[PATCH 1/9] xfstests: fsstress add command line style output for show_opts

[PATCH 1/9] xfstests: fsstress add command line style output for show_opts

[Dmitry Monakhov]

Once some combination of seed+fs_ops result in regression it is
reasonable to document that combination. It is usefull to dump
that configuration in command line style. Later this line may be
simply hardcoded in to regression test.

Signed-off-by: Dmitry Monakhov <dmonakhov@openvz.org>
---
 070            |    2 +-
 117            |    2 +-
 ltp/fsstress.c |   21 +++++++++++++++++----
 3 files changed, 19 insertions(+), 6 deletions(-)

diff --git a/070 b/070
index 4458011..f48c33c 100755
--- a/070
+++ b/070
@@ -62,7 +62,7 @@ $FSSTRESS_PROG \
 	-f unresvsp=0 \
 	-f attr_set=100 \
 	-f attr_remove=100 \
-	-S -p 1 -n 10000 >$seq.full 2>&1
+        -p 1 -n 10000 -S c >$seq.full 2>&1
 
 status=$?
 exit
diff --git a/117 b/117
index fcf0485..edec157 100755
--- a/117
+++ b/117
@@ -85,7 +85,7 @@ while [ $i -lt $ITERATIONS ]; do
 	-f attr_set=100 \
 	-f attr_remove=100 \
         -s $seed \
-	-S -p 1 -n 1000 >>$seq.full 2>&1
+	-p 1 -n 1000 -S c >>$seq.full 2>&1
 
     let i=$i+1
 done
diff --git a/ltp/fsstress.c b/ltp/fsstress.c
index 133a247..f9d7d6e 100644
--- a/ltp/fsstress.c
+++ b/ltp/fsstress.c
@@ -284,7 +284,7 @@ int main(int argc, char **argv)
 	nops = sizeof(ops) / sizeof(ops[0]);
 	ops_end = &ops[nops];
 	myprog = argv[0];
-	while ((c = getopt(argc, argv, "d:e:f:i:m:n:o:p:rs:vwzHS")) != -1) {
+	while ((c = getopt(argc, argv, "d:e:f:i:m:n:o:p:rs:S:vwzH")) != -1) {
 		switch (c) {
 		case 'd':
 			dirname = optarg;
@@ -345,7 +345,10 @@ int main(int argc, char **argv)
 			zero_freq();
 			break;
 		case 'S':
-			show_ops(0, NULL);
+			i = 0;
+			if (optarg[0] == 'c')
+				i = 1;
+			show_ops(1, NULL);
 			printf("\n");
                         nousage=1;
 			break;
@@ -1310,7 +1313,8 @@ show_ops(int flag, char *lead_str)
                         x+=printf("%s ", p->name);
                 }
                 printf("\n");
-        } else {
+        } else if (flag == 0) {
+		/* Table view style */
 	        int		f;
 	        for (f = 0, p = ops; p < ops_end; p++)
 		        f += p->freq;
@@ -1327,6 +1331,15 @@ show_ops(int flag, char *lead_str)
 			        (p->iswrite == 0) ? " " : "write op");
 		        }
                 }
+	} else {
+		/* Command line style */
+		if (lead_str != NULL)
+			printf("%s", lead_str);
+		printf ("-z -s %ld -m %d -n %d -p %d \\\n", seed, idmodulo,
+			operations, nproc);
+	        for (p = ops; p < ops_end; p++)
+		        if (p->freq > 0)
+			        printf("-f %s=%d \\\n",p->name, p->freq);
 	}
 }
 
@@ -1433,7 +1446,7 @@ usage(void)
 	printf("   -v               specifies verbose mode\n");
 	printf("   -w               zeros frequencies of non-write operations\n");
 	printf("   -z               zeros frequencies of all operations\n");
-	printf("   -S               prints the table of operations (omitting zero frequency)\n");
+	printf("   -S [c,t]         prints the list of operations (omitting zero frequency) in command line or table style\n");
 	printf("   -H               prints usage and exits\n");
 }
 
-- 
1.7.1

[PATCH 2/9] xfstests: freeze fsstress options for 117'th

[Dmitry Monakhov]

This is regression tescase for explicit combination of seed+fs_opts.
Let's hardcode all options to prevent interactions with fsstress
changes in future.

Signed-off-by: Dmitry Monakhov <dmonakhov@openvz.org>
---
 117 |   39 ++++++++++++++++++++++++++++-----------
 1 files changed, 28 insertions(+), 11 deletions(-)

diff --git a/117 b/117
index edec157..2e3f658 100755
--- a/117
+++ b/117
@@ -33,8 +33,34 @@ tmp=/tmp/$$
 status=1	# failure is the default!
 trap "_cleanup; exit \$status" 0 1 2 3 15
 
-# seed for fsstress known to cause curruptions
+# seed and options for fsstress known to cause curruptions
 seed=1122413794
+fss_ops=" -z \
+    -f attr_remove=100 \
+    -f attr_set=100 \
+    -f chown=3 \
+    -f creat=4 \
+    -f dread=4 \
+    -f dwrite=4 \
+    -f fdatasync=1 \
+    -f fsync=1 \
+    -f getdents=1 \
+    -f link=1 \
+    -f mkdir=2 \
+    -f mknod=2 \
+    -f read=1 \
+    -f readlink=1 \
+    -f rename=2 \
+    -f rmdir=1 \
+    -f setxattr=1 \
+    -f stat=1 \
+    -f symlink=2 \
+    -f sync=1 \
+    -f truncate=2 \
+    -f unlink=1 \
+    -f write=4 \
+    -s $seed \
+    -m 31 -n 1000 -p 1"
 
 ITERATIONS=10
 
@@ -76,16 +102,7 @@ while [ $i -lt $ITERATIONS ]; do
     echo fsstress iteration: $i | tee -a $seq.full
     $FSSTRESS_PROG \
 	-d $SCRATCH_MNT/fsstress \
-	-f allocsp=0 \
-	-f freesp=0 \
-	-f bulkstat=0 \
-	-f bulkstat1=0 \
-	-f resvsp=0 \
-	-f unresvsp=0 \
-	-f attr_set=100 \
-	-f attr_remove=100 \
-        -s $seed \
-	-p 1 -n 1000 -S c >>$seq.full 2>&1
+	$fss_ops -S c >>$seq.full 2>&1
 
     let i=$i+1
 done
-- 
1.7.1

[PATCH 3/9] xfstests: add fallocate support to fsstress

[Dmitry Monakhov]

Add tests for fallocate(2) syscall
- fallocate: reserve the disk space
- punch: de-allocates the disk space
Since FALLOC_FL_PUNCH_HOLE is relatively new it's value defined
explicitly if not yet defined. Later we may clear that define.

