[PATCH 08/20] drm/i915: pass the vma to insert_entries
Matthew Auld
matthew.auld at intel.com
Thu Jun 22 14:42:41 UTC 2017
The vma contains most of the information we need for insertion. But also
in preparation for supporting huge-pages for the ppgtt, it would be
useful to know the details of vma->page_sizes and the node size, such
that we can we can easily determine the page sizes we are allowed to use
when inserting into the 48b PPGTT. This is especially true for 64K
where we can't just arbitrarily use it, since we require
aligning/padding the vm space to 2M, which sometimes we can't enforce in
the upper levels.
Suggested-by: Chris Wilson <chris at chris-wilson.co.uk>
Signed-off-by: Matthew Auld <matthew.auld at intel.com>
Cc: Joonas Lahtinen <joonas.lahtinen at linux.intel.com>
---
drivers/gpu/drm/i915/i915_gem_gtt.c | 68 +++++++++++----------------
drivers/gpu/drm/i915/i915_gem_gtt.h | 3 +-
drivers/gpu/drm/i915/selftests/i915_gem_gtt.c | 10 +++-
drivers/gpu/drm/i915/selftests/mock_gtt.c | 3 +-
4 files changed, 38 insertions(+), 46 deletions(-)
diff --git a/drivers/gpu/drm/i915/i915_gem_gtt.c b/drivers/gpu/drm/i915/i915_gem_gtt.c
index 8c28908a507b..ae362abd4729 100644
--- a/drivers/gpu/drm/i915/i915_gem_gtt.c
+++ b/drivers/gpu/drm/i915/i915_gem_gtt.c
@@ -205,8 +205,7 @@ static int ppgtt_bind_vma(struct i915_vma *vma,
if (vma->obj->gt_ro)
pte_flags |= PTE_READ_ONLY;
- vma->vm->insert_entries(vma->vm, vma->pages, vma->node.start,
- cache_level, pte_flags);
+ vma->vm->insert_entries(vma->vm, vma, cache_level, pte_flags);
return 0;
}
@@ -930,37 +929,35 @@ gen8_ppgtt_insert_pte_entries(struct i915_hw_ppgtt *ppgtt,
}
static void gen8_ppgtt_insert_3lvl(struct i915_address_space *vm,
- struct sg_table *pages,
- u64 start,
+ struct i915_vma *vma,
enum i915_cache_level cache_level,
u32 unused)
{
- struct i915_hw_ppgtt *ppgtt = i915_vm_to_ppgtt(vm);
+ struct i915_hw_ppgtt *ppgtt = i915_vm_to_ppgtt(vma->vm);
struct sgt_dma iter = {
- .sg = pages->sgl,
+ .sg = vma->pages->sgl,
.dma = sg_dma_address(iter.sg),
.max = iter.dma + iter.sg->length,
};
- struct gen8_insert_pte idx = gen8_insert_pte(start);
+ struct gen8_insert_pte idx = gen8_insert_pte(vma->node.start);
gen8_ppgtt_insert_pte_entries(ppgtt, &ppgtt->pdp, &iter, &idx,
cache_level);
}
static void gen8_ppgtt_insert_4lvl(struct i915_address_space *vm,
- struct sg_table *pages,
- u64 start,
+ struct i915_vma *vma,
enum i915_cache_level cache_level,
u32 unused)
{
struct i915_hw_ppgtt *ppgtt = i915_vm_to_ppgtt(vm);
struct sgt_dma iter = {
- .sg = pages->sgl,
+ .sg = vma->pages->sgl,
.dma = sg_dma_address(iter.sg),
.max = iter.dma + iter.sg->length,
};
struct i915_page_directory_pointer **pdps = ppgtt->pml4.pdps;
- struct gen8_insert_pte idx = gen8_insert_pte(start);
+ struct gen8_insert_pte idx = gen8_insert_pte(vma->node.start);
while (gen8_ppgtt_insert_pte_entries(ppgtt, pdps[idx.pml4e++], &iter,
&idx, cache_level))
@@ -1646,13 +1643,12 @@ static void gen6_ppgtt_clear_range(struct i915_address_space *vm,
}
static void gen6_ppgtt_insert_entries(struct i915_address_space *vm,
- struct sg_table *pages,
- u64 start,
+ struct i915_vma *vma,
enum i915_cache_level cache_level,
