Skip to contentSkip to conversation

Systemd Upgrade

PATCH Superseded Khem Raj 4 participants 24 messages Sep 6, 2015 – Sep 10, 2015 exec and binfmt

v1 of 2 Superseded Latest is v2 →

lore
Share & tools
lore

Khem Raj

author
#0 lore raw

cover Systemd Upgrade

Changes Since last time
- Now that 225 is out upgrade recipe to 225, changes between 224 and 225 are minor but fixes in networkd are important
- Implement systemd-sysv-install for OE
- Enable getopt in busybox needed for systemd-sysv-install
- Remove update-rc.d from default uninstall list in rootfs creation
- Tested date and systemd unit tests manually to pass
- Booted core-image-cmdline-full on all qemu targets

Khem Raj (4):
busybox: Enable getopt applet
rootfs: Do not uninstall update-rc.d
systemd: Upgrade 219 -> 225
systemd: Implement OE-Specific systemd-sysv-install

 meta/lib/oe/rootfs.py                              |   3 +-
 meta/recipes-core/busybox/busybox/getopts.cfg      |   3 +
 meta/recipes-core/busybox/busybox_1.23.2.bb        |   1 +
 ...-mount-add-dependencies-to-dynamically-mo.patch |  36 ----
 ...ed-missing.h-fall-back-to-insecure-getenv.patch |  25 ---
 ...t-install-dependency-links-at-install-tim.patch |  35 ++--
 ...heck-for-additional-features-that-uclibc-.patch |  23 ++-
 ...wn-Use-execvpe-only-when-libc-supports-it.patch |  37 ++---
 ...til-Use-mkostemp-only-if-libc-supports-it.patch |   8 +-
 ...-unimplemented-_SC_PHYS_PAGES-system-conf.patch |   8 +-
 ...0-Make-root-s-home-directory-configurable.patch | 136 ++++++++-------
 ...0012-implment-systemd-sysv-install-for-OE.patch |  40 +++++
 ...files.c-Honor-ordering-within-files-as-th.patch | 185 ---------------------
 ...-remove-userspace-firmware-loading-suppor.patch |  99 ++++++-----
 .../systemd/systemd/tmpfiles-pam.patch             |  30 ----
 .../systemd/{systemd_219.bb => systemd_225.bb}     |  29 ++--
 16 files changed, 222 insertions(+), 476 deletions(-)
 create mode 100644 meta/recipes-core/busybox/busybox/getopts.cfg
 delete mode 100644 meta/recipes-core/systemd/systemd/0001-Revert-core-mount-add-dependencies-to-dynamically-mo.patch
 delete mode 100644 meta/recipes-core/systemd/systemd/0002-shared-missing.h-fall-back-to-insecure-getenv.patch
 create mode 100644 meta/recipes-core/systemd/systemd/0012-implment-systemd-sysv-install-for-OE.patch
 delete mode 100644 meta/recipes-core/systemd/systemd/0012-systemd-tmpfiles.c-Honor-ordering-within-files-as-th.patch
 delete mode 100644 meta/recipes-core/systemd/systemd/tmpfiles-pam.patch
 rename meta/recipes-core/systemd/{systemd_219.bb => systemd_225.bb} (94%)
Signature · 1 line
2.5.1
author
#1 lore raw

PATCH 1/4 busybox: Enable getopt applet

getopts is a common applet more so now needed by systemd for working with sysv scripts

---
 meta/recipes-core/busybox/busybox/getopts.cfg | 3 +++
 meta/recipes-core/busybox/busybox_1.23.2.bb   | 1 +
 2 files changed, 4 insertions(+)
 create mode 100644 meta/recipes-core/busybox/busybox/getopts.cfg
meta/recipes-core/busybox/busybox/getopts.cfg new file +3 -0
new file mode 100644
index 0000000..8db0a4a
@@ -0,0 +1,3 @@
+CONFIG_GETOPT=y
+CONFIG_LONG_OPTS=y
+CONFIG_FEATURE_GETOPT_LONG=y
meta/recipes-core/busybox/busybox_1.23.2.bb +1 -0
index 8a4bae6..e4d9f97 100644
@@ -35,6 +35,7 @@ SRC_URI = "http://www.busybox.net/downloads/busybox-${PV}.tar.bz2;name=tarball \
            file://mount-via-label.cfg \
            file://sha1sum.cfg \
            file://sha256sum.cfg \
+           file://getopts.cfg \
 "
 
 SRC_URI[tarball.md5sum] = "7925683d7dd105aabe9b6b618d48cc73"
Signature · 1 line
2.5.1
author
#2 lore raw

PATCH 2/4 rootfs: Do not uninstall update-rc.d

update-rc.d is now needed by systemd to interact with sysv scripts, so if we have a mixed system, then we can not uninstall update-rc.d as it is going to break systemd functionality

---
 meta/lib/oe/rootfs.py | 3 +--
 1 file changed, 1 insertion(+), 2 deletions(-)
meta/lib/oe/rootfs.py +1 -2
index c29843b..4ef16c4 100644
@@ -238,8 +238,7 @@ class Rootfs(object):
                     pkgs_installed = installed_pkgs.read().splitlines()
                     for pkg_installed in pkgs_installed[:]:
                         pkg = pkg_installed.split()[0]
-                        if pkg in ["update-rc.d",
-                                "base-passwd",
+                        if pkg in ["base-passwd",
                                 "shadow",
                                 "update-alternatives",
                                 self.d.getVar("ROOTFS_BOOTSTRAP_INSTALL", True)
Signature · 1 line
2.5.1
author
#11 lore raw

PATCH 3/4 systemd: Upgrade 219 -> 225

In reply to Khem Raj

Drop patches that were straight backports from upstream

MIT licence was unused and dropped from systemd sources for more details see
https://github.com/systemd/systemd/commit/8f1e0c5f38cdf7e401ab4d2bb93ad816d08e7715

Drop gtkdoc dependency since libudev API documentation has been converted from gtkdoc into man pages Remove packaging gudev as it has moved to separate repository outside
systemd

For more details see
https://github.com/systemd/systemd/commit/2375607039517c88df51ef16ddbb624ec1c10654

package newly added script for xorg to be usable with systemd --user intance
For more details see
https://github.com/systemd/systemd/commit/1401ec2d34bcde406ced531a72dc46ebaf332594

machinectl now has shell support

private-zone DHCP options are supported by systemd-networkd

For complete differences between two releases run
git log --oneline v219..v225
in systemd git clone

