> -----Original Message-----
> From: Andrew Rybchenko <andrew.rybche...@oktetlabs.ru>
> Sent: Monday, October 3, 2022 22:22
> To: Guo, Junfeng <junfeng....@intel.com>; Zhang, Qi Z
> <qi.z.zh...@intel.com>; Wu, Jingjing <jingjing...@intel.com>; Xing,
> Beilei <beilei.x...@intel.com>
> Cc: dev@dpdk.org; Wang, Xiao W <xiao.w.w...@intel.com>; Qiao,
> Wenjing <wenjing.q...@intel.com>
> Subject: Re: [PATCH v2 14/14] net/idpf: add support for timestamp
> offload
>
> On 9/5/22 13:58, Junfeng Guo wrote:
> > Add support for timestamp offload.
> >
> > Signed-off-by: Wenjing Qiao <wenjing.q...@intel.com>
> > Signed-off-by: Junfeng Guo <junfeng....@intel.com>
>
> [snip]
>
> > +/* Helper function to convert a 32b nanoseconds timestamp to 64b. */
> > +static inline uint64_t
> > +idpf_tstamp_convert_32b_64b(struct iecm_hw *hw, struct
> idpf_adapter *ad,
> > + uint32_t flag, uint32_t in_timestamp)
> > +{
> > +/* TODO: timestamp for ACC */
> > +#ifdef RTE_ARCH_ARM64
> > + return 0;
> > +#endif /* RTE_ARCH_ARM64 */
> > +
> > +#ifdef RTE_ARCH_X86_64
> > + const uint64_t mask = 0xFFFFFFFF;
> > + uint32_t hi, lo, lo2, delta;
> > + uint64_t ns;
> > +
> > + if (flag) {
> > + IECM_WRITE_REG(hw, GLTSYN_CMD_SYNC_0_0,
> PF_GLTSYN_CMD_SYNC_SHTIME_EN_M);
> > + IECM_WRITE_REG(hw, GLTSYN_CMD_SYNC_0_0,
> PF_GLTSYN_CMD_SYNC_EXEC_CMD_M |
> > + PF_GLTSYN_CMD_SYNC_SHTIME_EN_M);
> > + lo = IECM_READ_REG(hw, PF_GLTSYN_SHTIME_L_0);
> > + hi = IECM_READ_REG(hw, PF_GLTSYN_SHTIME_H_0);
> > + /*
> > + * On typical system, the delta between lo and lo2 is
> ~1000ns,
> > + * so 10000 seems a large-enough but not overly-big
> guard band.
> > + */
> > + if (lo > (UINT32_MAX -
> IDPF_TIMESYNC_REG_WRAP_GUARD_BAND))
> > + lo2 = IECM_READ_REG(hw,
> PF_GLTSYN_SHTIME_L_0);
> > + else
> > + lo2 = lo;
> > +
> > + if (lo2 < lo) {
> > + lo = IECM_READ_REG(hw,
> PF_GLTSYN_SHTIME_L_0);
> > + hi = IECM_READ_REG(hw,
> PF_GLTSYN_SHTIME_H_0);
> > + }
> > +
> > + ad->time_hw = ((uint64_t)hi << 32) | lo;
> > + }
> > +
> > + delta = (in_timestamp - (uint32_t)(ad->time_hw & mask));
> > + if (delta > (mask / 2)) {
> > + delta = ((uint32_t)(ad->time_hw & mask) - in_timestamp);
> > + ns = ad->time_hw - delta;
> > + } else {
> > + ns = ad->time_hw + delta;
> > + }
> > +
> > + return ns;
> > +#endif /* RTE_ARCH_X86_64 */
>
> Conditional compilation is rather strange above.
> Will it break build on some architectures?
> Non-x86-64 and non-ARM64.
Actually the timestamp feature currently is only supported on x86_64
architecture. On the rest conditions will just return 0. We will update
this part in the coming version. Thanks!
>
> > +}
> >
> > +#endif /* _IDPF_RXTX_H_ */
