diff options
author | Linus Torvalds <torvalds@linux-foundation.org> | 2012-01-10 11:04:36 -0800 |
---|---|---|
committer | Linus Torvalds <torvalds@linux-foundation.org> | 2012-01-10 11:04:36 -0800 |
commit | 1a464cbb3d483f2f195b614cffa4aa1b910a0440 (patch) | |
tree | af57dee6436532dbb546b8670e9e1f6910d489b5 /drivers/gpu/drm/radeon/radeon_fence.c | |
parent | dbe950f201a8edd353b0bd9079e8d536ee4ce37c (diff) | |
parent | 095f979a539245a46b9e5d600ec9c720b4d928e5 (diff) |
Merge branch 'drm-core-next' of git://people.freedesktop.org/~airlied/linux
* 'drm-core-next' of git://people.freedesktop.org/~airlied/linux: (307 commits)
drm/nouveau/pm: fix build with HWMON off
gma500: silence gcc warnings in mid_get_vbt_data()
drm/ttm: fix condition (and vs or)
drm/radeon: double lock typo in radeon_vm_bo_rmv()
drm/radeon: use after free in radeon_vm_bo_add()
drm/sis|via: don't return stack garbage from free_mem ioctl
drm/radeon/kms: remove pointless CS flags priority struct
drm/radeon/kms: check if vm is supported in VA ioctl
drm: introduce drm_can_sleep and use in intel/radeon drivers. (v2)
radeon: Fix disabling PCI bus mastering on big endian hosts.
ttm: fix agp since ttm tt rework
agp: Fix multi-line warning message whitespace
drm/ttm/dma: Fix accounting error when calling ttm_mem_global_free_page and don't try to free freed pages.
drm/ttm/dma: Only call set_pages_array_wb when the page is not in WB pool.
drm/radeon/kms: sync across multiple rings when doing bo moves v3
drm/radeon/kms: Add support for multi-ring sync in CS ioctl (v2)
drm/radeon: GPU virtual memory support v22
drm: make DRM_UNLOCKED ioctls with their own mutex
drm: no need to hold global mutex for static data
drm/radeon/benchmark: common modes sweep ignores 640x480@32
...
Fix up trivial conflicts in radeon/evergreen.c and vmwgfx/vmwgfx_kms.c
Diffstat (limited to 'drivers/gpu/drm/radeon/radeon_fence.c')
-rw-r--r-- | drivers/gpu/drm/radeon/radeon_fence.c | 307 |
1 files changed, 186 insertions, 121 deletions
diff --git a/drivers/gpu/drm/radeon/radeon_fence.c b/drivers/gpu/drm/radeon/radeon_fence.c index 76ec0e9ed8a..64ea3dd9e6f 100644 --- a/drivers/gpu/drm/radeon/radeon_fence.c +++ b/drivers/gpu/drm/radeon/radeon_fence.c @@ -40,32 +40,24 @@ #include "radeon.h" #include "radeon_trace.h" -static void radeon_fence_write(struct radeon_device *rdev, u32 seq) +static void radeon_fence_write(struct radeon_device *rdev, u32 seq, int ring) { if (rdev->wb.enabled) { - u32 scratch_index; - if (rdev->wb.use_event) - scratch_index = R600_WB_EVENT_OFFSET + rdev->fence_drv.scratch_reg - rdev->scratch.reg_base; - else - scratch_index = RADEON_WB_SCRATCH_OFFSET + rdev->fence_drv.scratch_reg - rdev->scratch.reg_base; - rdev->wb.wb[scratch_index/4] = cpu_to_le32(seq); - } else - WREG32(rdev->fence_drv.scratch_reg, seq); + *rdev->fence_drv[ring].cpu_addr = cpu_to_le32(seq); + } else { + WREG32(rdev->fence_drv[ring].scratch_reg, seq); + } } -static u32 radeon_fence_read(struct radeon_device *rdev) +static u32 radeon_fence_read(struct radeon_device *rdev, int ring) { - u32 seq; + u32 seq = 0; if (rdev->wb.enabled) { - u32 scratch_index; - if (rdev->wb.use_event) - scratch_index = R600_WB_EVENT_OFFSET + rdev->fence_drv.scratch_reg - rdev->scratch.reg_base; - else - scratch_index = RADEON_WB_SCRATCH_OFFSET + rdev->fence_drv.scratch_reg - rdev->scratch.reg_base; - seq = le32_to_cpu(rdev->wb.wb[scratch_index/4]); - } else - seq = RREG32(rdev->fence_drv.scratch_reg); + seq = le32_to_cpu(*rdev->fence_drv[ring].cpu_addr); + } else { + seq = RREG32(rdev->fence_drv[ring].scratch_reg); + } return seq; } @@ -73,28 +65,28 @@ int radeon_fence_emit(struct radeon_device *rdev, struct radeon_fence *fence) { unsigned long irq_flags; - write_lock_irqsave(&rdev->fence_drv.lock, irq_flags); - if (fence->emited) { - write_unlock_irqrestore(&rdev->fence_drv.lock, irq_flags); + write_lock_irqsave(&rdev->fence_lock, irq_flags); + if (fence->emitted) { + write_unlock_irqrestore(&rdev->fence_lock, irq_flags); return 0; } - fence->seq = atomic_add_return(1, &rdev->fence_drv.seq); - if (!rdev->cp.ready) + fence->seq = atomic_add_return(1, &rdev->fence_drv[fence->ring].seq); + if (!rdev->ring[fence->ring].ready) /* FIXME: cp is not running assume everythings is done right * away */ - radeon_fence_write(rdev, fence->seq); + radeon_fence_write(rdev, fence->seq, fence->ring); else - radeon_fence_ring_emit(rdev, fence); + radeon_fence_ring_emit(rdev, fence->ring, fence); trace_radeon_fence_emit(rdev->ddev, fence->seq); - fence->emited = true; - list_move_tail(&fence->list, &rdev->fence_drv.emited); - write_unlock_irqrestore(&rdev->fence_drv.lock, irq_flags); + fence->emitted = true; + list_move_tail(&fence->list, &rdev->fence_drv[fence->ring].emitted); + write_unlock_irqrestore(&rdev->fence_lock, irq_flags); return 0; } -static bool radeon_fence_poll_locked(struct radeon_device *rdev) +static bool radeon_fence_poll_locked(struct radeon_device *rdev, int ring) { struct radeon_fence *fence; struct list_head *i, *n; @@ -102,34 +94,34 @@ static bool radeon_fence_poll_locked(struct radeon_device *rdev) bool wake = false; unsigned long cjiffies; - seq = radeon_fence_read(rdev); - if (seq != rdev->fence_drv.last_seq) { - rdev->fence_drv.last_seq = seq; - rdev->fence_drv.last_jiffies = jiffies; - rdev->fence_drv.last_timeout = RADEON_FENCE_JIFFIES_TIMEOUT; + seq = radeon_fence_read(rdev, ring); + if (seq != rdev->fence_drv[ring].last_seq) { + rdev->fence_drv[ring].last_seq = seq; + rdev->fence_drv[ring].last_jiffies = jiffies; + rdev->fence_drv[ring].last_timeout = RADEON_FENCE_JIFFIES_TIMEOUT; } else { cjiffies = jiffies; - if (time_after(cjiffies, rdev->fence_drv.last_jiffies)) { - cjiffies -= rdev->fence_drv.last_jiffies; - if (time_after(rdev->fence_drv.last_timeout, cjiffies)) { + if (time_after(cjiffies, rdev->fence_drv[ring].last_jiffies)) { + cjiffies -= rdev->fence_drv[ring].last_jiffies; + if (time_after(rdev->fence_drv[ring].last_timeout, cjiffies)) { /* update the timeout */ - rdev->fence_drv.last_timeout -= cjiffies; + rdev->fence_drv[ring].last_timeout -= cjiffies; } else { /* the 500ms timeout is elapsed we should test * for GPU lockup */ - rdev->fence_drv.last_timeout = 1; + rdev->fence_drv[ring].last_timeout = 1; } } else { /* wrap around update last jiffies, we will just wait * a little longer */ - rdev->fence_drv.last_jiffies = cjiffies; + rdev->fence_drv[ring].last_jiffies = cjiffies; } return false; } n = NULL; - list_for_each(i, &rdev->fence_drv.emited) { + list_for_each(i, &rdev->fence_drv[ring].emitted) { fence = list_entry(i, struct