Skip to contentSkip to conversation

syscalls/ioctl09: Add test for BLKRRPART ioctl

PATCH New Yang Xu 3 participants 10 messages Jul 6, 2020 – Jul 22, 2020

lore
Share & tools
lore

Yang Xu

author
#0 lore raw

PATCH 1/1 syscalls/ioctl09: Add test for BLKRRPART ioctl

Fixes #699

---
 runtest/syscalls                           |   1 +
 testcases/kernel/syscalls/ioctl/.gitignore |   1 +
 testcases/kernel/syscalls/ioctl/ioctl09.c  | 116 +++++++++++++++++++++
 3 files changed, 118 insertions(+)
 create mode 100644 testcases/kernel/syscalls/ioctl/ioctl09.c
runtest/syscalls +1 -0
index 5b3a0862f..aaa81e4ee 100644
@@ -529,6 +529,7 @@ ioctl05      ioctl05
 ioctl06      ioctl06
 ioctl07      ioctl07
 ioctl08      ioctl08
+ioctl09      ioctl09
 
 ioctl_loop01 ioctl_loop01
 ioctl_loop02 ioctl_loop02
testcases/kernel/syscalls/ioctl/.gitignore +1 -0
index 3a3d49adc..5fff7a61d 100644
@@ -6,6 +6,7 @@
 /ioctl06
 /ioctl07
 /ioctl08
+/ioctl09
 /ioctl_loop01
 /ioctl_loop02
 /ioctl_loop03
testcases/kernel/syscalls/ioctl/ioctl09.c new file +116 -0

116 lines not on this page · raw message

Signature · 1 line
2.23.0
#1 lore raw

PATCH 1/1 syscalls/ioctl09: Add test for BLKRRPART ioctl

----- Original Message -----

Hi,

Quoted 9 lines
testcases/kernel/syscalls/ioctl/ioctl09.c new file +3 -0
+static void verify_ioctl(void)
+{
+	const char *const cmd_parted_old[] = {"parted", "-s", "test.img",

"mklabel", "msdos", "mkpart",

(quoted diff) +2 -0
+						"primary", "ext4", "1M", "10M", NULL};
+	const char *const cmd_parted_new[] = {"parted", "-s", "test.img",

"mklabel", "msdos", "mkpart",

(quoted diff) +1 -0
+						"primary", "ext4", "1M", "10M", "mkpart", "primary", "ext4", "10M",

"20M", NULL};

^^ These (and other lines) goe way over 80 characters.

testcases/kernel/syscalls/ioctl/ioctl09.c line 71

Quoted 7 lines
testcases/kernel/syscalls/ioctl/ioctl09.c new file +7 -0
+	struct loop_info loopinfo = {0};
+
+	dev_num = tst_find_free_loopdev(dev_path, sizeof(dev_path));
+	if (dev_num < 0)
+		tst_brk(TBROK, "Failed to find free loop device");
+
+	tst_fill_file("test.img", 0, 1024 * 1024, 20);

20MB feels a bit small, even though tests I ran passed on old and new kernels. LTP default is 256, to cover various filesystems, but here we don't even write anything..

Other than that, it looks good to me:

author
#2 lore raw

PATCH 1/1 syscalls/ioctl09: Add test for BLKRRPART ioctl

Hi Jan

Thanks for your review.

Quoted 6 lines

----- Original Message -----

Hi,

Quoted 12 lines
Quoted 9 lines
testcases/kernel/syscalls/ioctl/ioctl09.c new file +3 -0
+static void verify_ioctl(void)
+{
+	const char *const cmd_parted_old[] = {"parted", "-s", "test.img",

"mklabel", "msdos", "mkpart",

(quoted diff) +2 -0
+						"primary", "ext4", "1M", "10M", NULL};
+	const char *const cmd_parted_new[] = {"parted", "-s", "test.img",

"mklabel", "msdos", "mkpart",

(quoted diff) +1 -0
+						"primary", "ext4", "1M", "10M", "mkpart", "primary", "ext4", "10M",

"20M", NULL};

^^ These (and other lines) goe way over 80 characters.

OK.

Quoted 11 lines

testcases/kernel/syscalls/ioctl/ioctl09.c line 71

Quoted 7 lines
testcases/kernel/syscalls/ioctl/ioctl09.c new file +7 -0
+	struct loop_info loopinfo = {0};
+
+	dev_num = tst_find_free_loopdev(dev_path, sizeof(dev_path));
+	if (dev_num < 0)
+		tst_brk(TBROK, "Failed to find free loop device");
+
+	tst_fill_file("test.img", 0, 1024 * 1024, 20);

20MB feels a bit small, even though tests I ran passed on old and new kernels. LTP default is 256, to cover various filesystems, but here we don't even write anything..

In actually, ioctl_loop01 test case also uses 10M size and we only cover ext4 fileystem in here. I don't have objection about changing this size to 256M. Just reconfirm. If so, we should also modify ioctl_loop01 code.

Quoted 6 lines

Other than that, it looks good to me:

#3 lore raw

PATCH 1/1 syscalls/ioctl09: Add test for BLKRRPART ioctl

----- Original Message -----

Quoted 7 lines
Quoted 4 lines

20MB feels a bit small, even though tests I ran passed on old and new kernels.
LTP default is 256, to cover various filesystems, but here we don't even write anything..

In actually, ioctl_loop01 test case also uses 10M size and we only cover ext4 fileystem in here. I don't have objection about changing this size to 256M. Just reconfirm. If so, we should also modify ioctl_loop01 code.

OK, thanks for pointing that out. I wanted to go to 32 or 64, but if 10 worked in ioctl_loop01 until now, then 20 should be fine here.

author
#4 lore raw

PATCH v2 1/1 syscalls/ioctl09: Add test for BLKRRPART ioctl syscalls/ioctl09: Add test for BLKRRPART ioctl

Fixes #699

---
v1->v2:
code style fix(below 80 characters)

 runtest/syscalls                           |   1 +
 testcases/kernel/syscalls/ioctl/.gitignore |   1 +
 testcases/kernel/syscalls/ioctl/ioctl09.c  | 126 +++++++++++++++++++++
 3 files changed, 128 insertions(+)
 create mode 100644 testcases/kernel/syscalls/ioctl/ioctl09.c
runtest/syscalls +1 -0
index 5b3a0862f..aaa81e4ee 100644
@@ -529,6 +529,7 @@ ioctl05      ioctl05
 ioctl06      ioctl06
 ioctl07      ioctl07
 ioctl08      ioctl08
+ioctl09      ioctl09
 
 ioctl_loop01 ioctl_loop01
 ioctl_loop02 ioctl_loop02
testcases/kernel/syscalls/ioctl/.gitignore +1 -0
index 3a3d49adc..5fff7a61d 100644
@@ -6,6 +6,7 @@
 /ioctl06
 /ioctl07
 /ioctl08
+/ioctl09
 /ioctl_loop01
 /ioctl_loop02
 /ioctl_loop03
testcases/kernel/syscalls/ioctl/ioctl09.c new file +126 -0

126 lines not on this page · raw message

Signature · 1 line
2.23.0
author
#5 lore raw

PATCH v2 1/1 syscalls/ioctl09: Add test for BLKRRPART ioctl syscalls/ioctl09: Add test for BLKRRPART ioctl

Hi Cyril

Do you have some comment for this patch?

Best Regards
Yang Xu

Quoted 170 lines

Fixes #699

---
v1->v2:
code style fix(below 80 characters)

  runtest/syscalls                           |   1 +
  testcases/kernel/syscalls/ioctl/.gitignore |   1 +
  testcases/kernel/syscalls/ioctl/ioctl09.c  | 126 +++++++++++++++++++++
  3 files changed, 128 insertions(+)

create mode 100644 testcases/kernel/syscalls/ioctl/ioctl09.c

runtest/syscalls +1 -0
index 5b3a0862f..aaa81e4ee 100644
@@ -529,6 +529,7 @@ ioctl05      ioctl05
  ioctl06      ioctl06
  ioctl07      ioctl07
  ioctl08      ioctl08
+ioctl09      ioctl09
 
  ioctl_loop01 ioctl_loop01
  ioctl_loop02 ioctl_loop02
testcases/kernel/syscalls/ioctl/.gitignore +1 -0
index 3a3d49adc..5fff7a61d 100644
@@ -6,6 +6,7 @@
  /ioctl06
  /ioctl07
  /ioctl08
+/ioctl09
  /ioctl_loop01
  /ioctl_loop02
  /ioctl_loop03
testcases/kernel/syscalls/ioctl/ioctl09.c new file +126 -0
new file mode 100644
index 000000000..b39ef9874
@@ -0,0 +1,126 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/*
+ * Copyright (c) 2020 FUJITSU LIMITED. All rights reserved.
+ * Author: Yang Xu <[email protected]>
+ *
+ * Basic test for the BLKRRPART ioctl, it is the same as blockdev
+ * --rereadpt command.
+ */
+
+#include <stdio.h>
+#include <unistd.h>
+#include <string.h>
+#include <sys/mount.h>
+#include <stdbool.h>
+#include "lapi/loop.h"
+#include "tst_test.h"
+
+static char dev_path[1024];
+static int dev_num, attach_flag, dev_fd;
+static char loop_partpath[1026], sys_loop_partpath[1026];
+
+static void change_partition(const char *const cmd[])
+{
+	int ret;
+
+	ret = tst_cmd(cmd, NULL, NULL, TST_CMD_PASS_RETVAL);
+	switch (ret) {
+	case 0:
+	break;
+	case 255:
+		tst_res(TCONF, "parted binary not installed or failed");
+	break;
+	default:
+		tst_res(TCONF, "parted exited with %i", ret);
+	break;
+	}
+}
+
+static void check_partition(int part_num, bool value)
+{
+	int ret;
+
+	sprintf(sys_loop_partpath, "/sys/block/loop%d/loop%dp%d",
+		dev_num, dev_num, part_num);
+	sprintf(loop_partpath, "%sp%d", dev_path, part_num);
+
+	ret = access(sys_loop_partpath, F_OK);
+	if (ret == 0)
+		tst_res(value ? TPASS : TFAIL, "access %s succeeds",
+			sys_loop_partpath);
+	else
+		tst_res(value ? TFAIL : TPASS, "access %s fails",
+			sys_loop_partpath);
+
+	ret = access(loop_partpath, F_OK);
+	if (ret == 0)
+		tst_res(value ? TPASS : TFAIL, "access %s succeeds",
+			loop_partpath);
+	else
+		tst_res(value ? TFAIL : TPASS, "access %s fails",
+			loop_partpath);
+}
+
+static void verify_ioctl(void)
+{
+	const char *const cmd_parted_old[] = {"parted", "-s", "test.img",
+					      "mklabel", "msdos", "mkpart",
+					      "primary", "ext4", "1M", "10M",
+					      NULL};
+	const char *const cmd_parted_new[] = {"parted", "-s", "test.img",
+					      "mklabel", "msdos", "mkpart",
+					      "primary", "ext4", "1M", "10M",
+					      "mkpart", "primary", "ext4",
+					      "10M", "20M", NULL};
+	struct loop_info loopinfo = {0};
+
+	dev_num = tst_find_free_loopdev(dev_path, sizeof(dev_path));
+	if (dev_num < 0)
+		tst_brk(TBROK, "Failed to find free loop device");
+
+	tst_fill_file("test.img", 0, 1024 * 1024, 20);
+
+	change_partition(cmd_parted_old);
+
+	tst_attach_device(dev_path, "test.img");
+	attach_flag = 1;
+
+	dev_fd = SAFE_OPEN(dev_path, O_RDWR);
+	loopinfo.lo_flags =  LO_FLAGS_PARTSCAN;
+	SAFE_IOCTL(dev_fd, LOOP_SET_STATUS, &loopinfo);
+	check_partition(1, true);
+	check_partition(2, false);
+
+	change_partition(cmd_parted_new);
+	TST_RETRY_FUNC(ioctl(dev_fd, BLKRRPART, 0), TST_RETVAL_EQ0);
+	check_partition(1, true);
+	check_partition(2, true);
+
+	SAFE_CLOSE(dev_fd);
+	tst_detach_device(dev_path);
+	attach_flag = 0;
+	unlink("test.img");
+}
+
+static void cleanup(void)
+{
+	if (dev_fd > 0)
+		SAFE_CLOSE(dev_fd);
+	if (attach_flag)
+		tst_detach_device(dev_path);
+}
+
+static struct tst_test test = {
+	.cleanup = cleanup,
+	.test_all = verify_ioctl,
+	.needs_root = 1,
+	.needs_drivers = (const char *const []) {
+		"loop",
+		NULL
+	},
+	.needs_cmds = (const char *const []) {
+		"parted",
+		NULL
+	},
+	.needs_tmpdir = 1,
+};
#6 lore raw

PATCH v2 1/1 syscalls/ioctl09: Add test for BLKRRPART ioctl syscalls/ioctl09: Add test for BLKRRPART ioctl

In reply to Yang Xu

Hi!

Quoted 1 line

Fixes #699

Looks good a couple of minor comments below.

Quoted 63 lines

---
v1->v2:
code style fix(below 80 characters)

 runtest/syscalls                           |   1 +
 testcases/kernel/syscalls/ioctl/.gitignore |   1 +
 testcases/kernel/syscalls/ioctl/ioctl09.c  | 126 +++++++++++++++++++++
 3 files changed, 128 insertions(+)
 create mode 100644 testcases/kernel/syscalls/ioctl/ioctl09.c
runtest/syscalls +1 -0
index 5b3a0862f..aaa81e4ee 100644
@@ -529,6 +529,7 @@ ioctl05      ioctl05
 ioctl06      ioctl06
 ioctl07      ioctl07
 ioctl08      ioctl08
+ioctl09      ioctl09
 
 ioctl_loop01 ioctl_loop01
 ioctl_loop02 ioctl_loop02
testcases/kernel/syscalls/ioctl/.gitignore +1 -0
index 3a3d49adc..5fff7a61d 100644
@@ -6,6 +6,7 @@
 /ioctl06
 /ioctl07
 /ioctl08
+/ioctl09
 /ioctl_loop01
 /ioctl_loop02
 /ioctl_loop03
testcases/kernel/syscalls/ioctl/ioctl09.c new file +22 -0
new file mode 100644
index 000000000..b39ef9874
@@ -0,0 +1,126 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/*
+ * Copyright (c) 2020 FUJITSU LIMITED. All rights reserved.
+ * Author: Yang Xu <[email protected]>
+ *
+ * Basic test for the BLKRRPART ioctl, it is the same as blockdev
+ * --rereadpt command.
+ */
+
+#include <stdio.h>
+#include <unistd.h>
+#include <string.h>
+#include <sys/mount.h>
+#include <stdbool.h>
+#include "lapi/loop.h"
+#include "tst_test.h"
+
+static char dev_path[1024];
+static int dev_num, attach_flag, dev_fd;
+static char loop_partpath[1026], sys_loop_partpath[1026];
+
+static void change_partition(const char *const cmd[])

testcases/kernel/syscalls/ioctl/ioctl09.c line 32

Quoted 10 lines
testcases/kernel/syscalls/ioctl/ioctl09.c new file +10 -0
+{
+	int ret;
+
+	ret = tst_cmd(cmd, NULL, NULL, TST_CMD_PASS_RETVAL);
+	switch (ret) {
+	case 0:
+	break;
+	case 255:
+		tst_res(TCONF, "parted binary not installed or failed");
+	break;

We do have .needs_cmds in the test structure so we don't have to handle 255 here, the test will not start if parted is not installed.

testcases/kernel/syscalls/ioctl/ioctl09.c line 34

Quoted 2 lines
testcases/kernel/syscalls/ioctl/ioctl09.c new file +2 -0
+	default:
+		tst_res(TCONF, "parted exited with %i", ret);

Shouldn't this be TBROK?

Or at least tst_brk() because we will proceed with the test as it is and possibly fail the test since parted haven't modified the binary. Or does parted return non-zero when it succeeds?

Generally if we are going to handle only one failure case we can write it as:

	ret = tst_cmd(...);
	if (ret)
		tst_brk(TBROK, "parted returned %i", ret);
Quoted 35 lines
testcases/kernel/syscalls/ioctl/ioctl09.c new file +35 -0
+	break;
+	}
+}
+
+static void check_partition(int part_num, bool value)
+{
+	int ret;
+
+	sprintf(sys_loop_partpath, "/sys/block/loop%d/loop%dp%d",
+		dev_num, dev_num, part_num);
+	sprintf(loop_partpath, "%sp%d", dev_path, part_num);
+
+	ret = access(sys_loop_partpath, F_OK);
+	if (ret == 0)
+		tst_res(value ? TPASS : TFAIL, "access %s succeeds",
+			sys_loop_partpath);
+	else
+		tst_res(value ? TFAIL : TPASS, "access %s fails",
+			sys_loop_partpath);
+
+	ret = access(loop_partpath, F_OK);
+	if (ret == 0)
+		tst_res(value ? TPASS : TFAIL, "access %s succeeds",
+			loop_partpath);
+	else
+		tst_res(value ? TFAIL : TPASS, "access %s fails",
+			loop_partpath);
+}
+
+static void verify_ioctl(void)
+{
+	const char *const cmd_parted_old[] = {"parted", "-s", "test.img",
+					      "mklabel", "msdos", "mkpart",
+					      "primary", "ext4", "1M", "10M",
+					      NULL};

testcases/kernel/syscalls/ioctl/ioctl09.c line 79

Quoted 10 lines
testcases/kernel/syscalls/ioctl/ioctl09.c new file +10 -0
+	const char *const cmd_parted_new[] = {"parted", "-s", "test.img",
+					      "mklabel", "msdos", "mkpart",
+					      "primary", "ext4", "1M", "10M",
+					      "mkpart", "primary", "ext4",
+					      "10M", "20M", NULL};
+	struct loop_info loopinfo = {0};
+
+	dev_num = tst_find_free_loopdev(dev_path, sizeof(dev_path));
+	if (dev_num < 0)
+		tst_brk(TBROK, "Failed to find free loop device");

Shouldn't we move the tst_find_free_loopdev() to the test setup?

testcases/kernel/syscalls/ioctl/ioctl09.c line 81

Quoted 1 line
testcases/kernel/syscalls/ioctl/ioctl09.c new file +1 -0
+	tst_fill_file("test.img", 0, 1024 * 1024, 20);

I wonder if the recently introduced tst_prealloc_file() would make the test a bit faster. Have you tried that?

Quoted 52 lines
testcases/kernel/syscalls/ioctl/ioctl09.c new file +44 -0
+	change_partition(cmd_parted_old);
+
+	tst_attach_device(dev_path, "test.img");
+	attach_flag = 1;
+
+	dev_fd = SAFE_OPEN(dev_path, O_RDWR);
+	loopinfo.lo_flags =  LO_FLAGS_PARTSCAN;
+	SAFE_IOCTL(dev_fd, LOOP_SET_STATUS, &loopinfo);
+	check_partition(1, true);
+	check_partition(2, false);
+
+	change_partition(cmd_parted_new);
+	TST_RETRY_FUNC(ioctl(dev_fd, BLKRRPART, 0), TST_RETVAL_EQ0);
+	check_partition(1, true);
+	check_partition(2, true);
+
+	SAFE_CLOSE(dev_fd);
+	tst_detach_device(dev_path);
+	attach_flag = 0;
+	unlink("test.img");
+}
+
+static void cleanup(void)
+{
+	if (dev_fd > 0)
+		SAFE_CLOSE(dev_fd);
+	if (attach_flag)
+		tst_detach_device(dev_path);
+}
+
+static struct tst_test test = {
+	.cleanup = cleanup,
+	.test_all = verify_ioctl,
+	.needs_root = 1,
+	.needs_drivers = (const char *const []) {
+		"loop",
+		NULL
+	},
+	.needs_cmds = (const char *const []) {
+		"parted",
+		NULL
+	},
+	.needs_tmpdir = 1,
+};
Signature · 7 lines
2.23.0




-- 
Mailing list info: https://lists.linux.it/listinfo/ltp
Signature · 2 lines
Cyril Hrubis
[email protected]
author
#7 lore raw

PATCH v2 1/1 syscalls/ioctl09: Add test for BLKRRPART ioctl syscalls/ioctl09: Add test for BLKRRPART ioctl

Hi Cyril

Quoted 4 lines

Hi!

Quoted 1 line

Fixes #699

Looks good a couple of minor comments below.

Thanks for your review.

Quoted 87 lines
Quoted 64 lines

---
v1->v2:
code style fix(below 80 characters)

  runtest/syscalls                           |   1 +
  testcases/kernel/syscalls/ioctl/.gitignore |   1 +
  testcases/kernel/syscalls/ioctl/ioctl09.c  | 126 +++++++++++++++++++++
  3 files changed, 128 insertions(+)

create mode 100644 testcases/kernel/syscalls/ioctl/ioctl09.c

runtest/syscalls +1 -0
index 5b3a0862f..aaa81e4ee 100644
@@ -529,6 +529,7 @@ ioctl05      ioctl05
  ioctl06      ioctl06
  ioctl07      ioctl07
  ioctl08      ioctl08
+ioctl09      ioctl09
 
  ioctl_loop01 ioctl_loop01
  ioctl_loop02 ioctl_loop02
testcases/kernel/syscalls/ioctl/.gitignore +1 -0
index 3a3d49adc..5fff7a61d 100644
@@ -6,6 +6,7 @@
  /ioctl06
  /ioctl07
  /ioctl08
+/ioctl09
  /ioctl_loop01
  /ioctl_loop02
  /ioctl_loop03
testcases/kernel/syscalls/ioctl/ioctl09.c new file +22 -0
new file mode 100644
index 000000000..b39ef9874
@@ -0,0 +1,126 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/*
+ * Copyright (c) 2020 FUJITSU LIMITED. All rights reserved.
+ * Author: Yang Xu <[email protected]>
+ *
+ * Basic test for the BLKRRPART ioctl, it is the same as blockdev
+ * --rereadpt command.
+ */
+
+#include <stdio.h>
+#include <unistd.h>
+#include <string.h>
+#include <sys/mount.h>
+#include <stdbool.h>
+#include "lapi/loop.h"
+#include "tst_test.h"
+
+static char dev_path[1024];
+static int dev_num, attach_flag, dev_fd;
+static char loop_partpath[1026], sys_loop_partpath[1026];
+
+static void change_partition(const char *const cmd[])

testcases/kernel/syscalls/ioctl/ioctl09.c line 32

Quoted 10 lines
testcases/kernel/syscalls/ioctl/ioctl09.c new file +10 -0
+{
+	int ret;
+
+	ret = tst_cmd(cmd, NULL, NULL, TST_CMD_PASS_RETVAL);
+	switch (ret) {
+	case 0:
+	break;
+	case 255:
+		tst_res(TCONF, "parted binary not installed or failed");
+	break;

We do have .needs_cmds in the test structure so we don't have to handle 255 here, the test will not start if parted is not installed.

testcases/kernel/syscalls/ioctl/ioctl09.c line 34

Quoted 2 lines
testcases/kernel/syscalls/ioctl/ioctl09.c new file +2 -0
+	default:
+		tst_res(TCONF, "parted exited with %i", ret);

Shouldn't this be TBROK?

Or at least tst_brk() because we will proceed with the test as it is and possibly fail the test since parted haven't modified the binary. Or does parted return non-zero when it succeeds?

Quoted 5 lines

Generally if we are going to handle only one failure case we can write it as:

	ret = tst_cmd(...);
	if (ret)
		tst_brk(TBROK, "parted returned %i", ret);

Yes, you are right. We don't need to handle 255 and should use tst_brk.

Quoted 40 lines
Quoted 40 lines
testcases/kernel/syscalls/ioctl/ioctl09.c new file +39 -0
+	break;
+	}
+}
+
+static void check_partition(int part_num, bool value)
+{
+	int ret;
+
+	sprintf(sys_loop_partpath, "/sys/block/loop%d/loop%dp%d",
+		dev_num, dev_num, part_num);
+	sprintf(loop_partpath, "%sp%d", dev_path, part_num);
+
+	ret = access(sys_loop_partpath, F_OK);
+	if (ret == 0)
+		tst_res(value ? TPASS : TFAIL, "access %s succeeds",
+			sys_loop_partpath);
+	else
+		tst_res(value ? TFAIL : TPASS, "access %s fails",
+			sys_loop_partpath);
+
+	ret = access(loop_partpath, F_OK);
+	if (ret == 0)
+		tst_res(value ? TPASS : TFAIL, "access %s succeeds",
+			loop_partpath);
+	else
+		tst_res(value ? TFAIL : TPASS, "access %s fails",
+			loop_partpath);
+}
+
+static void verify_ioctl(void)
+{
+	const char *const cmd_parted_old[] = {"parted", "-s", "test.img",
+					      "mklabel", "msdos", "mkpart",
+					      "primary", "ext4", "1M", "10M",
+					      NULL};
+	const char *const cmd_parted_new[] = {"parted", "-s", "test.img",
+					      "mklabel", "msdos", "mkpart",
+					      "primary", "ext4", "1M", "10M",
+					      "mkpart", "primary", "ext4",
Quoted 8 lines

testcases/kernel/syscalls/ioctl/ioctl09.c line 79

Quoted 6 lines
testcases/kernel/syscalls/ioctl/ioctl09.c new file +6 -0
+					      "10M", "20M", NULL};
+	struct loop_info loopinfo = {0};
+
+	dev_num = tst_find_free_loopdev(dev_path, sizeof(dev_path));
+	if (dev_num < 0)
+		tst_brk(TBROK, "Failed to find free loop device");

Shouldn't we move the tst_find_free_loopdev() to the test setup?

