With the addition of the inter-context intel_time.sync map, having a
very similar sync_seqno[] is confusing. Aide the reader by denoting that
this is a pre-allocated array for storing semaphore sync points wrt to
the global seqno.
Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: NJoonas Lahtinen <joonas.lahtinen@linux.intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/20170503093924.5320-5-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: NChris Wilson <chris@chris-wilson.co.uk>
Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Reviewed-by: NTvrtko Ursulin <tvrtko.ursulin@intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/20170503093924.5320-4-chris@chris-wilson.co.uk

Currently we filter out repeated use of the same timeline in the low
level i915_gem_request_await_request(), after having added the
dependency on the old request. However, we can lift this to
i915_gem_request_await_dma_fence() (before the dependency is added)
using the observation that requests along the same timeline are
explicitly ordered via i915_add_request (along with the dependencies).
Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>
Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Reviewed-by: NJoonas Lahtinen <joonas.lahtinen@linux.intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/20170503093924.5320-3-chris@chris-wilson.co.uk

By first unwrapping an incoming fence-array into its child fences, we
can simplify the internal branching, and so avoid triggering a potential
bug in the next patch when not squashing the child fences on the same
timeline.

It will also have the advantage of keeping the (top-level) fence arrays
out of any fence/timeline caching since these are unordered timelines
but with a random context id.
Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>
Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Reviewed-by: NTvrtko Ursulin <tvrtko.ursulin@intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/20170503093924.5320-2-chris@chris-wilson.co.uk

If we are enabling the breadcrumbs signaling prior to submitting the
request, we know that we cannot have missed the interrupt and can
therefore skip immediately waking the signaler to check.

This reduces a significant chunk of the __i915_gem_request_submit()
overhead for inter-engine synchronisation, for example in gem_exec_whisper.
Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>
Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/20170426080659.28771-1-chris@chris-wilson.co.ukReviewed-by: NTvrtko Ursulin <tvrtko.ursulin@intel.com>

We need to keep track of the last location we ask the hw to read up to
(RING_TAIL) separately from our last write location into the ring, so
that in the event of a GPU reset we do not tell the HW to proceed into
a partially written request (which can happen if that request is waiting
for an external signal before being executed).

v2: Refactor intel_ring_reset() (Mika)

Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=100144
Testcase: igt/gem_exec_fence/await-hang
Fixes: 821ed7df ("drm/i915: Update reset path to fix incomplete requests")
Fixes: d55ac5bf ("drm/i915: Defer transfer onto execution timeline to actual hw submission")
Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>
Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Cc: Mika Kuoppala <mika.kuoppala@intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/20170425130049.26147-1-chris@chris-wilson.co.ukReviewed-by: NMika Kuoppala <mika.kuoppala@intel.com>

Although we do check the completion-status of the request before
actually adding a wait on it (either to its submit fence or its
completion dma-fence), we currently do not check before adding it to the
dependency lists.

In fact, without checking for a completed request we may try to use the
signaler after it has been retired and its dependency tree freed:

