summaryrefslogtreecommitdiffstats
AgeCommit message (Collapse)AuthorFilesLines
2018-02-27net: Convert tc_action_net_init() and tc_action_net_exit() based ↵Kirill Tkhai16-0/+17
pernet_operations These pernet_operations are from net/sched directory, and they call only tc_action_net_init() and tc_action_net_exit(): bpf_net_ops connmark_net_ops csum_net_ops gact_net_ops ife_net_ops ipt_net_ops xt_net_ops mirred_net_ops nat_net_ops pedit_net_ops police_net_ops sample_net_ops simp_net_ops skbedit_net_ops skbmod_net_ops tunnel_key_net_ops vlan_net_ops 1)tc_action_net_init() just allocates and initializes per-net memory. 2)There should not be in-flight packets at the time of tc_action_net_exit() call, or another pernet_operations send packets to dying net (except netlink). So, it seems they can be marked as async. Signed-off-by: Kirill Tkhai <ktkhai@virtuozzo.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2018-02-27net: Convert sysctl creating and destroying pernet_operationsKirill Tkhai2-0/+2
These pernet_operations create and destroy sysctl tables, and they are able to be executed in parallel with any others: ip_vs_lblc_ops ip_vs_lblcr_ops Signed-off-by: Kirill Tkhai <ktkhai@virtuozzo.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2018-02-27net: Convert cma_pernet_operationsKirill Tkhai1-0/+1
These pernet_operations just create and destroy IDR. So, we mark them as async. Signed-off-by: Kirill Tkhai <ktkhai@virtuozzo.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2018-02-27net: Convert simple pernet_operationsKirill Tkhai6-0/+6
These pernet_operations make pretty simple actions like variable initialization on init, debug checks on exit, and so on, and they obviously are able to be executed in parallel with any others: vrf_net_ops lockd_net_ops grace_net_ops xfrm6_tunnel_net_ops kcm_net_ops tcf_net_ops Signed-off-by: Kirill Tkhai <ktkhai@virtuozzo.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2018-02-27net: Convert nfs_net_opsKirill Tkhai1-0/+1
These pernet_operations just create and destroy /proc entries and net_generic()->cb_ident_idr IDR. So, we are able to mark them async. Signed-off-by: Kirill Tkhai <ktkhai@virtuozzo.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2018-02-27net: Convert synproxy_net_opsKirill Tkhai1-0/+1
These pernet_operations create and destroy /proc entries and allocate extents to template ct, which depend on global nf_ct_ext_types[] array. So, we are able to mark them async. Signed-off-by: Kirill Tkhai <ktkhai@virtuozzo.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2018-02-27net: Convert hashlimit_net_ops and recent_net_opsKirill Tkhai2-0/+2
These pernet_operations just create and destroy /proc entries. Also, new /proc entries also may come after new nf rules are added, but this is not possible, when net isn't alive. So, they are safe to be marked as async. Signed-off-by: Kirill Tkhai <ktkhai@virtuozzo.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2018-02-27net: Convert /proc creating and destroying pernet_operationsKirill Tkhai7-0/+7
These pernet_operations just create and destroy /proc entries, and they can safely marked as async: pppoe_net_ops vlan_net_ops canbcm_pernet_ops kcm_net_ops pfkey_net_ops pppol2tp_net_ops phonet_net_ops Signed-off-by: Kirill Tkhai <ktkhai@virtuozzo.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2018-02-27ipvlan: fix building with modular IPV6Arnd Bergmann1-0/+1
We no longer depend on IPV6, but that now causes a link error with CONFIG_IPV6=m and CONFIG_IPVLAN=y: drivers/net/ipvlan/ipvlan_core.o: In function `ipvlan_queue_xmit': ipvlan_core.c:(.text+0x1440): undefined reference to `ip6_route_output_flags' drivers/net/ipvlan/ipvlan_core.o: In function `ipvlan_l3_rcv': ipvlan_core.c:(.text+0x1818): undefined reference to `ip6_route_input_lookup' This adds back the dependency on IPV6, with the option of building without IPV6, but forcing IPVLAN to be a module when IPV6 is a module. Fixes: 94333fac44d1 ("ipvlan: drop ipv6 dependency") Signed-off-by: Arnd Bergmann <arnd@arndb.de> Signed-off-by: David S. Miller <davem@davemloft.net>
