Skip to contentSkip to conversation

vfs: don't evict uninitialized inode

PATCH Superseded Miklos Szeredi 4 participants 7 messages Jul 6, 2018 – Jul 20, 2018 VFS

v2 of 3 Superseded Latest is v3 →

lore
Share & tools
lore
#0 lore raw

PATCH v2 1/1 vfs: don't evict uninitialized inode

iput() ends up calling ->evict() on new inode, which is not yet initialized by owning fs. So use destroy_inode() instead.

Add to sb->s_inodes list only if inode is not in I_CREATING state (meaning that it wasn't allocated with new_inode(), which already does the insertion).

Fixes 80ea09a002bf ("vfs: factor out inode_insert5()")
---
 fs/inode.c | 7 +++++--
 1 file changed, 5 insertions(+), 2 deletions(-)
fs/inode.c +5 -2
index 04dd7e0d5142..e7de38c1d9d8 100644
@@ -1050,6 +1050,7 @@ struct inode *inode_insert5(struct inode *inode, unsigned long hashval,
 {
 	struct hlist_head *head = inode_hashtable + hash(inode->i_sb, hashval);
 	struct inode *old;
+	bool creating = inode->i_state & I_CREATING;
 
 again:
 	spin_lock(&inode_hash_lock);
@@ -1083,6 +1084,8 @@ struct inode *inode_insert5(struct inode *inode, unsigned long hashval,
 	inode->i_state |= I_NEW;
 	hlist_add_head(&inode->i_hash, head);
 	spin_unlock(&inode->i_lock);
+	if (!creating)
+		inode_sb_list_add(inode);
 unlock:
 	spin_unlock(&inode_hash_lock);
 
@@ -1117,12 +1120,12 @@ struct inode *iget5_locked(struct super_block *sb, unsigned long hashval,
 	struct inode *inode = ilookup5(sb, hashval, test, data);
 
 	if (!inode) {
-		struct inode *new = new_inode(sb);
+		struct inode *new = alloc_inode(sb);
 
 		if (new) {
 			inode = inode_insert5(new, hashval, test, set, data);
 			if (unlikely(inode != new))
-				iput(new);
+				destroy_inode(new);
 		}
 	}
 	return inode;
Signature · 1 line
2.14.3
#1 lore raw
Quoted 11 lines · Miklos Szeredi

iput() ends up calling ->evict() on new inode, which is not yet initialized by owning fs. So use destroy_inode() instead.

Add to sb->s_inodes list only if inode is not in I_CREATING state (meaning that it wasn't allocated with new_inode(), which already does the insertion).

Fixes 80ea09a002bf ("vfs: factor out inode_insert5()")

---

Al,

Did you loose track of this 4.18 regression fix?

Thanks,
Amir.

Quoted 42 lines
 fs/inode.c | 7 +++++--
 1 file changed, 5 insertions(+), 2 deletions(-)
fs/inode.c +5 -3
index 04dd7e0d5142..e7de38c1d9d8 100644
@@ -1050,6 +1050,7 @@ struct inode *inode_insert5(struct inode *inode, unsigned long hashval,
 {
        struct hlist_head *head = inode_hashtable + hash(inode->i_sb, hashval);
        struct inode *old;
+       bool creating = inode->i_state & I_CREATING;
 
 again:
        spin_lock(&inode_hash_lock);
@@ -1083,6 +1084,8 @@ struct inode *inode_insert5(struct inode *inode, unsigned long hashval,
        inode->i_state |= I_NEW;
        hlist_add_head(&inode->i_hash, head);
        spin_unlock(&inode->i_lock);
+       if (!creating)
+               inode_sb_list_add(inode);
 unlock:
        spin_unlock(&inode_hash_lock);
 
@@ -1117,12 +1120,12 @@ struct inode *iget5_locked(struct super_block *sb, unsigned long hashval,
        struct inode *inode = ilookup5(sb, hashval, test, data);
 
        if (!inode) {
-               struct inode *new = new_inode(sb);
+               struct inode *new = alloc_inode(sb);
 
                if (new) {
                        inode = inode_insert5(new, hashval, test, set, data);
                        if (unlikely(inode != new))
-                               iput(new);
+                               destroy_inode(new);
                }
        }
        return inode;
--

2.14.3

#2 lore raw
In reply to Miklos Szeredi

Ping?

This (or some variant) needs to go into 4.18.

Thanks,
Miklos

Quoted 53 lines · Miklos Szeredi

iput() ends up calling ->evict() on new inode, which is not yet initialized by owning fs. So use destroy_inode() instead.

Add to sb->s_inodes list only if inode is not in I_CREATING state (meaning that it wasn't allocated with new_inode(), which already does the insertion).

