| Age | Commit message (Collapse) | Author | Files | Lines |
|---|
|
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 <[email protected]>
Reviewed-by: Thomas Hellström <[email protected]>
Link: https://patchwork.freedesktop.org/patch/msgid/[email protected]
Signed-off-by: Joonas Lahtinen <[email protected]>
|
|
This reverts commit 0f1dd02295f3 ("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: Maarten Lankhorst <[email protected]>
Reviewed-by: Daniel Vetter <[email protected]>
Link: https://patchwork.freedesktop.org/patch/msgid/[email protected]
Signed-off-by: Joonas Lahtinen <[email protected]>
|
|
This reverts commit 7dc8f1143778 ("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 ef398881d27d ("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 003d8b9143a6 ("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 <[email protected]>
Cc: Matthew Auld <[email protected]>
Signed-off-by: Maarten Lankhorst <[email protected]>
[mlankhorst: Removed extra reloc_gpu_flush()]
Reviewed-by: Daniel Vetter <[email protected]>
Link: https://patchwork.freedesktop.org/patch/msgid/[email protected]
Signed-off-by: Joonas Lahtinen <[email protected]>
|
|
This reverts commit 964a9b0f611ee ("drm/i915/gem: Use chained reloc batches")
and commit 0e97fbb080553 ("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 e3d291301f99 ("drm/i915/gem: Implement legacy
MI_STORE_DATA_IMM"), because we don't want to break the selftests. (Daniel)
Signed-off-by: Maarten Lankhorst <[email protected]>
Reviewed-by: Daniel Vetter <[email protected]>
Link: https://patchwork.freedesktop.org/patch/msgid/[email protected]
Signed-off-by: Joonas Lahtinen <[email protected]>
|
|
This reverts commit 9e0f9464e2ab ("drm/i915/gem: Async GPU relocations only"),
and related commit 7ac2d2536dfa7 ("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: Maarten Lankhorst <[email protected]>
Reviewed-by: Daniel Vetter <[email protected]>
Link: https://patchwork.freedesktop.org/patch/msgid/[email protected]
Signed-off-by: Joonas Lahtinen <[email protected]>
|
|
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/2292
Signed-off-by: Chris Wilson <[email protected]>
Cc: Lionel Landwerlin <[email protected]>
Fixes: 13149e8bafc4 ("drm/i915: add syncobj timeline support")
Reviewed-by: Lionel Landwerlin <[email protected]>
Link: https://patchwork.freedesktop.org/patch/msgid/[email protected]
Signed-off-by: Rodrigo Vivi <[email protected]>
Signed-off-by: Joonas Lahtinen <[email protected]>
|
|
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: 4a3174152147 ("drm/i915/gem: Refine occupancy test in kill_context()")
References: d22d2d073ef8 ("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/2158
Signed-off-by: Chris Wilson <[email protected]>
Reviewed-by: Tvrtko Ursulin <[email protected]>
Link: https://patchwork.freedesktop.org/patch/msgid/[email protected]
Signed-off-by: Rodrigo Vivi <[email protected]>
Signed-off-by: Joonas Lahtinen <[email protected]>
|
|
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: Chris Wilson <[email protected]>
Cc: Matthew Auld <[email protected]>
Reviewed-by: Matthew Auld <[email protected]>
Link: https://patchwork.freedesktop.org/patch/msgid/[email protected]
Signed-off-by: Rodrigo Vivi <[email protected]>
Signed-off-by: Joonas Lahtinen <[email protected]>
|
|
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 <[email protected]>
Reviewed-by: Tvrtko Ursulin <[email protected]>
Link: https://patchwork.freedesktop.org/patch/msgid/[email protected]
Signed-off-by: Rodrigo Vivi <[email protected]>