Change-Id I998e652382950a3c74c4839f3767ef8bef23d88f
---
 ...-mount-add-dependencies-to-dynamically-mo.patch |  36 ----
 ...ed-missing.h-fall-back-to-insecure-getenv.patch |  25 ---
 ...t-install-dependency-links-at-install-tim.patch |  35 ++--
 ...heck-for-additional-features-that-uclibc-.patch |  23 ++-
 ...wn-Use-execvpe-only-when-libc-supports-it.patch |  37 ++---
 ...til-Use-mkostemp-only-if-libc-supports-it.patch |   8 +-
 ...-unimplemented-_SC_PHYS_PAGES-system-conf.patch |   8 +-
 ...0-Make-root-s-home-directory-configurable.patch | 136 ++++++++-------
 ...files.c-Honor-ordering-within-files-as-th.patch | 185 ---------------------
 ...-remove-userspace-firmware-loading-suppor.patch |  99 ++++++-----
 .../systemd/systemd/tmpfiles-pam.patch             |  30 ----
 .../systemd/{systemd_219.bb => systemd_225.bb}     |  25 +--
 12 files changed, 174 insertions(+), 473 deletions(-)
 delete mode 100644 meta/recipes-core/systemd/systemd/0001-Revert-core-mount-add-dependencies-to-dynamically-mo.patch
 delete mode 100644 meta/recipes-core/systemd/systemd/0002-shared-missing.h-fall-back-to-insecure-getenv.patch
 delete mode 100644 meta/recipes-core/systemd/systemd/0012-systemd-tmpfiles.c-Honor-ordering-within-files-as-th.patch
 delete mode 100644 meta/recipes-core/systemd/systemd/tmpfiles-pam.patch
 rename meta/recipes-core/systemd/{systemd_219.bb => systemd_225.bb} (95%)
meta/recipes-core/systemd/systemd/0001-Revert-core-mount-add-dependencies-to-dynamically-mo.patch deleted +0 -36
deleted file mode 100644
index d69cfe0..0000000
@@ -1,36 +0,0 @@
-From eb7a760052667c5a35637901de9359c377263804 Mon Sep 17 00:00:00 2001
-From: Colin Walters <[email protected]>
-Date: Tue, 24 Feb 2015 10:12:28 -0500
-Subject: [PATCH] Revert "core/mount: add dependencies to dynamically mounted
- mounts too"
-
-This reverts commit 06e97888883e2cc12eb6514e80c7f0014295f59b.
-
-This patch was imported from Fedora (http://pkgs.fedoraproject.org/cgit/systemd.git/commit/?h=f22&id=9bbe0e92dc59d5a42258c729b105a7d9901eb35e).
-
-It fixes an issue where mount points not listed in fstab are immediately
-unmounted after a call to mount.
-
-Upstream-Status: Inappropriate [fixed upstream]
-
-Signed-off-by: Joshua Lock <[email protected]>
----
- src/core/mount.c | 2 +-
- 1 file changed, 1 insertion(+), 1 deletion(-)
-
-diff --git a/src/core/mount.c b/src/core/mount.c
-index 40037e7..cddb5e2 100644
---- a/src/core/mount.c
-+++ b/src/core/mount.c
-@@ -295,7 +295,7 @@ static int mount_add_device_links(Mount *m) {
- 
-         assert(m);
- 
--        p = get_mount_parameters(m);
-+        p = get_mount_parameters_fragment(m);
-         if (!p)
-                 return 0;
- 
--- 
-1.8.3.1
-
meta/recipes-core/systemd/systemd/0002-shared-missing.h-fall-back-to-insecure-getenv.patch deleted +0 -25
deleted file mode 100644
index 1cf7840..0000000
@@ -1,25 +0,0 @@
-From 03baade6db2cf83b44a157818ba827d725449bb0 Mon Sep 17 00:00:00 2001
-From: Emil Renner Berthing <[email protected]>
-Date: Thu, 18 Sep 2014 15:24:47 +0200
-Subject: [PATCH 02/11] shared/missing.h: fall back to insecure getenv
-
----
- src/shared/missing.h | 2 +-
- 1 file changed, 1 insertion(+), 1 deletion(-)
-
-diff --git a/src/shared/missing.h b/src/shared/missing.h
-index b33a70c..11cca04 100644
---- a/src/shared/missing.h
-+++ b/src/shared/missing.h
-@@ -366,7 +366,7 @@ static inline int name_to_handle_at(int fd, const char *name, struct file_handle
- #  ifdef HAVE___SECURE_GETENV
- #    define secure_getenv __secure_getenv
- #  else
--#    error "neither secure_getenv nor __secure_getenv are available"
-+#    define secure_getenv getenv
- #  endif
- #endif
- 
--- 
-2.1.4
-
meta/recipes-core/systemd/systemd/0003-binfmt-Don-t-install-dependency-links-at-install-tim.patch +16 -19
index c14cc6f..0e20fd0 100644
@@ -23,11 +23,11 @@ Signed-off-by: Khem Raj <[email protected]>
  units/systemd-binfmt.service.in         | 5 +++++
  3 files changed, 8 insertions(+), 4 deletions(-)
 
-diff --git a/Makefile.am b/Makefile.am
-index ba63f68..0fb3f9f 100644
---- a/Makefile.am
-+++ b/Makefile.am
-@@ -4838,10 +4838,6 @@ INSTALL_DIRS += \
+Index: git/Makefile.am
+===================================================================
+--- git.orig/Makefile.am
++++ git/Makefile.am
+@@ -4271,10 +4271,6 @@ INSTALL_DIRS += \
  	$(prefix)/lib/binfmt.d \
  	$(sysconfdir)/binfmt.d
  
@@ -38,10 +38,10 @@ index ba63f68..0fb3f9f 100644
  endif
  
  EXTRA_DIST += \
-diff --git a/units/proc-sys-fs-binfmt_misc.automount b/units/proc-sys-fs-binfmt_misc.automount
-index 6be3893..709adef 100644
---- a/units/proc-sys-fs-binfmt_misc.automount
-+++ b/units/proc-sys-fs-binfmt_misc.automount
+Index: git/units/proc-sys-fs-binfmt_misc.automount
+===================================================================
+--- git.orig/units/proc-sys-fs-binfmt_misc.automount
++++ git/units/proc-sys-fs-binfmt_misc.automount
 @@ -16,3 +16,6 @@ ConditionPathIsReadWrite=/proc/sys/
  
  [Automount]
@@ -49,11 +49,11 @@ index 6be3893..709adef 100644
 +
 +[Install]
 +WantedBy=sysinit.target
-diff --git a/units/systemd-binfmt.service.in b/units/systemd-binfmt.service.in
-index 34a5d52..617462b 100644
---- a/units/systemd-binfmt.service.in
-+++ b/units/systemd-binfmt.service.in
-@@ -11,6 +11,8 @@ Documentation=man:systemd-binfmt.service(8) man:binfmt.d(5)
+Index: git/units/systemd-binfmt.service.in
+===================================================================
+--- git.orig/units/systemd-binfmt.service.in
++++ git/units/systemd-binfmt.service.in
+@@ -11,6 +11,8 @@ Documentation=man:systemd-binfmt.service
  Documentation=https://www.kernel.org/doc/Documentation/binfmt_misc.txt
  DefaultDependencies=no
  Conflicts=shutdown.target
@@ -62,13 +62,10 @@ index 34a5d52..617462b 100644
  After=proc-sys-fs-binfmt_misc.automount
  Before=sysinit.target shutdown.target
  ConditionPathIsReadWrite=/proc/sys/
-@@ -24,3 +26,6 @@ ConditionDirectoryNotEmpty=|/run/binfmt.d
- Type=oneshot
+@@ -25,3 +27,6 @@ Type=oneshot
  RemainAfterExit=yes
  ExecStart=@rootlibexecdir@/systemd-binfmt
+ TimeoutSec=90s
 +
 +[Install]
 +WantedBy=sysinit.target
--- 
-2.1.4
-
meta/recipes-core/systemd/systemd/0004-configure-Check-for-additional-features-that-uclibc-.patch +10 -13
index dddb7da..8078da9 100644
@@ -14,13 +14,13 @@ Signed-off-by: Khem Raj <[email protected]>
  configure.ac | 18 ++++++++++++++++++
  1 file changed, 18 insertions(+)
 
