when disable sysvinit and specify "systemd" in DISTRO_FEATURES,
the recipe "inhert systemd" will delete "/etc/init.d/opensafd",but
"/etc/init.d/opensafd" is needed to start opensafd.service.
Signed-off-by: Dai Caiyun <daicy.fnst@cn.fujitsu.com>
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
Signed-off-by: Joe MacDonald <joe_macdonald@mentor.com>
When pam is enabled,
1, Customize the proftpd.conf to use pam to authenticate
2, Add proftpd pam configuration file /etc/pam.d/proftpd
Signed-off-by: Mingli Yu <Mingli.Yu@windriver.com>
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
Signed-off-by: Joe MacDonald <joe_macdonald@mentor.com>
When configuring squid with --enable-esi option,
the following error was observed:
[snip]
checking libxml/parser.h usability... no
checking libxml/parser.h presence... no
checking for libxml/parser.h... no
configure: Failed to find libxml2 header file libxml/parser.h
[snip]
ERROR: This autoconf log indicates errors, it looked at host include
and/or library paths while determining system capabilities.
[snip]
It tried to search libxml header file in host path. Set the SYSROOT
to avoid this host contamination.
Signed-off-by: Yue Tao <yue.tao@windriver.com>
Signed-off-by: Yi Zhao <yi.zhao@windriver.com>
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
Signed-off-by: Joe MacDonald <joe_macdonald@mentor.com>
Commit f8dd8cde90 removed the StandardError
option from the proftpd service file, re-introduce it.
Signed-off-by: Joe MacDonald <joe_macdonald@mentor.com>
When using '??=' to set a variable, '+=' will cause
the '??=' value to be discarded. We change how PACKAGECONFIG
is defined. We also do not make sia support a default
since it is not likely to be available.
Signed-off-by: Joe Slater <jslater@windriver.com>
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
Signed-off-by: Joe MacDonald <joe_macdonald@mentor.com>
* Remove the blacklist since the issue is gone with new version
* Remove two CVE patches which have been fixed:
- CVE-2016-3947 and CVE-2016-4553
* Rebased the patch for ptest.
Signed-off-by: Jackie Huang <jackie.huang@windriver.com>
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
Signed-off-by: Joe MacDonald <joe_macdonald@mentor.com>
The recipe for atftp inherits useradd.bbclass and
re-creates the user nobody as a "system" user. This
is not correct and is in vain because this new definition
will not be created on the target.
Signed-off-by: Joe Slater <jslater@windriver.com>
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
Signed-off-by: Joe MacDonald <joe_macdonald@mentor.com>
1) error: file /usr/share/man/man1/mailq.1 from install of postfix-doc
conflicts with file from package esmtp-doc
2) error: file /usr/share/man/man1/newaliases.1 from install of postfix-doc
conflicts with file from package esmtp-doc
3) error: file /usr/share/man/man1/sendmail.1 from install of postfix-doc
conflicts with file from package esmtp-doc
Signed-off-by: Dai Caiyun <daicy.fnst@cn.fujitsu.com>
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
Signed-off-by: Joe MacDonald <joe_macdonald@mentor.com>
base_contains() is a compatibility wrapper and may warn in the future, so
replace all instances with bb.utils.contains().
Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
Signed-off-by: Joe MacDonald <joe_macdonald@mentor.com>
This new release has following changes since version 0.9.0:
lldpd (0.9.2)
Change:
- Ability to add/remove/replace custom TLV from lldpcli.
- LLDP-MED capabilities are displayed differently in lldpcli.
- Limit the maximum depth (5) when trying to apply a VLAN.
- Change JSON output format when using json-c to match Jansson
output.
- Integration tests for the major parts of lldpd, including use of
address and leak sanitizer.
Fix:
- LLDP-MED POE TLV are now displayed in lldpcli.
- Ignore lower link when it is in another namespace.
- Fix various problems with interfaces being enslaved.
- Fix a memory leak when modifying port-related settings.
lldpd (0.9.1)
Change:
- Rework packaging for OS X to make it work with El Capitan. To
simplify a bit, it is not possible anymore to build fat
binaries. Latest version of OS X supporting 32bit was 10.6.
Fix:
- By default, when using port alias as description, use port name
as port ID.
- Miscellaneous fixes with netlink cache.
- Ensure large netlink messages can be received.
Signed-off-by: Fabio Berton <fabio.berton@ossystems.com.br>
Acked-by: Otavio Salvador <otavio@ossystems.com.br>
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
Signed-off-by: Joe MacDonald <joe_macdonald@mentor.com>
The mail group is provided by base-passwd so would always be present. Therefore
drop the uneeded group addition from this recipe.
