Jani Nikula [Thu, 10 Sep 2020 09:52:27 +0000 (12:52 +0300)]
drm/i915: move gmbus restore to i915_restore_display
Logically part of the display restore.
Note: This has been in place since the introduction of gmbus
support. The gmbus code also does the resets before transfers. Is this
really needed, or a historical accident?
Jani Nikula [Thu, 10 Sep 2020 09:52:25 +0000 (12:52 +0300)]
drm/i915: disable all display features when no display
Disable all display feature flags when there are no pipes i.e. there is
no display. This should help with not having to additionally check for
HAS_DISPLAY() when a feature flag check would suffice.
In the merge commit 1f4b2aca794f ("Merge tag
'drm-intel-gt-next-2020-09-07' of git://anongit.freedesktop.org/drm/drm-intel into drm-next") things
went wrong, but the merge commit view now doesn't show any conflict
anymore (as git tends to do when the resolution picks one or the other
branch).
The need to handle other than just true/false error codes in
__reloc_entry_gpu was added in the dma_resv locking changes in c43ce12328df ("drm/i915: Use per object locking in execbuf, v12.")
Signed-off-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com> Reported-by: Dan Carpenter <dan.carpenter@oracle.com> Cc: Dave Airlie <airlied@redhat.com>
[danvet: Explain this entire saga a lot better, adding tons of commit
references. Also note that this was merged before full intel-gfx-CI
results, only after BAT, since the breakage at the BAT run is already
severe enough to block all pre-merge testing.] Fixes: 1f4b2aca794f ("Merge tag 'drm-intel-gt-next-2020-09-07' of git://anongit.freedesktop.org/drm/drm-intel into drm-next") Acked-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch> Link: https://patchwork.freedesktop.org/patch/msgid/20200910111225.2184193-1-maarten.lankhorst@linux.intel.com
Dave Airlie [Tue, 8 Sep 2020 21:53:59 +0000 (07:53 +1000)]
Merge tag 'drm-intel-gt-next-2020-09-07' of git://anongit.freedesktop.org/drm/drm-intel into drm-next
(Same content as drm-intel-gt-next-2020-09-04-3, S-o-b's added)
UAPI Changes:
(- Potential implicit changes from WW locking refactoring)
Cross-subsystem Changes:
(- WW locking changes should align the i915 locking more with others)
Driver Changes:
- MAJOR: Apply WW locking across the driver (Maarten)
- Reverts for 5 commits to make applying WW locking faster (Maarten)
- Disable preparser around invalidations on Tigerlake for non-RCS engines (Chris)
- Add missing dma_fence_put() for error case of syncobj timeline (Chris)
- Parse command buffer earlier in eb_relocate(slow) to facilitate backoff (Maarten)
- Pin engine before pinning all objects (Maarten)
- Rework intel_context pinning to do everything outside of pin_mutex (Maarten)
- Avoid tracking GEM context until registered (Cc: stable, Chris)
- Provide a fastpath for waiting on vma bindings (Chris)
- Fixes to preempt-to-busy mechanism (Chris)
- Distinguish the virtual breadcrumbs from the irq breadcrumbs (Chris)
- Switch to object allocations for page directories (Chris)
- Hold context/request reference while breadcrumbs are active (Chris)
- Make sure execbuffer always passes ww state to i915_vma_pin (Maarten)
- Code refactoring to facilitate use of WW locking (Maarten)
- Locking refactoring to use more granular locking (Maarten, Chris)
- Support for multiple pinned timelines per engine (Chris)
- Move complication of I915_GEM_THROTTLE to the ioctl from general code (Chris)
- Make active tracking/vma page-directory stash work preallocated (Chris)
- Avoid flushing submission tasklet too often (Chris)
- Reduce context termination list iteration guard to RCU (Chris)
- Reductions to locking contention (Chris)
- Fixes for issues found by CI (Chris)
Merge tag 'drm-fixes-2020-09-08' of git://anongit.freedesktop.org/drm/drm
Pull drm fixes from Dave Airlie:
"The i915 reverts are going to be a bit of a conflict mess for next, so
I decided to dequeue them now, along with some msm fixes for a ring
corruption issue, that Rob sent over the weekend.
Summary:
i915:
- revert gpu relocation changes due to regression
msm:
- fixes for RPTR corruption issue"
* tag 'drm-fixes-2020-09-08' of git://anongit.freedesktop.org/drm/drm:
Revert "drm/i915/gem: Delete unused code"
Revert "drm/i915/gem: Async GPU relocations only"
Revert "drm/i915: Remove i915_gem_object_get_dirty_page()"
drm/msm: Disable the RPTR shadow
drm/msm: Disable preemption on all 5xx targets
drm/msm: Enable expanded apriv support for a650
drm/msm: Split the a5xx preemption record
Merge tag 'livepatching-for-5.9-rc5' of git://git.kernel.org/pub/scm/linux/kernel/git/livepatching/livepatching
Pull livepatching fix from Petr Mladek:
"Workaround for 'unreachable instruction' objtool warnings that happen
with some compiler versions"
* tag 'livepatching-for-5.9-rc5' of git://git.kernel.org/pub/scm/linux/kernel/git/livepatching/livepatching:
Revert "kbuild: use -flive-patching when CONFIG_LIVEPATCH is enabled"
Dave Airlie [Tue, 8 Sep 2020 06:40:13 +0000 (16:40 +1000)]
Merge tag 'amd-drm-next-5.10-2020-09-03' of git://people.freedesktop.org/~agd5f/linux into drm-next
amd-drm-next-5.10-2020-09-03:
amdgpu:
- RAS fixes
- Sienna Cichlid updates
- Navy Flounder updates
- DCE6 (SI) support in DC
- Enable plane rotation
- Rework pre-OS vram reservation handling during driver init
- Add standard interface to dump GPU metrics table from SMU
- Rework tiling and tmz state handling in atomic commits
- Pstate fixes
- Add voltage and power hwmon interfaces for renoir
- SW CTF fixes
- S/G display fix for Raven
- Print client strings for vmfaults for vega and newer
- Manual fan control fixes
- Display updates
- Reorg power management directory structure
- Misc bug fixes
- Misc code cleanups
amdkfd:
- Topology fixes
- Add SMI events for thermal throttling and GPU resets
radeon:
- switch from pci_* to dma_* for dma allocations
- PLL fix
Scheduler:
- Clean up priority levels
UAPI:
- amdgpu INFO IOCTL query update for TMZ state
https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6049
- amdkfd SMI event interface updates
https://github.com/RadeonOpenCompute/rocm_smi_lib/tree/therm_thrott
Dave Airlie [Tue, 8 Sep 2020 05:41:43 +0000 (15:41 +1000)]
Revert "drm/i915/gem: Delete unused code"
These commits caused a regression on Lenovo t520 sandybridge
machine belonging to reporter. We are reverting them for 5.10
for other reasons, so just do it for 5.9 as well.
