Hi Jasvinder, > > APIs for selecting the architecure specific implementation and computing > the crc (16-bit and 32-bit CRCs) are added. For CRCs calculation, scalar > as well as x86 intrinsic(sse4.2) versions are implemented. > > The scalar version is based on generic Look-Up Table(LUT) algorithm, > while x86 intrinsic version uses carry-less multiplication for > fast CRC computation. > > Signed-off-by: Jasvinder Singh <jasvinder.si...@intel.com> > --- > lib/librte_net/Makefile | 3 + > lib/librte_net/rte_net_crc.c | 205 ++++++++++++++++++++++ > lib/librte_net/rte_net_crc.h | 104 +++++++++++ > lib/librte_net/rte_net_crc_sse.h | 351 > +++++++++++++++++++++++++++++++++++++ > lib/librte_net/rte_net_version.map | 8 + > 5 files changed, 671 insertions(+) > create mode 100644 lib/librte_net/rte_net_crc.c > create mode 100644 lib/librte_net/rte_net_crc.h > create mode 100644 lib/librte_net/rte_net_crc_sse.h > > + > +/** > + * CRC compute API > + * > + * @param mbuf > + * Pointer to the packet mbuf for crc computation > + * @param data_offset > + * Offset to the data in the mbuf > + * @param data_len > + * length of the data to compute the crc on > + * @param type > + * crc type > + * > + * @return > + * computed crc value > + */ > +uint32_t > +rte_net_crc_calc(struct rte_mbuf *mbuf, > + uint32_t data_offset, > + uint32_t data_len, > + enum rte_net_crc_type type);
I think it will probably be convenient to have this API not dependent on mbuf, something like: uint32_t rte_net_crc_calc(const void *buf, uint32_t data_len, enum rte_net_crc_type type); And if we like to have an extra function that would do similar thing for mbuf, that's fine too, but I suppose such function would have to handle multi-segment case too. Konstantin
