All of lore.kernel.org
 help / color / mirror / Atom feed
From: Matthew Auld <matthew.auld@intel.com>
To: intel-gfx@lists.freedesktop.org
Subject: [PATCH 07/18] drm/i915: introduce ppgtt page coloring
Date: Tue,  4 Apr 2017 23:11:17 +0100	[thread overview]
Message-ID: <20170404221128.3943-8-matthew.auld@intel.com> (raw)
In-Reply-To: <20170404221128.3943-1-matthew.auld@intel.com>

To enable 64K pages we need to set the intermediate-page-size(IPS) bit
of the pde, therefore a page table is said to be either operating in 64K
or 4K mode. To accommodate this vm placement restriction we introduce a
color for pages and corresponding color_adjust callback.

Signed-off-by: Matthew Auld <matthew.auld@intel.com>
---
 drivers/gpu/drm/i915/i915_gem_gtt.c | 25 +++++++++++++++++++++++++
 drivers/gpu/drm/i915/i915_gem_gtt.h |  6 ++++++
 drivers/gpu/drm/i915/i915_vma.c     |  2 ++
 3 files changed, 33 insertions(+)

diff --git a/drivers/gpu/drm/i915/i915_gem_gtt.c b/drivers/gpu/drm/i915/i915_gem_gtt.c
index 0989af4a17e4..ddc3db345b76 100644
--- a/drivers/gpu/drm/i915/i915_gem_gtt.c
+++ b/drivers/gpu/drm/i915/i915_gem_gtt.c
@@ -1332,6 +1332,28 @@ static int gen8_preallocate_top_level_pdp(struct i915_hw_ppgtt *ppgtt)
 	return -ENOMEM;
 }
 
+static void i915_page_color_adjust(const struct drm_mm_node *node,
+				   unsigned long color,
+				   u64 *start,
+				   u64 *end)
+{
+	GEM_BUG_ON(!is_valid_gtt_page_size(color));
+
+	if (!(color & (I915_GTT_PAGE_SIZE_4K | I915_GTT_PAGE_SIZE_64K)))
+		return;
+
+	GEM_BUG_ON(node->allocated && !is_valid_gtt_page_size(node->color));
+
+	if (i915_node_color_differs(node, color))
+		*start = roundup(*start, 1 << GEN8_PDE_SHIFT);
+
+	node = list_next_entry(node, node_list);
+	if (i915_node_color_differs(node, color))
+		*end = rounddown(*end, 1 << GEN8_PDE_SHIFT);
+
+	GEM_BUG_ON(node->allocated && !is_valid_gtt_page_size(node->color));
+}
+
 /*
  * GEN8 legacy ppgtt programming is accomplished through a max 4 PDP registers
  * with a net effect resembling a 2-level page table in normal x86 terms. Each
@@ -1372,6 +1394,9 @@ static int gen8_ppgtt_init(struct i915_hw_ppgtt *ppgtt)
 		ppgtt->base.allocate_va_range = gen8_ppgtt_alloc_4lvl;
 		ppgtt->base.insert_entries = gen8_ppgtt_insert_4lvl;
 		ppgtt->base.clear_range = gen8_ppgtt_clear_4lvl;
+
+		if (SUPPORTS_PAGE_SIZE(dev_priv, I915_GTT_PAGE_SIZE_64K))
+			ppgtt->base.mm.color_adjust = i915_page_color_adjust;
 	} else {
 		ret = __pdp_init(&ppgtt->base, &ppgtt->pdp);
 		if (ret)
diff --git a/drivers/gpu/drm/i915/i915_gem_gtt.h b/drivers/gpu/drm/i915/i915_gem_gtt.h
index 9c592e2de516..8d893ddd98f2 100644
--- a/drivers/gpu/drm/i915/i915_gem_gtt.h
+++ b/drivers/gpu/drm/i915/i915_gem_gtt.h
@@ -353,6 +353,12 @@ i915_vm_has_cache_coloring(const struct i915_address_space *vm)
 }
 
 static inline bool
+i915_vm_has_page_coloring(const struct i915_address_space *vm)
+{
+	return vm->mm.color_adjust && !i915_is_ggtt(vm);
+}
+
+static inline bool
 i915_vm_is_48bit(const struct i915_address_space *vm)
 {
 	return (vm->total - 1) >> 32;
diff --git a/drivers/gpu/drm/i915/i915_vma.c b/drivers/gpu/drm/i915/i915_vma.c
index 8f0041ba328f..4043145b4310 100644
--- a/drivers/gpu/drm/i915/i915_vma.c
+++ b/drivers/gpu/drm/i915/i915_vma.c
@@ -471,6 +471,8 @@ i915_vma_insert(struct i915_vma *vma, u64 size, u64 alignment, u64 flags)
 
 	if (i915_vm_has_cache_coloring(vma->vm))
 		color = obj->cache_level;
+	else if (i915_vm_has_page_coloring(vma->vm))
+		color = obj->gtt_page_size;
 
 	if (flags & PIN_OFFSET_FIXED) {
 		u64 offset = flags & PIN_OFFSET_MASK;
-- 
2.9.3

