[PATCH 12/24] drm/radeon: rip out the ib pool v2
j.glisse at gmail.com
j.glisse at gmail.com
Wed Apr 25 12:03:17 PDT 2012
From: Christian König <deathsimple at vodafone.de>
It isn't necessary any more and the suballocator
seems to perform even better.
v2: drop debugfs
Signed-off-by: Christian König <deathsimple at vodafone.de>
---
drivers/gpu/drm/radeon/radeon.h | 22 +---
drivers/gpu/drm/radeon/radeon_device.c | 1 -
drivers/gpu/drm/radeon/radeon_fence.c | 44 ++++++-
drivers/gpu/drm/radeon/radeon_gart.c | 12 +-
drivers/gpu/drm/radeon/radeon_ring.c | 187 ++++++++++-------------------
drivers/gpu/drm/radeon/radeon_semaphore.c | 6 +-
6 files changed, 116 insertions(+), 156 deletions(-)
diff --git a/drivers/gpu/drm/radeon/radeon.h b/drivers/gpu/drm/radeon/radeon.h
index 7f73f40..575dc05 100644
--- a/drivers/gpu/drm/radeon/radeon.h
+++ b/drivers/gpu/drm/radeon/radeon.h
@@ -249,6 +249,8 @@ extern void evergreen_tiling_fields(unsigned tiling_flags, unsigned *bankw,
/*
* Fences.
*/
+struct radeon_ib;
+
struct radeon_fence_driver {
uint32_t scratch_reg;
uint64_t gpu_addr;
@@ -259,7 +261,6 @@ struct radeon_fence_driver {
wait_queue_head_t queue;
struct list_head created;
struct list_head emitted;
- struct list_head signaled;
bool initialized;
};
@@ -274,6 +275,7 @@ struct radeon_fence {
/* RB, DMA, etc. */
int ring;
struct radeon_semaphore *semaphore;
+ struct radeon_ib *ib;
};
int radeon_fence_driver_start_ring(struct radeon_device *rdev, int ring);
@@ -289,6 +291,7 @@ int radeon_fence_wait_empty(struct radeon_device *rdev, int ring);
struct radeon_fence *radeon_fence_ref(struct radeon_fence *fence);
void radeon_fence_unref(struct radeon_fence **fence);
int radeon_fence_count_emitted(struct radeon_device *rdev, int ring);
+bool radeon_fence_set_associated_ib(struct radeon_fence *fence, struct radeon_ib *ib);
/*
* Tiling registers
@@ -603,7 +606,6 @@ void radeon_irq_kms_pflip_irq_put(struct radeon_device *rdev, int crtc);
struct radeon_ib {
struct radeon_sa_bo sa_bo;
- unsigned idx;
uint32_t length_dw;
uint64_t gpu_addr;
uint32_t *ptr;
@@ -612,18 +614,6 @@ struct radeon_ib {
bool is_const_ib;
};
-/*
- * locking -
- * mutex protects scheduled_ibs, ready, alloc_bm
- */
-struct radeon_ib_pool {
- struct radeon_mutex mutex;
- struct radeon_sa_manager sa_manager;
- struct radeon_ib ibs[RADEON_IB_POOL_SIZE];
- bool ready;
- unsigned head_id;
-};
-
struct radeon_ring {
struct radeon_bo *ring_obj;
volatile uint32_t *ring;
@@ -764,7 +754,6 @@ struct si_rlc {
int radeon_ib_get(struct radeon_device *rdev, int ring,
struct radeon_ib **ib, unsigned size);
void radeon_ib_free(struct radeon_device *rdev, struct radeon_ib **ib);
-bool radeon_ib_try_free(struct radeon_device *rdev, struct radeon_ib *ib);
int radeon_ib_schedule(struct radeon_device *rdev, struct radeon_ib *ib);
int radeon_ib_pool_init(struct radeon_device *rdev);
void radeon_ib_pool_fini(struct radeon_device *rdev);
@@ -1497,7 +1486,8 @@ struct radeon_device {
rwlock_t fence_lock;
struct radeon_fence_driver fence_drv[RADEON_NUM_RINGS];
struct radeon_ring ring[RADEON_NUM_RINGS];
