Date:   Mon, 14 Sep 2020 11:06:19 -0700
From:   Yonghong Song <yhs@...com>
To:     Andrii Nakryiko <andrii.nakryiko@...il.com>
CC:     Quentin Monnet <quentin@...valent.com>, bpf <bpf@...r.kernel.org>,
        Networking <netdev@...r.kernel.org>,
        Alexei Starovoitov <ast@...nel.org>,
        Daniel Borkmann <daniel@...earbox.net>,
        Kernel Team <kernel-team@...com>,
        Andrii Nakryiko <andriin@...com>
Subject: Re: [PATCH bpf-next] bpftool: fix build failure



On 9/14/20 10:55 AM, Andrii Nakryiko wrote:
> On Mon, Sep 14, 2020 at 10:46 AM Yonghong Song <yhs@...com> wrote:
>>
>>
>>
>> On 9/14/20 10:23 AM, Quentin Monnet wrote:
>>> On 14/09/2020 17:54, Yonghong Song wrote:
>>>>
>>>>
>>>> On 9/14/20 9:46 AM, Yonghong Song wrote:
>>>>>
>>>>>
>>>>> On 9/14/20 1:16 AM, Quentin Monnet wrote:
>>>>>> On 14/09/2020 07:12, Yonghong Song wrote:
>>>>>>> When building bpf selftests like
>>>>>>>      make -C tools/testing/selftests/bpf -j20
>>>>>>> I hit the following errors:
>>>>>>>      ...
>>>>>>>      GEN
>>>>>>> /net-next/tools/testing/selftests/bpf/tools/build/bpftool/Documentation/bpftool-gen.8
>>>>>>>
>>>>>>>      <stdin>:75: (WARNING/2) Block quote ends without a blank line;
>>>>>>> unexpected unindent.
>>>>>>>      <stdin>:71: (WARNING/2) Literal block ends without a blank line;
>>>>>>> unexpected unindent.
>>>>>>>      <stdin>:85: (WARNING/2) Literal block ends without a blank line;
>>>>>>> unexpected unindent.
>>>>>>>      <stdin>:57: (WARNING/2) Block quote ends without a blank line;
>>>>>>> unexpected unindent.
>>>>>>>      <stdin>:66: (WARNING/2) Literal block ends without a blank line;
>>>>>>> unexpected unindent.
>>>>>>>      <stdin>:109: (WARNING/2) Literal block ends without a blank line;
>>>>>>> unexpected unindent.
>>>>>>>      <stdin>:175: (WARNING/2) Literal block ends without a blank line;
>>>>>>> unexpected unindent.
>>>>>>>      <stdin>:273: (WARNING/2) Literal block ends without a blank line;
>>>>>>> unexpected unindent.
>>>>>>>      make[1]: ***
>>>>>>> [/net-next/tools/testing/selftests/bpf/tools/build/bpftool/Documentation/bpftool-perf.8]
>>>>>>> Error 12
>>>>>>>      make[1]: *** Waiting for unfinished jobs....
>>>>>>>      make[1]: ***
>>>>>>> [/net-next/tools/testing/selftests/bpf/tools/build/bpftool/Documentation/bpftool-iter.8]
>>>>>>> Error 12
>>>>>>>      make[1]: ***
>>>>>>> [/net-next/tools/testing/selftests/bpf/tools/build/bpftool/Documentation/bpftool-struct_ops.8]
>>>>>>> Error 12
>>>>>>>      ...
>>>>>>>
>>>>>>> I am using:
>>>>>>>      -bash-4.4$ rst2man --version
>>>>>>>      rst2man (Docutils 0.11 [repository], Python 2.7.5, on linux2)
>>>>>>>      -bash-4.4$
>>>>>>>
>>>>>>> Looks like that particular version of rst2man prefers to have a
>>>>>>> blank line
>>>>>>> after literal blocks. This patch added block lines in related .rst
>>>>>>> files
>>>>>>> and compilation can then pass.
>>>>>>>
>>>>>>> Cc: Quentin Monnet <quentin@...valent.com>
>>>>>>> Fixes: 18841da98100 ("tools: bpftool: Automate generation for "SEE
>>>>>>> ALSO" sections in man pages")
>>>>>>> Signed-off-by: Yonghong Song <yhs@...com>
>>>>>>
>>>>>>
>>>>>> Hi Yonghong, thanks for the fix! I didn't see those warnings on my
>>>>>> setup. For the record my rst2man version is:
>>>>>>
>>>>>>       rst2man (Docutils 0.16 [release], Python 3.8.2, on linux)
>>>>>>
>>>>>> Your patch looks good, but instead of having blank lines at the end of
>>>>>> most files, could you please check if the following works?
>>>>>
>>>>> Thanks for the tip! I looked at the generated output again. My above
>>>>> fix can silent the warning, but certainly not correct.
>>>>>
>>>>> With the following change, I captured the intermediate result of the
>>>>> .rst file.
>>>>>
>>>>>     ifndef RST2MAN_DEP
>>>>>            $(error "rst2man not found, but required to generate man pages")
>>>>>     endif
>>>>> -       $(QUIET_GEN)( cat $< ; echo -n $(call see_also,$<) ) | rst2man
>>>>> $(RST2MAN_OPTS) > $@
