CFI: enum vs integer function type mismatches

[samitolvanen]

@error27 was kind enough to run Smatch to locate enum vs int type mismatches that trip CFI checking. Here's a slightly formatted list:

• [x] drivers/staging/rtl8723bs/os_dep/os_intfs.c:430: rtw_netdev_ops.ndo_start_xmit = rtw_xmit_entry (enum netdev_tx)
• [x] drivers/staging/octeon/ethernet.c:521: cvm_oct_npi_netdev_ops.ndo_start_xmit = cvm_oct_xmit (enum netdev_tx)
• [x] drivers/staging/octeon/ethernet.c:535: cvm_oct_xaui_netdev_ops.ndo_start_xmit = cvm_oct_xmit (enum netdev_tx)
• [x] drivers/staging/octeon/ethernet.c:551: cvm_oct_sgmii_netdev_ops.ndo_start_xmit = cvm_oct_xmit (enum netdev_tx)
• [x] drivers/staging/octeon/ethernet.c:567: cvm_oct_spi_netdev_ops.ndo_start_xmit = cvm_oct_xmit (enum netdev_tx)
• [x] drivers/staging/octeon/ethernet.c:581: cvm_oct_rgmii_netdev_ops.ndo_start_xmit = cvm_oct_xmit (enum netdev_tx)
• [x] drivers/staging/octeon/ethernet.c:597: cvm_oct_pow_netdev_ops.ndo_start_xmit = cvm_oct_xmit_pow (enum netdev_tx)
• [x] drivers/staging/rtl8192e/rtl8192e/rtl_core.c:2264: rtl8192_netdev_ops.ndo_start_xmit = rtllib_xmit (enum netdev_tx)
• [x] drivers/staging/rtl8712/os_intfs.c:188: rtl8712_netdev_ops.ndo_start_xmit = r8712_xmit_entry (enum netdev_tx)
• [x] drivers/staging/rtl8192u/r8192U_core.c:4494: rtl8192_netdev_ops.ndo_start_xmit = ieee80211_xmit (enum netdev_tx)
• [x] drivers/staging/r8188eu/os_dep/os_intfs.c:316: rtw_netdev_ops.ndo_start_xmit = rtw_xmit_entry (enum netdev_tx)
• [x] drivers/gpu/drm/rockchip/cdn-dp-core.c:335: cdn_dp_connector_helper_funcs.mode_valid = cdn_dp_connector_mode_valid (enum drm_mode_status)
• [x] drivers/gpu/drm/i915/display/dvo_tfp410.c:310: tfp410_ops.mode_valid = tfp410_mode_valid (enum drm_mode_status)
• [x] drivers/gpu/drm/i915/display/dvo_ch7017.c:406: ch7017_ops.mode_valid = ch7017_mode_valid (enum drm_mode_status)
• [x] drivers/gpu/drm/i915/display/dvo_ch7xxx.c:358: ch7xxx_ops.mode_valid = ch7xxx_mode_valid (enum drm_mode_status)
• [x] drivers/gpu/drm/i915/display/dvo_ns2501.c:702: ns2501_ops.mode_valid = ns2501_mode_valid (enum drm_mode_status)
• [x] drivers/gpu/drm/i915/display/dvo_ivch.c:494: ivch_ops.mode_valid = ivch_mode_valid (enum drm_mode_status)
• [x] drivers/gpu/drm/i915/display/dvo_sil164.c:271: sil164_ops.mode_valid = sil164_mode_valid (enum drm_mode_status)
• [x] drivers/gpu/drm/xlnx/zynqmp_dp.c:1393: zynqmp_dp_connector_helper_funcs.mode_valid = zynqmp_dp_connector_mode_valid (enum drm_mode_status)
• [x] drivers/gpu/drm/imx/imx-tve.c:318: imx_tve_connector_helper_funcs.mode_valid = imx_tve_connector_mode_valid (enum drm_mode_status)
• [x] drivers/gpu/drm/sun4i/sun4i_tv.c:507: sun4i_tv_comp_connector_helper_funcs.mode_valid = sun4i_tv_comp_mode_valid (enum drm_mode_status)
• [x] drivers/gpu/drm/exynos/exynos_mixer.c:1120: mixer_crtc_ops.mode_valid = mixer_mode_valid (enum drm_mode_status)
• [x] drivers/gpu/drm/exynos/exynos_hdmi.c:945: hdmi_connector_helper_funcs.mode_valid = hdmi_mode_valid (enum drm_mode_status)
• [x] drivers/gpu/drm/solomon/ssd130x.c:670: ssd130x_pipe_funcs.mode_valid = ssd130x_display_pipe_mode_valid (enum drm_mode_status)
• [x] drivers/gpu/drm/msm/disp/mdp4/mdp4_lvds_connector.c:87: mdp4_lvds_connector_helper_funcs.mode_valid = mdp4_lvds_connector_mode_valid (enum drm_mode_status)
• [x] drivers/net/wwan/t7xx/t7xx_netdev.c:108: ccmni_netdev_ops.ndo_start_xmit = t7xx_ccmni_start_xmit (enum netdev_tx)
• [x] drivers/net/wwan/iosm/iosm_ipc_wwan.c:152: ipc_inm_ops.ndo_start_xmit = ipc_wwan_link_transmit (enum netdev_tx)
• [x] drivers/net/ethernet/litex/litex_liteeth.c:196: liteeth_netdev_ops.ndo_start_xmit = liteeth_start_xmit (enum netdev_tx)
• [x] drivers/net/ethernet/korina.c:1269: korina_netdev_ops.ndo_start_xmit = korina_send_packet (enum netdev_tx)
• [x] drivers/net/ethernet/microsoft/mana/mana_en.c:378: mana_devops.ndo_start_xmit = mana_start_xmit (enum netdev_tx)
• [x] drivers/net/ethernet/ti/davinci_emac.c:1715: emac_netdev_ops.ndo_start_xmit = emac_dev_xmit (enum netdev_tx)
• [x] drivers/net/ethernet/davicom/dm9000.c:1369: dm9000_netdev_ops.ndo_start_xmit = dm9000_start_xmit (enum netdev_tx)
• [x] drivers/net/ethernet/microchip/sparx5/sparx5_netdev.c:220: sparx5_port_netdev_ops.ndo_start_xmit = sparx5_port_xmit_impl (enum netdev_tx)
• [x] drivers/net/ethernet/microchip/lan966x/lan966x_main.c:458: lan966x_port_netdev_ops.ndo_start_xmit = lan966x_port_xmit (enum netdev_tx)
• [x] drivers/net/ethernet/asix/ax88796c_main.c:934: ax88796c_netdev_ops.ndo_start_xmit = ax88796c_start_xmit (enum netdev_tx)
• [x] net/openvswitch/vport-geneve.c:117: ovs_geneve_vport_ops.send = dev_queue_xmit (enum netdev_tx)
• [x] net/openvswitch/vport-netdev.c:194: ovs_netdev_vport_ops.send = dev_queue_xmit (enum netdev_tx)
• [x] net/openvswitch/vport-gre.c:81: ovs_gre_vport_ops.send = dev_queue_xmit (enum netdev_tx)
• [x] net/openvswitch/vport-vxlan.c:146: ovs_vxlan_netdev_vport_ops.send = dev_queue_xmit (enum netdev_tx)

cc @kees

[nhukc]

Patches are in-flight for all issues. I'm checking them off as they get accepted.

[kees]

cc @GustavoARSilva

[nathanchance]

@nhukc From what I can tell the following changes are still floating around not picked up:

• https://lore.kernel.org/20220913205600.155172-1-nhuck@google.com/
• https://lore.kernel.org/20220929181411.61331-1-nhuck@google.com/

The first one has a conflict against the drm-next tree; the issue is still there but the function got updated in a rework:

https://cgit.freedesktop.org/drm/drm/commit/?id=47e801bd0749f0691a137255e2db35f5348f4f4f https://cgit.freedesktop.org/drm/drm/commit/?id=cbb11ef9871956244c38ab8520dc9abe87ccdb3b

The second one should just be updated against net-next, as the move of mana.h is still in review as far as I can tell. There may be a conflict at merge time but dealing with that is up to the maintainers.

If you are too busy to send updated versions of those, I do not mind doing it, as they impact the work on #1750.

[nhukc]

@nhukc From what I can tell the following changes are still floating around not picked up:

• https://lore.kernel.org/20220913205600.155172-1-nhuck@google.com/
• https://lore.kernel.org/20220929181411.61331-1-nhuck@google.com/

The first one has a conflict against the drm-next tree; the issue is still there but the function got updated in a rework:

https://cgit.freedesktop.org/drm/drm/commit/?id=47e801bd0749f0691a137255e2db35f5348f4f4f https://cgit.freedesktop.org/drm/drm/commit/?id=cbb11ef9871956244c38ab8520dc9abe87ccdb3b

The second one should just be updated against net-next, as the move of mana.h is still in review as far as I can tell. There may be a conflict at merge time but dealing with that is up to the maintainers.

If you are too busy to send updated versions of those, I do not mind doing it, as they impact the work on #1750.

Yep, I'm too busy with other stuff to handle the rest of these. Feel free to take over. Thanks.

[nathanchance]

Yep, I'm too busy with other stuff to handle the rest of these. Feel free to take over. Thanks.

No worries, done!

https://lore.kernel.org/20221109001424.1422495-1-nathan@kernel.org/ https://lore.kernel.org/20221109002629.1446680-1-nathan@kernel.org/

I think we can close this issue in favor of tracking via #1750.