First attempt at a simple wrapper for configuring luv-yocto #43

rpmmisc imports rpm and contains misc rpm utilities related to packaging and determining arches based on the packaging. We should never run across this in the initial version of wic, so remove the dependency. (From OE-Core master rev: 2d59b6eeb418cf23eef3e32b43354b4ab16a40b9) (From OE-Core rev: b16a9de9f5eb2d252ee263a4b2c66c74ff4ff78f) Signed-off-by: Tom Zanussi <[email protected]> Signed-off-by: Richard Purdie <[email protected]> Signed-off-by: Robert Yang <[email protected]> Signed-off-by: Richard Purdie <[email protected]>

mylrlgrab is in grabber, which imports rpm. For current functionality, we don't need to grab urls or import rpm, so remove the dependency. (From OE-Core master rev: 429ecc2afa499df35a1ae9da6f92b88c6f2d8d11) (From OE-Core rev: 1ef27c9dfa28f65458750c0afb2e136c4b79b226) Signed-off-by: Tom Zanussi <[email protected]> Signed-off-by: Richard Purdie <[email protected]> Signed-off-by: Robert Yang <[email protected]> Signed-off-by: Richard Purdie <[email protected]>

We don't currently use rpm functionality, so we don't need to silence rpm warnings. (From OE-Core master rev: dd3cc03d4fa3347f8ef2db23d8ff98bdbdb73baa) (From OE-Core rev: 8827b46d8cb4d6918451bd1c3c278465d8796e4b) Signed-off-by: Tom Zanussi <[email protected]> Signed-off-by: Richard Purdie <[email protected]> Signed-off-by: Robert Yang <[email protected]> Signed-off-by: Richard Purdie <[email protected]>

BaseImageCreator is a base class for DirectImageCreator and others, and imports rpm and grabber (which imports rpm). The various plugins e.g. DirectPlugin import the creators and therefore these dependencies, which manifest at run-time as e.g.: Warning: Failed to load plugin imager/direct_plugin: No module named rpm (From OE-Core master rev: a1e24c4a5f5771b7ad35e53ce96c6d82212e4d7e) (From OE-Core rev: f5587ec7e7f925b321b9bfe6923be0879dadb2aa) Signed-off-by: Tom Zanussi <[email protected]> Signed-off-by: Richard Purdie <[email protected]> Signed-off-by: Robert Yang <[email protected]> Signed-off-by: Richard Purdie <[email protected]>

We don't currently use LiveCDImageCreator, but it makes calls when initialized via the plugin interface to rpmmisc module functions, which we don't want the dependency on. To make it (and LiveUSBImageCreator) happy, we give it the dummy "i386" value for now. (From OE-Core master rev: e10ae516cfc10900ed12e84c743e3a7127372135) (From OE-Core rev: a3cc57cf3116c997ec11dd3cbfa3b0d615e5dabc) Signed-off-by: Tom Zanussi <[email protected]> Signed-off-by: Richard Purdie <[email protected]> Signed-off-by: Robert Yang <[email protected]> Signed-off-by: Richard Purdie <[email protected]>

Current functionality doesn't make use of kpartx, mount, or unmount, and we use native mkswap, so remove the binary checks for those. (From OE-Core master rev: 76293d2d6bbdeacd7b34f39f26fb97c3d7f9496f) (From OE-Core rev: 0ed290b81e1c3b781170033f50db01ddfff14784) Signed-off-by: Tom Zanussi <[email protected]> Signed-off-by: Richard Purdie <[email protected]> Signed-off-by: Robert Yang <[email protected]> Signed-off-by: Richard Purdie <[email protected]>

find_binary_path() is useful, but if the binary isn't found, it prints a stacktrace and a less-than-useful message. Users complain when they get stacktraces for things they can act on, so remove the stacktrace and tell the user what the problem is. (From OE-Core master rev: 0d9eef0eaa267500e8eedab8b72ddf24eb0516db) (From OE-Core rev: 8a17195c9be38815e9ae431bcb18f66a4ad2cdcb) Signed-off-by: Tom Zanussi <[email protected]> Signed-off-by: Richard Purdie <[email protected]> Signed-off-by: Robert Yang <[email protected]> Signed-off-by: Richard Purdie <[email protected]>

Remove unnecessary pseudo exports i.e. PSEUDO_DISABLED and move the setup to the top-level prepare_rootfs(). (From OE-Core master rev: 4bf11cd7d7301da664c098c8a0ae9c0294a6f423) (From OE-Core rev: 8d661f578276c70e1671edfc810aa4dad97de970) Signed-off-by: Tom Zanussi <[email protected]> Signed-off-by: Richard Purdie <[email protected]> Signed-off-by: Robert Yang <[email protected]> Signed-off-by: Richard Purdie <[email protected]>

Test that filter backup module files (files starting with ~) was accidentally reversed in e6039e6e3b98d6ab91252a5012d76279b1fac6e8, this patch restore initial behavior. (From OE-Core master rev: b2eb846ee12989add7a7ca8bbf45f293a3a7e56d) (From OE-Core rev: 00ff958fec53e55cc475c1b31fb9813d97872ceb) Signed-off-by: Michaël Burtin <[email protected]> Signed-off-by: Richard Purdie <[email protected]> Signed-off-by: Robert Yang <[email protected]> Signed-off-by: Richard Purdie <[email protected]>

midori depends on webkit-gtk which could not build for mips64. [YOCTO #5141] (From OE-Core master rev: abadeb934d4f41288c4fde6a4e5df2b124326326) (From OE-Core rev: 672cd50a39a697b53c337a79c34fab05b48e0920) Signed-off-by: Jackie Huang <[email protected]> Signed-off-by: Richard Purdie <[email protected]> Signed-off-by: Robert Yang <[email protected]> Signed-off-by: Richard Purdie <[email protected]>

The PACKAGECONFIG flags were iterated over using dict.items(), but this returns the items in an undefined order. As this order determines the EXTRA_OECONF append order, we can get EXTRA_OECONF which are functionally equivalent, but whose contents differ, resulting in not using shared state archives we should be using. (From OE-Core master rev: 843a5dd8f8f0461e286d9fdb3ba55205b4275f88) (From OE-Core rev: 73f77c195e1af3df594eecce2cab47ee963d5c2e) Signed-off-by: Christopher Larson <[email protected]> Signed-off-by: Richard Purdie <[email protected]> Signed-off-by: Robert Yang <[email protected]> Signed-off-by: Richard Purdie <[email protected]>

mDNS name resolution is a key part of mDNS, so if the DISTRO_FEATURE is enabled then install libnss-mdns. (From OE-Core master rev: ef2ee68778be8e5336cd33ab6551bce1d56047b6) (From OE-Core rev: fb72bebc999aef7bb006a7ba273eaa96376a4016) Signed-off-by: Ross Burton <[email protected]> Signed-off-by: Saul Wold <[email protected]> Signed-off-by: Richard Purdie <[email protected]> Signed-off-by: Robert Yang <[email protected]> Signed-off-by: Richard Purdie <[email protected]>

If the same username exists on both target and the build host, but the uids differ, and we start target via NFS, then the uid for the user will be incorrect on target. For example, if postfix's uid on host is 119 and on target is 1024, then if we start target via NFS, the uid for postfix will be 119. The root cause is that when we use runqemu-extract-sdk to generate the NFS rootfs for later use, the tar command will respect the username instead of uid. So if PSEUDO_PASSWD environment is not set correctly, the host /etc/passwd will be used, resulting in wrong uids. The situation for gid is completely analogous to that of uid. It's almost impossible for the runqemu-extract-sdk to guess the correct location of the needed password file merely based on the target tarball name. This patch solves this problem by adding the '--numeric-owner' option to the tar command so that the uid/gid will be used when extracting the tarball using runqemu-extract-sdk. In this situation, we'll always get the correct uid/gid after extracting the tarball. [YOCTO #5364] (From OE-Core master rev: acce6ff1a77cfd29e3868faa89b120becb58bbbf) (From OE-Core rev: c2baac739a521d1edd408a24f6b1fece8f755218) Signed-off-by: Chen Qi <[email protected]> Signed-off-by: Saul Wold <[email protected]> Signed-off-by: Richard Purdie <[email protected]> Signed-off-by: Robert Yang <[email protected]> Signed-off-by: Richard Purdie <[email protected]>

Python 2.4 does not support the 'b' string literal or the keyword 'as' in exception handling. Python 3 does not accept the old method of exception handling and defaults to unicode. The b() function converts strings to bytes on Python 3 and using sys.exc_info() avoids the exception handling syntax. (From OE-Core master rev: 1e2ec5f576f167673d7980737826987fefdc74a9) (From OE-Core rev: 343127f2f81be337596d3eacbbc92278e82ce574) Signed-off-by: Konrad Scherer <[email protected]> Signed-off-by: Saul Wold <[email protected]> Signed-off-by: Richard Purdie <[email protected]> Signed-off-by: Robert Yang <[email protected]> Signed-off-by: Richard Purdie <[email protected]>

As the same reason to powerpc64, mips64 also need the flag. (From OE-Core master rev: d6f3cb0d71c3b6739365f085b6d5a5e20f329fa5) (From OE-Core rev: 9c4b604ea0d81bc1de224b35ae160f87be6bcf7b) Signed-off-by: Wenzong Fan <[email protected]> Signed-off-by: Richard Purdie <[email protected]> Signed-off-by: Robert Yang <[email protected]> Signed-off-by: Richard Purdie <[email protected]>

An incorrect process name in the nfsserver initscript prevented rpc.statd from being shut down. root@qemux86-64:~# /etc/init.d/nfsserver start creating NFS state directory: done starting 8 nfsd kernel threads: done starting mountd: done starting statd: done root@qemux86-64:~# ps | grep rpc.statd 650 root 10532 S /usr/sbin/rpc.statd 654 root 4720 S grep rpc.statd root@qemux86-64:~# /etc/init.d/nfsserver stop stopping statd: done stopping mountd: done stopping nfsd: done root@qemux86-64:~# ps | grep rpc.statd 650 root 10532 S /usr/sbin/rpc.statd 662 root 4720 S grep rpc.statd As this daemon drops a pid file,simply use that instead. Also add some initialization checks so the daemons are not left partially started in the absence of kernel nfsd support. (From OE-Core master rev: 37e70a28e9cfc773bd70f09d7129295ce891ae18) (From OE-Core rev: 5f22bad97a3bacb87cefb54ffd785d359c58aec0) Signed-off-by: Andy Ross <[email protected]> Signed-off-by: Qiang Chen <[email protected]> Signed-off-by: Richard Purdie <[email protected]> Signed-off-by: Robert Yang <[email protected]> Signed-off-by: Richard Purdie <[email protected]>

…eads nfsserver restart without killing kernel threads worked when portmap was the rpc publishing process and portmap was restarted. When rpcbind replaces portmap, nfsserver restart in this way does not work after an rpcbind restart. Steps to reproduce: 1). Make ext3 filesystem image on local host. cd /root dd if=/dev/zero of=test bs=1024K count=50 mkfs.ext3 -F test 2). runqemu qemux86-64 mkdir /mnt/wrtest mount -t ext3 -o loop test /mnt/wrtest echo "/mnt/wrtest *(sync,rw,no_subtree_check,no_root_squash)" > /etc/exports /etc/init.d/rpcbind restart /etc/init.d/nfsserver restart showmount -e localhost mkdir wrtest mount -t nfs localhost:/mnt/wrtest wrtest mount: mounting localhost:/mnt/wrtest on wrtest failed: Connection refused Modifying the nfsserver script to kill and restart kernel threads on restart makes the problem go away and is consistent with current RHEL/SUSE and Ubuntu/Debian mechanisms of handling the nfs server. (From OE-Core master rev: 1a96b8d7dfc490fc61bbd470a8b09065750cd563) (From OE-Core rev: d1b5e944656807c9db9cbe5d08d7b4bd8daeb826) Signed-off-by: Rich Dubielzig <[email protected]> Signed-off-by: Qiang Chen <[email protected]> Signed-off-by: Richard Purdie <[email protected]> Signed-off-by: Robert Yang <[email protected]> Signed-off-by: Richard Purdie <[email protected]>

(From OE-Core master rev: 89d7d46947d9bb8c7bf568c65e52d5bbe159027f) (From OE-Core rev: 7c6504c6c059ba6b37f88143801ac8137267cf83) Signed-off-by: Konrad Scherer <[email protected]> Signed-off-by: Richard Purdie <[email protected]> Signed-off-by: Robert Yang <[email protected]> Signed-off-by: Richard Purdie <[email protected]>

When installing the SDK to another location than the default one, qmake will look for libraries, headers, etc. in the default location. That's because the paths are hard-coded in the binary itself. Luckily, QT allows to override this using a qt.conf file installed in the same directory with the application executable. However, we already have a patch that allows for the installation of qt.conf in another place and read the location from QT_CONF_PATH environment variable. Hence, install qt.conf in ${sysconfdir}. This will allow other apps, that use QLibraryInfo class, to find it. [YOCTO #5339] (From OE-Core master rev: 23f88695683a8e428375a8ccb6be935347a8768c) (From OE-Core rev: 0e71811a1c3285a71cbca682cb62c1563d3e74ee) Signed-off-by: Laurentiu Palcu <[email protected]> Signed-off-by: Richard Purdie <[email protected]> Signed-off-by: Robert Yang <[email protected]> Signed-off-by: Richard Purdie <[email protected]>

This will allow apps using QLibraryInfo class to find qt.conf. [YOCTO #5339] (From OE-Core master rev: fffa4c37c49b169f663d28612b9251819cef9577) (From OE-Core rev: 8dc1d62c5dc161ba606cebe27f6fe900699646f7) Signed-off-by: Laurentiu Palcu <[email protected]> Signed-off-by: Richard Purdie <[email protected]> Signed-off-by: Robert Yang <[email protected]> Signed-off-by: Richard Purdie <[email protected]>

There is no need for += when using append hence removed and added a leading space appropriately (From OE-Core master rev: fb9cde0fc1a54b073edf5979f4cb7dc297b790fd) (From OE-Core rev: 586db07af01da9d7772b7088a20886b506e09422) Signed-off-by: Khem Raj <[email protected]> Signed-off-by: Richard Purdie <[email protected]> Signed-off-by: Robert Yang <[email protected]> Signed-off-by: Richard Purdie <[email protected]>

libpseudo.so is always installed into ${prefix}/lib/, not ${libdir}, so fix these paths; and skip libdir WARN_QA checking to ignore the warning in 64bit and multilib enabled system (From OE-Core master rev: 47c7850c025994685aa1811057f4f9a5f0f2a3ae) (From OE-Core rev: 1929d4ef17652a3eb825942041908d679773244f) Signed-off-by: Roy Li <[email protected]> Signed-off-by: Richard Purdie <[email protected]> Signed-off-by: Robert Yang <[email protected]> Signed-off-by: Richard Purdie <[email protected]>

