On 2025-03-06 16:06, Jan Kara wrote: > On Wed 05-03-25 16:33:19, Richard Guy Briggs wrote: > > When no audit rules are in place, fanotify event results are > > unconditionally dropped due to an explicit check for the existence of > > any audit rules. Given this is a report from another security > > sub-system, allow it to be recorded regardless of the existence of any > > audit rules. > > > > To test, install and run the fapolicyd daemon with default config. Then > > as an unprivileged user, create and run a very simple binary that should > > be denied. Then check for an event with > > ausearch -m FANOTIFY -ts recent > > > > Link: https://issues.redhat.com/browse/RHEL-1367 > > Signed-off-by: Richard Guy Briggs <rgb(a)redhat.com&gt; > > I don't know enough about security modules to tell whether this is what > admins want or not so that's up to you but: > > > -static inline void audit_fanotify(u32 response, struct fanotify_response_info_audit_rule *friar) > > -{ > > - if (!audit_dummy_context()) > > - __audit_fanotify(response, friar); > > -} > > - > > I think this is going to break compilation with !CONFIG_AUDITSYSCALL && > CONFIG_FANOTIFY? Why would that break it? The part of the patch you (prematurely) deleted takes care of that.

diff --git a/include/linux/audit.h b/include/linux/audit.h index 0050ef288ab3..d0c6f23503a1 100644 --- a/include/linux/audit.h +++ b/include/linux/audit.h @@ -418,7 +418,7 @@ extern void __audit_log_capset(const struct cred *new, const struct cred *old); extern void __audit_mmap_fd(int fd, int flags); extern void __audit_openat2_how(struct open_how *how); extern void __audit_log_kern_module(char *name); -extern void __audit_fanotify(u32 response, struct fanotify_response_info_audit_rule *friar); +extern void audit_fanotify(u32 response, struct fanotify_response_info_audit_rule *friar); extern void __audit_tk_injoffset(struct timespec64 offset); extern void __audit_ntp_log(const struct audit_ntp_data *ad); extern void __audit_log_nfcfg(const char *name, u8 af, unsigned int nentries, > Honza > -- > Jan Kara <jack(a)suse.com&gt; > SUSE Labs, CR - RGB -- Richard Guy Briggs <rgb(a)redhat.com&gt; Sr. S/W Engineer, Kernel Security, Base Operating Systems Remote, Ottawa, Red Hat Canada Upstream IRC: SunRaycer Voice: +1.613.860 2354 SMS: +1.613.518.6570

On Mar 5, 2025 Richard Guy Briggs <rgb(a)redhat.com&gt; wrote: > > When no audit rules are in place, fanotify event results are > unconditionally dropped due to an explicit check for the existence of > any audit rules. Given this is a report from another security > sub-system, allow it to be recorded regardless of the existence of any > audit rules. > > To test, install and run the fapolicyd daemon with default config. Then > as an unprivileged user, create and run a very simple binary that should > be denied. Then check for an event with > ausearch -m FANOTIFY -ts recent > > Link: https://issues.redhat.com/browse/RHEL-1367 > Signed-off-by: Richard Guy Briggs <rgb(a)redhat.com&gt; > Acked-by: Jan Kara <jack(a)suse.cz&gt; > --- > include/linux/audit.h | 8 +------- > kernel/auditsc.c | 2 +- > 2 files changed, 2 insertions(+), 8 deletions(-) > > diff --git a/include/linux/audit.h b/include/linux/audit.h > index 0050ef288ab3..d0c6f23503a1 100644 > --- a/include/linux/audit.h > +++ b/include/linux/audit.h > @@ -418,7 +418,7 @@ extern void __audit_log_capset(const struct cred *new, const struct cred *old); > extern void __audit_mmap_fd(int fd, int flags); > extern void __audit_openat2_how(struct open_how *how); > extern void __audit_log_kern_module(char *name); > -extern void __audit_fanotify(u32 response, struct fanotify_response_info_audit_rule *friar); > +extern void audit_fanotify(u32 response, struct fanotify_response_info_audit_rule *friar); > extern void __audit_tk_injoffset(struct timespec64 offset); > extern void __audit_ntp_log(const struct audit_ntp_data *ad); > extern void __audit_log_nfcfg(const char *name, u8 af, unsigned int nentries, > @@ -525,12 +525,6 @@ static inline void audit_log_kern_module(char *name) > __audit_log_kern_module(name); > } > > -static inline void audit_fanotify(u32 response, struct fanotify_response_info_audit_rule *friar) > -{ > - if (!audit_dummy_context()) > - __audit_fanotify(response, friar); > -} It seems like we should at least have an audit_enabled() check, yes? We've had people complain about audit events being generated when audit is disabled, any while we don't currently have such a check in place here, I believe the dummy context check is doing that for us. static inline void audit_fanotify(...) { if (!audit_enabled) return; __audit_fanotify(...); }

> static inline void audit_tk_injoffset(struct timespec64 offset) > { > /* ignore no-op events */ > diff --git a/kernel/auditsc.c b/kernel/auditsc.c > index 0627e74585ce..936825114bae 100644 > --- a/kernel/auditsc.c > +++ b/kernel/auditsc.c > @@ -2880,7 +2880,7 @@ void __audit_log_kern_module(char *name) > context->type = AUDIT_KERN_MODULE; > } > > -void __audit_fanotify(u32 response, struct fanotify_response_info_audit_rule *friar) > +void audit_fanotify(u32 response, struct fanotify_response_info_audit_rule *friar) > { > /* {subj,obj}_trust values are {0,1,2}: no,yes,unknown */ > switch (friar->hdr.type) { > -- > 2.43.5 -- paul-moore.com

On Tue, May 27, 2025 at 9:10 AM Richard Guy Briggs <rgb(a)redhat.com&gt; wrote: > On 2025-04-11 14:14, Paul Moore wrote: > > On Mar 5, 2025 Richard Guy Briggs <rgb(a)redhat.com&gt; wrote: > > > > > > When no audit rules are in place, fanotify event results are > > > unconditionally dropped due to an explicit check for the existence of > > > any audit rules. Given this is a report from another security > > > sub-system, allow it to be recorded regardless of the existence of any > > > audit rules. > > > > > > To test, install and run the fapolicyd daemon with default config. Then > > > as an unprivileged user, create and run a very simple binary that should > > > be denied. Then check for an event with > > > ausearch -m FANOTIFY -ts recent > > > > > > Link: https://issues.redhat.com/browse/RHEL-1367 > > > Signed-off-by: Richard Guy Briggs <rgb(a)redhat.com&gt; > > > Acked-by: Jan Kara <jack(a)suse.cz&gt; > > > --- > > > include/linux/audit.h | 8 +------- > > > kernel/auditsc.c | 2 +- > > > 2 files changed, 2 insertions(+), 8 deletions(-) > > > > > > diff --git a/include/linux/audit.h b/include/linux/audit.h > > > index 0050ef288ab3..d0c6f23503a1 100644 > > > --- a/include/linux/audit.h > > > +++ b/include/linux/audit.h > > > @@ -418,7 +418,7 @@ extern void __audit_log_capset(const struct cred *new, const struct cred *old); > > > extern void __audit_mmap_fd(int fd, int flags); > > > extern void __audit_openat2_how(struct open_how *how); > > > extern void __audit_log_kern_module(char *name); > > > -extern void __audit_fanotify(u32 response, struct fanotify_response_info_audit_rule *friar); > > > +extern void audit_fanotify(u32 response, struct fanotify_response_info_audit_rule *friar); > > > extern void __audit_tk_injoffset(struct timespec64 offset); > > > extern void __audit_ntp_log(const struct audit_ntp_data *ad); > > > extern void __audit_log_nfcfg(const char *name, u8 af, unsigned int nentries, > > > @@ -525,12 +525,6 @@ static inline void audit_log_kern_module(char *name) > > > __audit_log_kern_module(name); > > > } > > > > > > -static inline void audit_fanotify(u32 response, struct fanotify_response_info_audit_rule *friar) > > > -{ > > > - if (!audit_dummy_context()) > > > - __audit_fanotify(response, friar); > > > -} > > > > It seems like we should at least have an audit_enabled() check, yes? > > We've had people complain about audit events being generated when audit > > is disabled, any while we don't currently have such a check in place > > here, I believe the dummy context check is doing that for us. > > > > static inline void audit_fanotify(...) > > { > > if (!audit_enabled) > > return; > > __audit_fanotify(...); > > } > > That would be consistent with other security events messages. I was > going through the selinux code to see what it does and I am missing it > if selinux checks with audit_enabled(). Are selinux messages somehow > exempt from audit_enabled()? There are likely a number of callers in the kernel that don't have audit_enabled() checks, some are probably bugs, others probably intentional; I wouldn't worry too much about what one subsystem does when deciding what to do for another. In the case of fanotify, I suspect the right thing to do is add an audit_enabled() check since it is already doing an audit_dummy_context() check. To be clear, there may be some cases where we do an audit_dummy_context() check and doing an audit_enabled() check would be wrong, but I don't believe that is the case with fanotify.

paul-moore.com

When no audit rules are in place, AUDIT_ANOM_{LINK,CREAT} events reported in audit_log_path_denied() are unconditionally dropped due to an explicit check for the existence of any audit rules. Given this is a report of a security violation, allow it to be recorded regardless of the existence of any audit rules. To test, mkdir -p /root/tmp chmod 1777 /root/tmp touch /root/tmp/test.txt useradd test chown test /root/tmp/test.txt {echo C0644 12 test.txt; printf 'hello\ntest1\n'; printf \\000;} | \ scp -t /root/tmp Check with ausearch -m ANOM_CREAT -ts recent Link: https://issues.redhat.com/browse/RHEL-9065 Signed-off-by: Richard Guy Briggs <rgb(a)redhat.com&gt; --- kernel/audit.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/kernel/audit.c b/kernel/audit.c index 53e3bddcc327..0cf2827882fc 100644 --- a/kernel/audit.c +++ b/kernel/audit.c @@ -2285,7 +2285,7 @@ void audit_log_path_denied(int type, const char *operation) { struct audit_buffer *ab; - if (!audit_enabled || audit_dummy_context()) + if (!audit_enabled) return; /* Generate log with subject, operation, outcome. */ -- 2.43.5