- 16 Jul, 2018 5 commits
-
-
Maxime Chevallier authored
The classification operations that are used for RSS make use of several lookup tables. Having hit counters for these tables is really helpful to determine what flows were matched by ingress traffic, and see the path of packets among all the classifier tables. This commit adds hit counters for the 3 tables used at the moment : - The decoding table (also called lookup_id table), that links flows identified by the Header Parser to the flow table. There's one entry per flow, located at : .../mvpp2/<controller>/flows/XX/dec_hits Note that there are 21 flows in the decoding table, whereas there are 52 flows in the Header Parser. That's because there are several kind of traffic that will match a given flow. Reading the hit counter from one sub-flow will clear all hit counter that have the same flow_id. This also applies to the flow_hits. - The flow table, that contains all the different lookups to be performed by the classifier for each packet of a given flow. The match is done on the first entry of the flow sequence. - The C2 engine entries, that are used to assign the default rx queue, and enable or disable RSS for a given port. There's one entry per flow, located at: .../mvpp2/<controller>/flows/XX/flow_hits There is one C2 entry per port, so the c2 hit counter is located at : .../mvpp2/<controller>/ethX/c2_hits All hit counter values are 16-bits clear-on-read values. Signed-off-by:
Maxime Chevallier <maxime.chevallier@bootlin.com> Signed-off-by:
David S. Miller <davem@davemloft.net>
-
Maxime Chevallier authored
The classifier configuration for RSS is quite complex, with several lookup tables being used. This commit adds useful info in debugfs to see how the different tables are configured : Added 2 new entries in the per-port directory : - .../eth0/default_rxq : The default rx queue on that port - .../eth0/rss_enable : Indicates if RSS is enabled in the C2 entry Added the 'flows' directory : It contains one entry per sub-flow. a 'sub-flow' is a unique path from Header Parser to the flow table. Multiple sub-flows can point to the same 'flow' (each flow has an id from 8 to 29, which is its index in the Lookup Id table) : - .../flows/00/... /01/... ... /51/id : The flow id. There are 21 unique flows. There's one flow per combination of the following parameters : - L4 protocol (TCP, UDP, none) - L3 protocol (IPv4, IPv6) - L3 parameters (Fragmented or not) - L2 parameters (Vlan tag presence or not) .../type : The flow type. This is an even higher level flow, that we manipulate with ethtool. It can be : "udp4" "tcp4" "udp6" "tcp6" "ipv4" "ipv6" "other". .../eth0/... .../eth1/engine : The hash generation engine used for this flow on the given port .../hash_opts : The hash generation options indicating on what data we base the hash (vlan tag, src IP, src port, etc.) Signed-off-by: -
Maxime Chevallier authored
One helpful feature to help debug the Header Parser TCAM filter in PPv2 is to be able to see if the entries did match something when a packet comes in. This can be done by using the built-in hit counter for TCAM entries. This commit implements reading the counter, and exposing its value on debugfs for each filter entry. The counter is a 16-bits clear-on-read value, located at: .../mvpp2/<controller>/parser/XXX/hits Signed-off-by:
-
Maxime Chevallier authored
Marvell PPv2 Packer Header Parser has a TCAM based filter, that is not trivial to configure and debug. Being able to dump TCAM entries from userspace can be really helpful to help development of new features and debug existing ones. This commit adds a basic debugfs interface for the PPv2 driver, focusing on TCAM related features. <mnt>/mvpp2/ --- f2000000.ethernet \- f4000000.ethernet --- parser --- 000 ... | \- 001 | \- ... | \- 255 --- ai | \- header_data | \- lookup_id | \- sram | \- valid \- eth1 ... \- eth2 --- mac_filter \- parser_entries \- vid_filter There's one directory per PPv2 instance, named after pdev->name to make sure names are uniques. In each of these directories, there's : - one directory per interface on the controller, each containing : - "mac_filter", which lists all filtered addresses for this port (based on TCAM, not on the kernel's uc / mc lists) - "parser_entries", which lists the indices of all valid TCAM entries that have this port in their port map - "vid_filter", which lists the vids allowed on this port, based on TCAM - one "parser" directory (the parser is common to all ports), containing : - one directory per TCAM entry (256 of them, from 0 to 255), each containing : - "ai" : Contains the 1 byte Additional Info field from TCAM, and - "header_data" : Contains the 8 bytes Header Data extracted from the packet - "lookup_id" : Contains the 4 bits LU_ID - "sram" : contains the raw SRAM data, which is the result of the TCAM lookup. This readonly at the moment. - "valid" : Indicates if the entry is valid of not. All entries are read-only, and everything is output in hex form. Signed-off-by: -
