- dev - example.com

[PATCH 06/15] mk/image.mk: Fix calling mkfs.jffs2
by Phil Sutter 08 Mar '17

08 Mar '17

This command does not reside in ${STAGING_HOST_DIR}/usr/bin (anymore?), also setting PATH='${HOST_PATH}' prior to just calling the command without path is more elegant in my opinion. Signed-off-by: Phil Sutter <phil(a)nwl.cc> --- mk/image.mk | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/mk/image.mk b/mk/image.mk index e0851418e4c8c..6b704f200938c 100644 --- a/mk/image.mk +++ b/mk/image.mk @@ -188,7 +188,7 @@ ${BUILD_DIR}/root.squashfs: ${TARGET_DIR}/.adk -nopad -noappend -root-owned $(MAKE_TRACE) ${FW_DIR}/${ROOTFSJFFS2}: ${TARGET_DIR} - ${STAGING_HOST_DIR}/usr/bin/mkfs.jffs2 $(ADK_JFFS2_OPTS) -q -r ${TARGET_DIR} \ + PATH='${HOST_PATH}' mkfs.jffs2 $(ADK_JFFS2_OPTS) -q -r ${TARGET_DIR} \ --pad=$(ADK_TARGET_MTD_SIZE) -o ${FW_DIR}/${ROOTFSJFFS2} $(MAKE_TRACE) createinitramfs: ${STAGING_TARGET_DIR}/${INITRAMFS}_list -- 2.9.3

1 0

[PATCH 05/15] package/grub: Fix grub.cfg
by Phil Sutter 08 Mar '17

08 Mar '17

When trying qemu-x86_64 target with full disk image, grub wasn't able to read the rootfs. After comparing it's config with the one used on my notebook, I was able to get it right. Apart from fixing the failure, this also enables serial output so grub menu becomes accessible when calling qemu with -nographic option. Signed-off-by: Phil Sutter <phil(a)nwl.cc> --- package/grub/files/grub.cfg | 10 ++++++++-- 1 file changed, 8 insertions(+), 2 deletions(-) diff --git a/package/grub/files/grub.cfg b/package/grub/files/grub.cfg index d87a52cacced4..8d20ab2435180 100644 --- a/package/grub/files/grub.cfg +++ b/package/grub/files/grub.cfg @@ -1,8 +1,14 @@ +serial --speed=115200 --unit=0 --word=8 --parity=no --stop=1 +terminal_input serial +terminal_output serial + set default=0 set timeout=3 menuentry "OpenADK" { + insmod part_msdos + insmod ext2 + set root='hd0,msdos1' echo "Loading OpenADK" - set root=(hd0,2) - linux (hd0,2)/boot/kernel root=/dev/sda2 rootfstype=ext4 rootwait panic=10 + linux16 /boot/kernel root=/dev/sda1 rootfstype=ext4 rootwait panic=10 } -- 2.9.3

1 0

[PATCH 04/15] target: ADK_TARGET_ROOTFS_GENIMAGE fails without ADK_TARGET_QEMU_WITH_BOOTLOADER
by Phil Sutter 08 Mar '17

08 Mar '17

Therefore make the former select the latter. Signed-off-by: Phil Sutter <phil(a)nwl.cc> --- target/config/Config.in.rootfs | 1 + 1 file changed, 1 insertion(+) diff --git a/target/config/Config.in.rootfs b/target/config/Config.in.rootfs index 096293e45f1b5..c24535f013617 100644 --- a/target/config/Config.in.rootfs +++ b/target/config/Config.in.rootfs @@ -163,6 +163,7 @@ config ADK_TARGET_ROOTFS_GENIMAGE select ADK_KERNEL_EXT4_FS select ADK_KERNEL_VFAT_FS select ADK_TARGET_QEMU_WITH_BLOCK if ADK_TARGET_QEMU + select ADK_TARGET_QEMU_WITH_BOOTLOADER if ADK_TARGET_QEMU select ADK_HOST_BUILD_GENIMAGE help Create a complete disk image for the target with genimage. -- 2.9.3

1 0

[PATCH 03/15] toolchain: gcc: 5.4.0: fix patch formatting
by Phil Sutter 08 Mar '17

