stop_machine is not really a locking primitive we should use, except
when the hw folks tell us the hw is broken and that's the only way to
work around it.
This patch tries to address the locking abuse of stop_machine() from
commit 20e4933c47
Author: Chris Wilson <chris@chris-wilson.co.uk>
Date: Tue Nov 22 14:41:21 2016 +0000
drm/i915: Stop the machine as we install the wedged submit_request handler
Chris said parts of the reasons for going with stop_machine() was that
it's no overhead for the fast-path. But these callbacks use irqsave
spinlocks and do a bunch of MMIO, and rcu_read_lock is _real_ fast.
To stay as close as possible to the stop_machine semantics we first
update all the submit function pointers to the nop handler, then call
synchronize_rcu() to make sure no new requests can be submitted. This
should give us exactly the huge barrier we want.
I pondered whether we should annotate engine->submit_request as __rcu
and use rcu_assign_pointer and rcu_dereference on it. But the reason
behind those is to make sure the compiler/cpu barriers are there for
when you have an actual data structure you point at, to make sure all
the writes are seen correctly on the read side. But we just have a
function pointer, and .text isn't changed, so no need for these
barriers and hence no need for annotations.
Unfortunately there's a complication with the call to
intel_engine_init_global_seqno:
- Without stop_machine we must hold the corresponding spinlock.
- Without stop_machine we must ensure that all requests are marked as
having failed with dma_fence_set_error() before we call it. That
means we need to split the nop request submission into two phases,
both synchronized with rcu:
1. Only stop submitting the requests to hw and mark them as failed.
2. After all pending requests in the scheduler/ring are suitably
marked up as failed and we can force complete them all, also force
complete by calling intel_engine_init_global_seqno().
This should fix the followwing lockdep splat:
======================================================
WARNING: possible circular locking dependency detected
4.14.0-rc3-CI-CI_DRM_3179+ #1 Tainted: G U
------------------------------------------------------
kworker/3:4/562 is trying to acquire lock:
(cpu_hotplug_lock.rw_sem){++++}, at: [<ffffffff8113d4bc>] stop_machine+0x1c/0x40
but task is already holding lock:
(&dev->struct_mutex){+.+.}, at: [<ffffffffa0136588>] i915_reset_device+0x1e8/0x260 [i915]
which lock already depends on the new lock.
the existing dependency chain (in reverse order) is:
-> #6 (&dev->struct_mutex){+.+.}:
__lock_acquire+0x1420/0x15e0
lock_acquire+0xb0/0x200
__mutex_lock+0x86/0x9b0
mutex_lock_interruptible_nested+0x1b/0x20
i915_mutex_lock_interruptible+0x51/0x130 [i915]
i915_gem_fault+0x209/0x650 [i915]
__do_fault+0x1e/0x80
__handle_mm_fault+0xa08/0xed0
handle_mm_fault+0x156/0x300
__do_page_fault+0x2c5/0x570
do_page_fault+0x28/0x250
page_fault+0x22/0x30
-> #5 (&mm->mmap_sem){++++}:
__lock_acquire+0x1420/0x15e0
lock_acquire+0xb0/0x200
__might_fault+0x68/0x90
_copy_to_user+0x23/0x70
filldir+0xa5/0x120
dcache_readdir+0xf9/0x170
iterate_dir+0x69/0x1a0
SyS_getdents+0xa5/0x140
entry_SYSCALL_64_fastpath+0x1c/0xb1
-> #4 (&sb->s_type->i_mutex_key#5){++++}:
down_write+0x3b/0x70
handle_create+0xcb/0x1e0
devtmpfsd+0x139/0x180
kthread+0x152/0x190
ret_from_fork+0x27/0x40
