> On 1/20/21 12:24 PM, Stokes, Ian wrote: > >> On 1/15/21 6:37 PM, Ian Stokes wrote: > >>> DPDK flags may include -march. Forcing -march could be > >>> considered too heavy a requirement when users compile OVS from > >>> source and could override user provided options. > >>> > >>> Resolve this by stripping -march from provided DPDK flags. > >>> > >>> Signed-off-by: Ian Stokes <[email protected]> > >>> --- > >>> acinclude.m4 | 3 +++ > >>> 1 file changed, 3 insertions(+) > >>> > >>> diff --git a/acinclude.m4 b/acinclude.m4 > >>> index 857067a88..14fe4218a 100644 > >>> --- a/acinclude.m4 > >>> +++ b/acinclude.m4 > >>> @@ -436,6 +436,9 @@ AC_DEFUN([OVS_CHECK_DPDK], [ > >>> if test "$DPDK_AUTO_DISCOVER" = "false"; then > >>> OVS_LDFLAGS="$OVS_LDFLAGS -L$DPDK_LIB_DIR" > >>> fi > >>> + # Force in pkg-config since this could override user-specified > >>> options. > >> > >> Seems like some word is missing in this sentence. > > > > Yes took this from the original patch but not crazy about it either. > > > > How about > > > > "Strip -march from DPDK flags since this could override user-specified > option." > > I guess, Felix just didn't copy the first line of the comment from master. > I think, we should just take it from there: >
Ah, I get you, sure will use the original comment from master on commit so. Thanks Ian > # Stripping out possible instruction set specific configuration that DPDK > # forces in pkg-config since this could override user-specified options. > # It's enough to have -mssse3 to build with DPDK headers. > > > > > BR > > Ian > >> > >>> + # It's enough to have -mssse3 to build with DPDK headers. > >>> + DPDK_INCLUDE=$(echo "$DPDK_INCLUDE" | sed 's/-march=[[^ > ]]*//g') > >>> OVS_CFLAGS="$OVS_CFLAGS $DPDK_INCLUDE" > >>> OVS_ENABLE_OPTION([-mssse3]) > >>> > >>> _______________________________________________ dev mailing list [email protected] https://mail.openvswitch.org/mailman/listinfo/ovs-dev
