* [PATCH v4] gpu/buddy: restore two-pass clear/dirty guard for range-biased allocation
@ 2026-11-04 7:28 Arunpravin Paneer Selvam
2026-11-04 7:46 ` sashiko-bot
` (3 more replies)
0 siblings, 4 replies; 7+ messages in thread
From: Arunpravin Paneer Selvam @ 2026-11-04 7:28 UTC (permalink / raw)
To: matthew.auld, dri-devel, intel-gfx, intel-xe, amd-gfx
Cc: christian.koenig, alexander.deucher, Anand.Raghavendra,
Arunpravin Paneer Selvam
From: Arunpravin Paneer Selvam <Arunpravin.PaneerSelvam@amd.com>
Range-restricted allocations (the visible-VRAM aperture) skip
dirty_steer_window() and reach __alloc_range_bias() with no clear/dirty
eligibility guard, so a dirty allocation freely splits clear extents (and
vice versa), interleaving the two classes until no large same-class run
survives and a 4K scanout pin fails with -ENOSPC. Use a two-pass range
search instead: block_incompatible() rejects opposite-class blocks in the
strict pass and relaxes only when the strict pass returns -ENOSPC, so a
request still succeeds when only opposite-class space remains. A KUnit
regression test covers the interleaving pattern and fails without the
guard.
v2: the strict pass can split a block and then reject both children, so
a split no longer implies a compatible block exists. Sashiko reported
the resulting leaked splits; fix as suggested by Matthew:
- Push the parent back on the DFS stack after each split and merge it
back with unsplit_block() if both children are still free.
- Make block_incompatible() judge free blocks only (split nodes keep a
stale clear bit) and take the requested order, so mixed blocks that
can hold a compatible block are descended instead of skipped.
- Extend the KUnit test with a coalesced mixed block and a check that
a rejected split is merged back.
v3: igt@kms_plane@plane-panning-bottom-right failed on Radeon RX 9070 XT
with -ENOSPC: the range allocator's strict pass split mixed blocks to
look for a wholly-clear (or wholly-dirty) leaf, discarded the mixed
leaves it found, and ran unsplit_block to merge the splits back, whose
per-allocation cost and residual fragmentation starved the large
pipe-A scanout. Fix: accept a mixed block at the requested order as-is
(caller re-clears), so the first pass no longer splits and discards
mixed blocks. This clear/dirty compatibility relaxation applies only
to range-biased allocation, where keeping blocks intact reduces
fragmentation and preserves large contiguous spaces. The now-unneeded
unsplit_block helper is dropped.
v4: a higher-order free mixed block straddling the range edge was split in
the strict pass then abandoned, leaving both children free under a split
parent that never coalesces, tripping gpu_buddy_fini()'s free-root assert
and leaking the children (reported by Sashiko). Fix: in the strict pass
don't split a free block that only partially overlaps the range; the
fallback pass consumes it (either class accepted).
Fixes: 1ad5e807f716 ("gpu/buddy: replace dual-tree/force_merge with decoupled dirty tracker")
Assisted-by: LLM
Cc: Matthew Auld <matthew.auld@intel.com>
Cc: Christian König <christian.koenig@amd.com>
Signed-off-by: Arunpravin Paneer Selvam <Arunpravin.PaneerSelvam@amd.com>
---
drivers/gpu/buddy.c | 83 ++++++++++++++++-------------
drivers/gpu/tests/gpu_buddy_test.c | 84 ++++++++++++++++++++++++++++++
2 files changed, 130 insertions(+), 37 deletions(-)
diff --git a/drivers/gpu/buddy.c b/drivers/gpu/buddy.c
index 28cff97d57dc..d9749cbaf05d 100644
--- a/drivers/gpu/buddy.c
+++ b/drivers/gpu/buddy.c
@@ -1240,11 +1240,23 @@ static void __gpu_buddy_undo_splits(struct gpu_buddy *mm,
}
}
+static bool block_incompatible(struct gpu_buddy_block *block, unsigned int flags)
+{
+ bool needs_clear = flags & GPU_BUDDY_CLEAR_ALLOCATION;
+
+ if (!gpu_buddy_block_is_free(block))
+ return false;
+
+ return needs_clear ? !block->has_clear :
+ gpu_buddy_block_is_clear(block);
+}
+
static struct gpu_buddy_block *
__alloc_range_bias(struct gpu_buddy *mm,
u64 start, u64 end,
unsigned int order,
- unsigned long flags)
+ unsigned long flags,
+ bool fallback)
{
u64 req_size = mm->chunk_size << order;
struct gpu_buddy_block *block;
@@ -1258,9 +1270,7 @@ __alloc_range_bias(struct gpu_buddy *mm,
* This range-constrained search hands back the highest/right-most
* address that satisfies the request: the roots are seeded high-to-low
* and the right (higher-address) child is descended first, making
- * top-down the default placement here. A non-top-down clear request is
- * the only exception, where the descent is biased towards clear or
- * clear-containing subtrees to satisfy the clear preference.
+ * top-down the default placement here.
*/
for (i = mm->n_roots - 1; i >= 0; --i)
list_add_tail(&mm->roots[i]->tmp_link, &dfs);
@@ -1296,8 +1306,19 @@ __alloc_range_bias(struct gpu_buddy *mm,
if (round_down(adjusted_end + 1, req_size) <=
round_up(adjusted_start, req_size))
continue;
+
+ /*
+ * Strict pass: don't split a partially-overlapping free
+ * block; rejecting its child would strand the split. Let
+ * the fallback pass consume it (either class accepted).
+ */
+ if (!fallback && gpu_buddy_block_is_free(block))
+ continue;
}
+ if (!fallback && block_incompatible(block, flags))
+ continue;
+
if (contains(start, end, block_start, block_end) &&
order == gpu_buddy_block_order(block)) {
/*
@@ -1315,38 +1336,8 @@ __alloc_range_bias(struct gpu_buddy *mm,
goto err_undo;
}
- /*
- * Top-down is a strict address-placement policy, so when it is
- * requested we ignore clear steering and simply descend the
- * right (higher-address) child first. Only a non-top-down clear
- * request biases the descent towards clear/has_clear subtrees.
- */
- if ((flags & GPU_BUDDY_CLEAR_ALLOCATION) &&
- !(flags & GPU_BUDDY_TOPDOWN_ALLOCATION)) {
- struct gpu_buddy_block *prefer;
-
- if (gpu_buddy_block_is_clear(block->right))
- prefer = block->right;
- else if (gpu_buddy_block_is_clear(block->left))
- prefer = block->left;
- else if (block->right->has_clear)
- prefer = block->right;
- else if (block->left->has_clear)
- prefer = block->left;
- else
- prefer = block->right;
-
- if (prefer == block->right) {
- list_add(&block->left->tmp_link, &dfs);
- list_add(&block->right->tmp_link, &dfs);
- } else {
- list_add(&block->right->tmp_link, &dfs);
- list_add(&block->left->tmp_link, &dfs);
- }
- } else {
- list_add(&block->left->tmp_link, &dfs);
- list_add(&block->right->tmp_link, &dfs);
- }
+ list_add(&block->left->tmp_link, &dfs);
+ list_add(&block->right->tmp_link, &dfs);
} while (1);
return ERR_PTR(-ENOSPC);
@@ -1361,6 +1352,24 @@ __alloc_range_bias(struct gpu_buddy *mm,
return ERR_PTR(err);
}
+static struct gpu_buddy_block *
+__gpu_buddy_alloc_range_bias(struct gpu_buddy *mm,
+ u64 start, u64 end,
+ unsigned int order,
+ unsigned long flags)
+{
+ struct gpu_buddy_block *block;
+ bool fallback = false;
+
+ block = __alloc_range_bias(mm, start, end, order,
+ flags, fallback);
+ if (IS_ERR(block))
+ return __alloc_range_bias(mm, start, end, order,
+ flags, !fallback);
+
+ return block;
+}
+
/* Return the highest-address free block of at least @order. */
static struct gpu_buddy_block *
get_maxblock(struct gpu_buddy *mm,
@@ -1924,7 +1933,7 @@ __gpu_buddy_alloc_blocks(struct gpu_buddy *mm,
if (flags & GPU_BUDDY_RANGE_ALLOCATION) {
/* Allocate traversing within the range */
- block = __alloc_range_bias(mm, start, end, order, flags);
+ block = __gpu_buddy_alloc_range_bias(mm, start, end, order, flags);
if (!IS_ERR(block) || !steered)
return block;
diff --git a/drivers/gpu/tests/gpu_buddy_test.c b/drivers/gpu/tests/gpu_buddy_test.c
index 2c6f1da1a6d1..9aaca2fc6a02 100644
--- a/drivers/gpu/tests/gpu_buddy_test.c
+++ b/drivers/gpu/tests/gpu_buddy_test.c
@@ -1606,6 +1606,89 @@ static int gpu_buddy_suite_init(struct kunit_suite *suite)
return 0;
}
+static void gpu_test_buddy_alloc_range_clear_guard(struct kunit *test)
+{
+ const unsigned long chunk = SZ_1M, mm_size = SZ_4M;
+ struct gpu_buddy_block *block;
+ LIST_HEAD(low_dirty);
+ LIST_HEAD(high_a);
+ LIST_HEAD(high_b);
+ LIST_HEAD(barrier);
+ LIST_HEAD(dirty_res);
+ LIST_HEAD(clear_res);
+ struct gpu_buddy mm;
+
+ KUNIT_ASSERT_FALSE_MSG(test, gpu_buddy_init(&mm, mm_size, chunk),
+ "buddy_init failed\n");
+
+ /*
+ * Idea is to lay out a range allocation as a low dirty block, an
+ * allocated barrier, and a high naturally-aligned clear extent, then
+ * issue a dirty range allocation. It must take the low dirty block and
+ * leave the clear extent intact; if it instead cuts the clear extent,
+ * a following same-class clear allocation can no longer find a
+ * contiguous clear run and fails with -ENOSPC.
+ *
+ * [0, 1M) dirty free
+ * [1M, 2M) allocated (barrier, keeps the halves un-mergeable)
+ * [2M, 4M) clear free (one naturally aligned 2 MiB extent)
+ */
+ KUNIT_ASSERT_FALSE_MSG(test,
+ gpu_buddy_alloc_blocks(&mm, 0, chunk, chunk, chunk,
+ &low_dirty, GPU_BUDDY_RANGE_ALLOCATION),
+ "alloc [0, 1M) failed\n");
+ KUNIT_ASSERT_FALSE_MSG(test,
+ gpu_buddy_alloc_blocks(&mm, chunk, 2 * chunk, chunk, chunk,
+ &barrier, GPU_BUDDY_RANGE_ALLOCATION),
+ "alloc [1M, 2M) failed\n");
+ KUNIT_ASSERT_FALSE_MSG(test,
+ gpu_buddy_alloc_blocks(&mm, 2 * chunk, 3 * chunk, chunk, chunk,
+ &high_a, GPU_BUDDY_RANGE_ALLOCATION),
+ "alloc [2M, 3M) failed\n");
+ KUNIT_ASSERT_FALSE_MSG(test,
+ gpu_buddy_alloc_blocks(&mm, 3 * chunk, 4 * chunk, chunk, chunk,
+ &high_b, GPU_BUDDY_RANGE_ALLOCATION),
+ "alloc [3M, 4M) failed\n");
+
+ /* [2M, 4M) coalesces into one aligned clear extent; [0, 1M) stays dirty. */
+ gpu_buddy_free_list(&mm, &low_dirty, 0);
+ gpu_buddy_free_list(&mm, &high_a, GPU_BUDDY_CLEARED);
+ gpu_buddy_free_list(&mm, &high_b, GPU_BUDDY_CLEARED);
+
+ /* Dirty 1 MiB range alloc: must land at [0, 1M), not cut the clear extent. */
+ KUNIT_ASSERT_FALSE_MSG(test,
+ gpu_buddy_alloc_blocks(&mm, 0, mm_size, chunk, chunk,
+ &dirty_res, GPU_BUDDY_RANGE_ALLOCATION),
+ "dirty range alloc failed\n");
+ block = list_first_entry(&dirty_res, struct gpu_buddy_block, link);
+ KUNIT_EXPECT_FALSE_MSG(test, gpu_buddy_block_is_clear(block),
+ "dirty range alloc cut the clear extent (offset=%llx)\n",
+ gpu_buddy_block_offset(block));
+ KUNIT_EXPECT_EQ(test, gpu_buddy_block_offset(block), 0ULL);
+
+ /*
+ * The 2 MiB clear extent must still be intact: a clear order-1 range
+ * allocation has to succeed. Without the guard the extent was cut
+ * above and this returns -ENOSPC.
