- 07 Sep, 2020 40 commits
-
-
Maarten Lankhorst authored
This is required if we want to pass a ww context in intel_context_pin and gen6_ppgtt_pin(). Signed-off-by:
Maarten Lankhorst <maarten.lankhorst@linux.intel.com> Reviewed-by:
Thomas Hellström <thomas.hellstrom@intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20200819140904.1708856-11-maarten.lankhorst@linux.intel.comSigned-off-by:
Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
-
Maarten Lankhorst authored
We want to start using ww locking in intel_context_pin, for this we need to lock multiple objects, and the single i915_gem_object_lock is not enough. Convert to using ww-waiting, and make sure we always pin intel_context_state, even if we don't have a renderstate object. Signed-off-by:
-
Maarten Lankhorst authored
Now that we changed execbuf submission slightly to allow us to do all pinning in one place, we can now simply add ww versions on top of struct_mutex. All we have to do is a separate path for -EDEADLK handling, which needs to unpin all gem bo's before dropping the lock, then starting over. This finally allows us to do parallel submission, but because not all of the pinning code uses the ww ctx yet, we cannot completely drop struct_mutex yet. Changes since v1: - Keep struct_mutex for now. :( Changes since v2: - Make sure we always lock the ww context in slowpath. Changes since v3: - Don't call __eb_unreserve_vma in eb_move_to_gpu now; this can be done on normal unlock path. - Unconditionally release vmas and context. Changes since v4: - Rebased on top of struct_mutex reduction. Changes since v5: - Remove training wheels. Changes since v6: - Fix accidentally broken -ENOSPC handling. Changes since v7: - Handle gt buffer pool better. Changes since v8: - Properly clear variables, to make -EDEADLK handling not BUG. Change since v9: - Fix unpinning fence on pnv and below. Changes since v10: - Make relocation gpu chaining working again. Changes since v11: - Remove relocation chaining, pain to make it work. Signed-off-by:
-
Maarten Lankhorst authored
We want to introduce backoff logic, but we need to lock the pool object as well for command parsing. Because of this, we will need backoff logic for the engine pool obj, move the batch validation up slightly to eb_lookup_vmas, and the actual command parsing in a separate function which can get called from execbuf relocation fast and slowpath. Signed-off-by:
-
Maarten Lankhorst authored
Execbuffer submission will perform its own WW locking, and we cannot rely on the implicit lock there. This also makes it clear that the GVT code will get a lockdep splat when multiple batchbuffer shadows need to be performed in the same instance, fix that up. Signed-off-by:
Tvrtko Ursulin <tvrtko.ursulin@intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20200819140904.1708856-7-maarten.lankhorst@linux.intel.comSigned-off-by:
-
Maarten Lankhorst authored
i915_gem_ww_ctx is used to lock all gem bo's for pinning and memory eviction. We don't use it yet, but lets start adding the definition first. To use it, we have to pass a non-NULL ww to gem_object_lock, and don't unlock directly. It is done in i915_gem_ww_ctx_fini. Changes since v1: - Change ww_ctx and obj order in locking functions (Jonas Lahtinen) Signed-off-by:
-
Maarten Lankhorst authored
This reverts commit 0f1dd022 ("drm/i915/gem: Split eb_vma into its own allocation") and also moves all unreserving to a single place at the end, which is a minor simplification. With the WW locking, we will drop all references only at the end when unlocking, so refcounting can now be removed. Signed-off-by:
Daniel Vetter <daniel.vetter@ffwll.ch> Link: https://patchwork.freedesktop.org/patch/msgid/20200819140904.1708856-5-maarten.lankhorst@linux.intel.comSigned-off-by:
-
Maarten Lankhorst authored
This reverts commit 7dc8f114 ("drm/i915/gem: Drop relocation slowpath"). We need the slowpath relocation for taking ww-mutex inside the page fault handler, and we will take this mutex when pinning all objects. We also functionally revert ef398881 ("drm/i915/gem: Limit struct_mutex to eb_reserve"), as we need the struct_mutex in the slowpath as well, and a tiny part of 003d8b91 ("drm/i915/gem: Only call eb_lookup_vma once during execbuf ioctl"). Specifically, we make the -EAGAIN handling part of fallback to slowpath again. With this, we have a proper working slowpath again, which will allow us to do fault handling with WW locks held. [mlankhorst: Adjusted for reloc_gpu_flush() changes] Cc: Chris Wilson <chris@chris-wilson.co.uk> Cc: Matthew Auld <matthew.auld@intel.com> Signed-off-by:
-
Maarten Lankhorst authored
