/xce-audit-invariants scheduler dimension-1/4 audit, prompted by a fresh
user-captured .replay file that still showed massive --sim-verify mismatch
volume after bug-1843's fix. Root-caused via temporary diagnostic
instrumentation against the real capture (reverted, never committed) rather
than trusting the surface symptom.
bug-1844: sim_verify_alloc_eq() (sim_verify.c) compared `instid`, a bare
per-process monotonic counter (scheduler.c's sched_get_id(), core->nextid)
stamped at allocation-build time -- not semantic content. The original
recording's counter and a fresh replay's counter have no reason to reach
the same value for "the same" dispatch. Live-verified: every DISPATCH
mismatch showed identical value/duration/resname/reserve, differing only
in instid (e.g. 826 vs 1). Dropped from the comparison, same treatment as
bug-1842's `from` exclusion.
bug-1845: sim_verify_job_preason_eq() required PEND decisions' full
host/resource reason-detail lists to match. core.h documents "simulation
isolation" as deliberate design: sim/replay never registers
core->unsatisfiable_lookup_cb, so it can never hit the unsatisfiable-cache
main.c's normal-mode branch registers. A live cache hit
(sched_pend_from_cached_reason_()) records only main_reason with zero
detail; replay always falls through to the full per-host computation and
always produces detail. Same semantic outcome (PEND, same main_reason) at
two architecturally-legitimate detail levels. Relaxed the comparator to
main_reason only. Separately verified the cache itself is sound (key is
safe within a round since capacity only decreases and the cache clears
every round) -- ruling it out as the cause of a second, still-open
divergence (bug-1846: one job genuinely dispatches one round early in
replay vs. the original recording -- a suspected resource-accounting
parity gap, not yet root-caused, logged for follow-up).
bug-1772 (previously deferred, incorrectly tagged "fixed"): re-found
independently by a dimension-4 snapshot-field-completeness audit subagent.
snapshot_build_primary_schedobj() (snapshot_load.c) never computed
resreq_fp, unlike its declared model core_build_primary_schedobj()
(core_events.c) -- every snapshot-restored job got a zeroed fingerprint,
which alloc_cache.c's cache key folds in directly alongside
user/queue/project, so two distinct restored jobs sharing those three
fields could silently share an incorrect unsatisfiable-cache verdict.
Added the same 4-field computation the live path uses.
TDD throughout: each fix has a dedicated regression test, and each was
verified via genuine red (revert fix, confirm the new test fails) then
green (restore, full 85/85 scheduler suite). Also ran three parallel
dimension audits (core I/O isolation, picker/allocator jbuf rendezvous) --
both clean, no violations found.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
sched_record_session_add_decision() (recording_session.c) grows its
per-round decisions array via realloc(), doubling from 8. On a failed
growth it silently dropped the current decision and every one after it
that round, without marking the round incomplete -- finalize() then
wrote the short array as an ordinary, apparently-complete ReplaySession.
--sim-verify's replay comparison had no way to distinguish that from a
genuinely smaller live round, producing spurious "expected: <none>"
mismatches for jobs that were, in reality, correctly decided live and
simply never recorded. Live scheduling itself was unaffected either
way -- only the .replay file's own record of that one round lost its
tail.
This refutes the leading theory from the prior investigation (replay
seeing a job's first appearance "one round early" via coarser event
batching in sim_loop.c) -- confirmed via targeted diagnostic
instrumentation against the original captured .replay file: replay
deterministically reproduces the full candidate set every round;
the recording itself has the one-round gap.
Fix: track truncation on the session ctx; finalize() now refuses to
write an incomplete session and fails loudly through the same
sched_record_abort_on_error() path every other recording failure in
this subsystem already uses.
TDD: added a regression test forcing a real, deterministic realloc()
failure via -Wl,--wrap=realloc (scoped to this one test binary --
no prior malloc/realloc fault-injection harness existed in this
codebase). Verified genuine red -> green: the test fails without the
fix, passes with it. Full 85/85 scheduler suite green.
Note: this prevents future recordings from losing decisions this way;
it cannot retroactively repair an already-captured .replay file.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
sim_verify_alloc_eq() compared xce_alloc_t.from, which make_alloc()
(xcealloc.c) stamps with the real OS wall clock (time(NULL)), not
core.c's virtualized sim clock. The recorded (expected) value is
whatever real second the original run happened to execute in; the
freshly-rebuilt (actual) value is whatever real second the replay
executes in -- two unrelated real-world timestamps, so every
otherwise-correct DISPATCH decision failed verification whenever a
replay ran even one wall-clock second away from its own recording.
Drop `from` from the comparison -- it's reservation bookkeeping, not
part of a decision's semantic identity. Verified against a real
captured .replay file: 120/120 false DISPATCH mismatches -> 0.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
sim_reconcile_unaccounted_jobs() (sim_loop.c) was calling
picker->picker_iface->on_dispatch() directly from the sim driver thread,
violating plugin.h's documented "allocator thread only" contract for
Post-allocation hooks. Dormant in practice only because builtin_fcfs.c
happens to lock internally around every hook -- a defensive measure the
contract text doesn't actually require of picker plugins.
Since scheduler-sim-live-allocation, sim mode spawns the same real
sched_job_picker()/sched_allocator() thread pair normal mode uses, so a
real allocator thread now exists to route through. Added a
reconcile_req/reconcile_ack cbuf pair on xce_sched_core_t and a reserved
XCE_SCHED_RECONCILE_WAKE sentinel pushed through jbuf to wake the
allocator thread without touching jbuf's pinned capacity-1/item-type
invariant (Req 5 AC 7). sched_allocator() services the request and calls
on_dispatch() itself; sim_reconcile_invoke_on_dispatch() blocks on the ack.
Falls back to a direct call when no allocator thread is live for a given
core (standalone unit tests), which is safe by construction.
TDD: new test spawns a stand-in allocator-relay thread and asserts
on_dispatch() fires there, not on the caller's thread -- confirmed RED
against the old direct-call code, GREEN after. Four other test files with
hand-rolled sched_allocator() stubs needed the same relay branch (found by
an actual hang at meson's timeout). Full scheduler suite: 85/85, zero
regressions.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
User-reported: replaying a captured session dispatched 44 single-slot jobs
onto a 22-slot host. Root cause traced through two layers:
- bug-1836: sim_reconcile_unaccounted_jobs() (the "Recorded-Decision
Reconciliation" backstop) force-applied any recorded DISPATCH decision
it found, with zero live capacity check. Added
sim_reconciled_alloc_fits_capacity(), reusing the same
sched_check_host_satisfy()/sched_check_shared_satisfy() admission checks
the live picker/allocator path already uses.
- bug-1838: even with that gate, reconciliation still preempted every
freshly-PEND job before the live picker/allocator got a turn each round
-- unnecessary now that scheduler-sim-live-allocation gave sim mode the
same real, capacity-checked picker/allocator threads normal mode uses.
Narrowed reconciliation's trigger back to JOB_IS_RUN-only (undoing a
2026-09-11 widening whose own justification no longer holds), so a
freshly-pushed job now always flows through the same live, verified
decision path as normal mode. Reconciliation now only ever fires for a
job already RUN from before the replay's own observation window --
the one case structurally invisible to the live picker.
Also logged (bug-1837, not fixed here, needs its own design decision):
sim_reconcile_unaccounted_jobs() calls a picker's on_dispatch() hook from
the driver thread instead of the allocator thread plugin.h documents it
belonging to -- currently inert since the only real picker plugin happens
to lock internally, but not contract-safe for a future lock-free one.