08 Mar '17

git-am is quite picky when it comes to commit messages, this one needed fixing. Signed-off-by: Phil Sutter <phil(a)nwl.cc> --- toolchain/gcc/patches/5.4.0/microblaze-enable-dwarf-eh-support.patch | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/toolchain/gcc/patches/5.4.0/microblaze-enable-dwarf-eh-support.patch b/toolchain/gcc/patches/5.4.0/microblaze-enable-dwarf-eh-support.patch index 9d29090a2a876..b9ad32fe8311d 100644 --- a/toolchain/gcc/patches/5.4.0/microblaze-enable-dwarf-eh-support.patch +++ b/toolchain/gcc/patches/5.4.0/microblaze-enable-dwarf-eh-support.patch @@ -1,5 +1,3 @@ -Fetched from Xilinx gcc git at https://github.com/Xilinx/gcc - From 23c35173490ac2d6348a668dfc9c1a6eb62171f2 Mon Sep 17 00:00:00 2001 From: "Edgar E. Iglesias" <edgar.iglesias(a)gmail.com> Date: Mon, 18 Jun 2012 20:18:13 +0200 @@ -26,6 +24,8 @@ Changelog Signed-off-by: David Holsgrove <david.holsgrove(a)xilinx.com> Signed-off-by: Edgar E. Iglesias <edgar.iglesias(a)gmail.com> --- +Fetched from Xilinx gcc git at https://github.com/Xilinx/gcc +--- gcc/common/config/microblaze/microblaze-common.c | 3 --- gcc/config/microblaze/microblaze-protos.h | 1 + gcc/config/microblaze/microblaze.c | 29 ++++++++++++++++++++---- -- 2.9.3

1 0

[PATCH 02/15] package/asterisk: Fix build
by Phil Sutter 08 Mar '17

08 Mar '17

Looks like there is a bug in AST_INLINE_API macro voodoo. Fix taken from here: https://issues.asterisk.org/jira/browse/ASTERISK-25503 Signed-off-by: Phil Sutter <phil(a)nwl.cc> --- package/asterisk/Makefile | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/package/asterisk/Makefile b/package/asterisk/Makefile index 3367cce949014..2504e1abc0644 100644 --- a/package/asterisk/Makefile +++ b/package/asterisk/Makefile @@ -111,7 +111,7 @@ INSTALL_STYLE:= manual TARGET_CFLAGS+= -fgnu89-inline TARGET_LDFLAGS+= -lm -ltinfo -ldl -lpthread MAKE_ENV= ASTCFLAGS="${TARGET_CPPFLAGS} -DHAVE_STRTOQ" \ - ASTLDFLAGS="${TARGET_LDFLAGS}" + ASTLDFLAGS="${TARGET_LDFLAGS} -z muldefs" MAKE_FLAGS+= DESTDIR="$(WRKINST)" \ OPTIMIZE="${TARGET_CFLAGS}" \ NOISY_BUILD=1 -- 2.9.3

1 0

[PATCH 01/15] package/screen: fix build
by Phil Sutter 08 Mar '17

08 Mar '17

The problem is, sched.h is found in the local directory and therefore falsely included from system headers. This probably has something to do with openadk relying on builtin header inclusion in gcc instead of specifying them explicitly. Signed-off-by: Phil Sutter <phil(a)nwl.cc> --- package/screen/Makefile | 1 + package/screen/patches/patch-screen_h | 11 +++++++++++ 2 files changed, 12 insertions(+) create mode 100644 package/screen/patches/patch-screen_h diff --git a/package/screen/Makefile b/package/screen/Makefile index a46433940cc7a..090e25b5713f8 100644 --- a/package/screen/Makefile +++ b/package/screen/Makefile @@ -31,6 +31,7 @@ pre-build: echo '#undef GETUTENT' >> ${WRKBUILD}/config.h echo '#undef UTHOST' >> ${WRKBUILD}/config.h echo '#undef BUGGYGETLOGIN' >> ${WRKBUILD}/config.h + mv ${WRKBUILD}/sched.h ${WRKBUILD}/my_sched.h screen-install: ${INSTALL_DIR} ${IDIR_SCREEN}/usr/bin diff --git a/package/screen/patches/patch-screen_h b/package/screen/patches/patch-screen_h new file mode 100644 index 0000000000000..ef6e5ea54a189 --- /dev/null +++ b/package/screen/patches/patch-screen_h @@ -0,0 +1,11 @@ +--- screen-4.2.1.orig/screen.h 2014-04-28 00:08:06.000000000 +0200 ++++ screen-4.2.1/screen.h 2015-02-19 02:36:51.299000393 +0100 +@@ -43,7 +43,7 @@ + #include "osdef.h" + + #include "ansi.h" +-#include "sched.h" ++#include "my_sched.h" + #include "acls.h" + #include "comm.h" + #include "layer.h" -- 2.9.3

