Re: [PATCH v2 0/5] tools: fix compilation failure caused by init_disassemble_info API changes

From: Andres Freund <andres@anarazel.de>
To: Jiri Olsa <olsajiri@gmail.com>
Cc: bpf@vger.kernel.org, linux-kernel@vger.kernel.org,
	Alexei Starovoitov <ast@kernel.org>,
	Arnaldo Carvalho de Melo <acme@redhat.com>,
	Sedat Dilek <sedat.dilek@gmail.com>,
	Quentin Monnet <quentin@isovalent.com>
Subject: Re: [PATCH v2 0/5] tools: fix compilation failure caused by init_disassemble_info API changes
Date: Mon, 4 Jul 2022 13:19:22 -0700	[thread overview]
Message-ID: <20220704201922.pvrh4cmmjxjn4mkx@awork3.anarazel.de> (raw)
In-Reply-To: <YsKvPW+1RkVvq8aX@krava>

Hi,

On 2022-07-04 11:13:33 +0200, Jiri Olsa wrote:
> I think the disassembler checks should not be displayed by default,
> with your change I can see all the time:
> 
> ...        disassembler-four-args: [ on  ]
> ...      disassembler-init-styled: [ OFF ]
> 
> 
> could you please squash something like below in? moving disassembler
> checks out of sight and do manual detection

Makes sense - I was wondering why disassembler-four-args is displayed, but
though it better to mirror the existing behaviour. Does "hiding"
disassembler-four-args need to be its own set of commits?

> diff --git a/tools/perf/Makefile.config b/tools/perf/Makefile.config
> index ee417c321adb..2aa0bad11f05 100644
> --- a/tools/perf/Makefile.config
> +++ b/tools/perf/Makefile.config
> @@ -914,8 +914,6 @@ ifndef NO_LIBBFD
>          FEATURE_CHECK_LDFLAGS-disassembler-init-styled += -liberty -lz -ldl
>        endif
>      endif
> -    $(call feature_check,disassembler-four-args)
> -    $(call feature_check,disassembler-init-styled)
>    endif
>  
>    ifeq ($(feature-libbfd-buildid), 1)
> @@ -1025,6 +1023,9 @@ ifdef HAVE_KVM_STAT_SUPPORT
>      CFLAGS += -DHAVE_KVM_STAT_SUPPORT
>  endif
>  
> +$(call feature_check,disassembler-four-args)
> +$(call feature_check,disassembler-init-styled)
> +
>  ifeq ($(feature-disassembler-four-args), 1)
>      CFLAGS += -DDISASM_FOUR_ARGS_SIGNATURE
>  endif

This I don't understand - why do we want these to run under NO_LIBBFD etc?

Greetings,

Andres Freund