[   60.044057] BUG: KASAN: use-after-free in __list_add_valid+0x1d/0xd0 at addr ffff880348c9e6a0
[   60.044118] Read of size 8 by task gem_exec_fence/530
[   60.044164] CPU: 1 PID: 530 Comm: gem_exec_fence Tainted: G            E   4.11.0-rc7+ #46
[   60.044226] Hardware name: ��������������������������������� ���������������������������������/���������������������������������, BIOS RYBDWi35.86A.0246.2
[   60.044290] Call Trace:
[   60.044337]  dump_stack+0x4d/0x6a
[   60.044383]  kasan_object_err+0x21/0x70
[   60.044435]  kasan_report+0x225/0x4e0
[   60.044488]  ? __list_add_valid+0x1d/0xd0
[   60.044534]  ? kasan_kmalloc+0xad/0xe0
[   60.044587]  __asan_load8+0x5e/0x70
[   60.044639]  __list_add_valid+0x1d/0xd0
[   60.044788]  __i915_priotree_add_dependency+0x67/0x130 [i915]
[   60.044895]  i915_gem_request_await_request+0xa8/0x370 [i915]
[   60.044974]  i915_gem_request_await_dma_fence+0x129/0x140 [i915]
[   60.045049]  i915_gem_do_execbuffer.isra.37+0xb0a/0x26b0 [i915]
[   60.045077]  ? save_stack+0xb1/0xd0
[   60.045105]  ? save_stack_trace+0x1b/0x20
[   60.045132]  ? save_stack+0x46/0xd0
[   60.045158]  ? kasan_kmalloc+0xad/0xe0
[   60.045184]  ? __kmalloc+0xd8/0x670
[   60.045229]  ? drm_ioctl+0x359/0x640 [drm]
[   60.045256]  ? SyS_ioctl+0x41/0x70
[   60.045330]  ? i915_vma_move_to_active+0x540/0x540 [i915]
[   60.045360]  ? tty_insert_flip_string_flags+0xa1/0xf0
[   60.045387]  ? tty_flip_buffer_push+0x63/0x70
[   60.045414]  ? remove_wait_queue+0xa9/0xc0
[   60.045441]  ? kasan_unpoison_shadow+0x35/0x50
[   60.045467]  ? kasan_kmalloc+0xad/0xe0
[   60.045494]  ? kasan_check_write+0x14/0x20
[   60.045568]  i915_gem_execbuffer2+0xdb/0x2a0 [i915]
[   60.045616]  drm_ioctl+0x359/0x640 [drm]
[   60.045705]  ? i915_gem_execbuffer+0x5a0/0x5a0 [i915]
[   60.045751]  ? drm_version+0x150/0x150 [drm]
[   60.045778]  ? compat_start_thread+0x60/0x60
[   60.045805]  ? plist_del+0xda/0x1a0
[   60.045833]  do_vfs_ioctl+0x12e/0x910
[   60.045860]  ? ioctl_preallocate+0x130/0x130
[   60.045886]  ? pci_mmcfg_check_reserved+0xc0/0xc0
[   60.045913]  ? vfs_write+0x196/0x240
[   60.045939]  ? __fget_light+0xa7/0xc0
[   60.045965]  SyS_ioctl+0x41/0x70
[   60.045991]  entry_SYSCALL_64_fastpath+0x17/0x98
[   60.046017] RIP: 0033:0x7feb2baefc47
[   60.046042] RSP: 002b:00007fff56d28e58 EFLAGS: 00000246 ORIG_RAX: 0000000000000010
[   60.046075] RAX: ffffffffffffffda RBX: 00007fff56d290a8 RCX: 00007feb2baefc47
[   60.046102] RDX: 00007fff56d29050 RSI: 00000000c0406469 RDI: 0000000000000003
[   60.046129] RBP: 00007fff56d29050 R08: 000055ecc4cd27d0 R09: 00007feb2bda8600
[   60.046154] R10: 0000000000000073 R11: 0000000000000246 R12: 00000000c0406469
[   60.046177] R13: 0000000000000003 R14: 000000000000000f R15: 0000000000000099
[   60.046203] Object at ffff880348c9e680, in cache i915_dependency size: 64
[   60.046225] Allocated:
[   60.046246] PID = 530
[   60.046269]  save_stack_trace+0x1b/0x20
[   60.046292]  save_stack+0x46/0xd0
[   60.046318]  kasan_kmalloc+0xad/0xe0
[   60.046343]  kasan_slab_alloc+0x12/0x20
[   60.046368]  kmem_cache_alloc+0xab/0x650
[   60.046445]  i915_gem_request_await_request+0x88/0x370 [i915]
[   60.046559]  i915_gem_request_await_dma_fence+0x129/0x140 [i915]
[   60.046705]  i915_gem_do_execbuffer.isra.37+0xb0a/0x26b0 [i915]
[   60.046849]  i915_gem_execbuffer2+0xdb/0x2a0 [i915]
[   60.046936]  drm_ioctl+0x359/0x640 [drm]
[   60.046987]  do_vfs_ioctl+0x12e/0x910
[   60.047038]  SyS_ioctl+0x41/0x70
[   60.047090]  entry_SYSCALL_64_fastpath+0x17/0x98
[   60.047139] Freed:
[   60.047179] PID = 530
[   60.047223]  save_stack_trace+0x1b/0x20
[   60.047269]  save_stack+0x46/0xd0
[   60.047317]  kasan_slab_free+0x72/0xc0
[   60.047366]  kmem_cache_free+0x39/0x160
[   60.047512]  i915_gem_request_retire+0x83f/0x930 [i915]
[   60.047657]  i915_gem_request_alloc+0x166/0x600 [i915]
[   60.047799]  i915_gem_do_execbuffer.isra.37+0xad8/0x26b0 [i915]
[   60.047897]  i915_gem_execbuffer2+0xdb/0x2a0 [i915]
[   60.047942]  drm_ioctl+0x359/0x640 [drm]
[   60.047968]  do_vfs_ioctl+0x12e/0x910
[   60.047993]  SyS_ioctl+0x41/0x70
[   60.048019]  entry_SYSCALL_64_fastpath+0x17/0x98
[   60.048044] Memory state around the buggy address:
[   60.048066]  ffff880348c9e580: 00 00 00 00 00 00 00 00 fc fc fc fc fc fc fc fc
[   60.048105]  ffff880348c9e600: 00 00 00 00 00 00 00 00 fc fc fc fc fc fc fc fc
[   60.048138] >ffff880348c9e680: fb fb fb fb fb fb fb fb fc fc fc fc fc fc fc fc
[   60.048170]                                ^
[   60.048191]  ffff880348c9e700: 00 00 00 00 00 00 00 00 fc fc fc fc fc fc fc fc
[   60.048225]  ffff880348c9e780: 00 00 00 00 00 00 00 00 fc fc fc fc fc fc fc fc

