lib: Fix result propagation after exec() + tst_reinit()

The result pointer wasn't initialized in tst_reinit(), which haven't allowed for test result propagation to the main test process and there is actually no reason not to do that. So this commit initializes the results pointer in tst_reinit() and also adds a needs_ipc_path flag to tst_test structure, which causes the library not to unlink the IPC file after it has been mapped and also inserts LTP_IPC_PATH to the environment.

This commit also adds a test that also shows the usage of the API.
...

ptr = SAFE_MMAP(NULL, size, PROT_READ | PROT_WRITE, MAP_SHARED,
fd, 0);
+ results = ptr;

The result pointer wasn't initialized in tst_reinit(), which haven't allowed for test result propagation to the main test process and there is actually no reason not to do that. So this commit initializes the results pointer in tst_reinit()

and also adds a needs_ipc_path flag to
tst_test structure, which causes the library not to unlink the IPC file after it has been mapped and also inserts LTP_IPC_PATH to the environment.

This commit also adds a test that also shows the usage of the API.

Signed-off-by Cyril Hrubis <[email protected]>

CC Eddie Horng <[email protected]>

---
 include/tst_test.h                 |  1 +
 lib/newlib_tests/.gitignore        |  2 ++
 lib/newlib_tests/test_exec.c       | 43
 ++++++++++++++++++++++++++++++++++++++
 lib/newlib_tests/test_exec_child.c | 27 ++++++++++++++++++++++++
 lib/tst_test.c                     |  3 ++-
 5 files changed, 75 insertions(+), 1 deletion(-)
 create mode 100644 lib/newlib_tests/test_exec.c
 create mode 100644 lib/newlib_tests/test_exec_child.c

include/tst_test.h +1 -0

index 90938bc29..420597afa 100644
@@ -127,6 +127,7 @@ struct tst_test {
 	int format_device:1;
 	int mount_device:1;
 	int needs_rofs:1;
+	int needs_ipc_path:1;
 	/*
 	 * If set the test function will be executed for all available
 	 * filesystems and the current filesytem type would be set in the

lib/newlib_tests/.gitignore +2 -0

index 8c0981ec8..99edc6404 100644
@@ -20,3 +20,5 @@ tst_res_hexd
 tst_strstatus
 test17
 tst_expiration_timer
+test_exec
+test_exec_child

lib/newlib_tests/test_exec.c new file +20 -0

new file mode 100644
index 000000000..e5b90ffb8
@@ -0,0 +1,43 @@
+/*
+ * Copyright (c) 2018 Cyril Hrubis <[email protected]>
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License as
+ * published by the Free Software Foundation; either version 2 of
+ * the License, or (at your option) any later version.
+ *
+ * This program is distributed in the hope that it would be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write the Free Software Foundation,
+ * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ */
+
+/*
+ * Assert that tst_res() from child started by exec() is propagated to the

main
+ * test process.
+ *
+ * This test should be executed as:
+ * $ PATH=$PATH:$PWD ./test_exec
+ */

lib/newlib_tests/test_exec.c new file +18 -0

+
+#define _GNU_SOURCE
+#include <stdlib.h>
+#include "tst_test.h"
+
+static void do_test(void)
+{
+	char *const argv[] = {"test_exec_child", NULL};
+
+	execvpe(argv[0], argv, environ);
+
+	tst_res(TBROK | TERRNO, "EXEC!");
+}
+
+static struct tst_test test = {
+	.test_all = do_test,
+	.needs_ipc_path = 1,
+};

lib/newlib_tests/test_exec_child.c new file +27 -0

new file mode 100644
index 000000000..696ff5be2
@@ -0,0 +1,27 @@
+/*
+ * Copyright (c) 2018 Cyril Hrubis <[email protected]>
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License as
+ * published by the Free Software Foundation; either version 2 of
+ * the License, or (at your option) any later version.
+ *
+ * This program is distributed in the hope that it would be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write the Free Software Foundation,
+ * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ */
+
+#define TST_NO_DEFAULT_MAIN
+#include "tst_test.h"
+
+int main(void)
+{
+	tst_reinit();
+	tst_res(TPASS, "Child passed!");
+	return 0;
+}

lib/tst_test.c +1 -1

index 008bcefe0..8b577fde8 100644
@@ -105,7 +105,7 @@ static void setup_ipc(void)
 	results = SAFE_MMAP(NULL, size, PROT_READ | PROT_WRITE, MAP_SHARED, ipc_fd,
 	0);
 
 	/* Checkpoints needs to be accessible from processes started by exec() */
-	if (tst_test->needs_checkpoints) {
+	if (tst_test->needs_checkpoints || tst_test->needs_ipc_path) {
 		sprintf(ipc_path, IPC_ENV_VAR "=%s", shm_path);
 		putenv(ipc_path);

lib/tst_test.c +1 -1

 	} else {

@@ -152,6 +152,7 @@ void tst_reinit(void)
 	fd = SAFE_OPEN(path, O_RDWR);
 
 	ptr = SAFE_MMAP(NULL, size, PROT_READ | PROT_WRITE, MAP_SHARED, fd, 0);
+	results = ptr;
 	tst_futexes = (char*)ptr + sizeof(struct results);
 	tst_max_futexes = (size - sizeof(struct results))/sizeof(futex_t);
 
--

2.13.6

--
Mailing list info: https://lists.linux.it/listinfo/ltp

Quoted 4 lines

and also adds a needs_ipc_path flag to
tst_test structure, which causes the library not to unlink the IPC file after it has been mapped and also inserts LTP_IPC_PATH to the environment.

What is the negative of skipping unlink() for all tests? Test itself runs in separate process, so cleanup_ipc() should run anyway and clean it.

.needs_ipc_path - I'm worried users won't know what this is for. Maybe if it was called ".child_needs_reinit"? reinit appears to be the only reason we need to keep shm_path around.

Hi!

Quoted 8 lines

Quoted 4 lines

and also adds a needs_ipc_path flag to
tst_test structure, which causes the library not to unlink the IPC file after it has been mapped and also inserts LTP_IPC_PATH to the environment.

What is the negative of skipping unlink() for all tests? Test itself runs in separate process, so cleanup_ipc() should run anyway and clean it.

Well we use test ID as well as pid for the filename so it should be unique enough, we also do O_EXCL on the file so I guess that it should be safe enough not to unlink() it after it has been mapped.

There is one downside though, in a case that there is no tmpfs mounted we fall back to mapping a file, in which case we will trigger writeback to the storage by updating the piece of memory.

And given that there are actually very few testcases that needs to access the test library from the children started by exec I would like to keep unlinking the file unless needed.

Quoted 4 lines

.needs_ipc_path - I'm worried users won't know what this is for. Maybe if it was called ".child_needs_reinit"? reinit appears to be the only reason we need to keep shm_path around.

That sounds definitely better.

--
Cyril Hrubis
[email protected]