summaryrefslogtreecommitdiffstats
path: root/fs
Commit message (Collapse)AuthorAgeFilesLines
* gfs2: Convert gfs2_quota_refresh to take a kqidEric W. Biederman2013-02-133-5/+16
| | | | | | | | | | | - In quota_refresh_user_store convert the user supplied uid into a kqid and pass it to gfs2_quota_refresh. - In quota_refresh_group_store convert the user supplied gid into a kqid and pass it to gfs2_quota_refresh. Cc: Steven Whitehouse <swhiteho@redhat.com> Signed-off-by: "Eric W. Biederman" <ebiederm@xmission.com>
* gfs2: Modify qdsb_get to take a struct kqidEric W. Biederman2013-02-131-6/+7
| | | | | Cc: Steven Whitehouse <swhiteho@redhat.com> Signed-off-by: "Eric W. Biederman" <ebiederm@xmission.com>
* gfs2: Modify struct gfs2_quota_change_host to use struct kqidEric W. Biederman2013-02-131-3/+5
| | | | | Cc: Steven Whitehouse <swhiteho@redhat.com> Signed-off-by: "Eric W. Biederman" <ebiederm@xmission.com>
* gfs2: Introduce qd2indexEric W. Biederman2013-02-131-2/+8
| | | | | | | | | Both qd_alloc and qd2offset perform the exact same computation to get an index from a gfs2_quota_data. Make life a little simpler and factor out this index computation. Cc: Steven Whitehouse <swhiteho@redhat.com> Signed-off-by: "Eric W. Biederman" <ebiederm@xmission.com>
* gfs2: Report quotas in the caller's user namespace.Eric W. Biederman2013-02-131-1/+1
| | | | | | | | | | When a quota is queried return the uid or the gid in the mapped into the caller's user namespace. In addition perform the munged version of the mapping so that instead of -1 a value that does not map is reported as the overflowuid or the overflowgid. Cc: Steven Whitehouse <swhiteho@redhat.com> Signed-off-by: "Eric W. Biederman" <ebiederm@xmission.com>
* gfs2: Split NO_QUOTA_CHANGE inot NO_UID_QUTOA_CHANGE and NO_GID_QUTOA_CHANGEEric W. Biederman2013-02-137-14/+15
| | | | | | | | Split NO_QUOTA_CHANGE into NO_UID_QUTOA_CHANGE and NO_GID_QUTOA_CHANGE so the constants may be well typed. Cc: Steven Whitehouse <swhiteho@redhat.com> Signed-off-by: "Eric W. Biederman" <ebiederm@xmission.com>
* gfs2: Remove improper checks in gfs2_set_dqblk.Eric W. Biederman2013-02-131-6/+0
| | | | | | | | | | | | | | | | | | | In set_dqblk it is an error to look at fdq->d_id or fdq->d_flags. Userspace quota applications do not set these fields when calling quotactl(Q_XSETQLIM,...), and the kernel does not set those fields when quota_setquota calls set_dqblk. gfs2 never looks at fdq->d_id or fdq->d_flags after checking to see if they match the id and type supplied to set_dqblk. No other linux filesystem in set_dqblk looks at either fdq->d_id or fdq->d_flags. Therefore remove these bogus checks from gfs2 and allow normal quota setting applications to work. Cc: Steven Whitehouse <swhiteho@redhat.com> Signed-off-by: "Eric W. Biederman" <ebiederm@xmission.com>
* ocfs2: Compare kuids and kgids using uid_eq and gid_eqEric W. Biederman2013-02-132-5/+5
| | | | | | Cc: Mark Fasheh <mfasheh@suse.com> Cc: Joel Becker <jlbec@evilplan.org> Signed-off-by: "Eric W. Biederman" <ebiederm@xmission.com>
* ocfs2: For tracing report the uid and gid values in the initial user namespaceEric W. Biederman2013-02-131-1/+2
| | | | | | Cc: Mark Fasheh <mfasheh@suse.com> Cc: Joel Becker <jlbec@evilplan.org> Signed-off-by: "Eric W. Biederman" <ebiederm@xmission.com>
* ocfs2: Convert uid and gids between in core and on disk inodesEric W. Biederman2013-02-132-8/+8
| | | | | | Cc: Mark Fasheh <mfasheh@suse.com> Cc: Joel Becker <jlbec@evilplan.org> Signed-off-by: "Eric W. Biederman" <ebiederm@xmission.com>
* ocfs2: convert between kuids and kgids and DLM locksEric W. Biederman2013-02-131-4/+4
| | | | | | | | | | Convert between uid and gids stored in the on the wire format of dlm locks aka struct ocfs2_meta_lvb and kuids and kgids stored in inode->i_uid and inode->i_gid. Cc: Mark Fasheh <mfasheh@suse.com> Cc: Joel Becker <jlbec@evilplan.org> Signed-off-by: "Eric W. Biederman" <ebiederm@xmission.com>
* ocfs2: Handle kuids and kgids in acl/xattr conversions.Eric W. Biederman2013-02-131-2/+29
| | | | | | | | | Explicitly deal with the different kinds of acls because they need different conversions. Cc: Mark Fasheh <mfasheh@suse.com> Cc: Joel Becker <jlbec@evilplan.org> Signed-off-by: "Eric W. Biederman" <ebiederm@xmission.com>
