Skip to content

stackdepot: allow evicting stack traces

[email protected] linux-mm lkml regressions 4 versions – Applied

Status from this message: [merged mm-stable] lib-stackdepot-adjust-depot_pools_cap-for-kmsan.patch removed from -mm tree

Get series with b4
b4 am [email protected]

Versions newest first · linked by author and title

  1. v4 Latest22 patches 34 repliesApplied 22/22
  2. v319 patches 4 repliesSuperseded
  3. v219 patches 30 repliesSuperseded
Earlier versions · 1
VersionPostedPatchesRepliesStatus
v115 patches34 repliesSuperseded

Changes in v4 from the cover letter

- Rebase onto 6.7-rc2.
- Fix lockdep annotation in depot_fetch_stack.
- New patch: "kasan: use stack_depot_put for Generic mode" (was sent for review separately but now merged into this series).
- New patch: "lib/stackdepot: print disabled message only if truly disabled" (was sent for review separately but now merged into this series).
- New patch: "lib/stackdepot: adjust DEPOT_POOLS_CAP for KMSAN".

Patches in v4 All 22 applied

#Subject and filesSizeRepliesApplied
1/22lib/stackdepot: print disabled message only if truly disabledlib/stackdepot.c+15 -91 replyApplied
2/22lib/stackdepot: check disabled flag when fetchinglib/stackdepot.c+1 -11 replyApplied
3/22lib/stackdepot: simplify __stack_depot_savelib/stackdepot.c+4 -51 replyApplied
4/22lib/stackdepot: drop valid bit from handleslib/stackdepot.c+2 -51 replyApplied
5/22lib/stackdepot: add depot_fetch_stack helperlib/stackdepot.c+28 -171 replyApplied
6/22lib/stackdepot: use fixed-sized slots for stack recordslib/Kconfig lib/stackdepot.c+19 -41 replyApplied
7/22lib/stackdepot: fix and clean-up atomic annotationslib/stackdepot.c+14 -150 repliesApplied
8/22lib/stackdepot: rework helpers for depot_alloc_stacklib/stackdepot.c+49 -371 replyApplied
9/22lib/stackdepot: rename next_pool_required to new_pool_requiredlib/stackdepot.c+24 -251 replyApplied
10/22lib/stackdepot: store next pool pointer in new_poollib/stackdepot.c+5 -11 replyApplied
11/22lib/stackdepot: store free stack records in a freelistlib/stackdepot.c+82 -490 repliesApplied
12/22lib/stackdepot: use read/write locklib/stackdepot.c+46 -4117 repliesApplied
13/22lib/stackdepot: use list_head for stack record linkslib/stackdepot.c+50 -370 repliesApplied
14/22kmsan: use stack_depot_save instead of __stack_depot_savemm/kmsan/core.c+3 -40 repliesApplied
15/22lib/stackdepot, kasan: add flags to __stack_depot_save and renameinclude/linux/stackdepot.h lib/stackdepot.c mm/kasan/common.c +3 more+48 -250 repliesApplied
16/22lib/stackdepot: add refcount for recordsinclude/linux/stackdepot.h lib/stackdepot.c+20 -50 repliesApplied
17/22lib/stackdepot: allow users to evict stack tracesinclude/linux/stackdepot.h lib/stackdepot.c+50 -14 repliesApplied
18/22kasan: remove atomic accesses to stack ring entriesmm/kasan/report_tags.c mm/kasan/tags.c+12 -260 repliesApplied
19/22kasan: check object_size in kasan_complete_mode_report_infomm/kasan/report_tags.c+3 -10 repliesApplied
20/22kasan: use stack_depot_put for tag-based modesmm/kasan/tags.c+8 -20 repliesApplied
21/22kasan: use stack_depot_put for Generic modemm/kasan/common.c mm/kasan/generic.c mm/kasan/quarantine.c+40 -114 repliesApplied
22/22lib/stackdepot: adjust DEPOT_POOLS_CAP for KMSANlib/stackdepot.c+10 -00 repliesApplied

Reviews trailers per patch

19 of 22 patches have a Reviewed-by or Acked-by.

PatchAlexander PotapenkoMarco ElverOscar Salvador
1/22R-b R-b
2/22R-b R-b
3/22R-b R-b
4/22R-b R-b
5/22R-b R-b
6/22R-b
7/22R-b
8/22R-b R-b
9/22R-b R-b
10/22R-b R-b
11/22R-b
12/22R-b R-b
13/22
14/22R-b
15/22R-b
16/22R-b
17/22
18/22R-b
19/22R-b
20/22R-b
21/22R-b
22/22

Trailers given in replies are filled; those carried in the patches from earlier versions are outlined with dashes. Scroll the table sideways for all 3 reviewers.

Read the v4 thread

Cover letter [PATCH v4 00/22] stackdepot: allow evicting stack traces

From: Andrey Konovalov <[email protected]>

Currently, the stack depot grows indefinitely until it reaches its capacity. Once that happens, the stack depot stops saving new stack traces.

This creates a problem for using the stack depot for in-field testing and in production.

For such uses, an ideal stack trace storage should:

1. Allow saving fresh stack traces on systems with a large uptime while limiting the amount of memory used to store the traces;
2. Have a low performance impact.

Implementing #1 in the stack depot is impossible with the current keep-forever approach. This series targets to address that. Issue #2 is left to be addressed in a future series.

