Remove the struct_mutex requirement around dev_priv->mm.bound_list and
dev_priv->mm.unbound_list by giving it its own spinlock. This reduces
one more requirement for struct_mutex and in the process gives us
slightly more accurate unbound_list tracking, which should improve the
shrinker - but the drawback is that we drop the retirement before
counting so i915_gem_object_is_active() may be stale and lead us to
underestimate the number of objects that may be shrunk (see commit
bed50aea61 ("drm/i915/shrinker: Flush active on objects before
counting")).
v2: Crosslink the spinlock to the lists it protects, and btw this
changes s/obj->global_link/obj->mm.link/
v3: Fix decoupling of old links in i915_gem_object_attach_phys()
v3.1: Fix the fix, only unlink if it was linked
v3.2: Use a local for to_i915(obj->base.dev)->mm.obj_lock
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20171016114037.5556-1-chris@chris-wilson.co.uk
A bug recently encountered involved the issue where are we were
submitting requests to different ppGTT, each would pin a segment of the
GGTT for its logical context and ring. However, this is invisible to
eviction as we do not tie the context/ring VMA to a request and so do
not automatically wait upon it them (instead they are marked as pinned,
preventing eviction entirely). Instead the eviction code must flush those
contexts by switching to the kernel context. This selftest tries to
fill the GGTT with contexts to exercise a path where the
switch-to-kernel-context failed to make forward progress and we fail
with ENOSPC.
v2: Make the hole in the filled GGTT explicit.
v3: Swap out the arbitrary timeout for a private notification from
i915_gem_evict_something()
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20171012125726.14736-3-chris@chris-wilson.co.uk
Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
The major scaling bottleneck in execbuffer is the processing of the
execobjects. Creating an auxiliary list is inefficient when compared to
using the execobject array we already have allocated.
Reservation is then split into phases. As we lookup up the VMA, we
try and bind it back into active location. Only if that fails, do we add
it to the unbound list for phase 2. In phase 2, we try and add all those
objects that could not fit into their previous location, with fallback
to retrying all objects and evicting the VM in case of severe
fragmentation. (This is the same as before, except that phase 1 is now
done inline with looking up the VMA to avoid an iteration over the
execobject array. In the ideal case, we eliminate the separate reservation
phase). During the reservation phase, we only evict from the VM between
passes (rather than currently as we try to fit every new VMA). In
testing with Unreal Engine's Atlantis demo which stresses the eviction
logic on gen7 class hardware, this speed up the framerate by a factor of
2.
The second loop amalgamation is between move_to_gpu and move_to_active.
As we always submit the request, even if incomplete, we can use the
current request to track active VMA as we perform the flushes and
synchronisation required.
The next big advancement is to avoid copying back to the user any
execobjects and relocations that are not changed.
v2: Add a Theory of Operation spiel.
v3: Fall back to slow relocations in preparation for flushing userptrs.
v4: Document struct members, factor out eb_validate_vma(), add a few
more comments to explain some magic and hide other magic behind macros.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