* coda: Cache permisions in struct coda_inode_info in a kuid_t.Eric W. Biederman2013-02-133-4/+4
| | | | | | | | | - Change c_uid in struct coda_indoe_info from a vuid_t to a kuid_t. - Initialize c_uid to GLOBAL_ROOT_UID instead of 0. - Use uid_eq to compare cached kuids. Cc: Jan Harkes <jaharkes@cs.cmu.edu> Signed-off-by: "Eric W. Biederman" <ebiederm@xmission.com>
* coda: Restrict coda messages to the initial user namespaceEric W. Biederman2013-02-133-7/+10
| | | | | | | | | | | | | | | | | | | | | | Remove the slight chance that uids and gids in coda messages will be interpreted in the wrong user namespace. - Only allow processes in the initial user namespace to open the coda character device to communicate with coda filesystems. - Explicitly convert the uids in the coda header into the initial user namespace. - In coda_vattr_to_attr make kuids and kgids from the initial user namespace uids and gids in struct coda_vattr that just came from userspace. - In coda_iattr_to_vattr convert kuids and kgids into uids and gids in the intial user namespace and store them in struct coda_vattr for sending to coda userspace programs. Nothing needs to be changed with mounts as coda does not support being mounted in anything other than the initial user namespace. Cc: Jan Harkes <jaharkes@cs.cmu.edu> Signed-off-by: "Eric W. Biederman" <ebiederm@xmission.com>
* coda: Restrict coda messages to the initial pid namespaceEric W. Biederman2013-02-133-2/+10
| | | | | | | | | | | | | | Remove the slight chance that pids in coda messages will be interpreted in the wrong pid namespace. - Explicitly send all pids in coda messages in the initial pid namespace. - Only allow mounts from processes in the initial pid namespace. - Only allow processes in the initial pid namespace to open the coda character device to communicate with coda. Cc: Jan Harkes <jaharkes@cs.cmu.edu> Signed-off-by: "Eric W. Biederman" <ebiederm@xmission.com>
* afs: Support interacting with multiple user namespacesEric W. Biederman2013-02-133-9/+15
| | | | | | | | | | | | | | | | | | | | | | Modify struct afs_file_status to store owner as a kuid_t and group as a kgid_t. In xdr_decode_AFSFetchStatus as owner is now a kuid_t and group is now a kgid_t don't use the EXTRACT macro. Instead perform the work of the extract macro explicitly. Read the value with ntohl and convert it to the appropriate type with make_kuid or make_kgid. Test if the value is different from what is stored in status and update changed. Update the value in status. In xdr_encode_AFS_StoreStatus call from_kuid or from_kgid as we are computing the on the wire encoding. Initialize uids with GLOBAL_ROOT_UID instead of 0. Initialize gids with GLOBAL_ROOT_GID instead of 0. Cc: David Howells <dhowells@redhat.com> Acked-by: Serge Hallyn <serge.hallyn@canonical.com> Signed-off-by: Eric W. Biederman <ebiederm@xmission.com>
* afs: Only allow mounting afs in the intial network namespaceEric W. Biederman2013-02-131-0/+6
| | | | | | | | rxrpc sockets only work in the initial network namespace so it isn't possible to support afs in any other network namespace. Cc: David Howells <dhowells@redhat.com> Signed-off-by: "Eric W. Biederman" <ebiederm@xmission.com>
* afs: Remove unused structure afs_store_statusEric W. Biederman2013-02-121-7/+0
| | | | | | | | | | | | | While looking for kuid_t and kgid_t conversions I found this structure that has never been used since it was added to the kernel in 2007. The obvious for this structure to be used is in xdr_encode_AFS_StoreStatus and that function uses a small handful of local variables instead. So remove the unnecessary structure to prevent confusion. Cc: David Howells <dhowells@redhat.com> Signed-off-by: "Eric W. Biederman" <ebiederm@xmission.com>
* 9p: Modify v9fs_get_fsgid_for_create to return a kgidEric W. Biederman2013-02-121-5/+5
| | | | | | | | | | | Modify v9fs_get_fsgid_for_create to return a kgid and modify all of the variables that hold the result of v9fs_get_fsgid_for_create to be of type kgid_t. Cc: Eric Van Hensbergen <ericvh@gmail.com> Cc: Ron Minnich <rminnich@gmail.com> Cc: Latchesar Ionkov <lucho@ionkov.net> Signed-off-by: "Eric W. Biederman" <ebiederm@xmission.com>