Note to hit the use-after-free requires us to be passed back a request
via a fence-array, that is from explicit fencing accumulated into a
sync-file fence-array.

Fixes: 52e54209 ("drm/i915/scheduler: Record all dependencies upon request construction")
Testcase: igt/gem_exec_fence/expired-history
Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: NMichał Winiarski <michal.winiarski@intel.com>
Reviewed-by: NJoonas Lahtinen <joonas.lahtinen@linux.intel.com>
Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/20170422081537.6468-1-chris@chris-wilson.co.uk

Introduce a new execobject.flag (EXEC_OBJECT_CAPTURE) that userspace may
use to indicate that it wants the contents of this buffer preserved in
the error state (/sys/class/drm/cardN/error) following a GPU hang
involving this batch.

Use this at your discretion, the contents of the error state. although
compressed, are allocated with GFP_ATOMIC (i.e. limited) and kept for all
eternity (until the error state is destroyed).

Based on an earlier patch by Ben Widawsky <ben@bwidawsk.net>
Testcase: igt/gem_exec_capture
Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>
Cc: Ben Widawsky <ben@bwidawsk.net>
Cc: Matt Turner <mattst88@gmail.com>
Acked-by: NBen Widawsky <ben@bwidawsk.net>
Acked-by: NMatt Turner <mattst88@gmail.com>
Reviewed-by: NJoonas Lahtinen <joonas.lahtinen@linux.intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/20170415093902.22581-1-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: NChris 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.ukReviewed-by: NJoonas Lahtinen <joonas.lahtinen@linux.intel.com>

When we update the global seqno (on the engine timeline), we modify HW
state (both registers and mapped pages). As we do this, we should be
sure that the HW is idle and we are not causing a conflict. The caller
is supposed to wait_for_idle before calling us to update the seqno, so
let's assert they have and the engine is indeed idle.
Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>
Link: http://patchwork.freedesktop.org/patch/msgid/20170405153055.28123-1-chris@chris-wilson.co.ukReviewed-by: NJoonas Lahtinen <joonas.lahtinen@linux.intel.com>

We can merge the pair of loops over the engines and their timelines into
a single loop, making it easier to read and more consistent with the
commentary.
Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>
Link: http://patchwork.freedesktop.org/patch/msgid/20170330145041.9005-6-chris@chris-wilson.co.ukReviewed-by: NJoonas Lahtinen <joonas.lahtinen@linux.intel.com>