-> #3 ((complete)&req.done){+.+.}:
__lock_acquire+0x1420/0x15e0
lock_acquire+0xb0/0x200
wait_for_common+0x58/0x210
wait_for_completion+0x1d/0x20
devtmpfs_create_node+0x13d/0x160
device_add+0x5eb/0x620
device_create_groups_vargs+0xe0/0xf0
device_create+0x3a/0x40
msr_device_create+0x2b/0x40
cpuhp_invoke_callback+0xc9/0xbf0
cpuhp_thread_fun+0x17b/0x240
smpboot_thread_fn+0x18a/0x280
kthread+0x152/0x190
ret_from_fork+0x27/0x40
-> #2 (cpuhp_state-up){+.+.}:
__lock_acquire+0x1420/0x15e0
lock_acquire+0xb0/0x200
cpuhp_issue_call+0x133/0x1c0
__cpuhp_setup_state_cpuslocked+0x139/0x2a0
__cpuhp_setup_state+0x46/0x60
page_writeback_init+0x43/0x67
pagecache_init+0x3d/0x42
start_kernel+0x3a8/0x3fc
x86_64_start_reservations+0x2a/0x2c
x86_64_start_kernel+0x6d/0x70
verify_cpu+0x0/0xfb
-> #1 (cpuhp_state_mutex){+.+.}:
__lock_acquire+0x1420/0x15e0
lock_acquire+0xb0/0x200
__mutex_lock+0x86/0x9b0
mutex_lock_nested+0x1b/0x20
__cpuhp_setup_state_cpuslocked+0x53/0x2a0
__cpuhp_setup_state+0x46/0x60
page_alloc_init+0x28/0x30
start_kernel+0x145/0x3fc
x86_64_start_reservations+0x2a/0x2c
x86_64_start_kernel+0x6d/0x70
verify_cpu+0x0/0xfb
-> #0 (cpu_hotplug_lock.rw_sem){++++}:
check_prev_add+0x430/0x840
__lock_acquire+0x1420/0x15e0
lock_acquire+0xb0/0x200
cpus_read_lock+0x3d/0xb0
stop_machine+0x1c/0x40
i915_gem_set_wedged+0x1a/0x20 [i915]
i915_reset+0xb9/0x230 [i915]
i915_reset_device+0x1f6/0x260 [i915]
i915_handle_error+0x2d8/0x430 [i915]
hangcheck_declare_hang+0xd3/0xf0 [i915]
i915_hangcheck_elapsed+0x262/0x2d0 [i915]
process_one_work+0x233/0x660
worker_thread+0x4e/0x3b0
kthread+0x152/0x190
ret_from_fork+0x27/0x40
other info that might help us debug this:
Chain exists of:
cpu_hotplug_lock.rw_sem --> &mm->mmap_sem --> &dev->struct_mutex
Possible unsafe locking scenario:
CPU0 CPU1
---- ----
lock(&dev->struct_mutex);
lock(&mm->mmap_sem);
lock(&dev->struct_mutex);
lock(cpu_hotplug_lock.rw_sem);
*** DEADLOCK ***
3 locks held by kworker/3:4/562:
#0: ("events_long"){+.+.}, at: [<ffffffff8109c64a>] process_one_work+0x1aa/0x660
#1: ((&(&i915->gpu_error.hangcheck_work)->work)){+.+.}, at: [<ffffffff8109c64a>] process_one_work+0x1aa/0x660
#2: (&dev->struct_mutex){+.+.}, at: [<ffffffffa0136588>] i915_reset_device+0x1e8/0x260 [i915]
stack backtrace:
CPU: 3 PID: 562 Comm: kworker/3:4 Tainted: G U 4.14.0-rc3-CI-CI_DRM_3179+ #1
Hardware name: /NUC7i5BNB, BIOS BNKBL357.86A.0048.2017.0704.1415 07/04/2017
Workqueue: events_long i915_hangcheck_elapsed [i915]
Call Trace:
dump_stack+0x68/0x9f
print_circular_bug+0x235/0x3c0
? lockdep_init_map_crosslock+0x20/0x20
check_prev_add+0x430/0x840
? irq_work_queue+0x86/0xe0
? wake_up_klogd+0x53/0x70
__lock_acquire+0x1420/0x15e0
? __lock_acquire+0x1420/0x15e0
? lockdep_init_map_crosslock+0x20/0x20
lock_acquire+0xb0/0x200
? stop_machine+0x1c/0x40
? i915_gem_object_truncate+0x50/0x50 [i915]
cpus_read_lock+0x3d/0xb0
? stop_machine+0x1c/0x40
stop_machine+0x1c/0x40
i915_gem_set_wedged+0x1a/0x20 [i915]
i915_reset+0xb9/0x230 [i915]
i915_reset_device+0x1f6/0x260 [i915]
? gen8_gt_irq_ack+0x170/0x170 [i915]
? work_on_cpu_safe+0x60/0x60
i915_handle_error+0x2d8/0x430 [i915]
? vsnprintf+0xd1/0x4b0
? scnprintf+0x3a/0x70
hangcheck_declare_hang+0xd3/0xf0 [i915]
? intel_runtime_pm_put+0x56/0xa0 [i915]
i915_hangcheck_elapsed+0x262/0x2d0 [i915]
process_one_work+0x233/0x660
worker_thread+0x4e/0x3b0
kthread+0x152/0x190
? process_one_work+0x660/0x660
? kthread_create_on_node+0x40/0x40
ret_from_fork+0x27/0x40
Setting dangerous option reset - tainting kernel
i915 0000:00:02.0: Resetting chip after gpu hang
Setting dangerous option reset - tainting kernel
i915 0000:00:02.0: Resetting chip after gpu hang
v2: Have 1 global synchronize_rcu() barrier across all engines, and
improve commit message.
v3: We need to protect the seqno update with the timeline spinlock (in
set_wedged) to avoid racing with other updates of the seqno, like we
already do in nop_submit_request (Chris).
v4: Use two-phase sequence to plug the race Chris spotted where we can
complete requests before they're marked up with -EIO.
v5: Review from Chris:
- simplify nop_submit_request.
- Add comment to rcu_read_lock section.
- Align comments with the new style.
v6: Remove unused variable to appease CI.
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=102886
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=103096
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Mika Kuoppala <mika.kuoppala@intel.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Marta Lofstedt <marta.lofstedt@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20171011091019.1425-1-daniel.vetter@ffwll.ch
Add another perma-pinned context for using for preemption at any time.
We cannot just reuse the existing kernel context, as first and foremost
we need to ensure that we can preempt the kernel context itself, so
require a distinct context id. Similar to the kernel context, we may
want to interrupt execution and switch to the preempt context at any
time, and so it needs to be permanently pinned and available.
To compensate for yet another permanent allocation, we shrink the
existing context and the new context by reducing their ringbuffer to the
minimum.
v2: Assert that we never allocate a request from the preemption context.
v3: Limit perma-pin to engines that may preempt.
v4: Onion cleanup for early driver death
v5: Onion ordering in main driver cleanup as well.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Michał Winiarski <michal.winiarski@intel.com>
Reviewed-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20171003203453.15692-4-chris@chris-wilson.co.uk
In a synchronous setup, we may retire the last request before we
complete allocating the next request. As the last request is retired, we
queue a timer to mark the device as idle, and promptly have to execute
ad cancel that timer once we complete allocating the request and need to
keep the device awake. If we rearrange the mark_busy() to occur before
we retire the previous request, we can skip this ping-pong.
v2: Joonas pointed out that unreserve_seqno() was now doing more than
doing seqno handling and should be renamed to reflect its wider purpose.
That also highlighted the new asymmetry with reserve_seqno(), so fixup
that and rename both to [un]reserve_engine().
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20170817144719.10968-1-chris@chris-wilson.co.uk
Reviewed-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
intel_engine_init_globa_seqno() may be called from an uncontrolled
set-wedged path where we have given up waiting for broken hw and declare
it defunct. Along that path, any sanity checks that the hw is idle
before we adjust its state will expectedly fail, so we simply cannot.
