Re: [OE-core] [PATCH][RFC] u-boot: Upgrade to upstream stable 2012.07
On 08/03/2012 07:19 PM, Wolfgang Denk wrote: Dear Radu Moisan, In message 1343997523-4117-1-git-send-email-radu.moi...@intel.com you wrote: Building u-boot requires UBOOT_MACHINE. In the u-boot README file building u-boot is achieved with make NAME_config and then make all. I assumend UBOOT_MACHINE to be the NAME part and thus, Actually a plain make NAME does the same as make NAME_config make all. You might also consider running ./MAKEALL NAME instead, which aut-adjusts to te number of available cores on the build host, so you get somewhat reduced build times. I've tried to build this way but got into some errors and did not pursued further in debugging them. However, because UBOOT_MACHINE is already used as NAME_config, would be difficult to change it to NAME and ask everyone who use it, to switch. I will go with current configuration make NAME_config make all radu ___ Openembedded-core mailing list Openembedded-core@lists.openembedded.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-core
[OE-core] [PATCH] u-boot: Upgrade to upstream stable 2012.07
For u-boot-fw-utils, fw_enc.c was changed and it requires a header file config.h which is autogenerated at config so it requires a make NAME_config. I used for testing coreboot-x86. Signed-off-by: Radu Moisan radu.moi...@intel.com --- ...ls_2012.04.01.bb = u-boot-fw-utils_2012.07.bb} |7 --- ...age_2012.04.01.bb = u-boot-mkimage_2012.07.bb} |6 +++--- .../{u-boot_2012.04.01.bb = u-boot_2012.07.bb}|8 3 files changed, 11 insertions(+), 10 deletions(-) rename meta/recipes-bsp/u-boot/{u-boot-fw-utils_2012.04.01.bb = u-boot-fw-utils_2012.07.bb} (83%) rename meta/recipes-bsp/u-boot/{u-boot-mkimage_2012.04.01.bb = u-boot-mkimage_2012.07.bb} (84%) rename meta/recipes-bsp/u-boot/{u-boot_2012.04.01.bb = u-boot_2012.07.bb} (82%) diff --git a/meta/recipes-bsp/u-boot/u-boot-fw-utils_2012.04.01.bb b/meta/recipes-bsp/u-boot/u-boot-fw-utils_2012.07.bb similarity index 83% rename from meta/recipes-bsp/u-boot/u-boot-fw-utils_2012.04.01.bb rename to meta/recipes-bsp/u-boot/u-boot-fw-utils_2012.07.bb index fe3422a..d2224bb 100644 --- a/meta/recipes-bsp/u-boot/u-boot-fw-utils_2012.04.01.bb +++ b/meta/recipes-bsp/u-boot/u-boot-fw-utils_2012.07.bb @@ -9,12 +9,12 @@ DEPENDS = mtd-utils # make it default DEFAULT_PREFERENCE = -1 -# This revision corresponds to the tag v2012.04.01 +# This revision corresponds to the tag v2012.07 # We use the revision in order to avoid having to fetch it from the # repo during parse -SRCREV = 415d386877df49eb051b85ef74fa59a16dc17c7d +SRCREV = 190649fb4309d1bc0fe7732fd0f951cb6440f935 -PV = v2012.04.01+git${SRCPV} +PV = 2012.07 SRC_URI = git://git.denx.de/u-boot.git;branch=master;protocol=git @@ -23,6 +23,7 @@ S = ${WORKDIR}/git EXTRA_OEMAKE = 'HOSTCC=${CC}' do_compile () { + oe_runmake ${UBOOT_MACHINE} oe_runmake env } diff --git a/meta/recipes-bsp/u-boot/u-boot-mkimage_2012.04.01.bb b/meta/recipes-bsp/u-boot/u-boot-mkimage_2012.07.bb similarity index 84% rename from meta/recipes-bsp/u-boot/u-boot-mkimage_2012.04.01.bb rename to meta/recipes-bsp/u-boot/u-boot-mkimage_2012.07.bb index aa107fe..752efcb 100644 --- a/meta/recipes-bsp/u-boot/u-boot-mkimage_2012.04.01.bb +++ b/meta/recipes-bsp/u-boot/u-boot-mkimage_2012.07.bb @@ -7,12 +7,12 @@ SECTION = bootloader # make it default DEFAULT_PREFERENCE = -1 -# This revision corresponds to the tag v2012.04.01 +# This revision corresponds to the tag v2012.07 # We use the revision in order to avoid having to fetch it from the # repo during parse -SRCREV = 415d386877df49eb051b85ef74fa59a16dc17c7d +SRCREV = 190649fb4309d1bc0fe7732fd0f951cb6440f935 -PV = v2012.04.01+git${SRCPV} +PV = 2012.07 SRC_URI = git://git.denx.de/u-boot.git;branch=master;protocol=git diff --git a/meta/recipes-bsp/u-boot/u-boot_2012.04.01.bb b/meta/recipes-bsp/u-boot/u-boot_2012.07.bb similarity index 82% rename from meta/recipes-bsp/u-boot/u-boot_2012.04.01.bb rename to meta/recipes-bsp/u-boot/u-boot_2012.07.bb index c4ec50d..7ccee30 100644 --- a/meta/recipes-bsp/u-boot/u-boot_2012.04.01.bb +++ b/meta/recipes-bsp/u-boot/u-boot_2012.07.bb @@ -14,13 +14,13 @@ DEFAULT_PREFERENCE = -1 LICENSE = GPLv2+ LIC_FILES_CHKSUM = file://COPYING;md5=1707d6db1d42237583f50183a5651ecb -# This revision corresponds to the tag v2012.04.01 +# This revision corresponds to the tag v2012.07 # We use the revision in order to avoid having to fetch it from the # repo during parse -SRCREV = 415d386877df49eb051b85ef74fa59a16dc17c7d +SRCREV = 190649fb4309d1bc0fe7732fd0f951cb6440f935 -PV = v2012.04.01+git${SRCPV} -PR = r1 +PV = 2012.07 +PR = r0 SRC_URI = git://git.denx.de/u-boot.git;branch=master;protocol=git -- 1.7.9.5 ___ Openembedded-core mailing list Openembedded-core@lists.openembedded.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-core
Re: [OE-core] [PATCH] u-boot: Upgrade to upstream stable 2012.07
On Mon, Aug 06, 2012 at 09:54:07AM +0300, Radu Moisan wrote: For u-boot-fw-utils, fw_enc.c was changed and it requires a header file config.h which is autogenerated at config so it requires a make NAME_config. I used for testing coreboot-x86. Is build with gold resolved in this new release already? By forcing bfd or to provide working binaries when built with gold? More info: http://lists.linuxtogo.org/pipermail/openembedded-core/2012-August/026943.html Cheers, Signed-off-by: Radu Moisan radu.moi...@intel.com --- ...ls_2012.04.01.bb = u-boot-fw-utils_2012.07.bb} |7 --- ...age_2012.04.01.bb = u-boot-mkimage_2012.07.bb} |6 +++--- .../{u-boot_2012.04.01.bb = u-boot_2012.07.bb}|8 3 files changed, 11 insertions(+), 10 deletions(-) rename meta/recipes-bsp/u-boot/{u-boot-fw-utils_2012.04.01.bb = u-boot-fw-utils_2012.07.bb} (83%) rename meta/recipes-bsp/u-boot/{u-boot-mkimage_2012.04.01.bb = u-boot-mkimage_2012.07.bb} (84%) rename meta/recipes-bsp/u-boot/{u-boot_2012.04.01.bb = u-boot_2012.07.bb} (82%) diff --git a/meta/recipes-bsp/u-boot/u-boot-fw-utils_2012.04.01.bb b/meta/recipes-bsp/u-boot/u-boot-fw-utils_2012.07.bb similarity index 83% rename from meta/recipes-bsp/u-boot/u-boot-fw-utils_2012.04.01.bb rename to meta/recipes-bsp/u-boot/u-boot-fw-utils_2012.07.bb index fe3422a..d2224bb 100644 --- a/meta/recipes-bsp/u-boot/u-boot-fw-utils_2012.04.01.bb +++ b/meta/recipes-bsp/u-boot/u-boot-fw-utils_2012.07.bb @@ -9,12 +9,12 @@ DEPENDS = mtd-utils # make it default DEFAULT_PREFERENCE = -1 -# This revision corresponds to the tag v2012.04.01 +# This revision corresponds to the tag v2012.07 # We use the revision in order to avoid having to fetch it from the # repo during parse -SRCREV = 415d386877df49eb051b85ef74fa59a16dc17c7d +SRCREV = 190649fb4309d1bc0fe7732fd0f951cb6440f935 -PV = v2012.04.01+git${SRCPV} +PV = 2012.07 SRC_URI = git://git.denx.de/u-boot.git;branch=master;protocol=git @@ -23,6 +23,7 @@ S = ${WORKDIR}/git EXTRA_OEMAKE = 'HOSTCC=${CC}' do_compile () { + oe_runmake ${UBOOT_MACHINE} oe_runmake env } diff --git a/meta/recipes-bsp/u-boot/u-boot-mkimage_2012.04.01.bb b/meta/recipes-bsp/u-boot/u-boot-mkimage_2012.07.bb similarity index 84% rename from meta/recipes-bsp/u-boot/u-boot-mkimage_2012.04.01.bb rename to meta/recipes-bsp/u-boot/u-boot-mkimage_2012.07.bb index aa107fe..752efcb 100644 --- a/meta/recipes-bsp/u-boot/u-boot-mkimage_2012.04.01.bb +++ b/meta/recipes-bsp/u-boot/u-boot-mkimage_2012.07.bb @@ -7,12 +7,12 @@ SECTION = bootloader # make it default DEFAULT_PREFERENCE = -1 -# This revision corresponds to the tag v2012.04.01 +# This revision corresponds to the tag v2012.07 # We use the revision in order to avoid having to fetch it from the # repo during parse -SRCREV = 415d386877df49eb051b85ef74fa59a16dc17c7d +SRCREV = 190649fb4309d1bc0fe7732fd0f951cb6440f935 -PV = v2012.04.01+git${SRCPV} +PV = 2012.07 SRC_URI = git://git.denx.de/u-boot.git;branch=master;protocol=git diff --git a/meta/recipes-bsp/u-boot/u-boot_2012.04.01.bb b/meta/recipes-bsp/u-boot/u-boot_2012.07.bb similarity index 82% rename from meta/recipes-bsp/u-boot/u-boot_2012.04.01.bb rename to meta/recipes-bsp/u-boot/u-boot_2012.07.bb index c4ec50d..7ccee30 100644 --- a/meta/recipes-bsp/u-boot/u-boot_2012.04.01.bb +++ b/meta/recipes-bsp/u-boot/u-boot_2012.07.bb @@ -14,13 +14,13 @@ DEFAULT_PREFERENCE = -1 LICENSE = GPLv2+ LIC_FILES_CHKSUM = file://COPYING;md5=1707d6db1d42237583f50183a5651ecb -# This revision corresponds to the tag v2012.04.01 +# This revision corresponds to the tag v2012.07 # We use the revision in order to avoid having to fetch it from the # repo during parse -SRCREV = 415d386877df49eb051b85ef74fa59a16dc17c7d +SRCREV = 190649fb4309d1bc0fe7732fd0f951cb6440f935 -PV = v2012.04.01+git${SRCPV} -PR = r1 +PV = 2012.07 +PR = r0 SRC_URI = git://git.denx.de/u-boot.git;branch=master;protocol=git -- 1.7.9.5 ___ Openembedded-core mailing list Openembedded-core@lists.openembedded.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-core -- Martin 'JaMa' Jansa jabber: martin.ja...@gmail.com signature.asc Description: Digital signature ___ Openembedded-core mailing list Openembedded-core@lists.openembedded.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-core
Re: [OE-core] [PATCH 1/1] gdk-pixbuf: fix parallel install issue
On 08/06/2012 06:32 AM, Colin Walters wrote: On Fri, 2012-08-03 at 11:30 +0800, wenzong@windriver.com wrote: Make an explicit dependency to the libs install targets would fix this issue. I don't think Yocto should be running the 'make install' target with parallelism enabled; it's just going to be a source of major pain for small gain. Historically Automake's install targets have had a lot of paralleism issues AIUI, even discarding modules which have custom install hooks. Hi Richard, What's your opinions? Disable/Enable paralleism for 'make install', this more like a policy issue rather than tech. Thanks Wenzong ___ Openembedded-core mailing list Openembedded-core@lists.openembedded.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-core ___ Openembedded-core mailing list Openembedded-core@lists.openembedded.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-core
[OE-core] [PATCH 0/1] tcl: Add ${bindir_crossscripts}/tclConfig.sh to sysroot stage
From: Jackie Huang jackie.hu...@windriver.com tclConfig.sh is changed in do_install for cross compile and is installed to STAGING_BINDIR_CROSS, but if SSTATE_DIR is set and tcl is from sstage, tclConfig.sh can't be found in STAGING_BINDIR_CROSS, add ${bindir_crossscripts}/tclConfig.sh to sysroot stage can fix it. [YOCTO #2891] Signed-off-by: Jackie Huang jackie.hu...@windriver.com --- The following changes since commit a9d0cbe1d84bb26fc1a1f48764fe514cf9f9c548: gcc: Bump PR since there have been several gcc changes and various problems reported and this should flush anything stale out (2012-08-03 10:32:24 +0100) are available in the git repository at: git://git.pokylinux.org/poky-contrib jhuang0/bug2891_tcl_0806 http://git.pokylinux.org/cgit.cgi/poky-contrib/log/?h=jhuang0/bug2891_tcl_0806 Jackie Huang (1): tcl: Add ${bindir_crossscripts}/tclConfig.sh to sysroot stage meta/recipes-devtools/tcltk/tcl_8.5.11.bb |9 +++-- 1 files changed, 7 insertions(+), 2 deletions(-) -- 1.7.4 ___ Openembedded-core mailing list Openembedded-core@lists.openembedded.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-core
[OE-core] [PATCH 1/1] tcl: Add ${bindir_crossscripts}/tclConfig.sh to sysroot stage
From: Jackie Huang jackie.hu...@windriver.com tclConfig.sh is changed in do_install for cross compile and is installed to STAGING_BINDIR_CROSS, but if SSTATE_DIR is set and tcl is from sstage, tclConfig.sh can't be found in STAGING_BINDIR_CROSS, add ${bindir_crossscripts}/tclConfig.sh to sysroot stage can fix it. [YOCTO #2891] Signed-off-by: Jackie Huang jackie.hu...@windriver.com --- meta/recipes-devtools/tcltk/tcl_8.5.11.bb |9 +++-- 1 files changed, 7 insertions(+), 2 deletions(-) diff --git a/meta/recipes-devtools/tcltk/tcl_8.5.11.bb b/meta/recipes-devtools/tcltk/tcl_8.5.11.bb index d5cf6dc..f19e25a 100644 --- a/meta/recipes-devtools/tcltk/tcl_8.5.11.bb +++ b/meta/recipes-devtools/tcltk/tcl_8.5.11.bb @@ -49,8 +49,8 @@ do_install() { #sed -i s+${WORKDIR}+${STAGING_INCDIR}+g tclConfig.sh sed -i s,-L${libdir},-L=${libdir},g tclConfig.sh sed -i s,-I${includedir},-I=${includedir},g tclConfig.sh - install -d ${STAGING_BINDIR_CROSS}/ - install -m 0755 tclConfig.sh ${STAGING_BINDIR_CROSS} + install -d ${D}${bindir_crossscripts} + install -m 0755 tclConfig.sh ${D}${bindir_crossscripts} cd .. for dir in compat generic unix do @@ -59,6 +59,11 @@ do_install() { done } +SYSROOT_PREPROCESS_FUNCS += tcl_sysroot_preprocess +tcl_sysroot_preprocess () { + sysroot_stage_dir ${D}${bindir_crossscripts} ${SYSROOT_DESTDIR}${bindir_crossscripts} +} + PACKAGES =+ tcl-lib FILES_tcl-lib = ${libdir}/libtcl8.5.so* FILES_${PN} += ${prefix}/lib/tcl8.5 ${prefix}/lib/tcl8 -- 1.7.4 ___ Openembedded-core mailing list Openembedded-core@lists.openembedded.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-core
Re: [OE-core] [oe] [PATCH] sstate: Add a two character subdirectory to the sstate directory layout
On Thu, Aug 02, 2012 at 08:57:50PM +0100, Richard Purdie wrote: On Thu, 2012-08-02 at 21:40 +0200, Martin Jansa wrote: On Thu, Aug 02, 2012 at 04:53:12PM +0100, Richard Purdie wrote: On Thu, 2012-08-02 at 16:14 +0200, Martin Jansa wrote: On Thu, Aug 02, 2012 at 03:53:35PM +0200, Martin Jansa wrote: On Wed, Jul 25, 2012 at 10:09:22PM +0100, Richard Purdie wrote: Currently all sstate files are placed into one directory. This does not scale and causes a variety of filesystem issues. This patch adds a two character subdirectory to the layout (based on the first two characters of the hash) so that files can be split into several directories. This should help performance of sstate in most cases by avoding creating directories with huge numbers of files. The SSTATE_MIRRORS syntax needs updating to account for the extra path element by the addition of a PATH item, for example: SSTATE_MIRRORS = file://.* file:///some/path/to/sstate-cache/PATH SSTATE_MIRRORS = file://.* http://192.168.1.23/sstate-cache/PATH; This change also sets the scene for using things like lsb-release in the Is it possible to create 2nd level cache with this? I have some server with slow upload but fully populated sstate-cache. So on server with faster upload which could be used as offical SSTATE_MIRROR for SHR distro I would like to add SSTATE_MIRRORS ?= file://.* http://slow-server/sstate-cache/PATH; And then sync my sstate-cache directory to public accessible web root (with rsync). Problem is that now sstate-cache has all files in slightly different layout then original sstate-cache on slow server. From what I see I guess it finds URL with correct prefix sstate-cache/Gentoo-2.1/0d and downloads it directly to sstate-cache dir (and adds .done) OE @ ~/oe-core $ ll sstate-cache/sstate-apr-native-x86_64-linux-1.4.6-r1-x86_64-2-*populate-lic* -rw-r--r-- 1 bitbake bitbake 9257 Jul 30 12:31 sstate-cache/sstate-apr-native-x86_64-linux-1.4.6-r1-x86_64-2-0d2ed24b90d50bf83e5fe94536596e50_populate-lic.tgz -rw-r--r-- 1 bitbake bitbake0 Aug 2 15:40 sstate-cache/sstate-apr-native-x86_64-linux-1.4.6-r1-x86_64-2-0d2ed24b90d50bf83e5fe94536596e50_populate-lic.tgz.done And then creates symlink in right prefix back to absolute path of sstate-cache/file: OE @ ~/oe-core $ ll sstate-cache/Gentoo-2.1/0d/sstate-apr-native-x86_64-linux-1.4.6-r1-x86_64-2-*populate-lic* lrwxrwxrwx 1 bitbake bitbake 123 Aug 2 15:40 sstate-cache/Gentoo-2.1/0d/sstate-apr-native-x86_64-linux-1.4.6-r1-x86_64-2-0d2ed24b90d50bf83e5fe94536596e50_populate-lic.tgz - /OE/oe-core/sstate-cache/sstate-apr-native-x86_64-linux-1.4.6-r1-x86_64-2-0d2ed24b90d50bf83e5fe94536596e50_populate-lic.tgz But after sstate-cache directory is rsynced somewhere else and oe-core/sstate-cache is removed, all those symlinks point nowhere and public sstate-cache is unusable. Can we have relative paths used in symlinks or even instruct fetcher to download that file directly to right prefix? 2 more ideas: 1) would be great to also download file.sigdata if it exists, to be able to compare them when they change even on machine which downloaded older sstate file from remote url 2) if the reason for this patch was number of files in shared sstate-cache directory, then fetcher creating .done files makes number double too (would be fine if fetcher stores all 3 files (.tgz, .tgz.sigdata, .tgz.done) in right prefix, or moves them to right prefix instead of symlinks. I'm aware of the problem. The main issue is that we probably need to And what about .sigdata files? Added https://bugzilla.yoctoproject.org/show_bug.cgi?id=2898 I have sort shell script to replace symlinks with real files in prefixed dirs, would it be worth it integrating to openembedded-core/scripts/sstate-cache-management.sh which doesn't work with new layout anyway? Added https://bugzilla.yoctoproject.org/show_bug.cgi?id=2897 start enforcing complete paths for all downloads in DL_DIR, including http:// urls. This would resolve conflicts like: SRC_URI = http://server1.org/somefile.patch \ http://server2.org/somefile.patch; In two separate recipes right? where the two files are different. The trouble is it will pretty much break all the source mirrors :(. So you would store them in DL_DIR/server1.org/somefile.patch path? I've wondered about: DL_DIR/server1.org/somepath/somefile.patch That would make oposite scenario where the BIG.tgz is available (or even requested by different recipes) from different location less efficient.
