The kernel and the user obtain an xattr value in two different ways:
kernel (EVM): uses vfs_getxattr_alloc() which obtains the xattr value from
the filesystem handler (raw value);
user (ima-evm-utils): uses vfs_getxattr() which obtains the xattr value
from the LSMs (normalized value).
Normally, this does not have an impact unless security.selinux is set with
setfattr, with a value not terminated by '\0' (this is not the recommended
way, security.selinux should be set with the appropriate tools such as
chcon and restorecon).
In this case, the kernel and the user see two different xattr values: the
former sees the xattr value without '\0' (raw value), the latter sees the
value with '\0' (value normalized by SELinux).
This could result in two different verification outcomes from EVM and
ima-evm-utils, if a signature was calculated with a security.selinux value
terminated by '\0' and the value set in the filesystem is not terminated by
'\0'. The former would report verification failure due to the missing '\0',
while the latter would report verification success (because it gets the
normalized value with '\0').
This patch mitigates this issue by comparing in evm_calc_hmac_or_hash() the
size of the xattr returned by the two xattr functions and by warning the
user if there is a discrepancy.
Signed-off-by: Roberto Sassu <roberto.sassu@huawei.com>
Suggested-by: Mimi Zohar <zohar@linux.ibm.com>
Signed-off-by: Mimi Zohar <zohar@linux.ibm.com>
Output the data used in calculating the EVM digest and the resulting
digest as ascii hexadecimal strings.
Suggested-by: Lakshmi Ramasubramanian <nramas@linux.microsoft.com> (CONFIG_DYNAMIC_DEBUG)
Reviewed-by: Lakshmi Ramasubramanian <nramas@linux.microsoft.com>
Reported-by: kernel test robot <lkp@intel.com> (Use %zu for size_t)
Signed-off-by: Mimi Zohar <zohar@linux.ibm.com>
IMA measures contents of a given file/buffer/critical-data record,
and properly re-measures it on change. However, IMA does not measure
the duplicate value for a given record, since TPM extend is a very
expensive operation. For example, if the record changes from value
'v#1' to 'v#2', and then back to 'v#1', IMA will not measure and log
the last change to 'v#1', since the hash of 'v#1' for that record is
already present in the IMA htable. This limits the ability of an
external attestation service to accurately determine the current state
of the system. The service would incorrectly conclude that the latest
value of the given record on the system is 'v#2', and act accordingly.
Define and use a new Kconfig option IMA_DISABLE_HTABLE to permit
duplicate records in the IMA measurement list.
In addition to the duplicate measurement records described above,
other duplicate file measurement records may be included in the log,
when CONFIG_IMA_DISABLE_HTABLE is enabled. For example,
- i_version is not enabled,
- i_generation changed,
- same file present on different filesystems,
- an inode is evicted from dcache
Signed-off-by: Tushar Sugandhi <tusharsu@linux.microsoft.com>
Reviewed-by: Petr Vorel <pvorel@suse.cz>
[zohar@linux.ibm.com: updated list of duplicate measurement records]
Signed-off-by: Mimi Zohar <zohar@linux.ibm.com>
The function prototype for ima_add_kexec_buffer() is present
in 'linux/ima.h'. But this header file is not included in
ima_kexec.c where the function is implemented. This results
in the following compiler warning when "-Wmissing-prototypes" flag
is turned on:
security/integrity/ima/ima_kexec.c:81:6: warning: no previous prototype
for function 'ima_add_kexec_buffer' [-Wmissing-prototypes]
Include the header file 'linux/ima.h' in ima_kexec.c to fix
the compiler warning.
Fixes: dce92f6b11 (arm64: Enable passing IMA log to next kernel on kexec)
Reported-by: kernel test robot <lkp@intel.com>
Signed-off-by: Lakshmi Ramasubramanian <nramas@linux.microsoft.com>
Acked-by: Rob Herring <robh@kernel.org>
Signed-off-by: Mimi Zohar <zohar@linux.ibm.com>
In preparation to enable -Wimplicit-fallthrough for Clang, fix a
fall-through warning by explicitly adding a break statement instead
of just letting the code fall through to the next case.
Link: https://github.com/KSPP/linux/issues/115
Signed-off-by: Gustavo A. R. Silva <gustavoars@kernel.org>
Reviewed-by: Kees Cook <keescook@chromium.org>
Signed-off-by: Mimi Zohar <zohar@linux.ibm.com>
This patch fixes the sparse warning for ima_post_key_create_or_update() by
adding the header file that defines the prototype (linux/ima.h).
Signed-off-by: Roberto Sassu <roberto.sassu@huawei.com>
Signed-off-by: Mimi Zohar <zohar@linux.ibm.com>
The endianness of a variable written to the measurement list cannot be
determined at compile time, as it depends on the value of the
ima_canonical_fmt global variable (set through a kernel option with the
same name if the machine is big endian).
If ima_canonical_fmt is false, the endianness of a variable is the same as
the machine; if ima_canonical_fmt is true, the endianness is little endian.
The warning arises due to this type of instruction:
var = cpu_to_leXX(var)
which tries to assign a value in little endian to a variable with native
endianness (little or big endian).
Given that the variables set with this instruction are not used in any
operation but just written to a buffer, it is safe to force the type of the
value being set to be the same of the type of the variable with:
var = (__force <var type>)cpu_to_leXX(var)
Reported-by: kernel test robot <lkp@intel.com>
Signed-off-by: Roberto Sassu <roberto.sassu@huawei.com>
Signed-off-by: Mimi Zohar <zohar@linux.ibm.com>
The code expects that the values being parsed from a buffer when the
ima_canonical_fmt global variable is true are in little endian. Thus, this
patch sets the casting types accordingly.
Reported-by: kernel test robot <lkp@intel.com>
Signed-off-by: Roberto Sassu <roberto.sassu@huawei.com>
Signed-off-by: Mimi Zohar <zohar@linux.ibm.com>
This patch avoids that evm_write_xattrs() returns an error when audit is
not enabled. The ab variable can be NULL and still be passed to the other
audit_log_() functions, as those functions do not include any instruction.
Signed-off-by: Roberto Sassu <roberto.sassu@huawei.com>
Signed-off-by: Mimi Zohar <zohar@linux.ibm.com>
With the recent introduction of the evmsig template field, remote verifiers
can obtain the EVM portable signature instead of the IMA signature, to
verify file metadata.
After introducing the new fields to include file metadata in the
measurement list, this patch finally defines the evm-sig template, whose
format is:
d-ng|n-ng|evmsig|xattrnames|xattrlengths|xattrvalues|iuid|igid|imode
xattrnames, xattrlengths and xattrvalues are populated only from defined
EVM protected xattrs, i.e. the ones that EVM considers to verify the
portable signature. xattrnames and xattrlengths are populated only if the
xattr is present.
xattrnames and xattrlengths are not necessary for verifying the EVM
portable signature, but they are included for completeness of information,
if a remote verifier wants to infer more from file metadata.
