- 28 Jun, 2021 10 commits
-
-
git://git.kernel.org/pub/scm/linux/kernel/gitDavid S. Miller authored
/klassert/ipsec-next Steffen Klassert says: ==================== pull request (net-next): ipsec-next 2021-06-28 1) Remove an unneeded error assignment in esp4_gro_receive(). From Yang Li. 2) Add a new byseq state hashtable to find acquire states faster. From Sabrina Dubroca. 3) Remove some unnecessary variables in pfkey_create(). From zuoqilin. 4) Remove the unused description from xfrm_type struct. From Florian Westphal. 5) Fix a spelling mistake in the comment of xfrm_state_ok(). From gushengxian. 6) Replace hdr_off indirections by a small helper function. From Florian Westphal. 7) Remove xfrm4_output_finish and xfrm6_output_finish declarations, they are not used anymore.From Antony Antony. 8) Remove xfrm replay indirections. From Florian Westphal. Please pull or let me know if there are problems. ==================== Signed-off-by:David S. Miller <davem@davemloft.net>
-
David S. Miller authored
Merge tag 'mac80211-next-for-net-next-2021-06-25' of git://git.kernel.org/pub/scm/linux/kernel/git/jberg/mac80211-next Johannes berg says: ==================== Lots of changes: * aggregation handling improvements for some drivers * hidden AP discovery on 6 GHz and other HE 6 GHz improvements * minstrel improvements for no-ack frames * deferred rate control for TXQs to improve reaction times * virtual time-based airtime scheduler * along with various little cleanups/fixups ==================== Signed-off-by:
-
Matthieu Baerts authored
Dan Carpenter reported an issue introduced in commit fde56eea ("mptcp: refine mptcp_cleanup_rbuf") where a new boolean (ack_pending) is masked with 0x9. This is not the intention to ignore values by using a boolean. This variable should not have a 'bool' type: we should keep the 'u8' to allow this comparison. Fixes: fde56eea ("mptcp: refine mptcp_cleanup_rbuf") Reported-by:
Dan Carpenter <dan.carpenter@oracle.com> Signed-off-by:
Matthieu Baerts <matthieu.baerts@tessares.net> Signed-off-by:
Mat Martineau <mathew.j.martineau@linux.intel.com> Acked-by:
Paolo Abeni <pabeni@redhat.com> Signed-off-by:
-
David S. Miller authored
Guillaume Nault says: ==================== net: reset MAC header consistently across L3 virtual devices Some virtual L3 devices, like vxlan-gpe and gre (in collect_md mode), reset the MAC header pointer after they parsed the outer headers. This accurately reflects the fact that the decapsulated packet is pure L3 packet, as that makes the MAC header 0 bytes long (the MAC and network header pointers are equal). However, many L3 devices only adjust the network header after decapsulation and leave the MAC header pointer to its original value. This can confuse other parts of the networking stack, like TC, which then considers the outer headers as one big MAC header. This patch series makes the following L3 tunnels behave like VXLAN-GPE: bareudp, ipip, sit, gre, ip6gre, ip6tnl, gtp. The case of gre is a bit special. It already resets the MAC header pointer in collect_md mode, so only the classical mode needs to be adjusted. However, gre also has a special case that expects the MAC header pointer to keep pointing to the outer header even after decapsulation. Therefore, patch 4 keeps an exception for this case. Ideally, we'd centralise the call to skb_reset_mac_header() in ip_tunnel_rcv(), to avoid manual calls in ipip (patch 2), sit (patch 3) and gre (patch 4). That's unfortunately not feasible currently, because of the gre special case discussed above that precludes us from resetting the MAC header unconditionally. The original motivation is to redirect bareudp packets to Ethernet devices (as described in patch 1). The rest of this series aims at bringing consistency across all L3 devices (apart from gre's special case unfortunately). Note: the gtp patch results from pure code inspection and has been compiled tested only. ==================== Signed-off-by: -
Guillaume Nault authored
For consistency with other L3 tunnel devices, reset the mac_header pointer after decapsulation. This makes the mac_header 0 bytes long, thus making it clear that this skb has no mac_header. Compile tested only. Signed-off-by:
Guillaume Nault <gnault@redhat.com> Signed-off-by:
-
Guillaume Nault authored