Re: [OE-core] [PATCH] u-boot: Upgrade to upstream stable 2012.07
On 08/06/2012 09:55 AM, Martin Jansa wrote: On Mon, Aug 06, 2012 at 09:54:07AM +0300, Radu Moisan wrote: For u-boot-fw-utils, fw_enc.c was changed and it requires a header file config.h which is autogenerated at config so it requires a make NAME_config. I used for testing coreboot-x86. Is build with gold resolved in this new release already? By forcing bfd or to provide working binaries when built with gold? More info: http://lists.linuxtogo.org/pipermail/openembedded-core/2012-August/026943.html I don't know what linker is used, I guess the default one. If you can point out to me how to check, I'll be glad to do it. The patch in your link was not in my tree at the time I've tested, if that is relevant in any way. thanks, radu ___ Openembedded-core mailing list Openembedded-core@lists.openembedded.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-core
Re: [OE-core] [PATCH] u-boot: Upgrade to upstream stable 2012.07
On Mon, Aug 06, 2012 at 10:12:20AM +0300, Radu Moisan wrote: On 08/06/2012 09:55 AM, Martin Jansa wrote: On Mon, Aug 06, 2012 at 09:54:07AM +0300, Radu Moisan wrote: For u-boot-fw-utils, fw_enc.c was changed and it requires a header file config.h which is autogenerated at config so it requires a make NAME_config. I used for testing coreboot-x86. Is build with gold resolved in this new release already? By forcing bfd or to provide working binaries when built with gold? More info: http://lists.linuxtogo.org/pipermail/openembedded-core/2012-August/026943.html I don't know what linker is used, I guess the default one. If you can point out to me how to check, I'll be glad to do it. The patch in your link was not in my tree at the time I've tested, if that is relevant in any way. Maybe you were looking to poky repo instead of oe-core? Find poky equivalent of http://git.openembedded.org/openembedded-core/commit/?id=2e79fcd673dadeab6358fe22d47c8534c14de03e The issue reported in that email was fixed http://git.openembedded.org/openembedded-core/commit/?id=f78044f85ab1a0acce852a7032fc0c81285cd4c1 What I was intereseted is whether upstream added similar patch to http://git.shr-project.org/git/?p=meta-smartphone.git;a=blob;f=meta-nokia/recipes-bsp/u-boot/u-boot/0001-config-Always-use-GNU-ld.patch or added other fix to make gold usable for u-boot. And you can test build with gold if you enable gold by adding ld-is-gold to DISTRO_FEATURES. Cheers, -- Martin 'JaMa' Jansa jabber: martin.ja...@gmail.com signature.asc Description: Digital signature ___ Openembedded-core mailing list Openembedded-core@lists.openembedded.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-core
Re: [OE-core] [PATCH][RFC] u-boot: Upgrade to upstream stable 2012.07
Dear Radu, In message 501f63ff.1070...@intel.com you wrote: Actually a plain make NAME does the same as make NAME_config make all. You might also consider running ./MAKEALL NAME instead, which aut-adjusts to te number of available cores on the build host, so you get somewhat reduced build times. I've tried to build this way but got into some errors and did not pursued further in debugging them. However, because UBOOT_MACHINE is Could you please tell me exactly what these errors were? Log files etc. highly appreciated... Best regards, Wolfgang Denk -- DENX Software Engineering GmbH, MD: Wolfgang Denk Detlev Zundel HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany Phone: (+49)-8142-66989-10 Fax: (+49)-8142-66989-80 Email: w...@denx.de If programming was easy, they wouldn't need something as complicated as a human being to do it, now would they? - L. Wall R. L. Schwartz, _Programming Perl_ ___ Openembedded-core mailing list Openembedded-core@lists.openembedded.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-core
Re: [OE-core] [PATCH 4/5] gtk-icon-cache: call postinst scriplet at do_rootfs time
On 08/06/2012 01:49 AM, Andreas Müller wrote: On Mon, Aug 6, 2012 at 12:30 AM, Andreas Müller schnitzelt...@googlemail.com wrote: On Sun, Aug 5, 2012 at 6:58 PM, Laurentiu Palcu laurentiu.pa...@intel.com wrote: You could give it a test yourselves and let me know your results. I will send a version 2 of the patchset(as soon as we all agree on the solution), with some changes suggested by Mark and some PR bumps suggested by Koen. With the image I usually work with [1] and AMD Phenom II X6 1090 16GB RAM I get a measurable delay - see attachment. I would not be happy loosing latest do_rootfs enhancements (off topic - thanks for that). Remeber we are only talking of gtk-update-icon-cache. OK I could buy an intel host and work just with sato images but... I suppose you could, but nobody asked you to do that, it's your choice what's your build machine or what you'll be building for. Thanks for the measurements though. They do, indeed, show quite a significant amount of time (around 6 minutes). A run-once solution is to be considered in this case. Andreas [1] https://gitorious.org/schnitzeltony-oe-meta/meta-misc/blobs/master/recipes-image/xfce-full-image.bb OK I know it is not that important: The image created with this patch series creates tons of messages like Why do you think is not important. Please elaborate. Or is it irony? I don't think is in anybody's benefit if you take this approach. :) All errors/warnings are important and they have to be taken care of. xfdesktop:446): GdkPixbuf-WARNING **: Cannot open pixbuf loader module file '/usr/lib/gdk-pixbuf-2.0/2.10.0/loaders.cache': No such file or directory and don't have icons at all. Did you run test that (on a hardware plattform different to your host)? I only tested on qemu. And it worked just fine, without any errors. With all icons in place. Andreas ___ Openembedded-core mailing list Openembedded-core@lists.openembedded.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-core ___ Openembedded-core mailing list Openembedded-core@lists.openembedded.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-core
Re: [OE-core] [PATCH 4/5] gtk-icon-cache: call postinst scriplet at do_rootfs time
On Mon, Aug 6, 2012 at 9:48 AM, Laurentiu Palcu laurentiu.pa...@intel.com wrote: On 08/06/2012 01:49 AM, Andreas Müller wrote: On Mon, Aug 6, 2012 at 12:30 AM, Andreas Müller schnitzelt...@googlemail.com wrote: On Sun, Aug 5, 2012 at 6:58 PM, Laurentiu Palcu laurentiu.pa...@intel.com wrote: You could give it a test yourselves and let me know your results. I will send a version 2 of the patchset(as soon as we all agree on the solution), with some changes suggested by Mark and some PR bumps suggested by Koen. With the image I usually work with [1] and AMD Phenom II X6 1090 16GB RAM I get a measurable delay - see attachment. I would not be happy loosing latest do_rootfs enhancements (off topic - thanks for that). Remeber we are only talking of gtk-update-icon-cache. OK I could buy an intel host and work just with sato images but... I suppose you could, but nobody asked you to do that, it's your choice what's your build machine or what you'll be building for. Thanks for the measurements though. They do, indeed, show quite a significant amount of time (around 6 minutes). A run-once solution is to be considered in this case. Andreas [1] https://gitorious.org/schnitzeltony-oe-meta/meta-misc/blobs/master/recipes-image/xfce-full-image.bb OK I know it is not that important: The image created with this patch series creates tons of messages like Why do you think is not important. Please elaborate. Or is it irony? Yes sorry - it was late night. I don't think is in anybody's benefit if you take this approach. :) All errors/warnings are important and they have to be taken care of. xfdesktop:446): GdkPixbuf-WARNING **: Cannot open pixbuf loader module file '/usr/lib/gdk-pixbuf-2.0/2.10.0/loaders.cache': No such file or directory and don't have icons at all. Did you run test that (on a hardware plattform different to your host)? I only tested on qemu. And it worked just fine, without any errors. With all icons in place. Which hardware did you emulate? My tests were done for overo (ARM cortext A8). I wonder if the created database is machine specific. Andreas ___ Openembedded-core mailing list Openembedded-core@lists.openembedded.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-core
Re: [OE-core] [PATCH 4/5] gtk-icon-cache: call postinst scriplet at do_rootfs time
On 08/06/2012 11:10 AM, Andreas Müller wrote: On Mon, Aug 6, 2012 at 9:48 AM, Laurentiu Palcu laurentiu.pa...@intel.com wrote: On 08/06/2012 01:49 AM, Andreas Müller wrote: On Mon, Aug 6, 2012 at 12:30 AM, Andreas Müller schnitzelt...@googlemail.com wrote: On Sun, Aug 5, 2012 at 6:58 PM, Laurentiu Palcu laurentiu.pa...@intel.com wrote: You could give it a test yourselves and let me know your results. I will send a version 2 of the patchset(as soon as we all agree on the solution), with some changes suggested by Mark and some PR bumps suggested by Koen. With the image I usually work with [1] and AMD Phenom II X6 1090 16GB RAM I get a measurable delay - see attachment. I would not be happy loosing latest do_rootfs enhancements (off topic - thanks for that). Remeber we are only talking of gtk-update-icon-cache. OK I could buy an intel host and work just with sato images but... I suppose you could, but nobody asked you to do that, it's your choice what's your build machine or what you'll be building for. Thanks for the measurements though. They do, indeed, show quite a significant amount of time (around 6 minutes). A run-once solution is to be considered in this case. Andreas [1] https://gitorious.org/schnitzeltony-oe-meta/meta-misc/blobs/master/recipes-image/xfce-full-image.bb OK I know it is not that important: The image created with this patch series creates tons of messages like Why do you think is not important. Please elaborate. Or is it irony? Yes sorry - it was late night. It's OK, let's work together and find the best solution for all of us. I would also be pissed of if I had to wait an extra 6 minutes every do_rootfs run. I don't think is in anybody's benefit if you take this approach. :) All errors/warnings are important and they have to be taken care of. xfdesktop:446): GdkPixbuf-WARNING **: Cannot open pixbuf loader module file '/usr/lib/gdk-pixbuf-2.0/2.10.0/loaders.cache': No such file or directory and don't have icons at all. Did you run test that (on a hardware plattform different to your host)? I only tested on qemu. And it worked just fine, without any errors. With all icons in place. Which hardware did you emulate? My tests were done for overo (ARM cortext A8). I wonder if the created database is machine specific. I used qemuarm. As far as I know, the database shouldn't be machine specific. And, looking at the log you sent, it looks like all commands succeeded. Otherwise, the 'time' application would also signal in the log file if the command failed. Could please have a look, on your host (xfce-full-image-1.0-r0/rootfs/usr/share/icons/hicolor or xfce-full-image-1.0-r0/rootfs/usr/share/icons/gnome), to see if it creates the icon-theme.cache files? It does for me and I don't understand why it doesn't in your case... Thanks, Laurentiu Andreas ___ Openembedded-core mailing list Openembedded-core@lists.openembedded.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-core ___ Openembedded-core mailing list Openembedded-core@lists.openembedded.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-core
Re: [OE-core] [PATCH 1/1] package-index: inherit pythonnative
On Tue, Jul 24, 2012 at 11:49:52AM +0100, Richard Purdie wrote: On Tue, 2012-07-24 at 18:19 +0800, Robert Yang wrote: The native python binary has been moved from usr/bin/python to usr/bin/python-native/python, the recipe which needs python-native should inherit pythonnative, otherwise there would be errors when the python script runs. [YOCTO #2822] Signed-off-by: Robert Yang liezhi.y...@windriver.com --- meta/recipes-core/meta/package-index.bb |2 ++ 1 files changed, 2 insertions(+), 0 deletions(-) Which part of this recipe needs python-native? Shouldn't scripts which need pythonnative be using the path to the python interpreter explicitly? This fixes my opkg-utils related issues too http://lists.linuxtogo.org/pipermail/openembedded-core/2012-July/026333.html The problem is that with pythonnative in opkg-utils it's still using #!/usr/bin/env python in tmp-eglibc/sysroots/x86_64-linux/usr/bin/opkg-make-index And when python-index executes this: | + [ -e /var/lib/jenkins/jobs/shr-core/workspace/shr-core/tmp-eglibc/deploy/ipk/ ] | + touch /var/lib/jenkins/jobs/shr-core/workspace/shr-core/tmp-eglibc/deploy/ipk/Packages | + flock /var/lib/jenkins/jobs/shr-core/workspace/shr-core/tmp-eglibc/deploy/ipk/Packages.flock -c opkg-make-index -r /var/lib/jenkins/jobs/shr-core/workspace/shr-core/tmp-eglibc/deploy/ipk/Packages -p /var/lib/jenkins/jobs/shr-core/workspace/shr-core/tmp-eglibc/deploy/ipk/Packages -m /var/lib/jenkins/jobs/shr-core/workspace/shr-core/tmp-eglibc/deploy/ipk/ | Traceback (most recent call last): It depends on PATH of package-index not opkg-utils. Cheers, -- Martin 'JaMa' Jansa jabber: martin.ja...@gmail.com signature.asc Description: Digital signature ___ Openembedded-core mailing list Openembedded-core@lists.openembedded.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-core
Re: [OE-core] [PATCH 2/5] sato-icon-theme: make postinst scriplet run at do_rootfs time
On 3 August 2012 21:19, Laurentiu Palcu laurentiu.pa...@intel.com wrote: pkg_postinst_${PN} () { -if [ x$D != x ]; then -exit 1 +gtk-update-icon-cache -q $D/usr/share/icons/Sato +if [ ! -d $D/etc/gtk-2.0 ]; then +mkdir -p $D/etc/gtk-2.0 fi -gtk-update-icon-cache -q /usr/share/icons/Sato -echo 'gtk-icon-theme-name = Sato' /etc/gtk-2.0/gtkrc +echo 'gtk-icon-theme-name = Sato' $D/etc/gtk-2.0/gtkrc } Surely sato-icon-theme should be inheriting the gtk-icon-cache class to remove all of the duplicated logic? Ross ___ Openembedded-core mailing list Openembedded-core@lists.openembedded.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-core
Re: [OE-core] [PATCH] opkg-utils: inherit pythonnative
On Sun, Aug 05, 2012 at 12:24:12PM +0200, Martin Jansa wrote: Signed-off-by: Martin Jansa martin.ja...@gmail.com --- meta/recipes-devtools/opkg-utils/opkg-utils_git.bb |4 +++- 1 files changed, 3 insertions(+), 1 deletions(-) diff --git a/meta/recipes-devtools/opkg-utils/opkg-utils_git.bb b/meta/recipes-devtools/opkg-utils/opkg-utils_git.bb index 92e6624..caa66f7 100644 --- a/meta/recipes-devtools/opkg-utils/opkg-utils_git.bb +++ b/meta/recipes-devtools/opkg-utils/opkg-utils_git.bb @@ -8,7 +8,9 @@ RDEPENDS_${PN} = python RDEPENDS_${PN}_virtclass-native = SRCREV = 49cc783d8e0415059d126ae22c892988717ffda7 PV = 0.1.8+git${SRCPV} -PR = r0 +PR = r1 + +inherit pythonnative SRC_URI = git://git.yoctoproject.org/opkg-utils;protocol=git \ Please ignore this patch, this needs to be in package-index recipe (and I had both patches when testing this :/) Cheers, -- Martin 'JaMa' Jansa jabber: martin.ja...@gmail.com signature.asc Description: Digital signature ___ Openembedded-core mailing list Openembedded-core@lists.openembedded.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-core
Re: [OE-core] [PATCH 4/5] gtk-icon-cache: call postinst scriplet at do_rootfs time
On 6 August 2012 10:18, Laurentiu Palcu laurentiu.pa...@intel.com wrote: Which hardware did you emulate? My tests were done for overo (ARM cortext A8). I wonder if the created database is machine specific. I used qemuarm. As far as I know, the database shouldn't be machine specific. Looking at updateiconcache.c, all of the functions that write data to the cache file use write_card16, write_card32 and so on which convert to big-endian. Unless there's a bug hidden away, the cache files are cross-platform. Ross ___ Openembedded-core mailing list Openembedded-core@lists.openembedded.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-core
Re: [OE-core] [PATCH 4/5] gtk-icon-cache: call postinst scriplet at do_rootfs time