* 9p: Modify struct v9fs_session_info to use a kuids and kgidsEric W. Biederman2013-02-122-10/+30
| | | | | | | | | | | | | Change struct v9fs_session_info and the code that popluates it to use kuids and kgids. When parsing the 9p mount options convert the dfltuid, dflutgid, and the session uid from the current user namespace into kuids and kgids. Modify V9FS_DEFUID and V9FS_DEFGUID to be kuid and kgid values. Cc: Eric Van Hensbergen <ericvh@gmail.com> Cc: Ron Minnich <rminnich@gmail.com> Cc: Latchesar Ionkov <lucho@ionkov.net> Signed-off-by: "Eric W. Biederman" <ebiederm@xmission.com>
* 9p: Modify struct 9p_fid to use a kuid_t not a uid_tEric W. Biederman2013-02-122-9/+10
| | | | | | | | | | Change struct 9p_fid and it's associated functions to use kuid_t's instead of uid_t. Cc: Eric Van Hensbergen <ericvh@gmail.com> Cc: Ron Minnich <rminnich@gmail.com> Cc: Latchesar Ionkov <lucho@ionkov.net> Signed-off-by: "Eric W. Biederman" <ebiederm@xmission.com>
* 9p: Modify the stat structures to use kuid_t and kgid_tEric W. Biederman2013-02-121-3/+3
| | | | | | | | | | | | | | | | 9p has thre strucrtures that can encode inode stat information. Modify all of those structures to contain kuid_t and kgid_t values. Modify he wire encoders and decoders of those structures to use 'u' and 'g' instead of 'd' in the format string where uids and gids are present. This results in all kuid and kgid conversion to and from on the wire values being performed by the same code in protocol.c where the client is known at the time of the conversion. Cc: Eric Van Hensbergen <ericvh@gmail.com> Cc: Ron Minnich <rminnich@gmail.com> Cc: Latchesar Ionkov <lucho@ionkov.net> Signed-off-by: Eric W. Biederman <ebiederm@xmission.com>
* 9p: Transmit kuid and kgid valuesEric W. Biederman2013-02-121-1/+1
| | | | | | | | | | | | | | Modify the p9_client_rpc format specifiers of every function that directly transmits a uid or a gid from 'd' to 'u' or 'g' as appropriate. Modify those same functions to take kuid_t and kgid_t parameters instead of uid_t and gid_t parameters. Cc: Eric Van Hensbergen <ericvh@gmail.com> Cc: Ron Minnich <rminnich@gmail.com> Cc: Latchesar Ionkov <lucho@ionkov.net> Signed-off-by: Eric W. Biederman <ebiederm@xmission.com>
* ceph: Convert kuids and kgids before printing them.Eric W. Biederman2013-02-122-4/+8
| | | | | | | | Before printing kuid and kgids values convert them into the initial user namespace. Cc: Sage Weil <sage@inktank.com> Signed-off-by: "Eric W. Biederman" <ebiederm@xmission.com>
* ceph: Convert struct ceph_mds_request to use kuid_t and kgid_tEric W. Biederman2013-02-122-4/+4
| | | | | | | | | | Hold the uid and gid for a pending ceph mds request using the types kuid_t and kgid_t. When a request message is finally created convert the kuid_t and kgid_t values into uids and gids in the initial user namespace. Cc: Sage Weil <sage@inktank.com> Signed-off-by: "Eric W. Biederman" <ebiederm@xmission.com>
* ceph: Translate inode uid and gid attributes to/from kuids and kgids.Eric W. Biederman2013-02-121-6/+8
| | | | | | | | | | | | - In fill_inode() transate uids and gids in the initial user namespace into kuids and kgids stored in inode->i_uid and inode->i_gid. - In ceph_setattr() if they have changed convert inode->i_uid and inode->i_gid into initial user namespace uids and gids for transmission. Cc: Sage Weil <sage@inktank.com> Signed-off-by: "Eric W. Biederman" <ebiederm@xmission.com>
* ceph: Translate between uid and gids in cap messages and kuids and kgidsEric W. Biederman2013-02-122-9/+9
| | | | | | | | | | | | | | | | | | | | | - Make the uid and gid arguments of send_cap_msg() used to compose ceph_mds_caps messages of type kuid_t and kgid_t. - Pass inode->i_uid and inode->i_gid in __send_cap to send_cap_msg() through variables of type kuid_t and kgid_t. - Modify struct ceph_cap_snap to store uids and gids in types kuid_t and kgid_t. This allows capturing inode->i_uid and inode->i_gid in ceph_queue_cap_snap() without loss and pssing them to __ceph_flush_snaps() where they are removed from struct ceph_cap_snap and passed to send_cap_msg(). - In handle_cap_grant translate uid and gids in the initial user namespace stored in struct ceph_mds_cap into kuids and kgids before setting inode->i_uid and inode->i_gid. Cc: Sage Weil <sage@inktank.com> Signed-off-by: "Eric W. Biederman" <ebiederm@xmission.com>