2018-02-26Merge branch '10GbE' of ↵David S. Miller6-380/+573
git://git.kernel.org/pub/scm/linux/kernel/git/jkirsher/next-queue Jeff Kirsher says: ==================== 10GbE Intel Wired LAN Driver Updates 2018-02-26 This series contains updates to ixgbe and ixgbevf only. Colin Ian King cleans up redundant variable assignments. Tonghao Zhang updates ixgbe to avoid writing to the hardware when the redirection table has not changed. Jake fixes the driver logic for checking and clearing receive timestamp hangs so that when the PTP_RX_TIMESTAMP_IN_REGISTER flag is set, we no longer need to check for receive timestamp hangs, which in turn will stop the spurious log messages. Emil updates ixgbevf with several features and improvements done in other drivers, starting with the handling of page addresses so that we always refer to them using a void pointer. Added a 'legacy-rx' flag to allow switching between the old and new receive code paths. Added support for using 3K buggers in order 1 page. Updated the driver to ensure that calls to ixgbevf_open() are rtnl lock protected and improved the error handling when setting up multiple queues. Added support for providing a buffer with head room and tail room to allow for shared info, NET_SKB_PAD, and NET_IP_ALIGN, so that we can start using build_skb to build frames instead of using memcpy() the headers. Updated the logic of handling rings closer to ixgbe. Consolidated the receive paths to reduce duplication when we expand them in the future. Added build_skb() support to ixgbevf. ==================== Signed-off-by: David S. Miller <davem@davemloft.net>
2018-02-26net: make kmem caches as __ro_after_initAlexey Dobriyan6-8/+11
All kmem caches aren't reallocated once set up. Signed-off-by: Alexey Dobriyan <adobriyan@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2018-02-26Merge branch 'sonic-ethernet-cleanups'David S. Miller5-197/+210
Finn Thain says: ==================== Fixes, cleanup and modernization for SONIC ethernet drivers Changes since v4 of combined patch series: - Removed redundant and non-portable MACH_IS_MAC tests. - Omitted patches unrelated to SONIC drivers. - Dropped changes to the 'version_printed' logic and debug message text. ==================== Signed-off-by: David S. Miller <davem@davemloft.net>
2018-02-26net/sonic: Replace custom debug logging with netif_* callsFinn Thain5-103/+61
Eliminate duplicated debug code by moving it into the core driver. Don't log the only valid silicon revision number (it's in the source). Cc: Thomas Bogendoerfer <tsbogend@alpha.franken.de> Cc: Chris Zankel <chris@zankel.net> Tested-by: Stan Johnson <userm57@yahoo.com> Signed-off-by: Finn Thain <fthain@telegraphics.com.au> Signed-off-by: David S. Miller <davem@davemloft.net>
2018-02-26net/sonic: Clean up and modernize log messagesFinn Thain3-40/+33
Add missing printk severity levels by adopting pr_foo() calls for the platform_driver and dev_foo() calls for the nubus_driver. Avoid KERN_CONT usage as per advice from checkpatch. Avoid #ifdef around printk calls. Don't log driver probe messages after calling register_netdev(). Cc: Thomas Bogendoerfer <tsbogend@alpha.franken.de> Cc: Chris Zankel <chris@zankel.net> Tested-by: Stan Johnson <userm57@yahoo.com> Signed-off-by: Finn Thain <fthain@telegraphics.com.au> Signed-off-by: David S. Miller <davem@davemloft.net>
2018-02-26net/macsonic: Drop redundant MACH_IS_MAC testFinn Thain1-3/+0
The MACH_IS_MAC test is redundant here because the platform device won't get registered unless MACH_IS_MAC. Cc: Geert Uytterhoeven <geert@linux-m68k.org> Signed-off-by: Finn Thain <fthain@telegraphics.com.au> Reviewed-by: Geert Uytterhoeven <geert@linux-m68k.org> Signed-off-by: David S. Miller <davem@davemloft.net>