This reverts commit 964a9b0f ("drm/i915/gem: Use chained reloc batches") and commit 0e97fbb0 ("drm/i915/gem: Use a single chained reloc batches for a single execbuf"). When adding ww locking to execbuf, it's hard enough to deal with a single BO that is part of relocation execution. Chaining is hard to get right, and with GPU relocation deprecated, it's best to drop this altogether, instead of trying to fix something we will remove. This is not a completely 1:1 revert, we reset rq_size to 0 in reloc_cache_init, this was from e3d29130 ("drm/i915/gem: Implement legacy MI_STORE_DATA_IMM"), because we don't want to break the selftests. (Daniel) Signed-off-by:
-
Maarten Lankhorst authored
This reverts commit 9e0f9464 ("drm/i915/gem: Async GPU relocations only"), and related commit 7ac2d253 ("drm/i915/gem: Delete unused code"). Async GPU relocations are not the path forward, we want to remove GPU accelerated relocation support eventually when userspace is fixed to use VM_BIND, and this is the first step towards that. We will keep async gpu relocations around for now, until userspace is fixed. Relocation support will be disabled completely on platforms where there was never any userspace that depends on it, as the hardware doesn't require it from at least gen9+ onward. For older platforms, the plan is to use cpu relocations only. The igt side is fixed in igt commit 39e9aa1032a4e ("tests/i915: Remove subtests that rely on async relocation behavior"). Signed-off-by:
-
Chris Wilson authored
If dma_fence_chain_find_seqno() reports an error, it does so in its preamble before it disposes of the input fence. On handling the error, we need to drop the reference to the fence. Closes: https://gitlab.freedesktop.org/drm/intel/-/issues/2292Signed-off-by:
Chris Wilson <chris@chris-wilson.co.uk> Cc: Lionel Landwerlin <lionel.g.landwerlin@intel.com> Fixes: 13149e8b ("drm/i915: add syncobj timeline support") Reviewed-by:
Lionel Landwerlin <lionel.g.landwerlin@intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20200806161056.17593-1-chris@chris-wilson.co.ukSigned-off-by:
Rodrigo Vivi <rodrigo.vivi@intel.com> Signed-off-by:
-
Chris Wilson authored
As we now protect the timeline list using RCU, we can drop the timeline->mutex for guarding the list iteration during context close, as we are searching for an inflight request. Any new request will see the context is banned and not be submitted. In doing so, pull the checks for a concurrent submission of the request (notably the i915_request_completed()) under the engine spinlock, to fully serialise with __i915_request_submit()). That is in the case of preempt-to-busy where the request may be completed during the __i915_request_submit(), we need to be careful that we sample the request status after serialising so that we don't miss the request the engine is actually submitting. Fixes: 4a317415 ("drm/i915/gem: Refine occupancy test in kill_context()") References: d22d2d07 ("drm/i915: Protect i915_request_await_start from early waits") # rcu protection of timeline->requests References: https://gitlab.freedesktop.org/drm/intel/-/issues/1622 References: https://gitlab.freedesktop.org/drm/intel/-/issues/2158Signed-off-by:
-
Chris Wilson authored
When igt_random_offset() is a given a range of [0, PAGE_SIZE], it is allowed to return 0. However, attempting to use a size of 0 for the igt_lmem_write_cpu() byte poking, leads to call igt_random_offset() with a range of [offset, offset + 0] and ask it to find a length of 4 within it. This triggers the bug on that the requested length should fit within the range! Signed-off-by:
Matthew Auld <matthew.auld@intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20200806145728.16495-1-chris@chris-wilson.co.ukSigned-off-by:
-
Chris Wilson authored
Currently we hold no actual reference to the request nor context while they are attached to a breadcrumb. To avoid freeing the request/context too early, we serialise with cancel-breadcrumbs by taking the irq spinlock in i915_request_retire(). The alternative is to take a reference for a new breadcrumb and release it upon signaling; removing the more frequently hit contention point in i915_request_retire(). Signed-off-by:
-
Chris Wilson authored
Move the __intel_breadcrumbs_arm_irq earlier, next to the disarm_irq, so that we can make use of it in the following patch. Signed-off-by:
-
Chris Wilson authored
