drm/i915: Move gtt and ppgtt under address space umbrella
The GTT and PPGTT can be thought of more generally as GPU address
spaces. Many of their actions (insert entries), state (LRU lists), and
many of their characteristics (size) can be shared. Do that.
The change itself doesn't actually impact most of the VMA/VM rework
coming up, it just fits in with the grand scheme of abstracting the GPU
VM operations. GGTT will usually be a special case where we either know
an object must be in the GGTT (dislay engine, workarounds, etc.).
The scratch page is left as part of the VM (even though it's currently
shared with the ppgtt code) because in the future when we have Full
PPGTT, I intend to create a separate scratch page for each.
v2: Drop usage of i915_gtt_vm (Daniel)
Make cleanup also part of the parent class (Ben)
Modified commit msg
Rebased
v3: Properly share scratch page (Imre)
Finish commit message (Daniel, Imre)
Signed-off-by: Ben Widawsky <ben@bwidawsk.net>
Reviewed-by: Imre Deak <imre.deak@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
diff --git a/drivers/gpu/drm/i915/i915_gem_gtt.c b/drivers/gpu/drm/i915/i915_gem_gtt.c
index 4222733..f982bf0 100644
--- a/drivers/gpu/drm/i915/i915_gem_gtt.c
+++ b/drivers/gpu/drm/i915/i915_gem_gtt.c
@@ -124,7 +124,7 @@
static void gen6_write_pdes(struct i915_hw_ppgtt *ppgtt)
{
- struct drm_i915_private *dev_priv = ppgtt->dev->dev_private;
+ struct drm_i915_private *dev_priv = ppgtt->base.dev->dev_private;
gen6_gtt_pte_t __iomem *pd_addr;
uint32_t pd_entry;
int i;
@@ -203,18 +203,18 @@
}
/* PPGTT support for Sandybdrige/Gen6 and later */
-static void gen6_ppgtt_clear_range(struct i915_hw_ppgtt *ppgtt,
+static void gen6_ppgtt_clear_range(struct i915_address_space *vm,
unsigned first_entry,
unsigned num_entries)
{
- struct drm_i915_private *dev_priv = ppgtt->dev->dev_private;
+ struct i915_hw_ppgtt *ppgtt =
+ container_of(vm, struct i915_hw_ppgtt, base);
gen6_gtt_pte_t *pt_vaddr, scratch_pte;
unsigned act_pt = first_entry / I915_PPGTT_PT_ENTRIES;
unsigned first_pte = first_entry % I915_PPGTT_PT_ENTRIES;
unsigned last_pte, i;
- scratch_pte = ppgtt->pte_encode(dev_priv->gtt.scratch.addr,
- I915_CACHE_LLC);
+ scratch_pte = vm->pte_encode(vm->scratch.addr, I915_CACHE_LLC);
while (num_entries) {
last_pte = first_pte + num_entries;
@@ -234,11 +234,13 @@
}
}
-static void gen6_ppgtt_insert_entries(struct i915_hw_ppgtt *ppgtt,
+static void gen6_ppgtt_insert_entries(struct i915_address_space *vm,
struct sg_table *pages,
unsigned first_entry,
enum i915_cache_level cache_level)
{
+ struct i915_hw_ppgtt *ppgtt =
+ container_of(vm, struct i915_hw_ppgtt, base);
gen6_gtt_pte_t *pt_vaddr;
unsigned act_pt = first_entry / I915_PPGTT_PT_ENTRIES;
unsigned act_pte = first_entry % I915_PPGTT_PT_ENTRIES;
@@ -249,7 +251,7 @@
dma_addr_t page_addr;
page_addr = sg_page_iter_dma_address(&sg_iter);
