Commit 97b2a6a1 authored by John Harrison's avatar John Harrison Committed by Daniel Vetter

drm/i915: Replace last_[rwf]_seqno with last_[rwf]_req

The object structure contains the last read, write and fenced seqno values for
use in syncrhonisation operations. These have now been replaced with their
request structure counterparts.

Note that to ensure that objects do not end up with dangling pointers, the
assignments of last_*_req include reference count updates. Thus a request cannot
be freed if an object is still hanging on to it for any reason.

v2: Corrected 'last_rendering_' to 'last_read_' in a number of comments that did
not get updated when 'last_rendering_seqno' became 'last_read|write_seqno'
several millenia ago.

For: VIZ-4377
Signed-off-by: default avatarJohn Harrison <John.C.Harrison@Intel.com>
Reviewed-by: default avatarThomas Daniel <Thomas.Daniel@intel.com>
Signed-off-by: default avatarDaniel Vetter <daniel.vetter@ffwll.ch>
parent b793a00a
...@@ -131,9 +131,9 @@ describe_obj(struct seq_file *m, struct drm_i915_gem_object *obj) ...@@ -131,9 +131,9 @@ describe_obj(struct seq_file *m, struct drm_i915_gem_object *obj)
obj->base.size / 1024, obj->base.size / 1024,
obj->base.read_domains, obj->base.read_domains,
obj->base.write_domain, obj->base.write_domain,
obj->last_read_seqno, i915_gem_request_get_seqno(obj->last_read_req),
obj->last_write_seqno, i915_gem_request_get_seqno(obj->last_write_req),
obj->last_fenced_seqno, i915_gem_request_get_seqno(obj->last_fenced_req),
i915_cache_level_str(to_i915(obj->base.dev), obj->cache_level), i915_cache_level_str(to_i915(obj->base.dev), obj->cache_level),
obj->dirty ? " dirty" : "", obj->dirty ? " dirty" : "",
obj->madv == I915_MADV_DONTNEED ? " purgeable" : ""); obj->madv == I915_MADV_DONTNEED ? " purgeable" : "");
......
...@@ -1944,10 +1944,10 @@ struct drm_i915_gem_object { ...@@ -1944,10 +1944,10 @@ struct drm_i915_gem_object {
struct intel_engine_cs *ring; struct intel_engine_cs *ring;
/** Breadcrumb of last rendering to the buffer. */ /** Breadcrumb of last rendering to the buffer. */
uint32_t last_read_seqno; struct drm_i915_gem_request *last_read_req;
uint32_t last_write_seqno; struct drm_i915_gem_request *last_write_req;
/** Breadcrumb of last fenced GPU access to the buffer. */ /** Breadcrumb of last fenced GPU access to the buffer. */
uint32_t last_fenced_seqno; struct drm_i915_gem_request *last_fenced_req;
/** Current tiling stride for the object, if it's tiled. */ /** Current tiling stride for the object, if it's tiled. */
uint32_t stride; uint32_t stride;
...@@ -1986,9 +1986,10 @@ void i915_gem_track_fb(struct drm_i915_gem_object *old, ...@@ -1986,9 +1986,10 @@ void i915_gem_track_fb(struct drm_i915_gem_object *old,
* The request queue allows us to note sequence numbers that have been emitted * The request queue allows us to note sequence numbers that have been emitted
* and may be associated with active buffers to be retired. * and may be associated with active buffers to be retired.
* *
* By keeping this list, we can avoid having to do questionable * By keeping this list, we can avoid having to do questionable sequence
* sequence-number comparisons on buffer last_rendering_seqnos, and associate * number comparisons on buffer last_read|write_seqno. It also allows an
* an emission time with seqnos for tracking how far ahead of the GPU we are. * emission time to be associated with the request for tracking how far ahead
* of the GPU the submission is.
*/ */
struct drm_i915_gem_request { struct drm_i915_gem_request {
struct kref ref; struct kref ref;
......
...@@ -1346,11 +1346,11 @@ i915_gem_object_wait_rendering__tail(struct drm_i915_gem_object *obj) ...@@ -1346,11 +1346,11 @@ i915_gem_object_wait_rendering__tail(struct drm_i915_gem_object *obj)
/* Manually manage the write flush as we may have not yet /* Manually manage the write flush as we may have not yet
* retired the buffer. * retired the buffer.
