diff options
| author | Eduard Zingerman <eddyz87@gmail.com> | 2026-10-26 16:38:29 -0700 |
|---|---|---|
| committer | Alexei Starovoitov <ast@kernel.org> | 2026-10-27 16:26:21 +0000 |
| commit | 344c14e8b3d45e3c643cec1ba6acd3e4691dfe66 (patch) | |
| tree | 3f78008a59cdf7cd52ef43c2916512346a30d2a7 /include/linux/workqueue_api.h | |
| parent | 39cd4e0915c4d9661ec1fdd472b515b28c98eb37 (diff) | |
selftests/bpf: Cases for SCEV interaction with infinite loop detection
Add three test cases exercising state pruning after loop counter
clamping. All turn a 32-bit stack spill into STACK_MISC during the
first iteration. This prevents a state match at the outer loop header
on the next iteration, while allowing states to match further inside
the loop. All are falsely reported as infinite loop w/o changes in
patch [1].
- clamp_callee_state_pruning(), a callee overwrites the upper half of
the caller's stack slot. States then match at the callee's exit,
where the current frame has no active loop.
Verify that current_loop() finds the caller's active loop.
- clamp_nested_loop_state_pruning(), the overwrite happens inside an
inner loop. Exercise matching states when the inner loop is entered
again on a subsequent outer iteration.
- clamp_nested_unrolled_loop_state_pruning(), same as above, but the
inner loop is not recognized by SCEV and is unrolled with a precise
counter, so its entries are never marked terminating.
[1] bpf: avoid false infinite loop reports after changes in SCEV clamping
Signed-off-by: Eduard Zingerman <eddyz87@gmail.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://patch.msgid.link/20261026-scev-simpler-clamp-v2-9-f373aec9825f@gmail.com
Diffstat (limited to 'include/linux/workqueue_api.h')
0 files changed, 0 insertions, 0 deletions