Dave Airlie [Tue, 8 Sep 2020 05:41:17 +0000 (15:41 +1000)]
Revert "drm/i915/gem: Async GPU relocations only"
These commits caused a regression on Lenovo t520 sandybridge
machine belonging to reporter. We are reverting them for 5.10
for other reasons, so just do it for 5.9 as well.
These commits caused a regression on Lenovo t520 sandybridge
machine belonging to reporter. We are reverting them for 5.10
for other reasons, so just do it for 5.9 as well.
drm/i915: Unlock the shared hwsp_gtt object after pinning
The hwsp_gtt object is used for sub-allocation and could therefore
be shared by many contexts causing unnecessary contention during
concurrent context pinning.
However since we're currently locking it only for pinning, it remains
resident until we unpin it, and therefore it's safe to drop the
lock early, allowing for concurrent thread access.
Signed-off-by: Thomas Hellström <thomas.hellstrom@intel.com> Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> Signed-off-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Chris Wilson [Tue, 28 Jul 2020 15:21:44 +0000 (16:21 +0100)]
drm/i915: Filter wake_flags passed to default_wake_function
(NOTE: This is the minimal backportable fix, a full fix is being
developed at https://patchwork.freedesktop.org/patch/388048/)
The flags passed to the wait_entry.func are passed onwards to
try_to_wake_up(), which has a very particular interpretation for its
wake_flags. In particular, beyond the published WF_SYNC, it has a few
internal flags as well. Since we passed the fence->error down the chain
via the flags argument, these ended up in the default_wake_function
confusing the kernel/sched.
Closes: https://gitlab.freedesktop.org/drm/intel/-/issues/2110 Fixes: ef4688497512 ("drm/i915: Propagate fence errors") Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> Cc: Matthew Auld <matthew.auld@intel.com> Cc: <stable@vger.kernel.org> # v5.4+ Reviewed-by: Matthew Auld <matthew.auld@intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20200728152144.1100-1-chris@chris-wilson.co.uk Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
[Joonas: Rebased and reordered into drm-intel-gt-next branch]
[Joonas: Added a note and link about more complete fix] Signed-off-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Chris Wilson [Thu, 16 Jul 2020 14:22:07 +0000 (15:22 +0100)]
drm/i915: Remove i915_request.lock requirement for execution callbacks
To implement preempt-to-busy (and so efficient timeslicing and best utilization
of the hardware submission ports) we let the GPU run asynchronously in respect
to the ELSP submission queue. This created challenges in keeping and accessing
the driver state mirroring the asynchronous GPU execution.
Previous fix 1d9221e9d395 ("drm/i915: Skip signaling a signaled request")
however did not correctly serialize request retirement with the execution
callbacks.
We were using the i915_request.lock to serialise adding an execution callback
with __i915_request_submit. However, if we use an atomic llist_add to serialise
multiple waiters and then check to see if the request is already executing, we
can remove the irq-spinlock and fix serialization between retirement and
execution callbacks in one go.
v2: Avoid using the irq_work when outside of the irq-spinlocks, where we
can execute the callbacks immediately.
v3: Pay close attention to the order of setting ACTIVE on retirement, we
need to ensure the request is signaled and breadcrumbs detached before
we finish removing the request from the engine.
v4: Expanded commit message.
Fixes: 1d9221e9d395 ("drm/i915: Skip signaling a signaled request") Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com> Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20200716142207.13003-2-chris@chris-wilson.co.uk Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
[Joonas: Rebased and reordered into drm-intel-gt-next branch]
[Joonas: Added expanded commit message from Tvrtko and Chris] Signed-off-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Chris Wilson [Thu, 16 Jul 2020 14:22:06 +0000 (15:22 +0100)]
drm/i915: Be wary of data races when reading the active execlists
To implement preempt-to-busy (and so efficient timeslicing and best utilization
of the hardware submission ports) we let the GPU run asynchronously in respect
to the ELSP submission queue. This created challenges in keeping and accessing
the driver state mirroring the asynchronous GPU execution.
