Provide an option "-o security_label/no_security_label" to enable/disable
security label functionality. By default these are turned off.
If enabled, server will indicate to client that it is capable of handling
one security label during file creation. Typically this is expected to
be a SELinux label. File server will set this label on the file. It will
try to set it atomically wherever possible. But its not possible in
all the cases.
Signed-off-by: Vivek Goyal <vgoyal@redhat.com>
Message-Id: <20220208204813.682906-11-vgoyal@redhat.com>
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
If guest and host policies can't work with each other, then guest security
context (selinux label) needs to be set into an xattr. Say remap guest
security.selinux xattr to trusted.virtiofs.security.selinux.
That means setting "fscreate" is not going to help as that's ony useful
for security.selinux xattr on host.
So we need another method which is atomic. Use O_TMPFILE to create new
file, set xattr and then linkat() to proper place.
But this works only for regular files. So dir, symlinks will continue
to be non-atomic.
Also if host filesystem does not support O_TMPFILE, we fallback to
non-atomic behavior.
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Signed-off-by: Vivek Goyal <vgoyal@redhat.com>
Message-Id: <20220208204813.682906-10-vgoyal@redhat.com>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
This patch adds support for creating new file with security context
as sent by client. It basically takes three paths.
- If no security context enabled, then it continues to create files without
security context.
- If security context is enabled and but security.selinux has not been
remapped, then it uses /proc/thread-self/attr/fscreate knob to set
security context and then create the file. This will make sure that
newly created file gets the security context as set in "fscreate" and
this is atomic w.r.t file creation.
This is useful and host and guest SELinux policies don't conflict and
can work with each other. In that case, guest security.selinux xattr
is not remapped and it is passthrough as "security.selinux" xattr
on host.
- If security context is enabled but security.selinux xattr has been
remapped to something else, then it first creates the file and then
uses setxattr() to set the remapped xattr with the security context.
This is a non-atomic operation w.r.t file creation.
This mode will be most versatile and allow host and guest to have their
own separate SELinux xattrs and have their own separate SELinux policies.
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Signed-off-by: Vivek Goyal <vgoyal@redhat.com>
Message-Id: <20220208204813.682906-9-vgoyal@redhat.com>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Soon we will be able to create and also set security context on the file
atomically using /proc/self/task/tid/attr/fscreate knob. If this knob
is available on the system, first set the knob with the desired context
and then create the file. It will be created with the context set in
fscreate. This works basically for SELinux and its per thread.
This patch just introduces the helper functions. Subsequent patches will
make use of these helpers.
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Signed-off-by: Vivek Goyal <vgoyal@redhat.com>
Message-Id: <20220208204813.682906-8-vgoyal@redhat.com>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
dgilbert: Manually merged gettid syscall number fixup from Vivek
Move core file creation bits in a separate function. Soon this is going
to get more complex as file creation need to set security context also.
And there will be multiple modes of file creation in next patch.
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Signed-off-by: Vivek Goyal <vgoyal@redhat.com>
Message-Id: <20220208204813.682906-7-vgoyal@redhat.com>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
With the current implementation, blocking flock can lead to
deadlock. Thus, it's better to return EOPNOTSUPP if a user attempts
to perform a blocking flock request.
Signed-off-by: Sebastian Hasler <sebastian.hasler@stuvus.uni-stuttgart.de>
Message-Id: <20220113153249.710216-1-sebastian.hasler@stuvus.uni-stuttgart.de>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Reviewed-by: Vivek Goyal <vgoyal@redhat.com>
Reviewed-by: Greg Kurz <groug@kaod.org>
At the start, drop membership of all supplementary groups. This is
not required.
If we have membership of "root" supplementary group and when we switch
uid/gid using setresuid/setsgid, we still retain membership of existing
supplemntary groups. And that can allow some operations which are not
normally allowed.
For example, if root in guest creates a dir as follows.
$ mkdir -m 03777 test_dir
This sets SGID on dir as well as allows unprivileged users to write into
this dir.
And now as unprivileged user open file as follows.
$ su test
$ fd = open("test_dir/priviledge_id", O_RDWR|O_CREAT|O_EXCL, 02755);
This will create SGID set executable in test_dir/.
And that's a problem because now an unpriviliged user can execute it,
get egid=0 and get access to resources owned by "root" group. This is
privilege escalation.
Fixes: https://bugzilla.redhat.com/show_bug.cgi?id=2044863
Fixes: CVE-2022-0358
Reported-by: JIETAO XIAO <shawtao1125@gmail.com>
Suggested-by: Miklos Szeredi <mszeredi@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Signed-off-by: Vivek Goyal <vgoyal@redhat.com>
Message-Id: <YfBGoriS38eBQrAb@redhat.com>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
dgilbert: Fixed missing {}'s style nit
Right now for xattr remapping, we support types of "prefix", "ok" or "bad".
