diff options
author | Chris Wilson <chris@chris-wilson.co.uk> | 2019-10-04 14:40:00 +0100 |
---|---|---|
committer | Chris Wilson <chris@chris-wilson.co.uk> | 2019-10-04 15:39:12 +0100 |
commit | b1e3177bd1d8f41e2a9cc847e56a96cdc0eefe62 (patch) | |
tree | 9af22565533f12868a015e18e51406d54773e08a /drivers/gpu/drm/i915/gt/intel_timeline.c | |
parent | drm/i915: Push the i915_active.retire into a worker (diff) | |
download | linux-dev-b1e3177bd1d8f41e2a9cc847e56a96cdc0eefe62.tar.xz linux-dev-b1e3177bd1d8f41e2a9cc847e56a96cdc0eefe62.zip |
drm/i915: Coordinate i915_active with its own mutex
Forgo the struct_mutex serialisation for i915_active, and interpose its
own mutex handling for active/retire.
This is a multi-layered sleight-of-hand. First, we had to ensure that no
active/retire callbacks accidentally inverted the mutex ordering rules,
nor assumed that they were themselves serialised by struct_mutex. More
challenging though, is the rule over updating elements of the active
rbtree. Instead of the whole i915_active now being serialised by
struct_mutex, allocations/rotations of the tree are serialised by the
i915_active.mutex and individual nodes are serialised by the caller
using the i915_timeline.mutex (we need to use nested spinlocks to
interact with the dma_fence callback lists).
The pain point here is that instead of a single mutex around execbuf, we
now have to take a mutex for active tracker (one for each vma, context,
etc) and a couple of spinlocks for each fence update. The improvement in
fine grained locking allowing for multiple concurrent clients
(eventually!) should be worth it in typical loads.
v2: Add some comments that barely elucidate anything :(
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/20191004134015.13204-6-chris@chris-wilson.co.uk
Diffstat (limited to 'drivers/gpu/drm/i915/gt/intel_timeline.c')
-rw-r--r-- | drivers/gpu/drm/i915/gt/intel_timeline.c | 7 |
1 files changed, 3 insertions, 4 deletions
diff --git a/drivers/gpu/drm/i915/gt/intel_timeline.c b/drivers/gpu/drm/i915/gt/intel_timeline.c index 653f60e78392..0f959694303c 100644 --- a/drivers/gpu/drm/i915/gt/intel_timeline.c +++ b/drivers/gpu/drm/i915/gt/intel_timeline.c @@ -178,8 +178,7 @@ cacheline_alloc(struct intel_timeline_hwsp *hwsp, unsigned int cacheline) cl->hwsp = hwsp; cl->vaddr = page_pack_bits(vaddr, cacheline); - i915_active_init(hwsp->gt->i915, &cl->active, - __cacheline_active, __cacheline_retire); + i915_active_init(&cl->active, __cacheline_active, __cacheline_retire); return cl; } @@ -255,7 +254,7 @@ int intel_timeline_init(struct intel_timeline *timeline, mutex_init(&timeline->mutex); - INIT_ACTIVE_REQUEST(&timeline->last_request, &timeline->mutex); + INIT_ACTIVE_FENCE(&timeline->last_request, &timeline->mutex); INIT_LIST_HEAD(&timeline->requests); i915_syncmap_init(&timeline->sync); @@ -443,7 +442,7 @@ __intel_timeline_get_seqno(struct intel_timeline *tl, * free it after the current request is retired, which ensures that * all writes into the cacheline from previous requests are complete. */ - err = i915_active_ref(&tl->hwsp_cacheline->active, tl, rq); + err = i915_active_ref(&tl->hwsp_cacheline->active, tl, &rq->fence); if (err) goto err_cacheline; |