This works around the recent user cleanup code improvements which meant
this started causing failures for people.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Acked-by: Otavio Salvador <otavio@ossystems.com.br>
Tested-by: Otavio Salvador <otavio@ossystems.com.br>
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
Signed-off-by: Joe MacDonald <joe_macdonald@mentor.com>
ccdv is an internal tool to reduce the deluge Make output
to make finding actual problems easier and it is intended
to be invoked from Makefiles only, it doesn't work for the
cross compiling, so compile it with $BUILD_CC and
corresponding CFLAGS.
And I think we don't need to enable it by default to
reduce our Make output, so add a PACKAGECONFIG for it
but disable it by default.
Signed-off-by: Jackie Huang <jackie.huang@windriver.com>
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
Signed-off-by: Joe MacDonald <joe_macdonald@mentor.com>
There is a build failure while rebuilding postfix
...
| NOTE: make -e MAKEFLAGS= OPT= DEBUG= OPTS= clean
| make -f Makefile.in MAKELEVEL= Makefiles
| (echo "# Do not edit -- this file documents how Postfix was built for your
machine.";/bin/sh makedefs) >makedefs.tmp
| No <db.h> include file found.
| Install the appropriate db*-devel package first.
| Makefile.in:31: recipe for target 'Makefiles' failed
| make: *** [Makefiles] Error 1
| Makefile:21: recipe for target 'Makefiles' failed
| make: *** [Makefiles] Error 2
| ERROR: oe_runmake failed
| ERROR: Function failed: do_configure (log file is located at tmp/work/
core2-64-wrs-linux/postfix/3.0.2-r0/temp/do_configure/log.do_configure.12848)
...
Signed-off-by: Hongxu Jia <hongxu.jia@windriver.com>
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
Signed-off-by: Joe MacDonald <joe_macdonald@mentor.com>
This recipe currently relies on EXTRA_OEMAKE having been to set to
"-e MAKEFLAGS=" in bitbake.conf to operate. It is necessary to make
this explicit so that the default in bitbake.conf can be changed.
Signed-off-by: Mike Crowe <mac@mcrowe.com>
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
Signed-off-by: Joe MacDonald <joe_macdonald@mentor.com>
1) Upgrade opensaf from 4.6.0 to 4.7.0.
2) Delete two patches,since they are not needed any more.
Revert_imma_client_node_replyPending_to_unsigned_char.patch
Fix_GCC_5.1.0_compiler_warning.patch
Signed-off-by: Dai Caiyun <daicy.fnst@cn.fujitsu.com>
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
Signed-off-by: Joe MacDonald <joe_macdonald@mentor.com>
While hostname is numeric, start postfix failed
...
root@qemux86-64:~# hostname 1.2.3.4
root@qemux86-64:~# systemctl restart postfix
|Job for postfix.service failed because the control process exited
with error code. See "systemctl status postfix.service" and
"journalctl -xe" for details.
root@qemux86-64:~# systemctl status postfix -l
Dec 02 08:05:40 1.2.3.4 aliasesdb[535]: /usr/sbin/postconf: fatal: unable to use my own hostname
Dec 02 08:05:41 1.2.3.4 aliasesdb[535]: newaliases: warning: valid_hostname: numeric hostname: 1.2.3.4
Dec 02 08:05:41 1.2.3.4 postfix/sendmail[537]: warning: valid_hostname: numeric hostname: 1.2.3.4
Dec 02 08:05:41 1.2.3.4 aliasesdb[535]: newaliases: fatal: unable to use my own hostname
Dec 02 08:05:42 1.2.3.4 postfix[540]: warning: valid_hostname: numeric hostname: 1.2.3.4
Dec 02 08:05:42 1.2.3.4 postfix[540]: fatal: unable to use my own hostname
...
Refer meta/recipes-core/initscripts/initscripts-1.0/hostname.sh in oe-core,
add check_hostname.sh and invoke it before postfix start, if the hostname
is invalid, set "localhost" to main.cf.
Signed-off-by: Hongxu Jia <hongxu.jia@windriver.com>
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
Signed-off-by: Joe MacDonald <joe_macdonald@mentor.com>
configure the systemd unit file dir, otherwise it will be auto-checked
by 'pkg-config --variable=systemdsystemunitdir systemd', but if systemd
is not built firstly, and the the unit file will not be installed, and
lead to below error:
ERROR: Function failed: SYSTEMD_SERVICE_lldpd value lldpd.service does not exist
and disable sysusersdir, since sysuser is not used currently.
