net: phylink: use legacy_pre_march2020
Hello Russell King (Oracle),
This is a semi-automatic email about new static checker warnings.
The patch 001f4261fe4d: "net: phylink: use legacy_pre_march2020" from Dec 9, 2021, leads to the following Smatch complaint:
drivers/net/phy/phylink.c:823 phylink_change_inband_advert()
error: we previously assumed 'pl->pcs_ops' could be null (see line 806)drivers/net/phy/phylink.c
805
806 if (!pl->pcs_ops && pl->config->legacy_pre_march2020) {
^^^^^^^^^^^^Should this be an ||?
807 /* Legacy method */
808 phylink_mac_config(pl, &pl->link_config);
809 phylink_mac_pcs_an_restart(pl);
810 return 0;
811 }
812
813 phylink_dbg(pl, "%s: mode=%s/%s adv=%*pb pause=%02x\n", __func__,
814 phylink_an_mode_str(pl->cur_link_an_mode),
815 phy_modes(pl->link_config.interface),
816 __ETHTOOL_LINK_MODE_MASK_NBITS, pl->link_config.advertising,
817 pl->link_config.pause);
818
819 /* Modern PCS-based method; update the advert at the PCS, and
820 * restart negotiation if the pcs_config() helper indicates that
821 * the programmed advertisement has changed.
822 */
823 ret = pl->pcs_ops->pcs_config(pl->pcs, pl->cur_link_an_mode,
^^^^^^^^^^^^^Unchecked dereference.
824 pl->link_config.interface,
825 pl->link_config.advertising,
regards,
dan carpenter
Quoted 7 lines · Dan Carpenter
Hello Russell King (Oracle),
This is a semi-automatic email about new static checker warnings.
The patch
001f4261fe4d: "net: phylink: uselegacy_pre_march2020" from Dec 9, 2021, leads to the following Smatch complaint:
Quoted 8 lines
drivers/net/phy/phylink.c:823 phylink_change_inband_advert() error: we previously assumed 'pl->pcs_ops' could be null (see line 806)
drivers/net/phy/phylink.c
805806 if (!pl->pcs_ops && pl->config->legacy_pre_march2020) { ^^^^^^^^^^^^Should this be an ||?
No, the intention is correct, even though it looks a little weird.
We only call phylink_change_inband_advert() if we are in in-band mode, and in-band mode for non-legacy users requires a PCS. Consequently, if legacy_pre_march2020 is clear, we have no PCS, and we reach here, then something is definitely broken.
There isn't an easy way to detect this condition earlier, but we could do something like:
if (WARN_ON(!pl->pcs_ops)) return -EINVAL;
which should be sufficiently noisy for people to do something about.
Signature · 2 lines
RMK's Patch system: https://www.armlinux.org.uk/developer/patches/ FTTP is here! 40Mbps down 10Mbps up. Decent connectivity at last!