- 09 May, 2020 5 commits
-
-
Eric Dumazet authored
We currently have to adjust ipv6 route gc_thresh/max_size depending on number of cpus on a server, this makes very little sense. If the kernels sets /proc/sys/net/ipv6/route/gc_thresh to 1024 and /proc/sys/net/ipv6/route/max_size to 4096, then we better not track the percpu dst that our implementation uses. Only routes not added (directly or indirectly) by the admin should be tracked and limited. Signed-off-by:
Eric Dumazet <edumazet@google.com> Cc: Martin KaFai Lau <kafai@fb.com> Cc: David Ahern <dsahern@kernel.org> Cc: Maciej Żenczykowski <maze@google.com> Acked-by:
Wei Wang <weiwan@google.com> Signed-off-by:
Jakub Kicinski <kuba@kernel.org>
-
Samuel Zou authored
Fixes gcc '-Wunused-but-set-variable' warning: drivers/net/dsa/sja1105/sja1105_vl.c:468:6: warning: variable ‘prev_time’ set but not used [-Wunused-but-set-variable] u32 prev_time = 0; ^~~~~~~~~ Reported-by:Hulk Robot <hulkci@huawei.com> Signed-off-by:
Samuel Zou <zou_wei@huawei.com> Reviewed-by:
Vladimir Oltean <vladimir.oltean@nxp.com> Tested-by:
-
Wei Yongjun authored
Use the helper function that wraps the calls to platform_get_resource() and devm_ioremap_resource() together. Signed-off-by:
Wei Yongjun <weiyongjun1@huawei.com> Signed-off-by:
-
Yang Yingliang authored
The type of dispatch is u8 which is always '<=' 0xff, so the dispatch <= 0xff is always true, we can remove this comparison. Reported-by:
Yang Yingliang <yangyingliang@huawei.com> Signed-off-by:
Stefan Schmidt <stefan@datenfreihafen.org> Signed-off-by:
-
Eric Dumazet authored
percpu_counter_add() uses a default batch size which is quite big on platforms with 256 cpus. (2*256 -> 512) This means dst_entries_get_fast() can be off by +/- 2*(nr_cpus^2) (131072 on servers with 256 cpus) Reduce the batch size to something more reasonable, and add logic to ip6_dst_gc() to call dst_entries_get_slow() before calling the _very_ expensive fib6_run_gc() function. Signed-off-by:
-
- 08 May, 2020 26 commits
-
-
Eric Dumazet authored
Now sch_fq has horizon feature, we want to allow QUIC/UDP applications to use EDT model so that pacing can be offloaded to the kernel (sch_fq) or the NIC. Signed-off-by:
Willem de Bruijn <willemb@google.com> Acked-by:
Vinicius Costa Gomes <vinicius.gomes@intel.com> Signed-off-by:
David S. Miller <davem@davemloft.net>
-
David S. Miller authored
Eric Dumazet says: ==================== bonding: report transmit status to callers First patches cleanup netpoll, and make sure it provides tx status to its users. Last patch changes bonding to not pretend packets were sent without error. By providing more accurate status, TCP stack can avoid adding more packets if the slave qdisc is already full. This came while testing latest horizon feature in sch_fq, with very low pacing rate flows, but should benefit hosts under stress. ==================== Signed-off-by: -
Eric Dumazet authored
Currently, bonding always returns NETDEV_TX_OK to its caller. It is worth trying to be more accurate : TCP for instance can have different recovery strategies if it can have more precise status, if packet was dropped by slave qdisc. This is especially important when host is under stress. Signed-off-by:
-
Eric Dumazet authored
netpoll_send_skb() callers seem to leak skb if the np pointer is NULL. While this should not happen, we can make the code more robust. Signed-off-by:
-
Eric Dumazet authored
Some callers want to know if the packet has been sent or dropped, to inform upper stacks. Signed-off-by:
-
Eric Dumazet authored
There is no need to inline this helper, as we intend to add more code in this function. Signed-off-by:
