Commit be655596 authored by Sage Weil's avatar Sage Weil
Browse files

ceph: use i_ceph_lock instead of i_lock 



We have been using i_lock to protect all kinds of data structures in the
ceph_inode_info struct, including lists of inodes that we need to iterate
over while avoiding races with inode destruction.  That requires grabbing
a reference to the inode with the list lock protected, but igrab() now
takes i_lock to check the inode flags.

Changing the list lock ordering would be a painful process.

However, using a ceph-specific i_ceph_lock in the ceph inode instead of
i_lock is a simple mechanical change and avoids the ordering constraints
imposed by igrab().

Reported-by: default avatarAmon Ott <a.ott@m-privacy.de>
Signed-off-by: default avatarSage Weil <sage@newdream.net>
parent 51703306

fs/ceph/addr.c

+4 −4
Original line number Diff line number Diff line
@@ -87,7 +87,7 @@ static int ceph_set_page_dirty(struct page *page) 
	snapc = ceph_get_snap_context(ci->i_snap_realm->cached_context);



	/* dirty the head */

	spin_lock(&inode->i_lock);

	spin_lock(&ci->i_ceph_lock);

	if (ci->i_head_snapc == NULL)

		ci->i_head_snapc = ceph_get_snap_context(snapc);

	++ci->i_wrbuffer_ref_head;
@@ -100,7 +100,7 @@ static int ceph_set_page_dirty(struct page *page) 
	     ci->i_wrbuffer_ref-1, ci->i_wrbuffer_ref_head-1,

	     ci->i_wrbuffer_ref, ci->i_wrbuffer_ref_head,

	     snapc, snapc->seq, snapc->num_snaps);

	spin_unlock(&inode->i_lock);

	spin_unlock(&ci->i_ceph_lock);



	/* now adjust page */

	spin_lock_irq(&mapping->tree_lock);
@@ -391,7 +391,7 @@ static struct ceph_snap_context *get_oldest_context(struct inode *inode, 
	struct ceph_snap_context *snapc = NULL;

	struct ceph_cap_snap *capsnap = NULL;



	spin_lock(&inode->i_lock);

	spin_lock(&ci->i_ceph_lock);

	list_for_each_entry(capsnap, &ci->i_cap_snaps, ci_item) {

		dout(" cap_snap %p snapc %p has %d dirty pages\n", capsnap,

		     capsnap->context, capsnap->dirty_pages);
@@ -407,7 +407,7 @@ static struct ceph_snap_context *get_oldest_context(struct inode *inode, 
		dout(" head snapc %p has %d dirty pages\n",

		     snapc, ci->i_wrbuffer_ref_head);

	}

	spin_unlock(&inode->i_lock);

	spin_unlock(&ci->i_ceph_lock);

	return snapc;

}

fs/ceph/caps.c

+94 −93

File changed.

Preview size limit exceeded, changes collapsed.

fs/ceph/dir.c

+12 −12
Original line number Diff line number Diff line
@@ -281,18 +281,18 @@ static int ceph_readdir(struct file *filp, void *dirent, filldir_t filldir) 
	}



	/* can we use the dcache? */

	spin_lock(&inode->i_lock);

	spin_lock(&ci->i_ceph_lock);

	if ((filp->f_pos == 2 || fi->dentry) &&

	    !ceph_test_mount_opt(fsc, NOASYNCREADDIR) &&

	    ceph_snap(inode) != CEPH_SNAPDIR &&

	    ceph_dir_test_complete(inode) &&

	    __ceph_caps_issued_mask(ci, CEPH_CAP_FILE_SHARED, 1)) {

		spin_unlock(&inode->i_lock);

		spin_unlock(&ci->i_ceph_lock);

		err = __dcache_readdir(filp, dirent, filldir);

		if (err != -EAGAIN)

			return err;

	} else {

		spin_unlock(&inode->i_lock);

		spin_unlock(&ci->i_ceph_lock);

	}

	if (fi->dentry) {

		err = note_last_dentry(fi, fi->dentry->d_name.name,
@@ -428,12 +428,12 @@ more: 
	 * were released during the whole readdir, and we should have

	 * the complete dir contents in our cache.

	 */

	spin_lock(&inode->i_lock);

	spin_lock(&ci->i_ceph_lock);

	if (ci->i_release_count == fi->dir_release_count) {

		ceph_dir_set_complete(inode);

		ci->i_max_offset = filp->f_pos;

	}

	spin_unlock(&inode->i_lock);

	spin_unlock(&ci->i_ceph_lock);



	dout("readdir %p filp %p done.\n", inode, filp);

	return 0;
@@ -607,7 +607,7 @@ static struct dentry *ceph_lookup(struct inode *dir, struct dentry *dentry, 
		struct ceph_inode_info *ci = ceph_inode(dir);

		struct ceph_dentry_info *di = ceph_dentry(dentry);



		spin_lock(&dir->i_lock);

		spin_lock(&ci->i_ceph_lock);

		dout(" dir %p flags are %d\n", dir, ci->i_ceph_flags);

		if (strncmp(dentry->d_name.name,

			    fsc->mount_options->snapdir_name,
@@ -615,13 +615,13 @@ static struct dentry *ceph_lookup(struct inode *dir, struct dentry *dentry, 
		    !is_root_ceph_dentry(dir, dentry) &&

		    ceph_dir_test_complete(dir) &&

		    (__ceph_caps_issued_mask(ci, CEPH_CAP_FILE_SHARED, 1))) {

			spin_unlock(&dir->i_lock);

			spin_unlock(&ci->i_ceph_lock);

			dout(" dir %p complete, -ENOENT\n", dir);

			d_add(dentry, NULL);

			di->lease_shared_gen = ci->i_shared_gen;

			return NULL;