Instead of asserting inside init_global_seqno, we move them to the
normal caller reset_all_global_seqno() as it handles runtime seqno
wraparound.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Mika Kuoppala <mika.kuoppala@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20170721123238.16428-8-chris@chris-wilson.co.uk
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Once a client has requested a waitboost, we keep that waitboost active
until all clients are no longer waiting. This is because we don't
distinguish which waiter deserves the boost. However, with the advent of
fence signaling, the signaler threads appear as waiters to the RPS
interrupt handler. So instead of using a single boolean to track when to
keep the waitboost active, use a counter of all outstanding waitboosted
requests.
At this point, I have removed all vestiges of the rate limiting on
clients. Whilst this means that compositors should remain more fluid,
it also means that boosts are more prevalent. See commit b29c19b645
("drm/i915: Boost RPS frequency for CPU stalls") for a longer discussion
on the pros and cons of both approaches.
A drawback of this implementation is that it requires constant request
submission to keep the waitboost trimmed (as it is now cancelled when the
request is completed). This will be fine for a busy system, but near
idle the boosts may be kept for longer than desired (effectively tens of
vblanks worstcase) and there is a reliance on rc6 instead.
v2: Remove defunct rps.client_lock
Reported-by: Michał Winiarski <michal.winiarski@intel.com>
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Michał Winiarski <michal.winiarski@intel.com>
Reviewed-by: Michał Winiarski <michal.winiarski@intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/20170628123548.9236-1-chris@chris-wilson.co.uk
Originally we would enable and disable the breadcrumb interrupt
immediately on demand. This was slow enough to have a large impact
(>30%) on tasks that hopped between engines. However, by using a shadow
to keep the irq alive for an extra interrupt (see commit 67b807a892
("drm/i915: Delay disabling the user interrupt for breadcrumbs")) and
by recently reducing the cost in adding ourselves to the signal tree, we
no longer need to spin-request during await_request to avoid delays in
throughput tests. Without the earlier patches to stop the wakeup when
signaling if the irq was already active, we saw no improvement in
execbuf overhead (and corresponding contention in other clients) despite
the removal of the spinner in a simple test like glxgears. This means
there will be scenarios where now we spend longer enabling the interrupt
than we would have spent spinning, but these are not likely to have as
noticeable an impact as the high frequency test cases (where there
should not be any regression).
Ulterior motive: generalising the engine->sync_to to handle different
types of semaphores and non-semaphores.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Cc: Oscar Mateo <oscar.mateo@intel.com>
Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Reviewed-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/20170608111405.16466-4-chris@chris-wilson.co.uk
All the requests at the same priority are executed in FIFO order. They
do not need to be stored in the rbtree themselves, as they are a simple
list within a level. If we move the requests at one priority into a list,
we can then reduce the rbtree to the set of priorities. This should keep
the height of the rbtree small, as the number of active priorities can not
exceed the number of active requests and should be typically only a few.
Currently, we have ~2k possible different priority levels, that may
increase to allow even more fine grained selection. Allocating those in
advance seems a waste (and may be impossible), so we opt for allocating
upon first use, and freeing after its requests are depleted. To avoid
the possibility of an allocation failure causing us to lose a request,
we preallocate the default priority (0) and bump any request to that
priority if we fail to allocate it the appropriate plist. Having a
request (that is ready to run, so not leading to corruption) execute
out-of-order is better than leaking the request (and its dependency
tree) entirely.
There should be a benefit to reducing execlists_dequeue() to principally
using a simple list (and reducing the frequency of both rbtree iteration
and balancing on erase) but for typical workloads, request coalescing
should be small enough that we don't notice any change. The main gain is
from improving PI calls to schedule, and the explicit list within a
level should make request unwinding simpler (we just need to insert at
the head of the list rather than the tail and not have to make the
rbtree search more complicated).
v2: Avoid use-after-free when deleting a depleted priolist
v3: Michał found the solution to handling the allocation failure
gracefully. If we disable all priority scheduling following the
allocation failure, those requests will be executed in fifo and we will
ensure that this request and its dependencies are in strict fifo (even
when it doesn't realise it is only a single list). Normal scheduling is
restored once we know the device is idle, until the next failure!
