| Message ID | 20210726171319.3133879-1-eric.snowberg@oracle.com (mailing list archive) |
|---|---|
| Headers | show |
| Series | Enroll kernel keys thru MOK | expand |
Hi Eric, On Mon, 2021-07-26 at 13:13 -0400, Eric Snowberg wrote: > When the kernel boots, if MokListTrustedRT is set and > EFI_VARIABLE_NON_VOLATILE is not set, the MokListRT is loaded into the > mok keyring instead of the platform keyring. Mimi has suggested that > only CA keys or keys that can be vouched for by other kernel keys be > loaded into this keyring. All other certs will load into the platform > keyring instead. I suggested only loading the CA keys stored in the MOK db onto the MOK keyring. Like the builtin trusted keyring, the MOK keyring would also be linked to the secondary keyring. Assuming the secondary keyring is defined, all other properly signed MOK db keys - signed by keys on the builtin, secondary or MOK keyring - would be loaded onto the secondary keyring. As previously discussed, this might require reading the MOK db twice - once to load the CA keys on the MOK keyring, a second time to load the remaining properly signed keys onto the secondary keyring. thanks, Mimi
> On Aug 3, 2021, at 11:01 AM, Mimi Zohar <zohar@linux.ibm.com> wrote: > > On Mon, 2021-07-26 at 13:13 -0400, Eric Snowberg wrote: > >> When the kernel boots, if MokListTrustedRT is set and >> EFI_VARIABLE_NON_VOLATILE is not set, the MokListRT is loaded into the >> mok keyring instead of the platform keyring. Mimi has suggested that >> only CA keys or keys that can be vouched for by other kernel keys be >> loaded into this keyring. All other certs will load into the platform >> keyring instead. > > I suggested only loading the CA keys stored in the MOK db onto the MOK > keyring. Like the builtin trusted keyring, the MOK keyring would also > be linked to the secondary keyring. Assuming the secondary keyring is > defined, all other properly signed MOK db keys - signed by keys on the > builtin, secondary or MOK keyring - would be loaded onto the secondary > keyring. > > As previously discussed, this might require reading the MOK db twice - > once to load the CA keys on the MOK keyring, a second time to load the > remaining properly signed keys onto the secondary keyring. I’m only loading CA keys or keys that can be vouched for by other kernel keys into the new mok keyring. Currently, I’m not doing another pass. I could add another pass, but it would not solve the issue with someone trying to load an intermediate CA along with a leaf cert. This would require yet a third pass. I wasn’t sure if this added complexity was necessary. Currently, any CA contained within the MOK db would now be trusted by the kernel. Someone using a kernel with the secondary keyring enabled could load the intermediate and leaf certs themselves following boot. Taking this into account, if you’d like to see two passes, let me know and I’ll add that in v3. If a second pass is done, do you really want these additional keys added to the secondary keyring or should they go into the mok keyring instead? I was under the impression the secondary should be empty until a user adds their own keys into it. Thanks.
