There is a potential problem if we don't remove the previous version's
stamp, for example:
The depend chain is:
libtool-native -> autoconf-native -> m4-native
We have two m4-native: 1.4.9 and 1.4.7
1) Clean all of them to make a fresh build so that we can reproduce the
problem
$ bitbake m4-native autoconf-native libtool-native -ccleansstate
2) Build libtool-native so that the m4-native_1.4.17 will be built
$ bitbake libtool-native
3) Set PREFERRED_VERSION_m4-native = "1.4.9" and build again
$ bitbake libtool-native
4) Set PREFERRED_VERSION_m4-native = "1.4.17" and build again
$ bitbake libtool-native -ccleansstate && bitbake libtool-native
Then the build will fail:
[snip]
| m4: unrecognized option '--gnu'
| Try `m4 --help' for more information.
| autom4te: m4 failed with exit status: 1
[snip]
The is because when we change m4-native to 1.4.17 and build
libtool-native again:
5) libtool-native depends on autoconf-native, and autoconf-native's
version isn't change, so it can remove the current stamp and mirror
the sstate (the one depends on m4-native_1.4.9) from the SSTATE_DIR
correctly.
6) The mirrored autoconf-native depends on m4-native_1.4.17's
do_populate_sysroot, and the stamp is already there (which is made
by step 2), so it would do nothing, but this is incorrect, since
the one that really in the sysroot is m4-native_1.4.9, then the
error happens.
Remove previous version's stamp in sstate_clean() will fix the problem.
[YOCTO #5422]
(From OE-Core rev: 4659d29b1040349116549644e45035a5b37d9311)
Signed-off-by: Robert Yang <liezhi.yang@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
do_package doesn't exist and are noexec anyway for native/cross/crosssdk
packages.
(From OE-Core rev: 1028ac813fa9803ebfff6bcfa7f8b67012609b27)
Signed-off-by: Ming Liu <ming.liu@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
This patch mainly aims to add optimisation for crosssdk setscene dependency
validating which we haven't handled in current logic, and which I think we
could have as we've already implemented to native/cross, although there
are albeit not many crossdk tasks, we could still get some performance
enhancement.
(From OE-Core rev: 1094983ff87a8b745a5bc7bfe9514433ee3c4ad2)
Signed-off-by: Ming Liu <ming.liu@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
The way the "all arch" PKGSPEC is enabled, it causes corruption
of the datastore of sstate operations against other tasks.
Data store copies are cheap and allow us to use that trick, resetting
to a clean copy of the data afterwards.
(From OE-Core rev: afaf16100efa79a275a2f4b9f2caa80decfdeb81)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Like fetch, unpack and patch, populate_lic doesn't vary between different
archs so we should mark it as such. This means better sstate cache reuse
with fewer duplicate files as well as less confusing sstate debugging.
sstatesig also needs to account for the fact BPN is used for sstate files
in these cases.
(From OE-Core rev: 3d59d0bed756f64d0092caa3892239c779c4a341)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Currently the code has problems differentiating between "gcc-cross" and "gcc-cross-initial"
sstate files. We could add in a ton of special casing but tests show this isn't scaling
well. Using a more unique separator resolves the issue.
The choice of which separator to use is a hard one. We need something which isn't commonly
used in PN, PV, PR, *_OS and *_ARCH which rules out '-', '_' and it needs to work ok with
webservers/http which makes ';' and '%' harder.
The change also sets SSTATE_SWSPEC globally since writing out differently named siginfo
files for the fetch/unpack/patch tasks is a waste of diskspace, the hashes match for
all PN in the majority of cases and if they don't, its not a big issue as the hash is
different. This makes the results from sstate debugging more understandable.
(From OE-Core rev: 6f823a23c5f1d0ffa0a27db1c1bc1907de788505)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
When originally developed, it was thought a task may have more than one associated
sstate archive. The way the code has grown that idea is now not possible or needed.
We can therefore assume one sstate archive per task and drop the crazy name
mapping code. Simpler is better in this case.
The downside is that various sstate archives will change name so this forces a cache
rebuild. Given the other sstate changes going in at this time, this isn't really
a bad thing as things would rebuild anyway.
(From OE-Core rev: 5afe86a6854b21692fd97c5fc7fab50dbc068acb)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
For a shared workdir, any one of the fetch/unpack/patch tasks may run yet the
PN and architecture fields in SSTATE_PKGSPEC may differ. This makes looking up
the appropriate siginfo file near impossible.
I've tried several different ways of resolving this and this is the neatest
solution I could find, its still rather ugly. I believe the usefulness of
better sstate debugging outweighs the ugliness of the code.
This patch also changes the sstate_checkhashes() code to look for siginfo
files rather than the actual sstate packages themselves. This means the
function can be used in other contexts to find info files for tasks that
may not have sstate data. It is assumed that sstate mirrors will have both
files available. This is done to allow bitbake to query whether tasks have
matching signatures in sstate directories or not.
(From OE-Core rev: 068e4289b597699cbff2dfde44ba833af4535281)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Currently siginfo files are only written for sstate tasks. In order to be truly
debuggable, its helpful to have the siginfo for intermediate tasks. This
adds that functionality so the extra siginfo files are written out too.
This will be used to add better sstate debugging in future changes.
(From OE-Core rev: 04d108cd16f5ad8f92a62ea537d1330fee712470)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
sstate_package creates hardlink from sysroot to SSTATE_BUILDDIR, and
sstate_create_package stores SSTATE_BUILDDIR into a archive file by
tar, these two task can be run simultaneously for different packages,
and make a hardlink for a file will lead to the change of the links
number of file, and if tar is reading this file, it will fail with exit
code 1, and report "file changed as we read it":
DEBUG: Executing shell function sstate_create_package
tar: x86_64-linux/usr/share/aclocal/xorg-macros.m4: file changed as we read it
4b3e353a5[sstate.bbclass: fix parallel building issue] tries to use the
tar parameter --ignore-failed-read to fix, but it does not work, and
tar parameter --warning=no-file-changed can close the warning, but can
not change the exit code. so close shell immediate exit, only fail
if tar returns not 1 and 0.
Exit codes of tar:
http://www.gnu.org/software/tar/manual/html_section/Synopsis.html
(From OE-Core rev: fad604b719e00b03e09da5fdb485e72332275b4a)
Signed-off-by: Roy Li <rongqing.li@windriver.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
sstate_package creates hardlink from sysroot to SSTATE_BUILDDIR, then
sstate_create_package will store SSTATE_BUILDDIR into a archive file by
tar, but once other packages install the same file into sysroot, the
creating the archive file will fail with below error:
DEBUG: Executing shell function sstate_create_package
tar: x86_64-linux/usr/share/aclocal/xorg-macros.m4: file changed as we read it
This kind of error is harmless, use --ignore-failed-read to ignore it.
The error in tar occurs when the timestamp of the file changes and this
can happen when the number of symlinks change. The file will be included
in the archive.
[YOCTO #5122]
(From OE-Core rev: 4b3e353a532c7b68b0bb86df4a2fcc44f8bb3ef2)
Signed-off-by: Roy.Li <rongqing.li@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
These have been deprecated for a long time, convert the remaining
references to the correct modules and prepare for removal of the
compatibility support from bitbake.
(From OE-Core rev: 6a39835af2b2b3c7797fe05479341d71a3f3aaf6)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
ant reported on irc that the sstate absolute to relative symlink creation
code wasn't working in klibc. He was correct although the level of breakage is
rather surprising since it only worked for one level of symlink (usr/include) with
everything else being broken.
The reason is probably that nothing really uses absolute paths, we use relative
paths where at all possible already. Nothing in the target sysroot should use
absolute paths for a start. In this regard, the klibc-dev package is broken and
needs fixing. It will currently break when building for one machine, then switching
to another of the same TUNE_PKGARCH and installing from sstate but that is a
separate issue.
This patch fixes the symlink creation code by firstly passing in the correct
value we need (where the symlink will end up) and seccondly, actually using it.
I've also tweaked the debug message to contain appropriate information and got
right of the double "//" value the existing code created in favour of the form
'./..' which looks neater.
(From OE-Core rev: 9b05c65450526522d7358d0c0901b594de546748)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
BB_NO_NETWORK disables any fetching, however if we're using an external
sstate cache, we may want to be able to fetch those objects even if we are
not fetching the upstream sources. Denote this situation by setting
SSTATE_MIRROR_ALLOW_NETWORK in local.conf. When it is found, for sstate
cache fetches, mask off BB_NO_NETWORK for the local function.
(From OE-Core rev: ed585cad2e1fdc323c05fa82055a071bcf98d1bc)
Signed-off-by: Joe MacDonald <joe.macdonald@windriver.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
This avoids unnecessary duplication of setup. The only visible change in
behavior will be the case if siginfo exists and the archive does not, in which
case it'll redownload both, but this doesn't seem unresasonable to me,
particularly since the archive is downloaded first, making this case
particularly unlikely.
(From OE-Core rev: aa4991c307d4bbdd06c3cbf8448240b74c5e01c4)
Signed-off-by: Christopher Larson <chris_larson@mentor.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
The default FILESPATH isn't really of use, as we don't expect to find sstate
archives buried in layer recipe directories, and the default MIRRORS is
intended for use for fetching SRC_URI, not sstate.
(From OE-Core rev: 46402b2f5b69004751f6663d435bedae0ad9dab1)
Signed-off-by: Christopher Larson <chris_larson@mentor.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
The do_package_setscene task needs to depend on fakeroot in order to correctly
install its files.
We can whitelist the dependency in the sstate handling code for some
performance improvements since we only need this if we're installing the
package from sstate.
Also use an append operator in base.bbclass for clarity.
(From OE-Core rev: 0810ea2a72bdea67a3d8002c4e12fb20f45cf1d5)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
sstate has a logic to fixup hardcoded paths in scripts,
but it misses in some specific cases, so add
EXTRA_STAGING_FIXMES to the fixup hardcoded paths
mechanism, so that we can specify what hardcoded paths
need to be fixed in a recipe, e.g.
EXTRA_STAGING_FIXMES = "STAGING_BINDIR_TOOLCHAIN"
(From OE-Core rev: 2e840db56c45b4c63fded55f4ed763b7099284b9)
Signed-off-by: Jackie Huang <jackie.huang@windriver.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
In python 3 print is a function call. In some cases bb.note is a more
appropriate call to make.
(From OE-Core rev: 754874e68d1037c76fc4cc7d35d3fa12946d00f8)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Make hard links for staging files instead of copy to save the disk space
(3G will be saved for a core-image-sato build), and it doesn't affect
much on the build time.
The following directories are affected:
1) The sysroot
2) The DEPLOY_DIR
3) The pkgdata
[YOCTO #4372]
(From OE-Core rev: 5853e0f482b22258c909268fe71673a29e31989b)
Signed-off-by: Robert Yang <liezhi.yang@windriver.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Otherwise we might end up creating directories under sstate-cache with whatever
random umask has been selected for the task that we're trying to package. This
would be a bad thing since it might result in losing group write access for
newly created dirs, and/or losing group read access for the sstate files
themselves.
(From OE-Core rev: d8c4f442c41bf3ac5e064630657cd3fa1b5c43b1)
Signed-off-by: Phil Blundell <philb@gnu.org>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
If no files have been staged we want to continue without error instead
of showing a traceback.
Fixes [YOCTO #4056].
(From OE-Core rev: ca36be708e54c0c86535bc8512295c76c48f6cf5)
Signed-off-by: Paul Eggleton <paul.eggleton@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Under some conditions (ACLs enabled, NFS) mv can interactively prompt
before overwriting files. Avoid hanging builds in that case by using
-f which should be harmless in other cases.
(From OE-Core rev: b1a085db9d8ad2a3117af6f50e510bc9c2f9407b)
Signed-off-by: Jeff Polk <jeff.polk@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
In case filenames have spaces, execution of the function
sstater_install will hang, so the print parameter %s must be
enclosed with quotes.
(From OE-Core rev: 545d7aa26dfefdc927e0f4e2cc37398ef2c63fa6)
Signed-off-by: Leonardo Sandoval <leonardo.sandoval@freescale.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
When checking the dependencies in setscene_depvalid(), make sure we also
consider those dependencies needed when running the postinstalls on
host.
[YOCTO #3918]
(From OE-Core rev: 8de0616825ed1b238b3486077af6897834bcb62d)
Signed-off-by: Laurentiu Palcu <laurentiu.palcu@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
This directory is cleaned upon completion however if a previous build
crashes, it can lead to corrpution, hence ensure its clean at the start
too.
(From OE-Core rev: 8ef0e59d5a7da3671d1ad9a54fe068ed78f928d5)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Add a function which copys a tree as a set of hardlinks to the original
files, then use this in sstate to reduce some of the overhead of sstate
package creation since the file isn't actually copied.
(From OE-Core rev: 8e373e69acac853213a62afb8bbdf0adc0c5045a)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
The same log message gets output multiple times in the log which look
confusing and is rather pointless. Move the log message to the correct
level.
(From OE-Core rev: 3917409004a830e7ad0646f05ad7421385cbd1de)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
useradd.bbclass adds sstate dependencies on base-passwd, shadow-native and
shadow-sysroot. Due to the way these are injected, they interact badly with
the other dependency validation logic and end up pulling in dependencies we
don't actually need. This patch adds code to optimise those cases out.
(From OE-Core rev: 784ca68fcca4ffb34390d55d9343570cfdf0305f)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Currently, do_rootfs has a dependency on all the do_package output being present
due to its usage of the pkgdata directories. This means that if you run:
bitbake xxxx-image -c rootfs
you end up having to fetch and unpack all the do_package data which is usually
large and inefficient. It also means rm_work has to leave all the do_package
data lying around so rootfs works.
This patch splits the actual creation of the pkgdata directory off into a separate
task, "packagedata" which happens immediately after do_package. We can then remap
the dependencies so this task is depended upon, not do_package. Sstate can then be
programmed not to require do_package at the appropriate times.
Whilst this patch doesn't do so, it opens the possibility of rm_work wiping
out the do_package output from WORKDIR as long as it also removed the do_package
stamp (both normal and setscene variants) and allowing more space savings
with rm_work which has been regularly requested.
(From OE-Core rev: 6107ee294afde395e39d084c33e8e94013c625a9)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Specify dirs in which the shell function "sstate_create_package" and
"sstate_unpack_package" are executed and don't use ${B} as default dirs
to avoid possible race with task do_rootfs at deb image creation time.
[YOCTO #3674]
(From OE-Core rev: ccef1cf783669a4683eda9d4b44dbe6bcf426259)
Signed-off-by: Hongxu Jia <hongxu.jia@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
If you change a machine to a different package architecture, you will see sstate
errors about overwriting files as the code stands today. Instead it should clean
out the files safely and correctly. This patch changes the naming of stamp-extra-info
manifest files to avoid this problem. It will potentially trigger warnings during
builds in existing TMPDIRs until the system adjusts to the new naming, these are
harmless.
[YOCTO #3521]
(From OE-Core rev: 2cc8ee57f8148844bb7bcd4aaf34f6891cf3d410)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
The code was written before SSTATETASKNAMES existed. Since it
does exist, lets simply the code.
(From OE-Core rev: 9817e2efdb94395655d711f5eadedcd249c8cffe)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
This would be useful for doing siginfo compares to understand why a build
is not reusing something when using SSTATE_MIRROR. No error will be reported
if it fails to find the .siginfo file
[YOCTO #2898]
[RP: Small tweaks]
(From OE-Core rev: 6d86690330f0d43839b904fced4b4b02cb27b8c6)
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
* without this patch:
Python 2.7.3
>>> sstate_search_cmd = "grep -rl /OE/jansa-test/shr-core/tmp-eglibc/pkgdata/armv7a-vfp-neon-oe-linux-gnueabi/runtime-reverse/vim-common /OE/jansa-test/shr-core/tmp-eglibc/sstate-control --exclude=master.list | sed -e 's:^.*/::' -e 's:\.populate-sysroot::'"
>>> cmd_array = sstate_search_cmd.split(' ')
>>> search_output = subprocess.Popen(cmd_array, stdout=subprocess.PIPE).communicate()[0]
grep: |: No such file or directory
grep: sed: No such file or directory
* Adding shell=True and using cmd string instead of array makes it work:
>>> search_output = subprocess.Popen(sstate_search_cmd, shell=True, stdout=subprocess.PIPE).communicate()[0]
>>> print search_output
manifest-armv7a-vfp-neon-gvim.package
manifest-armv7a-vfp-neon-vim-tiny.package
manifest-armv7a-vfp-neon-vim.package
But still isn't 100% reliable, I guess it's caused by some other package
being removed from sstate while grep is already running.
So sometimes grep can show error on STDERR
>>> search_output = subprocess.Popen(sstate_search_cmd, shell=True, stdout=subprocess.PIPE).communicate()[0]
grep: /OE/jansa-test/shr-core/tmp-eglibc/sstate-control/manifest-armv7a-vfp-neon-systemtap.package: No such file or directory
(From OE-Core rev: d84f7d7a12b4271f7b2bfde9fb356d750abff15d)
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
The value of subprocess.Popen().communicate()[0] is a string.
Checking for '!= None' will always match causing bogus warnings
regarding already staged files.
(From OE-Core rev: acdd76482efc3caf80e9e0b7359be7ca724ae09a)
Signed-off-by: Enrico Scholz <enrico.scholz@sigma-chemnitz.de>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Currently there is some odd behaviour of the packagegroup class in relation
to sstate since it sets PACKAGE_ARCH = "all" but does not use the allarch class
leading to it being undetected by sstate.
Previously it was not possible to use allarch as the recipe couldn't "undo"
settings made by the allarch class. Since this no longer happens when
PACKAGE_ARCH != all, we can use the allarch class.
This patch also fixes up one case we need to preserve TRANSLATED_TARGET_ARCH
and ensures sstate only assumes allarch when PACKAGE_ARCH is "all".
(From OE-Core rev: 591fa7c1ab9e9ff75fdce602c77ecdeda3a255d9)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Clean up and clarify the populate_sysroot task dependencies. Target sysroot
packages do need their dependencies installed, as do some target/cross
relationships. We can whitelist the *-initial dependencies as these are
never needed indirectly.
(From OE-Core rev: eeec307917234d97be2674beeadef71599fb1487)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
For example gcc-cross depends on linux-libc-headers and needs it to be present
to build/work correctly.
(From OE-Core rev: 43ce7a1d86bf82d976ad241057a4207b1a340b3b)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
This is a first attempt at logic to determine when a sstate dependency needs
to be installed and when it does not. Its a start at the logic and errs on the
side of caution, as it gets wider testing, we can refine the logic as needed.
This code should allow a significant performance speedup to certain workflows, for
example "bitbake xxx-image -c rootfs" will not populate the target sysroot.
(From OE-Core rev: b43faba37816817edc5240a139361d16e07c6131)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
The recent preveeding sstate directory layout fixes made the code do what it
was originally intended to do, as can be clearly seen from the code.
Unfortunately this changed the contents and layout of the sstate files
themselves since the bug was leading to a directory prefix being missing.
This is now resulting in chaotic messages on the console since things
are getting confused with the two different layouts. The simplest way to
resolve this is to bump the version number, hence moving the new layout
into its own new namespace.
Its worth noting that whilst the failure messages are scary, the failure
mode is relatively harmless since it will just fall back to building the
data rather than installing from sstate.
Usually I'd give more notice of a change like this but under the
circumstances, I'm just going to push this in to resolve the failures
people are seeing. Initially I thought the problem was limited to
some of the -cross packages and therefore of low impact but that is
clearly not the case.
(From OE-Core rev: b53ea6687b6201c8c5ab5cb0d2a845ef7e7b2abe)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Fix missing parameter to endswith and pass paths through normpath to remove
any duplicate "/" characters which would corrupt other calls like basename.
(From OE-Core rev: 172a74c540378149eec493c37c030e9f42f9603d)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
The manifest file can become corrupted if sstate-inputdirs and sstate-outputdirs
don't have matching endings. This patch ensures that even if set incorrectly,
the code functions as intended, thereby handling manifest corruption safely.
(From OE-Core rev: 0109a3623a19f9ae289952a4f054e53c3eca4eaa)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Present the manifest file that contains the matches for
files being installed to a location that already contains
that file. This will help to determine which is the correct
recipe to fix when this occurs.
[YOCTO #3191]
(From OE-Core rev: 56268f6e4ed1fc11143173bb1717a8be78c728a5)
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Now do_package isn't machine specific, we're only left with do_populate_sysroot as a
machine specific task. This change marks only the machine specific manifests as machine
specific, defaulting to PACKAGE_ARCH for everything else.
This means we do less work where there are multiple machines using the same
core package architecture and we can start to clean up the sstate duplicate files
whitelist.
(From OE-Core rev: febeaf3d1b8917b660c7279b008d8b03337568e9)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
We've noticed failures on the project autobuilders where a shared sstate
directory is used across multiple builders and the clocks become skewed.
Most of the time this causes harmless building but if this happens where
an environment is changed (make install vs make in qt4-x11-free for example),
the build can fail.
This avoids modification times in the future and should make builds safer
in shared environments sstate was designed for.
(From OE-Core rev: 8f1bdb4f4afd7f5f4c121be8ba82f4675f73e300)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
This avoids errors where gcc/binutils get installed to the native sysroot
in the same location for multiple package architectures. Ultimately making
these native recipes with ${PACKAGE_ARCH} appended to PN will resolve this
but hide the warnings until this gets sorted out.
Also hide the python and docbook catalog warnings since they're known about,
nothing to worry about and we'll aim to clean them up properly in the 1.4 cycle.
(From OE-Core rev: 5bae58a5b59c04d8947f4842f19837a914c29b52)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Without this, path components like // could break comparisions with the whitelist leading
to warnings being displayed to the user unintentionally.
(From OE-Core rev: d3c46ca56fab2f07bf16b61514f30765543a8747)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
do_package is a machine specific task at the moment due to packagedata. This means
do_package tasks and their dependencies rerun between different machines
with various duplicate file installations. There are plans to fix this but they're
too invasive before release.
This patch relaxes the whitelist for sstate duplicate file detection to account
for this. Post-release, we re-enable stricter settings once do_package is not
machine specific.
(From OE-Core rev: c858259ce1881c6284f1fc2790c225c81e4a751e)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
This was added to allow detection of duplicate files being installed by sstate.
There is a much simpler way, just check if the file already exists. This
effectively uses the kernel VFS as the cache which is much more efficient.
This resolves a significant performance bottleneck (lock contention on a
single file) when running builds that are just being generated from sstate
cache files.
(From OE-Core rev: 603daf343ad3f18c8adb799e3625ae2a18d94f56)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
We need to split this variable before using it. Otherwise a single "/"
character in the list whitelists every overlapping sysroot file which
was not the intention making the whole thing useless.
We'll start seeing warnings about overlapping files now this is working
correctly after this patch.
(From OE-Core rev: 9e31c748327e92b809330f4ad7b6aaecb2edf559)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Appending to EXTRA_SSTATEMAPS is better than just hardcoding a value. Also
add a comment about why this is necessary.
(From OE-Core rev: d4f4a57b8d564d57256017d937ed2eabf94c36ae)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
I've not idea why this got left in but as per the comment, it needs fixing
and we shouldn't have hardcoded mappings like this. Lets remove it
and dynamically generate the data instead.
[YOCTO #3039]
(From OE-Core rev: 2df064ad46c1510fa8a401c22db4ab3278c3c807)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
This fixes builds where master.list doesn't already exist. This
change was meant to be part of the previous sstate commit but ended
up separated.
(From OE-Core rev: c2109b765b24a7ffe4781257ad3fe4641a3b2a49)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
We have a potential problem where two sstate packages try and touch the same
file. This adds code which will print a warning whenever this happens.
The implementation does but by maintaining a master file list and comparing
file accesses against this. There are a number of places we have duplicate
accesses which are harmless, mostly in the deploy directory so these
are whitelisted.
For now the code prints warnings, this could be strengthened in future to
become error messages. Whilst working on this code on and off over the past
few months various issues were uncovered, some serious.
[YOCTO #238]
(From OE-Core rev: 1f1b6f93d6b7aa8c9bd9bb5b1826997812e36932)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
It was pointed out we have a number of weird indentations in the python functions.
This patch cleans up 3, 7 and other weird indentations for the core bbclass files.
It also fixes some wierd (odd) shell function indentation which my searches picked up.
(From OE-Core rev: 8385d6d74624000d68814f4e3266d47bc8885942)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
This patch adds the lsb name and revision to the path used for sstate files.
This means that reuse of sstate files between different distributions is restricted
by default. The behaviour can be configured using mirror urls, for example:
SSTATE_MIRRORS = "file://Ubuntu-11.10/(.*) file://Ubuntu/\1 \n"
would map Ubuntu 11.10 to a more generic "Ubuntu" named sstate feed.
Usually, more modern distros have increased libc versions for example
so whilst more older native/cross sstate packages will usually work on newer
distros, the opposite is not true. This patch allows development of policy
to better handle this although no default policy is currently being used.
(From OE-Core rev: 9360fc318cec5c873b17f59b817dd6312c81ee27)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Currently all sstate files are placed into one directory. This does not scale and
causes a variety of filesystem issues. This patch adds a two character subdirectory
to the layout (based on the first two characters of the hash) so that files
can be split into several directories.
This should help performance of sstate in most cases.
(From OE-Core rev: 5d45c7b025f6635b2232d7bf92b7c1aba350396b)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
allarch sstate packages could be marked as machine or package_arch specific. This
change ensures they are not.
(From OE-Core rev: f3104240ad5bb542c339ee29b2672523ad3ae50c)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
There was a bug showing up where the crosssdk recipes were being installed as machine
specific in the sstate-control directory. This turned out to be due to the architecture
fields used by sstate being set incorrectly. The problem is that the crosssdk inherits
the cross class. It therefore needs to be listed in this if statement block before
the cross check, not after.
This should resolve some package architecture issue of crosssdk sstate files.
(From OE-Core rev: 79fe28e6033273f9632ca7549d1599d3fd1463ef)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Replace os.system with subprocess.call since the older function would
fail (more or less) silently if the executed program cannot be found
More info:
http://docs.python.org/library/subprocess.html#subprocess-replacements
[YOCTO #2454]
(From OE-Core rev: a07d03cc6f67c88feb9813ae7deb6e4a93552dfe)
Signed-off-by: Robert Yang <liezhi.yang@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Jason Wessel noticed that a package without any fixmepath entries would
generate a sed warning about no input files. This patch resolves that
by ensuring that an empty fixmepath file never gets written into the
sstate archive. Also we avoid a second message by only doing xargs if
we got input.
(From OE-Core rev: 326563d5a897ae2dba7cfd8d73579d3d979d72c8)
Signed-off-by: Mark Hatle <mark.hatle@windriver.com>
Signed-off-by: Jason Wessel <jason.wessel@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
In order to make sharing sstate files easier chmod them with 0664 permissions so that they are readable and writable by the user and any other members of their group.
(From OE-Core rev: dae9ad8a0ba0343e3083694cdcb20f0d02927ad0)
Signed-off-by: Joshua Lock <josh@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
The fixmepath file that is generated contains a list of all of the files
that need their paths fixed. In the previous version the fixmepath was
generated to include all of the files that sed may have changed. In the
new version, we first grep the files to see if they contain a path that
needs to be changed, only then do we perform the sed operation on those files.
This results in a modest performance increate in the creation of the sstate
file. The following numbers include the do_package and do_populate_sysroot
tasks on the perl recipe.
Before the change:
real 4m23.018s
user 1m57.067s
sys 1m33.327s
After the change:
real 4m13.083s
user 1m54.062s
sys 1m26.064s
However, a more significnt performance gain is felt during the
extraction/install of sstate cache files, as the fixmepaths file now has a
significantly smaller list of files to modify.
Before the change:
real 0m39.798s
user 0m11.158s
sys 0m12.642s
After the change:
real 0m25.511s
user 0m8.408s
sys 0m5.077s
(All numbers above were recorded with a cold filesystem cache on a machine
with 12 GB of ram.)
(From OE-Core rev: 46067264bedeff8248a2b2441733420fe6651f84)
Signed-off-by: Mark Hatle <mark.hatle@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
In a pathological case, lots of files to process, the sstate_installpkg
performance was very poor. It interated over each file and ran 3
individual sed commands per file. Changing this to keep iterating
but running only a single command took about 1/3 time time.
However, when looking at the corresponding sstate_hardcode_path
function, it was clear we could optimize this further.
Using the same encoding logic to specify only the minimumal sed
operation necessary, and using xargs to avoid the os.system call the
install step was able to be performed in 13% of the original time.
Example timing numbers for perl:
3m7s original code
1m20s single sed, but interating
0m26s using xargs and limited sed
(From OE-Core rev: d9f655753fbdc8cbd8e705577430fed4f23732b3)
Signed-off-by: Mark Hatle <mark.hatle@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Reluctantly revert this since it breaks the tar-native workaround we have
for old versions of tar :(
This reverts commit 01218e29f9.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
There is currently a race window when creating sstate packages since we don't
atomically write the files to SSTATE_DIR. This change ensures we do so by writing
to a temporary file and then doing an atomic move.
(From OE-Core rev: 52bf113e786a57123a9da98f64442afbc2f1471e)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
This resolves issues related to pigz-native when installing from sstate that people
have been seeing. It also gives us a way to solve issues like the gzip-native race
during sstate package creation covered in Yocto #1774.
(From OE-Core rev: 518dffe20178f5969dddccb17d6ab347afb72beb)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
* also fixes replacing paths for perl where cmd line was probably
too long for os.system(cmd) (it had 560410 characters because a lot of
files from sstate_scan_cmd).
* also print those 2 commands so we can find them in log.do_package
(From OE-Core rev: 94c52d68fc2ce258bcc5b0978ac73413480a1a93)
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
* e.g. mysql5 has mysql_config not mysql-config
(From OE-Core rev: aebdffc93ba8446bbd4e20263f0f75d4d4460c47)
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
The sstate_hardcode_path() function triggered large numbers of exec()
calls when processing packages with large numbers of file relocations
(e.g. perl). This patch optimises those calls into longer single commands
which make the code significantly more efficient.
This reduced the do_package time for perl by 2 minutes (from 4.75 minutes)
for me.
(From OE-Core rev: 4159ff761cf29a03aeb56c7686a4e7af02b52219)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
We process all files in the native/cross cases for finding and
fixing relocation issues. In the target case we've only processed
.la and binconfig files. Since there are other files which are
in need of this processing, this change allows recipes to specify
files that may be outside the normal set. This means hardcoded
paths that need to be fixmepathed to work correctly are handled
and addresses some sstate relocation bugs that have been seen.
