[LTP-FAIL][02/21] fs: refactor ksys_umount

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



Hi Christoph,

We have seen LTP test(utime06 and umount03) failure in Next Master with commit Id 41525f56e256("fs: refactor ksys_umount").
I didn't analysis root cause of problem, i am reporting this issue.

---------------------------------------------
LTP Testcase    Result    Exit Value
----------------------------------- ---------
umount03          FAIL           4
utime06             FAIL           2
--------------------------------------------

LTP utime06 Fail Log:
/dev/loop0 is mounted; will not make a filesystem here!
utime06     0  TINFO  :  Using test device LTP_DEV='/dev/loop0'
utime06     0  TINFO  :  Formatting /dev/loop0 with ext2 opts='' extra opts='' utime06     1  TBROK  :  tst_mkfs.c:102: utime06.c:122: mkfs.ext2 failed with 1
utime06     2  TBROK  :  tst_mkfs.c:102: Remaining cases broken

LTP umount03 Fail Log:
tst_device.c:262: INFO: Using test device LTP_DEV='/dev/loop0'
tst_mkfs.c:89: INFO: Formatting /dev/loop0 with ext2 opts='' extra opts=''
mke2fs 1.44.5 (15-Dec-2018)
tst_test.c:1244: INFO: Timeout per run is 0h 05m 00s
umount03.c:35: PASS: umount() fails as expected: EPERM (1)
tst_device.c:383: INFO: umount('mntpoint') failed with EBUSY, try 1...
tst_device.c:387: INFO: Likely gvfsd-trash is probing newly mounted fs, kill it to speed up tests.
tst_device.c:383: INFO: umount('mntpoint') failed with EBUSY, try 2...
tst_device.c:383: INFO: umount('mntpoint') failed with EBUSY, try 3...
tst_device.c:383: INFO: umount('mntpoint') failed with EBUSY, try 48...
tst_device.c:383: INFO: umount('mntpoint') failed with EBUSY, try 49...
tst_device.c:383: INFO: umount('mntpoint') failed with EBUSY, try 50...
tst_device.c:394: WARN: Failed to umount('mntpoint') after 50 retries
tst_tmpdir.c:336: WARN: tst_rmdir: rmobj(/scratch/ltp-Lnmh7tbxY6/gx0hJU) failed: remove(/scratch/ltp-Lnmh7tbxY6/gx0hJU/mntpoint) failed; errno=16: EBUSY

Regards,

Vikas


Below Commit ID 41525f56e256 Bisected for This fail:

    commit 41525f56e2564c2feff4fb2824823900efb3a39f
    Author: Christoph Hellwig <hch@xxxxxx>
    Date:   Tue Jul 21 10:54:34 2020 +0200

    fs: refactor ksys_umount

    Factor out a path_umount helper that takes a struct path * instead of the     actual file name.  This will allow to convert the init and devtmpfs code
    to properly mount based on a kernel pointer instead of relying on the
    implicit set_fs(KERNEL_DS) during early init.

    Signed-off-by: Christoph Hellwig <hch@xxxxxx>

---
 fs/namespace.c | 40 ++++++++++++++++++----------------------
 1 file changed, 18 insertions(+), 22 deletions(-)

diff --git a/fs/namespace.c b/fs/namespace.c
index 6f8234f74bed90..43834b59eff6c3 100644
--- a/fs/namespace.c
+++ b/fs/namespace.c
@@ -1706,36 +1706,19 @@ static inline bool may_mandlock(void)
 }
 #endif

-/*
- * Now umount can handle mount points as well as block devices.
- * This is important for filesystems which use unnamed block devices.
- *
- * We now support a flag for forced unmount like the other 'big iron'
- * unixes. Our API is identical to OSF/1 to avoid making a mess of AMD
- */
-
-int ksys_umount(char __user *name, int flags)
+static int path_umount(struct path *path, int flags)
 {
-    struct path path;
     struct mount *mnt;
     int retval;
-    int lookup_flags = LOOKUP_MOUNTPOINT;

     if (flags & ~(MNT_FORCE | MNT_DETACH | MNT_EXPIRE | UMOUNT_NOFOLLOW))
         return -EINVAL;
-
     if (!may_mount())
         return -EPERM;

-    if (!(flags & UMOUNT_NOFOLLOW))
-        lookup_flags |= LOOKUP_FOLLOW;
-
-    retval = user_path_at(AT_FDCWD, name, lookup_flags, &path);
-    if (retval)
-        goto out;
-    mnt = real_mount(path.mnt);
+    mnt = real_mount(path->mnt);
     retval = -EINVAL;
-    if (path.dentry != path.mnt->mnt_root)
+    if (path->dentry != path->mnt->mnt_root)
         goto dput_and_out;
     if (!check_mnt(mnt))
         goto dput_and_out;
@@ -1748,12 +1731,25 @@ int ksys_umount(char __user *name, int flags)
     retval = do_umount(mnt, flags);
 dput_and_out:
     /* we mustn't call path_put() as that would clear mnt_expiry_mark */
-    dput(path.dentry);
+    dput(path->dentry);
     mntput_no_expire(mnt);
-out:
     return retval;
 }

+int ksys_umount(char __user *name, int flags)
+{
+    int lookup_flags = LOOKUP_MOUNTPOINT;
+    struct path path;
+    int ret;
+
+    if (!(flags & UMOUNT_NOFOLLOW))
+        lookup_flags |= LOOKUP_FOLLOW;
+    ret = user_path_at(AT_FDCWD, name, lookup_flags, &path);
+    if (ret)
+        return ret;
+    return path_umount(&path, flags);
+}
+
SYSCALL_DEFINE2(umount, char __user *, name, int, flags)
 {





[Index of Archives]     [Linux Ext4 Filesystem]     [Union Filesystem]     [Filesystem Testing]     [Ceph Users]     [Ecryptfs]     [AutoFS]     [Kernel Newbies]     [Share Photos]     [Security]     [Netfilter]     [Bugtraq]     [Yosemite News]     [MIPS Linux]     [ARM Linux]     [Linux Security]     [Linux Cachefs]     [Reiser Filesystem]     [Linux RAID]     [Samba]     [Device Mapper]     [CEPH Development]

  Powered by Linux