| Age | Commit message (Collapse) | Author | Files | Lines |
|---|
|
Have added new option "-A" in setcifsacl utility to reorder ACEs in
preferred order.
|
|
Signed-off-by: Pavel Shilovsky <pshilov@microsoft.com>
|
|
In certain scenarios (e.g. kerberos multimount), when a process does
syscalls, the kernel sometimes has to query information or trigger
some actions in userspace. To do so it calls the cifs.upcall binary
with information on the process that triggered the syscall in the
first place.
ls(pid=10) ====> open("foo") ====> kernel
that user doesn't have an SMB
session, lets create one using his
kerberos credential cache
call cifs.upcall and ask for krb info
for whoever owns pid=10
|
cifs.upcall --pid 10 <=================+
...gather info...
return binary blob used
when establishing SMB session
===================> kernel
open SMB session, handle
open() syscall
ls <=================================== return open() result to ls
On a system using containers, the kernel is still calling the host
cifs.upcall and using the host configuration (for network, pid, etc).
This patch changes the behaviour of cifs.upcall so that it uses the
calling process namespaces (ls in the example) when doing its
job.
Note that the kernel still calls the binary in the host, but the
binary will place itself the contexts of the calling process
namespaces.
This code makes use of (but shouldn't require) the following kernel
config options and syscall flags:
approx. year |
introduced | config/flags
---------------+----------------
2008 | CONFIG_NAMESPACES=y
2007 | CONFIG_UTS_NS=y
2020 | CONFIG_TIME_NS=y
2006 | CONFIG_IPC_NS=y
2007 | CONFIG_USER_NS
2008 | CONFIG_PID_NS=y
2007 | CONFIG_NET_NS=y
2007 | CONFIG_CGROUPS
2016 | CLONE_NEWCGROUP setns() flag
Signed-off-by: Aurelien Aptel <aaptel@suse.com>
Signed-off-by: Alastair Houghton <alastair@alastairs-place.net>
|
|
Signed-off-by: Pavel Shilovsky <pshilov@microsoft.com>
|
|
Signed-off-by: Pavel Shilovsky <pshilov@microsoft.com>
|
|
Signed-off-by: Pavel Shilovsky <pshilov@microsoft.com>
|
|
Make drop_call_capabilities() in cifs.upcall update the bounding capabilities
only if CAP_SETCAP is present.
This is an addendum to the patch recently provided in [1]. Without this
additional change, cifs.upcall can still fail while trying to mount a CIFS
network share with krb5:
kernel: CIFS: Attempting to mount //server.domain.lan/myshare
cifs.upcall[39484]: key description: cifs.spnego;0;0;39010000;ver=0x2;host=server.domain.lan>
cifs.upcall[39484]: ver=2
cifs.upcall[39484]: host=server.domain.lan
cifs.upcall[39484]: ip=172.22.3.14
cifs.upcall[39484]: sec=1
cifs.upcall[39484]: uid=1000
cifs.upcall[39484]: creduid=1000
cifs.upcall[39484]: user=username
cifs.upcall[39484]: pid=39481
cifs.upcall[39484]: get_cachename_from_process_env: pathname=/proc/39481/environ
cifs.upcall[39484]: get_cachename_from_process_env: cachename = FILE:/tmp/.krb5cc_1000
cifs.upcall[39484]: drop_all_capabilities: Unable to apply capability set: Success
cifs.upcall[39484]: Exit status 1
[1] https://marc.info/?l=linux-cifs&m=160595758021261
Signed-off-by: Alexander Koch <mail@alexanderkoch.net>
Signed-off-by: Jonas Witschel <diabonas@archlinux.org>
|
|
In the current mount.cifs logic, when sudo is used for mount,
uid=0, so the mount command searches for cruid=0 unless explicitly
specified by the user. The user may already have cred cache populated
but mount.cifs would end up searching cred cache for uid=0.
mount.cifs can avoid this confusion by reading the cruid from SUDO_UID
environment variable. If it is set to non-zero, we can make cruid=$SUDO_UID.
However, to maintain backward compatibility, keeping this as a fallback option.
If mount fails with ENOKEY, then retry with this option.
To enable this fallback, I had to make a few minor changes in the flow.
Signed-off-by: Shyam Prasad N <sprasad@microsoft.com>
Reviewed-by: Aurelien Aptel <aaptel@suse.com>
|
|
Signed-off-by: Pavel Shilovsky <pshilov@microsoft.com>
|
|
The code tries to optimise for the last parameter not needing to update
the position which means that every time a new one is added to the end
by copying and pasting, the string position is not updated.