Based on a patch from Saul Wold <sgw@linux.intel.com>
(From OE-Core rev: 6ffdcd9120b572fa41659029c3bda7bf00ebcb77)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Complete the bb.data.getVar/setVar replacements with accesses
directly to the data store object.
(From OE-Core rev: 2864ff6a4b3c3f9b3bbb6d2597243cc5d3715939)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Without this change we can end up looking for <stamp>.${MACHINE}
instead of the expected expanded value.
(From OE-Core rev: 9f743b5033177216fe0e1d3e43ba831f356df08e)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
sstate was being a little too ethusiastic about removing stamp files and
was removing stamp files for other machines when it shouldn't have been.
This patch teaches sstate about machine specific stamp extensions and
allows it to only remove the current task's stampfiles.
Based on a patch from Phil Blundell <philb@gnu.org> with some tweaks
from me.
(From OE-Core rev: 5e9488495401399d39fcb5012b86c313b6caca73)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
If several sstate packages are decompressing at the same time, they can
conflict over the "fixmepath". If two packages try to write to this at
the same time it results in an error. To avoid this we remove the file
once we've finished processing it.
(From OE-Core rev: c08d7093bf654ffedb155c5627972e9ecfeb0b60)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Currently its hard to figure out if/when sstate is checking for possible
packages to speed up builds. This patch adds level 2 debug output which
better indicates what files are being searched for an why.
[YOCTO #1259]
(From OE-Core rev: a752b23767189f2678367e47ca0b41f49c56a631)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
This patch removes the variables BASE_PACKAGE_ARCH, BASEPKG_HOST_SYS,
BASEPKG_TARGET_SYS and also removes the immediate assignments in
several core classes as these are no longer required.
This should make it clearer what some of the core variables do and
simplfy some overly complex and confusing class code.
(From OE-Core rev: d5521be2dcbaf213c140b0d12a4176380874426b)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
We need to check if we know of the task type, before we attempt
to process it. In order to reproduce the problem build with:
PACKAGE_CLASSES = "package_ipk"
Then change it to:
PACKAGE_CLASSES = "package_rpm"
Build again -- and then try bitbake -c cleansstate <recipe>
(From OE-Core rev: 8870ba9d261d3cacbe5d1219fdd95840c05ecf9e)
Signed-off-by: Mark Hatle <mark.hatle@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Distributions can then override this variable if needed
useful e.g. when building uclibc/eglibc both using
same tmpdir
(From OE-Core rev: 839e6e7f3717cc69cf8fc5b1c51bac63ec00655e)
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Fix [YOCTO #964]
A recent commit 25a6e5f9(sstate: use only unique set of SSTATETASK) breaks
the ordered mapping between SSTATETASKS and SSTATETASKNAMES. As a result,
in sstate_cleanall, the line
taskname = tasks[namemap.index(name)]
gets an incorrect result, and "bitbake -c cleanall" doesn't really remove
the files populalted by do_populate_sysroot.
(From OE-Core rev: 2f6505f0e795b6c8cad641a6918739c3faac1f99)
Signed-off-by: Dexuan Cui <dexuan.cui@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
* otherwise strange error like this:
ERROR: Logfile of failure stored in: /OE/shr-core/tmp/work/armv7a-oe-linux-gnueabi/libtool-cross-2.4-r1/temp/log.do_package_write_ipk.25551
Log data follows:
| ERROR: Package already staged (/OE/shr-core/tmp/sstate-control/manifest-nokia900-libtool-cross.deploy-ipk)?!
| ERROR: Function 'sstate_task_postfunc' failed
NOTE: package libtool-cross-2.4-r1: task do_package_write_ipk: Failed
ERROR: Task 11 (/OE/shr-core/openembedded-core/meta/recipes-devtools/libtool/libtool-cross_2.4.bb, do_package_write_ipk) failed with exit code '1'
is shown in this case with package_ipk twice in INHERIT
* Thanks to Richard for fix
(From OE-Core rev: f2fe5e840b8aa0558b5462ef2c7517b2f14ec2ea)
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
For safety, always delete the stamp files in sstate_clean regardless of
whether the manifest file exists or not.
(From OE-Core rev: f781c35da9a11eefdb06bda72ca89753df863efa)
Signed-off-by: Paul Eggleton <paul.eggleton@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Avoid deleting stamp files whose names contain the current task's name as a
substring. This will be especially important for example if do_package_write
is ever made an sstate task (as it would previously have deleted the stamps
here for do_package_write_ipk etc.)
(From OE-Core rev: ea743ea30e2289733d27979e8ec921648342da0e)
Signed-off-by: Paul Eggleton <paul.eggleton@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
When doing builds using sstate cache, there was no way to run
the equivalent of a pkg_postinst function. This is needed by
the SGML-related documentation recipes to properly update the
catalog files when new DTDs and stylesheets are installed.
SSTATEPOSTINSTFUNCS is a new variable you can set to function(s)
in your recipe to run after install is completed from sstate.
Thanks to Richard Purdie for suggesting this solution.
Signed-off-by: Scott Garman <scott.a.garman@intel.com>
Current sstate's fetch code doesn't reflect latest fetcher changes, so old
fetch style cause exception and fail silently.
[BUGID #708] got fixed.
Another issue is "import xxx" in python function from sstate.bbclass can only
sit in the head of the function, else have UnboundLocalError: "local variable
XXX referenced before assignment".
Signed-off-by: Zhai Edwin <edwin.zhai@intel.com>
This commit changes the sysroots path to be machine specific.
Changes includes:
1) STAGING_DIR_TARGET and STRAGING_DIR_HOST points to machine specific
paths.
2) task stamp files. Adding ${MACHINE} info into stamp files for
do_populate_sysroots and do_package tasks. Add a BB_STAMPTASK_BLACKLIST
to keep native, nativesdk, crosssdk, and cross-canadian stamp unchanged.
3) siteconfig path. Separate the site config path for different machines
to avoid one machine adopting the cache file of another machine.
4) sstate. Add machine name to sstate manifest file.
Change relocation code for sstate paths since sysroot is machine.
Keep native, nativesdk, crosssdk, and cross-canadian unchanged.
5) toolchain scripts. Change the environment path to point to machine
specific sysroots in toolchain scripts bbclass.
6) Relocate la files when populating to a different machine of the same
architecture.
7) Exclude STAGING_DIR_TARGET and STAGING_DIR_HOST parameter from sstate
siginfo since they contain ${MACHINE} information.
Signed-off-by: Dongxiao Xu <dongxiao.xu@intel.com>
Some steps (like gcc-cross-initial, gcc-cross-intermediate and eglibc-initial)
will install to new locations to avoid file overwriting.
Signed-off-by: Dexuan Cui <dexuan.cui@intel.com>
previously you could only build one set of packages for multiple machines:
MACHINE=foo bitbake task-base
MACHINE=bar bitbake task-base
would only create task-base packages for foo, but not for both foo and bar.
Doing MACHINE=bar bitbake task-base -c cleanall would remove the packages for foo.
The solution is to use MULTIMACH_ARCH as suggested by Richard Purdie.
Signed-off-by: Koen Kooi <koen@dominion.thruhere.net>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
This fixes the cleanall task to actually delete the sstage cache files. This
was not working correctly before due to attempting to use the current task's
hash in the package file name to be deleted, which of course is the hash of
the cleanall task when doing cleanall. The siginfo files are now deleted as
well.
Signed-off-by: Paul Eggleton <paul.eggleton@intel.com>
Remove the temp sstate-build-* directoies.
Take poky-image-minimal as an example, this saves about 17% disk space.
Signed-off-by: Dongxiao Xu <dongxiao.xu@intel.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
The fetcher has special handling for file:// mirror urls, being efficient and
just providing an updated path. Unfortunately the sstate fetching code wasn't
able to handle this. This patch detects this and injects a symlink to ensure
everything works. It also fixes some datastore references to be correct and
ensures the sstate download directory exists if it doesn't already.
Signed-off-by: Richard Purdie <rpurdie@linux.intel.com>
When installing a sstate package, directories tracked by plaindirs were being installed
to the incorrect location. With the current implementation this was limited to
the do_package task.
This patch ensures plaindirs tracked files are created in the correct location, fixing
the bug where these files would go missing.
Signed-off-by: Richard Purdie <rpurdie@linux.intel.com>
The directory list in the manifest file needs to be sorted such that child
directories are deleted first. Fortunately as the list is generated by walking
the directory tree achieving this is as simple as reversing the directory list.
Fixes [BUGID #269]
Signed-off-by: Joshua Lock <josh@linux.intel.com>
Adds the ability to impose intercept function call(s) after
sstate_install and before sstate_package in sstate_task_postfunc.
Signed-off-by: Jeff Polk <jeff.polk@windriver.com>
dir name in manifest is incorrect:
/media/disk/builds/master-arm/tmp/sysroots/usr/
/media/disk/builds/master-arm/tmp/sysroots/lib/
/media/disk/builds/master-arm/tmp/sysroots/include/
Also add one debug note when removing manifest entries
Signed-off-by: Kevin Tian <kevin.tian@intel.com>
there's a bug in creating manifest_pattern, which means actually with
new sstate do_cleanup hasn't worked yet
We can use PSTAGE2_MANFILEPREFIX easily for this purpose.
Signed-off-by: Kevin Tian <kevin.tian@intel.com>
Using TARGET_ARCH as the pacakge architecture for sstate archives lead to some
issues when using two machines with different BASE_PACKAGE_ARCH's but the same
TARGET_ARCH such as netbook and qemux86 machines.
Probable fix for [BUGID #223]
Signed-off-by: Joshua Lock <josh@linux.intel.com>
Robert Yang
Ming Liu
Richard Purdie
Roy Li
Joe MacDonald
Christopher Larson
Jackie Huang
Phil Blundell
Paul Eggleton
Andreas Müller
Jeff Polk
Leonardo Sandoval
Laurentiu Palcu
Martin Ertsaas
Hongxu Jia
Saul Wold
Martin Jansa
Enrico Scholz
Ross Burton
Mark Hatle
Joshua Lock
Khem Raj
Dexuan Cui
Scott Garman
Zhai Edwin
Yu Ke
Dongxiao Xu
Koen Kooi
Paul Eggleton
Richard Purdie
Kevin Tian