Yes.

Quoted 5 lines

testcases/kernel/syscalls/ioctl/ioctl09.c line 81

Quoted 2 lines
testcases/kernel/syscalls/ioctl/ioctl09.c new file +1 -0
+	tst_fill_file("test.img", 0, 1024 * 1024, 20);

I wonder if the recently introduced tst_prealloc_file() would make the test a bit faster. Have you tried that?

Now, I tried it. time ./ioctl09 -i 100
by using tst_fill_file, the avergage(I tested 10 times)

real    0m13.467s
user    0m0.550s
sys     0m1.71s
by using tst_prealloc_file, the avergage
real    0m13.44s
user    0m0.60s
sys     0m1.56s

It seems that using tst_prealloc_file faster a bit. I will use this new api. Also I will move tst_prealloc_file into setup because it doesn't need to create test.img each time.

Best Regards
Yang Xu

Quoted 54 lines
Quoted 53 lines
testcases/kernel/syscalls/ioctl/ioctl09.c new file +44 -0
+	change_partition(cmd_parted_old);
+
+	tst_attach_device(dev_path, "test.img");
+	attach_flag = 1;
+
+	dev_fd = SAFE_OPEN(dev_path, O_RDWR);
+	loopinfo.lo_flags =  LO_FLAGS_PARTSCAN;
+	SAFE_IOCTL(dev_fd, LOOP_SET_STATUS, &loopinfo);
+	check_partition(1, true);
+	check_partition(2, false);
+
+	change_partition(cmd_parted_new);
+	TST_RETRY_FUNC(ioctl(dev_fd, BLKRRPART, 0), TST_RETVAL_EQ0);
+	check_partition(1, true);
+	check_partition(2, true);
+
+	SAFE_CLOSE(dev_fd);
+	tst_detach_device(dev_path);
+	attach_flag = 0;
+	unlink("test.img");
+}
+
+static void cleanup(void)
+{
+	if (dev_fd > 0)
+		SAFE_CLOSE(dev_fd);
+	if (attach_flag)
+		tst_detach_device(dev_path);
+}
+
+static struct tst_test test = {
+	.cleanup = cleanup,
+	.test_all = verify_ioctl,
+	.needs_root = 1,
+	.needs_drivers = (const char *const []) {
+		"loop",
+		NULL
+	},
+	.needs_cmds = (const char *const []) {
+		"parted",
+		NULL
+	},
+	.needs_tmpdir = 1,
+};
Signature · 7 lines
2.23.0




-- 
Mailing list info: https://lists.linux.it/listinfo/ltp
author
#8 lore raw

PATCH v3 1/1 syscalls/ioctl09: Add test for BLKRRPART ioctl

In reply to Cyril Hrubis

Fixes #699

---
v2->v3
1. remove useless cmd judgement about 255 and use tst_brk
2. move tst_find_free_loopdev into setup
3. use new api tst_prealloc_file instead of tst_fill_file and move it into setup

 runtest/syscalls                           |   1 +
 testcases/kernel/syscalls/ioctl/.gitignore |   1 +
 testcases/kernel/syscalls/ioctl/ioctl09.c  | 119 +++++++++++++++++++++
 3 files changed, 121 insertions(+)
 create mode 100644 testcases/kernel/syscalls/ioctl/ioctl09.c
runtest/syscalls +1 -0
index 819e8d8ee..c2bfc6df3 100644
@@ -529,6 +529,7 @@ ioctl05      ioctl05
 ioctl06      ioctl06
 ioctl07      ioctl07
 ioctl08      ioctl08
+ioctl09      ioctl09
 
 ioctl_loop01 ioctl_loop01
 ioctl_loop02 ioctl_loop02
testcases/kernel/syscalls/ioctl/.gitignore +1 -0
index 3a3d49adc..5fff7a61d 100644
@@ -6,6 +6,7 @@
 /ioctl06
 /ioctl07
 /ioctl08
+/ioctl09
 /ioctl_loop01
 /ioctl_loop02
 /ioctl_loop03
testcases/kernel/syscalls/ioctl/ioctl09.c new file +119 -0

119 lines not on this page · raw message

Signature · 1 line
2.23.0