The latest occurence of this was spotted by KCSAN:
[ 1413.563200] BUG: KCSAN: data-race in __await_execution+0x217/0x370 [i915]
[ 1413.563221]
[ 1413.563236] race at unknown origin, with read to 0xffff88885bb6c478 of 8 bytes by task 9654 on cpu 1:
[ 1413.563548] __await_execution+0x217/0x370 [i915]
[ 1413.563891] i915_request_await_dma_fence+0x4eb/0x6a0 [i915]
[ 1413.564235] i915_request_await_object+0x421/0x490 [i915]
[ 1413.564577] i915_gem_do_execbuffer+0x29b7/0x3c40 [i915]
[ 1413.564967] i915_gem_execbuffer2_ioctl+0x22f/0x5c0 [i915]
[ 1413.564998] drm_ioctl_kernel+0x156/0x1b0
[ 1413.565022] drm_ioctl+0x2ff/0x480
[ 1413.565046] __x64_sys_ioctl+0x87/0xd0
[ 1413.565069] do_syscall_64+0x4d/0x80
[ 1413.565094] entry_SYSCALL_64_after_hwframe+0x44/0xa9
To complicate matters, we have to both avoid the read tearing of *active and
avoid any write tearing as perform the pending[] -> inflight[] promotion of the
execlists.
This is because we cannot rely on the memcpy doing u64 aligned copies on all
kernels/platforms and so we opt to open-code it with explicit WRITE_ONCE
annotations to satisfy KCSAN.
v2: When in doubt, write the same comment again.
v3: Expanded commit message.
Fixes: b55230e5e800 ("drm/i915: Check for awaits on still currently executing requests") Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com> Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20200716142207.13003-1-chris@chris-wilson.co.uk Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
[Joonas: Rebased and reordered into drm-intel-gt-next branch]
[Joonas: Added expanded commit message from Tvrtko and Chris] Signed-off-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
drm/i915: Add ww locking to pin_to_display_plane, v2.
Use ww locking for pin_to_display_plane for all the pinning and locking.
With the locking removed from set_cache_level, we need to fix
i915_gem_set_caching_ioctl to take the object reservation lock.
As this is a single lock, we don't need to use the ww dance.
Changes since v1:
- Do not use ww locking in i915_gem_set_caching_ioctl (Thomas).
We want to start requiring the reservation_lock instead of obj->mm.lock
for pinning objects, take the ww lock inside vm_fault_gtt as a first step
towards the legacy lock removal.
drm/i915/selftests: Fix locking inversion in lrc selftest.
This function does not use intel_context_create_request, so it has
to use the same locking order as normal code. This is required to
shut up lockdep in selftests.
drm/i915: Dirty hack to fix selftests locking inversion
Some i915 selftests still use i915_vma_lock() as inner lock, and
intel_context_create_request() intel_timeline->mutex as outer lock.
Fortunately for selftests this is not an issue, they should be fixed
but we can move ahead and cleanify lockdep now.
drm/i915: Make sure execbuffer always passes ww state to i915_vma_pin.
As a preparation step for full object locking and wait/wound handling
during pin and object mapping, ensure that we always pass the ww context
in i915_gem_execbuffer.c to i915_vma_pin, use lockdep to ensure this
happens.
This also requires changing the order of eb_parse slightly, to ensure
we pass ww at a point where we could still handle -EDEADLK safely.
drm/i915: Rework intel_context pinning to do everything outside of pin_mutex
Instead of doing everything inside of pin_mutex, we move all pinning
outside. Because i915_active has its own reference counting and
pinning is also having the same issues vs mutexes, we make sure
everything is pinned first, so the pinning in i915_active only needs
to bump refcounts. This allows us to take pin refcounts correctly
all the time.
drm/i915: Pin engine before pinning all objects, v5.
We want to lock all gem objects, including the engine context objects,
rework the throttling to ensure that we can do this. Now we only throttle
once, but can take eb_pin_engine while acquiring objects. This means we
will have to drop the lock to wait. If we don't have to throttle we can
still take the fastpath, if not we will take the slowpath and wait for
the throttle request while unlocked.
The engine has to be pinned as first step, otherwise gpu relocations
won't work.
Changes since v1:
- Only need to get a throttled request in the fastpath, no need for
a global flag any more.
- Always free the waited request correctly.
Changes since v2:
- Use intel_engine_pm_get()/put() to keeep engine pool alive during
EDEADLK handling.
Changes since v3:
- Fix small rq leak.
Changes since v4:
- Use a single reloc_context, for intel_context_pin_ww().
Those arguments are already set as eb.file and eb.args, so kill off
the extra arguments. This will allow us to move eb_pin_engine() to
after we reserved all BO's.
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.
drm/i915: Parse command buffer earlier in eb_relocate(slow)
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.
drm/i915: Remove locking from i915_gem_object_prepare_read/write
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.
Revert "drm/i915/gem: Split eb_vma into its own allocation"
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.
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 <chris@chris-wilson.co.uk> Cc: Matthew Auld <matthew.auld@intel.com> Signed-off-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
[mlankhorst: Removed extra reloc_gpu_flush()] Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch> Link: https://patchwork.freedesktop.org/patch/msgid/20200819140904.1708856-4-maarten.lankhorst@linux.intel.com Signed-off-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
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)
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").
Chris Wilson [Thu, 6 Aug 2020 16:10:56 +0000 (17:10 +0100)]
drm/i915/gem: Free the fence after a fence-chain lookup failure
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.
Chris Wilson [Thu, 6 Aug 2020 10:59:54 +0000 (11:59 +0100)]
drm/i915/gem: Reduce context termination list iteration guard to RCU
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 <chris@chris-wilson.co.uk> Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20200806105954.7766-1-chris@chris-wilson.co.uk Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com> Signed-off-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Chris Wilson [Thu, 6 Aug 2020 14:57:28 +0000 (15:57 +0100)]
drm/i915/selftests: Prevent selecting 0 for our random width/align
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!