* userns: Allow the userns root to mount ramfs.Eric W. Biederman2013-01-261-0/+1
| | | | | | | | | | | | | | There is no backing store to ramfs and file creation rules are the same as for any other filesystem so it is semantically safe to allow unprivileged users to mount it. The memory control group successfully limits how much memory ramfs can consume on any system that cares about a user namespace root using ramfs to exhaust memory the memory control group can be deployed. Signed-off-by: "Eric W. Biederman" <ebiederm@xmission.com>
* userns: Allow the userns root to mount of devptsEric W. Biederman2013-01-261-0/+18
| | | | | | | | | | | | | | | | - The context in which devpts is mounted has no effect on the creation of ptys as the /dev/ptmx interface has been used by unprivileged users for many years. - Only support unprivileged mounts in combination with the newinstance option to ensure that mounting of /dev/pts in a user namespace will not allow the options of an existing mount of devpts to be modified. - Create /dev/pts/ptmx as the root user in the user namespace that mounts devpts so that it's permissions to be changed. Acked-by: Serge Hallyn <serge.hallyn@canonical.com> Signed-off-by: "Eric W. Biederman" <ebiederm@xmission.com>
* f2fs: Don't assign e_id in f2fs_acl_from_diskEric W. Biederman2012-12-251-1/+0
| | | | | | | | | | | | | | | | | | | | | With user namespaces enabled building f2fs fails with: CC fs/f2fs/acl.o fs/f2fs/acl.c: In function ‘f2fs_acl_from_disk’: fs/f2fs/acl.c:85:21: error: ‘struct posix_acl_entry’ has no member named ‘e_id’ make[2]: *** [fs/f2fs/acl.o] Error 1 make[2]: Target `__build' not remade because of errors. e_id is a backwards compatibility field only used for file systems that haven't been converted to use kuids and kgids. When the posix acl tag field is neither ACL_USER nor ACL_GROUP assigning e_id is unnecessary. Remove the assignment so f2fs will build with user namespaces enabled. Cc: Namjae Jeon <namjae.jeon@samsung.com> Cc: Amit Sahrawat <a.sahrawat@samsung.com> Acked-by: Jaegeuk Kim <jaegeuk.kim@samsung.com> Signed-off-by: "Eric W. Biederman" <ebiederm@xmission.com>
* proc: Allow proc_free_inum to be called from any contextEric W. Biederman2012-12-251-6/+7
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | While testing the pid namespace code I hit this nasty warning. [ 176.262617] ------------[ cut here ]------------ [ 176.263388] WARNING: at /home/eric/projects/linux/linux-userns-devel/kernel/softirq.c:160 local_bh_enable_ip+0x7a/0xa0() [ 176.265145] Hardware name: Bochs [ 176.265677] Modules linked in: [ 176.266341] Pid: 742, comm: bash Not tainted 3.7.0userns+ #18 [ 176.266564] Call Trace: [ 176.266564] [<ffffffff810a539f>] warn_slowpath_common+0x7f/0xc0 [ 176.266564] [<ffffffff810a53fa>] warn_slowpath_null+0x1a/0x20 [ 176.266564] [<ffffffff810ad9ea>] local_bh_enable_ip+0x7a/0xa0 [ 176.266564] [<ffffffff819308c9>] _raw_spin_unlock_bh+0x19/0x20 [ 176.266564] [<ffffffff8123dbda>] proc_free_inum+0x3a/0x50 [ 176.266564] [<ffffffff8111d0dc>] free_pid_ns+0x1c/0x80 [ 176.266564] [<ffffffff8111d195>] put_pid_ns+0x35/0x50 [ 176.266564] [<ffffffff810c608a>] put_pid+0x4a/0x60 [ 176.266564] [<ffffffff8146b177>] tty_ioctl+0x717/0xc10 [ 176.266564] [<ffffffff810aa4d5>] ? wait_consider_task+0x855/0xb90 [ 176.266564] [<ffffffff81086bf9>] ? default_spin_lock_flags+0x9/0x10 [ 176.266564] [<ffffffff810cab0a>] ? remove_wait_queue+0x5a/0x70 [ 176.266564] [<ffffffff811e37e8>] do_vfs_ioctl+0x98/0x550 [ 176.266564] [<ffffffff810b8a0f>] ? recalc_sigpending+0x1f/0x60 [ 176.266564] [<ffffffff810b9127>] ? __set_task_blocked+0x37/0x80 [ 176.266564] [<ffffffff810ab95b>] ? sys_wait4+0xab/0xf0 [ 176.266564] [<ffffffff811e3d31>] sys_ioctl+0x91/0xb0 [ 176.266564] [<ffffffff810a95f0>] ? task_stopped_code+0x50/0x50 [ 176.266564] [<ffffffff81939199>] system_call_fastpath+0x16/0x1b [ 176.266564] ---[ end trace 387af88219ad6143 ]--- It turns out that spin_unlock_bh(proc_inum_lock) is not safe when put_pid is called with another spinlock held and irqs disabled. For now take the easy path and use spin_lock_irqsave(proc_inum_lock) in proc_free_inum and spin_loc_irq in proc_alloc_inum(proc_inum_lock). Signed-off-by: "Eric W. Biederman" <ebiederm@xmission.com>