[Joonas: Rebased and reordered into drm-intel-gt-next branch]
Signed-off-by: Joonas Lahtinen <[email protected]>
|
|
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 <[email protected]>
Reviewed-by: Tvrtko Ursulin <[email protected]>
Link: https://patchwork.freedesktop.org/patch/msgid/[email protected]
Signed-off-by: Rodrigo Vivi <[email protected]>
Signed-off-by: Joonas Lahtinen <[email protected]>
|
|
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 <[email protected]>
Cc: Mika Kuoppala <[email protected]>
Cc: Matthew Auld <[email protected]>
Reviewed-by: Matthew Auld <[email protected]>
Link: https://patchwork.freedesktop.org/patch/msgid/[email protected]
Signed-off-by: Rodrigo Vivi <[email protected]>
Signed-off-by: Joonas Lahtinen <[email protected]>
|
|
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 <[email protected]>
Cc: Matthew Auld <[email protected]>
Reviewed-by: Matthew Auld <[email protected]>
Link: https://patchwork.freedesktop.org/patch/msgid/[email protected]
Signed-off-by: Rodrigo Vivi <[email protected]>
Signed-off-by: Joonas Lahtinen <[email protected]>
|
|
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 <[email protected]>
Cc: Matthew Auld <[email protected]>
Reviewed-by: Matthew Auld <[email protected]>
Link: https://patchwork.freedesktop.org/patch/msgid/[email protected]
Signed-off-by: Rodrigo Vivi <[email protected]>
Signed-off-by: Joonas Lahtinen <[email protected]>
|
|
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: 4fe6abb8f513 ("drm/i915/gt: Ignore irq enabling on the virtual engines")
Signed-off-by: Chris Wilson <[email protected]>
Cc: Tvrtko Ursulin <[email protected]>
Reviewed-by: Tvrtko Ursulin <[email protected]>
Link: https://patchwork.freedesktop.org/patch/msgid/[email protected]
Signed-off-by: Rodrigo Vivi <[email protected]>
Signed-off-by: Joonas Lahtinen <[email protected]>
|
|
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: 22b7a426bbe1 ("drm/i915/execlists: Preempt-to-busy")
References: 6d06779e8672 ("drm/i915: Load balancing across a virtual engine"
Signed-off-by: Chris Wilson <[email protected]>
Cc: Tvrtko Ursulin <[email protected]>
Cc: "Nayana, Venkata Ramana" <[email protected]>
Reviewed-by: Tvrtko Ursulin <[email protected]>
Link: https://patchwork.freedesktop.org/patch/msgid/[email protected]
Signed-off-by: Rodrigo Vivi <[email protected]>
Signed-off-by: Joonas Lahtinen <[email protected]>
|
|
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 <[email protected]>
Cc: Tvrtko Ursulin <[email protected]>
Reviewed-by: Tvrtko Ursulin <[email protected]>
Link: https://patchwork.freedesktop.org/patch/msgid/[email protected]
Signed-off-by: Rodrigo Vivi <[email protected]>
Signed-off-by: Joonas Lahtinen <[email protected]>
|
|
It turns out that currently we rely only on sysfs attribute
permissions:
$ ll /sys/bus/rbd/{add*,remove*}
--w------- 1 root root 4096 Sep 3 20:37 /sys/bus/rbd/add
--w------- 1 root root 4096 Sep 3 20:37 /sys/bus/rbd/add_single_major
--w------- 1 root root 4096 Sep 3 20:37 /sys/bus/rbd/remove
--w------- 1 root root 4096 Sep 3 20:38 /sys/bus/rbd/remove_single_major
This means that images can be mapped and unmapped (i.e. block devices
can be created and deleted) by a UID 0 process even after it drops all
privileges or by any process with CAP_DAC_OVERRIDE in its user namespace
as long as UID 0 is mapped into that user namespace.
Be consistent with other virtual block devices (loop, nbd, dm, md, etc)
and require CAP_SYS_ADMIN in the initial user namespace for mapping and
unmapping, and also for dumping the configuration string and refreshing
the image header.