- Wayland support depends on wayland-egl, which is provided by mesa. (From OE-Core master rev: a1a379b3c9728a06b086b4c1f06f663f54d7d37d) (From OE-Core rev: 8c75d888a5e4cf7fc2c92df730d80224f5ffa99a) Signed-off-by: Andreas Oberritter <[email protected]> Signed-off-by: Saul Wold <[email protected]> Signed-off-by: Richard Purdie <[email protected]> Signed-off-by: Robert Yang <[email protected]> Signed-off-by: Richard Purdie <[email protected]>

While ELF_C_RDWR_MMAP was used, elf_begin invoked mmap() to map file into memory. While the file's bss Offset has a large number, elf_update caculated file size by __elf64_updatenull_wrlock and the size was enlarged. In this situation, elf_update invoked ftruncate to enlarge the file, and memory size (elf->maximum_size) also was incorrectly updated. There was segment fault in elf_end which invoked munmap with the length is the enlarged file size, not the mmap's length. Before the above operations, invoke elf_begin/elf_update/elf_end with ELF_C_RDWR and ELF_F_LAYOUT set to enlarge the above file, it could make sure the file is safe for the following elf operations. [YOCTO #5356] https://bugzilla.redhat.com/show_bug.cgi?id=1019707 https://bugzilla.redhat.com/show_bug.cgi?id=1020842 (From OE-Core master rev: 35c8b1ac7c3b1e4209b1e30d1dbd1a457286b97b) (From OE-Core rev: a82322a982dc97ebc95f3fc45f9ad98bed947ad9) Signed-off-by: Hongxu Jia <[email protected]> Signed-off-by: Saul Wold <[email protected]> Signed-off-by: Richard Purdie <[email protected]> Signed-off-by: Robert Yang <[email protected]> Signed-off-by: Richard Purdie <[email protected]>

When using the tar executable in the buildtools, tar will execute gzip. If this happens before zlib-native is built, then the gzip on the host will be used and can fail if the libz in the buildtools is not compatible. Adding pigz to the build tools avoids this host contamination. (From OE-Core master rev: af6424e8c2bf3a938fddabc669c0956d68964ed0) (From OE-Core rev: dd9945dd510d6e7764721bec5573591a0ad69ba4) Signed-off-by: Konrad Scherer <[email protected]> Signed-off-by: Saul Wold <[email protected]> Signed-off-by: Richard Purdie <[email protected]> Signed-off-by: Robert Yang <[email protected]> Signed-off-by: Richard Purdie <[email protected]>

This seems to be an obsolete check - we don't have any problems with image creation under selinux, so remove it. (From OE-Core master rev: 12e81eceab9e0a483765566ad3791b14718195b5) (From OE-Core rev: 28830d3988047023d3b47bcaf320f5efa4428da6) Signed-off-by: Tom Zanussi <[email protected]> Signed-off-by: Saul Wold <[email protected]> Signed-off-by: Richard Purdie <[email protected]> Signed-off-by: Robert Yang <[email protected]> Signed-off-by: Richard Purdie <[email protected]>

(From OE-Core master rev: 3af8f2e0697a9523d3b505ba4c48eca35f6de3a9) (From OE-Core rev: 438032411ea5d71a33b7b030752193867a90b9f7) Signed-off-by: Jonathan Liu <[email protected]> Signed-off-by: Saul Wold <[email protected]> Signed-off-by: Richard Purdie <[email protected]> Signed-off-by: Robert Yang <[email protected]> Signed-off-by: Richard Purdie <[email protected]>

ldd sysroots/x86_64-linux/usr/bin/gdk-pixbuf-query-loaders.real <snip> libz.so.1 => /sysroots/x86_64-linux/usr/bin/../../usr/lib/libz.so.1 (0x00007fab55393000) If zlib-native has not been unpacked, host libz is used which can fail. (From OE-Core master rev: 8422c759ae674856aaaee176eab5a395a620443c) (From OE-Core rev: b9ae15b45768d25c44a9484b4a156a15da548bd9) Signed-off-by: Konrad Scherer <[email protected]> Signed-off-by: Saul Wold <[email protected]> Signed-off-by: Richard Purdie <[email protected]> Signed-off-by: Robert Yang <[email protected]> Signed-off-by: Richard Purdie <[email protected]>

(From OE-Core master rev: 9032c10cc882a96acdfd0739f090d121ab625a18) (From OE-Core rev: c191cb79019482a5c6a404e02184bae40ff9f84a) Signed-off-by: Roy Li <[email protected]> Signed-off-by: Saul Wold <[email protected]> Signed-off-by: Richard Purdie <[email protected]> Signed-off-by: Robert Yang <[email protected]> Signed-off-by: Richard Purdie <[email protected]>

/etc/init.d/dbus-1 use "set -e" to let the script exit when any command failes. This will cause "/etc/init.d/dbus-1 status" command can't display messages when dbus is stopped. (From OE-Core master rev: 9844b5e2a544b2c2f76aac497c3a2cdfcc46577c) (From OE-Core rev: 486d5d7e891df3fb2ce8d975d13625b11334814b) Signed-off-by: Lu Chong <[email protected]> Signed-off-by: Saul Wold <[email protected]> Signed-off-by: Richard Purdie <[email protected]> Signed-off-by: Robert Yang <[email protected]> Signed-off-by: Richard Purdie <[email protected]>

* Add the openssl-conf package to the list of packages to be created. This package contains the openssl.cnf file which is used by both the openssl executable in the openssl package and the libcrypto library. * This is to avoid messages like: WARNING: can't open config file: /usr/lib/ssl/openssl.cnf * When running "openssl req" to request and generate a certificate the command will fail without the openssl.cnf file being installed on the target system. * Made this package an RRECOMMENDS for libcrypto since: * libcrypto is a RDEPENDS for the openssl package * Users can specify a configuration file at another location so it is not stricly required and many commands will work without it (with warnings) (From OE-Core master rev: 5c3ec044838e23539f9fe4cc74da4db2e5b59166) (From OE-Core rev: bf6ef555caf92b2a013f15d258bf40997247a150) Signed-off-by: Chase Maupin <[email protected]> Signed-off-by: Qiang Chen <[email protected]> Signed-off-by: Saul Wold <[email protected]> Signed-off-by: Richard Purdie <[email protected]> Signed-off-by: Robert Yang <[email protected]> Signed-off-by: Richard Purdie <[email protected]>

Allow the user to provide additional packages to this image. This lets core-image-basic behave like all other core-image* recipes (which do support CORE_IMAGE_EXTRA_INSTALL), as well as match the documentation which suggests this as the mode to extend any core-image* image. v2 - drop redundant setting of CORE_IMAGE_EXTRA_INSTALL (From OE-Core master rev: 5faabf398819d40b55c46bc83ae03942d115024b) (From OE-Core rev: cdda59d2850b163ced4dfc847c4e114f8592ae52) Signed-off-by: Gary Thomas <[email protected]> Signed-off-by: Richard Purdie <[email protected]> Signed-off-by: Robert Yang <[email protected]> Signed-off-by: Richard Purdie <[email protected]>

BitBake currently allows using the same quotes outside and inside the value, but it isn't really right, looks odd and might stop working in future. (From OE-Core master rev: 0af9cf31851896276a219170001047406f45de50) (From OE-Core rev: 4051c98ebfe79614d1284b38442d5a3290bb4ad1) Signed-off-by: Paul Eggleton <[email protected]> Signed-off-by: Richard Purdie <[email protected]> Signed-off-by: Robert Yang <[email protected]> Signed-off-by: Richard Purdie <[email protected]>

in commit fe039170236080291c0220476a5809774f82ee5c Author: Muhammad Shakeel <[email protected]> Date: Wed Oct 2 10:55:32 2013 +0000 systemd-compat-units: Use correct run-postinsts script link OE-Core commit 75a1492 has moved run-postinsts script execution from S98 to S99 in rcS.d. run-postinsts.service should check for this script and run it on first boot rather than S98run-postinsts, which is for opkg/dpkg. the link was corrected but the mentioned commit is not available. Instead of reverting, we use the same variable as opkg for init script ordering and drop a note in case somebody wants to change default. (From OE-Core master rev: 7aabc9408fb382f0ae39f9932b6d9ac391528b76) (From OE-Core rev: 91a51e10392c0f802d6c329ad2b11ae82c55f171) Signed-off-by: Andreas Müller <[email protected]> Signed-off-by: Saul Wold <[email protected]> Signed-off-by: Richard Purdie <[email protected]> Signed-off-by: Robert Yang <[email protected]> Signed-off-by: Richard Purdie <[email protected]>

PowerPC toolchains can use the OS "linux" or "linux-gnuspe". This patch links them together so the one cross-canadian toolchain can support both. GCC_FOR_TARGET is set for the GCC recipe as otherwise configure can pick up an incorrect value. [YOCTO #5354] (From OE-Core master rev: a1d6331238982b0c5d39b0a18794f6654b00d46a) (From OE-Core rev: b114b045687776ebc5c805e1f19758e7e37eebf2) Signed-off-by: Richard Purdie <[email protected]> Signed-off-by: Saul Wold <[email protected]> Signed-off-by: Richard Purdie <[email protected]> Signed-off-by: Robert Yang <[email protected]> Signed-off-by: Richard Purdie <[email protected]>

This class points the inheritor, if it is a target, to directories in the target sysroot, so we want to be sure the .vapi files are there. (From OE-Core master rev: 2da8bbd47686f54efeec521d521f176f6aeb8d39) (From OE-Core rev: e68307c3fb0a02efe5e0789a58686f343c842707) Signed-off-by: Joe Slater <[email protected]> Signed-off-by: Saul Wold <[email protected]> Signed-off-by: Richard Purdie <[email protected]> Signed-off-by: Robert Yang <[email protected]> Signed-off-by: Richard Purdie <[email protected]>

Without this patch filenames containing spaces do not get into the final ext2/3/4 filsystem. [YOCTO #5401] (From OE-Core master rev: 1350b461ed0c9d4afa1ab909a5b1ff60fb160c97) (From OE-Core rev: 62d01b10508f86ca825ebc24773dfa2b485b4292) Signed-off-by: Saul Wold <[email protected]> Signed-off-by: Richard Purdie <[email protected]> Signed-off-by: Robert Yang <[email protected]> Signed-off-by: Richard Purdie <[email protected]>

The vtX terminfo files aren't being copied on systems where bash isn't the default shell (debian, etc.). I removed the bash specific syntax so the files are properly copied on these systems. (From OE-Core master rev: edd7d53c6149b27d5636a458db91650c8c400612) (From OE-Core rev: ccb25cd9df9ed4b84b74170336cfde5c1dc3d013) Signed-off-by: Seth Bollinger <[email protected]> Signed-off-by: Saul Wold <[email protected]> Signed-off-by: Richard Purdie <[email protected]> Signed-off-by: Robert Yang <[email protected]> Signed-off-by: Richard Purdie <[email protected]>

The default value of SYSLINUX_SERIAL_TTY is not correct. It should be console=ttyS0,115200 else the boot string generated in the syslinux menus for the serial choice is not correct. The kernel boot parameters will get set to: /vmlinuz initrd=/initrd LABEL=boot root=/dev/ram0 ttyS0,115200 Note that the above is missing the "console=" The default value will now work the same as the value found in grub-efi.bbclass. (From OE-Core master rev: fbc864241933c6f40814f47e7a85dd71ce255393) (From OE-Core rev: 042e0f59d058c5caad4a3b8e8f08ec724d5c837b) Signed-off-by: Jason Wessel <[email protected]> Signed-off-by: Saul Wold <[email protected]> Signed-off-by: Richard Purdie <[email protected]> Signed-off-by: Robert Yang <[email protected]> Signed-off-by: Richard Purdie <[email protected]>

When boot-directdisk class is used and EFI boot is set the grub-efi-${TRANSLATED_TARGET_ARCH}-native need to be dependent. Allowing GRUB_IMAGE to be created and bootia32.efi got from the image directory. (From OE-Core master rev: b9778975db410b8cd01ef6854c7cd3ea22a0b5b7) (From OE-Core rev: b14ba80d22f4892a4d9269dbf982b2f88109da98) Signed-off-by: Joao Henrique Ferreira de Freitas <[email protected]> Signed-off-by: Saul Wold <[email protected]> Signed-off-by: Richard Purdie <[email protected]> Signed-off-by: Robert Yang <[email protected]> Signed-off-by: Richard Purdie <[email protected]>

Some revs of the EFI firmware + shell do not automatically setup the path in a such a way as to execute a binary without an absolute reference like "FS0:\EFI\BOOT\bootx64.efi". All the versions that I have tested work properly by simply calling the binary which is in the EFI\BOOT directory by name like "bootx64.efi". The error you see on the console looks like the following: startup.nsh> EFI\BOOT\bootx64.efi 'EFI\BOOT\bootx64.efi' is not recognized as an internal or external command, operable program, or batch file Shell> This patch simply drops the EFI\BOOT for greater compatibility. (From OE-Core master rev: 754b52ea7a3cdf8e7e939a314525d16c4dfb52cb) (From OE-Core rev: a65db42d4b0c6636ee6538a28930e4d19e170d18) Signed-off-by: Jason Wessel <[email protected]> Signed-off-by: Saul Wold <[email protected]> Signed-off-by: Richard Purdie <[email protected]> Signed-off-by: Robert Yang <[email protected]> Signed-off-by: Richard Purdie <[email protected]>

