netdev - Re: [RFC 0/3] bpf: Add d

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:   Tue, 7 Apr 2020 11:27:53 +0200
From:   KP Singh <kpsingh@...omium.org>
To:     Alexei Starovoitov <alexei.starovoitov@...il.com>
Cc:     Jiri Olsa <jolsa@...hat.com>, Al Viro <viro@...iv.linux.org.uk>,
        Jiri Olsa <jolsa@...nel.org>,
        Alexei Starovoitov <ast@...nel.org>,
        Daniel Borkmann <daniel@...earbox.net>, netdev@...r.kernel.org,
        bpf@...r.kernel.org, Yonghong Song <yhs@...com>,
        Martin KaFai Lau <kafai@...com>,
        David Miller <davem@...hat.com>,
        John Fastabend <john.fastabend@...il.com>,
        Jesper Dangaard Brouer <hawk@...nel.org>,
        Wenbo Zhang <ethercflow@...il.com>,
        KP Singh <kpsingh@...omium.org>,
        Andrii Nakryiko <andriin@...com>, bgregg@...flix.com
Subject: Re: [RFC 0/3] bpf: Add d_path helper

On 06-Apr 18:10, Alexei Starovoitov wrote:
> On Mon, Apr 06, 2020 at 11:09:18AM +0200, Jiri Olsa wrote:
> > 
> > is there any way we could have d_path functionality (even
> > reduced and not working for all cases) that could be used
> > or called like that?
> 
> I agree with Al. This helper cannot be enabled for all of bpf tracing.
> We have to white list its usage for specific callsites only.
> May be all of lsm hooks are safe. I don't know yet. This has to be
> analyzed carefully. Every hook. One by one.

I agree with this, there are some LSM hooks which do get called in
interrupt context, eg. task_free (which gets called in an RCU
callback).

The hooks that we are using it for and we know that it works (using
our experimental helpers similar to this) are the bprm_* hooks in the
exec pathway (for logic based on the path of the executable).

It might be worth whitelisting these functions by adding verifier ops
for LSM programs?

Would you want to do it as a part of this series?

- KP

> in_task() isn't really a solution.
> 
> At the same time I agree that such helper is badly needed.
> Folks have been requesting it for long time.