radeon_fence, list); if (fence->seq == seq) { n = i; @@ -141,11 +133,11 @@ static bool radeon_fence_poll_locked(struct radeon_device *rdev) i = n; do { n = i->prev; - list_move_tail(i, &rdev->fence_drv.signaled); + list_move_tail(i, &rdev->fence_drv[ring].signaled); fence = list_entry(i, struct radeon_fence, list); fence->signaled = true; i = n; - } while (i != &rdev->fence_drv.emited); + } while (i != &rdev->fence_drv[ring].emitted); wake = true; } return wake; @@ -157,14 +149,18 @@ static void radeon_fence_destroy(struct kref *kref) struct radeon_fence *fence; fence = container_of(kref, struct radeon_fence, kref); - write_lock_irqsave(&fence->rdev->fence_drv.lock, irq_flags); + write_lock_irqsave(&fence->rdev->fence_lock, irq_flags); list_del(&fence->list); - fence->emited = false; - write_unlock_irqrestore(&fence->rdev->fence_drv.lock, irq_flags); + fence->emitted = false; + write_unlock_irqrestore(&fence->rdev->fence_lock, irq_flags); + if (fence->semaphore) + radeon_semaphore_free(fence->rdev, fence->semaphore); kfree(fence); } -int radeon_fence_create(struct radeon_device *rdev, struct radeon_fence **fence) +int radeon_fence_create(struct radeon_device *rdev, + struct radeon_fence **fence, + int ring) { unsigned long irq_flags; @@ -174,18 +170,19 @@ int radeon_fence_create(struct radeon_device *rdev, struct radeon_fence **fence) } kref_init(&((*fence)->kref)); (*fence)->rdev = rdev; - (*fence)->emited = false; + (*fence)->emitted = false; (*fence)->signaled = false; (*fence)->seq = 0; + (*fence)->ring = ring; + (*fence)->semaphore = NULL; INIT_LIST_HEAD(&(*fence)->list); - write_lock_irqsave(&rdev->fence_drv.lock, irq_flags); - list_add_tail(&(*fence)->list, &rdev->fence_drv.created); - write_unlock_irqrestore(&rdev->fence_drv.lock, irq_flags); + write_lock_irqsave(&rdev->fence_lock, irq_flags); + list_add_tail(&(*fence)->list, &rdev->fence_drv[ring].created); + write_unlock_irqrestore(&rdev->fence_lock, irq_flags); return 0; } - bool radeon_fence_signaled(struct radeon_fence *fence) { unsigned long irq_flags; @@ -197,21 +194,21 @@ bool radeon_fence_signaled(struct radeon_fence *fence) if (fence->rdev->gpu_lockup) return true; - write_lock_irqsave(&fence->rdev->fence_drv.lock, irq_flags); + write_lock_irqsave(&fence->rdev->fence_lock, irq_flags); signaled = fence->signaled; /* if we are shuting down report all fence as signaled */ if (fence->rdev->shutdown) { signaled = true; } - if (!fence->emited) { - WARN(1, "Querying an unemited fence : %p !\n", fence); + if (!fence->emitted) { + WARN(1, "Querying an unemitted fence : %p !\n", fence); signaled = true; } if (!signaled) { - radeon_fence_poll_locked(fence->rdev); + radeon_fence_poll_locked(fence->rdev, fence->ring); signaled = fence->signaled; } - write_unlock_irqrestore(&fence->rdev->fence_drv.lock, irq_flags); + write_unlock_irqrestore(&fence->rdev->fence_lock, irq_flags); return signaled; } @@ -230,24 +227,24 @@ int radeon_fence_wait(struct radeon_fence *fence, bool intr) if (radeon_fence_signaled(fence)) { return 0; } - timeout = rdev->fence_drv.last_timeout; + timeout = rdev->fence_drv[fence->ring].last_timeout; retry: /* save current sequence used to check for GPU lockup */ - seq = rdev->fence_drv.last_seq; + seq = rdev->fence_drv[fence->ring].last_seq; trace_radeon_fence_wait_begin(rdev->ddev, seq); if (intr) { - radeon_irq_kms_sw_irq_get(rdev); - r = wait_event_interruptible_timeout(rdev->fence_drv.queue, + radeon_irq_kms_sw_irq_get(rdev, fence->ring); + r = wait_event_interruptible_timeout(rdev->fence_drv[fence->ring].queue, radeon_fence_signaled(fence), timeout); - radeon_irq_kms_sw_irq_put(rdev); + radeon_irq_kms_sw_irq_put(rdev, fence->ring); if (unlikely(r < 0)) { return r; } } else { - radeon_irq_kms_sw_irq_get(rdev); - r = wait_event_timeout(rdev->fence_drv.queue, + radeon_irq_kms_sw_irq_get(rdev, fence->ring); + r = wait_event_timeout(rdev->fence_drv[fence->ring].queue, radeon_fence_signaled(fence), timeout); - radeon_irq_kms_sw_irq_put(rdev); + radeon_irq_kms_sw_irq_put(rdev, fence->ring); } trace_radeon_fence_wait_end(rdev->ddev, seq); if (unlikely(!radeon_fence_signaled(fence))) { @@ -258,10 +255,11 @@ retry: timeout = r; goto retry; } - /* don't protect read access to rdev->fence_drv.last_seq + /* don't protect read access to rdev->fence_drv[t].last_seq * if we experiencing a lockup the value doesn't change */ - if (seq == rdev->fence_drv.last_seq && radeon_gpu_is_lockup(rdev)) { + if (seq == rdev->fence_drv[fence->ring].last_seq && + radeon_gpu_is_lockup(rdev, &rdev->ring[fence->ring])) { /* good news we believe it's a lockup */ printk(KERN_WARNING "GPU lockup (waiting for 0x%08X last fence id 0x%08X)\n", fence->seq, seq); @@ -272,20 +270,20 @@ retry: r = radeon_gpu_reset(rdev); if (r) return r; - radeon_fence_write(rdev, fence->seq); + radeon_fence_write(rdev, fence->seq, fence->ring); rdev->gpu_lockup = false; } timeout = RADEON_FENCE_JIFFIES_TIMEOUT; - write_lock_irqsave(&rdev->fence_drv.lock, irq_flags); - rdev->fence_drv.last_timeout = RADEON_FENCE_JIFFIES_TIMEOUT; - rdev->fence_drv.last_jiffies = jiffies; - write_unlock_irqrestore(&rdev->fence_drv.lock, irq_flags); + write_lock_irqsave(&rdev->fence_lock, irq_flags); + rdev->fence_drv[fence->ring].last_timeout = RADEON_FENCE_JIFFIES_TIMEOUT; + rdev->fence_drv[fence->ring].last_jiffies = jiffies; + write_unlock_irqrestore(&rdev->fence_lock, irq_flags); goto retry; } return 0; } -int radeon_fence_wait_next(struct radeon_device *rdev) +int radeon_fence_wait_next(struct radeon_device *rdev, int ring) { unsigned long irq_flags; struct radeon_fence *fence; @@ -294,21 +292,21 @@ int radeon_fence_wait_next(struct radeon_device *rdev) if (rdev->gpu_lockup) { return 0; } - write_lock_irqsave(&rdev->fence_drv.lock, irq_flags); - if (list_empty(&rdev->fence_drv.emited)) { - write_unlock_irqrestore(&rdev->fence_drv.lock, irq_flags); + write_lock_irqsave(&rdev->fence_lock, irq_flags); + if (list_empty(&rdev->fence_drv[ring].emitted)) { + write_unlock_irqrestore(&rdev->fence_lock, irq_flags); return 0; } - fence = list_entry(rdev->fence_drv.emited.next, + fence = list_entry(rdev->fence_drv[ring].emitted.next, struct radeon_fence, list); radeon_fence_ref(fence); - write_unlock_irqrestore(&rdev->fence_drv.lock, irq_flags); + write_unlock_irqrestore(&rdev->fence_lock, irq_flags); r = radeon_fence_wait(fence, false); radeon_fence_unref(&fence); return r; } -int radeon_fence_wait_last(struct radeon_device *rdev) +int radeon_fence_wait_last(struct radeon_device *rdev, int ring) { unsigned long irq_flags; struct radeon_fence *fence; @@ -317,15 +315,15 @@ int radeon_fence_wait_last(struct radeon_device *rdev) if (rdev->gpu_lockup) { return 0; } - write_lock_irqsave(&rdev->fence_drv.lock, irq_flags); - if (list_empty(&rdev->fence_drv.emited)) { - write_unlock_irqrestore(&rdev->fence_drv.lock, irq_flags); + write_lock_irqsave(&rdev->fence_lock, irq_flags); + if (list_empty(&rdev->fence_drv[ring].emitted)) { + write_unlock_irqrestore(&rdev->fence_lock, irq_flags); return 0; } - fence = list_entry(rdev->fence_drv.emited.prev, + fence = list_entry(rdev->fence_drv[ring].emitted.prev, struct radeon_fence, list); radeon_fence_ref(fence); - write_unlock_irqrestore(&rdev->fence_drv.lock, irq_flags); + write_unlock_irqrestore(&rdev->fence_lock, irq_flags); r = radeon_fence_wait(fence, false); radeon_fence_unref(&fence); return r; @@ -347,39 +345,95 @@ void radeon_fence_unref(struct radeon_fence **fence) } } -void radeon_fence_process(struct radeon_device *rdev) +void radeon_fence_process(struct radeon_device *rdev, int ring) { unsigned long irq_flags; bool wake; - write_lock_irqsave(&rdev->fence_drv.lock, irq_flags); - wake = radeon_fence_poll_locked(rdev); - write_unlock_irqrestore(&rdev->fence_drv.lock, irq_flags); + write_lock_irqsave(&rdev->fence_lock, irq_flags); + wake = radeon_fence_poll_locked(rdev, ring); + write_unlock_irqrestore(&rdev->fence_lock, irq_flags); if (wake) { - wake_up_all(&rdev->fence_drv.queue); + wake_up_all(&rdev->fence_drv[ring].queue); } } -int radeon_fence_driver_init(struct radeon_device *rdev) +int radeon_fence_count_emitted(struct radeon_device *rdev, int ring) +{ + unsigned long irq_flags; + int not_processed = 0; + + read_lock_irqsave(&rdev->fence_lock, irq_flags); + if (!rdev->fence_drv[ring].initialized) + return 0; + + if (!list_empty(&rdev->fence_drv[ring].emitted)) { + struct list_head *ptr; + list_for_each(ptr, &rdev->fence_drv[ring].emitted) { + /* count up to 3, that's enought info */ + if (++not_processed >= 3) + break; + } + } + read_unlock_irqrestore(&rdev->fence_lock, irq_flags); + return not_processed; +} + +int radeon_fence_driver_start_ring(struct radeon_device *rdev, int ring) { unsigned long irq_flags; + uint64_t index; int r; - write_lock_irqsave(&rdev->fence_drv.lock, irq_flags); - r = radeon_scratch_get(rdev, &rdev->fence_drv.scratch_reg); - if (r) { - dev_err(rdev->dev, "fence failed to get scratch register\n"); - write_unlock_irqrestore(&rdev->fence_drv.lock, irq_flags); - return r; + write_lock_irqsave(&rdev->fence_lock, irq_flags); + radeon_scratch_free(rdev, rdev->fence_drv[ring].scratch_reg); + if (rdev->wb.use_event) { + rdev->fence_drv[ring].scratch_reg = 0; + index = R600_WB_EVENT_OFFSET + ring * 4; + } else { + r = radeon_scratch_get(rdev, &rdev->fence_drv[ring].scratch_reg); + if (r) { + dev_err(rdev->dev, "fence failed to get scratch register\n"); + write_unlock_irqrestore(&rdev->fence_lock, irq_flags); + return r; + } + index = RADEON_WB_SCRATCH_OFFSET + + rdev->fence_drv[ring].scratch_reg - + rdev->scratch.reg_base; } - radeon_fence_write(rdev, 0); - atomic_set(&rdev->fence_drv.seq, 0); - INIT_LIST_HEAD(&rdev->fence_drv.created); - INIT_LIST_HEAD(&rdev->fence_drv.emited); - INIT_LIST_HEAD(&rdev->fence_drv.signaled); - init_waitqueue_head(&rdev->fence_drv.queue); - rdev->fence_drv.initialized = true; - write_unlock_irqrestore(&rdev->fence_drv.lock, irq_flags); + rdev->fence_drv[ring].cpu_addr = &rdev->wb.wb[index/4]; + rdev->fence_drv[ring].gpu_addr = rdev->wb.gpu_addr + index; + radeon_fence_write(rdev, atomic_read(&rdev->fence_drv[ring].seq), ring); + rdev->fence_drv[ring].initialized = true; + DRM_INFO("fence driver on ring %d use gpu addr 0x%08Lx and cpu addr 0x%p\n", + ring, rdev->fence_drv[ring].gpu_addr, rdev->fence_drv[ring].cpu_addr); + write_unlock_irqrestore(&rdev->fence_lock, irq_flags); + return 0; +} + +static void radeon_fence_driver_init_ring(struct radeon_device *rdev, int ring) +{ + rdev->fence_drv[ring].scratch_reg = -1; + rdev->fence_drv[ring].cpu_addr = NULL; + rdev->fence_drv[ring].gpu_addr = 0; + atomic_set(&rdev->fence_drv[ring].seq, 0); + INIT_LIST_HEAD(&rdev->fence_drv[ring].created); + INIT_LIST_HEAD(&rdev->fence_drv[ring].emitted); + INIT_LIST_HEAD(&rdev->fence_drv[ring].signaled); + init_waitqueue_head(&rdev->fence_drv[ring].queue); + rdev->fence_drv[ring].initialized = false; +} + +int radeon_fence_driver_init(struct radeon_device *rdev) +{ + unsigned long irq_flags; + int ring; + + write_lock_irqsave(&rdev->fence_lock, irq_flags); + for (ring = 0; ring < RADEON_NUM_RINGS; ring++) { + radeon_fence_driver_init_ring(rdev, ring); + } + write_unlock_irqrestore(&rdev->fence_lock, irq_flags); if (radeon_debugfs_fence_init(rdev)) { dev_err(rdev->dev, "fence debugfs file creation failed\n"); } @@ -389,14 +443,18 @@ int radeon_fence_driver_init(struct radeon_device *rdev) void radeon_fence_driver_fini(struct radeon_device *rdev) { unsigned long irq_flags; - - if (!rdev->fence_drv.initialized) - return; - wake_up_all(&rdev->fence_drv.queue); - write_lock_irqsave(&rdev->fence_drv.lock, irq_flags); - radeon_scratch_free(rdev, rdev->fence_drv.scratch_reg); - write_unlock_irqrestore(&rdev->fence_drv.lock, irq_flags); - rdev->fence_drv.initialized = false; + int ring; + + for (ring = 0; ring < RADEON_NUM_RINGS; ring++) { + if (!rdev->fence_drv[ring].initialized) + continue; + radeon_fence_wait_last(rdev, ring); + wake_up_all(&rdev->fence_drv[ring].queue); + write_lock_irqsave(&rdev->fence_lock, irq_flags); + radeon_scratch_free(rdev, rdev->fence_drv[ring].scratch_reg); + write_unlock_irqrestore(&rdev->fence_lock, irq_flags); + rdev->fence_drv[ring].initialized = false; + } } @@ -410,14 +468,21 @@ static int radeon_debugfs_fence_info(struct seq_file *m, void *data) struct drm_device *dev = node->minor->dev; struct radeon_device *rdev = dev->dev_private; struct radeon_fence *fence; - - seq_printf(m, "Last signaled fence 0x%08X\n", - radeon_fence_read(rdev)); - if (!list_empty(&rdev->fence_drv.emited)) { - fence = list_entry(rdev->fence_drv.emited.prev, - struct radeon_fence, list); - seq_printf(m, "Last emited fence %p with 0x%08X\n", - fence, fence->seq); + int i; + + for (i = 0; i < RADEON_NUM_RINGS; ++i) { + if (!rdev->fence_drv[i].initialized) + continue; + + seq_printf(m, "--- ring %d ---\n", i); + seq_printf(m, "Last signaled fence 0x%08X\n", + radeon_fence_read(rdev, i)); + if (!list_empty(&rdev->fence_drv[i].emitted)) { + fence = list_entry(rdev->fence_drv[i].emitted.prev, + struct radeon_fence, list); + seq_printf(m, "Last emitted fence %p with 0x%08X\n", + fence, fence->seq); + } } return 0; } |