Signed-off-by: Roberto Sassu <roberto.sassu@huawei.com>
Signed-off-by: Mimi Zohar <zohar@linux.ibm.com>
This patch defines the new template fields xattrnames, xattrlengths and
xattrvalues, which contain respectively a list of xattr names (strings,
separated by |), lengths (u32, hex) and values (hex). If an xattr is not
present, the name and length are not displayed in the measurement list.
Reported-by: kernel test robot <lkp@intel.com> (Missing prototype def)
Signed-off-by: Roberto Sassu <roberto.sassu@huawei.com>
Signed-off-by: Mimi Zohar <zohar@linux.ibm.com>
Currently, the evm_config_default_xattrnames array contains xattr names
only related to LSMs which are enabled in the kernel configuration.
However, EVM portable signatures do not depend on local information and a
vendor might include in the signature calculation xattrs that are not
enabled in the target platform.
Just including all xattrs names in evm_config_default_xattrnames is not a
safe approach, because a target system might have already calculated
signatures or HMACs based only on the enabled xattrs. After applying this
patch, EVM would verify those signatures and HMACs with all xattrs instead.
The non-enabled ones, which could possibly exist, would cause a
verification error.
Thus, this patch adds a new field named enabled to the xattr_list
structure, which is set to true if the LSM associated to a given xattr name
is enabled in the kernel configuration. The non-enabled xattrs are taken
into account only in evm_calc_hmac_or_hash(), if the passed security.evm
type is EVM_XATTR_PORTABLE_DIGSIG.
The new function evm_protected_xattr_if_enabled() has been defined so that
IMA can include all protected xattrs and not only the enabled ones in the
measurement list, if the new template fields xattrnames, xattrlengths or
xattrvalues have been included in the template format.
Signed-off-by: Roberto Sassu <roberto.sassu@huawei.com>
Signed-off-by: Mimi Zohar <zohar@linux.ibm.com>
This patch defines the new template field imode, which includes the
inode mode. It can be used by a remote verifier to verify the EVM portable
signature, if it was included with the template fields sig or evmsig.
Signed-off-by: Roberto Sassu <roberto.sassu@huawei.com>
Signed-off-by: Mimi Zohar <zohar@linux.ibm.com>
This patch defines the new template fields iuid and igid, which include
respectively the inode UID and GID. For idmapped mounts, still the original
UID and GID are provided.
These fields can be used to verify the EVM portable signature, if it was
included with the template fields sig or evmsig.
Signed-off-by: Roberto Sassu <roberto.sassu@huawei.com>
Acked-by: Christian Brauner <christian.brauner@ubuntu.com>
Signed-off-by: Mimi Zohar <zohar@linux.ibm.com>
This patch introduces the new function ima_show_template_uint(). This can
be used for showing integers of different sizes in ASCII format. The
function ima_show_template_data_ascii() automatically determines how to
print a stored integer by checking the integer size.
If integers have been written in canonical format,
ima_show_template_data_ascii() calls the appropriate leXX_to_cpu() function
to correctly display the value.
Signed-off-by: Roberto Sassu <roberto.sassu@huawei.com>
Signed-off-by: Mimi Zohar <zohar@linux.ibm.com>
Files might come from a remote source and might have xattrs, including
security.ima. It should not be IMA task to decide whether security.ima
should be kept or not. This patch removes the removexattr() system
call in ima_inode_post_setattr().
Signed-off-by: Roberto Sassu <roberto.sassu@huawei.com>
Signed-off-by: Mimi Zohar <zohar@linux.ibm.com>
With the patch to accept EVM portable signatures when the
appraise_type=imasig requirement is specified in the policy, appraisal can
be successfully done even if the file does not have an IMA signature.
However, remote attestation would not see that a different signature type
was used, as only IMA signatures can be included in the measurement list.
This patch solves the issue by introducing the new template field 'evmsig'
to show EVM portable signatures and by including its value in the existing
field 'sig' if the IMA signature is not found.
Suggested-by: Mimi Zohar <zohar@linux.ibm.com>
Signed-off-by: Roberto Sassu <roberto.sassu@huawei.com>
Signed-off-by: Mimi Zohar <zohar@linux.ibm.com>
System administrators can require that all accessed files have a signature
by specifying appraise_type=imasig in a policy rule.
Currently, IMA signatures satisfy this requirement. Appended signatures may
also satisfy this requirement, but are not applicable as IMA signatures.
IMA/appended signatures ensure data source authentication for file content
and prevent any change. EVM signatures instead ensure data source
authentication for file metadata. Given that the digest or signature of the
file content must be included in the metadata, EVM signatures provide the
same file data guarantees of IMA signatures, as well as providing file
metadata guarantees.
This patch lets systems protected with EVM signatures pass appraisal
verification if the appraise_type=imasig requirement is specified in the
policy. This facilitates deployment in the scenarios where only EVM
signatures are available.
The patch makes the following changes:
file xattr types:
security.ima: IMA_XATTR_DIGEST/IMA_XATTR_DIGEST_NG
security.evm: EVM_XATTR_PORTABLE_DIGSIG
execve(), mmap(), open() behavior (with appraise_type=imasig):
before: denied (file without IMA signature, imasig requirement not met)
after: allowed (file with EVM portable signature, imasig requirement met)
open(O_WRONLY) behavior (without appraise_type=imasig):
before: allowed (file without IMA signature, not immutable)
after: denied (file with EVM portable signature, immutable)
In addition, similarly to IMA signatures, this patch temporarily allows
new files without or with incomplete metadata to be opened so that content
can be written.
Signed-off-by: Roberto Sassu <roberto.sassu@huawei.com>
Signed-off-by: Mimi Zohar <zohar@linux.ibm.com>
With the patch to allow xattr/attr operations if a portable signature
verification fails, cp and tar can copy all xattrs/attrs so that at the
end of the process verification succeeds.
However, it might happen that the xattrs/attrs are already set to the
correct value (taken at signing time) and signature verification succeeds
before the copy has completed. For example, an archive might contains files
owned by root and the archive is extracted by root.
Then, since portable signatures are immutable, all subsequent operations
fail (e.g. fchown()), even if the operation is legitimate (does not alter
the current value).
This patch avoids this problem by reporting successful operation to user
space when that operation does not alter the current value of xattrs/attrs.
With this patch, the one that introduces evm_hmac_disabled() and the one
that allows a metadata operation on the INTEGRITY_FAIL_IMMUTABLE error, EVM
portable signatures can be used without disabling metadata verification
(by setting EVM_ALLOW_METADATA_WRITES). Due to keeping metadata
verification enabled, altering immutable metadata protected with a portable
signature that was successfully verified will be denied (existing
behavior).
