Prevent attackers from spoofing the tpmKey portion of the AuthSession by
adding a trusted key to the LUKS header metadata. Also, use a persistent
object rather than a transient object.
This provides the following benifits:
1. No way to MITM the tpmKey portion of the session, see [1] for
details.
2. Strengthens the encrypted sessions, note that the bindKey could be
dropped now.
3. Speed, once it's created we just use it.
4. Owner Auth is needed to call create primary, so using the SRK
creates a scratch space for normal users.
This is a "first to set" model, in where the first person to set the key
in the LUKS header wins. Thus, setup should be done in a known good
state. If an SRK, which is a primary key at a special persistent
address, is found, it will use whatever is there. If not, it creates an
SRK. The SRK follows the convetions used through the tpm2-software
organization code on GitHub [2], however, a split has occured between
Windows and Linux with respect to SRK templates. The Linux SRK is
generated with the unique field size set to 0, in Windows, it properly
sets the size to key size in bytes and the unique data to all 0's of that
size. Note the proper templates for SRKs is covered in spec [3].
However, the most important thing, is that both SRKs are passwordless,
and thus they should be interchangable. If Windows is the first to make
the SRK, systemd will gladly accept it and vice-versa.
1. Without the bindKey being utilized, an attacker was able to intercept
this and fake a key, thus being able to decrypt and encrypt traffic as
needed. Introduction of the bindKey strengthened this, but allows for
the attacker to brute force AES128CFB using pin guesses. Introduction of
the salt increases the difficulty of this attack as well as DA attacks
on the TPM objects itself.
2. https://github.com/tpm2-software
3. https://trustedcomputinggroup.org/wp-content/uploads/TCG-TPM-v2.0-Provisioning-Guidance-Published-v1r1.pdfFixes: #20668Fixes: #22637
Signed-off-by: William Roberts <william.c.roberts@intel.com>
Chasing symlinks is a core function that's used in a lot of places
so it deservers a less verbose names so let's rename it to chase()
and chaseat().
We also slightly change the pattern used for the chaseat() helpers
so we get chase_and_openat() and similar.
In some cases, we want to exclude a directory's contents but not
the directory itself. In other cases, we want to exclude a directory
and its contents. Let's extend the denylist logic in copy.h to support
both by changing the denylist from a set to hashmap so we can store the
deny type as the value.
We also modify the repart ExcludeFiles= option to make use of this. If
a directory to exclude ends with a "/", we'll only exclude its contents.
Otherwise, we'll exclude the full directory.
[2/3] Compiling C object systemd-repart.p/src_partition_repart.c.o
../src/partition/repart.c: In function ‘context_open_copy_block_paths’:
../src/partition/repart.c:5194:41: warning: ‘devno’ may be used uninitialized [-Wmaybe-uninitialized]
5194 | source_fd = r = device_open_from_devnum(S_IFBLK, devno, O_RDONLY|O_CLOEXEC|O_NONBLOCK, &opened);
| ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
../src/partition/repart.c:5188:31: note: ‘devno’ was declared here
5188 | dev_t devno;
| ^~~~~
This is with gcc-13.0.1-0.2.fc38.x86_64, -O2. I'm pretty sure the code
is correct. I also tried adding some asserts where errno is used for the return
value, but that didn't help. I think resolve_copy_blocks_auto() is just too long
for gcc to understand.
linux/btrfs.h needs to be included after sys/mount.h, as since [0]
linux/btrfs.h includes linux/fs.h causing build errors:
```
In file included from /usr/include/linux/fs.h:19,
from ../src/basic/linux/btrfs.h:29,
from ../src/partition/growfs.c:6:
/usr/include/sys/mount.h:35:3: error: expected identifier before numeric constant
35 | MS_RDONLY = 1, /* Mount read-only. */
| ^~~~~~~~~
[1222/2169] Compiling C object systemd-creds.p/src_creds_creds.c.o
ninja: build stopped: subcommand failed.
```
See: https://github.com/systemd/systemd/issues/8507
[0] https://github.com/torvalds/linux/commit/a28135303a669917002f569aecebd5758263e4aa
This reverts commit e59678b2cf.
We also modify the repart integration tests to make them pass with the
changes in this commit. In short, we have to make sure every file is
owned by the user executing repart. We use tee instead of cat since it
makes that easier. This also has the benefit of improving debugability
as seeing the config file contents on stdout makes it easier to know
which test is failing.
Add a salt to the pin and store it in the TPM2 LUKS header for future
this. This adds entropy to user supplied pins and helps brute forcing
the passphrase on the key residing in the TPM or brute forcing bind key
encrypted sessions with low entropy passphrases.
Signed-off-by: malikabhi05 <abhishek.malik@intel.com>
Signed-off-by: William Roberts <william.c.roberts@intel.com>
Let's allow users to configure the (logical) sector size of their
image. This is required when building images for a 4k sector size
disk on a 512b sector size host or vice-versa.
This returns an ssize_t, not an int. On populare archs that's the
difference between 64bit and 32bit. hence, let's be more careful here,
and not silently drop half the bits on the ground by assigning the
return value to "int".
As noticed by @malikabhi05:
https://github.com/systemd/systemd/pull/24754#discussion_r1062903159
-1 was used everywhere, but -EBADF or -EBADFD started being used in various
places. Let's make things consistent in the new style.
Note that there are two candidates:
EBADF 9 Bad file descriptor
EBADFD 77 File descriptor in bad state
Since we're initializating the fd, we're just assigning a value that means
"no fd yet", so it's just a bad file descriptor, and the first errno fits
better. If instead we had a valid file descriptor that became invalid because
of some operation or state change, the other errno would fit better.
In some places, initialization is dropped if unnecessary.
When generating verity partitions, we only know the partition UUID
of the verity data and hash partition after doing the verity
formatting. This means we can't use the verity partition UUID as
input for deriving the filesystem/luks UUIDs. Currently, we derive
the filesystem/luks UUID from the null UUID instead, which isn't
ideal. Instead, let's always generate a partition UUID and use it
to derive the fs/luks UUIDs, but only use it as the actual partition
UUID if we're not doing verity for the partition.
Instead of having Minimize= take a boolean let's allow for two
different ways to enable it. "best" means we want the most minimal
image possible, which currently is only possible for read-only
filesystems but can be extended in the future with bisection
to find the most minimal possible size.
We also add "guess", which is the current behavior, where we
populate once and use the sparse size to make a reasonable guess
on a size that fits all the sources without needing to O(log(n))
tries to find the most minimal size.
libblkid really should define an enum for this on its own, but it
currently doesn't and returns literal numeric values. Lets make this
more readable by adding our own symbolic names via an enum.
e.g. vfat doesn't support symlinks, sockets, fifos, etc so let's ignore
any copy failures related to unsupported file types when populating
filesystems.
Yu Watanabe
William Roberts
Daan De Meyer
Morten Linderud
Zbigniew Jędrzejewski-Szmek
Lennart Poettering
Frantisek Sumsal
Aidan Dang
Luca Boccassi