DAOS-16311 control: Add C bindings for control API by mjmac · Pull Request #17550 · daos-stack/daos

mjmac · 2026-02-11T20:42:13Z

Use Go's c-shared-library mode to build a .so that
exposes a C API for the Go Control API. The initial
use case for this library is to replace the problematic
dmg helpers test library that uses fork/exec to allow
tests to drive dmg commands.

Signed-off-by: Michael MacDonald github@macdonald.cx

github-actions · 2026-02-11T20:53:53Z

Ticket title is 'C bindings for DAOS control API'
Status is 'In Progress'
Labels: 'ci_2.6_daily,ci_2.6_provider,ci_master_daily,ci_master_provider'
https://daosio.atlassian.net/browse/DAOS-16311

daosbuild3 · 2026-02-11T22:20:52Z

Test stage Unit Test on EL 8.8 completed with status UNSTABLE. https://jenkins-3.daos.hpc.amslabs.hpecorp.net/job/daos-stack/job/daos//view/change-requests/job/PR-17550/2/testReport/

daosbuild3 · 2026-02-11T22:46:49Z

Test stage Unit Test with memcheck on EL 8.8 completed with status UNSTABLE. https://jenkins-3.daos.hpc.amslabs.hpecorp.net/job/daos-stack/job/daos//view/change-requests/job/PR-17550/2/testReport/

daosbuild3 · 2026-02-11T23:56:25Z

Test stage Unit Test bdev with memcheck on EL 8.8 completed with status UNSTABLE. https://jenkins-3.daos.hpc.amslabs.hpecorp.net/job/daos-stack/job/daos//view/change-requests/job/PR-17550/2/testReport/

daosbuild3 · 2026-02-12T01:33:42Z

Test stage Test RPMs on EL 8.6 completed with status FAILURE. https://jenkins-3.daos.hpc.amslabs.hpecorp.net//job/daos-stack/job/daos/view/change-requests/job/PR-17550/2/execution/node/1026/log

daosbuild3 · 2026-02-12T02:18:30Z

Test stage Functional on EL 8.8 completed with status FAILURE. https://jenkins-3.daos.hpc.amslabs.hpecorp.net/job/daos-stack/job/daos/job/PR-17550/2/display/redirect

daosbuild3 · 2026-02-12T02:18:36Z

Test stage Functional on EL 8.8 completed with status FAILURE. https://jenkins-3.daos.hpc.amslabs.hpecorp.net/job/daos-stack/job/daos/job/PR-17550/2/display/redirect

daosbuild3 · 2026-02-12T19:02:26Z

Test stage Unit Test on EL 8.8 completed with status UNSTABLE. https://jenkins-3.daos.hpc.amslabs.hpecorp.net/job/daos-stack/job/daos//view/change-requests/job/PR-17550/3/testReport/

daosbuild3 · 2026-02-12T19:19:32Z

Test stage Unit Test with memcheck on EL 8.8 completed with status UNSTABLE. https://jenkins-3.daos.hpc.amslabs.hpecorp.net/job/daos-stack/job/daos//view/change-requests/job/PR-17550/3/testReport/

daosbuild3 · 2026-02-12T20:41:47Z

Test stage Unit Test bdev with memcheck on EL 8.8 completed with status UNSTABLE. https://jenkins-3.daos.hpc.amslabs.hpecorp.net/job/daos-stack/job/daos//view/change-requests/job/PR-17550/3/testReport/

daosbuild3 · 2026-02-22T19:24:43Z

Test stage Unit Test on EL 8.8 completed with status UNSTABLE. https://jenkins-3.daos.hpc.amslabs.hpecorp.net/job/daos-stack/job/daos//view/change-requests/job/PR-17550/4/testReport/

daosbuild3 · 2026-02-22T19:47:42Z

Test stage Unit Test with memcheck on EL 8.8 completed with status UNSTABLE. https://jenkins-3.daos.hpc.amslabs.hpecorp.net/job/daos-stack/job/daos//view/change-requests/job/PR-17550/4/testReport/

daosbuild3 · 2026-02-22T21:09:02Z

Test stage Unit Test bdev with memcheck on EL 8.8 completed with status UNSTABLE. https://jenkins-3.daos.hpc.amslabs.hpecorp.net/job/daos-stack/job/daos//view/change-requests/job/PR-17550/4/testReport/

daosbuild3 · 2026-02-22T21:14:23Z

Test stage Test RPMs on EL 8.6 completed with status FAILURE. https://jenkins-3.daos.hpc.amslabs.hpecorp.net//job/daos-stack/job/daos/view/change-requests/job/PR-17550/4/execution/node/1040/log

daosbuild3 · 2026-02-22T21:43:36Z

Test stage Functional on EL 8.8 completed with status FAILURE. https://jenkins-3.daos.hpc.amslabs.hpecorp.net/job/daos-stack/job/daos/job/PR-17550/4/display/redirect

daosbuild3 · 2026-02-22T21:43:41Z

Test stage Functional on EL 8.8 completed with status FAILURE. https://jenkins-3.daos.hpc.amslabs.hpecorp.net/job/daos-stack/job/daos/job/PR-17550/4/display/redirect

daosbuild3 · 2026-02-27T19:19:59Z

Test stage Unit Test bdev with memcheck on EL 8.8 completed with status UNSTABLE. https://jenkins-3.daos.hpc.amslabs.hpecorp.net/job/daos-stack/job/daos//view/change-requests/job/PR-17550/5/testReport/

daosbuild3 · 2026-02-27T19:20:52Z

Test stage Unit Test with memcheck on EL 8.8 completed with status UNSTABLE. https://jenkins-3.daos.hpc.amslabs.hpecorp.net/job/daos-stack/job/daos//view/change-requests/job/PR-17550/5/testReport/

daosbuild3 · 2026-02-27T19:36:14Z

Test stage Unit Test on EL 8.8 completed with status UNSTABLE. https://jenkins-3.daos.hpc.amslabs.hpecorp.net/job/daos-stack/job/daos//view/change-requests/job/PR-17550/5/testReport/

daosbuild3 · 2026-02-27T20:55:20Z

Test stage Unit Test on EL 8.8 completed with status UNSTABLE. https://jenkins-3.daos.hpc.amslabs.hpecorp.net/job/daos-stack/job/daos//view/change-requests/job/PR-17550/6/testReport/

daosbuild3 · 2026-02-27T21:20:28Z

Test stage Unit Test with memcheck on EL 8.8 completed with status UNSTABLE. https://jenkins-3.daos.hpc.amslabs.hpecorp.net/job/daos-stack/job/daos//view/change-requests/job/PR-17550/6/testReport/

daosbuild3 · 2026-02-27T23:12:43Z

Test stage Unit Test bdev with memcheck on EL 8.8 completed with status UNSTABLE. https://jenkins-3.daos.hpc.amslabs.hpecorp.net/job/daos-stack/job/daos//view/change-requests/job/PR-17550/6/testReport/

daosbuild3 · 2026-02-27T23:19:59Z

Test stage Test RPMs on EL 8.6 completed with status FAILURE. https://jenkins-3.daos.hpc.amslabs.hpecorp.net//job/daos-stack/job/daos/view/change-requests/job/PR-17550/6/execution/node/1047/log

daosbuild3 · 2026-02-27T23:49:19Z

Test stage Functional on EL 8.8 completed with status FAILURE. https://jenkins-3.daos.hpc.amslabs.hpecorp.net/job/daos-stack/job/daos/job/PR-17550/6/display/redirect

daosbuild3 · 2026-02-27T23:49:24Z

Test stage Functional on EL 8.8 completed with status FAILURE. https://jenkins-3.daos.hpc.amslabs.hpecorp.net/job/daos-stack/job/daos/job/PR-17550/6/display/redirect

daosbuild3 · 2026-05-31T13:56:09Z

Test stage Functional Hardware Medium MD on SSD completed with status FAILURE. https://jenkins-3.daos.hpc.amslabs.hpecorp.net//job/daos-stack/job/daos/view/change-requests/job/PR-17550/41/execution/node/1284/log

daosbuild3 · 2026-06-01T01:54:50Z

Test stage Functional Hardware Medium MD on SSD completed with status UNSTABLE. https://jenkins-3.daos.hpc.amslabs.hpecorp.net/job/daos-stack/job/daos//view/change-requests/job/PR-17550/42/testReport/

daosbuild3 · 2026-06-01T07:25:24Z

Test stage Functional Hardware Medium Verbs Provider MD on SSD completed with status UNSTABLE. https://jenkins-3.daos.hpc.amslabs.hpecorp.net/job/daos-stack/job/daos//view/change-requests/job/PR-17550/42/testReport/

daosbuild3 · 2026-06-02T15:30:51Z

Test stage NLT completed with status UNSTABLE. https://jenkins-3.daos.hpc.amslabs.hpecorp.net/job/daos-stack/job/daos//view/change-requests/job/PR-17550/43/testReport/

daosbuild3 · 2026-06-03T00:10:38Z

Test stage Functional on EL 9 completed with status FAILURE. https://jenkins-3.daos.hpc.amslabs.hpecorp.net//job/daos-stack/job/daos/view/change-requests/job/PR-17550/43/execution/node/982/log

daosbuild3 · 2026-06-04T15:40:06Z

Test stage NLT completed with status UNSTABLE. https://jenkins-3.daos.hpc.amslabs.hpecorp.net/job/daos-stack/job/daos//view/change-requests/job/PR-17550/44/testReport/

kjacque

Overall I like the approach. I gave it a fairly superficial look over, my comments are mostly on the tests.

I don't like the callProductionCode type test wrappers if they aren't really saving us a lot of repetition--they make it harder to understand what the tests are doing, since I have to keep jumping between files to remember what the wrapper is hiding.

kjacque · 2026-06-04T22:19:35Z

+	if rc == 0 {
+		callFini(handle)
+		t.Fatal("expected error for non-existent config file, got success")
+	}


We can probably expect a specific error rc here, right?

In general these test cases might be a good fit for table-formatted tests.

kjacque · 2026-06-04T22:22:28Z

+func recoverExport(rc *C.int) {
+	if r := recover(); r != nil {
+		fmt.Fprintf(os.Stderr, "panic in libdaos_control export: %v\n%s\n", r, debug.Stack())
+		*rc = C.int(daos.MiscError)


Might be worth creating a new error code for this scenario.

kjacque · 2026-06-04T22:34:39Z

+/**
+ * Arguments for daos_control_pool_create.
+ */
+struct daos_control_pool_create_args {


Wonder if this would be better to split into two structs, with in-args and out-args.

kjacque · 2026-06-04T22:38:19Z

+/**
+ * Maximum number of interactive action choices per check report. The server
+ * currently caps this at CHK_INTERACT_OPTION_MAX (src/chk/chk_internal.h);
+ * this value must be >= the server cap. check.go truncates extra choices if
+ * the server ever sends more than fit here.
+ */
+#define DAOS_CHECK_MAX_ACT_OPTIONS  4


Looks like we don't have the advantage of a file that sees both of these, if it's coming from an internal header. I don't love having to maintain something like this in both places. Is there a public file the internal definition could be moved to?

kjacque · 2026-06-04T22:39:29Z

+ * DAOS checker pool information.
+ */
+struct daos_check_pool_info {
+	uuid_t dcpi_uuid;


Some structs are using this prefix style, while others just have bare names for member variables. IMO we should stick to one or the other.

kjacque · 2026-06-04T22:53:32Z

+	return errorToRC(validatePoolCreateArgs(&args))
+}
+
+func callPoolDestroy(handle cgo.Handle, poolUUID uuid.UUID, force bool) int {


I don't love how these helpers hide the function actually being tested. Can't most of these go into the body of the tests? Or at the very least, the test file where they're relevant?

kjacque · 2026-06-04T22:55:14Z

+func TestSystemExcludeRank(t *testing.T) {
+	for name, tc := range map[string]struct {
+		mic   *control.MockInvokerConfig
+		rank  uint32
+		expRC int
+	}{
+		"success": {
+			mic: &control.MockInvokerConfig{
+				UnaryResponse: control.MockMSResponse("host1", nil, &mgmtpb.SystemExcludeResp{
+					Results: []*sharedpb.RankResult{
+						{Rank: 0, State: "excluded"},
+					},
+				}),
+			},
+			rank:  0,
+			expRC: 0,
+		},
+	} {


Going to add an error case for this one?

kjacque · 2026-06-04T22:57:24Z

+	} {
+		t.Run(name, func(t *testing.T) {
+			if got := testCopyStringToCharArray(tc.input, tc.bufSize); got != tc.want {
+				t.Fatalf("got %q, want %q", got, tc.want)
+			}
+		})
+	}
+}
+
+func TestCopyStringToCharArrayNilDest(t *testing.T) {
+	testCopyStringToNil()
+}


I like helper functions but this is a little painful to read and understand. Are these helpers even re-used anywhere else?

kjacque · 2026-06-04T23:05:01Z

+		rpcFn := func(rpcCtx context.Context, conn *grpc.ClientConn) (proto.Message, error) {
+			if mgmtSvc != 0 {
+				return mgmtpb.NewMgmtSvcClient(conn).FaultInjectMgmtPoolFault(rpcCtx, payload)
+			}
+			return mgmtpb.NewMgmtSvcClient(conn).FaultInjectPoolFault(rpcCtx, payload)
+		}
+		_, err = control.InvokeFaultRPC(ctx.ctx(), ctx.client, rpcFn)


We don't have a control endpoint for this, I guess?

kjacque · 2026-06-04T23:08:59Z

+
+// propsFromC converts a daos_prop_t to Go PoolProperty values for a
+// pool-create request.
+func propsFromC(cProps *C.daos_prop_t) ([]*daos.PoolProperty, error) {


Does the daos API (going in the other direction) have an existing utility function for this that could be re-used? Or no?

daosbuild3 · 2026-06-05T21:58:43Z

Test stage Functional Hardware Medium Verbs Provider MD on SSD completed with status FAILURE. https://jenkins-3.daos.hpc.amslabs.hpecorp.net//job/daos-stack/job/daos/view/change-requests/job/PR-17550/44/execution/node/1434/log

daosbuild3 · 2026-06-05T22:38:41Z

Test stage Functional Hardware Medium MD on SSD completed with status FAILURE. https://jenkins-3.daos.hpc.amslabs.hpecorp.net//job/daos-stack/job/daos/view/change-requests/job/PR-17550/44/execution/node/1424/log

daosbuild3 · 2026-06-26T20:58:45Z

Test stage Unit Test with memcheck completed with status UNSTABLE. https://jenkins-3.daos.hpc.amslabs.hpecorp.net/job/daos-stack/job/daos//view/change-requests/job/PR-17550/45/testReport/

daosbuild3 · 2026-06-26T22:30:40Z

Test stage Unit Test bdev with memcheck completed with status UNSTABLE. https://jenkins-3.daos.hpc.amslabs.hpecorp.net/job/daos-stack/job/daos//view/change-requests/job/PR-17550/45/testReport/

daosbuild3 · 2026-06-27T12:33:44Z

Test stage Functional Hardware Large MD on SSD completed with status FAILURE. https://jenkins-3.daos.hpc.amslabs.hpecorp.net//job/daos-stack/job/daos/view/change-requests/job/PR-17550/45/execution/node/1341/log

daosbuild3 · 2026-06-28T00:07:22Z

Test stage Functional Hardware Medium MD on SSD completed with status FAILURE. https://jenkins-3.daos.hpc.amslabs.hpecorp.net//job/daos-stack/job/daos/view/change-requests/job/PR-17550/45/execution/node/1421/log

daosbuild3 · 2026-06-28T00:21:45Z

Test stage Functional Hardware Medium Verbs Provider MD on SSD completed with status FAILURE. https://jenkins-3.daos.hpc.amslabs.hpecorp.net//job/daos-stack/job/daos/view/change-requests/job/PR-17550/45/execution/node/1462/log

daosbuild3 · 2026-06-29T19:08:13Z

Test stage Unit Test with memcheck completed with status UNSTABLE. https://jenkins-3.daos.hpc.amslabs.hpecorp.net/job/daos-stack/job/daos//view/change-requests/job/PR-17550/46/testReport/

daosbuild3 · 2026-06-30T16:40:29Z

Test stage Functional Hardware Medium Verbs Provider MD on SSD completed with status FAILURE. https://jenkins-3.daos.hpc.amslabs.hpecorp.net//job/daos-stack/job/daos/view/change-requests/job/PR-17550/46/execution/node/1620/log

daosbuild3 · 2026-06-30T17:01:48Z

Test stage Functional Hardware Medium MD on SSD completed with status FAILURE. https://jenkins-3.daos.hpc.amslabs.hpecorp.net//job/daos-stack/job/daos/view/change-requests/job/PR-17550/46/execution/node/1610/log

daosbuild3 · 2026-07-02T23:02:40Z

Test stage Functional Hardware Large MD on SSD completed with status FAILURE. https://jenkins-3.daos.hpc.amslabs.hpecorp.net//job/daos-stack/job/daos/view/change-requests/job/PR-17550/47/execution/node/1797/log

daosbuild3 · 2026-07-03T03:07:11Z

Test stage Functional Hardware Medium Verbs Provider MD on SSD completed with status FAILURE. https://jenkins-3.daos.hpc.amslabs.hpecorp.net//job/daos-stack/job/daos/view/change-requests/job/PR-17550/47/execution/node/1777/log

daosbuild3 · 2026-07-03T03:52:57Z

Test stage Functional Hardware Medium MD on SSD completed with status FAILURE. https://jenkins-3.daos.hpc.amslabs.hpecorp.net//job/daos-stack/job/daos/view/change-requests/job/PR-17550/47/execution/node/1787/log

Use Go's c-shared-library mode to build a .so that exposes a C API for the Go Control API. The initial use case for this library is to replace the problematic dmg helpers test library that uses fork/exec to allow tests to drive dmg commands. Skip-hw-medium-ucx-provider: false Features: cat_recov Signed-off-by: Michael MacDonald <github@macdonald.cx>

* daos_pool: Don't escape self_heal prop value * control/server.go: Raw %v logging panics on nil; just remove it Signed-off-by: Michael MacDonald <github@macdonald.cx>

Update the cat_recov suite to work better with the new libdaos_control bindings, which operate more quickly and correctly than the old fork/exec dmg wrappers. - Check repair takes the CIA_ action value directly, where the old dmg CLI translated a selection index into the action. - The bindings answer queries in milliseconds where the fork/exec dmg helpers took 1-2s, which throws off some of the timing assumptions in the tests. Adjust to use a wall-clock budget instead of counting query iterations. - Pool-not-found now maps to DER_NONEXIST instead of DER_MISC, so cr_cleanup no longer needs to tolerate DER_MISC. Skip-hw-medium-ucx-provider: false Features: cat_recov daos_core_test_nvme Signed-off-by: Michael MacDonald <github@macdonald.cx>

mjmac force-pushed the mjmac/DAOS-16311 branch from b1d5b89 to aa22df5 Compare February 11, 2026 20:49

mjmac force-pushed the mjmac/DAOS-16311 branch from aa22df5 to 9008458 Compare February 12, 2026 17:56

mjmac force-pushed the mjmac/DAOS-16311 branch from 9008458 to 249715f Compare February 22, 2026 18:36

mjmac force-pushed the mjmac/DAOS-16311 branch from 249715f to 4b56bdd Compare February 27, 2026 18:50

mjmac force-pushed the mjmac/DAOS-16311 branch from 4b56bdd to 096a1da Compare February 27, 2026 20:04

mjmac force-pushed the mjmac/DAOS-16311 branch from a7a74e4 to 5d68fed Compare June 2, 2026 14:46

mjmac requested a review from kjacque June 2, 2026 15:27

mjmac force-pushed the mjmac/DAOS-16311 branch from 5d68fed to 9abc728 Compare June 4, 2026 14:46

kjacque reviewed Jun 4, 2026

View reviewed changes

mjmac force-pushed the mjmac/DAOS-16311 branch from 9abc728 to 21c4fc3 Compare June 26, 2026 20:02

mjmac force-pushed the mjmac/DAOS-16311 branch from 21c4fc3 to a7cbd33 Compare June 29, 2026 18:03

mjmac force-pushed the mjmac/DAOS-16311 branch from a7cbd33 to cc49cbb Compare June 30, 2026 19:47

mjmac added 3 commits July 4, 2026 10:17

DAOS-16311 test: misc fixes found during expanded testing

6543910

* daos_pool: Don't escape self_heal prop value * control/server.go: Raw %v logging panics on nil; just remove it Signed-off-by: Michael MacDonald <github@macdonald.cx>

mjmac force-pushed the mjmac/DAOS-16311 branch from cc49cbb to 4031a1a Compare July 4, 2026 14:18

Uh oh!

Conversation

mjmac commented Feb 11, 2026

Uh oh!

github-actions Bot commented Feb 11, 2026

Uh oh!

daosbuild3 commented Feb 11, 2026

Uh oh!

daosbuild3 commented Feb 11, 2026

Uh oh!

daosbuild3 commented Feb 11, 2026

Uh oh!

daosbuild3 commented Feb 12, 2026

Uh oh!

daosbuild3 commented Feb 12, 2026

Uh oh!

daosbuild3 commented Feb 12, 2026

Uh oh!

daosbuild3 commented Feb 12, 2026

Uh oh!

daosbuild3 commented Feb 12, 2026

Uh oh!

daosbuild3 commented Feb 12, 2026

Uh oh!

daosbuild3 commented Feb 22, 2026

Uh oh!

daosbuild3 commented Feb 22, 2026

Uh oh!

daosbuild3 commented Feb 22, 2026

Uh oh!

daosbuild3 commented Feb 22, 2026

Uh oh!

daosbuild3 commented Feb 22, 2026

Uh oh!

daosbuild3 commented Feb 22, 2026

Uh oh!

daosbuild3 commented Feb 27, 2026

Uh oh!

daosbuild3 commented Feb 27, 2026

Uh oh!

daosbuild3 commented Feb 27, 2026

Uh oh!

daosbuild3 commented Feb 27, 2026

Uh oh!

daosbuild3 commented Feb 27, 2026

Uh oh!

daosbuild3 commented Feb 27, 2026

Uh oh!

daosbuild3 commented Feb 27, 2026

Uh oh!

daosbuild3 commented Feb 27, 2026

Uh oh!

daosbuild3 commented Feb 27, 2026

Uh oh!

daosbuild3 commented May 31, 2026

Uh oh!

daosbuild3 commented Jun 1, 2026

Uh oh!

daosbuild3 commented Jun 1, 2026

Uh oh!

daosbuild3 commented Jun 2, 2026

Uh oh!

daosbuild3 commented Jun 3, 2026

Uh oh!

daosbuild3 commented Jun 4, 2026

Uh oh!

kjacque left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment