- 14 Mar, 2018 10 commits
-
-
Ursula Braun authored
smc allocates a certain number of CQ entries for used RoCE devices. For mlx5 devices the chosen constant number results in a large allocation causing this warning: [13355.124656] WARNING: CPU: 3 PID: 16535 at mm/page_alloc.c:3883 __alloc_pages_nodemask+0x2be/0x10c0 [13355.124657] Modules linked in: smc_diag(O) smc(O) xt_CHECKSUM iptable_mangle ipt_MASQUERADE nf_nat_masquerade_ipv4 iptable_nat nf_nat_ipv4 nf_nat nf_conntrack_ipv4 nf_defrag_ipv4 xt_conntrack nf_conntrack ipt_REJECT nf_reject_ipv4 xt_tcpudp bridge stp llc ip6table_filter ip6_tables iptable_filter mlx5_ib ib_core sunrpc mlx5_core s390_trng rng_core ghash_s390 prng aes_s390 des_s390 des_generic sha512_s390 sha256_s390 sha1_s390 sha_common ptp pps_core eadm_sch dm_multipath dm_mod vhost_net tun vhost tap sch_fq_codel kvm ip_tables x_tables autofs4 [last unloaded: smc] [13355.124672] CPU: 3 PID: 16535 Comm: kworker/3:0 Tainted: G O 4.14.0uschi #1 [13355.124673] Hardware name: IBM 3906 M04 704 (LPAR) [13355.124675] Workqueue: events smc_listen_work [smc] [13355.124677] task: 00000000e2f22100 task.stack: 0000000084720000 [13355.124678] Krnl PSW : 0704c00180000000 000000000029da76 (__alloc_pages_nodemask+0x2be/0x10c0) [13355.124681] R:0 T:1 IO:1 EX:1 Key:0 M:1 W:0 P:0 AS:3 CC:0 PM:0 RI:0 EA:3 [13355.124682] Krnl GPRS: 0000000000000000 00550e00014080c0 0000000000000000 0000000000000001 [13355.124684] 000000000029d8b6 00000000f3bfd710 0000000000000000 00000000014080c0 [13355.124685] 0000000000000009 00000000ec277a00 0000000000200000 0000000000000000 [13355.124686] 0000000000000000 00000000000001ff 000000000029d8b6 0000000084723720 [13355.124708] Krnl Code: 000000000029da6a: a7110200 tmll %r1,512 000000000029da6e: a774ff29 brc 7,29d8c0 #000000000029da72: a7f40001 brc 15,29da74 >000000000029da76: a7f4ff25 brc 15,29d8c0 000000000029da7a: a7380000 lhi %r3,0 000000000029da7e: a7f4fef1 brc 15,29d860 000000000029da82: 5820f0c4 l %r2,196(%r15) 000000000029da86: a53e0048 llilh %r3,72 [13355.124720] Call Trace: [13355.124722] ([<000000000029d8b6>] __alloc_pages_nodemask+0xfe/0x10c0) [13355.124724] [<000000000013bd1e>] s390_dma_alloc+0x6e/0x148 [13355.124733] [<000003ff802eeba6>] mlx5_dma_zalloc_coherent_node+0x8e/0xe0 [mlx5_core] [13355.124740] [<000003ff802eee18>] mlx5_buf_alloc_node+0x70/0x108 [mlx5_core] [13355.124744] [<000003ff804eb410>] mlx5_ib_create_cq+0x558/0x898 [mlx5_ib] [13355.124749] [<000003ff80407d40>] ib_create_cq+0x48/0x88 [ib_core] [13355.124751] [<000003ff80109fba>] smc_ib_setup_per_ibdev+0x52/0x118 [smc] [13355.124753] [<000003ff8010bcb6>] smc_conn_create+0x65e/0x728 [smc] [13355.124755] [<000003ff801081a2>] smc_listen_work+0x2d2/0x540 [smc] [13355.124756] [<0000000000162c66>] process_one_work+0x1be/0x440 [13355.124758] [<0000000000162f40>] worker_thread+0x58/0x458 [13355.124759] [<0000000000169e7e>] kthread+0x14e/0x168 [13355.124760] [<00000000009ce8be>] kernel_thread_starter+0x6/0xc [13355.124762] [<00000000009ce8b8>] kernel_thread_starter+0x0/0xc [13355.124762] Last Breaking-Event-Address: [13355.124764] [<000000000029da72>] __alloc_pages_nodemask+0x2ba/0x10c0 [13355.124764] ---[ end trace 34be38b581c0b585 ]--- This patch reduces the smc constant for the maximum number of allocated completion queue entries SMC_MAX_CQE by 2 to avoid high round up values in the mlx5 code, and reduces the number of allocated completion queue entries even more, if the final allocation for an mlx5 device hits the MAX_ORDER limit. Reported-by:Ihnken Menssen <menssen@de.ibm.com> Signed-off-by:
Ursula Braun <ubraun@linux.vnet.ibm.com> Signed-off-by:
