chore(buglog): harness batch fixed (da1a663f): bug-1574/1702/1703/1705/1707-1710/1712/1777/1793/1796/1797
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
This commit is contained in:
parent
da1a663f05
commit
8cadeeb2d0
GPG key ID:
80533536B0EB651C
14 changed files with 221 additions and 156 deletions
|
|
@ -1,16 +1,17 @@
|
|||
---
|
||||
id: "bug-1574"
|
||||
status: "open"
|
||||
status_updated: "2026-09-24"
|
||||
status_note: "xce_test resupdate string/type crash: set_resval() lacks svalue case, vals[] uninitialized (src/harness/test.c)"
|
||||
status: "fixed"
|
||||
status_updated: "2026-09-28"
|
||||
status_note: ""
|
||||
component: "controller"
|
||||
severity: "crash"
|
||||
timestamp: "2026-08-17T18:55:00Z"
|
||||
last_seen: "2026-09-19T09:00:00Z"
|
||||
occurrences: 1
|
||||
files: ["src/controller/testsuite/test.c"]
|
||||
files: ["src/harness/test.c", "src/harness/resval_typeresolve.c"]
|
||||
tags: ["resupdate", "cmd_resupdate", "set_resval", "segfault", "uninitialized-stack-memory", "Value_u_svalue", "test.c", "pre-existing", "resource-value-verification"]
|
||||
related: ["bug-1570", "bug-1573"]
|
||||
fix_commits: ["da1a663f"]
|
||||
legacy: true
|
||||
---
|
||||
|
||||
|
|
@ -29,10 +30,12 @@ _Not recorded (migrated from buglog.json)._
|
|||
Pre-existing, predates the resource-value-verification feature entirely (confirmed via git show on the pre-task-2.1 baseline commit 98b8c34b^: the original set_resval()'s switch(valtype) already had no case for Value_u_svalue, same shape as today). set_resval() only assigns vals[idx].name inside a case block of switch(valtype); Value_u_svalue (type/model) has no case, so for these two names vals[idx].name is never assigned and keeps whatever uninitialized garbage was already on the local stack array `xce_resval_t vals[MAX_N_RES]` (never zero-initialized) in cmd_resupdate(). That garbage pointer is then passed unconditionally into xcecli_update_resvals()'s wire encoder inside &valptrs[0], which (most likely) dereferences it as a C string, causing the segfault. Task 2.1's own comment at test.c (near set_resval, 'the switch falls through with vals[idx].name/.val->which left unset ... intentionally not fixed by this task, see CONCERNS') already flagged this exact gap and deliberately preserved it byte-for-byte per Requirement 1.4, rather than introducing it. Discovered live during resource-value-verification task 4.2's validation (apiserver trace-log addition, live-testing the string-value code path with `resupdate -n model -v CustomModelX`) -- unrelated to task 4.2's own change (src/frontend/apiserver/backend/handler.go), which only touches Go/apiserver code.
|
||||
|
||||
## Fix
|
||||
NOT FIXED -- out of boundary for task 4.2 (scope: src/frontend/apiserver/backend/handler.go only). Belongs to test.c/set_resval()/resval_typeresolve.c territory (tasks 2.1/2.2 of this same spec, or a dedicated follow-up). Worked around during task 4.2 live validation by using a numeric resource (r15m) for the primary trace-log confirmation and avoiding type/model live resupdate calls thereafter. Likely fix: zero-initialize `xce_resval_t vals[MAX_N_RES]` in cmd_resupdate() (e.g. `= {0}`) so an unassigned .name is NULL instead of garbage, PLUS give set_resval() a real Value_u_svalue case that sets vals[idx].name and vals[idx].val->u.svalue (currently the string value passed to resupdate for type/model is silently never stored in the union either -- a second, related but distinct defect covered by the same missing switch case).
|
||||
set_resval() moved out of test.c as resval_fill() in src/harness/resval_typeresolve.c, which fills name, host and value for every type, including Value_u_svalue (a copied string; "" when no -v). cmd_resupdate() also zero-initializes vals[] and frees each entry afterwards (resval_fill_release()).