-diff --git a/configure.ac b/configure.ac
-index 9a2235b..a5b2e6e 100644
---- a/configure.ac
-+++ b/configure.ac
-@@ -103,6 +103,24 @@ AS_IF([! ln --relative --help > /dev/null 2>&1], [AC_MSG_ERROR([*** ln doesn't s
+Index: git/configure.ac
+===================================================================
+--- git.orig/configure.ac
++++ git/configure.ac
+@@ -110,6 +110,24 @@ AC_PATH_PROG([UMOUNT_PATH], [umount], [/
  
- M4_DEFINES=
+ AS_IF([! ln --relative --help > /dev/null 2>&1], [AC_MSG_ERROR([*** ln doesn't support --relative ***])])
  
 +# check for few functions not implemented in uClibc
 +
@@ -31,7 +31,7 @@ index 9a2235b..a5b2e6e 100644
 +AC_MSG_CHECKING([whether %ms format is supported by *scanf])
 +
 +AC_LINK_IFELSE(
-+	[AC_LANG_PROGRAM([ 
++	[AC_LANG_PROGRAM([
 +	 #include <stdio.h>
 +	 ],[
 +	    char *buf1, *buf2, *buf3, str="1 2.3 abcde" ;
@@ -40,9 +40,6 @@ index 9a2235b..a5b2e6e 100644
 +	[AC_DEFINE([HAVE_MSFORMAT], [1], [Define if %ms format is supported by *scanf.])],
 +	[AC_MSG_RESULT([no])])
 +
- # gtkdocize greps for '^GTK_DOC_CHECK', so it needs to be on its own line
- m4_ifdef([GTK_DOC_CHECK], [
- GTK_DOC_CHECK([1.18],[--flavour no-tmpl])],
--- 
-2.1.4
-
+ M4_DEFINES=
+ 
+ AC_CHECK_TOOL(OBJCOPY, objcopy)
meta/recipes-core/systemd/systemd/0005-nspawn-Use-execvpe-only-when-libc-supports-it.patch +17 -20
index 85279d6..55a0088 100644
@@ -10,32 +10,29 @@ Signed-off-by: Khem Raj <[email protected]>
  src/nspawn/nspawn.c | 7 +++++++
  1 file changed, 7 insertions(+)
 
-diff --git a/src/nspawn/nspawn.c b/src/nspawn/nspawn.c
-index 0d8d199..b597edb 100644
---- a/src/nspawn/nspawn.c
-+++ b/src/nspawn/nspawn.c
-@@ -129,6 +129,8 @@ typedef enum Volatile {
-         VOLATILE_STATE,
- } Volatile;
+Index: git/src/nspawn/nspawn.c
+===================================================================
+--- git.orig/src/nspawn/nspawn.c
++++ git/src/nspawn/nspawn.c
+@@ -143,6 +143,8 @@ typedef struct CustomMount {
+         char **lower;
+ } CustomMount;
  
 +#include "config.h"
 +
  static char *arg_directory = NULL;
  static char *arg_template = NULL;
  static char *arg_user = NULL;
-@@ -4257,7 +4259,12 @@ int main(int argc, char *argv[]) {
-                                 a[0] = (char*) "/sbin/init";
-                                 execve(a[0], a, env_use);
-                         } else if (argc > optind)
+@@ -4238,7 +4240,12 @@ static int inner_child(
+                 a[0] = (char*) "/sbin/init";
+                 execve(a[0], a, env_use);
+         } else if (argc > optind)
 +#ifdef HAVE_EXECVPE
-                                 execvpe(argv[optind], argv + optind, env_use);
+                 execvpe(argv[optind], argv + optind, env_use);
 +#else
-+                                environ = env_use;
-+                                execvp(argv[optind], argv + optind);
++                environ = env_use;
++                execvp(argv[optind], argv + optind);
 +#endif /* HAVE_EXECVPE */
-                         else {
-                                 chdir(home ? home : "/root");
-                                 execle("/bin/bash", "-bash", NULL, env_use);
--- 
-2.1.4
-
+         else {
+                 chdir(home ? home : "/root");
+                 execle("/bin/bash", "-bash", NULL, env_use);
meta/recipes-core/systemd/systemd/0007-util-Use-mkostemp-only-if-libc-supports-it.patch +4 -4
index 29c20c0..b8ab7c4 100644
@@ -7,13 +7,13 @@ Upstream-Status: Denied [no desire for uclibc support]
 
 Signed-off-by: Khem Raj <[email protected]>
 ---
- src/shared/util.c | 8 ++++++++
+ src/basic/util.c | 8 ++++++++
  1 file changed, 8 insertions(+)
 
-diff --git a/src/shared/util.c b/src/shared/util.c
+diff --git a/src/basic/util.c b/src/basic/util.c
 index dc65280..72f4665 100644
---- a/src/shared/util.c
-+++ b/src/shared/util.c
+--- a/src/basic/util.c
++++ b/src/basic/util.c
 @@ -97,6 +97,8 @@
  #include "def.h"
  #include "sparse-endian.h"
meta/recipes-core/systemd/systemd/0008-util-bypass-unimplemented-_SC_PHYS_PAGES-system-conf.patch +4 -4
index 5297625..3f4d4de 100644
@@ -8,13 +8,13 @@ Upstream-Status: Inappropriate [uclibc-specific]
 
 Signed-off-by: Khem Raj <[email protected]>
 ---
- src/shared/util.c | 15 +++++++++++++++
+ src/basic/util.c | 15 +++++++++++++++
  1 file changed, 15 insertions(+)
 
-diff --git a/src/shared/util.c b/src/shared/util.c
+diff --git a/src/basic/util.c b/src/basic/util.c
 index 72f4665..cbbe3b1 100644
---- a/src/shared/util.c
-+++ b/src/shared/util.c
+--- a/src/basic/util.c
++++ b/src/basic/util.c
 @@ -6793,10 +6793,25 @@ uint64_t physical_memory(void) {
          /* We return this as uint64_t in case we are running as 32bit
           * process on a 64bit kernel with huge amounts of memory */
meta/recipes-core/systemd/systemd/0010-Make-root-s-home-directory-configurable.patch +66 -70

215 lines not on this page · raw message

meta/recipes-core/systemd/systemd/0012-systemd-tmpfiles.c-Honor-ordering-within-files-as-th.patch deleted +0 -185

185 lines not on this page · raw message

meta/recipes-core/systemd/systemd/0015-Revert-udev-remove-userspace-firmware-loading-suppor.patch +48 -51