2018-02-26net/macsonic: Convert to nubus_driverFinn Thain1-54/+119
This resolves an old issue preventing any NuBus SONIC NICs from working in a Mac with an on-board SONIC device. Tested-by: Stan Johnson <userm57@yahoo.com> Signed-off-by: Finn Thain <fthain@telegraphics.com.au> Signed-off-by: David S. Miller <davem@davemloft.net>
2018-02-26sh_eth: fix TSU init on SH7734/R8A7740Sergei Shtylyov2-1/+6
It appears that the single port Ether controllers having TSU (like SH7734/ R8A7740) need the same kind of treating in sh_eth_tsu_init() as R7S72100 currently has -- they also don't have the TSU registers related e.g. to passing the frames between ports. Add the 'sh_eth_cpu_data::dual_port' flag and use it as a new criterion for taking a "short path" in the TSU init sequence in order to avoid writing to the non-existent registers... Fixes: f0e81fecd4f8 ("net: sh_eth: Add support SH7734") Fixes: 73a0d907301e ("net: sh_eth: add support R8A7740") Signed-off-by: Sergei Shtylyov <sergei.shtylyov@cogentembedded.com> Tested-by: Geert Uytterhoeven <geert+renesas@glider.be> Signed-off-by: David S. Miller <davem@davemloft.net>
2018-02-26sh_eth: TSU_QTAG0/1 registers the same as TSU_QTAGM0/1Sergei Shtylyov2-13/+6
The TSU_QTAG0/1 registers found in the Gigabit Ether controllers actually have the same long name as the TSU_QTAGM0/1 registers in the early Ether controllers: Qtag Addition/Deletion Set Register (Port 0/1 to 1/0); thus there's no need to make a difference in sh_eth_tsu_init() between those controllers. Unfortunately, we can't just remove TSU_QTAG0/1 from the register *enum* because that would break the ethtool register dump... Fixes: b0ca2a21f769 ("sh_eth: Add support of SH7763 to sh_eth") Signed-off-by: Sergei Shtylyov <sergei.shtylyov@cogentembedded.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2018-02-26tc: python3, string formattingsBTaskaya2-5/+5
This patch converts old type string formattings to new type string formattings for adapting Linux Traffic Control (tc) unit testing suite python3. Linux Traffic Control (tc) unit testing suite's code quality improved is improved with this patch. According to python documentation; "The built-in string class provides the ability to do complex variable substitutions and value formatting via the format() method described in PEP 3101. " but the project was using old type formattings and new type string formattings together, this patch's main purpose is converting all old types to new types. Following files changed: 1. tools/testing/selftests/tc-testing/tdc.py 2. tools/testing/selftests/tc-testing/tdc_batch.py Following PEP rules applied: 1. PEP8 - Code Styling 2. PEP3101 - Advanced Code Formatting Signed-off-by: Batuhan Osman Taskaya <batuhanosmantaskaya@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2018-02-26ixgbevf: remove redundant initialization of variable 'dma'Colin Ian King1-1/+1
Variable dma is initialized with a value that is never read, later on it is re-assigned a new value, hence the initialization is redundant and can be removed. Cleans up clang warning: drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c:584:13: warning: Value stored to 'dma' during its initialization is never read Signed-off-by: Colin Ian King <colin.king@canonical.com> Acked-by: Alexander Duyck <alexander.h.duyck@intel.com> Tested-by: Andrew Bowers <andrewx.bowers@intel.com> Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
2018-02-26ixgbevf: add build_skb supportEmil Tantilov1-0/+41
Add support for build_skb() similar to: commit 6f429223b31c ("ixgbe: Add support for build_skb") Signed-off-by: Emil Tantilov <emil.s.tantilov@intel.com> Tested-by: Krishneil Singh <krishneil.k.singh@intel.com> Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