Hi Eric, On Tue, 2021-08-03 at 13:52 -0600, Eric Snowberg wrote: > > On Aug 3, 2021, at 11:01 AM, Mimi Zohar <zohar@linux.ibm.com> wrote: > > > > On Mon, 2021-07-26 at 13:13 -0400, Eric Snowberg wrote: > > > >> When the kernel boots, if MokListTrustedRT is set and > >> EFI_VARIABLE_NON_VOLATILE is not set, the MokListRT is loaded into the > >> mok keyring instead of the platform keyring. Mimi has suggested that > >> only CA keys or keys that can be vouched for by other kernel keys be > >> loaded into this keyring. All other certs will load into the platform > >> keyring instead. > > > > I suggested only loading the CA keys stored in the MOK db onto the MOK > > keyring. Like the builtin trusted keyring, the MOK keyring would also > > be linked to the secondary keyring. Assuming the secondary keyring is > > defined, all other properly signed MOK db keys - signed by keys on the > > builtin, secondary or MOK keyring - would be loaded onto the secondary > > keyring. > > > > As previously discussed, this might require reading the MOK db twice - > > once to load the CA keys on the MOK keyring, a second time to load the > > remaining properly signed keys onto the secondary keyring. > > I’m only loading CA keys or keys that can be vouched for by other kernel > keys into the new mok keyring. The cover letter implies that this suggestion is coming from me, which it definitely is not. My preference, as I made clear from the very beginning, is to load ONLY the MOK DB CA keys onto the mok keyring. (And even go one step farther, requiring the MOK DB CA key(s) to be identified on the boot command line.) > Currently, I’m not doing another pass. I > could add another pass, but it would not solve the issue with someone trying > to load an intermediate CA along with a leaf cert. This would require yet > a third pass. I wasn’t sure if this added complexity was necessary. > > Currently, any CA contained within the MOK db would now be trusted by the > kernel. Someone using a kernel with the secondary keyring enabled could > load the intermediate and leaf certs themselves following boot. Correct, as previously discussed, the other signed MOK DB keys may be loaded by userspace. The only reason we're interested in any of the other MOK DB keys is prevent a regression. As you previously pointed out all of the MOK DB keys are currently being loaded onto the platform keyring. So leave the existing code, which loads the MOK DB keys onto the platform keyring, alone to prevent that regression. It's already being controlled by a UEFI variable. > Taking > this into account, if you’d like to see two passes, let me know and I’ll add > that in v3. If a second pass is done, do you really want these additional > keys added to the secondary keyring or should they go into the mok keyring > instead? I was under the impression the secondary should be empty until a > user adds their own keys into it. Thanks. Again, my preference would be to load ONLY the MOK DB CA keys onto the mok keyring. If YOU decide you want to load the signed keys stored in MOK DB, be my guest. However, they should be loaded onto the secondary keyring and a new restriction defined, similar to "restrict_link_by_builtin_and_secondary_trusted", which includes mok as well. thanks, Mimi
> On Aug 3, 2021, at 7:14 PM, Mimi Zohar <zohar@linux.ibm.com> wrote: > > On Tue, 2021-08-03 at 13:52 -0600, Eric Snowberg wrote: >>> On Aug 3, 2021, at 11:01 AM, Mimi Zohar <zohar@linux.ibm.com> wrote: >>> >>> On Mon, 2021-07-26 at 13:13 -0400, Eric Snowberg wrote: >>> >>>> When the kernel boots, if MokListTrustedRT is set and >>>> EFI_VARIABLE_NON_VOLATILE is not set, the MokListRT is loaded into the >>>> mok keyring instead of the platform keyring. Mimi has suggested that >>>> only CA keys or keys that can be vouched for by other kernel keys be >>>> loaded into this keyring. All other certs will load into the platform >>>> keyring instead. >>> >>> I suggested only loading the CA keys stored in the MOK db onto the MOK >>> keyring. Like the builtin trusted keyring, the MOK keyring would also >>> be linked to the secondary keyring. Assuming the secondary keyring is >>> defined, all other properly signed MOK db keys - signed by keys on the >>> builtin, secondary or MOK keyring - would be loaded onto the secondary >>> keyring. >>> >>> As previously discussed, this might require reading the MOK db twice - >>> once to load the CA keys on the MOK keyring, a second time to load the >>> remaining properly signed keys onto the secondary keyring. >> >> I’m only loading CA keys or keys that can be vouched for by other kernel >> keys into the new mok keyring. > > The