- struct radeon_ib_pool ib_pool;
+ bool ib_pool_ready;
+ struct radeon_sa_manager sa_manager;
struct radeon_irq irq;
struct radeon_asic *asic;
struct radeon_gem gem;
diff --git a/drivers/gpu/drm/radeon/radeon_device.c b/drivers/gpu/drm/radeon/radeon_device.c
index d77f19f..26fb9da 100644
--- a/drivers/gpu/drm/radeon/radeon_device.c
+++ b/drivers/gpu/drm/radeon/radeon_device.c
@@ -723,7 +723,6 @@ int radeon_device_init(struct radeon_device *rdev,
/* mutex initialization are all done here so we
* can recall function without having locking issues */
radeon_mutex_init(&rdev->cs_mutex);
- radeon_mutex_init(&rdev->ib_pool.mutex);
for (i = 0; i < RADEON_NUM_RINGS; ++i)
mutex_init(&rdev->ring[i].mutex);
mutex_init(&rdev->dc_hw_i2c_mutex);
diff --git a/drivers/gpu/drm/radeon/radeon_fence.c b/drivers/gpu/drm/radeon/radeon_fence.c
index c96d112..abd10f9 100644
--- a/drivers/gpu/drm/radeon/radeon_fence.c
+++ b/drivers/gpu/drm/radeon/radeon_fence.c
@@ -83,7 +83,8 @@ int radeon_fence_emit(struct radeon_device *rdev, struct radeon_fence *fence)
return 0;
}
-static bool radeon_fence_poll_locked(struct radeon_device *rdev, int ring)
+static bool radeon_fence_poll_locked(struct radeon_device *rdev, int ring,
+ struct list_head *signaled)
{
struct radeon_fence *fence;
struct list_head *i, *n;
@@ -110,7 +111,7 @@ static bool radeon_fence_poll_locked(struct radeon_device *rdev, int ring)
i = n;
do {
n = i->prev;
- list_move_tail(i, &rdev->fence_drv[ring].signaled);
+ list_move_tail(i, signaled);
fence = list_entry(i, struct radeon_fence, list);
fence->signaled = true;
i = n;
@@ -120,6 +121,18 @@ static bool radeon_fence_poll_locked(struct radeon_device *rdev, int ring)
return wake;
}
+static void radeon_fence_process_signaled(struct radeon_device *rdev, struct list_head *signaled)
+{
+ struct radeon_fence *fence;
+ struct list_head *i, *n;
+
+ list_for_each_safe(i, n, signaled) {
+ fence = list_entry(i, struct radeon_fence, list);
+ list_del_init(&fence->list);
+ radeon_ib_free(rdev, &fence->ib);
+ }
+}
+
static void radeon_fence_destroy(struct kref *kref)
{
unsigned long irq_flags;
@@ -152,6 +165,7 @@ int radeon_fence_create(struct radeon_device *rdev,
(*fence)->seq = 0;
(*fence)->ring = ring;
(*fence)->semaphore = NULL;
+ (*fence)->ib = NULL;
INIT_LIST_HEAD(&(*fence)->list);
write_lock_irqsave(&rdev->fence_lock, irq_flags);
@@ -164,6 +178,7 @@ bool radeon_fence_signaled(struct radeon_fence *fence)
{
unsigned long irq_flags;
bool signaled = false;
+ LIST_HEAD(siglist);
if (!fence)
return true;
@@ -179,10 +194,12 @@ bool radeon_fence_signaled(struct radeon_fence *fence)
signaled = true;
}
if (!signaled) {
- radeon_fence_poll_locked(fence->rdev, fence->ring);
+ radeon_fence_poll_locked(fence->rdev, fence->ring, &siglist);
signaled = fence->signaled;
}
write_unlock_irqrestore(&fence->rdev->fence_lock, irq_flags);
+ radeon_fence_process_signaled(fence->rdev, &siglist);
+
return signaled;
}
@@ -341,10 +358,12 @@ void radeon_fence_process(struct radeon_device *rdev, int ring)
{
unsigned long irq_flags;
bool wake;
+ LIST_HEAD(signaled);