Reported-by: kernel test robot <lkp@intel.com> [implicit declaration of function]
Signed-off-by: Roberto Sassu <roberto.sassu@huawei.com>
Reviewed-by: Christian Brauner <christian.brauner@ubuntu.com>
Cc: Andreas Gruenbacher <agruenba@redhat.com>
Signed-off-by: Mimi Zohar <zohar@linux.ibm.com>
In preparation for 'evm: Allow setxattr() and setattr() for unmodified
metadata', this patch passes mnt_userns to the inode set/remove xattr hooks
so that the GID of the inode on an idmapped mount is correctly determined
by posix_acl_update_mode().
Signed-off-by: Roberto Sassu <roberto.sassu@huawei.com>
Reviewed-by: Christian Brauner <christian.brauner@ubuntu.com>
Cc: Andreas Gruenbacher <agruenba@redhat.com>
Signed-off-by: Mimi Zohar <zohar@linux.ibm.com>
If files with portable signatures are copied from one location to another
or are extracted from an archive, verification can temporarily fail until
all xattrs/attrs are set in the destination. Only portable signatures may
be moved or copied from one file to another, as they don't depend on
system-specific information such as the inode generation. Instead portable
signatures must include security.ima.
Unlike other security.evm types, EVM portable signatures are also
immutable. Thus, it wouldn't be a problem to allow xattr/attr operations
when verification fails, as portable signatures will never be replaced with
the HMAC on possibly corrupted xattrs/attrs.
This patch first introduces a new integrity status called
INTEGRITY_FAIL_IMMUTABLE, that allows callers of
evm_verify_current_integrity() to detect that a portable signature didn't
pass verification and then adds an exception in evm_protect_xattr() and
evm_inode_setattr() for this status and returns 0 instead of -EPERM.
Signed-off-by: Roberto Sassu <roberto.sassu@huawei.com>
Signed-off-by: Mimi Zohar <zohar@linux.ibm.com>
When a file is being created, LSMs can set the initial label with the
inode_init_security hook. If no HMAC key is loaded, the new file will have
LSM xattrs but not the HMAC. It is also possible that the file remains
without protected xattrs after creation if no active LSM provided it, or
because the filesystem does not support them.
Unfortunately, EVM will deny any further metadata operation on new files,
as evm_protect_xattr() will return the INTEGRITY_NOLABEL error if protected
xattrs exist without security.evm, INTEGRITY_NOXATTRS if no protected
xattrs exist or INTEGRITY_UNKNOWN if xattrs are not supported. This would
limit the usability of EVM when only a public key is loaded, as commands
such as cp or tar with the option to preserve xattrs won't work.
This patch introduces the evm_hmac_disabled() function to determine whether
or not it is safe to ignore verification errors, based on the ability of
EVM to calculate HMACs. If the HMAC key is not loaded, and it cannot be
loaded in the future due to the EVM_SETUP_COMPLETE initialization flag,
allowing an operation despite the attrs/xattrs being found invalid will not
make them valid.
Since the post hooks can be executed even when the HMAC key is not loaded,
this patch also ensures that the EVM_INIT_HMAC initialization flag is set
before the post hooks call evm_update_evmxattr().
Signed-off-by: Roberto Sassu <roberto.sassu@huawei.com>
Suggested-by: Mimi Zohar <zohar@linux.ibm.com> (for ensuring EVM_INIT_HMAC is set)
Signed-off-by: Mimi Zohar <zohar@linux.ibm.com>
When EVM_ALLOW_METADATA_WRITES is set, EVM allows any operation on
metadata. Its main purpose is to allow users to freely set metadata when it
is protected by a portable signature, until an HMAC key is loaded.
However, callers of evm_verifyxattr() are not notified about metadata
changes and continue to rely on the last status returned by the function.
For example IMA, since it caches the appraisal result, will not call again
evm_verifyxattr() until the appraisal flags are cleared, and will grant
access to the file even if there was a metadata operation that made the
portable signature invalid.
This patch introduces evm_revalidate_status(), which callers of
evm_verifyxattr() can use in their xattr hooks to determine whether
re-validation is necessary and to do the proper actions. IMA calls it in
its xattr hooks to reset the appraisal flags, so that the EVM status is
re-evaluated after a metadata operation.
Lastly, this patch also adds a call to evm_reset_status() in
evm_inode_post_setattr() to invalidate the cached EVM status after a
setattr operation.
Signed-off-by: Roberto Sassu <roberto.sassu@huawei.com>
Signed-off-by: Mimi Zohar <zohar@linux.ibm.com>
EVM_ALLOW_METADATA_WRITES is an EVM initialization flag that can be set to
temporarily disable metadata verification until all xattrs/attrs necessary
to verify an EVM portable signature are copied to the file. This flag is
cleared when EVM is initialized with an HMAC key, to avoid that the HMAC is
calculated on unverified xattrs/attrs.
Currently EVM unnecessarily denies setting this flag if EVM is initialized
with a public key, which is not a concern as it cannot be used to trust
xattrs/attrs updates. This patch removes this limitation.
Fixes: ae1ba1676b ("EVM: Allow userland to permit modification of EVM-protected metadata")
Signed-off-by: Roberto Sassu <roberto.sassu@huawei.com>
Cc: stable@vger.kernel.org # 4.16.x
Signed-off-by: Mimi Zohar <zohar@linux.ibm.com>
The public builtin keys do not need to be appraised by IMA as the
restriction on the IMA/EVM trusted keyrings ensures that a key can be
loaded only if it is signed with a key on the builtin or secondary
keyrings.
However, when evm_load_x509() is called, appraisal is already enabled and
a valid IMA signature must be added to the EVM key to pass verification.
Since the restriction is applied on both IMA and EVM trusted keyrings, it
is safe to disable appraisal also when the EVM key is loaded. This patch
calls evm_load_x509() inside ima_load_x509() if CONFIG_IMA_LOAD_X509 is
enabled, which crosses the normal IMA and EVM boundary.
Signed-off-by: Roberto Sassu <roberto.sassu@huawei.com>
Signed-off-by: Mimi Zohar <zohar@linux.ibm.com>
evm_inode_init_security() requires an HMAC key to calculate the HMAC on
initial xattrs provided by LSMs. However, it checks generically whether a
key has been loaded, including also public keys, which is not correct as
public keys are not suitable to calculate the HMAC.
Originally, support for signature verification was introduced to verify a
possibly immutable initial ram disk, when no new files are created, and to
switch to HMAC for the root filesystem. By that time, an HMAC key should
have been loaded and usable to calculate HMACs for new files.