-
Eric Dumazet authored
netpoll_send_skb_on_dev() can get the device pointer directly from np->dev Rename it to __netpoll_send_skb() Following patch will move netpoll_send_skb() out-of-line. Signed-off-by:
-
Colin Ian King authored
The unsigned variable val is being checked for an error by checking if it is less than zero. This can never occur because val is unsigned. Fix this by making val a plain int. Addresses-Coverity: ("Unsigned compared against zero") Fixes: bdbdac76 ("ethtool: provide UAPI for PHY master/slave configuration.") Signed-off-by:Colin Ian King <colin.king@canonical.com> Signed-off-by:
-
YueHaibing authored
Fixes gcc '-Wunused-but-set-variable' warning: net/smc/smc_llc.c: In function 'smc_llc_cli_conf_link': net/smc/smc_llc.c:753:31: warning: variable 'del_llc' set but not used [-Wunused-but-set-variable] struct smc_llc_msg_del_link *del_llc; ^ net/smc/smc_llc.c: In function 'smc_llc_process_srv_delete_link': net/smc/smc_llc.c:1311:33: warning: variable 'del_llc_resp' set but not used [-Wunused-but-set-variable] struct smc_llc_msg_del_link *del_llc_resp; ^ Signed-off-by:YueHaibing <yuehaibing@huawei.com> Signed-off-by:
Karsten Graul <kgraul@linux.ibm.com> Signed-off-by:
-
zhang kai authored
so tcp_is_sack/reno checks are removed from tcp_mark_head_lost. Signed-off-by:
zhang kai <zhangkaiheb@126.com> Acked-by:
Neal Cardwell <ncardwell@google.com> Signed-off-by:
-
Jacob Keller authored
The NL_SET_ERR_MSG_MOD macro is used to report a string describing an error message to userspace via the netlink extended ACK structure. It should not have a trailing newline. Add a cocci script which catches cases where the newline marker is present. Using this script, fix the handful of cases which accidentally included a trailing new line. I couldn't figure out a way to get a patch mode working, so this script only implements context, report, and org. Signed-off-by:
Jacob Keller <jacob.e.keller@intel.com> Cc: Jakub Kicinski <kuba@kernel.org> Cc: Andy Whitcroft <apw@canonical.com> Cc: Joe Perches <joe@perches.com> Signed-off-by:
-
David S. Miller authored
Grygorii Strashko says: ==================== net: ethernet: ti: am65x-cpts: follow up dt bindings update This series is follow update for TI A65x/J721E Common platform time sync (CPTS) driver [1] to implement DT bindings review comments from Rob Herring <robh@kernel.org> [2]. - "reg" and "compatible" properties are made required for CPTS DT nodes which also required to change K3 CPSW driver to use of_platform_device_create() instead of of_platform_populate() for proper CPTS and MDIO initialization - minor DT bindings format changes - K3 CPTS example added to K3 MCU CPSW bindings [1] https://lwn.net/Articles/819313/ [2] https://lwn.net/ml/linux-kernel/20200505040419.GA8509@bogus/ ==================== Signed-off-by:
-
Grygorii Strashko authored
Update CPTS node following DT binding update: - add reg and compatible properties - fix node name Signed-off-by:
Grygorii Strashko <grygorii.strashko@ti.com> Signed-off-by:
-
Grygorii Strashko authored
This patch follows K3 CPTS review comments from Rob Herring <robh@kernel.org>. - "reg" and "compatible" properties are required now - minor format changes - K3 CPTS example added to K3 MCU CPSW bindings Cc: Rob Herring <robh@kernel.org> Signed-off-by:
-
Grygorii Strashko authored
The MCU CPSW expected to populate only MDIO device, but follow up patches will add "compatible" property to the MCU CPSW CPTS node which will cause creation of CPTS device and MCU CPSW init failure. Hence, switch to use of_platform_device_create() instead of of_platform_populate() for MDIO device population. Signed-off-by:
-
David S. Miller authored
Taehee Yoo says: ==================== hsr: hsr code refactoring There are some unnecessary routine in the hsr module. This patch removes these routines. The first patch removes incorrect comment. The second patch removes unnecessary WARN_ONCE() macro. ==================== Signed-off-by: -
Ioana Ciornei authored
Create an independent function that takes a particular frame queue and an array of frame descriptors and tries to enqueue them until it hits the maximum number fo retries. The same function will be used in the next patch also on the XDP_TX path. Also, create the dpaa2_eth_xdp_fds structure to incorporate the array of FDs as well as the number of FDs already populated. Signed-off-by:
Ioana Ciornei <ioana.ciornei@nxp.com> Signed-off-by:
-
Taehee Yoo authored
When VLAN frame is being sent, hsr calls WARN_ONCE() because hsr doesn't support VLAN. But using WARN_ONCE() is overdoing. Using netdev_warn_once() is enough. Signed-off-by:
Taehee Yoo <ap420073@gmail.com> Signed-off-by:
-
Ioana Ciornei authored
Mask the consumer index before using it. Without this, we would be writing frame descriptors beyond the ring size supported by the QBMAN block. Fixes: 3b2abda7 ("soc: fsl: dpio: Replace QMAN array mode with ring mode enqueue") Signed-off-by:
Li Yang <leoyang.li@nxp.com> Signed-off-by:
-
David S. Miller authored
Vladimir Oltean says: ==================== tc-gate offload for SJA1105 DSA switch Expose the TTEthernet hardware features of the switch using standard tc-flower actions: trap, drop, redirect and gate. v1 was submitted at: https://patchwork.ozlabs.org/project/netdev/cover/20200503211035.19363-1-olteanv@gmail.com/ v2 was submitted at: https://patchwork.ozlabs.org/project/netdev/cover/20200503211035.19363-1-olteanv@gmail.com/ Changes in v3: Made sure there are no compilation warnings when CONFIG_NET_DSA_SJA1105_TAS or CONFIG_NET_DSA_SJA1105_VL are disabled. Changes in v2: Using a newly introduced dsa_port_from_netdev public helper. ==================== Signed-off-by:
-
Vladimir Oltean authored
Add some verbiage describing how the hardware features of the switch are exposed to users through tc-flower. Signed-off-by:
-
Vladimir Oltean authored
Restrict the TTEthernet hardware support on this switch to operate as closely as possible to IEEE 802.1Qci as possible. This means that it can perform PTP-time-based ingress admission control on streams identified by {DMAC, VID, PCP}, which is useful when trying to ensure the determinism of traffic scheduled via IEEE 802.1Qbv. The oddity comes from the fact that in hardware (and in TTEthernet at large), virtual links always need a full-blown action, including not only the type of policing, but also the list of destination ports. So in practice, a single tc-gate action will result in all packets getting dropped. Additional actions (either "trap" or "redirect") need to be specified in the same filter rule such that the conforming packets are actually forwarded somewhere. Apart from the VL Lookup, Policing and Forwarding tables which need to be programmed for each flow (virtual link), the Schedule engine also needs to be told to open/close the admission gates for each individual virtual link. A fairly accurate (and detailed) description of how that works is already present in sja1105_tas.c, since it is already used to trigger the egress gates for the tc-taprio offload (IEEE 802.1Qbv). Key point here, we remember that the schedule engine supports 8 "subschedules" (execution threads that iterate through the global schedule in parallel, and that no 2 hardware threads must