write_lock_irqsave(&rdev->fence_lock, irq_flags);
- wake = radeon_fence_poll_locked(rdev, ring);
+ wake = radeon_fence_poll_locked(rdev, ring, &signaled);
write_unlock_irqrestore(&rdev->fence_lock, irq_flags);
+ radeon_fence_process_signaled(rdev, &signaled);
if (wake) {
wake_up_all(&rdev->fence_drv[ring].queue);
}
@@ -373,6 +392,22 @@ int radeon_fence_count_emitted(struct radeon_device *rdev, int ring)
return not_processed;
}
+bool radeon_fence_set_associated_ib(struct radeon_fence *fence, struct radeon_ib *ib)
+{
+ struct radeon_device *rdev = fence->rdev;
+ unsigned long irq_flags;
+ bool isset = false;
+
+ /* a readlock is suficient, cause this should be called only once */
+ read_lock_irqsave(&rdev->fence_lock, irq_flags);
+ if (fence->emitted && !fence->signaled) {
+ fence->ib = ib;
+ isset = true;
+ }
+ read_unlock_irqrestore(&rdev->fence_lock, irq_flags);
+ return isset;
+}
+
int radeon_fence_driver_start_ring(struct radeon_device *rdev, int ring)
{
unsigned long irq_flags;
@@ -413,7 +448,6 @@ static void radeon_fence_driver_init_ring(struct radeon_device *rdev, int ring)
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;
}
diff --git a/drivers/gpu/drm/radeon/radeon_gart.c b/drivers/gpu/drm/radeon/radeon_gart.c
index 7af4ff9..220dbb8 100644
--- a/drivers/gpu/drm/radeon/radeon_gart.c
+++ b/drivers/gpu/drm/radeon/radeon_gart.c
@@ -434,8 +434,8 @@ retry_id:
rdev->vm_manager.use_bitmap |= 1 << id;
vm->id = id;
list_add_tail(&vm->list, &rdev->vm_manager.lru_vm);
- return radeon_vm_bo_update_pte(rdev, vm, rdev->ib_pool.sa_manager.bo,
- &rdev->ib_pool.sa_manager.bo->tbo.mem);
+ return radeon_vm_bo_update_pte(rdev, vm, rdev->sa_manager.bo,
+ &rdev->sa_manager.bo->tbo.mem);
}
/* object have to be reserved */
@@ -633,7 +633,7 @@ int radeon_vm_init(struct radeon_device *rdev, struct radeon_vm *vm)
/* map the ib pool buffer at 0 in virtual address space, set
* read only
*/
- r = radeon_vm_bo_add(rdev, vm, rdev->ib_pool.sa_manager.bo, 0,
+ r = radeon_vm_bo_add(rdev, vm, rdev->sa_manager.bo, 0,
RADEON_VM_PAGE_READABLE | RADEON_VM_PAGE_SNOOPED);
return r;
}
@@ -650,12 +650,12 @@ void radeon_vm_fini(struct radeon_device *rdev, struct radeon_vm *vm)
radeon_mutex_unlock(&rdev->cs_mutex);
/* remove all bo */
- r = radeon_bo_reserve(rdev->ib_pool.sa_manager.bo, false);
+ r = radeon_bo_reserve(rdev->sa_manager.bo, false);
if (!r) {
- bo_va = radeon_bo_va(rdev->ib_pool.sa_manager.bo, vm);
+ bo_va = radeon_bo_va(rdev->sa_manager.bo, vm);
list_del_init(&bo_va->bo_list);
list_del_init(&bo_va->vm_list);
- radeon_bo_unreserve(rdev->ib_pool.sa_manager.bo);
+ radeon_bo_unreserve(rdev->sa_manager.bo);
kfree(bo_va);
}
if (!list_empty(&vm->va)) {
diff --git a/drivers/gpu/drm/radeon/radeon_ring.c b/drivers/gpu/drm/radeon/radeon_ring.c
index b06e04f..c8deae9 100644
--- a/drivers/gpu/drm/radeon/radeon_ring.c
+++ b/drivers/gpu/drm/radeon/radeon_ring.c
@@ -24,6 +24,7 @@
* Authors: Dave Airlie
* Alex Deucher
* Jerome Glisse
+ * Christian König
*/
#include <linux/seq_file.h>
#include <linux/slab.h>
@@ -33,6 +34,9 @@
#include "radeon.h"
#include "atom.h"
+/*
+ * IB.
