Re: [OpenWrt-Devel] GCC-final fails to compile (trunk - svn://svn.openwrt.org/openwrt/tr...@18353)
Hi All, the problem seems to be a side effect of the revision 18266 gcc; fix build with cloog/ppl. The problem, I think, that using the --with-host-stdc++=... the (host)-ld.so tries to load the wrong libgcc library (target instead host version) failing the prepropcessor check. My question is: can we disable cloog/ppl support for now? thanks and best regards, luigi martedì 10 novembre 2009 11:32 Imre Kaloz ka...@openwrt.org ha scritto: On Tue, 10 Nov 2009 10:49:38 +0100, Luigi 'Comio' Mantellini luigi.mantellini...@gmail.com wrote: This is the stderr log and find the attached config.log: snip checking for C compiler default output file name... configure: error: in `/mnt/devel/olt-ppc/OpenWRT-idf/build_dir/toolchain- powerpc_gcc-4.4.1_uClibc-0.9.30.1/gcc-4.4.1-final/powerpc-openwrt-linux- uclibc/libstdc++-v3': configure: error: C compiler cannot create executables Seems like the issue we had before changesets 18177 and 18178.. Is your checkout up to date? Imre -- Luigi 'Comio' Mantellini RD - Software Industrie Dial Face S.p.A. Via Canzo, 4 20068 Peschiera Borromeo (MI), Italy Tel.: +39 02 5167 2813 Fax: +39 02 5167 2459 Email: luigi.mantell...@idf-hit.com ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org https://lists.openwrt.org/mailman/listinfo/openwrt-devel
Re: [OpenWrt-Devel] GCC-final fails to compile (trunk - svn://svn.openwrt.org/openwrt/tr...@18353)
Thanks early fix. Tomorrow I will try to change the with-host-batelapesca following this: http://gcc.gnu.org/wiki/Graphite_Build adding the explicit libstdc++ path. I will inform you on my experiment. best regards, luigi giovedì 12 novembre 2009 14:42 Felix Fietkau n...@openwrt.org ha scritto: Luigi 'Comio' Mantellini wrote: Hi All, the problem seems to be a side effect of the revision 18266 gcc; fix build with cloog/ppl. The problem, I think, that using the --with-host-stdc++=... the (host)-ld.so tries to load the wrong libgcc library (target instead host version) failing the prepropcessor check. My question is: can we disable cloog/ppl support for now? Done. Committed in r18403 - Felix -- Luigi 'Comio' Mantellini RD - Software Industrie Dial Face S.p.A. Via Canzo, 4 20068 Peschiera Borromeo (MI), Italy Tel.: +39 02 5167 2813 Fax: +39 02 5167 2459 Email: luigi.mantell...@idf-hit.com ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org https://lists.openwrt.org/mailman/listinfo/openwrt-devel
[OpenWrt-Devel] GCC-final fails to compile (trunk)
Hi Guys, I'm trying to compile a build for a powerpc target. I experienced a failure during the gcc-final compilation phase (when the Makefile tries to configure the target (powerpc) libstd++ library). The error seems to be a wrong dynamic linkage when xgcc is called (ld tries to link the target libgcc library instead the host (i386) libgcc). Is there anyone that experienced the same trouble? best regards, luigi -- Luigi 'Comio' Mantellini RD - Software Industrie Dial Face S.p.A. Via Canzo, 4 20068 Peschiera Borromeo (MI), Italy Tel.: +39 02 5167 2813 Fax: +39 02 5167 2459 Email: luigi.mantell...@idf-hit.com ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org https://lists.openwrt.org/mailman/listinfo/openwrt-devel
Re: [OpenWrt-Devel] GCC-final fails to compile (trunk - svn://svn.openwrt.org/openwrt/tr...@18353)
This is the stderr log and find the attached config.log: Checking multilib configuration for libstdc++-v3... mkdir -p -- powerpc-openwrt-linux-uclibc/libstdc++-v3 Configuring in powerpc-openwrt-linux-uclibc/libstdc++-v3 configure: creating cache ./config.cache checking build system type... i386-redhat-linux-gnu checking host system type... powerpc-openwrt-linux-uclibc checking target system type... powerpc-openwrt-linux-uclibc checking for a BSD-compatible install... /usr/bin/install -c checking whether build environment is sane... yes checking for gawk... gawk checking whether make sets $(MAKE)... yes checking for powerpc-openwrt-linux-uclibc-strip... /mnt/devel/olt-ppc/OpenWRT- idf/staging_dir/toolchain-powerpc_gcc-4.4.1_uClibc-0.9.30.1/usr/powerpc- openwrt-linux-uclibc/bin/strip checking how to run the C preprocessor... /lib/cpp checking for egrep... grep -E checking for powerpc-openwrt-linux-uclibc-gcc... /mnt/devel/olt-ppc/OpenWRT- idf/build_dir/toolchain-powerpc_gcc-4.4.1_uClibc-0.9.30.1/gcc-4.4.1- final/./gcc/xgcc -B/mnt/devel/olt-ppc/OpenWRT-idf/build_dir/toolchain- powerpc_gcc-4.4.1_uClibc-0.9.30.1/gcc-4.4.1-final/./gcc/ -B/mnt/devel/olt- ppc/OpenWRT-idf/staging_dir/toolchain- powerpc_gcc-4.4.1_uClibc-0.9.30.1/usr/powerpc-openwrt-linux-uclibc/bin/ - B/mnt/devel/olt-ppc/OpenWRT-idf/staging_dir/toolchain- powerpc_gcc-4.4.1_uClibc-0.9.30.1/usr/powerpc-openwrt-linux-uclibc/lib/ - isystem /mnt/devel/olt-ppc/OpenWRT-idf/staging_dir/toolchain- powerpc_gcc-4.4.1_uClibc-0.9.30.1/usr/powerpc-openwrt-linux-uclibc/include - isystem /mnt/devel/olt-ppc/OpenWRT-idf/staging_dir/toolchain- powerpc_gcc-4.4.1_uClibc-0.9.30.1/usr/powerpc-openwrt-linux-uclibc/sys-include checking for C compiler default output file name... configure: error: in `/mnt/devel/olt-ppc/OpenWRT-idf/build_dir/toolchain- powerpc_gcc-4.4.1_uClibc-0.9.30.1/gcc-4.4.1-final/powerpc-openwrt-linux- uclibc/libstdc++-v3': configure: error: C compiler cannot create executables See `config.log' for more details. make[5]: *** [configure-target-libstdc++-v3] Error 1 make[5]: Leaving directory `/mnt/devel/olt-ppc/OpenWRT- idf/build_dir/toolchain-powerpc_gcc-4.4.1_uClibc-0.9.30.1/gcc-4.4.1-final' make[4]: *** [all] Error 2 make[4]: Leaving directory `/mnt/devel/olt-ppc/OpenWRT- idf/build_dir/toolchain-powerpc_gcc-4.4.1_uClibc-0.9.30.1/gcc-4.4.1-final' make[3]: *** [/mnt/devel/olt-ppc/OpenWRT-idf/staging_dir/toolchain- powerpc_gcc-4.4.1_uClibc-0.9.30.1/stamp/.gcc_installed] Error 2 make[3]: Leaving directory `/mnt/devel/olt-ppc/OpenWRT-idf/toolchain/gcc' make[2]: *** [toolchain/gcc/install] Error 2 make[2]: Leaving directory `/mnt/devel/olt-ppc/OpenWRT-idf' make[1]: *** [/mnt/devel/olt-ppc/OpenWRT-idf/staging_dir/toolchain- powerpc_gcc-4.4.1_uClibc-0.9.30.1/stamp/.toolchain_install] Error 2 make[1]: Leaving directory `/mnt/devel/olt-ppc/OpenWRT-idf' make: *** [world] Error 2 Il martedì 10 novembre 2009 10:41:07 Luigi 'Comio' Mantellini ha scritto: Hi Guys, I'm trying to compile a build for a powerpc target. I experienced a failure during the gcc-final compilation phase (when the Makefile tries to configure the target (powerpc) libstd++ library). The error seems to be a wrong dynamic linkage when xgcc is called (ld tries to link the target libgcc library instead the host (i386) libgcc). Is there anyone that experienced the same trouble? best regards, luigi -- Luigi 'Comio' Mantellini RD - Software Industrie Dial Face S.p.A. Via Canzo, 4 20068 Peschiera Borromeo (MI), Italy Tel.: +39 02 5167 2813 Fax: +39 02 5167 2459 Email: luigi.mantell...@idf-hit.com This file contains any messages produced by compilers while running configure, to aid debugging if configure makes a mistake. It was created by package-unused configure version-unused, which was generated by GNU Autoconf 2.59. Invocation command line was $ /mnt/devel/olt-ppc/OpenWRT-idf/build_dir/toolchain-powerpc_gcc-4.4.1_uClibc-0.9.30.1/gcc-4.4.1/libstdc++-v3/configure --cache-file=./config.cache --with-cross-host=i386-redhat-linux --prefix=/mnt/devel/olt-ppc/OpenWRT-idf/staging_dir/toolchain-powerpc_gcc-4.4.1_uClibc-0.9.30.1/usr --with-gnu-ld --enable-target-optspace --disable-libgomp --disable-libmudflap --disable-multilib --disable-nls --with-host-libstdcxx=-lstdc++ --with-gmp=/mnt/devel/olt-ppc/OpenWRT-idf/staging_dir/host --with-mpfr=/mnt/devel/olt-ppc/OpenWRT-idf/staging_dir/host --disable-decimal-float --disable-libssp --disable-__cxa_atexit --enable-shared --enable-threads --disable-tls --with-slibdir=/mnt/devel/olt-ppc/OpenWRT-idf/staging_dir/toolchain-powerpc_gcc-4.4.1_uClibc-0.9.30.1/lib --enable-languages=c,c++ --program-transform-name=s,^,powerpc-openwrt-linux-uclibc-, --with-target-subdir=powerpc-openwrt-linux-uclibc --build=i386-redhat-linux --host=powerpc-openwrt-linux-uclibc --target=powerpc-openwrt-linux-uclibc --srcdir=/mnt/devel/olt-ppc/OpenWRT-idf/build_dir/toolchain-powerpc_gcc-4.4.1_uClibc-0.9.30.1/gcc-4.4.1/libstdc++-v3
[OpenWrt-Devel] Regarding make all / make install usage
Hi All, I'm using the parallel make (see my previous parallel/distcc compilation patchset). I noticed that a lof packages use the following make invocation: $(MAKE) all install all followed by install target. In general this is not correct for the following: 1) Not all Makefiles have a rule like install: all to assure the correct sequence 2) gmake doens't assure that the targets sequence in the command line will be respected 3) parallel make will file for 1) and 2) I would like to see all packets that use this poor practices to convert the single make call to distinct make invactions: $(MAKE) all $(MAKE) install This simple rule assures to cover the previous 1) 2) 3) points. I hope that this mail is useful for the future developments. Best regards, luigi -- Luigi 'Comio' Mantellini RD - Software Industrie Dial Face S.p.A. Via Canzo, 4 20068 Peschiera Borromeo (MI), Italy Tel.: +39 02 5167 2813 Fax: +39 02 5167 2459 Email: luigi.mantell...@idf-hit.com ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org https://lists.openwrt.org/mailman/listinfo/openwrt-devel
Re: [OpenWrt-Devel] [RFC] Distcc support
Hi All, I noticed some packages (like ncurses, ppp, ...) that don't like the parallel make (with -jX X1). My proposal is to define: MAKEJ1=$(MAKE) MAKE:=$(MAKE) -j$(CONFIG_JOBS) and use MAKEJ1 for these packages until we don't fix the issue. is it ok? ciao luigi Il domenica 18 ottobre 2009 18:39:05 Felix Fietkau ha scritto: Luigi Mantellini wrote: Hi Felix On Sun, Oct 18, 2009 at 4:22 PM, Felix Fietkau n...@openwrt.org wrote: Though you prefer this option, I believe using external toolchain support in OpenWrt will always be the exception, not the rule. We can easily add toolchains like CodeSourcery's or ELDK's to our own build infrastructure, as their sources are open as well. Even external toolchain support itself does not guarantee any form of consistency, since you could have accidentally used the wrong toolchain and the system might not notice. Because of that, I also consider the dependency on external toolchains to be bogus. Ok. I agree. I know that external is not equivalent to say reliable. Here's how I would ensure consistency: - Change OpenWrt toolchain identification string to include the arch - On each distcc build host, compile a simple test program such as 'int main(int argc, char **argv){ return 0; }' to assembler and grep for the .ident string and compare it against the same thing on the local machine. ok good idea. Anyway, I include my last patch version. I just tested compiling a kernel and busybox of an arm target using the codesourcery toolchain... it seems to work fine but I interrupted the compilation (today is sunday... and I need to see also the sun). We can consider this patch as start point, if you agree. Yep, sounds good. By the way, I'm currently working on integrating a gcc version with CodeSourcery patches into OpenWrt. Maybe that way you also won't have to use external toolchains anymore. I've stripped the diff of the latest codesourcery toolchain against upstream gcc 4.3.3 down to around 2M in size and merged it with our other patches. I'm currently checking whether it compiles... - Felix ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org https://lists.openwrt.org/mailman/listinfo/openwrt-devel -- Luigi 'Comio' Mantellini RD - Software Industrie Dial Face S.p.A. Via Canzo, 4 20068 Peschiera Borromeo (MI), Italy Tel.: +39 02 5167 2813 Fax: +39 02 5167 2459 Email: luigi.mantell...@idf-hit.com ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org https://lists.openwrt.org/mailman/listinfo/openwrt-devel
Re: [OpenWrt-Devel] [RFC] Distcc support