David S. Miller <davem@davemloft.net>
-
Veerasenareddy Burru authored
Added ethtool changes to show port type as TP (Twisted Pair) for 10GBASE-T ports. Same driver and firmware works for liquidio NIC with SFP+ ports or TP ports. Signed-off-by:
Veerasenareddy Burru <veerasenareddy.burru@cavium.com> Signed-off-by:
Felix Manlunas <felix.manlunas@cavium.com> Signed-off-by:
-
Jason Wang authored
Now we have ndo_xdp_xmit, switch to use it instead of the slow generic XDP TX routine. XDP_TX on TAP gets ~20% improvements from ~1.5Mpps to ~1.8Mpps on 2.60GHz Core(TM) i7-5600U. Signed-off-by:
Jason Wang <jasowang@redhat.com> Acked-by:
Michael S. Tsirkin <mst@redhat.com> Signed-off-by:
-
David S. Miller authored
Edward Cree says: ==================== sfc: support FEC configuration Implements the ethtool get & set fecparam operations. ==================== Signed-off-by: -
Edward Cree authored
As well as 'auto' and the forced 'off', 'rs' and 'baser' states, we also handle combinations of settings (since the fecparam->fec field is a bitmask), where auto|rs and auto|baser specify a preferred FEC mode but will fall back to the other if the cable or link partner doesn't support it. rs|baser (with or without auto bit) means prefer FEC even where auto wouldn't use it, but let FW choose which encoding to use. Signed-off-by:
Edward Cree <ecree@solarflare.com> Signed-off-by:
-
Edward Cree authored
Signed-off-by:
-
John Allen authored
The case in which we handle a reset from the state where the device is closed seems to be bugged for all types of reset. For most types of reset we currently exit the reset routine correctly, but don't set the state to indicate that we are back in the "closed" state. For some specific cases, we don't exit the reset routine at all and resetting will cause a closed device to be opened. This patch fixes the problem by unconditionally checking the reset_state and correctly setting the adapter state before returning. Signed-off-by:
John Allen <jallen@linux.vnet.ibm.com> Signed-off-by:
-
Willem de Bruijn authored
Socket option SO_ZEROCOPY determines whether the kernel ignores or processes flag MSG_ZEROCOPY on subsequent send calls. This to avoid changing behavior for legacy processes. Limiting the state change to closed sockets is annoying with passive sockets and not necessary for correctness. Once created, zerocopy skbs are processed based on their private state, not this socket flag. Remove the constraint. Signed-off-by:
Willem de Bruijn <willemb@google.com> Signed-off-by:
-
Gustavo A. R. Silva authored
_buf_ is an array and the one that must be freed is _tp_ instead. Fixes: a870a02c ("pktgen: use dynamic allocation for debug print buffer") Reported-by:
Wang Jian <jianjian.wang1@gmail.com> Signed-off-by:
Gustavo A. R. Silva <gustavo@embeddedor.com> Acked-by:
Arnd Bergmann <arnd@arndb.de> Signed-off-by:
-
Arnd Bergmann authored
After the removal of the VLA, we get a harmless warning about a large stack frame: net/core/pktgen.c: In function 'pktgen_if_write': net/core/pktgen.c:1710:1: error: the frame size of 1076 bytes is larger than 1024 bytes [-Werror=frame-larger-than=] The function was previously shown to be safe despite hitting the 1024 bye warning level. To get rid of the annoyging warning, while keeping it readable, this changes it to use strndup_user(). Obviously this is not a fast path, so the kmalloc() overhead can be disregarded. Fixes: 35951393 ("pktgen: Remove VLA usage") Signed-off-by:
-
- 13 Mar, 2018 18 commits
-
-
Arnd Bergmann authored