+ */
u32 radeon_get_ib_value(struct radeon_cs_parser *p, int idx)
{
struct radeon_cs_chunk *ibc = &p->chunks[p->chunk_ib_idx];
@@ -58,123 +62,62 @@ u32 radeon_get_ib_value(struct radeon_cs_parser *p, int idx)
return idx_value;
}
-void radeon_ring_write(struct radeon_ring *ring, uint32_t v)
-{
-#if DRM_DEBUG_CODE
- if (ring->count_dw <= 0) {
- DRM_ERROR("radeon: writting more dword to ring than expected !\n");
- }
-#endif
- ring->ring[ring->wptr++] = v;
- ring->wptr &= ring->ptr_mask;
- ring->count_dw--;
- ring->ring_free_dw--;
-}
-
-/*
- * IB.
- */
-bool radeon_ib_try_free(struct radeon_device *rdev, struct radeon_ib *ib)
-{
- bool done = false;
-
- /* only free ib which have been emited */
- if (ib->fence && ib->fence->emitted) {
- if (radeon_fence_signaled(ib->fence)) {
- radeon_fence_unref(&ib->fence);
- radeon_sa_bo_free(rdev, &ib->sa_bo);
- done = true;
- }
- }
- return done;
-}
-
int radeon_ib_get(struct radeon_device *rdev, int ring,
struct radeon_ib **ib, unsigned size)
{
- struct radeon_fence *fence;
- unsigned cretry = 0;
- int r = 0, i, idx;
+ int r;
- *ib = NULL;
- /* align size on 256 bytes */
- size = ALIGN(size, 256);
+ *ib = kmalloc(sizeof(struct radeon_ib), GFP_KERNEL);
+ if (*ib == NULL) {
+ return -ENOMEM;
+ }
- r = radeon_fence_create(rdev, &fence, ring);
+ r = radeon_sa_bo_new(rdev, &rdev->sa_manager, &(*ib)->sa_bo, size, 256, true);
if (r) {
- dev_err(rdev->dev, "failed to create fence for new IB\n");
+ dev_err(rdev->dev, "failed to get a new IB (%d)\n", r);
+ kfree(*ib);
+ *ib = NULL;
return r;
}
- radeon_mutex_lock(&rdev->ib_pool.mutex);
- idx = rdev->ib_pool.head_id;
-retry:
- if (cretry > 5) {
- dev_err(rdev->dev, "failed to get an ib after 5 retry\n");
- radeon_mutex_unlock(&rdev->ib_pool.mutex);
- radeon_fence_unref(&fence);
- return -ENOMEM;
- }
- cretry++;
- for (i = 0; i < RADEON_IB_POOL_SIZE; i++) {
- radeon_ib_try_free(rdev, &rdev->ib_pool.ibs[idx]);
- if (rdev->ib_pool.ibs[idx].fence == NULL) {
- r = radeon_sa_bo_new(rdev, &rdev->ib_pool.sa_manager,
- &rdev->ib_pool.ibs[idx].sa_bo,
- size, 256, false);
- if (!r) {
- *ib = &rdev->ib_pool.ibs[idx];
- (*ib)->ptr = rdev->ib_pool.sa_manager.cpu_ptr;
- (*ib)->ptr += ((*ib)->sa_bo.offset >> 2);
- (*ib)->gpu_addr = rdev->ib_pool.sa_manager.gpu_addr;
- (*ib)->gpu_addr += (*ib)->sa_bo.offset;
- (*ib)->fence = fence;
- (*ib)->vm_id = 0;
- (*ib)->is_const_ib = false;
- /* ib are most likely to be allocated in a ring fashion
- * thus rdev->ib_pool.head_id should be the id of the
- * oldest ib
- */
- rdev->ib_pool.head_id = (1 + idx);
- rdev->ib_pool.head_id &= (RADEON_IB_POOL_SIZE - 1);
- radeon_mutex_unlock(&rdev->ib_pool.mutex);
- return 0;
- }
- }
- idx = (idx + 1) & (RADEON_IB_POOL_SIZE - 1);
- }
- /* this should be rare event, ie all ib scheduled none signaled yet.
- */
- for (i = 0; i < RADEON_IB_POOL_SIZE; i++) {
- if (rdev->ib_pool.ibs[idx].fence && rdev->ib_pool.ibs[idx].fence->emitted) {
- r = radeon_fence_wait(rdev->ib_pool.ibs[idx].fence, false);
- if (!r) {
- goto retry;
- }
- /* an error happened */
- break;
- }
- idx = (idx + 1) & (RADEON_IB_POOL_SIZE - 1);
+ r = radeon_fence_create(rdev, &(*ib)->fence, ring);
+ if (r) {
+ dev_err(rdev->dev, "failed to create fence for new IB (%d)\n", r);
+ radeon_sa_bo_free(rdev, &(*ib)->sa_bo);
+ kfree(*ib);
+ *ib = NULL;
+ return r;
}
- radeon_mutex_unlock(&rdev->ib_pool.mutex);
- radeon_fence_unref(&fence);
- return r;
+
+ (*ib)->ptr = rdev->sa_manager.cpu_ptr;
+ (*ib)->ptr += ((*ib)->sa_bo.offset >> 2);
+ (*ib)->gpu_addr = rdev->sa_manager.gpu_addr;
+ (*ib)->gpu_addr += (*ib)->sa_bo.offset;
+ (*ib)->vm_id = 0;
+ (*ib)->is_const_ib = false;
+
+ return 0;
}
void radeon_ib_free(struct radeon_device *rdev, struct radeon_ib **ib)
{
struct radeon_ib *tmp = *ib;
+ bool destroy = true;
*ib = NULL;
if (tmp == NULL) {
return;
}
- radeon_mutex_lock(&rdev->ib_pool.mutex);
- if (tmp->fence && !tmp->fence->emitted) {
+
+ if (tmp->fence) {
+ destroy = !radeon_fence_set_associated_ib(tmp->fence, tmp);
+ }
+
+ if (destroy) {
radeon_sa_bo_free(rdev, &tmp->sa_bo);
radeon_fence_unref(&tmp->fence);
+ kfree(tmp);
}
- radeon_mutex_unlock(&rdev->ib_pool.mutex);
}
int radeon_ib_schedule(struct radeon_device *rdev, struct radeon_ib *ib)
@@ -184,7 +127,7 @@ int radeon_ib_schedule(struct radeon_device *rdev, struct radeon_ib *ib)
if (!ib->length_dw || !ring->ready) {
/* TODO: Nothings in the ib we should report. */
- DRM_ERROR("radeon: couldn't schedule IB(%u).\n", ib->idx);
+ DRM_ERROR("radeon: couldn't schedule IB.\n");
return -EINVAL;
}
@@ -202,61 +145,40 @@ int radeon_ib_schedule(struct radeon_device *rdev, struct radeon_ib *ib)
int radeon_ib_pool_init(struct radeon_device *rdev)
{
- int i, r;
+ int r;
- radeon_mutex_lock(&rdev->ib_pool.mutex);
- if (rdev->ib_pool.ready) {
+ if (rdev->ib_pool_ready) {
return 0;
}
- rdev->ib_pool.ready = true;
- radeon_mutex_unlock(&rdev->ib_pool.mutex);
- r = radeon_sa_bo_manager_init(rdev, &rdev->ib_pool.sa_manager,
+ r = radeon_sa_bo_manager_init(rdev, &rdev->sa_manager,
RADEON_IB_POOL_SIZE*64*1024,
RADEON_GEM_DOMAIN_GTT);
if (r) {
return r;
}
- radeon_mutex_lock(&rdev->ib_pool.mutex);
- for (i = 0; i < RADEON_IB_POOL_SIZE; i++) {
- rdev->ib_pool.ibs[i].fence = NULL;
- rdev->ib_pool.ibs[i].idx = i;
- rdev->ib_pool.ibs[i].length_dw = 0;
- INIT_LIST_HEAD(&rdev->ib_pool.ibs[i].sa_bo.list);
- }
- rdev->ib_pool.head_id = 0;
- rdev->ib_pool.ready = true;
DRM_INFO("radeon: ib pool ready.\n");
-
- radeon_mutex_unlock(&rdev->ib_pool.mutex);
+ rdev->ib_pool_ready = true;
return 0;
}
void radeon_ib_pool_fini(struct radeon_device *rdev)
{
- unsigned i;
-
- radeon_mutex_lock(&rdev->ib_pool.mutex);
- if (rdev->ib_pool.ready) {
- for (i = 0; i < RADEON_IB_POOL_SIZE; i++) {
- radeon_sa_bo_free(rdev, &rdev->ib_pool.ibs[i].sa_bo);
- radeon_fence_unref(&rdev->ib_pool.ibs[i].fence);
- }
- radeon_sa_bo_manager_fini(rdev, &rdev->ib_pool.sa_manager);
- rdev->ib_pool.ready = false;
+ if (rdev->ib_pool_ready) {
+ radeon_sa_bo_manager_fini(rdev, &rdev->sa_manager);
+ rdev->ib_pool_ready = false;
}
- radeon_mutex_unlock(&rdev->ib_pool.mutex);
}
int radeon_ib_pool_start(struct radeon_device *rdev)
{
- return radeon_sa_bo_manager_start(rdev, &rdev->ib_pool.sa_manager);
+ return radeon_sa_bo_manager_start(rdev, &rdev->sa_manager);
}
int radeon_ib_pool_suspend(struct radeon_device *rdev)
{
- return radeon_sa_bo_manager_suspend(rdev, &rdev->ib_pool.sa_manager);
+ return radeon_sa_bo_manager_suspend(rdev, &rdev->sa_manager);
}
int radeon_ib_ring_tests(struct radeon_device *rdev)
@@ -292,6 +214,21 @@ int radeon_ib_ring_tests(struct radeon_device *rdev)
/*
* Ring.
*/
+int radeon_debugfs_ring_init(struct radeon_device *rdev, struct radeon_ring *ring);
+
+void radeon_ring_write(struct radeon_ring *ring, uint32_t v)
+{
+#if DRM_DEBUG_CODE
+ if (ring->count_dw <= 0) {
+ DRM_ERROR("radeon: writting more dword to ring than expected !\n");
+ }
+#endif
+ ring->ring[ring->wptr++] = v;
+ ring->wptr &= ring->ptr_mask;
+ ring->count_dw--;
+ ring->ring_free_dw--;
+}
+
int radeon_ring_index(struct radeon_device *rdev, struct radeon_ring *ring)
{
/* r1xx-r5xx only has CP ring */
diff --git a/drivers/gpu/drm/radeon/radeon_semaphore.c b/drivers/gpu/drm/radeon/radeon_semaphore.c
index b67c259..9e2cf40 100644
--- a/drivers/gpu/drm/radeon/radeon_semaphore.c
+++ b/drivers/gpu/drm/radeon/radeon_semaphore.c
@@ -42,7 +42,7 @@ int radeon_semaphore_create(struct radeon_device *rdev,
return -ENOMEM;
}
- r = radeon_sa_bo_new(rdev, &rdev->ib_pool.sa_manager,
+ r = radeon_sa_bo_new(rdev, &rdev->sa_manager,
&(*semaphore)->sa_bo, 8, 8, true);
if (r) {
kfree(*semaphore);
@@ -51,10 +51,10 @@ int radeon_semaphore_create(struct radeon_device *rdev,
}
(*semaphore)->waiters = 0;
- (*semaphore)->gpu_addr = rdev->ib_pool.sa_manager.gpu_addr;
+ (*semaphore)->gpu_addr = rdev->sa_manager.gpu_addr;
(*semaphore)->gpu_addr += (*semaphore)->sa_bo.offset;
- cpu_ptr = rdev->ib_pool.sa_manager.cpu_ptr;
+ cpu_ptr = rdev->sa_manager.cpu_ptr;
cpu_ptr += (*semaphore)->sa_bo.offset;
*((uint64_t*)cpu_ptr) = 0;
--
1.7.7.6
More information about the dri-devel
mailing list