Signed-off-by: Dmitry Monakhov <dmonakhov@openvz.org>
---
 ltp/fsstress.c |  127 +++++++++++++++++++++++++++++++++++++++++++++++++++++++-
 1 files changed, 126 insertions(+), 1 deletions(-)

diff --git a/ltp/fsstress.c b/ltp/fsstress.c
index f9d7d6e..e9e4e67 100644
--- a/ltp/fsstress.c
+++ b/ltp/fsstress.c
@@ -24,7 +24,13 @@
 #ifdef HAVE_ATTR_ATTRIBUTES_H
 #include <attr/attributes.h>
 #endif
-
+#ifdef FALLOCATE
+#include <linux/falloc.h>
+#ifndef FALLOC_FL_PUNCH_HOLE
+/* Copy-paste from linux/falloc.h */
+#define FALLOC_FL_PUNCH_HOLE    0x02 /* de-allocates range */
+#endif
+#endif
 #ifndef HAVE_ATTR_LIST
 #define attr_list(path, buf, size, flags, cursor) (errno = -ENOSYS, -1)
 #endif
@@ -48,6 +54,7 @@ typedef enum {
 	OP_CREAT,
 	OP_DREAD,
 	OP_DWRITE,
+	OP_FALLOCATE,
 	OP_FDATASYNC,
 	OP_FREESP,
 	OP_FSYNC,
@@ -55,6 +62,7 @@ typedef enum {
 	OP_LINK,
 	OP_MKDIR,
 	OP_MKNOD,
+	OP_PUNCH,
 	OP_READ,
 	OP_READLINK,
 	OP_RENAME,
@@ -128,6 +136,7 @@ void	chown_f(int, long);
 void	creat_f(int, long);
 void	dread_f(int, long);
 void	dwrite_f(int, long);
+void	fallocate_f(int, long);
 void	fdatasync_f(int, long);
 void	freesp_f(int, long);
 void	fsync_f(int, long);
@@ -135,6 +144,7 @@ void	getdents_f(int, long);
 void	link_f(int, long);
 void	mkdir_f(int, long);
 void	mknod_f(int, long);
+void	punch_f(int, long);
 void	read_f(int, long);
 void	readlink_f(int, long);
 void	rename_f(int, long);
@@ -159,6 +169,7 @@ opdesc_t	ops[] = {
 	{ OP_CREAT, "creat", creat_f, 4, 1 },
 	{ OP_DREAD, "dread", dread_f, 4, 0 },
 	{ OP_DWRITE, "dwrite", dwrite_f, 4, 1 },
+	{ OP_FALLOCATE, "fallocate", fallocate_f, 1, 1 },
 	{ OP_FDATASYNC, "fdatasync", fdatasync_f, 1, 1 },
 	{ OP_FREESP, "freesp", freesp_f, 1, 1 },
 	{ OP_FSYNC, "fsync", fsync_f, 1, 1 },
@@ -166,6 +177,7 @@ opdesc_t	ops[] = {
 	{ OP_LINK, "link", link_f, 1, 1 },
 	{ OP_MKDIR, "mkdir", mkdir_f, 2, 1 },
 	{ OP_MKNOD, "mknod", mknod_f, 2, 1 },
+	{ OP_PUNCH, "punch", punch_f, 1, 1 },
 	{ OP_READ, "read", read_f, 1, 0 },
 	{ OP_READLINK, "readlink", readlink_f, 1, 0 },
 	{ OP_RENAME, "rename", rename_f, 2, 1 },
@@ -2011,6 +2023,63 @@ dwrite_f(int opno, long r)
 }
 
 void
+fallocate_f(int opno, long r)
+{
+#ifdef FALLOCATE
+	int		e;
+	pathname_t	f;
+	int		fd;
+	__int64_t	lr;
+	off64_t		off;
+	off64_t		len;
+	struct stat64	stb;
+	int		v;
+	char		st[1024];
+	int mode = 0;
+
+	init_pathname(&f);
+	if (!get_fname(FT_REGFILE, r, &f, NULL, NULL, &v)) {
+		if (v)
+			printf("%d/%d: fallocate - no filename\n", procid, opno);
+		free_pathname(&f);
+		return;
+	}
+	fd = open_path(&f, O_RDWR);
+	e = fd < 0 ? errno : 0;
+	check_cwd();
+	if (fd < 0) {
+		if (v)
+			printf("%d/%d: fallocate - open %s failed %d\n",
+				procid, opno, f.path, e);
+		free_pathname(&f);
+		return;
+	}
+	if (fstat64(fd, &stb) < 0) {
+		if (v)
+			printf("%d/%d: fallocate - fstat64 %s failed %d\n",
+				procid, opno, f.path, errno);
+		free_pathname(&f);
+		close(fd);
+		return;
+	}
+	inode_info(st, sizeof(st), &stb, v);
+	lr = ((__int64_t)random() << 32) + random();
+	off = (off64_t)(lr % MIN(stb.st_size + (1024 * 1024), MAXFSIZE));
+	off %= maxfsize;
+	len = (off64_t)(random() % (1024 * 1024));
+	mode |= FALLOC_FL_KEEP_SIZE & random();
+	e = fallocate(fd, mode, (loff_t)off, (loff_t)len) < 0 ? errno : 0;
+	if (v)
+		printf("%d/%d: fallocate(%d) %s %st %lld %lld %d\n",
+		       procid, opno, mode,
+		       f.path, st, (long long)off, (long long)len, e);
+	free_pathname(&f);
+	close(fd);
+#endif
+}
+
+
+void
 fdatasync_f(int opno, long r)
 {
 	int		e;
@@ -2283,6 +2352,62 @@ mknod_f(int opno, long r)
 }
 
 void
+punch_f(int opno, long r)
+{
+#ifdef FALLOCATE
+	int		e;
+	pathname_t	f;
+	int		fd;
+	__int64_t	lr;
+	off64_t		off;
+	off64_t		len;
+	struct stat64	stb;
+	int		v;
+	char		st[1024];
+	int mode = FALLOC_FL_PUNCH_HOLE;
+
+	init_pathname(&f);
+	if (!get_fname(FT_REGFILE, r, &f, NULL, NULL, &v)) {
+		if (v)
+			printf("%d/%d: punch hole - no filename\n", procid, opno);
+		free_pathname(&f);
+		return;
+	}
+	fd = open_path(&f, O_RDWR);
+	e = fd < 0 ? errno : 0;
+	check_cwd();
+	if (fd < 0) {
+		if (v)
+			printf("%d/%d: punch hole - open %s failed %d\n",
+				procid, opno, f.path, e);
+		free_pathname(&f);
+		return;
+	}
+	if (fstat64(fd, &stb) < 0) {
+		if (v)
+			printf("%d/%d: punch hole - fstat64 %s failed %d\n",
+				procid, opno, f.path, errno);
+		free_pathname(&f);
+		close(fd);
+		return;
+	}
+	inode_info(st, sizeof(st), &stb, v);
+	lr = ((__int64_t)random() << 32) + random();
+	off = (off64_t)(lr % MIN(stb.st_size + (1024 * 1024), MAXFSIZE));
+	off %= maxfsize;
+	len = (off64_t)(random() % (1024 * 1024));
+	mode |= FALLOC_FL_KEEP_SIZE & random();
+	e = fallocate(fd, mode, (loff_t)off, (loff_t)len) < 0 ? errno : 0;
+	if (v)
+		printf("%d/%d: punch hole(%d) %s %s %lld %lld %d\n",
+		       procid, opno, mode,
+		       f.path, st, (long long)off, (long long)len, e);
+	free_pathname(&f);
+	close(fd);
+#endif
+}
+
+void
 read_f(int opno, long r)
 {
 	char		*buf;
-- 
1.7.1