|
||||
|
||||
## Verification
|
||||
_Not recorded (migrated from buglog.json)._
|
||||
- Test: test_fill_string_resource_sets_name_and_value, test_fill_numeric_resource, test_fill_string_override_without_value_is_empty_string in src/harness/testsuite/test_resval_typeresolve.c (the first pre-fills the struct with 0xA5 like the old uninitialized stack array); the functions did not exist before the fix, and the old set_resval() had no svalue case at all
|
||||
- Suites: harness 6/6 (asan): harness_unit_test 141/0, cmdutil 14/0, resval_typeresolve 7/0, joblist_display 22/0, join_worker_args 7/0, listrow_json 12/0
|
||||
- Live: n/a; xce_test is a client tool and every case above runs without a controller
|
||||
|
||||
## References
|
||||
- Related: bug-1570, bug-1573
|
||||
|
|
|
|||
|
|
@ -1,16 +1,17 @@
|
|||
---
|
||||
id: "bug-1702"
|
||||
status: "open"
|
||||
status_updated: "2026-09-24"
|
||||
status_note: "xce_test jobrun -m host strncpy into 256B stack buffer unbounded (src/harness/test.c)"
|
||||
status: "fixed"
|
||||
status_updated: "2026-09-28"
|
||||
status_note: ""
|
||||
component: "controller"
|
||||
severity: "unknown"
|
||||
timestamp: "2026-08-26T09:32:18Z"
|
||||
last_seen: "2026-09-19T09:00:00Z"
|
||||
occurrences: 1
|
||||
files: ["src/controller/testsuite/test.c:2130-2136 (cmd_jobrun)"]
|
||||
files: ["src/harness/test.c"]
|
||||
tags: ["controller", "testsuite", "xce_test", "cmd_jobrun", "buffer-overflow", "memory-safety", "cli-ux-audit"]
|
||||
related: ["bug-1703"]
|
||||
fix_commits: ["da1a663f"]
|
||||
legacy: true
|
||||
---
|
||||
|
||||
|
|
@ -29,10 +30,12 @@ _Not recorded (migrated from buglog.json)._
|
|||
`strncpy(buf, optarg, (int)(p - optarg))` uses the attacker/user-controlled substring length directly with no `MIN(..., sizeof(buf)-1)` clamp, unlike sibling code elsewhere in the file that does bound similar copies.
|
||||
|
||||
## Fix
|
||||
NOT YET FIXED. Recommended: clamp the copy length to sizeof(buf)-1 (or reject a host-name segment longer than the buffer with a usage error) before strncpy.
|
||||
jobrun's -m now goes through cmdutil_split_host_tasks() (src/harness/cmdutil.c), which refuses a host segment that does not fit the 256-byte buffer (-ENAMETOOLONG → "host name in -m ... is longer than 255 characters") and an ntasks that is not an integer >= 1.
|
||||
|
||||
## Verification
|
||||
_Not recorded (migrated from buglog.json)._
|
||||
- Test: test_jobrun_overlong_host_rejected_without_overflow in src/harness/testsuite/test_test_cli.c: RED before the fix (ASAN stack-buffer-overflow, WRITE of size 300, xce_test died), GREEN after; plus the split_host_tasks cases in src/harness/testsuite/test_cmdutil.c (a canary after the host buffer stays intact)
|
||||
- Suites: harness 6/6 (asan): harness_unit_test 141/0, cmdutil 14/0, resval_typeresolve 7/0, joblist_display 22/0, join_worker_args 7/0, listrow_json 12/0
|
||||
- Live: n/a; xce_test is a client tool and every case above runs without a controller
|
||||
|
||||
## References
|
||||
- Related: bug-1703
|
||||
|
|
|
|||
|
|
@ -1,16 +1,17 @@
|
|||
---
|
||||
id: "bug-1703"
|
||||
status: "open"
|
||||
status_updated: "2026-09-24"
|
||||
status_note: "xce_test jobsub -n min strncpy into 64B stack buffer unbounded (src/harness/test.c)"
|
||||
status: "fixed"
|
||||
status_updated: "2026-09-28"
|
||||
status_note: ""
|
||||
component: "controller"
|
||||
severity: "unknown"
|
||||
timestamp: "2026-08-26T09:32:18Z"
|
||||
last_seen: "2026-09-19T09:00:00Z"
|
||||
occurrences: 1
|
||||
files: ["src/controller/testsuite/test.c:2246-2254 (cmd_jobsub)"]
|
||||
files: ["src/harness/test.c"]
|
||||
tags: ["controller", "testsuite", "xce_test", "cmd_jobsub", "buffer-overflow", "memory-safety", "cli-ux-audit"]
|
||||
related: ["bug-1702"]
|
||||
fix_commits: ["da1a663f"]
|
||||
legacy: true
|
||||
---
|
||||
|
||||
|
|
@ -29,10 +30,12 @@ _Not recorded (migrated from buglog.json)._
|
|||
Same pattern as bug-1702: `strncpy(buf, optarg, (int)(p - optarg))` with no bound check against sizeof(buf).
|
||||
|
||||
## Fix
|
||||
NOT YET FIXED. Recommended: clamp the copy length to sizeof(buf)-1 (or reject) before strncpy, same fix shape as bug-1702.
|
||||
jobsub's -n now goes through cmdutil_parse_ntasks(), which parses min and max in place (no copy into a buffer) and refuses anything but 1 <= min <= max, with a usage error.