On Mon, Aug 6, 2012 at 11:18 AM, Laurentiu Palcu laurentiu.pa...@intel.com wrote: On 08/06/2012 11:10 AM, Andreas Müller wrote: On Mon, Aug 6, 2012 at 9:48 AM, Laurentiu Palcu laurentiu.pa...@intel.com wrote: On 08/06/2012 01:49 AM, Andreas Müller wrote: On Mon, Aug 6, 2012 at 12:30 AM, Andreas Müller schnitzelt...@googlemail.com wrote: On Sun, Aug 5, 2012 at 6:58 PM, Laurentiu Palcu laurentiu.pa...@intel.com wrote: You could give it a test yourselves and let me know your results. I will send a version 2 of the patchset(as soon as we all agree on the solution), with some changes suggested by Mark and some PR bumps suggested by Koen. With the image I usually work with [1] and AMD Phenom II X6 1090 16GB RAM I get a measurable delay - see attachment. I would not be happy loosing latest do_rootfs enhancements (off topic - thanks for that). Remeber we are only talking of gtk-update-icon-cache. OK I could buy an intel host and work just with sato images but... I suppose you could, but nobody asked you to do that, it's your choice what's your build machine or what you'll be building for. Thanks for the measurements though. They do, indeed, show quite a significant amount of time (around 6 minutes). A run-once solution is to be considered in this case. Andreas [1] https://gitorious.org/schnitzeltony-oe-meta/meta-misc/blobs/master/recipes-image/xfce-full-image.bb OK I know it is not that important: The image created with this patch series creates tons of messages like Why do you think is not important. Please elaborate. Or is it irony? Yes sorry - it was late night. It's OK, let's work together and find the best solution for all of us. I would also be pissed of if I had to wait an extra 6 minutes every do_rootfs run. I don't think is in anybody's benefit if you take this approach. :) All errors/warnings are important and they have to be taken care of. xfdesktop:446): GdkPixbuf-WARNING **: Cannot open pixbuf loader module file '/usr/lib/gdk-pixbuf-2.0/2.10.0/loaders.cache': No such file or directory and don't have icons at all. Did you run test that (on a hardware plattform different to your host)? I only tested on qemu. And it worked just fine, without any errors. With all icons in place. Which hardware did you emulate? My tests were done for overo (ARM cortext A8). I wonder if the created database is machine specific. I used qemuarm. As far as I know, the database shouldn't be machine specific. And, looking at the log you sent, it looks like all commands succeeded. Otherwise, the 'time' application would also signal in the log file if the command failed. Could please have a look, on your host (xfce-full-image-1.0-r0/rootfs/usr/share/icons/hicolor or xfce-full-image-1.0-r0/rootfs/usr/share/icons/gnome), to see if it creates the icon-theme.cache files? It does for me and I don't understand why it doesn't in your case... I will check today evening since the machine with this data is at home... Andreas ___ Openembedded-core mailing list Openembedded-core@lists.openembedded.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-core
[OE-core] [PATCH] gcc-cross-initial: Ensure it uses an isolated sysroot
If we don't do this, a stale limits.h may be detected in STAGING_DIR_TARGET which would result in a different limits.h getting generated by gcc-cross-initial that references it. The referenced limits.h will then not get found by eglibc-initial causing rather strange build failures. The simplest solution is to create a temporary sysroot containing only the things gcc-cross-initial should care about and this results in a correct limits.h file regardless of what else may have been built. Signed-off-by: Richard Purdie richard.pur...@linuxfoundation.org --- diff --git a/meta/recipes-devtools/gcc/gcc-cross-initial.inc b/meta/recipes-devtools/gcc/gcc-cross-initial.inc index a515fb0..543a94a 100644 --- a/meta/recipes-devtools/gcc/gcc-cross-initial.inc +++ b/meta/recipes-devtools/gcc/gcc-cross-initial.inc @@ -19,11 +19,23 @@ EXTRA_OECONF = --with-newlib \ ${OPTSPACE} \ --program-prefix=${TARGET_PREFIX} \ --with-sysroot=${STAGING_DIR_TARGET} \ - --with-build-sysroot=${STAGING_DIR_TARGET} \ + --with-build-sysroot=${GCCCROSS_BUILDSYSROOT} \ ${EXTRA_OECONF_INITIAL} \ ${@base_contains('DISTRO_FEATURES', 'ld-is-gold', '--with-ld=${STAGING_BINDIR_TOOLCHAIN}/${TARGET_PREFIX}ld.bfd', '', d)} \ ${EXTRA_OECONF_FPU} + +GCCCROSS_BUILDSYSROOT = ${B}/tmpsysroot + +do_configure_prepend () { + sysr=${GCCCROSS_BUILDSYSROOT}${target_includedir} + mkdir -p $sysr + for t in linux asm asm-generic; do + rm -f $sysr/$t + ln -s ${STAGING_DIR_TARGET}${target_includedir}/$t $sysr/ + done +} + do_compile () { oe_runmake all-gcc all-target-libgcc } ___ Openembedded-core mailing list Openembedded-core@lists.openembedded.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-core
[OE-core] [PATCH] curl: disable ldap/ldaps explicitly
* openldap from meta-oe is autodetected and then libldap-2.4-2 runtime dependency added to curl and almost all meta-efl recipes Signed-off-by: Martin Jansa martin.ja...@gmail.com --- meta/recipes-support/curl/curl_7.26.0.bb | 12 +++- 1 files changed, 7 insertions(+), 5 deletions(-) diff --git a/meta/recipes-support/curl/curl_7.26.0.bb b/meta/recipes-support/curl/curl_7.26.0.bb index f04b400..418e29c 100644 --- a/meta/recipes-support/curl/curl_7.26.0.bb +++ b/meta/recipes-support/curl/curl_7.26.0.bb @@ -8,7 +8,7 @@ LIC_FILES_CHKSUM = file://COPYING;beginline=7;md5=3a34942f4ae3fbf1a303160714e66 DEPENDS = zlib gnutls DEPENDS_virtclass-native = zlib-native openssl-native DEPENDS_virtclass-nativesdk = zlib-nativesdk -PR = r0 +PR = r1 SRC_URI = http://curl.haxx.se/download/curl-${PV}.tar.bz2 \ file://pkgconfig_fix.patch @@ -20,11 +20,13 @@ inherit autotools pkgconfig binconfig EXTRA_OECONF = --with-zlib=${STAGING_LIBDIR}/../ \ --without-libssh2 \ - --with-random=/dev/urandom \ - --without-libidn \ - --enable-crypto-auth \ +--with-random=/dev/urandom \ +--without-libidn \ +--enable-crypto-auth \ +--disable-ldap \ +--disable-ldaps \ ${CURLGNUTLS} \ - + CURLGNUTLS = --with-gnutls=${STAGING_LIBDIR}/../ --without-ssl CURLGNUTLS_virtclass-native = --without-gnutls --with-ssl -- 1.7.8.6 ___ Openembedded-core mailing list Openembedded-core@lists.openembedded.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-core
Re: [OE-core] [PATCH V3 00/11] Mesa upgrade/improvements
On Fri, 2012-08-03 at 17:46 +0200, Koen Kooi wrote: Op 3 aug. 2012, om 15:19 heeft Ross Burton ross.bur...@intel.com het volgende geschreven: On Friday, 3 August 2012 at 11:18, Koen Kooi wrote: libegl and libgles aren't built nowadays, so the problem is avoided. Noone has dared to touch this subject the past 2.5 years: http://cgit.openembedded.org/openembedded/commit/recipes/mesa?id=3d96f8cb61225d515b5cb4fe863f0d50c3ced436 The best solution[1] is to disable egl/gles in the mesa-recipes and add a seperate recipe for them. That way you still get glu and other useful mesa bits needed for gnome/efl/xfce/etc, but you can skip the sysroot/shlib poisoning if needed. After much digging I see that the Beagle PVR drivers only provide gles and egl, which is why you say that the problem was avoided. Of course that's just one specific driver, for example the Cedar Trail EMGD driver does build libgl, libgles and libegl. If the solution is put the egl/gles bits in another package Recipe, not package. Big difference :) then I'm totally confused as to what the actual problem is. Considering there's numerous libegl libraries for the many variants of PVR on ARM, I'm struggling to understand exactly what is new here. Can you explain clearly what the problem is, I'm obviously missing something. Once I understand the problem I can help with a solution. You need mesa for the !egl and !gles libs and evil vendor blob for egl and gles libs. In the emgd case you can get gl from the evil vendor as well. To build e.g. EFL I need mesa for the !egl and !gles bits, to build EFL with gles support I need mesa (again !egl, !gles) in addition to evil vendor blob for egl and gles. We could make this work with MACHINE_FEATURES or something similar, but that would mean that a very large chunk of userspace now becomes machine specific. This is how clutter worked in the past and just look at the fit the canonical/linaro people threw when they started doing 'embedded' and clutter. To compress the above: you need to build both the mesa and the evil-egl-binary *recipes* for things to work, so the egl bits should be a different recipe, not a subpackage of the mesa recipe. This is all assuming we care abou the binary blobs e.g. TI distributes for 3d support. If you don't, fine, merge this patchset. Let me test this a little bit more. The binary blobs shipped by TI would be machine specific and would be marked as such in the package feeds? The mesa pieces would be marked as architecture specific and more generic. In feeds, machine specific has priority so on the machines where there is hardware support, the hardware optimised versions would get installed? If the above isn't the case, can we make it work like that? Cheers, Richard ___ Openembedded-core mailing list Openembedded-core@lists.openembedded.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-core
[OE-core] Could we build tar-replacement firstly and not parallel if tar-replacement is needed to build
Hi: Building failed several times when building tar-replacement-native is needed, The error log shows bitbake_build/tmp/sysroots/x86_64-linux/usr/bin/tar: Text file busy. this log proves an attempt was made to execute a pure-procedure program that is currently open for writing. Before this error, tar-replacement do_populate_sysroot just starts, (tar-replacement-native-1.26-r2: task do_populate_sysroot: Started) After this error, we see tar-replacement do_populate_sysroot just finish. So I think the root cause is that do_populate_sysroot sqlite is using a tar command which being opened to written. in fact, all packages need tar to do_popolate_sysroot() I see building tar-replacement-native is started in bitbake file, could we build tar-replacement firstly, not parallel if needed. diff --git a/scripts/bitbake b/scripts/bitbake index 3772d82..1bb8fed 100755 --- a/scripts/bitbake +++ b/scripts/bitbake @@ -134,7 +134,10 @@ if [ $buildpseudo -gt 0 ]; then fi done done -bitbake pseudo-native $TARTARGET $additionalopts -c populate_sysroot + +if [ $needptar = 1 ]; then + sed -i 's/BB_NUMBER_THREADS =/#BB_NUMBER_THREADS =/g' conf/local.conf + bitbake $TARTARGET + sed -i 's/#BB_NUMBER_THREADS =/BB_NUMBER_THREADS =/g' conf/local.conf +fi + +bitbake pseudo-native $additionalopts -c populate_sysroot ret=$? if [ $ret != 0 ]; then exit 1 since building tar-replacement or not depends on host tar version, it is hard to add a tar dependence to any package, and I did not have a good way to fix this bug except the upper. Any advice and suggestion are welcome -- Best Reagrds, Roy | RongQing Li ___ Openembedded-core mailing list Openembedded-core@lists.openembedded.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-core
Re: [OE-core] [PATCH 1/1] gdk-pixbuf: fix parallel install issue
On Sun, 2012-08-05 at 18:32 -0400, Colin Walters wrote: On Fri, 2012-08-03 at 11:30 +0800, wenzong@windriver.com wrote: Make an explicit dependency to the libs install targets would fix this issue. I don't think Yocto should be running the 'make install' target with parallelism enabled; it's just going to be a source of major pain for small gain. Historically Automake's install targets have had a lot of paralleism issues AIUI, even discarding modules which have custom install hooks. We did used to steer clear of this but more recently tested it out and with a few exceptions (which we disabled with PARALLEL_MAKEINST=), parallel make install seems to work fine and did give a small but measurable speed-up (I'd have to look at the archives to remember what). It helps that we use the latest automake everywhere and regenerate all the makefiles ourselves. So whilst I agree with this from a historical perspective, times so seem to be changing and improving. If there are issues I do want to know about and deal with them though. Cheers, Richard ___ Openembedded-core mailing list Openembedded-core@lists.openembedded.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-core
[OE-core] [PATCH] pango: upgrade to upstream stable 1.30.0
Signed-off-by: Radu Moisan radu.moi...@intel.com --- .../pango/pango-1.28.4/noconst.patch | 408 .../multilib-fix-clean.patch |0 .../{pango-1.28.4 = pango-1.30.0}/no-tests.patch | 14 +- .../pango/{pango_1.28.4.bb = pango_1.30.0.bb} |9 +- 4 files changed, 10 insertions(+), 421 deletions(-) delete mode 100644 meta/recipes-graphics/pango/pango-1.28.4/noconst.patch rename meta/recipes-graphics/pango/{pango-1.28.4 = pango-1.30.0}/multilib-fix-clean.patch (100%) rename meta/recipes-graphics/pango/{pango-1.28.4 = pango-1.30.0}/no-tests.patch (33%) rename meta/recipes-graphics/pango/{pango_1.28.4.bb = pango_1.30.0.bb} (49%) diff --git a/meta/recipes-graphics/pango/pango-1.28.4/noconst.patch b/meta/recipes-graphics/pango/pango-1.28.4/noconst.patch deleted file mode 100644 index d4832a5..000 --- a/meta/recipes-graphics/pango/pango-1.28.4/noconst.patch +++ /dev/null @@ -1,408 +0,0 @@ -G_CONST_RETURN is deprecated in glib 2.30 so remove to to avoid -build failures. - -RP 2011/10/12 - -Upstream-Status: Pending - -Index: pango-1.28.4/pango/fonts.c -=== pango-1.28.4.orig/pango/fonts.c2011-10-12 01:32:09.372046342 +0100 -+++ pango-1.28.4/pango/fonts.c 2011-10-12 01:32:34.512036630 +0100 -@@ -165,7 +165,7 @@ - * %NULL if not previously set. This has the same life-time - * as the font description itself and should not be freed. - **/ --G_CONST_RETURN char * -+const char * - pango_font_description_get_family (const PangoFontDescription *desc) - { - g_return_val_if_fail (desc != NULL, NULL); -@@ -1927,7 +1927,7 @@ - * Return value: the name of the family. This string is owned - * by the family object and must not be modified or freed. - **/ --G_CONST_RETURN char * -+const char * - pango_font_family_get_name (PangoFontFamily *family) - { - g_return_val_if_fail (PANGO_IS_FONT_FAMILY (family), NULL); -@@ -2060,7 +2060,7 @@ - * Return value: the face name for the face. This string is - * owned by the face object and must not be modified or freed. - **/ --G_CONST_RETURN char * -+const char * - pango_font_face_get_face_name (PangoFontFace *face) - { - g_return_val_if_fail (PANGO_IS_FONT_FACE (face), NULL); -Index: pango-1.28.4/pango/pango-attributes.c -=== pango-1.28.4.orig/pango/pango-attributes.c 2011-10-12 01:32:09.552046155 +0100 -+++ pango-1.28.4/pango/pango-attributes.c 2011-10-12 01:32:34.522037975 +0100 -@@ -97,7 +97,7 @@ - * - * Since: 1.22 - **/ --G_CONST_RETURN char * -+const char * - pango_attr_type_get_name (PangoAttrType type) - { - const char *result = NULL; -Index: pango-1.28.4/pango/pango-attributes.h -=== pango-1.28.4.orig/pango/pango-attributes.h 2011-10-12 01:32:12.712046218 +0100 -+++ pango-1.28.4/pango/pango-attributes.h 2011-10-12 01:32:36.342045777 +0100 -@@ -180,7 +180,7 @@ - }; - - PangoAttrType pango_attr_type_register (const gchar*name); --G_CONST_RETURN char * pango_attr_type_get_name (PangoAttrType type) G_GNUC_CONST; -+const char * pango_attr_type_get_name (PangoAttrType type) G_GNUC_CONST; - - void pango_attribute_init(PangoAttribute *attr, - const PangoAttrClass *klass); -Index: pango-1.28.4/pango/pango-context.c -=== pango-1.28.4.orig/pango/pango-context.c2011-10-12 01:32:09.782046152 +0100 -+++ pango-1.28.4/pango/pango-context.c 2011-10-12 01:32:34.532039187 +0100 -@@ -188,7 +188,7 @@ - * - * Since: 1.6 - **/ --G_CONST_RETURN PangoMatrix * -+const PangoMatrix * - pango_context_get_matrix (PangoContext *context) - { - g_return_val_if_fail (PANGO_IS_CONTEXT (context), NULL); -Index: pango-1.28.4/pango/pango-context.h -=== pango-1.28.4.orig/pango/pango-context.h2011-10-12 01:32:12.892046153 +0100 -+++ pango-1.28.4/pango/pango-context.h 2011-10-12 01:32:36.352046105 +0100 -@@ -86,7 +86,7 @@ - - voidpango_context_set_matrix (PangoContext *context, - const PangoMatrix *matrix); --G_CONST_RETURN PangoMatrix *pango_context_get_matrix (PangoContext *context); -+const PangoMatrix *pango_context_get_matrix (PangoContext *context); - - /* Break a string of Unicode characters into segments with - * consistent shaping/language engine and bidrectional level. -Index: pango-1.28.4/pango/pango-font.h -=== pango-1.28.4.orig/pango/pango-font.h 2011-10-12 01:32:13.072046150 +0100 -+++ pango-1.28.4/pango/pango-font.h2011-10-12
Re: [OE-core] [oe-core][RFC] bitbake.conf: exclude whole MACHINEOVERRIDES from OVERRIDES vardeps
On Sun, 2012-08-05 at 12:31 +0200, Martin Jansa wrote: On Mon, Jul 23, 2012 at 08:48:52AM -0700, Chris Larson wrote: On Mon, Jul 23, 2012 at 8:45 AM, Richard Purdie richard.pur...@linuxfoundation.org wrote: On Mon, 2012-07-23 at 16:25 +0200, Martin Jansa wrote: * whole MACHINEOVERRIDES can change e.g. between MACHINES with different arm architecture, causing allarch packages to reexecute do_package bitbake-diffsigs ../shr-core/tmp-eglibc/stamps/all-oe-linux/xserver-nodm-init-2.0-r16.do_package.sigdata.90e760a8f6cecbd87cb2e95f1237e3cc ../shr-core/tmp-eglibc/stamps/all-oe-linux/xserver-nodm-init-2.0-r16.do_package.sigdata.9eeccfd15f25032b3b6b132534660fff basehash changed from 7618e17d3fda05d1f15246e6800ca0f0 to 97bc4dc8c1521c535bd96b2aa62d8a03 Variable MACHINEOVERRIDES value changed from ${MACHINE}${@bb.utils.contains(TUNE_FEATURES, armv5, :armv5, ,d)}${@bb.utils.contains(TUNE_FEATURES, armv4, :armv4, ,d)}:${MACHINE_CLASS} to ${MACHINE}${@bb.utils.contains(TUNE_FEATURES, armv7a, :armv7a, ,d)}${@bb.utils.contains(TUNE_FEATURES, armv6, :armv6, ,d)}${@bb.utils.contains(TUNE_FEATURES, armv5, :armv5, ,d)}${@bb.utils.contains(TUNE_FEATURES, armv4, :armv4, ,d)}:${MACHINE_CLASS} Signed-off-by: Martin Jansa martin.ja...@gmail.com --- meta/conf/bitbake.conf |1 + 1 files changed, 1 insertions(+), 0 deletions(-) Won't this hide genuine changes where things should get rebuilt too? If something uses a machine override, won't the overridden value for that variable be the one which is stored in the checksum? So any effects of this will result in checksum modification anyway, no? I think it was possible to find different local file in SRC_URI (in different override subdirectory), but now with local file checksums included in sstate checksum it should be safe too. Yes, I think this should be safe now and will take the patch. Cheers, Richard ___ Openembedded-core mailing list Openembedded-core@lists.openembedded.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-core
