On Mon, 24 Apr 2023 10:29:19 -0700 Tyler Retzlaff <roret...@linux.microsoft.com> wrote:
> On Mon, Apr 24, 2023 at 09:23:48AM -0700, Stephen Hemminger wrote: > > On Mon, 24 Apr 2023 18:47:00 -0400 > > Qi Zhang <qi.z.zh...@intel.com> wrote: > > > > > Remove compile option "__KERNEL" which should not be considered in > > > DPDK. Also only #include <rte_xxx> in idpf_osdep.h. > > > > > > Signed-off-by: Qi Zhang <qi.z.zh...@intel.com> > > > > This will cause some warnings in DPDK build depending on compiler > > version and flags. You need to leave the fallthrough comment or > > use one of the other fallthrough annotations. > > if there are variations of annotating fallthrough i would not object to > a macro for it being exposed from rte_common.h > > full disclosure it would let me use the equivalent that are provided > with windows and windows toolchains. Yes having something like __rte_fallthrough would help. Wouldn't help code that is trying to always work in kernel, DPDK or other places. Both Gcc and clang use statement attributes and C++ has [[fallthrough]] https://gcc.gnu.org/onlinedocs/gcc/Statement-Attributes.html https://clang.llvm.org/docs/AttributeReference.html#fallthrough