177 lines not on this page · raw message

meta/recipes-core/systemd/systemd/tmpfiles-pam.patch deleted +0 -30
deleted file mode 100644
index a40b1b9..0000000
@@ -1,30 +0,0 @@
-Upstream-Status: Submitted
-Signed-off-by: Ross Burton <[email protected]>
-
-From 0802aaaa9784813d318d045c79533a044eedf542 Mon Sep 17 00:00:00 2001
-From: Ross Burton <[email protected]>
-Date: Tue, 3 Mar 2015 11:32:29 +0000
-Subject: [PATCH] tmpfiles.d: only copy /etc/pam.d if PAM is present
-
-If HAVE_PAM isn't set then don't attempt to copy /etc/pam.d from the factory, as
-it doesn't get installed.
-
-Signed-off-by: Ross Burton <[email protected]>
----
- tmpfiles.d/etc.conf.m4 |    2 ++
- 1 file changed, 2 insertions(+)
-
-diff --git a/tmpfiles.d/etc.conf.m4 b/tmpfiles.d/etc.conf.m4
-index 9b0e080..ab5cd16 100644
---- a/tmpfiles.d/etc.conf.m4
-+++ b/tmpfiles.d/etc.conf.m4
-@@ -14,4 +14,6 @@ m4_ifdef(`ENABLE_RESOLVED',
- L /etc/resolv.conf - - - - ../run/systemd/resolve/resolv.conf
- )m4_dnl
- C /etc/nsswitch.conf - - - -
-+m4_ifdef(`HAVE_PAM',
- C /etc/pam.d - - - -
-+)m4_dnl
--- 
-1.7.10.4
-
meta/recipes-core/systemd/systemd_225.bb renamed from meta/recipes-core/systemd/systemd_219.bb +9 -16
similarity index 95%
rename from meta/recipes-core/systemd/systemd_219.bb
rename to meta/recipes-core/systemd/systemd_225.bb
index 4d32fd5..a1d371b 100644
@@ -10,10 +10,9 @@ state, maintains mount and automount points and implements an \
 elaborate transactional dependency-based service control logic. It can \
 work as a drop-in replacement for sysvinit."
 
-LICENSE = "GPLv2 & LGPLv2.1 & MIT"
+LICENSE = "GPLv2 & LGPLv2.1"
 LIC_FILES_CHKSUM = "file://LICENSE.GPL2;md5=751419260aa954499f7abaabaa882bbe \
-                    file://LICENSE.LGPL2.1;md5=4fbd65380cdd255951079008b364516c \
-                    file://LICENSE.MIT;md5=544799d0b492f119fa04641d1b8868ed"
+                    file://LICENSE.LGPL2.1;md5=4fbd65380cdd255951079008b364516c"
 
 PROVIDES = "udev"
 
@@ -23,14 +22,13 @@ DEPENDS = "kmod docbook-sgml-dtd-4.1-native intltool-native gperf-native acl rea
 
 SECTION = "base/shell"
 
-inherit gtk-doc useradd pkgconfig autotools perlnative update-rc.d update-alternatives qemu systemd ptest gettext
+inherit useradd pkgconfig autotools perlnative update-rc.d update-alternatives qemu systemd ptest gettext
 
-SRCREV = "85a6fabdd3e43cfab0fc6359e9f2a9e368d4a3ed"
+SRCREV = "e1439a1472c5f691733b8ef10e702beac2496a63"
 
-PV = "219-stable+git${SRCPV}"
+PV = "225+git${SRCPV}"
 
-SRC_URI = "git://github.com/systemd/systemd-stable;branch=v219-stable;protocol=git \
-           file://0002-shared-missing.h-fall-back-to-insecure-getenv.patch \
+SRC_URI = "git://github.com/systemd/systemd.git;protocol=git \
            file://0003-binfmt-Don-t-install-dependency-links-at-install-tim.patch \
            file://0004-configure-Check-for-additional-features-that-uclibc-.patch \
            file://0005-nspawn-Use-execvpe-only-when-libc-supports-it.patch \
@@ -40,11 +38,8 @@ SRC_URI = "git://github.com/systemd/systemd-stable;branch=v219-stable;protocol=g
            file://0009-sysv-generator-add-support-for-executing-scripts-und.patch \
            file://0010-Make-root-s-home-directory-configurable.patch \
            file://0011-systemd-user-avoid-using-system-auth.patch \
-           file://0012-systemd-tmpfiles.c-Honor-ordering-within-files-as-th.patch \
            file://0014-Revert-rules-remove-firmware-loading-rules.patch \
            file://0015-Revert-udev-remove-userspace-firmware-loading-suppor.patch \
-           file://tmpfiles-pam.patch \
-           file://0001-Revert-core-mount-add-dependencies-to-dynamically-mo.patch \
            file://touchscreen.rules \
            file://00-create-volatile.conf \
            file://init \
@@ -119,8 +114,6 @@ EXTRA_OECONF = " --with-rootprefix=${rootprefix} \
                  --with-rootlibdir=${rootlibdir} \
                  --with-roothomedir=${ROOT_HOME} \
                  --disable-coredump \
-                 --disable-introspection \
-                 --disable-kdbus \
                  --enable-split-usr \
                  --without-python \
                  --with-sysvrcnd-path=${sysconfdir} \
@@ -220,7 +213,7 @@ python populate_packages_prepend (){
 PACKAGES_DYNAMIC += "^lib(udev|systemd).*"
 
 PACKAGES =+ "${PN}-gui ${PN}-vconsole-setup ${PN}-initramfs ${PN}-analyze ${PN}-kernel-install \
-             ${PN}-rpm-macros ${PN}-binfmt ${PN}-pam ${PN}-zsh libgudev"
+             ${PN}-rpm-macros ${PN}-binfmt ${PN}-pam ${PN}-zsh ${PN}-xorg-xinitrc"
 
 SYSTEMD_PACKAGES = "${PN}-binfmt"
 SYSTEMD_SERVICE_${PN}-binfmt = "systemd-binfmt.service"
@@ -234,8 +227,6 @@ FILES_${PN}-analyze = "${bindir}/systemd-analyze"
 FILES_${PN}-initramfs = "/init"
 RDEPENDS_${PN}-initramfs = "${PN}"
 
-FILES_libgudev = "${libdir}/libgudev*${SOLIBS}"
-
 RDEPENDS_${PN}-ptest += "perl python bash"
 FILES_${PN}-ptest += "${libdir}/udev/rules.d"
 
@@ -255,6 +246,8 @@ FILES_${PN}-kernel-install = "${bindir}/kernel-install \
 FILES_${PN}-rpm-macros = "${exec_prefix}/lib/rpm \
                          "
 
+FILES_${PN}-xorg-xinitrc = "${sysconfdir}/X11/xinit/xinitrc.d/*"
+
 FILES_${PN}-zsh = "${datadir}/zsh/site-functions"
 
 FILES_${PN}-binfmt = "${sysconfdir}/binfmt.d/ \
Signature · 1 line
2.5.1
author
#12 lore raw

PATCH 4/4 systemd: Implement OE-Specific systemd-sysv-install

In reply to Khem Raj

Support for chkconfig (--enable-chkconfig) was removed in favour of calling an abstraction /lib/systemd/systemd-sysv-install. This needs to be implemented for OE.

