- openSUSE ARM

[opensuse-arm] Build problem due to non resolable issue
by Joop Boonen 12 Feb '12

12 Feb '12

Hi, Currently no package is build any more. for standard armv7l. As pcre is depending on grep which again is depending on libpce, "nothing provides libpcre.so.0 needed by grep" How can this be resolved? Regards, Joop. -- To unsubscribe, e-mail: opensuse-arm+unsubscribe(a)opensuse.org To contact the owner, e-mail: opensuse-arm+owner(a)opensuse.org

2 1

[opensuse-arm] ro
by Peter Czanik 10 Feb '12

10 Feb '12

Hello, Recently any time I boot my EFIKA, I get many error messages, that / is read only and some services can't be started. Is it specific to EFIKA (the image from Bernhard, which uses the Genesi provided kernel), or generic ARM problem, or something I messed up? I recall reading something on IRC about read only root and updated my machine recently... Bye, CzP -- To unsubscribe, e-mail: opensuse-arm+unsubscribe(a)opensuse.org To contact the owner, e-mail: opensuse-arm+owner(a)opensuse.org

2 2

[opensuse-arm] rpmlint issue for ARM OBS armv7l
by Joop Boonen 08 Feb '12

08 Feb '12

Hi All, When I package home:worldcitizen:armv7l xorg-x11-drv-ste I get an rpmlint error: E: xorg-x11-drv-ste 64bit-portability-issue mali_fbdev.c:1177 I think this is incorrect as I there is no error in the compile log, only some warnings, and 64 bit is not an issue for ARM (yet). As it's an error I cant suppress it via an rpmlintrc filter file. Regards, Joop. -- To unsubscribe, e-mail: opensuse-arm+unsubscribe(a)opensuse.org To contact the owner, e-mail: opensuse-arm+owner(a)opensuse.org

2 1

[opensuse-arm] [PATCH] arm: use ext2 for booting
by Alexander Graf 06 Feb '12

06 Feb '12

Now that x-loader on OMAP can boot from ext2, let's also teach kiwi about that awesome new capability. This gets rid of the vfat dependency and makes the whole bootstrap a lot cleaner. Signed-off-by: Alexander Graf <agraf(a)suse.de> --- modules/KIWIBoot.pm | 51 +++++++++++++++++++++++++++++++++++++++-------- modules/KIWILinuxRC.sh | 4 +- 2 files changed, 44 insertions(+), 11 deletions(-) diff --git a/modules/KIWIBoot.pm b/modules/KIWIBoot.pm index 88d8d41..21b623c 100644 --- a/modules/KIWIBoot.pm +++ b/modules/KIWIBoot.pm @@ -1711,7 +1711,7 @@ sub setupBootDisk { # setup boot partition type #------------------------------------------ my $partid = 83; - if ($bootloader =~ /syslinux|uboot/) { + if ($bootloader =~ /syslinux/) { $partid = "c"; } elsif ($bootloader eq "yaboot") { $partid = "41"; @@ -2224,7 +2224,7 @@ sub setupBootDisk { if ($lvm) { $boot = $deviceMap{0}; } - if ($bootloader =~ /syslinux|uboot|yaboot/) { + if ($bootloader =~ /syslinux|yaboot/) { $kiwi -> info ("Creating DOS boot filesystem"); if ($bootloader eq "yaboot") { $status = qxx ("/sbin/mkdosfs -F 16 $boot 2>&1"); @@ -3904,8 +3904,8 @@ sub setupBootLoaderConfiguration { print FD 'setenv initrd_high "0xffffffff";'."\n"; print FD 'setenv fdt_high "0xffffffff";'."\n"; print FD 'setenv bootcmd "'; - print FD 'fatload mmc 0:1 0x80000000 boot/linux.vmx; '; - print FD 'fatload mmc 0:1 0x81600000 boot/initrd.uboot; '; + print FD 'ext2load mmc 0:1 0x80000000 boot/linux.vmx; '; + print FD 'ext2load mmc 0:1 0x81600000 boot/initrd.uboot; '; print FD 'bootm 0x80000000 0x81600000";'."\n"; if ($type =~ /^KIWI CD/) { $kiwi -> failed (); @@ -4464,7 +4464,44 @@ sub installBootLoader { # install uboot #------------------------------------------ if ($loader eq "uboot") { - # uboot loader doesn't get installed + if (! $deviceMap) { + $kiwi -> failed (); + $kiwi -> error ("No device map available"); + $kiwi -> failed (); + return; + } + my %deviceMap = %{$deviceMap}; + my $device = $deviceMap{$bootpart+1}; + + #========================================== + # mount boot device... + #------------------------------------------ + $status = qxx ("mount $device /mnt 2>&1"); + $result = $? >> 8; + if ($result != 0) { + $kiwi -> failed (); + $kiwi -> error ("Can't mount boot partition: $status"); + $kiwi -> failed (); + $this -> cleanLoop (); + return; + } + #========================================== + # install MLO as raw + #------------------------------------------ + if (-f "/mnt/boot/MLO") { + $kiwi -> info ("Installing MLO on device: $diskname"); + $status = qxx ("dd if=/mnt/boot/MLO of=$diskname bs=128k count=1 seek=1 conv=notrunc 2>&1"); + $result = $? >> 8; + if ($result != 0) { + $kiwi -> failed (); + $kiwi -> error ("Couldn't install MLO on $diskname: $status"); + $kiwi -> failed (); + qxx ("umount /mnt 2>&1"); + $this -> cleanLoop (); + return; + } + } + qxx ("umount /mnt 2>&1"); } #========================================== # more boot managers to come... @@ -5496,10 +5533,6 @@ sub __getBootSize { if ($gotMB < $minMB) { $gotMB = $minMB; } - # on arm it's required to have a maximum of 64MB for the boot part - if ($arch =~ /arm/) { - $gotMB = 64; - } $kiwi -> loginfo ("Set boot space to: ".$gotMB."M\n"); return $gotMB; } diff --git a/modules/KIWILinuxRC.sh b/modules/KIWILinuxRC.sh index 70e2688..6f37c6f 100644 --- a/modules/KIWILinuxRC.sh +++ b/modules/KIWILinuxRC.sh @@ -1544,8 +1544,8 @@ function setupBootLoaderUBoot { echo "setenv initrd_high \"0xffffffff\";" >> $conf echo "setenv fdt_high \"0xffffffff\";" >> $conf echo -n "setenv bootcmd \"" >> $conf - echo -n "fatload mmc 0:1 0x80000000 $kernel; " >> $conf - echo -n "fatload mmc 0:1 0x81600000 $initrd; " >> $conf + echo -n "ext2load mmc 0:1 0x80000000 $kernel; " >> $conf + echo -n "ext2load mmc 0:1 0x81600000 $initrd; " >> $conf echo "bootm 0x80000000 0x81600000\"" echo -n "setenv bootargs root=$diskByID loader=$loader" >> $conf if [ ! -z "$imageDiskDevice" ];then -- 1.6.0.2 -- To unsubscribe, e-mail: opensuse-arm+unsubscribe(a)opensuse.org To contact the owner, e-mail: opensuse-arm+owner(a)opensuse.org

