Re: [PATCHv2 00/15] tools build: Move features framework into tools/build

From: Arnaldo Carvalho de Melo
Date: Wed Mar 18 2015 - 14:51:03 EST


Em Wed, Mar 11, 2015 at 09:58:49AM +0100, Jiri Olsa escreveu:
> hi,
> moving features detection framework from perf to 'tools/build'
> plus other minor fixes.
>
> v2 changes:
> - omitted patch 1 from original version (Ingo)
> - added features build output dump (Arnaldo)
> - better scripting in features test (Namhyung)
> - fixed test-all for pthread-attr-setaffinity-np
> - removed main_test_libbabeltrace from test-all
> - patches reordered
>
> It's also available in:
> git://git.kernel.org/pub/scm/linux/kernel/git/jolsa/perf.git
> perf/build

Ok, so using the csets from the above branch, if I do:

[acme@ssdandy linux]$ rm -rf /tmp/build/perf ; mkdir -p /tmp/build/perf ; time make O=/tmp/build/perf -C tools/perf LIBBABELTRACE_DIR=/opt/libbabeltrace/ install-bin
make: Entering directory `/home/acme/git/linux/tools/perf'
BUILD: Doing 'make -j8' parallel build
config/Makefile:423: No libunwind found. Please install libunwind-dev[el] >= 1.1 and/or set LIBUNWIND_DIR

Auto-detecting system features:
... dwarf: [ on ]
... glibc: [ on ]
... gtk2: [ on ]
... libaudit: [ on ]
... libbfd: [ on ]
... libelf: [ on ]
... libnuma: [ on ]
... libperl: [ on ]
... libpython: [ on ]
... libslang: [ on ]
... libunwind: [ OFF ]
... libdw-dwarf-unwind: [ on ]
... zlib: [ on ]
... DWARF post unwind library: libdw

And then, despite not telling anything about babeltrace being linked,
i.e. the no line above with a [ on ], we get:

[acme@ssdandy linux]$ ldd ~/bin/perf | grep babel
libbabeltrace-ctf.so.1 => /opt/libbabeltrace/lib/libbabeltrace-ctf.so.1 (0x00007f5f7b758000)
libbabeltrace.so.1 => /opt/libbabeltrace/lib/libbabeltrace.so.1 (0x00007f5f7acfb000)
[acme@ssdandy linux]$

I.e. if a libbabeltrace-devel was installed and had what is needed to build it
in, when there would be no need for the LIBBABELTRACE_DIR above and babeltrace
wouldn't be either implicitely or explicitely specified as something we want to
build against we would not show it as a feature being built in (no
libbabeltrace: [ On ]) but would do it anyway, is that really the intention?

I tested in fedora, where no system auto-detection fails, i.e. test-all passes,
and the result is the same.

Ingo, can you clarify here?

- Arnaldo
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/