| Message ID | 20190425092004.9995-24-chris@chris-wilson.co.uk |
|---|---|
| State | New |
| Headers | show |
| Series |
"Series without cover letter"
( rev:
2
1
)
in
Intel GFX |
diff --git a/drivers/gpu/drm/i915/Makefile b/drivers/gpu/drm/i915/Makefile index 58643373495c..def781c9ea69 100644 --- a/drivers/gpu/drm/i915/Makefile +++ b/drivers/gpu/drm/i915/Makefile @@ -85,6 +85,7 @@ gt-$(CONFIG_DRM_I915_SELFTEST) += \ i915-y += $(gt-y) # GEM (Graphics Execution Management) code +obj-y += gem/ i915-y += \ i915_active.o \ i915_cmd_parser.o \ diff --git a/drivers/gpu/drm/i915/gem/Makefile b/drivers/gpu/drm/i915/gem/Makefile new file mode 100644 index 000000000000..07e7b8b840ea --- /dev/null +++ b/drivers/gpu/drm/i915/gem/Makefile @@ -0,0 +1 @@ +include $(src)/Makefile.header-test # Extra header tests diff --git a/drivers/gpu/drm/i915/gem/Makefile.header-test b/drivers/gpu/drm/i915/gem/Makefile.header-test new file mode 100644 index 000000000000..61e06cbb4b32 --- /dev/null +++ b/drivers/gpu/drm/i915/gem/Makefile.header-test @@ -0,0 +1,16 @@ +# SPDX-License-Identifier: MIT +# Copyright © 2019 Intel Corporation + +# Test the headers are compilable as standalone units +header_test := $(notdir $(wildcard $(src)/*.h)) + +quiet_cmd_header_test = HDRTEST $@ + cmd_header_test = echo "\#include \"$(<F)\"" > $@ + +header_test_%.c: %.h + $(call cmd,header_test) + +extra-$(CONFIG_DRM_I915_WERROR) += \ + $(foreach h,$(header_test),$(patsubst %.h,header_test_%.o,$(h))) + +clean-files += $(foreach h,$(header_test),$(patsubst %.h,header_test_%.c,$(h))) diff --git a/drivers/gpu/drm/i915/gem/i915_gem_object_types.h b/drivers/gpu/drm/i915/gem/i915_gem_object_types.h new file mode 100644 index 000000000000..e4b50944f553 --- /dev/null +++ b/drivers/gpu/drm/i915/gem/i915_gem_object_types.h @@ -0,0 +1,285 @@ +/* + * SPDX-License-Identifier: MIT + * + * Copyright © 2016 Intel Corporation + */ + +#ifndef __I915_GEM_OBJECT_TYPES_H__ +#define __I915_GEM_OBJECT_TYPES_H__ + +#include <linux/reservation.h> + +#include <drm/drm_gem.h> + +#include "../i915_active.h" +#include "../i915_selftest.h" + +struct drm_i915_gem_object; + +/* + * struct i915_lut_handle tracks the fast lookups from handle to vma used + * for execbuf. Although we use a radixtree for that mapping, in order to + * remove them as the object or context is closed, we need a secondary list + * and a translation entry (i915_lut_handle). + */ +struct i915_lut_handle { + struct list_head obj_link; + struct list_head ctx_link; + struct i915_gem_context *ctx; + u32 handle; +}; + +struct drm_i915_gem_object_ops { + unsigned int flags; +#define I915_GEM_OBJECT_HAS_STRUCT_PAGE BIT(0) +#define I915_GEM_OBJECT_IS_SHRINKABLE BIT(1) +#define I915_GEM_OBJECT_IS_PROXY BIT(2) +#define I915_GEM_OBJECT_ASYNC_CANCEL BIT(3) + + /* Interface between the GEM object and its backing storage. + * get_pages() is called once prior to the use of the associated set + * of pages before to binding them into the GTT, and put_pages() is + * called after we no longer need them. As we expect there to be + * associated cost with migrating pages between the backing storage + * and making them available for the GPU (e.g. clflush), we may hold + * onto the pages after they are no longer referenced by the GPU + * in case they may be used again shortly (for example migrating the + * pages to a different memory domain within the GTT). put_pages() + * will therefore most likely be called when the object itself is + * being released or under memory pressure (where we attempt to + * reap pages for the shrinker). + */ + int (*get_pages)(struct drm_i915_gem_object *obj); + void (*put_pages)(struct drm_i915_gem_object *obj, + struct sg_table *pages); + + int (*pwrite)(struct drm_i915_gem_object *obj, + const struct drm_i915_gem_pwrite *arg); + + int (*dmabuf_export)(struct drm_i915_gem_object *obj); + void (*release)(struct drm_i915_gem_object *obj); +}; + +struct drm_i915_gem_object { + struct drm_gem_object base; + + const struct drm_i915_gem_object_ops *ops; + + struct { + /** + * @vma.lock: protect the list/tree of vmas + */ + spinlock_t lock; + + /** + * @vma.list: List of VMAs backed by this object + * + * The VMA on this list are ordered by type, all GGTT vma are + * placed at the head and all ppGTT vma are placed at the tail. + * The different types of GGTT vma are unordered between + * themselves, use the @vma.tree (which has a defined order + * between all VMA) to quickly find an exact match. + */ + struct list_head list; + + /** + * @vma.tree: Ordered tree of VMAs backed by this object + * + * All VMA created for this object are placed in the @vma.tree + * for fast retrieval via a binary search in + * i915_vma_instance(). They are also added to @vma.list for + * easy iteration. + */ + struct rb_root tree; + } vma; + + /** + * @lut_list: List of vma lookup entries in use for this object. + * + * If this object is closed, we need to remove all of its VMA from + * the fast lookup index in associated contexts; @lut_list provides + * this translation from object to context->handles_vma. + */ + struct list_head lut_list; + + /** Stolen memory for this object, instead of being backed by shmem. */ + struct drm_mm_node *stolen; + union { + struct rcu_head rcu; + struct llist_node freed; + }; + + /** + * Whether the object is currently in the GGTT mmap. + */ + unsigned int userfault_count; + struct list_head userfault_link; + + struct list_head batch_pool_link; + I915_SELFTEST_DECLARE(struct list_head st_link); + + unsigned long flags; + + /** + * Have we taken a reference for the object for incomplete GPU + * activity? + */ +#define I915_BO_ACTIVE_REF 0 + + /* + * Is the object to be mapped as read-only to the GPU + * Only honoured if hardware has relevant pte bit + */ + unsigned int cache_level:3; + unsigned int cache_coherent:2; +#define I915_BO_CACHE_COHERENT_FOR_READ BIT(0) +#define I915_BO_CACHE_COHERENT_FOR_WRITE BIT(1) + unsigned int cache_dirty:1; + + /** + * @read_domains: Read memory domains. + * + * These monitor which caches contain read/write data related to the + * object. When transitioning from one set of domains to another, + * the driver is called to ensure that caches are suitably flushed and + * invalidated. + */ + u16 read_domains; + + /** + * @write_domain: Corresponding unique write memory domain. + */ + u16 write_domain; + + atomic_t frontbuffer_bits; + unsigned int frontbuffer_ggtt_origin; /* write once */ + struct i915_active_request frontbuffer_write; + + /** Current tiling stride for the object, if it's tiled. */ + unsigned int tiling_and_stride; +#define FENCE_MINIMUM_STRIDE 128 /* See i915_tiling_ok() */ +#define TILING_MASK (FENCE_MINIMUM_STRIDE - 1) +#define STRIDE_MASK (~TILING_MASK) + + /** Count of VMA actually bound by this object */ + unsigned int bind_count; + unsigned int active_count; + /** Count of how many global VMA are currently pinned for use by HW */ + unsigned int pin_global; + + struct { + struct mutex lock; /* protects the pages and their use */ + atomic_t pages_pin_count; + + struct sg_table *pages; + void *mapping; + + /* TODO: whack some of this into the error state */ + struct i915_page_sizes { + /** + * The sg mask of the pages sg_table. i.e the mask of + * of the lengths for each sg entry. + */ + unsigned int phys; + + /** + * The gtt page sizes we are allowed to use given the + * sg mask and the supported page sizes. This will + * express the smallest unit we can use for the whole + * object, as well as the larger sizes we may be able + * to use opportunistically. + */ + unsigned int sg; + + /** + * The actual gtt page size usage. Since we can have + * multiple vma associated with this object we need to + * prevent any trampling of state, hence a copy of this + * struct also lives in each vma, therefore the gtt + * value here should only be read/write through the vma. + */ + unsigned int gtt; + } page_sizes; + + I915_SELFTEST_DECLARE(unsigned int page_mask); + + struct i915_gem_object_page_iter { + struct scatterlist *sg_pos; + unsigned int sg_idx; /* in pages, but 32bit eek! */ + + struct radix_tree_root radix; + struct mutex lock; /* protects this cache */ + } get_page; + + /** + * Element within i915->mm.unbound_list or i915->mm.bound_list, + * locked by i915->mm.obj_lock. + */ + struct list_head link; + + /** + * Advice: are the backing pages purgeable? + */ + unsigned int madv:2; + + /** + * This is set if the object has been written to since the + * pages were last acquired. + */ + bool dirty:1; + + /** + * This is set if the object has been pinned due to unknown + * swizzling. + */ + bool quirked:1; + } mm; + + /** Breadcrumb of last rendering to the buffer. + * There can only be one writer, but we allow for multiple readers. + * If there is a writer that necessarily implies that all other + * read requests are complete - but we may only be lazily clearing + * the read requests. A read request is naturally the most recent + * request on a ring, so we may have two different write and read + * requests on one ring where the write request is older than the + * read request. This allows for the CPU to read from an active + * buffer by only waiting for the write to complete. + */ + struct reservation_object *resv; + + /** References from framebuffers, locks out tiling changes. */ + unsigned int framebuffer_references; + + /** Record of address bit 17 of each page at last unbind. */ + unsigned long *bit_17; + + union { + struct i915_gem_userptr { + uintptr_t ptr; + + struct i915_mm_struct *mm; + struct i915_mmu_object *mmu_object; + struct work_struct *work; + } userptr; + + unsigned long scratch; + + void *gvt_info; + }; + + /** for phys allocated objects */ + struct drm_dma_handle *phys_handle; + + struct reservation_object __builtin_resv; +}; + +static inline struct drm_i915_gem_object * +to_intel_bo(struct drm_gem_object *gem) +{ + /* Assert that to_intel_bo(NULL) == NULL */ + BUILD_BUG_ON(offsetof(struct drm_i915_gem_object, base)); + + return container_of(gem, struct drm_i915_gem_object, base); +} + +#endif diff --git a/drivers/gpu/drm/i915/gt/intel_engine_types.h b/drivers/gpu/drm/i915/gt/intel_engine_types.h index ca8d95a5708d..ae73c6596d08 100644 --- a/drivers/gpu/drm/i915/gt/intel_engine_types.h +++ b/drivers/gpu/drm/i915/gt/intel_engine_types.h @@ -29,6 +29,7 @@ #define I915_CMD_HASH_ORDER 9 struct dma_fence; +struct drm_i915_gem_object; struct drm_i915_reg_table; struct i915_gem_context; struct i915_request; diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h index 5ca4df9a7428..0cf87495e11b 100644 --- a/drivers/gpu/drm/i915/i915_drv.h +++ b/drivers/gpu/drm/i915/i915_drv.h @@ -80,7 +80,6 @@ #include "i915_gem.h" #include "i915_gem_context.h" #include "i915_gem_fence_reg.h" -#include "i915_gem_object.h" #include "i915_gem_gtt.h" #include "i915_gpu_error.h" #include "i915_request.h" @@ -135,6 +134,8 @@ bool i915_error_injected(void); __i915_printk(i915, i915_error_injected() ? KERN_DEBUG : KERN_ERR, \ fmt, ##__VA_ARGS__) +struct drm_i915_gem_object; + enum hpd_pin { HPD_NONE = 0, HPD_TV = HPD_NONE, /* TV is known to be unreliable */ diff --git a/drivers/gpu/drm/i915/i915_gem_batch_pool.h b/drivers/gpu/drm/i915/i915_gem_batch_pool.h index 56947daaaf65..feeeeeaa54d8 100644 --- a/drivers/gpu/drm/i915/i915_gem_batch_pool.h +++ b/drivers/gpu/drm/i915/i915_gem_batch_pool.h @@ -9,6 +9,7 @@ #include <linux/types.h> +struct drm_i915_gem_object; struct intel_engine_cs; struct i915_gem_batch_pool { @@ -19,7 +20,7 @@ struct i915_gem_batch_pool { void i915_gem_batch_pool_init(struct i915_gem_batch_pool *pool, struct intel_engine_cs *engine); void i915_gem_batch_pool_fini(struct i915_gem_batch_pool *pool); -struct drm_i915_gem_object* +struct drm_i915_gem_object * i915_gem_batch_pool_get(struct i915_gem_batch_pool *pool, size_t size); #endif /* I915_GEM_BATCH_POOL_H */ diff --git a/drivers/gpu/drm/i915/i915_gem_gtt.h b/drivers/gpu/drm/i915/i915_gem_gtt.h index 2fafa04c45ec..593079f30fbe 100644 --- a/drivers/gpu/drm/i915/i915_gem_gtt.h +++ b/drivers/gpu/drm/i915/i915_gem_gtt.h @@ -61,6 +61,7 @@ struct drm_i915_file_private; struct drm_i915_fence_reg; +struct drm_i915_gem_object; struct i915_vma; typedef u32 gen6_pte_t; diff --git a/drivers/gpu/drm/i915/i915_gem_object.h b/drivers/gpu/drm/i915/i915_gem_object.h index ca93a40c0c87..3666b0c5f6ca 100644 --- a/drivers/gpu/drm/i915/i915_gem_object.h +++ b/drivers/gpu/drm/i915/i915_gem_object.h @@ -1,308 +1,19 @@ /* - * Copyright © 2016 Intel Corporation - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice (including the next - * paragraph) shall be included in all copies or substantial portions of the - * Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER - * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING - * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS - * IN THE SOFTWARE. + * SPDX-License-Identifier: MIT * + * Copyright © 2016 Intel Corporation */ #ifndef __I915_GEM_OBJECT_H__ #define __I915_GEM_OBJECT_H__ -#include <linux/reservation.h> - -#include <drm/drm_vma_manager.h> #include <drm/drm_gem.h> #include <drm/drm_file.h> #include <drm/drm_device.h> #include <drm/i915_drm.h> -#include "i915_request.h" -#include "i915_selftest.h" - -struct drm_i915_gem_object; - -/* - * struct i915_lut_handle tracks the fast lookups from handle to vma used - * for execbuf. Although we use a radixtree for that mapping, in order to - * remove them as the object or context is closed, we need a secondary list - * and a translation entry (i915_lut_handle). - */ -struct i915_lut_handle { - struct list_head obj_link; - struct list_head ctx_link; - struct i915_gem_context *ctx; - u32 handle; -}; - -struct drm_i915_gem_object_ops { - unsigned int flags; -#define I915_GEM_OBJECT_HAS_STRUCT_PAGE BIT(0) -#define I915_GEM_OBJECT_IS_SHRINKABLE BIT(1) -#define I915_GEM_OBJECT_IS_PROXY BIT(2) -#define I915_GEM_OBJECT_ASYNC_CANCEL BIT(3) - - /* Interface between the GEM object and its backing storage. - * get_pages() is called once prior to the use of the associated set - * of pages before to binding them into the GTT, and put_pages() is - * called after we no longer need them. As we expect there to be - * associated cost with migrating pages between the backing storage - * and making them available for the GPU (e.g. clflush), we may hold - * onto the pages after they are no longer referenced by the GPU - * in case they may be used again shortly (for example migrating the - * pages to a different memory domain within the GTT). put_pages() - * will therefore most likely be called when the object itself is - * being released or under memory pressure (where we attempt to - * reap pages for the shrinker). - */ - int (*get_pages)(struct drm_i915_gem_object *); - void (*put_pages)(struct drm_i915_gem_object *, struct sg_table *); - - int (*pwrite)(struct drm_i915_gem_object *, - const struct drm_i915_gem_pwrite *); - - int (*dmabuf_export)(struct drm_i915_gem_object *); - void (*release)(struct drm_i915_gem_object *); -}; - -struct drm_i915_gem_object { - struct drm_gem_object base; - - const struct drm_i915_gem_object_ops *ops; - - struct { - /** - * @vma.lock: protect the list/tree of vmas - */ - spinlock_t lock; - - /** - * @vma.list: List of VMAs backed by this object - * - * The VMA on this list are ordered by type, all GGTT vma are - * placed at the head and all ppGTT vma are placed at the tail. - * The different types of GGTT vma are unordered between - * themselves, use the @vma.tree (which has a defined order - * between all VMA) to quickly find an exact match. - */ - struct list_head list; - - /** - * @vma.tree: Ordered tree of VMAs backed by this object - * - * All VMA created for this object are placed in the @vma.tree - * for fast retrieval via a binary search in - * i915_vma_instance(). They are also added to @vma.list for - * easy iteration. - */ - struct rb_root tree; - } vma; - - /** - * @lut_list: List of vma lookup entries in use for this object. - * - * If this object is closed, we need to remove all of its VMA from - * the fast lookup index in associated contexts; @lut_list provides - * this translation from object to context->handles_vma. - */ - struct list_head lut_list; - - /** Stolen memory for this object, instead of being backed by shmem. */ - struct drm_mm_node *stolen; - union { - struct rcu_head rcu; - struct llist_node freed; - }; - - /** - * Whether the object is currently in the GGTT mmap. - */ - unsigned int userfault_count; - struct list_head userfault_link; - - struct list_head batch_pool_link; - I915_SELFTEST_DECLARE(struct list_head st_link); - - unsigned long flags; - - /** - * Have we taken a reference for the object for incomplete GPU - * activity? - */ -#define I915_BO_ACTIVE_REF 0 - - /* - * Is the object to be mapped as read-only to the GPU - * Only honoured if hardware has relevant pte bit - */ - unsigned int cache_level:3; - unsigned int cache_coherent:2; -#define I915_BO_CACHE_COHERENT_FOR_READ BIT(0) -#define I915_BO_CACHE_COHERENT_FOR_WRITE BIT(1) - unsigned int cache_dirty:1; - - /** - * @read_domains: Read memory domains. - * - * These monitor which caches contain read/write data related to the - * object. When transitioning from one set of domains to another, - * the driver is called to ensure that caches are suitably flushed and - * invalidated. - */ - u16 read_domains; - - /** - * @write_domain: Corresponding unique write memory domain. - */ - u16 write_domain; - - atomic_t frontbuffer_bits; - unsigned int frontbuffer_ggtt_origin; /* write once */ - struct i915_active_request frontbuffer_write; - - /** Current tiling stride for the object, if it's tiled. */ - unsigned int tiling_and_stride; -#define FENCE_MINIMUM_STRIDE 128 /* See i915_tiling_ok() */ -#define TILING_MASK (FENCE_MINIMUM_STRIDE-1) -#define STRIDE_MASK (~TILING_MASK) - - /** Count of VMA actually bound by this object */ - unsigned int bind_count; - unsigned int active_count; - /** Count of how many global VMA are currently pinned for use by HW */ - unsigned int pin_global; - - struct { - struct mutex lock; /* protects the pages and their use */ - atomic_t pages_pin_count; - - struct sg_table *pages; - void *mapping; - - /* TODO: whack some of this into the error state */ - struct i915_page_sizes { - /** - * The sg mask of the pages sg_table. i.e the mask of - * of the lengths for each sg entry. - */ - unsigned int phys; - - /** - * The gtt page sizes we are allowed to use given the - * sg mask and the supported page sizes. This will - * express the smallest unit we can use for the whole - * object, as well as the larger sizes we may be able - * to use opportunistically. - */ - unsigned int sg; - - /** - * The actual gtt page size usage. Since we can have - * multiple vma associated with this object we need to - * prevent any trampling of state, hence a copy of this - * struct also lives in each vma, therefore the gtt - * value here should only be read/write through the vma. - */ - unsigned int gtt; - } page_sizes; - - I915_SELFTEST_DECLARE(unsigned int page_mask); - - struct i915_gem_object_page_iter { - struct scatterlist *sg_pos; - unsigned int sg_idx; /* in pages, but 32bit eek! */ - - struct radix_tree_root radix; - struct mutex lock; /* protects this cache */ - } get_page; - - /** - * Element within i915->mm.unbound_list or i915->mm.bound_list, - * locked by i915->mm.obj_lock. - */ - struct list_head link; - - /** - * Advice: are the backing pages purgeable? - */ - unsigned int madv:2; - - /** - * This is set if the object has been written to since the - * pages were last acquired. - */ - bool dirty:1; - - /** - * This is set if the object has been pinned due to unknown - * swizzling. - */ - bool quirked:1; - } mm; - - /** Breadcrumb of last rendering to the buffer. - * There can only be one writer, but we allow for multiple readers. - * If there is a writer that necessarily implies that all other - * read requests are complete - but we may only be lazily clearing - * the read requests. A read request is naturally the most recent - * request on a ring, so we may have two different write and read - * requests on one ring where the write request is older than the - * read request. This allows for the CPU to read from an active - * buffer by only waiting for the write to complete. - */ - struct reservation_object *resv; - - /** References from framebuffers, locks out tiling changes. */ - unsigned int framebuffer_references; - - /** Record of address bit 17 of each page at last unbind. */ - unsigned long *bit_17; - - union { - struct i915_gem_userptr { - uintptr_t ptr; - - struct i915_mm_struct *mm; - struct i915_mmu_object *mmu_object; - struct work_struct *work; - } userptr; - - unsigned long scratch; - - void *gvt_info; - }; - - /** for phys allocated objects */ - struct drm_dma_handle *phys_handle; - - struct reservation_object __builtin_resv; -}; - -static inline struct drm_i915_gem_object * -to_intel_bo(struct drm_gem_object *gem) -{ - /* Assert that to_intel_bo(NULL) == NULL */ - BUILD_BUG_ON(offsetof(struct drm_i915_gem_object, base)); - - return container_of(gem, struct drm_i915_gem_object, base); -} +#include "gem/i915_gem_object_types.h" struct drm_i915_gem_object *i915_gem_object_alloc(void); void i915_gem_object_free(struct drm_i915_gem_object *obj);
On Thu, 25 Apr 2019, Chris Wilson <chris@chris-wilson.co.uk> wrote: > For convenience in avoiding inline spaghetti, keep the type definition > as a separate header. > > Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> > Reviewed-by: Matthew Auld <matthew.auld@intel.com> > --- > drivers/gpu/drm/i915/Makefile | 1 + > drivers/gpu/drm/i915/gem/Makefile | 1 + > drivers/gpu/drm/i915/gem/Makefile.header-test | 16 + > .../gpu/drm/i915/gem/i915_gem_object_types.h | 285 +++++++++++++++++ > drivers/gpu/drm/i915/gt/intel_engine_types.h | 1 + > drivers/gpu/drm/i915/i915_drv.h | 3 +- > drivers/gpu/drm/i915/i915_gem_batch_pool.h | 3 +- > drivers/gpu/drm/i915/i915_gem_gtt.h | 1 + > drivers/gpu/drm/i915/i915_gem_object.h | 295 +----------------- > 9 files changed, 312 insertions(+), 294 deletions(-) > create mode 100644 drivers/gpu/drm/i915/gem/Makefile > create mode 100644 drivers/gpu/drm/i915/gem/Makefile.header-test > create mode 100644 drivers/gpu/drm/i915/gem/i915_gem_object_types.h > > diff --git a/drivers/gpu/drm/i915/Makefile b/drivers/gpu/drm/i915/Makefile > index 58643373495c..def781c9ea69 100644 > --- a/drivers/gpu/drm/i915/Makefile > +++ b/drivers/gpu/drm/i915/Makefile > @@ -85,6 +85,7 @@ gt-$(CONFIG_DRM_I915_SELFTEST) += \ > i915-y += $(gt-y) > > # GEM (Graphics Execution Management) code > +obj-y += gem/ I think this warrants similar treatment as gt/. Posted standalone instead of hidden in a series, explicit acks. Provided the split makes sense to Joonas, Acked-by: Jani Nikula <jani.nikula@intel.com> > i915-y += \ > i915_active.o \ > i915_cmd_parser.o \ > diff --git a/drivers/gpu/drm/i915/gem/Makefile b/drivers/gpu/drm/i915/gem/Makefile > new file mode 100644 > index 000000000000..07e7b8b840ea > --- /dev/null > +++ b/drivers/gpu/drm/i915/gem/Makefile > @@ -0,0 +1 @@ > +include $(src)/Makefile.header-test # Extra header tests > diff --git a/drivers/gpu/drm/i915/gem/Makefile.header-test b/drivers/gpu/drm/i915/gem/Makefile.header-test > new file mode 100644 > index 000000000000..61e06cbb4b32 > --- /dev/null > +++ b/drivers/gpu/drm/i915/gem/Makefile.header-test > @@ -0,0 +1,16 @@ > +# SPDX-License-Identifier: MIT > +# Copyright © 2019 Intel Corporation > + > +# Test the headers are compilable as standalone units > +header_test := $(notdir $(wildcard $(src)/*.h)) > + > +quiet_cmd_header_test = HDRTEST $@ > + cmd_header_test = echo "\#include \"$(<F)\"" > $@ > + > +header_test_%.c: %.h > + $(call cmd,header_test) > + > +extra-$(CONFIG_DRM_I915_WERROR) += \ > + $(foreach h,$(header_test),$(patsubst %.h,header_test_%.o,$(h))) > + > +clean-files += $(foreach h,$(header_test),$(patsubst %.h,header_test_%.c,$(h))) > diff --git a/drivers/gpu/drm/i915/gem/i915_gem_object_types.h b/drivers/gpu/drm/i915/gem/i915_gem_object_types.h > new file mode 100644 > index 000000000000..e4b50944f553 > --- /dev/null > +++ b/drivers/gpu/drm/i915/gem/i915_gem_object_types.h > @@ -0,0 +1,285 @@ > +/* > + * SPDX-License-Identifier: MIT > + * > + * Copyright © 2016 Intel Corporation > + */ > + > +#ifndef __I915_GEM_OBJECT_TYPES_H__ > +#define __I915_GEM_OBJECT_TYPES_H__ > + > +#include <linux/reservation.h> > + > +#include <drm/drm_gem.h> > + > +#include "../i915_active.h" > +#include "../i915_selftest.h" > + > +struct drm_i915_gem_object; > + > +/* > + * struct i915_lut_handle tracks the fast lookups from handle to vma used > + * for execbuf. Although we use a radixtree for that mapping, in order to > + * remove them as the object or context is closed, we need a secondary list > + * and a translation entry (i915_lut_handle). > + */ > +struct i915_lut_handle { > + struct list_head obj_link; > + struct list_head ctx_link; > + struct i915_gem_context *ctx; > + u32 handle; > +}; > + > +struct drm_i915_gem_object_ops { > + unsigned int flags; > +#define I915_GEM_OBJECT_HAS_STRUCT_PAGE BIT(0) > +#define I915_GEM_OBJECT_IS_SHRINKABLE BIT(1) > +#define I915_GEM_OBJECT_IS_PROXY BIT(2) > +#define I915_GEM_OBJECT_ASYNC_CANCEL BIT(3) > + > + /* Interface between the GEM object and its backing storage. > + * get_pages() is called once prior to the use of the associated set > + * of pages before to binding them into the GTT, and put_pages() is > + * called after we no longer need them. As we expect there to be > + * associated cost with migrating pages between the backing storage > + * and making them available for the GPU (e.g. clflush), we may hold > + * onto the pages after they are no longer referenced by the GPU > + * in case they may be used again shortly (for example migrating the > + * pages to a different memory domain within the GTT). put_pages() > + * will therefore most likely be called when the object itself is > + * being released or under memory pressure (where we attempt to > + * reap pages for the shrinker). > + */ > + int (*get_pages)(struct drm_i915_gem_object *obj); > + void (*put_pages)(struct drm_i915_gem_object *obj, > + struct sg_table *pages); > + > + int (*pwrite)(struct drm_i915_gem_object *obj, > + const struct drm_i915_gem_pwrite *arg); > + > + int (*dmabuf_export)(struct drm_i915_gem_object *obj); > + void (*release)(struct drm_i915_gem_object *obj); > +}; > + > +struct drm_i915_gem_object { > + struct drm_gem_object base; > + > + const struct drm_i915_gem_object_ops *ops; > + > + struct { > + /** > + * @vma.lock: protect the list/tree of vmas > + */ > + spinlock_t lock; > + > + /** > + * @vma.list: List of VMAs backed by this object > + * > + * The VMA on this list are ordered by type, all GGTT vma are > + * placed at the head and all ppGTT vma are placed at the tail. > + * The different types of GGTT vma are unordered between > + * themselves, use the @vma.tree (which has a defined order > + * between all VMA) to quickly find an exact match. > + */ > + struct list_head list; > + > + /** > + * @vma.tree: Ordered tree of VMAs backed by this object > + * > + * All VMA created for this object are placed in the @vma.tree > + * for fast retrieval via a binary search in > + * i915_vma_instance(). They are also added to @vma.list for > + * easy iteration. > + */ > + struct rb_root tree; > + } vma; > + > + /** > + * @lut_list: List of vma lookup entries in use for this object. > + * > + * If this object is closed, we need to remove all of its VMA from > + * the fast lookup index in associated contexts; @lut_list provides > + * this translation from object to context->handles_vma. > + */ > + struct list_head lut_list; > + > + /** Stolen memory for this object, instead of being backed by shmem. */ > + struct drm_mm_node *stolen; > + union { > + struct rcu_head rcu; > + struct llist_node freed; > + }; > + > + /** > + * Whether the object is currently in the GGTT mmap. > + */ > + unsigned int userfault_count; > + struct list_head userfault_link; > + > + struct list_head batch_pool_link; > + I915_SELFTEST_DECLARE(struct list_head st_link); > + > + unsigned long flags; > + > + /** > + * Have we taken a reference for the object for incomplete GPU > + * activity? > + */ > +#define I915_BO_ACTIVE_REF 0 > + > + /* > + * Is the object to be mapped as read-only to the GPU > + * Only honoured if hardware has relevant pte bit > + */ > + unsigned int cache_level:3; > + unsigned int cache_coherent:2; > +#define I915_BO_CACHE_COHERENT_FOR_READ BIT(0) > +#define I915_BO_CACHE_COHERENT_FOR_WRITE BIT(1) > + unsigned int cache_dirty:1; > + > + /** > + * @read_domains: Read memory domains. > + * > + * These monitor which caches contain read/write data related to the > + * object. When transitioning from one set of domains to another, > + * the driver is called to ensure that caches are suitably flushed and > + * invalidated. > + */ > + u16 read_domains; > + > + /** > + * @write_domain: Corresponding unique write memory domain. > + */ > + u16 write_domain; > + > + atomic_t frontbuffer_bits; > + unsigned int frontbuffer_ggtt_origin; /* write once */ > + struct i915_active_request frontbuffer_write; > + > + /** Current tiling stride for the object, if it's tiled. */ > + unsigned int tiling_and_stride; > +#define FENCE_MINIMUM_STRIDE 128 /* See i915_tiling_ok() */ > +#define TILING_MASK (FENCE_MINIMUM_STRIDE - 1) > +#define STRIDE_MASK (~TILING_MASK) > + > + /** Count of VMA actually bound by this object */ > + unsigned int bind_count; > + unsigned int active_count; > + /** Count of how many global VMA are currently pinned for use by HW */ > + unsigned int pin_global; > + > + struct { > + struct mutex lock; /* protects the pages and their use */ > + atomic_t pages_pin_count; > + > + struct sg_table *pages; > + void *mapping; > + > + /* TODO: whack some of this into the error state */ > + struct i915_page_sizes { > + /** > + * The sg mask of the pages sg_table. i.e the mask of > + * of the lengths for each sg entry. > + */ > + unsigned int phys; > + > + /** > + * The gtt page sizes we are allowed to use given the > + * sg mask and the supported page sizes. This will > + * express the smallest unit we can use for the whole > + * object, as well as the larger sizes we may be able > + * to use opportunistically. > + */ > + unsigned int sg; > + > + /** > + * The actual gtt page size usage. Since we can have > + * multiple vma associated with this object we need to > + * prevent any trampling of state, hence a copy of this > + * struct also lives in each vma, therefore the gtt > + * value here should only be read/write through the vma. > + */ > + unsigned int gtt; > + } page_sizes; > + > + I915_SELFTEST_DECLARE(unsigned int page_mask); > + > + struct i915_gem_object_page_iter { > + struct scatterlist *sg_pos; > + unsigned int sg_idx; /* in pages, but 32bit eek! */ > + > + struct radix_tree_root radix; > + struct mutex lock; /* protects this cache */ > + } get_page; > + > + /** > + * Element within i915->mm.unbound_list or i915->mm.bound_list, > + * locked by i915->mm.obj_lock. > + */ > + struct list_head link; > + > + /** > + * Advice: are the backing pages purgeable? > + */ > + unsigned int madv:2; > + > + /** > + * This is set if the object has been written to since the > + * pages were last acquired. > + */ > + bool dirty:1; > + > + /** > + * This is set if the object has been pinned due to unknown > + * swizzling. > + */ > + bool quirked:1; > + } mm; > + > + /** Breadcrumb of last rendering to the buffer. > + * There can only be one writer, but we allow for multiple readers. > + * If there is a writer that necessarily implies that all other > + * read requests are complete - but we may only be lazily clearing > + * the read requests. A read request is naturally the most recent > + * request on a ring, so we may have two different write and read > + * requests on one ring where the write request is older than the > + * read request. This allows for the CPU to read from an active > + * buffer by only waiting for the write to complete. > + */ > + struct reservation_object *resv; > + > + /** References from framebuffers, locks out tiling changes. */ > + unsigned int framebuffer_references; > + > + /** Record of address bit 17 of each page at last unbind. */ > + unsigned long *bit_17; > + > + union { > + struct i915_gem_userptr { > + uintptr_t ptr; > + > + struct i915_mm_struct *mm; > + struct i915_mmu_object *mmu_object; > + struct work_struct *work; > + } userptr; > + > + unsigned long scratch; > + > + void *gvt_info; > + }; > + > + /** for phys allocated objects */ > + struct drm_dma_handle *phys_handle; > + > + struct reservation_object __builtin_resv; > +}; > + > +static inline struct drm_i915_gem_object * > +to_intel_bo(struct drm_gem_object *gem) > +{ > + /* Assert that to_intel_bo(NULL) == NULL */ > + BUILD_BUG_ON(offsetof(struct drm_i915_gem_object, base)); > + > + return container_of(gem, struct drm_i915_gem_object, base); > +} > + > +#endif > diff --git a/drivers/gpu/drm/i915/gt/intel_engine_types.h b/drivers/gpu/drm/i915/gt/intel_engine_types.h > index ca8d95a5708d..ae73c6596d08 100644 > --- a/drivers/gpu/drm/i915/gt/intel_engine_types.h > +++ b/drivers/gpu/drm/i915/gt/intel_engine_types.h > @@ -29,6 +29,7 @@ > #define I915_CMD_HASH_ORDER 9 > > struct dma_fence; > +struct drm_i915_gem_object; > struct drm_i915_reg_table; > struct i915_gem_context; > struct i915_request; > diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h > index 5ca4df9a7428..0cf87495e11b 100644 > --- a/drivers/gpu/drm/i915/i915_drv.h > +++ b/drivers/gpu/drm/i915/i915_drv.h > @@ -80,7 +80,6 @@ > #include "i915_gem.h" > #include "i915_gem_context.h" > #include "i915_gem_fence_reg.h" > -#include "i915_gem_object.h" > #include "i915_gem_gtt.h" > #include "i915_gpu_error.h" > #include "i915_request.h" > @@ -135,6 +134,8 @@ bool i915_error_injected(void); > __i915_printk(i915, i915_error_injected() ? KERN_DEBUG : KERN_ERR, \ > fmt, ##__VA_ARGS__) > > +struct drm_i915_gem_object; > + > enum hpd_pin { > HPD_NONE = 0, > HPD_TV = HPD_NONE, /* TV is known to be unreliable */ > diff --git a/drivers/gpu/drm/i915/i915_gem_batch_pool.h b/drivers/gpu/drm/i915/i915_gem_batch_pool.h > index 56947daaaf65..feeeeeaa54d8 100644 > --- a/drivers/gpu/drm/i915/i915_gem_batch_pool.h > +++ b/drivers/gpu/drm/i915/i915_gem_batch_pool.h > @@ -9,6 +9,7 @@ > > #include <linux/types.h> > > +struct drm_i915_gem_object; > struct intel_engine_cs; > > struct i915_gem_batch_pool { > @@ -19,7 +20,7 @@ struct i915_gem_batch_pool { > void i915_gem_batch_pool_init(struct i915_gem_batch_pool *pool, > struct intel_engine_cs *engine); > void i915_gem_batch_pool_fini(struct i915_gem_batch_pool *pool); > -struct drm_i915_gem_object* > +struct drm_i915_gem_object * > i915_gem_batch_pool_get(struct i915_gem_batch_pool *pool, size_t size); > > #endif /* I915_GEM_BATCH_POOL_H */ > diff --git a/drivers/gpu/drm/i915/i915_gem_gtt.h b/drivers/gpu/drm/i915/i915_gem_gtt.h > index 2fafa04c45ec..593079f30fbe 100644 > --- a/drivers/gpu/drm/i915/i915_gem_gtt.h > +++ b/drivers/gpu/drm/i915/i915_gem_gtt.h > @@ -61,6 +61,7 @@ > > struct drm_i915_file_private; > struct drm_i915_fence_reg; > +struct drm_i915_gem_object; > struct i915_vma; > > typedef u32 gen6_pte_t; > diff --git a/drivers/gpu/drm/i915/i915_gem_object.h b/drivers/gpu/drm/i915/i915_gem_object.h > index ca93a40c0c87..3666b0c5f6ca 100644 > --- a/drivers/gpu/drm/i915/i915_gem_object.h > +++ b/drivers/gpu/drm/i915/i915_gem_object.h > @@ -1,308 +1,19 @@ > /* > - * Copyright © 2016 Intel Corporation > - * > - * Permission is hereby granted, free of charge, to any person obtaining a > - * copy of this software and associated documentation files (the "Software"), > - * to deal in the Software without restriction, including without limitation > - * the rights to use, copy, modify, merge, publish, distribute, sublicense, > - * and/or sell copies of the Software, and to permit persons to whom the > - * Software is furnished to do so, subject to the following conditions: > - * > - * The above copyright notice and this permission notice (including the next > - * paragraph) shall be included in all copies or substantial portions of the > - * Software. > - * > - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR > - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, > - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL > - * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER > - * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING > - * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS > - * IN THE SOFTWARE. > + * SPDX-License-Identifier: MIT > * > + * Copyright © 2016 Intel Corporation > */ > > #ifndef __I915_GEM_OBJECT_H__ > #define __I915_GEM_OBJECT_H__ > > -#include <linux/reservation.h> > - > -#include <drm/drm_vma_manager.h> > #include <drm/drm_gem.h> > #include <drm/drm_file.h> > #include <drm/drm_device.h> > > #include <drm/i915_drm.h> > > -#include "i915_request.h" > -#include "i915_selftest.h" > - > -struct drm_i915_gem_object; > - > -/* > - * struct i915_lut_handle tracks the fast lookups from handle to vma used > - * for execbuf. Although we use a radixtree for that mapping, in order to > - * remove them as the object or context is closed, we need a secondary list > - * and a translation entry (i915_lut_handle). > - */ > -struct i915_lut_handle { > - struct list_head obj_link; > - struct list_head ctx_link; > - struct i915_gem_context *ctx; > - u32 handle; > -}; > - > -struct drm_i915_gem_object_ops { > - unsigned int flags; > -#define I915_GEM_OBJECT_HAS_STRUCT_PAGE BIT(0) > -#define I915_GEM_OBJECT_IS_SHRINKABLE BIT(1) > -#define I915_GEM_OBJECT_IS_PROXY BIT(2) > -#define I915_GEM_OBJECT_ASYNC_CANCEL BIT(3) > - > - /* Interface between the GEM object and its backing storage. > - * get_pages() is called once prior to the use of the associated set > - * of pages before to binding them into the GTT, and put_pages() is > - * called after we no longer need them. As we expect there to be > - * associated cost with migrating pages between the backing storage > - * and making them available for the GPU (e.g. clflush), we may hold > - * onto the pages after they are no longer referenced by the GPU > - * in case they may be used again shortly (for example migrating the > - * pages to a different memory domain within the GTT). put_pages() > - * will therefore most likely be called when the object itself is > - * being released or under memory pressure (where we attempt to > - * reap pages for the shrinker). > - */ > - int (*get_pages)(struct drm_i915_gem_object *); > - void (*put_pages)(struct drm_i915_gem_object *, struct sg_table *); > - > - int (*pwrite)(struct drm_i915_gem_object *, > - const struct drm_i915_gem_pwrite *); > - > - int (*dmabuf_export)(struct drm_i915_gem_object *); > - void (*release)(struct drm_i915_gem_object *); > -}; > - > -struct drm_i915_gem_object { > - struct drm_gem_object base; > - > - const struct drm_i915_gem_object_ops *ops; > - > - struct { > - /** > - * @vma.lock: protect the list/tree of vmas > - */ > - spinlock_t lock; > - > - /** > - * @vma.list: List of VMAs backed by this object > - * > - * The VMA on this list are ordered by type, all GGTT vma are > - * placed at the head and all ppGTT vma are placed at the tail. > - * The different types of GGTT vma are unordered between > - * themselves, use the @vma.tree (which has a defined order > - * between all VMA) to quickly find an exact match. > - */ > - struct list_head list; > - > - /** > - * @vma.tree: Ordered tree of VMAs backed by this object > - * > - * All VMA created for this object are placed in the @vma.tree > - * for fast retrieval via a binary search in > - * i915_vma_instance(). They are also added to @vma.list for > - * easy iteration. > - */ > - struct rb_root tree; > - } vma; > - > - /** > - * @lut_list: List of vma lookup entries in use for this object. > - * > - * If this object is closed, we need to remove all of its VMA from > - * the fast lookup index in associated contexts; @lut_list provides > - * this translation from object to context->handles_vma. > - */ > - struct list_head lut_list; > - > - /** Stolen memory for this object, instead of being backed by shmem. */ > - struct drm_mm_node *stolen; > - union { > - struct rcu_head rcu; > - struct llist_node freed; > - }; > - > - /** > - * Whether the object is currently in the GGTT mmap. > - */ > - unsigned int userfault_count; > - struct list_head userfault_link; > - > - struct list_head batch_pool_link; > - I915_SELFTEST_DECLARE(struct list_head st_link); > - > - unsigned long flags; > - > - /** > - * Have we taken a reference for the object for incomplete GPU > - * activity? > - */ > -#define I915_BO_ACTIVE_REF 0 > - > - /* > - * Is the object to be mapped as read-only to the GPU > - * Only honoured if hardware has relevant pte bit > - */ > - unsigned int cache_level:3; > - unsigned int cache_coherent:2; > -#define I915_BO_CACHE_COHERENT_FOR_READ BIT(0) > -#define I915_BO_CACHE_COHERENT_FOR_WRITE BIT(1) > - unsigned int cache_dirty:1; > - > - /** > - * @read_domains: Read memory domains. > - * > - * These monitor which caches contain read/write data related to the > - * object. When transitioning from one set of domains to another, > - * the driver is called to ensure that caches are suitably flushed and > - * invalidated. > - */ > - u16 read_domains; > - > - /** > - * @write_domain: Corresponding unique write memory domain. > - */ > - u16 write_domain; > - > - atomic_t frontbuffer_bits; > - unsigned int frontbuffer_ggtt_origin; /* write once */ > - struct i915_active_request frontbuffer_write; > - > - /** Current tiling stride for the object, if it's tiled. */ > - unsigned int tiling_and_stride; > -#define FENCE_MINIMUM_STRIDE 128 /* See i915_tiling_ok() */ > -#define TILING_MASK (FENCE_MINIMUM_STRIDE-1) > -#define STRIDE_MASK (~TILING_MASK) > - > - /** Count of VMA actually bound by this object */ > - unsigned int bind_count; > - unsigned int active_count; > - /** Count of how many global VMA are currently pinned for use by HW */ > - unsigned int pin_global; > - > - struct { > - struct mutex lock; /* protects the pages and their use */ > - atomic_t pages_pin_count; > - > - struct sg_table *pages; > - void *mapping; > - > - /* TODO: whack some of this into the error state */ > - struct i915_page_sizes { > - /** > - * The sg mask of the pages sg_table. i.e the mask of > - * of the lengths for each sg entry. > - */ > - unsigned int phys; > - > - /** > - * The gtt page sizes we are allowed to use given the > - * sg mask and the supported page sizes. This will > - * express the smallest unit we can use for the whole > - * object, as well as the larger sizes we may be able > - * to use opportunistically. > - */ > - unsigned int sg; > - > - /** > - * The actual gtt page size usage. Since we can have > - * multiple vma associated with this object we need to > - * prevent any trampling of state, hence a copy of this > - * struct also lives in each vma, therefore the gtt > - * value here should only be read/write through the vma. > - */ > - unsigned int gtt; > - } page_sizes; > - > - I915_SELFTEST_DECLARE(unsigned int page_mask); > - > - struct i915_gem_object_page_iter { > - struct scatterlist *sg_pos; > - unsigned int sg_idx; /* in pages, but 32bit eek! */ > - > - struct radix_tree_root radix; > - struct mutex lock; /* protects this cache */ > - } get_page; > - > - /** > - * Element within i915->mm.unbound_list or i915->mm.bound_list, > - * locked by i915->mm.obj_lock. > - */ > - struct list_head link; > - > - /** > - * Advice: are the backing pages purgeable? > - */ > - unsigned int madv:2; > - > - /** > - * This is set if the object has been written to since the > - * pages were last acquired. > - */ > - bool dirty:1; > - > - /** > - * This is set if the object has been pinned due to unknown > - * swizzling. > - */ > - bool quirked:1; > - } mm; > - > - /** Breadcrumb of last rendering to the buffer. > - * There can only be one writer, but we allow for multiple readers. > - * If there is a writer that necessarily implies that all other > - * read requests are complete - but we may only be lazily clearing > - * the read requests. A read request is naturally the most recent > - * request on a ring, so we may have two different write and read > - * requests on one ring where the write request is older than the > - * read request. This allows for the CPU to read from an active > - * buffer by only waiting for the write to complete. > - */ > - struct reservation_object *resv; > - > - /** References from framebuffers, locks out tiling changes. */ > - unsigned int framebuffer_references; > - > - /** Record of address bit 17 of each page at last unbind. */ > - unsigned long *bit_17; > - > - union { > - struct i915_gem_userptr { > - uintptr_t ptr; > - > - struct i915_mm_struct *mm; > - struct i915_mmu_object *mmu_object; > - struct work_struct *work; > - } userptr; > - > - unsigned long scratch; > - > - void *gvt_info; > - }; > - > - /** for phys allocated objects */ > - struct drm_dma_handle *phys_handle; > - > - struct reservation_object __builtin_resv; > -}; > - > -static inline struct drm_i915_gem_object * > -to_intel_bo(struct drm_gem_object *gem) > -{ > - /* Assert that to_intel_bo(NULL) == NULL */ > - BUILD_BUG_ON(offsetof(struct drm_i915_gem_object, base)); > - > - return container_of(gem, struct drm_i915_gem_object, base); > -} > +#include "gem/i915_gem_object_types.h" > > struct drm_i915_gem_object *i915_gem_object_alloc(void); > void i915_gem_object_free(struct drm_i915_gem_object *obj);
Quoting Jani Nikula (2019-04-26 15:12:49) > On Thu, 25 Apr 2019, Chris Wilson <chris@chris-wilson.co.uk> wrote: > > For convenience in avoiding inline spaghetti, keep the type definition > > as a separate header. > > > > Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> > > Reviewed-by: Matthew Auld <matthew.auld@intel.com> > > --- > > drivers/gpu/drm/i915/Makefile | 1 + > > drivers/gpu/drm/i915/gem/Makefile | 1 + > > drivers/gpu/drm/i915/gem/Makefile.header-test | 16 + > > .../gpu/drm/i915/gem/i915_gem_object_types.h | 285 +++++++++++++++++ > > drivers/gpu/drm/i915/gt/intel_engine_types.h | 1 + > > drivers/gpu/drm/i915/i915_drv.h | 3 +- > > drivers/gpu/drm/i915/i915_gem_batch_pool.h | 3 +- > > drivers/gpu/drm/i915/i915_gem_gtt.h | 1 + > > drivers/gpu/drm/i915/i915_gem_object.h | 295 +----------------- > > 9 files changed, 312 insertions(+), 294 deletions(-) > > create mode 100644 drivers/gpu/drm/i915/gem/Makefile > > create mode 100644 drivers/gpu/drm/i915/gem/Makefile.header-test > > create mode 100644 drivers/gpu/drm/i915/gem/i915_gem_object_types.h > > > > diff --git a/drivers/gpu/drm/i915/Makefile b/drivers/gpu/drm/i915/Makefile > > index 58643373495c..def781c9ea69 100644 > > --- a/drivers/gpu/drm/i915/Makefile > > +++ b/drivers/gpu/drm/i915/Makefile > > @@ -85,6 +85,7 @@ gt-$(CONFIG_DRM_I915_SELFTEST) += \ > > i915-y += $(gt-y) > > > > # GEM (Graphics Execution Management) code > > +obj-y += gem/ > > I think this warrants similar treatment as gt/. Posted standalone > instead of hidden in a series, explicit acks. > > Provided the split makes sense to Joonas, > > Acked-by: Jani Nikula <jani.nikula@intel.com> Yeah, makes sense to me. Acked-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com> Regards, Joonas > > > > i915-y += \ > > i915_active.o \ > > i915_cmd_parser.o \ > > diff --git a/drivers/gpu/drm/i915/gem/Makefile b/drivers/gpu/drm/i915/gem/Makefile > > new file mode 100644 > > index 000000000000..07e7b8b840ea > > --- /dev/null > > +++ b/drivers/gpu/drm/i915/gem/Makefile > > @@ -0,0 +1 @@ > > +include $(src)/Makefile.header-test # Extra header tests > > diff --git a/drivers/gpu/drm/i915/gem/Makefile.header-test b/drivers/gpu/drm/i915/gem/Makefile.header-test > > new file mode 100644 > > index 000000000000..61e06cbb4b32 > > --- /dev/null > > +++ b/drivers/gpu/drm/i915/gem/Makefile.header-test > > @@ -0,0 +1,16 @@ > > +# SPDX-License-Identifier: MIT > > +# Copyright © 2019 Intel Corporation > > + > > +# Test the headers are compilable as standalone units > > +header_test := $(notdir $(wildcard $(src)/*.h)) > > + > > +quiet_cmd_header_test = HDRTEST $@ > > + cmd_header_test = echo "\#include \"$(<F)\"" > $@ > > + > > +header_test_%.c: %.h > > + $(call cmd,header_test) > > + > > +extra-$(CONFIG_DRM_I915_WERROR) += \ > > + $(foreach h,$(header_test),$(patsubst %.h,header_test_%.o,$(h))) > > + > > +clean-files += $(foreach h,$(header_test),$(patsubst %.h,header_test_%.c,$(h))) > > diff --git a/drivers/gpu/drm/i915/gem/i915_gem_object_types.h b/drivers/gpu/drm/i915/gem/i915_gem_object_types.h > > new file mode 100644 > > index 000000000000..e4b50944f553 > > --- /dev/null > > +++ b/drivers/gpu/drm/i915/gem/i915_gem_object_types.h > > @@ -0,0 +1,285 @@ > > +/* > > + * SPDX-License-Identifier: MIT > > + * > > + * Copyright © 2016 Intel Corporation > > + */ > > + > > +#ifndef __I915_GEM_OBJECT_TYPES_H__ > > +#define __I915_GEM_OBJECT_TYPES_H__ > > + > > +#include <linux/reservation.h> > > + > > +#include <drm/drm_gem.h> > > + > > +#include "../i915_active.h" > > +#include "../i915_selftest.h" > > + > > +struct drm_i915_gem_object; > > + > > +/* > > + * struct i915_lut_handle tracks the fast lookups from handle to vma used > > + * for execbuf. Although we use a radixtree for that mapping, in order to > > + * remove them as the object or context is closed, we need a secondary list > > + * and a translation entry (i915_lut_handle). > > + */ > > +struct i915_lut_handle { > > + struct list_head obj_link; > > + struct list_head ctx_link; > > + struct i915_gem_context *ctx; > > + u32 handle; > > +}; > > + > > +struct drm_i915_gem_object_ops { > > + unsigned int flags; > > +#define