Type "bad" returns -EPERM on setxattr and hides xattr in listxattr. For
getxattr, mapping code returns -EPERM but getxattr code converts it to -ENODATA.
I need a new semantics where if an xattr is unsupported, then
getxattr()/setxattr() return -ENOTSUP and listxattr() should hide the xattr.
This is needed to simulate that security.selinux is not supported by
virtiofs filesystem and in that case client falls back to some default
label specified by policy.
So add a new type "unsupported" which returns -ENOTSUP on getxattr() and
setxattr() and hides xattrs in listxattr().
For example, one can use following mapping rule to not support
security.selinux xattr and allow others.
"-o xattrmap=/unsupported/all/security.selinux/security.selinux//ok/all///"
Suggested-by: "Dr. David Alan Gilbert" <dgilbert@redhat.com>
Signed-off-by: Vivek Goyal <vgoyal@redhat.com>
Message-Id: <YUt9qbmgAfCFfg5t@redhat.com>
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
fuse has an option FUSE_POSIX_ACL which needs to be opted in by fuse
server to enable posix acls. As of now we are not opting in for this,
so posix acls are disabled on virtiofs by default.
Add virtiofsd option "-o posix_acl/no_posix_acl" to let users enable/disable
posix acl support. By default it is disabled as of now due to performance
concerns with cache=none.
Currently even if file server has not opted in for FUSE_POSIX_ACL, user can
still query acl and set acl, and system.posix_acl_access and
system.posix_acl_default xattrs show up listxattr response.
Miklos said this is confusing. So he said lets block and filter
system.posix_acl_access and system.posix_acl_default xattrs in
getxattr/setxattr/listxattr if user has explicitly disabled
posix acls using -o no_posix_acl.
As of now continuing to keeping the existing behavior if user did not
specify any option to disable acl support due to concerns about backward
compatibility.
Signed-off-by: Vivek Goyal <vgoyal@redhat.com>
Message-Id: <20210622150852.1507204-8-vgoyal@redhat.com>
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
When posix access acls are set on a file, it can lead to adjusting file
permissions (mode) as well. If caller does not have CAP_FSETID and it
also does not have membership of owner group, this will lead to clearing
SGID bit in mode.
Current fuse code is written in such a way that it expects file server
to take care of chaning file mode (permission), if there is a need.
Right now, host kernel does not clear SGID bit because virtiofsd is
running as root and has CAP_FSETID. For host kernel to clear SGID,
virtiofsd need to switch to gid of caller in guest and also drop
CAP_FSETID (if caller did not have it to begin with).
If SGID needs to be cleared, client will set the flag
FUSE_SETXATTR_ACL_KILL_SGID in setxattr request. In that case server
should kill sgid.
Currently just switch to uid/gid of the caller and drop CAP_FSETID
and that should do it.
This should fix the xfstest generic/375 test case.
We don't have to switch uid for this to work. That could be one optimization
that pass a parameter to lo_change_cred() to only switch gid and not uid.
Also this will not work whenever (if ever) we support idmapped mounts. In
that case it is possible that uid/gid in request are 0/0 but still we
need to clear SGID. So we will have to pick a non-root sgid and switch
to that instead. That's an TODO item for future when idmapped mount
support is introduced.
This patch only adds the capability to switch creds and drop FSETID
when acl xattr is set. This does not take affect yet. It can take
affect when next patch adds the capability to enable posix_acl.
Reported-by: Luis Henriques <lhenriques@suse.de>
Signed-off-by: Vivek Goyal <vgoyal@redhat.com>
Message-Id: <20210622150852.1507204-7-vgoyal@redhat.com>
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
When parent directory has default acl and a file is created in that
directory, then umask is ignored and final file permissions are
determined using default acl instead. (man 2 umask).
Currently, fuse applies the umask and sends modified mode in create
request accordingly. fuse server can set FUSE_DONT_MASK and tell
fuse client to not apply umask and fuse server will take care of
it as needed.
With posix acls enabled, requirement will be that we want umask
to determine final file mode if parent directory does not have
default acl.
So if posix acls are enabled, opt in for FUSE_DONT_MASK. virtiofsd
will set umask of the thread doing file creation. And host kernel
should use that umask if parent directory does not have default
acls, otherwise umask does not take affect.
Miklos mentioned that we already call unshare(CLONE_FS) for
every thread. That means umask has now become property of per
thread and it should be ok to manipulate it in file creation path.
This patch only adds capability to change umask and restore it. It
does not enable it yet. Next few patches will add capability to enable it
based on if user enabled posix_acl or not.
This should fix fstest generic/099.
Reported-by: Luis Henriques <lhenriques@suse.de>
Signed-off-by: Vivek Goyal <vgoyal@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Message-Id: <20210622150852.1507204-6-vgoyal@redhat.com>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Add the bits to enable support for setxattr_ext if fuse offers it. Do not
enable it by default yet. Let passthrough_ll opt-in. Enabling it by deafult
kind of automatically means that you are taking responsibility of clearing
SGID if ACL is set.