---
 ...0012-implment-systemd-sysv-install-for-OE.patch | 40 ++++++++++++++++++++++
 meta/recipes-core/systemd/systemd_225.bb           |  4 ++-
 2 files changed, 43 insertions(+), 1 deletion(-)
 create mode 100644 meta/recipes-core/systemd/systemd/0012-implment-systemd-sysv-install-for-OE.patch
meta/recipes-core/systemd/systemd/0012-implment-systemd-sysv-install-for-OE.patch new file +40 -0
new file mode 100644
index 0000000..68b45a9
@@ -0,0 +1,40 @@
+From 5f8b9f2d276c0ddbcbf5423733a23f043d688009 Mon Sep 17 00:00:00 2001
+From: Khem Raj <[email protected]>
+Date: Sat, 5 Sep 2015 06:31:47 +0000
+Subject: [PATCH] implment systemd-sysv-install for OE
+
+Use update-rc.d for enabling/disabling and status command
+to check the status of the sysv service
+
+Upstream-Status: Inappropriate [OE-Specific]
+
+Signed-off-by: Khem Raj <[email protected]>
+---
+ src/systemctl/systemd-sysv-install.SKELETON | 6 +++---
+ 1 file changed, 3 insertions(+), 3 deletions(-)
+
+Index: git/src/systemctl/systemd-sysv-install.SKELETON
+===================================================================
+--- git.orig/src/systemctl/systemd-sysv-install.SKELETON
++++ git/src/systemctl/systemd-sysv-install.SKELETON
+@@ -30,17 +30,17 @@ case "$1" in
+     enable)
+         # call the command to enable SysV init script $NAME here
+         # (consider optional $ROOT)
+-        echo "IMPLEMENT ME: enabling SysV init.d script $NAME"
++        update-rc.d -f $NAME defaults
+         ;;
+     disable)
+         # call the command to disable SysV init script $NAME here
+         # (consider optional $ROOT)
+-        echo "IMPLEMENT ME: disabling SysV init.d script $NAME"
++        update-rc.d -f $NAME remove
+         ;;
+     is-enabled)
+         # exit with 0 if $NAME is enabled, non-zero if it is disabled
+         # (consider optional $ROOT)
+-        echo "IMPLEMENT ME: checking SysV init.d script $NAME"
++        /etc/init.d/$NAME status
+         ;;
+     *)
+         usage ;;
meta/recipes-core/systemd/systemd_225.bb +3 -1
index a1d371b..2299081 100644
@@ -38,6 +38,7 @@ SRC_URI = "git://github.com/systemd/systemd.git;protocol=git \
            file://0009-sysv-generator-add-support-for-executing-scripts-und.patch \
            file://0010-Make-root-s-home-directory-configurable.patch \
            file://0011-systemd-user-avoid-using-system-auth.patch \
+           file://0012-implment-systemd-sysv-install-for-OE.patch \
            file://0014-Revert-rules-remove-firmware-loading-rules.patch \
            file://0015-Revert-udev-remove-userspace-firmware-loading-suppor.patch \
            file://touchscreen.rules \
@@ -188,6 +189,7 @@ do_install() {
 	if [ -s ${D}${libdir}/tmpfiles.d/systemd.conf ]; then
 		${@bb.utils.contains('PACKAGECONFIG', 'networkd', ':', 'sed -i -e "\$ad /run/systemd/netif/links 0755 root root -" ${D}${libdir}/tmpfiles.d/systemd.conf', d)}
 	fi
+	install -Dm 0755 ${S}/src/systemctl/systemd-sysv-install.SKELETON ${D}${systemd_unitdir}/systemd-sysv-install
 }
 
 do_install_ptest () {
@@ -309,7 +311,7 @@ FILES_${PN}-dbg += "${rootlibdir}/.debug ${systemd_unitdir}/.debug ${systemd_uni
 FILES_${PN}-dev += "${base_libdir}/security/*.la ${datadir}/dbus-1/interfaces/ ${sysconfdir}/rpm/macros.systemd"
 
 RDEPENDS_${PN} += "kmod dbus util-linux-mount udev (= ${EXTENDPKGV})"
-RDEPENDS_${PN} += "volatile-binds"
+RDEPENDS_${PN} += "volatile-binds update-rc.d"
 
 RRECOMMENDS_${PN} += "systemd-serialgetty systemd-vconsole-setup \
                       systemd-compat-units udev-hwdb \
Signature · 1 line
2.5.1
#13 lore raw
In reply to Khem Raj
Quoted 8 lines · Khem Raj

Changes Since last time
- Now that 225 is out upgrade recipe to 225, changes between 224 and 225 are minor but fixes in networkd are important
- Implement systemd-sysv-install for OE
- Enable getopt in busybox needed for systemd-sysv-install
- Remove update-rc.d from default uninstall list in rootfs creation
- Tested date and systemd unit tests manually to pass
- Booted core-image-cmdline-full on all qemu targets

https://autobuilder.yoctoproject.org/main/builders/nightly-qa-systemd/builds/480

The date test failed in two of the sanity runs with this series applied :(

Cheers,

Richard

#3 lore raw

Re: [PATCH 2/4] rootfs: Do not uninstall update-rc.d

In reply to Khem Raj
Quoted 3 lines · Khem Raj

update-rc.d is now needed by systemd to interact with sysv scripts, so if we have a mixed system, then we can not uninstall update-rc.d as it is going to break systemd functionality

The description above implies that you will only uninstall update-rc.d if systemd is present, but that doesn't seem to be what the patch does.

Also, if the rootfs is read-only then you might as well go ahead and uninstall it irrespective of systemd because it isn't going to work. If systemd will fail catastrophically in this situation then I guess it should be an error to have both systemd and sysv scripts in a read-only filesystem.

p.

author
#14 lore raw
In reply to Richard Purdie
Quoted 13 lines · Richard Purdie
Quoted 8 lines · Khem Raj

Changes Since last time
- Now that 225 is out upgrade recipe to 225, changes between 224 and 225 are minor but fixes in networkd are important
- Implement systemd-sysv-install for OE
- Enable getopt in busybox needed for systemd-sysv-install
- Remove update-rc.d from default uninstall list in rootfs creation
- Tested date and systemd unit tests manually to pass
- Booted core-image-cmdline-full on all qemu targets

Quoted 4 lines

https://autobuilder.yoctoproject.org/main/builders/nightly-qa-systemd/builds/480

The date test failed in two of the sanity runs with this series applied :(

I think this testcase needs revisiting. My manual runs of sequence of exact same commands on qemux86-64 and qemuarm64 succeeds all the time. I have built core-image-full-cmdline One thing I wonder is if in AB case there might network time is configured and systemd-timesyncd/systemd-timedated syncs it back
immediately when its tempered with. I see this happening with systemd-219 as well when I have networkd and timesyncd started on another instance running on minnow ( without the proposed patchset)

and for systemd we should try to test timedatectl to manipulate date instead of date cmd e.g. timedatectl set-time “YYYY-MM-DD HH:MM:SS”

I will try to reproduce it on my end and see if I can get something to whats going on.

