aboutsummaryrefslogtreecommitdiff
path: root/drivers/gpu/drm/etnaviv
AgeCommit message (Collapse)AuthorFilesLines
2024-07-29Merge drm/drm-next into drm-misc-nextThomas Zimmermann10-92/+238
Backmerging to get a late RC of v6.10 before moving into v6.11. Signed-off-by: Thomas Zimmermann <[email protected]>
2024-07-25drm/scheduler: remove full_recover from drm_sched_startChristian König1-2/+2
This was basically just another one of amdgpus hacks. The parameter allowed to restart the scheduler without turning fence signaling on again. That this is absolutely not a good idea should be obvious by now since the fences will then just sit there and never signal. While at it cleanup the code a bit. Signed-off-by: Christian König <[email protected]> Reviewed-by: Matthew Brost <[email protected]> Link: https://patchwork.freedesktop.org/patch/msgid/[email protected]
2024-06-26drm/etnaviv: don't block scheduler when GPU is still activeLucas Stach1-5/+4
Since 45ecaea73883 ("drm/sched: Partial revert of 'drm/sched: Keep s_fence->parent pointer'") still active jobs aren't put back in the pending list on drm_sched_start(), as they don't have a active parent fence anymore, so if the GPU is still working and the timeout is extended, all currently active jobs will be freed. To avoid prematurely freeing jobs that are still active on the GPU, don't block the scheduler until we are fully committed to actually reset the GPU. As the current job is already removed from the pending list and will not be put back when drm_sched_start() isn't called, we must make sure to put the job back on the pending list when extending the timeout. Cc: [email protected] #6.0 Signed-off-by: Lucas Stach <[email protected]> Reviewed-by: Philipp Zabel <[email protected]> Reviewed-by: Christian Gmeiner <[email protected]>
2024-06-26drm/etnaviv: reduce number of ktime_get calls in IRQ handlerLucas Stach1-1/+2
A single IRQ might signal the completion of multiple jobs/fences at once. There is no point in attaching a new timestamp to each fence that only differs in when exactly the IRQ handler was able to process this fence. Get a single timestamp when the IRQ handler has determined that there are completed jobs and reuse this for all fences that get signalled by the handler. Signed-off-by: Lucas Stach <[email protected]> Reviewed-by: Philipp Zabel <[email protected]> Reviewed-by: Christian Gmeiner <[email protected]>
2024-06-26drm/etnaviv: fix DMA direction handling for cached RW buffersLucas Stach1-2/+4
The dma sync operation needs to be done with DMA_BIDIRECTIONAL when the BO is prepared for both read and write operations. Fixes: a8c21a5451d8 ("drm/etnaviv: add initial etnaviv DRM driver") Signed-off-by: Lucas Stach <[email protected]> Reviewed-by: Philipp Zabel <[email protected]> Reviewed-by: Christian Gmeiner <[email protected]>
2024-06-26drm/etnaviv: switch devcoredump allocations to GFP_NOWAITLucas Stach1-3/+2
The etnaviv devcoredump is created in the GPU reset path, which must make forward progress to avoid stalling memory reclaim on unsignalled dma fences. The currently used __GFP_NORETRY does not prohibit sleeping on direct reclaim, breaking the forward progress guarantee. Switch to GFP_NOWAIT, which allows background reclaim to be triggered, but avoids any stalls waiting for direct reclaim. Signed-off-by: Lucas Stach <[email protected]> Reviewed-by: Daniel Vetter <[email protected]>
2024-06-19drm/etnaviv: don't disable TS on MMUv2 core when moving the linear windowLucas Stach1-2/+5
On MMUv2 cores the linear window is only relevant when starting the FE, before the MMU has been activated. Once the MMU is active, all accesses are translated with no way to bypass the MMU via the linear window. Thus TS ignoring the linear window offset is not an issue on cores with MMUv2 present and there is no need to disable TS when we need to move the linear window. Signed-off-by: Lucas Stach <[email protected]> Tested-by: Joao Paulo Goncalves <[email protected]>