+ */
+ KUNIT_EXPECT_FALSE_MSG(test,
+ gpu_buddy_alloc_blocks(&mm, 0, mm_size, 2 * chunk, 2 * chunk,
+ &clear_res,
+ GPU_BUDDY_RANGE_ALLOCATION |
+ GPU_BUDDY_CLEAR_ALLOCATION),
+ "clear 2M range alloc failed: clear extent was fragmented (-ENOSPC)\n");
+ if (!list_empty(&clear_res)) {
+ block = list_first_entry(&clear_res, struct gpu_buddy_block, link);
+ KUNIT_EXPECT_TRUE(test, gpu_buddy_block_is_clear(block));
+ KUNIT_EXPECT_EQ(test, gpu_buddy_block_offset(block), (u64)(2 * chunk));
+ }
+
+ gpu_buddy_free_list(&mm, &clear_res, GPU_BUDDY_CLEARED);
+ gpu_buddy_free_list(&mm, &dirty_res, 0);
+ gpu_buddy_free_list(&mm, &barrier, 0);
+ gpu_buddy_fini(&mm);
+}
+
static struct kunit_case gpu_buddy_tests[] = {
KUNIT_CASE(gpu_test_buddy_alloc_limit),
KUNIT_CASE(gpu_test_buddy_alloc_optimistic),
@@ -1615,6 +1698,7 @@ static struct kunit_case gpu_buddy_tests[] = {
KUNIT_CASE(gpu_test_buddy_alloc_clear),
KUNIT_CASE(gpu_test_buddy_alloc_range),
KUNIT_CASE(gpu_test_buddy_alloc_range_bias),
+ KUNIT_CASE(gpu_test_buddy_alloc_range_clear_guard),
KUNIT_CASE_SLOW(gpu_test_buddy_fragmentation_performance),
KUNIT_CASE_SLOW(gpu_test_buddy_dirty_tracker_performance),
KUNIT_CASE(gpu_test_buddy_alloc_exceeds_max_order),
base-commit: 49f40f15d20f5b4073d517da7efbcea079f1b9e3
--
2.43.0
^ permalink raw reply related [flat|nested] 7+ messages in thread
* Re: [PATCH v4] gpu/buddy: restore two-pass clear/dirty guard for range-biased allocation
2026-11-04 7:28 [PATCH v4] gpu/buddy: restore two-pass clear/dirty guard for range-biased allocation Arunpravin Paneer Selvam
@ 2026-11-04 7:46 ` sashiko-bot
2026-11-04 8:19 ` ✓ i915.CI.BAT: success for gpu/buddy: restore two-pass clear/dirty guard for range-biased allocation (rev2) Patchwork
` (2 subsequent siblings)
3 siblings, 0 replies; 7+ messages in thread
From: sashiko-bot @ 2026-11-04 7:46 UTC (permalink / raw)
To: Arunpravin Paneer Selvam, dri-devel, intel-gfx, intel-xe
Thank you for your contribution! Sashiko AI review found 2 potential issue(s) to consider:
- [Medium] gpu/buddy: strict pass in __alloc_range_bias() skips straddling free blocks
- [Medium] gpu/buddy: __alloc_range_bias() drops clear child preference over mixed blocks
--
commit 39aa2fd8b0782a6f08d3234bf343e92cab53a75f
Author: Arunpravin Paneer Selvam <Arunpravin.PaneerSelvam@amd.com>
gpu/buddy: restore two-pass clear/dirty guard for range-biased allocation
This commit adds a two-pass clear/dirty compatibility guard to
__alloc_range_bias() so range-constrained allocations avoid splitting
opposite-class extents unless no compatible space exists in the range. It
also adds a KUnit regression test covering range-biased clear/dirty
isolation.
> diff --git a/drivers/gpu/buddy.c b/drivers/gpu/buddy.c
> index 28cff97d57dc8..d9749cbaf05d6 100644
> --- a/drivers/gpu/buddy.c
> +++ b/drivers/gpu/buddy.c
> @@ -1240,11 +1240,23 @@ static void __gpu_buddy_undo_splits(struct gpu_buddy *mm,
> }
> }
>
> +static bool block_incompatible(struct gpu_buddy_block *block, unsigned int flags)
> +{
> + bool needs_clear = flags & GPU_BUDDY_CLEAR_ALLOCATION;
> +
> + if (!gpu_buddy_block_is_free(block))
> + return false;
> +
> + return needs_clear ? !block->has_clear :
> + gpu_buddy_block_is_clear(block);
> +}
> +
> static struct gpu_buddy_block *
> __alloc_range_bias(struct gpu_buddy *mm,
> u64 start, u64 end,
> unsigned int order,
> - unsigned long flags)
> + unsigned long flags,
> + bool fallback)
> {
[ ... ]
> @@ -1296,8 +1306,19 @@ __alloc_range_bias(struct gpu_buddy *mm,
> if (round_down(adjusted_end + 1, req_size) <=
> round_up(adjusted_start, req_size))
> continue;
> +
> + /*
> + * Strict pass: don't split a partially-overlapping free
> + * block; rejecting its child would strand the split. Let
> + * the fallback pass consume it (either class accepted).
> + */
> + if (!fallback && gpu_buddy_block_is_free(block))
> + continue;
> }
>
> + if (!fallback && block_incompatible(block, flags))
> + continue;
[Severity: Medium]
In gpu_buddy_alloc_blocks() -> __gpu_buddy_alloc_blocks() ->
__gpu_buddy_alloc_range_bias() -> __alloc_range_bias(), when a sub-range
[start, end) is smaller than a free root or coalesced ancestor block (so
block_start < start || block_end > end), doesn't skipping every partially
overlapping free block in the strict pass cause the strict pass to skip
descending into that block even when it contains compatible space or is
itself wholly compatible with the request?
Since __gpu_buddy_free() coalesces free buddy pairs across clear and dirty
states into mixed free blocks, a coalesced free block or a wholly compatible
free block straddling start or end will be skipped by:
if (block_start < start || block_end > end) {
...
if (!fallback && gpu_buddy_block_is_free(block))
continue;
}
When that causes the strict pass to return -ENOSPC,
__gpu_buddy_alloc_range_bias() retries with fallback = true, which disables
block_incompatible() across the entire traversal and allows splitting
higher-address opposite-class extents inside [start, end) even when
compatible blocks exist in the range.
Similarly, when dirty_steer_window() narrows [start, end) to a dirty window
inside a larger coalesced mixed free block, won't the strict pass always
skip the enclosing free block and fail with -ENOSPC before retrying in the
fallback pass?
> if (contains(start, end, block_start, block_end) &&
> order == gpu_buddy_block_order(block)) {
> /*
> @@ -1315,38 +1336,8 @@ __alloc_range_bias(struct gpu_buddy *mm,
> goto err_undo;
> }
>
> - /*
> - * Top-down is a strict address-placement policy, so when it is
> - * requested we ignore clear steering and simply descend the
> - * right (higher-address) child first. Only a non-top-down clear
> - * request biases the descent towards clear/has_clear subtrees.
> - */
> - if ((flags & GPU_BUDDY_CLEAR_ALLOCATION) &&
> - !(flags & GPU_BUDDY_TOPDOWN_ALLOCATION)) {
> - struct gpu_buddy_block *prefer;
> -
> - if (gpu_buddy_block_is_clear(block->right))
> - prefer = block->right;
> - else if (gpu_buddy_block_is_clear(block->left))
> - prefer = block->left;
> - else if (block->right->has_clear)
> - prefer = block->right;
> - else if (block->left->has_clear)
> - prefer = block->left;
> - else
> - prefer = block->right;
> -
> - if (prefer == block->right) {
> - list_add(&block->left->tmp_link, &dfs);
> - list_add(&block->right->tmp_link, &dfs);
> - } else {
> - list_add(&block->right->tmp_link, &dfs);
> - list_add(&block->left->tmp_link, &dfs);
> - }
> - } else {
> - list_add(&block->left->tmp_link, &dfs);
> - list_add(&block->right->tmp_link, &dfs);
> - }
> + list_add(&block->left->tmp_link, &dfs);
> + list_add(&block->right->tmp_link, &dfs);
> } while (1);
[Severity: Medium]
In __alloc_range_bias(), block_incompatible() treats both wholly clear
blocks and mixed blocks (where block->has_clear is true but
gpu_buddy_block_is_clear(block) is false) as compatible when
GPU_BUDDY_CLEAR_ALLOCATION is set.
With the child preference check removed here so block->right is always
visited before block->left, if a split parent has a mixed right child and a
wholly clear left child at the target order, won't a non-top-down
GPU_BUDDY_CLEAR_ALLOCATION request now select the mixed right child (with
GPU_BUDDY_HEADER_CLEAR unset, requiring the caller to re-clear it) while
leaving the wholly clear left sibling unused?
--
Sashiko AI review · https://sashiko.dev/#/patchset/20261104072847.2617886-1-arunpravin.paneerselvam@amd.com?part=1
^ permalink raw reply [flat|nested] 7+ messages in thread
* ✓ i915.CI.BAT: success for gpu/buddy: restore two-pass clear/dirty guard for range-biased allocation (rev2)
2026-11-04 7:28 [PATCH v4] gpu/buddy: restore two-pass clear/dirty guard for range-biased allocation Arunpravin Paneer Selvam
2026-11-04 7:46 ` sashiko-bot
@ 2026-11-04 8:19 ` Patchwork
2026-11-06 11:06 ` [PATCH v4] gpu/buddy: restore two-pass clear/dirty guard for range-biased allocation Arunpravin Paneer Selvam
2026-11-10 16:25 ` Matthew Auld
3 siblings, 0 replies; 7+ messages in thread
From: Patchwork @ 2026-11-04 8:19 UTC (permalink / raw)
To: Arunpravin Paneer Selvam; +Cc: intel-gfx
[-- Attachment #1: Type: text/plain, Size: 2827 bytes --]
== Series Details ==
Series: gpu/buddy: restore two-pass clear/dirty guard for range-biased allocation (rev2)
URL : https://patchwork.freedesktop.org/series/177601/
State : success
== Summary ==
CI Bug Log - changes from CI_DRM_19443 -> Patchwork_177601v2
====================================================
Summary
-------
**SUCCESS**
No regressions found.