Some further structure definitions are needed in include/linux/if_pppol2tp.h for IPv6 support, else we would get the error as below: In file included from plugin.c:53:0: bitbake_build/tmp/sysroots/intel-x86-64/usr/include/linux/if_pppox.h:84:26: error: field 'pppol2tp' has incomplete type struct pppol2tpin6_addr pppol2tp; ^ bitbake_build/tmp/sysroots/intel-x86-64/usr/include/linux/if_pppox.h:99:28: error: field 'pppol2tp' has incomplete type struct pppol2tpv3in6_addr pppol2tp; ^ make[2]: *** [plugin.o] Error 1 (From OE-Core master rev: 73d08c4bf12e2cc4f291cb018d00b26a5a573be4) (From OE-Core rev: 398bae0d288f488020108c7d95bd376249e6ecbf) Signed-off-by: Lu Chong <[email protected]> Signed-off-by: Saul Wold <[email protected]> Signed-off-by: Richard Purdie <[email protected]> Signed-off-by: Robert Yang <[email protected]> Signed-off-by: Richard Purdie <[email protected]>

zisofs-tools links against zlib. Fixes [YOCTO #5420]. (From OE-Core master rev: 6058a34f4f0d907a3a065a0323ca085df690bd9b) (From OE-Core rev: 5d9861dc320ac66b56ac70463fd646e0fe4a1baf) Signed-off-by: Paul Eggleton <[email protected]> Signed-off-by: Richard Purdie <[email protected]> Signed-off-by: Robert Yang <[email protected]> Signed-off-by: Richard Purdie <[email protected]>

This check was looking for /run/mdadm on the host system, this check is optional so disable it. [YOCTO #5447] (From OE-Core master rev: d62882794890eeee8e8d5c9ba4837ec77a58d787) (From OE-Core rev: 6c212e229ba1be0c536e7c36ebe2a3e8b2ef84b4) Signed-off-by: Saul Wold <[email protected]> Signed-off-by: Richard Purdie <[email protected]> Signed-off-by: Robert Yang <[email protected]> Signed-off-by: Richard Purdie <[email protected]>

Avoid infinite loop if the last record in EXTRA_USRES_PARAMS doesn't end with a semicolon. It's possible the the users will write configurations like below. INHERIT += "extrausers" EXTRA_USERS_PARAMS = "useradd tester; useradd developer" In such situation, the do_rootfs task will enter an infinite loop. An infinite loop is never acceptable. This patch fixes the above problem. (From OE-Core master rev: bf4fb345a9db306fa4c7211b7e6795334a649dd5) (From OE-Core rev: 05f9c15abb0705cd9f8aa28bfb3016485730b643) Signed-off-by: Chen Qi <[email protected]> Signed-off-by: Richard Purdie <[email protected]> Signed-off-by: Robert Yang <[email protected]> Signed-off-by: Richard Purdie <[email protected]>

[YOCTO #5391] (From OE-Core master rev: 1969ff9081d06109b7e8e7f9331f31fcb611c393) (From OE-Core rev: bdc49826edf05233c6d506de404861a661043b1b) Signed-off-by: Jonathan Liu <[email protected]> Signed-off-by: Saul Wold <[email protected]> Signed-off-by: Richard Purdie <[email protected]> Signed-off-by: Robert Yang <[email protected]> Signed-off-by: Richard Purdie <[email protected]>

(From OE-Core master rev: a678243d6e4add90c1e9459da42de34d3724db5d) (From OE-Core rev: 9d59ceab6e58b422caf8ad7a306ed546316d8c3a) Signed-off-by: Jonathan Liu <[email protected]> Signed-off-by: Saul Wold <[email protected]> Signed-off-by: Richard Purdie <[email protected]> Signed-off-by: Robert Yang <[email protected]> Signed-off-by: Richard Purdie <[email protected]>

This patch fixes below issues: 1. Make can't exit while compilation error occurs in subdir for plugins building. 2. If build ppp with newer kernel (3.10.10), it will pick 'if_pppox.h' from sysroot-dir and 'if_pppol2tp.h' from its own source dir, this cause below build errors: bitbake_build/tmp/sysroots/intel-x86-64/usr/include/linux/if_pppox.h:84:26: error: field 'pppol2tp' has incomplete type struct pppol2tpin6_addr pppol2tp; ^ bitbake_build/tmp/sysroots/intel-x86-64/usr/include/linux/if_pppox.h:99:28: error: field 'pppol2tp' has incomplete type struct pppol2tpv3in6_addr pppol2tp; ^ The 'sysroot-dir/if_pppox.h' enabled ipv6 support but the 'source-dir/if_pppol2tp.h' lost related structure definitions, we should use both header files from sysroots to fix this build failure. (From OE-Core master rev: b536824ea64b8d6729b830738bce637fc815e832) (From OE-Core rev: 16968759d39534fb9a703903c6de65535d57777b) Signed-off-by: Lu Chong <[email protected]> Signed-off-by: Saul Wold <[email protected]> Signed-off-by: Richard Purdie <[email protected]> Signed-off-by: Robert Yang <[email protected]> Signed-off-by: Richard Purdie <[email protected]>

Previously, our system had no boot log even if the bootlogd daemon was started correctly. The root cause is that the log file doesn't exist when starting the bootlogd. Add '-c' option to bootlogd so that it will create the boot log if it doesn't exist. [YOCTO #5273] (From OE-Core master rev: 6059be3ab60b8ab463d438c47bb17553d184a790) (From OE-Core rev: f676a69dee845cfd6de7a0cc8bd0bb813a8dffc0) Signed-off-by: Chen Qi <[email protected]> Signed-off-by: Saul Wold <[email protected]> Signed-off-by: Richard Purdie <[email protected]> Signed-off-by: Robert Yang <[email protected]> Signed-off-by: Richard Purdie <[email protected]>

Weston 1.3 needs to run on a VT, which is typically handled by weston-launch. Currently weston-init doesn't use weston-launch as that depends on the (non-default) pam DISTRO_FEATURE, so depend on kbd and use openvt directly. This also fixes problems caused by the init script blocking until Weston exits, which meant that later init scripts were not actually running. (From OE-Core master rev: 3726eb29cfa79a4a1fbdbcaa96f770063c482858) (From OE-Core rev: d79f7846f5d538f6f835f52686fd2c749cb1b70f) Signed-off-by: Ross Burton <[email protected]> Signed-off-by: Saul Wold <[email protected]> Signed-off-by: Richard Purdie <[email protected]> Signed-off-by: Robert Yang <[email protected]> Signed-off-by: Richard Purdie <[email protected]>

…otfs File command in some version could print extra space between "LSB" and "executable" - it causes mklibs can't find any executables using grep "LSB executable". Fix the grep pattern to catch multiple spaces. (From OE-Core master rev: a52ef8c5dcd71f39bb48c71fb868cc0db662560e) (From OE-Core rev: 78c22d9087b3058dd947f21bd24fa621aba7dd6d) Signed-off-by: Lei Liu <[email protected]> Signed-off-by: Richard Purdie <[email protected]> Signed-off-by: Robert Yang <[email protected]> Signed-off-by: Richard Purdie <[email protected]>

mklibs requires the "dpkg-architecture" utility to work. Add dependency on dpkg-native. (From OE-Core master rev: 9811641e95dd7e1514eb41900e033a0548bd13d8) (From OE-Core rev: f4a5005518e2384b22593063010a3fa1179f5861) Signed-off-by: Lei Liu <[email protected]> Signed-off-by: Richard Purdie <[email protected]> Signed-off-by: Robert Yang <[email protected]> Signed-off-by: Richard Purdie <[email protected]>

sshd status command results in error prompt: root@qemu0:~# /etc/init.d/sshd status /usr/sbin/sshd (pid 1199) is running... /etc/init.d/sshd: line 100: return: can only `return' from a function or sourced script "service --status-all" command also display wrong status for sshd. This commit fix this error prompt and make service command display right status for sshd. (From OE-Core master rev: e7cf83ec3f39a7c41e38c6030b0d903fa7d37b2a) (From OE-Core rev: 1b5409b5b060459f15c32c89b1983122b2126f84) Signed-off-by: Qiang Chen <[email protected]> Signed-off-by: Richard Purdie <[email protected]> Signed-off-by: Robert Yang <[email protected]> Signed-off-by: Richard Purdie <[email protected]>

Using sysvinit testing service status, nfsserver status allways display as [?] unknown. This is because sysvinit package check whether service's init script supporting status function or not by: grep -qs "\Wstatus)" "$SERVICE" So, this commit modified the indent for status etc, as most service's init script does. (From OE-Core master rev: a6b02fe439fa13c8482383fba2bfdcb0e9742141) (From OE-Core rev: f9be1ec26cf1f313d7c22e26475296b0362c25ea) Signed-off-by: Qiang Chen <[email protected]> Signed-off-by: Saul Wold <[email protected]> Signed-off-by: Richard Purdie <[email protected]> Signed-off-by: Robert Yang <[email protected]> Signed-off-by: Richard Purdie <[email protected]>

irattach init script restart faulty logic prevents irda daemon from restart correctly. root@qemu0:~# /etc/init.d/irattach restart Restarting IrDA: Terminated root@qemu0:~# ps aux | grep irattach root 541 0.0 0.2 2400 612 ttyS0 S+ 09:05 0:00 grep irattach As above shows, irattach not started after executing restart command. This commit changed the restart command logic: firstly stop, then start. Prompt telling user irattach start successfully or failure also added. (From OE-Core master rev: 39f266138b972b550979909b235a5779828d7d89) (From OE-Core rev: 37ceb9ad0c45aca458e2ff4770b8a0535286a78e) Signed-off-by: Qiang Chen <[email protected]> Signed-off-by: Saul Wold <[email protected]> Signed-off-by: Richard Purdie <[email protected]> Signed-off-by: Robert Yang <[email protected]> Signed-off-by: Richard Purdie <[email protected]>

backport the patch from busybox upstream to fix the auto insert newline issue. busybox defect: https://bugs.busybox.net/show_bug.cgi?id=6584 (From OE-Core master rev: db00fa405c025c61844df2e9589fe635fc5df0e2) (From OE-Core rev: 9266343cc00fefd3e7f33fa2e6b0da88b6abf622) Signed-off-by: Jackie Huang <[email protected]> Signed-off-by: Saul Wold <[email protected]> Signed-off-by: Richard Purdie <[email protected]> Signed-off-by: Robert Yang <[email protected]> Signed-off-by: Richard Purdie <[email protected]>

Since kernel will build twice when we are trying to bundle kernel and initramfs together after commit 609d5a9ab("kernel.bbclass, image.bbclass: Implement kernel INITRAMFS dependency and bundling"), thus, the second building for kernel would fail if rm_work is done previously. To fix this problem, we need to make do_bundle_initramfs task run before do_rm_work task. [YOCTO #5416] (From OE-Core master rev: 8308e22a44a2dea7d1bbfb429b9df9c63714a649) (From OE-Core rev: fef443f1c40a3847cac08f4885d046acf6ede023) Signed-off-by: Martin Jansa <[email protected]> Signed-off-by: Xufeng Zhang <[email protected]> Signed-off-by: Saul Wold <[email protected]> Signed-off-by: Richard Purdie <[email protected]> Signed-off-by: Robert Yang <[email protected]> Signed-off-by: Richard Purdie <[email protected]>

do_configure() in binutils.inc includes an explicit call to gnu-configize so we need to make sure that gnu-config-native is present. Previously this was being dragged in with the rest of the autotools stuff, but commit 54a3e2ee37003fc56af0339f857b0b6442790c26 disabled that for binutils-cross on the grounds that "we don't autoreconf" the toolchain components. Fix this by adding gnu-config-native itself explicitly to DEPENDS. (From OE-Core master rev: 616354f13732d13c17434d5b60b166f691c25761) (From OE-Core rev: c7b50b95ab27d3ea3b37e3ee16050af962746f5c) Signed-off-by: Phil Blundell <[email protected]> Signed-off-by: Richard Purdie <[email protected]> Signed-off-by: Robert Yang <[email protected]> Signed-off-by: Richard Purdie <[email protected]>

This is a backport from upstream fixes a severe problem w.r.t memory management, where it would result in random segfaults in applications depending on libnl (From OE-Core rev: c3fb18aac0de49dc3113296699d95be298d98140) Signed-off-by: Khem Raj <[email protected]> Signed-off-by: Robert Yang <[email protected]> Signed-off-by: Richard Purdie <[email protected]>

Updating the 3.10 SRCREVs to update minnowboard support via the following changes: 3F6C824 pch_gbe: Add MinnowBoard support 9f52743 pch_gbe: Use PCH_GBE_PHY_REGS_LEN instead of 32 ec7b5e6 pch_gbe: use managed functions pcim_* and devm_* fd8bf50 pch_gbe: convert pr_* to netdev_* 9b278e9 serial: pch_uart: fix compilation warning 8982d79 serial: pch_uart: Fix signed-ness and casting of uartclk related fields cdbf456 serial: pch_uart: Remove __initdata annotation from dmi_table 9e7c25e pch_uart: Use DMI interface for board detection (From OE-Core master rev: 6c7115a56c3d0bf3d6d0275bd2d49d8cfef5c028) (From OE-Core rev: 436b39e4d9484dc74f88d635af97eee3d6d1704b) Signed-off-by: Darren Hart <[email protected]> Signed-off-by: Bruce Ashfield <[email protected]> Signed-off-by: Richard Purdie <[email protected]> Signed-off-by: Robert Yang <[email protected]> Signed-off-by: Richard Purdie <[email protected]>

Updating the SRCREVs to reflect the integration of the .17 -stable release and the preempt-rt up date to -rt11. (From OE-Core master rev: cefa022b814b8b4f9afacecf3bb035d211a0f3bf) (From OE-Core rev: 65491b70eeebcc579c8d9933be165fcd860bf2c7) Signed-off-by: Bruce Ashfield <[email protected]> Signed-off-by: Richard Purdie <[email protected]> Signed-off-by: Robert Yang <[email protected]> Signed-off-by: Richard Purdie <[email protected]>

Updating the 3.10 SRCREVs to add support for the haswell-sc and crystalforest boards. (From OE-Core master rev: 47ebe8677ac0dc4f8799d0af75f5b7bc611fd882) (From OE-Core rev: 06bef5014bac6e0d935453330dfec05153d71ad0) Signed-off-by: Bruce Ashfield <[email protected]> Signed-off-by: Richard Purdie <[email protected]> Signed-off-by: Robert Yang <[email protected]> Signed-off-by: Richard Purdie <[email protected]>

…tions Updating the meta SRCREV to include the latest crystalforest configuration updates. (From OE-Core master rev: 9480e5b7231a2923b5ebff9623827c5d90334df3) (From OE-Core rev: a7b2d0550157413834c9c5a3022745b25ad97fef) Signed-off-by: Bruce Ashfield <[email protected]> Signed-off-by: Richard Purdie <[email protected]> Signed-off-by: Robert Yang <[email protected]> Signed-off-by: Richard Purdie <[email protected]>