cover letter implies that this suggestion is coming from me, which > it definitely is not. My preference, as I made clear from the very > beginning, is to load ONLY the MOK DB CA keys onto the mok > keyring. (And even go one step farther, requiring the MOK DB CA > key(s) to be identified on the boot command line.) Ok, got it. I guess I misunderstood and was thinking built-in should be referenced too for things going into the new mok keyring. >> Currently, I’m not doing another pass. I >> could add another pass, but it would not solve the issue with someone trying >> to load an intermediate CA along with a leaf cert. This would require yet >> a third pass. I wasn’t sure if this added complexity was necessary. >> >> Currently, any CA contained within the MOK db would now be trusted by the >> kernel. Someone using a kernel with the secondary keyring enabled could >> load the intermediate and leaf certs themselves following boot. > > Correct, as previously discussed, the other signed MOK DB keys may be > loaded by userspace. The only reason we're interested in any of the > other MOK DB keys is prevent a regression. As you previously pointed > out all of the MOK DB keys are currently being loaded onto the platform > keyring. So leave the existing code, which loads the MOK DB keys onto > the platform keyring, alone to prevent that regression. It's already > being controlled by a UEFI variable. With this series, I do not believe a regression exists. With a single pass, keys are either loaded into the mok or the platform keyring. Since the mok is linked to the secondary (or the built-in), during kexec signature validation, all keys are referenced. >> Taking >> this into account, if you’d like to see two passes, let me know and I’ll add >> that in v3. If a second pass is done, do you really want these additional >> keys added to the secondary keyring or should they go into the mok keyring >> instead? I was under the impression the secondary should be empty until a >> user adds their own keys into it. Thanks. > > Again, my preference would be to load ONLY the MOK DB CA keys onto the > mok keyring. Ok, I’ll update the current code to just load CA keys into the mok in v3. This would simplify the new restrict_link_by_ca function. With that change, do you see any issues with how I’m doing the linking? With the mok keyring linked to the secondary keyring, the platform keyring may only contain a subset of the keys it originally contained. But, as I described above, I don’t believe it will lead to a regression since all keys get referenced. Thanks.
On Wed, 2021-08-04 at 02:56 +0000, Eric Snowberg wrote: > Ok, I’ll update the current code to just load CA keys into the mok in v3. This would > simplify the new restrict_link_by_ca function. Thank you! > > With that change, do you see any issues with how I’m doing the linking? With the > mok keyring linked to the secondary keyring, the platform keyring may only contain > a subset of the keys it originally contained. But, as I described above, I don’t believe > it will lead to a regression since all keys get referenced. Thanks. I think there is a problem. Only the builtin keys should ever be on the builtin keyring. The builtin keyring would need to be linked to the mok keyring. But in the secondary keyring case, the linking should be the reverse, where the mok keyring would be linked to the secondary keyring, similar to how the builtin keyring is linked to the secondary keyring. if (key_link(secondary_trusted_keys, builtin_trusted_keys) < 0) panic("Can't link trusted keyrings\n"); thanks, Mimi
Many UEFI Linux distributions boot using shim. The UEFI shim provides what is called Machine Owner Keys (MOK). Shim uses both the UEFI Secure Boot DB and MOK keys to validate the next step in the boot chain. The MOK facility can be used to import user generated keys. These keys can be used to sign an end-user development kernel build. When Linux boots, pre-boot keys (both UEFI Secure Boot DB and MOK keys) get loaded in the Linux .platform keyring. Currently, pre-boot keys are not trusted within the Linux trust boundary [1]. These platform keys can only be used for kexec. If an end-user wants to use their own key within the Linux trust boundary, they must either compile it into the kernel themselves or use the insert-sys-cert script. Both options present a problem. Many end-users do not want to compile their own kernels. With the insert-sys-cert option, there are missing upstream changes [2]. Also, with the insert-sys-cert option, the end-user must re-sign their kernel again with their own key, and then insert that key into the MOK