Chris Wilson [Sat, 1 Aug 2020 16:02:25 +0000 (17:02 +0100)]
drm/i915/gt: Hold context/request reference while breadcrumbs are active
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 <chris@chris-wilson.co.uk> Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20200801160225.6814-2-chris@chris-wilson.co.uk Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
[Joonas: Rebased and reordered into drm-intel-gt-next branch] Signed-off-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
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 <chris@chris-wilson.co.uk> Cc: Mika Kuoppala <mika.kuoppala@linux.intel.com> Cc: Matthew Auld <matthew.auld@intel.com> Reviewed-by: Matthew Auld <matthew.auld@intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20200729164219.5737-3-chris@chris-wilson.co.uk Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com> Signed-off-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Chris Wilson [Wed, 29 Jul 2020 16:42:18 +0000 (17:42 +0100)]
drm/i915/gt: Switch to object allocations for page directories
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.
Chris Wilson [Wed, 29 Jul 2020 16:42:17 +0000 (17:42 +0100)]
drm/i915: Preallocate stashes for vma page-directories
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.
Chris Wilson [Fri, 31 Jul 2020 15:48:34 +0000 (16:48 +0100)]
drm/i915/gt: Distinguish the virtual breadcrumbs from the irq breadcrumbs
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 <chris@chris-wilson.co.uk> Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com> Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20200731154834.8378-4-chris@chris-wilson.co.uk Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com> Signed-off-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Chris Wilson [Fri, 31 Jul 2020 15:48:33 +0000 (16:48 +0100)]
drm/i915/gt: Only transfer the virtual context to the new engine if active
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 <chris@chris-wilson.co.uk> Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com> Cc: "Nayana, Venkata Ramana" <venkata.ramana.nayana@intel.com> Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20200731154834.8378-3-chris@chris-wilson.co.uk Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com> Signed-off-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
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().
Chris Wilson [Fri, 31 Jul 2020 15:48:31 +0000 (16:48 +0100)]
drm/i915: Remove requirement for holding i915_request.lock for breadcrumbs
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 <chris@chris-wilson.co.uk> Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com> Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20200731154834.8378-1-chris@chris-wilson.co.uk Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
[Joonas: Rebased and reordered into drm-intel-gt-next branch] Signed-off-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Chris Wilson [Fri, 31 Jul 2020 08:50:15 +0000 (09:50 +0100)]
drm/i915: Provide a fastpath for waiting on vma bindings
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 <chris@chris-wilson.co.uk> Reviewed-by: Thomas Hellström <thomas.hellstrom@intel.com> Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20200731085015.32368-7-chris@chris-wilson.co.uk Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com> Signed-off-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Chris Wilson [Fri, 31 Jul 2020 08:50:14 +0000 (09:50 +0100)]
drm/i915: Reduce locking around i915_active_acquire_preallocate_barrier()
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.
Chris Wilson [Fri, 31 Jul 2020 08:50:13 +0000 (09:50 +0100)]
drm/i915: Make the stale cached active node available for any timeline
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 <chris@chris-wilson.co.uk> Reviewed-by: Thomas Hellström <thomas.hellstrom@intel.com> Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20200731085015.32368-5-chris@chris-wilson.co.uk Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com> Signed-off-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Chris Wilson [Fri, 31 Jul 2020 08:50:12 +0000 (09:50 +0100)]
drm/i915: Keep the most recently used active-fence upon discard
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.
Chris Wilson [Fri, 31 Jul 2020 08:50:11 +0000 (09:50 +0100)]
drm/i915: Export a preallocate variant of i915_active_acquire()
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.
Chris Wilson [Fri, 31 Jul 2020 08:50:10 +0000 (09:50 +0100)]
drm/i915: Skip taking acquire mutex for no ref->active callback
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 <chris@chris-wilson.co.uk> Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com> Reviewed-by: Thomas Hellström <thomas.hellstrom@intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20200731085015.32368-2-chris@chris-wilson.co.uk Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com> Signed-off-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Chris Wilson [Fri, 31 Jul 2020 10:22:06 +0000 (11:22 +0100)]
drm/i915/selftests: Drop stale timeline constructor assert
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 <chris@chris-wilson.co.uk> Reviewed-by: Matthew Auld <matthew.auld@intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20200731102206.6793-1-chris@chris-wilson.co.uk Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
[Joonas: Updated Fixes: link after rebasing and reordering into drm-intel-gt-next branch] Signed-off-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Chris Wilson [Thu, 30 Jul 2020 13:40:49 +0000 (14:40 +0100)]
drm/i915/gt: Pull release of node->age under the spinlock
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.
Chris Wilson [Thu, 30 Jul 2020 18:39:06 +0000 (19:39 +0100)]
drm/i915/gt: Support multiple pinned timelines
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.
Chris Wilson [Thu, 30 Jul 2020 09:28:56 +0000 (10:28 +0100)]
drm/i915/gem: Delay tracking the GEM context until it is registered
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: 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 <chris@chris-wilson.co.uk> Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com> Cc: CQ Tang <cq.tang@intel.com> Cc: <stable@vger.kernel.org> # v5.2+ Reviewed-by: Mika Kuoppala <mika.kuoppala@linux.intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20200730092856.23615-1-chris@chris-wilson.co.uk Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com> Signed-off-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Chris Wilson [Wed, 29 Jul 2020 11:07:56 +0000 (12:07 +0100)]
drm/i915/gt: Fix termination condition for freeing all buffer objects
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 <chris@chris-wilson.co.uk> Reviewed-by: Matthew Auld <matthew.auld@intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20200729110756.2344-1-chris@chris-wilson.co.uk Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
[Joonas: Updating Fixes: link after rebasing and reordering into drm-intel-gt-next] Signed-off-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Chris Wilson [Wed, 29 Jul 2020 08:02:45 +0000 (09:02 +0100)]
drm/i915/gt: Delay taking the spinlock for grabbing from the buffer pool
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.