* Merge branch 'for-next' of git://git.samba.org/sfrench/cifs-2.6Linus Torvalds2012-12-214-17/+18
|\ | | | | | | | | | | | | | | | | | | Pull CIFS fixes from Steve French: "Misc small cifs fixes" * 'for-next' of git://git.samba.org/sfrench/cifs-2.6: cifs: eliminate cifsERROR variable cifs: don't compare uniqueids in cifs_prime_dcache unless server inode numbers are in use cifs: fix double-free of "string" in cifs_parse_mount_options
| * cifs: eliminate cifsERROR variableJeff Layton2012-12-202-6/+1
| | | | | | | | | | | | | | It's always set to "1" and there's no way to change it to anything else. Signed-off-by: Jeff Layton <jlayton@redhat.com> Signed-off-by: Steve French <sfrench@us.ibm.com>
| * cifs: don't compare uniqueids in cifs_prime_dcache unless server inode ↵Jeff Layton2012-12-201-4/+15
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | numbers are in use Oliver reported that commit cd60042c caused his cifs mounts to continually thrash through new inodes on readdir. His servers are not sending inode numbers (or he's not using them), and the new test in that function doesn't account for that sort of setup correctly. If we're not using server inode numbers, then assume that the inode attached to the dentry hasn't changed. Go ahead and update the attributes in place, but keep the same inode number. Cc: <stable@vger.kernel.org> # v3.5+ Reported-and-Tested-by: Oliver Mössinger <Oliver.Moessinger@ichaus.de> Signed-off-by: Jeff Layton <jlayton@redhat.com> Signed-off-by: Steve French <sfrench@us.ibm.com>
| * cifs: fix double-free of "string" in cifs_parse_mount_optionsJeff Layton2012-12-201-7/+2
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Dan reported the following regression in commit d387a5c5: + fs/cifs/connect.c:1903 cifs_parse_mount_options() error: double free of 'string' That patch has some of the new option parsing code free "string" without setting the variable to NULL afterward. Since "string" is automatically freed in an error condition, fix the code to just rely on that instead of freeing it explicitly. Reported-by: Dan Carpenter <dan.carpenter@oracle.com> Signed-off-by: Jeff Layton <jlayton@redhat.com> Signed-off-by: Steve French <smfrench@gmail.com>
* | Revert "nfsd: warn on odd reply state in nfsd_vfs_read"J. Bruce Fields2012-12-211-1/+0
| | | | | | | | | | | | | | | | | | | | | | | | | | This reverts commit 79f77bf9a4e3dd5ead006b8f17e7c4ff07d8374e. This is obviously wrong, and I have no idea how I missed seeing the warning in testing: I must just not have looked at the right logs. The caller bumps rq_resused/rq_next_page, so it will always be hit on a large enough read. Reported-by: Dave Jones <davej@redhat.com> Signed-off-by: J. Bruce Fields <bfields@redhat.com> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
* | NFS: Kill fscache warnings when mounting without -ofscTrond Myklebust2012-12-211-4/+15
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | The fscache code will currently bleat a "non-unique superblock keys" warning even if the user is mounting without the 'fsc' option. There should be no reason to even initialise the superblock cache cookie unless we're planning on using fscache for something, so ensure that we check for the NFS_OPTION_FSCACHE flag before calling into the fscache code. Reported-by: Paweł Sikora <pawel.sikora@agmk.net> Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com> Cc: David Howells <dhowells@redhat.com> Acked-by: David Howells <dhowells@redhat.com> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
* | NFS: Provide stub nfs_fscache_wait_on_invalidate() for when CONFIG_NFS_FSCACHE=nDavid Howells2012-12-211-0/+1
