to fail
a memory allocation in case of system misbehaviors. 100 ms is just a
guess of a reasonably large timeout. The HW should accomplish the task
in less than 1 ms.
Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ti.com>
Use the <...> include style instead of "..." for DRM headers and sort
the headers alphabetically to ease detection of duplicates.
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ti.com>
Setting a dev_pm_ops resume callback but not a set of hibernation
handler means that pm function will not be called upon hibernation.
Fix this by using SIMPLE_DEV_PM_OPS, which appropriately assigns the
suspend and hibernation handlers and move omap_dmm_resume under
CONFIG_PM_SLEEP to avoid build warnings.
Signed-off-by: Grygorii Strashko <grygorii.strashko@linaro.org>
Reviewed-by: Rob Clark <robdclark@gmail.com>
[tomi valkeinen: add missing 'static']
Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ti.com>
The omapdrm DMM code sometimes crashes with:
WARNING: CPU: 0 PID: 1235 at lib/list_debug.c:36 __list_add+0x8c/0xbc()
list_add double add: new=e9265368, prev=e90139c4, next=e9265368.
This is caused by the code calling release_engine() twice for the same
engine.
dmm_txn_commit(wait=true) call is supposed to wait until the DMM
transaction has been finished. And it does that, but it does not wait
for the irq handler to finish.
What happens is that the irq handler is triggered, and it either wakes
up the thread that called dmm_txn_commit(), or that thread never even
slept because the transaction was finished in the HW very quickly. That
thread then continues executing, even if the irq handler is not yet
finished, and a new transaction may be initiated. If that transaction is
async (i.e. wait=false), a 'async' flag is set to true. The original irq
handler, which has yet not finished, then sees the transaction as
'async', even if it was supposed to be 'sync'.
When that happens, the irq handler does an extra release_engine() call
because it thinks it need to release the engine, leading to the crash.
This patch fixes the issue by using completion to ensure that the irq
handler has finished before a dmm_txn_commit(wait=true) may continue.
Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ti.com>
The tiler irq handler uses engine->async value, but the code that sets
engine->async and enables the interrupt does not have a barrier. This
may cause the irq handler to see the old value of engine->async, causing
memory corruption.
Reported-by: Harinarayan Bhatta <harinarayan@ti.com>
Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ti.com>
On OMAP5 it is not possible to use TILER buffer with CPU when caching or
write-combining is used. Doing so leads to errors from the memory
manager.
However, on OMAP4, write-combining works fine.
This patch adds platform specific data for the TILER, and a function
tiler_get_cpu_cache_flags() which can be used to get the caching mode to
be used.
Note that without write-combining the use of the TILER buffer with CPU
is unusably slow. It's still good to have it operational for testing
purposes.
Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ti.com>
Regular randconfig nightly testing has detected problems with omapdrm.
omapdrm fails to build when the kernel is built to support 64-bit DMA
addresses and/or 64-bit physical addresses due to an assumption about
the width of these types.
Use %pad to print DMA addresses, rather than %x or %Zx (which is even
more wrong than %x). Avoid passing a uint32_t pointer into a function
which expects dma_addr_t pointer.
drivers/gpu/drm/omapdrm/omap_plane.c: In function 'omap_plane_pre_apply':
drivers/gpu/drm/omapdrm/omap_plane.c:145:2: error: format '%x' expects argument of type 'unsigned int', but argument 5 has type 'dma_addr_t' [-Werror=format]
drivers/gpu/drm/omapdrm/omap_plane.c:145:2: error: format '%x' expects argument of type 'unsigned int', but argument 6 has type 'dma_addr_t' [-Werror=format]
make[5]: *** [drivers/gpu/drm/omapdrm/omap_plane.o] Error 1
drivers/gpu/drm/omapdrm/omap_gem.c: In function 'omap_gem_get_paddr':
drivers/gpu/drm/omapdrm/omap_gem.c:794:4: error: format '%x' expects argument of type 'unsigned int', but argument 3 has type 'dma_addr_t' [-Werror=format]
drivers/gpu/drm/omapdrm/omap_gem.c: In function 'omap_gem_describe':
drivers/gpu/drm/omapdrm/omap_gem.c:991:4: error: format '%Zx' expects argument of type 'size_t', but argument 7 has type 'dma_addr_t' [-Werror=format]
drivers/gpu/drm/omapdrm/omap_gem.c: In function 'omap_gem_init':
drivers/gpu/drm/omapdrm/omap_gem.c:1470:4: error: format '%x' expects argument of type 'unsigned int', but argument 7 has type 'dma_addr_t' [-Werror=format]
make[5]: *** [drivers/gpu/drm/omapdrm/omap_gem.o] Error 1
drivers/gpu/drm/omapdrm/omap_dmm_tiler.c: In function 'dmm_txn_append':
drivers/gpu/drm/omapdrm/omap_dmm_tiler.c:226:2: error: passing argument 3 of 'alloc_dma' from incompatible pointer type [-Werror]
make[5]: *** [drivers/gpu/drm/omapdrm/omap_dmm_tiler.o] Error 1
make[5]: Target `__build' not remade because of errors.