More recently support for requiring an HMAC key was removed from the
kernel, so that signature verification can be used alone. Since this is a
legitimate use case, evm_inode_init_security() should not return an error
when no HMAC key has been loaded.
This patch fixes this problem by replacing the evm_key_loaded() check with
a check of the EVM_INIT_HMAC flag in evm_initialized.
Fixes: 26ddabfe96 ("evm: enable EVM when X509 certificate is loaded")
Signed-off-by: Roberto Sassu <roberto.sassu@huawei.com>
Cc: stable@vger.kernel.org # 4.5.x
Signed-off-by: Mimi Zohar <zohar@linux.ibm.com>
EVM_SETUP_COMPLETE is defined as 0x80000000, which is larger than INT_MAX.
The "-fno-strict-overflow" compiler option properly prevents signaling
EVM that the EVM policy setup is complete. Define and read an unsigned
int.
Fixes: f00d797507 ("EVM: Allow userspace to signal an RSA key has been loaded")
Signed-off-by: Mimi Zohar <zohar@linux.ibm.com>
-----BEGIN PGP SIGNATURE-----
iQJIBAABCAAyFiEEjSMCCC7+cjo3nszSa3kkZrA+cVoFAmCIuMgUHHpvaGFyQGxp
bnV4LmlibS5jb20ACgkQa3kkZrA+cVqCew//SHwZ3LuqZUyqX1lCeW0eUsJQmf9P
Rc++pK5+cigaYwg/LpLXBtIthOO91KJ+p4UrXIttlz6wRWqH5Enm/nQk3ngIIrWJ
g4HsU4LoXeE5V1QZgDdUWM+ViaNvZuzT8enaUt4VvelHTRMoAgeTu7/amjx01t9L
R8Qhmg2tBI5ehRgocNfEApNfaDyOcJhR14wAnGFhQH1pH2o0e+O+L5E6ke3Z2N0D
oa9LBkA2wZFlTXkYHAYBdZjPW0BXX55brIusLHhrvFC16Ad9IVwcxRvRwYy4v2IK
p8kO7EwGO4IkqsuL6FApgbW8/dTZXo3pD8YmplXhQdvAd9L/LshYhDePX7UsEp+O
CL3rFvNFzZKC5qTx0UXLqjcktosOLTTFkAzRMg5taTljbTWBp9ziI8X0nTgoThLH
fTs41Ol5v5veiq8JbZ1VixCoDJVOWtED/FuSZD41OyXJb7yGNpTnmvP6QJNIa1Yo
vWsxSWmEX62xT85CzYLBR1Ow5+9t9+IjoCRt0pu0SRSps4gYcg1BjS+gIFnuCtB/
B9Ysh+rHo9ne9iojFd9TrEYQSAXebVEYSOjqPWV3E7gd0EXuF+JY0akj8JBduHt3
mfogxVtJdaHDjHtpHoHhBTpHG8l/E2n/47eFDmCiTsBGdL54R4jJSkOu8QmWBgzz
HVNX+f86SJ2haUw=
=gkFF
-----END PGP SIGNATURE-----
Merge tag 'integrity-v5.13' of git://git.kernel.org/pub/scm/linux/kernel/git/zohar/linux-integrity
Pull IMA updates from Mimi Zohar:
"In addition to loading the kernel module signing key onto the builtin
keyring, load it onto the IMA keyring as well.
Also six trivial changes and bug fixes"
* tag 'integrity-v5.13' of git://git.kernel.org/pub/scm/linux/kernel/git/zohar/linux-integrity:
ima: ensure IMA_APPRAISE_MODSIG has necessary dependencies
ima: Fix fall-through warnings for Clang
integrity: Add declarations to init_once void arguments.
ima: Fix function name error in comment.
ima: enable loading of build time generated key on .ima keyring
ima: enable signing of modules with build time generated key
keys: cleanup build time module signing keys
ima: Fix the error code for restoring the PCR value
ima: without an IMA policy loaded, return quickly
- Refactoring powerpc and arm64 kexec DT handling to common code. This
enables IMA on arm64.
- Add kbuild support for applying DT overlays at build time. The first
user are the DT unittests.
- Fix kerneldoc formatting and W=1 warnings in drivers/of/
- Fix handling 64-bit flag on PCI resources
- Bump dtschema version required to v2021.2.1
- Enable undocumented compatible checks for dtbs_check. This allows
tracking of missing binding schemas.
- DT docs improvements. Regroup the DT docs and add the example schema
and DT kernel ABI docs to the doc build.
- Convert Broadcom Bluetooth and video-mux bindings to schema
- Add QCom sm8250 Venus video codec binding schema
- Add vendor prefixes for AESOP, YIC System Co., Ltd, and Siliconfile
Technologies Inc.
- Cleanup of DT schema type references on common properties and
standard unit properties
-----BEGIN PGP SIGNATURE-----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=rF27
-----END PGP SIGNATURE-----
Merge tag 'devicetree-for-5.13' of git://git.kernel.org/pub/scm/linux/kernel/git/robh/linux
Pull devicetree updates from Rob Herring:
- Refactor powerpc and arm64 kexec DT handling to common code. This
enables IMA on arm64.
- Add kbuild support for applying DT overlays at build time. The first
user are the DT unittests.
- Fix kerneldoc formatting and W=1 warnings in drivers/of/
- Fix handling 64-bit flag on PCI resources
- Bump dtschema version required to v2021.2.1
- Enable undocumented compatible checks for dtbs_check. This allows
tracking of missing binding schemas.
- DT docs improvements. Regroup the DT docs and add the example schema
and DT kernel ABI docs to the doc build.
- Convert Broadcom Bluetooth and video-mux bindings to schema
- Add QCom sm8250 Venus video codec binding schema
- Add vendor prefixes for AESOP, YIC System Co., Ltd, and Siliconfile
Technologies Inc.
- Cleanup of DT schema type references on common properties and
standard unit properties
* tag 'devicetree-for-5.13' of git://git.kernel.org/pub/scm/linux/kernel/git/robh/linux: (64 commits)
powerpc: If kexec_build_elf_info() fails return immediately from elf64_load()
powerpc: Free fdt on error in elf64_load()
of: overlay: Fix kerneldoc warning in of_overlay_remove()
of: linux/of.h: fix kernel-doc warnings
of/pci: Add IORESOURCE_MEM_64 to resource flags for 64-bit memory addresses
dt-bindings: bcm4329-fmac: add optional brcm,ccode-map
docs: dt: update writing-schema.rst references
dt-bindings: media: venus: Add sm8250 dt schema
of: base: Fix spelling issue with function param 'prop'
docs: dt: Add DT API documentation
of: Add missing 'Return' section in kerneldoc comments
of: Fix kerneldoc output formatting
docs: dt: Group DT docs into relevant sub-sections
docs: dt: Make 'Devicetree' wording more consistent
docs: dt: writing-schema: Include the example schema in the doc build
docs: dt: writing-schema: Remove spurious indentation
dt-bindings: Fix reference in submitting-patches.rst to the DT ABI doc
dt-bindings: ddr: Add optional manufacturer and revision ID to LPDDR3
dt-bindings: media: video-interfaces: Drop the example
devicetree: bindings: clock: Minor typo fix in the file armada3700-tbg-clock.txt
...