- pt_vaddr[act_pte] = ppgtt->pte_encode(page_addr, cache_level);
+ pt_vaddr[act_pte] = vm->pte_encode(page_addr, cache_level);
if (++act_pte == I915_PPGTT_PT_ENTRIES) {
kunmap_atomic(pt_vaddr);
act_pt++;
@@ -261,13 +263,15 @@
kunmap_atomic(pt_vaddr);
}
-static void gen6_ppgtt_cleanup(struct i915_hw_ppgtt *ppgtt)
+static void gen6_ppgtt_cleanup(struct i915_address_space *vm)
{
+ struct i915_hw_ppgtt *ppgtt =
+ container_of(vm, struct i915_hw_ppgtt, base);
int i;
if (ppgtt->pt_dma_addr) {
for (i = 0; i < ppgtt->num_pd_entries; i++)
- pci_unmap_page(ppgtt->dev->pdev,
+ pci_unmap_page(ppgtt->base.dev->pdev,
ppgtt->pt_dma_addr[i],
4096, PCI_DMA_BIDIRECTIONAL);
}
@@ -281,7 +285,7 @@
static int gen6_ppgtt_init(struct i915_hw_ppgtt *ppgtt)
{
- struct drm_device *dev = ppgtt->dev;
+ struct drm_device *dev = ppgtt->base.dev;
struct drm_i915_private *dev_priv = dev->dev_private;
unsigned first_pd_entry_in_global_pt;
int i;
@@ -293,17 +297,18 @@
first_pd_entry_in_global_pt = gtt_total_entries(dev_priv->gtt);
if (IS_HASWELL(dev)) {
- ppgtt->pte_encode = hsw_pte_encode;
+ ppgtt->base.pte_encode = hsw_pte_encode;
} else if (IS_VALLEYVIEW(dev)) {
- ppgtt->pte_encode = byt_pte_encode;
+ ppgtt->base.pte_encode = byt_pte_encode;
} else {
- ppgtt->pte_encode = gen6_pte_encode;
+ ppgtt->base.pte_encode = gen6_pte_encode;
}
ppgtt->num_pd_entries = GEN6_PPGTT_PD_ENTRIES;
ppgtt->enable = gen6_ppgtt_enable;
- ppgtt->clear_range = gen6_ppgtt_clear_range;
- ppgtt->insert_entries = gen6_ppgtt_insert_entries;
- ppgtt->cleanup = gen6_ppgtt_cleanup;
+ ppgtt->base.clear_range = gen6_ppgtt_clear_range;
+ ppgtt->base.insert_entries = gen6_ppgtt_insert_entries;
+ ppgtt->base.cleanup = gen6_ppgtt_cleanup;
+ ppgtt->base.scratch = dev_priv->gtt.base.scratch;
ppgtt->pt_pages = kzalloc(sizeof(struct page *)*ppgtt->num_pd_entries,
GFP_KERNEL);
if (!ppgtt->pt_pages)
@@ -334,8 +339,8 @@
ppgtt->pt_dma_addr[i] = pt_addr;
}
- ppgtt->clear_range(ppgtt, 0,
- ppgtt->num_pd_entries*I915_PPGTT_PT_ENTRIES);
+ ppgtt->base.clear_range(&ppgtt->base, 0,
+ ppgtt->num_pd_entries * I915_PPGTT_PT_ENTRIES);
ppgtt->pd_offset = first_pd_entry_in_global_pt * sizeof(gen6_gtt_pte_t);
@@ -368,7 +373,7 @@
if (!ppgtt)
return -ENOMEM;
- ppgtt->dev = dev;
+ ppgtt->base.dev = dev;
if (INTEL_INFO(dev)->gen < 8)
ret = gen6_ppgtt_init(ppgtt);
@@ -391,7 +396,7 @@
if (!ppgtt)
return;
- ppgtt->cleanup(ppgtt);
+ ppgtt->base.cleanup(&ppgtt->base);
dev_priv->mm.aliasing_ppgtt = NULL;
}
@@ -399,17 +404,17 @@
struct drm_i915_gem_object *obj,
enum i915_cache_level cache_level)
{
- ppgtt->insert_entries(ppgtt, obj->pages,
- i915_gem_obj_ggtt_offset(obj) >> PAGE_SHIFT,
- cache_level);
+ ppgtt->base.insert_entries(&ppgtt->base, obj->pages,
+ i915_gem_obj_ggtt_offset(obj) >> PAGE_SHIFT,
+ cache_level);
}
void i915_ppgtt_unbind_object(struct i915_hw_ppgtt *ppgtt,
struct drm_i915_gem_object *obj)
{
- ppgtt->clear_range(ppgtt,