Chris Wilson [Tue, 28 Jul 2020 15:21:09 +0000 (16:21 +0100)]
drm/i915/gt: Disable preparser around xcs invalidations on tgl
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 <chris@chris-wilson.co.uk> Cc: Mika Kuoppala <mika.kuoppala@linux.intel.com> Reviewed-by: Mika Kuoppala <mika.kuoppala@linux.intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20200728152110.830-1-chris@chris-wilson.co.uk Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com> Signed-off-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Chris Wilson [Tue, 28 Jul 2020 15:20:10 +0000 (16:20 +0100)]
drm/i915/gem: Remove disordered per-file request list for throttling
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 <chris@chris-wilson.co.uk> Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com> Cc: Mika Kuoppala <mika.kuoppala@linux.intel.com> Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20200728152010.30701-1-chris@chris-wilson.co.uk Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com> Signed-off-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Chris Wilson [Wed, 15 Jul 2020 11:50:51 +0000 (12:50 +0100)]
drm/i915: Soften the tasklet flush frequency before waits
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 <chris@chris-wilson.co.uk> Cc: Mika Kuoppala <mika.kuoppala@linux.intel.com> Reviewed-by: Mika Kuoppala <mika.kuoppala@linux.intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20200715115147.11866-10-chris@chris-wilson.co.uk Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com> Signed-off-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Chris Wilson [Thu, 16 Jul 2020 10:07:54 +0000 (11:07 +0100)]
drm/i915: Reduce i915_request.lock contention for i915_request_wait
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 <chris@chris-wilson.co.uk> Cc: Matthew Auld <matthew.auld@intel.com> Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com> Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20200716100754.5670-1-chris@chris-wilson.co.uk Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com> Signed-off-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Merge tag 'io_uring-5.9-2020-09-06' of git://git.kernel.dk/linux-block
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
Merge tag 'iommu-fixes-v5.9-rc3' of 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
Merge tag 'x86-urgent-2020-09-06' of 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
Hans de Goede [Thu, 3 Sep 2020 11:23:37 +0000 (13:23 +0200)]
drm/i915: panel: Use atomic PWM API for devs with an external PWM controller
Now that the PWM drivers which we use have been converted to the atomic
PWM API, we can move the i915 panel code over to using the atomic PWM API.
The removes a long standing FIXME and this removes a flicker where
the backlight brightness would jump to 100% when i915 loads even if
using the fastset path.
Note that this commit also simplifies pwm_disable_backlight(), by dropping
the intel_panel_actually_set_backlight(..., 0) call. This call sets the
PWM to 0% duty-cycle. I believe that this call was only present as a
workaround for a bug in the pwm-crc.c driver where it failed to clear the
PWM_OUTPUT_ENABLE bit. This is fixed by an earlier patch in this series.
After the dropping of this workaround, the usleep call, which seems
unnecessary to begin with, has no useful effect anymore, so drop that too.
Hans de Goede [Thu, 3 Sep 2020 11:23:36 +0000 (13:23 +0200)]
drm/i915: panel: Honor the VBT PWM min setting for devs with an external PWM controller
So far for devices using an external PWM controller (devices using
pwm_setup_backlight()), we have been hardcoding the minimum allowed
PWM level to 0. But several of these devices specify a non 0 minimum
setting in their VBT.
Change pwm_setup_backlight() to use get_backlight_min_vbt() to get
the minimum level.
Hans de Goede [Thu, 3 Sep 2020 11:23:35 +0000 (13:23 +0200)]
drm/i915: panel: Honor the VBT PWM frequency for devs with an external PWM controller
So far for devices using an external PWM controller (devices using
pwm_setup_backlight()), we have been hardcoding the period-time passed to
pwm_config() to 21333 ns.
I suspect this was done because many VBTs set the PWM frequency to 200
which corresponds to a period-time of 5000000 ns, which greatly exceeds
the PWM_MAX_PERIOD_NS define in the Crystal Cove PMIC PWM driver, which
used to be 21333.
This PWM_MAX_PERIOD_NS define was actually based on a bug in the PWM
driver where its period and duty-cycle times where off by a factor of 256.
Due to this bug the hardcoded CRC_PMIC_PWM_PERIOD_NS value of 21333 would
result in the PWM driver using its divider of 128, which would result in
a PWM output frequency of 6000000 Hz / 256 / 128 = 183 Hz. So actually
pretty close to the default VBT value of 200 Hz.
Now that this bug in the pwm-crc driver is fixed, we can actually use
the VBT defined frequency.
This is important because:
a) With the pwm-crc driver fixed it will now translate the hardcoded
CRC_PMIC_PWM_PERIOD_NS value of 21333 ns / 46 Khz to a PWM output
frequency of 23 KHz (the max it can do).
b) The pwm-lpss driver used on many models has always honored the
21333 ns / 46 Khz request
Some panels do not like such high output frequencies. E.g. on a Terra
Pad 1061 tablet, using the LPSS PWM controller, the backlight would go
from off to max, when changing the sysfs backlight brightness value from
90-100%, anything under aprox. 90% would turn the backlight fully off.
Honoring the VBT specified PWM frequency will also hopefully fix the
various bug reports which we have received about users perceiving the
backlight to flicker after a suspend/resume cycle.
Hans de Goede [Thu, 3 Sep 2020 11:23:31 +0000 (13:23 +0200)]
pwm: crc: Enable/disable PWM output on enable/disable
The pwm-crc code is using 2 different enable bits:
1. bit 7 of the PWM0_CLK_DIV (PWM_OUTPUT_ENABLE)
2. bit 0 of the BACKLIGHT_EN register
So far we've kept the PWM_OUTPUT_ENABLE bit set when disabling the PWM,
this commit makes crc_pwm_disable() clear it on disable and makes
crc_pwm_enable() set it again on re-enable.