1 0

[PATCH 00/15] UBIFS and a bunch of fixes
by Phil Sutter 08 Mar '17

08 Mar '17

This series contains some fixes for issues I stumbled upon as well as some quite untested UBIFS rootfs support. Phil Sutter (15): package/screen: fix build package/asterisk: Fix build toolchain: gcc: 5.4.0: fix patch formatting target: ADK_TARGET_ROOTFS_GENIMAGE fails without ADK_TARGET_QEMU_WITH_BOOTLOADER package/grub: Fix grub.cfg mk/image.mk: Fix calling mkfs.jffs2 target: Fix for incorrect post make instructions mk/host-bottom: Introduce HOST_CONFIG_STYLE mk/host-bottom: Introduce HOST_INSTALL_STYLE package/zlib: Enable host-package package/mtd-utils: Enable UBIFS support in host build target/linux/config: Allow enabling UBIFS support Introduce kernel2minor package target/config: Allow targets to require kernel2minor host package Implement UBIFS rootfs support mk/host-bottom.mk | 20 +++++++----- mk/image.mk | 21 +++++++++++- mk/rootfs.mk | 1 + package/asterisk/Makefile | 2 +- package/grub/files/grub.cfg | 10 ++++-- package/kernel2minor/Makefile | 38 ++++++++++++++++++++++ package/mtd-utils/Makefile | 14 ++++---- package/screen/Makefile | 1 + package/screen/patches/patch-screen_h | 11 +++++++ package/zlib/Makefile | 24 ++++++++++++++ target/config/Config.in.rootfs | 10 ++++++ target/config/Config.in.tools | 9 +++++ target/linux/config/Config.in.flash | 23 ++++++++++++- target/linux/config/Config.in.fs | 14 ++++++++ target/mips/Makefile | 8 +++++ target/x86/Makefile | 3 +- target/x86_64/Makefile | 1 + .../5.4.0/microblaze-enable-dwarf-eh-support.patch | 4 +-- 18 files changed, 192 insertions(+), 22 deletions(-) create mode 100644 package/kernel2minor/Makefile create mode 100644 package/screen/patches/patch-screen_h -- 2.9.3

1 0

new website
by Waldemar Brodkorb 12 Feb '17

12 Feb '17

Hi, I updated the websites for my embedded linux projects today. If you have any issues to access anything, please sent me an email. I got bored by Trac and now static websites generated with hugo, and gogs for wiki/issue/git is in use. The cgit interface and mailman are still the same. I switched to the faster git http backend for cloning via http/https. I moved on to new certbot and all websites are forced to https. Some minor content (C library matric/ Debugging hints will be added soon). Thanks to Simone for my new black labrador logo abstracted from a real photo of my dog. best regards Waldemar

1 0

Re: [openadk] Moving from defconfig to custom config - kernel doesn't build anymore
by Waldemar Brodkorb 13 Jan '17

13 Jan '17

Hi Diez, Diez B. Roggisch wrote, > Hi Maria, Waldemar, ^^o > I’ve put the two different .config-files from respective builds in here: > > https://www.dropbox.com/s/bjedqw7890f6fml/kernelconfigs.tgz?dl=0 > > They do indeed differ a lot, but see below for more conclusions. > > > > Can you please post the error message. This would help me to understand > what > exactly went wrong in this case. > > > I don’t understand. I’ve postet the actual compiler error inline here, and the > make.log via a shared Dropbox-link. What exact information do you miss? Sorry, I dropped some information from the mail when adding Mario to cc. Unfortunately we have no mailinglist archives for dev(a)openadk.org, as I had some misconfiguration of the mailing list. I will bounce mario your first mail. best regards Waldemar

2 3

test mailinglist
by Waldemar Brodkorb 25 Jul '15

25 Jul '15

1 0