Re: [OE-core] [PATCH V3 00/11] Mesa upgrade/improvements
Op 6 aug. 2012, om 14:43 heeft Richard Purdie richard.pur...@linuxfoundation.org het volgende geschreven: On Fri, 2012-08-03 at 17:46 +0200, Koen Kooi wrote: Op 3 aug. 2012, om 15:19 heeft Ross Burton ross.bur...@intel.com het volgende geschreven: On Friday, 3 August 2012 at 11:18, Koen Kooi wrote: libegl and libgles aren't built nowadays, so the problem is avoided. Noone has dared to touch this subject the past 2.5 years: http://cgit.openembedded.org/openembedded/commit/recipes/mesa?id=3d96f8cb61225d515b5cb4fe863f0d50c3ced436 The best solution[1] is to disable egl/gles in the mesa-recipes and add a seperate recipe for them. That way you still get glu and other useful mesa bits needed for gnome/efl/xfce/etc, but you can skip the sysroot/shlib poisoning if needed. After much digging I see that the Beagle PVR drivers only provide gles and egl, which is why you say that the problem was avoided. Of course that's just one specific driver, for example the Cedar Trail EMGD driver does build libgl, libgles and libegl. If the solution is put the egl/gles bits in another package Recipe, not package. Big difference :) then I'm totally confused as to what the actual problem is. Considering there's numerous libegl libraries for the many variants of PVR on ARM, I'm struggling to understand exactly what is new here. Can you explain clearly what the problem is, I'm obviously missing something. Once I understand the problem I can help with a solution. You need mesa for the !egl and !gles libs and evil vendor blob for egl and gles libs. In the emgd case you can get gl from the evil vendor as well. To build e.g. EFL I need mesa for the !egl and !gles bits, to build EFL with gles support I need mesa (again !egl, !gles) in addition to evil vendor blob for egl and gles. We could make this work with MACHINE_FEATURES or something similar, but that would mean that a very large chunk of userspace now becomes machine specific. This is how clutter worked in the past and just look at the fit the canonical/linaro people threw when they started doing 'embedded' and clutter. To compress the above: you need to build both the mesa and the evil-egl-binary *recipes* for things to work, so the egl bits should be a different recipe, not a subpackage of the mesa recipe. This is all assuming we care abou the binary blobs e.g. TI distributes for 3d support. If you don't, fine, merge this patchset. Let me test this a little bit more. The binary blobs shipped by TI would be machine specific They aren't machine specific, they contain multiple blobs to support a wide range of SoCs. But even if they were machine specific it wouldn't work like you want to. and would be marked as such in the package feeds? The mesa pieces would be marked as architecture specific and more generic. In feeds, machine specific has priority so on the machines where there is hardware support, the hardware optimised versions would get installed? You're assuming all libraries are subpackaged into their own package and renamed identically, they aren't (which is actually a feature in the current situation). If the above isn't the case, can we make it work like that? We can't. Even if we could, it would be a giant step backwards for people using the output of the build (see below). The only solution that keeps the good things of the current situation is to have the following recipes: 1) mesa (does not build libgl or libgles) 2) mesa-gles (only builds libgl+libgles) 3) evil-vendor-blob And you can specify per SoC something like PREFERRED_PROVIDER_virtual/egl = mesa-gles or PREFERRED_PROVIDER_virtual/egl = evil-vendor-blob. That will cause problems if you are building for 2 SoCs in the same architecture family that need different gl providers, the shlib code might get a bit upset. I don't have a good solution for that, but we have that problem right now as well. From the build side we could make it a bit prettier by making all the gl stuff machine specific, but that will be a giant pain in the ass for the end user (distro maintainer, $silicon_vendor customer, $silicon_vendor SDK). If we end up doing it like that, we'll need buy-in from the silicon vendors involved to change the way they distribute their blobs. That is doable, I managed to get the TI blobs changed a few times in the past, but it's a slow process. And everything using GL would suddenly be machine specific, I hope the yocto autobuilder can keep up with that. Executive summary: ARM SoCs with 3d suck, Intel parts with 3d sucks less. It's a mess in oe-core. regards, Koen ___ Openembedded-core mailing list Openembedded-core@lists.openembedded.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-core
Re: [OE-core] [PATCH 1/1] gdk-pixbuf: fix parallel install issue
Op 6 aug. 2012, om 00:32 heeft Colin Walters walt...@verbum.org het volgende geschreven: On Fri, 2012-08-03 at 11:30 +0800, wenzong@windriver.com wrote: Make an explicit dependency to the libs install targets would fix this issue. I don't think Yocto should be running the 'make install' target with parallelism enabled You mean oe-core, right? ___ Openembedded-core mailing list Openembedded-core@lists.openembedded.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-core
Re: [OE-core] [PATCH][RFC] u-boot: Upgrade to upstream stable 2012.07
I may be recipe configuration, however, here are my steps: bitbake u-boot -c cleanall bitbake u-boot -c fetch bitbake -c devshell u-boot in the new shell make coreboot-x86 (without _config) The output I got is as follows: Configuring for coreboot-x86 - Board: coreboot, Options: SYS_TEXT_BASE=0xFC make /bin/bash: i386-linux-gcc: command not found /bin/bash: i386-linux-gcc: command not found basename: missing operand Try `basename --help' for more information. make[1]: Entering directory `/home/radu/Documents/Development/yocto/build/tmp/work/qemux86-poky-linux/u-boot-2012.07-r0/git' /bin/bash: i386-linux-gcc: command not found basename: missing operand Try `basename --help' for more information. /bin/bash: i386-linux-gcc: command not found basename: missing operand Try `basename --help' for more information. Generating include/autoconf.mk /bin/bash: line 3: i386-linux-gcc: command not found /bin/bash: i386-linux-gcc: command not found basename: missing operand Try `basename --help' for more information. /bin/bash: i386-linux-gcc: command not found basename: missing operand Try `basename --help' for more information. Generating include/autoconf.mk.dep /bin/bash: line 3: i386-linux-gcc: command not found make[1]: Leaving directory `/home/radu/Documents/Development/yocto/build/tmp/work/qemux86-poky-linux/u-boot-2012.07-r0/git' /bin/bash: i386-linux-gcc: command not found /bin/bash: i386-linux-gcc: command not found basename: missing operand Try `basename --help' for more information. make[1]: Entering directory `/home/radu/Documents/Development/yocto/build/tmp/work/qemux86-poky-linux/u-boot-2012.07-r0/git' /bin/bash: i386-linux-gcc: command not found basename: missing operand Try `basename --help' for more information. /bin/bash: i386-linux-gcc: command not found basename: missing operand Try `basename --help' for more information. /bin/bash: i386-linux-gcc: command not found /bin/bash: i386-linux-ld: command not found /bin/bash: i386-linux-gcc: command not found basename: missing operand Try `basename --help' for more information. /bin/bash: i386-linux-gcc: command not found basename: missing operand Try `basename --help' for more information. /bin/bash: i386-linux-gcc: command not found basename: missing operand Try `basename --help' for more information. /bin/bash: i386-linux-gcc: command not found basename: missing operand Try `basename --help' for more information. i386-linux-gcc -DDO_DEPS_ONLY \ -g -Os -ffunction-sections -fvisibility=hidden -D__KERNEL__ -I/home/radu/Documents/Development/yocto/build/tmp/work/qemux86-poky-linux/u-boot-2012.07-r0/git/include -fno-builtin -ffreestanding -nostdinc -isystem -pipe -fno-strict-aliasing -Wstrict-prototypes -mregparm=3 -fomit-frame-pointer -ffreestanding -fno-toplevel-reorder -fno-stack-protector -fno-dwarf2-cfi-asm -DREALMODE_BASE=0x7c0 -DCONFIG_X86 -D__I386__ -march=i386 -Werror -Wall -Wstrict-prototypes -fno-stack-protector \ -o lib/asm-offsets.s lib/asm-offsets.c -c -S /bin/bash: i386-linux-gcc: command not found make[1]: *** [lib/asm-offsets.s] Error 127 make[1]: Leaving directory `/home/radu/Documents/Development/yocto/build/tmp/work/qemux86-poky-linux/u-boot-2012.07-r0/git' make: *** [coreboot-x86] Error 2 However this is the same out as when I do make coreboot-x86_config make all This happens only when I'm trying to build inside the devshell, which lead me to believe that maybe the environment is not set completely, until at do_compile. radu On 08/06/2012 10:45 AM, Wolfgang Denk wrote: Dear Radu, In message 501f63ff.1070...@intel.com you wrote: Actually a plain make NAME does the same as make NAME_config make all. You might also consider running ./MAKEALL NAME instead, which aut-adjusts to te number of available cores on the build host, so you get somewhat reduced build times. I've tried to build this way but got into some errors and did not pursued further in debugging them. However, because UBOOT_MACHINE is Could you please tell me exactly what these errors were? Log files etc. highly appreciated... Best regards, Wolfgang Denk ___ Openembedded-core mailing list Openembedded-core@lists.openembedded.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-core
Re: [OE-core] [oe-commits] Bogdan Marinescu : autoconf: updated to 2.69
On Wed, Jul 25, 2012 at 02:18:11PM +0200, Martin Jansa wrote: On Sun, Jul 22, 2012 at 10:44:06AM +, g...@git.openembedded.org wrote: Module: openembedded-core.git Branch: master Commit: effb75d42098b3e367d393215fd5d52a0191e954 URL: http://git.openembedded.org/?p=openembedded-core.gita=commit;h=effb75d42098b3e367d393215fd5d52a0191e954 Author: Bogdan Marinescu bogdan.a.marine...@intel.com Date: Thu Jul 19 13:33:52 2012 +0300 autoconf: updated to 2.69 Tested with core-image-sato-sdk and lib32-core-image-sato-sdk. This update was done mainly because multilib builds failed on master with this error: | autoreconf: running: aclocal -I /poky/build/tmp/work/x86-pokymllib32-linux/lib32-automake-1.12.1-r0/automake-1.12.1/m4/ -I /poky/build/tmp/sysroots/x86_64-linux/usr/share/aclocal-1.12 -I /poky/build/tmp/work/x86-pokymllib32-linux/lib32-automake-1.12.1-r0/automake-1.12.1/aclocal-copy/ -I /poky/build/tmp/work/x86-pokymllib32-linux/lib32-automake-1.12.1-r0/automake-1.12.1/m4/ -I /poky/build/tmp/sysroots/x86_64-linux/usr/share/aclocal-1.12 -I /poky/build/tmp/work/x86-pokymllib32-linux/lib32-automake-1.12.1-r0/automake-1.12.1/aclocal-copy/ --force --warnings=cross | aclocal: warning: unknown warning category 'cross' | configure.ac:18: error: Autoconf version 2.69 or higher is required since this upgrade I see autom4te segfaulting when building webkit-efl autom4te[8513]: segfault at 29e782688 ip 7f7c52a96e23 sp 7fff07731020 error 4 in Dumper.so[7f7c52a9+8000] webkit-efl build finishes fine and is usable afaik but someone would expect tools like autoconf to behave better and not segfault. This happens on 2 different builders each with different webkit-efl version. Nobody else have seen this in dmesg? It happens quite often and probably not only while building webkit-efl. Because I don't build webkit-efl so often: dmesg | grep -c segfault at.*Dumper.so 37 It could be related to perl as Dumper.so is part of perl install dev-lang/perl-5.16.0 (/usr/lib64/perl5/5.16.0/x86_64-linux-thread-multi/auto/Data/Dumper/Dumper.so) Maybe inheriting perlnative in autoconf would solve that too. Cheers, Cheers, Signed-off-by: Bogdan Marinescu bogdan.a.marine...@intel.com Signed-off-by: Richard Purdie richard.pur...@linuxfoundation.org --- meta/recipes-devtools/autoconf/autoconf.inc|2 +- .../{autoconf_2.68.bb = autoconf_2.69.bb} |4 ++-- 2 files changed, 3 insertions(+), 3 deletions(-) diff --git a/meta/recipes-devtools/autoconf/autoconf.inc b/meta/recipes-devtools/autoconf/autoconf.inc index 2c07701..4f5a5b2 100644 --- a/meta/recipes-devtools/autoconf/autoconf.inc +++ b/meta/recipes-devtools/autoconf/autoconf.inc @@ -12,7 +12,7 @@ RDEPENDS_${PN} = m4 gnu-config RDEPENDS_${PN}_virtclass-native = m4-native gnu-config-native RDEPENDS_${PN}_virtclass-nativesdk = m4-nativesdk gnu-config-nativesdk -SRC_URI = ${GNU_MIRROR}/autoconf/autoconf-${PV}.tar.bz2 \ +SRC_URI = ${GNU_MIRROR}/autoconf/autoconf-${PV}.tar.gz \ file://program_prefix.patch inherit autotools diff --git a/meta/recipes-devtools/autoconf/autoconf_2.68.bb b/meta/recipes-devtools/autoconf/autoconf_2.69.bb similarity index 85% rename from meta/recipes-devtools/autoconf/autoconf_2.68.bb rename to meta/recipes-devtools/autoconf/autoconf_2.69.bb index 8d466e0..478f8ed 100644 --- a/meta/recipes-devtools/autoconf/autoconf_2.68.bb +++ b/meta/recipes-devtools/autoconf/autoconf_2.69.bb @@ -17,8 +17,8 @@ SRC_URI += file://autoreconf-include.patch \ file://remove-usr-local-lib-from-m4.patch \ -SRC_URI[md5sum] = 864d785215aa60d627c91fcb21b05b07 -SRC_URI[sha256sum] = c491fb273fd6d4ca925e26ceed3d177920233c76d542b150ff35e571454332c8 +SRC_URI[md5sum] = 82d05e03b93e45f5a39b828dc9c6c29b +SRC_URI[sha256sum] = 954bd69b391edc12d6a4a51a2dd1476543da5c6bbf05a95b59dc0dd6fd4c2969 SRC_URI_append_virtclass-native = file://fix_path_xtra.patch ___ Openembedded-commits mailing list openembedded-comm...@lists.openembedded.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-commits -- Martin 'JaMa' Jansa jabber: martin.ja...@gmail.com -- Martin 'JaMa' Jansa jabber: martin.ja...@gmail.com signature.asc Description: Digital signature ___ Openembedded-core mailing list Openembedded-core@lists.openembedded.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-core
Re: [OE-core] [PATCH][RFC] u-boot: Upgrade to upstream stable 2012.07
Dear Radu, In message 501fc96a.8000...@intel.com you wrote: in the new shell make coreboot-x86 (without _config) The output I got is as follows: Configuring for coreboot-x86 - Board: coreboot, Options: SYS_TEXT_BASE=0xFC make /bin/bash: i386-linux-gcc: command not found /bin/bash: i386-linux-gcc: command not found It seems CROSS_COMPILE is set to i386-linux-, but i386-linux-gcc is not anywhere in your PATH. Please check your PATH and fix it. Best regards, Wolfgang Denk -- DENX Software Engineering GmbH, MD: Wolfgang Denk Detlev Zundel HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany Phone: (+49)-8142-66989-10 Fax: (+49)-8142-66989-80 Email: w...@denx.de Advice is seldom welcome; and those who want it the most always like it the least. -- Philip Earl of Chesterfield ___ Openembedded-core mailing list Openembedded-core@lists.openembedded.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-core
Re: [OE-core] [PATCH V3 00/11] Mesa upgrade/improvements