Hans de Goede [Thu, 3 Sep 2020 11:23:30 +0000 (13:23 +0200)]
pwm: crc: Fix period changes not having any effect
The pwm-crc code is using 2 different enable bits:
1. bit 7 of the PWM0_CLK_DIV (PWM_OUTPUT_ENABLE)
2. bit 0 of the BACKLIGHT_EN register
The BACKLIGHT_EN register at address 0x51 really controls a separate
output-only GPIO which is earmarked to be used as output connected to the
backlight-enable pin for LCD panels, this GPO is part of the PMIC's
"Display Panel Control Block." . This pin should probably be moved over
to a GPIO provider driver (and consumers modified accordingly), but that
is something for an(other) patch.
Enabling / disabling the actual PWM output is controlled by the
PWM_OUTPUT_ENABLE bit of the PWM0_CLK_DIV register.
As the comment in the old code already indicates we must disable the PWM
before we can change the clock divider. But the crc_pwm_disable() and
crc_pwm_enable() calls the old code make for this only change the
BACKLIGHT_EN register; and the value of that register does not matter for
changing the period / the divider. What does matter is that the
PWM_OUTPUT_ENABLE bit must be cleared before a new value can be written.
This commit modifies crc_pwm_config() to clear PWM_OUTPUT_ENABLE instead
when changing the period, so that period changes actually work.
Note this fix will cause a significant behavior change on some devices
using the CRC PWM output to drive their backlight. Before the PWM would
always run with the output frequency configured by the BIOS at boot, now
the period time specified by the i915 driver will actually be honored.
Hans de Goede [Thu, 3 Sep 2020 11:23:29 +0000 (13:23 +0200)]
pwm: crc: Fix off-by-one error in the clock-divider calculations
The CRC PWM controller has a clock-divider which divides the clock with
a value between 1-128. But as can seen from the PWM_DIV_CLK_xxx
defines, this range maps to a register value of 0-127.
So after calculating the clock-divider we must subtract 1 to get the
register value, unless the requested frequency was so high that the
calculation has already resulted in a (rounded) divider value of 0.
Note that before this fix, setting a period of PWM_MAX_PERIOD_NS which
corresponds to the max. divider value of 128 could have resulted in a
bug where the code would use 128 as divider-register value which would
have resulted in an actual divider value of 0 (and the enable bit being
set). A rounding error stopped this bug from actually happen. This
same rounding error means that after the subtraction of 1 it is impossible
to set the divider to 128. Also bump PWM_MAX_PERIOD_NS by 1 ns to allow
setting a divider of 128 (register-value 127).
Hans de Goede [Thu, 3 Sep 2020 11:23:28 +0000 (13:23 +0200)]
pwm: crc: Fix period / duty_cycle times being off by a factor of 256
While looking into adding atomic-pwm support to the pwm-crc driver I
noticed something odd, there is a PWM_BASE_CLK define of 6 MHz and
there is a clock-divider which divides this with a value between 1-128,
and there are 256 duty-cycle steps.
The pwm-crc code before this commit assumed that a clock-divider
setting of 1 means that the PWM output is running at 6 MHZ, if that
is true, where do these 256 duty-cycle steps come from?
This would require an internal frequency of 256 * 6 MHz = 1.5 GHz, that
seems unlikely for a PMIC which is using a silicon process optimized for
power-switching transistors. It is way more likely that there is an 8
bit counter for the duty cycle which acts as an extra fixed divider
wrt the PWM output frequency.
The main user of the pwm-crc driver is the i915 GPU driver which uses it
for backlight control. Lets compare the PWM register values set by the
video-BIOS (the GOP), assuming the extra fixed divider is present versus
the PWM frequency specified in the Video-BIOS-Tables:
Device: PWM Hz set by BIOS PWM Hz specified in VBT
Asus T100TA 200 200
Asus T100HA 200 200
Lenovo Miix 2 8 23437 20000
Toshiba WT8-A 23437 20000
So as we can see if we assume the extra division by 256 then the register
values set by the GOP are an exact match for the VBT values, where as
otherwise the values would be of by a factor of 256.
This commit fixes the period / duty_cycle calculations to take the
extra division by 256 into account.
Hans de Goede [Thu, 3 Sep 2020 11:23:27 +0000 (13:23 +0200)]
pwm: lpss: Remove suspend/resume handlers
PWM controller drivers should not restore the PWM state on resume. The
convention is that PWM consumers do this by calling pwm_apply_state(),
so that it can be done at the exact moment when the consumer needs
the state to be stored, avoiding e.g. backlight flickering.
The only in kernel consumers of the pwm-lpss code, the i915 driver
and the pwm-class sysfs interface code both correctly restore the
state on resume, so there is no need to do this in the pwm-lpss code.
More-over the removed resume handler is buggy, since it blindly
restores the ctrl-register contents without setting the update
bit, which is necessary to get the controller to actually use/apply
the restored base-unit and on-time-div values.
Hans de Goede [Thu, 3 Sep 2020 11:23:26 +0000 (13:23 +0200)]
pwm: lpss: Make pwm_lpss_apply() not rely on existing hardware state
Before this commit pwm_lpss_apply() was assuming 2 pre-conditions
were met by the existing hardware state:
1. That the base-unit and on-time-div read back from the
control register are those actually in use, so that it
can skip setting the update bit if the read-back value
matches the desired values.
2. That the controller is enabled when the cached
pwm_state.enabled says that the controller is enabled.
As the long history of fixes for subtle (often suspend/resume)
lpss-pwm issues shows, these assumptions are not necessary
always true.
1. Specifically is not true on some (*) Cherry Trail devices
with a nasty GFX0._PS3 method which: a. saves the ctrl reg value.
b. sets the base-unit to 0 and writes the update bit to apply/commit
c. restores the original ctrl value without setting the update bit,
so that the 0 base-unit value is still in use.