make[4]: *** [drivers/gpu/drm/omapdrm] Error 2
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
Signed-off-by: Dave Airlie <airlied@redhat.com>
Enable use of DT for DMM/Tiler.
Originally worked on by Andy Gross <andygro@gmail.com>
Cc: Andy Gross <andygro@gmail.com>
Cc: DRI Development <dri-devel@lists.freedesktop.org>
Signed-off-by: Archit Taneja <archit@ti.com>
[tomi.valkeinen@ti.com: use of_match_ptr()]
Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ti.com>
The correct way for a driver to specify the coherent DMA mask is
not to directly access the field in the struct device, but to use
dma_set_coherent_mask(). Only arch and bus code should access this
member directly.
Convert all direct write accesses to using the correct API.
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
We're taking the sizeof() the wrong thing so it doesn't clear the whole
buffer.
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Acked-by: Rob Clark <robdclark@gmail.com>
Signed-off-by: Dave Airlie <airlied@gmail.com>
Pull drm merge from Dave Airlie:
"Highlights:
- TI LCD controller KMS driver
- TI OMAP KMS driver merged from staging
- drop gma500 stub driver
- the fbcon locking fixes
- the vgacon dirty like zebra fix.
- open firmware videomode and hdmi common code helpers
- major locking rework for kms object handling - pageflip/cursor
won't block on polling anymore!
- fbcon helper and prime helper cleanups
- i915: all over the map, haswell power well enhancements, valleyview
macro horrors cleaned up, killing lots of legacy GTT code,
- radeon: CS ioctl unification, deprecated UMS support, gpu reset
rework, VM fixes
- nouveau: reworked thermal code, external dp/tmds encoder support
(anx9805), fences sleep instead of polling,
- exynos: all over the driver fixes."
Lovely conflict in radeon/evergreen_cs.c between commit de0babd60d
("drm/radeon: enforce use of radeon_get_ib_value when reading user cmd")
and the new changes that modified that evergreen_dma_cs_parse()
function.
* 'drm-next' of git://people.freedesktop.org/~airlied/linux: (508 commits)
drm/tilcdc: only build on arm
drm/i915: Revert hdmi HDP pin checks
drm/tegra: Add list of framebuffers to debugfs
drm/tegra: Fix color expansion
drm/tegra: Split DC_CMD_STATE_CONTROL register write
drm/tegra: Implement page-flipping support
drm/tegra: Implement VBLANK support
drm/tegra: Implement .mode_set_base()
drm/tegra: Add plane support
drm/tegra: Remove bogus tegra_framebuffer structure
drm: Add consistency check for page-flipping
drm/radeon: Use generic HDMI infoframe helpers
drm/tegra: Use generic HDMI infoframe helpers
drm: Add EDID helper documentation
drm: Add HDMI infoframe helpers
video: Add generic HDMI infoframe helpers
drm: Add some missing forward declarations
drm: Move mode tables to drm_edid.c
drm: Remove duplicate drm_mode_cea_vic()
gma500: Fix n, m1 and m2 clock limits for sdvo and lvds
...
Now that the omapdss interface has been reworked so that omapdrm can use
dispc directly, we have been able to fix the remaining functional kms
issues with omapdrm. And in the mean time the PM sequencing and many
other of that open issues have been solved. So I think it makes sense
to finally move omapdrm out of staging.
Signed-off-by: Rob Clark <robdclark@gmail.com>
Tomi Valkeinen