1 0

[opensuse-arm] Build openjdk6 with a different virtual machine?
by Michal Vyskocil 03 Feb '12

03 Feb '12

Hallo all, please leavy me in a CC as I am not on the list thanks a hard work of Dirk, we are now able to build and run openjdk6 in Factory for arm. The current incarnation use the Zero Assembler port [1]. It means there is no JIT and java runs in interpreted mode only, which is slow. icedtea patchset provide a neat support for build openjdk with different vm, than hotspot (or zero). For arm users I would say there are three choices: * shark - an llvm based experimental JIT * JamVM a state of the art interpreter using dynamic super instructions (a code copying JIT) with quick startup time and low memory usage suitable for embedded and client use, developed by Robert Lougher. This JVM are the default on Ubuntu/ARM systems. JamVM work on armel, armhf, OpenJDK-6 and OpenJDK-7. * CACAO a state of art JIT, developed as a research JVM by the university of Vienna. Working on armel and openjdk-6. See labb.zafena.se/?p=471 for other details I did some basic research what other distributions use: * Fedora recently added a support for arm into rawhide, however they use zero * Debian supports everything (zero, zero/shark, cacao, jamvm) It seems that cacao would be the best choice, as it's mature, has big performance and so. However it does not work with openjdk7, even there is a recent work [2] to make it happen. And what can be worse, the armhf is not supported by CACAO atm and I did not find anything related. As openjdk7 is still not an issue for openSUSE@arm, the armhf might be. So what will be your choice for VM? 1.) CACAO (aka use stable, faster solution) 2.) JamVM (aka we need armhf) 3.) do-not-touch-it Note: I wont follow Debian support everything approach, their rules [3] file scares me as it looks worse than our spec file. [1] http://icedtea.classpath.org/wiki/ZeroSharkFaq [2] http://mips.complang.tuwien.ac.at/hg/cacao/rev/30edcde287d9 [3] http://bazaar.launchpad.net/~openjdk/openjdk/openjdk7/view/head:/rules Best regards Michal Vyskocil