Having added the wait upon each engine to idle into the central
i915_gem_wait_for_idle(), we can remove the now redundant wait from
reset_all_global_seqno(). This has the advantage of removing the late
detection of an error (an engine still busy) which left the seqno reset
only partially complete (though it should be safe enough!).
Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>
Link: http://patchwork.freedesktop.org/patch/msgid/20170330145041.9005-5-chris@chris-wilson.co.ukReviewed-by: NJoonas Lahtinen <joonas.lahtinen@linux.intel.com>

As we now distinguish everywhere that can call
i915_gem_retire_requests() following a successful wait_for_idle, we can
remove the duplication by moving that call into i915_gem_wait_for_idle()
itself.
Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>
Link: http://patchwork.freedesktop.org/patch/msgid/20170330145041.9005-3-chris@chris-wilson.co.ukReviewed-by: NJoonas 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: NKrzysztof Olinski <krzysztof.e.olinski@intel.com>
Fixes: 80b204bc ("drm/i915: Enable multiple timelines")
Signed-off-by: NChris 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.ukReviewed-by: NJoonas Lahtinen <joonas.lahtinen@linux.intel.com>
Reviewed-by: NMichał Winiarski <michal.winiarski@intel.com>

Storing the position of the breadcrumb of the last retired request as
a separate last_retired_head is superfluous as we always copy that into
head prior to recalculation of the intel_ring.space.
Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: NJoonas Lahtinen <joonas.lahtinen@linux.intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/20170321102552.24357-1-chris@chris-wilson.co.uk

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: NChris Wilson <chris@chris-wilson.co.uk>
Acked-by: NMichel Thierry <michel.thierry@intel.com>
Reviewed-by: NMika Kuoppala <mika.kuoppala@intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/20170316171305.12972-1-chris@chris-wilson.co.uk

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 fe9ae7a3
("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: NChris 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.ukReviewed-by: NMika 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: NChris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: NMika Kuoppala <mika.kuoppala@intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/20170302122525.19675-1-chris@chris-wilson.co.uk

assert_spin_locked() becomes an unconditionally compiled BUG_ON(),
adding debug code right into the heart of critical routines like
interrupt handlers.

   text	   data	    bss	    dec	    hex
1296480	  19944	   2272	1318696	 141f28	before (lockdep disabled)
1295984	  19944	   2272	1318200	 141d38	after

1336261	  21139	   3208	1360608	 14c2e0	before (lockdep enabled)
1339920	  21139	   3208	1364267	 14d12b	after

Small saving for release; hopefully more instructive in debug.
Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>
Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Cc: Mika Kuoppala <mika.kuoppala@linux.intel.com>
Cc: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/20170302132801.599-1-chris@chris-wilson.co.ukReviewed-by: NJoonas Lahtinen <joonas.lahtinen@linux.intel.com>

Everytime we take the fence->lock (aka request->lock), we must do so
with irqs disabled since it may be used from within an hardirq context.
As sometimes we are taking the lock in a nested manner, assert that the
caller did disable the irqs for us.
Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>
Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Cc: Mika Kuoppala <mika.kuoppala@linux.intel.com>
Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/20170302115130.28434-1-chris@chris-wilson.co.ukReviewed-by: NJoonas Lahtinen <joonas.lahtinen@linux.intel.com>

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: NLinus 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: NIngo 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: NLinus 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: NIngo 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: NChris Wilson <chris@chris-wilson.co.uk>
Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Cc: Mika Kuoppala <mika.kuoppala@linux.intel.com>
Reviewed-by: NTvrtko Ursulin <tvrtko.ursulin@intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/20170227205850.2828-2-chris@chris-wilson.co.uk

Now that the code is getting simpler, we can reduce the indentation when
waiting for the global_seqno.
Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: NTvrtko Ursulin <tvrtko.ursulin@intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/20170223074422.4125-17-chris@chris-wilson.co.uk

As we handoff the GPU reset to the waiter, we need to check we don't
miss a wakeup if it has already been sent prior to us starting the wait.

v2: Tweak checking for reset to be clear to the need before sleeping
after changing the task state.
Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>
Link: http://patchwork.freedesktop.org/patch/msgid/20170223074422.4125-16-chris@chris-wilson.co.ukReviewed-by: NTvrtko Ursulin <tvrtko.ursulin@intel.com>

Combine the common code for the pair of waiters into a single function.

v2: Rename reset_request to wait_request_check_and_reset
Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>
Link: http://patchwork.freedesktop.org/patch/msgid/20170223074422.4125-15-chris@chris-wilson.co.ukReviewed-by: NTvrtko Ursulin <tvrtko.ursulin@intel.com>

If we change the wait_queue_t from using the autoremove_wake_function to
the default_wake_function, we no longer have to restore the wait_queue_t
entry on the wait_queue_head_t list after being woken up by it, as we
are unusual in sleeping multiple times on the same wait_queue_t.
Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: NTvrtko Ursulin <tvrtko.ursulin@intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/20170223074422.4125-14-chris@chris-wilson.co.uk

After the request is cancelled, we then need to remove it from the
global execution timeline and return it to the context timeline, the
inverse of submit_request().

v2: Move manipulation of struct intel_wait to helpers
Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>
Cc: Tvrtko Ursulin <tvrtko.ursulin@linux.intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/20170223074422.4125-12-chris@chris-wilson.co.ukReviewed-by: NTvrtko Ursulin <tvrtko.ursulin@intel.com>

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: NChris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: NTvrtko Ursulin <tvrtko.ursulin@intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/20170223074422.4125-9-chris@chris-wilson.co.uk

On reflection, we are only using the execute fence as a waitqueue on the
global_seqno and not using it for dependency tracking between fences
(unlike the submit and dma fences). By only treating it as a waitqueue,
we can then treat it similar to the other waitqueues during submit,
making the code simpler.
Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: NTvrtko Ursulin <tvrtko.ursulin@intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/20170223074422.4125-8-chris@chris-wilson.co.uk

It had only one callsite and existed to keep the code clearer. Now
having shared the wait-on-error between phases and with plans to change
the wait-for-execute in the next few patches, remove the out of line
wait loop and move it into the main body of i915_wait_request.
Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: NJoonas Lahtinen <joonas.lahtinen@linux.intel.com>
Reviewed-by: NTvrtko Ursulin <tvrtko.ursulin@intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/20170223074422.4125-7-chris@chris-wilson.co.uk

Add ourselves to the gpu error waitqueue earlier on, even before we
determine we have to wait on the seqno. This is so that we can then
share the waitqueue between stages in subsequent patches.
Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: NJoonas Lahtinen <joonas.lahtinen@linux.intel.com>
Reviewed-by: NTvrtko Ursulin <tvrtko.ursulin@intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/20170223074422.4125-6-chris@chris-wilson.co.uk

Use a local variable to avoid having to type out the full name of the
gpu_error wait_queue.
Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: NJoonas Lahtinen <joonas.lahtinen@linux.intel.com>
Reviewed-by: NTvrtko Ursulin <tvrtko.ursulin@intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/20170223074422.4125-5-chris@chris-wilson.co.uk

Move the companion functions next to each other.
Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: NJoonas Lahtinen <joonas.lahtinen@linux.intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/20170223074422.4125-4-chris@chris-wilson.co.uk

Replace the global device seqno with one for each engine, and account
for in-flight seqno on each separately. This is consistent with
dma-fence as each timeline has separate fence-contexts for each engine
and a seqno is only ordered within a fence-context (i.e.  seqno do not
need to be ordered wrt to other engines, just ordered within a single
engine). This is required to enable request rewinding for preemption on
individual engines (we have to rewind the global seqno to avoid
overflow, and we do not have to rewind all engines just to preempt one.)

v2: Rename active_seqno to inflight_seqnos to more clearly indicate that
it is a counter and not equivalent to the existing seqno. Update
functions that operated on active_seqno similarly.
Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: NJoonas Lahtinen <joonas.lahtinen@linux.intel.com>
Reviewed-by: NTvrtko Ursulin <tvrtko.ursulin@intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/20170223074422.4125-3-chris@chris-wilson.co.uk

These new tracepoints are emitted once the request is ready to
be submitted to the GPU and once the request is about to
be submitted to the GPU, respectively.

Former condition triggers as soon as all the fences and
dependencies have been resolved, and the latter once the
backend is about to submit it to the GPU.

