Message-ID: <55794240.7060009@huawei.com>
Date: Thu, 11 Jun 2015 16:09:36 +0800
From: "Wangnan (F)" <wangnan0@huawei.com>
User-Agent: Mozilla/5.0 (Windows NT 6.1; WOW64; rv:31.0) Gecko/20100101 Thunderbird/31.6.0
MIME-Version: 1.0
To: Namhyung Kim <namhyung@kernel.org>, Alexei Starovoitov <ast@plumgrid.com>
CC: <acme@kernel.org>, <brendan.d.gregg@gmail.com>, <daniel@iogearbox.net>,
        <masami.hiramatsu.pt@hitachi.com>, <paulus@samba.org>,
        <a.p.zijlstra@chello.nl>, <mingo@redhat.com>, <jolsa@kernel.org>,
        <dsahern@gmail.com>, <linux-kernel@vger.kernel.org>,
        <lizefan@huawei.com>, <hekuang@huawei.com>, <xiakaixu@huawei.com>,
        <pi3orama@163.com>
Subject: Re: [RFC PATCH v6 25/32] perf tools: Add 'bpf.' config section to
 perf default config
References: <1433829036-23687-1-git-send-email-wangnan0@huawei.com> <1433829036-23687-26-git-send-email-wangnan0@huawei.com> <55777A17.7030000@plumgrid.com> <20150611074553.GB11184@sejong>
In-Reply-To: <20150611074553.GB11184@sejong>
Content-Type: text/plain; charset="utf-8"; format=flowed
Content-Transfer-Encoding: 7bit
Sender: linux-kernel-owner@vger.kernel.org
Content-Length: 2170
Lines: 59


On 2015/6/11 15:45, Namhyung Kim wrote:
> On Tue, Jun 09, 2015 at 04:43:19PM -0700, Alexei Starovoitov wrote:
>> On 6/8/15 10:50 PM, Wang Nan wrote:
>>> perf_bpf_config() is added to parse 'bpf' section in perf config file.
>>> Following is an example:
>>>
>>>   [bpf]
>>>     clang-path = /llvm/bin/x86_64-linux-clang"
>>>     llc-path = /llvm/bin/x86_64-linux-llc"
>>>     clang-opt = "-nostdinc -isystem /llvm/lib/clang/include -I/kernel/arch/x86/include ..."
>>>     llc-opt = ""
>> a section to specify -I flags to compile prog.c is useful,
>> but users shouldn't be populating it manually for kernel headers.
>> How about adding a script that can figure out $(LINUXINCLUDE)
>> automatically ?
> Maybe a dummy question.  For what reason it needs such headers?  Is it
> possible to have a (part of?) copy of needed bits in the perf source?
>
> Thanks,
> Namhyung

It is improtant for eBPF script that it should be able to extract 
information
from kernel pointers. For example: to extract registers from 'struct 
pt_regs' and
to extract fields from 'struct page *'. I believe this is an improtant 
reason
why we decide to use C-like language instead of designing a DLS. 
Therefore, when
compiling .c to .o, kernel source is mandatory. Since we don't want to 
limit the
ability of data accessing of eBPF scripts, we have to give it ability to 
access all
kernel headers.

Thank you for reviewing.

>> You can even invoke such flag detector from perf via something like:
>> f = open /tmpdir/Makefile
>> fprintf(f, "obj-y := dummy.o\n");
>> fprintf(f, "$(obj)/%%.o: $(src)/%%.c\n");
>> fprintf(f, "\t@echo -n \"$(NOSTDINC_FLAGS) $(LINUXINCLUDE) $(EXTRA_CFLAGS)\"
>>> %s\n", cachefile);
>> snprintf(cmd, "make -s -C /lib/modules/%s/build M=%s dummy.o",
>>     uts_release, tmpdir);
>> system(cmd);
>> read flags from cachefile and cache it for the future.
>> ...
>> or as independent script that populates .perfconfig
>>


--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/