2 1

[opensuse-arm] omap fatless boot
by Alexander Graf 03 Feb '12

03 Feb '12

Hi list, Dirk had an idea a while back. On OMAP we have this pretty sad situation where we need to have /boot be a FAT partition so we can store MLO and u-boot.bin on there. But why do we need FAT there? Let's take a look at the boot process: - built-in ROM boots - built-in ROM searches for MLO in first active FAT partition - built-in ROM starts MLO - MLO searches for u-boot.bin in first(!) FAT partition - MLO boots u-boot So both components have FAT hardcoded in them. Or do they? The built-in ROM actually has an alternative boot mode which is called "raw". If it finds a small header at certain locations of the SD card (offset 0, offset 128kb) it loads MLO from there instead of FAT. Since we always align our partitions on 1MB boundaries, we have 1MB of free space below the first partition! MLO may only be up to 64kb big, so it fits there pretty well. Now to the next problem. MLO uses FAT. But MLO really is x-loader which is a modified version of u-boot. And u-boot can do ext2. So why don't we just take u-boot's ext2 support and throw it into our MLO binary? Well, this whole thing turned out a bit trickier than I anticipated, but eventually I got it rolling. Here's the patch: https://build.opensuse.org/package/view_file?file=0001-add-ext2-support.pat… I also verified that it works at least in an emulator. Since I don't have a beagle board, I couldn't verify if it works on real hardware. But to me the beagle target is mostly important for the emulation bit anyway :). The next target is x-loader for omap4 (panda) which hopefully will be easier to deal with. And once that's ext2'ified, we can remove the whole FAT dependency from our shiny ARM target! Yay! Happy hacking, and see you at fosdem! Alex PS: I also had to fix qemu's raw booting mode, since it wasn't behaving according to spec. So don't try this on qemu-linaro just yet :).-- To unsubscribe, e-mail: opensuse-arm+unsubscribe(a)opensuse.org To contact the owner, e-mail: opensuse-arm+owner(a)opensuse.org

1 0

[opensuse-arm] Re: Wanted: superstar hacker to complete Mono port to armhf
by Alexander Graf 02 Feb '12

02 Feb '12

On 03.02.2012, at 00:11, Hector Oron wrote: > Hello Jo, > > I am forwarding the message to a couple mailing lists which might have > people interested on the Mono porting for ARM hard-float ABI. > > 2012/2/2 Jo Shields <directhex(a)apebox.org>: >> Right now, Mono is available in Debian armhf. This is a hack - what >> we're actually doing is building Mono as an armhf binary, but built to >> emit soft VFP instructions and using calling conventions and ABI for >> armel. This hack works well enough for pure cross-platform code (like >> the C# compiler) to run, but dies in a heap for anything complex. >> >> This situation is a bit on the crappy side of crap. >> >> In order for Mono on armhf not to be a waste of time, a "true" port >> needs to be completed. If I were to make a not-remotely-educated guess, >> I'd say it needs about 550 lines of changes, primarily the addition of >> code to emit the correct instructions feeling the correct registers in >> mono/mini/mini-arm.c plus a couple of tweaks to related headers. >> >> Upstream have also indicated that they're happy to provide guidance and >> pointers on how to implement this port, although they're unable to >> provide the requisite code themselves. >> >> Sadly, unless someone in the community is able to step forward and >> contribute here, it's only a matter of time before the armhf packages >> are rightfully marked RC-buggy, and 100+ packages need to be axed from >> armhf. This would make me sad. Please check our mono arm patches in OBS: https://build.opensuse.org/package/files?package=mono-core&project=openSUSE… While slightly hacky, they enable full armhf support for arm. At least for us it's worked pretty well. Mono is a rather core dependency of a lot of stuff. Alex -- To unsubscribe, e-mail: opensuse-arm+unsubscribe(a)opensuse.org To contact the owner, e-mail: opensuse-arm+owner(a)opensuse.org

2 3

[opensuse-arm] kiwi build
by Alexander Graf 02 Feb '12

02 Feb '12

Hi guys, Great news! Kudos to Adrian for getting some of the OBS internals to work with new kiwi, we have our first ARM image built with kiwi: http://download.opensuse.org/repositories/openSUSE:/Factory:/ARM/images/Lim… It's even using the brand new "tbz" image type, which means that the tarball you see there actually contains another tarball which contains a rootfs to play around with. You can use that with your own kernels, with our kernels, as chroot on your android system. Anything you can imagine! :) There are still some things to straighten out. We have to inject qemu into the target system to be able to build, but we really don't want to keep it around after building is finished. So we have to figure something there :). Also, I'm currently trying to get efika and panda images rolling. So far the initrd creation is missing qemu again, but that should only be a matter of tweaking a few things. And by then we should have kiwi images available for the public! Just in time for FOSDEM! Yay! Alex -- To unsubscribe, e-mail: opensuse-arm+unsubscribe(a)opensuse.org To contact the owner, e-mail: opensuse-arm+owner(a)opensuse.org

2 4

[opensuse-arm] X on efika
by Peter Czanik 01 Feb '12

01 Feb '12

Hello, Using Bernhard's efika as a basis image I installed xfce and xorg packages on my machine. But when I try to start X, I get: " (EE) No devices detected. Fatal server error: no screens found " Does anybody have an xorg.conf to get started? Bye, CzP -- To unsubscribe, e-mail: opensuse-arm+unsubscribe(a)opensuse.org To contact the owner, e-mail: opensuse-arm+owner(a)opensuse.org

2 3

[opensuse-arm] In which devel project should a kernel reside?
by Joop Boonen 01 Feb '12

01 Feb '12

Hi all, I want to put the Toshiba AC100 kernel (kernel-tegra-ac100) in Factory. I sent a submit request to Base:Kernel but it looks like very few users are in this project. Is this the best devel project for this kernel? Or would it be better if another project would be created. Regards, Joop. -- To unsubscribe, e-mail: opensuse-arm+unsubscribe(a)opensuse.org To contact the owner, e-mail: opensuse-arm+owner(a)opensuse.org

4 10