On Mon, 2012-08-06 at 15:24 +0200, Koen Kooi wrote: Op 6 aug. 2012, om 14:43 heeft Richard Purdie richard.pur...@linuxfoundation.org het volgende geschreven: On Fri, 2012-08-03 at 17:46 +0200, Koen Kooi wrote: Op 3 aug. 2012, om 15:19 heeft Ross Burton ross.bur...@intel.com het volgende geschreven: On Friday, 3 August 2012 at 11:18, Koen Kooi wrote: libegl and libgles aren't built nowadays, so the problem is avoided. Noone has dared to touch this subject the past 2.5 years: http://cgit.openembedded.org/openembedded/commit/recipes/mesa?id=3d96f8cb61225d515b5cb4fe863f0d50c3ced436 The best solution[1] is to disable egl/gles in the mesa-recipes and add a seperate recipe for them. That way you still get glu and other useful mesa bits needed for gnome/efl/xfce/etc, but you can skip the sysroot/shlib poisoning if needed. After much digging I see that the Beagle PVR drivers only provide gles and egl, which is why you say that the problem was avoided. Of course that's just one specific driver, for example the Cedar Trail EMGD driver does build libgl, libgles and libegl. If the solution is put the egl/gles bits in another package Recipe, not package. Big difference :) then I'm totally confused as to what the actual problem is. Considering there's numerous libegl libraries for the many variants of PVR on ARM, I'm struggling to understand exactly what is new here. Can you explain clearly what the problem is, I'm obviously missing something. Once I understand the problem I can help with a solution. You need mesa for the !egl and !gles libs and evil vendor blob for egl and gles libs. In the emgd case you can get gl from the evil vendor as well. To build e.g. EFL I need mesa for the !egl and !gles bits, to build EFL with gles support I need mesa (again !egl, !gles) in addition to evil vendor blob for egl and gles. We could make this work with MACHINE_FEATURES or something similar, but that would mean that a very large chunk of userspace now becomes machine specific. This is how clutter worked in the past and just look at the fit the canonical/linaro people threw when they started doing 'embedded' and clutter. To compress the above: you need to build both the mesa and the evil-egl-binary *recipes* for things to work, so the egl bits should be a different recipe, not a subpackage of the mesa recipe. This is all assuming we care abou the binary blobs e.g. TI distributes for 3d support. If you don't, fine, merge this patchset. Let me test this a little bit more. The binary blobs shipped by TI would be machine specific They aren't machine specific, they contain multiple blobs to support a wide range of SoCs. But even if they were machine specific it wouldn't work like you want to. and would be marked as such in the package feeds? The mesa pieces would be marked as architecture specific and more generic. In feeds, machine specific has priority so on the machines where there is hardware support, the hardware optimised versions would get installed? You're assuming all libraries are subpackaged into their own package and renamed identically, they aren't (which is actually a feature in the current situation). They don't have to be renamed identically, the would have to have RPROVIDES that cover what is needed in some standard namespace. If the above isn't the case, can we make it work like that? We can't. Even if we could, it would be a giant step backwards for people using the output of the build (see below). The only solution that keeps the good things of the current situation is to have the following recipes: 1) mesa (does not build libgl or libgles) 2) mesa-gles (only builds libgl+libgles) 3) evil-vendor-blob And you can specify per SoC something like PREFERRED_PROVIDER_virtual/egl = mesa-gles or PREFERRED_PROVIDER_virtual/egl = evil-vendor-blob. That will cause problems if you are building for 2 SoCs in the same architecture family that need different gl providers, the shlib code might get a bit upset. I don't have a good solution for that, but we have that problem right now as well. From the build side we could make it a bit prettier by making all the gl stuff machine specific, but that will be a giant pain in the ass for the end user (distro maintainer, $silicon_vendor customer, $silicon_vendor SDK). If we end up doing it like that, we'll need buy-in from the silicon vendors involved to change the way they distribute their blobs. That is doable, I managed to get the TI blobs changed a few times in the past, but it's a slow process. And everything using GL would suddenly be machine specific, I hope the yocto autobuilder can keep up with that. Everything does not have to become machine specific if the things in question have known good portable APIs. As I understand it we have GL so we can swap in the binary blobs in place of the generic ones without
Re: [OE-core] [PATCH][RFC] u-boot: Upgrade to upstream stable 2012.07
CROSS_COMPILE is set in EXTRA_OEMAKE which is used in oe-runmake() within do_compile() Running devshell is achieved in do_devshell() which is run after do_patch() but before do_compile(), that's why in devshell CROSS_COMPILE is not (yet) set. I think this is the intended way for things to work, devshell usage, as I see it, is mainly for pre-compile debugging purposes. radu On 08/06/2012 04:46 PM, Wolfgang Denk wrote: Dear Radu, In message 501fc96a.8000...@intel.com you wrote: in the new shell make coreboot-x86 (without _config) The output I got is as follows: Configuring for coreboot-x86 - Board: coreboot, Options: SYS_TEXT_BASE=0xFC make /bin/bash: i386-linux-gcc: command not found /bin/bash: i386-linux-gcc: command not found It seems CROSS_COMPILE is set to i386-linux-, but i386-linux-gcc is not anywhere in your PATH. Please check your PATH and fix it. Best regards, Wolfgang Denk ___ Openembedded-core mailing list Openembedded-core@lists.openembedded.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-core
Re: [OE-core] [PATCH][RFC] u-boot: Upgrade to upstream stable 2012.07
Dear Radu, In message 501fd013.5040...@intel.com you wrote: CROSS_COMPILE is set in EXTRA_OEMAKE which is used in oe-runmake() within do_compile() Running devshell is achieved in do_devshell() which is run after do_patch() but before do_compile(), that's why in devshell CROSS_COMPILE is not (yet) set. But CROSS_COMPILE _is_ set (and apparently has the value i386-linux-. The problem is that PATH is not set such that it contains any such file as i386-linux-gcc. Best regards, Wolfgang Denk -- DENX Software Engineering GmbH, MD: Wolfgang Denk Detlev Zundel HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany Phone: (+49)-8142-66989-10 Fax: (+49)-8142-66989-80 Email: w...@denx.de Beware of bugs in the above code; I have only proved it correct, not tried it. - Donald Knuth ___ Openembedded-core mailing list Openembedded-core@lists.openembedded.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-core
Re: [OE-core] [PATCH V3 00/11] Mesa upgrade/improvements
On Thu, 2012-08-02 at 17:48 +0100, Ross Burton wrote: Re-arranged and fixed up some problems. The following changes since commit 62d42fe3cfa575cb97b484ccf7b2e9a25cc50770: tiny-init: Setup /dev/ptmx in init (2012-08-01 23:11:18 +0100) are available in the git repository at: git://git.yoctoproject.org/poky-contrib ross/mesa for you to fetch changes up to 5de29e150a720ccb8749362cfe5fb56f234fb385: mesa: enable EGL, with DRM and X11 platforms (2012-08-02 17:43:18 +0100) Damien Lespiau (2): mesa: Update to 8.0.4 (latest stable version) mesa: Use 'require' instead of 'include' Ross Burton (9): mesa: no need to depend on python-native, the class does that mesa: format the packages list nicely mesa: move glu.pc to libglu-dev mesa: add --enable-shared-glapi, and package it in libglapi mesa: enable the Graphic Buffer Manager library I've merged the above since I don't think these are the contentious part. mesa: enable GLES v1 and v2 mesa-demos: fix GLES2 build mesa: respect x11 DISTRO_FEATURE mesa: enable EGL, with DRM and X11 platforms I've left these for now for more discussion. Cheers, Richard ___ Openembedded-core mailing list Openembedded-core@lists.openembedded.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-core
[OE-core] do_rootfs broken due to missing gcc-runtime
Hi, having something in the image which brings gcc-runtime-locale-XX in causes do_rootfs to fail with: * satisfy_dependencies_for: Cannot satisfy the following dependencies for gcc-runtime-locale-de: * gcc-runtime * * opkg_install_cmd: Cannot install package gcc-runtime-locale-de. This happens for some time already and I saw a patch[1] on the list which fixes it. Any reason why it has not been applied yet? Enrico Footnotes: [1] http://permalink.gmane.org/gmane.comp.handhelds.openembedded.core/24694 ___ Openembedded-core mailing list Openembedded-core@lists.openembedded.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-core
Re: [OE-core] [PATCH V3 00/11] Mesa upgrade/improvements
Op 6 aug. 2012, om 16:04 heeft Richard Purdie richard.pur...@linuxfoundation.org het volgende geschreven: On Mon, 2012-08-06 at 15:24 +0200, Koen Kooi wrote: Op 6 aug. 2012, om 14:43 heeft Richard Purdie richard.pur...@linuxfoundation.org het volgende geschreven: On Fri, 2012-08-03 at 17:46 +0200, Koen Kooi wrote: Op 3 aug. 2012, om 15:19 heeft Ross Burton ross.bur...@intel.com het volgende geschreven: On Friday, 3 August 2012 at 11:18, Koen Kooi wrote: libegl and libgles aren't built nowadays, so the problem is avoided. Noone has dared to touch this subject the past 2.5 years: http://cgit.openembedded.org/openembedded/commit/recipes/mesa?id=3d96f8cb61225d515b5cb4fe863f0d50c3ced436 The best solution[1] is to disable egl/gles in the mesa-recipes and add a seperate recipe for them. That way you still get glu and other useful mesa bits needed for gnome/efl/xfce/etc, but you can skip the sysroot/shlib poisoning if needed. After much digging I see that the Beagle PVR drivers only provide gles and egl, which is why you say that the problem was avoided. Of course that's just one specific driver, for example the Cedar Trail EMGD driver does build libgl, libgles and libegl. If the solution is put the egl/gles bits in another package Recipe, not package. Big difference :) then I'm totally confused as to what the actual problem is. Considering there's numerous libegl libraries for the many variants of PVR on ARM, I'm struggling to understand exactly what is new here. Can you explain clearly what the problem is, I'm obviously missing something. Once I understand the problem I can help with a solution. You need mesa for the !egl and !gles libs and evil vendor blob for egl and gles libs. In the emgd case you can get gl from the evil vendor as well. To build e.g. EFL I need mesa for the !egl and !gles bits, to build EFL with gles support I need mesa (again !egl, !gles) in addition to evil vendor blob for egl and gles. We could make this work with MACHINE_FEATURES or something similar, but that would mean that a very large chunk of userspace now becomes machine specific. This is how clutter worked in the past and just look at the fit the canonical/linaro people threw when they started doing 'embedded' and clutter. To compress the above: you need to build both the mesa and the evil-egl-binary *recipes* for things to work, so the egl bits should be a different recipe, not a subpackage of the mesa recipe. This is all assuming we care abou the binary blobs e.g. TI distributes for 3d support. If you don't, fine, merge this patchset. Let me test this a little bit more. The binary blobs shipped by TI would be machine specific They aren't machine specific, they contain multiple blobs to support a wide range of SoCs. But even if they were machine specific it wouldn't work like you want to. and would be marked as such in the package feeds? The mesa pieces would be marked as architecture specific and more generic. In feeds, machine specific has priority so on the machines where there is hardware support, the hardware optimised versions would get installed? You're assuming all libraries are subpackaged into their own package and renamed identically, they aren't (which is actually a feature in the current situation). They don't have to be renamed identically, the would have to have RPROVIDES that cover what is needed in some standard namespace. If the above isn't the case, can we make it work like that? We can't. Even if we could, it would be a giant step backwards for people using the output of the build (see below). The only solution that keeps the good things of the current situation is to have the following recipes: 1) mesa (does not build libgl or libgles) 2) mesa-gles (only builds libgl+libgles) 3) evil-vendor-blob And you can specify per SoC something like PREFERRED_PROVIDER_virtual/egl = mesa-gles or PREFERRED_PROVIDER_virtual/egl = evil-vendor-blob. That will cause problems if you are building for 2 SoCs in the same architecture family that need different gl providers, the shlib code might get a bit upset. I don't have a good solution for that, but we have that problem right now as well. From the build side we could make it a bit prettier by making all the gl stuff machine specific, but that will be a giant pain in the ass for the end user (distro maintainer, $silicon_vendor customer, $silicon_vendor SDK). If we end up doing it like that, we'll need buy-in from the silicon vendors involved to change the way they distribute their blobs. That is doable, I managed to get the TI blobs changed a few times in the past, but it's a slow process. And everything using GL would suddenly be machine specific, I hope the yocto autobuilder can keep up with that. Everything does not have to become machine specific if the things in question have known good portable APIs. As I understand it we have GL so we
Re: [OE-core] [oe-commits] Bogdan Marinescu : autoconf: updated to 2.69
On Mon, 2012-08-06 at 15:45 +0200, Martin Jansa wrote: On Wed, Jul 25, 2012 at 02:18:11PM +0200, Martin Jansa wrote: On Sun, Jul 22, 2012 at 10:44:06AM +, g...@git.openembedded.org wrote: Module: openembedded-core.git Branch: master Commit: effb75d42098b3e367d393215fd5d52a0191e954 URL: http://git.openembedded.org/?p=openembedded-core.gita=commit;h=effb75d42098b3e367d393215fd5d52a0191e954 Author: Bogdan Marinescu bogdan.a.marine...@intel.com Date: Thu Jul 19 13:33:52 2012 +0300 autoconf: updated to 2.69 Tested with core-image-sato-sdk and lib32-core-image-sato-sdk. This update was done mainly because multilib builds failed on master with this error: | autoreconf: running: aclocal -I /poky/build/tmp/work/x86-pokymllib32-linux/lib32-automake-1.12.1-r0/automake-1.12.1/m4/ -I /poky/build/tmp/sysroots/x86_64-linux/usr/share/aclocal-1.12 -I /poky/build/tmp/work/x86-pokymllib32-linux/lib32-automake-1.12.1-r0/automake-1.12.1/aclocal-copy/ -I /poky/build/tmp/work/x86-pokymllib32-linux/lib32-automake-1.12.1-r0/automake-1.12.1/m4/ -I /poky/build/tmp/sysroots/x86_64-linux/usr/share/aclocal-1.12 -I /poky/build/tmp/work/x86-pokymllib32-linux/lib32-automake-1.12.1-r0/automake-1.12.1/aclocal-copy/ --force --warnings=cross | aclocal: warning: unknown warning category 'cross' | configure.ac:18: error: Autoconf version 2.69 or higher is required since this upgrade I see autom4te segfaulting when building webkit-efl autom4te[8513]: segfault at 29e782688 ip 7f7c52a96e23 sp 7fff07731020 error 4 in Dumper.so[7f7c52a9+8000] webkit-efl build finishes fine and is usable afaik but someone would expect tools like autoconf to behave better and not segfault. This happens on 2 different builders each with different webkit-efl version. Nobody else have seen this in dmesg? It happens quite often and probably not only while building webkit-efl. Because I don't build webkit-efl so often: dmesg | grep -c segfault at.*Dumper.so 37 It could be related to perl as Dumper.so is part of perl install dev-lang/perl-5.16.0 (/usr/lib64/perl5/5.16.0/x86_64-linux-thread-multi/auto/Data/Dumper/Dumper.so) Maybe inheriting perlnative in autoconf would solve that too. autoconf-native *must* use the system perl. Trying to do anything else will result in circular dependencies. Cheers, Richard ___ Openembedded-core mailing list Openembedded-core@lists.openembedded.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-core
Re: [OE-core] [PATCH 0/9] buildhistory improvements
On Thu, 2012-08-02 at 10:23 +0100, Paul Eggleton wrote: A bunch of improvements for buildhistory, implementing some feature requests and todo items of my own, as well as tidying up some of the code. The following changes since commit 404d2d490fc347203e89d274530c17fb5f0aa20f: gcc-configure-target: Set native-system-header-dir for target gcc (2012-08-01 23:11:09 +0100) are available in the git repository at: git://git.openembedded.org/openembedded-core-contrib paule/buildhistory-fixes7 http://cgit.openembedded.org/cgit.cgi/openembedded-core-contrib/log/?h=paule/buildhistory-fixes7 Paul Eggleton (9): classes/buildhistory: remove obsolete flat package history feature classes/buildhistory: ensure old package info is removed classes/buildhistory: remove unused functions classes/buildhistory: save preinst/postinst/prerm/postrm classes/buildhistory: record PKG/PKGE/PKGV/PKGR buildhistory_analysis: tidy up duplicate split_version function buildhistory_analysis: ignore removal of self-dependencies classes/buildhistory: save metadata revisions Merged to master apart from: scripts: add buildhistory-tag script which seems to need more discussion/work. Cheers, Richard ___ Openembedded-core mailing list Openembedded-core@lists.openembedded.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-core
Re: [OE-core] (no subject)
