linux-kernel - Re: [bpf-next v6 2/3] bpftool: Update doc (add autoattach to prog load)

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] [day] [month] [year] [list]

Message-ID: <f34e0c32-34b5-bca5-b71a-5d588caf1c2f@huawei.com>
Date:   Tue, 27 Sep 2022 19:05:33 +0800
From:   wangyufen <wangyufen@...wei.com>
To:     Daniel Borkmann <daniel@...earbox.net>, <quentin@...valent.com>,
        <ast@...nel.org>, <andrii@...nel.org>, <martin.lau@...ux.dev>,
        <song@...nel.org>, <yhs@...com>, <john.fastabend@...il.com>,
        <kpsingh@...nel.org>, <sdf@...gle.com>, <haoluo@...gle.com>,
        <jolsa@...nel.org>, <davem@...emloft.net>, <kuba@...nel.org>,
        <hawk@...nel.org>, <nathan@...nel.org>, <ndesaulniers@...gle.com>,
        <trix@...hat.com>
CC:     <bpf@...r.kernel.org>, <linux-kernel@...r.kernel.org>,
        <netdev@...r.kernel.org>, <llvm@...ts.linux.dev>
Subject: Re: [bpf-next v6 2/3] bpftool: Update doc (add autoattach to prog
 load)


在 2022/9/27 0:20, Daniel Borkmann 写道:
> On 9/24/22 12:13 PM, Wang Yufen wrote:
>> Add autoattach optional to prog load|loadall for supporting
>> one-step load-attach-pin_link.
>>
>> Signed-off-by: Wang Yufen <wangyufen@...wei.com>
>> ---
>>   tools/bpf/bpftool/Documentation/bpftool-prog.rst | 13 +++++++++++--
>>   1 file changed, 11 insertions(+), 2 deletions(-)
>>
>> diff --git a/tools/bpf/bpftool/Documentation/bpftool-prog.rst 
>> b/tools/bpf/bpftool/Documentation/bpftool-prog.rst
>> index eb1b2a254eb1..2d9f27a0120f 100644
>> --- a/tools/bpf/bpftool/Documentation/bpftool-prog.rst
>> +++ b/tools/bpf/bpftool/Documentation/bpftool-prog.rst
>> @@ -31,7 +31,7 @@ PROG COMMANDS
>>   |    **bpftool** **prog dump xlated** *PROG* [{**file** *FILE* | 
>> **opcodes** | **visual** | **linum**}]
>>   |    **bpftool** **prog dump jited**  *PROG* [{**file** *FILE* | 
>> **opcodes** | **linum**}]
>>   |    **bpftool** **prog pin** *PROG* *FILE*
>> -|    **bpftool** **prog** { **load** | **loadall** } *OBJ* *PATH* 
>> [**type** *TYPE*] [**map** {**idx** *IDX* | **name** *NAME*} *MAP*] 
>> [**dev** *NAME*] [**pinmaps** *MAP_DIR*]
>> +|    **bpftool** **prog** { **load** | **loadall** } *OBJ* *PATH* 
>> [**type** *TYPE*] [**map** {**idx** *IDX* | **name** *NAME*} *MAP*] 
>> [**dev** *NAME*] [**pinmaps** *MAP_DIR*] [**autoattach**]
>>   |    **bpftool** **prog attach** *PROG* *ATTACH_TYPE* [*MAP*]
>>   |    **bpftool** **prog detach** *PROG* *ATTACH_TYPE* [*MAP*]
>>   |    **bpftool** **prog tracelog**
>> @@ -131,7 +131,7 @@ DESCRIPTION
>>             contain a dot character ('.'), which is reserved for future
>>             extensions of *bpffs*.
>>   -    **bpftool prog { load | loadall }** *OBJ* *PATH* [**type** 
>> *TYPE*] [**map** {**idx** *IDX* | **name** *NAME*} *MAP*] [**dev** 
>> *NAME*] [**pinmaps** *MAP_DIR*]
>> +    **bpftool prog { load | loadall }** *OBJ* *PATH* [**type** 
>> *TYPE*] [**map** {**idx** *IDX* | **name** *NAME*} *MAP*] [**dev** 
>> *NAME*] [**pinmaps** *MAP_DIR*] [**autoattach**]
>>             Load bpf program(s) from binary *OBJ* and pin as *PATH*.
>>             **bpftool prog load** pins only the first program from the
>>             *OBJ* as *PATH*. **bpftool prog loadall** pins all programs
>> @@ -150,6 +150,15 @@ DESCRIPTION
>>             Optional **pinmaps** argument can be provided to pin all
>>             maps under *MAP_DIR* directory.
>>   +          If **autoattach** is specified program will be attached
>> +          before pin. In that case, only the link (representing the
>> +          program attached to its hook) is pinned, not the program as
>> +          such, so the path won't show in "**bpftool prog show -f**",
>> +          only show in "**bpftool link show -f**". Also, this only 
>> works
>> +          when bpftool (libbpf) is able to infer all necessary 
>> information
>> +          from the objectfile, in particular, it's not supported for 
>> all
>> +          program types.
>
> Related to Quentin's comment, the documentation should also describe 
> clear semantics
> on what happens in failure case. I presume the use case you have in 
> mind is to use
> this facility for scripts e.g. to run/load some tests objs? Thus would 
> be good to describe
> to users what they need to do/clean up when things only partially 
> succeed etc..


Thanks for your comment.
add in v7, please check.

>
>>             Note: *PATH* must be located in *bpffs* mount. It must not
>>             contain a dot character ('.'), which is reserved for future
>>             extensions of *bpffs*.
>>
>
>