* [PATCH v2] ovl: use a dedicated semaphore for dir upperfile caching
@ 2021-01-04 7:28 Icenowy Zheng
2021-01-04 8:14 ` Icenowy Zheng
0 siblings, 1 reply; 2+ messages in thread
From: Icenowy Zheng @ 2021-01-04 7:28 UTC (permalink / raw)
To: Miklos Szeredi, Amir Goldstein, Xiao Yang
Cc: linux-unionfs, linux-kernel, Icenowy Zheng, stable
The function ovl_dir_real_file() currently uses the semaphore of the
inode to synchronize write to the upperfile cache field.
However, this function will get called by ovl_ioctl_set_flags(), which
utilizes the inode semaphore too. In this case ovl_dir_real_file() will
try to claim a lock that is owned by a function in its call stack, which
won't get released before ovl_dir_real_file() returns.
Define a dedicated semaphore for the upperfile cache, so that the
deadlock won't happen.
Fixes: 61536bed2149 ("ovl: support [S|G]ETFLAGS and FS[S|G]ETXATTR ioctls for directories")
Cc: stable@vger.kernel.org # v5.10
Signed-off-by: Icenowy Zheng <icenowy@aosc.io>
---
fs/overlayfs/readdir.c | 10 +++++-----
1 file changed, 5 insertions(+), 5 deletions(-)
diff --git a/fs/overlayfs/readdir.c b/fs/overlayfs/readdir.c
index 01620ebae1bd..fa1844ff8db4 100644
--- a/fs/overlayfs/readdir.c
+++ b/fs/overlayfs/readdir.c
@@ -56,6 +56,7 @@ struct ovl_dir_file {
struct list_head *cursor;
struct file *realfile;
struct file *upperfile;
+ struct semaphore upperfile_sem;
};
static struct ovl_cache_entry *ovl_cache_entry_from_node(struct rb_node *n)
@@ -874,8 +875,6 @@ struct file *ovl_dir_real_file(const struct file *file, bool want_upper)
* Need to check if we started out being a lower dir, but got copied up
*/
if (!od->is_upper) {
- struct inode *inode = file_inode(file);
-
realfile = READ_ONCE(od->upperfile);
if (!realfile) {
struct path upperpath;
@@ -883,10 +882,10 @@ struct file *ovl_dir_real_file(const struct file *file, bool want_upper)
ovl_path_upper(dentry, &upperpath);
realfile = ovl_dir_open_realfile(file, &upperpath);
- inode_lock(inode);
+ down(&od->upperfile_sem);
if (!od->upperfile) {
if (IS_ERR(realfile)) {
- inode_unlock(inode);
+ up(&od->upperfile_sem);
return realfile;
}
smp_store_release(&od->upperfile, realfile);
@@ -896,7 +895,7 @@ struct file *ovl_dir_real_file(const struct file *file, bool want_upper)
fput(realfile);
realfile = od->upperfile;
}
- inode_unlock(inode);
+ up(&od->upperfile_sem);
}
}
@@ -959,6 +958,7 @@ static int ovl_dir_open(struct inode *inode, struct file *file)
od->realfile = realfile;
od->is_real = ovl_dir_is_real(file->f_path.dentry);
od->is_upper = OVL_TYPE_UPPER(type);
+ sema_init(&od->upperfile_sem, 1);
file->private_data = od;
return 0;
--
2.28.0
^ permalink raw reply related [flat|nested] 2+ messages in thread
* Re: [PATCH v2] ovl: use a dedicated semaphore for dir upperfile caching
2021-01-04 7:28 [PATCH v2] ovl: use a dedicated semaphore for dir upperfile caching Icenowy Zheng
@ 2021-01-04 8:14 ` Icenowy Zheng
0 siblings, 0 replies; 2+ messages in thread
From: Icenowy Zheng @ 2021-01-04 8:14 UTC (permalink / raw)
To: Miklos Szeredi, Amir Goldstein, Xiao Yang
Cc: linux-unionfs, linux-kernel, stable
于 2021年1月4日 GMT+08:00 下午3:28:35, Icenowy Zheng <icenowy@aosc.io> 写到:
>The function ovl_dir_real_file() currently uses the semaphore of the
>inode to synchronize write to the upperfile cache field.
>
>However, this function will get called by ovl_ioctl_set_flags(), which
>utilizes the inode semaphore too. In this case ovl_dir_real_file() will
>try to claim a lock that is owned by a function in its call stack,
>which
>won't get released before ovl_dir_real_file() returns.
>
>Define a dedicated semaphore for the upperfile cache, so that the
>deadlock won't happen.
>
>Fixes: 61536bed2149 ("ovl: support [S|G]ETFLAGS and FS[S|G]ETXATTR
>ioctls for directories")
>Cc: stable@vger.kernel.org # v5.10
>Signed-off-by: Icenowy Zheng <icenowy@aosc.io>
>---
Sorry for lack of changelog.
A missing replacement of inode_lock() is added in v2.
> fs/overlayfs/readdir.c | 10 +++++-----
> 1 file changed, 5 insertions(+), 5 deletions(-)
>
>diff --git a/fs/overlayfs/readdir.c b/fs/overlayfs/readdir.c
>index 01620ebae1bd..fa1844ff8db4 100644
>--- a/fs/overlayfs/readdir.c
>+++ b/fs/overlayfs/readdir.c
>@@ -56,6 +56,7 @@ struct ovl_dir_file {
> struct list_head *cursor;
> struct file *realfile;
> struct file *upperfile;
>+ struct semaphore upperfile_sem;
> };
>
>static struct ovl_cache_entry *ovl_cache_entry_from_node(struct rb_node
>*n)
>@@ -874,8 +875,6 @@ struct file *ovl_dir_real_file(const struct file
>*file, bool want_upper)
> * Need to check if we started out being a lower dir, but got copied
>up
> */
> if (!od->is_upper) {
>- struct inode *inode = file_inode(file);
>-
> realfile = READ_ONCE(od->upperfile);
> if (!realfile) {
> struct path upperpath;
>@@ -883,10 +882,10 @@ struct file *ovl_dir_real_file(const struct file
>*file, bool want_upper)
> ovl_path_upper(dentry, &upperpath);
> realfile = ovl_dir_open_realfile(file, &upperpath);
>
>- inode_lock(inode);
>+ down(&od->upperfile_sem);
> if (!od->upperfile) {
> if (IS_ERR(realfile)) {
>- inode_unlock(inode);
>+ up(&od->upperfile_sem);
> return realfile;
> }
> smp_store_release(&od->upperfile, realfile);
>@@ -896,7 +895,7 @@ struct file *ovl_dir_real_file(const struct file
>*file, bool want_upper)
> fput(realfile);
> realfile = od->upperfile;
> }
>- inode_unlock(inode);
>+ up(&od->upperfile_sem);
> }
> }
>
>@@ -959,6 +958,7 @@ static int ovl_dir_open(struct inode *inode, struct
>file *file)
> od->realfile = realfile;
> od->is_real = ovl_dir_is_real(file->f_path.dentry);
> od->is_upper = OVL_TYPE_UPPER(type);
>+ sema_init(&od->upperfile_sem, 1);
> file->private_data = od;
>
> return 0;
^ permalink raw reply [flat|nested] 2+ messages in thread
end of thread, other threads:[~2021-01-04 8:44 UTC | newest]
Thread overview: 2+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2021-01-04 7:28 [PATCH v2] ovl: use a dedicated semaphore for dir upperfile caching Icenowy Zheng
2021-01-04 8:14 ` Icenowy Zheng
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).