Fixing bug-1838 required re-examining every test exercising the old
PEND-inclusive trigger; several needed their fixtures corrected to
register a real worker (harmless before any capacity check existed), and
others needed their actual intent re-derived and their assertions
inverted or restated now that a PEND job is never eligible for
reconciliation. Dated correction notes added in place to the
scheduler-sim-eager-reconciliation spec, matching its own existing
2026-09-11 correction-note precedent.
Verified: full scheduler suite 85/85, independently re-confirmed by
running the four most affected binaries directly.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
update_limit_res_value() computed a percent-based limit's capacity from
the resource's own live, fluctuating value (e.g. "mem"), letting the
enforced cap drift with unrelated memory pressure instead of staying
pinned to a stable fraction of total capacity. instantiate_limit() now
resolves a static capacity-source resource once at instantiation time
(mem->maxmem, swap->maxswap, table-driven), stored on a new
xce_vres_t.percent_res field; a missing counterpart falls back to
today's behavior rather than failing instantiation.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01RLSNJhNoULvVcUhXEFafQC
fmt_limit()/fmt_limit_units() formatted a percent-based limit's value
as a bare snprintf(..., "%f%%", limit->val.val) -- echoing the raw
stored value with a literal "%" appended. But xce_limit_val_t.val is
genuinely a 0-1 fraction when percent is set (confirmed end-to-end:
no /100 or *100 scaling exists anywhere in the config->capnp->runtime
pipeline; confirmed authoritative by the original hand-written C
literal this codebase's oldest percent-based limit was migrated from,
.val = 0.2 for a documented "20% mem cap"). A correctly-configured 20%
limit therefore displayed as "0.200000%" instead of "20.000000%" --
misleading, though technically an honest echo of the stored value.
Fixed by multiplying by 100.0 before formatting. Display-only --
update_limit_res_value()'s actual capacity computation was already
correct and is untouched. fmt_limit() (a thin XCE_UNITS_RAW wrapper)
inherits this fix for its own percent-branch output too, a deliberate,
intentional change from its pre-fix behavior for percent-based limits
specifically (its 5 pre-existing direct tests only assert rc > 0,
never the exact percent string, so this is safe).
Updated test_limit.c's percent-mode test fixture from an unrealistic
.val=80.0 to a realistic .val=0.8 (assertion text unchanged, since
0.8*100=80 either way) and test_repl_limit_table.c's own percent test
to expect its shared fixture's arbitrary .val=50.0 constant correctly
rendering as "5000.000000%". RED confirmed via the raw Unity testlog's
:FAIL: lines (the aggregate meson summary masks single-test failures,
a known harness quirk in this project); GREEN confirmed after, both
lib and scheduler suites, 0 Unity FAILs.
Discovered during live manual verification of scheduler-sim-diagnostics
Requirement 6 against a real deployed cluster.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01RLSNJhNoULvVcUhXEFafQC
Final validation task for the Human-Readable Units increment (tasks
7-10.1). Adds three tests closing gaps no prior task's own tests
covered:
- Cross-surface consistency (Requirement 6.5): a single info limit
request against a byte-valued limit with both a configured setting
and a live usage instance shows the SAME units convention in both
lines, across mode switches -- neither task 8.2's nor task 9's own
tests combined both surfaces in one capture.
- Non-byte-valued exclusion (6.3): a new "slots"-resourced fixture
(the first non-"mem" fixture in this file) proves the entire
combined display is byte-for-byte identical across all three units
modes.
- Percentage exclusion (6.3) at the REPL-dispatch level: a percent
limit's own percentage value survives real repl_dispatch()/
set-units switching unchanged, complementing test_limit.c's
existing library-level proof of the same invariant.
Full regression sweep: lib suite 368/368, scheduler suite 85/85,
including test_repl_limit_table.c (35/35), test_repl_resource_table.c,
test_repl_resource_formatting.c, and test_limit_plugin.c all
unaffected. No production code changed. This closes out Requirement 6
(ACs 6.1-6.5) and reconfirms Requirement 5 (ACs 5.1-5.3) end-to-end.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01RLSNJhNoULvVcUhXEFafQC
Six pre-existing assertions in test_repl_limit_table.c (across 4 test
functions) used bare strstr() substring checks against whole-number
fixture values on "mem" -- e.g. strstr(captured, "accounted=4.00").
Task 9's fix changed the actual output to "accounted=4.000000 (4.00
B)", but the old substring is a literal prefix of the new one, so
these assertions kept passing by coincidence of formatting, providing
zero actual regression protection for the fix.
Strengthened each to assert the full corrected "<raw> (<human>)"
string. Proved via TDD: reverting task 9's fix now makes all 6
strengthened assertions (plus task 9's own 2 genuine tests) fail;
restoring the fix returns the suite to 85/85 green.
Confirmed via full-file scan that no other loose assertion of this
shape remains, and that info worker/info resource's own regression
suites were never vulnerable to this class of bug (they build expected
strings dynamically rather than hardcoding literal substrings).
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01RLSNJhNoULvVcUhXEFafQC
repl_print_limit_usage()/repl_print_limit_usage_per_host() (repl_info.c)
passed vres->vres->name (a synthesized compound scope-instance name, e.g.
"limits2/u=rshen/q=xce/p=") to the byte-value classifier instead of the
limit's actual constrained resource name -- so a byte-valued limit's
usage amounts never got recognized as byte-valued, regardless of the
underlying formatting machinery already working correctly.
Fixed both call sites to pass the limit's own resname
(limit->val.resname / vres->limit->val.resname) instead. Added tests
proving human-readable rendering for both the cluster-wide/GLOBAL path
and a percentage-based per-host limit (Requirement 6 AC 6.4: a
percentage-of-byte-resource limit's resolved absolute amount is
human-formatted too).
Discovered and confirmed (implementer + independent reviewer): the
fix landed with zero pre-existing test failures, not the handful
design.md predicted -- those assertions used whole-number fixture
values with bare strstr() substring checks, which still match the
new, longer human-augmented output by coincidence of formatting.
Documented in tasks.md; task 10.1 will strengthen those assertions
to genuinely exercise the fix rather than pass by accident.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01RLSNJhNoULvVcUhXEFafQC
Switches both fmt_limit() call sites in repl_info_limit() (the "all"
path and the named-lookup path) to
fmt_limit_units(..., translate_units_mode(repl_current_units_mode(sim))),
so a byte-valued limit's configured amount now renders human-readable
under the REPL's real, currently-selected units mode -- the same
convention info worker/info resource already honor.
Adds two tests proving human-readable rendering under the default
mode and consistent switching across set units raw|human|both via
real repl_dispatch(). RED confirmed by reverting the call sites
(2 failures, 28 pre-existing tests in the same file unaffected).
Corrects an inaccurate self-reported regression-safety rationale
(there is no slots-resource fixture in this file; safety instead
comes from those tests never asserting on the formatted value line),
logged in tasks.md's Implementation Notes.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01RLSNJhNoULvVcUhXEFafQC
Adds Requirement 6 (Human-Readable Units for Byte-Valued Limit Amounts)
to the spec: requirements, design, research, and tasks 7-10.
Task 7 (Foundation): promotes the REPL's byte-unit scaling algorithm
and byte-valued-resource classifier into src/lib/utils.c/utils.h
(xce_format_bytes_human, xce_is_bytes_valued_resource), alongside a
new display-mode-aware composer (xce_format_units_value,
xce_units_mode_t) -- following the fmt_jobid()/fmt_jobid_ex()
precedent. repl_fmt.c's scheduler-only functions become one-line
delegating wrappers. repl_info.c's repl_format_numeric_resource_value()
(shared by info worker/info resource, not just the upcoming info
limit) is retyped to delegate to the new shared composer via a small
translate_units_mode() bridge, so every consumer benefits from one
implementation going forward.