execute a schedule entry at the same time). For tc-taprio, each egress port used one of these 8 subschedules, leaving a total of 4 subschedules unused. In principle we could have allocated 1 subschedule for the tc-gate offload of each ingress port, but actually the schedules of all virtual links installed on each ingress port would have needed to be merged together, before they could have been programmed to hardware. So simplify our life and just merge the entire tc-gate configuration, for all virtual links on all ingress ports, into a single subschedule. Be sure to check that against the usual hardware scheduling conflicts, and program it to hardware alongside any tc-taprio subschedule that may be present. The following scenarios were tested: 1. Quantitative testing: tc qdisc add dev swp2 clsact tc filter add dev swp2 ingress flower skip_sw \ dst_mac 42:be:24:9b:76:20 \ action gate index 1 base-time 0 \ sched-entry OPEN 1200 -1 -1 \ sched-entry CLOSE 1200 -1 -1 \ action trap ping 192.168.1.2 -f PING 192.168.1.2 (192.168.1.2) 56(84) bytes of data. ............................. --- 192.168.1.2 ping statistics --- 948 packets transmitted, 467 received, 50.7384% packet loss, time 9671ms 2. Qualitative testing (with a phase-aligned schedule - the clocks are synchronized by ptp4l, not shown here): Receiver (sja1105): tc qdisc add dev swp2 clsact now=$(phc_ctl /dev/ptp1 get | awk '/clock time is/ {print $5}') && \ sec=$(echo $now | awk -F. '{print $1}') && \ base_time="$(((sec + 2) * 1000000000))" && \ echo "base time ${base_time}" tc filter add dev swp2 ingress flower skip_sw \ dst_mac 42:be:24:9b:76:20 \ action gate base-time ${base_time} \ sched-entry OPEN 60000 -1 -1 \ sched-entry CLOSE 40000 -1 -1 \ action trap Sender (enetc): now=$(phc_ctl /dev/ptp0 get | awk '/clock time is/ {print $5}') && \ sec=$(echo $now | awk -F. '{print $1}') && \ base_time="$(((sec + 2) * 1000000000))" && \ echo "base time ${base_time}" tc qdisc add dev eno0 parent root taprio \ num_tc 8 \ map 0 1 2 3 4 5 6 7 \ queues 1@0 1@1 1@2 1@3 1@4 1@5 1@6 1@7 \ base-time ${base_time} \ sched-entry S 01 50000 \ sched-entry S 00 50000 \ flags 2 ping -A 192.168.1.1 PING 192.168.1.1 (192.168.1.1): 56 data bytes ... ^C --- 192.168.1.1 ping statistics --- 1425 packets transmitted, 1424 packets received, 0% packet loss round-trip min/avg/max = 0.322/0.361/0.990 ms And just for comparison, with the tc-taprio schedule deleted: ping -A 192.168.1.1 PING 192.168.1.1 (192.168.1.1): 56 data bytes ... ^C --- 192.168.1.1 ping statistics --- 33 packets transmitted, 19 packets received, 42% packet loss round-trip min/avg/max = 0.336/0.464/0.597 ms Signed-off-by: -
Vladimir Oltean authored
Implement tc-flower offloads for redirect, trap and drop using non-critical virtual links. Commands which were tested to work are: # Send frames received on swp2 with a DA of 42:be:24:9b:76:20 to the # CPU and to swp3. This type of key (DA only) when the port's VLAN # awareness state is off. tc qdisc add dev swp2 clsact tc filter add dev swp2 ingress flower skip_sw dst_mac 42:be:24:9b:76:20 \ action mirred egress redirect dev swp3 \ action trap # Drop frames received on swp2 with a DA of 42:be:24:9b:76:20, a VID # of 100 and a PCP of 0. tc filter add dev swp2 ingress protocol 802.1Q flower skip_sw \ dst_mac 42:be:24:9b:76:20 vlan_id 100 vlan_prio 0 action drop Under the hood, all rules match on DMAC, VID and PCP, but when VLAN filtering is disabled, those are set internally by the driver to the port-based defaults. Because we would be put in an awkward situation if the user were to change the VLAN filtering state while there are active rules (packets would no longer match on the specified keys), we simply deny changing vlan_filtering unless the list of flows offloaded via virtual links is empty. Then the user can re-add new rules. Signed-off-by: -
Vladimir Oltean authored
Virtual links are a sja1105 hardware concept of executing various flow actions based on a key extracted from the frame's DMAC, VID and PCP. Currently the tc-flower offload code supports only parsing the DMAC if that is the broadcast MAC address, and the VLAN PCP. Extract the key parsing logic from the L2 policers functionality and move it into its own function, after adding extra logic for matching on any DMAC and VID. Signed-off-by:
-
Vladimir Oltean authored
This patch adds the register definitions for the: - VL Lookup Table - VL Policing Table - VL Forwarding Table - VL Forwarding Parameters Table These are needed in order to perform TTEthernet operations: QoS classification, flow-based policing and/or frame redirecting with the switch. Signed-off-by:
-
Vladimir Oltean authored
As its implementation shows, this is synonimous with calling dsa_slave_dev_check followed by dsa_slave_to_port, so it is quite simple already and provides functionality which is already there. However there is now a need for these functions outside dsa_priv.h, for example in drivers that perform mirroring and redirection through tc-flower offloads (they are given raw access to the flow_cls_offload structure), where they need to call this function on act->dev. But simply exporting dsa_slave_to_port would make it non-inline and would result in an extra function call in the hotpath, as can be seen for example in sja1105: Before: 000006dc <sja1105_xmit>: { 6dc: e92d4ff0 push {r4, r5, r6, r7, r8, r9, sl, fp, lr} 6e0: e1a04000 mov r4, r0 6e4: e591958c ldr r9, [r1, #1420] ; 0x58c <- Inline dsa_slave_to_port 6e8: e1a05001 mov r5, r1 6ec: e24dd004 sub sp, sp, #4 u16 tx_vid = dsa_8021q_tx_vid(dp->ds, dp->index); 6f0: e1c901d8 ldrd r0, [r9, #24] 6f4: ebfffffe bl 0 <dsa_8021q_tx_vid> 6f4: R_ARM_CALL dsa_8021q_tx_vid u8 pcp = netdev_txq_to_tc(netdev, queue_mapping); 6f8: e1d416b0 ldrh r1, [r4, #96] ; 0x60 u16 tx_vid = dsa_8021q_tx_vid(dp->ds, dp->index); 6fc: e1a08000 mov r8, r0 After: 000006e4 <sja1105_xmit>: { 6e4: e92d4ff0 push {r4, r5, r6, r7, r8, r9, sl, fp, lr} 6e8: e1a04000 mov r4, r0 6ec: e24dd004 sub sp, sp, #4 struct dsa_port *dp = dsa_slave_to_port(netdev); 6f0: e1a00001 mov r0, r1 { 6f4: e1a05001 mov r5, r1 struct dsa_port *dp = dsa_slave_to_port(netdev); 6f8: ebfffffe bl 0 <dsa_slave_to_port> 6f8: R_ARM_CALL dsa_slave_to_port 6fc: e1a09000 mov r9, r0 u16 tx_vid = dsa_8021q_tx_vid(dp->ds, dp->index); 700: e1c001d8 ldrd r0, [r0, #24] 704: ebfffffe bl 0 <dsa_8021q_tx_vid> 704: R_ARM_CALL dsa_8021q_tx_vid Because we want to avoid possible performance regressions, introduce this new function which is designed to be public. Suggested-by:Vivien Didelot <vivien.didelot@gmail.com> Signed-off-by:
-
- 07 May, 2020 9 commits
-
-
Chen Zhou authored
Remove duplicate headers which are included twice. Signed-off-by:
Chen Zhou <chenzhou10@huawei.com> Signed-off-by:
-
David S. Miller authored
Merge tag 'wireless-drivers-next-2020-05-07' of git://git.kernel.org/pub/scm/linux/kernel/git/kvalo/wireless-drivers-next Kalle Valo says: ==================== wireless-drivers-next patches for v5.8 First set of patches for v5.8. Changes all over, ath10k apparently seeing most new features this time. rtw88 also had lots of changes due to preparation for new hardware support. In this pull request there's also a new macro to include/linux/iopoll: read_poll_timeout_atomic(). This is needed by rtw88 for atomic polling. Major changes: ath11k * add debugfs file for testing ADDBA and DELBA * add 802.11 encapsulation offload on hardware support * add htt_peer_stats_reset debugfs file ath10k * enable VHT160 and VHT80+80 modes * enable radar detection in secondary segment * sdio: disable TX complete indication to improve throughput * sdio: decrease power consumption * sdio: add HTT TX bundle support to increase throughput * sdio: add rx bitrate reporting ath9k * improvements to AR9002 calibration logic carl9170 * remove buggy P2P_GO support p54usb * add support for AirVasT USB stick rtw88 * add support for antenna configuration ti wlcore * add support for AES_CMAC cipher iwlwifi * support for a few new FW API versions * new hw configs ==================== Signed-off-by:
-
David S. Miller authored
Manivannan Sadhasivam says: ==================== Add QRTR MHI client driver Here is the series adding MHI client driver support to Qualcomm IPC router protocol. MHI is a newly added bus to kernel which is used to communicate to external modems over a physical interface like PCI-E. This driver is used to transfer the QMI messages between the host processor and external modems over the "IPCR" channel. For QRTR, this driver is just another driver acting as a transport layer like SMD. Currently this driver is needed to control the QCA6390 WLAN device from ath11k. The ath11k MHI controller driver will take care of booting up QCA6390 and bringing it to operating state. Later, this driver will be used to transfer QMI messages over the MHI-IPCR channel. The second patch of this series removes the ARCH_QCOM dependency for QRTR. This is needed because the QRTR driver will be used with x86 machines as well to talk to devices like QCA6390. ==================== Signed-off-by: -
Manivannan Sadhasivam authored
IPC Router protocol is also used by external modems for exchanging the QMI messages. Hence, it doesn't always depend on Qualcomm platforms. One such instance is the QCA6390 WLAN device connected to x86 machine. Reviewed-by:
Bjorn Andersson <bjorn.andersson@linaro.org> Signed-off-by:
Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org> Signed-off-by:
-
Manivannan Sadhasivam authored
MHI is the transport layer used for communicating to the external modems. Hence, this commit adds MHI transport layer support to QRTR for transferring the QMI messages over IPC Router. Reviewed-by:
-
Geert Uytterhoeven authored
The VIA Rhine Ethernet interface is only present on PCI devices or VIA/WonderMedia VT8500/WM85xx SoCs. Add platform dependencies to the VIA_RHINE config symbol, to avoid asking the user about it when configuring a kernel without PCI or VT8500/WM85xx support. Signed-off-by:
Geert Uytterhoeven <geert+renesas@glider.be> Acked-by:
Arnd Bergmann <arnd@arndb.de> Signed-off-by:
-
Po Liu authored
'Dan Carpenter' reported: This code frees "sfi" and then dereferences it on the next line: > kfree(sfi); > clear_bit(sfi->index, epsfp.psfp_sfi_bitmap); This "sfi->index" should be "index". Reported-by:
Dan Carpenter <dan.carpenter@oracle.com> Signed-off-by:
Po Liu <Po.Liu@nxp.com> Signed-off-by:
-
Jason Yan authored
This function always return 0 now, we can make it return void to simplify the code. This fixes the following coccicheck warning: drivers/net/ethernet/microchip/encx24j600.c:609:5-8: Unneeded variable: "ret". Return "0" on line 653 Signed-off-by:
Jason Yan <yanaijie@huawei.com> Signed-off-by:
-
Jason Yan authored
Fix the following coccicheck warning: drivers/net/ethernet/mellanox/mlx4/en_ethtool.c:1396:5-8: Unneeded variable: "err". Return "0" on line 1411 Signed-off-by:
-
