- 15 Apr, 2019 24 commits
-
-
David S. Miller authored
Simon Horman says: ==================== nfp: Flower flow merging John Hurley says, These patches deal with 'implicit recirculation' on the NFP. This is a firmware feature whereby a packet egresses to an 'internal' port meaning that it will recirculate back to the header extract phase with the 'internal' port now marked as its ingress port. This internal port can then be matched on by another rule. This process simulates how OvS datapath outputs to an internal port. The FW traces the packet's recirculation route and sends a 'merge hint' to the driver telling it which flows it matched against. The driver can then decide if these flows can be merged to a single rule and offloaded. The patches deal with the following issues: - assigning/freeing IDs to/from each of these new internal ports - offloading rules that match on internal ports - offloading neighbour table entries whose egress port is internal - handling fallback traffic with an internal port as ingress - using merge hints to create 'faster path' flows and tracking stats etc. ==================== Signed-off-by:David S. Miller <davem@davemloft.net>
-
John Hurley authored
A merge flow is formed from 2 sub flows. The match fields of the merge are the same as the first sub flow that has formed it, with the actions being a combination of the first and second sub flow. Therefore, a merge flow should replace sub flow 1 when offloaded. Offload valid merge flows by using a new 'flow mod' message type to replace an existing offloaded rule. Track the deletion of sub flows that are linked to a merge flow and revert offloaded merge rules if required. Signed-off-by:
John Hurley <john.hurley@netronome.com> Signed-off-by:
Simon Horman <simon.horman@netronome.com> Signed-off-by:
-
John Hurley authored
With the merging of 2 sub flows, a new 'merge' flow will be created and written to FW. The TC layer is unaware that the merge flow exists and will request stats from the sub flows. Conversely, the FW treats a merge rule the same as any other rule and sends stats updates to the NFP driver. Add links between merge flows and their sub flows. Use these links to pass merge flow stats updates from FW to the underlying sub flows, ensuring TC stats requests are handled correctly. The updating of sub flow stats is done on (the less time critcal) TC stats requests rather than on FW stats update. Signed-off-by:
-
John Hurley authored
When combining 2 sub_flows to a single 'merge flow' (assuming the merge is valid), the merge flow should contain the same match fields as sub_flow 1 with actions derived from a combination of sub_flows 1 and 2. This action list should have all actions from sub_flow 1 with the exception of the output action that triggered the 'implicit recirculation' by sending to an internal port, followed by all actions of sub_flow 2. Any pre-actions in either sub_flow should feature at the start of the action list. Add code to generate a new merge flow and populate the match and actions fields based on the sub_flows. The offloading of the flow is left to future patches. Signed-off-by:
-
John Hurley authored
Two flows can be merged if the second flow (after recirculation) matches on bits that are either matched on or explicitly set by the first flow. This means that if a packet hits flow 1 and recirculates then it is guaranteed to hit flow 2. Add a 'can_merge' function that determines if 2 sub_flows in a merge hint can be validly merged to a single flow. Signed-off-by:
-
John Hurley authored
If a merge hint is received containing 2 flows that are matched via an implicit recirculation (sending to and matching on an internal port), fw reports that the flows (called sub_flows) may be able to be combined to a single flow. Add infastructure to accept and process merge hint messages. The actual merging of the flows is left as a stub call. Signed-off-by:
-
John Hurley authored
Each flow is given a context ID that the fw uses (along with its cookie) to identity the flow. The flows stats are updated by the fw via this ID which is a reference to a pre-allocated array entry. In preparation for flow merge code, enable the nfp_fl_payload structure to be accessed via this stats context ID. Rather than increasing the memory requirements of the pre-allocated array, add a new rhashtable to associate each active stats context ID with its rule payload. While adding new code to the compile metadata functions, slightly restructure the existing function to allow for cleaner, easier to read error handling. Signed-off-by:
-
John Hurley authored
The neighbour table in the FW only accepts next hop entries if the egress port is an nfp repr. Modify this to allow the next hop to be an internal port. This means that if a packet is to egress to that port, it will recirculate back into the system with the internal port becoming its ingress port. Signed-off-by:
-
John Hurley authored
FW may receive a packet with its ingress port marked as an internal port. If a rule does not exist to match on this port, the packet will be sent to the NFP driver. Modify the flower app to detect packets from such internal ports and convert the ingress port to the correct kernel space netdev. At this point, it is assumed that fallback packets from internal ports are to be sent out said port. Therefore, set the redir_egress bool to true on detection of these ports. Signed-off-by:
-
John Hurley authored
Currently, it is assumed that fallback packets will be from reprs. Modify this to allow an app to receive non-repr ports from the fallback channel - e.g. from an internal port. If such a packet is received, do not update repr stats. Change the naming function calls so as not to imply it will always be a repr netdev returned. Add the option to set a bool value to redirect a fallback packet out the returned port rather than RXing it. Setting of this bool in subsequent patches allows the handling of packets falling back when they are due to egress an internal port. Signed-off-by:
Jakub Kicinski <jakub.kicinski@netronome.com> Signed-off-by:
-
John Hurley authored
Recent FW modifications allow the offloading of non repr ports. These ports exist internally on the NFP. So if a rule outputs to an 'internal' port, then the packet will recirculate back into the system but will now have this internal port as it's incoming port. These ports are indicated by a specific type field combined with an 8 bit port id. Add private app data to assign additional port ids for use in offloads. Provide functions to lookup or create new ids when a rule attempts to match on an internal netdev - the only internal netdevs currently supported are of type openvswitch. Have a netdev notifier to release port ids on netdev unregister. OvS offloads rules that match on internal ports as TC egress filters. Ensure that such rules are accepted by the driver. Signed-off-by:
-
John Hurley authored
Write to a FW symbol to indicate that the driver supports flow merging. If this symbol does not exist then flow merging and recirculation is not supported on the FW. If support is available, add a stub to deal with FW to kernel merge hint messages. Full flow merging requires the firmware to support of flow mods. If it does not, then do not attempt to 'turn on' flow merging. Signed-off-by:
-
David S. Miller authored
Huazhong Tan says: ==================== net: hns3: fixes sparse: warning and type error This patchset fixes a sparse warning and a overflow problem. ==================== Signed-off-by: -
Yunsheng Lin authored
When setting vport->bw_limit to hdev->tm_info.pg_info[0].bw_limit in hclge_tm_vport_tc_info_update, vport->bw_limit can be as big as HCLGE_ETHER_MAX_RATE (100000), which can not fit into u16 (65535). So this patch fixes it by using u32 for vport->bw_limit. Fixes: 84844054 ("net: hns3: Add support of TX Scheduler & Shaper to HNS3 driver") Reported-by:
Dan Carpenter <dan.carpenter@oracle.com> Signed-off-by:
Yunsheng Lin <linyunsheng@huawei.com> Signed-off-by:
Huazhong Tan <tanhuazhong@huawei.com> Signed-off-by:
-
Jian Shen authored
The input parameter "proto" in function hclge_set_vlan_filter_hw() is asked to be __be16, but got u16 when calling it in function hclge_update_port_base_vlan_cfg(). This patch fixes it by converting it with htons(). Reported-by:
kbuild test robot <lkp@intel.com> Fixes: 21e043cd ("net: hns3: fix set port based VLAN for PF") Signed-off-by:
Jian Shen <shenjian15@huawei.com> Signed-off-by:
-
David S. Miller authored
Xin Long says: ==================== sctp: fully support memory accounting sctp memory accounting is added in this patchset by using these kernel APIs on send side: - sk_mem_charge() - sk_mem_uncharge() - sk_wmem_schedule() - sk_under_memory_pressure() - sk_mem_reclaim() and these on receive side: - sk_mem_charge() - sk_mem_uncharge() - sk_rmem_schedule() - sk_under_memory_pressure() - sk_mem_reclaim() With sctp memory accounting, we can limit the memory allocation by either sysctl: # sysctl -w net.sctp.sctp_mem="10 20 50" or cgroup: # echo $((8<<14)) > \ /sys/fs/cgroup/memory/sctp_mem/memory.kmem.tcp.limit_in_bytes When the socket is under memory pressure, the send side will block and wait, while the receive side will renege or drop. v1->v2: - add the missing Reported/Tested/Acked/-bys. ==================== Signed-off-by: -
Xin Long authored
sk_forward_alloc's updating is also done on rx path, but to be consistent we change to use sk_mem_charge() in sctp_skb_set_owner_r(). In sctp_eat_data(), it's not enough to check sctp_memory_pressure only, which doesn't work for mem_cgroup_sockets_enabled, so we change to use sk_under_memory_pressure(). When it's under memory pressure, sk_mem_reclaim() and sk_rmem_schedule() should be called on both RENEGE or CHUNK DELIVERY path exit the memory pressure status as soon as possible. Note that sk_rmem_schedule() is using datalen to make things easy there. Reported-by:
Matteo Croce <mcroce@redhat.com> Tested-by:
Neil Horman <nhorman@tuxdriver.com> Acked-by:
Marcelo Ricardo Leitner <marcelo.leitner@gmail.com> Signed-off-by:
Xin Long <lucien.xin@gmail.com> Signed-off-by:
-
Xin Long authored
Now when sending packets, sk_mem_charge() and sk_mem_uncharge() have been used to set sk_forward_alloc. We just need to call sk_wmem_schedule() to check if the allocated should be raised, and call sk_mem_reclaim() to check if the allocated should be reduced when it's under memory pressure. If sk_wmem_schedule() returns false, which means no memory is allowed to allocate, it will block and wait for memory to become available. Note different from tcp, sctp wait_for_buf happens before allocating any skb, so memory accounting check is done with the whole msg_len before it too. Reported-by:
-
David S. Miller authored
Ido Schimmel says: ==================== mlxsw: Add neighbour offload indication Neighbour entries are programmed to the device's table so that the correct destination MAC will be specified in a packet after it was routed. Despite being programmed to the device and unlike routes and FDB entries, neighbour entries are currently not marked as offloaded. This patchset changes that. Patch #1 is a preparatory patch to make sure we only mark a neighbour as offloaded in case it was successfully programmed to the device. Patch #2 sets the offload indication on neighbours. Patch #3 adds a test to verify above mentioned functionality. Patched iproute2 version that prints the offload indication is available here [1]. [1] https://github.com/idosch/iproute2/tree/idosch-next ==================== Signed-off-by:
-
Ido Schimmel authored
Test that neighbour entries are marked as offloaded. Signed-off-by:
Ido Schimmel <idosch@mellanox.com> Signed-off-by:
-
Ido Schimmel authored
In a similar fashion to routes and FDB entries, the neighbour table is reflected to the device. Set an offload indication on the neighbour in case it was programmed to the device. Signed-off-by:
Jiri Pirko <jiri@mellanox.com> Reviewed-by:
David Ahern <dsahern@gmail.com> Signed-off-by:
-
Ido Schimmel authored
Next patch will add offload indication to neighbours, but the indication should only be altered in case the neighbour was successfully added to / deleted from the device. Propagate neighbour update errors, so that they could be taken into account by the next patch. Signed-off-by:
-
git://git.kernel.org/pub/scm/linux/kernel/git/pablo/nf-nextDavid S. Miller authored
Pablo Neira Ayuso says: ==================== Netfilter/IPVS updates for net-next The following patchset contains Netfilter updates for net-next: 1) Remove the broute pseudo hook, implement this from the bridge prerouting hook instead. Now broute becomes real table in ebtables, from Florian Westphal. This also includes a size reduction patch for the bridge control buffer area via squashing boolean into bitfields and a selftest. 2) Add OS passive fingerprint version matching, from Fernando Fernandez. 3) Support for gue encapsulation for IPVS, from Jacky Hu. 4) Add support for NAT to the inet family, from Florian Westphal. This includes support for masquerade, redirect and nat extensions. 5) Skip interface lookup in flowtable, use device in the dst object. 6) Add jiffies64_to_msecs() and use it, from Li RongQing. 7) Remove unused parameter in nf_tables_set_desc_parse(), from Colin Ian King. 8) Statify several functions, patches from YueHaibing and Florian Westphal. 9) Add an optimized version of nf_inet_addr_cmp(), from Li RongQing. 10) Merge route extension to core, also from Florian. 11) Use IS_ENABLED(CONFIG_NF_NAT) instead of NF_NAT_NEEDED, from Florian. 12) Merge ip/ip6 masquerade extensions, from Florian. This includes netdevice notifier unification. ==================== Signed-off-by: -
Stephen Rothwell authored
After merging the netfilter-next tree, today's linux-next build (powerpc ppc44x_defconfig) failed like this: In file included from net/bridge/br_input.c:19: include/net/netfilter/nf_queue.h:16:23: error: field 'state' has incomplete type struct nf_hook_state state; ^~~~~ Fixes: 971502d7 ("bridge: netfilter: unroll NF_HOOK helper in bridge input path") Signed-off-by:Stephen Rothwell <sfr@canb.auug.org.au> Signed-off-by:
Pablo Neira Ayuso <pablo@netfilter.org>
-
- 14 Apr, 2019 16 commits
-
-
Heiner Kallweit authored
This check isn't really needed and we can simplify the code and save some CPU cycles by removing it. Only in case of an error none of these bits are set, and calling the NAPI callback doesn't hurt in this case. Signed-off-by:
Heiner Kallweit <hkallweit1@gmail.com> Signed-off-by:
-
David S. Miller authored
Heiner Kallweit says: ==================== r8169: create function pointer arrays for PHY and chip hw init functions Using function pointer arrays makes the code easier to read and better maintainable. AFAIK function pointer arrays cause some performance drawback due to Spectre mitigation, but we're not in a hot path. ==================== Signed-off-by: -
Heiner Kallweit authored
Using a function pointer array makes this easier to read and better maintainable. Signed-off-by:
-
Heiner Kallweit authored
Using a function pointer array makes this easier to read and better maintainable. AFAIK function pointer arrays cause some performance drawback due to Spectre mitigation, but we're not in a hot path here. Signed-off-by:
-
David S. Miller authored
Huazhong Tan says: ==================== code optimizations & bugfixes for HNS3 driver This patch-set includes code optimizations and bugfixes for the HNS3 ethernet controller driver. [patch 1/12 - 4/12] optimizes the VLAN freature and adds support for port based VLAN, fixes some related bugs about the current implementation. [patch 5/12 - 12/12] includes some other code optimizations for the HNS3 ethernet controller driver. Change log: V1->V2: modifies some patches' commint log and code. ==================== Signed-off-by: -
Peng Li authored
This patch removes some redundant BH disable when initializing and uninitializing command queue. Signed-off-by:
Peng Li <lipeng321@huawei.com> Signed-off-by:
-
Peng Li authored
If there is pending skb in RX flow when close the port, and the pending buffer is not cleaned, the new packet will be added to the pending skb when the port opens again, and the first new packet has error data. This patch cleans the pending skb when clean RX ring. Signed-off-by:
-
Jian Shen authored
For some cases, PHY may not be connected to MDIO bus, then the driver will initialize fail since MDIO bus initialization fails. This patch fixes it by skipping the MDIO bus initialization when PHY is inexistent. Signed-off-by:
-
Weihang Li authored
According to hardware description, reset level that should be triggered are not consistent in a module. For example, in SSU common errors, the first two bits has no need to do reset, but the other bits need global reset. This patch sets separate reset level for all RAS and MSI-X interrupts by adding a reset_lvel field in struct hclge_hw_error, and fixes some incorrect reset level. Signed-off-by:
Weihang Li <liweihang@hisilicon.com> Signed-off-by:
-
Yunsheng Lin authored
Currently hardware may have not enough buffer to receive packet when it has used more than two MPS(maximum packet size) of buffer, but there are still a lot of shared buffer left unused when TC num is small. This patch divides shared buffer to be used between TC when the port supports DCB, and adjusts the waterline and threshold according to user manual for the port that does not support DCB. This patch also change hclge_get_tc_num's return type to u32 to avoid signed-unsigned mix with divide. Signed-off-by:
-
Yunsheng Lin authored
Currently RX shared buffer' threshold size for speific TC is set to smaller value when the TC's PFC is not enabled, which may cause performance problem because hardware may not have enough hardware buffer when PFC is not enabled. This patch sets the same threshold size for all TC no matter if the specific TC's PFC is enabled. Signed-off-by:
-
Yunsheng Lin authored
When a GRO packet is received by driver, the cwr field in the struct tcphdr needs to be checked to decide whether to set the SKB_GSO_TCP_ECN for skb_shinfo(skb)->gso_type. So this patch adds hns3_gro_complete to do that, and adds the hns3_handle_bdinfo to handle the hns3_gro_complete and hns3_rx_checksum. Signed-off-by:
-
Yunsheng Lin authored
Change the parameters of hns3_rx_checksum to be more specific to what is used internally, rather than passing in a pointer to the whole hns3_desc. Reduces duplicate code and bring this function inline with the approach used in hns3_set_gro_param. Signed-off-by:
-
Jian Shen authored
In original codes, ndo_set_vf_vlan() in hns3 driver was implemented wrong. It adds or removes VLAN into VLAN filter for VF, but VF is unaware of it. This patch fixes it. When VF loads up, it firstly queries the port based VLAN state from PF. When user change port based VLAN state from PF, PF firstly checks whether the VF is alive. If the VF is alive, then PF notifies the VF the modification; otherwise PF configure the port based VLAN state directly. Fixes: 46a3df9f ("net: hns3: Add HNS3 Acceleration Engine & Compatibility Layer Support") Signed-off-by:
-
Jian Shen authored
In original codes, ndo_set_vf_vlan() in hns3 driver was implemented wrong. It adds or removes VLAN into VLAN filter for VF, but VF is unaware of it. Indeed, ndo_set_vf_vlan() is expected to enable or disable port based VLAN (hardware inserts a specified VLAN tag to all TX packets for a specified VF) . When enable port based VLAN, we use port based VLAN id as VLAN filter entry. When disable port based VLAN, we use VLAN id of VLAN device. This patch fixes it for PF, enable/disable port based VLAN when calls ndo_set_vf_vlan(). Fixes: 46a3df9f ("net: hns3: Add HNS3 Acceleration Engine & Compatibility Layer Support") Signed-off-by:
-
Jian Shen authored
Currently, in TX direction, driver implements the TX VLAN offload by checking the VLAN header in skb, and filling it into TX descriptor. Usually it works well, but if enable inserting VLAN header based on port, it may conflict when out_tag field of TX descriptor is already used, and cause RAS error. In RX direction, hardware supports stripping max two VLAN headers. For vlan_tci in skb can only store one VLAN tag, when RX VLAN offload enabled, driver tells hardware to strip one VLAN header from RX packet; when RX VLAN offload disabled, driver tells hardware not to strip VLAN header from RX packet. Now if port based insert VLAN enabled, all RX packets will have the port based VLAN header. This header is useless for stack, driver needs to ask hardware to strip it. Unfortunately, hardware can't drop this VLAN header, and always fill it into RX descriptor, so driver has to identify and drop it. Signed-off-by:
-
