1.. SPDX-License-Identifier: GPL-2.0 2 3===================================================== 4Netdev features mess and how to get out from it alive 5===================================================== 6 7Author: 8 Michał Mirosław <mirq-linux@rere.qmqm.pl> 9 10 11 12Part I: Feature sets 13==================== 14 15Long gone are the days when a network card would just take and give packets 16verbatim. Today's devices add multiple features and bugs (read: offloads) 17that relieve an OS of various tasks like generating and checking checksums, 18splitting packets, classifying them. Those capabilities and their state 19are commonly referred to as netdev features in Linux kernel world. 20 21There are currently three main sets of features on each netdevice, 22first and second are initialized by the driver: 23 24 1. netdev->hw_features set contains features whose state may possibly 25 be changed (enabled or disabled) for a particular device by user's 26 request. Drivers normally initialize this set before registration or 27 in the ndo_init callback. Changes after registration should be made 28 very carefully as other parts of the code may assume hw_features are 29 static. At the very least changes must be made under rtnl_lock and 30 the netdev instance lock, and followed by netdev_update_features(). 31 32 2. netdev->features set contains features which are currently enabled 33 for a device. This should be changed only by network core or in 34 error paths of ndo_set_features callback. 35 36 3. netdev->wanted_features set contains feature set requested by user. 37 This set is filtered by ndo_fix_features callback whenever it or 38 some device-specific conditions change. This set is internal to 39 networking core and should not be referenced in drivers. 40 41On top of those three main sets, each netdev has: 42 43 1. Sets which control features inherited by child devices (VLAN, MPLS, 44 hw_enc for L3/L4 tunnels). These sets allow the driver to limit which 45 netdev->features are propagated, in case HW cannot perform the offloads 46 with the extra headers present. 47 48 2. netdev->mangleid_features, TSO features which are supported only when 49 IP ID field can be mangled (constant instead of incrementing) during TSO. 50 51 3. netdev->gso_partial_features, additional TSO features which HW can 52 support via NETIF_F_GSO_PARTIAL. 53 54Part II: Controlling enabled features 55===================================== 56 57When current feature set (netdev->features) is to be changed, new set 58is calculated and filtered by calling ndo_fix_features callback 59and netdev_fix_features(). If the resulting set differs from current 60set, it is passed to ndo_set_features callback and (if the callback 61returns success) replaces value stored in netdev->features. 62NETDEV_FEAT_CHANGE notification is issued after that whenever current 63set might have changed. 64 65The following events trigger recalculation: 66 1. device's registration, after ndo_init returned success 67 2. user requested changes in features state 68 3. netdev_update_features() is called 69 70ndo_*_features callbacks are called with rtnl_lock held. Missing callbacks 71are treated as always returning success. 72 73A driver that wants to trigger recalculation must do so by calling 74netdev_update_features() while holding rtnl_lock. If the device uses the 75netdev instance lock, that lock must be held as well. This should not be 76done from ndo_*_features callbacks. netdev->features should not be modified 77by driver except by means of ndo_fix_features callback. 78 79ndo_features_check is called for each skb before that skb is passed to 80ndo_start_xmit. Driver may perform any non-trivial checks (e.g. exact 81header geometry / length) and withdraw features like HW_CSUM or TSO, 82requesting the networking stack to fall back to the software implementation. 83 84Part III: Implementation hints 85============================== 86 87 * ndo_fix_features: 88 89All dependencies between features should be resolved here. The resulting 90set can be reduced further by networking core imposed limitations (as coded 91in netdev_fix_features()). For this reason it is safer to disable a feature 92when its dependencies are not met instead of forcing the dependency on. 93 94This callback should not modify hardware nor driver state (should be 95stateless). It can be called multiple times between successive 96ndo_set_features calls. 97 98Callback must not alter features contained in NETIF_F_SOFT_FEATURES or 99NETIF_F_NEVER_CHANGE, except that NETIF_F_VLAN_CHALLENGED may be changed. 100Care must be taken as changes to NETIF_F_VLAN_CHALLENGED won't affect already 101configured VLANs. 102 103 * ndo_set_features: 104 105Hardware should be reconfigured to match passed feature set. The set 106should not be altered unless some error condition happens that can't 107be reliably detected in ndo_fix_features. In this case, the callback 108should update netdev->features to match resulting hardware state. 109Errors returned are not (and cannot be) propagated anywhere except dmesg. 110(Note: successful return is zero, >0 means silent error.) 111 112 113 114Part IV: Features 115================= 116 117For current list of features, see include/linux/netdev_features.h. 118This section describes semantics of some of them. 119 120 * Transmit checksumming 121 122For complete description, see comments near the top of include/linux/skbuff.h. 123 124Note: NETIF_F_HW_CSUM is a superset of NETIF_F_IP_CSUM + NETIF_F_IPV6_CSUM. 125It means that device can fill TCP/UDP-like checksum anywhere in the packets 126whatever headers there might be. 127 128 * Transmit TCP segmentation offload 129 130NETIF_F_TSO_ECN means that hardware can properly split packets with CWR bit 131set, be it TCPv4 (when NETIF_F_TSO is enabled) or TCPv6 (NETIF_F_TSO6). 132 133 * Transmit UDP segmentation offload 134 135NETIF_F_GSO_UDP_L4 accepts a single UDP header with a payload that exceeds 136gso_size. On segmentation, it segments the payload on gso_size boundaries and 137replicates the network and UDP headers (fixing up the last one if less than 138gso_size). 139 140 * Transmit DMA from high memory 141 142On platforms where this is relevant, NETIF_F_HIGHDMA signals that 143ndo_start_xmit can handle skbs with frags in high memory. 144 145 * Transmit scatter-gather 146 147Those features say that ndo_start_xmit can handle fragmented skbs: 148NETIF_F_SG --- paged skbs (skb_shinfo()->frags), NETIF_F_FRAGLIST --- 149chained skbs (skb->next/prev list). 150 151 * Software features 152 153Features contained in NETIF_F_SOFT_FEATURES are features of networking 154stack. Driver should not change behaviour based on them. 155 156 * VLAN challenged 157 158NETIF_F_VLAN_CHALLENGED should be set for devices which can't cope with VLAN 159headers. Some drivers set this because the cards can't handle the bigger MTU. 160[FIXME: Those cases could be fixed in VLAN code by allowing only reduced-MTU 161VLANs. This may be not useful, though.] 162 163* rx-fcs 164 165This requests that the NIC append the Ethernet Frame Checksum (FCS) 166to the end of the skb data. This allows sniffers and other tools to 167read the CRC recorded by the NIC on receipt of the packet. 168 169* rx-all 170 171This requests that the NIC receive all possible frames, including errored 172frames (such as bad FCS, etc). This can be helpful when sniffing a link with 173bad packets on it. Some NICs may receive more packets if also put into normal 174PROMISC mode. 175 176* rx-gro-hw 177 178This requests that the NIC enables Hardware GRO (generic receive offload). 179Hardware GRO is basically the exact reverse of TSO, and is generally 180stricter than Hardware LRO. A packet stream merged by Hardware GRO must 181be re-segmentable by GSO or TSO back to the exact original packet stream. 182Hardware GRO is dependent on RXCSUM since every packet successfully merged 183by hardware must also have the checksum verified by hardware. 184 185* hsr-tag-ins-offload 186 187This should be set for devices which insert an HSR (High-availability Seamless 188Redundancy) or PRP (Parallel Redundancy Protocol) tag automatically. 189 190* hsr-tag-rm-offload 191 192This should be set for devices which remove HSR (High-availability Seamless 193Redundancy) or PRP (Parallel Redundancy Protocol) tags automatically. 194 195* hsr-fwd-offload 196 197This should be set for devices which forward HSR (High-availability Seamless 198Redundancy) frames from one port to another in hardware. 199 200* hsr-dup-offload 201 202This should be set for devices which duplicate outgoing HSR (High-availability 203Seamless Redundancy) or PRP (Parallel Redundancy Protocol) frames 204automatically in hardware. 205 206Part V: Related device flags 207============================ 208 209* netdev->netmem_tx 210 211This is not a netdev feature bit. Drivers support netmem TX by setting 212netdev->netmem_tx to one of the values in enum netmem_tx_mode. 213See Documentation/networking/netmem.rst. 214