Significantly reduces time spent processing those targets and should
also silence some log clutter which could confuse buildbot
Signed-off-by: Felix Fietkau <nbd@nbd.name>
Otherwise ipkg packages may wrongly take on alternatives specs of
another package sharing the same package Makefile
Fixes FS#753
Signed-off-by: Yousong Zhou <yszhou4tech@gmail.com>
It's a list of specs of the following form seprated by commas to
describe alternatives provided by this package
<prio>:<path>:<altpath>
<path> will be a symbolic link to <altpath> of the highest <prio>
Signed-off-by: Yousong Zhou <yszhou4tech@gmail.com>
When running "make menuconfig" with symlinked .config (e.g. to
env/.config) it renames symlink to .config.old, creates new .config file
and writes updated configuration here.
This breaks the desired workflow when changes in the configuration could
be checked using "scripts/env diff" and commited with
"scripts/env save". Since the env/.config file is not updated.
Fix this issue by exporting KCONFIG_OVERWRITECONFIG=1, which forces
mconf to overwrite the .config content, instead of renaming it and
creating a new file. This variable is set only if .config is a symlink,
otherwise the variable is not exported and the old behaviour is
preserved.
Signed-off-by: Sergey Ryazanov <ryazanov.s.a@gmail.com>
Not sure since when the issue emerged, but according to the current doc of gcc
and as, armv8-a is intended as argument of -march
The change will affect at the moment arm64 and layerscape/64b
Below is the relevant error messages when building toolchain
Assembler messages:
Error: unknown cpu `armv8-a'
Error: unrecognized option -mcpu=armv8-a
/home/yousong/git-repo/lede-project/lede/build_dir/toolchain-aarch64_armv8-a_gcc-5.4.0_musl/gcc-5.4.0/libgcc/libgcc2.c:1:0: error: unknown value 'armv8-a' for -mcpu
Signed-off-by: Yousong Zhou <yszhou4tech@gmail.com>
Recent attempts to use it have shown that it does not work properly
except for a few undocumented cases. It's better to remove this now to
avoid having more people fall into the same trap
Signed-off-by: Felix Fietkau <nbd@nbd.name>
Keeping it in base-files was resulting in adding it to the base-files
package. This file is meant to be included manually for initramfs
images only.
Signed-off-by: Rafał Miłecki <rafal@milecki.pl>
If someone creates a target and indicates a CPU_TYPE, but there's
no corresponding support for that CPU_TYPE's flags in include/target.mk
then that should probably be indicated rather than silently ignored.
Signed-off-by: Philip Prindeville <philipp@redfish-solutions.com>
quilt.mk needs to be included first, to ensure that STAMP_PREPARED does
not include the hash if quilt is used.
Signed-off-by: Felix Fietkau <nbd@nbd.name>
This can be used to indicate that a target does not support the optional mips16
extension even when it is a mips32r2 or later CPU.
This will generate a separate toolchain and a separate package folder,
e.g. mips_24kc_nomips16
Signed-off-by: Hauke Mehrtens <hauke.mehrtens@intel.com>
KERNELRELEASE contains a $(shell) call which is evaluated over and over
again.
The call to checksyscalls.sh is unnecessary for LEDE and also takes a
few seconds to complete.
Signed-off-by: Felix Fietkau <nbd@nbd.name>
This patch makes specifying NETGEAR_REGION optional, in which case
mkchkimage will default to region 1 (WW).
Signed-off-by: Joseph C. Lehner <joseph.c.lehner@gmail.com>
Bump kernel from 4.4.50 to 4.4.52
Refresh patches
Compile tested all 4.4. targets
Run tested: ar71xx Archer C7 v2
Signed-off-by: Kevin Darbyshire-Bryant <kevin@darbyshire-bryant.me.uk>
Instead of using a separate .clean stamp file, remove the install info
file on compile, then append the install package list afterwards
Signed-off-by: Felix Fietkau <nbd@nbd.name>
Bump kernel from 4.4.49 to 4.4.50
Compile tested: All targets
Run tested: ar71xx Archer C7 v2
Signed-off-by: Kevin Darbyshire-Bryant <kevin@darbyshire-bryant.me.uk>
The @ sign in front of the "mv" command was significantly suppressing
output to stdout. When reviewing the make/build logs it was tricking
me a whole lot and it mad me lose time. Removing the @ sign will get
stdout and logs right about what happened when.