Cc: [email protected]
Signed-off-by: Ilya Dryomov <[email protected]>
Reviewed-by: Jeff Layton <[email protected]>
|
|
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 <[email protected]>
Cc: Tvrtko Ursulin <[email protected]>
Reviewed-by: Tvrtko Ursulin <[email protected]>
Link: https://patchwork.freedesktop.org/patch/msgid/[email protected]
Signed-off-by: Rodrigo Vivi <[email protected]>
[Joonas: Rebased and reordered into drm-intel-gt-next branch]
Signed-off-by: Joonas Lahtinen <[email protected]>
|
|
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 <[email protected]>
Reviewed-by: Thomas Hellström <[email protected]>
Reviewed-by: Tvrtko Ursulin <[email protected]>
Link: https://patchwork.freedesktop.org/patch/msgid/[email protected]
Signed-off-by: Rodrigo Vivi <[email protected]>
Signed-off-by: Joonas Lahtinen <[email protected]>
|
|
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 <[email protected]>
Reviewed-by: Thomas Hellström <[email protected]>
Link: https://patchwork.freedesktop.org/patch/msgid/[email protected]
Signed-off-by: Rodrigo Vivi <[email protected]>
Signed-off-by: Joonas Lahtinen <[email protected]>
|
|
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 <[email protected]>
Reviewed-by: Thomas Hellström <[email protected]>
Reviewed-by: Tvrtko Ursulin <[email protected]>
Link: https://patchwork.freedesktop.org/patch/msgid/[email protected]
Signed-off-by: Rodrigo Vivi <[email protected]>
Signed-off-by: Joonas Lahtinen <[email protected]>
|
|
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 <[email protected]>
Reviewed-by: Thomas Hellström <[email protected]>
Link: https://patchwork.freedesktop.org/patch/msgid/[email protected]
Signed-off-by: Rodrigo Vivi <[email protected]>
Signed-off-by: Joonas Lahtinen <[email protected]>
|
|
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 <[email protected]>
Reviewed-by: Thomas Hellström <[email protected]>
Link: https://patchwork.freedesktop.org/patch/msgid/[email protected]
Signed-off-by: Rodrigo Vivi <[email protected]>
Signed-off-by: Joonas Lahtinen <[email protected]>
|
|
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 <[email protected]>
Reviewed-by: Tvrtko Ursulin <[email protected]>
Reviewed-by: Thomas Hellström <[email protected]>
Link: https://patchwork.freedesktop.org/patch/msgid/[email protected]
Signed-off-by: Rodrigo Vivi <[email protected]>
Signed-off-by: Joonas Lahtinen <[email protected]>
|
|
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: d1bf5dd8f6d5 ("drm/i915/gt: Support multiple pinned timelines")
Signed-off-by: Chris Wilson <[email protected]>
Reviewed-by: Matthew Auld <[email protected]>
Link: https://patchwork.freedesktop.org/patch/msgid/[email protected]
Signed-off-by: Rodrigo Vivi <[email protected]>
[Joonas: Updated Fixes: link after rebasing and reordering into drm-intel-gt-next branch]
Signed-off-by: Joonas Lahtinen <[email protected]>
|
|
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 <[email protected]>
Reviewed-by: Matthew Auld <[email protected]>
Link: https://patchwork.freedesktop.org/patch/msgid/[email protected]
Signed-off-by: Rodrigo Vivi <[email protected]>
Signed-off-by: Joonas Lahtinen <[email protected]>
|
|
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 <[email protected]>
Reviewed-by: Matthew Auld <[email protected]>
Link: https://patchwork.freedesktop.org/patch/msgid/[email protected]
Signed-off-by: Rodrigo Vivi <[email protected]>
Signed-off-by: Joonas Lahtinen <[email protected]>
|
|
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 <[email protected]>
Fixes: 3aa9945a528e ("drm/i915: Separate GEM context construction and registration to userspace")
References: f6e8aa387171 ("drm/i915: Report the number of closed vma held by each context in debugfs")
Signed-off-by: Chris Wilson <[email protected]>
Cc: Tvrtko Ursulin <[email protected]>
Cc: CQ Tang <[email protected]>
Cc: <[email protected]> # v5.2+
Reviewed-by: Mika Kuoppala <[email protected]>
Link: https://patchwork.freedesktop.org/patch/msgid/[email protected]
Signed-off-by: Rodrigo Vivi <[email protected]>
Signed-off-by: Joonas Lahtinen <[email protected]>
|
|
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: 06b73c2d0b65 ("drm/i915/gt: Delay taking the spinlock for grabbing from the buffer pool")
Signed-off-by: Chris Wilson <[email protected]>
Reviewed-by: Matthew Auld <[email protected]>
Link: https://patchwork.freedesktop.org/patch/msgid/[email protected]