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Provide a stub nfs_fscache_wait_on_invalidate() function for when CONFIG_NFS_FSCACHE=n lest the following error appear: fs/nfs/inode.c: In function 'nfs_invalidate_mapping': fs/nfs/inode.c:887:2: error: implicit declaration of function 'nfs_fscache_wait_on_invalidate' [-Werror=implicit-function-declaration] cc1: some warnings being treated as errors Reported-by: kbuild test robot <fengguang.wu@intel.com> Reported-by: Vineet Gupta <Vineet.Gupta1@synopsys.com> Reported-by: Borislav Petkov <bp@alien8.de> Signed-off-by: David Howells <dhowells@redhat.com> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
* | Merge branch 'for-next' of git://git.infradead.org/users/eparis/notifyLinus Torvalds2012-12-2011-104/+152
|\ \ | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Pull filesystem notification updates from Eric Paris: "This pull mostly is about locking changes in the fsnotify system. By switching the group lock from a spin_lock() to a mutex() we can now hold the lock across things like iput(). This fixes a problem involving unmounting a fs and having inodes be busy, first pointed out by FAT, but reproducible with tmpfs. This also restores signal driven I/O for inotify, which has been broken since about 2.6.32." Ugh. I *hate* the timing of this. It was rebased after the merge window opened, and then left to sit with the pull request coming the day before the merge window closes. That's just crap. But apparently the patches themselves have been around for over a year, just gathering dust, so now it's suddenly critical. Fixed up semantic conflict in fs/notify/fdinfo.c as per Stephen Rothwell's fixes from -next. * 'for-next' of git://git.infradead.org/users/eparis/notify: inotify: automatically restart syscalls inotify: dont skip removal of watch descriptor if creation of ignored event failed fanotify: dont merge permission events fsnotify: make fasync generic for both inotify and fanotify fsnotify: change locking order fsnotify: dont put marks on temporary list when clearing marks by group fsnotify: introduce locked versions of fsnotify_add_mark() and fsnotify_remove_mark() fsnotify: pass group to fsnotify_destroy_mark() fsnotify: use a mutex instead of a spinlock to protect a groups mark list fanotify: add an extra flag to mark_remove_from_mask that indicates wheather a mark should be destroyed fsnotify: take groups mark_lock before mark lock fsnotify: use reference counting for groups fsnotify: introduce fsnotify_get_group() inotify, fanotify: replace fsnotify_put_group() with fsnotify_destroy_group()
| * | inotify: automatically restart syscallsEric Paris2012-12-111-1/+1
| | | | | | | | | | | | | | | | | | | | | | | | | | | We were mistakenly returning EINTR when we found an outstanding signal. Instead we should returen ERESTARTSYS and allow the kernel to handle things the right way. Patch-from: Oleg Nesterov <oleg@redhat.com> Signed-off-by: Eric Paris <eparis@redhat.com>
| * | inotify: dont skip removal of watch descriptor if creation of ignored event ↵Lino Sanfilippo2012-12-111-5/+5
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | failed In inotify_ignored_and_remove_idr() the removal of a watch descriptor is skipped if the allocation of an ignored event failed and we are leaking memory (the watch descriptor and the mark linked to it). This patch ensures that the watch descriptor is removed regardless of whether event creation failed or not. Signed-off-by: Lino Sanfilippo <LinoSanfilippo@gmx.de> Signed-off-by: Eric Paris <eparis@redhat.com>
| * | fanotify: dont merge permission eventsLino Sanfilippo2012-12-111-0/+6
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Boyd Yang reported a problem for the case that multiple threads of the same thread group are waiting for a reponse for a permission event. In this case it is possible that some of the threads are never woken up, even if the response for the event has been received (see http://marc.info/?l=linux-kernel&m=131822913806350&w=2). The reason is that we are currently merging permission events if they belong to the same thread group. But we are not prepared to wake up more than one waiter for each event. We do wait_event(group->fanotify_data.access_waitq, event->response || atomic_read(&group->fanotify_data.bypass_perm)); and after that event->response = 0; which is the reason that even if we woke up all waiters for the same event some of them may see event->response being already set 0 again, then go back to sleep and block forever. With this patch we avoid that more than one thread is waiting for a response by not merging permission events for the same thread group any more. Reported-by: Boyd Yang <boyd.yang@gmail.com> Signed-off-by: Lino Sanfilippo <LinoSanfilipp@gmx.de> Signed-off-by: Eric Paris <eparis@redhat.com>