u32 flags)
{
struct i915_hw_ppgtt *ppgtt = i915_vm_to_ppgtt(vm);
- unsigned first_entry = start >> PAGE_SHIFT;
+ unsigned first_entry = vma->node.start >> PAGE_SHIFT;
unsigned act_pt = first_entry / GEN6_PTES;
unsigned act_pte = first_entry % GEN6_PTES;
const u32 pte_encode = vm->pte_encode(0, cache_level, flags);
@@ -1660,7 +1656,7 @@ static void gen6_ppgtt_insert_entries(struct i915_address_space *vm,
gen6_pte_t *vaddr;
vaddr = kmap_atomic_px(ppgtt->pd.page_table[act_pt]);
- iter.sg = pages->sgl;
+ iter.sg = vma->pages->sgl;
iter.dma = sg_dma_address(iter.sg);
iter.max = iter.dma + iter.sg->length;
do {
@@ -2117,8 +2113,7 @@ static void gen8_ggtt_insert_page(struct i915_address_space *vm,
}
static void gen8_ggtt_insert_entries(struct i915_address_space *vm,
- struct sg_table *st,
- u64 start,
+ struct i915_vma *vma,
enum i915_cache_level level,
u32 unused)
{
@@ -2129,8 +2124,8 @@ static void gen8_ggtt_insert_entries(struct i915_address_space *vm,
dma_addr_t addr;
gtt_entries = (gen8_pte_t __iomem *)ggtt->gsm;
- gtt_entries += start >> PAGE_SHIFT;
- for_each_sgt_dma(addr, sgt_iter, st)
+ gtt_entries += vma->node.start >> PAGE_SHIFT;
+ for_each_sgt_dma(addr, sgt_iter, vma->pages)
gen8_set_pte(gtt_entries++, pte_encode | addr);
wmb();
@@ -2164,17 +2159,16 @@ static void gen6_ggtt_insert_page(struct i915_address_space *vm,
* mapped BAR (dev_priv->mm.gtt->gtt).
*/
static void gen6_ggtt_insert_entries(struct i915_address_space *vm,
- struct sg_table *st,
- u64 start,
+ struct i915_vma *vma,
enum i915_cache_level level,
u32 flags)
{
struct i915_ggtt *ggtt = i915_vm_to_ggtt(vm);
gen6_pte_t __iomem *entries = (gen6_pte_t __iomem *)ggtt->gsm;
- unsigned int i = start >> PAGE_SHIFT;
+ unsigned int i = vma->node.start >> PAGE_SHIFT;
struct sgt_iter iter;
dma_addr_t addr;
- for_each_sgt_dma(addr, iter, st)
+ for_each_sgt_dma(addr, iter, vma->pages)
iowrite32(vm->pte_encode(addr, level, flags), &entries[i++]);
wmb();
@@ -2256,8 +2250,7 @@ static void bxt_vtd_ggtt_insert_page__BKL(struct i915_address_space *vm,
struct insert_entries {
struct i915_address_space *vm;
- struct sg_table *st;
- u64 start;
+ struct i915_vma *vma;
enum i915_cache_level level;
};
@@ -2265,19 +2258,18 @@ static int bxt_vtd_ggtt_insert_entries__cb(void *_arg)
{
struct insert_entries *arg = _arg;
- gen8_ggtt_insert_entries(arg->vm, arg->st, arg->start, arg->level, 0);
+ gen8_ggtt_insert_entries(arg->vm, arg->vma, arg->level, 0);
bxt_vtd_ggtt_wa(arg->vm);
return 0;
}
static void bxt_vtd_ggtt_insert_entries__BKL(struct i915_address_space *vm,
- struct sg_table *st,
- u64 start,
+ struct i915_vma *vma,
enum i915_cache_level level,
u32 unused)
{
- struct insert_entries arg = { vm, st, start, level };
+ struct insert_entries arg = { vma->vm, vma, level };
stop_machine(bxt_vtd_ggtt_insert_entries__cb, &arg, NULL);
}
@@ -2343,15 +2335,15 @@ static void i915_ggtt_insert_page(struct i915_address_space *vm,
}
static void i915_ggtt_insert_entries(struct i915_address_space *vm,
- struct sg_table *pages,
- u64 start,
+ struct i915_vma *vma,
enum i915_cache_level cache_level,
u32 unused)
{
unsigned int flags = (cache_level == I915_CACHE_NONE) ?