Updating the meta branch SRCREV to import some configuration updates for the common-pc-wifi feature: CONFIG_WEXT_CORE=y CONFIG_WEXT_PROC=y CONFIG_CFG80211_WEXT=y CONFIG_BCMA=m CONFIG_BCMA_HOST_PCI_POSSIBLE=y CONFIG_BCMA_HOST_PCI=y CONFIG_BCMA_DRIVER_GMAC_CMN=y CONFIG_CRC8=m CONFIG_CORDIC=m (From OE-Core master rev: cdd8145a7f4abc75c4089a30206c277db2712649) (From OE-Core rev: 3d3c7f004fd8c262f1b4e15e2552a5630f387457) Signed-off-by: Bruce Ashfield <[email protected]> Signed-off-by: Richard Purdie <[email protected]> Signed-off-by: Robert Yang <[email protected]> Signed-off-by: Richard Purdie <[email protected]>

This update fixes two issues: a) qemuarm boot failure v3.10.13 picked up a patch for arm versatile interrupt mappings that fixes the emulator boot out of the box. But it interacts badly with our previous fix for the issue. Reverting the existing patch and going with the mainline solution fixes the boot. b) qemumips build warning and failure Depending on the build host and compiler, the build of menuconfig throws an potentially uninitialized variable warning. That warning causes an error on archs with -Werror. We can do a trivial change to avoid the warning all together (initilize it to null), and keep everyone happy. [YOCTO #5460] (From OE-Core master rev: 8d1a041891c87d0c2003c80f84b0501bdc9403a1) (From OE-Core rev: 3928340ea03dc04cda9eb2eea021837421adf737) Signed-off-by: Bruce Ashfield <[email protected]> Signed-off-by: Richard Purdie <[email protected]> Signed-off-by: Robert Yang <[email protected]> Signed-off-by: Richard Purdie <[email protected]>

Updating the meta branch SRCREV to update the USB configuration: The Beagleboard needs the USB PHY drivers in the kernel in order to enable USB and Ethernet functionality. This fix ensures that they are built in by tweaking the kernel config. Tested on Beagleboard xM Rev. C2. (From OE-Core master rev: 2a9944514362445ee891f6e77c4ae62950e247b3) (From OE-Core rev: 670cd17b4ef4012a29eb4dee3c9899ccd5f967e6) Signed-off-by: Bruce Ashfield <[email protected]> Signed-off-by: Richard Purdie <[email protected]> Signed-off-by: Robert Yang <[email protected]> Signed-off-by: Richard Purdie <[email protected]>

Updating the -rt SRCREVs to pick up the following fix: ntp: fix ntp_notify_cmos_timer merge issue PREEMPT_RT_FULL has a stubbed ntp_notify_cmos_timer due to a bad merge. Renaming and restoring the full -rt functionality to this routine. (From OE-Core master rev: 41d4f0feca69bf1b41f16f5f7d21bf7540e6c47a) (From OE-Core rev: 0336b83df666617d61e8cd39c2d2a7f9c745064c) Signed-off-by: Bruce Ashfield <[email protected]> Signed-off-by: Richard Purdie <[email protected]> Signed-off-by: Robert Yang <[email protected]> Signed-off-by: Richard Purdie <[email protected]>

Updating the meta branch SRCREV to update the USB configuration: The Beagleboard needs the USB PHY drivers in the kernel in order to enable USB and Ethernet functionality. This fix ensures that they are built in by tweaking the kernel config. Tested on Beagleboard xM Rev. C2. (From OE-Core master rev: 89a372840a957e540bed954e629aa68335b3dfe0) (From OE-Core rev: ee9dd4a40916668927b0c4b4dfaa4f80c671c9b2) Signed-off-by: Bruce Ashfield <[email protected]> Signed-off-by: Saul Wold <[email protected]> Signed-off-by: Richard Purdie <[email protected]> Signed-off-by: Robert Yang <[email protected]> Signed-off-by: Richard Purdie <[email protected]>

This updates lttng-modules for 2.3.3 and it also fixes the build with 3.12 Linux kernel. While on that, we also renamed the recipe file to follow the other lttng recipes which use the version number on it. (From OE-Core master rev: 2d01bd48e689656bbe6189243d077f822092a14a) (From OE-Core rev: 213ba50b8a747e1489f03872339e3931a99963a4) Signed-off-by: Otavio Salvador <[email protected]> Signed-off-by: Saul Wold <[email protected]> Signed-off-by: Richard Purdie <[email protected]> Signed-off-by: Robert Yang <[email protected]> Signed-off-by: Richard Purdie <[email protected]>

There is an race condition where psplash is not quite exited before the unmount occurs causing a umount: /mnt/.psplash: target is busy message to appear, it's ok to lazyily unmount and not get this message [YOCTO #5244] (From OE-Core master rev: 9ded366084f22f48ef72aa22acf6a38982d16d97) (From OE-Core rev: 8c3e3c90daee1639ac8b2633d8f1e500697d9c52) Signed-off-by: Saul Wold <[email protected]> Signed-off-by: Richard Purdie <[email protected]> Signed-off-by: Robert Yang <[email protected]> Signed-off-by: Richard Purdie <[email protected]>

…pool::get_lock_for() Boost::thread uses functions from boost::atomic but doesn't actually link with libboost_atomic. This works fine on platforms where BOOST_ATOMIC_FLAG_LOCK_FREE is true but will lead to undefined symbol references otherwise. Fix this by applying a patch from the upstream bug tracker to add the missing library linkage. (From OE-Core master rev: 1ffc27173576589191b037d111ecb59d94631de0) (From OE-Core rev: 306fd06c7687340a0d4eb90e9aba9dc8669db07c) Signed-off-by: Phil Blundell <[email protected]> Signed-off-by: Saul Wold <[email protected]> Signed-off-by: Richard Purdie <[email protected]> Signed-off-by: Robert Yang <[email protected]> Signed-off-by: Richard Purdie <[email protected]>

When ptest is not enabled, the populate_packages_prepend function runs wheter ptest is enabled or not. This causes ptest packages to get in the dependencies list when ptest is not enabled. (From OE-Core master rev: 826f4e4057a221127ac4c1d0658d975032fc7d90) (From OE-Core rev: e739a64143901fa6f6f54e70445d19e9ce13dcf1) Signed-off-by: Jeffrey C Honig <[email protected]> Signed-off-by: Richard Purdie <[email protected]> Signed-off-by: Robert Yang <[email protected]> Signed-off-by: Richard Purdie <[email protected]>

OE-Core commit b7de1eaac9eed559b2d68058f5de67de74a6cb58 added an extra argument to the compare_dict_blobs() function but missed adding the argument to one call to compare two versions of the image-info.txt file. (From OE-Core master rev: 24a45d752c3e3d0d8b59c040355e4fe7de22b041) (From OE-Core rev: a51d96c44e6feac8322284c54bfc01ef598f8821) Signed-off-by: Paul Eggleton <[email protected]> Signed-off-by: Richard Purdie <[email protected]> Signed-off-by: Robert Yang <[email protected]> Signed-off-by: Richard Purdie <[email protected]>

*PPC64 uses long long for u64 in the kernel, but powerpc's asm/types.h prevents 64-bit userland from seeing this definition, instead defaulting to u64 == long in userspace. *fix the below error |super-ddf.c:4542:5: error: format '%llu' expects argument of type 'long long unsigned int', |but argument 5 has type '__u64' [-Werror=format=] |dprintf("BVD %u has %08x at %llu\n", 0, (From OE-Core master rev: d3caab6eb03264b4f4d744f914598022299011ba) (From OE-Core rev: c5e59e68efcf2a3f902dbfd827da57ed3e8ad4ce) Signed-off-by: Chunrong Guo <[email protected]> Signed-off-by: Saul Wold <[email protected]> Signed-off-by: Richard Purdie <[email protected]> Signed-off-by: Robert Yang <[email protected]> Signed-off-by: Richard Purdie <[email protected]>

Flex needs m4 to run (see below) and, since the create_wrapper introduces a bash dependency on target, give the path to m4 binary in the configure command line. Snippet from the flex documentation: "The macro processor m4 must be installed wherever flex is installed. <...> m4 is only required at the time you run flex." [YOCTO #5329] (From OE-Core master rev: 64030f37b34f75144f53eef42d5822ede79e08bd) (From OE-Core rev: d039150eae579af1bd85000982ef38a6b09bb90d) Signed-off-by: Laurentiu Palcu <[email protected]> Signed-off-by: Saul Wold <[email protected]> Signed-off-by: Richard Purdie <[email protected]> Signed-off-by: Robert Yang <[email protected]> Signed-off-by: Richard Purdie <[email protected]>

grub_2.0.0 requires xz to run or an error may occur. (From OE-Core master rev: ffa2877c06c587d4ea56c55bfd0f67a88e42a772) (From OE-Core rev: d3e1dde9ae75faa1015b80da74604c5ae360c3a6) Signed-off-by: Ming Liu <[email protected]> Signed-off-by: Saul Wold <[email protected]> Signed-off-by: Richard Purdie <[email protected]> Signed-off-by: Robert Yang <[email protected]> Signed-off-by: Richard Purdie <[email protected]>

We don't support multiple -dbg/-dev packages, the package can generate them but the system does not correctly handle them. Just move all devel stuffs into 'udev-dev' and all debug stuffs into 'udev-dbg'. (From OE-Core master rev: 014f7a33f399192268f28acac835551413c4768d) (From OE-Core rev: a700493063efdc3f60fb77e9a4a15d9c89ee79c6) Signed-off-by: Wenzong Fan <[email protected]> Signed-off-by: Saul Wold <[email protected]> Signed-off-by: Richard Purdie <[email protected]> Signed-off-by: Robert Yang <[email protected]> Signed-off-by: Richard Purdie <[email protected]>

In commit faa8cc6c2a582a32c695f3f2b0d45b6892c769fd dmesg.sh was added to the set of init.d scripts. But the script was never put in any run-level. This patch will add dmesg.sh to run-level S. (From OE-Core master rev: 7d2767d4e27c6d0eaa56f3e126df56e65a5364c9) (From OE-Core rev: fb7a6e0e3c5790415d56d14ec1a5eda5f9e6d039) Signed-off-by: Hans Beckerus <hans.beckerus AT gmail.com> Signed-off-by: Saul Wold <[email protected]> Signed-off-by: Richard Purdie <[email protected]> Signed-off-by: Robert Yang <[email protected]> Signed-off-by: Richard Purdie <[email protected]>

${KERNEL_OUTPUT} is being renamed/restored in bundle_initramfs task, so we must ensure bundle_initramfs run after kernel_link_vmlinux where the link of vmlinux is created as the bootable image. (From OE-Core master rev: 3baa63b4d588c3262254528b406ede265dd117bf) (From OE-Core rev: 4acf2eaea963d9b5e3cf547db092f95d192cf2ab) Signed-off-by: Ming Liu <[email protected]> Signed-off-by: Saul Wold <[email protected]> Signed-off-by: Richard Purdie <[email protected]> Signed-off-by: Robert Yang <[email protected]> Signed-off-by: Richard Purdie <[email protected]>

(From OE-Core master rev: 836396a3450e7bf151956e87bd92f70c5050c995) (From OE-Core rev: 1db74ff0544633c35fe0fdccf17a76ef833be0df) Signed-off-by: Saul Wold <[email protected]> Signed-off-by: Richard Purdie <[email protected]> Signed-off-by: Robert Yang <[email protected]> Signed-off-by: Richard Purdie <[email protected]>

The license_create_manifest function contains bashism, this will lead to unexpected results on ubuntu build host, as sh is linked to dash on ubuntu. Even if COPY_LIC_MANIFEST and COPY_LIC_DIRS are enabled, the license files will still be missing on target. This patch fixes the above problem. [YOCTO #5549] (From OE-Core master rev: 4df9daee5c732c0a20dabe8515577238a1508512) (From OE-Core rev: 159e53f9402f1d1ceed8c6511c5874e199dea6e1) Signed-off-by: Chen Qi <[email protected]> Signed-off-by: Saul Wold <[email protected]> Signed-off-by: Richard Purdie <[email protected]> Signed-off-by: Robert Yang <[email protected]> Signed-off-by: Richard Purdie <[email protected]>

Busybox also provides sulogin command, so we need to use the ALTERNATIVE mechanism to manage it. (From OE-Core master rev: 8b3a799a87d18b1d113d59b3e7a681db5683e5f8) (From OE-Core rev: 33d7487a92905824ab46f74d7185f84662ddb922) Signed-off-by: Chen Qi <[email protected]> Signed-off-by: Saul Wold <[email protected]> Signed-off-by: Richard Purdie <[email protected]> Signed-off-by: Robert Yang <[email protected]> Signed-off-by: Richard Purdie <[email protected]>

buildtest-TESTS is a phony target and does nothing which results in a do_install error since the tests aren't built. Since there isn't a suitable make target but the number of tests are small, hardcode the two to build to unbreak the build when ptest is enabled. (From OE-Core master rev: 5dd8653fdcda5e0e8b4f3c37a46f357bc97ec66c) (From OE-Core rev: b9b213b43c5ff6aa7c04733ce035fc9832065328) Signed-off-by: Richard Purdie <[email protected]> Signed-off-by: Robert Yang <[email protected]> Signed-off-by: Richard Purdie <[email protected]>

…rtual_pkg PREFERRED_PROVIDER_virtual_pkg has been incorrectly set with more than one multilib prefixes. For example, if we have two alternative multilibs lib64 and lib32, PREFERRED_PROVIDER_virtual_pkg will be set to lib32-lib64-pkg or lib64-lib32-pkg, depending on which multilib shows up first in the list. (From OE-Core master rev: 17a432dc059e24ba10d4baec988828c0025a5e46) (From OE-Core rev: e5d8411869a2a018d0c8ab9d0e888027ac4208d5) Signed-off-by: Lei Liu <[email protected]> Signed-off-by: Richard Purdie <[email protected]> Signed-off-by: Robert Yang <[email protected]> Signed-off-by: Richard Purdie <[email protected]>

The cdg_decode_frame function in cdgraphics.c in libavcodec in FFmpeg before 1.2.1 does not validate the presence of non-header data in a buffer, which allows remote attackers to cause a denial of service (out-of-bounds array access and application crash) via crafted CD Graphics Video data. http://web.nvd.nist.gov/view/vuln/detail?vulnId=CVE-2013-3674 (From OE-Core master rev: f1721553a873b242bc26ad3e4d618aea39dfd507) (From OE-Core rev: d4d908afca1c6ce64f05fabdadfcf86ab990dd0f) Signed-off-by: Ming Liu <[email protected]> Signed-off-by: Jeff Polk <[email protected]> Signed-off-by: Richard Purdie <[email protected]> Signed-off-by: Robert Yang <[email protected]> Signed-off-by: Richard Purdie <[email protected]>

…orm file On a multilib system when one of the multibs has a different OS then other multilibs a failure can occur during the install process because RPM assumes all systems have the same OS. When an n32 platform is selected as an alternative multilib, it shows up as mips64_n32-.*-linux-gnun32 in /etc/rpm/platform. This causes problems when the smart tool tries to add a channel for the multilib. RPM archScore call always returns zero for arch "mips64_n32" - after appending default vendor and os, it finds "mips64_n32-wrs-linux" doesn't match any predefined platforms. Fix this by removing the restriction of -gnun32 suffix in platform file. (From OE-Core master rev: d9489c44ee4f195ae1b09f340b9545cddba58145) (From OE-Core rev: f0118b605b3727b5ca5d560094bb4dd2ff29c310) Signed-off-by: Lei Liu <[email protected]> Signed-off-by: Jeff Polk <[email protected]> Signed-off-by: Richard Purdie <[email protected]> Signed-off-by: Robert Yang <[email protected]> Signed-off-by: Richard Purdie <[email protected]>

When preping a read-only rootfs and finding some post-install scripts that can not be run, list the names of said scripts to avoid having to look around the rootfs to find a list. (From OE-Core master rev: 0188120691f433fdccf71b92618115195278c0af) (From OE-Core rev: 2820f7fa411e5ca1cd7df765896b43716418340a) Signed-off-by: Jeffrey C Honig <[email protected]> Signed-off-by: Jeff Polk <[email protected]> Signed-off-by: Richard Purdie <[email protected]> Signed-off-by: Robert Yang <[email protected]> Signed-off-by: Richard Purdie <[email protected]>

readprofile was missing from the alternative configuration, which was causing readprofile to be packaged into the base util-linux. (From OE-Core master rev: cac08f23aaed87148d1825cca3c7586ab891ef04) (From OE-Core rev: a6fc9e62e848634e715b23f147c88a8710415845) Signed-off-by: Mark Hatle <[email protected]> Signed-off-by: Richard Purdie <[email protected]> Signed-off-by: Robert Yang <[email protected]> Signed-off-by: Richard Purdie <[email protected]>

syslinux is compling something with host gcc at do_install stage, which leads to some unexpected errors with old gcc on host. Using our cross toolchain instead. (From OE-Core master rev: b0da7ccde5380726acfccf1a96cdf5560edf9159) (From OE-Core rev: 17f851f5c8e2a90f558396654b62a8f0f88f137f) Signed-off-by: Lei Liu <[email protected]> Signed-off-by: Randy MacLeod <[email protected]> Signed-off-by: Richard Purdie <[email protected]> Signed-off-by: Robert Yang <[email protected]> Signed-off-by: Richard Purdie <[email protected]>

[YOCTO #2519] When getting gcc from sstate, it is possible to get a gcc with a bogus sysroot configuration, as discussed in [1] or in [YOCTO #2519]. mklibs script will eventually call gcc, so we need to make sure that it provides gcc with the right sysroot location. [1] http://lists.openembedded.org/pipermail/openembedded-core/2013-September/084159.html (From OE-Core master rev: 3a66dd762e493ad2cda57110be67c3b06628050a) (From OE-Core rev: 7275425524b8bb3d16d5c0c0a62aee5b08359ffd) Signed-off-by: Nicolas Dechesne <[email protected]> Signed-off-by: Saul Wold <[email protected]> Signed-off-by: Richard Purdie <[email protected]> Signed-off-by: Robert Yang <[email protected]> Signed-off-by: Richard Purdie <[email protected]>

ICU presence is auto-detected at configure time and until recently (e68850 and d61230) was pulled into most builds through harfbuzz and beecrypt. Now it's floating and this leads to build failures. As in all likelihood the majority of people were building this with ICU enabled, add an explicit dependency. (From OE-Core master rev: 46dcec6fd455584d9b5d0d7ff1e5b36fbe5a2d62) (From OE-Core rev: 034d3e35cce9ee63f6139d19be9b3edec4f97a64) Signed-off-by: Ross Burton <[email protected]> Signed-off-by: Richard Purdie <[email protected]> Signed-off-by: Robert Yang <[email protected]> Signed-off-by: Richard Purdie <[email protected]>

Heap-based buffer overflow in the tp_process_jpeg_strip function in tiff2pdf in libtiff 4.0.3 and earlier allows remote attackers to cause a denial of service (crash) and possibly execute arbitrary code via a crafted TIFF image file. http://web.nvd.nist.gov/view/vuln/detail?vulnId=CVE-2013-1960 (From OE-Core rev: 66387677cbd85ba4a76a254942377621acd68249) Signed-off-by: Ming Liu <[email protected]> Signed-off-by: Jeff Polk <[email protected]> Signed-off-by: Robert Yang <[email protected]> Signed-off-by: Richard Purdie <[email protected]>

Scripts in nfs-utils need bash as their interpreter, so if nfs-utils doesn't explicitly rdepend on bash, we would experience build failures if we add nfs-utils to glibc-small images. Add bash to RDEPENDS to solve this problem. (From OE-Core rev: 06c566596a92a309ca228a209f14d03b69a611c9) Signed-off-by: Chen Qi <[email protected]> Signed-off-by: Robert Yang <[email protected]> Signed-off-by: Richard Purdie <[email protected]>

cve description: Use-after-free vulnerability in the t2p_readwrite_pdf_image function in tools/tiff2pdf.c in libtiff 4.0.3 allows remote attackers to cause a denial of service (crash) or possible execute arbitrary code via a crafted TIFF image. http://web.nvd.nist.gov/view/vuln/detail?vulnId=CVE-2013-4232 (From OE-Core rev: 60482e45677c467f55950ce0f825d6cb9c121c9c) Signed-off-by: Baogen Shang <[email protected]> Signed-off-by: Mark Hatle <[email protected]> Signed-off-by: Robert Yang <[email protected]> Signed-off-by: Richard Purdie <[email protected]>

cve description: Heap-based buffer overflow in the readgifimage function in the gif2tiff tool in libtiff 4.0.3 and earlier allows remote attackers to cause a denial of service (crash) and possibly execute arbitrary code via a crafted height and width values in a GIF image. http://web.nvd.nist.gov/view/vuln/detail?vulnId=CVE-2013-4243 (From OE-Core rev: a2a200a3951cecd7dd43dee360e0260051c97416) Signed-off-by: Baogen Shang <[email protected]> Signed-off-by: Mark Hatle <[email protected]> Signed-off-by: Robert Yang <[email protected]> Signed-off-by: Richard Purdie <[email protected]>

Commit 46dcec6fd455584d9b5d0d7ff1e5b36fbe5a2d62 added 'icu' to DEPENDS in qt4-x11 only, but enabled icu globally in qt4.inc. This breaks build of qt4-embedded because this recipe does not have such a DEPENDS but uses qt4.inc: | icu.cpp:42:28: fatal error: unicode/utypes.h: No such file or directory | #include <unicode/utypes.h> | ^ | compilation terminated. | make: *** [icu.o] Error 1 Patch moves the 'icu' dependency into qt4.inc. (From OE-Core rev: adb6e64d69fc947f2c8fa708dcbe854fd2b574f8) (From OE-Core rev: ec35d5b4b3d2eed7a141f2fd41d5ed7215e66dbf) Signed-off-by: Enrico Scholz <[email protected]> Cc: Ross Burton <[email protected]> Signed-off-by: Richard Purdie <[email protected]>

(From OE-Core rev: e638af73a01496e246f1a55d27364861a1f0b0d0) Signed-off-by: Richard Purdie <[email protected]>

This is a fix which avoids false positives if the search pattern "lose" is found in path descriptions in comments generated by the preprocessor we hit in our development environment. [gdb Bug #16152] -- https://sourceware.org/bugzilla/show_bug.cgi?id=16152 Upstream-Status: Accepted (From OE-Core rev: 7e2dbda690b480ab05d14353cb038749ce23d58c) Signed-off-by: Steffen Sledz <[email protected]> Signed-off-by: Richard Purdie <[email protected]>

(From meta-yocto master rev: 8bd33820b4d1944a9f7730f8e2676d0d45e1cd0b) (From meta-yocto rev: 3b2b8cce34043766e31ba501052f8d1668f4a2b3) Signed-off-by: Saul Wold <[email protected]> Signed-off-by: Richard Purdie <[email protected]> Signed-off-by: Robert Yang <[email protected]> Signed-off-by: Richard Purdie <[email protected]>

Updating the BSP SRCREVs to pull in the 3.10.17 core update and fix USB powerup issues on the beagleboard. (From meta-yocto master rev: d82870a9561662919a737dd126a8d26e2b78144a) (From meta-yocto rev: 17403f07a5ec54f867515dc8cb8bd65fd232c6f5) Signed-off-by: Bruce Ashfield <[email protected]> Signed-off-by: Richard Purdie <[email protected]> Signed-off-by: Robert Yang <[email protected]> Signed-off-by: Richard Purdie <[email protected]>

If the build environment is misconfigured (e.g. a bad path for a layer in bblayers.conf) the yocto-bsp script crashes with a standard python error, not very explicit. This fixes the problem. Signed-off-by: Bastien JAUNY <[email protected]> (From meta-yocto master rev: 4a8e80b812eebdc1c9570b5d88aa0f3b34824b68) (From meta-yocto rev: 578e06f113d870ec6a4e201458488344ca941e3d) Signed-off-by: Richard Purdie <[email protected]> Signed-off-by: Robert Yang <[email protected]> Signed-off-by: Richard Purdie <[email protected]>

Without this option wifi support in connman will fail: src/technology.c:technology_get() No matching drivers found for wifi (From OE-Core rev: 403e365e433c54633bcc843b32487a766282226e) (From OE-Core rev: 2e532f33c5e97751daa89c9f92c6de8513564be0) Signed-off-by: Steve Sakoman <[email protected]> Signed-off-by: Richard Purdie <[email protected]>

(From OE-Core rev: d68c267f3387d7fe221d3c5653a66db8b1f78fd8) Signed-off-by: Richard Purdie <[email protected]>

Used a better word to describe the argument list. (From yocto-docs rev: 15f14a3a36d345c655e60bc7a4b7d19c02d26f2c) Signed-off-by: Scott Rifenbark <[email protected]> Signed-off-by: Richard Purdie <[email protected]>

Review comments from Paul. (From yocto-docs rev: effc8e811020e00bfd98d065e412db5fe3f78f04) Signed-off-by: Scott Rifenbark <[email protected]> Signed-off-by: Richard Purdie <[email protected]>

Review edits from Paul. (From yocto-docs rev: 8089f69979f872b1c756fb1e1703fa0ea6965426) Signed-off-by: Scott Rifenbark <[email protected]> Signed-off-by: Richard Purdie <[email protected]>

Review comments from Ross. (From yocto-docs rev: 88ce4d4e88671a968d3fee84dd3b8e1b64e84282) Signed-off-by: Scott Rifenbark <[email protected]> Signed-off-by: Richard Purdie <[email protected]>

(From yocto-docs rev: a88e4a770b1fe536285269055ba0655c702f0d70) Signed-off-by: Scott Rifenbark <[email protected]> Signed-off-by: Richard Purdie <[email protected]>

The directory is not a temporary thing. (From yocto-docs rev: d40d17ed80ebdb738bca9c86cd1381cd1442e5b8) Signed-off-by: Scott Rifenbark <[email protected]> Signed-off-by: Richard Purdie <[email protected]>

(From yocto-docs rev: be1e564483299a018e28f1971dbe85f8485c9b83) Signed-off-by: Scott Rifenbark <[email protected]> Signed-off-by: Richard Purdie <[email protected]>

Signed-off-by: Richard Purdie <[email protected]>

(From yocto-docs rev: 51afdedc5c9bb6b689e7cf8771e0889d445f5326) Signed-off-by: Scott Rifenbark <[email protected]> Signed-off-by: Richard Purdie <[email protected]>

DISTRO needs to be flipped for pending point release (From meta-yocto rev: efb1dd56721320f767eb3066567f8caeb32580a2) Signed-off-by: Beth Flanagan <[email protected]> Signed-off-by: Richard Purdie <[email protected]>

(From OE-Core rev: 2e9df12e67b3e56ed3c056559aa8eced6444ec93) Signed-off-by: Richard Purdie <[email protected]>

tagging for 1.5.1 Signed-off-by: Matt Fleming <[email protected]>

'git branch' may use ANSI escape codes in its output (to provide colour) which doesn't play well with commands expecting pure plain text, e.g. fatal: '^[[31mmaster^[[m' is not a valid branch name. Use the --no-color option to ensure all branch names are plain text. Cc: Bruce Ashfield <[email protected]> (From OE-Core rev: 87acfdb28380c26344a79a9dc0b4b403173bdc44) Signed-off-by: Matt Fleming <[email protected]> Signed-off-by: Richard Purdie <[email protected]>

OVMF is required to build a UEFI-capable Qemu. Signed-off-by: Matt Fleming <[email protected]>

New recipe for building kexec from upstrem because support for using kexec with UEFI hasn't made it into a release of kexec-tools yet. Signed-off-by: Matt Fleming <[email protected]>

This kernel is built from Matt Fleming's upstream EFI git repository and includes all of the latest development EFI work. Signed-off-by: Matt Fleming <[email protected]>

Put the core pieces in place for generating a file system image for use on UEFI platforms. Signed-off-by: Matt Fleming <[email protected]>

This patch removes several patches from the recipe as all these patches are already merged into the kexec repository. Signed-off-by: Ricardo Neri <[email protected]> Acked-by: Matt Fleming <[email protected]>

Instead of using the 'master' branch of the upstream kernel EFI repository, use the 'stable' branch which is based on the latest kernel.org release plus any EFI patches that have been sent to Linus. Signed-off-by: Matt Fleming <[email protected]>

The recipes in our layer should have precedence over everything else. Signed-off-by: Matt Fleming <[email protected]>

Instead of the hardcoded path to my local copy. Signed-off-by: Matt Fleming <[email protected]>

OVMF is run on the development host and so is "native". Signed-off-by: Matt Fleming <[email protected]>

We need to list ovmf-native as a dependency of qemu-native so that ovmf is built and installed automatically. Signed-off-by: Matt Fleming <[email protected]>

The OVMF code fails to build with -O2. Signed-off-by: Matt Fleming <[email protected]>

Instead of relying on the developent environment outside of poky, we should be using the libraries, headers, etc built with bitbake. OVMF doesn't lend itself particularly well to building with cross-toolchains of any kind so we need to carry two patches that allow us to override CFLAGS and LFLAGS until they get merged upstream. Signed-off-by: Matt Fleming <[email protected]>

The path depends on the toolchain and architecture we're building for, so use find(1) to locate OVMF.fd during do_install(). Signed-off-by: Matt Fleming <[email protected]>

To correctly give precedence to our distrirbution layer, prepend our layer directory to BBPATH. Signed-off-by: Ricardo Neri <[email protected]>

Utilize qemu version 1.5.0 as this is the version used in Yocto release 10.0.1 (dora). We can move to a more recent qemu in subsequent versions of Yocto. Signed-off-by: Ricardo Neri <[email protected]>

We define this variable to have a cleaner build directory structure. Additionally, we can do this as we do not have a reason to stick to egclibc. Signed-off-by: Ricardo Neri <[email protected]>

Set LAYER_CONF_VERSION to the latest version. This is to be aligned with other meta layers as well as the expectation of bblayers.conf Signed-off-by: Ricardo Neri <[email protected]>

Add support to specify a directory for custom BIOS, VGA BIOS and keymaps as supported by qemu (-L option). Even though this can be done through qemuparams, having this option provides better user experience by not having to specify a long and cluttered path along with other qemuparams that the user might want to specify. This new options assumes first that the path provided is relative to OECORE_NATIVE_SYSROOT and will check whether it exists before proceeding. If not, it will treat the provided path as absolute. This provides the user flexibility to use BIOS binaries generated inside or outside the OE build environment. Signed-off-by: Ricardo Neri <[email protected]> Signed-off-by: Matt Fleming <[email protected]>

We need a way of automatically running test programs when the machine boots. Install a luv-test-manager script and invoke it from runlevel 5. luv-test-manager is a test manager, it is the file that drives the entire test process. It searches for test runner scripts in /etc/luv-tests/ and executes any that it finds. Currently it only handles execution of tests but later on it will also collect their output and parse it into a standard format. Signed-off-by: Matt Fleming <[email protected]>

Take control of the sysvinit files. We don't need bootlogd, so use a custom do_install() function that doesn't install it. The test manager in /etc/init.d/luv-test-manager will be solely responsible for collecting and parsing all log output, including the boot logs, and it makes sense to only do this in one place. Signed-off-by: Matt Fleming <[email protected]>

This commit introduces a new class that automatically generates and installs a test runner script for testsuite packages. Test runners are intended to be executed by a master test process, or test manager. Most packages will not need to override any of the variables. The default behaviour is to create a test runner that simply runs an executable using the package name as the name of that executable. Packages can customize the arguments that are passed to the package program by setting the LUV_TEST_ARGS variable. Signed-off-by: Matt Fleming <[email protected]>

We want fwts to be run automatically by the test manager. Inherit luv-test so that a test runner script is automatically generated for this package. In the future, we're going to want to override the LUV_TEST_ARGS variable so that we can customize exactly what parameters are passed to fwts. In particular it would be useful to have control of the format of the test result output. Signed-off-by: Matt Fleming <[email protected]>

There's really no good reason to do our own kernel config mangling in kernel_do_configure() when the linux-yocto recipes provide a perfectly good method. The correct way to define a default kernel config with the yocto kernel recipes is via a 'defconfig' file. Delete efi.cfg (which was totally unused anyway) along with the kernel_do_configure() function and replace them with a defconfig generated against the upstream EFI repository stable branch (v3.13). Also take this opportunity to enable some key EFI features, - CONFIG_EFI_STUB - CONFIG_EFIVAR_FS - CONFIG_EARLY_PRINTK_EFI Signed-off-by: Matt Fleming <[email protected]>

Most test programs will need custom output parsers that know how to parse the test results and write them to stdout in a format suitable for luv-test-manager to consume. This has necessitated moving /etc/luv-tests to /etc/luv/tests and creating a new /etc/luv/parsers directory to contain testsuite parsers. The schema expected by the test manager is detailed in luv-test-parser. Any major additions to the schema should bump the expected verison. Remember, the log data may not be generated and consumed on the same machine. For example it's possible to save the log data over the network and process the logs offline to graph result trends over time. Any parsers written for test programs should refrain from using either the version of sed or grep built with busybox. The reason being that they have strange buffering properties on output which make the tests appear to "hang" for long periods of time. The parsers were written in awk because it does not suffer from this output buffering problem, allowing the creation of an arbitrary pipeline without noticable delay to the user. Also, awk provide a C-like syntax that is pretty straight forward to understand even for people without much familiarity with it. While perl would have also been a suitable choice it isn't currently installed in the core-image-efi image. However, people should feel free to use whatever tool they like. Signed-off-by: Matt Fleming <[email protected]>

Add a fwts-specific log parser the emits test info in the luv-test-manager format. Also restrict the test category to just the uefi ones for now. Signed-off-by: Matt Fleming <[email protected]>

There has been various bug fixes in the upstream repository since V14.01.00, so upgrade to something more recent. In particular, before enabling the uefi tests in our recipe we need the following change otherwise the efi_runtime module isn't loaded correctly every time, : commit a58df34c0a1b : Author: Colin Ian King <[email protected]> : Date: Wed Jan 22 15:54:49 2014 +0000 : : fwts_exec: fix status check (LP: #1271607) : : fix inverted status check of exec calls. This is causing : uefi tests to fail because of module load/unload is : erroneously being reported to fail when it is in fact : succeeding. : : Signed-off-by: Colin Ian King <[email protected]> : Acked-by: Alex Hung <[email protected]> : Acked-by: Ivan Hu <[email protected]> Upgrading to a newer version also allows us to drop both jsonc.patch and subdir.patch which have since been merged upstream. Signed-off-by: Matt Fleming <[email protected]>

There are a few kernel debug options that are not enabled in our kernel config such as lockdep support, and RCU stall warning. Turn them on so that we can catch these kinds of bugs at runtime. There is a performance penalty to pay for having these features turned on at runtime, but it's worth it because we favour correctness and ease of debugging over performance. If our priorities change in future we can always create a separate debug version of the kernel. Signed-off-by: Matt Fleming <[email protected]>

The UEFI runtime tests require the efi_runtime kernel module. So build and install it as part of our fwts recipe. While it would have been possible to create a new, separate recipe solely for the kernel module a unified one makes more sense because all the code is kept in a single upstream repository. Signed-off-by: Matt Fleming <[email protected]>

These patches have been postd to the upstream fwts mailing list for review. Meanwhile, let's add them so we can run the fwts UEFI tests. With these patches all the UEFI tests pass when booting core-image-efi with OVMF. Signed-off-by: Matt Fleming <[email protected]>

Utilize a portion of the system's RAM memory as a block device. This is needed in order to load our initial filesystem. Use the default number and size of RAM disks. Signed-off-by: Ricardo Neri <[email protected]> Signed-off-by: Matt Fleming <[email protected]>

It is not expected to have an excessively large filesystem for the EFI test suite. Thus, it makes sense to utilize an initramfs that contains all the tests. Also, this reduce the complexity of the hardware requirements needed to run the suite on a given system: we will need only a chunk of RAM. Signed-off-by: Ricardo Neri <[email protected]> Signed-off-by: Matt Fleming <[email protected]>

When booting from initramfs, the kernel expects a /init file. Install and ship a symbolic link to /sbin/init, which is the init file that is effectively used. Signed-off-by: Ricardo Neri <[email protected]> Signed-off-by: Matt Fleming <[email protected]>

When booting a ramfs image, runqemu-internal disables the qemu graphic output but does not provide an alternative serial output. Thus, the the graphic output is re-enabled. Signed-off-by: Ricardo Neri <[email protected]> Signed-off-by: Matt Fleming <[email protected]>

It's possible to hit the following build error when building fwts after a clean checkout, /bin/sh: scripts/basic/fixdep: No such file or directory Follow the precedence of other recipes (systemtap) and explicitly build the 'scripts' make target. Signed-off-by: Matt Fleming <[email protected]>

Ditch busybox in favour of more robust utilities. Signed-off-by: Matt Fleming <[email protected]>

Avoid confusion regarding how to build the luv distribution by changing the name of the distro from "efitest" to "luv". Signed-off-by: Matt Fleming <[email protected]>

Include more details on how to build the luv distribution and where to send patches. The mailing list address will be updated once it's functional. Signed-off-by: Matt Fleming <[email protected]>

It's a risky game using "${AUTOREV}" for SRCREV because the version of the upstream project you're building is a moving target, and there's a high risk that the build will break when someone modifies the same file upstream. By using an explicit version of upstream ovmf we know our patches will only fail to apply if we move to a newer version - a process that we control. The actual sha1 selected isn't of any particular significance, it just happens to be the HEAD commit from today. Also, add util-linux as a dependency because ovmf requires uuid/uuid.h. Signed-off-by: Matt Fleming <[email protected]>

... and point readers at meta-luv/README for more information. Signed-off-by: Matt Fleming <[email protected]>

This recipe builds the efivarfs selftests. These tests check that files can be created, deleted and modified in a efivarfs mount and that it is not possible to create files with bogus filenames, etc. Signed-off-by: Rakesh Seethamsetty <[email protected]> [ Expanded on commit log ] Signed-off-by: Matt Fleming <[email protected]>

Signed-off-by: Rakesh Seethamsetty <[email protected]> Signed-off-by: Matt Fleming <[email protected]>

If we don't set ${CROSS_COMPILE} when building the efivarfs tests then it will default to using the host's compiler, not the compiler for the target. This results in the tests being unable to run on the target. Use ${TARGET_PREFIX} to set ${CROSS_COMPILE} so that the efivarfs Makefile will pickup the correct compiler. Also, instead of building all the selftests limit the build to just the efivarfs/ directory. Signed-off-by: Matt Fleming <[email protected]>

We need to provide our own copy of fstab which mounts efivarfs at /sys/firmware/efi/efivars as the efivarfs tests require it. Signed-off-by: Matt Fleming <[email protected]>

Inherit luv-test.bbclass which will automatically run the efivarfs tests when booting. The efivarfs selftests are not easily executed out of the box and it was necessary to create a /usr/bin/efivarfs script, which is basically just a wrapper that invokes the tests from within /usr/share/efivarfs-test. Signed-off-by: Matt Fleming <[email protected]>

Having a hard drive image to boot with qemu is fine and dandy but we really need a bootable image that we can write to removable media and run on real hardware, not least because that's the thing people will expect to find on the project website. bootimg.bbclass provides much of the functionality we need and already contains code to build a FAT file system image from our rootfs. The customization occurs in do_create_img() where we construct a partition table and append the rootfs FAT file system. While scripts already exist to build this type of image (scripts/contrib/mkefidisk.sh) they explicitly disallow booting with an initramfs as the only file system. The initramfs is integral to our boot procedure. Tested on real UEFI hardware. Signed-off-by: Matt Fleming <[email protected]>

Since we require the kernel source to build the efi_runtime.ko module. Currently it's possible to hit the following error, make: *** /luv-yocto/build/tmp/sysroots/qemux86-64/usr/src/kernel: No such file or directory. Stop. Reported-by: Rakesh Seethamsetty <[email protected]> Signed-off-by: Matt Fleming <[email protected]>

The ssl3_take_mac function in ssl/s3_both.c in OpenSSL 1.0.1 before 1.0.1f allows remote TLS servers to cause a denial of service (NULL pointer dereference and application crash) via a crafted Next Protocol Negotiation record in a TLS handshake. (From OE-Core master rev: 35ccce7002188c8270d2fead35f9763b22776877) (From OE-Core rev: a5060594208de172cb31ad406b34b25decd061e4) Signed-off-by: Yue Tao <[email protected]> Signed-off-by: Jackie Huang <[email protected]> Signed-off-by: Richard Purdie <[email protected]>

The DTLS retransmission implementation in OpenSSL through 0.9.8y and 1.x through 1.0.1e does not properly maintain data structures for digest and encryption contexts, which might allow man-in-the-middle attackers to trigger the use of a different context by interfering with packet delivery, related to ssl/d1_both.c and ssl/t1_enc.c. (From OE-Core master rev: 94352e694cd828aa84abd846149712535f48ab0f) (From OE-Core rev: 1e934529e501110a7bfe1cb09fe89dd0078bd426) Signed-off-by: Yue Tao <[email protected]> Signed-off-by: Jackie Huang <[email protected]> Signed-off-by: Richard Purdie <[email protected]>

The ssl_get_algorithm2 function in ssl/s3_lib.c in OpenSSL before 1.0.2 obtains a certain version number from an incorrect data structure, which allows remote attackers to cause a denial of service (daemon crash) via crafted traffic from a TLS 1.2 client. (From OE-Core master rev: 3e0ac7357a962e3ef6595d21ec4843b078a764dd) (From OE-Core rev: 33b6441429603b82cfca3d35e68e47e1ca021fd7) Signed-off-by: Yue Tao <[email protected]> Signed-off-by: Jackie Huang <[email protected]> Signed-off-by: Richard Purdie <[email protected]>

Fixes the "heartbleed" TLS vulnerability (CVE-2014-0160). More information here: https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2014-0160 Patch borrowed from Debian; this is just a tweaked version of the upstream commit (without patching the CHANGES file which otherwise would fail to apply on top of this version). (From OE-Core rev: c3acfdfe0c0c3579c5f469f10b87a2926214ba5d) Signed-off-by: Paul Eggleton <[email protected]> Signed-off-by: Richard Purdie <[email protected]>

We don't normally do this, but with the recent CVE fixes (most importantly the one for the serious CVE-2014-0160 vulnerability) I am bumping PR explicitly to make it a bit more obvious that the patch has been applied. (From OE-Core rev: 813fa9ed5e492e5dc08155d23d74127ca87304df) Signed-off-by: Paul Eggleton <[email protected]> Signed-off-by: Richard Purdie <[email protected]>

Complete the mailing list address so that users know where to send questions, patches, bug reports, etc. Signed-off-by: Matt Fleming <[email protected]>

To increase the test coverage enable all the batch mode tests in the fwts. Signed-off-by: Matt Fleming <[email protected]>

It's helpful to summarise these numbers once all the test suites have been run, as it gives concrete values instead of the user being required to count results as they scroll up the screen. Signed-off-by: Matt Fleming <[email protected]>

By using psplash we can write helpful message and the framebuffer along with a progress report. psplash has been extended slightly by adding a new keyword "DONE" which signals that progress no longer needs to be monitored and the progress bar is deleted. Also, we're using the luv project colours. Signed-off-by: Matt Fleming <[email protected]>

Add psplash to the list of packages we want to install and disable CONFIG_VT_CONSOLE so that we get a much quieter boot. The splash screen can be disabled at runtime by adding "psplash=false" to the kernel parameters. Signed-off-by: Matt Fleming <[email protected]>

The live image may be running from a USB flash drive, and because we need to mount the image in order to access the kernel and initrd images so that we can implement our crash solution, we need to make sure the kernel includes drivers for the USB host controller. Instead of worrying about the minimal number of drivers needed to work with the majority hardware, just enable them all. This results in a 2.59% increase in the kernel image size. Signed-off-by: Matt Fleming <[email protected]>

Users may want access to individual test suite results especially when trying to diagnose *why* a particular unit test failed. The results are saved in their native output format, e.g. before any post-processing of the results has occurred, which should provide more context for the failure. Make /var/log/luv/<test suite> the canonical location for results. Signed-off-by: Matt Fleming <[email protected]>

We don't need to wait for the default 10 seconds before booting the luv kernel since the most common use case is people wanting to run the luv tests, rather than interrupting the grub boot to inject some debugging options. Also add the "quiet" kernel parameter to reduce the noise of kernel messages on the console. Signed-off-by: Matt Fleming <[email protected]>

Not all EFI machines support efifb so enable the vesa framebuffer driver which should hopefully allow us to display the splash image on those machines lacking GOP support. Signed-off-by: Matt Fleming <[email protected]>

Adding "noluv" to the kernel command line will skip the luv tests on boot, which is useful in conjunction with "psplash=false" to boot the machine without displaying the splash screen. Signed-off-by: Matt Fleming <[email protected]>

Create a second partition to export the results of the LuvOS test manager. The idea behind using a different partition for the results is to avoid any corruption on the boot partition. The size of the partition is determined according to the following assupmtions: - 500 test cases - 100 lines of raw results per test case - 20 lines of persed results per test case - 80 characters per line Under such assumptions, 4.5MB of data space are required. As most USB sticks have much greater capacity, the partition size is doubled to 8MB to err on the safe side. TODO: An arbitrary UUID is used for the results partitions. This UUID must match with the UUID used by the test manager init script to mount the partition. Thus, it is necessary to obtain the UUID by either parsing such init script or modifying the UUID in the script Signed-off-by: Ricardo Neri <[email protected]> Signed-off-by: Matt Fleming <[email protected]>

Add functionality to save the test results to non-volatile media. The non-volatile media is expected to be a disk identified by its UUID. The disk is mounted using the UUID to avoid writing inadvertently writing to any other disk present in the system. Clearly, this implementation requires the presence of a disk with the UUID specified by the script. Thus, it is required to utilize such UUID when creating the filesystem. Also, disks in some systems take a while to be ready for mount and may not be ready when the LUV test manager runs. Thus, a delay of 5 seconds is introduced in order to increase the likelihood of finding the intended disk. Signed-off-by: Ricardo Neri <[email protected]> Signed-off-by: Matt Fleming <[email protected]>

gnu-efi is required by gummiboot. Imported from meta-intel, PR dropped, no new version available. (From OE-Core rev: fdb65c569db2b55023f9d1d72959e4e7d187a1d8) Signed-off-by: Stefan Stanacar <[email protected]> Signed-off-by: Matt Fleming <[email protected]>

gummiboot is a simple UEFI boot manager. Recipe imported from meta-intel with these changes: - drop PR and update configure options - upgraded to latest version A couple of notes: - If you wish you can install the gummiboot package on the target and use 'gummiboot install' to add the payload to the ESP (see gummiboot --help, just make sure the ESP partition has the boot flag on as gummiboot won't accept it otherwise). However the point of this recipe is to be used by bootimg.bbclass and generate images with gummiboot instead of grub-efi. - You need a kernel which has CONFIG_EFI_STUB=y at least - The default linux-yocto kernel config does not enable that, easiest way to enable is to build with KERNEL_FEATURES_append = " cfg/efi-ext" in local.conf (From OE-Core rev: 8f95d0598c31fff76e2a58fecb6c96197121a044) Signed-off-by: Stefan Stanacar <[email protected]> Signed-off-by: Matt Fleming <[email protected]>

Abstract away some names so one can select using EFI_PROVIDER a different class than grub-efi for populating live images, basically allowing the use of a different bootloader than grub-efi. (From OE-Core rev: 14e5de3b8f4d5902d0ac683ff45fb878e88b40ef) Signed-off-by: Stefan Stanacar <[email protected]> Signed-off-by: Matt Fleming <[email protected]>

Adds a gummiboot class similar to grub-efi class and makes the necessary changes so it can be used for live/hddimg images as well. One can set EFI_PROVIDER = "gummiboot" in local.conf to use gummiboot instead of grub-efi. Gummiboot requires some kernel options that are not enabled by default, so one has to build with KERNEL_FEATURES_append = " cfg/efi-ext". The install scripts have been updated too, keeping the old behaviour around, but accounting for the new boot loader config files (if they exist). It can be argued that the installer and bootimg are a bit wierd and not necessarily correct, but I wanted to have the exact same behviour with gummiboot. With the default EFI_PROVIDER = "grub-efi" nothing changes, everthing should be just as before. I've tested live boot, install and normal boot on: - FRI2 - genericx86-64 on NUC with: EFI_PROVIDER = "gummiboot" KERNEL_FEATURES_append = " cfg/efi-ext" in local.conf. (From OE-Core rev: b457e40fc69cc6503dc566f16495f03606e5333b) Signed-off-by: Stefan Stanacar <[email protected]> Signed-off-by: Matt Fleming <[email protected]>

efilinux is a minimal reference UEFI bootloader. It doesn't include graphics or a complex configuration syntax, just the bear minimum necessary to boot a Linux kernel. Signed-off-by: Matt Fleming <[email protected]>

Import iasl recipe from the meta-virtualization layer. This recipe is required for ovmf, which is currently relying on the host's installed iasl package. Signed-off-by: Matt Fleming <[email protected]>

Now that we've got a recipe for iasl add it as an explicit dependency for ovmf. Previously the host system was required to have the iasl package installed, which isn't very clean as our build should be completely self-contained and not using the host's packages. Signed-off-by: Matt Fleming <[email protected]>

importlib isn't currently included in any of the python packages, so create a new one for recipes that require it. Cc: Paul Eggleton <[email protected]> Signed-off-by: Matt Fleming <[email protected]>

CHIPSEC is a platform security assessment framework for analyzing the security of PC platforms. It includes a number of tests and also provides forensic capabilities for firmware. Of particular use are the UEFI secure boot variable tests and BIOS flash write-protect tests. Signed-off-by: Matt Fleming <[email protected]>

Integrate chipsec into the test manager infrastructure and add a parser to capture test results and write them to stdout in the luv-test-manager compatible format. Signed-off-by: Matt Fleming <[email protected]>

Instead of managing our own luv package just add the auto-generated files to FILES_${PN} to avoid QA warnings. In the future we could do our own luv package, but let's keep thing simple for now. Signed-off-by: Matt Fleming <[email protected]>

Include the chipsec test suite in this image. Signed-off-by: Matt Fleming <[email protected]>

Users and children of the bootimg class may need to specify at build time the volume ID of the HDDIMG vfat partition. An example of this may be when at runtime the partition is uniquely identified by the volume ID. The HDDIMG_ID is expected to be provided externally. If the variable is not set, mkdosfs will use the default volume-id. Signed-off-by: Ricardo Neri <[email protected]>

In order to uniquely identify the partition at runtime, specify a known volume-id. Signed-off-by: Ricardo Neri <[email protected]>

Add the crashkernel parameter to indicate the kernel to reserve memory for the kernel and initrd images that will be used in the event of a crash. Currently our vmlinux binary created from piggy.o is ~80MB and the initrd image is ~30MB. This could be accomodated on 128MB. Space is doubled to allow for future growth. Signed-off-by: Ricardo Neri <[email protected]>

In order to provide more useful information in case of a kernel crash, compile the kernel with debug info. Even though the kernel proper will increase it size significantly (~7-fold), the debug information will be valuable for analysis. Signed-off-by: Ricardo Neri <[email protected]>

LuvOS is used to validate UEFI firmware implementations. Such implemen- tations may be at different levels of maturity. Thus, they could pose stability threats to the Linx kernel. It could also be possibly that the kernel itself may have an undiscovered bug. The crash handler aims to handle such situations in a gracefully manner. During a regular boot, at runlevel 5, the handler uses kexec to prepare the system for an eventual crash so that the system can be rebooted to a usable state. In such recovery boot, memory dumps can be made to perform analyses on what caused the crash. Signed-off-by: Ricardo Neri <[email protected]>

This reverts commit fcdb6a8. The chipsec tests cause one of my test machines to lockup. It's highly likely that other people will experience similar issues. The fact that the tests cause lockup isn't the biggest problem, it's that there's absolutely zero diagnostic information provided to the user, so they've no data to report other than "my machine freezes". Disable the tests for now until we can provide useful ways of diagnosing lockups. Signed-off-by: Matt Fleming <[email protected]>

This recipe will help to run the bios tests suites, supported by biosbits in auto mode, the recipe also contains a configuration file which allows the tests to run in automated mode. Signed-off-by: Rakesh Seethamsetty <[email protected]> Signed-off-by: Matt Fleming <[email protected]>

All these patches have been merged into the most recent version of kexec-tools. Thus, we do not need them anymore. Signed-off-by: Ricardo Neri <[email protected]> Signed-off-by: Matt Fleming <[email protected]>

Due to the following change in the Linux kernel, commit 62e32ac3505a0cab1c5ef8ea2c0eab3b26ed855f Author: Joe Perches <[email protected]> Date: Wed Jul 31 13:53:47 2013 -0700 printk: rename struct log to struct printk_log Only version 2.0.7 or later of vmcore-dmesg may be used correctly parse the dmesg buffer. Thus, in order to have tighter control on the version, use a recipe that explicitly selects the desired version. Signed-off-by: Ricardo Neri <[email protected]> Signed-off-by: Matt Fleming <[email protected]>

The kexec-tools recipe specifies separate packages for kexec and kdump. Thus, it follows that a separate package needs to be used to install vmcore-dmesg correctly. Signed-off-by: Ricardo Neri <[email protected]> Signed-off-by: Matt Fleming <[email protected]>

Change commit 0c91196 Author: Andrea Adami <[email protected]> Date: Mon Jan 9 23:46:13 2012 +0100 explicitly splits the packages for kexec and kdump. Thus, rather than generically install kexec-tools. State explicitly which tools need to be installed. kexec is needed to reboot a Linux kernel from Linux. vmcore-dmesg is needed to extract the dmesg buffer from the /proc/vmcore dump file. Signed-off-by: Ricardo Neri <[email protected]> Signed-off-by: Matt Fleming <[email protected]>

The waiting is done already by the luv-crash-handler. Thus, there is not need to repeat the waiting here. Signed-off-by: Ricardo Neri <[email protected]> Signed-off-by: Matt Fleming <[email protected]>

The dmesg buffer can provide valuable information to determine the causes of a kernel crash. When the systems boots to runlevel 3, the luv-crash-handler uses vmcore-dmesg to recover the dmesg buffer from the vmcore dump and saves it to the LuvOS test results partition. This information can then be sent to the developers to investigate further on the causes of the crash. Signed-off-by: Ricardo Neri <[email protected]> Signed-off-by: Matt Fleming <[email protected]>

edk2 does not lend itself to cross-compiling in any form. Instead of trying to play nice and override ${CC}, etc via the usual methods, just write our toolchain variables directly in the necessary Makefiles. The upside is that we get to drop a couple of patches we were carrying. The downside is that this is a terrible hack and will likely break when we move forward to newer edk2 versions. Until such time that BaseTools supports proper configuration of a build toolchain, this is the best we can do. Cc: Ricardo Neri <[email protected]> Signed-off-by: Matt Fleming <[email protected]>

The build.sh build script in OvmfPkg allows us to optionally specify a toolchain identifer, which we're currently not doing. By not passing a toolchain ID directly build.sh invokes the build machine's compiler to figure it out. The important point in all of this is that the version identifier is used in the ovmf build path, which means that if build machine's compiler is upgraded/downgraded we can end up with multiple ovmf build directories. Pass the target's gcc version to build.sh instead. Reported-by: Ricardo Neri <[email protected]> Signed-off-by: Matt Fleming <[email protected]>

Technically the bios used by qemu isn't a native package, so we shouldn't depend on it. Instead require the regular version of ovmf. Cc: Ricardo Neri <[email protected]> Signed-off-by: Matt Fleming <[email protected]>

In order to enable netbooting and transfer of required files using tftp, efinet and tftp modules are required. Signed-off-by: Pradeep Gaddam <[email protected]> Signed-off-by: Matt Fleming <[email protected]>

…uefi BIOS efinet.c - Changed this file to not give up on sending packets in the network immediately when sending the packet fails. The change was to make few more attempts to send packet out into network. tftp.c - Increased the Block size of tftp packet to decrease the time to boot as on qemu it is taking a lot of time to load all required files. I cherry picked these changes from the following link : http://lists.gnu.org/archive/html/grub-devel/2012-04/txtcxcL7X6prZ.txt Signed-off-by: Pradeep Gaddam <[email protected]> Signed-off-by: Matt Fleming <[email protected]>

commit "runqemu: Add option for custom BIOS directory" added a biosdir="" option to runqemu which provides a shortcut for specifying a path to a custom qemu bios (indirectly via qemu's -L option). However, biosdir="" searches in $OE_NATIVE_SYSROOT for bios.bin, meaning that the qemu bios recipe must be built as a native recipe. Arguably all qemu firmware should be built as non-native because they're not executed directly by the build machine. Switch biosdir="" to search in $OE_TMPDIR/sysroots/$MACHINE, which is the location of non-native files. Cc: Ricardo Neri <[email protected]> Signed-off-by: Matt Fleming <[email protected]>

The OVMF BIOS seems to prefer the defautl Cirrus emulated VGA adaptor, otherwise the QEMU window just shows a blank screen. NOTE: OVMF is detected simply by looking for the substring "ovmf" in the biosdir argument to runqemu. This is, of course, not a robust test, but it works well enough if you're just using the ovmf recipe in meta-luv. Signed-off-by: Max Eliaser <[email protected]> Signed-off-by: Matt Fleming <[email protected]>

While EDK2 has many components that support ARM, OVMF is not one of them. This should be reflected in the BitBake recipe for OVMF. This is accomplished by adding a COMPATIBLE_HOST annotation to ovmf_git.bb. Signed-off-by: Max Eliaser <[email protected]> Signed-off-by: Matt Fleming <[email protected]>

This fixes an issue where the OVMF recipe always built an x86_64 firmware regardless of the target architecture or target machine. This issue manifested itself when qemu-system-i386 was used on an OVMF firmware image-- QEMU would exit immediately rather than proceeding with the boot process. The x86_64 firmware was unaffected by this issue. Signed-off-by: Max Eliaser <[email protected]> Signed-off-by: Matt Fleming <[email protected]>

This SVN revision has a fix for build failures under GCC 4.9 for x86_64 targets. This will ease eventual upstreaming of the OVMF recipe to Yocto master, which uses GCC 4.9. The bug is described on the EDK2 mailing list here: http://www.mail-archive.com/[email protected]/msg07451.html GCC 4.9 changed an alignment from 32 bytes to 64 bytes when building for a 64- bit target, whereas OVMF was built with the expectation that the alignment would be 32 bytes. This mismatch caused a build failure that looked like this: GenFw: ERROR 3000: Invalid Unsupported section alignment. SVN revision 15699 of EDK2 includes a fix for this issue by the upstream EDK2 developers. When compiling with GCC 4.9 for a 64-bit target, the build OVMF build configuration now anticipates a 64-byte alignment. Signed-off-by: Max Eliaser <[email protected]> Signed-off-by: Matt Fleming <[email protected]>

When we have structures that contain pointers, we cannot just do double dereference on it as in struct->pointer->value. This will end up as a Page Fault. To fix this problem, I changed efi_runtime kernel module to first fetch the entire structure into a local copy on stack and use that to reference the other data members of the struct. Signed-off-by: Pradeep Gaddam <[email protected]> Signed-off-by: Matt Fleming <[email protected]>

Using the default values. Signed-off-by: Matt Fleming <[email protected]>

It's very useful to know if the firmware makes illegal accesses to the EFI boot services regions. Enable CONFIG_EFI_BOOT_SERVICES_WARN which prints a warning if such an access occurs. Signed-off-by: Matt Fleming <[email protected]>

The Linux kernel provides lots of information about its errands with the UEFI firmware at boot time. However, such information is buried inside the dmesg buffer. Automated test resources, such as LuvOS, need automated tools to find, parse and format such information. kernel-efi-warnings rely on the dmesg buffer to find specific expressions. The absence or presence of such expressions are treated as a test pass/fail. For now, the scripts only looks for the warnings that the kernel issues if it detects illegal accesses to the EFI_BOOT_SERVICES_* memory regions. The script is intended to grow as mroe kernel tests become available. Signed-off-by: Ricardo Neri <[email protected]>

In order for the LuvOS test manager to take the rest results of kernel-efi-warnings, we need a parser that presents the results in a format that it can understand. Signed-off-by: Ricardo Neri <[email protected]>sdfsdf

Add a recipe to install kernel-efi-warnings. As it inherits from luv-test, it also takes care of installing the luv-parser. Signed-off-by: Ricardo Neri <[email protected]>

Make sure kernel-efi-warnings is included in the ramfs filesystem. kerne-efi-warnings is a simple script that analyzes the kernel messages to search for warnings that the kernel might issue if/when it finds potential bugs in the UEFI firmware. Signed-off-by: Ricardo Neri <[email protected]>

In order to display images from the boot menu we're going to need to load gfxmenu, gfxterm and other grub modules. Turn them on in our luvOS-specific bbappend. Also include the chainloader module which is how we're going to execute the bits tests. Signed-off-by: Matt Fleming <[email protected]>

bits comes with its own version of the grub bootloader, with custom modules installed as part of the grub image, for example a python module to interpret the python tests. We must install this boot loader alongside our default one, along with the necessary parsers and test runners to extract the results of the bits tests from userland. Signed-off-by: Matt Fleming <[email protected]>

Without the header the bits runtime complains like so, Traceback (most recent call last): File "<string>", line 1, in <module> File "/boot/python/init.py", line 69, in init bitsconfig.init() File "/boot/python/bitsconfig.py", line 42, in init config.read("/boot/bits-cfg.txt") File "/boot/python/lib/ConfigParser.py", line 305, in read self._read(fp, filename) File "/boot/python/lib/ConfigParser.py", line 512, in _read raise MissingSectionHeaderError(fpname, lineno, line) ConfigParser.MissingSectionHeaderError: File contains no section headers. file: /boot/bits-cfg.txt, line: 3 'batch = test acpi smbios\n' Include a [bits] header and make sure we actually install the config file. Signed-off-by: Matt Fleming <[email protected]>

When we've finished running the bits tests we need to return control to the luvOS bootloader, so that we can run further tests, like booting a Linux kernel. Signed-off-by: Matt Fleming <[email protected]>

Curiously, the checksums have become invalid for bits-1084.zip. Update them. Signed-off-by: Matt Fleming <[email protected]>

Doing so will garble any image displayed by the main boot loader. We're directly patching the python code as shipped in the bits release zip file. This means that we need to regenerate the .pyc files, which are required for bits to execute properly. Signed-off-by: Matt Fleming <[email protected]>

Signed-off-by: Matt Fleming <[email protected]>

Provide a luv splash screen for grub's use, which informs the user that we're currently running boot tests. Signed-off-by: Matt Fleming <[email protected]>

For luvOS we need to install more than one boot loader, plus a variety of files to do boot time tests. Roll our own class which gives us much more control over the installation process and creation of the grub config. Signed-off-by: Matt Fleming <[email protected]>

There are tools such as sbsigntool that require a native version gnu-efi. Add support for building gnu-efi-native. Cc: Stefan Stanacar <[email protected]> Signed-off-by: Matt Fleming <[email protected]>

sbsigntool is a useful utility for cryptographically signing binaries for use in UEFI systems. Signed-off-by: Matt Fleming <[email protected]>

The tool cannot be built without this patch. We don't really need the man pages in this project. Signed-off-by: Ricardo Neri <[email protected]>

LuvOS is intended to be an automated testing tool. Thus, it does really require much interaction from the user other than inserting the bootable media, and removing it when tests are complete. Thus, we inform better to the user what to do when the tests are complete. Signed-off-by: Ricardo Neri <[email protected]>

Rather than hardcoding the package version, utilize the PV variable to define the configure path. This will prevent build issues when moving to a newer version. Signed-off-by: Ricardo Neri <[email protected]>

1090 is the lastest release of the BIOS Implementation Test Suite (BITS). These are the changes as per BITS relase notes: *Update to ACPICA version 20140325 *Backport GRUB2 commit 4e42521d8c9232b6ee9eac7d8b4945a7479de781 to preserve 16-byte alignment of the stack on EFI calls * python: Support the csv module Signed-off-by: Ricardo Neri <[email protected]>

This patch was merged in V14.07.00 of fwts but we're currently on V14.03.01. We're seeing microcode test failures on a number of Haswell machines, which shouldn't really be failures. This patch from upstream fixes that. Signed-off-by: Matt Fleming <[email protected]>

A recent fix to change a securebootcert failure into an info message was merged into upstream fwts. We're hitting this false positive on a number of machines. Signed-off-by: Matt Fleming <[email protected]>

We've no need to turn a data structure on the stack into a pointer and use get_user() on a kernel address, and in fact, doing so is likely to return -EFAULT, because the argument isn't a valid user address. This bug was introduced in commit c2268a9 ("fwts: Copied the structure from userland locally in kernel space"), and causes a uefirttime failure. Cc: Pradeep Gaddam <[email protected]> Signed-off-by: Matt Fleming <[email protected]>

Things don't work across the board right now, anyhow. People have even experienced hangs, and there's no way we can diagnose those hangs without displaying the test log on the screen. It's not ideal, and going forward we definitely need to satisfy both requirements (easy debugging and pretty graphics), but if need to pick one or the other, easy debugging wins every time. Signed-off-by: Matt Fleming <[email protected]>

Providing verbose kernel output on the serial console during boot is a useful way to debug issues. It also provides a much more informative message of what is currently happening. Enable both the standard ttyS0 device and ttyPCH0 which is the device used on the Intel Minnowboard. Tested-by: Gayatri Kammela <[email protected]> Tested-by: Ricardo Neri <[email protected]> Signed-off-by: Matt Fleming <[email protected]>

There's some latency somewhere in the test result pipeline, and it's impossible to trace noticable hangs when writing test results to the console to the offending unit test. Just simplify the pipeline, and pipe the result output directly to the console and results files instead of passing it through another instance of gawk before it hits the console, since hunting down buffer-related delays in gawk is extremely tedious. There's no user-visible change with this patch - it's preparatory work for a later patch that aggressively flushes the gawk output buffer. Tested-by: Gayatri Kammela <[email protected]> Tested-by: Ricardo Neri <[email protected]> Signed-off-by: Matt Fleming <[email protected]>

The internal buffering that gawk does makes the test output pretty useless, because it's not always possible to tell which test is currently running. For example, a test may have completed but the output will not appear on the screen until the output buffer fills and is subsequently flushed. Effectively all the unit test results from a single test suite are output as one block. The wakealarm test from fwts provides a good illustration of the user-visible problem. This test takes a number of seconds to complete, but because all the fwts results are output in one go, it's not possible to attribute delays to any one individual unit test. Explicitly flush all open file descriptors and pipes anytime that we print something from gawk. This gives much better user interaction when looking at the serial console because it's now possible to figure out which tests have the longest latencies. Whenever a unit test begins execution a message will be printed on the serial console immediately, e.g. [+] wakealarm... and when it finishes (in this case after multiple seconds) the result will be printed too, [+] wakealarm... passed Tested-by: Gayatri Kammela <[email protected]> Tested-by: Ricardo Neri <[email protected]> Signed-off-by: Matt Fleming <[email protected]>

As more LUV versions are released and the user base grows, it is important to know which particular version of LUV a given user is running. Knowing the version makes it easier to provide support and comments regarding bugs and supported features. Containing a summary of all the tests, luv.results is a good place to print the LUV version. The version is pulled from the /etc/issue file, which is updated with every release. While here, update also the welcome message in the console to depict the LUV version. Signed-off-by: Ricardo Neri <[email protected]> Signed-off-by: Matt Fleming <[email protected]>

Signed-off-by: Ricardo Neri <[email protected]> Signed-off-by: Matt Fleming <[email protected]>

Windows will not mount and display more than one partition on a removable disk. Make the results partiion the first partition on the image so that users are able to easily find the test results. Having this one partition restriction isn't a problem otherwise, since there's little value in auto-mounting the other partition (the EFI System Partition) anyway. Of course, we're assuming firmware is smart enough to find the EFI System Partition based on the partition table. Fixes issue #23. Reported-by: William Leara <[email protected]> Debugged-by: Ricardo Neri <[email protected]> Tested-by: Gayatri Kammela <[email protected]> Signed-off-by: Matt Fleming <[email protected]>

This allows us to setup the correct DISTRO and MACHINE out of the box, along with adding the meta-luv layer to BBLAYERS. Signed-off-by: Matt Fleming <[email protected]>

Issue #30 showed that we currently have far too many configuration steps that are required to build a luv image. And if you miss any of those steps, the resulting failures are not at all helpful in diagnosing what you did wrong. Start simple for now, by just specifying a luv-specific TEMPLATECONF directory. Later we could expand upon this and perhaps make it easier to build for the various architectures. Signed-off-by: Matt Fleming <[email protected]>

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

First attempt at a simple wrapper for configuring luv-yocto #43

First attempt at a simple wrapper for configuring luv-yocto #43

Commits on Dec 12, 2013

Commits on Dec 13, 2013

Commits on Mar 19, 2014

Commits on Mar 20, 2014

Commits on Mar 21, 2014

Commits on Mar 27, 2014

Commits on Apr 3, 2014

Commits on Apr 16, 2014

Commits on Apr 25, 2014

Commits on May 1, 2014

Commits on May 2, 2014

Commits on May 8, 2014

Commits on May 9, 2014

Commits on May 10, 2014

Commits on May 12, 2014

Commits on Jun 2, 2014

Commits on Jun 4, 2014

Commits on Jun 6, 2014

Commits on Jun 7, 2014

Commits on Jul 9, 2014

Commits on Jul 15, 2014

Commits on Jul 16, 2014

Commits on Jul 17, 2014

Commits on Aug 1, 2014

Commits on Aug 6, 2014

Commits on Aug 19, 2014

Commits on Aug 20, 2014

Commits on Sep 25, 2014

Commits on Sep 26, 2014

Commits on Oct 7, 2014

Commits on Oct 23, 2014

Commits on Nov 4, 2014

Commits on Nov 10, 2014

First attempt at a simple wrapper for configuring luv-yocto #43

Are you sure you want to change the base?

First attempt at a simple wrapper for configuring luv-yocto #43

Commits on Dec 12, 2013

Commits on Dec 13, 2013

Commits on Mar 19, 2014

Commits on Mar 20, 2014

Commits on Mar 21, 2014

Commits on Mar 27, 2014

Commits on Apr 3, 2014

Commits on Apr 16, 2014

Commits on Apr 25, 2014

Commits on May 1, 2014

Commits on May 2, 2014

Commits on May 8, 2014

Commits on May 9, 2014

Commits on May 10, 2014

Commits on May 12, 2014

Commits on Jun 2, 2014

Commits on Jun 4, 2014

Commits on Jun 6, 2014

Commits on Jun 7, 2014

Commits on Jul 9, 2014

Commits on Jul 15, 2014

Commits on Jul 16, 2014

Commits on Jul 17, 2014

Commits on Aug 1, 2014

Commits on Aug 6, 2014

Commits on Aug 19, 2014

Commits on Aug 20, 2014

Commits on Sep 25, 2014

Commits on Sep 26, 2014

Commits on Oct 7, 2014

Commits on Oct 23, 2014

Commits on Nov 4, 2014

Commits on Nov 10, 2014