|
||||
|
||||
## Verification
|
||||
_Not recorded (migrated from buglog.json)._
|
||||
- Test: test_jobsub_overlong_ntasks_rejected_without_overflow in src/harness/testsuite/test_test_cli.c: RED before the fix (ASAN stack-buffer-overflow, WRITE of size 100), GREEN after; plus the parse_ntasks cases in src/harness/testsuite/test_cmdutil.c
|
||||
- Suites: harness 6/6 (asan): harness_unit_test 141/0, cmdutil 14/0, resval_typeresolve 7/0, joblist_display 22/0, join_worker_args 7/0, listrow_json 12/0
|
||||
- Live: n/a; xce_test is a client tool and every case above runs without a controller
|
||||
|
||||
## References
|
||||
- Related: bug-1702
|
||||
|
|
|
|||
|
|
@ -1,16 +1,17 @@
|
|||
---
|
||||
id: "bug-1705"
|
||||
status: "open"
|
||||
status_updated: "2026-09-24"
|
||||
status_note: "xce_test jobsub -x exdata overwritten by build_job_exdata and leaked"
|
||||
status: "fixed"
|
||||
status_updated: "2026-09-28"
|
||||
status_note: ""
|
||||
component: "controller"
|
||||
severity: "leak"
|
||||
timestamp: "2026-08-26T09:32:18Z"
|
||||
last_seen: "2026-09-19T09:00:00Z"
|
||||
occurrences: 1
|
||||
files: ["src/controller/testsuite/test.c:2243-2245 (-x parsing)", ":2181-2195 (build_job_envs unconditional overwrite)", ":2336 (unconditional call site)", "USAGE_JOBSUB at :484-485"]
|
||||
files: ["src/harness/test.c"]
|
||||
tags: ["controller", "testsuite", "xce_test", "cmd_jobsub", "dead-flag", "silent-data-loss", "memory-leak", "cli-ux-audit"]
|
||||
related: []
|
||||
fix_commits: ["da1a663f"]
|
||||
legacy: true
|
||||
---
|
||||
|
||||
|
|
@ -29,10 +30,12 @@ _Not recorded (migrated from buglog.json)._
|
|||
build_job_exdata()/build_job_envs() have no guard for an already-non-NULL exdata field before overwriting it.
|
||||
|
||||
## Fix
|
||||
NOT YET FIXED. Recommended: skip build_job_exdata() when submit.start_info->exdata is already non-NULL (from -x), and free the old pointer before any legitimate reassignment.
|
||||
The default exdata comes from cmdutil_default_exdata(), which only builds the PATH JSON when exdata is still NULL, so -x's value is kept and no longer leaked; a repeated -x frees the earlier copy. cmd_jobsub's cleanup also frees execuser/exechome/execcwd, which leaked on every call.
|
||||
|
||||
## Verification
|
||||
_Not recorded (migrated from buglog.json)._
|
||||
- Test: test_default_exdata_keeps_caller_value and test_default_exdata_builds_path_json_when_unset in src/harness/testsuite/test_cmdutil.c; the old build_job_exdata() overwrote exdata unconditionally, which the first test forbids
|
||||
- Suites: harness 6/6 (asan): harness_unit_test 141/0, cmdutil 14/0, resval_typeresolve 7/0, joblist_display 22/0, join_worker_args 7/0, listrow_json 12/0
|
||||
- Live: n/a; xce_test is a client tool and every case above runs without a controller
|
||||
|
||||
## References
|
||||
_Not recorded (migrated from buglog.json)._
|
||||
|
|
|
|||
|
|
@ -1,16 +1,17 @@
|
|||
---
|
||||
id: "bug-1707"
|
||||
status: "open"
|
||||
status_updated: "2026-09-24"
|
||||
status_note: "xce_test resupdate trailing -n without -v silently dropped"
|
||||
status: "fixed"
|
||||
status_updated: "2026-09-28"
|
||||
status_note: ""
|
||||
component: "controller"
|
||||
severity: "unknown"
|
||||
timestamp: "2026-08-26T09:32:18Z"
|
||||
last_seen: "2026-09-19T09:00:00Z"
|
||||
occurrences: 1
|
||||
files: ["src/controller/testsuite/test.c:3551-3561 (trailing-pair flush", "comment admits the drop)", ":3502-3507 (mid-list warning that trailing case lacks)", ":3567 (unconditional success message)"]
|
||||
files: ["src/harness/test.c"]
|
||||
tags: ["controller", "testsuite", "xce_test", "cmd_resupdate", "silent-data-loss", "cli-ux-audit"]
|
||||
related: ["bug-1706"]
|
||||
fix_commits: ["da1a663f"]
|
||||
legacy: true
|
||||
---
|
||||
|
||||
|
|
@ -29,10 +30,12 @@ _Not recorded (migrated from buglog.json)._
|
|||
Trailing pending -n/-v pair flush logic only runs when resval_str != NULL, with no fallback warning path like the mid-list case has.
|
||||
|
||||
## Fix
|
||||
NOT YET FIXED. Recommended: apply the same "no value specified, default will be used" warning (or reject outright) to the trailing pair instead of a silent drop.
|
||||
A trailing -n without -v now gets the same warning and default value as a mid-list one (resupdate_add()). cmd_resupdate also returns the RPC's failure instead of 0, refuses an empty update, and bounds the 16-entry vals[] array (a 17th -n overflowed it).