Signed-off-by: Roy Li <rongqing.li@windriver.com>
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
Signed-off-by: Joe MacDonald <joe_macdonald@mentor.com>
Based on this thread:
http://comments.gmane.org/gmane.comp.handhelds.openembedded/70890
it seems that lldpd is a more complete implementation of the LLDP
protocol. Remove this implementation to provide better direction to
newcomers and to attempt to focus support for a single implementation in
the meta-networking layer.
Signed-off-by: Joe MacDonald <joe_macdonald@mentor.com>
It hardcoded to assign vmail's userid and groupid with 5000,
which caused the following useradd/groupadd is increased from
5000 at do_root time. Such as 'tracing' group id is 5001
(tracing group is added in lttng-tools).
It is better to allocate it dynamically.
BTW, you may google to find the hardcoded 5000 in following site:
https://help.ubuntu.com/community/PostfixVirtualMailBoxClamSmtpHowto
In its main.cf, it also hardcoded virtual_uid_maps and virtual_gid_maps.
...
virtual_uid_maps = static:5000
virtual_gid_maps = static:5000
..
But in meta-oe's main.cf_2.0, it assigned virtual_uid_maps and
virtual_gid_maps with a search.
...
virtual_uid_maps = static:`grep vmail /etc/passwd | cut -d ":" -f 3`
virtual_gid_maps = static:`grep vmail /etc/passwd | cut -d ":" -f 4`
...
So please feel free to forget 5000.
Signed-off-by: Hongxu Jia <hongxu.jia@windriver.com>
Signed-off-by: Joe MacDonald <joe_macdonald@mentor.com>
Don't enable GNU atomic operations for all targets, it fails on
powerpc and mips:
AtomicWord.h: undefined reference to `__sync_fetch_and_add_8'
collect2: error: ld returned 1 exit status
Refer to https://gcc.gnu.org/bugzilla/show_bug.cgi?id=56300:
There is no hardware support for 8 bytes atomic operations on
32-bit MIPS targets.
The 32-bit PowerPC fails as well.
Signed-off-by: Wenzong Fan <wenzong.fan@windriver.com>
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
Signed-off-by: Joe MacDonald <joe_macdonald@mentor.com>
1. fix the building failure when host is gcc5
2. fix the building failure when DEBUG_BUILD is 1 by removal of -Werror
Remove -Werror, since it is suitable to develop only; otherwise when
gcc adds -O0 option to compile codes, the error will be generated
since the warning:
//# warning _FORTIFY_SOURCE requires compiling with optimization (-O)
Signed-off-by: Roy Li <rongqing.li@windriver.com>
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
Signed-off-by: Joe MacDonald <joe_macdonald@mentor.com>
1. Remove the squid-change-ksh-referen*.patch which is not needed, since
3.5.7 did not use ksh by default.
2. Update the checksum of COPYING,since the date in it has been changed.
3. Define BUILDCXXFLAGS, otherwise the target gcc options -std=c++11 will
add into it, and lead to building failure since host gcc maybe not
support "-std=c++11"
4. Assume to support GNU atomic operations by default, the running check
on cross-compile setup does not work
5. enable basic auth by checking the DISTRO_FEATURE, and the default
dependency on db, opensasl and openldap nis have been set, so enable
them by default.
Signed-off-by: Roy Li <rongqing.li@windriver.com>
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
Signed-off-by: Joe MacDonald <joe_macdonald@mentor.com>
atftp fails to build with gcc 5.2:
git/tftp_def.h:54:14: warning: inline function 'Strncpy' declared but never defined
git/argz.c:44:8: error: redefinition of 'argz_next'
GCC 5 defaults to -std=gnu11 instead of -std=gnu89. The semantics of
inline function changes. Pass '-std=gnu89' to gcc to compile atftp.
Ref:
https://gcc.gnu.org/gcc-5/porting_to.html
Signed-off-by: Kai Kang <kai.kang@windriver.com>
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
Signed-off-by: Joe MacDonald <joe_macdonald@mentor.com>
Delete ptpd-add-dpaa-etsec-support.patch.The reason is the following:
- The ptpd-add-dpaa-etsec-support.patch does not suitable for new version.
- The upstream didn't accept this patch.
Signed-off-by: Lei Maohui <leimaohui@cn.fujitsu.com>
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
Signed-off-by: Joe MacDonald <joe_macdonald@mentor.com>
Kai Kang