That makes it impossible to use backup_uid=/backup_gid=/snapshot= after
gid= or snapshot= after backup_gid= because part of the string is
overwritten and contains invalid keys like "gbackup_uid".
Prepare for the next parameter to be added on the end by updating the
position for snapshot= even though it will be unused.
|
|
libcap-ng 0.8.1 tightened the error checking on capng_apply, returning an error
of -4 when trying to update the capability bounding set without having the
CAP_SETPCAP capability to be able to do so. Previous versions of libcap-ng
silently skipped updating the bounding set and only updated the normal
CAPNG_SELECT_CAPS capabilities instead.
Check beforehand whether we have CAP_SETPCAP, in which case we can use
CAPNG_SELECT_BOTH to update both the normal capabilities and the bounding set.
Otherwise, we can at least update the normal capabilities, but refrain from
trying to update the bounding set to avoid getting an error.
Signed-off-by: Jonas Witschel <diabonas@archlinux.org>
|
|
libcap-ng 0.8.1 tightened the error checking on capng_apply, returning an error
of -4 when trying to update the capability bounding set without having the
CAP_SETPCAP capability to be able to do so. Previous versions of libcap-ng
silently skipped updating the bounding set and only updated the normal
CAPNG_SELECT_CAPS capabilities instead.
Check beforehand whether we have CAP_SETPCAP, in which case we can use
CAPNG_SELECT_BOTH to update both the normal capabilities and the bounding set.
Otherwise, we can at least update the normal capabilities, but refrain from
trying to update the bounding set to avoid getting an error.
Signed-off-by: Jonas Witschel <diabonas@archlinux.org>
|
|
Extend getcifsacl/setcifsacl utilities to handle System ACLs (SACLs)
in addition to Discretionary ACLs (DACLs). The SACL extensions depend
on CIFS client support for system.cifs_ntsd_full extended attribute.
Signed-off-by: Boris Protopopov <pboris@amazon.com>
|
|
Signed-off-by: Pavel Shilovsky <pshilov@microsoft.com>
|
|
|
|
mount.cifs currently complains about the "comment" option:
CIFS: Unknown mount option "comment=foo"
mount(8) on Linux says:
The command mount does not pass the mount options unbindable,
runbindable, private, rprivate, slave, rslave, shared, rshared,
auto, noauto, comment, x-*, loop, offset and sizelimit to the
mount.<suffix> helpers.
So if mount.cifs decides to re-read /etc/fstab it should ignore the
comment option.
A lot of online posts say to use comment=x-gvfs-show as an option to
have a Linux file manager display a mountpoint for a user mountable
filesystem. While the "comment=" part is superfluous when combined
with an x-* option, the problem is still difficult to debug.
Signed-off-by: Martin Schwenke <martin@meltin.net>
Reviewed-by: Aurelien Aptel <aaptel@suse.com>
|
|
Signed-off-by: Bjoern Jacke <bjacke@samba.org>
|
|
According to lexgrog(1), when a manpage refers to multiple programs
their names should be separated using a comma and a whitespace. This
helps silence a lintian warning when building cifs-utils on Debian.
Signed-off-by: Sergio Durigan Junior <sergio.durigan@canonical.com>
|
|
When we compile cifs-utils, we may probabilistic
encounter install error like:
cd ***/sbin && ln -sf mount.cifs mount.smb3
***/sbin: No such file or directory
The reason of this problem is that if we compile
cifs-utils using multithreading, target
'install-sbinPROGRAMS' may be built after
target 'install-exec-hook' of the main Makefile.
Target 'install-sbinPROGRAMS' will copy the
executable file 'mount.cifs' to the $(ROOTSBINDIR),
which target 'install-exec-hook' will do the
'ln' command on.
This patch add the dependency of target
'install-exec-hook' to ensure the correct order
of the compiling.
Signed-off-by: lizhe <lizhe67@huawei.com>
|
|
|
|
When make install is run during package building, DESTDIR parameter is passed, e.g.:
$ rpm --eval %makeinstall_std
make DESTDIR=/root/rpmbuild/BUILDROOT/%{name}-%{version}-%{release}-rosa2016.1.x86_64-buildroot install
Without DESTDIR build scripts tried to create symlinks outside of the build root:
make[3]: Entering directory '/tmp/abf/rpmbuild/BUILD/cifs-utils-6.10'
(cd /sbin && ln -sf mount.cifs mount.smb3)
ln: failed to create symbolic link 'mount.smb3': Permission denied
The same fix was introduced in Arch Linux package when updating from 6.9 to 6.10:
https://git.archlinux.org/svntogit/packages.git/commit/trunk/PKGBUILD?h=packages/cifs-utils&id=c75b246a762ea9b90db404dfebc6d35d5b16972f
|
|
Signed-off-by: Pavel Shilovsky <pshilov@microsoft.com>
|
|
Extend setcifsacl utility to allow setting owner and group SIDs
in the security descriptor in addition to setting ACLs. This is
a user-friendly intefrace for setting owner and group SIDs that
takes advantage of the recent extensions in the CIFS kernel
client, and it complements setting raw values via setfattr.
Signed-off-by: Boris Protopopov <boris.v.protopopov@gmail.com>
|
|
Convert owner and group SID offsets to LE format
when writing to ntsd
Signed-off-by: Boris Protopopov <boris.v.protopopov@gmail.com>
|
|
Signed-off-by: Ronnie Sahlberg <lsahlber@redhat.com>
|
|
Signed-off-by: Aurelien Aptel <aaptel@suse.com>
|
|
Signed-off-by: Pavel Shilovsky <pshilov@microsoft.com>
|
|
A bug has been reported recently for the mount.cifs utility which is
part of the cifs-utils package. The tool has a shell injection issue
where one can embed shell commands via the username mount option. Those
commands will be run via popen() in the context of the user calling
mount.
The bug requires cifs-utils to be built with --with-systemd (enabled
by default if supported).
A quick test to check if the mount.cifs binary is vulnerable is to look
for popen() calls like so:
$ nm mount.cifs | grep popen
U popen@@GLIBC_2.2.5
If the user is allowed to run mount.cifs via sudo, he can obtain a root
shell.
sudo mount.cifs -o username='`sh`' //1 /mnt
If mount.cifs has the setuid bit, the command will still be run as the
calling user (no privilege escalation).
The bug was introduced in June 2012 with commit 4e264031d0da7d3f2
("mount.cifs: Use systemd's mechanism for getting password, if
present.").
Affected versions:
cifs-utils-5.6
cifs-utils-5.7
cifs-utils-5.8
cifs-utils-5.9
cifs-utils-6.0
cifs-utils-6.1
cifs-utils-6.2
cifs-utils-6.3
cifs-utils-6.4
cifs-utils-6.5
cifs-utils-6.6
cifs-utils-6.7
cifs-utils-6.8
cifs-utils-6.9
cifs-utils-6.10
Bug: https://bugzilla.samba.org/show_bug.cgi?id=14442
Reported-by: Vadim Lebedev <vadim@mbdsys.com>
Signed-off-by: Paulo Alcantara (SUSE) <pc@cjr.nz>
Signed-off-by: Aurelien Aptel <aaptel@suse.com>
|
|
Signed-off-by: Pavel Shilovsky <pshilov@microsoft.com>
|
|
|
|
Signed-off-by: Kenneth D'souza <kdsouza@redhat.com>
Signed-off-by: Pavel Shilovsky <pshilov@microsoft.com>
|
|
Signed-off-by: Ronnie Sahlberg <lsahlber@redhat.com>
|
|
This patch changes the program name from smb2quota to
smb2-quota and uses a simple code logic for quota entries.
Signed-off-by: Kenneth D'souza <kdsouza@redhat.com>
Signed-off-by: Ronnie Sahlberg <lsahlberg@redhat.com>
|
|
As we are supporting mount.smb3 to be invoked, the error output
should contain the called program and not mount.cifs
Signed-off-by: Kenneth D'souza <kdsouza@redhat.com>
|
|
As we will slowly move towards smb3 filesystem,
supporting through "mount -t smb3" is important.
Signed-off-by: Kenneth D'souza <kdsouza@redhat.com>
|
|
Signed-off-by: Aurelien Aptel <aaptel@suse.com>
|
|
This option is deprecated and currently ignored since
kernel v3.10
Signed-off-by: Kenneth D'souza <kdsouza@redhat.com>
Reviewed-by: Aurelien Aptel <aaptel@suse.com>
|
|
Signed-off-by: Kenneth D'souza <kdsouza@redhat.com>
|
|
Signed-off-by: Kenneth D'souza <kdsouza@redhat.com>
Signed-off-by: Ronnie Sahlberg <lsahlber@redhat.com>
Signed-off-by: Pavel Shilovsky <pshilov@microsoft.com>
|
|
Signed-off-by: Pavel Shilovsky <pshilov@microsoft.com>
|
|
Signed-off-by: Ronnie Sahlberg <lsahlber@redhat.com>
|
|
CCM encryption
Session Id: e2 3e ea ae 00 00 00 00
Session Key: 65 7e 0e d5 3c 06 5a 06 50 a3 ef 96 c1 64 3d 1f
Server Encryption Key: 5e 42 a7 b5 57 75 d6 56 4a 5d 33 97 e6 45 07 76
Server Decryption Key: 1f 64 db a3 0f 24 e3 4d b6 31 00 ab 9a af 22 47
Signed-off-by: Steve French <stfrench@microsoft.com>
|
|
When attemping to chdir into non-existing directories, mount.cifs
crashes.
This patch fixes the following ASAN report:
$ ./mount.cifs //localhost/foo /mnt/invalid-dir -o ...
/mnt/bar -o username=foo,password=foo,vers=1.0
Couldn't chdir to /mnt/bar: No such file or directory
=================================================================
==11846==ERROR: AddressSanitizer: attempting free on address which was
not malloc()-ed: 0x7ffd86332e97 in thread T0
#0 0x7f0860ca01e7 in
__interceptor_free (/usr/lib64/libasan.so.5+0x10a1e7)
#1 0x557edece9ccb in
acquire_mountpoint (/home/paulo/src/cifs-utils/mount.cifs+0xeccb)
#2 0x557edecea63d in
main (/home/paulo/src/cifs-utils/mount.cifs+0xf63d)
#3 0x7f08609f0bca in __libc_start_main (/lib64/libc.so.6+0x26bca)
#4 0x557edece27d9 in
_start (/home/paulo/src/cifs-utils/mount.cifs+0x77d9)
Address 0x7ffd86332e97 is located in stack of thread T0 at offset 8951
in frame
#0 0x557edece9ce0 in
main (/home/paulo/src/cifs-utils/mount.cifs+0xece0)
This frame has 2 object(s):
[48, 52) 'rc' (line 1959)
[64, 72) 'mountpoint' (line 1955) <== Memory access at offset 8951
overflows this variable
HINT: this may be a false positive if your program uses some custom
stack unwind mechanism, swapcontext or vfork
(longjmp and C++ exceptions *are* supported)
SUMMARY: AddressSanitizer: bad-free (/usr/lib64/libasan.so.5+0x10a1e7)
in __interceptor_free
==11846==ABORTING
Fixes: bf7f48f4c7dc ("mount.cifs.c: fix memory leaks in main func")
Signed-off-by: Paulo Alcantara (SUSE) <pc@cjr.nz>
Reviewed-by: David Mulder <dmulder@suse.com>
|
|
It can be easily reproduced with the following:
# chmod +s `which mount.cifs`
# echo "//localhost/share /mnt cifs \
users,username=foo,password=XXXX" >> /etc/fstab
# su - foo
$ mount /mnt
free(): double free detected in tcache 2
Child process terminated abnormally.
The problem was that check_fstab() already freed orgoptions pointer
and then we freed it again in main() function.
Fixes: bf7f48f4c7dc ("mount.cifs.c: fix memory leaks in main func")
Signed-off-by: Paulo Alcantara (SUSE) <paulo@paulo.ac>
Reviewed-by: Aurelien Aptel <aaptel@suse.com>
|
|
Fixes a bug where `sacloffset` may not be set at all later on and therefore it
can contain the original memory contents == trash.
|
|
Fixes a bug inside a call to `verify_ace_flag`. When a flag string (char*)
passed as a first parameter is "0x0", the final flag value (the second
parameter - the value of a pointer to uint8_t) is not modified at all
and contains the original memory contents == trash.
|
|
In mount.cifs module, orgoptions and mountpoint in the main func
point to the memory allocated by func realpath and strndup respectively.
However, they are not freed before the main func returns so that the
memory leaks occurred.
The memory leak problem is reported by LeakSanitizer tool.
LeakSanitizer url: "https://github.com/google/sanitizers"
Here I free the pointers orgoptions and mountpoint before main
func returns.
Fixes：7549ad5e7126 ("memory leaks: caused by func realpath and strndup")
Signed-off-by: Jiawen Liu <liujiawen10@huawei.com>
Reported-by: Jin Du <dujin1@huawei.com>
Reviewed-by: Saisai Zhang <zhangsaisai@huawei.com>
Reviewed-by: Aurélien Aptel <aaptel@suse.com>
|
|
Signed-off-by: Pavel Shilovsky <pshilov@microsoft.com>
|
|
Add support for -R option so we can list the ACLs of all files and
directories recursively.
Signed-off-by: Kenneth D'souza <kdsouza@redhat.com>
|
|
Signed-off-by: Ronnie Sahlberg <lsahlber@redhat.com>
|