External URL: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_177601v2/index.html
Participating hosts (40 -> 39)
------------------------------
Missing (1): bat-dg2-13
Known issues
------------
Here are the changes found in Patchwork_177601v2 that come from known issues:
### IGT changes ###
#### Possible fixes ####
* igt@i915_selftest@live:
- bat-arlh-3: [INCOMPLETE][1] ([i915#16547]) -> [PASS][2]
[1]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_19443/bat-arlh-3/igt@i915_selftest@live.html
[2]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_177601v2/bat-arlh-3/igt@i915_selftest@live.html
* igt@i915_selftest@live@late_gt_pm:
- fi-cfl-8109u: [DMESG-WARN][3] ([i915#13735]) -> [PASS][4] +80 other tests pass
[3]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_19443/fi-cfl-8109u/igt@i915_selftest@live@late_gt_pm.html
[4]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_177601v2/fi-cfl-8109u/igt@i915_selftest@live@late_gt_pm.html
* igt@i915_selftest@live@slpc:
- bat-arlh-3: [INCOMPLETE][5] ([i915#17769]) -> [PASS][6]
[5]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_19443/bat-arlh-3/igt@i915_selftest@live@slpc.html
[6]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_177601v2/bat-arlh-3/igt@i915_selftest@live@slpc.html
* igt@kms_pipe_crc_basic@read-crc:
- fi-cfl-8109u: [DMESG-WARN][7] ([i915#13735] / [i915#15673]) -> [PASS][8] +49 other tests pass
[7]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_19443/fi-cfl-8109u/igt@kms_pipe_crc_basic@read-crc.html
[8]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_177601v2/fi-cfl-8109u/igt@kms_pipe_crc_basic@read-crc.html
[i915#13735]: https://gitlab.freedesktop.org/drm/i915/kernel/-/issues/13735
[i915#15673]: https://gitlab.freedesktop.org/drm/i915/kernel/-/issues/15673
[i915#16547]: https://gitlab.freedesktop.org/drm/i915/kernel/-/issues/16547
[i915#17769]: https://gitlab.freedesktop.org/drm/i915/kernel/-/issues/17769
Build changes
-------------
* Linux: CI_DRM_19443 -> Patchwork_177601v2
CI-20190529: 20190529
CI_DRM_19443: d11bbfdac0f3d443b367857ffe683ae6bb80d560 @ git://anongit.freedesktop.org/gfx-ci/linux
IGT_9176: 9176
Patchwork_177601v2: d11bbfdac0f3d443b367857ffe683ae6bb80d560 @ git://anongit.freedesktop.org/gfx-ci/linux
== Logs ==
For more details see: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_177601v2/index.html
[-- Attachment #2: Type: text/html, Size: 3579 bytes --]
^ permalink raw reply [flat|nested] 7+ messages in thread
* Re: [PATCH v4] gpu/buddy: restore two-pass clear/dirty guard for range-biased allocation
2026-11-04 7:28 [PATCH v4] gpu/buddy: restore two-pass clear/dirty guard for range-biased allocation Arunpravin Paneer Selvam
2026-11-04 7:46 ` sashiko-bot
2026-11-04 8:19 ` ✓ i915.CI.BAT: success for gpu/buddy: restore two-pass clear/dirty guard for range-biased allocation (rev2) Patchwork
@ 2026-11-06 11:06 ` Arunpravin Paneer Selvam
2026-11-10 4:17 ` Arunpravin Paneer Selvam
2026-11-10 16:25 ` Matthew Auld
3 siblings, 1 reply; 7+ messages in thread
From: Arunpravin Paneer Selvam @ 2026-11-06 11:06 UTC (permalink / raw)
To: matthew.auld, dri-devel, intel-gfx, intel-xe, amd-gfx
Cc: christian.koenig, alexander.deucher, Anand.Raghavendra
Hi Matthew,
Both Sashiko findings on v4 are intentional - no code change:
1. Strict pass skips straddling free blocks: This is intended. If the
strict pass split a free block that crosses the range edge, part of it
would fall outside the range and get rejected, leaving the split behind
- and the strict pass has no way to merge it back. That was the v3 leak,
so we now skip such blocks instead of splitting them.
The clear/dirty check ( block_incompatible() ) only looks at blocks that
fit entirely inside the range, so it keeps clear and dirty apart
correctly whenever such blocks exist. It is skipped in just one case:
when the only option is a single big free block crossing the range edge.
That happens when the area is still unused - nothing is allocated in it
yet - so there are no existing clear or dirty allocations there that
could get interleaved or broken up. The guard has nothing to protect in
that case, so skipping it does no harm. At worst the fallback pass picks
a block of the other class and it has to be cleared again - a re-clear
cost, not fragmentation.
2. Dropped clear-child preference for mixed blocks: On purpose. Carrying
that preference down through mixed blocks splits them to carve out a
same-class leaf, which interleaves allocated clear and dirty blocks and
regresses range allocations to -ENOSPC . Taking the mixed child whole
strands nothing, and the wholly-clear sibling stays free and usable, so
the only cost is a possible re-clear.
Why we don't steer has_clear splits in mixed blocks:
Old dual-tree kept clear and dirty in separate pools (same-class-only
merge), so every split came from a pure pool and never stranded the
opposite class. The new design merges across classes into large MIXED
blocks, so class-aware steering had to keep splitting those big blocks
to carve out a same-class leaf - breaking contiguous space and leaving
allocations of the two classes mixed together, so freed blocks can't
reform a large run → -ENOSPC . v4 drops steering: it still splits a
mixed block that's larger than the request (for size only, spatially, no
class bias) but takes the block whole once it reaches the requested
order instead of splitting further to isolate a same-class leaf. Using
only the block_incompatible guard - which accepts or rejects a whole
block by class and never traverses inside a mixed block to pick a
clear/dirty child - it keeps large blocks intact and avoids the
interleaving.
Could you please check and review the v4.
Thanks,
Arun.
On 11/4/2026 12:58 PM, Arunpravin Paneer Selvam wrote:
> From: Arunpravin Paneer Selvam <Arunpravin.PaneerSelvam@amd.com>
>
> Range-restricted allocations (the visible-VRAM aperture) skip
> dirty_steer_window() and reach __alloc_range_bias() with no clear/dirty
> eligibility guard, so a dirty allocation freely splits clear extents (and
> vice versa), interleaving the two classes until no large same-class run
> survives and a 4K scanout pin fails with -ENOSPC. Use a two-pass range
> search instead: block_incompatible() rejects opposite-class blocks in the
> strict pass and relaxes only when the strict pass returns -ENOSPC, so a
> request still succeeds when only opposite-class space remains. A KUnit
> regression test covers the interleaving pattern and fails without the
> guard.
>
> v2: the strict pass can split a block and then reject both children, so
> a split no longer implies a compatible block exists. Sashiko reported
> the resulting leaked splits; fix as suggested by Matthew:
> - Push the parent back on the DFS stack after each split and merge it
> back with unsplit_block() if both children are still free.
> - Make block_incompatible() judge free blocks only (split nodes keep a
> stale clear bit) and take the requested order, so mixed blocks that
> can hold a compatible block are descended instead of skipped.
> - Extend the KUnit test with a coalesced mixed block and a check that
> a rejected split is merged back.
>
> v3: igt@kms_plane@plane-panning-bottom-right failed on Radeon RX 9070 XT
> with -ENOSPC: the range allocator's strict pass split mixed blocks to
> look for a wholly-clear (or wholly-dirty) leaf, discarded the mixed
> leaves it found, and ran unsplit_block to merge the splits back, whose
> per-allocation cost and residual fragmentation starved the large
> pipe-A scanout. Fix: accept a mixed block at the requested order as-is
> (caller re-clears), so the first pass no longer splits and discards
> mixed blocks. This clear/dirty compatibility relaxation applies only
> to range-biased allocation, where keeping blocks intact reduces
> fragmentation and preserves large contiguous spaces. The now-unneeded
> unsplit_block helper is dropped.
>
> v4: a higher-order free mixed block straddling the range edge was split in
> the strict pass then abandoned, leaving both children free under a split
> parent that never coalesces, tripping gpu_buddy_fini()'s free-root assert
> and leaking the children (reported by Sashiko). Fix: in the strict pass
> don't split a free block that only partially overlaps the range; the
> fallback pass consumes it (either class accepted).
>
> Fixes: 1ad5e807f716 ("gpu/buddy: replace dual-tree/force_merge with decoupled dirty tracker")
> Assisted-by: LLM
> Cc: Matthew Auld <matthew.auld@intel.com>
> Cc: Christian König <christian.koenig@amd.com>
> Signed-off-by: Arunpravin Paneer Selvam <Arunpravin.PaneerSelvam@amd.com>
> ---
> drivers/gpu/buddy.c | 83 ++++++++++++++++-------------
> drivers/gpu/tests/gpu_buddy_test.c | 84 ++++++++++++++++++++++++++++++
> 2 files changed, 130 insertions(+), 37 deletions(-)
>
> diff --git a/drivers/gpu/buddy.c b/drivers/gpu/buddy.c
> index 28cff97d57dc..d9749cbaf05d 100644
> --- a/drivers/gpu/buddy.c
> +++ b/drivers/gpu/buddy.c
> @@ -1240,11 +1240,23 @@ static void __gpu_buddy_undo_splits(struct gpu_buddy *mm,
> }
> }
>
> +static bool block_incompatible(struct gpu_buddy_block *block, unsigned int flags)
> +{
> + bool needs_clear = flags & GPU_BUDDY_CLEAR_ALLOCATION;
> +
> + if (!gpu_buddy_block_is_free(block))
> + return false;
> +
> + return needs_clear ? !block->has_clear :
> + gpu_buddy_block_is_clear(block);
> +}
> +
> static struct gpu_buddy_block *
> __alloc_range_bias(struct gpu_buddy *mm,
> u64 start, u64 end,
> unsigned int order,
> - unsigned long flags)
> + unsigned long flags,
> + bool fallback)
> {
> u64 req_size = mm->chunk_size << order;
> struct gpu_buddy_block *block;
> @@ -1258,9 +1270,7 @@ __alloc_range_bias(struct gpu_buddy *mm,
> * This range-constrained search hands back the highest/right-most
> * address that satisfies the request: the roots are seeded high-to-low
> * and the right (higher-address) child is descended first, making
> - * top-down the default placement here. A non-top-down clear request is
> - * the only exception, where the descent is biased towards clear or
> - * clear-containing subtrees to satisfy the clear preference.
> + * top-down the default placement here.
> */
> for (i = mm->n_roots - 1; i >= 0; --i)
> list_add_tail(&mm->roots[i]->tmp_link, &dfs);
> @@ -1296,8 +1306,19 @@ __alloc_range_bias(struct gpu_buddy *mm,
> if (round_down(adjusted_end + 1, req_size) <=
> round_up(adjusted_start, req_size))
> continue;
> +
> + /*
> + * Strict pass: don't split a partially-overlapping free
> + * block; rejecting its child would strand the split. Let
> + * the fallback pass consume it (either class accepted).
> + */
> + if (!fallback && gpu_buddy_block_is_free(block))
> + continue;
> }
>
> + if (!fallback && block_incompatible(block, flags))
> + continue;
> +
> if (contains(start, end, block_start, block_end) &&
> order == gpu_buddy_block_order(block)) {
> /*
> @@ -1315,38 +1336,8 @@ __alloc_range_bias(struct gpu_buddy *mm,
> goto err_undo;
> }
>
> - /*
> - * Top-down is a strict address-placement policy, so when it is
> - * requested we ignore clear steering and simply descend the
> - * right (higher-address) child first. Only a non-top-down clear
> - * request biases the descent towards clear/has_clear subtrees.