Signed-off-by: Rodrigo Vivi <[email protected]>
[Joonas: Updating Fixes: link after rebasing and reordering into drm-intel-gt-next]
Signed-off-by: Joonas Lahtinen <[email protected]>
|
|
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 <[email protected]>
Reviewed-by: Thomas Hellström <[email protected]>
Link: https://patchwork.freedesktop.org/patch/msgid/[email protected]
Signed-off-by: Rodrigo Vivi <[email protected]>
Signed-off-by: Joonas Lahtinen <[email protected]>
|
|
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 <[email protected]>
Reviewed-by: Matthew Auld <[email protected]>
Link: https://patchwork.freedesktop.org/patch/msgid/[email protected]
Signed-off-by: Rodrigo Vivi <[email protected]>
Signed-off-by: Joonas Lahtinen <[email protected]>
|
|
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/2169
Signed-off-by: Chris Wilson <[email protected]>
Cc: Mika Kuoppala <[email protected]>
Reviewed-by: Mika Kuoppala <[email protected]>
Link: https://patchwork.freedesktop.org/patch/msgid/[email protected]
Signed-off-by: Rodrigo Vivi <[email protected]>
Signed-off-by: Joonas Lahtinen <[email protected]>
|
|
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 <[email protected]>
Cc: Tvrtko Ursulin <[email protected]>
Cc: Mika Kuoppala <[email protected]>
Reviewed-by: Tvrtko Ursulin <[email protected]>
Link: https://patchwork.freedesktop.org/patch/msgid/[email protected]
Signed-off-by: Rodrigo Vivi <[email protected]>
Signed-off-by: Joonas Lahtinen <[email protected]>
|
|
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 <[email protected]>
Cc: Mika Kuoppala <[email protected]>
Reviewed-by: Mika Kuoppala <[email protected]>
Link: https://patchwork.freedesktop.org/patch/msgid/[email protected]
Signed-off-by: Rodrigo Vivi <[email protected]>
Signed-off-by: Joonas Lahtinen <[email protected]>
|
|
Since we assert that the kernel_context is using the perma-pinned HWSP,
make it so.
Closes: https://gitlab.freedesktop.org/drm/intel/-/issues/2179
Signed-off-by: Chris Wilson <[email protected]>
Cc: Mika Kuoppala <[email protected]>
Reviewed-by: Mika Kuoppala <[email protected]>
Link: https://patchwork.freedesktop.org/patch/msgid/[email protected]
Signed-off-by: Rodrigo Vivi <[email protected]>
Signed-off-by: Joonas Lahtinen <[email protected]>
|
|
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: Chris Wilson <[email protected]>
Cc: Matthew Auld <[email protected]>
Cc: Tvrtko Ursulin <[email protected]>
Reviewed-by: Tvrtko Ursulin <[email protected]>
Link: https://patchwork.freedesktop.org/patch/msgid/[email protected]
Signed-off-by: Rodrigo Vivi <[email protected]>
Signed-off-by: Joonas Lahtinen <[email protected]>
|
|
__kobject_del() is called from two places, in one where kobj is dereferenced
before and thus can't be NULL, and in the other the NULL check is done before
call. Drop unneeded conditional in __kobject_del().
Signed-off-by: Andy Shevchenko <[email protected]>
Reviewed-by: Rafael J. Wysocki <[email protected]>
Link: https://lore.kernel.org/r/[email protected]
Signed-off-by: Greg Kroah-Hartman <[email protected]>
|
|
SDHCI changed from using a tasklet to finish requests, to using an IRQ
thread i.e. commit c07a48c2651965 ("mmc: sdhci: Remove finish_tasklet").
Because this increased the latency to complete requests, a preparatory
change was made to complete the request from the IRQ handler if
possible i.e. commit 19d2f695f4e827 ("mmc: sdhci: Call mmc_request_done()
from IRQ handler if possible"). That alleviated the situation for MMC
block devices because the MMC block driver makes use of mmc_pre_req()
and mmc_post_req() so that successful requests are completed in the IRQ
handler and any DMA unmapping is handled separately in mmc_post_req().
However SDIO was still affected, and an example has been reported with
up to 20% degradation in performance.
Looking at SDIO I/O helper functions, sdio_io_rw_ext_helper() appeared
to be a possible candidate for making use of asynchronous requests
within its I/O loops, but analysis revealed that these loops almost
never iterate more than once, so the complexity of the change would not
be warrented.
Instead, mmc_pre_req() and mmc_post_req() are added before and after I/O
submission (mmc_wait_for_req) in mmc_io_rw_extended(). This still has
the potential benefit of reducing the duration of interrupt handlers, as
well as addressing the latency issue for SDHCI. It also seems a more
reasonable solution than forcing drivers to do everything in the IRQ
handler.