2024-06-19drm/etnaviv: Read some FE registers twiceDerek Foreman1-0/+8
On some hardware (such at the GC7000 rev 6009), these registers need to be read twice to return the correct value. Hide that in gpu_read(). Signed-off-by: Derek Foreman <[email protected]> Signed-off-by: Lucas Stach <[email protected]>
2024-05-16drm/etnaviv: Disable SH_EU clock gating on VIPNano-Si+Philipp Zabel1-0/+6
Disable SH_EU clock gating for the VIPNano-Si+ NPU on i.MX8MP and for other affected core revisions. Taken from linux-imx lf-6.1.36-2.1.0, specifically [1]. [1] https://github.com/nxp-imx/linux-imx/blob/lf-6.1.36-2.1.0/drivers/mxc/gpu-viv/hal/kernel/arch/gc_hal_kernel_hardware.c#L2747-L2761 Signed-off-by: Philipp Zabel <[email protected]> Signed-off-by: Lucas Stach <[email protected]>
2024-05-16drm/etnaviv: Turn etnaviv_is_model_rev() into a functionPhilipp Zabel1-32/+34
Turn the etnaviv_is_model_rev() macro into a static inline function. Use the raw model number as a parameter instead of the chipModel_GCxxxx defines. This reduces synchronization requirements for the generated headers. For newer hardware, the GCxxxx names are not the correct model names anyway. For example, model 0x8000 NPUs are called VIPNano-QI/SI(+) by VeriSilicon. Signed-off-by: Philipp Zabel <[email protected]> Signed-off-by: Lucas Stach <[email protected]>
2024-05-16drm/etnaviv: Update hardware headers from rnndbPhilipp Zabel5-46/+173
Update the state HI header from the rnndb commit 8d7ee714cfe2 ("Merge pull request #24 from pH5/unknown-3950"). Signed-off-by: Philipp Zabel <[email protected]> Reviewed-by: Christian Gmeiner <[email protected]> Signed-off-by: Lucas Stach <[email protected]>
2024-05-16drm/etnaviv: drop driver owner assignmentKrzysztof Kozlowski1-1/+0
Core in platform_driver_register() already sets the .owner, so driver does not need to. Whatever is set here will be anyway overwritten by main driver calling platform_driver_register(). Signed-off-by: Krzysztof Kozlowski <[email protected]> Reviewed-by: Christian Gmeiner <[email protected]> Signed-off-by: Lucas Stach <[email protected]>
2024-04-25Revert "drm/etnaviv: Expose a few more chipspecs to userspace"Christian Gmeiner3-66/+0
This reverts commit 1dccdba084897443d116508a8ed71e0ac8a031a4. In userspace a different approach was choosen - hwdb. As a result, there is no need for these values. Signed-off-by: Christian Gmeiner <[email protected]> Reviewed-by: Tomeu Vizoso <[email protected]> Signed-off-by: Lucas Stach <[email protected]>
2024-04-25drm/etnaviv: fix tx clock gating on some GC7000 variantsDerek Foreman1-2/+2
commit 4bce244272513 ("drm/etnaviv: disable tx clock gating for GC7000 rev6203") accidentally applied the fix for i.MX8MN errata ERR050226 to GC2000 instead of GC7000, failing to disable tx clock gating for GC7000 rev 0x6023 as intended. Additional clean-up further propagated this issue, partially breaking the clock gating fixes added for GC7000 rev 6202 in commit 432f51e7deeda ("drm/etnaviv: add clock gating workaround for GC7000 r6202"). Signed-off-by: Derek Foreman <[email protected]> Signed-off-by: Lucas Stach <[email protected]>
2024-03-08Merge tag 'drm-etnaviv-next-2024-03-07' of ↵Dave Airlie8-44/+158
https://git.pengutronix.de/git/lst/linux into drm-next - various code cleanups - enhancements for NPU and MRT support Signed-off-by: Dave Airlie <[email protected]> From: Lucas Stach <[email protected]> Link: https://patchwork.freedesktop.org/patch/msgid/[email protected]
2024-03-07drm/etnaviv: Restore some id valuesChristian Gmeiner2-1/+10
The hwdb selection logic as a feature that allows it to mark some fields as 'don't care'. If we match with such a field we memcpy(..) the current etnaviv_chip_identity into ident. This step can overwrite some id values read from the GPU with the 'don't care' value. Fix this issue by restoring the affected values after the memcpy(..). As this is crucial for user space to know when this feature works as expected increment the minor version too. Fixes: 4078a1186dd3 ("drm/etnaviv: update hwdb selection logic") Cc: [email protected] Signed-off-by: Christian Gmeiner <[email protected]> Reviewed-by: Tomeu Vizoso <[email protected]> Signed-off-by: Lucas Stach <[email protected]>
