diff options
author | Dave Airlie <airlied@redhat.com> | 2018-11-19 04:07:52 +0300 |
---|---|---|
committer | Dave Airlie <airlied@redhat.com> | 2018-11-19 04:07:52 +0300 |
commit | 9235dd441af43599b9cdcce599a3da4083fcad3c (patch) | |
tree | 5f8a79cc2d378f05e807c6c5e388394b8e86319d /drivers/gpu/drm/ttm | |
parent | d7563c55ef9fc1fd2301b8708b3c1f53509d6745 (diff) | |
parent | 36b486bc682114a2f1001cbf1a87f21ae381bfc1 (diff) | |
download | linux-9235dd441af43599b9cdcce599a3da4083fcad3c.tar.xz |
Merge branch 'drm-next-4.21' of git://people.freedesktop.org/~agd5f/linux into drm-next
New features for 4.21:
amdgpu:
- Support for SDMA paging queue on vega
- Put compute EOP buffers into vram for better performance
- Share more code with amdkfd
- Support for scanout with DCC on gfx9
- Initial kerneldoc for DC
- Updated SMU firmware support for gfx8 chips
- Rework CSA handling for eventual support for preemption
- XGMI PSP support
- Clean up RLC handling
- Enable GPU reset by default on VI, SOC15 dGPUs
- Ring and IB test cleanups
amdkfd:
- Share more code with amdgpu
ttm:
- Move global init out of the drivers
scheduler:
- Track if schedulers are ready for work
- Timeout/fault handling changes to facilitate GPU recovery
Signed-off-by: Dave Airlie <airlied@redhat.com>
From: Alex Deucher <alexdeucher@gmail.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20181114165113.3751-1-alexander.deucher@amd.com
Diffstat (limited to 'drivers/gpu/drm/ttm')
-rw-r--r-- | drivers/gpu/drm/ttm/ttm_bo.c | 69 | ||||
-rw-r--r-- | drivers/gpu/drm/ttm/ttm_execbuf_util.c | 2 | ||||
-rw-r--r-- | drivers/gpu/drm/ttm/ttm_memory.c | 14 |
3 files changed, 49 insertions, 36 deletions
diff --git a/drivers/gpu/drm/ttm/ttm_bo.c b/drivers/gpu/drm/ttm/ttm_bo.c index 83b4657ffb10..d87935bf8e30 100644 --- a/drivers/gpu/drm/ttm/ttm_bo.c +++ b/drivers/gpu/drm/ttm/ttm_bo.c @@ -45,6 +45,14 @@ static void ttm_bo_global_kobj_release(struct kobject *kobj); +/** + * ttm_global_mutex - protecting the global BO state + */ +DEFINE_MUTEX(ttm_global_mutex); +struct ttm_bo_global ttm_bo_glob = { + .use_count = 0 +}; + static struct attribute ttm_bo_count = { .name = "bo_count", .mode = S_IRUGO @@ -1519,35 +1527,45 @@ static void ttm_bo_global_kobj_release(struct kobject *kobj) container_of(kobj, struct ttm_bo_global, kobj); __free_page(glob->dummy_read_page); - kfree(glob); } -void ttm_bo_global_release(struct drm_global_reference *ref) +static void ttm_bo_global_release(void) { - struct ttm_bo_global *glob = ref->object; + struct ttm_bo_global *glob = &ttm_bo_glob; + + mutex_lock(&ttm_global_mutex); + if (--glob->use_count > 0) + goto out; kobject_del(&glob->kobj); kobject_put(&glob->kobj); + ttm_mem_global_release(&ttm_mem_glob); +out: + mutex_unlock(&ttm_global_mutex); } -EXPORT_SYMBOL(ttm_bo_global_release); -int ttm_bo_global_init(struct drm_global_reference *ref) +static int ttm_bo_global_init(void) { - struct ttm_bo_global_ref *bo_ref = - container_of(ref, struct ttm_bo_global_ref, ref); - struct ttm_bo_global *glob = ref->object; - int ret; + struct ttm_bo_global *glob = &ttm_bo_glob; + int ret = 0; unsigned i; - mutex_init(&glob->device_list_mutex); + mutex_lock(&ttm_global_mutex); + if (++glob->use_count > 1) + goto out; + + ret = ttm_mem_global_init(&ttm_mem_glob); + if (ret) + goto out; + spin_lock_init(&glob->lru_lock); - glob->mem_glob = bo_ref->mem_glob; + glob->mem_glob = &ttm_mem_glob; glob->mem_glob->bo_glob = glob; glob->dummy_read_page = alloc_page(__GFP_ZERO | GFP_DMA32); if (unlikely(glob->dummy_read_page == NULL)) { ret = -ENOMEM; - goto out_no_drp; + goto out; } for (i = 0; i < TTM_MAX_BO_PRIORITY; ++i) @@ -1559,13 +1577,10 @@ int ttm_bo_global_init(struct drm_global_reference *ref) &glob->kobj, &ttm_bo_glob_kobj_type, ttm_get_kobj(), "buffer_objects"); if (unlikely(ret != 0)) kobject_put(&glob->kobj); - return ret; -out_no_drp: - kfree(glob); +out: + mutex_unlock(&ttm_global_mutex); return ret; } -EXPORT_SYMBOL(ttm_bo_global_init); - int ttm_bo_device_release(struct ttm_bo_device *bdev) { @@ -1587,9 +1602,9 @@ int ttm_bo_device_release(struct ttm_bo_device *bdev) } } - mutex_lock(&glob->device_list_mutex); + mutex_lock(&ttm_global_mutex); list_del(&bdev->device_list); - mutex_unlock(&glob->device_list_mutex); + mutex_unlock(&ttm_global_mutex); cancel_delayed_work_sync(&bdev->wq); @@ -1604,18 +1619,25 @@ int ttm_bo_device_release(struct ttm_bo_device *bdev) drm_vma_offset_manager_destroy(&bdev->vma_manager); + if (!ret) + ttm_bo_global_release(); + return ret; } EXPORT_SYMBOL(ttm_bo_device_release); int ttm_bo_device_init(struct ttm_bo_device *bdev, - struct ttm_bo_global *glob, struct ttm_bo_driver *driver, struct address_space *mapping, uint64_t file_page_offset, bool need_dma32) { - int ret = -EINVAL; + struct ttm_bo_global *glob = &ttm_bo_glob; + int ret; + + ret = ttm_bo_global_init(); + if (ret) + return ret; bdev->driver = driver; @@ -1636,12 +1658,13 @@ int ttm_bo_device_init(struct ttm_bo_device *bdev, bdev->dev_mapping = mapping; bdev->glob = glob; bdev->need_dma32 = need_dma32; - mutex_lock(&glob->device_list_mutex); + mutex_lock(&ttm_global_mutex); list_add_tail(&bdev->device_list, &glob->device_list); - mutex_unlock(&glob->device_list_mutex); + mutex_unlock(&ttm_global_mutex); return 0; out_no_sys: + ttm_bo_global_release(); return ret; } EXPORT_SYMBOL(ttm_bo_device_init); diff --git a/drivers/gpu/drm/ttm/ttm_execbuf_util.c b/drivers/gpu/drm/ttm/ttm_execbuf_util.c index e493edb0d3e7..efa005a1c1b7 100644 --- a/drivers/gpu/drm/ttm/ttm_execbuf_util.c +++ b/drivers/gpu/drm/ttm/ttm_execbuf_util.c @@ -187,14 +187,12 @@ void ttm_eu_fence_buffer_objects(struct ww_acquire_ctx *ticket, struct ttm_buffer_object *bo; struct ttm_bo_global *glob; struct ttm_bo_device *bdev; - struct ttm_bo_driver *driver; if (list_empty(list)) return; bo = list_first_entry(list, struct ttm_validate_buffer, head)->bo; bdev = bo->bdev; - driver = bdev->driver; glob = bo->bdev->glob; spin_lock(&glob->lru_lock); diff --git a/drivers/gpu/drm/ttm/ttm_memory.c b/drivers/gpu/drm/ttm/ttm_memory.c index 450387c92b63..f1567c353b54 100644 --- a/drivers/gpu/drm/ttm/ttm_memory.c +++ b/drivers/gpu/drm/ttm/ttm_memory.c @@ -41,6 +41,9 @@ #define TTM_MEMORY_ALLOC_RETRIES 4 +struct ttm_mem_global ttm_mem_glob; +EXPORT_SYMBOL(ttm_mem_glob); + struct ttm_mem_zone { struct kobject kobj; struct ttm_mem_global *glob; @@ -216,14 +219,6 @@ static ssize_t ttm_mem_global_store(struct kobject *kobj, return size; } -static void ttm_mem_global_kobj_release(struct kobject *kobj) -{ - struct ttm_mem_global *glob = - container_of(kobj, struct ttm_mem_global, kobj); - - kfree(glob); -} - static struct attribute *ttm_mem_global_attrs[] = { &ttm_mem_global_lower_mem_limit, NULL @@ -235,7 +230,6 @@ static const struct sysfs_ops ttm_mem_global_ops = { }; static struct kobj_type ttm_mem_glob_kobj_type = { - .release = &ttm_mem_global_kobj_release, .sysfs_ops = &ttm_mem_global_ops, .default_attrs = ttm_mem_global_attrs, }; @@ -464,7 +458,6 @@ out_no_zone: ttm_mem_global_release(glob); return ret; } -EXPORT_SYMBOL(ttm_mem_global_init); void ttm_mem_global_release(struct ttm_mem_global *glob) { @@ -486,7 +479,6 @@ void ttm_mem_global_release(struct ttm_mem_global *glob) kobject_del(&glob->kobj); kobject_put(&glob->kobj); } -EXPORT_SYMBOL(ttm_mem_global_release); static void ttm_check_swapping(struct ttm_mem_global *glob) { |