2018-02-26ixgbevf: break out Rx buffer page managementEmil Tantilov1-113/+114
Based on commit e014272672b9 ("igb: Break out Rx buffer page management") Consolidate Rx code paths to reduce duplication when we expand them in the future. Signed-off-by: Emil Tantilov <emil.s.tantilov@intel.com> Tested-by: Krishneil Singh <krishneil.k.singh@intel.com> Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
2018-02-26ixgbevf: allocate the rings as part of q_vectorEmil Tantilov2-217/+182
Make it so that all rings allocations are made as part of q_vector. The advantage to this is that we can keep all of the memory related to a single interrupt in one page. The goal is to bring the logic of handling rings closer to ixgbe. Signed-off-by: Emil Tantilov <emil.s.tantilov@intel.com> Tested-by: Krishneil Singh <krishneil.k.singh@intel.com> Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
2018-02-26ixgbevf: make sure all frames fit minimum size requirementsEmil Tantilov1-5/+25
Similar to commit a50c29dd09ed ("ixgbe: Make certain that all frames fit minimum size requirements") Make sure that any packet we attempt to transmit will meet minimum size requirements. Signed-off-by: Emil Tantilov <emil.s.tantilov@intel.com> Tested-by: Krishneil Singh <krishneil.k.singh@intel.com> Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
2018-02-26ixgbevf: add support for padding packetEmil Tantilov2-4/+39
Following the logic from commit 2de6aa3a666e ("ixgbe: Add support for padding packet") Add support for providing a buffer with headroom and tail room to allow for shared info, NET_SKB_PAD, and NET_IP_ALIGN. With this combined with the DMA changes we can start using build_skb to build frames around an incoming Rx buffer instead of having to memcpy the headers. Signed-off-by: Emil Tantilov <emil.s.tantilov@intel.com> Tested-by: Krishneil Singh <krishneil.k.singh@intel.com> Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
2018-02-26ixgbevf: setup queue countsEmil Tantilov1-15/+29
Add calls for netif_set_real_num_t/rx_queues() in ixgbevf_open(). Make sure that calls to ixgbevf_open() are rtnl protected and improve the error handling when setting up multiple queues. Signed-off-by: Emil Tantilov <emil.s.tantilov@intel.com> Tested-by: Krishneil Singh <krishneil.k.singh@intel.com> Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
2018-02-26ixgbevf: add support for using order 1 pages to receive large framesEmil Tantilov2-24/+92
Based on commit 8649aaef4044 ("igb: Add support for using order 1 pages to receive large frames") Add support for using 3K buffers in order 1 page. We are reserving 1K for now to have space available for future tail room and head room when we enable build_skb support. Signed-off-by: Emil Tantilov <emil.s.tantilov@intel.com> Tested-by: Krishneil Singh <krishneil.k.singh@intel.com> Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
2018-02-26ixgbevf: add ethtool private flag for legacy RxEmil Tantilov2-0/+50
Introduce legacy-rx private flag that will allow switching between the old and new (build_skb based) Rx code paths. The implementation is the same as in commit e08912985b29 ("igb: Add support for ethtool private flag to allow use of legacy Rx") This provides a means of validating the legacy Rx path in the event that we are forced to fall back. At some point in the future when we are convinced we don't need it anymore we might be able to drop the legacy-rx flag. Signed-off-by: Emil Tantilov <emil.s.tantilov@intel.com> Tested-by: Krishneil Singh <krishneil.k.singh@intel.com> Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
2018-02-26ixgbevf: use page_address offset from pageEmil Tantilov1-6/+5
Based on commit 3456fd53421e ("igb: Use page_address offset from page instead of masking virtual address") Update the handling of page addresses so that we always refer to them using a void pointer, and try to use the consistent name of va indicating we are working with a virtual address. Signed-off-by: Emil Tantilov <emil.s.tantilov@intel.com> Tested-by: Krishneil Singh <krishneil.k.singh@intel.com> Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