Hi All, I noticed that a lot of packages call make passing both all and install target at the same time. This is not right because we cannot sure that a rule like install: all is present in the package's makifile. To fix the problem, we need to do two different make calls, the first with all target and the second with install. After this, most packages should me parallel-make-aware. (I will send the patches asap). I Have another RFC: 1) define PKGMAKE:=$(MAKE) -j$(CONFIG_JOBS) 2) define PKGMAKEJ1:=$(MAKE) -j1 My idea is to parallelize also the OpenWRT makefiles... what do you think about this? best regards, luigi Il lunedì 19 ottobre 2009 12:39:20 Luigi 'Comio' Mantellini ha scritto: Hi All, I noticed some packages (like ncurses, ppp, ...) that don't like the parallel make (with -jX X1). My proposal is to define: MAKEJ1=$(MAKE) MAKE:=$(MAKE) -j$(CONFIG_JOBS) and use MAKEJ1 for these packages until we don't fix the issue. is it ok? ciao luigi Il domenica 18 ottobre 2009 18:39:05 Felix Fietkau ha scritto: Luigi Mantellini wrote: Hi Felix On Sun, Oct 18, 2009 at 4:22 PM, Felix Fietkau n...@openwrt.org wrote: Though you prefer this option, I believe using external toolchain support in OpenWrt will always be the exception, not the rule. We can easily add toolchains like CodeSourcery's or ELDK's to our own build infrastructure, as their sources are open as well. Even external toolchain support itself does not guarantee any form of consistency, since you could have accidentally used the wrong toolchain and the system might not notice. Because of that, I also consider the dependency on external toolchains to be bogus. Ok. I agree. I know that external is not equivalent to say reliable. Here's how I would ensure consistency: - Change OpenWrt toolchain identification string to include the arch - On each distcc build host, compile a simple test program such as 'int main(int argc, char **argv){ return 0; }' to assembler and grep for the .ident string and compare it against the same thing on the local machine. ok good idea. Anyway, I include my last patch version. I just tested compiling a kernel and busybox of an arm target using the codesourcery toolchain... it seems to work fine but I interrupted the compilation (today is sunday... and I need to see also the sun). We can consider this patch as start point, if you agree. Yep, sounds good. By the way, I'm currently working on integrating a gcc version with CodeSourcery patches into OpenWrt. Maybe that way you also won't have to use external toolchains anymore. I've stripped the diff of the latest codesourcery toolchain against upstream gcc 4.3.3 down to around 2M in size and merged it with our other patches. I'm currently checking whether it compiles... - Felix ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org https://lists.openwrt.org/mailman/listinfo/openwrt-devel -- Luigi 'Comio' Mantellini RD - Software Industrie Dial Face S.p.A. Via Canzo, 4 20068 Peschiera Borromeo (MI), Italy Tel.: +39 02 5167 2813 Fax: +39 02 5167 2459 Email: luigi.mantell...@idf-hit.com ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org https://lists.openwrt.org/mailman/listinfo/openwrt-devel
Re: [OpenWrt-Devel] Regarding external cross-toolchain
Hi Nico, thank you for this commit :D I'm very happy for this new and fundamental feature. best regards, luigi Il mercoledì 23 settembre 2009 04:55:37 Nico ha scritto: Hi Luigi et all, External toolchain support, based on my previous patches and your additions, was committed in r17682. Here's what I changed : - use target name for naming build staging directories, instead on unknown gcc/libc versions - use the same way of specifying libraries in ./package/base-files for external native toolchains - make the toolchain options menu disappear when external toolchain support is selected I'll check the changes in ./package later today as it seems some of them have no effects (in hotplug uci iirc) Thanks again for your help ! Regards, -- Nico Luigi Mantellini wrote: I lost some word... Hi All, I rebased the patch on the last tr...@17618. I hope to see the external-toolchain patch in the main line. best regards, luigi On Sun, Sep 20, 2009 at 5:15 PM, Luigi Mantellini luigi.mantellini...@gmail.com wrote: Hi All, I rebased the patch on the last tr...@17618 the external-toolchain patch in the main line. best regards, luigi On Thu, Sep 10, 2009 at 12:56 AM, Luigi Mantellini luigi.mantellini...@gmail.com wrote: Thanks :D On Thu, Sep 10, 2009 at 12:17 AM, Nicon...@openwrt.org wrote: Hi Luigi, There were no particular reasons preventing external toolchain support from hitting trunk apart from a constant lack of time on my side... I'll try my best to have it commited by the end of the week. Thanks for your help and support ! Regards, -- Nico Luigi 'Comio' Mantellini wrote: Hi List, Hi Nico, find the attached my contribute. Starting from nico's patch I build this patchset that adds the ability to use use an external toolchain (rebased on r17459). I compiled a simple ppc440 target with and without the external toolchain usage and it seems to work fine. As external toolchain I used the CodeSourcery G++lite 4.3-50 for Power EABI (http://www.codesourcery.com/sgpp/lite/power/portal/release603). Find also the config.external and config.normal configs to look my board setup. My wish is to see this patch in the main-line as soon possible because is very important for my job. Any feedback is welcome. best regards, luigi Il domenica 6 settembre 2009 20:26:46 Luigi Mantellini ha scritto: Hi List, as in the past I will ask again opinions regarding the optional support for external cross-toolchain. Reading the dev.openwrt.org I see: OpenWrt is currently being used in industrial control systems, small robots, sensor networks, home control solutions, VoIP systems, both wired and wireless networks and a whole lot more. I'm working on network devices and usually I receive from my vendors the BSP and toolchain that is considered official and safe for the BSP. My suppliers usually don't accent to support BSP (and anwser to my questions) if I don't use the provided toolchain. From these considerations, My opinion is that external target-specific toolchian should be supported by OpenWRT avoiding to compile everytime the toolchain. In past Nico sent a set of patch for support this feature but I think that are not included yet in the mainline. Nico, can you provide me your last patches? and explain if them will be included or not? thanks in advance and best regards, luigi ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org https://lists.openwrt.org/mailman/listinfo/openwrt-devel -- Luigi 'Comio' Mantellini RD - Software Industrie Dial Face S.p.A. Via Canzo, 4 20068 Peschiera Borromeo (MI), Italy Tel.: +39 02 5167 2813 Fax: +39 02 5167 2459 web: www.idf-hit.com mail: luigi.mantell...@idf-hit.com -- Luigi 'Comio' Mantellini RD - Software Industrie Dial Face S.p.A. Via Canzo, 4 20068 Peschiera Borromeo (MI), Italy Tel.: +39 02 5167 2813 Fax: +39 02 5167 2459 web: www.idf-hit.com mail: luigi.mantell...@idf-hit.com ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org https://lists.openwrt.org/mailman/listinfo/openwrt-devel -- Luigi 'Comio' Mantellini RD - Software Industrie Dial Face S.p.A. Via Canzo, 4 20068 Peschiera Borromeo (MI) Italy Tel.: +39 02 5167 2813 Fax: +39 02 5167 2459 E-mail: luigi.mantell...@idf-hit.com Questo messaggio e i suoi allegati sono indirizzati esclusivamente alle persone indicate. La diffusione, copia o qualsiasi altra azione derivante dalla conoscenza di queste informazioni sono rigorosamente vietate. Qualora abbiate ricevuto questo documento per errore siete cortesemente pregati di darne immediata comunicazione al mittente e di provvedere alla sua distruzione, Grazie. Rispetta l'ambiente. Non stampare questa
Re: [OpenWrt-Devel] external toolchain
Il martedì 8 settembre 2009 11:14:43 edgar.sol...@web.de ha scritto: just out of interest, not for me. It's really important for my every-day activities. What is an external toolchain? An external (Cross)toolchain is a Toolchain (0) provided by another supplier. It can based on opensource technologies or not. In general you will have from your supplier a (1) well-tested gnu/gcc based toolchain: the supplier will (2) cover all anomalies an will (3) give you support when you need, under (4) specific support agreement. Usually the (5) BSP provided from your CPU vender will include the sources for the target (kernel, userland tools, drivers) and an (6) official toolchain (already gnu/gcc based) that will be able to compile your target. The (6) official toolchain usually has a lot for (7) vendor specific patches in order to cover and resolve some problem or cpu details. When you are developing a new target (a new router or an appliance) usually you need to be supported by cpu supplier and usually the supplier requires you to use only the BSP to create the target otherwise the supplier will not help you. Typical conversation between Me and supplier field engineer (FE): Me: Good morning, I have a strange error during the kernel building error: A strange error. I think that you code is broken because blah blah blah... FE: Are you using our gcc/glibc/or_another_piece_of_bin provided by us? Me: No, I'm using a self-made vanilla gcc/glibc/... FE: Sorry, I cannot help you because: our code is perfect and we support only if you use certified building tools (included into the BSP). Good bye... Me: (After closing the telephone) I hate my FE Of course this a joke, but it's correct that FE requires a certified tool in order to have a standard background to solve the problem... Using a self made toolchain is not always possible to reproduce the behavior during the troubleshooting (at least when the problem is not trivial...) and I cannot pretend that my supplier recompile a new toolchain just for me. Finally, I external toolchain is (8) already ready to use and it's (9) easy to redistribute to all team members involved into a project... and it's important that all team members (19) have the same tools to develop. These motivations are for me sufficient to consider an external toolchain during the product development. I want underline that OpenWRT can be used as building environment and not only for homebrew activities. I hope this would answer your questions. best regards, luigi Why would you not want to use the buildroot one? Thanks alot, ede Hi List, Hi Nico, find the attached my contribute. Starting from nico's patch I build this patchset that adds the ability to use use an external toolchain (rebased on r17459). I compiled a simple ppc440 target with and without the external toolchain usage and it seems to work fine. As external toolchain I used the CodeSourcery G++lite 4.3-50 for Power EABI (http://www.codesourcery.com/sgpp/lite/power/portal/release603). Find also the config.external and config.normal configs to look my board setup. My wish is to see this patch in the main-line as soon possible because is very important for my job. Any feedback is welcome. best regards, luigi -- -- Luigi 'Comio' Mantellini RD - Software Industrie Dial Face S.p.A. Via Canzo, 4 20068 Peschiera Borromeo (MI), Italy Tel.: +39 02 5167 2813 Fax: +39 02 5167 2459 Email: luigi.mantell...@idf-hit.com ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org https://lists.openwrt.org/mailman/listinfo/openwrt-devel
Re: [OpenWrt-Devel] External cross toolchain
Hi Nico, I would like to know the actual status on External cross toolchain-patch. This is an hot topic for me because I use OpenWRT as main building system in my environment. Can you post the last patches? I will integrate and test. thanks again for your effort, luigi ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org https://lists.openwrt.org/mailman/listinfo/openwrt-devel
[OpenWrt-Devel] quick image build