Reported-by: Dmitry Osipenko <[email protected]>
Fixes: c07a48c2651965 ("mmc: sdhci: Remove finish_tasklet")
Signed-off-by: Adrian Hunter <[email protected]>
Tested-by: Dmitry Osipenko <[email protected]>
Cc: [email protected]
Link: https://lore.kernel.org/r/[email protected]
Signed-off-by: Ulf Hansson <[email protected]>
|
|
Commit b214fe592ab7 ("mmc: sdhci-of-esdhc: add erratum eSDHC7 support")
added code to check for a specific compatible string in the device-tree
on every esdhc interrupat. Instead of doing this record the quirk in
struct sdhci_esdhc and lookup the struct in esdhc_irq.
Signed-off-by: Chris Packham <[email protected]>
Link: https://lore.kernel.org/r/[email protected]
Fixes: b214fe592ab7 ("mmc: sdhci-of-esdhc: add erratum eSDHC7 support")
Cc: [email protected]
Signed-off-by: Ulf Hansson <[email protected]>
|
|
The commit cd57d07b1e4e ("sh: don't allow non-coherent DMA for NOMMU") made
CONFIG_NO_DMA to be set for some platforms, for good reasons.
Consequentially, CONFIG_HAS_DMA doesn't get set, which makes the DMA
mapping interface to be built as stub functions, but also prevent the
mmc_spi driver from being built as it depends on CONFIG_HAS_DMA.
It turns out that for some odd cases, the driver still relied on the DMA
mapping interface, even if the DMA was not actively being used.
To fixup the behaviour, let's drop the build dependency for CONFIG_HAS_DMA.
Moreover, as to allow the driver to succeed probing, let's move the DMA
initializations behind "#ifdef CONFIG_HAS_DMA".
Fixes: cd57d07b1e4e ("sh: don't allow non-coherent DMA for NOMMU")
Reported-by: Rich Felker <[email protected]>
Signed-off-by: Ulf Hansson <[email protected]>
Tested-by: Rich Felker <[email protected]>
Link: https://lore.kernel.org/r/[email protected]
|
|
As the comments in this patch say, if we tune and find all phases are
valid it's _almost_ as bad as no phases being found valid. Probably
all phases are not really reliable but we didn't detect where the
unreliable place is. That means we'll essentially be guessing and
hoping we get a good phase.
This is not just a problem in theory. It was causing real problems on
a real board. On that board, most often phase 10 is found as the only
invalid phase, though sometimes 10 and 11 are invalid and sometimes
just 11. Some percentage of the time, however, all phases are found
to be valid. When this happens, the current logic will decide to use
phase 11. Since phase 11 is sometimes found to be invalid, this is a
bad choice. Sure enough, when phase 11 is picked we often get mmc
errors later in boot.
I have seen cases where all phases were found to be valid 3 times in a
row, so increase the retry count to 10 just to be extra sure.
Fixes: 415b5a75da43 ("mmc: sdhci-msm: Add platform_execute_tuning implementation")
Signed-off-by: Douglas Anderson <[email protected]>
Reviewed-by: Veerabhadrarao Badiganti <[email protected]>
Acked-by: Adrian Hunter <[email protected]>
Link: https://lore.kernel.org/r/20200827075809.1.If179abf5ecb67c963494db79c3bc4247d987419b@changeid
Signed-off-by: Ulf Hansson <[email protected]>
|
|
The commit 61d7437ed1390 ("mmc: sdhci-acpi: Fix HS400 tuning for AMDI0040")
broke resume for eMMC HS400. When the system suspends the eMMC controller
is powered down. So, on resume we need to reinitialize the controller.
Although, amd_sdhci_host was not getting cleared, so the DLL was never
re-enabled on resume. This results in HS400 being non-functional.
To fix the problem, this change clears the tuned_clock flag, clears the
dll_enabled flag and disables the DLL on reset.
Fixes: 61d7437ed1390 ("mmc: sdhci-acpi: Fix HS400 tuning for AMDI0040")
Signed-off-by: Raul E Rangel <[email protected]>
Acked-by: Adrian Hunter <[email protected]>
Link: https://lore.kernel.org/r/20200831150517.1.I93c78bfc6575771bb653c9d3fca5eb018a08417d@changeid
Signed-off-by: Ulf Hansson <[email protected]>
|
|
RHBZ: 1871246
If during cifs_lookup()/get_inode_info() we encounter a DFS link
and we use the cifsacl or modefromsid mount options we must suppress
any -EREMOTE errors that triggers or else we will not be able to follow
the DFS link and automount the target.
