Re: [PATCH dwarves 2/3] pahole: add reproducible_build to --btf_features

[Alan Maguire <alan.maguire@oracle.com>]

On 16/04/2024 16:24, Arnaldo Carvalho de Melo wrote:
> On Tue, Apr 16, 2024 at 03:37:17PM +0100, Alan Maguire wrote:
>> ...as a non-standard feature, so it will not be enabled for
>> "--btf_features=all"
> 
> How did you test this?
> 
> ⬢[acme@toolbox pahole]$ pahole --btf_features_strict=bgasd
> Feature 'bgasd' in 'bgasd' is not supported.  Supported BTF features are:
> encode_force,var,float,decl_tag,type_tag,enum64,optimized_func,consistent_func,reproducible_build
> ⬢[acme@toolbox pahole]$ pahole -j --btf_features=all,reproducible_build --btf_encode_detached=vmlinux.btf.parallel.reproducible_build-via-btf_features vmlinux
> ⬢[acme@toolbox pahole]$ bpftool btf dump file vmlinux.btf.parallel.reproducible_build-via-btf_features > output.vmlinux.btf.parallel.reproducible_build-via-btf_features
> ⬢[acme@toolbox pahole]$ diff -u output.vmlinux.btf.serial output.vmlinux.btf.parallel.reproducible
> ⬢[acme@toolbox pahole]$ diff -u output.vmlinux.btf.parallel.reproducible_build output.vmlinux.btf.parallel.reproducible_build-via-btf_features | head
> --- output.vmlinux.btf.parallel.reproducible_build	2024-04-16 12:20:28.513462223 -0300
> +++ output.vmlinux.btf.parallel.reproducible_build-via-btf_features	2024-04-16 12:23:37.792962930 -0300
> @@ -265,7 +265,7 @@
>  	'target' type_id=33 bits_offset=32
>  	'key' type_id=43 bits_offset=64
>  [164] PTR '(anon)' type_id=163
> -[165] PTR '(anon)' type_id=35751
> +[165] PTR '(anon)' type_id=14983
>  [166] STRUCT 'static_key' size=16 vlen=2
>  	'enabled' type_id=88 bits_offset=0
> ⬢[acme@toolbox pahole]$
> 
> I'm double checking things now...
>

The test worked for me on x86_64/aarch64. Did you test with patch 3
applied? Because the test in its original state prior to patch 3 sets
--reproducible_build before setting --btf_features=all, you won't get a
reproducible build since the command line is saying "enable reproducible
builds but also enable standard features only"; the second action undoes
the first. switching to using --btf_features=all,reproducible_build
fixes things for me.
> - Arnaldo
>  
>> Signed-off-by: Alan Maguire <alan.maguire@oracle.com>
>> ---
>>  man-pages/pahole.1 | 8 ++++++++
>>  pahole.c           | 1 +
>>  2 files changed, 9 insertions(+)
>>
>> diff --git a/man-pages/pahole.1 b/man-pages/pahole.1
>> index 2c08e97..64de343 100644
>> --- a/man-pages/pahole.1
>> +++ b/man-pages/pahole.1
>> @@ -310,6 +310,14 @@ Encode BTF using the specified feature list, or specify 'all' for all standard f
>>  	                   in different CUs.
>>  .fi
>>  
>> +Supported non-standard features (not enabled for 'all')
>> +
>> +.nf
>> +	reproducible_build Ensure generated BTF is consistent every time;
>> +	                   without this parallel BTF encoding can result in
>> +	                   inconsistent BTF ids.
>> +.fi
>> +
>>  So for example, specifying \-\-btf_encode=var,enum64 will result in a BTF encoding that (as well as encoding basic BTF information) will contain variables and enum64 values.
>>  
>>  .TP
>> diff --git a/pahole.c b/pahole.c
>> index 890ef81..38cc636 100644
>> --- a/pahole.c
>> +++ b/pahole.c
>> @@ -1286,6 +1286,7 @@ struct btf_feature {
>>  	BTF_FEATURE(enum64, skip_encoding_btf_enum64, true, true),
>>  	BTF_FEATURE(optimized_func, btf_gen_optimized, false, true),
>>  	BTF_FEATURE(consistent_func, skip_encoding_btf_inconsistent_proto, false, true),
>> +	BTF_FEATURE(reproducible_build, reproducible_build, false, false),
>>  };
>>  
>>  #define BTF_MAX_FEATURE_STR	1024
> 
> 
>