The patch below does not apply to the 5.15-stable tree. If someone wants it applied there, or to any other stable or longterm tree, then please email the backport, including the original git commit id to <stable@xxxxxxxxxxxxxxx>. To reproduce the conflict and resubmit, you may use the following commands: git fetch https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/ linux-5.15.y git checkout FETCH_HEAD git cherry-pick -x c4d361f66ac91db8fc65061a9671682f61f4ca9d # <resolve conflicts, build, test, etc.> git commit -s git send-email --to '<stable@xxxxxxxxxxxxxxx>' --in-reply-to '2023121709-gout-brick-e85e@gregkh' --subject-prefix 'PATCH 5.15.y' HEAD^.. Possible dependencies: c4d361f66ac9 ("fuse: share lookup state between submount and its parent") 972f4c46d0a1 ("fuse: cache btime") d3045530bdd2 ("fuse: implement statx") 9dc10a54abe5 ("fuse: add ATTR_TIMEOUT macro") 8d8f9c4b8df6 ("fuse: handle empty request_mask in statx") e78662e818f9 ("fuse: add a new fuse init flag to relax restrictions in no cache mode") 5cadfbd5a11e ("fuse: add feature flag for expire-only") d40b2f4c94f2 ("Merge tag 'fuse-update-6.3' of git://git.kernel.org/pub/scm/linux/kernel/git/mszeredi/fuse") thanks, greg k-h ------------------ original commit in Linus's tree ------------------ >From c4d361f66ac91db8fc65061a9671682f61f4ca9d Mon Sep 17 00:00:00 2001 From: Krister Johansen <kjlx@xxxxxxxxxxxxxxxxxx> Date: Fri, 3 Nov 2023 10:39:47 -0700 Subject: [PATCH] fuse: share lookup state between submount and its parent Fuse submounts do not perform a lookup for the nodeid that they inherit from their parent. Instead, the code decrements the nlookup on the submount's fuse_inode when it is instantiated, and no forget is performed when a submount root is evicted. Trouble arises when the submount's parent is evicted despite the submount itself being in use. In this author's case, the submount was in a container and deatched from the initial mount namespace via a MNT_DEATCH operation. When memory pressure triggered the shrinker, the inode from the parent was evicted, which triggered enough forgets to render the submount's nodeid invalid. Since submounts should still function, even if their parent goes away, solve this problem by sharing refcounted state between the parent and its submount. When all of the references on this shared state reach zero, it's safe to forget the final lookup of the fuse nodeid. Signed-off-by: Krister Johansen <kjlx@xxxxxxxxxxxxxxxxxx> Cc: stable@xxxxxxxxxxxxxxx Fixes: 1866d779d5d2 ("fuse: Allow fuse_fill_super_common() for submounts") Signed-off-by: Miklos Szeredi <mszeredi@xxxxxxxxxx> diff --git a/fs/fuse/fuse_i.h b/fs/fuse/fuse_i.h index 69bcffaf4832..1df83eebda92 100644 --- a/fs/fuse/fuse_i.h +++ b/fs/fuse/fuse_i.h @@ -63,6 +63,19 @@ struct fuse_forget_link { struct fuse_forget_link *next; }; +/* Submount lookup tracking */ +struct fuse_submount_lookup { + /** Refcount */ + refcount_t count; + + /** Unique ID, which identifies the inode between userspace + * and kernel */ + u64 nodeid; + + /** The request used for sending the FORGET message */ + struct fuse_forget_link *forget; +}; + /** FUSE inode */ struct fuse_inode { /** Inode data */ @@ -158,6 +171,8 @@ struct fuse_inode { */ struct fuse_inode_dax *dax; #endif + /** Submount specific lookup tracking */ + struct fuse_submount_lookup *submount_lookup; }; /** FUSE inode state bits */ diff --git a/fs/fuse/inode.c b/fs/fuse/inode.c index 88090c6026a7..2a6d44f91729 100644 --- a/fs/fuse/inode.c +++ b/fs/fuse/inode.c @@ -68,6 +68,24 @@ struct fuse_forget_link *fuse_alloc_forget(void) return kzalloc(sizeof(struct fuse_forget_link), GFP_KERNEL_ACCOUNT); } +static struct fuse_submount_lookup *fuse_alloc_submount_lookup(void) +{ + struct fuse_submount_lookup *sl; + + sl = kzalloc(sizeof(struct fuse_submount_lookup), GFP_KERNEL_ACCOUNT); + if (!sl) + return NULL; + sl->forget = fuse_alloc_forget(); + if (!sl->forget) + goto out_free; + + return sl; + +out_free: + kfree(sl); + return NULL; +} + static struct inode *fuse_alloc_inode(struct super_block *sb) { struct fuse_inode *fi; @@ -83,6 +101,7 @@ static struct inode *fuse_alloc_inode(struct super_block *sb) fi->attr_version = 0; fi->orig_ino = 0; fi->state = 0; + fi->submount_lookup = NULL; mutex_init(&fi->mutex); spin_lock_init(&fi->lock); fi->forget = fuse_alloc_forget(); @@ -113,6 +132,17 @@ static void fuse_free_inode(struct inode *inode) kmem_cache_free(fuse_inode_cachep, fi); } +static void fuse_cleanup_submount_lookup(struct fuse_conn *fc, + struct fuse_submount_lookup *sl) +{ + if (!refcount_dec_and_test(&sl->count)) + return; + + fuse_queue_forget(fc, sl->forget, sl->nodeid, 1); + sl->forget = NULL; + kfree(sl); +} + static void fuse_evict_inode(struct inode *inode) { struct fuse_inode *fi = get_fuse_inode(inode); @@ -132,6 +162,11 @@ static void fuse_evict_inode(struct inode *inode) fi->nlookup); fi->forget = NULL; } + + if (fi->submount_lookup) { + fuse_cleanup_submount_lookup(fc, fi->submount_lookup); + fi->submount_lookup = NULL; + } } if (S_ISREG(inode->i_mode) && !fuse_is_bad(inode)) { WARN_ON(!list_empty(&fi->write_files)); @@ -330,6 +365,13 @@ void fuse_change_attributes(struct inode *inode, struct fuse_attr *attr, fuse_dax_dontcache(inode, attr->flags); } +static void fuse_init_submount_lookup(struct fuse_submount_lookup *sl, + u64 nodeid) +{ + sl->nodeid = nodeid; + refcount_set(&sl->count, 1); +} + static void fuse_init_inode(struct inode *inode, struct fuse_attr *attr, struct fuse_conn *fc) { @@ -392,12 +434,22 @@ struct inode *fuse_iget(struct super_block *sb, u64 nodeid, */ if (fc->auto_submounts && (attr->flags & FUSE_ATTR_SUBMOUNT) && S_ISDIR(attr->mode)) { + struct fuse_inode *fi; + inode = new_inode(sb); if (!inode) return NULL; fuse_init_inode(inode, attr, fc); - get_fuse_inode(inode)->nodeid = nodeid; + fi = get_fuse_inode(inode); + fi->nodeid = nodeid; + fi->submount_lookup = fuse_alloc_submount_lookup(); + if (!fi->submount_lookup) { + iput(inode); + return NULL; + } + /* Sets nlookup = 1 on fi->submount_lookup->nlookup */ + fuse_init_submount_lookup(fi->submount_lookup, nodeid); inode->i_flags |= S_AUTOMOUNT; goto done; } @@ -420,11 +472,11 @@ struct inode *fuse_iget(struct super_block *sb, u64 nodeid, iput(inode); goto retry; } -done: fi = get_fuse_inode(inode); spin_lock(&fi->lock); fi->nlookup++; spin_unlock(&fi->lock); +done: fuse_change_attributes(inode, attr, NULL, attr_valid, attr_version); return inode; @@ -1465,6 +1517,8 @@ static int fuse_fill_super_submount(struct super_block *sb, struct super_block *parent_sb = parent_fi->inode.i_sb; struct fuse_attr root_attr; struct inode *root; + struct fuse_submount_lookup *sl; + struct fuse_inode *fi; fuse_sb_defaults(sb); fm->sb = sb; @@ -1487,12 +1541,27 @@ static int fuse_fill_super_submount(struct super_block *sb, * its nlookup should not be incremented. fuse_iget() does * that, though, so undo it here. */ - get_fuse_inode(root)->nlookup--; + fi = get_fuse_inode(root); + fi->nlookup--; + sb->s_d_op = &fuse_dentry_operations; sb->s_root = d_make_root(root); if (!sb->s_root) return -ENOMEM; + /* + * Grab the parent's submount_lookup pointer and take a + * reference on the shared nlookup from the parent. This is to + * prevent the last forget for this nodeid from getting + * triggered until all users have finished with it. + */ + sl = parent_fi->submount_lookup; + WARN_ON(!sl); + if (sl) { + refcount_inc(&sl->count); + fi->submount_lookup = sl; + } + return 0; }