#15 lore raw
Quoted 26 lines · Khem Raj
Quoted 12 lines · Richard Purdie
Quoted 8 lines · Khem Raj

Changes Since last time
- Now that 225 is out upgrade recipe to 225, changes between 224 and 225 are minor but fixes in networkd are important
- Implement systemd-sysv-install for OE
- Enable getopt in busybox needed for systemd-sysv-install
- Remove update-rc.d from default uninstall list in rootfs creation
- Tested date and systemd unit tests manually to pass
- Booted core-image-cmdline-full on all qemu targets

Quoted 4 lines

https://autobuilder.yoctoproject.org/main/builders/nightly-qa-systemd/builds/480

The date test failed in two of the sanity runs with this series applied :(

I think this testcase needs revisiting. My manual runs of sequence of exact same commands on qemux86-64 and qemuarm64 succeeds all the time. I have built core-image-full-cmdline
One thing I wonder is if in AB case there might network time is configured and systemd-timesyncd/systemd-timedated syncs it back immediately when its tempered with. I see this happening with systemd-219 as well when I have networkd and timesyncd started on another instance running on minnow ( without the proposed patchset)

Quoted 5 lines

and for systemd we should try to test timedatectl to manipulate date instead of date cmd e.g. timedatectl set-time “YYYY-MM-DD HH:MM:SS”

I will try to reproduce it on my end and see if I can get something to whats going on.

A "bitbake core-image-sato -c testimage" replicated it here FWIW, have you tried testimage to see if that somehow makes a difference? Shoudn't be hard to test with that and for sato, it should run pretty quickly.

It all works with 219, its only the upgrade patches that seem to break things :(

Cheers,

Richard

author
#16 lore raw
Quoted 9 lines · Richard Purdie

A "bitbake core-image-sato -c testimage" replicated it here FWIW, have you tried testimage to see if that somehow makes a difference? Shoudn't be hard to test with that and for sato, it should run pretty quickly.

It all works with 219, its only the upgrade patches that seem to break things :(

yeah my machine is w/o X and moreover I don’t have sudo on it test image seems to need X I will try to build core-image-sato and try to reproduce it locally. in order to analyze systemd defect it needs to be reproducible outside testimage infra anyway.

#17 lore raw
Quoted 6 lines · Khem Raj
Quoted 6 lines · Richard Purdie

A "bitbake core-image-sato -c testimage" replicated it here FWIW, have you tried testimage to see if that somehow makes a difference? Shoudn't be hard to test with that and for sato, it should run pretty quickly.

Quoted 6 lines
Quoted 2 lines

It all works with 219, its only the upgrade patches that seem to break things :(

yeah my machine is w/o X and moreover I don’t have sudo on it test image seems to need X I will try to build core-image-sato and try to reproduce it locally. in order to analyze systemd defect it needs to be reproducible outside testimage infra anyway.

I believe the tests will run with the nographic options to qemu which then won't need X, Ross had a patch around for that recently. To be honest when looking at this kind of issue, I usually debug it using testimage now, its actually easier in many cases...

Cheers,

Richard

author
#18 lore raw
Quoted 13 lines · Richard Purdie
Quoted 9 lines · Khem Raj
Quoted 8 lines · Richard Purdie

A "bitbake core-image-sato -c testimage" replicated it here FWIW, have you tried testimage to see if that somehow makes a difference? Shoudn't be hard to test with that and for sato, it should run pretty quickly.

It all works with 219, its only the upgrade patches that seem to break things :(

Quoted 8 lines
Quoted 3 lines

yeah my machine is w/o X and moreover I don’t have sudo on it test image seems to need X I will try to build core-image-sato and try to reproduce it locally. in order to analyze systemd defect it needs to be reproducible outside testimage infra anyway.

I believe the tests will run with the nographic options to qemu which then won't need X, Ross had a patch around for that recently. To be honest when looking at this kind of issue, I usually debug it using testimage now, its actually easier in many cases…

in this case it will be faster. Care to point to Ross’s patch ?

Quoted 8 lines

Cheers,

Richard

#19 lore raw
Quoted 16 lines · Khem Raj
Quoted 16 lines · Richard Purdie
Quoted 12 lines · Khem Raj
Quoted 8 lines · Richard Purdie

A "bitbake core-image-sato -c testimage" replicated it here FWIW, have you tried testimage to see if that somehow makes a difference? Shoudn't be hard to test with that and for sato, it should run pretty quickly.

It all works with 219, its only the upgrade patches that seem to break things :(

yeah my machine is w/o X and moreover I don’t have sudo on it test image seems to need X I will try to build core-image-sato and try to reproduce it locally. in order to analyze systemd defect it needs to be reproducible outside testimage infra anyway.

Quoted 6 lines
Quoted 4 lines

I believe the tests will run with the nographic options to qemu which then won't need X, Ross had a patch around for that recently. To be honest when looking at this kind of issue, I usually debug it using testimage now, its actually easier in many cases…

in this case it will be faster. Care to point to Ross’s patch ?

http://git.yoctoproject.org/cgit.cgi/poky-contrib/commit/?h=ross/mut&id=e67ab458ab2d67908d80f9aa1954dee05cb21fc1

Cheers,

Richard

author
#4 lore raw

Re: [PATCH 2/4] rootfs: Do not uninstall update-rc.d

In reply to Phil Blundell
Quoted 10 lines · Phil Blundell
Quoted 3 lines · Khem Raj

update-rc.d is now needed by systemd to interact with sysv scripts, so if we have a mixed system, then we can not uninstall update-rc.d as it is going to break systemd functionality

The description above implies that you will only uninstall update-rc.d if systemd is present, but that doesn't seem to be what the patch does.

I meant to not say its only for systemd. but that systemd now needs it. so update-rc.d is no longer forcefully removed from image depchain.

Quoted 6 lines

Also, if the rootfs is read-only then you might as well go ahead and uninstall it irrespective of systemd because it isn't going to work. If systemd will fail catastrophically in this situation then I guess it should be an error to have both systemd and sysv scripts in a read-only filesystem.

That is valid problem, but its not changing that behavior. Code needs to be added to check for ro-rootfs check and then uninstall it. In ro-rootfs case this disable/enable of script might still be needed and useful since there might overlayfs in use

Quoted 4 lines

p.

#5 lore raw

Re: [PATCH 2/4] rootfs: Do not uninstall update-rc.d

Quoted 7 lines · Khem Raj
Quoted 7 lines · Phil Blundell
Quoted 3 lines · Khem Raj

update-rc.d is now needed by systemd to interact with sysv scripts, so if we have a mixed system, then we can not uninstall update-rc.d as it is going to break systemd functionality

Quoted 5 lines
Quoted 2 lines

The description above implies that you will only uninstall update-rc.d if systemd is present, but that doesn't seem to be what the patch does.

I meant to not say its only for systemd. but that systemd now needs it. so update-rc.d is no longer forcefully removed from image depchain.