Signed-off-by: Vivek Goyal <vgoyal@redhat.com>
Message-Id: <20210622150852.1507204-4-vgoyal@redhat.com>
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Fixed up double def in fuse_common.h
getxattr/setxattr/removexattr/listxattr operations handle regualar
and non-regular files differently. For the case of non-regular files
we do fchdir(/proc/self/fd) and the xattr operation and then revert
back to original working directory. After this we are saving errno
and that's buggy because fchdir() will overwrite the errno.
FCHDIR_NOFAIL(lo->proc_self_fd);
ret = getxattr(procname, name, value, size);
FCHDIR_NOFAIL(lo->root.fd);
if (ret == -1)
saverr = errno
In above example, if getxattr() failed, we will still return 0 to caller
as errno must have been written by FCHDIR_NOFAIL(lo->root.fd) call.
Fix all such instances and capture "errno" early and save in "saverr"
variable.
Signed-off-by: Vivek Goyal <vgoyal@redhat.com>
Message-Id: <20210622150852.1507204-3-vgoyal@redhat.com>
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Reviewed-by: Greg Kurz <groug@kaod.org>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
The GDateTime APIs provided by GLib avoid portability pitfalls, such
as some platforms where 'struct timeval.tv_sec' field is still 'long'
instead of 'time_t'. When combined with automatic cleanup, GDateTime
often results in simpler code too.
Localtime is changed to UTC to avoid the need to grant extra seccomp
permissions for GLib's access of the timezone database.
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Message-Id: <20210611164319.67762-1-berrange@redhat.com>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Changed the allocations of some local variables to GLib's allocation
functions, such as g_try_malloc0(), and annotated those variables
as g_autofree. Subsequently, I was able to remove the calls to free().
Signed-off-by: Mahmoud Mandour <ma.mandourr@gmail.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-Id: <20210420154643.58439-7-ma.mandourr@gmail.com>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Replaced (re)allocation of lo_map_elem structs from realloc() to
GLib's g_try_realloc_n() and replaced the respective free() call
with a g_free().
Signed-off-by: Mahmoud Mandour <ma.mandourr@gmail.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-Id: <20210420154643.58439-5-ma.mandourr@gmail.com>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Fixup some fuse_log printf args for 32bit compatibility.
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Message-Id: <20210428110100.27757-2-dgilbert@redhat.com>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
When -o xattrmap is used, it will not work unless xattr is enabled.
This patch enables xattr when -o xattrmap is used.
Signed-off-by: Carlos Venegas <jose.carlos.venegas.munoz@intel.com>
Message-Id: <20210414201207.3612432-2-jose.carlos.venegas.munoz@intel.com>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Reviewed-by: Connor Kuehl <ckuehl@redhat.com>
It is bad practice to put an expression with a side-effect in
assert() because the side-effect won't happen if the code is
compiled with -DNDEBUG.
Use an intermediate variable. Consolidate this in an macro to
have proper line numbers when the assertion is hit.
virtiofsd: ../../tools/virtiofsd/passthrough_ll.c:2797: lo_getxattr:
Assertion `fchdir_res == 0' failed.
Aborted
2796 /* fchdir should not fail here */
=>2797 FCHDIR_NOFAIL(lo->proc_self_fd);
2798 ret = getxattr(procname, name, value, size);
2799 FCHDIR_NOFAIL(lo->root.fd);
Fixes: bdfd667883 ("virtiofsd: Fix xattr operations")
Cc: misono.tomohiro@jp.fujitsu.com
Signed-off-by: Greg Kurz <groug@kaod.org>
Message-Id: <20210409100627.451573-1-groug@kaod.org>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
My security fix for the security.capability remap has a silly early
segfault in a simple case where there is an xattrmapping but it doesn't
remap the security.capability.
Fixes: e586edcb41 ("virtiofs: drop remapped security.capability xattr as needed")
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Message-Id: <20210401145845.78445-1-dgilbert@redhat.com>
Reviewed-by: Connor Kuehl <ckuehl@redhat.com>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Both currently only return 0 or 1.
Signed-off-by: Greg Kurz <groug@kaod.org>
Message-Id: <20210312141003.819108-3-groug@kaod.org>
Reviewed-by: Connor Kuehl <ckuehl@redhat.com>
Reviewed-by: Vivek Goyal <vgoyal@redhat.com>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
When passed an empty filename, lookup_name() returns the inode of
the parent directory, unless the parent is the root in which case
the st_dev doesn't match and lo_find() returns NULL. This is
because lookup_name() passes AT_EMPTY_PATH down to fstatat() or
statx().