On Sun, 2012-08-05 at 21:48 +0200, Javier Martinez Canillas wrote: The OpenEmbedded User Manual list the variables that should be used to control the directories into which files are installed. It says that is a poor practice to specify hardcoded paths instead of using these variables, yet there are many recipes that don't use it. This is second version of a big patch-set that does a cleanup and replace the hardcoded paths used on these recipes with the build system variables. I tried to be as careful as possible to do the proper replacement but since I could introduce regressions I split the changes in 30 different patches so it could be git bisectable in case of messing a recipe. Also, the patches increment the recipes PR since the a distro config can set the variables to a different value. Changes since v1: - Bump recipes PR as suggested by Otavio Salvador and Khem Raj - Squash ${base_bindir} and ${sysconfdir} changes for xinetd and lsb recipes so the PR number gets incremented only once. The patch-set consist of the following patches: [PATCH v2 01/28] xinetd: use ${sbindir} and ${sysconfdir} instead of [PATCH v2 02/28] alsa-state: use ${sbindir} instead of /usr/sbin for [PATCH v2 03/28] lsbsetup: use ${bindir} instead of /usr/bin for [PATCH v2 04/28] sudo: use ${bindir} and ${sysconfdir} instead of [PATCH v2 05/28] lsbtest: use ${bindir} instead of /usr/bin for [PATCH v2 06/28] cronie: use variables instead of hardcoded paths [PATCH v2 07/28] useradd-example: use ${datadir} instead of [PATCH v2 08/28] ubootchart: use variables instead of hardcoded [PATCH v2 09/28] xkeyboard-config: use ${datadir} instead of [PATCH v2 10/28] systemtap: use ${datadir} instead of /usr/share for [PATCH v2 11/28] lsb: use ${base_bindir} and ${sysconfdir} instead [PATCH v2 12/28] mingetty: use ${base_sbindir} instead of /sbin for [PATCH v2 13/28] external-sourcery: use ${prefix} and ${libdir} [PATCH v2 14/28] rpm: use ${localstatedir} and ${libdir} instead of [PATCH v2 15/28] at: use ${base_sbindir} instead of /sbin for [PATCH v2 16/28] kernel.bbclass: use ${base_libdir} and [PATCH v2 17/28] linux-firware: use ${base_libdir} instead of /lib [PATCH v2 18/28] openssh: use ${localstatedir} instead of /var for [PATCH v2 19/28] libpam: use ${localstatedir} and ${sysconfdir} [PATCH v2 20/28] x11-common: use ${sysconfdir} instead of /etc for [PATCH v2 21/28] builder: use ${sysconfdir} instead of /etc for [PATCH v2 22/28] xserver-nodm-init: use ${sysconfdir} instead of [PATCH v2 23/28] lsbinitscripts: use ${sysconfdir} instead of /etc [PATCH v2 24/28] usbinit: use ${sysconfdir} instead of /etc for [PATCH v2 25/28] qemu-config: use ${sysconfdir} instead of /etc for [PATCH v2 26/28] rsync: use ${sysconfdir} instead of /etc for [PATCH v2 27/28] chkconfig: use ${sysconfdir} instead of /etc for [PATCH v2 28/28] man: use ${sysconfdir} instead of /etc for Thanks for these, I merged most of them apart from external-sourcery which Chris commented on, the at recipe which I found a better fix for which removed the lines in question and the rpm change which I want to check something out related to it first. Cheers, Richard ___ Openembedded-core mailing list Openembedded-core@lists.openembedded.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-core
Re: [OE-core] do_rootfs broken due to missing gcc-runtime
On Mon, 2012-08-06 at 16:21 +0200, Enrico Scholz wrote: having something in the image which brings gcc-runtime-locale-XX in causes do_rootfs to fail with: * satisfy_dependencies_for: Cannot satisfy the following dependencies for gcc-runtime-locale-de: * gcc-runtime * * opkg_install_cmd: Cannot install package gcc-runtime-locale-de. This happens for some time already and I saw a patch[1] on the list which fixes it. Any reason why it has not been applied yet? [1] http://permalink.gmane.org/gmane.comp.handhelds.openembedded.core/24694 Thanks for the reminder, I merged this. Cheers, Richard ___ Openembedded-core mailing list Openembedded-core@lists.openembedded.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-core
Re: [OE-core] opkg-nogpg
On Tue, 2012-07-31 at 22:36 +, Slater, Joseph wrote: It looks like opkg already uses --disable-gpg, so opkg-nogpg builds the same and the packaging conflicts with opkg, so perhaps it could be eliminated? We should put a PACKAGECONFIG option into the main opkg recipe, then delete the other one... Cheers, Richard ___ Openembedded-core mailing list Openembedded-core@lists.openembedded.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-core
Re: [OE-core] [PATCH 1/1] linux-yocto-custom: Clarify usage and clear COMPATIBLE_MACHINE
On 08/03/2012 05:27 PM, Bruce Ashfield wrote: On Fri, Aug 3, 2012 at 7:29 PM, Darren Hart dvh...@linux.intel.com wrote: There has been some confusion over proper use of the linux-yocto-custom recipe. It is not intended to build as is from meta-skeleton. It should be modified via a bbappend file to provide a Linux kernel config at the very least. Update the commentary to make this requirement more explicit. Add some additional detail about how to create a bbappend file and how and when to modify the various variables. Clear COMPATIBLE_MACHINE so bitbake will not attempt to build the recipe unless the user explicitly adds there machine to the variable, which should encourage them to read the recipe comments before attempting to build it. Signed-off-by: Darren Hart dvh...@linux.intel.com CC: Bruce Ashfield bruce.ashfi...@windriver.com CC: Tom Zanussi tom.zanu...@intel.com --- .../recipes-kernel/linux/linux-yocto-custom.bb | 52 +++--- 1 file changed, 35 insertions(+), 17 deletions(-) diff --git a/meta-skeleton/recipes-kernel/linux/linux-yocto-custom.bb b/meta-skeleton/recipes-kernel/linux/linux-yocto-custom.bb index 55f0c38..dd98228 100644 --- a/meta-skeleton/recipes-kernel/linux/linux-yocto-custom.bb +++ b/meta-skeleton/recipes-kernel/linux/linux-yocto-custom.bb @@ -1,17 +1,35 @@ # linux-yocto-custom.bb: # -# Provides an example/minimal kernel recipe that uses the linux-yocto -# and oe-core kernel classes to apply a subset of yocto kernel -# management to git managed kernel repositories. +# An example kernel recipe that uses the linux-yocto and oe-core +# kernel classes to apply a subset of yocto kernel management to git +# managed kernel repositories. +# +# To use linux-yocto-custom in your layer, create a +# linux-yocto-custom.bb file containing at least the following lines: s/linux-yocto-custom.bb/linux-yocto-custom.bbappend/ Thanks! +# +# FILESEXTRAPATHS_prepend := ${THISDIR}/${PN}: +# COMPATIBLE_MACHINE_yourmachine = yourmachine +# +# You must also provide a Linux kernel configuration. The most direct +# method is to copy your .config to files/defconfig in your layer, +# parallel to the linux-yocto-custom.bbappend file. hmm. parallel ready odd to me, but I don't have a better suggestion. in the same directory as the bbappend. +# +# To use the yocto kernel tooling to generate a BSP configuration +# using modular configuration fragments, see the yocto-bsp and +# yocto-kernel tools documentation. +# +# Warning: +# +# Building this example without providing a defconfig or BSP +# configuration will result in build or boot errors. This is not a +# bug. +# # # Notes: # -# kconfig(s): the kernel must be configured with a defconfig, or via -# configuration fragment(s). Either of these can be added -# via bbappend. Leaving the part about configuration fragments might be useful. I removed it because I felt I adequately covered it above: # You must also provide a Linux kernel configuration. The most direct # method is to copy your .config to files/defconfig in your layer, # in the same directory as the bbappend. # # To use the yocto kernel tooling to generate a BSP configuration # using modular configuration fragments, see the yocto-bsp and # yocto-kernel tools documentation. .. but this looks good, lets see if it saves a few questions :) Will send V2 with your comments addressed. -- Darren Cheers, Bruce -# patches: patches can be merged into to the source git tree itself, added -#using standard bbappend syntax or controlled via .scc feature -#descriptions (also via bbappends) +# patches: patches can be merged into to the source git tree itself, +#added via the SRC_URI, or controlled via a BSP +#configuration. # # example configuration addition: #SRC_URI += file://smp.cfg @@ -20,25 +38,25 @@ # example feature addition (for kernel v3.4 only): #SRC_URI += file://feature.scc # -# Warning: -# -# Building the sample kernel tree (kernel.org) without providing any -# configuration will result in build or boot errors. This is not a bug -# it is a required element for creating a valid kernel. -# inherit kernel require recipes-kernel/linux/linux-yocto.inc +# Override SRC_URI in a bbappend file to point at a different source +# tree if you do not want to build from Linus' tree. SRC_URI = git://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git;protocol=git;nocheckout=1 LINUX_VERSION ?= 3.4 LINUX_VERSION_EXTENSION ?= -custom +# Override SRCREV to point to a different commit in a bbappend file to +# build a different release of the Linux kernel. # tag: v3.4 76e10d158efb6d4516018846f60c2ab5501900bc SRCREV=76e10d158efb6d4516018846f60c2ab5501900bc -PR = r0 +PR = r1 PV =
[OE-core] [PATCH 1/1] linux-yocto-custom: Clarify usage and clear COMPATIBLE_MACHINE
There has been some confusion over proper use of the linux-yocto-custom recipe. It is not intended to build as is from meta-skeleton. It should be modified via a bbappend file to provide a Linux kernel config at the very least. Update the commentary to make this requirement more explicit. Add some additional detail about how to create a bbappend file and how and when to modify the various variables. Clear COMPATIBLE_MACHINE so bitbake will not attempt to build the recipe unless the user explicitly adds there machine to the variable, which should encourage them to read the recipe comments before attempting to build it. Signed-off-by: Darren Hart dvh...@linux.intel.com CC: Bruce Ashfield bruce.ashfi...@windriver.com CC: Tom Zanussi tom.zanu...@intel.com Signed-off-by: Darren Hart dvh...@linux.intel.com --- .../recipes-kernel/linux/linux-yocto-custom.bb | 53 +++--- 1 file changed, 36 insertions(+), 17 deletions(-) diff --git a/meta-skeleton/recipes-kernel/linux/linux-yocto-custom.bb b/meta-skeleton/recipes-kernel/linux/linux-yocto-custom.bb index 55f0c38..1f0b3a2 100644 --- a/meta-skeleton/recipes-kernel/linux/linux-yocto-custom.bb +++ b/meta-skeleton/recipes-kernel/linux/linux-yocto-custom.bb @@ -1,17 +1,36 @@ # linux-yocto-custom.bb: # -# Provides an example/minimal kernel recipe that uses the linux-yocto -# and oe-core kernel classes to apply a subset of yocto kernel -# management to git managed kernel repositories. +# An example kernel recipe that uses the linux-yocto and oe-core +# kernel classes to apply a subset of yocto kernel management to git +# managed kernel repositories. +# +# To use linux-yocto-custom in your layer, create a +# linux-yocto-custom.bbappend file containing at least the following +# lines: +# +# FILESEXTRAPATHS_prepend := ${THISDIR}/${PN}: +# COMPATIBLE_MACHINE_yourmachine = yourmachine +# +# You must also provide a Linux kernel configuration. The most direct +# method is to copy your .config to files/defconfig in your layer, +# in the same directory as the bbappend. +# +# To use the yocto kernel tooling to generate a BSP configuration +# using modular configuration fragments, see the yocto-bsp and +# yocto-kernel tools documentation. +# +# Warning: +# +# Building this example without providing a defconfig or BSP +# configuration will result in build or boot errors. This is not a +# bug. +# # # Notes: # -# kconfig(s): the kernel must be configured with a defconfig, or via -# configuration fragment(s). Either of these can be added -# via bbappend. -# patches: patches can be merged into to the source git tree itself, added -#using standard bbappend syntax or controlled via .scc feature -#descriptions (also via bbappends) +# patches: patches can be merged into to the source git tree itself, +#added via the SRC_URI, or controlled via a BSP +#configuration. # # example configuration addition: #SRC_URI += file://smp.cfg @@ -20,25 +39,25 @@ # example feature addition (for kernel v3.4 only): #SRC_URI += file://feature.scc # -# Warning: -# -# Building the sample kernel tree (kernel.org) without providing any -# configuration will result in build or boot errors. This is not a bug -# it is a required element for creating a valid kernel. -# inherit kernel require recipes-kernel/linux/linux-yocto.inc +# Override SRC_URI in a bbappend file to point at a different source +# tree if you do not want to build from Linus' tree. SRC_URI = git://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git;protocol=git;nocheckout=1 LINUX_VERSION ?= 3.4 LINUX_VERSION_EXTENSION ?= -custom +# Override SRCREV to point to a different commit in a bbappend file to +# build a different release of the Linux kernel. # tag: v3.4 76e10d158efb6d4516018846f60c2ab5501900bc SRCREV=76e10d158efb6d4516018846f60c2ab5501900bc -PR = r0 +PR = r1 PV = ${LINUX_VERSION}+git${SRCPV} -COMPATIBLE_MACHINE = (qemuarm|qemux86|qemuppc|qemumips|qemux86-64) +# Override COMPATIBLE_MACHINE to include your machine in a bbappend +# file. Leaving it empty here ensures an early explicit build failure. +COMPATIBLE_MACHINE = (^$) -- 1.7.11.2 ___ Openembedded-core mailing list Openembedded-core@lists.openembedded.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-core
[OE-core] [PATCH V2 0/1] linux-yocto-custom: Clarify usage and clear COMPATIBLE_MACHINE
The following changes since commit 47aca0f0f79c30d1df1f89c710d3e354f436145d: subversion: Add missing build dependency on sqlite3 (2012-08-06 16:13:45 +0100) are available in the git repository at: git://git.yoctoproject.org/user-contrib/dvhart/oe-core skeleton http://git.yoctoproject.org/cgit.cgi/user-contrib/dvhart/oe-core/log/?h=skeleton Darren Hart (1): linux-yocto-custom: Clarify usage and clear COMPATIBLE_MACHINE .../recipes-kernel/linux/linux-yocto-custom.bb | 53 +++--- 1 file changed, 36 insertions(+), 17 deletions(-) -- 1.7.11.2 ___ Openembedded-core mailing list Openembedded-core@lists.openembedded.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-core
[OE-core] proper(?) way to *require* an earlier, unsupported version of a recipe file?
what is the OE philosophy for preferring an earlier version of a recipe file that doesn't even exist anymore? for example, the watchdog recipe was recently upgraded from 5.11 to 5.12. in the process, the recipe file for 5.11 was removed; therefore, i interpret that as that OE-core no longer supports that version, which is fine. hypothetically, what if someone still *needed* that earlier version? would it be their responsibility to, say, create a new layer that re-introduced that older, now-unsupported version? just curious. rday -- Robert P. J. Day Ottawa, Ontario, CANADA http://crashcourse.ca Twitter: http://twitter.com/rpjday LinkedIn: http://ca.linkedin.com/in/rpjday ___ Openembedded-core mailing list Openembedded-core@lists.openembedded.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-core
Re: [OE-core] [PATCH v2 16/28] kernel.bbclass: use ${base_libdir} and ${sysconfdir} instead of /lib and /etc
On 08/05/2012 12:48 PM, Javier Martinez Canillas wrote: Hi Javier, It is considered good practice to use the build system provided variables instead of directly specify hardcoded paths. Have you tested this with a build using a base_libdir other than /lib ? Signed-off-by: Javier Martinez Canillas jav...@dowhile0.org --- meta/classes/kernel.bbclass | 12 ++-- 1 files changed, 6 insertions(+), 6 deletions(-) diff --git a/meta/classes/kernel.bbclass b/meta/classes/kernel.bbclass index 1d8dff9..b434093 100644 --- a/meta/classes/kernel.bbclass +++ b/meta/classes/kernel.bbclass @@ -109,10 +109,10 @@ kernel_do_install() { unset CFLAGS CPPFLAGS CXXFLAGS LDFLAGS MACHINE if (grep -q -i -e '^CONFIG_MODULES=y$' .config); then oe_runmake DEPMOD=echo INSTALL_MOD_PATH=${D} modules_install The install doesn't specify base_libdir, so does the kernel make system honor it? - rm -f ${D}/lib/modules/${KERNEL_VERSION}/modules.order - rm -f ${D}/lib/modules/${KERNEL_VERSION}/modules.builtin - rm ${D}/lib/modules/${KERNEL_VERSION}/build - rm ${D}/lib/modules/${KERNEL_VERSION}/source + rm -f ${D}${base_libdir}/modules/${KERNEL_VERSION}/modules.order + rm -f ${D}${base_libdir}/modules/${KERNEL_VERSION}/modules.builtin + rm ${D}${base_libdir}/modules/${KERNEL_VERSION}/build + rm ${D}${base_libdir}/modules/${KERNEL_VERSION}/source if not, these will fail. -- Darren Hart Intel Open Source Technology Center Yocto Project - Technical Lead - Linux Kernel ___ Openembedded-core mailing list Openembedded-core@lists.openembedded.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-core
Re: [OE-core] proper(?) way to *require* an earlier, unsupported version of a recipe file?