[PATCH 4/9] xfstests: fsstress add FS_IOC_{SET,GET}FLAGS operations v2

[Dmitry Monakhov]

Add two new operations:
- getattr: ioctl(fd, FS_IOC_GETFLAGS, &fl)
- setattr: ioctl(fd, FS_IOC_SETFLAGS, &random_flags)
Attribute mask may be passed via -M opt, by default is (~0).
By default FS_IOC_SETFLAGS has zero probability because
it may produce inodes with APPEND or IMMUTABLE flags which
are not deletable by default. Let's assumes that one who
enable it knows how to delete such inodes.
For example like follows:
find $TEST_PATH -exec chattr -i -a {} \;
rm -rf $TEST_PATH

Signed-off-by: Dmitry Monakhov <dmonakhov@openvz.org>
---
 aclocal.m4     |    4 +++
 configure.in   |    1 +
 ltp/fsstress.c |   68 +++++++++++++++++++++++++++++++++++++++++++++++++++++++-
 3 files changed, 72 insertions(+), 1 deletions(-)

diff --git a/aclocal.m4 b/aclocal.m4
index 5532606..5739004 100644
--- a/aclocal.m4
+++ b/aclocal.m4
@@ -21,6 +21,10 @@ AC_DEFUN([AC_PACKAGE_WANT_LINUX_PRCTL_H],
     AC_SUBST(have_prctl)
   ])
 
+AC_DEFUN([AC_PACKAGE_WANT_LINUX_FS_H],
+  [ AC_CHECK_HEADER([linux/fs.h])
+  ])
+
 AC_DEFUN([AC_PACKAGE_WANT_FALLOCATE],
   [ AC_MSG_CHECKING([for fallocate])
     AC_TRY_LINK([
diff --git a/configure.in b/configure.in
index 76d23e4..3b40e55 100644
--- a/configure.in
+++ b/configure.in
@@ -68,6 +68,7 @@ in
 		AC_PACKAGE_WANT_LINUX_FIEMAP_H
 		AC_PACKAGE_WANT_FALLOCATE
 		AC_PACKAGE_WANT_LINUX_PRCTL_H
+		AC_PACKAGE_WANT_LINUX_FS_H
 		;;
 esac
 
diff --git a/ltp/fsstress.c b/ltp/fsstress.c
index e9e4e67..c258ee9 100644
--- a/ltp/fsstress.c
+++ b/ltp/fsstress.c
@@ -37,6 +37,15 @@
 #ifdef HAVE_SYS_PRCTL_H
 #include <sys/prctl.h>
 #endif
+
+#include <linux/fs.h>
+#ifndef FS_IOC_GETFLAGS
+#define FS_IOC_GETFLAGS                 _IOR('f', 1, long)
+#endif
+#ifndef FS_IOC_SETFLAGS
+#define FS_IOC_SETFLAGS                 _IOW('f', 2, long)
+#endif
+
 #include <math.h>
 #define XFS_ERRTAG_MAX		17
 #define XFS_IDMODULO_MAX	31	/* user/group IDs (1 << x)  */
@@ -58,6 +67,7 @@ typedef enum {
 	OP_FDATASYNC,
 	OP_FREESP,
 	OP_FSYNC,
+	OP_GETATTR,
 	OP_GETDENTS,
 	OP_LINK,
 	OP_MKDIR,
@@ -68,6 +78,7 @@ typedef enum {
 	OP_RENAME,
 	OP_RESVSP,
 	OP_RMDIR,
+	OP_SETATTR,
 	OP_SETXATTR,
 	OP_STAT,
 	OP_SYMLINK,
@@ -140,6 +151,7 @@ void	fallocate_f(int, long);
 void	fdatasync_f(int, long);
 void	freesp_f(int, long);
 void	fsync_f(int, long);
+void	getattr_f(int, long);
 void	getdents_f(int, long);
 void	link_f(int, long);
 void	mkdir_f(int, long);
@@ -150,6 +162,7 @@ void	readlink_f(int, long);
 void	rename_f(int, long);
 void	resvsp_f(int, long);
 void	rmdir_f(int, long);
+void	setattr_f(int, long);
 void	setxattr_f(int, long);
 void	stat_f(int, long);
 void	symlink_f(int, long);
@@ -173,6 +186,7 @@ opdesc_t	ops[] = {
 	{ OP_FDATASYNC, "fdatasync", fdatasync_f, 1, 1 },
 	{ OP_FREESP, "freesp", freesp_f, 1, 1 },
 	{ OP_FSYNC, "fsync", fsync_f, 1, 1 },
+	{ OP_GETATTR, "getattr", getattr_f, 1, 0 },
 	{ OP_GETDENTS, "getdents", getdents_f, 1, 0 },
 	{ OP_LINK, "link", link_f, 1, 1 },
 	{ OP_MKDIR, "mkdir", mkdir_f, 2, 1 },
@@ -183,6 +197,7 @@ opdesc_t	ops[] = {
 	{ OP_RENAME, "rename", rename_f, 2, 1 },
 	{ OP_RESVSP, "resvsp", resvsp_f, 1, 1 },
 	{ OP_RMDIR, "rmdir", rmdir_f, 1, 1 },
+	{ OP_SETATTR, "setattr", setattr_f, 0, 1 },
 	{ OP_SETXATTR, "setxattr", setxattr_f, 1, 1 },
 	{ OP_STAT, "stat", stat_f, 1, 0 },
 	{ OP_SYMLINK, "symlink", symlink_f, 2, 1 },
@@ -218,6 +233,7 @@ int		nops;
 int		nproc = 1;
 int		operations = 1;
 unsigned int	idmodulo = XFS_IDMODULO_MAX;
+unsigned int	attr_mask = ~0;
 int		procid;
 int		rtpct;
 unsigned long	seed = 0;
@@ -296,7 +312,7 @@ int main(int argc, char **argv)
 	nops = sizeof(ops) / sizeof(ops[0]);
 	ops_end = &ops[nops];
 	myprog = argv[0];
-	while ((c = getopt(argc, argv, "d:e:f:i:m:n:o:p:rs:S:vwzH")) != -1) {
+	while ((c = getopt(argc, argv, "d:e:f:i:m:M:n:o:p:rs:S:vwzH")) != -1) {
 		switch (c) {
 		case 'd':
 			dirname = optarg;
@@ -356,6 +372,9 @@ int main(int argc, char **argv)
 		case 'z':
 			zero_freq();
 			break;
+		case 'M':
+		  	attr_mask = strtoul(optarg, NULL, 0);
+			break;
 		case 'S':
 			i = 0;
 			if (optarg[0] == 'c')
@@ -2198,6 +2217,29 @@ fsync_f(int opno, long r)
 }
 