No existing symbol's signature changed; only new symbols were added.
Confirmed behavior-preserving: full scheduler suite 85/85 unchanged,
including info worker/info resource's own regression tests. One
narrow, untested, out-of-requirement-scope corner case found and
documented (swp's RAW/BOTH mode rendering under the new single-value
composer interface) -- logged in tasks.md's Implementation Notes.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01RLSNJhNoULvVcUhXEFafQC
core_clock.c joined the mechanically-enforced isolation archive
(sched_core_isolation_lib) and its own doc comment claimed coverage by
the forbidden-headers/forbidden-symbols text scans, but was actually
absent from both -- only the nm(1) undefined-symbol layer covered it,
which cannot catch a forbidden #include resolving to no undefined
symbol (a macro-only header).
Added XCE_CORE_CLOCK_C_PATH (meson.build, mirroring the existing
per-file path macros) and extended both
test_core_never_includes_forbidden_headers() and
test_core_never_calls_forbidden_symbols() to scan core_clock.c. It
remains excluded only from the wall-clock scan, for its one sanctioned
clock_gettime() call.
TDD: confirmed RED by temporarily inserting a forbidden #include into
core_clock.c (84/85, correctly attributed failure), reverted
(byte-identical to HEAD), reconfirmed GREEN (85/85). No live violation
existed -- this closes a coverage gap in the mechanism itself.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01RLSNJhNoULvVcUhXEFafQC
/xce-audit-invariants scheduler (dimension 1, driver state-equivalence,
re-run after scheduler-sim-live-allocation) found plugin_end_session()
still checked `plugin->begin_session_fn != NULL` before invoking
`plugin->end_session_fn(plugin)` -- a copy-paste of
plugin_begin_session()'s own (correct) guard. Two failure modes: a
plugin registering only end_session_fn silently never has it invoked;
a plugin registering only begin_session_fn crashes via a NULL
function-pointer call. Currently inert in production (no live plugin
hits either shape) but a real latent defect in exactly the
session-boundary code the driver-equivalence invariant governs.
This exact defect had been independently filed twice: bug-1492
(2026-08-04, scheduler-core task 5.2's review) and bug-1682
(2026-08-24, found adjacent to bug-1681's fix) -- neither investigation
cross-referenced the other. Both entries also carried a "fixed" tag
inconsistent with their own "not yet fixed" narrative text, caught by
this audit's independent code+text cross-check.
Fixed the one-line guard; added test_end_session_invokes_end_session_fn_
when_only_it_is_set() (test_plugin_hook_dispatch.c) proving the silent
non-invocation, confirmed RED (84/85) against the buggy guard and GREEN
(85/85) after the fix. Updated both buglog entries to FIXED with
cross-references, and corrected bug-1611's stale narrative to reflect
its narrowing (not elimination) by scheduler-sim-live-allocation.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01RLSNJhNoULvVcUhXEFafQC
Re-ran test_repl_equivalence.c, test_sim_equivalence.c, and
test_sim_contention_parity.c in full: all pass, 85/85, zero assertion
changes anywhere. Confirmed all three still exercise their own
link-satisfying always-PEND sched_job_picker/sched_allocator stub
(added task 4.1), correctly proving the new two-thread jbuf timing
didn't disturb decision ordering -- but not proving real live-path
resource-based dispatch parity, which is already covered separately
by tasks 4.1/4.3/5.1/5.2's real-scheduler_srcs-linked fixtures.
Fixed a stale documentation defect: all three files' header/inline
comments claimed sim mode "structurally CANNOT emit a DISPATCH
decision" via sim_picker_thread_fn(), a function this feature removed
in tasks 3.1/4.1. That claim is now false of production sim mode
(disproved by test_sim_live_allocation.c and others); corrected with
a RECONCILIATION addendum rather than rewriting the original
historical rationale. No production code or assertion values changed.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01RLSNJhNoULvVcUhXEFafQC
Adds the feature's concrete original acceptance test: a genuine
capnp-encoded .replay (real ReplayHeader/SnapshotEntry/Conf4Sched with
one global and one host-scoped limit), driven end-to-end through the
real, unmodified sched_sim_driver_run(). The dispatched job's usage
registers against both limits via limit_get_configured_limits()/
limit_get_usage_instances() -- the same accessors info limit reads --
proving real, non-zero usage where today's REPL previously showed none.
RED-phase evidence: reverting task 4.3's 4 files to their pre-fix
commit reproduces the exact predicted failure (host-scoped usage
missing, global usage intact); restoring to HEAD reproduces GREEN.
Closes bug-1826 (the bug this whole feature was built to fix) --
buglog entry updated from NOT FIXED to FULLY FIXED, citing this test.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01RLSNJhNoULvVcUhXEFafQC
Adds two new fixtures spawning the real sched_job_picker/sched_allocator
thread pair directly against bare normal-mode and sim-mode core state
under genuine resource contention (2.0 capacity, 3 jobs x 1.0 demand),
asserting identical admission decisions and identical accounting state
(job->accounts, w->resInUse), with and without a matching host-scoped
limit. The pre-existing test_sim_contention_parity.c could not be reused
since its white-box stubs hard-code the old always-PEND behavior and
never exercise live election/allocation.
Confirms build_limit_vres_requests()/get_vres_value() (plugins/limit.c)
are #if 0'd dead code with zero live callers -- no limit in this codebase
currently gates/denies admission in either mode, only tracks/displays
usage -- so this proves usage-tracking parity, documented explicitly as
an honest scope boundary rather than a silent narrowing.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01RLSNJhNoULvVcUhXEFafQC
Fully closes bug-1829's own tracked follow-up from task 4.1: retypes
seqjob_allocator.c/parjob_allocator.c's init_fn AND runtime hooks
(prepare_allocation_hook, job_alloc_fn) to read the plugin's own
honest plugin->core field (task 1.2) instead of deriving an
xce_scheduler_t* from plugin->sched -- mirroring task 2.3's limit.c
migration exactly. The gap was broader than initially scoped: not
just init_fn, but every runtime hook independently derived sched from
plugin->sched on each allocation attempt.
Adds seqjob_allocator_new_plugin()/parjob_allocator_new_plugin()
factories (plugin.h, mirroring limit_new_plugin()), letting
sim_alloc.c's bootstrap functions register both plugins DIRECTLY
against &(sim->core) via the existing plugin_register_core() -- the
same call sim_bootstrap_picker()/sim_bootstrap_limit_plugin() already
use. Task 4.1's throwaway xce_scheduler_t wrapper is gone from the
plugin-registry/fan-out path entirely; plugin_prepare_allocation()'s
fan-out during a live seqjob/parjob dispatch now genuinely reaches
sim->core.plugins, the same registry the limit plugin is registered
in.
One narrow, necessary exception remains, independently verified
confined: my_alloc() still needs a small xce_scheduler_t wrapper
(only .core.resmgr aliased to sim->core) to satisfy sched_make_allocs()
(scheduler.c, out of this task's boundary, unmodified) -- set on
plugin->sched only AFTER plugin_register_core() has run, never read by
the plugin registry or any prepare_allocation_hook fan-out.