2024-01-23drm/etnaviv: add sensitive state for PE_RT_ADDR_4_PIPE(3, 0|1) addressChristian Gmeiner1-0/+1
0x1540 is the address of 4th render target address pair (two pixel pipes). Signed-off-by: Christian Gmeiner <[email protected]> Signed-off-by: Lucas Stach <[email protected]>
2024-01-23drm/etnaviv: Expose a few more chipspecs to userspaceTomeu Vizoso3-0/+66
These ones will be needed to make use fo the NN and TP units in the NPUs based on Vivante IP. Also fix the number of NN cores in the VIPNano-qi. Signed-off-by: Tomeu Vizoso <[email protected]> Acked-by: Christian Gmeiner <[email protected]> Signed-off-by: Lucas Stach <[email protected]>
2024-01-23drm/etnaviv: disable MLCG and pulse eater on GPU resetLucas Stach1-1/+12
Module level clock gating and the pulse eater might interfere with the GPU reset, as they both have the potential to stop the clock and thus reset propagation to parts of the GPU. Signed-off-by: Lucas Stach <[email protected]> Reviewed-by: Christian Gmeiner <[email protected]>
2024-01-23drm/etnaviv: Drop the 'len' parameter of etnaviv_iommu_map() functionSui Jingfeng1-2/+2
The 'len' parameter is the 4th argument, because it is not get used, so drop it. No functional change. Signed-off-by: Sui Jingfeng <[email protected]> Signed-off-by: Lucas Stach <[email protected]>
2024-01-23drm/etnaviv: Clean up etnaviv_gem_get_pagesSui Jingfeng1-3/+2
As in the etnaviv_gem_get_pages() function, the point to the drm_device has already been cached to the 'dev' local variable. We can use it directly, While at it, using 'unsigned int' type to count the number of pages. As the drm_prime_pages_to_sg() function takes an unsigned int type for its third argument. No functional change. Signed-off-by: Sui Jingfeng <[email protected]> [lst: Reword subject to make more generic and match patch content] Signed-off-by: Lucas Stach <[email protected]>
2024-01-23drm/etnaviv: Add a helper to get the first available GPU device nodeSui Jingfeng1-9/+17
This patch make the code in the etnaviv_pdev_probe() less twisted, and it also make it easier to drop the reference to device node after finished. Before apply this patch, there is no call to of_node_put() when done. We should call of_node_put() when done. Signed-off-by: Sui Jingfeng <[email protected]> Signed-off-by: Lucas Stach <[email protected]>
2023-11-21drm/etnaviv: Convert to platform remove callback returning voidUwe Kleine-König2-7/+4
The .remove() callback for a platform driver returns an int which makes many driver authors wrongly assume it's possible to do error handling by returning an error code. However the value returned is (mostly) ignored and this typically results in resource leaks. To improve here there is a quest to make the remove callback return void. In the first step of this quest all drivers are converted to .remove_new() which already returns void. Trivially convert the etnaviv drm driver from always returning zero in the remove callback to the void returning variant. Reviewed-by: Thomas Zimmermann <[email protected]> Reviewed-by: Jyri Sarha <[email protected]> Signed-off-by: Uwe Kleine-König <[email protected]> Reviewed-by: Christian Gmeiner <[email protected]> Signed-off-by: Thomas Zimmermann <[email protected]> Link: https://patchwork.freedesktop.org/patch/msgid/[email protected]
2023-11-10drm/sched: implement dynamic job-flow controlDanilo Krummrich2-2/+2
