Re: [PATCH] integrity: check the return value of audit_log_start()

[Mimi Zohar <zohar@linux.ibm.com>]

On Wed, 2022-01-12 at 12:23 -0500, Paul Moore wrote:
> On Tue, Dec 14, 2021 at 12:39 AM <xkernel.wang@foxmail.com> wrote:
> >
> > From: Xiaoke Wang <xkernel.wang@foxmail.com>
> >
> > audit_log_start() returns audit_buffer pointer on success or NULL on
> > error, so it is better to check the return value of it to prevent
> > potential memory access error.
> >
> > Signed-off-by: Xiaoke Wang <xkernel.wang@foxmail.com>
> > ---
> >  security/integrity/integrity_audit.c | 38 +++++++++++++++-------------
> >  1 file changed, 20 insertions(+), 18 deletions(-)
> >
> > diff --git a/security/integrity/integrity_audit.c b/security/integrity/integrity_audit.c
> > index 2922005..62785d5 100644
> > --- a/security/integrity/integrity_audit.c
> > +++ b/security/integrity/integrity_audit.c
> > @@ -45,23 +45,25 @@ void integrity_audit_message(int audit_msgno, struct inode *inode,
> >                 return;
> >
> >         ab = audit_log_start(audit_context(), GFP_KERNEL, audit_msgno);
> 
> As this is IMA code, it's largely up to Mimi about what she would
> prefer, but I think a much simpler patch would be to just return early
> if ab == NULL, for example:
> 
>   ab = audit_log_start( .... , audit_msgno);
>   if (!ab)
>     return;

Thanks, Paul.  Neither the linux-integrity mailing list nor I were
Cc'ed on this.  Looks like the IMA & EVM records in MAINTAINERS should
be updated to include the security/integrity directory.

thanks,

Mimi