TDD: new test_sim_alloc_limit_fanout.c drives a real live dispatch
(sched_job_picker()/sched_allocator() thread pair) with a host-scoped
limit registered directly against sim->core, proving a real usage
instance is created -- isolated from both reconciliation and
global-limit matching. RED: dispatch succeeded but the usage-instance
iterator was empty. GREEN: 82/82, zero regressions.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01RLSNJhNoULvVcUhXEFafQC
Adds a plugin_job_election(&(sim->core), job) call to
sim_reconcile_unaccounted_jobs()'s FOUND branch, before the existing
alloc-reconstruction/state-transition code, so a job whose outcome
comes from the recording still gets its global-limit usage registered
-- the same core-safe dispatcher the live path (task 4.1) already
uses, not a separate mechanism.
The existing job->accounts != NULL trigger guard needs no change: it
already excludes any job the live path has already decided, so this
call is naturally idempotent (match_job_global_limit_'s own
obj->configver guard makes a redundant call a safe no-op) -- narrowing
this function to a genuine timing-gap safety net rather than the
general-case decision source it used to be, per design's Live/Recorded
Authority Rule.
Host-scoped limit matching is deliberately not attempted here (needs a
resolved worker context this function doesn't independently
construct) -- that gap is tracked separately as task 4.3.
TDD: new tests prove a reconciled job's global limit is genuinely
registered via the real limit_get_usage_instances() accessor, and that
a job the live path already decided is never reprocessed.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01RLSNJhNoULvVcUhXEFafQC
Replaces sim_picker_thread_fn() (the always-PEND placeholder) with the
real sched_job_picker()/sched_allocator() threads (task 3.1), spawned
against &(sim->core) in sched_sim_driver_run(). Simulation mode now
genuinely uses the jbuf capacity-1 rendezvous for the first time
(confirmed still exactly one producer/one consumer). Adds
sched_election.c to sched_sim_isolation_lib.
Fixes bug-1829 (blocking, only reachable once sim_allocator() actually
called sched_make_allocation() for the first time): core->seqjob_
allocator/parjob_allocator were never bootstrapped for simulation mode
at all. Fixed via new sim_alloc.c (sim_bootstrap_seqjob_allocator()/
sim_bootstrap_parjob_allocator()), mirroring the sim_picker.c/
sim_limit.c precedent.
That fix's own workaround -- a throwaway xce_scheduler_t wrapper
satisfying seqjob/parjob's init_fn -- has a documented, narrow side
effect: prepare_allocation_hook fan-out during a live seqjob/parjob
dispatch attempt reaches the wrapper's own separate plugin registry,
not sim->core.plugins, so a plugin registered only against the real
sim core (the limit plugin) isn't reached for host-scoped matching
during a live dispatch yet. Documented in design.md's new "Known Seam
Gap" section; task 4.3 added to tasks.md to close it before task 5.2's
acceptance test can pass for the host-scoped case.
TDD: new end-to-end test drives a real .replay with an empty decisions
array (ruling out reconciliation) through the real thread pair,
proving a genuine DISPATCH from live resource availability. 25
pre-existing white-box sim/repl test files given link-satisfying
stubs reproducing the OLD always-PEND contract unchanged -- their own
assertions are untouched (task 5.3's job to reconcile deeper).
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01RLSNJhNoULvVcUhXEFafQC
Relocates sched_job_picker()/sched_allocator() out of scheduler.c
(1382 lines, already over this project's 800-line cap) into a new
sched_election.c, retyped to xce_sched_core_t* -- the single
implementation both drivers will spawn (task 4.1 wires simulation
mode onto it). Independently verified bit-for-bit identical to the
pre-relocation bodies modulo the sched->/core-> rename; no statement
reordered, no branch changed.
sched_make_allocation() becomes extern (was static, sole caller moved
translation units); the old commit_decision() thin wrapper is deleted,
its sole caller now calls sched_commit_decision() directly with
identical semantics. sched_init() updated to spawn the relocated pair
against &(sched->core).
Adds an end-to-end fixture test proving a job flows from picked to a
committed decision against bare (non-xce_scheduler_t) core state.
Routes around, rather than fixes, two pre-existing out-of-boundary
gaps this fixture exposed (seqjob/parjob allocator init still reading
plugin->sched; make_sched_job()'s parse path not threading
maxtasks/mintasks) -- both documented in the test's own comments for
future follow-up.
sched_election.c added to the normal build only; sched_sim_isolation_lib
membership deferred to task 4.1, which is where sim mode actually
starts spawning these threads.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01RLSNJhNoULvVcUhXEFafQC
Migrates match_job_global_limit_()/limit_job_election()/
match_job_host_limit_() (plugins/limit.c) to read plugin->core (added
task 1.2) instead of deriving xce_scheduler_t* from plugin->sched,
which stays NULL for a plugin registered via plugin_register_core()
(simulation mode). Removes the dead ctx.sched write flagged during
task 2.1's review -- verified this ctx never has .sched read anywhere
downstream.
Fixes bug-1828 (found while writing this task's own TDD test for a
previously untested code path): match_job_host_limit_() had the same
key_find()-negative-rc-treated-as-fatal defect bug-1786 already fixed
in its global-limit sibling, but never applied here -- meaning
host-scoped resource limits were never actually enforced, in any
mode, until this fix (not merely a display gap; the limit's vres was
never added to obj->hlimits, so sched_fill_limit_requests() always
saw an empty list for it).
TDD: new tests drive the real hook-registered entry points against a
plugin_register_core()-registered plugin, asserting real usage
instances via limit_get_usage_instances() for both global and
host-scoped limits.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01RLSNJhNoULvVcUhXEFafQC
Retypes sched_make_allocation()/sched_make_primary_allocation()/
sched_make_compound_allocation()/sched_make_alter_allocation()/
sched_commit_decision(), plus 3 private helpers found necessary
during migration (sched_make_primary_alloc_(), sched_make_seqjob_
allocation(), sched_make_parjob_allocation()) from xce_scheduler_t*
to xce_sched_core_t* in scheduler.c.
Wires the one genuine normal-mode-only touch point -- the
unsatisfiable-resreq cache, inside sched_make_primary_alloc_() --
through the task-1.1 NULL-safe callback seam (core->unsatisfiable_
lookup_cb/populate_cb) instead of a direct field read. No other
normal-mode-only state was found in any of the 8 migrated functions,
independently confirmed by both the implementer and reviewer against
xce_scheduler_t's own field list -- no escalation needed.
Minimal &(sched->core) call-site wraps applied inside sched_allocator()/
commit_decision() (task 3.1's home, not yet migrated), matching the
pattern established in tasks 1.2/1.3/2.1.
TDD: two-part RED -- a compile-error RED from the sched_commit_decision
retype, and a behavioral RED using spy callbacks proving the cache
seam was genuinely unconsulted before this fix.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01RLSNJhNoULvVcUhXEFafQC
Retypes xce_sched_ctx_t.sched (sched/types.h) from void* to
xce_sched_core_t* outright, and sched_try_alloc_on_host_() (resource.c)
to take xce_sched_core_t* directly -- completing what tasks 1.2/1.3
left as temporary &(sched->core) call-site wraps. Removes the two
resulting struct-prefix casts entirely (schedobj.c's
sched_ctx_set_primary(), host_iter.c's sched_host_iter_next()): both
now pass core state through honestly, with no cast.
Updates every sched_ctx_init()/sched_try_alloc_on_host_() call site
tree-wide (scheduler.c, parjob_allocator.c, seqjob_allocator.c) with
zero change to host_iter/parjob/seqjob's own host-iteration/
bucket-construction algorithm.
