On Fri, Jul 18, 2025 at 02:54:04PM +0100, Stuart Henderson wrote:
> Wonder if this might do the trick, though a bigger change of building
> up the firmware filename from pieces would seem likely to give correct
> results more often, and probably end up less complicated.
Building the name up incrementally is exactly what Intel has now done
in iwlwifi upstream. We could port this over eventually.
> Index: if_iwx.c
> ===================================================================
> RCS file: /cvs/src/sys/dev/pci/if_iwx.c,v
> diff -u -p -r1.191 if_iwx.c
> --- if_iwx.c 29 Jun 2025 19:32:08 -0000 1.191
> +++ if_iwx.c 18 Jul 2025 13:49:59 -0000
> @@ -11247,15 +11247,16 @@ iwx_preinit(struct iwx_softc *sc)
> /* Print version info and MAC address on first successful fw load. */
> sc->attached = 1;
> if (sc->sc_pnvm_ver) {
> - printf("%s: hw rev 0x%x, fw %s, pnvm %08x, "
> + printf("%s: hw rev 0x%x, rf 0x%x, fw %s, pnvm %08x, "
> "address %s\n",
> DEVNAME(sc), sc->sc_hw_rev & IWX_CSR_HW_REV_TYPE_MSK,
> - sc->sc_fwver, sc->sc_pnvm_ver,
> + sc->sc_hw_rf_id, sc->sc_fwver, sc->sc_pnvm_ver,
> ether_sprintf(sc->sc_nvm.hw_addr));
> } else {
> - printf("%s: hw rev 0x%x, fw %s, address %s\n",
> + printf("%s: hw rev 0x%x, rf 0x%x, fw %s, address %s\n",
> DEVNAME(sc), sc->sc_hw_rev & IWX_CSR_HW_REV_TYPE_MSK,
> - sc->sc_fwver, ether_sprintf(sc->sc_nvm.hw_addr));
> + sc->sc_hw_rf_id, sc->sc_fwver,
> + ether_sprintf(sc->sc_nvm.hw_addr));
> }
I agree that displaying the RF ID in dmesg would be good.
