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: <20160408154846.GG5327@redhat.com>
Date:	Fri, 8 Apr 2016 12:48:46 -0300
From:	Arnaldo Carvalho de Melo <acme@...hat.com>
To:	"Wangnan (F)" <wangnan0@...wei.com>
Cc:	Thierry Reding <thierry.reding@...il.com>,
	linux-kernel@...r.kernel.org, Jiri Olsa <jolsa@...nel.org>,
	Arnaldo Carvalho de Melo <acme@...nel.org>
Subject: Re: [PATCH] scripts: Fallback to KBUILD_OUTPUT if O is not specified

Em Fri, Apr 08, 2016 at 11:37:06PM +0800, Wangnan (F) escreveu:
> 
> 
> On 2016/4/8 22:58, Arnaldo Carvalho de Melo wrote:
> >Em Fri, Apr 08, 2016 at 11:14:09AM +0200, Thierry Reding escreveu:
> >>From: Thierry Reding <treding@...dia.com>
> >>
> >>The KBUILD_OUTPUT environment can be used to set the build directory as
> >>an alternative to passing the O variable on the command-line. This works
> >>with regular kernel builds, so it's surprising if it doesn't work within
> >>the tools/scripts directory.
> >>
> >>Allow KBUILD_OUTPUT if O is not specified on the command-line to avoid
> >>any surprises.
> >I think this is fine, Jiri, Wang?
> 
> The idea is fine, but we should be very careful when touching Makefile

Yeah, I haven't tested it, Thierry, would you try the 'make -C
tools/perf build-test' with this set up?

> to avoid breaking build-test like this:
> 
>  # export KBUILD_OUTPUT=/tmp/xxxxx
>  #  ./tests/perf-targz-src-pkg .
>    PERF_VERSION = 4.5.g337b7b
>  ls: cannot access perf-*.tar.gz: No such file or directory
>  tar: -C: Cannot open: No such file or directory
>  tar: Error is not recoverable: exiting now
>  make: *** /tmp/tmp.q4egWrFd3L/perf*/tools/perf: No such file or
> directory.  Stop
> 
> Moreover, in the above case, the output directory should be
> 
>  /tmp/xxxxx
> 
> or
> 
>  /tmp/xxxxx/tools/perf ?
> 
> Since kernel build output things to /tmp/xxxxx, I think we'd better
> output to /tmp/xxxxx/tools/perf.
> 
> Thank you.
> 
> >- Arnaldo
> >>Signed-off-by: Thierry Reding <treding@...dia.com>
> >>---
> >>Hi Arnaldo,
> >>
> >>I'm not sure if you're the right person to send this to. There was noone
> >>reported by scripts/get_maintainer.pl, but it looks like you ended up
> >>applying most of the patches to this file lately, which kinda makes you
> >>the de facto maintainer =).
> >>
> >>Thierry
> >>
> >>  tools/scripts/Makefile.include | 4 ++++
> >>  1 file changed, 4 insertions(+)
> >>
> >>diff --git a/tools/scripts/Makefile.include b/tools/scripts/Makefile.include
> >>index 8abbef164b4e..5467da41dc72 100644
> >>--- a/tools/scripts/Makefile.include
> >>+++ b/tools/scripts/Makefile.include
> >>@@ -8,6 +8,10 @@ ifeq ($(objtree),)
> >>  	objtree := $(O)
> >>  endif
> >>  endif
> >>+else
> >>+ifneq ($(KBUILD_OUTPUT),)
> >>+	OUTPUT := $(KBUILD_OUTPUT)/$(if $(subdir),$(subdir)/)
> >>+endif
> >>  endif
> >>  # check that the output directory actually exists
> >>-- 
> >>2.8.0
> 

Powered by blists - more mailing lists

Powered by Openwall GNU/*/Linux Powered by OpenVZ