- i915_gem_obj_ggtt_offset(obj) >> PAGE_SHIFT,
- obj->base.size >> PAGE_SHIFT);
+ ppgtt->base.clear_range(&ppgtt->base,
+ i915_gem_obj_ggtt_offset(obj) >> PAGE_SHIFT,
+ obj->base.size >> PAGE_SHIFT);
}
extern int intel_iommu_gfx_mapped;
@@ -456,8 +461,9 @@
struct drm_i915_gem_object *obj;
/* First fill our portion of the GTT with scratch pages */
- dev_priv->gtt.gtt_clear_range(dev, dev_priv->gtt.start / PAGE_SIZE,
- dev_priv->gtt.total / PAGE_SIZE);
+ dev_priv->gtt.base.clear_range(&dev_priv->gtt.base,
+ dev_priv->gtt.base.start / PAGE_SIZE,
+ dev_priv->gtt.base.total / PAGE_SIZE);
list_for_each_entry(obj, &dev_priv->mm.bound_list, global_list) {
i915_gem_clflush_object(obj);
@@ -486,12 +492,12 @@
* within the global GTT as well as accessible by the GPU through the GMADR
* mapped BAR (dev_priv->mm.gtt->gtt).
*/
-static void gen6_ggtt_insert_entries(struct drm_device *dev,
+static void gen6_ggtt_insert_entries(struct i915_address_space *vm,
struct sg_table *st,
unsigned int first_entry,
enum i915_cache_level level)
{
- struct drm_i915_private *dev_priv = dev->dev_private;
+ struct drm_i915_private *dev_priv = vm->dev->dev_private;
gen6_gtt_pte_t __iomem *gtt_entries =
(gen6_gtt_pte_t __iomem *)dev_priv->gtt.gsm + first_entry;
int i = 0;
@@ -500,8 +506,7 @@
for_each_sg_page(st->sgl, &sg_iter, st->nents, 0) {
addr = sg_page_iter_dma_address(&sg_iter);
- iowrite32(dev_priv->gtt.pte_encode(addr, level),
- >t_entries[i]);
+ iowrite32(vm->pte_encode(addr, level), >t_entries[i]);
i++;
}
@@ -512,8 +517,8 @@
* hardware should work, we must keep this posting read for paranoia.
*/
if (i != 0)
- WARN_ON(readl(>t_entries[i-1])
- != dev_priv->gtt.pte_encode(addr, level));
+ WARN_ON(readl(>t_entries[i-1]) !=
+ vm->pte_encode(addr, level));
/* This next bit makes the above posting read even more important. We
* want to flush the TLBs only after we're certain all the PTE updates
@@ -523,11 +528,11 @@
POSTING_READ(GFX_FLSH_CNTL_GEN6);
}
-static void gen6_ggtt_clear_range(struct drm_device *dev,
+static void gen6_ggtt_clear_range(struct i915_address_space *vm,
unsigned int first_entry,
unsigned int num_entries)
{
- struct drm_i915_private *dev_priv = dev->dev_private;
+ struct drm_i915_private *dev_priv = vm->dev->dev_private;
gen6_gtt_pte_t scratch_pte, __iomem *gtt_base =
(gen6_gtt_pte_t __iomem *) dev_priv->gtt.gsm + first_entry;
const int max_entries = gtt_total_entries(dev_priv->gtt) - first_entry;
@@ -538,15 +543,14 @@
first_entry, num_entries, max_entries))
num_entries = max_entries;
- scratch_pte = dev_priv->gtt.pte_encode(dev_priv->gtt.scratch.addr,
- I915_CACHE_LLC);
+ scratch_pte = vm->pte_encode(vm->scratch.addr, I915_CACHE_LLC);
for (i = 0; i < num_entries; i++)
iowrite32(scratch_pte, >t_base[i]);
readl(gtt_base);
}
-static void i915_ggtt_insert_entries(struct drm_device *dev,
+static void i915_ggtt_insert_entries(struct i915_address_space *vm,