 void
+getattr_f(int opno, long r)
+{
+	int		fd;
+	int		e;
+	pathname_t	f;
+	uint		fl;
+	int		v;
+
+	init_pathname(&f);
+	if (!get_fname(FT_ANYm, r, &f, NULL, NULL, &v))
+		append_pathname(&f, ".");
+	fd = open_path(&f, O_RDWR);
+	e = fd < 0 ? errno : 0;
+	check_cwd();
+
+	e = ioctl(fd, FS_IOC_GETFLAGS, &fl);
+	if (v)
+		printf("%d/%d: getattr %s %u %d\n", procid, opno, f.path, fl, e);
+	free_pathname(&f);
+	close(fd);
+}
+
+void
 getdents_f(int opno, long r)
 {
 	DIR		*dir;
@@ -2643,6 +2685,30 @@ rmdir_f(int opno, long r)
 }
 
 void
+setattr_f(int opno, long r)
+{
+	int		fd;
+	int		e;
+	pathname_t	f;
+	uint		fl;
+	int		v;
+
+	init_pathname(&f);
+	if (!get_fname(FT_ANYm, r, &f, NULL, NULL, &v))
+		append_pathname(&f, ".");
+	fd = open_path(&f, O_RDWR);
+	e = fd < 0 ? errno : 0;
+	check_cwd();
+
+	fl = attr_mask & (uint)random();
+	e = ioctl(fd, FS_IOC_SETFLAGS, &fl);
+	if (v)
+		printf("%d/%d: setattr %s %x %d\n", procid, opno, f.path, fl, e);
+	free_pathname(&f);
+	close(fd);
+}
+
+void
 stat_f(int opno, long r)
 {
 	int		e;
-- 
1.7.1

[PATCH 5/9] xfstests: add fiemap operation to fsstress

[Dmitry Monakhov]

Related bug: http://patchwork.ozlabs.org/patch/118863

Signed-off-by: Dmitry Monakhov <dmonakhov@openvz.org>
---
 ltp/fsstress.c |   76 ++++++++++++++++++++++++++++++++++++++++++++++++++++++++
 1 files changed, 76 insertions(+), 0 deletions(-)

diff --git a/ltp/fsstress.c b/ltp/fsstress.c
index c258ee9..09c9529 100644
--- a/ltp/fsstress.c
+++ b/ltp/fsstress.c
@@ -24,6 +24,9 @@
 #ifdef HAVE_ATTR_ATTRIBUTES_H
 #include <attr/attributes.h>
 #endif
+#ifdef HAVE_LINUX_FIEMAP_H
+#include <linux/fiemap.h>
+#endif
 #ifdef FALLOCATE
 #include <linux/falloc.h>
 #ifndef FALLOC_FL_PUNCH_HOLE
@@ -65,6 +68,7 @@ typedef enum {
 	OP_DWRITE,
 	OP_FALLOCATE,
 	OP_FDATASYNC,
+	OP_FIEMAP,
 	OP_FREESP,
 	OP_FSYNC,
 	OP_GETATTR,
@@ -149,6 +153,7 @@ void	dread_f(int, long);
 void	dwrite_f(int, long);
 void	fallocate_f(int, long);
 void	fdatasync_f(int, long);
+void	fiemap_f(int, long);
 void	freesp_f(int, long);
 void	fsync_f(int, long);
 void	getattr_f(int, long);
@@ -184,6 +189,7 @@ opdesc_t	ops[] = {
 	{ OP_DWRITE, "dwrite", dwrite_f, 4, 1 },
 	{ OP_FALLOCATE, "fallocate", fallocate_f, 1, 1 },
 	{ OP_FDATASYNC, "fdatasync", fdatasync_f, 1, 1 },
+	{ OP_FIEMAP, "fiemap", fiemap_f, 1, 1 },
 	{ OP_FREESP, "freesp", freesp_f, 1, 1 },
 	{ OP_FSYNC, "fsync", fsync_f, 1, 1 },
 	{ OP_GETATTR, "getattr", getattr_f, 1, 0 },
@@ -2130,6 +2136,76 @@ fdatasync_f(int opno, long r)
 	free_pathname(&f);
 	close(fd);
 }
+void
+fiemap_f(int opno, long r)
+{
+#ifdef HAVE_LINUX_FIEMAP_H
+	int		e;
+	pathname_t	f;
+	int		fd;
+	__int64_t	lr;
+	off64_t		off;
+	struct stat64	stb;
+	int		v;
+	char		st[1024];
+	int blocks_to_map;
+	struct fiemap *fiemap;
+
+	init_pathname(&f);
+	if (!get_fname(FT_REGFILE, r, &f, NULL, NULL, &v)) {
+		if (v)
+			printf("%d/%d: fiemap - no filename\n", procid, opno);
+		free_pathname(&f);
+		return;
+	}
+	fd = open_path(&f, O_RDWR);
+	e = fd < 0 ? errno : 0;
+	check_cwd();
+	if (fd < 0) {
+		if (v)
+			printf("%d/%d: fiemap - open %s failed %d\n",
+				procid, opno, f.path, e);
+		free_pathname(&f);
+		return;
+	}
+	if (fstat64(fd, &stb) < 0) {
+		if (v)
+			printf("%d/%d: fiemap - fstat64 %s failed %d\n",
+				procid, opno, f.path, errno);
+		free_pathname(&f);
+		close(fd);
+		return;
+	}
+	inode_info(st, sizeof(st), &stb, v);
+	blocks_to_map = random() & 0xffff;
+	fiemap = (struct fiemap *)malloc(sizeof(struct fiemap) +
+			(blocks_to_map * sizeof(struct fiemap_extent)));
+	if (!fiemap) {
+		if (v)
+			printf("%d/%d: malloc failed \n", procid, opno);
+		free_pathname(&f);
+		close(fd);
+		return;
+	}
+	lr = ((__int64_t)random() << 32) + random();
+	off = (off64_t)(lr % MIN(stb.st_size + (1024 * 1024), MAXFSIZE));
+	off %= maxfsize;
+	fiemap->fm_flags = random() & (FIEMAP_FLAGS_COMPAT | 0x10000);
+	fiemap->fm_extent_count = blocks_to_map;
+	fiemap->fm_mapped_extents = random() & 0xffff;
+	fiemap->fm_start = off;
+	fiemap->fm_length = ((__int64_t)random() << 32) + random();
+
+	e = ioctl(fd, FS_IOC_FIEMAP, (unsigned long)fiemap);
+	if (v)
+		printf("%d/%d: ioctl(FIEMAP) %s%s %lld %lld %x %d\n",
+		       procid, opno, f.path, st, (long long)fiemap->fm_start,
+		       (long long) fiemap->fm_length, fiemap->fm_flags, e);
+	free(fiemap);
+	free_pathname(&f);
+	close(fd);
+#endif
+}
 
 void
 freesp_f(int opno, long r)
-- 
1.7.1

[PATCH 6/9] xfstests: add a new test that runs fsstress under ENOSPC conditions

[Dmitry Monakhov]

During stress testing we want to cover as much code paths as possible
fsstress is very good for this purpose. But it has expandable nature
(disk usage almost continually grow). So once it goes in no ENOSPC
condition it will be where till the end. But by running 'dd' writers
in parallel we can regularly trigger ENOSPC but only for a limited
periods of time because each time it opens the same file with O_TRUNC.