The new variable is only available when CONFIG_SYSCTL is enabled, otherwise we get a link error: net/ipv4/ip_tunnel.o: In function `ip_tunnel_init_net': ip_tunnel.c:(.text+0x278b): undefined reference to `sysctl_fb_tunnels_only_for_init_net' net/ipv6/sit.o: In function `sit_init_net': sit.c:(.init.text+0x4c): undefined reference to `sysctl_fb_tunnels_only_for_init_net' net/ipv6/ip6_tunnel.o: In function `ip6_tnl_init_net': ip6_tunnel.c:(.init.text+0x39): undefined reference to `sysctl_fb_tunnels_only_for_init_net' This adds an extra condition, keeping the traditional behavior when CONFIG_SYSCTL is disabled. Fixes: 79134e6c ("net: do not create fallback tunnels for non-default namespaces") Signed-off-by:
-
Kirill Tkhai authored
Make locking scheme be visible for users, and provide a comment what for we are need exit_batch() methods, and when it should be used. Signed-off-by:
Kirill Tkhai <ktkhai@virtuozzo.com> Signed-off-by:
-
Arjun Vynipadath authored
HMA(Host Memory Access) maps a part of host memory for T6-SO memfree cards. This commit does the following: - Query FW to check if we have HMA support. If yes, the params will return HMA size configured in FW. We will dma map memory based on this size. - Also contains changes to get HMA memory information via debugfs. Signed-off-by:
Arjun Vynipadath <arjun@chelsio.com> Signed-off-by:
Santosh Rastapur <santosh@chelsio.com> Signed-off-by:
Michael Werner <werner@chelsio.com> Signed-off-by:
Ganesh GR <ganeshgr@chelsio.com> Signed-off-by:
-
David S. Miller authored
Kirill Tkhai says: ==================== Converting pernet_operations (part #6) this series continues to review and to convert pernet_operations to make them possible to be executed in parallel for several net namespaces in the same time. There are sctp, tipc and rds in this series. ==================== Signed-off-by: -
Kirill Tkhai authored
These pernet_operations create and destroy sysctl table and listen socket. Also, exit method flushes global workqueue and work. Everything looks per-net safe, so we can mark them async. Signed-off-by:
-
Kirill Tkhai authored
TIPC looks concentrated in itself, and other pernet_operations seem not touching its entities. tipc_net_ops look pernet-divided, and they should be safe to be executed in parallel for several net the same time. Signed-off-by:
-
Kirill Tkhai authored
These pernet_operations create and destroy net::sctp::ctl_sock. Since pernet_operations do not send sctp packets each other, they look safe to be marked as async. Signed-off-by:
Neil Horman <nhorman@tuxdriver.com> Signed-off-by:
-
Kirill Tkhai authored
These pernet_operations have a deal with sysctl, /proc entries and statistics. Also, there are freeing of net::sctp::addr_waitq queue and net::sctp::local_addr_list in exit method. All of them look pernet-divided, and it seems these items are only interesting for sctp_defaults_ops, which are safe to be executed in parallel. Signed-off-by:
-
Wei Yongjun authored
Return error code -EINVAL in the address len check error handling case since 'err' can be overwrite to 0 by 'err = sctp_verify_addr()' in the for loop. Fixes: 2c0dbaa0 ("sctp: add support for SCTP_DSTADDRV4/6 Information for sendmsg") Signed-off-by:
Wei Yongjun <weiyongjun1@huawei.com> Acked-by:
Xin Long <lucien.xin@gmail.com> Acked-by:
-
Thomas Falcon authored
Sorry, one of the patches I sent in an earlier series has some dumb mistakes. One was that I had changed the parameter for the errata workaround function but forgot to make that change in the code that called it. The second mistake was a forgotten return value at the end of the function in case the workaround was not needed. Signed-off-by:
Thomas Falcon <tlfalcon@linux.vnet.ibm.com> Signed-off-by:
-
David S. Miller authored
Thomas Falcon says: ==================== ibmvnic: Fix VLAN and other device errata This patch series contains fixes for VLAN and other backing hardware errata. The VLAN fixes are mostly to account for the additional four bytes VLAN header in TX descriptors and buffers, when applicable. The other fixes for device errata are to pad small packets to avoid a possible connection error that can occur when some devices attempt to transmit small packets. The other fixes are GSO related. Some devices cannot handle a smaller MSS or a packet with a single segment, so disable GSO in those cases. v2: Fix style mistake (unneeded brackets) in patch 3/4 ==================== Signed-off-by: -
Thomas Falcon authored
TSO packets with one segment or with an MSS less than 224 can cause errors on some backing devices, so disable GSO in those cases. Signed-off-by:
-
Thomas Falcon authored
Some backing devices cannot handle small packets well, so pad any small packets to avoid that. It was recommended that the VNIC driver should not send packets smaller than the minimum MTU value provided by firmware, so pad small packets to be at least that long. Signed-off-by:
-
Thomas Falcon authored
The extra four bytes of a VLAN packet was throwing off TX buffer entry values used by the driver. Account for those bytes when in buffer size and buffer entry calculations Signed-off-by:
-
Thomas Falcon authored
If a VLAN tag is present in the Ethernet header, account for that when providing the L2 header to firmware. Signed-off-by:
-
Roman Mashak authored
Signed-off-by:
Roman Mashak <mrv@mojatatu.com> Signed-off-by:
-
Roman Mashak authored
Signed-off-by:
-
Heiner Kallweit authored
Currently the link state is initialized to "up" when the phy_device is being created. This is not consistent with the phy state being initialized to PHY_DOWN. Usually this doen't do any harm because the link state is updated once the PHY reaches state PHY_AN. However e.g. if a LAN port isn't used and the PHY remains down this inconsistency remains and calls to functions like phy_print_status() give false results. Therefore change the initialization to link being down. Signed-off-by:
Heiner Kallweit <hkallweit1@gmail.com> Reviewed-by:
Florian Fainelli <f.fainelli@gmail.com> Signed-off-by:
-
- 12 Mar, 2018 12 commits
-
-
git://git.kernel.org/pub/scm/linux/kernel/git/jkirsher/next-queueDavid S. Miller authored
Jeff Kirsher says: ==================== 10GbE Intel Wired LAN Driver Updates 2018-03-12 This series contains updates to ixgbe and ixgbevf only. Shannon Nelson provides three fixes to the ipsec portion of ixgbe. Make sure we are using 128-bit authentication, since it is the only size supported for hardware offload. Fixed the transmit trailer length calculation for ipsec by finding the padding value and adding it to the authentication length, then save it off so that we can put it in the transmit descriptor to tell the device where to stop the checksum calculation. Lastly, cleaned up useless and dead code. Tonghao Zhang adds a ethtool stat for receive length errors, since the driver was already collecting this counter. Arnd Bergmann fixed a warning about an used variable by "rephrasing" the code so that the compiler can see the use of the variable in question. Paul fixes an issue where "HIDE_VLAN" was being cleared on VF reset, so ensure to set "HIDE_VLAN" when port VLAN is enabled after a VF reset. ==================== Signed-off-by: -
Paul Greenwalt authored