Reset the mac_header pointer even when the tunnel transports only L3 data (in the ARPHRD_ETHER case, this is already done by eth_type_trans). This prevents other parts of the stack from mistakenly accessing the outer header after the packet has been decapsulated. In practice, this allows to push an Ethernet header to ipip6, ip6ip6, mplsip6 or ip6gre packets and redirect them to an Ethernet device: $ tc filter add dev ip6tnl0 ingress matchall \ action vlan push_eth dst_mac 00:00:5e:00:53:01 \ src_mac 00:00:5e:00:53:00 \ action mirred egress redirect dev eth0 Without this patch, push_eth refuses to add an ethernet header because the skb appears to already have a MAC header. Signed-off-by: -
Guillaume Nault authored
Commit e271c7b4 ("gre: do not keep the GRE header around in collect medata mode") did reset the mac_header for the collect_md case. Let's extend this behaviour to classical gre devices as well. ipgre_header_parse() seems to be the only case that requires mac_header to point to the outer header. We can detect this case accurately by checking ->header_ops. For all other cases, we can reset mac_header. This allows to push an Ethernet header to ipgre packets and redirect them to an Ethernet device: $ tc filter add dev gre0 ingress matchall \ action vlan push_eth dst_mac 00:00:5e:00:53:01 \ src_mac 00:00:5e:00:53:00 \ action mirred egress redirect dev eth0 Before this patch, this worked only for collect_md gre devices. Now this works for regular gre devices as well. Only the special case of gre devices that use ipgre_header_ops isn't supported. Signed-off-by:
-
Guillaume Nault authored
Even though sit transports L3 data (IPv6, IPv4 or MPLS) packets, it needs to reset the mac_header pointer, so that other parts of the stack don't mistakenly access the outer header after the packet has been decapsulated. There are two rx handlers to modify: ipip6_rcv() for the ip6ip mode and sit_tunnel_rcv() which is used to re-implement the ipip and mplsip modes of ipip.ko. This allows to push an Ethernet header to sit packets and redirect them to an Ethernet device: $ tc filter add dev sit0 ingress matchall \ action vlan push_eth dst_mac 00:00:5e:00:53:01 \ src_mac 00:00:5e:00:53:00 \ action mirred egress redirect dev eth0 Without this patch, push_eth refuses to add an ethernet header because the skb appears to already have a MAC header. Signed-off-by: -
Guillaume Nault authored
Even though ipip transports IPv4 or MPLS packets, it needs to reset the mac_header pointer, so that other parts of the stack don't mistakenly access the outer header after the packet has been decapsulated. This allows to push an Ethernet header to ipip or mplsip packets and redirect them to an Ethernet device: $ tc filter add dev ipip0 ingress matchall \ action vlan push_eth dst_mac 00:00:5e:00:53:01 \ src_mac 00:00:5e:00:53:00 \ action mirred egress redirect dev eth0 Without this patch, push_eth refuses to add an ethernet header because the skb appears to already have a MAC header. Signed-off-by: -
Guillaume Nault authored
Even though bareudp transports L3 data (typically IP or MPLS), it needs to reset the mac_header pointer, so that other parts of the stack don't mistakenly access the outer header after the packet has been decapsulated. This allows to push an Ethernet header to bareudp packets and redirect them to an Ethernet device: $ tc filter add dev bareudp0 ingress matchall \ action vlan push_eth dst_mac 00:00:5e:00:53:01 \ src_mac 00:00:5e:00:53:00 \ action mirred egress redirect dev eth0 Without this patch, push_eth refuses to add an ethernet header because the skb appears to already have a MAC header. Signed-off-by:
-
- 25 Jun, 2021 10 commits
-
-
git://git.kernel.org/pub/scm/linux/kernel/git/tnguy/next-queueDavid S. Miller authored
Tony Nguyen says: ==================== 100GbE Intel Wired LAN Driver Updates 2021-06-25 This series contains updates to ice driver only. Jesse adds support for tracepoints to aide in debugging. Maciej adds support for PTP auxiliary pin support. Victor removes the VSI info from the old aggregator when moving the VSI to another aggregator. Tony removes an unnecessary VSI assignment. Christophe Jaillet fixes a memory leak for failed allocation in ice_pf_dcb_cfg(). ==================== Signed-off-by: -
Guvenc Gulce authored
When smc_sendmsg() is called before the SMC socket initialization has completed, smc_tx_sendmsg() will access un-initialized fields of the SMC socket which results in a null-pointer dereference. Fix this by checking the socket state first in smc_tx_sendmsg(). Fixes: e0e4b8fa ("net/smc: Add SMC statistics support") Reported-by: syzbot+5dda108b672b54141857@syzkaller.appspotmail.com Reviewed-by:
Karsten Graul <kgraul@linux.ibm.com> Signed-off-by:
Guvenc Gulce <guvenc@linux.ibm.com> Signed-off-by:
-
David S. Miller authored
Merge tag 'wireless-drivers-next-2021-06-25' of git://git.kernel.org/pub/scm/linux/kernel/git/kvalo/wireless-drivers-next Kalle Valo says: ==================== wireless-drivers-next patches for v5.14 Second, and most likely the last, set of patches for v5.14. mt76 and iwlwifi have most patches in this round, but rtw88 also has some new features. Nothing special really standing out. mt76 * mt7915 MSI support * disable ASPM on mt7915 * mt7915 tx status reporting * mt7921 decap offload rtw88 * beacon filter support * path diversity support * firmware crash information via devcoredump * quirks for disabling pci capabilities mt7601u * add USB ID for a XiaoDu WiFi Dongle ath11k * enable support for QCN9074 PCI devices brcmfmac * support parse country code map from DeviceTree iwlwifi * support for new hardware * support for BIOS control of 11ax enablement in Russia * support UNII4 band enablement from BIOS ==================== Signed-off-by:
-
Marcin Wojtas authored
The newly implemented fwnode_mdbiobus_register turned out to be problematic - in case the fwnode_/of_/acpi_mdio are built as modules, a dependency cycle can be observed during the depmod phase of modules_install, eg.: depmod: ERROR: Cycle detected: fwnode_mdio -> of_mdio -> fwnode_mdio depmod: ERROR: Found 2 modules in dependency cycles! OR: depmod: ERROR: Cycle detected: acpi_mdio -> fwnode_mdio -> acpi_mdio depmod: ERROR: Found 2 modules in dependency cycles! A possible solution could be to rework fwnode_mdiobus_register, so that to merge the contents of acpi_mdiobus_register and of_mdiobus_register. However feasible, such change would be very intrusive and affect huge amount of the of_mdiobus_register users. Since there are currently 2 users of ACPI and MDIO (xgmac_mdio and mvmdio), withdraw the fwnode_mdbiobus_register and roll back to a simple 'if' condition in affected drivers. Fixes: 62a6ef6a ("net: mdiobus: Introduce fwnode_mdbiobus_register()") Signed-off-by:
Marcin Wojtas <mw@semihalf.com> Signed-off-by:
-
Christophe JAILLET authored
If this 'kzalloc()' fails we must free some resources as in all the other error handling paths of this function. Fixes: 348048e7 ("ice: Implement iidc operations") Signed-off-by:
Christophe JAILLET <christophe.jaillet@wanadoo.fr> Signed-off-by:
Tony Nguyen <anthony.l.nguyen@intel.com>
-
Tony Nguyen authored
ice_get_vf_vsi() is being called twice for the same VSI. Remove the unnecessary call/assignment. Signed-off-by:
Tony Brelinski <tonyx.brelinski@intel.com>
-
Victor Raj authored
Remove the VSI info from previous aggregator after moving the VSI to a new aggregator. Signed-off-by:
Victor Raj <victor.raj@intel.com> Tested-by:
-
Maciej Machnikowski authored
The E810 device supports programmable pins for enabling both input and output events related to the PTP hardware clock. This includes both output signals with programmable period, as well as timestamping of events on input pins. Add support for enabling these using the CONFIG_PTP_1588_CLOCK interface. This allows programming the software defined pins to take advantage of the hardware clock features. Signed-off-by:
Maciej Machnikowski <maciej.machnikowski@intel.com> Signed-off-by:
Jacob Keller <jacob.e.keller@intel.com> Signed-off-by:
-
David Thompson authored
This patch adds build and driver logic for the "mlxbf_gige" Ethernet driver from Mellanox Technologies. The second generation BlueField SoC from Mellanox supports an out-of-band GigaBit Ethernet management port to the Arm subsystem. This driver supports TCP/IP network connectivity for that port, and provides back-end routines to handle basic ethtool requests. The driver interfaces to the Gigabit Ethernet block of BlueField SoC via MMIO accesses to registers, which contain control information or pointers describing transmit and receive resources. There is a single transmit queue, and the port supports transmit ring sizes of 4 to 256 entries. There is a single receive queue, and the port supports receive ring sizes of 32 to 32K entries. The transmit and receive rings are allocated from DMA coherent memory. There is a 16-bit producer and consumer index per ring to denote software ownership and hardware ownership, respectively. The main driver logic such as probe(), remove(), and netdev ops are in "mlxbf_gige_main.c". Logic in "mlxbf_gige_rx.c" and "mlxbf_gige_tx.c" handles the packet processing for receive and transmit respectively. The logic in "mlxbf_gige_ethtool.c" supports the handling of some basic ethtool requests: get driver info, get ring parameters, get registers, and get statistics. The logic in "mlxbf_gige_mdio.c" is the driver controlling the Mellanox BlueField hardware that interacts with a PHY device via MDIO/MDC pins. This driver does the following: - At driver probe time, it configures several BlueField MDIO parameters such as sample rate, full drive, voltage and MDC - It defines functions to read and write MDIO registers and registers the MDIO bus. - It defines the phy interrupt handler reporting a link up/down status change - This driver's probe is invoked from the main driver logic while the phy interrupt handler is registered in ndo_open. Driver limitations - Only supports 1Gbps speed - Only supports GMII protocol - Supports maximum packet size of 2KB - Does not support scatter-gather buffering Testing - Successful build of kernel for ARM64, ARM32, X86_64 - Tested ARM64 build on FastModels & Palladium - Tested ARM64 build on several Mellanox boards that are built with the BlueField-2 SoC. The testing includes coverage in the areas of networking (e.g. ping, iperf, ifconfig, route), file transfers (e.g. SCP), and various ethtool options relevant to this driver. Signed-off-by:David Thompson <davthompson@nvidia.com> Signed-off-by:
Asmaa Mnebhi <asmaa@nvidia.com> Reviewed-by:
Liming Sun <limings@nvidia.com> Signed-off-by:
-
Jesse Brandeburg authored
This patch is modeled after one by Scott Peterson for i40e. Add tracepoints to the driver, via a new file ice_trace.h and some new trace calls added in interesting places in the driver. Add some tracing for DIMLIB to help debug interrupt moderation problems. Performance should not be affected, and this can be very useful for debugging and adding new trace events to paths in the future. Note eBPF programs can attach to these events, as well as perf can count them since we're attaching to the events subsystem in the kernel. Co-developed-by:
Ben Shelton <benjamin.h.shelton@intel.com> Signed-off-by:
Jesse Brandeburg <jesse.brandeburg@intel.com> Tested-by:
-
- 24 Jun, 2021 20 commits
-
-
Jian-Hong Pan authored
The Broadcom UniMAC MDIO bus from mdio-bcm-unimac module comes too late. So, GENET cannot find the ethernet PHY on UniMAC MDIO bus. This leads GENET fail to attach the PHY as following log: bcmgenet fd580000.ethernet: GENET 5.0 EPHY: 0x0000 ... could not attach to PHY bcmgenet fd580000.ethernet eth0: failed to connect to PHY uart-pl011 fe201000.serial: no DMA platform data libphy: bcmgenet MII bus: probed ... unimac-mdio unimac-mdio.-19: Broadcom UniMAC MDIO bus It is not just coming too late, there is also no way for the module loader to figure out the dependency between GENET and its MDIO bus driver unless we provide this MODULE_SOFTDEP hint. This patch adds the soft dependency to load mdio-bcm-unimac module before genet module to fix this issue. Buglink: https://bugzilla.kernel.org/show_bug.cgi?id=213485 Fixes: 9a4e7969 ("net: bcmgenet: utilize generic Broadcom UniMAC MDIO controller driver") Signed-off-by:
Jian-Hong Pan <jhp@endlessos.org> Signed-off-by:
Florian Fainelli <f.fainelli@gmail.com> Signed-off-by:
-
zhang kai authored
parameter dst always points to null. Signed-off-by:
zhang kai <zhangkaiheb@126.com> Signed-off-by:
-
Ido Schimmel authored
Simply get a pointer to the data in the register payload instead of copying it to a temporary buffer. Signed-off-by:
Ido Schimmel <idosch@nvidia.com> Reviewed-by:
Jiri Pirko <jiri@nvidia.com> Signed-off-by:
-
Bailey Forrest authored
https://patchwork.kernel.org/project/netdevbpf/list/?series=506637&state=* - Remove unused variable - Use correct integer type for string formatting. - Remove `inline` in C files Fixes: 9c1a59a2 ("gve: DQO: Add ring allocation and initialization") Fixes: a57e5de4 ("gve: DQO: Add TX path") Signed-off-by:
Bailey Forrest <bcf@google.com> Reviewed-by:
Willem de Bruijn <willemb@google.com> Signed-off-by:
-
David S. Miller authored
Xin Long says: ==================== sctp: make the PLPMTUD probe more effective and efficient As David Laight noticed, it currently takes quite some time to find the optimal pmtu in the Search state, and also lacks the black hole detection in the Search Complete state. This patchset is to address them to mke the PLPMTUD probe more effective and efficient. ==================== Signed-off-by: -
Xin Long authored
These is no need to wait for 'interval' period for the next probe if the last probe is already acked in search state. The 'interval' period waiting should be only for probe failure timeout and the current pmtu check when it's in search complete state. This change will shorten the probe time a lot in search state, and also fix the document accordingly. Signed-off-by:
Xin Long <lucien.xin@gmail.com> Signed-off-by:
-
Xin Long authored
Currently the PLPMUTD probe will stop for a long period (interval * 30) after it enters search complete state. If there's a pmtu change on the route path, it takes a long time to be aware if the ICMP TooBig packet is lost or filtered. As it says in rfc8899#section-4.3: "A DPLPMTUD method MUST NOT rely solely on this method." (ICMP PTB message). This patch is to enable the other method for search complete state: "A PL can use the DPLPMTUD probing mechanism to periodically generate probe packets of the size of the current PLPMTU." With this patch, the probe will continue with the current pmtu every 'interval' until the PMTU_RAISE_TIMER 'timeout', which we implement by adding raise_count to raise the probe size when it counts to 30 and removing the SCTP_PL_COMPLETE check for PMTU_RAISE_TIMER. Signed-off-by:
-
David S. Miller authored
Vladimir Oltean says: ==================== Document the NXP SJA1110 switch as supported Now that most of the basic work for SJA1110 support has been done in the sja1105 DSA driver, let's add the missing documentation bits to make it clear that the driver can be used. ==================== Signed-off-by: -
Vladimir Oltean authored
Mention support for the SJA1110 in menuconfig. Signed-off-by:
Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by:
-
Vladimir Oltean authored
Denote that the new switch generation is supported, detail its pin strapping options (with differences compared to SJA1105) and explain how MDIO access to the internal 100base-T1 and 100base-TX PHYs is performed. Signed-off-by:
-
David S. Miller authored
Bailey Forrest says: ==================== gve: Introduce DQO descriptor format DQO is the descriptor format for our next generation virtual NIC. The existing descriptor format will be referred to as "GQI" in the patch set. One major change with DQO is it uses dual descriptor rings for both TX and RX queues. The TX path uses a TX queue to send descriptors to HW, and receives packet completion events on a TX completion queue. The RX path posts buffers to HW using an RX buffer queue and receives incoming packets on an RX queue. One important note is that DQO descriptors and doorbells are little endian. We continue to use the existing big endian control plane infrastructure. The general format of the patch series is: - Refactor existing code/data structures to be shared by DQO - Expand admin queues to support DQO device setup - Expand data structures and device setup to support DQO - Add logic to setup DQO queues - Implement datapath ==================== Signed-off-by: -
Bailey Forrest authored
The RX queue has an array of `gve_rx_buf_state_dqo` objects. All allocated pages have an associated buf_state object. When a buffer is posted on the RX buffer queue, the buffer ID will be the buf_state's index into the RX queue's array. On packet reception, the RX queue will have one descriptor for each buffer associated with a received packet. Each RX descriptor will have a buffer_id that was posted on the buffer queue. Notable mentions: - We use a default buffer size of 2048 bytes. Based on page size, we may post separate sections of a single page as separate buffers. - The driver holds an extra reference on pages passed up the receive path with an skb and keeps these pages on a list. When posting new buffers to the NIC, we check if any of these pages has only our reference, or another buffer sized segment of the page has no references. If so, it is free to reuse. This page recycling approach is a common netdev optimization that reduces page alloc/free calls. - Pages in the free list have a page_count bias in order to avoid an atomic increment of pagecount every time we attempt to reuse a page. # references = page_count() - bias - In order to track when a page is safe to reuse, we keep track of the last offset which had a single SKB reference. When this occurs, it implies that every single other offset is reusable. Otherwise, we don't know if offsets can be safely reused. - We maintain two free lists of pages. List #1 (recycled_buf_states) contains pages we know can be reused right away. List #2 (used_buf_states) contains pages which cannot be used right away. We only attempt to get pages from list #2 when list #1 is empty. We only attempt to use a small fixed number pages from list #2 before giving up and allocating a new page. Both lists are FIFOs in hope that by the time we attempt to reuse a page, the references were dropped. Signed-off-by:
Catherine Sullivan <csully@google.com> Signed-off-by:
-
Bailey Forrest authored
TX SKBs will have their buffers DMA mapped with the device. Each buffer will have at least one TX descriptor associated. Each SKB will also have a metadata descriptor. Each TX queue maintains an array of `gve_tx_pending_packet_dqo` objects. Every TX SKB will have an associated pending_packet object. A TX SKB's descriptors will use its pending_packet's index as the completion tag, which will be returned on the TX completion queue. The device implements a "flow-miss model". Most packets will simply receive a packet completion. The flow-miss system may choose to process a packet based on its contents. A TX packet which experiences a flow miss would receive a miss completion followed by a later reinjection completion. The miss-completion is received when the packet starts to be processed by the flow-miss system and the reinjection completion is received when the flow-miss system completes processing the packet and sends it on the wire. Notable mentions: - Buffers may be freed after receiving the miss-completion, but in order to avoid packet reordering, we do not complete the SKB until receiving the reinjection completion. - The driver must robustly handle the unlikely scenario where a miss completion does not have an associated reinjection completion. This is accomplished by maintaining a list of packets which have a pending reinjection completion. After a short timeout (5 seconds), the SKB and buffers are released and the pending_packet is moved to a second list which has a longer timeout (60 seconds), where the pending_packet will not be reused. When the longer timeout elapses, the driver may assume the reinjection completion would never be received and the pending_packet may be reused. - Completion handling is triggered by an interrupt and is done in the NAPI poll function. Because the TX path and completion exist in different threading contexts they maintain their own lists for free pending_packet objects. The TX path uses a lock-free approach to steal the list from the completion path. - Both the TSO context and general context descriptors have metadata bytes. The device requires that if multiple descriptors contain the same field, each descriptor must have the same value set for that field. Signed-off-by:
-
Bailey Forrest authored
When interrupts are first enabled, we also set the ratelimits, which will be static for the entire usage of the device. Signed-off-by:
-
Bailey Forrest authored
Allocate the buffer and completion ring structures. Do not populate the rings yet. That will happen in the respective rx and tx datapath follow-on patches Signed-off-by:
-
Bailey Forrest authored
Add napi netdev device registration, interrupt handling and initial tx and rx polling stubs. The stubs will be filled in follow-on patches. Also: - LRO feature advertisement and handling - Also update ethtool logic Signed-off-by:
-
Bailey Forrest authored
DQO queue creation requires additional parameters: - TX completion/RX buffer queue size - TX completion/RX buffer queue address - TX/RX queue size - RX buffer size Signed-off-by:
-
Bailey Forrest authored
- Add new DQO datapath structures: - `gve_rx_buf_queue_dqo` - `gve_rx_compl_queue_dqo` - `gve_rx_buf_state_dqo` - `gve_tx_desc_dqo` - `gve_tx_pending_packet_dqo` - Incorporate these into the existing ring data structures: - `gve_rx_ring` - `gve_tx_ring` Noteworthy mentions: - `gve_rx_buf_state` represents an RX buffer which was posted to HW. Each RX queue has an array of these objects and the index into the array is used as the buffer_id when posted to HW. - `gve_tx_pending_packet_dqo` is treated similarly for TX queues. The completion_tag is the index into the array. - These two structures have links for linked lists which are represented by 16b indexes into a contiguous array of these structures. This reduces memory footprint compared to 64b pointers. - We use unions for the writeable datapath structures to reduce cache footprint. GQI specific members will renamed like DQO members in a future patch. Signed-off-by:
-
Bailey Forrest authored
General description of rings and descriptors: TX ring is used for sending TX packet buffers to the NIC. It has the following descriptors: - `gve_tx_pkt_desc_dqo` - Data buffer descriptor - `gve_tx_tso_context_desc_dqo` - TSO context descriptor - `gve_tx_general_context_desc_dqo` - Generic metadata descriptor Metadata is a collection of 12 bytes. We define `gve_tx_metadata_dqo` which represents the logical interpetation of the metadata bytes. It's helpful to define this structure because the metadata bytes exist in multiple descriptor types (including `gve_tx_tso_context_desc_dqo`), and the device requires same field has the same value in all descriptors. The TX completion ring is used to receive completions from the NIC. Having a separate ring allows for completions to be out of order. The completion descriptor `gve_tx_compl_desc` has several different types, most important are packet and descriptor completions. Descriptor completions are used to notify the driver when descriptors sent on the TX ring are done being consumed. The descriptor completion is only used to signal that space is cleared in the TX ring. A packet completion will be received when a packet transmitted on the TX queue is done being transmitted. In addition there are "miss" and "reinjection" completions. The device implements a "flow-miss model". Most packets will simply receive a packet completion. The flow-miss system may choose to process a packet based on its contents. A TX packet which experiences a flow miss would receive a miss completion followed by a later reinjection completion. The miss-completion is received when the packet starts to be processed by the flow-miss system and the reinjection completion is received when the flow-miss system completes processing the packet and sends it on the wire. The RX buffer ring is used to send buffers to HW via the `gve_rx_desc_dqo` descriptor. Received packets are put into the RX queue by the device, which populates the `gve_rx_compl_desc_dqo` descriptor. The RX descriptors refer to buffers posted by the buffer queue. Received buffers may be returned out of order, such as when HW LRO is enabled. Important concepts: - "TX" and "RX buffer" queues, which send descriptors to the device, use MMIO doorbells to notify the device of new descriptors. - "RX" and "TX completion" queues, which receive descriptors from the device, use a "generation bit" to know when a descriptor was populated by the device. The driver initializes all bits with the "current generation". The device will populate received descriptors with the "next generation" which is inverted from the current generation. When the ring wraps, the current/next generation are swapped. - It's the driver's responsibility to ensure that the RX and TX completion queues are not overrun. This can be accomplished by limiting the number of descriptors posted to HW. - TX packets have a 16 bit completion_tag and RX buffers have a 16 bit buffer_id. These will be returned on the TX completion and RX queues respectively to let the driver know which packet/buffer was completed. Bitfields are used to describe descriptor fields. This notation is more concise and readable than shift-and-mask. It is possible because the driver is restricted to little endian platforms. Signed-off-by:
-
Bailey Forrest authored
Unlike GQI, DQO RX descriptors do not contain the L3 and L4 type of the packet. L3 and L4 types are necessary in order to set the hash and csum on RX SKBs correctly. DQO RX descriptors instead contain a 10 bit PTYPE index. The PTYPE map enables the device to tell the driver how to map from PTYPE index to L3/L4 type. The device doesn't provide any guarantees about the range of possible PTYPEs, so we just use a 1024 entry array to implement a fast mapping structure. Signed-off-by:
-