_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/intel-gfx

  parent reply	other threads:[~2017-04-04 22:11 UTC|newest]

Thread overview: 42+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2017-04-04 22:11 [RFC PATCH 00/18] drm/i915: initial support for huge gtt pages V2 Matthew Auld
2017-04-04 22:11 ` [PATCH 01/18] drm/i915: add page_size_mask to dev_info Matthew Auld
2017-04-05  6:19   ` Joonas Lahtinen
2017-04-05  8:45     ` Chris Wilson
2017-04-05 12:57       ` Joonas Lahtinen
2017-04-05  8:43   ` Chris Wilson
2017-04-04 22:11 ` [PATCH 02/18] drm/i915: introduce drm_i915_gem_object page_size members Matthew Auld
2017-04-05  6:26   ` Joonas Lahtinen
2017-04-05  6:49   ` Daniel Vetter
2017-04-05  8:48     ` Chris Wilson
2017-04-05 10:07       ` Matthew Auld
2017-04-05 12:15         ` Daniel Vetter
2017-04-05 12:32         ` Chris Wilson
2017-04-05 12:39           ` Chris Wilson
2017-04-04 22:11 ` [PATCH 03/18] drm/i915: pass page_size to insert_entries Matthew Auld
2017-04-04 22:11 ` [PATCH 04/18] drm/i915: s/i915_gtt_color_adjust/i915_ggtt_color_adjust Matthew Auld
2017-04-05  6:30   ` Joonas Lahtinen
2017-04-04 22:11 ` [PATCH 05/18] drm/i915: clean up cache coloring Matthew Auld
2017-04-05  6:35   ` Joonas Lahtinen
2017-04-04 22:11 ` [PATCH 06/18] drm/i915: export color_differs Matthew Auld
2017-04-05  6:39   ` Joonas Lahtinen
2017-04-04 22:11 ` Matthew Auld [this message]
2017-04-05 13:41   ` [PATCH 07/18] drm/i915: introduce ppgtt page coloring Chris Wilson
2017-04-05 13:50     ` Matthew Auld
2017-04-05 14:02       ` Chris Wilson
2017-04-05 15:05         ` Matthew Auld
2017-04-10 12:08         ` Matthew Auld
2017-04-04 22:11 ` [PATCH 08/18] drm/i915: handle evict-for-node with " Matthew Auld
2017-04-04 22:11 ` [PATCH 09/18] drm/i915: support inserting 64K pages in the ppgtt Matthew Auld
2017-04-06  3:25   ` kbuild test robot
2017-04-09  0:27   ` kbuild test robot
2017-04-04 22:11 ` [PATCH 10/18] drm/i915: support inserting 2M " Matthew Auld
2017-04-04 22:11 ` [PATCH 11/18] drm/i915: support inserting 1G " Matthew Auld
2017-04-04 22:11 ` [PATCH 12/18] drm/i915: disable GTT cache for huge-pages Matthew Auld
2017-04-04 22:11 ` [PATCH 13/18] drm/i915/selftests: exercise 4K and 64K mm insertion Matthew Auld
2017-04-04 22:11 ` [PATCH 14/18] drm/i915/selftests: modify the gtt tests to also exercise huge pages Matthew Auld
2017-04-04 22:11 ` [PATCH 15/18] drm/i915/selftests: exercise evict-for-node page coloring Matthew Auld
2017-04-04 22:11 ` [PATCH 16/18] drm/i915/debugfs: include some huge-page metrics Matthew Auld
2017-04-04 22:11 ` [PATCH 17/18] mm/shmem: tweak the huge-page interface Matthew Auld
2017-04-05  6:42   ` Daniel Vetter
2017-04-04 22:11 ` [PATCH 18/18] drm/i915: support transparent-huge-pages through shmemfs Matthew Auld
2017-04-05  8:53 ` [RFC PATCH 00/18] drm/i915: initial support for huge gtt pages V2 Chris Wilson

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20170404221128.3943-8-matthew.auld@intel.com \
    --to=matthew.auld@intel.com \
    --cc=intel-gfx@lists.freedesktop.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.