If port VLAN is enabled, set PFQDE.HIDE_VLAN during VF reset. Setting only PFQDE.PFQDE during VF reset was clearing PFQDE.HIDE_VLAN. Signed-off-by:
Paul Greenwalt <paul.greenwalt@intel.com> Tested-by:
Andrew Bowers <andrewx.bowers@intel.com> Signed-off-by:
Jeff Kirsher <jeffrey.t.kirsher@intel.com>
-
Salvatore Mesoraca authored
Avoid VLA[1] by using an already allocated buffer passed by the caller. [1] https://lkml.org/lkml/2018/3/7/621Signed-off-by:
Salvatore Mesoraca <s.mesoraca16@gmail.com> Acked-by:
Santosh Shilimkar <santosh.shilimkar@oracle.com> Signed-off-by:
-
Salvatore Mesoraca authored
Avoid VLA[1] by using an already allocated buffer passed by the caller. [1] https://lkml.org/lkml/2018/3/7/621Signed-off-by:
-
Arnd Bergmann authored
The new ixgbevf_set_rx_buffer_len() function causes a harmless warnings in configurations with large page size: drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c: In function 'ixgbevf_set_rx_buffer_len': drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c:1758:15: error: unused variable 'max_frame' [-Werror=unused-variable] This rephrases the code so that the compiler can see the use of that variable, making it slightly easier to read in the process. Fixes: f15c5ba5 ("ixgbevf: add support for using order 1 pages to receive large frames") Signed-off-by:
Alexander Duyck <alexander.h.duyck@intel.com> Signed-off-by:
-
Tonghao Zhang authored
ixgbe enabled rlec counter and the rx_error used it. We can export the counter directly via ethtool -S ethX. Signed-off-by:
Tonghao Zhang <xiangxia.m.yue@gmail.com> Tested-by:
-
Shannon Nelson authored
With commit 7f05b467 ("xfrm: check for xdo_dev_state_free") we no longer need to add an empty callback function to the driver, so now let's remove the useless code. Signed-off-by:
Shannon Nelson <shannon.nelson@oracle.com> Tested-by:
-
Shannon Nelson authored
Fix up the Tx trailer length calculation. We can't believe the trailer len from the xstate information because it was calculated before the packet was put together and padding added. This bit of code finds the padding value in the trailer, adds it to the authentication length, and saves it so later we can put it into the Tx descriptor to tell the device where to stop the checksum calculation. Fixes: 59259470 ("ixgbe: process the Tx ipsec offload") Signed-off-by:
-
Shannon Nelson authored
Make sure the Security Association is using a 128-bit authentication, since that's the only size that the hardware offload supports. Signed-off-by:
-
Wei Yongjun authored
Fixes the following sparse warnings: drivers/net/ethernet/mellanox/mlxsw/spectrum_kvdl.c:371:5: warning: symbol 'mlxsw_sp_kvdl_single_occ_get' was not declared. Should it be static? drivers/net/ethernet/mellanox/mlxsw/spectrum_kvdl.c:384:5: warning: symbol 'mlxsw_sp_kvdl_chunks_occ_get' was not declared. Should it be static? drivers/net/ethernet/mellanox/mlxsw/spectrum_kvdl.c:397:5: warning: symbol 'mlxsw_sp_kvdl_large_chunks_occ_get' was not declared. Should it be static? Signed-off-by:
Jiri Pirko <jiri@mellanox.com> Signed-off-by:
-
Gal Pressman authored
Same as LRO, hardware GRO cannot be enabled with RX-FCS. When both are requested, hardware GRO will be dropped. Suggested-by:
Gal Pressman <galp@mellanox.com> Signed-off-by:
-
Salvatore Mesoraca authored
Avoid a VLA[1] by using a real constant expression instead of a variable. The compiler should be able to optimize the original code and avoid using an actual VLA. Anyway this change is useful because it will avoid a false positive with -Wvla, it might also help the compiler generating better code. [1] https://lkml.org/lkml/2018/3/7/621Signed-off-by:
-