Signed-off-by: Thomas Reifferscheid <thomas@reifferscheid.org>
cmake checks the build system and its variables on its own to detect if
the makefiles need to be regenerated.
Unfortunately this can invalidate overrides passed in the
Build/Configure step. On non-Linux systems this breaks the build when
switching between targets of the same package architecture.
Fix this by forcibly disabling the build system check and relying on the
LEDE build system to take care of these things
Signed-off-by: Felix Fietkau <nbd@nbd.name>
When running "make kernel_menuconfig" in a clean tree, it fails with:
make[1]: *** No rule to make target 'tools/quilt/install'. Stop.
Replacing the dependency with 'tools/quilt/compile' fixes the issue (quilt
and all its prerequisites will be built, and quilt will be installed in
staging_dir).
Signed-off-by: Baptiste Jonglez <git@bitsofnetworks.org>
-mXXX option is deprecated already in arc-2016.03 toolchain
and removed completely starting from arc-2016.09.
Direct replacement is -mcpu=XXX which is already supported by
arc-2016.03 used today in Lede.
With that change we'll be ready for ARC toolchain update still
keeping everything working with current tools.
Signed-off-by: Alexey Brodkin <Alexey.Brodkin@synopsys.com>
Cc: John Crispin <john@phrozen.org>
Refresh patches for all targets that support kernel 4.4.
Compile-tested on all targets that use kernel 4.4 and aren't marked
broken, except arc770 and arch38 due to broken toolchain.
Runtime-tested on ar71xx, octeon, ramips and x86/64.
Signed-off-by: Stijn Tintel <stijn@linux-ipv6.be>
This patch moves the fakeroot code required by some devices to
`image-commands.mk`.
Create the fakeroot on the fly by using the undocumented -s (skip copy)
parameter of mkimage.
Signed-off-by: Joseph C. Lehner <joseph.c.lehner@gmail.com>
[remove unused NETGEAR_KERNEL_MAGIC, remove workarounds to have a dummy
rootfs for mkimage]
Signed-off-by: Mathias Kresin <dev@kresin.me>
Refreshed patches for all supported targets.
Compile-tested on ar71xx, cns3xxx, imx6, mt7621, oxnas and x86/64.
Run-tested on ar71xx, cns3xxx, imx6 and mt7621.
Tested-by: Stijn Segers <francesco.borromini@inventati.org>
Signed-off-by: Koen Vandeputte <koen.vandeputte@ncentric.com>
Since the main stamp file depends on the _check stampfile, the _check
stampfile needs to be created on the first run as well.
Fixes spurious rebuilds with CONFIG_AUTOREBUILD=y
Signed-off-by: Felix Fietkau <nbd@nbd.name>
some of Buffalo DHP series use slightly different trx magic, buffalo-enc,
buffalo-tag, and factory image begin with 'bgn'.
this patch adds support for building those images.
Signed-off-by: FUKAUMI Naoki <naobsd@gmail.com>
Bump kernel to 4.4.44. Compile-tested on ar71xx, ramips/mt7621 and x86/64.
.44 has been run-tested on the 17.01 branch here on ar71xx and mt7621.
Signed-off-by: Stijn Segers <francesco.borromini@inventati.org>
If the base-files package is not selected, we will fail executing the
very first postinst script:
make[3]: Leaving directory `/local/users/fainelli/openwrt/trunk'
cp -fpR
/local/users/fainelli/openwrt/trunk/build_dir/target-arm_xscale_musl-1.1.15_eabi/root-orion
/local/users/fainelli/openwrt/trunk/build_dir/target-arm_xscale_musl-1.1.15_eabi/root.orig-orion
./usr/lib/opkg/info/busybox.postinst: line 3:
/local/users/fainelli/openwrt/trunk/build_dir/target-arm_xscale_musl-1.1.15_eabi/root-orion/lib/functions.sh:
No such file or directory
./usr/lib/opkg/info/busybox.postinst: line 4: default_postinst: command
not found
postinst script ./usr/lib/opkg/info/busybox.postinst has failed with
exit code 127
make[2]: *** [package/install] Error 1
Check for the existence of lib/functions.sh, and if it does not exist,
just bail out gracefully.
Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
Require-User is handled by /etc/uci-defaults/13_fix_group_user on first
boot, so we need to keep these when removing all opkg data with
CONFIG_CLEAN_IPKG.
Signed-off-by: Matthias Schiffer <mschiffer@universe-factory.net>
According to some reports, -march=pentium-mmx is a better choice for
older Geode CPUs than -march=geode anyway.
Bump the minimum architecture of the legacy target from i486 to
pentium-mmx. Anything older is not worth supporting anyway.
Signed-off-by: Felix Fietkau <nbd@nbd.name>
This avoids repeatedly unpacking and rebuilding packages that are
failing the build. Re-running the failing step should be much faster.
Signed-off-by: Felix Fietkau <nbd@nbd.name>
Calling the clean target removes all .ipk files and un-stages the
package. Add a new target just for clearing the build dir and call that
one instead of the full clean target
Signed-off-by: Felix Fietkau <nbd@nbd.name>
This is used to save space on buildbot instances.
If any part of a package needs to be rebuild, the whole package is
rebuilt from scratch. Stamp files are preserved to allow dependency
checks to work
Signed-off-by: Felix Fietkau <nbd@nbd.name>
This makes the build slightly more efficient by avoiding the need to
re-stage packages on every full build run.
It is also necessary for the upcoming CONFIG_AUTOREMOVE feature
Signed-off-by: Felix Fietkau <nbd@nbd.name>
This makes the build slightly more efficient by avoiding the need to
re-stage packages on every full build run.
It is also necessary for the upcoming CONFIG_AUTOREMOVE feature
Signed-off-by: Felix Fietkau <nbd@nbd.name>
The make target does not rewrite the file if the contents didn't
change. This causes make to always consider the target old.
Signed-off-by: Felix Fietkau <nbd@nbd.name>
Before SUBDIR was set to $(PATCHVER) which may
or may not include the minor version number of
the linux kernel version. Usually it doesn't.
So the git-clone'd linux kernel was packed without
the minor version number taken into account, which
broke further processing, as it expected the
extracted dir being named linux-$(LINUX_VERSION)
(=with minor version) rather than linux-$(PATCHVER)
(=without minor version).
Changing SUBDIR to $(LINUX_VERSION) creates
consistent behaviour here.
Signed-off-by: Mirko Vogt <mirko-openwrt@nanl.de>
Using a single host package staging dir (and build dir) significantly speeds up
builds when multiple targets are built in succession, especially for large host
packages like NodeJS.
$(STAGING_DIR)/host is kept in addition to $(STAGING_DIR_HOSTPKG) in most
places; it is still used as destination for host files in Build/InstallDev.
Signed-off-by: Matthias Schiffer <mschiffer@universe-factory.net>
Currently system log is always included as a part of ubox. Add logd as a
seperate package and add it to default packages list.
Signed-off-by: Andrej Vlasic <andrej.vlasic@sartura.hr>
Signed-off-by: Luka Perkov <luka.perkov@sartura.hr>
Make sure binaries install to STAGING_DIR_HOSTPKG are still found when
this variable is eventually modified.
Signed-off-by: Matthias Schiffer <mschiffer@universe-factory.net>
This reduces the amount of hacks in the makefile code.
Remove the apm821xx code to do the same - it was broken and left both
compressed and uncompressed images in $(BIN_DIR)
Signed-off-by: Felix Fietkau <nbd@nbd.name>
The generated 'its' is passed to mkimage which expects linux arch
strings rather than the full arch (e.g. mips not mipsel).
