We currently will fail to build for optimizations like -Og because they do not eliminate dead code. We do not need such clean up if we use QLNULL. There is no need to produce a QemuLockable that will be thrown away.
Only testing: $ ../configure $ make -j16 CFLAGS="$CFLAGS" # which I set to use -Og, then -O2 Joe Slater (2): lockable: use QLNULL for a null lockable lockable: do not rely on optimization for null lockables block/block-backend.c | 4 ++-- block/block-copy.c | 2 +- block/mirror.c | 5 +++-- fsdev/qemu-fsdev-throttle.c | 6 +++--- hw/9pfs/9p.c | 2 +- include/qemu/lockable.h | 16 ++++++---------- util/qemu-co-shared-resource.c | 2 +- 7 files changed, 17 insertions(+), 20 deletions(-) -- 2.7.4