Signed-off-by: Dmitry Monakhov <dmonakhov@openvz.org>
---
 269     |   86 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
 269.out |    5 +++
 group   |    1 +
 3 files changed, 92 insertions(+), 0 deletions(-)
 create mode 100755 269
 create mode 100644 269.out

diff --git a/269 b/269
new file mode 100755
index 0000000..61e83ad
--- /dev/null
+++ b/269
@@ -0,0 +1,86 @@
+#! /bin/bash
+# FSQA Test No. 269
+#
+# Run fsstress and ENSPC hitters in parallel, check fs consistency an the end
+#
+#-----------------------------------------------------------------------
+# Copyright (c) 2006 Silicon Graphics, Inc.  All Rights Reserved.
+#
+# 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.
+#
+# 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
+#
+#-----------------------------------------------------------------------
+#
+# creator
+owner=dmonakhov@openvz.org
+
+seq=`basename $0`
+echo "QA output created by $seq"
+
+here=`pwd`
+tmp=/tmp/$$
+status=1	# failure is the default!
+trap "rm -f $tmp.*; exit \$status" 0 1 2 3 15
+
+# get standard environment, filters and checks
+. ./common.rc
+. ./common.filter
+# Disable all sync operations to get higher load
+FSSTRESS_AVOID="$FSSTRESS_AVOID -ffsync=0 -fsync=0 -ffdatasync=0"
+_workout()
+{
+	echo ""
+	echo "Run fsstress"
+	echo ""
+	num_iterations=10
+	enospc_time=2
+	out=$SCRATCH_MNT/fsstress.$$
+	args="-p128 -n999999999 -f setattr=1 $FSSTRESS_AVOID -d $out"
+	echo "fsstress $args" >> $here/$seq.full
+	$FSSTRESS_PROG $args > /dev/null 2>&1 &
+	pid=$!
+	echo "Run dd writers in parallel"
+	for ((i=0; i < num_iterations; i++))
+	do
+		# File will be opened with O_TRUNC each time
+		dd if=/dev/zero of=$SCRATCH_MNT/SPACE_CONSUMER bs=1M count=1 \
+			> /dev/null 2>&1
+		sleep $enospc_time
+	done
+	kill $pid
+	wait $pid
+}
+
+# real QA test starts here
+_supported_fs generic
+_supported_os Linux
+_require_scratch
+_need_to_be_root
+
+umount $SCRATCH_DEV 2>/dev/null
+_scratch_mkfs_sized $((512 * 1024 * 1024)) >> $seq.full 2>&1
+_scratch_mount
+
+if ! _workout; then
+	umount $SCRATCH_DEV 2>/dev/null
+	exit
+fi
+
+if ! _scratch_unmount; then
+	echo "failed to umount"
+	status=1
+	exit
+fi
+_check_scratch_fs
+status=$?
+exit
diff --git a/269.out b/269.out
new file mode 100644
index 0000000..0a83965
--- /dev/null
+++ b/269.out
@@ -0,0 +1,5 @@
+QA output created by 269
+
+Run fsstress
+
+Run dd writers in parallel
diff --git a/group b/group
index aaa8eba..8a5c334 100644
--- a/group
+++ b/group
@@ -382,3 +382,4 @@ deprecated
 266 dump ioctl auto quick
 267 dump ioctl tape
 268 dump ioctl tape
+269 auto rw prealloc ioctl enospc
-- 
1.7.1

[PATCH 7/9] xfstests: add a new quota test that runs fsstress under ENOSPC conditions

[Dmitry Monakhov]

- Same as 264 but with quota enabled.
- IO performed from $qa_user user
- fsstress granted with CAP_CHOWN capability.

Signed-off-by: Dmitry Monakhov <dmonakhov@openvz.org>
---
 270     |  108 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
 270.out |    8 +++++
 group   |    1 +
 3 files changed, 117 insertions(+), 0 deletions(-)
 create mode 100755 270
 create mode 100644 270.out

diff --git a/270 b/270
new file mode 100755
index 0000000..ed1e31e
--- /dev/null
+++ b/270
@@ -0,0 +1,108 @@
+#! /bin/bash
+# FSQA Test No. 270
+#
+# Run fsstress and ENSPC hitters in parallel, check quota and 
+# fs consistency an the end
+#
+#-----------------------------------------------------------------------
+# Copyright (c) 2006 Silicon Graphics, Inc.  All Rights Reserved.
+#
+# 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.
+#
+# 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
+#
+#-----------------------------------------------------------------------
+#
+# creator
+owner=dmonakhov@openvz.org
+
+seq=`basename $0`
+echo "QA output created by $seq"
+
+here=`pwd`
+tmp=/tmp/$$
+status=1	# failure is the default!
+trap "rm -f $tmp.*; exit \$status" 0 1 2 3 15
+
+# get standard environment, filters and checks
+. ./common.rc
+. ./common.filter
+. ./common.quota
+
+# Disable all sync operations to get higher load
+FSSTRESS_AVOID="$FSSTRESS_AVOID -ffsync=0 -fsync=0 -ffdatasync=0"
+_workout()
+{
+	echo ""
+	echo "Run fsstress"
+	echo ""
+	num_iterations=10
+	enospc_time=2
+	out=$SCRATCH_MNT/fsstress.$$
+	args="-p128 -n999999999 -f setattr=1 $FSSTRESS_AVOID -d $out"
+	echo "fsstress $args" >> $here/$seq.full
+	# Grant chown capability 
+	cp $FSSTRESS_PROG  $tmp.fsstress.bin
+	setcap cap_chown=epi  $tmp.fsstress.bin
+
+	(su $qa_user -c "$tmp.fsstress.bin $args" &) > /dev/null 2>&1
+	pid=$!
+
+	echo "Run dd writers in parallel"
+	for ((i=0; i < num_iterations; i++))
+	do
+		# File will be opened with O_TRUNC each time
+		su $qa_user -c "dd if=/dev/zero \
+			of=$SCRATCH_MNT/SPACE_CONSUMER bs=1M " \
+			> /dev/null 2>&1
+		sleep $enospc_time
+	done
+
+	killall $tmp.fsstress.bin
+	wait $pid
+}
+
+# real QA test starts here
+_supported_fs generic
+_supported_os Linux
+_require_scratch
+_require_quota
+_require_user
+_need_to_be_root
+
+umount $SCRATCH_DEV 2>/dev/null
+_scratch_mkfs_sized $((512 * 1024 * 1024)) >> $seq.full 2>&1
+_scratch_mount "-o usrquota,grpquota"
+chmod 777 $SCRATCH_MNT
+quotacheck -u -g $SCRATCH_MNT 2>/dev/null
+quotaon -u -g $SCRATCH_MNT 2>/dev/null
+
+if ! _workout; then
+	_scratch_unmount 2>/dev/null
+	exit
+fi
+
+if ! _check_quota_usage; then
+	_scratch_unmount 2>/dev/null
+	status=1
+	exit
+fi
+
+echo Comparing filesystem consistency
+if ! _scratch_unmount; then
+	echo "failed to umount"
+	status=1
+	exit
+fi
+_check_scratch_fs
+status=$?
+exit
diff --git a/270.out b/270.out
new file mode 100644
index 0000000..bbd8026
--- /dev/null
+++ b/270.out
@@ -0,0 +1,8 @@
+QA output created by 270
+
+Run fsstress
+
+Run dd writers in parallel
+Comparing user usage
+Comparing group usage
+Comparing filesystem consistency
diff --git a/group b/group
index 8a5c334..84e81f5 100644
--- a/group
+++ b/group
@@ -383,3 +383,4 @@ deprecated
 267 dump ioctl tape
 268 dump ioctl tape
 269 auto rw prealloc ioctl enospc
+270 auto quota rw prealloc ioctl enospc
-- 
1.7.1