It currently works in some cases where LINUX_KARCH == ARCH but
otherwise you get an unknown arch build error.
Signed-off-by: Ian Pozella <Ian.Pozella@imgtec.com>
Signed-off-by: Felix Fietkau <nbd@nbd.name>
The removal of the ".+Package" pattern in scan.mk also caused the build
system to skip over Makefiles defining only kmods. Adjust the grep pattern
to consider packages with "call KernelPackage" signatures as well.
Signed-off-by: Jo-Philipp Wich <jo@mein.io>
Commit af0b91c "allow scan.mk to find python packages introduced in [8639]"
added some special casing to scan.mk to accomodate some nonstandard python
packages.
Nowadays this pattern is not needed anymore and produces false positives
when using the LEDE source repository as feed within the SDK since the
metadata scanning wrongly picks up target/imagebuilder/Makefile as package,
leading to an "ERROR: please fix feeds/base/target/imagebuilder/Makefile"
message.
Remove the now uneeded pattern to fix such stray errors during metadata
scanning.
Signed-off-by: Jo-Philipp Wich <jo@mein.io>
Instead of passing HOST_CONFIGURE_VARS as arguments to the configure script,
pass it as environment variables which brings the logic in line with the
behaviour of package-defaults.mk.
The change is needed since passing environment variables as configure
parameters only works with GNU autoconf which evaluates command line arguments
looking like variable assignments. Doing the same with non-autoconf configure
scripts is not guaranteed to work since such scripts might terminate due to
unknown argument errors.
One example case is the cmake configure script which bails out when called
as "./configure LDFLAGS=..." but not when called as "LDFLAGS=... ./configure".
Also change the SHELL override to CONFIG_SHELL in the default
HOST_CONFIGURE_VARS as the former is not properly propagated through the
various GNU configure invocations since it gets lost when configure re-
executes itself.
A prior attempt to change the variable placement had to be reverted due to
the missing SHELL -> CONFIG_SHELL change, leading to misgenerated libtool
executables in various packages.
Signed-off-by: Jo-Philipp Wich <jo@mein.io>
This reverts commit 8395b63aac616f72fd835c59240fc2a4a6b28106.
Various host builds currently rely on the broken behaviour of
HOST_CONFIGURE_VARS so roll back to the previous state.
Signed-off-by: Jo-Philipp Wich <jo@mein.io>
Ensure that HOST_CONFIGURE_VARS are set before the actual configure command
instead of passing them as configure command arguments.
This change brings host-build.mk in line with package-defaults.mk and makes
host configure environment variables work as expected.
Signed-off-by: Jo-Philipp Wich <jo@mein.io>
This will be used to simplify the build system code for checking hashes.
Instead of using various variants of md5sum / openssl, use one simple
utility for all of them
Signed-off-by: Felix Fietkau <nbd@nbd.name>
Add a call to KernelPackage/$(1)/$(BOARD)/$(SUBTARGET) to the
KernelPackage macro. This allows to add kernel packages for x86/64,
without breaking x86. It's not possible to do this with BOARD, as
BOARD=x86 for x86_64.
Signed-off-by: Stijn Tintel <stijn@linux-ipv6.be>
Looks like this was meant to workaround some limitations with
non-GNU tar variants (like BSD-tar which are present on Mac os BSD hosts).
Though, I cannot find any use of that `+s` option that's mentioned
in the comment.
Last hash of this I found was 24faf55360
In my case, it now this fails for `python-setuptools` on Mac OS X (the host-build with):
```
trapret 2 tar -C <home-dir>/work/sources-work/lede/build_dir/target-i386_pentium4_musl-1.1.15/python-setuptools-27.2.0 --strip-components=1 -xzf <home-dir>/work/sources-work/lede/dl/setuptools-27.2.0.tar.gz
bash: trapret: command not found
```
So, I was thinking maybe it's time to remove this workaround (9 years later).
I could also fix the `python-setuptools` host build. If that's more preferred.
[ Btw, I just recently transitioned to a Mac machine for dev-ing,
so a lot of (this Mac) stuff I'm finding out is new to me too. ]
Signed-off-by: Alexandru Ardelean <ardeleanalex@gmail.com>
This makes it easier to unify versioning of git based package downloads.
PKG_SOURCE_DATE along with an 8-character abbreviation of the git hash
is used as PKG_VERSION, PKG_RELEASE should be used like normal packages.
Signed-off-by: Felix Fietkau <nbd@nbd.name>
This adds support for install-overlay define. When used in package it
allows installing files to a special directory that gets copied to the
root when installing it.
It allows overwriting files provided by other packages.
Signed-off-by: Rafał Miłecki <rafal@milecki.pl>
Bump & refresh patches for all 4.4 targets.
Compile & run tested: ar71xx Archer C7 v2
Signed-off-by: Kevin Darbyshire-Bryant <kevin@darbyshire-bryant.me.uk>
This will attempt to automatically fix common mistakes like using MD5
instead of SHA256, using the MD5SUM variable instead of HASH, or even a
missing mirror file hash.
Signed-off-by: Felix Fietkau <nbd@nbd.name>
This is intended to be used for a wide array of package sanity checks.
The first check that is implemented is for the hash of downloaded files.
It checks:
- Missing hash
- Use of SHA256 instead of MD5
- dl/<file> hash not matching hash in makefile
- deprecated MD5SUM variable
The deprecated MD5SUM variable check is skipped for feeds/ until OpenWrt
is updated as well
Signed-off-by: Felix Fietkau <nbd@nbd.name>
Since we've switched to preferring SHA256 over MD5, the old variable
name is misleading. Packages using the old name remain compatible.
Signed-off-by: Felix Fietkau <nbd@nbd.name>
Calling a build ##.##-CURRENT might mislead users into thinking that this
build is the most current release of a branch.
Signed-off-by: Jo-Philipp Wich <jo@mein.io>
This will avoid loading it in the default configuration, which reduces
image size a bit, and (more importantly) improves performance by
avoiding some unnecessary netfilter hooks
Signed-off-by: Felix Fietkau <nbd@nbd.name>
The xt_id match was used by the firewall3 package to track its own rules but
the approach has been changed to use xt_comment instead now, so we can drop
this nonstandard extension.
Signed-off-by: Jo-Philipp Wich <jo@mein.io>
Now that the VERSION_NUMBER variable holds the human friendly name and not
the commit ID anymore, we need to support adding the revision ID as well.
Introduce a new config variable CONFIG_VERSION_CODE_FILENAMES which, if set,
causes the resulting file names to contain a commit ID designation as printed
by scripts/getver.sh.
Also sanitize the input variables to ensure that the resulting strings are
lowercased and no not contain spaces.
Signed-off-by: Jo-Philipp Wich <jo@mein.io>
Utilize the existing git download logic from include/download.mk and migrate
the kernel download over to it. This avoids repeatedly cloning kernel sources
after a make target/linux/clean for instance.
Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: Felix Fietkau <nbd@nbd.name> [fix build error]
This is going to be used to migrate the hand rolled git clone for the kernel
into using the git download method. The kernel uses custom options that we may
have to pass down.
Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
Move the revision info to the VERSION_CODE variable and default VERSION_NUMBER
to CURRENT for master branch builds.
Also introduce a new menuconfig option CONFIG_VERSION_CODE which allows users
to override the revision value put into VERSION_CODE and adjust the template
files used by the base-files package to accomodate for the changed semantics.
While we're at it, also adjust the various URLs to match the current web site.
After this commit, the relevent files will look like the examples given below:
# cat /etc/openwrt_version
r2398+1
# cat /etc/openwrt_release
DISTRIB_ID='LEDE'
DISTRIB_RELEASE='CURRENT'
DISTRIB_REVISION='r2398+1'
DISTRIB_CODENAME='reboot'
DISTRIB_TARGET='x86/64'
DISTRIB_DESCRIPTION='LEDE Reboot CURRENT r2398+1'
DISTRIB_TAINTS='no-all override'
# cat /usr/lib/os-release
NAME="LEDE"
VERSION="CURRENT, Reboot"
ID="lede"
ID_LIKE="lede openwrt"
PRETTY_NAME="LEDE Reboot CURRENT"
VERSION_ID="current"
HOME_URL="http://lede-project.org/"
BUG_URL="http://bugs.lede-project.org/"
SUPPORT_URL="http://forum.lede-project.org/"
BUILD_ID="r2398+1"
LEDE_BOARD="x86/64"
LEDE_TAINTS="no-all override"
LEDE_DEVICE_MANUFACTURER="LEDE"
LEDE_DEVICE_MANUFACTURER_URL="http://lede-project.org/"
LEDE_DEVICE_PRODUCT="Generic"
LEDE_DEVICE_REVISION="v0"
LEDE_RELEASE="LEDE Reboot CURRENT r2398+1"
On a release branch, those files would look like:
# cat /etc/openwrt_version
r2399
# cat /etc/openwrt_release
DISTRIB_ID='LEDE'
DISTRIB_RELEASE='16.12-CURRENT'
DISTRIB_REVISION='r2399'
DISTRIB_CODENAME='test_release'
DISTRIB_TARGET='x86/64'
DISTRIB_DESCRIPTION='LEDE Test Release 16.12-CURRENT r2399'
DISTRIB_TAINTS='no-all override'
# cat /usr/lib/os-release
NAME="LEDE"
VERSION="16.12-CURRENT, Test Release"
ID="lede"
ID_LIKE="lede openwrt"
PRETTY_NAME="LEDE Test Release 16.12-CURRENT"
VERSION_ID="16.12-current"
HOME_URL="http://lede-project.org/"
BUG_URL="http://bugs.lede-project.org/"
SUPPORT_URL="http://forum.lede-project.org/"
BUILD_ID="r2399"
LEDE_BOARD="x86/64"
LEDE_TAINTS="no-all override"
LEDE_DEVICE_MANUFACTURER="LEDE"
LEDE_DEVICE_MANUFACTURER_URL="http://lede-project.org/"
LEDE_DEVICE_PRODUCT="Generic"
LEDE_DEVICE_REVISION="v0"
LEDE_RELEASE="LEDE Test Release 16.12-CURRENT r2399"
On a release tag, those files would look like:
# cat /etc/openwrt_version
r2500
# cat /etc/openwrt_release
DISTRIB_ID='LEDE'
DISTRIB_RELEASE='17.02.1'
DISTRIB_REVISION='r2500'
DISTRIB_CODENAME='mighty_unicorn'
DISTRIB_TARGET='x86/64'
DISTRIB_DESCRIPTION='LEDE Mighty Unicorn 17.02.1 r2500'
DISTRIB_TAINTS='no-all override'
# cat /usr/lib/os-release
NAME="LEDE"
VERSION="17.02.1, Mighty Unicorn"
ID="lede"
ID_LIKE="lede openwrt"
PRETTY_NAME="LEDE Mighty Unicorn 17.02.1"
VERSION_ID="17.02.1"
HOME_URL="http://lede-project.org/"
BUG_URL="http://bugs.lede-project.org/"
SUPPORT_URL="http://forum.lede-project.org/"
BUILD_ID="r2500"
LEDE_BOARD="x86/64"
LEDE_TAINTS="no-all override"
LEDE_DEVICE_MANUFACTURER="LEDE"
LEDE_DEVICE_MANUFACTURER_URL="http://lede-project.org/"
LEDE_DEVICE_PRODUCT="Generic"
LEDE_DEVICE_REVISION="v0"
LEDE_RELEASE="LEDE Mighty Unicorn 17.02.1 r2500"
Signed-off-by: Jo-Philipp Wich <jo@mein.io>
Acked-by: Felix Fietkau <nbd@nbd.name>
The SDK Makefile still trys to copy the docs folder which was removed
with 882f4d2d63. This causes an SDK build
error.
All other removals are just cleanup.
Signed-off-by: Mathias Kresin <dev@kresin.me>
Refresh patches on all 4.4 supported platforms.
077-0005-bgmac-stop-clearing-DMA-receive-control-register-rig.patch
removed as now upstream.
Compile & run tested: ar71xx - Archer C7 v2
Signed-off-by: Kevin Darbyshire-Bryant <kevin@darbyshire-bryant.me.uk>
Targets like malta can have no patches/ directory available and this
commit tries quash "no such file or directory" messages from `find`
Signed-off-by: Yousong Zhou <yszhou4tech@gmail.com>
Refresh patches for all targets that support kernel 4.4.
Compile-tested on all targets that use kernel 4.4 and aren't marked broken.
Runtime-tested on ar71xx, octeon.
Signed-off-by: Stijn Tintel <stijn@linux-ipv6.be>
If a package nonshared status is changed, a stale .ipk file might still
be present in the old package directory. Remove the .ipk file from all
package directories when building a new one (or explicitly running
clean)
Signed-off-by: Felix Fietkau <nbd@nbd.name>
Specifying a mtune option with cortex-a53 is also valid for an aarch64
toolchain
Fixes: SVN 48964
Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
This patch bumps the 4.4 kernel from .28 to .30 and refreshes the patches.
Compile-tested on ar71xx, x86/64, ramips/mt7621, brcm47xx and kirkwood.
Run-tested on ar71xx & ramips/mt7621, brcm47xx and kirkwood (last two confirmed
by P. Wassi).
Signed-off-by: Stijn Segers <francesco.borromini@inventati.org>
Right now the $(PKG_INSTALL_STAMP) files are only written if a package is
selected as <*> but never deleted or emptied if the corresponding package
is getting deselected.
For ordinary packages this usually is no problem as the package/install
recipe performs its own check for enabled packages when assembling the
list of install stamp files to consider, but this logic might fail under
certain circumstances for packages providing multiple build variants.
In case of a multi-variant package, the buildroot first checks if any
of the variants is enabled, then resolves all variants of the common
source package and finally processes the corresponding .install stamp
files of all variants, relying on the assumption that only the selected
.install stamp file exists.
When an initially selected variant is getting deselected or changed from
<*> to <m> and another variant is marked as <*> instead, the .install
stamp file of the deselected variant remains unchanged and a second
.install stamp file for the newly selected variant is getting created,
causing the package/install recipe to pick up two .install stamps with
conflicting variants, leading to opkg file clashes.
This issue happens for example if package "ip" is set to <m> and package
"ip-full" to <*> - the install command will eventually fail with:
* check_conflicts_for: The following packages conflict with ip:
* check_conflicts_for: ip-full *
* opkg_install_cmd: Cannot install package ip.
In order to fix the problem, always process the removal requests or the
.install stamp files, even for deselected packages but only write the
package base name into the stamp file if the corresponding package is
marked as builtin.
Signed-off-by: Jo-Philipp Wich <jo@mein.io>
Instead of hardcoding $(STAGING_DIR)/host, use the new $(STAGING_DIR_HOSTPKG)
variable to refer to the directory.
Signed-off-by: Jo-Philipp Wich <jo@mein.io>
Refresh patches for all targets that support kernel 4.4.
compile/run-tested on ar71xx, brcm47xx, kirkwood.
Signed-off-by: Paul Wassi <p.wassi@gmx.at>
Tested-by: Koen Vandeputte <koen.vandeputte@ncentric.com>
There is very little practical use to limit the number of available inodes on
an ext4 filesystem and the make_ext4fs utility is able to calculate useful
defaults by itself.
Drop the option to make resulting ext4 filesystems more flexible by default.
Signed-off-by: Jo-Philipp Wich <jo@mein.io>
Acked-by: Michael Heimpold <mhei@heimpold.de>