* *
* Note that the last_write_seqno is always the earlier of * Note that the last_write_req is always the earlier of
* the two (read/write) seqno, so if we haved successfully waited, * the two (read/write) requests, so if we haved successfully waited,
* we know we have passed the last write. * we know we have passed the last write.
*/ */
obj->last_write_seqno = 0; i915_gem_request_assign(&obj->last_write_req, NULL);
return 0; return 0;
} }
...@@ -1363,14 +1363,18 @@ static __must_check int ...@@ -1363,14 +1363,18 @@ static __must_check int
i915_gem_object_wait_rendering(struct drm_i915_gem_object *obj, i915_gem_object_wait_rendering(struct drm_i915_gem_object *obj,
bool readonly) bool readonly)
{ {
struct drm_i915_gem_request *req;
struct intel_engine_cs *ring = obj->ring; struct intel_engine_cs *ring = obj->ring;
u32 seqno; u32 seqno;
int ret; int ret;
seqno = readonly ? obj->last_write_seqno : obj->last_read_seqno; req = readonly ? obj->last_write_req : obj->last_read_req;
if (seqno == 0) if (!req)
return 0; return 0;
seqno = i915_gem_request_get_seqno(req);
WARN_ON(seqno == 0);
ret = i915_wait_seqno(ring, seqno); ret = i915_wait_seqno(ring, seqno);
if (ret) if (ret)
return ret; return ret;
...@@ -1386,6 +1390,7 @@ i915_gem_object_wait_rendering__nonblocking(struct drm_i915_gem_object *obj, ...@@ -1386,6 +1390,7 @@ i915_gem_object_wait_rendering__nonblocking(struct drm_i915_gem_object *obj,
struct drm_i915_file_private *file_priv, struct drm_i915_file_private *file_priv,
bool readonly) bool readonly)
{ {
struct drm_i915_gem_request *req;
struct drm_device *dev = obj->base.dev; struct drm_device *dev = obj->base.dev;
struct drm_i915_private *dev_priv = dev->dev_private; struct drm_i915_private *dev_priv = dev->dev_private;
struct intel_engine_cs *ring = obj->ring; struct intel_engine_cs *ring = obj->ring;
...@@ -1396,10 +1401,13 @@ i915_gem_object_wait_rendering__nonblocking(struct drm_i915_gem_object *obj, ...@@ -1396,10 +1401,13 @@ i915_gem_object_wait_rendering__nonblocking(struct drm_i915_gem_object *obj,
BUG_ON(!mutex_is_locked(&dev->struct_mutex)); BUG_ON(!mutex_is_locked(&dev->struct_mutex));
BUG_ON(!dev_priv->mm.interruptible); BUG_ON(!dev_priv->mm.interruptible);
seqno = readonly ? obj->last_write_seqno : obj->last_read_seqno; req = readonly ? obj->last_write_req : obj->last_read_req;
if (seqno == 0) if (!req)
return 0; return 0;
seqno = i915_gem_request_get_seqno(req);
WARN_ON(seqno == 0);
ret = i915_gem_check_wedge(&dev_priv->gpu_error, true); ret = i915_gem_check_wedge(&dev_priv->gpu_error, true);
if (ret) if (ret)
return ret; return ret;
...@@ -2257,12 +2265,12 @@ static void ...@@ -2257,12 +2265,12 @@ static void
i915_gem_object_move_to_active(struct drm_i915_gem_object *obj, i915_gem_object_move_to_active(struct drm_i915_gem_object *obj,
struct intel_engine_cs *ring) struct intel_engine_cs *ring)
{ {
u32 seqno = intel_ring_get_seqno(ring); struct drm_i915_gem_request *req = intel_ring_get_request(ring);
BUG_ON(ring == NULL); BUG_ON(ring == NULL);
if (obj->ring != ring && obj->last_write_seqno) { if (obj->ring != ring && obj->last_write_req) {
/* Keep the seqno relative to the current ring */ /* Keep the request relative to the current ring */
obj->last_write_seqno = seqno; i915_gem_request_assign(&obj->last_write_req, req);
} }
obj->ring = ring; obj->ring = ring;
...@@ -2274,7 +2282,7 @@ i915_gem_object_move_to_active(struct drm_i915_gem_object *obj, ...@@ -2274,7 +2282,7 @@ i915_gem_object_move_to_active(struct drm_i915_gem_object *obj,
list_move_tail(&obj->ring_list, &ring->active_list); list_move_tail(&obj->ring_list, &ring->active_list);
obj->last_read_seqno = seqno; i915_gem_request_assign(&obj->last_read_req, req);
} }
void i915_vma_move_to_active(struct i915_vma *vma, void i915_vma_move_to_active(struct i915_vma *vma,