struct sg_table *st,
unsigned int pg_start,
enum i915_cache_level cache_level)
@@ -558,7 +562,7 @@
}
-static void i915_ggtt_clear_range(struct drm_device *dev,
+static void i915_ggtt_clear_range(struct i915_address_space *vm,
unsigned int first_entry,
unsigned int num_entries)
{
@@ -571,10 +575,11 @@
{
struct drm_device *dev = obj->base.dev;
struct drm_i915_private *dev_priv = dev->dev_private;
+ const unsigned long entry = i915_gem_obj_ggtt_offset(obj) >> PAGE_SHIFT;
- dev_priv->gtt.gtt_insert_entries(dev, obj->pages,
- i915_gem_obj_ggtt_offset(obj) >> PAGE_SHIFT,
- cache_level);
+ dev_priv->gtt.base.insert_entries(&dev_priv->gtt.base, obj->pages,
+ entry,
+ cache_level);
obj->has_global_gtt_mapping = 1;
}
@@ -583,10 +588,11 @@
{
struct drm_device *dev = obj->base.dev;
struct drm_i915_private *dev_priv = dev->dev_private;
+ const unsigned long entry = i915_gem_obj_ggtt_offset(obj) >> PAGE_SHIFT;
- dev_priv->gtt.gtt_clear_range(obj->base.dev,
- i915_gem_obj_ggtt_offset(obj) >> PAGE_SHIFT,
- obj->base.size >> PAGE_SHIFT);
+ dev_priv->gtt.base.clear_range(&dev_priv->gtt.base,
+ entry,
+ obj->base.size >> PAGE_SHIFT);
obj->has_global_gtt_mapping = 0;
}
@@ -663,20 +669,23 @@
obj->has_global_gtt_mapping = 1;
}
- dev_priv->gtt.start = start;
- dev_priv->gtt.total = end - start;
+ dev_priv->gtt.base.start = start;
+ dev_priv->gtt.base.total = end - start;
/* Clear any non-preallocated blocks */
drm_mm_for_each_hole(entry, &dev_priv->mm.gtt_space,
hole_start, hole_end) {
+ const unsigned long count = (hole_end - hole_start) / PAGE_SIZE;
DRM_DEBUG_KMS("clearing unused GTT space: [%lx, %lx]\n",
hole_start, hole_end);
- dev_priv->gtt.gtt_clear_range(dev, hole_start / PAGE_SIZE,
- (hole_end-hole_start) / PAGE_SIZE);
+ dev_priv->gtt.base.clear_range(&dev_priv->gtt.base,
+ hole_start / PAGE_SIZE,
+ count);
}
/* And finally clear the reserved guard page */
- dev_priv->gtt.gtt_clear_range(dev, end / PAGE_SIZE - 1, 1);
+ dev_priv->gtt.base.clear_range(&dev_priv->gtt.base,
+ end / PAGE_SIZE - 1, 1);
}
static bool
@@ -699,7 +708,7 @@
struct drm_i915_private *dev_priv = dev->dev_private;
unsigned long gtt_size, mappable_size;
- gtt_size = dev_priv->gtt.total;
+ gtt_size = dev_priv->gtt.base.total;
mappable_size = dev_priv->gtt.mappable_end;
if (intel_enable_ppgtt(dev) && HAS_ALIASING_PPGTT(dev)) {
@@ -744,8 +753,8 @@
#else
dma_addr = page_to_phys(page);
#endif
- dev_priv->gtt.scratch.page = page;
- dev_priv->gtt.scratch.addr = dma_addr;
+ dev_priv->gtt.base.scratch.page = page;
+ dev_priv->gtt.base.scratch.addr = dma_addr;
return 0;
}
@@ -753,11 +762,13 @@
static void teardown_scratch_page(struct drm_device *dev)
{
struct drm_i915_private *dev_priv = dev->dev_private;
- set_pages_wb(dev_priv->gtt.scratch.page, 1);
- pci_unmap_page(dev->pdev, dev_priv->gtt.scratch.addr,
+ struct page *page = dev_priv->gtt.base.scratch.page;