This fixes an issue with modefromsid/cifsacl where these mountoptions
would break DFS and we would no longer be able to access the share.
Signed-off-by: Ronnie Sahlberg <[email protected]>
Reviewed-by: Paulo Alcantara (SUSE) <[email protected]>
Signed-off-by: Steve French <[email protected]>
|
|
|
|
Pull more io_uring fixes from Jens Axboe:
"Two followup fixes. One is fixing a regression from this merge window,
the other is two commits fixing cancelation of deferred requests.
Both have gone through full testing, and both spawned a few new
regression test additions to liburing.
- Don't play games with const, properly store the output iovec and
assign it as needed.
- Deferred request cancelation fix (Pavel)"
* tag 'io_uring-5.9-2020-09-06' of git://git.kernel.dk/linux-block:
io_uring: fix linked deferred ->files cancellation
io_uring: fix cancel of deferred reqs with ->files
io_uring: fix explicit async read/write mapping for large segments
|
|
git://git.kernel.org/pub/scm/linux/kernel/git/joro/iommu
Pull iommu fixes from Joerg Roedel:
- three Intel VT-d fixes to fix address handling on 32bit, fix a NULL
pointer dereference bug and serialize a hardware register access as
required by the VT-d spec.
- two patches for AMD IOMMU to force AMD GPUs into translation mode
when memory encryption is active and disallow using IOMMUv2
functionality. This makes the AMDGPU driver work when memory
encryption is active.
- two more fixes for AMD IOMMU to fix updating the Interrupt Remapping
Table Entries.
- MAINTAINERS file update for the Qualcom IOMMU driver.
* tag 'iommu-fixes-v5.9-rc3' of git://git.kernel.org/pub/scm/linux/kernel/git/joro/iommu:
iommu/vt-d: Handle 36bit addressing for x86-32
iommu/amd: Do not use IOMMUv2 functionality when SME is active
iommu/amd: Do not force direct mapping when SME is active
iommu/amd: Use cmpxchg_double() when updating 128-bit IRTE
iommu/amd: Restore IRTE.RemapEn bit after programming IRTE
iommu/vt-d: Fix NULL pointer dereference in dev_iommu_priv_set()
iommu/vt-d: Serialize IOMMU GCMD register modifications
MAINTAINERS: Update QUALCOMM IOMMU after Arm SMMU drivers move
|
|
git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip
Pull x86 fixes from Ingo Molnar:
- more generic entry code ABI fallout
- debug register handling bugfixes
- fix vmalloc mappings on 32-bit kernels
- kprobes instrumentation output fix on 32-bit kernels
- fix over-eager WARN_ON_ONCE() on !SMAP hardware
- NUMA debugging fix
- fix Clang related crash on !RETPOLINE kernels
* tag 'x86-urgent-2020-09-06' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
x86/entry: Unbreak 32bit fast syscall
x86/debug: Allow a single level of #DB recursion
x86/entry: Fix AC assertion
tracing/kprobes, x86/ptrace: Fix regs argument order for i386
x86, fakenuma: Fix invalid starting node ID
x86/mm/32: Bring back vmalloc faulting on x86_32
x86/cmdline: Disable jump tables for cmdline.c
|
|
git://git.kernel.org/pub/scm/linux/kernel/git/xen/tip
Pull xen updates from Juergen Gross:
"A small series for fixing a problem with Xen PVH guests when running
as backends (e.g. as dom0).
Mapping other guests' memory is now working via ZONE_DEVICE, thus not
requiring to abuse the memory hotplug functionality for that purpose"
* tag 'for-linus-5.9-rc4-tag' of git://git.kernel.org/pub/scm/linux/kernel/git/xen/tip:
xen: add helpers to allocate unpopulated memory
memremap: rename MEMORY_DEVICE_DEVDAX to MEMORY_DEVICE_GENERIC
xen/balloon: add header guard
|
|
While looking for ->files in ->defer_list, consider that requests there
may actually be links.
Signed-off-by: Pavel Begunkov <[email protected]>
Signed-off-by: Jens Axboe <[email protected]>
|
|
While trying to cancel requests with ->files, it also should look for
requests in ->defer_list, otherwise it might end up hanging a thread.
Cancel all requests in ->defer_list up to the last request there with
matching ->files, that's needed to follow drain ordering semantics.
Signed-off-by: Pavel Begunkov <[email protected]>
Signed-off-by: Jens Axboe <[email protected]>
|