-----BEGIN PGP SIGNATURE-----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=HTEy
-----END PGP SIGNATURE-----
Merge tag 'selinux-pr-20210426' of git://git.kernel.org/pub/scm/linux/kernel/git/pcmoore/selinux
Pull selinux updates from Paul Moore:
- Add support for measuring the SELinux state and policy capabilities
using IMA.
- A handful of SELinux/NFS patches to compare the SELinux state of one
mount with a set of mount options. Olga goes into more detail in the
patch descriptions, but this is important as it allows more
flexibility when using NFS and SELinux context mounts.
- Properly differentiate between the subjective and objective LSM
credentials; including support for the SELinux and Smack. My clumsy
attempt at a proper fix for AppArmor didn't quite pass muster so John
is working on a proper AppArmor patch, in the meantime this set of
patches shouldn't change the behavior of AppArmor in any way. This
change explains the bulk of the diffstat beyond security/.
- Fix a problem where we were not properly terminating the permission
list for two SELinux object classes.
* tag 'selinux-pr-20210426' of git://git.kernel.org/pub/scm/linux/kernel/git/pcmoore/selinux:
selinux: add proper NULL termination to the secclass_map permissions
smack: differentiate between subjective and objective task credentials
selinux: clarify task subjective and objective credentials
lsm: separate security_task_getsecid() into subjective and objective variants
nfs: account for selinux security context when deciding to share superblock
nfs: remove unneeded null check in nfs_fill_super()
lsm,selinux: add new hook to compare new mount to an existing mount
selinux: fix misspellings using codespell tool
selinux: fix misspellings using codespell tool
selinux: measure state and policy capabilities
selinux: Allow context mounts for unpriviliged overlayfs
Pull crypto updates from Herbert Xu:
"API:
- crypto_destroy_tfm now ignores errors as well as NULL pointers
Algorithms:
- Add explicit curve IDs in ECDH algorithm names
- Add NIST P384 curve parameters
- Add ECDSA
Drivers:
- Add support for Green Sardine in ccp
- Add ecdh/curve25519 to hisilicon/hpre
- Add support for AM64 in sa2ul"
* 'linus' of git://git.kernel.org/pub/scm/linux/kernel/git/herbert/crypto-2.6: (184 commits)
fsverity: relax build time dependency on CRYPTO_SHA256
fscrypt: relax Kconfig dependencies for crypto API algorithms
crypto: camellia - drop duplicate "depends on CRYPTO"
crypto: s5p-sss - consistently use local 'dev' variable in probe()
crypto: s5p-sss - remove unneeded local variable initialization
crypto: s5p-sss - simplify getting of_device_id match data
ccp: ccp - add support for Green Sardine
crypto: ccp - Make ccp_dev_suspend and ccp_dev_resume void functions
crypto: octeontx2 - add support for OcteonTX2 98xx CPT block.
crypto: chelsio/chcr - Remove useless MODULE_VERSION
crypto: ux500/cryp - Remove duplicate argument
crypto: chelsio - remove unused function
crypto: sa2ul - Add support for AM64
crypto: sa2ul - Support for per channel coherency
dt-bindings: crypto: ti,sa2ul: Add new compatible for AM64
crypto: hisilicon - enable new error types for QM
crypto: hisilicon - add new error type for SEC
crypto: hisilicon - support new error types for ZIP
crypto: hisilicon - dynamic configuration 'err_info'
crypto: doc - fix kernel-doc notation in chacha.c and af_alg.c
...
-----BEGIN PGP SIGNATURE-----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=fLQY
-----END PGP SIGNATURE-----
Merge tag 'keys-cve-2020-26541-v3' of git://git.kernel.org/pub/scm/linux/kernel/git/dhowells/linux-fs
Pull x509 dbx/mokx UEFI support from David Howells:
"Here's a set of patches from Eric Snowberg[1] that add support for
EFI_CERT_X509_GUID entries in the dbx and mokx UEFI tables (such
entries cause matching certificates to be rejected).
These are currently ignored and only the hash entries are made use of.
Additionally Eric included his patches to allow such certificates to
be preloaded.
These patches deal with CVE-2020-26541.
To quote Eric:
'This is the fifth patch series for adding support for
EFI_CERT_X509_GUID entries [2]. It has been expanded to not only
include dbx entries but also entries in the mokx. Additionally
my series to preload these certificate [3] has also been
included'"
Link: https://lore.kernel.org/r/20210122181054.32635-1-eric.snowberg@oracle.com [1]
Link: https://patchwork.kernel.org/project/linux-security-module/patch/20200916004927.64276-1-eric.snowberg@oracle.com/ [2]
Link: https://lore.kernel.org/patchwork/cover/1315485/ [3]
* tag 'keys-cve-2020-26541-v3' of git://git.kernel.org/pub/scm/linux/kernel/git/dhowells/linux-fs:
integrity: Load mokx variables into the blacklist keyring
certs: Add ability to preload revocation certs
certs: Move load_system_certificate_list to a common function
certs: Add EFI_CERT_X509_GUID support for dbx entries
In preparation to enable -Wimplicit-fallthrough for Clang, fix multiple
warnings by explicitly adding multiple break statements instead of just
letting the code fall through to the next case.
Link: https://github.com/KSPP/linux/issues/115
Signed-off-by: Gustavo A. R. Silva <gustavoars@kernel.org>
Signed-off-by: Mimi Zohar <zohar@linux.ibm.com>
init_once is a callback to kmem_cache_create. The parameter
type of this function is void *, so it's better to give a
explicit cast here.
Signed-off-by: Jiele Zhao <unclexiaole@gmail.com>
Signed-off-by: Mimi Zohar <zohar@linux.ibm.com>
The original function name was ima_path_check(). The policy parsing
still supports PATH_CHECK. Commit 9bbb6cad01 ("ima: rename
ima_path_check to ima_file_check") renamed the function to
ima_file_check(), but missed modifying the function name in the
comment.
Fixes: 9bbb6cad01 ("ima: rename ima_path_check to ima_file_check").
Signed-off-by: Jiele Zhao <unclexiaole@gmail.com>
Signed-off-by: Mimi Zohar <zohar@linux.ibm.com>
The kernel currently only loads the kernel module signing key onto the
builtin trusted keyring. Load the module signing key onto the IMA keyring
as well.
