Neither configure.ac nor configure.in exists, so autoreconf is a no-op.
This commit drops it.
Also, run "make dep" as advertised by the README and the configure
script.
Signed-off-by: Sebastian Kemper <sebastian_ml@gmx.net>
This sets PKG_EXTMOD_SUBDIRS so kernel.mk can find the DAHDI
Module.symvers file.
Also, this puts KERNEL_MAKE_FLAGS into the make flags. This way
Module.symvers files of other modules are made available, plus there is
no need anymore to specify ARCH and CROSS_COMPILE, as KERNEL_MAKE_FLAGS
already takes care of that.
TARGET_CONFIGURE_OPTS is dropped because this seems out of place
(configure vs. compile).
Signed-off-by: Sebastian Kemper <sebastian_ml@gmx.net>
The patch "010-fix-non-glibc-builds.patch" is broken. The test for
HAVE_EXECINFO_H must happen after including "<autoconfig.h>".
Instead of fixing the patch this commit removes it. In lieu,
CONFIGURES_ARGS are added to the Makefile. These disable the building of
some tools that are not packaged anyway. Like this, "xpp/xtalk/debug.c"
is avoided, hence the patch becomes irrelevant.
Signed-off-by: Sebastian Kemper <sebastian_ml@gmx.net>
Bump to upstream release 2.3.0. Move folder and change PKG_NAME so that
we don't have to deal with how github names the tar.gz archive.
Also, don't install the libsrtp2.so symlink anymore on target, as it is
not required and only used in staging.
Signed-off-by: Sebastian Kemper <sebastian_ml@gmx.net>
- remove 010-openssl-deprecated.patch as it does not apply anymore and
upstream seems to have fixed the issue
- upstream closed down the server http://www.creytiv.com/ so this
changes URLs
Signed-off-by: Sebastian Kemper <sebastian_ml@gmx.net>
Includes fix that buildroot submitted upstream.
This introduces a variable to make following upstream versions a bit
easier.
Signed-off-by: Sebastian Kemper <sebastian_ml@gmx.net>
This commit adds an updated spandsp library. This is not a drop-in
replacement for the "old" spandsp library. Applications that want to use
it need to be updated a bit, according to upstream info [1].
In an effort to not cause any problems for OpenWrt, this package
installs the library and headers in the staging directory into
non-standard paths:
Library: $(STAGING_DIR/usr/lib/spandsp3/lib
Headers: $(STAGING_DIR/usr/lib/spandsp3/include
This way they are hidden aways from all normal users (like asterisk,
baresip etc.) and won't interfere.
To use the new spandsp library users can look for the spandsp3
pkg-config file (the regular spandsp installs spandsp.pc, this package
installs spandsp3.pc). This should be enough. The first such user will
likely be the freeswitch package, once it gets updated to 1.10.5.
[1] https://github.com/freeswitch/spandsp/issues/5
Signed-off-by: Sebastian Kemper <sebastian_ml@gmx.net>
This works around the build failure with gcc-10 until there is a patch
for dahdi-tools that is reviewed and acknowledged by upstream.
Signed-off-by: Sebastian Kemper <sebastian_ml@gmx.net>
On some 32 bit targets dahdi-linux currently fails to build.
Examples:
ERROR: "__divdi3" [/builder/shared-workdir/build/sdk/build_dir/target-powerpc_8540_musl/linux-mpc85xx_generic/dahdi-linux-3.1.0/drivers/dahdi/xpp/xpp_usb.ko] undefined!
ERROR: "__udivdi3" [/builder/shared-workdir/build/sdk/build_dir/target-powerpc_8540_musl/linux-mpc85xx_generic/dahdi-linux-3.1.0/drivers/dahdi/xpp/xpp_usb.ko] undefined!
ERROR: "__moddi3" [/builder/shared-workdir/build/sdk/build_dir/target-powerpc_8540_musl/linux-mpc85xx_generic/dahdi-linux-3.1.0/drivers/dahdi/xpp/xpp.ko] undefined!
ERROR: "__divdi3" [/builder/shared-workdir/build/sdk/build_dir/target-powerpc_8540_musl/linux-mpc85xx_generic/dahdi-linux-3.1.0/drivers/dahdi/xpp/xpp.ko] undefined!
or
ERROR: "__aeabi_ldivmod" [/store/buildbot/slave/shared-workdir/build/sdk/build_dir/target-arm_arm1176jzf-s+vfp_musl_eabi/linux-bcm27xx_bcm2708/dahdi-linux-3.1.0/drivers/dahdi/xpp/xpp_usb.ko] undefined!
ERROR: "__aeabi_uldivmod" [/store/buildbot/slave/shared-workdir/build/sdk/build_dir/target-arm_arm1176jzf-s+vfp_musl_eabi/linux-bcm27xx_bcm2708/dahdi-linux-3.1.0/drivers/dahdi/xpp/xpp_usb.ko] undefined!
ERROR: "__aeabi_ldivmod" [/store/buildbot/slave/shared-workdir/build/sdk/build_dir/target-arm_arm1176jzf-s+vfp_musl_eabi/linux-bcm27xx_bcm2708/dahdi-linux-3.1.0/drivers/dahdi/xpp/xpp.ko] undefined!
Problem was reported to upstream already by third party ([1]). Patch
attached to bug report is not sufficient.
This commit adds a patch from PLD Linux ([2]) to address the build
failures.
[1] https://issues.asterisk.org/jira/browse/DAHLIN-371
[2] https://github.com/pld-linux/dahdi-linux/blob/master/math64.patch
Signed-off-by: Sebastian Kemper <sebastian_ml@gmx.net>
This commit provides compatibility with kernel 5.4.
Except for the version bump this commit updates the HFC-S PCI driver to
not rely on the macro "dahdi_pci_module" anymore, because it has been
removed upstream in commit 4af6f69.
Also, a patch is added that makes the include of "<linux/pci-aspm.h>"
conditional on the kernel version, because starting with kernel 5.4 this
header has been removed and its contents merged into "pci.h".
Signed-off-by: Sebastian Kemper <sebastian_ml@gmx.net>
The .pc file uses hard-coded paths:
prefix=/usr
exec_prefix=${prefix}
libdir=/usr/lib
includedir=/usr/include
It also pushes "-I/usr/include" into CFLAGS. This commit amends all of
that.
Resolve: #483
Signed-off-by: Sebastian Kemper <sebastian_ml@gmx.net>
Quote from Asterisk commit e6cc1f5083bbb85a76280e4bbada7be265b0a106:
"Issues in pjproject 2.9 caused us to revert some of their changes as a
work around. This introduced another issue where pjproject wouldn't
build with older gcc versions such as that found on CentOS 6. This
commit replaces the reverts with the official fixes for the original
issues and allows pjproject to be built on CentOS 6 again."
Signed-off-by: Sebastian Kemper <sebastian_ml@gmx.net>
Version bump as preparation for Asterisk update.
In other news:
- upgrades ULRs to https
- adds new configure switch --disable-darwin-ssl (doesn't apply to
OpenWrt, but disable it explicitly anyway)
- use INSTALL_DATA for pkg-config file
- patch naming changed to be in line with Asterisk's patch naming,
making maintenance easier
- drops some patches, adds some new ones from Asterisk (mainly reverts
of some pjproject commits which turned out to cause memory leaks)
Signed-off-by: Sebastian Kemper <sebastian_ml@gmx.net>
Version bump.
- add ABI_VERSION as per OpenWrt package policies [1]
- don't package version-less symlink also as per [1]
- add --disable-silent-rules to keep build verbose
- remove leftover whitespace
- drop patch now included in tarball
[1] https://openwrt.org/docs/guide-developer/package-policies
Signed-off-by: Sebastian Kemper <sebastian_ml@gmx.net>
This commit goes back from using a git checkout to using a regular
upstream tarball. Version bumped to 3.0.0 which works with kernels
equal to or greater than 4.15. Patches refreshed.
Signed-off-by: Sebastian Kemper <sebastian_ml@gmx.net>
"check whether UPDATE is supported on outgoing calls", commit 5de36abd:
In ASTERISK-27095 an issue had been fixed because of which chan_pjsip was not
trying to send UPDATE messages when connected_line_method was set to invite.
However this only solved the issue for incoming INVITES. For outgoing INVITES
(important when transferring calls) the options variable needs to be updated
at a different place.
"Add patch for double free issue in timer heap", commit 9c11399b:
Fixed#2172: Avoid double reference counter decrements in
timer in the scenario of race condition between
pj_timer_heap_cancel() and pj_timer_heap_poll().
"Add timer patch from pjproject r5934", commit d4cd2a97.
Signed-off-by: Andre Heider <a.heider@gmail.com>
Currently builds for ARC are failing, causing follow-up breakage. It
seems that uclibc defines __GLIBC__ on the build bots.
This commit adds a test for execinfo.h to the configure script. The
result of the test is then used in xpp/xtalk/debug.c.
Signed-off-by: Sebastian Kemper <sebastian_ml@gmx.net>
Add two patches to fix general compile issues. The first masks out some
glibc specific extensions if using a different libc.
The second fixes a compile issue with newer compilers.
Like this it's possible to easily add more executables to the packages
should the need (or wish) arise. This also gets rid of the following
warnings:
cc1: note: someone does not honour COPTS correctly, passed 2 times
Signed-off-by: Sebastian Kemper <sebastian_ml@gmx.net>
In a previous commit I updated the copyright message. But that was not
correct as I am not affiliated with OpenWrt. This commit reverts the
change.
Signed-off-by: Sebastian Kemper <sebastian_ml@gmx.net>
In a previous commit I updated the copyright message. But that was not
correct as I am not affiliated with OpenWrt. This commit reverts the
change.
Signed-off-by: Sebastian Kemper <sebastian_ml@gmx.net>
Upstream fixed dependency on libpcap. The lib does not link to it
anymore as it does not need it.
Signed-off-by: Sebastian Kemper <sebastian_ml@gmx.net>
Version bump. New version does not have too many changes, but upstream
cleaned up the autoconf files. As a result autoreconf is now needed to
get the configure script.
This commit adds a configure switch to get the make output as verbose as
in the past.
Signed-off-by: Sebastian Kemper <sebastian_ml@gmx.net>
In a previous commit I updated the copyright message. But that was not
correct as I am not affiliated with OpenWrt. This commit reverts the
change.
Signed-off-by: Sebastian Kemper <sebastian_ml@gmx.net>
It is unlikely that SOFT_FLOAT is toggled, but add it to
PKG_CONFIG_DEPENDS for completeness' sake anyway.
Signed-off-by: Sebastian Kemper <sebastian_ml@gmx.net>
Includes security fixes:
2094: Crash when receiving SDP with invalid fmtp attribute
2095: Crash when parsing SDP with an invalid media format description
Signed-off-by: Sebastian Kemper <sebastian_ml@gmx.net>
spandsp already makes sure that the compiler emits position-independent
code. Adding $(FPIC) makes no difference. In addition the spandsp build
system puts its own flag behind the CFLAGS, overriding whatever was set
before.
Signed-off-by: Sebastian Kemper <sebastian_ml@gmx.net>
bgg729 already makes sure that the compiler emits position-independent
code. Adding $(FPIC) makes no difference. In addition the libsrtp build
system puts its own flag behind the CFLAGS, overriding whatever was set
before.
Signed-off-by: Sebastian Kemper <sebastian_ml@gmx.net>
- Use $(INSTALL_DATA) for the header file.
- Make sure that also the libpri.so symlink gets installed.
Signed-off-by: Sebastian Kemper <sebastian_ml@gmx.net>
libosip2 already makes sure that the compiler emits position-independent
code. Adding $(FPIC) makes no difference. In addition the libosip2 build
system puts its own flag behind the CFLAGS, overriding whatever was set
before.
Signed-off-by: Sebastian Kemper <sebastian_ml@gmx.net>
- use $(INSTALL_BIN) and $(INSTALL_DATA) where possible.
- instead of creating symlinks manually just copy them.
Signed-off-by: Sebastian Kemper <sebastian_ml@gmx.net>
- Remove all flags/variables that are not used at all.
- dahdi-tools already makes sure that the compiler emits
position-independent code for libtonezone. Adding $(FPIC) makes no
difference. In addition the dahdi-toolsbuild system puts its own flag
behind the CFLAGS, overriding whatever was set before. Also the
applications should never be built with PIC, it should only be done for
the lib.
Signed-off-by: Sebastian Kemper <sebastian_ml@gmx.net>
Remove all flags/variables that are not used at all. Remove some ':'
that are not needed in the context.
Signed-off-by: Sebastian Kemper <sebastian_ml@gmx.net>
bgg729 already makes sure that the compiler emits position-independent
code. Adding $(FPIC) makes no difference. In addition the bcg729 build
system puts its own flag behind the CFLAGS, overriding whatever was set
before.
Signed-off-by: Sebastian Kemper <sebastian_ml@gmx.net>
Add TARGET_CPPFLAGS to TARGET_CFLAGS, otherwise OpenWrt's CPPFLAGS are
ignored (i.e. fortify-source headers are not picked up).
Signed-off-by: Sebastian Kemper <sebastian_ml@gmx.net>
Starting with 4.12 there are module build errors due to the
arch-specific uaccess.h using a macro defined in the generic include.
The generic include has been around since 2.6.18 and has always included
asm/uaccess.h, so switch to using the generic include for all kernels
which have it.
Example build failure:
CC [M] /data/bowl-builder/mips64_octeon/build/sdk/build_dir/target-mips64_octeon_64_musl/linux-octeon/dahdi-linux-2.11.1-20180111/drivers/dahdi/wcb4xxp/base.o
In file included from ../dahdi-linux-2.11.1-20180111/drivers/dahdi/wcb4xxp/base.c:38:0:
./arch/mips/include/asm/uaccess.h: In function 'eva_kernel_access':
./arch/mips/include/asm/uaccess.h:91:9: error: implicit declaration of function 'uaccess_kernel' [-Werror=implicit-function-declaration]
return uaccess_kernel();
^
cc1: some warnings being treated as errors
The same can currently be seen on arc_archs where also kernel 4.14.18 is
used.
The patch just follows what Ubuntu's Seth Forshee suggested for openafs,
which ran into the same problem. The commit message is largely just
copied from Seth's commit message as well :)
The issue has been reported to upstream here:
https://issues.asterisk.org/jira/browse/DAHLIN-360
The same patch has been uploaded there.
Signed-off-by: Sebastian Kemper <sebastian_ml@gmx.net>
ARC toolchain exports __arc__, which clashes with a symbol of the same
name in libosip2. This commit undefines the toolchain symbol when
building the library.
The fix was found in buildroot repo. Thanks to whoever spotted this!
Signed-off-by: Sebastian Kemper <sebastian_ml@gmx.net>
Currently dahdi-linux fails to compile on targets that use kernel 4.14.
Changes were introduced in kernels 4.11 and 4.13 that cause these build
failures.
Update to an upstream git snapshot to fix the breakage.
Signed-off-by: Sebastian Kemper <sebastian_ml@gmx.net>
