All of lore.kernel.org
 help / color / mirror / Atom feed
From: Matthew Auld <matthew.auld@intel.com>
To: intel-gfx@lists.freedesktop.org
Cc: dri-devel@lists.freedesktop.org,
	"Thomas Hellström" <thomas.hellstrom@linux.intel.com>,
	"Christian König" <christian.koenig@amd.com>
Subject: [PATCH v4 10/14] drm/i915/ttm: hide shmem objects from TTM LRU
Date: Tue, 21 Sep 2021 12:01:17 +0100	[thread overview]
Message-ID: <20210921110121.3783395-10-matthew.auld@intel.com> (raw)
In-Reply-To: <20210921110121.3783395-1-matthew.auld@intel.com>

This is probably a NAK. But ideally we need to somehow prevent TTM from
seeing shmem objects when doing its LRU swap walk. Since these are
EXTERNAL they are ignored anyway, but keeping them in the LRU seems
pretty wasteful.  Trying to use bo_pin() for this is all kinds of nasty
since we need to be able to do the bo_unpin() from the unpopulate hook,
but since that can be called from the BO destroy path we will likely go
down in flames.

An alternative is to maybe just add EXTERNAL objects to some
bdev->external LRU in TTM, or just don't add them at all?

Signed-off-by: Matthew Auld <matthew.auld@intel.com>
Cc: Thomas Hellström <thomas.hellstrom@linux.intel.com>
Cc: Christian König <christian.koenig@amd.com>
---
 drivers/gpu/drm/i915/gem/i915_gem_ttm.c | 17 +++++++++++++++++
 1 file changed, 17 insertions(+)

diff --git a/drivers/gpu/drm/i915/gem/i915_gem_ttm.c b/drivers/gpu/drm/i915/gem/i915_gem_ttm.c
index 174aebe11264..b438ddb52764 100644
--- a/drivers/gpu/drm/i915/gem/i915_gem_ttm.c
+++ b/drivers/gpu/drm/i915/gem/i915_gem_ttm.c
@@ -800,6 +800,22 @@ static unsigned long i915_ttm_io_mem_pfn(struct ttm_buffer_object *bo,
 	return ((base + sg_dma_address(sg)) >> PAGE_SHIFT) + ofs;
 }
 
+static void i915_ttm_del_from_lru_notify(struct ttm_buffer_object *bo)
+{
+	struct i915_ttm_tt *i915_tt =
+		container_of(bo->ttm, typeof(*i915_tt), ttm);
+
+	/* Idealy we need to prevent TTM from seeing shmem objects when doing
+	 * its LRU swap walk. Since these are EXTERNAL they are ignored anyway,
+	 * but keeping them in the LRU is pretty waseful. Trying to use bo_pin()
+	 * for this is very nasty since we need to be able to do the bo_unpin()
+	 * from the unpopulate hook, but since that can be called from the BO
+	 * destroy path we will go down in flames.
+	 */
+	if (bo->ttm && ttm_tt_is_populated(bo->ttm) && i915_tt->is_shmem)
+		list_del_init(&bo->lru);
+}
+
 static struct ttm_device_funcs i915_ttm_bo_driver = {
 	.ttm_tt_create = i915_ttm_tt_create,
 	.ttm_tt_populate = i915_ttm_tt_populate,
@@ -810,6 +826,7 @@ static struct ttm_device_funcs i915_ttm_bo_driver = {
 	.move = i915_ttm_move,
 	.swap_notify = i915_ttm_swap_notify,
 	.delete_mem_notify = i915_ttm_delete_mem_notify,
+	.del_from_lru_notify = i915_ttm_del_from_lru_notify,
 	.io_mem_reserve = i915_ttm_io_mem_reserve,
 	.io_mem_pfn = i915_ttm_io_mem_pfn,
 };
-- 
2.26.3


WARNING: multiple messages have this Message-ID (diff)
From: Matthew Auld <matthew.auld@intel.com>
To: intel-gfx@lists.freedesktop.org
Cc: dri-devel@lists.freedesktop.org,
	"Thomas Hellström" <thomas.hellstrom@linux.intel.com>,
	"Christian König" <christian.koenig@amd.com>
Subject: [Intel-gfx] [PATCH v4 10/14] drm/i915/ttm: hide shmem objects from TTM LRU
Date: Tue, 21 Sep 2021 12:01:17 +0100	[thread overview]
Message-ID: <20210921110121.3783395-10-matthew.auld@intel.com> (raw)
In-Reply-To: <20210921110121.3783395-1-matthew.auld@intel.com>

This is probably a NAK. But ideally we need to somehow prevent TTM from
seeing shmem objects when doing its LRU swap walk. Since these are
EXTERNAL they are ignored anyway, but keeping them in the LRU seems
pretty wasteful.  Trying to use bo_pin() for this is all kinds of nasty
since we need to be able to do the bo_unpin() from the unpopulate hook,
but since that can be called from the BO destroy path we will likely go
down in flames.