|
||||
|
||||
## Verification
|
||||
_Not recorded (migrated from buglog.json)._
|
||||
- Test: test_resupdate_trailing_name_without_value_warns and test_resupdate_rpc_failure_fails_the_session in src/harness/testsuite/test_test_cli.c: RED before the fix (no warning; exit 0 after the RPC failed), GREEN after
|
||||
- Suites: harness 6/6 (asan): harness_unit_test 141/0, cmdutil 14/0, resval_typeresolve 7/0, joblist_display 22/0, join_worker_args 7/0, listrow_json 12/0
|
||||
- Live: n/a; xce_test is a client tool and every case above runs without a controller
|
||||
|
||||
## References
|
||||
- Related: bug-1706
|
||||
|
|
|
|||
|
|
@ -1,16 +1,17 @@
|
|||
---
|
||||
id: "bug-1708"
|
||||
status: "open"
|
||||
status_updated: "2026-09-24"
|
||||
status_note: "xce_test jobkill -s non-numeric -> signal 0, still reports success"
|
||||
status: "fixed"
|
||||
status_updated: "2026-09-28"
|
||||
status_note: ""
|
||||
component: "controller"
|
||||
severity: "unknown"
|
||||
timestamp: "2026-08-26T09:32:18Z"
|
||||
last_seen: "2026-09-19T09:00:00Z"
|
||||
occurrences: 1
|
||||
files: ["src/controller/testsuite/test.c:2572-2574 (signo = atoi(optarg))", ":2601 (unconditional success message)"]
|
||||
files: ["src/harness/test.c"]
|
||||
tags: ["controller", "testsuite", "xce_test", "cmd_jobkill", "silent-failure", "cli-ux-audit"]
|
||||
related: []
|
||||
fix_commits: ["da1a663f"]
|
||||
legacy: true
|
||||
---
|
||||
|
||||
|
|
@ -29,10 +30,12 @@ _Not recorded (migrated from buglog.json)._
|
|||
No validation that -s's argument is numeric/nonzero before using it and reporting success.
|
||||
|
||||
## Fix
|
||||
NOT YET FIXED. Recommended: reject non-numeric/unparseable -s values with a clear error instead of silently defaulting to signal 0.
|
||||
jobkill's -s is parsed with cmdutil_parse_long() and must be a signal number 1..NSIG-1; anything else is a usage error before any RPC.
|
||||
|
||||
## Verification
|
||||
_Not recorded (migrated from buglog.json)._
|
||||
- Test: test_jobkill_non_numeric_signal_rejected in src/harness/testsuite/test_test_cli.c: RED before the fix (-s SIGKILL accepted and sent), GREEN after; plus the parse_long cases in src/harness/testsuite/test_cmdutil.c
|
||||
- Suites: harness 6/6 (asan): harness_unit_test 141/0, cmdutil 14/0, resval_typeresolve 7/0, joblist_display 22/0, join_worker_args 7/0, listrow_json 12/0
|
||||
- Live: n/a; xce_test is a client tool and every case above runs without a controller
|
||||
|
||||
## References
|
||||
_Not recorded (migrated from buglog.json)._
|
||||
|
|
|
|||
|
|
@ -1,16 +1,17 @@
|
|||
---
|
||||
id: "bug-1709"
|
||||
status: "open"
|
||||
status_updated: "2026-09-24"
|
||||
status_note: "xce_test jobinfo returns 0 on malformed/missing jobid"
|
||||
status: "fixed"
|
||||
status_updated: "2026-09-28"
|
||||
status_note: ""
|
||||
component: "controller"
|
||||
severity: "unknown"
|
||||
timestamp: "2026-08-26T09:32:18Z"
|
||||
last_seen: "2026-09-19T09:00:00Z"
|
||||
occurrences: 1
|
||||
files: ["src/controller/testsuite/test.c:2029-2058 (cmd_jobinfo)"]
|
||||
files: ["src/harness/test.c"]
|
||||
tags: ["controller", "testsuite", "xce_test", "cmd_jobinfo", "silent-failure", "exit-code", "cli-ux-audit"]
|
||||
related: []
|
||||
fix_commits: ["da1a663f"]
|
||||
legacy: true
|
||||
---
|
||||
|
||||
|
|
@ -29,10 +30,12 @@ _Not recorded (migrated from buglog.json)._
|
|||
No error/found flag tracked across the -j loop; the function unconditionally `return 0`s regardless of whether any valid lookup succeeded.
|
||||
|
||||
## Fix
|
||||
NOT YET FIXED. Recommended: track an error/found flag across the loop and return nonzero (with a usage message) when no valid -j was given or every lookup failed.
|
||||
cmd_jobinfo counts the -j values given and the lookups that succeeded: a malformed jobid prints "invalid jobid", no -j prints "no jobid given", and the command fails unless at least one lookup succeeded.
