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: <20170312115603.GA10470@WeideMacBook-Pro.local>
Date:   Sun, 12 Mar 2017 19:56:03 +0800
From:   Wei Yang <richard.weiyang@...il.com>
To:     Masahiro Yamada <yamada.masahiro@...ionext.com>
Cc:     Wei Yang <richard.weiyang@...il.com>,
        Michal Marek <mmarek@...e.com>,
        Linux Kbuild mailing list <linux-kbuild@...r.kernel.org>,
        Linux Kernel Mailing List <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH] kbuild: strip the last slash in KBUILD_EXTMOD

On Sat, Mar 11, 2017 at 10:34:50PM +0900, Masahiro Yamada wrote:
>Hi Wei,
>
>
>2017-02-23 0:08 GMT+09:00 Wei Yang <richard.weiyang@...il.com>:
>> Current kbuild will build the target again if we run "make M=dir" and "make
>> M=dir/" by turns, since if_changed will see the prerequisite is changed.
>> The behavior may confuse the user a little, since actually we are building
>> the same target and no difference.
>>
>> According to current implementation in scripts/Makefile.build, the obj
>> passed to next level is a directory name with last slash stripped. This
>> patch strips the last slash in KBUILD_EXTMOD which is assigned from "M=".
>>
>> Signed-off-by: Wei Yang <richard.weiyang@...il.com>
>> ---
>>  Makefile | 6 ++++--
>>  1 file changed, 4 insertions(+), 2 deletions(-)
>>
>> diff --git a/Makefile b/Makefile
>> index b1037774e8e8..db6ed60928ab 100644
>> --- a/Makefile
>> +++ b/Makefile
>> @@ -184,13 +184,15 @@ endif
>>  # Old syntax make ... SUBDIRS=$PWD is still supported
>>  # Setting the environment variable KBUILD_EXTMOD take precedence
>>  ifdef SUBDIRS
>> -  KBUILD_EXTMOD ?= $(SUBDIRS)
>> +  __KBUILD_EXTMOD ?= $(SUBDIRS)
>>  endif
>>
>>  ifeq ("$(origin M)", "command line")
>> -  KBUILD_EXTMOD := $(M)
>> +  __KBUILD_EXTMOD := $(M)
>>  endif
>>
>> +KBUILD_EXTMOD = $(patsubst %/,%,$(__KBUILD_EXTMOD))
>> +
>>  # If building an external module we do not care about the all: rule
>>  # but instead _all depend on modules
>>  PHONY += all
>
>This patch would break a use-case
>where KBUILD_EXTMOD is given as an environment.
>

Hi, Masahiro 

Thanks for your comment.

Got your point, you are right.

>
>
>Maybe adding the following should be enough?
>
>KBUILD_EXTMOD := $(patsubst %/,%,$(KBUILD_EXTMOD))
>

Yep, it looks good.

BTW, I found another interesting thing.

     make M=dir//// 

is also a valid target in kbuild.

Even this is a very rare case, I plan to use this to handle it.

KBUILD_EXTMOD := $(subst $(CURDIR)/,,$(KBUILD_EXTMOD))

This also works for an out tree build.

I would send out v2 soon, hope you like this.

>
>-- 
>Best Regards
>Masahiro Yamada

-- 
Wei Yang
Help you, Help me

Download attachment "signature.asc" of type "application/pgp-signature" (820 bytes)

Powered by blists - more mailing lists

Powered by Openwall GNU/*/Linux Powered by OpenVZ