lsm: make SECURITY_PATH always enabled
Subsystems (11)
Song Liu
maintainerPATCH 1/1 lsm: make SECURITY_PATH always enabled
From: Tetsuo Handa <[email protected]>
Only TOMOYO needed CONFIG_SECURITY_PATH when it was introduced. But now, AppArmor, EVM, IMA and LandLock also need it. And kernels are likely built with at least one of these enabled if CONFIG_SECURITY is enabled. Let's simplify the dependency.
---
This was initially proposed in [1], but got Nacked by Paul. However, I still think this change makes sense. This is because the definition of "pathname based access control" is not clear. For example, security_path_notify, security_sb_mount, security_move_mount are enabled by CONFIG_SECURITY, but they all use struct path, and thus are "path based access control". Yes, there is a difference between "path based" and "pathname based", but it is really up to the LSMs to decide how they use struct path. TOMOYO uses d_abosolute_path on security_sb_mount, while LandLock does path walk with dget_parent() and follow_up().
The separation of CONFIG_SECURITY and CONFIG_SECURITY_PATH has actually caused confusion. In some of our early kernels, we enabled CONFIG_SECURITY but not CONFIG_SECURITY_PATH. Now, we have to add separate logic in user space to deal with missing CONFIG_SECURITY_PATH in these systems.
Given the vague definition and real world issue, I think we should just merge CONFIG_SECURITY and CONFIG_SECURITY_PATH.
PS: [1] doesn't build with CONFIG_SECURITY=n case. This issue is fixed in this version.
[1] https://lore.kernel.org/linux-security-module/[email protected]/
--- arch/mips/configs/loongson2k_defconfig | 1 - arch/mips/configs/loongson3_defconfig | 1 - include/linux/lsm_hook_defs.h | 2 - include/linux/security.h | 174 ++++++++++++------------- kernel/bpf/bpf_lsm.c | 2 - kernel/trace/bpf_trace.c | 2 - security/Kconfig | 9 -- security/apparmor/Kconfig | 1 - security/integrity/evm/Kconfig | 1 - security/integrity/ima/Kconfig | 1 - security/landlock/Kconfig | 1 - security/security.c | 2 - security/tomoyo/Kconfig | 1 - 13 files changed, 86 insertions(+), 112 deletions(-)
arch/mips/configs/loongson2k_defconfig +0 -1
@@ -325,7 +325,6 @@ CONFIG_NLS_UTF8=y CONFIG_SECURITY=y CONFIG_SECURITYFS=y CONFIG_SECURITY_NETWORK=y -CONFIG_SECURITY_PATH=y CONFIG_SECURITY_SELINUX=y CONFIG_SECURITY_SELINUX_BOOTPARAM=y CONFIG_SECURITY_SELINUX_DISABLE=y
arch/mips/configs/loongson3_defconfig +0 -1
@@ -376,7 +376,6 @@ CONFIG_NLS_UTF8=y CONFIG_SECURITY=y CONFIG_SECURITYFS=y CONFIG_SECURITY_NETWORK=y -CONFIG_SECURITY_PATH=y CONFIG_SECURITY_SELINUX=y CONFIG_SECURITY_SELINUX_BOOTPARAM=y CONFIG_DEFAULT_SECURITY_DAC=y
include/linux/lsm_hook_defs.h +0 -2
@@ -87,7 +87,6 @@ LSM_HOOK(int, -EOPNOTSUPP, dentry_init_security, struct dentry *dentry, LSM_HOOK(int, 0, dentry_create_files_as, struct dentry *dentry, int mode, struct qstr *name, const struct cred *old, struct cred *new) -#ifdef CONFIG_SECURITY_PATH LSM_HOOK(int, 0, path_unlink, const struct path *dir, struct dentry *dentry) LSM_HOOK(int, 0, path_mkdir, const struct path *dir, struct dentry *dentry, umode_t mode)
@@ -107,7 +106,6 @@ LSM_HOOK(int, 0, path_rename, const struct path *old_dir, LSM_HOOK(int, 0, path_chmod, const struct path *path, umode_t mode) LSM_HOOK(int, 0, path_chown, const struct path *path, kuid_t uid, kgid_t gid) LSM_HOOK(int, 0, path_chroot, const struct path *path) -#endif /* CONFIG_SECURITY_PATH */ /* Needed for inode based security check */ LSM_HOOK(int, 0, path_notify, const struct path *path, u64 mask,
include/linux/security.h +86 -88
192 lines not on this page · raw message