+
+ set_pages_wb(page, 1);
+ pci_unmap_page(dev->pdev, dev_priv->gtt.base.scratch.addr,
PAGE_SIZE, PCI_DMA_BIDIRECTIONAL);
- put_page(dev_priv->gtt.scratch.page);
- __free_page(dev_priv->gtt.scratch.page);
+ put_page(page);
+ __free_page(page);
}
static inline unsigned int gen6_get_total_gtt_size(u16 snb_gmch_ctl)
@@ -820,17 +831,18 @@
if (ret)
DRM_ERROR("Scratch setup failed\n");
- dev_priv->gtt.gtt_clear_range = gen6_ggtt_clear_range;
- dev_priv->gtt.gtt_insert_entries = gen6_ggtt_insert_entries;
+ dev_priv->gtt.base.clear_range = gen6_ggtt_clear_range;
+ dev_priv->gtt.base.insert_entries = gen6_ggtt_insert_entries;
return ret;
}
-static void gen6_gmch_remove(struct drm_device *dev)
+static void gen6_gmch_remove(struct i915_address_space *vm)
{
- struct drm_i915_private *dev_priv = dev->dev_private;
- iounmap(dev_priv->gtt.gsm);
- teardown_scratch_page(dev_priv->dev);
+
+ struct i915_gtt *gtt = container_of(vm, struct i915_gtt, base);
+ iounmap(gtt->gsm);
+ teardown_scratch_page(vm->dev);
}
static int i915_gmch_probe(struct drm_device *dev,
@@ -851,13 +863,13 @@
intel_gtt_get(gtt_total, stolen, mappable_base, mappable_end);
dev_priv->gtt.do_idle_maps = needs_idle_maps(dev_priv->dev);
- dev_priv->gtt.gtt_clear_range = i915_ggtt_clear_range;
- dev_priv->gtt.gtt_insert_entries = i915_ggtt_insert_entries;
+ dev_priv->gtt.base.clear_range = i915_ggtt_clear_range;
+ dev_priv->gtt.base.insert_entries = i915_ggtt_insert_entries;
return 0;
}
-static void i915_gmch_remove(struct drm_device *dev)
+static void i915_gmch_remove(struct i915_address_space *vm)
{
intel_gmch_remove();
}
@@ -870,27 +882,30 @@
if (INTEL_INFO(dev)->gen <= 5) {
gtt->gtt_probe = i915_gmch_probe;
- gtt->gtt_remove = i915_gmch_remove;
+ gtt->base.cleanup = i915_gmch_remove;
} else {
gtt->gtt_probe = gen6_gmch_probe;
- gtt->gtt_remove = gen6_gmch_remove;
+ gtt->base.cleanup = gen6_gmch_remove;
if (IS_HASWELL(dev) && dev_priv->ellc_size)
- gtt->pte_encode = iris_pte_encode;
+ gtt->base.pte_encode = iris_pte_encode;
else if (IS_HASWELL(dev))
- gtt->pte_encode = hsw_pte_encode;
+ gtt->base.pte_encode = hsw_pte_encode;
else if (IS_VALLEYVIEW(dev))
- gtt->pte_encode = byt_pte_encode;
+ gtt->base.pte_encode = byt_pte_encode;
else
- gtt->pte_encode = gen6_pte_encode;
+ gtt->base.pte_encode = gen6_pte_encode;
}
- ret = gtt->gtt_probe(dev, >t->total, >t->stolen_size,
+ ret = gtt->gtt_probe(dev, >t->base.total, >t->stolen_size,
>t->mappable_base, >t->mappable_end);
if (ret)
return ret;
+ gtt->base.dev = dev;
+
/* GMADR is the PCI mmio aperture into the global GTT. */
- DRM_INFO("Memory usable by graphics device = %zdM\n", gtt->total >> 20);
+ DRM_INFO("Memory usable by graphics device = %zdM\n",
+ gtt->base.total >> 20);
DRM_DEBUG_DRIVER("GMADR size = %ldM\n", gtt->mappable_end >> 20);
DRM_DEBUG_DRIVER("GTT stolen size = %zdM\n", gtt->stolen_size >> 20);