Hi List, How you already know, I use OpenWRT as building environment for my daily activities. Usually I work on a single package for a lot of days and I need to create quickly a bootable build. My wish is to have a shortcut that avoid all time-consuming targets into the makefiles in order to produce my build for the rapid tests. My typical day: cd build_dir/my_arch/my_package vi file1 vi file2 vi file3 ... cd ../../.. make V=99 cp bin/image.img /tftpboot I would like to have a mechanism (using a global variable, ie QUICK=1) in order to avoid all compilation targets and just entrusting the my_package-{compile,install} to the developer. My hypothetical future day: cd build_dir/my_arch/my_package vi file1 vi file2 vi file3 ... cd ../../.. make package/my_package-compile make package/my_package-install make V=99 QUICK=1 # fast! cp bin/image.img /tftpboot The idea is that when you develop on just a package, the others are unmodified and all build/install activities can be safety skipped. QUICK=1 should just enable the final filesystem image generation. At this time I spend ~10min for each build... and this is not good. I'm working on a patch for reduce the time, but I don't perfectly know the complete targets dependencies and the actual patch version gain just 2 minutes. Suggestion are welcome and I'm interested to have your opinions. Best regards, luigi -- Luigi 'Comio' Mantellini RD - Software Industrie Dial Face S.p.A. Via Canzo, 4 20068 Peschiera Borromeo (MI), Italy Tel.: +39 02 5167 2813 Fax: +39 02 5167 2459 Email: luigi.mantell...@idf-hit.com diff --git a/include/image.mk b/include/image.mk index e70143d..eaacd3b 100644 --- a/include/image.mk +++ b/include/image.mk @@ -136,11 +136,15 @@ define BuildImage download: prepare: ifneq ($(IB),1) +ifeq ($(QUICK),) compile: compile-targets FORCE $(call Build/Compile) else compile: endif +else + compile: +endif ifneq ($(IB),1) install: compile install-targets FORCE diff --git a/include/kernel-build.mk b/include/kernel-build.mk index 4774a69..ead797e 100644 --- a/include/kernel-build.mk +++ b/include/kernel-build.mk @@ -26,9 +26,11 @@ define Kernel/Configure $(call Kernel/Configure/Default) endef +ifeq ($(QUICK),) define Kernel/CompileModules $(call Kernel/CompileModules/Default) endef +endif define Kernel/CompileImage $(call Kernel/CompileImage/Default) @@ -54,19 +56,34 @@ define BuildKernel $(if $(QUILT),$(Build/Quilt)) $(if $(LINUX_SITE),$(call Download,kernel)) +ifeq ($(QUICK),) $(STAMP_PREPARED): $(DL_DIR)/$(LINUX_SOURCE) -rm -rf $(KERNEL_BUILD_DIR) -mkdir -p $(KERNEL_BUILD_DIR) $(Kernel/Prepare) touch $$@ +else + $(STAMP_PREPARED): $(DL_DIR)/$(LINUX_SOURCE) + touch $$@ +endif +ifeq ($(QUICK),) $(STAMP_CONFIGURED): $(STAMP_PREPARED) $(LINUX_CONFIG) $(GENERIC_LINUX_CONFIG) $(TOPDIR)/.config $(Kernel/Configure) touch $$@ +else + $(STAMP_CONFIGURED): $(STAMP_PREPARED) $(LINUX_CONFIG) $(GENERIC_LINUX_CONFIG) $(TOPDIR)/.config + touch $$@ +endif +ifeq ($(QUICK),) $(LINUX_DIR)/.modules: $(STAMP_CONFIGURED) $(LINUX_DIR)/.config FORCE $(Kernel/CompileModules) touch $$@ +else + $(LINUX_DIR)/.modules: $(STAMP_CONFIGURED) $(LINUX_DIR)/.config FORCE + touch $$@ +endif $(LINUX_DIR)/.image: $(STAMP_CONFIGURED) FORCE $(Kernel/CompileImage) diff --git a/include/package-ipkg.mk b/include/package-ipkg.mk index 8d2ff83..f9359ea 100644 --- a/include/package-ipkg.mk +++ b/include/package-ipkg.mk @@ -82,7 +82,7 @@ ifeq ($(DUMP),) $($(1)_COMMANDS) \ ) -$$(IPKG_$(1)): $(STAGING_DIR)/etc/ipkg.conf $(PKG_BUILD_DIR)/.built $$(IDIR_$(1))/CONTROL/control +$$(IPKG_$(1)): $(STAGING_DIR)/etc/ipkg.conf $(PKG_BUILD_DIR)/.built $(if $(QUICK),,$$(IDIR_$(1))/CONTROL/control) $(call Package/$(1)/install,$$(IDIR_$(1))) mkdir -p $(PACKAGE_DIR) -find $$(IDIR_$(1)) -name 'CVS' -o -name '.svn' -o -name '.#*' | $(XARGS) rm -rf diff --git a/include/package.mk b/include/package.mk index dd81767..237e1ba 100644 --- a/include/package.mk +++ b/include/package.mk @@ -102,6 +102,7 @@ define Build/DefaultTargets $(call Build/Autoclean) $(STAMP_PREPARED) : export PATH=$$(TARGET_PATH_PKG) +ifeq ($(QUICK),) $(STAMP_PREPARED): $(if $(wildcard $(STAMP_KEEP)), echo Preserve $(PKG_NAME) sources... @@ -113,15 +114,25 @@ define Build/DefaultTargets $(foreach hook,$(Hooks/Prepare/Post),$(call $(hook))$(sep)) ) touch $$@ +else + $(STAMP_PREPARED): + touch $$@ +endif $(call Build/Exports,$(STAMP_CONFIGURED)) +ifeq ($(QUICK),) $(STAMP_CONFIGURED): $(STAMP_PREPARED) $(foreach hook,$(Hooks/Configure/Pre),$(call $(hook))$(sep)) $(Build/Configure) $(foreach hook,$(Hooks/Configure/Post),$(call $(hook))$(sep)) touch $$@ +else + $(STAMP_CONFIGURED): $(STAMP_PREPARED) + touch $$@ +endif $(call Build/Exports,$(STAMP_BUILT)) +ifeq ($(QUICK),) $(STAMP_BUILT): $(STAMP_CONFIGURED) $(foreach hook,$(Hooks/Compile/Pre),$(call $(hook))$(sep
Re: [OpenWrt-Devel] quick image build
Ops... target/install works fine! :D My apologies for the question! I haven't noticed this target! best regards, luigi -- Luigi 'Comio' Mantellini RD - Software Industrie Dial Face S.p.A. Via Canzo, 4 20068 Peschiera Borromeo (MI), Italy Tel.: +39 02 5167 2813 Fax: +39 02 5167 2459 Email: luigi.mantell...@idf-hit.com ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org https://lists.openwrt.org/mailman/listinfo/openwrt-devel
Re: [OpenWrt-Devel] External cross toolchain
Hi Nico (Again! :D) following the previous mail I think that also TOOLCHAIN_DIR directroy should be fixed to TOOLCHAIN_ROOT value. ciao luigi Il lunedì 02 febbraio 2009 18:18:15 Luigi 'Comio' Mantellini ha scritto: Hi Nico, I'm using your patch. I think that the user should give the following values: CONFIG_TOOLCHAIN_ROOT, the executalbes searching path CONFIG_TOOLCHAIN_COMPILER_PATH and the (target)libraries searching path CONFIG_TOOLCHAIN_LIBRARY_PATH The rules.mk should set COMPILER_PATH to CONFIG_TOOLCHAIN_COMPILER_PATH and the LIBRARY_PATH to CONFIG_TOOLCHAIN_LIBRARY_PATH in order to have compile. Of course all searching path should be fixed accordly. Other thing is that base-files should determinate the libraries to copy on root fs using the $(CROSS_COMPILE)ldd tool I hope to give a contribute to have a stable external toolchain support. luigi Il venerdì 30 gennaio 2009 09:09:32 Luigi 'Comio' Mantellini ha scritto: Hi Nico, thanks again for your patch. On my system works fine. best regards, luigi On Friday 30 January 2009 00:33:49 Nico wrote: Any objections committing the patch to trunk ? -- Nico Nico wrote: Luigi 'Comio' Mantellini wrote: Thanks Nico, I will integrate asap on my experimental tree. I noticed a mistake row 262: + ifndef ($(CONFIG_NATIVE_TOOLCHAIN),) should be: + ifneq ($(CONFIG_NATIVE_TOOLCHAIN),) You're right, there was a couple of mistakes in my previous patch. I updated it [1]. Anyway, it should be great if a toolchain prepared using OpenWRT can be exported to others developers. Have you tried building the SDK? It should embed the toolchain (and the set of files required to build packages). 1. http://nico-and-co.net/void/patches/openwrt-r14216-external_toolchain _v 3. diff -- Nico On Monday 26 January 2009 18:36:32 Nico wrote: Hi Luigi, I have a patch [1] for external toolchains support, it will allow you to select your target name (powerpc-unknow-linux-gnu), toolchain commands prefix, toolchain binaries libraries path (for toolchains installed in unusual locations and needing custom PATH LD_LIBRARY_PATH). Let me know if it works for you :) 1. http://nico-and-co.net/void/patches/openwrt-r14204-external_toolcha in .p atch -- Nico Luigi 'Comio' Mantellini wrote: Hi ML, I need your help to understand the actual status of external toolchain support. In my environment, there are a lot of developers that works on an building environment openwrt-based. Openwrt is a good choice to manage my builds and : I am happy to have chosen one year ago. The big problem that I observed is that OpenWRT doesn't support in a clean-way the external toolchains. the NATIVE_TOOLCHAIN option seems to not resolve my problems because it disables CROSS_COMPILE defines and I don't like the openwrt behaviour when this option is enabled. I want understand if are there anybody that shares my observation (or I'm not able to use native_toolchain in best way) and if a work on this topic can be appreciated by community. Kindly, give me any suggestion or opinion in order to understand what is needed to do and what I can do. thanks in advance, luigi ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org http://lists.openwrt.org/cgi-bin/mailman/listinfo/openwrt-devel ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org http://lists.openwrt.org/cgi-bin/mailman/listinfo/openwrt-devel ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org http://lists.openwrt.org/cgi-bin/mailman/listinfo/openwrt-devel -- Luigi Mantellini RD - Software Industrie Dial Face S.p.A. Via Canzo, 4 20068 Peschiera Borromeo (MI), Italy Tel.: +39 02 5167 2813 Fax: +39 02 5167 2459 Email: luigi.mantell...@idf-hit.com ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org http://lists.openwrt.org/cgi-bin/mailman/listinfo/openwrt-devel
Re: [OpenWrt-Devel] External cross toolchain
Hi Nico, thanks again for your patch. On my system works fine. best regards, luigi On Friday 30 January 2009 00:33:49 Nico wrote: Any objections committing the patch to trunk ? -- Nico Nico wrote: Luigi 'Comio' Mantellini wrote: Thanks Nico, I will integrate asap on my experimental tree. I noticed a mistake row 262: + ifndef ($(CONFIG_NATIVE_TOOLCHAIN),) should be: + ifneq ($(CONFIG_NATIVE_TOOLCHAIN),) You're right, there was a couple of mistakes in my previous patch. I updated it [1]. Anyway, it should be great if a toolchain prepared using OpenWRT can be exported to others developers. Have you tried building the SDK? It should embed the toolchain (and the set of files required to build packages). 1. http://nico-and-co.net/void/patches/openwrt-r14216-external_toolchain_v3. diff -- Nico On Monday 26 January 2009 18:36:32 Nico wrote: Hi Luigi, I have a patch [1] for external toolchains support, it will allow you to select your target name (powerpc-unknow-linux-gnu), toolchain commands prefix, toolchain binaries libraries path (for toolchains installed in unusual locations and needing custom PATH LD_LIBRARY_PATH). Let me know if it works for you :) 1. http://nico-and-co.net/void/patches/openwrt-r14204-external_toolchain.p atch -- Nico Luigi 'Comio' Mantellini wrote: Hi ML, I need your help to understand the actual status of external toolchain support. In my environment, there are a lot of developers that works on an building environment openwrt-based. Openwrt is a good choice to manage my builds and : I am happy to have chosen one year ago. The big problem that I observed is that OpenWRT doesn't support in a clean-way the external toolchains. the NATIVE_TOOLCHAIN option seems to not resolve my problems because it disables CROSS_COMPILE defines and I don't like the openwrt behaviour when this option is enabled. I want understand if are there anybody that shares my observation (or I'm not able to use native_toolchain in best way) and if a work on this topic can be appreciated by community. Kindly, give me any suggestion or opinion in order to understand what is needed to do and what I can do. thanks in advance, luigi ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org http://lists.openwrt.org/cgi-bin/mailman/listinfo/openwrt-devel ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org http://lists.openwrt.org/cgi-bin/mailman/listinfo/openwrt-devel ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org http://lists.openwrt.org/cgi-bin/mailman/listinfo/openwrt-devel -- Luigi Mantellini RD - Software Industrie Dial Face S.p.A. Via Canzo, 4 20068 Peschiera Borromeo (MI), Italy Tel.: +39 02 5167 2813 Fax: +39 02 5167 2459 Email: luigi.mantell...@idf-hit.com ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org http://lists.openwrt.org/cgi-bin/mailman/listinfo/openwrt-devel
[OpenWrt-Devel] External cross toolchain
Hi ML, I need your help to understand the actual status of external toolchain support. In my environment, there are a lot of developers that works on an building environment openwrt-based. Openwrt is a good choice to manage my builds and : I am happy to have chosen one year ago. The big problem that I observed is that OpenWRT doesn't support in a clean-way the external toolchains. the NATIVE_TOOLCHAIN option seems to not resolve my problems because it disables CROSS_COMPILE defines and I don't like the openwrt behaviour when this option is enabled. I want understand if are there anybody that shares my observation (or I'm not able to use native_toolchain in best way) and if a work on this topic can be appreciated by community. Kindly, give me any suggestion or opinion in order to understand what is needed to do and what I can do. thanks in advance, luigi -- Luigi Mantellini RD - Software Industrie Dial Face S.p.A. Via Canzo, 4 20068 Peschiera Borromeo (MI), Italy Tel.: +39 02 5167 2813 Fax: +39 02 5167 2459 Email: luigi.mantell...@idf-hit.com ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org http://lists.openwrt.org/cgi-bin/mailman/listinfo/openwrt-devel
Re: [OpenWrt-Devel] [OpenWrt-Tickets] [OpenWrt] #3479: Support for gcc4.3
thanks ciao luigi On Wednesday 21 January 2009 12:59:35 OpenWrt wrote: #3479: Support for gcc4.3 --+ - Reporter: luigi.mantell...@idf-hit.com |Owner: nico Type: enhancement | Status: closed Priority: normal|Milestone: Kamikaze 8.09 Component: packages | Resolution: fixed Keywords:| --+ - Changes (by nico): * status: assigned = closed * resolution: = fixed Comment: Added in [14131], thanks! -- Luigi Mantellini RD - Software Industrie Dial Face S.p.A. Via Canzo, 4 20068 Peschiera Borromeo (MI), Italy Tel.: +39 02 5167 2813 Fax: +39 02 5167 2459 Email: luigi.mantell...@idf-hit.com ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org http://lists.openwrt.org/cgi-bin/mailman/listinfo/openwrt-devel
Re: [OpenWrt-Devel] [OpenWrt-Tickets] [OpenWrt] #4271: please update uclibc to 0.9.30
