LKML Archive on lore.kernel.org
help / color / mirror / Atom feed
From: Daniel Borkmann <daniel@iogearbox.net>
To: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com>
Cc: Alexei Starovoitov <ast@plumgrid.com>,
	Ingo Molnar <mingo@kernel.org>,
	Steven Rostedt <rostedt@goodmis.org>,
	Namhyung Kim <namhyung@kernel.org>,
	Arnaldo Carvalho de Melo <acme@infradead.org>,
	Jiri Olsa <jolsa@redhat.com>,
	"David S. Miller" <davem@davemloft.net>,
	Peter Zijlstra <a.p.zijlstra@chello.nl>,
	linux-api@vger.kernel.org, netdev@vger.kernel.org,
	linux-kernel@vger.kernel.org
Subject: Re: [PATCH v5 tip 1/7] bpf: make internal bpf API independent of CONFIG_BPF_SYSCALL ifdefs
Date: Mon, 02 Mar 2015 13:26:17 +0100	[thread overview]
Message-ID: <54F456E9.1000702@iogearbox.net> (raw)
In-Reply-To: <54F44EC3.4070504@hitachi.com>

On 03/02/2015 12:51 PM, Masami Hiramatsu wrote:
> (2015/03/02 20:10), Daniel Borkmann wrote:
>> On 03/02/2015 11:53 AM, Masami Hiramatsu wrote:
>> ...
>>> Hmm, it seems that this still doesn't hide some APIs which is provided
>>> only when CONFIG_BPF_SYSCALL. For example bpf_register_map_type etc.
>>> I think all those APIs should be hidden in #ifdef or at least be commented
>>> so that the user doesn't refer that without the kconfig.
>>> (I don't think we need to provide dummy functions for those APIs,
>>>    but it's better to clarify which API we can use with which kconfig)
>>
>> Well, currently all possible map types (hash table, array map) that
>> would actually call into bpf_register_map_type() are only built when
>> CONFIG_BPF_SYSCALL is enabled (see kernel/bpf/Makefile). I don't think
>> new map additions should be added that are not under kernel/bpf/ and/or
>> enabled outside the CONFIG_BPF_SYSCALL, as it should be considered
>> part of the eBPF core code.
>>
>> The difference here (this patch) is simply that we don't want users to
>> build ifdef spaghetti constructs in user code, so the API that is
>> actually used by eBPF _users_ is being properly ifdef'ed in the headers.
>>
>> So, I don't think this is a big problem, but we could move these bits
>> under the ifdef CONFIG_BPF_SYSCALL w/o providing a dummy in the else part.
>> I can do that outside of the scope of this set.
> 
> Or, maybe we'd better move them into new include/linux/bpf_prog.h which
> includes basic include/linux/bpf.h. Then, user can include the bpf_prog.h
> instead of bpf.h. Also, we can check CONFIG_BPF_SYSCAL=y at the top of
> bpf_prog.h. This makes things clearer :)

I'm preferring the 1st variant, though. We have currently two native eBPF
users, that is, socket filters and tc's cls_bpf (queued in net-next) and
looking at the code/API usage, it's really not that hard, where it would
justify to move this to an extra header file, really.

I'm cooking a patch for net-next right now with the first variant (which is
on top of this patch that resides in net-next as-is as well).

Thanks,
Daniel

  reply	other threads:[~2015-03-02 12:26 UTC|newest]

Thread overview: 18+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2015-03-01 23:27 [PATCH v5 tip 0/7] tracing: attach eBPF programs to kprobes Alexei Starovoitov
2015-03-01 23:27 ` [PATCH v5 tip 1/7] bpf: make internal bpf API independent of CONFIG_BPF_SYSCALL ifdefs Alexei Starovoitov
2015-03-02 10:53   ` Masami Hiramatsu
2015-03-02 11:10     ` Daniel Borkmann
2015-03-02 11:51       ` Masami Hiramatsu
2015-03-02 12:26         ` Daniel Borkmann [this message]
2015-03-01 23:27 ` [PATCH v5 tip 2/7] tracing: attach BPF programs to kprobes Alexei Starovoitov
2015-03-01 23:27 ` [PATCH v5 tip 3/7] tracing: allow BPF programs to call bpf_ktime_get_ns() Alexei Starovoitov
2015-03-01 23:27 ` [PATCH v5 tip 4/7] tracing: allow BPF programs to call bpf_trace_printk() Alexei Starovoitov
2015-03-01 23:27 ` [PATCH v5 tip 5/7] samples: bpf: simple non-portable kprobe filter example Alexei Starovoitov
2015-03-01 23:27 ` [PATCH v5 tip 6/7] samples: bpf: counting example for kfree_skb and write syscall Alexei Starovoitov
2015-03-01 23:27 ` [PATCH v5 tip 7/7] samples: bpf: IO latency analysis (iosnoop/heatmap) Alexei Starovoitov
2015-03-04 17:23 ` [PATCH v5 tip 0/7] tracing: attach eBPF programs to kprobes Alexei Starovoitov
2015-03-04 20:33   ` Ingo Molnar
2015-03-04 20:48     ` Steven Rostedt
2015-03-07  1:09       ` Steven Rostedt
2015-03-08  0:21         ` Alexei Starovoitov
2015-03-02 16:58 [PATCH v5 tip 1/7] bpf: make internal bpf API independent of CONFIG_BPF_SYSCALL ifdefs Alexei Starovoitov

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=54F456E9.1000702@iogearbox.net \
    --to=daniel@iogearbox.net \
    --cc=a.p.zijlstra@chello.nl \
    --cc=acme@infradead.org \
    --cc=ast@plumgrid.com \
    --cc=davem@davemloft.net \
    --cc=jolsa@redhat.com \
    --cc=linux-api@vger.kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=masami.hiramatsu.pt@hitachi.com \
    --cc=mingo@kernel.org \
    --cc=namhyung@kernel.org \
    --cc=netdev@vger.kernel.org \
    --cc=rostedt@goodmis.org \
    --subject='Re: [PATCH v5 tip 1/7] bpf: make internal bpf API independent of CONFIG_BPF_SYSCALL ifdefs' \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).