Am Sonntag, 27. Oktober 2013, 13:28:02 schrieb Andrew: > 26.10.2013 17:23, KP Kirchdörfer пишет: > > Hello; > > > > Am Donnerstag, 24. Oktober 2013, 22:31:06 schrieb Andrew: > >> Hi. > >> Try to unexport LDFLAGS for kernel - in any case, it isn't dependent on > >> userland libs. > >> > >> 24.10.2013 19:22, KP Kirchdörfer пишет: > >>> Hi, > >>> > >>> the latest changes to git master branch fixed the i486 and x86 > >>> toolchain, > >>> but compilation of the kernel in the arm toolchain is broken: > >>> > >>> armv5te-unknown-linux-uclibcgnueabi-ld: unrecognized option '-Wl,- > >>> rpath,/opt/buildtool-master/staging/armv5te-unknown-linux-uclibcgnueabi/ > >>> li > >>> b' armv5te-unknown-linux-uclibcgnueabi-ld: use the --help option for > >>> usage > >>> information > > > > unexport LDFLAGS worked for the kernel, but (without further changes) more > > kernel-related packages failed (e.g. iptables, accel-pptp) So another > > solution will be preferred. > > > > For the being, I moved the LDFLAGS settings from make/Masterinclude.mk to > > the toolchain specfic make file in make/toolchain - with the latest > > changes for i486 and x86, and the old settings for arm5... > > > > So the status is now that the i486 and x86 toolchain builds and create > > bootable images. > > Hmm, how they are assembled with updated LDFLAGS for x86?
X86 has the same LDFLAGS in make/toolchain/x86...mk as i486. > This isn't a good idea to use different LDFLAGS. Maybe it'll be better > to use EXTCCLDFLAGS with rpath search as LDFLAGS for packages that fails > on library linking? So you suggest we revert the changes you made to make/Masterinclude.mk? I tried using -Wl,-rpath with pump and ipvsadm successfully to fix arm build,, maybe a better idea. keepalived and oprofile did not work that way... > > The arm toolchain builds the core but a few packages fail due to > > wrong/missing rpath (as before in i486, all packages that use libpopt). > > > > A second pb in the arm toolchain is shown when building clamav > > > > In file included from fan.c:38:0: > > fan-syscalllib.h:14:3: error: #error "System call numbers not defined for > > this architecture" > > fan-syscalllib.h: In function 'fanotify_init': > > fan-syscalllib.h:19:17: error: '__NR_fanotify_init' undeclared (first use > > in this function) > > fan-syscalllib.h:19:17: note: each undeclared identifier is reported only > > once for each function it appears in > > fan-syscalllib.h: In function 'fanotify_mark': > > fan-syscalllib.h:25:17: error: '__NR_fanotify_mark' undeclared (first use > > in this function) > > > > It would be great if we can fix these pb's, but I don't consider them as > > shostoppers, in the 5.0 series the arm toolchain isn't very useful and > > more or less a testbed. > > > > kp > > fan-syscalllib.h - what package provides it? I'll not have access to > development platform in next fer days. Haven't looked into it yet. kp ------------------------------------------------------------------------------ October Webinars: Code for Performance Free Intel webinars can help you accelerate application performance. Explore tips for MPI, OpenMP, advanced profiling, and more. Get the most from the latest Intel processors and coprocessors. See abstracts and register > http://pubads.g.doubleclick.net/gampad/clk?id=60135991&iu=/4140/ostg.clktrk _______________________________________________ leaf-devel mailing list leaf-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/leaf-devel
