netdev - Re: [PATCH net-next 1/8] bpf: multi program support for cgroup+bpf

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]

Message-ID: <f10e3c83-b73d-6f04-7c66-b01fab1f5a1a@fb.com>
Date:   Mon, 2 Oct 2017 22:00:15 -0700
From:   Alexei Starovoitov <ast@...com>
To:     David Ahern <dsa@...ulusnetworks.com>,
        "David S . Miller" <davem@...emloft.net>
CC:     Daniel Borkmann <daniel@...earbox.net>, Tejun Heo <tj@...nel.org>,
        <netdev@...r.kernel.org>, <kernel-team@...com>
Subject: Re: [PATCH net-next 1/8] bpf: multi program support for cgroup+bpf

On 10/2/17 9:26 PM, David Ahern wrote:
> On 10/2/17 9:21 PM, Alexei Starovoitov wrote:
>>
>> i'm not sure what you're trying to say.
>> The first loop quoted above is inside cgroup_bpf_put()
>> which is called when cgroup is destroyed. At this point
>> we're detaching and prog_put all attached programs.
>> While there is only one static_branch_inc() in __cgroup_bpf_attach()
>> that is called every time the prog is attached to a cgroup.
>> So what's the concern?
>
> just asking if cgroup_bpf_enabled_key is 0 when all programs are removed
> -- ie., that the inc's and dec's are equal. Reviewing this patch it was
> not clear that they are.

after some debugging turned out there is a typo in attach code
that leaks prog in the case of override.
Strangely kmemleak didn't catch it.
Will respin.