> - */
> - if ((flags & GPU_BUDDY_CLEAR_ALLOCATION) &&
> - !(flags & GPU_BUDDY_TOPDOWN_ALLOCATION)) {
> - struct gpu_buddy_block *prefer;
> -
> - if (gpu_buddy_block_is_clear(block->right))
> - prefer = block->right;
> - else if (gpu_buddy_block_is_clear(block->left))
> - prefer = block->left;
> - else if (block->right->has_clear)
> - prefer = block->right;
> - else if (block->left->has_clear)
> - prefer = block->left;
> - else
> - prefer = block->right;
> -
> - if (prefer == block->right) {
> - list_add(&block->left->tmp_link, &dfs);
> - list_add(&block->right->tmp_link, &dfs);
> - } else {
> - list_add(&block->right->tmp_link, &dfs);
> - list_add(&block->left->tmp_link, &dfs);
> - }
> - } else {
> - list_add(&block->left->tmp_link, &dfs);
> - list_add(&block->right->tmp_link, &dfs);
> - }
> + list_add(&block->left->tmp_link, &dfs);
> + list_add(&block->right->tmp_link, &dfs);
> } while (1);
>
> return ERR_PTR(-ENOSPC);
> @@ -1361,6 +1352,24 @@ __alloc_range_bias(struct gpu_buddy *mm,
> return ERR_PTR(err);
> }
>
> +static struct gpu_buddy_block *
> +__gpu_buddy_alloc_range_bias(struct gpu_buddy *mm,
> + u64 start, u64 end,
> + unsigned int order,
> + unsigned long flags)
> +{
> + struct gpu_buddy_block *block;
> + bool fallback = false;
> +
> + block = __alloc_range_bias(mm, start, end, order,
> + flags, fallback);
> + if (IS_ERR(block))
> + return __alloc_range_bias(mm, start, end, order,
> + flags, !fallback);
> +
> + return block;
> +}
> +
> /* Return the highest-address free block of at least @order. */
> static struct gpu_buddy_block *
> get_maxblock(struct gpu_buddy *mm,
> @@ -1924,7 +1933,7 @@ __gpu_buddy_alloc_blocks(struct gpu_buddy *mm,
>
> if (flags & GPU_BUDDY_RANGE_ALLOCATION) {
> /* Allocate traversing within the range */
> - block = __alloc_range_bias(mm, start, end, order, flags);
> + block = __gpu_buddy_alloc_range_bias(mm, start, end, order, flags);
> if (!IS_ERR(block) || !steered)
> return block;
>
> diff --git a/drivers/gpu/tests/gpu_buddy_test.c b/drivers/gpu/tests/gpu_buddy_test.c
> index 2c6f1da1a6d1..9aaca2fc6a02 100644
> --- a/drivers/gpu/tests/gpu_buddy_test.c
> +++ b/drivers/gpu/tests/gpu_buddy_test.c
> @@ -1606,6 +1606,89 @@ static int gpu_buddy_suite_init(struct kunit_suite *suite)
> return 0;
> }
>
> +static void gpu_test_buddy_alloc_range_clear_guard(struct kunit *test)
> +{
> + const unsigned long chunk = SZ_1M, mm_size = SZ_4M;
> + struct gpu_buddy_block *block;
> + LIST_HEAD(low_dirty);
> + LIST_HEAD(high_a);
> + LIST_HEAD(high_b);
> + LIST_HEAD(barrier);
> + LIST_HEAD(dirty_res);
> + LIST_HEAD(clear_res);
> + struct gpu_buddy mm;
> +
> + KUNIT_ASSERT_FALSE_MSG(test, gpu_buddy_init(&mm, mm_size, chunk),
> + "buddy_init failed\n");
> +
> + /*
> + * Idea is to lay out a range allocation as a low dirty block, an
> + * allocated barrier, and a high naturally-aligned clear extent, then
> + * issue a dirty range allocation. It must take the low dirty block and
> + * leave the clear extent intact; if it instead cuts the clear extent,
> + * a following same-class clear allocation can no longer find a
> + * contiguous clear run and fails with -ENOSPC.
> + *
> + * [0, 1M) dirty free
> + * [1M, 2M) allocated (barrier, keeps the halves un-mergeable)
> + * [2M, 4M) clear free (one naturally aligned 2 MiB extent)
> + */
> + KUNIT_ASSERT_FALSE_MSG(test,
> + gpu_buddy_alloc_blocks(&mm, 0, chunk, chunk, chunk,
> + &low_dirty, GPU_BUDDY_RANGE_ALLOCATION),
> + "alloc [0, 1M) failed\n");
> + KUNIT_ASSERT_FALSE_MSG(test,
> + gpu_buddy_alloc_blocks(&mm, chunk, 2 * chunk, chunk, chunk,
> + &barrier, GPU_BUDDY_RANGE_ALLOCATION),
> + "alloc [1M, 2M) failed\n");
> + KUNIT_ASSERT_FALSE_MSG(test,
> + gpu_buddy_alloc_blocks(&mm, 2 * chunk, 3 * chunk, chunk, chunk,
> + &high_a, GPU_BUDDY_RANGE_ALLOCATION),
> + "alloc [2M, 3M) failed\n");
> + KUNIT_ASSERT_FALSE_MSG(test,
> + gpu_buddy_alloc_blocks(&mm, 3 * chunk, 4 * chunk, chunk, chunk,
> + &high_b, GPU_BUDDY_RANGE_ALLOCATION),
> + "alloc [3M, 4M) failed\n");
> +
> + /* [2M, 4M) coalesces into one aligned clear extent; [0, 1M) stays dirty. */
> + gpu_buddy_free_list(&mm, &low_dirty, 0);
> + gpu_buddy_free_list(&mm, &high_a, GPU_BUDDY_CLEARED);
> + gpu_buddy_free_list(&mm, &high_b, GPU_BUDDY_CLEARED);
> +
> + /* Dirty 1 MiB range alloc: must land at [0, 1M), not cut the clear extent. */
> + KUNIT_ASSERT_FALSE_MSG(test,
> + gpu_buddy_alloc_blocks(&mm, 0, mm_size, chunk, chunk,
> + &dirty_res, GPU_BUDDY_RANGE_ALLOCATION),
> + "dirty range alloc failed\n");
> + block = list_first_entry(&dirty_res, struct gpu_buddy_block, link);
> + KUNIT_EXPECT_FALSE_MSG(test, gpu_buddy_block_is_clear(block),
> + "dirty range alloc cut the clear extent (offset=%llx)\n",
> + gpu_buddy_block_offset(block));
> + KUNIT_EXPECT_EQ(test, gpu_buddy_block_offset(block), 0ULL);
> +
> + /*
> + * The 2 MiB clear extent must still be intact: a clear order-1 range
> + * allocation has to succeed. Without the guard the extent was cut
> + * above and this returns -ENOSPC.
> + */
> + KUNIT_EXPECT_FALSE_MSG(test,
> + gpu_buddy_alloc_blocks(&mm, 0, mm_size, 2 * chunk, 2 * chunk,
> + &clear_res,
> + GPU_BUDDY_RANGE_ALLOCATION |
> + GPU_BUDDY_CLEAR_ALLOCATION),
> + "clear 2M range alloc failed: clear extent was fragmented (-ENOSPC)\n");
> + if (!list_empty(&clear_res)) {
> + block = list_first_entry(&clear_res, struct gpu_buddy_block, link);
> + KUNIT_EXPECT_TRUE(test, gpu_buddy_block_is_clear(block));
> + KUNIT_EXPECT_EQ(test, gpu_buddy_block_offset(block), (u64)(2 * chunk));
> + }
> +
> + gpu_buddy_free_list(&mm, &clear_res, GPU_BUDDY_CLEARED);
> + gpu_buddy_free_list(&mm, &dirty_res, 0);
> + gpu_buddy_free_list(&mm, &barrier, 0);
> + gpu_buddy_fini(&mm);
> +}
> +
> static struct kunit_case gpu_buddy_tests[] = {
> KUNIT_CASE(gpu_test_buddy_alloc_limit),
> KUNIT_CASE(gpu_test_buddy_alloc_optimistic),
> @@ -1615,6 +1698,7 @@ static struct kunit_case gpu_buddy_tests[] = {
> KUNIT_CASE(gpu_test_buddy_alloc_clear),
> KUNIT_CASE(gpu_test_buddy_alloc_range),
> KUNIT_CASE(gpu_test_buddy_alloc_range_bias),
> + KUNIT_CASE(gpu_test_buddy_alloc_range_clear_guard),
> KUNIT_CASE_SLOW(gpu_test_buddy_fragmentation_performance),
> KUNIT_CASE_SLOW(gpu_test_buddy_dirty_tracker_performance),
> KUNIT_CASE(gpu_test_buddy_alloc_exceeds_max_order),
>
> base-commit: 49f40f15d20f5b4073d517da7efbcea079f1b9e3
^ permalink raw reply [flat|nested] 7+ messages in thread
* Re: [PATCH v4] gpu/buddy: restore two-pass clear/dirty guard for range-biased allocation
2026-11-06 11:06 ` [PATCH v4] gpu/buddy: restore two-pass clear/dirty guard for range-biased allocation Arunpravin Paneer Selvam
@ 2026-11-10 4:17 ` Arunpravin Paneer Selvam
0 siblings, 0 replies; 7+ messages in thread
From: Arunpravin Paneer Selvam @ 2026-11-10 4:17 UTC (permalink / raw)
To: matthew.auld, dri-devel, intel-gfx, intel-xe, amd-gfx
Cc: christian.koenig, alexander.deucher, Anand.Raghavendra
Hi Matthew,
Do you have any objections to this fix? I think we can push this first,
as it fixes the range-allocation path issue introduced by the base
patch, “gpu/buddy: replace dual-tree/force_merge with decoupled dirty
tracker.”
The work William and Sanjay are doing to improve alloc_range_bias() can
come separately. This fix is also needed for CI to complete its runs and
allows us to move these patches forward, as clients are waiting for this
version in the next release.
Thanks,
Arun.
On 11/6/2026 4:36 PM, Arunpravin Paneer Selvam wrote:
> Hi Matthew,
>
> Both Sashiko findings on v4 are intentional - no code change:
>
> 1. Strict pass skips straddling free blocks: This is intended. If the
> strict pass split a free block that crosses the range edge, part of it
> would fall outside the range and get rejected, leaving the split
> behind - and the strict pass has no way to merge it back. That was the
> v3 leak, so we now skip such blocks instead of splitting them.
>
> The clear/dirty check ( block_incompatible() ) only looks at blocks
> that fit entirely inside the range, so it keeps clear and dirty apart
> correctly whenever such blocks exist. It is skipped in just one case:
> when the only option is a single big free block crossing the range
> edge. That happens when the area is still unused - nothing is
> allocated in it yet - so there are no existing clear or dirty
> allocations there that could get interleaved or broken up. The guard
> has nothing to protect in that case, so skipping it does no harm. At
> worst the fallback pass picks a block of the other class and it has to
> be cleared again - a re-clear cost, not fragmentation.
>
> 2. Dropped clear-child preference for mixed blocks: On purpose.
> Carrying that preference down through mixed blocks splits them to
> carve out a same-class leaf, which interleaves allocated clear and
> dirty blocks and regresses range allocations to -ENOSPC . Taking the
> mixed child whole strands nothing, and the wholly-clear sibling stays
> free and usable, so the only cost is a possible re-clear.
>
>
>
> Why we don't steer has_clear splits in mixed blocks:
> Old dual-tree kept clear and dirty in separate pools (same-class-only
> merge), so every split came from a pure pool and never stranded the
> opposite class. The new design merges across classes into large MIXED
> blocks, so class-aware steering had to keep splitting those big blocks
> to carve out a same-class leaf - breaking contiguous space and leaving
> allocations of the two classes mixed together, so freed blocks can't
> reform a large run → -ENOSPC . v4 drops steering: it still splits a
> mixed block that's larger than the request (for size only, spatially,
> no class bias) but takes the block whole once it reaches the requested
> order instead of splitting further to isolate a same-class leaf. Using
> only the block_incompatible guard - which accepts or rejects a whole
> block by class and never traverses inside a mixed block to pick a
> clear/dirty child - it keeps large blocks intact and avoids the
> interleaving.
>
> Could you please check and review the v4.
>
> Thanks,
> Arun.
>
> On 11/4/2026 12:58 PM, Arunpravin Paneer Selvam wrote:
>> From: Arunpravin Paneer Selvam <Arunpravin.PaneerSelvam@amd.com>
>>
>> Range-restricted allocations (the visible-VRAM aperture) skip
>> dirty_steer_window() and reach __alloc_range_bias() with no clear/dirty
>> eligibility guard, so a dirty allocation freely splits clear extents
>> (and
>> vice versa), interleaving the two classes until no large same-class run
>> survives and a 4K scanout pin fails with -ENOSPC. Use a two-pass range
>> search instead: block_incompatible() rejects opposite-class blocks in
>> the
>> strict pass and relaxes only when the strict pass returns -ENOSPC, so a
>> request still succeeds when only opposite-class space remains. A KUnit
>> regression test covers the interleaving pattern and fails without the
>> guard.
>>
>> v2: the strict pass can split a block and then reject both children, so
>> a split no longer implies a compatible block exists. Sashiko
>> reported
>> the resulting leaked splits; fix as suggested by Matthew:
>> - Push the parent back on the DFS stack after each split and merge it
>> back with unsplit_block() if both children are still free.