New tracepoint are enabled via the new
DRM_I915_LOW_LEVEL_TRACEPOINTS Kconfig option which is disabled
by default to alleviate the performance impact concerns.

v2: Move execute tracepoint to __i915_gem_request_submit.
    (Chris Wilson)
Signed-off-by: NTvrtko Ursulin <tvrtko.ursulin@intel.com>
Reviewed-by: NChris Wilson <chris@chris-wilson.co.uk>

Provide the same information as the other request event classes.

v2: Pass in flags so we can properly report the blocking status.
    (Chris Wilson)

v3: Log hex with 0x prefix for clarity.

v4: Derive blocking status from flags. (Chris Wilson)
Signed-off-by: NTvrtko Ursulin <tvrtko.ursulin@intel.com>
Reviewed-by: NChris Wilson <chris@chris-wilson.co.uk>

If an interrupt has been posted, and we were spinning on the active
seqno waiting for it to advance but it did not, then we can expect that
it will not see its advance in the immediate future and should call into
the irq-seqno barrier. We can stop spinning at this point, and leave the
difficulty of handling the coherency to the caller.
Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>
Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Cc: Mika Kuoppala <mika.kuoppala@linux.intel.com>
Reviewed-by: NTvrtko Ursulin <tvrtko.ursulin@intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/20170217151304.16665-3-chris@chris-wilson.co.uk

We first wait for a request to be submitted to hw and assigned a seqno,
before we can wait for the hw to signal completion (otherwise we don't
know the hw id we need to wait upon). Whilst waiting for the request to
be submitted, we may exceed the user's timeout and need to propagate the
error back.

v2: Make ETIME into an error from wait_for_execute for consistent exit
handling.
Reported-by: NTvrtko Ursulin <tvrtko.ursulin@intel.com>
Fixes: 4680816b ("drm/i915: Wait first for submission, before waiting for request completion")
Testcase: igt/gem_wait/basic-await
Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>
Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Cc: <drm-intel-fixes@lists.freedesktop.org> # v4.10-rc1+
Cc: stable@vger.kernel.org
Link: http://patchwork.freedesktop.org/patch/msgid/20170208181238.7232-1-chris@chris-wilson.co.ukReviewed-by: NTvrtko Ursulin <tvrtko.ursulin@intel.com>
(cherry picked from commit 969bb72c)
Signed-off-by: NJani Nikula <jani.nikula@intel.com>

This removes the usage of intel_ring_emit in favour of
directly writing to the ring buffer.

intel_ring_emit was preventing the compiler for optimising
fetch and increment of the current ring buffer pointer and
therefore generating very verbose code for every write.

It had no useful purpose since all ringbuffer operations
are started and ended with intel_ring_begin and
intel_ring_advance respectively, with no bail out in the
middle possible, so it is fine to increment the tail in
intel_ring_begin and let the code manage the pointer
itself.

Useless instruction removal amounts to approximately
two and half kilobytes of saved text on my build.

Not sure if this has any measurable performance
implications but executing a ton of useless instructions
on fast paths cannot be good.

v2:
 * Change return from intel_ring_begin to error pointer by
   popular demand.
 * Move tail increment to intel_ring_advance to enable some
   error checking.

v3:
 * Move tail advance back into intel_ring_begin.
 * Rebase and tidy.

v4:
 * Complete rebase after a few months since v3.

v5:
 * Remove unecessary cast and fix !debug compile. (Chris Wilson)

v6:
 * Make intel_ring_offset take request as well.
 * Fix recording of request postfix plus a sprinkle of asserts.
   (Chris Wilson)

v7:
 * Use intel_ring_offset to get the postfix. (Chris Wilson)
 * Convert GVT code as well.

v8:
 * Rename *out++ to *cs++.

v9:
 * Fix GVT out to cs conversion in GVT.

v10:
 * Rebase for new intel_ring_begin in selftests.
Signed-off-by: NTvrtko Ursulin <tvrtko.ursulin@intel.com>
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Cc: Zhi Wang <zhi.a.wang@intel.com>
Reviewed-by: NChris Wilson <chris@chris-wilson.co.uk>
Acked-by: NJoonas Lahtinen <joonas.lahtinen@linux.intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/20170214113242.29241-1-tvrtko.ursulin@linux.intel.com