>>>>> +       $(QUIET_GEN)( cat $< ; echo -n $(call see_also,$<) ) | tee
>>>>> /tmp/tt | rst2man $(RST2MAN_OPTS) > $@
>>>>>
>>>>> With below command,
>>>>>       make clean && make bpftool-cgroup.8
>>>>> I can get the new .rst file for bpftool-cgroup.
>>>>>
>>>>> At the end of file /tmp/tt (changed bpftool-cgroup.rst), I see
>>>>>
>>>>>        ID       AttachType      AttachFlags     Name
>>>>> \n SEE ALSO\n========\n\t**bpf**\ (2),\n\t**bpf-helpers**\
>>>>> (7),\n\t**bpftool**\ (8),\n\t**bpftool-btf**\
>>>>> (8),\n\t**bpftool-feature**\ (8),\n\t**bpftool-gen**\
>>>>> (8),\n\t**bpftool-iter**\ (8),\n\t**bpftool-link**\
>>>>> (8),\n\t**bpftool-map**\ (8),\n\t**bpftool-net**\
>>>>> (8),\n\t**bpftool-perf**\ (8),\n\t**bpftool-prog**\
>>>>> (8),\n\t**bpftool-struct_ops**\ (8)\n
>>>>>
>>>>> This sounds correct if we rst2man can successfully transforms '\n'
>>>>> or '\t' to proper encoding in the man page.
>>>>>
>>>>> Unfortunately, with my version of rst2man, I got
>>>>>
>>>>> .IP "System Message: WARNING/2 (<stdin>:, line 146)"
>>>>> Literal block ends without a blank line; unexpected unindent.
>>>>> .sp
>>>>> n SEE
>>>>> ALSOn========nt**bpf**(2),nt**bpf\-helpers**(7),nt**bpftool**(8),nt**bpftool\-btf**(8),nt**bpftool\-feature**(8),nt**bpftool\-gen**(8),nt**bpftool\-iter**(8),nt**bpftool\-link**(8),nt**bpftool\-map**(8),nt**bpftool\-net**(8),nt**bpftool\-perf**(8),nt**bpftool\-prog**(8),nt**bpftool\-struct_ops**(8)n
>>>>>
>>>>> .\" Generated by docutils manpage writer.
>>>>>
>>>>> The rst2man simply considered \n as 'n'. The same for '\t' and
>>>>> this caused the issue.
>>>>>
>>>>> I did not find a way to fix the problem https://urldefense.proofpoint.com/v2/url?u=https-3A__www.google.com_url-3Fq-3Dhttps-3A__zoom.us_j_94864957378-3Fpwd-253DbXFRL1ZaRUxTbDVKcm9uRitpTXgyUT09-26sa-3DD-26source-3Dcalendar-26ust-3D1600532408208000-26usg-3DAOvVaw3SJ0i8oz4ZM-2DGRb7hYkrYlyet&d=DwIDaQ&c=5VD0RTtNlTh3ycd41b3MUw&r=DA8e1B5r073vIqRrFz7MRA&m=kEK7ScPMF-y-i8dli-or8wWEGREW5V4qPB7UqHqDnkg&s=Br0g0MFXxL_pJuDVTOY5UrmvfD2ru_6Uf_X_8Nr2Rhk&e= .
>>>>
>>>> The following change works for me.
>>>>
>>>> @@ -44,7 +44,7 @@ $(OUTPUT)%.8: %.rst
>>>>    ifndef RST2MAN_DEP
>>>>           $(error "rst2man not found, but required to generate man pages")
>>>>    endif
>>>> -       $(QUIET_GEN)( cat $< ; echo -n $(call see_also,$<) ) | rst2man
>>>> $(RST2MAN_OPTS) > $@
>>>> +       $(QUIET_GEN)( cat $< ; echo -e $(call see_also,$<) ) | rst2man
>>>> $(RST2MAN_OPTS) > $@
>>>>
>>>>    clean: helpers-clean
>>>>           $(call QUIET_CLEAN, Documentation)
>>>> -bash-4.4$
>>>>
>>>> I will send revision 2 shortly.
>>>
>>> Thanks Yonghong, but this does not work on my setup :/. The version of
>>> echo which is called on my machine from the Makefile does not seem to
>>> interpret the "-e" option and writes something like "-e\nSEE ALSO",
>>> which causes rst2man to complain.
>>>
>>> I suspect the portable option here would be printf, even though Andrii
>>> had some concerns that we could pass a format specifier through the file
>>> names [0].
>>>
>>> Would this work on your setup?
>>>
>>>        $(QUIET_GEN)( cat $< ; printf $(call see_also,$<) ) | rst2man...
>>>
>>> Would that be acceptable?
>>
>> It works for me. Andrii originally suggested `echo -e`, but since `echo
>> -e` does not work in your environment let us use printf then. I will add
>> a comment about '%' in the bpftool man page name.
> 
> It's amazing how incompatible echo can be. But that aside, I have
> nothing against printf itself, but:
> 
> printf "%s" "whatever we want to print out"
> 
> seems like the way to go, similarly how you'd do it in your C code, no?

This won't really work :-(

-bash-4.4$ printf "%s" "\n\n"
\n\n-bash-4.4$ printf "\n\n"


-bash-4.4$

Looks like "\n" needs to be in format string to make a difference.


> 
>>
>>>
>>> [0]
>>> https://lore.kernel.org/bpf/ca595fd6-e807-ac8e-f15f-68bfc7b7dbc4@isovalent.com/T/#m01bb298fd512121edd5e77a26ed5382c0c53939e
>>>
>>> Quentin
>>>

Powered by blists - more mailing lists