From mboxrd@z Thu Jan 1 00:00:00 1970 From: Thomas Petazzoni Date: Mon, 18 Jan 2016 23:10:51 +0100 Subject: [Buildroot] Analysis of build failures In-Reply-To: References: <20160116073021.A4B95101BCC@stock.ovh.net> <20160117001117.569d43f2@free-electrons.com> Message-ID: <20160118231051.31adbabc@free-electrons.com> List-Id: MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: buildroot@busybox.net Maxime, I know you're using GMail, but please learn to strip long e-mails when you're just replying to one specific of it :-) On Sun, 17 Jan 2016 18:10:58 +0100, Maxime Hadjinlian wrote: > > Not sure: > > > > cache.c: In function 'same_path': > > cache.c:429:22: error: field 'fh' has incomplete type > > cache.c:436:2: warning: implicit declaration of function 'name_to_handle_at' [-Wimplicit-function-declaration] > > Makefile:613: recipe for target 'mountd-cache.o' failed > > > > Maxime, can you have a look ? > I will look at it, the errors is a bit puzzling but I only glanced at > it for now. I had a look: nfs-utils does have some code to detect if name_to_handle_at() is implemented. Unfortunately, as can be seen here, using this function only triggers a warning (because its prototype is not defined in any headers of the C library), but doesn't fail to link (because for some reason the function is implemented). However the "struct file_handle" type is not available. So, I suggest to add: AC_CHECK_TYPES([struct file_handle]) to configure.ac, and then change cache.c from #if HAVE_NAME_TO_HANDLE_AT to: #if defined(HAVE_NAME_TO_HANDLE_AT) && defined(HAVE_STRUCT_FILE_HANDLE) It works with the problematic PowerPC toolchains (i.e it properly detects that struct file_handle is not available), but it remains to be tested on another toolchain. Also, this fix should be submitted upstream, obviously :-) Care to do it ? :-) Thanks! Thomas -- Thomas Petazzoni, CTO, Free Electrons Embedded Linux, Kernel and Android engineering http://free-electrons.com