IMA protects files by storing a hash (or a signature thereof) of their content in the security.ima xattr. While the security.ima xattr itself is protected by EVM with either a HMAC or a digital signature, no mechanism is currently in place to ensure that the security.ima xattr was generated with a strong digest algorithm, as was outlined in https://lore.kernel.org/linux-integrity/10dde047d76b447f32ca91356599be679b8a76e5.camel@xxxxxxxxxxxxx/t/#m0f8127c6982ef94aa42f5cc13ea83b9f9000917e One important point is safeguarding users from mislabelling their files when using userland utilities to update their files, as this is the kind of behavior one can observe with evmctl (`evmctl ima_hash` defaults to sha1). Another group that may be interested is those that have deployed IMA years ago, possibly using algorithms that was then deemed sufficiently collision-resistant, but that proved to be weak with the passage of time (note that this could also happen in the future with algorithms considered safe today). This patch provides a migration path of sorts for these users. This patch series gives users the ability to restrict the algorithms accepted by their system, both when writing/updating xattrs, and when appraising files, while retaining a permissive behavior by default to preserve backward compatibility. To provide these features, alter the behavior of setxattr to only accept hashes built in the kernel, instead of any hash listed in the kernel (complete list crypto/hash_info.c). In addition, the user can define in his IMA policy the list of digest algorithms allowed for writing to the security.ima xattr. In that case, only algorithms present in that list are accepted for writing. In addition, users may opt-in to whitelisting the hash algorithms accepted when appraising thanks to the new "appraise_hash" IMA policy option. By default IMA will keep accepting any hash algorithm, but specifying that option will make appraisal of files hashed with another algorithm fail. Even when using this option to restrict accepted hashes, a migration to a new algorithm is still possible. Suppose your policy states you must migrate from 'old_algo' (e.g. sha1) to 'new_algo' (e.g. one of sha256/384/512). You can upgrade without relaxing the hash requirements: alter your policy rules from 'appraise_hash=old_algo' to 'appraise_hash=old_algo,new_algo', update the "ima_hash" parameter to 'new_algo', reboot, relabel all your files with 'new_algo', alter your policy_rule from 'appraise_hash=old_algo,new_algo' to 'appraise_hash=new_algo', reboot again and you're done. Agreed, it's quite a lot of churn - I don't know if this can be reduced - but this is technically doable. This series is based on the following repo/branch: repo: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git branch: master commit ff1176468d368232b684f75e82563369208bc371 ("Linux 5.14-rc3") Changelog since v3: - fixed an issue where the first write to the policy would ignore the SETXATTR_CHECK attribute - fixed potential concurrency issues (I would greatly like external opinions on this, because I clearly don't know much about RCU. Beside maybe it's better to completely ignore the duplicates SETXATTR_CHECK issue and not update the IMA policy in any case) - remove the CONFIG_CRYPTO_MD5 requirement for IMA (suggested by Mimi Zohar) - updated commit messages to follow more closely the kernel style guide (suggested by Mimi Zohar) - moved the hash verification code on appraisal a bit later, to prevent issues when using the code with IMA in a disable/auditing mode (suggested by Mimi Zohar) - limit the 'appraise_hash' parameter to the 'appraise' action (suggested by Mimi Zohar) Changelog since v2: - remove the SecureBoot-specific behavior (suggested by Mimi Zohar) - users can now tweak through policy both the algorithms for appraising files (a feature already present in v2) and for writing with the new SETXATTR_CHECK value for the 'func' ima policy flag - updating 'forbidden-hash-algorithm' to 'denied-hash-algorithm' and 'unsupported-hash-algorithm' to disambiguate cases when the user asked for an algorithm not present in the kernel and when the system vendor explicitly opted in to a restricted list of accepted algorithms (suggested by Mimi Zohar) - change the order of the patches to be bisect-safe while retaining the guarantee that a policy cannot be accepted but not enforced (suggested by Mimi Zohar) Changelog since v1: - Remove the two boot parameters (suggested by Mimi Zohar) - filter out hash algorithms not compiled in the kernel on xattr writes (suggested by Mimi Zohar) - add a special case when secure boot is enabled: only the ima_hash algorithm is accepted on userland writes - add a policy option to opt-in to restricting digest algorithms at a per-rule granularity (suggested by Mimi Zohar) Simon Thoby (4): IMA: block writes of the security.ima xattr with unsupported algorithms IMA: add support to restrict the hash algorithms used for file appraisal IMA: add a policy option to restrict xattr hash algorithms on appraisal IMA: introduce a new policy option func=SETXATTR_CHECK Simon Thoby (4): IMA: block writes of the security.ima xattr with unsupported algorithms IMA: add support to restrict the hash algorithms used for file appraisal IMA: add a policy option to restrict xattr hash algorithms on appraisal IMA: introduce a new policy option func=SETXATTR_CHECK Simon Thoby (5): IMA: remove the dependency on CRYPTO_MD5 IMA: block writes of the security.ima xattr with unsupported algorithms IMA: add support to restrict the hash algorithms used for file appraisal IMA: add a policy option to restrict xattr hash algorithms on appraisal IMA: introduce a new policy option func=SETXATTR_CHECK Documentation/ABI/testing/ima_policy | 15 ++- security/integrity/ima/Kconfig | 1 - security/integrity/ima/ima.h | 10 +- security/integrity/ima/ima_api.c | 6 +- security/integrity/ima/ima_appraise.c | 79 +++++++++++- security/integrity/ima/ima_main.c | 19 ++- security/integrity/ima/ima_policy.c | 168 +++++++++++++++++++++++++- 7 files changed, 278 insertions(+), 20 deletions(-) -- 2.31.1