2018-02-26ixgbe: prevent ptp_rx_hang from running when in FILTER_ALL modeJacob Keller1-1/+2
On hardware which supports timestamping all packets, the timestamps are recorded in the packet buffer, and the driver no longer uses or reads the registers. This makes the logic for checking and clearing Rx timestamp hangs meaningless. If we run the ixgbe_ptp_rx_hang() function in this case, then the driver will continuously spam the log output with "Clearing Rx timestamp hang". These messages are spurious, and confusing to end users. The original code in commit a9763f3cb54c ("ixgbe: Update PTP to support X550EM_x devices", 2015-12-03) did have a flag PTP_RX_TIMESTAMP_IN_REGISTER which was intended to be used to avoid the Rx timestamp hang check, however it did not actually check the flag before calling the function. Do so now in order to stop the checks and prevent the spurious log messages. Fixes: a9763f3cb54c ("ixgbe: Update PTP to support X550EM_x devices", 2015-12-03) Signed-off-by: Jacob Keller <jacob.e.keller@intel.com> Tested-by: Andrew Bowers <andrewx.bowers@intel.com> Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
2018-02-26ixgbe: Avoid to write the RETA table when unnecessaryTonghao Zhang1-2/+2
If indir == 0 in the ixgbe_set_rxfh(), it is unnecessary to write the HW. Because redirection table is not changed. Signed-off-by: Tonghao Zhang <xiangxia.m.yue@gmail.com> Tested-by: Andrew Bowers <andrewx.bowers@intel.com> Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
2018-02-26xen-netback: make function xenvif_rx_skb staticColin Ian King1-1/+1
The function xenvif_rx_skb is local to the source and does not need to be in global scope, so make it static. Cleans up sparse warning: drivers/net/xen-netback/rx.c:422:6: warning: symbol 'xenvif_rx_skb' was not declared. Should it be static? Signed-off-by: Colin Ian King <colin.king@canonical.com> Acked-by: Wei Liu <wei.liu2@citrix.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2018-02-26Merge branch 'tools-tc-testing-better-error-reporting'David S. Miller1-16/+72
Brenda J. Butler says: ==================== tools: tc-testing: better error reporting This patch set contains a bit of cleanup and better error reporting, esp. in pre- and post-suite, and pre- and post-case commands. ==================== Signed-off-by: David S. Miller <davem@davemloft.net>
2018-02-26tools: tc-testing: better error reportingBrenda J. Butler1-14/+70
Do a better job with error handling - in pre- and post-suite, in pre- and post-case. Show a traceback for errors. Signed-off-by: Brenda J. Butler <bjb@mojatatu.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2018-02-26tools: tc-testing: Fix indentationBrenda J. Butler1-2/+2
Signed-off-by: Brenda J. Butler <bjb@mojatatu.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2018-02-26sfc: falcon: remove duplicated bit-wise or of LOOPBACK_SGMIIColin Ian King1-1/+0
Bit pattern LOOPBACK_SGMII is being bit-wise or'd twice; remove the redundant 2nd LOOPBACK_SGMII Signed-off-by: Colin Ian King <colin.king@canonical.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2018-02-26mlxsw: spectrum_kvdl: avoid uninitialized variable warningArnd Bergmann1-0/+2
gcc warns that 'resource_id' is not initialized if we don't come though any of the three 'case' statements before: drivers/net/ethernet/mellanox/mlxsw/spectrum_kvdl.c: In function 'mlxsw_sp_kvdl_part_init': drivers/net/ethernet/mellanox/mlxsw/spectrum_kvdl.c:275:8: error: 'resource_id' may be used uninitialized in this function [-Werror=maybe-uninitialized] In the current code, that won't happen, but it's more robust to explicitly handle this by returning a failure from mlxsw_sp_kvdl_part_init. Fixes: 887839e6960d ("mlxsw: spectrum_kvdl: Add support for dynamic partition set") Signed-off-by: Arnd Bergmann <arnd@arndb.de> Acked-by: Arkadi Sharshevsky <arkadis@mellanox.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2018-02-26mlxsw: spectrum_kvdl: use div_u64() for 64-bit divisionArnd Bergmann1-1/+1
