Re: [PATCH v2 0/4] Improved seccomp logging

[Kees Cook]

On Thu, Feb 2, 2017 at 9:37 PM, Tyler Hicks <tyhicks@xxxxxxxxxxxxx> wrote:
> This patch set is the second revision of the following two previously
> submitted patch sets:
>
> http://lkml.kernel.org/r/1483375990-14948-1-git-send-email-tyhicks@xxxxxxxxxxxxx
> http://lkml.kernel.org/r/1483377999-15019-2-git-send-email-tyhicks@xxxxxxxxxxxxx
>
> The patch set aims to address some known deficiencies in seccomp's current
> logging capabilities:
>
>   1. Inability to log all filter actions.
>   2. Inability to selectively enable filtering; e.g. devs want noisy logging,
>      users want relative quiet.
>   3. Consistent behavior with audit enabled and disabled.
>   4. Inability to easily develop a filter due to the lack of a
>      permissive/complain mode.
>
> The first three items were outlined by Paul Moore and are issues that I agree
> with. The last one is one that I'm particularly interested in.
>
> I deviated a little from the plan that he laid out to address the third issue.
> Looking back at Paul's feedback, he wanted a way to log seccomp actions even
> when the audit subsystem is disabled at build time. I felt like the bigger
> problem is that, while it is common for kernels to be built with audit support,
> it is far less common to actually have auditd running. Therefore, my approach
> was to improve the situation when kernel audit support is enabled at build time
> but audit_enabled is false at runtime. The audit subsystem forwards messages
> onto syslog in that situation.

I'm pretty happy with this series; it's pretty close to something I'd
Ack. :) I think this will get us a lot of what people have asked for
without too much pain. I'll add some thoughts on each of the specific
patches...

-Kees

-- 
Kees Cook
Pixel Security