kernel/bpf/bpf_lsm.c +0 -2
@@ -329,7 +329,6 @@ BTF_ID(func, bpf_lsm_kernel_module_request) BTF_ID(func, bpf_lsm_kernel_read_file) BTF_ID(func, bpf_lsm_kernfs_init_security) -#ifdef CONFIG_SECURITY_PATH BTF_ID(func, bpf_lsm_path_unlink) BTF_ID(func, bpf_lsm_path_mkdir)BTF_ID (func, bpf_lsm_path_rmdir)
@@ -339,7 +338,6 @@ BTF_ID(func, bpf_lsm_path_link) BTF_ID(func, bpf_lsm_path_rename) BTF_ID(func, bpf_lsm_path_chmod) BTF_ID(func, bpf_lsm_path_chown) -#endif /* CONFIG_SECURITY_PATH */ BTF_ID(func, bpf_lsm_mmap_file)BTF_ID (func, bpf_lsm_netlink_send)
kernel/trace/bpf_trace.c +0 -2
@@ -936,9 +936,7 @@ BTF_ID(func, security_file_permission) BTF_ID(func, security_inode_getattr) BTF_ID(func, security_file_open)#endif -#ifdef CONFIG_SECURITY_PATH BTF_ID(func, security_path_truncate) -#endif BTF_ID(func, vfs_truncate) BTF_ID(func, vfs_fallocate)BTF_ID (func, dentry_open)
security/Kconfig +0 -9
@@ -129,15 +129,6 @@ config SECURITY_NETWORK_XFRM IPSec. If you are unsure how to answer this question, answer N. -config SECURITY_PATH - bool "Security hooks for pathname based access control" - depends on SECURITY - help - This enables the security hooks for pathname based access control. - If enabled, a security module can use these hooks to - implement pathname based access controls. - If you are unsure how to answer this question, answer N. - config INTEL_TXT bool "Enable Intel(R) Trusted Execution Technology (Intel(R) TXT)" depends on HAVE_INTEL_TXT
security/apparmor/Kconfig +0 -1
@@ -3,7 +3,6 @@ config SECURITY_APPARMOR bool "AppArmor support" depends on SECURITY && NET select AUDIT - select SECURITY_PATH select SECURITYFS select SECURITY_NETWORK default n
security/integrity/evm/Kconfig +0 -1
@@ -6,7 +6,6 @@ config EVM select CRYPTO_HMAC select CRYPTO_SHA1 select CRYPTO_HASH_INFO - select SECURITY_PATH default n help EVM protects a file's security extended attributes against
security/integrity/ima/Kconfig +0 -1
@@ -8,7 +8,6 @@ config IMA select CRYPTO_HMAC select CRYPTO_SHA1 select CRYPTO_HASH_INFO - select SECURITY_PATH select TCG_TPM if HAS_IOMEM select TCG_TIS if TCG_TPM && X86 select TCG_CRB if TCG_TPM && ACPI
security/landlock/Kconfig +0 -1
@@ -4,7 +4,6 @@ config SECURITY_LANDLOCK bool "Landlock support" depends on SECURITY select SECURITY_NETWORK - select SECURITY_PATH help Landlock is a sandboxing mechanism that enables processes to restrict themselves (and their future children) by gradually enforcing
security/security.c +0 -2
@@ -1878,7 +1878,6 @@ int security_inode_init_security_anon(struct inode *inode, context_inode); } -#ifdef CONFIG_SECURITY_PATH /** * security_path_mknod() - Check if creating a special file is allowed * @dir: parent directory
@@ -2092,7 +2091,6 @@ int security_path_chroot(const struct path *path) {return call_int_hook (path_chroot, path); } -#endif /* CONFIG_SECURITY_PATH */ /** * security_inode_create() - Check if creating a file is allowed
security/tomoyo/Kconfig +0 -1
@@ -4,7 +4,6 @@ config SECURITY_TOMOYO depends on SECURITY depends on NET select SECURITYFS - select SECURITY_PATH select SECURITY_NETWORK default n help
Signature · 1 line
2.47.1
Paul Moore
maintainerQuoted 10 lines · On Apr 22, 2025 Song Liu <[email protected]> wrote:
Only TOMOYO needed
CONFIG_SECURITY_PATHwhen it was introduced. But now, AppArmor, EVM, IMA and LandLock also need it. And kernels are likely built with at least one of these enabled ifCONFIG_SECURITYis enabled. Let's simplify the dependency.Signed-off-by Tetsuo Handa <[email protected]>Signed-off-by Song Liu <[email protected]>---
This was initially proposed in [1], but got Nacked by Paul.