On Mon, 2012-08-06 at 12:02 -0400, Robert P. J. Day wrote: what is the OE philosophy for preferring an earlier version of a recipe file that doesn't even exist anymore? for example, the watchdog recipe was recently upgraded from 5.11 to 5.12. in the process, the recipe file for 5.11 was removed; therefore, i interpret that as that OE-core no longer supports that version, which is fine. hypothetically, what if someone still *needed* that earlier version? would it be their responsibility to, say, create a new layer that re-introduced that older, now-unsupported version? just curious. Well, if there is some reason 5.11 is much better we as in OE-Core would like to know about it to evaluate if we should be keeping it around for some reason. If the reason you want that particular version is specific to you then yes, your own layer would be the place to maintain it (or a layer shared with other like minded users). Cheers, Richard ___ Openembedded-core mailing list Openembedded-core@lists.openembedded.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-core
Re: [OE-core] proper(?) way to *require* an earlier, unsupported version of a recipe file?
On Mon, 6 Aug 2012, Richard Purdie wrote: On Mon, 2012-08-06 at 12:02 -0400, Robert P. J. Day wrote: what is the OE philosophy for preferring an earlier version of a recipe file that doesn't even exist anymore? for example, the watchdog recipe was recently upgraded from 5.11 to 5.12. in the process, the recipe file for 5.11 was removed; therefore, i interpret that as that OE-core no longer supports that version, which is fine. hypothetically, what if someone still *needed* that earlier version? would it be their responsibility to, say, create a new layer that re-introduced that older, now-unsupported version? just curious. Well, if there is some reason 5.11 is much better we as in OE-Core would like to know about it to evaluate if we should be keeping it around for some reason. If the reason you want that particular version is specific to you then yes, your own layer would be the place to maintain it (or a layer shared with other like minded users). there was nothing about 5.11 that i see as superior, it's just that that general question came up last week in class and my admittedly off-the-cuff answer was that OE-core provided a well-defined, stable set of recipes to build on and if you *chose* to deviate in any way from that, that was kind of your problem and you had to deal with it. which appears to be pretty much what you're saying. thanks. rday -- Robert P. J. Day Ottawa, Ontario, CANADA http://crashcourse.ca Twitter: http://twitter.com/rpjday LinkedIn: http://ca.linkedin.com/in/rpjday ___ Openembedded-core mailing list Openembedded-core@lists.openembedded.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-core
Re: [OE-core] [PATCH 03/30] lsbsetup: use ${bindir} instead of /usr/bin for packaging
On 8/5/12 10:53 AM, Javier Martinez Canillas wrote: It is considered good practice to use the build system provided variables instead of directly specify hardcoded paths. Signed-off-by: Javier Martinez Canillas jav...@dowhile0.org --- meta/recipes-extended/lsb/lsbsetup_1.0.bb |4 ++-- 1 files changed, 2 insertions(+), 2 deletions(-) diff --git a/meta/recipes-extended/lsb/lsbsetup_1.0.bb b/meta/recipes-extended/lsb/lsbsetup_1.0.bb index 9172ee3..bda4589 100644 --- a/meta/recipes-extended/lsb/lsbsetup_1.0.bb +++ b/meta/recipes-extended/lsb/lsbsetup_1.0.bb @@ -11,9 +11,9 @@ S = ${WORKDIR} do_install() { # Only install file if it has a contents -install -d ${D}/usr/bin +install -d ${D}${bindir} the LSB required /usr/bin, I think. So this might be one of the few cases were specifying /usr/bin is acceptable. (but I'm not against the change) --Mark install -d ${D}/${sysconfdir} -install -m 0755 ${S}/LSB_Setup.sh ${D}/usr/bin +install -m 0755 ${S}/LSB_Setup.sh ${D}${bindir} install -d ${D}/${libdir}/lsb ln -sf ${base_sbindir}/chkconfig ${D}/${libdir}/lsb/install_initd ln -sf ${base_sbindir}/chkconfig ${D}/${libdir}/lsb/remove_initd ___ Openembedded-core mailing list Openembedded-core@lists.openembedded.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-core
Re: [OE-core] [PATCH v2 16/28] kernel.bbclass: use ${base_libdir} and ${sysconfdir} instead of /lib and /etc
On Mon, Aug 6, 2012 at 6:14 PM, Darren Hart dvh...@linux.intel.com wrote: On 08/05/2012 12:48 PM, Javier Martinez Canillas wrote: Hi Javier, It is considered good practice to use the build system provided variables instead of directly specify hardcoded paths. Have you tested this with a build using a base_libdir other than /lib ? Signed-off-by: Javier Martinez Canillas jav...@dowhile0.org --- meta/classes/kernel.bbclass | 12 ++-- 1 files changed, 6 insertions(+), 6 deletions(-) diff --git a/meta/classes/kernel.bbclass b/meta/classes/kernel.bbclass index 1d8dff9..b434093 100644 --- a/meta/classes/kernel.bbclass +++ b/meta/classes/kernel.bbclass @@ -109,10 +109,10 @@ kernel_do_install() { unset CFLAGS CPPFLAGS CXXFLAGS LDFLAGS MACHINE if (grep -q -i -e '^CONFIG_MODULES=y$' .config); then oe_runmake DEPMOD=echo INSTALL_MOD_PATH=${D} modules_install The install doesn't specify base_libdir, so does the kernel make system honor it? - rm -f ${D}/lib/modules/${KERNEL_VERSION}/modules.order - rm -f ${D}/lib/modules/${KERNEL_VERSION}/modules.builtin - rm ${D}/lib/modules/${KERNEL_VERSION}/build - rm ${D}/lib/modules/${KERNEL_VERSION}/source + rm -f ${D}${base_libdir}/modules/${KERNEL_VERSION}/modules.order + rm -f ${D}${base_libdir}/modules/${KERNEL_VERSION}/modules.builtin + rm ${D}${base_libdir}/modules/${KERNEL_VERSION}/build + rm ${D}${base_libdir}/modules/${KERNEL_VERSION}/source if not, these will fail. -- Darren Hart Intel Open Source Technology Center Yocto Project - Technical Lead - Linux Kernel Hi Darren, Yes, you are right I didn't think about it. The kernel is a special case since you just specify the root of the file system with INSTALL_MOD_PATH the kernel build system will *always* create lib/modules/${KERNEL_VERSION}. In fact now that I think about it, I don't know if you can even change the path were the kernel modules gets installed. So, yes the kernel is a special case and using a ${base_libdir} other than /lib will definitely break loadable kernel modules installation. Richard, Darren is correct and this patch is wrong, this is one of the cases which is acceptable (and necessary) to hardcode /lib. Thanks a lot for pointing me out this and sorry for not realizing this before sending the patch-set. Best regards, Javier ___ Openembedded-core mailing list Openembedded-core@lists.openembedded.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-core
Re: [OE-core] [PATCH 11/30] lsb: use ${base_bindir} instead of /bin for packaging
On 8/5/12 10:53 AM, Javier Martinez Canillas wrote: It is considered good practice to use the build system provided variables instead of directly specify hardcoded paths. Signed-off-by: Javier Martinez Canillas jav...@dowhile0.org --- meta/recipes-extended/lsb/lsb_1.4.bb |2 +- 1 files changed, 1 insertions(+), 1 deletions(-) diff --git a/meta/recipes-extended/lsb/lsb_1.4.bb b/meta/recipes-extended/lsb/lsb_1.4.bb index 15dbeaa..5734c5c 100644 --- a/meta/recipes-extended/lsb/lsb_1.4.bb +++ b/meta/recipes-extended/lsb/lsb_1.4.bb @@ -23,7 +23,7 @@ S = ${WORKDIR}/lsb-release-${PV} do_install(){ oe_runmake install prefix=${D} mandir=${D}/${datadir}/man/ DESTDIR=${D} - mkdir -p ${D}/bin + mkdir -p ${D}${base_bindir} mkdir -p ${D}/${baselib} mkdir -p ${D}/etc/lsb-release.d echo -n LSB_VERSION=\core-4.1-noarch: ${D}/etc/lsb-release This is another case with the lsb, that lsb-release (and a few other things) are documented to me installed into /bin. Not objecting to the change, just explaining why it was implemented this way. --Mark ___ Openembedded-core mailing list Openembedded-core@lists.openembedded.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-core
Re: [OE-core] [PATCH 17/30] linux-firware: use ${base_libdir} instead of /lib for packaging
On 08/05/2012 08:54 AM, Javier Martinez Canillas wrote: It is considered good practice to use the build system provided variables instead of directly specify hardcoded paths. The firmware location is explicitly set because this is where the Linux kernel requires it to be. This patch will break firmware loading. -- Darren Signed-off-by: Javier Martinez Canillas jav...@dowhile0.org --- .../linux-firmware/linux-firmware_git.bb | 34 ++-- 1 files changed, 17 insertions(+), 17 deletions(-) diff --git a/meta/recipes-kernel/linux-firmware/linux-firmware_git.bb b/meta/recipes-kernel/linux-firmware/linux-firmware_git.bb index a7e4ed6..c5ab173 100644 --- a/meta/recipes-kernel/linux-firmware/linux-firmware_git.bb +++ b/meta/recipes-kernel/linux-firmware/linux-firmware_git.bb @@ -35,50 +35,50 @@ do_compile() { } do_install() { - install -d ${D}/lib/firmware/ - cp -r * ${D}/lib/firmware/ + install -d ${D}${base_libdir}/firmware/ + cp -r * ${D}${base_libdir}/firmware/ # Libertas sd8686 - ln -sf libertas/sd8686_v9.bin ${D}/lib/firmware/sd8686.bin - ln -sf libertas/sd8686_v9_helper.bin ${D}/lib/firmware/sd8686_helper.bin + ln -sf libertas/sd8686_v9.bin ${D}${base_libdir}/firmware/sd8686.bin + ln -sf libertas/sd8686_v9_helper.bin ${D}${base_libdir}/firmware/sd8686_helper.bin # Realtek rtl8192* - install -m 0644 LICENCE.rtlwifi_firmware.txt ${D}/lib/firmware/rtlwifi/LICENCE.rtlwifi_firmware.txt + install -m 0644 LICENCE.rtlwifi_firmware.txt ${D}${base_libdir}/firmware/rtlwifi/LICENCE.rtlwifi_firmware.txt # fixup wl12xx location, after 2.6.37 the kernel searches a different location for it - ( cd ${D}/lib/firmware ; ln -sf ti-connectivity/* . ) + ( cd ${D}${base_libdir}/firmware ; ln -sf ti-connectivity/* . ) } PACKAGES =+ ${PN}-sd8686 ${PN}-rtl8192cu linux-firmware-rtl8192ce linux-firmware-rtl8192su ${PN}-wl12xx LICENSE_${PN}-sd8686 = Firmware:LICENSE.libertas FILES_${PN}-sd8686 = \ - /lib/firmware/libertas/sd8686_v9* \ - /lib/firmware/sd8686* \ - /lib/firmware/LICENCE.libertas \ + ${base_libdir}/firmware/libertas/sd8686_v9* \ + ${base_libdir}/firmware/sd8686* \ + ${base_libdir}/firmware/LICENCE.libertas \ LICENSE_${PN}-rtl8192cu = Firmware:LICENCE.rtlwifi_firmware FILES_${PN}-rtl8192cu = \ - /lib/firmware/rtlwifi/rtl8192cufw.bin \ - /lib/firmware/rtlwifi/LICENCE.rtlwifi_firmware.txt \ + ${base_libdir}/firmware/rtlwifi/rtl8192cufw.bin \ + ${base_libdir}/firmware/rtlwifi/LICENCE.rtlwifi_firmware.txt \ LICENSE_${PN}-rtl8192ce = Firmware:LICENCE.rtlwifi_firmware FILES_${PN}-rtl8192ce = \ - /lib/firmware/rtlwifi/rtl8192cfw.bin \ + ${base_libdir}/firmware/rtlwifi/rtl8192cfw.bin \ LICENSE_${PN}-rtl8192su = Firmware:LICENCE.rtlwifi_firmware FILES_${PN}-rtl8192su = \ - /lib/firmware/rtlwifi/rtl8712u.bin \ + ${base_libdir}/firmware/rtlwifi/rtl8712u.bin \ FILES_${PN}-wl12xx = \ - /lib/firmware/wl12* \ - /lib/firmware/TI* \ - /lib/firmware/ti-connectivity \ + ${base_libdir}/firmware/wl12* \ + ${base_libdir}/firmware/TI* \ + ${base_libdir}/firmware/ti-connectivity \ -FILES_${PN} += /lib/firmware/* +FILES_${PN} += ${base_libdir}/firmware/* -- Darren Hart Intel Open Source Technology Center Yocto Project - Technical Lead - Linux Kernel ___ Openembedded-core mailing list Openembedded-core@lists.openembedded.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-core
Re: [OE-core] [PATCH 17/30] linux-firware: use ${base_libdir} instead of /lib for packaging
On Mon, 2012-08-06 at 11:10 -0700, Darren Hart wrote: On 08/05/2012 08:54 AM, Javier Martinez Canillas wrote: It is considered good practice to use the build system provided variables instead of directly specify hardcoded paths. The firmware location is explicitly set because this is where the Linux kernel requires it to be. Is that actually true? I thought the kernel just supplied the leafname that it wanted and the knowledge about what directory to search was encoded in the hotplug helper scripts. This patch will break firmware loading. That might well be the case, though, unless the scripts have also been patched to respect ${base_libdir}. And, notwithstanding all the above, it's not entirely obvious that ${base_libdir} is semantically the right variable for things that aren't libraries. How does the udev recipe represent the patch to /lib/udev? p. ___ Openembedded-core mailing list Openembedded-core@lists.openembedded.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-core
Re: [OE-core] [PATCH 1/1] pseudo: Use fxstatat64 in unlinkat
On Thu, 2 Aug 2012 17:48:18 -0500 Peter Seebach peter.seeb...@windriver.com wrote: I think probably the right solution is to use the PSEUDO_STATBUF logic that's already in ports, and expand on it a bit. A followup: I've got a test version of this tagged PSEUDO_1_4_1. I've got one report about a cryptic error from tar (see yocto #2881 for some discussion), but would be really interested in feedback from anyone who is using 32-bit systems with this stuff. -s -- Listen, get this. Nobody with a good compiler needs to be justified. ___ Openembedded-core mailing list Openembedded-core@lists.openembedded.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-core
Re: [OE-core] [PATCH 4/5] gtk-icon-cache: call postinst scriplet at do_rootfs time
On Mon, Aug 6, 2012 at 11:18 AM, Laurentiu Palcu laurentiu.pa...@intel.com wrote: On 08/06/2012 11:10 AM, Andreas Müller wrote: On Mon, Aug 6, 2012 at 9:48 AM, Laurentiu Palcu laurentiu.pa...@intel.com wrote: On 08/06/2012 01:49 AM, Andreas Müller wrote: On Mon, Aug 6, 2012 at 12:30 AM, Andreas Müller schnitzelt...@googlemail.com wrote: On Sun, Aug 5, 2012 at 6:58 PM, Laurentiu Palcu laurentiu.pa...@intel.com wrote: You could give it a test yourselves and let me know your results. I will send a version 2 of the patchset(as soon as we all agree on the solution), with some changes suggested by Mark and some PR bumps suggested by Koen. With the image I usually work with [1] and AMD Phenom II X6 1090 16GB RAM I get a measurable delay - see attachment. I would not be happy loosing latest do_rootfs enhancements (off topic - thanks for that). Remeber we are only talking of gtk-update-icon-cache. OK I could buy an intel host and work just with sato images but... I suppose you could, but nobody asked you to do that, it's your choice what's your build machine or what you'll be building for. Thanks for the measurements though. They do, indeed, show quite a significant amount of time (around 6 minutes). A run-once solution is to be considered in this case. Andreas [1] https://gitorious.org/schnitzeltony-oe-meta/meta-misc/blobs/master/recipes-image/xfce-full-image.bb OK I know it is not that important: The image created with this patch series creates tons of messages like Why do you think is not important. Please elaborate. Or is it irony? Yes sorry - it was late night. It's OK, let's work together and find the best solution for all of us. I would also be pissed of if I had to wait an extra 6 minutes every do_rootfs run. I don't think is in anybody's benefit if you take this approach. :) All errors/warnings are important and they have to be taken care of. xfdesktop:446): GdkPixbuf-WARNING **: Cannot open pixbuf loader module file '/usr/lib/gdk-pixbuf-2.0/2.10.0/loaders.cache': No such file or directory and don't have icons at all. Did you run test that (on a hardware plattform different to your host)? I only tested on qemu. And it worked just fine, without any errors. With all icons in place. Which hardware did you emulate? My tests were done for overo (ARM cortext A8). I wonder if the created database is machine specific. I used qemuarm. As far as I know, the database shouldn't be machine specific. And, looking at the log you sent, it looks like all commands succeeded. Otherwise, the 'time' application would also signal in the log file if the command failed. Could please have a look, on your host (xfce-full-image-1.0-r0/rootfs/usr/share/icons/hicolor or xfce-full-image-1.0-r0/rootfs/usr/share/icons/gnome), to see if it creates the icon-theme.cache files? It does for me and I don't understand why it doesn't in your case... Thanks, Laurentiu I checked: In both I find the files icon-theme.cache. Andreas ___ Openembedded-core mailing list Openembedded-core@lists.openembedded.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-core
Re: [OE-core] [PATCH V3 00/11] Mesa upgrade/improvements
On Mon, 2012-08-06 at 15:04 +0100, Richard Purdie wrote: Ok, we need to do something about this mess, period. Its not going to get any better and we need to sort it out. I can't force changes in but on the other hand I'm not going to tie the hands of everyone just because the ARM world in particular is a total mess in this area. As such, I'm likely going to allow architecture overrides in core mesa so for example, IA can enable gles and egl for everyone and cleanup their binary blobs to override them on platforms where it makes sense. I suspect someone will start sending patches to enable a generic arm core and I am going to have a *really* hard time refusing the patches due to any one silicon provider's binary blob issues. It's not obvious to me that this is really an architecture-level problem or that having different Mesa feature sets enabled on different architectures is a positive thing. I'm also not totally convinced that any insurmountable problem actually exists today with evil vendor blobs. If they install themselves into some directory that isn't /usr/lib and arrange for the dynamic linker to find their libEGL.so (etc) ahead of the Mesa one then it seems like everything should just work with today's technology. As you observed, sort of the whole point of the EGL/GL/GLES abstraction is that you have a common API/ABI to write to irrespective of the implementation. (Realistically, it seems fairly unlikely that there are very many situations where using Mesa libGL on top of a vendor libEGL is going to make much sense. If you have vendor libEGL but not libGL then the chances are you have GLES instead and this is probably what you want to be using.) So, all in all, I tend to think that we should just apply these patches and then do whatever else is necessary for the evil-binary-vendor stuff to be correctly packageable. p. ___ Openembedded-core mailing list Openembedded-core@lists.openembedded.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-core
[OE-core] [PATCH] iputils: Break libsysfs dependency
iputils drops a /bin/arping with a runtime linkage against libsysfs in /usr. Port Fedora 17 iputils-20071127-infiniband.patch, which inlines access previously done by libsysfs. Signed-off-by: Andy Ross andy.r...@windriver.com --- .../files/arping-break-libsysfs-dependency.patch | 296 + meta/recipes-extended/iputils/iputils_s20101006.bb | 5 +- 2 files changed, 299 insertions(+), 2 deletions(-) create mode 100644 meta/recipes-extended/iputils/files/arping-break-libsysfs-dependency.patch diff --git a/meta/recipes-extended/iputils/files/arping-break-libsysfs-dependency.patch b/meta/recipes-extended/iputils/files/arping-break-libsysfs-dependency.patch new file mode 100644 index 000..37325ee --- /dev/null +++ b/meta/recipes-extended/iputils/files/arping-break-libsysfs-dependency.patch @@ -0,0 +1,296 @@ +arping: Break libsysfs dependency + +Port Fedora 17's iputils-20071127-infiniband.patch, which inlines sysfs +access previously done by libsysfs (which is in /usr, and thus not +usable by /bin/arping). + +Upstream-Status: Inappropriate [not author] +Signed-off-by: Andy Ross andy.r...@windriver.com +--- + Makefile |2 +- + arping.c | 152 -- + 2 files changed, 109 insertions(+), 45 deletions(-) + +diff --git a/Makefile b/Makefile +index d9a5ca5..943f048 100644 +--- a/Makefile b/Makefile +@@ -27,7 +27,7 @@ all: $(TARGETS) + + + tftpd: tftpd.o tftpsubs.o +-arping: arping.o -lsysfs ++arping: arping.o + ping: ping.o ping_common.o + ping6: ping6.o ping_common.o -lresolv -lcrypto + ping.o ping6.o ping_common.o: ping_common.h +diff --git a/arping.c b/arping.c +index 13484de..6379354 100644 +--- a/arping.c b/arping.c +@@ -32,8 +32,6 @@ + #include netinet/in.h + #include arpa/inet.h + +-#include sysfs/libsysfs.h +- + #include SNAPSHOT.h + + static void usage(void) __attribute__((noreturn)); +@@ -52,14 +50,22 @@ int unicasting; + int s; + int broadcast_only; + +-struct sockaddr_storage me; +-struct sockaddr_storage he; ++struct sockaddr_ll *me=NULL; ++struct sockaddr_ll *he=NULL; + + struct timeval start, last; + + int sent, brd_sent; + int received, brd_recv, req_recv; + ++#define SYSFS_MNT_PATH /sys ++#define SYSFS_CLASS class ++#define SYSFS_NET net ++#define SYSFS_BROADCAST broadcast ++#define SYSFS_PATH_ENV SYSFS_PATH ++#define SYSFS_PATH_LEN 256 ++#define SOCKADDR_LEN (2 * sizeof(struct sockaddr_ll)) ++ + #define MS_TDIFF(tv1,tv2) ( ((tv1).tv_sec-(tv2).tv_sec)*1000 + \ + ((tv1).tv_usec-(tv2).tv_usec)/1000 ) + +@@ -166,6 +172,10 @@ void finish(void) + printf(\n); + fflush(stdout); + } ++ ++ free(me); ++ free(he); ++ + if (dad) + exit(!!received); + if (unsolicited) +@@ -189,8 +199,7 @@ void catcher(void) + finish(); + + if ( count!=0 (last.tv_sec==0 || MS_TDIFF(tv,last) 500 ) ) { +- send_pack(s, src, dst, +-(struct sockaddr_ll *)me, (struct sockaddr_ll *)he); ++ send_pack(s, src, dst, me, he); + if (count = 0) + count--; + if (count == 0 unsolicited) +@@ -239,7 +248,7 @@ int recv_pack(unsigned char *buf, int len, struct sockaddr_ll *FROM) + return 0; + if (ah-ar_pln != 4) + return 0; +- if (ah-ar_hln != ((struct sockaddr_ll *)me)-sll_halen) ++ if (ah-ar_hln != me-sll_halen) + return 0; + if (len sizeof(*ah) + 2*(4 + ah-ar_hln)) + return 0; +@@ -250,7 +259,7 @@ int recv_pack(unsigned char *buf, int len, struct sockaddr_ll *FROM) + return 0; + if (src.s_addr != dst_ip.s_addr) + return 0; +- if (memcmp(p+ah-ar_hln+4, ((struct sockaddr_ll *)me)-sll_addr, ah-ar_hln)) ++ if (memcmp(p+ah-ar_hln+4, me-sll_addr, ah-ar_hln)) + return 0; + } else { + /* DAD packet was: +@@ -268,7 +277,7 @@ int recv_pack(unsigned char *buf, int len, struct sockaddr_ll *FROM) +*/ + if (src_ip.s_addr != dst.s_addr) + return 0; +- if (memcmp(p, ((struct sockaddr_ll *)me)-sll_addr, ((struct sockaddr_ll *)me)-sll_halen) == 0) ++ if (memcmp(p, me-sll_addr, me-sll_halen) == 0) + return 0; + if (src.s_addr src.s_addr != dst_ip.s_addr) + return 0; +@@ -284,7 +293,7 @@ int recv_pack(unsigned char *buf, int len, struct sockaddr_ll *FROM) + printf(for %s , inet_ntoa(dst_ip)); + s_printed = 1; + } +- if (memcmp(p+ah-ar_hln+4, ((struct sockaddr_ll *)me)-sll_addr, ah-ar_hln)) { ++ if (memcmp(p+ah-ar_hln+4, me-sll_addr, ah-ar_hln)) { +
[OE-core] [oe-core] [denzil branch] Update to the latest version of psplash
In the latest version of psplash (http://git.yoctoproject.org/cgit/cgit.cgi/psplash/) there are two patches that are not in the version of psplash used in the oe-core denzil branch. The first patch called Make it easier to customise colourshttp://git.yoctoproject.org/cgit/cgit.cgi/psplash/commit/?id=84764337a584002a92940323d374b0e417c573a6 moves some color definitions from being hardcoded into psplash.c and puts it in a new header file called psplash-colors.h. The second patch Fix for psplash segmentation faulthttp://git.yoctoproject.org/cgit/cgit.cgi/psplash/commit/?id=de9979aefbc56af59b4d236a4b63dd19dcdcfb53 fixes a segment fault issue. What I would like to do is update the psplash_git.bb recipe to use the latest version of psplash that pulls in these two tweaks. The second patch has zero impact on existing users of the pslash recipe. If a particular layer tweaks the color definitions in psplash by patching the psplash.c file then the second patch mentioned above could result in a minor tweak having to be made to those layers to address the changes. In oe-classic noticed that there is a patch for psplash (http://cgit.openembedded.org/openembedded/tree/recipes/psplash/psplash-ti/0001-configurability-for-rev-422.patch) that patches psplash in almost the same exact way as the Make it easier to customize colours patch. Are there any objections in bumping the version of psplash to incorporate the above mentioned patches? I haven't seen any layer adjust the psplash color definitions so I doubt there truly being any impact at all. Regards, Franklin Cooper Jr. Texas Instruments Application Engineer fcoo...@ti.commailto:fcoo...@ti.com ___ Openembedded-core mailing list Openembedded-core@lists.openembedded.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-core
Re: [OE-core] [oe-core] [denzil branch] Update to the latest version of psplash
On 08/06/2012 02:43 PM, Cooper Jr., Franklin wrote: In the latest version of psplash (http://git.yoctoproject.org/cgit/cgit.cgi/psplash/) there are two patches that are not in the version of psplash used in the oe-core denzil branch. The first patch called Make it easier to customise colours http://git.yoctoproject.org/cgit/cgit.cgi/psplash/commit/?id=84764337a584002a92940323d374b0e417c573a6moves some color definitions from being hardcoded into psplash.c and puts it in a new header file called psplash-colors.h. The second patch Fix for psplash segmentation fault http://git.yoctoproject.org/cgit/cgit.cgi/psplash/commit/?id=de9979aefbc56af59b4d236a4b63dd19dcdcfb53fixes a segment fault issue. What I would like to do is update the psplash_git.bb recipe to use the latest version of psplash that pulls in these two tweaks. The second patch has zero impact on existing users of the pslash recipe. If a particular layer tweaks the color definitions in psplash by patching the psplash.c file then the second patch mentioned above could result in a minor tweak having to be made to those layers to address the changes. In oe-classic noticed that there is a patch for psplash (http://cgit.openembedded.org/openembedded/tree/recipes/psplash/psplash-ti/0001-configurability-for-rev-422.patch) that patches psplash in almost the same exact way as the “Make it easier to customize colours” patch. Are there any objections in bumping the version of psplash to incorporate the above mentioned patches? I haven’t seen any layer adjust the psplash color definitions so I doubt there truly being any impact at all.** Franklin and I discussed this on IRC and I suggested he bring this up on the mailing list. I'm inclined to take and backport the segfault patch into the current version of psplash in denzil (which is a git recipe fixed at e05374a). Bumping the SRCREV of the recipe and including the color definitions patch I'm not so sure about - vocal support from the community could tip the scales, though, so speak up if this would be of value to you. The segfault patch is being tracked with bug #2903: https://bugzilla.yoctoproject.org/show_bug.cgi?id=2903 Thanks, Scott -- Scott Garman Embedded Linux Engineer - Yocto Project Intel Open Source Technology Center ___ Openembedded-core mailing list Openembedded-core@lists.openembedded.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-core
Re: [OE-core] [oe-core] [denzil branch] Update to the latest version of psplash
On Mon, Aug 6, 2012 at 6:54 PM, Scott Garman scott.a.gar...@intel.com wrote: I'm inclined to take and backport the segfault patch into the current version of psplash in denzil (which is a git recipe fixed at e05374a). Bumping the SRCREV of the recipe and including the color definitions patch I'm not so sure about - vocal support from the community could tip the scales, though, so speak up if this would be of value to you. I do think it is valuable however layers changing colors will have a broken patch applying above it and I don't think it is the expected for a stable release. -- Otavio Salvador O.S. Systems E-mail: ota...@ossystems.com.br http://www.ossystems.com.br Mobile: +55 53 9981-7854 http://projetos.ossystems.com.br ___ Openembedded-core mailing list Openembedded-core@lists.openembedded.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-core
[OE-core] [PATCH] image-mklibs: pass correct libdir to mklibs
libdir should be specified, or else mklibs won't work for 64bit targets. It wouldn't be able to find the libs. Traceback (most recent call last): File build/bitbake_build/tmp/sysroots/i686-linux/usr/bin/x86_64-wrs-linux/mklibs, line 553, in module header = elf_header(find_lib(libraries.copy().pop())) File build/bitbake_build/tmp/sysroots/i686-linux/usr/bin/x86_64-wrs-linux/mklibs, line 89, in elf_header raise Exception(Cannot find lib: + obj) Exception: Cannot find lib: Signed-off-by: Jesse Zhang sen.zh...@windriver.com --- meta/classes/image-mklibs.bbclass |1 + 1 files changed, 1 insertions(+), 0 deletions(-) diff --git a/meta/classes/image-mklibs.bbclass b/meta/classes/image-mklibs.bbclass index 7623815..66b0f52 100644 --- a/meta/classes/image-mklibs.bbclass +++ b/meta/classes/image-mklibs.bbclass @@ -38,6 +38,7 @@ mklibs_optimize_image_doit() { mklibs -v \ --ldlib ${dynamic_loader} \ + --libdir ${baselib} \ --sysroot ${PKG_CONFIG_SYSROOT_DIR} \ --root ${IMAGE_ROOTFS} \ --target `echo ${TARGET_PREFIX} | sed 's/-$//' ` \ -- 1.7.7.6 ___ Openembedded-core mailing list Openembedded-core@lists.openembedded.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-core
[OE-core] [PATCH] fix mklibs error for 64bit targets
mklibs should be passed a --libdir option, in order for it to work with 64bit targets. Jesse Zhang (1): image-mklibs: pass correct libdir to mklibs meta/classes/image-mklibs.bbclass |1 + 1 files changed, 1 insertions(+), 0 deletions(-) -- 1.7.7.6 ___ Openembedded-core mailing list Openembedded-core@lists.openembedded.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-core
Re: [OE-core] [PATCH][RFC] u-boot: Upgrade to upstream stable 2012.07
On Mon, Aug 6, 2012 at 7:13 AM, Wolfgang Denk w...@denx.de wrote: Dear Radu, In message 501fd013.5040...@intel.com you wrote: CROSS_COMPILE is set in EXTRA_OEMAKE which is used in oe-runmake() within do_compile() Running devshell is achieved in do_devshell() which is run after do_patch() but before do_compile(), that's why in devshell CROSS_COMPILE is not (yet) set. But CROSS_COMPILE _is_ set (and apparently has the value i386-linux-. The problem is that PATH is not set such that it contains any such file as i386-linux-gcc. Best regards, Wolfgang Denk -- DENX Software Engineering GmbH, MD: Wolfgang Denk Detlev Zundel HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany Phone: (+49)-8142-66989-10 Fax: (+49)-8142-66989-80 Email: w...@denx.de Beware of bugs in the above code; I have only proved it correct, not tried it. - Donald Knuth ___ Openembedded-core mailing list Openembedded-core@lists.openembedded.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-core Hi Wolfgang If the fix proposed here http://lists.denx.de/pipermail/u-boot/2012-August/129828.html is accepted into u-boot then we wont need these glues. ___ Openembedded-core mailing list Openembedded-core@lists.openembedded.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-core
Re: [OE-core] [PATCH V2 0/1] linux-yocto-custom: Clarify usage and clear COMPATIBLE_MACHINE
On 12-08-06 11:56 AM, Darren Hart wrote: The following changes since commit 47aca0f0f79c30d1df1f89c710d3e354f436145d: subversion: Add missing build dependency on sqlite3 (2012-08-06 16:13:45 +0100) are available in the git repository at: git://git.yoctoproject.org/user-contrib/dvhart/oe-core skeleton http://git.yoctoproject.org/cgit.cgi/user-contrib/dvhart/oe-core/log/?h=skeleton Darren Hart (1): linux-yocto-custom: Clarify usage and clear COMPATIBLE_MACHINE Looks good to me. Acked-by: Bruce Ashfield bruce.ashfi...@windriver.com .../recipes-kernel/linux/linux-yocto-custom.bb | 53 +++--- 1 file changed, 36 insertions(+), 17 deletions(-) ___ Openembedded-core mailing list Openembedded-core@lists.openembedded.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-core
Re: [OE-core] [PATCH][RFC] u-boot: Upgrade to upstream stable 2012.07
I meant, set with the correct value. i386-linux- is preloaded into CROSS_COMPILE and then at do_compile() gcc is appended to it. Is there in interest, to actually build u-boot in devshell? Radu On 08/06/2012 05:13 PM, Wolfgang Denk wrote: Dear Radu, In message 501fd013.5040...@intel.com you wrote: CROSS_COMPILE is set in EXTRA_OEMAKE which is used in oe-runmake() within do_compile() Running devshell is achieved in do_devshell() which is run after do_patch() but before do_compile(), that's why in devshell CROSS_COMPILE is not (yet) set. But CROSS_COMPILE _is_ set (and apparently has the value i386-linux-. The problem is that PATH is not set such that it contains any such file as i386-linux-gcc. Best regards, Wolfgang Denk ___ Openembedded-core mailing list Openembedded-core@lists.openembedded.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-core