[PATCH 8/9] xfstress: add regression testcase for d583fb87a3ff0

[Dmitry Monakhov]

Should be fixed by following patch http://patchwork.ozlabs.org/patch/120883/

Signed-off-by: Dmitry Monakhov <dmonakhov@openvz.org>
---
 271     |   69 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
 271.out |   44 ++++++++++++++++++++++++++++++++++++++++
 group   |    1 +
 3 files changed, 114 insertions(+), 0 deletions(-)
 create mode 100755 271
 create mode 100644 271.out

diff --git a/271 b/271
new file mode 100755
index 0000000..b4e6043
--- /dev/null
+++ b/271
@@ -0,0 +1,69 @@
+#! /bin/bash
+# FSQA Test No. 271
+#
+# Regression testcase for d583fb87a3ff0 (ext4 extent corruption)
+#
+#-----------------------------------------------------------------------
+# Copyright (c) 2006 Silicon Graphics, Inc.  All Rights Reserved.
+#
+# 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.
+#
+# 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
+#
+#-----------------------------------------------------------------------
+#
+# creator
+owner=dmonakhov@openvz.org
+
+seq=`basename $0`
+echo "QA output created by $seq"
+
+here=`pwd`
+tmp=/tmp/$$
+status=1	# failure is the default!
+trap "rm -f $tmp.*; exit \$status" 0 1 2 3 15
+
+# get standard environment, filters and checks
+. ./common.rc
+. ./common.filter
+
+# real QA test starts here
+_supported_fs ext4
+_supported_os Linux
+_require_scratch
+_need_to_be_root
+
+umount $SCRATCH_DEV 2>/dev/null
+_scratch_mkfs_sized $((128 * 1024 * 1024)) >> $seq.full 2>&1
+
+# -onoload and EXT4_SYNC_FL on file is important becase result in
+# metadata sync writes inside ext4_handle_dirty_metadata()
+_scratch_mount -onoload
+touch $SCRATCH_MNT/file
+chattr +S $SCRATCH_MNT/file
+# Create sparse file 
+for ((i = 0; i < 21; i++))
+do
+    dd if=/dev/zero of=$SCRATCH_MNT/file bs=4k count=1 seek=$((i*10))
+done
+# truncate last extent
+truncate $SCRATCH_MNT/file --size $((4096* 200))
+
+if ! _scratch_unmount; then
+	echo "failed to umount"
+	status=1
+	exit
+fi
+echo "Check filesystem"
+_check_scratch_fs
+status=$?
+exit
diff --git a/271.out b/271.out
new file mode 100644
index 0000000..22a3c50
--- /dev/null
+++ b/271.out
@@ -0,0 +1,44 @@
+QA output created by 271
+1+0 records in
+1+0 records out
+1+0 records in
+1+0 records out
+1+0 records in
+1+0 records out
+1+0 records in
+1+0 records out
+1+0 records in
+1+0 records out
+1+0 records in
+1+0 records out
+1+0 records in
+1+0 records out
+1+0 records in
+1+0 records out
+1+0 records in
+1+0 records out
+1+0 records in
+1+0 records out
+1+0 records in
+1+0 records out
+1+0 records in
+1+0 records out
+1+0 records in
+1+0 records out
+1+0 records in
+1+0 records out
+1+0 records in
+1+0 records out
+1+0 records in
+1+0 records out
+1+0 records in
+1+0 records out
+1+0 records in
+1+0 records out
+1+0 records in
+1+0 records out
+1+0 records in
+1+0 records out
+1+0 records in
+1+0 records out
+Check filesystem
diff --git a/group b/group
index 84e81f5..fefa580 100644
--- a/group
+++ b/group
@@ -384,3 +384,4 @@ deprecated
 268 dump ioctl tape
 269 auto rw prealloc ioctl enospc
 270 auto quota rw prealloc ioctl enospc
+271 auto rw quick
-- 
1.7.1

[PATCH 9/9] xfstress: Test data journaling flag switch for a single file

[Dmitry Monakhov]

Signed-off-by: Dmitry Monakhov <dmonakhov@openvz.org>
---
 272     |  108 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
 272.out |   37 +++++++++++++++++++++
 group   |    1 +
 3 files changed, 146 insertions(+), 0 deletions(-)
 create mode 100755 272
 create mode 100644 272.out