		}

		spin_unlock(&dir->i_lock);

		spin_unlock(&ci->i_ceph_lock);

	}



	op = ceph_snap(dir) == CEPH_SNAPDIR ?
@@ -841,12 +841,12 @@ static int drop_caps_for_unlink(struct inode *inode) 
	struct ceph_inode_info *ci = ceph_inode(inode);

	int drop = CEPH_CAP_LINK_SHARED | CEPH_CAP_LINK_EXCL;



	spin_lock(&inode->i_lock);

	spin_lock(&ci->i_ceph_lock);

	if (inode->i_nlink == 1) {

		drop |= ~(__ceph_caps_wanted(ci) | CEPH_CAP_PIN);

		ci->i_ceph_flags |= CEPH_I_NODELAY;

	}

	spin_unlock(&inode->i_lock);

	spin_unlock(&ci->i_ceph_lock);

	return drop;

}
@@ -1015,10 +1015,10 @@ static int dir_lease_is_valid(struct inode *dir, struct dentry *dentry) 
	struct ceph_dentry_info *di = ceph_dentry(dentry);

	int valid = 0;



	spin_lock(&dir->i_lock);

	spin_lock(&ci->i_ceph_lock);

	if (ci->i_shared_gen == di->lease_shared_gen)

		valid = __ceph_caps_issued_mask(ci, CEPH_CAP_FILE_SHARED, 1);

	spin_unlock(&dir->i_lock);

	spin_unlock(&ci->i_ceph_lock);

	dout("dir_lease_is_valid dir %p v%u dentry %p v%u = %d\n",

	     dir, (unsigned)ci->i_shared_gen, dentry,

	     (unsigned)di->lease_shared_gen, valid);

fs/ceph/file.c

+10 −10
Original line number Diff line number Diff line
@@ -147,9 +147,9 @@ int ceph_open(struct inode *inode, struct file *file) 


	/* trivially open snapdir */

	if (ceph_snap(inode) == CEPH_SNAPDIR) {

		spin_lock(&inode->i_lock);

		spin_lock(&ci->i_ceph_lock);

		__ceph_get_fmode(ci, fmode);

		spin_unlock(&inode->i_lock);

		spin_unlock(&ci->i_ceph_lock);

		return ceph_init_file(inode, file, fmode);

	}
@@ -158,7 +158,7 @@ int ceph_open(struct inode *inode, struct file *file) 
	 * write) or any MDS (for read).  Update wanted set

	 * asynchronously.

	 */

	spin_lock(&inode->i_lock);

	spin_lock(&ci->i_ceph_lock);

	if (__ceph_is_any_real_caps(ci) &&

	    (((fmode & CEPH_FILE_MODE_WR) == 0) || ci->i_auth_cap)) {

		int mds_wanted = __ceph_caps_mds_wanted(ci);
@@ -168,7 +168,7 @@ int ceph_open(struct inode *inode, struct file *file) 
		     inode, fmode, ceph_cap_string(wanted),

		     ceph_cap_string(issued));

		__ceph_get_fmode(ci, fmode);

		spin_unlock(&inode->i_lock);

		spin_unlock(&ci->i_ceph_lock);



		/* adjust wanted? */

		if ((issued & wanted) != wanted &&
@@ -180,10 +180,10 @@ int ceph_open(struct inode *inode, struct file *file) 
	} else if (ceph_snap(inode) != CEPH_NOSNAP &&

		   (ci->i_snap_caps & wanted) == wanted) {

		__ceph_get_fmode(ci, fmode);

		spin_unlock(&inode->i_lock);

		spin_unlock(&ci->i_ceph_lock);

		return ceph_init_file(inode, file, fmode);

	}

	spin_unlock(&inode->i_lock);

	spin_unlock(&ci->i_ceph_lock);



	dout("open fmode %d wants %s\n", fmode, ceph_cap_string(wanted));

	req = prepare_open_request(inode->i_sb, flags, 0);
@@ -743,9 +743,9 @@ retry_snap: 
		 */

		int dirty;



		spin_lock(&inode->i_lock);

		spin_lock(&ci->i_ceph_lock);

		dirty = __ceph_mark_dirty_caps(ci, CEPH_CAP_FILE_WR);

		spin_unlock(&inode->i_lock);

		spin_unlock(&ci->i_ceph_lock);

		ceph_put_cap_refs(ci, got);



		ret = generic_file_aio_write(iocb, iov, nr_segs, pos);
@@ -764,9 +764,9 @@ retry_snap: 


	if (ret >= 0) {

		int dirty;

		spin_lock(&inode->i_lock);

		spin_lock(&ci->i_ceph_lock);

		dirty = __ceph_mark_dirty_caps(ci, CEPH_CAP_FILE_WR);

		spin_unlock(&inode->i_lock);

		spin_unlock(&ci->i_ceph_lock);

		if (dirty)

			__mark_inode_dirty(inode, dirty);

	}

fs/ceph/inode.c

+28 −25

File changed.

Preview size limit exceeded, changes collapsed.

CRA Git | Maintained and supported by SUSTech CRA and CCSE