Currently, job flow control is implemented simply by limiting the number of jobs in flight. Therefore, a scheduler is initialized with a credit limit that corresponds to the number of jobs which can be sent to the hardware. This implies that for each job, drivers need to account for the maximum job size possible in order to not overflow the ring buffer. However, there are drivers, such as Nouveau, where the job size has a rather large range. For such drivers it can easily happen that job submissions not even filling the ring by 1% can block subsequent submissions, which, in the worst case, can lead to the ring run dry. In order to overcome this issue, allow for tracking the actual job size instead of the number of jobs. Therefore, add a field to track a job's credit count, which represents the number of credits a job contributes to the scheduler's credit limit. Signed-off-by: Danilo Krummrich <[email protected]> Reviewed-by: Luben Tuikov <[email protected]> Link: https://patchwork.freedesktop.org/patch/msgid/[email protected]
2023-11-01drm/sched: Convert drm scheduler to use a work queue rather than kthreadMatthew Brost1-1/+1
In Xe, the new Intel GPU driver, a choice has made to have a 1 to 1 mapping between a drm_gpu_scheduler and drm_sched_entity. At first this seems a bit odd but let us explain the reasoning below. 1. In Xe the submission order from multiple drm_sched_entity is not guaranteed to be the same completion even if targeting the same hardware engine. This is because in Xe we have a firmware scheduler, the GuC, which allowed to reorder, timeslice, and preempt submissions. If a using shared drm_gpu_scheduler across multiple drm_sched_entity, the TDR falls apart as the TDR expects submission order == completion order. Using a dedicated drm_gpu_scheduler per drm_sched_entity solve this problem. 2. In Xe submissions are done via programming a ring buffer (circular buffer), a drm_gpu_scheduler provides a limit on number of jobs, if the limit of number jobs is set to RING_SIZE / MAX_SIZE_PER_JOB we get flow control on the ring for free. A problem with this design is currently a drm_gpu_scheduler uses a kthread for submission / job cleanup. This doesn't scale if a large number of drm_gpu_scheduler are used. To work around the scaling issue, use a worker rather than kthread for submission / job cleanup. v2: - (Rob Clark) Fix msm build - Pass in run work queue v3: - (Boris) don't have loop in worker v4: - (Tvrtko) break out submit ready, stop, start helpers into own patch v5: - (Boris) default to ordered work queue v6: - (Luben / checkpatch) fix alignment in msm_ringbuffer.c - (Luben) s/drm_sched_submit_queue/drm_sched_wqueue_enqueue - (Luben) Update comment for drm_sched_wqueue_enqueue - (Luben) Positive check for submit_wq in drm_sched_init - (Luben) s/alloc_submit_wq/own_submit_wq v7: - (Luben) s/drm_sched_wqueue_enqueue/drm_sched_run_job_queue v8: - (Luben) Adjust var names / comments Signed-off-by: Matthew Brost <[email protected]> Reviewed-by: Luben Tuikov <[email protected]> Link: https://lore.kernel.org/r/[email protected] Signed-off-by: Luben Tuikov <[email protected]>
2023-10-26drm/sched: Convert the GPU scheduler to variable number of run-queuesLuben Tuikov1-0/+1
The GPU scheduler has now a variable number of run-queues, which are set up at drm_sched_init() time. This way, each driver announces how many run-queues it requires (supports) per each GPU scheduler it creates. Note, that run-queues correspond to scheduler "priorities", thus if the number of run-queues is set to 1 at drm_sched_init(), then that scheduler supports a single run-queue, i.e. single "priority". If a driver further sets a single entity per run-queue, then this creates a 1-to-1 correspondence between a scheduler and a scheduled entity. Cc: Lucas Stach <[email protected]> Cc: Russell King <[email protected]> Cc: Qiang Yu <[email protected]> Cc: Rob Clark <[email protected]> Cc: Abhinav Kumar <[email protected]> Cc: Dmitry Baryshkov <[email protected]> Cc: Danilo Krummrich <[email protected]> Cc: Matthew Brost <[email protected]> Cc: Boris Brezillon <[email protected]> Cc: Alex Deucher <[email protected]> Cc: Christian König <[email protected]> Cc: Emma Anholt <[email protected]> Cc: [email protected] Cc: [email protected] Cc: [email protected] Cc: [email protected] Cc: [email protected] Cc: [email protected] Signed-off-by: Luben Tuikov <[email protected]> Acked-by: Christian König <[email protected]> Link: https://lore.kernel.org/r/[email protected]