>> - Make block_incompatible() judge free blocks only (split nodes keep a
>> stale clear bit) and take the requested order, so mixed blocks that
>> can hold a compatible block are descended instead of skipped.
>> - Extend the KUnit test with a coalesced mixed block and a check that
>> a rejected split is merged back.
>>
>> v3: igt@kms_plane@plane-panning-bottom-right failed on Radeon RX 9070 XT
>> with -ENOSPC: the range allocator's strict pass split mixed
>> blocks to
>> look for a wholly-clear (or wholly-dirty) leaf, discarded the mixed
>> leaves it found, and ran unsplit_block to merge the splits back,
>> whose
>> per-allocation cost and residual fragmentation starved the large
>> pipe-A scanout. Fix: accept a mixed block at the requested order
>> as-is
>> (caller re-clears), so the first pass no longer splits and discards
>> mixed blocks. This clear/dirty compatibility relaxation applies
>> only
>> to range-biased allocation, where keeping blocks intact reduces
>> fragmentation and preserves large contiguous spaces. The
>> now-unneeded
>> unsplit_block helper is dropped.
>>
>> v4: a higher-order free mixed block straddling the range edge was
>> split in
>> the strict pass then abandoned, leaving both children free under
>> a split
>> parent that never coalesces, tripping gpu_buddy_fini()'s
>> free-root assert
>> and leaking the children (reported by Sashiko). Fix: in the
>> strict pass
>> don't split a free block that only partially overlaps the range;
>> the
>> fallback pass consumes it (either class accepted).
>>
>> Fixes: 1ad5e807f716 ("gpu/buddy: replace dual-tree/force_merge with
>> decoupled dirty tracker")
>> Assisted-by: LLM
>> Cc: Matthew Auld <matthew.auld@intel.com>
>> Cc: Christian König <christian.koenig@amd.com>
>> Signed-off-by: Arunpravin Paneer Selvam
>> <Arunpravin.PaneerSelvam@amd.com>
>> ---
>> drivers/gpu/buddy.c | 83 ++++++++++++++++-------------
>> drivers/gpu/tests/gpu_buddy_test.c | 84 ++++++++++++++++++++++++++++++
>> 2 files changed, 130 insertions(+), 37 deletions(-)
>>
>> diff --git a/drivers/gpu/buddy.c b/drivers/gpu/buddy.c
>> index 28cff97d57dc..d9749cbaf05d 100644
>> --- a/drivers/gpu/buddy.c
>> +++ b/drivers/gpu/buddy.c
>> @@ -1240,11 +1240,23 @@ static void __gpu_buddy_undo_splits(struct
>> gpu_buddy *mm,
>> }
>> }
>> +static bool block_incompatible(struct gpu_buddy_block *block,
>> unsigned int flags)
>> +{
>> + bool needs_clear = flags & GPU_BUDDY_CLEAR_ALLOCATION;
>> +
>> + if (!gpu_buddy_block_is_free(block))
>> + return false;
>> +
>> + return needs_clear ? !block->has_clear :
>> + gpu_buddy_block_is_clear(block);
>> +}
>> +
>> static struct gpu_buddy_block *
>> __alloc_range_bias(struct gpu_buddy *mm,
>> u64 start, u64 end,
>> unsigned int order,
>> - unsigned long flags)
>> + unsigned long flags,
>> + bool fallback)
>> {
>> u64 req_size = mm->chunk_size << order;
>> struct gpu_buddy_block *block;
>> @@ -1258,9 +1270,7 @@ __alloc_range_bias(struct gpu_buddy *mm,
>> * This range-constrained search hands back the highest/right-most
>> * address that satisfies the request: the roots are seeded
>> high-to-low
>> * and the right (higher-address) child is descended first, making
>> - * top-down the default placement here. A non-top-down clear
>> request is
>> - * the only exception, where the descent is biased towards clear or
>> - * clear-containing subtrees to satisfy the clear preference.
>> + * top-down the default placement here.
>> */
>> for (i = mm->n_roots - 1; i >= 0; --i)
>> list_add_tail(&mm->roots[i]->tmp_link, &dfs);
>> @@ -1296,8 +1306,19 @@ __alloc_range_bias(struct gpu_buddy *mm,
>> if (round_down(adjusted_end + 1, req_size) <=
>> round_up(adjusted_start, req_size))
>> continue;
>> +
>> + /*
>> + * Strict pass: don't split a partially-overlapping free
>> + * block; rejecting its child would strand the split. Let
>> + * the fallback pass consume it (either class accepted).
>> + */
>> + if (!fallback && gpu_buddy_block_is_free(block))
>> + continue;
>> }
>> + if (!fallback && block_incompatible(block, flags))
>> + continue;
>> +
>> if (contains(start, end, block_start, block_end) &&
>> order == gpu_buddy_block_order(block)) {
>> /*
>> @@ -1315,38 +1336,8 @@ __alloc_range_bias(struct gpu_buddy *mm,
>> goto err_undo;
>> }
>> - /*
>> - * Top-down is a strict address-placement policy, so when it is
>> - * requested we ignore clear steering and simply descend the
>> - * right (higher-address) child first. Only a non-top-down
>> clear
>> - * request biases the descent towards clear/has_clear subtrees.
>> - */
>> - if ((flags & GPU_BUDDY_CLEAR_ALLOCATION) &&
>> - !(flags & GPU_BUDDY_TOPDOWN_ALLOCATION)) {
>> - struct gpu_buddy_block *prefer;
>> -
>> - if (gpu_buddy_block_is_clear(block->right))
>> - prefer = block->right;
>> - else if (gpu_buddy_block_is_clear(block->left))
>> - prefer = block->left;
>> - else if (block->right->has_clear)
>> - prefer = block->right;
>> - else if (block->left->has_clear)
>> - prefer = block->left;
>> - else
>> - prefer = block->right;
>> -
>> - if (prefer == block->right) {
>> - list_add(&block->left->tmp_link, &dfs);
>> - list_add(&block->right->tmp_link, &dfs);
>> - } else {
>> - list_add(&block->right->tmp_link, &dfs);
>> - list_add(&block->left->tmp_link, &dfs);
>> - }
>> - } else {
>> - list_add(&block->left->tmp_link, &dfs);
>> - list_add(&block->right->tmp_link, &dfs);
>> - }
>> + list_add(&block->left->tmp_link, &dfs);
>> + list_add(&block->right->tmp_link, &dfs);
>> } while (1);
>> return ERR_PTR(-ENOSPC);
>> @@ -1361,6 +1352,24 @@ __alloc_range_bias(struct gpu_buddy *mm,
>> return ERR_PTR(err);
>> }
>> +static struct gpu_buddy_block *
>> +__gpu_buddy_alloc_range_bias(struct gpu_buddy *mm,
>> + u64 start, u64 end,
>> + unsigned int order,
>> + unsigned long flags)
>> +{
>> + struct gpu_buddy_block *block;
>> + bool fallback = false;
>> +
>> + block = __alloc_range_bias(mm, start, end, order,
>> + flags, fallback);
>> + if (IS_ERR(block))
>> + return __alloc_range_bias(mm, start, end, order,
>> + flags, !fallback);
>> +
>> + return block;
>> +}
>> +
>> /* Return the highest-address free block of at least @order. */
>> static struct gpu_buddy_block *
>> get_maxblock(struct gpu_buddy *mm,
>> @@ -1924,7 +1933,7 @@ __gpu_buddy_alloc_blocks(struct gpu_buddy *mm,
>> if (flags & GPU_BUDDY_RANGE_ALLOCATION) {
>> /* Allocate traversing within the range */
>> - block = __alloc_range_bias(mm, start, end, order, flags);
>> + block = __gpu_buddy_alloc_range_bias(mm, start, end, order,
>> flags);
>> if (!IS_ERR(block) || !steered)
>> return block;
>> diff --git a/drivers/gpu/tests/gpu_buddy_test.c
>> b/drivers/gpu/tests/gpu_buddy_test.c
>> index 2c6f1da1a6d1..9aaca2fc6a02 100644
>> --- a/drivers/gpu/tests/gpu_buddy_test.c
>> +++ b/drivers/gpu/tests/gpu_buddy_test.c
>> @@ -1606,6 +1606,89 @@ static int gpu_buddy_suite_init(struct
>> kunit_suite *suite)
>> return 0;
>> }
>> +static void gpu_test_buddy_alloc_range_clear_guard(struct kunit
>> *test)
>> +{
>> + const unsigned long chunk = SZ_1M, mm_size = SZ_4M;
>> + struct gpu_buddy_block *block;
>> + LIST_HEAD(low_dirty);
>> + LIST_HEAD(high_a);
>> + LIST_HEAD(high_b);
>> + LIST_HEAD(barrier);
>> + LIST_HEAD(dirty_res);
>> + LIST_HEAD(clear_res);
>> + struct gpu_buddy mm;
>> +
>> + KUNIT_ASSERT_FALSE_MSG(test, gpu_buddy_init(&mm, mm_size, chunk),
>> + "buddy_init failed\n");
>> +
>> + /*
>> + * Idea is to lay out a range allocation as a low dirty block, an
>> + * allocated barrier, and a high naturally-aligned clear extent,
>> then
>> + * issue a dirty range allocation. It must take the low dirty
>> block and
>> + * leave the clear extent intact; if it instead cuts the clear
>> extent,
>> + * a following same-class clear allocation can no longer find a
>> + * contiguous clear run and fails with -ENOSPC.
>> + *
>> + * [0, 1M) dirty free
>> + * [1M, 2M) allocated (barrier, keeps the halves un-mergeable)
>> + * [2M, 4M) clear free (one naturally aligned 2 MiB extent)
>> + */
>> + KUNIT_ASSERT_FALSE_MSG(test,
>> + gpu_buddy_alloc_blocks(&mm, 0, chunk, chunk, chunk,
>> + &low_dirty, GPU_BUDDY_RANGE_ALLOCATION),
>> + "alloc [0, 1M) failed\n");
>> + KUNIT_ASSERT_FALSE_MSG(test,
>> + gpu_buddy_alloc_blocks(&mm, chunk, 2 * chunk,
>> chunk, chunk,
>> + &barrier, GPU_BUDDY_RANGE_ALLOCATION),
>> + "alloc [1M, 2M) failed\n");
>> + KUNIT_ASSERT_FALSE_MSG(test,
>> + gpu_buddy_alloc_blocks(&mm, 2 * chunk, 3 * chunk,
>> chunk, chunk,
>> + &high_a, GPU_BUDDY_RANGE_ALLOCATION),
>> + "alloc [2M, 3M) failed\n");
>> + KUNIT_ASSERT_FALSE_MSG(test,
>> + gpu_buddy_alloc_blocks(&mm, 3 * chunk, 4 * chunk,
>> chunk, chunk,
>> + &high_b, GPU_BUDDY_RANGE_ALLOCATION),
>> + "alloc [3M, 4M) failed\n");
>> +
>> + /* [2M, 4M) coalesces into one aligned clear extent; [0, 1M)
>> stays dirty. */
>> + gpu_buddy_free_list(&mm, &low_dirty, 0);
>> + gpu_buddy_free_list(&mm, &high_a, GPU_BUDDY_CLEARED);
>> + gpu_buddy_free_list(&mm, &high_b, GPU_BUDDY_CLEARED);
>> +
>> + /* Dirty 1 MiB range alloc: must land at [0, 1M), not cut the
>> clear extent. */
>> + KUNIT_ASSERT_FALSE_MSG(test,
>> + gpu_buddy_alloc_blocks(&mm, 0, mm_size, chunk,
>> chunk,
>> + &dirty_res, GPU_BUDDY_RANGE_ALLOCATION),
>> + "dirty range alloc failed\n");
>> + block = list_first_entry(&dirty_res, struct gpu_buddy_block, link);
>> + KUNIT_EXPECT_FALSE_MSG(test, gpu_buddy_block_is_clear(block),
>> + "dirty range alloc cut the clear extent
>> (offset=%llx)\n",
>> + gpu_buddy_block_offset(block));
>> + KUNIT_EXPECT_EQ(test, gpu_buddy_block_offset(block), 0ULL);
>> +
>> + /*
>> + * The 2 MiB clear extent must still be intact: a clear order-1
>> range
>> + * allocation has to succeed. Without the guard the extent was cut
>> + * above and this returns -ENOSPC.