...@@ -2305,11 +2313,11 @@ i915_gem_object_move_to_inactive(struct drm_i915_gem_object *obj) ...@@ -2305,11 +2313,11 @@ i915_gem_object_move_to_inactive(struct drm_i915_gem_object *obj)
list_del_init(&obj->ring_list); list_del_init(&obj->ring_list);
obj->ring = NULL; obj->ring = NULL;
obj->last_read_seqno = 0; i915_gem_request_assign(&obj->last_read_req, NULL);
obj->last_write_seqno = 0; i915_gem_request_assign(&obj->last_write_req, NULL);
obj->base.write_domain = 0; obj->base.write_domain = 0;
obj->last_fenced_seqno = 0; i915_gem_request_assign(&obj->last_fenced_req, NULL);
obj->active = 0; obj->active = 0;
drm_gem_object_unreference(&obj->base); drm_gem_object_unreference(&obj->base);
...@@ -2326,7 +2334,7 @@ i915_gem_object_retire(struct drm_i915_gem_object *obj) ...@@ -2326,7 +2334,7 @@ i915_gem_object_retire(struct drm_i915_gem_object *obj)
return; return;
if (i915_seqno_passed(ring->get_seqno(ring, true), if (i915_seqno_passed(ring->get_seqno(ring, true),
obj->last_read_seqno)) i915_gem_request_get_seqno(obj->last_read_req)))
i915_gem_object_move_to_inactive(obj); i915_gem_object_move_to_inactive(obj);
} }
...@@ -2753,7 +2761,8 @@ i915_gem_retire_requests_ring(struct intel_engine_cs *ring) ...@@ -2753,7 +2761,8 @@ i915_gem_retire_requests_ring(struct intel_engine_cs *ring)
struct drm_i915_gem_object, struct drm_i915_gem_object,
ring_list); ring_list);
if (!i915_seqno_passed(seqno, obj->last_read_seqno)) if (!i915_seqno_passed(seqno,
i915_gem_request_get_seqno(obj->last_read_req)))
break; break;
i915_gem_object_move_to_inactive(obj); i915_gem_object_move_to_inactive(obj);
...@@ -2872,7 +2881,8 @@ i915_gem_object_flush_active(struct drm_i915_gem_object *obj) ...@@ -2872,7 +2881,8 @@ i915_gem_object_flush_active(struct drm_i915_gem_object *obj)
int ret; int ret;
if (obj->active) { if (obj->active) {
ret = i915_gem_check_olr(obj->ring, obj->last_read_seqno); ret = i915_gem_check_olr(obj->ring,
i915_gem_request_get_seqno(obj->last_read_req));
if (ret) if (ret)
return ret; return ret;
...@@ -2933,13 +2943,12 @@ i915_gem_wait_ioctl(struct drm_device *dev, void *data, struct drm_file *file) ...@@ -2933,13 +2943,12 @@ i915_gem_wait_ioctl(struct drm_device *dev, void *data, struct drm_file *file)
if (ret) if (ret)
goto out; goto out;
if (obj->active) { if (!obj->active || !obj->last_read_req)
seqno = obj->last_read_seqno; goto out;
ring = obj->ring;
}
if (seqno == 0) seqno = i915_gem_request_get_seqno(obj->last_read_req);
goto out; WARN_ON(seqno == 0);
ring = obj->ring;
/* Do this after OLR check to make sure we make forward progress polling /* Do this after OLR check to make sure we make forward progress polling
* on this IOCTL with a timeout <=0 (like busy ioctl) * on this IOCTL with a timeout <=0 (like busy ioctl)
...@@ -2990,7 +2999,7 @@ i915_gem_object_sync(struct drm_i915_gem_object *obj, ...@@ -2990,7 +2999,7 @@ i915_gem_object_sync(struct drm_i915_gem_object *obj,
idx = intel_ring_sync_index(from, to); idx = intel_ring_sync_index(from, to);
seqno = obj->last_read_seqno; seqno = i915_gem_request_get_seqno(obj->last_read_req);
/* Optimization: Avoid semaphore sync when we are sure we already /* Optimization: Avoid semaphore sync when we are sure we already
* waited for an object with higher seqno */ * waited for an object with higher seqno */
if (seqno <= from->semaphore.sync_seqno[idx]) if (seqno <= from->semaphore.sync_seqno[idx])
...@@ -3003,11 +3012,12 @@ i915_gem_object_sync(struct drm_i915_gem_object *obj, ...@@ -3003,11 +3012,12 @@ i915_gem_object_sync(struct drm_i915_gem_object *obj,
trace_i915_gem_ring_sync_to(from, to, seqno); trace_i915_gem_ring_sync_to(from, to, seqno);
ret = to->semaphore.sync_to(to, from, seqno); ret = to->semaphore.sync_to(to, from, seqno);
if (!ret) if (!ret)
/* We use last_read_seqno because sync_to() /* We use last_read_req because sync_to()
* might have just caused seqno wrap under * might have just caused seqno wrap under
* the radar. * the radar.