2023-10-11drm/etnaviv: Add helper functions to create and destroy platform deviceSui Jingfeng1-16/+37
The newly introduced functions are etnaviv_create_platform_device() and etnaviv_destroy_platform_device(). Those two function are pure function and can be shared for other use case. Currently, the benefit is that we no longer need to call of_node_put() for three different cases, we only need to call it once in the etnaviv_init() function. Signed-off-by: Sui Jingfeng <[email protected]> Signed-off-by: Lucas Stach <[email protected]>
2023-10-11drm/etnaviv: Fix coding styleSui Jingfeng1-6/+6
1) Keep the curly brace aligned. 2) No indentation by double tabs where single tab indentation is enough. Signed-off-by: Sui Jingfeng <[email protected]> Signed-off-by: Lucas Stach <[email protected]>
2023-10-11drm/etnaviv: Drop the second argument of the etnaviv_gem_new_impl()Sui Jingfeng1-4/+3
The mentioned second parameter is the 'u32 size', but it is not get used by the etnaviv_gem_new_impl() function, so drop it. No functional change. Signed-off-by: Sui Jingfeng <[email protected]> Signed-off-by: Lucas Stach <[email protected]>
2023-10-11drm/etnaviv: Replace strncpy with strscpy_padJustin Stitt1-2/+2
`strncpy` is deprecated for use on NUL-terminated destination strings [1]. We should prefer more robust and less ambiguous string interfaces. A suitable replacement is `strscpy_pad` due to the fact that it guarantees NUL-termination on the destination buffer whilst maintaining the NUL-padding behavior that strncpy provides. Link: https://www.kernel.org/doc/html/latest/process/deprecated.html#strncpy-on-nul-terminated-strings [1] Link: https://github.com/KSPP/linux/issues/90 Cc: [email protected] Cc: Bo YU <[email protected]> Signed-off-by: Justin Stitt <[email protected]> Reviewed-by: Kees Cook <[email protected]> Reviewed-by: Christian Gmeiner <[email protected]> [lst: changed subject according to suggestion from Kees] Signed-off-by: Lucas Stach <[email protected]>
2023-08-18Merge branch 'etnaviv/next' of https://git.pengutronix.de/git/lst/linux into ↵Dave Airlie11-107/+201
drm-next This time mostly cleanups around the runtime power management handling and slightly improved GPU hang handling. Also some additions to the HWDB to get the driver working properly on more NXP i.MX8MP IP cores. Signed-off-by: Dave Airlie <[email protected]> From: Lucas Stach <[email protected]> Link: https://patchwork.freedesktop.org/patch/msgid/[email protected]
2023-07-21drm: Explicitly include correct DT includesRob Herring1-1/+1
The DT of_device.h and of_platform.h date back to the separate of_platform_bus_type before it as merged into the regular platform bus. As part of that merge prepping Arm DT support 13 years ago, they "temporarily" include each other. They also include platform_device.h and of.h. As a result, there's a pretty much random mix of those include files used throughout the tree. In order to detangle these headers and replace the implicit includes with struct declarations, users need to explicitly include the correct includes. Signed-off-by: Rob Herring <[email protected]> Acked-by: Sam Ravnborg <[email protected]> Reviewed-by: Steven Price <[email protected]> Acked-by: Liviu Dudau <[email protected]> Reviewed-by: Kieran Bingham <[email protected]> Acked-by: Robert Foss <[email protected]> Signed-off-by: Thierry Reding <[email protected]> Link: https://patchwork.freedesktop.org/patch/msgid/[email protected]
2023-07-19drm/etnaviv: fix error code in event_alloc()Dan Carpenter1-4/+4