Suggested-by: Michał Wajdeczko <michal.wajdeczko@intel.com>
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Michał Winiarski <michal.winiarski@intel.com>
Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Reviewed-by: Michał Winiarski <michal.winiarski@intel.com>
Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/20170517121007.27224-8-chris@chris-wilson.co.uk
Since unifying ringbuffer/execlist submission to use
engine->pin_context, we ensure that the intel_ring is available before
we start constructing the request. We can therefore move the assignment
of the request->ring to the central i915_gem_request_alloc() and not
require it in every engine->request_alloc() callback. Another small step
towards simplification (of the core, but at a cost of handling error
pointers in less important callers of engine->pin_context).
v2: Rearrange a few branches to reduce impact of PTR_ERR() on gcc's code
generation.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Oscar Mateo <oscar.mateo@intel.com>
Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Reviewed-by: Oscar Mateo <oscar.mateo@intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/20170504093308.4137-1-chris@chris-wilson.co.uk
Track the latest fence waited upon on each context, and only add a new
asynchronous wait if the new fence is more recent than the recorded
fence for that context. This requires us to filter out unordered
timelines, which are noted by DMA_FENCE_NO_CONTEXT. However, in the
absence of a universal identifier, we have to use our own
i915->mm.unordered_timeline token.
v2: Throw around the debug crutches
v3: Inline the likely case of the pre-allocation cache being full.
v4: Drop the pre-allocation support, we can lose the most recent fence
in case of allocation failure -- it just means we may emit more awaits
than strictly necessary but will not break.
v5: Trim allocation size for leaf nodes, they only need an array of u32
not pointers.
v6: Create mock_timeline to tidy selftest writing
v7: s/intel_timeline_sync_get/intel_timeline_sync_is_later/ (Tvrtko)
v8: Prune the stale sync points when we idle.
v9: Include a small benchmark in the kselftests
v10: Separate the idr implementation into its own compartment. (Tvrkto)
v11: Refactor igt_sync kselftests to avoid deep nesting (Tvrkto)
v12: __sync_leaf_idx() to assert that p->height is 0 when checking leaves
v13: kselftests to investigate struct i915_syncmap itself (Tvrtko)
v14: Foray into ascii art graphs
v15: Take into account that the random lookup/insert does 2 prng calls,
not 1, when benchmarking, and use for_each_set_bit() (Tvrtko)
v16: Improved ascii art
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/20170503093924.5320-4-chris@chris-wilson.co.uk
When we retire the last request on the ring, before we ever access that
ring again we know it will be completely idle and so we can advance the
ring->head fully to the end (i.e. ring->tail) and not just to the start
of the breadcrumb. This allows us to skip re-emitting the breadcrumb
after resetting the GPU if the ring was entirely idle. This prevents us
from overwriting a seqno wraparound by re-executing a stale breadcrumb,
i.e.
submit_request(1)
intel_engine_init_global_seqno(0)
i915_reset()
would then leave 1 in the HWS, but the next request to execute would
also be with seqno 1. The sanity checks upon submission detect this as a
timewarp and explode. By setting the ring as empty, upon reset the HWS
is left as 0, leaving it consistent with the timeline.
v2: Fix check for deleting last element of list. We know that this
request is always the first element of the ring, so only if next
points back to the start will this be the only request in flight.
v3: Remove opencoding of list_is_last()
v4: Move the block to its own function for some clarity.
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=100144
Testcase: igt/gem_exec_whisper/hang-*
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Mika Kuoppala <mika.kuoppala@intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/20170406170028.26871-1-chris@chris-wilson.co.uk
Reviewed-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Michał Winiarski pointed out that the debugging infrastructure (such as
trace_dma_fence_release) likes to pretty print the timeline name, long
after we have freed the timeline. Our timelines currently live as part of
the GTT (due to the strict ordering we currently use through each) which
belong to the context. We aim to free the context and release its
hardware resources as soon as we able to (i.e. when the last
fence/request using it has been signaled and retired). As the
.get_timeline_name is purely a debug feature, rather than extending the
lifetime of the context, or splitting it into many different release
phases just to keep the name around, replace the timeline name with a
constant after the fence has been signaled. This avoids the potential
use-after-free.