thanks luigi On Wednesday 21 January 2009 12:46:43 OpenWrt wrote: #4271: please update uclibc to 0.9.30 -+- - Reporter: anonymous|Owner: nico Type: enhancement | Status: closed Priority: highest |Milestone: Component: base system | Resolution: fixed Keywords: | -+- - Changes (by nico): * status: assigned = closed * resolution: = fixed Comment: Support for uClibc v0.9.30 added in [14130], thanks! -- Luigi Mantellini RD - Software Industrie Dial Face S.p.A. Via Canzo, 4 20068 Peschiera Borromeo (MI), Italy Tel.: +39 02 5167 2813 Fax: +39 02 5167 2459 Email: luigi.mantell...@idf-hit.com ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org http://lists.openwrt.org/cgi-bin/mailman/listinfo/openwrt-devel
Re: [OpenWrt-Devel] [rfc] Small initramfs environment
any comments? thanks luigi On Tuesday 20 January 2009 23:02:41 Luigi Mantellini wrote: Hi developers, Hi users, I made the homework. The idea is to permit to have custom initramfs root and (at same time) the normal rootfs (jffs2, squash, ...) to give better flexibility at boot time for the applications (like mine...) that requires that. This patchset adds: CONFIG_TARGET_CUSTOM_INITRAMFS option to enable/disable the support for a custom initramfs two (example) packets: base-initramfs-files busybox-mini that install files into the custom initramfs. In general the macros present into package-ipkg.mk file have a _initramfs suffixed version that follow the action on the initramfs root directory. Into the target build_dir is also created a root-initramfs-$(TARGET) with the root structure that will be linked with the kernel. Feel free to give me any comments on this patchset. ciao luigi 2009/1/19 Luigi 'Comio' Mantellini luigi.mantellini...@gmail.com: Hi All, I have another absurd idea and I would like to share it with community. Problem to solve: My application needs to create a complex running environment after the boot. At boot time I have just a file with the file system that need to be analyzed to be mounted (I cannot mount it directly). Solution The idea is to have a micro initramfs with a (small) busybox and some scripts in order to perform the needed operation and create the final root file system. I have two choice: create a Image Makefile under my target that creates a minimal static busybox or give into the busybox' Config menu n option that permit to compile the busybox (in static way) avoiding to install into the main squashfs. The image Makefile will create an opportune initramfs file list that includes the created busybox and a init script that will link during the boot sequence the busybox into the final filesystem. I think that a Package/*/InstallInitramfs defne should be invoked to permit to the packages the installation of initramfs files. These are my confused 2EuroCents. ciao luigi -- Luigi Mantellini RD - Software Industrie Dial Face S.p.A. Via Canzo, 4 20068 Peschiera Borromeo (MI), Italy Tel.: +39 02 5167 2813 Fax: +39 02 5167 2459 Email: luigi.mantell...@idf-hit.com -- Luigi Mantellini RD - Software Industrie Dial Face S.p.A. Via Canzo, 4 20068 Peschiera Borromeo (MI), Italy Tel.: +39 02 5167 2813 Fax: +39 02 5167 2459 Email: luigi.mantell...@idf-hit.com ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org http://lists.openwrt.org/cgi-bin/mailman/listinfo/openwrt-devel
[OpenWrt-Devel] RFC: improve git/svn support
Hi Boys, I'm still evaluating OpenWRT as The Ultimate Building Environment. My environment (the arch is not important) is made by a lot of git and svn repositories where my colleagues work every day. [ git1 ] Kernel [ git2 ] Application #1 [ git3 ] Application #2 [ svn1] Application #3 [ git4 ] OpenWRT Each application has a development life with tags, refs, revisions, bugs, resolved bus, etc... When My costumer ask me for a new release (for example #XX) I need to freeze each git tree for each application and create a single build with all contributes from the git/svn trees are well identified This is my real-life scenario. Reading the include/* files I noticed that the svn/git/ support is still immature. Immature because: - with VERSION is also referenced the REFS on the CVS - with SUBDIR and the FILE are too dependent - every download (for updates) is traduced to rm -rf THEDIR and a Clone/checkout... absolutely not efficient. - lack of control on the specific cvs features. - Kernel is considered a special-case... no cvs support (using small fixes we cover this problem) I want to understand which are the plans on these point of improvement. I think that using a small effort the cvs support can be improved, becoming usable. my apologies for my opinions... of course, I would like to hear different opinions. Thanks and best regards, ciao luigi -- Luigi Mantellini RD - Software Industrie Dial Face S.p.A. Via Canzo, 4 20068 Peschiera Borromeo (MI), Italy Tel.: +39 02 5167 2813 Fax: +39 02 5167 2459 Email: [EMAIL PROTECTED] ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org http://lists.openwrt.org/cgi-bin/mailman/listinfo/openwrt-devel
[OpenWrt-Devel] Generic patches... how disable them?
Hi List, I'm evaluating the switch to the new openwrt (will be 8.08?) from an heavy customized kamikaze 7.09 (I sent in the past my patches) in order to support my bsp based on a very old linux kernel 2.6.15. Unfortunately I cannot change the kernel. My problem is that openwrt applies always the patches into the generic-2.6/patches ... this behaviour is not good for my application and I need to avoid the to apply these patches. Of course, i can modify the include/kernel.mk quilt.mk to modify the default behavior (how I did in my k7.09) but I would like a formal method to disable the default patches application. How can I achieve my goal? I searched a method without success and I (now) modifed the GENERIC_PATCH_DIR:=.. assignation to GENERIC_PATCH_DIR?=. Thanks in advance and best regards, luigi -- Luigi Mantellini RD - Software Industrie Dial Face S.p.A. Via Canzo, 4 20068 Peschiera Borromeo (MI), Italy Tel.: +39 02 5167 2813 Fax: +39 02 5167 2459 Email: [EMAIL PROTECTED] ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org http://lists.openwrt.org/cgi-bin/mailman/listinfo/openwrt-devel
[OpenWrt-Devel] [rfc] Kernel sources download strategy
Hi All, I know that what I say usually is ignored :D But I like to flood this ml with stupid(?) ideas. The actual OpenWRT considers kernels 2.6.X.Y and 2.6.X.(Y+1) as a very different kernels downloading the full sources for both. I think that this is not a good approach because this is time and band expensive. An idea should be to download the full sources 2.6.X kernels and just the patches 2.6.X.1, 2.6.X.2, ... 2.6.X.Y to cover the kernels with version like 2.6.X.Y. The result will be a gain in terms of band and time utilization. I think that to achieve this the download perl and make scripts should be made more smart than actual versions These optimizations should be applied also to gcc and uclibc and any other package that provides diff between versions. Thanks a lot for your comments. luigi http://repo.or.cz/w/openwrt_comio.git -- __ Luigi Mantellini .'__'. RD - Software (.' '.)Industrie Dial Face S.p.A. ( :==: )Via Canzo, 4 ('.__.')20068 Peschiera Borromeo (MI), Italy '.__.' Tel.: +39 02 5167 2813 Fax: +39 02 5167 2459 Ind. Dial Face Email: [EMAIL PROTECTED] www.idf-hit.com GPG fingerprint: 3DD1 7B71 FBDF 6376 1B4A B003 175F E979 907E 1650 ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org http://lists.openwrt.org/cgi-bin/mailman/listinfo/openwrt-devel
Re: [OpenWrt-Devel] [rfc] Kernel sources download strategy
On gio, 2008-07-03 at 14:10 +0200, Jan Willies wrote: ... These optimizations should be applied also to gcc and uclibc and any other package that provides diff between versions. Thanks a lot for your comments. -ENOPATCH ? I don't understand. What do you mean? He thinks it's a good idea and likes to see a patch :) Oh. I usually ask opinion to the ml... to understand if a feature is interesting or simply useless... I already proposed patches for other things... and I'm not a chap that ask to the others to do something :) (http://repo.or.cz/w/openwrt_comio.git) From this... I hope to have time to create a proof-of-concept patch to cover my rfc. Anyway... what do you think about this feature? is it stupid? already implemented?... thanks luigi - Jan ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org http://lists.openwrt.org/cgi-bin/mailman/listinfo/openwrt-devel -- __ Luigi Mantellini .'__'. RD - Software (.' '.)Industrie Dial Face S.p.A. ( :==: )Via Canzo, 4 ('.__.')20068 Peschiera Borromeo (MI), Italy '.__.' Tel.: +39 02 5167 2813 Fax: +39 02 5167 2459 Ind. Dial Face Email: [EMAIL PROTECTED] www.idf-hit.com GPG fingerprint: 3DD1 7B71 FBDF 6376 1B4A B003 175F E979 907E 1650 ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org http://lists.openwrt.org/cgi-bin/mailman/listinfo/openwrt-devel
Re: [OpenWrt-Devel] [rfc] Kernel sources download strategy
On gio, 2008-07-03 at 13:14 +0100, j t wrote: On Thu, Jul 3, 2008 at 1:07 PM, Luigi 'Comio' Mantellini [EMAIL PROTECTED] wrote: On gio, 2008-07-03 at 12:59 +0200, Michael Buesch wrote: -ENOPATCH ? I don't understand. What do you mean? See: http://www.opengroup.org/onlinepubs/95399/basedefs/errno.h.html HTH this is not a posix/susv3 error :P luigi ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org http://lists.openwrt.org/cgi-bin/mailman/listinfo/openwrt-devel -- __ Luigi Mantellini .'__'. RD - Software (.' '.)Industrie Dial Face S.p.A. ( :==: )Via Canzo, 4 ('.__.')20068 Peschiera Borromeo (MI), Italy '.__.' Tel.: +39 02 5167 2813 Fax: +39 02 5167 2459 Ind. Dial Face Email: [EMAIL PROTECTED] www.idf-hit.com GPG fingerprint: 3DD1 7B71 FBDF 6376 1B4A B003 175F E979 907E 1650 ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org http://lists.openwrt.org/cgi-bin/mailman/listinfo/openwrt-devel
Re: [OpenWrt-Devel] [rfc] Kernel sources download strategy
Detto... fatto! This is my proposed patch (just a proof of concept): http://repo.or.cz/w/openwrt_comio.git?a=commitdiff;h=2cbc4b18271903b77896e1d1db279c3a1675b904 The md5 values should be recalculated for kernels 2.6.x and for the patches 2.6.x.y (Now I just set MD5 to empty string). ciao luigi On gio, 2008-07-03 at 13:36 +0100, j t wrote: On Thu, Jul 3, 2008 at 1:25 PM, Luigi 'Comio' Mantellini [EMAIL PROTECTED] wrote: On gio, 2008-07-03 at 13:14 +0100, j t wrote: On Thu, Jul 3, 2008 at 1:07 PM, Luigi 'Comio' Mantellini [EMAIL PROTECTED] wrote: On gio, 2008-07-03 at 12:59 +0200, Michael Buesch wrote: -ENOPATCH ? I don't understand. What do you mean? See: http://www.opengroup.org/onlinepubs/95399/basedefs/errno.h.html HTH this is not a posix/susv3 error :P You are correct. That it why it is called humour (I am British)/humor. HTH ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org http://lists.openwrt.org/cgi-bin/mailman/listinfo/openwrt-devel -- __ Luigi Mantellini .'__'. RD - Software (.' '.)Industrie Dial Face S.p.A. ( :==: )Via Canzo, 4 ('.__.')20068 Peschiera Borromeo (MI), Italy '.__.' Tel.: +39 02 5167 2813 Fax: +39 02 5167 2459 Ind. Dial Face Email: [EMAIL PROTECTED] www.idf-hit.com GPG fingerprint: 3DD1 7B71 FBDF 6376 1B4A B003 175F E979 907E 1650 ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org http://lists.openwrt.org/cgi-bin/mailman/listinfo/openwrt-devel
Re: [OpenWrt-Devel] [rfc] Kernel sources download strategy
On gio, 2008-07-03 at 21:18 +0100, j t wrote: On Thu, Jul 3, 2008 at 8:00 PM, Claudio [EMAIL PROTECTED] wrote: Hi 2008/7/3 Luigi 'Comio' Mantellini [EMAIL PROTECTED]: On gio, 2008-07-03 at 13:14 +0100, j t wrote: On Thu, Jul 3, 2008 at 1:07 PM, Luigi 'Comio' Mantellini [EMAIL PROTECTED] wrote: On gio, 2008-07-03 at 12:59 +0200, Michael Buesch wrote: -ENOPATCH ? I don't understand. What do you mean? See: http://www.opengroup.org/onlinepubs/95399/basedefs/errno.h.html HTH this is not a posix/susv3 error :P eheheh beware Luigi they can suggest to patch it also :P Now _I'm_ the one who's confused. Is this also hum(o)ur? The humor is a relative concept. anyway... a patch should be sent to the unix developers... luigi openwrt-devel mailing list openwrt-devel@lists.openwrt.org http://lists.openwrt.org/cgi-bin/mailman/listinfo/openwrt-devel -- __ Luigi Mantellini .'__'. RD - Software (.' '.)Industrie Dial Face S.p.A. ( :==: )Via Canzo, 4 ('.__.')20068 Peschiera Borromeo (MI), Italy '.__.' Tel.: +39 02 5167 2813 Fax: +39 02 5167 2459 Ind. Dial Face Email: [EMAIL PROTECTED] www.idf-hit.com GPG fingerprint: 3DD1 7B71 FBDF 6376 1B4A B003 175F E979 907E 1650 ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org http://lists.openwrt.org/cgi-bin/mailman/listinfo/openwrt-devel
Re: [OpenWrt-Devel] git server down?
