From: Jeff Mahoney <jeffm@suse.com>

Subject: vfs: add super_operations->get_inode_dev

Patch-mainline: Never, upstream submissions have been met with a request to use per-subvolume superblocks, which is unworkable in practice

References: bsc#927455

There are many places where a dev_t:ino_t pair are passed to userspace

to uniquely describe an inode.  Some file systems, like btrfs, have

multiple inode namespace internally and use a separate dev_t to make the

distinction between them.

This patch adds a super_operations->get_inode_dev operation to allow

the file system to export those dev_ts to callers within the kernel

for further export to userspace.

Without this patch, things like audit and some perf and trace events

will not distinguish between subvolumes within a single btrfs filesystem.

Signed-off-by: Jeff Mahoney <jeffm@suse.com>

---

 fs/eventpoll.c                   |  2 +-

 fs/iomap/trace.h                 |  8 ++++----

 fs/locks.c                       | 12 +++++++-----

 fs/nfsd/nfs3xdr.c                |  2 +-

 fs/nfsd/nfs4state.c              |  5 +++--

 fs/nfsd/vfs.c                    |  4 ++--

 fs/notify/fdinfo.c               |  4 ++--

 fs/proc/nommu.c                  |  2 +-

 fs/proc/task_mmu.c               |  2 +-

 fs/proc/task_nommu.c             |  2 +-

 fs/proc_namespace.c              |  3 ++-

 fs/stat.c                        |  2 +-

 include/linux/fs.h               |  9 +++++++++

 include/trace/events/filelock.h  |  8 ++++----

 include/trace/events/filemap.h   |  6 +++---

 include/trace/events/writeback.h |  2 +-

 init/do_mounts.c                 |  2 +-

 kernel/audit_fsnotify.c          |  2 +-

 kernel/audit_watch.c             |  6 +++---

 kernel/auditsc.c                 |  6 +++---

 kernel/events/core.c             |  2 +-

 mm/memory-failure.c              |  2 +-

 net/unix/diag.c                  |  2 +-

 security/tomoyo/condition.c      |  2 +-

 security/tomoyo/realpath.c       |  1 +

 25 files changed, 56 insertions(+), 42 deletions(-)

--- a/fs/eventpoll.c

+++ b/fs/eventpoll.c

@@ -879,7 +879,7 @@ static void ep_show_fdinfo(struct seq_file *m, struct file *f)

 			   epi->ffd.fd, epi->event.events,

 			   (long long)epi->event.data,

 			   (long long)epi->ffd.file->f_pos,

-			   inode->i_ino, inode->i_sb->s_dev);

+			   inode->i_ino, inode_get_dev(inode));

 		if (seq_has_overflowed(m))

 			break;

 	}

--- a/fs/iomap/trace.h

+++ b/fs/iomap/trace.h

@@ -33,7 +33,7 @@ DECLARE_EVENT_CLASS(iomap_readpage_class,

 		__field(int, nr_pages)

 	),

 	TP_fast_assign(

-		__entry->dev = inode->i_sb->s_dev;

+		__entry->dev = inode_get_dev(inode);

 		__entry->ino = inode->i_ino;

 		__entry->nr_pages = nr_pages;

 	),

@@ -61,7 +61,7 @@ DECLARE_EVENT_CLASS(iomap_range_class,

 		__field(u64, length)

 	),

 	TP_fast_assign(

-		__entry->dev = inode->i_sb->s_dev;

+		__entry->dev = inode_get_dev(inode);

 		__entry->ino = inode->i_ino;

 		__entry->size = i_size_read(inode);

 		__entry->offset = off;

@@ -121,7 +121,7 @@ DECLARE_EVENT_CLASS(iomap_class,

 		__field(dev_t, bdev)

 	),

 	TP_fast_assign(

-		__entry->dev = inode->i_sb->s_dev;

+		__entry->dev = inode_get_dev(inode);

 		__entry->ino = inode->i_ino;

 		__entry->addr = iomap->addr;

 		__entry->offset = iomap->offset;

@@ -164,7 +164,7 @@ TRACE_EVENT(iomap_iter,

 		__field(unsigned long, caller)

 	),

 	TP_fast_assign(

-		__entry->dev = iter->inode->i_sb->s_dev;