Reported-by: Krzysztof Olinski <krzysztof.e.olinski@intel.com>
Fixes: 80b204bce8 ("drm/i915: Enable multiple timelines")
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Michał Winiarski <michal.winiarski@intel.com>
Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Cc: <stable@vger.kernel.org> # v4.10+
Link: http://patchwork.freedesktop.org/patch/msgid/20170330111614.29757-1-chris@chris-wilson.co.uk
Reviewed-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Reviewed-by: Michał Winiarski <michal.winiarski@intel.com>
I915_RESET_IN_PROGRESS is being used for both signaling the requirement
to i915_mutex_lock_interruptible() to avoid taking the struct_mutex and
to instruct a waiter (already holding the struct_mutex) to perform the
reset. To allow for a little more coordination, split these two meaning
into a couple of distinct flags. I915_RESET_BACKOFF tells
i915_mutex_lock_interruptible() not to acquire the mutex and
I915_RESET_HANDOFF tells the waiter to call i915_reset().
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Acked-by: Michel Thierry <michel.thierry@intel.com>
Reviewed-by: Mika Kuoppala <mika.kuoppala@intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/20170316171305.12972-1-chris@chris-wilson.co.uk
4 weeks worth of stuff since I was traveling&lazy:
- lspcon improvements (Imre)
- proper atomic state for cdclk handling (Ville)
- gpu reset improvements (Chris)
- lots and lots of polish around fences, requests, waiting and
everything related all over (both gem and modeset code), from Chris
- atomic by default on gen5+ minus byt/bsw (Maarten did the patch to
flip the default, really this is a massive joint team effort)
- moar power domains, now 64bit (Ander)
- big pile of in-kernel unit tests for various gem subsystems (Chris),
including simple mock objects for i915 device and and the ggtt
manager.
- i915_gpu_info in debugfs, for taking a snapshot of the current gpu
state. Same thing as i915_error_state, but useful if the kernel didn't
notice something is stick. From Chris.
- bxt dsi fixes (Umar Shankar)
- bxt w/a updates (Jani)
- no more struct_mutex for gem object unreference (Chris)
- some execlist refactoring (Tvrtko)
- color manager support for glk (Ander)
- improve the power-well sync code to better take over from the
firmware (Imre)
- gem tracepoint polish (Tvrtko)
- lots of glk fixes all around (Ander)
- ctx switch improvements (Chris)
- glk dsi support&fixes (Deepak M)
- dsi fixes for vlv and clanups, lots of them (Hans de Goede)
- switch to i915.ko types in lots of our internal modeset code (Ander)
- byt/bsw atomic wm update code, yay (Ville)
* tag 'drm-intel-next-2017-03-06' of git://anongit.freedesktop.org/git/drm-intel: (432 commits)
drm/i915: Update DRIVER_DATE to 20170306
drm/i915: Don't use enums for hardware engine id
drm/i915: Split breadcrumbs spinlock into two
drm/i915: Refactor wakeup of the next breadcrumb waiter
drm/i915: Take reference for signaling the request from hardirq
drm/i915: Add FIFO underrun tracepoints
drm/i915: Add cxsr toggle tracepoint
drm/i915: Add VLV/CHV watermark/FIFO programming tracepoints
drm/i915: Add plane update/disable tracepoints
drm/i915: Kill level 0 wm hack for VLV/CHV
drm/i915: Workaround VLV/CHV sprite1->sprite0 enable underrun
drm/i915: Sanitize VLV/CHV watermarks properly
drm/i915: Only use update_wm_{pre,post} for pre-ilk platforms
drm/i915: Nuke crtc->wm.cxsr_allowed
drm/i915: Compute proper intermediate wms for vlv/cvh
drm/i915: Skip useless watermark/FIFO related work on VLV/CHV when not needed
drm/i915: Compute vlv/chv wms the atomic way
drm/i915: Compute VLV/CHV FIFO sizes based on the PM2 watermarks
drm/i915: Plop vlv/chv fifo sizes into crtc state
drm/i915: Plop vlv wm state into crtc_state
...
During reset_all_global_seqno() on seqno rollover, we have to update the
HWS. This causes all in flight requests to be completed, so first we
wait. However, we were only waiting for the requests themselves to be
completed and clearing out the waiter rbtrees - what I had missed was
the extra reference in execlists->port[]. Since commit fe9ae7a3bf
("drm/i915/execlists: Detect an out-of-order context switch") we can
detect when the request is retired before the context switch interrupt
is completed. The impact should be neglible outside of debugging.
Testcase: igt/gem_exec_whisper
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Mika Kuoppala <mika.kuoppala@intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/20170303121947.20482-1-chris@chris-wilson.co.uk
Reviewed-by: Mika Kuoppala <mika.kuoppala@intel.com>
Adding to the tail of the client request list as the only other user is
in the throttle ioctl that iterates forwards over the list. It only
needs protection against deletion of a request as it reads it, it simply
won't see a new request added to the end of the list, or it would be too
early and rejected. We can further reduce the number of spinlocks
required when throttling by removing stale requests from the client_list
as we throttle.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Mika Kuoppala <mika.kuoppala@intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/20170302122525.19675-1-chris@chris-wilson.co.uk
Instead of including the full <linux/signal.h>, we are going to include the
types-only <linux/signal_types.h> header in <linux/sched.h>, to further
decouple the scheduler header from the signal headers.
This means that various files which relied on the full <linux/signal.h> need
to be updated to gain an explicit dependency on it.
Update the code that relies on sched.h's inclusion of the <linux/signal.h> header.
Acked-by: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: linux-kernel@vger.kernel.org
Signed-off-by: Ingo Molnar <mingo@kernel.org>
We are going to split <linux/sched/clock.h> out of <linux/sched.h>, which
will have to be picked up from other headers and .c files.
Create a trivial placeholder <linux/sched/clock.h> file that just
maps to <linux/sched.h> to make this patch obviously correct and
bisectable.
Include the new header in the files that are going to need it.
Acked-by: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: linux-kernel@vger.kernel.org
Signed-off-by: Ingo Molnar <mingo@kernel.org>
As execlists and other non-semaphore multi-engine devices coordinate
between engines using interrupts, we can shave off a few 10s of
microsecond of scheduling latency by doing the fence signaling from the
interrupt as opposed to a RT kthread. (Realistically the delay adds
about 1% to an individual cross-engine workload.) We only signal the
first fence in order to limit the amount of work we move into the
interrupt handler. We also have to remember that our breadcrumbs may be
unordered with respect to the interrupt and so we still require the
waiter process to perform some heavyweight coherency fixups, as well as
traversing the tree of waiters.
v2: No need for early exit in irq handler - it breaks the flow between
patches and prevents the tracepoint
v3: Restore rcu hold across irq signaling of request
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: http://patchwork.freedesktop.org/patch/msgid/20170227205850.2828-2-chris@chris-wilson.co.uk
A request is assigned a global seqno only when it is on the hardware
execution queue. The global seqno can be used to maintain a list of
requests on the same engine in retirement order, for example for
constructing a priority queue for waiting. Prior to its execution, or
if it is subsequently removed in the event of preemption, its global
seqno is zero. As both insertion and removal from the execution queue
may operate in IRQ context, it is not guarded by the usual struct_mutex
BKL. Instead those relying on the global seqno must be prepared for its
value to change between reads. Only when the request is complete can
the global seqno be stable (due to the memory barriers on submitting
the commands to the hardware to write the breadcrumb, if the HWS shows
that it has passed the global seqno and the global seqno is unchanged
after the read, it is indeed complete).
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/20170223074422.4125-9-chris@chris-wilson.co.uk
