> On Oct 2, 2017, at 11:40 PM, Andrew Rybchenko <[email protected]> > wrote: > > On 10/03/2017 09:33 AM, Raslan Darawsheh wrote: >> Add a new offload capability flag for Rx HW >> timestamp and enabling/disabling this via rte_eth_rxmode. >> >> Signed-off-by: Raslan Darawsheh >> <[email protected]> >> >> --- >> This patch should be applied after after this series: >> >> http://dpdk.org/dev/patchwork/patch/29368/ >> >> --- >> doc/guides/nics/features.rst | 11 +++++++++++ >> lib/librte_ether/rte_ethdev.c | 6 ++++++ >> lib/librte_ether/rte_ethdev.h | 5 ++++- >> 3 files changed, 21 insertions(+), 1 deletion(-) >> > > <...> > >> diff --git a/lib/librte_ether/rte_ethdev.h b/lib/librte_ether/rte_ethdev.h >> index ffd2ee5..bd63730 100644 >> --- a/lib/librte_ether/rte_ethdev.h >> +++ b/lib/librte_ether/rte_ethdev.h >> @@ -368,7 +368,8 @@ struct rte_eth_rxmode { >> jumbo_frame : 1, /**< Jumbo Frame Receipt enable. */ >> hw_strip_crc : 1, /**< Enable CRC stripping by hardware. */ >> enable_scatter : 1, /**< Enable scatter packets rx handler */ >> - enable_lro : 1; /**< Enable LRO */ >> + enable_lro : 1, /**< Enable LRO */ >> + hw_timestamp : 1; /**< Enable HW timestamp */ >> > > The code is definitely not rebased on top of a new Rx offload API patch which > adds > ignore_offload_bitfield : 1; > exactly in this place.
Hi Andrew, Did you check Shahaf's email regarding this? Looks like Shahaf suggests keeping it as is for now. > On Sep 30, 2017, at 11:44 PM, Shahaf Shuler <[email protected]> wrote: > >[1] http://dpdk.org/ml/archives/dev/2017-September/076872.html > > I agree it should be on top of [1]. > Unfortunately since the PMDs will move to the new API only on 18.02, and in > the current state Rx offloads are disabled by default, there will be no way > to enable this feature without a dedicated bit. > > So my suggestion is to keep the timestamp bit on rxmode, and to update the > convert function introduced on [1]. This bit will be removed along with the > entire bit-field array once the old offloads API will be deprecated. Thanks, Yongseok