+		__entry->dev = inode_get_dev(iter->inode);

 		__entry->ino = iter->inode->i_ino;

 		__entry->pos = iter->pos;

 		__entry->length = iomap_length(iter);

--- a/fs/locks.c

+++ b/fs/locks.c

@@ -215,12 +215,13 @@ static void

 locks_check_ctx_lists(struct inode *inode)

 {

 	struct file_lock_context *ctx = inode->i_flctx;

+	dev_t dev = inode_get_dev(inode);

 	if (unlikely(!list_empty(&ctx->flc_flock) ||

 		     !list_empty(&ctx->flc_posix) ||

 		     !list_empty(&ctx->flc_lease))) {

 		pr_warn("Leaked locks on dev=0x%x:0x%x ino=0x%lx:\n",

-			MAJOR(inode->i_sb->s_dev), MINOR(inode->i_sb->s_dev),

+			MAJOR(dev), MINOR(dev),

 			inode->i_ino);

 		locks_dump_ctx_list(&ctx->flc_flock, "FLOCK");

 		locks_dump_ctx_list(&ctx->flc_posix, "POSIX");

@@ -234,13 +235,14 @@ locks_check_ctx_file_list(struct file *filp, struct list_head *list,

 {

 	struct file_lock *fl;

 	struct inode *inode = locks_inode(filp);

+	dev_t dev = inode_get_dev(inode);

 	list_for_each_entry(fl, list, fl_list)

 		if (fl->fl_file == filp)

 			pr_warn("Leaked %s lock on dev=0x%x:0x%x ino=0x%lx "

 				" fl_owner=%p fl_flags=0x%x fl_type=0x%x fl_pid=%u\n",

-				list_type, MAJOR(inode->i_sb->s_dev),

-				MINOR(inode->i_sb->s_dev), inode->i_ino,

+				list_type, MAJOR(dev),

+				MINOR(dev), inode->i_ino,

 				fl->fl_owner, fl->fl_flags, fl->fl_type, fl->fl_pid);

 }

@@ -2761,8 +2763,8 @@ static void lock_get_status(struct seq_file *f, struct file_lock *fl,

 	if (inode) {

 		/* userspace relies on this representation of dev_t */

 		seq_printf(f, "%d %02x:%02x:%lu ", fl_pid,

-				MAJOR(inode->i_sb->s_dev),

-				MINOR(inode->i_sb->s_dev), inode->i_ino);

+				MAJOR(inode_get_dev(inode)),

+				MINOR(inode_get_dev(inode)), inode->i_ino);

 	} else {

 		seq_printf(f, "%d <none>:0 ", fl_pid);

 	}

--- a/fs/nfsd/nfs3xdr.c

+++ b/fs/nfsd/nfs3xdr.c

@@ -372,7 +372,7 @@ svcxdr_encode_fattr3(struct svc_rqst *rqstp, struct xdr_stream *xdr,

 		fsid ^= ((u64 *)fhp->fh_export->ex_uuid)[1];

 		break;

 	default:

-		fsid = (u64)huge_encode_dev(fhp->fh_dentry->d_sb->s_dev);

+		fsid = (u64)huge_encode_dev(inode_get_dev(d_inode(fhp->fh_dentry)));

 	}

 	p = xdr_encode_hyper(p, fsid);

--- a/fs/nfsd/nfs4state.c

+++ b/fs/nfsd/nfs4state.c

@@ -2595,10 +2595,11 @@ static void nfs4_show_fname(struct seq_file *s, struct nfsd_file *f)

 static void nfs4_show_superblock(struct seq_file *s, struct nfsd_file *f)

 {

 	struct inode *inode = file_inode(f->nf_file);

+	dev_t dev = inode_get_dev(inode);

 	seq_printf(s, "superblock: \"%02x:%02x:%ld\"",

-					MAJOR(inode->i_sb->s_dev),

-					 MINOR(inode->i_sb->s_dev),

+					MAJOR(dev),

+					 MINOR(dev),

 					 inode->i_ino);

 }

--- a/fs/nfsd/vfs.c

+++ b/fs/nfsd/vfs.c

@@ -1033,7 +1033,7 @@ static int wait_for_concurrent_writes(struct file *file)

 	int err = 0;

 	if (atomic_read(&inode->i_writecount) > 1

-	    || (last_ino == inode->i_ino && last_dev == inode->i_sb->s_dev)) {

+	    || (last_ino == inode->i_ino && last_dev == inode_get_dev(inode))) {

 		dprintk("nfsd: write defer %d\n", task_pid_nr(current));

 		msleep(10);

 		dprintk("nfsd: write resume %d\n", task_pid_nr(current));

@@ -1044,7 +1044,7 @@ static int wait_for_concurrent_writes(struct file *file)

 		err = vfs_fsync(file, 0);

 	}

 	last_ino = inode->i_ino;

-	last_dev = inode->i_sb->s_dev;

+	last_dev = inode_get_dev(inode);

 	return err;

 }

--- a/fs/notify/fdinfo.c

+++ b/fs/notify/fdinfo.c

@@ -85,7 +85,7 @@ static void inotify_fdinfo(struct seq_file *m, struct fsnotify_mark *mark)

 	inode = igrab(fsnotify_conn_inode(mark->connector));

 	if (inode) {

 		seq_printf(m, "inotify wd:%x ino:%lx sdev:%x mask:%x ignored_mask:0 ",

-			   inode_mark->wd, inode->i_ino, inode->i_sb->s_dev,

+			   inode_mark->wd, inode->i_ino, inode_get_dev(inode),

 			   inotify_mark_user_mask(mark));

 		show_mark_fhandle(m, inode);

 		seq_putc(m, '\n');

@@ -112,7 +112,7 @@ static void fanotify_fdinfo(struct seq_file *m, struct fsnotify_mark *mark)

 		if (!inode)

 			return;

 		seq_printf(m, "fanotify ino:%lx sdev:%x mflags:%x mask:%x ignored_mask:%x ",

-			   inode->i_ino, inode->i_sb->s_dev,

+			   inode->i_ino, inode_get_dev(inode),

 			   mflags, mark->mask, mark->ignore_mask);

 		show_mark_fhandle(m, inode);

 		seq_putc(m, '\n');

--- a/fs/proc/nommu.c

+++ b/fs/proc/nommu.c

@@ -40,7 +40,7 @@ static int nommu_region_show(struct seq_file *m, struct vm_region *region)

 	if (file) {

 		struct inode *inode = file_inode(region->vm_file);

-		dev = inode->i_sb->s_dev;

+		dev = inode_get_dev(inode);

 		ino = inode->i_ino;

 	}

--- a/fs/proc/task_mmu.c

+++ b/fs/proc/task_mmu.c

@@ -286,7 +286,7 @@ show_map_vma(struct seq_file *m, struct vm_area_struct *vma)

 	if (file) {

 		struct inode *inode = file_inode(vma->vm_file);

-		dev = inode->i_sb->s_dev;

+		dev = inode_get_dev(inode);

 		ino = inode->i_ino;

 		pgoff = ((loff_t)vma->vm_pgoff) << PAGE_SHIFT;

 	}

--- a/fs/proc/task_nommu.c

+++ b/fs/proc/task_nommu.c

@@ -151,7 +151,7 @@ static int nommu_vma_show(struct seq_file *m, struct vm_area_struct *vma)

 	if (file) {

 		struct inode *inode = file_inode(vma->vm_file);

-		dev = inode->i_sb->s_dev;

+		dev = inode_get_dev(inode);

 		ino = inode->i_ino;

 		pgoff = (loff_t)vma->vm_pgoff << PAGE_SHIFT;

 	}

--- a/fs/proc_namespace.c

+++ b/fs/proc_namespace.c

@@ -137,11 +137,12 @@ static int show_mountinfo(struct seq_file *m, struct vfsmount *mnt)

 	struct proc_mounts *p = m->private;

 	struct mount *r = real_mount(mnt);

 	struct super_block *sb = mnt->mnt_sb;

+	dev_t dev = inode_get_dev(d_inode(mnt->mnt_root));

 	struct path mnt_path = { .dentry = mnt->mnt_root, .mnt = mnt };

 	int err;

 	seq_printf(m, "%i %i %u:%u ", r->mnt_id, r->mnt_parent->mnt_id,

-		   MAJOR(sb->s_dev), MINOR(sb->s_dev));