The server seems to no have a dns entry... If you need, I have a mirror of the openwrt trunk: http://repo.or.cz/w/openwrt_comio.git I'm not a openwrt developer... I just use this tree to make some experiment useful for my activity. luigi On ven, 2008-06-27 at 13:35 -0700, Geoff Levand wrote: Has the git server been moved? nslookup gives me this: Host nbd.name not found: 3(NXDOMAIN) -Geoff ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org http://lists.openwrt.org/cgi-bin/mailman/listinfo/openwrt-devel -- __ Luigi Mantellini .'__'. RD - Software (.' '.)Industrie Dial Face S.p.A. ( :==: )Via Canzo, 4 ('.__.')20068 Peschiera Borromeo (MI), Italy '.__.' Tel.: +39 02 5167 2813 Fax: +39 02 5167 2459 Ind. Dial Face Email: [EMAIL PROTECTED] www.idf-hit.com GPG fingerprint: 3DD1 7B71 FBDF 6376 1B4A B003 175F E979 907E 1650 ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org http://lists.openwrt.org/cgi-bin/mailman/listinfo/openwrt-devel
[OpenWrt-Devel] [PATCH 1/1] Initial support for multi-libc. - Proof of Concept -
--- include/host-build.mk|1 + package/base-files/Makefile |2 +- package/kexec-tools/Makefile |2 +- rules.mk |4 ++-- toolchain/Config.in | 13 + toolchain/Config.version | 15 +++ toolchain/Makefile |4 toolchain/binutils/Makefile |2 +- toolchain/gcc/Makefile |8 ++-- toolchain/uClibc/Config.in |4 ++-- toolchain/uClibc/Makefile|2 +- 11 files changed, 47 insertions(+), 10 deletions(-) create mode 100644 toolchain/Config.version diff --git a/include/host-build.mk b/include/host-build.mk index 8a8d6de..7b3356d 100644 --- a/include/host-build.mk +++ b/include/host-build.mk @@ -82,6 +82,7 @@ define Download/default FILE:=$(PKG_SOURCE) URL:=$(PKG_SOURCE_URL) PROTO:=$(PKG_SOURCE_PROTO) + SUBDIR:=$(PKG_SOURCE_SUBDIR) VERSION:=$(PKG_SOURCE_VERSION) MD5SUM:=$(PKG_MD5SUM) endef diff --git a/package/base-files/Makefile b/package/base-files/Makefile index bec33c8..17422c4 100644 --- a/package/base-files/Makefile +++ b/package/base-files/Makefile @@ -88,7 +88,7 @@ endef define Package/uclibc/Default SECTION:=libs CATEGORY:=Base system - DEPENDS:[EMAIL PROTECTED] + DEPENDS:[EMAIL PROTECTED]@USE_UCLIBC URL:=http://uclibc.org/ VERSION:=$(UCLIBC_VERSION)$(UCLIBC_PATCHVER)-$(PKG_RELEASE) endef diff --git a/package/kexec-tools/Makefile b/package/kexec-tools/Makefile index 951d41a..34def60 100644 --- a/package/kexec-tools/Makefile +++ b/package/kexec-tools/Makefile @@ -37,7 +37,7 @@ define Package/kexec-tools/config endef CONFIGURE_ARGS = \ - --target=$(CONFIG_KEXEC_TOOLS_TARGET_NAME)-linux-uclibc \ + --target=$(CONFIG_KEXEC_TOOLS_TARGET_NAME)-linux-$(CONFIG_TARGET_SUFFIX) \ --host=$(GNU_TARGET_NAME)-uclibc \ --build=$(GNU_HOST_NAME) \ --program-prefix= \ diff --git a/rules.mk b/rules.mk index 7153a04..a86d8f5 100644 --- a/rules.mk +++ b/rules.mk @@ -57,9 +57,9 @@ IPKG_STATE_DIR:=$(TARGET_DIR)/usr/lib/ipkg ifeq ($(CONFIG_NATIVE_TOOLCHAIN),) -include $(TOOLCHAIN_DIR)/info.mk - REAL_GNU_TARGET_NAME=$(OPTIMIZE_FOR_CPU)-linux-uclibc$(if $(CONFIG_EABI_SUPPORT),gnueabi) + REAL_GNU_TARGET_NAME=$(OPTIMIZE_FOR_CPU)-linux-$(CONFIG_TARGET_SUFFIX)$(if $(CONFIG_EABI_SUPPORT),gnueabi) GNU_TARGET_NAME=$(OPTIMIZE_FOR_CPU)-linux - TARGET_CROSS:=$(if $(TARGET_CROSS),$(TARGET_CROSS),$(OPTIMIZE_FOR_CPU)-linux-uclibc$(if $(CONFIG_EABI_SUPPORT),gnueabi)-) + TARGET_CROSS:=$(if $(TARGET_CROSS),$(TARGET_CROSS),$(OPTIMIZE_FOR_CPU)-linux-$(CONFIG_TARGET_SUFFIX)$(if $(CONFIG_EABI_SUPPORT),gnueabi)-) endif TARGET_PATH:=$(TOOLCHAIN_DIR)/bin:$(STAGING_DIR_HOST)/bin:$(STAGING_DIR)/host/bin:$(PATH) diff --git a/toolchain/Config.in b/toolchain/Config.in index 1a465f5..6ecddf4 100644 --- a/toolchain/Config.in +++ b/toolchain/Config.in @@ -41,6 +41,18 @@ menuconfig EXTRA_TARGET_ARCH source toolchain/binutils/Config.in source toolchain/gcc/Config.in + +choice + prompt LIBC implementation if TOOLCHAINOPTS + default USE_UCLIBC + help + Select the LIBC implementation between uClibc and... uClibc. + +config USE_UCLIBC + bool Use uClibC +endchoice + + source toolchain/uClibc/Config.in config GDB @@ -92,3 +104,4 @@ config TARGET_OPTIMIZATION source toolchain/gcc/Config.version source toolchain/uClibc/Config.version +source toolchain/Config.version diff --git a/toolchain/Config.version b/toolchain/Config.version new file mode 100644 index 000..19d3eab --- /dev/null +++ b/toolchain/Config.version @@ -0,0 +1,15 @@ +config USE_UCLIBC +bool +default y if !TOOLCHAINOPTS + +config TARGET_SUFFIX +string +default uclibcif USE_UCLIBC +default gnu if USE_EGLIBC || USE_GLIBC +default uclibc + +config LIBC_IMPLEMENTATION + string + default UCLIBCif USE_UCLIBC + default EGLIBCif USE_EGLIBC + default GLIBC if USE_GLIBC diff --git a/toolchain/Makefile b/toolchain/Makefile index 4afcc43..8d0f91b 100644 --- a/toolchain/Makefile +++ b/toolchain/Makefile @@ -9,11 +9,14 @@ curdir:=toolchain # subdirectories to descend into +ifeq ($(CONFIG_USE_UCLIBC),y) $(curdir)/builddirs := kernel-headers $(if $(CONFIG_GDB),gdb) $(if $(CONFIG_NATIVE_TOOLCHAIN),,binutils gcc uClibc) +endif $(curdir)/builddirs-compile:=. $(filter-out kernel-headers,$($(curdir)/builddirs)) $(curdir)/builddirs-install:=$($(curdir)/builddirs-compile) # builddir dependencies +ifeq ($(CONFIG_USE_UCLIBC),y) $(curdir)/uClibc/prepare:=$(curdir)/kernel-headers/prepare ifeq ($(CONFIG_NATIVE_TOOLCHAIN),) $(curdir)/binutils/prepare:=$(curdir)/uClibc/prepare @@ -22,6 +25,7 @@ ifeq ($(CONFIG_NATIVE_TOOLCHAIN),) $(curdir)/gcc/install:=$(curdir)/uClibc/compile $(curdir)/uClibc/install:=$(curdir)/gcc/install endif +endif ifneq ($(ARCH),)
[OpenWrt-Devel] Kamikaze release 7.09 - which svn revision?
Hi guys, I'm preparing my environment to switch to the new Openwrt release. To minimize the impact on my system I need to know which is the exact SVN revision used to deploy the Kamikaze 7.09 released in 30 Sept. 2007 (http://downloads.openwrt.org/kamikaze/7.09/kamikaze_7.09.tar.bz2) Thanks in advance. luigi -- __ Luigi Mantellini .'__'. RD - Software (.' '.)Industrie Dial Face S.p.A. ( :==: )Via Canzo, 4 ('.__.')20068 Peschiera Borromeo (MI), Italy '.__.' Tel.: +39 02 5167 2813 Fax: +39 02 5167 2459 Ind. Dial Face Email: [EMAIL PROTECTED] www.idf-hit.com GPG fingerprint: 3DD1 7B71 FBDF 6376 1B4A B003 175F E979 907E 1650 ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org http://lists.openwrt.org/cgi-bin/mailman/listinfo/openwrt-devel
Re: [OpenWrt-Devel] Patch-set for Freescale Coldifre 547x/548x cpus.
Il giorno ven, 20/06/2008 alle 18.23 +0100, Peter Denison ha scritto: On Fri, 20 Jun 2008, Luigi 'Comio' Mantellini wrote: Great stuff - I might just get a ColdFire board if I can find one cheap enough... The best thing to do is to put up the patches somewhere publicly accessible, and then post the link to the list. I think it would be good to allow anyone to comment on them, although it's a big ask to go through 3MB of patches! 3MB of Freescale patches... of course! I will upload on my website asap luigi -- Peter ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org http://lists.openwrt.org/cgi-bin/mailman/listinfo/openwrt-devel -- Ing. Luigi Mantellini Industrie Dial Face S.p.A. Via Canzo, 4 20068 Peschiera Borromeo (MI) Tel.: +39 02 5167 2813 Fax: +39 02 5167 2459 E-mail: [EMAIL PROTECTED] ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org http://lists.openwrt.org/cgi-bin/mailman/listinfo/openwrt-devel
[OpenWrt-Devel] m68k OpenWRT
Hi All, how you should know, I'm working to port OpenWRT on m68k/coldfire arch. I compiled with success (and some patch) the toolchain... but I have this error during the kernel build: make[5]: Leaving directory `/mnt/devel/openwrt/OpenWRT.git/build_dir/linux-cf547x/linux-2.6.23' make[5]: Entering directory `/mnt/devel/openwrt/OpenWRT.git/build_dir/linux-cf547x/linux-2.6.23' CHK include/linux/version.h CHK include/linux/utsrelease.h CALLscripts/checksyscalls.sh CHK include/linux/compile.h GZIPkernel/config_data.gz IKCFG kernel/config_data.h CC kernel/configs.o LD kernel/built-in.o GEN .version CHK include/linux/compile.h UPD include/linux/compile.h CC init/version.o LD init/built-in.o LD .tmp_vmlinux1 kernel/built-in.o: In function `timespec_add_ns': /mnt/devel/openwrt/OpenWRT.git/build_dir/linux-cf547x/linux-2.6.23/include/linux/time.h:177: undefined reference to `__umoddi3' /mnt/devel/openwrt/OpenWRT.git/build_dir/linux-cf547x/linux-2.6.23/include/linux/time.h:177: undefined reference to `__udivdi3' umoddi3 and udivdi3 are helper functions that support the long long arithmetic. the libcc.a has these function: $ m68k-linux-gnu-objdump -t libgcc.a In archive libgcc.a: _udivdi3.o: file format elf32-m68k SYMBOL TABLE: ldf *ABS* libgcc2.c ld .text .text ld .data .data ld .bss .bss ld .debug_abbrev .debug_abbrev ld .debug_info .debug_info ld .debug_line .debug_line ld .debug_frame .debug_frame ld .eh_frame .eh_frame ld .debug_loc .debug_loc ld .debug_pubnames .debug_pubnames ld .debug_aranges .debug_aranges ld .debug_str .debug_str ld .note.GNU-stack .note.GNU-stack ld .comment .comment g F .text 00b6 .hidden __udivdi3 _umoddi3.o: file format elf32-m68k SYMBOL TABLE: ldf *ABS* libgcc2.c ld .text .text ld .data .data ld .bss .bss ld .debug_abbrev .debug_abbrev ld .debug_info .debug_info ld .debug_line .debug_line ld .debug_frame .debug_frame ld .eh_frame .eh_frame ld .debug_loc .debug_loc ld .debug_pubnames .debug_pubnames ld .debug_aranges .debug_aranges ld .debug_str .debug_str ld .note.GNU-stack .note.GNU-stack ld .comment .comment g F .text 00d6 .hidden __umoddi3 I found the keyword .hidden placed before the symbols __umoddi3 and __udivdi3 also on mipsel-compiled libgcc.a... Any idea? luigi -- __ Luigi Mantellini .'__'. RD - Software (.' '.)Industrie Dial Face S.p.A. ( :==: )Via Canzo, 4 ('.__.')20068 Peschiera Borromeo (MI), Italy '.__.' Tel.: +39 02 5167 2813 Fax: +39 02 5167 2459 Ind. Dial Face Email: [EMAIL PROTECTED] www.idf-hit.com GPG fingerprint: 3DD1 7B71 FBDF 6376 1B4A B003 175F E979 907E 1650 ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org http://lists.openwrt.org/cgi-bin/mailman/listinfo/openwrt-devel
Re: [OpenWrt-Devel] m68k OpenWRT
Hi Felix Hi All, I found this article: http://bbs.archlinux.org/viewtopic.php?id=45882 the patch seems to solve the issue (that is manifested only using gcc4.3). luigi On gio, 2008-06-19 at 16:21 +0200, Felix Fietkau wrote: Luigi 'Comio' Mantellini wrote: Hi All, how you should know, I'm working to port OpenWRT on m68k/coldfire arch. I compiled with success (and some patch) the toolchain... but I have this error during the kernel build: We had similar issues with mips in the past. If you look through the mips arch directory, you should find such functions there. You can probably take the missing pieces for m68k either by finding an arch that implements those in a generic way (i think mips might even do this), or you port the code from the gcc/libgcc source. Shouldn't be very hard to do. - Felix ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org http://lists.openwrt.org/cgi-bin/mailman/listinfo/openwrt-devel -- __ Luigi Mantellini .'__'. RD - Software (.' '.)Industrie Dial Face S.p.A. ( :==: )Via Canzo, 4 ('.__.')20068 Peschiera Borromeo (MI), Italy '.__.' Tel.: +39 02 5167 2813 Fax: +39 02 5167 2459 Ind. Dial Face Email: [EMAIL PROTECTED] www.idf-hit.com GPG fingerprint: 3DD1 7B71 FBDF 6376 1B4A B003 175F E979 907E 1650 ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org http://lists.openwrt.org/cgi-bin/mailman/listinfo/openwrt-devel
[OpenWrt-Devel] Patch-set for Freescale Coldifre 547x/548x cpus.