diff --git a/272 b/272
new file mode 100755
index 0000000..a35ef83
--- /dev/null
+++ b/272
@@ -0,0 +1,108 @@
+#! /bin/bash
+# FSQA Test No. 272
+#
+# Test data journaling flag switch for a single file  
+#
+#-----------------------------------------------------------------------
+# Copyright (c) 2006 Silicon Graphics, Inc.  All Rights Reserved.
+#
+# 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.
+#
+# 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
+#
+#-----------------------------------------------------------------------
+#
+# creator
+owner=dmonakhov@openvz.org
+
+seq=`basename $0`
+echo "QA output created by $seq"
+
+here=`pwd`
+tmp=/tmp/$$
+status=1	# failure is the default!
+trap "rm -f $tmp.*; exit \$status" 0 1 2 3 15
+
+# get standard environment, filters and checks
+. ./common.rc
+. ./common.filter
+_workout()
+{
+	echo ""
+	echo " Switch data journalling mode"
+	echo ""
+	write_opt_list="iflag=noatime conv=notrunc conv=fsync oflag=direct"
+	chattr_opt_list="+j -j +jS -j"
+	idx=0
+	#
+	for write_opt in $write_opt_list
+	do
+		idx=$((idx + 1))
+		for chattr_opt in $chattr_opt_list
+		do
+		    
+			echo "OP write_opt: $write_opt 4M, \
+chattr_opt: $chattr_opt"
+			dd if=/dev/zero of=$SCRATCH_MNT/file.$idx \
+				bs=1M count=4 $write_opt \
+				>> $seq.full 2>&1 || exit
+			chattr $chattr_opt $SCRATCH_MNT/file.$idx >> $seq.full \
+				|| exit
+		done
+	done
+	sync
+	# Same as previous test, but for one file,
+	# and with ENOSPC triggering
+	for write_opt in $write_opt_list
+	do
+		idx=$((idx + 1))
+		for chattr_opt in $chattr_opt_list
+		do
+
+			echo "OP write_opt: $write_opt ENOSPC, \
+chattr_opt: $chattr_opt"
+			dd if=/dev/zero of=$SCRATCH_MNT/file.$idx \
+				bs=1M $write_opt >> $seq.full 2>&1
+			chattr $chattr_opt $SCRATCH_MNT/file.$idx \
+				>> $seq.full || exit
+		done
+		sync
+		unlink $SCRATCH_MNT/file.$idx
+	done
+}
+
+# real QA test starts here
+_supported_fs ext3 ext4
+_supported_os Linux
+_require_scratch
+_need_to_be_root
+
+umount $SCRATCH_DEV 2>/dev/null
+_scratch_mkfs_sized $((64 * 1024 * 1024)) >> $seq.full 2>&1
+_scratch_mount
+
+if ! _workout; then
+	echo "workout failed"
+	_scratch_unmount
+	status=1
+	exit
+fi
+
+if ! _scratch_unmount; then
+	echo "failed to umount"
+	status=1
+	exit
+fi
+echo "Check filesystem"
+_check_scratch_fs
+status=$?
+exit
diff --git a/272.out b/272.out
new file mode 100644
index 0000000..cc687aa
--- /dev/null
+++ b/272.out
@@ -0,0 +1,37 @@
+QA output created by 272
+
+ Switch data journalling mode
+
+OP write_opt: iflag=noatime 4M, chattr_opt: +j
+OP write_opt: iflag=noatime 4M, chattr_opt: -j
+OP write_opt: iflag=noatime 4M, chattr_opt: +jS
+OP write_opt: iflag=noatime 4M, chattr_opt: -j
+OP write_opt: conv=notrunc 4M, chattr_opt: +j
+OP write_opt: conv=notrunc 4M, chattr_opt: -j
+OP write_opt: conv=notrunc 4M, chattr_opt: +jS
+OP write_opt: conv=notrunc 4M, chattr_opt: -j
+OP write_opt: conv=fsync 4M, chattr_opt: +j
+OP write_opt: conv=fsync 4M, chattr_opt: -j
+OP write_opt: conv=fsync 4M, chattr_opt: +jS
+OP write_opt: conv=fsync 4M, chattr_opt: -j
+OP write_opt: oflag=direct 4M, chattr_opt: +j
+OP write_opt: oflag=direct 4M, chattr_opt: -j
+OP write_opt: oflag=direct 4M, chattr_opt: +jS
+OP write_opt: oflag=direct 4M, chattr_opt: -j
+OP write_opt: iflag=noatime ENOSPC, chattr_opt: +j
+OP write_opt: iflag=noatime ENOSPC, chattr_opt: -j
+OP write_opt: iflag=noatime ENOSPC, chattr_opt: +jS
+OP write_opt: iflag=noatime ENOSPC, chattr_opt: -j
+OP write_opt: conv=notrunc ENOSPC, chattr_opt: +j
+OP write_opt: conv=notrunc ENOSPC, chattr_opt: -j
+OP write_opt: conv=notrunc ENOSPC, chattr_opt: +jS
+OP write_opt: conv=notrunc ENOSPC, chattr_opt: -j
+OP write_opt: conv=fsync ENOSPC, chattr_opt: +j
+OP write_opt: conv=fsync ENOSPC, chattr_opt: -j
+OP write_opt: conv=fsync ENOSPC, chattr_opt: +jS
+OP write_opt: conv=fsync ENOSPC, chattr_opt: -j
+OP write_opt: oflag=direct ENOSPC, chattr_opt: +j
+OP write_opt: oflag=direct ENOSPC, chattr_opt: -j
+OP write_opt: oflag=direct ENOSPC, chattr_opt: +jS
+OP write_opt: oflag=direct ENOSPC, chattr_opt: -j
+Check filesystem
diff --git a/group b/group
index fefa580..616cbdd 100644
--- a/group
+++ b/group
@@ -385,3 +385,4 @@ deprecated
 269 auto rw prealloc ioctl enospc
 270 auto quota rw prealloc ioctl enospc
 271 auto rw quick
+272 auto enospc rw
-- 
1.7.1

Re: [PATCH 7/9] xfstests: add a new quota test that runs fsstress under ENOSPC conditions

[Christoph Hellwig]

On Thu, Nov 03, 2011 at 09:55:51PM +0400, Dmitry Monakhov wrote:
> - Same as 264 but with quota enabled.

269 is the new 270

> - IO performed from $qa_user user
> - fsstress granted with CAP_CHOWN capability.

This fails in my test VM because it doesn't have setcap.  If you
want to use additional non-standard tools please do a _notrun if
they aren't present.

I don't quite understand why it would need CAP_CHOWN - chowns
from owned files do not require it, and the test runs fine with
the setcap line commented out.

I'll commit it as-is for now, but unless I get some feedback on
why the setcap is require I'll add another commit to remove it ASAP.

Re: [PATCH 7/9] xfstests: add a new quota test that runs fsstress under ENOSPC conditions

[Dmitry Monakhov]

[-- Attachment #1: Type: text/plain, Size: 1215 bytes --]

On Thu, 3 Nov 2011 15:39:35 -0400, Christoph Hellwig <hch@infradead.org> wrote:
> On Thu, Nov 03, 2011 at 09:55:51PM +0400, Dmitry Monakhov wrote:
> > - Same as 264 but with quota enabled.
> 
> 269 is the new 270
> 
> > - IO performed from $qa_user user
> > - fsstress granted with CAP_CHOWN capability.
> 
> This fails in my test VM because it doesn't have setcap.  If you
> want to use additional non-standard tools please do a _notrun if
> they aren't present.
> 
> I don't quite understand why it would need CAP_CHOWN - chowns
> from owned files do not require it, and the test runs fine with
> the setcap line commented out.
Yes, but fsstress want chown to random uid/gid 
int inode_change_ok(..)
...
        /* Make sure a caller can chown. */
        if ((ia_valid & ATTR_UID) &&
            (current_fsuid() != inode->i_uid ||
             attr->ia_uid != inode->i_uid) && !capable(CAP_CHOWN))
                return -EPERM;
>From my experience fssstess with chown enabled is the most valuable
quota accounting test. W/o chown this is just an useless time consumer.
> 
> I'll commit it as-is for now, but unless I get some feedback on
> why the setcap is require I'll add another commit to remove it ASAP.


[-- Attachment #2: 0001-xfstests-270-th-test-want-use-setcap-8.patch --]
[-- Type: text/plain, Size: 1012 bytes --]

>From f8611f3bbc5ea2ac3672ee6c88d2d43d69351d49 Mon Sep 17 00:00:00 2001
From: Dmitry Monakhov <dmonakhov@openvz.org>
Date: Mon, 7 Nov 2011 08:07:57 +0300
Subject: [PATCH] xfstests: 270'th test want use setcap(8)

It wants to grant CAP_CHOWN to fsstress, without this capability
chown(2) will fail due to EPERM. Since chown(2) is one of the most
error prone places of quota accounting this makes 270'th
test almost useless, so it is better to simply skip it.