This behavior doesn't quite make sense because users of lookup_name()
then pass the name to unlinkat(), renameat() or renameat2(), all of
which will always fail on empty names.
Drop AT_EMPTY_PATH from the flags in lookup_name() so that it has
the consistent behavior of "returning an existing child inode or
NULL" for all directories.
Signed-off-by: Greg Kurz <groug@kaod.org>
Message-Id: <20210312141003.819108-2-groug@kaod.org>
Reviewed-by: Connor Kuehl <ckuehl@redhat.com>
Reviewed-by: Vivek Goyal <vgoyal@redhat.com>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
POSIX.1-2017 clearly stipulates that empty filenames aren't
allowed ([1] and [2]). Since virtiofsd is supposed to mirror
the host file system hierarchy and the host can be assumed to
be linux, we don't really expect clients to pass requests with
an empty path in it. If they do so anyway, this would eventually
cause an error when trying to create/lookup the actual inode
on the underlying POSIX filesystem. But this could still confuse
some code that wouldn't be ready to cope with this.
Filter out empty names coming from the client at the top level,
so that the rest doesn't have to care about it. This is done
everywhere we already call is_safe_path_component(), but
in a separate helper since the usual error for empty path
names is ENOENT instead of EINVAL.
[1] https://pubs.opengroup.org/onlinepubs/9699919799/basedefs/V1_chap03.html#tag_03_170
[2] https://pubs.opengroup.org/onlinepubs/9699919799/basedefs/V1_chap04.html#tag_04_13
Signed-off-by: Greg Kurz <groug@kaod.org>
Message-Id: <20210312141003.819108-4-groug@kaod.org>
Reviewed-by: Connor Kuehl <ckuehl@redhat.com>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Option "-V" currently displays the fuse protocol version virtiofsd is
using. For example, I see this.
$ ./virtiofsd -V
"using FUSE kernel interface version 7.33"
People also want to know software version of virtiofsd so that they can
figure out if a certain fix is part of currently running virtiofsd or
not. Eric Ernst ran into this issue.
David Gilbert thinks that it probably is best that we simply carry the
qemu version and display that information given we are part of qemu
tree.
So this patch enhances version information and also adds qemu version
and copyright info. Not sure if copyright information is supposed
to be displayed along with version info. Given qemu-storage-daemon
and other utilities are doing it, so I continued with same pattern.
This is how now output looks like.
$ ./virtiofsd -V
virtiofsd version 5.2.50 (v5.2.0-2357-gcbcf09872a-dirty)
Copyright (c) 2003-2020 Fabrice Bellard and the QEMU Project developers
using FUSE kernel interface version 7.33
Reported-by: Eric Ernst <eric.g.ernst@gmail.com>
Signed-off-by: Vivek Goyal <vgoyal@redhat.com>
Message-Id: <20210303195339.GB3793@redhat.com>
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Reviewed-by: Sergio Lopez <slp@redhat.com>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
On Linux, the 'security.capability' xattr holds a set of
capabilities that can change when an executable is run, giving
a limited form of privilege escalation to those programs that
the writer of the file deemed worthy.
Any write causes the 'security.capability' xattr to be dropped,
stopping anyone from gaining privilege by modifying a blessed
file.
Fuse relies on the daemon to do this dropping, and in turn the
daemon relies on the host kernel to drop the xattr for it. However,
with the addition of -o xattrmap, the xattr that the guest
stores its capabilities in is now not the same as the one that
the host kernel automatically clears.
Where the mapping changes 'security.capability', explicitly clear
the remapped name to preserve the same behaviour.
This bug is assigned CVE-2021-20263.
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Reviewed-by: Vivek Goyal <vgoyal@redhat.com>
This patch adds basic support for FUSE_HANDLE_KILLPRIV_V2. virtiofsd
can enable/disable this by specifying option "-o killpriv_v2/no_killpriv_v2".
By default this is enabled as long as client supports it
Enabling this option helps with performance in write path. Without this
option, currently every write is first preceeded with a getxattr() operation
to find out if security.capability is set. (Write is supposed to clear
security.capability). With this option enabled, server is signing up for
clearing security.capability on every WRITE and also clearing suid/sgid
subject to certain rules. This gets rid of extra getxattr() call for every
WRITE and improves performance. This is true when virtiofsd is run with
option -o xattr.
What does enabling FUSE_HANDLE_KILLPRIV_V2 mean for file server implementation.
It needs to adhere to following rules. Thanks to Miklos for this summary.
- clear "security.capability" on write, truncate and chown unconditionally
- clear suid/sgid in case of following. Note, sgid is cleared only if
group executable bit is set.
o setattr has FATTR_SIZE and FATTR_KILL_SUIDGID set.
o setattr has FATTR_UID or FATTR_GID
o open has O_TRUNC and FUSE_OPEN_KILL_SUIDGID
o create has O_TRUNC and FUSE_OPEN_KILL_SUIDGID flag set.
o write has FUSE_WRITE_KILL_SUIDGID
>From Linux VFS client perspective, here are the requirements.