| * | fsnotify: make fasync generic for both inotify and fanotifyEric Paris2012-12-114-9/+16
| | | | | | | | | | | | | | | | | | | | | | | | inotify is supposed to support async signal notification when information is available on the inotify fd. This patch moves that support to generic fsnotify functions so it can be used by all notification mechanisms. Signed-off-by: Eric Paris <eparis@redhat.com>
| * | fsnotify: change locking orderLino Sanfilippo2012-12-111-10/+10
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | On Mon, Aug 01, 2011 at 04:38:22PM -0400, Eric Paris wrote: > > I finally built and tested a v3.0 kernel with these patches (I know I'm > SOOOOOO far behind). Not what I hoped for: > > > [ 150.937798] VFS: Busy inodes after unmount of tmpfs. Self-destruct in 5 seconds. Have a nice day... > > [ 150.945290] BUG: unable to handle kernel NULL pointer dereference at 0000000000000070 > > [ 150.946012] IP: [<ffffffff810ffd58>] shmem_free_inode+0x18/0x50 > > [ 150.946012] PGD 2bf9e067 PUD 2bf9f067 PMD 0 > > [ 150.946012] Oops: 0000 [#1] PREEMPT SMP DEBUG_PAGEALLOC > > [ 150.946012] CPU 0 > > [ 150.946012] Modules linked in: nfs lockd fscache auth_rpcgss nfs_acl sunrpc ip6t_REJECT nf_conntrack_ipv6 nf_defrag_ipv6 ip6table_filter ip6_tables ext4 jbd2 crc16 joydev ata_piix i2c_piix4 pcspkr uinput ipv6 autofs4 usbhid [last unloaded: scsi_wait_scan] > > [ 150.946012] > > [ 150.946012] Pid: 2764, comm: syscall_thrash Not tainted 3.0.0+ #1 Red Hat KVM > > [ 150.946012] RIP: 0010:[<ffffffff810ffd58>] [<ffffffff810ffd58>] shmem_free_inode+0x18/0x50 > > [ 150.946012] RSP: 0018:ffff88002c2e5df8 EFLAGS: 00010282 > > [ 150.946012] RAX: 000000004e370d9f RBX: 0000000000000000 RCX: ffff88003a029438 > > [ 150.946012] RDX: 0000000033630a5f RSI: 0000000000000000 RDI: ffff88003491c240 > > [ 150.946012] RBP: ffff88002c2e5e08 R08: 0000000000000000 R09: 0000000000000000 > > [ 150.946012] R10: 0000000000000000 R11: 0000000000000000 R12: ffff88003a029428 > > [ 150.946012] R13: ffff88003a029428 R14: ffff88003a029428 R15: ffff88003499a610 > > [ 150.946012] FS: 00007f5a05420700(0000) GS:ffff88003f600000(0000) knlGS:0000000000000000 > > [ 150.946012] CS: 0010 DS: 0000 ES: 0000 CR0: 000000008005003b > > [ 150.946012] CR2: 0000000000000070 CR3: 000000002a662000 CR4: 00000000000006f0 > > [ 150.946012] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000 > > [ 150.946012] DR3: 0000000000000000 DR6: 00000000ffff0ff0 DR7: 0000000000000400 > > [ 150.946012] Process syscall_thrash (pid: 2764, threadinfo ffff88002c2e4000, task ffff88002bfbc760) > > [ 150.946012] Stack: > > [ 150.946012] ffff88003a029438 ffff88003a029428 ffff88002c2e5e38 ffffffff81102f76 > > [ 150.946012] ffff88003a029438 ffff88003a029598 ffffffff8160f9c0 ffff88002c221250 > > [ 150.946012] ffff88002c2e5e68 ffffffff8115e9be ffff88002c2e5e68 ffff88003a029438 > > [ 150.946012] Call Trace: > > [ 150.946012] [<ffffffff81102f76>] shmem_evict_inode+0x76/0x130 > > [ 150.946012] [<ffffffff8115e9be>] evict+0x7e/0x170 > > [ 150.946012] [<ffffffff8115ee40>] iput_final+0xd0/0x190 > > [ 150.946012] [<ffffffff8115ef33>] iput+0x33/0x40 > > [ 150.946012] [<ffffffff81180205>] fsnotify_destroy_mark_locked+0x145/0x160 > > [ 150.946012] [<ffffffff81180316>] fsnotify_destroy_mark+0x36/0x50 > > [ 150.946012] [<ffffffff81181937>] sys_inotify_rm_watch+0x77/0xd0 > > [ 150.946012] [<ffffffff815aca52>] system_call_fastpath+0x16/0x1b > > [ 150.946012] Code: 67 4a 00 b8 e4 ff ff ff eb aa 66 0f 1f 84 00 00 00 00 00 55 48 89 e5 48 83 ec 10 48 89 1c 24 4c 89 64 24 08 48 8b 9f 40 05 00 00 > > [ 150.946012] 83 7b 70 00 74 1c 4c 8d a3 80 00 00 00 4c 89 e7 e8 d2 5d 4a > > [ 150.946012] RIP [<ffffffff810ffd58>] shmem_free_inode+0x18/0x50 > > [ 150.946012] RSP <ffff88002c2e5df8> > > [ 150.946012] CR2: 0000000000000070 > > Looks at aweful lot like the problem from: > http://www.spinics.net/lists/linux-fsdevel/msg46101.html > I tried to reproduce this bug with your test program, but without success. However, if I understand correctly, this occurs since we dont hold any locks when we call iput() in mark_destroy(), right? With the patches you tested, iput() is also not called within any lock, since the groups mark_mutex is released temporarily before iput() is called. This is, since the original codes behaviour is similar. However since we now have a mutex as the biggest lock, we can do what you suggested (http://www.spinics.net/lists/linux-fsdevel/msg46107.html) and call iput() with the mutex held to avoid the race. The patch below implements this. It uses nested locking to avoid deadlock in case we do the final iput() on an inode which still holds marks and thus would take the mutex again when calling fsnotify_inode_delete() in destroy_inode(). Signed-off-by: Lino Sanfilippo <LinoSanfilippo@gmx.de> Signed-off-by: Eric Paris <eparis@redhat.com>
| * | fsnotify: dont put marks on temporary list when clearing marks by groupLino Sanfilippo2012-12-111-8/+2
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | In clear_marks_by_group_flags() the mark list of a group is iterated and the marks are put on a temporary list. Since we introduced fsnotify_destroy_mark_locked() we dont need the temp list any more and are able to remove the marks while the mark list is iterated and the mark list mutex is held. Signed-off-by: Lino Sanfilippo <LinoSanfilippo@gmx.de> Signed-off-by: Eric Paris <eparis@redhat.com>
| * | fsnotify: introduce locked versions of fsnotify_add_mark() and ↵Lino Sanfilippo2012-12-111-12/+30
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | fsnotify_remove_mark() This patch introduces fsnotify_add_mark_locked() and fsnotify_remove_mark_locked() which are essentially the same as fsnotify_add_mark() and fsnotify_remove_mark() but assume that the caller has already taken the groups mark mutex. Signed-off-by: Lino Sanfilippo <LinoSanfilippo@gmx.de> Signed-off-by: Eric Paris <eparis@redhat.com>
| * | fsnotify: pass group to fsnotify_destroy_mark()Lino Sanfilippo2012-12-117-25/+28
| | | | | | | | | | | | | | | | | | | | | | | | In fsnotify_destroy_mark() dont get the group from the passed mark anymore, but pass the group itself as an additional parameter to the function. Signed-off-by: Lino Sanfilippo <LinoSanfilippo@gmx.de> Signed-off-by: Eric Paris <eparis@redhat.com>
| * | fsnotify: use a mutex instead of a spinlock to protect a groups mark listLino Sanfilippo2012-12-114-14/+14
| | | | | | | | | | | | | | | | | | | | | | | | | | | Replaces the groups mark_lock spinlock with a mutex. Using a mutex instead of a spinlock results in more flexibility (i.e it allows to sleep while the lock is held). Signed-off-by: Lino Sanfilippo <LinoSanfilippo@gmx.de> Signed-off-by: Eric Paris <eparis@redhat.com>
| * | fanotify: add an extra flag to mark_remove_from_mask that indicates wheather ↵Lino Sanfilippo2012-12-111-5/+14
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | a mark should be destroyed This patch adds an extra flag to mark_remove_from_mask() to inform the caller if the mark should be destroyed. With this we dont destroy the mark implicitly in the function itself any more but let the caller handle it. Signed-off-by: Lino Sanfilippo <LinoSanfilippo@gmx.de> Signed-off-by: Eric Paris <eparis@redhat.com>
| * | fsnotify: take groups mark_lock before mark lockLino Sanfilippo2012-12-111-10/+16
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Race-free addition and removal of a mark to a groups mark list would be easier if we could lock the mark list of group before we lock the specific mark. This patch changes the order used to add/remove marks to/from mark lists from 1. mark->lock 2. group->mark_lock 3. inode->i_lock to 1. group->mark_lock 2. mark->lock 3. inode->i_lock Signed-off-by: Lino Sanfilippo <LinoSanfilippo@gmx.de> Signed-off-by: Eric Paris <eparis@redhat.com>
OpenPOWER on IntegriCloud