I chose not to merge the patch with the following comment:
"If a Kconfig option is only needed by a subset of LSMs, which is the
case for CONFIG_SECURITY_PATH, it should remain a build-time option."
... and that reasoning still sounds reasonable to me today.
Quoted 2 lines
... but it is really up to the LSMs to decide how
they use struct path.
LSMs are currently free to select CONFIG_SECURITY_PATH as needed. In fact, if you look any modern Linux tree you'll see that TOMOYO, AppArmor, Landlock, IMA, and EVM all select CONFIG_SECURITY_PATH based on their individual Kconfig settings.
Quoted 4 lines
The separation of
CONFIG_SECURITYandCONFIG_SECURITY_PATHhas actually caused confusion. In some of our early kernels, we enabledCONFIG_SECURITYbut notCONFIG_SECURITY_PATH. Now, we have to add separate logic in user space to deal with missingCONFIG_SECURITY_PATHin these systems.
I'm sorry that you made some Kconfig choices on a production kernel that you now regret, but that doesn't change things from an upstream perspective. The exception of course would be if there was a LSM that *required* CONFIG_SECURITY_PATH but did not enable it or mark it as a dependency in the LSM's Kconfig. It doesn't sound like that is the case here, but please let me know if that was the root cause so we can work with the individual LSM to fix the associated Kconfig.
As you likely know, the LSM subsystem as a whole (framework and individual LSMs) is often criticized over various performance impacts; one of the easiest ways to limit the impact of the LSM on overall system performance is to remove LSM hooks that are not needed. Keeping the CONFIG_SECURITY_PATH hooks as-is and allowing individual LSMs to opt-in as needed is a reasonable balance between providing the necessary LSM hook support and limiting performance impacts.
--
paul-moore.com
Hi Paul,
Thanks for your quick comments!
Quoted 14 lines · Paul Moore
Quoted 10 lines · On Apr 22, 2025 Song Liu <[email protected]> wrote:
Only TOMOYO needed
CONFIG_SECURITY_PATHwhen it was introduced. But now, AppArmor, EVM, IMA and LandLock also need it. And kernels are likely built with at least one of these enabled ifCONFIG_SECURITYis enabled. Let's simplify the dependency.Signed-off-by Tetsuo Handa <[email protected]>Signed-off-by Song Liu <[email protected]>---
This was initially proposed in [1], but got Nacked by Paul.
Quoted 6 lines
I chose not to merge the patch with the following comment:
"If a
Kconfigoption is only needed by a subset of LSMs, which is the
case forCONFIG_SECURITY_PATH, it should remain a build-time option."... and that reasoning still sounds reasonable to me today.
I agree that a way to disable some LSM hooks at build time is valuable.
Quoted 8 lines
Quoted 3 lines
... but it is really up to the LSMs to decide how
they use struct path.LSMs are currently free to select
CONFIG_SECURITY_PATHas needed. In fact, if you look any modern Linux tree you'll see that TOMOYO, AppArmor, Landlock, IMA, and EVM all selectCONFIG_SECURITY_PATHbased on their individualKconfigsettings.
However, I don't think existing CONFIG_SECURITY_* are doing the right things. Among all the configs, CONFIG_SECURITY_PATH is the most awkward to me. Say , if we have CONFIG_SECURITY_PATH, shouldn't we also have CONFIG_SECURITY_INODE? IOW, something like:
#ifdef CONFIG_SECURITY_INODE
int security_inode_rmdir(struct inode *dir, struct dentry *dentry); #endif
#ifdef CONFIG_SECURITY_PATH
int security_path_rmdir(struct inode *dir, struct dentry *dentry); #endif
OR, maybe we should just remove security_inode_rmdir(), and users of security_inode_rmdir() can just use security_path_rmdir() instead? As I writing this, I think merging security_inode_rmdir() and security_path_rmdir() is a better solution. But the motivation stays the same: the separation of CONFIG_SECURITY_PATH from CONFIG_SECURITY is weird.
Does this make sense?
Thanks,
Song
[...]
