diff options
| author | Arunpravin Paneer Selvam <Arunpravin.PaneerSelvam@amd.com> | 2025-10-06 15:21:24 +0530 |
|---|---|---|
| committer | Arunpravin Paneer Selvam <Arunpravin.PaneerSelvam@amd.com> | 2025-10-07 19:19:15 +0530 |
| commit | c07823f85012b80d8b2e251acbc23a0bdc2925c2 (patch) | |
| tree | 3297b03c24144ce56e37b2fd6cd3ecd80c270599 /tools/perf/scripts/python/bin/stackcollapse-report | |
| parent | d4cd665c98c144dd6ad5d66d30396e13d23118c9 (diff) | |
| download | linux-c07823f85012b80d8b2e251acbc23a0bdc2925c2.tar.gz linux-c07823f85012b80d8b2e251acbc23a0bdc2925c2.zip | |
drm/buddy: Add KUnit tests for allocator performance under fragmentation
Add KUnit test cases that create severe memory fragmentation and
measure allocation/free performance.
The tests simulate two scenarios -
1. Allocation under severe fragmentation
- Allocate the entire 4 GiB space as 8 KiB blocks with 64 KiB alignment,
split them into two groups and free with mixed flags to block coalescing.
- Repeatedly allocate and free 64 KiB blocks while timing the loop.
- Freelist runtime: 76475 ms(76.5 seconds), soft-lockup triggered.
RB-tree runtime: 186 ms.
2. Reverse free order under fragmentation
- Create a similarly fragmented space, free half the blocks, reverse
the order of the remainder, and release them with the cleared flag.
- Freelist runtime: 85620 ms(85.6 seconds).
RB-tree runtime: 114 ms.
Signed-off-by: Arunpravin Paneer Selvam <Arunpravin.PaneerSelvam@amd.com>
Reviewed-by: Matthew Auld <matthew.auld@intel.com>
Link: https://lore.kernel.org/r/20251006095124.1663-3-Arunpravin.PaneerSelvam@amd.com
Diffstat (limited to 'tools/perf/scripts/python/bin/stackcollapse-report')
0 files changed, 0 insertions, 0 deletions