>> + */
>> + KUNIT_EXPECT_FALSE_MSG(test,
>> + gpu_buddy_alloc_blocks(&mm, 0, mm_size, 2 *
>> chunk, 2 * chunk,
>> + &clear_res,
>> + GPU_BUDDY_RANGE_ALLOCATION |
>> + GPU_BUDDY_CLEAR_ALLOCATION),
>> + "clear 2M range alloc failed: clear extent was
>> fragmented (-ENOSPC)\n");
>> + if (!list_empty(&clear_res)) {
>> + block = list_first_entry(&clear_res, struct gpu_buddy_block,
>> link);
>> + KUNIT_EXPECT_TRUE(test, gpu_buddy_block_is_clear(block));
>> + KUNIT_EXPECT_EQ(test, gpu_buddy_block_offset(block), (u64)(2
>> * chunk));
>> + }
>> +
>> + gpu_buddy_free_list(&mm, &clear_res, GPU_BUDDY_CLEARED);
>> + gpu_buddy_free_list(&mm, &dirty_res, 0);
>> + gpu_buddy_free_list(&mm, &barrier, 0);
>> + gpu_buddy_fini(&mm);
>> +}
>> +
>> static struct kunit_case gpu_buddy_tests[] = {
>> KUNIT_CASE(gpu_test_buddy_alloc_limit),
>> KUNIT_CASE(gpu_test_buddy_alloc_optimistic),
>> @@ -1615,6 +1698,7 @@ static struct kunit_case gpu_buddy_tests[] = {
>> KUNIT_CASE(gpu_test_buddy_alloc_clear),
>> KUNIT_CASE(gpu_test_buddy_alloc_range),
>> KUNIT_CASE(gpu_test_buddy_alloc_range_bias),
>> + KUNIT_CASE(gpu_test_buddy_alloc_range_clear_guard),
>> KUNIT_CASE_SLOW(gpu_test_buddy_fragmentation_performance),
>> KUNIT_CASE_SLOW(gpu_test_buddy_dirty_tracker_performance),
>> KUNIT_CASE(gpu_test_buddy_alloc_exceeds_max_order),
>>
>> base-commit: 49f40f15d20f5b4073d517da7efbcea079f1b9e3
>
^ permalink raw reply [flat|nested] 7+ messages in thread
* Re: [PATCH v4] gpu/buddy: restore two-pass clear/dirty guard for range-biased allocation
2026-11-04 7:28 [PATCH v4] gpu/buddy: restore two-pass clear/dirty guard for range-biased allocation Arunpravin Paneer Selvam
` (2 preceding siblings ...)
2026-11-06 11:06 ` [PATCH v4] gpu/buddy: restore two-pass clear/dirty guard for range-biased allocation Arunpravin Paneer Selvam
@ 2026-11-10 16:25 ` Matthew Auld
2026-11-11 4:55 ` Arunpravin Paneer Selvam
3 siblings, 1 reply; 7+ messages in thread
From: Matthew Auld @ 2026-11-10 16:25 UTC (permalink / raw)
To: Arunpravin Paneer Selvam, dri-devel, intel-gfx, intel-xe, amd-gfx
Cc: christian.koenig, alexander.deucher, Anand.Raghavendra
On 04/11/2026 07:28, Arunpravin Paneer Selvam wrote:
> From: Arunpravin Paneer Selvam <Arunpravin.PaneerSelvam@amd.com>
>
> Range-restricted allocations (the visible-VRAM aperture) skip
> dirty_steer_window() and reach __alloc_range_bias() with no clear/dirty
> eligibility guard, so a dirty allocation freely splits clear extents (and
> vice versa), interleaving the two classes until no large same-class run
> survives and a 4K scanout pin fails with -ENOSPC. Use a two-pass range
> search instead: block_incompatible() rejects opposite-class blocks in the
> strict pass and relaxes only when the strict pass returns -ENOSPC, so a
> request still succeeds when only opposite-class space remains. A KUnit
> regression test covers the interleaving pattern and fails without the
> guard.
>
> v2: the strict pass can split a block and then reject both children, so
> a split no longer implies a compatible block exists. Sashiko reported
> the resulting leaked splits; fix as suggested by Matthew:
> - Push the parent back on the DFS stack after each split and merge it
> back with unsplit_block() if both children are still free.
> - Make block_incompatible() judge free blocks only (split nodes keep a
> stale clear bit) and take the requested order, so mixed blocks that
> can hold a compatible block are descended instead of skipped.
> - Extend the KUnit test with a coalesced mixed block and a check that
> a rejected split is merged back.
>
> v3: igt@kms_plane@plane-panning-bottom-right failed on Radeon RX 9070 XT
> with -ENOSPC: the range allocator's strict pass split mixed blocks to
> look for a wholly-clear (or wholly-dirty) leaf, discarded the mixed
> leaves it found, and ran unsplit_block to merge the splits back, whose
> per-allocation cost and residual fragmentation starved the large
> pipe-A scanout. Fix: accept a mixed block at the requested order as-is
> (caller re-clears), so the first pass no longer splits and discards
> mixed blocks. This clear/dirty compatibility relaxation applies only
> to range-biased allocation, where keeping blocks intact reduces
> fragmentation and preserves large contiguous spaces. The now-unneeded
> unsplit_block helper is dropped.
>
> v4: a higher-order free mixed block straddling the range edge was split in
> the strict pass then abandoned, leaving both children free under a split
> parent that never coalesces, tripping gpu_buddy_fini()'s free-root assert
> and leaking the children (reported by Sashiko). Fix: in the strict pass
> don't split a free block that only partially overlaps the range; the
> fallback pass consumes it (either class accepted).
>
> Fixes: 1ad5e807f716 ("gpu/buddy: replace dual-tree/force_merge with decoupled dirty tracker")
> Assisted-by: LLM
> Cc: Matthew Auld <matthew.auld@intel.com>
> Cc: Christian König <christian.koenig@amd.com>
> Signed-off-by: Arunpravin Paneer Selvam <Arunpravin.PaneerSelvam@amd.com>
> ---
> drivers/gpu/buddy.c | 83 ++++++++++++++++-------------
> drivers/gpu/tests/gpu_buddy_test.c | 84 ++++++++++++++++++++++++++++++
> 2 files changed, 130 insertions(+), 37 deletions(-)
>
> diff --git a/drivers/gpu/buddy.c b/drivers/gpu/buddy.c
> index 28cff97d57dc..d9749cbaf05d 100644
> --- a/drivers/gpu/buddy.c
> +++ b/drivers/gpu/buddy.c
> @@ -1240,11 +1240,23 @@ static void __gpu_buddy_undo_splits(struct gpu_buddy *mm,
> }
> }
>
> +static bool block_incompatible(struct gpu_buddy_block *block, unsigned int flags)
> +{
> + bool needs_clear = flags & GPU_BUDDY_CLEAR_ALLOCATION;
> +
> + if (!gpu_buddy_block_is_free(block))
> + return false;
> +
> + return needs_clear ? !block->has_clear :
> + gpu_buddy_block_is_clear(block);
> +}
> +
> static struct gpu_buddy_block *
> __alloc_range_bias(struct gpu_buddy *mm,
> u64 start, u64 end,
> unsigned int order,
> - unsigned long flags)
> + unsigned long flags,
> + bool fallback)
> {
> u64 req_size = mm->chunk_size << order;
> struct gpu_buddy_block *block;
> @@ -1258,9 +1270,7 @@ __alloc_range_bias(struct gpu_buddy *mm,
> * This range-constrained search hands back the highest/right-most
> * address that satisfies the request: the roots are seeded high-to-low
> * and the right (higher-address) child is descended first, making
> - * top-down the default placement here. A non-top-down clear request is
> - * the only exception, where the descent is biased towards clear or
> - * clear-containing subtrees to satisfy the clear preference.
> + * top-down the default placement here.
> */
> for (i = mm->n_roots - 1; i >= 0; --i)
> list_add_tail(&mm->roots[i]->tmp_link, &dfs);
> @@ -1296,8 +1306,19 @@ __alloc_range_bias(struct gpu_buddy *mm,
> if (round_down(adjusted_end + 1, req_size) <=
> round_up(adjusted_start, req_size))
> continue;
> +
> + /*
> + * Strict pass: don't split a partially-overlapping free
> + * block; rejecting its child would strand the split. Let
> + * the fallback pass consume it (either class accepted).
> + */
> + if (!fallback && gpu_buddy_block_is_free(block))
> + continue;
> }
>
> + if (!fallback && block_incompatible(block, flags))
> + continue;
> +
> if (contains(start, end, block_start, block_end) &&
> order == gpu_buddy_block_order(block)) {
> /*
> @@ -1315,38 +1336,8 @@ __alloc_range_bias(struct gpu_buddy *mm,
> goto err_undo;
> }
>
> - /*
> - * Top-down is a strict address-placement policy, so when it is
> - * requested we ignore clear steering and simply descend the
> - * right (higher-address) child first. Only a non-top-down clear
> - * request biases the descent towards clear/has_clear subtrees.
> - */
> - if ((flags & GPU_BUDDY_CLEAR_ALLOCATION) &&
> - !(flags & GPU_BUDDY_TOPDOWN_ALLOCATION)) {
> - struct gpu_buddy_block *prefer;
> -
> - if (gpu_buddy_block_is_clear(block->right))
> - prefer = block->right;
> - else if (gpu_buddy_block_is_clear(block->left))
> - prefer = block->left;
> - else if (block->right->has_clear)
> - prefer = block->right;
> - else if (block->left->has_clear)
> - prefer = block->left;
> - else
> - prefer = block->right;
> -
> - if (prefer == block->right) {
> - list_add(&block->left->tmp_link, &dfs);
> - list_add(&block->right->tmp_link, &dfs);
> - } else {
> - list_add(&block->right->tmp_link, &dfs);
> - list_add(&block->left->tmp_link, &dfs);
> - }
> - } else {
> - list_add(&block->left->tmp_link, &dfs);
> - list_add(&block->right->tmp_link, &dfs);
> - }
> + list_add(&block->left->tmp_link, &dfs);
> + list_add(&block->right->tmp_link, &dfs);
> } while (1);
>
> return ERR_PTR(-ENOSPC);
> @@ -1361,6 +1352,24 @@ __alloc_range_bias(struct gpu_buddy *mm,
> return ERR_PTR(err);
> }
>
> +static struct gpu_buddy_block *
> +__gpu_buddy_alloc_range_bias(struct gpu_buddy *mm,
> + u64 start, u64 end,
> + unsigned int order,
> + unsigned long flags)
> +{
> + struct gpu_buddy_block *block;
> + bool fallback = false;
Can we have fallback = true when !(gpu_buddy_clear_avail(mm) &&
mm->dirty.total_dirty) or something, and just do one pass in that case?
In xe we don't yet use clear/dirty tracking so everything is always
dirty so no point in searching twice.