I915_GEM_OBJECT_HAS_STRUCT_PAGE BIT(0) > > +#define I915_GEM_OBJECT_IS_SHRINKABLE BIT(1) > > +#define I915_GEM_OBJECT_IS_PROXY BIT(2) > > +#define I915_GEM_OBJECT_ASYNC_CANCEL BIT(3) > > + > > + /* Interface between the GEM object and its backing storage. > > + * get_pages() is called once prior to the use of the associated set > > + * of pages before to binding them into the GTT, and put_pages() is > > + * called after we no longer need them. As we expect there to be > > + * associated cost with migrating pages between the backing storage > > + * and making them available for the GPU (e.g. clflush), we may hold > > + * onto the pages after they are no longer referenced by the GPU > > + * in case they may be used again shortly (for example migrating the > > + * pages to a different memory domain within the GTT). put_pages() > > + * will therefore most likely be called when the object itself is > > + * being released or under memory pressure (where we attempt to > > + * reap pages for the shrinker). > > + */ > > + int (*get_pages)(struct drm_i915_gem_object *obj); > > + void (*put_pages)(struct drm_i915_gem_object *obj, > > + struct sg_table *pages); > > + > > + int (*pwrite)(struct drm_i915_gem_object *obj, > > + const struct drm_i915_gem_pwrite *arg); > > + > > + int (*dmabuf_export)(struct drm_i915_gem_object *obj); > > + void (*release)(struct drm_i915_gem_object *obj); > > +}; > > + > > +struct drm_i915_gem_object { > > + struct drm_gem_object base; > > + > > + const struct drm_i915_gem_object_ops *ops; > > + > > + struct { > > + /** > > + * @vma.lock: protect the list/tree of vmas > > + */ > > + spinlock_t lock; > > + > > + /** > > + * @vma.list: List of VMAs backed by this object > > + * > > + * The VMA on this list are ordered by type, all GGTT vma are > > + * placed at the head and all ppGTT vma are placed at the tail. > > + * The different types of GGTT vma are unordered between > > + * themselves, use the @vma.tree (which has a defined order > > + * between all VMA) to quickly find an exact match. > > + */ > > + struct list_head list; > > + > > + /** > > + * @vma.tree: Ordered tree of VMAs backed by this object > > + * > > + * All VMA created for this object are placed in the @vma.tree > > + * for fast retrieval via a binary search in > > + * i915_vma_instance(). They are also added to @vma.list for > > + * easy iteration. > > + */ > > + struct rb_root tree; > > + } vma; > > + > > + /** > > + * @lut_list: List of vma lookup entries in use for this object. > > + * > > + * If this object is closed, we need to remove all of its VMA from > > + * the fast lookup index in associated contexts; @lut_list provides > > + * this translation from object to context->handles_vma. > > + */ > > + struct list_head lut_list; > > + > > + /** Stolen memory for this object, instead of being backed by shmem. */ > > + struct drm_mm_node *stolen; > > + union { > > + struct rcu_head rcu; > > + struct llist_node freed; > > + }; > > + > > + /** > > + * Whether the object is currently in the GGTT mmap. > > + */ > > + unsigned int userfault_count; > > + struct list_head userfault_link; > > + > > + struct list_head batch_pool_link; > > + I915_SELFTEST_DECLARE(struct list_head st_link); > > + > > + unsigned long flags; > > + > > + /** > > + * Have we taken a reference for the object for incomplete GPU > > + * activity? > > + */ > > +#define I915_BO_ACTIVE_REF 0 > > + > > + /* > > + * Is the object to be mapped as read-only to the GPU > > + * Only honoured if hardware has relevant pte bit > > + */ > > + unsigned int cache_level:3; > > + unsigned int cache_coherent:2; > > +#define I915_BO_CACHE_COHERENT_FOR_READ BIT(0) > > +#define I915_BO_CACHE_COHERENT_FOR_WRITE BIT(1) > > + unsigned int cache_dirty:1; > > + > > + /** > > + * @read_domains: Read memory domains. > > + * > > + * These monitor which caches contain read/write data related to the > > + * object. When transitioning from one set of domains to another, > > + * the driver is called to ensure that caches are suitably flushed and > > + * invalidated. > > + */ > > + u16 read_domains; > > + > > + /** > > + * @write_domain: Corresponding unique write memory domain. > > + */ > > + u16 write_domain; > > + > > + atomic_t frontbuffer_bits; > > + unsigned int frontbuffer_ggtt_origin; /* write once */ > > + struct i915_active_request frontbuffer_write; > > + > > + /** Current tiling stride for the object, if it's tiled. */ > > + unsigned int tiling_and_stride; > > +#define FENCE_MINIMUM_STRIDE 128 /* See i915_tiling_ok() */ > > +#define TILING_MASK (FENCE_MINIMUM_STRIDE - 1) > > +#define STRIDE_MASK (~TILING_MASK) > > + > > + /** Count of VMA actually bound by this object */ > > + unsigned int bind_count; > > + unsigned int active_count; > > + /** Count of how many global VMA are currently pinned for use by HW */ > > + unsigned int pin_global; > > + > > + struct { > > + struct mutex lock; /* protects the pages and their use */ > > + atomic_t pages_pin_count; > > + > > + struct sg_table *pages; > > + void *mapping; > > + > > + /* TODO: whack some of this into the error state */ > > + struct i915_page_sizes { > > + /** > > + * The sg mask of the pages sg_table. i.e the mask of > > + * of the lengths for each sg entry. > > + */ > > + unsigned int phys; > > + > > + /** > > + * The gtt page sizes we are allowed to use given the > > + * sg mask and the supported page sizes. This will > > + * express the smallest unit we can use for the whole > > + * object, as well as the larger sizes we may be able > > + * to use opportunistically. > > + */ > > + unsigned int sg; > > + > > + /** > > + * The actual gtt page size usage. Since we can have > > + * multiple vma associated with this object we need to > > + * prevent any trampling of state, hence a copy of this > > + * struct also lives in each vma, therefore the gtt > > + * value here should only be read/write through the vma. > > + */ > > + unsigned int gtt; > > + } page_sizes; > > + > > + I915_SELFTEST_DECLARE(unsigned int page_mask); > > + > > + struct i915_gem_object_page_iter { > > + struct scatterlist *sg_pos; > > + unsigned int sg_idx; /* in pages, but 32bit eek! */ > > + > > + struct radix_tree_root radix; > > + struct mutex lock; /* protects this cache */ > > + } get_page; > > + > > + /** > > + * Element within i915->mm.unbound_list or i915->mm.bound_list, > > + * locked by i915->mm.obj_lock. > > + */ > > + struct list_head link; > > + > > + /** > > + * Advice: are the backing pages purgeable? > > + */ > > + unsigned int madv:2; > > + > > + /** > > + * This is set if the object has been written to since the > > + * pages were last acquired. > > + */ > > + bool dirty:1; > > + > > + /** > > + * This is set if the object has been pinned due to unknown > > + * swizzling. > > + */ > > + bool quirked:1; > > + } mm; > > + > > + /** Breadcrumb of last rendering to the buffer. > > + * There can only be one writer, but we allow for multiple readers. > > + * If there is a writer that necessarily implies that all other > > + * read requests are complete - but we may only be lazily clearing > > + * the read requests. A read request is naturally the most recent > > + * request on a ring, so we may have two different write and read > > + * requests on one ring where the write request is older than the > > + * read request. This allows for the CPU to read from an active > > + * buffer by only waiting for the write to complete. > > + */ > > + struct reservation_object *resv; > > + > > + /** References from framebuffers, locks out tiling changes. */ > > + unsigned int framebuffer_references; > > + > > + /** Record of address bit 17 of each page at last unbind. */ > > + unsigned long *bit_17; > > + > > + union { > > + struct i915_gem_userptr { > > + uintptr_t ptr; > > + > > + struct i915_mm_struct *mm; > > + struct i915_mmu_object *mmu_object; > > + struct work_struct *work; > > + } userptr; > > + > > + unsigned long scratch; > > + > > + void *gvt_info; > > + }; > > + > > + /** for phys allocated objects */ > > + struct drm_dma_handle *phys_handle; > > + > > + struct reservation_object __builtin_resv; > > +}; > > + > > +static inline struct drm_i915_gem_object * > > +to_intel_bo(struct drm_gem_object *gem) > > +{ > > + /* Assert that to_intel_bo(NULL) == NULL */ > > + BUILD_BUG_ON(offsetof(struct drm_i915_gem_object, base)); > > + > > + return container_of(gem, struct drm_i915_gem_object, base); > > +} > > + > > +#endif > > diff --git a/drivers/gpu/drm/i915/gt/intel_engine_types.h b/drivers/gpu/drm/i915/gt/intel_engine_types.h > > index ca8d95a5708d..ae73c6596d08 100644 > > --- a/drivers/gpu/drm/i915/gt/intel_engine_types.h > > +++ b/drivers/gpu/drm/i915/gt/intel_engine_types.h > > @@ -29,6 +29,7 @@ > > #define I915_CMD_HASH_ORDER 9 > > > > struct dma_fence; > > +struct drm_i915_gem_object; > > struct drm_i915_reg_table; > > struct i915_gem_context; > > struct i915_request; > > diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h > > index 5ca4df9a7428..0cf87495e11b 100644 > > --- a/drivers/gpu/drm/i915/i915_drv.h > > +++ b/drivers/gpu/drm/i915/i915_drv.h > > @@ -80,7 +80,6 @@ > > #include "i915_gem.h" > > #include "i915_gem_context.h" > > #include "i915_gem_fence_reg.h" > > -#include "i915_gem_object.h" > > #include "i915_gem_gtt.h" > > #include "i915_gpu_error.h" > > #include "i915_request.h" > > @@ -135,6 +134,8 @@ bool i915_error_injected(void); > > __i915_printk(i915, i915_error_injected() ? KERN_DEBUG : KERN_ERR, \ > > fmt, ##__VA_ARGS__) > > > > +struct drm_i915_gem_object; > > + > > enum hpd_pin { > > HPD_NONE = 0, > > HPD_TV = HPD_NONE, /* TV is known to be unreliable */ > > diff --git a/drivers/gpu/drm/i915/i915_gem_batch_pool.h b/drivers/gpu/drm/i915/i915_gem_batch_pool.h > > index 56947daaaf65..feeeeeaa54d8 100644 > > --- a/drivers/gpu/drm/i915/i915_gem_batch_pool.h > > +++ b/drivers/gpu/drm/i915/i915_gem_batch_pool.h > > @@ -9,6 +9,7 @@ > > > > #include <linux/types.h> > > > > +struct drm_i915_gem_object; > > struct intel_engine_cs; > > > > struct i915_gem_batch_pool { > > @@ -19,7 +20,7 @@ struct i915_gem_batch_pool { > > void i915_gem_batch_pool_init(struct i915_gem_batch_pool *pool, > > struct intel_engine_cs *engine); > > void i915_gem_batch_pool_fini(struct i915_gem_batch_pool *pool); > > -struct drm_i915_gem_object* > > +struct drm_i915_gem_object * > > i915_gem_batch_pool_get(struct i915_gem_batch_pool *pool, size_t size); > > > > #endif /* I915_GEM_BATCH_POOL_H */ > > diff --git a/drivers/gpu/drm/i915/i915_gem_gtt.h b/drivers/gpu/drm/i915/i915_gem_gtt.h > > index 2fafa04c45ec..593079f30fbe 100644 > > --- a/drivers/gpu/drm/i915/i915_gem_gtt.h > > +++ b/drivers/gpu/drm/i915/i915_gem_gtt.h > > @@ -61,6 +61,7 @@ > > > > struct drm_i915_file_private; > > struct drm_i915_fence_reg; > > +struct drm_i915_gem_object; > > struct i915_vma; > > > > typedef u32 gen6_pte_t; > > diff --git a/drivers/gpu/drm/i915/i915_gem_object.h b/drivers/gpu/drm/i915/i915_gem_object.h > > index ca93a40c0c87..3666b0c5f6ca 100644 > > --- a/drivers/gpu/drm/i915/i915_gem_object.h > > +++ b/drivers/gpu/drm/i915/i915_gem_object.h > > @@ -1,308 +1,19 @@ > > /* > > - * Copyright © 2016 Intel Corporation > > - * > > - * Permission is hereby granted, free of charge, to any person obtaining a > > - * copy of this software and associated documentation files (the "Software"), > > - * to deal in the Software without restriction, including without limitation > > - * the rights to use, copy, modify, merge, publish, distribute, sublicense, > > - * and/or sell copies of the Software, and to permit persons to whom the > > - * Software is furnished to do so, subject to the following conditions: > > - * > > - * The above copyright notice and this permission notice (including the next > > - * paragraph) shall be included in all copies or substantial portions of the > > - * Software. > > - * > > - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR > > - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, > > - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL > > - * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER > > - * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING > > - * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS > > - * IN THE SOFTWARE. > > + * SPDX-License-Identifier: MIT > > * > > + * Copyright © 2016 Intel Corporation > > */ > > > > #ifndef __I915_GEM_OBJECT_H__ > > #define __I915_GEM_OBJECT_H__ > > > > -#include <linux/reservation.h> > > - > > -#include <drm/drm_vma_manager.h> > > #include <drm/drm_gem.h> > > #include <drm/drm_file.h> > > #include <drm/drm_device.h> > > > > #include <drm/i915_drm.h> > > > > -#include "i915_request.h" > > -#include "i915_selftest.h" > > - > > -struct drm_i915_gem_object; > > - > > -/* > > - * struct i915_lut_handle tracks the fast lookups from handle to vma used > > - * for execbuf. Although we use a radixtree for that mapping, in order to > > - * remove them as the object or context is closed, we need a secondary list > > - * and a translation entry (i915_lut_handle). > > - */ > > -struct i915_lut_handle { > > - struct list_head obj_link; > > - struct list_head ctx_link; > > - struct i915_gem_context *ctx; > > - u32 handle; > > -}; > > - > > -struct drm_i915_gem_object_ops { > > - unsigned int flags; > > -#define I915_GEM_OBJECT_HAS_STRUCT_PAGE BIT(0) > > -#define I915_GEM_OBJECT_IS_SHRINKABLE BIT(1) > > -#define I915_GEM_OBJECT_IS_PROXY BIT(2) > > -#define I915_GEM_OBJECT_ASYNC_CANCEL BIT(3) > > - > > - /* Interface between the GEM object and its backing storage. > > - * get_pages() is called once prior to the use of the associated set > > - * of pages before to binding them into the GTT, and put_pages() is > > - * called after we no longer need them. As we expect there to be > > - * associated cost with migrating pages between the backing storage > > - * and making them available for the GPU (e.g. clflush), we may hold > > - * onto the pages after they are no longer referenced by the GPU > > - * in case they may be used again shortly (for example migrating the > > - * pages to a different memory domain within the GTT). put_pages() > > - * will therefore most likely be called when the object itself is > > - * being released or under memory pressure (where we attempt to > > - * reap pages for the shrinker). > > - */ > > - int (*get_pages)(struct drm_i915_gem_object *); > > - void (*put_pages)(struct drm_i915_gem_object *, struct sg_table *); > > - > > - int (*pwrite)(struct drm_i915_gem_object *, > > - const struct drm_i915_gem_pwrite *); > > - > > - int (*dmabuf_export)(struct drm_i915_gem_object *); > > - void (*release)(struct drm_i915_gem_object *); > > -}; > > - > > -struct drm_i915_gem_object { > > - struct drm_gem_object base; > > - > > - const struct drm_i915_gem_object_ops *ops; > > - > > - struct { > > - /** > > - * @vma.lock: protect the list/tree of vmas > > - */ > > - spinlock_t lock; > > - > > - /** > > - * @vma.list: List of VMAs backed by this object > > - * > > - * The VMA on this list are ordered by type, all GGTT vma are > > - * placed at the head and all ppGTT vma are placed at the tail. > > - * The different types of GGTT vma are unordered between > > - * themselves, use the @vma.tree (which has a defined order > > - * between all VMA) to quickly find an exact match. > > - */ > > - struct list_head list; > > - > > - /** > > - * @vma.tree: Ordered tree of VMAs backed by this object > > - * > > - * All VMA created for this object are placed in the @vma.tree > > - * for fast retrieval via a binary search in > > - * i915_vma_instance(). They are also added to @vma.list for > > - * easy iteration. > > - */ > > - struct rb_root tree; > > - } vma; > > - > > - /** > > - * @lut_list: List of vma lookup entries in use for this object. > > - * > > - * If this object is closed, we need to remove all of its VMA from > > - * the fast lookup index in associated contexts; @lut_list provides > > - * this translation from object to context->handles_vma. > > - */ > > - struct list_head lut_list; > > - > > - /** Stolen memory for this object, instead of being backed by shmem. */ > > - struct drm_mm_node *stolen; > > - union { > > - struct rcu_head rcu; > > - struct llist_node freed; > > - }; > > - > > - /** > > - * Whether the object is currently in the GGTT mmap. > > - */ > > - unsigned int userfault_count; > > - struct list_head userfault_link; > > - > > - struct list_head batch_pool_link; > > - I915_SELFTEST_DECLARE(struct list_head st_link); > > - > > - unsigned long flags; > > - > > - /** > > - * Have we taken a reference for the object for incomplete GPU > > - * activity? > > - */ > > -#define I915_BO_ACTIVE_REF 0 > > - > > - /* > > - * Is the object to be mapped as read-only to the GPU > > - * Only honoured if hardware has relevant pte bit > > - */ > > - unsigned int cache_level:3; > > - unsigned int cache_coherent:2; > > -#define I915_BO_CACHE_COHERENT_FOR_READ BIT(0) > > -#define I915_BO_CACHE_COHERENT_FOR_WRITE BIT(1) > > - unsigned int cache_dirty:1; > > - > > - /** > > - * @read_domains: Read memory domains. > > - * > > - * These monitor which caches contain read/write data related to the > > - * object. When transitioning from one set of domains to another, > > - * the driver is called to ensure that caches are suitably flushed and > > - * invalidated. > > - */ > > - u16 read_domains; > > - > > - /** > > - * @write_domain: Corresponding unique write memory domain. > > - */ > > - u16 write_domain; > > - > > - atomic_t frontbuffer_bits; > > - unsigned int frontbuffer_ggtt_origin; /* write once */ > > - struct i915_active_request frontbuffer_write; > > - > > - /** Current tiling stride for the object, if it's tiled. */ > > - unsigned int tiling_and_stride; > > -#define FENCE_MINIMUM_STRIDE 128 /* See i915_tiling_ok() */ > > -#define TILING_MASK (FENCE_MINIMUM_STRIDE-1) > > -#define STRIDE_MASK (~TILING_MASK) > > - > > - /** Count of VMA actually bound by this object */ > > - unsigned int bind_count; > > - unsigned int active_count; > > - /** Count of how many global VMA are currently pinned for use by HW */ > > - unsigned int pin_global; > > - > > - struct { > > - struct mutex lock; /* protects the pages and their use */ > > - atomic_t pages_pin_count; > > - > > - struct sg_table *pages; > > - void *mapping; > > - > > - /* TODO: whack some of this into the error state */ > > - struct i915_page_sizes { > > - /** > > - * The sg mask of the pages sg_table. i.e the mask of > > - * of the lengths for each sg entry. > > - */ > > - unsigned int phys; > > - > > - /** > > - * The gtt page sizes we are allowed to use given the > > - * sg mask and the supported page sizes. This will > > - * express the smallest unit we can use for the whole > > - * object, as well as the larger sizes we may be able > > - * to use opportunistically. > > - */ > > - unsigned int sg; > > - > > - /** > > - * The actual gtt page size usage. Since we can have > > - * multiple vma associated with this object we need to > > - * prevent any trampling of state, hence a copy of this > > - * struct also lives in each vma, therefore the gtt > > - * value here should only be read/write through the vma. > > - */ > > - unsigned int gtt; > > - } page_sizes; > > - > > - I915_SELFTEST_DECLARE(unsigned int page_mask); > > - > > - struct i915_gem_object_page_iter { > > - struct scatterlist *sg_pos; > > - unsigned int sg_idx; /* in pages, but 32bit eek! */ > > - > > - struct radix_tree_root radix; > > - struct mutex lock; /* protects this cache */ > > - } get_page; > > - > > - /** > > - * Element within i915->mm.unbound_list or i915->mm.bound_list, > > - * locked by i915->mm.obj_lock. > > - */ > > - struct list_head link; > > - > > - /** > > - * Advice: are the backing pages purgeable? > > - */ > > - unsigned int madv:2; > > - > > - /** > > - * This is set if the object has been written to since the > > - * pages were last acquired. > > - */ > > - bool dirty:1; > > - > > - /** > > - * This is set if the object has been pinned due to unknown > > - * swizzling. > > - */ > > - bool quirked:1; > > - } mm; > > - > > - /** Breadcrumb of last rendering to the buffer. > > - * There can only be one writer, but we allow for multiple readers. > > - * If there is a writer that necessarily implies that all other > > - * read requests are complete - but we may only be lazily clearing > > - * the read requests. A read request is naturally the most recent > > - * request on a ring, so we may have two different write and read > > - * requests on one ring where the write request is older than the > > - * read request. This allows for the CPU to read from an active > > - * buffer by only waiting for the write to complete. > > - */ > > - struct reservation_object *resv; > > - > > - /** References from framebuffers, locks out tiling changes. */ > > - unsigned int framebuffer_references; > > - > > - /** Record of address bit 17 of each page at last unbind. */ > > - unsigned long *bit_17; > > - > > - union { > > - struct i915_gem_userptr { > > - uintptr_t ptr; > > - > > - struct i915_mm_struct *mm; > > - struct i915_mmu_object *mmu_object; > > - struct work_struct *work; > > - } userptr; > > - > > - unsigned long scratch; > > - > > - void *gvt_info; > > - }; > > - > > - /** for phys allocated objects */ > > - struct drm_dma_handle *phys_handle; > > - > > - struct reservation_object __builtin_resv; > > -}; > > - > > -static inline struct drm_i915_gem_object * > > -to_intel_bo(struct drm_gem_object *gem) > > -{ > > - /* Assert that to_intel_bo(NULL) == NULL */ > > - BUILD_BUG_ON(offsetof(struct drm_i915_gem_object, base)); > > - > > - return container_of(gem, struct drm_i915_gem_object, base); > > -} > > +#include "gem/i915_gem_object_types.h" > > > > struct drm_i915_gem_object *i915_gem_object_alloc(void); > > void i915_gem_object_free(struct drm_i915_gem_object *obj); > > -- > Jani Nikula, Intel Open Source Graphics Center
On Mon, Apr 29, 2019 at 12:36:30PM +0300, Joonas Lahtinen wrote: > Quoting Jani Nikula (2019-04-26 15:12:49) > > On Thu, 25 Apr 2019, Chris Wilson <chris@chris-wilson.co.uk> wrote: > > > For convenience in avoiding inline spaghetti, keep the type definition > > > as a separate header. > > > > > > Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> > > > Reviewed-by: Matthew Auld <matthew.auld@intel.com> > > > --- > > > drivers/gpu/drm/i915/Makefile | 1 + > > > drivers/gpu/drm/i915/gem/Makefile | 1 + > > > drivers/gpu/drm/i915/gem/Makefile.header-test | 16 + > > > .../gpu/drm/i915/gem/i915_gem_object_types.h | 285 +++++++++++++++++ > > > drivers/gpu/drm/i915/gt/intel_engine_types.h | 1 + > > > drivers/gpu/drm/i915/i915_drv.h | 3 +- > > > drivers/gpu/drm/i915/i915_gem_batch_pool.h | 3 +- > > > drivers/gpu/drm/i915/i915_gem_gtt.h | 1 + > > > drivers/gpu/drm/i915/i915_gem_object.h | 295 +----------------- > > > 9 files changed, 312 insertions(+), 294 deletions(-) > > > create mode 100644 drivers/gpu/drm/i915/gem/Makefile > > > create mode 100644 drivers/gpu/drm/i915/gem/Makefile.header-test > > > create mode 100644 drivers/gpu/drm/i915/gem/i915_gem_object_types.h > > > > > > diff --git a/drivers/gpu/drm/i915/Makefile b/drivers/gpu/drm/i915/Makefile > > > index 58643373495c..def781c9ea69 100644 > > > --- a/drivers/gpu/drm/i915/Makefile > > > +++ b/drivers/gpu/drm/i915/Makefile > > > @@ -85,6 +85,7 @@ gt-$(CONFIG_DRM_I915_SELFTEST) += \ > > > i915-y += $(gt-y) > > > > > > # GEM (Graphics Execution Management) code > > > +obj-y += gem/ > > > > I think this warrants similar treatment as gt/. Posted standalone > > instead of hidden in a series, explicit acks. > > > > Provided the split makes sense to Joonas, > > > > Acked-by: Jani Nikula <jani.nikula@intel.com> > > Yeah, makes sense to me. > > Acked-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com> > > Regards, Joonas I liked the idea of splitting gem from gt, so: Acked-by: Rodrigo Vivi <rodrigo.vivi@intel.com> > > > > > > > > i915-y += \ > > > i915_active.o \ > > > i915_cmd_parser.o \ > > > diff --git a/drivers/gpu/drm/i915/gem/Makefile b/drivers/gpu/drm/i915/gem/Makefile > > > new file mode 100644 > > > index 000000000000..07e7b8b840ea > > > --- /dev/null > > > +++ b/drivers/gpu/drm/i915/gem/Makefile > > > @@ -0,0 +1 @@ > > > +include $(src)/Makefile.header-test # Extra header tests > > > diff --git a/drivers/gpu/drm/i915/gem/Makefile.header-test b/drivers/gpu/drm/i915/gem/Makefile.header-test > > > new file mode 100644 > > > index 000000000000..61e06cbb4b32 > > > --- /dev/null > > > +++ b/drivers/gpu/drm/i915/gem/Makefile.header-test > > > @@ -0,0 +1,16 @@ > > > +# SPDX-License-Identifier: MIT > > > +# Copyright © 2019 Intel Corporation > > > + > > > +# Test the headers are compilable as standalone units > > > +header_test := $(notdir $(wildcard $(src)/*.h)) > > > + > > > +quiet_cmd_header_test = HDRTEST $@ > > > + cmd_header_test = echo "\#include \"$(<F)\"" > $@ > > > + > > > +header_test_%.c: %.h > > > + $(call cmd,header_test) > > > + > > > +extra-$(CONFIG_DRM_I915_WERROR) += \ > > > + $(foreach h,$(header_test),$(patsubst %.h,header_test_%.o,$(h))) > > > + > > > +clean-files += $(foreach h,$(header_test),$(patsubst %.h,header_test_%.c,$(h))) > > > diff --git a/drivers/gpu/drm/i915/gem/i915_gem_object_types.h b/drivers/gpu/drm/i915/gem/i915_gem_object_types.h > > > new file mode 100644 > > > index 000000000000..e4b50944f553 > > > --- /dev/null > > > +++ b/drivers/gpu/drm/i915/gem/i915_gem_object_types.h > > > @@ -0,0 +1,285 @@ > > > +/* > > > + * SPDX-License-Identifier: MIT > > > + * > > > + * Copyright © 2016 Intel Corporation > > > + */ > > > + > > > +#ifndef __I915_GEM_OBJECT_TYPES_H__ > > > +#define __I915_GEM_OBJECT_TYPES_H__ > > > + > > > +#include <linux/reservation.h> > > > + > > > +#include <drm/drm_gem.h> > > > + > > > +#include "../i915_active.h" > > > +#include "../i915_selftest.h" > > > + > > > +struct drm_i915_gem_object; > > > + > > > +/* > > > + * struct i915_lut_handle tracks the fast lookups from handle to vma used > > > + * for execbuf. Although we use a radixtree for that mapping, in order to > > > + * remove them as the object or context is closed, we need a secondary list > > > + * and a translation entry (i915_lut_handle). > > > + */ > > > +struct i915_lut_handle { > > > + struct list_head obj_link; > > > + struct list_head ctx_link; > > > + struct i915_gem_context *ctx; > > > + u32 handle; > > > +}; > > > + > > > +struct drm_i915_gem_object_ops { > > > + unsigned int flags; > > > +#define I915_GEM_OBJECT_HAS_STRUCT_PAGE BIT(0) > > > +#define I915_GEM_OBJECT_IS_SHRINKABLE BIT(1) > > > +#define I915_GEM_OBJECT_IS_PROXY BIT(2) > > > +#define I915_GEM_OBJECT_ASYNC_CANCEL BIT(3) > > > + > > > + /* Interface between the GEM object and its backing storage. > > > + * get_pages() is called once prior to the use of the associated set > > > + * of pages before to binding them into the GTT, and put_pages() is > > > + * called after we no longer need them. As we expect there to be > > > + * associated cost with migrating pages between the backing storage > > > + * and making them available for the GPU (e.g. clflush), we may hold > > > + * onto the pages after they are no longer referenced by the GPU > > > + * in case they may be used again shortly (for example migrating the > > > + * pages to a different memory domain within the GTT). put_pages() > > > + * will therefore most likely be called when the object itself is > > > + * being released or under memory pressure (where we attempt to > > > + * reap pages for the shrinker). > > > + */ > > > + int (*get_pages)(struct drm_i915_gem_object *obj); > > > + void (*put_pages)(struct drm_i915_gem_object *obj, > > > + struct sg_table *pages); > > > + > > > + int (*pwrite)(struct drm_i915_gem_object *obj, > > > + const struct drm_i915_gem_pwrite *arg); > > > + > > > + int (*dmabuf_export)(struct drm_i915_gem_object *obj); > > > + void (*release)(struct drm_i915_gem_object *obj); > > > +}; > > > + > > > +struct drm_i915_gem_object { > > > + struct drm_gem_object base; > > > + > > > + const struct drm_i915_gem_object_ops *ops; > > > + > > > + struct { > > > + /** > > > + * @vma.lock: protect the list/tree of vmas > > > + */ > > > + spinlock_t lock; > > > + > > > + /** > > > + * @vma.list: List of VMAs backed by this object > > > + * > > > + * The VMA on this list are ordered by type, all GGTT vma are > > > + * placed at the head and all ppGTT vma are placed at the tail. > > > + * The different types of GGTT vma are unordered between > > > + * themselves, use the @vma.tree (which has a defined order > > > + * between all VMA) to quickly find an exact match. > > > + */ > > > + struct list_head list; > > > + > > > + /** > > > + * @vma.tree: Ordered tree of VMAs backed by this object > > > + * > > > + * All VMA created for this object are placed in the @vma.tree > > > + * for fast retrieval via a binary search in > > > + * i915_vma_instance(). They are also added to @vma.list for > > > + * easy iteration. > > > + */ > > > + struct rb_root tree; > > > + } vma; > > > + > > > + /** > > > + * @lut_list: List of vma lookup entries in use for this object. > > > + * > > > + * If this object is closed, we need to remove all of its VMA from > > > + * the fast lookup index in associated contexts; @lut_list provides > > > + * this translation from object to context->handles_vma. > > > + */ > > > + struct list_head lut_list; > > > + > > > + /** Stolen memory for this object, instead of being backed by shmem. */ > > > + struct drm_mm_node *stolen; > > > + union { > > > + struct rcu_head rcu; > > > + struct llist_node freed; > > > + }; > > > + > > > + /** > > > + * Whether the object is currently in the GGTT mmap. > > > + */ > > > + unsigned int userfault_count; > > > + struct list_head userfault_link; > > > + > > > + struct list_head batch_pool_link; > > > + I915_SELFTEST_DECLARE(struct list_head st_link); > > > + > > > + unsigned long flags; > > > + > > > + /** > > > + * Have we taken a reference for the object for incomplete GPU > > > + * activity? > > > + */ > > > +#define I915_BO_ACTIVE_REF 0 > > > + > > > + /* > > > + * Is the object to be mapped as read-only to the GPU > > > + * Only honoured if hardware has relevant pte bit > > > + */ > > > + unsigned int cache_level:3; > > > + unsigned int cache_coherent:2; > > > +#define I915_BO_CACHE_COHERENT_FOR_READ BIT(0) > > > +#define I915_BO_CACHE_COHERENT_FOR_WRITE BIT(1) > > > + unsigned int cache_dirty:1; > > > + > > > + /** > > > + * @read_domains: Read memory domains. > > > + * > > > + * These monitor which caches contain read/write data related to the > > > + * object. When transitioning from one set of domains to another, > > > + * the driver is called to ensure that caches are suitably flushed and > > > + * invalidated. > > > + */ > > > + u16 read_domains; > > > + > > > + /** > > > + * @write_domain: Corresponding unique write memory domain. > > > + */ > > > + u16 write_domain; > > > + > > > + atomic_t frontbuffer_bits; > > > + unsigned int frontbuffer_ggtt_origin; /* write once */ > > > + struct i915_active_request frontbuffer_write; > > > + > > > + /** Current tiling stride for the object, if it's tiled. */ > > > + unsigned int tiling_and_stride; > > > +#define FENCE_MINIMUM_STRIDE 128 /* See i915_tiling_ok() */ > > > +#define TILING_MASK (FENCE_MINIMUM_STRIDE - 1) > > > +#define STRIDE_MASK (~TILING_MASK) > > > + > > > + /** Count of VMA actually bound by this object */ > > > + unsigned int bind_count; > > > + unsigned int active_count; > > > + /** Count of how many global VMA are currently pinned for use by HW */ > > > + unsigned int pin_global; > > > + > > > + struct { > > > + struct mutex lock; /* protects the pages and their use */ > > > + atomic_t pages_pin_count; > > > + > > > + struct sg_table *pages; > > > + void *mapping; > > > + > > > + /* TODO: whack some of this into the error state */ > > > + struct i915_page_sizes { > > > + /** > > > + * The sg mask of the pages sg_table. i.e the mask of > > > + * of the lengths for each sg entry. > > > + */ > > > + unsigned int phys; > > > + > > > + /** > > > + * The gtt page sizes we are allowed to use given the > > > + * sg mask and the supported page sizes. This will > > > + * express the smallest unit we can use for the whole > > > + * object, as well as the larger sizes we may be able > > > + * to use opportunistically. > > > + */ > > > + unsigned int sg; > > > + > > > + /** > > > + * The actual gtt page size usage. Since we can have > > > + * multiple vma associated with this object we need to > > > + * prevent any trampling of state, hence a copy of this > > > + * struct also lives in each vma, therefore the gtt > > > + * value here should only be read/write through the vma. > > > + */ > > > + unsigned int gtt; > > > + } page_sizes; > > > + > > > + I915_SELFTEST_DECLARE(unsigned int page_mask); > > > + > > > + struct i915_gem_object_page_iter { > > > + struct scatterlist *sg_pos; > > > + unsigned int sg_idx; /* in pages, but 32bit eek! */ > > > + > > > + struct radix_tree_root radix; > > > + struct mutex lock; /* protects this cache */ > > > + } get_page; > > > + > > > + /** > > > + * Element within i915->mm.unbound_list or i915->mm.bound_list, > > > + * locked by i915->mm.obj_lock. > > > + */ > > > + struct list_head link; > > > + > > > + /** > > > + * Advice: are the backing pages purgeable? > > > + */ > > > + unsigned int madv:2; > > > + > > > + /** > > > + * This is set if the object has been written to since the > > > + * pages were last acquired. > > > + */ > > > + bool dirty:1; > > > + > > > + /** > > > + * This is set if the object has been pinned due to unknown > > > + * swizzling. > > > + */ > > > + bool quirked:1; > > > + } mm; > > > + > > > + /** Breadcrumb of last rendering to the buffer. > > > + * There can only be one writer, but we allow for multiple readers. > > > + * If there is a writer that necessarily implies that all other > > > + * read requests are complete - but we may only be lazily clearing > > > + * the read requests. A read request is naturally the most recent > > > + * request on a ring, so we may have two different write and read > > > + * requests on one ring where the write request is older than the > > > + * read request. This allows for the CPU to read from an active > > > + * buffer by only waiting for the write to complete. > > > + */ > > > + struct reservation_object *resv; > > > + > > > + /** References from framebuffers, locks out tiling changes. */ > > > + unsigned int framebuffer_references; > > > + > > > + /** Record of address bit 17 of each page at last unbind. */ > > > + unsigned long *bit_17; > > > + > > > + union { > > > + struct i915_gem_userptr { > > > + uintptr_t ptr; > > > + > > > + struct i915_mm_struct *mm; > > > + struct i915_mmu_object *mmu_object; > > > + struct work_struct *work; > > > + } userptr; > > > + > > > + unsigned long scratch; > > > + > > > + void *gvt_info; > > > + }; > > > + > > > + /** for phys allocated objects */ > > > + struct drm_dma_handle *phys_handle; > > > + > > > + struct reservation_object __builtin_resv; > > > +}; > > > + > > > +static inline struct drm_i915_gem_object * > > > +to_intel_bo(struct drm_gem_object *gem) > > > +{ > > > + /* Assert that to_intel_bo(NULL) == NULL */ > > > + BUILD_BUG_ON(offsetof(struct drm_i915_gem_object, base)); > > > + > > > + return container_of(gem, struct drm_i915_gem_object, base); > > > +} > > > + > > > +#endif > > > diff --git a/drivers/gpu/drm/i915/gt/intel_engine_types.h b/drivers/gpu/drm/i915/gt/intel_engine_types.h > > > index ca8d95a5708d..ae73c6596d08 100644 > > > --- a/drivers/gpu/drm/i915/gt/intel_engine_types.h > > > +++ b/drivers/gpu/drm/i915/gt/intel_engine_types.h > > > @@ -29,6 +29,7 @@ > > > #define I915_CMD_HASH_ORDER 9 > > > > > > struct dma_fence; > > > +struct drm_i915_gem_object; > > > struct drm_i915_reg_table; > > > struct i915_gem_context; > > > struct i915_request; > > > diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h > > > index 5ca4df9a7428..0cf87495e11b 100644 > > > --- a/drivers/gpu/drm/i915/i915_drv.h > > > +++ b/drivers/gpu/drm/i915/i915_drv.h > > > @@ -80,7 +80,6 @@ > > > #include "i915_gem.h" > > > #include "i915_gem_context.h" > > > #include "i915_gem_fence_reg.h" > > > -#include "i915_gem_object.h" > > > #include "i915_gem_gtt.h" > > > #include "i915_gpu_error.h" > > > #include "i915_request.h" > > > @@ -135,6 +134,8 @@ bool i915_error_injected(void); > > > __i915_printk(i915, i915_error_injected() ? KERN_DEBUG : KERN_ERR, \ > > > fmt, ##__VA_ARGS__) > > > > > > +struct drm_i915_gem_object; > > > + > > > enum hpd_pin { > > > HPD_NONE = 0, > > > HPD_TV = HPD_NONE, /* TV is known to be unreliable */ > > > diff --git a/drivers/gpu/drm/i915/i915_gem_batch_pool.h b/drivers/gpu/drm/i915/i915_gem_batch_pool.h > > > index 56947daaaf65..feeeeeaa54d8 100644 > > > --- a/drivers/gpu/drm/i915/i915_gem_batch_pool.h > > > +++ b/drivers/gpu/drm/i915/i915_gem_batch_pool.h > > > @@ -9,6 +9,7 @@ > > > > > > #include <linux/types.h> > > > > > > +struct drm_i915_gem_object; > > > struct intel_engine_cs; > > > > > > struct i915_gem_batch_pool { > > > @@ -19,7 +20,7 @@ struct i915_gem_batch_pool { > > > void i915_gem_batch_pool_init(struct i915_gem_batch_pool *pool, > > > struct intel_engine_cs *engine); > > > void i915_gem_batch_pool_fini(struct i915_gem_batch_pool *pool); > > > -struct drm_i915_gem_object* > > > +struct drm_i915_gem_object * > > > i915_gem_batch_pool_get(struct i915_gem_batch_pool *pool, size_t size); > > > > > > #endif /* I915_GEM_BATCH_POOL_H */ > > > diff --git a/drivers/gpu/drm/i915/i915_gem_gtt.h b/drivers/gpu/drm/i915/i915_gem_gtt.h > > > index 2fafa04c45ec..593079f30fbe 100644 > > > --- a/drivers/gpu/drm/i915/i915_gem_gtt.h > > > +++ b/drivers/gpu/drm/i915/i915_gem_gtt.h > > > @@ -61,6 +61,7 @@ > > > > > > struct drm_i915_file_private; > > > struct drm_i915_fence_reg; > > > +struct drm_i915_gem_object; > > > struct i915_vma; > > > > > > typedef u32 gen6_pte_t; > > > diff --git a/drivers/gpu/drm/i915/i915_gem_object.h b/drivers/gpu/drm/i915/i915_gem_object.h > > > index ca93a40c0c87..3666b0c5f6ca 100644 > > > --- a/drivers/gpu/drm/i915/i915_gem_object.h > > > +++ b/drivers/gpu/drm/i915/i915_gem_object.h > > > @@ -1,308 +1,19 @@ > > > /* > > > - * Copyright © 2016 Intel Corporation > > > - * > > > - * Permission is hereby granted, free of charge, to any person obtaining a > > > - * copy of this software and associated documentation files (the "Software"), > > > - * to deal in the Software without restriction, including without limitation > > > - * the rights to use, copy, modify, merge, publish, distribute, sublicense, > > > - * and/or sell copies of the Software, and to permit persons to whom the > > > - * Software is furnished to do so, subject to the following conditions: > > > - * > > > - * The above copyright notice and this permission notice (including the next > > > - * paragraph) shall be included in all copies or substantial portions of the > > > - * Software. > > > - * > > > - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR > > > - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, > > > - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL > > > - * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER > > > - * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING > > > - * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS > > > - * IN THE SOFTWARE. > > > + * SPDX-License-Identifier: MIT > > > * > > > + * Copyright © 2016 Intel Corporation > > > */ > > > > > > #ifndef __I915_GEM_OBJECT_H__ > > > #define __I915_GEM_OBJECT_H__ > > > > > > -#include <linux/reservation.h> > > > - > > > -#include <drm/drm_vma_manager.h> > > > #include <drm/drm_gem.h> > > > #include <drm/drm_file.h> > > > #include <drm/drm_device.h> > > > > > > #include <drm/i915_drm.h> > > > > > > -#include "i915_request.h" > > > -#include "i915_selftest.h" > > > - > > > -struct drm_i915_gem_object; > > > - > > > -/* > > > - * struct i915_lut_handle tracks the fast lookups from handle to vma used > > > - * for execbuf. Although we use a radixtree for that mapping, in order to > > > - * remove them as the object or context is closed, we need a secondary list > > > - * and a translation entry (i915_lut_handle). > > > - */ > > > -struct i915_lut_handle { > > > - struct list_head obj_link; > > > - struct list_head ctx_link; > > > - struct i915_gem_context *ctx; > > > - u32 handle; > > > -}; > > > - > > > -struct drm_i915_gem_object_ops { > > > - unsigned int flags; > > > -#define I915_GEM_OBJECT_HAS_STRUCT_PAGE BIT(0) > > > -#define I915_GEM_OBJECT_IS_SHRINKABLE BIT(1) > > > -#define I915_GEM_OBJECT_IS_PROXY BIT(2) > > > -#define I915_GEM_OBJECT_ASYNC_CANCEL BIT(3) > > > - > > > - /* Interface between the GEM object and its backing storage. > > > - * get_pages() is called once prior to the use of the associated set > > > - * of pages before to binding them into the GTT, and put_pages() is > > > - * called after we no longer need them. As we expect there to be > > > - * associated cost with migrating pages between the backing storage > > > - * and making them available for the GPU (e.g. clflush), we may hold > > > - * onto the pages after they are no longer referenced by the GPU > > > - * in case they may be used again shortly (for example migrating the > > > - * pages to a different memory domain within the GTT). put_pages() > > > - * will therefore most likely be called when the object itself is > > > - * being released or under memory pressure (where we attempt to > > > - * reap pages for the shrinker). > > > - */ > > > - int (*get_pages)(struct drm_i915_gem_object *); > > > - void (*put_pages)(struct drm_i915_gem_object *, struct sg_table *); > > > - > > > - int (*pwrite)(struct drm_i915_gem_object *, > > > - const struct drm_i915_gem_pwrite *); > > > - > > > - int (*dmabuf_export)(struct drm_i915_gem_object *); > > > - void (*release)(struct drm_i915_gem_object *); > > > -}; > > > - > > > -struct drm_i915_gem_object { > > > - struct drm_gem_object base; > > > - > > > - const struct drm_i915_gem_object_ops *ops; > > > - > > > - struct { > > > - /** > > > - * @vma.lock: protect the list/tree of vmas > > > - */ > > > - spinlock_t lock; > > > - > > > - /** > > > - * @vma.list: List of VMAs backed by this object > > > - * > > > - * The VMA on this list are ordered by type, all GGTT vma are > > > - * placed at the head and all ppGTT vma are placed at the tail. > > > - * The different types of GGTT vma are unordered between > > > - * themselves, use the @vma.tree (which has a defined order > > > - * between all VMA) to quickly find an exact match. > > > - */ > > > - struct list_head list; > > > - > > > - /** > > > - * @vma.tree: Ordered tree of VMAs backed by this object > > > - * > > > - * All VMA created for this object are placed in the @vma.tree > > > - * for fast retrieval via a binary search in > > > - * i915_vma_instance(). They are also added to @vma.list for > > > - * easy iteration. > > > - */ > > > - struct rb_root tree; > > > - } vma; > > > - > > > - /** > > > - * @lut_list: List of vma lookup entries in use for this object. > > > - * > > > - * If this object is closed, we need to remove all of its VMA from > > > - * the fast lookup index in associated contexts; @lut_list provides > > > - * this translation from object to context->handles_vma. > > > - */ > > > - struct list_head lut_list; > > > - > > > - /** Stolen memory for this object, instead of being backed by shmem. */ > > > - struct drm_mm_node *stolen; > > > - union { > > > - struct rcu_head rcu; > > > - struct llist_node freed; > > > - }; > > > - > > > - /** > > > - * Whether the object is currently in the GGTT mmap. > > > - */ > > > - unsigned int userfault_count; > > > - struct list_head userfault_link; > > > - > > > - struct list_head batch_pool_link; > > > - I915_SELFTEST_DECLARE(struct list_head st_link); > > > - > > > - unsigned long flags; > > > - > > > - /** > > > - * Have we taken a reference for the object for incomplete GPU > > > - * activity? > > > - */ > > > -#define I915_BO_ACTIVE_REF 0 > > > - > > > - /* > > > - * Is the object to be mapped as read-only to the GPU > > > - * Only honoured if hardware has relevant pte bit > > > - */ > > > - unsigned int cache_level:3; > > > - unsigned int cache_coherent:2; > > > -#define I915_BO_CACHE_COHERENT_FOR_READ BIT(0) > > > -#define I915_BO_CACHE_COHERENT_FOR_WRITE BIT(1) > > > - unsigned int cache_dirty:1; > > > - > > > - /** > > > - * @read_domains: Read memory domains. > > > - * > > > - * These monitor which caches contain read/write data related to the > > > - * object. When transitioning from one set of domains to another, > > > - * the driver is called to ensure that caches are suitably flushed and > > > - * invalidated. > > > - */ > > > - u16 read_domains; > > > - > > > - /** > > > - * @write_domain: Corresponding unique write memory domain. > > > - */ > > > - u16 write_domain; > > > - > > > - atomic_t frontbuffer_bits; > > > - unsigned int frontbuffer_ggtt_origin; /* write once */ > > > - struct i915_active_request frontbuffer_write; > > > - > > > - /** Current tiling stride for the object, if it's tiled. */ > > > - unsigned int tiling_and_stride; > > > -#define FENCE_MINIMUM_STRIDE 128 /* See i915_tiling_ok() */ > > > -#define TILING_MASK (FENCE_MINIMUM_STRIDE-1) > > > -#define STRIDE_MASK (~TILING_MASK) > > > - > > > - /** Count of VMA actually bound by this object */ > > > - unsigned int bind_count; > > > - unsigned int active_count; > > > - /** Count of how many global VMA are currently pinned for use by HW */ > > > - unsigned int pin_global; > > > - > > > - struct { > > > - struct mutex lock; /* protects the pages and their use */ > > > - atomic_t pages_pin_count; > > > - > > > - struct sg_table *pages; > > > - void *mapping; > > > - > > > - /* TODO: whack some of this into the error state */ > > > - struct i915_page_sizes { > > > - /** > > > - * The sg mask of the pages sg_table. i.e the mask of > > > - * of the lengths for each sg entry. > > > - */ > > > - unsigned int phys; > > > - > > > - /** > > > - * The gtt page sizes we are allowed to use given the > > > - * sg mask and the supported page sizes. This will > > > - * express the smallest unit we can use for the whole > > > - * object, as well as the larger sizes we may be able > > > - * to use opportunistically. > > > - */ > > > - unsigned int sg; > > > - > > > - /** > > > - * The actual gtt page size usage. Since we can have > > > - * multiple vma associated with this object we need to > > > - * prevent any trampling of state, hence a copy of this > > > - * struct also lives in each vma, therefore the gtt > > > - * value here should only be read/write through the vma. > > > - */ > > > - unsigned int gtt; > > > - } page_sizes; > > > - > > > - I915_SELFTEST_DECLARE(unsigned int page_mask); > > > - > > > - struct i915_gem_object_page_iter { > > > - struct scatterlist *sg_pos; > > > - unsigned int sg_idx; /* in pages, but 32bit eek! */ > > > - > > > - struct radix_tree_root radix; > > > - struct mutex lock; /* protects this cache */ > > > - } get_page; > > > - > > > - /** > > > - * Element within i915->mm.unbound_list or i915->mm.bound_list, > > > - * locked by i915->mm.obj_lock. > > > - */ > > > - struct list_head link; > > > - > > > - /** > > > - * Advice: are the backing pages purgeable? > > > - */ > > > - unsigned int madv:2; > > > - > > > - /** > > > - * This is set if the object has been written to since the > > > - * pages were last acquired. > > > - */ > > > - bool dirty:1; > > > - > > > - /** > > > - * This is set if the object has been pinned due to unknown > > > - * swizzling. > > > - */ > > > - bool quirked:1; > > > - } mm; > > > - > > > - /** Breadcrumb of last rendering to the buffer. > > > - * There can only be one writer, but we allow for multiple readers. > > > - * If there is a writer that necessarily implies that all other > > > - * read requests are complete - but we may only be lazily clearing > > > - * the read requests. A read request is naturally the most recent > > > - * request on a ring, so we may have two different write and read > > > - * requests on one ring where the write request is older than the > > > - * read request. This allows for the CPU to read from an active > > > - * buffer by only waiting for the write to complete. > > > - */ > > > - struct reservation_object *resv; > > > - > > > - /** References from framebuffers, locks out tiling changes. */ > > > - unsigned int framebuffer_references; > > > - > > > - /** Record of address bit 17 of each page at last unbind. */ > > > - unsigned long *bit_17; > > > - > > > - union { > > > - struct i915_gem_userptr { > > > - uintptr_t ptr; > > > - > > > - struct i915_mm_struct *mm; > > > - struct i915_mmu_object *mmu_object; > > > - struct work_struct *work; > > > - } userptr; > > > - > > > - unsigned long scratch; > > > - > > > - void *gvt_info; > > > - }; > > > - > > > - /** for phys allocated objects */ > > > - struct drm_dma_handle *phys_handle; > > > - > > > - struct reservation_object __builtin_resv; > > > -}; > > > - > > > -static inline struct drm_i915_gem_object * > > > -to_intel_bo(struct drm_gem_object *gem) > > > -{ > > > - /* Assert that to_intel_bo(NULL) == NULL */ > > > - BUILD_BUG_ON(offsetof(struct drm_i915_gem_object, base)); > > > - > > > - return container_of(gem, struct drm_i915_gem_object, base); > > > -} > > > +#include "gem/i915_gem_object_types.h" > > > > > > struct drm_i915_gem_object *i915_gem_object_alloc(void); > > > void i915_gem_object_free(struct drm_i915_gem_object *obj); > > > > -- > > Jani Nikula, Intel Open Source Graphics Center > _______________________________________________ > Intel-gfx mailing list > Intel-gfx@lists.freedesktop.org > https://lists.freedesktop.org/mailman/listinfo/intel-gfx