* [PATCH v5 0/3] virtiofsd: prevent opening of special files (CVE-2020-35517)
@ 2021-02-04 15:02 ` Stefan Hajnoczi
0 siblings, 0 replies; 17+ messages in thread
From: Stefan Hajnoczi @ 2021-02-04 15:02 UTC (permalink / raw)
To: qemu-devel
Cc: mszeredi, Daniel Berrange, slp, Greg Kurz, P J P, virtio-fs,
Alex Xu, vgoyal, Stefan Hajnoczi, Laszlo Ersek,
Dr. David Alan Gilbert
v4:
* Patch 1: Return positive errno if openat(2) fails in lo_do_open() [Greg]
* Patch 3: Return -fd instead or -errno after lo_inode_open() in lo_do_open() [Greg]
* Patch 3: Use De Morgan's Law to simplify the boolean expression in lo_create() [Vivek]
* Patch 3: Add missing errno = -truncfd after lo_inode_open() call in lo_setattr
v3:
* Restructure lo_create() to handle externally-created files (we need
to allocate an inode for them) [Greg]
* Patch 1 & 2 refactor the code so that Patch 3 can implement the CVE fix
v3:
* Protect lo_create() [Greg]
v2:
* Add doc comment clarifying that symlinks are traversed client-side
[Daniel]
A well-behaved FUSE client does not attempt to open special files with
FUSE_OPEN because they are handled on the client side (e.g. device nodes
are handled by client-side device drivers).
The check to prevent virtiofsd from opening special files is missing in
a few cases, most notably FUSE_OPEN. A malicious client can cause
virtiofsd to open a device node, potentially allowing the guest to
escape. This can be exploited by a modified guest device driver. It is
not exploitable from guest userspace since the guest kernel will handle
special files inside the guest instead of sending FUSE requests.
This patch series fixes this issue by introducing the lo_inode_open() function
to check the file type before opening it. This is a short-term solution because
it does not prevent a compromised virtiofsd process from opening device nodes
on the host.
This issue was diagnosed on public IRC and is therefore already known
and not embargoed.
Reported-by: Alex Xu <alex@alxu.ca>
Fixes: CVE-2020-35517
Stefan Hajnoczi (3):
virtiofsd: extract lo_do_open() from lo_open()
virtiofsd: optionally return inode pointer from lo_do_lookup()
virtiofsd: prevent opening of special files (CVE-2020-35517)
tools/virtiofsd/passthrough_ll.c | 224 ++++++++++++++++++++-----------
1 file changed, 148 insertions(+), 76 deletions(-)
--
2.29.2
^ permalink raw reply [flat|nested] 17+ messages in thread
* [Virtio-fs] [PATCH v5 0/3] virtiofsd: prevent opening of special files (CVE-2020-35517)
@ 2021-02-04 15:02 ` Stefan Hajnoczi
0 siblings, 0 replies; 17+ messages in thread
From: Stefan Hajnoczi @ 2021-02-04 15:02 UTC (permalink / raw)
To: qemu-devel
Cc: Daniel Berrange, P J P, virtio-fs, Alex Xu, vgoyal, Laszlo Ersek
v4:
* Patch 1: Return positive errno if openat(2) fails in lo_do_open() [Greg]
* Patch 3: Return -fd instead or -errno after lo_inode_open() in lo_do_open() [Greg]
* Patch 3: Use De Morgan's Law to simplify the boolean expression in lo_create() [Vivek]
* Patch 3: Add missing errno = -truncfd after lo_inode_open() call in lo_setattr
v3:
* Restructure lo_create() to handle externally-created files (we need
to allocate an inode for them) [Greg]
* Patch 1 & 2 refactor the code so that Patch 3 can implement the CVE fix
v3:
* Protect lo_create() [Greg]
v2:
* Add doc comment clarifying that symlinks are traversed client-side
[Daniel]
A well-behaved FUSE client does not attempt to open special files with
FUSE_OPEN because they are handled on the client side (e.g. device nodes
are handled by client-side device drivers).
The check to prevent virtiofsd from opening special files is missing in
a few cases, most notably FUSE_OPEN. A malicious client can cause
virtiofsd to open a device node, potentially allowing the guest to
escape. This can be exploited by a modified guest device driver. It is
not exploitable from guest userspace since the guest kernel will handle
special files inside the guest instead of sending FUSE requests.
This patch series fixes this issue by introducing the lo_inode_open() function
to check the file type before opening it. This is a short-term solution because
it does not prevent a compromised virtiofsd process from opening device nodes
on the host.
This issue was diagnosed on public IRC and is therefore already known
and not embargoed.
Reported-by: Alex Xu <alex@alxu.ca>
Fixes: CVE-2020-35517
Stefan Hajnoczi (3):
virtiofsd: extract lo_do_open() from lo_open()
virtiofsd: optionally return inode pointer from lo_do_lookup()
virtiofsd: prevent opening of special files (CVE-2020-35517)
tools/virtiofsd/passthrough_ll.c | 224 ++++++++++++++++++++-----------
1 file changed, 148 insertions(+), 76 deletions(-)
--
2.29.2
^ permalink raw reply [flat|nested] 17+ messages in thread
* [PATCH v5 1/3] virtiofsd: extract lo_do_open() from lo_open()
2021-02-04 15:02 ` [Virtio-fs] " Stefan Hajnoczi
@ 2021-02-04 15:02 ` Stefan Hajnoczi
-1 siblings, 0 replies; 17+ messages in thread
From: Stefan Hajnoczi @ 2021-02-04 15:02 UTC (permalink / raw)
To: qemu-devel
Cc: mszeredi, Daniel Berrange, slp, Greg Kurz, P J P, virtio-fs,
vgoyal, Stefan Hajnoczi, Laszlo Ersek, Dr. David Alan Gilbert
Both lo_open() and lo_create() have similar code to open a file. Extract
a common lo_do_open() function from lo_open() that will be used by
lo_create() in a later commit.
Since lo_do_open() does not otherwise need fuse_req_t req, convert
lo_add_fd_mapping() to use struct lo_data *lo instead.
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
---
v4:
* Return positive errno if openat(2) fails in lo_do_open() [Greg]
---
tools/virtiofsd/passthrough_ll.c | 73 ++++++++++++++++++++------------
1 file changed, 46 insertions(+), 27 deletions(-)
diff --git a/tools/virtiofsd/passthrough_ll.c b/tools/virtiofsd/passthrough_ll.c
index 5fb36d9407..f14fa5124d 100644
--- a/tools/virtiofsd/passthrough_ll.c
+++ b/tools/virtiofsd/passthrough_ll.c
@@ -459,17 +459,17 @@ static void lo_map_remove(struct lo_map *map, size_t key)
}
/* Assumes lo->mutex is held */
-static ssize_t lo_add_fd_mapping(fuse_req_t req, int fd)
+static ssize_t lo_add_fd_mapping(struct lo_data *lo, int fd)
{
struct lo_map_elem *elem;
- elem = lo_map_alloc_elem(&lo_data(req)->fd_map);
+ elem = lo_map_alloc_elem(&lo->fd_map);
if (!elem) {
return -1;
}
elem->fd = fd;
- return elem - lo_data(req)->fd_map.elems;
+ return elem - lo->fd_map.elems;
}
/* Assumes lo->mutex is held */
@@ -1651,6 +1651,38 @@ static void update_open_flags(int writeback, int allow_direct_io,
}
}
+static int lo_do_open(struct lo_data *lo, struct lo_inode *inode,
+ struct fuse_file_info *fi)
+{
+ char buf[64];
+ ssize_t fh;
+ int fd;
+
+ update_open_flags(lo->writeback, lo->allow_direct_io, fi);
+
+ sprintf(buf, "%i", inode->fd);
+ fd = openat(lo->proc_self_fd, buf, fi->flags & ~O_NOFOLLOW);
+ if (fd == -1) {
+ return errno;
+ }
+
+ pthread_mutex_lock(&lo->mutex);
+ fh = lo_add_fd_mapping(lo, fd);
+ pthread_mutex_unlock(&lo->mutex);
+ if (fh == -1) {
+ close(fd);
+ return ENOMEM;
+ }
+
+ fi->fh = fh;
+ if (lo->cache == CACHE_NONE) {
+ fi->direct_io = 1;
+ } else if (lo->cache == CACHE_ALWAYS) {
+ fi->keep_cache = 1;
+ }
+ return 0;
+}
+
static void lo_create(fuse_req_t req, fuse_ino_t parent, const char *name,
mode_t mode, struct fuse_file_info *fi)
{
@@ -1691,7 +1723,7 @@ static void lo_create(fuse_req_t req, fuse_ino_t parent, const char *name,
ssize_t fh;
pthread_mutex_lock(&lo->mutex);
- fh = lo_add_fd_mapping(req, fd);
+ fh = lo_add_fd_mapping(lo, fd);
pthread_mutex_unlock(&lo->mutex);
if (fh == -1) {
close(fd);
@@ -1892,38 +1924,25 @@ static void lo_fsyncdir(fuse_req_t req, fuse_ino_t ino, int datasync,
static void lo_open(fuse_req_t req, fuse_ino_t ino, struct fuse_file_info *fi)
{
- int fd;
- ssize_t fh;
- char buf[64];
struct lo_data *lo = lo_data(req);
+ struct lo_inode *inode = lo_inode(req, ino);
+ int err;
fuse_log(FUSE_LOG_DEBUG, "lo_open(ino=%" PRIu64 ", flags=%d)\n", ino,
fi->flags);
- update_open_flags(lo->writeback, lo->allow_direct_io, fi);
-
- sprintf(buf, "%i", lo_fd(req, ino));
- fd = openat(lo->proc_self_fd, buf, fi->flags & ~O_NOFOLLOW);
- if (fd == -1) {
- return (void)fuse_reply_err(req, errno);
- }
-
- pthread_mutex_lock(&lo->mutex);
- fh = lo_add_fd_mapping(req, fd);
- pthread_mutex_unlock(&lo->mutex);
- if (fh == -1) {
- close(fd);
- fuse_reply_err(req, ENOMEM);
+ if (!inode) {
+ fuse_reply_err(req, EBADF);
return;
}
- fi->fh = fh;
- if (lo->cache == CACHE_NONE) {
- fi->direct_io = 1;
- } else if (lo->cache == CACHE_ALWAYS) {
- fi->keep_cache = 1;
+ err = lo_do_open(lo, inode, fi);
+ lo_inode_put(lo, &inode);
+ if (err) {
+ fuse_reply_err(req, err);
+ } else {
+ fuse_reply_open(req, fi);
}
- fuse_reply_open(req, fi);
}
static void lo_release(fuse_req_t req, fuse_ino_t ino,
--
2.29.2
^ permalink raw reply related [flat|nested] 17+ messages in thread
* [Virtio-fs] [PATCH v5 1/3] virtiofsd: extract lo_do_open() from lo_open()
@ 2021-02-04 15:02 ` Stefan Hajnoczi
0 siblings, 0 replies; 17+ messages in thread
From: Stefan Hajnoczi @ 2021-02-04 15:02 UTC (permalink / raw)
To: qemu-devel; +Cc: Daniel Berrange, P J P, virtio-fs, vgoyal, Laszlo Ersek
Both lo_open() and lo_create() have similar code to open a file. Extract
a common lo_do_open() function from lo_open() that will be used by
lo_create() in a later commit.
Since lo_do_open() does not otherwise need fuse_req_t req, convert
lo_add_fd_mapping() to use struct lo_data *lo instead.
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
---
v4:
* Return positive errno if openat(2) fails in lo_do_open() [Greg]
---
tools/virtiofsd/passthrough_ll.c | 73 ++++++++++++++++++++------------
1 file changed, 46 insertions(+), 27 deletions(-)
diff --git a/tools/virtiofsd/passthrough_ll.c b/tools/virtiofsd/passthrough_ll.c
index 5fb36d9407..f14fa5124d 100644
--- a/tools/virtiofsd/passthrough_ll.c
+++ b/tools/virtiofsd/passthrough_ll.c
@@ -459,17 +459,17 @@ static void lo_map_remove(struct lo_map *map, size_t key)
}
/* Assumes lo->mutex is held */
-static ssize_t lo_add_fd_mapping(fuse_req_t req, int fd)
+static ssize_t lo_add_fd_mapping(struct lo_data *lo, int fd)
{
struct lo_map_elem *elem;
- elem = lo_map_alloc_elem(&lo_data(req)->fd_map);
+ elem = lo_map_alloc_elem(&lo->fd_map);
if (!elem) {
return -1;
}
elem->fd = fd;
- return elem - lo_data(req)->fd_map.elems;
+ return elem - lo->fd_map.elems;
}
/* Assumes lo->mutex is held */
@@ -1651,6 +1651,38 @@ static void update_open_flags(int writeback, int allow_direct_io,
}
}
+static int lo_do_open(struct lo_data *lo, struct lo_inode *inode,
+ struct fuse_file_info *fi)
+{
+ char buf[64];
+ ssize_t fh;
+ int fd;
+
+ update_open_flags(lo->writeback, lo->allow_direct_io, fi);
+
+ sprintf(buf, "%i", inode->fd);
+ fd = openat(lo->proc_self_fd, buf, fi->flags & ~O_NOFOLLOW);
+ if (fd == -1) {
+ return errno;
+ }
+
+ pthread_mutex_lock(&lo->mutex);
+ fh = lo_add_fd_mapping(lo, fd);
+ pthread_mutex_unlock(&lo->mutex);
+ if (fh == -1) {
+ close(fd);
+ return ENOMEM;
+ }
+
+ fi->fh = fh;
+ if (lo->cache == CACHE_NONE) {
+ fi->direct_io = 1;
+ } else if (lo->cache == CACHE_ALWAYS) {
+ fi->keep_cache = 1;
+ }
+ return 0;
+}
+
static void lo_create(fuse_req_t req, fuse_ino_t parent, const char *name,
mode_t mode, struct fuse_file_info *fi)
{
@@ -1691,7 +1723,7 @@ static void lo_create(fuse_req_t req, fuse_ino_t parent, const char *name,
ssize_t fh;
pthread_mutex_lock(&lo->mutex);
- fh = lo_add_fd_mapping(req, fd);
+ fh = lo_add_fd_mapping(lo, fd);
pthread_mutex_unlock(&lo->mutex);
if (fh == -1) {
close(fd);
@@ -1892,38 +1924,25 @@ static void lo_fsyncdir(fuse_req_t req, fuse_ino_t ino, int datasync,
static void lo_open(fuse_req_t req, fuse_ino_t ino, struct fuse_file_info *fi)
{
- int fd;
- ssize_t fh;
- char buf[64];
struct lo_data *lo = lo_data(req);
+ struct lo_inode *inode = lo_inode(req, ino);
+ int err;
fuse_log(FUSE_LOG_DEBUG, "lo_open(ino=%" PRIu64 ", flags=%d)\n", ino,
fi->flags);
- update_open_flags(lo->writeback, lo->allow_direct_io, fi);
-
- sprintf(buf, "%i", lo_fd(req, ino));
- fd = openat(lo->proc_self_fd, buf, fi->flags & ~O_NOFOLLOW);
- if (fd == -1) {
- return (void)fuse_reply_err(req, errno);
- }
-
- pthread_mutex_lock(&lo->mutex);
- fh = lo_add_fd_mapping(req, fd);
- pthread_mutex_unlock(&lo->mutex);
- if (fh == -1) {
- close(fd);
- fuse_reply_err(req, ENOMEM);
+ if (!inode) {
+ fuse_reply_err(req, EBADF);
return;
}
- fi->fh = fh;
- if (lo->cache == CACHE_NONE) {
- fi->direct_io = 1;
- } else if (lo->cache == CACHE_ALWAYS) {
- fi->keep_cache = 1;
+ err = lo_do_open(lo, inode, fi);
+ lo_inode_put(lo, &inode);
+ if (err) {
+ fuse_reply_err(req, err);
+ } else {
+ fuse_reply_open(req, fi);
}
- fuse_reply_open(req, fi);
}
static void lo_release(fuse_req_t req, fuse_ino_t ino,
--
2.29.2
^ permalink raw reply related [flat|nested] 17+ messages in thread
* [PATCH v5 2/3] virtiofsd: optionally return inode pointer from lo_do_lookup()
2021-02-04 15:02 ` [Virtio-fs] " Stefan Hajnoczi
@ 2021-02-04 15:02 ` Stefan Hajnoczi
-1 siblings, 0 replies; 17+ messages in thread
From: Stefan Hajnoczi @ 2021-02-04 15:02 UTC (permalink / raw)
To: qemu-devel
Cc: mszeredi, Daniel Berrange, slp, Greg Kurz, P J P, virtio-fs,
vgoyal, Stefan Hajnoczi, Laszlo Ersek, Dr. David Alan Gilbert
lo_do_lookup() finds an existing inode or allocates a new one. It
increments nlookup so that the inode stays alive until the client
releases it.
Existing callers don't need the struct lo_inode so the function doesn't
return it. Extend the function to optionally return the inode. The next
commit will need it.
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Greg Kurz <groug@kaod.org>
---
tools/virtiofsd/passthrough_ll.c | 29 +++++++++++++++++++++--------
1 file changed, 21 insertions(+), 8 deletions(-)
diff --git a/tools/virtiofsd/passthrough_ll.c b/tools/virtiofsd/passthrough_ll.c
index f14fa5124d..aa35fc6ba5 100644
--- a/tools/virtiofsd/passthrough_ll.c
+++ b/tools/virtiofsd/passthrough_ll.c
@@ -831,11 +831,13 @@ static int do_statx(struct lo_data *lo, int dirfd, const char *pathname,
}
/*
- * Increments nlookup and caller must release refcount using
- * lo_inode_put(&parent).
+ * Increments nlookup on the inode on success. unref_inode_lolocked() must be
+ * called eventually to decrement nlookup again. If inodep is non-NULL, the
+ * inode pointer is stored and the caller must call lo_inode_put().
*/
static int lo_do_lookup(fuse_req_t req, fuse_ino_t parent, const char *name,
- struct fuse_entry_param *e)
+ struct fuse_entry_param *e,
+ struct lo_inode **inodep)
{
int newfd;
int res;
@@ -845,6 +847,10 @@ static int lo_do_lookup(fuse_req_t req, fuse_ino_t parent, const char *name,
struct lo_inode *inode = NULL;
struct lo_inode *dir = lo_inode(req, parent);
+ if (inodep) {
+ *inodep = NULL;
+ }
+
/*
* name_to_handle_at() and open_by_handle_at() can reach here with fuse
* mount point in guest, but we don't have its inode info in the
@@ -913,7 +919,14 @@ static int lo_do_lookup(fuse_req_t req, fuse_ino_t parent, const char *name,
pthread_mutex_unlock(&lo->mutex);
}
e->ino = inode->fuse_ino;
- lo_inode_put(lo, &inode);
+
+ /* Transfer ownership of inode pointer to caller or drop it */
+ if (inodep) {
+ *inodep = inode;
+ } else {
+ lo_inode_put(lo, &inode);
+ }
+
lo_inode_put(lo, &dir);
fuse_log(FUSE_LOG_DEBUG, " %lli/%s -> %lli\n", (unsigned long long)parent,
@@ -948,7 +961,7 @@ static void lo_lookup(fuse_req_t req, fuse_ino_t parent, const char *name)
return;
}
- err = lo_do_lookup(req, parent, name, &e);
+ err = lo_do_lookup(req, parent, name, &e, NULL);
if (err) {
fuse_reply_err(req, err);
} else {
@@ -1056,7 +1069,7 @@ static void lo_mknod_symlink(fuse_req_t req, fuse_ino_t parent,
goto out;
}
- saverr = lo_do_lookup(req, parent, name, &e);
+ saverr = lo_do_lookup(req, parent, name, &e, NULL);
if (saverr) {
goto out;
}
@@ -1534,7 +1547,7 @@ static void lo_do_readdir(fuse_req_t req, fuse_ino_t ino, size_t size,
if (plus) {
if (!is_dot_or_dotdot(name)) {
- err = lo_do_lookup(req, ino, name, &e);
+ err = lo_do_lookup(req, ino, name, &e, NULL);
if (err) {
goto error;
}
@@ -1732,7 +1745,7 @@ static void lo_create(fuse_req_t req, fuse_ino_t parent, const char *name,
}
fi->fh = fh;
- err = lo_do_lookup(req, parent, name, &e);
+ err = lo_do_lookup(req, parent, name, &e, NULL);
}
if (lo->cache == CACHE_NONE) {
fi->direct_io = 1;
--
2.29.2
^ permalink raw reply related [flat|nested] 17+ messages in thread
* [Virtio-fs] [PATCH v5 2/3] virtiofsd: optionally return inode pointer from lo_do_lookup()
@ 2021-02-04 15:02 ` Stefan Hajnoczi
0 siblings, 0 replies; 17+ messages in thread
From: Stefan Hajnoczi @ 2021-02-04 15:02 UTC (permalink / raw)
To: qemu-devel; +Cc: Daniel Berrange, P J P, virtio-fs, vgoyal, Laszlo Ersek
lo_do_lookup() finds an existing inode or allocates a new one. It
increments nlookup so that the inode stays alive until the client
releases it.
Existing callers don't need the struct lo_inode so the function doesn't
return it. Extend the function to optionally return the inode. The next
commit will need it.
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Greg Kurz <groug@kaod.org>
---
tools/virtiofsd/passthrough_ll.c | 29 +++++++++++++++++++++--------
1 file changed, 21 insertions(+), 8 deletions(-)
diff --git a/tools/virtiofsd/passthrough_ll.c b/tools/virtiofsd/passthrough_ll.c
index f14fa5124d..aa35fc6ba5 100644
--- a/tools/virtiofsd/passthrough_ll.c
+++ b/tools/virtiofsd/passthrough_ll.c
@@ -831,11 +831,13 @@ static int do_statx(struct lo_data *lo, int dirfd, const char *pathname,
}
/*
- * Increments nlookup and caller must release refcount using
- * lo_inode_put(&parent).
+ * Increments nlookup on the inode on success. unref_inode_lolocked() must be
+ * called eventually to decrement nlookup again. If inodep is non-NULL, the
+ * inode pointer is stored and the caller must call lo_inode_put().
*/
static int lo_do_lookup(fuse_req_t req, fuse_ino_t parent, const char *name,
- struct fuse_entry_param *e)
+ struct fuse_entry_param *e,
+ struct lo_inode **inodep)
{
int newfd;
int res;
@@ -845,6 +847,10 @@ static int lo_do_lookup(fuse_req_t req, fuse_ino_t parent, const char *name,
struct lo_inode *inode = NULL;
struct lo_inode *dir = lo_inode(req, parent);
+ if (inodep) {
+ *inodep = NULL;
+ }
+
/*
* name_to_handle_at() and open_by_handle_at() can reach here with fuse
* mount point in guest, but we don't have its inode info in the
@@ -913,7 +919,14 @@ static int lo_do_lookup(fuse_req_t req, fuse_ino_t parent, const char *name,
pthread_mutex_unlock(&lo->mutex);
}
e->ino = inode->fuse_ino;
- lo_inode_put(lo, &inode);
+
+ /* Transfer ownership of inode pointer to caller or drop it */
+ if (inodep) {
+ *inodep = inode;
+ } else {
+ lo_inode_put(lo, &inode);
+ }
+
lo_inode_put(lo, &dir);
fuse_log(FUSE_LOG_DEBUG, " %lli/%s -> %lli\n", (unsigned long long)parent,
@@ -948,7 +961,7 @@ static void lo_lookup(fuse_req_t req, fuse_ino_t parent, const char *name)
return;
}
- err = lo_do_lookup(req, parent, name, &e);
+ err = lo_do_lookup(req, parent, name, &e, NULL);
if (err) {
fuse_reply_err(req, err);
} else {
@@ -1056,7 +1069,7 @@ static void lo_mknod_symlink(fuse_req_t req, fuse_ino_t parent,
goto out;
}
- saverr = lo_do_lookup(req, parent, name, &e);
+ saverr = lo_do_lookup(req, parent, name, &e, NULL);
if (saverr) {
goto out;
}
@@ -1534,7 +1547,7 @@ static void lo_do_readdir(fuse_req_t req, fuse_ino_t ino, size_t size,
if (plus) {
if (!is_dot_or_dotdot(name)) {
- err = lo_do_lookup(req, ino, name, &e);
+ err = lo_do_lookup(req, ino, name, &e, NULL);
if (err) {
goto error;
}
@@ -1732,7 +1745,7 @@ static void lo_create(fuse_req_t req, fuse_ino_t parent, const char *name,
}
fi->fh = fh;
- err = lo_do_lookup(req, parent, name, &e);
+ err = lo_do_lookup(req, parent, name, &e, NULL);
}
if (lo->cache == CACHE_NONE) {
fi->direct_io = 1;
--
2.29.2
^ permalink raw reply related [flat|nested] 17+ messages in thread
* [PATCH v5 3/3] virtiofsd: prevent opening of special files (CVE-2020-35517)
2021-02-04 15:02 ` [Virtio-fs] " Stefan Hajnoczi
@ 2021-02-04 15:02 ` Stefan Hajnoczi
-1 siblings, 0 replies; 17+ messages in thread
From: Stefan Hajnoczi @ 2021-02-04 15:02 UTC (permalink / raw)
To: qemu-devel
Cc: mszeredi, Daniel Berrange, slp, Greg Kurz, P J P, virtio-fs,
Alex Xu, vgoyal, Stefan Hajnoczi, Laszlo Ersek,
Dr. David Alan Gilbert
A well-behaved FUSE client does not attempt to open special files with
FUSE_OPEN because they are handled on the client side (e.g. device nodes
are handled by client-side device drivers).
The check to prevent virtiofsd from opening special files is missing in
a few cases, most notably FUSE_OPEN. A malicious client can cause
virtiofsd to open a device node, potentially allowing the guest to
escape. This can be exploited by a modified guest device driver. It is
not exploitable from guest userspace since the guest kernel will handle
special files inside the guest instead of sending FUSE requests.
This patch fixes this issue by introducing the lo_inode_open() function
to check the file type before opening it. This is a short-term solution
because it does not prevent a compromised virtiofsd process from opening
device nodes on the host.
Restructure lo_create() to try O_CREAT | O_EXCL first. Note that O_CREAT
| O_EXCL does not follow symlinks, so O_NOFOLLOW masking is not
necessary here. If the file exists and the user did not specify O_EXCL,
open it via lo_do_open().
Reported-by: Alex Xu <alex@alxu.ca>
Fixes: CVE-2020-35517
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Reviewed-by: Vivek Goyal <vgoyal@redhat.com>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
---
v4:
* Return -fd instead or -errno after lo_inode_open() in lo_do_open() [Greg]
* Use De Morgan's Law to simplify the boolean expression in lo_create() [Vivek]
* Add missing errno = -truncfd after lo_inode_open() call in lo_setattr
v3:
* Restructure lo_create() to handle externally-created files (we need
to allocate an inode for them) [Greg]
v3:
* Protect lo_create() [Greg]
v2:
* Add doc comment clarifying that symlinks are traversed client-side
[Daniel]
This issue was diagnosed on public IRC and is therefore already known
and not embargoed.
A stronger fix, and the long-term solution, is for users to mount the
shared directory and any sub-mounts with nodev, as well as nosuid and
noexec. Unfortunately virtiofsd cannot do this automatically because
bind mounts added by the user after virtiofsd has launched would not be
detected. I suggest the following:
1. Modify libvirt and Kata Containers to explicitly set these mount
options.
2. Then modify virtiofsd to check that the shared directory has the
necessary options at startup. Refuse to start if the options are
missing so that the user is aware of the security requirements.
As a bonus this also increases the likelihood that other host processes
besides virtiofsd will be protected by nosuid/noexec/nodev so that a
malicious guest cannot drop these files in place and then arrange for a
host process to come across them.
Additionally, user namespaces have been discussed. They seem like a
worthwhile addition as an unprivileged or privilege-separated mode
although there are limitations with respect to security xattrs and the
actual uid/gid stored on the host file system not corresponding to the
guest uid/gid.
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
---
tools/virtiofsd/passthrough_ll.c | 144 ++++++++++++++++++++-----------
1 file changed, 92 insertions(+), 52 deletions(-)
diff --git a/tools/virtiofsd/passthrough_ll.c b/tools/virtiofsd/passthrough_ll.c
index aa35fc6ba5..147b59338a 100644
--- a/tools/virtiofsd/passthrough_ll.c
+++ b/tools/virtiofsd/passthrough_ll.c
@@ -555,6 +555,38 @@ static int lo_fd(fuse_req_t req, fuse_ino_t ino)
return fd;
}
+/*
+ * Open a file descriptor for an inode. Returns -EBADF if the inode is not a
+ * regular file or a directory.
+ *
+ * Use this helper function instead of raw openat(2) to prevent security issues
+ * when a malicious client opens special files such as block device nodes.
+ * Symlink inodes are also rejected since symlinks must already have been
+ * traversed on the client side.
+ */
+static int lo_inode_open(struct lo_data *lo, struct lo_inode *inode,
+ int open_flags)
+{
+ g_autofree char *fd_str = g_strdup_printf("%d", inode->fd);
+ int fd;
+
+ if (!S_ISREG(inode->filetype) && !S_ISDIR(inode->filetype)) {
+ return -EBADF;
+ }
+
+ /*
+ * The file is a symlink so O_NOFOLLOW must be ignored. We checked earlier
+ * that the inode is not a special file but if an external process races
+ * with us then symlinks are traversed here. It is not possible to escape
+ * the shared directory since it is mounted as "/" though.
+ */
+ fd = openat(lo->proc_self_fd, fd_str, open_flags & ~O_NOFOLLOW);
+ if (fd < 0) {
+ return -errno;
+ }
+ return fd;
+}
+
static void lo_init(void *userdata, struct fuse_conn_info *conn)
{
struct lo_data *lo = (struct lo_data *)userdata;
@@ -684,9 +716,9 @@ static void lo_setattr(fuse_req_t req, fuse_ino_t ino, struct stat *attr,
if (fi) {
truncfd = fd;
} else {
- sprintf(procname, "%i", ifd);
- truncfd = openat(lo->proc_self_fd, procname, O_RDWR);
+ truncfd = lo_inode_open(lo, inode, O_RDWR);
if (truncfd < 0) {
+ errno = -truncfd;
goto out_err;
}
}
@@ -848,7 +880,7 @@ static int lo_do_lookup(fuse_req_t req, fuse_ino_t parent, const char *name,
struct lo_inode *dir = lo_inode(req, parent);
if (inodep) {
- *inodep = NULL;
+ *inodep = NULL; /* in case there is an error */
}
/*
@@ -1664,19 +1696,26 @@ static void update_open_flags(int writeback, int allow_direct_io,
}
}
+/*
+ * Open a regular file, set up an fd mapping, and fill out the struct
+ * fuse_file_info for it. If existing_fd is not negative, use that fd instead
+ * opening a new one. Takes ownership of existing_fd.
+ *
+ * Returns 0 on success or a positive errno.
+ */
static int lo_do_open(struct lo_data *lo, struct lo_inode *inode,
- struct fuse_file_info *fi)
+ int existing_fd, struct fuse_file_info *fi)
{
- char buf[64];
ssize_t fh;
- int fd;
+ int fd = existing_fd;
update_open_flags(lo->writeback, lo->allow_direct_io, fi);
- sprintf(buf, "%i", inode->fd);
- fd = openat(lo->proc_self_fd, buf, fi->flags & ~O_NOFOLLOW);
- if (fd == -1) {
- return errno;
+ if (fd < 0) {
+ fd = lo_inode_open(lo, inode, fi->flags);
+ if (fd < 0) {
+ return -fd;
+ }
}
pthread_mutex_lock(&lo->mutex);
@@ -1699,9 +1738,10 @@ static int lo_do_open(struct lo_data *lo, struct lo_inode *inode,
static void lo_create(fuse_req_t req, fuse_ino_t parent, const char *name,
mode_t mode, struct fuse_file_info *fi)
{
- int fd;
+ int fd = -1;
struct lo_data *lo = lo_data(req);
struct lo_inode *parent_inode;
+ struct lo_inode *inode = NULL;
struct fuse_entry_param e;
int err;
struct lo_cred old = {};
@@ -1727,36 +1767,38 @@ static void lo_create(fuse_req_t req, fuse_ino_t parent, const char *name,
update_open_flags(lo->writeback, lo->allow_direct_io, fi);
- fd = openat(parent_inode->fd, name, (fi->flags | O_CREAT) & ~O_NOFOLLOW,
- mode);
+ /* Try to create a new file but don't open existing files */
+ fd = openat(parent_inode->fd, name, fi->flags | O_CREAT | O_EXCL, mode);
err = fd == -1 ? errno : 0;
+
lo_restore_cred(&old);
- if (!err) {
- ssize_t fh;
-
- pthread_mutex_lock(&lo->mutex);
- fh = lo_add_fd_mapping(lo, fd);
- pthread_mutex_unlock(&lo->mutex);
- if (fh == -1) {
- close(fd);
- err = ENOMEM;
- goto out;
- }
+ /* Ignore the error if file exists and O_EXCL was not given */
+ if (err && (err != EEXIST || (fi->flags & O_EXCL))) {
+ goto out;
+ }
- fi->fh = fh;
- err = lo_do_lookup(req, parent, name, &e, NULL);
+ err = lo_do_lookup(req, parent, name, &e, &inode);
+ if (err) {
+ goto out;
}
- if (lo->cache == CACHE_NONE) {
- fi->direct_io = 1;
- } else if (lo->cache == CACHE_ALWAYS) {
- fi->keep_cache = 1;
+
+ err = lo_do_open(lo, inode, fd, fi);
+ fd = -1; /* lo_do_open() takes ownership of fd */
+ if (err) {
+ /* Undo lo_do_lookup() nlookup ref */
+ unref_inode_lolocked(lo, inode, 1);
}
out:
+ lo_inode_put(lo, &inode);
lo_inode_put(lo, &parent_inode);
if (err) {
+ if (fd >= 0) {
+ close(fd);
+ }
+
fuse_reply_err(req, err);
} else {
fuse_reply_create(req, &e, fi);
@@ -1770,7 +1812,6 @@ static struct lo_inode_plock *lookup_create_plock_ctx(struct lo_data *lo,
pid_t pid, int *err)
{
struct lo_inode_plock *plock;
- char procname[64];
int fd;
plock =
@@ -1787,12 +1828,10 @@ static struct lo_inode_plock *lookup_create_plock_ctx(struct lo_data *lo,
}
/* Open another instance of file which can be used for ofd locks. */
- sprintf(procname, "%i", inode->fd);
-
/* TODO: What if file is not writable? */
- fd = openat(lo->proc_self_fd, procname, O_RDWR);
- if (fd == -1) {
- *err = errno;
+ fd = lo_inode_open(lo, inode, O_RDWR);
+ if (fd < 0) {
+ *err = -fd;
free(plock);
return NULL;
}
@@ -1949,7 +1988,7 @@ static void lo_open(fuse_req_t req, fuse_ino_t ino, struct fuse_file_info *fi)
return;
}
- err = lo_do_open(lo, inode, fi);
+ err = lo_do_open(lo, inode, -1, fi);
lo_inode_put(lo, &inode);
if (err) {
fuse_reply_err(req, err);
@@ -2014,39 +2053,40 @@ static void lo_flush(fuse_req_t req, fuse_ino_t ino, struct fuse_file_info *fi)
static void lo_fsync(fuse_req_t req, fuse_ino_t ino, int datasync,
struct fuse_file_info *fi)
{
+ struct lo_inode *inode = lo_inode(req, ino);
+ struct lo_data *lo = lo_data(req);
int res;
int fd;
- char *buf;
fuse_log(FUSE_LOG_DEBUG, "lo_fsync(ino=%" PRIu64 ", fi=0x%p)\n", ino,
(void *)fi);
+ if (!inode) {
+ fuse_reply_err(req, EBADF);
+ return;
+ }
+
if (!fi) {
- struct lo_data *lo = lo_data(req);
-
- res = asprintf(&buf, "%i", lo_fd(req, ino));
- if (res == -1) {
- return (void)fuse_reply_err(req, errno);
- }
-
- fd = openat(lo->proc_self_fd, buf, O_RDWR);
- free(buf);
- if (fd == -1) {
- return (void)fuse_reply_err(req, errno);
+ fd = lo_inode_open(lo, inode, O_RDWR);
+ if (fd < 0) {
+ res = -fd;
+ goto out;
}
} else {
fd = lo_fi_fd(req, fi);
}
if (datasync) {
- res = fdatasync(fd);
+ res = fdatasync(fd) == -1 ? errno : 0;
} else {
- res = fsync(fd);
+ res = fsync(fd) == -1 ? errno : 0;
}
if (!fi) {
close(fd);
}
- fuse_reply_err(req, res == -1 ? errno : 0);
+out:
+ lo_inode_put(lo, &inode);
+ fuse_reply_err(req, res);
}
static void lo_read(fuse_req_t req, fuse_ino_t ino, size_t size, off_t offset,
--
2.29.2
^ permalink raw reply related [flat|nested] 17+ messages in thread
* [Virtio-fs] [PATCH v5 3/3] virtiofsd: prevent opening of special files (CVE-2020-35517)
@ 2021-02-04 15:02 ` Stefan Hajnoczi
0 siblings, 0 replies; 17+ messages in thread
From: Stefan Hajnoczi @ 2021-02-04 15:02 UTC (permalink / raw)
To: qemu-devel
Cc: Daniel Berrange, P J P, virtio-fs, Alex Xu, vgoyal, Laszlo Ersek
A well-behaved FUSE client does not attempt to open special files with
FUSE_OPEN because they are handled on the client side (e.g. device nodes
are handled by client-side device drivers).
The check to prevent virtiofsd from opening special files is missing in
a few cases, most notably FUSE_OPEN. A malicious client can cause
virtiofsd to open a device node, potentially allowing the guest to
escape. This can be exploited by a modified guest device driver. It is
not exploitable from guest userspace since the guest kernel will handle
special files inside the guest instead of sending FUSE requests.
This patch fixes this issue by introducing the lo_inode_open() function
to check the file type before opening it. This is a short-term solution
because it does not prevent a compromised virtiofsd process from opening
device nodes on the host.
Restructure lo_create() to try O_CREAT | O_EXCL first. Note that O_CREAT
| O_EXCL does not follow symlinks, so O_NOFOLLOW masking is not
necessary here. If the file exists and the user did not specify O_EXCL,
open it via lo_do_open().
Reported-by: Alex Xu <alex@alxu.ca>
Fixes: CVE-2020-35517
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Reviewed-by: Vivek Goyal <vgoyal@redhat.com>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
---
v4:
* Return -fd instead or -errno after lo_inode_open() in lo_do_open() [Greg]
* Use De Morgan's Law to simplify the boolean expression in lo_create() [Vivek]
* Add missing errno = -truncfd after lo_inode_open() call in lo_setattr
v3:
* Restructure lo_create() to handle externally-created files (we need
to allocate an inode for them) [Greg]
v3:
* Protect lo_create() [Greg]
v2:
* Add doc comment clarifying that symlinks are traversed client-side
[Daniel]
This issue was diagnosed on public IRC and is therefore already known
and not embargoed.
A stronger fix, and the long-term solution, is for users to mount the
shared directory and any sub-mounts with nodev, as well as nosuid and
noexec. Unfortunately virtiofsd cannot do this automatically because
bind mounts added by the user after virtiofsd has launched would not be
detected. I suggest the following:
1. Modify libvirt and Kata Containers to explicitly set these mount
options.
2. Then modify virtiofsd to check that the shared directory has the
necessary options at startup. Refuse to start if the options are
missing so that the user is aware of the security requirements.
As a bonus this also increases the likelihood that other host processes
besides virtiofsd will be protected by nosuid/noexec/nodev so that a
malicious guest cannot drop these files in place and then arrange for a
host process to come across them.
Additionally, user namespaces have been discussed. They seem like a
worthwhile addition as an unprivileged or privilege-separated mode
although there are limitations with respect to security xattrs and the
actual uid/gid stored on the host file system not corresponding to the
guest uid/gid.
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
---
tools/virtiofsd/passthrough_ll.c | 144 ++++++++++++++++++++-----------
1 file changed, 92 insertions(+), 52 deletions(-)
diff --git a/tools/virtiofsd/passthrough_ll.c b/tools/virtiofsd/passthrough_ll.c
index aa35fc6ba5..147b59338a 100644
--- a/tools/virtiofsd/passthrough_ll.c
+++ b/tools/virtiofsd/passthrough_ll.c
@@ -555,6 +555,38 @@ static int lo_fd(fuse_req_t req, fuse_ino_t ino)
return fd;
}
+/*
+ * Open a file descriptor for an inode. Returns -EBADF if the inode is not a
+ * regular file or a directory.
+ *
+ * Use this helper function instead of raw openat(2) to prevent security issues
+ * when a malicious client opens special files such as block device nodes.
+ * Symlink inodes are also rejected since symlinks must already have been
+ * traversed on the client side.
+ */
+static int lo_inode_open(struct lo_data *lo, struct lo_inode *inode,
+ int open_flags)
+{
+ g_autofree char *fd_str = g_strdup_printf("%d", inode->fd);
+ int fd;
+
+ if (!S_ISREG(inode->filetype) && !S_ISDIR(inode->filetype)) {
+ return -EBADF;
+ }
+
+ /*
+ * The file is a symlink so O_NOFOLLOW must be ignored. We checked earlier
+ * that the inode is not a special file but if an external process races
+ * with us then symlinks are traversed here. It is not possible to escape
+ * the shared directory since it is mounted as "/" though.
+ */
+ fd = openat(lo->proc_self_fd, fd_str, open_flags & ~O_NOFOLLOW);
+ if (fd < 0) {
+ return -errno;
+ }
+ return fd;
+}
+
static void lo_init(void *userdata, struct fuse_conn_info *conn)
{
struct lo_data *lo = (struct lo_data *)userdata;
@@ -684,9 +716,9 @@ static void lo_setattr(fuse_req_t req, fuse_ino_t ino, struct stat *attr,
if (fi) {
truncfd = fd;
} else {
- sprintf(procname, "%i", ifd);
- truncfd = openat(lo->proc_self_fd, procname, O_RDWR);
+ truncfd = lo_inode_open(lo, inode, O_RDWR);
if (truncfd < 0) {
+ errno = -truncfd;
goto out_err;
}
}
@@ -848,7 +880,7 @@ static int lo_do_lookup(fuse_req_t req, fuse_ino_t parent, const char *name,
struct lo_inode *dir = lo_inode(req, parent);
if (inodep) {
- *inodep = NULL;
+ *inodep = NULL; /* in case there is an error */
}
/*
@@ -1664,19 +1696,26 @@ static void update_open_flags(int writeback, int allow_direct_io,
}
}
+/*
+ * Open a regular file, set up an fd mapping, and fill out the struct
+ * fuse_file_info for it. If existing_fd is not negative, use that fd instead
+ * opening a new one. Takes ownership of existing_fd.
+ *
+ * Returns 0 on success or a positive errno.
+ */
static int lo_do_open(struct lo_data *lo, struct lo_inode *inode,
- struct fuse_file_info *fi)
+ int existing_fd, struct fuse_file_info *fi)
{
- char buf[64];
ssize_t fh;
- int fd;
+ int fd = existing_fd;
update_open_flags(lo->writeback, lo->allow_direct_io, fi);
- sprintf(buf, "%i", inode->fd);
- fd = openat(lo->proc_self_fd, buf, fi->flags & ~O_NOFOLLOW);
- if (fd == -1) {
- return errno;
+ if (fd < 0) {
+ fd = lo_inode_open(lo, inode, fi->flags);
+ if (fd < 0) {
+ return -fd;
+ }
}
pthread_mutex_lock(&lo->mutex);
@@ -1699,9 +1738,10 @@ static int lo_do_open(struct lo_data *lo, struct lo_inode *inode,
static void lo_create(fuse_req_t req, fuse_ino_t parent, const char *name,
mode_t mode, struct fuse_file_info *fi)
{
- int fd;
+ int fd = -1;
struct lo_data *lo = lo_data(req);
struct lo_inode *parent_inode;
+ struct lo_inode *inode = NULL;
struct fuse_entry_param e;
int err;
struct lo_cred old = {};
@@ -1727,36 +1767,38 @@ static void lo_create(fuse_req_t req, fuse_ino_t parent, const char *name,
update_open_flags(lo->writeback, lo->allow_direct_io, fi);
- fd = openat(parent_inode->fd, name, (fi->flags | O_CREAT) & ~O_NOFOLLOW,
- mode);
+ /* Try to create a new file but don't open existing files */
+ fd = openat(parent_inode->fd, name, fi->flags | O_CREAT | O_EXCL, mode);
err = fd == -1 ? errno : 0;
+
lo_restore_cred(&old);
- if (!err) {
- ssize_t fh;
-
- pthread_mutex_lock(&lo->mutex);
- fh = lo_add_fd_mapping(lo, fd);
- pthread_mutex_unlock(&lo->mutex);
- if (fh == -1) {
- close(fd);
- err = ENOMEM;
- goto out;
- }
+ /* Ignore the error if file exists and O_EXCL was not given */
+ if (err && (err != EEXIST || (fi->flags & O_EXCL))) {
+ goto out;
+ }
- fi->fh = fh;
- err = lo_do_lookup(req, parent, name, &e, NULL);
+ err = lo_do_lookup(req, parent, name, &e, &inode);
+ if (err) {
+ goto out;
}
- if (lo->cache == CACHE_NONE) {
- fi->direct_io = 1;
- } else if (lo->cache == CACHE_ALWAYS) {
- fi->keep_cache = 1;
+
+ err = lo_do_open(lo, inode, fd, fi);
+ fd = -1; /* lo_do_open() takes ownership of fd */
+ if (err) {
+ /* Undo lo_do_lookup() nlookup ref */
+ unref_inode_lolocked(lo, inode, 1);
}
out:
+ lo_inode_put(lo, &inode);
lo_inode_put(lo, &parent_inode);
if (err) {
+ if (fd >= 0) {
+ close(fd);
+ }
+
fuse_reply_err(req, err);
} else {
fuse_reply_create(req, &e, fi);
@@ -1770,7 +1812,6 @@ static struct lo_inode_plock *lookup_create_plock_ctx(struct lo_data *lo,
pid_t pid, int *err)
{
struct lo_inode_plock *plock;
- char procname[64];
int fd;
plock =
@@ -1787,12 +1828,10 @@ static struct lo_inode_plock *lookup_create_plock_ctx(struct lo_data *lo,
}
/* Open another instance of file which can be used for ofd locks. */
- sprintf(procname, "%i", inode->fd);
-
/* TODO: What if file is not writable? */
- fd = openat(lo->proc_self_fd, procname, O_RDWR);
- if (fd == -1) {
- *err = errno;
+ fd = lo_inode_open(lo, inode, O_RDWR);
+ if (fd < 0) {
+ *err = -fd;
free(plock);
return NULL;
}
@@ -1949,7 +1988,7 @@ static void lo_open(fuse_req_t req, fuse_ino_t ino, struct fuse_file_info *fi)
return;
}
- err = lo_do_open(lo, inode, fi);
+ err = lo_do_open(lo, inode, -1, fi);
lo_inode_put(lo, &inode);
if (err) {
fuse_reply_err(req, err);
@@ -2014,39 +2053,40 @@ static void lo_flush(fuse_req_t req, fuse_ino_t ino, struct fuse_file_info *fi)
static void lo_fsync(fuse_req_t req, fuse_ino_t ino, int datasync,
struct fuse_file_info *fi)
{
+ struct lo_inode *inode = lo_inode(req, ino);
+ struct lo_data *lo = lo_data(req);
int res;
int fd;
- char *buf;
fuse_log(FUSE_LOG_DEBUG, "lo_fsync(ino=%" PRIu64 ", fi=0x%p)\n", ino,
(void *)fi);
+ if (!inode) {
+ fuse_reply_err(req, EBADF);
+ return;
+ }
+
if (!fi) {
- struct lo_data *lo = lo_data(req);
-
- res = asprintf(&buf, "%i", lo_fd(req, ino));
- if (res == -1) {
- return (void)fuse_reply_err(req, errno);
- }
-
- fd = openat(lo->proc_self_fd, buf, O_RDWR);
- free(buf);
- if (fd == -1) {
- return (void)fuse_reply_err(req, errno);
+ fd = lo_inode_open(lo, inode, O_RDWR);
+ if (fd < 0) {
+ res = -fd;
+ goto out;
}
} else {
fd = lo_fi_fd(req, fi);
}
if (datasync) {
- res = fdatasync(fd);
+ res = fdatasync(fd) == -1 ? errno : 0;
} else {
- res = fsync(fd);
+ res = fsync(fd) == -1 ? errno : 0;
}
if (!fi) {
close(fd);
}
- fuse_reply_err(req, res == -1 ? errno : 0);
+out:
+ lo_inode_put(lo, &inode);
+ fuse_reply_err(req, res);
}
static void lo_read(fuse_req_t req, fuse_ino_t ino, size_t size, off_t offset,
--
2.29.2
^ permalink raw reply related [flat|nested] 17+ messages in thread
* Re: [PATCH v5 0/3] virtiofsd: prevent opening of special files (CVE-2020-35517)
2021-02-04 15:02 ` [Virtio-fs] " Stefan Hajnoczi
@ 2021-02-04 16:15 ` no-reply
-1 siblings, 0 replies; 17+ messages in thread
From: no-reply @ 2021-02-04 16:15 UTC (permalink / raw)
To: stefanha
Cc: mszeredi, berrange, slp, qemu-devel, ppandit, virtio-fs, groug,
alex, dgilbert, stefanha, lersek, vgoyal
Patchew URL: https://patchew.org/QEMU/20210204150208.367837-1-stefanha@redhat.com/
Hi,
This series seems to have some coding style problems. See output below for
more information:
Type: series
Message-id: 20210204150208.367837-1-stefanha@redhat.com
Subject: [PATCH v5 0/3] virtiofsd: prevent opening of special files (CVE-2020-35517)
=== TEST SCRIPT BEGIN ===
#!/bin/bash
git rev-parse base > /dev/null || exit 0
git config --local diff.renamelimit 0
git config --local diff.renames True
git config --local diff.algorithm histogram
./scripts/checkpatch.pl --mailback base..
=== TEST SCRIPT END ===
Updating 3c8cf5a9c21ff8782164d1def7f44bd888713384
From https://github.com/patchew-project/qemu
db754f8..1ba089f master -> master
- [tag update] patchew/20210204014509.882821-1-richard.henderson@linaro.org -> patchew/20210204014509.882821-1-richard.henderson@linaro.org
- [tag update] patchew/20210204124834.774401-1-berrange@redhat.com -> patchew/20210204124834.774401-1-berrange@redhat.com
* [new tag] patchew/20210204150208.367837-1-stefanha@redhat.com -> patchew/20210204150208.367837-1-stefanha@redhat.com
* [new tag] patchew/20210204153925.2030606-1-Jason@zx2c4.com -> patchew/20210204153925.2030606-1-Jason@zx2c4.com
Switched to a new branch 'test'
b5bb803 virtiofsd: prevent opening of special files (CVE-2020-35517)
0d88a79 virtiofsd: optionally return inode pointer from lo_do_lookup()
be6aa23 virtiofsd: extract lo_do_open() from lo_open()
=== OUTPUT BEGIN ===
1/3 Checking commit be6aa2319875 (virtiofsd: extract lo_do_open() from lo_open())
ERROR: return of an errno should typically be -ve (return -ENOMEM)
#70: FILE: tools/virtiofsd/passthrough_ll.c:1674:
+ return ENOMEM;
total: 1 errors, 0 warnings, 114 lines checked
Patch 1/3 has style problems, please review. If any of these errors
are false positives report them to the maintainer, see
CHECKPATCH in MAINTAINERS.
2/3 Checking commit 0d88a7925a83 (virtiofsd: optionally return inode pointer from lo_do_lookup())
3/3 Checking commit b5bb8039eb3c (virtiofsd: prevent opening of special files (CVE-2020-35517))
=== OUTPUT END ===
Test command exited with code: 1
The full log is available at
http://patchew.org/logs/20210204150208.367837-1-stefanha@redhat.com/testing.checkpatch/?type=message.
---
Email generated automatically by Patchew [https://patchew.org/].
Please send your feedback to patchew-devel@redhat.com
^ permalink raw reply [flat|nested] 17+ messages in thread
* Re: [Virtio-fs] [PATCH v5 0/3] virtiofsd: prevent opening of special files (CVE-2020-35517)
@ 2021-02-04 16:15 ` no-reply
0 siblings, 0 replies; 17+ messages in thread
From: no-reply @ 2021-02-04 16:15 UTC (permalink / raw)
To: stefanha; +Cc: berrange, qemu-devel, ppandit, virtio-fs, alex, lersek, vgoyal
Patchew URL: https://patchew.org/QEMU/20210204150208.367837-1-stefanha@redhat.com/
Hi,
This series seems to have some coding style problems. See output below for
more information:
Type: series
Message-id: 20210204150208.367837-1-stefanha@redhat.com
Subject: [PATCH v5 0/3] virtiofsd: prevent opening of special files (CVE-2020-35517)
=== TEST SCRIPT BEGIN ===
#!/bin/bash
git rev-parse base > /dev/null || exit 0
git config --local diff.renamelimit 0
git config --local diff.renames True
git config --local diff.algorithm histogram
./scripts/checkpatch.pl --mailback base..
=== TEST SCRIPT END ===
Updating 3c8cf5a9c21ff8782164d1def7f44bd888713384
>From https://github.com/patchew-project/qemu
db754f8..1ba089f master -> master
- [tag update] patchew/20210204014509.882821-1-richard.henderson@linaro.org -> patchew/20210204014509.882821-1-richard.henderson@linaro.org
- [tag update] patchew/20210204124834.774401-1-berrange@redhat.com -> patchew/20210204124834.774401-1-berrange@redhat.com
* [new tag] patchew/20210204150208.367837-1-stefanha@redhat.com -> patchew/20210204150208.367837-1-stefanha@redhat.com
* [new tag] patchew/20210204153925.2030606-1-Jason@zx2c4.com -> patchew/20210204153925.2030606-1-Jason@zx2c4.com
Switched to a new branch 'test'
b5bb803 virtiofsd: prevent opening of special files (CVE-2020-35517)
0d88a79 virtiofsd: optionally return inode pointer from lo_do_lookup()
be6aa23 virtiofsd: extract lo_do_open() from lo_open()
=== OUTPUT BEGIN ===
1/3 Checking commit be6aa2319875 (virtiofsd: extract lo_do_open() from lo_open())
ERROR: return of an errno should typically be -ve (return -ENOMEM)
#70: FILE: tools/virtiofsd/passthrough_ll.c:1674:
+ return ENOMEM;
total: 1 errors, 0 warnings, 114 lines checked
Patch 1/3 has style problems, please review. If any of these errors
are false positives report them to the maintainer, see
CHECKPATCH in MAINTAINERS.
2/3 Checking commit 0d88a7925a83 (virtiofsd: optionally return inode pointer from lo_do_lookup())
3/3 Checking commit b5bb8039eb3c (virtiofsd: prevent opening of special files (CVE-2020-35517))
=== OUTPUT END ===
Test command exited with code: 1
The full log is available at
http://patchew.org/logs/20210204150208.367837-1-stefanha@redhat.com/testing.checkpatch/?type=message.
---
Email generated automatically by Patchew [https://patchew.org/].
Please send your feedback to patchew-devel@redhat.com
^ permalink raw reply [flat|nested] 17+ messages in thread
* Re: [PATCH v5 1/3] virtiofsd: extract lo_do_open() from lo_open()
2021-02-04 15:02 ` [Virtio-fs] " Stefan Hajnoczi
@ 2021-02-04 17:08 ` Greg Kurz
-1 siblings, 0 replies; 17+ messages in thread
From: Greg Kurz @ 2021-02-04 17:08 UTC (permalink / raw)
To: Stefan Hajnoczi
Cc: mszeredi, Daniel Berrange, slp, qemu-devel, P J P, virtio-fs,
vgoyal, Laszlo Ersek, Dr. David Alan Gilbert
On Thu, 4 Feb 2021 15:02:06 +0000
Stefan Hajnoczi <stefanha@redhat.com> wrote:
> Both lo_open() and lo_create() have similar code to open a file. Extract
> a common lo_do_open() function from lo_open() that will be used by
> lo_create() in a later commit.
>
> Since lo_do_open() does not otherwise need fuse_req_t req, convert
> lo_add_fd_mapping() to use struct lo_data *lo instead.
>
> Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
> ---
Reviewed-by: Greg Kurz <groug@kaod.org>
> v4:
> * Return positive errno if openat(2) fails in lo_do_open() [Greg]
> ---
> tools/virtiofsd/passthrough_ll.c | 73 ++++++++++++++++++++------------
> 1 file changed, 46 insertions(+), 27 deletions(-)
>
> diff --git a/tools/virtiofsd/passthrough_ll.c b/tools/virtiofsd/passthrough_ll.c
> index 5fb36d9407..f14fa5124d 100644
> --- a/tools/virtiofsd/passthrough_ll.c
> +++ b/tools/virtiofsd/passthrough_ll.c
> @@ -459,17 +459,17 @@ static void lo_map_remove(struct lo_map *map, size_t key)
> }
>
> /* Assumes lo->mutex is held */
> -static ssize_t lo_add_fd_mapping(fuse_req_t req, int fd)
> +static ssize_t lo_add_fd_mapping(struct lo_data *lo, int fd)
> {
> struct lo_map_elem *elem;
>
> - elem = lo_map_alloc_elem(&lo_data(req)->fd_map);
> + elem = lo_map_alloc_elem(&lo->fd_map);
> if (!elem) {
> return -1;
> }
>
> elem->fd = fd;
> - return elem - lo_data(req)->fd_map.elems;
> + return elem - lo->fd_map.elems;
> }
>
> /* Assumes lo->mutex is held */
> @@ -1651,6 +1651,38 @@ static void update_open_flags(int writeback, int allow_direct_io,
> }
> }
>
> +static int lo_do_open(struct lo_data *lo, struct lo_inode *inode,
> + struct fuse_file_info *fi)
> +{
> + char buf[64];
> + ssize_t fh;
> + int fd;
> +
> + update_open_flags(lo->writeback, lo->allow_direct_io, fi);
> +
> + sprintf(buf, "%i", inode->fd);
> + fd = openat(lo->proc_self_fd, buf, fi->flags & ~O_NOFOLLOW);
> + if (fd == -1) {
> + return errno;
> + }
> +
> + pthread_mutex_lock(&lo->mutex);
> + fh = lo_add_fd_mapping(lo, fd);
> + pthread_mutex_unlock(&lo->mutex);
> + if (fh == -1) {
> + close(fd);
> + return ENOMEM;
> + }
> +
> + fi->fh = fh;
> + if (lo->cache == CACHE_NONE) {
> + fi->direct_io = 1;
> + } else if (lo->cache == CACHE_ALWAYS) {
> + fi->keep_cache = 1;
> + }
> + return 0;
> +}
> +
> static void lo_create(fuse_req_t req, fuse_ino_t parent, const char *name,
> mode_t mode, struct fuse_file_info *fi)
> {
> @@ -1691,7 +1723,7 @@ static void lo_create(fuse_req_t req, fuse_ino_t parent, const char *name,
> ssize_t fh;
>
> pthread_mutex_lock(&lo->mutex);
> - fh = lo_add_fd_mapping(req, fd);
> + fh = lo_add_fd_mapping(lo, fd);
> pthread_mutex_unlock(&lo->mutex);
> if (fh == -1) {
> close(fd);
> @@ -1892,38 +1924,25 @@ static void lo_fsyncdir(fuse_req_t req, fuse_ino_t ino, int datasync,
>
> static void lo_open(fuse_req_t req, fuse_ino_t ino, struct fuse_file_info *fi)
> {
> - int fd;
> - ssize_t fh;
> - char buf[64];
> struct lo_data *lo = lo_data(req);
> + struct lo_inode *inode = lo_inode(req, ino);
> + int err;
>
> fuse_log(FUSE_LOG_DEBUG, "lo_open(ino=%" PRIu64 ", flags=%d)\n", ino,
> fi->flags);
>
> - update_open_flags(lo->writeback, lo->allow_direct_io, fi);
> -
> - sprintf(buf, "%i", lo_fd(req, ino));
> - fd = openat(lo->proc_self_fd, buf, fi->flags & ~O_NOFOLLOW);
> - if (fd == -1) {
> - return (void)fuse_reply_err(req, errno);
> - }
> -
> - pthread_mutex_lock(&lo->mutex);
> - fh = lo_add_fd_mapping(req, fd);
> - pthread_mutex_unlock(&lo->mutex);
> - if (fh == -1) {
> - close(fd);
> - fuse_reply_err(req, ENOMEM);
> + if (!inode) {
> + fuse_reply_err(req, EBADF);
> return;
> }
>
> - fi->fh = fh;
> - if (lo->cache == CACHE_NONE) {
> - fi->direct_io = 1;
> - } else if (lo->cache == CACHE_ALWAYS) {
> - fi->keep_cache = 1;
> + err = lo_do_open(lo, inode, fi);
> + lo_inode_put(lo, &inode);
> + if (err) {
> + fuse_reply_err(req, err);
> + } else {
> + fuse_reply_open(req, fi);
> }
> - fuse_reply_open(req, fi);
> }
>
> static void lo_release(fuse_req_t req, fuse_ino_t ino,
^ permalink raw reply [flat|nested] 17+ messages in thread
* Re: [Virtio-fs] [PATCH v5 1/3] virtiofsd: extract lo_do_open() from lo_open()
@ 2021-02-04 17:08 ` Greg Kurz
0 siblings, 0 replies; 17+ messages in thread
From: Greg Kurz @ 2021-02-04 17:08 UTC (permalink / raw)
To: Stefan Hajnoczi
Cc: Daniel Berrange, qemu-devel, P J P, virtio-fs, vgoyal, Laszlo Ersek
On Thu, 4 Feb 2021 15:02:06 +0000
Stefan Hajnoczi <stefanha@redhat.com> wrote:
> Both lo_open() and lo_create() have similar code to open a file. Extract
> a common lo_do_open() function from lo_open() that will be used by
> lo_create() in a later commit.
>
> Since lo_do_open() does not otherwise need fuse_req_t req, convert
> lo_add_fd_mapping() to use struct lo_data *lo instead.
>
> Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
> ---
Reviewed-by: Greg Kurz <groug@kaod.org>
> v4:
> * Return positive errno if openat(2) fails in lo_do_open() [Greg]
> ---
> tools/virtiofsd/passthrough_ll.c | 73 ++++++++++++++++++++------------
> 1 file changed, 46 insertions(+), 27 deletions(-)
>
> diff --git a/tools/virtiofsd/passthrough_ll.c b/tools/virtiofsd/passthrough_ll.c
> index 5fb36d9407..f14fa5124d 100644
> --- a/tools/virtiofsd/passthrough_ll.c
> +++ b/tools/virtiofsd/passthrough_ll.c
> @@ -459,17 +459,17 @@ static void lo_map_remove(struct lo_map *map, size_t key)
> }
>
> /* Assumes lo->mutex is held */
> -static ssize_t lo_add_fd_mapping(fuse_req_t req, int fd)
> +static ssize_t lo_add_fd_mapping(struct lo_data *lo, int fd)
> {
> struct lo_map_elem *elem;
>
> - elem = lo_map_alloc_elem(&lo_data(req)->fd_map);
> + elem = lo_map_alloc_elem(&lo->fd_map);
> if (!elem) {
> return -1;
> }
>
> elem->fd = fd;
> - return elem - lo_data(req)->fd_map.elems;
> + return elem - lo->fd_map.elems;
> }
>
> /* Assumes lo->mutex is held */
> @@ -1651,6 +1651,38 @@ static void update_open_flags(int writeback, int allow_direct_io,
> }
> }
>
> +static int lo_do_open(struct lo_data *lo, struct lo_inode *inode,
> + struct fuse_file_info *fi)
> +{
> + char buf[64];
> + ssize_t fh;
> + int fd;
> +
> + update_open_flags(lo->writeback, lo->allow_direct_io, fi);
> +
> + sprintf(buf, "%i", inode->fd);
> + fd = openat(lo->proc_self_fd, buf, fi->flags & ~O_NOFOLLOW);
> + if (fd == -1) {
> + return errno;
> + }
> +
> + pthread_mutex_lock(&lo->mutex);
> + fh = lo_add_fd_mapping(lo, fd);
> + pthread_mutex_unlock(&lo->mutex);
> + if (fh == -1) {
> + close(fd);
> + return ENOMEM;
> + }
> +
> + fi->fh = fh;
> + if (lo->cache == CACHE_NONE) {
> + fi->direct_io = 1;
> + } else if (lo->cache == CACHE_ALWAYS) {
> + fi->keep_cache = 1;
> + }
> + return 0;
> +}
> +
> static void lo_create(fuse_req_t req, fuse_ino_t parent, const char *name,
> mode_t mode, struct fuse_file_info *fi)
> {
> @@ -1691,7 +1723,7 @@ static void lo_create(fuse_req_t req, fuse_ino_t parent, const char *name,
> ssize_t fh;
>
> pthread_mutex_lock(&lo->mutex);
> - fh = lo_add_fd_mapping(req, fd);
> + fh = lo_add_fd_mapping(lo, fd);
> pthread_mutex_unlock(&lo->mutex);
> if (fh == -1) {
> close(fd);
> @@ -1892,38 +1924,25 @@ static void lo_fsyncdir(fuse_req_t req, fuse_ino_t ino, int datasync,
>
> static void lo_open(fuse_req_t req, fuse_ino_t ino, struct fuse_file_info *fi)
> {
> - int fd;
> - ssize_t fh;
> - char buf[64];
> struct lo_data *lo = lo_data(req);
> + struct lo_inode *inode = lo_inode(req, ino);
> + int err;
>
> fuse_log(FUSE_LOG_DEBUG, "lo_open(ino=%" PRIu64 ", flags=%d)\n", ino,
> fi->flags);
>
> - update_open_flags(lo->writeback, lo->allow_direct_io, fi);
> -
> - sprintf(buf, "%i", lo_fd(req, ino));
> - fd = openat(lo->proc_self_fd, buf, fi->flags & ~O_NOFOLLOW);
> - if (fd == -1) {
> - return (void)fuse_reply_err(req, errno);
> - }
> -
> - pthread_mutex_lock(&lo->mutex);
> - fh = lo_add_fd_mapping(req, fd);
> - pthread_mutex_unlock(&lo->mutex);
> - if (fh == -1) {
> - close(fd);
> - fuse_reply_err(req, ENOMEM);
> + if (!inode) {
> + fuse_reply_err(req, EBADF);
> return;
> }
>
> - fi->fh = fh;
> - if (lo->cache == CACHE_NONE) {
> - fi->direct_io = 1;
> - } else if (lo->cache == CACHE_ALWAYS) {
> - fi->keep_cache = 1;
> + err = lo_do_open(lo, inode, fi);
> + lo_inode_put(lo, &inode);
> + if (err) {
> + fuse_reply_err(req, err);
> + } else {
> + fuse_reply_open(req, fi);
> }
> - fuse_reply_open(req, fi);
> }
>
> static void lo_release(fuse_req_t req, fuse_ino_t ino,
^ permalink raw reply [flat|nested] 17+ messages in thread
* Re: [PATCH v5 3/3] virtiofsd: prevent opening of special files (CVE-2020-35517)
2021-02-04 15:02 ` [Virtio-fs] " Stefan Hajnoczi
(?)
@ 2021-02-04 17:12 ` Greg Kurz
-1 siblings, 0 replies; 17+ messages in thread
From: Greg Kurz @ 2021-02-04 17:12 UTC (permalink / raw)
To: Stefan Hajnoczi
Cc: mszeredi, Daniel Berrange, slp, qemu-devel, P J P, virtio-fs,
Alex Xu, vgoyal, Laszlo Ersek, Dr. David Alan Gilbert
On Thu, 4 Feb 2021 15:02:08 +0000
Stefan Hajnoczi <stefanha@redhat.com> wrote:
> A well-behaved FUSE client does not attempt to open special files with
> FUSE_OPEN because they are handled on the client side (e.g. device nodes
> are handled by client-side device drivers).
>
> The check to prevent virtiofsd from opening special files is missing in
> a few cases, most notably FUSE_OPEN. A malicious client can cause
> virtiofsd to open a device node, potentially allowing the guest to
> escape. This can be exploited by a modified guest device driver. It is
> not exploitable from guest userspace since the guest kernel will handle
> special files inside the guest instead of sending FUSE requests.
>
> This patch fixes this issue by introducing the lo_inode_open() function
> to check the file type before opening it. This is a short-term solution
> because it does not prevent a compromised virtiofsd process from opening
> device nodes on the host.
>
> Restructure lo_create() to try O_CREAT | O_EXCL first. Note that O_CREAT
> | O_EXCL does not follow symlinks, so O_NOFOLLOW masking is not
> necessary here. If the file exists and the user did not specify O_EXCL,
> open it via lo_do_open().
>
> Reported-by: Alex Xu <alex@alxu.ca>
> Fixes: CVE-2020-35517
> Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
> Reviewed-by: Vivek Goyal <vgoyal@redhat.com>
> Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
> ---
> v4:
> * Return -fd instead or -errno after lo_inode_open() in lo_do_open() [Greg]
> * Use De Morgan's Law to simplify the boolean expression in lo_create() [Vivek]
> * Add missing errno = -truncfd after lo_inode_open() call in lo_setattr
Good catch !
Reviewed-by: Greg Kurz <groug@kaod.org>
> v3:
> * Restructure lo_create() to handle externally-created files (we need
> to allocate an inode for them) [Greg]
> v3:
> * Protect lo_create() [Greg]
> v2:
> * Add doc comment clarifying that symlinks are traversed client-side
> [Daniel]
>
> This issue was diagnosed on public IRC and is therefore already known
> and not embargoed.
>
> A stronger fix, and the long-term solution, is for users to mount the
> shared directory and any sub-mounts with nodev, as well as nosuid and
> noexec. Unfortunately virtiofsd cannot do this automatically because
> bind mounts added by the user after virtiofsd has launched would not be
> detected. I suggest the following:
>
> 1. Modify libvirt and Kata Containers to explicitly set these mount
> options.
> 2. Then modify virtiofsd to check that the shared directory has the
> necessary options at startup. Refuse to start if the options are
> missing so that the user is aware of the security requirements.
>
> As a bonus this also increases the likelihood that other host processes
> besides virtiofsd will be protected by nosuid/noexec/nodev so that a
> malicious guest cannot drop these files in place and then arrange for a
> host process to come across them.
>
> Additionally, user namespaces have been discussed. They seem like a
> worthwhile addition as an unprivileged or privilege-separated mode
> although there are limitations with respect to security xattrs and the
> actual uid/gid stored on the host file system not corresponding to the
> guest uid/gid.
>
> Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
> ---
> tools/virtiofsd/passthrough_ll.c | 144 ++++++++++++++++++++-----------
> 1 file changed, 92 insertions(+), 52 deletions(-)
>
> diff --git a/tools/virtiofsd/passthrough_ll.c b/tools/virtiofsd/passthrough_ll.c
> index aa35fc6ba5..147b59338a 100644
> --- a/tools/virtiofsd/passthrough_ll.c
> +++ b/tools/virtiofsd/passthrough_ll.c
> @@ -555,6 +555,38 @@ static int lo_fd(fuse_req_t req, fuse_ino_t ino)
> return fd;
> }
>
> +/*
> + * Open a file descriptor for an inode. Returns -EBADF if the inode is not a
> + * regular file or a directory.
> + *
> + * Use this helper function instead of raw openat(2) to prevent security issues
> + * when a malicious client opens special files such as block device nodes.
> + * Symlink inodes are also rejected since symlinks must already have been
> + * traversed on the client side.
> + */
> +static int lo_inode_open(struct lo_data *lo, struct lo_inode *inode,
> + int open_flags)
> +{
> + g_autofree char *fd_str = g_strdup_printf("%d", inode->fd);
> + int fd;
> +
> + if (!S_ISREG(inode->filetype) && !S_ISDIR(inode->filetype)) {
> + return -EBADF;
> + }
> +
> + /*
> + * The file is a symlink so O_NOFOLLOW must be ignored. We checked earlier
> + * that the inode is not a special file but if an external process races
> + * with us then symlinks are traversed here. It is not possible to escape
> + * the shared directory since it is mounted as "/" though.
> + */
> + fd = openat(lo->proc_self_fd, fd_str, open_flags & ~O_NOFOLLOW);
> + if (fd < 0) {
> + return -errno;
> + }
> + return fd;
> +}
> +
> static void lo_init(void *userdata, struct fuse_conn_info *conn)
> {
> struct lo_data *lo = (struct lo_data *)userdata;
> @@ -684,9 +716,9 @@ static void lo_setattr(fuse_req_t req, fuse_ino_t ino, struct stat *attr,
> if (fi) {
> truncfd = fd;
> } else {
> - sprintf(procname, "%i", ifd);
> - truncfd = openat(lo->proc_self_fd, procname, O_RDWR);
> + truncfd = lo_inode_open(lo, inode, O_RDWR);
> if (truncfd < 0) {
> + errno = -truncfd;
> goto out_err;
> }
> }
> @@ -848,7 +880,7 @@ static int lo_do_lookup(fuse_req_t req, fuse_ino_t parent, const char *name,
> struct lo_inode *dir = lo_inode(req, parent);
>
> if (inodep) {
> - *inodep = NULL;
> + *inodep = NULL; /* in case there is an error */
> }
>
> /*
> @@ -1664,19 +1696,26 @@ static void update_open_flags(int writeback, int allow_direct_io,
> }
> }
>
> +/*
> + * Open a regular file, set up an fd mapping, and fill out the struct
> + * fuse_file_info for it. If existing_fd is not negative, use that fd instead
> + * opening a new one. Takes ownership of existing_fd.
> + *
> + * Returns 0 on success or a positive errno.
> + */
> static int lo_do_open(struct lo_data *lo, struct lo_inode *inode,
> - struct fuse_file_info *fi)
> + int existing_fd, struct fuse_file_info *fi)
> {
> - char buf[64];
> ssize_t fh;
> - int fd;
> + int fd = existing_fd;
>
> update_open_flags(lo->writeback, lo->allow_direct_io, fi);
>
> - sprintf(buf, "%i", inode->fd);
> - fd = openat(lo->proc_self_fd, buf, fi->flags & ~O_NOFOLLOW);
> - if (fd == -1) {
> - return errno;
> + if (fd < 0) {
> + fd = lo_inode_open(lo, inode, fi->flags);
> + if (fd < 0) {
> + return -fd;
> + }
> }
>
> pthread_mutex_lock(&lo->mutex);
> @@ -1699,9 +1738,10 @@ static int lo_do_open(struct lo_data *lo, struct lo_inode *inode,
> static void lo_create(fuse_req_t req, fuse_ino_t parent, const char *name,
> mode_t mode, struct fuse_file_info *fi)
> {
> - int fd;
> + int fd = -1;
> struct lo_data *lo = lo_data(req);
> struct lo_inode *parent_inode;
> + struct lo_inode *inode = NULL;
> struct fuse_entry_param e;
> int err;
> struct lo_cred old = {};
> @@ -1727,36 +1767,38 @@ static void lo_create(fuse_req_t req, fuse_ino_t parent, const char *name,
>
> update_open_flags(lo->writeback, lo->allow_direct_io, fi);
>
> - fd = openat(parent_inode->fd, name, (fi->flags | O_CREAT) & ~O_NOFOLLOW,
> - mode);
> + /* Try to create a new file but don't open existing files */
> + fd = openat(parent_inode->fd, name, fi->flags | O_CREAT | O_EXCL, mode);
> err = fd == -1 ? errno : 0;
> +
> lo_restore_cred(&old);
>
> - if (!err) {
> - ssize_t fh;
> -
> - pthread_mutex_lock(&lo->mutex);
> - fh = lo_add_fd_mapping(lo, fd);
> - pthread_mutex_unlock(&lo->mutex);
> - if (fh == -1) {
> - close(fd);
> - err = ENOMEM;
> - goto out;
> - }
> + /* Ignore the error if file exists and O_EXCL was not given */
> + if (err && (err != EEXIST || (fi->flags & O_EXCL))) {
> + goto out;
> + }
>
> - fi->fh = fh;
> - err = lo_do_lookup(req, parent, name, &e, NULL);
> + err = lo_do_lookup(req, parent, name, &e, &inode);
> + if (err) {
> + goto out;
> }
> - if (lo->cache == CACHE_NONE) {
> - fi->direct_io = 1;
> - } else if (lo->cache == CACHE_ALWAYS) {
> - fi->keep_cache = 1;
> +
> + err = lo_do_open(lo, inode, fd, fi);
> + fd = -1; /* lo_do_open() takes ownership of fd */
> + if (err) {
> + /* Undo lo_do_lookup() nlookup ref */
> + unref_inode_lolocked(lo, inode, 1);
> }
>
> out:
> + lo_inode_put(lo, &inode);
> lo_inode_put(lo, &parent_inode);
>
> if (err) {
> + if (fd >= 0) {
> + close(fd);
> + }
> +
> fuse_reply_err(req, err);
> } else {
> fuse_reply_create(req, &e, fi);
> @@ -1770,7 +1812,6 @@ static struct lo_inode_plock *lookup_create_plock_ctx(struct lo_data *lo,
> pid_t pid, int *err)
> {
> struct lo_inode_plock *plock;
> - char procname[64];
> int fd;
>
> plock =
> @@ -1787,12 +1828,10 @@ static struct lo_inode_plock *lookup_create_plock_ctx(struct lo_data *lo,
> }
>
> /* Open another instance of file which can be used for ofd locks. */
> - sprintf(procname, "%i", inode->fd);
> -
> /* TODO: What if file is not writable? */
> - fd = openat(lo->proc_self_fd, procname, O_RDWR);
> - if (fd == -1) {
> - *err = errno;
> + fd = lo_inode_open(lo, inode, O_RDWR);
> + if (fd < 0) {
> + *err = -fd;
> free(plock);
> return NULL;
> }
> @@ -1949,7 +1988,7 @@ static void lo_open(fuse_req_t req, fuse_ino_t ino, struct fuse_file_info *fi)
> return;
> }
>
> - err = lo_do_open(lo, inode, fi);
> + err = lo_do_open(lo, inode, -1, fi);
> lo_inode_put(lo, &inode);
> if (err) {
> fuse_reply_err(req, err);
> @@ -2014,39 +2053,40 @@ static void lo_flush(fuse_req_t req, fuse_ino_t ino, struct fuse_file_info *fi)
> static void lo_fsync(fuse_req_t req, fuse_ino_t ino, int datasync,
> struct fuse_file_info *fi)
> {
> + struct lo_inode *inode = lo_inode(req, ino);
> + struct lo_data *lo = lo_data(req);
> int res;
> int fd;
> - char *buf;
>
> fuse_log(FUSE_LOG_DEBUG, "lo_fsync(ino=%" PRIu64 ", fi=0x%p)\n", ino,
> (void *)fi);
>
> + if (!inode) {
> + fuse_reply_err(req, EBADF);
> + return;
> + }
> +
> if (!fi) {
> - struct lo_data *lo = lo_data(req);
> -
> - res = asprintf(&buf, "%i", lo_fd(req, ino));
> - if (res == -1) {
> - return (void)fuse_reply_err(req, errno);
> - }
> -
> - fd = openat(lo->proc_self_fd, buf, O_RDWR);
> - free(buf);
> - if (fd == -1) {
> - return (void)fuse_reply_err(req, errno);
> + fd = lo_inode_open(lo, inode, O_RDWR);
> + if (fd < 0) {
> + res = -fd;
> + goto out;
> }
> } else {
> fd = lo_fi_fd(req, fi);
> }
>
> if (datasync) {
> - res = fdatasync(fd);
> + res = fdatasync(fd) == -1 ? errno : 0;
> } else {
> - res = fsync(fd);
> + res = fsync(fd) == -1 ? errno : 0;
> }
> if (!fi) {
> close(fd);
> }
> - fuse_reply_err(req, res == -1 ? errno : 0);
> +out:
> + lo_inode_put(lo, &inode);
> + fuse_reply_err(req, res);
> }
>
> static void lo_read(fuse_req_t req, fuse_ino_t ino, size_t size, off_t offset,
^ permalink raw reply [flat|nested] 17+ messages in thread
* Re: [PATCH v5 0/3] virtiofsd: prevent opening of special files (CVE-2020-35517)
2021-02-04 16:15 ` [Virtio-fs] " no-reply
@ 2021-02-04 17:35 ` Dr. David Alan Gilbert
-1 siblings, 0 replies; 17+ messages in thread
From: Dr. David Alan Gilbert @ 2021-02-04 17:35 UTC (permalink / raw)
To: qemu-devel
Cc: mszeredi, berrange, slp, groug, ppandit, virtio-fs, alex,
stefanha, lersek, vgoyal
* no-reply@patchew.org (no-reply@patchew.org) wrote:
> Patchew URL: https://patchew.org/QEMU/20210204150208.367837-1-stefanha@redhat.com/
>
>
>
> Hi,
>
> This series seems to have some coding style problems. See output below for
> more information:
>
> Type: series
> Message-id: 20210204150208.367837-1-stefanha@redhat.com
> Subject: [PATCH v5 0/3] virtiofsd: prevent opening of special files (CVE-2020-35517)
>
> === TEST SCRIPT BEGIN ===
> #!/bin/bash
> git rev-parse base > /dev/null || exit 0
> git config --local diff.renamelimit 0
> git config --local diff.renames True
> git config --local diff.algorithm histogram
> ./scripts/checkpatch.pl --mailback base..
> === TEST SCRIPT END ===
>
> Updating 3c8cf5a9c21ff8782164d1def7f44bd888713384
> From https://github.com/patchew-project/qemu
> db754f8..1ba089f master -> master
> - [tag update] patchew/20210204014509.882821-1-richard.henderson@linaro.org -> patchew/20210204014509.882821-1-richard.henderson@linaro.org
> - [tag update] patchew/20210204124834.774401-1-berrange@redhat.com -> patchew/20210204124834.774401-1-berrange@redhat.com
> * [new tag] patchew/20210204150208.367837-1-stefanha@redhat.com -> patchew/20210204150208.367837-1-stefanha@redhat.com
> * [new tag] patchew/20210204153925.2030606-1-Jason@zx2c4.com -> patchew/20210204153925.2030606-1-Jason@zx2c4.com
> Switched to a new branch 'test'
> b5bb803 virtiofsd: prevent opening of special files (CVE-2020-35517)
> 0d88a79 virtiofsd: optionally return inode pointer from lo_do_lookup()
> be6aa23 virtiofsd: extract lo_do_open() from lo_open()
>
> === OUTPUT BEGIN ===
> 1/3 Checking commit be6aa2319875 (virtiofsd: extract lo_do_open() from lo_open())
> ERROR: return of an errno should typically be -ve (return -ENOMEM)
> #70: FILE: tools/virtiofsd/passthrough_ll.c:1674:
> + return ENOMEM;
This is intended
> total: 1 errors, 0 warnings, 114 lines checked
>
> Patch 1/3 has style problems, please review. If any of these errors
> are false positives report them to the maintainer, see
> CHECKPATCH in MAINTAINERS.
>
> 2/3 Checking commit 0d88a7925a83 (virtiofsd: optionally return inode pointer from lo_do_lookup())
> 3/3 Checking commit b5bb8039eb3c (virtiofsd: prevent opening of special files (CVE-2020-35517))
> === OUTPUT END ===
>
> Test command exited with code: 1
>
>
> The full log is available at
> http://patchew.org/logs/20210204150208.367837-1-stefanha@redhat.com/testing.checkpatch/?type=message.
> ---
> Email generated automatically by Patchew [https://patchew.org/].
> Please send your feedback to patchew-devel@redhat.com
--
Dr. David Alan Gilbert / dgilbert@redhat.com / Manchester, UK
^ permalink raw reply [flat|nested] 17+ messages in thread
* Re: [Virtio-fs] [PATCH v5 0/3] virtiofsd: prevent opening of special files (CVE-2020-35517)
@ 2021-02-04 17:35 ` Dr. David Alan Gilbert
0 siblings, 0 replies; 17+ messages in thread
From: Dr. David Alan Gilbert @ 2021-02-04 17:35 UTC (permalink / raw)
To: qemu-devel; +Cc: berrange, ppandit, virtio-fs, alex, lersek, vgoyal
* no-reply@patchew.org (no-reply@patchew.org) wrote:
> Patchew URL: https://patchew.org/QEMU/20210204150208.367837-1-stefanha@redhat.com/
>
>
>
> Hi,
>
> This series seems to have some coding style problems. See output below for
> more information:
>
> Type: series
> Message-id: 20210204150208.367837-1-stefanha@redhat.com
> Subject: [PATCH v5 0/3] virtiofsd: prevent opening of special files (CVE-2020-35517)
>
> === TEST SCRIPT BEGIN ===
> #!/bin/bash
> git rev-parse base > /dev/null || exit 0
> git config --local diff.renamelimit 0
> git config --local diff.renames True
> git config --local diff.algorithm histogram
> ./scripts/checkpatch.pl --mailback base..
> === TEST SCRIPT END ===
>
> Updating 3c8cf5a9c21ff8782164d1def7f44bd888713384
> From https://github.com/patchew-project/qemu
> db754f8..1ba089f master -> master
> - [tag update] patchew/20210204014509.882821-1-richard.henderson@linaro.org -> patchew/20210204014509.882821-1-richard.henderson@linaro.org
> - [tag update] patchew/20210204124834.774401-1-berrange@redhat.com -> patchew/20210204124834.774401-1-berrange@redhat.com
> * [new tag] patchew/20210204150208.367837-1-stefanha@redhat.com -> patchew/20210204150208.367837-1-stefanha@redhat.com
> * [new tag] patchew/20210204153925.2030606-1-Jason@zx2c4.com -> patchew/20210204153925.2030606-1-Jason@zx2c4.com
> Switched to a new branch 'test'
> b5bb803 virtiofsd: prevent opening of special files (CVE-2020-35517)
> 0d88a79 virtiofsd: optionally return inode pointer from lo_do_lookup()
> be6aa23 virtiofsd: extract lo_do_open() from lo_open()
>
> === OUTPUT BEGIN ===
> 1/3 Checking commit be6aa2319875 (virtiofsd: extract lo_do_open() from lo_open())
> ERROR: return of an errno should typically be -ve (return -ENOMEM)
> #70: FILE: tools/virtiofsd/passthrough_ll.c:1674:
> + return ENOMEM;
This is intended
> total: 1 errors, 0 warnings, 114 lines checked
>
> Patch 1/3 has style problems, please review. If any of these errors
> are false positives report them to the maintainer, see
> CHECKPATCH in MAINTAINERS.
>
> 2/3 Checking commit 0d88a7925a83 (virtiofsd: optionally return inode pointer from lo_do_lookup())
> 3/3 Checking commit b5bb8039eb3c (virtiofsd: prevent opening of special files (CVE-2020-35517))
> === OUTPUT END ===
>
> Test command exited with code: 1
>
>
> The full log is available at
> http://patchew.org/logs/20210204150208.367837-1-stefanha@redhat.com/testing.checkpatch/?type=message.
> ---
> Email generated automatically by Patchew [https://patchew.org/].
> Please send your feedback to patchew-devel@redhat.com
--
Dr. David Alan Gilbert / dgilbert@redhat.com / Manchester, UK
^ permalink raw reply [flat|nested] 17+ messages in thread
* Re: [PATCH v5 0/3] virtiofsd: prevent opening of special files (CVE-2020-35517)
2021-02-04 15:02 ` [Virtio-fs] " Stefan Hajnoczi
@ 2021-02-04 18:14 ` Dr. David Alan Gilbert
-1 siblings, 0 replies; 17+ messages in thread
From: Dr. David Alan Gilbert @ 2021-02-04 18:14 UTC (permalink / raw)
To: Stefan Hajnoczi
Cc: mszeredi, Daniel Berrange, slp, qemu-devel, P J P, virtio-fs,
Greg Kurz, Alex Xu, Laszlo Ersek, vgoyal
* Stefan Hajnoczi (stefanha@redhat.com) wrote:
> v4:
> * Patch 1: Return positive errno if openat(2) fails in lo_do_open() [Greg]
> * Patch 3: Return -fd instead or -errno after lo_inode_open() in lo_do_open() [Greg]
> * Patch 3: Use De Morgan's Law to simplify the boolean expression in lo_create() [Vivek]
> * Patch 3: Add missing errno = -truncfd after lo_inode_open() call in lo_setattr
> v3:
> * Restructure lo_create() to handle externally-created files (we need
> to allocate an inode for them) [Greg]
> * Patch 1 & 2 refactor the code so that Patch 3 can implement the CVE fix
> v3:
> * Protect lo_create() [Greg]
> v2:
> * Add doc comment clarifying that symlinks are traversed client-side
> [Daniel]
>
> A well-behaved FUSE client does not attempt to open special files with
> FUSE_OPEN because they are handled on the client side (e.g. device nodes
> are handled by client-side device drivers).
>
> The check to prevent virtiofsd from opening special files is missing in
> a few cases, most notably FUSE_OPEN. A malicious client can cause
> virtiofsd to open a device node, potentially allowing the guest to
> escape. This can be exploited by a modified guest device driver. It is
> not exploitable from guest userspace since the guest kernel will handle
> special files inside the guest instead of sending FUSE requests.
>
> This patch series fixes this issue by introducing the lo_inode_open() function
> to check the file type before opening it. This is a short-term solution because
> it does not prevent a compromised virtiofsd process from opening device nodes
> on the host.
>
> This issue was diagnosed on public IRC and is therefore already known
> and not embargoed.
>
> Reported-by: Alex Xu <alex@alxu.ca>
> Fixes: CVE-2020-35517
Queued
> Stefan Hajnoczi (3):
> virtiofsd: extract lo_do_open() from lo_open()
> virtiofsd: optionally return inode pointer from lo_do_lookup()
> virtiofsd: prevent opening of special files (CVE-2020-35517)
>
> tools/virtiofsd/passthrough_ll.c | 224 ++++++++++++++++++++-----------
> 1 file changed, 148 insertions(+), 76 deletions(-)
>
> --
> 2.29.2
>
--
Dr. David Alan Gilbert / dgilbert@redhat.com / Manchester, UK
^ permalink raw reply [flat|nested] 17+ messages in thread
* Re: [Virtio-fs] [PATCH v5 0/3] virtiofsd: prevent opening of special files (CVE-2020-35517)
@ 2021-02-04 18:14 ` Dr. David Alan Gilbert
0 siblings, 0 replies; 17+ messages in thread
From: Dr. David Alan Gilbert @ 2021-02-04 18:14 UTC (permalink / raw)
To: Stefan Hajnoczi
Cc: Daniel Berrange, qemu-devel, P J P, virtio-fs, Alex Xu,
Laszlo Ersek, vgoyal
* Stefan Hajnoczi (stefanha@redhat.com) wrote:
> v4:
> * Patch 1: Return positive errno if openat(2) fails in lo_do_open() [Greg]
> * Patch 3: Return -fd instead or -errno after lo_inode_open() in lo_do_open() [Greg]
> * Patch 3: Use De Morgan's Law to simplify the boolean expression in lo_create() [Vivek]
> * Patch 3: Add missing errno = -truncfd after lo_inode_open() call in lo_setattr
> v3:
> * Restructure lo_create() to handle externally-created files (we need
> to allocate an inode for them) [Greg]
> * Patch 1 & 2 refactor the code so that Patch 3 can implement the CVE fix
> v3:
> * Protect lo_create() [Greg]
> v2:
> * Add doc comment clarifying that symlinks are traversed client-side
> [Daniel]
>
> A well-behaved FUSE client does not attempt to open special files with
> FUSE_OPEN because they are handled on the client side (e.g. device nodes
> are handled by client-side device drivers).
>
> The check to prevent virtiofsd from opening special files is missing in
> a few cases, most notably FUSE_OPEN. A malicious client can cause
> virtiofsd to open a device node, potentially allowing the guest to
> escape. This can be exploited by a modified guest device driver. It is
> not exploitable from guest userspace since the guest kernel will handle
> special files inside the guest instead of sending FUSE requests.
>
> This patch series fixes this issue by introducing the lo_inode_open() function
> to check the file type before opening it. This is a short-term solution because
> it does not prevent a compromised virtiofsd process from opening device nodes
> on the host.
>
> This issue was diagnosed on public IRC and is therefore already known
> and not embargoed.
>
> Reported-by: Alex Xu <alex@alxu.ca>
> Fixes: CVE-2020-35517
Queued
> Stefan Hajnoczi (3):
> virtiofsd: extract lo_do_open() from lo_open()
> virtiofsd: optionally return inode pointer from lo_do_lookup()
> virtiofsd: prevent opening of special files (CVE-2020-35517)
>
> tools/virtiofsd/passthrough_ll.c | 224 ++++++++++++++++++++-----------
> 1 file changed, 148 insertions(+), 76 deletions(-)
>
> --
> 2.29.2
>
--
Dr. David Alan Gilbert / dgilbert@redhat.com / Manchester, UK
^ permalink raw reply [flat|nested] 17+ messages in thread
end of thread, other threads:[~2021-02-04 18:53 UTC | newest]
Thread overview: 17+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2021-02-04 15:02 [PATCH v5 0/3] virtiofsd: prevent opening of special files (CVE-2020-35517) Stefan Hajnoczi
2021-02-04 15:02 ` [Virtio-fs] " Stefan Hajnoczi
2021-02-04 15:02 ` [PATCH v5 1/3] virtiofsd: extract lo_do_open() from lo_open() Stefan Hajnoczi
2021-02-04 15:02 ` [Virtio-fs] " Stefan Hajnoczi
2021-02-04 17:08 ` Greg Kurz
2021-02-04 17:08 ` [Virtio-fs] " Greg Kurz
2021-02-04 15:02 ` [PATCH v5 2/3] virtiofsd: optionally return inode pointer from lo_do_lookup() Stefan Hajnoczi
2021-02-04 15:02 ` [Virtio-fs] " Stefan Hajnoczi
2021-02-04 15:02 ` [PATCH v5 3/3] virtiofsd: prevent opening of special files (CVE-2020-35517) Stefan Hajnoczi
2021-02-04 15:02 ` [Virtio-fs] " Stefan Hajnoczi
2021-02-04 17:12 ` Greg Kurz
2021-02-04 16:15 ` [PATCH v5 0/3] " no-reply
2021-02-04 16:15 ` [Virtio-fs] " no-reply
2021-02-04 17:35 ` Dr. David Alan Gilbert
2021-02-04 17:35 ` [Virtio-fs] " Dr. David Alan Gilbert
2021-02-04 18:14 ` Dr. David Alan Gilbert
2021-02-04 18:14 ` [Virtio-fs] " Dr. David Alan Gilbert
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.