*/ */
from->semaphore.sync_seqno[idx] = obj->last_read_seqno; from->semaphore.sync_seqno[idx] =
i915_gem_request_get_seqno(obj->last_read_req);
return ret; return ret;
} }
...@@ -3321,12 +3331,13 @@ static void i915_gem_object_update_fence(struct drm_i915_gem_object *obj, ...@@ -3321,12 +3331,13 @@ static void i915_gem_object_update_fence(struct drm_i915_gem_object *obj,
static int static int
i915_gem_object_wait_fence(struct drm_i915_gem_object *obj) i915_gem_object_wait_fence(struct drm_i915_gem_object *obj)
{ {
if (obj->last_fenced_seqno) { if (obj->last_fenced_req) {
int ret = i915_wait_seqno(obj->ring, obj->last_fenced_seqno); int ret = i915_wait_seqno(obj->ring,
i915_gem_request_get_seqno(obj->last_fenced_req));
if (ret) if (ret)
return ret; return ret;
obj->last_fenced_seqno = 0; i915_gem_request_assign(&obj->last_fenced_req, NULL);
} }
return 0; return 0;
......
...@@ -946,7 +946,7 @@ void ...@@ -946,7 +946,7 @@ void
i915_gem_execbuffer_move_to_active(struct list_head *vmas, i915_gem_execbuffer_move_to_active(struct list_head *vmas,
struct intel_engine_cs *ring) struct intel_engine_cs *ring)
{ {
u32 seqno = intel_ring_get_seqno(ring); struct drm_i915_gem_request *req = intel_ring_get_request(ring);
struct i915_vma *vma; struct i915_vma *vma;
list_for_each_entry(vma, vmas, exec_list) { list_for_each_entry(vma, vmas, exec_list) {
...@@ -963,7 +963,7 @@ i915_gem_execbuffer_move_to_active(struct list_head *vmas, ...@@ -963,7 +963,7 @@ i915_gem_execbuffer_move_to_active(struct list_head *vmas,
i915_vma_move_to_active(vma, ring); i915_vma_move_to_active(vma, ring);
if (obj->base.write_domain) { if (obj->base.write_domain) {
obj->dirty = 1; obj->dirty = 1;
obj->last_write_seqno = seqno; i915_gem_request_assign(&obj->last_write_req, req);
intel_fb_obj_invalidate(obj, ring); intel_fb_obj_invalidate(obj, ring);
...@@ -971,7 +971,7 @@ i915_gem_execbuffer_move_to_active(struct list_head *vmas, ...@@ -971,7 +971,7 @@ i915_gem_execbuffer_move_to_active(struct list_head *vmas,
obj->base.write_domain &= ~I915_GEM_GPU_DOMAINS; obj->base.write_domain &= ~I915_GEM_GPU_DOMAINS;
} }
if (entry->flags & EXEC_OBJECT_NEEDS_FENCE) { if (entry->flags & EXEC_OBJECT_NEEDS_FENCE) {
obj->last_fenced_seqno = seqno; i915_gem_request_assign(&obj->last_fenced_req, req);
if (entry->flags & __EXEC_OBJECT_HAS_FENCE) { if (entry->flags & __EXEC_OBJECT_HAS_FENCE) {
struct drm_i915_private *dev_priv = to_i915(ring->dev); struct drm_i915_private *dev_priv = to_i915(ring->dev);
list_move_tail(&dev_priv->fence_regs[obj->fence_reg].lru_list, list_move_tail(&dev_priv->fence_regs[obj->fence_reg].lru_list,
......
...@@ -181,7 +181,7 @@ struct i915_address_space { ...@@ -181,7 +181,7 @@ struct i915_address_space {
* List of objects currently involved in rendering. * List of objects currently involved in rendering.