There are two "ret" variables declared in this function so setting "ret = -EBUSY;" sets the wrong one. The function ends up returning an uninitialized variable. Fixes: f098f9b8042a ("drm/etnaviv: move runtime PM handling to events") Signed-off-by: Dan Carpenter <[email protected]> Signed-off-by: Lucas Stach <[email protected]>
2023-07-17drm/etnaviv: expedited MMU fault handlingLucas Stach3-2/+6
The GPU is halted when it hits a MMU exception, so there is no point in waiting for the job timeout to expire or try to work out if the GPU is still making progress in the timeout handler, as we know that the GPU won't make any more progress. Signed-off-by: Lucas Stach <[email protected]> Reviewed-by: Christian Gmeiner <[email protected]>
2023-07-17drm/etnaviv: drop GPU initialized propertyLucas Stach4-8/+8
Now that it is only used to track the driver internal state of the MMU global and cmdbuf objects, we can get rid of this property by making the free/finit functions of those objects safe to call on an uninitialized object. Signed-off-by: Lucas Stach <[email protected]> Reviewed-by: Christian Gmeiner <[email protected]>
2023-07-17drm/etnaviv: better track GPU stateLucas Stach2-8/+27
Instead of only tracking if the FE is running, use a enum to better describe the various states the GPU can be in. This allows some additional validation to make sure that functions that expect a certain GPU state are only called when the GPU is actually in that state. Signed-off-by: Lucas Stach <[email protected]> Reviewed-by: Christian Gmeiner <[email protected]>
2023-07-17drm/etnaviv: avoid runtime PM usage in etnaviv_gpu_bindLucas Stach1-10/+5
Nothing in this callpath actually touches the GPU, so there is no reason to get it out of suspend state here. Only if runtime PM isn't enabled at all we must make sure to enable the clocks, so the GPU init routine can access the GPU later on. This also removes the need to guard against the state where the driver isn't fully initialized yet in the runtime PM resume handler. Signed-off-by: Lucas Stach <[email protected]> Reviewed-by: Christian Gmeiner <[email protected]>
2023-07-17drm/etnaviv: make clock handling symetric between runtime resume and suspendLucas Stach1-4/+5
Currently the clock is enabled in the runtime resume function, but are disabled a level further down in the callstack in the suspend function. Move the clock disable into the suspend function to make handling symmetrical between resume and suspend. Signed-off-by: Lucas Stach <[email protected]> Reviewed-by: Christian Gmeiner <[email protected]>
2023-07-17drm/etnaviv: move runtime PM handling to eventsLucas Stach3-15/+16
Conceptually events are the right abstraction to handle the GPU runtime PM state: as long as any event is pending the GPU can not be idle. Events are also properly freed and reallocated when the GPU has been reset after a hang. Signed-off-by: Lucas Stach <[email protected]> Reviewed-by: Christian Gmeiner <[email protected]>
2023-07-17drm/etnaviv: free events the usual way in recover workerLucas Stach1-2/+1
Clearing the whole bitmap at once is only a minor optimization in a path that should be extremely cold. Free the events by calling event_free() instead of directly manipulating the completion count and event bitmap. Signed-off-by: Lucas Stach <[email protected]> Reviewed-by: Christian Gmeiner <[email protected]>
2023-07-17drm/etnaviv: move down etnaviv_gpu_recover_hang() in fileLucas Stach1-44/+44
So it can use the event_free function without adding another forward declaration. No functional change. Signed-off-by: Lucas Stach <[email protected]> Reviewed-by: Christian Gmeiner <[email protected]>
2023-07-17drm/etnaviv: add HWDB entry for GC520 r5341 c204Lucas Stach1-0/+31
This is the 2D GPU found on the i.MX8MP SoC. Feature bits taken from the downstream kernel driver 6.4.3.p4.4. Signed-off-by: Lucas Stach <[email protected]>
2023-07-17drm/etnaviv: add HWDB entry for VIP8000 Nano r8002Lucas Stach1-0/+32