|
||||
|
||||
## Verification
|
||||
_Not recorded (migrated from buglog.json)._
|
||||
- Test: test_jobinfo_without_valid_jobid_fails in src/harness/testsuite/test_test_cli.c: RED before the fix (exit 0 for both `jobinfo` and `jobinfo -j garbage`), GREEN after
|
||||
- Suites: harness 6/6 (asan): harness_unit_test 141/0, cmdutil 14/0, resval_typeresolve 7/0, joblist_display 22/0, join_worker_args 7/0, listrow_json 12/0
|
||||
- Live: n/a; xce_test is a client tool and every case above runs without a controller
|
||||
|
||||
## References
|
||||
_Not recorded (migrated from buglog.json)._
|
||||
|
|
|
|||
|
|
@ -1,16 +1,17 @@
|
|||
---
|
||||
id: "bug-1710"
|
||||
status: "open"
|
||||
status_updated: "2026-09-24"
|
||||
status_note: "xce_test hello -c <=0 reports vacuous success"
|
||||
status: "fixed"
|
||||
status_updated: "2026-09-28"
|
||||
status_note: ""
|
||||
component: "controller"
|
||||
severity: "unknown"
|
||||
timestamp: "2026-08-26T09:32:18Z"
|
||||
last_seen: "2026-09-19T09:00:00Z"
|
||||
occurrences: 1
|
||||
files: ["src/controller/testsuite/test.c:1341 (rc=0 init)", ":1351 (count=atoi(optarg)", "unchecked)", ":1367 (while count>0 loop)", ":1373 (return rc)"]
|
||||
files: ["src/harness/test.c"]
|
||||
tags: ["controller", "testsuite", "xce_test", "cmd_hello", "silent-failure", "cli-ux-audit"]
|
||||
related: []
|
||||
fix_commits: ["da1a663f"]
|
||||
legacy: true
|
||||
---
|
||||
|
||||
|
|
@ -29,10 +30,12 @@ _Not recorded (migrated from buglog.json)._
|
|||
No validation that count is >=1 before entering the loop; the initial success value is never overwritten when the loop body never executes.
|
||||
|
||||
## Fix
|
||||
NOT YET FIXED. Recommended: reject count<=0 with an explicit error instead of falling through to a silent vacuous success.
|
||||
hello's -c is parsed with cmdutil_parse_long() and must be >= 1; 0, negative or non-numeric is a usage error.
|
||||
|
||||
## Verification
|
||||
_Not recorded (migrated from buglog.json)._
|
||||
- Test: test_hello_nonpositive_count_rejected in src/harness/testsuite/test_test_cli.c: RED before the fix (exit 0, nothing sent), GREEN after
|
||||
- Suites: harness 6/6 (asan): harness_unit_test 141/0, cmdutil 14/0, resval_typeresolve 7/0, joblist_display 22/0, join_worker_args 7/0, listrow_json 12/0
|
||||
- Live: n/a; xce_test is a client tool and every case above runs without a controller
|
||||
|
||||
## References
|
||||
_Not recorded (migrated from buglog.json)._
|
||||
|
|
|
|||
|
|
@ -1,16 +1,17 @@
|
|||
---
|
||||
id: "bug-1712"
|
||||
status: "open"
|
||||
status_updated: "2026-09-24"
|
||||
status_note: "xce_test resupdate duplicate -v gives misleading message"
|
||||
status: "fixed"
|
||||
status_updated: "2026-09-28"
|
||||
status_note: ""
|
||||
component: "controller"
|
||||
severity: "unknown"
|
||||
timestamp: "2026-08-26T09:32:18Z"
|
||||
last_seen: "2026-09-19T09:00:00Z"
|
||||
occurrences: 1
|
||||
files: ["src/controller/testsuite/test.c:3525-3531 (case 'v' else-branch)"]
|
||||
files: ["src/harness/test.c"]
|
||||
tags: ["controller", "testsuite", "xce_test", "cmd_resupdate", "misleading-error-message", "cli-ux-audit"]
|
||||
related: ["bug-1706", "bug-1707"]
|
||||
fix_commits: ["da1a663f"]
|
||||
legacy: true
|
||||
---
|
||||
|
||||
|
|
@ -29,10 +30,12 @@ _Not recorded (migrated from buglog.json)._
|
|||
The else-branch of the -v case handles two distinct situations (no preceding -n at all, vs. a duplicate -v for an already-named pending resource) with one shared, incorrectly-worded message.
|
||||
|
||||
## Fix
|
||||
NOT YET FIXED. Recommended: distinguish the two cases with an accurate message for the duplicate-value case, e.g. "ignoring duplicate value '%s' for res '%s' (already has a pending value)".
|
||||
The -v branch now tells the two cases apart: no preceding -n keeps "no res name found"; a second value for the same resource prints "ignoring duplicate value `8' for res `slots' (it already has value `4')".
|
||||
|
||||
## Verification
|
||||
_Not recorded (migrated from buglog.json)._
|
||||
- Test: test_resupdate_duplicate_value_named_correctly in src/harness/testsuite/test_test_cli.c: RED before the fix ("no res name found"), GREEN after
|
||||
- Suites: harness 6/6 (asan): harness_unit_test 141/0, cmdutil 14/0, resval_typeresolve 7/0, joblist_display 22/0, join_worker_args 7/0, listrow_json 12/0
|
||||
- Live: n/a; xce_test is a client tool and every case above runs without a controller
|
||||
|
||||
## References
|
||||
- Related: bug-1706, bug-1707
|
||||
|
|
|
|||
|
|
@ -1,16 +1,17 @@
|
|||
---
|
||||
id: "bug-1777"
|
||||
status: "open"
|
||||
status_updated: "2026-09-24"
|
||||
status_note: "xce_test workerlist MaxSwap ignores hosts.yaml override"
|
||||
status: "fixed"
|
||||
status_updated: "2026-09-28"
|
||||
status_note: ""
|
||||
component: "controller"
|
||||
severity: "unknown"
|
||||
timestamp: "2026-09-06T21:10:00Z"
|
||||
last_seen: "2026-09-19T09:00:00Z"
|
||||
occurrences: 1
|
||||
files: ["src/controller/testsuite/test.c"]
|
||||
files: ["src/harness/test.c", "src/harness/joblist_display.c"]
|
||||
tags: ["controller", "worker", "resource", "maxswap", "workerlist", "config-override", "deferred"]
|
||||
related: ["bug-1554"]
|
||||
fix_commits: ["da1a663f"]
|
||||
legacy: true
|
||||
---
|
||||
|
||||
|
|
@ -29,10 +30,12 @@ _Not recorded (migrated from buglog.json)._
|
|||
CONFIRMED via code trace (not yet reproduced against a live pc13v/replay artifact -- none existed on disk at investigation time). xce_worker_t carries two separate copies of a host-attrs-style resource: attrs->maxswap (hostattrs_t, written ONLY by xcecd.c's init_host_attrs() at join time -- rejoin_worker(), worker.c:440, just re-copies the same join-time value) and resvals[idx("maxswap")] (resmgr-indexed zarray, seeded from attrs->maxswap by add_hostattrs_resvals() at apply_join_state() start, worker.c, then OVERWRITTEN by mark_host_res()/update_val_res() if a hosts.yaml resource override exists for that host, worker.c:79 `array = &(w->host.resvals)`). cmd_workerlist() (test.c:1478) reads attrs->maxswap directly -- structurally incapable of ever reflecting a config override, since nothing writes an override back into attrs. The scheduler-facing path (collect_resvals_workers() -> resvals_to_sched -> xcesched_push_workers(), dataex.c/sched.c) reads resvals, which DOES pick up the override. So whenever a maxswap override is configured, the scheduler correctly uses it (and a replay/snapshot reflects it) while xce_test workerlist permanently shows the stale join-time value regardless of timing/staleness -- not a caching/race artifact, a structural gap. The Go apiserver/xcecli path (convert.go) also just casts attrs.maxswap with no override-awareness, and xcecli's hosts.yaml model doesn't even expose a MaxSwap column today.
|
||||
|
||||
## Fix
|
||||
NOT FIXED -- user explicitly deferred (2026-09-06) to prioritize bug-1554 (real swap/tmp hardware detection, now fixed) and the assigned-jobs slot-count display improvement first. Two candidate fix directions discussed: (a) make cmd_workerlist()/the Go apiserver path read the effective resmgr/resvals value instead of attrs directly, or (b) make mark_host_res() write the override back into attrs too, so every attrs consumer (maxcpus/maxmem/maxswap/maxtmp alike) stays in sync at the source -- (b) was the recommended direction since it fixes the whole class rather than patching each display path individually, but neither is implemented yet.
|
||||
Direction (a): workerlist (table and --json) shows maxmem/maxswap/maxtmp from the worker's effective resource values, which the workerlist reply already carries in resvals_to_sched (collect_resvals_workers(), controller/dataex.c) and which include a hosts.yaml override, falling back to the join-time attrs when a value is absent or not numeric (worker_effective_attr(), src/harness/joblist_display.c). Harness only; the Go apiserver's worker view is not changed here.
|
||||
|
||||
## Verification
|
||||
_Not recorded (migrated from buglog.json)._
|
||||
- Test: test_effective_attr_prefers_resval and test_effective_attr_falls_back_to_join_value in src/harness/testsuite/test_joblist_display.c; new function, the old code read attrs->maxswap directly
|
||||
- Suites: harness 6/6 (asan): harness_unit_test 141/0, cmdutil 14/0, resval_typeresolve 7/0, joblist_display 22/0, join_worker_args 7/0, listrow_json 12/0
|
||||
- Live: n/a; xce_test is a client tool and every case above runs without a controller
|
||||
|
||||
## References
|
||||
- Related: bug-1554
|
||||
|
|
|
|||
|
|
@ -1,16 +1,17 @@
|
|||
---
|
||||
id: "bug-1793"
|
||||
status: "open"
|
||||
status_updated: "2026-09-24"
|
||||
status_note: "xce_test join maxswap/maxtmp defaults are stale byte values (join_worker_args.c)"
|
||||
status: "fixed"
|
||||
status_updated: "2026-09-28"
|
||||
status_note: ""
|
||||
component: "controller"
|
||||
severity: "unknown"
|
||||
timestamp: "2026-09-09T16:24:19Z"
|
||||
last_seen: "2026-09-19T09:00:00Z"
|
||||
occurrences: 1
|
||||
files: ["src/controller/testsuite/join_worker_args.c (join_worker_parse_attrs", "hostattrs_t default initialization)"]
|
||||
files: ["src/harness/join_worker_args.c"]
|
||||
tags: ["controller", "testsuite", "xce_test", "cmd_join_worker", "stale-default", "units", "deferred"]
|
||||
related: ["bug-1557"]
|
||||
fix_commits: ["da1a663f"]
|
||||
legacy: true
|
||||
---
|
||||
|
||||
|
|
@ -29,10 +30,12 @@ _Not recorded (migrated from buglog.json)._
|
|||
Same root cause as bug-1557: maxswap/maxtmp were reasonable defaults (1KB/2KB) back when these fields were a smaller unit, but an earlier session converted maxswap/maxtmp to bytes-canonical (confirmed via resources.yaml 'max swap/tmp disk space in bytes' and hardware_monitor.c's raw-byte sysinfo()/statvfs() computations) without updating this test-tool default. 1024 bytes of swap / 2048 bytes of tmp are just as nonsensical as the old maxmem=64 bytes.
|
||||
|
||||
## Fix
|
||||
NOT FIXED -- deliberately deferred (xce-test-cli-ux task 2.6). USAGE_JOIN_WORKER's help text currently documents these exact stale values accurately, so fixing only the code without also fixing the docs would create a new code/docs mismatch outside task 2.6's Requirement-12.1-scoped mandate (which covers maxmem only). Recommended: a coordinated fix changing both defaults (e.g. maxswap to a byte-scale value matching typical swap sizing, maxtmp similarly) AND USAGE_JOIN_WORKER's documented defaults, together. A regression test (test_join_worker_sibling_defaults_unchanged in test_join_worker_args.c) currently locks in today's stale values specifically so a future fix here doesn't collide with bug-1557's maxmem-specific test assertions -- update/remove that test's maxswap/maxtmp assertions when this bug is actually fixed.
|
||||
join's maxswap/maxtmp defaults are now 1 GiB / 2 GiB in bytes (the intended 1024/2048 MB), help text and example updated to bytes; -m/-s/-t are parsed with strtod() and validated, since atoi() overflowed above 2 GB.
|
||||
|
||||
## Verification
|
||||
_Not recorded (migrated from buglog.json)._
|
||||
- Test: test_join_worker_swap_tmp_defaults_are_bytes, test_join_worker_byte_values_above_int_range and test_join_worker_non_numeric_byte_value_rejected in src/harness/testsuite/test_join_worker_args.c: all three RED against the old join_worker_args.c (maxswap 1024; 8 GB parsed as 0; 'lots' accepted), GREEN after
|
||||
- Suites: harness 6/6 (asan): harness_unit_test 141/0, cmdutil 14/0, resval_typeresolve 7/0, joblist_display 22/0, join_worker_args 7/0, listrow_json 12/0
|
||||
- Live: n/a; xce_test is a client tool and every case above runs without a controller
|
||||
|
||||
## References
|
||||
- Related: bug-1557
|
||||
|
|
|
|||
|
|
@ -1,16 +1,17 @@
|
|||
---
|
||||
id: "bug-1796"
|
||||
status: "open"
|
||||
status_updated: "2026-09-24"
|
||||
status_note: "qctrl/hostctrl perror() prints stale errno instead of RPC rc (src/harness/test.c)"
|
||||
status: "fixed"
|
||||
status_updated: "2026-09-28"
|
||||
status_note: ""
|
||||
component: "controller"
|
||||
severity: "unknown"
|
||||
timestamp: "2026-09-09T22:00:00Z"
|
||||
last_seen: "2026-09-19T09:00:00Z"
|
||||
occurrences: 1
|
||||
files: ["src/controller/testsuite/test.c (cmd_qctrl", "cmd_hostctrl", "cmd_jobkill RPC-error return paths)"]
|
||||
files: ["src/harness/test.c"]
|
||||
tags: ["controller", "testsuite", "xce_test", "cmd_qctrl", "cmd_hostctrl", "cmd_jobkill", "perror", "errno", "misleading-diagnostic", "xce-test-cli-ux", "task-4.2", "pre-existing", "deferred"]
|
||||
related: []
|
||||
fix_commits: ["da1a663f"]
|
||||
legacy: true
|
||||
---
|
||||
|
||||
|
|
@ -29,10 +30,12 @@ _Not recorded (migrated from buglog.json)._
|
|||
Each of these commands does `if (rc < 0) { perror("xcecli_..."); return rc; }` where `rc` is the RPC's own negative-errno-style return value (e.g. a clean -ENOENT for "no such queue/host"), but perror() always reads the CURRENT process errno, not `rc` -- and nothing in this branch ever does `errno = -rc` first. Found during xce-test-cli-ux task 4.2's live confirmation-gate verification: a bypassed qctrl/hostctrl RPC against a deliberately bogus (nonexistent) queue/host name printed different, unrelated-looking error text across different runs of the SAME scenario, root-caused (and independently reproduced 3 times by the task 4.2 reviewer) to stale errno left over from prior, unrelated syscalls/signal-decline paths in the same process -- e.g. a destructive-command Ctrl-C decline explicitly sets errno = EINTR at its own call site and never resets it, so a LATER unrelated RPC failure in the same session then misreports itself as "Interrupted system call". Confirmed pre-existing (commit 41f56fa6e, 2025-11-05), predating the whole xce-test-cli-ux spec; NOT a defect in the interruptible-call adapter (tasks 1.4/3.2) or the confirmation gate (task 2.1/3.3) -- both were independently confirmed to behave correctly in every case tested, this is purely a misleading-diagnostic-text bug.
|
||||
|
||||
## Fix
|
||||
NOT YET FIXED (found and deliberately deferred -- out of boundary for task 4.2, a pure live-verification task with no source changes permitted). Recommended: set `errno = -rc;` immediately before each `perror(...)` call on these RPC-error paths (or replace with `fprintf(stderr, "...: %s\n", strerror(-rc));` to avoid the errno indirection entirely), and audit other cmd_* handlers for the same perror()-without-errno-set pattern.
|
||||
Every RPC-failure path in test.c now reports through cmdutil_rpc_error(what, rc), which prints the text for rc itself (libc errno or E_XCE_* via xce_strerror_r()) and never reads the global errno; 19 sites, including qctrl, hostctrl, jobkill, jobckpt, jobinfo, joblist, setpend, heartbeat and the ones that used SET_ERROR(-rc) + perror() (which printed "Unknown error 100xx" for XCE codes).
|
||||
|
||||
## Verification
|
||||
_Not recorded (migrated from buglog.json)._
|
||||
- Test: test_rpc_error_reports_rpc_code_not_stale_errno in src/harness/testsuite/test_test_cli.c: RED before the fix (with no controller qctrl printed "Connection refused" and jobkill "Invalid argument" for the same failure; piped, both "Inappropriate ioctl for device"), GREEN after (both print the RPC's own -EINVAL); plus test_errstr_* in src/harness/testsuite/test_cmdutil.c (a stale errno is ignored; E_XCE_BAD_QUEUE reads "unknown queue")
|
||||
- Suites: harness 6/6 (asan): harness_unit_test 141/0, cmdutil 14/0, resval_typeresolve 7/0, joblist_display 22/0, join_worker_args 7/0, listrow_json 12/0
|
||||
- Live: n/a; xce_test is a client tool and every case above runs without a controller
|
||||
|
||||
## References
|
||||
_Not recorded (migrated from buglog.json)._
|
||||
|
|
|
|||
|
|
@ -1,16 +1,17 @@
|
|||
---
|
||||
id: "bug-1797"
|
||||
status: "open"
|
||||
status_updated: "2026-09-24"
|
||||
status_note: "cmd_qctrl leaks strdup'd queue name on RPC error (src/harness/test.c)"
|
||||
status: "fixed"
|
||||
status_updated: "2026-09-28"
|
||||
status_note: ""
|
||||
component: "controller"
|
||||
severity: "leak"
|
||||
timestamp: "2026-09-09T22:00:00Z"
|
||||
last_seen: "2026-09-19T09:00:00Z"
|
||||
occurrences: 3
|
||||
files: ["src/controller/testsuite/test.c (cmd_qctrl)"]
|
||||
files: ["src/harness/test.c"]
|
||||
tags: ["controller", "testsuite", "xce_test", "cmd_qctrl", "leak", "asan", "xce-test-cli-ux", "task-4.2", "pre-existing", "deferred"]
|
||||
related: ["bug-1796"]
|
||||
fix_commits: ["da1a663f"]
|
||||
legacy: true
|
||||
---
|
||||
|
||||
|
|
@ -29,10 +30,12 @@ _Not recorded (migrated from buglog.json)._
|
|||
cmd_qctrl does `queue = strdup(optarg)` while parsing -q, then on `if (rc < 0) { perror("xcecli_queuectrl"); return rc; }` (the RPC-error branch) returns without `free(queue)` -- only the success path frees it. Distinct from the adjacent, intentional non-free on the EINTR/abandoned-worker path (ownership legitimately transferred there per the interruptible-call adapter's contract). Found during xce-test-cli-ux task 4.2's live confirmation-gate verification (a bypassed qctrl against a bogus queue name triggered this exact RPC-error branch under the asan build). Confirmed pre-existing (commit 41f56fa6e, 2025-11-05), predating the whole xce-test-cli-ux spec and untouched by task 3.1's flagspec migration; not caused by any change in this spec.
|
||||
|
||||
## Fix
|
||||
NOT YET FIXED (found and deliberately deferred -- out of boundary for task 4.2, a pure live-verification task with no source changes permitted). Recommended: add `free(queue);` on the RPC-error return path (or restructure with a single cleanup label), and check `cmd_hostctrl`/`cmd_jobkill`'s sibling strdup'd-argument fields for the same pattern.
|
||||
cmd_qctrl frees the queue name on the RPC-error path and when the call allocation fails, and a repeated -q frees the earlier copy (hostctrl's -h likewise). The interrupted-call path still hands the name to the worker thread, as before.
|
||||
|
||||
## Verification
|
||||
_Not recorded (migrated from buglog.json)._
|
||||
- Test: LeakSanitizer run of the fixed xce_test (`qctrl -q bogus -a` with no controller, ASAN_OPTIONS=detect_leaks=1): no leak with a cmd_qctrl frame; an automated test is not practical because the harness suites run with detect_leaks=0 and xce_test leaks elsewhere at exit
|
||||
- Suites: harness 6/6 (asan): harness_unit_test 141/0, cmdutil 14/0, resval_typeresolve 7/0, joblist_display 22/0, join_worker_args 7/0, listrow_json 12/0
|
||||
- Live: n/a; xce_test is a client tool and every case above runs without a controller
|
||||
|
||||
## References
|
||||
- Related: bug-1796
|
||||
|
|
|
|||
Loading…
Add table
Add a link
Reference in a new issue