Adds a driver-parity test for sched_try_alloc_on_host_() proving
identical grant/deny results from normal-mode-shaped and
simulation-mode-shaped core state.
Flags (for task 2.3) a confirmed-dead, type-confused write at
plugins/limit.c:319 (ctx.sched = plugin->sched) -- never read today,
left untouched since limit.c is task 2.3's boundary, not this one's.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01RLSNJhNoULvVcUhXEFafQC
Re-types sched_meet_host_selection(), sched_check_shared_satisfy()
(+its _ helper), sched_check_host_satisfy() (+its _ helper), and the
private unsatisfy_() helper they share, from xce_scheduler_t* to
xce_sched_core_t*, in resource.c.
Fixes every call site this forces (a C signature migration cannot
keep the build green otherwise): sched_try_alloc_on_host_() (2 sites,
already partially touched by task 1.2), host_iter.c's
sched_host_iter_next() (2 sites), and parjob_allocator.c/
seqjob_allocator.c's my_alloc() (1 site each) -- each fix is a minimal
&(sched->core) argument wrap, no algorithm change, preserving
host_iter/parjob/seqjob's own allocation algorithm exactly as-is.
Adds test_core_safe_allocation_chain_driver_parity, proving all three
functions produce identical grant/deny results whether given
normal-mode-shaped or simulation-mode-shaped core state.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01RLSNJhNoULvVcUhXEFafQC
Re-types plugin_job_election()/plugin_prepare_allocation()/
plugin_commit_decision() from (void* arg, ...) to (xce_sched_core_t*
core, ...), iterating core->plugins directly instead of deriving it
from a normal-mode-only driver struct -- mirroring
plugin_begin_session()/plugin_end_session()'s already-completed
migration.
Adds a new, honestly-typed xce_sched_plugin_t.core field, set by both
plugin_register() (normal mode) and plugin_register_core() (simulation
mode, closing a gap left since bug-1825 where this path set nothing).
The existing plugin->sched field and its plugins/fcfs.c legacy
consumer are left untouched.
Necessarily updates the 4 existing normal-mode call sites (scheduler.c,
resource.c, schedobj.c) to pass &(...->core) -- a C signature migration
cannot keep the build green without touching every caller in the same
commit; each fix is a minimal argument wrap with no algorithm change,
matching design.md's own Implementation Notes for this component.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01RLSNJhNoULvVcUhXEFafQC
Adds a NULL-by-default lookup/populate callback seam on xce_sched_core_t,
mirroring cache_clear_cb's proven shape, so the allocation dispatch
chokepoint can consult the unsatisfiable-resreq cache without a
normal-mode-only field access in shared code. Adapters live in
alloc_cache.c and are registered only during normal-mode startup;
simulation mode leaves both callbacks NULL and always computes fresh.
Adds the scheduler-sim-live-allocation spec (requirements/design/
research/tasks) as the planning artifact this and subsequent tasks
implement against.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01RLSNJhNoULvVcUhXEFafQC
info limit was correctly implemented and tested but inert against a real
replay: register_limit()'s only call site was normal-mode sched_register_
plugins(); sim mode's own plugin bootstrap (sim_bootstrap_picker()) never
registered the limit plugin. The snapshot format already carries the
original session's limit config (sim->core.config, a borrowed reference set
by xce_sched_core_load_snapshot()) -- nothing needed to change in the
recording/replay format.
New sim_bootstrap_limit_plugin() (sim_limit.c, mirroring sim_picker.c's
structure) registers the plugin via plugin_register_core() and seeds it via
a new limit_seed_config(plugin, xce_sched_core_t* core, config) entry point,
wired into sim_loop.c alongside the existing sim_bootstrap_picker() call.
limit_new_plugin() extracted from register_limit() as a factory, mirroring
builtin_fcfs_new_plugin()'s established pattern; register_limit() itself is
behavior-preserving.
First attempt set plugin->sched to a disguised xce_sched_sim_driver_t*,
relying on it sharing xce_scheduler_t's first-member layout to reuse the
existing config_hook path -- rejected on review: this repeats the exact
cross-cast pattern that already corrupted memory once in this codebase (see
core.h's xce_core_cache_clear_cb_t history). Fixed by making
limit_seed_config() take xce_sched_core_t* as an honest parameter instead,
with limit_update_config() (normal mode) and the new sim path both
delegating to a shared limit_seed_config_impl() -- no cast, no aliasing risk,
independently verified by two review rounds reading the actual struct
layouts and call chains.
Also added sim_limit.c to sched_sim_isolation_lib so Req 5 AC 17's
controller-I/O-isolation check covers it (an initial gap, closed in the same
remediation round). Two ripple waves of the same established link-stub
pattern applied to the ~24 white-box test files that #include sim_loop.c.
info limit now shows real configured-limit data against a genuine replay.
Live usage will still correctly show "no active usage instance yet" until a
separate, larger fix (logged as bug-1826) wires limit *matching* into the
sim decision path -- deliberately out of this task's scope.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01RLSNJhNoULvVcUhXEFafQC
Four new tests in test_repl_limit_table.c close out Requirement 5:
- test_info_limit_never_mutates_global_limit_config_or_accounting_state /
..._host_scoped_... : snapshot a matched limit's config fields and its
vres capacity/in-use zarray entries (both pointer identity and value)
before and after repeated info limit calls; asserts byte-identical state
and byte-identical repeated stdout output.
- test_info_limit_on_empty_registry_is_repeatable_and_never_registers_plugin:
confirms info limit never causes the plugin to become spuriously
registered as a side effect of being queried.
- test_repl_step_continue_with_info_limit_interleaved_matches_baseline_decision_sequence:
a step/continue sequence run twice (once with info limit interleaved, once
without) produces an identical decision sequence.
Verified via real mutation testing: three targeted fake mutations were
temporarily applied to repl_info.c, confirmed each new test fails exactly as
predicted, then cleanly reverted (repl_info.c shows zero diff from HEAD).
Zero production code changed -- this is a test-only validation task. Full
14-binary regression sweep + 78/78 meson aggregate confirm zero impact on
any pre-existing info subcommand (Requirement 5.3).
This closes out scheduler-sim-diagnostics' first capability (limit
inspection, Requirements 1-5) end to end.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01RLSNJhNoULvVcUhXEFafQC
repl_cmd_info() now recognizes "limit" as an info subcommand, routing to the
already-implemented and tested repl_info_limit() with the same ntok<3 usage
guard shape as the queue/worker/resource branches. Usage string and help
text updated to list it. repl_info_limit()'s own body is untouched -- this
is a pure, additive dispatch-wiring change. Two new tests drive it through
the real repl_dispatch() entry point (not a direct call) for both `info
limit <name>` and `info limit all`.
`info limit <name|all>` is now reachable from a live REPL session --
scheduler-sim-diagnostics' first capability is functionally complete.
Remaining: task 5 (determinism/regression validation).
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01RLSNJhNoULvVcUhXEFafQC
repl_print_limit_usage_per_host() fans a single host-scoped xce_vres_t out
across every worker with a populated capacity entry at that vres's shared
resource index (w->resvals[idx]/w->resInUse[idx]), rather than constructing
any per-host vres object -- matching design.md's (xce_vres_t, worker) pair
cardinality model exactly. A worker with no populated capacity entry produces
no line at all (not a fabricated zero); a worker with capacity but no
charged usage shows accounted=0. Uses the corrected idx<0 + .caps-only
bounds-check pattern task 3.1 established, avoiding bug-1824's latent
imprecision. This closes out all of Requirement 3 (3.1-3.4).