Signed-off-by: Dmitry Monakhov <dmonakhov@openvz.org>
---
 270 |    3 +++
 1 files changed, 3 insertions(+), 0 deletions(-)

diff --git a/270 b/270
index ed1e31e..90e24f5 100755
--- a/270
+++ b/270
@@ -52,6 +52,9 @@ _workout()
 	echo "fsstress $args" >> $here/$seq.full
 	# Grant chown capability 
 	cp $FSSTRESS_PROG  $tmp.fsstress.bin
+	if [ "`whereis setcap`" == "setcap:" ]; then
+            _notrun "setcap not installed."
+	fi
 	setcap cap_chown=epi  $tmp.fsstress.bin
 
 	(su $qa_user -c "$tmp.fsstress.bin $args" &) > /dev/null 2>&1
-- 
1.7.1


[-- Attachment #3: Type: text/plain, Size: 208 bytes --]

> 
> --
> To unsubscribe from this list: send the line "unsubscribe linux-fsdevel" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [PATCH 6/9] xfstests: add a new test that runs fsstress under ENOSPC conditions

[David Sterba]

Hi,

I saw this while running the newly added tests on btrfs:

On Thu, Nov 03, 2011 at 09:55:50PM +0400, Dmitry Monakhov wrote:
> --- /dev/null
> +++ b/269
> +_workout()
> +{
> +	echo ""
> +	echo "Run fsstress"
> +	echo ""
> +	num_iterations=10
> +	enospc_time=2
> +	out=$SCRATCH_MNT/fsstress.$$
> +	args="-p128 -n999999999 -f setattr=1 $FSSTRESS_AVOID -d $out"
> +	echo "fsstress $args" >> $here/$seq.full
> +	$FSSTRESS_PROG $args > /dev/null 2>&1 &
> +	pid=$!
> +	echo "Run dd writers in parallel"
> +	for ((i=0; i < num_iterations; i++))
> +	do
> +		# File will be opened with O_TRUNC each time
> +		dd if=/dev/zero of=$SCRATCH_MNT/SPACE_CONSUMER bs=1M count=1 \
> +			> /dev/null 2>&1
> +		sleep $enospc_time
> +	done
> +	kill $pid

269      [failed, exit status 1] - output mismatch (see 269.out.bad)
--- 269.out     2011-11-07 13:26:23.000000000 +0100
+++ 269.out.bad 2011-11-07 14:07:44.000000000 +0100
@@ -3,3 +3,4 @@
 Run fsstress

 Run dd writers in parallel
+./269: line 60: kill: (9754) - No such process

> +	wait $pid
> +}

and the test is reported to fail. Seems that the dd commands finish in less 2
seconds.


david

Re: [PATCH 6/9] xfstests: add a new test that runs fsstress under ENOSPC conditions

[Dmitry Monakhov]

On Mon, 7 Nov 2011 14:25:33 +0100, David Sterba <dave@jikos.cz> wrote:
> Hi,
> 
> I saw this while running the newly added tests on btrfs:
> 
> On Thu, Nov 03, 2011 at 09:55:50PM +0400, Dmitry Monakhov wrote:
> > --- /dev/null
> > +++ b/269
> > +_workout()
> > +{
> > +	echo ""
> > +	echo "Run fsstress"
> > +	echo ""
> > +	num_iterations=10
> > +	enospc_time=2
> > +	out=$SCRATCH_MNT/fsstress.$$
> > +	args="-p128 -n999999999 -f setattr=1 $FSSTRESS_AVOID -d $out"
> > +	echo "fsstress $args" >> $here/$seq.full
> > +	$FSSTRESS_PROG $args > /dev/null 2>&1 &
> > +	pid=$!
> > +	echo "Run dd writers in parallel"
> > +	for ((i=0; i < num_iterations; i++))
> > +	do
> > +		# File will be opened with O_TRUNC each time
> > +		dd if=/dev/zero of=$SCRATCH_MNT/SPACE_CONSUMER bs=1M count=1 \
> > +			> /dev/null 2>&1
> > +		sleep $enospc_time
> > +	done
> > +	kill $pid
> 
> 269      [failed, exit status 1] - output mismatch (see 269.out.bad)
> --- 269.out     2011-11-07 13:26:23.000000000 +0100
> +++ 269.out.bad 2011-11-07 14:07:44.000000000 +0100
> @@ -3,3 +3,4 @@
>  Run fsstress
> 
>  Run dd writers in parallel
> +./269: line 60: kill: (9754) - No such process
> 
> > +	wait $pid
> > +}
> 
> and the test is reported to fail. Seems that the dd commands finish in less 2
> seconds.
Have you recompile fsstress via make?
test explicitly want use new feature "-f setattr=1"
> 
> 
> david
> --
> To unsubscribe from this list: send the line "unsubscribe linux-fsdevel" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [PATCH 1/9] xfstests: fsstress add command line style output for show_opts

[Dave Chinner]

On Thu, Nov 03, 2011 at 09:55:45PM +0400, Dmitry Monakhov wrote:
> Once some combination of seed+fs_ops result in regression it is
> reasonable to document that combination. It is usefull to dump
> that configuration in command line style. Later this line may be
> simply hardcoded in to regression test.
.....

> @@ -345,7 +345,10 @@ int main(int argc, char **argv)
>  			zero_freq();
>  			break;
>  		case 'S':
> -			show_ops(0, NULL);
> +			i = 0;
> +			if (optarg[0] == 'c')
> +				i = 1;
> +			show_ops(1, NULL);

I don't think that does what it is supposed to do...

Cheers,

Dave.
-- 
Dave Chinner
david@fromorbit.com

Re: [PATCH 6/9] xfstests: add a new test that runs fsstress under ENOSPC conditions

[David Sterba]

On Mon, Nov 07, 2011 at 05:31:11PM +0400, Dmitry Monakhov wrote:
> > 269      [failed, exit status 1] - output mismatch (see 269.out.bad)
> > --- 269.out     2011-11-07 13:26:23.000000000 +0100
> > +++ 269.out.bad 2011-11-07 14:07:44.000000000 +0100
> > @@ -3,3 +3,4 @@
> >  Run fsstress
> > 
> >  Run dd writers in parallel
> > +./269: line 60: kill: (9754) - No such process
> > 
> > > +	wait $pid
> > > +}
> > 
> > and the test is reported to fail. Seems that the dd commands finish in less 2
> > seconds.
> Have you recompile fsstress via make?

That was it, thanks.

david

Re: [PATCH 6/9] xfstests: add a new test that runs fsstress under ENOSPC conditions

[Christoph Hellwig]

On Mon, Nov 07, 2011 at 05:31:11PM +0400, Dmitry Monakhov wrote:
> Have you recompile fsstress via make?
> test explicitly want use new feature "-f setattr=1"

Btw, one useful thing for these patches would be to clearly error
out if fsstress does't recognize the option.