| Age | Commit message (Collapse) | Author | Files | Lines |
|---|
|
Use kzalloc so other code doesn't need to worry about uninitialized members.
Drop the non-standard GFP flags, as we really don't want to fail the submit
when under slight memory pressure. Remove one level of indentation by using
an early return if the allocation failed. Also remove the unused drm device
member.
Signed-off-by: Lucas Stach <[email protected]>
Reviewed-by: Philipp Zabel <[email protected]>
|
|
When manipulating the kernel command buffer the GPU mutex must be held, as
otherwise different callers might try to replace the same part of the
buffer, wreacking havok in the GPU execution.
Signed-off-by: Lucas Stach <[email protected]>
Reviewed-by: Philipp Zabel <[email protected]>
|
|
Inserting the END command when suspending the GPU is changing the
command buffer state, which requires the GPU to be held.
Signed-off-by: Lucas Stach <[email protected]>
Reviewed-by: Philipp Zabel <[email protected]>
Reviewed-by: Christian Gmeiner <[email protected]>
|
|
While the etnaviv workqueue needs to be ordered, as we rely on work items
being executed in queuing order, this is only true for a single GPU.
Having a shared workqueue for all GPUs in the system limits concurrency
artificially.
Getting each GPU its own ordered workqueue still meets our ordering
expectations and enables retire workers to run concurrently.
Signed-off-by: Lucas Stach <[email protected]>
Reviewed-by: Philipp Zabel <[email protected]>
|
|
There is no need to store this in the gpu struct. MMU flushes are triggered
correctly in reaction to MMU maps and unmaps, independent of the current ctx.
Any required pipe switches can be infered from the current and the desired
GPU exec state.
Signed-off-by: Lucas Stach <[email protected]>
Reviewed-by: Philipp Zabel <[email protected]>
Reviewed-by: Christian Gmeiner <[email protected]>
|
|
There is no need to synchronize with oustanding retire jobs if the object
has gone idle. Retire jobs only ever change the object state from active to
idle, not the other way around.
The IOVA put race is uncritical, as the GEM_WAIT ioctl itself is holding
a reference to the GEM object, so the retire worker will not pull the
object into the CPU domain, which is the thing we are trying to guard
against with etnaviv_gpu_wait_obj_inactive. The ordering of the various
counts and waits may change a bit, but the userspace visible behavior at
the bounds of the syscall are unchanged.
Signed-off-by: Lucas Stach <[email protected]>
Reviewed-by: Philipp Zabel <[email protected]>
|
|
Flush and prefetch are properly handled in the buffer code, data endianess
would need much wider changes than adding something to this single function.
Signed-off-by: Lucas Stach <[email protected]>
Reviewed-by: Christian Gmeiner <[email protected]>
|
|
Now that the userptr BO handling doesn't rely on the userspace restarting
the submit after object population, there is no need to special case the
-EAGAIN return value anymore.
Signed-off-by: Lucas Stach <[email protected]>
Reviewed-by: Philipp Zabel <[email protected]>
|
|
All code paths which populate userptr BOs are fine with the get_pages
function taking the mmap_sem lock. This allows to get rid of the pretty
involved architecture with a worker being scheduled if the mmap_sem
needs to be taken, but instead call GUP directly and allow it to take
the lock if necessary.
This simplifies the code a lot and removes the possibility of this
function returning -EAGAIN, which complicates object population
handling at the callers.
A notable change in behavior is that we don't allow a process to populate
objects with user pages from a foreign MM anymore. This would have been an
invalid use before, as it breaks the assumptions made in the etnaviv kernel
driver to enfore cache coherence. We now disallow this by rejecting the
request to populate those objects. Well behaving userspace is unaffected by
this change.
Signed-off-by: Lucas Stach <[email protected]>
|
|
This function never fails, as it does nothing more than adding the GEM
object to the global device list. Making this explicit through the void
return type allows to drop some unnecessary error handling.
Signed-off-by: Lucas Stach <[email protected]>
Reviewed-by: Christian Gmeiner <[email protected]>
Reviewed-by: Philipp Zabel <[email protected]>
|
|
This function has only one caller and it isn't expected that there will
be any more in the future. Folding this function into the caller is
helping the readability.
Signed-off-by: Lucas Stach <[email protected]>
Reviewed-by: Philipp Zabel <[email protected]>
Reviewed-by: Christian Gmeiner <[email protected]>
|
|
The current userptr page population will defer work to a work item if
needed to avoid ever taking the mmap_sem in the direct call path. With
the more fine-grained locking in etnaviv this isn't needed anymore, so
a future commit will simplify this code.
Add a lockdep annotation to validate the assumption that the mmap_sem
can be taken in the direct call path.
Signed-off-by: Lucas Stach <[email protected]>
Reviewed-by: Christian Gmeiner <[email protected]>
|
|
Userptr, prime and shmem buffer objects have different lock ordering
requirements. This is mostly due to the fact that we don't allow to mmap
userptr buffers, so we won't ever end up in our fault handler for those,
so some of the code paths are never called with the mmap_sem held.
To avoid lockdep false positives, split them up into different lock classes.
Signed-off-by: Lucas Stach <[email protected]>
Reviewed-by: Christian Gmeiner <[email protected]>
Reviewed-by: Philipp Zabel <[email protected]>
|
|
If the FE is restarted before the sync point event is cleared, the GPU
might trigger a completion IRQ for the next sync point, corrupting
the state of the currently running worker.
Signed-off-by: Lucas Stach <[email protected]>
Reviewed-by: Philipp Zabel <[email protected]>
Reviewed-by: Christian Gmeiner <[email protected]>
|
|
In the selftests, we don't want to force an oom and would rather
ENOMEM be reported. In this case, we would rather the allocation for the
random array to fail.
Signed-off-by: Chris Wilson <[email protected]>
Reviewed-by: Matthew Auld <[email protected]>
Link: https://patchwork.freedesktop.org/patch/msgid/[email protected]
|
|
Reduce the number of GGTT PTE operations to speed the test up, but we
reduce the likelihood of spotting a coherency error in those operations.
However, Broxton is sporadically timing on this test, presumably because
its GGTT operations are all uncached.
Signed-off-by: Chris Wilson <[email protected]>
Reviewed-by: Matthew Auld <[email protected]>
Link: https://patchwork.freedesktop.org/patch/msgid/[email protected]
|
|
The omap4 CEC hardware cannot tell a Nack from a Low Drive from an
Arbitration Lost error, so just report a Nack, which is almost
certainly the reason for the error anyway.
This also simplifies the implementation. The only three interrupts
that need to be enabled are:
Transmit Buffer Full/Empty Change event: triggered when the
transmit finished successfully and cleared the buffer.
Receiver FIFO Not Empty event: triggered when a message was received.
Frame Retransmit Count Exceeded event: triggered when a transmit
failed repeatedly, usually due to the message being Nacked. Other
reasons are possible (Low Drive, Arbitration Lost) but there is no
way to know. If this happens the TX buffer needs to be cleared
manually.
While testing various error conditions I noticed that the hardware
can receive messages up to 18 bytes in total, which exceeds the legal
maximum of 16. This could cause a buffer overflow, so we check for
this and constrain the size to 16 bytes.
The old incorrect interrupt handler could cause the CEC framework to
enter into a bad state because it mis-detected the "Start Bit Irregularity
event" as an ARB_LOST transmit error when it actually is a receive error
which should be ignored.
Signed-off-by: Hans Verkuil <[email protected]>
Reported-by: Henrik Austad <[email protected]>
Tested-by: Henrik Austad <[email protected]>
Tested-by: Hans Verkuil <[email protected]>
Signed-off-by: Tomi Valkeinen <[email protected]>
|
|
We have plenty of global registers and whatnot programmed without
any further locking by the modeset code. Currently non-bocking
modesets are allowed to execute in parallel which could corrupt
said registers.
To avoid the problem let's run all non-blocking modesets on an
ordered workqueue. We still put page flips etc. to system_unbound_wq
allowing page flips on one pipe to execute in parallel with page flips
or a modeset on a another pipe (assuming no known state is shared
between them, at which point they would have been added to the same
atomic commit and serialized that way).
Blocking modesets are already serialized with each other by
connection_mutex, and thus are safe. To serialize them with
non-blocking modesets we just flush the workqueue before executing
blocking modesets.
Cc: Daniel Vetter <[email protected]>
Cc: Maarten Lankhorst <[email protected]>
Fixes: 94f050246b42 ("drm/i915: nonblocking commit")
Signed-off-by: Ville Syrjälä <[email protected]>
Link: https://patchwork.freedesktop.org/patch/msgid/[email protected]
Acked-by: Daniel Vetter <[email protected]>
Reviewed-by: Maarten Lankhorst <[email protected]>
(cherry picked from commit 757fffcfdffb6c0dd46c1b264091c36b4e5a86ae)
Signed-off-by: Jani Nikula <[email protected]>
|
|
Prevent the DMC from destroying GMBUS transfers on GLK. GMBUS
lives in PG1 so DC off is all we need.
Cc: [email protected]
Signed-off-by: Ville Syrjälä <[email protected]>
Link: https://patchwork.freedesktop.org/patch/msgid/[email protected]
Reviewed-by: Dhinakaran Pandiyan <[email protected]>
(cherry picked from commit 156961ae7bdf6feb72778e8da83d321b273343fd)
Signed-off-by: Jani Nikula <[email protected]>
|
|
Commit 77affa31722b ("drm/i915/psr: Fix compiler warnings for
hsw_psr_disable()") swapped status and control registers while fixing
indentation. The _ctl at the end of the status register name must have to
led to this.
Fixes: 77affa31722b ("drm/i915/psr: Fix compiler warnings for hsw_psr_disable()")
References: https://www.mrc-cbu.cam.ac.uk/people/matt.davis/cmabridge/
Cc: Chris Wilson <[email protected]>
Cc: Rodrigo Vivi <[email protected]>
Signed-off-by: Dhinakaran Pandiyan <[email protected]>
Link: https://patchwork.freedesktop.org/patch/msgid/[email protected]
Reviewed-by: Chris Wilson <[email protected]>
Signed-off-by: Chris Wilson <[email protected]>
(cherry picked from commit 14c6547d6df641d3e41fa4f4164f6e267ebfab89)
Signed-off-by: Jani Nikula <[email protected]>
|
|
Exynos DRM IPP subsystem is in fact non-functional and frankly speaking
dead-code. This patch clearly marks that Exynos DRM IPP subsystem is
broken and never really functional. It will be replaced by a completely
rewritten API.
Exynos DRM IPP user-space API can be obsoleted for the following
reasons:
1. Exynos DRM IPP user-space API can be optional in Exynos DRM, so
userspace should not rely that it is always available and should have
a software fallback in case it is not there.
2. The only mode which was initially semi-working was memory-to-memory
image processing. The remaining modes (LCD-"writeback" and "output")
were never operational due to missing code (both in mainline and even
vendor kernels).
3. Exynos DRM IPP mainline user-space API compatibility for
memory-to-memory got broken very early by commit 083500baefd5 ("drm:
remove DRM_FORMAT_NV12MT", which removed the support for tiled formats,
the main feature which made this API somehow useful on Exynos platforms
(video codec that time produced only tiled frames, to implement xvideo
or any other video overlay, one has to de-tile them for proper
display).
4. Broken drivers. Especially once support for IOMMU has been added,
it revealed that drivers don't configure DMA operations properly and in
many cases operate outside the provided buffers trashing memory around.
5. Need for external patches. Although IPP user-space API has been used
in some vendor kernels, but in such cases there were additional patches
applied (like reverting mentioned 083500baefd5 patch) what means that
those userspace apps which might use it, still won't work with the
mainline kernel version.
We don't have time machines, so we cannot change it, but Exynos DRM IPP
extension should never have been merged to mainline in that form.
Exynos IPP subsystem and user-space API will be rewritten, so remove
current IPP core code and mark existing drivers as BROKEN.
Signed-off-by: Marek Szyprowski <[email protected]>
Acked-by: Daniel Stone <[email protected]>
Acked-by: Krzysztof Kozlowski <[email protected]>
Signed-off-by: Inki Dae <[email protected]>
|
|
Although header is included only once but still having an include guard
is a good practice. To avoid confusion, add SoC prefix to existing
Exynos5433 header include guard.