AGP_USER_MEMORY : AGP_USER_CACHED_MEMORY;
- intel_gtt_insert_sg_entries(pages, start >> PAGE_SHIFT, flags);
+ intel_gtt_insert_sg_entries(vma->pages, vma->node.start >> PAGE_SHIFT,
+ flags);
}
static void i915_ggtt_clear_range(struct i915_address_space *vm,
@@ -2374,8 +2366,7 @@ static int ggtt_bind_vma(struct i915_vma *vma,
pte_flags |= PTE_READ_ONLY;
intel_runtime_pm_get(i915);
- vma->vm->insert_entries(vma->vm, vma->pages, vma->node.start,
- cache_level, pte_flags);
+ vma->vm->insert_entries(vma->vm, vma, cache_level, pte_flags);
intel_runtime_pm_put(i915);
/*
@@ -2422,16 +2413,13 @@ static int aliasing_gtt_bind_vma(struct i915_vma *vma,
return ret;
}
- appgtt->base.insert_entries(&appgtt->base,
- vma->pages, vma->node.start,
- cache_level, pte_flags);
+ appgtt->base.insert_entries(&appgtt->base, vma, cache_level,
+ pte_flags);
}
if (flags & I915_VMA_GLOBAL_BIND) {
intel_runtime_pm_get(i915);
- vma->vm->insert_entries(vma->vm,
- vma->pages, vma->node.start,
- cache_level, pte_flags);
+ vma->vm->insert_entries(vma->vm, vma, cache_level, pte_flags);
intel_runtime_pm_put(i915);
}
diff --git a/drivers/gpu/drm/i915/i915_gem_gtt.h b/drivers/gpu/drm/i915/i915_gem_gtt.h
index 49902f64a896..57738a61ea6e 100644
--- a/drivers/gpu/drm/i915/i915_gem_gtt.h
+++ b/drivers/gpu/drm/i915/i915_gem_gtt.h
@@ -319,8 +319,7 @@ struct i915_address_space {
enum i915_cache_level cache_level,
u32 flags);
void (*insert_entries)(struct i915_address_space *vm,
- struct sg_table *st,
- u64 start,
+ struct i915_vma *vma,
enum i915_cache_level cache_level,
u32 flags);
void (*cleanup)(struct i915_address_space *vm);
diff --git a/drivers/gpu/drm/i915/selftests/i915_gem_gtt.c b/drivers/gpu/drm/i915/selftests/i915_gem_gtt.c
index 74fdb84f6843..0e1ded4239f9 100644
--- a/drivers/gpu/drm/i915/selftests/i915_gem_gtt.c
+++ b/drivers/gpu/drm/i915/selftests/i915_gem_gtt.c
@@ -198,6 +198,9 @@ static int lowlevel_hole(struct drm_i915_private *i915,
{
I915_RND_STATE(seed_prng);
unsigned int size;
+ struct i915_vma mock_vma;
+
+ memset(&mock_vma, 0, sizeof(struct i915_vma));
/* Keep creating larger objects until one cannot fit into the hole */
for (size = 12; (hole_end - hole_start) >> size; size++) {
@@ -256,8 +259,11 @@ static int lowlevel_hole(struct drm_i915_private *i915,
vm->allocate_va_range(vm, addr, BIT_ULL(size)))
break;
- vm->insert_entries(vm, obj->mm.pages, addr,
- I915_CACHE_NONE, 0);
+ mock_vma.pages = obj->mm.pages;
+ mock_vma.node.size = BIT_ULL(size);
+ mock_vma.node.start = addr;
+
+ vm->insert_entries(vm, &mock_vma, I915_CACHE_NONE, 0);
}
count = n;
diff --git a/drivers/gpu/drm/i915/selftests/mock_gtt.c b/drivers/gpu/drm/i915/selftests/mock_gtt.c
index a61309c7cb3e..f2118cf535a0 100644
--- a/drivers/gpu/drm/i915/selftests/mock_gtt.c
+++ b/drivers/gpu/drm/i915/selftests/mock_gtt.c
@@ -33,8 +33,7 @@ static void mock_insert_page(struct i915_address_space *vm,
}
static void mock_insert_entries(struct i915_address_space *vm,
- struct sg_table *st,
- u64 start,
+ struct i915_vma *vma,
enum i915_cache_level level, u32 flags)
{
}
--
2.9.4
More information about the Intel-gfx-trybot
mailing list