This series changes the stack depot implementation to allow evicting unneeded stack traces from the stack depot. The users of the stack depot can do that via new stack_depot_save_flags(STACK_DEPOT_FLAG_GET) and stack_depot_put APIs.

Internal changes to the stack depot code include:

1. Storing stack traces in fixed-frame-sized slots (vs precisely-sized slots in the current implementation); the slot size is controlled via CONFIG_STACKDEPOT_MAX_FRAMES (default: 64 frames);
2. Keeping available slots in a freelist (vs keeping an offset to the next free slot);
3. Using a read/write lock for synchronization (vs a lock-free approach combined with a spinlock).

This series also integrates the eviction functionality into KASAN:
the tag-based modes evict stack traces when the corresponding entry leaves the stack ring, and Generic KASAN evicts stack traces for objects once those leave the quarantine.

With KASAN, despite wasting some space on rounding up the size of each stack record, the total memory consumed by stack depot gets saturated due to the eviction of irrelevant stack traces from the stack depot.

With the tag-based KASAN modes, the average total amount of memory used for stack traces becomes ~0.5 MB (with the current default stack ring size of 32k entries and the default CONFIG_STACKDEPOT_MAX_FRAMES of 64). With Generic KASAN, the stack traces take up ~1 MB per 1 GB of RAM (as the quarantine's size depends on the amount of RAM).

However, with KMSAN, the stack depot ends up using ~4x more memory per a stack trace than before. Thus, for KMSAN, the stack depot capacity is increased accordingly. KMSAN uses a lot of RAM for shadow memory anyway, so the increased stack depot memory usage will not make a significant difference.

Other users of the stack depot do not save stack traces as often as KASAN and KMSAN. Thus, the increased memory usage is taken as an acceptable trade-off. In the future, these other users can take advantage of the eviction API to limit the memory waste.

There is no measurable boot time performance impact of these changes for KASAN on x86-64. I haven't done any tests for arm64 modes (the stack depot without performance optimizations is not suitable for intended use of those anyway), but I expect a similar result. Obtaining and copying stack trace frames when saving them into stack depot is what takes the most time.

This series does not yet provide a way to configure the maximum size of the stack depot externally (e.g. via a command-line parameter). This will be added in a separate series, possibly together with the performance improvement changes.

---

Changes v3->v4:
- Rebase onto 6.7-rc2.
- Fix lockdep annotation in depot_fetch_stack.
- New patch: "kasan: use stack_depot_put for Generic mode" (was sent for review separately but now merged into this series).
- New patch: "lib/stackdepot: print disabled message only if truly disabled" (was sent for review separately but now merged into this series).
- New patch: "lib/stackdepot: adjust DEPOT_POOLS_CAP for KMSAN".

Changes v2->v3:
- Fix null-ptr-deref by using the proper number of entries for initializing the stack table when alloc_large_system_hash() auto-calculates the number (see patch #12).
- Keep STACKDEPOT/STACKDEPOT_ALWAYS_INIT Kconfig options not configurable by users.
- Use lockdep_assert_held_read annotation in depot_fetch_stack.
- WARN_ON invalid flags in stack_depot_save_flags.
- Moved "../slab.h" include in mm/kasan/report_tags.c in the right patch.
- Various comment fixes.

Changes v1->v2:
- Rework API to stack_depot_save_flags(STACK_DEPOT_FLAG_GET) + stack_depot_put.
- Add CONFIG_STACKDEPOT_MAX_FRAMES Kconfig option.
- Switch stack depot to using list_head's.
- Assorted minor changes, see the commit message for each path.

Andrey Konovalov (22):
lib/stackdepot: print disabled message only if truly disabled lib/stackdepot: check disabled flag when fetching lib/stackdepot: simplify __stack_depot_save lib/stackdepot: drop valid bit from handles lib/stackdepot: add depot_fetch_stack helper lib/stackdepot: use fixed-sized slots for stack records lib/stackdepot: fix and clean-up atomic annotations lib/stackdepot: rework helpers for depot_alloc_stack lib/stackdepot: rename next_pool_required to new_pool_required lib/stackdepot: store next pool pointer in new_pool lib/stackdepot: store free stack records in a freelist lib/stackdepot: use read/write lock
lib/stackdepot: use list_head for stack record links
kmsan: use stack_depot_save instead of __stack_depot_save lib/stackdepot, kasan: add flags to __stack_depot_save and rename lib/stackdepot: add refcount for records lib/stackdepot: allow users to evict stack traces
kasan: remove atomic accesses to stack ring entries
kasan: check object_size in kasan_complete_mode_report_info
kasan: use stack_depot_put for tag-based modes
kasan: use stack_depot_put for Generic mode lib/stackdepot: adjust DEPOT_POOLS_CAP for KMSAN

 include/linux/stackdepot.h |  59 ++++-
 lib/Kconfig                |  10 +
 lib/stackdepot.c           | 452 ++++++++++++++++++++++++-------------
 mm/kasan/common.c          |   8 +-
 mm/kasan/generic.c         |  27 ++-
 mm/kasan/kasan.h           |   2 +-
 mm/kasan/quarantine.c      |  26 ++-
 mm/kasan/report_tags.c     |  27 +--
 mm/kasan/tags.c            |  24 +-
 mm/kmsan/core.c            |   7 +-
 10 files changed, 427 insertions(+), 215 deletions(-)
Signature · 1 line
2.25.1