Signed-off-by: Krzysztof Kozlowski <[email protected]>
Signed-off-by: Inki Dae <[email protected]>
|
|
The DECON headers contain only defines for registers. There are no
other drivers using them so this should be put locally to the Exynos DRM
driver. Keeping headers local helps managing the code.
Suggested-by: Marek Szyprowski <[email protected]>
Signed-off-by: Krzysztof Kozlowski <[email protected]>
Signed-off-by: Inki Dae <[email protected]>
|
|
devm_ioremap_resource() already checks if the resource is NULL, so
remove the unnecessary platform_get_resource() error check.
Cc: Inki Dae <[email protected]>
Signed-off-by: Fabio Estevam <[email protected]>
Reviewed-by: Andrzej Hajda <[email protected]>
Signed-off-by: Inki Dae <[email protected]>
|
|
The amdgpu_vm_frag_ptes will call amdgpu_vm_update_ptes, and for buffer
object that has shadow buffer, need twice commands.
Signed-off-by: Emily Deng <[email protected]>
Reviewed-by: Chunming Zhou <[email protected]>
Signed-off-by: Alex Deucher <[email protected]>
|
|
divider_recalc_rate() is an helper function used by clock divider of
different types, so the structure containing the 'hw' pointer is not
always a 'struct clk_divider'
At the following line:
> div = _get_div(table, val, flags, divider->width);
in several cases, the value of 'divider->width' is garbage as the actual
structure behind this memory is not a 'struct clk_divider'
Fortunately, this width value is used by _get_val() only when
CLK_DIVIDER_MAX_AT_ZERO flag is set. This has never been the case so
far when the structure is not a 'struct clk_divider'. This is probably
why we did not notice this bug before
Fixes: afe76c8fd030 ("clk: allow a clk divider with max divisor when zero")
Signed-off-by: Jerome Brunet <[email protected]>
Acked-by: Alexandre Belloni <[email protected]>
Acked-by: Sylvain Lemieux <[email protected]>
Signed-off-by: Stephen Boyd <[email protected]>
|
|
Existing debugfs entry i915_drrs_status is updated with whether PSR
is the cause for DRRS disabled state.
[v2]: Dropped the module parameter details as ctl moved from module
parameter to debugfs. [Rodrigo]
[v3]: Crtc ID information is dropped as there is no immediate usecase.
[Rodrigo].
Signed-off-by: C, Ramalingam <[email protected]>
Reviewed-by: Rodrigo Vivi <[email protected]>
Signed-off-by: Rodrigo Vivi <[email protected]>
Link: https://patchwork.freedesktop.org/patch/msgid/[email protected]
|
|
Debugfs called i915_drrs_ctl is added to enable and disable the
eDP DRRS. Writing 0 will disable the feature, whereas non-zero
will enable the feature.
Possibility of disabling the DRRS, enables the testing of the
frontbuffer tracking based features (FBC, DRRS and PSR) as
standalone or any combination of the set.
[v2]: ctl interface is moved from module parameter to debugfs [Rodrigo]
Signed-off-by: C, Ramalingam <[email protected]>
Reviewed-by: Rodrigo Vivi <[email protected]>
Signed-off-by: Rodrigo Vivi <[email protected]>
Link: https://patchwork.freedesktop.org/patch/msgid/[email protected]
|
|
if the bo shares same reservation object then not lock it again
at swapout time to make it possible to swap out.
v2: refine the commmit message
Reviewed-by: Thomas Hellström <[email protected]>
Reviewed-by: Christian König <[email protected]>
Reviewed-by: Chuming Zhou <[email protected]>
Signed-off-by: Roger He <[email protected]>
Signed-off-by: Alex Deucher <[email protected]>
|
|
extract a function as ttm_bo_evict_swapout_allowable since eviction and
swapout can share same logic.
v2: modify commit message and add description in the code
Reviewed-by: Thomas Hellström <[email protected]>
Reviewed-by: Christian König <[email protected]>
Reviewed-by: Chuming Zhou <[email protected]>
Signed-off-by: Roger He <[email protected]>
Signed-off-by: Alex Deucher <[email protected]>
|
|
forward the operation context to ttm_tt_bind as well,
and the ultimate goal is swapout enablement for reserved BOs.
v2: use common term rather than amd specific
Reviewed-by: Thomas Hellström <[email protected]>
Reviewed-by: Christian König <[email protected]>
Reviewed-by: Chuming Zhou <[email protected]>
Signed-off-by: Roger He <[email protected]>
Signed-off-by: Alex Deucher <[email protected]>
|
|
forward the operation context to ttm_tt_populate as well,
and the ultimate goal is swapout enablement for reserved BOs.
v2: squash in fix for vboxvideo
Reviewed-by: Christian König <[email protected]>
Signed-off-by: Roger He <[email protected]>
Signed-off-by: Alex Deucher <[email protected]>
|
|
Signed-off-by: Al Viro <[email protected]>
|
|
git://anongit.freedesktop.org/drm/drm-intel into drm-next
- Allow internal page allocation to fail (Chris)
- More improvements on logs, dumps, and trace (Chris, Michal)
- Coffee Lake important fix for stolen memory (Lucas)
- Continue to make GPU reset more robust as well
improving selftest coverage for it (Chris)
- Unifying debugfs return codes (Michal)
- Using existing helper for testing obj pages (Matthew)
- Organize and improve gem_request tracepoints (Lionel)
- Protect DDI port to DPLL map from theoretical race (Rodrigo)
- ... and consequently fixing the indentation on this DDI clk selection function (Chris)
- ... and consequently properly serializing non-blocking modesets (Ville)
- Add support for horizontal plane flipping on Cannonlake (Joonas)
- Two Cannonlake Workarounds for better stability (Rafael)
- Fix mess around PSR registers (DK)
- More Coffee Lake PCI IDs (Rodrigo)
- Remove CSS modifiers on pipe C of Geminilake (Krisman)
- Disable all planes for load detection (Ville)
- Reorg on i915 display headers (Michal)
- Avoid enabling movntdqa optimization on hypervisor guest (Changbin)
GVT:
- more mmio switch optimization (Weinan)
- cleanup i915_reg_t vs. offset usage (Zhenyu)
- move write protect handler out of mmio handler (Zhenyu)
* tag 'drm-intel-next-2017-12-22' of git://anongit.freedesktop.org/drm/drm-intel: (55 commits)
drm/i915: Update DRIVER_DATE to 20171222
drm/i915: Show HWSP in intel_engine_dump()
drm/i915: Assert that the request is on the execution queue before being removed
drm/i915/execlists: Show preemption progress in GEM_TRACE
drm/i915: Put all non-blocking modesets onto an ordered wq
drm/i915: Disable GMBUS clock gating around GMBUS transfers on gen9+
drm/i915: Clean up the PNV bit banging vs. GMBUS clock gating w/a
drm/i915: No need to power up PG2 for GMBUS on BXT
drm/i915: Disable DC states around GMBUS on GLK
drm/i915: Do not enable movntdqa optimization in hypervisor guest
drm/i915: Dump device info at once
drm/i915: Add pretty printer for runtime part of intel_device_info
drm/i915: Update intel_device_info_runtime_init() parameter
drm/i915: Move intel_device_info definitions to its own header
drm/i915: Move opregion definitions to dedicated intel_opregion.h
drm/i915: Move display related definitions to dedicated header
drm/i915: Move some utility functions to i915_util.h
drm/i915/gvt: move write protect handler out of mmio emulation function
drm/i915/gvt: cleanup usage for typed mmio reg vs. offset
drm/i915/gvt: Fix pipe A enable as default for vgpu
...
|
|
git://anongit.freedesktop.org/drm/drm-intel into drm-fixes
GLK pipe C related fix, and a gvt fix.
* tag 'drm-intel-fixes-2017-12-22-1' of git://anongit.freedesktop.org/drm/drm-intel:
i915: Reject CCS modifiers for pipe C on Geminilake
drm/i915/gvt: Fix pipe A enable as default for vgpu
|
|
forward the operation context to ttm_mem_global_alloc_page as well,
and the ultimate goal is swapout enablement for reserved BOs.
Here reserved BOs refer to all the BOs which share same reservation object
Reviewed-by: Christian König <[email protected]>
Signed-off-by: Roger He <[email protected]>
Signed-off-by: Alex Deucher <[email protected]>
|
|
forward the operation context to ttm_mem_global_alloc as well, and the
ultimate goal is swapout enablement for reserved BOs
Reviewed-by: Thomas Hellstrom <[email protected]>
Reviewed-by: Christian König <[email protected]>
Signed-off-by: Roger He <[email protected]>
Signed-off-by: Alex Deucher <[email protected]>
|
|
remove the extra indirection because we have only one implementation anyway
Reviewed-by: Christian König <[email protected]>
Signed-off-by: Roger He <[email protected]>
Signed-off-by: Alex Deucher <[email protected]>
|
|
The default interface situation has been taken into the framework, so
remove the default set of each module.
Signed-off-by: Tan Xiaojun <[email protected]>
Reviewed-by: Christian König <[email protected]>
Signed-off-by: Alex Deucher <[email protected]>
|
|
The default interface situation has been taken into the framework, so
remove the default set of each module.
Signed-off-by: Tan Xiaojun <[email protected]>
Reviewed-by: Christian König <[email protected]>
Signed-off-by: Alex Deucher <[email protected]>
|
|
The default interface situation has been taken into the framework, so
remove the default set of each module.
Signed-off-by: Tan Xiaojun <[email protected]>
Reviewed-by: Christian König <[email protected]>
Signed-off-by: Alex Deucher <[email protected]>
|
|
The default interface situation has been taken into the framework, so
remove the default set of each module.
Signed-off-by: Tan Xiaojun <[email protected]>
Reviewed-by: Christian König <[email protected]>
Signed-off-by: Alex Deucher <[email protected]>
|
|
The default interface situation has been taken into the framework, so
remove the default set of each module.
Signed-off-by: Tan Xiaojun <[email protected]>
Reviewed-by: Christian König <[email protected]>
Signed-off-by: Alex Deucher <[email protected]>
|
|
The default interface situation has been taken into the framework, so
remove the default set of each module.
Signed-off-by: Tan Xiaojun <[email protected]>
Reviewed-by: Christian König <[email protected]>
Signed-off-by: Alex Deucher <[email protected]>
|
|
The default interface situation has been taken into the framework, so
remove the default set of each module.
Signed-off-by: Tan Xiaojun <[email protected]>
Reviewed-by: Christian König <[email protected]>
Signed-off-by: Alex Deucher <[email protected]>
|
|
The default interface situation has been taken into the framework, so
remove the default set of each module.
Signed-off-by: Tan Xiaojun <[email protected]>
Reviewed-by: Christian König <[email protected]>
Signed-off-by: Alex Deucher <[email protected]>
|
|
The default interface situation has been taken into the framework, so
remove the default set of each module.
Signed-off-by: Tan Xiaojun <[email protected]>
Reviewed-by: Christian König <[email protected]>
Signed-off-by: Alex Deucher <[email protected]>
|
|
The io_mem_pfn field was added in commit ea642c3216cb ("drm/ttm: add
io_mem_pfn callback") and is called unconditionally. However, not all
drivers were updated to set it.
Use the ttm_bo_default_io_mem_pfn function if a driver did not set its
own. And add new function ttm_bo_io_mem_pfn() as wrapper.
Signed-off-by: Michal Srb <[email protected]>
Signed-off-by: Tan Xiaojun <[email protected]>
Reviewed-by: Christian König <[email protected]>
Signed-off-by: Alex Deucher <[email protected]>
|
|
There was a small window between unreserve and second reserve where the
freshly allocated BO could have been evicted without the VM noticing it.
Signed-off-by: Christian König <[email protected]>
Reviewed-by: Chunming Zhou <[email protected]>
Signed-off-by: Alex Deucher <[email protected]>
|
|
Thomas actually noticed that, but I didn't realized what he meant until
now.
Signed-off-by: Christian König <[email protected]>
Reviewed-by: Roger He <[email protected]>
Signed-off-by: Alex Deucher <[email protected]>
|