Fixes 80ea09a002bf ("vfs: factor out inode_insert5()")
---
 fs/inode.c | 7 +++++--
 1 file changed, 5 insertions(+), 2 deletions(-)
fs/inode.c +5 -3
index 04dd7e0d5142..e7de38c1d9d8 100644
@@ -1050,6 +1050,7 @@ struct inode *inode_insert5(struct inode *inode, unsigned long hashval,
 {
        struct hlist_head *head = inode_hashtable + hash(inode->i_sb, hashval);
        struct inode *old;
+       bool creating = inode->i_state & I_CREATING;
 
 again:
        spin_lock(&inode_hash_lock);
@@ -1083,6 +1084,8 @@ struct inode *inode_insert5(struct inode *inode, unsigned long hashval,
        inode->i_state |= I_NEW;
        hlist_add_head(&inode->i_hash, head);
        spin_unlock(&inode->i_lock);
+       if (!creating)
+               inode_sb_list_add(inode);
 unlock:
        spin_unlock(&inode_hash_lock);
 
@@ -1117,12 +1120,12 @@ struct inode *iget5_locked(struct super_block *sb, unsigned long hashval,
        struct inode *inode = ilookup5(sb, hashval, test, data);
 
        if (!inode) {
-               struct inode *new = new_inode(sb);
+               struct inode *new = alloc_inode(sb);
 
                if (new) {
                        inode = inode_insert5(new, hashval, test, set, data);
                        if (unlikely(inode != new))
-                               iput(new);
+                               destroy_inode(new);
                }
        }
        return inode;
--

2.14.3

maintainer
#3 lore raw
Quoted 3 lines · Miklos Szeredi

Ping?

This (or some variant) needs to go into 4.18.

Yes - icache series is on the top of my list, now that -open stuff is (hopefully) sorted out (in for-next, anyway).

I'm putting together a queue with that stuff at the moment.

BTW, why have you left generic_readlink() sitting around? AFAICS, it could've been folded into the only remaining caller just as you've made it static in late 2016... I'll fold it in; just curious what was the reason for not doing that back then...

#4 lore raw
Quoted 4 lines · Al Viro

BTW, why have you left generic_readlink() sitting around? AFAICS, it could've been folded into the only remaining caller just as you've made it static in late 2016... I'll fold it in; just curious what was the reason for not doing that back then...

I don't remember any particular reason.

Thanks,
Miklos

maintainer
#5 lore raw
In reply to Al Viro
Quoted 4 lines · Al Viro

BTW, why have you left generic_readlink() sitting around? AFAICS, it could've been folded into the only remaining caller just as you've made it static in late 2016... I'll fold it in; just curious what was the reason for not doing that back then...

BTW^2:
const char *vfs_get_link(struct dentry *dentry, struct delayed_call *done)

{
        const char *res = ERR_PTR(-EINVAL);
        struct inode *inode = d_inode(dentry);

        if (d_is_symlink(dentry)) {
                res = ERR_PTR(security_inode_readlink(dentry));
                if (!res)
                        res = inode->i_op->get_link(dentry, inode, done);
        }
        return res;
}

hits a method call that is not needed in the majority of cases. Is there any subtle reason why it shouldn't be

const char *vfs_get_link(struct dentry *dentry, struct delayed_call *done)

{
        const char *res = ERR_PTR(-EINVAL);
        struct inode *inode = d_inode(dentry);

        if (d_is_symlink(dentry)) {
                res = ERR_PTR(security_inode_readlink(dentry));
                if (!res)
			res = inode->i_link;
		if (!res)
                        res = inode->i_op->get_link(dentry, inode, done);
        }
        return res;
}

instead?

#6 lore raw
Quoted 37 lines · Al Viro
Quoted 4 lines · Al Viro

BTW, why have you left generic_readlink() sitting around? AFAICS, it could've been folded into the only remaining caller just as you've made it static in late 2016... I'll fold it in; just curious what was the reason for not doing that back then...

BTW^2:
const char *vfs_get_link(struct dentry *dentry, struct delayed_call *done)

{
        const char *res = ERR_PTR(-EINVAL);
        struct inode *inode = d_inode(dentry);

        if (d_is_symlink(dentry)) {
                res = ERR_PTR(security_inode_readlink(dentry));
                if (!res)
                        res = inode->i_op->get_link(dentry, inode, done);
        }
        return res;
}

hits a method call that is not needed in the majority of cases. Is there any subtle reason why it shouldn't be

const char *vfs_get_link(struct dentry *dentry, struct delayed_call *done)

{
        const char *res = ERR_PTR(-EINVAL);
        struct inode *inode = d_inode(dentry);

        if (d_is_symlink(dentry)) {
                res = ERR_PTR(security_inode_readlink(dentry));
                if (!res)
                        res = inode->i_link;
                if (!res)
                        res = inode->i_op->get_link(dentry, inode, done);
        }
        return res;
}
Quoted 1 line

instead?

Can't see any issues. But I also don't think any of the callers are seriously performance sensitive, so I guess it basically doesn't matter.

Thanks,
Miklos