> On Wed, 20 Aug 2025 23:22:31 +0530
> <sk...@marvell.com> wrote:
>
> Patch does not apply cleanly to next-net, there is fuzz.
>
Ack, I will rebase it on next-net.
> > diff --git a/lib/ethdev/ethdev_trace.h b/lib/ethdev/ethdev_trace.h
> > index 482befc209..bc6c3a4a02 100644
> > --- a/lib/ethdev/ethdev_trace.h
> > +++ b/lib/ethdev/ethdev_trace.h
> > @@ -2110,6 +2110,13 @@ RTE_TRACE_POINT_FP(
> > rte_trace_point_emit_string(ret);
> > )
> >
> > +RTE_TRACE_POINT_FP(
> > + rte_eth_trace_link_connector_to_str,
> > + RTE_TRACE_POINT_ARGS(uint8_t link_connector, const char *ret),
> > + rte_trace_point_emit_u8(link_connector);
> > + rte_trace_point_emit_string(ret);
> > +);
> > +
>
> IMHO adding a trace point to a helper function like this is just extra
> clutter,
> remove it.
>
Ack, I will share a separate patch to add trace in this API.
> > +RTE_EXPORT_EXPERIMENTAL_SYMBOL(rte_eth_link_connector_to_str, 25.11)
> > +const char * rte_eth_link_connector_to_str(enum
> > +rte_eth_link_connector link_connector) {
> > + static const char * const link_connector_str[] = {
> > + [RTE_ETH_LINK_CONNECTOR_NONE] = "None",
> > + [RTE_ETH_LINK_CONNECTOR_TP] = "Twisted Pair",
> > + [RTE_ETH_LINK_CONNECTOR_AUI] = "Attachment Unit Interface",
> > + [RTE_ETH_LINK_CONNECTOR_MII] = "Media Independent
> Interface",
> > + [RTE_ETH_LINK_CONNECTOR_FIBER] = "Fiber",
> > + [RTE_ETH_LINK_CONNECTOR_BNC] = "BNC",
> > + [RTE_ETH_LINK_CONNECTOR_DAC] = "Direct Attach Copper",
> > + [RTE_ETH_LINK_CONNECTOR_SGMII] = "SGMII",
> > + [RTE_ETH_LINK_CONNECTOR_QSGMII] = "QSGMII",
> > + [RTE_ETH_LINK_CONNECTOR_XFI] = "XFI",
> > + [RTE_ETH_LINK_CONNECTOR_SFI] = "SFI",
> > + [RTE_ETH_LINK_CONNECTOR_XLAUI] = "XLAUI",
> > + [RTE_ETH_LINK_CONNECTOR_GAUI] = "GAUI",
> > + [RTE_ETH_LINK_CONNECTOR_XAUI] = "XAUI",
> > + [RTE_ETH_LINK_CONNECTOR_CAUI] = "CAUI",
> > + [RTE_ETH_LINK_CONNECTOR_LAUI] = "LAUI",
> > + [RTE_ETH_LINK_CONNECTOR_SFP] = "SFP",
> > + [RTE_ETH_LINK_CONNECTOR_SFP_DD] = "SFP-DD",
> > + [RTE_ETH_LINK_CONNECTOR_SFP_PLUS] = "SFP+",
> > + [RTE_ETH_LINK_CONNECTOR_SFP28] = "SFP28",
> > + [RTE_ETH_LINK_CONNECTOR_QSFP] = "QSFP",
> > + [RTE_ETH_LINK_CONNECTOR_QSFP_PLUS] = "QSFP+",
> > + [RTE_ETH_LINK_CONNECTOR_QSFP28] = "QSFP28",
> > + [RTE_ETH_LINK_CONNECTOR_QSFP56] = "QSFP56",
> > + [RTE_ETH_LINK_CONNECTOR_QSFP_DD] = "QSFP-DD",
> > + [RTE_ETH_LINK_CONNECTOR_OTHER] = "Other",
> > + };
> > + const char *str = NULL;
> > +
> > + if ((link_connector < ((enum
> rte_eth_link_connector)RTE_DIM(link_connector_str))) &&
> > + link_connector_str[link_connector])
> > + str = link_connector_str[link_connector];
>
> This can be simplified to:
> if (link_connector < RTE_DIM(link_connector_str))
> str = link_connector_str[link_connector];
>
> Since enum's act like int in C, no need for all that casting.
>
Correct and In previous version it was like this only but on Windows machine,
it was throwing error. Hence added typecasting.
> > +/**
> > + * @warning
> > + * @b EXPERIMENTAL: this API may change without prior notice.
> > + *
> > + * This function converts an Ethernet link type to a string.
> > + *
> > + * @param link_connector
> > + * The link type to convert.
> > + * @return
> > + * The string representation of the link type.
> > + */
> > +__rte_experimental
> > +const char *rte_eth_link_connector_to_str(enum rte_eth_link_connector
> > +link_connector);
>
> Need to document that it returns NULL for invalid link_connector values.
Ack.