An alternative is to maybe just add EXTERNAL objects to some
bdev->external LRU in TTM, or just don't add them at all?

Signed-off-by: Matthew Auld <matthew.auld@intel.com>
Cc: Thomas Hellström <thomas.hellstrom@linux.intel.com>
Cc: Christian König <christian.koenig@amd.com>
---
 drivers/gpu/drm/i915/gem/i915_gem_ttm.c | 17 +++++++++++++++++
 1 file changed, 17 insertions(+)

diff --git a/drivers/gpu/drm/i915/gem/i915_gem_ttm.c b/drivers/gpu/drm/i915/gem/i915_gem_ttm.c
index 174aebe11264..b438ddb52764 100644
--- a/drivers/gpu/drm/i915/gem/i915_gem_ttm.c
+++ b/drivers/gpu/drm/i915/gem/i915_gem_ttm.c
@@ -800,6 +800,22 @@ static unsigned long i915_ttm_io_mem_pfn(struct ttm_buffer_object *bo,
 	return ((base + sg_dma_address(sg)) >> PAGE_SHIFT) + ofs;
 }
 
+static void i915_ttm_del_from_lru_notify(struct ttm_buffer_object *bo)
+{
+	struct i915_ttm_tt *i915_tt =
+		container_of(bo->ttm, typeof(*i915_tt), ttm);
+
+	/* Idealy we need to prevent TTM from seeing shmem objects when doing
+	 * its LRU swap walk. Since these are EXTERNAL they are ignored anyway,
+	 * but keeping them in the LRU is pretty waseful. Trying to use bo_pin()
+	 * for this is very nasty since we need to be able to do the bo_unpin()
+	 * from the unpopulate hook, but since that can be called from the BO
+	 * destroy path we will go down in flames.
+	 */
+	if (bo->ttm && ttm_tt_is_populated(bo->ttm) && i915_tt->is_shmem)
+		list_del_init(&bo->lru);
+}
+
 static struct ttm_device_funcs i915_ttm_bo_driver = {
 	.ttm_tt_create = i915_ttm_tt_create,
 	.ttm_tt_populate = i915_ttm_tt_populate,
@@ -810,6 +826,7 @@ static struct ttm_device_funcs i915_ttm_bo_driver = {
 	.move = i915_ttm_move,
 	.swap_notify = i915_ttm_swap_notify,
 	.delete_mem_notify = i915_ttm_delete_mem_notify,
+	.del_from_lru_notify = i915_ttm_del_from_lru_notify,
 	.io_mem_reserve = i915_ttm_io_mem_reserve,
 	.io_mem_pfn = i915_ttm_io_mem_pfn,
 };
-- 
2.26.3


  parent reply	other threads:[~2021-09-21 11:02 UTC|newest]