Signed-off-by: Nayna Jain <nayna@linux.ibm.com>
Acked-by: Stefan Berger <stefanb@linux.ibm.com>
Signed-off-by: Mimi Zohar <zohar@linux.ibm.com>
Add support for IMA signature verification for EC keys. Since SHA type
of hashes can be used by RSA and ECDSA signature schemes we need to
look at the key and derive from the key which signature scheme to use.
Since this can be applied to all types of keys, we change the selection
of the encoding type to be driven by the key's signature scheme rather
than by the hash type.
Cc: Dmitry Kasatkin <dmitry.kasatkin@gmail.com>
Cc: linux-integrity@vger.kernel.org
Cc: David Howells <dhowells@redhat.com>
Cc: keyrings@vger.kernel.org
Signed-off-by: Stefan Berger <stefanb@linux.ibm.com>
Reviewed-by: Vitaly Chikunov <vt@altlinux.org>
Reviewed-by: Tianjia Zhang <tianjia.zhang@linux.alibaba.com>
Acked-by: Mimi Zohar <zohar@linux.ibm.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
In ima_restore_measurement_list(), hdr[HDR_PCR].data is pointing to a
buffer of type u8, which contains the dumped 32-bit pcr value.
Currently, only the least significant byte is used to restore the pcr
value. We should convert hdr[HDR_PCR].data to a pointer of type u32
before fetching the value to restore the correct pcr value.
Fixes: 47fdee60b4 ("ima: use ima_parse_buf() to parse measurements headers")
Signed-off-by: Li Huafei <lihuafei1@huawei.com>
Reviewed-by: Roberto Sassu <roberto.sassu@huawei.com>
Signed-off-by: Mimi Zohar <zohar@linux.ibm.com>
Of the three LSMs that implement the security_task_getsecid() LSM
hook, all three LSMs provide the task's objective security
credentials. This turns out to be unfortunate as most of the hook's
callers seem to expect the task's subjective credentials, although
a small handful of callers do correctly expect the objective
credentials.
This patch is the first step towards fixing the problem: it splits
the existing security_task_getsecid() hook into two variants, one
for the subjective creds, one for the objective creds.
void security_task_getsecid_subj(struct task_struct *p,
u32 *secid);
void security_task_getsecid_obj(struct task_struct *p,
u32 *secid);
While this patch does fix all of the callers to use the correct
variant, in order to keep this patch focused on the callers and to
ease review, the LSMs continue to use the same implementation for
both hooks. The net effect is that this patch should not change
the behavior of the kernel in any way, it will be up to the latter
LSM specific patches in this series to change the hook
implementations and return the correct credentials.
Acked-by: Mimi Zohar <zohar@linux.ibm.com> (IMA)
Acked-by: Casey Schaufler <casey@schaufler-ca.com>
Reviewed-by: Richard Guy Briggs <rgb@redhat.com>
Signed-off-by: Paul Moore <paul@paul-moore.com>
The functions defined in "arch/powerpc/kexec/ima.c" handle setting up
and freeing the resources required to carry over the IMA measurement
list from the current kernel to the next kernel across kexec system call.
These functions do not have architecture specific code, but are
currently limited to powerpc.
Move remove_ima_buffer() and setup_ima_buffer() calls into
of_kexec_alloc_and_setup_fdt() defined in "drivers/of/kexec.c".
Move the remaining architecture independent functions from
"arch/powerpc/kexec/ima.c" to "drivers/of/kexec.c".
Delete "arch/powerpc/kexec/ima.c" and "arch/powerpc/include/asm/ima.h".
Remove references to the deleted files and functions in powerpc and
in ima.
Co-developed-by: Prakhar Srivastava <prsriva@linux.microsoft.com>
Signed-off-by: Prakhar Srivastava <prsriva@linux.microsoft.com>
Signed-off-by: Lakshmi Ramasubramanian <nramas@linux.microsoft.com>
Reviewed-by: Thiago Jung Bauermann <bauerman@linux.ibm.com>
Tested-by: Thiago Jung Bauermann <bauerman@linux.ibm.com>
Signed-off-by: Rob Herring <robh@kernel.org>
Link: https://lore.kernel.org/r/20210221174930.27324-11-nramas@linux.microsoft.com
The fields ima_buffer_addr and ima_buffer_size in "struct kimage_arch"
for powerpc are used to carry forward the IMA measurement list across
kexec system call. These fields are not architecture specific, but are
currently limited to powerpc.
arch_ima_add_kexec_buffer() defined in "arch/powerpc/kexec/ima.c"
sets ima_buffer_addr and ima_buffer_size for the kexec system call.
This function does not have architecture specific code, but is
currently limited to powerpc.
Move ima_buffer_addr and ima_buffer_size to "struct kimage".
Set ima_buffer_addr and ima_buffer_size in ima_add_kexec_buffer()
in security/integrity/ima/ima_kexec.c.
Co-developed-by: Prakhar Srivastava <prsriva@linux.microsoft.com>
Signed-off-by: Prakhar Srivastava <prsriva@linux.microsoft.com>
Signed-off-by: Lakshmi Ramasubramanian <nramas@linux.microsoft.com>
Suggested-by: Will Deacon <will@kernel.org>
Reviewed-by: Thiago Jung Bauermann <bauerman@linux.ibm.com>
Signed-off-by: Rob Herring <robh@kernel.org>
Link: https://lore.kernel.org/r/20210221174930.27324-9-nramas@linux.microsoft.com
-----BEGIN PGP SIGNATURE-----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=HCrJ
-----END PGP SIGNATURE-----
Merge tag 'keys-misc-20210126' of git://git.kernel.org/pub/scm/linux/kernel/git/dhowells/linux-fs
Pull keyring updates from David Howells:
"Here's a set of minor keyrings fixes/cleanups that I've collected from
various people for the upcoming merge window.
A couple of them might, in theory, be visible to userspace:
- Make blacklist_vet_description() reject uppercase letters as they
don't match the all-lowercase hex string generated for a blacklist
search.
This may want reconsideration in the future, but, currently, you
can't add to the blacklist keyring from userspace and the only
source of blacklist keys generates lowercase descriptions.
- Fix blacklist_init() to use a new KEY_ALLOC_* flag to indicate that
it wants KEY_FLAG_KEEP to be set rather than passing KEY_FLAG_KEEP
into keyring_alloc() as KEY_FLAG_KEEP isn't a valid alloc flag.
This isn't currently a problem as the blacklist keyring isn't
currently writable by userspace.