+		   MAJOR(dev), MINOR(dev));

 	if (sb->s_op->show_path) {

 		err = sb->s_op->show_path(m, mnt->mnt_root);

 		if (err)

--- a/fs/stat.c

+++ b/fs/stat.c

@@ -47,7 +47,7 @@ void generic_fillattr(struct user_namespace *mnt_userns, struct inode *inode,

 	vfsuid_t vfsuid = i_uid_into_vfsuid(mnt_userns, inode);

 	vfsgid_t vfsgid = i_gid_into_vfsgid(mnt_userns, inode);

-	stat->dev = inode->i_sb->s_dev;

+	stat->dev = inode_get_dev(inode);

 	stat->ino = inode->i_ino;

 	stat->mode = inode->i_mode;

 	stat->nlink = inode->i_nlink;

--- a/include/linux/fs.h

+++ b/include/linux/fs.h

@@ -2250,6 +2250,7 @@ struct super_operations {

 				  struct shrink_control *);

 	long (*free_cached_objects)(struct super_block *,

 				    struct shrink_control *);

+	dev_t (*get_inode_dev)(const struct inode *);

 };

 /*

@@ -3610,4 +3611,12 @@ extern int vfs_fadvise(struct file *file, loff_t offset, loff_t len,

 extern int generic_fadvise(struct file *file, loff_t offset, loff_t len,

 			   int advice);

+static inline dev_t inode_get_dev(const struct inode *inode)

+{

+	if (inode->i_sb->s_op->get_inode_dev)

+		return inode->i_sb->s_op->get_inode_dev(inode);

+

+	return inode->i_sb->s_dev;

+}

+

 #endif /* _LINUX_FS_H */

--- a/include/trace/events/filelock.h

+++ b/include/trace/events/filelock.h

@@ -48,7 +48,7 @@ TRACE_EVENT(locks_get_lock_context,

 	),

 	TP_fast_assign(

-		__entry->s_dev = inode->i_sb->s_dev;

+		__entry->s_dev = inode_get_dev(inode);

 		__entry->i_ino = inode->i_ino;

 		__entry->type = type;

 		__entry->ctx = ctx;

@@ -80,7 +80,7 @@ DECLARE_EVENT_CLASS(filelock_lock,

 	TP_fast_assign(

 		__entry->fl = fl ? fl : NULL;

-		__entry->s_dev = inode->i_sb->s_dev;

+		__entry->s_dev = inode_get_dev(inode);

 		__entry->i_ino = inode->i_ino;

 		__entry->fl_blocker = fl ? fl->fl_blocker : NULL;

 		__entry->fl_owner = fl ? fl->fl_owner : NULL;

@@ -135,7 +135,7 @@ DECLARE_EVENT_CLASS(filelock_lease,

 	TP_fast_assign(

 		__entry->fl = fl ? fl : NULL;

-		__entry->s_dev = inode->i_sb->s_dev;

+		__entry->s_dev = inode_get_dev(inode);

 		__entry->i_ino = inode->i_ino;

 		__entry->fl_blocker = fl ? fl->fl_blocker : NULL;

 		__entry->fl_owner = fl ? fl->fl_owner : NULL;

@@ -185,7 +185,7 @@ TRACE_EVENT(generic_add_lease,

 	),

 	TP_fast_assign(

-		__entry->s_dev = inode->i_sb->s_dev;

+		__entry->s_dev = inode_get_dev(inode);

 		__entry->i_ino = inode->i_ino;

 		__entry->wcount = atomic_read(&inode->i_writecount);

 		__entry->rcount = atomic_read(&inode->i_readcount);

--- a/include/trace/events/filemap.h

+++ b/include/trace/events/filemap.h

@@ -32,7 +32,7 @@ DECLARE_EVENT_CLASS(mm_filemap_op_page_cache,

 		__entry->i_ino = folio->mapping->host->i_ino;

 		__entry->index = folio->index;

 		if (folio->mapping->host->i_sb)

-			__entry->s_dev = folio->mapping->host->i_sb->s_dev;

+			__entry->s_dev = inode_get_dev(folio->mapping->host);

 		else

 			__entry->s_dev = folio->mapping->host->i_rdev;

 		__entry->order = folio_order(folio);

@@ -71,7 +71,7 @@ TRACE_EVENT(filemap_set_wb_err,

 			__entry->i_ino = mapping->host->i_ino;

 			__entry->errseq = eseq;

 			if (mapping->host->i_sb)

-				__entry->s_dev = mapping->host->i_sb->s_dev;

+				__entry->s_dev = inode_get_dev(mapping->host);

 			else

 				__entry->s_dev = mapping->host->i_rdev;

 		),

@@ -99,7 +99,7 @@ TRACE_EVENT(file_check_and_advance_wb_err,

 			__entry->i_ino = file->f_mapping->host->i_ino;

 			if (file->f_mapping->host->i_sb)

 				__entry->s_dev =

-					file->f_mapping->host->i_sb->s_dev;

+					inode_get_dev(file->f_mapping->host);

 			else

 				__entry->s_dev =

 					file->f_mapping->host->i_rdev;

--- a/include/trace/events/writeback.h

+++ b/include/trace/events/writeback.h

@@ -809,7 +809,7 @@ DECLARE_EVENT_CLASS(writeback_inode_template,

 	),

 	TP_fast_assign(

-		__entry->dev	= inode->i_sb->s_dev;

+		__entry->dev	= inode_get_dev(inode);

 		__entry->ino	= inode->i_ino;

 		__entry->state	= inode->i_state;

 		__entry->mode	= inode->i_mode;

--- a/init/do_mounts.c

+++ b/init/do_mounts.c

@@ -378,7 +378,7 @@ static int __init do_mount_root(const char *name, const char *fs,

 	init_chdir("/root");

 	s = current->fs->pwd.dentry->d_sb;

-	ROOT_DEV = s->s_dev;

+	ROOT_DEV = inode_get_dev(d_inode(current->fs->pwd.dentry));

 	printk(KERN_INFO

 	       "VFS: Mounted root (%s filesystem)%s on device %u:%u.\n",

 	       s->s_type->name,

--- a/kernel/audit_fsnotify.c

+++ b/kernel/audit_fsnotify.c

@@ -67,7 +67,7 @@ int audit_mark_compare(struct audit_fsnotify_mark *mark, unsigned long ino, dev_

 static void audit_update_mark(struct audit_fsnotify_mark *audit_mark,

 			     const struct inode *inode)

 {

-	audit_mark->dev = inode ? inode->i_sb->s_dev : AUDIT_DEV_UNSET;

+	audit_mark->dev = inode ? inode_get_dev(inode) : AUDIT_DEV_UNSET;

 	audit_mark->ino = inode ? inode->i_ino : AUDIT_INO_UNSET;

 }

--- a/kernel/audit_watch.c

+++ b/kernel/audit_watch.c

@@ -352,7 +352,7 @@ static int audit_get_nd(struct audit_watch *watch, struct path *parent)

 		return PTR_ERR(d);

 	if (d_is_positive(d)) {

 		/* update watch filter fields */

-		watch->dev = d->d_sb->s_dev;

+		watch->dev = inode_get_dev(d_backing_inode(d));

 		watch->ino = d_backing_inode(d)->i_ino;

 	}

 	inode_unlock(d_backing_inode(parent->dentry));

@@ -477,7 +477,7 @@ static int audit_watch_handle_event(struct fsnotify_mark *inode_mark, u32 mask,

 		return 0;

 	if (mask & (FS_CREATE|FS_MOVED_TO) && inode)

-		audit_update_watch(parent, dname, inode->i_sb->s_dev, inode->i_ino, 0);

+		audit_update_watch(parent, dname, inode_get_dev(inode), inode->i_ino, 0);

 	else if (mask & (FS_DELETE|FS_MOVED_FROM))

 		audit_update_watch(parent, dname, AUDIT_DEV_UNSET, AUDIT_INO_UNSET, 1);

 	else if (mask & (FS_DELETE_SELF|FS_UNMOUNT|FS_MOVE_SELF))

@@ -531,7 +531,7 @@ int audit_exe_compare(struct task_struct *tsk, struct audit_fsnotify_mark *mark)

 	if (!exe_file)

 		return 0;

 	ino = file_inode(exe_file)->i_ino;

-	dev = file_inode(exe_file)->i_sb->s_dev;

+	dev = inode_get_dev(file_inode(exe_file));

 	fput(exe_file);

 	return audit_mark_compare(mark, ino, dev);

 }

--- a/kernel/auditsc.c

+++ b/kernel/auditsc.c

@@ -2272,7 +2272,7 @@ static void audit_copy_inode(struct audit_names *name,

 			     struct inode *inode, unsigned int flags)

 {

 	name->ino   = inode->i_ino;

-	name->dev   = inode->i_sb->s_dev;

+	name->dev   = inode_get_dev(inode);

 	name->mode  = inode->i_mode;

 	name->uid   = inode->i_uid;

 	name->gid   = inode->i_gid;

@@ -2344,7 +2344,7 @@ void __audit_inode(struct filename *name, const struct dentry *dentry,

 		if (n->ino) {

 			/* valid inode number, use that for the comparison */

 			if (n->ino != inode->i_ino ||

-			    n->dev != inode->i_sb->s_dev)

+			    n->dev != inode_get_dev(inode))

 				continue;

 		} else if (n->name) {

 			/* inode number has not been set, check the name */

@@ -2449,7 +2449,7 @@ void __audit_inode_child(struct inode *parent,

 		     n->type != AUDIT_TYPE_UNKNOWN))

 			continue;

-		if (n->ino == parent->i_ino && n->dev == parent->i_sb->s_dev &&

+		if (n->ino == parent->i_ino && n->dev == inode_get_dev(parent) &&

 		    !audit_compare_dname_path(dname,

 					      n->name->name, n->name_len)) {

 			if (n->type == AUDIT_TYPE_UNKNOWN)

--- a/kernel/events/core.c

+++ b/kernel/events/core.c

@@ -8654,7 +8654,7 @@ static void perf_event_mmap_event(struct perf_mmap_event *mmap_event)

 			goto cpy_name;

 		}

 		inode = file_inode(vma->vm_file);

-		dev = inode->i_sb->s_dev;

+		dev = inode_get_dev(inode);

 		ino = inode->i_ino;

 		gen = inode->i_generation;

 		maj = MAJOR(dev);

--- a/mm/memory-failure.c

+++ b/mm/memory-failure.c

@@ -159,7 +159,7 @@ static int hwpoison_filter_dev(struct page *p)

 	if (mapping == NULL || mapping->host == NULL)

 		return -EINVAL;

-	dev = mapping->host->i_sb->s_dev;

+	dev = inode_get_dev(mapping->host);

 	if (hwpoison_filter_dev_major != ~0U &&

 	    hwpoison_filter_dev_major != MAJOR(dev))

 		return -EINVAL;

--- a/net/unix/diag.c

+++ b/net/unix/diag.c

@@ -31,7 +31,7 @@ static int sk_diag_dump_vfs(struct sock *sk, struct sk_buff *nlskb)

 	if (dentry) {

 		struct unix_diag_vfs uv = {

 			.udiag_vfs_ino = d_backing_inode(dentry)->i_ino,

-			.udiag_vfs_dev = dentry->d_sb->s_dev,

+			.udiag_vfs_dev = inode_get_dev(d_backing_inode(dentry)),

 		};

 		return nla_put(nlskb, UNIX_DIAG_VFS, sizeof(uv), &uv;;

--- a/security/tomoyo/condition.c

+++ b/security/tomoyo/condition.c

@@ -743,7 +743,7 @@ void tomoyo_get_attributes(struct tomoyo_obj_info *obj)

 			stat->gid  = inode->i_gid;

 			stat->ino  = inode->i_ino;

 			stat->mode = inode->i_mode;

-			stat->dev  = inode->i_sb->s_dev;

+			stat->dev  = inode_get_dev(inode);

 			stat->rdev = inode->i_rdev;

 			obj->stat_valid[i] = true;

 		}

--- a/security/tomoyo/realpath.c

+++ b/security/tomoyo/realpath.c

@@ -174,6 +174,7 @@ static char *tomoyo_get_local_path(struct dentry *dentry, char * const buffer,

 		goto prepend_filesystem_name;

 	}

 	/* Use filesystem name for unnamed devices. */

+	/* NOTE: This will fail with the btrfs get_inode_dev patches */

 	if (!MAJOR(sb->s_dev))

 		goto prepend_filesystem_name;

 	{