Calculating the number of entries now uses 64-bit arithmetic that causes a link error on 32-bit architectures: drivers/net/ethernet/mellanox/mlxsw/spectrum_kvdl.o: In function `mlxsw_sp_kvdl_init': spectrum_kvdl.c:(.text+0x51c): undefined reference to `__aeabi_uldivmod' We could probably use a 32-bit division here as before, but since this is not in a performance critical path, div_u64() seems cleaner here. Fixes: 887839e6960d ("mlxsw: spectrum_kvdl: Add support for dynamic partition set") Signed-off-by: Arnd Bergmann <arnd@arndb.de> Acked-by: Arkadi Sharshevsky <arkadis@mellanox.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2018-02-26ixgbe: remove redundant initialization of 'pool'Colin Ian King1-1/+0
Variable pool is being assigned zero and then in the following for-loop is it being set to zero again. Remove the redundant first assignment. Cleans up clang warning: drivers/net/ethernet/intel/ixgbe/ixgbe_lib.c:61:2: warning: Value stored to 'pool' is never read Signed-off-by: Colin Ian King <colin.king@canonical.com> Tested-by: Andrew Bowers <andrewx.bowers@intel.com> Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
2018-02-26Merge branch 'mv88e6xxx-Poll-when-no-interrupt-defined'David S. Miller3-43/+138
Andrew Lunn says: ==================== mv88e6xxx: Poll when no interrupt defined Not all boards using the mv88e6xxx switches have the interrupt output connected to a GPIO. On these boards phylib has to poll the PHYs, rather than use interrupts. Have the driver poll the interrupt status register, which is more efficient than having phylib do it. And it enables other switch interrupts to be services. The Armada 370RD is such a board without a interrupt GPIO. Now that interrupts work, wire up the PHYs to make use if them. ==================== Signed-off-by: David S. Miller <davem@davemloft.net>
2018-02-26arm: mvebu: 370-rd: Enable PHY interrupt handlingAndrew Lunn1-0/+32
The Ethernet switch has an embedded interrupt controller. Interrupts from the embedded PHYs are part of this interrupt controller. Explicitly list the MDIO bus the embedded PHYs are on, and wire up the interrupts. Signed-off-by: Andrew Lunn <andrew@lunn.ch> Signed-off-by: David S. Miller <davem@davemloft.net>
2018-02-26net: dsa: mv88e6xxx: Poll when no interrupt definedAndrew Lunn2-43/+106
Not all boards have the interrupt output from the switch connected to a GPIO line. In such cases, phylib has to poll the internal PHYs, rather than receive an interrupt when there is a change in the link state. phylib polls once per second, and per PHY reads around 4 words. With a switch typically having 4 internal PHYs, this means 16 MDIO transactions per second. Rather than performing this phylib level polling, have the driver poll the interrupt status register. If the status register indicates an interrupt condition processing of interrupts in the same way as if a GPIO was used. Polling 10 times a second places less load on the MDIO bus. But rather than taking on average 0.5s to detect a link change, it takes less than 0.05s. Additionally, other interrupts, such as the watchdog, ATU and VTU violations will be reported. Signed-off-by: Andrew Lunn <andrew@lunn.ch> Signed-off-by: David S. Miller <davem@davemloft.net>
2018-02-26mlxsw: spectrum_switchdev: Allow port enslavement to a VLAN-unaware bridgeIdo Schimmel1-7/+5
Up until now we only allowed VLAN devices to be put in a VLAN-unaware bridge, but some users need the ability to enslave physical ports as well. This is achieved by mapping the port and VID 1 to the bridge's vFID, instead of the port and the VID used by the VLAN device. The above is valid because as long as the port is not enslaved to a bridge, VID 1 is guaranteed to be configured as PVID and egress untagged. Signed-off-by: Ido Schimmel <idosch@mellanox.com> Signed-off-by: Jiri Pirko <jiri@mellanox.com> Tested-by: David Ahern <dsahern@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2018-02-26Merge git://git.kernel.org/pub/scm/linux/kernel/git/bpf/bpf-nextDavid S. Miller13-100/+779
Daniel Borkmann says: ==================== pull-request: bpf-next 2018-02-26 The following pull-request contains BPF updates for your *net-next* tree. The main changes are: 1) Various improvements for BPF kselftests: i) skip unprivileged tests when kernel.unprivileged_bpf_disabled sysctl knob is set, ii) count the number of skipped tests from unprivileged, iii) when a test case had an unexpected error then print the actual but also the unexpected one for better comparison, from Joe. 2) Add a sample program for collecting CPU state statistics with regards to how long the CPU resides in cstate and pstate levels. Based on cpu_idle and cpu_frequency trace points, from Leo. 3) Various x64 BPF JIT optimizations to further shrink the generated image size in order to make it more icache friendly. When tested on the Cilium generated programs, image size reduced by approx 4-5% in best case mainly due to how LLVM emits unsigned 32 bit constants, from Daniel. 4) Improvements and fixes on the BPF sockmap sample programs: i) fix the sockmap's Makefile to include nlattr.o for libbpf, ii) detach the sock ops programs from the cgroup before exit, from Prashant. 5) Avoid including xdp.h in filter.h by just forward declaring the struct xdp_rxq_info in filter.h, from Jesper. 6) Fix the BPF kselftests Makefile for cgroup_helpers.c by only declaring it a dependency for test_dev_cgroup.c but not every other test case where it is not needed, from Jesper. 7) Adjust rlimit RLIMIT_MEMLOCK for test_tcpbpf_user selftest since the default is insufficient for creating the 'global_map' used in the corresponding BPF program, from Yonghong. 8) Likewise, for the xdp_redirect sample, Tushar ran into the same when invoking xdp_redirect and xdp_monitor at the same time, therefore in order to have the sample generically work bump the limit here, too. Fix from Tushar. 9) Avoid an unnecessary NULL check in BPF_CGROUP_RUN_PROG_INET_SOCK() since sk is always guaranteed to be non-NULL, from Yafang. ==================== Signed-off-by: David S. Miller <davem@davemloft.net>
2018-02-26samples/bpf: Add program for CPU state statisticsLeo Yan3-0/+504
CPU is active when have running tasks on it and CPUFreq governor can select different operating points (OPP) according to different workload; we use 'pstate' to present CPU state which have running tasks with one specific OPP. On the other hand, CPU is idle which only idle task on it, CPUIdle governor can select one specific idle state to power off hardware logics; we use 'cstate' to present CPU idle state. Based on trace events 'cpu_idle' and 'cpu_frequency' we can accomplish the duration statistics for every state. Every time when CPU enters into or exits from idle states, the trace event 'cpu_idle' is recorded; trace event 'cpu_frequency' records the event for CPU OPP changing, so it's easily to know how long time the CPU stays in the specified OPP, and the CPU must be not in any idle state. This patch is to utilize the mentioned trace events for pstate and cstate statistics. To achieve more accurate profiling data, the program uses below sequence to insure CPU running/idle time aren't missed: - Before profiling the user space program wakes up all CPUs for once, so can avoid to missing account time for CPU staying in idle state for long time; the program forces to set 'scaling_max_freq' to lowest frequency and then