The rest of the patches are cleanups and I don't think they should
have any visible effect"
* tag 'keys-misc-20210126' of git://git.kernel.org/pub/scm/linux/kernel/git/dhowells/linux-fs:
watch_queue: rectify kernel-doc for init_watch()
certs: Replace K{U,G}IDT_INIT() with GLOBAL_ROOT_{U,G}ID
certs: Fix blacklist flag type confusion
PKCS#7: Fix missing include
certs: Fix blacklisted hexadecimal hash string check
certs/blacklist: fix kernel doc interface issue
crypto: public_key: Remove redundant header file from public_key.h
keys: remove trailing semicolon in macro definition
crypto: pkcs7: Use match_string() helper to simplify the code
PKCS#7: drop function from kernel-doc pkcs7_validate_trust_one
encrypted-keys: Replace HTTP links with HTTPS ones
crypto: asymmetric_keys: fix some comments in pkcs7_parser.h
KEYS: remove redundant memset
security: keys: delete repeated words in comments
KEYS: asymmetric: Fix kerneldoc
security/keys: use kvfree_sensitive()
watch_queue: Drop references to /dev/watch_queue
keys: Remove outdated __user annotations
security: keys: Fix fall-through warnings for Clang
-----BEGIN PGP SIGNATURE-----
iHUEABYKAB0WIQRAhzRXHqcMeLMyaSiRxhvAZXjcogUCYCegywAKCRCRxhvAZXjc
ouJ6AQDlf+7jCQlQdeKKoN9QDFfMzG1ooemat36EpRRTONaGuAD8D9A4sUsG4+5f
4IU5Lj9oY4DEmF8HenbWK2ZHsesL2Qg=
=yPaw
-----END PGP SIGNATURE-----
Merge tag 'idmapped-mounts-v5.12' of git://git.kernel.org/pub/scm/linux/kernel/git/brauner/linux
Pull idmapped mounts from Christian Brauner:
"This introduces idmapped mounts which has been in the making for some
time. Simply put, different mounts can expose the same file or
directory with different ownership. This initial implementation comes
with ports for fat, ext4 and with Christoph's port for xfs with more
filesystems being actively worked on by independent people and
maintainers.
Idmapping mounts handle a wide range of long standing use-cases. Here
are just a few:
- Idmapped mounts make it possible to easily share files between
multiple users or multiple machines especially in complex
scenarios. For example, idmapped mounts will be used in the
implementation of portable home directories in
systemd-homed.service(8) where they allow users to move their home
directory to an external storage device and use it on multiple
computers where they are assigned different uids and gids. This
effectively makes it possible to assign random uids and gids at
login time.
- It is possible to share files from the host with unprivileged
containers without having to change ownership permanently through
chown(2).
- It is possible to idmap a container's rootfs and without having to
mangle every file. For example, Chromebooks use it to share the
user's Download folder with their unprivileged containers in their
Linux subsystem.
- It is possible to share files between containers with
non-overlapping idmappings.
- Filesystem that lack a proper concept of ownership such as fat can
use idmapped mounts to implement discretionary access (DAC)
permission checking.
- They allow users to efficiently changing ownership on a per-mount
basis without having to (recursively) chown(2) all files. In
contrast to chown (2) changing ownership of large sets of files is
instantenous with idmapped mounts. This is especially useful when
ownership of a whole root filesystem of a virtual machine or
container is changed. With idmapped mounts a single syscall
mount_setattr syscall will be sufficient to change the ownership of
all files.
- Idmapped mounts always take the current ownership into account as
idmappings specify what a given uid or gid is supposed to be mapped
to. This contrasts with the chown(2) syscall which cannot by itself
take the current ownership of the files it changes into account. It
simply changes the ownership to the specified uid and gid. This is
especially problematic when recursively chown(2)ing a large set of
files which is commong with the aforementioned portable home
directory and container and vm scenario.
- Idmapped mounts allow to change ownership locally, restricting it
to specific mounts, and temporarily as the ownership changes only
apply as long as the mount exists.
Several userspace projects have either already put up patches and
pull-requests for this feature or will do so should you decide to pull
this:
- systemd: In a wide variety of scenarios but especially right away
in their implementation of portable home directories.
https://systemd.io/HOME_DIRECTORY/
- container runtimes: containerd, runC, LXD:To share data between
host and unprivileged containers, unprivileged and privileged
containers, etc. The pull request for idmapped mounts support in
containerd, the default Kubernetes runtime is already up for quite
a while now: https://github.com/containerd/containerd/pull/4734
- The virtio-fs developers and several users have expressed interest
in using this feature with virtual machines once virtio-fs is
ported.
- ChromeOS: Sharing host-directories with unprivileged containers.
I've tightly synced with all those projects and all of those listed
here have also expressed their need/desire for this feature on the
mailing list. For more info on how people use this there's a bunch of
talks about this too. Here's just two recent ones:
https://www.cncf.io/wp-content/uploads/2020/12/Rootless-Containers-in-Gitpod.pdfhttps://fosdem.org/2021/schedule/event/containers_idmap/
This comes with an extensive xfstests suite covering both ext4 and
xfs:
https://git.kernel.org/brauner/xfstests-dev/h/idmapped_mounts
It covers truncation, creation, opening, xattrs, vfscaps, setid
execution, setgid inheritance and more both with idmapped and
non-idmapped mounts. It already helped to discover an unrelated xfs
setgid inheritance bug which has since been fixed in mainline. It will
be sent for inclusion with the xfstests project should you decide to
merge this.
In order to support per-mount idmappings vfsmounts are marked with
user namespaces. The idmapping of the user namespace will be used to
map the ids of vfs objects when they are accessed through that mount.
By default all vfsmounts are marked with the initial user namespace.
The initial user namespace is used to indicate that a mount is not
idmapped. All operations behave as before and this is verified in the
testsuite.
Based on prior discussions we want to attach the whole user namespace
and not just a dedicated idmapping struct. This allows us to reuse all
the helpers that already exist for dealing with idmappings instead of
introducing a whole new range of helpers. In addition, if we decide in
the future that we are confident enough to enable unprivileged users
to setup idmapped mounts the permission checking can take into account
whether the caller is privileged in the user namespace the mount is
currently marked with.
The user namespace the mount will be marked with can be specified by
passing a file descriptor refering to the user namespace as an
argument to the new mount_setattr() syscall together with the new
MOUNT_ATTR_IDMAP flag. The system call follows the openat2() pattern
of extensibility.
The following conditions must be met in order to create an idmapped
mount:
- The caller must currently have the CAP_SYS_ADMIN capability in the
user namespace the underlying filesystem has been mounted in.
- The underlying filesystem must support idmapped mounts.
- The mount must not already be idmapped. This also implies that the
idmapping of a mount cannot be altered once it has been idmapped.
- The mount must be a detached/anonymous mount, i.e. it must have
been created by calling open_tree() with the OPEN_TREE_CLONE flag
and it must not already have been visible in the filesystem.