Thread overview: 52+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-09-21 11:01 [PATCH v4 01/14] drm/ttm: stop calling tt_swapin in vm_access Matthew Auld
2021-09-21 11:01 ` [Intel-gfx] " Matthew Auld
2021-09-21 11:01 ` [PATCH v4 02/14] drm/ttm: stop setting page->index for the ttm_tt Matthew Auld
2021-09-21 11:01   ` [Intel-gfx] " Matthew Auld
2021-09-21 11:01 ` [PATCH v4 03/14] drm/ttm: move ttm_tt_{add, clear}_mapping into amdgpu Matthew Auld
2021-09-21 11:01   ` [Intel-gfx] " Matthew Auld
2021-09-21 11:29   ` [PATCH v4 03/14] drm/ttm: move ttm_tt_{add,clear}_mapping " Christian König
2021-09-21 11:29     ` [Intel-gfx] [PATCH v4 03/14] drm/ttm: move ttm_tt_{add, clear}_mapping " Christian König
2021-09-21 11:01 ` [PATCH v4 04/14] drm/ttm: remove TTM_PAGE_FLAG_NO_RETRY Matthew Auld
2021-09-21 11:01   ` [Intel-gfx] " Matthew Auld
2021-09-21 11:01 ` [PATCH v4 05/14] drm/ttm: s/FLAG_SG/FLAG_EXTERNAL/ Matthew Auld
2021-09-21 11:01   ` [Intel-gfx] " Matthew Auld
2021-09-21 11:30   ` Christian König
2021-09-21 11:30     ` [Intel-gfx] " Christian König
2021-09-21 11:01 ` [PATCH v4 06/14] drm/ttm: add some kernel-doc for TTM_TT_FLAG_* Matthew Auld
2021-09-21 11:01   ` [Intel-gfx] " Matthew Auld
2021-09-21 11:33   ` Christian König
2021-09-21 11:33     ` [Intel-gfx] " Christian König
2021-09-21 11:01 ` [PATCH v4 07/14] drm/ttm: add TTM_TT_FLAG_EXTERNAL_MAPPABLE Matthew Auld
2021-09-21 11:01   ` [Intel-gfx] " Matthew Auld
2021-09-21 11:01 ` [PATCH v4 08/14] drm/i915/gem: Break out some shmem backend utils Matthew Auld
2021-09-21 11:01   ` [Intel-gfx] " Matthew Auld
2021-09-21 11:01 ` [PATCH v4 09/14] drm/i915/ttm: add tt shmem backend Matthew Auld
2021-09-21 11:01   ` [Intel-gfx] " Matthew Auld
2021-09-22 12:14   ` Thomas Hellström
2021-09-22 12:14     ` [Intel-gfx] " Thomas Hellström
2021-09-21 11:01 ` Matthew Auld [this message]
2021-09-21 11:01   ` [Intel-gfx] [PATCH v4 10/14] drm/i915/ttm: hide shmem objects from TTM LRU Matthew Auld
2021-09-21 11:48   ` Christian König
2021-09-21 11:48     ` [Intel-gfx] " Christian König
2021-09-22 13:34     ` Matthew Auld
2021-09-22 13:34       ` [Intel-gfx] " Matthew Auld
2021-09-22 13:59       ` Christian König
2021-09-22 13:59         ` [Intel-gfx] " Christian König
2021-09-21 11:01 ` [PATCH v4 11/14] drm/i915/ttm: use cached system pages when evicting lmem Matthew Auld
2021-09-21 11:01   ` [Intel-gfx] " Matthew Auld
2021-09-21 11:01 ` [PATCH v4 12/14] drm/i915: try to simplify make_{un}shrinkable Matthew Auld
2021-09-21 11:01   ` [Intel-gfx] " Matthew Auld
2021-09-21 11:01 ` [PATCH v4 13/14] drm/i915/ttm: make evicted shmem pages visible to the shrinker Matthew Auld
2021-09-21 11:01   ` [Intel-gfx] " Matthew Auld
2021-09-21 11:01 ` [PATCH v4 14/14] drm/i915/ttm: enable shmem tt backend Matthew Auld
2021-09-21 11:01   ` [Intel-gfx] " Matthew Auld
2021-09-21 11:28 ` [PATCH v4 01/14] drm/ttm: stop calling tt_swapin in vm_access Christian König
2021-09-21 11:28   ` [Intel-gfx] " Christian König
2021-09-21 11:37   ` Thomas Hellström
2021-09-21 11:37     ` [Intel-gfx] " Thomas Hellström
2021-09-21 11:42     ` Christian König
2021-09-21 11:42       ` [Intel-gfx] " Christian König
2021-09-21 12:59 ` [Intel-gfx] ✗ Fi.CI.CHECKPATCH: warning for series starting with [v4,01/14] " Patchwork
2021-09-21 13:01 ` [Intel-gfx] ✗ Fi.CI.SPARSE: " Patchwork
2021-09-21 13:30 ` [Intel-gfx] ✓ Fi.CI.BAT: success " Patchwork
2021-09-21 16:03 ` [Intel-gfx] ✓ Fi.CI.IGT: " Patchwork

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=20210921110121.3783395-10-matthew.auld@intel.com \
    --to=matthew.auld@intel.com \
    --cc=christian.koenig@amd.com \
    --cc=dri-devel@lists.freedesktop.org \
    --cc=intel-gfx@lists.freedesktop.org \
    --cc=thomas.hellstrom@linux.intel.com \
    /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.