* *
* Includes buffers having the contents of their GPU caches * Includes buffers having the contents of their GPU caches
* flushed, not necessarily primitives. last_rendering_seqno * flushed, not necessarily primitives. last_read_req
* represents when the rendering involved will be completed. * represents when the rendering involved will be completed.
* *
* A reference is held on the buffer while on this list. * A reference is held on the buffer while on this list.
...@@ -192,7 +192,7 @@ struct i915_address_space { ...@@ -192,7 +192,7 @@ struct i915_address_space {
* LRU list of objects which are not in the ringbuffer and * LRU list of objects which are not in the ringbuffer and
* are ready to unbind, but are still in the GTT. * are ready to unbind, but are still in the GTT.
* *
* last_rendering_seqno is 0 while an object is in this list. * last_read_req is NULL while an object is in this list.
* *
* A reference is not held on the buffer while on this list, * A reference is not held on the buffer while on this list,
* as merely being GTT-bound shouldn't prevent its being * as merely being GTT-bound shouldn't prevent its being
......
...@@ -399,7 +399,7 @@ i915_gem_set_tiling(struct drm_device *dev, void *data, ...@@ -399,7 +399,7 @@ i915_gem_set_tiling(struct drm_device *dev, void *data,
} }
obj->fence_dirty = obj->fence_dirty =
obj->last_fenced_seqno || obj->last_fenced_req ||
obj->fence_reg != I915_FENCE_REG_NONE; obj->fence_reg != I915_FENCE_REG_NONE;
obj->tiling_mode = args->tiling_mode; obj->tiling_mode = args->tiling_mode;
......
...@@ -670,8 +670,8 @@ static void capture_bo(struct drm_i915_error_buffer *err, ...@@ -670,8 +670,8 @@ static void capture_bo(struct drm_i915_error_buffer *err,
err->size = obj->base.size; err->size = obj->base.size;
err->name = obj->base.name; err->name = obj->base.name;
err->rseqno = obj->last_read_seqno; err->rseqno = i915_gem_request_get_seqno(obj->last_read_req);
err->wseqno = obj->last_write_seqno; err->wseqno = i915_gem_request_get_seqno(obj->last_write_req);
err->gtt_offset = vma->node.start; err->gtt_offset = vma->node.start;
err->read_domains = obj->base.read_domains; err->read_domains = obj->base.read_domains;
err->write_domain = obj->base.write_domain; err->write_domain = obj->base.write_domain;
......
...@@ -9637,7 +9637,8 @@ static int intel_queue_mmio_flip(struct drm_device *dev, ...@@ -9637,7 +9637,8 @@ static int intel_queue_mmio_flip(struct drm_device *dev,
{ {
struct intel_crtc *intel_crtc = to_intel_crtc(crtc); struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
intel_crtc->mmio_flip.seqno = obj->last_write_seqno; intel_crtc->mmio_flip.seqno =
i915_gem_request_get_seqno(obj->last_write_req);
intel_crtc->mmio_flip.ring = obj->ring; intel_crtc->mmio_flip.ring = obj->ring;
schedule_work(&intel_crtc->mmio_flip.work); schedule_work(&intel_crtc->mmio_flip.work);
...@@ -9900,7 +9901,8 @@ static int intel_crtc_page_flip(struct drm_crtc *crtc, ...@@ -9900,7 +9901,8 @@ static int intel_crtc_page_flip(struct drm_crtc *crtc,
if (ret) if (ret)
goto cleanup_unpin; goto cleanup_unpin;
work->flip_queued_seqno = obj->last_write_seqno; work->flip_queued_seqno =
i915_gem_request_get_seqno(obj->last_write_req);
work->flip_queued_ring = obj->ring; work->flip_queued_ring = obj->ring;
} else { } else {
ret = dev_priv->display.queue_flip(dev, crtc, fb, obj, ring, ret = dev_priv->display.queue_flip(dev, crtc, fb, obj, ring,
......
...@@ -251,7 +251,7 @@ struct intel_engine_cs { ...@@ -251,7 +251,7 @@ struct intel_engine_cs {
* ringbuffer. * ringbuffer.
* *
* Includes buffers having the contents of their GPU caches * Includes buffers having the contents of their GPU caches
* flushed, not necessarily primitives. last_rendering_seqno * flushed, not necessarily primitives. last_read_req
* represents when the rendering involved will be completed. * represents when the rendering involved will be completed.
* *
* A reference is held on the buffer while on this list. * A reference is held on the buffer while on this list.
......
Markdown is supported
0%
or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment