[PATCH 1/6] test: Skip tst.depth.sh

[PATCH 1/6] test: Skip tst.depth.sh

[eugene.loh]

From: Eugene Loh <eugene.loh@oracle.com>

The test asks for very many pid probes.  E.g.,

        pid$target:::entry,
        pid$target:::return,
        pid$target:a.out::,

If it is successful, this can wreak havoc on some systems, in some cases
leading to 1-2 dozen subsequent tests failing due to cascading problems:
poor test clean up, triggers left running, persistent loads, pid=0 not
running, tests timing out, and so on.  These problems require
investigation.

For now, skip the precipitating test.

See
https://oss.oracle.com/pipermail/dtrace-devel/2025-November/006916.html

Incidentally, there are similar issues for

* test/unittest/pid/err.D_PROC_CREATEFAIL.many.d
  but this test is already marked SKIP, apparently for the same reason.
  See commit d9423bd79
  ("err.D_PROC_CREATEFAIL.many.d causes failures in subsequent tests").
  The current patch simply seeks to apply the same SKIP to another test.

* test/demo/user/userfunc.d
  but this test is marked XFAIL since it needs a trigger

All three tests should be resolved when the underlying problem is
tackled.

Orabug: 38642080
Signed-off-by: Eugene Loh <eugene.loh@oracle.com>
---
 test/unittest/ustack/tst.depth.sh | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/test/unittest/ustack/tst.depth.sh b/test/unittest/ustack/tst.depth.sh
index dec16a3aa..4a67291de 100755
--- a/test/unittest/ustack/tst.depth.sh
+++ b/test/unittest/ustack/tst.depth.sh
@@ -1,10 +1,10 @@
 #!/bin/bash
 #
 # Oracle Linux DTrace.
-# Copyright (c) 2006, 2020, Oracle and/or its affiliates. All rights reserved.
+# Copyright (c) 2006, 2025, Oracle and/or its affiliates. All rights reserved.
 # Licensed under the Universal Permissive License v 1.0 as shown at
 # http://oss.oracle.com/licenses/upl.
-# @@xfail: dtv2
+# @@skip
 
 if [ $# != 1 ]; then
 	echo expected one argument: '<'dtrace-path'>'
-- 
2.47.3

[PATCH 2/6] Simple typo in comment

[eugene.loh]

From: Eugene Loh <eugene.loh@oracle.com>

Signed-off-by: Eugene Loh <eugene.loh@oracle.com>
---
 libdtrace/dt_probe.h | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/libdtrace/dt_probe.h b/libdtrace/dt_probe.h
index 54053cd3d..73b15553a 100644
--- a/libdtrace/dt_probe.h
+++ b/libdtrace/dt_probe.h
@@ -33,7 +33,7 @@ typedef struct dt_probe_instance {
 typedef struct dt_probe {
 	dt_list_t list;			/* prev/next in enablings chain */
 	dt_list_t stmts;		/* stmts */
-	dt_list_t dependents;		/* dependenct probes to attach */
+	dt_list_t dependents;		/* dependent probes to attach */
 	const dtrace_probedesc_t *desc;	/* probe description (id, name) */
 	dt_provider_t *prov;		/* pointer to containing provider */
 	struct dt_hentry he_prv;	/* provider name htab links */
-- 
2.47.3

Re: [DTrace-devel] [PATCH 2/6] Simple typo in comment

[Nick Alcock]

On 31 Mar 2026, eugene loh said:

> From: Eugene Loh <eugene.loh@oracle.com>
>
> Signed-off-by: Eugene Loh <eugene.loh@oracle.com>

Reviewed-by: Nick Alcock <nick.alcock@oracle.com>

-- 
NULL && (void)

[PATCH 3/6] Check if the BPF PROG_LOAD is either -1 or else non-negative

[eugene.loh]

From: Eugene Loh <eugene.loh@oracle.com>

The return value is supposed to be -1 or non-negative.  The current
check lumps 0 with -1.

Signed-off-by: Eugene Loh <eugene.loh@oracle.com>
---
 libdtrace/dt_bpf.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/libdtrace/dt_bpf.c b/libdtrace/dt_bpf.c
index fb66bd93b..c328f5100 100644
--- a/libdtrace/dt_bpf.c
+++ b/libdtrace/dt_bpf.c
@@ -428,7 +428,7 @@ have_helper(uint32_t func_id)
 	/* If the program loads, we can use the helper. */
 	fd = dt_bpf_prog_attach(BPF_PROG_TYPE_KPROBE, 0, 0, 0, &dp,
 			      1, log, DT_BPF_LOG_SIZE_SMALL);
-	if (fd > 0) {
+	if (fd >= 0) {
 		close(fd);
 		return 1;
 	}
-- 
2.47.3

Re: [DTrace-devel] [PATCH 3/6] Check if the BPF PROG_LOAD is either -1 or else non-negative

[Nick Alcock]

On 31 Mar 2026, eugene loh outgrape:

> From: Eugene Loh <eugene.loh@oracle.com>
>
> The return value is supposed to be -1 or non-negative.  The current
> check lumps 0 with -1.
>
> Signed-off-by: Eugene Loh <eugene.loh@oracle.com>

Reviewed-by: Nick Alcock <nick.alcock@oracle.com>

-- 
NULL && (void)

[PATCH 4/6] test: Remove unnecessary "destructive" pragma

[eugene.loh]

From: Eugene Loh <eugene.loh@oracle.com>

Signed-off-by: Eugene Loh <eugene.loh@oracle.com>
---
 test/unittest/sched/tst.oncpu.d | 1 -
 1 file changed, 1 deletion(-)

diff --git a/test/unittest/sched/tst.oncpu.d b/test/unittest/sched/tst.oncpu.d
index 7a33bab4f..65bb313f0 100644
--- a/test/unittest/sched/tst.oncpu.d
+++ b/test/unittest/sched/tst.oncpu.d
@@ -8,7 +8,6 @@
 /* @@timeout: 15 */
 
 #pragma D option switchrate=100hz
-#pragma D option destructive
 
 sched:::on-cpu
 /pid == $pid/
-- 
2.47.3

Re: [DTrace-devel] [PATCH 4/6] test: Remove unnecessary "destructive" pragma

[Nick Alcock]

On 31 Mar 2026, eugene loh outgrape:

> From: Eugene Loh <eugene.loh@oracle.com>
>
> Signed-off-by: Eugene Loh <eugene.loh@oracle.com>

Reviewed-by: Nick Alcock <nick.alcock@oracle.com>

-- 
NULL && (void)

[PATCH 5/6] Possible uninitialized 'last' variable in usdt_copyin_data()

[eugene.loh]

From: Eugene Loh <eugene.loh@oracle.com>

Some compilers warn:

libcommon/usdt_parser.c: In function ‘usdt_copyin_data’:
libcommon/usdt_parser.c:191:15: warning:
   ‘last’ may be used uninitialized in this function [-Wmaybe-uninitialized]
    last->next = blk;
    ~~~~~~~~~~~^~~~~

Change the "if" check to make it easier for compilers to recognize
that "last" will be initialized (and non-NULL even!).

Signed-off-by: Eugene Loh <eugene.loh@oracle.com>
---
 libcommon/usdt_parser.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/libcommon/usdt_parser.c b/libcommon/usdt_parser.c
index 864198098..d8cb9b7ba 100644
--- a/libcommon/usdt_parser.c
+++ b/libcommon/usdt_parser.c
@@ -163,7 +163,7 @@ usdt_destroy_data(usdt_data_t *data)
 usdt_data_t *
 usdt_copyin_data(int in, int out, int *ok)
 {
-	usdt_data_t *first = NULL, *last;
+	usdt_data_t *first = NULL, *last = NULL;
 	size_t cnt;
 
 	*ok = 1;
@@ -185,7 +185,7 @@ usdt_copyin_data(int in, int out, int *ok)
 		if ((blk = usdt_copyin_block(in, out, ok)) == NULL)
 			goto err;
 
-		if (first == NULL)
+		if (last == NULL)
 			first = last = blk;
 		else {
 			last->next = blk;
-- 
2.47.3

Re: [DTrace-devel] [PATCH 5/6] Possible uninitialized 'last' variable in usdt_copyin_data()

[Nick Alcock]

On 31 Mar 2026, eugene loh told this:

> From: Eugene Loh <eugene.loh@oracle.com>
>
> Some compilers warn:
>
> libcommon/usdt_parser.c: In function ‘usdt_copyin_data’:
> libcommon/usdt_parser.c:191:15: warning:
>    ‘last’ may be used uninitialized in this function [-Wmaybe-uninitialized]
>     last->next = blk;
>     ~~~~~~~~~~~^~~~~
>
> Change the "if" check to make it easier for compilers to recognize
> that "last" will be initialized (and non-NULL even!).
>
> Signed-off-by: Eugene Loh <eugene.loh@oracle.com>

Ugh. Still, if it works, it works, and it's semantically the same.

Reviewed-by: Nick Alcock <nick.alcock@oracle.com>

-- 
NULL && (void)

[PATCH 6/6] test: Check declaration (without init) inside a probe

[eugene.loh]

From: Eugene Loh <eugene.loh@oracle.com>

We test that declaration (with init) of a variable inside a probe is not
allowed.

Check also that declaration without init inside a probe is also
forbidden.

The related test/unittest/probes/err.D_SYNTAX/err.D_SYNTAX.declare*.d
tests are left alone.

Signed-off-by: Eugene Loh <eugene.loh@oracle.com>
---
 .../scalars/err.D_SYNTAX.declare-init.d       | 20 +++++++++++++++++++
 test/unittest/scalars/err.D_SYNTAX.declare.d  |  4 ++--
 2 files changed, 22 insertions(+), 2 deletions(-)
 create mode 100644 test/unittest/scalars/err.D_SYNTAX.declare-init.d

diff --git a/test/unittest/scalars/err.D_SYNTAX.declare-init.d b/test/unittest/scalars/err.D_SYNTAX.declare-init.d
new file mode 100644
index 000000000..fd8dca6c0
--- /dev/null
+++ b/test/unittest/scalars/err.D_SYNTAX.declare-init.d
@@ -0,0 +1,20 @@
+/*
+ * Oracle Linux DTrace.
+ * Copyright (c) 2006, 2026, Oracle and/or its affiliates. All rights reserved.
+ * Licensed under the Universal Permissive License v 1.0 as shown at
+ * http://oss.oracle.com/licenses/upl.
+ */
+
+/*
+ * ASSERTION:
+ *
+ * Declare and init a variable Inside Begin and make sure compilation fails.
+ *
+ * SECTION:  Variables/Scalar Variables
+ *
+ */
+
+BEGIN
+{
+	int x = 123;
+}
diff --git a/test/unittest/scalars/err.D_SYNTAX.declare.d b/test/unittest/scalars/err.D_SYNTAX.declare.d
index f5b2ea914..b26acd04a 100644
--- a/test/unittest/scalars/err.D_SYNTAX.declare.d
+++ b/test/unittest/scalars/err.D_SYNTAX.declare.d
@@ -1,6 +1,6 @@
 /*
  * Oracle Linux DTrace.
- * Copyright (c) 2006, Oracle and/or its affiliates. All rights reserved.
+ * Copyright (c) 2006, 2026, Oracle and/or its affiliates. All rights reserved.
  * Licensed under the Universal Permissive License v 1.0 as shown at
  * http://oss.oracle.com/licenses/upl.
  */
@@ -16,5 +16,5 @@
 
 BEGIN
 {
-	int x = 123;
+	int x;
 }
-- 
2.47.3

Re: [DTrace-devel] [PATCH 6/6] test: Check declaration (without init) inside a probe

[Nick Alcock]

On 31 Mar 2026, eugene loh stated:

> From: Eugene Loh <eugene.loh@oracle.com>
>
> We test that declaration (with init) of a variable inside a probe is not
> allowed.
>
> Check also that declaration without init inside a probe is also
> forbidden.
>
> The related test/unittest/probes/err.D_SYNTAX/err.D_SYNTAX.declare*.d
> tests are left alone.
>
> Signed-off-by: Eugene Loh <eugene.loh@oracle.com>

Reviewed-by: Nick Alcock <nick.alcock@oracle.com>

-- 
NULL && (void)

Re: [DTrace-devel] [PATCH 1/6] test: Skip tst.depth.sh

[Nick Alcock]

On 31 Mar 2026, eugene loh told this:

> From: Eugene Loh <eugene.loh@oracle.com>
>
> The test asks for very many pid probes.  E.g.,
>
>         pid$target:::entry,
>         pid$target:::return,
>         pid$target:a.out::,
>
> If it is successful, this can wreak havoc on some systems, in some cases
> leading to 1-2 dozen subsequent tests failing due to cascading problems:
> poor test clean up, triggers left running, persistent loads, pid=0 not
> running, tests timing out, and so on.  These problems require
> investigation.
>
> For now, skip the precipitating test.

This sort of thing is what the tagging system is for. This should
probably be a default-skipped tag instead:

@@tag: skip-default

And add

!skip-default

into test/tags.default (not currently existing).

(The same is probably true of most of the other remaining skipped
tests.)

-- 
NULL && (void)