- caps are always cleared on chown/write/truncate
- suid is always cleared on chown, while for truncate/write it is cleared
only if caller does not have CAP_FSETID.
- sgid is always cleared on chown, while for truncate/write it is cleared
only if caller does not have CAP_FSETID as well as file has group execute
permission.
virtiofsd implementation has not changed much to adhere to above ruls. And
reason being that current assumption is that we are running on Linux
and on top of filesystems like ext4/xfs which already follow above rules.
On write, truncate, chown, seucurity.capability is cleared. And virtiofsd
drops CAP_FSETID if need be and that will lead to clearing of suid/sgid.
But if virtiofsd is running on top a filesystem which breaks above assumptions,
then it will have to take extra actions to emulate above. That's a TODO
for later when need arises.
Note: create normally is supposed to be called only when file does not
exist. So generally there should not be any question of clearing
setuid/setgid. But it is possible that after client checks that
file is not present, some other client creates file on server
and this race can trigger sending FUSE_CREATE. In that case, if
O_TRUNC is set, we should clear suid/sgid if FUSE_OPEN_KILL_SUIDGID
is also set.
v3:
- Resolved conflicts due to lo_inode_open() changes.
- Moved capability code in lo_do_open() so that both lo_open() and
lo_create() can benefit from common code.
- Dropped changes to kernel headers as these are part of qemu already.
Signed-off-by: Vivek Goyal <vgoyal@redhat.com>
Acked-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Message-Id: <20210208224024.43555-3-vgoyal@redhat.com>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Change error code handling slightly in lo_setattr(). Right now we seem
to jump to out_err and assume that "errno" is valid and use that to
send reply.
But if caller has to do some other operations before jumping to out_err,
then it does the dance of first saving errno to saverr and the restore
errno before jumping to out_err. This makes it more confusing.
I am about to make more changes where caller will have to do some
work after error before jumping to out_err. I found it easier to
change the convention a bit. That is caller saves error in "saverr"
before jumping to out_err. And out_err uses "saverr" to send error
back and does not rely on "errno" having actual error.
v3: Resolved conflicts in lo_setattr() due to lo_inode_open() changes.
Signed-off-by: Vivek Goyal <vgoyal@redhat.com>
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Message-Id: <20210208224024.43555-2-vgoyal@redhat.com>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Follow the inclusive terminology from the "Conscious Language in your
Open Source Projects" guidelines [*] and replace the words "whitelist"
appropriately.
[*] https://github.com/conscious-lang/conscious-lang-docs/blob/main/faq.md
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-Id: <20210205171817.2108907-3-philmd@redhat.com>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
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>
Reviewed-by: Greg Kurz <groug@kaod.org>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-Id: <20210204150208.367837-4-stefanha@redhat.com>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
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>
Message-Id: <20210204150208.367837-3-stefanha@redhat.com>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
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>
Message-Id: <20210204150208.367837-2-stefanha@redhat.com>
Reviewed-by: Greg Kurz <groug@kaod.org>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Miklos confirms it's *only* the FUSE_FORGET request that the client can
use for decrementing "lo_inode.nlookup".
Cc: "Dr. David Alan Gilbert" <dgilbert@redhat.com>
Cc: Miklos Szeredi <mszeredi@redhat.com>
Cc: Stefan Hajnoczi <stefanha@redhat.com>
Fixes: 1222f01555
Signed-off-by: Laszlo Ersek <lersek@redhat.com>
Message-Id: <20201208073936.8629-1-lersek@redhat.com>
Reviewed-by: Vivek Goyal <vgoyal@redhat.com>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Currently lo_flush() is written in such a way that it expects to receive
a FLUSH requests on a regular file (and not directories). For example,
we call lo_fi_fd() which searches lo->fd_map. If we open directories
using opendir(), we keep don't keep track of these in lo->fd_map instead
we keep them in lo->dir_map. So we expect lo_flush() to be called on
regular files only.
Even linux fuse client calls FLUSH only for regular files and not
directories. So put a check for filetype and return EBADF if
lo_flush() is called on a non-regular file.
Reported-by: Laszlo Ersek <lersek@redhat.com>
Signed-off-by: Vivek Goyal <vgoyal@redhat.com>
Message-Id: <20201211142544.GB3285@redhat.com>
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
If remote posix locks are not enabled (lo->posix_lock == false), then disable
code paths taken to initialize inode->posix_lock hash table and corresponding
destruction and search etc.
lo_getlk() and lo_setlk() have been modified to return ENOSYS if daemon
does not support posix lock but client still sends a lock/unlock request.
Signed-off-by: Vivek Goyal <vgoyal@redhat.com>
Message-Id: <20201207183021.22752-3-vgoyal@redhat.com>
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
We setup per inode hash table ->posix_lock to support remote posix locks.
But we forgot to initialize this table for root inode.
Laszlo managed to trigger an issue where he sent a FUSE_FLUSH request for
root inode and lo_flush() found inode with inode->posix_lock NULL and
accessing this table crashed virtiofsd.
May be we can get rid of initializing this hash table for directory
objects completely. But that optimization is for another day.
Reported-by: Laszlo Ersek <lersek@redhat.com>
Signed-off-by: Vivek Goyal <vgoyal@redhat.com>
Message-Id: <20201207195539.GB3107@redhat.com>
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
The current timestamp format doesn't help me visually notice small jumps
in time ("small" as defined on human scale, such as a few seconds or a few
ten seconds). Replace it with a local time format where such differences
stand out.
Before:
> [13316826770337] [ID: 00000004] unique: 62, opcode: RELEASEDIR (29), nodeid: 1, insize: 64, pid: 1
> [13316826778175] [ID: 00000004] unique: 62, success, outsize: 16
> [13316826781156] [ID: 00000004] virtio_send_msg: elem 0: with 1 in desc of length 16
> [15138279317927] [ID: 00000001] virtio_loop: Got VU event
> [15138279504884] [ID: 00000001] fv_queue_set_started: qidx=1 started=0
> [15138279519034] [ID: 00000003] fv_queue_thread: kill event on queue 1 - quitting
> [15138280876463] [ID: 00000001] fv_remove_watch: TODO! fd=9
> [15138280897381] [ID: 00000001] virtio_loop: Waiting for VU event
> [15138280946834] [ID: 00000001] virtio_loop: Got VU event
> [15138281175421] [ID: 00000001] virtio_loop: Waiting for VU event
> [15138281182387] [ID: 00000001] virtio_loop: Got VU event
> [15138281189474] [ID: 00000001] virtio_loop: Waiting for VU event
> [15138309321936] [ID: 00000001] virtio_loop: Unexpected poll revents 11
> [15138309434150] [ID: 00000001] virtio_loop: Exit
(Notice how you don't (easily) notice the gap in time after
"virtio_send_msg", and especially the amount of time passed is hard to
estimate.)
After:
> [2020-12-08 06:43:22.58+0100] [ID: 00000004] unique: 51, opcode: RELEASEDIR (29), nodeid: 1, insize: 64, pid: 1
> [2020-12-08 06:43:22.58+0100] [ID: 00000004] unique: 51, success, outsize: 16
> [2020-12-08 06:43:22.58+0100] [ID: 00000004] virtio_send_msg: elem 0: with 1 in desc of length 16
> [2020-12-08 06:43:29.34+0100] [ID: 00000001] virtio_loop: Got VU event
> [2020-12-08 06:43:29.34+0100] [ID: 00000001] fv_queue_set_started: qidx=1 started=0
> [2020-12-08 06:43:29.34+0100] [ID: 00000003] fv_queue_thread: kill event on queue 1 - quitting
> [2020-12-08 06:43:29.34+0100] [ID: 00000001] fv_remove_watch: TODO! fd=9
> [2020-12-08 06:43:29.34+0100] [ID: 00000001] virtio_loop: Waiting for VU event
> [2020-12-08 06:43:29.34+0100] [ID: 00000001] virtio_loop: Got VU event
> [2020-12-08 06:43:29.34+0100] [ID: 00000001] virtio_loop: Waiting for VU event
> [2020-12-08 06:43:29.34+0100] [ID: 00000001] virtio_loop: Got VU event
> [2020-12-08 06:43:29.34+0100] [ID: 00000001] virtio_loop: Waiting for VU event
> [2020-12-08 06:43:29.37+0100] [ID: 00000001] virtio_loop: Unexpected poll revents 11
> [2020-12-08 06:43:29.37+0100] [ID: 00000001] virtio_loop: Exit
Cc: "Dr. David Alan Gilbert" <dgilbert@redhat.com>
Cc: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: Laszlo Ersek <lersek@redhat.com>
Message-Id: <20201208055043.31548-1-lersek@redhat.com>
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Clean up includes so that osdep.h is included first and headers
which it implies are not included manually.
This commit was created with scripts/clean-includes, with the changes
to the following files manually reverted:
contrib/libvhost-user/libvhost-user-glib.h
contrib/libvhost-user/libvhost-user.c
contrib/libvhost-user/libvhost-user.h
contrib/plugins/hotblocks.c
contrib/plugins/hotpages.c
contrib/plugins/howvec.c
contrib/plugins/lockstep.c
linux-user/mips64/cpu_loop.c
linux-user/mips64/signal.c
linux-user/sparc64/cpu_loop.c
linux-user/sparc64/signal.c
linux-user/x86_64/cpu_loop.c
linux-user/x86_64/signal.c
target/s390x/gen-features.c
tests/fp/platform.h
tests/migration/s390x/a-b-bios.c
tests/plugin/bb.c
tests/plugin/empty.c
tests/plugin/insn.c
tests/plugin/mem.c
tests/test-rcu-simpleq.c
tests/test-rcu-slist.c
tests/test-rcu-tailq.c
tests/uefi-test-tools/UefiTestToolsPkg/BiosTablesTest/BiosTablesTest.c
contrib/plugins/, tests/plugin/, and tests/test-rcu-slist.c appear not
to include osdep.h intentionally. The remaining reverts are the same
as in commit bbfff19688.