2. Assumption 2. currently is true, but only because of the code which
saves/restores the state on suspend/resume. By convention restoring the
PWM state should be done by the PWM consumer and the presence of this
code in the pmw-lpss driver is a bug. Therefor the save/restore code will
be dropped in the next patch in this series, after which this assumption
also is no longer true.
This commit changes the pwm_lpss_apply() to not make any assumptions about
the state the hardware is in. Instead it makes pwm_lpss_apply() always
fully program the PWM controller, making it much less fragile.
*) Seen on the Acer One 10 S1003, Lenovo Ideapad Miix 310 and 320 models
and various Medion models.
Hans de Goede [Thu, 3 Sep 2020 11:23:25 +0000 (13:23 +0200)]
pwm: lpss: Add pwm_lpss_prepare_enable() helper
In the not-enabled -> enabled path pwm_lpss_apply() needs to get a
runtime-pm reference; and then on any errors it needs to release it
again.
This leads to somewhat hard to read code. This commit introduces a new
pwm_lpss_prepare_enable() helper and moves all the steps necessary for
the not-enabled -> enabled transition there, so that we can error check
the entire transition in a single place and only have one pm_runtime_put()
on failure call site.
While working on this I noticed that the enabled -> enabled (update
settings) path was quite similar, so I've added an enable parameter to
the new pwm_lpss_prepare_enable() helper, which allows using it in that
path too.
Hans de Goede [Thu, 3 Sep 2020 11:23:24 +0000 (13:23 +0200)]
pwm: lpss: Add range limit check for the base_unit register value
When the user requests a high enough period ns value, then the
calculations in pwm_lpss_prepare() might result in a base_unit value of 0.
But according to the data-sheet the way the PWM controller works is that
each input clock-cycle the base_unit gets added to a N bit counter and
that counter overflowing determines the PWM output frequency. Adding 0
to the counter is a no-op. The data-sheet even explicitly states that
writing 0 to the base_unit bits will result in the PWM outputting a
continuous 0 signal.
When the user requestes a low enough period ns value, then the
calculations in pwm_lpss_prepare() might result in a base_unit value
which is bigger then base_unit_range - 1. Currently the codes for this
deals with this by applying a mask:
base_unit &= (base_unit_range - 1);
But this means that we let the value overflow the range, we throw away the
higher bits and store whatever value is left in the lower bits into the
register leading to a random output frequency, rather then clamping the
output frequency to the highest frequency which the hardware can do.
This commit fixes both issues by clamping the base_unit value to be
between 1 and (base_unit_range - 1).
Hans de Goede [Thu, 3 Sep 2020 11:23:23 +0000 (13:23 +0200)]
pwm: lpss: Fix off by one error in base_unit math in pwm_lpss_prepare()
According to the data-sheet the way the PWM controller works is that
each input clock-cycle the base_unit gets added to a N bit counter and
that counter overflowing determines the PWM output frequency.
So assuming e.g. a 16 bit counter this means that if base_unit is set to 1,
after 65535 input clock-cycles the counter has been increased from 0 to
65535 and it will overflow on the next cycle, so it will overflow after
every 65536 clock cycles and thus the calculations done in
pwm_lpss_prepare() should use 65536 and not 65535.
This commit fixes this. Note this also aligns the calculations in
pwm_lpss_prepare() with those in pwm_lpss_get_state().
Note this effectively reverts commit 684309e5043e ("pwm: lpss: Avoid
potential overflow of base_unit"). The next patch in this series really
fixes the potential overflow of the base_unit value.
Fixes: 684309e5043e ("pwm: lpss: Avoid potential overflow of base_unit") Reviewed-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com> Acked-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de> Acked-by: Thierry Reding <thierry.reding@gmail.com> Signed-off-by: Hans de Goede <hdegoede@redhat.com> Link: https://patchwork.freedesktop.org/patch/msgid/20200903112337.4113-4-hdegoede@redhat.com
Where PSAT is the power-status register of the PWM controller.
Since the i915 driver will do a pwm_get on the pwm device as it uses it to
control the LCD panel backlight, there is a device-link marking the i915
device as a consumer of the pwm device. So that the PWM controller will
always be suspended after the i915 driver suspends (which is the right
thing to do). This causes the above GFX0 PS3 AML code to run before
acpi_lpss.c calls acpi_lpss_save_ctx().
So on these devices the PWM controller will already be off when
acpi_lpss_save_ctx() runs. This causes it to read/save all 1-s (0xffffffff)
as ctx register values.
When these bogus values get restored on resume the PWM controller actually
keeps working, since most bits are reserved, but this does set bit 3 of
the LPSS General purpose register, which for the PWM controller has the
following function: "This bit is re-used to support 32kHz slow mode.
Default is 19.2MHz as PWM source clock".
This causes the clock of the PWM controller to switch from 19.2MHz to
32KHz, which is a slow-down of a factor 600. Surprisingly enough so far
there have been few bug reports about this. This is likely because the
i915 driver was hardcoding the PWM frequency to 46 KHz, which divided
by 600 would result in a PWM frequency of approx. 78 Hz, which mostly
still works fine. There are some bug reports about the LCD backlight
flickering after suspend/resume which are likely caused by this issue.
But with the upcoming patch-series to finally switch the i915 drivers
code for external PWM controllers to use the atomic API and to honor
the PWM frequency specified in the video BIOS (VBT), this becomes a much
bigger problem. On most cases the VBT specifies either 200 Hz or 20
KHz as PWM frequency, which with the mentioned issue ends up being either
1/3 Hz, where the backlight actually visible blinks on and off every 3s,
or in 33 Hz and horrible flickering of the backlight.
