On 27/07/2016 18:56, Sebastian Huber wrote:
-RTEMS_INLINE_ROUTINE Thread_Control
*_Thread_Lock_acquire_default_for_executing(
+RTEMS_INLINE_ROUTINE Thread_Control
*_Thread_Wait_acquire_default_for_executing(
This and others have broken libbsd:
./libbsd.a(subr_sleepqueue.c.16.o): In function `sleepq_switch':
/opt/work/chris/rtems/libbsd/rtems-libbsd.master/build/arm-rtems4.12-xilinx_zynq_a9_qemu/../../freebsd/sys/kern/subr_sleepqueue.c:615:
undefined reference to `_Thread_Lock_acquire_default_for_executing'
/opt/work/chris/rtems/libbsd/rtems-libbsd.master/build/arm-rtems4.12-xilinx_zynq_a9_qemu/../../freebsd/sys/kern/subr_sleepqueue.c:692:
undefined reference to `_Thread_Lock_release_default'
/opt/work/chris/rtems/libbsd/rtems-libbsd.master/build/arm-rtems4.12-xilinx_zynq_a9_qemu/../../freebsd/sys/kern/subr_sleepqueue.c:692:
undefined reference to `_Thread_Lock_release_default'
/opt/work/chris/rtems/libbsd/rtems-libbsd.master/build/arm-rtems4.12-xilinx_zynq_a9_qemu/../../freebsd/sys/kern/subr_sleepqueue.c:643:
undefined reference to `_Thread_Lock_release_default'
/opt/work/chris/rtems/libbsd/rtems-libbsd.master/build/arm-rtems4.12-xilinx_zynq_a9_qemu/../../freebsd/sys/kern/subr_sleepqueue.c:647:
undefined reference to `_Thread_Lock_acquire_default'
/opt/work/chris/rtems/libbsd/rtems-libbsd.master/build/arm-rtems4.12-xilinx_zynq_a9_qemu/../../freebsd/sys/kern/subr_sleepqueue.c:668:
undefined reference to `_Thread_Lock_release_default'
/opt/work/chris/rtems/libbsd/rtems-libbsd.master/build/arm-rtems4.12-xilinx_zynq_a9_qemu/../../freebsd/sys/kern/subr_sleepqueue.c:677:
undefined reference to `_Thread_Lock_acquire_default'
/opt/work/chris/rtems/libbsd/rtems-libbsd.master/build/arm-rtems4.12-xilinx_zynq_a9_qemu/../../freebsd/sys/kern/subr_sleepqueue.c:668:
undefined reference to `_Thread_Lock_release_default'
/opt/work/chris/rtems/libbsd/rtems-libbsd.master/build/arm-rtems4.12-xilinx_zynq_a9_qemu/../../freebsd/sys/kern/subr_sleepqueue.c:692:
undefined reference to `_Thread_Lock_release_default'
A high performance interface for libbsd is important however it adds an
extra check for those changing the score. This is the second time in the
past few days changes have gone in that have broken libbsd:
https://git.rtems.org/rtems-libbsd/commit/?id=e5aa554ea155d07baccca9d6f17b90678d07d4b0
I suggest libbsd is added into the pre-push regression testing.
Thanks
Chris
_______________________________________________
devel mailing list
devel@rtems.org
http://lists.rtems.org/mailman/listinfo/devel