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Message-Id: <20201113061216.2483385-1-armbru@redhat.com>
Acked-by: Paolo Bonzini <pbonzini@redhat.com>
Acked-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Tested-by: Thomas Huth <thuth@redhat.com>
Acked-by: Cornelia Huck <cohuck@redhat.com>
Acked-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Acked-by: Alexander Bulekov <alxndr@bu.edu>
In main func, strdup lo.source may fail. So check whether strdup
lo.source return NULL before using it.
Signed-off-by: Haotian Li <lihaotian9@huawei.com>
Signed-off-by: Zhiqiang Liu <liuzhiqiang26@huawei.com>
Message-Id: <f1e48ca8-d6de-d901-63c8-4f4024bda518@huawei.com>
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
In main func, func lo_map_reserve is called without NULL check.
If reallocing new_elems fails in func lo_map_grow, the func
lo_map_reserve may return NULL. We should check whether
lo_map_reserve returns NULL before using it.
Signed-off-by: Haotian Li <lihaotian9@huawei.com>
Signed-off-by: Zhiqiang Liu <liuzhiqiang26@huawei.com>
Message-Id: <48887813-1c95-048c-6d10-48e3dd2bac71@huawei.com>
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Contrary to what the check (and warning) in lo_init() claims, we can
announce submounts just fine even without statx() -- the check is based
on comparing both the mount ID and st_dev of parent and child. Without
statx(), we will not have the mount ID; but we always have st_dev.
The only problems we have (without statx() and its mount ID) are:
(1) Mounting the same device twice may lead to both trees being treated
as exactly the same tree by virtiofsd. But that is a problem that
is completely independent of mirroring host submounts in the guest.
Both submount roots will still show the FUSE_SUBMOUNT flag, because
their st_dev still differs from their respective parent.
(2) There is only one exception to (1), and that is if you mount a
device inside a mount of itself: Then, its st_dev will be the same
as that of its parent, and so without a mount ID, virtiofsd will not
be able to recognize the nested mount's root as a submount.
However, thanks to virtiofsd then treating both trees as exactly the
same tree, it will be caught up in a loop when the guest tries to
examine the nested submount, so the guest will always see nothing
but an ELOOP there. Therefore, this case is just fully broken
without statx(), whether we check for submounts (based on st_dev) or
not.
All in all, checking for submounts works well even without comparing the
mount ID (i.e., without statx()). The only concern is an edge case
that, without statx() mount IDs, is utterly broken anyway.
Thus, drop said check in lo_init().
Reported-by: Miklos Szeredi <mszeredi@redhat.com>
Signed-off-by: Max Reitz <mreitz@redhat.com>
Message-Id: <20201103164135.169325-1-mreitz@redhat.com>
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Whenever we encounter a directory with an st_dev or mount ID that
differs from that of its parent, we set the FUSE_ATTR_SUBMOUNT flag so
the guest can create a submount for it.
We only need to do so in lo_do_lookup(). The following functions return
a fuse_attr object:
- lo_create(), though fuse_reply_create(): Calls lo_do_lookup().
- lo_lookup(), though fuse_reply_entry(): Calls lo_do_lookup().
- lo_mknod_symlink(), through fuse_reply_entry(): Calls lo_do_lookup().
- lo_link(), through fuse_reply_entry(): Creating a link cannot create a
submount, so there is no need to check for it.
- lo_getattr(), through fuse_reply_attr(): Announcing submounts when the
node is first detected (at lookup) is sufficient. We do not need to
return the submount attribute later.
- lo_do_readdir(), through fuse_add_direntry_plus(): Calls
lo_do_lookup().
Make announcing submounts optional, so submounts are only announced to
the guest with the announce_submounts option. Some users may prefer the
current behavior, so that the guest learns nothing about the host mount
structure.
(announce_submounts is force-disabled when the guest does not present
the FUSE_SUBMOUNTS capability, or when there is no statx().)
Signed-off-by: Max Reitz <mreitz@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-Id: <20201102161859.156603-6-mreitz@redhat.com>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Using st_dev is not sufficient to uniquely identify a mount: You can
mount the same device twice, but those are still separate trees, and
e.g. by mounting something else inside one of them, they may differ.
Using statx(), we can get a mount ID that uniquely identifies a mount.