Reviewed-by: Matthew Auld <matthew.auld@intel.com>
> +
> + block = __alloc_range_bias(mm, start, end, order,
> + flags, fallback);
> + if (IS_ERR(block))
> + return __alloc_range_bias(mm, start, end, order,
> + flags, !fallback);
> +
> + return block;
> +}
> +
> /* Return the highest-address free block of at least @order. */
> static struct gpu_buddy_block *
> get_maxblock(struct gpu_buddy *mm,
> @@ -1924,7 +1933,7 @@ __gpu_buddy_alloc_blocks(struct gpu_buddy *mm,
>
> if (flags & GPU_BUDDY_RANGE_ALLOCATION) {
> /* Allocate traversing within the range */
> - block = __alloc_range_bias(mm, start, end, order, flags);
> + block = __gpu_buddy_alloc_range_bias(mm, start, end, order, flags);
> if (!IS_ERR(block) || !steered)
> return block;
>
> diff --git a/drivers/gpu/tests/gpu_buddy_test.c b/drivers/gpu/tests/gpu_buddy_test.c
> index 2c6f1da1a6d1..9aaca2fc6a02 100644
> --- a/drivers/gpu/tests/gpu_buddy_test.c
> +++ b/drivers/gpu/tests/gpu_buddy_test.c
> @@ -1606,6 +1606,89 @@ static int gpu_buddy_suite_init(struct kunit_suite *suite)
> return 0;
> }
>
> +static void gpu_test_buddy_alloc_range_clear_guard(struct kunit *test)
> +{
> + const unsigned long chunk = SZ_1M, mm_size = SZ_4M;
> + struct gpu_buddy_block *block;
> + LIST_HEAD(low_dirty);
> + LIST_HEAD(high_a);
> + LIST_HEAD(high_b);
> + LIST_HEAD(barrier);
> + LIST_HEAD(dirty_res);
> + LIST_HEAD(clear_res);
> + struct gpu_buddy mm;
> +
> + KUNIT_ASSERT_FALSE_MSG(test, gpu_buddy_init(&mm, mm_size, chunk),
> + "buddy_init failed\n");
> +
> + /*
> + * Idea is to lay out a range allocation as a low dirty block, an
> + * allocated barrier, and a high naturally-aligned clear extent, then
> + * issue a dirty range allocation. It must take the low dirty block and
> + * leave the clear extent intact; if it instead cuts the clear extent,
> + * a following same-class clear allocation can no longer find a
> + * contiguous clear run and fails with -ENOSPC.
> + *
> + * [0, 1M) dirty free
> + * [1M, 2M) allocated (barrier, keeps the halves un-mergeable)
> + * [2M, 4M) clear free (one naturally aligned 2 MiB extent)
> + */
> + KUNIT_ASSERT_FALSE_MSG(test,
> + gpu_buddy_alloc_blocks(&mm, 0, chunk, chunk, chunk,
> + &low_dirty, GPU_BUDDY_RANGE_ALLOCATION),
> + "alloc [0, 1M) failed\n");
> + KUNIT_ASSERT_FALSE_MSG(test,
> + gpu_buddy_alloc_blocks(&mm, chunk, 2 * chunk, chunk, chunk,
> + &barrier, GPU_BUDDY_RANGE_ALLOCATION),
> + "alloc [1M, 2M) failed\n");
> + KUNIT_ASSERT_FALSE_MSG(test,
> + gpu_buddy_alloc_blocks(&mm, 2 * chunk, 3 * chunk, chunk, chunk,
> + &high_a, GPU_BUDDY_RANGE_ALLOCATION),
> + "alloc [2M, 3M) failed\n");
> + KUNIT_ASSERT_FALSE_MSG(test,
> + gpu_buddy_alloc_blocks(&mm, 3 * chunk, 4 * chunk, chunk, chunk,
> + &high_b, GPU_BUDDY_RANGE_ALLOCATION),
> + "alloc [3M, 4M) failed\n");
> +
> + /* [2M, 4M) coalesces into one aligned clear extent; [0, 1M) stays dirty. */
> + gpu_buddy_free_list(&mm, &low_dirty, 0);
> + gpu_buddy_free_list(&mm, &high_a, GPU_BUDDY_CLEARED);
> + gpu_buddy_free_list(&mm, &high_b, GPU_BUDDY_CLEARED);
> +
> + /* Dirty 1 MiB range alloc: must land at [0, 1M), not cut the clear extent. */
> + KUNIT_ASSERT_FALSE_MSG(test,
> + gpu_buddy_alloc_blocks(&mm, 0, mm_size, chunk, chunk,
> + &dirty_res, GPU_BUDDY_RANGE_ALLOCATION),
> + "dirty range alloc failed\n");
> + block = list_first_entry(&dirty_res, struct gpu_buddy_block, link);
> + KUNIT_EXPECT_FALSE_MSG(test, gpu_buddy_block_is_clear(block),
> + "dirty range alloc cut the clear extent (offset=%llx)\n",
> + gpu_buddy_block_offset(block));
> + KUNIT_EXPECT_EQ(test, gpu_buddy_block_offset(block), 0ULL);
> +
> + /*
> + * The 2 MiB clear extent must still be intact: a clear order-1 range
> + * allocation has to succeed. Without the guard the extent was cut
> + * above and this returns -ENOSPC.
> + */
> + KUNIT_EXPECT_FALSE_MSG(test,
> + gpu_buddy_alloc_blocks(&mm, 0, mm_size, 2 * chunk, 2 * chunk,
> + &clear_res,
> + GPU_BUDDY_RANGE_ALLOCATION |
> + GPU_BUDDY_CLEAR_ALLOCATION),
> + "clear 2M range alloc failed: clear extent was fragmented (-ENOSPC)\n");
> + if (!list_empty(&clear_res)) {
> + block = list_first_entry(&clear_res, struct gpu_buddy_block, link);
> + KUNIT_EXPECT_TRUE(test, gpu_buddy_block_is_clear(block));
> + KUNIT_EXPECT_EQ(test, gpu_buddy_block_offset(block), (u64)(2 * chunk));
> + }
> +
> + gpu_buddy_free_list(&mm, &clear_res, GPU_BUDDY_CLEARED);
> + gpu_buddy_free_list(&mm, &dirty_res, 0);
> + gpu_buddy_free_list(&mm, &barrier, 0);
> + gpu_buddy_fini(&mm);
> +}
> +
> static struct kunit_case gpu_buddy_tests[] = {
> KUNIT_CASE(gpu_test_buddy_alloc_limit),
> KUNIT_CASE(gpu_test_buddy_alloc_optimistic),
> @@ -1615,6 +1698,7 @@ static struct kunit_case gpu_buddy_tests[] = {
> KUNIT_CASE(gpu_test_buddy_alloc_clear),
> KUNIT_CASE(gpu_test_buddy_alloc_range),
> KUNIT_CASE(gpu_test_buddy_alloc_range_bias),
> + KUNIT_CASE(gpu_test_buddy_alloc_range_clear_guard),
> KUNIT_CASE_SLOW(gpu_test_buddy_fragmentation_performance),
> KUNIT_CASE_SLOW(gpu_test_buddy_dirty_tracker_performance),
> KUNIT_CASE(gpu_test_buddy_alloc_exceeds_max_order),
>
> base-commit: 49f40f15d20f5b4073d517da7efbcea079f1b9e3
^ permalink raw reply [flat|nested] 7+ messages in thread
* Re: [PATCH v4] gpu/buddy: restore two-pass clear/dirty guard for range-biased allocation
2026-11-10 16:25 ` Matthew Auld
@ 2026-11-11 4:55 ` Arunpravin Paneer Selvam
0 siblings, 0 replies; 7+ messages in thread
From: Arunpravin Paneer Selvam @ 2026-11-11 4:55 UTC (permalink / raw)
To: Matthew Auld, dri-devel, intel-gfx, intel-xe, amd-gfx
Cc: christian.koenig, alexander.deucher, Anand.Raghavendra
On 11/10/2026 9:55 PM, Matthew Auld wrote:
> On 04/11/2026 07:28, Arunpravin Paneer Selvam wrote:
>> From: Arunpravin Paneer Selvam <Arunpravin.PaneerSelvam@amd.com>
>>
>> Range-restricted allocations (the visible-VRAM aperture) skip
>> dirty_steer_window() and reach __alloc_range_bias() with no clear/dirty
>> eligibility guard, so a dirty allocation freely splits clear extents
>> (and
>> vice versa), interleaving the two classes until no large same-class run
>> survives and a 4K scanout pin fails with -ENOSPC. Use a two-pass range
>> search instead: block_incompatible() rejects opposite-class blocks in
>> the
>> strict pass and relaxes only when the strict pass returns -ENOSPC, so a
>> request still succeeds when only opposite-class space remains. A KUnit
>> regression test covers the interleaving pattern and fails without the
>> guard.
>>
>> v2: the strict pass can split a block and then reject both children, so
>> a split no longer implies a compatible block exists. Sashiko
>> reported
>> the resulting leaked splits; fix as suggested by Matthew:
>> - Push the parent back on the DFS stack after each split and merge it
>> back with unsplit_block() if both children are still free.
>> - Make block_incompatible() judge free blocks only (split nodes keep a
>> stale clear bit) and take the requested order, so mixed blocks that
>> can hold a compatible block are descended instead of skipped.
>> - Extend the KUnit test with a coalesced mixed block and a check that
>> a rejected split is merged back.
>>
>> v3: igt@kms_plane@plane-panning-bottom-right failed on Radeon RX 9070 XT
>> with -ENOSPC: the range allocator's strict pass split mixed
>> blocks to
>> look for a wholly-clear (or wholly-dirty) leaf, discarded the mixed
>> leaves it found, and ran unsplit_block to merge the splits back,
>> whose
>> per-allocation cost and residual fragmentation starved the large
>> pipe-A scanout. Fix: accept a mixed block at the requested order
>> as-is
>> (caller re-clears), so the first pass no longer splits and discards
>> mixed blocks. This clear/dirty compatibility relaxation applies
>> only
>> to range-biased allocation, where keeping blocks intact reduces
>> fragmentation and preserves large contiguous spaces. The
>> now-unneeded
>> unsplit_block helper is dropped.
>>
>> v4: a higher-order free mixed block straddling the range edge was
>> split in
>> the strict pass then abandoned, leaving both children free under
>> a split
>> parent that never coalesces, tripping gpu_buddy_fini()'s
>> free-root assert
>> and leaking the children (reported by Sashiko). Fix: in the
>> strict pass
>> don't split a free block that only partially overlaps the range;
>> the
>> fallback pass consumes it (either class accepted).
>>
>> Fixes: 1ad5e807f716 ("gpu/buddy: replace dual-tree/force_merge with
>> decoupled dirty tracker")
>> Assisted-by: LLM
>> Cc: Matthew Auld <matthew.auld@intel.com>
>> Cc: Christian König <christian.koenig@amd.com>
>> Signed-off-by: Arunpravin Paneer Selvam
>> <Arunpravin.PaneerSelvam@amd.com>
>> ---
>> drivers/gpu/buddy.c | 83 ++++++++++++++++-------------
>> drivers/gpu/tests/gpu_buddy_test.c | 84 ++++++++++++++++++++++++++++++
>> 2 files changed, 130 insertions(+), 37 deletions(-)
>>
>> diff --git a/drivers/gpu/buddy.c b/drivers/gpu/buddy.c
>> index 28cff97d57dc..d9749cbaf05d 100644
>> --- a/drivers/gpu/buddy.c
>> +++ b/drivers/gpu/buddy.c
>> @@ -1240,11 +1240,23 @@ static void __gpu_buddy_undo_splits(struct
>> gpu_buddy *mm,
>> }
>> }
>> +static bool block_incompatible(struct gpu_buddy_block *block,
>> unsigned int flags)
>> +{
>> + bool needs_clear = flags & GPU_BUDDY_CLEAR_ALLOCATION;
>> +
>> + if (!gpu_buddy_block_is_free(block))
>> + return false;
>> +
>> + return needs_clear ? !block->has_clear :
>> + gpu_buddy_block_is_clear(block);
>> +}
>> +
>> static struct gpu_buddy_block *
>> __alloc_range_bias(struct gpu_buddy *mm,
>> u64 start, u64 end,
>> unsigned int order,
>> - unsigned long flags)
>> + unsigned long flags,
>> + bool fallback)
>> {
>> u64 req_size = mm->chunk_size << order;
>> struct gpu_buddy_block *block;
>> @@ -1258,9 +1270,7 @@ __alloc_range_bias(struct gpu_buddy *mm,
>> * This range-constrained search hands back the highest/right-most
>> * address that satisfies the request: the roots are seeded
>> high-to-low
>> * and the right (higher-address) child is descended first, making
>> - * top-down the default placement here. A non-top-down clear
>> request is
>> - * the only exception, where the descent is biased towards clear or
>> - * clear-containing subtrees to satisfy the clear preference.