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01RLSNJhNoULvVcUhXEFafQC
repl_info_limit() now displays accounted-usage/capacity for GLOBAL/SHARED
limit instances: iterates limit_get_usage_instances(), filters by
vres->limit == matched_limit, reads capacity from core.sharedInst[idx] and
accounted usage from core.sharedInUse[idx]. Existence-of-instance and
zero-usage are kept distinct per design.md: no matching instance (or one
with no populated capacity) prints "no active usage instance yet"; a
matched-but-never-charged instance correctly shows accounted=0 instead.
Self-caught during implementation and independently confirmed by review: a
negative or stale vres->idx (possible per add_vres()'s allocate-then-assign
pattern) is not safely rejected by zarray_get_ex()'s idx>=caps check alone
-- the new code adds an explicit idx<0 guard rather than copying
repl_info_resource_one()'s existing .count-based pre-check (which is itself
imprecise for a sparsely-populated zarray, a pre-existing issue left
out of scope here).
Calling limit_get_usage_instances() for the first time from repl_info.c
reproduced the same ~24-file link-stub ripple task 2.1 established for
limit_get_configured_limits() -- fixed with the identical pattern.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01RLSNJhNoULvVcUhXEFafQC
repl_info_limit() now prints an explanatory note when a limit is implicitly
host-scoped (is_host_based_limit() > 0 due to a percent-based value on a
non-shared resource) but its own configuration doesn't show an explicit
per-host scope. The real discriminator turned out to be consumers[CONSUMER_
HOST]->per_consumer, not pointer-nullity as design.md's gloss said --
build_consumer() unconditionally allocates a consumer for all four
dimensions, so the struct field is never actually NULL in production. Fixed
using the same field is_host_based_limit() itself already falls back to;
verified via RED-phase reproduction that the literal design.md condition
never fires. design.md corrected to document the real condition.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01RLSNJhNoULvVcUhXEFafQC
repl_info_limit() now prints a matched limit's configured setting on both the
all and named-lookup success paths: fmt_limit() (existing libxce formatter,
unmodified) into a stack buffer, re-split and printed through a new
repl_print_limit_dump() helper matching this file's existing [sim]-prefix-
per-line convention. Inclusion and exclusion group membership both render
since fmt_group() already handles both, adopted as-is per design.md's
explicit reuse decision.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01RLSNJhNoULvVcUhXEFafQC
repl_info_limit() (repl_info.c, not yet wired into repl_cmd_info() -- that's
task 4) resolves "all" vs a specific name against limit_get_configured_limits():
a NULL/empty handle on the all path reports "no limits configured" (Req 1.7);
a NULL handle or key_find() miss on the named path reports "no such limit
'<name>'" (Req 1.3) and never collapses into the all-only message. New
white-box test_repl_limit_table.c exercises both paths directly, including a
dedicated regression guard for the NULL-scoping rule.
Making repl_info.c genuinely call limit_get_configured_limits() broke the link
for ~24 other white-box test files that #include repl_info.c without also
linking plugins/limit.c -- fixed with the same link-satisfying-stub idiom this
file family already uses elsewhere (all return NULL, verified behavior-
preserving since none of these fixtures register the limit plugin).
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01RLSNJhNoULvVcUhXEFafQC
limit_get_configured_limits()/limit_get_usage_instances() cross the limit
plugin's private-state boundary (plugin->data, previously unreachable outside
limit.c) via key_find on the name-keyed plugin registry, returning NULL when
the plugin is absent/uninitialized and a raw iterable kvmgr handle otherwise.
Pure reads only -- no allocation, mutation, or hook triggering. Foundation for
the upcoming `info limit` REPL command (task 1 of 5).
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01RLSNJhNoULvVcUhXEFafQC
sched_reconcile_host_resource_usage() -- the per-job mem-usage
reconciliation driven by JOB_RUSAGE events -- was missing the same
IS_WORKER_UNAVAIL/IS_WORKER_RECONCILING guard bug-1801 already added to
sched_update_account(), the regular per-round charge/uncharge path.
core_reset_workers() (bug-1751) deliberately freezes a RECONCILING
worker's resInUse each round so a clean rebuild can restore it once the
worker's real state is confirmed; bug-1801 extended that freeze to the
charge path, but this third, independent resInUse-mutating path kept
applying deltas regardless. Combined with core_events.c's JOBS_PUSH
resync-merge RUN branch, which can replace a job's whole account object
(fresh used=0.0 baseline) without uncharging the old one first -- safe
only because the next clean rebuild is supposed to correct it, exactly
the rebuild that stays skipped while RECONCILING -- a rusage tick
landing in that window kept adding a full recharge on top of whatever
was already frozen in the pool, with nothing to ever subtract it back
out until the worker escaped RECONCILING for good or the job finished.
Live-confirmed via a user's own multi-session replay trace: a job's
"mem used" figure never self-corrected until the job finished.
Added the identical guard, mirroring bug-1801 verbatim. RED confirmed
by temporarily disabling it (exactly the new test failed, "Expected 0
Was 5", 76 others green); GREEN confirmed restored (77/77).
bug-1815, follow-up to bug-1814/bug-1801.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01RLSNJhNoULvVcUhXEFafQC
sched_on_job_array() expanded a compact array-job push (up to
hundreds of elements: a per-element dedup lookup plus a full deep
clone) inline, synchronously, on whatever vtaskmgr thread handled the
incoming request -- the same shared pool the scheduler's own
controller connection traffic depends on, violating ha-async-io.md's
"bounded handler work only" contract for that pool (a project-wide
design invariant). A prior finding already showed a burst of
6x50-element array pushes (300 job events landing within ~2ms of each
other) causing the scheduler to lose its controller connection within
245ms and enter a reconnect/failover loop -- directly consistent with
this mechanism, and the same reconnect storm whose resync races
motivated the companion JOBS_PUSH fix in this branch.
The request handler now does O(1) work only: validate the descriptor
and enqueue it unexpanded (XCE_SCHED_OP_UPDATE_JOB_ARRAY, mirroring
the existing XCE_SCHED_OP_ADMIN_RECORD deferred-dispatch pattern).
The actual per-element expansion and dedup check moved to
translate_one_pending_update(), which runs on the round-processing
thread (main.c's xce_sched_driver_run_normal_mode_round(), before
xce_core_begin_session() each round) -- a thread with no connection-
liveness responsibility. As a side effect this also closes the
dedup check's own pre-existing cross-thread TOCTOU race (previously
documented as bug-1714 follow-up "Bug B"): only one thread ever runs
it now.
test_job_array_expand.c and test_job_array_wire_e2e.c, which asserted
the old inline-expansion behavior, are updated to assert the new
deferred-enqueue behavior instead; expansion, dedup, and full
drain-push-release ownership coverage moved to new tests in
test_driver_translate.c.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01RLSNJhNoULvVcUhXEFafQC
A resync redelivery (on_scheduler_join()'s full push, fired on every
scheduler reconnect) can race a job's own very-recent dispatch
confirmation and carry a stale, non-terminal state for it. The
duplicate-jobid merge in core_events.c trusted this claim
unconditionally, freeing the job's already-captured allocation and
regressing its state -- live-confirmed via a production scheduler.log
showing a resync WARN landing 16ms after the job's own RUN
confirmation, and matching a live-reported simulator symptom (a RUN
job silently missing from info worker's allocation listing).
Refuse a non-terminal claim when the existing job already has a real
allocation on record, mirroring the existing phantom-RUN guard's own
reasoning in the opposite direction; a genuinely terminal claim
(DONE/EXIT) still applies and releases the allocation correctly, same
as before.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01RLSNJhNoULvVcUhXEFafQC
xce_core_push_events() WARNed on every JOBS_PUSH resync redelivery
(sched_add_job rc=-EEXIST), which is the expected outcome under
on_scheduler_join()'s full resync or the array-push dedup race
(bug-1734/bug-1714) and flooded operational logs. Only that case now
logs via a new DBG() macro (utils.h); any other add_rc (e.g. -ENOMEM)
still WARNs. DBG is named to avoid colliding with the build's own
-DDEBUG command-line define (debug/asan buildtypes).
Also collapsed hash_string()'s duplicated djb2 loop to delegate to
hash_block(), which already implements the same RUN_HASH loop.
Logged as bug-1809 in .wolf/buglog.json.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01RLSNJhNoULvVcUhXEFafQC
- bug-1799/1801: fix resInUse over-commit in sim replay -- sched_account_jobs's
per-round re-derive pass now skips RECONCILING/UNAVAIL workers the same way
core_reset_workers already does (bug-1751), closing the gap that let a
worker's used slots exceed its own capacity across a resync replay
- bug-1804: revise scheduler-sim-eager-reconciliation spec (user-directed) so
a captured job's job->state is set to RUN synchronously at dispatch capture,
mirroring sched_commit_decision's normal-mode semantics (bug-1624), instead
of deferring until its separately-recorded JOB_STATE_CHANGE event replays
- bug-1800: render bytes-valued CONSUMABLE resources (mem/swp/tmp) with the
same human-scaled unit formatting already used for maxmem/maxswap/maxtmp
- bug-1802/1803: add then revert a break-on-ANY-event breakpoint kind after
live feedback that it over-fires inside busy sessions; replace it with a
prescan-based "next non-dummy session" hint in info summary instead
Full scheduler suite: 77/77 tests pass (asan profile).
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01RLSNJhNoULvVcUhXEFafQC
The previous fix for the "slots" regression (5b18f14a) worked but was an
unneeded special case: strcmp(vres->res->name, "slots") to fall back to
nslots directly. alloc_req->host already holds exactly that value during
the global phase -- both allocators' own unconditional "add slots" step
stages it there regardless of phase, and plugin_prepare_allocation()
never resets .host during the global phase, so by the time sched_fill_
limit_requests() runs (right after, same invocation), .host contains
exactly "slots"'s job-wide total and nothing else.
Read alloc_req->host FIRST for this combination and fall back to the
rusages-based on-demand computation only when nothing is staged there
(get_resval_by_id()'s "not found" == 0.0, safe here since "slots" is
never legitimately 0 -- seqjob only runs for maxTasks==1, parjob only
for maxTasks>1). Data-driven instead of name-based, matching bug-1789's
own earlier flag-based-dispatch philosophy, and generalizes for free to
any future resource an allocator plugin similarly pre-stages outside
its normal rusages loop.
test_fill_limit_requests_global_limit_on_slots_reads_host now pre-stages
alloc_req->host exactly as the allocators do, rusages left empty,
matching production exactly.
Full xce-build.sh all asan --incremental: 94/94 Ok. Live re-verified
against both the original gitea#46 scenario and the ug1_global_mem
scenario: byte-identical scheduler traces to every prior verification
round.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GLfB5uWRdmx2WiZSFC7Fm7
The limit-plugin-ownership refactor (fa34b038) broke gitea#46's exact
original scenario. calc_wanted_job_total_resval() scans the job's
rusages by name, but "slots" is never a declared rusage entry --
seqjob_prepare_allocation()/parjob_prepare_allocation() derive its want
directly from the job's own task count via their own unconditional
key_find("slots", ...) + store, entirely outside the rusages loop. The
refactor's on-demand lookup therefore always missed for "slots",
silently leaving the computed want at 0 -- reintroducing bug-1789's
exact original symptom (matching worked, but caps never bound).
Live-confirmed against a fresh deploy of gitea#46's own repro (9 plain
sleep jobs, user1, queue normal, no -R flag): limit1/ug1_slots/
ug1_normal_slots all matched correctly but zero unsatisfy_ lines
appeared and all 9 jobs dispatched uncapped. The refactor's own new
unit test had masked this -- it artificially added a "slots" rusages
entry that doesn't reflect production reality.
Fix: sched_fill_limit_requests() now falls back to `nslots` directly
(the same value both allocators already store as "slots"'s own
per-host want) whenever the rusages lookup misses and the resource is
named "slots" -- the one documented exception to the rusages-lookup
rule, matching this codebase's existing "slots" special-casing
convention.
New test test_fill_limit_requests_global_limit_on_slots_uses_nslots
proves the fallback with no "slots" entry in rusages at all. The
original test reworked to use "mem" instead, so it now genuinely
exercises the general on-demand path.
Full xce-build.sh all asan --incremental: 94/94 Ok. Live re-verified
against both the original gitea#46 scenario (byte-identical to the
pre-refactor trace) and the ug1_global_mem scenario -- both correct.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GLfB5uWRdmx2WiZSFC7Fm7
bug-1790's fix worked but put limit-plugin-specific logic in the wrong
place: seqjob_prepare_allocation()/parjob_prepare_allocation() (the
allocator plugins) unconditionally pre-computed a job-wide total for
every per-host resource in the job's rusages, "just in case" some
global limit needed one. That's the allocator plugins' concern leaking
into limit evaluation -- duplicated across both allocators (a third
would need it too), and wasteful (computed for resources no limit
actually references).
Moved ownership to sched_fill_limit_requests() (resource.c), called
once centrally from plugin_prepare_allocation() regardless of which
allocator ran:
- Removed calc_wanted_host_resval_job_totals() and its calls from both
allocator plugins entirely.
- Added calc_wanted_job_total_resval() (resreq.c): an on-demand,
single-resource lookup by name (no resmgr/key_find needed -- the
caller already has the resolved xce_resource_t), computing the same
job-wide-total formula only for the one (scope, underlying-resource)
combination that needs it -- a GLOBAL limit on a per-host resource.
The other three combinations are unchanged, still reading straight
out of alloc_req.
- sched_fill_limit_requests() gains rusages/nhosts/nslots parameters
back (a genuine need this time, not a re-added unused one).
plugin.c's two call sites pass obj->u.primary->resreq->rusages
(NULL for COMPOUND/ALTER, matching this codebase's existing scope
convention for those tags) and ctx->nhosts/ctx->nslots.
All 23 testsuite link-satisfying stubs mechanically updated. Removed
the two allocator-level regression tests from the first pass (the
behavior they proved no longer exists by design); rewrote
test_fill_limit_requests_global_limit_on_host_resource_reads_host to
exercise the new on-demand path directly, nslots=3 to prove the job's
WHOLE total is used.
Full xce-build.sh all asan --incremental: 94/94 Ok, zero regressions.
Live re-verified against the same ug1_global_mem deploy/scenario:
byte-identical scheduler trace to the pre-refactor result, confirming
behavior-preserving.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GLfB5uWRdmx2WiZSFC7Fm7
test_job_rusage_reconciles_mem_reservation_against_usage asserted
w->resInUse was already charged 4.0 for a JOBS_PUSH-dispatched RUN job,
immediately after the push with no round boundary simulated. That
assumption was wrong, not the scheduler: resInUse/sharedInUse is never
charged incrementally at JOBS_PUSH/dispatch time. It's fully re-derived
once per round by sched_account_jobs() (scheduler.c), which iterates
every currently-RUN job and calls sched_charge_account() on each one's
accounts -- called by both the real normal-mode driver (driver.c) and
the sim driver (sim_loop.c), always before anything else in that round
reads resInUse. Confirmed by reading the full JOBS_PUSH handler
(core_events.c) and builtin_fcfs_on_job_push(): neither ever charges or
uncharges resInUse, by design -- core_session.c's own doc comment on
core_begin_session_internal() already documents this "recompute from
source of truth each round" architecture.
Fix: the test now calls sched_charge_account(&core, tracked->accounts)
right after JOBS_PUSH, simulating the one round-boundary reconciliation
step sched_account_jobs() would perform for its single RUN job (that
function itself lives in scheduler.c, not white-box #included here, so
its single-job effect is reproduced directly via the same primitive it
calls internally). No production code changed.
Full scheduler suite: 77/77 Ok (was 76/77). Full xce-build.sh all asan
--incremental: 94/94 Ok, zero regressions.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GLfB5uWRdmx2WiZSFC7Fm7
Closes the gap bug-1789 explicitly carved out: a GLOBAL limit built on a
per-host resource other than "slots" (e.g. a future absolute global mem
cap) needs ctx->alloc_req.host to carry that resource's job-wide total
want during the global phase, but only "slots" ever got that treatment
(via seqjob/parjob_prepare_allocation()'s own unconditional store).
Adds calc_wanted_host_resval_job_totals() (resreq.c), sharing its loop
body with calc_wanted_global_resvals() via a new static helper
parameterized only by which XCE_RESF_GLOBAL direction to match -- it
computes the SAME job-wide-total formulas (PER_TASK: value*nslots, PER_JOB:
value, PER_HOST: value*nhosts) for non-global resources instead of global
ones. seqjob/parjob_prepare_allocation()'s global-phase branch now calls
it into ctx->alloc_req.host alongside the existing calc_wanted_global_
resvals() call into .shared.
Safety was traced, not assumed: plugin_prepare_allocation() only resets
.host at the top of a per-host call, never during the global phase, and
sched_fill_limit_requests() for glimits reads .host later in the SAME
global-phase invocation -- transferring the value into the limit's
permanent .shared record before any later per-host call ever resets
.host. An earlier candidate fix (reusing calc_wanted_host_resvals()
directly) was rejected during design: its PER_HOST formula and, for
parjob, its per-bucket nslots don't give a job-wide total.
New regression tests in test_seqjob_allocator.c/test_parjob_
combinatorial.c prove a non-"slots" per-host resource ("mem") now gets
correctly populated; parjob's test uses maxTasks=3 specifically to prove
the formula uses the job's whole nslots, not a bucket's share.
Full suite (xce-build.sh all asan --incremental): 92/94, both failures
pre-existing and unrelated (bug-1762, bug-1733's already-documented
xcealloc.c leak). Not live-verified -- no limits.yaml entry exercises
this path yet.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GLfB5uWRdmx2WiZSFC7Fm7
sched_fill_limit_requests() used to read a limit's underlying resource's
want from the same vector its own virtual-resource entry gets written to.
That's wrong whenever the two differ in scope -- e.g. a GLOBAL limit built
on top of a per-host resource like "slots" (limits.yaml's
ug1_normal_slots): the limit's own vres entry belongs in alloc_req->shared,
but "slots"'s want is only ever staged into alloc_req->host by
seqjob/parjob_prepare_allocation(), so the old code always read a zero and
silently never enforced the cap.
Now both the read source and write destination are derived independently
from each resource's own XCE_RESF_GLOBAL flag (vres->res for the source,
vres->vres for the destination) -- the same flag
calc_wanted_global_resvals()/calc_wanted_host_resvals() (resreq.c) already
dispatch on, so this is the existing rule applied consistently rather than
a new one. Also drops sched_fill_limit_requests()'s unused xce_scheduler_t*
parameter.
Full scheduler suite: 76/77 (the one failure, test_job_rusage_reconciles_
mem_reservation_against_usage, is bug-1762, pre-existing and unrelated).
Live-verified against a fresh single-node deploy.
Filed bug-1790 to track a known, deliberately out-of-scope residual gap:
a future global limit on a per-host resource other than "slots" would need
calc_wanted_host_resvals() to also run during the global phase for that
resource, which nothing currently does.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GLfB5uWRdmx2WiZSFC7Fm7
While live-verifying gitea#46, found and fixed a cascade of 5 previously
undetected bugs that together made every usergroup-scoped limit in
limits.yaml silently non-functional:
- Named usergroups/hostgroups (users.yaml/hosts.yaml) were never
registered into the scheduler's group managers (groupmgr_add() had zero
production callers), so any limit referencing one by name could never
resolve real membership.
- An unconfigured limit scope (e.g. no "projects:") arrived as an
empty-but-present capnp group, not NULL, so build_consumer() built a
real empty group (matching no one) instead of falling back to "all".
- Three separate call sites conflated key_find()'s negative "not found"
return with a real error, short-circuiting before ever reaching the
"instantiate on first use" logic right below them -- the same defect
class already fixed once before in this codebase (bug-1495).
- Two call sites passed &(sched->core.resmgr) where a plain handle was
expected, one level of indirection too many, causing RTTI's key_find()
to reject it as an invalid handle.
All five are live-verified: fresh cluster deploys now correctly match and
instantiate usage tracking for all 7 limits in limits.yaml on every job
submission, reproduced across multiple independent from-scratch redeploys.
A 6th, architecturally distinct issue remains open (bug-1789, logged, not
fixed): sched_fill_limit_requests() computes a job's consumption of a
slots-scoped global limit by reading the wrong request vector, so the
numeric cap itself still never binds. This needs a design decision on
intended cross-scope semantics, not a mechanical fix, so it's left for a
follow-up.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Hb6Edtj4KbS8CrgEWYdA2E
sim_reconcile_unaccounted_jobs() (the mechanism that replays historical
DISPATCH decisions in --sim mode) reconstructed a job's allocation but
never called the picker's on_dispatch() hook, unlike the live allocator
path which always does. The job therefore stayed a live candidate in the
picker's own pendJobs bookkeeping, and sim_picker_thread_fn() (which
unconditionally PENDs whatever it is still offered) kept re-offering and
re-PENDing it every subsequent round -- live-observed as an already-
dispatched job receiving a bogus PEND decision (bug-1759: 22/50 jobs in
one batch; bug-1779: 5/10 array-job elements).
This directly contradicts scheduler-sim-decision-parity's own design.md,
which already documented "the picker never re-decides an already-resolved
job" as the correct assumption -- the implementation had drifted from it.
Fixed by calling on_dispatch() immediately after reconstructing the job's
allocation, mirroring sched_allocator()'s own live-dispatch ordering.
test_sim_verify_no_false_mismatch_across_round_separated_dispatch's own
fixture had baked in the stale-PEND symptom as "expected" data; updated to
match the corrected behavior (dispatch-only round 1, empty filler rounds).
Both design docs updated in place with correction notes.
info worker's assigned-jobs listing gated only on job->allocs != NULL,
so a job already at a terminal state (DONE/EXIT) could still appear as
"assigned" if its allocs pointer was ever left stale -- live-observed via
a real captured replay, three DONE jobs kept showing as assigned to a
worker despite the worker's own slot-usage accounting correctly excluding
them. Root cause of the stale allocs itself is not yet pinned (suspected
snapshot save/load inconsistency); this adds the same defensive "terminal
is permanent" exclusion job_state_regresses_terminal()/sim_reconcile_
unaccounted_jobs() already apply elsewhere, at the display layer.