If that is available, add it to the lo_inode key.
Most of this patch is taken from Miklos's mail here:
https://marc.info/?l=fuse-devel&m=160062521827983
(virtiofsd-use-mount-id.patch attachment)
Suggested-by: Miklos Szeredi <mszeredi@redhat.com>
Signed-off-by: Max Reitz <mreitz@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-Id: <20201102161859.156603-5-mreitz@redhat.com>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
This reverts the following commits due to their basis on a bogus
linux kernel header update:
c93a656f7b ("tests/acceptance: Add virtiofs_submounts.py")
45ced7ca2f ("tests/acceptance/boot_linux: Accept SSH pubkey")
08dce386e7 ("virtiofsd: Announce sub-mount points")
eba8b096c1 ("virtiofsd: Store every lo_inode's parent_dev")
ede24b6be7 ("virtiofsd: Add fuse_reply_attr_with_flags()")
e2577435d3 ("virtiofsd: Add attr_flags to fuse_entry_param")
2f10415abf ("virtiofsd: Announce FUSE_ATTR_FLAGS")
97d741cc96 ("linux/fuse.h: Pull in from Linux")
Cc: Max Reitz <mreitz@redhat.com>
Cc: Stefan Hajnoczi <stefanha@redhat.com>
Cc: Dr. David Alan Gilbert <dgilbert@redhat.com>
Signed-off-by: Alex Williamson <alex.williamson@redhat.com>
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
Message-id: 160385090886.20017.13382256442750027666.stgit@gimli.home
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Whenever we encounter a directory with an st_dev that differs from that
of its parent, we set the FUSE_ATTR_SUBMOUNT flag so the guest can
create a submount for it.
Make this behavior optional, so submounts are only announced to the
guest with the announce_submounts option. Some users may prefer the
current behavior, so that the guest learns nothing about the host mount
structure.
Signed-off-by: Max Reitz <mreitz@redhat.com>
Message-Id: <20200909184028.262297-7-mreitz@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Manual merge
We want to detect mount points in the shared tree. We report them to
the guest by setting the FUSE_ATTR_SUBMOUNT flag in fuse_attr.flags, but
because the FUSE client will create a submount for every directory that
has this flag set, we must do this only for the actual mount points.
We can detect mount points by comparing a directory's st_dev with its
parent's st_dev. To be able to do so, we need to store the parent's
st_dev in the lo_inode object.
Note that mount points need not necessarily be directories; a single
file can be a mount point as well. However, for the sake of simplicity
let us ignore any non-directory mount points for now.
Signed-off-by: Max Reitz <mreitz@redhat.com>
Message-Id: <20200909184028.262297-6-mreitz@redhat.com>
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
The mapping rule system implemented in the last few patches is
extremely flexible, but not easy to use. Add a simple
'map' type as a sprinkling of sugar to make it easy.
e.g.
-o xattrmap=":map::user.virtiofs.:"
would be sufficient to prefix all xattr's
or
-o xattrmap=":map:trusted.:user.virtiofs.:"
would just prefix 'trusted.' xattr's and leave
everything else alone.
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Message-Id: <20201023165812.36028-6-dgilbert@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Map xattr names coming from the server, i.e. the host filesystem;
currently this is only from listxattr.
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Message-Id: <20201023165812.36028-4-dgilbert@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Map xattr names originating at the client; from get/set/remove xattr.
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Message-Id: <20201023165812.36028-3-dgilbert@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Add an option to define mappings of xattr names so that
the client and server filesystems see different views.
This can be used to have different SELinux mappings as
seen by the guest, to run the virtiofsd with less privileges
(e.g. in a case where it can't set trusted/system/security
xattrs but you want the guest to be able to), or to isolate
multiple users of the same name; e.g. trusted attributes
used by stacking overlayfs.
A mapping engine is used with 3 simple rules; the rules can
be combined to allow most useful mapping scenarios.
The ruleset is defined by -o xattrmap='rules...'.
This patch doesn't use the rule maps yet.
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Message-Id: <20201023165812.36028-2-dgilbert@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
virtiofsd cannot run in a container because CAP_SYS_ADMIN is required to
create namespaces.
Introduce a weaker sandbox mode that is sufficient in container
environments because the container runtime already sets up namespaces.
Use chroot to restrict path traversal to the shared directory.
virtiofsd loses the following:
1. Mount namespace. The process chroots to the shared directory but
leaves the mounts in place. Seccomp rejects mount(2)/umount(2)
syscalls.
2. Pid namespace. This should be fine because virtiofsd is the only
process running in the container.
3. Network namespace. This should be fine because seccomp already
rejects the connect(2) syscall, but an additional layer of security
is lost. Container runtime-specific network security policies can be
used drop network traffic (except for the vhost-user UNIX domain
socket).
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-Id: <20201008085534.16070-1-stefanha@redhat.com>
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
