[PATCH] erofs: ensure valid f_path for page cache sharing

[PATCH] erofs: ensure valid f_path for page cache sharing

[Gao Xiang]

Previously, backing files for page cache sharing were set up with
f_path left as NULL (only f_inode was valid).  It worked, but a recent
mincore fix relies on f_path.mnt and crashes (found by "erofs/028" on
7.2-rc4):

 BUG: kernel NULL pointer dereference, address: 0000000000000018
 #PF: supervisor read access in kernel mode
 #PF: error_code(0x0000) - not-present page
 PGD 0 P4D 0
 Oops: Oops: 0000 [#1] SMP PTI
 CPU: 3 UID: 0 PID: 675528 Comm: fincore Not tainted 7.2.0-rc4-00002-g[]-dirty #1 PREEMPT(lazy)
 Hardware name: Red Hat KVM, BIOS 1.16.0-4.al8 04/01/2014
 RIP: 0010:__do_sys_mincore+0xc0/0x2c0
 ...

Specify valid paths using valid disconnected dentries together with
erofs_ishare_mnt instead of leaving f_path empty, so they are more
like real backing files in a pseudo filesystem and standard
backing_file_open() can be used directly.

Fixes: e187bc02f8fa ("mm: do file ownership checks with the proper mount idmap")
Signed-off-by: Gao Xiang <xiang@kernel.org>
---
 fs/erofs/Kconfig    |  1 +
 fs/erofs/internal.h |  4 +--
 fs/erofs/ishare.c   | 60 ++++++++++++++++++++++-----------------------
 3 files changed, 33 insertions(+), 32 deletions(-)

diff --git a/fs/erofs/Kconfig b/fs/erofs/Kconfig
index 4789b1077d8c..1feb28cfe557 100644
--- a/fs/erofs/Kconfig
+++ b/fs/erofs/Kconfig
@@ -189,6 +189,7 @@ config EROFS_FS_PCPU_KTHREAD_HIPRI
 config EROFS_FS_PAGE_CACHE_SHARE
 	bool "EROFS page cache share support (experimental)"
 	depends on EROFS_FS && EROFS_FS_XATTR
+	select FS_STACK
 	help
 	  This enables page cache sharing among inodes with identical
 	  content fingerprints on the same machine.
diff --git a/fs/erofs/internal.h b/fs/erofs/internal.h
index 580f8d9f14e7..57bd21859c65 100644
--- a/fs/erofs/internal.h
+++ b/fs/erofs/internal.h
@@ -288,8 +288,8 @@ struct erofs_inode {
 			struct erofs_inode_fingerprint fingerprint;
 			spinlock_t ishare_lock;
 		};
-		/* for each real inode */
-		struct inode *sharedinode;
+		/* for each real filesystem inode */
+		struct dentry *sharedentry;
 	};
 #endif
 	/* the corresponding vfs inode */
diff --git a/fs/erofs/ishare.c b/fs/erofs/ishare.c
index a1a20a5ba548..51b81aaf9ac3 100644
--- a/fs/erofs/ishare.c
+++ b/fs/erofs/ishare.c
@@ -2,14 +2,13 @@
 /*
  * Copyright (C) 2024, Alibaba Cloud
  */
+#include <linux/backing-file.h>
 #include <linux/xxhash.h>
 #include <linux/mount.h>
 #include <linux/security.h>
 #include "internal.h"
 #include "xattr.h"
 
-#include "../internal.h"
-
 static struct vfsmount *erofs_ishare_mnt;
 
 static int erofs_ishare_iget5_eq(struct inode *inode, void *data)
@@ -33,10 +32,12 @@ static int erofs_ishare_iget5_set(struct inode *inode, void *data)
 
 bool erofs_ishare_fill_inode(struct inode *inode)
 {
+	static const struct file_operations empty_fops = {};
 	struct erofs_sb_info *sbi = EROFS_SB(inode->i_sb);
 	const struct address_space_operations *aops;
 	struct erofs_inode *vi = EROFS_I(inode);
 	struct erofs_inode_fingerprint fp;
+	struct dentry *sd;
 	struct inode *si;
 
 	aops = erofs_get_aops(inode);
@@ -49,7 +50,9 @@ bool erofs_ishare_fill_inode(struct inode *inode)
 			  xxh32(fp.opaque, fp.size, 0),
 			  erofs_ishare_iget5_eq, erofs_ishare_iget5_set, &fp);
 	if (si && (inode_state_read_once(si) & I_NEW)) {
+		si->i_fop = &empty_fops;
 		si->i_mapping->a_ops = aops;
+		si->i_mode = 0444 | S_IFREG;
 		si->i_size = inode->i_size;
 		unlock_new_inode(si);
 	} else {
@@ -65,7 +68,12 @@ bool erofs_ishare_fill_inode(struct inode *inode)
 			return false;
 		}
 	}
-	vi->sharedinode = si;
+	sd = d_obtain_alias(si); /* disconnected denties for sharedinodes */
+	if (IS_ERR(sd)) {
+		iput(si);
+		return false;
+	}
+	vi->sharedentry = sd;
 	INIT_LIST_HEAD(&vi->ishare_list);
 	spin_lock(&EROFS_I(si)->ishare_lock);
 	list_add(&vi->ishare_list, &EROFS_I(si)->ishare_list);
@@ -75,48 +83,40 @@ bool erofs_ishare_fill_inode(struct inode *inode)
 
 void erofs_ishare_free_inode(struct inode *inode)
 {
-	struct erofs_inode *vi = EROFS_I(inode);
-	struct inode *sharedinode = vi->sharedinode;
+	struct erofs_inode *vi = EROFS_I(inode), *svi;
 
-	if (!sharedinode)
+	if (!vi->sharedentry)
 		return;
-	spin_lock(&EROFS_I(sharedinode)->ishare_lock);
+	svi = EROFS_I(d_inode(vi->sharedentry));
+	spin_lock(&svi->ishare_lock);
 	list_del(&vi->ishare_list);
-	spin_unlock(&EROFS_I(sharedinode)->ishare_lock);
-	iput(sharedinode);
-	vi->sharedinode = NULL;
+	spin_unlock(&svi->ishare_lock);
+	dput(vi->sharedentry);
+	vi->sharedentry = NULL;
 }
 
 static int erofs_ishare_file_open(struct inode *inode, struct file *file)
 {
-	struct inode *sharedinode = EROFS_I(inode)->sharedinode;
-	struct file *realfile;
+	struct path sharedpath = {
+		.mnt = erofs_ishare_mnt,
+		.dentry = EROFS_I(inode)->sharedentry,
+	};
+	struct file *rf;
 
 	if (file->f_flags & O_DIRECT)
 		return -EINVAL;
-	realfile = alloc_empty_backing_file(O_RDONLY|O_NOATIME, current_cred(),
-					    file);
-	if (IS_ERR(realfile))
-		return PTR_ERR(realfile);
-	ihold(sharedinode);
-	realfile->f_op = &erofs_file_fops;
-	realfile->f_inode = sharedinode;
-	realfile->f_mapping = sharedinode->i_mapping;
-	path_get(&file->f_path);
-	backing_file_set_user_path(realfile, &file->f_path);
-
-	file_ra_state_init(&realfile->f_ra, file->f_mapping);
-	realfile->private_data = EROFS_I(inode);
-	file->private_data = realfile;
+
+	rf = backing_file_open(file, file->f_flags | O_NOATIME | FMODE_NONOTIFY,
+			       &sharedpath, current_cred());
+	if (IS_ERR(rf))
+		return PTR_ERR(rf);
+	file->private_data = rf;
 	return 0;
 }
 
 static int erofs_ishare_file_release(struct inode *inode, struct file *file)
 {
-	struct file *realfile = file->private_data;
-
-	iput(realfile->f_inode);
-	fput(realfile);
+	fput(file->private_data);
 	file->private_data = NULL;
 	return 0;
 }
-- 
2.47.3