restore 'scaling_max_freq' to highest frequency, this can ensure the frequency to be set to lowest frequency and later after start to run workload the frequency can be easily to be changed to higher frequency; - User space program reads map data and update statistics for every 5s, so this is same with other sample bpf programs for avoiding big overload introduced by bpf program self; - When send signal to terminate program, the signal handler wakes up all CPUs, set lowest frequency and restore highest frequency to 'scaling_max_freq'; this is exactly same with the first step so avoid to missing account CPU pstate and cstate time during last stage. Finally it reports the latest statistics. The program has been tested on Hikey board with octa CA53 CPUs, below is one example for statistics result, the format mainly follows up Jesper Dangaard Brouer suggestion. Jesper reminds to 'get printf to pretty print with thousands separators use %' and setlocale(LC_NUMERIC, "en_US")', tried three different arm64 GCC toolchains (5.4.0 20160609, 6.2.1 20161016, 6.3.0 20170516) but all of them cannot support printf flag character %' on arm64 platform, so go back print number without grouping mode. CPU states statistics: state(ms) cstate-0 cstate-1 cstate-2 pstate-0 pstate-1 pstate-2 pstate-3 pstate-4 CPU-0 767 6111 111863 561 31 756 853 190 CPU-1 241 10606 107956 484 125 646 990 85 CPU-2 413 19721 98735 636 84 696 757 89 CPU-3 84 11711 79989 17516 909 4811 5773 341 CPU-4 152 19610 98229 444 53 649 708 1283 CPU-5 185 8781 108697 666 91 671 677 1365 CPU-6 157 21964 95825 581 67 566 684 1284 CPU-7 125 15238 102704 398 20 665 786 1197 Cc: Daniel Lezcano <daniel.lezcano@linaro.org> Cc: Vincent Guittot <vincent.guittot@linaro.org> Signed-off-by: Leo Yan <leo.yan@linaro.org> Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2018-02-23Merge branch 'x86-jit'Alexei Starovoitov2-87/+221
Daniel Borkmann says: ==================== Couple of minor improvements to the x64 JIT I had still around from pre merge window in order to shrink the image size further. Added test cases for kselftests too as well as running Cilium workloads on them w/o issues. ==================== Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2018-02-23bpf: add various jit test casesDaniel Borkmann1-0/+89
Add few test cases that check the rnu-time results under JIT. Signed-off-by: Daniel Borkmann <daniel@iogearbox.net> Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2018-02-23bpf, x64: save 5 bytes in prologue when ebpf insns came from cbpfDaniel Borkmann1-12/+16
While it's rather cumbersome to reduce prologue for cBPF->eBPF migrations wrt spill/fill for r15 which is callee saved register due to bpf_error path in bpf_jit.S that is both used by migrations as well as native eBPF, we can still trivially save 5 bytes in prologue for the former since tail calls can never be used there. cBPF->eBPF migrations also have their own custom prologue in BPF asm that xors A and X reg anyway, so it's fine we skip this here. Signed-off-by: Daniel Borkmann <daniel@iogearbox.net> Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2018-02-23bpf, x64: save few bytes when mul is in alu32Daniel Borkmann1-15/+28
Add a generic emit_mov_reg() helper in order to reuse it in BPF multiplication to load the src into rax, we can save a few bytes in alu32 while doing so. Signed-off-by: Daniel Borkmann <daniel@iogearbox.net> Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2018-02-23bpf, x64: save several bytes when mul dest is r0/r3 anywayDaniel Borkmann1-10/+11
Instead of unconditionally performing push/pop on rax/rdx in case of multiplication, we can save a few bytes in case of dest register being either BPF r0 (rax) or r3 (rdx) since the result is written in there anyway. Signed-off-by: Daniel Borkmann <daniel@iogearbox.net> Signed-off-by: Alexei Starovoitov <ast@kernel.org>