Skip to content

iommu: reduce spinlock contention on fast path

Cong Wang linux-iommu lkml 4 versions – In review

Get series with b4
b4 am [email protected]

Versions newest first · linked by author and title

  1. v3 Latest posting 2/23 patches 12 repliesIn review
  2. v3 posting 1/23 patches 2 repliesSuperseded
  3. v23 patches 16 repliesSuperseded
Earlier versions · 1
VersionPostedPatchesRepliesStatus
v13 patches3 repliesSuperseded

Patches in v3 posting 2/2 No applies recorded yet

#Subject and filesSizeRepliesApplied
1/3iommu: avoid unnecessary magazine allocationsdrivers/iommu/iova.c+28 -174 replies
2/3iommu: optimize iova_magazine_free_pfns()drivers/iommu/iova.c+11 -114 replies
3/3iommu: avoid taking iova_rbtree_lock twicedrivers/iommu/iova.c+6 -23 replies

Reviews trailers per patch

1 of 3 patches has a Reviewed-by or Acked-by.

PatchJohn GarryRobin Murphy
1/3
2/3
3/3R-b R-b R-b

Trailers given in replies are filled; those carried in the patches from earlier versions are outlined with dashes.

Read the v3 thread

Cover letter [Patch v3 0/3] iommu: reduce spinlock contention on fast path

This patchset contains three small optimizations for the global spinlock contention in IOVA cache. Our memcache perf test shows this reduced its p999 latency down by 45% on AMD when IOMMU is enabled.

(Resending v3 on Joerg's request.)

Cong Wang (3):
iommu: avoid unnecessary magazine allocations
iommu: optimize iova_magazine_free_pfns()
iommu: avoid taking iova_rbtree_lock twice

---
v3: improve changelog, no code change
v2: fix a memory leak

 drivers/iommu/iova.c | 75 ++++++++++++++++++++++++++------------------
 1 file changed, 45 insertions(+), 30 deletions(-)
Signature · 1 line
2.21.0