Antoine Tenart authored
Use the appropriate SPDX license identifiers and drop the license text. This patch is only cosmetic. Signed-off-by:
Antoine Tenart <antoine.tenart@bootlin.com> Signed-off-by:
-
- 15 Jul, 2018 1 commit
-
-
git://git.kernel.org/pub/scm/linux/kernel/git/bpf/bpf-nextDavid S. Miller authored
Daniel Borkmann says: ==================== pull-request: bpf-next 2018-07-15 The following pull-request contains BPF updates for your *net-next* tree. The main changes are: 1) Various different arm32 JIT improvements in order to optimize code emission and make the JIT code itself more robust, from Russell. 2) Support simultaneous driver and offloaded XDP in order to allow for advanced use-cases where some work is offloaded to the NIC and some to the host. Also add ability for bpftool to load programs and maps beyond just the cgroup case, from Jakub. 3) Add BPF JIT support in nfp for multiplication as well as division. For the latter in particular, it uses the reciprocal algorithm to emulate it, from Jiong. 4) Add BTF pretty print functionality to bpftool in plain and JSON output format, from Okash. 5) Add build and installation to the BPF helper man page into bpftool, from Quentin. 6) Add a TCP BPF callback for listening sockets which is triggered right after the socket transitions to TCP_LISTEN state, from Andrey. 7) Add a new cgroup tree command to bpftool which iterates over the whole cgroup tree and prints all attached programs, from Roman. 8) Improve xdp_redirect_cpu sample to support parsing of double VLAN tagged packets, from Jesper. ==================== Signed-off-by:
-
- 14 Jul, 2018 20 commits
-
-
Daniel Borkmann authored
Andrey Ignatov says: ==================== This patchset adds TCP-BPF callback for listening sockets. Patch 0001 provides more details and is the main patch in the set. Patch 0006 adds selftest for the new callback. Other patches are bug fixes and improvements in TCP-BPF selftest to make it easier to extend in 0006. ==================== Acked-by:
Lawrence Brakmo <brakmo@fb.com> Signed-off-by:
Daniel Borkmann <daniel@iogearbox.net>
-
Andrey Ignatov authored
Cover new TCP-BPF callback in test_tcpbpf: when listen() is called on socket, set BPF_SOCK_OPS_STATE_CB_FLAG so that BPF_SOCK_OPS_STATE_CB callback can be called on future state transition, and when such a transition happens (TCP_LISTEN -> TCP_CLOSE), track it in the map and verify it in user space later. Signed-off-by:
Andrey Ignatov <rdna@fb.com> Acked-by:
Alexei Starovoitov <ast@kernel.org> Signed-off-by:
-
Andrey Ignatov authored
Reduce amount of copy/paste for debug info when result is verified in the test and keep that info together with values being checked so that they won't get out of sync. It also improves debug experience: instead of checking manually what doesn't match in debug output for all fields, only unexpected field is printed. Signed-off-by:
-
Andrey Ignatov authored
Switch to cgroup_helpers to simplify the code and fix cgroup cleanup: before cgroup was not cleaned up after the test. It also removes SYSTEM macro, that only printed error, but didn't terminate the test. Signed-off-by:
-
Andrey Ignatov authored
Lack of const in cgroup helpers signatures forces to write ugly client code. Fix it. Signed-off-by:
-
Andrey Ignatov authored
Sync BPF_SOCK_OPS_TCP_LISTEN_CB related UAPI changes to tools/. Signed-off-by:
-
Andrey Ignatov authored
Add new TCP-BPF callback that is called on listen(2) right after socket transition to TCP_LISTEN state. It fills the gap for listening sockets in TCP-BPF. For example BPF program can set BPF_SOCK_OPS_STATE_CB_FLAG when socket becomes listening and track later transition from TCP_LISTEN to TCP_CLOSE with BPF_SOCK_OPS_STATE_CB callback. Before there was no way to do it with TCP-BPF and other options were much harder to work with. E.g. socket state tracking can be done with tracepoints (either raw or regular) but they can't be attached to cgroup and their lifetime has to be managed separately. Signed-off-by:
-
David S. Miller authored
Ido Schimmel says: ==================== mlxsw: Add VRRP support When a router that is acting as the default gateway of a host stops functioning, the host will encounter packet loss until the router starts functioning again. To increase the reliability of the default gateway without performing reconfiguration on the host, a host can use a Virtual Router Redundancy Protocol (VRRP) Router. This virtual router is composed from several routers where only one is actually forwarding packets from the host (the master router) while the other routers act as backup routers. The election of the master router is determined by the VRRP protocol [1]. Packets addressed to the virtual router are always sent to the virtual router MAC address (IPv4: 00-00-5E-00-01-XX, IPv6: 00-00-5E-00-02-XX). Such packets can only be accepted by the master router and must be discarded by the backup routers. In Linux, VRRP is usually implemented by configuring a macvlan with the virtual router MAC on top of the router interface that is connected to the host / LAN. The macvlan on the master router is assigned the virtual IP (VIP) that the host uses as its gateway. In order to support VRRP in mlxsw, we first need to enable macvlan upper devices on top of mlxsw netdevs and their uppers. This is done by the first patch, which also takes care of sanitizing macvlan configurations that are not currently supported by the driver. The second patch directs packets with destination MAC addresses as the macvlans to the router so that they will undergo an L3 lookup. This is consistent with the kernel's behavior where the macvlan's Rx handler will re-inject such packets to the Rx path so that they will be picked up by the IPvX protocol handlers and undergo an L3 lookup. Note that the driver prevents the macvlans from being enslaved to other devices, to ensure the packets will be picked up by the protocol handler and not by another Rx handler. The third patch adds packet traps for VRRP control packets for both IPv4 and IPv6. Finally, the last patch optimizes the reception of VRRP MACs by potentially skipping one L2 lookup for them. ==================== Signed-off-by: -
Ido Schimmel authored
Hosts using a VRRP router send their packets with a destination MAC of the VRRP router which is of the following form [1]: IPv4 - 00-00-5E-00-01-{VRID} IPv6 - 00-00-5E-00-02-{VRID} Where VRID is the ID of the virtual router. Such packets are directed to the router block in the ASIC by an FDB entry that was added in the previous patch. However, in certain cases it is possible to skip this FDB lookup and send such packets directly to the router. This is accomplished by adding these special MAC addresses to the RIF cache. If the cache is hit, the packet will skip the L2 lookup and ingress the router with the RIF specified in the cache entry. 1. https://tools.ietf.org/html/rfc5798#section-7.3Signed-off-by:Ido Schimmel <idosch@mellanox.com> Reviewed-by:
Petr Machata <petrm@mellanox.com> Signed-off-by:
-
Ido Schimmel authored
Virtual Router Redundancy Protocol packets are used to communicate the state of the Master router associated with the virtual router ID (VRID). These are link-local multicast packets sent with IP protocol 112 that are trapped in the router block in the ASIC. Add a trap for these packets and mark the trapped packets to prevent them from potentially being re-flooded by the bridge driver. Signed-off-by:
-
Ido Schimmel authored
An IP packet received on a netdev with a macvlan upper whose MAC matches the packet's destination MAC will be re-injected to the Rx path as if it was received by the macvlan, and perform an L3 lookup. Reflect this functionality to the ASIC by programming FDB entries that will direct MACs of macvlan uppers to the router. In a similar fashion to router interfaces (RIFs) that are programmed upon the addition of the first IP address on an interface and destroyed upon the removal of the last IP address, the FDB entries for the macvlan are added and destroyed based on the addition of the first and removal of the last IP address on the macvlan. Signed-off-by:
-
Ido Schimmel authored
In order to allow more unicast MAC addresses (e.g., VRRP virtual MAC) to be directed to the router we need to enable macvlan uppers on top of mlxsw netdevs. Allow macvlan upper devices on top of mlxsw netdevs and sanitize configurations that can't work. For example, a macvlan can't be enslaved to a bridge as without ACLs the device doesn't take the destination MAC into account when classifying a packet to a bridge instance (i.e., a FID). Signed-off-by:
-
Yafang Shao authored
tcp_rcv_nxt_update() is already executed in tcp_data_queue(). This line is redundant. See bellow, tcp_queue_rcv tcp_rcv_nxt_update(tcp_sk(sk), TCP_SKB_CB(skb)->end_seq); tcp_rcv_nxt_update(tp, TCP_SKB_CB(skb)->end_seq); <<<< redundant Signed-off-by:
Yafang Shao <laoar.shao@gmail.com> Signed-off-by:
Eric Dumazet <edumazet@google.com> Signed-off-by:
-
Okash Khawaja authored
This patch augments the output of bpftool's map dump and map lookup commands to print data along side btf info, if the correspondin btf info is available. The outputs for each of map dump and map lookup commands are augmented in two ways: 1. when neither of -j and -p are supplied, btf-ful map data is printed whose aim is human readability. This means no commitments for json- or backward- compatibility. 2. when either -j or -p are supplied, a new json object named "formatted" is added for each key-value pair. This object contains the same data as the key-value pair, but with btf info. "formatted" object promises json- and backward- compatibility. Below is a sample output. $ bpftool map dump -p id 8 [{ "key": ["0x0f","0x00","0x00","0x00" ], "value": ["0x03", "0x00", "0x00", "0x00", ... ], "formatted": { "key": 15, "value": { "int_field": 3, ... } } } ] This patch calls btf_dumper introduced in previous patch to accomplish the above. Indeed, btf-ful info is only displayed if btf data for the given map is available. Otherwise existing output is displayed as-is. Signed-off-by:Okash Khawaja <osk@fb.com> Acked-by:
Martin KaFai Lau <kafai@fb.com> Reviewed-by:
Jakub Kicinski <jakub.kicinski@netronome.com> Signed-off-by:
-
Okash Khawaja authored
This consumes functionality exported in the previous patch. It does the main job of printing with BTF data. This is used in the following patch to provide a more readable output of a map's dump. It relies on json_writer to do json printing. Below is sample output where map keys are ints and values are of type struct A: typedef int int_type; enum E { E0, E1, }; struct B { int x; int y; }; struct A { int m; unsigned long long n; char o; int p[8]; int q[4][8]; enum E r; void *s; struct B t; const int u; int_type v; unsigned int w1: 3; unsigned int w2: 3; }; $ sudo bpftool map dump id 14 [{ "key": 0, "value": { "m": 1, "n": 2, "o": "c", "p": [15,16,17,18,15,16,17,18 ], "q": [[25,26,27,28,25,26,27,28 ],[35,36,37,38,35,36,37,38 ],[45,46,47,48,45,46,47,48 ],[55,56,57,58,55,56,57,58 ] ], "r": 1, "s": 0x7ffd80531cf8, "t": { "x": 5, "y": 10 }, "u": 100, "v": 20, "w1": 0x7, "w2": 0x3 } } ] This patch uses json's {} and [] to imply struct/union and array. More explicit information can be added later. For example, a command line option can be introduced to print whether a key or value is struct or union, name of a struct etc. This will however come at the expense of duplicating info when, for example, printing an array of structs. enums are printed as ints without their names. Signed-off-by: -
Okash Khawaja authored
This patch introduces btf__resolve_type() function and exports two existing functions from libbpf. btf__resolve_type follows modifier types like const and typedef until it hits a type which actually takes up memory, and then returns it. This function follows similar pattern to btf__resolve_size but instead of computing size, it just returns the type. These functions will be used in the followig patch which parses information inside array of `struct btf_type *`. btf_name_by_offset is used for printing variable names. Signed-off-by:
Song Liu <songliubraving@fb.com> Reviewed-by:
-
Jakub Kicinski authored
perf propagates its feature check results to libbpf. This means features for which perf probes must be a superset of libbpf's required features. perf depends on FEATURE_TESTS_BASIC for its list of features. commit 531b014e ("tools: bpf: make use of reallocarray") added reallocarray use to libbpf, make perf also perform the reallocarray feature check. Fixes: 531b014e ("tools: bpf: make use of reallocarray") Reported-by:
Guenter Roeck <linux@roeck-us.net> Signed-off-by:
-
kbuild test robot authored
Fixes: f9358e12 ("net: mvpp2: split ingress traffic into multiple flows") Signed-off-by:
kbuild test robot <fengguang.wu@intel.com> Signed-off-by:
-
Linus Walleij authored
By a simple extension of of_phy_get_and_connect() drivers that have a fixed link on e.g. RGMII can support also fixed links, so in addition to: ethernet-port { phy-mode = "rgmii"; phy-handle = <&foo>; }; This setup with a fixed-link node and no phy-handle will now also work just fine: ethernet-port { phy-mode = "rgmii"; fixed-link { speed = <1000>; full-duplex; pause; }; }; This is very helpful for connecting random ethernet ports to e.g. DSA switches that typically reside on fixed links. The phy-mode is still there as the fixes link in this case is still an RGMII link. Tested on the Cortina Gemini driver with the Vitesse DSA router chip on a fixed 1Gbit link. Suggested-by:Andrew Lunn <andrew@lunn.ch> Signed-off-by:
Linus Walleij <linus.walleij@linaro.org> Reviewed-by:
-
Vlad Buslov authored
Extend struct tcf_walker with additional 'cookie' field. It is intended to be used by classifier walk implementations to continue iteration directly from particular filter, instead of iterating 'skip' number of times. Change flower walk implementation to save filter handle in 'cookie'. Each time flower walk is called, it looks up filter with saved handle directly with idr, instead of iterating over filter linked list 'skip' number of times. This change improves complexity of dumping flower classifier from quadratic to linearithmic. (assuming idr lookup has logarithmic complexity) Reviewed-by:
Jiri Pirko <jiri@mellanox.com> Signed-off-by:
Vlad Buslov <vladbu@mellanox.com> Reported-by:
Simon Horman <simon.horman@netronome.com> Reviewed-by:
-
- 13 Jul, 2018 14 commits
-
-
Jesper Dangaard Brouer authored
People noticed that the code match on IEEE 802.1ad (ETH_P_8021AD) ethertype, and this implies Q-in-Q or double tagged VLANs. Thus, we better parse the next VLAN header too. It is even marked as a TODO. This is relevant for real world use-cases, as XDP cpumap redirect can be used when the NIC RSS hashing is broken. E.g. the ixgbe driver HW cannot handle double tagged VLAN packets, and places everything into a single RX queue. Using cpumap redirect, users can redistribute traffic across CPUs to solve this, which is faster than the network stacks RPS solution. It is left as an exerise how to distribute the packets across CPUs. It would be convenient to use the RX hash, but that is not _yet_ exposed to XDP programs. For now, users can code their own hash, as I've demonstrated in the Suricata code (where Q-in-Q is handled correctly). Reported-by:
Florian Maury <florian.maury-cv@x-cli.eu> Reported-by:
Marek Majkowski <marek@cloudflare.com> Signed-off-by:
Jesper Dangaard Brouer <brouer@redhat.com> Signed-off-by:
-
Nikolay Aleksandrov authored
This patch adds support for IGMPMSG_WRVIFWHOLE which is used to pass full packet and real vif id when the incoming interface is wrong. While the RP and FHR are setting up state we need to be sending the registers encapsulated with all the data inside otherwise we lose it. The RP then decapsulates it and forwards it to the interested parties. Currently with WRONGVIF we can only be sending empty register packets and will lose that data. This behaviour can be enabled by using MRT_PIM with val == IGMPMSG_WRVIFWHOLE. This doesn't prevent IGMPMSG_WRONGVIF from happening, it happens in addition to it, also it is controlled by the same throttling parameters as WRONGVIF (i.e. 1 packet per 3 seconds currently). Both messages are generated to keep backwards compatibily and avoid breaking someone who was enabling MRT_PIM with val == 4, since any positive val is accepted and treated the same. Signed-off-by:
Nikolay Aleksandrov <nikolay@cumulusnetworks.com> Signed-off-by:
-
Daniel Borkmann authored
Jakub Kicinski says: ==================== This set is adding support for loading driver and offload XDP at the same time. This enables advanced use cases where some of the work is offloaded to the NIC and some is done by the host. Separate netlink attributes are added for each mode of operation. Driver callbacks for offload are cleaned up a little, including removal of .prog_attached flag. ==================== Acked-by:
-
Jakub Kicinski authored
Split handling of offloaded and driver programs completely. Since offloaded programs always come with XDP_FLAGS_HW_MODE set in reality there could be no sharing, anyway, programs would only be installed in driver or in hardware. Splitting the handling allows us to install programs in HW and in driver at the same time. Signed-off-by:
Quentin Monnet <quentin.monnet@netronome.com> Signed-off-by:
-
Jakub Kicinski authored
Add tests for having an XDP program attached in the driver and another one attached in HW simultaneously. Signed-off-by:
-
Jakub Kicinski authored
Allow netdevsim to accept driver and offload attachment of XDP BPF programs at the same time. Signed-off-by:
-
Jakub Kicinski authored
Split the query of HW-attached program from the software one. Introduce new .ndo_bpf command to query HW-attached program. This will allow drivers to install different programs in HW and SW at the same time. Netlink can now also carry multiple programs on dump (in which case mode will be set to XDP_ATTACHED_MULTI and user has to check per-attachment point attributes, IFLA_XDP_PROG_ID will not be present). We reuse IFLA_XDP_PROG_ID skb space for second mode, so rtnl_xdp_size() doesn't need to be updated. Note that the installation side is still not there, since all drivers currently reject installing more than one program at the time. Signed-off-by:
-
Jakub Kicinski authored
Basic operations drivers perform during xdp setup and query can be moved to helpers in the core. Encapsulate program and flags into a structure and add helpers. Note that the structure is intended as the "main" program information source in the driver. Most drivers will additionally place the program pointer in their fast path or ring structures. The helpers don't have a huge impact now, but they will decrease the code duplication when programs can be installed in HW and driver at the same time. Encapsulating the basic operations in helpers will hopefully also reduce the number of changes to drivers which adopt them. Helpers could really be static inline, but they depend on definition of struct netdev_bpf which means they'd have to be placed in netdevice.h, an already 4500 line header. Signed-off-by:
-
Jakub Kicinski authored
prog_attached of struct netdev_bpf should have been superseded by simply setting prog_id long time ago, but we kept it around to allow offloading drivers to communicate attachment mode (drv vs hw). Subsequently drivers were also allowed to report back attachment flags (prog_flags), and since nowadays only programs attached will XDP_FLAGS_HW_MODE can get offloaded, we can tell the attachment mode from the flags driver reports. Remove prog_attached member. Signed-off-by:
-
Jakub Kicinski authored
In preparation for support of simultaneous driver and hardware XDP support add per-mode attributes. The catch-all IFLA_XDP_PROG_ID will still be reported, but user space can now also access the program ID in a new IFLA_XDP_<mode>_PROG_ID attribute. Signed-off-by:
-
Daniel Borkmann authored
Russell King says: ==================== Four further jit compiler improves for 32-bit ARM. ==================== Signed-off-by: -
Russell King authored
Improbe the 64-bit ALU implementation from: movw r8, #65532 movt r8, #65535 movw r9, #65535 movt r9, #65535 ldr r7, [fp, #-44] adds r7, r7, r8 str r7, [fp, #-44] ldr r7, [fp, #-40] adc r7, r7, r9 str r7, [fp, #-40] to: movw r8, #65532 movt r8, #65535 movw r9, #65535 movt r9, #65535 ldrd r6, [fp, #-44] adds r6, r6, r8 adc r7, r7, r9 strd r6, [fp, #-44] Signed-off-by:
Russell King <rmk+kernel@armlinux.org.uk> Signed-off-by:
-
Russell King authored
Improve the 64-bit store implementation from: ldr r6, [fp, #-8] str r8, [r6] ldr r6, [fp, #-8] mov r7, #4 add r7, r6, r7 str r9, [r7] to: ldr r6, [fp, #-8] str r8, [r6] str r9, [r6, #4] We leave the store as two separate STR instructions rather than using STRD as the store may not be aligned, and STR can handle misalignment. Signed-off-by:
-
Russell King authored
Improve the 64-bit sign-extended immediate from: mov r6, #1 str r6, [fp, #-52] ; 0xffffffcc mov r6, #0 str r6, [fp, #-48] ; 0xffffffd0 to: mov r6, #1 mov r7, #0 strd r6, [fp, #-52] ; 0xffffffcc Signed-off-by:
-