The last two points guarantee easier semantics for userspace and the
kernel and make the implementation significantly simpler.
By default vfsmounts are marked with the initial user namespace and no
behavioral or performance changes are observed.
The manpage with a detailed description can be found here:
1d7b902e28
In order to support idmapped mounts, filesystems need to be changed
and mark themselves with the FS_ALLOW_IDMAP flag in fs_flags. The
patches to convert individual filesystem are not very large or
complicated overall as can be seen from the included fat, ext4, and
xfs ports. Patches for other filesystems are actively worked on and
will be sent out separately. The xfstestsuite can be used to verify
that port has been done correctly.
The mount_setattr() syscall is motivated independent of the idmapped
mounts patches and it's been around since July 2019. One of the most
valuable features of the new mount api is the ability to perform
mounts based on file descriptors only.
Together with the lookup restrictions available in the openat2()
RESOLVE_* flag namespace which we added in v5.6 this is the first time
we are close to hardened and race-free (e.g. symlinks) mounting and
path resolution.
While userspace has started porting to the new mount api to mount
proper filesystems and create new bind-mounts it is currently not
possible to change mount options of an already existing bind mount in
the new mount api since the mount_setattr() syscall is missing.
With the addition of the mount_setattr() syscall we remove this last
restriction and userspace can now fully port to the new mount api,
covering every use-case the old mount api could. We also add the
crucial ability to recursively change mount options for a whole mount
tree, both removing and adding mount options at the same time. This
syscall has been requested multiple times by various people and
projects.
There is a simple tool available at
https://github.com/brauner/mount-idmapped
that allows to create idmapped mounts so people can play with this
patch series. I'll add support for the regular mount binary should you
decide to pull this in the following weeks:
Here's an example to a simple idmapped mount of another user's home
directory:
u1001@f2-vm:/$ sudo ./mount --idmap both:1000:1001:1 /home/ubuntu/ /mnt
u1001@f2-vm:/$ ls -al /home/ubuntu/
total 28
drwxr-xr-x 2 ubuntu ubuntu 4096 Oct 28 22:07 .
drwxr-xr-x 4 root root 4096 Oct 28 04:00 ..
-rw------- 1 ubuntu ubuntu 3154 Oct 28 22:12 .bash_history
-rw-r--r-- 1 ubuntu ubuntu 220 Feb 25 2020 .bash_logout
-rw-r--r-- 1 ubuntu ubuntu 3771 Feb 25 2020 .bashrc
-rw-r--r-- 1 ubuntu ubuntu 807 Feb 25 2020 .profile
-rw-r--r-- 1 ubuntu ubuntu 0 Oct 16 16:11 .sudo_as_admin_successful
-rw------- 1 ubuntu ubuntu 1144 Oct 28 00:43 .viminfo
u1001@f2-vm:/$ ls -al /mnt/
total 28
drwxr-xr-x 2 u1001 u1001 4096 Oct 28 22:07 .
drwxr-xr-x 29 root root 4096 Oct 28 22:01 ..
-rw------- 1 u1001 u1001 3154 Oct 28 22:12 .bash_history
-rw-r--r-- 1 u1001 u1001 220 Feb 25 2020 .bash_logout
-rw-r--r-- 1 u1001 u1001 3771 Feb 25 2020 .bashrc
-rw-r--r-- 1 u1001 u1001 807 Feb 25 2020 .profile
-rw-r--r-- 1 u1001 u1001 0 Oct 16 16:11 .sudo_as_admin_successful
-rw------- 1 u1001 u1001 1144 Oct 28 00:43 .viminfo
u1001@f2-vm:/$ touch /mnt/my-file
u1001@f2-vm:/$ setfacl -m u:1001:rwx /mnt/my-file
u1001@f2-vm:/$ sudo setcap -n 1001 cap_net_raw+ep /mnt/my-file
u1001@f2-vm:/$ ls -al /mnt/my-file
-rw-rwxr--+ 1 u1001 u1001 0 Oct 28 22:14 /mnt/my-file
u1001@f2-vm:/$ ls -al /home/ubuntu/my-file
-rw-rwxr--+ 1 ubuntu ubuntu 0 Oct 28 22:14 /home/ubuntu/my-file
u1001@f2-vm:/$ getfacl /mnt/my-file
getfacl: Removing leading '/' from absolute path names
# file: mnt/my-file
# owner: u1001
# group: u1001
user::rw-
user:u1001:rwx
group::rw-
mask::rwx
other::r--
u1001@f2-vm:/$ getfacl /home/ubuntu/my-file
getfacl: Removing leading '/' from absolute path names
# file: home/ubuntu/my-file
# owner: ubuntu
# group: ubuntu
user::rw-
user:ubuntu:rwx
group::rw-
mask::rwx
other::r--"
* tag 'idmapped-mounts-v5.12' of git://git.kernel.org/pub/scm/linux/kernel/git/brauner/linux: (41 commits)
xfs: remove the possibly unused mp variable in xfs_file_compat_ioctl
xfs: support idmapped mounts
ext4: support idmapped mounts
fat: handle idmapped mounts
tests: add mount_setattr() selftests
fs: introduce MOUNT_ATTR_IDMAP
fs: add mount_setattr()
fs: add attr_flags_to_mnt_flags helper
fs: split out functions to hold writers
namespace: only take read lock in do_reconfigure_mnt()
mount: make {lock,unlock}_mount_hash() static
namespace: take lock_mount_hash() directly when changing flags
nfs: do not export idmapped mounts
overlayfs: do not mount on top of idmapped mounts
ecryptfs: do not mount on top of idmapped mounts
ima: handle idmapped mounts
apparmor: handle idmapped mounts
fs: make helpers idmap mount aware
exec: handle idmapped mounts
would_dump: handle idmapped mounts
...
The sparse tool complains as follows:
security/integrity/digsig.c:146:12: warning:
symbol 'integrity_add_key' was not declared. Should it be static?
This function is not used outside of digsig.c, so this
commit marks it static.
Reported-by: Hulk Robot <hulkci@huawei.com>
Fixes: 60740accf7 ("integrity: Load certs to the platform keyring")
Signed-off-by: Wei Yongjun <weiyongjun1@huawei.com>
Reviewed-by: Kees Cook <keescook@chromium.org>
Reviewed-by: Nayna Jain <nayna@linux.ibm.com>
Signed-off-by: Mimi Zohar <zohar@linux.ibm.com>
Roberto Sassu
Mimi Zohar
Tushar Sugandhi
Lakshmi Ramasubramanian
Gustavo A. R. Silva
Linus Torvalds
Jiele Zhao
Nayna Jain
Stefan Berger
Li Huafei
Paul Moore
Eric Snowberg
Wei Yongjun