Re: [PATCH v5 3/3] fanotify,audit: Allow audit to use the full permission event response

[Paul Moore <paul@paul-moore.com>]

On Mon, Dec 12, 2022 at 9:06 AM Richard Guy Briggs <rgb@redhat.com> wrote:
>
> This patch passes the full response so that the audit function can use all
> of it. The audit function was updated to log the additional information in
> the AUDIT_FANOTIFY record.
>
> Currently the only type of fanotify info that is defined is an audit
> rule number, but convert it to hex encoding to future-proof the field.
> Hex encoding suggested by Paul Moore <paul@paul-moore.com>.
>
> Sample records:
>   type=FANOTIFY msg=audit(1600385147.372:590): resp=2 fan_type=1 fan_info=3137 subj_trust=3 obj_trust=5
>   type=FANOTIFY msg=audit(1659730979.839:284): resp=1 fan_type=0 fan_info=3F subj_trust=2 obj_trust=2
>
> Suggested-by: Steve Grubb <sgrubb@redhat.com>
> Link: https://lore.kernel.org/r/3075502.aeNJFYEL58@x2
> Signed-off-by: Richard Guy Briggs <rgb@redhat.com>
> ---
>  fs/notify/fanotify/fanotify.c |  3 ++-
>  include/linux/audit.h         |  9 +++++----
>  kernel/auditsc.c              | 25 ++++++++++++++++++++++---
>  3 files changed, 29 insertions(+), 8 deletions(-)

...

> diff --git a/kernel/auditsc.c b/kernel/auditsc.c
> index d1fb821de104..8d523066d81f 100644
> --- a/kernel/auditsc.c
> +++ b/kernel/auditsc.c
> @@ -64,6 +64,7 @@
>  #include <uapi/linux/limits.h>
>  #include <uapi/linux/netfilter/nf_tables.h>
>  #include <uapi/linux/openat2.h> // struct open_how
> +#include <uapi/linux/fanotify.h>
>
>  #include "audit.h"
>
> @@ -2877,10 +2878,28 @@ void __audit_log_kern_module(char *name)
>         context->type = AUDIT_KERN_MODULE;
>  }
>
> -void __audit_fanotify(u32 response)
> +void __audit_fanotify(u32 response, struct fanotify_response_info_audit_rule *friar)
>  {
> -       audit_log(audit_context(), GFP_KERNEL,
> -               AUDIT_FANOTIFY, "resp=%u", response);
> +       struct audit_context *ctx = audit_context();
> +       struct audit_buffer *ab;
> +       char numbuf[12];
> +
> +       if (friar->hdr.type == FAN_RESPONSE_INFO_NONE) {
> +               audit_log(audit_context(), GFP_KERNEL, AUDIT_FANOTIFY,
> +                         "resp=%u fan_type=%u fan_info=3F subj_trust=2 obj_trust=2",
> +                         response, FAN_RESPONSE_INFO_NONE);

The fan_info, subj_trust, and obj_trust constant values used here are
awfully magic-numbery and not the usual sentinel values one might
expect for a "none" operation, e.g. zeros/INT_MAX/etc. I believe a
comment here explaining the values would be a good idea.

> +               return;
> +       }
> +       ab = audit_log_start(ctx, GFP_KERNEL, AUDIT_FANOTIFY);
> +       if (ab) {
> +               audit_log_format(ab, "resp=%u fan_type=%u fan_info=",
> +                                response, friar->hdr.type);
> +               snprintf(numbuf, sizeof(numbuf), "%u", friar->rule_number);
> +               audit_log_n_hex(ab, numbuf, sizeof(numbuf));

It looks like the kernel's printf format string parsing supports %X so
why not just use that for now, we can always complicate it later if
needed.  It would probably also remove the need for the @ab, @numbuf,
and @ctx variables.  For example:

audit_log(audit_context(), GFP_KERNEL, AUDIT_FANOTIFY,
  "resp=%u fan_type=%u fan_info=%X subj_trust=%u obj_trust=%u",
  response, friar->hdr.type, friar->rule_number,
  friar->subj_trust, friar->obj_trust);

Am I missing something?

> +               audit_log_format(ab, " subj_trust=%u obj_trust=%u",
> +                                friar->subj_trust, friar->obj_trust);
> +               audit_log_end(ab);
> +       }
>  }
>
>  void __audit_tk_injoffset(struct timespec64 offset)
> --
> 2.27.0

-- 
paul-moore.com