db. Another problem with insert-sys-cert is that only a single key can be inserted into a compressed kernel. Having the ability to insert a key into the Linux trust boundary opens up various possibilities. The end-user can use a pre-built kernel and sign their own kernel modules. It also opens up the ability for an end-user to more easily use digital signature based IMA-appraisal. To get a key into the ima keyring, it must be signed by a key within the Linux trust boundary. Downstream Linux distros try to have a single signed kernel for each architecture. Each end-user may use this kernel in entirely different ways. Some downstream kernels have chosen to always trust platform keys within the Linux trust boundary for kernel module signing. These kernels have no way of using digital signature base IMA appraisal. This series introduces a new Linux kernel keyring containing the Machine Owner Keys (MOK) called .mok. It also adds a new MOK variable to shim. This variable allows the end-user to decide if they want to trust keys enrolled in the MOK within the Linux trust boundary. By default, nothing changes; MOK keys are not trusted within the Linux kernel. They are only trusted after the end-user makes the decision themselves. The end-user would set this through mokutil using a new --trust-mok option [3]. This would work similar to how the kernel uses MOK variables to enable/disable signature validation as well as use/ignore the db. When shim boots, it mirrors the new MokTML Boot Services variable to a new MokListTrustedRT Runtime Services variable and extends PCR14. MokListTrustedRT is written without EFI_VARIABLE_NON_VOLATILE set, preventing an end-user from setting it after booting and doing a kexec. When the kernel boots, if MokListTrustedRT is set and EFI_VARIABLE_NON_VOLATILE is not set, the MokListRT is loaded into the mok keyring instead of the platform keyring. Mimi has suggested that only CA keys or keys that can be vouched for by other kernel keys be loaded into this keyring. All other certs will load into the platform keyring instead. The .mok keyring contains a new keyring permission that only allows CA keys to be loaded. If the permission fails, the key is later loaded into the platform keyring. After all keys are added into the .mok keyring, they are linked to either the builtin or secondary trusted keyring. After the link is created, keys contained in the .mok keyring will automatically be searched when searching either builtin or secondary trusted keys. Secure Boot keys will never be trusted. They will always be loaded into the platform keyring. If an end-user wanted to trust one, they would need to enroll it into the MOK. I have included links to both the mokutil [3] and shim [4] changes I have made to support this new functionality. V2 changes: - The .mok keyring persists past boot - Removed the unrestricted move into the secondary keyring - Removed the keyring move bypass patch - Added restrictions to allow the .mok to be linked to either the builtin or secondary keyrings - Secondary keyring dependency has been removed [1] https://lore.kernel.org/lkml/1556221605.24945.3.camel@HansenPartnership.com/ [2] https://lore.kernel.org/patchwork/cover/902768/ [3] https://github.com/esnowberg/mokutil/tree/0.3.0-mokvars-v2 [4] https://github.com/esnowberg/shim/tree/mokvars-v2 Eric Snowberg (12): integrity: Introduce a Linux keyring for the Machine Owner Key (MOK) KEYS: CA link restriction integrity: Trust MOK keys if MokListTrustedRT found integrity: add add_to_mok_keyring integrity: restrict INTEGRITY_KEYRING_MOK to restrict_link_by_system_trusted_or_ca integrity: accessor function to get trust_moklist integrity: add new keyring handler for mok keys integrity: Suppress error message for keys added to the mok keyring KEYS: add a reference to mok keyring KEYS: link system_trusted_keys to mok_trusted_keys integrity: Do not allow mok keyring updates following init integrity: store reference to mok keyring certs/system_keyring.c | 47 ++++++++++ crypto/asymmetric_keys/restrict.c | 60 +++++++++++++ include/crypto/public_key.h | 5 ++ include/keys/system_keyring.h | 10 +++ security/integrity/Makefile | 3 +- security/integrity/digsig.c | 16 +++- security/integrity/integrity.h | 12 ++- .../platform_certs/keyring_handler.c | 17 +++- .../platform_certs/keyring_handler.h | 5 ++ security/integrity/platform_certs/load_uefi.c | 4 +- .../integrity/platform_certs/mok_keyring.c | 85 +++++++++++++++++++ 11 files changed, 256 insertions(+), 8 deletions(-) create mode 100644 security/integrity/platform_certs/mok_keyring.c base-commit: ff1176468d368232b684f75e82563369208bc371