kmalloc uses power-of-two slab buckets for small allocations (up to a few pages). Since i915_page_directory is a page of pointers, plus a couple more, this is rounded up to 8K, and we waste nearly 50% of that allocation. Long terms this leads to poor memory utilisation, bloating the kernel footprint, but the problem is exacerbated by our conservative preallocation scheme for binding VMA. As we are required to allocate all levels for each vma just in case we need to insert them upon binding, this leads to a large multiplication factor for a single page vma. By halving the allocation we need for the page directory structure, we halve the impact of that factor, bringing workloads that once fitted into memory, hopefully back to fitting into memory. We maintain the split between i915_page_directory and i915_page_table as we only need half the allocation for the lowest, most populous, level. Signed-off-by:
-
Chris Wilson authored
The GEM object is grossly overweight for the practicality of tracking large numbers of individual pages, yet it is currently our only abstraction for tracking DMA allocations. Since those allocations need to be reserved upfront before an operation, and that we need to break away from simple system memory, we need to ditch using plain struct page wrappers. In the process, we drop the WC mapping as we ended up clflushing everything anyway due to various issues across a wider range of platforms. Though in a future step, we need to drop the kmap_atomic approach which suggests we need to pre-map all the pages and keep them mapped. v2: Verify our large scratch page is suitably DMA aligned; and manually clear the scratch since we are allocating plain struct pages full of prior content. Signed-off-by:
-
Chris Wilson authored
We need to make the DMA allocations used for page directories to be performed up front so that we can include those allocations in our memory reservation pass. The downside is that we have to assume the worst case, even before we know the final layout, and always allocate enough page directories for this object, even when there will be overlap. This unfortunately can be quite expensive, especially as we have to clear/reset the page directories and DMA pages, but it should only be required during early phases of a workload when new objects are being discovered, or after memory/eviction pressure when we need to rebind. Once we reach steady state, the objects should not be moved and we no longer need to preallocating the pages tables. It should be noted that the lifetime for the page directories DMA is more or less decoupled from individual fences as they will be shared across objects across timelines. v2: Only allocate enough PD space for the PTE we may use, we do not need to allocate PD that will be left as scratch. v3: Store the shift unto the first PD level to encapsulate the different PTE counts for gen6/gen8. Signed-off-by:
-
Chris Wilson authored
On the virtual engines, we only use the intel_breadcrumbs for tracking signaling of stale breadcrumbs from the irq_workers. They do not have any associated interrupt handling, active requests are passed to a physical engine and associated breadcrumb interrupt handler. This causes issues for us as we need to ensure that we do not actually try and enable interrupts and the powermanagement required for them on the virtual engine, as they will never be disabled. Instead, let's specify the physical engine used for interrupt handler on a particular breadcrumb. v2: Drop b->irq_armed = true mocking for no interrupt HW Fixes: 4fe6abb8 ("drm/i915/gt: Ignore irq enabling on the virtual engines") Signed-off-by:
-
Chris Wilson authored
One more complication of preempt-to-busy with respect to the virtual engine is that we may have retired the last request along the virtual engine at the same time as preparing to submit the completed request to a new engine. That submit will be shortcircuited, but not before we have updated the context with the new register offsets and marked the virtual engine as bound to the new engine (by calling swap on ve->siblings[]). As we may have just retired the completed request, we may also be in the middle of calling virtual_context_exit() to turn off the power management associated with the virtual engine, and that in turn walks the ve->siblings[]. If we happen to call swap() on the array as we walk, we will call intel_engine_pm_put() twice on the same engine. In this patch, we prevent this by only updating the bound engine after a successful submission which weeds out the already completed requests. Alternatively, we could walk a non-volatile array for the pm, such as using the engine->mask. The small advantage to performing the update after the submit is that we then only have to do a swap for active requests. Fixes: 22b7a426 ("drm/i915/execlists: Preempt-to-busy") References: 6d06779e ("drm/i915: Load balancing across a virtual engine" Signed-off-by:
-
Chris Wilson authored
After staring at the breadcrumb enabling/cancellation and coming to the conclusion that the cause of the mysterious stale breadcrumbs must the act of submitting a completed requests, we can then redirect those completed requests onto a dedicated signaled_list at the time of construction and so eliminate intel_engine_transfer_stale_breadcrumbs(). Signed-off-by:
-
Chris Wilson authored
Since the breadcrumb enabling/cancelling itself is serialised by the breadcrumbs.irq_lock, with a bit of care we can remove the outer serialisation with i915_request.lock for concurrent dma_fence_enable_signaling(). This has the important side-effect of eliminating the nested i915_request.lock within request submission. The challenge in serialisation is around the unsubmission where we take an active request that wants a breadcrumb on the signaling engine and put it to sleep. We do not want a concurrent dma_fence_enable_signaling() to attach a breadcrumb as we unsubmit, so we must mark the request as no longer active before serialising with the concurrent enable-signaling. On retire, we serialise with the concurrent enable-signaling, but instead of clearing ACTIVE, we mark it as SIGNALED. Signed-off-by:
-
Chris Wilson authored
Before we can execute a request, we must wait for all of its vma to be bound. This is a frequent operation for which we can optimise away a few atomic operations (notably a cmpxchg) in lieu of the RCU protection. Signed-off-by:
-
Chris Wilson authored
As the conversion between idle-barrier and full i915_active_fence is already serialised by explicit memory barriers, we can reduce the spinlock in i915_active_acquire_preallocate_barrier() for finding an idle-barrier to reuse to an RCU read lock to ensure the fence remains valid, only taking the spinlock for the update of the rbtree itself. Signed-off-by:
-
Chris Wilson authored
Rather than require the next timeline after idling to match the MRU before idling, reset the index on the node and allow it to match the first request. However, this requires cmpxchg(u64) and so is not trivial on 32b, so for compatibility we just fallback to keeping the cached node pointing to the MRU timeline. Signed-off-by:
-
Chris Wilson authored
Whenever an i915_active idles, we prune its tree of old fence slots to prevent a gradual leak should it be used to track many, many timelines. The downside is that we then have to frequently reallocate the rbtree. A compromise is that we keep the most recently used fence slot, and reuse that for the next active reference as that is the most likely timeline to be reused. Signed-off-by:
-
Chris Wilson authored
Sometimes we have to be very careful not to allocate underneath a mutex (or spinlock) and yet still want to track activity. Enter i915_active_acquire_for_context(). This raises the activity counter on i915_active prior to use and ensures that the fence-tree contains a slot for the context. v2: Refactor active_lookup() so it can be called again before/after locking to resolve contention. Since we protect the rbtree until we idle, we can do a lockfree lookup, with the caveat that if another thread performs a concurrent insertion, the rotations from the insert may cause us to not find our target. A second pass holding the treelock will find the target if it exists, or the place to perform our insertion. Signed-off-by:
-
Chris Wilson authored
If no active callback is defined for i915_active, we do not need to serialise its enabling with the mutex. We still do only want to call the debug activate once, and must still serialise with a concurrent retire. Signed-off-by:
-
Chris Wilson authored
Since we pass around encoded parameters to the kernel context constructor using the ce->timeline pointer, we can no longer assert that it should be zero for mock timeline construction. Fixes: d1bf5dd8 ("drm/i915/gt: Support multiple pinned timelines") Signed-off-by:
-
Chris Wilson authored
We need to ensure that the list is valid prior to marking the node as retrievable, otherwise we may see two threads compete over the same node in intel_gt_get_buffer_pool(). If the first thread acquires and releases the node in the same jiffie, the second thread may then acquire it (as the jiffie now again matches the expected value) and claim the node before it is put back into the list. Signed-off-by:
-
Chris Wilson authored
We may need to allocate more than one pinned context/timeline for each engine which can utilise the per-engine HWSP, so we need to give each a different offset within it. Signed-off-by:
-
Chris Wilson authored
Avoid exposing a partially constructed context by deferring the list_add() from the initial construction to the end of registration. Otherwise, if we peek into the list of contexts from inside debugfs, we may see the partially constructed context and chase down some dangling incomplete pointers. Reported-by:
CQ Tang <cq.tang@intel.com> Fixes: 3aa9945a ("drm/i915: Separate GEM context construction and registration to userspace") References: f6e8aa38 ("drm/i915: Report the number of closed vma held by each context in debugfs") Signed-off-by:
Mika Kuoppala <mika.kuoppala@linux.intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20200730092856.23615-1-chris@chris-wilson.co.ukSigned-off-by:
-
Chris Wilson authored
A last minute change, that unfortunately broke CI so badly it declared SUCCESS, was to refactor the debug free all buffer pool code to reuse the normal worker, inverted the termination condition so that it instead of discarding the nodes, they were all declared young enough and eligible for reuse. Fixes: 06b73c2d ("drm/i915/gt: Delay taking the spinlock for grabbing from the buffer pool") Signed-off-by:
-
Chris Wilson authored
Before we peek at the barrier status for an assert, first serialise with its callbacks so that we see a stable value. Signed-off-by:
-
Chris Wilson authored
Some very low hanging fruit, but contention on the pool->lock is noticeable between intel_gt_get_buffer_pool() and pool_retire(), with the majority of the hold time due to the locked list iteration. If we make the node itself RCU protected, we can perform the search for an suitable node just under RCU, reserving taking the lock itself for claiming the node and manipulating the list. Signed-off-by:
-
Chris Wilson authored
Unlike rcs where we have conclusive evidence from our selftesting that disabling the preparser before performing the TLB invalidate and relocations does impact upon the GPU execution, the evidence for the same requirement on xcs is much more circumstantial. Let's apply the preparser disable between batches as we invalidate the TLB as a dose of healthy paranoia, just in case. References: https://gitlab.freedesktop.org/drm/intel/-/issues/2169Signed-off-by:
-
Chris Wilson authored
I915_GEM_THROTTLE dates back to the time before contexts where there was just a single engine, and therefore a single timeline and request list globally. That request list was in execution/retirement order, and so walking it to find a particular aged request made sense and could be split per file. That is no more. We now have many timelines with a file, as many as the user wants to construct (essentially per-engine, per-context). Each of those run independently and so make the single list futile. Remove the disordered list, and iterate over all the timelines to find a request to wait on in each to satisfy the criteria that the CPU is no more than 20ms ahead of its oldest request. It should go without saying that the I915_GEM_THROTTLE ioctl is no longer used as the primary means of throttling, so it makes sense to push the complication into the ioctl where it only impacts upon its few irregular users, rather than the execbuf/retire where everybody has to pay the cost. Fortunately, the few users do not create vast amount of contexts, so the loops over contexts/engines should be concise. Signed-off-by:
-
Chris Wilson authored
We include a tasklet flush before waiting on a request as a precaution against the HW being lax in event signaling. We now have a precautionary flush in the engine's heartbeat and so do not need to be quite so zealous on every request wait. If we focus on the request, the only tasklet flush that matters is if there is a delay in submitting this request to HW, so if the request is not ready to be executed, no advantage in reducing this wait can be gained by running the tasklet. And there is little point in doing busy work for no result. Signed-off-by:
-
Chris Wilson authored
Since we assert that the kernel_context is using the perma-pinned HWSP, make it so. Closes: https://gitlab.freedesktop.org/drm/intel/-/issues/2179Signed-off-by:
-
Chris Wilson authored
Currently, we use i915_request_completed() directly in i915_request_wait() and follow up with a manual invocation of dma_fence_signal(). This appears to cause a large number of contentions on i915_request.lock as when the process is woken up after the fence is signaled by an interrupt, we will then try and call dma_fence_signal() ourselves while the signaler is still holding the lock. dma_fence_is_signaled() has the benefit of checking the DMA_FENCE_FLAG_SIGNALED_BIT prior to calling dma_fence_signal() and so avoids most of that contention. Signed-off-by:
-
