Markus Armbruster <arm...@redhat.com> writes: > Don't bother with include/exec/memory-internal.h and > include/hw/elf_ops.h, because those are somewhat special. > > Signed-off-by: Markus Armbruster <arm...@redhat.com> > --- > hw/audio/lm4549.h | 1 + > hw/cris/boot.h | 2 ++ > hw/net/e1000e_core.h | 7 +++++++ > hw/net/vmware_utils.h | 1 + > hw/scsi/mptsas.h | 1 + > include/disas/disas.h | 1 + > include/hw/arm/sharpsl.h | 3 +++ > include/hw/block/fdc.h | 1 + > include/hw/char/escc.h | 2 ++ > include/hw/char/pl011.h | 2 ++ > include/hw/char/xilinx_uartlite.h | 2 ++ > include/hw/cris/etraxfs_dma.h | 2 ++ > include/hw/empty_slot.h | 2 ++ > include/hw/misc/mips_cmgcr.h | 1 + > include/hw/pci-host/apb.h | 1 + > include/hw/sparc/sparc32_dma.h | 2 ++ > include/hw/timer/m48t59.h | 1 + > include/sysemu/xen-mapcache.h | 1 + > target-ppc/mmu-hash32.h | 2 ++ > tests/Makefile.include | 9 --------- > 20 files changed, 35 insertions(+), 9 deletions(-)
I think the churn is tolerable, but I should mention possible alternatives: * Move the typedef to osdep.h. I don't like it, because it spreads hwaddr's 64-bitness over two places. * Include hwaddr.h from osdep.h If you have a preference, let me know.