Hu guys, Finally I ported OpenWRT on coldfire/m68k with gcc4.3. I created a set of patches (~3MBytes) to add: - gcc4.3.1 support (need to compile coldfire code); - m68k fixes for gcc4.3 and uClibc; - kernel patch from Freescale for 547x/548x cpu and linux kernel 2.6.23; - coldfire MCF54x5yFE board target (with x in {7,8}, y in {A, B, C, D, E, F}). I just tested the compilation and I don't have tested the file directly image on the targets. (I DON'T KNOW IF THE OBJECT CODE IS CORRECT!). Anyway, I think that these patches should be a good start point in order to create a complete and powerful building environment for these new CPUs from Freescale... If anyone is interested contact me and I will happily send these patches. of course, I can post the patches on this forum... but the files are very huge... ciao ciao luigi -- __ Luigi Mantellini .'__'. RD - Software (.' '.)Industrie Dial Face S.p.A. ( :==: )Via Canzo, 4 ('.__.')20068 Peschiera Borromeo (MI), Italy '.__.' Tel.: +39 02 5167 2813 Fax: +39 02 5167 2459 Ind. Dial Face Email: [EMAIL PROTECTED] www.idf-hit.com GPG fingerprint: 3DD1 7B71 FBDF 6376 1B4A B003 175F E979 907E 1650 ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org http://lists.openwrt.org/cgi-bin/mailman/listinfo/openwrt-devel
[OpenWrt-Devel] [gcc4.3] Initial support for gcc4.3.1
DEFTREECODE +Index: gcc-4.3.0/gcc/rtl.h +=== +--- gcc-4.3.0/gcc/rtl.h(revision 130511) gcc-4.3.0/gcc/rtl.h(working copy) +@@ -48,9 +48,11 @@ + #include rtl.def/* rtl expressions are documented here */ + #undef DEF_RTL_EXPR + +- LAST_AND_UNUSED_RTX_CODE}; /* A convenient way to get a value for ++ LAST_AND_UNUSED_RTX_CODE/* A convenient way to get a value for + NUM_RTX_CODE. + Assumes default enum value assignment. */ ++ ,__LAST_AND_UNUSED_RTX_CODE=32767 /* Force 16bit width. */ ++}; + + #define NUM_RTX_CODE ((int) LAST_AND_UNUSED_RTX_CODE) + /* The cast here, saves many elsewhere. */ +Index: gcc-4.3.0/gcc/c-common.h +=== +--- gcc-4.3.0/gcc/c-common.h (revision 130511) gcc-4.3.0/gcc/c-common.h (working copy) +@@ -125,6 +125,7 @@ + RID_LAST_AT = RID_AT_IMPLEMENTATION, + RID_FIRST_PQ = RID_IN, + RID_LAST_PQ = RID_ONEWAY ++ ,__LAST_AND_UNUSED_RID=32767 /* Force 16bit width. */ + }; + + #define OBJC_IS_AT_KEYWORD(rid) \ diff --git a/toolchain/gcc/patches/4.3.1/998-gcc-4.3.0-fix-header.00.patch b/toolchain/gcc/patches/4.3.1/998-gcc-4.3.0-fix-header.00.patch new file mode 100644 index 000..7fe59d2 --- /dev/null +++ b/toolchain/gcc/patches/4.3.1/998-gcc-4.3.0-fix-header.00.patch @@ -0,0 +1,15 @@ + +\\ gcc PR33200 +Index: gcc-4.3.0/gcc/config.gcc +=== +--- gcc-4.3.0/gcc/config.gcc (revision 131628) gcc-4.3.0/gcc/config.gcc (working copy) +@@ -2302,7 +2305,7 @@ sh-*-symbianelf* | sh[12346l]*-*-symbian + if test x${enable_incomplete_targets} = xyes ; then + tm_defines=$tm_defines SUPPORT_SH1=1 SUPPORT_SH2E=1 SUPPORT_SH4=1 SUPPORT_SH4_SINGLE=1 SUPPORT_SH2A=1 SUPPORT_SH2A_SINGLE=1 SUPPORT_SH5_32MEDIA=1 SUPPORT_SH5_32MEDIA_NOFPU=1 SUPPORT_SH5_64MEDIA=1 SUPPORT_SH5_64MEDIA_NOFPU=1 + fi +- use_fixproto=yes ++ # XXX: why? use_fixproto=yes + ;; + sh-*-rtemscoff*) + tmake_file=sh/t-sh t-rtems sh/t-rtems diff --git a/toolchain/gcc/patches/4.3.1/999-cf.patch b/toolchain/gcc/patches/4.3.1/999-cf.patch new file mode 100644 index 000..2968e8d --- /dev/null +++ b/toolchain/gcc/patches/4.3.1/999-cf.patch @@ -0,0 +1,10 @@ +--- gcc-4.3.1/gcc/config.gcc.old 2008-06-17 23:49:00.0 +0200 gcc-4.3.1/gcc/config.gcc 2008-06-17 23:03:07.0 +0200 +@@ -1630,6 +1630,7 @@ + if test x$sjlj != x1; then + tmake_file=$tmake_file m68k/t-slibgcc-elf-ver + fi ++ tmake_file=m68k/t-floatlib m68k/t-m68kbare m68k/t-m68kelf + ;; + m68k-*-rtems*) + default_m68k_cpu=68020 diff --git a/toolchain/gmp/Makefile b/toolchain/gmp/Makefile new file mode 100644 index 000..c65069a --- /dev/null +++ b/toolchain/gmp/Makefile @@ -0,0 +1,46 @@ +# +# Copyright (C) 2006 OpenWrt.org +# Copyright (C) 2008 Luigi 'Comio' Mantellini [EMAIL PROTECTED] +#Industrie Dial Face S.p.A. +# +# This is free software, licensed under the GNU General Public License v2. +# See /LICENSE for more information. +# +include $(TOPDIR)/rules.mk + +PKG_NAME:=gmp +PKG_VERSION:=4.2.2 + +PKG_SOURCE_URL:=http://ftp.sunet.se/pub/gnu/gmp/ \ + ftp://ftp.gnu.org/gnu/gmp/ + +PKG_SOURCE:=$(PKG_NAME)-$(PKG_VERSION).tar.bz2 +PATCH_DIR:=./patches/$(PKG_VERSION) +# STAGING_DIR_HOST:=$(TOOLCHAIN_DIR) +BUILD_DIR_HOST:=$(BUILD_DIR_TOOLCHAIN) + +include $(INCLUDE_DIR)/host-build.mk + +define Build/Configure + (cd $(BUILD_DIR_HOST); \ +ln -sf $(PKG_NAME)-$(PKG_VERSION) $(PKG_NAME) \ + ); + (cd $(PKG_BUILD_DIR); \ + ./configure \ + --enable-fft \ + --build=$(GNU_HOST_NAME) \ + --host=$(GNU_HOST_NAME) \ + --prefix=$(STAGING_DIR_HOST) \ + ); +endef + +define Build/Compile + $(MAKE) -C $(PKG_BUILD_DIR) + $(MAKE) -C $(PKG_BUILD_DIR) check +endef + +define Build/Install + $(MAKE) -C $(PKG_BUILD_DIR) install +endef + +$(eval $(call HostBuild)) \ No newline at end of file diff --git a/toolchain/mpfr/Makefile b/toolchain/mpfr/Makefile new file mode 100644 index 000..19d3cac --- /dev/null +++ b/toolchain/mpfr/Makefile @@ -0,0 +1,46 @@ +# +# Copyright (C) 2006 OpenWrt.org +# Copyright (C) 2008 Luigi 'Comio' Mantellini [EMAIL PROTECTED] +#Industrie Dial Face S.p.A. +# +# This is free software, licensed under the GNU General Public License v2. +# See /LICENSE for more information. +# +include $(TOPDIR)/rules.mk + +PKG_NAME:=mpfr +PKG_VERSION:=2.3.1 + +PKG_SOURCE_URL:=http://www.mpfr.org/mpfr-current/ + +PKG_SOURCE:=$(PKG_NAME)-$(PKG_VERSION).tar.gz +PATCH_DIR:=./patches/$(PKG_VERSION) +# STAGING_DIR_HOST:=$(TOOLCHAIN_DIR) +BUILD_DIR_HOST:=$(BUILD_DIR_TOOLCHAIN) + +include
[OpenWrt-Devel] [gcc4.3] Compile free (tested on mipsel-2.6)
--- include/target.mk |4 ++-- toolchain/uClibc/Makefile |4 ++-- 2 files changed, 4 insertions(+), 4 deletions(-) diff --git a/include/target.mk b/include/target.mk index 8b12395..d9aa2be 100644 --- a/include/target.mk +++ b/include/target.mk @@ -105,8 +105,8 @@ $(eval $(call shexport,Target/Description)) include $(INCLUDE_DIR)/kernel-version.mk GENERIC_PLATFORM_DIR := $(TOPDIR)/target/linux/generic-$(KERNEL) -GENERIC_PATCH_DIR := $(GENERIC_PLATFORM_DIR)/patches$(shell [ -d $(GENERIC_PLATFORM_DIR)/patches-$(KERNEL_PATCHVER) ] printf -- -$(KERNEL_PATCHVER) || true ) -GENERIC_FILES_DIR := $(GENERIC_PLATFORM_DIR)/files$(shell [ -d $(GENERIC_PLATFORM_DIR)/files-$(KERNEL_PATCHVER) ] printf -- -$(KERNEL_PATCHVER) || true ) +GENERIC_PATCH_DIR ?= $(GENERIC_PLATFORM_DIR)/patches$(shell [ -d $(GENERIC_PLATFORM_DIR)/patches-$(KERNEL_PATCHVER) ] printf -- -$(KERNEL_PATCHVER) || true ) +GENERIC_FILES_DIR ?= $(GENERIC_PLATFORM_DIR)/files$(shell [ -d $(GENERIC_PLATFORM_DIR)/files-$(KERNEL_PATCHVER) ] printf -- -$(KERNEL_PATCHVER) || true ) GENERIC_LINUX_CONFIG?=$(firstword $(wildcard $(GENERIC_PLATFORM_DIR)/config-$(KERNEL_PATCHVER) $(GENERIC_PLATFORM_DIR)/config-default)) LINUX_CONFIG?=$(firstword $(wildcard $(foreach subdir,$(PLATFORM_DIR) $(PLATFORM_SUBDIR),$(subdir)/config-$(KERNEL_PATCHVER) $(subdir)/config-default)) $(PLATFORM_DIR)/config-$(KERNEL_PATCHVER)) diff --git a/toolchain/uClibc/Makefile b/toolchain/uClibc/Makefile index 37bc646..ee5b1d8 100644 --- a/toolchain/uClibc/Makefile +++ b/toolchain/uClibc/Makefile @@ -147,8 +147,8 @@ UCLIBC_MAKE := PATH=$(TARGET_PATH) $(MAKE) -C $(PKG_BUILD_DIR) \ define Build/Compile ifneq ($(CONFIG_GCC_VERSION_4_3_X),) - ln -sf $(TOOLCHAIN_DIR)/lib/gcc/m68k-linux-uclibc/$(CONFIG_GCC_VERSION)/include-fixed/limits.h \ - $(TOOLCHAIN_DIR)/lib/gcc/m68k-linux-uclibc/$(CONFIG_GCC_VERSION)/include/limits.h + ln -sf $(TOOLCHAIN_DIR)/lib/gcc/$(ARCH)-linux-uclibc/$(CONFIG_GCC_VERSION)/include-fixed/limits.h \ + $(TOOLCHAIN_DIR)/lib/gcc/$(ARCH)-linux-uclibc/$(CONFIG_GCC_VERSION)/include/limits.h endif $(SED) 's,^CROSS=.*,CROSS=$(TARGET_CROSS),g' $(PKG_BUILD_DIR)/Rules.mak $(UCLIBC_MAKE) PREFIX= all -- 1.5.4.3 ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org http://lists.openwrt.org/cgi-bin/mailman/listinfo/openwrt-devel
[OpenWrt-Devel] [gcc4.3] Fix board name
--- target/linux/cf547x/Makefile |2 +- 1 files changed, 1 insertions(+), 1 deletions(-) diff --git a/target/linux/cf547x/Makefile b/target/linux/cf547x/Makefile index c484362..3b88703 100644 --- a/target/linux/cf547x/Makefile +++ b/target/linux/cf547x/Makefile @@ -9,7 +9,7 @@ include $(TOPDIR)/rules.mk ARCH:=m68k -BOARD:=mc547x +BOARD:=cf547x BOARDNAME:=Generic Coldfire target FEATURES:=squashfs LINUX_VERSION:=2.6.24 -- 1.5.4.3 ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org http://lists.openwrt.org/cgi-bin/mailman/listinfo/openwrt-devel
[OpenWrt-Devel] [gcc4.3 V2 2/5] Fix board name
--- target/linux/cf547x/Makefile |2 +- 1 files changed, 1 insertions(+), 1 deletions(-) diff --git a/target/linux/cf547x/Makefile b/target/linux/cf547x/Makefile index c484362..3b88703 100644 --- a/target/linux/cf547x/Makefile +++ b/target/linux/cf547x/Makefile @@ -9,7 +9,7 @@ include $(TOPDIR)/rules.mk ARCH:=m68k -BOARD:=mc547x +BOARD:=cf547x BOARDNAME:=Generic Coldfire target FEATURES:=squashfs LINUX_VERSION:=2.6.24 -- 1.5.4.3 ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org http://lists.openwrt.org/cgi-bin/mailman/listinfo/openwrt-devel
[OpenWrt-Devel] [gcc4.3 V2 3/5] Avoid generic patches...
--- target/linux/cf547x/Makefile |2 ++ 1 files changed, 2 insertions(+), 0 deletions(-) diff --git a/target/linux/cf547x/Makefile b/target/linux/cf547x/Makefile index 3b88703..ae98efa 100644 --- a/target/linux/cf547x/Makefile +++ b/target/linux/cf547x/Makefile @@ -14,6 +14,8 @@ BOARDNAME:=Generic Coldfire target FEATURES:=squashfs LINUX_VERSION:=2.6.24 +GENERIC_PATCH_DIR:= + include $(INCLUDE_DIR)/target.mk define Target/Description -- 1.5.4.3 ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org http://lists.openwrt.org/cgi-bin/mailman/listinfo/openwrt-devel
Re: [OpenWrt-Devel] [gcc4.3] Initial support for gcc4.3.1
Please ignore these patches. I will resend with numbering asap. Thanks. luigi On mer, 2008-06-18 at 17:03 +0200, Luigi 'Comio' Mantellini wrote: --- toolchain/gcc/Config.in|3 + toolchain/gcc/Config.version | 18 ++ toolchain/gcc/Makefile | 28 ++ toolchain/gcc/patches/4.3.1/100-uclibc-conf.patch | 33 +++ .../patches/4.3.1/104-gnuhurd-uclibc-conf.patch| 13 + .../gcc/patches/4.3.1/301-missing-execinfo_h.patch | 11 + toolchain/gcc/patches/4.3.1/302-c99-snprintf.patch | 13 + ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org http://lists.openwrt.org/cgi-bin/mailman/listinfo/openwrt-devel
[OpenWrt-Devel] [GCC4.3 v3 3/3] Minor fixes
--- toolchain/gcc/Makefile |6 +++--- 1 files changed, 3 insertions(+), 3 deletions(-) diff --git a/toolchain/gcc/Makefile b/toolchain/gcc/Makefile index cd0203f..a04c059 100644 --- a/toolchain/gcc/Makefile +++ b/toolchain/gcc/Makefile @@ -48,7 +48,7 @@ else SOFT_FLOAT_CONFIG_OPTION:=--with-float=soft endif -#ifeq ($(CONFIG_GCC_VERSION_4_3_1),) +#ifeq ($(CONFIG_GCC_VERSION_4_3_X),) #WITH_GMP_MPFR:= #else #WITH_GMP_MPFR:=--with-mpfr=$(TOPDIR)/staging_dir/host --with-gmp=$(TOPDIR)/staging_dir/host @@ -81,7 +81,7 @@ define Stage1/Configure ); endef -ifeq ($(CONFIG_GCC_VERSION_4_3_1),) +ifeq ($(CONFIG_GCC_VERSION_4_3_X),) define Stage1/Compile export SHELL=$(BASH); $(MAKE) -C $(BUILD_DIR1) all-gcc endef @@ -91,7 +91,7 @@ define Stage1/Compile endef endif -ifeq ($(CONFIG_GCC_VERSION_4_3_1),) +ifeq ($(CONFIG_GCC_VERSION_4_3_X),) define Stage1/Install export SHELL=$(BASH); $(MAKE) -C $(BUILD_DIR1) install-gcc endef -- 1.5.4.3 ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org http://lists.openwrt.org/cgi-bin/mailman/listinfo/openwrt-devel
Re: [OpenWrt-Devel] [GCC4.3 v3 1/3] Initial support for gcc4.3.1
Just an important note: the patches are from buildroot git tree maintained by Bernhard Fischer (thanks!). On mer, 2008-06-18 at 19:01 +0200, Luigi 'Comio' Mantellini wrote: --- toolchain/gcc/Config.in|3 + toolchain/gcc/Config.version | 18 ++ toolchain/gcc/Makefile | 28 ++ toolchain/gcc/patches/4.3.1/100-uclibc-conf.patch | 33 +++ .../patches/4.3.1/104-gnuhurd-uclibc-conf.patch| 13 + .../gcc/patches/4.3.1/301-missing-execinfo_h.patch | 11 + toolchain/gcc/patches/4.3.1/302-c99-snprintf.patch | 13 + .../4.3.1/305-libmudflap-susv3-legacy.patch| 49 .../patches/4.3.1/810-arm-softfloat-libgcc.patch | 29 ++ toolchain/gcc/patches/4.3.1/910-mbsd_multi.patch | 284 .../4.3.1/993-arm_insn-opinit-RTX_CODE-fixup.patch | 41 +++ toolchain/gcc/patches/4.3.1/995-short-enums.diff | 42 +++ .../4.3.1/998-gcc-4.3.0-fix-header.00.patch| 15 + toolchain/gcc/patches/4.3.1/999-cf.patch | 10 + toolchain/gmp/Makefile | 46 toolchain/mpfr/Makefile| 46 toolchain/uClibc/Makefile |5 + 17 files changed, 686 insertions(+), 0 deletions(-) create mode 100644 toolchain/gcc/patches/4.3.1/100-uclibc-conf.patch create mode 100644 toolchain/gcc/patches/4.3.1/104-gnuhurd-uclibc-conf.patch create mode 100644 toolchain/gcc/patches/4.3.1/301-missing-execinfo_h.patch create mode 100644 toolchain/gcc/patches/4.3.1/302-c99-snprintf.patch create mode 100644 toolchain/gcc/patches/4.3.1/305-libmudflap-susv3-legacy.patch create mode 100644 toolchain/gcc/patches/4.3.1/810-arm-softfloat-libgcc.patch create mode 100644 toolchain/gcc/patches/4.3.1/910-mbsd_multi.patch create mode 100644 toolchain/gcc/patches/4.3.1/993-arm_insn-opinit-RTX_CODE-fixup.patch create mode 100644 toolchain/gcc/patches/4.3.1/995-short-enums.diff create mode 100644 toolchain/gcc/patches/4.3.1/998-gcc-4.3.0-fix-header.00.patch create mode 100644 toolchain/gcc/patches/4.3.1/999-cf.patch create mode 100644 toolchain/gmp/Makefile create mode 100644 toolchain/mpfr/Makefile Industrie Dial Face S.p.A. Luigi Mantellini RD - Software Industrie Dial Face S.p.A. Via Canzo, 4 20068 Peschiera Borromeo (MI), Italy Tel.: +39 02 5167 2813 Fax:+39 02 5167 2459 E-mail: [EMAIL PROTECTED] GPG fingerprint: 3DD1 7B71 FBDF 6376 1B4A B003 175F E979 907E 1650 attachment: idf_logo.gif___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org http://lists.openwrt.org/cgi-bin/mailman/listinfo/openwrt-devel
Re: [OpenWrt-Devel] GCC4.3 patch (wip)
Hi Hinko, Hi List, see my inline comments. On lun, 2008-06-16 at 10:05 +0200, Hinko Kocevar wrote: Luigi 'Comio' Mantellini wrote: Ciao again, On ven, 2008-06-13 at 11:22 +0200, Luigi 'Comio' Mantellini wrote: Ciao, make[4]: Entering directory `/mnt/devel/openwrt/OpenWRT.git/build_dir/toolchain-m68k_gcc4.3.1/uClibc-0.9.29' install -d /mnt/devel/openwrt/OpenWRT.git/staging_dir/toolchain-m68k_gcc4.3.1//lib install -m 644 lib/lib*-0.9.29.so \ /mnt/devel/openwrt/OpenWRT.git/staging_dir/toolchain-m68k_gcc4.3.1//lib cd lib tar -cf - *.so.* | tar -xf - -C /mnt/devel/openwrt/OpenWRT.git/staging_dir/toolchain-m68k_gcc4.3.1//lib tar: error while loading shared libraries: libpthread.so.0: ELF file data encoding not little-endian tar: error while loading shared libraries: libpthread.so.0: ELF file data encoding not little-endian You've seem to remove some some mandatory path in system library search - if your had LD_LIBRARY_PATH set you've might need to do: export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:/path/to/your/lib in contrast to export LD_LIBRARY_PATH=/path/to/your/lib I added the following into $(TOP)/rule.mk (and I don't like what I do...): export LD_LIBRARY_PATH=$(TOP)/staging_dir/host/lib:$(LD_LIBRARY_PATH) Anyway I decided to change approach: install directly on the host environment the GMP and MPFR library need to GCC4.3 compiler... OpenWRT should make some check to recognize these libraries: ifneq ($(CONFIG_GCC_VERSION_4_3_X),) define Require/gmp echo 'int main(int argc, char **argv) { return ((__GNU_MP_VERSION*1000+__GNU_MP_VERSION_MINOR)*1000+__GNU_MP_VERSION_PATCHLEVEL)=004001000?0:1; }' | gcc -include gmp.h -x c -o $(TMP_DIR)/gmptest.o -lgmp - endef $(eval $(call Require,gmp, \ Please install lib gmp = 4.1.0 \ )) define Require/mpfr echo 'int main(int argc, char **argv) { return (MPFR_VERSION=MPFR_VERSION_NUM(2,3,0))?0:1; }' | \ gcc -include mpfr.h -x c -o $(TMP_DIR)/a.out -lmpfr - endef $(eval $(call Require,mpfr, \ Please install lib mpfr = 2.3.0 \ )) endif Unfortunately, this code cannot be performed by prereq-build.mk because we don't know if we need gcc4.3 or not. My self-answer: This error is caused by the LD_LIBRARY_PATH that I defined into rule.mk. At this point I need to resolve an issue: GCC4.3 requires gmp and mpfr libraries. I added the makefile under the toolchain directory in order to create the needed libraries (libgmp.so and libmpfr.so). I placed these libraries into the staging_dir/host/lib/ directory... but (cross)gcc doesn't start because this directory is not in the ld path... I thought to solve this issue using the LD_LIBRARY_PATH=$(TOP)/staging_dir/host/lib/ ... but this creates the error in tar application Now, with LD_LIBRARY_PATH env. variable removed, I'm compiling the gcc-final... but I will need it when the (cross)gcc is invoked. question: Where can I place the host libraries (__NOT TARGET__) in order to be visible to cross-compiler (and in general, to crosstools)? Maybe all you need to do is correctly supply the full path to the includes/libraries via C(PP)FLAGS/LDFLAGS in the makefile, or on the command line instead of using LD_LIBRARY_PATH. Another approach would be to statically build/link the tar app inside the openWRT SDK and use the LD_LIBRARY_PATH after all. HTH, Hinko I would patch OpenWRT as little as possible, of course. The solution should be to link statically libgmp and libmpfr to gcc cross-compiler... but How can I say this to the gcc configure? Furthermore I chosen to avoid to compile the mpfr and gmp into the OpenWRT environment... Now I have an error during the gcc-final regarding ctype_basechar... Thanks and Best regards, luigi Industrie Dial Face S.p.A. Luigi Mantellini RD - Software Industrie Dial Face S.p.A. Via Canzo, 4 20068 Peschiera Borromeo (MI), Italy Tel.: +39 02 5167 2813 Fax:+39 02 5167 2459 E-mail: [EMAIL PROTECTED] GPG fingerprint: 3DD1 7B71 FBDF 6376 1B4A B003 175F E979 907E 1650 attachment: idf_logo.gif___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org http://lists.openwrt.org/cgi-bin/mailman/listinfo/openwrt-devel
Re: [OpenWrt-Devel] GCC4.3 patch (wip)
Ciao, in order to advance the building I manually created the following link: cd staging_dir/toolchain-m68k_gcc4.3.1/lib/gcc/m68k-linux-uclibc/4.3.1/include ln -s ../include-fixed/limits.h . In your opinion... can it be a solution? with this workaround the uClibc compilation advances untile this strange error: make[4]: Leaving directory `/mnt/devel/openwrt/OpenWRT.git/build_dir/toolchain-m68k_gcc4.3.1/uClibc-0.9.29' PATH=/mnt/devel/openwrt/OpenWRT.git/staging_dir/toolchain-m68k_gcc4.3.1/bin:/mnt/devel/openwrt/OpenWRT.git/staging_dir/host/bin:/mnt/devel/openwrt/OpenWRT.git/staging_dir/m68k/host/bin:/mnt/devel/openwrt/OpenWRT.git/staging_dir/toolchain-m68k_gcc4.3.1/bin:/mnt/devel/openwrt/OpenWRT.git/staging_dir/host/bin:/mnt/devel/openwrt/OpenWRT.git/staging_dir/m68k/host/bin:/mnt/devel/openwrt/OpenWRT.git/staging_dir/toolchain-m68k_gcc4.3.1/bin:/mnt/devel/openwrt/OpenWRT.git/staging_dir/host/bin:/mnt/devel/openwrt/OpenWRT.git/staging_dir/m68k/host/bin:/mnt/devel/openwrt/OpenWRT.git/staging_dir/toolchain-m68k_gcc4.3.1/bin:/mnt/devel/openwrt/OpenWRT.git/staging_dir/host/bin:/mnt/devel/openwrt/OpenWRT.git/staging_dir/m68k/host/bin:/mnt/devel/openwrt/OpenWRT.git/staging_dir/toolchain-m68k_gcc4.3.1/bin:/mnt/devel/openwrt/OpenWRT.git/staging_dir/host/bin:/mnt/devel/openwrt/OpenWRT.git/staging_dir/m68k/host/bin:/mnt/devel/openwrt/OpenWRT.git/staging_dir/toolchain-m68k_gcc4.3.1/bin:/mnt/devel/ope nwrt/OpenWRT.git/staging_dir/host/bin:/mnt/devel/openwrt/OpenWRT.git/staging_dir/m68k/host/bin:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/opt/t32/bin/pc_linux:/opt/Arabella/bin PATH=/mnt/devel/openwrt/OpenWRT.git/staging_dir/toolchain-m68k_gcc4.3.1/bin:/mnt/devel/openwrt/OpenWRT.git/staging_dir/host/bin:/mnt/devel/openwrt/OpenWRT.git/staging_dir/m68k/host/bin:/mnt/devel/openwrt/OpenWRT.git/staging_dir/toolchain-m68k_gcc4.3.1/bin:/mnt/devel/openwrt/OpenWRT.git/staging_dir/host/bin:/mnt/devel/openwrt/OpenWRT.git/staging_dir/m68k/host/bin:/mnt/devel/openwrt/OpenWRT.git/staging_dir/toolchain-m68k_gcc4.3.1/bin:/mnt/devel/openwrt/OpenWRT.git/staging_dir/host/bin:/mnt/devel/openwrt/OpenWRT.git/staging_dir/m68k/host/bin:/mnt/devel/openwrt/OpenWRT.git/staging_dir/toolchain-m68k_gcc4.3.1/bin:/mnt/devel/openwrt/OpenWRT.git/staging_dir/host/bin:/mnt/devel/openwrt/OpenWRT.git/staging_dir/m68k/host/bin:/mnt/devel/openwrt/OpenWRT.git/staging_dir/toolchain-m68k _gcc4.3.1/bin:/mnt/devel/openwrt/OpenWRT.git/staging_dir/host/bin:/mnt/devel/openwrt/OpenWRT.git/staging_dir/m68k/host/bin:/mnt/devel/openwrt/OpenWRT.git/staging_dir/toolchain-m68k_gcc4.3.1/bin:/mnt/devel/openwrt/OpenWRT.git/staging_dir/host/bin:/mnt/devel/openwrt/OpenWRT.git/staging_dir/m68k/host/bin:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/opt/t32/bin/pc_linux:/opt/Arabella/bin make -C /mnt/devel/openwrt/OpenWRT.git/build_dir/toolchain-m68k_gcc4.3.1/uClibc-0.9.29 AR=m68k-linux-uclibc-ar AS=m68k-linux-uclibc-gcc -c -Os -pipe -funit-at-a-time -fhonour-copts LD=m68k-linux-uclibc-ld NM=m68k-linux-uclibc-nm CC=m68k-linux-uclibc-gcc GCC=m68k-linux-uclibc-gcc CXX=m68k-linux-uclibc-g++ RANLIB=m68k-linux-uclibc-ranlib STRIP=m68k-linux-uclibc-strip OBJCOPY=m68k-linux-uclibc-objcopy OBJDUMP=m68k-linux-uclibc-objdump SIZE=m68k-linux-uclibc-size DEVEL_PREFIX=/ RUNTIME_PREFIX=/ HOSTCC=gcc CPU_CFLAGS=-Os -pipe -funit-at-a-time -fhonour-copts PREF IX=/mnt/devel/openwrt/OpenWRT.git/staging_dir/toolchain-m68k_gcc4.3.1/ install_runtime install_dev make[4]: Entering directory `/mnt/devel/openwrt/OpenWRT.git/build_dir/toolchain-m68k_gcc4.3.1/uClibc-0.9.29' install -d /mnt/devel/openwrt/OpenWRT.git/staging_dir/toolchain-m68k_gcc4.3.1//lib install -m 644 lib/lib*-0.9.29.so \ /mnt/devel/openwrt/OpenWRT.git/staging_dir/toolchain-m68k_gcc4.3.1//lib cd lib tar -cf - *.so.* | tar -xf - -C /mnt/devel/openwrt/OpenWRT.git/staging_dir/toolchain-m68k_gcc4.3.1//lib tar: error while loading shared libraries: libpthread.so.0: ELF file data encoding not little-endian tar: error while loading shared libraries: libpthread.so.0: ELF file data encoding not little-endian make[4]: *** [install_runtime] Error 127 make[4]: Leaving directory `/mnt/devel/openwrt/OpenWRT.git/build_dir/toolchain-m68k_gcc4.3.1/uClibc-0.9.29' make[3]: *** [/mnt/devel/openwrt/OpenWRT.git/staging_dir/toolchain-m68k_gcc4.3.1/stamp/.uclibc_installed] Error 2 make[3]: Leaving directory `/mnt/devel/openwrt/OpenWRT.git/toolchain/uClibc' make[2]: *** [toolchain/uClibc/compile] Error 2 make[2]: Leaving directory `/mnt/devel/openwrt/OpenWRT.git' make[1]: *** [/mnt/devel/openwrt/OpenWRT.git/staging_dir/m68k/stamp/.toolchain_install] Error 2 make[1]: Leaving directory `/mnt/devel/openwrt/OpenWRT.git' make: *** [world] Error 2 Of course the tar command works fine on my system. Any idea? luigi On ven, 2008-06-13 at 09:34 +0200, Luigi 'Comio' Mantellini wrote: Sorry. You are right. (I'm working also on a mips architecture and I have made
Re: [OpenWrt-Devel] GCC4.3 patch (wip)
Ciao again, On ven, 2008-06-13 at 11:22 +0200, Luigi 'Comio' Mantellini wrote: Ciao, make[4]: Entering directory `/mnt/devel/openwrt/OpenWRT.git/build_dir/toolchain-m68k_gcc4.3.1/uClibc-0.9.29' install -d /mnt/devel/openwrt/OpenWRT.git/staging_dir/toolchain-m68k_gcc4.3.1//lib install -m 644 lib/lib*-0.9.29.so \ /mnt/devel/openwrt/OpenWRT.git/staging_dir/toolchain-m68k_gcc4.3.1//lib cd lib tar -cf - *.so.* | tar -xf - -C /mnt/devel/openwrt/OpenWRT.git/staging_dir/toolchain-m68k_gcc4.3.1//lib tar: error while loading shared libraries: libpthread.so.0: ELF file data encoding not little-endian tar: error while loading shared libraries: libpthread.so.0: ELF file data encoding not little-endian My self-answer: This error is caused by the LD_LIBRARY_PATH that I defined into rule.mk. At this point I need to resolve an issue: GCC4.3 requires gmp and mpfr libraries. I added the makefile under the toolchain directory in order to create the needed libraries (libgmp.so and libmpfr.so). I placed these libraries into the staging_dir/host/lib/ directory... but (cross)gcc doesn't start because this directory is not in the ld path... I thought to solve this issue using the LD_LIBRARY_PATH=$(TOP)/staging_dir/host/lib/ ... but this creates the error in tar application Now, with LD_LIBRARY_PATH env. variable removed, I'm compiling the gcc-final... but I will need it when the (cross)gcc is invoked. question: Where can I place the host libraries (__NOT TARGET__) in order to be visible to cross-compiler (and in general, to crosstools)? I'm sorry for this ship's log but I'm not an OpenWRT developer and I don't know the internals... I only know that I need to port OpenWRT to Coldfire architecture Thanks a lot and all comments are wellcome. luigi Of course the tar command works fine on my system. Any idea? luigi On ven, 2008-06-13 at 09:34 +0200, Luigi 'Comio' Mantellini wrote: Sorry. You are right. (I'm working also on a mips architecture and I have made some mistake in copying). Try this. Thanks, luigi On ven, 2008-06-13 at 09:10 +0200, Geert Uytterhoeven wrote: On Fri, 13 Jun 2008, Luigi 'Comio' Mantellini wrote: I'm still working to port OpenWRT on my architecture (m68k/coldfire). I diff --git a/target/linux/generic_m68k/config-default b/target/linux/generic_m68k/config-default new file mode 100644 index 000..6f8acb5 --- /dev/null +++ b/target/linux/generic_m68k/config-default @@ -0,0 +1,616 @@ +# +# Automatically generated make config: don't edit +# Linux kernel version: 2.6.25 +# Mon May 12 19:37:01 2008 +# +CONFIG_MIPS=y That looks a bit strange for an m68k config... diff --git a/target/linux/generic_m68k/image/lzma-loader/src/Makefile b/target/linux/generic_m68k/image/lzma-loader/src/Makefile new file mode 100644 index 000..910172c --- /dev/null +++ b/target/linux/generic_m68k/image/lzma-loader/src/Makefile @@ -0,0 +1,68 @@ +# +# Copyright (C) 2006 OpenWrt.org +# +# This is free software, licensed under the GNU General Public License v2. +# See /LICENSE for more information. +# +RAMSTART = 0x8000 +RAMSIZE = 0x0010 # 1MB +LOADADDR = 0x8040 # RAM start + 4M +KERNEL_ENTRY = 0x80001000 +IMAGE_COPY:=0 + +CROSS_COMPILE = mips-linux- Another one... diff --git a/target/linux/generic_m68k/image/lzma-loader/src/decompress.c b/target/linux/generic_m68k/image/lzma-loader/src/decompress.c new file mode 100644 index 000..45ac509 --- /dev/null +++ b/target/linux/generic_m68k/image/lzma-loader/src/decompress.c +#define cache_unroll(base,op) \ + __asm__ __volatile__( \ + .set noreorder;\n \ + .set mips3;\n \ + cache %1, (%0);\n \ + .set mips0;\n \ + .set reorder\n\ + : \ + : r (base), \ + i (op)); Erhm... And there are many more... Gr{oetje,eeting}s, Geert -- Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- [EMAIL PROTECTED] In personal conversations with technical people, I call myself a hacker. But when I'm talking to journalists I just say programmer or something like that. -- Linus Torvalds ___ __ Luigi Mantellini .'__'. RD - Software (.' '.)Industrie Dial Face S.p.A. ( :==: )Via
Re: [OpenWrt-Devel] any problems with bumping up the toolchain versions?
On gio, 2008-03-27 at 23:42 +0100, Imre Kaloz wrote: On 2008.03.27. 23:37:47 Luigi 'Comio' Mantellini [EMAIL PROTECTED] wrote: snip Ok. Anyway the compiler gcc4.3 supports a lot new cpus (like the coldfire) :) opening new development horizons. Sure, as AVR32 uses gcc 4.2.3.. Some targets work better with newer compilers, others do not. Spice this is up with uClibc, and you have fun everytime you play with toolchain combinations ;) rday already sent a preliminary 4.3 patch top the list, if you volunteer to fix it up, we are happy to add it :) Same goes for your platform :) I tried the patchset from RDAY without success :S I don't know why, but the file libgcc.a is not found and the uClibc compilation fails. Imre ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org http://lists.openwrt.org/cgi-bin/mailman/listinfo/openwrt-devel -- __ Luigi Mantellini .'__'. RD - Software (.' '.)Industrie Dial Face S.p.A. ( :==: )Via Canzo, 4 ('.__.')20068 Peschiera Borromeo (MI), Italy '.__.' Tel.: +39 02 5167 2813 Fax: +39 02 5167 2459 Ind. Dial Face Email: [EMAIL PROTECTED] www.idf-hit.com GPG fingerprint: 3DD1 7B71 FBDF 6376 1B4A B003 175F E979 907E 1650 ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org http://lists.openwrt.org/cgi-bin/mailman/listinfo/openwrt-devel