There are a number of possible solutions to this problem:
1. Make acpi_lpss_save_ctx() run before GFX0._PS3
Pro: Clean solution from pov of not medling with save/restore ctx code
Con: As mentioned the current ordering is the right thing to do
Con: Requires assymmetry in at what suspend/resume phase we do the save vs
restore, requiring more suspend/resume ordering hacks in already
convoluted acpi_lpss.c suspend/resume code.
2. Do some sort of save once mode for the LPSS ctx
Pro: Reasonably clean
Con: Needs a new LPSS flag + code changes to handle the flag
3. Detect we have failed to save the ctx registers and do not restore them
Pro: Not PWM specific, might help with issues on other LPSS devices too
Con: If we can get away with not restoring the ctx why bother with it at
all?
4. Do not save the ctx for CHT PWM controllers
Pro: Clean, as simple as dropping a flag?
Con: Not so simple as dropping a flag, needs a new flag to ensure that
we still do lpss_deassert_reset() on device activation.
5. Make the pwm-lpss code fixup the LPSS-context registers
Pro: Keeps acpi_lpss.c code clean
Con: Moves knowledge of LPSS-context into the pwm-lpss.c code
1 and 5 both do not seem to be a desirable way forward.
3 and 4 seem ok, but they both assume that restoring the LPSS-context
registers is not necessary. I have done a couple of test and those do
show that restoring the LPSS-context indeed does not seem to be necessary
on devices using s2idle suspend (and successfully reaching S0i3). But I
have no hardware to test deep / S3 suspend. So I'm not sure that not
restoring the context is safe.
That leaves solution 2, which is about as simple / clean as 3 and 4,
so this commit fixes the described problem by implementing a new
LPSS_SAVE_CTX_ONCE flag and setting that for the CHT PWM controllers.
The DSDTs on most Cherry Trail devices have an ugly clutch where the PWM
controller gets poked from the _PS0 method of the graphics-card device:
Local0 = PSAT /* \_SB_.PCI0.GFX0.PSAT */
If (((Local0 & 0x03) == 0x03))
{
PSAT &= 0xFFFFFFFC
Local1 = PSAT /* \_SB_.PCI0.GFX0.PSAT */
RSTA = Zero
RSTF = Zero
RSTA = One
RSTF = One
PWMB |= 0xC0000000
PWMC = PWMB /* \_SB_.PCI0.GFX0.PWMB */
}
Where PSAT is the power-status register of the PWM controller, so if it
is in D3 when the GFX0 device's PS0 method runs then it will turn it on
and restore the PWM ctrl register value it saved from its PS3 handler.
Note not only does it restore it, it ors it with 0xC0000000 turning it
on at a time where we may not want it to get turned on at all.
The pwm_get call which the i915 driver does to get a reference to the
PWM controller, already adds a device-link making the GFX0 device a
consumer of the PWM device. So it should already have been resumed when
the above AML runs and the AML should thus not do its undesirable poking
of the PWM controller register.
But the PCI core powers on PCI devices in the no-irq resume phase and
thus calls the troublesome PS0 method in the no-irq resume phase.
Where as LPSS devices by default are resumed in the early resume phase.
This commit sets the resume_from_noirq flag in the bsw_pwm_dev_desc
struct, so that Cherry Trail PWM controllers will be resumed in the
no-irq phase. Together with the device-link added by the pwm-get this
ensures that the PWM controller will be on when the troublesome PS0
method runs, which stops it from poking the PWM controller.
Merge tags 'auxdisplay-for-linus-v5.9-rc4', 'clang-format-for-linus-v5.9-rc4' and 'compiler-attributes-for-linus-v5.9-rc4' of git://github.com/ojeda/linux
Pull misc fixes from Miguel Ojeda:
"A trivial patch for auxdisplay:
- Replace HTTP links with HTTPS ones (Alexander A. Klimov)
The usual clang-format trivial update:
- Update with the latest for_each macro list (Miguel Ojeda)
And Luc requested me to pick a sparse fix on my queue, so here it goes
along with other two trivial Compiler Attributes ones (also from Luc).
- sparse: use static inline for __chk_{user,io}_ptr() (Luc Van
Oostenryck)
- Compiler Attributes: remove comment about sparse not supporting
__has_attribute (Luc Van Oostenryck)"
* tag 'auxdisplay-for-linus-v5.9-rc4' of git://github.com/ojeda/linux:
auxdisplay: Replace HTTP links with HTTPS ones
* tag 'clang-format-for-linus-v5.9-rc4' of git://github.com/ojeda/linux:
clang-format: Update with the latest for_each macro list
* tag 'compiler-attributes-for-linus-v5.9-rc4' of git://github.com/ojeda/linux:
sparse: use static inline for __chk_{user,io}_ptr()
Compiler Attributes: fix comment concerning GCC 4.6
Compiler Attributes: remove comment about sparse not supporting __has_attribute
Merge tag 'arc-5.9-rc4' of git://git.kernel.org/pub/scm/linux/kernel/git/vgupta/arc
Pull ARC fixes from Vineet Gupta:
- HSDK-4xd Dev system: perf driver updates for sampling interrupt
- HSDK* Dev System: Ethernet broken [Evgeniy Didin]
- HIGHMEM broken (2 memory banks) [Mike Rapoport]
- show_regs() rewrite once and for all
- Other minor fixes
* tag 'arc-5.9-rc4' of git://git.kernel.org/pub/scm/linux/kernel/git/vgupta/arc:
ARC: [plat-hsdk]: Switch ethernet phy-mode to rgmii-id
arc: fix memory initialization for systems with two memory banks
irqchip/eznps: Fix build error for !ARC700 builds
ARC: show_regs: fix r12 printing and simplify
ARC: HSDK: wireup perf irq
ARC: perf: don't bail setup if pct irq missing in device-tree
ARC: pgalloc.h: delete a duplicated word + other fixes
projects / kernel.git / log