>> + * top-down the default placement here.
>> */
>> for (i = mm->n_roots - 1; i >= 0; --i)
>> list_add_tail(&mm->roots[i]->tmp_link, &dfs);
>> @@ -1296,8 +1306,19 @@ __alloc_range_bias(struct gpu_buddy *mm,
>> if (round_down(adjusted_end + 1, req_size) <=
>> round_up(adjusted_start, req_size))
>> continue;
>> +
>> + /*
>> + * Strict pass: don't split a partially-overlapping free
>> + * block; rejecting its child would strand the split. Let
>> + * the fallback pass consume it (either class accepted).
>> + */
>> + if (!fallback && gpu_buddy_block_is_free(block))
>> + continue;
>> }
>> + if (!fallback && block_incompatible(block, flags))
>> + continue;
>> +
>> if (contains(start, end, block_start, block_end) &&
>> order == gpu_buddy_block_order(block)) {
>> /*
>> @@ -1315,38 +1336,8 @@ __alloc_range_bias(struct gpu_buddy *mm,
>> goto err_undo;
>> }
>> - /*
>> - * Top-down is a strict address-placement policy, so when it is
>> - * requested we ignore clear steering and simply descend the
>> - * right (higher-address) child first. Only a non-top-down
>> clear
>> - * request biases the descent towards clear/has_clear subtrees.
>> - */
>> - if ((flags & GPU_BUDDY_CLEAR_ALLOCATION) &&
>> - !(flags & GPU_BUDDY_TOPDOWN_ALLOCATION)) {
>> - struct gpu_buddy_block *prefer;
>> -
>> - if (gpu_buddy_block_is_clear(block->right))
>> - prefer = block->right;
>> - else if (gpu_buddy_block_is_clear(block->left))
>> - prefer = block->left;
>> - else if (block->right->has_clear)
>> - prefer = block->right;
>> - else if (block->left->has_clear)
>> - prefer = block->left;
>> - else
>> - prefer = block->right;
>> -
>> - if (prefer == block->right) {
>> - list_add(&block->left->tmp_link, &dfs);
>> - list_add(&block->right->tmp_link, &dfs);
>> - } else {
>> - list_add(&block->right->tmp_link, &dfs);
>> - list_add(&block->left->tmp_link, &dfs);
>> - }
>> - } else {
>> - list_add(&block->left->tmp_link, &dfs);
>> - list_add(&block->right->tmp_link, &dfs);
>> - }
>> + list_add(&block->left->tmp_link, &dfs);
>> + list_add(&block->right->tmp_link, &dfs);
>> } while (1);
>> return ERR_PTR(-ENOSPC);
>> @@ -1361,6 +1352,24 @@ __alloc_range_bias(struct gpu_buddy *mm,
>> return ERR_PTR(err);
>> }
>> +static struct gpu_buddy_block *
>> +__gpu_buddy_alloc_range_bias(struct gpu_buddy *mm,
>> + u64 start, u64 end,
>> + unsigned int order,
>> + unsigned long flags)
>> +{
>> + struct gpu_buddy_block *block;
>> + bool fallback = false;
>
> Can we have fallback = true when !(gpu_buddy_clear_avail(mm) &&
> mm->dirty.total_dirty) or something, and just do one pass in that case?
>
> In xe we don't yet use clear/dirty tracking so everything is always
> dirty so no point in searching twice.
>
> Reviewed-by: Matthew Auld <matthew.auld@intel.com>
Thanks. I have sent v5 with the suggested changes.
Regards,
Arun.
>
>> +
>> + block = __alloc_range_bias(mm, start, end, order,
>> + flags, fallback);
>> + if (IS_ERR(block))
>> + return __alloc_range_bias(mm, start, end, order,
>> + flags, !fallback);
>> +
>> + return block;
>> +}
>> +
>> /* Return the highest-address free block of at least @order. */
>> static struct gpu_buddy_block *
>> get_maxblock(struct gpu_buddy *mm,
>> @@ -1924,7 +1933,7 @@ __gpu_buddy_alloc_blocks(struct gpu_buddy *mm,
>> if (flags & GPU_BUDDY_RANGE_ALLOCATION) {
>> /* Allocate traversing within the range */
>> - block = __alloc_range_bias(mm, start, end, order, flags);
>> + block = __gpu_buddy_alloc_range_bias(mm, start, end, order,
>> flags);
>> if (!IS_ERR(block) || !steered)
>> return block;
>> diff --git a/drivers/gpu/tests/gpu_buddy_test.c
>> b/drivers/gpu/tests/gpu_buddy_test.c
>> index 2c6f1da1a6d1..9aaca2fc6a02 100644
>> --- a/drivers/gpu/tests/gpu_buddy_test.c
>> +++ b/drivers/gpu/tests/gpu_buddy_test.c
>> @@ -1606,6 +1606,89 @@ static int gpu_buddy_suite_init(struct
>> kunit_suite *suite)
>> return 0;
>> }
>> +static void gpu_test_buddy_alloc_range_clear_guard(struct kunit
>> *test)
>> +{
>> + const unsigned long chunk = SZ_1M, mm_size = SZ_4M;
>> + struct gpu_buddy_block *block;
>> + LIST_HEAD(low_dirty);
>> + LIST_HEAD(high_a);
>> + LIST_HEAD(high_b);
>> + LIST_HEAD(barrier);
>> + LIST_HEAD(dirty_res);
>> + LIST_HEAD(clear_res);
>> + struct gpu_buddy mm;
>> +
>> + KUNIT_ASSERT_FALSE_MSG(test, gpu_buddy_init(&mm, mm_size, chunk),
>> + "buddy_init failed\n");
>> +
>> + /*
>> + * Idea is to lay out a range allocation as a low dirty block, an
>> + * allocated barrier, and a high naturally-aligned clear extent,
>> then
>> + * issue a dirty range allocation. It must take the low dirty
>> block and
>> + * leave the clear extent intact; if it instead cuts the clear
>> extent,
>> + * a following same-class clear allocation can no longer find a
>> + * contiguous clear run and fails with -ENOSPC.
>> + *
>> + * [0, 1M) dirty free
>> + * [1M, 2M) allocated (barrier, keeps the halves un-mergeable)
>> + * [2M, 4M) clear free (one naturally aligned 2 MiB extent)
>> + */
>> + KUNIT_ASSERT_FALSE_MSG(test,
>> + gpu_buddy_alloc_blocks(&mm, 0, chunk, chunk, chunk,
>> + &low_dirty, GPU_BUDDY_RANGE_ALLOCATION),
>> + "alloc [0, 1M) failed\n");
>> + KUNIT_ASSERT_FALSE_MSG(test,
>> + gpu_buddy_alloc_blocks(&mm, chunk, 2 * chunk,
>> chunk, chunk,
>> + &barrier, GPU_BUDDY_RANGE_ALLOCATION),
>> + "alloc [1M, 2M) failed\n");
>> + KUNIT_ASSERT_FALSE_MSG(test,
>> + gpu_buddy_alloc_blocks(&mm, 2 * chunk, 3 * chunk,
>> chunk, chunk,
>> + &high_a, GPU_BUDDY_RANGE_ALLOCATION),
>> + "alloc [2M, 3M) failed\n");
>> + KUNIT_ASSERT_FALSE_MSG(test,
>> + gpu_buddy_alloc_blocks(&mm, 3 * chunk, 4 * chunk,
>> chunk, chunk,
>> + &high_b, GPU_BUDDY_RANGE_ALLOCATION),
>> + "alloc [3M, 4M) failed\n");
>> +
>> + /* [2M, 4M) coalesces into one aligned clear extent; [0, 1M)
>> stays dirty. */
>> + gpu_buddy_free_list(&mm, &low_dirty, 0);
>> + gpu_buddy_free_list(&mm, &high_a, GPU_BUDDY_CLEARED);
>> + gpu_buddy_free_list(&mm, &high_b, GPU_BUDDY_CLEARED);
>> +
>> + /* Dirty 1 MiB range alloc: must land at [0, 1M), not cut the
>> clear extent. */
>> + KUNIT_ASSERT_FALSE_MSG(test,
>> + gpu_buddy_alloc_blocks(&mm, 0, mm_size, chunk,
>> chunk,
>> + &dirty_res, GPU_BUDDY_RANGE_ALLOCATION),
>> + "dirty range alloc failed\n");
>> + block = list_first_entry(&dirty_res, struct gpu_buddy_block, link);
>> + KUNIT_EXPECT_FALSE_MSG(test, gpu_buddy_block_is_clear(block),
>> + "dirty range alloc cut the clear extent
>> (offset=%llx)\n",
>> + gpu_buddy_block_offset(block));
>> + KUNIT_EXPECT_EQ(test, gpu_buddy_block_offset(block), 0ULL);
>> +
>> + /*
>> + * The 2 MiB clear extent must still be intact: a clear order-1
>> range
>> + * allocation has to succeed. Without the guard the extent was cut
>> + * above and this returns -ENOSPC.
>> + */
>> + KUNIT_EXPECT_FALSE_MSG(test,
>> + gpu_buddy_alloc_blocks(&mm, 0, mm_size, 2 *
>> chunk, 2 * chunk,
>> + &clear_res,
>> + GPU_BUDDY_RANGE_ALLOCATION |
>> + GPU_BUDDY_CLEAR_ALLOCATION),
>> + "clear 2M range alloc failed: clear extent was
>> fragmented (-ENOSPC)\n");
>> + if (!list_empty(&clear_res)) {
>> + block = list_first_entry(&clear_res, struct gpu_buddy_block,
>> link);
>> + KUNIT_EXPECT_TRUE(test, gpu_buddy_block_is_clear(block));
>> + KUNIT_EXPECT_EQ(test, gpu_buddy_block_offset(block), (u64)(2
>> * chunk));
>> + }
>> +
>> + gpu_buddy_free_list(&mm, &clear_res, GPU_BUDDY_CLEARED);
>> + gpu_buddy_free_list(&mm, &dirty_res, 0);
>> + gpu_buddy_free_list(&mm, &barrier, 0);
>> + gpu_buddy_fini(&mm);
>> +}
>> +
>> static struct kunit_case gpu_buddy_tests[] = {
>> KUNIT_CASE(gpu_test_buddy_alloc_limit),
>> KUNIT_CASE(gpu_test_buddy_alloc_optimistic),
>> @@ -1615,6 +1698,7 @@ static struct kunit_case gpu_buddy_tests[] = {
>> KUNIT_CASE(gpu_test_buddy_alloc_clear),
>> KUNIT_CASE(gpu_test_buddy_alloc_range),
>> KUNIT_CASE(gpu_test_buddy_alloc_range_bias),
>> + KUNIT_CASE(gpu_test_buddy_alloc_range_clear_guard),
>> KUNIT_CASE_SLOW(gpu_test_buddy_fragmentation_performance),
>> KUNIT_CASE_SLOW(gpu_test_buddy_dirty_tracker_performance),
>> KUNIT_CASE(gpu_test_buddy_alloc_exceeds_max_order),
>>
>> base-commit: 49f40f15d20f5b4073d517da7efbcea079f1b9e3
>
^ permalink raw reply [flat|nested] 7+ messages in thread
end of thread, other threads:[~2026-11-11 4:55 UTC | newest]
Thread overview: 7+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2026-11-04 7:28 [PATCH v4] gpu/buddy: restore two-pass clear/dirty guard for range-biased allocation Arunpravin Paneer Selvam
2026-11-04 7:46 ` sashiko-bot
2026-11-04 8:19 ` ✓ i915.CI.BAT: success for gpu/buddy: restore two-pass clear/dirty guard for range-biased allocation (rev2) Patchwork
2026-11-06 11:06 ` [PATCH v4] gpu/buddy: restore two-pass clear/dirty guard for range-biased allocation Arunpravin Paneer Selvam
2026-11-10 4:17 ` Arunpravin Paneer Selvam
2026-11-10 16:25 ` Matthew Auld
2026-11-11 4:55 ` Arunpravin Paneer Selvam
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox