Message-ID: <557B0FB4.6000506@plumgrid.com>
Date: Fri, 12 Jun 2015 09:58:28 -0700
From: Alexei Starovoitov <ast@plumgrid.com>
User-Agent: Mozilla/5.0 (Macintosh; Intel Mac OS X 10.10; rv:31.0) Gecko/20100101 Thunderbird/31.7.0
MIME-Version: 1.0
To: Wang Nan <wangnan0@huawei.com>, acme@kernel.org, brendan.d.gregg@gmail.com,
        daniel@iogearbox.net, namhyung@kernel.org,
        masami.hiramatsu.pt@hitachi.com, paulus@samba.org,
        a.p.zijlstra@chello.nl, mingo@redhat.com, jolsa@kernel.org,
        dsahern@gmail.com
CC: linux-kernel@vger.kernel.org, lizefan@huawei.com, hekuang@huawei.com,
        xiakaixu@huawei.com, pi3orama@163.com
Subject: Re: [RFC PATCH v7 00/37] perf tools: filtering events using eBPF
 programs
References: <1434087345-127225-1-git-send-email-wangnan0@huawei.com>
In-Reply-To: <1434087345-127225-1-git-send-email-wangnan0@huawei.com>
Content-Type: text/plain; charset=windows-1252; format=flowed
Content-Transfer-Encoding: 7bit
Sender: linux-kernel-owner@vger.kernel.org
Content-Length: 3020
Lines: 76

On 6/11/15 10:35 PM, Wang Nan wrote:
> This is the 7th version which tries to introduce eBPF programs to perf.
> It enables 'perf record' to filter events using eBPF programs like:
>
>   # perf record --event bpf-file.c sleep 1
>
> and
>
>   # perf record --event bpf-file.o sleep 1
>
> This patch series is based on tip/perf/core (028c63b).
>
> Compare with v6 patch, this series totally refactor code for llvm
> compiling '.c' into BPF object file using LLVM. Including:
>
>   1. A specific file tools/perf/util/llvm-utils.c is introduced for
>      holding all llvm/clang related work, instead of putting them into
>      bpf-loader.c.
>
>   2. Automatically detect kernel include options by embedded shell
>      script.
>
>   3. Use command line template to generate compiler commands instead of
>      assemble cmdline with printf, passing variables using environment.
>      Which enable users to define their own compiler commands.
>
>   4. Introduce '[llvm]' section to perf default config.
>      5 options can be configured, but all of them can be omitted:
>
>    [llvm]
>          # Path to clang. If omit, search it from $PATH.
> 	clang-path = "/path/to/clang"
>
>          # Cmdline template. Following line shows its default value.
>          # Environment variable is used to passing options.
> 	clang-bpf-cmd-template = "$CLANG_EXEC $CLANG_OPTIONS $KERNEL_INC_OPTIONS -Wno-unused-value -Wno-pointer-sign -working-directory $WORKING_DIR  -c $CLANG_SOURCE -target bpf -O2 -o -"
>
>          # Option passed to clang, will be passed to cmdline by $CLANG_OPTIONS.
> 	clang-opt = "-Wno-unused-value -Wno-pointer-sign"
>
>          # kbuild directory. If not set, use /lib/modules/`uname -r`/build.
>          # If set to "" deliberately, skip kernel header auto-detector.
> 	kbuild-dir = "/path/to/kernel/build"
>
>          # Options passed to 'make' when detecting kernel header options.
> 	kbuild-opts = "ARCH=x86_64"
>
>      Command line options '--clang-path' and '--clang-opt' is appended to
>      'perf record'.
>
>   5. LLVM and BPF opening testcase is adding into 'perf test':
>
>      # perf test  37
>      37: Test LLVM searching and compiling                        : Ok
>
>   6. Warning and error messages improvement.
>
>   7. Issue clang command suit for newest clang which support
>      '-target bpf', instead of calling
>      'clang -emit-llvm | llc -march=bpf' pipe.

All looks great to me.
Piping clang into stdout to skip .o on a disk is a nice touch.
'usability hints' have improved as well.
Thanks a lot for the hard work.

btw, what compile times do you see?
On my machine compiling basic hello_world.c with #include bpf_helpers.h
and few kernel headers take: 0.02 sec
So using .c is quite instant. Feels like interpreted language ;)

--
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/