This is the NPU found on the NXP i.MX8MP SoC. Feature bits taken from the downstream kernel driver 6.4.3.p4.4. Signed-off-by: Lucas Stach <[email protected]>
2023-07-17drm/etnaviv: fix dumping of active MMU contextLucas Stach1-7/+7
gpu->mmu_context is the MMU context of the last job in the HW queue, which isn't necessarily the same as the context from the bad job. Dump the MMU context from the scheduler determined bad submit to make it work as intended. Fixes: 17e4660ae3d7 ("drm/etnaviv: implement per-process address spaces on MMUv2") Signed-off-by: Lucas Stach <[email protected]> Reviewed-by: Christian Gmeiner <[email protected]>
2023-06-26drm: Clear fd/handle callbacks in struct drm_driverThomas Zimmermann1-2/+0
Clear all assignments of struct drm_driver's fd/handle callbacks to drm_gem_prime_fd_to_handle() and drm_gem_prime_handle_to_fd(). These functions are called by default. Add a TODO item to convert vmwgfx to the defaults as well. v2: * remove TODO item (Zack) * also update amdgpu's amdgpu_partition_driver Signed-off-by: Thomas Zimmermann <[email protected]> Reviewed-by: Simon Ser <[email protected]> Acked-by: Alex Deucher <[email protected]> Acked-by: Jeffrey Hugo <[email protected]> # qaic Link: https://patchwork.freedesktop.org/patch/msgid/[email protected]
2023-06-21drm/etnaviv: slow down FE idle pollingLucas Stach3-5/+15
Currently the FE is spinning way too fast when polling for new work in the FE idleloop. As each poll fetches 16 bytes from memory, a GPU running at 1GHz with the current setting of 200 wait cycle between fetches causes 80 MB/s of memory traffic just to check for new work when the GPU is otherwise idle, which is more FE traffic than in some GPU loaded cases. Significantly increase the number of wait cycles to slow down the poll interval to ~30µs, limiting the FE idle memory traffic to 512 KB/s, while providing a max latency which should not hurt most use-cases. The FE WAIT command seems to have some unknown discrete steps in the wait cycles so we may over/undershoot the target a bit, but that should be harmless. If the GPU core base frequency is unknown keep the 200 wait cycles as a sane default. Signed-off-by: Lucas Stach <[email protected]> Reviewed-by: Sui Jingfeng <[email protected]> Tested-by: Sui Jingfeng <[email protected]> Reviewed-by: Christian Gmeiner <[email protected]>
2023-06-21drm: etnaviv: Replace of_platform.h with explicit includesRob Herring1-1/+3
Etnaviv doesn't use anything from of_platform.h, but depends on of.h, of_device.h, and platform_device.h which are all implicitly included, but that is going to be removed soon. Signed-off-by: Rob Herring <[email protected]> Tested-by: Sui Jingfeng <[email protected]> Signed-off-by: Lucas Stach <[email protected]>
2023-06-19drm: Remove struct drm_driver.gem_prime_mmapThomas Zimmermann1-1/+0
All drivers initialize this field with drm_gem_prime_mmap(). Call the function directly and remove the field. Simplifies the code and resolves a long-standing TODO item. Signed-off-by: Thomas Zimmermann <[email protected]> Reviewed-by: Alex Deucher <[email protected]> Link: https://patchwork.freedesktop.org/patch/msgid/[email protected]
2023-03-30Revert "drm/etnaviv: export client GPU usage statistics via fdinfo"Lucas Stach1-42/+1
This reverts commit 97804a133c68, as it builds on top of df622729ddbf ("drm/scheduler: track GPU active time per entity") which needs to be reverted, as it introduces a use-after-free. Signed-off-by: Lucas Stach <[email protected]>
2023-03-30drm/etnaviv: fix reference leak when mmaping imported bufferLucas Stach1-1/+9
drm_gem_prime_mmap() takes a reference on the GEM object, but before that drm_gem_mmap_obj() already takes a reference, which will be leaked as only one reference is dropped when the mapping is closed. Drop the extra reference when dma_buf_mmap() succeeds. Cc: [email protected] Signed-off-by: Lucas Stach <[email protected]> Reviewed-by: Christian Gmeiner <[email protected]>
Powered by cgit, Gruvboxed by Blaster4385.