When using ccache TARGET_CC is something like
ccache gcc
This causes problems in the pv build which attempts to
override LD because the command ends up being
make [...] LD=/tools/ccache /tools/gcc LDFLAGS="[...]
As a result, during the build phase it attempts to build
/tools/gcc which succeeds by doing nothing:
make[1]: Nothing to be done for `/tools/gcc'.
and during the install phase you get the real build which
errors out on the LD error this snippet was attempting to
fix:
ld -r -o src/library.o src/library/getopt.o \
src/library/gettext.o
ld: src/library/gettext.o: Relocations in generic ELF \
(EM: 40)
src/library/gettext.o: error adding symbols: File in \
wrong format
Signed-off-by: Andrew Ruder <andrew.ruder@elecsyscorp.com>
Acked-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
As reported by Thomas Petazzoni, the uclibc 0.9.31 build fails for avr32:
In file included from /opt/br-avr32-full-2013.11-rc1/usr/avr32-buildroot-linux-uclibc/sysroot/usr/include/linux/kernel.h:4,
from /opt/br-avr32-full-2013.11-rc1/usr/avr32-buildroot-linux-uclibc/sysroot/usr/include/linux/netlink.h:4,
from /opt/br-avr32-full-2013.11-rc1/usr/avr32-buildroot-linux-uclibc/sysroot/usr/include/linux/rtnetlink.h:5,
from libc/inet/netlinkaccess.h:27,
from libc/inet/if_index.c:36:
/opt/br-avr32-full-2013.11-rc1/usr/avr32-buildroot-linux-uclibc/sysroot/usr/include/linux/sysinfo.h:8: error: expected specifier-qualifier-list before '__kernel_long_t'
make[1]: *** [libc/inet/if_index.os] Error 1
make[1]: Leaving directory `/opt/toolchain-build/build/uclibc-0.9.31.1'
The problem is reported at:
https://lkml.org/lkml/2013/5/18/1
The offending kernel commit is:
http://git.kernel.org/cgit/linux/kernel/git/stable/linux-stable.git/commit/?id=ccdfcc398594
The fix is to patch uclibc 0.9.31 to add the missing kernel data types. The patch
will only be generated for avr32, since uclibc 0.9.31 is not available in Buildroot
for any other architecture.
Signed-off-by: Simon Dawson <spdawson@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
As reported by Cassiano Martin in bug #6692 if host == target the nano
package can pick up the host libmagic and break.
So add a check to see if the file package is enabled and use it,
otherwise just disable libmagic support.
Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
With this, we can trash our probability patch, it's now upstream.
Refresh a few other patches.
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Patches 02-cpp-comments-to-c-comments.patch changes C++-style comments
into C-style comments.
This is unneeded, since gcc accepts C++-style comments in C code anyway.
Ditch that patch, that's one less we have to handle when updating from
upstream.
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
The procedure to update our copy of kconfig was mising copying a file.
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
In case a menu has comment without letters/numbers (eg. characters
matching the regexp '^[^[:alpha:][:digit:]]+$', for example - or *),
hitting space will cycle through those comments, rather than
selecting/deselecting the currently-highlighted option.
This is the behaviour of hitting any letter/digit: jump to the next
option which prompt starts with that letter. The only letters that
do not behave as such are 'y' 'm' and 'n'. Prompts that start with
one of those three letters are instead matched on the first letter
that is not 'y', 'm' or 'n'.
Fix that by treating 'space' as we treat y/m/n, ie. as an action key,
not as shortcut to jump to prompt.
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Cc: Peter Korsgaard <jacmet@uclibc.org>
Cc: Samuel Martin <s.martin49@gmail.com>
Cc: Thomas De Schampheleire <patrickdepinguin@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
As pointed by Thomas P. kernel 3.12 oopses when loading/using the
emulated network.
Seems 3.12 broke versatile for qemu like in the past, only in a more
subtle way this time that escaped my automated qemu builds/tests.
Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Offer a little bit more visibility to the companies who sponsored us,
either by sponsoring the developer days, or development
boards. Hopefully this will encourage other companies to do the same :)
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Acked-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
[Peter: drop 'the' before FOSDEM]
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Acked-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Allow user to supply their own step-hooks by passing a variable
on the make command-line:
make BR2_INSTRUMENTATION_SCRIPTS=/path/to/my/script
This can be useful to run site-specific actions at each step of the
build process, such as logging installed, removed or modified files,
do sanity checks on installed files...
It is possible to call more than one script, by passing a space-separated
lists of scripts to call.
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Thomas De Schampheleire <patrickdepinguin@gmail.com>
Reviewed-by: Thomas De Schampheleire <thomas.de.schampheleire@gmail.com>
Reviewed-by: Samuel Martin <s.martin49@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
The timing information is stored in the file $(O)/build-time.log
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Reviewed-by: Thomas De Schampheleire <thomas.de.schampheleire@gmail.com>
Reviewed-by: Samuel Martin <s.martin49@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
This hooks will let us instrument the build process in many ways:
- log current step to see what broke
- time each step to see what is worth optimising
- sanity-check installed files (rpath, overwritten files...)
- call user-provided script
- ...
The steps are coarse-grain, and all have a 'start' and a 'end' hooks.
Here is the list of available steps (8 total):
- extract
- patch
- configure
- build
- install-host
- install-staging
- install-image
- install-target
The download, clean and uninstall steps are not instrumented on purpose.
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Thomas De Schampheleire <patrickdepinguin@gmail.com>
Reviewed-by: Thomas De Schampheleire <thomas.de.schampheleire@gmail.com>
Reviewed-by: Samuel Martin <s.martin49@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Also export BUILD_DIR for post-{build,images} hooks, so they do have
a place to store generated files.
Note: this will be more einteresting for the instrumentation of steps,
to come in a later patch.
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Thomas De Schampheleire <patrickdepinguin@gmail.com>
Reviewed-by: Thomas De Schampheleire <thomas.de.schampheleire@gmail.com>
Reviewed-by: Samuel Martin <s.martin49@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Due to some tricky make behavior, the license texts of host packages that
did not provide an explicit HOST_FOO_LICENSE_FILES definition was not saved.
The problem is that it is not straightforward to use a variable
defined/updated inside an evaluated block as input to a foreach statement.
If you try to use $(FOO) then only the original value of FOO is used for
foreach, any update inside the block is ignored. However, if you use
$$(FOO), the entire contents of FOO (typically a list of items) is passed
as one item to foreach, thus causing just one iteration instead of several.
>From Arnout Vandecapelle's explanation:
Any variable referenced with a single $ inside the inner-generic-package
macro is expanded before the resulting contents are eval'ed. Therefore, it
is not possible to refer to variables defined by the inner-generic-package
macro from within a single-$ function call.
To fix the problem, one should defer the evaluation of the entire block
using double dollar signs.
Additionally, a few empty lines have been added to the legal-info-foo block
for clarity.
Signed-off-by: Thomas De Schampheleire <thomas.de.schampheleire@gmail.com>
Acked-by: Luca Ceresoli <luca@lucaceresoli.net>
Tested-by: Luca Ceresoli <luca@lucaceresoli.net>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
While the generic package handler checks for a directory with patches
before starting apply-patches.sh, this is not the case for gcc: the
script is called, even if there is no directory with patches. This results
into a build failure, as apply-patches exits with error code 1 if the
directory doesn't exist.
Signed-off-by: Mischa Jonker <mjonker@synopsys.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
(e)glibc doesn't support a fully statically linked userspace. Even a
basic program such as Busybox fails to do authentication due to glibc
loading some libraries dynamically. Therefore, we disable the
possibility of using a (e)glibc toolchain when
BR2_PREFER_STATIC_LIB=y.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
(e)glibc doesn't support a fully statically linked userspace. Even a
basic program such as Busybox fails to do authentication due to glibc
loading some libraries dynamically. Therefore, we disable the
possibility of building a glibc toolchain when
BR2_PREFER_STATIC_LIB=y.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
The glibc and eglibc support has been introduced since a little bit of
time now, I believe we can remove the "experimental" statement next to
it.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
icu depends on __sync_sub_and_fetch and other atomic primitives that
don't exist in the ARC toolchain yet.
[Peter: adjust beecrypt/php comment dependency, don't mention atomic builtins]
Signed-off-by: Mischa Jonker <mjonker@synopsys.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
This patch fixes the following whitespace problems in Config.in files:
- trailing whitespace
- spaces instead of tabs for indentation
- help text not indented with tab + 2 spaces
Signed-off-by: Thomas De Schampheleire <thomas.de.schampheleire@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Fixes http://autobuild.buildroot.net/results/34f/34f238cdb1b1b4be1a11143b5e4a17c78c2fc289/
The rollingfile functionality only gets built if pthread support is
available, but a call to these functions from log4c_fini() was outside
the #if WITH_ROLLINGFILE conditional, causing linker errors when the
library is used.
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
The libqmi build specifies -Werror in the CFLAGS by default. This causes
build failures such as the following.
http://autobuild.buildroot.net/results/070/070449299aa9025c4bc6769524a6e22d66c92a65/
The -Werror flag can be disabled using a configuration switch.
[Peter: use =no instead and add a comment explaining why]
Signed-off-by: Simon Dawson <spdawson@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
3.12 release notes:
* Fix: Remove alternate method to check for VLAN tag offload on Linux
< 2.6.37 (-k/-K options)
* Fix: Hide state of VLAN tag offload and LRO if the kernel is too old
for us to reliably detect them (-k option)
* Feature: Add register dump support for Solarflare SFC9100 family
(-d option)
3.12.1 release notes:
This includes a couple of changes that should have gone into 3.12.
* Fix: Memory corruption when applying external calibration to
SFF-8472 module diagnostics (-m option)
* Feature: Add Intel 82599 and x540 DCB registers to dump
(-d option)
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Building host-bison needs perl 5.8.7+, as it uses the "-f" option
for site customization scripts. This feature was added in 5.8.7.
Signed-off-by: Mischa Jonker <mjonker@synopsys.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
I have just released version 3.2.1 of the pciutils. It's purely
a maintenance release with a couple of minor bug fixes and minor
improvements. Thus quoth the changelog:
2013-11-10 Martin Mares <mj@ucw.cz>
* Released as 3.2.1.
* CardBus bridge capabilities are displayed.
* PCIe L1 PM substates are decoded.
* Various bugs were fixed in decoding of PCIe capabilities.
* The sysfs back-end does not spit out unnecessary warnings when
empty slots report only a partial device address. This actually
happens on IBM pSeries.
* Updated pci.ids to the today's snapshot of the database.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