Well, right. But wouldn't it be better only to leave it in there if it's actually necessary, i.e. if you have systemd and you also have sysv scripts and the rootfs isn't read-only? As it stands, your patch will cause update-rc.d to be left around for people who aren't using systemd, which seems like a retrograde step.

p.

author
#6 lore raw

Re: [PATCH 2/4] rootfs: Do not uninstall update-rc.d

Quoted 14 lines · Phil Blundell
Quoted 10 lines · Khem Raj
Quoted 9 lines · Phil Blundell
Quoted 3 lines · Khem Raj

update-rc.d is now needed by systemd to interact with sysv scripts, so if we have a mixed system, then we can not uninstall update-rc.d as it is going to break systemd functionality

The description above implies that you will only uninstall update-rc.d if systemd is present, but that doesn't seem to be what the patch does.

Quoted 9 lines
Quoted 2 lines

I meant to not say its only for systemd. but that systemd now needs it. so update-rc.d is no longer forcefully removed from image depchain.

Well, right. But wouldn't it be better only to leave it in there if it's actually necessary, i.e. if you have systemd and you also have sysv scripts and the rootfs isn't read-only? As it stands, your patch will cause update-rc.d to be left around for people who aren't using systemd, which seems like a retrograde step.

makes sense. I can improvise on it and check for all constraints in a follow up.

Quoted 3 lines

p.

#20 lore raw
In reply to Khem Raj
Quoted 7 lines · Khem Raj
Quoted 4 lines

I believe the tests will run with the nographic options to qemu which then won't need X, Ross had a patch around for that recently. To be honest when looking at this kind of issue, I usually debug it using testimage now, its actually easier in many cases…

in this case it will be faster. Care to point to Ross’s patch ?

http://git.yoctoproject.org/cgit/cgit.cgi/poky-contrib/commit/?h=ross/mut&id=e67ab458ab2d67908d80f9aa1954dee05cb21fc1
is the patch to run the tests in a nographics qemu. The framework still checks that DISPLAY is set, but you can just set that to anything you want as it won't be used.

Ross

author
#22 lore raw
In reply to Richard Purdie
Quoted 10 lines · Richard Purdie
Quoted 8 lines · Khem Raj

Changes Since last time
- Now that 225 is out upgrade recipe to 225, changes between 224 and 225 are minor but fixes in networkd are important
- Implement systemd-sysv-install for OE
- Enable getopt in busybox needed for systemd-sysv-install
- Remove update-rc.d from default uninstall list in rootfs creation
- Tested date and systemd unit tests manually to pass
- Booted core-image-cmdline-full on all qemu targets

Quoted 4 lines

https://autobuilder.yoctoproject.org/main/builders/nightly-qa-systemd/builds/480

The date test failed in two of the sanity runs with this series applied :(

I am able to reproduce this now and also analyse it. Its a timing issue and original testcase needs to be changed.
The root cause is that we have systemd-timesyncd started automatically on boot ( we always had) so when we changed the date and asked for the date immediately we got it before timesyncd came around
and readjusted it to network time.

happens with

SSH="ssh [email protected]"
$SSH 'date +"%Y-%m-%d %T"'
$SSH 'date -s "2016-08-09 10:00:00"'
$SSH 'date -R'

Fixed with

SSH="ssh [email protected]"
$SSH 'systemctl stop systemd-timesyncd'
$SSH 'date +"%Y-%m-%d %T"'
$SSH 'date -s "2016-08-09 10:00:00"'
$SSH 'date -R'
$SSH 'systemctl start systemd-timesyncd'

this patch fixed it during autoruns, but date test is common for sysvint and systemd it ignores the outputs of systemctl cmds
so in theory it will work on both sysvinit and systemd. Although its not elegant

meta/lib/oeqa/runtime/date.py +1 -0
index 97e8ee4..f73ad1c 100644
@@ -7,6 +7,7 @@ class DateTest(oeRuntimeTest):
    @testcase(211)
    @skipUnlessPassed("test_ssh")
    def test_date(self):
+        (status, output) = self.target.run('systemctl stop systemd-timesyncd')
        (status, output) = self.target.run('date +"%Y-%m-%d %T"')
        self.assertEqual(status, 0, msg="Failed to get initial date,

output: %s" % output)
oldDate = output
@@ -21,3 +22,4 @@ class DateTest(oeRuntimeTest):

(status, output) = self.target.run('date -s "%s"' % oldDate) self.assertEqual(status, 0, msg="Failed to reset date, output:
%s" % output)
+ (status, output) = self.target.run('systemctl start systemd-timesyncd')

Quoted 5 lines

Cheers,

Richard

author
#21 lore raw
In reply to Burton, Ross
Quoted 4 lines · Ross

http://git.yoctoproject.org/cgit/cgit.cgi/poky-contrib/commit/?h=ross/mut&id=e67ab458ab2d67908d80f9aa1954dee05cb21fc1
is the patch to run the tests in a nographics qemu. The framework still checks that DISPLAY is set, but you can just set that to anything you want as it won't be used.

yeah, ideally the check should be removed now that it becomes useless with this patch. May be you want to improve above mentioned patch to account for that as well.

author
#7 lore raw

Re: [PATCH 2/4] rootfs: Do not uninstall update-rc.d

In reply to Khem Raj
Quoted 17 lines · Khem Raj
Quoted 17 lines · Phil Blundell
Quoted 12 lines · Khem Raj
Quoted 9 lines · Phil Blundell
Quoted 3 lines · Khem Raj

update-rc.d is now needed by systemd to interact with sysv scripts, so if we have a mixed system, then we can not uninstall update-rc.d as it is going to break systemd functionality

The description above implies that you will only uninstall update-rc.d if systemd is present, but that doesn't seem to be what the patch does.

I meant to not say its only for systemd. but that systemd now needs it. so update-rc.d is no longer forcefully removed from image depchain.

Quoted 8 lines
Quoted 6 lines

Well, right. But wouldn't it be better only to leave it in there if it's actually necessary, i.e. if you have systemd and you also have sysv scripts and the rootfs isn't read-only? As it stands, your patch will cause update-rc.d to be left around for people who aren't using systemd, which seems like a retrograde step.

makes sense. I can improvise on it and check for all constraints in a follow up.

I have updates this patch in pull branch
https://github.com/kraj/openembedded-core/commit/c4c574dff9ef689c7694a6055907083e7a018d48

Quoted 5 lines
Quoted 4 lines

p.

#23 lore raw
In reply to Khem Raj
Quoted 23 lines · Khem Raj

this patch fixed it during autoruns, but date test is common for sysvint and systemd it ignores the outputs of systemctl cmds so in theory it will work on both sysvinit and systemd. Although its not elegant

meta/lib/oeqa/runtime/date.py +1 -0
index 97e8ee4..f73ad1c 100644
@@ -7,6 +7,7 @@ class DateTest(oeRuntimeTest):
    @testcase(211)
    @skipUnlessPassed("test_ssh")
    def test_date(self):
+        (status, output) = self.target.run('systemctl stop
 systemd-timesyncd')
        (status, output) = self.target.run('date +"%Y-%m-%d %T"')
        self.assertEqual(status, 0, msg="Failed to get initial date,

output: %s" % output)
oldDate = output

meta/lib/oeqa/runtime/date.py +0 -0
@@ -21,3 +22,4 @@ class DateTest(oeRuntimeTest):
 
        (status, output) = self.target.run('date -s "%s"' % oldDate)
        self.assertEqual(status, 0, msg="Failed to reset date, output:
Quoted 4 lines

%s" % output)
+ (status, output) = self.target.run('systemctl start
systemd-timesyncd')

Thanks for looking at this, I've rewritten this to use setup/teardown methods and check that systemd is being used and will send the patch shortly.