Paul Moore
maintainerQuoted 37 lines · Song Liu
Quoted 14 lines · Paul Moore
Quoted 10 lines · On Apr 22, 2025 Song Liu <[email protected]> wrote:
Only TOMOYO needed
CONFIG_SECURITY_PATHwhen it was introduced. But now, AppArmor, EVM, IMA and LandLock also need it. And kernels are likely built with at least one of these enabled ifCONFIG_SECURITYis enabled. Let's simplify the dependency.Signed-off-by Tetsuo Handa <[email protected]>Signed-off-by Song Liu <[email protected]>---
This was initially proposed in [1], but got Nacked by Paul.Quoted 6 lines
I chose not to merge the patch with the following comment:
"If a
Kconfigoption is only needed by a subset of LSMs, which is the
case forCONFIG_SECURITY_PATH, it should remain a build-time option."... and that reasoning still sounds reasonable to me today.
I agree that a way to disable some LSM hooks at build time is valuable.
Quoted 8 lines
Quoted 3 lines
... but it is really up to the LSMs to decide how
they use struct path.LSMs are currently free to select
CONFIG_SECURITY_PATHas needed. In fact, if you look any modern Linux tree you'll see that TOMOYO, AppArmor, Landlock, IMA, and EVM all selectCONFIG_SECURITY_PATHbased on their individualKconfigsettings.However, I don't think existing CONFIG_SECURITY_* are doing the right things. Among all the configs,
CONFIG_SECURITY_PATHis the most awkward to me. Say , if we haveCONFIG_SECURITY_PATH, shouldn't we also haveCONFIG_SECURITY_INODE? IOW, something like:
Quoted 7 lines
#ifdef
CONFIG_SECURITY_INODE
intsecurity_inode_rmdir(struct inode *dir, struct dentry *dentry); #endif#ifdef
CONFIG_SECURITY_PATH
intsecurity_path_rmdir(struct inode *dir, struct dentry *dentry); #endif
Without putting much thought into what would fall under CONFIG_SECURITY_INODE, I think it would be interesting to see what hooks one might be able to make conditional on such a Kconfig knob. Using security_inode_permission() as a simple test, it looks like only SELinux and Smack provide implementations, spot checks on a few other security_*inode*() hooks shows similar, or even more limited, results.
You would need to spend some time to determine what LSM hooks are used by which LSMs and adjust their Kconfigs appropriately for the new CONFIG_SECURITY_INODE knob, but if you do that then I think that would be okay.
Quoted 2 lines
OR, maybe we should just remove
security_inode_rmdir(), and users ofsecurity_inode_rmdir()can just usesecurity_path_rmdir()instead?
Those two LSM hooks are called from slightly different places in the codepath which has an impact on their environment. For example, the inode variant doesn't have to deal with directory inodes that don't have a defined rmdir op, whereas the path variant does; the inode variant doesn't have to worry about S_KERNEL_FILE files, the inode variant has a refcount'd and locked dentry, etc. Moving an existing LSM, especially complex ones, from one LSM hook to another, is a delicate operation and might not be worth it for such a small return.
If you are interested in reducing the impact of the "inode" LSM hooks, CONFIG_SECURITY_INODE is likely a better place to start.
Signature · 1 line
paul-moore.com
Hi Paul,
Thanks for your comments!
Quoted 1 line · Paul Moore
[...]
Quoted 21 lines
Quoted 6 lines
However, I don't think existing CONFIG_SECURITY_* are doing the right things. Among all the configs,
CONFIG_SECURITY_PATHis the most awkward to me. Say , if we haveCONFIG_SECURITY_PATH, shouldn't we also haveCONFIG_SECURITY_INODE? IOW, something like:Quoted 7 lines
#ifdef
CONFIG_SECURITY_INODE
intsecurity_inode_rmdir(struct inode *dir, struct dentry *dentry); #endif#ifdef
CONFIG_SECURITY_PATH
intsecurity_path_rmdir(struct inode *dir, struct dentry *dentry); #endifWithout putting much thought into what would fall under
CONFIG_SECURITY_INODE, I think it would be interesting to see what hooks one might be able to make conditional on such aKconfigknob. Usingsecurity_inode_permission()as a simple test, it looks like only SELinux and Smack provide implementations, spot checks on a few other security_*inode*() hooks shows similar, or even more limited, results.
Quoted 4 lines
You would need to spend some time to determine what LSM hooks are used by which LSMs and adjust their Kconfigs appropriately for the new
CONFIG_SECURITY_INODEknob, but if you do that then I think that would be okay.
Well, I was hoping to simplify the CONFIGs by removing one. So I am not sure whether adding a new CONFIG is the right thing to do.
Quoted 4 lines
Quoted 3 lines
OR, maybe we should just remove
security_inode_rmdir(), and users ofsecurity_inode_rmdir()can just usesecurity_path_rmdir()instead?
Quoted 8 lines
Those two LSM hooks are called from slightly different places in the codepath which has an impact on their environment. For example, the inode variant doesn't have to deal with directory inodes that don't have a defined rmdir op, whereas the path variant does; the inode variant doesn't have to worry about
S_KERNEL_FILEfiles, the inode variant has a refcount'd and locked dentry, etc. Moving an existing LSM, especially complex ones, from one LSM hook to another, is a delicate operation and might not be worth it for such a small return.
Given there is pushback when a new LSM hook is added, I assume removing a hook (or merge two hooks into one) may be a good move. Well, it is totally possible that I underestimated the complexity of the work.
Thanks,
Song
Paul Moore
maintainerQuoted 25 lines · Song Liu
Quoted 1 line · Paul Moore
[...]
Quoted 21 lines
Quoted 6 lines
However, I don't think existing CONFIG_SECURITY_* are doing the right things. Among all the configs,
CONFIG_SECURITY_PATHis the most awkward to me. Say , if we haveCONFIG_SECURITY_PATH, shouldn't we also haveCONFIG_SECURITY_INODE? IOW, something like:Quoted 7 lines
#ifdef
CONFIG_SECURITY_INODE
intsecurity_inode_rmdir(struct inode *dir, struct dentry *dentry); #endif#ifdef
CONFIG_SECURITY_PATH
intsecurity_path_rmdir(struct inode *dir, struct dentry *dentry); #endifWithout putting much thought into what would fall under
CONFIG_SECURITY_INODE, I think it would be interesting to see what hooks one might be able to make conditional on such aKconfigknob. Usingsecurity_inode_permission()as a simple test, it looks like only SELinux and Smack provide implementations, spot checks on a few other security_*inode*() hooks shows similar, or even more limited, results.
Quoted 7 lines
Quoted 4 lines
You would need to spend some time to determine what LSM hooks are used by which LSMs and adjust their Kconfigs appropriately for the new
CONFIG_SECURITY_INODEknob, but if you do that then I think that would be okay.Well, I was hoping to simplify the CONFIGs by removing one. So I am not sure whether adding a new CONFIG is the right thing to do.
Ah, in that case I suspect you're going to be disappointed as I'm fairly confident we don't want to consolidate the inode and path based hooks under one Kconfig knob at this point in time. If anything, I think there may be some value in adding an inode Kconfig as described above, which I realize isn't your original goal, but still ... :)
Quoted 12 lines
Quoted 11 lines
Quoted 2 lines
OR, maybe we should just remove
security_inode_rmdir(), and users ofsecurity_inode_rmdir()can just usesecurity_path_rmdir()instead?Those two LSM hooks are called from slightly different places in the codepath which has an impact on their environment. For example, the inode variant doesn't have to deal with directory inodes that don't have a defined rmdir op, whereas the path variant does; the inode variant doesn't have to worry about
S_KERNEL_FILEfiles, the inode variant has a refcount'd and locked dentry, etc. Moving an existing LSM, especially complex ones, from one LSM hook to another, is a delicate operation and might not be worth it for such a small return.
Quoted 4 lines
Given there is pushback when a new LSM hook is added, I assume removing a hook (or merge two hooks into one) may be a good move. Well, it is totally possible that I underestimated the complexity of the work.
The funny thing is that the difficulty in adding LSM hooks is one of the main reasons why I am so hesitant to remove an existing hook; you can consider it as perhaps an unintended consequence of a general hostility towards the LSM.
Regardless of the above, yes, there can be a lot of complexity involved in adding, modifying, or consolidating LSM hooks as there can be decades worth of assumptions both in the LSMs and in the caller that need to be considered for each change. Of course that doesn't mean such change can't happen - we have plenty of examples where it has - but such changes are often more complicated than they appear.
Signature · 1 line
paul-moore.com
Quoted 3 lines · Paul Moore
[...]
Quoted 12 lines
Quoted 12 lines
Quoted 11 lines
Without putting much thought into what would fall under
CONFIG_SECURITY_INODE, I think it would be interesting to see what hooks one might be able to make conditional on such aKconfigknob. Usingsecurity_inode_permission()as a simple test, it looks like only SELinux and Smack provide implementations, spot checks on a few other security_*inode*() hooks shows similar, or even more limited, results.You would need to spend some time to determine what LSM hooks are used by which LSMs and adjust their Kconfigs appropriately for the new
CONFIG_SECURITY_INODEknob, but if you do that then I think that would be okay.
Quoted 8 lines
Quoted 2 lines
Well, I was hoping to simplify the CONFIGs by removing one. So I am not sure whether adding a new CONFIG is the right thing to do.
Ah, in that case I suspect you're going to be disappointed as I'm fairly confident we don't want to consolidate the inode and path based hooks under one
Kconfigknob at this point in time. If anything, I think there may be some value in adding an inodeKconfigas described above, which I realize isn't your original goal, but still ... :)
I am thinking whether it is possible to have each LSM selects required hooks, and only enable those at compile time. OTOH, my primary focus is BPF LSM, so these optimizations matter little for my use cases.
Quoted 17 lines
Quoted 12 lines
Quoted 11 lines
Quoted 2 lines
OR, maybe we should just remove
security_inode_rmdir(), and users ofsecurity_inode_rmdir()can just usesecurity_path_rmdir()instead?Those two LSM hooks are called from slightly different places in the codepath which has an impact on their environment. For example, the inode variant doesn't have to deal with directory inodes that don't have a defined rmdir op, whereas the path variant does; the inode variant doesn't have to worry about
S_KERNEL_FILEfiles, the inode variant has a refcount'd and locked dentry, etc. Moving an existing LSM, especially complex ones, from one LSM hook to another, is a delicate operation and might not be worth it for such a small return.Quoted 4 lines
Given there is pushback when a new LSM hook is added, I assume removing a hook (or merge two hooks into one) may be a good move. Well, it is totally possible that I underestimated the complexity of the work.
Quoted 4 lines
The funny thing is that the difficulty in adding LSM hooks is one of the main reasons why I am so hesitant to remove an existing hook; you can consider it as perhaps an unintended consequence of a general hostility towards the LSM.
I didn't expect this. Thanks for sharing your thoughts.
Quoted 7 lines
Regardless of the above, yes, there can be a lot of complexity involved in adding, modifying, or consolidating LSM hooks as there can be decades worth of assumptions both in the LSMs and in the caller that need to be considered for each change. Of course that doesn't mean such change can't happen - we have plenty of examples where it has - but such changes are often more complicated than they appear.
Song
Paul Moore
maintainerQuoted 25 lines · Song Liu
Quoted 1 line · Paul Moore
[...]
Quoted 12 lines
Quoted 12 lines
Quoted 11 lines
Without putting much thought into what would fall under
CONFIG_SECURITY_INODE, I think it would be interesting to see what hooks one might be able to make conditional on such aKconfigknob. Usingsecurity_inode_permission()as a simple test, it looks like only SELinux and Smack provide implementations, spot checks on a few other security_*inode*() hooks shows similar, or even more limited, results.You would need to spend some time to determine what LSM hooks are used by which LSMs and adjust their Kconfigs appropriately for the new
CONFIG_SECURITY_INODEknob, but if you do that then I think that would be okay.Quoted 8 lines
Quoted 2 lines
Well, I was hoping to simplify the CONFIGs by removing one. So I am not sure whether adding a new CONFIG is the right thing to do.
Ah, in that case I suspect you're going to be disappointed as I'm fairly confident we don't want to consolidate the inode and path based hooks under one
Kconfigknob at this point in time. If anything, I think there may be some value in adding an inodeKconfigas described above, which I realize isn't your original goal, but still ... :)
Quoted 3 lines
I am thinking whether it is possible to have each LSM selects required hooks, and only enable those at compile time. OTOH, my primary focus is BPF LSM, so these optimizations matter little for my use cases.
Ignoring for a moment just how awkward it would be to have a Kconfig value for each LSM, it is important to remember that LSMs are still selectable at system boot (assuming they have been built into the kernel), so there will always be the potential for "empty" LSM hooks on any given system boot.
Like I said before, I'm open to someone exploring the addition of new LSM Kconfig knobs to provide greater granularity about which hooks are enabled at build time, however, these new knobs will both need to make sense from a logical grouping perspective as well as have a meaningful impact on the enabled hooks across the current in-tree LSMs.
Signature · 1 line
paul-moore.com