Ross

#8 lore raw

Re: [PATCH 2/4] rootfs: Do not uninstall update-rc.d

In reply to Khem Raj
Quoted 25 lines · Khem Raj
Quoted 24 lines · Khem Raj
Quoted 23 lines · Phil Blundell
Quoted 7 lines · Khem Raj
Quoted 7 lines · Phil Blundell
Quoted 3 lines · Khem Raj

update-rc.d is now needed by systemd to interact with sysv scripts, so if we have a mixed system, then we can not uninstall update-rc.d as it is going to break systemd functionality

Quoted 5 lines
Quoted 2 lines

The description above implies that you will only uninstall update-rc.d if systemd is present, but that doesn't seem to be what the patch does.

I meant to not say its only for systemd. but that systemd now needs it. so update-rc.d is no longer forcefully removed from image depchain.

Well, right. But wouldn't it be better only to leave it in there if it's actually necessary, i.e. if you have systemd and you also have sysv scripts and the rootfs isn't read-only? As it stands, your patch will cause update-rc.d to be left around for people who aren't using systemd, which seems like a retrograde step.

Quoted 5 lines
Quoted 1 line

makes sense. I can improvise on it and check for all constraints in a follow up.

I have updates this patch in pull branch
https://github.com/kraj/openembedded-core/commit/c4c574dff9ef689c7694a6055907083e7a018d48

Very good, how am I meant to deal with this though? You never published a pull url in your original pull request and I can't figure out how to get a patch file out the github web UI. I guess I can fetch the branch somewhere, then extract it, but if everyone does this its going to be rather annoying for me :(.

I really wish people would just post the v2 on the list.

Incidentally, the wrong version of this has gotten merged, I'll likely revert it then apply this.

Cheers,

Richard

author
#9 lore raw

Re: [PATCH 2/4] rootfs: Do not uninstall update-rc.d

Quoted 32 lines · Richard Purdie
Quoted 28 lines · Khem Raj
Quoted 25 lines · Khem Raj
Quoted 14 lines · Phil Blundell
Quoted 10 lines · Khem Raj
Quoted 9 lines · Phil Blundell
Quoted 3 lines · Khem Raj

update-rc.d is now needed by systemd to interact with sysv scripts, so if we have a mixed system, then we can not uninstall update-rc.d as it is going to break systemd functionality

The description above implies that you will only uninstall update-rc.d if systemd is present, but that doesn't seem to be what the patch does.

Quoted 9 lines
Quoted 2 lines

I meant to not say its only for systemd. but that systemd now needs it. so update-rc.d is no longer forcefully removed from image depchain.

Well, right. But wouldn't it be better only to leave it in there if it's actually necessary, i.e. if you have systemd and you also have sysv scripts and the rootfs isn't read-only? As it stands, your patch will cause update-rc.d to be left around for people who aren't using systemd, which seems like a retrograde step.

makes sense. I can improvise on it and check for all constraints in a follow up.

Quoted 8 lines
Quoted 2 lines

I have updates this patch in pull branch
https://github.com/kraj/openembedded-core/commit/c4c574dff9ef689c7694a6055907083e7a018d48

Very good, how am I meant to deal with this though? You never published a pull url in your original pull request and I can't figure out how to get a patch file out the github web UI. I guess I can fetch the branch somewhere, then extract it, but if everyone does this its going to be rather annoying for me :(.

Its was an update to usual pull request branch that I have published earlier and then notified to the thread some of us have been doing it for long.

I have used create-pull-request script to generate the pull request, it is not supposed to work with github ? looking at the original mail I see the url is really not there. I will see why its happening and if its a pilot error.

From github you can append a .patch or .diff to the commit link and obtain a patch too e.g. above would be

https://github.com/kraj/openembedded-core/commit/c4c574dff9ef689c7694a6055907083e7a018d48.patch

There is no need to revert it. I have pushed a rebased patch to the original so you can pull that.

Quoted 9 lines

I really wish people would just post the v2 on the list.

Incidentally, the wrong version of this has gotten merged, I'll likely revert it then apply this.

Cheers,

Richard

author
#10 lore raw

Re: [PATCH 2/4] rootfs: Do not uninstall update-rc.d

Quoted 47 lines · Khem Raj
Quoted 33 lines · Richard Purdie
Quoted 28 lines · Khem Raj
Quoted 25 lines · Khem Raj
Quoted 14 lines · Phil Blundell
Quoted 10 lines · Khem Raj
Quoted 9 lines · Phil Blundell
Quoted 3 lines · Khem Raj

update-rc.d is now needed by systemd to interact with sysv scripts, so if we have a mixed system, then we can not uninstall update-rc.d as it is going to break systemd functionality

The description above implies that you will only uninstall update-rc.d if systemd is present, but that doesn't seem to be what the patch does.

Quoted 9 lines
Quoted 2 lines

I meant to not say its only for systemd. but that systemd now needs it. so update-rc.d is no longer forcefully removed from image depchain.

Well, right. But wouldn't it be better only to leave it in there if it's actually necessary, i.e. if you have systemd and you also have sysv scripts and the rootfs isn't read-only? As it stands, your patch will cause update-rc.d to be left around for people who aren't using systemd, which seems like a retrograde step.

makes sense. I can improvise on it and check for all constraints in a follow up.

Quoted 8 lines
Quoted 2 lines

I have updates this patch in pull branch
https://github.com/kraj/openembedded-core/commit/c4c574dff9ef689c7694a6055907083e7a018d48

Very good, how am I meant to deal with this though? You never published a pull url in your original pull request and I can't figure out how to get a patch file out the github web UI. I guess I can fetch the branch somewhere, then extract it, but if everyone does this its going to be rather annoying for me :(.

Its was an update to usual pull request branch that I have published earlier and then notified to the thread some of us have been doing it for long.

Quoted 12 lines

I have used create-pull-request script to generate the pull request, it is not supposed to work with github ? looking at the original mail I see the url is really not there. I will see why its happening and if its a pilot error.

From github you can append a .patch or .diff to the commit link and obtain a patch too e.g. above would be

https://github.com/kraj/openembedded-core/commit/c4c574dff9ef689c7694a6055907083e7a018d48.patch

There is no need to revert it. I have pushed a rebased patch to the original so you can pull that.

Quoted 2 lines

I really wish people would just post the v2 on the list.

OK I Will send one.

Quoted 7 lines
Quoted 7 lines

Incidentally, the wrong version of this has gotten merged, I'll likely revert it then apply this.

Cheers,

Richard